Structural and genetic basis of HIV-1 envelope V2 apex recognition by rhesus broadly neutralizing antibodies

Abstract

Broadly neutralizing antibodies targeting the V2 apex of HIV-1 envelope are desired as vaccine design templates, but few have been described. Here, we report 11 lineages of V2 apex-neutralizing antibodies from simian-human immunodeficiency virus (SHIV)-infected rhesus macaques and determine cryo-EM structures for 9. A single V2 apex-neutralizing lineage accounted for cross-clade breadth in most macaques, and somatic hypermutation relative to breadth was generally low, exemplified by antibody V033-a.01 with <5% nucleotide mutation and 37% breadth (208-strain panel). Envelope complex structures revealed eight different antibody classes (one multi-donor) and the complete repertoire of all five possible recognition topologies, recapitulating canonical human modes of apex insertion and C-strand hydrogen bonding. Despite this diversity in recognition, all rhesus-V2 apex antibodies were derived from reading frame two of the DH3-15*01 gene. Collectively, these results define-in rhesus-the structural and genetic basis of HIV-1 V2 apex recognition and demonstrate unprecedented structural plasticity of a highly selected immunogenetic element.

Figure 1.

Single-cell sorting identifies 11 cross-clade–neutralizing lineages from 10 SHIV-infected rhesus macaques with HIV-1 V2 apex–directed heterologous neutralization breadth. (a) Schematic for the present study. (b) Macaque host ID, infecting SHIV strain, and immunogenetics of a representative mAb from each of the 11 rhesus lineages reported in this study. (c) Neutralization breadth and potency of representative monoclonal lineage antibodies. Left: Neutralization activity against a 19-member panel of cross-clade tier-2 HIV-1 strains and a simian immunodeficiency virus infecting chimpanzees (SIVcpz), MT145K, which has been shown to bear the conserved HIV-1 V2 apex epitope. Data are reported as IC₅₀ titer (µg/ml) and colored according to the legend. Bold boxes demarcate activity against autologous virus; for example, 6070-a.01 was isolated from an animal infected with an SHIV bearing the HIV-1 CH505.TF Env (SHIV.CH505). All small panel neutralization experiments were performed in duplicate and repeated twice. Right: Neutralization breadth and geomean IC₅₀ against one or two large cross-clade panels of HIV-1 strains. # denotes activity against 119 viruses (Seaman panel); & denotes activity against 208 viruses (VRC panel). Bottom: Previously published rhesus and human V2 apex broadly neutralizing antibodies are included below the gray row for comparison; human antibodies are denoted with *. IC₅₀ data for these antibodies are shown in italics when obtained from their respective publications. 119 virus panel data (#) for CH01, VRC26.08, and PCT64-35M were derived from CATNAP (https://www.hiv.lanl.gov/components/sequence/HIV/neutralization/). All large panel neutralization assays were performed in duplicate. (d) SHM versus antibody breadth on a 119-isolate panel is shown for representative antibodies, with V2 apex rhesus antibodies in red. (e) SHM versus antibody category is shown for antibodies with over 30% breadth on a 208-strain HIV-1 panel. V2 apex antibodies have SHM levels lower than other categories, though similar to those of the MPER category. Notably, V033-a.01 with 37% breadth showed substantially lower SHM.

Figure S1.

Identification of V2 apex broadly neutralizing antibodies. (a) Rhesus macaques from which V2 apex lineages have been isolated in this study are grouped by their respective infecting SHIV strain. Animals previously described by Roark et al. (2021) for polyclonal V2 apex mutational mapping are denoted with #. (b) Neutralization of two sets of heterologous wild-type and V2 apex epitope mutant viruses by rhesus macaque plasma or purified polyclonal IgG (RM41328) for SHIV-infected rhesus macaques reported in this study. Plasma neutralization assays were repeated twice. (c) Representative FACS gating schemes for the two different single-cell sorting strategies. Top: Identification of 5695-b lineage members by collecting double-positive cells stained with a heterologous Env SOSIP probe conjugated with two different fluorophores. Bottom: Identification of 42056-a lineage members by collecting single-positive cells staining with heterologous wild-type and C-strand mutant Env SOSIP probe pairs.

Figure S2.

