Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 May 10.
Published in final edited form as: Structure. 2011 May 11;19(5):691–699. doi: 10.1016/j.str.2011.02.012

Structural Analysis of Human and Macaque Monoclonal Antibodies 2909 and 2.5B: Implications for the Configuration of the Quaternary Neutralizing Epitope of HIV-1 gp120

Brett Spurrier 1, Jared M Sampson 1, Maxim Totrov 2, Huiguang Li 1, Timothy O'Neal 3, Constance Williams 3, James Robinson 4, Miroslaw K Gorny 3, Susan Zolla-Pazner 3,5, Xiang-Peng Kong 1,*
PMCID: PMC3096878  NIHMSID: NIHMS284718  PMID: 21565703

Summary

The quaternary neutralizing epitope (QNE) of HIV-1 gp120 is preferentially expressed on the trimeric envelope spikes of intact HIV virions, and QNE-specific monoclonal antibodies (mAbs) potently neutralize HIV-1. Here we present the crystal structures of the Fabs of human mAb 2909 and macaque mAb 2.5B. Both mAbs have long beta hairpin CDR H3 regions >20Å in length that are each situated at the center of their respective antigen-binding sites. Computational analysis showed that the paratopes include the whole CDR H3, while additional CDR residues form shallow binding pockets. Structural modeling suggests a way to understand the configuration of QNEs and the antigen antibody interaction for QNE mAbs. Our data will be useful in designing immunogens that may elicit potent neutralizing QNE Abs.

Keywords: HIV/AIDS, gp120, quaternary neutralizing epitope (QNE), monoclonal antibody (mAb), crystal structure, immunogen design, vaccine

Introduction

Many monoclonal antibodies (mAbs) specific for the envelope (Env) glycoproteins of HIV-1 have been isolated that are able to neutralize virus infectivity. Yet none has displayed neutralization potencies comparable to the sub-nanogram neutralizing activity of a recently identified family of mAbs isolated from HIV-infected humans and SHIV-infected macaques (Gorny et al., 2005; Walker et al., 2009; Robinson et al., 2010). These mAbs have all been selected with neutralization assays, and each targets an epitope selectively displayed on the surface of intact virus particles and infected cells. They react weakly, or not at all, with monomeric forms of gp120, and appear to react preferentially with the gp120 trimeric form of the mature Env spike; consequently, these mAbs are thought to react with a complex region on the Env spike referred to as the “quaternary neutralizing epitope” (QNE) (Gorny et al., 2005).

Antibodies to the QNE were first suggested to be present in SHIV89.6P-infected macaques. Though difficult to characterize in polyclonal sera, their activity appeared to map to a discontinuous epitope formed by the second and third variable regions (V2 and V3) on the virus' trimeric Env spikes (Etemad-Moghadam et al., 1998). Similar Abs mapping to variable regions in gp120 were described in an HIV-infected chimpanzee (Cho et al., 2000; Chen et al., 2001). Polyclonal Abs from these animals displayed extreme neutralizing potency for the viruses infecting the donors. The first QNE-specific human mAb, 2909, was isolated from peripheral blood cells of a clade B HIV-infected long-term non-progressor on the basis of its potent neutralization of HIVSF162; the epitope of this mAb was mapped to portions of the V2 and V3 regions of gp120 (Gorny et al., 2005). While targeting several residues in V2 and V3, the narrow neutralizing capacity of mAb 2909 for HIVSF162 maps primarily to a single residue in V2 at residue 160 (HXB2 numbering (Ratner et al., 1987) where the common glycosylated asparagine is replaced by lysine, an unusual substitution at this position (Honnen et al., 2007). After the description of mAb 2909, two clonally related broadly neutralizing human mAbs, PG9 and PG16, were isolated from a single clade A-infected volunteer on the basis of their ability to neutralize HIVJR-CSF; again, their epitopes were mapped to regions in V2 and V3, but the reactivity of these mAbs was much broader (Walker et al., 2009). More recently, several macaque QNE mAbs were isolated from three macaques infected with SHIVSF162P4 (Robinson et al., 2010). These mAbs again target the QNE composed of V2 and V3; they are narrow in their neutralizing activity and display a variety of patterns responsible for epitope–mAb interaction.

A common factor of these QNE-specific mAbs is their recognition of both V2 and V3 regions, which are highly variable portions of gp120 known to play critical roles in epitope masking and co-receptor binding, respectively. Each QNE-specific mAb appears to target a unique but related antigenic determinant (epitope), and as such, these mAbs are likely to belong to a single family (Walker et al., 2010). There are other Ab families found to have reactivity to distinct but overlapping epitopes, such as the class of anti-V3 and the Abs specific for cluster I and for cluster II of gp41 (Xu et al., 1991; Jiang et al., 2010a).

Due to the dependence of the QNE on the formation of gp120 trimer, crystallization of the antigen-antibody complexes of QNE mAbs is extremely challenging, if not impossible, with the currently available tools. The crystal structures of the uncomplexed form of the Fab fragments of PG16 and of the light chain of PG9 have recently been described (Pancera et al., 2010; Pejchal et al., 2010), but no structural information about either the QNE itself or how it interacts with these mAbs has been forthcoming. Here we describe the crystallization of the Fab fragments of two QNE mAbs, human mAb 2909 and rhesus mAb 2.5B. Our structural analysis and computational modeling suggest a structural model for the contours of QNE and a rational hypothesis for how these QNE mAbs interact with their epitopes. The results also have implications pertaining to the epitope masking of V3 and CD4-induced conformational changes.

