Investigations of the effect of antigen binding upon tryptophan reactivity in immunoglobulin G Citation Baldwin-Zuschke, Brenda (1986) Investigations of the effect of antigen binding upon tryptophan reactivity in immunoglobulin G. Master's thesis, California Institute of Technology. doi:10.7907/pwp3-7532. https://resolver.caltech.edu/CaltechTHESIS:04052010-153132274 Abstract A ^3H/^(14)C-labelled reagent that reacts specifically with tryptophan residues, was used to monitor the reactivity of tryptophans in the Fc and Fab portions of IgG upon binding DNP-gly, (DNP)_1-cytochrome c, and (DNP)_8-cytochrome c. Binding of monovalent antigen caused a decrease in the reactivity of tryptophan residues in the Fc by 18%; binding of multivalent antigen caused decreases in the reactivity of tryptophan residues in Fc by 21% and Fab by 12%. Experiments to elucidate the interaction between IgG and complement component C4 were attempted. Future studies include further characterization of (DNP)_1-cytochrome c and-the complement fixation capability of IgM-ABPC22 with monovalent antigen. Also planned is an investigation of the interaction of C4 with IgM in soluble antibody-antigen complexes. Item Type: Thesis (Master's thesis) 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): Unknown, Unknown Thesis Committee: Unknown, Unknown Defense Date: June 1985 Record Number: CaltechTHESIS:04052010-153132274 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:04052010-153132274 DOI: 10.7907/pwp3-7532 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 5691 Collection: CaltechTHESIS Deposited By: Tony Diaz Deposited On: 19 Apr 2010 17:20 Last Modified: 16 Apr 2021 22:14 Thesis Files Preview PDF - Final Version See Usage Policy. 2MB Repository Staff Only: item control page

INVESTIGATIONS OF THE EFFECT OF ANTIGEN BINDING UPON TRYPTOPHAN REACTIVITY IN IMMUNOGLOBULIN G Thesis by Brenda Baldwin-Zuschke In Partial Fullfillment of the Requirements for the Degree of Master of Science California Institute of Technology Pasadena, California 1986 (Submitted June 1985) Table of Contents List of Tables and Figures Abbreviations Abstract Introduction Part A: Studies on the Reactivity of Tryptophan Residues in IgG upon Binding Antigen Introduction Experimental Results Discussion Part B: Studies on the Nature of the Interaction between Complement Component C4 and Immmoglobulin G Introduction Experimental Results Discussion Future Research References Tables and Figures Proposition Abstracts Proposition I Proposition II Proposition III ll 13 15 17 19 20 21 24 27 37 38 44 48 List of Tables and Figures Table 1 3H:1"C Ratios obtained from labelling studies Table 2 Classical Pathway of Complement Activation Figure 2 Reaction sequence of the complement system, depicting both the Classic and Alternate Pathways Figure 3 Peptide chain structure of Clq Figure 4 Schematic models of Immmoglobulin G molecule and Immmoglobulin M molecule Figure 5 Domain structure of IgG molecule Figure 6 Diagram of the basic immmoglobulin fold Figure 7 Synthetic Route to dimethyl (2-hydroxy-3-methyl-5- nitrobenzyl)sulfonium bromide Figure 8 Summary of changes in tryptophan reactivities of IgG in the presence of hapten, monovalent antigen, and mitivalent antigen page 27 28 30 31 32 33 33 34 35 Abbreviations G2 DNP (DNP) ,-cytochrome c (DNP) ,-cytochrome c EAC] cells EDTA Fab Fe Fd HMNB-DMSBr IgG IgM SDS tris iv second constant heavy chain domain unit in antibody structure dinitrophenyl group cytochrome c carrier with one dinitrophenyl group covalently attached - cytochrome c carrier with an average of eight dinitro- phenyl groups per cytochrome c molecule antibody coated erythrocytes with complement component C1 bound ethylenediamine tetra-acetic acid product of papain digestion of IgG that possesses the antigen-binding site product of papain digestion of IgG analogous to the crystallizable fragment from rabbit IgG portion of the Fab derived from the heavy chain dimethy] (2-hydroxy-3-methyl-5-nitrobenzyl) sulfonium bromide; modified Koshland's reagent immunoglobulin class G immmoglobulin class M sodium dodecyl sulfate tris (hydroxymethyl) amino methane; tris base Abstract A 3H/}*C-labelled reagent that reacts specifically with tryptophan residues, was used to monitor the reactivity of tryptophans in the Fc and Fab portions of IgG upon binding DNP-gly, (DNP),-cytochrome c, and (DNP),-cytochrome c. Binding of monovalent antigen caused a decrease in the reactivity of tryptophan residues in the Fc by 18%; binding of multivalent antigen caused decreases in the reactivity of tryptophan residues in Fc by 21% and Fab by 12%. Experiments to elucidate the interaction between IgG and complement component C4 were attempted, Future studies include further characterization of (DNP) ,-~cytochrome c and-the complement fixation capability of IgM-ABPC22 with monovalent antigen. Also planned is an investigation of the interaction of C4 with IgM in soluble antibody-antigen complexes. Introduction As early as 1890 the destruction of foreign cells in immmized animals was found to require both antibodies provided through inmumization and some heat labile portion of the serum. (Bordet, 1896, 1898) The unstable serum factor was subsequently called complement. Antibodies may bind to the offending foreign substance, but they have no capacity to degrade it. Instead, the antibodies may effect the destruction of _a foreign cell by initiating an interaction involving complement. Complement is a collective name for a complex set of proteins that, when activated in an ordered sequence, directly damages the cell and causes cytolysis (figure 1). Control proteins are included as well as proteins of the activation sequence. A cascade system with each step able to be inhibited offers the possibility of rapid activation coupled to a series of fail-safe controls. (Porter and Reid, 1978) The classical pathway of complement activation is initiated by the associ- ation of immmoglobulins, IgG or IgM, to soluble antigens or antigens on cell surfaces. One of the proteins of the Cl complex, Cls, is triggered to become an active protease that cleaves and thereby activates C4 and C2. The activated complex of Cab and CZa constitutes C3 convertase, which subsequently cleaves C3. . The resulting complex, C4bZa3b, activates C5 again through cleavage. Self | assembly of C5b67895 on the cell surface then leads to lysis. (Eisen, 1980) Further, an alternate pathway of complement activation, also called the - properdin pathway exists; this pathway does not require antibody-antigen complexes. Instead, certain polysaccharides and other substances effect the cleavage of C3 via the activation of a set of serum proteins, including properdin, factor B, and factor D. Activated C3b thus formed then interacts with C5; the late steps of the pathway (C5-C9) are shared by the classical and alternate pathways (figure 2). (Eisen, 1980) The precise mechanism by which associations of antibody and antigen trigger the initiation of the classical cascade has long been of interest. The Clq subunit of Cl binds via one of its six antibody combining sites (figure 3) to ‘IgG or IgM immunoglobulins of the appropriate subclass. (Reid and Porter, 1976) The interaction between Clq and antibody has been observed to involve the CH, domain of the Fc region of IgG, while binding of antigen takes place at the tips of the Fab portions of the molecule. (Kehoe and Fougereau, 1969, 1974; _ Colomb and Porter, 1975; Yasmeen et al., 1976; Isenman et al., 1977) Since the antigen binding site is spatially removed from the site at which Clq binds, one needs to consider how these two sites on the antibody molecule interact. There are two current models to explain the activation of complement by anti- body-antigen complexes. First, binding of antigen by the Fab portion may cause conformational changes in the Fe region; this model proposes that the Clq binding site in the Fc is exposed or formed by an allosteric change when antibody binds antigen. (Brown and Koshland, 1975, 1977) Circular polarization of fluorescence has detected spectral changes in IgG upon binding either monovalent or multivalent antigen. (Givol et al., 1974; Jaton et al., 1975; Schlessinger et al., 1975) These changes were interpreted as taking place in the Fc region since quite different patterns were observed when Fab units alone were allowed to interact with antigen. Also, the hinge portion of IgG shows a greater constriction _ upon antigen binding. (Colman et al., 1976; Deisenhofer et al., 1976; Huber et al., 1976) This may provide an avenue for transmission of information from Fab to Fc. Second, the associative model proposes that antigen binding mainly causes clustering of the Fc regions of antibody molecules. Optimal complement acti- vation is certainly accomplished by antibody-antigen complexes of more than one antibody molecule. The clustering of Fc regions could facilitate multiple binding by the hexavalent Clq molecule. (Howard et al., 1978; Metzger, 1978) Aggregation is important to complement activation; it is not clear whether it is