Phenotypic analysis of isolated broadly neutralizing antibodies. (a) Heatmap and phylogram based on hierarchical clustering of (log10-transformed) IC₅₀ neutralization titers against a 119-heterologous virus panel. Epitope classes are shown next to the phylogram, and branch splits with >50% bootstrap support are indicated. (b) Hierarchical clustering of Pearson correlations of (log10-transformed) IC₅₀ titers for the same pseudoviruses compared across broadly neutralizing antibodies. Of the seven clusters identified, three included V2 apex lineages. All axes, the majority of needles, and the only combined VRC26 clustered in the central cluster, while the remaining needles with two combined lineages formed the cluster at the bottom. DH1020 lineage members formed their own cluster. (c) Potent rhesus and human V2 apex–targeted lineages can be divided into five neutralization groups (I–V) based on mutant virus epitope mapping. Groups III and IV have not been previously described. The envelope trimer (PDB ID 4ZMJ) highlights the location of N-linked glycan and protein residue substitutions used for V2 apex mapping. Neutralization data values from mapping experiments are provided in Table S1. (d) Correlations between plasma neutralization ID₅₀s and predicted neutralization by isolated mAbs at the specific concentration, “C,” which is provided in µg/ml. Neutralization data values for plasma and antibodies are provided in Table S1.

Figure 2.

All SHIV-induced V2 apex–directed lineages are derived from the same rhesus DH3-15*01 gene in reading frame two and invariantly acquire a minimal five-residue motif. (a) Germline VH, DH, and JH genes for all potently neutralizing rhesus V2 apex–directed lineages described here and previously (13 total). (b) VDJ junction analysis of a representative antibody from each rhesus lineage. The respective germline VH, DH, and JH gene sequences are truncated and aligned to each VDJ junction, and each VDJ junction is aligned with respect to the DH gene. VH and JH nucleotides and residues are colored gray, non-templated nucleotides and residues (insertions and N/P additions) are colored blue, DH-gene nucleotides and residues are colored red, and SHM is colored black; we do not interpret SHM within non-templated regions. The reading frame in which each DH gene has been incorporated is denoted. The five-residue DH3-15*01 motif (EDDYG) acquired by all 13 lineages during VDJ recombination is highlighted with transparent red shading. (c) List of conserved DH3-15*01 residue motifs of varying length that are acquired by at least half of the rhesus lineages during VDJ recombination. The five-residue EDDYG motif described in panel b is written in red.

Figure S3.

Rhesus macaque and human immunoglobulin sequence analysis. (a–c) Ancestral clonal sequences identify HCDR3 indels and VDJ gene contributions and confirm the acquisition of a five-residue EDDYG motif for rhesus lineages with significant somatic mutation within D gene segments. HCDR3 alignments of the rhesus antibodies (a) 42056-a.01, (b) 6561-a.01, and (c) V031-a.01 to their respective VDJ germline genes and ancestral lineage intermediate sequences identified from peripheral memory B cells at the indicated time points. Mismatches to the germline V, D, and J genes are highlighted. (d) Human V2 apex broadly neutralizing lineage D genes are aligned with their rhesus D gene homologs. Nucleotide differences in rhesus homologs are colored in blue. Non-homolog rhesus D genes encoding the “YYD” motif in human DH3-3*01 are included below that alignment. Rhesus DH3-15*01 is included for reference with the unique three-residue anionic motif (EDD) underlined. All rhesus sequences are labeled in red. (e) Left, proportion of reading frame usage among naïve rhesus B cells derived from DH4-25*01, the rhesus homolog of the human PGT145 lineage DH4-17*01 gene, in the peripheral Indian rhesus macaque repertoire. Right, box plots showing in order: the frequency of DH4-17*01 usage in reading frame two among all naïve B cells in the peripheral rhesus repertoire; HCDR3 length distribution of naïve rhesus B cells in the peripheral rhesus repertoire derived from DH4-17*01 in reading frame two; HCDR3 net charge distribution of naïve rhesus B cells in the peripheral repertoire derived from DH4-17*01 in reading frame two. Charge calculations only consider amino acid residues and not predicted sites of tyrosine sulfation.

Figure 3.

