Use of 2D NMR, protein engineering, and molecular modeling to study the hapten-binding site of an antibody Fv fragment against 2-phenyloxazolone

Two-dimensional (2D) 1H NMR spectroscopy was used to study the hapten-binding site of a recombinant antibody Fv fragment expressed in Escherichia coli. Point mutations of residues in the CDR loops of the Fv fragment were designed in order to investigate their influence on hapten binding and to make site-specific assignments of aromatic NMR proton signals. Two tyrosines giving NOEs to the ligand 2-phenyloxazolone were identified, residue 33 in CDR1 of the heavy chain and residue 32 in CDR1 of the light chain. The benzyl portion of 2-phenyloxazolone is located between these two residues. The binding site is close to the surface of the Fv fragment. Comparison with a different anti-2-phenyloxazolone antibody, the crystal structure of which has recently been solved, shows that the general location of the hapten-binding site in both antibodies is similar. However, in the crystallographically solved antibody, the hapten is bound farther from the surface in a pocket created by a short CDR3 loop of the heavy chain. In the binding site identified in the Fv fragment studied in this report, this space is probably filled by the extra seven residues of the CDR3.     

Identification of binding epitope of a monoclonal antibody (Z12) against human TNF-α using computer modeling and deletion mutant technique

The genes of the heavy and light chain variable region (VH, VL) of Z12 antibody against hTNF-α were cloned, and according to the translated sequence of amino acids, the spatial structures of VH and VL domains were modeled by using homology-based modeling method, followed by constructing the whole three-dimensional structure of Fv fragment. The complex model of Fv interacting with hTNF-α was gained with computer-guided molecular docking method, based on which, it was predicted that the epitope recognized by Z12 was from 141 to 146 of hTNF-α. hTNF-α molecule was divided into two fragments of N-terminal region from 1 to 91 and C-terminal region from 92 to 157 with prokaryotic expression. The measured results suggested that the antigenic epitope recognized by Z12 antibody was located in the C-terminal region 92-157 of hTNF-α, proving the predicted result reliable preliminarily. Further experimental results showed that after hTNF-α 141-146 residues were deleted, Z12 antibody almost lost the ability to recognize the mutant, suggesting that the amino acid residues from 141 to 146 of hTNF-α were specially recognized by Z12 antibody.     

Identification of binding epitope of a monoclonal antibody (Z12) against human TNF-α using computer modeling and deletion mutant technique

Humantumornecrosisfactor-a(hTNF-a)isanunglycosylated,pleiotropiccytokinewithnumerousbiologicaleffectsincludingcytotoxicandproinflam-matoryactivities[1].Asrevealedbytheresearchwithmurinemodel,hTNF-playedakeyroleinmanydis-easessuchasrheumatoidarthritis,multiplesclerosisandinflammatoryboweldisease[2],andtherefore,be-cameausefultargetoftherapyforthediseases.Neu-tralizingmonoclonalantibody(mAb)againsthTNF-,asanagentblockinghTNF-activity,hasbeenusedfortherapyofthosediseasesabove[3,4].ThemAb,de…     

Impact of variable domain glycosylation on antibody clearance: an LC/MS characterization

Variable (Fv) domain N-glycosylation sites are found in approximately 20% of human immunoglobulin Gs (IgGs) in addition to the conserved N-glycosylation sites in the C(H)2 domains. The carbohydrate structures of the Fv glycans and their impact on in vivo half-life are not well characterized. Oligosaccharide structures in a humanized anti-Abeta IgG1 monoclonal antibody (Mab) with an N-glycosylation site in the complementary determining region (CDR2) of the heavy chain variable region were elucidated by LC/MS analysis following sequential exoglycosidase treatments of the endoproteinase Lys-C digest. Results showed that the major N-linked oligosaccharide structures in the Fv region have three characteristics (core-fucosylated biantennary oligosaccharides with one or two N-glycolylneuraminic acid [NeuGc] residues, zero or one alpha-linked Gal residue, and zero or one beta-linked GalNAc residue), whereas N-linked oligosaccharides in the Fc region contained typical Fc glycans (core-fucosylated, biantennary oligosaccharides with zero to two Gal residues). To elucidate the contribution of Fv glycans to the half-life of the antibody, a method that allows capture of the Mab and determination of its glycan structures at various time points after administration to mice was developed. Anti-Abeta antibody in mouse serum was immunocaptured by immobilized goat anti-human immunoglobulin Fc(gamma) antibody resin, and the captured material was treated with papain to generate Fab and Fc for LC/MS analysis. Different glycans in the Fc region showed the same clearance rate as demonstrated previously. In contrast to many other non-antibody glycosylated therapeutics, there is no strong correlation between oligosaccharide structures in the Fv region and their clearance rates in vivo. Our data indicated that biantennary oligosaccharides lacking galactosylation had slightly faster clearance rates than other structures in the Fv domain.     

