Structure of the Fab Fragment of F105, a Broadly R
http://www.100md.com
病菌学杂志 2005年第20期
Department of Chemistry and Biochemistry, Montana State University, Bozeman, Montana 59717
Beth Israel-Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02115
ABSTRACT
We have determined the crystal structure of the Fab fragment from F105, a broadly reactive human antibody with limited potency that recognizes the CD4 binding site of gp120. The structure reveals an extended CDR H3 loop with a phenylalanine residue at the apex and shows a striking pattern of serine and tyrosine residues. Modeling the interaction between gp120 and F105 suggests that the phenylalanine may recognize the binding pocket of gp120 used by Phe43 of CD4 and that numerous tyrosine and serine residues form hydrogen bonds with the main chain atoms of gp120. A comparison of the F105 structure to that of immunoglobulin G1 b12, a much more potent and broadly neutralizing antibody with an overlapping epitope, suggests similarities that contribute to the broad recognition of human immunodeficiency virus by both antibodies. While the putative epitope for F105 shows significant overlap with that predicted for b12, it appears to differ from the b12 epitope in extending across the interface between the inner and outer domains of gp120. In contrast, the CDR loops of b12 appear to interact predominantly with the outer domain of gp120. The difference between the predicted epitopes for b12 and F105 suggests that the unique potency of b12 may arise from its ability to avoid the interface between the inner and outer domains of gp120.
INTRODUCTION
A key step in the development of a successful vaccine against human immunodeficiency virus (HIV) will be the design of immunogens capable of generating an effective humoral immune response (3). Antibodies with two features characterize such a response. They must show, at the same time, both potency and broad specificity. A number of human antibodies that recognize elements of the conserved CD4 binding site of HIV type 1 (HIV-1) gp120 have now been isolated from HIV-infected patients (2, 11, 16, 24, 38, 40, 47, 49, 59, 65). This includes immunoglobulin G1 (IgG1) b12 (38, 47, 53), one of the most potent and broadly reactive anti-HIV antibodies known. In contrast, other CD4 binding-site antibodies are less potent towards many clinical isolates, even though they may be broadly cross-reactive. F105, the subject of this paper, is representative of the latter group.
F105 is an IgG1 human monoclonal antibody isolated from an HIV-infected individual (49). It binds to the CD4 binding sites of both trimeric and monomeric gp120 and is capable of neutralizing various strains of HIV (e.g., IIIB [HXBc2], MN, RF, and SF2) (7, 49, 58) but is less successful against many primary clinical isolates (12, 32). F105 did not show evidence of anti-HIV-1 activity or a viral load decrease in a phase I dose-escalation study (5, 70). However, in triple and quadruple combination therapies with anti-HIV monoclonal antibodies (2F5, 2G12, and 694/98D) with other specificities, a complete and synergistic neutralization of the SHIV-Vpu+ chimeric simian-human immunodeficiency virus was seen in macaque peripheral blood mononuclear cells in vitro and in an in vivo macaque model that mimics mucosal exposure during intrapartum virus transmission (1, 31).
Crystallographic studies of the ternary complex of the HIV gp120 core, CD4, and antibody 17b provided the first look at the structure of gp120 and its interactions with CD4 (26-28, 65). CD4 was found to bind at the nexus of the inner domain, the outer domain, and the bridging sheet of gp120 (26, 27). A large body of biochemical and biophysical data indicates a considerable conformational change in gp120 upon binding to CD4 (4, 6, 15, 24, 26, 27, 39, 41, 50, 55, 60, 62-69, 72, 73). The conformational change results in the formation and/or exposure of the chemokine receptor sites (62, 64), thus promoting further viral attachment and membrane fusion.
The molecular reorganization that results upon binding of CD4 is revealed by the structure of an unliganded simian immunodeficiency virus (SIV) gp120 core (6). With a few important exceptions, the structure of the outer domain is quite similar to that seen in the CD4-bound state. In contrast, the structure of the inner domain is markedly different. A comparison of CD4-bound and unliganded gp120 shows that the conformational change is not a simple movement of the inner domain as a rigid body. Rather, the inner domain is comprised of a set of distinct substructures that move relatively independently of one another (6). The binding of CD4 results in a rearrangement of these secondary structural elements within the inner domain (6). As predicted (24), the bridging sheet is not present in the structure of unliganded gp120 (6).
The structure of antibody b12 was determined by Saphire et al. (52-54). The structure revealed an extended CDR H3 loop with an apical tryptophan residue that is thought to recognize the Phe43 binding pocket of gp120 (53, 75). This putative interaction places significant constraints on possible gp120/b12 interactions, allowing an interaction between b12 and the CD4-bound conformation of gp120 to be modeled (53, 75). This model suggests that the broad neutralizing activity of b12 lies in its ability to interact with conserved features of the CD4 binding site through recognition of the Phe43 pocket and interactions with main chain atoms of gp120 and in its ability to recognize trimeric gp120 on the native viral surface.
A further understanding of the molecular properties that confer broad reactivity and potency upon a CD4 binding-site antibody are of substantial interest. The underlying principles are likely to impact the design of new immunogens with the potential to elicit a successful immune response. In this regard, the structure of F105 provides an opportunity to compare and contrast the structure of a broadly reactive but nonpotent CD4 binding-site antibody (F105) with that of a broadly neutralizing antibody with an overlapping epitope (b12). The comparison provides significant insight into the molecular properties that impart broad reactivity and potency upon a CD4 binding-site antibody.
MATERIALS AND METHODS
Expression and purification. The expression of F105 was performed as previously described (49). F105 was purified over protein G and eluted with 0.1 M glycine-HCl, pH 2.7, which was immediately neutralized upon collection with 1 M Tris, pH 9. Previous work had shown some contamination of protein G-purified F105 with bovine Ig from the fetal calf serum used for growing the hybridoma cells (19). Therefore, we further purified F105 using Immunopure immobilized protein L (Pierce), eluted it with 0.1 M glycine-HCl, pH 2.5, and neutralized the eluant with 1 M Na2HPO4, pH 7.2. The elution fractions containing F105 were combined and concentrated (Amicon Centricon filter with 30K molecular weight cutoff) to 4.5 mg/ml.
Fab fragments of the purified antibody were prepared by digestion with papain (0.4 μM) (ICN Biomedicals) in the presence of 20 mM ?-mercaptoethanol. The reaction was stopped by dialyzing against phosphate-buffered saline to remove the ?-mercaptoethanol, and the completeness of the digestion was verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis.
Fab fragments were purified by anion-exchange chromatography using a POROS HQ10 column on a BioCAD Sprint perfusion chromatography system (Perseptive BioSystems, Inc.) using a linear gradient from 0 to 1.0 M NaCl in 10 mM Tris, pH 8.1. Fractions containing the F105 Fab fragment were concentrated to 6.0 mg/ml in an Amicon Ultra 10K-molecular-weight-cutoff centrifugal filter. Protein concentrations were determined by bicinchoninic acid analysis using bovine serum albumin as a standard (56). The overall yield was approximately 2.1 mg of purified Fab from 10 mg of protein G-purified antibody.
Crystallization and data collection. Purified Fab from F105 was crystallized by hanging-drop vapor diffusion at 18°C. Drops were assembled with 2 μl of F105 Fab fragment mixed with 2 μl of well solution containing 15.5% polyethylene glycol 4000, 25% isopropanol, 100 mM NaCl, and 75 mM sodium citrate, pH 5.6, with or without 1 μl of 600 μM peptide 44 (19). Crystals typically appeared in 1 to 7 weeks. Single crystals were frozen in liquid nitrogen. Data were collected at 100 K using a Rigaku RUH3R rotating anode, a CuK X-ray source, and a MAR345 image plate detector. F105 Fab fragment crystals diffracted to a 2.8-? resolution and belonged to space group P43212, with a = b = 120.4 ? and c = 72.98 ?, with a single Fab per asymmetric unit. Data were integrated and reduced using the HKL software package (42). Statistics on data completeness and quality are presented in Table 1.
Structure determination and refinement. The structure of the F105 Fab was determined by molecular replacement. Due to large variations in the elbow angles of crystallized Fabs, the molecular replacement procedure typically breaks the chosen Fab search model into constant and variable domains, and each domain is independently positioned. However, in this case, we chose to use the family search option of COMO and a library of 195 Fab search models. This allowed us to utilize entire Fabs as search models, and due to the large number of models, we were able to effectively search the available "elbow angle space" to identify the correct solution. The library was composed of 195 human, mouse, and rat IgG Fab structures taken from the protein data bank. Each downloaded structure file was processed with a series of script files to generate a single Fab search model free of hetero compounds. Upon completion of the library, the automated search process was rapid and successful, placing a humanized anti-CD18 Fab (PDB entry 2FGW) (13) into the F105 asymmetric unit to yield a correlation coefficient of 43.5% and an initial R factor of 40.2%.
Subsequently, nonconserved residues in the positioned 2FGW model were mutated to alanine or glycine, as appropriate, and the variable and constant domains were subjected to independent rigid body refinement with CNS, followed by positional and independent B-factor refinement. Sigma A-weighted 2mFo-dFc and mFo-dFc maps (51) then showed clearly interpretable density for much of the missing model, including very nice density for the large CDR H3 loop present in F105. It is noteworthy that the CDR H3 loop was disordered in 2FGW and therefore was not present in the search model. All data between 15.0 and 2.8 ? were used for the refinement, with 5% of the data randomly chosen for Rfree calculations. Rounds of iterative model building with O (20) and refinement with CNS resulted in a final model with an R factor of 21.0% (Rfree, 25.1%). The statistics on model quality are summarized in Table 1. The model exhibits good geometry, with 88.8% of residues in the most favored regions of the Ramachandran plot, 11.5% in additionally allowed regions, and one residue in a generously allowed region (29). One residue, AlaL51, is found in a disallowed region. It is located in a highly conserved turn and contains geometry typical of CDR L2 structures seen in many other antibodies (53). The figures were generated with PYMOL (http://www.pymol.org).
Docking. The F105 Fab structure was manually docked to core gp120 from the structure of the gp120-CD4-Fab17b complex (PDB entry 1GM9) in an orientation that places the antigen binding site of the Fab in the CD4 binding pocket of gp120, minimizes bad contacts, and maximizes overlap of the antigen-combining site with residues implicated in binding of F105, as described in Results and Discussion. b12 docking was manually recreated using information provided by Saphire et al. (53) and Zwick et al. (75). Specifically, b12 was positioned such that Trp100 of the H3 loop was inserted into the hydrophobic pocket of gp120, Tyr53 of the H2 loop was fit into the gap between Thr373 and Asn386 of gp120, and Arg29 and Arg31 of the L1 loop were in close proximity to Asn276/Lys282 and Asn280/Ala281 of the D loop of gp120, respectively (75). The relative position of hand-docked b12 to gp120 was consistent with the proposed contacts described by Saphire et al. (53), Zwick et al. (75), and Pantophlet et al. (45). Following the addition of hydrogen atoms and capping of the polypeptide chains, both the F105/gp120 and b12/gp120 complexes were subjected to energy minimization using the programs InsightII and Discover (Accelrys, San Diego, Calif.). An initial 100 iterations were carried out using the steepest descent algorithm to minimize bad contacts, followed by 1,000 iterations using conjugate gradient minimization. In the minimized F105/gp120 complex, PheH100A of F105 retains an interaction with residues forming the Phe43 pocket of gp120 and ArgL31 forms a salt bridge to gp120 residues Asp474 and Asp477.
Amino acid content of human CDR loops. The amino acid sequences of CDR loops in complete human antibodies present in the Kabat sequence database were downloaded using Kabatman software (35). The frequency of occurrence of each amino acid, as well as that of serine/tyrosine pairs, was determined using a small Fortran program written for this purpose.
Protein structure accession number. Atomic coordinates and structure factors have been deposited in the Protein Data Bank under accession code 1U6A.
RESULTS AND DISCUSSION
Structure of F105 Fab fragment. The F105 model includes light chain (chain L) residues 1 through 213 and heavy chain (chain H) residues 2 through 213. There are no breaks in the main chain electron density, and the side chain density for most residues, including those in the CDR loops, is also quite good. The structure of the F105 Fab fragment reveals several notable features. The most obvious of these is a relatively long CDR H3 loop (Fig. 1). This 15-residue loop is well ordered and extends approximately 7 ? beyond the other CDR loops, with residues PheH100A and TyrH100B (Kabat numbering) at the apex. This conformation of the H3 loop is stabilized by a hydrogen bond between the main chain carbonyl of GlyH100C and the side chain of ArgL31A, which is also prominent at the top of the antigen recognition site. Further stabilization of the H3 loop may arise from its three proline residues (Fig. 1A). The increased occurrence of proline in CDR H3 loops has been suggested to confer stability to longer CDR H3 regions (71).
The PheH100A and TyrH100B side chains are involved in lattice contacts within the crystal. One face of the PheH100A side chain packs against the main chain atoms of ProH136, SerH137, and SerH138 in a neighboring molecule, while the tyrosine side chain contacts the C atom of GlyH200 in an adjacent F105 molecule. Overall, the lattice contacts do not appear to provide substantial stabilization of the ordered H3 loop. Thus, it is likely that an ordered H3 loop is intrinsic to F105, rather than a result of lattice contacts within the crystal.