Results

Structure determination of Fab fragments of mAbs 2909 and 2.5B

The antigen binding fragments of mAbs 2909 (Fab 2909) and 2.5B (Fab 2.5B) were crystallized, and their structures determined and refined to resolutions of 3.2 Å and 1.9 Å, respectively (Table 1 and Figure1). Fab 2909 crystallized in the trigonal space group P3221 with six Fabs in the asymmetric unit (ASU), while Fab 2.5B crystallized in the monoclinic space group P21 containing two Fabs in the ASU. The structure of Fab 2909, though at a relatively low 3.2 Å resolution, has all residues of its complementarity determining regions (CDRs) accountable in the electron densities for all six Fab complexes in the ASU. On the other hand, the 1.9 Å resolution structure of Fab 2.5B has six residues (a.a. sequence EDDFGD) at the tip of its CDR H3 with poor electron densities, indicating flexibility of the apex of its H3. The densities of these six residues of one of the Fab (Fab1) were only traceable at the 0.7 σ level, while that for the other Fab (Fab2) were completely missing; we have therefore left these residues out for Fab2 in the final model. Non-crystallographic symmetry (NCS) restraints were applied in the refinement of the low-resolution 2909 structure, but three of the six Fabs in the crystal have slightly better electron densities; nevertheless all six Fabs have essentially identical structures. We therefore, henceforth, refer to only one of the 2909 Fabs. Similarly, the two Fabs in the 2.5B crystal are overall the same except for the side chains of several key residues at, possibly, a part of the antigen-binding site (see below). For clarity, we named these two Fabs Fab1 and Fab2 of 2.5B. Two Tyr residues (YH100a and YH100c; in the Kabat numbering scheme (Kabat et al., 1991), and we use L and H in the superscripts to refer residues in the light and heavy chains, respectively) at the distal end of mAb 2909 CDR H3 were shown recently to be sulfated (Changela et al., 2010), and we were able to fit the sulfate group of these residues although three of the six YH100c in the crystal did not have good side chain densities. We have shown here by mass spectrometry (Figure S1) that TyrH100i, located near the apex of its CDR H3, of mAb 2.5B is also sulfated. However, we were not able to place its sulfate group in the structure without generating much negative density in the refinement. The structures of Fabs 2909 and 2.5B are highly similar, and the RMSD of Cα superimposition for the variable regions (Fv) is 0.71 Å (for 100% of the residues) for the light chains, and 1.40 Å (for 99% of the residues) for the heavy chains; hence they superimpose well with the exception of the distal ends of CDR H3 regions (Figure1E). The structural similarity may reflect their similar Ig gene usages (Figure S1); both mAbs, though one is derived from a human patient and the other from a rhesus monkey, use the same VL3-21 for their light chains.

Table 1. Crystallographic Data Collection and Refinement Statistics.

Fab 2909 Fab 2.5B
Data collection
Space group P3221 P21
Cell dimensions
 a, b, c (Å) 178.50, 178.50, 218.90 77.83, 72.82, 88.90
 α, β, γ (°) 90.0, 90.0, 120.0 90.0, 115.64, 90.0
Resolution (Å) 46.6-3.2 (3.31-3.2) 29.7-1.9 (1.97-1.90)
Rmerge 0.129 (0.695) 0.085 (0.599)
I / σI 15.5 (1.8) 17.9 (2.0)
Completeness (%) 99.9 (99.1) 98.2 (92.8)
Redundancy 7.2 (5.4) 3.9 (3.1)
Refinement
Resolution (Å) 3.2 1.9
No. reflections/test set 66876/3391 68816/3482
Rwork / Rfree 19.9 / 24.3 19.1 / 23.1
No. atoms:
 Protein 20042 6554
 Solvent Not placed 869
B-factors
 Protein  96.04 26.97
 Solvent 34.54
R.m.s. deviations
 Bond lengths (Å) 0.010 0.007
 Bond angles (°) 1.234 1.123
Ramachandran plot (Procheck)
 Favored and allowed regions (%) 99.3 99.7
 Disallowed regions (%) 0.7 0.3
Open in a new tab

Values in parentheses are for the highest resolution shell.

Figure 1. Crystal structures of Fabs 2909 and 2.5B.

Figure 1

Open in a new tab

(A) Ribbon representation of the structure of Fab 2909 in a side view with CDR H3 (red) pointing up. The light chain and the rest of heavy chain are colored cyan and green, respectively. (B) Ribbon representation of the structure of Fab 2.5B. (C) A top view of the CDRs of Fab 2909. The CDRs are labeled and colored as magenta (L1), pink (L2), violet (L3), yellow orange (H1), orange (H2) and red (H3), respectively. (D) A top view of the CDRs of Fab 2.5B. (E) Superimposition of the variable domains of Fabs 2909 (cyan and green) and 2.5B (blue and dark green). Note that the two structures superimposed well except at the CDR H3 tips. See also Figure S1.

Structural analysis of the antigen binding sites of Fabs 2909 and 2.5B

The presumed antigen binding sites of both mAbs 2909 and 2.5B, formed by the six CDRs of the light and heavy chains (Figure 1C and 1D), have very unusual features, the most striking of which is their long, extended CDR H3. The CDR H3 of Fab 2909 is 21 residues in length while that of 2.5B is 25 residues (Figure S1). CDR H3s of both Fabs 2909 and 2.5B form a hairpin structure standing at the center of the CDR region and extend away from the rest of the Fab with a distance greater than 20 Å as measured from the apex Cα atom to its base. The tips of all the CDR H3 loops of both Fabs 2909 and 2.5B have some contacts with neighboring molecules in the crystal packing, however, these contacts are minimal and vary from molecule to molecule. For example, the only tip region of CDR H3 of 2909 that make contacts with neighboring molecules are the side chains of the sulfated YH100c and YH100a, located at the very distal end of CDR H3. Their contact areas ranged 34-59Å2 for YH100c and 11-25Å2 for YH100a, respectively. The largest of these values (59 and 25Å2) account for only ∼28% and ∼12% of the available surface areas of these two particular sulfated Tyr residues. Similarly, two of the tip residues, FH100f and YH100i, of Fab1 of mAb 2.5B are involved in protein-protein contacts with contact areas of 46 Å2 and 37 Å2, again ∼30% of their available surface areas. Thus these contacts are likely to have helped stabilize the CDR H3 loops in the crystal structures, but unlikely to have affected the hairpin nature of these loops. CDR H3 of both mAbs 2909 and 2.5B contain five negatively charged residues, and together with the negative charges of the other CDR loops give the encompassing antigen binding sites an overall negative electrostatic surface. This is similar to all the known structures of human anti-V3 mAbs, whose negatively charged antigen binding sites can bind to the positively charged V3 crown (Stanfield et al., 2004; Stanfield et al., 2006; Bell et al., 2008; Burke et al., 2009; Jiang et al., 2010b). These two CDR H3s are also populated with Tyr residues: CDR H3 of mAb 2909 has six Tyr while that of mAb 2.5B has four; this, again, is similar to two human anti-V3 mAbs 447-52D and 537-10D, both of which also have elongated CDR H3s (Stanfield et al., 2004; Burke et al., 2009). Five Tyr residues (YH100k, YH100i, YH100g and the sulfated YH100a and YH100c) of mAb 2909's CDR H3 line along the hairpin structure in such a way so that their side chains form a configuration of a circular staircase.