sufficient alone for activation. Aggregation may be necessary, but not sufficient; for example, aggregation may be required to initiate or enhance conformational changes. Possibly both allosteric changes initiated by antigen binding and aggre- gation of Fe regions are processes that are important in the initiation of the classical complement cascade. (Porter and Reid, 1979) For example, Brown and Koshland in 1975 reacted an anti-lac dye eM with monovalent antigen, and observed complement activation. (Brown and Koshland, 1975) Each molecule of IgM contains five Fc regions; one could perceive IgM as an intramolecular aggregation of Fc regions. Still, without bound antigen, IgM does not activate complement. Hoffman's model for Cl-antibody-antigen interaction proposes that immmo- globulin is an allosteric protein acting on Clq as substrate, and that antigen is the allosteric regulator. (Hoffman, 1976) The antibody is in equilibrium between a tensed state and a relaxed state. The antibody's affinity for Clq is greater in the relaxed state. In the absence of antigen, the tense state predominates, but in the presence of antigen, the proportion of relaxed anti- body to tense antibody is increased. The degree of enhancement of the relaxed state is related to the number of IgG-like units associated. In other words, | the antibody molecules in an aggregate act cooperatively in their transition from the tense state to the relaxed state. Part A Studies on the Reactivity of Tryptophan Residues in IgG upon Binding Antigen Introduction Immunoglobulins are multichained proteins that are described by the formula (HL). where H represents the heavy chain (MW 50K) and L the light chain (MW 25K). For IgG, n=1 and for IgM, n=5 (figure 4). Figure 5 illus- trates this chain structure for IgG and also the "domain" units which make up the chains for IgG. Each domain has roughly the same length, ~ 110 amino acids, and each has a centrally placed disulfide bridge. (Hood et al., 1978) In addition, each domain exhibits the characteristic "immmoglobulin fold" "(figure 6). | The allosteric model of complement activation asserts that a conforma- tional change occurs in the Fe region upon antigen binding at the Fab. Much of the spectroscopic evidence for such a conformational change has been based upon the differences observed between spectra of IgG and spectra of the Fab portion alone. A method that could observe the Fe portion of the molecule directly would be helpful. If a conformational change that initiates the complement cascade does occur in the Fc, one would expect that the binding site for Clq might be a prime site to look for such a change. As mentioned previously, Ciq is believed to interact with IgG via the immmoglobulin's (,2 domain. It has been sug- gested that a tryptophan residue in the Cy2 domain is likely to be involved in Clq binding, though perhaps not directly. Allan and Isliker modified two tryptophans in the Fe with 2-hydroxy-5-nitrobenzyl bromide and observed a dramatic drop in complement-fixing capacity of rabbit IgG aggregates. They suggested that it is the tryptophan at residue 277 which is concerned with the Clq binding site. (Allan and Isliker, 1974) Johnson and Thames have performed binding studies of Clq with synthetic analogues of the C.,2 region around tryptophan 277 and their results suggest that Clq binding site is located in this vicinity. (Johnson and Thames, 1976} The same conclusion was drawn from work done with natural polypeptides that mimic G,2. (Lee and Painter, 1980) Circularly polarized fluorescence has provided some evidence for a conforma- tional change in the Fc region, and this technique actually reflects local environments of tryptophan residues. (Schlessinger, 1975) (In this technique the sample is irradiated at the absorption maxima of tryptophan, and primarily tryptophan's fluorescence is observed.) In these experiments, a tryptophan specific reagent, dimethyl (2-hydroxy- 3-methyl-5-nitrobenzyl) sulfonium bromide, was employed to monitor the reactivity of tryptophans in the immmoglobulin. This reagent, abbreviated HMNB-DMSBr, specifically benzylates tryptophan in solution as a free amino acid or in a protein. It is a modification of a Koshland's reagent, and was developed and characterized by Dan Sand in this laboratory. (Sand, 1982) A double radiolabeling technique was employed to compare tryptophan reactivity in the presence and absence of antigen. Immumoglobulin with antigen bound was labelled with tritiated HMNB-DMSBr. Then this antibody was mixed with an equal amount of antibody labelled with 1“C_HMNB-DMSBr without antigen present. The inmunoglobulin under investigation was a hybridoma IgG, specific’ for dinitrophenyl haptens. Dan Sand's labelling experiments in the presence and absence of hapten and in the presence and absence of monovalent antigen -were repeated, and the results were verified. In addition, a miltivalent antigen was synthesized, and labelling experiments were undertaken to examine its effect on tryptophan reactivity in the Fc portion of IgG. Experimental HMNB-DMSBr The synthesis and characterization of dimethyl (2-hydroxy-3-methyl1-5- nitrobenzyl) sulfonium bromide was developed by Dan Sand in this laboratory and was described in his thesis. (Sand, 1982) The synthetic route can incorporate *H or '*C, via introduction of tritiated NaBH, (New England Nuclear) or ?"C paraformaldehyde (Amersham Searle) at appropriate times in the synthesis (figure 7}. Both tritiated and ‘*C labelled reagent were _ prepared and purified by this method. Isolation of 29Bl IgG5, The mouse hybridoma 29Bl produces an anti-dinitrophenyl IgGy,. (Oi, 1979) Immunoglobulin from this tumor was harvested in SJL x BALBc mice. Ammonium sulfate precipitation was performed on ascites at 4°C. A saturated ammonium sulfate solution (pH 7.4) was added dropwise to the stirring ascites until the final solution was at 45% saturation. After 2 hrs. of stirring, precipitate was collected by centrifugation. The pellet was resuspended in a minimm of 50 mM tris buffer (pH 8.3), and this solution was applied to a Whatman DEAE-Sephacel column which had been equilibrated with the same buffer. All unbound protein was eluted with two bed volumes of buffer. A linear ionic gradient consisting of 0.5 M NaCl in the static chamber was used to elute the immunoglobulin. The first peak with an absorbance of 280 nm was pooled as the 29B] IgG. The peak was concentrated by ultrafiltration at 4°C and dialyzed _ against 0.15 M NaCl, 0.02 M NaH,PO, (pH 7.4) buffer. The protein was then chromatographed on LKB Ultragel AcA 34, and its purity determined by reducing SDS-polyacrylamide gel electrophoresis. If necessary, the immmoglobul.in was rechromatographed on LKB Ultragel AcA 34 to improve purity. Synthesis of DNP-antigens DNP- glycine DNP-glycine was synthesized according to the procedure of Porter. (Porter, 1950) Cytochrome C Horse heart cytochrome c (Sigma Chemical Co.) was first purified via chromatography on Sephadex G-75, eluting with 0.05 M NaH,PO, (pH 6.8) buffer. (DNP) , -cytochrome_c The synthesis of DNP,-cychrome c was described by Dan Sand. (Sand, 1982) A nitro group was introduced into the protein at its only tryptophan residue via Koshland's reagent, The nitro group was reduced to an amine with Na,S,0,. Dinitrofluorobenzene was then added at a pH of 6.3. The resulting modified protein was then purified by affinity chromatography on a column of ABPC-22 IgM coupled to Sepharose-4B. (Porathetal, 1973) This IgM is an anti-DNP protein. (Jarvis, 1981) The purified DNP,-cytochrome C was desalted, lyophilized, and stored at -20°C. (DNP) ,-cytochrome c The synthesis of multivalent antigen, (DNP),-cytochrome c was accomplished by adding a 15-fold excess of 2,4-dinitrofluorobenzene (Aldrich Chemical Co.) to the purified cytochrome c at pH 10.1. A solution of one part dinitrofluorobenzene per 50 parts water and 50 parts dioxane was prepared, and a volume of this solution which represented a 15-fold excess was added to a sample of cytochrome c in 0.5 M NaHCO, (pH 10.1). The reaction mixture was allowed to react for 24 hours with no stirring in the dark. Excess reagent was then removed by exhaustive dialysis against fresh 0.5 M NaHCO, buffer. Labelling of IgG with HMNB-DMSBr Immunoglobulin was labelled in the presence or absence of antigen. Solutions of cytochrome c carrier antigens were prepared by dissolving 1.85 mg DNP-glycine, 41.3 mg (DNP),-cytochrome c, 5.2 mg (DNP),-cytochrome c, 0.5 mg glycine, or 5.2 mg cytochrome c in 4 ml of 0.15 M NaCl, 0.02 M NaH,PO,, pH 6.8 buffer. One milliliter of a 5 mg/ml solution of 29Bl IgG2a was added. A six-fold excess of tritiated HMNB-DMSBr (4 x 10°" M in 1 x 10°3 N HCl) was added dropwise at 0° with stirring. The solution was gently agitated for 24 hours at 4°C, A solution of IgG in the absence of antigen was labelled in | the identical manner with ?