Cryo-EM structures reveal needle-like modes of V2 apex recognition to be a reproducible antibody extended class in rhesus macaques. (a) Top: Cryo-EM reconstruction of 6070-a.01 in complex with Q23.17 MD39 Env at 3.6-Å resolution. The 6070-a.01 heavy and light chains are colored blue and gray, respectively. Envelope gp120, gp41, and N-linked glycans are colored turquoise, pink, and purple, respectively. Middle: Expanded interface view of 6070-a.01 from the top panel to highlight binding position and interactions with apical envelope glycans. Glycans bound by 6070-a.01 are shown in stick representation with transparent surfaces. The N160 glycan reoriented outward and away from the threefold trimer axis is denoted with *. Sulfated tyrosine residues are shown in stick representation to highlight their position within the trimer. Bottom: Further expanded interface view of 6070-a.01 to highlight interactions with apical envelope residues. Interacting residues are depicted in stick representation. Residues at positions corresponding to the conserved five-residue DH3-15*01 gene motif are colored dark red, while the remaining D gene residues are colored pink. Conserved motif position labels are italicized when subjected to SHM. Nitrogen atoms are colored blue, oxygen atoms are colored bright red, and sulfur atoms are colored yellow. Hydrogen bonds and salt bridges (distance < 3.3 Å) are depicted with dashed lines. (b) Top: Cryo-EM reconstruction of T646-a.01 in complex with Q23.17 MD39 Env at 3.5-Å resolution. The T646-a.01 heavy chain is colored orange, and the remainder of the complex is colored similarly to panel a. Middle: Expanded interface view of T646-a.01 from the top panel to highlight binding position and apical glycan interactions is shown similarly to panel a, including the N160 glycan reoriented into a horizontal conformation denoted with *. Bottom: Further expanded interface view of T646-a.01 to highlight apical residue interactions is shown similarly to panel a. (c) Top: Cryo-EM reconstruction of 42056-a.01 in complex with CAP256.wk34.c80 RnS2 SOSIP determined at 4.1-Å resolution. The 42056-a.01 heavy chain is colored light green, and the remainder of the complex is colored similarly to panel a. Middle: Expanded interface view of 42056-a.01 from the top panel to highlight binding position and apical glycan interactions is shown similarly to panel a. Bottom: Further expanded interface view of 42056-a.01 highlight apical residue interactions is shown similarly to panel a. (d) Top: Cryo-EM reconstruction of 44715-a.01 in complex with BG505 DS-SOSIP at 3.9-Å resolution. The 44715-a.01 heavy chain is colored teal, and the remainder of the complex is colored similarly to panel a. Middle: Expanded interface view of 44715-a.01 from the top panel to highlight binding position and apical glycan interactions is shown similarly to panel a. Bottom: Further expanded interface view of 44715-a.01 to highlight apical residue interactions is shown similarly to panel a. (e) Top: Expanded HCDR3 interface view of PGT145 (PDB ID 5V8L) to highlight apical residue interactions is shown similarly to panel a. The tyrosine sulfation posttranslational modification of Y100i was not included in this structure and therefore modeled here. Bottom: Expanded HCDR3 interface side view of the alignment of envelope complex structures of 6070-a.01, T646-a.01, 42056-a.01, and 44715-a.01 determined here to envelope complexes with human Fab’s PCT64-35S (PDB ID 7T74) and PGT145 (PDB ID 5V8L) and rhesus Fab RHA1.V2.01 (PDB ID 6XRT). Alignments were made with gp120 from each complex. Only gp120 of the 6070-a.01 complex is shown for clarity. Sulfated tyrosine residues are shown to highlight their positioning within the trimer.

Figure S4.

Rhesus V2 apex lineages bear tyrosine sulfation and O-linked glycosylation posttranslational modifications. (a) Summary of posttranslational modifications detected on F(ab’)2-digested rhesus and human antibodies by mass spectroscopy. Human antibodies are denoted with *. The number of sulfation groups detected per proteoform is written in red. All rhesus antibodies contained various types of O-linked glycosylation, while human antibody controls did not. Antibodies without structural data are denoted with n/a (not applicable). (b) Full list of individual detected experimental masses and corresponding deconvoluted posttranslational modifications that were summarized in panel (a). The number of sulfation groups detected per modification is written in red (#xSO3). (c) Representative deconvoluted mass spectra for rhesus lineage F(ab’)2 subunits with tyrosine sulfation peaks. Y axis: relative intensity. X axis: mass (Da). (d) Representative deconvoluted mass spectra for rhesus lineage F(ab’)2 subunits without tyrosine sulfation peaks. Y axis: relative intensity. X axis: mass (Da). (e) Deconvoluted mass spectra for human F(ab’)2-digested antibodies PGDM1400 and ACS202, which serve as positive and negative controls for tyrosine sulfation, respectively. Y axis: relative intensity. X axis: mass (Da).

Figure 4.