Identification of binding epitope of a monoclonal antibody (Z12) against human TNF-α using computer modeling and deletion mutant technique

The genes of the heavy and light chain variable region (VH, VL) of Z12 antibody against hTNF-α were cloned, and according to the translated sequence of amino acids, the spatial structures of VH and VL domains were modeled by using homology-based modeling method, followed by constructing the whole three-dimensional structure of Fv fragment. The complex model of Fv interacting with hTNF-α was gained with computer-guided molecular docking method, based on which, it was predicted that the epitope recognized by Z12 was from 141 to 146 of hTNF-α. hTNF-α molecule was divided into two fragments of N-terminal region from 1 to 91 and C-terminal region from 92 to 157 with prokaryotic expression. The measured results suggested that the antigenic epitope recognized by Z12 antibody was located in the C-terminal region 92–157 of hTNF-α, proving the predicted result reliable preliminarily. Further experimental results showed that after hTNF-α 141–146 residues were deleted, Z12 antibody almost lost the ability to recognize the mutant, suggesting that the amino acid residues from 141 to 146 of hTNF-α were specially recognized by Z12 antibody.     

Localization within the heavy chain sequence of tyrosyl peptides from the combining sites in a rabbit anti-3-azopyridine antibody preparation

A relatively homogeneous rabbit heavy chain was cleaved by CNBr. Fragment C-1 (the N-terminal half of the heavy chain) was isolated. Reduction and alkylation of C-1 liberated three fragments and partial sequence analysis of these isolated fragments showed that C-1 had been split on the carboxyl-side of Met 84. Similar results were obtained with another anti-hapten antibody preparation in which tyrosyl residues in the combining sites had been labeled. The labeled tyrosyl residues were found in the fragment representing residues 85–253. Since the constant region begins at about residue 120 and the sequences of the tyrosyl peptides from the combining sites are not present in published constant region sequences, these peptides appear to be derived from a variable region between residues 85 and 120.     

Novel single-chain Fv' formats for the generation of immunoliposomes by site-directed coupling

Immunoliposomes generated by coupling of antibodies to the liposomal surface allow for an active targeting of entrapped compounds to diseased areas. Single-chain Fv fragments (scFv) represent the smallest part of an antibody containing the entire antigen-binding site. They can be coupled in a defined and site-directed manner through genetically engineered cysteine residues, for example, those added at the C-terminus. Here, we have performed a comparative analysis of various scFv' variants with cysteine residues present at the end of a C-terminal extension of varying length and composition (HC variants) or introduced in the linker sequence connecting the variable heavy and light chain domain (LC variants). Using a scFv fragment directed against fibroblast activation protein (FAP) as a model antibody, we could show that all variants can be employed for the generation of active immunoliposomes, although the presence of three additional cysteine residues in one scFv' molecule resulted in decreased binding of immunoliposomes compared to that of immunoliposomes generated with scFv' molecules containing only one additional cysteine residue. In order to further improve the scFv' format by reducing the number of additional amino acid residues, we also generated molecules with the hexahistidyl-tag incorporated into the linker sequence together with a cysteine residue either at position 1 or 3 of the linker sequence (LCH variants). These newly designed scFv' molecules may be particularly suitable for the generation of immunoliposomes and other antibody conjugates, limiting the number of additional residues in these antibody molecules to a minimum.     