*C-HMNB-DMSBr. Five milligrams of this '*C label- led protein were added to each tritiated sample solution. The doubly labelled solution was immediately dialyzed against 0.10 M tris buffer, pH 8.0 to remove excess reagent. Thirty milligrams of completely unlabelled IgG was added to each sample at this point to minimize loss of radioactive label in subsequent steps. The solution was then made 0.03 M in Na,S,0, to remove DNP groups. One hour of reaction under nitrogen was followed by one half hour of oxygen bubbling through to remove excess Na,5,0,. Dialysis against 0.15 M NaCl, 0.02 M NaH,PO0,, pH 7,4 was followed by chromatography on Sephadex G-75. The fractions containing antibody free of hapten and antigen were collected and combined. Papain Digestion The immmoglobulin samples in 0.05 M sodium acetate, pH 5.5 were concen- trated to 5 mg/ml by ultracentrifugation. The samples were made 0.001 M in _ cysteine and 0.002 M in EDTA, and 0.5 mg papain was added to each sample. Incubation for 45 minutes at 37°C was followed by addition of 3 mg p-chloro- mercuribenzoate to stop the reaction. Separation of Fc and Fab Fragments Each sample was dialyzed against 0.15 M NaCl, 0.02 M NaH,PO,, pH 7.4. The 10 Fe fragment was eluted with this same buffer on a DNP-sepharose colum. The buffer was made 0.03 M in DNP-glycine and the Fab fragment was collected. The fractions containing Fc and Fab fragments were concentrated to one milliliter and their purity was checked by SDS-polyacrylamide gel electrophoresis. Liquid Scintillation Counting Each one milliliter sample was added to 10 ml Aquasol 2 scintillation counting fluid. Each sample was counted for fifty minutes on a Beckman LS-250 liquid scintillation counter programmed for a 3% counting error. li Results Characterization of (DNP),-cytochrome c Initially the absorbances at 280 nm and 365 nm were used to calculate the average ratio of DNP groups per protein molecule. This was the method used previously in this laboratory. The extinction coefficient of DNP at 365 nm is 16000, and the extinction coefficient of cytochrome c at 280 nm is 21200. (Mayer et al, 1970) Multiple modifications of cytochrome were performed in order to arrive at a procedure that produced several dinitrophenyl groups on a cytochrome c carrier. Cytochrome c is rich in lysine residues, containing | 19 residues of lysine per molecule. (Margoliash et al, 1962) Thus modification of too many residues resulted in protein polymerization and precipitation. Optimization of experimental procedure was attempted by varying both the excess of dinitrofluorobenzene and the pH at which the reaction was carried out. Under all conditions where a soluble product was obtained, the ratio of DNP groups to cytochrome c molecules did not exceed 1.7:1. This ratio was obtained using the absorbance at 280 nm for the protein and the absorbance at 365 nm for dinitrophenyl groups. Ultraviolet and visible spectra revealed an absorbance of the dinitrophenyl moiety at about 260 mm, which distorted the protein absorbance at 280 nm. In addition, the main absorbance of cytochrome c at 400 nm interfered with the absorbance at 365 nm for DNP. To correct for these difficulties, the visible absorbance of cytochrome _¢ at 400 nm was used to estimate the protein concentration. Given A,,,, the absorbance at 365 nm due to cytochrome c can be calculated and used to correct the observed value A,;,;. Thus the absorbance at 365 nm due solely to dinitro- phenyl groups was obtained. With this procedure involving corrections for quantitating the extent of dinitrophenyl incorporation, the multivalent antigen 12 cytochrome c, with an average of 8 DNP groups per protein molecule, was pre- pared and characterized. In addition, the absorbance of cytochrome c at 528 mm confirmed the protein concentration that we had calculated. Labelling of IgG with HMNB-DMSBr Immunoglobulin 29Bl was labelled in the presence and absence of unmodified cytochrome, DNP-glycine, (DNP),-cytochrome c, and (DNP),-cytochrome c. The labelled antibody was then digested with papain, and the Fc and Fab fragments separated. The results of the double labelling experiments are given in Table 1. Each value given on the table is-the average of duplicate experiments, and half of the difference between the duplicate values is given as the error. The concentration of dinitrophenyl moieties was sufficient to occupy greater than ninety percent of the antigen binding sites. 13 Discussion Figure 8 illustrates the changes in tryptophan reactivity of IgG,, 29B1 in the presence of hapten, monovalent antigen, and multivalent antigen. The presence of the hapten, DNP-glycine, only slightly increases tryptophan reactivity. There is a significant decrease in the reactivity of the Fc region's tryptophan residues upon binding either monovalent or multivalent antigen. These experiments confirm the trends observed for the monovalent antigen system that were observed by Dan Sand in this laboratory. In addition, a multivalent antigen utilizing the same protein carrier was examined for its | effect on tryptophan reactivity. The presence of multivalent (DNP),-cytochrome c decreased tryptophan reactivity in both the Fab and Fc portions of the molecule. The significant reduction in the reactivities of the tryptophan residues in the Fab portion of the molecule is noteworthy. It is possible to offer an explanation for this 12% reduction in reactivity in terms of the movement of the two Fab arms when both antigen-combining sites bind haptens on the same protein carrier. These results strongly suggest that antigen binding causes a change in conformation or motional freedom of regions of the Fc portion such that tryptophan residues are less available to labelled HMNB-DMSBr. The monovalent (DNP) ,-cytochrome c is probably too small to interfere sterically with the availability of the antibody's tryptophan residues; also, aggregation is avoided in the monovalent antigen system. These experiments support a mechanism of complement activation in which a change in the Fe occurs upon binding of antigen, and that change may represent a Significant contribution to the triggering mechanism for the complement cascade, Part B Studies on the Nature of the Interaction between Complement Component C4 and Immmoglobulin G 15 Introduction Complement component C4 is made up of three disulfide-linked peptide chains of molecular weight 93,000(a), 78,000(B), and 33,000(y). (Schreiber and Muller-Eberhard, 1974; Nagasawa and Stroud, 1976) There are also non- covalent interactions between the chains, and 6 M guanidine or 1% sodium dodecyl sulfate is required to separate the chains after reduction. (Gigli et al, 1977) When hydrolysed by CIs, C4 loses C4a, a 8800 molecular weight piece from the amino terminus of its a chain. (Gorski et al, 1981) Active ' CBB remains and combines with activated (a to form C3 convertase. Only a small fraction (<10%) of activated C4b is found bound to antibody aggregates or antibody-coated cells; the rest becomes inactivated in the serum. (Goers and Porter, 1978; Muller-Eberhard et al, 1966) Component C4 labelled with '*5I was reacted with EAC] cells, and only 5-10% of the C4 activated was observed bound to the cells. (Cooper and Muller-Eberhard, 1968) Heat-ageregated IgG binds C4, and this complex participates in the classical complement pathway (Ishizaka et al., 1961), at least through C3. (Sand, 1982) The radioimmunoassay results of Goers and Porter suggested that the C4b molecules that are effective in complement are those bound to the antibody. It was found that rabbit IgG on sheep erythrocytes could participate in a hemoly- tically active IgG/C4/C2 complex that was transferable from cell to cell. (Goers and Porter, 1978) A reactive group, possibly an acyl group, may be generated when C4 is activated by CI and that this reactive group may react with IgG as well as with other adjacent structures. (Campbell, et al., 1980). The binding of Cab to IgG has been demonstrated to be a covalent interaction between the a' 16 chain of C4b and the heavy chain of the Fab portion of the immmoglobulin. (Campbell et al., 1980) Thus the IgG Fab arms provide binding sites for activated C4. The nature of the binding site for activated C4 on antibody-antigen aggregates or antibody-coated red cells is far from clear, The question of whether the activated C4b moiety binds to the Fd at a specific site or whether the binding site is nonspecific, and simply reflects diffusion of the reactive C4b protein is addressed in this section. The early components (Cl and C4) of the classical pathway were combined and allowed to react with aggregates of IgG., 29B1 in order to effect activation. We hoped that binding of C4b to the Fd portion of the immmoglobulin would occur and that subsequent enzyme digests of the antibody molecule with C4b attached would elucidate whether C4b was bound discretely or randomly to the Fab. 17 Experimental 29B1 IgG... The TgG., hybridoma 29B1 was used to produce antibody, and the antibody was isolated and purified as described previously. Complement Component C1 . Component Cl was obtained from fresh human serum. (Gigli, Porter, and Sim, 1976) The euglobin precipitate of serum which contained Cl also contained a high concentration of proteolylytic