Cryo-EM structures reveal axe-like modes of V2 apex recognition to be a reproducible antibody extended class in rhesus macaques. (a) Top: Cryo-EM reconstruction of 41328-a.01 in complex with BG505 DS-SOSIP at 2.9-Å resolution. The 41328-a.01 heavy and light chains are colored blue and gray, respectively. Envelope gp120, gp41, and N-linked glycans are colored turquoise, pink, and purple, respectively. Middle: Expanded interface view of 41328-a.01 from the top panel to highlight binding position and interactions with apical envelope glycans. Glycans bound by 41328-a.01 are shown in stick representation with transparent surfaces. Bottom: Further expanded interface view of 41328-a.01 to highlight interactions with apical envelope residues. Interacting residues are depicted in stick representation. Residues at positions corresponding to the conserved five-residue DH3-15*01 gene motif are colored dark red, while the remaining D gene residues are colored pink. Conserved motif position labels are italicized when subjected to SHM. Nitrogen atoms are colored blue, and oxygen atoms are colored bright red. Hydrogen bonds and salt bridges (distance < 3.3 Å) are depicted with dashed lines. The orientations of the C-strand and HCDR3 β-strand mainchain interactions are labeled in the top left corner. (b) Top: Cryo-EM reconstruction of V033-a.01 in complex with BG505 DS-SOSIP at 3.1-Å resolution. The V033-a.01 heavy chain is colored orange, and the remainder of the complex is colored similarly to panel a. Middle: Expanded interface view of V033-a.01 from the top panel to highlight binding position and apical glycan interactions is shown similarly to panel a. Bottom: Further expanded interface view of V033-a.01 to highlight apical residue interactions is shown similarly to panel a. (c) Right: Expanded HCDR3 interface view of CH03 (PDB ID 5ESV) to highlight apical residue interactions is shown similarly to panel A. Left: Expanded HCDR3 interface side view of the alignment of SOSIP complex structures of 41328-a.01 and V033-a.01 determined here to an envelope complex with human Fab PG9 (PDB ID 8FL1) and V1V2–scaffold complex with human Fab CH03 (PDB ID 5ESV). Alignments were made with the V1V2 region from each complex. Only gp120 of the 41328-a.01 complex is shown for clarity.

Figure 5.

Cryo-EM structures reveal combined modes of V2 apex recognition to be a reproducible antibody extended class in rhesus macaques. (a) Top: Cryo-EM reconstruction of V031-a.01 in complex with BG505 DS-SOSIP at 3.1-Å resolution. The V031-a.01 heavy and light chains are colored blue and gray, respectively. Env gp120, gp41, and N-linked glycans are colored turquoise, pink, and purple, respectively. Bottom: Expanded interface view of V031-a.01 from the top panel to highlight binding position and interactions with apical envelope glycans. Glycans bound by V031-a.01 are shown in stick representation with transparent surfaces. Sulfated tyrosine residues are shown in stick representation to highlight their position within the trimer. (b) Further expanded interface views of V031-a.01 to highlight interactions with apical envelope residues. Interacting residues are depicted in stick representation. Residues at positions corresponding to the conserved five-residue DH3-15*01 gene motif are colored dark red, while the remaining D gene residues are colored pink. Conserved motif position labels are italicized when subjected to SHM. Nitrogen atoms are colored blue, oxygen atoms are colored bright red, and sulfur atoms are colored yellow. Hydrogen bonds and salt bridges (distance < 3.3 Å) are depicted with dashed lines. Top: interactions made with the primary recognized C strand. The orientations of the C-strand and HCDR3 β-strand mainchain interactions are labeled in the top left corner. Bottom: Interactions mediated by HCDR3 residues inserted into the trimer hole. (c) Top: Cryo-EM reconstruction of 6561-a.01 in complex with Ce1176 RnS2 SOSIP at 4.1-Å resolution. The 6561-a.01 heavy chain is colored orange, and the remainder of the complex is depicted similarly to panel a. Bottom: Expanded interface view of 6561-a.01 from the top panel to highlight binding position and interactions with apical glycans is shown similarly to panel a. (d) Further expanded interface views of 6561-a.01 to highlight interactions with apical residues are shown similarly to panel b. (e) Top: Cryo-EM reconstruction of 40591-a.01 in complex with T250.4 RnS2 SOSIP at 4.2-Å resolution. The 40591-a.01 heavy chain is colored orange, and the remainder of the complex is colored similarly to panel a. Bottom: Expanded interface view of 40591-a.01 from the top panel to highlight binding position and interactions with apical Env glycans is shown similarly to panel a. (f) Further expanded interface views of 40591-a.01 to highlight interactions with apical Env residues are shown similarly to panel b. (g) Expanded HCDR3 interface views of VRC26.25 (PDB ID 6VTT) to highlight apical envelope residue interactions are shown similarly to panel b. Left: Interactions are mediated by HCDR3 residues inserted into the trimer hole. Right: Interactions mediated by all other Fab residues. (h) Expanded HCDR3 interface side view of the alignment of envelope complex structures of V031-a.01, 6561-a.01, and 40591-a.01 determined here to the envelope complex of VRC26.25 (PDB ID 6VTT). Alignments were made with gp120 from each complexes, while only gp120 of the VRC26.25 complex is shown for clarity. Sulfated tyrosine residues are shown to highlight their positioning within the trimer. Residue F100e of 40591-a.01 is also shown since it is similarly inserted into the trimer hole but cannot be modified posttranslationally.