As a first step in analyzing the antigen binding site of these two mAbs in order to define their paratopes, the regions that are most likely to be involved in protein–protein contacts were identified using the Optimal Docking Area (ODA) function of ICM (Fernandez-Recio et al., 2005). The ODA algorithm analyzes the physico-chemical characteristics of protein surfaces by computing desolvation energy upon burial of various patches on the protein surface. Contributions to desolvation from aromatic, aliphatic, polar and charged buried atoms are optimized for best recognition of protein-protein interaction sites. Probing spheres associated with the most favorable desolvation score are predicted to indicate putative protein interaction partner locations. This is a fast and accurate method to predict regions of protein-protein interactions; in a test study, this method was used and correctly located 80% of the binding sites for general proteins and 100% for Abs (Fernandez-Recio et al., 2005). As an additional test, we used ODA analysis first with the structures of two anti-V3 mAbs, 447-52D and 537-10D, for which structures of antigen-antibody complexes are available (Stanfield et al., 2004; Burke et al., 2009); ODA accurately predicted the antigen binding sites Figure S2).

Application of the ODA method to the Fab fragments of mAbs 2909 and 2.5B showed that the “hotspots” for protein–protein interactions are almost exclusively concentrated on their CDR H3s (Figure 2). Hence, the apexes of the CDR H3 regions in both anti-QNE mAbs are likely to play central roles in these interactions, and all sides of the CDR H3s are involved in interacting with their epitopes. This property of the QNE mAbs contrasts with that of the anti-V3 mAbs 447-52D and 537-10D, whose antigen-binding sites are located only on one side of their long CDR H3, sufficient for binding the relatively simpler epitope – V3 crown alone (Figure S2).

Figure 2. ICM Optimal Docking Area (ODA) analyses of Fabs 2909 and 2.5B.

Figure 2

Open in a new tab

ODA analysis of Fabs 2909 (A) and 2.5B (B). ODA uses desolvation energy to accurately predict sites of protein-protein interaction, and spheres are placed at the surface of the protein. The redness and size of the spheres are proportional to the region's likelihood to be involved in protein-protein interactions. The results indicate that CDR H3 of both mAbs is predominantly involved in binding their epitopes. See also Figure S2

We further analyzed the antigen binding sites of the mAbs 2909 and 2.5B using a complementary method that determines the existence of binding pockets in regions formed by the CDR loops. For this, we used the PocketFinder function of ICM (An et al., 2005). The PocketFinder algorithm uses protein surface geometry as well as its van der Waals interaction potentials to identify ligand-binding pockets. It was developed for small molecules and had a 96.8% success rate when tested with 5,616 binding sites collected from ligand–protein complexes, and 11,510 apo binding sites inferred from the complexes by homology (An et al., 2005). Since this ICM function was originally developed for small molecular ligands, we tested its ability first in determining the antigen binding pockets on mAbs crystallized in complex with their cognate peptide epitopes. The binding sites for all eight human anti-V3 mAbs with known structures (Stanfield et al., 2004; Stanfield et al., 2006; Bell et al., 2008; Burke et al., 2009; Jiang et al., 2010b) were correctly identified. We show for example, in Figure 3D and Figure S3, the results for mAbs 447-52D and 537-10D.

Figure 3. Possible antigen binding pockets of Fabs 2909 and 2.5B identified by the PocketFinder function of ICM.

Figure 3

Open in a new tab

(A) Two pockets were found for the Fab 2909 antigen binding site (inset: surfaces illustrating pocket 1 (P1) and pocket 2 (P2) are colored blue and yellow, respectively). The side chains of the residues lining the pockets are displayed and labeled, and they are colored the same as the pockets in the inset, except TyrH100i which is involved in both pockets. Arrowheads point to three residues, TrpL91, AspH95a and TyrH100k, that form the aromatic-acidic triad proposed to be involved in binding the arch of V3 (see text). (B) Two pockets are also identified for Fab1 of 2.5B. Again the side chains of residues lining the pockets are displayed. Only two of three residues forming the aromatic-acidic triad (arrowheads), TrpL91 and TyrH100n are involved in lining P1. However, these two residues and the third residue, DH58 (light blue), together with another light chain residue, HL95a (light blue) collectively made a conformational change between Fab1 and Fab2 (see Figure 3C). (C) Two pockets of Fab2 of 2.5B. The aromatic-acidic triad (arrowheads) has slightly different conformation from that of Fab1. (D) Potential paratope residues of Fabs 2009 and 2.5B identified by ICM PocketFinder. The top panel is a comparison of the actual paratope residues (yellow) of mAbs 447-52D (PDB: 3GHB) and 537-10D (PDB: 3GHE) and that predicted by ICM PocketFinder (green). The predicted paratope residues for mAbs 2909 and 2.5B are indicated in the lower panel. CDR residues of light (L1-L3) and heavy (H1-H3) chains are numbered according to the Kabat definitions. Letter insertions following a numeral (e.g. 27a, 27b, and 27c in CDR L1) are indicated by lowercase letters above the corresponding amino acids. See also Figure S3