enzymes. The difficulty of isolating component Cl lay in its easy conversion to activated CI by these proteases. ‘In this particular preparation, repeated additions of the protease inhibitor, phenylmethylsulfonyl chloride, during the isolation steps greatly enhanced the proportion of Cl isolated in its unactivated form... The isolation included addition of the serum to an excess of 5 mM aqueous CaCl,, after which the euglobin precipitate was collected by centrifugation. The precipitate was resuspended in 0.04M sodium acetate buffer and run down a column of Sepharose 6B. In addition human Cl in functionally pure form was obtained from Cordis Laboratories in order to compare its behavior to that of the protein isolated. Complement Component C4 Human C4 was obtained from Cordis Laboratories in functionally pure forn. Preparation of Antibody Aggregates IgG a 29B1 was used to make antibody aggregates, by the method described 2 by Ziccardi. (Ziccardi, 1981) This procedure involved incubation of a 10 mg/ml sample at 63°C for 20 minutes followed by centrifugation. Cyanogen Bromide Cleavage The antibody after reduction with dithiothreitol was dissolved in 98-100% formic acid and then diluted with water. (Steers et al, 1965) Quantitative 18 cleavage of the heavy chain at methionine residues was achieved by adding solid CNBr and allowing the mixture to stand at 5°C for 24 hours. (Givol and Porter, 1965) Chemical Crosslinking of IgG Immunoglobulin G from hybridoma 29Bl was crosslinked by a modification of "Method B” described by Wright. (Wright et al, 1980) The antibody mole- cules were polymerized in the presence of the bifunctional cross-linking reagent dithiobis (succinimidylpropionate) in the absence of antigen. The reagent was added dropwise at 0°C to the IgG in sodium borate buffer (pH 9.0). The mixture was stirred at 4°C for exactly two hours, followed by immediate chromatography on Sephadex G-25. Peak fractions corresponding to IgG dimer were combined and further purified on Ultragel AcA 34 colum., SDS~ PAGE . SDS-polyacrylamide slab gel electrophoresis was used to gauge the purity of proteins and to separate components of mixtures. The gels were stained generally with Coomassie Blue, or with silver when appropriate. (Merril, 1981) Activation of Complement Components The activation of Cl and C4 under physiological conditions has been described. (Ziccardi, 1981) ‘The early classical complement pathway was reconstituted from the purified components at concentrations corresponding to their normal serum concentrations: 135 mg/ml Cl and 409 mg/ml C4. The purified component mixture was in tris-buffered saline (pH 7.5) that was 1.5mM in Ca** and 0.6mM in Mgt+. Aggregated 29Bl or dimers of IgG 29Bl were added to the complement mixture to effect activation. 19 Results Attempts were made to characterize the cyanogen bromide digest of IgG. 29B1 by SDS-PAGE. The difficulty lay in the range of molecular size of the fragments to be separated. SDS/polyacrylamide gels of various acrylamide concentrations (12% to 16%) failed to effect satisfactory separation. The early classical complement pathway was reconstituted with Cl and C4 as described in the experimental section. Addition of IgG., 29B1 aggregates should have caused activation of Cl and C4, and thus resulted in binding of C4 to the aggregates. In addition, the activation of complement components C1 “and C4 by chemically crosslinked dimers of 29Bl was attempted. In all cases, no results have been obtained to date since a satisfactory resolution of the CNBr fragments of IgG has not been achieved. 20 Discussion Complement components Cl and C4 of the classical cascade were reassembled from purified components in physiological solution. Either heat-aggregated IgG or chemically cross-linked dimers of IgG were used to effect activation. Attempts to characterize and separate peptides resulting from the cyan- ogen bromide digest of IgG have been unsuccessful to date. Cyanogen bromide was chosen because it has been demonstrated to produce six fragments from the heavy chain. (Givol and Porter, 1965) If C4 attachment could be demonstrated on one of the six peptides, via comparison .of SDS-PAGE of IgG/CaB digests with ‘SDS-PAGE of IgG digests, further digests of that particular peptide with other enzymes might determine whether C4b binds to a specific site on the IgG molecule. If binding were specific, one would expect to be able to detect C4b only on one specific peptide resulting from any one enzyme digestion of the immunoglobulin. (Incomplete enzyme digestion might distort the results; further digestion could alleviate the problem.) Note that this would not prove spectfic binding, but it was intended as a preliminary experiment. The peptide produced by CNBr cleavage which includes the N-terminus of the heavy chain has been named C-1 (not to be confused with a complement com- ponent), and it includes the Fd portion of the molecule. (Givol and Porter, 1965) Since the site of the C4b attachment to the immmoglobulin has been shown to occur on the Fd portion, one would expect to observe C4b attached to C-1 after cyanogen bromide digestion. Attempts to observe satisfactory separ- ation on SDS-PAGE of the fragments have failed thus far. HPLC may make separ- ation of the fragments achievable. Fragments can be separated on a Sephadex - G-100 colum. However, separation on a column requires a large scaling up of the experiment and requires prohibitive quantities of C4. Isolation of C4 £rom large quantities of outdated human serum may make this possible. 21 Future Research Experiments testing the ability of (DNP),-cytochrome c to activate comple- ment are called for. The combined classical-alternate pathway assay (Sand, 1982) will be used to test for activation of the classical pathway and check for diminution of components. Further characterization of (DNP),-cytochrome is necessary to absolutely confirm that it is indeed monovalent. The modification of this protein involves reaction of its single tryptophan residue with Koshland's reagent, after which the nitro group is reduced to an amine with Na,S,0,. This amine is modified - _ by dinitroflourobenzene at pH 6.3. A control experiment will be carried out under identical conditions, but omitting reduction with Na,S,0, and using tritiated dinitrofluorobenzene. Lack of incorporation of radioactive label into the protein-would indicate that the reaction conditions are such that dinitrofluorobenzene reacts only at the modified tryptophan site. Studies of the interaction between the IgG... 29B1 and (DNP), -cytochrome c have been carried out (Sand, 1982), and confirmation of these results have been described in this report. The interaction of this monovalent antigen with immuno- - globin of class M is also of interest. An IgM antibody secreted by the myeloma ABPC22 is available in this laboratory. This tumor provides a homogeneous population of immmoglobulin M which binds dansyl- and DNP- derivatized proteins. The investigations of Richard Hardy in this laboratory indicate that the mono- valent antigens (DNP),-LBTI or (DNP),-BSA can trigger ABPC22 to activate complement. (Hardy, 1981) This result is in agreement with the experiments -of Brown and Koshland (1975) in which a heterogeneous population of IgM raised in rabbits against lac dye could activate complement upon binding monovalent antigen. The observation that IgM with monovalent antigen can initiate comple- ment activation is consistant with a mechanism of complement activation in which antigen induces a conformational change in the Fc portion of the antibody. The ability of complexes of IgM ABPC22 and monovalent antigen (DNP), - cytochrome c to activate the classical complement cascade will be investigated, 22 Cytochrome c is viewed as a better protein carrier than bovine serum albumin and lima bean trypsin inhibitor for several reasons. (DNP),-cytochrome c has been characterized to be likely monovalent, and future investigation will improve this characterization. The procedure for modifying BSA leaves some doubt as to its monofunctionality, and spectrophotometric analysis assures only that the average ratio of DNP groups per BSA molecule is one. Lima bean trypsin inhibitor was modified at its only tryosine residue, but the monomeric form of this protein undergoes a concentration-dependent dimerization (Krahn and Stevens, 1971) which would result in a divalent antigen. In addition, cyto-" chrome c is a very compact molecule with a low molecular weight of 12384 daltons. Its small size decreases the possibility of blocking sites on the Fe region of the immmoglobulin. The classical complement activation capability of IgM ABPC22 with (DNP),- cytochrome c will be observed employing a standard microcomplement 51Cr release assay (Bengali et al, 1980), and the method of Riches and Stanworth (1980) will be used to gauge any possible capacity of the system to activate the alternate complement pathway. A related facet to the classical complement activation problem involving C4 is particularly intriguing. Much evidence exists that C4b does bind to immmoglobulin G and that this IgG-bound C4b can participate in the complement cascade. (Ishizaka, 1961; Willoughby and Mayer, 1965; Goers and Porter, 1978; Campbell, 1980} Evidence has suggested that hemolytic activity is due primarily to C4b bound to the IgG molecule and not to C4b bound to the cell surface. Recently, Circolo and Borsos have found that cell bound IgM does not bind detectable amounts of Cab. The hemolytically active C4b can be shown to be bound to the cell surface. (Circolo and Borsos, 1982) In order to elucidate the interaction between immunoglobulin and C4, the 23 location of C4b in soluble complexes of IgM ABPC22, C4b and (DNP) ,-cytochrome c will be probed. Previous studies of IgM and C4 have utilized antigens on cell surfaces; the investigation of the interaction between IgM and C4 in soluble complexes may reveal an interaction that is quite different. In this system, in which the antigen (DNP),-cytochrome c is small and monovalent, will C4b be bound to the IgM molecule as indicated by studies of IgG, or will no C4b be bound to the IgM as the results of Circolo and Borsos (1982) suggest for cellular antigens? These experiments will utilize C4 labelled with 3?