Crystallographic Investigations of Respiratory Proteins Citation Iverson, Tina Michelle (2000) Crystallographic Investigations of Respiratory Proteins. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/ftcb-ry72. https://resolver.caltech.edu/CaltechTHESIS:11252025-213630272 Abstract All organisms require a respiratory process to produce energy. In eukaryotes, this process occurs in the mitochondria, and requires a respiratory chain of four integral membrane proteins as well as a membrane-soluble quinone pool and cytochrome c. The respiratory proteins transfer electrons to oxygen as the terminal electron acceptor with the electron transfer coupled to the translocation of protons across the mitochondrial membrane. Not all organisms use oxygen as the terminal electron acceptor of their electron transport chain. One of the more common alternative electron acceptors is fumarate, but other common electron acceptors include nitrogen-containing compounds, the transformation of which represents an important step in the biological nitrogen cycle. This thesis discusses the structural investigations of proteins involved in diverse respiratory processes. The crystal structure of the Escherichia coli fumarate reductase, an integral-membrane enzyme complex involved in anaerobic respiration with fumarate as the terminal electron acceptor, has been solved. This structure both suggests the mechanism of the terminal step of anaerobic fumarate respiration and gives a model for the function of the homologous protein succinate dehydrogenase from mitochondrial respiration. The crystal structure of cytochrome c554 from the chemoautotrophic nitrifer Nitrosomonas europaea shows a heme-packing motif that may be important in respiratory pathways that require the simultaneous transfer of multiple electrons. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Biochemistry) Degree Grantor: California Institute of Technology Division: Biology Major Option: Biochemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Rees, Douglas C. Thesis Committee: Bjorkman, Pamela J. (chair) Rees, Douglas C. Gray, Harry B. Rothenberg, Ellen V. Defense Date: 5 May 2000 Record Number: CaltechTHESIS:11252025-213630272 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11252025-213630272 DOI: 10.7907/ftcb-ry72 ORCID: Author ORCID Iverson, Tina Michelle 0000-0001-8816-6352 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17775 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 02 Dec 2025 19:07 Last Modified: 02 Dec 2025 19:07 Thesis Files PDF - Final Version See Usage Policy. 81MB Repository Staff Only: item control page

Crystallographic Investigations of Respiratory Proteins Thesis by Tina Michelle Iverson In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, Califomia 2000 (Submitted May 5, 2000) 11 ©2000 Tina Michelle Iverson All Rights Reserved 111 Acknowledgements I'd like to thank: Doug for financial and intellectual support Glenn for emotional support Amy, Eric, Grandma and Grandpa And all the lab membtffs, collaborators, and others I've met along the way that have made this a truly fun experience. (i) (i) !V Abstract All organisms require a respiratory process to produce energy. In eukaryotes, this process occurs in the mitochondria, and requires a respiratory chain of four integral membrane proteins as well as a membrane-soluble quinone pool and cytochrome c. The respiratory proteins transfer electrons to oxygen as the terminal electron acceptor with the electron transfer coupled to the translocation of protons across the mitochondrial membrane. Not all organisms use oxygen as the terminal electron acceptor of their electron transport chain. One of the more common alternative electron acceptors is fumarate, but other common electron acceptors include nitrogen-containing compounds, the transformation of which represents an important step in the biological nitrogen cycle. This thesis discusses the structural investigations of proteins involved in diverse respiratory processes. The crystal structure of the Escherichia coli fumarate reductase, an integral-membrane enzyme complex involved in anaerobic respiration with fumarate as the terminal electron acceptor, has been solved. This structure both suggests the mechanism of the terminal step of anaerobic fumarate respiration and gives a model for the function of the homologous protein succinate dehydrogenase from mitochondrial respiration. The crrstal structure of cytochrome c554 from the chemoautotrophic nitrifer Nitrosomonas europaea shows a heme-packing motif that may be important in respiratory pathways that require the simultaneous transfer of multiple electrons. V Table of Contents Page Acknowledgements lll Abstract IV Table of Contents V List of Figures Vll List of Tables Xll Chapter 1 Mitochondrial Aerobic Respiration and Bacterial 1 Anaerobic Respiration with Fumarate as the Terminal Electron Acceptor Chapter 2 Overexpression, Purification and Crystallization 43 of the Membrane-bound Fumarate Reductase from Escherichia coli Chapter 3 Structure of the Escherichia coli Fumarate 77 Reductase Respiratory Complex Chapter 4 Analyzing your Complexes: Structures of the Quinol- 84 Fumarate Reductase Respiratory Complex Chapter 5 Biological Nitrification and the Nitrogen Cycle 115 Vl Chapter 6 Heme-packing Motifs Revealed by the Structure of 147 Cytochrome c554 from Nitrosomonas europaea Chapter 7 High-resolution Structures of the Oxidized and Reduced 156 States of Cytochrome c554 from Nitrosomonas europaea Chapter 8 A review of Cytochrome c554 182 Archaeal Methanogenesis and they-class Carbonic 212 Appendix A A-1 Anhydrase from Methanosarcina thermophila A-2 A Closer Look at the Active Site of y-class Carbonic 220 Anhydrases: High-Resolution Structures of the Carbonic Anhydrase from Methanosarcina thermophila A-3 Crystallographic Studies of they-class Carbonic 262 Anhydrase from Methanosarcina thermophila in Complex with inhibitors A-4 Crystallization and Preliminary X-ray Analysis of a ~-class Carbonic Anhydrase from Methanobacterium thermoautotrophicum LlH 279 Vll List of Figures Chapter 1: 1.1 Mitochondrial electron transport chain 19 1.2 Structural understanding of the mitochondrial respiratory chain 20 1.3 Topological predictions and cluster location of complex I 21 1.4 The pH dependent midpoint of cluster N2 21 1.5 Q-cycling by cytochrome bc 1 22 1.6 Movement of the Rieske iron-sulfur protein during Q-cycling 22 1.7 Proposed mechanism of hydroquinone oxidation at Q0 23 1.8 Stereoview of cofactors of complex IV and their ligands 24 1.9 The catalytic cycle of complex IV 24 1.10 Schematic drawing of the F 1F ATPase 25 1.11 Model for generation of rotation by proton transport through F 0 25 1.12 Binding change mechanism of the ATP synthase 26 1.13 Model for the control of the FNR transcription factor 26 1.14 Structures of the menaquinone and ubiquinone cofactors 27 1. 15 Schematic representation of anaerobic respiration with fumarate 27 1.16 Structural understanding of anaerobic respiration with fumarate 28 O Vlll Chapter 2: 2.1 Optimization of bacterial growth 63 2.2 SDS-PAGE analysis of E. coli enriched membranes and purified protein 64 2.3 Chromatograms of Thesit solubilized QFR 65 2.4 Phospholipid concentration and QFR activity during purification 66 2.5 Effect of guanidinium hydrochloride on tryptophan fluorescence 67 2.6 Crystals obtained from purified QFR 68 Chapter 3: 3.1 Structure of fumarate reductase 79 3.2 Proteins of the fumarate reductase complex 81 3.3 Cofactor location and pathway of electron transfer 81 3.4 Quinone binding pockets and active site residues 82 Chapter 4: 4.1 Ribbon diagrams of the E. coli and W. succinogenes QFR complexes 99 4.2 Distances between cofactors in the QFR complexes 100 4.3 Stereoview illustrating movement of the capping domain 101 4.4 Schematic diagram of active site interactions and proposed mechanism 102 4.5 Stereoview illustrating active site movement between fumarate reductases 103 lX 4.6 Alignment of the transmembrane helices of the QFR complexes 104 4.7 Comparison of QFR Qp to cytochrome bc 1 Q 1 105 Chapter 5: 5.1 Enzymatic components of the biological nitrogen cycle 131 5.2 The cofactors of the nitrogenase system 132 5.3 Heme-packing motifs 133 5.4 Molybdenum-containing cofactors 134 5.5 Stereoview of the siroheme of sulfite reductase 135 5.6 Ligand switching to the c-heme of the cd 1-nitrite reductase 136 5.7 Movement of Tyr 25 during catalysis 137 5.8 Stereoview of the Cuz center of nitrous oxide reductase 138 Chapter 6: 6.1 The oxidation of NH 3 to No 2- by Nitrosomonas europaea 148 6.2 The fold and secondary structure of cyt c554 149 6.3 Heme configuration and superposition of heme-packing motifs 150 6.4 Unusual aspects of the structure of cyt c554 151 6.5 GRASP electrostatic surfaces and docking to HAO 152 X Chapter 7: 7.1 RMS deviation between crystal forms and conformational change 173 7.2 Stereoview of the conformational change associated with reduction 175 7.3 Conserved water molecules in heme-packing motifs 175 7.4 Disallowed angles between His 129 and Phe 130 176 Chapter 8: 8.3D The fold of cytochrome c554 201 8.1 The biological nitrogen cycle 201 8.2 Biological nitrification catalyzed by Nitrosomonas europaea 202 8.3 Heme organization and interheme iron distances 202 8.4 No ligaiton of heme I 203 8.5 Parallel stacked heme pairs 204 8.6 Alignment of the heme cores of cyt c554, HAO, and cyt c NR 204 8.7 The His129-Phe130 peptide bond 205 8.8 Conformational change accompanying reduction in cyt c554 206 8.9 Electrostatic representation and docking model of cyt c554 and HAO 206 Xl Appendix A-1: A-1.1 Microbial food chain 216 A-1.2 Pathway for conversion of acetate to methane by methanoarchaea 216 A-1.3 Methanogenic fermentation of acetate by M. thermophila 217 Appendix A-2: A-2.1 Stereoview of the Cam trimer 247 A-2.2 Active site coordination in Cam 248 A-2.3 Stereoview of catalytically important residues 249 A-2.4 Stereoview of electrostatic surface potential representation 250 A-2.5 Proposed reaction mechanism 252 Appendix A-3: A-3. l Binding of inhibitors to the active site of Cam 274 Appendix A-4: A-4.1 Crystals of ~-carbonic anhydrase (CAB) 287 A-4.2 Typical diffraction from CAB 287 A-4.3 Rotation function from CAB data suggests a tetrameric protein 288 xii List of Tables Chapter 1: 1.1 Midpoint potentials of Bos taurus complex II 28 1.2 Oxido-reductases of the anaerobic respiratory chains of E. coli 29 Chapter 2: 2.1 Purification of the QFR complex from Escherichia coli 69 2.2 Effect of detergents on enzyme activity and fluorescence emission 70 Chapter 3: 3.1 Summary of data collection and refinement statistics 80 Chapter 6: 5.1 Summary of crystallographic statistics 153 5.2 Interheme angles and distances 154 Chapter 7: 7.1 Data collection and refinement statistics 177 7.2 Heme iron ligation distances 178 xiii Appendix 2: A-2.1 Data collection and refinement statistics 253 A-2.2 Summary of metal ligation in Cam 254 A-2.3 Active site geometry in Zn-Cam 255 A-2.4 Active site geometry of Co-Cam 255 A-2.5 Active site geometry of Zn-Cam+ HCO 3- 256 A-2.6 Active site geometry of Co-Cam+ HCO 3- 256 A-2.7 Active site geometry of Zn-Cam+ SO/ 257 A-2.8 Active site geometry of Co-Cam+ SO/ · 257 Appendix 3: A-3.1 Data collection and refinement statistics of Cam-inhibitor complexes 275 Appendix 4: A-4.1 Data collection statistics for ~-carbonic anhydrase 289 1 Chapter 1: Mitochondrial Aerobic Respiration and Bacterial Anaerobic Respiration with Fumarate as the Terminal Electron Acceptor Mitochondrial Aerobic Respiration Almost four billion years ago, some unicellular organisms developed the ability to harvest light energy from the sun (I). As water is exceedingly stable and therefore plentiful, the primitive photosynthetic organisms evolved mechanisms to oxidize water to gaseous 0 2. The evolutionary consequences 9f oxygen-evolving photosynthesis were tremendous, as a reducing atmosphere became oxidizing, and strategies for dealing with a new, reactive environment had to be developed (2). Unlike water, oxygen is capable of participating in catastrophic reactions that can lead to disruption of cellular processes and cell death. On the other hand, the appearance of a significant amount of 0 2 in the atmosphere must have exerted a very strong evolutionary pressure favoring organisms able to utilize 0 2 as an oxidant of foodstuffs , as much more energy can be released with 0 2 than with less effective oxidants. Dioxygen has been harnessed effectively as the terminal oxidant in aerobic respiration. One of the cornerstones of biology involves understanding oxidative phosphorylation, the process that couples generation of adenine triphosphate (ATP) to oxygen reduction through a transmembrane proton gradient. In mammalian aerobic respiration, this process occurs at the inner membrane of the mitochondria and involves four transmembrane respiratory complexes (complexes I - IV), the soluble protein 2 cytochrome c and membrane soluble ubiquinone molecules, as well as the ATP synthase (complex V; Fig. 1.1). The electron transfer steps catalyzed by complexes I-IV each harvest energy by coupling the electron transfer to cofactors of increasingly higher reduction potential to the formation of a proton gradient across the membrane. The electron transfer requires a variety of redox-active cofactors associated with each complex. As endergonic tunneling of electrons occurs at physiologically relevant rates when the cofactor separation is < ~ 14 A, each of these large respiratory complexes requires multiple redox-active centers for efficient transfer of the electron from the cytoplasm to the membrane. All of the respiratory complexes are large, transmembrane molecules. One theme evident in the organization of these complexes is that they are often modular structures arising from the conjunction of different proteins during evolution. This modular complex formation may be an evolutionarily efficient mechanism for an organism to respond to the various chemicals that can be used as electron donors and acceptors in respiration. Ferredoxin folds are commonly seen in proteins joining a catalytic domain and a transmembrane domain. Additionally, the catalytic domain often exhibits folds seen in other proteins. Complex I Complex I (NADH-ubiquinone oxidoreductase; NUO) represents the largest, most complicated, and least understood of all the respiratory enzymes (3-5). The oxidation of NADH by this enzyme couples the transfer of two electrons to the ubiquinone pool to the vectorial translocation of 4H+ (6-9) across the mitochondrial membrane. Bovine mitochondrial complex I is composed of at least 43 different 3 subunits, (J 0-13) seven of which are mitochondrially encoded (14 ), while bacterial homologues of complex I contain 13 to 14 subunits (15-18) that form the functional core of the enzyme. Electron microscopy of complex I from Escherichia coli and Neurospora crassa shows that complex I forms an L-shaped overall structure ( 19, 20) (Fig. 1.2). This structure can be divided into two major parts: the membrane arm that extends 180 A within the membrane and the peripheral arm that protrudes 80 A from the mitochondrial membrane. The peripheral arm can be biochemically divided into two subcomplexes: the flavoprotein (Fp) and the iron protein (Ip) (21 ). The flavoprotein is composed of three subunits of molecular weight 51 kD, 24 kD and 10 kD and contains noncovalently-associated flavin mononucleotide (FMN) as well as several [Fe:S] clusters. The Ip subcomplex contains six subunits of molecular weights 75 kD, 49 kD, 30 kD, 18 kD, 15 kD, and 13 kD as well as nonheme iron and acid-labile sulfur, suggesting the presence of iron-sulfur clusters. The sequence similarity between the Fp of complex I and the NAD-reducing hydrogenase from Alcalugenes eutrophus (22), the sequence similarity between the Ip and ferredoxins, and the ability to resolve complex I into several independently folding subcomplexes (21 ), suggest complex I has a modular structure resulting from its evolutionary origins. Respiratory complex II, hydrogenase and formate dehydrogenase, described below, similarly appear to have evolved from smaller subcomponents; thus, modular complex formation represents a theme in the evolution of many respiratory enzymes. Associated with the peripheral arm of complex I is at least one flavin mononucleotide (FMN) (23) that likely forms the active site for NADH oxidation. Until the recent discovery of spectroscopic signals arising from two [4Fe:4S] clusters associated with a truncated form of the complex I homologue from Paracoccus denitrificians (24), electron paramagnetic resonance (EPR) had shown the presence of 4 only six distinct [Fe:S] clusters associated with complex I (25-28) despite the presence of eight consensus [Fe:S] cluster sequences (11 , 15-17). All the [Fe:S] clusters are believed to be associated with the peripheral arm of the complex or lie at the interface between the peripheral arm and the membrane arm. Standard nomenclature defined by Ohnishi (29) labels the six well-studied [Fe:S] clusters Nla, Nlb, N2, N3, N4 and N5 in order of increasing spin relaxation rates ; the two recently discovered [4Fe:4S] clusters have not yet been assigned specific nomenclature. Of the [Fe:S] clusters that have been identified, two (Nla, Nlb) have EPR spectra characteristic of [2Fe:2S] clusters, three (N3, N4, N5) exhibit classic [4Fe:4S] spectra, and one (N2) is believed to be a [4Fe:4S] with pH dependent EPR spectra. Associated with the membrane arm of complex I are at least two ubiquinonebinding sites that have been identified by the presence of two distinct semiquinone signals in the EPR spectra (30, 31) that are quenched by different quinone-binding inhibitors (32-34). The presence of additional quinone binding sites in complex I is debated ( 35 ). A model for the functional core of complex I, including the putative cofactor locations, is shown in Fig. 1.3. Theories of proton translocation by complex I cannot fully explain the experimentally observed proton to electron stoichiometry of 4H+12e- (21, 33, 36). However, proton pumping may include a combination of vectorial translocation involving the N2 [4Fe:4S] cluster and a modification of the quinone-cycling (Q-cycling) mechanism ( 37) described in detail below for complex III, as well as an additional undetermined mechanism. Cluster N2 has been proposed to play an important role for energy conservation in complex I (29, 37) since it exhibits several unique properties. First, it has the highest Em value of all the clusters that have been observed so far in 5 complex I (25, 38) suggesting it is the last [Fe:S] cluster to be reduced before the electrons are passed to the quinones. Second, its one electron reduction or oxidation is coupled with the binding and release of one proton in the physiological pH range (Fig. 1.4) (3, 25, 38-40) . Third, the location of the cluster N2 is likely near the interface between the peripheral arm (24) and the membrane arm. Both EPR interactions showing a distance of < 11 A from one of the semiquinone species (31, 41) and the observation that delipidation of purified complex I affects the reduction potential of cluster N2 (42) suggest this interface position. A location at the membrane interface allows cluster N2 to be optimally positioned for involvement in proton translocation. Complex II Complex II (succinate-quinone oxidoreductase; SQR) oxidizes succinate to fumarate at a flavin adenine dinucleotide (FAD) and transports the electrons generated from this process into the ubiquinone pool (43, 44 ). By catalyzing oxidation of succinate and participating in electron transfer, complex II acts as important member of both the Krebs cycle and the mitochondrial aerobic respiratory chain (Figures 1. 1; 1.2 (45 )). Complex II is organized into four polypeptide chains (43 ): the flavoprotein (Mr -60 kD) contains the covalently-linked active site FAD; the iron protein (Mr -25kD) covalently links one [2Fe:2S] (46), one [3Fe:4S] (47), and one [4Fe:4S] center (48); and the membrane anchors (Mr -13 kD and -11 kD) bind one b-type heme and are associated with ubiquinone molecules. Like complex I, sequence and biochemical analyses of complex II suggest the structure has evolved as a conjunction of modules (49 ). Indeed, complex II can be 6 biochemically resolved into a membrane-spanning domain and a soluble domain; the soluble domain containing the flavoprotein and the iron protein retains catalytic activity in succinate oxidation under appropriate conditions (50) . The flavoprotein exhibits significant sequence similarity to soluble flavocytochrome c3 fumarate reductases (5153 ), L-aspartate oxidase (54-60) , a protein involved in the biosynthesis of NADH, 3- ketosteroid-delta-1-dehydrogenase (61-63) a protein involved in the first step of steroid catabolism, and an osmotic regulatory protein from yeast (64, 65). Similarly, the Nterminus of the iron protein has sequence similar to [2Fe:2S] plant-type ferredoxins, while the C-terminus contains [Fe:S] binding cysteine motifs identical to those seen in bacterial ferredoxins (43 ). All of the flavin and iron-sulfur cluster prosthetic groups of complex II are located in the soluble domain of the complex. The reduction potentials of each prosthetic group are tuned to catalyze the oxidation of succinate and to facilitate the transfer of electrons into the ubiquinone molecules associated with the membrane anchors (Table 1. 1). However, under appropriate conditions the enzyme can also act in the reduction of fumarate (66, 67). Because of the redox couple between succinate and fumarate (+25 mV), electron transfer to a non-covalently attached FAD (<-200 mV) would be extremely unfavorable. Site-directed mutagenesis altering the histidine that forms the covalent linkage to the FAD results in an enzyme that cannot catalyze the oxidation of succinate (68), but can function in the reduction of fumarate (69), which is opposite of the physiological direction of the reaction in mitochondria. The 8cx-FAD covalent attachment to the NE of a histidine (70) raises the reduction potential to -79 mV. Transfer of an electron from succinate to the FAD still appears somewhat unfavorable; however this energy barrier can be overcome during reaction turnover. By analogy, chemical 7 reactions can overcome the barrier of high energy intermediates during the formation of lower energy products. Based on a sequence identity of ~40% between the soluble domain of complex II and the soluble domain of bacterial anaerobic homologues of complex II, the fold of the soluble domain should be nearly identical between the two complexes . Recent crystallographic analysis of the anaerobic homologue of complex II (71, 72) has confirmed the modular nature of the enzyme and the similar fold of the flavoprotein and soluble homologues (73-76) as well as the iron protein and bacterial ferredoxins . In contrast, the membrane transmembrane anchors between complex II and bacterial homologues have virtually no sequence similarity. However, both complexes are believed to associate with quinone molecules in a similar fashion (77). Complex II is predicted to be associated with at least two quinone molecules situated on opposite sides of the membrane (78-81) in a fashion reminiscent of that utilized for the Q-cycle mechanism of complex III ( 82- 84 ). Recent crystallographic studies of the anaerobic complex II homologue from E. coli (71) confi rm the transmembrane positions of two quinone molecules. Despite this suggestive arrangement of quinones, it has long been held that complex II is not energy coupled. Calculations suggest that the energetics of proton translocation by complex II would be unfavorable during aerobic respiration (85). Complex III Complex III (cytochrome be,; ubiquinol:cytochrome c oxidoreductase) accepts electrons from the reduced ubiquinol pool and transfers them to the soluble cytochrome c with the concomitant generation of a proton gradient across the mitochondrial membrane. Complex III is composed of a functional core containing cytochrome b, cytochrome c,, 8 and the Rieske [2Fe:2S] iron-sulfur protein. Essential for the function of the complex are the cofactors associated with these subunits: two b-type hemes (bL and bH); one c-type heme; the Rieske [2Fe:2S] center; and the binding sites for two ubiquinone molecules (86, 87). Mitochondrial complex III contains additional subunits not present in bacterial complexes that may be involved in mitochrondrial targeting and processing ( 88-90 ). Based on the organization of cofactors seen in the crystal structure as well as the buried surface area between monomers (82-84), complex III functions as a physiological dimer (Fig. 1.2). The protonmotive quinone cycle (Q-cycle) originally proposed by Peter Mitchell (91) best explains the experimental results on electron transfer through the redox centers of complex III. The Q-cycle requires two separate quinone binding sites (Q 0 and Q;) positioned on opposite sides of the membrane with a bifurcated electron flow from the Q0 quinone bound near the intermembrane space (positive side of the membrane; Fig. 1.5). The unique bifurcated oxidation of ubiquinol at Q0 of complex III is the reaction within the Q-cycle reaction scheme that is most critical for the link between electron transfer and vectorial proton translocation (92). Bifurcated electron flow occurs during hydroquinone oxidation at the Q0 site (93) , with one electron transferred through a high-potentialchain involving the Rieske [2Fe:2S] cluster and heme c 1 to cytochrome c, while the second electron is transferred by the two b-type hemes across the membrane to reduce the quinone bound to the Qi site (94-98). Oxidation of reduced quinone (hydroquinone; QH 2) to ubiquinol (Q) requires the release of the two protons that are transferred to the intermembrane space. Likewise, the reduction of quinone in the Qi site occurs with the uptake of two protons from the matrix. The resultant H+/e- stoichiometry of the Q-cycle is 2H+ translocated for each electron that reduces cytochrome c. 9 While there is a general consensus about the overall reaction at Q0 , the nature of the intermediates and the way the reaction is controlled to ensure obligatory bifurcation is still controversial (99) . Crystallographic studies of complex III (83, 84, 100) suggest that the bifurcated pathway of electron transfer is associated with physical movement of the Rieske iron-sulfur protein (101, 102). A hinge rotation around two ~-strands combined with the isomerization of Pro 175 ( 84) can shift the location of the Rieske [2Fe:2S] cluster over 15 A (83, 84 ). In one stabilized position, the cluster lies in close contact to the Q0 site, in a second position, the cluster is proximal to heme c 1, and in a third position the Rieske iron-sulfur protein is found at an intermediate location (Fig. 1.6). In addition to the movement by the Rieske iron-sulfur protein, the binding sites stabilizing the quinones are critical for function of complex III. The Qi site must preferentially bind fully oxidized quinone and stabilize a ubisemiquinone species after the transfer of one electron from the Q0 site. After reduction by the second electron and the uptake of two protons, binding to the hydroquinone species must weaken such that it is released back into the Q-pool. Several mechanisms have been suggested to explain the bifurcated electron flow from the Q0 site (103-106). The proposal by Brandt and von Jagow (104, 105) of a "catalytic switch" mechanism perhaps agrees best with the recent crystallographic data. Crystallographic studies of complex III have not yet shown the presence of a quinone in the Q0 site, but its location has been mapped by inhibitors . The inhibitor binding sites show a bi-lobed pocket for quinone binding (107, 108), and the position of the Rieske iron-sulfur protein is dependent upon the lobe in which the inhibitor binds ( 107) . This would suggest that hydroquinone would bind at Q0 triggering the conformation of the Rieske iron-sulfur protein to shift to a location proximal to the quinone (97). Before QH2 can be oxidized by the Reiske iron-sulfur cluster, it must first be deprotonated to Qff, which represents the activation barrier of hydroquinone oxidation (I 09 ). This is believed to occur through the conserved His 161 and Glu 272 residues that make hydrogenbonding contacts to polar groups of inhibitor molecules (100, 110 ). Upon oxidation, the newly formed ubisemiquinone shifts into a second site (97), concomitantly pushing the Rieske protein away to interact with cytochrome c 1 (Fig. 1.7). Qff is deprotonated to a ubisemiquinone species (SQ•) before electron transfer fully oxidizes the quinone by the low potential pathway. The time scale of the movement of the Rieske iron sulfur protein may be slow enough to prevent both of the electrons from being transferred from the Q0 site by the high-potential pathway, but is not thought to be a rate-limiting step in the Qcycle ( 111 ). Rather, the reaction rate is limited by the ubihydroquinone deprotonation ( 109) or the stability of the semiquinone intermediate (97). Complex JV Complex IV (cytochrome c oxidase; CcO) catalyzes the final step in the electron transport chain (112-115) by accepting electrons from the soluble cytochrome c and using them to reduce molecular dioxygen to water, with the concomitant vectorial translocation of four protons (116-119), such that the overall reaction is: where H\ 0 and H+ out indicate protons that are taken up and released vectorially with respect to the respiratory membrane. Complex IV, an aa 3 -type cytochrome c oxidase, resides in a superfamily of terminal oxidases (120, 121) including the cytochrome bo3 11 complexes and the aa 3-type cytochrome c oxidases, that require similar cofactors for reaction turnover. Complex IV from all organisms is comprised of at least two integralmembrane subunits (122, 123) that contain a di-nuclear copper center (Cu A) and a lowspin six-coordinate a -type heme (heme a) that mediate electron input to a heme-copper dioxygen activating and reduction center (heme a3 - Cu 8 ). Crystal structures of bacterial (124, 125) and mitochondrial (126-128) complex IV (Fig. 1.2) have shown that the CuA is located in core subunit II while the rest of the cofactors known to be essential for catalytic activity are located in core subunit I (129) . This functional core appears capable of independently catalyzing the reduction of molecular oxygen in bacterial systems. The peripheral subunits of mitochondrial complex IV form binding sites for zinc, sodium, and magnesium (126-128, 130) and may be necessary for assembly of the complex ( 131) or for allosteric regulation of activity by ATP (127, 132, 133). Site directed mutagenesis of the side chains that bind the peripheral metals (134, 135) suggest they are necessary for catalysis in the mitochondria ( 136) , but it is possible that these metals merely serve a structural role. Electrons enter complex IV from reduced cytochrome c, which binds above CuA and heme a (137-139), and proceed to the active site. The active site of complex IV, often called the binuclear center, is composed of heme a 3 -Cu 8 . Heme a3 is a fivecoordinate heme with the open ligation site suggested to be the dioxygen binding site. Cu 8 lies 4.9 A away from the heme iron in the unliganded form of the enzyme and 5.3 A away if the enzyme is bound to molecular dioxygen, cyanide, or azide. The change in distance is associated with movement of the Cu 8 center (128), while the heme position remains fixed, regardless of the oxidation state of the enzyme. Cu 8 forms a bond to the substrate or substrate analog bound to the heme iron ( 124, 128), changing the coordination from trigonal planar to square planar. Near the active site, a catalytically 12 essential tyrosine (140) (Tyr 244 using bovine numbering) appears to make a covalent bond to one of the histidine ligands (His 240) to the CuB center (Fig. 1.8) in both the bacterial (125) and mitochondrial complexes (128) . This covalent bond may raise the pK. of the tyrosine hydroxyl to a value that would function in proton shuttling to the oxygen species bound at the active site. The catalytic cycle of complex IV appears to have multiple intermediates (Fig. 1.9). In the first step, the oxidized enzyme accepts two electrons and pumps two protons. The second step binds molecular oxygen to the active site at the binuclear center. One oxygen atom forms a double bond to the iron of heme a 3 in what is historically known as the peroxy (P) intermediate. The other oxygen atom associates with CuB. During the peroxy intermediate, Tyr 244 donates a proton to CuB forming a tyrosyl radical (•0-Tyr) and a CuB bound to -OH. It is thought that the spectral signals arising from the tyrosyl radical have mistakenly caused the interpretation of this state being a peroxy radical. Coupled to the uptake of one electron and the pumping of one proton, the next step in the catalytic cycle, called ferry (F) reprotonates Tyr 244, such that both CuB and Tyr 244 are both protonated, and the iron is in a ferroxy state. With the pumping of a final proton and reduction by a fourth electron, the oxidative (0) state is formed. In this state, all of the catalytically active species are bound to hydroxide. The correlation of proton translocation to the catalytic cycle is highly controversial. Recent electrometric measurements suggest that protons may be pumped in other phases of the catalytic cycle (141). No channel providing access for the substrate molecular oxygen has been observed in the crystal structure, as is the case for hemoglobin and myoglobin (142). In analogy to these globins, the conformation for an open 0 2 channel may be attained 13 through transient conformational changes . A candidate that could serve as a water channel for the removal of the product may lie between subunit I and II (127). Residues lining this channel have been shown by site-directed mutagenesis to be necessary for enzyme function and are highly conserved. Complex IV is not associated with quinones, excluding a Q-cycling mechanism for proton pumping. Thus channels for proton uptake and translocation are essential for enzyme function. It is proposed that two separate proton channels are present (143, 144 ). One pathway delivers protons from the negative side of the membrane to the active site to combine with the reduced oxygen substrate. A second pathway is utilized for vectorial proton translocation and requires a full pathway through the complex. Based on the structure of complex IV (Fig. 1.2) three compelling candidates for the proton channels have been proposed ( 127) . Two of these candidates, dubbed the H-channel and Dchannel for the conserved external residues lining the channel (His and Asp respectively) have been suggested as candidates for vectorial proton translocation. Site-directed mutagenesis shows the D-channel is absolutely required for activity (145 ), making it a stronger candidate for proton translocation. The third channel, dubbed the K-channel for the lysine residue positioned at its exterior, likely delivers protons into the bi-nuclear active site. Despite the seemingly few possible proton transfer routes, no consensus on the mechanism of proton translocation through complex IV has been reached, and currently, different models are intensely debated (123, 141, 145-149). Complex V Complex V (F0 F 1 ATP synthase) catalyzes the formation of ATP from the precursor components adenine diphosphate (ADP) and inorganic phosphate (Pi) (150). 14 The complex is formed from a soluble domain (F 1) and a membrane-spanning domain (F0 ; Fig. 1.2). The functional core of F0 portion consists of three types of subunits in bacteria: a, b, and c. It is believed that one a-subunit, two b-subunits, and ten c-subunits are present ( 151 ). In contrast, the more complex mitochondrial F0 has additional subunits. The F 1 part of the ATP synthase consists of five types of subunits: a, p, y, 8, and £. While there are three each of a and p, there is only one of each of the others. Despite the sequence similarity between the a and p subunits, ATP synthesis only occurs in the p subunits. The most detailed studies of ATP synthesis have highlighted the F 1 part of the complex (152, 153 ). It is now widely accepted that the asymmetric y, 8, and £ subunits rotate in a cylinder formed of alternating a and p subunits (152) (Fig. 1.10). This rotational catalysis has been recently confirmed by visualization of the F 1 subunit rotating an actin filament (154-156) . The transmembrane proton gradient achieved by the oxidation-reduction reactions catalyzed by complexes I-IV in the electron transport chain drives the rotation of the F 1 subunit (91) . The membrane-spanning F0 subunit forms a proton channel between the asubunit and the ring of ten c-subunits. Each c-subunit is composed of two membranespanning helices (157). Protons are delivered to a conserved carboxylate side chain (Asp 61 in E. coli and Glu 59 in S. cerevisiae) by subunit a, resulting in a neutral state of the side chain that is stable in the center of the membrane. The uptake of each proton and concomitant release of a proton on the other side of the membrane adds torque to the central y-subunit (Fig. 1.11 ). After three to four protons have been passed back along their gradient, enough torque has built up on the y-subunit to force its rotation within the a and p subunits, causing a change in conformation in this a/p subunit cylinder (158). This conformational change sterically forces ADP and Pi together in the binding pocket of the p subunit to form ATP (Fig. 1.12) in what is referred to as the binding change 15 mechanism (159 ). This binding change mechanism facilitates release of the tightly bound ATP after formation. Bacterial Anaerobic Respiration with Fumarate as the Terminal Electron Acceptor Not all organisms require oxygen to produce ATP (160). In facultative anaerobes, like E. coli, the switch from aerobic respiration to anaerobic respiration requires the expression of a variety of proteins dedicated to anaerobic respiration and the repression of proteins involved in aerobic respiration. In the 1970s, the characterization of mutants that failed to carry out anaerobic respiration by fumarate and nitrate reduction were designated FNR mutants (161) . Since then, it has been determined that FNR is a single protein that is a global regulator responsible for controlling the anaerobic-aerobic regulation of over 120 target genes in E. coli (162). Under anaerobic conditions, FNR exists as a homodimeric transcription factor with each monomer containing a [4Fe:4S] cluster (163, 164) (Fig. 1.13). A growth shift from anaerobic to aerobic conditions causes the [4Fe:4S] clusters to disassemble rapidly to form two [2Fe:2S] clusters (165, 166). The loss of the [4Fe:4S] clusters converts FNR from a homodimer with high DNAbinding affinity to a monomer with low DNA-binding affinity. One important switch that occurs during the transition from aerobic to anaerobic respiration involves the identity of the quinone acting as the membrane-soluble electron carrier between respiratory complexes. Aerobic respiration usually utilizes ubiquinone, while menaquinone molecules mediate most anaerobic respiration (Fig. 1.14). This switch results in the use of quinone molecules that exhibit midpoint potentials less than that of the terminal electron acceptor, facilitating the directional electron transfer. 16 Ubiquinone has a resting potential of+70 mV, while menaquinone has a potential of -74 mV. Although the potentials may be modulated when bound to the protein complexes, the resting potential generally indicate the direction the reaction will proceed in vivo . Many different primary electron donors and terminal electron acceptors can be used in the absence of oxygen ( 167)(Table 1.2). One of the more common terminal electron acceptors in anaerobic respiration is fumarate (43, 44 ), which is reduced to succinate by the quinol-fumarate reductase (QFR) complex located in the periplasmic membrane. The electrons required for fumarate reduction are supplied by a reduced menaquinol pool. A membrane-bound hydrogenase and formate dehydrogenase (168170) can supply all of the reducing equivalents necessary in both W. succinogenes and E. coli and represents one of the simplest systems for anaerobic respiration (Fig. 1. 15). During this type of anaerobic respiration, the transmembrane proton gradient used to drive ATP synthesis is generated by selective catalysis of reactions on opposite sides of the membrane, where formate dehydrogenase and hydrogenase release protons on one side of the membrane (171 ), while QFR utilizes protons on the other. Hydrogenase The capacity to oxidize molecular hydrogen is a central metabolic feature of a wide variety of microorganisms (172-177). Hydrogenases can be classified into four types according to metal composition (178): the iron-only, or Fe-hydrogenases; Ni-Fe hydrogenases; Ni-Fe-Se hydrogenases; and one enzyme that does not contain any bound transition metals. The membrane-bound Ni-Fe hydrogenases act in anaerobic respiration when fumarate is the terminal electron acceptor and contain three redox active iron-sulfur -clusters, two [4Fe:4S] clusters and one [3Fe:4S] cluster, as well the Ni-Fe binuclear 17 center at the active site (Fig. 1.16). A high-resolution crystal structure of the purified soluble domain of the Ni-Fe hydrogenase from Desulfovibrio gigas (179) shows the NiFe binuclear center as ligated to the protein by four cysteine ligands, two to the nickel, and two that bridge the nickel and iron; three diatomic ligands stabilize the iron. Additionally, the binuclear center is bridged by an unassigned mononuclear ligand (180), the position of which is believed to be the binding site for hydrogen during reaction turnover. The periplasmically-located soluble domain of the Ni-Fe hydrogenase contains two subunits of molecular weights 63 kD and 29 kD, with the active site Ni-Fe center covalently bound to the large subunit. The small subunit of the Ni-Fe hydrogenase contains the ligands to the [Fe:S] clusters that form a linear chain for electron transfer ( 180 ). Hydrophobic analysis and sequence similarity to the hydrophobic membrane anchors of complex I suggest six different proteins are capable of forming the membraneanchors in E. coli ( 170). Since E. coli contains genes for three isozymes of hydrogenase, it is thought that each isozyme is exclusively associated with two membrane anchors. However, the properties exhibited by the membrane anchors have not been well characterized. Formate dehydrogenase Formate dehydrogenase oxidizes formate to carbon dioxide at a molybdopterin guanine dinucleotide (MGD) cofactor (181 , 182). The fold of the 79 kD soluble subunit contains four domains ( 183) (Fig. 1.16) and exhibits high structural similarity to DMSO reductase ( 184, 185 ), which also contains an MGD cofactor and can act as a terminal electron acceptor during anaerobic respiration in the presence of DMSO ( 186). 18 Interestingly, a nonsense (UGA) codon is present at a position corresponding to amino acid 140 of the polypeptide ( 187) . This codon is required for cotranslational incorporation of selenocysteine into formate dehydrogenase (188). The crystal structure of the soluble domain ( 183) showed that this selenocysteine acts as a ligand to the molybdenum in the MGD cofactor. The gene cluster for formate dehydrogenase additionally contains two peripheral proteins that are believed to be responsible for membrane association. These two genes encode an [Fe:S] protein and a cytochrome b like polypeptide. Quinol-fumarate reductase During anaerobic respiration with fumarate as the terminal electron acceptor, the quinol-fumarate reductase (QFR) complex is expressed and catalyzes the reduction of fumarate to succinate, which is opposite of the reaction catalyzed by complex II during aerobic respiration. QFR and complex II exhibit significant sequence similarity, thus the structure of either complex provides information about both aerobic and anaerobic respiration. The structure of the Escherichia coli QFR respiratory complex has been solved at 3.3 A resolution ( 69) using the method of mulitwavelength anomalous dispersion. Analysis of this structure has revealed insights into the formation of the complex, catalysis, and the potential role this enzyme might play in energy coupling. Structural aspects of this complex as well as mechanistic implications for both aerobic and anaerobic respiration will be discussed in detail in the next three chapters. 19 Fig. 1.1: Mitochondrial electron transport chain. The electron transport chain is initiated at either complex I (NADH ubiquinone oxidoreductase; NUO) or complex II (succinatequinone oxidoreductase; SQR), which reduce the ubiquinone pool. Electron transport by complex I is coupled to proton translocation across the mitochondrial membrane. Reduced ubiquinol then binds to complex III (cytochrome be,), which uses a Q-cycling mechanism to translocate protons across the mitochondrial membrane. Electrons are transferred from complex III to complex IV (cytochrome c oxidase; CcO) with the soluble cytochrome c as the intermediate electron carrier. Complex IV catalyzes the final step in respiration by transferring electrons to molecular dioxygen, thus reducing it to water. Fig. adapted from (83 ). l outer membrane inner membran folded into • 0.5 µm l the matrix space the terminal stages of oxidation occur at the inner membrane ~Fum Succ H+ Complex I Complex II Complex IV Complex V 20 Fig. 1.2: The structural understanding the mitochondrial respiratory chain. Complexes I, III, and IV pump protons out of the mitochondrial matrix to form a transmembrane proton gradient. Following this, complex V (the ATP synthase) couples the favorable process of bringing a proton back along its gradient with the unfavorable process of ATP synthesis. Complex I has been solved to ~35 A resolution by electron microscopy and exhibits an L-shaped structure both in mitochondria and bacteria. The remaining complexes have been solved to atomic resolution by x-ray diffraction. Only bacterial homologues of complex II have been solved (Escherichia coli and Woline/la succinogenes). Complex III has been solved from the mitochondria (bovine and chicken). Complex IV has been solved from both bacterial and mitochondrial sources (Paracoccus denitrificians and bovine). Complex V has been solved from yeast mitochondria and the F 1 portion has been solved from bovine. In all cases, the structures have revealed insights into the mechanism of function by the respiratory complex. The structures that have been solved from multiple sources have agreed in the overall fold, but yielded complementary insight into function. ATP synthase Respiratory chain complexes complex I complex II complex Ill complex IV succinate fumarate F; ATP synthase F0 ATP synthase cytochrome c 21 Fig. 1.3: Topological prediction and cluster location of complex I. Only the functional core of the enzyme is shown with the subunits labeled according to the nomenclature used for the E. coli enzyme. NuoE : c r~:o:o NuoF ~,. o ~ N3 NuoG <:e1 b f/tN4 ~ NS NuoCD ~ 7 ~ Nuol ? NuoB N2? (-) ~ Fig. 1.4: The pH dependent midpoint of cluster N2 in Oxidation (Emin mV) complex I. It is believed that cluster N2 is protonated [W]N2 0 x I upon reduction and may play a role in proton translocation. The limits for the pKA and ~ values were estimated from experimental data given in (36). CJ > -100 [W]N2red > 851 N2red <6 <-250 N2ox Cl> -0 a 0 ::, a 5· ::, 'o ;,:: ~ 22 Fig. 1.5: Q-cycling by cytochrome bc 1 requires bifurcated electron flow at the Q0 Rieske ISP 2H+ binding site. As the quinone changes its oxidation state, protons must disassociate (when quinone is oxidized) or associate (when quinone is reduced) to form a stable molecule. The quinones found at the (-) membrane interface will acquire or deposit protons into the adjacent solvent when driven by the bifurcated electron transfer shown. Fig. 1.6: Movement of the Rieske iron-sulfur protein (ISP) is required for catalysis of partial reactions during the Q-cycle. When cyt CJ interacting with either ubiquinone or heme Ci, hydrogen bonds are formed between the ISP and the cofactor binding site, while the intermediate state does not make specific interactions with either the cofactors or the other proteins in the complex. crystal contacts have Although stabilized an intermediate position of the Rieske ISP, the Heme bl binding switch between ubiquinone and heme c 1 may involve random exploration of cyt b three-dimensional space rather than discrete movement. The binding of the ubiquinone likely depends on whether the quinone is in the Qff or SQ• oxidation state. 23 Fig. 1.7: Proposed mechanism of hydroquinone oxidation at the Q site of 0 the bc 1 complex based on the multiple conformations of the Rieske ISP. Red iron cofactors are reduced while orange iron cofactors are oxidized. The ISP starts at an intermediate position with the Q0 binding site unoccupied (step 1; upper left). Hydroquinone (QH2) binds to Q0 (step 2), followed by ionization with the proton shuttled to the intermembrane space. The resultant Qff forms a ! hydrogen-bond to the shifted position of the ISP (step 3), stabilizing the protein in that conformation while an electron is transferred to the Rieske [2Fe:2S] cluster, SQ• c:()::l bl forming an unstable semiquinone (SQ•; step 4). The ISP changes conformations and is stabilized by hydrogen bonds to the cytochrome c1 protein subunit, while a second hydrogen is lost from the SQ• species, with the concomitant transfer of an electron to bL (step 5). The ISP then reduces heme c 1 (step 6) as the electron is passed from bL to bH (not shown). As the reduced cytochrome c 1 passes the electron to cytochrome c, the starting state is regenerated. 