¹⁵N Nuclear Magnetic Resonance Studies of the Active Sites of Enzymes and In Vivo Nitrogen Metabolism Citation Kanamori, Keiko (1981) ¹⁵N Nuclear Magnetic Resonance Studies of the Active Sites of Enzymes and In Vivo Nitrogen Metabolism. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/ryxe-9852. https://resolver.caltech.edu/CaltechTHESIS:09182025-000702960 Abstract To determine whether the catalytic zinc of carboxypeptidase A (CPA) is complexed to the active-site tyrosine (Tyr-248) in solution, a 15 N probe, an arsanilazo group enriched in 15 N at both azo nitrogens, has been selectively coupled to Tyr-248 with retention of full catalytic activity, and the resulting arsanilazotyrosyl-248 carboxypeptidase A (Zn.Azo-CPA) studied by 15 N NMR spectroscopy over pH range from 7.0 to 10.3. Its azo nitrogen resonances, Na (proximal to Tyr-248) and NB (distal), show 8.3 ppm and 23.5 ppm shielding respectively relative to the corresponding resonances of apoarsanilazotyrosyl-248 carboxypeptidase A at pH 8.8. By comparison with the 15 N shifts of a model azotyrosine and its zinc complex, these shift differences indicate substantial complexation of the azotyrosyl-248 with the catalytic zinc at pH 8.8. Consistent with such complexation is the fact that the azo nitrogen resonances of Zn.Azo-CPA show large deshieldings on addition of a slowly hydrolyzing substrate glycyl-L-tyrosine or an inhibitor β -phenylpropionate, each of which binds to the catalytic zinc and therefore is expected to break any azotyrosine-Zn complex. The results demonstrate unequivocally that the molecular basis for the properties of Zn.Azo-CPA observed by other spectral techniques is a Zn-azotyrosyl-248 complex, with the N β nitrogen (and presumably the phenolic oxygen) as ligand to zinc. Furthermore, the high sensitivity of 15 N shifts to coordination with zinc permits at least a semi-quantitative estimate of the degree of azotyrosine-Zn complexation and indicates that the major conformation of ζ n.Azo-CPA, and by implication, of CPA in solution could be the one in which the active-site tyrosine is complexed to zinc, in contrast to the major conformation in the crystalline enzyme in which the Tyr-248 hydroxyl is 17 Å from the zinc. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Roberts, John D. Thesis Committee: Unknown, Unknown Defense Date: 14 August 1980 Record Number: CaltechTHESIS:09182025-000702960 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:09182025-000702960 DOI: 10.7907/ryxe-9852 ORCID: Author ORCID Kanamori, Keiko 0000-0001-6931-724X Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17686 Collection: CaltechTHESIS Deposited By: Ben Maggio Deposited On: 04 Oct 2025 12:11 Last Modified: 04 Oct 2025 12:33 Thesis Files PDF - Final Version See Usage Policy. 52MB Repository Staff Only: item control page

15 N NUCLEAR MAGNETIC RESONANCE STUDIES OF THE ACTIVE SITES OF ENZYMES AND IN VIVO NITROGEN METABOLISM Thesis by Keiko Kanamori In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (Submitted August 14, 1980) ii To Hirao, Atty and Teddy iii ACKNOWLEDGMENTS I would like to express my deep gratitude to my research advisor, Professor John D. Roberts, for his guidance, encouragement and patience during the course of this study. His insight, know- ledge and attitude towards science have made it a truly rewarding experience to work for him. I am grateful to all members of his group, past and present, for their support and numerous assistance along the way. I would like to express my sincere thanks to Professor John H. Richards and his group for their kindness and generosity in permitting me to use many of their facilities over the years; the work described in Part I would not have been possible without their generous assistance. I would like to thank Professor William W. Bachovchin for many helpful suggestions and collaboration in the work described in Part I and for teaching me many useful techniques during my first year. The contribution of Professor Bert L, Vallee to the project is also gratefully acknowledged. I would also like to thank Timothy L. Legerton and Professor Richard L, Weiss for their valuable contributions to the work described in Part IV. iv I am grateful to Mrs. Rose Meldrum for skillful illustrations and typing, and Ms Vere Snell for her excellent typing of most of this thesis. Support was received from the Institute in the form of Graduate Research Assistantship from 1977 to 80; this support is greatly appreciated. Finally, I would like to thank the members of my family for their patience and companionship during the busy years. V ABSTRACT PART I A 15 N Nuclear Magnetic Resonance Study of the Active-site Conformation of Arsanilazo-carboxypeptidase A To determine whether the catalytic zinc of carboxypeptidase A (CPA) is complexed to the active-site tyrosine (Tyr-248) in solution, a 15 N probe, an arsanilazo group enriched in 15 N at both azo nitrogens, has been selectively coupled to Tyr-248 with retention of full catalytic activity, and the resulting arsanilazotyrosyl-248 carboxypeptidase A (Zn.Azo-CPA) studied by 15 N NMR spectroscopy over pH range from 7.0 to 10.3. Its azo nitrogen resonances, Na (proximal to Tyr-248) and NB (distal), show 8.3 ppm and 23.5 ppm shielding respectively relative to the corresponding resonances of apoarsanilazotyrosyl-248 carboxypeptidase A at pH 8.8. By comparison with the 15 N shifts of a model azotyrosine and its zinc complex, these shift differences indicate substantial complexation of the azotyrosyl-248 with the catalytic zinc at pH 8.8. Con- sistent with such complexation is the fact that the azo nitrogen resonances of Zn.Azo-CPA show large deshieldings on addition of a slowly hydrolyzing substrate glycyl-.!::_-tyrosine or an inhibitor B-phenylpropionate, each of which binds to the catalytic zinc and therefore is expected to break any azotyrosine-Zn complex. The results demonstrate unequivocally that the molecular basis for the properties of Zn.Azo-CPA observed by other spec- vi tral techniques is a Zn-azotyrosyl-248 complex, with the NS nitrogen (and presumably the phenolic oxygen) as ligand to zinc. Furthermore, the high sensitivity of 15 N shifts to coordination with zinc permits at least a semi-quantitative estimate of the degree of azotyrosine-Zn complexation and indicates that the major conformation of zn.Azo-CPA, and by implication, of CPA in solution could be the one in which the active-site tyrosine is complexed to zinc, in contrast to the major conformation in 0 the crystalline enzyme in which the Tyr-248 hydroxyl is 17 A from the zinc. The binding of the quasi-substrate, glycyl-.!:._-tyrosine, or the inhibitor, B-phenylpropionate, to Zn.Azo-CPA causes a greater deshielding of the NB resonance than can be accounted for solely by the disruption of azotyrosylZn complex. The NB shift is also a sensitive probe of intramolecular hydrogen bonding of this nitrogen with Tyr-248 hydroxyl whose postulated catalytic function is to donate a proton to the leaving amine group of the substrate. The NB resonance of Zn.Azo-CPA-Gly-Tyr complex is deshielded by 8.9 ppm relative to that of Apo.Azo-CPA at pH 8.8, a result consistent with a partial hydrogen bonding of Tyr-248 hydroxyl with the amide nitrogen of the scissile peptide bond of the substrate. In the Zn.Azo-CPA-B-phenyl- propionate complex, a 28.2 ppm deshielding of the NB resonance relative to that of the apoenzyme indicates disruption of the hydrogen bond between NB and Tyr-248 hydroxyl; the most reasonable explanation is that the tyrosine hydroxyl hydrogen-bonds to the carboxylate group of the inhibitor. vii PART II Benzoate Catalysis in the Hydrolysis of endo-5-[4'(5' )imidazolyl]-bicyclo[2.2.l] hept-endo-2-yl transcinnamate: Implications for the Charge-transfer Mechanism of Catalysis by Serine Proteases The acceleration, by a factor of 2500, of the hydrolysis of endo-5-[4'(5') imidazolyl]bicyclo[2,2.l]hept-endo-2-yl trans-cinnamate - - by 0.5 M sodium benzoate in 42 mol % dioxane in water can be explained without resort to operation of a "charge-relay" mechanism similar to that often postulated to account for the enzymatic activity of serine proteases. The degree of ionization of 4-methylimidazole and of sodium benzoate in 42 mol %dioxane in water at 60°c have been measured by NMR spectroscopy. viii PART III Studies of pH and Anion Complexation Effects on !:_-Arginine by Natural Abundance 15 N Nuclear Magnetic Resonance Spectroscopy Natural-abundance nitrogen-15 nuclear magnetic resonance (NMR) spectroscopy has been used to investigate (1) the effect of pH on the 15 N chemical shifts of 1-arginine and (2) the possible effects on 15 N chemical shifts and T1 values of the complexation of 1-arginine with chloride ions, phosphate ions, and adenosine triphosphate (ATP) in aqueous solutions. Guanidine carbonate solutions and protamine sulfate solutions containing some of these ions have also been examined. On protonation of the guanidinium group of 1-arginine , the guanidino nitrogens and the -NH-nitrogen show upfield shifts of 15 and 3 ppm, respectively, as a result of change in the second-order paramagnetic effect, while the a-amino nitrogen shows an 8-ppm downfield shift on protonation as a result of decreased shielding. Complexation with equimolar phosphate ions produces a small downfield shift of approximately 0.8 ppm in the chemical shift of the guanidino nitrogens of~arginine. No significant changes in 15 N shifts or T1 values were observed on complexation with chloride ions or adenosine triphosphate . ix PART IV A 15 N nuclear magnetic resonance study of amino acid metabolism in Neurospora crassa 15 N nuclear magnetic resonance spectra of suspensions of intact Neurospora crassa mycelia, cultured in 15NH 4Cl-containing medium, have been obtained at 18.25 MHz. Well-resolved 15 N resonances of free metabolites that play crucial roles in intermediary nitrogen metabolism, such as the amide nitrogen of glutamine, the a-amino nitrogens of glutamate and other amino acids, the guanidino nitrogens of arginine and the ureido nitrogens of citrulline are observed, as well as those of cellular components such as chitin and uridine diphosphates. The guanidino nitrogens of intracellular arginine are found to have a spin-lattice relaxation time of l. 06 sec and nuclear Overhauser enhancement of -3.6, indicating that these nuclei have a correlation time of the order of 10-lO sec in the intracellular environment. This suggests that a pulsing rate of a few seconds with proton decoupling is favorable for the observation of protonated nitrogens of small metabolites in intact microorganisms. The time course of the assimilation of 15 NH: into glutamine and glutamate was followed in vivo in two !i_. crassa strains, one capable of synthesizing glutamate directly from NH + 4 and a-ketogl utarate, and the other incapable of this synthesis. Preliminary X results demonstrate that an alternative pathway of NH: assimilation into glutamate exists in Ji. crassa, and suggests that this takes place via the amide nitrogen of glutamine. Potential utility of 15 N NMR spectroscopy for study of the .i.!J.. vivo turnover rates of nitrogenous metabolites, for elucidation of previously unknown pathways and for cellular regulation of nitrogen metabolism in intact microorganisms is discussed. xi ABSTRACTS OF THE PROPOSITIONS PROP OS IT ION I A 15 N NMR study of the ionization state of the coenzyme pyridoxal phosphate at the active site of aspartate aminotransferase is proposed. PROPOSITION I I A 15 N NMR study of the mode of binding of the inhibitor sulfanilamide to the catalytic zinc of carbonic anhydrase is proposed. PROPOSITION III Chromosomal localization of the gene for the s chain of human embryonic hemoglobin is proposed, with a view to obtaining clues to the evolutionary relationship of the s chain to the a globin chain. PROPOSITION IV A resonance Raman study of electron polarization in a chromophoric substrate 4-dimethylaminocinnamaldehyde bound to alcohol dehydrogenase is proposed. PROPOSITION V A study of the mechanism of action of an anticancer drug, streptozotocin, by radiochromatography is proposed. xii TABLE OF CONTENTS PART I A 15 N Nuclear Magnetic Resonance Study of the Active-site Conformation of Arsanilazo-carboxypeptidase A Page Abstract ........................................................... V Introduction 2 Experimental 8 Results and Discussion .................. . .......................... 11 ~~N shifts of arsanilazo-ii_-acetyltyrosine (DAT) ................. 11 N shifts of tetrazolylazo-N-acetyltyrosine (TAT) and its ,···:········-:.................................... 16 15 zinc.complex N sh, fts of arsan, l azotyrosyl-248 CPA (Zn .Azo-CPA) . . . . . . . . . . . . 18 15 N shifts of apoarsanilazotyrosyl-248 CPA (Apo.Azo-CPA) ........ 25 15 N shifts of Gly-Tyr and S-phenylpropionate complexes of Zn .Azo-CPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Conclusion ......................................................... 36 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 PART II Benzoate Catalysis in the Hydrolysis of endo-5-[4'(5')imidazolyl]-bicyclo[2.2.l] hept-endo-2-yl transcinnamate: Implications for the Charge-transfer Mechanism of Catalysis by Serine Proteases Abstract ......................................... , ................. v, 1 Introduction ................................................... • • • • 43 xiii Page Experimental ....................................................... 46 Results and Discussion ............................................. 46 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 PART III Studies of pH and Anion Complexation Effects on L-Arginine by Natural Abundance 15 N Nuclear Magnetic Resonance Spectroscopy Abs tract .......................................................... vii; Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Experimenta 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Results and Discussion ............................................. 56 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 PART IV A 15 N Nuclear Magnetic Resonance Study of Amino Acid Metabolism in Neurospora crassa Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,x Introduction ....................................................... 68 Experimenta 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 xiv Page Results and Discussion . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 1. The 74 15 N shift, the spin-lattice relaxation time and the nuclear Overhauser enhancement of intracellular arg1n1ne ( 15N J • N. crassa ........................................ ........w?w ,n ... 74 2. l 5N .spectra of intact tL crassa myce l i a cultured in 15NH conta i ni ng medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 3. The time course of assimilation of 15NH! into glutamine and glutamate in two !i_. crassa strains ........................... 101 4- References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 PROPOSITIONS Abstracts ........................................................... xi Proposition I .............••.•••••••• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 122 Proposition II ...................................................... 138 Proposition III ..................................................... 148 Proposition IV ...................................................... 158 Proposition V ................................................... • • • • 164 xv Part I Table I II I LIST OF TABLES Title Page 15 N shifts of arsanilazotyrosyl-248 carboxypeptidase A (Zn.Azo-CPA), its Gly-Tyr and B-phenylpropionate complexes, Apo.Azo-CPA and model compounds ................. 32 15 N and 13 c shifts of 4-methylimidazole and benzoic acid in 42 mol % dioxane in water ....... 48 III I 15 N NMR shifts of guanidine carbonate, L-arginine and protamine sulfate in presence of anions 60 IV II T1 values for !:_-arginine nitrogens in the presence of anions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 I T1 values of intracellular and extracellular • • (15N • N. crassa .................. 85 arg1n1ne , ) 1n w,w II Assignments of 15 N-labeled metabolites observed in N. crassa mycelia ............................ 89 xvi LIST OF FIGURES Part Figure I 1 The active site of crystalline carboxypeptidase A 3 2 The active site of crystalline carboxypeptidase A on binding of Gly-Tyr ......................... . 4 A probable mechanism of catalysis by carboxypeptidase A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 The structures of arsanilazotyrosyl-248 carboxypeptidase A (Zn.Azo-CPA) and model compounds ..... 12 3 4 5 Title 15 N shifts of arsanilazo-!!--acetyltyrosine (DAT) as functions of pH .. .. . .. .. . .. .. . . . . . . . . . .. .. .. .. 6 13 13 c shifts of arsanilazo-_!i-chloroacetyltyrosine as functi ans of pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Page 15 15 N shifts of tetrazolylazo-_!i-acetyltyrosine (TAT) and its zinc complex as functions of pH 17 8 15N spectra of Zn.Azo-CPA at pH 7.0, 8.8, and 9.6. 19 9 15N shifts of Zn.Azo-CPA, Apo.Azo-CPA, DAT and 10 TAT.Zn complex as functions of pH ................ 20 Probable states of Tyr-248 and Zn in Zn .Azo-CPA... 24 xvii Part Figure I 11 Title 15 N spectra of Zn.Azo-CPA, its Gly-Tyr and Page S-phenylpropionate complexes and Apo.Azo-CPA at pH 8. 8 ........................................ . 12 IV N stli fts of Zn .Azo-CPA and its Gly-Tyr and S-phenylpropionate complexes as functions of pH 29 13 Possible modes of binding of Gly-Tyr to Zn.Azo-CPA 31 14 Possible mode of binding of s-phenylpropionate to Zn.Azo-CPA ................................... . 34 13c shifts of benzoate carboxyl carbon as a II I II 15 26 l 1 function of pH in 42 mol % dioxane in water....... 49 The pH dependence of 15 N chemical shifts of !:_-arginine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 15 N chemical shift spectrum of intracellular arginine ( 15 N ,) in intact N. crassa mycelia ..... 75 w,w 2 3 4 15 N spectra of intracellular arginine ( 15N w,w ,) in N. crassa for NOE measurement ................ . 77 a. Plot of NOE against Tc . for nitrogen-15 ........ . b. Plot of T1 against Tc for nitrogen-15...... .... 79 79 Representative 15 N spectra for the T1 measurement of arginine ( 15 N , ) in N. crassa ........ . ..... . w,w 81 xviii Part IV Figure Title Page 5 Plots of ln(S co -S) against T for the r 1 measureT 15 ment of arginine ( N , ) in N. crassa . . . . . . . . . . . . . 82 w,w 6 a. 15 N spectrum of mycelia suspension of N. crassa ure-1 strain grown in 15NH 4Cl-containi;g medium b. spectrum of the same mycelia after additional 14NHll containing medium..... 2 hours' growth in 87 Anabolic pathways of some nitrogenous metabolites in!:!_. crassa mycelia .............................. •. 90 15 N·spectrum of cel 1 extracts of!:!_. crassa mycel ia... 96 15 N spectrum of mycelia suspension of N. crassa ure-1 strain grown in 15 NH 4Cl-containi;g medium after addition of cycloheximide . . . . . . . . . . . . . . . . . . . . . 99 • course of ass1m1 • •1at.,on of 15,NH+ ,n • to a. Time 4 glutamine and glutamate in !i• crassa am-1 strain b. The same in N. crassa ure-1 strain •········ •· •··· 105 rs; 7 8 9 10 11 Comparison of the peak intensities of glutamine Ny and glutamate N as functions of time in N. crassa a am- l and ure- l strain ...............•• • • • • • • • • • • • • • • l 08 1 PART I A 15 N Nuclear Magnetic Resonance Study of the Active-site Conformation of Arsanilazo-carboxypeptidase A 2 INTRODUCTION Carboxypeptidase A (CPA) catalyzes the cleavage of C-terminal amino acid from a peptide chain. * Its active-site residues have been identified as zinc, tyrosine (Tyr-248), glutamic acid (Glu-270), and arginine (Arg-145) on the basis of chemical modificationla-c and X-ray 2a,b studies. In recent years, there has been considerable controve~?Y 3a-c regarding the conformation of the active site, specifically the location of Tyr-248 relative to zinc, because the conformation of the enzyme in solution might be significantly different from that in the crystalline enzyme. According to the most recent interpretation of the electron-density maps, crystalline CPA at pH 7.5 exhibits two conformations; the predominant conformation being the one in which Tyr-248 hydroxyl is at the surface of the molecule 0 and is separated by 17 A from the zinc (Figure 1), and a second conformation constituting up to 25% of the enzymes in which the tyrosine hydroxyl 0 is within 2 A of the zinc. 4 On binding of a slowly-hydrolyzing substrate, glycyl -b_-tyrosine, to the enzyme, there is a large confo rmational change of the predominant conformation whereby the tyrosine hydroxyl moves from 0 0 the surface of the molecule to within 4~5A of the zinc and 3A from the NH group of the scissile peptide bond 2b (Figure 2). This movement of Tyr-248 has been regarded as one of the most striking examples of the induced-fit theory of enzyme-substrate binding . *The pH dependence of the activity of CPA at ionic strength 0.5 shows a broad plateau of optimum activity in pH 7.0 to 9.0 region, th relative velocity of hydrolysis at pH 8.6 being 93% of that at pH 7.5. 1d 3 Zn. lv , Fig. l The predominant conformation of the active site of crystalline carboxypeptidase A according to X-ray structure. The carbon, oxygen and nitrogen atoms of the active-site residues (including the C atoms) a are shown by filled circles, and parts of the peptide backbones around the active site are shown by blank circles (Adapted from R. E. Dickerson and I. Geis, "The Structure and Action of Proteins, Harper and Row, New York, l 969) 11 , 4 Fig. 2 The active-site conformation of crystalline carboxypeptidase A on binding of a quasi-substrate glycyl-L-tyrosine. Glycyl-L-tyrosine is shown by shaded circles and the active-site residues by filled circles. (Adapted from R. E. Dickerson and I. Geis, "The Structure and Action of Proteins~ Harper and Row, New York, 1969) 5 However, Vallee and coworkers 3a, b have studied the active-site confonnation of the enzyme in solution and proposed that the Tyr-248 hydroxyl is complexed to zinc before substrate binding, and therefore has a confonnation significantly different from that of the major conformation of the crystalline enzyme. To determine the confonnation in solution, Vallee and coworkers used a modified CPA with the active-site tyrosine-248 residue specifically labeled with an arsanilazo group. The coupling of arsanilazo group to Tyr-248 is selective, 5 and the resulting arsanilazo tyrosyl-248 CPA (Zn.Azo-CPA) has full catalytic activity, 6 indicating that the activesite conformation necessary for catalysis is preserved. Vallee and co- workers showed, by comparison with model azotyrosine compound and their • • • v1s1 • •bl ea bsorp t.,on, 3b' 6 c1rcu • 1ar zinc comp 1exes, th at Zn. Azo- CPA exh1b1ts dichroic, 3b, 6 and resonance Raman spectra 7 similar to those of azophenol, azophenolate-Zn complex and azophenolate ion at pH 6.2, 8.5 and 10.8 respectively and therefore they proposed that Tyr-248 in Zn.Azo-CPA is complexed to the catalytic zinc around pH 8.5 in solution. Valuable as the speitral techniques have proved to be, the techniques employed so far provide no unequivocal molecular basis for the observed spectral properties of Zn.Azo-CPA, and if azotyrosine-Zn complexation is taking place, the mode of complexation on molecular level--whether the proximal or the distal azo nitrogen is involved in chelation--is not known. Moreover, previous techniques provide no quantitative estimate of the degree of Tyr-248-Zn complexation, if it is occurring, at the active-site of Zn-Azo-CPA in solution. Another interesting point is the detailed con- formational change that occurs at the active-site on binding of a substrate or an inhibitor to Zn.Azo-CPA in solution. 6- The present research involves the application of 15 N nuclear magnetic resonance spectroscopy to study the active-site conformation of arsanilazotyrosyl-248 CPA ( 15 N enriched in both azo nitrogens) (1) as free enzyme in solution, (2) on binding of a quasi-substrate, glycyl-L-tyrosine and an inhibitor S-phenylpropionate, and (3) on removal of the catalytic zinc. The potential of 15 N NMR spectroscopy as probe of enzyme active sites has been successfully demonstrated for a-lytic protease. 8 15 N chemical shift is potentially a very sensitive probe of azotyrosine-Zn complexation in arsanilazotyrosyl-248 CPA, because, with a model azotyrosine compound, the azo nitrogen shows an upfield shift of fully 90 ppm on complexation with zinc. The high sensitivity of 15 N shift to coordination with zinc permits a semi-quantitative estimate of the degree of azotyrosineZn complexation in arsanilazotyrosyl-248 CPA in solution which so far has not been possible by other spectral techniques. With Zn.Azo-CPA 15 N enriched in both azo nitrogens, it is also possible to determine which azo nitrogen, if any, is involved in complexation. Moreover, the 15 N shifts of azo nitrogens can be a sensitive probe, not only of zinc complexation but also of intramolecular hydrogen bonding with tyrosyl-248 hydroxyl whose catalytic function, according to proposed mechanism of catalysis 9 is to donate a proton to the leaving amine group (Figure 3). Thus, 15 N chemical shifts of Zn.Azo-CPA-Gly-Tyr complex were expected to provide important information on the state of hydrogen-bonding of the Tyr-248 hydroxyl in the enzyme-substrate complex in solution. 7 OH Hydrophobic pocket Afll45 N-( Tyr 248 + N ~ V-o . . . H (a) CH_coo- / H-N Glu 270 ~ C _,,,,,.. \\ 0 o- \=o I @ /CH R2 \ N-H I ~l \. ___ cooCH (d) / H-N I H HO \ c=o I ..