°I by the method of Reboul (1979) and Campbell (1980). It may be that the roles of the IgG molecule and the IgM molecule in the classical complement cascade are quite different. Much research done on the mechanism of complement activation has assumed IgM to be a functional pentamer of IgG; such may not be the case. Indeed, many functions of IgM and IgG are quite different in the immune system. IgM is produced first in the primary immune response, while IgG is produced later and is the main antibody produced in a secondary immune response. Class M antibody usually has a lower affinity for antigen per site, but an overall high avidity for antigen due to its multiplicity of binding sites. Class G antibody also crosses the placenta to enter the fetal circulation, but IgM does not. It is possible that IgM behaves in a manner fundamentally different than IgG in the activation of the complement cascade. 24 References Allen, R. and Isliker, H. Immmochem. 1974, 11, 175. Allen, R. and Isliker, H. Immmochem. 1974, 11, 243. Bengali, Z.H.; Das, S.R. and Levine, P.H. J. Immmol. Methods 1980, 33, 63. Bordet, J. Ann. Inst. Pasteur Paris 1896, 10, 193. Bordet, J. Ann. Inst. Pasteur Paris 1898, 12, 688. Borsos, T. and Opferkuch, Biologic Activities of Complement 1972, 27 Brown, J.C. and Koshland, M.E. Proc. Nat. Acad. Sci. 1975, 72, 5111. Brown, J.C. and Koshland, M.E. Proc. Nat. Acad. Sci. 1977, 74, 5682. Campbell, R.D.; Dodds, A.W. and Porter, R.R. Biochem. J. 1980, 189, 67. Circolo, A. and Borsos, R. J. Immmol. 1982, 129, 1485. Colman, P.M.; Deisenhofer, J.; Huber, R. and Palm, W. J. Mol. Biol. 1976, 100, 257. Colomb, M.G. and Porter, R.R. Biochem J. 1975, 145, 177. Cooper, N.R. and Muller-Eberhard, H.J. Immumochem. 1968, 5, 155. Deisenhofer, Je; Colman, P.M.; Epp, O. and Huber, R. Z. Physiol. Chem. 1976, | 357, 1421. Eisen, H.M. Immmology, 2nd Edition, Harper and Row, 1980. Gigli, I.; Porter, R.R. and Sim, R.B. Biochem. J. 1976, 157, 541. Gigli, I.; Zabern, I. and Porter, R.R. Biochem. J. 1977, 165, 439. Givol, D; Pecht, I.; Hochman, J.; Schlessinger, J. and Steinberg, I.Z. Progress in Immm. II, Vol. 1, p. 39, Brent and Holborow, J., 1974. Goers, J. and Porter, R.R. Biochem. J. 1978, 175, 675. Gorski, J.P.; Hugli, T.E. and Muller-Eberhard, H.J. J. Biol. Chem. 1981, 256, 2707. 25 Hardy, R. Caltech Ph.D. Thesis, 1981. Hoffman, L.G. Immmochem 1976, 13, 725, 731, 737. Hood, L.E.; Weissman, I.L. and Wood, W.B. Immumology, Benjamin/Cummings Pub- lishing, 1978. Howard, J.C.; Butcher, G.W.; Galfre, G. and Gilstein, C. Curr. Top. Microbiol. Immunol. 1978, 81, 54. Huber, R.; Deisenhofer, J.; Colman, P.M.; Matsushima, M. and Palm, W. Nature +1976, 264, 415. Hyslop, N.E.; Dourmashkin, R.R.; Green, N.M. and Porter, R.R. J. Exp. Med. 1970, 131, 783. Isenman, D.; Painter, R. and Dorrington, K. Proc. Natl. Acad. Sci. USA 1975, 72, 548. Isenman, D.; Ellerson, J.R.; Painter, R.H. and Dorrington, K.J. Biochem. 1977, 16, 233. Ishizaka, T.; Ishizaka, K.; Borsos, T. J. Immmol. 1961, 87, 433. Jarvis, M.R. and Voss, E.W. Mol. Immmol. 1981, 18, 261. Jaton, J.C.; Huster, H.; Braun, D.J.; Givol, D.; Pecht, I. and Schlessinger, J. Biochem. 1975, 14, 5312. Johnson, B.J. andThames, K.E. J. Immunol. 1976, 117, 1491. Kehoe, J.M. and Fougereau, M. Nature 1969, 224, 1212. Kehoe, J.M.; Bourgois, A.; Capra, J.D. and Fougereau, M. Biochem. 1974, 13, 2499. Krahn, J. and Stevens, F.C. FEBS Lett. 1971, 13, 339. Lee, J.P. and Painter, R.H. Mol. Immmol. 1980, 17, 1155. Margoliash, E.; Kimmel, J.R.; Hill, R.L. and Schmidt, W.R. J. Biol. Chem. 1962, 237, 2148. 26 Mayer, M.M. and Miller, J.A. Anal. Biochem. 1970, 36, 91. Merrill, Science 1981, 211, 1437. Metzger, H. Contemp. Top. Mol. Immmol. 1978, 7, 119. Muller-Eberhard, H.J.; Dalmasso, A.P. and Calcott, M.A. J. Exp. Med. 1966, 123, 33. Muller-Eberhard, H.J. and Lepow, I.H. J. Exp. Med. , 121, 819. Nagasawa, S. and Stroud, R.M. Fed. Proc. 1976, 35, 654. Ol, V.T. and Herzenberg, L.A. Mol. Immmol:. 1979, 16, 1005. Porath, J.; Aspberg, K.; Drevin, H. and Axen, R. J. Chromatography 1973, 86, 53. Porter, R.R. Methods in Med. Res. 1950, 3, 256. Porter, R.R. and Reid, K.B.M. Nature 1978, 275, 699. Porter, R.R. and Reid, K.B.M. Adv, Protein Chem. 1979, 33,1. Reboul, A.; Thielens, N.; Villiers, M.B. and Colomb, M. FEBS Lett. 1979, 103, 156. Reid, K.B.M. and Porter, R.R. Biochem. J. 1976, 155, 19. Riches, D.W.H. and Stanworth, D.R. Immumol. Lett. 1980, 1, 363. Sand, I1.D. Caltech Ph.D. Thesis, 1982. Schlessinger, J.; Steinberg, I.Z.; Givol. D.; Hochman, J. and Pecht, I. Proc. Natl. Acad. Sci. USA 1975, 72, 2775. Schreiber, R.D. and Muller-Eberhard, H.J. J. Exp. Med. 1974, 140, 1324. Steers, E.; Craven, G.R.; Anfisen, C.B. and Bethune, J.L. J. Biol. Chem. 1965, 240, 2478. Willoughby, W.F. and Mayer, M.M. Science 1965, 150, 907. Wright, J.K.; Tschopp, J. and Jaton, J.C. Biochem. J. 1980, 187, 767. Yasmeen, D.; Ellerson, J.R.; Dovington, K.J. and Painter, R.H. J. Immmol. 1976, 116, 518. Ziccardi, R.J. J. Immmol, 1981, 126, 1769. 27 Table 1: 3H:/"C Ratios obtained from labelling studies Controls: 29B1 + unmodified cytc 29B1 + glycine Antibody with hapten: 29B1 + DNP-glycine Antibody with monovalent antigen: 29B1 + (DNP),-cytc Antibody with mltivalent antigen: 29B1 + (DNP),-cytc 0.99 0.97 0.99 0.81 0.78 Fe I+ t+ + 0.02 0.09 0.07 0.03 0.07 1.00 0.98 1.02 0,95 0.88 Fab I+ I+ 0.07 0.05 0.07 0.05 0.06 28 CELL MEMBRANE rae BINDING 7 SITE Figure 1. Classical Pathway of Complement Activation a. Complement factor Cl, consisting of subunits Clq, Cls and Clr, binds to the antibody. b.c. The Cl complex becomes enzymatically active and cleaves C4. d. C4b binding to the cell surface is illustrated, although it is also found covalently bound to the Fd portion of IgG. e.f. C2 is cleaved by the CIs subunit and combines with C4b to form C3 29 & Figure 1 (cont.) _ g. C3 convertase splits C3 into C3a and C3b h. The complex C4b-C2a-C3b cleaves C5 in the last enzymatic step of the cascade. i.j.k.1. The remaining components assemble themselves in such a way that a small hole is formed in the membrane. From M. Mayer, ''The Complement System,'' Sci. Am. 229, 54 (1973). 30 454 IMMUNOLOGY Ciq-Cir-Cis Classic Ab: Ag pathway complexes Cla: Cif Cig c4 c2 . 3a Anaphylatoxin C4atC4b. C2a+C2b C3a py C3b Immune adherence reactions: C4b 2a C3 phagocytosis, secretion amplification PC Sb-Bb | C4b-da-3b cs Ch _ Cab 2a°3b H—> fecss emotaxis, P-C3b Bb Cab anaphylatoxin Surface | C5b P| polysac- cé charide C5b,6 C3b Bb+Ba c7 D C3b-B C5b,6, 7 Reactive lysis Alternate cs C5b,.6.7 Chemotaxis C3b+B pathway Figure 2: C5b-6 Partial membrane co lesion C5b-9 Complete membrane tesion Reaction sequence of the complement system, depicting both the Classic and Alternate Pathways (P=properdin, B=factor B, D=factor D) (Eisen, 1980) 31 Equimolar amounts of A, B and C chains (Each ~ 23,000 mol. wt.) A NYY YY N00 S-S B i VAC AU We aN COOH Cc 8 YY 008 4 — Ss C xle SS SS COOH B NYY 0008 A YY 100 H _St T t ' Collagen-like i Non-Collagen-like 88 Residues 1403 to 108 Residues 1 / — Globular C-Terminal Regions Peptide chain structure of Clq Dimensions are from electron microscope studies. (Porter, R.R. and Reid, K.B.B. Nature 1978, 275, 699) Figure 3: 32 -~SS— Disulfide bonds [——— Variabie region GEMM Constant region Figure 4: Schematic models of immmoglobulin G molecule and immunoglobulin M molecule (Eisen, 1980) 33 Figure 5: Domain structure of IgG molecule (Poljak, R. Nature 1972, 235, 137) COO- oz marker Figure 6: Diagram of the basic immunoglobulin fold. The domain shown here is Cy which starts at residue 110 and extends to 214, the carboxyl terminus, The Cyl domain has a very similar structure. The Vz and Vy domains contain an additional loop of poly- peptide (dasked lines). (Poljak, R. Proc. Natl. Acad. Sci. 1973, 70, 3306) 34 (zg6T*pues) *‘(be)aqH pue apAyapTewaojsered YyAFA -[AyQou-z Jo vofFIeTAYyJowouorq aya UF partanoo Ont jo uofqei1od10ouT 