The F105 structure also reveals a striking constellation of nine tyrosine residues aligned across the top of the antigen recognition site (Fig. 1). This includes several tyrosine residues nestled around the base of the H3 loop that might provide steric constraints and favorable interactions leading to reduced conformational flexibility of the H3 loop. For example, ValH97 packs tightly against the ring of TyrH33. The extended H3 loop, the nine tyrosine residues, and ArgL31 appear to be principle features of the F105 paratope.
However, the CDR sequences are also rich in serine (Fig. 1). We found that 14 of 64 residues present in the CDRs are serine. Unlike the tyrosine residues, which are distributed across the top of the antigen binding site, the serine residues are concentrated along the periphery (Fig. 1C). Combined, the serine and tyrosine residues account for 23 of 64 CDR residues (36%).
Structural comparison of F105 and b12. An extended H3 loop is also seen in antibody b12, with a tryptophan at its apex. It is thought to protrude into the recessed CD4 binding site of gp120, with the tryptophan occupying the same pocket that accommodates Phe43 of CD4 (53, 75). However, the H3 loop in b12 is slightly longer than that in F105, as it is 18 residues in length versus 15 residues in F105. Two views of a superposition of F105 and b12 are shown in Fig. 1D and E. This superposition suggests that F105 PheH100A might also be able to reach the Phe43 binding pocket of gp120 in a manner analogous to that of Trp100 in the b12 antibody and Phe43 of CD4.
However, this does not necessarily indicate that the extended H3 loop of F105, with its apical phenylalanine, recognizes the Phe43 pocket of gp120. In contrast to the case for murine antibodies, extended H3 loops are apparently more common in human antibodies that recognize virus (8, 71). Collis et al. have found an average length of 16.5 residues in antiviral human H3 loops (8), and many non-CD4 binding-site antibodies capable of neutralizing HIV show extended loops (9, 17, 26, 27, 57, 74). Mutational analyses and antibody competition studies suggest that CD4 and CD4 binding-site antibodies recognize different conformations of gp120 (50, 69, 73). In addition, the binding of b12, F105, and CD4 to core gp120 results in entropic changes of 5.7 kcal/mol, 18.9 kcal/mol, and 35 kcal/mol, respectively (24, 41). Together, these data suggest that the F105-bound conformation of gp120 is significantly different from the unliganded or CD4-bound conformation of gp120.
The entropic argument is also relevant to b12, where an entropic difference of 30 kcal/mol is seen for binding of b12 versus CD4 (24). Indeed, reliance on entropic considerations alone would suggest that b12 should be docked to the unliganded conformation of gp120 rather than the CD4-bound conformation. However, the model for b12 docked to the CD4-bound conformation of gp120 is consistent with subsequent mutational and structural studies and appears to be substantially correct, whereas b12 is unlikely to bind the unliganded gp120 structure (6, 45, 53, 75). One explanation for this is that elements of the b12 epitope are present within multiple conformations of gp120, including the CD4-bound conformation. Thus, while binding of b12 does not induce the CD4-bound conformation, the CD4-bound conformation of gp120 does represent a starting point for modeling the interaction between b12 and gp120.
Thus, it is not clear whether the F105-bound conformation of gp120 bears the greatest similarity to the CD4-bound state, the unliganded state, an intermediate conformation, or a conformation that is altogether different. We have addressed this question by mapping gp120 residues critical for binding of F105 to the CD4-bound and unliganded structures of gp120 (Fig. 2) (45, 59, 61). For reference, we also mapped 26 residues (26, 27) known to make direct contact with CD4. In the unliganded structure, only a few residues implicated in interactions with F105 are apparent on the surface of gp120; many of the critical residues are occluded by residues N-terminal to helix 1 and by helix 1 itself (Fig. 2B and D).
In the CD4-bound state, however, many of the critical residues are solvent exposed where they are found along the lip of the Phe43 binding pocket, ?-strand 15 (CD4 binding loop), the N-terminal end of helix 5, and its preceding loop. In contrast to the unliganded state of gp120, in the CD4-bound state these residues form a nearly contiguous solvent-exposed surface, one that intersects the surface involved in the recognition of CD4 (26, 27, 59, 61, 65). Figure 2C illustrates the recessed nature of these residues, many of which lie at or near the interface of the inner and outer domains of gp120 (Fig. 2A), strongly suggesting that residues forming the rim of the Phe43 binding pocket are recognized by F105. Similarly, mutations that alter the binding of antibody b12 also map to a contiguous surface on CD4-bound gp120 (45, 75) but are occluded or more dispersed across the surface of the unliganded structure (6).
Further examination of the ligand-free and CD4-bound conformations suggests that F105 is likely to recognize a conformation in which the transition of the inner domain to the CD4-bound conformation is nearing completion rather than an inner domain conformation that is intermediate between the two states. As gp120 transits from the ligand-free conformation to the CD4-bound conformation, the center of mass of helix 1 moves 15 ? away from the outer domain (Fig. 2B) (6). An incomplete inner domain transition, at least with respect to 1, would presumably occlude critical residues in helix 5, restricting antibody access to residues on the floor of the CD4 binding site. Furthermore, Chen and coworkers pointed out that 1 is an amphipathic helix and that an intermediate position of 1 is likely to bury a set of charged residues. They suggested that intermediate conformations of 1 would therefore be unstable (6).
Thus, surface mapping of residues critical for F105 binding to the ligand-free and CD4-bound conformations suggests that the F105-bound conformation of gp120 will have an inner domain structure approaching that seen in the CD4-bound state. This does not imply that the interface between the inner and outer domains is identical to that in the CD4-bound state; conformational differences are likely to exist, perhaps resulting in the exposure of additional residues critical to the recognition of F105 and other CD4 binding-site antibodies (65). Furthermore, while the inner domain transition appears largely complete, the bridging sheet may not be formed (6), consistent with the calorimetric data of Kwong et al. (24). In this respect, it is likely that the F105-bound state is intermediate between the unliganded and CD4-bound conformations of gp120.
This also suggests that the inner and outer domains of CD4-bound gp120 are the preferred starting point to examine possible interactions between F105 and gp120. For example, by docking F105 to the CD4-bound conformation of gp120, we can test whether the extended H3 loop and other elements of the F105 paratope are able to access critical residues on the floor of the CD4 binding site. If so, we can ask whether the interacting surfaces appear to be complementary with respect to shape, charge, hydrophilicity, and hydrophobicity and how these interactions compare to those suggested for b12 (45, 53, 75).
As the starting point for the docking exercise, we considered the possibility that the apical PheH100A of F105 functions in a manner analogous to that of Phe43 of CD4, i.e., that PheH100A binds within the Phe43 pocket of gp120. Thus, superposition of the phenylalanine side chains of F105 and CD4 provided an initial docking of F105 to gp120 from strain HXBc2, an interaction that is known to occur (7, 24, 25). This was followed by rotation of F105 in order to minimize bad contacts, to maximize surface complementarity, and to maximize interactions with gp120 residues critical to the binding of F105. Following energy minimization of the F105/gp120 model, we found that PheH100A of F105 remains in the Phe43 binding pocket of gp120. This demonstrates that even though the H3 loop of F105 is three residues shorter than that of b12, it is long enough to access the Phe43 pocket of gp120.
Importantly, the minimized structure also suggests logical roles for other elements in the F105 paratope. Many of the tyrosine and serine side chains demonstrate the ability to form hydrogen bonds with gp120, including numerous interactions with exposed main chain atoms in the CD4 binding site, while ArgL31 can form salt bridges to acidic residues in helix 5 of gp120. Thus, the docking exercise suggests that elements of the F105 epitope might indeed reside within the CD4-bound conformation of gp120.
As expected, there are similarities and differences between the footprint of docked F105 on gp120 and that predicted for b12 (45, 46, 53, 75). For both F105 and b12, the heavy chains are predicted to straddle a ridge on the surface of gp120 formed by ?-strand 15. The H3 loops project down one side of the ridge towards the Phe43 pocket of gp120, while residues in the H2 loops interact with the opposite side. In the case of F105, however, the cleft between the H2 and H3 loops is significantly less pronounced than that seen in b12, and thus the extension of the H2 loop down the opposite side of the ?15 ridge is reduced. ?-Strand 15 is one outer domain element that does undergo significant rearrangement upon binding of CD4, and it is referred to as the CD4 binding loop by Chen et al. (6). Surface mapping suggests that the CD4-bound conformation of this loop, rather than the ligand-free conformation, is more likely to be recognized by F105. However, alternate conformations of the CD4 binding loop that expose additional residues implicated in the F105/gp120 interaction could also be considered.
Distinct differences between the putative surfaces recognized by the b12 and F105 light chains are also apparent. In the case of b12, interactions between the light chain CDRs and outer domain residues, including residues within the D loop (Fig. 2), are thought to be critical (45, 75). In contrast, the model for F105 suggests that in addition to interactions with outer domain residues, the light chain also recognizes inner domain residues within helix 5. Mutations to gp120 residues 473 through 476 (GDMR) at the N-terminal end of helix 5 inhibit the recognition of gp120 by F105 but do not inhibit the binding of antibody b12 (45). This indicates an important difference between the F105 and b12 epitopes. In this regard, the docking model for F105 is consistent with mutational data indicating the importance of these residues (45, 59).
Neutralizing ability of b12 versus that of F105. The broad neutralizing ability of b12 has been attributed to several properties (53). These include interactions with conserved elements of the CD4 binding site, many of which are mediated through contacts with main chain atoms of gp120, and the interaction of the extended H3 loop with the rim of the Phe43 binding pocket. In addition, b12 is thought to bind in a manner that allows it to access the epitope not only on monomeric gp120, but on the native viral surface as well (45, 53). In the case of F105, the docking exercise makes specific suggestions regarding the mechanisms that provide for broad recognition of the CD4 binding site by F105. These include an extended H3 loop that interacts with conserved residues surrounding the Phe43 binding pocket and the use of tyrosine and serine residues of F105 to hydrogen bond to exposed main chain atoms of the CD4 binding site, properties thought to contribute to the unique potency of b12. When the exercise is extended to the trimeric gp120 model of Kwong et al. (28), it suggests that F105 is also capable of accessing the epitope on the native viral surface, a property that F105 is known to possess (48, 49). Thus, many of the properties suggested to contribute to the potency of b12 are also seemingly present in F105. In this regard, these shared properties provide an explanation for the broad reactivities of both F105 and b12. It is less clear, however, why they confer such potency to b12, while F105 is relatively nonpotent.
Perhaps it is a matter of degree. In addition to a possible interaction with the Phe43 pocket of gp120, the longer H3 loop of b12 results in a lateral extension that may be more effective in its interactions with the protruding ridge formed by the CD4 binding loop of gp120. The deeper cleft between the H2 and H3 loops might also result in stronger interactions between b12 and gp120. However, we believe that the proposed interaction between the light chain of F105 and the inner domain of gp120 is another obvious difference between these two antibodies.
The binding of F105 across the domain interface would be expected to result in a significant ordering of residues at the inner domain/outer domain junction. It is also likely to lead to reduced group motions for other inner domain substructures, consistent with the large decrease in entropy that is observed (6, 24). In contrast to the case for F105, the reduced entropic component (24) and the modeled b12 interactions (45, 53, 75) both suggest that b12 largely avoids the domain interface, restricting primary recognition to the outer domain of gp120. In this regard, the combination of the deeper H2/H3 cleft and the lateral component of the extended H3 loop may be critical in allowing b12 to recognize an epitope centered on the outer domain (Fig. 1E and 2). The deeper cleft between the H2 and H3 loops potentially allows b12 to "tip back," away from the inner domain/outer domain junction (Fig. 2E and F), while maintaining the interaction between the tip of the H3 loop and the Phe43 pocket of gp120. Furthermore, the relative orientations of the H2/H3 cleft and the lateral extension of the H3 loop may position the b12 light chain to contact the D loop of gp120. The net result is movement of the b12 light chain away from the interdomain junction.
We would like to emphasize that the CD4-bound conformation of gp120 certainly differs from the F105-bound conformation. In particular, the bridging sheet is unformed in the unliganded state of monomeric gp120 and may also be absent in the F105-bound state. In addition, residues in the CD4 binding loop (?15) or those at the inner domain/outer domain junction might differ significantly from the CD4-bound conformation (65). Indeed, several residues near the interdomain junction that affect the binding of F105 are not surface exposed, even in the CD4-bound conformation (59, 61, 65). Thus, the docking of F105 to gp120 should be considered only a low-resolution model for the interaction of F105 with gp120, but one that is consistent with known constraints from previous mutational, structural, and calorimetric analyses. Importantly, it suggests that the greater potency of b12 can be attributed to an epitope that lies primarily on the outer domain of gp120, allowing b12 to avoid the conformational or entropic masking that is apparently so effective at protecting the HIV virus from most other CD4 binding-site antibodies, including F105 (24).