-CH R2 \ N-H I Fig. 3 A probable mechanism for catalysis by carboxypeptidase A (Adapted from R. E. Dickerson and I. Geis. "Structure and Action of Proteins, 11 Harper and Row, New York, 1969) 8 EXPERIMENTAL Sodium nitrite (99% enriched in 15 N) and Q_-arsanilic acid (99% enriched in 15 N) were obtained from Isotope Labeling Corporation. Monoarsanilazo-N-acetyltyrosine (DAT) was prepared by reacting diazo. 7c l 0 tized arsanilic acid with H_-acetyltyros,ne. ' Selective nitrogen-15 isotopic enrichment of Na or NB was effected by using sodium ( 15 N) nitrite or ( 15 N)Q_-arsanilic acid, respectively in the diazotization reaction. The ultraviolet-visible absorption spectrum of each preparation of 15 Nenriched DAT was carefully examined over the pH range 4.0 to 11.0. general, all spectral parameters were In found to be identical to those reported previously . le, lO Monotetrazolylazo-!!_-acetyltyrosine (TAT) was prepared by reacting diazonium-lH-tetrazole with !i_-acetyltyrosine as described by Sokolovsky and Vallee. 11 The 15 N derivative was synthesized by using sodium ( 15 N) nitrite in the preparation of the diazonium-lH-tetrazole. Ultraviolet- visible absorption spectra of the nitrogen-15 enriched product were obtained as a function of pH and of added zinc ion; observed spectral properties agreed closely with those previously reported. 3b Carboxypeptidase A12 was obtained as a crystal suspension from Sigma Chemical Company and was recrystallized before use. Arsanilazotyrosyl- 248 carboxypeptidase A was prepared by reacting diazotized Q_-arsanilic acid with carboxypeptidase A crystals as described by Johansen and Vallee. Nitrogen-15 enrichment at Na or NB was effected through the use of sodium ( 15 N) nitrite or ( 15 N)Q_-arsanilic acid, respectively, as described above for DAT. Ultraviolet-visible absorption spectroscopy was used to 3a 9 demonstrate that each preparation of nitrogen-15 enriched azo-enzyme responded to pH, removal of zinc, inhibitors and substrates, and denaturing agents in the manner expected. All spectral properties were in good agreement with previously reported results. Peptidase activity was determined spectrophotometrically using (furylacryloyl)glycylphenylalanine. Activity of the modified enzyme as measured with this substrate was not discernibly different from that of the unmodified enzyme. Apoarsanilazotyr-248-CPA was prepared by suspending the crystals of arsanilazotyr-248-CPA (5 mg/ml) in 0.01 M 1,10 phenanthroline, 0.01 M 2o (fi-morpholino)ethanesulfonic acid (Mes) buffer (pH 7.0) at 20 C for 1 hr. for four successive times followed by four 0.5 hr. washings with 0.01 M Mes buffer (pH 7.0). 9b The apoenzyme prepared by this method contained less than 0.016 g-atom of zinc/mol of enzyme as determined by atomic absorption spectroscopy. 10-Phenylazo-9-phenanthrol was synthesized by refluxing 9,10phenanthraquinone and ph~nylhydrazine hydrochloride in acetic acid. 9c Crystallization from ethanol gave pure 10-phenylazo-9-phenanthrol 0 (melting point 163~164 C). Its anion was prepared by addition of 1.5 equivalent of sodium hydroxide in dimethyl sulfoxide. 15 N chemical shift spectra were obtained with a Bruker WH180 spectromA 1 M solution of H15 No in o o contained 3 2 in capillary provided both reference signal and field-frequency lock. eter operating at 18.25 MHz. For the model compounds arsanilazo-!i_-acetyltyrosine (DAT), tetrazolylazo!i_-acetyltyrosine (TAT) and its zinc complex (TAT.Zn), the spectra were 10 taken of 5 mM aqueous solution. 0 The operating conditions were 90 pulse and 15 sec. delay .. For arsanilazotyrosyl-248-CPA, the spectra were taken of 1.5~2.0 mM aqueous solutions in 1.0 M NaCl-0.02M tris(hydroxymethyl) 0 aminomethane buffer at 15 C. Glycyl-.h_-tyrosine (7.5 mM) or S.,-phenyl- propionate (7.5 mM) was added to the enzyme solution to make enzymesubstrate or enzyme-inhibitor complexes, respectively. For apoar- sanilazotyrosyl-248-CPA, the spectra were obtained of 0.3 mM aqueous 0 solution in the same buffer at 5 C. Operating conditions for the 15 N spectra of the enzyme, its substrate and inhibitor complexes and the 0 apoenzyme were 90 pulse and 0.8 sec. delay and pH 7.0 and 7.8 and 0.29 sec. delay at other pH values. accumulation time. Each spectrum required 8~16 hours of The stabilities of the enzyme and its complexes during NMR measurement were monitored before and after by the visible absorption spectrum of each solution. All solutions were found to be stable during the NMR measurements. The peptidase activity of Zn.Azo-CPA was checked before and after each NMR measurement. For NMR experiments at pH 7.0~9.6, there was no appreciable change in activity before and after the experiments. At pH 10.3, Zn.Azo-CPA has no catalytic activity, presum- ably due to binding of hydroxide ion to catalytic Zn, but the threedimensiona1 conformation of the enzyme is preserved as indicated by recovery of catalytically active Zn.Azo-CPA on lowering pH. 13 c chemical shift spectra were obtained with a Varian XLlOO spectromo eter operating at 25.14 MHz'. Operating conditions were 90 pulse and 3 sec. delay. 11 Enzyme concentrations were measured by the absorbance at 278 nm, based on a molar absorptivity at 278 nm of 7.32 x 10 4 M-l cm: 1 Visible absorption spectra were obtained on Beckman Acta CIII spectrophotometer, and/or on a Varian 219 spectrophotometer. RESULTS AND DISCUSSION Figure 4 shows the structures of arsanilazotyrosyl-248 residue of the modified CPA (Zn.Azo-CPA), and two model compounds arsanilazo!i_-acetyltyrosine (DAT) and tetrazolylazo-N-acetyltyrosine (TAT). DAT is the model for free azotyrosine, but because the bidentate DAT does not form a soluble 1:1 zinc complex, the zinc complex of tridentate TAT was used as a model for azotyrosine-Zn complex as in previous studies of arsanilazotyrosyl-248 CPA by visible absorption and resonance Raman spectroscopy. 15 N NMR shifts of DAT The 15 N chemical shifts of the two azo nitrogens of DAT in water as a function of pH is shown in Figure 5. The assignment of Na and NS nitrogens was straightforward because DAT was 15 N-enriched in Na and NS separately. While the N shift of DAT at -126~-128 ppm is close to that a of azobenzene, -133 ppm, 13 the NS shift of DAT at -82 ppm is approximately 45 ppm shielded relative to Na. This shielding is the result of an intra- molecular hydrogen bond between NS and the phenolic proton (Figure 5-I). Although such hydrogen-bond is well-known for a-hydroxy azo compound,s , 14 , 15 it has not been previously determined whether the hydrogen bond is to Na 12 Tyr-248 tJNet. ,(i~p A•O,H, . ~ OH »'\ 17llffel/ll7 A-r~~viilazotyrosi~e-2't8 carboxypeptidase A (Zn·A-zo-(PA) 0 0 II CH3-C-NH-CH-coe I z (Hz II 8 CH1-C-NH-CH-C0 ~ I 1 CMz. OH Arsanilazo-N-Acetyltyrosine (DAT) Tetr~zoly13zo-N-Acetyltyros,ne (TAT) Fig. 4 The structures of arsanilazotyrosyl-248 residue of the modified CPA (Zn.Azo-CPA) and the model compounds arsanilazo-!i acetyltyrosine (DAT) and tetrazolylazo-!:!_-acetyltyrosine (TAT). 13 0 II 0 CH ,-C-NH-CH-CO I I ' CH, ~ '5Ny I fla 0 11~ / a 'H HAao, I II m o-o-o-o-oo-o-0-0---0 Np z -"' • 120 □ -□-o-□--o-o--□-□ --o-a N"'- 6 8 10 pH Fig. 5 15 N chemical shifts of the azo nitrogens of arsanilazoN-acetyltyrosine (DAT) as function of pH in water. 14 or to NB in the model compound DAT. 7b The 15 NB shifts of DAT in H o and 2 in dimethyl sulfoxide (discussed on p.30 ) clearly establish NB as the acceptor of the hydrogen bond from the phenolic hydroxyl. The 15 N chemical shift of DAT shows small but real changes with a pH, a shielding of 2.2 ppm between pH 5.8 and 8.8 and a deshielding of 1.7 ppm between pH 8.8 and 11.9. The shielding is probably due to the dissociation of the second proton from the arsenate group, and the deshielding can be attributed to the ionization of the tyrosine hydroxyl. The 13 c shift-pH titration curves of arsanilazo-N-chloroacetyltyrosine in H20, shown in Figure 6, support this interpretation; between pH 6 and 8.1, only the C4 carbon adjacent to the arsonate group shows appreciable shift change with pH,-while at higher pH, the aromatic carbons on the tyrosine moiety show large shift changes with pH, with a pK value of a approximately 9.5, due to ionization of the tyrosine hydroxyl. For a structurally similar model compound monoarsanilazocresol, pK values of a 8.4 and 9.5 have been reported for the ionization of the arsonic and the • 7b tyrosine hydroxyl respectively. The apparent insensitivity of the NB shift of DAT to ionization of tyrosine hydroxyl is surprising but can be interpreted as follows. When the phenolic proton ionizes, the resulting anion can be represented as a resonance hybrid of the principal structures II and III (Figure 5). If the azophenolate (II) were the dominant resonance form, the loss of the hydrogen bond would cause a significant deshielding of NB. However, a predominant contribution from structure III would be expected to be 15 • 0 C3.,J' I ,_ • C.lf. i s .sinqlet } in pl'oton0 s doublet co'4pled spect-Y~ 6 7 9 10 11 13 Fig . 6 c chemical shifts of arsanilazo-,!i-cl1loroacetyltyrosine .as function 13 of pH in water. The c resonances were assigned by comparison with those of tyrosine, azobenzene and arsanilic acid. The assignments of C~uz and CE1 above pH 9.2 are uncertain due to overlap. 16 associated with a pronounced shielding of N . These effects oppose each 8 other and if more or less equal could thus result in an essentially unchanged 15 N shift. Further evidence in favor of this interpretation will be presented later in this paper. 15N NMR shifts of TAT and TAT-Zn complex The 15 N shifts of TAT and its zinc complex as a ,function of pH are a shown in Figure 7. The 15N shifts of TAT and DAT in the absence of zinc a are similar except for the small shielding around pH 7.7 in DAT due to the ionization of the arsonate group. The 15 Na r~sonances of both TAT and DAT show slight deshielding around pH 8.5 due to the ionization of the ~yrosine hydroxyl. On addition of one equivalent of zinc chloride, the 15 N resonance of TAT shows a large shielding as pH is increased from 3 to 6; clearly the complexation with zinc is facilitated by ionization of the proton from the tetrazolering which has an estimated pK of 5.4. 7b Further shielding a is observed between pH 8 and 9 on ionization of the phenolic proton. On full zinc complexation at pH 9, the azo nitrogen of TAT is shielded to -40 ppm, an upfield shift of 90 ppm compared to that of TAT alone. Increasing the TAT-Zn ratio to 1:3 caused no further shielding of Na. Above pH 10.0, hydroxide ions begin to complex with zinc, and precipitation of zinc hydroxide occurs. It is known that an a-hydroxy azo compound with a heterocyclic substituent such as TAT acts as a tridentate ligand, complexing with most 17 z -"' \ ,() • 100 ~ • 120 /· A TAT /.,...-•-·-·-·-·--·---· 6 8 10 pH Fig. 7 15 N shifts of tetrazolylazo-I!-acetyltyrosine (TAT) and its zinc complex (TAT.Zn) as functions of pH. 18 metals through the 2_-hydroxyl group, the azo nitrogen nearest to the phenolic ring (N) and the heterocyclic nitrogen atom, giving two stable a five-membered chelate rings as shown in Figure 7. 16 a, b Therefore, the shift of Na in TAT.Zn complex at -40 ppm can be regarded as the shift of an azo nitrogen directly complexed to Zn. Thus, if azotyrosine-Zn complex- ation actually occurs in arsanilazotyrosyl-248-CPA, we can expect one of the azo nitrogens, the one directly complexed to zinc, to shift upfield to the vicinity of -40 ppm. 15 N shifts of arsanilazotyrosyl-248 CPA (Zn.Azo-CPA) The 15 N NMR spectra of arsanilazotyrosyl-248 CPA enriched in 15 N at both azo nitrogens, at pH 7.0, 8.8 and 9.6 are shown in Figure 8. This enzyme, with a molecular weight of 34,472 17 is the largest enzyme from which 15 N spectra have so far been obtained. The assignment of Na and NS was made by first obtaining 15 N spectra of arsanilazotyrosyl-248 CPA 15 N enriched in N only; the 15 N shift so obtained was identical to the a one assigned to Na in the doubly enriched enzyme. The 15 N shifts of the two azo nitrogens of arsanilazotyrosyl-248 CPA (Zn.Azo-CPA) as function of pH are compared with those of DAT and TAT.Zn complex in Figure 9, and are listed in Table 1. At pH 7.0 and 7.8, the shifts of Na at -130 ppm and NS at -78 ppm in the enzyme are very close to the corresponding shifts of DAT without zinc and therefore there is no indication of complexation with zinc at this pH; the shielding of NS relative to Na indicates that in the enzyme as in DAT, NS is hydrogenbonded to the Tyr-248 hydroxyl. 19 8.8 ! 9.6 -120 -BO -40 o ppm Fig. 8 15 N spectra of arsarrilazotyrosyl-248 CPA (Zn.Azo-CPA) enri.ched in 15 N at both azo nitrogens obtained at pH 7.0 (43,799 transients), pH 8.8 (70,010 transients) and at pH 9.6 (52,057 transients). 20 A t'rpo•Pa.O· - -•-- • ,r • ~ ·CP,r Nf O-0-O-O.0-0•0-o-0 E -ao ~ ~ rree kZo-Tyr Np '- z ~ '° •100 A I I 3 7 6 8 9 10 pH Fig. 9 Thi pH dependence of the 15 N shifts of the two azo nitrogens of arsanilazotyrosyl-248 CPA (Zn.Azo-CPA; ■: Na: •; NJ; apoarsanilazotyrosyl-248 CPA (Apo.Azo-CPA,*: Na; X: N ), and the model compounds 8 DAT (free Azo-Tyr, c: Na; o: N8) and TAT.Zn complex (Zn.Azo-Tyr, -D-: Na) 21 At pH 8.8, however, the NS nitrogen shows a large upfield shift to -56 ppm and Na shows a small upfield shift to -118 ppm. These results strongly suggest that at this pH, the azotyrosine in Zn.Azo-CPA is at least partially complexed with zinc. The larger shift observed for Ns relative to Na suggests that in arsanilazotyrosyl-248 CPA, zinc complexes to the phenolic oxygen and Ns, resulting in a six-membered ring (Figure 10) as in other bidentate azophenol ~etal complexes whose structures are known from X-ray studies. 18a, b This is in contrast to the tridentate TAT where Zn complexes to Na , phenolic oxygen and heterocyclic nitrogen to form two five-membered rings as mentioned earlier. Because the azo nitrogen directly complexed to zinc is Na in TAT but NS in Zn.Azo-CPA, the 15 N shifts of these two nitrogens are the ones that should be compared to estimate the extent of zinc complexation in the enzyme, although such estimate will be approximate because of the unknown effect of the presence of the third ligand, the tetrazole nitrogen, on the magnitude of the shielding of the azo nitrogen on complexation with zinc. A semi-quantitative estimate of the extent of complexation in· Zn.Azo-CPA can be attempted as follows. One approach is to compare the observed 15 N chemical shifts (o 15 N) of the free azotyrosine,model azotyrosine-Zn complex and the enzyme. If we assume from Figure 9 that -40 ppm represents the chemical shift of an azo nitrogen fully complexed to zinc, and -82.5 ppm represents that of an azo nitrogen hydrogen-bonded to tyrosine hydroxyl, then the chemical shift of -55.6 ppm observed for the azotyrosine in Zn.Azo-CPA suggests that about 63% of the time, the active-site tyrosine in the enzyme is complexed to catalytic zinc. Another approach is to compare the 22 magnitude of shielding, 6o 15 N, on complexation, in the model compound and the enzyme. If we assume that o15 N of Zn.Azo-CPA in the absence of hydrogen-bonding or zinc complexation is equal to that the NB of DAT in dimethyl sulfoxide, -103.9 ppm, or of NB in Zn.Azo-CPA~B-phenylpropionate ~ ( complex, -107.3 ppm (Table I, discussed on p. 30 ), and further assume that only zinc complexation contributes to the shielding of NB in Zn.Azo-CPA, 15 the observed o NB of -55.6 ppm indicates that zinc complexation causes 15 a shielding (6o NB) of 48.3 ppm (or 51.7 ppm) in Zn.Azo-CPA. Compared 15 to the 6o N of 90 ppm for the azo nitrogen of TAT on full complexation to zinc, the observed 6o 15 N in Zn.Azo-CPA indicates 54~57% complexation. The two approaches, based on slightly different assumptions as indicated above, yield reasonably close values, and suggest that the major activesite conformation of Zn.Azo-CPA at this pH is the one in which the active-site tyrosine is complexed to zinc. Such partial complexation can arise from either (1) the coexistence of at least two different conformations for a flexible tyrosyl-248 side chain in which full zinc-azotyrosyl complex occurs 54~63% of the time, or (2) a single conformation in which the azotyrosyl-248 Zn complex is weak due to steric constraints at the active site and/or partial neutralization of zinc by other ligands on the enzyme. Although one cannot distinguish between these two possibilities at present, the 15 N results clearly show that in solution the conformation of tyrosyl-248 side chain is such as to allow 54~63% azotyrosyl-Zn complexation at pH 8.8. At pH 9.6~10.3, both azo nitrogens in Zn.Azo-CPA shift downfield to values close to those of free azophenolate ion. This is in accord with loss of the azotyrosyl-Zn complexation because of the competition of 23 hydroxide ions with azophenolate for zinc, just as was observed in TAT.Zn mixtures above pH 10.0. Figure 10 shows schematically the probable states of the activesite tyrosine and zinc in Zn.Azo-CPA at various pH values. At pH 7.0, the azo nitrogen (NS) is intramolecularly hydrogen-bonded to the phenolic proton. At pH 8.8, on partial ionization of phenolic proton, the azo- phenolate complexes with zinc. At pH 10.3, hydroxide ion replaces azo- phenolate as ligand to zinc. The X-ray structure of crystalline CPA has been studied only at pH 7.5, and we cannot determine from the 15 N results alone whether at this pH the azotyrosyl residue is close to the catalytic zinc but is not complexed because the catalytic zinc does not compete effectively with proton for the azophenolate at that pH, or the tyrosine hydroxyl in solution is more distant from zinc as in crystalline CPA. However, the 15 N results clearly indicate that the tyrosine-248 side chain is mobile enough to assume a conformation in which the azophenolate is close enough to zinc to form a complex in the pH range where it is an effective ligand to zinc. Thus, caution is required in interpreting the conformation of crystalline enzyme as being directly applicable to solution conformation. ~ ~ iJJJJ.;JJ_ CHL CH.i lHz. f'N/1 15N,Y 1sN 1, 11 1,~ oe ' 11A R,_.. N~ '1-t ,,o HzO ' I I /\ R \\ I I ~ ,~/ 01-t f N .f.:,, - 'l///7f/l/ 7Tff'T pl-I 7, 0 pl-I 8.8 pl~ 9, 6 Fig. 10 Schematic diagram of the probable state of the active-site tyrosine and zinc in arsanilazotyrosyl-248 CPA at various pH values. 25 15 N shifts of apoarsanilazotyrosyl-248 CPA (Apo.Azo-CPA) To investigate whether the observed shielding of the azo nitrogens of arsanilazotyrosyl-248 CPA (Zn.Azo-CPA) at pH 8.8 is caused by complexation with the catalytic zinc and not by possible interactions with other active-site groups, the zinc was removed from the enzyme to obtain apoarsanilazotyrosyl-248 CPA (Apo.Azo-CPA) (Experimental section). The 15N spectrum of Apo.Azo-CPA at pH 8.8 is shown in Figure ll(D) and the 15 N shifts of the two azo nitrogens are compared to those of Zn.Azo-CPA in Figure 9. Removal of zinc causes a 23.5 ppm deshielding of NS resonance, from -55.6 ppm for the Zn.Azo-CPA to -79.1 ppm for the Apo.Azo-CPA, a value close to that of free azotyrosine DAT (-82.5 ppm). The Na resonance also s~ifts downfield, from -117.9 ppm for Zn.Azo-CPA to -126.2 ppm for Apo.Azo-CPA, a value almost identical with that of DAT (-126.0 ppm). These results provide strong evidence that the large shielding of the azo nitrogens observed in Zn.Azo-CPA relative to that of Apo.Azo-CPA and of free DAT at pH 8.8 is caused by partial complexation of the azotyrosyl group of the enzyme with the catalytic zinc. 26 Zn t,,,,z.o-CPA (pH 8.8} CA) • ~ ,~~ M ~ with inkibitor ~-pkenylpropionate cc> Apo·Azo-CPA {D) I -120 I -40 I 0 Fig. 11 15 N spectra at pH 8.8 of (a) arsanilazotyrosyl-248 CPA (Zn.Azo-CPA, 70,010 transients); (b) Zn.Azo-CPA with quasi-substrate glycyl-k-tyrosine (245,587 transients); (c) Zn.Azo-CPA with inhibitor B-phenylpropionate (131,054 transients); (d) apoarsanilazotyrosyl-248 CPA (Apo.Azo-CPA, 255,327 transients). 27 15 N shifts of Gly-Tyr and s-phenylpropionate complexes of Zn.Azo-CPA Further evidence that the shielding of the azo nitrogens of the enzyme at pH 8.8 is due to complexation with zinc is found in the significant changes observed in the 15 N shifts on addition of a substrate glycyl-k-tyrosine (Gly-Tyr) or an inhibitor S-phenylpropionate, each of which is known to bind to the active site and cause loss of the absorption maximum at 510 nm which has been attributed to Zn.azotyr-248 complex. 6 The mode of binding of Gly-Tyr to the active-site of carboxypeptidase A, according to X-ray structure at • reso l ut ,on, • 2b 1s • sown h • F1gure • 2; zinc is found to be bound 2A ,n to the carbonyl oxygen of the substrate and Tyr-248 has its hydroxyl close to the amide nitrogen of the peptide bond. Therefore the binding of Gly-Tyr is expected to break the azotyrosine-Zn complex and cause deshielding of the azo nitrogens. This is indeed what occurs. Figure 11 shows how the 15N spectrum of Zn.Azo-CPA' at pH 8,8 (A) compares with that of the substrate-enzyme complex at the same pH (B). The NS nitrogen which was shielded to -56 ppm in Zn,Azo-CPA shifts downfield, on binding of Gly-Tyr, to -88 ppm, a value close to that of DAT. The resonance of the N nitrogen also shifts downfield to a a value characteristic of DAT. These results provide further evidence that the 15 N shifts of the azo nitrogens in Zn.Azo~CPA are a sensitive 28 probe of the formation and dissociatton of azot.yrosine-:--zinc cornplexa,tton\ The inhibitor--s-phenylpropionate is also known to bind to the catalytic zinc atom through tts carboxylate group and is therefore expected to break the Tyr"Zn association.. In presence of the inhibitor at pH 8,8, the NS peak agatn ts substanttally downfield relative to that of the enzyme alone which indicates the expected dissociation of the Tyr-Zn association (_Figure 11 CC)}, The 15 N shifts of the s-ubstrate and inhibitor comp l e_xes of Zn,Azo-CPA at pH 8.8 are sensitive probes~ not only of the dissociation of azotyrosine-Zn complex, but also of the formation or dissociation of hydrogen bonding of NS with neighboring groups~ and thus provide interesting information as to the changes that occur at the active site of the enzyme in solution on binding of the sub.s.trate or the inhibitor. Figure 12 provides a comparison of the 15 N shifts of the substrate and inhi !:ii.tor complexes of Zn \,Azo.,.CPA over the range of pH from 8.8 to 10,3 wtt~ those of free Zn,Azo"'.CPA, Apo~zo.,..CPA and DAT, Significantly, at pH 8,.8, the NS shift of the enzyme i_nhi_bitor complex is deshielded by 28,2 ppm while that of the e.nzyme~substrate complex is slightly deshielded by 8,,9- ppm relattve to that of Apo,Azo-CPA, In the Gly .. Tyr complex of Zn.Azo-CPA, the 8,9 ppm deshielding of the NS resonance relattve to that of Apo.Azo-CPA suggests the 29 • Zt1·A-zo-CPA o A-zo-TyYi~ H2.0 -bO • Apo· MC· CPA * Z 11 •A-io -CPA + 4\y-TyY -70 -so ~ Zn ,/Jao-CPA -t itthibitol' ·-• A 0 --0-0-0-0-0-0-- /* -90 -100 -110 -120 7 6 9 10 pH Fig. 12 The pH dependence of the 15 N shifts of the two azo nitrogens of arsanilazotyrosyl-248 CPA (Zn.Azo-CPA) and its complex with glycyl-L-tyrosine and with inhibitor S-phenyl-- propionate. The l 5N shifts of apoarsanil azotyrosyl -24-8 CP.l\ (Apo-Azo-CPA) and of the model compound DAT (Azo-Tyr} are also shown for comparison. 30 following possibilities. (1) The dominant conformation at the active site is the one in which NB is hydrogen bonded to Tyr-248 hydroxyl (Figure 13 Ia) but there is a minor conformation in which Tyr-248 hydroxyl is intermolecularly hydrogen-bonded to another neighboring group, possibly to the NH group of the substrate (Figure 13 Ib). (2) The NB is hydrogen bonded to NH group of the substrate while the Tyr.,.248 hydroxyl is hydrogen bonded to the same amide nitrogen (Figure 13 II). If the true catalytic function of Tyr-248 hydroxyl is to donate a proton to the leaving amine group, conformation Ib or II in which the tyrosine hydroxyl is hydrogen bonded to the amide nitrogen of the substrate should occur in Zn.Azo-CPA which has full catalytic activity. Further study is required to distinguish between these possibiiities, for example, by use of N-methyl and ester analogues of Gly-Tyr. In the B-phenylpropionate complex of Zn,Azo-CPA, the NS resonance is deshielded to -107 ppm at pH 8.8 (Figure 12). This shift is seen to be comparable to that of the NB of DAT (-103. 9 ppm) in dimethyl sulfoxide (Table I), It is also clear that the change in NB shift of the inhibitor complex, from -107 ppm at pH 8,8 to -76,9 ppm at pH 10.3 closely resembles the change in the NS shift of DAT in dimethyl sulfoxide from -103,9 ppm to .,.78,l ppm* on ionization of Tyr-0H by addition of sodium hydroxide. 15 N shift of ionized DAT was measured in dimethyl sulfoxide-water * The in order to keep the ionized species in solution (see Table I). 31 Fig. 13 Possible modes of binding of Gly-Tyr to Zn.Azo-CPA eH 7.0 8.8 9.6 10.3 Table I 15 N shifts of Zn.Azo-CPA, its Gly-Tyr and 8-phenylpropionate complex25:, Apo.Azo-CPA and DAT o15 N for N5 ~ a Zn.Azo-CPA Zn.Azo-CPA Zn.Azo-CPA + Apo.Azo-CPA DAT l 0-phenyl + Gly-Tyr H20 DMSO S-cp-propi onate DMSO/H 20 azo-9-phenanthrol anion 3:1 v/v (DMSO -78.0 -82. l -107. 3-55.6 -88,0 -79. l -82.5 -103.9 - l 01 . 