24 Fig. 1.8: A Met 8207 Met 8207 c::;:; stereoview of the cofactors of complex IV and their ligands. Electrons enter complex IV above the CuA copper Since Tyr A244 Tyr A244 center. CuA 1s equidistant between heme a and heme £½, it is unclear how the electron proceeds to the active site heme ll:3-Cu . 8 0 is believed to bind at the open site of heme £½ in contact with Cu 2 Fig. 1.9: The catalytic cycle of complex IV. . 8 Cus2+ Fea33+ OH-Tyr Although portions are widely debated, the catalytic cycle of complex IV has several generally accepted 2H; 2e - w Feai+ Cus1+ OH-Tyr intermediates that have been characterized. Four protons are vectorially translocated during the 02 2+Fe 3- 0 a '-.... Cus2+ ·~ A 0 OH-Tyr catalytic cycle (shown over the reaction arrows). In addition, four protons are combined with one 4+Fea3= 0 OH-Cu 82+ p •O-Tyr molecule of oxygen to form water (shown in boxes). H;e4+Fe 8 3= 0 OH-Cu 82+ F OH- Tyr H;e3+Fe 8 3- 0 H OH-Cu 82+ OH -Tyr 0 25 Fig. 1.10: Schematic drawing of the F,F0 ATP synthase. The F0 portion (a subunit in light green, b subunit in green and c subunits in lavender) is located in the membrane, while the F, portion is located in the mitochondrial matrix. Protons that flow through the cylindrical c subunits forces a rotation of the c subunits as well as y, while the a, b, and 6 subunits stabilize the a/~ heterotrimer. The rotation of the asymmetric y subunit within the stationery a/~ heterotrimer forces conformational changes within the a and ~ subunits. The conformational changes in the active site of the ~ subunit result in the formation of ATP. Fig. 1.11: Model for the generation of rotation by proton transport through the F0 portion of the ATPase. The central cylinder (blue) consists of ten c subunits while the external part (green) is composed of a single a subunit. The magenta arrow indicates the path of proton transfer. The flow of protons allows the c subunit to rotate with respect to the a subunit. This rotation continues through the y subunit of the F 1 portion. H+ 26 Fig. 1.12: Binding change mechanism of the ATP synthase. The F 1 ATP synthase is represented as a circle with alternating a and (3 subunits at four different stages of ATP synthesis. The asymmetric y subunit causes the conformational changes in the a and (3 subunits, as shown in the center in yellow. The three different conformational states of the 13 subunit are termed open (13 0 ), loose (13J, and tight (f½,). In the first step, an ATP molecule is bound to 13-r,, while 130 and (3L are empty. The ATP synthase then binds to ADP and Pi. After an input of energy by the F0 subunit, the F 1 part rotates 120° around the asymmetric y subunit causing conformational changes in the (3 subunits such that the ADP and Pi are now bound in 13-r, and ATP is bound to 130 • The 130 conformation has a low affinity for ATP, which is released in the final step, regenerating the initial state. 0 Fig. 1.13: Model for the control 0 0 0 of the FNR transcription factor in 0 response to the presence of 0 2 • 0 Oxidizing DNA binding of FNR 1s promoted by the formation of a Aeduc~ homodimer with each monomer containing a single [4Fe:4S] cluster. Within minutes of oxidation, the homodimer is converted to monomers, with the [4Fe:4S] clusters oxidized to [2Fe:2S] clusters. After several hours in oxidizing conditions, the [2Fe:2S] clusters are also lost. 27 Fig. 1.14: Structures of menaqumone and ubiquinone cofactors. the In the membrane, menaqumone hydrophobic tail has 0 0 a contains ubiquinone seven isoprenol groups while ubiquinone has ten. Fig. 1.15: Schematic representation of anaerobic respiration with formate and hydrogen as the electron donors and fumarate as the terminal electron acceptor. Anaerobic respiration occurs in the periplasmic membrane in bacteria. Formate dehydrogenase and hydrogenase catalyze the oxidation of formate and hydrogen, respectively, and transfer the electrons through the membrane-soluble menaquinone pool to quinol-fumarate reductase. Fig. adapted from (83 ). 1 µm 2H+ + Furn Ouinol Fumarate Reductase (QFR) HC0 2• E. Coli CO 2 + H+ H2 Formate Dehydrogenase Hydrogenase 2H+ 28 Fig. 1.16: The structural understanding of anaerobic respiration with fumarate as the terminal electron acceptor. The soluble domain of the Ni-Fe hydrogenase has been + 2H + Fumarate solved from Desulfovibrio gigas, and domain the from dehydrogenase '-..___/ soluble formate has Succinate Formate dehydrogenase Hydrogenase been solved from Escherichia Q coli. The complex II homologue quinol- fumarate reductase has been solved from Escherichia coli and Wolinella succinogenes. Table 1.1: Reduction potentials for Bos taurus mitochondrial complex II succinate +25mV FAD -79mV [2Fe:2S] OmV [4Fe:4S] -260mV [3Fe:4S] +60mV hemeb -185mV ubiquinone +113 mV "')' Quinol-fumarate reductase (QFR) 29 Table 1.2: Oxido-reductases of the anaerobic respiratory chains of Escherichia coli =E=n=.zy--'m=e_ _ _ _ _ _ _ _ _ _ _R"'-='-ed=o=x~co"'"'u"'-lp=l=-e-E'p-=a1= ·r_ _ _ _ _ _----'E=mmM Primary electron donors Formate dehydrogenase HCO 3-JCO2 + H+ -0.43 Hydrogenase H/H+ -0.42 NADH dehydrogenase NAD+/NADH -0.32 Glycerol 3-P dehydrogenase DHAP/Gly-3-P -0.19 Pyruvate oxidase acetate + CO/pyruvate Lactate dehydrogenase pyruvate/lactate -0.19 Glucose dehydrogenase glucose/gluconate -0.14 Succinate dehydrogenase succinate/fumarate +0.03 Nitrate reductase NO 3-/NO 2- +0.42 Nitrite reductase No 2-/NH4+ +0.36 DMSO reductase DMSO/DMS +0.16 TMAO reductase TMAO/TMA +0.13 Fumarate reductase fumarate/succinate +0.03 Termainal reductases 30 REFERENCES 1. Szalai, A., and Brudvig, G. W. (1998) American Scientist 86, 542-551. 2. Babcock, G. T. (1999) Proc. Natl. Acad. Sci. USA 96, 12971-12973. 3. Brandt, U. (1997) Biochim. Biophys. Acta 1318, 79-91. 4. Friedrich, T., Steinmutter, K. , and Weiss, H. (1995) FEBS Lett. 367, 107-111. 5. Walker, J. E., Skehel, J. M., and Buchanan, S. K. (1995) Methods Enzymol. 260, 1434. 6. Vercesi, A., Reynafarje, B., and Lehninger, A. L. (1978) J. Biol. Chem. 253, 63796385 . 7. DiVirgilio, F., and Azzone, G. F. (1982) J. Biol. Chem. 257, 4106-4113. 8. Wikstrom, M. (1984) FEBS Lett. 169, 300-304. 9. Brown, G. C., and Brand, M. D. (1988) Biochem. J. 252, 473-479. 10. Walker, J.E. (1992) Q. Rev. Biophys. 25, 253-324. 11. Fearnley, I. M., and Walker, J. E. (1992) Biochim. Biophys Acta 1140, 105-134. 12. Buchanan, S. K., and Walker, J.E. (1996) Biochem. J. 318, 343-349. 13. Skehel, J.M., Fearnley, I. M., and Walker, J. E. (1998) FEBS Lett. 438, 301-305. 14. Chomyn, A., Cleeter, M. W. J. , Ragan, C. I., Riley, M., Doolittle, R. F., and Attardi, G. (1986) Science 234, 614-618. 15. Weidner, U., Geier, S., Ptock, A., Friedrich, T., Leif, H., and Weiss, H. (1993) J. Mol. Biol. 233, 109-122. 16. Yagi, T., Yano, T., and Matsunoyagi, A. (1993) J. Bioenerg. Biomembr. 25, 339-345. 17. Yano, T., Chu, S.S ., Sled, V. D., Ohnishi, T., and Yagi, J. (1997) J. Biol. Chem. 272, 4201-4211. 18. Blattner, F. R., Plunkett, G., Bloch, C. A., Perna, N. T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J. D., Rode, C. K., Mayhew, G. A. F., Gregor, J., Davis, N. 31 W., Kirkpatrick, H. A., Goeden, M.A., Rose, D. J., Mau, B., and Shao, Y. (1997) Science 277, 1453-1474. 19. Guenebaut, V., Vincentelli, R., Mills, D., Weiss, H., and Leonard, K. R. (1997) J. Mol. Biol. 265, 409-418. 20. Guenebaut, V., Schlitt, A., Weiss, H., Leonard, K., and Friedrich, T. (1998) J. Mol. Biol. 276, 105-112. 21. Ragan, C. I. (1987) Curr. Top. Bioenerg. 15, 1-36. 22. Pilkington, S. J., Skehel, J.M. , Gennis, R. B., and Walker, J.E. (1991) Biochemistry 30, 2166-2175. 23 . Rao, N. A., Felton, S. P., Huennekens, F. M. , and Mackler, B. (1963) J. Biol. Chem. 238, 449-455. 24. Yano, T., and Yani, T. (1999) J. Biol. Chem. 274, 28606-28611. 25. Ohnishi, T. (1979) in membrane proteins in energy transduction (Capaldi, R. A., Ed.), Marcel Dekker, New York. 26. Ohnishi, T., and Salerno, J. C. (1982) in Iron-sulfur proteins (Spiro, T. G., Ed.), Wiley, New York. 27 . Beinert, H., and Albracht, S. P. J. (1982) Biochim. Biophys. Acta 683, 245-277. 28. Yano, T., Magnitsky, S., Sled, V. D., Ohnishi, T., and Yagi, T. (1999) J. Biol. Chem. 274, 28598-28605. 29. Ohnishi, T. (1998) Biochim. Biophys. Acta 1364, 186-206. 30. Suzuki, H., and King, T. E. (1983) J. Biol. Chem. 268, 352-358. 31. Vinogradov, A. D. , Sled, V. D., Burbaev, D. S., Grivennikova, V. G., Moroz, I. A., and Ohnishi, T. (1995) FEBS Lett. 370, 83-87. 32. Burbaev, D. S., Moroz, I. A., Kotlyar, A. B., Sled, V. D. , and Vinogradov, A. D. (1989) FEBS Lett. 254, 47-51. 32 33. Kotlyar, A. B. , Sled, V. D., D.S., B., Moroz, I. A., and Vinogradov, A. D. (1990) FEBS Lett. 264, 17-20. 34. Ohnishi, T., Magnitsky, S., Toulokhonova, L., Yano, T., Yagi, T., Burbaev, D. S., and Vinogradov, A. D. (1999) Biochem. Soc. T. 27, 586-591. 35. Albracht, S. P. J., Mariette, A., and deJong, P. (1997) Biochim. Biophys. Acta 1318, 92-106. 36. Krishnamoorthy, G. , and Hinkle, P. C. (1988) J. Biol. Chem. 263, 17566-17575. 37. Dutton, P. L., Moser, C. C., Sled, V. D., Daldal, F., and Ohnishi, T. (1998) Biochim. Biophys. Acta 1364, 245-257. 38. Ingledew, W. J., and Ohnishi, T. (1980) Biochem. J. 186, 111-117. 39. Gutman, M., T.P., S., and Beinert, H. (1972) Biochemistry 11, 556-562. 40. Sled, V. D. , Friedrich, T., Leif, H., Weiss, H., Meinhardt, S. W., Fukumori, Y., Calhoun, M. W., Gennis, R. B., and Ohnishi, T. (1993) J. Bioenerg. Biomembr. 25, 347356. 41. Salerno, J. C. , Blum, H. , and Ohnishi, T. (1979) Biochim. Biophys. Acta 547, 270281. 42. Ohnishi, T., Leigh, J. S., Ragan, C. I., and Racker, E. (1974) Biochem. Biophys. Res. Commun. 56, 775-782. 43 . Ackrell, B. A. C., Johnson, M. K., Gunsalus, R. P., and Cecchini, G. (1992) Chemistry and biochemistry offlavoenzymes, Vol. 3, CRC Press, Boca Raton. 44. Hagerhall, C. (1997) Biochim. Biophys. Acta 1320, 107-141. 45. Guest, J. R., and Russell, G. C. (1992) Curr. Top. Cell. Regul. 33, 231-247. 46. Kowal, A. T., Werth, M. T., Moanodori, A., Cecchini, G., Schroder, I., Gunsalus, R. P., and Johnson, M. K. (1995) Biochemsitry 34, 12284-12293. 47. Morningstar, J. E., Johnson, M. K., Cecchini, G., Ackrell, B. A. C., and Kearney, E. B. ( 1985) J. Biol. Chem. 260, 3631-3638. 33 48. Cammack, R., Patil, D. S., and Weiner, J. H. (1986) Biochim. Biophys. Acta 870, 545551. 49. Gest, H. (1980) FEMS Lett. 7, 73-77. 50. Singer, T. P., Kearney, E. B., and Kenney, W. C. (1973) Adv. Enzymol. 37, 189-272. 51. Pealing, S. L., Black, A. C. , Manson, F. D., Ward, F. B. , Chapman, S. K., and Reid, G. A. (1992) Biochemistry 31, 12132-12140. 52. Simon, J., Gross , R. , Klimmek, 0. , Ringel , M. , and Kroger, A. (1998) Arch. Microbial. 169, 424-433. 53 . Pealing, S. L. , Lysek, D. A. , Taylor, P., Alexeev, D., Reid, G. A., Chapman, S. K. , and Walkinshaw, M. D. (1999) J. Structu. Biol. 127, 76-78. 54. Flachmann, R. , Kunz, N. , Seifert, J. , Gutlich, M., Wientjes, F. J., Laufer, A., and Gassen, H. G. (1988) Eur. J. Biochem. 175, 221-228. 55. Sun, D., and Setlow, P. (1993) J. Bacterial. 175, 1423-1432. 56. Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E. , Nakamura, Y., Miyajima, N., Hirosawa, M., Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N., Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M., Yasuda, M., and Tabata, S. (1996) DNA Res. 3, 109-136. 57. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S. V., Eiglmeier, K., Gas, S., Barry, C. E. I. I. I., Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T. , Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin, N., Holroyd, S., Hornsby, T., Jagels, K., Krogh, A., McLean, J., Moule, S., Murphy, L., Oliver, S., Osborne, J., Quail, M.A., Rajandream, M. A., Rogers, J., Rutter, S., Seeger, K., Skelton, S., Squares, S., Sqares, R., Sulston, J.E., Taylor, K., Whitehead, S., and Barrell, B. G. (1998) Nature 393, 537-544. 34 58. Deckert, G., Warren, P. V. , Gaasterland, T., Young, W. G., Lenox, A. L., Graham, D. E., Overbeek, R., Snead, M.A., Keller, M. , Aujay, M., Huber, R. , Feldman, R. A., Short, J. M. , Olson, G. J., and Swanson, R. V. (1998) Nature 392, 353-358. 59. Redenbach, M., Kieser, H. M., Denapaite, D., Eichner, A. , Cullum, J., Kinashi, H., and Hopwood, D. A. (1996) Mal. Microbial. 21, 77-96. 60. Tettelin, H., Saunders, N. J., Heidelberg, J., Jeffries, A. C., Nelson, K. E., Eisen, J. A., Ketchum, K. A. , Hood, D. W. , Peden, J. F. , Dodson, R. J., Nelson, W. C., Gwinn, M. L. , DeBoy, R., Peterson, J. D., Hickey, E. K. , Haft, D. H., Salzberg, S. L., White, 0., Fleischmann, R. D. , Dougherty, B. A., Mason, T., Ciecko, A., Parksey, D. S., Blair, E., Cittone, H., Clark, E. B., Cotton, M. D., Utterback, T. R., Khouri, H., Qin, H., Vamathevan, J., Gill, J., Scarlato, V., Masignani, V., Pizza, M., Grandi, G., Sun, L., Smith, H. 0 ., Fraser, C. M., Moxon, E. R., Rappuoli, R., and Venter, J.C. (2000) Science 287, 1809-1815. 61. Plesiat, P., Grandguillot, M. , Harayama, S., Vragar, S., and Michel-Briand, Y. (1991) J. Bacterial. 173, 7219-7227. 62. Morii, S., Fujii, C. , Miyoshi, T., Iwami, M., and Itagaki, E. (1998) J. Biochem. 124, 1026-1032. 63. Molnar, I., Choi, K. P., Yamashita, M., and Murooka, Y. (1995) Mal. Microbial. 15, 895-905. 64. Galibert, F., Alexandraki, D., Baur, A., Boles, E., Chalwatzis, N., Chuat, J. C., Coster, F., Cziepluch, C., De Haan, M., Domdey, H., Durand, P., Entian, K. D., Gatius, M., Goffeau, A., Grivell, L.A., Hennemann, A., Herbert, C. J., Heumann, K., Hilger, F., Hollenberg, C. P., Huang, M. E., Jacq, C., Jauniaux, J.C., Katsoulou, C., and KarpfingerHartl, L. e. a. (1996) EMBO J. 15, 2031-2049. 65. Goffeau, A., Barrell, B. G., Bussey, H., Davis, R. W., Dujon, B., Feldmann, H., Galibert, F., Hoheisel, J. D., Jacq, C., Johnston, M., Louis, E. J., Mewes, H. W., 35 Murakami, Y., Philippsen, P., Tettelin, H., and Oliver, S. G. (1996) Science 274, 546550. 66. Maklashina, E., Berthold, D. K., and Cecchini, G. (1998) J. Bacterial. 180, 59895996. 67. Pershad, H. R. , Hirst, J., Cochran, B. , Ackrell, B. A. C., and Armstrong, F. A. (1999) Biochim. Biophys. Acta 1412, 262-272. 68. Blaut, M. , Whittaker, K., Valdovinos, A., Ackrell, B. A. C., Gunsalus, R. P., and Cecchini, G. (1989) J. Biol. Chem. 264, 13599-13604. 69. Robinson, K. M., Rothery, R. A., Weiner, J. H., and Lemire, B. D. (1994) Eur. J. Biochem. 222, 983-990. 70. Walker, W. H., and Singer, T. P. (1970) J. Biol. Chem. 245, 4224-4225. 71. Iverson, T. M. , Luna-Chavez, C. , Cecchini, G., and Rees, D. C. (1999) Science 284, 1961-1966. 72. Lancaster, C.R. D., Kroger, A., Auer, M., and Michel, H. (1999) Nature 402, 377385. 73. Matttevi, A., Tedeschi, G., Bacchella, L., Coda, A., Negri, A., and Ronchi, S. (1999) Structure 7, 745-756. 74. Bamford, V., Dobbin, P. S., Richardson, D. J., and Hemmings, A. M. (1999) Nat. Struct. Biol. 6, 1104-1109. 75 . Leys, D., Tsapin, A. S., Nealson, K. H., Meyer, T. E., Cusanovich, M. A., and VanBeeumen, J. J. (1999) Nat. Struct. Biol. 6, 1113-1117. 76. Taylor, P., Pealing, S. L., Reid, G. A., Chapman, S. K., and Walkinshaw, M. D. (1999) Nat. Struct. Biol. 6, 1108-1112. 77. Hagerha.11, C., and Hederstedt, L. (1996) FEBS Lett. 389, 25-31. 78. Westenberg, D. J., Gunsalus, R. P., Ackrell, B. A. C. , Sices, H., and Cecchini, G. (1993) J. Biol. Chem. 268, 815-822. 36 79. Lee, G. Y., He, D. Y., Yu, L. , and Yu, C. A. (1995) J. Biol. Chem. 270, 6193-6198. 80. Shenoy, S. K., Yu, L. , and Yu, C. A. (1997) J. Biol. Chem. 272, 17867-17872. 81. Hagerhall, C., Magnitsky, S., Sled, V. D., Schroder, I., Gunsalus, R. P., Cecchini, G., and Ohnishi, T. (1999) J. Biol. Chem. 274, 26157-26164. 82. Xia, D., Yu, C., Kim, H., Xia, J., Kachurin, A. M. , Zhang, L., L. Yu, and Deisenhofer, J. (1997) Science 277, 60-66. 83 . Zhang, Z. (1998) Nature 392, 677-684. 84. Iwata, S., Lee, J. W., Okada, K., Lee, J. K., Iwata, M., Rasmussen, B., Link, T. A., Ramaswamy, S., and Jap, B. K. (1998) Science 281, 64-71. 85. Ohnishi, T. , Moser, C. C., Page, C. C., Dutton, P. L., and Yano, T. (2000) Structure 8, R23-R32. 86. Yang, X. H., and Trumpower, B. L. (1986) J. Biol. Chem. 262, 8702-8706. 87. Robertson, D. E., Ding, H. G., Chelminski, P.R., Slaughter, C., Hsu, J., Moomaw, C., Tokita, M., Daldal, F., and Dutton, P. L. (1993) Biochemistry 32, 1310-1317. 88. Schagger, H., Brandt, U., Gencic, S., and von Jagow, G. (1995) Method. Enzymol. 260, 82-96. 89. Braun, H.P., and Schmitz, U. K. (1997) Int. J. Biochem. Cell. B. 29, 1043-1045. 90. Glaser, E., and Dessi, P. (1999) J. Bioenerg. Biomemb. 31, 259-274. 91. Mitchell, P. (1961) Nature 191, 144-148. 92. Cramer, W. A., and Soriano, G. M. (1999) Biofactors 2, 81-6. 93. Wikstrom, M. K., and Berden, J. A. (1972) Biochim. Biophys. Acta 283, 403-407. 94. Trumpower, B. L. (1990) J. Biol. Chem. 265, 11409-11412. 95. Brandt, U., and Trumpower, B. (1994) Crit. Rev. Biochem. Mol. 29, 165-197. 96. Brandt, I. (1996) FEES Lett. 387, 1-6. 97. Link, T. A. (1997) FEES Lett 412, 257-264. 98. Brandt, U. (1998) Biochim. Biophys. Acta 1365, 261-268. 37 99. Brandt, U. (1999) J. Bioenerg. Biomembr. 31, 243-250. 100. Crofts, A. R., Barquera, B., Gennis, R. B., Kuras, R. , Guergova-Kuras, M ., and Berry, E. A. (1999) Biochemistry 38, 15807-15926. 101. Yu, C. A., Zhang, L. , Deng, K. P., Tian, H., Xia, D., Kim, H., Deisenhofer, J., and Yu, L. (1999) Biofactors 9, 103-109. 102. Brugna, M., Rodgers, S., Schricker, A. , Montoya, G., Kazmeier, M. , Nitschke, W., and Sinning, I. (2000) Proc. Natl. Acad. Sci. USA 97, 2069-2074. 103. Crofts, A. R., and Wang, Z. G. (1989) Photosynth. Res. 22, 69-87. 104. Brandt, U., and vonJargow, G. (1991) Eur. J. Biochem. 195, 163-170. 105. Brandt, U., Haase, U., Schagger, H., and vonJargow, G. (1991) J. Biol. Chem. 266, 19958-19964. 106. Ding, H., Moser, C. C., Robertson, D. E. , Tokito, M. K., Daldal, F., and Dutton, P. L. (1995) Biochemistry 34, 15979-15996. 107. Kim, H., Xia, D. , Yu, C., Xia, J., Kachurin, A. M. , Zhang, L., Yu, L., and Deisenhofer, J. (1998) Proc. Natl. Acad. Sci. USA 95, 8026-8033. 108. Crofts, A. R., Barquera, B. , Gennis, R. B., Kuras, R., Guergova-Kuras, M., and Berry, E. A. (1999) Biochemistry 38, 15807-15826. 109. Brandt, U., and Okun, J. G. (1997) Biochemistry 36, 11234-11240. 110. Iwata, S. (personal communication) . 111. Crofts, A. R., Hong, S., Zhang, Z., and Berry, E. A. (1999) Biochemistry 38, 1582715839. 112. Saraste, M. (1990) Q. Rev. Biophys. 23, 331-366. 113. Babcock, G. T., and Wikstrom, M. (1992) Nature 356, 301-309. 114. Calhoun, M. W., Thomas, J. W., and Gennis, R. B. (1994) Trends. Biochem. Sci. 19, 325-330. 115. Ferguson-Miller, S., and Babcock, G. T. (1996) Chem. Rev. 96, 2889-2907. 38 116. Wikstrom, M. (1977) Nature 266, 271-273. 117. Solioz, M., Carafoli, E., and Ludwig, B. (1982) J. Biol. Chem. 257, 1579-1582. 118. Sone, N., and Hinkle, P. C. (1982) J. Biol. Chem. 257, 2600-2604. 119. Puustinen, A., Finel, M. , Virkki, M., and Wikstrom, M. (1989) FEBS Lett. 249, 163167. 120. Musser, S. M., and Chan, S. I. (1998) J. Mal. Evol. 46, 508-520. 121. Garcia-Horsman, J. A., Barquera, B. , Rumbley, J., Ma, J. X., and Gennis, R. B. (1994) J. Bacterial. 176, 5587-5600. 122. Kadenbach, B., Kuhnnentwig , L., and Buge, U. (1987) Curr. Top. Bioenerg. 15, 113-161. 123. Musser, S. M., Stowell, M. H. B. , and Chan, S. I. (1995) Adv. Enzymol. RAMB 71, 79-208. 124. Iwata, S., Ostermeier, C., Ludwig, B., and Michel, H. (1995) Nature 376, 660-669. 125. Ostermeier, C., Harrenga, A. , Ermler, U. , and Michel, H. (1997) Proc. Natl. Acad. Sci. USA 94 , 10547-10553. 126. Tsukihara, T., Aoyama, H. , Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawaltoh, K., Nakashima, R., Yaono, R. , and Yoshikawa, S. (1995) Science 269, 1069-1074. 127. Tsukihara, T., Aoyama, H., Tomizaki, E. Y. Y., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., and Yoshikawa, S. (1996) Science 272, 1136-1144. 128. Yoshikawa, S., Shinzawa-ltoh, K., Nakashima, R., Yaono, R. , Yamashita, E., Inoue, N., Yao, M., Fie, M. J., Libeu, C. P., Mizushima, T., Yamaguchi, H., Tomizaki, T., and Tsukihara, T. (1998) Science 280, 1723-1729. 129. Moller, M., Schlapfer, B., and Azzi, A. (1988) Proc. Natl. Acad. Sci. USA 85, 66476651. 130. Yoshikawa, S., Tera, T., Takahashi, Y., T., T., and Caughey, W. S. (1988) Proc. natl. Acad. Sci. USA 85, 1354-1358. 39 131. Hiser, L., Di Valentin, M. , Hamer, A. G., and Hosler, J. P. (2000) J. Biol. Chem. 275, 619-623 . 132. Napiwotzki, J., Shinzawa-Itoh, K., Yoshikawa, S., and Kadenbach, B. (1997) Biol. Chem. 378, 1013-1021. 133. Bender, E., and Kadenbach, B. (2000) FEBS Lett. 466, 130-134. 134. Hosler, J. P., Espe, M. P., Zhen, Y. J., Babcock, G. T., and Ferguson-Miller, S. (1995) Biochemistry 34, 7586-7592. 135. Espe, M. P., Hosler, J. P., Ferguson-Miller, S., Babcock, G. T., and McCracken, J. (1995) Biochemistry 34, 7593-7602. 136. Kelly, M., Lappalainen, P., Talbo, G., Haltia, T., Vanderoost, J., and Saraste, M. (1993) J. Biol. Chem. 268, 16781-16787. .. 137. Zhen, Y. J., Hoganson, C. W., Babcock, G. T., and Ferguson-Miller, S. (1999) J. Biol. Chem. 274, 38032-38041. 138. Wang, K. F., Zhen, Y. J. , Sadoski, R., Grinnell, S. , Geren, L., G., Ferguson-Miller, S., Durham, B., and Millett, F. (1999) J. Biol. Chem. 274, 38042-38050. 139. Roberts, V. A., and Pique, M. E. (1999) J. Biol. Chem. 274, 38051-38060. 140. Thomas, J. W., Calhoun, M . W ., Lemieux, L. J. , Puutstinen, A., Wikstrom, M., Alben, J. 0., and Gennis, R. B. (1994) Biochemistry 33, 13013-13021. 141. Verkhovsky, M. I., Jasaitis, A., Verkovskaya, M. L., Morgan, J.E., and Wikstrom, M. (1999) Nature 400, 480-483. 142. Perutz, M. F. (1990) Mechanism of Cooperativity and Allosteric Regulation in Proteins, Cambridge University Press, Cambridge. 143. Williams, R. J.P. (1995) Nature 376, 643-643. 144. Gennis, R., and S., F.-M. (1995) Science 269, 1063-1064. 40 145. Mitchell, D. M., Fetter, J. R. , Mills, D. A., Adelroth, P., Pressler, M. A., Kim, Y., Aasa, R. , Brzezinski, P., Malmstrom, B. G., Alben, J. 0 ., Babcock, G. T., FergusonMiller, S. , and Gennis, R. B. (1996) Biochemistry 35, 13089-13093. 146. Michel, H. (1998) Proc. Natl. Acad. Aci. USA 95, 107-112. 147. Michel, H. (1999) Biochemistry 38, 15129-15140. 148. Michel, H. (1999) Nature 402 , 602-603 . 149. Verkhovsky, M. I., Jasaitis, A., Verkhovskaya, M. L. , Morgan, J.E. , and Wikstrom, M . (1999) Nature 402. 150. Boyer, P. D. (1997) Annu. Rev. Biochem. 66, 717-749. 151. Stock, D., Leslie, A. G. W. , and Walker, J.E. (1999) Science 286, 1700-1705. 152. Abrahams, J. P., Leslie, A. G. W., Lutter, R., and Walker, J. E. (1994) Nature 370, 621-628. 153 . Shirakihara, Y., Leslie, A. G. W., Abrahams, J. P., Walker, J. E., Ueda, T., Sekimoto, Y., Kambara, M., Saika, K., Kagawa, Y., and Yoshida, M. (1997) Structure 5, 825-836. 154. Noji, H., Yasuda, R. , Yoshida, M., and Kinoshita, K. (1997) Nature 386, 8215-8221. 155. Yasuda, R., Noji, H., Kinosita, K., and Yoshida, M. (1998) Cell 93, 1117-1124. 156. Sambongi, Y., Iko, Y., Tanabe, M., Omote, H., Iwamoto-Kigara, A., Ueda, I. , Yanagida, T., Wada, Y., and Futai, M. (1999) Science 286, 1722-1724. 157. Girvin, M. E., Rastogi, V. K., Abildgaard, F., Markley, J. L., and Fillingame, R.H. (1998) Biochemistry 37, 8817-8824. 158. Sabbert, D., Engelbrecht, S., and Junge, W. (1996) Nature 381, 623-625. 159. Boyer, P. D. (1993) Biochim. Biophys. Acta 1140, 215-250. 160. Bauer, C. E., Elsen, S., and Bird, T. H. (1999) Annu. Rev. Microbial. 53, 495-523. 161. Guest, J. R., and Lambden, P. R. (1976) J. Gen. Microbial. 180. 162. Sawers, R. G., Zalelein, E., and Bock, A. (1988) Arch. Microbial. 149, 240-244. 41 163. Khoroshilova, N., Beinert, H. , and Kiley, P. J. (1995) Proc. Natl. Acad. Sci. USA 92, 2499-2503. 164. Lazazzera, B. A., Beinert, H., Khoroshilova, N., Kennedy, M . C., and Kiley, P. J. (1996) J. Biol. Chem. 271, 2762-2768. 165. Khoroshilova, N., Popescu, C., Mtinck, E., Beinert, H ., and Kiley, P. J. (1997) Proc. Natl. Acad. Sci. USA 94, 6087-6092. 166. Popescu, C., Bates, D. M., Beinert, H., Milnck, E., and Kiley, P. J. (1998) Proc. Natl. Acad. Sci. USA 95, 13431-13435. 167. Unden, G., and Bongaerts, J. (1997) Biochim. Biophys. Acta 1320, 217-234. 168. Mell, H., Bronder, M., and Kroger, A. (1982) Arch. Microbial. 131, 224-228. 169. Kroger, A. , Geisler, V., Lemma, E., Theis, F., and Lenger, R. (1992) Arch. Microbial. 158, 477-483. 170. Sawers, G. (1994) Anton van Leeuwenhoek 66, 57-88. 171. Geisler, V., Ullmann, R., and Kroger, A. (1994) Biochim. Biophys. Acta 1184, 219226. 172. Stephenson, M., and Stickland, L. H. (1931) Biochem. J. 25. 173. Adams, M. M. W. (1990) Biochim. Biophys. Acta 1020, 115-145. 174. Pryzbyla, A. E., Robbins, J., Menon, N., and Peck, H. D. (1992) FEMS Microbial. Rev. 88, 109-135. 175. Wu, L. F., and Mandrand, M.A. (1993) FEMS Microbial. Rev. 104, 243-270. 176. Schauder, R., and Kroger, A. (1993) Arch. Microbial. 159, 238-242. 177. Hedderich, R. , Klimmek, 0., Kroger, A., Dirmeier, R., Keller, M., and Stetter, K. 0. (1999) FEMS Microbial. Rev. 22, 353-381. 178. Reeve, J. N., and Beckler, G. S. (1990) FEMS Microbial. Rev. 87, 419-424. 179. Higuchi, Y., Yagi, T., and Yasuoka, N. (1997) Structure 5, 1671-1680. 42 180. Volbeda, A., Charon, M. H., Piras, C., Hatchiikian, E. C., Frey, M., and FontecillaCamps, J.C. (1995) Nature 373, 580-587. 181. Axley, M. J., Grahame, D. A., and Stadtman, T. C. (1990) J. Biol. Chem. 265, 18213-18218. 182. Gladyshev, V. N., Khangulov, S. V., Axley, M. J., and Stadtman, T. C. (1994) Proc. Natl. Acad. Sci. USA 91 , 7708-7711 . 183. Boyington, J. C., Gladyshev, V. N., Khangolov, S. V., Stadtman, T. C., and Sun, P. D. (1997) Science 275, 1305-1308. 184. Schindelin, H. , Kisker, C. , Hilton, J., Rajagopalan, K. V., and Rees , D. C. (1996) Science 272, 1615-1621. 185. Schneider, F., Lowe, J., Huber, R., Schindelin, H. , Kisker, C., and Knablein, J. (1996) J. Mol. Biol. 263, 53-69. 186. McEwan, A.G. (1994) Anton. van Leeuwenhoek 66, 151-164. 187. Zinoni, F., Birkmann, A., Stadtman, T. C., and Bock, A. (1986) Proc. Natl. Acad. Sci. USA 83 , 4650-4654. 188. Zinoni, F., Birkmann, A., Leinfelder, W., and Bock, A. (1987) Proc. Natl. Acad. Sci. USA 84, 3156-3160. 43 Chapter 2 Luna-Chavez, C., Iverson, T .M., Rees , D.C., Cecchini, G. (2000) Overexpression, Purification and Crystallization of the Membrane-bound Fumarate Reductase from Escherichia coli. Protein Exp. Pur. , in press. 44 Overexpression, Purification and Crystallization of the Membrane-bound Fumarate Reductase from Escherichia coli Running Title: Crystallization of Fumarate Reductase Cesar Luna-Chavez, 1 Tina M. Iverson, 2 Douglas C. Rees, 3 and Gary Cecchini 1* 1 Molecular Biology Division (151-S), VA Medical Center, San Francisco, CA 94121 and Department of Biochemistry and Biophysics, University of California, San Francisco, CA 94143 2 Graduate Option in Biochemistry, 147-75CH, California Institute of Technology, Pasadena, CA 91125 3 Howard Hughes Medical Institute, Division of Chemistry and Chemical Engineering, 147-75CH, California Institute of Technology, Pasadena, CA 91125 *To whom correspondence should be addressed at Molecular Biology Div. (151-S), VA Medical Center, 4150 Clement Street, San Francisco, CA 94121. Telephone: (415) 7529676, FAX: (415) 750-6959, E-mail: ceccini@itsa.ucsf.edu. Key Words: Fumarate reductase, succinate dehydrogenase, crystallography, membrane proteins. 45 4 Abbreviations used: Ap', ampicillin resistant; CHAPS, (3-[(3-cholamidopropyl) dimethy lammonio ]-1-propane-sulfonate; cmc, critical micelle concentration; DCIP, 2,6dichlorophenolindophenol; Fp, flavoprotein subunit; GuHCl, guanidinium hydrochloride; Ip, iron-sulfur protein subunit; LB , Luria-Bertani; PEG, polyethylene glycol; PES, phenazine ethosulfate; QFR, quinol:fumarate reductase; SQR, succinate:quinone reductase 46 ABSTRACT Quinol-fumarate reductase (QFR) from Escherichia coli is a membrane-bound four subunit respiratory protein that shares many physical and catalytic properties with succinate-quinone oxidoreductase (SQR) (EC 1.3.99.1) commonly referred to as complex II. The E. coli QFR has been overexpressed using plasmid vectors so that more than 50% of the cytoplasmic membrane fraction is composed of the four subunit enzyme complex. The growth characteristics required for optimal levels of expression with minimal degradation by host cell proteases and oxidation factors were determined for the strains harboring the recombinant plasmid. The enzyme is extracted from the enriched membrane fraction using the non-ionic detergent Thesit (polyoxyethylene(9)dodecyl ether) in a monodisperse form and then purified by a combination of anion exchange, perfusion, and gel filtration chromatography. The purified enzyme is highly active and contains all types of redox cofactors expected to be associated with the enzyme. Crystallization screening of the purified QFR by vapor diffusion resulted in the formation of crystals within 24 hours using a sodium citrate buffer and polyethylene glycol precipitant. The crystals contain the complete four subunit QFR complex, diffract to 3.3 A resolution and were found to be in space group P2 12 12 1 with unit cell dimensions a=96.6 A, b=138.l A, and c=275 .3 A. The purification and crystallization procedures are 47 highly reproducible and the general procedure may prove useful for complex II' s from other sources. 48 INTRODUCTION Complex II' s (succinate:quinone reductase, SQR and quinol:fumarate reductase, QFR) 4 are structurally and functionally similar membrane-bound enzymes involved in aerobic and anaerobic respiration, respectively. The enzymes catalyze the oxidation of succinate or the reduction of fumarate in the mitochondrial matrix or bacterial cytoplasm and the reduction or oxidation of quinone/quinol in the membrane. All complex II enzymes are composed of a large flavoprotein subunit (Fp, 66-70 kDa) which contains an 8a-N(3)-histidyl-FAD cofactor and the dicarboxylate-binding site and an iron-sulfur protein subunit (Ip, -27 kDa) containing three distinct iron-sulfur clusters (1, 2). The Fp and Ip catalytic dimer is linked to the membrane in most cases by two hydrophobic polypeptides. The membrane anchor domain of the enzyme (FrdC/SdhC and FrdD/SdhD) is more varied in composition and in some cases consists of only one polypeptide. Depending on the species from which the enzyme is isolated, the membrane anchors also may contain and also may contain zero, one, or two b hemes. The primary amino acid sequence similarity is lower in the membrane domain; however, evidence indicates that the membrane anchors have a conserved general structure (2, 3). The membrane anchor domain of complex II also provides binding sites for quinones as has been demonstrated in SQR/QFR by various methods (4-8). Purified preparations of SQR (9) and QFR (10) in a membrane-bound form have been achieved for a number of years; however, the enzymes were often isolated under 49 harsh conditions that gave varied stoichiometry of subunits and cofactors (1, 11 ). The SQR and QFR complexes from Escherichia coli have been cloned and overexpressed and this has proven useful in obtaining highly purified preparations (12, 13). The E. coli complexes have the additional advantage for studying the properties of complex II that they are easily genetically manipulated. In many respects this has made them the model systems of choice for investigating complex II structure and the function of the various redox cofactors from the complex. In the past few years high resolution structures for several of the membranebound electron transport chain respiratory complexes have become available (14-19). Nevertheless, the number of membrane proteins that have been crystallized in forms that diffract to high resolution remains relatively small in comparison to the number of soluble proteins for which there are well-ordered crystals. There are a number of reasons for the lack of high resolution crystals structures for membrane proteins. These include the fact that many are relatively unstable when extracted from their native membrane environment by the use of detergents where they become exposed to conditions quite different than their native environment (20). Membrane proteins also are often difficult to overexpress often becoming toxic to host cells or forming inclusion bodies. The most recent x-ray structures that have become available for membrane-bound electron transport proteins are the four-subunit non-heme containing fumarate reductase from E. coli (21) and the three subunit diheme enzyme from Wolinella succinogenes (22). This 50 manuscript describes the modified and improved procedures that were used to achieve expression and purification of E. coli QFR that resulted in a protein complex that was amenable to crystallization and also contains two tightly bound menaquinone molecules. MATERIALS AND METHODS Materials . "Complete" protease inhibitor cocktail tablets, the detergent Thesit (polyoxyethylene(9)dodecyl ether), and the perfusion chromatography anion exchange column, POROS 50HQ were obtained from Roche Molecular Biochemicals (Indianapolis, IN) . DEAE Sepharose Fast Flow, Sephacryl S-300, PhastGel isoelectrofocusing gels, and high and low molecular weight protein standards were obtained from Amersham Pharmacia Biotech (Piscataway, NJ). SDS-PAGE (10-20%) and native PAGE precast Tris-HCl gels were obtained from Bio-Rad (Hercules, CA). Protein concentration was accomplished with YM30 membranes and Centriprep-30 concentrators from Amicon (Bradford, MA). The Membfac screening kit from Hampton Research (Laguna Hills, CA) was used for crystallization trials. Reagents used for optimization of crystallization conditions were from Fluka (Milwaukee, WI), and all other reagents were analytical grade and obtained from Sigma (St. Louis, MO). Double distilled water was used for preparing all solutions. Growth and overexpression of recombinant QFR. The construction of E. coli strain DW35 (!ifrdABCD sdhC:: kan) has been described elsewhere (7) and this strain is 51 completely deficient in both fumarate reductase and succinate dehydrogenase activity. Plasmid pH3 (Apr) encodes wild type E. coli fumarate reductase (frdA + B+ c+ D+) and is a derivative of pBR322 (23). Growth of bacteria was initiated from a single colony of DW35 transformed with pH3 by transferring to 5-ml of LB medium containing 100 µg ampicillin/ml and growing overnight with shaking at 275 rpm at 37 °C. A 3-rnl inoculum from the overnight culture was then added to 50-ml of LB medium plus ampicillin and grown as above for 6 h. Appropriate inocula (25/OD 600 = ml per 20-liter carboy flask) were then added to 20-liters of anaerobic growth medium contained in a 20-liter carboy filled to the neck. The growth medium contained glycerol-fumarate medium (24) minus manganese, 35 µg ampicillin/ml, and 0.2% (w/v) and 0.1 % (w/v) tryptone and yeast extract, respectively. The cells were grown anaerobically with slow stirring overnight at 37 °C. The cells were harvested by continuous flow centrifugation at 4 °C when they had reached late log phase (OD600 = 1.1-1.2, - 17 h) and then kept on ice for immediate preparation of crude membranes. Typically the yield of cells (wet wt.) is 1.5 g per liter of culture. Preparation of membranes. All subsequent steps are done at 4 °C . Crude membrane fractions enriched in E. coli inner membranes and fumarate reductase containing tubules were prepared as follows: 90 g of cells were resuspended by homogenization in 300-ml of buffer A ( 100 mM potassium phosphate, 0.1 mM EDT A, pH 6.8) and in the presence of small amounts of DNAse, RNAse, and six Complete 52 protease inhibitor tablets and then ruptured by two passages through a French pressure cell at 15,000 p.s.i. The cell lysate was centrifuged at 10,000 x g for 15 min. The supernatant was carefully removed and centrifuged for 1 hour at 120,000 x g. The pellets from this centrifugation were resuspended in 180-ml of buffer A containing 3 protease inhibitor tablets and centrifuged for 15 min at 10,000 x g to remove any remaining cellular debris. A final centrifugation for 1 h at 120,000 x g yielded a dark brown membrane fraction enriched for QFR. The membrane fraction contained about one gram of total protein and the QFR is approximately 30-50% pure. The membranes could be stored at -80 °C for use within one month. Solubilization of QFR. The membranes described above were resuspended in 120 ml of buffer B (20 mM Tris-HCl, 0.1 mM EDTA, pH 7.4) with 3 Complete protease inhibitor tablets. Thesit was added from a 20% (w/v) stock to give a final concentration of 2% and 5-10 mg protein/ml. Fresh detergent stock was prepared for each enzyme preparation and used within 7-10 days. The membrane suspension was stirred briefly and then centrifuged at 100,000 x g for 30 min. The supernatant containing the solubilized QFR was filtered using a 0.2 µm nylon filter and kept at 4 °C for immediate chromatography. FPLC Chromatography for purification of QFR. The dark brownish supernatant from the previous step was applied to a DEAE Sepharose FF column (3.6 x 30 cm, 300ml bed volume) previously equilibrated with buffer C (20 mM Tris-HCl, 0.1 mM EDTA, 53 0.05% (w/v) Thesit, pH 7.4). The column was then washed with two bed volumes of buffer C followed by two bed volumes of the same buffer containing 0.1 M NaCl. The QFR complex was eluted with seven bed volumes of buffer C using a linear gradient from 0.1 - 0.25 M NaCl at a flow rate of 2 ml/min. The eluant peak containing QFR appears between 0.11 - 0.25 M NaCl as a broad peak; however, only the central fractions of these samples were pooled for subsequent steps. The salt background of the pooled samples was reduced by repeated exchange in an Amicon stirred cell with a YM30 membrane against buffer C to a final concentration of less than 50 mM NaCl, and concentrated for the next step. The pooled and concentrated fractions were further purified by perfusion chromatography using a POROS 50HQ column using a strategy identical to that described for the DEAE Sepharose FF column. The fractionation obtained during this step, although modest, is important since the protein obtained from the NaCl gradient contains the material that can be successfully crystallized. Minimal enzyme activity was recovered from material that remains bound to the column and eluted with high salt; such material contains QFR fragments or other denatured proteins (data not shown). The pooled fractions from the center of the peak from the POROS column are subsequently purified on a Sephacryl S-300 gel filtration column. Fractions containing excess detergent aggregates elute slightly earlier than the QFR containing fractions as confirmed by passing detergent alone through the column and analysis for the presence of detergent (data not shown). The dark brown QFR containing fraction was 54 pooled, concentrated, filtered by passage through 0.2 µm nylon filters and finally concentrated using sterile Centiprep-30 concentrators to a final concentration of 30-40 mg protein/ml. Characterization of QFR. Protein determination was performed according to Markwell et al., (25) . The enzyme was judged to be >95% pure according to native PAGE and SDS-PAGE using both the methods of von Jagow/Schagger and Laemmli (26,27). The homogeneity of the QFR preparation was also checked by isoelectrofocusing. Protein was stained using coomassie blue R-250 and gels were permanently stained and fixed with 5% (w/v) acetic acid. Enzyme activity was measured as previously described (28). Endogenous lipids were measured by the method of Bartlett (29). Fluorescence spectroscopy. Relative changes in the environment of the tryptophan residues of QFR were followed by monitoring variation in the maximum emission wavelength (AmaJ as previously described (30) using a Hitachi F-4010 fluorescence spectrophotometer. The excitation wavelength was 286 nm with a bandwidth of 5 nm for both excitation and emission slits. All measurements were done at a concentration of 50 µg protein/ml in 30 mM NaHCO3 , 45 mM NaCl, at pH 8.5. Various concentrations of detergent and guanidinium hydrochloride (GuHCl) were added to the enzyme and then incubated for 72 h at 4 °C prior to measurement of changes in 55 fluorescence. Similar results were seen if the samples were incubated briefly or overnight. Crystallization of QFR. The initial search for suitable crystallization conditions of QFR was performed by vapor diffusion using the Hampton sparse matrix screening kit. A droplet of 5 µl protein and 5 µl reservoir solution were mixed and equilibrated against 0.8 ml of a reservoir solution. The final protein concentration was 30 mg/ml. Several screening solutions gave growth of crystals ; however, the crystals that were optimized from starting conditions containing 12% (w/v) PEG 12K and 100 rnM sodium citrate, pH 5.8 eventually displayed the highest resolution diffraction. Crystallization using PEG 12K was very reproducible. Nevertheless, a further screening of precipitant molecular weight was carried out. A screen of different detergent and divalent cation additives produced crystals with varying degrees of resolution. Consequently, hanging drop vapordiffusion experiments were performed using 12-20% (w/v) PEG in the average molecular weight range of 4-20K. The crystallization conditions resulting in the best diffracting crystals were buffered using sodium citrate in the pH range of 5.8-6.2 and included 85 mM magnesium acetate, 0.1 mM EDTA and 60 µM dithiothreitol. Under these conditions, crystal growth can be observed overnight and mature within 3-10 days. For cryoprotection experiments crystals were soaked in the mother liquor solutions containing 30% (v/v) ethylene glycol. RESULTS AND DISCUSSION 56 Production of recombinant QFR. Expression of the membrane-bound fumarate reductase has been amplified more than 30-fold over native levels of the enzyme utilizing a recombinant plasmid carrying the frdA+B+c+D+ operon as has been previously described (12,28). The enzyme can account for up to 60% of the total crude membrane protein fraction when cells are allowed to grow until late stationary phase (12). This very high level of overexpression has been shown to result in the formation of tubular structures branching from the cytoplasmic membrane that are composed of the four subunit QFR complex and phospholipid, primarily cardiolipin (31 ). A growth profile for E. coli cells overexpressing QFR (Fig. 2. lA) is ·shown with the hatched region indicating where cells can be harvested and the resulting isolated enzyme proving amenable to crystallization. Although intracellular lipid-protein tubules enriched in QFR are produced in late-log phase (Fig. 2.lB.1) their concentration is less than in stationary phase cells (Fig. 2. lB.2). SDS-PAGE and enzyme specific activity analyses of QFR produced from late stationary phase cells indicates that the enzyme has become partially degraded (Fig. 2.2). A specific proteolytic fragment of the flavoprotein (Fp) subunit can be identified of approximately 50 kDa. A smaller fragment of approximately 15 kDa contains covalently bound flavin. Analysis of the larger peptide showed that proteolytic cleavage occurs between a pair of arginine residues (residues 114 and 115) in the Fp sequence. The concentration of this proteolytic peptide increases the later in stationary phase the cells were harvested (Fig. 2.2). In order to minimize the in vivo proteolysis of 57 QPR it was necessary to harvest cells in log phase or as they approached stationary phase. This approach yielded enzyme with the highest specific activity (Table 2.1). A significant decrease in the microheterogeneity and proteolytic nicking of the enzyme was observed when the membranes were prepared from fresh rather than frozen cells harvested at late log phase. The observations are consistent with those of others showing that protein microheterogeneity can occur as a result of protein degradation during growth or purification (32). Figure 2.2 shows an SDS-PAGE profile of membranes prepared from late stationary phase (lane 2) and the late log phase membranes (lane 3) that proved best for enzyme that was amenable to crystallization (lane 7) . The preparation of inner-membranes from late-log phase cells yielded a membrane fraction that contained 50-60% of its protein as QPR (Table 2.1, Fig. 2.2). Extraction of the enzyme from the membrane using the non-ionic detergent Thesit readily solubilizes the enzyme complex in a stable form as determined by enzyme activity measurements. Purification and characterization of QFR. QFR from the Thesit extract was purified by a combination of ion exchange and gel filtration chromatography. The Thesit extraction is done using 2% (w/v) detergent, however, during elution of the enzyme from the DEAE Sepharose FF column the concentration of Thesit which results in the best resolution of QPR is 0.05% (Fig. 2.3A). The specific fumarate reductase/succino- oxidase activity of the QPR eluted from column both to the left and right of the arrows shown in Fig. 2.3A was very similar to the specific activity between the arrows, however, 58 only the protein fractions located between the arrows resulted in enzyme that was amenable to crystallization. On this basis it appears that specific enzyme activity is not a key criterion for discrimination of samples with the ability to crystallize. The material eluting between the arrows in Fig. 2.3A was then subjected to a second ion exchange step using perfusion chromatography media (POROS 50HQ) . Fig. 2.3B shows that this perfusion chromatography step separates the enzyme into three major fractions with the major peak located between the arrows providing enzyme suitable for crystallization. The peak eluting at 1 M NaCl contains material that is less pure (observed by SDSPAGE) and is rich in endogenous lipid (data not shown). Trials with other purification resins indicated that this step using the perfusion media was required in order to obtain enzyme that crystallizes. The QFR protein from the central peak in Fig. 2.3B was then concentrated and applied to a Sephacryl S-300 gel filtration column. The resultant elution profile (Fig. 2.3C) shows a major protein peak with shoulders on both sides. The material in the shoulders appears to be degraded enzyme based on SDS-PAGE. The central fraction, however, produces protein that is highly active (Fig. 2.4) and amenable to crystallization. Figure 2.4 also shows that the initial anion exchange chromatography step removes most of the endogenous lipids associated with the protein complex and they are apparently completely removed by the final gel filtration step. The data show that considerable amounts of phospholipids are present in the enzyme complex during the extraction process, however, they are removed during the subsequent purification 59 process. The enzyme activity of isolated QFR (Fig. 2.4) does not appear dependent upon the presence of phospholipids. It was also found that the purified enzyme would not crystallize if it contained significant amounts of the native lipids. The purified QFR complex was found to be >95 % pure based on densitometer analysis of SDS-PAGE, native gel electrophoresis and isoelectric focusing (data not shown). The purified enzyme shows a single band during native gel electrophoresis, which was found necessary for subsequent crystallization. The purified QFR complex following extensive dialysis, to remove excess detergent present following concentration of the protein, was found to have pl of 4.8. The purified enzyme was evaluated as to its stability by determining effects of various detergents and the chaotrope guanidinium hydrochloride on enzyme activity and conformation of the enzyme based on tryptophan fluorescence. As shown in Fig. 2.5 the purified QFR complex is much more stable in guanidinium hydrochloride than is the soluble domain (FrdAB) (28 and manuscript in preparation) of fumarate reductase. The FrdCD hydrophobic peptides stabilize the QFR complex and appear to protect the [3Fe4S] cluster from the oxidative denaturation of the enzyme that occurs when this ironsulfur cluster is destroyed (1,12,28). The slow unfolding of QFR is indicative of irreversible non-cooperative denaturation due to strong hydrophobic interactions in the complex (33) . Table 2.2 shows the change in tryptophan fluorescence maximum wavelength O"maJ (30) of QFR as a function of different detergents. Moderate unfolding 60 of the complex is observed using 0.05% Thesit as is the case for the majority of other detergents. As shown in Table2.2, the specific activity of QFR is enhanced by the presence of Thesit in the assay mixture. By contrast the detergents, octylglucoside, CHAPS, and sodium cholate all inhibited enzyme activity to varying extents while 0.2% (w/v) SDS inactivated the enzyme almost completely. Similar effects on protein conformation and enzyme activity have been shown for membrane-bound proteins in the presence of detergents or chaotropes (34,35). Crystallization of QFR. Crystallization trials using standard screens in vapordiffusion experiments yielded a number of conditions producing QFR crystals. As shown in Fig. 2.6, the crystals were dark brown, birefringent, and assumed a uniform shape with distinct edges and vertices. For all protein preparations, crystals grown in 100 mM sodium citrate, pH 5.8, 85 mM magnesium acetate, 0.1 mM EDTA, and 60 µM dithiothreitol provided the highest resolution diffraction. In contrast, the molecular weight of the polyethylene glycol that grew the best diffracting crystals varied from preparation to preparation. A concentration of 30 mg protein/ml in the final droplet was also required for optimal crystal growth. SDS-PAGE analysis of serially washed crystals showed four distinct bands of the correct molecular size representative of the FrdABCD subunits. Crystal growth was apparent within 24 hours and crystals 24-72 hours old gave the best diffraction. Crystals used for the initial structure determination were obtained 61 from crystals grown from l0K PEG with ethylene glycol as cryoprotectant and immediate cryo-cooling in liquid nitrogen. The QFR structure has been solved to 3.3 A using these crystals with the space group symmetry P2 12 12 1 with a=96.6 A, b= 138.1 A, and c=275.3 A (21). The presence of Thesit in concentrations of 140-fold above its cmc does not prevent growth of crystals as has been shown for other proteins (36). The concentration of micelles in the solvent was about 0.6 per protein molecule (mol/mol) with no excess of "empty" micelles (37) that might otherwise inhibit the growth of crystals. At least one molecule of Thesit is found per QFR monomer from the structure of the protein complex (21). The refined structure suggests that the FrdAB soluble domain constitute the largest part of the structure as compared to the detergent coated hydrophobic (FrdCD) domain (21) . The QFR complex essentially may behave as a soluble protein with the colloidal state becoming crucial when upon assembly of the crystal lattice the detergent coats come into close contact (38). This arrangement is atypical of either type I or type II crystals (39,40). Unlike prostaglandin H synthase (41), where the relative protein mass located on the cytoplasmic side of the membrane is 90% and in bacteriorhodopsin is as low as 10%; in QFR 75% of the complex is a rather extensive globular domain projecting from the membrane to make the necessary interlayer interactions in the crystal lattice. 