Maturation and amino acid content of the CDR loops. Another notable difference between F105 and b12 is the number of mutations from the germ line V and J region sequences. Huang et al. calculated the number of such mutations in 25 human gp120-reactive antibodies, finding an average of 22 (17). F105 has among the fewest, at 13, whereas b12 has one of the highest, at 45 (17), many of which are thought to be important for its ability to bind gp120 (75). Thus, b12 represents an antibody that has undergone extensive affinity maturation by somatic mutation, while F105 is more representative of the germ line immune repertoire. Extensive mutation of b12 in response to a changing gp120 protein in the patient may have increased its potency while maintaining broad specificity. It is notable that 2G12, one of the few antibodies to gp120 that is potent and broadly neutralizing like b12, has 51 V and J somatic mutations (17). Furthermore, studies of broadly neutralizing human anti-HIV antibodies directed against gp41 epitopes have also noted that their potency is correlated with long CDR H3 loops and extensive affinity maturation by somatic mutation (23).
The frequent occurrence of serine and tyrosine residues within CDRs has been noted before and is in fact a general phenomenon (8, 10, 14, 18, 21, 22, 30, 33, 37, 43, 44, 71). For example, Mian et al. (37) reported a mean composition for vertebrate CDRs that shows an average serine content of 17.27%, a tyrosine content of 11.08%, and thus a combined serine/tyrosine content of 28%. In our own analysis, we found the combined serine/tyrosine content in CDRs of human antibodies to be 27.0% ± 6.6%. Relative to these values, the combined serine and tyrosine content of 36% for F105 is significantly above average. The heavy use of serine and tyrosine residues in F105 raises several questions deserving of further comment.
(i) Why are serine and tyrosine residues favorable to recognition of the CD4 binding site? The docking exercise suggests that the numerous tyrosine and serine residues allow F105 to maximize hydrogen bonding to the exposed main chain atoms of the CD4 binding site, thus helping to confer broad reactivity to F105. These residues are capable of acting as both hydrogen bond donors and acceptors (36) and can thus hydrogen bond to both carbonyl oxygen atoms and amide protons of the polypeptide backbone. Several additional properties of tyrosine are thought to contribute to its frequent use in antigen recognition (14, 18, 37). Together, these properties make tyrosine particularly well suited for the recognition of variable epitopes.
The relative positions of the serine and tyrosine residues within the F105 CDRs are illustrative of those in antibodies in general (8, 37). For example, Mian et al. (37) have noted an increased serine and tyrosine content for residues in known binding positions. They also found that while the tyrosine content was greater at known binding positions than in the remainder of the CDR, the greatest enrichment of serine was in regions of the CDR loops distal to the known binding sites.
(ii) What is the purpose of the serine residues along the periphery of the antigen combining site? Mian et al. suggested that these residues play a structural role (37). However, the frequency with which serine is found in the CDRs outside of the contact residues is much greater than that generally seen in loops or ?-strands (8), suggesting that serine residues in this region of the CDR are likely to serve a functional role as well. The use of the small polar serine side chain along the periphery of the antigen recognition site should maximize access to a recessed antigen binding site. Furthermore, in the case of incidental contact between these residues and the antigen, the chances of a favorable contact are also maximized, while the entropic penalty for ordering of the side chain is minimal. Since the CD4 binding site is recessed and the opening to this site is quite narrow, at least in one dimension, the serine residues along the periphery of F105 might play a role in promoting the interaction of F105 with gp120.
(iii) Has maturation of F105 resulted in an increased occurrence of serine and tyrosine residues in the CDRs, or are the germ line genes naturally rich in serine and tyrosine? The answer appears to be the latter. The V and J gene precursors for the F105 heavy chain are VH4-59 and JH4, respectively, while the light chain precursors are A27a and JK2. There are 10 mutations that fall within the Kabat CDRs, resulting in a net loss of one tyrosine and two serine residues.
(iv) Are other CD4 binding-site antibodies rich in serine and tyrosine? We found an even greater serine/tyrosine content in the CD4 binding-site antibody 15e (2, 16, 38), where fully 40% of the CDR residues are serine or tyrosine. In contrast, the serine/tyrosine content of antibody b12 is only about average (25%). However, the germ line precursors for antibody b12 do exhibit an elevated serine/tyrosine content (40%). Thus, the decreased occurrence reflects the extensive affinity maturation that has taken place in b12 relative to F105 and 15e.
In summary, the structure of F105 suggests a common molecular basis for the broad recognition of HIV-1 by F105 and b12. It reaffirms the need for immunogens capable of eliciting antibodies with extended H3 loops and suggests that the unique potency of antibody b12 resides in its ability to avoid the interface between the inner and outer domains of gp120. Thus, approaches to generating a b12-like immune response might consider ways to elicit extended H3 loops while minimizing interactions with the inner domain residues of gp120. The subtle differences between the H3 loops of b12 and F105 speak to the difficulty of doing so in a vaccine setting.
ACKNOWLEDGMENTS
This work was supported by grants AI49753 (M.T.) and AI26926 (M.R.P.) from the National Institutes of Health.
We sincerely thank all reviewers for their helpful suggestions.
REFERENCES
Baba, T. W., V. Liska, R. Hofmann-Lehmann, J. Vlasak, W. Xu, S. Ayehunie, L. A. Cavacini, M. R. Posner, H. Katinger, G. Stiegler, B. J. Bernacky, T. A. Rizvi, R. Schmidt, L. R. Hill, M. E. Keeling, Y. Lu, J. E. Wright, T. C. Chou, and R. M. Ruprecht. 2000. Human neutralizing monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human immunodeficiency virus infection. Nat. Med. 6:200-206.
Bagley, J., P. J. Dillon, C. Rosen, J. Robinson, J. Sodroski, and W. A. Marasco. 1994. Structural characterization of broadly neutralizing human monoclonal antibodies against the CD4 binding site of HIV-1 gp120. Mol. Immunol. 31:1149-1160.
Burton, D. R., R. C. Desrosiers, R. W. Doms, W. C. Koff, P. D. Kwong, J. P. Moore, G. J. Nabel, J. Sodroski, I. A. Wilson, and R. T. Wyatt. 2004. HIV vaccine design and the neutralizing antibody problem. Nat. Immunol. 5:233-236.
Cao, J., N. Sullivan, E. Desjardin, C. Parolin, J. Robinson, R. Wyatt, and J. Sodroski. 1997. Replication and neutralization of human immunodeficiency virus type 1 lacking the V1 and V2 variable loops of the gp120 envelope glycoprotein. J. Virol. 71:9808-9812.
Cavacini, L. A., C. L. Emes, A. V. Wisnewski, J. Power, G. Lewis, D. Montefiori, and M. R. Posner. 1998. Functional and molecular characterization of human monoclonal antibody reactive with the immunodominant region of HIV type 1 glycoprotein 41. AIDS Res. Hum. Retrovir. 14:1271-1280.
Chen, B., E. M. Vogan, H. Gong, J. J. Skehel, D. C. Wiley, and S. C. Harrison. 2005. Structure of an unliganded simian immunodeficiency virus gp120 core. Nature 433:834-841.
Choe, H., W. Li, P. L. Wright, N. Vasilieva, M. Venturi, C. C. Huang, C. Grundner, T. Dorfman, M. B. Zwick, L. Wang, E. S. Rosenberg, P. D. Kwong, D. R. Burton, J. E. Robinson, J. G. Sodroski, and M. Farzan. 2003. Tyrosine sulfation of human antibodies contributes to recognition of the CCR5 binding region of HIV-1 gp120. Cell 114:161-170.
Collis, A. V., A. P. Brouwer, and A. C. Martin. 2003. Analysis of the antigen combining site: correlations between length and sequence composition of the hypervariable loops and the nature of the antigen. J. Mol. Biol. 325:337-354.
Darbha, R., S. Phogat, A. F. Labrijn, Y. Shu, Y. Gu, M. Andrykovitch, M. Y. Zhang, R. Pantophlet, L. Martin, C. Vita, D. R. Burton, D. S. Dimitrov, and X. Ji. 2004. Crystal structure of the broadly cross-reactive HIV-1-neutralizing Fab X5 and fine mapping of its epitope. Biochemistry 43:1410-1417.
Davies, D. R., and G. H. Cohen. 1996. Interactions of protein antigens with antibodies. Proc. Natl. Acad. Sci. USA 93:7-12.
Ditzel, H. J., J. M. Binley, J. P. Moore, J. Sodroski, N. Sullivan, L. S. Sawyer, R. M. Hendry, W. P. Yang, C. F. Barbas III, and D. R. Burton. 1995. Neutralizing recombinant human antibodies to a conformational V2- and CD4-binding site-sensitive epitope of HIV-1 gp120 isolated by using an epitope-masking procedure. J. Immunol. 154:893-906.
D'Souza, M. P., D. Livnat, J. A. Bradac, and S. H. Bridges. 1997. Evaluation of monoclonal antibodies to human immunodeficiency virus type 1 primary isolates by neutralization assays: performance criteria for selecting candidate antibodies for clinical trials. AIDS Clinical Trials Group Antibody Selection Working Group. J. Infect. Dis. 175:1056-1062.
Eigenbrot, C., T. Gonzalez, J. Mayeda, P. Carter, W. Werther, T. Hotaling, J. Fox, and J. Kessler. 1994. X-ray structures of fragments from binding and nonbinding versions of a humanized anti-CD18 antibody: structural indications of the key role of VH residues 59 to 65. Proteins 18:49-62.
Fellouse, F. A., C. Wiesmann, and S. S. Sidhu. 2004. Synthetic antibodies from a four-amino-acid code: a dominant role for tyrosine in antigen recognition. Proc. Natl. Acad. Sci. USA 101:12467-12472.
Hart, T. K., R. Kirsh, H. Ellens, R. W. Sweet, D. M. Lambert, S. R. Petteway, Jr., J. Leary, and P. J. Bugelski. 1991. Binding of soluble CD4 proteins to human immunodeficiency virus type 1 and infected cells induces release of envelope glycoprotein gp120. Proc. Natl. Acad. Sci. USA 88:2189-2193.
Ho, D. D., J. A. McKeating, X. L. Li, T. Moudgil, E. S. Daar, N. C. Sun, and J. E. Robinson. 1991. Conformational epitope on gp120 important in CD4 binding and human immunodeficiency virus type 1 neutralization identified by a human monoclonal antibody. J. Virol. 65:489-493.
Huang, C. C., M. Venturi, S. Majeed, M. J. Moore, S. Phogat, M. Y. Zhang, D. S. Dimitrov, W. A. Hendrickson, J. Robinson, J. Sodroski, R. Wyatt, H. Choe, M. Farzan, and P. D. Kwong. 2004. Structural basis of tyrosine sulfation and VH-gene usage in antibodies that recognize the HIV type 1 coreceptor-binding site on gp120. Proc. Natl. Acad. Sci. USA 101:2706-2711.
Ivanov, I., J. Link, G. C. Ippolito, and H. W. Schroeder. 2002. Constraints on the hydropathicity and sequence composition of HCDR3 are conserved across evolution, p. 43-67. In M. Zanetti and J. Capra (ed.), The antibodies. Taylor and Francis, London, United Kingdom.
Jacobs, J. M., B. W. Bailey, J. B. Burritt, S. G. Morrison, R. P. Morrison, E. A. Dratz, A. J. Jesaitis, and M. Teintze. 2003. QSYP peptide sequence is selected from phage display libraries by bovine IgG contaminants in monoclonal antibody preparations. BioTechniques 34:132-134, 137-141.
Jones, T. A., J. Y. Zou, S. W. Cowan, and M. Kjeldgaard. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47:110-119.
Kabat, E. A., T. T. Wu, and H. Bilofsky. 1977. Unusual distributions of amino acids in complementarity-determining (hypervariable) segments of heavy and light chains of immunoglobulins and their possible roles in specificity of antibody-combining sites. J. Biol. Chem. 252:6609-6616.
Kabat, E. A., T. T. Wu, M. Reid-Miller, H. M. Perry, and K. S. Gottesman. 1987. Sequences of proteins of immunological interest. U.S. Department of Health and Human Services, Washington, D.C.
Kunert, R., S. Wolbank, G. Stiegler, R. Weik, and H. Katinger. 2004. Characterization of molecular features, antigen-binding, and in vitro properties of IgG and IgM variants of 4E10, an anti-HIV type 1 neutralizing monoclonal antibody. AIDS Res. Hum. Retrovir. 20:755-762.
Kwong, P. D., M. L. Doyle, D. J. Casper, C. Cicala, S. A. Leavitt, S. Majeed, T. D. Steenbeke, M. Venturi, I. Chaiken, M. Fung, H. Katinger, P. W. Parren, J. Robinson, D. Van Ryk, L. Wang, D. R. Burton, E. Freire, R. Wyatt, J. Sodroski, W. A. Hendrickson, and J. Arthos. 2002. HIV-1 evades antibody-mediated neutralization through conformational masking of receptor-binding sites. Nature 420:678-682.
Kwong, P. D., R. Wyatt, E. Desjardins, J. Robinson, J. S. Culp, B. D. Hellmig, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 1999. Probability analysis of variational crystallization and its application to gp120, the exterior envelope glycoprotein of type 1 human immunodeficiency virus (HIV-1). J. Biol. Chem. 274:4115-4123.