4 (phenol) -78.5 -84.4 -82.4 -80.8 -80.5 -76.9 -82.6 -78. l -33.8 (phenol ate) o15 N for N~ {eQm}a eH Zn.Azo-CPA Zn.Azo-CPA + Gly-Tyr 7.0 8.8 -129.8 -117.9 -125.l 9.6 10.3 -122.3 -123.8 -125.5 -125.9 a. from HN0 3. Zn.Azo-CPA + Apo.Azo-CPA B-cp-propionate -125.9 -126.2 - 125. l Negative sign indicates downfield. H20 -127.8 -126.0 -126.8 -127.4 DAT DMSO DMSO/H~O 3; l v/ 10-phenyl azo-9-phenan"." throl anion DMSO) -137.0 (phenol) -130. l (phenol ate) -110.l w N 33 The NB shifts of DAT in dimethyl sulfoxide can be interpreted as follows. The position of NB resonance of DAT in dimethyl sulfoxide at -103.9 ppm is much more deshielded than that in water (-82.5 ppm) because dimethyl sulfoxide is an excellent hydrogen-bond acceptor and can compete effectively with the NB as the acceptor for the hydrogen bond to the phenolic proton (see Figure 14). Arguing from this, the close similarity between the NB shift of DAT in dimethyl sulfoxide (-103.9 ppm) and that of the Zn.Azo-CPA- B-phenyloropionate complex (-107.3 ppm) at pH 8.8 strongly suggests that in the enzyme-inhibitor complex, the tyrosine-248 hydroxyl is not hydrogen bonded to NB but instead is hydrogen bonded to another neighboring group, the most likely acceptor being the second carboxylate oxygen of the inhibitor as shown schematically in Figure 14. Interaction of the carboxylate group of B-phenylpropionate with the catalytic zinc of CPA has been inferred from (a) X~ray structure 0 of CPA-B-(£-iodophenyl)propionate complex at 6 A resolution 19 ; (b) the decreased rate of exchange of zinc with 65 zn in CPA on inhibitor addition and the failure of the inhibitor to bind to zinc-free CPA. 20 It is further known that the Tyr-248 hydroxyl of the free CPA can be acetylated by treatment with H-acetylimidazole and that this acetylation is blocked if B-phenylpropionate is bound 0 ~ e II CJ-13-C -NH-CH-COi I Cl-fz Cl-ti pH 8.8 ' N -107.3~11 /N 0 R ,,. 'N -103,'l~ II N '' ' I 'o e:I ~-pke"yt- / pYo i,ionate Mz:O,A-.s ,t;:-c,R."' : o, H. ···o~ ,,,,,c MJ s I Cl·(, • Zn Fig. 14 Probable mode of binding of the inhibitor 13-phenylpropionate to the arsanilazotyrosyl-248 residue of Zn.Azo-CPA (left). The observed 15 N shifts in the enzyme-inhibitor complex are similar to those of the model compound OAT in dimethyl sulfoxide (right). w .i,.,. 35 • •t yo f t he carboxylate t o th e enzyme;21 th·1s suggests a c l ose prox1m1 group of the inhibitor not only to zinc, but also to the Tyr-248 hydroxyl of CPA. Thus the chemical and NMR evidence are consistent in sugg~sting that the carboxylate group of the inhibitor coordinates with zinc through one oxygen and forms a hydrogen bond with Tyr-248 hydroxyl through the other as shown schematically in Figure 14. On ionization of the tyrosine hydroxyl, the 15 N shift of Zn.Azo-CPA-B-phenylpropionate complex at pH 10.3 shifts upfield to -76.9 ppm, close to the value for ionized DAT in dimethyl sulfoxide/water, -78.l ppm (Table 1). On'i~nization of the phenolic proton, DAT can be represented as a resonance hybrid of structures such as II and III in Figure 5, and the observed shielding of NB might be attributed to some contribution of structure III. To obtain an estimate of the NB shift expected for an ionized hydrazone of structure III, a model compound, the anionic form of '10-phenylazo-9-phenanthrol which is known to exist completely in the hydrazone form, 22 was synthesized and its 15 N shifts were measured in dimethyl sulfoxide. Its observed NB shift of -33.8 ppm (Table I) indicates that NB resonance of an ionized hydrazone is considerably shielded relative to that of azobenzene (-133 ppm) and suggests that the observed shielding of NB resonance of DAT from -101,4 ppm to -78.l ppm on ionization of phenolic proton can be attrtbuted to some contribution from structure III. In water, the ionized DAT shows NB shift (-82 ppm) very close 36 to the corresponding shift in dimethyl sulfoxide/water (Table I), suggesting that in water also, the resonance structure III (Figure 5) makes contribution to the phenolate species of DAT. CONCLUSION The results presented here clearly indicate that the 15 N shifts of azo nitrogens are highly sensitive to both complexation with zinc and to hydrogen-bonding and are therefore useful probes for studying such interactions in model systems and biomolecules. The present study demonstrates that it is feasible, by introducing an 15 N-enriched probe, to study the active-site conformation of an enzyme such as carboxypeptidase A (CPA), having molecular weight of -v34,500, by 15 N NMR, despite the low sensitivity of 15 N nucleus and line broadening expected for slowly tumbling macromolecules. The 15 N shifts of an arsanilazo group, specifically coupled to the active site tyrosine-248 of CPA with retention of full catalytic activity, serve as sensitive probes of the possible complexation of azotyrosyl-248 with the catalytic zinc in the enzyme in solution, In arsanilazotyrosyl-248 CPA (Zn.Azo-CPA), the 23.5 ppm and 8.3 ppm shielding observed for the NB and Na nitrogens relative to those of apo.Azo-CPA at pH 8,8 clearly indicate that complexation of zinc with azotyrosyl-248, with NB as ligand to zinc is taking place, and that such complex constitutes the molecular 36 to the corresponding shift in dimethyl sulfoxide/water (Table I), suggesting that in water also, the resonance structure III (Figure 5) makes contribution to the phenolate species of DAT. CONCLUSION The results presented here clearly indicate that the 15 N shifts of azo nitrogens are highly sensitive to both complexation with zinc and to hydrogen-bonding and are therefore useful probes for studying such interactions in model systems and biomolecules. The present study demonstrates that it is feasible, by introducing an 15 N-enriched probe, to study the active-site conformation of an enzyme such as carboxypeptidase A (CPA), having molecular weight of -v34,500, by 15 N NMR, despite the low sensitivity of 15 N nucleus and line broadening expected for slowly tumbling macromolecules. The 15 N shifts of an arsanilazo group, specifically coupled to the active site tyrosine-248 of CPA with retention of full catalytic activity, serve as sensitive probes of the possible complexation of azotyrosyl-248 with the catalytic zinc in the enzyme in solution, In arsanilazotyrosyl-248 CPA (Zn.Azo-CPA), the 23.5 ppm and 8.3 ppm shielding observed for the NB and Na nitrogens relative to those of apo.Azo-CPA at pH 8.8 clearly indicate that complexation of zinc with azotyrosyl-248, with NB as ligand to zinc is taking place, and that such complex constitutes the molecular 37 basis for the spectral properties of Zn.Azo-CPA observed in this pH range by visible absorption, circular dichroic and resonance Raman techniques. Moreover, the high sensitivity of 15 N shift to 8 coordination with zinc permits a semi-quantitative estimate of the degree of azotyrosyl-248-Zn complexation which is calculated to be 54"--63% at pH 8.8. Thus the major conformation of Zn.Azo-CPA, and by implication, of CPA in solution at this pH is the one in which the active-site tyrosine is complexed to zinc, It has generally been accepted on the basis of X-ray structures of crystalline CPA and its substrate complex at pH 7.5 that the major conformation of free CPA is the one in which the Tyr-248 hydroxyl is • 17 A from the zinc and that, on binding of a substrate, a large conformational change involving an inward movement of the Tyr-248 hydroxyl from the surface of the molecule toward zinc takes place. This has been regarded as one of the most striking examples of the induced fit theory of enzyme-substrate binding. The 15 N NMR results indicate that, in the absence of substrate binding, the Tyr~248 can assume a conformation in which its hydroxyl is close to the catalytic zinc. Thus caution is required in interpreting the conformation of crystalline enzyme as being applicable to solution conformation. The 15 N shift of azotyrosine is also a sensitive probe of intramolecular hydrogen bonding of this nitrogen with phenolic 38 proton, as indicated by the large deshielding observed in the model compound DAT on disruption of the hydrogen bond by dimethyl sulfoxide. The binding of a quasi-substrate, glycyl-1_-tyrosine to Zn.Azo-CPA not only breaks the azotyrosyl-248-Zn complex but also partially disrupts the intramolecular hydrogen-bonding of N with tyrosyl-248 hydroxyl 8 as indicated by the 8.9 ppm deshielding of this nitrogen relative to that of Apo.Azo-CPA at pH 8.8. This result is consistent with partial hydrogen bonding of Tyr-248 hydroxyl, whose postulated catalytic function is to donate a proton to the leaving amine group, with the amide nitrogen of the scissile peptide bond of the quasi-substrate. The binding of an inhibitor 8-phenylpropionate causes a nearly complete disruption of the intramolecular hydrogen bond between N8 and the tyrosyl-248 hydroxyl as indicated by 28.2 ppm deshielding of NB relative to that of Apo.Azo-CPA. The greater deshielding of the N8 resonance in the enzyme inhibitor complex than in the enzyme-quasisubstrate complex indicates that the carboxylate group of the inhibitor competes much more effectively with the azo nitrogen for the Tyr-248 hydroxyl than does the quasi-substrate. 39 REFERENCES 1. a. J. F. Riordan and B. L. Vallee, Biochemistry £, 1460 (1963) b. R. T. Simpson and B. L. Vallee, Biochemistry ~' 1760 ( 1966) c. J. P. Felber, T. L. Coombs and B. L. Vallee, Biochemistry l, 231 (1962) d. R. Lumry, E. L. Smith and R. R. Glantz, J. Am. Chem. Soc. 73, 4330,(1951) 2. a. G. N. Reeke, J. A, Hartsuck, M. L. Ludwig, F, A, Quiocho, T. A. Steitz and W. N. Lipscomb, Proc. Natl. Acad. Sci. USA, 58, 2220 ( 1967) b. W. N. Lipscomb, J. A. Hartsuck, G. N. Reeke, F. A. Quiocho, P. H. Bethge, M. N. Ludwig, T. A. Steitz, H. Muirhead and J. C. Coppola, Brookhaven Symp. Biol.~' 24 (1968) 3. a. J. T. Johansen and B. L. Vallee, Proc. Natl, Acad. Sci. USA, 68, 2532 (1971) b. J. T. Johansen and B. L, Vallee, Proc. Natl. Acad. Sci. USA, ZQ_, 2006 (1973) c. F. A. Quiocho, C. H. McMurray and W. N, Lipscomb, Proc. Natl. Acad. Sci. USA, .§2_, 2850 (1972) Lipscomb, Proc. Natl. Acad, Sci. USA, IQ_, 3797 (1973) 4. W. N. 5. J. T.Johansen, D. M. Livingston and B, L. Vallee, Biochemistry, lJ, 2584 (1972) 6. J. T. Johansen and B. L. Vallee, Biochemistry, li, 649 (1975) ,,. 40 7. a. R. K. Scheule, H. E. Van Wart, B. L. Vallee and H. A. Scheraga, Proc. Natl. Acad. Sci. USA 74, 3273 (1977) b. R. K. Scheule, H. E. Van Wart, B, 0. Zweigel, B. L, Vallee and H. A. Scheraga, J, Inorg, Biochem, .l_l, 283 (1979) c. R. K. Scheule, H. E. Van Wart, B, L, Vallee and H. A, Scheraga, Biochemistry,~' 759 (1980) 8. W.W. Bachov<;hin and J. D. Roberts, J. Am. Chem. Soc, 100, 8041, (1978) 9. a. J. A. Hartsuck and W. N. Lipscomb, nThe Enzymes 11 ed, P, D, Boyer, vol. III p. l (1971) b. D. s. Auld and B. Holmquist, Biochemistry .J.l, 4355 (1974) c. R. G. R. Bacon and D. C, H. Biggs, J, Chem, Soc, Perkins Trans. l, 2156 ( 1974) 10. M. Tabachni~k and H.Sobotka, J . Biol, Chem, 234, 1726(1959) 11. M. Sokolovsky and B. L. Vallee, Biochemistry~. 3574 ll966) 12. D. J. Cox, F. c. Bovard, J, P, Bargetze, K. A, Walsh and H. Neurath, Biochemistry l, 44 (1964) 13. R. 0. Duthaler and J, DI Roberts, J, Am. Chem. Soc i 100, 4969 (J 978) 14. s. B. Hendrichs, O. R. Wulf, G. E. Hilbert and U, Liddel, J. Am. Chem . Soc, 58, 1991 (1963) 15. A. Burawoy and I, Ma rkowi tsch ,:- Burawoy, J ., Ch.em. Soc ,. 36 (J 936} 42 PART II Benzoate Catalysis in the Hydrolysis of endo-5-[4'(5')imidazolyl]-bicyclo[2.2.l] hept-endo-2-yl transcinnamate: Implications for the Charge-transfer Mechanism of Catalysis by Serine Proteases 43 INTRODUCTION A 11 charge-relay" mechanism hps been postulated to be crucial to the effectiveness of operation of the "catalytic tri.ad (sertne, histidine, and aspartic acid residues}" in the enzymatic cleavage of peptide bonds by serine proteases(l~ The key feature of this mechanism (Eq. l) is formation of tetrahedral intermediate by attack of serine on the peptide carbonyl, assisted by proton transfers from serinyl OH to histidyl imidazole and of H3 of the histidyl imidazole to the aspartate carboxylate anion . That the effectiveness of the second of these transfers requires the aspartate carboxylate to be a stronger base than the histidyl imidazole has been clearly delineated by Hunkapiller et ~( 2~ However, a 15 N NMR investigation ( 3) of a-lytic protease clearly showed th.at the histidyl imidazole of this enzyme has a normal base strength and the decrease of activity that occurs at low pH values is fully consistent with protonation of this - ~ h ,.... I' 0 ~,. . . _0-H ="y"H H 01 . . . ._ / R' C /R NH / '-(!> kfrAhedral itdermediaTe -- ll l 44 imidazole rather th_M th_e asp~rtate carboxylate ani_on of the tri.ad C2 ) [2) An alternative mechanism (Eq , 2} has been proposed ( 3), which might be called a carboxylate-assisted process. This mechanism features (i) a sufficient degree of hydrogen bonding of histidyl imidazole in the tautomeric form with the proton on N3 and the aspartate carboxylate to make this tautomer the predominant one, and (ii) assistance of transfer of the serine proton to Nl of the imidazole by both hydrogen and electrostatic bonding between the resulting imidazolium cation and aspartate carboxylate. This mechanism is completed to the serine ester stage by transfer of a proton from the imidazolium cation to the nitrogen of the leaving RNH - group ( 3). The elegant model compound studies by Bender and coworkers ( 4, 5), however, seem to provide strong support for the charge-relay formulation through discovery of a very pronounced benzoate catalysis of the hydrolysis of endo-5{4 -(5 )imidazolyl]bicyclc:{2,2,l]hept-endo-2-yl 1 1 45 trans-cinnarnate lIL The postulqted roech.anis,rn i:s shown i_n ~q,1 3, wi_th water playing th.e role of sertne) the pend9,nt i:rntdazole the role of histidine, and benzoate fon Ute role of the aspartate carboxylate, - slow 131 I The argument for this mechanism is derived from the fact that, although there is no benzoate catalysis, even at 0.5 M benzoate in water, there is a steadily increasing catalytic effect with increasing dioxane content in the water. This increase in catalytic effectiveness of benzoate ion can be attributed to an increase in the pKa of the benzoic acid and, thus, a concomitant increase in the basicity of the benzoate ion (5) Exactly the same kind of argument has been used to account for the postulated effectiveness of the aspartate carboxylate in the charge-relay mechanism--the carboxylate in question being located in a "hydrophobic pocket 11 of the enzyme (l), thus leading to the expectation of a larger pKa and greater basicity of the aspartate carboxylate ( 2) We believe that there are alternative interpretations of the results. Benzoate ion in 42 mol % dioxane in water has a high catalytic effectiveness. The ratio of catalyzed rate to uncatalyzed rate is about 400 for 0.1 M benzoate (2500 for 0.5 M benzoate) 46 between pH 6 and 8 1 wh.ere th.e reported pKa is 9, 4 ( 5) 1 Nonetheless, the reported pKa value of 9.4 implies that the effective concentration of benzoate ion must be small (_calculated for a 0, 1 M solution to be 0,004 Mat pH 8 and 4xlo'" 5 Mat pH. 6 ).We have used 13c and 15 N NMR to probe the species present in the solutions at these concentrations with 4..fllethyl imtdazo le as model for I, EXPERIMENTAL SECTION The 13 C NMR spectra were determined at 25,14 MHz with a Varian XLlOO spectrometer using the pulse Fourier transfonn technique with a 35.,.µsec pulse width and 3-sec repetition rate with proton decoupling, and 12-mm sample tubes, tert-Butyl alcohol and 2H20 in a capillary provided both the shift reference and the lock signal. The 15 N NMR spectra were obtained with a Bruker WH180 spectrometer operating at 18,25 MHz, using a 55-µsec pulse width, 5-sec delay, and full proton decoupling (6\ • 1 • To remove traces of Fe 3+ and Cu 2+ , the dioxane was dis- tilled from 8-hydroxyquinoline. This procedure ensured being able to detect the nitrogen resonances of 4-methylimidazole. ( 3 , 7 ) RESULTS AND DISCUSSION The use of 13 c and 15 N NMR spectroscopy to determine the degree of ionization of carboxylic acids and imidazoles, respectively, is well established (3 , 7 •3). The results obtained for several solutions in 42 mol % dioxane in water containing benzoic acid and 4-methylimidazole at different pH (9 ) and ionic-strength values are given in Table 47 L Clearly, the concentration of benzoate ton in these solutions greatly exceeds that whJch might be expected from the renorted pK ' . a value. The discrepancy is more real th.an apparent because the reported pKa is the thermodynamic value derived with proper regard for the activity coefficients of the ions in this solvent system C5 ,lol, Con centration equilibrium constants for benzoic acid correspond i ng to pKa values of 6,78 and about 6,2 can be derived, respectively, for 0,1 Msolutions (with an approximate i.onic strength of Q, l l and 0,5 M solutions (ionic strength, 0.51 from the 13 c NMR titration curves of Fig. 1, The pKa values were not changed significantly by addi.tion of equivalent concentrations of 4-methylimidazole . The curve in Fig, 1 for ionic strength 0,5 could not be determined over the whole range of concentrations because of phase separations, but , unquestionably, there is the expected substantial decrease in pKa with increasing ionic strengths. The pKa of 4-rnethylimidazolium cation calculated from the concentrations listed in Table I is 6,4±0.2, which is, within experimental error, the same as the thermodynamic value for I reported by Komiyama et tl (5) . This is as expected for the ionization of a cationic acid, BH+ + H o I B + H+ 3o, where changes in the activity 2 coefficient of BH+ and H+ 3o with solvent and ionic strength will be reasonably parallel. The hydrolysis rate of I substantially decreases in the absence of benzoate as the proportion of dioxane is increased in the solvent. This decrease has been attributed to preferential salvation of the trans-cinnamate group by dioxane, causing steric hindrance to the 48 Table 1. 4-Methyli_mi_dazole And benzoi_c aci.d (_0,5 Meach)_ i_n 42 mol % dioxane/58% H20 at 60°c 4~Methylimidazole % imidazBenzoate o 15N' ppm * olium 13 C,ppm·a -b %C 6H5co 2 Nl N3 cationt 0,5c ~141.0 56,6 190.6 183,5 60. 1 6.4 0. 5c -141.8 71. 7 188.6 180.9 52.9 6.7 0. 5c -142.2 79.2 185.9 177, 9 44.0 6. l 0. ld -140.2 41.5 189 .4 182. 5 56.6 pH JJ 6, 1 0 * Upfield from 1 M H15 N0 3 , t Calculated from the fo 11 owing 15 N shifts of 0. 5 M 4-methyl imi dazo le (JJ=0.l) in 42 mol % ~ioxane/58% H20: 0 15N, ppm pH Nl N3 4.3 202.0 197. 7 4-methylimidazolium cation 8.5 174.2 161. 1 4-methylimidazole a. Shift of carboxylcarbon downfield of the methyl carbon of tertbutyl alcohol b. Calculated from the 13 c shifts of carboxyl carbon of 0. 1 M benzoic acid in Fig. 1 c. c H COO-Na+ (0.5 M) and 4-methylimidazole (0.5 M) dissolved in 6 5 dioxane/H 0 and adjusted to pH by addition of HCl 2 d. c H C00H (0.5 M), 4-methylimidazole (0.5 M), and KCl (0.1 M) 6 5 dissolved in dioxane/H 20 and adjusted to pH by addition of Na0H. 49 :i -140 u.. ....0 ,t! ~ ll:: ~ -142 Phase separation -144-------~-~---' 4 5 6 7 8 9 pH 13 c NMR chemi.ca-1 shi.fts of benzoate carboxyl carbon as a func..,. F,•g. 1 •. tion of pH at 60°c in 42 mo l % di oxane in water, o , 0, l M c H COOH 6 5 with 0.1 M KCl; the pH was changed by addition of Na0H ,e, 0, 5 M c6H5COO- Na+; the pH was adjusted by addition of HCl . The pH values for this solvent mixture are corrected from glass electrode readings by the procedure of Van Uitert and Fernelius ( 9}_ The shifts are in ppm downfield of the methyl carbon of tert-butyl alcohol. The curves were calculated for pKa values of 6,78 and 6. 2 for 0. 1 Mand 0. 5 M benzoate concentrations, respectively. 50 concerted a.ttack of water and i.11Jidazole 1 as i_n Eq 1 3 \ A more im.portant factor is likely to be the decreased ability of the increasingly nonpolar solvent to accommodate the charge separation produced in the fonnatton of the transition state for Eq, 4. In order for benzoate ion to be effecttve in the mechanisms such as Eq. 3, 14 1 it should be of comparable basicity to the imidazole group~ and, yet, even in 17%mol fraction dioxane in water , where benzoate is only 1/l0th as strong a base as imidazole~ the rate at 0.5 M benzoate is about 300 times faster than when benzoate is absent ( 5) It is our view that benzoate ion can increase the hydrolysis rate of I by virtue of acting much like the asparate carboxylate group in the catalytic triad of a serine protease: first, by hydrogen bonding to the imidazole before the transition state, which will make the imidazole ring more nucleophilic, and, second, by additional stabilization of the transition state by a further degree of hydrogen bonding associated with building up the imidazolium charge as the result of proton transfer (Eq . 5}. In addition, there are likely to be secondary influences of benzoate ion at these large concentrations 51 M M e,,,< M 15i - I ·( \_ :N~"'-M R M which contribute to the general polarity of the medium. That the combination of adding benzoate and dioxane could make the reaction rate faster than in pure water is in accord with the speculations of Warshel (ll). Formulations such as given in Eq. 5 also are supported by the failure to obtain evidence for a two-proton transfer mechanism in the benzoate-assisted hydrolysis of I, * although such two-proton transfers have been suggested by proton-inventory studies for some, but not all, reactions catalyzed by a-lytic protease (l 2) and trypsin (l 3 ) We thank Profs. J. H. Richards and W.W. Bachovchin for many helpful suggestions. This is contribution no. 6173 from the Gates and Crellin Laboratories of Chemistry. These studies were supported by U.S. Public Health Service Grant GM-11072 from the Division of General Medical Sciences and by the National Science Foundation. * Bender, M. L., presented at the Fourth IUPAC Conference on Physical Organic Chemistry, York, England> September 4-8~ 1978, 52 REFERENCES 1. D, M, Blow, J, J, Bi:rktoft, and B, S, Hartley (19.69} Nature (.London) 221, 337, 2. M. W, Hunkapiller, S, H, Smallcombe, D, E, Whitaker and J, H. Richards (1973} Btochemistry JI~ 4732. 3, W, W, Bachovchin and J, D, Roberts (1978} J. Am, Chem._ Soc. 100, 8041. 4, M. Korniyama, T. R. Rosel and M, L, Bender (.1977} Proc. Natl . Acad, Sci. USA 74, 23, 5. M. Komiyama) M, L. Bender! M, Utaka and A, Takeda (.1977} Proc. Natl, Acad, Sci. USA 74, 2634 . 6. D. Gust, R. B. Moon and J, D, Roberts (1975} Proc . Natl . Acad , Sci. USA ?.1_, 4696, 7. I. I. Schuster and J. D, Roberts (1979} J, Org. Chem, 44 3864 Am. Chern. Soc. 2}_, 4504. 8. R. Hagen and J. D, Roberts (J 969). J, 9. L. G. Van Uitert and W. C. Ferneli.us (1954). J , Am. Chem . Soc, 76 , 5887. 10, L. G, Van Uitert and C, G. Haas C,953)_ J, Am. Chem, Soc, ~. 451, 11. A. Warshel (1978). Proc, Natl .. Acad, Sci. USA 75, 5250. 12. M. W. Hunkapiller, M, D, Forgac and J. H. Richards (1976). Biochemistry ~. 5581. 13. R. L. Schowen (1977} in Solvent Isotope Effects on Enzyme-Catalyzed Reactions, eds . Cleland, W.W. ,O ' Leary, M. H. and Northrop, D. B. (University Park Press, Baltimore, MD}, pp, 64~99 53 PART III Studies of pH and Anion Complexation Effects on !:_-Arginine by Natural Abundance 15 N Nuclear Magnetic Resonance Spectroscopy 1 54 INTRODUCTION Several enzymes have been found to contain arginyl residues which appear to interact with negatively charged phosphate or carboxylat~ groups of substrates or cofactors. 2 Chemical modification of positively charged arginyl residues (and the resulting loss of charge) has been shown to result in deactivation of the enzymes 2. Riordan and co-workers have further suggested that enzymes which interact with anionic substances or cofactors will probably contain arginyl residues at the active site. 3 The binding of DNA with histone and protamine is considered to involve electrostatic interactions between the phosphate anions of DNA and the positively charged arginyl and lysyl residues of these nucleoproteins. The guanidinium group of methyl- guanidine has been shQWn by x-ray crystallography to complex with a phosphate anion through multiple hydrogen bonds. 4 The complexation of b-arginine with phosphate and chloride ions in aqueous solution has been studied by 31 P.and 35 c1 nuclear magnetic resonance spectroKatz and co-workers 5 have reported a downfield shift of 0.5 ppm in 31 P chemical shift of methyl phosphate on addition of k-arginine scopy. in D20 solution at pH 6.15. Jonsson and co-workers 6 have measured the pH dependence of 35 c1 chemical shift in an aqueous mixture of NaCl and k-arginine and observed a distinct change in 35 c1 shift on deprotonation of the guanidinium group. The present study was designed (1) to determine a complete 15 N chemical shift-pH titration curve for 55 k-arginine,and (2) to investigate the utilit,y of 15 N NMR for detection and characterization of complexations of guanidinium ion (1) and barginine (2) with fluoroborate ions, chloride ions, phosphate ions, and ATP in water solutions, guanidinium ion, .!_ arginine, .,.... 2 EXPERIMENTAL SECTION Natural-abundance 15 N NMR spectra were recorded using the pulse Fourier.-transform technique with a Bruker WH-180 spectrometer operating at 18 ,25 MHz. An external l M H15 NO/H 20 capillary reference was used in 25-mm spinning sample tubes, Normal operating condi.tions for guanidine carbonate and arginine solutions (for spectra other than T1 measurements) were 55-µs pulses (70° flip angle) with a 15-,.s pulse delay, and for protamine solutions were 60-µs pulses with a 0.58-,-s delay . With full proton decoupling, the sample temperatures were about ss 0 c, Spectra were obtained of 2 M aqueous solutions of guanidine 56 carbonate with internal D20 field ... frequency locks, and the b-arginine spectra were similarly taken of l Maqueous solutions, sulfate spectra were of 0.0315 M aqueous solutions. Protamine k-Arginine was purchased from Sigma Chemi.cal Co,, and guanidine carbonate from Matheson Coleman and Bell, Clupeine, a protamine i.solated from herring, was obtained as protamine sulfate from the Sigma Chemical Co, 15 The N T1 values were obtained u~ing the fast inversion recovery (FIRFT) technique. 