62 The technique described above for optimizing preparations of highly pure QFR and its crystallization has been highly reproducible. Although variation of crystallization conditions yielded various crystal shapes, the space group remained the same. The isolated and crystallized enzyme show the same subunit stoichiometry and contains all types of redox cofactors known to be present in E. coli QFR (1-3) including two molecules of menaquinone. The general procedures described above may be applicable to the isolation and crystallization of other membrane-bound QFR's and SQR's. ACKNOWLEDGMENTS This work was supported by the Department of Veterans Affairs, and grants from the NIH (HL-16251 and GM45162), the NSF (MCB-9728778), and the Howard Hughes Medical Institute. T.M.I. is supported by an NIH training grant. C.L.C. would like to acknowledge the support and continuing mentorship of Professor Alex V. Nichols and his helpful discussions and suggestions and Dr. Ciaran N. Cronin for useful discussions. 63 Figure 2.1: Optimization of bacterial growth . A. E. coli strain DW35 transformed with the QFR encoding plasmid pH3 was grown anaerobically as described in Materials and Methods. The hatched regions between points II and III produced cells containing QFR that were amenable to crystallization. Growth of cells beyond point III although producing even greater quantities of QFR results in enzyme that does not crystallize. B. Thin section electron micrographs of longitudinal cross sections of E. coli DW35 transformed with pH3 grown to (B 1) late log phase and (B2) late stationary phase. The tubules produced in the cell are composed almost entirely of QFR and cardiolipin as shown in Refs.12 and 31. a I I[ Ill + + + 1.0 0 0 IO ~ 0.5 0 0 16 hours m 48 64 Figure 2.2 : SDS-PAGE analysis of E. coli QFR enriched membranes and purified protein. SDS-PAGE were performed with 10-20% polyacrylamide gels and stained with Coomassie brilliant blue. The amount of protein loaded ranged from 25 to 30 µg for lanes 2 and 3 and 10 to 15 µg for lanes 4-7. Lane 1, molecular markers; lane 2, membrane preparations from late stationary phase; lane 3, membranes from late log phase; lane 4, Thesit extract; lane 5, DEAE Sepharose FF fraction; lane 6, POROS 50HQ chromatography fraction; lane 7, purified QFR from Sephacryl S-300 column. 12 94kDa ~ 67kDa ~ 43kDa ~ 30kDa ~ 34567 +- FrdA ..._ FrdB 20.lkD ~ 14.4kD ~ +- FrdC/D 65 Chromatograms of Thesit solubilized QFR. A. DEAE Sepharose FF Figure 2.3: chromatography showing effect of different concentrations of Thesit used in elution buffer. Equivalent amounts of protein were loaded and elution carried out at 4 °C with 0% (dashed line, ---), 0.05% (solid line, _), and 0.8% (dotted line, ...... ), (w/v) Thesit included in the elution buffer. The gradient of 100-250 mM NaCl is indicated by the hatched line. The QFR elutes at approximately 125-175 mM NaCl. The solid arrows in all chromatograms (~) indicate the beginning and end points of the fractions of QFR that were isolated and pooled for subsequent steps. B. Perfusion chromatography with PO ROS 50HQ ion exchange of the isolated QFR from the previous step. Conditions as described in Materials and Methods. C. Sephacryl S-300 gel filtration chromatography of QFR from step B. The area between the arrows indicates QFR that is amenable to crystallization. l.00 0.7! 0.50 z. (')· = ~ 0;25 0.00. Ma z. .Q ~ 2 ◄ C Vo \ 66 Figure 2.4: Phospholipid concentration and QFR activity during purification. CM refers to membrane fraction ; TE, Thesit extract; and DEAE, POROS, and Sephacryl, refer to the respective chromatography steps. The open rectangles (D) indicate enzyme activity and the solid diamonds ( ♦) indicate the percent of the original fraction of lipids remaining. The lipid concentration and enzyme activity measurements were as described in Methods. 100 ..... ..... .... 25 ..... ..... ..... 20 Q) ~ \ Q) l,.i "'0 15 \ ··- 40 \ - ~ \ ~ \ 10 5 -s \ 0 \ \ \ \ :::::l • --- - - :::+ - - .- .-- -........... ................................................ 0 TE u ::s r,J Cl.I \ 20 .... ...= u \ 0 ~ 0 Cl.I \ Q. ...... "O \ 60 "O ----Cl.I N \ CM s .= .. s ----- = 80 Q. oJl DEAE .. ..- ....... .... .. POROS .. SEPHACRYL 0 67 Figure 2.5: Effect of guanidinium hydrochloride (GuHCl) on tryptophan fluorescence maximum emission wavelength (A111 a) (30). Soluble fumarate reductase (FrdAB subunits) (28) is shown by the solid circle (• ). QFR is indicated by the open circle (o ). The concentration used during analysis was 50 µg protein/ml in both cases. The final GuHCl concentration in the samples was obtained by addition of 8 M GuHCl. Samples were incubated for 72 h at 4 °C prior to fluorescence measurement. Excitation wavelength was 286 nm. 354 ,-..... e c)50 ~ 9346 «· 342 338 334 0 1 2 3 4 5 GuHCl(M) 6 7 8 68 Figure 2.6: Crystals obtained from purified QFR. Protein concentration was 30 mg/ml. Crystals were grown for 48 h in 100 mM sodium citrate, pH 5.8, 85 mM magnesium acetate, 0.1 mM EDTA, and 60 µM dithiothreitol with 10 kDa polyethylene glycol as the precipitating agent. 69 Table 2.1 Purification of QFR Complex from Escherichia col? Total Protein (mg) Total Activitl Specific Activity Foldc µmol/min/mg purific. µmol/min Membranes 742 4,600 9 2-3 100 Thesit extract 384 2,600 20 4.4 56 DEAE Seph FF 129 2,300 21 4.7 49 POROS 50HQ 112 2,300 22 4.9 49 220 22 4.9 5 Sephacryl S-300 20 % yield a 90 g wet weight of anaerobically grown E. coli DW35 containing plasmid pH3. b Enzyme activity was determined by following the succinate:phenazine ethosulfate (PES) reaction in the presence of dicholorphenolindophenol (DCIP) (E600 = 21.8 mM- 1 cm1) using 1.5 mM PES and 50 µM DCIP at 37 °C as previously described (28). c Washing of membranes to remove light brownish material prior to extraction improves the initial yield of QFR some 2-3 fold, thus the initial purification is greater than 1-fold. 70 Table 2.2 Effect of detergentsa on enzyme activity and fluorescence emission. Detergent Type % Specific cmc (L1Amax? Activityc (mM/ Nonef O 25±3 Micellar M/ 0.05% Thesit 14 33±2 0.1 -64,000 0.1 % Triton X-100 n.d. 22±1 0.3 -90,000 0.09% Dodecylmaltoside 4 27±1 0.2 -50,000 0.8% Octylglucoside 13 4±1 15 8,800-29,200 0.5% CHAPS 16 9±2 4-6 -6,200 0.6% Sodium cholate 16 13±2 3-10 800-2,000 0.2% SDS 20 0.2±0.1 8.1 17,000-28,800 a The concentration of detergent used was 3-10 fold above their corresponding cmc. b Relative change of "-max at 286 nm of a 50 µg /ml sample, calculated as % change to QFR in the absence of detergent. Detergent was incubated with sample for 72 h at 4 °C. c Specific activity determined as in Table 2.1. d The data for cmc are taken from tables reported by Roche Molecular Biochemicals. e The average micellar molecular weights are as reported by Roche Molecular Biochemicals. f Although QFR is purified in the presence of Thesit no additional detergent was added to sample prior to measurement. n.d. = not determined 71 REFERENCES 1. Ackrell, B.A.C. , Johnson, M.K. , Gunsalus, R.P. , and Cecchini, G. (1992) Structure and function of succinate dehydrogenase and fumarate reductase in "Chemistry and Biochemistry of Flavoenzymes" (Muller, F., Ed.), pp. 229-297, CRC Press, Inc ., Boca Raton. 2. Hagerhall, C. (1997) Succinate:quinone oxidoreductases; variations on a conserved theme. Biochim. Biophys. Acta 1320, 107-141 . 3 . Hagerhall, C., and Hederstedt, L. (1996) A structural model for the membraneintegral domain of succinate:quinone oxidoreductases. FEBS Lett. 389, 25-31. 4 . Ruzicka, F.J., Beinert, H., Schepler, K.L., Dunham, W.R., and Sands, R.H. (1975) Interaction of ubisemiquinone with a paramagnetic component in heart tissue. Proc. Natl. Acad. Sci. USA 72, 2886-2890. 5. Yu, C.-A., and Yu, L. (1982) Specific interaction between protein and ubiquinone in succinate-ubiquinone reductase. J. Biol. Chem. 257, 6127-6131. 6. Yang, X., Yu, L., He, D., and Yu, C.-A. (1998) The quinone-binding site in succinate-ubiquinone reductase from Escherichia coli. J. Biol. Chem. 273, 3191631923. 72 7. Westenberg, D.J., Gunsalus, RP., Ackrell, B.A.C., Sices, H., and Cecchini, G. (1993) Escherichia coli fumarate reductasefrdC andfrdD mutants. J. Biol. Chem. 268, 815- 822. 8 . Yankovskaya, V., Sablin, S.O. , Ramsay, R.R., Singer, T.P., Ackrell, B.A.C ., Cecchini, G., and Miyoshi, H. (1996) Inhibitor probes of the quinone binding sites of mammalian complex II and Escherichia coli fumarate reductase. J. Biol. Chem. 271, 21020-21024. 9 . Ziegler, D.M., and Doeg, K.A . (1959) Studies on the electron transport system: XLIII. The isolation of a succinate-coenzyme Q reductase from beef heart mitochondria. Biochem. Biophys. Res. Commun. 1, 344-349. 10. Unden, G., Hackenberg, H., and Kroger, A. (1980) Isolation and functional aspects of the fumarate reductase involved in the phosphorylative electron transport of Vibrio succinogenes. Biochim. Biophys. Acta 591, 275-288. 11.Singer, T.P., and Johnson, M.K. (1985) The prosthetic groups of succinate dehydrogenase: 30 years from discovery to identification. FEBS Lett. 190, 189-198. 12.Lemire, B.D., and Weiner, J.H. (1986) Fumarate reductase of Escherichia coli. Methods Enzymol. 126, 377-386. 13. Kita, K., Vibat, C.R., Meinhardt, S., Guest, J.R., and Gennis, RB. (1989) One-step purification from Escherichia coli of complex II (succinate:ubiquinone 73 oxidoreductase) associated with succinate-reducible cytochrome b 556 . J. Biol. Chem. 264, 2672-2677. 14 . Iwata, S., Ostermeier, C. , Ludwig, B., and Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376, 660669. 15 . Tsukihara, T. , Aoyama, H. , Yamashita, E. , Tomizaki, T., Yamaguchi, H., Shinzawaltoh, K., Nakashima, R. , Yaono, R., and Yoshikawa, S. (1996) The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. Science 272, 1136-1144. 16. Yoshikawa, S. , Shinzawa-ltoh, K., Nakashima, R., Yaono, R., Yamashita, E., Inoue, N., Yao, M., Fei, M.J. , Libeu, C.P. , Mizushima, T., Yamaguchi, H., Tomizaki, T., and Tsukihara, T. (1998) Redox-coupled crystal structural changes in bovine heart cytochrome c oxidase. Science 280, 1723-1729. 17. Xia, D. , Yu, C.-A. , Kim, H., Xia, J.Z., Kachurin, A.M., Zhang, L., Yu, L. , and Deisenhofer, J. (1997) Crystal structure of the cytochrome bc 1 complex from bovine heart mitochondria. Science 277, 60-66. 18. Zhang, Z., Huang, L., Shulmeister, V.M., Chi, Y.I., Kim, K.K., Hung, L.W., Crofts, A.R., Berry, E.A., and Kim, S.H. (1998) Electron transfer by domain movement in cytochrome bc 1. Nature 392, 677-684. 74 19 . Iwata, S., Lee, J.W., Okada, K., Lee, J.K., Iwata, M., Rasmussen, B., Link, T.A. , Ramaswamy, S., and Jap, B.K. (1998) Complete structure of the 11-subunit bovine mitochondrial cytochrome bc 1 complex. Science 281, 64-71. 20. Stowell, M.H. and Rees, D.C. (1995) Structure and stability of membrane proteins. Adv. Protein Chem. 46, 279-311. 21. Iverson, T.M., Luna-Chavez, C. , Cecchini, G., and Rees, D.C. (1999) Structure of the Escherichia coli fumarate reductase respiratory complex. Science 284, 1961 -1966. 22 .Lancaster, C.R.D., Kroger, A., Auer, M., and Michel, H. (1999) Structure of fumarate reductase from Wolinella succinogenes at 2.2 A resolution. Nature 402, 377-385. 23 .Blaut, M. , Whittaker, K. , Valdovinos, A., Ackrell, B.A.C., Gunsalus, R.P., and Cecchini, G. (1989) Fumarate reductase mutants f of Escherichia coli that lack covalently bound flavin. J. Biol. Chem. 264, 13599-13604. 24. Spencer, M.E., and Guest, J.R. (1973) Isolation and properties of fumarate reductase mutants of Escherichia coli. J. Bacterial. 114, 563-570. 25 . Markwell, M.A., Haas, S.M., Bieber, L.L., and Tolbert, N.E. (1978) A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples. Anal. Biochem. 87, 206-210. 26. von Jagow, G., and Schagger, H. (1994) in "A practical guide to membrane protein purification." Academic Press, San Diego, CA. 75 27 . Laemmli, U .K. ( 1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. 2 8 .Cecchini, G ., Ackrell , B.A.C. , Deshler, J.O., and Gunsalus, R.P. (1986) Reconstitution of quinone reduction and characterization of Escherichia coli fumarate reductase activity. J. Biol. Chem. 261, 1808-1814. 29 . Bartlett, G.F. (1959) Phosphorus assay in column chromatography. J. Biol. Chem. 234, 466-468 . 30. Luna-Chavez, C., Gong, E.L. , Forte, T.M. , and Nichols, AV. (1994) Sodium oleatefacilitated reassembly of apolipoprotein A-I with phosphatidylcholine. Biochim. Biophys. Acta 1215, 141-149. 31. Weiner, J.H., Lemire, B.D., Elmes, M.L., Bradley, R.D., and Scraba, D.G. (1984) Overproduction of fumarate reductase in Escherichia coli induces a novel intracellular lipid-protein organelle. J. Bacterial. 158, 590-596. 32.Park, J.-1., Grant, C.M., Davies, M.J., and Dawes, I.W. (1998) The cytoplasmic Cu,:&i superoxide dismutase of Saccharomyces cerevisiae is required for resistance to freeze-thaw stress. J. Biol. Chem. 273, 22921-22928. 3 3. Yeates, T. (1993) in "Thermodynamics of Membrane Receptors and Channels" (Jackson, M.D., Ed.), pp. 1-25, CRC Press, Inc., Boca Raton. 76 34 . Yun, C.-H., Song, M. , and Kim, H. (1997) Conformational change of cytochrome P450 1A2 induced by phospholipids and detergents. J. Biol. Chem. 272, 1972519730. 35 . Ushakova, A.V., Grivennikova, V.G., Ohnishi, T. , and Vinogradov, A.D. (1999) Triton X-100 as a specific inhibitor of the mammalian NADH-ubiquinone oxidoreductases (complex I). Biochim. Biophys. Acta 1409, 143-153. 36. Garavito, R.M., Picot, D., and Loll, P.J. (1996) Strategies for crystallizing membrane proteins. J. Bioenerget. Biomemb. 28, 13-27. 37 . Song, L., and Gouaux, J.E. (1997) Membrane protein crystallization: Application of sparse matrices to the a-hemolysin heptamer. Methods Enzymol. 276, 60-74. 38. Zulauf, M. (1990) in "Crystallization of membrane proteins" (Michel, H., ed.), pp. 53-72, CRC Press, Boca Raton. 39. Michel, H. (1983) Crystallization of membrane proteins. Trends Biochem. Sci. 8, 5659. 40. Ostermeier, C., and Michel, H. (1997) Crystallization of membrane proteins. Curr. Op. Str. Biol. 7, 697-701. 41. Picot, D., Loll, P.J., and Garavito, R.M. (1994) The x-ray crystal structure of the membrane protein prostaglandin H2 synthase-1. Nature 367, 243-249. 77 Chapter 3 Reprinted with permission from : Iverson, T.M., Luna-Chavez, C., Cecchini, G., Rees, D.C. (1999) Structure of the Escherichia coli Fumarate Reductase Respiratory Complex, Science 284: 1961-1966. Copyright 1999, American Association for the Advancement of Science 78 RESEARCH ARTICLES Structure of the Escherichia coli Fumarate Reductase Respiratory Complex Tina M. lverson, 1 Cesar Luna-Chavez.2 Gary Cecchini,2* Douglas C. Rees3 * The integral membrane protein fumarate reductase catalyzes the final step of anaerobic respiration when fumarate is the terminal electron acceptor. The homologous enzyme succinate dehydrogenase also plays a prominent role in cellular energetics as a member of the Krebs cycle and as complex II of the aerobic respiratory chain. Fumarate reductase consists of four subunits that contain a covalently linked flavin adenine dinucleotide, three different ironsulfur clusters, and at least two quinones. The crystal structure of intact fumarate reductase has been solved at 3.3 angstrom resolution and demonstrates that the cofactors are arranged in a nearly linear manner from the membrane-bound quinone to the active site flavin. Although fumarate reductase is not associated with any proton-pumping function, the two quinones are positioned on opposite sides of the membrane in an arrangement similar to that of the Q-cycle organization observed for cytochrome be,. Because oxygen has a high affinity for electrons. aerobic respiration represents a very favorable forrn of energy metabolism. However, in the absence of oxygen, many microorganisms can obtain energy through anaerobic respiratory processes that result in the reduction of alternate terminal acceptors (]). One of the most widespread acceptors is fumarate (2), which is reduced to succinate by fumarate reductase, an integral membrane protein containing flavin adenine dinucleotide (FAD) and iron-sulfur clusters (3). The electron donor for this reaction is reduced menaquinone, which commonly serves as a membrane-soluble, mobile electron carrier between respiratory complexes. The most extensively characterized fumarate reductase, from Escherichia coli, has a total molecular mass of 121 kD in four subunits. It consists of two water-soluble subunits, the flavoprotein (66 kD) and iron-sulfur protein (27 kD) subunits, and two membrane anchor subunits ( 15 and 13 kD ), which are the products of the .fi·dABCD genes, respectively (4 ). The flavoprotcin (Fp) subunit contains the catalytic site for fumarate reduction and succinate oxidation at a covalently linked FAD 'Graduate Option in Biochemistry, 147-75CH, California Institute of Technology, Pasadena, CA 91125, USA. 'Molecular Biology Division, Department of Veterans Affairs Medical Center, San Francisco, CA 94121, and Department of Biochemistry and Biophysics, University of California, San Francisco, CA 94143, USA. 'Howard Hughes Medical Institute, Division of Chemistry and Chemical Engineering, 147-75CH, California Institute of Technology, Pasadena, CA 91125, USA. *To whom correspondence should be addressed. Emait ceccini@itsa.ucsf.edu: dcrees@caltech.edu (5 ), while the iron-sulfur protein subunit (Ip) contains three different types of ironsulfur clusters, [2Fe:2S)., [4Fe:4S], and [3Fe:4S], which have been spectroscopically characterized ( 6). At least two sites associated with the membrane anchor subunits have been proposed to bind the quinones Iha! arc involved in electron transfer reactions of the enzyme (7). Fumarate reductase catalyzes the reverse reaction of succinate dehydrogenase, which participates in both the aerobic respiratory chain as complex II and in the Krebs cycle (3). These two proteins exhibit substantial similarities in amino acid sequence, cofactor composition, and mechanism. Indeed, under certain conditions, one enzyme can functionally replace the other and support bacterial growth ( 8). Because of the central role of fumarate reductase and succinate dehydrogenase in respiration, mutations in these complexes can have substantial metabolic consequences. In bacteria, mutations in fumarate reductase can significantly retard growth under appropriate conditions (9). In higher organisms, mutations of succinate dchydrogenase have been linked lo oxidative stress and aging in nematodes (/0) and to Leigh's syndrome in humans (I/). Historically, succinate dehydrogenase was one of the most widely studied enzymes during the development of enzymology. Early studies resulted in the discoveries of nonheme iron and covalently bound flavin in proteins (/ 2). To provide a framework for addressing the functional properties of fumarate reductasc and suceinate dchydrogenase, we have solved the strucnire of the £. coli fumarate reductase at 3.3 A resolution. Here we describe the fold of the polypeptides and location of the cofactors, and the functional implications of this structural arrangement. Structure Determination and Overall Fold Fumarate reductase from £. coli was purified and crystallized in the presence of the nonionic detergent Thesit ( / 3). The structure was solved by multiple wavelength anomalous diffraction (MAD), with data collected at three wavelengths near the Fe K edge (14) (Table I). The iron-sulfur clusters and transmembrane helices were striking in the initial maps calculated at 4 A resolution, and the structure was sol vcd by iterati vc combination of density modification, noncrystallographic symmetry averaging, model building, and refinement (/ 5). The final model has been refined to values of Rccyso of22.2% and R,'ccc of 29.2% at 3.3 A resolution with reasonable stcrcochcmistry (/ 5, I 6). The four subunits in fumarate rcductase are arranged in a complex resembling the letter "q," with the top of the "q" generated by the Fp and Ip subunits (diameter -70 A), while the tail of the "q" (length 110 A) contains the membrane anchor subunits (Fig. I, A and B). The orientation of fumarate reductase in the cell membrane is such that the Fp and Ip subunits are located in the cytoplasm (equivalent to the mitochrondrial matrix for succinatc dehydrogenasc). In these crystals, two fumarate rcduclase complexes, which are related by a twofold axis approximately parallel to the membrane normal, are present per asymmetric unit. These two complexes associate through their transmembrane regions. Contacts with neighboring molecules related by crystallographic symmetry also occur in the membrane-spanning region, creating a continuous membrane-spanning region throughout the crystal (Fig. IC). Despite the suggestiveness of this arrangement, there is no evidence that a dimer is physiologically relevant, unlike the situation with cytochrome be, (/7). Additionally, the contact region between fmnarate reductase molecules in the crystals is relatively small (-325 A2 ) (/8) and is unlikely to support formation of a stable dimer (Fig. 1B). The Fp (FrdA) subunit is organized around an FAD/NAD(P) (nicotinamidc adenine dinucleotide phosphate) binding domain formed by residues Al to 11.50, 11.130 to A231, and A354 to A4l4 (Fig. 2A). This domain sttucture includes a Rossmann-type fold that provides the binding site for FAD. The FAD is further associated with the flavoprotcin through a covalent bond between the flavin C8A methyl group and the Ne atom of the side chain of His A44. The remaining residues of this subunit, AS I to Al 29, A232 to A353, and A415 to A575, are inserted into www.sciencemag.org SCIENCE VOL 284 18 JUNE 1999 19 79 RESEARCH ARTICLES Fig. 1. Structure of fumarate reductase. (A) Stereoview of a fumarate reductase monomer. The flavoprotein is in blue, the iron protein is in red, and the membrane anchors are in green (FrdC) and purple (FrdD). The [Fe:S] clusters are shown as purple (Fe atoms) and yellow (S atoms), while the menaquinones and FAD are shown in yellow. (B) Space-filling model of fumarate reductase showing the crystal contacts between the two complexes in the membrane-spanning portion. In the fumarate reductase complex, oxygen atoms are shown in red, nitrogen atoms are shown in blue, sulfur atoms are shown in yellow. Menaquinone molecules are shown in magenta (see right-hand monomer). The location of the membrane-spanning region can be inferred from the coloring of the atoms In this representation. The more hydrophilic (soluble) region contains many polar oxygen and nitrogen atoms (red and blue) while the hydrophobic (membrane-spanning) region contain mostly apolar carbon atoms (gray). (C) Crystal packing of fumarate reductase Is through the transmembrane regions of the protein (green and purple) and forms a continuous membrane-spanning portion In the crystal (36). In this representation, the FAD and menaqulnone are shown In gray (37) . A B 110A 1962 18 JUNE 1999 VOL 284 SCIENCE www.sciencemag.org 80 RESEARCH ARTICLES the core FAD binding domain and adopt compact folds that do not exhibit significant structural similarities to known fold s in the Protein Data Bank, as assessed by the DALI server (19). Fumarate reductase contains three ironsulfur clusters that are coordinated by cysteine residues in the Ip (FrdB) subunit as follows: [2Fc:2S] (Cys residues B57, B62, B65, and B77); [4Fc:4S] (Cys B148, B151, B154, and B214); and [3Fc:4S] (Cys B158, B204, and B210). Consistent with the conclusions of sequence and electron paramagnetic resonance (EPR) analyses (3, 6), the Ip subunit is organized into t\vo domains (Fig. 2B), one characteristic of [2Fc:2S]-containing ferredoxins (residues BI to B9 I) and the other characteristic of bacterial ferredoxins that contain [3Fe:4S] or [4Fe:4S] type clusters (residues B145 to B221). The [2Fe:2S] domain superimposes closely with plant-type fcrrcdoxins, while more substantial changes have occurred in the bacterial fcrrcdoxin domain. Although bacterial ferredoxins containing both a [3Fe:4S) and a [4Fe:4S) cluster are relatively common, the cysteine ligands for the [3Fe:4S) cluster always appear first in the protein sequence. As anticipated from sequence analyses, the cluster arrangement in fumarate reductase is reversed relative to these fcrredoxins. In addition, die four-stranded antiparallel 13 sheet found on one side of the clusters in ferredoxins has been replaced by a helical hairpin in fumaratc reductase. The two membrane anchor subunits, FrdC and FrdD, exhibit similar folds, each with three transmembrane helices connected by extra-membrane loops (Fig. 2C). These helices, designated I to VI (3), consist of residues C22 to C49, which is actually composed of two kinked, helical segments, C66 to C90, CI 05 to C 128, D9 to D35 , D61 to D89 , and D97 to D 115, and are in reasonable agreement with transmembrane segments predicted by hydropathy analysis and mutagenesis (20) . Helices I, II, IV, and V are tilted -3 0° to 40° from the membrane normal, as defined by the dimer t\vofold axis, and are arranged in a right-handed, helical bundle, with helix crossing angles of - 120°. In contrast, helices III and VI arc more parallel to the membrane normal with a tilt of - 10° to 25°. The NH 2 - and COOH-tcrrnini arc on opposite sides of the membrane-spanning region, which correspond to the cytoplasm and periplasm, respectively. The overall arrangement is such that the t\vo subunits could be cova lently connected upon deletion of helix 111, consistent with the observation that homologous enzymes have been identified that contain only a single transmembrane domain with five membranespanning helices (20). Two menaquinone molecules, which are located on opposite sides of the membranespanning region, arc present in this fumaratc reductase structure. The menaquinone (Q,,) positioned proximal to the [3Fc:4S] cluster of the Ip subunit binds in a relatively polar pocket formed by helices I, II, IV, and V, while the second menaquinone (Q 0 ) distal to this cluster, is positioned - 27 A from the first and binds in a relatively hydrophobic pocket near the other ends ofhcliccs I, II, IV, and V. Site-directed mutagcncsis and labeling with azido-quinones have implicated residues in both these regions as involved in quinone binding in both fumarate reductase and succinate dehydrogenase (9, 21). Although evidence suggests that complex II contains a stabilized semi"quinone pair in close proximity to the [3Fc-4S) cluster with the quinone rings perpendicular to the membrane plane (7, 22), only single quinone molecules were identified at the spatially distinct Qr and Q 0 sites in the fumarate reductasc structure. In the absence of a conformational rearrangement of the protein in the Qp region, it does not appear that this site can accommodate more than one quinone molecule. Redox Centers and Electron Transfer Pathway The six redox cofactors offumarate reductase arc organized into a chain with the sequence FAD-[2Fe:2S]-[4Fe:4S]-[3Fe:4S]-Qp-Q 0 (Fig. 3). With the exception of the -27 A spacing between the t\vo menaquinones, the redox cofactors are all separated by - I I to 14 A center-to-center distances, which are common cofactor separation distances observed in multiccntcrcd electron transfer proteins. In its physiologically relevant reaction, electrons enter fumaratc rcductasc in the form of reduced mcnaquinonc, although the enzyme Table 1. Summary of data collection and refinement statistics. Numbers in parentheses indicate values for the highest resolution bin. The figure of merit was 0.48 for all data to 4.0 A. R,.,m= l Ji, - (/)]/l[(l)i. Res, resolution {in angstroms); URef, number of unique reflections; and Red, redundancy. (PP) denotes overall phasing power to 4 A resolution. Completeness (%) Energy (keV) f' f" Res URef Red 7.500 {high remote) 7.127 {peak) 7.120 {inflection) - 2.2 - 6.6 - 4.2 3.6 4.5 2.7 3.3 3.5 4.0 49,332 39,139 31,125 4.4 87.2 (90.0) 0.093 (0.277) 5.3 94.1 (91.0) 0.096 (0.236) 4.1 97.2 (97.1) 0.118(0.269) Rsym lier (PP) 15.1 (6.3) 14.5 (7.9) 15.8 (10.2) 2.6 2.8 1.6 www.sciencemag.org SCIENCE VOL 284 will physiologically reduce ubiquinonc at rates similar to native succinatc dehydrogcnasc (8). Examination of a space-filling model of fumarate reductase indicates that both quinone binding sites are exposed and should be accessible to the exterior of the complex (Fig. 1B). Electron transfer to die iron-sulfur cluster almost ce1tainly would occur from the Qr site, which is adjacent to the [3Fc:4S] cluster. Although Qr primarily interacts with residues from both membrane anchor subunits, it has limited contact with Ip through Lys B238. The polar environment of the Qr (Fig. 4A) site resembles the Q0 site of bacterial photosynthetic reaction centers (23), which can accommodate all three quinonc oxidation states (reduced, oxidized, and semiquinone). The Q., site has an apolar character (Fig. 4B) that resembles the QA site of photosynthetic reaction centers, which can accommodate only the oxidized and scmiquinonc states. Indeed, residues in the Qr and Q 0 binding pockets had been identified with Q 8 and QA, respectively, based on the consequences of residue substitution ( 7). However, the assignments of quinone oxidation states to specific binding sites during enzyme turnover cannot be unambiguously established at present. During fumarntc reduction, electrons from reduced quinone are transferred to the ironsulfur clusters. Both the oxygen sensitivity upon removal of the membrane anchor subunits and perturbation of the EPR spectrum of the [3Fe:4S] cluster by quinone site inhibitors and mutants suggest that the [3Fe:4S) cluster interacts with the quinone binding subunits to initially accept electrons from the menaquinone (9, 24 ). Spectroscopic studies have indicated that the [2Fe:2S) cluster is in close proximity to the FAD (25), and this cluster is the likely donor of electrons to the flavin . The crystallographic structure demonstrates that His A44, covalently linked to the flavin, intervenes between the flavin and [2Fe:2S) cluster. Although there was speculation that the [4Fe:4S] cluster was "off-pathway," in pa,t due to the low reduction potential of this center, the structure confirms the proposal from EPR studies that the clusters arc arranged in the sequence [3 Fe:4S]-[4Fe:4S]-[2Fc:2S] (6). The consequences of this arrangement for kinetics of electron transfer through complex II have recently been discussed (26) . Many of the residues observed to pa1ticipate in the binding of FAD to fumarate reductase had been previously identified from sequence analysis, and from molecular biological and biochemical studies. The binding site for fumarate can be inferred from the location of oxaloacctatc observed in the structure. Oxaloacctatc is a physiological inhibitor of fumaratc rcductase (inhibition constant K, < I µ,M), and purified preparations of enzyme contain oxaloacetate that remains 18 JUNE 1999 19 81 RESEARCH ARTICLES Fig. 2 (left). Proteins of the fumarate reductase complex. (A) The flavoprotein. Stereoview of the Ca trace of the flavoprotein (blue) with the Rossmann fold highlighted in dark blue. The view is looking down onto the plane of the membrane and rotated 90 A from the view in Fig. 1A. (B) The iron protein. Stereoview of the Ca trace of the iron protein (red) aligned with the 8Fe ferredoxin from Peptococcus aerogenes (light green) (35) and the 2Fe ferredoxin from Spiru/ina platensis (cyan) (34). The nnsd for the Ca atoms in these alignments is 0.8 and 1.6 A, respectively. (C) The membrane anchor proteins. The view is down the center of the fourhelix bundle, approximately normal to the plane of the membrane. FrdC (green) consists of helices I to Ill, and FrdD (purple) consists of helices IV to VI. Fig. 3 (below, right). Cofactor location and pathway of electron transfer. Center-to-center distances between each of the cofactors are indicated. The oxaloacetate was visible in the experimental electron density maps and was positioned to provide the best fit to the electron density and to avoid steric clashes. However. at this resolution, the binding mode for oxaloacetate cannot be unambiguously established. B . .,. .e~ .. •... ...... FAD OAA <: (... .... /4A [2Fe:2S] 7,.A [4Fe:4S] C 7,A [3Fe:4S] /11 A 0 Op ... 21 A 0 1964 18 JUNE 1999 VOL 284 SCIENCE www.sciencemag.org . • .. 82 RESEARCH ARTICLES bound until an excess of substrate is added (27 ). Oxaloacetate interacts with a pocket of arginines and histidines near the side of the flavin opposite to the [2Fe:2S] cluster (Fig. 4C). This group is positioned near conserved residues His A232, Glu A245 , and His A355, which may serve as proton donors, and the N5 of the Jlavin, which likely functions as a hydride donor to fumarate. Although the Jlavin is buried, access to the active site could take place at the interface between two domains of the Fp. A B C .. PheA116 Phe A116 Arg A287 ... HlsA44 HisA44 ArgA287 t · Glu A245 • "' - ·2 ' L ' : A248 o • Glu A245 • ~ 3' A248 Fig. 4. Quinone binding pockets and active site residues. (A) Stereoview of the Qp binding site shows Qp is bound in a polar pocket likely positioned just above the membrane bilayer. (B) Stereoview of the Q 0 site shows Q0 is in a relatively apolar pocket within the membrane bilayer. (C) Binding site for the physiological inhibitor oxaloacetate adjacent to the FAD. Oxaloacetate lies beneath the isoalloxazine ring of the flavin. The flavin ring and inhibitor are shown superimposed onto a 21F0 I - IF,I map contoured at 1cr. The adenine has been omitted for clarity. Side chains that appear to interact directly with the inhibitor are labeled. www.sciencemag.org SCIENCE VOL 284 Implications for Energy Transduction Processes in Respiration Quinones play a central role in respiration because they serve as membrane-soluble electron carriers that can couple proton and electron transfer reactions. In the placement of quinones on opposite sides of the membrane-spanning region, fumaratc rcductasc resembles tl1c arrangement observed in cytochrome be, (17) more than the photosynthetic reaction center (23), where the quinones are on the same side of the membrane. This distinction can be functionally significant, because the quinone arrangement in cytochrome be, (/7) allows proton translocation to be coupled to electron transfer through operation of the Q-cycle, as first recognized by Mitchell (28) . When quinones are on opposite sides of the membrane, the Q-cycle couples oxidation at one quinonc site to reduction at the second site, with the net result that protons are transported across the membrane. In contrast, if the quinones are on the same side of the membrane as in the reaction center (23), there may be proton release associated with hydroquinone oxidation, or proton uptake associated with quinone oxidation, but d1crc cannot be quinonc-mcdiatcd proton trnnslocation across the membrane. In cytochrome be,, two heme groups are positioned between the quinone binding sites to mediate electron transfer between these centers. While the £. coli fumaratc rcductase lacks heme, other fumarate reductases and succinate dehydrogenases contain one or two b-type hemes. These hemes arc likely coordinated by histidine residues positioned between the quinone sites in the membrane-spanning region. Although fumarate reductase and succinate dehydrogenasc are not known to couple proton translocation to electron transfer, the similar arrangement of redox groups across the membrane like that found in the cytochrome be, complex raises the possibility that at some point these enzymes may have participated in proton translocation. Fumaratc reductasc and succinate dchydrogcnase occupy central positions in cellular energy metabolism; fumarate reductase serves as the terminal acceptor for a major anaerobic respiratory pathway, while succinate dehydrogenase participates in both the Krebs cycle and as complex 11 of the aerobic respiratory chain. Although fumaratc rcductasc and succinatc dehydrogcnasc catalyze the same reaction (but in different physiological directions) and are predicted to have similar structures, organisms with both types of respiratory chains usc distinct proteins for each purpose for reason s not understood . In terms of the overall process of respiration, exciting progress has been made recently in structurally characterizing membrane-associated members of respiratory pathways (29) . Fumarate reductase 18 JUNE 1999 19 83 RESEARCH ARTICLES (complex II) now joins structures available for cytochrome bc 1 [complex III (1 7)], cytochrome c oxidasc (complex IV (30)], and the F 1 component of the ATP synthase [complex V (31)] . This gives a more complete view of the respiratory chain at the atomic level and increases our understanding of one of the most fundamental processes of biological systems. References and Notes 1. For a review, see G. Gottschalk. Bacterial Metabolism (Springer-Verlag. New Yori<, ed. 2, 1986). 2. For a review, see A Kroger, V. Geisler, E. Lemma, F. Theis. R. Lenger. Arch. Microbial. 158. 311 (1992). 3. For reviews, see B. A. C. Ackrell, M. K. Johnson, R. P. Gunsalus, G. Cecchini, in Chemistry and Biochemistry of Flavoenzymes, F. MUller, Ed. (CRC Press, Boca Raton. FL. 1992), vol. 3, pp. 229-297; C. Hagerhall. Biochim. Biophys. Acta 1320, 107 (1997): J. J. Van Hellemond and A.G. M. Tielens. Biochem.}. 304. 321 (1994). 4. T. Grundstr0m and B. Jaurin, Proc. Natl. Acad. Sci. U.S.A . 79. 1111 (1982): 5. T. Cole. Eur.}. Biochem. 122, 479 (1982): - - , T. Grundstrom, B. Jaurin, J. J. Robinson, J. H. Weiner. ibid. 126, 211 (1982). 5. W. H. Walker and T. P. Singer, }. Biol. Chem. 245, 4224 (1970): J. H. Weiner and P. Dickie. Can. }. Biochem. 254. 8590 (1979). 6. J. E. Morningstar, M. I(. Johnson, G. Cecchini, B. A C. Ackrell, E. B. Kearney. /. Biol. Chem. 260, 13631 (1985); R. Cammack, D. 5. Patil, J. H. Weiner, Biochim. Biophys. Acta 870, 545 (1986): A. T. Kowal et al., Biochemirtry 34, 12284 (1995). 7. F. J. Ruzicka, H. Beinert. K. L Schepler, W. R. Dunham, n. R. H. Sands, Proc. Natl. Acad. Sci. U.S.A. 2886 (1975): J.C. Salemo and T. Ohnishi, Biochem. j. 192, 769 (1980): D. J. Westenberg. R. P. Gunsalus, B. A. C. Ackrell. H. Sices. G. Cecchini, j . Biol. Chem. 268, 815 (1993). 8. E. Maklashina, D. K. Berthold, G. Cecchini, J. Bacteriol. 180, 5989 (1998): J. R. Guest.}. Gen. Microbiol. 122. 171 (1981). 9. J. H. Weiner et al., Proc. Natl. Acad. Sci. U.S.A. 83, 2056 (1986); G. Cecchini et al.. ibid.. p. 8898; D. J. Westenberg. R. P. Gunsalus, 8. A C. Ackrell G. Cecchini,/. Biol. Chem. 265, 19560 (1990); I. Schroder, R. P. Gunsalus, B. A. C. Ackrell, B. Cochran, G. Cecchini, ibid. 266. 13572 (1991). 10. N. Ishii et al.. Nature 394. 694 (1998). 11. T. Bourgeron et al., Nature Genet 11, 144 (1995). 12. T. P. Singer, Biological Oxidations, T. P. Singer, Ed. (lnterscience, New York, 1968). pp. 339-377. 13. Fumarate reducta se from E. co li was produced and purified as de sc ribed {C. Luna-Chavez and G. Cecchini, in preparation). In brief, protein was e>etracted into 0.05% (w/v) Thesit (polyoxyethylene 9-do- decyl ether) and concentrated to 30 mg/ml in a buffer of 20 mM Iris (pH 7.4) and 0.7% Thesit. Crystals were obtai ned by vapor diffu sion with hanging drops equilibrated against a reservoir solution [8% PEG 10,000, 0.1 M sodium citrate (pH 5.8). 0.085 M magnesi um acetate, 0. 1 mM EDTA. 0.001 % (w/v) dithiothreitol]. Menaquinone and oxaloacetate were not added to either the protein preparation or the crystallization conditions. Crystals grew in the space group P2,2,2, with unit cell constants a = 96.6 A. b = 138.1 A. and c = 275.3 A and two complexes per asymmetric unit. The oxidation state of the enzyme cannot be clearly established in this structure. 14. Data were collected by the inverse beam method at - 180°C at beam line 5.0.2 at the Advanced light Source with an AOSC Quantum4 charge-coupled device. Wavelengths for optimal data collection were determined with a single-crystal EXAFS (extended x-ray absorption fine structure) scan. Data were pro- program SHARP [E. De La Fortelle and G. Bricogne, Methods Enzymol. 276A, 472 (1997)] yielding ini- tial phases. Phases were improved with solvent flattening and multiple domain noncrystallographic averaging in DM [(32); K. Cowtan. Joint CCP4 EACBM News/ett. Crysta/log,. 31. 34 (1994)] initially with the transformation matrix determined based on the positions of the iron-sulfur clusters with the program FINDNCS [G. Lu.}. Appl. Crysta/log,. 32. 365 (1999)]. Masks used in averaging were calculated with MAMA [G. J. Kleywegt and T. A. Jones, Acta Crysta/log,. D55. 941 (1999)1 The model was built as a polyalanine chain at 4 reso lution with the program O version 6.22 (T. A. Jones and M. Kjeldgaard. Methods Enzymol. 277. 173 (1997)]. Subsequent rounds of modeling used the structures of NADH peroxidase (Protein Data Bank (PD8) accession number ljoa.pdb: the root mean square deviation (rmsd} is 1.4 A for the 125 Ca atoms in this alignment] (33), the 2Fe ferredoxin from SpiruUna platensis (accession number 4fxc.pdb) (34), and the 8Fe ferredoxin from Peptococcus aerogenes (accession number 1fdx.pdb) (35) as guides for structurally similar regions. This modeling used phase-combined maps produced by SIGMAA (32) and REFMAC [(32): G. N. Murshudov, A. A. Yagin, E. j. Dodson. Acta Crysta/log,. D53 , 240 (1997)]. Side chai ns were added to the model when the effective resolution reached 3.5 A resolution and the R,rl!f: dropped below 40%. Side chain assignment in the iron prote in was greatly aided by sequence analysis identifying the iron-sulfur cluster ligands (3}. Side chains were initially assigned to the fla vo protein based on sequence similarity with NADH peroxidase in the flavin-binding region . Refinement was carried out with alternating cycles of REF MAC [(32): G. N. Murshudov, A. A. Yagin, E. J. Dodson, Acta Crysta/log,. D53, 240 (1997)] and X-PLOR [N . 5. Pannu and R. J. Read. ibid. A52. 659 (1996)] with tight noncrystallographic symmetry (NCS) restraints on the soluble region until the Rfrcc dropped below 35%, when the NCS restraints were only loosely used. The rmsd of the soluble and mem brane domains of the two NCS related complexes in the asymmetric unit is 0.85 and 0.79 A, re spectively, as determined with the program LSQKA 8 (32). To take advantage of the quality and completeness of the peak data set and the higher resolution of the high remote data set, we chose the same test set for Rtrcc for these data sets and alternated refinement between them. The values for Rcryst and Rfrcc after refinement are 22.2 and 29.2%, respectively, where Rcryst = ~:IF0 i - IFc"I/ l !F0 1 and Rrru is Rcryst for 2% of the reflections omitted from the refinement totaling 1005 Bijvoet pairs. The rmsd of bond lengths and angles is 0.02 A and 1.9°, respectively, with less than 1% of the residues having di sa llowed values for the <I> and 'I' angles according to a Ramachandran diagram. The final model for each complex contains 8467 atoms in residues 1 to 575 (of 601 total residues) of FrdA, 1 to 243 (of 243 total residues) of FrdB. 1 to 130 (of 130 total residues) of FrdC. and 1 to 118 (of 118 residues) of FrdD, as well as a [2Fe:25] cluster. a [3Fe:4S] cluster, a [4Fe:45] cluster, a FAD, an oxaloacetate molecule, two menaquinone molecules, and one molecule of ordered detergent. The initiator methionine has been excluded from residue numbering because it is believed to be cleaved from the mature pro tein (3), consistent with the lack of electron density seen before residue 1 of each protein in the complex. 16. The coordinates have been deposited in the Protein Data Bank (accession number lfum). cessed with DENZO and scaled with SCALEPACK [Z. 17. D. Xia et al., Science 277, 60 (1997): Z. Zhang et al.. Nature 392,677 (1998): S. Iwata et al., Science 281, 64 (1998). Otwinoski and W. Minor, Methods Enzymol. 276,307 18. Surface area calculation was done with the program (1997)1 and the CCP4 suite of programs (32). 1S. Sites for the iron-sulfur clusters were determined 1966 from anomalous Patterson maps calculated at 6 A reso lution with the program FFT (32). Positions and occupancies of the sites, as well as values for f' and f" for two wavelengths, were refined with the GRASP [A. Nicholls, K. Sharp. B. Honig. Proteins 11. 281 (1991)]. 18 JUNE 1999 VOL 284 SCIENCE 19. A search of the Protein Data Bank with the Europea n Molecular Biology Laboratory DALI server [L Holm and C. Sander,/. Mo/. Biol. 233, 123 (1993)] did not reveal significant structural similarity to any previously determined folds for domains of the flavoprotein other than the flavin binding domain mentioned in the text. 20. C. Hagerhall and L. Heder.;tedt, FEBS Lett. 389, 25 (1996). 21. 5.K.Shenoy,L. Yu,C.A. Yu./. Biol. Chem. 272, 17867 (1997): X. D. Yang. L Yu, D. Y. He, C. A. Yu, ibid. 273, 31916 (1998). 22. J. C. Salemo, H. j. Harmon. H. Blum. J. 5. Leigh, T. Ohnishi, FEBS Lett. 82, 179 (1977); T. Miki, L Yu, C.A. Yu. Arch. Biochem. Biophys. 293. 61 (1992). A. R. Waldeck et al. , J. Biol. Chem. in. 19373 (1997). 23. J. Deisenhofer, 0. Epp, K. Miki, R. Huber, H. Mi chel, Nature 318, 618 (1985): G. Feher. J.P. Allen. M. Y. Okamura. D. C. Rees. ibid. 339. 111 {1989). 24. H. Beinert, 8. A. C. Ackrell, A. D. Vinogradov, E. Kearney, T. P. Singer.Arch. Biochem. Biophys. 182, 95 (1977); T. Ohnishi and B. l. Trumpower.}. Biol. Chem. 255. 3278 (1980). 25. T. Ohnishi et al .. J. Biol. Chem. 256, 5577 (1981). 26. P. l. Dutton et al., Biological Electron Transfer Chains: Genetics, Composition and Mode of Operation, G. W. Canters and E. Vijgenboom, Eds. {Kluwer, Netherlands, 1988), pp. 3-7. 27. G. Cecchi ni, 8. A. C. Ackrell, J. 0. Deshler, R. P. Gunsalus, /. Biol. Chem. 261 , 1808 (1986); 8. A. C. AckrelL B. Cochran, G. Cecchini, Arch. Biochem. Biophys. 268, 26 (1989). 28. P. Mitchell, Biol. Rev. 41, 445 (1966); For a review, see U. Brandt and 8. l. Trumpower, Crit. Rev. Biochem. Mo/. Biol. 29, 165 (1994). 29. For a review. see M. Saraste, Science 283, 1488 (1999). 30. S. Iwata, C. Ostermeier, B. Ludwig, H. Michel, Nature, 376, 660 (1995); T. Tsukihara et al.. Science 269, 1069 (1995). 31. J.P. Abrahams. A.G. W. Leslie, R. Lutter, J. E. Walker, Nature 370, 621 (1994). 32. 5. Bailey, Acta Crysta/log,. D50, 760 (1994). 33. J. I. Yeh, A. Claiborne, W. G. Hot. Biochemistry 35, 9951 (1996). 34. K. Fukuyama, N. Ueki, H. Nakamura, T. Tsuk.ihara, H. Matsubara.}. Biochem (Tokyo) 117. 1017 (1995). 35. E. T. Adman, L T. Sieker, L H. Jensen, j. Biol. Chem. 251. 3801 (1976). 36. Crystal contacts in the membrane-spanning regio n of a protein are relatively unusual, but have previously been seen in the structure of the iron transporter FhuA (K. P. Locher, B. Rees, D. Moras, J. P. Rosenbusch, in G-Protein-Coupled Receptors T. Haga, Ed. (CRC Monographs, CRC Press. Boca Ra - ton, FL, in press); K. P. Locher et al., Cell 95, 771 (1998)], porins [S. W. Cowen et al., Nature 358, 727 (1992)]. and bacteriorhodopsin, which was crystallized in the presence of cubic lipids [E. Pebay-Peyroula. G. Rummel, J. P. Rosenbusch. E. M. landau, Science 277, 1676 (1997)]. 37. All figures were made with MOLSCRIPT [P. J. Kraulis, }. Appl. Crysta/log,. 24,946 (1991)] or BO8SCRIPT [R. Esnouf.}. Mo/. Graph. 15. 133 (1997)] and rendered with RASTER3D [E. A. Merritt and M. E. P. Murphy, Acta Cryrtallogr. D50, B69 (1994)]. 38. Supported by the Department of Veterans Affairs, NIH, and NSF (G.C. and C.L-c.), and the Howard Hughes Medical Institute and NIH (D.C.R.). T.M.I. is supported by an NIH training grant. We thank I. Sch roder, S. I. Chan, S. C. Hung. and the members of the Rees group for discussions; K. H. Tubman and T. D. Tubman for critical reading: and J. J. Ottesen and T. N. Earnest for experimental ass istance. The Advanced light Source is supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, Materials Sciences Division, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098 at Lawrence Berkeley National Laboratory. 3 May 1999; accepted 21 May 1999 www.sciencemag.org 84 Chapter 4 Iverson, T.M., Luna-Chavez, C., Schroder, I. , Cecchini, G., Rees, D.C. (2000) Analyzing your complexes: Structure of the quinol-fumarate reductase respiratory complex. Curr. Op. Str. Biol. in press. 