Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Dec 15.
Published in final edited form as: Immunity. 2015 Dec 15;43(6):1053–1063. doi: 10.1016/j.immuni.2015.11.007

Affinity maturation of a potent family of HIV antibodies is primarily focused on accommodating or avoiding glycans

Fernando Garces 1,2,3, Jeong Hyun Lee 1,2,3, Natalia de Val 1,2,3, Alba Torrents de la Pena 6, Leopold Kong 1,2,3, Cristina Puchades 1, Yuanzi Hua 1, Robyn L Stanfield 1,2,3, Dennis R Burton 2,3,5,7, John P Moore 8, Rogier W Sanders 6,8, Andrew B Ward 1,2,3, Ian A Wilson 1,2,3,4,*
PMCID: PMC4692269  NIHMSID: NIHMS744079  PMID: 26682982

SUMMARY

The high mannose patch on the HIV-1 Envelope (Env) glycoprotein is the epicenter for binding of the potent broadly neutralizing PGT121 family of antibodies, but strategies for generating such antibodies by vaccination have not been defined. We generated structures of inferred antibody intermediates by X-ray crystallography and electron microscopy to elucidate the molecular events that occurred during evolution of this family. Binding analyses revealed that affinity maturation was primarily focused on avoiding, accommodating, or binding the N137 glycan. The overall antibody approach angle to Env was defined very early in the maturation process, yet differences evolved in the PGT121 family branches that led to differences in glycan specificities in their respective epitopes. Furthermore, we determined a crystal structure of the recombinant BG505 SOSIP.664 HIV-1 trimer with a PGT121 family member at 3.0 Å that, in concert with these antibody intermediate structures, provide insights to advance design of HIV vaccine candidates.

Introduction

The envelope glycoprotein trimer (Env) on the surface of HIV-1 is the sole target for neutralizing antibodies (nAbs) during the host immune response. Env is heavily glycosylated and the host-derived glycans conceal much of the protein surface from antibody recognition. Despite the extensive “glycan shield”, the immune system is capable of evolving an antibody response over time that is both potent and broadly neutralizing. A number of these broadly neutralizing antibodies (bnAbs) penetrate the glycan shield and engage both carbohydrate and protein components (Burton and Mascola, 2015; Ward and Wilson, 2015). To date, several bnAbs have been isolated from HIV-infected individuals that bind to glycans located in a high-mannose patch centered on the N332 glycan in gp120, and to protein components at the base of the V3 loop (Garces et al., 2014; Kong et al., 2013; Mouquet et al., 2012; Pejchal et al., 2011).

Members of the PGT121 family of Abs (PGT121, 122, 123 and 124) all recognize this high mannose patch and were isolated from a single donor (Donor 17) infected with a clade A HIV virus (Walker et al., 2011). Another set of PGT121 variants, including 10-1074, have also been isolated and characterized (Garces et al., 2014; Mouquet et al., 2012). PGT121, 122, and 123 are closely related members of the largest branch of the family, whereas PGT124 and 10-1074 reside on the other main branch (Sok et al., 2013). This family of Abs, although preferentially binding to the glycan at N332, can counter viral escape in the context of some isolates by binding to alternate, proximal glycans when the N332 glycan is absent (Garces et al., 2014; Sok et al., 2014). However, this promiscuity is complex and strain dependent. Neutralization assays have shown that the PGT121 family members engage or avoid different glycans throughout the affinity maturation process (Garces et al., 2014; Sok et al., 2014). Affinity maturation of HIV-1 bnAbs is driven by extensive somatic hypermutation (SHM) (Eisen and Siskind, 1964; Goidl et al., 1968; McKean et al., 1984) over a period of one to several years (Doria-Rose et al., 2009; Sather et al., 2009) through a process that may prove challenging to mimic during vaccine design. Nonetheless, inferred Ab pairs of the PGT121 family (3H+3L, 9H+3L and 32H+3L), identified by next-generation sequencing and exhibiting considerably less SHM than the corresponding mature Abs such as PGT121 or PGT124, still have notable broadly neutralizing capability (Sok et al., 2013).

Here, we present a comprehensive study of the molecular events associated with increased SHM and affinity maturation in the PGT121 family. Using site-directed mutagenesis, combined with Isothermal Titration Calorimetry (ITC), we measured Ab-antigen dissociation constants (Kd) at several distinct stages of affinity maturation. Moreover, crystal structures of putative precursor Abs of the PGT121 family have revealed the molecular details that led to affinity and specificity changes. Using x-ray crystallography and electron microscopy (EM), we have determined several structures of inferred intermediate Abs in complex with the soluble BG505 SOSIP.664 trimer (Env trimer corresponding to an HIV-1 clade A virus strain) and examined their specificity and angle of approach to Env throughout key steps in maturation of the PGT121 family. We have also determined the crystal structure of the 3H+109L Ab, whose heavy chain has the least SHM from the inferred germline PGT121 heavy chain, in complex with the Env trimer at 3.0 Å resolution (overall data completeness of 100%). Previous BG505 SOSIP.664 trimer crystals (Julien et al., 2013a; Pancera et al., 2014) have suffered from anisotropic diffraction that has limited the resolution achievable, but we overcame that problem by complexing the trimer with 3H+109L Abs and 35O22 Fabs. The outcome is an increased structural and functional understanding of this HIV-1 vaccine candidate and how it is recognized by the human immune system.

Results

Affinity maturation is focused on accommodating the N137 glycan

To study the molecular events associated with affinity maturation of the PGT121 family, we paired each of the inferred intermediate heavy chains 3H, 9H and 32H with the intermediate light chain 3L, as previously described (Sok et al., 2013), and used ITC to quantify Ab binding to BG505 SOSIP.664 trimers (produced in 293S cells that yield primarily high mannose (Man5-9) glycans). Although the trimer binding of the inferred PGT121 germline Ab was too weak to measure (Figure S1A), we successfully characterized the binding of other inferred precursors including 3H+3L, which had heavy and light chains that most closely resembled the inferred germline Ab (Figure 1 and Figure S1). During the maturation process, binding affinity was differentially gained in each of the two major branches of the PGT121 family (Figure 1 and Figure S1B). Indeed, when tested against BG505 SOSIP.664, the 9H+3L antibody displayed a ~6-fold increase in binding affinity for the BG505 SOSIP.664 trimer compared to 3H+3L, an earlier inferred Ab pair with low SHM common to both branches of the PGT121 family. In contrast, the affinity gain for 32H+3L from the PGT124 branch was modest compared that of 3H+3L (Figure 1 and Figure S1B). A very different scenario arose in further affinity maturation. The PGT124 affinity for the Env trimer was ~13 times higher than that of 32H+3L, whereas the PGT122 affinity decreased slightly compared to that of 9H+3L (Figure 1 and Figure S1B). Nevertheless, for PGT122, a large gain in enthalpy was observed but compensated for by a large loss in entropy (Figure S1B), which may in part have been due to slight differences in interaction with the N137 glycan (Garces et al., 2014) that decreased its conformational mobility. As expected, we observed an overall increase in binding affinity at every step of affinity maturation of the PGT121 family.

Figure 1. Representation of the relative binding affinities of antibody PGT121 family members to wild-type and glycan-variant SOSIP trimers.