Visual inspection shows that there are small pockets on the surfaces of Fabs 2909 and 2.5B at surface areas formed by the CDRs (not shown), but they are relatively shallow. We therefore lowered the PocketFinder algorithm tolerance to 3.0 (which gave reasonable results for anti-V3 mAbs) instead of using the default value of 4.6. The PocketFinder function found two small pockets (P1 and P2) at the antigen-binding site of Fab 2909 with volumes of 243 Å3 (P1) and 127 Å3 (P2) (inset, Figure 3A). (In comparison, the 3D volumes for the antigen binding site pockets of mAbs 447-52D and 537-10D are, at the 4.6 tolerance, 458 Å3 and 485 Å3, respectively; Figure S3.) Both pockets of Fab 2909 are located at the base of CDR H3 with the first one formed by residues from both light and heavy chains, while the second pocket is formed by only heavy chain residues (Figures 3A and 3D). Drawing analogies from the structures of mAbs 447-52D and 537-10D (Stanfield et al., 2004; Burke et al., 2009) (Figures 3D and S3), these pockets appear as sites of interaction with the apexes of V3 and V2. The two Fabs in the 2.5B crystal gave slightly different results (Figures 3B, 3C and 3D), derived from the different side chain conformations of several residues: TrpL91, AspH58, TyrH100n, and HisH95a (Figures 3B and 3C, arrowheads). We term residues TrpL91, AspH58 and TyrH100n in 2.5B the aromatic-acidic triad of the antigen binding site; it is also present in Fab 2909 at the same position (Figure 3A, arrowheads). In Fab2 of 2.5B, the side chain of TrpL91 points down towards the interior of the molecule (Figure 3C), while in Fab1, its side chain is flipped upward and pushes away the side chain of HisH95a (Figure 3B). Comparison of the pockets identified by ICM PocketFinder showed that the residues forming these pockets are very similar in position between Fabs 2909 and 2.5B (Figure 3).

Signatures of V3 binding

The epitopes recognized by all QNE mAbs described to date include the V3 region of gp120. Since the structures for a panel of anti-V3 mAbs in complex with V3 peptides have been determined (Stanfield et al., 2004; Stanfield et al., 2006; Bell et al., 2008; Burke et al., 2009; Jiang et al., 2010b), we looked for elements in the CDRs of Fabs 2909 and 2.5B that have characteristics similar to those of anti-V3 mAbs. The structures of Fabs 2909 and 2.5B resemble those of mAbs 447-52D and 537-10D (Burke et al., 2009), having rather extended structures of CDR H3. The Fabs of 2909 and 2.5B superimposed well with those of 447-52D and 537-10D, except that the CDR H3s have different orientations (Figure S2C). The paratopes of both mAbs 447-52D and 537-10D are of the so-called “ladle” type, i.e., their antigen binding sites are shaped like a soup ladle with the CDR H3 forming the handle (Burke et al., 2009) and the N-terminal side of the V3 crown interacting with the CDR H3 by main-chain interactions, while the apex of the V3 crown falls into the bowl of the ladle. The bowl of the antigen binding site has a unique feature allowing it to bind the GPGR arch motif of the V3 crown. In particular, both Fabs 447-52D and 537-10D have an aromatic-acidic triad of Tyr-Trp-Asp/Glu (Figure S3) (Stanfield et al., 2004; Burke et al., 2009). Inspection of the antigen binding sites of Fabs 2909 and 2.5B shows that both also have this aromatic-acidic group at the base of the antigen-binding site (Figure 3). In the case of mAb 2.5B, the side chains of this triad adopted two conformations in Fab1 and Fab2, suggesting that this region is somewhat flexible and may allow an induced fit upon epitope binding.

We then attempted to computationally dock V3 into the antigen binding sites of Fabs 2909 and 2.5B, using a two step docking procedure in ICM – rigid-body docking followed by side-chain optimization (Fernandez-Recio et al., 2002, 2004). We first created a structure of V3 with SF162 sequence by homology modeling based of a V3 structure in the context of gp120 core (PDB: 2B4C). This full-length V3 was then docked to the antigen binding sites of Fabs 2909 and 2.5B and the lowest energy results are presented in Figure 4. The docked complexes of both Fabs 2909 and 2.5B have very similar structural features: the N-terminal side of the V3 crown forms main chain interactions with the long CDR H3 with the GPGR arch of the crown docked at CDR regions of those V3 binding signature residues.

Figure 4. Computational docking of V3 into the antigen binding site of Fabs 2909 and 2.5B.

Figure 4

Open in a new tab

A full-length V3 is docked into the antigen binding sites of Fabs 2909 (A) and 2.5B (B) using a two-step soft docking procedure of ICM (Fernandez-Recio et al., 2002, 2004). The lowest energy complexes are presented here and, for clarity, only selected residues of V3 and the antigen binding sites are shown. The light, heavy chains of the mAbs and V3 are colored cyan, green and magenta, respectively. The sulfated TyrH100i in Fab 2.5B (panel B) was modeled in the structure according to the mass spectrometry data (Figure S1D). Note the docked modes of both Fabs 2909 and 2.5B are similar to each other and to that of V3 mAbs 447-52D and 537-10D (Figures S2A and S2B). See also Figure S4

Discussion

Atomic structures of mAbs specific for HIV-1 gp120 QNE can shed light on how they bind to their cognate epitopes. We have determined the atomic structures of two QNE-specific mAbs, one derived from cells of an HIV-infected human and the other from a SHIVSF162P4- infected rhesus macaque. Our results show that both mAbs have long, extended CDR H3 regions with a straight beta hairpin structure and with shallow binding pockets formed by the base region of CDRH3 and other CDRs (Figures 1-3). ICM ODA analyses of the structures showed that the CDR H3 regions of both mAbs display key protein-protein interaction surfaces (Figure 2). Hence the CDR H3 portions of these mAbs appears to play a dominant role in epitope binding. A complementary computational analysis of the antigen binding sites formed by all six CDRs revealed pockets at the base of the CDR H3 region which are likely to be part of the paratopes of these two mAbs (Figure 3).