85 Analyzing your complexes: Structure of the quinolfumarate reductase respiratory complex Tina M. Iverson*, Cesar Luna-Chavezt, Imke Schrodert, Gary Cecchinit, and Douglas C. Rees§ * Graduate Option in Biochemistry, 14 7-75 CH, California Institute of Technology, Pasadena, CA 91125 t Molecular Biology Division, Department of Veterans Affairs Medical Center, San Francisco, CA 91421 and the Department of Biochemistry and Biophysics, University of California, San Francisco, CA 94143 :j: Department of Microbiology and Molecular Genetics and the Molecular Biology Institute, University of California, Los Angeles, CA 90024 86 § Howard Hughes Medical Institute, Division of Chemistry and Chemical Engineering, 147-75 CH, California Institute of Technology, Pasadena, CA 91125; telephone (626) 395-8393, FAX (626) 744-9524, e-mail dcrees@caltech.edu. Abbreviations QFR quinol-fumarate reductase SQR succinate-quinone oxidoreductase FAD flavin adenine dinucleotide Qr menaquinone binding site proximal to the QFR soluble domain Q0 menaquinone binding site distal to the QPR-soluble domain Qi quinone-reducing site from the cytochrome bc 1 complex Q0 quinol-oxidizing site from the cytochrome bc 1 complex The integral-membrane protein complex quinol-fumarate reductase (QFR) catalyzes the terminal step in a major anaerobic respiratory pathway. The homologous enzyme succinate-quinone oxidoreductase participates in aerobic respiration as both complex II and as a member of the Krebs cycle. Last year, two structures of QFRs were reported. These structures revealed the cofactor organization linking the fumarate and quinol sites and showed a cofactor arrangement across the membrane that is suggestive of a possible energy coupling function. 87 Introduction Bacteria such as Escherichia coli can utilize multiple respiratory pathways, and although aerobic respiration is preferred due to the highly favorable energetics of oxygen reduction, respiration can proceed in the absence of oxygen [l]. Fumarate represents one of the more common alternative terminal electron acceptors [2], and is reduced to succinate during anaerobic respiration by quinol-fumarate reductase (QFR), an integral membrane protein located in the bacterial inner membrane. A pool of reduced quinones in the membrane serves as the electron source for fumarate reduction by QFR. The reduced quinones are ultimately generated through the action of an integral membrane protein donor, such as hydrogenase or a formate dehydrogenase. For comparison, mitochondrial aerobic respiration [3] requires four integral-membrane proteins: NADH-ubiquinone oxidoreductase (NUO; complex I); succinate-quinone oxidoreductase (SQR; complex II); cytochrome bc 1 (complex III); and cytochrome c oxidase (complex IV); as well as the involvement of soluble cytochrome c and the ubiquinone pool to transport electrons between complexes. SQR and QFR represent a common element of both aerobic and anaerobic respiration since they exhibit significant sequence similarity to one another. Both complexes can catalyze either the oxidation of succinate or the reduction of fumarate [4-6), although physiologically they function to 88 catalyze this reaction in opposite directions [7-9] . In addition to the involvement in both aerobic and anaerobic respiration, complex II (SQR) also participates in the citric acid cycle, reflecting the central importance of this enzyme in basic metabolism [ 1O] . QFR and SQR are composed of three or four subunits: the flavoprotein; the iron protein; and one or two membrane-bound subunits, depending on the organism. The total molecular weight of a single complex is -120 kD. Both QFR and SQR can be separated into two components: a water-soluble domain and the membrane anchor subunit(s). The purified soluble domain consists of the flavoprotein subunit and the iron protein subunit, and retains catalytic activity for fumarate reduction when a suitable source of reducing equivalents is provided [ 11, 12]. The subunits of the soluble domain exhibit strong sequence conservation throughout all species and contain a variety of redox cofactors. The flavoprotein contains flavin adenine dinucleotide (FAD) covalently linked to the Ne atom of a conserved histidine residue [13], while the iron protein contains three iron sulfur clusters, a [2Fe:2S] cluster, a [3Fe:4S] cluster, and a [4Fe:4S] cluster [7]. In contrast to the soluble proteins, the sequence and cofactor composition of the membrane anchors vary between different organisms and even between QFR and SQR of the same organism. These membrane anchors have been assigned to four classes [ 14] that differ in the number of transmembrane subunits (one or two), the number of transmembrane 89 helices (five or six), and the number of associated b-type heme moieties (zero, one, or two). The preceding year has witnessed an explosion in the structural understanding of complex II, with structures for the complete QFR complex from both E. coli [ 15] and Wolinella succinogenes [16] reported. Additionally, the structures of four soluble homologs of the flavoprotein were described [17-20], which allow a detailed comparison of flavoprotein structure, and increase insight into catalysis by complex II. Organization of the complex As predicted from biochemical and sequence analyses, complex II exists as an essentially modular protein (Figure 1). Each subunit in the soluble domain exhibits homology to other known proteins, suggesting that the formation of this complex may have proceeded by the assembly of other proteins [21]. For example, soluble homologs of the flavoprotein have been seen in multiple pathways (NADH biosynthesis and fumarate reduction), and the iron protein has sequence and structural similarity to both plant and bacterial ferredoxins. Furthermore, the iron protein exhibits weak sequence homology and similar cluster characteristics to the iron protein subunits found in other terminal respiratory complexes, including the multi-subunit membrane-bound DMSO reductase 90 [22] and heterodisulfide reductase [23] . Thus, modular complex formation may represent an evolutionary mechanism to adjust to the availability of dissimilar terminal electron acceptors. In the transmembrane region, the membrane anchor subunits of the E. coli and W. succinogenes QFR complexes are organized around a central four-helix bundle, with two or one additional helices outside this central core, respectively. The E. coli and W. succinogenes QFRs (Figure 1) contain different cofactors associated with the membrane spanning helices. In the E. coli QFR structure, two menaquinone molecules have been identified in the electron density, while the W. succinogenes enzyme contains two b-type hemes. As a result of the varied cofactor composition in the transmembrane region , the intercofactor distances (Figure 2) differ between the E. coli and W. succinogenes enzymes in the membrane-spanning region. This situation may be contrasted to the conserved distances between cofactors in the soluble domain. In the crystal structures of both the E. coli and W. succinogenes QFRs, two complexes are present in the asymmetric unit that associate through their transmembrane regions in a fashion suggestive of dimer formation. In the E. coli structure, this crystal contact buries 325 A2 of surface area, and is mediated by two ordered detergent molecules (C 12E9 ; [15]). In the W. succinogenes QFR structure, almost 3700 A2 of surface area is buried [16] in 91 this interface, suggesting this enzyme exists as a dimer. Although detergent molecules may have disassociated a physiological dimer in the E. coli enzyme, the C 12E9 detergent used in crystallization yields optimal enzyme activity when used in biochemical assays [24] . Additionally, it is possible that the crystals of the W. succinogenes enzyme contain a deceivingly tight crystal contact, and that the enzyme is indeed a monomer. Based on the intercofactor distances of the crystal packing interaction, if the W. succinogenes enzyme exists as a dimer, it likely acts as a structural and not a functional dimer, unlike the situation in the cytochrome bc 1 complex [25-27]. Fumarate reduction by the flavoprotein and comparison of flavoprotein conformations The flavoprotein subunit contains two major domains, a flavin-binding domain based on a Rossmann type fold, and a capping domain (Figure 1). The flavin-binding domain exhibits sequence and structural conservation between the two full length QPR structures, the soluble fumarate reductases [ 18-20], and L-aspartate oxidase (LAS PO; [ 17]). The pairwise deviations between Ca atoms in these proteins never exceeds 1. 7 A, consistent with a sequence identity that is never less than 30%. 92 The flavin-binding domain and capping domain of the flavoprotein are connected by a small hinge region consisting of two ~-strands, with the active site located at the interface between these two domains. An overlay of the recently solved structures containing the flavoprotein [17-20, 28] fold shows that the relative angle between the flavin-binding and capping domains can vary (Figure 3). It is proposed that during catalysis, fumarate enters the active site, with concomitant closure of the capping domain, sterically causing a rotation of the substrate carboxylate around the double bond (Figure 4). Interaction with the protein have been shown to distort the substrate in complexes with the W. succinogenes enzyme [ 16] and the soluble Shewanella enzyme [19] . Fumarate reduction then proceeds by hydride transfer from the NS of the flavin, followed by proton transfer from a nearby side chain. Four active site residues (His 232, Arg 287, His 355 and Arg 390 in the E. coli sequence) that directly contact substrate or inhibitor are absolutely conserved in all available sequences. Three of these residues, His 232, Arg 287 and Arg 390, have been substituted by site directed mutagenesis and are known to be critical for enzyme activity [29; Maklashina, Schroder and Cecchini, unpublished observation] . Both His 355 and Arg 390 have been suggested as proton donors for this reaction [ 16, 19]; however, the role each side chain plays in the reaction mechanism remains to be conclusively established. 93 In an overlay of the ensemble of available structures (Figure 5), the side chains of His 232, His 355 and Arg 390 appear structurally conserved both in location and conformation. In contrast, Arg 287, which is located in the capping domain, exhibits different conformations depending on the orientation of the capping domain and the nature of the species bound at the active site. This suggests Arg 287 may be involved with recruiting substrate into or moving product out of the active site. The iron protein As predicted by sequence similarity to plant and bacterial ferredoxins, the iron protein contains two domains. The N-terminus has a fold similar to plant-type ferredoxins surrounding the [2Fe:2S] cluster, while the C-terminal domain exhibits a core similar to bacterial ferredoxins and contains the cysteine ligands to the [4Fe:4S] and the [3Fe:4S] clusters. The C-terminal domain additionally contains several other helices that associate with the membrane anchor subunits. In agreement with EPR spectroscopy [30-33], the three clusters are arranged in a nearly linear fashion leading from the membrane anchor to the active site flavin (Figure 2). This structural organization indicates that, despite the low reduction potential of the [4Fe:4S] cluster [7, 8], it is likely that all [Fe:S] clusters participate in electron transfer between the quinone pool and fumarate [34, 44]. 94 Organization of cofactors in the membrane: a possible role in energy transduction Two sets of transmembrane associated cofactors are present in the two available QFR structures: two menaquinone molecules (Qr and Q0 ) are observed in the E. coli enzyme, while two b-type hemes bind to the W. succinogenes enzyme (Figure 1). Despite the differences in subunit composition, sequences, and cofactor components, the transmembrane anchors of the E. coli and W. succinogenes QFRs are both organized around an antiparallel four helix bundle in the membrane, with two or one additional peripheral transmembrane helices, respectively. Irr each structure, one of the helices of the four helix bundle contains a distinctive bend in the transmembrane section. As an ideal four-helix bundle is two-fold symmetric, the transmembrane helices of the QFR structures can be aligned in two distinct ways. In one alignment, the C-subunit of the E. coli enzyme is superimposed with the N-terminus of the C-subunit of the W. succinogenes enzyme, yielding an rms deviation of 2.4 A for 137 Ca atoms between corresponding residues in the five transmembrane helices of the Wolinella enzyme and the E. coli enzyme. In this alignment, the connections between the transmembrane helices are similar, and the bent helices superimpose (Figure 6A). However, if the menaquinone molecules from the E. coli QFR were transferred into the W. succinogenes enzyme, they would sterically clash with the distal heme of the latter. In the second alignment, which is rotated ~180° from the first alignment, the N-terminus of the D- 95 subunit of the E. coli enzyme becomes superimposed with the N-terminus of the Csubunit of the W. succinogenes enzyme (Figure 6B). Although this alignment appears quantitatively inferior (2.2 A for only 62 Ca atoms), it preserves the relative orientation between the transmembrane helices and the soluble domains in the two QFR structures. Furthermore, the Q 0 menaquinone from the E. coli enzyme no longer sterically clashes with the distal heme, suggesting that this position might correspond to where Q0 binds in the W. succinogenes enzyme. Indeed, a cavity is present in the Wolinella QFR near this position that might represent the cofactor binding site for that complex. The separation distance between the two menaquinone molecules in the E. coli enzyme (~25 A; Figure 2) is too far to support electron transfer at physiological relevant rates [34] . While this may indicate that the Q 0 site is not catalytically relevant, an alternate explanation is that a third cofactor binding site is positioned between the Qp and Q 0 sites and would be ~ 13 A from each of the established quinone sites. Although the involvement of an additional cofactor cannot be conclusively proven, several factors point to its possible existence. Biochemical evidence [35-37] suggests that Q0 is necessary for physiological catalysis, and site-directed mutagenesis has implicated a cluster of residues between the two known quinone binding sites as necessary for enzyme function [35]. This cluster of residues lies near a cavity between the transmembrane helices in the E. coli enzyme and is associated with density from the structure 96 determination that could not be assigned. Spectroscopic evidence suggests that a stabilized semiquinone pair separated by ~8 A is localized near the [3Fe:4S] cluster [33, 38-41], which could be a result of Qp interacting with a quinone in a central binding site that could not be identified in the density. Indeed, several residues (Trp C86, Arg D51) lining the pocket of the potential quinone binding site also line the QP binding site. Although a menaquinone molecule could not be docked into the density observed in the E. coli structure determination, it is possible that this site is only partially occupied in the crystal structure. One intriguing observation concerning the quinone locations in the E. coli QFR is that they are positioned on opposite sides of the membrane, as observed for the two quinone sites in the cytochrome be, complex [25-27]. In cytochrome be,, this transmembrane arrangement is essential for proton pumping across the mitochondrial membrane via a Qcycling mechanism [42]. In addition to the location of the quinone binding sites, the environment of the Qp site shares certain features in common with the Q 0 and Qi sites of cytochrome be, . Proton translocation at the Q0 site in cytochrome be, is thought to occur through His 161 and Glu 272 (bovine numbering). These two residues only form hydrogen bonds to inhibitors when the Reiske iron sulfur center is at a position near the Q0 site [43; S. Iwata, personal communication] thereby mechanically coupling proton translocation to electron transfer to the Reiske center. Analysis of the Q; site suggests 97 that the conserved Asp C228 and His C20 1 (bovine numbering) residues, both of which form hydrogen bonds to the quinone (Figure 7 A), participate in the proton translocation mechanism. In the E. coli QFR structure, Glu C29 and Lys B228 are each within hydrogen bonding distance of the quinone (Figure 7B). Glu C29 has been shown by sitedirected mutagenesis [35] to be critical for enzyme function . Proton translocation has not been observed for QFR or SQR, although the energetics of proton translocation by QFR have been calculated to be favorable for the W. succinogenes complex [44]. In contrast, similar calculations indicate that energetics are less favorable for the E. coli enzyme. If other modular terminal respiratory acceptors, such as DMSO reductase or heterodisulfide reductase, have evolved with a similar transmembrane arrangement of quinone binding sites, it is possible that these proteins may employ a Q-cycling mechanism to pump protons. Conclusions Structural information is now available for most members of the aerobic respiratory chain, including atomic-level resolution crystal structures of the respiratory complexes IIIV [25-27, 45, 46], much of the ATP synthase [47, 48] and electron microscopic reconstruction of complex I [49-51] at approximately 30 A resolution. In the last year, structures have been reported for two intact two full-length QFRs, one from E. coli [15] 98 and one from W. succinogenes [16], as well as several soluble homologs [17-20]. This ensemble of fumarate reductase structures allows a better understanding of the reaction mechanism catalyzed by the flavoprotein as well as an understanding of how electrons are passed from the membrane soluble quinones into the active site of the enzyme. Additionally, these structures open the question of whether QFR and SQR can potentially couple electron transfer to proton translocation, thereby representing an additional mechanism for energy conservation in respiratory chains. Acknowledgements This work has been supported by the Department of Veterans Affairs, and funding from HHMI (DCR), the NIH (GM45162, DCR; HL-16251, GC), and the NSF (MCB-9729778, GC) . We would like to acknowledge stimulating discussions with T. Ohnishi, P. L. Dutton, H.B. Gray, and S. Iwata. 99 Figure Ribbon 4.1: diagrams of the E. coli (lFUM) and w. succinogenes (IQLA) QFR complexes. The flavoprotein is colored in purple, capping highlighted with the domain m dark purple; the iron protein is colored in teal; the C-subunit membrane anchor is in peach; while the D-subunit membrane anchor is in gray (E. coli enzyme only). The left column shows a view that is rotated 90° from the right column. The W. succinogenes enzyme contains a C-terminal extension of the flavoprotein (lower right figure) that was not observed in the E. coli structure determination. Figures 1-3 and 5-7 were made using MOLSCRlPT [52], BOBSCRlPT [53], and RASTER3D. 100 Figure 4.2: Distances between cofactors in the E. coli and W. succinogenes QFR complexes. Distances between redox centers are measured between the closest redox active atom of each cluster, rather than center-to-center distance as given in Iverson et al [15]. The distances to hemes are measured to the heme iron rather than the edge of the porphyrin ring. The distances between cofactors in the soluble domains ([Fe:S] clusters and FAD) are conserved, whereas the distances between the cofactors associated with the membrane anchors varies as a result of the differences in cofactor composition. ., •,..• .• ,. •.t..·.:. ,t . • .. .. • ' 12.3A C ,. ,. 0 • • FAD OAA [2Fe:2S] [3Fe:4S] 9.1 A a.2 A [4Fe:4S] \ .. 17.6 A 15.6 A On 0 .J 101 Figure 4.3: Stereoview illustrating the movement of the capping domain with respect to the flavin-binding domain in the flavoprotein. The capping domain from S. putrefaciens (green; 1D4D; [19]) S. frigidimarina (pink; 1Q08; [18]) and W. succinogenes (blue; lQLA; [16]) flavoenzymes are shown relative to a least squares superposition of the flavin-binding domain (black), with the position of the FAD indicated (red). The maximal displacement for the capping domain suggests that the hinge connecting the two domains can bend over 30 °. Other flavoproteins having intermediate hinge angles have been omitted for clarity. 102 Figure 4.4: diagram of Schematic active site R la:; c~yo interactions with proposed reaction rl mechanism. NH N I Hydrogen bond interactions : ~\(" ~ -··!\ 0/, _•• ~; ✓ '<(/ between the protein and :r"" A,g 390 bound ligands are indicated, with the exception of Arg which exhibits His 232 N His 355 significant conformational variability in the different structural determinations. Hydride transfer from the NS of FAD (green arrows) represents the initial step of fumarate reduction, followed by proton transfer to the substrate. The side chains of His 232, Arg 'lKl, His 355 or Arg 390 are likely candidates for the immediate proton donor. Arg 287 and Arg 390 have also been shown to be necessary for covalent attachment of the FAD to His 44 [Maklashina, Schroder and Cecchini, unpublished observation]. 103 Figure 4.5: Stereoview illustrating the superposition of active site residues in different fumarate reductase structures. The positions of residues from the E. coli (lFUM) enzyme are red, from the Wolinella (lQLB) enzyme are orange, from L-aspartate oxidase (lCHU) are yellow, from the S. frigidimarina soluble fumarate reductase are green and blue (1Q08; lQJD), and from the S. putrefaciens fumarate reductase (1D4D) are magenta. The cofactor FAD and substrate fumarate substrate fumarate are black and gray, respectively. Although the positions of His 232, His 355 and Arg 390 are structurally conserved, the positions of the Arg 287 side chain and the surrounding main chain are more variable. His 44 His 44 104 Figure 4.6: Alignment of transmembrane helices from the E. coli (lFUM) and W. succinogenes (lQLB) QFR complexes. A. Superposition of the transmembrane helices with the N-terminus of the C-subunit of the E. coli enzyme (C subunit in blue, D subunit in green) superimposed with the N-terminus from the W. succinogenes enzyme (pink). Quinones from the E. coli enzyme are shown in yellow, while the hemes from the Wolinella enzyme are in black. B. Alternate alignment superimposing the N-terminus of the E. coli enzyme D subunit with the N-terminus of the W. succinogenes C subunit. A cavity located by the program VOIDOO [55] in the W. succinogenes enzyme is shown in brown mesh. If this cavity represents the quinone •binding site for the Wolinella enzyme, the location would be structurally preserved in both enzymes. 105 Figure 4.7: A. Binding pocket for QP in the E. coli QFR reductase complex. The quinone binding environment includes hydrogen bonds to the side chains of Glu C29 and Lys B228 B. Binding pocket for Q0 in the cytochrome bc 1 complex. The structural arrangement of the hydrogen bonds relative to the quinone is similar to those seen in the E. coli QFR. Additionally, several other r•esidues (Phe D17 (QFR) and Phe C32 (bc 1); Trp D14 (QFR) and Leu C22 (bc1); [3Fe:4S] (QFR) and heme bH (bc 1)) are observed in conserved locations with resides of similar property in the E. coli enzyme. 106 References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: • of special interest •• of outstanding interest 1. Gottschalk, G: Bacterial Metabolism. New York: Springer-Verlag; 1986. 2. Kroger, A, Geisler, V, Lemma, E, Theis, F and Lenger, R: Bacterial Fumarate Respiration. Arch Microbial 1992, 158:311-314. •• 3. Saraste, M: Oxidative phosphorylation at the fin de siecle. Science 1999, 283:14881493. 4. Guest, JR: Partial replacement of succinate dehydrogenase function in phage and plasmid-specified fumarate reductase in Escherichia coli. J Gen Microbial 1981, 122: 171-179. 5. Maklashina, E, Berthold, DK and Cecchini, G: Anaerobic expression of Escherichia coli succinate dehydrogenase: functional replacement of fumarate reductase in the respiratory chain during anaerobic growth. J Bacterial 1998, 180:5989-5996. 6. Pershad, HR, Hirst, J, Cochran, B, Ackrell, BAC and Armstrong, FA: Voltammeric studies of bidirectional catalytic electron transport in Escherichia coli succinate 107 dehydrogenase: comparison with the enzyme from beef heart mitochondria. Biochim Biophys Acta 1999, 1412:262-272. 7. Ackrell, BAC, Johnson, MK, Gunsalus, RP and Cecchini, G: Chemistry and biochemistry offlavoenzymes. Boca Raton: CRC Press; 1992. 8. Hagerhall, C: Succinate:quinone oxidoreductases. Variations on a conserved theme. Biochim Biophys Acta 1997, 1320:107-141. 9. Hellemond, JJV and Tielens, AGM: Expression and functional properties of fumarate reductase. Biochem J 1994, 304:321-331. 10. Guest, JR and Russell, GC: Complexes and complexities of the citric acid cycle in Escherichia coli. Curr Top Cell Regul 1992, 33:231-247. 11. Singer, TP, Kearney, EB and Kenney, WC: Succinate dehydrogenase. Adv Enzymol 1973, 37:189-272. 12. Ackrell, BAC, Kearney, EB and Singer, TP: Mammalian succinate dehydrogenase. Methods Enzymol 1978, 53:466-483. 13. Walker, WH and Singer, TP: Identification of the covalently-bound flavin of succinate dehydrogenas as 8a-(histidyl) flavin adenine dinucleotide. J Biol Chem 1970, 245:4224-4225. 14. Hagerhall, C and Hederstedt, L: A general structural model for the membrane- domain of succinate:quinone oxidoreductases. FEBS Lett 1996, 389:25-31. 108 ••15. Iverson, TM, Luna-Chavez, C, Cecchini, G and Rees, DC: Structure of the Escherichia coli fumarate reductase respiratory complex. Science 1999, 284:1961- 1966. The first structural analysis of quinol-fumarate reductase details the organization of the complex and reveals bound quinones in the membrane spanning regions. The locations of these quinones are reminiscent of quinones involved in energy transduction . .. 16. Lancaster, CRD, Kroger, A, Auer, Mand Michel, H: Structure of fumarate reductase from Wolinella succinogenes at 2.2 A. resolution. Nature 1999, 402:377-385. The structure determination of the intact quinol-fumarate reductase from W. succinogenes allows analysis of b-type hemes bound to the membrane anchor subunit and reveals common elements of QPR structure through comparisons to the E. coli enzyme. •17. Matttevi, A, Tedeschi, G, Bacchella, L, Coda, A, Negri, A and Ronchi, S: Structure of L-aspartate oxidase: implication for the succinate dehydrogenase/fumarate reductase oxidoreductase family. Structure 1999, 7:745-756. The first structure determination of a soluble homologue of the flavoprotein suggests flexibility exists between the flavin-binding domain and the capping domain. 109 •18 . Bamford, V, Dobbin, PS, Richardson, DJ and Hemmings, AM: Open conformation of a flavocytochrome c3 fumarate reductase. Nat Struct Biol 1999, 6: 1104-1109. •19. Leys, D, Tsapin, AS, Nealson, KH, Meyer, TE, Cusanovich, MA and VanBeeumen, JJ : Structure and Mechanism of the flavocytochrome c fumarate reductase of Shewanella putrefaciens MR-1. Nat Struct Biol 1999, 6: 1113-1117. •20. Taylor, P, Pealing, SL, Reid, GA, Chapman, SK and Walkinshaw, MD: Structural and mechanistic mapping of a unique fumarate reductase. Nat Struct Biol 1999, 6:1108-1112. This report and the preceding two [18, 19], structure determinations of soluble fumarate reductases, published simultaneously, provide further support for domain movement in the flavoprotein. 21. Gest, H: The evolution of biological energy-transducing systems. FEMS Lett 1980, 7:73-77. 22. Bilous, PT, Cole, ST, Andersib, WF and Weiner, JH: Nucleotide sequence of the dmsABC operon encoding the anaerobic dimethylsulphoxide reductase of Escherichia coli. Mol Microbiol 1988, 2:785-795. 23 . Kunkel, A, Vaupel, M, Heim, S, Thauer, RK and Hedderich, R: Heterodisulfide reductase from methanol-grown cells of Methanosarcina barkeri is not a flavoenzyme. Eur J Biochem 1997, 244:226-234. 110 24. Luna-Chavez, C, Iverson, TM, Rees , DC and Cecchini, G: Overexpression, purification, and crystallization of the membrane-bound fumarate reductase from Eschericia coli. Protein Expr Purif2000, in press. 25 . Xia, D, Yu, C, Kim, H, Xia, J, Kachurin, AM, Zhang, L, L. Yu and Deisenhofer, J: Crystal structure of the cytochrome bq complex from bovine heart mitochondria. Science 1997, 277:60-66. 26. Zhang, Z: Electron transfer by domain movement in cytochrome bq. Nature 1998, 392:677-684. 27. Iwata, S, Lee, JW, Okada, K, Lee, JK, Iwata, M, Rasmussen, B, Link, TA, Ramaswamy , S and Jap, BK: Complete structure of the 11-subunit bovine mitochondrial bq complex. Science 1998, 281:64-71. 28 . Ackrell, BAC: Progress in understanding structure-function relationships in respiratory chain complex II. FEBS Lett 2000, 466: 1-5. 29. Schroder, I, Gunsalus, R, Ackrell, BAC, Cochran, Band Cecchini, G: Identification of active site residues of Eschericia coli fumarate reductase by site directed mutagenesis. J Biol Chem 1991, 266:13572-13579. 30. Morningstar, JE, Johnson, MK, Cecchini, G, Ackrell, BAC and Kearney, EB: The high-potential iron-sulfur center in Escherichia coli fumarate reductase is a 3-iron cluster. J Biol Chem 1985, 260:3631-3638. 111 31. Cammack, R, Patil, DS and Weiner, JH: Evidence that center-2 in Escherichia coli fumarate reductase is a [4Fe-4S] cluster. Biochim Biophys Acta 1986, 870:545-551. 32. Kowal, AT, Werth, MT, Moanodori, A, Cecchini, G, Schroder, I, Gunsalus, RP and Johnson, MK: Effect of cysteine to serine mutation on the properties of the [4Fe-4S] center in Escherichia coli fumarate reductase. Biochemsitry 1995, 34: 12284-12293. 33 . Waldeck, AR, Stowell, MHB , Lee, HK, Hung, SC, Matsson, M, Hederstedt, L, Ackrell, BAC and Chan, SI: Electron paramagnetic resonance studies of succinate:ubiquinone oxidoreductase from Paracoccus denitrificans. J Biol Chem 1997, 272: 19373-19382. ••34. Page, CC, Moser, CC, Chen, X and Dutton, PL: Natural engineering principles of electron tunneling in biological molecules. Nature 1999, 402:47-52. This thorough compilation of electron transfer proteins discusses the mechanism of endergonic tunneling as well as explaining the biological tendency for redox cofactors to be positioned at distances ~ 13 A apart in proteins. 35. Westenberg, DJ, Gunsalus, RP, Ackrell, BAC, Sices, Hand Cecchini, G: Escherichia coli fumarate reductase FrdC and FrdD mutants. Identification of amino acid residues involved in catalytic activity with quinones. J Biol Chem 1993, 268:815-822. 36. Lee, GY, He, DY, Yu, Land Yu, CA: Identification of the ubiquinone-binding domain in QPsl of succinate-ubiquinone reductase. J Biol Chem 1995, 270:61936198. 112 37. Shenoy, SK, Yu, L and Yu, CA: The smallest membrane anchoring subunit (QPs3) of bovine heart mitochondrial succinate-ubiquinone reductase. Cloning, sequencing, topology, and Q-binding domain. J Biol Chem 1997, 272:17867-17872. 38. Salerno, JC, Harmon, HJ, Blum, H, Leigh, JS and Ohnishi, T: A transmembrane quinone pair in the succinate dehydrogenase cytochrome b region. FEES Lett 1977, 82: 179-182. 39. Rich, PR, Moore, AL, Ingledew, WJ and Bonner, WDJ: EPR studies of higher plant mitochondria. I ubisemiquinone and its relation to alternative respiratory oxidations. Biochim Biophys Acta 1977, 504:345-363. 40. Rich, PR and Bonner, WDJ: An EPR analysis of cyanide-resistant mitochondria isolated from the mutant poky strain of Neurospora crassa. Biochim Biophys Acta 1978, 504:345-363. 41. Salerno, JC and Ohnishi, T: Studies on the stabilized ubisemiquinone species in the succinate-cytochrome c reductasesegment of the intact mitochondrial membrane system. Biochem J 1980, 192:769-781. 42. Mitchell, P: Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nature 1961, 191:144-148. 43. Crofts, AR, Barquera, B, Gennis, RB, Kuras, R, Guergova-Kuras, Mand Berry, EA: Mechanism of ubiquinol oxidation by the bet complex: different domains of the 113 quinol binding pocket and their role in the mechanism and binding of inhibitors. Biochemistry 1999, 38:15807-15926. ••44. Ohnishi, T, Moser, CC, Page, CC, Dutton, PL and Yano, T: Simple redox-linked proton-transfer design: new insights from structures of quinol-fumarate reductase. Structure 2000, 8:R23-R32. This insightful mini-review focuses on the bioenergetics of QFR and SQR. 45 . Iwata, S, Ostermeier, C, Ludwig, B and Michel, H: Structure at 2.8-Angstrom resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 1995, 376:660-669. 46. Tsukihara, T, Aoyama, H, Tomizaki, T, Yamaguchi, H, Sinzawaltoh, K, Nakashima, R, Yaono, R and Yoshikawa, S: The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 Angstrom. Science 1996, 272: 1136-1144. 47. Abrahams, JP, Leslie, AGW, Lutter, Rand Walker, JE: Structure at 2.8-Angstrom resolution of FlATPase from bovine heart-mitochondria. Nature 1994, 370:621-628. ••48. Stock, D, Leslie, AGW and Walker, JE: Molecular architecture of the rotary motor in ATP synthase. Science 1999, 286:1700-1705. The structure determination of the c subunit of the ATP synthase shows homodecameric arrangement associated with the trimeric water soluble Fl component, suggesting a non-integral number of protons is pumped for each ATP synthesized. 11 4 49. Guenebaut, V, Vincentelli, R, Mills, D, Weiss , H and Leonard, KR: Three dimensional structure of NADH-dehydrogenase from Neurospora crassa by electron microscopy and conical tilt reconstruction. J Mol Biol 1997, 265:409-418. 50. Guenebaut, V, Schlitt, A, Weiss, H, Leonard, K and Friedrich, T: Consistent structure between bacterial and mitochondrial NADH:ubiquinone oxidoreductase (complex I). J Mol Biol 1998, 276: 105-112. 51. Grigorieff, N: Three-dimensional structure of bovine NADH:ubiquinone oxidoreductase (complex I) at 22.A in ice. J Mol Biol 1998, 277:1033-1046. 52. Kraulis , PJ: MOLSCRIPT - a program to produce both detailed and schematic plots of protein structures. J Appl Cryst 1991, 24:946-950. 53 . Esnouf, R: BOBSCRIPT -- An extensively modified version of molscript that includes greatly enhanced coloring capabilities. J Mol Graph 1997, 15:133-138. 54. Merritt, EA and Murphy , MEP: Raster3D Version 2.0 - a program for photorealistic molecular graphics. Acta Cryst 1994, D50:869-873. with a site-specific mutant. Biochemistry 1989, 28:7913-7918 . 55. Kleywegt, GJ and Jones, TA: Detection, delineation, measurement and display of cavities in macromolecular structures. Acta Cry st 1994, D50: 17 5-185. 115 Chapter 5: Biological Nitrification and the Nitrogen Cycle The biological nitrogen cycle involves a network of microorganisms that interconvert inorganic nitrogen between different redox states (+5 to -3) in five reduction reactions and three oxidation reactions ( 1-3 )(Figure 5.1). Aside from the obvious ecological relevance of the biological nitrogen cycle, the application of the reactions catalyzed by the nitrogen cycle has been a major biotechnological advance in waste removal of NO 3- , No 2-, and NH/ (4, 5) . One byproduct of the nitrogen cycle, N2 O represents the third largest contributor to greenhouse gasses, following CH4 and CO2 , and has been implicated in ozone depletion (6) . Further advances in microbial technology could help reduce atmospheric pollution by N2O. The metabolism of organisms that participate in the nitrogen cycle are relatively unusual, thus the enzymes that catalyze these reaction contain various metal cofactors that are often unique in composition, ligation, or both (1-3 ). In particular, molybdenumcontaining cofactors are used for both nitrogen fixation and nitrite reduction (7, 8); stacked-heme motifs and unusual heme ligation are involved in the catalytically opposite reaction of nitrification and nitrate ammonification (9-11 ); and novel four-copper center provides the active site for nitrous oxide reduction (12 ). These active site architectures are discussed in detail below. 116 Nitrogen fixation and nitrogenase Biological nitrogen fixation catalyzes the conversion of gaseous dinitrogen (N 2) to ammonia (NH 3) (7, 8). The extremely stable triple bond of N 2 provides a kinetic challenge to oxidize or reduce. The industrial method of doing so, called the HaberBosch process, requires extreme temperatures (~800 K) and pressures (~500 atm) as well as an iron catalyst to drive what should be a thermodynamically favorable process (13, 14). In contrast, obligate anaerobic nitrogen-fixing bacteria can accomplish this reaction at ambient temperatures and pressures using ·a variety of interesting iron-containing cofactors and produces a significantly greater amount of ammonia (~ 170 x 109 kg) per year than does the industrial process (~80 x 109 kg). Several types of nitrogenases are recognized. Although the molybdenum-containing nitrogenase, described in detail below, is by far the best-studied, some organisms have alternative nitrogenase systems containing vanadium and iron, or iron exclusively (15). The reaction catalyzed by the molybdenum-containing nitrogenase requires the input of approximately 16 ATP molecules (16), but still acts as the only energy source for nitrogen-fixing bacteria (Equation 5.1). N2 + 8H+ + 16 MgATP ➔ 2NH3 + H2 + 16 MgADP + 16 Pi (5.1) 117 The molybdenum-containing nitrogenase system involves two main components: the molybdenum-iron (MoFe) protein and the iron (Fe) protein. The MoFe-protein exists as a ~300 kD exiP 2 heterotetrameric complex (17-19) (Figure 5.1). Two unusual ironcontaining clusters are associated with the MoFe-protein: an [8Fe:7S] EPR-silent cluster (P-cluster) thought to participate in electron transfer (20-22) ; and the active site ironmolybdenum cofactor (20, 21) (FeMoCo; Figure 5.2). The FeMoCo contains 7 irons, 1 molybdenum atom, 1 homocitrate, and 9 bridging sulfur atoms and is ligated to the protein by only two protein ligands, one cysteine to an iron atom and one histidine to the molybdenum. The coordination sphere of the molybdenum is octahedral, and the sphere is completed by three of the bridging sulfurs of the FeMoCo and two contacts to the homocitrate. Although it is largely accepted that the FeMoCo is the active site (23-25 ), evidence for the catalytic roles of each of the two cofactors of the MoFe-protein remains largely circumstantial. However, changes in the cofactor or the surrounding environment of the FeMoCo can drastically affect reduction of the substrate. The Fe-protein couples the hydrolysis of MgATP (26, 27) to the electron transfer to the MoFe-protein that provides the reducing equivalents necessary for the reduction of N2 . The Fe-protein acts as a homodimer with a dimeric molecular weight of ~60,000 and contains one [4Fe:4S] cluster at the dimer interface (28). Two ATP molecules bind per 118 dimer of reduced Fe-protein, allowing a conformational change upon association with the MoFe-protein (29) . This necessary conformational change pushes the [4Fe:4S] cluster out so that it lies within 14 A of the P-cluster of the MoFe-protein and the ability of the protein to drastically alter its structure may be reflected in the conformational variability of the Fe-protein (30). Upon ATP hydrolysis , the complex between the MoFe-protein and Fe-protein disassociates, allowing the cycle to repeat until eight electrons are incorporated into N 2 . Concomitant with the reducing equivalents is the association of the substrate with protons, however a proton channel has not yet been conclusively established in the MoFe-protein. Nitrification and Nitrate Ammonification Nitrification and nitrite ammonification are essentially opposite reactions interconverting No 2- and NH/ (equation 5.2); nitrification cycles the electrons to terminal oxidases, whereas nitrite ammonification uses No2- as the terminal electron acceptor in the electron transport chain. (5.2) Nitrification removes biologically active NH/ from the soil in an obligatory aerobic process and converts it to nitrite with hydroxylamine as the intermediate. Although 119 farmers may resent the loss of NH/-containing fertilizer by nitrification, it is necessary for the production of a second important inorganic nitrogen compound that can be readily assimilated into the biomass: N03- ( 31). The bioenergetic pathway of the organism Nitrosomonas europaea represents the most extensively studied system of nitrification, however gene hybridization studies ( 32) suggest that similar enzymes are used for nitrification in other nitrifying bacteria, including Nitrobacter and Nitrosospira . Two enzymes function in the nitrification pathway: hydroxylamine oxidoreductase (HAO) and ammonia monooxygenase (AMO) (4, 33). The former is a trimer (Figure 5.1) containing eight hemes per monomer arranged in stacked pairs (Figure 5.3a), with the active site P460 heme covalently attached to a tyrosine (34 ). The latter is an integral-membrane protein with sequence identity to particulate methane monooxygenase that contains copper and iron cofactors ( 35-37). Electrons released by the oxidation of NH/ are transferred by a tetraheme cytochrome ( 38) (cytochrome c554) and a monoheme cytochrome ( 39) (cytochrome c552) to the terminal oxidases of the electron transport chain (40, 41). Intriguingly, the heme core of cytochrome c554 appears to be entirely contained within its redox partner HAO despite the lack of sequence identity (9 )(Figure 5.3b). The evolutionary relationship between HAO and cytochrome c554 will be discussed in detail in the following three chapters. 120 Nitrite ammonification catalyzed by cytochrome c nitrite reductase requires a protein acting as a physiological dimer (Figure 5.1) with five hemes per monomer (JO) . The hemes are stacked in the same manner as observed in hydroxylamine oxidoreductase and cytochrome c554 (10, 11, 42)(Figure 5.3c), suggesting evolutionary convergence between these proteins catalyzing the same reaction in opposite directions. The active site lies above a five-coordinate heme, where the fifth ligand is a lysine rather than a histidine that is normally found in the C-x-y-C-H c-heme binding motif (JO). Nitrite oxidation Nitrite oxidation converts No2- to N0 3-, which is more readily incorporated into the biomass (Figure 5.1). This reaction is catalyzed by an integral-membrane protein, nitrite oxidase, about which very little is known. Nitrite oxidase exhibits significant sequence similarity to a membrane-bound nitrate reductase that catalyzes the reaction in the opposite direction. As other soluble proteins are involved in nitrate reduction, most studies have focussed on these more facile systems. Denitrification Denitrification 1s the process by which bacteria that use nitrogen oxidized compounds rather than oxygen as the terminal electron acceptor in respiration. These metabolic processes return N2 to the atmosphere, completing the nitrogen cycle. The 121 commercial importance of denitrification lies in the depletion of N03- from fixed nitrogen fertilizer. Conversely, denitrification fulfills an important role in wastewater treatment (43, 44) . One disadvantage of the "leaky" denitrification process is the release of large amounts of the greenhouse gas N2 O into the atmosphere. Denitrification occurs in both assimilatory and dissimilatory pathways (44) . The assimilatory pathways (Figure 5. 1) reduce nitrogen oxide compounds in the physiologically opposite direction as nitrification such that the end product is NH/ (45, 46). The annual regeneration of NH 3 from NO{ and No 2- by assimilatory denitrification returns approximately 100 times the amount of NH3 that is generated by nitrogen fixation (47). The dissimilatory pathway of nitrification (Figure 5.1) reduces nitrogen to N2. Nitrate reduction The reduction of nitrate to nitrite is catalyzed by a heterogeneous group of molybdoenzymes that are similar in active site organization. Nitrate reductases (NaR) can be broadly categorized as assimilatory and dissimilatory. The assimilatory NaRs catalyze the reduction of NO 3- to NH/. The enzyme forms homodimers with a molecular weight of ~220 kD, and contains three cofactors: a molybdenum cofactor (MoCo; Figure 5.4a), ab-type heme, and an FAD/NADP+ binding site (48). Although no structural information is available for an assimilatory nitrate reductase, sequence analyses show that 122 sulfite oxidase can act as a structural model for the majority of this enzyme (49, 50) . Sulfite oxidase contains the domains associated with the MoCo and the b-type heme, but does not contain the domain that binds FAD/NADP+. The fold of sulfite oxidase can be divided into three domains that contain a mixed a/~ fold (Figure 5.1); the first domain has structural similarity to cytochrome b 5 , the second domain binding the MoCo appears structurally unique, while the third domain contains a greek-key motif (51) . The active site consists of a single molybdopterin bound deeply within the protein. Residues involved in cofactor binding are conserved between the sequences of nitrate reductases and sulfite oxidase, thus the active site of sulfite oxidase should represent a reasonable model for nitrate reductases. In sulfite oxidase, the substrate binds near the vicinity of the oxo-ligand to the MoCo in a positively charged pocket, suggesting its position during catalytic turnover. In contrast, the dissimilatory nitrate reductases (Figure 5.1) catalyze the reduction of NO3- to No 2-, and have significant sequence and structural similarity to DMSO reductase (52, 53) and formate dehydrogenase (54). The dissimilatory nitrate reductases (DNaR) act as the primary respiratory enzymes in nitrate reduction. DNaR folds into four distinct domains ( 5 5) with pronounced structural similarity to formate dehydrogenase (rms deviation of 1.47 A) but exhibits a slightly more divergent fold to DMSO reductase. DNaR acts as an ~80 kD monomeric enzyme with a molybdenum 123 atom coordinated to a molybdopterin guanine dinucleotide (MGD) as the active site (Figure 5.4b) . A [4Fe:4S] cluster positioned ~12 A away from the molybdenum transports electrons into the active site for nitrate reduction. The reaction catalyzed by DNaR involves binding of the substrate NO 3- to an open ligation position of the molybdenum, with the molybdenum removing one oxygen to form an oxo-ligand previously observed in several MGD containing protein structures. Nitrite reduction Like nitrate reduction, nitrite reduction can occur as an assimilatory or dissimilatory process. The assimilatory nitrite reductases (NiR) reduce No 2• to NH/ in a process with no detectable intermediates, with a flavodoxin likely acting as the phsyiological electron donor. Assimilatory NiRs have an active site containing a [4Fe:4S] cluster linked to a siroheme (56, 57) (Figure 5.5) and share significant sequence similarity to sulfite reductases. No structural information is currently available for an assimilatory NiR, although the structure of sulfite reductase (SiR) (58) can yield insight into the function of the enzyme. Indeed SiR can catalyze assimilatory denitrification (59, 60 ), albeit with a much slower rate than it catalyzes sulfite reduction or that physiological assimilatory NiRs catalyze denitrification. SiR comprises a monomer that folds into three domains with pseudo-twofold symmetry (58) (Figure 5.1). The non-covalently bound siroheme is positioned at the center of the three domains and acts as the active site. 124 The pocket surrounding the open ligation site of the five-coordinate siroheme contains numerous basic residues as well as many solvent molecules that participate in a complicated hydrogen-bonding network. This positively charged active site should have a high affinity for oxo-ligands, as reflected by the presence of a phosphate acquired from the crystallization conditions binding in the reduced form of the enzyme. In the presence of the substrate, the phosphate is replaced by sulfite. Numerous organisms and enzymes can catalyze the dissimilatory reduction of No 2- to NO, and two distinct classes of nitrite oxide reductases have been recognized: the heme cd 1-containing and the copper-containing nitrite reductases ( 1, 44). The heme cd 1containing nitrite reductases (CDNiR) are the more abundant in nature, whereas the copper-nitrite reductases (CuNiR) are perhaps better studied. CDNiRs are bifunctional enzymes that catalyze the one-electron reduction of nitrite to nitric oxide (equation 5.3) or the four-electron reduction of oxygen to water, however, No 2- reduction represents the main physiological reaction ( 61) . The physiological electron donor for this reaction appears to be cytochrome c551 (62, 63) . (5.3) 125 CDNiR contains two hemes: a c-type and an unusual non-covalently bound d 1-type heme that acts as the active site (61) . The enzyme acts as a functional dimer (64) with each monomer containing a helical domain ligating the c-type heme and a ~-propeller domain containing the d 1-heme (65) . Based on comparisons to c-cytochromes, the c-heme was predicted to have His-Met ligation. It was therefore a surprise when the crystal structure showed His-His coordination in the oxidized state. Later analysis of CDNiR under reducing conditions showed that a large conformational change occurs in the domain binding the c-heme, switching the distal ligand from His 17 to Met 106 (66) (Figure 5.6). Based on previous small angle x-ray scattering experimetns (67), a large conformational change of the protein had been expected. Like the c-heme, the catalytic d 1-heme undergoes ligand switching during catalysis (66) . The original structure determination showed that a tyrosyl phenoxy group ligated the open ligation position of the d 1-heme during the resting, oxidized state. In the presence of reaction intermediates or product, the tyrosine side chain rotates around chi 1 either to stabilize the product through a hydrogen bond, or to form an alternative hydrogen bond with the side chain of a threonine when the reaction intermediates are present (Figure 5.7) 126 Copper-containing NiRs (CuNiRs) appear pink, blue, or green depending on the state of the enzyme and physiologically accept electrons from azurin to perform nitrite reduction (2, 3, 44) (equation 5.3) . CuNiRs exist as trimers (monomeric molecular weight of ~35 kD) with two distinct copper centers ligated at the interface between the monomers (68). Copper center 1 exhibits similar ligation as the plastocyanin or azurin copper ligation and is known as the "blue" copper (69). The second, "non-blue" copper center exhibits tetrahedral coordination by three hisitidine side chains, two of which are provided by one monomer, while the third is provided by a second monomer, and a water molecule (68) . This second copper center is believed to represent the active site of the enzyme. Crystal structures show that N02• replaces the copper-bound water molecule, changing the almost perfect tetrahedral coordination to asymmetric five-coordinate ligation through the binding of two oxygen atoms of the No 2• (69-71) . However, most proposed reaction mechanisms involve the N-bound to the active site copper. Protons that reduce the oxygen to water may be shuttled to the active site by a conserved aspartate side chain. Nitric oxide reduction Nitric oxide (NO) serves as a messenger molecule for a variety of biological functions, however can be toxic to cells. In lower organisms, nitric oxide levels are diminished immediately after nitrite reduction by further reduction to nitrous oxide. 