Protein docking using continuum electrostatics and geometric fit.

The computer program DOT quickly finds low-energy docked structures for two proteins by performing a systematic search over six degrees of freedom. A novel feature of DOT is its energy function, which is the sum of both a Poisson-Boltzmann electrostatic energy and a van der Waals energy, each represented as a grid-based correlation function. DOT evaluates the energy of interaction for many orientations of the moving molecule and maintains separate lists scored by either the electrostatic energy, the van der Waals energy or the composite sum of both. The free energy is obtained by summing the Boltzmann factor over all rotations at each grid point. Three important findings are presented. First, for a wide variety of protein-protein interactions, the composite-energy function is shown to produce larger clusters of correct answers than found by scoring with either van der Waals energy (geometric fit) or electrostatic energy alone. Second, free-energy clusters are demonstrated to be indicators of binding sites. Third, the contributions of electrostatic and attractive van der Waals energies to the total energy term appropriately reflect the nature of the various types of protein-protein interactions studied.

Stabilization of bound polycyclic aromatic hydrocarbons by a pi-cation interaction.

Proteins can use aromatic side-chains to stabilize bound cationic ligands through cation-pi interactions. Here, we report the first example of the reciprocal process, termed pi-cation, in which a cationic protein side-chain stabilizes a neutral aromatic ligand. Site-directed mutagenesis revealed that an arginine side-chain located in the deep binding pocket of a monoclonal antibody (4D5) is essential for binding the neutral polynuclear aromatic hydrocarbon benzo[a]pyrene. This Arg was very likely selected for in the primary response, further underscoring the importance of the pi-cation interaction for ligand binding, which should be considered in protein analysis and design when ligands include aromatic groups.

Definition of the interaction domain for cytochrome c on cytochrome c oxidase. III. Prediction of the docked complex by a complete, systematic search.

The electron transfer complex between bovine cytochrome c oxidase and horse cytochrome c has been predicted with the docking program DOT, which performs a complete, systematic search over all six rotational and translational degrees of freedom. Energies for over 36 billion configurations were calculated, providing a free-energy landscape showing guidance of positively charged cytochrome c to the negative region on the cytochrome c oxidase surface formed by subunit II. In a representative configuration, the solvent-exposed cytochrome c heme edge is within 4 A of the indole ring of subunit II residue Trp(104), indicating a likely electron transfer path. These two groups are surrounded by a small, hydrophobic contact region, which is surrounded by electrostatically complementary hydrophilic interactions. Cytochrome c/cytochrome c oxidase interactions of Lys(13) with Asp(119) and Lys(72) with Gln(103) and Asp(158) are the most critical polar interactions due to their proximity to the hydrophobic region and exclusion from bulk solvent. The predicted complex matches previous mutagenesis, binding, and time-resolved kinetics studies that implicate Trp(104) in electron transfer and show the importance of specific charged residues to protein affinity. Electrostatic forces not only enhance long range protein/protein association; they also predominate in short range alignment, creating the transient interaction needed for rapid turnover.

Unraveling the effect of changes in conformation and compactness at the antibody V(L)-V(H) interface upon antigen binding.

We have analyzed conformational changes that occur at the interface between the light (V(L)) and heavy (V(H)) chains in antibody variable fragments upon binding to antigens. We wrote and applied the Tiny Probe program that computes the buried atomic contact surface area of three-dimensional structures to evaluate changes in compactness of the V(L)-V(H) interface between bound and unbound antibodies. We found three categories of these changes, which correlated with the size of the antigen. Upon binding, medium-sized nonprotein antigens cause an opening of the V(L)-V(H) interface (less compact), small antigens or haptens cause a closure of the interface (more compact), whereas large protein antigens have little effect on the compactness of the V(L)-V(H) interface. The largest changes in the atomic buried contact surface area at the V(L)-V(H) interface occur in residue pairs providing two 'shock absorbers' between the edge beta-strands of the V(L) and V(H) beta-sheets forming the antibody binding site. Importantly, the correlation between the size of antigens and conformational changes indicates that the V(L)-V(H) interface in antibodies plays a significant role in the antigen binding process. Furthermore, as the energy involved in such a motion is significant (up to 3 kcal/mol), these results provide a general mechanism for how residues distant from the combining site can significantly alter the affinity of an antibody for its antigen. Thus, mutations introduced at the V(L)-V(H) interface can be used to change antibody binding affinity with antigens. Due to the tightly packed V(L)-V(H) interface, the introduction of random mutations is not advisable. Rather our analysis suggests that concerted mutations of residues preceding CDRL2 and following CDRH3 or residues preceding CDRH2 and at the end of CDRL3 are most likely to alter or improve antigen binding affinity.