Structure, function and properties of antibody binding sites

Do antibody combining sites possess general properties that enable them to bind different antigens with varying affinities and to bind novel antigens? Here, we address this question by examining the physical and chemical characteristics most favourable for residues involved in antigen accommodation and binding. Amphipathic amino acids could readily tolerate the change of environment from hydrophilic to hydrophobic that occurs upon antibody-antigen complex formation. Residues that are large and can participate in a wide variety of van der Waals' and electrostatic interactions would permit binding to a range of antigens. Amino acids with flexible side-chains could generate a structurally plastic region, i.e. a binding site possessing the ability to mould itself around the antigen to improve complementarity of the interacting surfaces. Hence, antibodies could bind to an array of novel antigens using a limited set of residues interspersed with more unique residues to which greater binding specificity can be attributed. An individual antibody molecule could thus be cross-reactive and have the capacity to bind structurally similar ligands. The accommodation of variations in antigenic structure by modest combining site flexibility could make an important contribution to immune defence by allowing antibody binding to distinct but closely related pathogens. Tyr and Trp most readily fulfil these catholic physicochemical requirements and thus would be expected to be common in combining sites on theoretical grounds. Experimental support for this comes from three sources, (1) the high frequency of participation by these amino acids in the antigen binding observed in six crystallographically determined antibody-antigen complexes, (2) their frequent occurrence in the putative binding regions of antibodies as determined from structural and sequence data and (3) the potential for movement of their side-chains in known antibody binding sites and model systems. The six bound antigens comprise two small different haptens, non-overlapping regions of the same large protein and a 19 amino acid residue peptide. Out of a total of 85 complementarity determining region positions, only 37 locations (plus 3 framework) are directly involved in antigen interaction. Of these, light chain residue 91 is utilized by all the complexes examined, whilst light chain 32, light chain 96 and heavy chain 33 are employed by five out of the six. The binding sites in known antibody-antigen complexes as well as the postulated combining sites in free Fab fragments show similar characteristics with regard to the types of amino acids present. The possible role of other amino acids is also assessed.(ABSTRACT TRUNCATED AT 400 WORDS)     

NMR-derived model for a peptide-antibody complex

The TE34 monoclonal antibody against cholera toxin peptide 3 (CTP3; VEVPGSQHIDSQKKA) was sequenced and investigated by two-dimensional transferred NOE difference spectroscopy and molecular modeling. The VH sequence of TE34, which does not bind cholera toxin, shares remarkable homology to that of TE32 and TE33, which are both anti-CTP3 antibodies that bind the toxin. However, due to a shortened heavy chain CDR3, TE34 assumes a radically different combining site structure. The assignment of the combining site interactions to specific peptide residues was completed by use of AcIDSQRKA, a truncated peptide analogue in which lysine-13 was substituted by arginine, specific deuteration of individual polypeptide chains of the antibody, and a computer model for the Fv fragment of TE34. NMR-derived distance restraints were then applied to the calculated model of the Fv to generate a three-dimensional structure of the TE34/CTP3 complex. The combining site was found to be a very hydrophobic cavity composed of seven aromatic residues. Charged residues are found in the periphery of the combining site. The peptide residues HIDSQKKA form a beta-turn inside the combining site. The contact area between the peptide and the TE34 antibody is 388 A2, about half of the contact area observed in protein-antibody complexes.     

Dynamical structure of the antibody combining site as studied by 1H-15N shift correlation NMR spectroscopy.

The Fv fragment, which is a smallest antigen-binding unit of immunoglobulin, has been used for a 1H-15N shift correlation NMR study of the dynamical structure of the antibody combining site. Fv has been prepared by clostripain digestion of a mouse anti-dansyl IgG2a monoclonal antibody that lacks the entire CH1 domain. We have previously reported that of the six hypervariable regions, three each from the heavy chain (H1, H2, and H3) and the light chain (L1, L2, and L3), H3 is primarily responsible for the antigen binding in the anti-dansyl Fv fragment. The backbone amide nitrogens of all non-proline amino acid residues in H3 have been multiply labeled with 15N. [15N]T2 relaxation times and hydrogen-deuterium exchange rates of the amide groups of the main chain were measured in the absence and presence of epsilon-dansyl-L-lysine (DNS-Lys). It has been shown that (1) in the absence of DNS-Lys H3 displays a significant degree of internal motion and (2) antigen binding induces a significant change in the dynamical structure of H3.     