Figure 1

Open in a new tab

Schematic representation of the PGT121 family phylogenetic tree with heavy chains paired with the corresponding light chains. Dissociation constants (Kd) were determined by ITC for selected members of this family from inferred germline to intermediates (precursors) and to affinity-matured Abs in complex with SOSIP.664 trimers and N137A or N332A glycan knock-out variants, expressed in either 293S or 293F cells (Figure S1). The fold changes in binding affinity (nM) (Figure S1B) on glycan removal were used to estimate how the N137 glycan contributes to affinity maturation. A relative scale for both positive and negative contributions to the Ab affinity was determined based on the extent of the changes in dissociation constants (Kd’s) (see also Figure S1B for complete thermodynamic data). Top right: surface representation of the location of glycans (colored) that interact with PGT121 family with the variable heavy chain in grey and variable light chain in light brown.

One particular feature of the PG121 family of bnAbs is the “open face” formed by complementarity determining regions (CDR) H1, H2 and H3, which is used either to bind the N137 glycan (PGT121-123) or to simply accommodate or avoid this sugar (PGT124) (Garces et al., 2014; Julien et al., 2013b; Mouquet et al., 2012). Thus, we also investigated Ab binding to a variant trimer with a point substitution that knocked-out the N137 glycan (BG505 SOSIP.664-N137A). For all Abs tested, the affinity and stoichiometry of trimer binding was increased when the N137 glycan was absent (Figure 1 and Figure S1B). Indeed, the Kd of 35 nM for the binding of early precursor 3H+3L Ab to the SOSIP.664-N137A trimer represented an ~15-fold enhanced affinity vs. its binding to SOSIP.664 (i.e., wild-type trimer) and was comparable to the Kd of 32 nM for the mature PGT124 bnAb and wild-type trimer (Figure 1 and Figure S1B). The 32H+3L antibody, a more mature Ab within the PGT124 branch, also exhibited an enhanced affinity (~16-fold) upon N137 removal, which was higher than the affinity of the final stage PGT124 affinity-matured Ab. Thus, perhaps not surprisingly, the affinity of PGT124 for the SOSIP.664-N137A trimer was also ~5-fold enhanced (Figure 1 and Figure S1B). The same trend was also observed for the PGT122 branch, where the affinities of the intermediate 9H+3L and the mature PGT122 bnAbs were enhanced by ~11- and ~5-fold, respectively, for SOSIP.664-N137A vs. the wild-type trimer (Figure 1 and Figure S1B). Hence, the N137 glycan impeded Env binding of all PGT121 family members tested, implying that accommodating this glycan was the major driving force behind affinity maturation.

To correlate our ITC data with neutralization assays, we also measured binding of PGT121 family Abs to wild-type and N137A SOSIP.664 trimers that were produced in 293F cells, where the glycans could be fully processed into complex and/or hybrid forms that likely more closely mimic the glycan composition on the viruses used in the neutralization assays. The resulting ITC data were similar to those obtained using 293S cell-produced trimers, in that the affinities of all the intermediate and mature Abs were increased for the SOSIP.664-N137A trimer versus wild type (Figure 1 and Figure S1B). However, the PGT122 affinity decreased when Env was produced in 293F cells (Figure 1 and Figure S1B), which may reflect differences in binding or avoiding complex glycans (i.e. larger and more complex carbohydrate structures that would be naturally present in some glycosylation sites on the virus and found in proteins expressed in 293F cells, as compared to high mannose only glycans in 293S cells) in the vicinity. Indeed, the PGT122-BG505 SOSIP.664 trimer structure showed that the light chain was in close proximity to the N156 and N301 glycans, and distal from the N137 glycan (Figure 1) (Julien et al., 2013a). To investigate whether pairing of precursor Ab heavy chains (3H, 9H and 32H) with a mature light chain (109L) (Sok et al., 2013) would affect trimer binding, we tested binding of hybrid Abs to the N137A SOSIP.664 trimer that contained the following heavy and light chains: 3H+109L, 9H+109L and 32H+109L (Figure 1 and Figure S1B). For each hybrid Ab, a slight increase in binding was observed over an Ab containing the precursor 3L light chain (Figure S1B), implying that the more mature 109 light chain has evolved to accommodate the N156 and N301 glycans. Hence, these two glycans also likely play a role in the affinity maturation of the PGT121 family.

The PGT121 family angle of approach to Env is largely defined at early stages of affinity maturation

The precise angle of approach to the trimer is important for bnAbs that recognize the CD4 binding site, as small variations can have a substantial adverse effect on HIV-1 neutralization (Chen et al., 2009). However, bnAbs that interact with the central N332 glycan can approach their target from several directions because access to the oligomannose patch is not as restricted (Kong et al., 2013). It is not clear how and when the approach angle (rotation and tilt) is set for any particular family of bnAbs and whether it changes during the affinity maturation process. Because this information would be valuable for immunogen design, we determined negative-stain EM structures of BG505 SOSIP.664 trimer complexes of the inferred intermediates and affinity-matured Abs described above (Figure 2 and Figure S2). To prepare a stable, higher affinity, complex with 3H+3L, we used the SOSIP.664-N137A trimer. The resulting negative-stain EM reconstruction could be fitted to a similar reconstruction of 32H+3L with the wild-type SOSIP.664 trimer with a correlation of 0.92, indicating that both Abs adopt the same overall angle of approach. Thus, the presence or absence of the N137 glycan does not grossly affect the approach angle. This observation also applies to the mature PGT124 Ab, as the equivalent correlation was 0.91 (Figure 2). For the PGT122 branch, correlations of the 3H+3L:SOSIP.664-N137A reconstruction with the 9H+3L and PGT122 reconstructions with SOSIP.664 were 0.87 and 0.95, respectively, again suggesting a general conservation of the approach angle of the antibodies to the Env trimer. These findings indicate that the PGT121 family angle of approach to Env is largely defined at early stages of antibody selection and affinity maturation.

Figure 2. Angle of approach to the Env trimer by PGT121 family members as determined by EM.

Figure 2

Open in a new tab

Schematic representation of Abs from the PGT121 phylogenetic tree (Sok et al., 2013) (displayed horizontally) in complex with SOSIP trimer as determined from negative-stain EM reconstructions for Ab:SOSIP complexes. The correlation values were calculated by fitting the EM reconstruction of the 3H+3L:SOSIP.664-N137A complex into the corresponding reconstructions of the 32H+3L:SOSIP.664, PGT124:SOSIP.664, 9H+109L:SOSIP.664 and PGT122:SOSIP.664 complexes (please see Figure S2).

Crystal structure of the BG505 SOSIP.664 trimer at 3.0 Å in complex with the 3H+109L Ab precursor

Crystal structures for mature PGT124 and PGT122 bnAbs in complex with their respective trimer epitopes have been previously reported (Garces et al., 2014; Julien et al., 2013a; Pancera et al., 2014). Here, we sought to compare the crystal structure of antibodies with less mature H or L chains of the PGT121 family with more mature antibodies of the family. The high binding affinity of the inferred heavy chain precursor 3H (paired with the mature 109L light chain) for the BG505.SOSIP.664 trimer in the absence of the N137 glycan (Figure 1 and Figure S1B) provided an opportunity to crystallize that complex. Addition of 35O22 Fab that binds to gp41 (Pancera et al., 2014) further facilitated formation of x-ray quality crystals and enabled the crystal structure of 3H+109L in complex with BG505 SOSIP.664-N137A and 35O22 Fab to be determined at 3.0 Å resolution (Figure 3A and Table S1).