Kwong, P. D., R. Wyatt, S. Majeed, J. Robinson, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 2000. Structures of HIV-1 gp120 envelope glycoproteins from laboratory-adapted and primary isolates. Struct. Fold Des. 8:1329-1339.
Kwong, P. D., R. Wyatt, J. Robinson, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 1998. Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature 393:648-659.
Kwong, P. D., R. Wyatt, Q. J. Sattentau, J. Sodroski, and W. A. Hendrickson. 2000. Oligomeric modeling and electrostatic analysis of the gp120 envelope glycoprotein of human immunodeficiency virus. J. Virol. 74:1961-1972.
Laskowski, R. A., M. W. MacArthur, D. S. Moss, and J. M. Thornton. 1993. PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26:283-291.
Lea, S., and D. Stuart. 1995. Analysis of antigenic surfaces of proteins. FASEB J. 9:87-93.
Li, A., H. Katinger, M. R. Posner, L. Cavacini, S. Zolla-Pazner, M. K. Gorny, J. Sodroski, T. C. Chou, T. W. Baba, and R. M. Ruprecht. 1998. Synergistic neutralization of simian-human immunodeficiency virus SHIV-Vpu + by triple and quadruple combinations of human monoclonal antibodies and high-titer anti-human immunodeficiency virus type 1 immunoglobulins. J. Virol. 72:3235-3240.
Litwin, V., K. A. Nagashima, A. M. Ryder, C. H. Chang, J. M. Carver, W. C. Olson, M. Alizon, K. W. Hasel, P. J. Maddon, and G. P. Allaway. 1996. Human immunodeficiency virus type 1 membrane fusion mediated by a laboratory-adapted strain and a primary isolate analyzed by resonance energy transfer. J. Virol. 70:6437-6441.
Lo Conte, L., C. Chothia, and J. Janin. 1999. The atomic structure of protein-protein recognition sites. J. Mol. Biol. 285:2177-2198.
MacCallum, R. M., A. C. Martin, and J. M. Thornton. 1996. Antibody-antigen interactions: contact analysis and binding site topography. J. Mol. Biol. 262:732-745.
Martin, A. C. 1996. Accessing the Kabat antibody sequence database by computer. Proteins 25:130-133.
McDonald, I., and J. M. Thornton. Atlas of side-chain and main-chain hydrogen bonding. [Online.] http://www.biochem.ucl.ac.uk/bsm/atlas/.
Mian, I. S., A. R. Bradwell, and A. J. Olson. 1991. Structure, function and properties of antibody binding sites. J. Mol. Biol. 217:133-151.
Moore, J. P., F. E. McCutchan, S. W. Poon, J. Mascola, J. Liu, Y. Cao, and D. D. Ho. 1994. Exploration of antigenic variation in gp120 from clades A through F of human immunodeficiency virus type 1 by using monoclonal antibodies. J. Virol. 68:8350-8364.
Moore, J. P., J. A. McKeating, R. A. Weiss, and Q. J. Sattentau. 1990. Dissociation of gp120 from HIV-1 virions induced by soluble CD4. Science 250:1139-1142.
Moran, M. J., J. S. Andris, Y. Matsumato, J. D. Capra, and E. M. Hersh. 1993. Variable region genes of anti-HIV human monoclonal antibodies: non-restricted use of the V gene repertoire and extensive somatic mutation. Mol. Immunol. 30:1543-1551.
Myszka, D. G., R. W. Sweet, P. Hensley, M. Brigham-Burke, P. D. Kwong, W. A. Hendrickson, R. Wyatt, J. Sodroski, and M. L. Doyle. 2000. Energetics of the HIV gp120-CD4 binding reaction. Proc. Natl. Acad. Sci. USA 97:9026-9031.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode, p. 307-326. In C. Carter and R. Sweet (ed.), Macromolecular crystallography, part A, vol. 276. Academic Press, San Diego, Calif.
Padlan, E. A. 1994. Anatomy of the antibody molecule. Mol. Immunol. 31:169-217.
Padlan, E. A. 1990. On the nature of antibody combining sites: unusual structural features that may confer on these sites an enhanced capacity for binding ligands. Proteins 7:112-124.
Pantophlet, R., E. Ollmann Saphire, P. Poignard, P. W. Parren, I. A. Wilson, and D. R. Burton. 2003. Fine mapping of the interaction of neutralizing and nonneutralizing monoclonal antibodies with the CD4 binding site of human immunodeficiency virus type 1 gp120. J. Virol. 77:642-658.
Pantophlet, R., I. A. Wilson, and D. R. Burton. 2004. Improved design of an antigen with enhanced specificity for the broadly HIV-neutralizing antibody b12. Protein Eng. Des. Sel. 17:749-758.
Parren, P. W., I. Mondor, D. Naniche, H. J. Ditzel, P. J. Klasse, D. R. Burton, and Q. J. Sattentau. 1998. Neutralization of human immunodeficiency virus type 1 by antibody to gp120 is determined primarily by occupancy of sites on the virion irrespective of epitope specificity. J. Virol. 72:3512-3519.
Posner, M. R., L. A. Cavacini, C. L. Emes, J. Power, and R. Byrn. 1993. Neutralization of HIV-1 by F105, a human monoclonal antibody to the CD4 binding site of gp120. J. Acquir. Immune Defic. Syndr. 6:7-14.
Posner, M. R., T. Hideshima, T. Cannon, M. Mukherjee, K. H. Mayer, and R. A. Byrn. 1991. An IgG human monoclonal antibody that reacts with HIV-1/gp120, inhibits virus binding to cells, and neutralizes infection. J. Immunol. 146:4325-4332.
Raja, A., M. Venturi, P. Kwong, and J. Sodroski. 2003. CD4 binding site antibodies inhibit human immunodeficiency virus gp120 envelope glycoprotein interaction with CCR5. J. Virol. 77:713-718.
Read, R. J. 1997. Model phases: probabilities and bias. Methods Enzymol. 277:110-128.
Saphire, E. O., P. W. Parren, C. F. Barbas III, D. R. Burton, and I. A. Wilson. 2001. Crystallization and preliminary structure determination of an intact human immunoglobulin, b12: an antibody that broadly neutralizes primary isolates of HIV-1. Acta Crystallogr. D 57:168-171.
Saphire, E. O., P. W. Parren, R. Pantophlet, M. B. Zwick, G. M. Morris, P. M. Rudd, R. A. Dwek, R. L. Stanfield, D. R. Burton, and I. A. Wilson. 2001. Crystal structure of a neutralizing human IgG against HIV-1: a template for vaccine design. Science 293:1155-1159.
Saphire, E. O., R. L. Stanfield, M. D. Crispin, G. Morris, M. B. Zwick, R. A. Pantophlet, P. W. Parren, P. M. Rudd, R. A. Dwek, D. R. Burton, and I. A. Wilson. 2003. Crystal structure of an intact human IgG: antibody asymmetry, flexibility, and a guide for HIV-1 vaccine design. Adv. Exp. Med. Biol. 535:55-66.
Sattentau, Q. J., and J. P. Moore. 1991. Conformational changes induced in the human immunodeficiency virus envelope glycoprotein by soluble CD4 binding. J. Exp. Med. 174:407-415.
Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. Klenk. 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150:76-85.
Stanfield, R. L., M. K. Gorny, C. Williams, S. Zolla-Pazner, and I. A. Wilson. 2004. Structural rationale for the broad neutralization of HIV-1 by human monoclonal antibody 447-52D. Structure (Cambridge) 12:193-204.
Sullivan, N., Y. Sun, J. Li, W. Hofmann, and J. Sodroski. 1995. Replicative function and neutralization sensitivity of envelope glycoproteins from primary and T-cell line-passaged human immunodeficiency virus type 1 isolates. J. Virol. 69:4413-4422.
Thali, M., C. Furman, D. D. Ho, J. Robinson, S. Tilley, A. Pinter, and J. Sodroski. 1992. Discontinuous, conserved neutralization epitopes overlapping the CD4-binding region of human immunodeficiency virus type 1 gp120 envelope glycoprotein. J. Virol. 66:5635-5641.
Thali, M., J. P. Moore, C. Furman, M. Charles, D. D. Ho, J. Robinson, and J. Sodroski. 1993. Characterization of conserved human immunodeficiency virus type 1 gp120 neutralization epitopes exposed upon gp120-CD4 binding. J. Virol. 67:3978-3988.
Thali, M., U. Olshevsky, C. Furman, D. Gabuzda, M. Posner, and J. Sodroski. 1991. Characterization of a discontinuous human immunodeficiency virus type 1 gp120 epitope recognized by a broadly reactive neutralizing human monoclonal antibody. J. Virol. 65:6188-6193.
Trkola, A., M. Purtscher, T. Muster, C. Ballaun, A. Buchacher, N. Sullivan, K. Srinivasan, J. Sodroski, J. P. Moore, and H. Katinger. 1996. Human monoclonal antibody 2G12 defines a distinctive neutralization epitope on the gp120 glycoprotein of human immunodeficiency virus type 1. J. Virol. 70:1100-1108.
Wei, X., J. M. Decker, S. Wang, H. Hui, J. C. Kappes, X. Wu, J. F. Salazar-Gonzalez, M. G. Salazar, J. M. Kilby, M. S. Saag, N. L. Komarova, M. A. Nowak, B. H. Hahn, P. D. Kwong, and G. M. Shaw. 2003. Antibody neutralization and escape by HIV-1. Nature 422:307-312.
Wu, L., N. P. Gerard, R. Wyatt, H. Choe, C. Parolin, N. Ruffing, A. Borsetti, A. A. Cardoso, E. Desjardin, W. Newman, C. Gerard, and J. Sodroski. 1996. CD4-induced interaction of primary HIV-1 gp120 glycoproteins with the chemokine receptor CCR-5. Nature 384:179-183.
Wyatt, R., P. D. Kwong, E. Desjardins, R. W. Sweet, J. Robinson, W. A. Hendrickson, and J. G. Sodroski. 1998. The antigenic structure of the HIV gp120 envelope glycoprotein. Nature 393:705-711.
Wyatt, R., J. Moore, M. Accola, E. Desjardin, J. Robinson, and J. Sodroski. 1995. Involvement of the V1/V2 variable loop structure in the exposure of human immunodeficiency virus type 1 gp120 epitopes induced by receptor binding. J. Virol. 69:5723-5733.
Wyatt, R., and J. Sodroski. 1998. The HIV-1 envelope glycoproteins: fusogens, antigens, and immunogens. Science 280:1884-1888.
Wyatt, R., N. Sullivan, M. Thali, H. Repke, D. Ho, J. Robinson, M. Posner, and J. Sodroski. 1993. Functional and immunologic characterization of human immunodeficiency virus type 1 envelope glycoproteins containing deletions of the major variable regions. J. Virol. 67:4557-4565.
Xiang, S. H., P. D. Kwong, R. Gupta, C. D. Rizzuto, D. J. Casper, R. Wyatt, L. Wang, W. A. Hendrickson, M. L. Doyle, and J. Sodroski. 2002. Mutagenic stabilization and/or disruption of a CD4-bound state reveals distinct conformations of the human immunodeficiency virus type 1 gp120 envelope glycoprotein. J. Virol. 76:9888-9899.
Xu, W., R. Hofmann-Lehmann, H. M. McClure, and R. M. Ruprecht. 2002. Passive immunization with human neutralizing monoclonal antibodies: correlates of protective immunity against HIV. Vaccine 20:1956-1960.
Zemlin, M., M. Klinger, J. Link, C. Zemlin, K. Bauer, J. A. Engler, H. W. Schroeder, Jr., and P. M. Kirkham. 2003. Expressed murine and human CDR-H3 intervals of equal length exhibit distinct repertoires that differ in their amino acid composition and predicted range of structures. J. Mol. Biol. 334:733-749.
Zhang, W., G. Canziani, C. Plugariu, R. Wyatt, J. Sodroski, R. Sweet, P. Kwong, W. Hendrickson, and I. Chaiken. 1999. Conformational changes of gp120 in epitopes near the CCR5 binding site are induced by CD4 and a CD4 miniprotein mimetic. Biochemistry 38:9405-9416.
Zwick, M. B., R. Kelleher, R. Jensen, A. F. Labrijn, M. Wang, G. V. Quinnan, Jr., P. W. Parren, and D. R. Burton. 2003. A novel human antibody against human immunodeficiency virus type 1 gp120 is V1, V2, and V3 loop dependent and helps delimit the epitope of the broadly neutralizing antibody immunoglobulin G1 b12. J. Virol. 77:6965-6978.
Zwick, M. B., H. K. Komori, R. L. Stanfield, S. Church, M. Wang, P. W. Parren, R. Kunert, H. Katinger, I. A. Wilson, and D. R. Burton. 2004. The long third complementarity-determining region of the heavy chain is important in the activity of the broadly neutralizing anti-human immunodeficiency virus type 1 antibody 2F5. J. Virol. 78:3155-3161.