127 Nitric oxide reduction occurs via a cytochrome be complex in prokaryotes, and by a heme-containing enzyme in eukaryotic fungal systems (44) . The bacterial nitric oxide reductases (NORs) consist of an integral-membrane heterodimer with subunit molecular weights of 53 and 17 kD (72-76) , and appear to contain both low- and high-spin hemes as well as non-heme iron (76). The high-spin heme is likely five-coordinate thus may serve as the active site of the enzyme. Little else is known regarding the mechanism of reduction of NO by the bacterial NOR complex In the fungus Fusarium oxysporum, a monomeric water-soluble heme enzyme (77) with a molecular weight of 46,000 (Figure 5.1) catalyzes NO reduction (equation 5.4). (5.4) The fungal NO reductase (NOR) exhibits significant sequence and structural similarity to monooxygenase P450 superfamily (78). However, compared to structures of the active site in the P450 superfamily (79), the helices surrounding the open ligation site of the heme are pushed back, allowing a larger cavity for the reaction to take place (80). The reduction of NO occurs at a five-coordinate heme as the active site with a cysteine thiolate as the fifth axial ligand rather than a histidine normally present in the C-x-y-C-H 128 c-heme binding motif. The ferric state of the enzyme appears to be a non-functional resting state, while the ferrous state forms a water molecule network between the active site heme iron and solvent, suggesting a plausible proton transfer pathway . Crystallographic investigation has implicated one of these water molecules as a proton donor to the Fe-NO moiety during catalysis (81 ). There does not appear to be an obvious binding-site for NADH, and it has been suggested that NADH physically enters the active site after NO is bound to the heme, donating the electron directly to the substrate (80). Nitrous oxide reduction Nitrous oxide is a major greenhouse gas (6), thus understanding the biological reduction has environmental implications. The penultimate step of denitrification requires the conversion of N20 to N2 (equation 5.5). (5.5) Nitrous oxide reductase (N20R) has long been known to contain two copper centers: CuA and the active site Cuz (82). CuA has spectral properties similar to CuA from respiratory complex IV(83, 84), and the sequence ligating CuA to N20R shares significant similarity to the sequence that ligates Cu A of complex IV ( 85). It was therefore not surprising to find a similar fold associated with the ligation of CuA from 129 both proteins ( 12). The nature of the Cuz copper center of N2OR has remained more mysterious. Metal quantification and spectroscopy had suggested that Cuz likely existed as either a mono- or bi-nuclear cluster (86, 87). The novel tetra-nuclear cluster revealed by the crystal structure (12) was therefore a surprise. This cluster is ligated with distorted geometry to the polypeptide chain by seven conserved histidine ligands (Figure 5.8) and is associated with two additional water molecule ligands. It is believed that substrate binding occurs by replacement of one of the two water molecule ligands and that the tetra-nuclear center provides and electron-sink for catalysis. N2OR acts as a physiological dimer with dimeric molecular weight of 130 kD. Each monomer folds into a ~-propeller, however the dimeric associated is involves a rotation of the second monomer 90° with respect to that observed for dimeric association of the cytochrome cd 1 nitric oxide reductase (65) (Figure 5.1). Anaerobic ammonia Oxidation (ANAMOX) The process of anaerobic ammonia oxidation (ANAMOX) is the most recent nitrogen cycle reaction to be described (88). Bacteria that can catalyze this process can grow under chemolithoautotrophic conditions using ammonia as an electron donor and nitrite as an electron acceptor (equation 5.6) 130 (5 .6) The reaction intermediates for this pathway may include hydrazine (89), and can generate transmembrane potential if hydrazine oxidation occurs at the periplasmic face of the cytoplasmic membrane. Although the ANAMOX phenomenon has been known for two years, the first organism catalyzing ANAMOX was only discovered last year (90) and represents a new type of bacterium. The enzymology of the reaction(s) involved remains uncertain. The discovery of new pathways and organisms involved in the biological nitrogen cycle acts as a reminder that the current understanding of the interconversion of inorganic nitrogen may include many as yet undiscovered components. 131 Figure 5.1 (next page): The enzymatic components of the biological nitrogen cycle. Protein components of nitrogenase have been solved from Azotobacter vinelandii (lNIP, 2NIP (28) ; 2MIN, 3MIN (22) ; lDE0, (97)), Clostridium pasteurianum (lMIO, (18); 1CP2, (30)), and Klebsiella pneumonae (lQGU, lQHl, 1QH8, ( 19)). The Azotobacter vinelandii (lN2C, (29)) complex is shown. Hydroxylamine oxidoreductase has been solved from Nitrosomonas europaea (lFGJ, (34)). Cytochrome c nitrite reductase has been solved from Sulfurospirillum deleyianum (lQDB, ( 10)). The assirnilitory nitrite and nitrate reductases have been modeled from the structures of the Escherichia coli sulfite reductase (lAOP, (58); 2AOP, 3AOP, 2GEP, 3GEO, 4AOP, 4GEP, 5AOP, 5GEP, 6GEP, 7GEP, 8GEP, (96)) and the chicken liver sulfite oxidase (lSOX, (51)), respectively. Nitrate reducase has been solved from Desulfovibrio desulfuricans (2NAP). Cytochrome cd 1 nitrite reductase has been solved from both Thiospaera pantotropha (lQKS, (65); lAOF, lAOM, lAOQ, (66)) and Pseudomonas aeruginosa (1BL9, lNNO, (92); 1N15, 1N50, 1N90, (94 ); lNIR, (64 )). Copper nitrite reductases have been solved from Alcaligenes faecalis (lAFN, lNTD, (70); 1AQ8, 1AS6, 1AS7, 1AS8, (91)), Alcaligenes xylosoxidans (1BQ5 (93); lNDR, INDS, lNDT (71)), and Achromobacter cycloclastes (lNIA, lNIB, lNIC, lNID, lNIE, lNIF, (95)). Nitric oxide reductase has been solved from Fusarium oxysporum (lROM, 2ROM, (80); 1CL6 lCMJ, lCMN (81 )). Nitrous oxide reductase has been solved from Pseudomonas nautica (lQNI, ( 12)). 132 assimilatory nitrate reductase chrome c ite reductase nitrate reduct cytochrome cd1 , 1"- ~.~ nitrite reductase \~: ; ,/ ;, e.b, "✓ nitrogenase nitrous oxide reductse Figure 5.2: The cofactors of the nitrogenase FeMoCo system. The two views are separate by 90°. P-cluster The FeMoCo and P-cluster are located in the [4Fe:4S] Mo-Fe protein while the [4Fe:4S] cluster and ADP·ALF4 - are located in the iron protein. i.. ADP-AIF4- , . .. "' 6 _, . • e l • ~ ~ 133 Figure 5.3: Heme-packing motifs of proteins involved in nitrification and nitrite ammonification. The hemes of each protein are labeled in the order of their ligation to the polypeptide chain using the author's numbering scheme. A. The eight hemes of HAO. B. The four hemes of B cytochrome c554. C. The five hemes of cytochrome c nitrite reductase. All proteins are shown from the same relative C view. Intriguingly, a five- coordinate heme aligns in each of these proteins. In HAO (heme P460) and cytochrome c nitrite reductase (heme 515), the five-coordinate heme is associated with catalytic activity, whereas in cytochrome c554 (heme II), the only known function of the fivecoordinate heme is in electron transfer. 134 Figure 5.4: Molybdenum-containing cofactors (98). A. Stereoview of the molybdenum cofactor (MoCo) from sulfite oxidase. The pterin ring is highly bent. B. The MGD cofactor found in nitrate reductase requires two pterins with a nucleotide extention. The two cofactors are not shown to scale. A • • • p • • ., • V P ) • • B " ~ 135 Figure 5.5: Stereoview of the siroheme of sulfite reductase. The structure of the cofactor will be similar to that of the assimilatory nitrite reductase. A siroheme differs from a c-type heme in that the pyrrole rings are saturated with dicarboxylate groups. The iron of the siroheme is bridged to the [4Fe:4S] cluster by Cys 483 of sulfite reductase. The binding of phosphate to the distal side of the siroheme shows where substrate binds . • 136 Figure S.6: Ligand switching to the c-heme of the cd 1 nitrite reductase. The panel on the left shows the c-heme ligation under oxidizing conditions where His 17 acts as the heme iron ligand. Under reducing conditions (right panel), the long loop changes conformation with the simultaneous unfolding of the N-terrninus to residue 26. Met 106 then acts and the heme iron ligand. 137 Figure 5.7: Stereoview of the movement of Tyr 25 during catalysis by cytochrome cd1 nitrite reductase. In the oxidized form of the enzyme, Tyr 25 directly ligates the heme iron (blue position). Upon the addition of substrate, the tyrosine swings away (green position) to allow access to the binding site. During catalysis, Tyr 25 swings back such that it makes contact to the reaction intermediates (red position), possibly stabilizing the transition state. • ...) • 138 Figure 5.8: Stereoview of the CUz center of nitrous oxide reductase. The four copper ions of CUz form a diamond and are coordinated by seven histidine residues from the protein. The substrate N2O has been modeled to interact with Cu IV, replacing a position filled by an ordered water molecule in the structure determination. Pho 561'11> Pha S54b Hl11 437 fl Hi a 4!'!7 ' Ly~ 3!l "f HI!: JTG i 139 REFERENCES 1. Berks, B. C., Ferguson, S. J., Moir, J. W. B., and Richardson, D. J. (1995) Biochim. Biophys. Acta 1231, 97-173. 2. Ferguson, S. J. (1998) Curr. Op. Chem. Biol. 2, 182-193. 3. Richardson, D. J., and Watmough, N. J. (1999) Curr. Op. Chem. Biol. 3, 207-219. 4. Wood, P. M. (1986) in Nitrification (Prosser, J. I., Ed.) pp 39-62, IRL Press, Oxford. 5. Hutchins, H. (1991) Appl. Environ. Microbial. 57, 2403-2407. 6. Rasmussen, R. A. , and Khalil, M.A. K. (1986) Science 232, 1623-1624. 7. Howard, J.B ., and Rees, D. C. (1996) Chem. Rev. 96, 2965-2982. 8. Burgess, B. K., and Lowe, D. J. (1 996) Chem. Rev. 96, 2983-3011. 9. Iverson, T. M., Arciero, D. M., Hsu, B. T., Logan, M. S. P., Hooper, A. B., and Rees, D. C. (1998) Nat. Struct. Biol. 5, 1005-1012. 10. Einsle, 0., Messerschmidt, A. , Stach, P., Bourenkov, G. P., Bartunik, H. D., Huber, R., and Kroneck, P. M. H. (1999) Nature 400, 476-480. 11. Iverson, T. M., Arciero, D. M., Hooper, A. B., and Rees, D. C. (in preparation) J. Biol. Inorg. Chem. 12. Brown, K., Tegoni, M., Prudencio, M ., Oereira, A. S., Besson, S., Moura, J. J., Moura, I., and Cambillau, C. (2000) . 13. Topham, S. A. (1985) in Catalysis: Science and Technology (Anderson, J. R., and Boudart, M., Eds.), Springer-Verlag, Berlin. 140 14. Jennings, J. R. (1991) , Plenum Press, New York. 15. Eady, R.R. (1996) Chem. Rev. 96, 301 3-3030. 16. Alberty, R. A. (1994) J. Biol. Chem. 269, 7099-7102. 17. Kim, J., and Rees, D. C. (1992) Nature 360, 553-560. 18. Kim, J., Woo, D. , and Rees, D. C. (1993) Biochemistry 32, 7104-7115. 19. Mayer, S. M. , Lawson, D. M., Gormal, C. A. , Roe, S. M., and Smith, B. E. (1999) J. Mal. Biol. 292 , 871-89 1. 20. Kim, J., and Rees, D. C. (1992) Science 257, 1677-1682. 21. Chan, M. K., Kim, J., and Rees, D. C. (1993) Science 260, 792-794. 22. Peters, J. W., Stowell, M. H. B., Soltis, M. , Finnegan, M. G., Johnson, M. K., and Rees, D. C. (1997) Biochemistry 36, 1181-1187. 23. Hawkes, T. R., Lowe, D. J., and Smith, B. E. (1983) Biochem. J. 211, 495-497. 24. Hawkes, T. R. , McLean, P. A., and Smith, B. E. (1984) Biochem. J. 217, 317-321. 25 . Imperial, J., Hoover, T. R. , Madden, M. S. , Ludden, P. W. , and Shah, V. K. (1989) Biochemistry 28, 7796-7799. 26. Howard, J.B ., and Rees, D. C. (1994) Annu. Rev. Biochem. 63, 235-264. 27. Seefedt, L. C., and Dean, D.R. (1997) Acc. Chem. Res. 30, 260-266. 28. Georgiadis, M. M., Komiya, H., Chakrabarti, P., Woo, D., Kornuc, J. J., and Rees, D. C. (1992) Science 257, 1653-1659. 141 29. Schindelin, H., Kisker, C., Schlessman, J. L. , Howard, J.B ., and Rees, D. C. (1997) Nature 387, 370-376. 30. Schlessman, J. L., Woo, D., Joshua-Tor, L., Howard, J. B. , and Rees, D. C. (1998) J. Mal. Biol. 280, 669-685. 31. Prince, R. C. , and George, G. N. (1997) Nat. Str. Biol. 4, 247-250. 32. Bruns, M. A., Fries, J.M., Tiedje, J.M., and Paul, E. A. (1998) Microb. Ecol. 36, 293-302. 33. Hooper, A. B., Vannelli, T., Bergmann, C. J. , and Arciero, D. M. (1997) Anton. van Lieuwenhoek 71 , 56-67. 34. Igarashi, N., Moriyama, H. , Fujiwara, T. , Fukumori, Y., and Tanaka, N. (1997) Nat. Struct. Biol. 4, 276-284. 35. Shears, J. H., and Wood, P. M. (1985) Biochem. J. 226, 499-507. 36. Zahn, J. A., and Dispirito, A. A. (1996) J. Bacterial. 178, 1018-1029. 37. Zahn, J. A., Arciero, D. M., Hooper, A. B., and DiSpirito, A. A. (1996) FEBS Lett. 397, 35-38. 38. Andersson, K., Lipscomb, J., Valentine, M., Mtinck, E., and Hooper, A. (1986) J. Biol. Chem. 261, 1126-1138. 39. Tirnkovich, R., Bergmann, D., Arciero, D. M., and Hooper, A. B. (1998) Biophys. J. 75, 1964-1972. 40. Yamanaka, T., and Shima, M. (1974) J. Biochem. 75, 1265-1273. 142 41. Yamanaka, T. (1996) Plant Cell Physiol. 37, 569-574. 42. Iverson, T. M., Hendrich, M. P., Arciero, D. M., Hooper, A. B., and Rees, D. C. (in preparation) in Metalloprotein handbook (Poulos, T. , Ed.). 43 . Mellor, R. B. , Ronnenberg, J., Campbell, W. H., and Diekmann, S. (1992) Nature 355, 717-719. 44. Averill, B. A. (1 996) Chem. Rev. 96, 2951-2964. 45 . Cole, J. A. , and Brown, C. M. (1980) FEMS Microbiol. Lett. 7, 65-72. 46. Wray, J. L. , and Kinghorn, J. R. (1989), Oxford Science Publications, Oxford. 47. Campbell, W. H., and Kinghorn, J. R. (1990) Trends. Biochem. Sci. 15, 315-319. 48. Kubo, Y., Ogura, N., and Nakagawa, H. (1988) J. Biol. Chem. 263, 19684-19689. 49. Neame, P. J., and Barber, M. J. (1 989) J. Biol. Chem. 264, 20894-20901. 50. Wooton, J. C., Nicholson, R. E., Cock, J.M., Walters, D. E., Burke, J. F., Doyle, W. A. , and Bray, R. C. (1991) Biochim. Biophys. Acta 1057, 157-185. 51. Kisker, C. , Schindelin, H. , Pacheco, A., Wehbi, W. A. , Garrett, R. M., Rajagopalan, K. V., Enemark, J. H., and Rees, D. C. (1997) Cell 91, 973-983. 52. Schindelin, H. , Kisker, C. , Hilton, J., Rajagopalan, K. V., and Rees, D. C. (1996) Science 272, 1615-1621. 53. Schneider, F., Lowe, J., Huber, R., Schindelin, H., Kisker, C., and Knablein, J. (1996) J. Mol. Biol. 263, 53-69. 143 54. Boyington, J. c .; Gladyshev, V. N., Khangolov, S. V., Stadtman, T. C., and Sun, P. D. (1997) Science 275, 1305-1308. 55 . Dias, J. M., Than, M. E., Humm, A. , Huber, R., Bourenkov, G. P., Bartunik, H. D. , Bursakov, S., Calvete, J., Caldeira, J. , Carneiro, C., Moura, J. J. G., Moura, I., and Romao, M. J. ( 1999) Structure 7, 65-79. 56. Janick, P. A., Rueger, D. C., Krueger, R. J., Barber, M. J., and Siegel, L. M. (1982) Biochemistry 21, 3538-3547. 57. Wilkerson, J. 0 ., Janick, P. A. , and Siegel, L. M. (1983) Biochemistry 22, 5048-5054. 58. Crane, B. R., Siegel, L. M., and Getzoff, E. D. (1995) Science 270, 59-67. 59. Siegel, L. M., and Davis, P. S. (1974) J. Biol. Chem. 249, 1587-1593. 60. Janick, P. A., Rueger, D. C. , Krueger, R. J., Barber, M. J., and Siegel, L. M. (1983) Biochemistry 22, 396-408. 61. Moir, J. W. B., Baratta, D. , Richardson, D. J., and Ferguson, S. J. (1993) Eur. J. Biochem. 212, 377-385. 62. Greenwood, C., Barber, D., Parr, S. R., Antonini, E., Brunori, M., and Colosimo, A. (1978) Biochem. J. 173, 11-17. 63. Silverstrini, M. C., Tordi, M.G., Colosimo, A. A., E., and Brunori, M. (1982) Biochem. J. 203, 445-451. 64. Nurizzo, D., Silvestrni, M.-C., Mathieu, M., Cutruzzola, F., Bourgeois, s. D., Fulop, V., Hajdu, J., Brunori, M., Tegoni, M., and Cambillau, C. (1997) Structure 5, 1157-1171. 144 65. Fi.ilop, V., Moir, J. W. B., Ferguson, S. J., and Hajdu, J. (1995) Cell 81, 369-377. 66. Williams, P. A. , Fi.ilop, V. , Garman, E. F. , Saunders, N. F. W. , Ferguson, S. J., and Hajdu, J. (1997) Nature 389, 406-41 2. 67. Berger, H., and Wharton, D. C. (1980) Biochim. Biophys. Acta 622, 355-359. 68. Godden, J. W., Turley, S. , Teller, D. C., Adman, E.T., Liu, M. Y., Payne, W. J., and LeGall, J. (1991) Science 253, 438-442. 69. Dodd, F. E. , VanBeeumen, J., Eady, R. R., and Hasnain, S. S. (1998) J. Mal. Biol. 282 , 369-382. 70. Murphy, M. E. P., Turley, S., Kukimoto, M., Nishiyama, M., Horinouchi, S., Sasaki, H., Tanokura, M., and Adman, E.T. (1995) Biochemistry 34, 12107-12117. 71. Dodd, F. E., Hasnain, S.S., Abraham, Z. H. L., Eady, R.R., and Smith, B. E. (1997) Acta Cryst. D553, 406-418. 72. Heiss, B., Frunzke, K., and Zumft, W. G. (1989) J. Bacterial. 171, 3288-3297. 73. Carr, G. J., and Ferguson, S. J. (1990) Biochem. J. 269, 423-429. 74. Dermastia, M., Turk, T., and Hollocher, T. C. (1991) J. Biol. Chem. 266, 1089910905. 75. Jones, A. M., and Hollocher, T. C. (1993) Biochim. Biophys. Acta 1144, 359-366. 76. Kastrau, D. H . W ., Heiss, B. , Kroneck, P. M . H ., and Zumft, W. G. (1994) Eur. J. Biochem. 222, 293-303. 145 77 . Shiro, Y., Fujii, M. , Isogai, Y., Adachi, S., Iizuka, T ., Obayashi, R., Makino, R. , Nakahara, K., and Shoun, H. (1 995) Biochemistry 34, 9052-9058. 78. Kizawa, H., Tomura, D., Oda, M., Fukamizu, A. , Hoshino, T., Gotoh, 0 ., Yasui, T. , and Shoun, H. (1991) J. Biol. Chem. 266, 10632-10637. 79 . Hasemann, C. A., Ravichandran, K. G. , Boddupalli, S. S., Peterson, J. A., and Deisenhofer, J. (1 995) Structure 3, 41-62. 80. Park, S.-Y., Shimizu, H. , Adachi, S. , Nakagawa, A. , Tanaka, I., Nakahara, K. , Shoun, H., Obayashi, E. , Nakamura, H. , Iizuka, T., and Shiro, Y. (1997) Nat. Struct. Biol. 4, 827832. 8 1. Shimizu, H. , Obayashi, E. , Gomi, Y., Arakawa, H., Park, S.-Y. , Nakamura, H. , Adachi, S., Shoun, H. , and Shiro, Y. (2000) J. Biol. Chem. 275, 4816-4826. 82. Zumft, W. G. (1997) Microbial. Mo l. Biol. Rev. 61 , 533-616. 83. Kroneck, P. M. H. , Antholine, W. E., Riester, J., and Zumft, W. G. (1988) FEBS Lett. 242, 70-74. 84. Antholine, W. E., Kastrau, D. H. W., Steffens, G. C. M., Buse, G., Zumft, W . G., and Kroneck, P. M. H. (1992) Eur. J. Biochem. 209, 875-881. 85. Holm, L. , Saraste, M. , and Wikstrom, M. (1987) EMBO J. 6, 2819-2823. 86. Farrar, J. A., Thomson, A. J. , Cheesman, M. R., Dooley, D. M., and Zumft, W. G. (1991) FEBS Lett. 294, 11-15. 146 87. Farrar, J. A. , Zurnft, W. G., and Thomson, A. J. (1998) Proc. Natl. Acad. Sci. USA 95, 9891-9896. 88 . Van de Graf, A. , deBruijn, P., Robertson, L. A., Jetten, M. S. M., and Kuenen, J. G. (1998) Microbiology 143, 2415-2421. 89. Schalk, J., Outstad, H., Kuenen, J. G., and Jetten, M. S. (1998) FEMS Microbial. Lett. 158, 61-67. 90. Strous, M., Fuerst, J. A., Kramer, E. H. M. , Logemann, S., Muyzer, G., van de PasSchoonen, K. T., Webb, R., Kuenen, J. G., and Jetten, M. S. M. (1999) Nature 400, 446449. 91. Murphy, M. E.P., Turley, S., Adman, E.T. (1997) J Biol Chem 272, 28455-28462. 92. Nurizzo, D., Cutruzzola, F., Arese, M., Bourgeois, D., Brunori, M., Cambillau, C., Tegoni, M. (1998) Biochemistry 37, 13987-13995. 93 . Inoue, T. , Gotowda, M., Deligeer, Kataoka, K. , Yamaguchi, K., Suzuki, S. , Watanabe, H., Gohow, M., Kai, Y. (1998) J.Biochem.(Tokyo) 124, 876-889. 94. Nurizzo, D., Cutruzzola, F. , Arese, M., Bourgeois, D., Brunori, M., Cambillau, C., Tegoni, M. (1999) I.Biol.Chem. 274, 14997-15008. 95. Adman, E.T., Godden, J. W., Turley, S. (1995) J Biol Chem 270, 27458-27470. 96. Crane, B. R., Siegel, L. M., Getzoff, E. D. (1997) Biochemistry 36, 12120-12135. 97. Jang, S. B., Seefeldt, L. C., Peters, J. W. (2000) Biochemistry 39, 641-650. 98. Schindelin, H., Kisker, C., Rees, D.C. (1997) J. Biol. Inorg. Chem. 2, 773-781. 147 Chapter 6 Reprinted with permission from: Iverson, T.M. , Arciero, D .M ., Hsu, B.T., Logan, M .S.P., Hooper, A.B ., Rees, D.C. (1998) Heme-packing motifs revealed by the structure of the tetra-heme cytochrome c554 from Nitrosomonas europaea. Nat. Struct. Biol. 5: 1005-1012. Copyright 1998, Nature America, Inc. 148 - -- - - -- - - - ~~ 1998 Nature America Inc. • http://structbio.nature.com articles Heme packing motifs revealed by th e crystal structure of the tetra -heme cytoch rome c554 from Nitrosomonas europaea Tina M. Iverson1. David M. Arciero 2, Barbara T. Hsu3. Michael S. P. Logan2.4, Alan B. Hooper2 and Douglas C. Rees 3-5 E 8 ~ .3 ('(I C: Cytochrome c554 (cyt c554) , a tetra -heme cytochrome from Nitrosomonas europaea, is an essential component in the biological nitrification pathway. In N. europaea, ammonia is converted to hydroxylamine, which is t hen oxidized to nitrite by hydroxylamine oxidoreductase (HAO). Cyt c554 functions in the latter process by accepting pairs of electrons from HAO and transferring them to a cytochrome acceptor. The crystal structure of cyt c554 at 2.6 A resolution shows a predominantly a-helical protein with four covalently attached hemes. The four hemes are arranged in two pairs such that the planes of the porphyrin rings are almost parallel and overlapping at the edge; corresponding heme arrangements are observed in other multi-heme proteins. Striking structural similarities are evident between the tetra-heme core of cyt c554 and hemes 3- 6 of HAO, which suggests an evolutionary relationship between these redox partners. 0 ~ ! C. t: .J::. Cyt c554 (cyt c554) ls a tetra-heme 1 cytochrome involved in the oxidation of ammonia to nitrite in the chemoautotrophic bacterium Nitrosomonas curopat•a2. Oxidation of ammonia is the sole energy source for N. europaea and the pathway by which this occurs is summarized schematically in Fig. I. As the mobile electron carrier between hydroxylamine oxidoreductasc (HAO) and the monoheme cytochrome c552 (cyt c552) 1, cyt c554 plays a critir.al role in the process of oitriflrntion In the blologlral nitrogen rycle. Cyt c554 does not exhibit significant sequence homology lo any other known protein, including other tetra-heme cytochromes, and is considered to be in its own class3 . The name reflects the absorbance peak of the a -band at 554 nm in its fully reduced form' . Cyt c554 has unusual spectroscopic properties that change dramatically above pl-I IO or below pH 4. These studies have established that three of the hemes are low-spin while the fourt h is high-spin, and that the four hemes Interact magnetically!. The reduction potentials of the four hemes at pH 7. 0 have been determined' by a combination of voltam metry and spectroscopy, using an optically transparent th in layer electrode. The high-spin heme has a red uction potentia l of +4 7 mV, while the low-spin hemes titrate with midpo int potentia ls of +47 mV, - 147 mV, and -276 mV. Stopped-flow kinetic studies of electron transfer at physiological pH revealed that one electron is passed from HAO lo cyt c554 with a rate constant of 250-300 s· 1, whi le a second electron is passed ten times more slowly with a rate constant of 25-30 s 1 (ref. 5). The order of reduction of the hemes under physiological conditions Is unknown . + H20 Fig. 1 The oxidation of NH, to NO,· by Nitrosomonas europaea. Ammonia is first oxidized to hydroxylamine (NH20H) by ammonia monooxygenase (AMO.) The product. NH20 H. is oxidized to N0 2• by HAO. The released electrons are transferred to cyt c554 (a two electron acceptor) and then possi- bly to cyt c552 (a one electron acceptor). which ultimately passes electrons to terminal oxidases. The electron transfer pathway following the oxidation of NH 20H to NO,- is not fu lly understood and alternative electron transfer routes may exist". including the transfer of electrons from cyt c554 directly to a membrane-bound electron transport chain". 'Graduate Oplion in Biochemistry, 14 7-75 CH. California lnslilule of Technology. Pasadena, California 91125, USA. 2Departmenl of Biochemistry. University of Minnesota, St. Paul. Minnesota 55108, USA. ' Division of Chemistry and Chemical Engineering, 147-75 CH, California Institute ofTechnology. Pasadena, California 91125, USA. 'Present address: Biotechnology Center for Agriculture and the Environment. Rutgers Univer~ty. New Brunswick. New Jersey 08903. USA Howard Hughes Medical Institute. Pasadena. ca1ifornia 91125. USA. Correspondence should be addressed to D.C.R. email: dcrees@caltech.edu nature structural biology• volume 5 number 11 • november 1998 1005 149 - - - - - - - - - - -~ 1998 Nature America Inc. • http://structbio.nature.com articles Fig. 2 a. A schematic drawing illustrating the overall fold ofcytc554. The chain can be traced by starting at the blue N-terminus and ending at the red C-terminus. b, Stereo view of the Cu. trace in the same orientation as (a). c. Alignment of the sequence of cyt c554 with secondary structure elements. a-helices are shown in green, 3, 0-helices are shown in blue, and ~-strands are indicated by yellow arrows. Hemes are named according to the order in which they are linked to the sequence. Sites for the covalent attachment of the hemes are highlighted in red, with the heme number indicated underneath. Likewise, histidines that serve as the distal ligand for a heme are in red, with the heme to which they are connected indicated by a Roman numeral underneath. Figs. 2-4 were made using MOLSCRIPT". BOBSCRIPT 33 and RASTERJD". 200 200 E 8 e! ~C: c:i ~ s ~ i:i . :i::: .,;;. c.i ..!: 75 75 C (1J (.) -~ E <( ADAPFEGRKKCSSCHKAQAQSWKDTAHAKAMESLKPNVKKEAKQKAKLDPAKDYTQDKDCVGCHV 65 Heme I Q) 5 "' z 00 a, a, a IV Heme II DGFGQKGGYTIESPKPMLTGVGCESCHGPGRNFRGDHRKSGQAFEKSGKKTPRKDLAKKGQDFHF 130 Heme III I ->- ---- EERCSACH LNYEGSPWKGAKAPYTPFTPEVDAKYTFKFDEMVKEVKAMHEHYKLEGVFEGEPKEK 195 Heme IV III FHDEFQASAKPAKKGK 211 Crystallographic studies of cyt c554 were undertaken ln order to provide a structural basis for the participation of this protein In the electron transfer reactions associated with ammonia oxi· dation. The arrangement of the heme groups and the interactions with the surrounding polypeptide chain will further provide a structural framework for a more detailed analysis of the specLroscopic and electron transfer properties of the protein. Overall fold The structure of cyt c554 crystallized at pH IO.I has been solved to 2.6 A resolution by the method of multiple isomorphous replacement. Data collection and refinem ent statistics arc listed in Table l. The secondary structure of cyt c554 consists primarily of a-helices, wilh lwo slrelches of 3 10-helix and two short, parallel ~-strands (Fig. 2a,b) . A search of the Brookhaven Protein Data bank using the EMBL Dali servcr 6 did 1006 not reveal significant similarity to any previously determined protein folds; however. as detailed below, striking similarities arc evident in the heme arrangement and the fold of the polypeptide chain su rrou nding the hemes between cyt c554 and HAO. Heme configuration All four hemes in cyt c554 are attached lo the protein through thioether linkages to the c-heme binding motif, -C-x-y-C-H-. As shown in Fig. 2c, where the sequence of cyt c554 has been aligned with its secondary structure elements, the hemes have been numbered in the order that they are covalently connected to the polypeptide chain. The overall heme arrangement, with the inter-heme iron distances indicated, is shown in Fig. 3a. The four hemes arc arranged as two di -heme pairs. Within each pair, the heme planes are roughly parallel with each other. nature structural biology • volume 5 number 11 • november 1998 150 _ _ _ _ _ _ _ _ ___,~ 1998 Nature America Inc. • http://structbio.nature.com articles b E 8 ~ ::, roC: 0 :a c3 ~ :::: ii . t:: .,:;;; ci -= Ol {) -~ E e <( fl! ::, ro z CX) C) C) a Fig. 3 Heme configuration and superposition of heme stacking motifs. a, Inter-heme iron distances and overall heme configuration. Hemes are shown in the same view as in Fig. 2a. b, Overlay of hemes I and Ill from cyt c554 (red) with hemes 1 and 2 from HAO (teal), hemes 3 and 5 from HAO (magenta), hemes 6 and 7 from HAO (green) and hemes 1 and 2 from t he split Soret cytochrome (yellow) . The pseudo two-fold axis has been ca lculated for the hemes of cyt c554 and is shown as a black line. c, Overlay of hemes II and IV from cyt c554 (red) w ith P460 and 6 of HAO (teal) . The 5coordinate hemes II (cyt c554) and P460 (HAO) are on the right. The pseudo two-fold axis ca lculated for hemes II and IV of cyt c554 is indicated as a black line. d, Overlay of t he two central hemes, hemes Ill and IV, from cyt c554 with hemes 5 and 6 from HAO (teal), hemes 7 and 8 from HAO (green), hemes 69 and 70 from cyt c551.5 (magenta) and hemes 201 and 203 from cyt c, (yellow). e, Stereoviewof the overlay of all hemes and surrounding structural elements from cyt c554 (red) with hemes 3-6 of HAO (teal). Heme I is located at one end of the molecule and heme II is at the opposite end. Hemes III and IV are located towards the middle of the mo lecule, with heme III adjacent lo heme I and heme IV adjacent to heme II. Table 2 summarizes the interplanar geometry between the hemes. The heme organization of cyt r.554 differs substantially from that observed in either the lelraheme cytochrome c/ or the tetra-heme cytochrome associated with the photosynthetic reaction center8. Two distinct types of di-heme packing arrangements, parallel and perpendicular, occur in cyt c554 with the hemes in van der Waals contact. Significantly, both types of packing arrangements nature structural biology • volume 5 number 11 • november 1998 are present in other multi-heme cy tochromes7·" 11 , and hence likely represent common motifs that will continue to emerge as more structures of multi-heme cytochromes become available. These heme-packing motifs may reflect favorable heme-heme interactions that facilitate electron transfer processes between hemes. At present, all multi-heme proteins exhibiting these types of packing arrangement~ have been iso lated from organisms that either oxidize or reduce nitrogen or sulfur containing compounds as their primary energy source; this correlation may refl ect the multi -electron rcdox processes associated with the energy metabolism of these organisms. 1007 151 - - - - - - - - - - -~ 1998 Nature America Inc.• http://structbio.nature.com articles . a .. .--~.. Glu 126 His 102 • • Hcmcl Heme I b c..i .s ~ -~ Glul31 ~ ~ :::, 1ii C z <Xl Cl) Cl) Fig. 4 Unusual aspects of the structure of cyt c554. a Stereoview of heme I. Note the iron coordination by the N6 atom of His 102. The model is superimposed onto a"• weighted 21fol • IFd style omit map around His 102 calculated using the program RfFMAC" after the truncation of His 102 to an Ala. b.c. The His 129-Phe 130 cis-peptide bond. (b) The final model i11 the region of Phe 130 superimposed onto a "• weighted 21f0 I · IFJ omit style map calculated using the program REFMAC after the deletion of His 129 and Phe 130 from the model. (c) Crystal contacts between His 129 and Glu 131 show that crystal contacts could influence the formation of the cis-peptide bond. I and III of cyt c554 with a root-mean-square (r.m.s.) deviation of 0.5-0.7 A, and the interheme iron distances vary between 9.0- 9.8 A.. The methyl carbon on pyrrole ring C is in direct van dcr Waals contact with the histidine coordinating the adjacent heme, making closer pack· ing in this manner impossible. In the second parallel stacking arrangement adopted hy hemes II and IV of cyt c554, the hemes arc again approximately related by .a two-fold rotation axis that is nearly parallel no the heme planes. Relative lo the llrst parallel arrangement, pyrrolc rings B and D arc adja· cent to the two-fold axis, and each heme Is then further rotated by ~30' about the normal to its plane (Fig. 3c) . Hemes P460 and 6 of HAO exhibit this arrangement, and may be superimposed with hemes II and IV of cyt c554 with an r.m.s. deviation of 0.51 A. Interestingly, in both of these pairs, one of the hemes is 11ve-r.oordinate (heme TI or cyt c554 and heme P460 of HAO) . A third parallel stacking arrangement may be identified in the photosynthetic reaction center 8 between the two bactPriochiorophyH groups that form the speci al pair. This arrangement is similar to the preceding one except that pyrrole rings Band C sandwich the rotation axis . In the reaction center, there Is a much more extensive overlap between the chlorophyll planes, with the magnesiums separated by 7.6 A. This closer distance is a consequence of,the acetyl group of each chlorophyll coordinating the other magnesium. In all known protein-associated hemes , Including cyt c554, non -porphyrin groups serve as the fifth and sixth ligands (when present) that prevent this close of an interaction. The second distinct type of heme packing arrangement seen in cyl c554, the perpendicular packing arrangement, is exemplified by the two central hemes, III and IV. The perpendicular packing arrangement orients the heme planes at a 11 0' angle to each other, with a heme iron separation of 12.2 A. This type of heme packing has previously been seen in HAO 10 , cyt c55 L5 9 and ,cyt c/ (Fig. 3d) . These di-heme clusters each superimpose with hemes II[ and IV of cyt c554 with r.m.s. deviations of The first of these di-heme packing arrangements involves parallel but offset packing of the heme groups. Two variants of the parallel packing arrangement occur In cyt c554 (Fig. 3b,c) , with counterpart~ present in other multi-heme proteins. In both cases, heme planes arc separated by a distance of closest approach of 3.5-3.8 A between ring atoms, while the distance between irons in the heme pair is -9 A. In the first parallel arrangement, adopted by hemes I and III of cyt c554. the two hemes are approximately related by a two-fold rotation axis that is nearly parallel to the heme planes. Pyrrole rings A and B of each heme straddle the two-fold axis, with lhe heme edges alternatively parallel and perpendicular to this axis (Fig. 3b) . The two pairs of hemes of the dimeric split Soret o.s-0.9 A. Intriguingly, the four hemes of cyt c554 may be structurally cytochrome", and three heme pairs in HAO 10 (hemes I and 2; hemes 3 and S; and hemes 6 and 7) all exhibit this packing aligned with hemes 3-6 of HAO (r.m.s. deviation 1.2 A) and arc arrangement. Each of these heme pairs su perimposes with hemes connected co the protein in the same order, which is suggestive 1008 nature structural biology • volume 5 number 11 • november 1998 152 - - - - - - - - - - -~ 1998 Nature America Inc. • http ://structbio.nature.com articles of an evolutionary relationship between these two redox partners. The exact relationship is diff1cult to establish, however, since no significant sequence and structural similarities can be detected between cyt c554 and HAO by automated searches. A side-by-side alignment of the structures using the EMBL Dali server6 results in a Z-score of 0.5, where a Z-score less than 2.0 is considered structurally dissimilar. However, when the hemes of cyt c554 are overlaid with hemes 3-6 of HAO, there is secondary structure agreement between the two molecules in the immediate heme core environment (Fig 3e). Furthermore, the polypeptide chains surrounding these hemes in cyt c554 and HAO follow the same general pal hs, even though they cannot always be superimposed . Although the overall sequence similarity is poor in the regions where t he secondary structure overlays, side chains with similar properties overlap, lead ing to a similar heme core. Given the structural and sequence differences. a common evolutionary ancestry cannot be convincingly established, but it is striking that the only two examples of this tetraheme arrangement have been found in two proteins that arc redox partners. b cyt c554 unoccupied binding site cyt c554 unoccupied binding site Unusual heme properties The distal histidine ligand of heme I, His 102. is coordinated lo the iron through Nii rather than through NE (Fig. 4a). A survey of heme iron coordination using Fig. 5 a, Stereoview of the GRASP 35 electrostatic surface potential representation of cyt c554 PROMISE 12 shows that none of the cur- illustrating the large positive ridge (right side) proposed to represent the HAO docking site. The rently available heme-protein structures color scheme shows positively charged surface area as blue, negatively charged surface area as has a heme iron coordinated with the Nii red, and uncharged surface area as white . b, Stereovicw of the docking model between cyt c554 of a histidine. It has been proposed that and the HAO trimer. The view is down the three-fold axis of the HAO trimer. One molecule of cyt c554 (yellow) is shown docked to HAO, while a second molecule of cyt c554 is pulled away heme iron coordination by the histidine from HAO to illustrate the complementary surfaces of the two molecules, in terms of both Nii is sterically unfavorable 13 ; the structure shape and electrostatic potential. The third putative binding site for cytochrome c554 on HAO of cyt c554 rev~ais that this need not be the has been left unoccupied at the bottom of the figure . c, Further rotation of cyt c554 from the view in (a) shows a negative cavity on the reverse side of the molecule . The function of this negcase. The side chain torsion angles for His ative cavity is not known . 102 have favorable values. However, there appear lo be limited backbone conformations that will allow heme iron coordination by the histidine No. As a result, It does not seem that a No Phe 156 pack against the porphyrin ring, shielding the sixth coordination would be achievable In a -C-x-y-C-H - heme bind- coordination site. It has been postulated that the five-coordiing motif where the backbone Is conformationally constrained. nate heme would have an open s ite capable of binding to a ligThis is reflected in the great preference for NE heme iron coordi- and , possibly NO, and therefore may participate directly in the nation that is seen. In addition to the Fe coordination, the orien- reaction oxidizing ammonia to nitrite as well as being an electation of this residue is stabilized by a hydrogen bond between tron carrier'. In Lhe comparison between cy l c554 and HAO NE and the side chain of Glu 126. Glu 126 is further hydrogen discussed previously, heme II corresponds to heme P460 which bonded to a water {Wl3) that is in turn hydrogen bonded to Lhe forms the active site for HAO and is also five-coordinate. While Nii of His 92, which acts as the proximal ligand to heme Ill . This this suggests that heme 11 may have a catalytic function , the hydrogen bonding network between the two hemes may con- crystal structure of cyt c554 indicates that this is unlikely in the tribute to the pH dependent interaction between hemes that has absence of conformational rearrangements involving residues that limit access to this heme . been observed spectroscopically. In the vicinity of heme III, a els-peptide bond occurs between In contrast to the other three hemes, heme II is five -coordinate and thus likely the +47 mV high-spin heme 14 . No amino His 129 and Phe 130. This cis-peplide bond is localed in a loop acid side chains or solvent molecules arc close enough to coor- at the end of one of the two ~-strands, with Phc 130 beginning dinate the heme iron. The side chains ofThr 154, Pro 155, and the next a -helix {Fig. 4b) . The aromatic side chain of Phc 130 nature structural biology• volume 5 number 11 • november 1998 1009 153 - - - - - - - -- - -~ 1998 Nature America Inc. • http ://structbio.nature.com articles Data sets High resolution (A) Completeness(%) Reflections Unique reflections I/a R,,.,'(%) Phasing Sta'tistics E 8 ~ .2 «J c:: d B E (.) Table1 Summary of crystallographic statistics Nat1 HoCI, Nat2 Ir'· TMLA 2.6 3.0 3.7 2.8 3.3 92.7 (92 .1) 97.8 (97 .2) 97 .o (100.0) 95.6 (88.3) 94.3 (95.9) 51,408 33,016 12,161 26,175 21,005 11,956 8,303 4,497 9,925 6,175 10.3 (3 .4)' 10.6 (3.7) 5.03 (3.1 ) 9.6 (2.5) 10.4 (5.1) 9.7 (29.1) 14.7 (32 .7) 18.4 (30.9) 10.6 (23.0) 11 .4 (24 .3) Phasing power' for each resolution bin (A) No. of sites 9.99 6.5 7 5.23 4.47 3.95 Rcu111s" 4 0.91 1.99 1.72 1.44 1. 14 1.10 0.74 2.95 2.39 3 1.89 1.52 1.31 0.85 1.19 0.62 3 1.39 0.88 0.65 10 0.61 3.75 3.78 3.06 2.27 2.04 10 0.82 2.57 2.31 2.07 1.74 1.43 0.79 2.64 1.34 7 2.63 2.10 1.45 0.75 2.27 2.04 1.63 1.27 1.14 0.84 0.73 0.61 0.80 0.65 Native Anomalous HoCl 3 HoCl 3 (anomalous) lr 3 • lr 3• (anomalous) TMLA TMLA (anomalous) Figure of mer·i t Refinement statistics Resolution range R,..,,.5 Rtrac6 R.m.s. deviation bond lengths R.m.s. deviation 'bond angles o.oisA 1.89' ~ ii :i:: .c: u ' R,,.,. is the same as R,,, .. for 10% of the data omitted from refinement totaling 1,167 reflections. Q) E <( ~ :, 1il z co O> O> Overall (3.6 A) 1.45 2.04 0.99 2.93 1.88 1.94 1.62 0.68 20-2.6 A 20.8% 25.4% 'Numbers in parentheses indicate values for the highest resolution bin. . 'R,,.., = l:(11 • <l>)i:E[<I>] where i is itt, measurement and <I> is the weighted mean of 1. ' Phasing power= The mean value of the heavy atom structure factor divided by the lack of closure error. •Ra,,,., = l:l\lF"•l · IF~-IF11ll i J:!IFa,I•IF11II where Fe and Fa , are the native and derivative structure factor amplitudes. E ~ ·;:: 3.60 0.916 1.14 0.58 1.94 1.18 1.23 0.97 0.53 'Ra,,, =i:ur....1•jF w•ll / l:jF .~1 lies within 4 Aof the porphyrin ring of heme III. There is the possibility that the cis conformation of the peptide bond is preferentially stabilized by a nearby crystal contact. Crystal contacts have previously been implicated in stabilization of a functionally relevant cis-peptide bond in the structure of flavodoxin , where Asn 137 from a symmetry-related molecule is directly hydrogen bonded to the backbone 15 . In cyt c554, the two residues adjacent to the els-peptide bond , His 129 and Ciu 131. interact around a crystallographic two-fold axis, which may stabilize the cisconformation (Fig.4c) . A fundam enta l property of electron transfer proteins such as cyt c554 is the electrochemical behavior of the associated cofactors. This is typically characterized In terms of the reduction potentials that measure the thermodynamic tendency for these groups to accept an electron. The factors that influen<:1! the reduction potentials of electron transfer proteins have been extensively studied (reviewed in refs 16,17) . and an important objective is to understand the overall redox behavior of the component cofactors. In the case of cyt c554, it is likely that the five coordinate heme represents the high spin heme that titrates al +47 mV 11 . Given the sensitivity of the reduction potentials to various structural and environmental factors. including imeraclion between the hemes (cooperalivity effects), pH and oxidation state linked structural changes, however, a specific assignment between the remaining low-spin hemes and the three six-coordinate hemes does not seem warranted, in the absence of more detailed information co ncerning these effects. energetics of protein-protein interactions. Surface potential representations were generated to search for reasonable interaction sites. There is a large ridge of posilively charged surface area above heme l of cyt c554, which could be involved in binding to HAO (Fig. 5a). This positive ridge is 30 Awide, which is almost exactly the size of the negative cavity between the monomers in the HAO trimer, where it was proposed that cyt c554 may bind 10 . Although speculative, docking between cyt c554 and HAO at this site appears qualitatively reasonable (Fig. 5 b). In this model, the solvent exposed edge of heme I of HAO is adjacent to the solvent exposed edge of hcnir I of cyt c554, with an iron-iron distance between the two hemes of 20 A, and a closest distance between heme rings of 8 A. Although electrostatic effects have been proposed to influence the binding between cyt c552 and cyt c554, there is no clear binding region on cyt c552 that could take advantage of their respective electrostatic charges. Despite the high pl of cyt c554 , there ls a negatively charged cavity on the opposite side of the molecu le from the positive ridge (Fig. Sc) . The function of this negative cavity is unknown. The 2.6 Aresolution crystal structure of cyt c554 has revealed striking structural similarities in the structure of the telraheme core lo that observed in its redox partner HAO. Furthermore, this tetra-heme arrangement is composed of diheme interactions that have been identified in other multiheme proteins that part icipate in electron transfer reactions. From the structures of the individua l cyt c554 and HAO mole cules, initial proposals Into the electron transfer processes Binding of cytochrome c554 to HAO and cytochrome c552 between these components can be formulated: more substantial Based on t he pl's of HAO (3.4) 18 • cyt c554 (10 .7) and cyt c552 progress will undoubtedly require determination of the (3 .7)2. it is plausible that electrostatic effects contribute to the HAO-cyt c554 complex. This structure may also provide further 1010 nature structural l)iology • volume 5 number 11 • november 1998 154 ~ 1998 Nature America Inc. • http://structbio.nature.com ------- ----~ insight into the question of why so many hemes are required for the oxidation of hydroxylamine. Methods Crystallization. Cyt c554 was produced and purified as described'. Crystals of oxidized eye c554 were obtained at 22 ' C by the vapor diffusion method using sitting drops with 5 µI of 40 mg ml- 1 protein plus 5 µI of the reservoir solution (62.5% w/vol potassium phosphate pH 10.1, prepared by titrating a solution of 92.5% w/vol K2HPO, to pH 10.1 with KOH, then diluting). In a solution of 35% w/vol potassium phosphate pH 10.1, oxidized native protein shows an EPR spectrum identical to that seen at physiological pH, with a g-value of 3.31 (data not shown) . Crystals belonging to the tetragonal space group P4 3 22 with a = b = 68.20 A and c = 168.83 A grew within 2 weeks to a maximal size of0.5 mm x 0.5 mm x 1.5 mm. Although the volume of the unit cell suggested two molecules per asymmetric unit, the crystals had one monomer per asymmetric unit. with 68% solvent content (V,,, = 3.8) , Crystals were cryo-cooled in the mother liquor using no additional cryo protectant. E 8 ~ .3 ca C: 0 :.c t5 g Cf) ::::: ci. :I:: ,:;;. rl -~ E <( I!! :, "' z a) a, a, articles Table 2 lnterheme angles and distances Heme pair I and II I and Ill I and IV II and 111 II and IV Ill and IV Angle' 88.4' 13.6' 110.8' 84 .1' 24 .7' 108.3' Pa ir? no yes no no yes yes lnterheme Fe'' distance (A) 28.0 9.3 20.7 20.5 9.5 12.2 'Inter-heme plane angles were calculated using the program MG86". Patterson map of the Ho'· derivative was interpreted using SHELX 25 • Sites for the other two heavy atom derivatives were determined using d.ifference Fourier maps phased with the Ho 1 • derivative. Heavy atom positions were initially refined using MLPHARE" and the refined positions and occupancies were further refined using SHARP 26 to 3.6 A resolution. The phases from SHARP were improved using solvent flattening with OM" to 3.5 A resolution. Preparation of derivatives. The Ho'· derivative w as prepared by Phasing statistics are given in Table 1. The positions of the hemes soaking mother liquor containing 5 mM HoCI, for 5 h. Interestingly, in the initial experimental maps were striking and each heme iron Ho,. formed a derivative even though the crystallization conditions had experimental density greater than 5cr associated with it. Using were high in phosphate. as phosphate has been shown to compete the center of the 5cr density as the location of each of the heme with proteins for the binding of lanthanides 19. Although three irons, the anomalous scattering contribution of the irons was phosphates are included in the final model , none of the Ho'' ions is added to the phase calculation . bound to any of these phosphates. The trimethyl lead acetate (TMLA) derivative was prepared by soaking crystals in mother liquor Model building and refinement. All 211 residues of cytochrome containing 30 mM TMLA for three weeks 20 . Because of the extended <;554 were built into the maps at 3.5 A resolution using the prolength of time of the derivative soak, a small amount of uncrystal- gram 0 27 . Subsequent rounds of modeling employed phase comlized protein was added to the soak. The Ir'' derivative was pre- bined maps produced with SIGMAA'' until the Rr,.,,, dropped to pared by soaking in mother liquor saturated with cis-dinitro 32%. Refinement continued using crA weighted maps ca lculated bis(ethylene-diamine)lr(III) chloride for 18 h. To minimize non-iso- using the program REFMAC" until the final model was obtained . morphism between native data set 2 and all the derivative data sets, Refinement was initially carried out using REFMAC", and was later native cyt c554 crystals were soaked in protein-free artificial mother continued in X-PLOR" using maximum likelihood simulated liquor for 24 h prior to data collection. annealing. The final model contains 1,908 atoms including four hemes, 66 water molecules, and three phosphates. The crystalloData collection. Native data set 1 was collected at -180 'C on beam graphic R-factor is 20.8% and the R.,,. is 25.4%. The final model line 7-1 at the Stanford Synchrotron Radiation Laboratories using a has r.m.s. deviations for bond lengths and angles of 0.015 A and wavelength of 1.08 A on a 180 mm MAR research image plate 1.89' respectively, with 84 .2% of the residues in the most favored detector. All other data sets were collected on an RAXIS II image regions of the Ramachandran plot, 15.3% in the additionally plate using a Rigaku rotating anode generating Cu Ka radiation . allowed regions, and 0.6% of the residues (Lys 209) in the generCuriously, R,,,. values for data sets collected using Cu Ka radiation ously a1Iowed regions and no residues in the disallowed regions as were particularly high compared to data sets collected using syn- calculated using the program PROCHECK 29 . chrotron radiation (Table 1). Often, R,,m was over 10% even in the lowest resolution bins, where 1/cr was greater than 15. The high val- R.m.s. deviations of heme overlays were calculated using the proues of R..,,,, were not significantly improved by using stringent rejec- gram LSQKAB" , and the two-fold axes drawn in Fig . 