Figure 3. The HIV-1 SOSIP Env trimer structure at 3.0 Å resolution.

Figure 3

Open in a new tab

(A) Ribbon representation of gp41 in the context of the BG505 SOSIP.664 trimer. The gp120 subunits are shown in surface representation while gp41 is shown in ribbon representation. (B) Ribbon representation of one protomer with the new region of 20 residues, located within the HR1 helix, and connecting the I548 and I568 residues, highlighted with a yellow surface. The position of the I559P mutation in the trimer is also indicated. On the right hand side is an enlarged image of the region between residues I548 and I568 (in red) and a 2Fo-Fc electron density map contoured at 0.5σ (yellow mesh). (C) Detailed representation of mutations introduced into the gp41 structure to stabilize the SOSIP.664 trimer.

This 3.0 Å crystal structure allowed us to interpret residues that were not visible or interpretable in previous Env trimer structures (Julien et al., 2013a; Pancera et al., 2014), as well as to improve the definition of side-chain rotamers and, hence, obtain further improvements to the Env trimer model. The overall structure (Fig. 3A,B) was very similar to that in Pancera et al. (0.4 Å Cα root-mean-square deviation (RMSD)), but some additional features were now visible, especially in gp41. The previous BG505 SOSIP.664 crystal structures have poor or no density in the central region of the gp41 heptad repeat 1 (HR1) sequence (residues I548 to I568) that connects the top of the HR1 helix with the fusion peptide proximal region (FPPR) (Julien et al., 2013a; Pancera et al., 2014). While there were still some interpretative difficulties, we saw significantly more density in this region (Figure 3B and Figure S3A). Even so, the more diffuse electron density compared to the surrounding areas suggests that this region was either inherently very flexible in the pre-fusion state of the trimer or that the trimer-stabilizing I559P mutation disrupted the native secondary structure. Overall, the electron density map suggested that we may have captured a structure that transitions between an α-helix and a more extended loop (Figure 3B). Moreover, it also illustrates the key role that the I559P mutation plays in stabilizing the trimer in a pre-fusion conformation.

We used this new structural information on HR1 to design mutations to verify the amino-acid register that we built for the now complete HR1 region and to possibly increase the stability of the SOSIP.664 trimers. The L555K substitution was designed to stabilize the trimer by introducing a hydrogen bond between gp41 in one protomer with gp120 from an adjacent protomer, and an L556K substitution to create a hydrogen bond between HR1 in gp41 and α-helix 0 in gp120 within a protomer. When co-expressed with furin under standard conditions, the resulting L555K and L556K trimers were fully cleaved as shown by the conversion of gp140 bands to gp120 and gp41 on a reducing SDS-PAGE gel (Figure S3B). Moreover, analysis by Native-PAGE gel showed that these mutants have also the same molecular weight as the wild type (Figure S3C). Hence, the point substitutions do not adversely affect the overall trimer conformation. Although the melting temperatures were essentially unchanged, the yields of the L555K and L556K trimers were increased by ~3 and ~2-fold, respectively, compared to wild type, when purified on a PGT145 bnAb affinity column (Figure 3C). A SOSIP.664-L555C/Q49C double mutant was then designed to introduce a new disulfide bond to covalently link gp41 with the gp120 N-terminus of the adjacent protomer (Figure 3C). Although analysis by SDS-PAGE and Native-PAGE gel showed that this double mutant exhibits a native conformation (Figure S3C), the expression yield of well-ordered trimers was 4-fold less than wild type. However, its melting temperature of 75.2° is the highest that we have observed to date for a SOSIP.664 trimer variant (Figure 3C).

We used a panel of bnAbs, non-neutralizing antibodies (non-nAbs) and CD4 to probe the antigenic structure of the above trimer mutants by ELISA, and found bnAbs and CD4 bound equivalently to wild-type BG505 SOSIP.664 (Figure S3D). Importantly, for the double mutant L555C/Q49C, binding of the 17b non-bnAb to its CD4-induced epitope was decreased 5-fold (Figure S3D).

Thus, the HR1 model based on the higher resolution structure (Figure 3B) was strongly validated by the structure-based mutants, implying that additional mutations can now be designed to further stabilize the pre-fusion form of the trimer and/or facilitate its production for vaccine trials.

Interaction of 3H+109L with the BG505 SOSIP.664 trimer

The 3H+109L Fab binds to the highly conserved “GDIR” peptide motif at the base of V3 (residues 324-327) and to the oligomannose glycan attached to N332 (Figure 4A). Although the electron density suggested a Man8GlcNAc2 glycan (Man8), the Ab could also accommodate a Man9 glycan without steric clashes. Additional electron density at N156 and N301 could be assigned to Man3GlcNAc2 and Man2GlcNAc2 moieties, respectively, (Figure 4A), but neither appears to establish a close contact with the Fab when the trimer is expressed in a high mannose form in 293S cells. As the 3H+3L combination Ab did not yield any trimer crystals despite extensive attempts, we compared its negative-stain EM structure in complex with BG505 SOSIP.664-N137A with the 3H+109L:BG505 SOSIP.664-N137A crystal structure (Figure S2E and F). Despite the differences in resolution, the two structures were readily compared due to the availability of higher resolution models that can be fit into the EM reconstruction (Figure S2E and F). The x-ray and EM structures are in agreement, suggesting that pairing 3H with either the 3L (inferred precursor) or the 109L (mature) light chain does not substantially affect antibody approach to the trimer and, hence, interaction with the epitope. To further analyze the 109L conformation, we superimposed it on the 3L light chain derived from the 9H+3L Fab crystal structure (Figure S2G and Table S1). Both structures aligned well, with a Cα RMSD of 0.6 Å. Moreover, 109L shared 82% identity of its VL with 3L, 95% with PGT124, but only 69% with PGT122 (Figure S2H). This indicates that 109L is phylogenetically more related to the PGT124LC than to the PGT122LC (Sok et al., 2013).

Figure 4. Crystal structure of the BG505 SOSIP.664 trimer in complex with Fab 3H+109L and Fab 35O22.

Figure 4

Open in a new tab

(A) A ribbon representation of the ternary complex of Fab 3H+109L (orange and brown for heavy and light chains, respectively) and Fab 35O22 (light gray and dark gray for heavy and light chains, respectively) bound to an Env protomer (gp120 is in light blue with N137 deleted from the trimer, and gp41 in green), and with the glycans represented as spheres. (B) Superimposition of 3H+109L with PGT122 (on top) and with PGT124 (center), and PGT122 with PGT124 (bottom) on the BG505 SOSIP.664 trimer to show their relative dispositions. Only the CDR loops and LFR3 of 3H+109L (in orange) and PGT122 (in gray) are shown. Right, schematic representation of the relative rotation in degrees (°) of the FV regions of the different Abs on the epitope surface. (C) Superimposition of the structures of the three complexes based on the gp120 subunits only (in ribbon representation), with the variable region of each Ab represented as a solid surface and the glycans in spheres. The vertical axes that point towards the center of epitope were used to calculate the comparative tilt angle for each Ab in relation to each other.