Zwick, M. B., P. W. Parren, E. O. Saphire, S. Church, M. Wang, J. K. Scott, P. E. Dawson, I. A. Wilson, and D. R. Burton. 2003. Molecular features of the broadly neutralizing immunoglobulin G1 b12 required for recognition of human immunodeficiency virus type 1 gp120. J. Virol. 77:5863-5876.(Royce A. Wilkinson, Chayn)
Beth Israel-Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02115
ABSTRACT
We have determined the crystal structure of the Fab fragment from F105, a broadly reactive human antibody with limited potency that recognizes the CD4 binding site of gp120. The structure reveals an extended CDR H3 loop with a phenylalanine residue at the apex and shows a striking pattern of serine and tyrosine residues. Modeling the interaction between gp120 and F105 suggests that the phenylalanine may recognize the binding pocket of gp120 used by Phe43 of CD4 and that numerous tyrosine and serine residues form hydrogen bonds with the main chain atoms of gp120. A comparison of the F105 structure to that of immunoglobulin G1 b12, a much more potent and broadly neutralizing antibody with an overlapping epitope, suggests similarities that contribute to the broad recognition of human immunodeficiency virus by both antibodies. While the putative epitope for F105 shows significant overlap with that predicted for b12, it appears to differ from the b12 epitope in extending across the interface between the inner and outer domains of gp120. In contrast, the CDR loops of b12 appear to interact predominantly with the outer domain of gp120. The difference between the predicted epitopes for b12 and F105 suggests that the unique potency of b12 may arise from its ability to avoid the interface between the inner and outer domains of gp120.
INTRODUCTION
A key step in the development of a successful vaccine against human immunodeficiency virus (HIV) will be the design of immunogens capable of generating an effective humoral immune response (3). Antibodies with two features characterize such a response. They must show, at the same time, both potency and broad specificity. A number of human antibodies that recognize elements of the conserved CD4 binding site of HIV type 1 (HIV-1) gp120 have now been isolated from HIV-infected patients (2, 11, 16, 24, 38, 40, 47, 49, 59, 65). This includes immunoglobulin G1 (IgG1) b12 (38, 47, 53), one of the most potent and broadly reactive anti-HIV antibodies known. In contrast, other CD4 binding-site antibodies are less potent towards many clinical isolates, even though they may be broadly cross-reactive. F105, the subject of this paper, is representative of the latter group.
F105 is an IgG1 human monoclonal antibody isolated from an HIV-infected individual (49). It binds to the CD4 binding sites of both trimeric and monomeric gp120 and is capable of neutralizing various strains of HIV (e.g., IIIB [HXBc2], MN, RF, and SF2) (7, 49, 58) but is less successful against many primary clinical isolates (12, 32). F105 did not show evidence of anti-HIV-1 activity or a viral load decrease in a phase I dose-escalation study (5, 70). However, in triple and quadruple combination therapies with anti-HIV monoclonal antibodies (2F5, 2G12, and 694/98D) with other specificities, a complete and synergistic neutralization of the SHIV-Vpu+ chimeric simian-human immunodeficiency virus was seen in macaque peripheral blood mononuclear cells in vitro and in an in vivo macaque model that mimics mucosal exposure during intrapartum virus transmission (1, 31).
Crystallographic studies of the ternary complex of the HIV gp120 core, CD4, and antibody 17b provided the first look at the structure of gp120 and its interactions with CD4 (26-28, 65). CD4 was found to bind at the nexus of the inner domain, the outer domain, and the bridging sheet of gp120 (26, 27). A large body of biochemical and biophysical data indicates a considerable conformational change in gp120 upon binding to CD4 (4, 6, 15, 24, 26, 27, 39, 41, 50, 55, 60, 62-69, 72, 73). The conformational change results in the formation and/or exposure of the chemokine receptor sites (62, 64), thus promoting further viral attachment and membrane fusion.
The molecular reorganization that results upon binding of CD4 is revealed by the structure of an unliganded simian immunodeficiency virus (SIV) gp120 core (6). With a few important exceptions, the structure of the outer domain is quite similar to that seen in the CD4-bound state. In contrast, the structure of the inner domain is markedly different. A comparison of CD4-bound and unliganded gp120 shows that the conformational change is not a simple movement of the inner domain as a rigid body. Rather, the inner domain is comprised of a set of distinct substructures that move relatively independently of one another (6). The binding of CD4 results in a rearrangement of these secondary structural elements within the inner domain (6). As predicted (24), the bridging sheet is not present in the structure of unliganded gp120 (6).
The structure of antibody b12 was determined by Saphire et al. (52-54). The structure revealed an extended CDR H3 loop with an apical tryptophan residue that is thought to recognize the Phe43 binding pocket of gp120 (53, 75). This putative interaction places significant constraints on possible gp120/b12 interactions, allowing an interaction between b12 and the CD4-bound conformation of gp120 to be modeled (53, 75). This model suggests that the broad neutralizing activity of b12 lies in its ability to interact with conserved features of the CD4 binding site through recognition of the Phe43 pocket and interactions with main chain atoms of gp120 and in its ability to recognize trimeric gp120 on the native viral surface.
A further understanding of the molecular properties that confer broad reactivity and potency upon a CD4 binding-site antibody are of substantial interest. The underlying principles are likely to impact the design of new immunogens with the potential to elicit a successful immune response. In this regard, the structure of F105 provides an opportunity to compare and contrast the structure of a broadly reactive but nonpotent CD4 binding-site antibody (F105) with that of a broadly neutralizing antibody with an overlapping epitope (b12). The comparison provides significant insight into the molecular properties that impart broad reactivity and potency upon a CD4 binding-site antibody.
MATERIALS AND METHODS
Expression and purification. The expression of F105 was performed as previously described (49). F105 was purified over protein G and eluted with 0.1 M glycine-HCl, pH 2.7, which was immediately neutralized upon collection with 1 M Tris, pH 9. Previous work had shown some contamination of protein G-purified F105 with bovine Ig from the fetal calf serum used for growing the hybridoma cells (19). Therefore, we further purified F105 using Immunopure immobilized protein L (Pierce), eluted it with 0.1 M glycine-HCl, pH 2.5, and neutralized the eluant with 1 M Na2HPO4, pH 7.2. The elution fractions containing F105 were combined and concentrated (Amicon Centricon filter with 30K molecular weight cutoff) to 4.5 mg/ml.
Fab fragments of the purified antibody were prepared by digestion with papain (0.4 μM) (ICN Biomedicals) in the presence of 20 mM ?-mercaptoethanol. The reaction was stopped by dialyzing against phosphate-buffered saline to remove the ?-mercaptoethanol, and the completeness of the digestion was verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis.
Fab fragments were purified by anion-exchange chromatography using a POROS HQ10 column on a BioCAD Sprint perfusion chromatography system (Perseptive BioSystems, Inc.) using a linear gradient from 0 to 1.0 M NaCl in 10 mM Tris, pH 8.1. Fractions containing the F105 Fab fragment were concentrated to 6.0 mg/ml in an Amicon Ultra 10K-molecular-weight-cutoff centrifugal filter. Protein concentrations were determined by bicinchoninic acid analysis using bovine serum albumin as a standard (56). The overall yield was approximately 2.1 mg of purified Fab from 10 mg of protein G-purified antibody.
Crystallization and data collection. Purified Fab from F105 was crystallized by hanging-drop vapor diffusion at 18°C. Drops were assembled with 2 μl of F105 Fab fragment mixed with 2 μl of well solution containing 15.5% polyethylene glycol 4000, 25% isopropanol, 100 mM NaCl, and 75 mM sodium citrate, pH 5.6, with or without 1 μl of 600 μM peptide 44 (19). Crystals typically appeared in 1 to 7 weeks. Single crystals were frozen in liquid nitrogen. Data were collected at 100 K using a Rigaku RUH3R rotating anode, a CuK X-ray source, and a MAR345 image plate detector. F105 Fab fragment crystals diffracted to a 2.8-? resolution and belonged to space group P43212, with a = b = 120.4 ? and c = 72.98 ?, with a single Fab per asymmetric unit. Data were integrated and reduced using the HKL software package (42). Statistics on data completeness and quality are presented in Table 1.
Structure determination and refinement. The structure of the F105 Fab was determined by molecular replacement. Due to large variations in the elbow angles of crystallized Fabs, the molecular replacement procedure typically breaks the chosen Fab search model into constant and variable domains, and each domain is independently positioned. However, in this case, we chose to use the family search option of COMO and a library of 195 Fab search models. This allowed us to utilize entire Fabs as search models, and due to the large number of models, we were able to effectively search the available "elbow angle space" to identify the correct solution. The library was composed of 195 human, mouse, and rat IgG Fab structures taken from the protein data bank. Each downloaded structure file was processed with a series of script files to generate a single Fab search model free of hetero compounds. Upon completion of the library, the automated search process was rapid and successful, placing a humanized anti-CD18 Fab (PDB entry 2FGW) (13) into the F105 asymmetric unit to yield a correlation coefficient of 43.5% and an initial R factor of 40.2%.
Subsequently, nonconserved residues in the positioned 2FGW model were mutated to alanine or glycine, as appropriate, and the variable and constant domains were subjected to independent rigid body refinement with CNS, followed by positional and independent B-factor refinement. Sigma A-weighted 2mFo-dFc and mFo-dFc maps (51) then showed clearly interpretable density for much of the missing model, including very nice density for the large CDR H3 loop present in F105. It is noteworthy that the CDR H3 loop was disordered in 2FGW and therefore was not present in the search model. All data between 15.0 and 2.8 ? were used for the refinement, with 5% of the data randomly chosen for Rfree calculations. Rounds of iterative model building with O (20) and refinement with CNS resulted in a final model with an R factor of 21.0% (Rfree, 25.1%). The statistics on model quality are summarized in Table 1. The model exhibits good geometry, with 88.8% of residues in the most favored regions of the Ramachandran plot, 11.5% in additionally allowed regions, and one residue in a generously allowed region (29). One residue, AlaL51, is found in a disallowed region. It is located in a highly conserved turn and contains geometry typical of CDR L2 structures seen in many other antibodies (53). The figures were generated with PYMOL (http://www.pymol.org).
Docking. The F105 Fab structure was manually docked to core gp120 from the structure of the gp120-CD4-Fab17b complex (PDB entry 1GM9) in an orientation that places the antigen binding site of the Fab in the CD4 binding pocket of gp120, minimizes bad contacts, and maximizes overlap of the antigen-combining site with residues implicated in binding of F105, as described in Results and Discussion. b12 docking was manually recreated using information provided by Saphire et al. (53) and Zwick et al. (75). Specifically, b12 was positioned such that Trp100 of the H3 loop was inserted into the hydrophobic pocket of gp120, Tyr53 of the H2 loop was fit into the gap between Thr373 and Asn386 of gp120, and Arg29 and Arg31 of the L1 loop were in close proximity to Asn276/Lys282 and Asn280/Ala281 of the D loop of gp120, respectively (75). The relative position of hand-docked b12 to gp120 was consistent with the proposed contacts described by Saphire et al. (53), Zwick et al. (75), and Pantophlet et al. (45). Following the addition of hydrogen atoms and capping of the polypeptide chains, both the F105/gp120 and b12/gp120 complexes were subjected to energy minimization using the programs InsightII and Discover (Accelrys, San Diego, Calif.). An initial 100 iterations were carried out using the steepest descent algorithm to minimize bad contacts, followed by 1,000 iterations using conjugate gradient minimization. In the minimized F105/gp120 complex, PheH100A of F105 retains an interaction with residues forming the Phe43 pocket of gp120 and ArgL31 forms a salt bridge to gp120 residues Asp474 and Asp477.
Amino acid content of human CDR loops. The amino acid sequences of CDR loops in complete human antibodies present in the Kabat sequence database were downloaded using Kabatman software (35). The frequency of occurrence of each amino acid, as well as that of serine/tyrosine pairs, was determined using a small Fortran program written for this purpose.
Protein structure accession number. Atomic coordinates and structure factors have been deposited in the Protein Data Bank under accession code 1U6A.
RESULTS AND DISCUSSION
Structure of F105 Fab fragment. The F105 model includes light chain (chain L) residues 1 through 213 and heavy chain (chain H) residues 2 through 213. There are no breaks in the main chain electron density, and the side chain density for most residues, including those in the CDR loops, is also quite good. The structure of the F105 Fab fragment reveals several notable features. The most obvious of these is a relatively long CDR H3 loop (Fig. 1). This 15-residue loop is well ordered and extends approximately 7 ? beyond the other CDR loops, with residues PheH100A and TyrH100B (Kabat numbering) at the apex. This conformation of the H3 loop is stabilized by a hydrogen bond between the main chain carbonyl of GlyH100C and the side chain of ArgL31A, which is also prominent at the top of the antigen recognition site. Further stabilization of the H3 loop may arise from its three proline residues (Fig. 1A). The increased occurrence of proline in CDR H3 loops has been suggested to confer stability to longer CDR H3 regions (71).
The PheH100A and TyrH100B side chains are involved in lattice contacts within the crystal. One face of the PheH100A side chain packs against the main chain atoms of ProH136, SerH137, and SerH138 in a neighboring molecule, while the tyrosine side chain contacts the C atom of GlyH200 in an adjacent F105 molecule. Overall, the lattice contacts do not appear to provide substantial stabilization of the ordered H3 loop. Thus, it is likely that an ordered H3 loop is intrinsic to F105, rather than a result of lattice contacts within the crystal.