7 Before measurement of the T values, the so.,. 1 lutions were treated with Chelex-100, using the method of Irving and Lapidot, 8 and purged with nitrogen. The pH of each solution was determined with a Radiometer PHM-26C pH meter using a -::ombined glass reference electrode and adjusted with hydrochloric acid and potassium hydroxide. Viscosity measurements were made in an Ostwald viscosi.- meter immersed in a constant-temperature bath. RESULTS AND DISCUSSION Aqueous l Msolutions of h-argini.ne show three 15 N absorptions. Of these, the resonance at 304 ppm can be assigned to the two equivalent guanidino nitrogens through comparison with the resonance in guanidine carbonate (301,7 ppm); the a-amino nitrogen resonance is surely the one at 342 ppm through comparison with a ... amino chemical shifts for other amino acids, and the -NH-absorption is, by elimination, the one at 289 ppm. (A compatible, more rigorous, assignment of peaks has been reported by Pregosin and co-workers. 9 ) 57 "'',·~ \ 10 l .---. ~ •• pHr 5 , ( ~ { ) /. _,.. / i• .. ~ I (·•·"·• I I I r • , i-NH· I •• \ - •I 21S 295 305 15N che111lcal shift, pp n, • - Q a111lno 335 0 345 figure l. The pH dependence of 15 N chemical shifts of b~~rginine at 25 0 c. The chemical shifts are tn parts per million upfie~d of H15 No 3. The accuracy of pH values is ±0,05 pH unit for pH:_ 13,5, For pH > 13.5, the error may be as much as 0,1-0.3 pH unit because of the high potassium ion concentrations, 58 The pH dependence of the chemical sh.Ht of each of the three resonances is shown i_n Figure l. The a . . ppm downfield shift observed for the a~amino nitrogen at 342 ppm as the pH is lowered from 10.6 to 7 .3 (.pKa = 9,0 for the a ... amino group of b--arginine) 10 is about as expected from decreased shielding of a saturated amine nitrogen on protonation. The 15-ppm upfield shift of the guanidino nitrogens and the 3-ppm upfield shift of the ..--NH ... nitrogen, which result when pH is lowered from 14.l to 12.0 and protonation occurs, are probably due to a change in the second-order paramagnetic effect of the neutral guanidino nitrogens. Previous studies 11 of the pH dependence of the 15 N chemical shift of 1-arginine found a similar shift in the highly basic region, However, it was also reported that above pH 13,5 the guanidino nitrogen resonance separates into two resonances, one at approximately 285 ppm and the other at 342 ppm, while the a-amino resonance could not be observed above pH 5.8, The present work and a recent paper by Blomberg and co-workers 12 show clearly that the single resonance at 288 ppm observed in the highly basic region represents the neutral guanidino nitrogens, and that the 342. ppm resonance represents the a-amino nitrogen; the quenching of this nitrogen resonance near its pK a was almost certainly- the result of enhanced relaxation arising from traces of paramagnetic ions, as has been observed for glycine, 13 On protonation of the carboxyl group (pKa = 2.2), the a~amino nitrogen resonance shifts 0,6 ppm upfield. This may be due to in- 59 creased shi_eldi_ng of th.e nitrogen on loss of hydrogen bondi.ng between the ... NH 3+ group and the -coos:- groups as has been suggested by Blomberg and co ...workers 12 for glycine, The 0.7-ppm upfield shift of the -NHgroup on protonation of thea-amino group (.PKa = 9.0)_ may be similarly ascribed to increased s·hielding resulting from the loss of interactions, such as hydrogen bonding, with the neutral a-,,ami_no group , To determine the effects of anions on the resonance of guanidine and b-arginine nitrogens, equivalent amounts of salts or acids were added and the shifts compared to those produced by fluoroborate, which was not expected to complex significantly- in aqueous solution. As can be seen from Table I, no very significant changes in shifts were observed on addition of chloride ions. For phosphate-arginine mixtures, the 15 N shifts were measured at pH 6, 3 at arginine-phosphate mole ratios of 1:1, l :2, and l :3 and compared to those produced by fluoroborate . The guanidino nitrogens show a small downfield shift of 0.8 ppm on addition of equimolar phosphate, which may be due to a change in the second-order paramagnetic effect of the guanidino nitrogens on hydrogen bonding with phosphate oxygens. Increasing the phosphate-arginine mole ratio, however, caused no further increase in guanidino nitrogen shifts. Possible effect of phosphate ani.ons on the arginyl residues of protamines was investigated by comparing the nitrogen shifts of arginyl residues in clupeine, a protamine isolated from herring, in 14 the presence and absence of ATP, This protamine has a molecular weight of approximately 5000 and contains 20 arginyl residues per 60 TABLE I. 15 N NMR Shiftsa of Guanidi.ne Carbonate, k-Arginine, and Protamine Sul fate i.n Presence of Antons sample (mole ratio) alone NaCl (l:l) . . . . pH l M arginine guanidino guanidi.ne carbonate 11. l 30 l . 7, 301.0 290.3 289.7 304.4 304 . 0 342.7 342.7 HP04 2- (l :l) 10.6 300.6 289.6 303.3 341.4 HBF 4 ( l: l) 8.9 302.7 291. 2 304 . 2 336. l b HP04 2- and 7. 3 290 . 5 303. l 334. 4 b 290.8 291 .0 290.7 303.8 303.7 304 . 0 e 335.l b 334.8 b 6,3 d 290 . 4 303. 2 e 334 . 4 b 6.3 d 290.4 303.0 e 334.3 b 6,3 d 290,6 303. l e 334.5 b Protamine 2.8 290.2 9 303.5 h alone f ATP ( l : l) i 2.8 290.4 9 H/0 -c (l:l) 4 HCl (l:l) ATP ( l : l) HBF 4 HP0 2- and 4 11. 5 6,0 4.4 6,3 d 301 .7 R/0 4- (l:l) HP04 2- and H/0 4- (l :2) HP04 2- and H/0 4- (1:3) a Chemical shifts are in parts per million upfield from nitric acid. b This shift, compared to those above, results from the change in pH 61 from 11,5 on adding acid (iee Figure ll, c HCl was added to adjust pH to 7,3, d pH adjusted by Na0H, e. Precision is about ±0,l ppm, f Protamine sulfate at 0,03 M. This substance was too insoluble in H2o to take its 15 N spectrum at higher pH values, g Chemical shift of arginyl ~NH ~ resonances in protamine sulfate, h Chemical shift of arginyl guanidino nitrogens of protamine sulfate. i Arginyl residue: ATP= 1:1. 62 molecule. As can be seen i_n Table!, no significant ch.i'!nges in shifts were observed on addi.tion of ATP, If complexation of anions with guanidinium nitrogens does i.n fact occur in water, the T1 relaxatton ti.mes would be expected to be shortened. The T1 data obtained in the presence and absence of various anions are summarized in Table II. It appears that the pH effect on the nitrogen T1 values is much more signi:ficant than the effects arising from complexation with chlori.de anions. Protonati:on changes are not expected to occur in the pH range studied here for guanidino and -NH-nitrogens which have pKa = 12 . 48 1 lO but possible conformation changes resulting from the deprotonatinn of the Q~amino nitrogen (pKa = 9,0) and its interaction with the guanidin-fum group may contribute to the observed change in T1 of guanidino and -NH~nitrogens between pH 6,0 and 9.4. The T1 changes may also be due to traces of heavy-metal ions not removed by the Chelex treatment, and such changes in amino acid relaxati_on times with pH associated with heavy metal ions are well documented . 8, 13 Trace metal ions were in fact responsible for the quenching of the a,-amino resonance at high pH; after a second treatment with Chelex and addition of 10 mM EDTA, the a-amino resonance was clearly observed at pH 10.1. On addition of ATP to the arginine solutions, the viscosity increased substantially. In view of the well-known effects of viscosity on T , the T1 ts of the arginine nitrogens in the presence of ATP were 1 compared at pH 10,l with those in arginine solutions made isoviscous TABLE I I. T1 Values for b-Arginine Nitrogens in the Presence of Anions concn, a SAMPLE no addends k-arginine with NaCl !,;-arginine with HC1 .L-arginine with HC1 no addends c ,k-arginine with glycerol c ~-arginine with ATP c temp, viscos ity' cP Tl, s - M EH OC 1.0 1 .o 10. 1 9.4 56 58 0.66 0.80 guanidino 8.3 7.7 1.0 6.0 51 0.82 4.8 9.7 7. l 1.0 2.7 53 0.85 6.6 9.7 6,9 -NH- a-amino 14,0 13. 5 b b Ol w 2.5 1.0 0.6 each 10. 1 10. 1 51 51 1.33 1.35 2.9 3.6 5.7 6.6 4.5 4.2 10. 1 49 1.32 2.5 5,9 2.3 a Concentration of k-arginine unless otherwise specified, b No resonance observed. c 10 mM EDTA added. 64 with (l) increasing arginine concentration and (Jl addition of ~lycerol , The results provide no evidence for significant changes in T1 due to complexation with ATP, as can be seen in Table II, The low sensitivity· of guanidino and arginyl nitrogen ch.emi.cal~ shift and T1 data for aqueous solution toward anionic substances, which have been postulated to complex with such residues -of enzymes, do not necessarily mean that 15 N NMR cannot be useful for such studies, Arginine in water may well have reduced tendency for complexation compared to arginyl residues in enzymes and, in fact, Borders and co~ workers 15 have shown that free arginine reacts wtth 2,3 ... butanedi.one about 15 times more slowly than the essential arginY'l residue of creatin kinase. Hydrophobic envi.ronments near the active site of the enzyme could very well enhance the binding capabilities of arginyl residues toward anions, and this idea wtll be investigated further, 65 REF ERE NC ES AND NOTES Supported by th.e Nati.anal Insti.tutes of Health., PubHc Heqlth Service, Grant GM.,...11073 from th.e Divtsi.on of General Medical Sciences, and by the National Science Foundation, 2. L. Patthy and E, L. Smith, J . Biol . Chem, 250, 565 (1975), and references cited therein, 3. J. F. Riordan, K, D. McElvany , and C, L. Borders, Jr., Science, 195, 884 (1977}. 4. F. A. Cotton, E, E, Hazen, Jr , t V. W. Day, s . Larsen, J , G, Nonnan, Jr., S. T. K. Wong, and K, H, Johnson, J. Am, Chem . Soc, 95 2367 (1973) . 5. R. Katz, H, J. c. Yeh, and D, F. Johnson, Bioch.em . Biophys . Res. Conmun , 58, 316 (.1974 ). 6. B. Jonsson and B, Lindman, FEBS Lett, 78, 67 (1977) , 7. The FIRFT method of T1 measurement compares quite favorably with the standard inversion recovery method (and less favorahly v-lith the progressive saturation method) in studies carried out in· this laboratory . See D. Canet, G. C, Levy, and I. A. Peat, J. Magn. Reson. 1§_, 199 (1975). 8. C. S. Irving and A. _Lapidot, J. Am. Chem. Soc.'}]_, 5945 (1975). 9. P. S. Pregosin, E. W. Randall, and A. I . White, Chem. Commun. 1602 ( 1971) . 10. H. E. Mahler and E. H. Cordes, Biological Chemistry Harper and Row, New York, N. Y. 1966, p. 34 . 11. T. Suzuki, T. Yamaguchi, and M. Iamanari, Tetrahedron Lett., 1809 (1974). 12. F. Blomberg, W. Maurer, and H. Ruterjans, Proc. Natl. Acad. Sci. U.S.A., 73 1409 (1976). 13. (a) T. K. Leipert and J. H. Noggle, J. Am. Chem , Soc. C£]__, 269 (1975); (b) H. Pearson, D. Gust, I . M. Armitage, H. Huber, J. D. Roberts, R. E. Stark, R. R. Void, and R. L. Void, Proc. Natl. Acad. Sci. U.S.A. ?1_, 1599 (1975). 1. 11 11 , 66 14, The protamine-DNA complex could not be studied because of low solubilHy i_n water, See D, E, Olins~ A, L, Oltns, and P, H, von Hippel, J, Mol, Biol~ §_, 265 (.1968)_, 15. C, L. Borders, Jr., and J, F, Riordan, Biochemi.stry ,!i, 4699 C,975). 67 • PART IV A 15 N Nuclear Magnetic Resonance Study of Amino Acid Metabolism in Neurospora crassa 58 INTRODUCTION Isotopic tracers, radioactive and stable,of C, H, and P have been one of the most valuable tools for studying the pathways and rates of metabolic processes. For nitrogen, which lacks a radioisotope of adequately long lifetime, 15 N is of special importance as a tracer of nitrogenous metabolites, and 15 N nuclear magnetic resonance spectroscopy is potentially a very powerful technique for use of this isotope. The advantage of nuclear magnetic resonance spectroscopy in stable isotope labeling experiments is that metabolic processes can be studied .i!!_ vivo in intact organisms. Thus the uncertainties encountered in the extraction and separation of labeled metabolites necessary for the detection of the label by mass spectrometry, can be avoided. The advantage of nuclear magnetic resonance spectroscopy in studying in vivo systems has been well demonstrated in the application of 31 P NMR for the measurement of intracellular pH 1 and for study of adenosine triphosphate metabolism in intact cells and tissues, 2 and in recent 13 c NMR studies of glycolysis in microorganisms 2 . There have already been a few reports of 15 N NMR studies of intact organisms, although difficulties have been encountered because of 69 the low sensitivity and relatively long relaxation times of the 15 N nucleus. J. Schaefer and coworkers 3 used cross-polarization solidstate 15 N NMR to study the change in the relative concentrations of amino acid residues in proteins at various stages of growth of 15 N-labeled soybeans. M. Llinas and coworkers have reported the observation of a single peptidyl amide nitrogen resonance in 15 Nenriched culture of fungus Ustilago sphaerogena 4 . A. Lapidot and C. S. Irving have studied the dynamic properties of cell wall components of intact bacteria cells by 15 N NMR 5. In these works on microorganisms, only 15 N resonances of macromolecules such as proteins and cell-wall peptidoglycan were detected. So far, no 15 N NMR studies of small metabolites such as amino acids in intact microorganisms have been reported. While a wealth of valuable information has been obtained in the past three decades regarding the intermediates and the enzymes involved in amino acid metabolism and the mechanisms of metabolic regulation, there remain intriguing questions about the actual functioning of metabolic processes and their regulation in vivo. Some of the questions that can be studied by labeling the metabolites of micro- 70 15 N and tracing their fate i.!!_ vivo by 15 N NMR follow. • an,·t rogen source, sue h as 15 NH + , d"1str1"b uted among l . How 1s 4 competing anabolic pathways? organisms with 2. Where alternative pathways from a precursor to a product exist, which pathway is favored? 3. What are the turnover rates of nitrogenous metabolites in vivo? The aim of this study is to explore the potential of 15 N NMR as tracer of in vivo nitrogen metabolism in microorganisms. The organism chosen for the study is fungus Neurospora crassa which has the following advantages. 1. It is easily cultured in the laboratory. Wild-type Neurospora can biosythesize all nitrogenous metabolites from NH 4Cl as the sole nitrogen source, 2. It is among the genetically most extensively studied organisms; a number of well-characterized mutant strains, each deficient in an enzyme for a specific metabolic reaction, are known. 3. Unlike procaryotes in which intracellular pools of amino acids are maintained at low levels, this eucaryotic organism contains large free pools of several amino acids 6 , This facilitates the study of their metabolic fates by 15 N NMR. This paper describes some preliminary results which demonstrate the feasibility of observing small 15 N-labeled metabolites such as g~utamine, glutamate and arginine in intact N, crassa mycelia and the 71 types of information on in vivo metabolism that can be obtained 15 by N nuclear magnetic resonance spectroscopy. The following experiments are described. 15 1. The N chemical shift, the spin-lattice relaxation time T , 1 and nuclear Overhauser enhancement (NOE) of intracellular arginine (enriched in 15 N at Nw,w ,) in intact Neurospora mycelia were measured to determine the optimum NMR operating conditions for observing 15 N resonances of small metabolites in intact mycelia. 2. 15 N spectra of !i• crassa mycelia grown in minimal medium with 15 NH 4Cl as the sole nitrogen source were obtained, and the distribution of 15 NH 4+ into various anabolic pathways of nitrogenous cell components was studied under different growth conditions. The turnover times of some 15 N-labeled metabolites were studied by the isotope dilution method. 3. The time course of assimilation of 15 NH 4+ into 15 N-labeled metabolites was studied in vivo. With aeration, the mycelia were found to metabolize at an observable rate in a "miniature culture" within the NMR tube. The time dependence of the fonnation of glutamine and glutamate was compared in two Neuro&pora strains, one capable of synthesizing glutamate directly from 15 NH 4+ and a-ketoglutarate and one incapable of this synthesis, with the purpose of finding out whether an alternative pathway of glutamate synthesis via glutamine occurs j__!!_ vivo. 72 EXPERIMENTAL The mutant strains of Neurospora crassa used in these experiments are ure-1 allele 9 (FGSC 1230) (ureaseless) and aga UM906 (FGSC 1699) (arginaseless) obtained from R. H. Davis, and am allele 32213 (FGSC 521) (glutamate dehydrogenase deficient) obtained from the Fungal Genetics Stock Center at Arcata, California. The medium used was Vogel 1 s minimal medium supplemented with 1.5% sucrose 7 . When we speak of a 11 nitrogen-free medium 11 , we mean the same medium without ammonium nitrate. The 15 NH 4Cl (99% enriched in 15 N) and .!:_-arginine hydrochloride (99% enriched in 15 N at Nw,w ,) were purchased from KOR Isotopes. Cultures were inoculated with an aqueous suspension of washed conidia to a final concentration of lxlo 7 conidia/ml. Conidia were genninated in l liter baffled flasks containing 500 ml of minimal medium with aeration provided by shaking at 150 rpm at 16°c for 16 ~ 20 hours. The temperature was then raised to 30°c for 3 hours or until logarithmic growth was established as judged by turbidity. In some cases, mycelia were exposed to 15 NH 4 Cl (0.2%) for the entire germination period. Alternatively, the mycelia were transferred to nitrogen-free medium for 3 hours prior to the addition of 15 NH 4 Cl. Where specified, cycloheximide was added to the culture medium to a final concentration of 20 µg/ml, a few minutes before the addition of 15 NH 4 Cl or 15 N-labeled arginine, to inhibit the incorporation of 15 N-labeled amino acids into proteins. The mycelia suspension for the NMR measurements was prepared 73 by collecting the mycelia by filtration and resuspending them in enough medium to make 18 ml of mycelia suspension in 25 mm-diameter NMR tube. The volume ratio of mycelia pellets to the medium in NMR sample was approximately 1:1, unless otherwise specified. The 15 N NMR spectra were obtained with a Bruker WH-180 spectrometer operating at 18.25 MHz. Chemical shifts are reported in ppm upfield of al Msolution of H15 No 3 made by dissolving in o2o and used as an external standard. The operating conditions, unless otherwise specified, were 70 µsec pulses (90° flip angle) with a 2-sec delay and with full proton decoupling. The sample temperature was maintained at 10 ± 2°c. To check the viability of mycelia during the NMR experiments, the doubling time and the ability to take up arginine from the medium were compared in mycelia before and after exposure to similar environment; no change was observed in either cellular property. that the mycelia are fully viable under these conditions. We conclude 74 RESULTS AND DISCUSSION 1. The 15 N shift, the spin-lattice relaxation time and the nuclear Overhauser enhancement of intracellular arginine ( 15 Nw,w ,) in N. crassa Wild-type!!_. crassa, cultured on minimal medium with ammonium nitrate as the nitrogen source, accumulate large pools of amino acids such as arginine (8 mM in cell water) and ornithine. 8 On addition of high levels of arginine to the growth medium, the intracellular arginine pool expands; the increase is several-fold during steadystate growth in arginine-supplemented medium 9 . To test the feasibility and determine the optimum conditions for observing small nitrogenous metabolites in!!_. crassa, it is advantageous to start with mycelia which have incorporated arginine (15 N-labeled at Nw,w ,) and in which the intracellular level of free arginine is maintained at a constant level, without degradation by arginase or incorporation into cellular proteins. Thus the~ (arginaseless) strain of · N. crassa was grown in 500 ml of minimal 15 medium and 200 mg of L-arginine ( N-enriched in N 1 ) was added = w,w to the culture medium after addition of cycloheximide. After approximately 90% of arginine ( 15 Nw,w ,) had been taken up (1-2 hours), the mycelia were collected by filtration, washed extensively with water and resuspended in arginine-free medium for NMR measurement. Figure 1 shows the proton-decoupled 15 N spectrum of intracellular arginine ( 15 N ,) in intact N. crassa mycelia. w,w - An intense 15 N 75 280 300 320 340 ppm Proton-decoupled 15 N chemical shift spectrum of intracellular arginine ( 15 Nw,w 1 ) in intact N. crassa mycelia. The spectrum was obtained in 100 transients, with 70° pulse and 4-sec delay. Figure l. 76 resonance at 303.5 ppm is observed after only 100 scans (7 minutes of accumulation time) lO The 15 N shift of 303.5 ppm for intracellular 15 arginine ( Nw,w') in N. crassa is only 0.3 ppm downfield of that of arginine ( 15 Nw,w' ) in the mycelia free medium (Table 1). Thus an intracellular environment causes no significant change in the 15 Nw,w' shift of arginine. These results demonstrate the feasibility of observing small 15 N-labeled metabolites in intact mycelia by 15 N nuclear magnetic resonance spectroscopy. The observation is facilitated by (l) the high level of free intracellular amino acids in eucaryotic cells and (2) the high sensitivity of our Bruker WH180 spectrometer which has been described elsewhere 11 . The high sensitivity gives a good signal-to-noise ratio in a relatively short time even when a relatively long repetition rate of 4 sec is used. In previous 15 N NMR studies of microorganisms, 4 , 5 15 N resonances of small metabolites were not detected presumably because these nuclei have relatively long relaxation times and were saturated as a result of using fast repetition rates. To obtain information on the correlation time Tc of intracellular arginine and to estimate the optimum NMR operating condition for obtaining its 15 N spectra efficiently, the nuclear Overhauser enhancement (NOE) and the spin-lattice relaxation time (T 1) of 15 of intracellular arginine were measured. 15 Figure 2 compares the N spectra of intracellular arginine Nw,w' 77 (a) (b) Figure 2. 15 N spectra of intracellular arginine ( 15 N w,w' ) in N• crassa for NOE measurement, obtained in 80 transients with 90° pulse and 32-sec delay (a) with gated proton decoupling (b) with continuous proton decoupling. 78 obtained with gated proton decoupling (2a) and continuous proton decoupling (2b) for NOE measurement. The NOE value, calculated from the ratio of the two peak intensities is -3.6 ± 0.2. The NOE value can be used to determine the dipolar contribution to relaxation by the equation NOE T =T max ldd lobs NOEobs ( 1) where Tlobs is the measured T1 value and Tldd is the dipolar contri9 bution to T1 . Substituting NOEmax = -3.96 and NOEobs = -3.6, we obtain Tlobs Tldd ~ - = 0.92 This indicates that the relaxation of 15 N~,w• of intracellular arginine is predominantly by dipole-dipole interaction with protons, with little contribution from · other relaxation mechanisms, On the assumption that only dipolar relaxation to protons is important and that isotropic reorientation depends on the correlation time Tc' the dependence of NOE on Tc at 42 kG has been calculated by D. Gust, R. B. Moon and J. D. Roberts 11 and is shown in Figure 3a. From the observed NOE of -3.6 for Nw,w , of intracellular arginine in~- crassa, its correlation time can be estimated to be less than 4xlO-lO sec . The spin-lattice relaxation time T1 of 15 Nw,w , of intracellular arginine in N. crassa was measured by the fast-inversion-recovery 79 3a Correlation time (Tel in sec 10-2 '----''--__.__ 10-12 _.__.....__~~-- Correlation time (sec) Figure 3a. Plot of NOE against Tc for nitrogen-15, assuming dipolar relaxation from protons and isotropic rotation. (After Fi_gure 3 i_n reference 11) Figure 3b. Plot of r 1 against Tc for nitrogen-15 at 18.25 MHz, as- suming dipolar relaxation and isotropic rotation:-, relaxation to 0 one proton with internuclear distance of 1.02 A; ----, relaxation to 0 two protons with internuclear distance 1.03 A; -· -· - relaxation to 0 three protons with internuclear distance 1.03 A. (After Figure 4 in reference i1) 80 12 method, using the pulse sequence (180° -T- 90° - T) . Some n representative spectra from which the signal intensity S was measured as a function of Tare shown in Figure 4. Semilogarithmic plots of (S S) vs Tare shown in Figure 5. The T1 value of 15 NW,W T of intracellular arginine calculated from the slope of the computer00 - 1 fitted least-squares line through the plots is 1.06 sec, with a correlation coefficient of 0.9986. The dependence of T on correlation 1 11 time Tc at 42 kG has been calculated , assuming isotropic reorientation and dipolar relaxation to protons, and is shown in Figure 3b. From the observed T1 of 1.06 sec for 15 Nw,w , of intracellular arginine in ij, crassa, its correlation time can be estimated to be approximately lxlo-lO sec. The large negative NOE of -3.6 and the relaxation time of 1.06 sec observed for the Nw,w , of intracellular arginine indicate that to obtain spectra of protonated nitrogens of arginine, and by inference, of other small metabolites in _!i. crassa, it is advantageous to use continuous proton decoupling and a repetition rate of a few seconds. These operating conditions were used throughout the remaining experiments. The r value of 1.06 for the 15 Nw,w , of intracellular arginine in 1 N. crassa is considerably shorter than the T1 value of 4.25 sec obtained for 15 Nw,w of 12 mM arginine in the mycelia-free medium at the same temperature (Table I). Assuming dipolar relaxation, T1 I depends on molecular size and the viscosity of the medium, because -r = 1 sec Figure 4. Representative 15 N spectra for the measurement of T1 of 15N , of intracellular arginine in!!