Comparisons of 3H+109L with PGT122 (4TVP) and PGT124 (4R2G), in complex with the BG505 Env trimer and JRCSF gp120 core, respectively, (Figure 4B), showed the CDR H3, L1 and L3 loops that contact the epitope core (GDIR and N332 glycan) were highly conserved (0.38 and 0.42 Å Cα RMSD for 3H+109L vs. PGT124 and 3H+109L vs. PGT122, respectively) (Figure S4A). The CDR H1 and H2 loops that comprise the “open face” were also conserved in shape among the 3H, PGT122 and PGT124 Abs (Figure S4C), but differed substantially in their relative disposition on the trimer (Figure 4B). To further investigate this variation, we determined the rotation angle of the Fab variable region (FV), which consisted of variable heavy and variable light chain domains, around the GDIR motif (after superposition of their gp120 chains). We observed that PGT124-FV is rotated by 10.4° and 10.8° relative to 3H+109L-FV and PGT122-FV on gp120, respectively, so that CDRs H1 and H2 of PGT124-Fv are displaced away from the N137 glycan. In contrast, for PGT122-FV, the corresponding deviation from the orientation of 3H+109L-FV on gp120 is only 1.4° (Figure 4B).

Differences in the approach angle are defined not only by the rotation of one Ab relative to the other, but also by differences in the Ab tilt angle relative to the trimer surface. The PGT124 tilt angle differed by ~5° compared to PGT122 and 3H+109L whereas PGT122 and 3H+109L differed by only 1° in this regard (Figure 4C). Thus, PGT124 also tilted away from the N137 glycan and towards N301, consistent with the same trend seen in the rotation angle. These differences in Ab rotations and tilt suggested the PGT124 branch of the family reduced contact with the N137 glycan during affinity maturation. Indeed, calculations of the surface areas that are buried when these Abs bind to the BG505 SOSIP.664 trimer (Table S2) indicated that PGT124 buries ~3-fold less heavy-chain surface area on N137 (108 Å2), compared to PGT122 (300Å2) or 3H (260 Å2).

We then compared the BG505 SOSIP.664-N137A high-resolution structure with the wild-type BG505 trimer crystal structure (4TVP) to identify any differences, especially in the glycans near N137 (Figure S4B). The conformation and/or orientation of the surrounding glycans (N332, N301 and N156) were not substantially affected by the absence of the N137 glycan. Indeed, the only changes were in the V1 loop upstream from A137, in a region that does not interact with PGT121 family Abs (Figure S4B). Hence, deletion of the N137 glycan from the Env surface increased antibody affinity without greatly affecting the native structure of PGT121 epitope.

The PGT121 family of antibodies vary their binding sites and angle of approach to accommodate the N137, N301 and N156 glycans during affinity maturation

Crystal structures show that Abs of the PGT121 lineage have a common overall architecture: an “open face” comprised of the three heavy chain CDRs and an “elongated face” formed by the light chain along with portions of CDR H3 (Figure S4C) (Julien et al., 2013b; Mouquet et al., 2012). While the elongated face mainly interacts with the GDIR peptide and the N332 glycan, the open face interacts with the N137 glycan. Moreover, while the interactions with the GDIR motif and the N332 glycan are mostly defined at the earliest stages of the affinity maturation process (Figure S4A) (Garces et al., 2014), the binding affinity data suggested different strategies have evolved for interacting with the N137 glycan in the different branches of the PGT121 lineage (Figure 1). Accordingly, we determined crystal structures of key heavy chain precursors (3H+109L, 9H+3L, 32H+109L) at branch points in the inferred phylogenetic tree of the PGT121 family to compare with the mature bnAbs (PGT122 and PGT124) (Table S1 and Figure 5). We then modeled a N137 complex glycan in the open face of the precursor Ab by superimposition with the PGT121 structure (Julien et al., 2013b; Mouquet et al., 2012). We identified potential clashes between the earliest precursor Ab 3H and the complex glycan (Figure 6A and Figure S5A) that were reduced or eliminated in 9H+3L and PGT121 (Figure 6A and Figure S5A and B). Indeed, residues (Y32, D53 and Q97) in the early precursor 3H that clashed with the glycan as well as others in the vicinity (Q96, Y100l and Y100n) are either mutated to residues with a smaller side-chain or adopt a different rotamer (H32, D53, Q96, H97, T100l and F100n (Figure S5) in 9H. In PGT121, some residues were further mutated (K53) or acquired a different rotamer (H97) in PGT121 to interact better with the N137 glycan (Figure 6A). Indeed, K53 and H97 were previously defined as key residues for interaction of PGT121 with a complex glycan (Julien et al., 2013b; Mouquet et al., 2012). In contrast, in the PGT124 branch, these residues are conserved from the 3H precursor (Figure 6A and Figure S5A and B).

Figure 5. Crystal structures of the PGT121 family members displayed from germline to mature antibodies.

Figure 5

Open in a new tab

Mapping of the Somatic Hypermutations (SHM) (in red spheres) onto the variable region of both heavy-chain and light-chain of the precursors and mature antibodies of the PGT121 lineage. The number of SHM (no. residues) for each antibody at each step in evolution is indicated.

Figure 6. Molecular details of the interaction between N332 proximal glycans and PGT121 family members.

Figure 6

Open in a new tab

(A) Interaction of N137 and the “open face” of the members of the PGT121 family. The glycan, in red, is modeled from the PGT121 structure crystalized with a complex glycan in the open face. The evolution of critical residues (in stick-and ball representation) throughout affinity maturation of the PGT122 and PGT124 branches is highlighted in enlarged boxes. The role each residue plays in the glycan interaction at different stages in the affinity maturation is highlighted to the right of each enlarged images. (B) Interaction between the N301 glycan (in yellow ball-and-stick) and the light-chain FR3 (cartoon representation). (C) The N156 glycan (in blue ball-and-stick), represented only as its Man3 form.

The light chain framework 3 (LFR3) of the PGT121 family has a three-residue insertion after residue L67 (67a,b,c) that is important for HIV-1 neutralization (Sok et al., 2013). This insertion occurs early in affinity maturation and enabled F67c at the tip of LFR3 to interact with I323 and G324 in the Env GDIR motif (Figure S5C), which is shielded in part by the N301 glycan (Figure 6B). A high degree of conformational flexibility was found at the LFR3 tip (Figure 6B) in the three mature light chains (PGT122, PGT124 and 109L), when comparing the free and bound forms of these Abs. Furthermore, F67c in the 109L and PGT124 Abs must change its side-chain rotamer from its unliganded form to avoid a clash with the N301 glycan. Although the LFR3 tip appeared to be more conserved between the unbound and bound states of PGT122 (Figure 6B), F67c adopted a different rotamer compared to the 109L and PGT124 Abs.

The light chain N-terminus of PGT121 family Abs acquired a deletion of as many as 7 residues during affinity maturation that increases neutralization potency (Sok et al., 2013). Our structural data show that N156 is proximal to this deleted N- terminus. Hence, although the glycan at N156 is only partially observed in the crystal structure, we can infer that it would likely clash with the N-terminus if present, which explains the deletion (Figure 6C). Thus key modifications to the antibody paratopes were able to accommodate the spectrum of glycans in vicinity of their epitopes and thereby increase antibody binding affinity.