The F105 structure also reveals a striking constellation of nine tyrosine residues aligned across the top of the antigen recognition site (Fig. 1). This includes several tyrosine residues nestled around the base of the H3 loop that might provide steric constraints and favorable interactions leading to reduced conformational flexibility of the H3 loop. For example, ValH97 packs tightly against the ring of TyrH33. The extended H3 loop, the nine tyrosine residues, and ArgL31 appear to be principle features of the F105 paratope.
However, the CDR sequences are also rich in serine (Fig. 1). We found that 14 of 64 residues present in the CDRs are serine. Unlike the tyrosine residues, which are distributed across the top of the antigen binding site, the serine residues are concentrated along the periphery (Fig. 1C). Combined, the serine and tyrosine residues account for 23 of 64 CDR residues (36%).
Structural comparison of F105 and b12. An extended H3 loop is also seen in antibody b12, with a tryptophan at its apex. It is thought to protrude into the recessed CD4 binding site of gp120, with the tryptophan occupying the same pocket that accommodates Phe43 of CD4 (53, 75). However, the H3 loop in b12 is slightly longer than that in F105, as it is 18 residues in length versus 15 residues in F105. Two views of a superposition of F105 and b12 are shown in Fig. 1D and E. This superposition suggests that F105 PheH100A might also be able to reach the Phe43 binding pocket of gp120 in a manner analogous to that of Trp100 in the b12 antibody and Phe43 of CD4.
However, this does not necessarily indicate that the extended H3 loop of F105, with its apical phenylalanine, recognizes the Phe43 pocket of gp120. In contrast to the case for murine antibodies, extended H3 loops are apparently more common in human antibodies that recognize virus (8, 71). Collis et al. have found an average length of 16.5 residues in antiviral human H3 loops (8), and many non-CD4 binding-site antibodies capable of neutralizing HIV show extended loops (9, 17, 26, 27, 57, 74). Mutational analyses and antibody competition studies suggest that CD4 and CD4 binding-site antibodies recognize different conformations of gp120 (50, 69, 73). In addition, the binding of b12, F105, and CD4 to core gp120 results in entropic changes of 5.7 kcal/mol, 18.9 kcal/mol, and 35 kcal/mol, respectively (24, 41). Together, these data suggest that the F105-bound conformation of gp120 is significantly different from the unliganded or CD4-bound conformation of gp120.
The entropic argument is also relevant to b12, where an entropic difference of 30 kcal/mol is seen for binding of b12 versus CD4 (24). Indeed, reliance on entropic considerations alone would suggest that b12 should be docked to the unliganded conformation of gp120 rather than the CD4-bound conformation. However, the model for b12 docked to the CD4-bound conformation of gp120 is consistent with subsequent mutational and structural studies and appears to be substantially correct, whereas b12 is unlikely to bind the unliganded gp120 structure (6, 45, 53, 75). One explanation for this is that elements of the b12 epitope are present within multiple conformations of gp120, including the CD4-bound conformation. Thus, while binding of b12 does not induce the CD4-bound conformation, the CD4-bound conformation of gp120 does represent a starting point for modeling the interaction between b12 and gp120.
Thus, it is not clear whether the F105-bound conformation of gp120 bears the greatest similarity to the CD4-bound state, the unliganded state, an intermediate conformation, or a conformation that is altogether different. We have addressed this question by mapping gp120 residues critical for binding of F105 to the CD4-bound and unliganded structures of gp120 (Fig. 2) (45, 59, 61). For reference, we also mapped 26 residues (26, 27) known to make direct contact with CD4. In the unliganded structure, only a few residues implicated in interactions with F105 are apparent on the surface of gp120; many of the critical residues are occluded by residues N-terminal to helix 1 and by helix 1 itself (Fig. 2B and D).
In the CD4-bound state, however, many of the critical residues are solvent exposed where they are found along the lip of the Phe43 binding pocket, ?-strand 15 (CD4 binding loop), the N-terminal end of helix 5, and its preceding loop. In contrast to the unliganded state of gp120, in the CD4-bound state these residues form a nearly contiguous solvent-exposed surface, one that intersects the surface involved in the recognition of CD4 (26, 27, 59, 61, 65). Figure 2C illustrates the recessed nature of these residues, many of which lie at or near the interface of the inner and outer domains of gp120 (Fig. 2A), strongly suggesting that residues forming the rim of the Phe43 binding pocket are recognized by F105. Similarly, mutations that alter the binding of antibody b12 also map to a contiguous surface on CD4-bound gp120 (45, 75) but are occluded or more dispersed across the surface of the unliganded structure (6).
Further examination of the ligand-free and CD4-bound conformations suggests that F105 is likely to recognize a conformation in which the transition of the inner domain to the CD4-bound conformation is nearing completion rather than an inner domain conformation that is intermediate between the two states. As gp120 transits from the ligand-free conformation to the CD4-bound conformation, the center of mass of helix 1 moves 15 ? away from the outer domain (Fig. 2B) (6). An incomplete inner domain transition, at least with respect to 1, would presumably occlude critical residues in helix 5, restricting antibody access to residues on the floor of the CD4 binding site. Furthermore, Chen and coworkers pointed out that 1 is an amphipathic helix and that an intermediate position of 1 is likely to bury a set of charged residues. They suggested that intermediate conformations of 1 would therefore be unstable (6).
Thus, surface mapping of residues critical for F105 binding to the ligand-free and CD4-bound conformations suggests that the F105-bound conformation of gp120 will have an inner domain structure approaching that seen in the CD4-bound state. This does not imply that the interface between the inner and outer domains is identical to that in the CD4-bound state; conformational differences are likely to exist, perhaps resulting in the exposure of additional residues critical to the recognition of F105 and other CD4 binding-site antibodies (65). Furthermore, while the inner domain transition appears largely complete, the bridging sheet may not be formed (6), consistent with the calorimetric data of Kwong et al. (24). In this respect, it is likely that the F105-bound state is intermediate between the unliganded and CD4-bound conformations of gp120.
This also suggests that the inner and outer domains of CD4-bound gp120 are the preferred starting point to examine possible interactions between F105 and gp120. For example, by docking F105 to the CD4-bound conformation of gp120, we can test whether the extended H3 loop and other elements of the F105 paratope are able to access critical residues on the floor of the CD4 binding site. If so, we can ask whether the interacting surfaces appear to be complementary with respect to shape, charge, hydrophilicity, and hydrophobicity and how these interactions compare to those suggested for b12 (45, 53, 75).
As the starting point for the docking exercise, we considered the possibility that the apical PheH100A of F105 functions in a manner analogous to that of Phe43 of CD4, i.e., that PheH100A binds within the Phe43 pocket of gp120. Thus, superposition of the phenylalanine side chains of F105 and CD4 provided an initial docking of F105 to gp120 from strain HXBc2, an interaction that is known to occur (7, 24, 25). This was followed by rotation of F105 in order to minimize bad contacts, to maximize surface complementarity, and to maximize interactions with gp120 residues critical to the binding of F105. Following energy minimization of the F105/gp120 model, we found that PheH100A of F105 remains in the Phe43 binding pocket of gp120. This demonstrates that even though the H3 loop of F105 is three residues shorter than that of b12, it is long enough to access the Phe43 pocket of gp120.
Importantly, the minimized structure also suggests logical roles for other elements in the F105 paratope. Many of the tyrosine and serine side chains demonstrate the ability to form hydrogen bonds with gp120, including numerous interactions with exposed main chain atoms in the CD4 binding site, while ArgL31 can form salt bridges to acidic residues in helix 5 of gp120. Thus, the docking exercise suggests that elements of the F105 epitope might indeed reside within the CD4-bound conformation of gp120.
As expected, there are similarities and differences between the footprint of docked F105 on gp120 and that predicted for b12 (45, 46, 53, 75). For both F105 and b12, the heavy chains are predicted to straddle a ridge on the surface of gp120 formed by ?-strand 15. The H3 loops project down one side of the ridge towards the Phe43 pocket of gp120, while residues in the H2 loops interact with the opposite side. In the case of F105, however, the cleft between the H2 and H3 loops is significantly less pronounced than that seen in b12, and thus the extension of the H2 loop down the opposite side of the ?15 ridge is reduced. ?-Strand 15 is one outer domain element that does undergo significant rearrangement upon binding of CD4, and it is referred to as the CD4 binding loop by Chen et al. (6). Surface mapping suggests that the CD4-bound conformation of this loop, rather than the ligand-free conformation, is more likely to be recognized by F105. However, alternate conformations of the CD4 binding loop that expose additional residues implicated in the F105/gp120 interaction could also be considered.
Distinct differences between the putative surfaces recognized by the b12 and F105 light chains are also apparent. In the case of b12, interactions between the light chain CDRs and outer domain residues, including residues within the D loop (Fig. 2), are thought to be critical (45, 75). In contrast, the model for F105 suggests that in addition to interactions with outer domain residues, the light chain also recognizes inner domain residues within helix 5. Mutations to gp120 residues 473 through 476 (GDMR) at the N-terminal end of helix 5 inhibit the recognition of gp120 by F105 but do not inhibit the binding of antibody b12 (45). This indicates an important difference between the F105 and b12 epitopes. In this regard, the docking model for F105 is consistent with mutational data indicating the importance of these residues (45, 59).
Neutralizing ability of b12 versus that of F105. The broad neutralizing ability of b12 has been attributed to several properties (53). These include interactions with conserved elements of the CD4 binding site, many of which are mediated through contacts with main chain atoms of gp120, and the interaction of the extended H3 loop with the rim of the Phe43 binding pocket. In addition, b12 is thought to bind in a manner that allows it to access the epitope not only on monomeric gp120, but on the native viral surface as well (45, 53). In the case of F105, the docking exercise makes specific suggestions regarding the mechanisms that provide for broad recognition of the CD4 binding site by F105. These include an extended H3 loop that interacts with conserved residues surrounding the Phe43 binding pocket and the use of tyrosine and serine residues of F105 to hydrogen bond to exposed main chain atoms of the CD4 binding site, properties thought to contribute to the unique potency of b12. When the exercise is extended to the trimeric gp120 model of Kwong et al. (28), it suggests that F105 is also capable of accessing the epitope on the native viral surface, a property that F105 is known to possess (48, 49). Thus, many of the properties suggested to contribute to the potency of b12 are also seemingly present in F105. In this regard, these shared properties provide an explanation for the broad reactivities of both F105 and b12. It is less clear, however, why they confer such potency to b12, while F105 is relatively nonpotent.
Perhaps it is a matter of degree. In addition to a possible interaction with the Phe43 pocket of gp120, the longer H3 loop of b12 results in a lateral extension that may be more effective in its interactions with the protruding ridge formed by the CD4 binding loop of gp120. The deeper cleft between the H2 and H3 loops might also result in stronger interactions between b12 and gp120. However, we believe that the proposed interaction between the light chain of F105 and the inner domain of gp120 is another obvious difference between these two antibodies.
The binding of F105 across the domain interface would be expected to result in a significant ordering of residues at the inner domain/outer domain junction. It is also likely to lead to reduced group motions for other inner domain substructures, consistent with the large decrease in entropy that is observed (6, 24). In contrast to the case for F105, the reduced entropic component (24) and the modeled b12 interactions (45, 53, 75) both suggest that b12 largely avoids the domain interface, restricting primary recognition to the outer domain of gp120. In this regard, the combination of the deeper H2/H3 cleft and the lateral component of the extended H3 loop may be critical in allowing b12 to recognize an epitope centered on the outer domain (Fig. 1E and 2). The deeper cleft between the H2 and H3 loops potentially allows b12 to "tip back," away from the inner domain/outer domain junction (Fig. 2E and F), while maintaining the interaction between the tip of the H3 loop and the Phe43 pocket of gp120. Furthermore, the relative orientations of the H2/H3 cleft and the lateral extension of the H3 loop may position the b12 light chain to contact the D loop of gp120. The net result is movement of the b12 light chain away from the interdomain junction.
We would like to emphasize that the CD4-bound conformation of gp120 certainly differs from the F105-bound conformation. In particular, the bridging sheet is unformed in the unliganded state of monomeric gp120 and may also be absent in the F105-bound state. In addition, residues in the CD4 binding loop (?15) or those at the inner domain/outer domain junction might differ significantly from the CD4-bound conformation (65). Indeed, several residues near the interdomain junction that affect the binding of F105 are not surface exposed, even in the CD4-bound conformation (59, 61, 65). Thus, the docking of F105 to gp120 should be considered only a low-resolution model for the interaction of F105 with gp120, but one that is consistent with known constraints from previous mutational, structural, and calorimetric analyses. Importantly, it suggests that the greater potency of b12 can be attributed to an epitope that lies primarily on the outer domain of gp120, allowing b12 to avoid the conformational or entropic masking that is apparently so effective at protecting the HIV virus from most other CD4 binding-site antibodies, including F105 (24).