3 were calcution criteria when processing data. To assess whether the high lated using the rotation matr ix determined in LSQKAB 24 • values of R..,,,, were a result of photoreduction by the X-rays during Coordinates for HAO", cyt c551.59 • the photosynthetic reaction the course of data collection 21 , single crystal spectra were measured center• and cyt c, 1 were obtained from PDB entries 1FGJ, 1NEW, on a microspectrophotometer. After 20 h exposure to Cu Ka radia- 1PRC and 3CYR respectively. tion at room temperature, initially oxidized crystals exhibited a spectrum with characteristics of reduced protein, indicating pho- Coordinates. Coordinates have been deposited in the Protein toreduction was potentially occurring on all data sets taken using Data Bank (accession number 1BVB) . Cu Ka radiation (data not shown). It is unclear whether the photoreduction is a result of the longer wavelength of Cu Ka radiation, the longer exposure time necessary with t he less intense X-rays, or a Acknowledgments combination of both. Although the single crystal spectra were We thank J. Hu, L Joshua-Tor, M. Stowell. D. Cascio, M. Soltis, H. Schindclin, taken on a crystal exposed to radiation at room temperature. simi- C. KiskeJ" and C. Kielkopf for experimental assistance: C. Drennan and H lar photoreduction phenomena occur at cryogenic temperatures". Axelrod for critical reading and helpful discussions: K. Matthews and B. Crane Because of this consistent discrepancy in symmetry related renec- for the microspectrophotometery: M. Carrondo for the split Sorel cytochrome tions, 1/cr rather than R,,,,, was used as the primary indicator of data coordin ates: and R. Timkovich for the cytochrome c552 coordinates. Thiswork quality, and was therefore used as the primary statistic to determine was supported by an NSF grant to A B Hand an NIH grant to D C.R. IM.I is the high resolution limit, for data sets taken using Cu Ka radiation. supported by an NIH fellowship. The rotation camera facility at SSRL is supported by the NIH and Department of Energy Data processing and phasing. Data were processed using DENZO and Scalepack", and the CCP4 su ite of programs". The difference Received 26 June, 1998: accepted 14 September, 1998. nature structural biology • volume 5 number 11 • november 1998 101 1 155 - - - - - - - - - -~ 1998 Nature America Inc. • http://structbio.nature.com articles E 8 ~ ~C: ci :0 t, 2 ~ C. ~ ,:;: c.i ..5: ~ 1. Andersson. K., Lipscomb, J., Valentine , M., M0nck, E. & Hooper, A. Tetraheme cytochrome c-55'1 from Nitrosomonas europaea: heme -heme interactions and ligand binding. J. Biol. Chem. 261, 1126-1138 (1986). 2. Yamanaka. T. & Shlnra, M. Cytochrome c-552 and cytochrome c-554 derived from Nlrrosomonas europaea: purification. properties, and their function in hydroxylamlnc oxidation. J. Biochem. 75, 1265- 1273 {1974) . 3. Bergmann, D., Arciero, D. & Hooper, A. Organization of the hao gene cluster or Nirrosomonas europaea: genes for two tetraheme c cytochromes. J. Bacterial. 176. 3148-3153 (1994). 4. Arclero. 0 ., Collins. M., Haladjlan, J., Blanco. P. & Hooper. A. Resolutlon of the four hemes of cytochrome c554 from Nlcrosomonas europaea by redox potcntiometry and optical spectroscopy. Biochemistry 30, 11459-11465 (1991) . 5. Arciero. 0 .. Balny, C. & Hooper, A. Spectroscopic and rapid kinelic studies of reduction of cytochrome c554 by hydroxylamine oxidoreductase from Nltrosomonas europaea. Biochemistry 30. 11466-114 72 (1991). 6. Holm, l. & Sander, C. Protein structure comparison by alignment of distance matrices. J. Mo/. Biol. 233. 123-138 (1993). 7. Matias. P., Frazao, C., Morals, J.. Coll. M. & Carrondo, M. Structure analysls of cytochrome cl from Dcsulfovibrio vulgaris Hildcnborough at 1.9 A resolution. J Mo/. Biol. 234, 680-699 (1993). 8. Deisenhofer, J.. Epp, 0 ., Miki. K.. Huber, R. & Michel, H. Structure of t he protein subunits in the photosynthet ic reaction center of Rhodopseudomonas viridis at 3 A resolution . Nature 318, 618-624 (1985). 9. Coutinho, 1.8., Turner, D.l., Liu, M. Y., LeGall, J. & Xavier. A.V. Structure of the three-haem core of cytochrome c~~u determined by 'H NMR. J. Biol. Jnorg. Chem. 1. 305-311 (1996). 10. lgarashl, N., Moriyama, H., Fujiwara, T., Fukumori. Y. & Tanaka, N. The 2.8 A structure of hydroxlyamlne oxidoreductase from a nltrlfying chemoautotrophlc bacterium. Nltrosomonas europea. Nature Strucr. 8/ol. 4, 276- 284 (1997) . 11 . Matias. P. et al. A preliminary analysts of the three-dlmenslonal structure of dlmerlc di-haem spUt-Soret cytochrome c from Desulfovlbrlo desutfurlcans ATCC 27774 at 2.5 A resolutton using the MAO phasing method: a novel cytochrome fold with a stacked -haem arrangement. J. Biol. lnorg. Chem. 2, 507-514 (1997). 12. Oegtyarenko, K.N., North, A.C.T.. Perkins. D.N. & Findlay, J.B.C. PROMISE: a database of Information on prosthetic centres and meta l ions In protein active sites. Nucleic Acids Res. 28. 376-381 (1998). 13. Moore, G. & Pettigrew, G. Cytochromes c- evolutJonary. structural and physiochemical aspects (Springer-Verlag, Berlin; 1990). 14. Loew. G.H. In Iron porphyrins (eds Lever. A.P.B. & Gray, H.B.) (Addison-Wesley Publishi ng Co.. Reading: 1983). l 5. Ludwig , M. et al. Control of oxidation-reduction potentials in flavodoxln from Clostridium bcijcrinckit. the role of conformatiomil changes. Biochemistry 36, 1259-1280 (1997) . 16. Rees, D.C. & Farrelly, D. Biological electron transfer. The enzymes 19, 37- 97 (1990). 17. Stephens, P.J., Jolli e, D.R. & Warshel, A. Protein control of redox potentlal of iron-sulfur proteins. Chem. Rev. 96 , 2491-2513 (1996). 18. Hooper, A .. Maxwell, P. & Terry, K. Hydroxylamine oxidoreductase from Nltrosomonas: absorption spectra and content of heme and metal. Biochemistry 17. 2984-2989 (1978). 19. Petsko, G. Preparation of isomorphous heavy-atom derivatives. Meth. Enz. 114, 147-157 (1985). 20. Holden, H.M. & Rayment, I. Trimethyl lead acetate-a first choice heavy atom derivative for protein crystallography. Arch. Blochcm. 291 , 187-194 (1991 ). 21 . Nave, C. Radiation damage In protein crystallography. Radfat. Phys. Chem. 45, 483-490 (1995). 22. Gonzales, A. & Nave, C. Radiation damage in protein crysUlls at low temperature . Acta CrysUJ/lgr. D 50. 874-877 (1994}. 23. Otwinowski, Z. In CCP4 study weekend data collection and processing (eds Sawyer. L. . Isaacs. N. & Balley. S.) 56-62 (SERC Daresbury Laboratory. UK: 1993). 24 . Bailey, S. The CCP4 suite-programs for protein crystallography. Acta Cry,tsl/ogr. D 50, 760-763 (1994). 25. Sheldrick, G.M. Phase annealing in SHELX-90-dlrect methods for larger structures. Acta Crystallogr. A 46. 467-473 (1990). 26. De La Fortellc, E. & Brlcognc , G. Maximum likelihood heavy atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Meth. Enz. 276, 472-494 (1997). 27. Jones, T.A.. Zou, J.Y.. Cowan, S.W. & Kjeldgaard, M. Improved methods tor building protein models In electron density maps and the location of errors In these models. Acta Crystal/ogr. A 47 , 110-119 (1991). 28. BrO nger. A.T. X-PLOR version 3.1-A system for x-ray crystallography and NMR (Yale University Press, New Haven and London; 1992). 29. Laskowski, R.A., Mcarthur, M.W., Moss. D.S. & Thornton, J.M. PROCHECK • a program to check the stereochemlcal quality of protein structures. J. Appl. Cry,tallogr. 26, 283-291 (1993). 30. Pri nce . R.C. & George, G.N. The remarkable complexity of hydroxylamlne oxldoreductase. Nature Struct. Bio/. 4. 247- 250 (1997). 31 . McTavish, H., Arctero, D. & Hooper. A. Interaction with membranes of cytochrome c554 from Nitrruomonas europaea. Arch. Biochem. Biophys. 324, 53-58 (1995). 32. Kraulls, P.J. MOLSCRIPT-a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystal/gr. 24, 946-950 {1991) . 33. Esnouf, R. BOBSCRIPT-An extensively modiried version of Molscript that includes greally enhanced coloring capabilities. J. Mol. Graph. 15, 133-138 (1997). 34. Merritt, E.A. & Murphy, M.E.P. Raster3D Version 2.0-a program for photoreallstlc molecular graphics. Acea Crystellogr. 050, 869-873 (1994) . 35. Sharp, K.. Fine, R. & Honig, 8. Computer simulations of lhe diffusion of a substrate to an active site of an enzyme. Science 236, 1460- 1463 (1987). 36. Allen, F. & Kennard, O. 3D search and research using the Cambridge Data Base. Chemical Design and Automation News 8, 1&31 - 37 (1993). "ffi ~ ~ ::, 1ii z a:, en en 1012 nature structural biology• volume 5 number 9 • september 1998 156 Chapter 7 Iverson, T.M., Arciero, D.M., Hooper, A.B., Rees , D.C . (2000) High-resolution Structures of the Oxidized and Reduced States of Cytochrome c554 from Nitrosomonas europaea. J. Biol. Inorg. Chem. in preparation. 157 High-resolution Structures of the Oxidized and Reduced States of Cytochrome c554 from Nitrosomonas europaea Tina M. Iverson, David M. Arciero, Alan B. Hooper and Douglas C. Rees T. M. Iverson Graduate Option in Biochemistry, MC 147-75 CH, California Institute of Technology, Pasadena, CA 91125 D. M. Arciero • A. B. Hooper Department of Biochemistry, University of Minnesota, St. Paul, MN 55108 D. C. Rees (!8J) Division of Chemistry and Chemical Engineering and Howard Hughes Medical Institute, MC 147-75 CH, California Institute of Technology, Pasadena, CA 91125 e-mail: dcrees@caltech.edu Tel: (626) 395-8393 Fax: (626) 744-9524 158 Abstract Cytochrome c554 (cyt c554) is a tetra-heme cytochrome involved in the oxidation of NH 3 by Nitrosomonas europaea. The x-ray crystal structures of cyt c554 in the oxidized and dithionite reduced forms have been solved by molecular replacement at 1.6 A and 1.8 A resolution respectively. Upon reduction, a conformational change occurs in the conformation of the polypeptide chain between residues 175 and 179, adjacent to hemes III and IV. Cyt c554 displays conserved heme-packing motifs that are continuing to emerge in other heme-containing proteins. By comparison to hydroxylamine oxidoreductase, the redox partner, and cytochrome c nitrite reductase, an enzyme involved in nitrite ammonification, substantial structural similarity has been observed in the heme core environment. At the resolution of these structures, the structural determinants of these heme packing are shown to extend to the buried water molecules that orient the histidine ligands to the heme iron. In the original structure determination of a tetragonal crystal form, a cis-peptide bond between His 129 and Phe 130 was identified that appeared to be stabilized by crystal contacts. In the rhombohedral crystal form used in the present high-resolution structure determination, this peptide bond is clearly trans-, but with disallowed angles of <1> and 'I'. 159 Key Words cytochrome· electron transfer · heme · nitrification • Nitrosomonas europaea • x-ray crystallography · conformational change· oxidation • reduction Abbreviations cyt c554 cytochrome c554 • HAO hydroxylamine oxidoreductase · AMO ammonia monooxygenase • cyt c552 cytochrome c552 • cyt c NR cytochrome c nitrite reductase 160 Introduction Cytochrome c554 (cyt c554) is a tetra-heme [1] cytochrome involved in ammonia oxidation by the chemoautotrophic soil bacterium Nitrosomonas europaea [2, 3]. This transformation of ammonia represents a key component of the process of nitrification in the biological nitrogen cycle. Oxidation of NH 3 serves as the sole energy source for N. europaea [4] and the physiological process occurs in two enzymatic steps followed by the energy yielding electron transfer steps. The first enzymatic step, catalyzed by the integral-membrane protein ammonia monooxygenase (AMO), converts NH 3 to hydroxylamine (NH20H) and water and requires the input of two electrons as well as molecular dioxygen. Following this, hydroxylamine oxidoreductase (HAO) converts NH 2 0H to No 2- at a catalytic phlorin. This latter reaction releases four electrons that are passed pairwise to two molecules of cyt c554. Although there is controversy over the exact nature of the electron transport proteins following cyt c554 in vivo, the pathway can be reconstructed in vitro with cytochrome c552 (cyt c552) [2]. Crystallographic studies of cyt c554 [10] revealed a primarily a-helical secondary structure. The four covalently attached hemes are stacked into two types of pairs [10]. The first type of heme pair stacks with the planes of each pair of porphyrin rings nearly parallel and overlapping at the edge, and occurs in two variants. In the first variant, two 161 hemes exhibiting bis-histidine coordinate stack with the heme edges parallel to a pseudo two-fold axis. This type of heme stacking has been observed in the split-Soret cytochrome [11], HAO [12], cytochrome c nitrite reductase (cyt c NR) [13] and cytochrome c fumarate reductase (cyt c FR) [ 14-16]. In the second variant of parallel stacked hemes exhibited by cyt c554, one of the two stacked hemes has bis-histidine iron coordination, while the other is a five-coordinate iron. The planes of the hemes remain parallel, but the edges now lie at a 30° angle to the pseudo two-fold axis. This hemestacking motif has been observed in HAO [12], cytochrome c nitrite reductase (cyt c NR) [13], where both of these proteins have catalytic activity associated with the fivecoordinate heme. The second type of heme pair is arranged with the planes of the porphyrins nearly perpendicular to one another as has previously been observed in cytochrome c3 [ 17] and proteins with c3-like repeating units [18, 19] as well as in HAO [12] and cyt c NR [13]. All of the proteins exhibiting these heme-stacking are involved in the metabolic pathways that require multiple electron transfer steps, suggesting that the stacked heme motifs may aid the rapid transfer of multiple electrons [10] . In view of the function of cyt c554 as a multi-electron transfer protein displaying a conserved heme arrangement, we have determined the structure in the fully oxidized 162 and dithionite reduced states to characterize oxidation-state dependent conformational changes. These studies serve to identify regions of cyt c554 that may be functionally important for the oxidation-reduction processes required for biological nitrification, and allow a better structural understanding of the heme packing motifs. Materials and Methods Crystallization and Data Collection Cyt c554 was produced, purified [9] and crystallized [10] as previously described. Instead of the tetragonal crystals produced from early preparations, one protein preparation yielded rhombohedral crystals. These crystals were of superior quality to the tetragonal crystals with a diffraction limit up to 1.6 A resolution. Fully oxidized crystals, grown aerobically, were reduced by adding an excess of Na 2S2O4 to the mother liquor containing the crystals. The crystals were soaked in the dithionite solution for several hours prior to cryo cooling and exhibited a color change during this time. Data sets were collected at -180° C on Stanford Synchrotron Radiation Laboratories beam line 7-1 using a wavelength of 1.08 Aand a 180 mm MAR Research image plate detector. 163 Data Processing and Structure Solution Data were processed and scaled using DENZO, SCALEPACK [20], and the CCP4 suite of programs [21]. The oxidized structure was solved by molecular replacement with AMORE [22] using the structure of cyt c554 from the tetragonal crystal form (lBVB) as the search model. Since the dithionite reduced crystals were isomorphous with the oxidized crystals, the same molecular replacement solution was subjected to additional rigid body refinement against the reduced data. To ensure accurate comparison of the statistics, the same set oftest reflections were selected for the Rfree from each data set. Model Building and Refinement The program O [23] was used for model building of the cyt c554 structure. Refinement was carried out using REFMAC [21 , 24], XPLOR [25] and SHELX [26] . Maps for model building were calculated with the CCP4 suite of programs [21] and were weighted using cra-coefficients determined in REFMAC [21, 24]. As a result of anisotropic density associated with the heme irons, the histidine ligation distances to the irons were difficult to determine using standard temperature factors when calculating maps. Thus, anisotropic temperature factors for the heme irons were refined in SHELX 164 [26], and a-a-weighted maps from this calculation were used to determine the iron ligation distances. The resultant ligation distances were constrained in X-PLOR [25], which refined the overall protein model to significantly better geometry. Data collection and refinement statistics are summarized in Table 7.1. Results and Discussion Overall Fold The crystal structures of oxidized and dithionite reduced cyt c554 have been refined to 1.6 A and 1.8 A resolution respectively. When cyt c554 from the original, tetragonal space group was superimposed onto oxidized cyt c554 from the rhombohedral space group described here, the rms~ was 0.6 Afor the Ca atoms. The structural changes that occur between the two crystal forms are located in areas of the protein either involved in crystal contacts (Fig. 7.la - residues 142 to 153) or that have higher temperature factors and are therefore expected to be more mobile (Fig. 7 .1 b ). 165 Changes in polypeptide fold upon reduction Alignment of oxidized and dithionite reduced cyt c554 in the rhombohedral space group resulted in an rmsl\ of 0.4 A overall, thus upon dithionite reduction, the majority of the protein remains in the same conformation. However, one interesting conformational change occurs between residues 175 and 179 (Fig. 7. le). This loop lies above Hemes III and IV with His 179 serving as a ligand to Heme III. The conformational change displaces the backbone over 2 A and may be triggered by a change in the peptide bond orientation between Met 178 and His 179 (Fig. 7.2). In the oxidized form of cyt c554, the peptide carbonyl is in van der Waals contact (4.7 A) with the porphyrin of Heme III. Upon reduction, reorientation of the peptide bond places the backbone carbonyl 8.3 A away from heme III, with an ordered water molecule replacing the peptide carbonyl. The relative rigidity of the region between residues 17 5 and 179 is reflected in the low temperature factors for this loop (all < 20 A2) in both the oxidized and reduced protein (Fig. 7 .1b). This suggests that the observed structural changes are significant and not associated with inherent flexibility of cyt c554 in that region. The conformational change was observed in three different crystals of dithionite reduced cyt c554 (data not shown). It is possible that this conformational change is reflected in the biphasic kinetic of dithionite reduction [5]. 166 Spectroscopic studies [1, 7] have established that three of the four hemes in cyt c554 are low-spin, while the fourth is high-spin and thus likely 5-coordinate [8]. The high-spin heme has a reduction potential of +47 mV , while the low-spin hemes titrate with midpoint potentials of +47 mV, -147 mV, and -276 mV [9]. While the five-coordinate heme II likely represents the +47 mV high-spin heme, the remainder of the reduction potentials have not yet been assigned. As the backbone conformational change is probably associated with the reduction of heme III, which has an unknown reduction potential, it is difficult to determine the physiological relevance of this movement. A second conformational change associated with the reduction of cyt c554 is a change in the position of the propionate group of heme IV. The determinants of this conformational change are unclear. Other side chains that change position between the structures of the two oxidation states lie on the surface of the protein, and it is unclear if the change in side chain position reflects actual oxidation-state dependent changes or inherent mobility. 167 Role of ordered water molecules in the stacked-heme motif Alignment of the hemes common to cyt c554, HAO [12] and cyt c NR [13] shows that in addition to the conserved heme arrangement, there is similarity in the conformation of the surrounding polypeptide chain. As cyt c NR and cyt c554 have been solved at resolutions where ordered water molecules can be unambiguously assigned, the role of conserved, buried water molecules involved in heme-stacking can be investigated. Cyt c554 and cyt c NR can be aligned such that hemes II through IV of cyt c554 are superimposed onto hemes 515 through 517 of cyt ·c NR. In this alignment polypeptide chains surrounding the hemes overlay. However, disappointingly few water molecules are in conserved locations. The water molecules that do appear in conserved locations in the two structures form hydrogen bond contacts to the histidine ligands of the two perpendicular stacked (central) hemes, stabilizing the orientation of the plane of the imidazole ring such that they are similar in the two structures (Fig. 7.3). As modulation of the reduction potential [27] and the electronic configuration of hemes [28] may be affected by the orientation of the imidazole ring of the histidine ligand, the structural conservation of the ligand side chain may be involved may reflect a necessary component for heme-packing motifs. 168 Although HAO has not been solved at a resolution where ordered water molecules can be distinguished, the planes of the imidazole rings of the histidine ligands are similar to those from cyt c554 and cyt c NR. Surprisingly, the distal ligand to heme I of cyt c554 exhibits a similar orientation to heme 3 of HAO despite that histidine 102 coordinates the heme iron with the N8 atom, while all histidine ligands to heme irons in HAO coordinate with the Ne atom. The similarities between the polypeptide chain surrounding the hemes as well as the extent of the heme packing motifs themselves ·suggest an evolutionary relationship exists between cyt c554, HAO, and cyt c NR. Cyt c NR is involved in nitrite ammonifiction, and catalyzes essentially the same reaction as occurs during nitrification by N. europaea, but in the physiologically reverse direction. In both cyt c NR and HAO, the reaction is catalyzed at a five-coordinate heme, however in cyt c NR, the protein ligand is a lysine, rather than a histidine. The similarities between the heme packing of cyt c NR and the two proteins involved in nitrification thus seem to be an example of convergent evolution. The evolutionary relationship between HAO and cyt c554 is less clear, despite the apparent lack of significant sequence identity, it seems remarkable that this motif would have evolved separately in the two redox partners. 169 Heme Iron Ligation At the resolution of these structures, the heme iron ligation distances can be compared and are listed in Table 7 .2. As expected [8] , the six-coordinate low spin heme irons found in hemes I, III and IV have shorter histidine ligation distances than the fivecoordinate high-spin iron found in heme II. Additionally, the iron of heme II lies~ 0.3 A out of the plane of the heme resulting in longer average Fe-NP (pyrrole nitrogen) ligation distances than the six-coordinate low-spin counterparts. Reduced heme iron (Fe 2+) would have a slightly larger radius than oxidized heme iron (Fe3+) and thus is expected to have longer coordination distances to the histidine ligands. In these crystal structures, the coordination distances appear to obey a general trend of lengthening upon reduction, however, the small change in coordination distance between oxidized and reduced heme iron (Table 7.2) is smaller than the estimated coordinate error of the crystal structures presented here (Table 7 .1 ). His 129 - Phe 130 Peptide Bond The overall geometry and refinement for these structures of cyt c554 is reasonable (Table 7 .1) and only a few Ramachandran outliers are seen. In the oxidized crystal, Ala 170 207 and Lys 209 are in the generously allowed and disallowed regions of the Ramachandran plot respectively. These two residues lie at the C-terminus of the protein and are relatively disordered, thus the poor geometry is a reflection of the poor electron density in this region. Because of this disorder, the C-terrninus has been omitted from the model of the reduced crystal, as there was no clear density beyond residue 208 . In contrast, in both the oxidized and reduced crystals, the peptide bond of Phe 130 is in clear density and has <I> and o/ angles in the disallowed region of the Ramachandran diagram (Fig. 7.4a). In the tetragonal crystal form of cyt c554, the bond between His 129 and Phe 130 appeared to be in a cis- conformation (Fig. 7.4b) [10]. This cis-peptide bond was stabilized by a crystal contact involving the side chains of His 129 and Asp 131 around the crystallographic 2-fold axis, whereas the rhombohedral crystal form has no crystal contact in this area. One possible explanation for the behavior of this peptide bond is that the cispeptide bond is a crystal packing artifact, and that the peptide bond in that region is normally trans- but has disallowed <I> and o/ angles. A more intriguing possibility is that this peptide bond is functionally significant. By having a high-energy disallowed transpeptide bond, it would facilitate isomerization to cis-. This cis- form may have been 171 fortuitously stabilized in the tetragonal crystal form by crystal contacts. Crystal contacts have previously been implicated in the stabilization of functionally relevant cis- peptide bonds [29] . The location of Phe 130 is in a pocket above Hemes I and III near the proposed binding site for HAO. Thus, the His 129 - Phe 130 peptide bond may have implications for electron transfer between cyt c554 and HAO, or could be involved in complex formation . Conclusions This study shows a high-resolution view of cyt c554, a member of the biological nitrogen cycle, in two oxidation states of the protein. A conformational change accompanies four-electron dithionite reduction. The structural determinants of hemepacking motifs extend to the orientation of the imidazole rings of the histidine ligands to the heme irons. This involves the structural conservation of buried water molecules that stabilize the imidazole orientation, however the remainder of the water molecules are not conserved between these proteins. Disallowed Ramachandran angles at Phe 130 suggest functional significance of this residue, which has yet to be investigated biochemically. Further crystallographic studies will attempt to trap a two electron reduced state of cyt c554. Additionally, it would be of great benefit to determine the structure of the HAOcyt c554 complex. A structure of this complex would illuminate specific inter-molecular interactions between the two proteins. 172 Acknowledgements We thank R. Dieckmann, J. Chiu, and F. A. Tezcan for experimental assistance and J.A. Shelnutt for calculations of the distortions of the porphyrins. This work was supported by an NSF grant(#) to ABH and an NIH grant (GM45162) to DCR. TMI is supported by an NIH fellowship (GM07737) . This work is based upon research conducted at the Stanford Synchrotron Radiation Laboratory (SSRL), which is funded by the Department of Energy, Office of Basic Energy Sciences. 173 Fig. 7.1 (next page) : a. RMS deviations of oxidized, rhombohedral cyt c554 and cyt c554 from the tetragonal crystal form. The structure of cyt c554 in the tetragonal crystal form is colored according to the differences between the two crystal forms, with large differences in red, negligible differences in blue, and intermediate differences scaled in orange, yellow, and green. The arrow indicates a loop between residues 152 and 153 that shifts over 1 A between the two crystal forms that is involved in a crystal contact in the tetragonal crystal form . b. Variation in temperature factors of cyt c554. The model of oxidized cyt c554 in the rhombohedral crystals form is colored according to variation of temperature factors. Red indicates the most mobile regions of the model and blue indicates the most stable regions, with intermediate mobility scaled in orange, yellow, and green. Reduced cyt c554 has a similar distribution of temperature factors (data not shown). c. Differences in oxidized and reduced rhombohedral cyt c554. The rms!l of the two models have been superpositioned onto the structure of reduced cyt c554 in the rhombohedral crystal form. The largest shifts in Ca coordinates (red) occur between residues 175 and 179 and maximally displace the loop 2.3 A. Careful inspection of this loop in Fig. 1b and Fig. le shows the nature of the conformational change with respect to the rest of the protein. All molecular representations were made using MOLSCRIPT [30], BOBSCRIPT [31] and RASTER3D [32]. 174 A B > 1.05A o.8 A to 1.05 A o.55 A to o.8 A o.3 A to o.55 A o.o5 A to o.3 A 28 A2 to 33 A2 23 A2 to 28 A2 18 A2 to 23 A2 13 A2 to 18 A2 > 1.5 A C 1.175 Ato 1.5 A o.8 Ato 1.175 A 0.425 Ato o.8 A o.o5 Ato 0.425 A 175 Fig. 7.2: Stereoview of the conformational change in the loop between 175 and 179. Reduced cyt c5.54 is shown in red, while oxidized cyt c5.54 is shown in blue. The side chains of His 179 and Met 178 as well as the peptide bond between those two residues are shown. This peptide bond changes conformation upon reduction. Fig. 7.3: Alignment of hemes and surrounding secondary structure of cyt c554 and cyt c NR. Water molecules located in conserved positions are shown. Cyt c5.54 is colored red and cyt c NR is colored green. 176 Fig. 7.4: Peptide bond at Phe 130. a. The His 129 - Phe 130 peptide bond is shown superpositioned onto aa weighted 21F0 I - IFcl density calculated after the omission of both His 129 and Phe 130. b. The His 129-Phe 130 peptide bond in the tetragonal crystal form shows a cis-conformation, and is influenced by crystal contacts around the crystallographic 2-fold axis of symmetry. His 129 Glu 131 • ~ . •"' '\,.; . ......--· w • Glu131 8 e:1 • His129 '- ~.S 177 Table 7.1: Data Collection and Refinement Statistics Unit Cell Constants a=b (A) C (A) Data Resolution (A) Observations Unique Observations Completeness (%) Rsym2 Ila Oxidized Reduced 147.89 33.91 147.24 33.88 1.6 109,601 35,562 97.3 (86.1)1 0.045 (0.174) 28.0 (7.3) 1.8 112,676 24,588 95.5 (85.9) 0.051 (0.288) 28.5 (5.1) Refinement 0.188 Rcrys/ 4 0.215 Rfree 5 Estimated Coordinate Error o.o9A No. of free-R reflections 1808 RMS~ bond lengths 0.021 A RMS~ bond angles 1.99° Number of atoms Residues Water molecules 1988 1-211 Other HET atoms Ramachandran statistics % Most favored % Allowed % Generously allowed % Disallowed 0.197 0.236 0.13 A 1242 0.016 A 1.60 142 3 ips* 1954 1-208 142 1 ips*, 1 dtn*, 1 SO/ 87.0 11.3 0.6 (Ala 207) 1.1 (Phe 130, Lys 209) 87.4 12.0 0.0 0.6 (Phe 130) 1 Numbers in parentheses indicate values for the highest resolution bin. 2Rsym=I:[1 - <I> ]IL[ <I>] where i is the ith measurement and <I> is the weighted mean of 1 I. Rcryst = r11Fob5' - IFcalcll/LIFobsl Rrree is the same as Rcryst for a selection of randomly omitted reflections . 5 Estimated Coordinate Error is the Cruickshanks value determined in REFMAC. *ips is inorganic phosphate (PO/-) and dtn is dithionite (S 2O/-). SO 3 is the oxidative breakdown product of dithionite. 3 4 178 Table 2: Heme Iron Ligation distances Oxidized proximal Oxidized distal Oxidized NP Heme! 1,95 A 2.osA 1.98 A Heme II 2.17 A n/a 2.03 A Reduced proximal Reduced distal Reduced NP 2.ooA 2.01 A 1.98 A 2.20A n/a 2.04A 0 0 Heme III 2.03A 2.01 A 1,99 A 0 2.04A 2.10A 1.98 A 0 HemeIV 1.97 A 2.03 A 1.97 A 2.06A 2.06A 1.98 A Proximal refers to the histidine involved in the -C-x-y-C-H- heme binding motif, while distal refers to the histidine found elsewhere in the sequence. Heme II is 5-coordinate and therefore has no distal ligand. NP refers to pyrrole nitrogen atoms and is an average of the four NP-Fe bond lengths. 179 References 1. Andersson, K, Lipscomb, J, Valentine , M, Mi.inck, E and Hooper, A (1986) J Biol Chem 261 : 1126-1138 2. Yamanaka, T and Shima, M (1974) J Biochem 75 : 1265-1273 3. Suzuki, I and S-C., K (1981) Can J Biochem 59: 484-488 4. Wood, PM (1986) In: Prosser, JI (ed) Nitrification. IRL Press, Oxford, pp 39-62 5. DiSpirito, AA, Balny, C and Hooper, AB (1987) Eur J Biochem 162: 299-304 6. Tsang, DCY and Suzuki, I (1982) Can J Biochem 60: 1018-1024 7. Petersson, Land Andersson, KK (1987) Biochim Biophys Acta 915: 261-266 8. Scheidt, WR and Gouterman, M (1983) In: Lever, ABP and Gray, HB (ed) Iron Porphyrins Part One. Addison - Wesley Publishing Co, Inc., London, pp 89-139 9. Arciero, D, Collins, M, Haladjian, J, Bianco, P and Hooper, A (1991) Biochemistry 30: 11459-11465 10. Iverson, TM, Arciero, DM, Hsu, BT, Logan, MSP, Hooper, AB and Rees, DC (1998) Nat Struct Biol 5: 1005-1012 11. Matias, P, Morias, J, Coelho, A, Meijers, R, Gonzalez, A, Thompson, A, Sieker, L, LeGall, J and Carrondo, M (1997) JBIC 2: 507-514 12. Igarashi, N, Moriyama, H, Fujiwara, T, Fukumori, Y and Tanaka, N (1997) Nat Struct Biol 4: 276-284 180 13 . Einsle, 0, Messerschmidt, A, Stach, P, Bourenkov, GP, Bartunik, HD, Huber, Rand Kroneck, PMH (1999) Nature 400: 476-480 14. Bamford, V, Dobbin, PS, Richardson, DJ and Hemmings, AM (1999) Nat Struct Biol 6: 1104-1109 15. Taylor, P, Pealing, SL, Reid, GA, Chapman, SK and Walkinshaw, MD (1999) Nat Struct Biol 6: 1108-1112 16. Leys, D, Tsapin, AS, Nealson, KH, Meyer, TE, Cusanovich, MA and VanBeeumen, JJ (1999) Nat Struct Biol 6: 1113-1117 17. Matias, P, Frazao, C, Morais, J, Coll, Mand Carrondo, M (1993) J Mol Biol 234: 680-699 18. Matias, PM, Saraiva, LM, Soares, CM, Coelho, AV, LeGall, J and Carrondo, MA (1999) JBIC 4: 478-494 19. Coutinho, IB , Turner, DL, Liu, MY, LeGall, J and Xavier, AV (1996) J Biol Inorg Chem 1: 305-311 20. Otwinowski, Z (1993) In: Sawyer, L, Isaacs, N and Bailey, S (ed) CCP4 Study Weekend Data Collection and Processing. SERC Daresbury Laboratory, UK, pp56-62 21. Bailey, S (1994) Acta Cryst D50: 760-763 22. Navaza, J (1994) Acta Cryst A50: 157-163 23. Jones, TA, Zou, JY, Cowan, SW and Kjeldgaard, M (1991) Acta Cryst A47 : 110-119 24. Murshudov, GN, Yagin, AA and Dodson, EJ (1997) Acta Crystallogr D53: 240-255 181 25. Bri.inger, AT (1992) X-PLOR version 3.1 - A system for x-ray crystallography and NMR. Yale University Press, New Haven and London 26. Sheldrick, GM (1990) Acta Cryst A46: 467-473 27. Rousseau, DG and Rousseau, DL (1992) J Struct Biol 109: 13-17 28. Shokhirev, NV and Walker, FA (1998) J Biol Inorg Chem 3: 581-594 29. Ludwig, M, Pattridge, K, Metzger, A, Dixon, M, Eren, M, Feng, Y and Swenson, R (1997) Biochemistry 36: 1259-1280 30. Kraulis, PJ (1991) J Appl Cryst 24: 946-950 31. Esnouf, RM (1997) J Mol Graph 15: 133-138 32. Merritt, EA and Murphy, MEP (1994) Acta Cryst D50: 869-873 182 Chapter 8 Tina M. Iverson, Michael P. Hendrich, David M. Arciero, Alan B. Hooper, Douglas C. Rees (2002) In: The Metalloprotein handbook, T. Poulos, ed., in preparation. 183 Cytochrome c554 Tina M. Iverson 1, Michael P. Hendrich2 , David M. Arciero 3 , Alan B. Hooper3 , Douglas C. Rees 4 1. Graduate Option in Biochemistry, 147-75 CH, California Institute of Technology, Pasadena, CA 91125. 2. Department of Chemistry, Carnegie Mellon University, 4400 5th Ave., 268-1061, Pittsburgh, PA 15213-2683 . 3. Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, St. Paul, MN 55108. E-mail hooper@umn.edu. 4. Division of Chemistry and Chemical Engineering and Howard Hughes Medical Institute, MC 147-75 CH, California Institute of Technology, Pasadena, CA 91125 . Email: dcrees@caltech.edu. Tel: (626) 395-8393. Fax: (626) 744-9524. Supported by NSF Research to ABH # MCB-9316906. 184 Cytochrome c554 (cyt c554) is a tetra-heme cytochrome from the chemolithotrophic bacterium, Nitrosomonas europaea, which derives energy for growth from oxidation of ammonia to nitrite with hydroxylamine as the free intermediate. Cyt c554 participates in this pathway as an electron acceptor of the hydroxylamine-oxidizing enzyme. The hemes of cyt c554 display conserved packing motifs involving stacked pairs with the planes of the porphyrin rings approximately parallel and stacked at a planeto-plane distance of 3.6 A to 3.8 A. Currently, the stacked-heme motifs have been seen exclusively in other bacterial proteins involved in the metabolism of sulfur or nitrogen containing compounds, and may represent a structural arrangement necessary for the transfer of multiple electrons associated with these types of metabolic pathways. (3D structure near here) INVOLVEMENT IN THE BIOLOGICAL NITROGEN CYCLE (Figure 8.1 near here) The chemoautotrophic soil bacterium Nitrosomonas europaea generates all of its energy from the oxidation of NH 3 to No2-. 1 As a result, N. europaea plays a key role in the biological nitrogen cycle (Figure 8.1). In addition to maintaining balanced ecosystems 2 , applications of the nitrification (the oxidation of NH3 to No2-) by the organism include the removal of ammonia in sewage treatment plants 3 and rivers polluted by NH3-containing compounds4, e.g., fertilizer runoff or ammonia- 185 based industrial waste. Molecular oxygen is required for the oxidation of NH3 to NH 20H (Figure 8.2), consequently N. europaea is an obligate aerobe. (Figure 8.2 near here) THE BIOCHEMISTRY OF NITRIFICATION The pathway by which N. europaea oxidizes NH 3 (Figure 8.2) involves two enzymatic steps. Ammonia monooxygenase (AMO) , an integral-membrane protein 5, 6 , first oxidizes ammonia to hydroxylamine (NH 20H)7. This oxidation requires the input of two electrons and molecular dioxygen. NH 20H is further oxidized, to No 2- by hydroxylamine oxidoreductase (HA0) 1 which removes four electrons and inserts of an atom of oxygen from water. Thus, two electrons per NH 3 oxidized enter the electron transport chain leading to a cytochrome oxidase 10 or to reduce nitrite or nitric oxide when the oxygen concentration is low 11 . The physiological acceptor of electrons from HAO is cyt c554 (Figure 8.2), which contains four covalently attached heme moieties 8 having a range of midpoint potentials, but apparently accepts only two electrons in the two lower potential hemes per turnover9 . Cyt c554 cannot pass electrons directly to the terminal electron acceptors, indicating the existence of at least one intervening electron transfer protein. The identity of the intervening electron carrier(s) in vivo is not known 12- 14 . The soluble, monoheme cytochrome c552 (cyt c552) can be rapidly reduced in the presence of hydroxylamine and catalytic amounts of HAO and cyt c554 in vitro , but the role for cyt c552 as an electron donor to terminal electron acceptors has not been established 15 . In 186 keeping with what appears to be a key role in the central energy yielding redox interactions of this organism, the approximate cellular content of AMO, HAO, cyt c554, cytochrome aa3 fall within a five fold range and cyt c552concentrations are significantly higher 11 . Additionally, the intermediate(s) in the electron donor pathway to AMO that begin the biological nitrification process remains unclear 16 • 17 . Although cyt c554 normally localizes to the periplasm 11 as does HAO, exogenous electrostatic association of cyt c554 with the membrane has been observed 18 • SEQUENCE INFORMATION The genome of N. europaea contains three gene copies of cycA (also called hey), the gene encoding cyt c554 19, 20 each of which contain four -C-x-y-C-H heme-binding motifs. The gene copies themselves differ by only a few bases and the amino acid sequence predicted by the genes that have been sequenced to date are identical1 9, 20• 21 . The pathway utilized by N. europaea to oxidize NH 3 represents a highly specialized process of energy harvesting, and currently no other proteins, including other tetra-heme cytochromes, exhibit significant sequence similarity to cyt c554. Thus, cyt c554 has long been considered to represent its own class2 1. 187 The gene amoA, coding for one of the subunits of AMO has been sequenced in a number of ammonia-oxidizing autotrophic bacteria in the p or y subdivision of proteobacteria. Genes of AMO from three different members of the p subdivision show a minimum of 83% identity. Additionally, amoA of this subdivision exhibits significant sequence similarity ( ~44%) to the genes encoding the particulate methane monooxygenase (pMMO) of the methanotrophs. The gene amoA of the nitrifier of the y subdivision shows 51 % identity to the corresponding members of the p subdivision and 83% identity to the pMMO genes. Genes for HAO and cyt c554 have been detected by gene hybridization in Paracoccus denitrificans 23 and Nitrosospira spp 22 , another autotrophic ammonia oxidizing bacteria of the p subdivision, suggesting these organisms may use a similar pathway for NH 3 oxidation as Nitrosomonas . As more sequence information becomes available, it seems likely that other autotrophic ammonia oxidizers of the p and y subdivisions will be shown to contain homologs of both cyt c554 and HAO. PROTEIN PRODUCTION AND PURIFICATION The first purification protocol 15 for cyt c554 employed the extreme solubility and stability of the protein. When a cell extract of N. europaea is brought to 80% saturation with (NH 4 ) 2S04 , only two proteins remain soluble: cyt c554 and HAO. As the saturation 188 increases from 80% to 90%, HAO precipitates, leaving cyt c554 in solution and relatively pure. To increase purity, additional chromatography can be performed24 . This series involves an octyl-Sepharose column followed by dialysis to remove the (NH 4 ) 2SO4 followed by an Amberlite CG-50 column and gel filtration using Sephadex G-100. THREE-DIMENSIONAL STRUCTURE DETERMINATION Crystallization Cyt c554 has been crystallized in three distinct crystal forms. High concentrations of phosphate at pH 10.1 resulted in the growth of two of these crystal forms : one in the tetragonal space group P432225 , and the other in the rhombohedral space group R3 26 . The preliminary structure was determined by the method of multiple isomorphous replacement and based on the tetragonal crystal form at 2.6 A resolution 25 . The rhombohedral crystal form has subsequently been solved by molecular replacement and has been refined to 1.6 A and 1.8A resolution in the oxidized and dithionite reduced states, respectively 26 . A third distinct crystal form diffracting to 2.3 A resolution grew from ammonium sulfate precipitant in the monoclinic space group P2i27 . A structural analysis of this crystal form has not yet been reported. 189 Description of the polypeptide fold Cyt c554 is oblong in shape with dimensions 60 x 30 x 30 A. The polypeptide chain primarily consists of helices, most of which adopt a-helical conformations, but additionally the fold contains two stretches of 3 10 helix, and two short, parallel ~-strands (see 8.3D representation). Consistent with the specialized nature of cyt c554, the fold of the polypeptide chain initially appeared to be unlike any previously determined folds, however, remarkable similarities exist between the heme arrangement of cyt c554, HAO, and cytochrome c NR (cyt c NR) . These similarities extend to the polypeptide chain surrounding the hemes, despite the lack of any detectable sequence identity between these proteins. Heme iron coordination and heme configuration The polypeptide chain of cyt c554 compactly binds the four hemes such that the porphyrins are positioned within van der Waals contact of one another and the interheme iron distances are relatively short (Figure 8.3). (Figure 8.3 near here) Each of the four heme moieties of cyt c554 has the heme iron coordinated by histidine side chains. The heme iron of hemes III and IV are coordinated with standard bis-histidine coordination, whereas the heme iron ligation seen in hemes I and II appears more unusual. Although two histidines coordinate the iron of heme I, this appears to be the first example of a heme iron coordinated by the No atom of a histidine side chain (Figure 8.4) while heme II 190 is a true 5-coordinate heme, with no known role other than electron transfer. Cyt c554 is not known to bind small molecules physiologically; it neither is associated with enzymatic activity nor is it regulated allosterically by small molecules such as hydroxylamine. Recent unpublished data indicates that several small molecules (NO, CN-, F-) that typically bind to 5-coordinate hemes do not bind to the open position of heme II. The crystal structure of cyt c554 shows that the side chains of Thr 154, Pro 155 and Phe 156 shield the open ligation site on the iron of heme II. The inaccessibility of this heme site affords a heme reduction potential that is not significantly affected by direct solvent interactions. (Figure 8.4 near here) •• High-resolution structural information for cyt c55426 has allowed measurement of the bond lengths of the histidine ligands to the heme iron, which are listed in Table 1. The bond lengths vary depending on the oxidation-state of the protein, as expected. The most interesting aspect of the hemes is their packing arrangement in the protein. The four hemes of cyt c554 stack into distinct types of pairs (Figure 8.5). The first type of pair appears stacked with the porphyrin planes approximately parallel to one another and has two slight variants, depending of the coordination of the heme iron. The second type of pair shows two hemes in van der Waals contact with the heme planes approximately perpendicular. 191 The first variant of the stacked heme pair occurs when two hemes of bis-histidine iron coordination are related by a two-fold axis passing between the hemes, and separated by a plane-to-plane distance of ~3 .6 A to ~3 .8 A (Figure 8.5A). The heme orientation of one site is inverted relative to the other site. Other proteins exhibiting this type of stacked heme pairs include the split-Soret cytochrome 28 , cyt c NR 29 , cytochrome c fumarate reductase (cyt c FR) 30-32 and HAO 33 , the redox partner of cyt c554. A second variant of the stacked heme paif occurs when a six-coordinate bishistidine heme stacks onto a five-coordinate heme with a ~30° angle between the heme edge and the two-fold axis (Figure 8.5B). This stacking arrangement has been observed in HAO 33 and cyt c NR29 , however in each of these cases, catalytic activity is associated with the five-coordinate heme29 , 34 , and in cyt c NR, the fifth ligand is a lysine, albeit with a bent conformation reminiscent of a histidine ring (Figure 8.6). (Figure 8.5 near here) In contrast to the stacked pairs exhibited by hemes I and III and hemes II and IV of cyt c554, the central hemes (hemes III and IV) of cyt c554 are packed in with the heme planes at an angle of 110° to one another (Figure 8.5C). This approximately perpendicular arrangement has packing angles similar to those exhibited by cyt c335 and cyt c3 like proteins 36-39 , as well as HAO 33 , and cyt c NR29 . Both types of the heme 192 packing observed in cyt c554 appear to be emerging motifs in multi-heme proteins involved in metabolic processes requiring the rapid transfer of multiple electrons. The heme arrangements in cyt c554, HAO and cyt c NR exhibit more extensive similarities beyond the pairs of hemes 25 , 29 . In these cases, the heme packing motifs extends to three or four aligned hemes, with the five-coordinate heme overlaying. An intriguing aspect of the heme arrangement in cyt c554, HAO and cyt c NR is that the polypeptide chain immediately adjacent to the heme environment exhibits structural similarity 25 , despite the lack of sequence identity (Figure 8.6). Although the polypeptide chain is similar for these three proteins, the conservation of buried water molecules between cyt c554 and cyt c NR appears limited to those that orient the histidine ligands to the heme irons26 . Water molecules cannot clearly be identified in the structure of HAO, however, the imidazole rings of the histidine ligands to the heme iron have the same relative orientation as those seen in cyt c554 and cyt c NR. The relative orientations of histidine ligands to heme irons has been suggested to alter the electronic configuration of the porphyrin40 , and thus may be functionally relevant in the formation of the stackedheme pairs. (Figure 8.6 near here) The heme alignment and corresponding secondary structure similarity between cyt c554, HAO, and cyt c NR strongly suggest a possible evolutionary relationship 193 between these three proteins. While cyt c554 and HAO are involved in the nitrification process, oxidizing NH 3 to No 2·, cyt c NR acts in the reverse process of nitrite ammonification, reducing No 2- to NH 3 at a catalytic heme. Cyt c NR is additionally capable of catalyzing the reduction of sulfite to hydrogen sulfide. The remarkable similarity in the heme packing motifs of these three proteins combined with the similar active site architecture between HAO and cyt c NR support evolutionary relatedness between these proteins that act in catalyzing a similar reaction in physiologically opposite directions. Although cyt c554 and HAO have virtually no sequence identity in the structurally similar heme core environment, it is striking that cyt c554 contains a motif that is a subset of the structure of HAO. Similar internal structure of these proteins acting in the same pathway suggests a common evolutionary ancestor, however the lack of sequence identity precludes a definitive determination of the evolutionary relationship between the redox partners. Since the reactants for HAO and cyt c NR are so similar and the geochemical milieu supporting their evolution had these reactants in common, it is possible that HAO, cyt c554, and cyt c NR share a common ancestry. The divergence between these proteins may have occurred long enough ago that the sequence similarity has all but disappeared, but a significant similarity of secondary structure has been retained. 