Comparison of the antigen binding sites of Fabs 2909 and 2.5B with that of anti-V3 human mAbs indicated that signatures of V3 binding exist at their antigen binding sites. Computational protein-protein docking allowed us to place a V3 into their binding sites in a mode highly similar to that of anti-V3 mAbs 447-52D and 537-10D (Figure 4). However, due to the shallowness of the pockets at their antigen binding sites, mAbs 2909 and 2.5B are not able to bind V3 alone. Our docking results, together with the available gp120 structures derived from X-ray crystallography and cryo-electron microscopy (Huang et al., 2005; Huang et al., 2007; Liu et al., 2008), allowed us to further construct a model of antigen-antibody interactions for 2909/2.5B (Supplementary Data and Figure S4A). This model, while highly speculative, provides several clues of these interactions and suggests a way in configuring the QNEs. First, this model suggests a spatial relationship between the V1/V2 and V3 loops, and it explains how mAbs 2909 and 2.5B can interact simultaneously with both V2 and V3 loops: one side of the CDR H3 binds to V3, the other side binds V2, and the tips of the V2 and V3 loops reaches down into the pockets at the base of CDR H3, agreeing with the ODA predicted regions involved in the antigen-antibody interactions (Figure 2). Second, this model indicates that the QNE occurs within a single gp120 monomer, i.e., it consists of the V1/V2 and V3 loops from a single molecule. This is in agreement with the proposed nature of the epitope recognized by mAbs PG9 and PG16, and with the finding that mAb PG9 can interact, albeit weakly, with gp120 of strain DU422 (Walker et al., 2009; Doores and Burton, 2010). However, due to the flexibility of the gp120 variable loops, it appears that the stable QNE can only form optimally in the gp120 trimeric spike on the surface of virions and infected cells. Third, this model indicates that the tip of the CDR H3 of mAbs 2909 and 2.5B can reach the co-receptor binding site of gp120 and that this site may be part of the QNE. This is where the N-terminus of CCR5 interacts through sulfated Tyr residues (Huang et al., 2007), and the Tyr residues at the apex of the CDR H3 of these mAbs may serve the same purpose. This is supported by that fact that Tyr residues at the apex of their CDR H3 are sulfated Figure S1D) (Changela et al., 2010). Fourth, the spatial relationship between V1/V2 and V3 in our model also suggests a mechanistic view for how V1/V2 might mask V3, and why exposure to CD4 enhances V3 exposure (Figure S4A). In the unbound or b12-bound form, V1/V2 and V3 are located in spatial proximity and may allow the former to mask the latter (Figures S4A). It is known that, upon CD4 binding, the inner domain of gp120 undergoes major conformational changes (Zhou et al., 2007; Liu et al., 2008; White et al., 2010), V3 exposure is enhanced (Mbah et al., 2001), and many HIV strains become more sensitive to anti-V3 mAbs (Wu et al., 2008). This can be visualized as the V1/V2 loop moves away from V3 upon binding to CD4 and the bridging sheet forms as depicted in Figure S4(Zhou et al., 2007; Liu et al., 2008; White et al., 2010).

We also constructed a model for the antigen-antibody interaction for the QNE mAb PG16 (Figure S4), whose Fab crystal structure was published recently (Pancera et al., 2010; Pejchal et al., 2010). These structures showed that the long CDR H3 of PG16 does not form a straight beta hairpin like that of Fabs 2909 and 2.5B presented here. Instead, the CDR H3 apex of PG16 bends sideways forming an additional secondary structure that was termed a “hammerhead” (Pejchal et al., 2010), or an “axe” (Pancera et al., 2010). If this configuration is a stable structure, as the authors argue, and remains in the same configuration upon epitope binding, the CDR H3 of PG16 will have a slightly different way to approach the co-receptor binding site of gp120 (Supplemental Results and Figure S4B). However, the configuration of the QNE for PG16 in our model is essentially the same as that of 2909/2.5B (Figures S4A and S4B). This is supported by recent data that that a K160N mutation in V2 is primarily responsible for limiting the neutralization breadth of mAb 2909, while a reverse mutation will substantially reduce the breadth of PG16 (Dr. X. Wu, personal communications).

The fact that the only mAbs that have been selected to date using neutralization assays are QNE-specific and that these mAbs are all extremely potent and target overlapping epitopes suggests that Abs of this type may be highly desirable to induce with a prophylactic vaccine. It is not yet clear, however, whether these mAbs are rare or commonly made by infected humans (Walker et al., 2010). If they can be found only rarely, it would suggest that virions and native gp120 are not ideal immunogens for the induction of highly potent anti-QNE Abs; moreover, a trimeric gp140-based vaccine would need to assume the correct conformation to allow formation of the QNE, a feat that has yet to be achieved. The weak reactivity of PG9 with monomeric gp120 and trimerized gp140 constructs, and the absence of reactivity with these ligands by PG16 (Walker et al., 2009), indicates that such immunogens need further refinement if they are to be used to induce QNE Abs. A modified monomeric gp120 that stabilizes the V2/V3 spatial relationship, as suggested by our model (Figure S4), or an engineered protein which presents a structural mimotope of the QNE may be better suited to elicit these types of neutralizing Abs. The structural data and models presented here may serve as a first step in the design of such immunogens.

Experimental Procedures

Monoclonal antibodies

Monoclonal Ab 2909 was developed from the cells of an individual chronically infected with clade B HIV-1, using a previously described cellular method (Gorny et al., 2005). Monoclonal Ab 2909 belongs to the IgG1 subclass and bears lambda light chains. Monoclonal Ab 2.5B was generated using a cellular method from a rhesus macaque seven month post-infection with SHIVSF162P4 (Robinson et al., 2010). This mAb is also an IgG1 with lambda light chains. The variable domains of the heavy and light chains of both mAbs were sequenced as previously described (Jeffs et al., 2001; Gorny et al., 2009) and the sequences were analyzed using the International ImMunoGeneTics (IMGT) information system (http://imgt.cines.fr). The immunoglobulin (Ig) gene usage by the mAbs is shown in Figure S1 Identification of tyrosine sulfation of TyrH100i in CDR H3 of 2.5B was performed by MS Bioworks LLC (Ann Arbor, MI). A gel band of Fab 2.5B heavy chain was digested with elastase at 37°C for 18h and quenched with formic acid; the supernatant was analyzed by MS/MS mass spectrometry directly without further processing.