Maturation and amino acid content of the CDR loops. Another notable difference between F105 and b12 is the number of mutations from the germ line V and J region sequences. Huang et al. calculated the number of such mutations in 25 human gp120-reactive antibodies, finding an average of 22 (17). F105 has among the fewest, at 13, whereas b12 has one of the highest, at 45 (17), many of which are thought to be important for its ability to bind gp120 (75). Thus, b12 represents an antibody that has undergone extensive affinity maturation by somatic mutation, while F105 is more representative of the germ line immune repertoire. Extensive mutation of b12 in response to a changing gp120 protein in the patient may have increased its potency while maintaining broad specificity. It is notable that 2G12, one of the few antibodies to gp120 that is potent and broadly neutralizing like b12, has 51 V and J somatic mutations (17). Furthermore, studies of broadly neutralizing human anti-HIV antibodies directed against gp41 epitopes have also noted that their potency is correlated with long CDR H3 loops and extensive affinity maturation by somatic mutation (23).
The frequent occurrence of serine and tyrosine residues within CDRs has been noted before and is in fact a general phenomenon (8, 10, 14, 18, 21, 22, 30, 33, 37, 43, 44, 71). For example, Mian et al. (37) reported a mean composition for vertebrate CDRs that shows an average serine content of 17.27%, a tyrosine content of 11.08%, and thus a combined serine/tyrosine content of 28%. In our own analysis, we found the combined serine/tyrosine content in CDRs of human antibodies to be 27.0% ± 6.6%. Relative to these values, the combined serine and tyrosine content of 36% for F105 is significantly above average. The heavy use of serine and tyrosine residues in F105 raises several questions deserving of further comment.
(i) Why are serine and tyrosine residues favorable to recognition of the CD4 binding site? The docking exercise suggests that the numerous tyrosine and serine residues allow F105 to maximize hydrogen bonding to the exposed main chain atoms of the CD4 binding site, thus helping to confer broad reactivity to F105. These residues are capable of acting as both hydrogen bond donors and acceptors (36) and can thus hydrogen bond to both carbonyl oxygen atoms and amide protons of the polypeptide backbone. Several additional properties of tyrosine are thought to contribute to its frequent use in antigen recognition (14, 18, 37). Together, these properties make tyrosine particularly well suited for the recognition of variable epitopes.
The relative positions of the serine and tyrosine residues within the F105 CDRs are illustrative of those in antibodies in general (8, 37). For example, Mian et al. (37) have noted an increased serine and tyrosine content for residues in known binding positions. They also found that while the tyrosine content was greater at known binding positions than in the remainder of the CDR, the greatest enrichment of serine was in regions of the CDR loops distal to the known binding sites.
(ii) What is the purpose of the serine residues along the periphery of the antigen combining site? Mian et al. suggested that these residues play a structural role (37). However, the frequency with which serine is found in the CDRs outside of the contact residues is much greater than that generally seen in loops or ?-strands (8), suggesting that serine residues in this region of the CDR are likely to serve a functional role as well. The use of the small polar serine side chain along the periphery of the antigen recognition site should maximize access to a recessed antigen binding site. Furthermore, in the case of incidental contact between these residues and the antigen, the chances of a favorable contact are also maximized, while the entropic penalty for ordering of the side chain is minimal. Since the CD4 binding site is recessed and the opening to this site is quite narrow, at least in one dimension, the serine residues along the periphery of F105 might play a role in promoting the interaction of F105 with gp120.
(iii) Has maturation of F105 resulted in an increased occurrence of serine and tyrosine residues in the CDRs, or are the germ line genes naturally rich in serine and tyrosine? The answer appears to be the latter. The V and J gene precursors for the F105 heavy chain are VH4-59 and JH4, respectively, while the light chain precursors are A27a and JK2. There are 10 mutations that fall within the Kabat CDRs, resulting in a net loss of one tyrosine and two serine residues.
(iv) Are other CD4 binding-site antibodies rich in serine and tyrosine? We found an even greater serine/tyrosine content in the CD4 binding-site antibody 15e (2, 16, 38), where fully 40% of the CDR residues are serine or tyrosine. In contrast, the serine/tyrosine content of antibody b12 is only about average (25%). However, the germ line precursors for antibody b12 do exhibit an elevated serine/tyrosine content (40%). Thus, the decreased occurrence reflects the extensive affinity maturation that has taken place in b12 relative to F105 and 15e.
In summary, the structure of F105 suggests a common molecular basis for the broad recognition of HIV-1 by F105 and b12. It reaffirms the need for immunogens capable of eliciting antibodies with extended H3 loops and suggests that the unique potency of antibody b12 resides in its ability to avoid the interface between the inner and outer domains of gp120. Thus, approaches to generating a b12-like immune response might consider ways to elicit extended H3 loops while minimizing interactions with the inner domain residues of gp120. The subtle differences between the H3 loops of b12 and F105 speak to the difficulty of doing so in a vaccine setting.
ACKNOWLEDGMENTS
This work was supported by grants AI49753 (M.T.) and AI26926 (M.R.P.) from the National Institutes of Health.
We sincerely thank all reviewers for their helpful suggestions.
REFERENCES
Baba, T. W., V. Liska, R. Hofmann-Lehmann, J. Vlasak, W. Xu, S. Ayehunie, L. A. Cavacini, M. R. Posner, H. Katinger, G. Stiegler, B. J. Bernacky, T. A. Rizvi, R. Schmidt, L. R. Hill, M. E. Keeling, Y. Lu, J. E. Wright, T. C. Chou, and R. M. Ruprecht. 2000. Human neutralizing monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human immunodeficiency virus infection. Nat. Med. 6:200-206.
Bagley, J., P. J. Dillon, C. Rosen, J. Robinson, J. Sodroski, and W. A. Marasco. 1994. Structural characterization of broadly neutralizing human monoclonal antibodies against the CD4 binding site of HIV-1 gp120. Mol. Immunol. 31:1149-1160.
Burton, D. R., R. C. Desrosiers, R. W. Doms, W. C. Koff, P. D. Kwong, J. P. Moore, G. J. Nabel, J. Sodroski, I. A. Wilson, and R. T. Wyatt. 2004. HIV vaccine design and the neutralizing antibody problem. Nat. Immunol. 5:233-236.
Cao, J., N. Sullivan, E. Desjardin, C. Parolin, J. Robinson, R. Wyatt, and J. Sodroski. 1997. Replication and neutralization of human immunodeficiency virus type 1 lacking the V1 and V2 variable loops of the gp120 envelope glycoprotein. J. Virol. 71:9808-9812.
Cavacini, L. A., C. L. Emes, A. V. Wisnewski, J. Power, G. Lewis, D. Montefiori, and M. R. Posner. 1998. Functional and molecular characterization of human monoclonal antibody reactive with the immunodominant region of HIV type 1 glycoprotein 41. AIDS Res. Hum. Retrovir. 14:1271-1280.
Chen, B., E. M. Vogan, H. Gong, J. J. Skehel, D. C. Wiley, and S. C. Harrison. 2005. Structure of an unliganded simian immunodeficiency virus gp120 core. Nature 433:834-841.
Choe, H., W. Li, P. L. Wright, N. Vasilieva, M. Venturi, C. C. Huang, C. Grundner, T. Dorfman, M. B. Zwick, L. Wang, E. S. Rosenberg, P. D. Kwong, D. R. Burton, J. E. Robinson, J. G. Sodroski, and M. Farzan. 2003. Tyrosine sulfation of human antibodies contributes to recognition of the CCR5 binding region of HIV-1 gp120. Cell 114:161-170.
Collis, A. V., A. P. Brouwer, and A. C. Martin. 2003. Analysis of the antigen combining site: correlations between length and sequence composition of the hypervariable loops and the nature of the antigen. J. Mol. Biol. 325:337-354.
Darbha, R., S. Phogat, A. F. Labrijn, Y. Shu, Y. Gu, M. Andrykovitch, M. Y. Zhang, R. Pantophlet, L. Martin, C. Vita, D. R. Burton, D. S. Dimitrov, and X. Ji. 2004. Crystal structure of the broadly cross-reactive HIV-1-neutralizing Fab X5 and fine mapping of its epitope. Biochemistry 43:1410-1417.
Davies, D. R., and G. H. Cohen. 1996. Interactions of protein antigens with antibodies. Proc. Natl. Acad. Sci. USA 93:7-12.
Ditzel, H. J., J. M. Binley, J. P. Moore, J. Sodroski, N. Sullivan, L. S. Sawyer, R. M. Hendry, W. P. Yang, C. F. Barbas III, and D. R. Burton. 1995. Neutralizing recombinant human antibodies to a conformational V2- and CD4-binding site-sensitive epitope of HIV-1 gp120 isolated by using an epitope-masking procedure. J. Immunol. 154:893-906.
D'Souza, M. P., D. Livnat, J. A. Bradac, and S. H. Bridges. 1997. Evaluation of monoclonal antibodies to human immunodeficiency virus type 1 primary isolates by neutralization assays: performance criteria for selecting candidate antibodies for clinical trials. AIDS Clinical Trials Group Antibody Selection Working Group. J. Infect. Dis. 175:1056-1062.
Eigenbrot, C., T. Gonzalez, J. Mayeda, P. Carter, W. Werther, T. Hotaling, J. Fox, and J. Kessler. 1994. X-ray structures of fragments from binding and nonbinding versions of a humanized anti-CD18 antibody: structural indications of the key role of VH residues 59 to 65. Proteins 18:49-62.
Fellouse, F. A., C. Wiesmann, and S. S. Sidhu. 2004. Synthetic antibodies from a four-amino-acid code: a dominant role for tyrosine in antigen recognition. Proc. Natl. Acad. Sci. USA 101:12467-12472.
Hart, T. K., R. Kirsh, H. Ellens, R. W. Sweet, D. M. Lambert, S. R. Petteway, Jr., J. Leary, and P. J. Bugelski. 1991. Binding of soluble CD4 proteins to human immunodeficiency virus type 1 and infected cells induces release of envelope glycoprotein gp120. Proc. Natl. Acad. Sci. USA 88:2189-2193.
Ho, D. D., J. A. McKeating, X. L. Li, T. Moudgil, E. S. Daar, N. C. Sun, and J. E. Robinson. 1991. Conformational epitope on gp120 important in CD4 binding and human immunodeficiency virus type 1 neutralization identified by a human monoclonal antibody. J. Virol. 65:489-493.
Huang, C. C., M. Venturi, S. Majeed, M. J. Moore, S. Phogat, M. Y. Zhang, D. S. Dimitrov, W. A. Hendrickson, J. Robinson, J. Sodroski, R. Wyatt, H. Choe, M. Farzan, and P. D. Kwong. 2004. Structural basis of tyrosine sulfation and VH-gene usage in antibodies that recognize the HIV type 1 coreceptor-binding site on gp120. Proc. Natl. Acad. Sci. USA 101:2706-2711.
Ivanov, I., J. Link, G. C. Ippolito, and H. W. Schroeder. 2002. Constraints on the hydropathicity and sequence composition of HCDR3 are conserved across evolution, p. 43-67. In M. Zanetti and J. Capra (ed.), The antibodies. Taylor and Francis, London, United Kingdom.
Jacobs, J. M., B. W. Bailey, J. B. Burritt, S. G. Morrison, R. P. Morrison, E. A. Dratz, A. J. Jesaitis, and M. Teintze. 2003. QSYP peptide sequence is selected from phage display libraries by bovine IgG contaminants in monoclonal antibody preparations. BioTechniques 34:132-134, 137-141.
Jones, T. A., J. Y. Zou, S. W. Cowan, and M. Kjeldgaard. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47:110-119.
Kabat, E. A., T. T. Wu, and H. Bilofsky. 1977. Unusual distributions of amino acids in complementarity-determining (hypervariable) segments of heavy and light chains of immunoglobulins and their possible roles in specificity of antibody-combining sites. J. Biol. Chem. 252:6609-6616.
Kabat, E. A., T. T. Wu, M. Reid-Miller, H. M. Perry, and K. S. Gottesman. 1987. Sequences of proteins of immunological interest. U.S. Department of Health and Human Services, Washington, D.C.
Kunert, R., S. Wolbank, G. Stiegler, R. Weik, and H. Katinger. 2004. Characterization of molecular features, antigen-binding, and in vitro properties of IgG and IgM variants of 4E10, an anti-HIV type 1 neutralizing monoclonal antibody. AIDS Res. Hum. Retrovir. 20:755-762.
Kwong, P. D., M. L. Doyle, D. J. Casper, C. Cicala, S. A. Leavitt, S. Majeed, T. D. Steenbeke, M. Venturi, I. Chaiken, M. Fung, H. Katinger, P. W. Parren, J. Robinson, D. Van Ryk, L. Wang, D. R. Burton, E. Freire, R. Wyatt, J. Sodroski, W. A. Hendrickson, and J. Arthos. 2002. HIV-1 evades antibody-mediated neutralization through conformational masking of receptor-binding sites. Nature 420:678-682.
Kwong, P. D., R. Wyatt, E. Desjardins, J. Robinson, J. S. Culp, B. D. Hellmig, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 1999. Probability analysis of variational crystallization and its application to gp120, the exterior envelope glycoprotein of type 1 human immunodeficiency virus (HIV-1). J. Biol. Chem. 274:4115-4123.
Kwong, P. D., R. Wyatt, S. Majeed, J. Robinson, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 2000. Structures of HIV-1 gp120 envelope glycoproteins from laboratory-adapted and primary isolates. Struct. Fold Des. 8:1329-1339.