194 The Phe130-His131 peptide bond One interesting feature discovered in the structure determination of the tetragonal crystal form of cyt c554 was the presence of a non-proline cis-peptide bond between Phe 130 and His 131 . This cis-peptide bond appeared to be stabilized by crystal contacts (Figure 8.7 A) that were not present in he rhombohedral crystal form. In the absence of the stabilizing crystal contacts, the peptide bond appeared to be trans-, but with disallowed angles of <l> and ~ (Figure 8.7B). Cis-peptide bonds can be functionally relevant 41 , 42 , and the transient cis-peptide bonds stabilized by crystal contacts in flavodoxin have been suggested to be necessary for catalytic activity, however a functional role for the cis-peptide bond in cyt c554 has not yet been conclusively determined. (Figure 8. 7 near here) SPECTROSCOPIC PROPERTIES The relatively unusual spectral properties of cyt c554 may reflect the interactions between the stacked heme arrangements. Other proteins with similar heme packing arrangements also exhibit complex spectral properties, suggesting the hemes are coupled in some manner, but the detailed spectra of each protein is distinct. Although the splitSoret cytochrome c 43 has not been thoroughly investigated spectroscopically, it exhibits unusual optical properties. In the remainder of the proteins known to have stacked 195 hemes, i.e. cyt c554 HAO 8 , 4 4 , and cyt c NR 4 5, 46 , the EPR or Mossbauer spectra suggest the hemes are coupled in some manner. Optical spectroscopy Cyt c554 was named for the 554 nm absorbence of the a-band 15 seen upon reduction at physiological pH, however the spectral property used for the protein's nomenclature is pH dependent, and can vary between 55 1 nm and 556 nm8 . Additionally, both the Soret maximum and spectral features associated with the spin-state of the heme irons are pH dependent. This includes a shoulder to the Soret maximum at 360 nm, which indicates the presence of low-spin heme iron, and the presence of the charge transfer band at 645 nm associated with the presence of high-spin heme iron. EPR spectroscopy The EPR spectra of cyt c554 8, 47 differ fundamentally from the spectra of classic hemecontaining proteins48 . Like the visible spectra, they are highly pH dependent and are additionally sensitive to ionic strength 8 . g-values of 2.85, 2.2, 1.55 and 2.67, 2.2, 1.76 at pH 10.5 indicate the presence of two distinct low-spin species. In contrast, the spectra show peaks near g=6 and g=2 at low pH, indicating predominantly high-spin heme. A dominant g-value of 3.3 at neutral pH cannot be ascribed to any known spin state of heme iron or to any mixture of spin states for four heme irons. As there are no other EPR 196 active species in the protein, it has been suggested that g=3 .3 results from some unknown combination of heme-heme interactions. Herne coupling could be affected by the tight heme packing motifs seen in the crystal structure. The packing of the four hemes in cyt c55425 is similar to the packing of hemes 3-6 in HA0 33 and the EPR spectra from both proteins generally reflect the presence of hemeheme interactions. The EPR spectra of HAO are consistent with pair-wise interactions between the heme pairs 3 and 5, and pairs 4 and 644, 49 . The strength of the interactions is indicative of an exchange coupling between the iron sites. Surprisingly however, no bond pathway is observed in the crystal structure that would be consistent with the strength of the exchange interaction. Consequently, it appears that the exchange coupling is derived from an electronic overlap between the n-orbitals of the porphyrins. The EPR spectra of cyt c554 are also consistent with a similar exchange pathway between all four of the hemes. The main differences between the cyt c554 and HAO EPR spectra may be due to the accessibility of solvent to the 5-coordinate heme site in HAO. Mossbauer spectroscopy Mossbauer spectroscopy of cyt c554 reveals the presence of both high-spin and low-spin heme iron. At physiological pH, 75% of the hemes are low-spin while 25% of the hemes are high spin 8 , suggesting the presence of three six-coordinate hemes and one five- 197 coordinate heme. This is consistent with the optical and EPR spectra described above as well as Resonance Raman spectroscopy50 that showed the presence of both low-spin and high-spin heme. The Mossbauer spectra do not contain hyperfine interactions 8 suggesting the hemes are magnetically coupled, and further supporting coupling as the rationale for the appearance of the atypical EPR spectra. This coupling of the heme irons may occur in pairs, consistent with contact shifted NMR spectra51 and the stacking of heme pairs observed in the crystal structure25 . The Mossbauer spectrum of the high-spin iron of oxidized cyt c554 (8=0.40 mm/s, ~Eq=l.10 mm/s) originates from a species that is not observed in the Mossbauer spectra of oxidized HAO. Furthermore, the high-spin site of reduced cyt c554 (8=0.94 mm/s, ~Eq=2.43 mm/s) has substantially different iron parameters than that of reduced HAO (8=0.96 mm/s, ~Eq=4.21 mm/s) 52 . These differences indicate that while the heme pair containing the 5-coordinate heme in both proteins stacks similarly, the electronic structure of the 5-coordinate hemes is significantly different. HAO contains a covalently linked tyrosine to the 5-coordinate active site heme, but how this would affect the electronic configuration is uncertain. 198 FUNCTIONAL ASPECTS Oxidation and Reduction of Cyt c554 The reduction potentials of the four hemes of cyt c554 at pH 7.0 have been determined by a combination of voltammetry and spectroscopy, using an optically transparent thin layer electrode24 . The high-spin heme has a reduction potential of +47 mV, while the low-spin hemes titrate with midpoint potentials of +47 mV, -147 mV, and -276 mV. Five-coordination results in a high-spin heme iron 53 , thus heme II can be assigned a reduction potential of +4 7 m V; however, the remainder of the reduction potentials have not yet been assigned. Electron transfer kinetics As cyt c554 physiologically acts as a two-electron acceptor9 , the rates of electron transfer of two electrons are of interest. Stopped-flow kinetic studies of electron transfer at physiological pH revealed that one electron is passed from HAO to cyt c554 with a rate constant of 250-300 s- 1 , while a second electron is passed ten times more slowly with a rate constant of 25-30 s- 1_s4 Kinetics of dithionite reduction of cyt c554 appear biphasic and result in nonphysiological four-electron reduction of the protein. The first phase mimics the 199 physiological reduction rates, and involves the reduction of two hemes. In contrast, the second phase, involving the reduction of the remaining two hemes, proceeds over 300 times more slowly 17 . The multi-phasic reduction kinetics is consistent with negative cooperativity between the hemes, or a conformation change occurring during reduction and may reflect a large backbone movement observed between the oxidized and reduced states of cyt c55426 . The conformational change occurs in a stable loop (residues 175 179) between heme III and heme IV and displaces the loop over 2 A (Figure 8.8). Residue 179 serves as the histidine ligand to heme III, and the conformational change may be triggered by a change in peptide orientation between Met 178 and His 179. In the oxidized crystal form, the carbonyl group of this peptide bond appears to be in van der Waals contact with the porphyrin of heme III. Upon the reduction of heme III, this carbonyl group flips away from the porphyrin. As the reduction potential for heme III has not been assigned, it is difficult to interpret the relevance of the shift in backbone position upon the change in oxidation state. Interaction of cyt c554 with other members of the nitrification pathway (Figure 8.9 near here) Based on the pl values for HAO (3.4) 55 , cyt c554 (10.7) 15 and cyt c5 52 (3. 7) 15 , electrostatic effects probably contribute to intermolecular interactions between the soluble components of the nitrification pathway. Electrostatic surface representations of HA0 33 and cyt c55425 show that HAO and cyt c554 contain 200 apparently complementary surfaces that are negatively and positively charged, respectively, indicating a suitable binding mode for these two molecules (Figure 8.9). Although the solution structure of cyt c552 has been solved56 , a clear binding mode for cyt c552 and cyt c554 remains to be established. 201 8.3D representation Two views of the fold of cyt c554 are shown with a 90° rotation between them. The N-terminus is colored in black, and the protein can be traced by following the color gradient to the C-terminus (red). Figure 8.1 The biological nitrogen cycle. cytochrome c nitrite reductase assimilatory nitrate reductase cytochrome cd1 nitrite reductase copper nitrite reductase nitrous oxide reductse assimilatory nitrite reductase 202 Figure 8.2 Biological nitrification catalyzed by Nitrosomonas europaea. The role of cyt c552 is uncertain in this pathway. HAO \ ~ cyt c554 4e- • cyt c552 --•.- terminal oxidases Figure 8.3 Heme organization and interheme iron distances. 203 Figure 8.4 The unusual histidine Nb ligation of heme I of cytochrome c554. The model is superimposed onto an IF) IFcl omit map at 1.6 A resolution calculated after the removal of His 102. The histidine 1s stabilized m this orientation by a hydrogen-bond from the Ne atom to Gln 126. 204 Figure 8.5 Two variants of parallel stacked heme pairs. On the left are 0° and 90° views of the stacked heme pairs, where both hemes are six-coordinate with bis-histidine hemeiron ligation. On the right are 0° and 90° views of the stacked heme pairs when one hemes has six-coordinate bis-histidine ligation and the other has five-coordinate ligation. Figure 8.6 Stereoview of the alignment of the heme cores of HAO (blue), cyt c554 (red), and cyt c NR (green). 205 Figure 8.7 The His 129-Phe130 peptide bond appears cis- in the 1.6 A resolution crystal structure determination with the rhombohedral crystal form (top) and trans- in the 2.6 A resolution tetragonal crystal form (bottom). Glu 131 His 129 Phe 128 Glu 131 206 Figure 8.8 A conformational change occurs near heme III and heme IV upon reduction. In the oxidized state (black), the Met 178-his 179 peptide bond is within 5 A of the heme porphyrin. Upon reduction (white) the peptide bond flips away, causing a >2 A displacement of the loop between residues 173 and 179. Figure 8.9 Stereoview of a GRASP electrostatic surface potential representation of HAO (center) and cyt c554. The complementary negatively and positively charged surfaces respectively indicate a likely binding mode (yellow). cyt c554 cyt c554 cyt c554 unoccupied binding site unoccupied binding site 207 REFERENCES 1. P. M. Wood, in J. I. Prosser (ed), Nitrification, IRL Press, Oxford, pp 39-62 (1986) 2. R. A. Herbert, FEMS Microbial. Rev., 23, 563 - 590 (1999). 3. H. A. Painter, in J. I. Prosser (ed), Nitrification, IRL Press, Oxford, pp 185-211 (1986) 4. J. E. Alleman and K. Preston, in A. Swann (ed.), Proceedings of the Second Annual Conference of Commercial Aquaculture, Bloomington, IL (1991). Also available on the W or 1d Wide Web : http://ce.ecn.purdue.edu/~alleman/w3-articles/nitrifier- physiology/nitrifier-behavior.html 5. J. A. Zahn and A. A. Dispirito, J. Bacterial., 178, 1018-1029 (1996). 6. J. A. Zahn, D. M. Arciero, A. B. Hooper and A. A. DiSpirito, FEBS Lett., 397, 35-38 (1996). 7. J. H. Shears and P. M. Wood, Biochem. J., 226, 499-507 (1985). 8. K. Andersson, J. Lipscomb, M. Valentine, E. Mi.inck and A. Hooper, J. Biol. Chem., 261, 1126-1138 (1986). 9. D. M. Arciero, C. Balny and A. B. Hooper, J. Inorg. Biochem., 36, 217-217 (1989) . 10. T. Yamazaki, Y. Fukumori and T. Yamanaka, J. Biochem., 103, 499-503 (1988). 11. A. A. DiSpirito, L. R. Taaffe, J. D. Lipscomb and A. B. Hooper, Biochim. Biophys. Acta, 827, 320-326 (1985). 12. R. C. Prince and G. N. George, Nat. Str. Biol., 4, 247-250 (1997). 13. D. J. Richardson and N. J. Watmough, Curr. Op. Chem. Biol., 3, 207-219 (1999). 208 14. T. Yamanaka, Plant Cell Physiol. , 37, 569-574 (1996). 15. T. Yamanaka and M. Shinra, J. Biochem., 75, 1265-1273 (1974). 16. D. C. Y. Tsang and I. Suzuki, Can. J. Biochem., 60, 1018-1024 (1982). 17. A. A. DiSpirito, C. Balny and A. B. Hooper, Eur. J. Biochem, 162, 299-304 (1987). 18. H. McTavish, D. M. Arciero and A. B. Hooper, Arch Biochem Biophys, 324, 53-58 (1995). 19. N. G. Hommes, L.A. Sayavedra-Soto and D. J. Arp, Gene, 146, 87-89 (1994). 20. H. McTavish, F. LaQuier, D. M. Arciero, M. S. P. Logan, G. Mundfrom, J. A. Fuchs and A. B. Hooper, J. Bacterial., 175, 2445-2447 (1993). 21. D. Bergmann, D. Arciero and A. Hooper, J. Bact. , 176, 3148-3153 (1994). 22. M. A. Bruns, J. M. Fries, J. M. Tiedje and E. A. Paul, Microb. Ecol. , 36, 293-302 (1998). 23 . L. C. Crossman, J. W. B. Moir, J. J. Enticknap, D. J. Richardson and S. Spiro, Microbiology, 143, 3775-3783 (1997). 24. D. M. Arciero, M. Collins, J. Haladjian, P. Bianco and A. B. Hooper, Biochemistry, 30, 11459-11465 (1991). 25 . T. M. Iverson, D. M. Arciero, B. T. Hsu, M. S. P. Logan, A. B. Hooper and D. C. Rees, Nat. Struct. Biol., 5, 1005-1012 (1998). 26. T. M. Iverson, D. M. Arciero, A. B. Hooper and D. C. Rees, (in preparation). 209 27. C. Nagata, N. lgarashi, H. Moriyama, T. Fujiwara, Y. Fukumori and N. Tanaka, J. Biochem (Tokyo), 117, 931-932 (1995). 28. P. Matias, J. Morias, A. Coelho, R. Meijers, A. Gonzalez, A. Thompson, L. Sieker, J. LeGall and M. Carrondo, J. Biol. Inorg. Chem. , 2, 507-514 (1997). 29. 0. Einsle, A. Messerschmidt, P. Stach, G. P. Bourenkov, H. D. Bartunik, R. Huber and P. M. H. Kroneck, Nature, 400, 476-480 (1999). 30. V. Bamford, P. S. Dobbin, D. J. Richardson and A. M. Hemmings, Nat. Struct. Biol., 6, 1104-1109 (1999). 31. D. Leys, A. S. Tsapin, K. H. Nealson, T. E. ··Meyer, M. A. Cusanovich and J. J. VanBeeumen, Nat. Struct. Biol., 6, 1113-1117 (1999). 32. P. Taylor, S. L. Pealing, G. A. Reid, S. K. Chapman and M. D. Walkinshaw, Nat. Struct. Biol., 6, 1108-1112 (1999). 33. N. Igarashi, H. Moriyama, T. Fujiwara, Y. Fukumori and N. Tanaka, Nat. Struct. Biol., 4, 276-284 (1997). 34. A. B. Hooper and K. R. Terry, Biochemistry, 16, 455-459 (1977). 35. P. M. Matias, C. Frazao, J. Morais, M. Coll and M. A. Carrondo, J Mal Biol, 234 , 680-699 (1993). 36. I. B. Coutinho, D. L. Turner, M. Y. Liu, J. LeGall and A. V. Xavier, J. Biol. lnorg. Chem., 1, 305-311 (1996). 210 37. L. Banci, I. Bertini, M. Bruschi, P. Sompornpisut and P. Turano, Proc. Natl. Acad. Sci. USA, 93, 14396-14400 (1996). 38. M. Czjzek, F. Guerleqiun, M. Bruschi and R. Haser, Structure, 4, 395-404 (1996). 39. P. M. Matias, L. M. Saraiva, C. M. Soares, A. V. Coelho, J. LeGall and M. A. Carrondo, J. Biol. Inorg. Chem. , 4, 478-494 (1999). 40. N. V. Shokhirev and F. A. Walker, J. Biol. lnorg. Chem., 3, 581-594 (1998). 41. M. L. Ludwig, K. A. Pattridge, A. L. Metzger, M. M. Dixon, M. Eren, Y. Feng and R. P. Swenson, Biochemistry, 36, 1259-1280 (1997). 42. A. Jabs, M. S. Weiss and H. R, J. Mal. Biol., 286, 291-304 (1999). 43. M.-C. Liu, C. Costa, I. B. Coutinho, J. J. G. Moura, I. Moura, A. V. Xavier and J. LeGall, J. Bacterial., 170, 5545-5551 (1988). 44. M. Hendrich, M. S. P. Logan, K. Andersson, D. M. Arciero, J. Lipscomb and A. B. Hooper, J. Am. Chem. Soc., 116, 11961-11968 (1994 ). , 45. W. Schumacher and P. M. H. Kroneck, Arch.Microbial., 156, 70-74 (1991). 46. W. Schumacher, U. Hole and P. M. H. Kroneck, Biochem. Biophys. Res. Commun., 205, 911-916 (1994). 47. K. K. Andersson, J. D. Lipscomb and A. B. Hooper, Inorg. Chim. Acta, 79, 181-182 (1983). 48. G. Palmer, in A. B. P. Lever and H. B. Gray (ed), Iron Porphyrins part Two, AddisonWesley Publishing Company, London, pp 43-85 (1983). 21 1 49. D. M. Arciero, A. Golombek, M. P. Hendrich and A. B. Hooper, Biochemistry, 37, 523-529 ( 1998). 50. K. K. Andersson, G. T. Babcock and A. B. Hooper, FEBS Letters, 170, 331-334 (1984) . 5 1. K. K. Andersson, J. D. Lipscomb and A. B. Hooper, Inorg. Chim. Acta, 79, 181-182 (1983). 52. K. K. Andersson, T. A. Kent, J. D. Lipscomb, A. B. Hooper and E. Mtinck, J. Biol. Chem., 259, 6833-6840 (1984). 53 . W. R. Scheidt and M. Gouterman, in A. B. P . Lever and H. B. Gray (ed), Iron Porphyrins Part One, Addison - Wesley Publishing Co, Inc., London, pp 89-139 (1983) 54. D. Arciero, C. Balny and A. Hooper, Biochemistry, 30, 11466-11472 (1991). 55. A. B. Hooper, P. Maxwell and K. Terry, Biochemistry, 17, 2984-2989 (1978). 56. R. Timkovich, D. Bergmann, D. M. Arciero and A. B. Hooper, Biophys. J., 75 , 19641972 (1998). 212 Appendices: Carbonic Anhydrases A-1. Archaeal methanogenesis and the y-class carbonic anhydrase from Methanosarcina thermophila Methane producing microbes release about 400 million metric tons of methane into the atmosphere each year, thus significantly contributing to the global greenhouse effect. Microbial methane production requires strictly anaerobic environments ( 1 ), which can include the interior of animals (such as the rumen of cattle), watery landscapes (natural wetlands, bogs, or rice paddies), or human-made sites (landfills or sewage digestors). In any of these locations, the biological production of methane requires a food chain of at least three interacting metabolic groups of obligate anaerobes (Fig. A-1.1) (2) . One type of anaerobe, the methanogenic archaeon, derives all its energy from the methanogenesis of a very restricted number of substrates: acetate; carbon dioxide; hydrogen; formate; methanol; and methylamines (3). Depending upon the substrate available, methanogenic archaea use separate pathways (4) in which methane is produced from the methyl group of acetate (reaction 1), the reduction of carbon 213 dioxide (reactions 2a and 2b ), or the reduction of methanol or methylamines (reaction 3). The pathways of methane formation described by each equation involve a number of enzymes and coenzymes that are found exclusively in archael methanogenic pathways (5). The existence of these unique enzymes has led to the proposal of the three-domain phylogeny: archaea; bacteria; and eukarya (6). CH3Coo- + H+ ➔ CH4 + CO2 (1) CO 2 + 4H 2 ➔ CH4 + 2H 2O (2a) 4H2CO 2 ➔ 3CO2 + CH 4 + 2H2O (2b) 4(R-CH3 ) + 2H2O ➔ 4RH + 3CH4 + CO 2 (3)" The methanogenic archaeon Methanosarcina thermophila can derive all of its energy from the conversion of acetate to methane (reaction 1, Fig. A-1.2) in conditions where other substrates are not available. In the first step of the fermentation of acetate, acetate kinase (Ack) phosphorylates the carboxyl group of acetate (7, 8), consuming one molecule of ATP (Fig. A.1.3 step 1). Following this, the substrate is activated by phosphotransacetylase (Pta) (9) yielding acetyl-CoA as the product (Fig. A-1.3 step 2). The key enzyme in the aceticlastic pathway is the five-subunit carbon monoxide dehydrogenase (Cdh). Cdh cleaves the C-C and C-S bonds in the acetyl moiety of acetyl-CoA (10, 11 ), oxidizes the carbonyl group to CO2, and transfers the methyl group to tetrahydrosarcinapterin (12) (H4SPT; Fig. 214 A.1.3 step 3). Electrons derived from the carboxyl group of acetate are passed from Cdh to a transmembrane cytochrome b (cyt b) complex by ferredoxin (FdxA) (13) (Fig. A-1.3 step 4). It has been suggested that the integral-membrane cyt b complex may aid in the formation of a transmembrane proton gradient used to drive ATP synthesis ( 14). The CH 3 -H4 SPT molecule is activated by a membrane-bound methyltransferase (Tase) to form CH 3-S-CoM (Fig. A-1.3 step 5). This reaction may be coupled to the transport of a sodium ion across the membrane of M. thermophila, which could help drive ATP synthesis ( 15 ). The activated CH 3-S-CoM can now be acted on by methyl-CoM reductase (Mer; Fig. A-1.3 step 6), an enzyme common to all methanogenic pathways (reactions 1 - 3) and the final chemical reaction occurring during methanogenesis ( 16). Electrons are supplied to this last reaction by heterodisulfide reductase (Hdr) (17) and coenzyme B (Fig. A-1.3 step 7). During the fermentation of acetate by M. thermophila, the activity of a secreted carbonic anhydrase (Cam) increases, suggesting involvement of this enzyme in acetate catabolism (18). It is proposed that CO 2, a byproduct of the reaction catalyzed by Cdh, is converted to HC0 3- by Cam outside of the cell to facilitate either the uptake of acetate or the removal of CO2 from the cytoplasm (Fig. A-1.3 step 8). 215 The structure of Cam has previously been solved at 2.8 A resolution (19) revealing an unusual left-handed P-helix fold, and showing the evolutionary convergence of the ligation of the active site zinc between Cam and mammalian carbonic anhydrases. Using a new crystal form, the structure of Cam has been solved in nine states of the enzyme at resolutions between 1.46 A and 1.95 A by molecular replacement (20; Iverson et al. in press). The high-resolution structure of Cam, combined with previous kinetic studies (18, 21, 22, 23 ), has revealed mechanistic insights in the function of this enzyme. 216 Fig. A-1.1: Microbial Food Chain. Three of the four types of microbes shown are required for the complete decomposition of complex biomass to methane and carbon dioxide. Complex organics i CH3 CH2 COOCH3 CH2 CH 2 COO· 4 ! .. l H2 -producing I acetogenic rCO -reducing 2 ! acetogenic CH COO- - - - - - L_ bac~_"~-- _ 3 bacteria Methanogenic archaea • "CH 3C0 2H Fig. A-1.2 Pathway for the convers10n of acetate to methane by methanoarchaea. l Chemically modified acetate intermediates are indicated with necessary CoASH · CH3COSC0A 2e-,2w + co2 CoASH THSPt cofactors, intermediate products, and reactants indicated. "CH 3-THSPt HSCoM - - t -THSPt Enzymes used in this pathway are shown in Fig. A.1.3. · cH 3-SCoM HSCoB _j 2e- ! - - - - CoM-S-S-CoB + 2W 217 Fig. A-1.3: Methanogenic fermentation of acetate by M. thermophila. Steps described in the text are indicated by number. Electron transfer steps are indicated with a black arrows, while chemical pathways are indicated with colored arrows. - C~am - - CO, hD / FdxA Cfi,-THSPt ' THSP 4• 218 REFERENCES 1. Whitman, W., Coleman, D. , and Wiebe, W. (1998) Proc. Natl. Acad. Sci. USA 95, 6578-6583. 2. Zinder, S. (1993) in Methanogenesis (Ferry, J., Ed.) pp 128-206, Chapman and Hall, New York. 3. Ferry, J. (1997) BioFactors 6, 25-35. 4. Blaut, M. (1994) Antoine van Leeuwenhoek 66, 187-208. 5. DiMarco, A., Bobik, T., and Wolfe, R. (1990) Annu. Rev. Biochem. 59, 355-394. 6. Woese, C. , Kandler, 0., and Wheelis, M. (1990) Proc. Natl. Acad. Sci. USA 87, 4576-4579. 7. Aceti, D., and Ferry, J. (1988) J. Biol. Chem. 263, 15444-15488. 8. Fisher, R., and RK, T. (1988) FEES Lett. 228, 249-253. 9. Lundie, L., and Ferry, J. (1989) J. Biol. Chem. 264, 18392-18396. 10. Abbanat, D., and Ferry, J. (1991) Proc. Nat. Acad. Sci. USA 88, 3272-3276. 11. Raybuck, S., Ramer, S., Abbanat, D., Peters, J., Orme-Johnson, W., Ferry, J., and Walsh, C. (1991) J. Bacterial. 173, 929-932. 12. Grahame, D. (1991) J. Biol. Chem. 266, 22227-22233. 13. Fisher, R., and Thauer, R. (1990) FEES Lett. 269, 368-372. 14. Mountfort, D. (1978) Biochem. Biophys. Res. Commun. 85, 1346-1350. 15. Al-Mahrouq, H., Carper, S., and Lancaster, J. J. (1986) FEES Lett. 207, 262-265. 219 16. Jablonski, P., and Ferry, J. (199 1) J. Bacterial. 173, 2481-2487. 17. Hedderich, R., Berkessel, A., and Thauer, R. (1990) Eur. J. Biochem. 193 , 255261. 18. Alber, B. E., and Ferry, J. G. (1 994) Proc. Natl. Acad. Sci. USA 91, 6909-6913. 19. Kisker, C., Schindelin, H., Alber, B. E. , Ferry, J. G., and Rees, D. (1996) EMBO ]. 15, 2323-2330. 20. Iverson, T., Alber, B. , Kisker, C., Ferry, J., and Rees, D. (submitted) Biochemistry. 21. Alber, B. E. , and Ferry, J. G. (1996) J. Bacterial. 178, 3270-3274. 22. Alber, B. E., Colangelo, C. M., Dong, J., Stalhandske, C. M. V., Baird, T. T., Tu, C. , Fierke, C. A., Silverman, D. N., Scott, R. A., and Ferry, J. G. (1999) Biochemistry 38, 13119-13128. 23. Tripp, B. C. and Ferry, J.G. (submitted) Biochemistry. 220 Appendix A-2 Iverson, T.M., Alber, B.E., Kisker, C., Ferry, J.G., Rees, D.C. (2000) A closer look at the active site of y-class carbonic anhydrases: high-resolution crystallographic studies of the carbonic anhydrase from Methanosarcina thermophila. revision. Biochemistry. In 221 A Closer Look at the Active Site of y-Class Carbonic Anhydrases: High-resolution Crystallographic Studies of the Carbonic Anhydrase from M ethanosarcina thermophila t ,§§ Tina M. Iverson+, Birgit E. Alber§·®, Caroline Kisker 11 ·#, James G. Ferry\ and Douglas C. Rees 11 ·.1• :j: Graduate Option in Biochemistry, 147-75CH, California Institute of Technology, Pasadena, CA, 91125 § Department of Biochemistry and Molecular Biology, Pennsylvania State University, University Park, PA 16802 @ Present address: Institut fi.i r Biologie II, Mikrobiologie, Albert-LudwigsUniversitiit Freiburg, Schiinzlestr. 1, D-79104 Freiburg, Germany Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125 II # Present address: State University of New York At Stony Brook, Department of Pharmacological Sciences, Stony Brook, NY 11794-8651 ..L Howard Hughes Medical Institute, Division of Chemistry and Chemical Engineering, 147-75CH, California Institute of Technology, Pasadena, CA 91125 *To whom correspondence should be addressed at dcrees@caltech.edu. Phone (626) 395-8393 FAX (626) 744-9524 t This work was supported by a grant from the National Institutes of Health to J.G.F. (GM44661). T.M.I. is supported by a NIH training grant (GM07737) . §§ Coordinates for all structures described here have been deposited with the PDB and have been released. PDB accession numbers are listed in Table A-2.1. RUNNING TITLE: Crystallographic studies of y-carbonic anhydrase 222 Abbreviations: Cam, Methanosarcina thermophila carbonic anhydrase; Zn-Cam, Cam that has been heterologously produced in Eschericia coli, metal-depleted and reconstituted with zinc; Co-Cam, Cam that has been heterologously produced in E. coli, metal-depleted and reconstituted with cobalt; HCAII, human carbonic anhydrase isozyme II; E, enzyme; Bet, bicarbonate; PEG, polyethylene glycol; EXAFS, extended x-ray absorption fine structure; PDB, Protein Data Bank. 223 ABSTRACT: The prototype of the y-class of carbonic anhydrase has been characterized from the methanogenic archaeon Methanosarcina thermophila. Previously reported kinetic studies of the y-class carbonic anhydrase are consistent with this enzyme having a similar reaction mechanism as the mammalian a-class of carbonic anhydrase. However, the overall folds of these two enzymes are dissimilar and, apart from the zinccoordinating histidines, the active site residues bear little resemblance. The crystal structures of zinc-containing and cobalt-substituted y-class carbonic anhydrase from M. thermophila are reported here between 1.46 A and 1.95 Aresolution in the unbound form and co-crystallized with either SO/ or HC03-. Relative to the tetrahedral coordination geometry seen at the active site in the a-class of carbonic anhydrases, the active site of the y-class enzyme contains additional metal-bound water ligands, so the overall coordination geometry is trigonal bipyramidal for the zinc-containing enzyme and octahedral for the cobalt-substituted enzyme. Ligands bound to the active site all make contacts with the side chain of Glu 62 in manners that suggest the side chain is likely protonated. In the uncomplexed zinc-containing enzyme, the side chains of Glu 62 and Glu 84 appear to share a proton; additionally Glu 84 exhibits multiple conformations. This suggests that Glu 84 may act as a proton shuttle, which is an important aspect of the reaction mechanism of a-class carbonic anhydrases. A hydrophobic pocket on the surface of the enzyme may participate in the trapping of CO 2 at the active site. Based on 224 the coordination geometry at the active site, ligand binding modes, the behavior of the side chains of Glu 62 and Glu 84, and analogies to the well-characterized a-class of carbonic anhydrase, a more-defined reaction mechanism is proposed for the y-class of carbonic anhydrase. 225 Carbonic anhydrases are zinc-containing enzymes that were first noted for their physiological role in interconverting CO 2 and HCO 3- ( 1 ). Based on sequence similarities, there are three distinct classes of carbonic anhydrases: a, p, and y. The a-class is generally thought of as mammalian, and includes at least fourteen mammalian isozymes, along with two isozymes from Chlamydomonas reinhardtii, and two prokaryotic isozymes (2, 3 ). The P-class is primarily found in phototrophic organisms, including plants. The prototype for the y-class of carbonic anhydrase has been discovered in the methanoarchaeon Methanosarcina thermophila (Cam') (4). Crystal structures have been reported for Cam (5) and several isozymes of the a-class of carbonic anhydrase (6-11 ), but there is currently no crystallographic information available for any representative of the p-class of carbonic anhydrase. In M. thermophila, Cam activity increases when the substrate for growth is switched from methanol to acetate (12), suggesting involvement of this enzyme in acetate catabolism. M. thermophila obtains energy for growth by cleaving the C-C bond of acetate and reducing the methyl group to CH with electrons derived from the oxidation 4 of the carbonyl group to CO . 2 It is proposed that CO is converted to HCO 2 3 by Cam outside the cell to facilitate either the uptake of acetate or the removal of CO2 from the cytoplasm. Cam catalyzes the interconversion of CO 2 and HCO 3- with turnover numbers 226 as high as 6 x 104 sec- 1 (13 ), which approaches the turnover number of 106 sec- 1 observed for a-carbonic anhydrases (14) . Kinetic studies (13) suggest that the hydration of CO2 by Cam may occur by a mechanism similar to that proposed for human carbonic anhydrase II (HCAII) (14-17). HCAII, the best studied carbonic anhydrase and the isozyme with highest activity (18 ), belongs to the a-class of carbonic anhydrases and utilizes a "zinc-hydroxide" mechanism for catalysis. The overall enzyme-catalyzed reaction occurs in two distinct half-reactions. The first half-reaction is the interconversion of CO 2 and HC0 3- (eqs la and lb, where E signifies the enzyme) and involves nucleophilic attack of the zinc hydroxyl ion on CO 2 . This is followed by exchange of zinc bound HC03- with water. The second half reaction corresponds to the rate-determining intramolecular and intermolecular proton transfer steps (eqs 2a and 2b), which regenerates the zinc-hydroxide at the active site. E-Zn2+-0ff + CO2 H E-Zn2+-Hco 3- (la) E-Zn2+-HC0 3- + H20 H E-Zn2+-H 20 + HC0 3- (lb) E-Zn2+-H 20 H +H-E-Zn 2+-0ff (2a) +H-E-Zn 2+-0ff + B H E-Zn 2+-0ff + BH+ (2b) 227 The proton indicated in eqs 2a and 2b represents protonation of His 64 of HCAII, which shuttles protons between the active site zinc and the buffer molecules (B) in the solvent (19, 20) . Intramolecular proton transfer (eq 2a) is rate-limiting at saturating buffer concentrations. At low buffer concentrations, intermolecular transfer (eq 2b) is rate limiting. In addition to the similar kinetic properties, Cam and HCAII exhibit structural similarities in coordination of the zinc ion. The crystal structure of HCAII (8) reveals an active site containing a zinc ion tetrahedrally coordinated to three histidines and a highly conserved water molecule. The structure of Cam has been previously solved at 2.8 A resolution (5) . Although the overall folds of Cam and HCAII are unrelated, the histidines coordinating the Zn2+ ion of Cam superimpose on the corresponding histidines of HCAII. Neighboring residues in the active site of Cam differ completely from those in HCAII, however, and it is not obvious how the residues surrounding the metal site in Cam function, by analogy to catalytically essential residues in HCAII. One key residue for the catalytic mechanism in HCAII is Thr 199, which has no obvious structurally analogous residue in Cam ( 5) . Thr 199 plays two known mechanistic roles. The first is in selecting protonated molecules to bind to the active site zinc. Because of this role, Thr 199 is often referred to as the "gatekeeper" residue. To 228 achieve this selection, the side chain hydroxyl is hydrogen-bonded to the side chain carboxyl of Glu 106, so that the Thr 199 OH must be a hydrogen-bond donor in that interaction. Consequently, Thr 199 acts as a hydrogen-bond acceptor in any other hydrogen bond interactions. Thr 199 OH is directly hydrogen bonded to the Zn-bound water molecule and therefore selects only protonated molecules to bind in that position (21). As a result of this directional hydrogen-bond, a lone pair of the Zn-bound OH is optimally oriented fo r attack on CO 2 (22). Thr 199 may play other roles in the interconversion of HCO3- and CO 2 . Crystallographic studies of HCAII with the inhibitors CN- and CNo- have shown that, rather than binding to the active site metal as anticipated, both cN- and CNo- bind to the backbone amide nitrogen of Thr 199 (23). The results from this study suggest that the backbone amide nitrogen of Thr 199 is also involved in catalysis, possibly to both orient and/or polarize the CO 2 , preparing it for attack by the Zn-bound OH. There are also differences in the kinetic behavior of HCAII and Cam. In HCAII, a higher coordination number of the metal ion is postulated to decrease the turnover rate. Indeed, only cobalt substituted HCAII retains any activity, with 50% of the activity of the native enzyme (24 ). Crystal structures of HCAII metal-substituted with Co, Cu, Ni or Mn, show an increased number of metal coordinating ligands and longer average coordination distances (25), consistent with the theory that the coordination number is a 229 primary factor influencing the turnover rate in HCAII. In contrast to mammalian carbonic anhydrases, substitution of the zinc ion by cobalt doubles Cam activity ( 13 ), even though EXAFS analysis indicates a higher coordination number for cobaltsubstituted Cam (Co-Cam) relative to zinc-containing Cam (Zn-Cam). Consequently, while the mechanisms of the a- and y-classes of carbonic anhydrase enzymes may be similar, they are unlikely to be identical. Further understanding of the similarities and differences of a- and y-class carbonic anhydrases will identify chemical constraints within which the different classes evolved, and will provide further insight to the fundamental features of the catalytic mechanism of carbonic anhydrases. The crystal structures of uncomplexed Co-Cam, as well as Co-Cam cocrystallized with HC0 3- and SO/ were determined at high resolution to investigate whether the change in activity for Co-Cam is an effect of active site coordination, and to learn more about the mechanism of Cam. For accurate comparison to Zn-Cam, highresolution structures of Zn-Cam, as well as Zn-Cam co-crystallized with HC03- and SO/ have also been determined. These high-resolution structures allow a first look at the solvent structure of the enzyme. As solvent molecules are used as a catalytic group in the reaction, only high-resolution structures can give a complete picture of the active site. These structures allow assignment of several catalytically important residues and solvent molecules in the active site of Cam. 230 EXPERIMENTAL PROCEDEURES Crystallization . "Cam" refers to native carbonic anhydrase isolated from Methanosarcina thermophila (12) . "Zn-Cam" and "Co-Cam" designate enzymes in which the indicated metal was incorporated into the active site of as-purified enzyme produced in E. coli, by preparation of the apoenzyme using the denaturant guanidine-HCl and the metal chelator dipicolinate followed by reconstitution in the presence of ZnSO4 or CoC1 2 ( 1 3). For all crystallizations described, the stock protein concentration is 10 mg/mL buffered in 5 mM phosphate pH 7.0. Cubic crystals of Zn-Cam were obtained in hanging drops using 3% PEG 8000 and 0.1 M (NH 4 ) 2SO4 at 22 °C. Cubic crystals of CoCam were obtained in hanging drops containing 5% PEG 8000 and 0.5 M (NH 4 ) 2SO4 . The final pH of an artificial crystallization solution was measured as pH 6.2. The pH of the crystallization reaction including Co-Cam was measured using a rnicroelectrode as pH 5.8. The crystals appeared within 3 days and grew to a maximal size of 0.5 mm x 0.5 mm x 0.5 mm. Cubic crystals belonged to the space group P2 13 with unit cell dimensions between a=82.3 A and a=82.7 A and one monomer per asymmetric unit. For cryocooling, all crystals described in this paper were soaked in a solution containing all of the crystallization components plus 25% glycerol. 231 Co-crystals of Cam with bicarbonate were obtained in hanging drops containing 7.5 % PEG 8000, 0.8 M (NH 4 ) 2 SO4 , and 20 mM NaHCO 3 . The crystals were isomorphous with the substrate free crystals. Zn-Cam crystals only formed using PEG 8000 from Hampton, and formation of all bicarbonate co-crystals was dependent on the lot of PEG 8000. This is presumably due to the batch-to-batch variability of chemical impurities in PEG 8000 (26). When the uncomplexed crystals were left in the crystallization conditions longer than one day after crystal formation, a large, unanticipated feature appeared in the electron density maps adjacent to the metal ion. In view of the size and tetrahedral shape of this density, and the fact that (NH 4 ) 2 SO4 was present in the crystallization conditions, this density was modeled as SO/-. Data collection. Data for Zn-Cam complexed with so/-, as well as a low resolution pass for Zn-Cam were collected at 113 K using CuKa radiation from an RU 200 rotating anode at a wavelength of 1.54 A on an Raxis Ile image plate. All other data sets were collected at 93 K at the Stanford Synchrotron Radiation Laboratory beam line 7-1 (tl,=1.08 A) using a MAR Research image plate detector. All data were processed using DENZO and scaled with SCALEPACK (27). The resolutions of the data sets merged from 1.46 Ato 1.95 A (Table A-2.1). 232 The structure of Co-Cam was solved by molecular replacement in AMORE (28) using the Cam monomer (5) as a search model. The other cubic crystals of Cam were isomorphous with cubic Co-Cam. Therefore, the refined monomer of Co-Cam from the cubic space group was used as a starting model in those cases. Model Building and Refinement. All Cam models were built using the program 0 (29). Refinement was carried out using REFMAC (30, 31) and X-PLOR (32). Uncomplexed structures were first refined with bond lengths of the metal coordinating ligands restrained to the distances determined by EXAFS ( 13 ). After the Rcryst dropped below 19%, the refinement proceeded unrestrained. The Rrree was composed of 1093 randomly selected reflections from the native Co-Cam data set (6% ). The reflections for Rrree for the remaining data sets are identical to those from the Co-Cam data set with additional randomly selected reflections selected for data sets of higher resolution. PDB accession codes and statistics are given in Table A-2.1. RESULTS Overall fold of the enzyme. The overall fold of Zn-Cam is a left-handed 13-helix (Fig. A-2.1) as previously reported (5). Superposition of residues 9-213 of the monomers of Zn-Cam and Co-Cam resulted in an rms deviation of 0.22 A overall and 0.14 A for the 233 residues in the P-helix, indicating no significant overall structural change occurs upon metal substitution. When a trimer of Zn-Cam from the previously reported tetragonal space group (P432 12) (5) was superimposed onto a trimer of Co-Cam, therms~ was 0.37 A for the Ca-atoms of residues 6-212, and 0.23 A for the Ca-atoms of residues of the Phelix. Thus, with the exception of the N-termini described below, no significant structural changes occur in the trimer with the change in space group. The structure of Zn-Cam in the tetragonal space group (5) shows two trimers related by a two-fold axis with an interaction between the N-termini (residues 0-6) creating a six-stranded p-barrel. In the cubic space group P2 13, there is no two-fold axis relating the trimers, indicating that Cam forms a trimer and not a hexamer in the crystal. In the model of Zn-Cam in complex with HCO 3-, the N-termini are lying across the top of the trimer in an extended conformation. However, in the remainder of the models, the Ntermini are disordered. The active site. As previously described (5), the active site contains a zinc ion coordinated to the protein by three histidine side chains. In the structures described here, the active site coordination geometry differs depending on the divalent cation at the metal center and the presence or absence of additional ligands. The coordination geometry relative to the metal center is summarized in Table A-2.2. 234 1. Active site of uncomplexed Zn-Cam. In the active site of uncomplexed Zn- Cam, the Zn 2+ has three protein and two solvent ligands arranged in trigonal bipyramidal geometry (Fig. A-2.2A, Table A-2.3) . As was determined previously for Cam (5), the zinc is coordinated between two monomers with His 81 and His 122 contributed by one monomer, and His 117 contributed from an adjacent monomer. In the previously described structure, only one Zn-bound water molecule was identified, due to difficulties in conclusively assigning water molecules at 2.8 A resolution. The water molecule assigned in the 2.8 A resolution structure corresponds to Wat 2 in the high-resolution structure of Zn-Cam. In the current high-resolution study, an additional solvent molecule coordinating the zinc ion could be identified in the electron density maps. The coordination geometry for the Zn2+ is summarized in Table A-2.3. Penta-coordination is consistent with the EXAFS data, which shows a Zn 2+ coordination sphere of Zn(N,O) 2_3 (imidazole) 3 ( 13). The average bond distance of 2.14 A is within experimental error of the distance determined by EXAFS of 2.06 A, where the errors in EXAFS determination are approximately 0.02 A and the coordinate error for the crystal structures are listed in Table A-2.1. 235 An interesting aspect of these structures is the multiple conformations for the side chain of Glu 84. In Zn-Carn, two discrete conformations of Glu 84 are present, one of which points toward the active site and lies within hydrogen-bonding distance of the side chain of Glu 62. In the other structures described in this paper, a third conformation for the side chain of Glu 84 was observed (Fig. A-2.3A). Since the side chains of Glu 84 and Glu 62 lie within hydrogen-bonding distance, it is likely that a proton lies between these two side chains. This indicates that the protein environment has elevated the pKa value for one or both of these side chains above that observed in solution. The observation further suggests that protons may be shuttled ··out of the active site via a mechanism involving both Glu 62 and Glu 84. Glu 84 may be analogous to the proton shuttle residue His 64 of HCAII (Fig. A-3.3B) , which also exhibits two discreet conformations in the crystal structure (33 ). Replacement of Glu 84 with alanine in Carn decreases protein activity to 1% of wild type levels, and is restored to 53% of wild type activity by the addition of irnidazole (34 ). Irnidazole has been shown to act as an alternative proton shuttle in HCAII (20), lending biochemical support to the crystallographic observation that Glu 84 may act as a proton shuttle in Carn. 2. Active Site of Uncomplexed Co-Cam . The active site of Co-Cam shows the cobalt to be hexa-coordinate in a distorted octahedral arrangement, with ligands provided by the three histidine side chains and three water molecules (Fig. A-2.2B, Table A-2.4). 236 Hexa-coordination is consistent with the EXAFS data that suggests Co(N,O)i_ 3 (imidazole) 2_3 coordination ( 13) . Only one of the three water molecules (Wat 1) coordinating the cobalt corresponds to a water molecule (Wat 1) coordinating the zinc in Zn-Cam (Table A-2.2) . Thus, Wat 2 of Zn-Cam and Wat 2 and Wat 3 of Co-Cam are unique to each structure, respectively. In contrast, almost 70% of the water molecules not located in the active site have equivalent locations in these two uncomplexed structures. The average Co 2+-ligand bond length of 2.32 Aobserved for Co-Cam is similar to the average ligation distance of 2.2 A observed for other six-coordinate Co2+ in protein structures (35-39). The average Co-ligand bond length determined by EXAFS analysis of Co-Cam is 2.09 A (13) and the disagreement in bond lengths is just outside of the experimental error of both techniques. This disagreement may arise from pH differences between the crystallization conditions (pH 5.8) and EXAFS conditions (pH 7 .0). In HCAII, lowered pH lengthens metal-ligand bond lengths (8, 25, 40), so that a comparison of ligation bond lengths may not be accurate when the pH of the buffer differs. 3. Active Site of Zn-Cam Co-crystallized with Bicarbonate. Crystallographically, carbonic acid (H 2CO3 ) and bicarbonate (HCO 3-) are indistinguishable at the resolution of these structures. Although HCO3• was added the crystallization conditions, the species 237 bound to the active site might be H 2 CO 3 rather than HCO 3-. First, the pH of the crystallization conditions is very near the pKa of bicarbonate, so that protonation may have occurred in solution. Second, the pKa of bicarbonate could be elevated when bound to the protein, as appears to be the case for the neighboring Glu 62 and Glu 84 side chains. Third, HCO3- has a very low affinity for the enzyme, similar to that of HCAII. In HCAII, it has not been possible so far to co-crystallize native enzyme with the substrate due to the weak binding. H2CO3 might have more stable contacts than HCO 3 -, which could facilitate trapping of this species in the crystal. The hydrogen-bonding contacts of the species observed at the active site cannot unambiguously identify the nature of the ligand. The ligand has three oxygens within hydrogen-bonding distance (< 3.2 A ) of the side chain of Glu 62. It is not possible, however, to establish the location and nature of the bound protons, since the potential hydrogen-bonds are not optimally oriented. Furthermore, it is possible that the interaction between the side chain of Glu 62 and the active site ligand is primarily electrostatic, where the partial positive charge of one hydrogen stabilizes multiple oxygen-oxygen contacts. For the purposes of describing the interaction, the species will be referred to as bicarbonate (HCO 3-), despite its ambiguous identity. 238 Bicarbonate binds in a bidentate fashion to the Zn 2+ replacing both Wat 1 and Wat 2 (Fig. A-2.2C, Table A-2.5). The HC03- is stabilized by the coordination of one oxygen (03) directly to the active site zinc and by a hydrogen bond contact of this same oxygen (03) to the side chain of Glu 62. A second bicarbonate oxygen (01) makes a long contact to the Zn2+ . The third oxygen (02) is within hydrogen-bonding distance to both carboxylate oxygens of the Glu 62 side chain. 4. Active site of Co-Cam Co-Crystallized with Bicarbonate. HC0 3- binds in a monodentate fashion to the cobalt, replacing one of the coordinating water molecules (Wat 3 from Co-Cam) as well as a second water molecule (Wat 4 from Co-Cam) that is part of a water molecule network in the active site (Fig. A-2.2D, Table A-2.6). The water molecule replaced in the Co-Cam-HC0 3- complex is not present in Zn-Cam, so that the HC0 3- binding differs between Zn-Cam and Co-Cam. Multiple binding modes of HC0 3suggest that the active site cavity can accommodate a continuum of binding modes for the substrate, as is proposed for HCAII (41, 42). The ligand is partially stabilized in this binding mode by a hydrogen-bond contact to the side chain of Glu 62. 239 5. Active Site of Zn-Cam in Complex with Sulfate. Sulfate has been shown not to inhibit Cam activity at concentrations up to 0.3 M at pH 5.9 ( 34 ). Therefore it was surprising to find that in Cam, sulfate can bind directly to the Zn 2+ in a manner suggestive of an inhibitor. The binding replaces both zinc-coordinating waters and thus retains the distorted trigonal bipyramidal coordination (Fig. A-2.2E, Table A-2.7). One oxygen (01) is within hydrogen-bonding distance of the side chain of Glu 62. This suggests that a proton is shared between the side chain of Glu 62 and the active site ligand, and further supports an elevation of the pKa value for the GH.1 62 side chain. 