Structural basis for amide hydrolysis catalyzed by the 43C9 antibody.

Among catalytic antibodies, the well-characterized antibody 43C9 is unique in its ability to catalyze the difficult, but desirable, reaction of selective amide hydrolysis. The crystallographic structures that we present here for the single-chain variable fragment of the 43C9 antibody, both with and without the bound product p -nitrophenol, strongly support and extend the structural and mechanistic information previously provided by a three-dimensional computational model, together with extensive biochemical, kinetics, and mutagenesis results. The structures reveal an unexpected extended beta-sheet conformation of the third complementarity determining region of the heavy chain, which may be coupled to the novel indole ring orientation of the adjacent Trp H103. This unusual conformation creates an antigen-binding site that is significantly deeper than predicted in the computational model, with a hydrophobic pocket that encloses the p -nitrophenol product. Despite these differences, the previously proposed roles for Arg L96 in transition-state stabilization and for His L91 as the nucleophile that forms a covalent acyl-antibody intermediate are fully supported by the crystallographic structures. His L91 is now centered at the bottom of the antigen-binding site with the imidazole ring poised for nucleophilic attack. His L91, Arg L96, and the bound p -nitrophenol are linked into a hydrogen-bonding network by two well-ordered water molecules. These water molecules may mimic the positions of the phosphonamidate oxygen atoms of the antigen, which in turn mimic the transition state of the reaction. This network also contains His H35, suggesting that this residue may also stabilize the transition-states. A possible proton-transfer pathway from His L91 through two tyrosine residues may assist nucleophilic attack. Although transition-state stabilization is commonly observed in esterolytic antibodies, nucleophilic attack appears to be unique to 43C9 and accounts for the unusually high catalytic activity of this antibody.

Monoclonal antibody-based ELISAs for part-per-billion determination of polycyclic aromatic hydrocarbons: effects of haptens and formats on sensitivity and specificity.

As a first step toward developing sensitive enzyme-linked immunosorbent assays (ELISAs) for multianalyte detection of polycyclic aromatic hydrocarbons (PAHs), haptens with different lengths of carboxylic acid spacers at various positions were derived from naphthalene, fluorene, anthracene, phenanthrene, pyrene, fluoranthene, chrysene, and benzo[a]pyrene (BaP). These haptens were coupled with bovine serum albumin (BSA) to form competitor conjugates. All of these haptens were recognized to different extents by monoclonal antibodies (MAbs) 4D5 and 10C10 originally derived by Gomes and Santella (Chem. Res. Toxicol. 1990, 3, 307-310). The most sensitive indirect ELISAs were obtained by coating wells with the least competitive conjugates. Direct ELISAs using horseradish peroxidase conjugates of pyrene and BaP were less sensitive. The MAbs bound BaP with spacers at either C1 or C6. The cross-reactivity profiles of the eight PAHs were different with each PAH-BSA conjugate used as coating antigen. The ELISA results for BaP closely correlated with those by gas chromatography (GC), but the detection limit of the ELISA was approximately 150-fold more sensitive than that of GC, with 2-600 nM spike recoveries of 80-127% from human urine and canal and tap water.

Complex formation between Azotobacter vinelandii ferredoxin I and its physiological electron donor NADPH-ferredoxin reductase.