Kwong, P. D., R. Wyatt, J. Robinson, R. W. Sweet, J. Sodroski, and W. A. Hendrickson. 1998. Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature 393:648-659.
Kwong, P. D., R. Wyatt, Q. J. Sattentau, J. Sodroski, and W. A. Hendrickson. 2000. Oligomeric modeling and electrostatic analysis of the gp120 envelope glycoprotein of human immunodeficiency virus. J. Virol. 74:1961-1972.
Laskowski, R. A., M. W. MacArthur, D. S. Moss, and J. M. Thornton. 1993. PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26:283-291.
Lea, S., and D. Stuart. 1995. Analysis of antigenic surfaces of proteins. FASEB J. 9:87-93.
Li, A., H. Katinger, M. R. Posner, L. Cavacini, S. Zolla-Pazner, M. K. Gorny, J. Sodroski, T. C. Chou, T. W. Baba, and R. M. Ruprecht. 1998. Synergistic neutralization of simian-human immunodeficiency virus SHIV-Vpu + by triple and quadruple combinations of human monoclonal antibodies and high-titer anti-human immunodeficiency virus type 1 immunoglobulins. J. Virol. 72:3235-3240.
Litwin, V., K. A. Nagashima, A. M. Ryder, C. H. Chang, J. M. Carver, W. C. Olson, M. Alizon, K. W. Hasel, P. J. Maddon, and G. P. Allaway. 1996. Human immunodeficiency virus type 1 membrane fusion mediated by a laboratory-adapted strain and a primary isolate analyzed by resonance energy transfer. J. Virol. 70:6437-6441.
Lo Conte, L., C. Chothia, and J. Janin. 1999. The atomic structure of protein-protein recognition sites. J. Mol. Biol. 285:2177-2198.
MacCallum, R. M., A. C. Martin, and J. M. Thornton. 1996. Antibody-antigen interactions: contact analysis and binding site topography. J. Mol. Biol. 262:732-745.
Martin, A. C. 1996. Accessing the Kabat antibody sequence database by computer. Proteins 25:130-133.
McDonald, I., and J. M. Thornton. Atlas of side-chain and main-chain hydrogen bonding. [Online.] http://www.biochem.ucl.ac.uk/bsm/atlas/.
Mian, I. S., A. R. Bradwell, and A. J. Olson. 1991. Structure, function and properties of antibody binding sites. J. Mol. Biol. 217:133-151.
Moore, J. P., F. E. McCutchan, S. W. Poon, J. Mascola, J. Liu, Y. Cao, and D. D. Ho. 1994. Exploration of antigenic variation in gp120 from clades A through F of human immunodeficiency virus type 1 by using monoclonal antibodies. J. Virol. 68:8350-8364.
Moore, J. P., J. A. McKeating, R. A. Weiss, and Q. J. Sattentau. 1990. Dissociation of gp120 from HIV-1 virions induced by soluble CD4. Science 250:1139-1142.
Moran, M. J., J. S. Andris, Y. Matsumato, J. D. Capra, and E. M. Hersh. 1993. Variable region genes of anti-HIV human monoclonal antibodies: non-restricted use of the V gene repertoire and extensive somatic mutation. Mol. Immunol. 30:1543-1551.
Myszka, D. G., R. W. Sweet, P. Hensley, M. Brigham-Burke, P. D. Kwong, W. A. Hendrickson, R. Wyatt, J. Sodroski, and M. L. Doyle. 2000. Energetics of the HIV gp120-CD4 binding reaction. Proc. Natl. Acad. Sci. USA 97:9026-9031.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode, p. 307-326. In C. Carter and R. Sweet (ed.), Macromolecular crystallography, part A, vol. 276. Academic Press, San Diego, Calif.
Padlan, E. A. 1994. Anatomy of the antibody molecule. Mol. Immunol. 31:169-217.
Padlan, E. A. 1990. On the nature of antibody combining sites: unusual structural features that may confer on these sites an enhanced capacity for binding ligands. Proteins 7:112-124.
Pantophlet, R., E. Ollmann Saphire, P. Poignard, P. W. Parren, I. A. Wilson, and D. R. Burton. 2003. Fine mapping of the interaction of neutralizing and nonneutralizing monoclonal antibodies with the CD4 binding site of human immunodeficiency virus type 1 gp120. J. Virol. 77:642-658.
Pantophlet, R., I. A. Wilson, and D. R. Burton. 2004. Improved design of an antigen with enhanced specificity for the broadly HIV-neutralizing antibody b12. Protein Eng. Des. Sel. 17:749-758.
Parren, P. W., I. Mondor, D. Naniche, H. J. Ditzel, P. J. Klasse, D. R. Burton, and Q. J. Sattentau. 1998. Neutralization of human immunodeficiency virus type 1 by antibody to gp120 is determined primarily by occupancy of sites on the virion irrespective of epitope specificity. J. Virol. 72:3512-3519.
Posner, M. R., L. A. Cavacini, C. L. Emes, J. Power, and R. Byrn. 1993. Neutralization of HIV-1 by F105, a human monoclonal antibody to the CD4 binding site of gp120. J. Acquir. Immune Defic. Syndr. 6:7-14.
Posner, M. R., T. Hideshima, T. Cannon, M. Mukherjee, K. H. Mayer, and R. A. Byrn. 1991. An IgG human monoclonal antibody that reacts with HIV-1/gp120, inhibits virus binding to cells, and neutralizes infection. J. Immunol. 146:4325-4332.
Raja, A., M. Venturi, P. Kwong, and J. Sodroski. 2003. CD4 binding site antibodies inhibit human immunodeficiency virus gp120 envelope glycoprotein interaction with CCR5. J. Virol. 77:713-718.
Read, R. J. 1997. Model phases: probabilities and bias. Methods Enzymol. 277:110-128.
Saphire, E. O., P. W. Parren, C. F. Barbas III, D. R. Burton, and I. A. Wilson. 2001. Crystallization and preliminary structure determination of an intact human immunoglobulin, b12: an antibody that broadly neutralizes primary isolates of HIV-1. Acta Crystallogr. D 57:168-171.
Saphire, E. O., P. W. Parren, R. Pantophlet, M. B. Zwick, G. M. Morris, P. M. Rudd, R. A. Dwek, R. L. Stanfield, D. R. Burton, and I. A. Wilson. 2001. Crystal structure of a neutralizing human IgG against HIV-1: a template for vaccine design. Science 293:1155-1159.
Saphire, E. O., R. L. Stanfield, M. D. Crispin, G. Morris, M. B. Zwick, R. A. Pantophlet, P. W. Parren, P. M. Rudd, R. A. Dwek, D. R. Burton, and I. A. Wilson. 2003. Crystal structure of an intact human IgG: antibody asymmetry, flexibility, and a guide for HIV-1 vaccine design. Adv. Exp. Med. Biol. 535:55-66.
Sattentau, Q. J., and J. P. Moore. 1991. Conformational changes induced in the human immunodeficiency virus envelope glycoprotein by soluble CD4 binding. J. Exp. Med. 174:407-415.
Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. Klenk. 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150:76-85.
Stanfield, R. L., M. K. Gorny, C. Williams, S. Zolla-Pazner, and I. A. Wilson. 2004. Structural rationale for the broad neutralization of HIV-1 by human monoclonal antibody 447-52D. Structure (Cambridge) 12:193-204.
Sullivan, N., Y. Sun, J. Li, W. Hofmann, and J. Sodroski. 1995. Replicative function and neutralization sensitivity of envelope glycoproteins from primary and T-cell line-passaged human immunodeficiency virus type 1 isolates. J. Virol. 69:4413-4422.
Thali, M., C. Furman, D. D. Ho, J. Robinson, S. Tilley, A. Pinter, and J. Sodroski. 1992. Discontinuous, conserved neutralization epitopes overlapping the CD4-binding region of human immunodeficiency virus type 1 gp120 envelope glycoprotein. J. Virol. 66:5635-5641.
Thali, M., J. P. Moore, C. Furman, M. Charles, D. D. Ho, J. Robinson, and J. Sodroski. 1993. Characterization of conserved human immunodeficiency virus type 1 gp120 neutralization epitopes exposed upon gp120-CD4 binding. J. Virol. 67:3978-3988.
Thali, M., U. Olshevsky, C. Furman, D. Gabuzda, M. Posner, and J. Sodroski. 1991. Characterization of a discontinuous human immunodeficiency virus type 1 gp120 epitope recognized by a broadly reactive neutralizing human monoclonal antibody. J. Virol. 65:6188-6193.
Trkola, A., M. Purtscher, T. Muster, C. Ballaun, A. Buchacher, N. Sullivan, K. Srinivasan, J. Sodroski, J. P. Moore, and H. Katinger. 1996. Human monoclonal antibody 2G12 defines a distinctive neutralization epitope on the gp120 glycoprotein of human immunodeficiency virus type 1. J. Virol. 70:1100-1108.
Wei, X., J. M. Decker, S. Wang, H. Hui, J. C. Kappes, X. Wu, J. F. Salazar-Gonzalez, M. G. Salazar, J. M. Kilby, M. S. Saag, N. L. Komarova, M. A. Nowak, B. H. Hahn, P. D. Kwong, and G. M. Shaw. 2003. Antibody neutralization and escape by HIV-1. Nature 422:307-312.
Wu, L., N. P. Gerard, R. Wyatt, H. Choe, C. Parolin, N. Ruffing, A. Borsetti, A. A. Cardoso, E. Desjardin, W. Newman, C. Gerard, and J. Sodroski. 1996. CD4-induced interaction of primary HIV-1 gp120 glycoproteins with the chemokine receptor CCR-5. Nature 384:179-183.
Wyatt, R., P. D. Kwong, E. Desjardins, R. W. Sweet, J. Robinson, W. A. Hendrickson, and J. G. Sodroski. 1998. The antigenic structure of the HIV gp120 envelope glycoprotein. Nature 393:705-711.
Wyatt, R., J. Moore, M. Accola, E. Desjardin, J. Robinson, and J. Sodroski. 1995. Involvement of the V1/V2 variable loop structure in the exposure of human immunodeficiency virus type 1 gp120 epitopes induced by receptor binding. J. Virol. 69:5723-5733.
Wyatt, R., and J. Sodroski. 1998. The HIV-1 envelope glycoproteins: fusogens, antigens, and immunogens. Science 280:1884-1888.
Wyatt, R., N. Sullivan, M. Thali, H. Repke, D. Ho, J. Robinson, M. Posner, and J. Sodroski. 1993. Functional and immunologic characterization of human immunodeficiency virus type 1 envelope glycoproteins containing deletions of the major variable regions. J. Virol. 67:4557-4565.
Xiang, S. H., P. D. Kwong, R. Gupta, C. D. Rizzuto, D. J. Casper, R. Wyatt, L. Wang, W. A. Hendrickson, M. L. Doyle, and J. Sodroski. 2002. Mutagenic stabilization and/or disruption of a CD4-bound state reveals distinct conformations of the human immunodeficiency virus type 1 gp120 envelope glycoprotein. J. Virol. 76:9888-9899.
Xu, W., R. Hofmann-Lehmann, H. M. McClure, and R. M. Ruprecht. 2002. Passive immunization with human neutralizing monoclonal antibodies: correlates of protective immunity against HIV. Vaccine 20:1956-1960.
Zemlin, M., M. Klinger, J. Link, C. Zemlin, K. Bauer, J. A. Engler, H. W. Schroeder, Jr., and P. M. Kirkham. 2003. Expressed murine and human CDR-H3 intervals of equal length exhibit distinct repertoires that differ in their amino acid composition and predicted range of structures. J. Mol. Biol. 334:733-749.
Zhang, W., G. Canziani, C. Plugariu, R. Wyatt, J. Sodroski, R. Sweet, P. Kwong, W. Hendrickson, and I. Chaiken. 1999. Conformational changes of gp120 in epitopes near the CCR5 binding site are induced by CD4 and a CD4 miniprotein mimetic. Biochemistry 38:9405-9416.
Zwick, M. B., R. Kelleher, R. Jensen, A. F. Labrijn, M. Wang, G. V. Quinnan, Jr., P. W. Parren, and D. R. Burton. 2003. A novel human antibody against human immunodeficiency virus type 1 gp120 is V1, V2, and V3 loop dependent and helps delimit the epitope of the broadly neutralizing antibody immunoglobulin G1 b12. J. Virol. 77:6965-6978.
Zwick, M. B., H. K. Komori, R. L. Stanfield, S. Church, M. Wang, P. W. Parren, R. Kunert, H. Katinger, I. A. Wilson, and D. R. Burton. 2004. The long third complementarity-determining region of the heavy chain is important in the activity of the broadly neutralizing anti-human immunodeficiency virus type 1 antibody 2F5. J. Virol. 78:3155-3161.
Zwick, M. B., P. W. Parren, E. O. Saphire, S. Church, M. Wang, J. K. Scott, P. E. Dawson, I. A. Wilson, and D. R. Burton. 2003. Molecular features of the broadly neutralizing immunoglobulin G1 b12 required for recognition of human immunodeficiency virus type 1 gp120. J. Virol. 77:5863-5876.(Royce A. Wilkinson, Chayn)