鼠抗人纤维蛋白抗体单链Fv片段的人源化分子设计

产生免疫原性的残基主要是位于蛋白表面的暴露残基,为了消除鼠抗体对人的免疫原性,利用表面再塑的方法对本室克隆的鼠抗人纤维蛋白抗体单链Ｆｖ片段进行了人源化分子设计．首先确定了鼠及人Ｆｖ片段的表面残基,在此基础上分析了鼠与人抗体Ｆｖ片段表面残基的差异,将存在差异的鼠抗体的表面残基换成人的,从而实现鼠抗体的人源化．提出了残基最高频率人源化及最相似链人源化两种分子设计方案．人源化的鼠抗人纤维蛋白抗体单链Ｆｖ片段的结构经Ｐｒｏｆｉｌｅｓ－３Ｄ检测证明合理,替换的表面残基的溶剂可及性未变,而且未对ＣＤＲｓ的空间构象产生明显影响,应不会影响与纤维蛋白的亲和力,为鼠抗体人源化实验研究奠定了基础．     

鼠抗人纤维蛋白抗体单链Fv片段的人源化分子设计

产生免疫原性的残基都是位于蛋白表面的暴露残基,为了消除鼠抗体对人的免疫原性,利用表面再塑方法对本室克隆的鼠抗人纤维蛋白抗体单链Fv片断进行了人源化分子设计。首先确定了鼠及人Fv表面残基,在此基础上分析了鼠与人Fv间表面残基的差异,将有差异的鼠表面残基换成人的。提出了残基最高频率人源化及最相似链人源化两种人源化方案。人源化后鼠抗人纤维蛋白抗体单链Fv的结构经Profile-3D验证是合理的,置换的表面残基溶液可及性未变,且未影响CDRs的结构,应不会影响与纤维蛋白的亲和力,为鼠抗体人源化实验研究奠定了基础。     

NMR structure of an anti-gp120 antibody complex with a V3 peptide reveals a surface important for co-receptor binding

BACKGROUND: The protein 0.5beta is a potent strain-specific human immunodeficiency virus type 1 (HIV-1) neutralizing antibody raised against the entire envelope glycoprotein (gp120) of the HIV-1(IIIB) strain. The epitope recognized by 0.5beta is located within the third hypervariable region (V3) of gp120. Recently, several HIV-1 V3 residues involved in co-receptor utilization and selection were identified. RESULTS: Virtually complete sidechain assignment of the variable fragment (Fv) of 0.5beta in complex with the V3(IIIB) peptide P1053 (RKSIRIQRGPGRAFVTIG, in single-letter amino acid code) was accomplished and the combining site structure of 0.5beta Fv complexed with P1053 was solved using multidimensional nuclear magnetic resonance (NMR). Five of the six complementarity determining regions (CDRs) of the antibody adopt standard canonical conformations, whereas CDR3 of the heavy chain assumes an unexpected fold. The epitope recognized by 0.5beta encompasses 14 of the 18 P1053 residues. The bound peptide assumes a beta-hairpin conformation with a QRGPGR loop located at the very center of the binding pocket. The Fv and peptide surface areas buried upon binding are 601 A and 743 A(2), respectively, in the 0.5beta Fv-P1053 mean structure. The surface of P1053 interacting with the antibody is more extensive and the V3 peptide orientation in the binding site is significantly different compared with those derived from the crystal structures of a V3 peptide of the HIV-1 MN strain (V3(MN)) complexed to three different anti-peptide antibodies. CONCLUSIONS: The surface of P1053 that is in contact with the anti-protein antibody 0.5beta is likely to correspond to a solvent-exposed region in the native gp120 molecule. Some residues of this region of gp120 are involved in co-receptor binding, and in discrimination between different chemokine receptors utilized by the protein. Several highly variable residues in the V3 loop limit the specificity of the 0.5beta antibody, helping the virus to escape from the immune system. The highly conserved GPG sequence might have a role in maintaining the beta-hairpin conformation of the V3 loop despite insertions, deletions and mutations in the flanking regions.     

Specificity of interactions of hapten side chains with the combining site of the myeloma protein MOPC 315.

The pKa values of the three histidine residues in the Fv fragment (variable region of the heavy and light chains) of the mouse myeloma protein MOPC 315, measured by high resolution n.m.r. (nuclear magnetic resonance), are 5.9, 6.9 and 8.2. The perturbation of the pKa of one of the histidines (pKa 6.9) on the addition of hapten and the narrow linewidth of its proton resonances suggests that it is at the edge of the combining site. References to the model of the Fv fragment [Padlan, Davies, Pecht, Givol & Wright (1976) Cold Spring Harbor Symp. Quant. Biol. 41, in the press] allows assignment of the three histidine residues, histidine-102H, histidine-97L and histidine-44L. The determination of the pKa of the phosphorus group, by 31P n.m.r., of a homologous series of Dnp- and Tnp- (di- and tri-nitrophenyl) haptens has located a positively charged residue. Molecular-model studies on the conformations of these haptens show that the residue is at the edge of the site. The model suggests that the positively charged residue is either arginine-95L or lysine-52H.     