In Azotobacter vinelandii, deletion of the fdxA gene, which encodes ferredoxin I (FdI), leads to activation of the expression of the fpr gene, which encodes NADPH-ferredoxin reductase (FPR). In order to investigate the relationship of these two proteins further, the interactions of the two purified proteins have been examined. AvFdI forms a specific 1:1 cross-linked complex with AvFPR through ionic interactions formed between the Lys residues of FPR and Asp/Glu residues of FdI. The Lys in FPR has been identified as Lys258, a residue that forms a salt bridge with one of the phosphate oxygens of FAD in the absence of FdI. UV-Vis and circular dichroism data show that on binding FdI, the spectrum of the FPR flavin is hyperchromatic and red-shifted, confirming the interaction region close to the FAD. Cytochrome c reductase assays and electron paramagnetic resonance data show that electron transfer between the two proteins is pH-dependent and that the [3Fe-4S]+ cluster of FdI is specifically reduced by NADPH via FPR, suggesting that the [3Fe-4S] cluster is near FAD in the complex. To further investigate the FPR:FdI interaction, the electrostatic potentials for each protein were calculated. Strongly negative regions around the [3Fe-4S] cluster of FdI are electrostatically complementary with a strongly positive region overlaying the FAD of FPR, centered on Lys258. These proposed interactions of FdI with FPR are consistent with cross-linking, peptide mapping, spectroscopic, and electron transfer data and strongly support the suggestion that the two proteins are physiological redox partners.

Harmful somatic mutations: lessons from the dark side.

The ability of somatic mutation to modify the course of an immune response is well documented. However, emphasis has been placed almost exclusively on the ability of somatic mutation to improve the functional characteristics of representative antibodies. The harmful effects of somatic mutation, its dark side, have been far less well characterized. Yet evidence suggests that the number of B cells directed to wastage pathways as a result of harmful somatic mutation probably far exceeds the number of cells whose antibodies have been improved. Here we review our recent findings in understanding the structural and functional consequences of V-region mutation.

Consensus chemistry and beta-turn conformation of the active core of the insect kinin neuropeptide family.

BACKGROUND: Neuropeptides are examples of small, flexible molecules that bind to receptors and induce signal transduction, thereby eliciting biological activity. The multifunctional insect kinin neuropeptides retain full activity when reduced to only their carboxy-terminal pentapeptide (Phe1-X2-X3-Trp4-Gly5-NH2), thereby allowing extensive structure-function studies and conformational analysis. RESULTS: A combined experimental and theoretical analysis of the insect kinin carboxy-terminal pentapeptide was used to probe the role of each residue, define the bioactive conformation, and design a constrained bioactive analog. Coupling receptor-binding data with two biological activity assays allowed receptor binding and signal transduction to be differentiated. A preferred beta-turn conformation, found for residues 1-4 by molecular dynamics simulations, was tested by designing a conformationally restricted cyclic hexapeptide. This cyclic analog showed a preference for the beta-turn conformation, as shown by a conformational search and nuclear magnetic resonance spectroscopy, and it showed stronger receptor binding but decreased activity relative to highly active linear analogs. CONCLUSIONS: Each residue of the insect kinin carboxy-terminal pentapeptide has a distinct role in conformational preference, specific receptor interactions or signal transduction. The beta-turn preference of residues Phe1-X2-X3-Trp4 implicates this as the bioactive conformation. The amidated carboxyl terminus, required for activity in many neuropeptide families, may be generally important for signal transduction and its inclusion may therefore be essential for agonist design.

Tolerance of single, but not multiple, amino acid replacements in antibody VH CDR 2: a means of minimizing B cell wastage from somatic hypermutation?