_. crassa. The pulse sequence We,W (180 O --r- 90 - T)n . was used, with T ;:: 4 sec and 1, 2, 4, and 32 sec. transients. -r;:: Q t 1, 0.3, 0'. 7~ Each spectrum represents an accumulation of 80 82 -- l-' V) 8 V) s:: ,.... 1 2 3 4 -r (sec) Figure 5. of 15 N W,W Plots of ln(S00 I S-r) against -r for the T1 measurement of arginine in Neurospora. S1: represents the signal intensity at each -r value shown, and S00 is the signal intensity at -r = 32 sec. From the s l bpe of the 1east-squa.res 1i ne through the plots, T1 of 1.06 ±0.02sec was obtained, 83 l --cx:T Tldd c in mobile liquids where the relation T!ff(wN + wH) 2 << l holds~ and Tc is given to a first approximation by Tc 3 = 4TTn a 3kT where n is the viscosity of the liquid,~ is the radius of the molecule (assumed to be spherical), Tis the absolute temperature and k is the Boltzmann's constant. Thus the shorter Tldd' and the longer Tc' 15 of Nw,w , of intracellular arginine could be due to the higher viscosity of intracellular environment and/or association of arginine with other cellular components which increases the effective molecular radius a. It has been demonstrated that the bulk intracellular pool • • • N. crassa 1s • seques t ere d 1n • vacuo l es 13 . of arg1n1ne ,n Because the subcellular fraction representing the vacuole contains large quantities of polyphosphates as well as arginine, it has been speculated that the positively-charged guanidino groups of arginine are associated with the negatively-charged polyphosphates in the vacuole; the estimated concentration of arginine in the vacuole is approximately lM and that of polyphosphate equivalent to approximately lM of monophosphate Unl. t S 14 . In this regard, it is significant that a r value of 0.96 1 sec was obtained for lM aqueous solution of arginine in the presence 84 of 0.2M sodium pentaphosphate (arginine-monophosphate ratio of l :1) (Table f ). This value is quite close to the T1 of 1.06 sec ob- tained for intracellular arginine in N. crassa. TABLE I Comparison of the Sample In trace 11 ul ar arginine in N. crassa 15 N shifts and T1 values of intracellular and extracellular arginine ( 15 Nw,w ,) in N.crassa Solvent & Addends pH T mp 5 ( C) mycelia sus- medium pended in 5.8 minimal medium Viscosity (cp) 615N (ppm) 303.5 10 T 1 (sec) 1.06 (X) u, Arginine ( 12 mM) in minimal medium 5.8 10 Arginine Aqueous solution with 0,2M NayP5016 5,8 10 ( l M) 303.8 2\3 4.25 0\9.6 86 2. 15 N Spectra of Intact N. crassa Mycelia Cultured in 15 NH+- _ _.....___ _ _ _ _ _ _- - ' - - ' - - - ' - - - - ' - ' - - - - - - - - - - ' - - ~ 4 Containing Medium To investigate the types of 15 N-labeled metabolites in !i_. crassa that can be observed in vivo by 15 N NMR, spectra were taken -. -- of intact mycelia grown in 15 NH 4Cl-containing medium. The ure-1 strain of N. crassa was grown in minimal medium containing 0.2% 15 NH 4Cl for 20 hours, and the mycelia suspension for NMR measurements was prepared as described in the experimental section. Figure 6a shows the broadband proton-decoupled 15 N spectrum of the mycelia. A number of well-resolved resonances of 15 N-labeled intermediates and products are observed. The assignments of the 15 N resonances of the intracellular metabolites by comparison with those of the corresponding compounds in aqueous solution are summarized in Table II. The structures of these metabolites and the anabolic pathways by which they are synthesized from 15 NH 4+ and carbon sources in N. crassa are shown in Figure 7. The intense peaks observed in the spectrum in Figure 6a are the side-chain nitrogens of amino acids (both as free amino acids and residues in proteins), viz glutamine NY (262.4 ppm), arginine N0 (291,0 ppm) and Nw,w' (303.5 ppm), and lysine N£ and/or ornithine N0 (342.6 ppm), as well as the a-amino nitrogens of free amino acids such as praline and/or hydroxyproline (321.9 ppm), alanine ('\, 332,3 ppm}, glutamate, glutamine, lysine, and arginine ('\, 334.3 ppm), and valine and/or serine (338,9 ppm) 18 The most shielded peak at 354,5 ppm represents ammoni.um ion. 87 (a) val , cit N N i6 ~~~~~ U ~ ; · NAG \arg) N.s: ,N u w,w , lys, orn ...,__------=~...1..-.1---L-l..__!._..L--L.-.J.__l__i.......J.._,,J._L._jL..i_L._. I r I I I 200 250 Fi 9. 6 ' I 300 ppm \from 1M HN0 ) 3 (Caption on the fol lmtiing pa9e) 350 88 Fig. 6. (a) Broadband proton-decoupled 15 N spectrum of intact mycelia suspension of !i, crassa ijre-1 strain grown in 15 NH 4Cl containing medium for 20 hours. The spectrum represents an accumulation of 5867 transients. The inset shows the amide region of the same sample obtained with gated proton decoupling using 4-sec repetition rate. (b) Broadband protondecoupled 15 N spectrum of the same mycelia after additional 2 hours' growth in 14 NH 4Cl-containing medium, The spectrum represents an accumulation of 21,790 transients. Abbreviations: U:uridine; NAG:N-acetylglucosamine; gln:glutamine; cit:citrulline; arg:arginine; pro:proline; lys:lysine; orn:ornithine; ser:serine; val :valine. 89 Table II 15 Assignments of N-labeled metabolites observed in N. crassa mycelia grown .i.!}_ 15 NH Cl-containing medium4 o15N ( ppm from 1 MHN0 3 ) Metabolites Uridine N3 Guanosine Nl Uridine Nl J t!-acetylglucosamine Peptide amide N Glutamine Ny Cytidine NH 2 Citrull ine N 0 Arginine H0 Adenosine NH 2 Citrull i ne N w Guanosine NH 2 Arginine N , w,w Proline N 4-hydroxy~roline NCL Alanine N crassa -N. -Cell extract after Intact mycelia 15 heat treatment grown with NH 4Cl (Figure 8) (Figure 6a) 216.7a 217.8 228.8 1 251.5a,b 262 .4a,b 287.7 291.0a,b 1 253.2 263.3 281.8 291. 2 297.3 303.1 304.5 24 321 .9 J 332.3a,b CL Lysine Na } Glutamate N Glutamine N: Arginine N CL Valine N } Serine / I CL Lysine N£/Ornithine N 0 Glycine No.of dipeptides Ammonium ion 216 . 5c 229.0h 252.4d 300 . 6 303.5a,b H20 332 .8 334.3a,b 334.9 338.9 337.3 342.6 347.2 354.5 343.4 348.1 354 . 8 263. 1e 282.0h 287. 9e 290.6f 297.9h 300.8e 303.2h 303 .8f 319.8e 322.8 9 332 . 39 334./ 334.3e 334.7f 334. 9f 339 .09 339.0 9 341. 7f 348. 1g 354.5 a. Resonances also observed in the mycelia after isotope dilution by further growth on 14 NH 4Cl (Figure 6b) b. Resonances also observed in mycelia grown on 14 NH 4Cl-containing medium after cycloheximide addition (Figure 9) c. Reference 15 d. Reference 16 e. Measured by the author in mycelia-free, nitrogen-free medium at 25°C f. Measured by the author in aqueous solution at pH 5-7 g. Reference 17 h. Shifts of nucleotides in yeast transfer RNA (Reference 11) 90 a-ketoripote- a - o m i n / \ I ysine '[00- L-c;ilutomote a-ketoQlutorote C=O H It2 - - - - . ~~~R-c-cooI chitin r- cH2 1 coo- uri dine d1phosphoteN-ocetylgl ucoso mine ~H+3 a-Ui09lulorote i L- omino acids 0 coo- II R-c-coo- ~--~ I 9H2 CH2 I + HT-NH3 ATP 4 - cooL-Qlutomote l N-ocetyl-L-Qlutomic).-Hffiioldehyde ; ~ L-glutomote ~ a-ketoglutorote NH+ I 3 l.......----------9H2 .f H2N CH2 'c=o H-r- I CH 2 H-t-NH+ CH2 I 3 coo- I CH2 I CH2 I H-C-NH; I coo- L-ornilhine t I I L-cilru I line ,_,.,.,,,,.=:i I I fumarole H N_. 2 C=NH2 H-N-' I CH2 I CH2 I CH2 I I I I + H-C-NH3 I coo- ________.___ !!!!!. L-arginine Fig. 7, Anabolic pathways of some nitrogenous metabolites in N. crassa. 91 Significantly, one can also detect small 15 N resonances from some intermediate in arginine biosynthesis such as citrulline N0 (287.7 ppm) and Nw (300.6 ppm). Observed resonances from other cellular components include N-acetyl-D-glucosamine (NAG, 251 ,5 ppm) which is a component of fungal cell wall;[. crassa cell wall contains 4.9 mg glucose polymer and 0.5 mg chitin, a homopolymer of N-acetyl-Q_-glucosamine, per 10 mg cell wall 19 . Among the nitrogens in cellular nucleic acids and free nucleotides, only the N3 resonance of uridine (U, 216.7 ppm) is observed; this probably arises from uridine diphosphoacetyl-glucosamine (UDP-GNAc) and uridine diphospho:Jlucose (UDP-glucose} which are intennediates in the biosynthesis of chitin and have been reported to occur in significant amounts (60 µmoles UDP-GNAc and 45 , 5 µmoles UDP-glucose compared to 7.8 µmoles ATP per 100 g wet-weight cell) in N. crassa 20 . The broadband proton-decoupled 15 N spectrum in Figure 6a shows no peaks corresponding to the numerous backbone amide nitrogens of cellular proteins, although they appear as a broad peak around 253 ppm when the spectrum is obtained with gated proton-decoupling as shown for comparison in the inset. The nulling of these resonances on broadband proton decoupling could be the result of the amide nitrogens 0f cellular proteins having on the average an NOE value of approximately zero. A similar effect was observed in the 15 N spectra of bacterial cell walls Sa The relatively sharp resonance at 257.7 ppm can be due to the amide nitrogen of dipeptides such as leucylglycine (257.l ppm), although one would expect a much broader peak from a variety of oligopeptide amide 92 nitrogens present in growing mycelia, on the basis of marked dependence 15 of the amide N shifts on the nature of the adjacent amino acid (_ranging from 242.6 ppm '\, 260,4 ppm for aliphatic dipeptides 21 .) Alternatively, the peak can represent the pyridine ring nitrogen of reduced nicotinamide adenine dinucleotide (256,8 ppm in D 0), 22 2 which is an important cofactor in a number of oxidation-reduction reactions. Definitive assignment of the 257.7 ppm resonance must await further study. The anabolic pathways by which these 15 N~labeled metabolites are synthesized from 15 NH + and carbon sources in N. crassa are 4 summarized in Figure 7. Glutamate and glutamine play crucial roles in assimilating NH 4+ and acting as precursors for many important nitrogenous cellular components, Glutamate, synthesized from NH 4+ and a-ketoglutarate, acts (1) as an a-amino group donor in the synthesis of essentially all the common amino acids, (.2) as a precursor for the ornithine backbone of arginine, and (.3) as the source of nitrogen for the N£H 3+ group of lysine. Glutamine is synthesized from NH 4+ and glutamate. Its amide nitrogen is the source of nitrogen (.1) for one of the guanidino nitrogens (.Nw,w') of arginine (via carbamyl phosphate and citrullinel, (2) for N-acetyl ~_Q_-glucosamine, (3) for imidazole N of histidine and the indole nitrogen of tryptophan, and (4) for 1T N3 and N9 of purines and N3 of pyrimidines. The 15 N spectrum of intact li• crassa mycelia in Figure 6a clearly demonstrates the feasibility of observing 15 N resonances of the important intermediates, glutamine and glutamate, as well as some 93 intermediates in the biosynthesis of arginine and chitin. It is to be noted, however, that in this spectrum, the 15 N resonances of ami.no acid side chains such as glutamine Ny, arginine N0 and Nw~• and lysine NE: represent contributions from both free amino acids and amino acid residues in proteins. An interesting way to check whether free amino acids or protein residues make predominant contribution to the resonances is to test, by the isotope dilution method, how rapidly the components represented in these resonances turn over. For this purpose, the 15 N-labeled mycelia whose spectrum is shown in Figure 6a were collected by filtration, washed and resuspended in 14 NH 4Cl~containing medium (500 ml) for additional 2 hours' growth. During this period, degradation of 15 Nlabeled cellular proteins is expected to be negligible since most proteins turn over slowly with half lives of the order of days 23 , By contrast, free 15 N-labeled amino acids that metabolize rapidly into other cellular components would be expected to show decrease in 15 N peak intensities due to dilution of 15 N by 14 N as the result of metabolic turnover. of those mycelia, Figure 6b shows the proton~decoupled 15 N spectrum The most significant change compared to Figure 6a is the marked loss in the intensities of the glutamine Ny peak (262.4 ppm) and the Na peaks (.332,3, 334,3 ppm) of amino acids 24 compared to other peaks. This result clearly indicates that the glutamine 15 Ny peak observed in Figure 6a before isotope dilution arose predominantly from free intracellular glutamine molecules, and not from glutaminyl residues in proteins, Moreover, the glutamine Ny 94 shows a very rapid turnover rate. It i.s clear, from a comparison of the glutamine Ny peak intensities in Figures 6a and 6b (after correction for the difference in the number of transients accumulatedl that the half-life of this nitrogen of free glutamine in~- crassa growing in NH 4Cl must be less than l hour, although for a more accurate determination, it is necessary to determine the exact, albeit small, contribution of glutamine residues in proteins to the peak intensities. It is also interesting to note that the intermediates in arginine biosynthesis such as citrulline N0 and Nw turnover rapidly as indicated by the disappearance of these 15 N peaks in Figure 6b. The decrease in the intensities of the a.-amino nitrogen resonances of amino acids is no doubt mainly due to their incorporation into proteins, and to transaminations of the glutamate a-amino group to other nitrogenous metabolites. This preliminary result suggests that 15 N NMR can be a valuable technique for studying the~ vivo turnover time of metabolically important functional groups such as the amide nitrogen of glutamine and a-amino groups of amino acids which are not amenable to radioisotope labeling. It is useful to check whether any 15 N-labeled cellular components that were present in sufficient quantities in the intact mycelia could not be detected in the broadband proton-decoupled spectrum of Figure 6 because of unfavorable NOE caused by low mobility and long correlation time (Figure 3a). The mobility of cellular components such as globular proteins and transfer RNA 1 s can be increased by heat denaturation 11 . Thus the mycelia suspension whose spectrum is shown 95 in Figure 6b was subsequently heated at 100°c for one hour, and after removal of the mycelia by filtration, the filtrate containing soluble 15 cell extracts was studied by N NMR. Figure 8 shows the broadband proton ~ decoupled 15 N spectrum obtained at 5o 0 c, The broad peak around 253.2 ppm can be assigned to the amide nitrogens of soluble oligopeptides which have greater mobility, and hence larger negative NOE, than the globular cellular proteins in intact mycelia, In addition, several new peaks are distinctly observable, and can be assigned to Nl of uridine and guanosine (228,8 ppm}, cytidine NH 2 (281,8 ppm), adenosine NH 2 (297.3 ppm), and guanosine NH 2 (.303.l ppm}. which appears as a distinct shoulder on the larger arginine Nw wt peak (~04,5 ppm} 25 (_Table II,)_ ' The uridine N3 resonance {_217.8 ppm) is also substantially increased in intensity. Thus the resonances of all the protonated nitrogens of the four major bases of ribonucleic acids are o5servable in the cell extracts. These probably represent bases in transfer RNA's which, on loss of base-pairing by heat denaturation, assumed a more flexible confonnation in which the individual nitrogen nuclei have shorter correlation times, and hence larger negative NOE 11 . The fact that in intact mycelia resonances from the less mobile, structured cellular components such as tRNA, DNA and the peptide backbone of globular proteins cannot be detected in proton-decoupled spectrum places a restriction on the types of cellular components that can be studied.:!.!!_ vivo by 15 N NMR. Yet, for the study of nonspecifically 15 N~labeled intact microorganisms, this has the advantage of facilitating the observation and assignment of smaller, metabolically 96 C A U,N3 G,U N1 .._ 200 250 .' I 300 I • I 350 15 Fig. 8. Broadband proton-decoup1ed N spectru~ of the ce11 extracts obtained by heat treatment of the mycelia in Fig. 6b .. The spectrum represents an accumulation of 17,607 transients. Abbreviations: A:adenine; C:cytidine; G:guanosine; U:uridine;(the three labeled upfield peaks are the NH resonances of the specified base) . 2 97 interesting molecules such as amino acids, In the intact N, crassa myceli.a, tt is possible to study the assimilation of 15 NH 4+ into free amino acids, while restricting their incorporation into proteins. This can 6e achieved by the addition of cycloheximide to the culture medium before the addition of 15 NH 4Cl to inhibit protein synthesis; under such condttions, the rnycelia can biosynthesize metabolites such as amino acids from 15 NH 4Cl with preexisting enzymes, but the d e ~ synthesis of proteins using 15 Nlabeled amino acids will be inhibited. Thus the biosynthesis of free 15 N-labeled amino acids in vi.vo can be studied by NMR, unobscured by the contributions of 15 N.. labeled amino acid residues in proteins, Possible effects of the inhi.biti.on of protein synthesis on regulati.on of cellular metabolism are as follows: 1, If the biosynthesis of an amino acid is regulated through a feedback inhibition of regulating enzymes at or near the beginning of the pathway by the amino acid al ready formed, the biosynthes is wi.11 cease as the intracellular amino acid concentration approaches the f_i value of the regulatory enzyme, Such feedback inhibition can occur sooner or later in all N. crassa mycelia growing in 15 NH 4Cl, either with or without cycloheximide addition, but with cycloheximide, the inhibition may occur sooner because the synthesized amino acid is not utilized for protein synthesis, 2. If a metabolic pathway is regulated by induction or repres- sion of the enzyme(s) involved, no de novo synthesis of inducible enzymes can occur~ even in the presence of the inducer~ after 98 cycloheximide has been added. However, because most known inducible enzymes are involved in catabolic, not anabolic pathways, the rate of biosynthesis of an amino acid is not likely to be affected by this factor. It is possi'ble, h.owever, that an individual enzyme (as a protein} may have a rapid turnover rate; then the addition of cycloheximide would lead to decrease in the concentration of that enzyme. Repression, which results in the inhibitton of enzyme synthesis, will have no addi.tional effect on mycelia in which protein synthesis is already inhibited by cycloheximide, !!, crassa fil-l strain was grown in minimal medium for 17 hours, then on nitrogen-free medium for 3 hours to deplete the intracellular pool of unlabeled amino acids. 26 After addition of cycloheximide (20 µg/ml), the mycelia were grown in a medium containing 0.4% 15 NH Cl as the sole nitrogen source for 3 hours. Figure 9 shows the 4 proton-decoupled 15 N spectrum of the mycelia . The most conspicuous differences from the spectra of mycelia grown without cycloheximide in Figure 6a are that (l) the glutamine Ny peak (262.4 ppm) is much larger relative to those of arginine N6 (291.0 ppm) and Nw,wt (303.5 ppm) and (2) the peak corresponding to lysine NE and/or ornithine N6 (342.6 ppm) is extremely small, being just distinguishable from the baseline (the right inset). Glutamine can be biosynthesized i_n a large quantity as observed by the intense Ny peak in Pigure 9 because the enzyme catalyzing its synthesis, glutamine synthetase, is not feedback inhibited by glutamine itself 27 . The larger pool of glutamine observed in Figure 9 compared 99 ....... 200 250 • ' l 300 I t I 350 ppm tfrom 1M HN0 ) 3 Fig. 9, Broadband proton~decoupled 15 N spectrum of intact mycelia suspension of Ji. crassa~r~-1 strain grown in 15 NH 4Cl-containing medium after addition of cycloheximide, The spectrum represents an accumulation of 4854 transients. 100 to Figure 6a could be due to decreased outflow of glutamine for the synthesis of cellular components such as chitin when growth is restricted by cycloheximide addition, Arginine, too, is biosynthesized under these condiUons as observed in Figure 9, Significantly, this occurs despite the presence of a feedback inhibition mechanism, In N~ cr~ssa, arginine biosyn thesis is controlled by feedback inhibition of an enzyme acetylglutamate kinase 28 and repression of carbamylphosphate synthetase 29 Both regulatory responses react to the cytosol ic arginine concentNti.on. This concentration is maintained at a low level because most of the arginine pool of cells growing in minimal medium is sequestered in vacuoles 12 . Thus, large pools of arginine must be accumulated before the cytosolic level of the amino acid is high enough to be an effective inhibitor, By contrast, neither ornithine nor lysine occurs i.n significant quantities in those mycel ia as observed by the very small peak at 342.5 ppm, Because ornithine 15 N6 is a precursor for arginine 15 N6 (Figure 7), the presence of arginine 15 N6 peak (291,0 ppm) in Figure 9 indicates that ornithi_ne was synthesized but has been mostly uti 1ized for arginine synthesis, Because, under normal growth condition, ornithine is accumulated in vacuoles, 8 and a small pool of lysine is also accumulated, the above result suggests that cycloheximide may have the effect of preventing the accumulation of these amino acids . 101 These results indicate that cycloheximi.de addi.tion is useful for observing the assimilation of 15 NH 4+ into amino acids such as glutamine, glutamate and arginine which can be biosynthesized in large quantities without inhibiting its own synthesis, For these amino acids, the addition of cycloheximide has the advantage of permitting normal synthesis and yet restricting their outflow into proteins, 3. The time course of assimilation of 15 NH! into glutamine and glutamate in two N. crassa strains. As indicated in the previous section, glutamate and glutamine play central roles in assimilating NH: and acting as precursors for other nitrogenous metabolites (Figure 7). For many years, it was assumed that the primary metabolic pathway for the biosynthesis of glutamate involved the reductive amination of a-ketoglutarate, catalyzed by nicotinamide adenine dinucleotide phosphate (NADP)dependent glutamate dehydrogenase: 31 a-ketoglutarate +NH:+ NADPH + H - - ~ L-glutamate + NADP++ H20 (1) Recently, interest has been shown for an alternate pathway for glutamate biosynthesis because a previously unknown enzyme, glutamate synthase, has been found in procaryotes, 32 yeast, 33 and higher plants. 34 102 This enzyme catalyzes the formation of two molecules of !:_:-glutamate from a-ketoglutarate and !,,_-glutamine (reaction 3). By coupling this reaction with glutamine synthetase (reaction 2), an essentially irreversible pathway for the formation of glutamate from NH +4 and a-ketoglutarate is achieved (reaction 4). k-glutamate +ATP+ NH: ) b~glutamine +ADP+ P a -ketoglutarate + k-glutamine + NAD(P)H ~ 2 k-glutamate + NAD(P) (2) (3) + Sum: a -ketoglutarate + NH 4 +ATP+ NAD(P)H ) L-alutamate +ADP+ P + NAD(P)+ = ~ (4) Because of the irreversibility and the high affinity of bacterial glutamine synthetase for ammonia, this pathway can function in a very efficient manner when the cellular levels of free ammonia are low. By contrast, the reaction catalyzed by glutamate dehydro- genase is freely reversible and proceeds towards biosynthesis of glutamate only in the presence of relatively high concentration of ammonium salts, 35 The cellular regulation of reaction (4) in procaryotes suggests that its primary function in these organisms is for the synthesis of glutamate when growth is limited by the availability of ammonia. 32 In plants, by contrast, reaction (4) 103 is suggested as the primary pathway for ammonia assimilation and glutamate formation; recent studies on the changes in the specific activities .of glutamine synthetase, glutamate synthase and glutamate dehydrogenase in response to the availability of nitrogen source in the medium support this suggestion. 34 d Very recently, glutamate synthase was partially purified from N.crassa and shown to catalyze reaction (3) in vitro. 36 It is therefore of great interest to investigate whether this reaction is coupled with the glutamine synthetase reaction to form an alternate pathway for the assimilation of NH; into glutamate in vivo. This can be studied using the am-1 strain of N.crassa which lacks NADP-dependent glutamate dehydrogenase for catalyzing reaction (1). If this strain is found to assimilate exogenous NH: to form glutamate, this would provide strong support for the occurrence of glutamate biosynthesis via reaction (4). in vivo. 15 N NMR spectroscopy can be a valuable technique for such studies. If N. crassa am-1 strain is first germinated in medium containing unlabeled glutamate, and then is transferred to minimal medium containing 15 NH: as the sole nitrogen source, only glutamate synthesized from exogenous 15 NH:, and not the preexisting unlabeled glutamate, would give rise to a 15 N resonance, Moreover, if the incorporation of 15 N-label into the glutamine amide nitrogen and glutamate Na is 104 studied as a function of time, one might obtain interesting clues as to whether glutamate biosynthesis occurs via glutamine as postulated in reaction (4). N. crassa am-1 strain was cultured in minimal medium containing unlabeled glutamate for 17 hours, transferred to nitrogen-free medium for 3 hours, then, after addition of cycloheximide (20 µg/ml), 15 NH 4Cl was added to the medium to the final concentration of 0.2% at time zero. Figure lOa shows the proton-decoupled 15 N spectra of the mycelial suspension, prepared as described in figure caption, as a function of time starting from the 15 NH 4Cl addition. The course of the assimilation of 15 NH; (resonance at 354.5 ppm) into the amide nitrogen of glutamine (262.4 ppm) and glutamate Na(334.3 ppm) 37 is clearly observed. Clearly the N. crassa am-1 strain, which is incapable of glutamate synthesis via reaction (1), is capable of using an alternate pathway for ass imi l ati ng 15 trn; into glutamate. Because wild-type N. crassa carries all the genes that are functional in a mutant, wild type N. crassa too must be capable of using this pathway. 