6. Active Site of Co-Cam Complexed with Sulfate. In the complex of Co-Cam and SO/, the SO/ binds to the cobalt in a bidentate fashion, replacing Wat 1 and Wat 3, but retaining Wat 2 (Fig. A-2.2F, Table A-2.8). Although the SO/ contacts the side chain of Glu 62 as in the Zn-Cam-SO4 - complex, in the Co-Cam-SO/ structure, there are two contacts within hydrogen-bonding distance, rather than one. It is possible that these two contacts are sharing a single proton, most-likely donated by a protonated Glu 62. DISCUSSION 240 Reaction mechanism implications. As discussed above, kinetic analyses indicate that Cam and HCAII each catalyze the interconversion of CO 2 and HCO 3- through a zinchydroxide mechanism ( 13 ). Comparison of the active site structures of Cam and HCAII highlight side chains that may function analogously in these two enzymes. As HCAII has been extensively studied, this comparison helps construct a preliminary structure-based mechanism for Cam. Differences in the active sites are also of interest as they highlight aspects of the reaction mechanism that have evolved to proceed using different components, yet are still compatible with high rates in both cases. In the first half-reaction (eqs. la and lb), a metal-bound hydroxyl group directly attacks CO 2 . In HCAII, it is proposed that CO 2 binds in a hydrophobic pocket of the active site lined with the side chains of Val121, Val143, Leu 198, and Trp209 (43). The active site of Cam is located at the base of a cleft between the two monomers. An electrostatic surface potential representation (Fig. A-2.4) shows the HCO 3- from the ZnCam-HCO 3- complex appears solvent accessible. Below the active site cavity of Cam, there appears to be a relatively apolar surface (Fig. A-2.4). This surface, composed of the side chains of Leu 83, Phe 132b (where b denotes that the residue belongs to the adjacent monomer), Met 135b, Phe 138, Phe 140, Ile 157, and Val 172, may be analogous to the hydrophobic pocket in HCAII, thus serving as a binding site for CO 2 in the active site. 241 In HCAII, following the trapping of CO 2 by the hydrophobic pocket, CO 2 may interact with the amide nitrogen of Thr 199 (23, 44), replacing the deep water molecule (8). In the Co-Cam-HCO 3- structure, Gln 75b NE and Asn 202b N8 serve as hydrogen- bond donors for the 02 and 03 of the metal-bound HCO 3- (Fig. A-3.3A). However, they do not contact the HCO 3- in the Zn-Cam-HCO3- structure. These two residues are possible candidates for orienting CO 2 for attack or stabilizing a transition state. Since the active site solvent structure is very different between Zn-Cam and Co-Cam, candidates for a deep water molecule are less clear. Although water molecules are missing from each structure upon binding of HCO 3-, (Wat 4 from Co-Cam and Wat 24 from Zn-Cam) these water molecules are not analogous. In HCAII, the first half-reaction culminates in the formation of a bond between the metal-bound OH and the CO 2 to form metal-bound HCO3-, followed by an exchange of the HCO 3- with water (eq. lb). During this transition state, the side chain of Thr 199 helps orient the product. Although an exact analogue of Thr 199 does not exist in Cam, Glu 62 does act to stabilize the binding of ligands to the active site, and thus could partially fulfill a gatekeeper function. Studies of HCAII have previously suggested that bicarbonate may isomerize to form a leaving group (45 ). Multiple binding modes for HCO 3- in the active site of Cam suggest there may be a continuum of binding modes 242 available to the substrate, and perhaps that isomerization could occur before product release. Active site coordination has been postulated to be a primary factor in determining the first committed step of catalysis by HCAII (eqs la and lb) (8, 46). The most obvious structural similarity between the a- and y-classes of carbonic anhydrase is the conserved metal coordination by three histidines. Two of the histidines use the NE nitrogen for metal coordination, while the third uses the No nitrogen (Fig. A-2.3A and A-2.3B). However, as discussed above, solvent-ligation bf the metal center of Cam differs from HCAII with respect to the water ligands. Both native Zn-containing HCAII and Cosubstituted HCAII have distorted tetrahedral geometries, while Zn-Cam is trigonal bipyramidal and Co-Cam is octahedral. The active site coordination and geometry in ZnCam most closely resembles that of Ni-substituted HCAII, while the geometry of CoCam most closely resembles that of Mn-substituted HCAII (25). Both Ni- and Mnsubstituted HCAII are essentially inactive enzymes. It is a fascinating mystery why the Cam catalyzed reaction proceeds at a high rate with an increased number of active site ligands relative to HCAII, and dramatically illustrates the complexity of protein-solventmetal interactions and their implications for reaction mechanisms. 243 In HCAII, the second half- reaction (eqs. 2a and 2b) contains the proton transfer step, which shuttles a proton from the metal-bound water molecule into the solvent, regenerating the metal-hydroxide active site. Residues that potentially act analogously in HCAII and Cam are more clear for this half-reaction. The proton shuttle His 64 of HCAII has two discrete conformations in the crystal structure ( 3 3) . In this ensemble of structures, Glu 84 has three discrete conformations, and in one of those conformations, a proton apparently is shared between the side chains of Glu 62 and Glu 84 (Fig. A-2.3). Glu 62 and Glu 84 are located at the base of the active site in a negatively charged tunnel (Fig. A-2.4) which may aid in the protein-solvent transfer of protons. The direct contact between Glu 84 and Glu 62, combined with the likely protonation of Glu 62 when a ligand is bound to the active site, suggest that the proton shuttle involves both side chains. However, a proton could alternatively be shuttled from the active site to Glu 84 through an intervening water molecule (Wat 67 in Zn-Cam) that provides a hydrogenbond bridge between the active site ligands and the side chain of Glu 84. Based on the finding in this study, as well as previous biochemical evidence, a preliminary reaction mechanism is proposed (Fig. A-2.5) which provides a foundation for future studies. This mechanism is missing aspects that are involved in the first halfreaction as they cannot be determined conclusively with the findings to date. In the proposed mechanism, CO2 binds adjacent to the zinc-bound hydroxyl (representations A 244 and B), the position of which is unknown. The hydroxyl attacks CO 2 leading to a transition state in which the protonated oxygen of the bicarbonate is coordinated to the metal, possibly stabilized by either Gln 75 or Asn 202. These steps are primarily based on analogy to the well-characterized HCAII and a kinetic analysis of Cam suggesting a zinc-hydroxide mechanism. The subsequent steps are based on structures revealed by this study. In the next step, the bicarbonate may isomerize (representations C and D) before replacement by a water molecule (representation E). In the final steps, Glu 62 shuttles a proton from one of the metal-bound water molecules, transferring it to Glu 84 and regenerating the active zinc-bound hydroxyr(representations F and G). Summary and Conclusions. This study, presenting a high-resolution structure for Zn-Cam and the first structure of the Co-substituted Cam, has advanced our structural and functional understanding of the prototype of the novel y-class of carbonic anhydrase. The results identify five and six coordination of the metals in Zn-Cam and Co-Cam, a distinct departure from the active site coordination of the a-class of carbonic anhydrases. The high-resolution structures also identify active site water molecules involved both in metal ligation and the hydrogen-bonding network with adjacent residues that are likely to be catalytically important. The structures presented here have identified multiple positions for Glu 84 consistent with a role for this residue in proton transfer from the zinc-bound water molecule to bulk solvent. 245 The results presented here also raise questions concerning the mechanism of Cam. For example, many of the active site residues (Glu 62, Glu 84, Asn 202) are poorly conserved between Cam and homologs sharing significant sequence similarity with Cam, which is surprising since these appear to be critical for the catalytic mechanism of Cam. However, to date none of these homologous proteins have yet been shown to actually exhibit carbonic anhydrase activity. Additionally, there are several residues near the active site that are well conserved in homologs (Arg 59, Asp 61, Gln 75) where it is unclear how they function in the catalytic mechanism. The results identify residues that potentially function in proton transport; however, residues with the potential to function in CO 2 hydration have yet to be identified. Crystallographic analysis of inhibitor complexes may suggest the location of the CO 2 binding site. Site-directed mutagenesis is expected to identify residues essential for the first half-reaction and the so-called "gatekeeper" function; indeed, this study has suggested Gln 75 and Asn 202 as potential targets for site-specific replacement. 246 ACKNOWLEDGEMENTS We are grateful to C. Fierke, D. Silverman and R. Cheng for critical reading and helpful discussions, B. Tripp for unpublished results and H. Schindelin for experimental assistance. This work is based upon research conducted at the Stanford Synchrotron Radiation Laboratory (SSRL), which is funded by the Department of Energy, Office of Basic Energy Sciences. 247 Fig. A-2.1: Stereoview illustrating the Cam trimer. The overall fold is a left-handed t3- helix, consisting of three untwisted, parallel t3-sheets connected by left-handed crossovers. Figures 1, 2, and 3 were made using MOLSCRIPT (47) , BOBSCRIPT (48) and RASTER3D (49). C C 248 Fig. A-2.2: Active site coordination. Zn-Cam A Models His 122 "' are superimposed onto • the . . .. . '- • 1' His 11 7b -~ Wat2 \'. Wat1 ~ • --,. d ') "' ~ .. contoured at 4a. .. A. '-' His 12~ C . His 117b uncomplexed Wat 3 o) Gin 75b -, ,.• v His 81 -, Glu 62 • 1' . D '"> Wat 2 ' Wat 1 ~ o- e ') Gin 75b _,. • f 1 •• • - His 117b Glu62 His 122 active site ligands and Wat~ - ~ ·- · • His 81 Glu 62 removal of the relevant 1 .~ • $ __.f'• e? Gin 75b t,, '1 His 81 . •. • ~ -.Mb" • , after B His 122 . Wat 1 calculated Co-Cam .• - His 117b • •• • • __ t: • .,. ., e>Gin 75b I> • Glu 62 ., Coordination of Zn- Asn 202b Cam exhibits distorted C His 122 "' trigonal E F His 122 • 1' JI . bipyramidal 1 . ~~·- ~ ~ . Wat 36 geometry. B. IS ~ .p Coordination of Co- • Glu 62 Cam shows distorted octahedral geometry. Glu 62 His 117b sot • . -.') Gin; sb His 81 C. Coordination of Zn-Cam bound to bicarbonate. D . The model of Co-Cam bound to HCO 3 •• E. The model of Zn-Cam bound to SO/ . F. The model of Co-Cam bound to SO/. 249 Fig. A-2.3: A. Asn 202b Stereoview of • Gin 75b ~ \ ,-~ ~ some catalytically important residues. Asn 202b Glu 62 His 81 ,.\.. : il . ,~• '-, · -~ · . ..• . ♦ -- 1 .. • .. .. Glu84 ~ r. ·' ,, His 117b ~ • Glu 62 His 81 Gin 75b ( : .. -. • -l......t~ _. ...... '\ ... . --. .. c... •· His 117b • · • A V Glu 84 •...__ rj 1 His 122 His 122 The active site of Co-Cam . in .. Glu 106 • • 11 ,.. complex with His96 I.I .. - ~ with residues that may be His64 Thr199 • • 11 Thr 199 • His 119 ~ ~~ . Glu 106 • •~ ~ • .. His 94 fl • 1 .. ,~ His 119 •. ~l - . .. B His94 His64 functionally important. The hydrogen bonds between HCO3 • and the candidates for a Thr 199 analog (Gln 75b and Asn 202b) are shown. Glu 84 is shown in three conformations, two of which are alternative conformations in the Zn-Cam structure, while the third conformation is present in the remaining structures. B. Stereoview of the active site of HCAII. The view is shown with the metal-ligating histidines in the same relative orientation as in 3A, i.e., His 119 of HCAII, which coordinates using the No atom, is shown in the same relative location as His 81 of Cam, which also ligates the metal using the No atom. His 64, which is thought to act as the proton shuttle is shown in both conformations. 250 Fig. A-2.4: Stereoview of electrostatic surface potential representation calculated using the program GRASP (50) at an ionic strength of 100 mM and contoured at ±12k8 T (k8 is the Boltzmann constant and T is the absolute temperature). The N-terminus of the protein is at the top of the representation. The surface was made from the Zn-Cam-HC03trimer with the bicarbonate omitted and shows that the activ,e site cavity is weakly negatively charged. The HC0 3- is shown as a stick representation. Directly below the active site cavity (lower right) is a funnel-shaped hydrophobic pocket, which may serve to trap CO2 into the active site. The negatively charged tunnel above the active site contains Glu 62 and Glu 84, and is believed to.be the proton shuttle pathway. The view is the same as in Fig. 1. active site active site hydrophobic pocket hydrophobic pocket 251 Fig. A-2.5 (next page): Proposed reaction mechanism for Cam. The reaction mechanism is drawn using Zn-Cam as the template. Co-Cam should have a similar reaction mechanism with an additional water molecule as an active site ligand. A. Zn2+ is coordinated to one water molecule and one hydroxide ion at the beginning of the first half-reaction (eqs. la and lb). B . Carbon dioxide enters the active site along the hydrophobic pocket. C. Carbon dioxide is attacked by the hydroxide bound to the zinc. D . The bicarbonate may have several stable binding modes. This bidentate binding mode, which requires loss of a metal ligand water molecule, is similar to that seen in the structure of Zn-Cam in complex with HCO 3-. E. The first half-reaction ends with exchange of bicarbonate and a water molecule from the solvent. This state is crystallographically indistinguishable from that in (A), and may be represented by the structures of Zn-Cam or Co-Cam. F. The second half-reaction (eqs. 2a and 2b) begins with the deprotonation of one zinc-bound water molecule, with the proton transferred to Glu 62. During this process, the side chain of Glu 84 swings in so that it may accept the proton. This step is represented by the structure of water-liganded Zn-Cam with the Glu 84 side chain in conformation 1. G. The proton is passed from Glu 62 to Glu 84. This state is represented by the structure of Zn-Cam with the Glu 84 side chain in conformation 2. With the transfer of proton to the solvent, the second half-reaction is complete and state (A) is regenerated. 252 N \,.: ./"" ""1 Gln 75 ◄/► i O ·n2o \ His I 17 ' / OH , , zl'"' N "o 28A •. il2O co • 2 His 122 ✓"'"" B His! 17 '- pl ofi .. - o ,'.::'\o l \';° o 1-i "c ,,,,zln"----- 'o His l 17 His 122 His8 1 I,. ~ co,H,O Gin 75 Gin 75 N\ "~ ~;A " ' \ / O'(i Zn Hi s1 17 / -) Glu 84 O~ I"----- His 122 His Bl - N\ ,.: J"'"~ -----;► H,O\ / (jff Zn Hisl l7 /I"----- His 122 Hi s 81 D ,_.,,zln"----Hi s 122 His 81 1-Iis 81 Glu62 \ - Glu 84 ( Gln75 \ lqq0.pdb 89.1 9.7 0.6 0.6 (1) 0 lqre.pdb 89.8 10.2 0 0 0 Co-Cam Co-Cam+HCOL 82.34 A 82.58 A 20 - 1.76 A 20 - 1.46 A 95,188 128,698 18,449 32,731 1,093 1,697 98.7 (98.7) 99.4 (99.1) 0.054 (0.283) 0.055 (0.297) 21 (4.2) 20 (3.2) 0.181 0.185 0.217 0.206 0.011 A 0.012 A 1.60 1.60 0.12A 0.01 A 6-21 3 5-213 1641 1681 79 118 lqrf.pdb 88.4 11.6 0 0.6(1) 0 Co-Cam+SO1._2 • 82.61 A 20 - 1.55 A 91,423 25,247 1,028 91.6 (90.3) 0.071 (0.276) 15 (3.5) 0.205 0.229 0.014 A 1.70 o.o9A 8-213 N 1629 Vl w 88 Numbers in parenthesis indicate values for the highest resolution bin. Rsym=Lilli-<1>1/Hi where i is the ith measurement and <I> is the weighted mean of I. Rcrysr= LIIF 0 b5' - IF0 a1cll/LIF0 bsl, Rrree is the same as Rcrysr for data omitted from the refinement. Coordinate error is the Cruickshanks value (50) determined in REFMAC (31) and based on the Rrreevalue. lqrl.pdb lqrm.pdb lqrg.pdb PDB Accession 83.4 15.5 1.1 0 89.8 9.7 0.6 0 0 0 90.8 9.2 0 0 0 Ramachandran statistics % Most favored % Allowed % Generous % Disallowed 0 0 Zn-Cam+SO 4 2 • 82.68 A 20 - 1.95 A 51 ,109 13,667 760 97.5 (95.0) 0.081 (0.284) 14 (4.1) 0.189 0.236 0.019 A 1.90 0.18 A 6-213 1601 40 Zn-Cam Zn-Cam+HCOL 82.64 A 82.51 A Unit Cell (a=b=c) 20 - 1.85 A Resolution 20 - 1.72 A 55,426 Observations 128,420 15,750 Unique Reflections 19,889 Free R reflections 1,134 937 Completeness ( %) 98.0 (97.6) 96.9 (82.3) 0.097 (0.240) 0.050 (0.238) Rsym 19(4.9) 16 (4.3) 1/cr 0.183 0.189 Rcryst 0.227 0.236 Rfree RMS~ Bond Lengths 0.021 A 0.021 A 1.80 1.70 RMS~ Bond angles Coordinate Error 0.12A 0.15 A Residues 8-213 0-213 Number of atoms 1674 1617 Water molecules 64 75 Table 1: Data Collection and Refinement Statistics 254 Table 2: Summary of Metal Ligation in Cam ,3 , , I I 1 .. - - , I I I . -2. 1 .. - - ,3 , I .. I ,, , ' 2 '4 '4 I•· Ligation position 2· Zn-Cam uncomplexed Co-Cam uncomplexed Zn-Cam- HC03Co-Cam- HCQ3 Zn-Cam-SO/ Co-Cam-So/ - 1 Wat 1 Wat 1 HC03 - Q3 Watl so42- 04 so42- 04 3 4 Wat2 X X X Wat2 Wat3 HC03 - 0l* X X X Wat2 HC03 01 X X X Wat 36* 2 S04 - 03 so4 2- 03 * denotes a significant shift in the ligation position relative to the uncomplexed form. In the Zn-Cam-HC03 - complex, the HC03 - 01 is displaced 1.5 A relative to the location of the water molecule in Zn-Cam. In the Co-Cam-SO/ complex, water 36 is displaced 0.9 A from the location of Wat 3 in uncomplexed Co-Cam. Of note is that Co-Cam and ZnCam always fill ligation position 1, while ligation position 2 tends to be occupied in ZnCam (resulting in trigonal bipyramidal coordination geometry) and ligation positions 3 and 4 tend to be occupied in Co-Cam (resulting in octahedral geometry). The exception is the Co-Cam-S042- complex, which fills ligation positions 1, 2, and 4, resulting in more distorted coordination geometry. 0 2.14 A 2.13 A 2.15 A 2.23 A 2.06A 2.14A 0 Distance to Zinc Angle x-Zn-W2 71.9 2.31 A 2.soA 2.42 A 2.29A 2.19 A 2.25 A 2.32A Wl W2 W3 His81 His117 His122 Average 0 Distance to Cobalt Atom Angle x-Co-W2 90.9 Table 4: Active Site Geometry of Co-Cam Wl W2 His81 Hisl 17 His122 Average Atom Table 3: Active site geometry of Zn-Cam x-Co-W3 81.2 82.4 x-Zn-81 92.8 121.4 x-Co-81 89.4 169.6 87.4 x-Zn-117 161.4 90.1 118.0 x-Co-117 169.0 84.7 88.2 93.2 x-Zn-122 93 .1 119.1 92.3 100.3 x-Co-122 91.0 82.5 162.8 107.9 98.4 N VI VI Distance to Zinc Angle x-Zn-Bct 01 45.9 x-Zn-117 152.0 110.1 90.3 Distance to Cobalt Wl 2.25 A W2 2.31 A BctOl 2.35 A His81 2.22 A Hisl 17 2.30 A His122 2.19 A Average 2.27 A Bet denotes the HCO 3•. Atom Angle x-Co-W2 98.0 x-Co-Bct 78.3 78.9 x-Co-81 89.3 162.5 87.1 Table 6: Active Site Geometry of Co-Cam Complexed with HCO3 • x-Zn-81 Bet 03 2.27 A 92.8 Bet 01 3.24 A 121.2 His 81 2.18 A His 117 2.30 A His 122 2.05 A Average 2.41 A (2.2 A without Bet O 1) Bet denotes the HCO 3•. AtQm Table 5: Active site geometry of Zn-Cam+ HCQ3• x-CQ-117 172.5 81.8 94.3 89.0 x-Zn-122 94.9 106.2 119.2 107.8 x-Co-122 88.0 87.4 159.0 108.9 99.4 N Vt 0\ 2.40A 2.11 A 2.17 A 2.3oA 2.11 A 2.23A SO 4 04 SO 4 03 His81 His117 His122 Average Angle x-Zn-O3 62.7 x-Zn-81 98.3 117.1 x-Zn-117 163.3 100.7 90.1 2.29A 2.32A 2.09A 2.14 A 2.31 A 2.13A 2.21 A WI SO 4 03 SO 4 04 His81 His117 His122 Average 0 0 0 Distance to Cobalt Atom Angle x-Co-O3 55.5 x-Co-O4 91.7 64.2 x-Co-81 153.0 103.8 93 .6 Table 8: Active Site Geometry of Co-Cam Complexed with SO/" Distance to Zinc Atom Table 7: Active Site Geometry of Zn-Cam Complexed with SO/ x-Co-117 76.5 94.2 158.3 89.3 x-Zn-122 85.9 122.9 113.4 104.1 x-Co-122 92.4 141.4 100.2 112.7 98.4 N Vt -....) 258 REFERENCES 1. Meldrum, H., and Roughton, F. (1933) J. Physiol. 80, 113-42. 2. Hewett-Emmett, D., and Tashian, R. (1996) Mal. Phylogenet. Evol. 5, 50-77. 3. Soltes-Rak, E., Mulligan, M. E., and Coleman, J. R. (1997) J. Bacterial. 179, 769-774. 4. Alber, B. E. , and Ferry, J. G. (1994) Proc. Natl. Acad. Sci. USA 91, 6909-6913. 5. Kisker, C., Schindelin, H., Alber, B. E., Ferry, J. G., and Rees, D. (1996) EMBO J. 15, 2323-2330. 6. Kannan, K . K., Notstrand, B., Fridborg, K.; Lovgren, S., Ohlsson, A., and Petef, M. (1974) Proc. Natl. Acad. Sci. USA 72, 51-55. 7. Kannan, K., Ramanadham, M., and Jones, T. (1984) Ann. NY Acad. Sci. 429, 49-60. 8. Hakansson, K., Carlsson, M. , Svensson, L.A., and Liljas, A. (1992) J. Mo!. Biol. 227, 1192-1204. 9. Eriksson, A. E., and Liljas, A. (1993) Proteins 16, 29-42. 10. Boriack-Sjodin, P.A. , Heck, R. W., Laipis, P. J., Silverman, D. N., and Christianson, D. W. (1995) Proc. Natl. Acad. Sci. USA 92, 10949-10953. 11. Starns, T., Nair, S. K., Okuyama, T., Waheed, A., Sly, W. S., and Christianson, D. W. (1996) Proc. Natl. Acad. Sci USA 93, 13589-13594. 12. Alber, B. E., and Ferry, J. G. (1996) J. Bacterial. 178, 3270-3274. 259 13. Alber, B. E., Colangelo, C. M. , Dong, J., Stalhandske, C. M. V., Baird, T. T., Tu, C., Fierke, C. A. , Silverman, D. N. , Scott, R. A., and Ferry, J. G. (1999) Biochemistry 38, 13119-13128. 14. Silverman, D. N., and Lindskog, S. (1988) Acc. Chem. Res. 21, 30-36. 15. Khalifah, R. G. (1971) J. Biol. Chem. 246, 2561-2573. 16. Steiner, H., Jonsson, B. H., and Lindskog, S. (1975) Eur. J. Biochem. 59, 253-259. 17. Christianson, D. W., and Fierke, C. A. (1996) Acc. Chem. Res. 29, 331-339. 18. Sly, W. S., and Hu, P. Y. (1995) Annu. Rev. Biochem. 64, 375-401. 19. Taoka, S. , Tu, C. K., Kistler, K. A., and Silverman, D. N. (1994) J. Biol. Chem. 269, 17988-17992. 20. Tu, C. K., Silverman, D. N., Forsman, C., Jonsson, B. H., and Lindskog, S. (1989) Biochemistry 28, 7913-7918. 21. Xue, Y., Liljas, A., Jonsson, B.-H., and Lindskog, S. (1993) PROTEINS: Struct. Funct. Genet. 17, 93-106. 22. Merz, K. M. (1990) J. Mal. Biol. 214, 799-802. 23. Lindahl, M., Svensson, L. A., and Liljas, A. (1993) Proteins 15, 177-182. 24. Hunt, J., Rhee, M., and Storm, C. (1977) Anal. Biochem. 79, 614-617. 25. Hakansson, K., Wehnert, A., and Liljas, A. (1994) Acta Cryst. D50, 93-100. 26. Jurnak, F. (1986) J. Cryst. Growth 76, 577-82. 260 27 . Otwinowski, Z. (1993) in CCP4 Study Weekend Data Collection and Processing (Sawyer, L., Isaacs, N., and Bailey, S., Eds.) pp 56-62, SERC Daresbury Laboratory, UK. 28. Navaza, J. (1994) Acta Cryst. A50, 157-163. 29. Jones, T. A. , Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Cryst. A47, 110-119. 30. Bailey, S. (1994) Acta Cryst. D50, 760-763. 31. Murshudov, G. N., Yagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr. D53, 240-255. 32. Brtinger, A. T. (1992) X-PLOR version 3.1 -·A system for X-ray crystallography and NMR , Yale University Press, New Haven and London. 33. Nair, S. K. , and Christianson, D. W. (1991) J. Am. Chem. Soc. 113, 406-411. 34. Tripp, B., and Ferry, J. (submitted) Biochemistry. 35 . Jenkins, J., Janin, J., Rey, F. , Chiadmi, M., vanTilbeurgh, H., Lasters, I., DeMaeyer, M., VanBelle, D., Wodak, S. J., Lauwereys, M., Stanssens, P., Mrabet, N. T., Snauwaert, J., Matthysens, G., and Lambeir, A. (1992) Biochemistry 31, 5449-5458 . 36. Lloyd, L. F., Gallay, 0. S., Akins, J., and Zeikus, J. G. (1994) J. Mal. Biol. 240, 504506. 37. Holland, D.R., Hausrath, A. C., Juers, D., and Matthews, B. W. (1995) Protein Sci. 4, 1955-1965. 261 38. Loll, P. J., Quirk, S., Lattman, E. E., and Garavito, R. M. (1995) Biochemistry 34, 4216-4324. 39. Mancia, F., Keep, N. H. , Nakagawa, A., Leadlay, P. F. , McSweeney, S., Rasmussen, B., Bosecke, P., Diat, 0., and Evans, P. R. (1996) Structure 4, 339-350. 40. Jonsson, B. M. , Hakansson, K., and Liljas, A. (1993) FEBS Letters 322, 186-190. 41. Lindskog, S., and Liljas, A. (1993) Curr. Op. Struct. Biol. 3, 915-920. 42. Merz, K. M., and Banci, L. (1997) J. Am. Chem. Soc. 119, 863-871. 43 . Lindskog, S. (1997) Pharmacol. Thera. 74, 1-20. 44. Liljas, A., Hakansson, K. , Jonsson, B. H. , arid Xue, Y. (1994) Eur. J. Biochem. 219, 1-10. 45. Liang, J. Y., and Lipscomb, W. N. (1987) Biochemistry 26, 5293-5301. 46. Garmer, D., and Krauss, M. (1992) J. Am. Chem Soc. 114, 6487-6493 . 47. Bacon, D., and Anderson, W. F. (1988) J. Mal. Graph. 6, 219-220. 48. Esnouf, R. (1997) J. Mal. Graph. 15, 133-138. 49. Merritt, E. A., and Murphy, M. E. P. (1994) Acta Cryst. D50, 869-873. 50. Sharp, K., Fine, R., and Honig, B. (1987) Science 236, 1460-1463. 51. Cruickshank, D. W. J. (1999) Acta Cryst. D55, 583-601. 262 A- 3. Crystallographic studies of they-class carbonic anhydrase frorn Methanosarcina thermoplila in complex with inhibitors t Tina M . Iverson+, Anthony Gianetti+, Jessica Wuu\ Birgit E. Alber®,U, J. Greg Ferry®, and Douglas C. Rees§ +Graduate Option in Biochemistry, 147-75, California Institute of Technology, Pasadena, CA 91125. § MC 147-75 CH, Howard Hughes Medical Institute and Division of Chemistry, California Institute of Technology, Pasadena, CA 91125 @ Department of Biochemistry and Molecular Biology, Eberly College of Science, The Pennsylvania State University, University Park, PA 16802-4500 II Present address: Institut flir Biologie II, Mikrobiologie, Albert-Ludwigs-Universitat Freiburg, Schanzlestr. 1, D-79104 Freiburg, Germany t This work was supported by a grant from the National Institutes of Health to J.G.F. (GM44661). RUNNING TITLE: Inhibitor complexes of y-carbonic anhydrase 263 Abbreviations: CAM, Methanosarcina thermophila carbonic anhydrase; Zn-Cam, Cam that has been heterologously produced in Escherichia coli, metal-depleted and reconstituted with zinc; Co-Cam, Cam that has been heterologously produced in Escherichia coli, metal-depleted and reconstituted with cobalt; HCAII, human carbonic anhydrase II; EXAFS, extended x-ray absorption fine structure. 264 ABSTRACT: Carbonic anhydrases catalyze the interconversion of CO 2 and HC03- by a mechanism involving two half-reactions: conversion of the substrate to product followed by the regeneration of the active form of the enzyme by a proton shuttle mechanism. Kinetic analyses and high-resolution crystal structures of the y-class carbonic anhydrase have suggested the proton shuttle mechanism for that enzyme. To explore the CO 2 hydration step of the reaction, the three-dimensional crystal structures of Co-substituted y-class carbonic anhydrase in complex with the inhibitors acetazolamide and cyanide have been solved at 1.68 A and 1.75 A resolution respectively and are compared to similar enzyme-inhibitor complexes of a-class carbonic anhydrases. The ring of the acetazolamide binds over the metal, removing all metal-bound water molecule ligands. In contrast, acetazolamide has been shown to bind to human carbonic anhydrase II (HCAII), an a-class carbonic anhydrase through contacts of the sulphonamide group. Similarly, cyanide appears to bind directly to the active site metal, a distinct departure from the binding to the HCAII. Although diffraction data has been collected from crystals of Cam that have been exposed to the substrate CO2 , there does not appear to be density for the substrate in the maps at the current state of refinement. 265 Carbonic anhydrases (CAs) are zinc-containing enzymes noted for their ability to interconvert CO 2 and HCO 3- with a k ca 1 approaching 106 per second ( 1 ). Sequence and structural analyses have indicated the existence of three distinct classes of CA: a- p- and y- (2-5) . The a -class of CA can be exemplified by Human Carbonic Anhydrase II (HCAII) which is the best-studied isozyme of this class and has the highest activity. Turnover of the reaction is accomplished by a "zinc-hydroxide" mechanism. In HCAII, the overall enzyme-catalyzed reaction occurs in two distinct half-reactions. The first half-reaction involves the interconversion of CO2 and HCO 3- (eqs la and lb, where E signifies the enzyme) and involves nucleophilic attack of the zinc hydroxyl ion on CO2 . This is followed by exchange of zinc-bound HCO 3- with water. The second half reaction corresponds to the rate-determining intramolecular and intermolecular proton transfer steps (eqs 2a and 2b), which regenerates the zinc-hydroxide at the active site. E-Zn2+-Off + CO 2 H E-Zn2+-HCO 3- (la) E-Zn2+-Hco3- + H 2O H E-Zn 2+-H 2O + HCO 3- (lb) E-Zn 2+-H 2O H +H-E-Zn2+-Off (2a) +H-E-Zn 2+-Off + B H E-Zn2+-Off + BH+ (2b) 266 The prototype of the y-class of carbonic anhydrase is the carbonic anhydrase from Methanosarcina thermophila (Carn). Kinetic studies suggest that the hydration of CO 2 by Carn may occur by a mechanism similar to that proposed for human carbonic anhydrase II (HCAII) (6, 7). Indeed, crystallographic (8) and biochemical (9) studies have suggested that the proton shuttle mechanism catalyzed by Carn in the second halfreaction (eqs. 2a and 2b) is quite similar in nature to the proton shuttle mechanism catalyzed by HCAII. However, the biochemical behavior of the two enzymes is not identical. One distinct difference involves the coordination number of the active site metal. HCAII contains a four-coordinate zinc ion in distorted tetrahedral geometry, while Carn shows a five-coordinate zinc ion in trigonal bipyrarnidal geometry. In HCAII, substitution of the active site zinc to almost any other metal ion knocks out enzyme activity. Indeed only cobalt-substituted HCAII retains any activity, with a rate of ~50% of the native enzyme (10). Crystal structures of metal-substituted HCAII show that the presence of Co, Cu, Ni, or Mn have an increased number of active site ligands (11 ), leading to the postulation that a higher coordination number of the metal ion decreases the turnover rate. In contrast to HCAII, substitution of the zinc ion by cobalt doubles Carn activity (7), even though 267 EXAFS analysis indicates a higher coordination number for cobalt-substituted Cam (CoCam) relative to Zn-containing Cam (Zn-Cam). Currently, there is not any structural or biochemical information that reliably suggests the mechanism of CO 2 hydration (eqs. la and lb) . One of the great advances in understanding the substrate-product interconversion catalyzed by HCAII came through the crystallographic study of enzyme-inhibitor complexes. Most of the inhibitors, including sulphonamide type inhibitors (11-15 ), bisulfite, and formate (3), are bound directly to the active site zinc, thus explaining their inhibitory mechanism. However, cyanide and cyanate bound instead to the backbone amide nitrogen of Thr 199 (16) , suggesting a functional role for this amide nitrogen. It has been proposed that cyanide and cyanate act as substrate analogs in HCAII. The current proposed reaction mechanism includes Thrl 99 backbone amide nitrogen acting to polarize the substrate CO 2 , preparing it for attack by the active site Zn-OH. Additionally, complexes of HCAII and the non-protonated anion inhibitors bromide and azide (17) show that these inhibitors replace the "deep" water molecule that is required for catalysis by HCAII. These inhibitor studies have suggested catalytically important structural features that are independent of the chemical properties of the protein side chains, thus 268 site directed mutagenesis could not have been used to identify the functional aspects in these cases. In order to learn more regarding the CO 2 hydration step in Cam, two enzymeinhibitor complexes have been crystallized. The binding of the inhibitors show a distinct departure from their interaction with HCAII, suggesting that the mechanism of CO 2 hydration may differ between Cam and HCAII, despite the utilization of a similar proton shuttle mechanism by both enzymes. Additionally, Co-substituted crystals of Cam have been exposed to a CO 2 atmosphere prior to cryo::.cooling in hopes of trapping an enzymesubstrate complex. EXPERIMENTAL PROCEDEURES Crystallization . "Cam" refers to native carbonic anhydrase isolated from Methanosarcina thermophila (6) . "Co-Cam" designate enzymes in which the indicated metal was incorporated into the active site of as-purified enzyme produced in E. coli, by preparation of the apoenzyme using the denaturant guanidine-HCl and the metal chelator dipicolinate followed by reconstitution in the presence of C0Cl 2 as previously described (7). For crystallizations, the stock protein concentration was 10 mg/mL buffered in 5 269 mM phosphate pH 7 .0. Crystallization was performed in conditions previously described with the addition of 4 mM acetazolamide or 3 mM cyanide to the protein prior to crystallization. The crystals appeared within 3 days and grew to a maximal size of 0.5 mm x 0.5 mm x 0.5 mm. Cubic crystals belonged to the space group P2 13 with unit cell dimensions between a=82.3 A and a=82.9 A and one monomer per asymmetric unit. Prior to cryo-cooling, crystals were soaked in a solution containing all of the crystallization components plus 25% glycerol. CO 2 containing crystals were prepared by growing Co-Cam crystals as previously with no inhibitor present, and using a gas pressurizing device (18, 19) to subject the crystal to 0.5 MPa of CO 2 for 30 seconds prior to cryo-cooling. Longer exposures to the gas or higher pressures resulted in and poor diffraction, possibly as a result of internal pH change in the crystal during reaction turnover. Data for the inhibitor complexes were collected on an R-axis Ile image plate detector using x-rays at a wavelength of 1.54 A generated by a RU-200 rotating anode. Data for the CO2 treated crystals were collected at SSRL beamline 9-1 at a wavelength of 0.98 Aon a MAR 345 image plate detector. Data collection. Data for Co-Cam complexed with inhibitors were collected at 113 K using CuKcx radiation from an RU 200 rotating anode at a wavelength of 1.54 Aon an Raxis Ile image plate. Data for the CO 2 treated crystals were collected at 93 K at the 270 Stanford Synchrotron Radiation Laboratory beamline 9-1 at a wavelength of 0.98 A using a MAR Research image plate detector. All data were processed using DENZO and scaled with SCALEPACK (20) . The crystals of the inhibitor complexes were isomorphous with the cubic crystals described previously; therefore, the refined model of Co-Cam ( 8) with all water molecules removed was used as a starting model in each case. Model Building and Refinement. All Cam models were built using the program 0 (21). Refinement was carried out using REFMAC (22, 23) and X-PLOR (24). The reflections for Rfree for the data sets are identical to those from the Co-Cam data set described elsewhere (8) with additional randomly selected reflections selected for data sets of higher resolution. Data collection and refinement statistics are given in Table A3.1. RESULTS Co-Cam crystallized with inhibitors or exposed to the substrate CO2 diffracted between 1.68 A and 1.76 A resolution, depending on the crystal. Data collection and 271 current refinement statistics are listed in Table A-3.1. Each of these structures is still in the process of refinement. In both the Co-Cam inhibitor complexes, the inhibitor appears to be bound directly to the active site. Density for each inhibitor at the active site is shown in Fig. A3.1. The position of the inhibitor at the active site was particularly disappointing in the case of the Co-Cam-CN complex, since crystal structures of HCAII with this inhibitor highlighted residues involved in CO 2 binding. Co-Cam was also subjected to pressurization with the substrate CO 2 . This represents the first known attempt to crystallize a gaseous substrate with an enzyme. However, the active site of this enzyme appears similar to the native state and does not appear to contain density that could be interpreted as a CO 2 molecule. Neither can density suggestive of bound CO 2 be found in other areas of the maps. It is possible that the binding affinity for the substrate is weak enough that the CO2 could not be trapped at the active site. One additional explanation is that the crystals were not exposed to the substrate long enough to have a high occupancy of the CO 2 binding site. 272 DISCUSSION Although the structures are not yet fully refined, preliminary analysis suggests that Cam binds to inhibitors in a different manner than was seen for HCAII. This suggests differences may exist in the mechanism CO 2 hydration between the a-class and y-class carbonic anhydrases, however, neither a mechanism for CO2 hydration, nor a site for CO 2 binding by Cam can yet be proposed. CONCLUSIONS The structures of Co-Cam in complex with known carbonic anhydrase inhibitors suggest that the CO2 hydration half-reaction proceeds by a different mechanism than seen for a-class carbonic anhydrases, despite similar kinetic characteristics. Further work will focus on different inhibitor-complexes that may elucidate functional residues and water molecules, as well as exposing inhibitor-inactivated Cam crystals to CO2 , potentially exposing a CO2-bound intermediate state of the enzyme. 273 ACKNOWLEDGEMENTS This work is based upon research conducted at the Stanford Synchrotron Radiation Laboratory (SSRL), which is funded by the Department of Energy, Office of Basic Energy Sciences. 274 Figure A-3.1 Binding of inhibitors to the A active site of Cam. Contours in blue represent 12F0 I - IFcl denzity while contours in green show IF0 I - IFcl difference density. A. Acetazolamide co-crystals. B. CN- cocrystals. C. CO2 exposed crystals. Acetazolamide has been placed into this density for that structure, while the other B two structures, have only active site water ligands added. C 275 Table A-3.1: Data collection and refinement statistics 0 Unit cell (A) Resolution Total reflections Unique reflections Completeness ( % ) ~ym Residues Number of atoms Water Molecules Inhibitor added Rcryst Rrree Co-Cam+AZM 82.9 1.68 A 101064 20694 94.4 (82.9) 0.059 (0.158) 8-213 1611 56 no 0.202 0.236 Co-Cam+CN 82.6 1.75 A 60987 18546 96.1 (86.3) 0.064 (0.217) 8-213 1558 17 no 0.221 0.260 Co-Cam+ CO1 82.3 1.76A 46086 18407 98.2 (98.2) 0.067 (0.253) 9-213 67 no 0.190 0.227 276 REFERENCES 1. Meldrum, H., and Roughton, F. (1933) J. Physiol. 80, 113-42. 2. Hewett-Emmett, D., and Tashian, R. (1996) Mol. Phylogenet. Evol. 5, 50-77. 3. Hakansson, K. , Carlsson, M. , Svensson, L.A., and Liljas, A. (1992) J. Mol. Biol. 227, 1192-1204. 4. Kisker, C., Schindelin, H., Alber, B. E., Ferry, J. G., and Rees, D. (1996) EMBO J. 15, 2323-2330. 5. Mitsuhasi, S. , Mizushima, T ., Yamashita, E., Yamamoto, M., Kumasaka, T., Moriyana, H., Ueki, T., Miyachi, S., and Tsukihara, T. (2000) J. Biol. Chem. 275, 55215526. 6. Alber, B. E., and Ferry, J. G. (1996) J. Bacterial. 178, 3270-3274. 7. Alber, B. E., Colangelo, C. M., Dong, J., Stalhandske, C. M. V., Baird, T. T., Tu, C., Fierke, C. A., Silverman, D. N., Scott, R. A., and Ferry, J. G. (1999) Biochemistry 38, 13119-13128. 8. Iverson, T., Alber, B., Kisker, C., Ferry, J., and Rees, D. (submitted) Biochemistry. 9. Tripp, B., and Ferry, J. (submitted) Biochemistry. 10. Hunt, J., Rhee, M., and Storm, C. (1977) Anal. Biochem. 79, 614-617. 11. Hakansson, K., and Liljas, A. (1994) FEBS Lett. 350, 319-322. 277 12. Baldwin, J., Ponticello, G. , Anderson , P., Christy, M. , Murcko, M., Randall, W., Schwam, H., Sugrue, M., Springer, J. , Gautheron, P., Grove, J., Mallorga, P., Viader, M.P., McKeever, B., and Navia, M. (1989) J. Med. Chem. 32, 2510-? 13. Vidgren, J., Liljas, A., and Walker, N. (1990) Int. J. Biol. Macromol. 12, 342-? 14. Eriksson, A. E., and Liljas, A. (1993) Proteins 16, 29-42. 15. Vidgren, J., Svensson, A. , and Liljas, A. (1993) Int. J. Biol. Macromol. 15, 97-100. 16. Lindahl, M., Svensson, L.A. , and Liljas, A. (1993) Proteins 15, 177-182. 17. Jonsson, B. M., Hakansson, K. , and Liljas, A. (1993) FEES Letters 322, 186-190. 18. Stowell, M. H. B., Soltis, S. M., Kisker, C.;Peters, J. W., Schindelin, H., Rees, D. C., Cascio, D., Beamer, L., Hart, P. J., Wiemer, M. C., and Whitby, F. G. (1996) J. Appl. Cryst. 29, 608-613 . 19. Soltis, S. M., Stowell, M. H. B. , Wiener, M. C., Phillips, G. N., and Rees, D. C. (1997) J. Appl. Cryst. 30, 190-194. 20. Otwinowski, Z. (1993) in CCP4 Study Weekend Data Collection and Processing (Sawyer, L., Isaacs, N., and Bailey, S., Eds.) pp 56-62, SERC Daresbury Laboratory, UK. 21. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Cryst. A47, 110-119. 22. Bailey, S. (1994) Acta Cryst. D50, 760-763 . 278 23 . Murshudov, G. N., Yagin, A. A. , and Dodson, E. J. (1997) Acta Crystallogr. D53, 240-255. 24. Brlinger, A. T. (1992) X-PLOR version 3.1 - A system for X-ray crystallography and NMR, Yale University Press, New Haven and London. 279 A-4. Crystallization and preliminary x-ray analysis of a Pclass carbonic anhydrase from M ethanobacterium thermoautotrophicum ~Ht Pavel Stropt, Kerry S. Smith\ Tina M. Iversont, Jessica Wuu@, J. Greg Ferry§, Douglas C. Rees@ :j: Graduate Option in Biochemistry, 147-75 , California Institute of Technology, Pasadena, CA 91125 . § Department of Biochemistry and Molecular Biology, Eberly College of Science, The Pennsylvania State University, University Park, PA 16802-4500 @ MC 147-75 CH, Howard Hughes Medical Institute and Division of Chemistry, California Institute of Technology, Pasadena, CA 91125 t This work was supported by a grant from the National Institutes of Health to J.G.F. (GM44661). P.S. is supported by a fellowship from the National Science Foundation. RUNNING TITLE: Crystallization of ~-carbonic anhydrase 280 ABSTRACT Carbonic anhydrases are-zinc containing enzymes that catalyze the interconversion of CO2 and HCO 3- . Carbonic anhydrases are found ubiquitously throughout nature and currently three distinct classes are recognized: a- ~- and y-. The ~class carbonic anhydrases are present in higher plants, algae, and prokaryotes and are involved in inorganic carbon utilization as well as the Calvin-Benson cycle. Crystals of the ~-class carbonic anhydrase from Methanobacterium thermoautotrophicum 11H belonging to the primitive orthorhombic space group P2 12 12 1 and having unit cell constants of a=54.6 A b=l 13.8 A c=156.8 A have been obtained using ethanol and methyl-pentanediol as precipitants. A complete data set to 2.7 A resolution has been collected using CuKa radiation. A Matthew's calculation is consistent with the presence of four molecules in the asymmetric unit, while a rotation function of these data suggest 222 point group symmetry. 281 Abbreviations : CA, carbonic anhydrase; Cam, y-class CA from Methanisarcina thermophila; Cab, p-class CA from Methanobacterium thermoautotrophicum !lH; MPD, methyl pentane diol. 282 Carbonic anhydrases (CAs) are zinc-containing enzymes found in all phyla ( 1) that were first noted for their physiological role in interconverting CO2 and HCO3- (2) . Based on sequence and structure analyses, three classes of carbonic anhydrases are currently recognized: a-, P-, and y-. The a-class includes seven mammalian isozymes and two isozymes from Chlamydomonas reinhardtii (3, 4 ), and has been characterized both biochemically and structurally. Analyses of the crystal structures (5-9) and biochemical evidence have suggested a detailed reaction mechanism for this enzyme. The structure of Cam, the prototype of the y-class of CAs has been solved (10 ) , and unlike the eukaryotic counterparts was found to be a trimer and have a unique left handed P-helix overall fold . Recent kinetic analyses ( 11, 12), combined with highresolution crystal structures ( 13 ), suggest that Cam utilizes a reaction mechanism similar to that of the a-class CAs. p-CAs are present in higher plants, algae, and prokaryotes. Sequence and EXAFS ( 14, 15) analyses of p-class CAs suggest that the ligands to the active site zinc include two cysteine thiols and one histidine, as well as a fourth N/O ligand that was presumed to be a catalytically active water molecule. P-class CAs typically catalyze CO2 hydration with kcat values between 105 and 106 s- 1, similar to the rates observed for both the a- and y- 283 classes of carbonic anhydrase. The first known ~-CA in the archaea was purified from the methanogenic archaeon Methanobacte rium thermoautotrophicum L'lH ( 16) and exhibits 34% sequence identity to the ~-class carbonic anhydrase from Escherichia coli. The calculated molecular weight of this enzyme, designated Cab, is 20 kD. Based upon native gel-filtration experiments suggesting a total molecular weight of ~90 kD it is likely that Cab exists as a tetramer. The enzyme has CO2 hydration activity with a kcat ofl.7 X 104, similar to the values seen for the other classes of carbonic anhydrases. Recent crystallographic studies of the· ~-CA from the red algae Porphyridium purpureum ( 17) have proved fruitful. The gene encoding the P. purpureum ~-CA has undergone fusion so that it contains two nearly identical repeats, each with significant identity to a consensus ~-CA sequence. The structure of the P. purpureum ~-CA shows a dimer with pseudo 2 2 2 point group symmetry, thus an enzyme composed of monomers would form a tetramer. The active site shows zinc ligation by two cystiene ligands, one histidine ligand, and one aspartic acid, suggesting a departure from the catalytic mechanism proposed for the other two classes of CA, which require a catalytic water molecule bound to the active site. Crystals of Cab have been grown in the primitive orthorhombic space group P2 12 12 1• Crystallographic analysis of these crystals suggests the molecular structure is consistent with the structure of P. purpureum enzyme. A 284 molecular replacement solution should provide interesting insights by comparison of the ~-CA structure from two different organisms. EXPERIMENTALPROCEDEURES Crystallization. Cab has been overexpressed in E. coli and purified by heat denaturation combined with ion exchange as previously described ( 16) . Crystals of the ~-CA from Methanobacterium thermoautotrophicum ~H have been obtained from conditions containing 35% ethanol, 12% methyl pentanediol (MPD) and 50 mM Calcium acetate buffered in 100 mM HEPES pH 7.5 at 22 °C (Fig. A-4.1). Crystals form in the primitive orthorhombic space group P2 12 12 1 and appear well ordered, diffracting to 2.7 A resolution using Cu Ka generated by a Rigaku RHU3RHB rotating anode and collected on an Raxis II image plate (Fig. A-4.2). The diffraction may be limited by crystal size (0.05 X 0.05 X 0.1 mm). Attempts have been made to grow larger crystals using a variety of seeding techniques, however, these techniques have not substantially improved the crystals in a reproducible manner. 285 Data collection Cryo cooling of the crystals involved screening of a large variety of potential cryo-protectants including glycerol ethylene glycol, glucose, and low molecular weight polyethylene glycol compounds. The best cryo-protection was achieved by substituting all of the water volume of the crystallization conditions with either ethanol or MPD resulting in concentrations of -50% ethanol or -35% MPD, and the resultant diffraction for the crystals in each type of cryoprotectant was approximately equivalent. As small volumes of solutions containing low-boiling point compounds tend to evaporate quickly, crystals were routinely cryo-cooled using 35% MPD to promote ease of handling. RESULTS AND DISCUSSION SDS-PAGE analysis of carefully washed and dissolved crystals shown that Cab is the exclusive constituent. A complete data set merging to 2.7 A resolution has been collected on Raxis IV image plates using a RHU3RHB rotating anode. A summary of the data collection and processing statistics is shown in Table A-4.1. The data were processed using DENZO and scaled using SCALEP ACK (18). Systematic absences combined with lattice spacings suggest that the crystals belong to the primitive orthorhombic space group P2 12 12 1 with unit cell constants a=54.6 A b=l 13.8 A c=156.8 A. According to the cell size and symmetry, calculation of the solvent content suggests four molecules are present in the asymmetric unit, with a corresponding Matthew's 286 coefficient of 2. 7 ( 19) . To search for the non-crystallographic axes, a self-rotation function (Fig. A.4.1) was calculated using POLARRFN (20) . The rotation function (Fig. A-4.3) indicates a 2-fold axis along x and 2-fold axes in the yz plane 45° from the z-axis. Combined with a 16 cr packing peak in the native Patterson at x,y,z=(0,0,0.36), analysis of the native data indicate the asymmetric unit probably contains a tetramer with 2 2 2 point group symmetry, consistent with the pseudo 2 2 2 symmetry seen in the ~-CA from P. purpureum. CONCLUSIONS Preliminary x-ray analysis of Cab crystals suggests that the enzyme forms a tetramer. Molecular replacement of the data with the model from P. purpureum will allow a comparison of the two enzymes that may provide further insight into catalysis. 287 Fig. A-4.1. Crystals of M. thermoautotrophicum AH ~-carbonic anhydrase (Cab). Fig. A-4.2. Typical diffraction pattern of a crystal of Cab. The edge of the image plate corresponds to 2.7 A resolution. . . 288 Fig. A-4.3. CAB self rotation function . The left panel show the K=180° section. The peaks at (<j>,w,K) = (±90°, ±45°, 180°) indicate a 2-fold in the yz plane 45° from the z-axis. Remaining peaks are origin peaks. The right panel shows the K=90° section. The peak at <1>=0, w=O is consistent with a 2-fold along the x-axis. ~ 90.0 - -·----- 1' )» /1/1/1 1/1( - -----------~✓' -90.0 y cab self-rotation R=20A, 4 - 50 A Se cti on kappa = 90 289 Table A-4.1 Data collection and processing statistics Crystal Data P2 12 12 1 Space group Unit cell dimensions a=54.6, b=l 13.8, c=156.8 Data collection 20 - 2.7 Resolution (A) 2.8 - 2.7 Last resolution shell (A) 96% (83%) Completeness 206,646 Number of reflections Number of unique reflections 27,526 Redundancy 7 .5 Rsym 0.086 (0.322) I/cr 15 (3.4) Numbers in parentheses indicate values for the last resolution shell. 290 REFERENCES 1. Smith, K. S., Jakubzick, C., Whittam, T. S., and Ferry, J. G. (1999) Proc. Nat. Acad. Sci. USA 96, 15184-15189. 2. Meldrum, H. , and Roughton, F. (1933) J. Physiol. 80, 113-42. 3. Hewett-Emmett, D., and Tashian, R. (1996) Mol. Phylogenet. Evol. 5, 50-77. 4. Soltes-Rak, E. , Mulligan, M. E., and Coleman, J. R. (1997) J. Bacterial. 179, 769-774. 5. Boriack-Sjodin, P. A. , Heck, R. W., Laipis, P. J., Silverman, D. N., and Christianson, D. W. (1995) Proc. Natl. Acad. Sci. USA 92 , 10949-10953. 6. Eriksson, A. E., and Liljas, A. (1993) Proteins 16, 29-42. 7. Hakansson, K. , Carlsson, M., Svensson, L.A. , and Liljas, A. (1992) J. Mol. Biol. 227, 1192-1204. 8. Kannan, K. , Ramanadham, M. , and Jones, T. (1984) Ann. NY Acad. Sci. 429, 49-60. 9. Starns, T., Nair, S. K., Okuyama, T., Waheed, A., Sly, W. S., and Christianson, D. W. (1996) Proc. Natl. Acad. Sci USA 93, 13589-13594. 10. Kisker, C., Schindelin, H., Alber, B. E., Ferry, J. G., and Rees, D. (1996) EMBO J. 15, 2323-2330. 11. Alber, B. E., and Ferry, J. G. (1996) J. Bacterial. 178, 3270-3274. 12. Alber, B. E., Colangelo, C. M., Dong, J., Stalhandske, C. M. V., Baird, T. T., Tu, C., Fierke, C. A., Silverman, D. N., Scott, R. A., and Ferry, J. G. (1999) Biochemistry 38, 13119-13128. 29 1 13. Iverson, T., Alber, B., Kisker, C., Ferry, J., and Rees, D. (submitted) Biochemistry. 14. Bracey, M. H. , Christiansen, J., Tovar, P., Cramer, S. P., and Bartlett, S. G. (1 994) Biochemistry 33, 13126-13131. 15 . Rowlett, R. S. , Chance, M. R., Wirt, M. D., Sidelinger, D. E., Royal , J. R. , Woodroffe, M. , Wang, Y. F., Saha, R. P., and Lam, M. G. (1994) Biochemistry 33, 13967-13976. 16. Smith, K., and Ferry, J. (1999) J. Bacterial. 181, 6247-6253 . 17 . Mitsuhasi, S., Mizushima, T., Yamashita, E. , Yamamoto, M. , Kumasaka, T., Moriyana, H. , Ueki, T., Miyachi, S., and Tsukihara, T. (2000) J. Biol. Chem. 275, 55215526. 18. Otwinowski, Z. (1993) in CCP4 Study Weekend Data Collection and Processing (Sawyer, L., Isaacs, N., and Bailey, S., Eds.) pp 56-62, SERC Daresbury Laboratory, UK. 19. Matthews, B. W. (1968) J. Mol. Biol. 33, 491-497. 20. Bailey, S. (1994)Acta Cryst. D50, 760-763.