DISCUSSION

Affinity maturation normally involves several rounds of exposure to an antigen in order to drive B cells to produce antibodies with successively greater affinities (Berek et al., 1991; Eisen and Siskind, 1964; Goidl et al., 1968; Pappas et al., 2014). For the PGT121 family, affinity maturation not only selects for more productive contacts between antibody and antigen or more rigid conformations of their complementarity determining regions (CDRs) (Sok et al., 2013), but also removes clashes with glycan(s) on the epitope perimeter. The binding data suggested that many of the productive contacts between the final bnAb and its peptide-glycan epitope are established at a very early stage in maturation of the lineage, and that the subsequent steps are largely focused on how to cope with neighboring glycans, mainly N137. Removal of the N137 glycan from the trimer had a profound effect on antibody binding affinity at all stages in the evolution of PGT121 lineage suggesting that several rounds of affinity maturation are required to minimize the blocking effect of this glycan.

While binding studies can be very informative on the course of affinity maturation, here we sought more detailed molecular information through the determination of crystal and EM structures of the Env trimer in complex with inferred intermediates in the PGT121 lineage. We found that the two branches of the lineage follow different strategies to deal with blocking glycans. The PGT124 branch evolved so that: 1) the antibody approaches the epitope at a ~5° tilt away from N137 and towards the N301 glycan compared to PGT122; and 2) the variable region (FV) correspondingly rotated 10.8° clockwise so that the CDR H1 and H2 loops move away from the N137 glycan. For the PGT121-123 branch, the strategy differs: 1) antibodies residues whose side chains clash with the N137 glycan are replaced by smaller side chains (9H+3L); and 2) mutations that establish productive contacts with the N137 glycan are then selected (PGT121-123). The shift in the focus of the PGT121-23 branch towards the N137 glycan may have been a consequence of the loss of the N332 glycan from the infecting virus at some stage during its evolution in the patient (Figure S6) since N137 can substitute for N332 in the context of certain isolates (Sok et al., 2014). The high affinity interaction of PGT124 that involves only a single glycan (the N332 glycan site is ~72% conserved across 3,045 HIV-1 strains in the Los Alamos HIV Sequence Database) could have exerted pressure on the virus to escape by deleting the N332 glycan (Figure S6). In this scenario, the evolution of antibodies in the PGT122 branch to productively bind the N137 glycan (~72% conserved) using the ‘open’ face would sustain an overall neutralization response against the virus (Figure S6). This mechanism of cooperation between different antibodies has previously been suggested for CD4 binding site Abs from two different lineages (Gao et al., 2014). Although our observations involved a single lineage of antibodies, they demonstrated how different antibodies in the same family can evolve synergistically. Our hypothesis is supported by the structural data showing that the early inferred precursor, 3H, binds the GDIR motif and N332 glycan in the same way as the mature bnAbs PGT122 and PGT124. Although the PGT122 branch depends on the N332 glycan throughout its evolution, the affinity of mature PGT122 was ~2 fold higher for the BG505 SOSIP trimer when N332 was absent (Kd values: SOSIP.664 174.0 nM vs SOSIP.664-N137A 82.4 nM), but we do not know if this is the case for other strains. Notwithstanding, these data reinforce the argument that PGT122, in contrast to PGT124, became more focused on the N137 glycan (Figure S6).

The 3H precursor recognizes the N332 glycan and the GDIR motif, which acted as a double anchor for this lineage. These interactions likely define the initial angle of approach for PGT121 family of Abs. Indeed, whether the N137 blocking glycan was present or not, the angle of approach in the PGT122 branch remained unchanged during affinity maturation. However, PGT124 must acquire additional mutations to avoid the N137 glycan and boost its affinity for the trimer.

We can now utilize such lessons for designing an immunogen that triggers this antibody lineage. For example, to elicit PGT121, 122 and 123 bnAbs (all in the same evolutionary branch), several steps could be designed to drive precursors to mature during the course of immunization. Firstly, exposing B cells to a stabilized trimer that displays the PGT121-124 epitope(s) may allow an appropriately shaped Ab to be selected; i.e. a precursor with open and closed faces that will set the angle of approach for the lineage. Immunogens lacking the N332 glycan could then mimic escape of the infected virus from PGT124, so as to enhance interaction with surrounding glycans, such as N137, and eventually induce PGT122-like bnAbs. In contrast, to drive maturation of PGT124 Abs (in the other evolutionary branch), the continuous presence of not only N332 but also the N137 glycan might be crucial to slightly modify the angle of approach and thereby accommodate or minimize interaction with N137. On the other hand, selection of the appropriate germline Ab may be aided by removal of the N137 glycan.

Our efforts to unravel the molecular details that drove the affinity maturation of the PGT121 lineage yielded an Env trimer structure at 3.0 Å resolution. Previous BG505 SOSIP.664 trimer crystals (Julien et al., 2013a; Pancera et al., 2014), have suffered from severe anisotropy, which limits the resolution in some directions and hence the overall resolution and experimental completeness of the data. If we compare the ‘effective resolution’ (Reseff) (Weiss, 2001; Weiss, 1997), where the high resolution limit is adjusted for data incompleteness (i.e. Resolution * Completeness −1/3), we see a substantial improvement, with Reseff =3.0 Å (this work), 4.9 Å (PDB ID:4NCO) (Julien et al., 2013a) and 3.8 Å (PDB ID:4TVP) (Pancera et al., 2014).

The quality of this crystal structure therefore not only improved the overall accuracy of the three-dimensional model of the Env trimer, but also added more structural information for regions not resolved in previous crystal and cryo-EM structures (Julien et al., 2013a; Lyumkis et al., 2013; Pancera et al., 2014). Although the structure of the post-fusion conformation of gp41 has long been known, this is not so for the pre-fusion state, which has proven to be difficult to determine by x-ray crystallography. An almost complete trace of gp41 was reported (Pancera et al., 2014). However, the connecting loop between helices α6 (FPPR) and HR1 (α7) was not resolved. We can now observe electron density for this region, and the resulting extensive modeling suggests that we may have captured a structure that alternates between a helix and an extended structure. Our structure also explained why the Ile559Pro (‘IP’ in SOSIP) mutation in this region destabilizes the post-fusion form and stabilizes the pre-fusion form of the Env trimer. The mutation is likely at the end of helix 6 and, therefore, prevented formation of a longer helix as in the fusogenic form, where the fusion peptide becomes accessible for insertion into the host cell membrane. Thus, this new structural information can be used for vaccine design by guiding further structure-based Env trimer variants.

EXPERIMENTAL PROCEDURES

Expression and purification of proteins

All Fabs were produced by transient transfection of FreeStyle™ 293F cells (Invitrogen) and purified by affinity chromatography followed by cation exchange chromatography and size exclusion chromatography (SEC) on a Superdex 200 16/60 column. BG505 SOSIP.664 wild-type and mutant trimers (SOSIP-137A and SOSIP-N332A) cloned in a phMCV3 vector, were expressed in FreeStyle™ 293F or 293S cells and purified using a 2G12-coupled affinity matrix followed by SEC, with exception for the BG505 SOSIP.664 N332A mutant for which a GN lectin affinity column was used in the initial stage. The structure-based SOSIP.664 L555K, L556K and L555C-Q49C mutants were purified using a PGT145-coupled affinity matrix and then by SEC.

Antibody and envelope substitutions

Substitutions in the Env glycoprotein constructs were introduced using QuikChange site-directed mutagenesis (Stratagene, La Jolla, CA). Substitutions were verified by DNA sequencing (Retrogen, San Diego, CA).