Mutations in the heavy chain complementarity determining region 2 (CDR2) of the phosphocholine-specific T15 Ab can have a dramatic effect on the ability of the Ab to bind Ag. A panel of multisite mutants that had lost detectable binding to phosphocholine-containing Ags was previously created by saturation mutagenesis of the CDR2 region of T15. Based on the predicted importance of amino acid changes represented in the multisite mutants, we have created single-site mutations, yielding a panel of Abs with which to test 17 of the 19 CDR2 residues. Of the 17 positions examined, only one, Arg52, is intolerant to change, yielding a nonbinder phenotype even with conservative amino acid replacement. Mutation at two other sites, Ala50 and Tyr55, can yield a nonbinder phenotype depending on the amino acid replacement. Single-site mutations of the remaining 14 positions allowed retention of binding ability. Thus, except for positions 50, 52, and 55, multiple mutations must be introduced into the CDR2 region to create a nonbinder phenotype. We provide a newly refined model of T15, illustrating the structure and the interactions of the CDR2 region. Our results imply that introduction of point mutations would not normally delete Ag-binding ability until two or more mutations had accumulated. This would minimize potentially harmful effects of somatic mutation on Ig V region genes and improve the chance of survival for an Ab such as T15, which in its unmutated form is already well suited to bind Ag.

Humanization and molecular modeling of the anti-CD4 monoclonal antibody, OKT4A.

OKT4A, a murine mAb that recognizes an epitope on the CD4 receptor, is a potent immunosuppressive agent in vitro and in a variety of nonhuman primate models of graft rejection and autoimmune disease. Initial human cardiac transplant trials suggest that OKT4A does not cause either cytokine release syndrome or CD4+ cell depletion, but does induce a human anti-mouse Ab (HAMA) response despite strong concurrent immunosuppression. To further investigate the potential of OKT4A as an immunomodulator, it was necessary to decrease its immunogenicity. Therefore, we developed a humanized version of this Ab (gOKT4A-4), which has the same binding affinity and in vitro immunosuppressive properties of OKT4A, but retains only three murine sequence-derived amino acid residues outside of the complementarity-determining regions (CDRs). Detailed computer modeling of both OKT4A and gOKT4A-4 provided a computational rationale for the changes necessary to regain activity after humanization. This has also provided a plausible representation of the Ag binding site. Preliminary clinical results with gOKT4A-4 suggest that we have eliminated the immunogenicity observed in the parent murine Ab.

A nonpeptidal peptidomimetic agonist of the insect FLRFamide myosuppressin family.

Benzethonium chloride (Bztc) is the first totally nonpeptide ligand for an insect, indeed an invertebrate, peptide receptor. Bztc mimics the inhibitory physiological activity of the myosuppressins, a subfamily of the FLRFamides, in three different insect bioassay systems. The inhibitory action of leucomyosuppressin and the nonpeptide Bztc in both the cockroach hindgut and the mealworm neuromuscular junction can be blocked by the lipoxygenase inhibitor, nordihydroguaiaretic acid, providing evidence for similar modes of action. Lipoxygenase metabolites of arachidonic acid may mediate inhibition of neuromuscular transmission by these two factors. In addition, Bztc competitively displaces a radiolabeled myosuppressin analogue from high- and low-affinity receptors of the locust oviduct. Thus, the nonpeptide interacts with both binding and activating regions of myosuppressin receptors. Molecular dynamics experiments in which selected functional groups of Bztc were fit onto corresponding functional groups of low-energy myosuppressin pentapeptide structures indicate how Bztc may mimic the myosuppressins at a molecular level. The discovery of Bztc as a nonpeptidal peptidomimetic analogue provides an opportunity to develop new pest management strategies by targeting an insect's own peptide receptor.

Enhancement and destruction of antibody function by somatic mutation: unequal occurrence is controlled by V gene combinatorial associations.