4 aul - ugeN YEA VOFJONp|r TEFIFUF 9yQ UP poZerOdTOSUF Sem wNnFEITAL apfworig wnzuoz[Ns (TAzuaqoijyu ~c—[Ayaul—¢-Axorapéy~z) TAYWEP OF ephyopTeTAoy;Tes worz 9 ynor oFqeyqudg *yf aansty qTouaydo1yfu-y £H9 | 2 2 2 2 of es —*H9 NO ares ne N a "yo - fip-s-fip wee *ont HO HO (0%) HO Fu EHD fu9 NO Post \OW-IN \ YgeN Zz HOBN HO ONPN Lg * Ho HO fy9 fn Ho?H9 oo 35 Figure 8. Summary of changes in tryptophan reactivities of IgG in the presence of hapten, monovalent atigen, and multivalent antigen The values expressed are percentage change of incor- poration of label in the presence of ligand compared to incorporation of label in absence of ligand. Fe fragments: 2981 + DNP-gly +2.1% 2981 + (DNP), cyt c -18.2% 2981 + (DNP) geyt c 21.2% - + ‘Fab fragments: 2981 + DNP-gly +4.1% 2981 + (DNP), cyt ec -5.0% 29B1 + (DXP) gcyt c ~-12.0% - + 36 PROPOSITIONS 37 Proposition Abstracts Proposition I Cooperativity among metal atoms in a cluster has been used to explain the rate enhancement of carbonyl dissociation with progressive phosphine substitution, Ir, (00), Ir, (00), , (PPh,) 23225 rr, (co), ,(PPh,), HE" y Ir, (co) PPh), Another explanation for this enhancement is the great steric bulk of the triphenyl phosphine ligand, The investigation of CO exchange processes in complexes of the type Ir,(CO},,PX, is proposed in order to assess the relative contributions of electronic cooperativity and spatial consideration. Phosphines could be employed that have roughly equivalent size, but very different m-accepting characteristics. Proposition II Upon activation, C4b exhibits a metastable binding site for the Fd portion of IgG in immume aggregates. Evidence suggests that C4 contains an internal thioester, which upon cleavage of C4 by Cls is exposed and activated. The activate acyl of the thioester may react with nucleophiles on the Fd. Experi- ments are proposed to investigate the effect of antigen-binding and Cl-binding on the availability/reactivity of nucleophilic groups in the Fd portion of the immmoglobulin. Double labelling experiments with a small, activated ester are suggested. Proposition III The study of the globin gene family of vertebrates has contributed much to knowledge of evolution, development, and gene clusters. The isolation and sequencing of a hemoglobin gene from the annelid worm, Glycera dibranchiata is proposed because comparison to vertebrate globin, vertebrate nyoglobin, and leghemoglobin may help explain the evolution of iron heme-bearing pigments that function in oxygen storage. 38 Proposition I Transition-metal cluster compounds are of interest to chemists largely because of their potential applications in catalysis. The special properties that adjacent metals give the molecule or the catalytically active fragments that the molecule releases may explain the reactivity of clusters as homo- geneous catalysts. (Muetterties, 1976; Smith and Basset, 1977; Pittman and Ryan, 1978) Clusters are also being utilized to model substrate bonding and the reactions of substrates on metal surfaces. (Muetterties, 1975; Muetterties, _ 1979; Muetterties et al, 1979) Because it has been suggested that the tetranuclear phosphine-substituted carbonyls of iridium may serve as catalysts in the hydroformylation of olefins (Drakesmith and Whyman, 1973), this class of compounds is of particular interest. The tetrairidium dodecacarbonyl molecule is a tetrahedral cluster of metal atoms with carbonyl groups attached linearly to the apices of the iridium eolyhedron. The four iridium atoms are held together solely by metal-metal bonds (figure 1,A). (Churchhill and Hutchinson, 1978) Substitution of phos- phine groups for carbonyls results in a change in the skeletal structure of the complex: three of the carbonyl groups occupy bridging positions. The diphosphine-substituted cluster, Ir,(CO),,(PPh;),., sports one basal axial phosphine and one basal equatorial phosphine (figure 1,B). The phosphines in Ir, (CO),(PPh,), occupy basal axial, basal equatorial, and basal equatorial positions (figure 1,C). (Albano et al, 1967) These stereochemistries persist in solution for the analogous Ir,(CO),,(PPh,Me), and Ir, (CO), (PPh,Me), as revealed by *°C-NYR studies. (Stuntz and Shapley, 1977) Carbonyl scrambling has been demonstrated by NMR in the phosphine-substituted clusters, indicating facile terminal-bridging exchange. (Cattermole et al, 1974) 39 Norton reported a cooperative effect in the reactions of the triphenyl phosphine-substituted iridium carbonyls, suggesting the importance of inter- actions among metal atoms in clusters. The kinetic parameters for progressive carbonyl substitution in the cluster revealed a substantial rate enhancement for carbonyl dissociation with progressive phosphine substitution. (Karel and Norton, 1974) Ir, (CO);, —2 Ir, (CO) , , (PPh) 2224 Ir, (CO),,(PPh,), —*8%) tr, (co), @PPh,), These results are often interpreted as being supportive evidence that electronic interactions can be transmitted from neighboring metal atoms to the active catalytic center in a cluster. (Pittman and Ryan, 1978) In contrast, in the analogous cobalt system, no such enormous rate enhancement was noted. (Darensbourg and Incorvia, 1979, 1980). Co (CO) 17 —25 Co, (CO) 1 [P(OMe) 3} —2224 Co, (CO), ,LP (Me) ,], 2-42) co, (co) ,[P(oMe) ,], In this case the ligand replacing CO is the small, good m-acceptor P(QMe),. One possible explanation for the much greater enhancement in the iridium system is the greater steric bulk of the triphenyl phosphine ligand as com- pared to the small trimethyl phosphite group. Sizeable steric acceleration _is well-documented in mononuclear carbonyl phosphine and phosphite derivatives. (Darensbourg and Graves, 1979} The study of CO dissociation in the complexes Ir, (CO) ,[PR,],, R=Me,nBu, Et, i-Pr, Ph has demonstrated that carbonyls are more 40 easily lost when the phosphine is larger. (Darensbourg and Baldwin-Zuschke, 1982) An interesting question follows: In clusters of the type Try, (C0), ,-n (PX), X=OR, R, is the ease with which a carbonyl dissociates related to a electronic cooperativity among the metal atoms, and/or b) spatial confrontations between ligands dictated by the cluster's framework? Thus, it would be helpful to know to what extent the steric qualities or the electronic qualities of the ligand affect the cluster. It would be interesting to investigate CO dissociation in clusters using 1i- gands of varying electronic properties. Such studies could help discern whether any cooperation does take place among metal atoms in the cluster. Mono-phosphine substituted clusters of the type Ir,(CO),,L would be most useful since the steric characteristics of L will be minimized. The first four ligands of interest would be perhaps PF;, P(OCH2CH2C1):, P(OMe),, and P(OEt),. These ligands have very different m-accepting charac- teristics, as judged by the infrared carbonyl stretching frequencies of Ni(CO) ,L where L=PF,, P(OCH,CH,C1),, P(OMe),, P(OEt),. (‘Tolman parameter," Tolman, 1970) The m-accepting capability ordering is: PF, > P(OCH,CH,Cl), > P(OMe), > P(OEt), These four phosphine ligands also have roughly equivalent size as indicated by their respective ligand cone angles: 104°, 110°, 107°, and 109°. (Tolman, 1977) Given ligands of equivalent size but different electronic properties, the rates of dissociation of CO from Ir (CO),,L, L=PF,, P(OCH,CH,Cl),, P(OMe),, P(OEt), would be very enlightening. The rate of carbonyl dissociation could 41 measured by observing the rate of exchange of carbonyls in the complex with a 13CO atmosphere. Carbon-13 enrichment could be detected via carbonyl infrared stretching frequencies. The absorption region for the bridging carbonyls (1820-1700 cm™!) is an especially sensitive indicator of such enrichment; absorption peaks shift to lower frequencies as '°CO is incorporated. If the metal atoms are cooperating, then a better t-acceptor on one metal center should decrease the rate of exchange of carbonyls. One would see the fastest rate for Ir,(CO),,[P(OEt),] and the slowest rate for Ir,(CO),,[PF,]. _ Subsequent studies could investigate these.same ligands in complexes of the type Ir,(CO),;)L2 in an effort to demonstrate the relative contributions: of electronic cooperativity and spatial consideration. References Albano, V.; Bellont, P. and Scatturin, V. Chem. Comm. 1967, 730. Cattermole, P.E.; Orrel, K.G. and Osborne, A.G. J. Chem. Soc., Dalton 1974, 328. Churchhill, M.R. and Hutchinson, J.P. Inorg. Chem. 1978, 17, 3528. Darensbourg, D.J. and Baldwin-Zuschke, B.J. J. Am. Chem. Soc. 1982, 104, 3906. Darensbourg, D.J. and Incorvia, M.J. J. Organometallic Chem. 1979, 171, 89. Darensbourg, D.J. and Incorvia, M.J. Inorg. Chem. 1980, 19, 2585. _Darensbourg, D.J. and Graves, A.H. Inorg. .Chem. 1979, 18, 1257. Drakesmith, A.J. and Whyman, R. J. Chem. Soc., Dalton 1973, 362. Karel, K.J. and Norton, J.R. J. Am. Chem. Soc. 1974, 96, 6812. Muetterties, E.L. Bull. Soc. Chem. Belg. 1975, 84, 959. Muetterties, E.L. Bull. Soc. Chem. Belg. 1976, 85, 451. Muetterties, E.L. Science 1979, 194, 1150. Meutterties, E.L.; Rhodin, T.N.; Bond, E.; Brucker, C.F. and Pretzer, W.R. Chem Rev. 1979, 79, 91. Pittman, C.U. and Ryan, R.C. Chemtech 1978, 170. Smith, A.K. and Basset, J.M. J. Mol. Catalysis 1977, 2, 229. Stumtz, G.F. and Shapley, J.R. J. Am. Chem. Soc. 1977, 99, 607. Tolman, C.A. J. Am. Chem. Soc. 1970, 92, 2953. Tolman, C.A. Chem. Rev. 1977, 77, 313. 