Crystallization and structure determination

Fabs of mAbs 2909 and 2.5B were generated by papain digestion and purified using size exclusion chromatography (Burke et al., 2009). The purified Fabs in 20 mM Tris, pH 6.5, and 100 mM NaCl were concentrated for crystallization. Fab 2909 was crystallized by the micro-batch method under paraffin oil with 0.5 μl Fab at 30 mg/ml mixed with an equal amount of buffer containing 14% PEG 8000, 0.1M HEPES, pH6.8, 0.1M glycine, and 8% ethylene glycol. Fab 2.5B was crystallized by the vapor diffusion hanging drop method with 0.5 μl of Fab at 20 mg/ml mixed with an equal volume of well solution containing 18% PEG 3350 and 100 mM ammonium nitrate, pH 6.3.

Several synchrotron beamlines were used for characterization of the crystals and diffraction data collection, including X4C and X6A at the National Synchrotron Light Source (NSLS) and IMCA-CAT, and GM/CA-CAT at the Advanced Photo Source (APS). The final data sets for both Fabs were collected at GM/CA-CAT. The structures were determined by molecular replacement using MOLREP in the CCP4 software package with a starting models built using ICM (Abagyan et al., 1994) on a structure of the homologous Fab-Hyb3 (Hulsmeyer et al., 2005) (PDB: 1W72), and refined by cycles of manual adjustments and computational minimizations using Coot {Emsley, 2004 #1051}, O {Jones, 2004 #718}, CNS {Brunger, 1998 #601}, Refmac {Murshudov, 1997 #1052}, and Phenix (version 1.6.4) {Adams, 2002 #1069}. NCS restraints were applied by Phenix in the refinement of the 3.2 Å resolution structure of Fab 2909. Due the elbow differences (up to 15°; not shown) of the 6 Fabs in the asymmetric unit, the restraints were applied separately for the variable and constant regions of each chain. However, the restraints were not imposed on the side chains of CDR H3 due to the different packing environments of the tip residues. Rfree was calculated for both structures with 5% of randomly selected reflections, which distributed about equally in all resolution shells (not shown). Structural analysis and modeling were carried out using ICM and the figures were created using ICM and Pymol (DeLano, 2002).

Computational analysis of the antigen binding sites (paratopes) of Fabs 2909 and 2.5B

The ICM Optical Docking Area (ODA) function was used to identify regions of the Fabs likely involved in antigen-antibody interactions of Fabs 2909 and 2.5B. The algorithm identifies surface patches with optimal docking desolvation energy based on atomic solvation parameters adjusted for protein–protein interactions (Fernandez-Recio et al., 2005). It has been benchmarked on a panel of non-homologous proteins involved in non-obligate protein–protein interactions. The ICM PocketFinder function was used to identify the potential binding pockets on the surfaces of Fabs 2909 and 2.5B. The algorithm builds a grid map of the van der Waals potential, and the position and size of the antigen binding pocket(s) are determined based on the construction of equipotential surfaces after a smoothing transformation of the map that emphasizes buried continuous regions of low potential (An et al., 2005). A tolerance parameter of 3.0 was used for mAbs 2909 and 2.5B instead of the default value, which is 4.6; a lower value results in a greater sensitivity.

A two-step soft protein-protein docking procedure of ICM was used to automatically dock the V3 model into the antigen binding sites of Fabs 2909 and 2.5B (Fernandez-Recio et al., 2002, 2004). The V3 model was first docked as a rigid body to the antigen binding sites. Then, a modified Monte-Carlo docking procedure was used to refine the docking position and conformation of the V3 model as well as the interacting residues of the Fabs, allowing for marginal flexibility in the backbone torsion angle and full side chain flexibility. Each iteration of the Monte-Carlo procedure was followed by a pseudo-Brownian movement of the V3 position and a local energy minimization. The iterative process was repeated several million times until a minimum energy was converged upon.

Accession Numbers

The atomic coordinates of both structures have been deposited in RCSB Protein Data Bank with PDB IDs 3Q6F for Fab 2909 and 3Q6G for Fab 2.5B.

Supplementary Material

01

Acknowledgments

We thank Nick Cowan for critically reading the manuscript, and staff members at synchrotron beamlines X6A at NSLS, and IMCA-CAT and GM/CA-CAT at APS for help with X-ray diffraction data collections. This work was supported by an NIH supplementary grant (X.P.K.) as part of the New York University Center for AIDS Research (CFAR, AI027742), by NIH grants AI084119, AI082274 (X.P.K.), AI036085, HL059725 (S.Z.P.) and AI077451 (M.K.G.), by grants from the Bill & Melinda Gates Foundation (Zolla-Pazner & Stamatatos CAVD VDCs), and by research funds from the Department of Veterans Affairs. B.S. is supported by NIH training grant GM088118.