We examined the positive and negative effects of somatic mutation on antibody function using saturation mutagenesis in vitro to mimic the potential of the in vivo process to diversify antibodies. Identical mutations were introduced into the second complementarity determining region of two anti-phosphocholine antibodies, T15 and D16, which share the same germline VH gene sequence. T15 predominates in primary responses and does not undergo affinity maturation. D16 is representative of antibodies that co-dominate in memory responses and do undergo affinity maturation. We previously reported that > 50% of T15 mutants had decreased antigen binding capacity. To test if this high frequency of binding loss was unique to T15 or a consequence of random point mutations applicable to other combining sites, we analyzed the same mutations in D16. We show that D16 suffers a similar loss of function, indicating an equally high potential for B-cell wastage. However, only D16 displayed the capacity for somatic mutation to improve antigen binding, which should enhance its persistence in memory responses. Mutation of residues contacting the haptenic group, as determined by molecular modeling, did not improve binding. Instead, productive mutations occurred in residues that either contacted carrier protein or were distant from the antigen binding site, possibly increasing binding site flexibility through long-range effects. Targeting such residues for mutation should aid in the rational design of improved antibodies.

Pseudodipeptide analogs of the pyrokinin/PBAN (FXPRLa) insect neuropeptide family containing carbocyclic Pro-mimetic conformational components.

Three N-terminal amino acid residues of the C-terminal core pentapeptide Phe-X-Pro-Arg-Leu-NH2 (X = Gly, Ser, Thr, Val) of the pryokinin/PBAN insect neuropeptide family were replaced by nonpeptide moieties. To reestablish some of the conformational properties lost upon removal of the peptide bonds and Pro of the three amino acid residue block, carbocyclic Pro-mimetic components were incorporated into pseudodipeptide analogs. The most active analog contained a trans-DL-1,2-cyclopentanedicarboxyl carbocyclic component and proved to be over 3 orders of magnitude more potent than a simple, straight chain pseudodipeptide analog and approached the potency of the pentapeptide core in a cockroach hindgut myotropic bioassay. The pseudodipeptide analog retains a critical carbonyl residue which can participate in a hydrogen bond that stabilizes a beta-turn conformation in the active core region of the pyrokinin/PBAN peptides. This study demonstrates that knowledge of active conformation can be used to enhance the biological potency of pseudopeptide mimetic analogs of insect neuropeptides. The analogs represent a milestone in the development of pseudopeptide and nonpeptide mimetic analogs of this peptide family, which has been associated with such critical physiological processes as hindgut and oviduct contraction, pheromone biosynthesis, diapause induction, and induction of melanization and reddish coloration in a variety of insects. Mimetic analogs are potentially valuable tools to insect neuroendocrinologists studying these physiological processes and/or engaged in the development of future pest management strategies.

Metalloantibody design.

Metal-binding sites were designed within the antigen-binding pocket of the catalytic antibody 43C9 based on a 3-dimensional antibody model and crystallographic structures of Zn-binding metalloenzymes. These tetrahedral Zn-binding sites were designed to mimic both secondary and tertiary structural characteristics of catalytic metal sites in enzymes. Each site was planned to have two His ligands across from each other on adjacent antiparallel beta-strands. Sites were selected to sequester the metal ion from bulk solvent and place an open metal coordination position next to the antigen or potential substrates. Three distinct metal-site designs, with ligands in the variable light domain, in the variable heavy domain, and in both domains, were later implemented experimentally and shown spectroscopically to bind metal ions as predicted. These results demonstrate the success of our design approach, the versatility of the antibody structure for metalloprotein design, and the validity of the 3-dimensional model. The ability to predictably design multiple metal sites in the ordered antigen-recognition region at the bottom of the pocket allows tuning of metal ion placement and enhances the likelihood of interaction with putative substrates.

Predicting molecular interactions and inducible complementarity: fragment docking of Fab-peptide complexes.