SDS-PAGE and Blue Native-Page

BG505 SOSIP.664 and mutant trimers were analyzed using SDS-PAGE and BN-PAGE and bands were revealed after staining with Coomassie blue. Each sample was pre-mixed with loading buffer and shortly after loaded onto a 4-12% Bis-Tris NuPAGE gel (Invitrogen). Each run was for 1.5h at 200V.

Formation of protein complexes and partial deglycosylation

Many combinations of ligand-SOSIP.664 trimer complexes, with and without deglycosylation, were subjected to crystal trials. Generally, trimers and ligands were mixed in a 1:3.2 molar ratio before SEC purification of the complex. To decrease heterogeneity on SOSIP.664 trimer:ligand complexes, deglycosylation was carried out on SOSIP.664 produced in 293S cells with Endoglycosidase H (New England Biolabs) in 200 mM NaCl, 50 mM sodium citrate pH 5.5 for 35 minutes at 37° C. Thus, the glycans protected by an interaction with a Fab are preserved intact, but those accessible to EndoH are trimmed to the core GlcNAc moiety.

Crystallization and data collection

Each Fab was concentrated to ~15 mg/ml before being screened against 960 crystallization conditions at both 4 and 20 °C. High quality crystals for 9H+3L Fab were obtained from 5% (w/v) PEG 1000, 28% (v/v) PEG 600, 10% (v/v) glycerol and 0.1 M MES pH 6.0. For 32H+109L, Fab crystals appeared in 40% (v/v) PEG 600, 0.1M NaCl and 0.1 M Na citrate pH 5.5. Crystals were harvested and cryoprotected by brief immersion in well solution augmented with 30% glycerol, followed by immediate flash cooling in liquid nitrogen. Data were collected at Advanced Light Source (ALS) and Advanced Photon Source (APS), respectively. The best 9H+3L crystal diffracted to 2.1 Å resolution and the diffraction data were processed with HKL-2000 (Otwinowski and Minor, 1997) to an overall Rsym of 0.06% and completeness of 98.4% (Table S1) in space group P21 with unit cell parameters: a = 70.9 Å, b =108.7 Å, c = 72.7 Å. Data were collected from a 32H+109L crystal to 2.4 Å and processed with HKL-2000 (Otwinowski and Minor, 1997) to an overall Rsym of 0.13% and completeness of 94.2% (Table S1) in space group P212121 with unit cell parameters: a = 40.7 Å, b = 338.1 Å, c = 62.1 Å.

To determine the molecular footprint and angle of approach of the inferred 3H precursor to the Env trimer, we screened a wide range of crystallization conditions with many combinations of Fab precursors with the BG505 SOSIP.664 constructs (SOSIP.664 wild type and SOSIP.664-N137A mutant). In addition, we used other crystallization additives, such as Fab 35O22, to facilitate crystal packing (Pancera et al., 2014). The most successful combination was the 3H+109L Fab with the BG505 SOSIP.664-N137A and the 35O22 Fab. Purified ternary complex samples were concentrated to ~12 mg/ml and screened against 480 crystallization conditions at both 4 and 20 °C. Suitable crystals were grown in 0.2 M CaCl2, 28% (v/v) PEG 400 and 0.1 M Hepes pH 7.5. The complex crystals were cryoprotected by brief immersion in well solution containing 20% glycerol, followed by immediate flash cooling in liquid nitrogen. Data were collected at APS beamline 23IDB. Although the best crystal has a few strong reflections visible at 2.7Å resolution, the final dataset was processed with HKL-2000 (Otwinowski and Minor, 1997) to 3.0 Å with an overall Rsym of 0.11 % and 100% completeness in space group P63 with unit cell parameters: a = b= 128.0 Å, c = 316.1 Å (Table S1).

Structure determination and refinement

The unliganded 9H+3L and 32H+109L structures were solved by molecular replacement (MR) using Phaser with the PGT124 Fab structure (PDB 4R26), from which the CDR loops had been deleted, as the initial model. For the ternary complex, multiple components were used for phasing by MR: one protomer of 35O22:BG505 SOSIP.664 (PDB ID:4TVP) and the high-resolution unliganded PGT124 Fab. Model building was carried out using Coot (Emsley and Cowtan, 2004) and refinement was carried out with phenix.refine using reference model restraints calculated from the structure of Fab 35O22 (PDB ID: 4TOY) (Adams et al., 2010), respectively.

Final Rcryst and Rfree values for 9H+3L and 32H+109L Fabs structure are 20.0% and 24.4% and 22.4% and 27.3%, respectively. For the complex structure, the corresponding values are 22.7% and 25.9% (Table S1). For the Fabs, residues were numbered according to Kabat (Martin, 1996), and gp120 numbered following the HXBc2 system (Ratner et al., 1987).

Isothermal titration calorimetry

ITC experiments were performed using a MicroCal Auto-iTC200 instrument (GE). All proteins were extensively dialyzed against a buffer containing 20 mM Tris, 150 mM NaCl, pH 7.4 before conducting the titrations. Subsequently, protein concentrations were adjusted and confirmed by using calculated extinction coefficients and absorbance at 280 nm. Ligands representing precursor and mature Abs of the PGT121 family were present in the syringe at concentrations ranging between 60-125 μM. The BG505 SOSIP.664 wild type and mutants were in the cell at concentrations ranging between 3.0-5.0 μM. Two-protein binding experiments were performed with the following parameters: cell at 25°C, 16 injections of 2.5 μl each, injection interval of 180 s, injection duration of 5 s, and reference power of 5 μcals. To calculate the dissociation constants (Kd), the molar reaction enthalpy (ΔH) and the stoichiometry of binding (N), Origin 7.0 software was used to fit and integrate the titration peaks using a single-site binding model.

Electron microscopy (EM)

All complexes were analyzed by negative-stain EM. The 3H+3L Fab complex was formed with the BG505 SOSIP.664-N137A mutant, while 32H+3L Fab was in complex with the wild type BG505 SOSIP.664. Both Env proteins were produced in HEK 293S cells. For the 9H+3L complexes, the BG505 SOSIP.664 and BG505 SOSIP.664-N137A trimers were produced in HEK 293F cells. A 3 μL aliquot of 10 μg/ml of the complex was applied for 15s onto a glow discharged, carbon coated 400 Cu mesh grid and stained with 2% uranyl formate for 20s. Grids were imaged using a FEI Tecnai T12 electron microscope operating at 120 kV using 52,000 x magnification and electron dose of 25 e−/Å2, resulting in a pixel size of 2.05 Å at the specimen plane. Images were acquired with a Tietz 4k × 4k CCD camera in 5° tilt increments from 0° to 55° at a defocus of 1000 nm using LEGINON (Suloway et al., 2005).

Image processing

Particles were picked automatically by using DoG Picker and put into a particle stack using the Appion software package (Lander et al., 2009; Voss et al., 2009). Initial reference-free 2D class averages were calculated using unbinned particles via the Xmipp Clustering 2D Alignment and sorted into 400 classes (Sorzano et al., 2010). Particles corresponding to the complexes were selected into a substack and another round of reference-free alignment was carried out using Xmipp Clustering 2D alignment and IMAGIC software (van Heel et al., 1996). To generate an ab initio 3D starting model, a template stack of 120 images of 2D class averages was used with imposing C3 symmetry. This starting model was refined against 12,245 raw particles for 30 cycles using EMAN (Ludtke et al., 1999). The resolution of the final reconstruction for 3H+3L with BG505 SOSIP.664-N137A was calculated to be 17Å, for 9H+3L with BG505 SOSIP.664 was 21 Å, for 9H+3L with BG505 SOSIP-N137A was 19 Å and for 32H+109L with BG505 SOSIP.664 was 22 Å using an FSC cut-off of 0.5 (Figure S2).