Humanization of a murine monoclonal antibody by simultaneous optimization of framework and CDR residues.

Optimal protein function often depends on co-operative interactions between amino acid residues distant in the protein primary sequence yet spatially near one another following protein folding. For example, antibody affinity is influenced by interactions of framework residues with complementarity-determining region (CDR) residues. However, despite the abundance of antibody structural information and computational tools the humanization of rodent antibodies for clinical use often results in a significant loss of affinity. To date, antibody engineering efforts have focused either on optimizing CDR residues involved in antigen binding or on optimizing antibody framework residues that serve critical roles in preserving the conformation of CDRs. In the present study a new approach which permits the rapid identification of co-operatively interacting framework and CDR residues was used to simultaneously humanize and optimize a murine antibody directed against CD40. Specifically, a combinatorial library that examined eight potentially important framework positions concomitantly with focused CDR libraries consisting of variants containing random single amino acid mutations in the third CDR of the heavy and light chains was expressed. Multiple anti-CD40 Fab variants containing as few as one murine framework residue and displaying up to approximately 500-fold higher affinity than the initial chimeric Fab were identified. The higher affinity humanized variants demonstrated a co-operative interaction between light chain framework residue Y49 and heavy chain CDR3 residue R/K101 (coupling energy, DeltaGI=0.9 kcal/mol). Screening of combinatorial framework-CDR libraries permits identification of monoclonal antibodies (mAb) with structures optimized for function, including instances in which the antigen induces conformational changes in the mAb. Moreover, the enhanced humanized variants contain fewer murine framework residues and could not be identified by sequential in vitro humanization and affinity muturation strategies. This approach to identifying co-operatively interacting residues is not restricted to antibody-antigen interactions and consequently, may be used broadly to gain insight into protein structure-function relationships, including proteins that serve as catalysts.     

DNA binding by the VH domain of anti-Z-DNA antibody and its modulation by association of the VL domain

mAb Z22 is a highly selective IgG anti-Z-DNA Ab from an immunized C57BL/6 mouse. Previous studies showed that heavy chain CDR3 amino acids are critical for Z-DNA binding by the single chain variable fragment (scFv) comprising both V region heavy chain (VH) and V region light chain (VL) of mAb Z22 and that the VH domain alone binds Z-DNA with an affinity similar to that of whole variable fragment (Fv). To determine whether Z-DNA binding by VH alone and by Fv involves identical complementarity determining region residues, we tested effects of single or multiple amino acid substitutions in recombinant VH, scFv, and associated VH-VL heterodimers. Each recombinant product was a fusion protein with a B domain of Staphylococcal protein A (SPA). Z22VH-SPA alone was not highly selective; it bound strongly to other polynucleotides, particularly polypyrimidines, and ssDNA as well as to Z-DNA. In contrast, scFv-SPA or associated VH-VL dimers bound only to Z-DNA. VL-SPA domains bound weakly to Z-DNA; SPA alone did not bind. Introduction of multiple substitutions revealed that the third complementarity determining region of the heavy chain (CDR3H) was critical for both VH and scFv binding to Z-DNA. However, single substitutions that eliminated or markedly reduced Z-DNA binding by scFv instead caused a modest increase or no reduction in binding by VH alone. Association of VH-SPA with Z22VL-SPA restored both the effects of single substitutions and Z-DNA selectivity seen with Fv and intact Ab. Polypyrimidine and ssDNA binding by the isolated VH domain of immunization-induced anti-Z-DNA Ab resembles the activity of natural autoantibodies and suggests that VH-dependent binding to a ligand mimicked by polypyrimidines may play a role in B cell selection before immunization with Z-DNA.     