15 + NH 4 assimilation occurs via glutamine as postulated in reaction (4) in am-1 strain, a comparison 15 N spectrum was obtained for ~ l strain To obtain clues as to whether the alternate pathway of of N. crassa which is capable of reaction (1). The ure-1 strain was grown on minimal medium for 17 hours, transferred to nitrogen-free medium for 3 hours, then after addition of cycloheximide (20 µg/ml), 105 (a) am-1 t = 30 min. t 120 min. 105 min. 240 min • 150 min. . 1 •• r\~ :f'f[~Mi~ 1 360 min. gln Z50 = 40 min. rr --r71rl 285 min• glu ,00 :,50 Z50 :,OD Fig. 10 (Caption on the following page) .550 106 Fig. 10 (a) The proton-decoupled 15 N NMR spectra of N. crassa am-1 strain as a function of time from 15 NHll addition. The myceli-;were suspended in 500 ml of nitrogen-free me.dium in a flask and aerated by bubbling air through them at 25 -30°C. 15 NH 4Cl (lg) was added to the culture medium at time 0. The NMR sample was prepared by collecting the mycelia by filtration and adding to the mycelia pellets, 9 ml of the culture medium to make 18 ml of mycelia suspension in the NMR tube. After each NMR measurement, the mycelia were resuspended in the same culture medium for further assimilation of 15 NH 4Cl . Each spectrum represents 143 transients accumulated in 5 minutes. (b) The proton-decoupled 15 N NMR spectra of N. crassa ure-1 strain at 28°C as a_ function of time from 15 NH 4c1 addition. The cells were suspended in 18 ml of nitrogen-free medium in the NMR tube and were aerated by bubbling in air throughout the NMR measurement. 15 NHll (200 mg) was added directly to the NMR sample at time 0. Each spectrum represents 1004 transients accumulated in 33 minutes. The time given for each spectrum indicates the middle of the accumulation period referred to 15 NH 4Cl addition. Both spectra were obtain~d with 70 µsec pulse width and 2-sec delay. Abbreviations: gln:glutamine; glu:glutamate, 107 15 + NH 4 was added to the mycelia suspension to a final concentration of 1% at time zero. Figure lOb shows the 15 N spectrum of ure-1 strain as a function of time from 15 NH; addition, obtained with the same NMR parameters as in Figure 10a. Comparison of the two spectra (Figures lOa and lOb) clearly indicates that the rate of increase of resonance intensities of glutamine Ny and glutamate Na are different for the two strains; Ny in am-1 strain, the resonance intensity of glutamine increases faster than that of glutamate Na during the first 120 min., whereas in ure-1 strain, the resonance intensities of glutamine Ny and glutamate Na increase at the same rate. This can also be seen in Figure 11~ and~ which plot the resonance intensities of glutamine Ny and glutamate Na as function of time for am-1 and ~re-1 strain respectively. In general, the intensities of resonances in the broadband protondecoupled 15 N spectra cannot be used as a measure of the absolute concentrations of the metabolites because the pulse repetition time may not be long enough to allow full relaxation of all nuclei observed, and the NOE contributions to the intensities may be different for each resonance. However, because the two spectra in Figure 10 were obtained with the same pulse width and repetition rate, the following equation is valid at a given time!= (Sgln 1?glu)am-1 (Sgln/Sglu)ure-1 = ([glnJ/[glu])am-l ([gln]/[glu])ure~l (5) (b) am-1 ure-1 15 _,,..--• (I) ~ ·~10 C: <l) +-' .!I:: ro <l) 0.. -:, b.O ~04 -+-' .....C: ( C) (a) 5 I °/giu ~1/· 1 180 t (min) (I) c:: 52 / (I) b.O 2 ,, , ,I A ./l.-,am-1 ' '' ' .A ,. .... 1 -A-A-A-A ','A ure-1 1Iii 360 180 t (min) 360 180 360 t (min) Fig. 11 (a) Peak intensities S of the observed 15 N resonances in N. crassa am-1 strain as a function of time from 15 NH 4Cl addition, taken from Fig. l0a;O:glutamine amide nitrogen; e: glutamate a-amino nitrogen. Peak intensities are given in arbitrary units. (b) The same for_!!. crassa ure-1 strain; ■: glutamine amide nitrogen and glutamate a-amino nitrogen which have the same intensities as can be seen in Fig. l0b. (c) The peak intensity ratio Sgn ure-1 - strains where 1 /S gu 1 as a function of time in -am-1 and Sgln and Sglu are the peak intensities of glutamine amide nitrogen and glutamate a-amino nitrogen respectively, taken from Fig. lla and b. s n/Sglu is proportional to the con91 centration ratio [gln]/[glu] (see text) __. 0 CX) 109 where Sgln and Sgluare the peak intensities of glutamine NY and glutamate N and [gln] and [glu] are the intracellular concentrations of the a indicated :specfes, and the subscripts ~- l and ure- l refer to the respective strain. From Figure l0b, we can see that (Sgln/Sglu)ure-l = 1 throughout the observed period, indicating that ([gln]/[glu])ure-l must have remained constant during this period. Therefore, from equation 5, (S g1n/S g1u) am- 1 is proportional to ([gln]/[glu])am-l' There- fore, the peak intensity ratio s n;s u can be used as a measure of 91 91 the concentration ratio [gln]/[glu] for each strain. Figure llc compares (Sgln/Sglu)am-l and (Sgln/Sglu)ure-l as function of time. Clearly, in am-1 strain, glutamine concentration increases at a faster rate than glutamate concentration during the first 120 min,. then over the next 240 min., the rate of glutamate synthesis speeds up relative to that of glutamine synthesis. By contrast, in ure-1 strain, the relative rates of glutamine and glutamate synthesis are constant through the entire observed period (285 min.). This result gives support to the hypothesis that in ~~1 strain 15 NH; is assimilated via the glutamine amide nitrogen to form the a-amino group of glutamate. Moreover, our result with ~-1 strain indicates that in vivo the assimilation of 15 NH; into glutamate by the alternate pathway occurs at a relatively high ammonium ion concentration of 35 mM 110 in the medium whereas 'in vitro the partially purifiied glutamate synthase was reported to be inhibited by the presence of 2,5 mM NH + . 36 4 15 Although the intracellular NH: concentration may be different from that in the medium, our result demonstrates that .i!!_ ~ the enzyme is active in N. crassa mycelia growing in excess NH: and suggests that in wild-type N. crassa too, NH: assimilation by the alternative pathway (4) may make greater contribution to glutamate biosynthesis than has been inferred from in vitro studies, 38 The spectrum in Figure lOb shows, for the first time, that it is possible to follow the time course of the assimilation of 15 NH: into nitrogenous metabolites in intact mycelia suspension growing in a "miniature culture" in an NMR tube; the mycelia were found to assimilate 15 NH: at an observable rate at 28°C when they are aerated by bubbling air through the sample and when the volume ratio of mycelia pellet to medium is approximately 1:10. Because of the large sample volume (18 ml) used in our wide-bore Bruker WH180 spectrometer, such mycelia concentration is sufficient to give 15 N resonances of the major metabolites with good signal~to-noise ratios in a reasonable amount of time. The preliminary results described in this paper suggest that NMR can be a promising new technique for studying --.--~ in vivo nitrogen 15 N 111 metabolism and its regulation in intact microorganisms. It is particularly valuable for tracing the metabolism of nitrogenous groups such as the amide group of glutamine, the a-amino group of glutamate and the guanidino group of arginine which play crucial roles in intermediary nitrogen metabolism but are not amenable to the traditional method of radioisotope labeling. This paper demonstrates the feasibility of (l) tracing the time course of the biosynthesis of these metabolites from NH4 in vivo in intact~-~~ mycelia by the noninvasive 15 N NMR technique; (2) estimating their turnover time; and (3) measuring their relaxation times T1 and negative Overhauser enhancements NOE in the intracellular environment. A few of the numerous areas to which this technique can be applied are as fo 11 ows. l. Valuable information on the functioning of metabolic regulation in vivo can be obtained by comparing the patterns of the distribution of NH: into competing anabolic pathways such as is shown in the spectrum in Figure 6a under different nutritional conditions such as the availability of the nitrogen source and the presence of specific repressors or inducers. For major metabolites such as glutamine, glutamate and arginine, their peak intensities in the 15 N spectrum can be used to obtain a quantitative estimate of the relative amounts of 15 NH+ channeled into their biosynthesis under a specific condition, 4 112 if we first measure the intracellular T1 and NOE 6f the 15 N nuclei of these metabolites, then use a pulsing rate that permits full relaxation of all nuclei under consideration and correct for the different contributions of NOE to the peak intensities. 2. 15 N NMR is particularly useful for distinguishing between alternative pathways from a precursor to a product, and for obtaining clues to the existence of a previously unknown or suspected pathway, in cases where the pathways result in different labeling patterns for the product. A specific example for N. cras-sa is the following. N. crassa cultured on minimal medium biosynthesize free glutamine far in excess of their immediate need for this metabolite as a protein component and as a precursor for other nitrogenous cellular components. 30 (Figure 6). Unlike arginine, only a small part of the accumulated glutamine is localized in vacuoles, the bulk of glutamine pool being found in the cytoso1. 30 Does the organism store up glutamine as a nitrogen reservoir to be mobilized during nitrogen deficiency, for the synthesis of glutamate and other amino acids, in addition to other metabolites for which it is a known precursor? Glutamine can be converted to glutamate either by reaction (.3) or, in many organisms, by simple hydrolysis by glutaminase: 113 glutamine + H 0 glutaminase ) glutamate+ NH: (6) 2 Although glutaminase occurs in a wide variety of bacteria, and some species of fungi, 39 glutaminase activity has not been reported in ~- crassa to date. However, it is conceivable that glutaminase is produced in N. crassa but has not been detected because a significant level of production occurs only during certain phases of growth as in E. Coli ~~O or in the presence of an inducer. One known pathway by which the amide nitrogen of glutamine is . ,,~ . crass a ,. s: 41 re l ease d as r1JH+4 ,n glutamine ~ =-::::::;:i > a-ketoglutaramate--➔ 1rn+4 + a-ketoglutarate (7) a-ketoasparagine succinamate Recently, G. Espin and coworkers 42 proposed, on the basis of studies on the changes in the intracellular concentrations of glutamine, glutamate and arginine and release of NH: during nitrogen deficiency, that glutamine amide nitrogen is degraded to NH +4 by reaction (7) during nitrogen starvation and reutilized for glutamate synthesis. However, in the absence of isotope labeling, it is uncertain whether the released NH: derives from glutamine or from other nitrogenous metabolites such as arginine and purine which are known to degrade via urea to • n,·t rogen s t arva t·,on. 26 NH+ dur,ng 4 114 15 N NMR, combined with 13 c NMR, can provide a definitive method for elucidating the pathway(s) of glutamine utilization during nitrogen deficiency .i!l vivo. N. crassa am-1 strain can be grown on glutamine and arginine as nitrogen source. 43 The mycelia can then be transferred to medium containing ( 15 ~Y)glutamine as the sole nitrogen source and the metabolic fate of the 15 N-label can be traced as function of time. Incorporation of the 15 N-label mainly into the a-amino group of glutamate and other amino acids would provide strong evidence that the primary pathway for glutamate synthesis from glutamine is via reaction (3): 0 t- 15 NH 2 cooI I (?H2)~ H-C-NH I coo + (yH2)2 C=O 3 15 ( NY)glutamine I cooa-ketoglutarate NADH "-- NAO+ __./' coo ' I > (CHf)2 + 1-l-t- 5NH+ 3 coo I. cooI (CH 2)~ H-l-NH3 (3 I ) l coo (15 Na)glutamate unlabeled glutamate On the other hand, appearance of the 15 N-label mostly as 15 NH; would imply that the glutamine amide group was hydrolyzed by one of the following processes: 115 ·O 15 ~- NH2 (.CH 2 )~ fOO .. - - - - - = - - - - - - - - ~ (yH ).~ H13{-NH 3 coo- 2 l 5NH+ Hl 3c -NH 3 coo- 4 1 (.15N lg l utami ne glutamate y or a-ketosuccinamate ( 15 Ny )glutamine a,:- keto.gl utaramate a-ketogl utarate Pathways (6 1) and (7 r ) can be distinguished by repeating the experiment with ( 13 c )glutamine and monitoring the appearance of the 13 c label a as either ( 13 c )glutamate or as carbonyl 13 c of a-ketoglutaramate and CX44 15 + 13 a-ketoglutarate. If NH 4 and ( Ca)glutamate are found to be released from glutamine, this would provide strong evidence for the existence of glutaminase activity in N. crassa. Since 15 NH; would be partly re~~ased ~ into the medium, it is important that the entire metabolic process takes place within a miniature culture medium in the NMR tube. The feasibility of such experiment by Bruker WHl 80 spectrometer has been demonstrated. Thus 15 N and 13 c NMR can be used to distinguish 116 between alternative metabolic pathways where they result in different labeling patterns for the products, and possibly provide evidence for the existence of previously unknown pathways. 117 REFERENCES AND NOTES 1. R. B. Moon and J. H. Richards, J. Bio1. Chem. 248, 7276 (1973). 2. R. G. Schulman, T. R. Brown, K. Ugurbi1, S. Ogawa, S. M. Cohen, J. A. den Hollander, Science, 205, 160 (1979) and references cited therein, 3. J. Schaefer, E, O. Stejskal, R. A. McKay, Biochem. Biophys. Res. Commun. 88, 274 4. (1979). M. L1inas, K. Wuthrich, W. Schwotzer and W. von Philipsborn, Nature 257, 817 (1975). 5. a. A. Lapidot and C. S. Irving, Proc. Nat1. Acad. Sci. USA 74, 1988 ( 1977). b. A. Lapidot and C. S. Irving,Biochemistry 1.§_, 704 (1979). c. A. Lapidot and C. S. Irving, Biochemistry 1.§_, 1788 (1979). 6. R. L. Weiss and R. H. Davis, J. Biol. Chem. 248, 5403 (1973). 7. T. J. Fack1am and G. A. Marzluf, Biochem. Genet . .l§_, 343 (1978). 8. R. L. Weiss, J. Bacteriol.126 {3), 1173 (1976). 9. R. H. Davis, M. B. Lawless and L.A. Port, J. Bacterial. 102, 299 (1970). 10. When the myce1ia are extensively washed prior to suspension in arginine-free medium, no leakage of intracellular 15 N-arginine into the medium occurs during the NMR measurement. This was checked by recovering the mycelia by filtration and checking the filtrate .. forte h presence of 15N -arg1n1ne. 118 11. D. Gust, R. B. Moon and J, D. Roberts, Proc. Natl. Acad. Sci. USA 72, 4696 ( 1975). 12. D. Canet, G. C. Levy and I. A. Peat, J. Magn. Reson. ~' 199 ( 1975). 13. R. L. Weiss, J. Biol. Chem., 248, 5409 (.1973), 14. a. C. E. Cramer, L. E. Vaughn and R.H. Davis, (1980) in preparation. b. E. Martinoia, U. Heck, Th. Boller, A. Wiemken and Ph. Matile, Arch. Microbiol. 120, 31 (1979). 15. V. Markowski, G, R, SulHvan and J, D, Roberts, J, Am, Chem, Soc, 99, 715 (1977). 16. R. E. Botto and J. D. Roberts, J. Org, Chem, 42(13), 2247 (1977}. 17. A. Abe, Ph. D, Thesis ,California Institute of-Technology. 18., Although asparagine NS has a very similar shift to that of glutamine NY its concentration in _!i, crassa mycelia has been found to be less than 5% of that of glutamine (see also reference 30}_. Simi- larly, amino acids which have Na shifts in the regions specified but are known to occur in very low amounts relative to the others in _!i. crassa are not included in the assignments, 19. N. de Terra and E. L. Tatum, Sci_ence IB, 1066 (1961). 20. E. J. Smith and R. W. Wheat, Archives Biochem. Biophys . 86, 267 (1960). 21. T, B. Posner, V. Markowski, P. Lotus and J, D. Roberts, J, s. C. Chem. Commun. 768 (1975). 119 22. N. J. Oppenheimer & R. M, Davidson, Org. Mag. Res, ]1_, 14 (1980). 23. E. Martegani and L, Alberghina, J, Biol, Chem, 254, 7047 (1979}. To ascertain that no leakage of 15 N~labeled metabolites into the 24. 14NH Cl~containing medi_um had occurred, the medium used for 4 washing and growing the 15 N-labeled mycelia \'las concentrated and its 15 N spectrum was taken , No 15 N resonance was detected after 3000 transients, 25. The higher chemical shift of 304,5 ppm observed for Nw~w' of arginine in cell extract at 5o 0 c compared to 303.5 ppm in mycelia a.t 10°c is probably due to the hi. gh.e r temperature s i.nce for N , of free arginine in aqueous solution, too, a ~hift of w~w · 304,4 ppm is observed at 55°C (_Table I, in Part III ) compared to 303 ,8 ppm at room temperature. Thus the peak at 303,l ppm can most reasonably- be assigned to guanosine NH 2 . 26, T. L, Legerton and R. L. Weiss, J. Bacteriol. 138, 909 (1979) 27 . J. Hubbard and E. R, Stadtman ~ J. Bacteriol, 93, 1045 (1967). Wolf and R. L. Weiss ( in press 1. 28. E. 29. J. Cybis and R. H, Davis, J. Bacteriol, 123, 196 (.1975}. 30. Y. Mora, G, Espin, K, Wi 11 ns and J, Mora, J, Gen, Microbiol. 104, 241 (1978).. 120 31. Wild-type i!_. crassa synthesizes two glutamate dehydrogenases; one specific for nicotinamide adenine dinucleotide phosphate (NADP-GDH) and the other specific for nicotinamide adenine dinucleotide (NAD-GDH). NADP-GDH is biosynthetic, while NAD-GDH is oxidative. ,,. _., , . 32. D. W. Tempest, J. L. Meers and C. M. Brown in "Enzymes of G!utamine ttetabolism'~ eds.S. Prusiner and E. R, Stadtman, Academic Press, N.Y. 1973. 33. R. J. Roon, H. J. Even and F. Larimore, J. Bacteriol 118, 89 (1974) 34. a. P. J. Lea and B. J. Miflin, Nature 251, 614 (1974) b. M. J. Boland and A. G. Benny, Eur. J. Biochem. ?J.._, 3-55 (1977) c. B, J. Miflin and P. Lea in "Progress in Phytochemistry .1:_, eds. 11 L. Reinhold, J, B. Harborne, and T, Swain, Pergamon Press, Oxford, 1977 d. J. Y. Chiu and P. D. Shargool, Plant Physiol, §i, 409 (1979) 35. P. P. Trotta, K. E. B. Platzer, R, H. Haschemeyer and S. Meister, Proc. Natl. Acad. Sci. USA, Zl, 4607 (1974) 36. G. Hummelt and J. Mora, Biochem. Biophys. Res. Commun.~' 127, (1980) 37. In addition to N of glutamate, N of other amino acids, synthesized a a from glutamate by transamination (Figure 7) may make small contribution to this resonance at later stage in the experiment, Only the N of lysine and glutamine, however, have 15 N shifts within !o.4 ppm a. of that of glutamate (Table II), and concentration of lysine has been shown to be very low in N. crassa growing under these conditions (p.100) 121 38. Little is known about the regulation of formation of glutamate synthase in response to the availability of NH: in other organisms. The enzyme level is higher in limited NH: than in excess NH: in some bacterial strains, whereas the reverse has been reported for other strains (see B. ' Tyler, Ann. Rev. Biochem. 47, 1127 (1978)) 39. A. Imada, S. Igarasi, K. Nakahama and M. Isono, J. Gen. Microbio1. 76, 85 (1973) 40. P. D. Boyer Enzymes 41. C. Monder and A. Meister, Biochem. Biophys. Acta 28, 202 (1958) 42. G. Espin, R. Palacios and J. Mora, J. Gen. Microbiol. 115, 59 (1979) 43. G. Vaca and J. Mora, J. Bacteriol . .!}.l., 719 (1977) Although 13 c -glutamate in reaction 6 1 can be metabolized to 44. 11 11 vol. 4, 1971, p.81, Academic Press, New York a 13 c-a-ketoglutarate by subsequent transamination, such process 1 would occur after release of 15 NH + 4 in reaction 6 whereas in reaction 7' appearance of 13 c-a-ketoglutaramate would precede the appearance of 15 NH+4, 122 PROPOSITION I A 15 N nuclear magnetic resonance study of the ionization states of the coenzyme pyridoxal phosphate and quasi-substrate at the active site of aspartate aminotransferase Pyridoxal phosphate-dependent aminotransferases, which play an essential role in intermediary nitrogen metabolism, catalyze the reversible transfer of thea--amino group of an amino acid to ana--keto acid, with the coenzyme pyridoxal phosphate acting as an intermediary carrier of the amino group. The pyridoxal phosphate-enzyme complex (PLP-E) accepts an amino group from an amino acid, via two intermediate Schiff's bases, to form the pyridoxamine phosphate-enzyme complex (PMP-E) which then donates its amino group to an a-keto acid (Figure l). Of the ami notransferases, aspartate - arrli notransferase, which 1 catalyzes the reaction,1-aspartate + a-ketoglutarate t oxaloacetate + k-glutaimate, has been .studied most extensively. Aspartate amino- transferase from pig-heart cytosol is a dimer (MW"' 90,000) consisting of two identical subunits; each subunit contains one molecule of the coenzyme, pyridoxal phosphate (PLP), bound to the €-amino group of Lys258 in an aldimine linkage (Figure 2a), A dynamic model for the mechanism of catalysis of aspartate aminotransferase · developed by Karpeisky and coworkers 1 with partial revision in the light of more recent evidence 2 , 3 , is shown in Figure 2. A key feature of the model is the assumption that cooperative proton 123 R1 I C=O I coo- f1 ~I _; H-C-N=C-® --> C=N-CH2® I I COCf I coo- H Ketimine Aldimine 0 * II ©-C-H ( PLP-E) Fi:gure l * The aldehyde group of the pyridoxal phosphate forms a Schiff's base with the E-amino group of a lysine residue of the enzyme whenever it is not occupied by an amino group of the substrate. 124 "max 360 nm HZ . z ~430nm ~max 490nm (e) ( f) (d) (c) (b) (a) i z Amax 330nm ( g) Figure 2 ( h) 125 pK 0 = 10. 5 Amax= 367 nm ( b} R I H, ~N 00 HOH2c* pK 0 =8.0 N I CH3 CH3 Amax=378nm (d} Enzyme pK 0 =6.3 Amax =430 nm (e} Figure 3 ~max= 360 nm (f) 126 shifts involved in the transaldimination are linked with reversible reorientation of the coenzyme. The catalytically active form of the pyridoxal phosphate-aspartate aminotfansfetase complex (PLP-E) at neutral pH has an absorption maximum at 360 nm, while the inactive form at acidic pH has Amax ~430 nm; the pKa of this transition is 6.3 4 . By contrast, in free PLP-aldimines, the pKa of this transition (from Amax= 367 nm to 414 nm), attributed to protonation of the phenolate group of pyridoxal, is 10.5 (Fig. 3a,b) 5 . In the proposed mechanism of catalysis, the phenolic group is considered to be ionized in the active form of PLP-E with Amax= 360 nm (Fig. 2a), and the large decrease in pKa, from 10.5 for free PLP-aldimine to the observed 6.3 for the PLP-E, is partially attributed to a postulated hydrogen bonding (or protonation) of the pyridine N atom by a proton-donating group in the enzyme 6 A positive charge 6n this atom, however, lowers the pKa only to ~8.0, as observed in free PLP~aldimine methylated at pyridine N (Figure 3c,d) 7 , and further lowering to 6.3 in PLP-E is attributed to the presence of a cationic group (possibly an imidazole group) close to the phenolic group (Fig. 3e,f) 1 . It is this postulated hydrogen bonding (or protonation)_ of the pyridine nitrogen that can be investigated by 15 N chemical-snfft measurement of 15 N-labeled PLP bound to the enzyme. On addition of the substrate (Figure 2b), the carboxylate group presumably binds to the same cationic site on the enzyme; neutralization of charge on the carboxylate lowers the basicity of the a-amino group, 127 allowing the transfer of a proton from the substrate NH + group to 3 the Schiff-base nitrogen (or phenolate oxygen) of PLP-E (Figure 2c). Now the unionized nucleophilic amino group is ready to attack the highly reactive C=N bond of the protonated Schiff base. However, to form the tetrahedral intermediate (Figure 2d), the pyridine ring of the coenzyme must rotate "40° around an axis passing through its C2 and CS , necessitating the disruption of the hydrogen bond between the pyridine N and proton-donating group on the enzyme. On formation of the substrate aldimine (Figure 2e), its a-proton is abstracted, possibly by Lys-E-NH 2 , resulting in the ketimine formation (Figure 2f) . • I On protonation, the C4.atom acquires a tetrahedral configuration, allowing the coenzyme ring to rotate back to the original plane with concomitant reprotonation of the pyridine N by the proton-donating group HZ (Figure 2f-g). On attack of H2_o (or OH-) at the Ca atom, the a-oxoacid and enzyme-bound pyridoxamine phosphate are formed, completing the first half of the transamination (Figure 2h). The assignment of absorption maxima to reaction intermediates in Figure 2 is based on the following evidence: (a) Study of transient normal reaction intermediates by the temperature-jump method indicated 8 that intermediates with Amax at 360, 430, 490 and 330 nm are formed (b) Amax of 430 nm is assigned to substrate aldimine (Figure 2e) because this absorption maximum is also observed on binding of a quasisubstrate, a-methyl aspartate, which forms a stable aldimine with PLP-E and does not undergo further reaction due to the lack of an a-H atom 9 128 (c) The PMP-E (Figure 2h) has a pH-independent Amax at 330 nm. This model, based on kinetic, spectrophotometric and stereochemical evidence to date, is the most favored model proposed so far. Yet ' • some of the key features of the model remain hypotheses which await experimental verification, viz (1) the hydrogen bonding of pyridine N of the coenzyme with a proton-donating group on the enzyme, which lowers the pKa of the phenolic proton and allows it to abstract a proton from the NH; group of the substrate, and which then acts as a nucleophile (Figure 2b,c); (2) the breaking of this hydrogen bond on binding of the substrate to the enzyme (Figure 2d,e) to allow the formation of the tetrahedral intermediate. In the absorption spectra of the phenolate form of free PL-aldimines, the introduction of a positive charge on pyridine N per~ causes very little change in Amax (367 nm for pyridine (Figure 3b) -vs 378 nm for N-methylpyridinium (Figure 3d) 7. Therefore, the observed absorption spectra of active PLP-E •with Amax at 360 nm provides no evidence for or against protonation or hydrogen bonding at pyridine N. The hydrogen bonding (or protonation) is postulated to explain the large decrease in pKa of the transition from Amax= 430 nm to 360 nm (attributed to phenol ionization) from 10.5 for free PL-aldimine to 6.3 for PLP-E, because the presence of a positive charge at pyridine N was shown to lower the pKa to ~s.o (Figure 3c,d) 7 . The postulated structure of the PLP-aldimine in aspartate aminotransferase in Figure 2a can be thought of as the dipolar tautomer, 129 lb, of PLP-aldimine in Figure 4. For free PLP-aldimine in H o, 2 13 however, recent c NMR studies showed that the neutral tautomer, le, is predominant at neutral pH, with deprotonation occurring predominantly at pyridine N with pKa ~ 6.6 11 , as reflected in large 13 c shift changes at C-6 and C-2 between pH 4 and 7 12 . Thus, for the dipolar tautomer to predominate in PLP-E as postulated in the proposed mechanism, the coenzyme-binding site of the enzyme must be a highly polar environment, with a proton-donating group near the pyridine N and a cationic group near the phenol group of the coenzyme. Although crystallization of pig mitochondrial aspartate aminotransferase for X-ray studies has been reported recently 13 , the X-ray structure of the active site of the enzyme has not yet been elucidated, and the proximity of these groups remain hypothetical. 