Determination of the approach angles of the Fabs to gp120 on the Env trimer

To measure the differences in angles of approach of different Fabs to the gp120, we used two complementary methods. First, we superimposed all the coordinates onto the 4TVP reference structure by their gp120 chains, using SSM superpose in Coot. Then, to calculate the overall angular difference in approach, the variable (VL-VH) domains of each Fab were superimposed using LSQMAN, with the overall angle for that superposition reported. This value represents the total rotational difference in Fab binding to gp120. The second measurement again started with all coordinates superimposed by their gp120 chains onto 4TVP. We then calculated an orthogonal set of axes bisecting each Fab variable domain, with one axis along the pseudo 2-fold axis relating VL and VH, a second perpendicular axis approximately bisecting the two canonical disulfide bonds in the variable domain, and a third, mutually perpendicular axis. We next calculated the angle between the pseudo-2-fold axes (these axes point towards the epicenter of the epitope) and also between the axes bisecting the canonical disulfides. These two angles define the difference in the overall angular approach (rotation and tilt) of the Fab to the gp120, and then how much the Fabs rotate to interact with antigen. The sum of these two angles is roughly equivalent to the total angular difference in binding orientation calculated by the first method.

Differential scanning calorimetry (DSC)

The thermal stabilities of BG505 SOSIP.664 and mutant trimers were analyzed on a VP-DSC calorimeter (GE Healthcare).

Supplementary Material

Acknowledgements

We are very grateful to X. Dai for assistance with data processing; M. Elsliger for computer support; M. Deller and H. Tien for the crystallization screening. X-ray data sets were collected at the ALS and APS. The ALS is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DEAC02-05CH11231. The APS is supported by the U.S. Department of Energy (DOE), Basic Energy Sciences, Office of Science, under contract no. DE-AC02-06CH11357. This work was supported by the HIV Vaccine Research and Design (HIVRAD) program (P01 AI082362 and P01 AI110657) (J.P.M., R.W.S., A.B.W., I.A.W.), R37 AI036082 (J.P.M.), by the Center for HIV/AIDS Vaccine Immunology and Immunogen Discovery (CHAVI-ID UM1 AI00663) and International AIDS Vaccine Initiative Neutralizing Antibody Center and Collaboration for AIDS Vaccine Discovery (CAVD) (A.B.W., D.R.B., I.A.W.), by NIH RO1 GM046192 (I.A.W.), by NIH 2R56 AI084817 (A.B.W., I.A.W.) and by the Joint Center of Structural Genomics (JCSG) funded by the NIH NIGMS, Protein Structure Initiative (U54 GM094586) (I.A.W.). R.W.S. is a recipient of a Vidi grant from the Netherlands Organization for Scientific Research (NWO) and a Starting Investigator Grant from the European Research Council (ERC-StG-2011–280829-SHEV). J.H.L. is supported by the California HIV/AIDS Research Program Dissertation Award D12-SRI-353.

Footnotes

Accession Codes

Coordinates and structure factors for Fabs 32H+109L and 9H+3L and for the ternary complex of 3H+109L with BG505 SOSIP.664-N137A mutant and 35O22 have been deposited in the PDB under accession codes 5CEX, 5CEY and 5CEZ, respectively. The 9H+3L:SOSIP.664, 32H+3L:SOSIP.664, 3H+3L:SOSIP.664-N137A and 9H+3L:SOSIP.664-N137A EM reconstructions are deposited in the EMDB under accession codes 6379, 3092, 3093 and 6380, respectively.

Author Contributions: Project design by F.G. and I.A.W.; X-ray work and analysis by F.G., L.K., R.L.S. and I.A.W.; EM work by J.H.L., N.V. and A.B.W.; ITC binding assays and analysis by F.G. and I.A.W.; SOSIP mutational work by F.G., A.T.P. and R.W.S.; Blue-Native and SDS-PAGE analysis by A.T.P and R.W.S.; DSC experiments by A.T.P and R.W.S.; Protein expression and purification by F.G., C.P. and Y.H.; manuscript written or edited by F.G., D.R.B., J.P.M., R.W.S., A.B.W. and I.A.W. All authors were asked to comment on the manuscript. This is manuscript # 29116 from The Scripps Research Institute.

SUPPLEMENTAL DATA

Supplemental Data includes 6 Figures and 2 Tables.