Footnotes

Supplemental Data include computational modeling of QNE and four figures can be found with this article online at http://www.cell.com/structure/supplemental/xxx.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Abagyan RA, Totrov M, Kuznetsov D. ICM - A new method for protein modeling and design: Applications to docking and structure prediction from the distorted native conformation. Journal of Computational Chemistry. 1994;15:488–506. [Google Scholar]
  2. Adams PD, Grosse-Kunstleve RW, Hung LW, Ioerger TR, McCoy AJ, Moriarty NW, Read RJ, Sacchettini JC, Sauter NK, Terwilliger TC. PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr D Biol Crystallogr. 2002;58:1948–1954. doi: 10.1107/s0907444902016657. [DOI] [PubMed] [Google Scholar]
  3. An J, Totrov M, Abagyan R. Pocketome via comprehensive identification and classification of ligand binding envelopes. Mol Cell Proteomics. 2005;4:752–761. doi: 10.1074/mcp.M400159-MCP200. [DOI] [PubMed] [Google Scholar]
  4. Bell CH, Pantophlet R, Schiefner A, Cavacini LA, Stanfield RL, Burton DR, Wilson IA. Structure of antibody F425-B4e8 in complex with a V3 peptide reveals a new binding mode for HIV-1 neutralization. J Mol Biol. 2008;375:969–978. doi: 10.1016/j.jmb.2007.11.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr. 1998;54:905–921. doi: 10.1107/s0907444998003254. [DOI] [PubMed] [Google Scholar]
  6. Burke V, Williams C, Sukumaran M, Kim S, Li H, Wang XH, Gorny MK, Zolla-Pazner S, Kong XP. Structural Basis of the Cross-Reactivity of Genetically Related Human Anti-HIV-1 mAbs: Implications for Design of V3-Based Immunogens. Structure. 2009;17:1538–1546. doi: 10.1016/j.str.2009.09.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Changela A, Wu X, Yang Y, Zhang B, Zhu J, Nardone GA, O'Dell S, Pancera M, Gorny MK, Phogat S, et al. Crystal structure of human antibody 2909 reveals conserved features of quaternary-specific antibodies that potently neutralize HIV-1. J Virol. 2010 doi: 10.1128/JVI.02335-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen CH, Jin L, Zhu C, Holz-Smith S, Matthews TJ. Induction and characterization of neutralizing antibodies against a human immunodeficiency virus type 1 primary isolate. J Virol. 2001;75:6700–6704. doi: 10.1128/JVI.75.14.6700-6704.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cho MW, Lee MK, Chen CH, Matthews T, Martin MA. Identification of gp120 regions targeted by a highly potent neutralizing antiserum elicited in a chimpanzee inoculated with a primary human immunodeficiency virus type 1 isolate. J Virol. 2000;74:9749–9754. doi: 10.1128/jvi.74.20.9749-9754.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. DeLano WL. The PyMOL User's Manual. Palo Alto: DeLano Scientific; 2002. [Google Scholar]
  11. Doores KJ, Burton DR. Variable loop glycan dependency of the broad and potent HIV-1 neutralizing antibodies PG9 and PG16. J Virol. 2010 doi: 10.1128/JVI.00552-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Emsley P, Cowtan K. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr. 2004;60:2126–2132. doi: 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  13. Etemad-Moghadam B, Karlsson GB, Halloran M, Sun Y, Schenten D, Fernandes M, Letvin NL, Sodroski J. Characterization of simian-human immunodeficiency virus envelope glycoprotein epitopes recognized by neutralizing antibodies from infected monkeys. J Virol. 1998;72:8437–8445. doi: 10.1128/jvi.72.10.8437-8445.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fernandez-Recio J, Totrov M, Abagyan R. Soft protein-protein docking in internal coordinates. Protein Sci. 2002;11:280–291. doi: 10.1110/ps.19202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fernandez-Recio J, Totrov M, Abagyan R. Identification of protein-protein interaction sites from docking energy landscapes. J Mol Biol. 2004;335:843–865. doi: 10.1016/j.jmb.2003.10.069. [DOI] [PubMed] [Google Scholar]
  16. Fernandez-Recio J, Totrov M, Skorodumov C, Abagyan R. Optimal docking area: a new method for predicting protein-protein interaction sites. Proteins. 2005;58:134–143. doi: 10.1002/prot.20285. [DOI] [PubMed] [Google Scholar]
  17. Gorny MK, Stamatatos L, Volsky B, Revesz K, Williams C, Wang XH, Cohen S, Staudinger R, Zolla-Pazner S. Identification of a new quaternary neutralizing epitope on human immunodeficiency virus type 1 virus particles. J Virol. 2005;79:5232–5237. doi: 10.1128/JVI.79.8.5232-5237.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gorny MK, Wang XH, Williams C, Volsky B, Revesz K, Witover B, Burda S, Urbanski M, Nyambi P, Krachmarov C, et al. Preferential use of the VH5-51 gene segment by the human immune response to code for antibodies against the V3 domain of HIV-1. Mol Immunol. 2009;46:917–926. doi: 10.1016/j.molimm.2008.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Honnen WJ, Krachmarov C, Kayman SC, Gorny MK, Zolla-Pazner S, Pinter A. Type-specific epitopes targeted by monoclonal antibodies with exceptionally potent neutralizing activities for selected strains of human immunodeficiency virus type 1 map to a common region of the V2 domain of gp120 and differ only at single positions from the clade B consensus sequence. J Virol. 2007;81:1424–1432. doi: 10.1128/JVI.02054-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Huang CC, Lam SN, Acharya P, Tang M, Xiang SH, Hussan SS, Stanfield RL, Robinson J, Sodroski J, Wilson IA, et al. Structures of the CCR5 N terminus and of a tyrosine-sulfated antibody with HIV-1 gp120 and CD4. Science. 2007;317:1930–1934. doi: 10.1126/science.1145373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Huang CC, Tang M, Zhang MY, Majeed S, Montabana E, Stanfield RL, Dimitrov DS, Korber B, Sodroski J, Wilson IA, et al. Structure of a V3-containing HIV-1 gp120 core. Science. 2005;310:1025–1028. doi: 10.1126/science.1118398. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hulsmeyer M, Chames P, Hillig RC, Stanfield RL, Held G, Coulie PG, Alings C, Wille G, Saenger W, Uchanska-Ziegler B, et al. A major histocompatibility complex-peptide-restricted antibody and t cell receptor molecules recognize their target by distinct binding modes: crystal structure of human leukocyte antigen (HLA)-A1-MAGE-A1 in complex with FAB-HYB3. J Biol Chem. 2005;280:2972–2980. doi: 10.1074/jbc.M411323200. [DOI] [PubMed] [Google Scholar]