15 N nuclear magnetic resonance spectroscopy provides a sensitive technique for experimental verification of the postulated hydrogen bonding or protonation of the pyridine nitrogen in PLP-E and its disruption in the PLP-E-substrate complex, because the pyridine nitrogen chemical shift is very sensitive to hydrogen bonding and protonation . J 4a,b R. 0. Duthaler and J. D. Roberts 15 showed that, relative to cyclohexane as solvent, the 15 N shift of pyridine in H2o is shielded by 26 .6 ppm (Table l). Compared to the shift in H2o, a solvent such as trifluoroethanol, capable of stronger hydrogen bonding to the nitrogen unshared pair, causes a further shielding of 11.8 ppm, while protonation at the pyridine N causes a dramatic shielding of 93.9 ppm. 130 ~ H--..... ~N C ®OH2~0- l~J__CH3 I CH3 {Ie) (If) figure 4 131 Table 1 15 N Chemical Shifts of Pyridine o 15 N (ppm from H15 No 3) Compound Solvent Pyridine Cyclohexane 52.9 H20 Tri fl uoroethano 1 79.5 91.3 Pyridine hydrochloride H20 173.4 N-methylpyridinium hydrochloride H20 174.5 II It is proposed that the postulated hydrogen bonding (or protonation) of the pyridine nitrogen of pyridoxal phosphate to aspartate aminotransferase in the active fonn, which is an important feature of the proposed mechanism of transamination, be investigated by 15 N NMR. The pyridine 15 N shift of 15 N-enriched PLP-aspartate transaminase can be measured and compared to that of free PLP-aldimine in H20 at pH 8.5 * to investigate whether any shielding indicative of hydrogen bonding or protonation can be observed in the coenzyme-enzyme complex. For free PLP-aldimine, absorption and 13 c NMR spectra have been obtained in equilibrium mixtures of PLP and amine in H20. At a concentration of 0.5 Min each, the formation of aldimine is virtually * This enzyme shows maximum activity over pH 6. 3 '\, 9. 5. 132 complete over pH 8.3 r..; 12.05, and 15 N spectra of these complexes can be obtained at the natural-abundance level. Because the free PLP- aldimine in H2o at pH 8.5 can exist as a mixture of tautomers, lb and Ic,in Figure 4, it is desirable to obtain 15 N shifts of methylated derivatives, le and If, as models for lb and Ic, respectively. A comparison of the 15 N shift of free PLP-aldimine at pH 8.5 with those of the model compounds, le and If, will then indicate the relative contributions of tautomers lb and le in H2o. To obtain 15 N spectra of the PLP-enzyme complex, it is necessary to synthesize PLP enriched with 15 N at the pyridine nitrogen. Among the various methods for chemical synthesis of pyridoxal phosphate, a procedure in which the 15 N label is introduced at a relatively late stage in the synthesis is shown in Figure 5; reported yield for each 16, 17 step is shown in parentheses To form the 15 N-PLP-E complex, the unlabeled coenzyme in the PLP-E complex(cornmercially available) can be replaced by 15 N-PLP, using the following procedure. PLP-E cysteinesulfinate > PMP PMP-E dia,;;;: against 3> Apo E phosphate pH 9,0 Conversion to PMP ... E facilitates dissociation of the coenzyme to form a stable apoenzyme, and addition of equivalent amounts of PLP (two Ac0~ CH20Ac Ac0H 2CUCH20Ac Ac0H2C)c(CH20Ac 'ti--.( 15NH 20H I I H2 , Roney Ni I ~O~COCH (90%}a > 0 C-CH Ac20 (91%}a> II 3 H 'c~ N-4>-0Et H 'c~ I 3. HCI, ~O (76%} N-<#>-OEt CH20H OH I I lH20 H, 0 CH3 Pyridoxine (65%)° 0 'c~ ~-11-0H2C~OH 0 15NHAc I. Electrolysis in Me0H 2. Na OH, H20 II -o-P-0H 2C'(XOH cyanoethylHOH2COOH H0H2C II phosphate I. Mn 02 O N <N, N dicyclohexyl 15N~ < 2. NH2-4>-0Et CH3 carbodiimide CH3 (70%)b i I 3 15N I OH 0 ~ CH- CH3 ~~I~, N CH 3 Pyridoxol phosphate (Total yield= 28%} Figure 5 a. reference 16 b. reference 17a c. reference 17b __, w w 134 moles of PLP per one mole of dimeric apoenzyme) has been shown to result in reconstitution of catalytically active PLP-E 18 . The pyridine 15 N shift of PLP-E, when compared with that of free aldimine in H2o, should be a very sensitive indicator of the postulated hydrogen bonding to (or protonation by) a donor group in the enzyme. On the basis of 15 N shifts of pyridine in Table 1,relative to free aldimine in H 0, we would expect the pyridine Nin PLP-E 2 to be shielded by more than 10 ppm for hydrogen bonding to a donor group in the enzyme, and by ~go ppm for full protonation. If the 15 N-PLP-E experiment proves to be fruitful, a further interesting experiment that can be attempted is the investigation of the postulated disruption of such hydrogen bonding (if it is occurring in PLP-E) on binding of the substrate. A quasi-substrate, a-methyl-L-aspartate, forms a stable aldimine with the enzyme-bound pyridoxal phosphate and, therefore, provides a stabilized analogue of the substrate-enzyme complex (Fig. 2e). By using 15 N-labeled a-methylaspartate, which can be synthesized by the procedure 19 , 135 we could obtain a 15 N spectrum of the enzyme-quasi-substrate complex 15 N-enriched at both pyridine N of the coenzyme and at the substrate aldimine nitrogen. The pyridine 15 N shift will indicate whether the postulated disruption of hydrogen bonding or deprotonation actually occurs on substrate binding. In addition, the aldimine 15 N shift, which is sensitive to hydrogen bonding and protonation, will indicate whether the protonated aldimine structure for the enzymesubstrate complex (Figure 2e) inferred on the basis of spectrophotometric evidence is borne out by the 15 N chemical-shift data. 136 REFERENCES 1. V. I. Ivanov and M. Ya, Karpeisky, Advan. Enzymol. 32, 21 (1969). 2. A. E. Braunstein in "The Enzymes" ed. P. D. Boyer, vol. IX, p. 379, Academic Press 3. ' New York, 1973 (a) D. E. Metzler, Advan. Enzymol. 50, 1 (1979); (b) J. W. Ledbetter, H. W. Askins and R. S. Hartman, J. Amer. Chem. Soc. 101, 4284 (1979). 4. W. T. Jenkins, D. A. Yphantis and J. W. Sizer, J. Biol. Chem. 234, 51 ( 1958) . 5. D. E. Metzler, J. Amer. Chem. Soc. 79, 485 (1957). 6. I. W. Sizer and W. T. Jenkins in "Chemical & Biological Aspects of Pyri doxa 1 Catalysis 11 , ed. E. E. Sne 11 . et tl, p. 123, Pergamon Press, Oxford 1963. 7. C. C. Johnston, H. G. Brooks, J. D. Albert and D. E. Metzler in "Chemical & Biological Aspects of Pyridoxal Catalysis", ed. E. E. Snell, et tl, p. 69, Pergamon Press, Oxford 1963. 8. P. Fasella and G. G. Hammes, Biochemistry,§_, 1798 (1967). 9. P. Fasella, A. Giartosio and G. G. Hammes, Biochemistry i, 197 (1966). 10. W. T. Jenkins, J. Biol. Chem. 236, 1121 (1961). 11. B. H. Jo, V. Nair and L. Davis, J. Amer. Chem. Soc. 99, 4467 (1977). 12. R. C. Harruff and W. T. Jenkins, Org. Mag. Res.~' 548 (1976). 137 13. G. Eichele, G, C. Ford, J. N. Jansonius, J. Mol. Biol. 135(2), 513 (1979). 14. (a) R. L. Lichter and J. D. Roberts, J. Am. Chem. Soc. 93, 5218 (1971). (b) H. Saito, Y. Tanaka and S. Nagata, J. Am. Chem, Soc. 95, 324 (1973). 15. R. 0. Dutha l er and J. D. Roberts, J. Am. Chem. Soc. l 00, 4969 (1978). 16. N. Elming and N. Clauson-Kaas, Acta Chem. Scan. 1, 23 (1955). 17. (a) V. L. Florentiev, V. J. Ivanov and M. Ya Karpeisky in 11 Methods in Enzymology 11 , ed. D. B. McConnick and L, D. Wright, vol. XVIII, p. 567, Academic Press, London Cl970). (b) M. H. OtLeary and J. R. Payne, J. Biol. Chem. 251, 2248 (1976). 18. M. Arri o-Dupont, Bi ochem. Bi ophys. Res. Commun. 36, 306 C1969) . 19. P. Pfeiffer and E. .Heinrich, J, Prakt. Chem. 146,105 C,936). 138 PROPOSITION II A 15 N nuclear magnetic resonance study of the binding of the inhibitor sulfanilamide to carbonic anhydrase Carbonic anhydrase, a zinc metalloenzyme widespread in nature, is a highJy efficient catalyst of the reversible hydration of co 2 in erythrocytes: Thus, it serves an important physiological role in facilitating the transport of metabolic CO2 between the lung and tissues 1 . The activity of carbonic anhydrase around a neutral pH is governed by the ionization of a group with a pKa near 7. In the generally accepted model of the mechanism of catalysis of this enzyme, this group is thought to be the zinc-bound water molecule which, on ionization to a hydroxide ion, acts as a nucleophile to the substrate CO 2 molecule (Figure la) 2 However, it has also been suggested that the group with pKa 7 may represent the im;-dazole moiety of a histidine residue that is involved in a general-base-assisted attack by ZnOH 2 3 on CO 2 (Figure 1~) , The activity of carbonic anhydrase is strongly inhibited by aromatic sulfonamides such as sulfanilamide (1). For this inhibitor, 139 0~ C ' ~0 ,,,Zn-0 '~ '-H '-. - hQ I ~o' , Zn 11111 OH111 C ~ >, <.. '*~ H..,......o......,_H H ...., possible transition state #0 '-. /C--o- Zn-0 ,j 'H ...., ~Zn-OH2 + HC03 ' \~ H..,......o......,_H (a) 0 /H II ,,jzn -o,H C <.. II 0 '-. /o, H H N"-N-H " ::::::.. '\. /H ,,,Zn-o, <.. \'l'i C=O / oH ::::::.. ' ,,, Zn-OH2 '"7 HCO-3 ' /0 H H'-. +N~N-H (b~ Fi.gure l +~ H-,-H 140 I. the dissociation constant of the enzyme-inhibitor complex is Recent x-ray data s~ow that, in a sulfonamidecarbonic anhydrase complex, the so 2NH 2 group lies within the primary 4a -d coordination sphere of the enzyme ls single Zn atom. The presence of Zn is required for strong binding of sulfonamides; apocarbonic anhydrase shows only very weak binding 2 . The aromatic ring of the sulfonamide also plays an important part in binding, because aliphatic sulfonamides such as methane-sulfonamides show only weak inhibitory effects 5 The x-ray structure of the enzyme-inhibitor complex shows that the aromatic part of the inhibitor is in van der Waals contact with a number of side chains, especially those situated 4d on the hydrophobic wall of the active site, . Ultraviolet 6 fluorescence7 and resonance Raman 8 spectra of the enzyme-inhibitor complexes suggest that the sulfonamide inhibitors are bound as anions 141 . These results suggest the mode of complexation shown in Figure 2. , _,.;Ar ,,, Zn-0 • _ '''l 11111 S111110 ~ N-H ~Ar o=s nmo ''7 'o ' N-H ~,, Zn" H/ I H ( b) Figure 2 Two possible modes of coordination of the deprotonated sulfanilamide to Zn have been suggested: (1) coordination through one of 3 the oxygen atoms (Figure 2a) , and (2) coordination through the nitrogen atom (Figure 2b} 8 . The x-ray structure of a sulfonamide-carbonic anhydrase complex at the resolution obtained to date cannot distinguish 142 4d between the two possible modes of binding and the detailed geometry of coordination remains unknown. !:!_-substituted sulfanilamides such as sulfanilacetamide (NH 2C6H4so 2NHCOCH ) 3 9 have no inhibitory activity . This provides favorable, but by no means definitive, evidence for coordination through the nitrogen, because if coordination takes place through the oxygen, such substitution is also expected to cause steric hindrance between the N-substituted moiety and the aryl group (Figure 2a) 3 . The precise mode of binding of sulfonamide to the active-site of carbonic anhydrase is of great interest for the following reasons: 1) The binding of sulfonamides to carbonic anhydrase and another zinc enzyme, alkaline phosphatase 10 , is highly specific; other wellcharacterized zinc enzymes such as carboxypeptidase and alcohol dehydrogenase do not bind sulfonamides. 2) Sulfonamides are clinically used as an anticonvulsant and cerebrovasodilator drug because they cause an increase in cerebral blood flow by raising CO 2 levels through inhibition of brain carbonic . 11 any . h drase 3) It has been suggested 8 that the sulfonamide in the bound state (Figure 2b) mimics the transition state of the substrate (Figure la), and this may account in part for the extremely high affinity of sulfonamides to carbonic anhydrase. Dissociation constants for sulfonamide-carbonic anhydrase complexes vary between 2xl0- 9 M for ethoxzolamide 3 and 2.6xl0-S Mfor sulfqnilamide compared to 143 10 -3 ~ 10 -2 Mfor the CO 2-carbonic anhydrase complex 2 Stable compounds that are chemically reasonable models for the transition state of the substrate (transition-state analogues) have been shown to be potent inhibitors with affinities three to four orders of magnitude higher than the substrate,for enzymes such as elastase and aldolase 12 . It is proposed that the precise mode of binding of sulfanilamide to the active-site Zn of carbonic anhydrase be investigated by 15 N nuclear magnetic resonance spectroscopy. 15 N NMR spectroscopy can be a sensitive technique for distinguishing between the two possible modes of binding shown in Figure 2, because the 15 N chemical shift of a deprotonated sulfonamide group bound to carbonic anhydrase is expected to be quite different depending on whether it is bound to zinc through the oxygen or the nitrogen atom of the sulfonamide. The 15 N chemical shift of N1 of free sulfanilamide in the neutral form in dimethylsulfoxide is 278 , l ppm upfield of H15 No 3 13 . The 15 N shift of free deprotonated sulfanilamide (pKa1 = 10.4) has not been measured, but is expected to be deshielded by more than 10 ppm relative to that of the neutral form on the basis of shift changes observed in other sulfonamides. On deprotonation of the sulfonamide group, the amide nitrogen shows a downfield shift of 12.8 ppm in £_-toluenesulfonamide (CH 3c6H4so 2NH 2) in dimethylsulfoxide 14 , and of 11.0 ppm in methanesulfonamide (CH 3so 2NH 2) in water 15 . Binding of the deprotonated sulfanilamide through the oxygen atom 144 to Zn of carbonic anhydrase, as schematized in Figure 2a, is expected to cause a substantial deshielding of the sulfonamide nitrogen chemical shift compared to that of the free deprotonated sulfanilamide due to the increased sulfur-nitrogen double-bond character on binding 16 . By contrast, binding of the deprotonated sulfanilamide through the nitrogen atom to the Zn is expected to cause a substantial shielding of the sulfonamide nitrogen shift compared to that of the free deprotonated sulfanilamide on the basis of previous studies on the effect of metal binding on 15 N chemical shifts 17 •18 . Sulfanilamide enriched with 15 N at N1 can be synthesized from£_acetamidobenzene sulfonylchloride and 15 N-enri"ched ammonia by the following procedure: 145 With sulfanilamide 99% enriched i.n sulfonamide nitrogen, it is quite feasible to obtain 15 N spectra of the carbonic anhydraseThe enzyme is sufficiently soluble (>lxl0- 3M in CH3CN:Tris-so4 buffer 5:95 v/v 8 ), and, because of its relatively small size (MW 29,300 for human carbonic anhydrase B), the molecular correlation time would be within a range where the 15 N signal from sulfanilamide complex. the bound sulfanilamide can be obtained without excessive linebroadenOur work on arsanilazotyrosyl 248 carboxypeptidase A (99 % enriched with 15 N at both azo nitrogens, MW 34,700) showed that, for ing. 1.5 mM solution of the enzyme, 15 N resonances with linewidths of 18 ~100 Hz can be obtained after 8-16 hours of accumulation time Thus, 15 N NMR should be a valuable technique for determining the precise mode of binding of the inhibitor, sulfanilamide, to the activesite zinc of carbonic anhydrase. 146 REFERENCES 1. T. H. Maren, Physiol, Rev, 47, 595 (1967} , 2. S. Lindskog, L. E. Henderson, K. K. Kannan, A. Lilyas, P. 0, Nyman, B. Strandberg in "The Enzymes ed. P. D. Boyer, vol . 5, 585 1 \ (1971). 3. Y. Pock.e r and S. Sarkanen in Adv. in Enzymol., ed. A. Meister 47, 149 (1978). 4. (a) K. Fridborg, K. K. Kannan, A. Lilyas, J. Lundin, B. Strandberg, R. Strandberg, B. Ti lander and G. Wir~n, J. Mol. Biol . 25 505 ( 1967). (b) P. C. Beyst~n, I. Vaara, S. Lovgren, A. Liljas, K. K. Kannan and U. Bengtsson in 11 0xygen Affinity of Hemoglobin and Red Cell Acid Base Status, Alfred Benzon Symposium IV, M. 11 RHrth, D. Astrup, eds, Munksgaard, Copenhagen, p. 363 (1972). (c) I. Vaara The Molecular Structure of Human Carbonic An_hydrase 11 Form C and Inhibitor Complexes", Inaugural Dissertation, UUIC-B22-2, Uppsala University (1974). (d) K. K. Kannan, I. Vaara, B. Natstrand, S. Lovgren, A. Borell, K. Fridborg and M. Petef, in 11 Proceedings of the Symposium 11. oo Drug Action at the Molecular Level, G. C. K. Roberts , ed .1 MacMi 11 an, London, p. 73, ( 1977) 147 5. T. H. Maren and C. E. Wi.ley, J, Med, Chern, .l.l, 228 (1968). 6. R. W. King and A. S. V. Burgen, Biochim. Biophys. Acta, 207, 278 ( 1970). 7. R. F. Chen and J. C. Kernohan, J. Biol. Chem. 242, 5813 (1967). 8. K. Kumar, R. W. King and P.R. Carey, Biochemistry .J2_, 2195 (1976). 9. T. Mann and D. Keilin, Nature 146, 164 (1940). 10. G. H. Price, Clin. Chim. Acta 94(3) 211 (1979). ll. I. T. Barnish, P. E. Cross, R. P. Dickinson, B. Gadsby, M. J. Parry, M. J. Randall and I. W. Sinclair, J. Med. Chem. _g]_, 177 ( 1980) . 12. G. E. Lienhard, Science 180, 149 (1973). 13. I. J. Schuster, S. H. Doss and J. D. Roberts, J. Org. Chem. 43, 4693 (1978). 14. C. Casewit, personal communication. 15. H. R. Kricheldorf, _Angew Chem. Int., ed Engl. ll_, 442 (1978). 16. G. C. Levy and R. L. Lichter, 11 Nitrogen-15 Nuclear Magnetic Resonance Spectroscopy 11 , Wiley & Sons, New York, 1979. 17. M. Alei, L. 0. Morgan, W. E. Wageman, Inorg. Chem . ll_, 2289 (1978). 18. The present thesis. 148 PROPOSITION III Chromosomal localization of human embryonicz; globin gene Proteins are chemical fingerprints of evolutionary history, bearing in their amino-acid sequences records of past mutations. Through investigation of the differences in the amino-acid sequences of a protein in a large number of different species, it becomes possible to estimate the rate of occurrence of mutations and, hence, the rate of evolution of the protein. The rates of evolution of cytochrome C, the globins and fibrinopepetides have been interpreted in terms of the biological roles of the proteins 1. For hemoglobins, the ubiquitous oxygen carriers in higher organisms, the amino-acid sequences of the component globin chains have provided some intriguing clues to their evolutionary relationships 2 . In human hemoglobin, six different types of globin chains, a, S, y, o, E and z; , have been identified so far 3. The globin chain composition of HbA, the predominant adult hemoglobin~ is a 2S2 : while that of HbA 2 , a minor adult hemoglobin component of unknown function, is a o . HbF, the major fetal hemoglobin with a higher oxygen 2 2 affinity than HbA, is a 2y 2. Embryonic hemoglobins are considered to be of three types: Hb Portland with the well-established composition z; y , Hb Gower 2(a 2c2) and Hb Gower 1 (z; 2c2 or c4 ) 4 . The types of 22 globin chains synthesized at different stages of human development are summarized in Figure 1 5 149 While adult and fetal hemoglobins have been subjects of extensive study, little is known about the physiological properties of embryonic hemoglobins. In nonnal embryos, the embryonic hemoglobins are produced in nucleated erythrocytes derived from the yolk sac in the first 10 weeks of gestation, prior to the appearance of the a and y chains (Figure l). In a fetus incapable of a-chain synthesis because of genetic defects, Hb Portland (s 2y 2) continues to be produced during fetal life 5 , 7 _ These observations, along with the similarity of the s- and a-chains in amino-acid composition (discussed below), suggest that the s chain functions as an a-type chain in the early stage of normal embryonic development and is capable of continued activity in an a-deficient fetus. The s chain is the only known globin chain that resembles the a-chain in amino acid Composition and function and its evolutionary relationship to the a-chain is therefore of great interest. In their amino acid sequences, the four chains of fetal and adult hemoglobins, a, 8, y and o, show varying degrees of homology: a and S chains have identical amino acids in 42% of their residues, On the basis of these 8 and y chains 71 % and 8 and o chains 96% 8 homologies, it is believed that these globin chains evolved from a common ancestral globin gene by gene duplication followed by independent evolution caused by mutations. The evolutionary relationships of the four chains inferred from their homologies are shown in Figure 2. 2 150 0 o e .!'.' ., . , . • ~-~one •.-.auow • f ;:; Jb::nlX'ylt . •• . , .._:. :/~;~ ~r&i----:::======:t:==========~ ex ex y f3 y 6 12 18 24 30 rost~conccptual c2e (wcel1s) 36 t Birth 6 12 18 24 30 36 42 ~ Post~natal ago lweeks) TJU ffOLUTlON o:r nm GLOBINS Fig. 1 (After W.G. Wood, Br. Med. Bull. E, 2s2 ( 1976)) Fig. 2 (After R.E. Dickerson, 11 The Structure &Action of Proteins" Harper & Row, 1969) For the s chain of embryonic hemoglobin, the complete amino acid sequence has yet to be waked out, but the amino acid sequences of peptide fragments obtained by trypsin and pepsin digestion of the s chain are known 9 When these peptide fragments are aligned so as to maximize homology with the a-chain, the sand a-chains have identical amino acids in 61% of the residues. This homology is greater than that of a-B (42%), but less than that of B-y (71%). This gives rise to a speculation that separation of the s-gene from the a-gene occurred prior to the separation of the Y and B genes (Figure 2). Lending support to this speculation is the finding that, while 151 embryonic hemoglobins are found in a large variety of mammals, including primates, artiodactyls, perissodactyls, carnivores and rodents, fetal hemoglobins are present in primates and artiodactyls only, not in perissodactyls, carnivores and rodents 10 . Within the last few years, molecular hybridization assay using somatic cell hybrids has opened the way to chromosomal localization of human structural genes. The human a~globin gene has been assigned to chromosume 16 and the Sandy globin genes to chromos~me 11 ll,l 2 The 6 globin gene is known to be located next to the S gene on the same chromosome, because globin chains containing segments of both the 6and S chains as a result of unequal crossing~over between the Sand 6 gene loci have been found in hemoglobin variants (Hb Leopor.e and Anti- leopore) 3. If the a and S globin genes originated from a common ancestral gene, their present loci in separate chromosomes suggest that in the course of evolution. they were formed by one of the following possible processes: (1) Tandem duplication of the ancestral gene followed by translocation of one of the genes to another chromosnme. (2) Chromos6me duplication either of the single chromosome carrying the ancestral globin gene (aneuploidy) or of the total number of chromos'Qmes (polyploidy). The first process, while possible, is not very likely, because translocation often results in an inviable offspring. More plausible is the assumption that during evolution, nuclear genetic material has 152 increased by chromo~me duplications. In animals, polyploidy would most likely have taken place early in vertebrate evolution, before the chromosomal sex pattern had become firmly established. Subse- quent to whatever events may have caused the separation of the a and S globin genes, tandem duplication of the S gene to fonn they gene, and, at a later period, another tandem duplication to form the o globin gene must have taken place. Such gene duplication is a continuing process; the present~day y-chain gene is itself duplicated, and the a- gene is duplicated in some populations 3 , 8 . Where, then, is the s globin gene of embryonic hemoglobin located? If this chain separated from the a-chain after the a-S separation, but prior to the S-y separation, as suggested by the relative homologies of the chains, it is possible that the s gene was formed by duplication of the entire chromosome carrying the ancestral a-gene rather than by gene duplication on the same chromosome, as in the formation of the y from the S gene. It would therefore be interesting to attempt a chromoS,Omal assignment of the humans globin gene, to determine whether it is located on chromosome 16, like the a gene, or on a separate chromosome. The fact that individuals totally lacking the a-gene loci can produce Hb Portland (s 2Y2) 13 suggests that the a ands gene loci are either tn separate chromosomes or, at some distance from each other, in the same chromosome. Chromosomal assignment of the humans globin gene can be attempted by the following steps: 153 (1) Isolation of mixed globin mRNA fort, Bandy chains from the erythrocytes of newborns with Bart 1 s hydrops syndrome, which can synthesize only these three chains. (2) Synthesis of single-stranded [ 3HJ-complementary DNA for s, Bandy globin genes by RNA-directed DNA polymerase, using the mixed globin mRNA as templates. Purification of t-cDNA by removing By- cDNA through hybridization with By-mRNA. (3) Assay the hybridization of the t-cDNA with DNA extracted from a series of human-mouse hybrid cell lines, each of which contains only certain ones of human chromosomes. If the t-cDNA hybridizes only with the DNA fragments extracted from the hybrid cell lines that contain ; th human chromos~me, evidence will be provided that the s gene is located on the ; th chromosoo1e. The feasibility of each step is discussed below. l. Isolation of Mixed Globin mRNA for s, Bandy Chains For unambiguous chromosomal assignment of the t gene, it is essential to obtain globin mRNA which contains t-mRNA and is uncontaminated by a-mRNA. In a normal fetus, synthesis of the s globin chain is mostly limited to the first 10 weeks of gestation (Figure 1), with only a trace amount (up to 0.2%) of Hb Portland (t 2y2) found in cord blood at birth 6 . However, newborns with Hb Bart s hydrops 1 syndrome, which is often lethal, totally lack the a-gene loci 7 ; their erythrocytes contain no a-chain, 10 ~ 15% of Hb Portland (t y 2), 80% of Hb Bart ( y4), and a small percentage of HbH ( 84) 2 13 154 Therefore thei.r erythrocytes will yield globin mRNA for the r;, S and y chains only. Globin mRNA can be obtained from reticulocytes by cell lysis, phenol extraction of protein to obtain crude RNA followed, by sucrose gradient centrifu~ation, oligo (dT) cellulose column chromatography and gel electrophoresis 14 , 15 2. Synthesis of [ 3H] cDNA for the r; Globin Gene The mixed globin c;Sy-mRNA can be incubated with [ 3H] deoxyri- bonucleotides and RNA-directed DNA polymerase (reverse transcriptase) to synthesize [ 3H]-complementary DNA for r;, Bandy globin genes 11 ,ls_ To separate the r;-cDNA from By-cDNA, the latter can be annealed to globin aBy-mRNA from normal newborns and the resulting duplex (BycDNA-By-mRNA) separatedfrom single-stranded r;-cDNA by passage over a column of hydroxylapatite 16 . 