Structural consequences of mutations in interfacial Tyr residues of a protein antigen-antibody complex. The case of HyHEL-10-HEL

Tyrosine is an important amino acid in protein-protein interaction hot spots. In particular, many Tyr residues are located in the antigen-binding sites of antibodies and endow high affinity and high specificity to these antibodies. To investigate the role of interfacial Tyr residues in protein-protein interactions, we performed crystallographic studies and thermodynamic analyses of the interaction between hen egg lysozyme (HEL) and the anti-HEL antibody HyHEL-10 Fv fragment. HyHEL-10 has six Tyr residues in its antigen-binding site, which were systematically mutated to Phe and Ala using site-directed mutagenesis. The crystal structures revealed several critical roles for these Tyr residues in the interaction between HEL and HyHEL-10 as follows: 1) the aromatic ring of Tyr-50 in the light chain (LTyr-50) was important for the correct ternary structure of variable regions of the immunoglobulin light chain and heavy chain and of HEL; 2) deletion of the hydroxyl group of Tyr-50 in the heavy chain (HTyr-50) resulted in structural changes in the antigen-antibody interface; and 3) the side chains of HTyr-33 and HTyr-53 may help induce fitting of the antibody to the antigen. Hot spot Tyr residues may contribute to the high affinity and high specificity of the antigen-antibody interaction through a diverse set of structural and thermodynamic interactions.     

The V-region sequence of the H chain from a third rabbit anti-pneumococcal antibody.

The amino acid sequence of the V (variable) region of the heavy (H) chain of rabbit antibody BS-1, raised against type III pneumococcal vaccine, is reported. Together with the sequence data of the V region of the light (L) chain previously determined [Jaton (1974a) Biochem. J. 141, 1-13], the present work completes the analysis of the V domain of the homogeneous antibody BS-1. The V domains (VL + VH regions) of this antibody are compared with those of two other anti-(type III) pneumococcal antibodies BS-5 and K-25 [Jaton (1975) Biochem. J. 147, 235-247]. Except for the second hypervariable section of the L chains, these antibodies have very different sequences in the hypervariable segments of the V domains. Within the third hypervariable region of the H chain, each antibody has a different length: BS-1 is three amino acids shorter than K-25 and two amino acids shorter than BS-5. When the sequences in that section are aligned for maximal homology, only two residues, glycine-97 and leucine-101, are common to the three antibodies. On the basis of the amino acid sequences of these three anti-pneumococcal antibodies, the results do not support the concept of a simple correlation between primary structure in the hypervariable sections (known to determine the shape of the combining site) and antigen-binding specificity.     

Construction of a semisynthetic antibody library using trinucleotide oligos.

A semisynthetic antibody library composed of single chain Fv fragments (scFv) was constructed by replacing the heavy chain CDR3 region of a human scFv by a random sequence of eight amino acids using trinucleotide codons. After cloning into a phage display vector, an antibody library was generated with a complexity of 8 x 10(8) independent clones. The library was screened for binders to dinitrophenol, fluorescein isothiocyanate and 3-nitro-4-hydroxy-5-iodophenylacetic acid. scFv antibodies that specifically bound the antigen were obtained in each case.     

Two-Dimensional nmr studies of the interactions between a peptide of cholera toxin and monoclonal antibodies

To increase our understanding of the molecular basis for antibody specificity and for the cross-reactivity of antipeptide antibodies with native proteins, it is important to study the three-dimensional structure of antibody complexes with their peptide antigens. For this purpose it may not be necessary to solve the structure of the whole antibody complex but rather to concentrate on elucidating the combining site structure, the interactions of the antibody with its antigen, and the bound peptide conformation. To extract the information about antibody–peptide interactions and intramolecular interactions in the bound ligand from the complicated and unresolved spectrum of the Fab–peptide complex (Fab: antibody fragment made of Fv—the antibody fragment composed of the variable regions of the light and heavy chains forming a single combining site for the antigen—the light chain, and the first heavy chain constant regions), an nmr methodology based on measurements of two-dimensional transferred nuclear Overhauser effect (NOE) difference spectra was developed. Using this methodology the interactions of three monoclonal antibodies with a cholera toxin peptide were studied. The observed interactions were assigned to the antibody protons involved by specific deuteration of aromatic amino acids and specific chain labeling, and by using a predicted model for the structure of the antibody combining site. The assigned NOE interactions were translated to restraints on interproton distances in the complex that were used to dock the peptide into calculated models for the antibodies combining sites. Comparison of the interactions of three antibodies against a cholera toxin peptide (CTP3). which differ in their cross-reactivity with the toxin, yields information about the size and conformation of antigenic determinants recognized by the antibodies, the structure of their combining sites, and relationships between antibodies' primary structure and their interactions with peptide antigens. © 1994 John Wiley & Sons, Inc.