  23. Jeffs SA, Gorny MK, Williams C, Revesz K, Volsky B, Burda S, Wang XH, Bandres J, Zolla-Pazner S, Holmes H. Characterization of human monoclonal antibodies selected with a hypervariable loop-deleted recombinant HIV-1(IIIB) gp120. Immunol Lett. 2001;79:209–213. doi: 10.1016/s0165-2478(01)00289-9. [DOI] [PubMed] [Google Scholar]
  24. Jiang X, Burke V, Totrov M, Williams C, Cardozo T, Gorny MK, Zolla-Pazner S, Kong XP. Conserved structural elements in the V3 crown of HIV-1 gp120. Nat Struct Mol Biol. 2010a;17:955–961. doi: 10.1038/nsmb.1861. [DOI] [PubMed] [Google Scholar]
  25. Jiang X, Burke V, Totrov M, Williams C, Cardozo T, Gorny MK, Zolla-Pazner S, Kong XP. Conserved structural elements in the V3 crown of HIV-1 gp120. Nat Struct Mol Biol. 2010b doi: 10.1038/nsmb.1861. [DOI] [PubMed] [Google Scholar]
  26. Jones TA. Interactive electron-density map interpretation: from INTER to O. Acta Crystallogr D Biol Crystallogr. 2004;60:2115–2125. doi: 10.1107/S0907444904023509. [DOI] [PubMed] [Google Scholar]
  27. Kabat EA, Wu TT, Perry HM, Gottesman KS. Sequences of Proteins of Immunological Interest. 5th. Bethesda, MD: National Institutes of Health; 1991. [Google Scholar]
  28. Liu J, Bartesaghi A, Borgnia MJ, Sapiro G, Subramaniam S. Molecular architecture of native HIV-1 gp120 trimers. Nature. 2008;455:109–113. doi: 10.1038/nature07159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mbah HA, Burda S, Gorny MK, Williams C, Revesz K, Zolla-Pazner S, Nyambi PN. Effect of soluble CD4 on exposure of epitopes on primary, intact, native human immunodeficiency virus type 1 virions of different genetic clades. J Virol. 2001;75:7785–7788. doi: 10.1128/JVI.75.16.7785-7788.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Murshudov GN, Vagin AA, Dodson EJ. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr. 1997;53:240–255. doi: 10.1107/S0907444996012255. [DOI] [PubMed] [Google Scholar]
  31. Pancera M, McLellan JS, Wu X, Zhu J, Changela A, Schmidt SD, Yang Y, Zhou T, Phogat S, Mascola JR, Kwong PD. Crystal structure of PG16 and chimeric dissection with somatically related PG9: Structure-function analysis of two quaternary-specific antibodies that effectively neutralize HIV-1. J Virol. 2010 doi: 10.1128/JVI.00966-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pejchal R, Walker LM, Stanfield RL, Phogat SK, Koff WC, Poignard P, Burton DR, Wilson IA. Structure and function of broadly reactive antibody PG16 reveal an H3 subdomain that mediates potent neutralization of HIV-1. Proc Natl Acad Sci U S A. 2010;107:11483–11488. doi: 10.1073/pnas.1004600107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ratner L, Fisher A, Jagodzinski LL, Mitsuya H, Liou RS, Gallo RC, Wong-Staal F. Complete nucleotide sequences of functional clones of the AIDS virus. AIDS Res Hum Retroviruses. 1987;3:57–69. doi: 10.1089/aid.1987.3.57. [DOI] [PubMed] [Google Scholar]
  34. Robinson JE, Franco K, Elliott DH, Maher MJ, Reyna A, Montefiori DC, Zolla-Pazner S, Gorny MK, Kraft Z, Stamatatos L. Quaternary epitope specificities of anti-HIV-1 neutralizing antibodies generated in rhesus macaques infected by the simian/human immunodeficiency virus SHIVSF162P4. J Virol. 2010;84:3443–3453. doi: 10.1128/JVI.02617-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Stanfield RL, Gorny MK, Williams C, Zolla-Pazner S, Wilson IA. Structural rationale for the broad neutralization of HIV-1 by human monoclonal antibody 447-52D. Structure. 2004;12:193–204. doi: 10.1016/j.str.2004.01.003. [DOI] [PubMed] [Google Scholar]
  36. Stanfield RL, Gorny MK, Zolla-Pazner S, Wilson IA. Crystal structures of human immunodeficiency virus type 1 (HIV-1) neutralizing antibody 2219 in complex with three different V3 peptides reveal a new binding mode for HIV-1 cross-reactivity. J Virol. 2006;80:6093–6105. doi: 10.1128/JVI.00205-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Walker LM, Phogat SK, Chan-Hui PY, Wagner D, Phung P, Goss JL, Wrin T, Simek MD, Fling S, Mitcham JL, et al. Broad and potent neutralizing antibodies from an African donor reveal a new HIV-1 vaccine target. Science. 2009;326:285–289. doi: 10.1126/science.1178746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Walker LM, Simek MD, Priddy F, Gach JS, Wagner D, Zwick MB, Phogat SK, Poignard P, Burton DR. A Limited Number of Antibody Specificities Mediate Broad and Potent Serum Neutralization in Selected HIV-1 Infected Individuals. PLoS Pathog. 2010;6:e1001028. doi: 10.1371/journal.ppat.1001028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. White TA, Bartesaghi A, Borgnia MJ, Meyerson JR, de la Cruz MJV, Bess JW, Nandwani R, Hoxie JA, Lifson JD, Milne JLS, Subramaniam S. Molecular Architectures of Trimeric SIV and HIV-1 Envelope Glycoproteins on Intact Viruses: Strain-Dependent Variation in Quaternary Structure. PLoS Pathog. 2010;6:e1001249. doi: 10.1371/journal.ppat.1001249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Wu X, Sambor A, Nason MC, Yang ZY, Wu L, Zolla-Pazner S, Nabel GJ, Mascola JR. Soluble CD4 broadens neutralization of V3-directed monoclonal antibodies and guinea pig vaccine sera against HIV-1 subtype B and C reference viruses. Virology. 2008;380:285–295. doi: 10.1016/j.virol.2008.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Xu JY, Gorny MK, Palker T, Karwowska S, Zolla-Pazner S. Epitope mapping of two immunodominant domains of gp41, the transmembrane protein of human immunodeficiency virus type 1, using ten human monoclonal antibodies. J Virol. 1991;65:4832–4838. doi: 10.1128/jvi.65.9.4832-4838.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhou T, Xu L, Dey B, Hessell AJ, Van Ryk D, Xiang SH, Yang X, Zhang MY, Zwick MB, Arthos J, et al. Structural definition of a conserved neutralization epitope on HIV-1 gp120. Nature. 2007;445:732–737. doi: 10.1038/nature05580. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS284718-supplement-01.pdf (7.3MB, pdf)

RESOURCES