3. Hybridization of s-cDNA with DNA from Human-mouse Hybrid Cells When human-mouse hybrid cells are formed by fusion of human and mouse somatic cells, the nuclei of the hybrid cells contain the chromosomes of both parent cells. In subsequent cell divisions. there is a progressive and preferential loss of human chromosomes from the hybrid cells; the resulting clones of hybrid cells have only certain chromosomes of the original parent cell 17 . For the hybridization assay, a series of about 20 cell lines can be chosen on the basis of complementarity of the overlapping subsets of human chromosomes that they contain. The human and mouse chromosomes can be distinguished, and the human chromosomes present in each hybrid clone can be identified 155 by a variety of differential staining and banding techniques 11 Single-stranded DNA extracted by standard procedure from the hybrid cell lines of various human chromosomal compositions can be tested for hybridization with the [ 3HJ-z:;-cDNA. The formation of the [ 3HJ-z:;-cDNA-DNA hybrid can be monitored by measuring the radioactivity 3 of the hybrid, after removal of unhybridized [ H]cDNA by digestion with single-stranded specific s1 nuclease, and comparing the radioactivity to that of the total [ 3H]-cDNA initially added 18 . If the [ 3H] -cDNA for z:; gene is found to hybridize only with those DNA fragments extracted from the cell lines that contain ; th human chromos.0'me, it would provide evidence that the z:; gene is located in the ; th chromosome. 156 REFERENCES l. R. E. Dickerson, J, Mal. Evol. l, 26 (1971), 2. R. E. Dickerson, Row , 3. 11 The Structure and Action of Proteins", Harper & New York, 1969 W. P. Winter, S. M. Hanash and D. L, Rucknagel, Adv. in Human Genetics .2_, 229 (1979). 4. E. R. Huehns and A. M. Farooqui, Nature 254, 335 (1975}. 5. \.1. G. Wood, Br. Med. Bull. 32, 282 (1976). 6. F. Hecht, R. T. Jones and R. D. Kaler, Ann, Hum. Genet, (London) I!_, 215 ( 1967). 7. J.M. Taylor, A. Dozy, Y. W. Kan, H. E. Varmus, L. E. Lie .. Inyo, J. Ganesau and D. Todd, Nature 251, 392 (1974), 8. H. Lehmann and R. G. Huntsman, 11 Man 1·s Hemoglobins including the Hemoglobinopathies and their Investigation", 2nd ed.~ J , B. Lippincott, Philadelphia,1974. 9. H. Kamuzora and H. Lehmann, Nature 256, 511 (1975). 10. H. Kitchen and I. Brett, Ann. N. Y, Acad. Science 241, 653 (1974). 11. A. Deisseroth, A. Nienhuis, P. Turner, R. Veles, W, F. Anderson, F. Ruddle, J. Lawrence, R, Greagan, and R, Kucherlapati, Cell _J_g_, 205 (1977) . 12. A. Deisseroth, A. Nienhuis 1 J. Lawrence, R, Giles, P, Turner and F. H. Ruddle, Proc. Natl, Acad. Sci, USA 7.l, 1456 (l978L 13. D. J. Weatherall, J. B. Glegg and W, H, Boon, Br, J, of Haematol, 1§., 357 (1970). 157 14. A. W. Nienhuis, A, K, Falvey and W, F. Anderson in "Methods in Enzymology", ed. S. P. Colowick and N, 0. Kaplan, vol. 30, p. 621 (1974). 15. J. T. Wilson, B. G. Forget, L. B. Wilson and S, M. Weissman, Science 196, 200 (.1977). 16. A. W. Nienhuis, P. Turner and E, J, Benz, Proc, Natl . Acad. Sci. USA 94, 3960 (.1977}. 17. V. A. McKusick and F. H. Ruddle, Science 196, 390 (1977). • 18. E. J. Benz, C. E. Geist, A. W. Steggles, J. E. Barker and A. W. Nienhuis, J. Biol. Chem. 252, 1908 (1977). 158 PROPOSITION IV A study of an alcohol dehydrogenase-substrate complex by resonance Raman spectroscopy Within the last decade, resonance Raman spectroscopy has become a valuable technique for observing selectively the vibrational spectra of a chromophoric group in a biomolecule; it has been used extensively • for the study of hemeproteins l , visual pigments 2 and, more recently, flavoproteins 3 An advantage of Raman over infrared spectroscopy for the study of biomolecules is that water, a poor solvent for the infrared, is an excellent solvent for Raman spectroscopy. For the study of a chromophoric group in a macromolecule, the resonance Raman effect is especially useful. Excitation of the molecule with a laser having a frequency close to the absorption frequencies of the chromophore causes an enormous enhancement of the Raman lines due to the chromophore vibrations which can then be observed at a low concentration, unobscured by vibrational features of the nonchromophoric part of the macromolecule. A new promising area of application for this technique is the study of interaction of a chromophoric natural substrate with an enzyme, because resonance Raman spectra are expected to be sensitive to changes such as bond distortion and electron polarization that are induced in the substrate on binding to the active site, and are postulated to be crucial steps in the mechanism of enzyme catalysis 4 159 Recently, Carey and co..,,.workers 5 studied the resonance Raman spectra of a chromophoric substrate for the enzyme papain, 4..-dimethylamino-3-nitro-a-benzamido-cinnamic acid methyl ester (1), In the process of esterolysis, this substrate forms an acyl-enzyme intermediate, 4-dimethylamino-3-nitro-a ... benzamido..-cinnamoylpapain, in which the cinnamoyl group is covalently bound at the S atom of the cysteine-25 residue of the enzyme. H The acyl-enzyme intermediate, 0 II I C=C-C-OMe I NH I C=O I cp I. 2. which remains stable when maintained at pH3 and 4°C, shows spectral features quite distinct from that of the substrate. In the absorption spectrum, Amax shifts from 350 nm for the substrate to 412 nm for the intermediate, and in the resonance Raman spectrum, the frequency of the _, ethylenic stretching vibration, vC=C' shifts from 1642 cm - for the substrate to an intense peak at 1570 cm-l in the intermediate. The 160 spectral properties of the acyl enzyme are attrtbuted not to the for~ mation of the acyl-cy-steine 25 Hnkage per se (since neither the denatured acyl-enzyme nor thioethyl ester of the substrate show these properties) but to interaction of the acyl moiety with the intact active site. The absorption and resonance Raman properties of the acyl enzyme closely resemble those of 4-dimethylaminocinnambyl imida• zole, which are characterized by a very strong electron-attracting group (imidazole) at the carbonyl end and an electron...donating group (e.g., dimethyl-amino} at the other extremity of the molecule. Thus these residues, actias through chemical bonds, set up a highly polarized TI-electron system, Carey and co-workers proposed that a similar polarization of TI electrons is produced in the acyl group by papain's active site, either by strong hydrogen bonding of a neighboring group to the carbonyl group of the substrate, or by proximity of a cation such as the imidazolium ion of His 159, The implication of .their study is that the electron polarization induced in the cinnamoyl substrate by interaction of its carbonyl group with the active site of the enzyme can be detected by a marked decrease in the frequency of the ethylenic stretching mode in the resonance Raman spectrum. If this proves to be true, resonance Raman spectroscopy can be a sensitive and useful technique for detecting such electron polarization in a suitable substrate or quasi-substrateenzyme complex where such polarization is postulated to be a crucial step in catalysis. 161 An interesting enzyme to which the resonance Raman technique can be applied is horse liver alcohol dehydrogenase (LADH}, a NADHdependent zinc meta 11 oenzyme whi'ch cata 1yzes the transfer of a hydride from the reduced nicotinamide adenine dinuc1eotide (NADH) to the carbonyl carbon of an aldehyde: 0 R-~-H + NADH + H+ ~ RCH 0H +NAO+ 2 In the proposed mechanism of catalysis, it is postulated that the catalytic zinc of the enzyme coordinates to the carbonyl oxygen of the substrate, thereby polarizing the carbonyl carbon and making it more susceptible to attack by the hydride ion, pound, 4-dimethylaminocinnamaldehyde The chromophoric com- (2) (\max 398 nm),* is a natural substrate for alcohol dehydrogenase and has been found to react rapidly with the LADH-NADH complex to form a transient chemical intermediate (Amax 464 nm) which has been detected by a combined stoppedflow temperature-jump rapid kinetic technique 6 In a model study, ** the Amax of this compound (367 nm) shifts to 431 nm on addition of excess Znc1 27 It was proposed that the magnitude of the red shift in the chromophore spectrum, the pH independence of the intermediate formation between pH 6 and 9, and the absence of a deuterium isotope rate effect on substituting NADD for NADH are consistent only with a structure for the intermediate that involves a coordination bond between the catalytic zinc and the carbonyl oxygen 8 In a recent X-ray study of the alcohol dehydra,genase-H 2NADH-dimethylaminocinna- 8 * in water ** in diethyl ether 162 0 maldehyde complex at 3,7 A resoluti.on, the carbonyl oxygen of the substrate appears to be displacing the water molecule coordinated to Zn of the enzyme 9 . Because the chromophoric substrate, 4-dimethylaminocinnamaldehyde, shows an intense resonance Raman band for vC=C at 1591 cm- 1 , 5 , the LADH .. NADH-4-dimethylaminocinnamaldehyde intermediate (which can be stabilized at pH> 9 and 4°c when the enzyme concentration is high) 6 is well suited for study by resonance Raman spectroscopy, It is proposed that the model system, a 4-dimethylaminocinnamaldehydeZnC1 2 mixture, and the enzyme-substrate complex, LADH-NADH-4dimethylaminocinnamaldehyde intermediate, be studied by resonance Raman spectroscopy to investigate whether the proposed coordination of the carbonyl oxygen of the substrate to the zn 2+ in the model system and to the catalytic zinc in the enzyme can be detected by a decrease in the frequency of the ethylenic stretching vibration. Because the sub- strate chromophore in the LADH-NADH-substrate complex has Amax~ 464 nm, laser excitation at, for example, 458 nm, is expected to enhance only the Raman lines that are due to the substrate and not those due to the NADH chromophore which has Amax at 325 nm. If a significant decrease in the frequency of the ethylenic vibration is observed in both the model system and the LADH-NADH-4-dimethylaminocinnamaldehyde intermediate, it will provide strong evidence that TI electrons in the substrate are polarized on binding to the enzyme active site, due to the interaction of the carbonyl oxygen of the substrate with the catalytic zinc of the enzyme. 163 REFERENCES 1. T.. G. Spiro, Biochem, Biophys _, Acta 416, 169 (1975). 2, R. Callender and B. Honig, Ann. Fev. Biophys. Bioen_g, §_, 33 (1977), 3, (a) P. K. Delta, J. R. Nestor and T. G, Spiro, Proc. Natl. Acad . Sci. USA ]_i, 4146 (1977). (b) Y. Nishina, T. Kilagawa, K, Shiga, K. Horiike, Y, Matsumura~ H, Watan and T. Yamano, J. Biochem. 84, 925 (1978). (c) T. Kitagawa, Y, Nishina, K, Shiga, H, Watain, Y, Matsumura and T. Yamam, J. Amer, Chem, Soc. l 01, 3376 {_1979). (d) M, Benechy, T. Y. Li, J. Schmidt, F. Freeman, K, L, Walters Biochemistry ~' 3471 {_1979), 4. P. R, Carey and H, Schneider~ Ace, of Chem .. Res, ll_, 122 {_1978), 5. P. R. Carey, R. G, Caniere, D._ J, Phelps and H. Schneider, Biochemistry .lZ., 1081 (1978), 6. M, F. Dunn and J, S, Hutchison, Bi.ochernistry g, 4882 {_1973). 7. C, T. Angelis, M. F, Dunn, D, C. Muchmore and R. M, Wong, Biochemistry ]_§_, 2922 (1977).. 8, R. G. Morris, G, Saliman and M. F. Dunn, Biochemistry .!2_, 725 ( 1980). 9. C. I, Br~nd~n, H. Eklund, J, P, Samana and L, Wallen, 11th International Congress of Biochemistry, Toronto, Canada, July 8.,,.13, 1979 Abstr, No. 03-4-S95, p, 194, 164 PROPOSITION V Mechanism of Action of an Anticancer Drug The nitrosoureas are an important class of antitumor agents with a broad spectrum of activity in human cancer 1 . However, all the clinically available agents, with the exception of streptozotocin, produce bone marrow depression that limits the clinical usefulness of these agents 2 Streptozotocin, an antibiotic produced by Streptomyces achromogenes, is a 2-deoxy-D-glucose derivative of N-methyl N-nitrosourea (Figure 1) H,OH HN-CO-N-CH 3 I NO Figure 1. Structure of Streptozotocin It is an anticancer agent currently being used clinically, which is particularly effective in the treatment of malignant islet cell cancer of the pancreas, gastrinomas and Hodgkin's disease 3 While the clinical and therapeutic aspects of this drug have been extensively studied, there have been relatively few studies on its mechanism of action. F, Reusser 4 found that streptozotocin induces rapid degradation of deoxyribonucleic acid (DNA) in actively dividing or resting Bacillus subtilis cells. H. S. Rosenkranz and coworkers 5 observed a similar degradation of cellular DNA in E. Coli cells treated 165 with streptozotocin; the drug was also found to inhibit the incorporation 3 of ( H)-labeled precursors into DNA. B. K, Bhuyan 6 studied the action of streptozotocin on mammalian cells and found that streptozotocin inhibits the incorporation of precursors into DNA more than into RNA or protein. It also inhibits the progress of cells into mitosis and cell replication. However, enzymes involved in DNA synthesis, such as DNA polymerase, deoxycytidine kinase, thymidine kinase and thymidylate synthetase, were not significantly inhibited at a concentrati'on of streptozotocin much higher than that needed for a complete inhibition of DNA synthesis. These studies suggest that the inhibition of DNA synthesis, possibly through degradation of template DNA, and the resulting inhibition of cell replication, are the basis for the carcinostatic activity of streptozotocin, However, the precise mechanism by which streptozotocin causes inhibition of DNA synthesis and cell replication is not presently understood, Because !:!_-methyl -!:!-nitrosourea (.MNU) is known to alkyl ate DNA, it has been speculated that streptozotocin, its 2-deoxy-Q-glucose derivative, acts as an alkylating agent. However, no direct evidence of alkylation, such as the isolation of alkylated bases from the DNA of cells treated with streptozotocin, has yet been obtained, Moreover, MNU possesses mutagenic and carcinogenic activities, whereas streptozotocin is primarily carcinostatic, This raises an interesting possibility that the glucose moiety may endow streptozotocin with a mechanism of action which is basically different from that of MNU, Alternatively, strepto- 166 zotocin may alkylate DNA, but th.e effect of DNA alkylation may be potentially mutagenic or potentially template~inactivating~ depending on the site and extent of alkylation, To gain some insight into these problems, it is proposed that possible alkylation of DNA in mammalian cells by streptozotocin be studied by the following steps: (.l} administration of (methyl- 14 C) streptozotocin to mouse ~:eukemia Ll210 cells; (.2) extraction of DNA followed by mi.ld acid hydrolysis, and (3} separation, identification and determination of the relative abundance of (methyh 14 C) bases by radiochromatography, (Methyh 14 C) streptozotocin can be synthesi'zed from tetra.,.Qacetyl glucosamine and (methyl-,. 14 C) methyl isocyanate by the method of Herr and coworkers 7 (Figure 2). Figure 2. Synthesis of Streptozotocin 167 (Methyl~ 14 C}-streptozotocin can be administered to a culture of mouse leukemia Ll210 cells at a concentrati.on high enough to inhibit DNA synthesis. After a suitable incubation period, the cells can be lysed by sonication, and DNA extracted by standard procedures. DNA can be hydrolyzed in mild acid, and, after the addition of authentic methy-lated bases as nonradioactive markers, the bases can be separated, on an ion-exchange column such as Dowex 50. The chromatographic positions of methylated bases can be identified by analysis of the respective UV absorption spectra, By measuring the radioactivity of the chromatographic fractions containing each methylated base, we can determine the relative abundance of methylated bases, if any, produced by treatment with (methyl- 14 C) streptozotocin. If detectable quantities of (methyl- 14 C) bases are recovered from DNA by this method, they would provide direct evidence that streptozotocin alkylates DNA in mammalian cells. In addition, it would be interesting to compare the relative abundance of me_thylated bases produced by streptozotocin with that produced by other alkylating agents such as methylmethane sul- ,,, fonate and MNU; treatment with these drugs yields 7~methylguanine as the major product and 3-methyladenine, 1-methyladenine and 06methylguanine as minor products 8 , Alkylation of DNA can lead to (a) anomalous base pairing in DNA replication, which is potentially mutagenic, or (b} depurination and cleavage of DNA strands at the site of alkylation, which can inactivate the template DNA, Anomalous base pairing involving 06-methylguanine 168 has been observed by L, L, Gerchnrnn i:l,nd coworkers 9 ~ the presence of 6 0 -rnethylgua,nine in the template for polynucleotide synthesis has been shown to cause misincorporation of UMP or AMP tnto the complementary strand. Moreover, the inability of an organ to remove 06 - methylguanine from its DNA has been shown to correlate with th.e organ-specific carcinogenic action of alkylating agents 10 , Such evidence suggests that 06-rnethylguanine plays an important role in the mutagenic and carcinogenic activities of i:llkylating agents. Particularly important is the finding that, in fibroblasts from individuals with a hereditary ski_n disease, xeroderma pigmentosum, which is associated with defects in DNA repair mechanisms, removal of 6 0 -methylguanine from DNA i.s significantly slower than i_n normal individuals 11 , and the incidence of cancer in such individuals is • •f.,can tl y h.19 her 12 , s1gn1 Another possible effect of DNA alkylation is the depurination and breakage of DNA strand(s} at the site of alkylation, Depurination of the methylated base can occur enzymatically or by spontaneous hydrolysis. P, D. Lawley and coworkers 13 showed that alkylation of guanine at the N7 position labilizes the N~glycosidic bond and leads to spontaneous depurination of the base at neutral pH. On the other hand, removal of 3-methyladenine, which is lost more rapidly from DNA than expected from spontaneous hydrolysis, arid of 06-methylguanine, which is chemically stable at neutral pH, is considered to be enzyme-catalyzed. Endonuclease II which recognizes the alkylated site on DNA has been isolated from 169 f. Coli l 4a and DNA Ji ..glycosidase, wh.ich removes 3.,fTlethyladenine from DNA by cleavage of Ji-glycosidic bond, has been purified from~. Luteus 14 b, Depurination can lead to enzymatic cleavage of DNA at apurinic sites; bacterial and mammalian endonucleases which cleave DNA at apurinic sites have also been isolated 15 , These considerations suggest that, if streptozotocin acts as an alkylating agent, depurination and cleavage of template DNA at the site of alkylation is a possible cause of the inhibiti_on of DNA synthesis observed in streptozotocin~treated cells. To test whether streptozotocin induces single .. strand breaks in DNA of mammalian cells, it is proposed that DNA of Ll210 cells prelabeled with (_methyl- 3H) thymidine and exposed to an appropriate dose of streptozotocin be studied by centrifugation in alkaline sucrose density gradients to detect single-strand breaks. When the normal, heavy DNA chains are cleaved by single-strand breaks into lighter chains, a change in the sedimentation pattern, as monitored by radioactivity~ is observed; the radioactive peak shifts from the heavy to the lighter side of the gradients. If streptozotocin is found to induce single-strand breaks in DNA~ it would be interesting to measure the extent of these breaks, and the extent of the inhibition of incorporation of ( 3H)-labeled precursors into DNA at various concentrations of streptozotocin, Finding that incorporation of (methyl-3i:-1) thymidine into DNA is increasingly inhibited with increasing incidence of single-strand breaks in DNA 170 would lend support to the h.ypothesis that streptozotocin inhibits DNA synthesis by inducing single ...stNnd bre~ks in the template DNA. Alkylated DNA can 5e repaired by the excision,,-repair process. Excision of alkylated purine and incision of the phosphodiester chain at the apurinic site by the enzymes described above is followed by removal of the defective sugar phosphate by an exonuclease, formation of the correct nucleoti.de by DNA polymerase and, finally, by joining of the repaired chain segment to the intact chain by DNA ligase 16 , Partial repair of depurinated DNA in vitro by incubation for 1 hour with nuclease and DNA polymerase from a human lymphoblastoid line and DtlA ligase from T-4-infected ~- Coli has been reported 17 , Such excision-repair process, occurring in vivo would remove at least some of the alkylated purines and single-strand breaks in DNA induced by treatment with streptozotocin. It would be interesting, therefore, to vary the incubation period after streptozotocin treatment in the experiments described above to permit varying periods for repair, and to look for variations in the amounts of methylated bases or the single-strand breaks in DNA relevant to the period allowed for~ vivo excision-repair process. If those studies on Ll210 cells prove to be fruitful, they can be extended to in vivo studies in rats. A particularly intriguing study would be a comparison of the site of alkylation and the extent of DNA cleavage induced by equimolar concentration of MNU and streptozotocin in the brain and pancreas of rats. MNU induces malignant tumors in rat brain while streptozotocin is carcinostatic in rat pancreas. 171 Any signi.ficant differences fQund tn th.e si:te of alkylation or in the extent of DNA cleavage. between DNA from the pancrei3S of streptozo.,. tocin..-treated rats and the DNA from the brain of MNU.-treated rats, may provide valuable clues to the different effects, carcinostatic Y2._ carcinogenic, of the two alk.alating agents, 172 REFERENCES ! 35 (1973) . l. S. K. Carter, Cancer Chemotherapy Rep, Part 3, 2. L. C. Panasci, P.A .. Fox , and P. S, Schein, Cancer Res. 37, 3321 (1977). 3. (a) R. W, DuPtiest, M. C, Huntington, W, H. Massey, A. J . Weiss, W. L. Wilson and W. S. Hetcher, Cancer~. 358 (.1975). (.b) P. S. Schein, M, J. O'Connell, J, Blom, S, Hubbard , I.. T. Magrath, P. Bergevin, P. H. Wiernik, J. L . Ziegler and V. T. De Vita, Cancer 34 , 993 (J 974 L (c) P. Lins and E. Suad , Di.abetes 28,190 (1979). (d) J. A, Van Heerden, A. J, Edis and F. J. Service, Ann, Surg, 189, 677 (1979), (e). F. Stadil , G. Stage, J . F, Rehfeld , F, Efsen and K. Fischennan, N. Engl. J . Med . 294, 1440 (1976) 4. F. Reusser, J . Bacteriology 105 , 580 {1971), 5. H. S. Rosenkranz and H, S, Carr, Cancer Res. 30, 112 (1970}. 6. B. K, Bhuyan, Cancer Res. 30, 2017 (1970). 7. (a) R. R. Herr and H. K, Jahnke, J , Amer . Chem , Soc, 89 , 4808 (1967). (.b) P. F. Wiley, R, R, Herr, H, K, Jahnke, C, G, Chidester, $, A. Mizsak, L . B. Spaulding and A. D. Argoudelis, J. Org, Chem, 44, 9 (1979), 8. D, M. Kirtikar and D, A, Goldthwait, Proc .. Natl, Acad . Sci. U. S.A. Zl, 2022 (1974). 173 9. L. L. Gerchman and D, B, Ludlum~ Bi.ochem~ Bi.ophys, Acta~, 310 (1973}. 10. (a) J, W, Nicoll, P, F. Swann and A, E, Pegg, Nature 254., 261 (1975). lb) P. Kleihues and G, P. Margtson, J, Natl, Cancer Inst, ·53, 1839 (197 4 L (c} R. Goth and M, f, Rajewsky, Proc, Natl, Acad, Sci, U.S,A, Zl, 639 (l974L ll. R. Goth-Goldstein, Nature 267, 81 (1977). 12. R. B. Setlow, Nature 271, 713 (1978). 13. P. D. Lawley and P. Brookes, Biochem. J. 89, 127 (1963). 14. (a) D. M. Kirtikar, G. R. Cathcart and D. A. Goldthwait, Proc. Natl. Acad. Sci. U.S.A. 73, 4324 (1976). 15. (b) J. Laval, Nature 269, 829 (1977). (a) S. Ljungquist and T. Lindahl, J. Biol. Chem. 249, 1536 (1974). (b) S. Ljungquist, J. Biol. Chem. 252, 2808 (1977). (c) V. Bibor and W. G. Verley, J. Biol. Chem. 850 (1978). 16. J. Laval, Biochimie 60, 1123 (1978). 17. K. Bose, P. Karran and B. Strauss, Proc. Natl. Acad. Sci. U.S.A. ZE._, 794 (1978).