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. Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. 2010;66:213–221. doi: 10.1107/S0907444909052925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Berek C, Berger A, Apel M. Maturation of the immune response in germinal centers. Cell. 1991;67:1121–1129. doi: 10.1016/0092-8674(91)90289-b. [DOI] [PubMed] [Google Scholar]
  3. Burton DR, Mascola JR. Antibody responses to envelope glycoproteins in HIV-1 infection. Nat. Immunol. 2015;16:571–576. doi: 10.1038/ni.3158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen L, Kwon YD, Zhou T, Wu X, O'Dell S, Cavacini L, Hessell AJ, Pancera M, Tang M, Xu L, et al. Structural basis of immune evasion at the site of CD4 attachment on HIV-1 gp120. Science. 2009;326:1123–1127. doi: 10.1126/science.1175868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Doria-Rose NA, Klein RM, Manion MM, O'Dell S, Phogat A, Chakrabarti B, Hallahan CW, Migueles SA, Wrammert J, Ahmed R, et al. Frequency and phenotype of human immunodeficiency virus envelope-specific B cells from patients with broadly cross-neutralizing antibodies. J. Virol. 2009;83:188–199. doi: 10.1128/JVI.01583-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Eisen HN, Siskind GW. Variations in affinities of antibodies during the immune response. Biochemistry. 1964;3:996–1008. doi: 10.1021/bi00895a027. [DOI] [PubMed] [Google Scholar]
  7. Emsley P, Cowtan K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. 2004;60:2126–2132. doi: 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  8. Gao F, Bonsignori M, Liao HX, Kumar A, Xia SM, Lu X, Cai F, Hwang KK, Song H, Zhou T, et al. Cooperation of B cell lineages in induction of HIV-1-broadly neutralizing antibodies. Cell. 2014;158:481–491. doi: 10.1016/j.cell.2014.06.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Garces F, Sok D, Kong L, McBride R, Kim HJ, Saye-Francisco KF, Julien JP, Hua Y, Cupo A, Moore JP, et al. Structural evolution of glycan recognition by a family of potent HIV antibodies. Cell. 2014;159:69–79. doi: 10.1016/j.cell.2014.09.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Goidl EA, Paul WE, Siskind GW, Benacerraf B. The effect of antigen dose and time after immunization on the amount and affinity of anti-hapten antibody. J. Immunol. 1968;100:371–375. [PubMed] [Google Scholar]
  11. Julien JP, Cupo A, Sok D, Stanfield RL, Lyumkis D, Deller MC, Klasse PJ, Burton DR, Sanders RW, Moore JP, et al. Crystal structure of a soluble cleaved HIV-1 envelope trimer. Science. 2013a;342:1477–1483. doi: 10.1126/science.1245625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Julien JP, Sok D, Khayat R, Lee JH, Doores KJ, Walker LM, Ramos A, Diwanji DC, Pejchal R, Cupo A, et al. Broadly neutralizing antibody PGT121 allosterically modulates CD4 binding via recognition of the HIV-1 gp120 V3 base and multiple surrounding glycans. PLoS Pathog. 2013b;9:e1003342. doi: 10.1371/journal.ppat.1003342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kong L, Lee JH, Doores KJ, Murin CD, Julien JP, McBride R, Liu Y, Marozsan A, Cupo A, Klasse PJ, et al. Supersite of immune vulnerability on the glycosylated face of HIV-1 envelope glycoprotein gp120. Nat. Struct. Mol. Biol. 2013:796–803. doi: 10.1038/nsmb.2594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lander GC, Stagg SM, Voss NR, Cheng A, Fellmann D, Pulokas J, Yoshioka C, Irving C, Mulder A, Lau PW, et al. Appion: an integrated, database-driven pipeline to facilitate EM image processing. J. Struct. Biol. 2009;166:95–102. doi: 10.1016/j.jsb.2009.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ludtke SJ, Baldwin PR, Chiu W. EMAN: semiautomated software for high-resolution single-particle reconstructions. J. Struct. Biol. 1999;128:82–97. doi: 10.1006/jsbi.1999.4174. [DOI] [PubMed] [Google Scholar]
  16. Lyumkis D, Julien JP, de Val N, Cupo A, Potter CS, Klasse PJ, Burton DR, Sanders RW, Moore JP, Carragher B, et al. Cryo-EM structure of a fully glycosylated soluble cleaved HIV-1 envelope trimer. Science. 2013;342:1484–1490. doi: 10.1126/science.1245627. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Martin AC. Accessing the Kabat antibody sequence database by computer. Proteins. 1996;25:130–133. doi: 10.1002/(SICI)1097-0134(199605)25:1<130::AID-PROT11>3.0.CO;2-L. [DOI] [PubMed] [Google Scholar]
  18. McKean D, Huppi K, Bell M, Staudt L, Gerhard W, Weigert M. Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin. Proc. Natl. Acad. Sci. USA. 1984;81:3180–3184. doi: 10.1073/pnas.81.10.3180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Mouquet H, Scharf L, Euler Z, Liu Y, Eden C, Scheid JF, Halper-Stromberg A, Gnanapragasam PN, Spencer DI, Seaman MS, et al. Complex-type N-glycan recognition by potent broadly neutralizing HIV antibodies. Proc. Natl. Acad. Sci. USA. 2012;109:E3268–3277. doi: 10.1073/pnas.1217207109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Otwinowski Z, Minor W. Processing of X-ray diffraction data collected in oscillation mode. Meth. Enzymol. 1997;276:307–326. doi: 10.1016/S0076-6879(97)76066-X. [DOI] [PubMed] [Google Scholar]
  21. Pancera M, Zhou T, Druz A, Georgiev IS, Soto C, Gorman J, Huang J, Acharya P, Chuang GY, Ofek G, et al. Structure and immune recognition of trimeric pre-fusion HIV-1 Env. Nature. 2014;514:455–461. doi: 10.1038/nature13808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Pappas L, Foglierini M, Piccoli L, Kallewaard NL, Turrini F, Silacci C, Fernandez-Rodriguez B, Agatic G, Giacchetto-Sasselli I, Pellicciotta G, et al. Rapid development of broadly influenza neutralizing antibodies through redundant mutations. Nature. 2014;516:418–422. doi: 10.1038/nature13764. [DOI] [PubMed] [Google Scholar]
  23. Pejchal R, Doores KJ, Walker LM, Khayat R, Huang PS, Wang SK, Stanfield RL, Julien JP, Ramos A, Crispin M, et al. A potent and broad neutralizing antibody recognizes and penetrates the HIV glycan shield. Science. 2011;334:1097–1103. doi: 10.1126/science.1213256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. 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. Retrovir. 1987;3:57–69. doi: 10.1089/aid.1987.3.57. [DOI] [PubMed] [Google Scholar]
  25. Sather DN, Armann J, Ching LK, Mavrantoni A, Sellhorn G, Caldwell Z, Yu X, Wood B, Self S, Kalams S, Stamatatos L. Factors associated with the development of cross-reactive neutralizing antibodies during human immunodeficiency virus type 1 infection. J. Virol. 2009;83:757–769. doi: 10.1128/JVI.02036-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sok D, Doores KJ, Briney B, Le KM, Saye-Francisco KL, Ramos A, Kulp DW, Julien JP, Menis S, Wickramasinghe L, et al. Promiscuous glycan site recognition by antibodies to the high-mannose patch of gp120 broadens neutralization of HIV. Science Trans. Med. 2014;6:236ra263. doi: 10.1126/scitranslmed.3008104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Sok D, Laserson U, Laserson J, Liu Y, Vigneault F, Julien JP, Briney B, Ramos A, Saye KF, Le K, et al. The effects of somatic hypermutation on neutralization and binding in the PGT121 family of broadly neutralizing HIV antibodies. PLoS Pathog. 2013;9:e1003754. doi: 10.1371/journal.ppat.1003754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Sorzano CO, Bilbao-Castro JR, Shkolnisky Y, Alcorlo M, Melero R, Caffarena-Fernandez G, Li M, Xu G, Marabini R, Carazo JM. A clustering approach to multireference alignment of single-particle projections in electron microscopy. J. Struct. Biol. 2010;171:197–206. doi: 10.1016/j.jsb.2010.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Suloway C, Pulokas J, Fellmann D, Cheng A, Guerra F, Quispe J, Stagg S, Potter CS, Carragher B. Automated molecular microscopy: the new Leginon system. J. Struct. Biol. 2005;151:41–60. doi: 10.1016/j.jsb.2005.03.010. [DOI] [PubMed] [Google Scholar]
  30. van Heel M, Harauz G, Orlova EV, Schmidt R, Schatz M. A new generation of the IMAGIC image processing system. J. Struct. Biol. 1996;116:17–24. doi: 10.1006/jsbi.1996.0004. [DOI] [PubMed] [Google Scholar]
  31. Voss NR, Yoshioka CK, Radermacher M, Potter CS, Carragher B. DoG Picker and TiltPicker: software tools to facilitate particle selection in single particle electron microscopy. J. Struct. Biol. 2009;166:205–213. doi: 10.1016/j.jsb.2009.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Walker LM, Huber M, Doores KJ, Falkowska E, Pejchal R, Julien JP, Wang SK, Ramos A, Chan-Hui PY, Moyle M, et al. Broad neutralization coverage of HIV by multiple highly potent antibodies. Nature. 2011;477:466–470. doi: 10.1038/nature10373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ward AB, Wilson IA. Insights into the trimeric HIV-1 envelope glycoprotein structure. Trends Biochem. Sci. 2015;40:101–107. doi: 10.1016/j.tibs.2014.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Weiss MS. Global indicators of X-ray data quality. J. Appl. Crystallogr. 2001;34:130–135. [Google Scholar]
  35. Weiss MS, Hilgenfeld R. On the use of the merging R factor as a quality indicator for X-ray data J. Appl. Cryst. 1997;30:203–205. [Google Scholar]

Associated Data

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

Supplementary Materials

NIHMS744079-supplement.pdf (23.3MB, pdf)

RESOURCES