43 FIGURE 1 A. Ir,(CO)s5 B. Ir, (CO) o(PPh,)> Proposition II The generation of a conformational change in the Fc region upon antigen binding in the Fab region has long been considered a model for activation of the classical complement cascade. Since C3b and Cab seem to bind only to the Fd portion of IgG in antibody-antigen aggregates (Gadd and Reid, 1981; Campbell, et al., 1980), it seems worthwhile to investigate the possibility that a conformational change in the Fab region upon antigen binding is a necessary condition for C3b or C45 to bind and to take part in the rest of the classical _complement cascade. There is good evidence that unactivated complement components C3 and C4 each contain an internal thioester linkage. Upon cleavage of C4 into Cda and Cab by the active proteinase CIS, a transient reactive group is generated elsewhere on C4b's a-chain that allows C4b to bind covalently to the Fd portion of IgG (Campbell, et al., 1980) or to the surface of erythrocytes. (Law, et al., 1980) The same phenomenon is observed for C3b upon activation by C3 convertase (C4b7Za). (Gadd and Reid, 1981; Law and Levine, 1977; Law, et al., 1979) This reactive group exposed upon activation reacts rapidly with water; thus C4b and C3b are rapidly inactivated unless binding to antibody or cell surface occurs. The interaction between C3b and antibody or cell surface is through an ester or amide bond (Gadd and Reid, 1981; Law and Levine, 1977) in which the acyl group is contributed by a glutamic acid residue from the a-chain of C3b. (Law, et al., 1979a; Law, et al., 1979b) The glutamyl residue and cysteinyl residue that make up this thioester in C3 and C4 are separated by two amino acids in the primary sequence. The thioester site is contained in a sequence of eight amino acids that are almost completely conserved among C3, C4, and a,-macroglobulin. (Thomas, et al., 1982; Campbell, et al., 1981; Swenson and 45 Howard, 1980) The reactivities of the thioester linkage in the unactivated components C3 and C4 toward nucleophiles are much less than the reactivities of the components upon activation. These observations have been explained by a thioester "buried" within the protein; upon cleavage of C4 by CIs or C3 by C3 convertase, the thioester is exposed and provides a very reactive acyl group. (Tack, et al., 1980; Harrison, et al., 1981) This acyl group can then react with nucleophiles on antibodies or cell surfaces, or it can be inactivated through reaction with water. As mentioned earlier in this report, there is evidence that C4b binds to the Fd portion of antibody in antibody-antigen aggregates and that it is this C4B bound to antibody which is hemolytically active. (Campbell, et al., 1980) Binding of C4B elsewhere on the antibody molecule was not detected. Component C3b is also covalently bound to the Fd in antibody-antigen aggregates. (Gadd and Reid, 1981) If activation of C3 or C4 generates a reactive thioester sub- ject to nucleophilic acyl substitution, why does reaction only occur in the Fd portion of the immmoglobulin molecule? A few speculations come to mind immediately. First, perhaps Cl with its long collagen-like extensions actually Cleaves C4 at a place nearer the Fab arms; since the thioester is hydrolysed rapidly, C4b either reacts with Fab (or cell surface) which is near at hand or is inactivated by water. This does not easily explain why binding of C4b to the light chain of the Fab region is not detected. Second, perhaps Cl's presence at the Fc region blocks certain nucleophilic groups of the Fe from reacting with the nascent thioester of C3b or C4b. Third, perhaps binding of antigen increases the reactivity/availability of nucleophilic groups in the Fd portion of IgG. 46 With these questions in mind, it is appropriate to investigate the avail- ability of nucleophilic groups in the IgG molecule. A reactive ester such as pentachlorophenyl acetate might be used to label IgG. The reaction conditions (pH, reagent excess, etc.) could be varied in order to discover conditions under which the more reactive/available nucleophilic groups react with the ester. Once such conditions are established, a double-labelling experiment could be carried out in the presence and absence of antigen. For example, tritiated pentachlorophenyl acetate is allowed to react with antibody with antigen bound. _ Then this *H-labelled IgG is mixed with IgG which has been labelled with '*C- pentachlorophenyl acetate in the absence of antigen. Comparison of *H:!*C ratios obtained should indicate whether the binding of antigen affects nucleo- philic groups on the IgG molecule. Digestion with papain and affinity chroma- tography can separate Fc and Fab regions so that the °H:’"C ratios for the two fragments can be determined. One should realize that the physical presence of antigen may very well affect the availability of nucleophiles at the antigen-binding site. Another phase of the project would be to perform the same experiments under conditions where Cl is also bound to the antibody. The results may reveal whether the presence of Cl on the Fc might be blocking or otherwise affecting nucleophilic groups of the IgG molecule. 47 References Campbell, R.D., Dodds, A.W., and Porter, R.R. Biochem. J. 1980, 189, 67. Campbell, R.D., Gagnon, J., and Porter, R.R. Biochem J. 1981, 199, 359. Gadd, K.J., and Reid, K.B.M. Biochem. J, 1981, 195, 471. Harrison, R.A,, Thomas, M.L., and Tack, B.F. Proc. Natl. Acad. Sci. USA 1981, "78, 7388, Law, S.K. and Levine, R.P. Proc. Natl, Acad. Sci. USA 1977, 74, 2701. Law, S.K., Fearon, D.T., and Levine, R.P. J. Immmol. 1979, 122, 759. (a) | Law, S.K., Lichtenberg, N.A., and Levine, R.P., J. Immunol. 1979, 123, 1388. (b) Law, S.K., Lichtenberg, N.A., Holcombe, F.H., and Levine, R.P. J. Immmol. 1980, 125, 634, Swenson, R., and Howard, J.B. J. Biol. Chem. 1980, 255, 8087. Tack, B.F,, Harrison, R.A., Janatova, J., Thomas, M.L., and Prahl, J.W. Proc. Natl. Acad. Sci. USA 1980, 77, 5764. Thomas, M.L., Janatova, J., Gray, W.R., and Tack, B.F. Proc. Natl. Acad. Sci. USA 1982, 79, 1054, 48 Proposition III More is probably known of the evolution of the mammalian globin gene family than any other system. Vertebrate hemoglobin is composed of two chains each of a-like and 8-like globin polypeptides. The initial interest in these genes was Chiefly their developmental aspects: each type of chain is encoded by a gene under differential, but coordinate control. (Proudfoot et al., 1980; Efstratiadis et al., 1980) Investigation of the structure of globin genes revealed a 3 exon/2 intron _ Structure that was characteristic of every. functional globin gene studied. Further, every gene contained two introns which interrupted the protein-coding sequence at homologous positions. Intron positions and lengths for globin genes of mouse, man, rabbit, chicken, sheep, goat, and frog are given in Table 1. The homology in sequence and similarity in structure of these genes has led to the suggestion that the two introns were present before the duplication that gave rise to separate a- and 8-globin genes, an event estimated to have occurred 500 million years ago. (Dayhoff et al., 1972) The nitrogen-fixing nodules of legumes contain a myoglobin-like protein termed leghemoglobin. (Baulcombe and Verma, 1978} The amino acid sequence homology that leghemoglobin genes share with vertebrate globin genes suggests that leghemoglobins and globins have a common evolutionary origin. (Hunt et al., 1978} The leghemoglobin genes contain three introns, the first and third of which are in homologous positions to the globin gene introns. (Hyldig- Nielsen et al., 1982) The middle "extra't intron has provoked considerable attention. Mitika Go used C%-C* distance maps to demonstrate four structural units in the globin polypeptide. Two of these peptide units correspond to the first 49 and second exon in the globin chains. The remaining two peptide units are coded by the middle exon. If that "extra'’ intron in leghemoglobin were present in globin, each of the four structural mits in the protein would correspond to a separate exon. (Go, 1981) Given these intriguing facts on globin and leghemoglobin gene structure, examination of the gene coding for myoglobin should prove interesting. Compari- sons between the protein structure and amino acid sequence suggest that the myoglobin and globin genes originated from a common phylogenetic ancestor about 600-800 million years ago, before the divergence of o- and 8-globin genes. (Hunt et al., 1978; Czelusniak et al., 1982) Would the myoglobin gene contain three introns at positions homologous to the intron positions in the leghemo- globin gene? Very recently (Blanchetot, et al., February 24, 1983), the myoglobin gene from grey seal has been isolated and characterized. The gene contains two introns at positions precisely homologous to globin introns; however, the introns of length 4800 base pairs and 3400 base pairs are unusually long. These lengths are totally unlike the globin gene intron lengths (Table 1), and the complete myoglobin gene was determined to be 9,200 base pairs long. (Blanchetot, et al., 1983) The "extra'’ intron in the leghemoglobin genes and these surprisingly long introns in the myoglobin genes should focus more attention on the study of the hemoglobin-like proteins of invertebrates. The nomenclature is difficult here. Such proteins in invertebrates are called hemoglobins in the literature; it may or may not be correct to do so. In any case, in the following paragraphs hemoglobin-like proteins in invertebrates will be referred to as hemoglobin. The distribution in invertebrates of the red fluid variously called cruorine, erythrocruorin, hemoglobin, or myoglobin, depending on when it was described and where it is located, has been the object of many studies. (Terwilliger,. 