Antibody-antigen interactions are representative of a broad class of receptor-ligand interactions involving both specificity and potential inducible complementarity. To test possible mechanisms of antigen-antibody recognition and specificity computationally, we have used a Metropolis Monte Carlo algorithm to dock fragments of the epitope Glu-Val-Val-Pro-His-Lys-Lys to the X-ray structures of both the free and the complexed Fab of the antibody B13I2 (raised against the C-helix of myohemerythrin). The fragments Pro-His and Val-Pro-His, which contain residues experimentally identified as important for binding, docked correctly to both structures, but all tetrapeptide and larger fragments docked correctly only to the complexed Fab, even when torsional flexibility was added to the ligand. However, only tetrapeptide and larger fragments showed significantly more favorable energies when docked to the complexed Fab coordinates than when docked to either the free Fab or a non-specific site remote from the combining site. Comparison of the free and complexed B13I2 structures revealed that atoms within 5 A of Val-Pro-His showed little movement upon peptide binding, but atoms within 5 A of the other four epitope residues showed greater movements. These results computationally distinguished recognition and binding processes with practical implications for drug design strategies. Overall, this new fragment docking approach establishes distinct roles for the "lock-and-key" (recognition) and the "handshake" (binding) paradigms in antibody-antigen interaction, suggests an incremental approach to incorporating flexibility in computational docking, and identifies critical regions within receptor binding sites for ligand recognition.

Site-directed mutagenesis of a catalytic antibody: an arginine and a histidine residue play key roles.

Individual residues important for ligand binding and catalytic activity were identified by computer modeling and investigated by site-directed mutagenesis for catalytic antibody 43C9, which accelerates amide hydrolysis by a factor of 10(6). On the basis of a computer model, Tyr L32, His L91, Arg L96, His H35, and Tyr H95 were chosen for replacement by site-directed mutagenesis. To facilitate these studies, an expression system was developed in which properly folded 43C9 single-chain antibody was secreted from an engineered Escherichia coli host. Substitution of His L91 by Gln produced a mutant with no catalytic activity, but whose affinities for ligands were nearly the same as those of the wild-type, identifying His L91 as the nucleophile that forms the acyl intermediate implicated by previous kinetic studies. Arg L96 is also critical for catalytic activity and appears to function as a oxyanion hole for the tetrahedral transition states. Two substitutions for His H35 resulted in mutant proteins with no catalytic activity as well as altered affinities for ligands, indicating an important structural role for this residue. Substitutions for Tyr L32 and Tyr H95 were made in an attempt to improve the catalytic efficiency of 43C9. The results of these mutations allow us to propose a mechanism for 43C9-catalyzed hydrolysis: Substrate binding to 43C9 orients the scissile carbonyl group adjacent to both the His L91 and Arg L96 side chains. The imidazole of His L91 acts as a nucleophile, forming an acyl-antibody intermediate that breaks down by hydroxide attack to afford the products and regenerate the catalyst.

Catalytic antibody model and mutagenesis implicate arginine in transition-state stabilization.

To probe the mechanism of the catalytic antibody NPN43C9, we have constructed a three-dimensional model of the NPN43C9 variable region using our antibody structural database (ASD), which takes maximal advantage of immunoglobulin sequence and structural information. The ASD contains separately superimposed variable light and variable heavy chains, which reveal not only conserved backbone structure, but also structurally conserved side-chain conformations. The NPN43C9 model revealed that the guanidinium group of light chain Arg L96 was positioned at the bottom of the antigen-binding site and formed a salt bridge with the antigen's phosphonamidate group, which mimics the negatively charged, tetrahedral transition states in the hydrolysis reaction. Thus, the model predicts both binding and catalytic functions for Arg L96, which previously had not been implicated in either. First, Arg L96 should enhance antigen binding by electrostatically complementing the negative charge of the antigen, which is buried upon complex formation. Second, Arg L96 should promote catalysis by electrostatically stabilizing the negatively charged transition states formed during catalysis. These hypotheses were tested experimentally by design and characterization of the R-L96-Q mutant, in which Arg L96 was replaced with Gln by site-directed mutagenesis. As predicted, antigen binding in the R-L96-Q mutant was decreased relative to that in the parent NPN43C9 antibody, but binding of antigen fragments lacking the phosphonamidate group was retained. In addition, the R-L96-Q mutant had no detectable esterase activity. Thus, the computational model and experimental results together suggest a mechanism by which the catalytic antibody NPN43C9 stabilizes high-energy transition states during catalysis.