1980) In the 1800's, E. Ray Lankester described the apparent irregular occurence of this "red pigment'' among the invertebrates. (Lankester, 1872) Does this intermittent distribution represent the descent from an ancestral hemoglobin- like protein or the repeated development of oxygen-binding heme proteins as different organisms respond to similar environmental pressures? Svedberg suggested that all hemoglobin structures were based on an oxygen-binding poly- peptide unit with molecular weight 14-17,000 daltons. (Svedberg, 1933) Thus far, the most primitive phylum for which the presence of hemoglobin has been demonstrated is Phylum Protozoa. Two species of Paramecium, P. caudatum (Sato and Tamiya, 1937) and P. tetraurelia (Smith, et al., 1962) have so far been shown to contain hemoglobin. Paramecium hemoglobin is a monoheme poly- peptide of molecular weight 13,400 daltons (Smith, et al., 1962), similar to myoglobin's molecular weight. It exhibits in general spectroscopic and chemical reactivity typical of vertebrate globin and myoglobin. (Smith et al., 1962) The structure of this hemoglobin has not been studied; so its relation to the vertebrate globins cannot be assessed. The complete amino acid sequence (147 residues) has been determined for the monomeric hemoglobin of the common bloodworm, Glycera dibranchiata. (Imamura et al., 1972) Amino acids at 23% of the sites are identical to those in sperm whale myoglobin, and 20% of the positions are identical to human B-globin. Although these degrees of correspondence seem low, the overall conformation of the annelid hemoglobin chain is remarkably similar to the arrangement found in vertebrate globins. Crystallographic studies reveal seven helical segments which are disposed in the familiar "myoglobin fold". (figure 1} (Padlan and Love, 1974) Notably the D helix is absent and the distal 51 histidine in vertebrate globins is occupied by leucine. (Padlan and Love, 1974) The protein has a molecular weight of 15,590 daltons, (Imamura et al., 1972) Since the sequence of this invertebrate hemoglobin gene may prove invalu- able to the understanding of globin evolution, it is proposed that the gene be isolated and sequenced. The protein is not of prohibitively low concentration in the organism; it should be possible to include the gene in a DNA library made from mRNA. (Grunstein and Hogness, 1975; Benton and Davis, 1977) The cDNA library could be screened with a *?P- labelled copy of a vertebrate globin or myoglobin gene. If the degree of homology is not sufficient for the probe to pick out the hemoglobin gene, an alternate method could be attempted. Since the amino acid sequence for the G. dibranchiata hemoglobin is known (Imamura, et al., 1972), a mixture of synthetic oligonucleotides that could code for a portion of the peptide chain could be used as probes (as in Singer- Sam et al., 1983). A problem with this technique is that there will be a range of melting temperatures for a mixture of probes. This might make it difficult to distinguish a low melting, but well-matched probe from a probe that contains more mismatches but that melts at a higher temperature, due to its base composition. Once the cDNA clone is isolated, it can be used to screen a genomic library » constructed from either restriction enzyme fragments or sheared DNA. (Maniatis et al., 1978} The genomic recombinant molecules can be isolated, prepared and cloned (Lacy et al., 1979), and the appropriate fragment could be isolated and sequenced. (Sanger et al., 1980) Knowledge of the sequence of the G. dibranchiata hemoglobin gene may lead to better understanding of the evolution of hemoglobin-like proteins. Possible sequence homologies between this gene and the vertebrate globin and myoglobin genes would be important, as would similarities in intron/exon organization, position, and length. 52 References Baulcombe, D., and Verma, D. Nucl, Acids Res. 1978, 5, 4141. Benton, W.D., and Davis, R.W. Science 1977, 196, 180. Blanchetot, A., Wilson, V., Wood, D., and Jeffreys, Nature 1983, 301, 732. Czelusniak, J., Goodman, M., Hewett-Emmett, D., Weiss, M., Venta, P., and Tashian, R. Nature 1982, 298, 297. Dayhoff, M., Hunt, L., McLaughlin, P., and Jones, D. Atlas of Protein Sequence and Structure, 1972, ed. M. Dayhoff, pp. 17-30. Efstratiadis, A., Posakony, J., Maniatis, T., Lawn, R., O'Connell, C., Spritz, R., DeRiel, J., Forget, B., Weissman, S., Slightom, J., Blechl, A., Smithies, O., Baralle, I., Shoulders, C., Proudfoot, N. Cell 1980, 21, 653. Go, M. Nature 1981, 291, 90. Gmmstein, M., and Hogness, D. Proc. Natl. Acad. Sci. USA 1975, 72, 3961. Hunt, T.L., Hurst-Calderone, S., and Dayhoff, M. Atlas of Protein Sequence and Structure 5, Supplement 3 1978, pp. 229-251. Hyldig-Nielsen, J., Jensen, E., Paludan, K,, Wiborg, O., Garrett, R., Jorgensen, P., Marcker, K. Nucl. Acids Res. 1982, 10, 689. Imamura, T., Baldwin, T., and Riggs, A. J. Biol. Chem. 1972, 247, 2785. Lacy, E., Hardison, R.C., Quon, D., and Maniatis, T. Cell 1979, 18, 1273. Lankester, E.R. Proc. Roy. Soc. Lond., B. 1872, 21, 70. Maniatis, T., Hardison, R., Lacy, E., Lauer, J., O'Connell, C., Quon, D., Sim, D., and Efstratiadis, A. Cell 1978, 15, 687. Padlan, E.A., and Love, W.E. J. Biol. Chem. 1974, 249, 4067. Proudfoot, N.J., Shander, M., Manley, J.L., Gefter, M.L., and Maniatis, T. Science 1980, 209, 1329. Sanger, F., Coulson, A., Barrell, B.Q., Smith, A.J.H., and Roe, B.A. J. Mol. Biol. 1980, 143, 161. 53 Sato, T., and Tamiya, H. Cytologia Fuju Jubilaei Volumen 1937, 1133. Singer-Sam, J., Simmer, R.L., Keith, D.H., Shively, L., Teplitz, M., Itakura, K., Gartler, S., and Riggs, A.D. Proc. Natl. Acad. Sci. USA 1983, 80. Smith, M.H., George, P., and Preer, J. Arch. Biochem. Biophys. 1962, 99, 313. Svedberg, T. J. Biol. Chem. 1933, 103, 311. Terwilliger, R.C. Am. Zool. 1980, 20, 53. 54 cl Tt oT OT 622 L6T S82 089 66 06T ve? £22 6 ely ¢°8 z°9 S ele ¢ v4 if pol’sor 69°89 ze‘ T¢ pOl‘Sor 69°89 ze ‘Te pot. cor 69°89 ze ‘Ts? yo‘ cor 69°89 zo Te goUalajol uot}ISsod proe uotytsod proe uorytsod proe OUTWE ¢SAT ourwe ZSAT ourwe [SAT 9O0T*SOT ze° ts voL‘coL Of ‘62 pot ‘cor O¢ ‘62 SOT‘ VOT Te ‘0 SOT‘ POT Te ‘O¢ SOL‘ vOT Te Oe 001‘ 66 ze* Te O01‘ 66 ze‘ T¢ 0OT ‘66 Ze‘ Ts sdudLayet uot TSod proe outue ZSAI ooPre™ Lvs 906 22S £69 OS8 Bre <OT scl Ovt uor}Esod proe outTwe TSAT (dq) yQsueqT ZSAI (dq) ya8uetT (dq) yqsuaT SAL éSAl SOUS‘) BYT[-ULGOT LO} SUOTIESOg puke syysUaT uoTQUY 611 ‘oq 611 “qT 691 oq 6IT BGT uTgo{souseYysaT ueaqco (dq) yzsusT TSAI 008b~u qn utqoTsoAu [eas Add (sTAgeT *X) Z6L 9 utqoTs-g dol 82- Mig urqo[s-g 38e0O 92t Tg utqoTs-g 31qQqe OTT Ceu! urTqo[s-g asno O<T J urgo[3-g uEuN (staoeT *X) TZT i” utgoTds-0 dol 80T 0 UTqOTS-10 2eO eal 0 uTqoTs-0 asno LIT a9) utqgo[s-0 ueWN (dq) yQsuaz, suod TISAI -T OTqeE 55 "SOS “TZ ‘O86T JP) ‘9 ‘£_ “Swett tM pue “‘wy ‘Aey S'y'e ‘uo 8UTYTA “wu § ‘quatied "$1 —. "SSOv “64 ‘2861 VSN “ToS ‘pesy “TVEN "901d “S*d°G ‘vuto, pue ‘‘N ‘uossTag ‘7] “LES OT “786 “Sey Spry “TOMN “W"y ‘toyoieW pue ‘*y ‘uepnteg ‘-g ‘uasuar ‘*pP SuasTarN- Sip[AH ‘'o £8410qTM “TT "689 ‘OT “2g61 "soy SPtoy *TONN “y ‘ateyoteW pue ‘+g “uesuedlor Sy qqoraey *° O “810qTM ‘“*y ‘uepnteg a “uasua “re ‘UasTOTN- StpIAH ‘OT "2¢d “Tog ‘S86l einzeN ery “Skoljjor pue ‘"( ‘poom ‘*A “UOSTIM “"y yoROYDURTY "6 “L21L “4L “0861 VSO_WTOS "PPV "TIWN “P20Ld ae “Tersury pure “+g ‘yoeqsoy ‘yr ‘seudey °g "Lee “907 ‘6LZ6L 90U9TIS *4y ‘SuRUISSTOM pue ‘‘y ‘toque ‘*¢ ‘310g uag wea ‘uskog ueA */ —, 998 “8T ‘6261 TT90, “dS 10pa7T pue ‘*A I ‘Tazrey ‘ ‘V'd “TexUoy °9 "L£y9 “IZ ‘O86T TI99 ‘stqeruepy pue *9 ‘ Treuuo) 0 "V Sstperqeis3d "Wu ‘UMBT °S "S86L “* Te 39 JoJSYyoULTY UT pazeotpuT ‘e eae poystyqndun “o'r ‘swettTtm pue “*y"y ‘quatzeg “wry ‘Aey tp "$789 ‘252 ‘2861 “woud “Tord ‘fr ‘d'c¢ ‘Teasury pue ‘ws ‘oyutom S*y'g fuoyss “¢ __ "S28 “81 “6L6T 1190 ‘d ‘tepay pue "x ‘eyoTYyDIN °z “pSOL “LZ ‘O86T VSN “19S “Pedy “T2eN “S01g ‘*M*A ‘uRy pue ‘*C*p ‘suassoo9 ‘y's ‘aaqeyqeT] “T I P1qeyt, 02 saduaL1ozoy Figure 1. The "Myoglobin Fold" Atlas of Protein Sequence and Structure, Vol.5, M.0. Dayhoff ,ed. , 1972.