Figure S5.

Structural superimpositions and sequence signatures for rhesus DH3-15–encoded V2 apex antibody classes. (a) Structural and sequence definition of the RHA1 reproducible antibody class. (b–h) Unique structural and sequence definitions of the 42056-a, 44715-a, 41328-a, V033-a, V031-a, 40691-a, and 6561-a antibody classes. In the sequence signatures, “x” represents any amino acid, while the numbers indicate the Kabat positions of the respective residues.

Figure 6.

The rhesus DH3-15*01 exhibits structural plasticity and encodes a unique anionic motif. (a) HCDR3 structures from rhesus PGT145-like Fab’s in complex with envelope trimers. DH3-15*01 conserved five-residue motif (EDDYG) positions are colored and labeled in dark red, and the remaining D gene positions are colored and labeled in pink. Conserved motif position labels are italicized when subjected to SHM. D gene position side chains are shown in ball-and-stick representation with nitrogen atoms colored blue, oxygen atoms colored red, and sulfur atoms colored yellow (Kabat numbering). The remaining HCDR3 residues are colored gray with side chains hidden. Below each structure is an alignment of the germline DH3-15*01 coding fragment that was acquired during VDJ recombination (bottom) with the sequence at these positions in the mature antibody (top; Kabat numbering). Somatic mutations that conserve side chain aromaticity or anionic charge are depicted in bold, while discordant somatic mutations are underlined. Sites of tyrosine sulfation are highlighted in green. The functional role of the conserved five-residue motif segment is written above the alignment. (b) HCDR3 structures from rhesus PG9-like Fab’s in complex with envelope trimers. Structures are depicted similarly to panel a. (c) HCDR3 structures from rhesus VRC26-like Fab’s in complex with SOSIP trimers. Structures are depicted similarly to panel a. Residue insertions in these structures and sequence alignments are colored blue. (d) Proportion of reading frame usage among naïve rhesus B cells derived from DH3 genes in the peripheral Indian rhesus macaque repertoire. DH3-15*01 is highlighted in red here and in the remaining panels. B cells derived from rhesus DH3-19*01 are exceedingly rare and therefore excluded from analysis. (e) Frequency of DH3 family usage in reading frame two among all naïve B cells in the peripheral rhesus repertoire. (f) HCDR3 length distributions of naïve rhesus B cells in the peripheral rhesus repertoire derived from DH3 genes in reading frame two. (g) HCDR3 net charge distributions of naïve rhesus B cells in the peripheral repertoire derived from DH3 genes in reading frame two. The net charge of DH3-15*01–derived HCDR3s is more anionic (****, P < 0.00001 Student’s t test) than all other groups. Charge calculations only consider amino acid residues and not predicted sites of tyrosine sulfation.

Figure 7.

The rhesus DH3-15*01 combines residue features that are advantageous for V2 apex epitope recognition. (a and b) Features of the Indian rhesus macaque and human DH-coding sequence repertoire partitioned by the D gene family. Each dot represents a germline D gene expressed in one of the three forward reading frames with a value corresponding to the total number of residues (length), the number of Tyr, Trp, and Phe residues (aromatic), and the number of Glu and Asp residues (anionic). Boxes extend to the respective D gene family mean value, and whiskers show the standard deviation. Sequences were excluded if stop codons were found in the middle of the coding sequence in any particular reading frame. In panel a, the rhesus DH3-15*01 gene expressed in reading frame two is marked with a red symbol, and rhesus D genes with length and aromatic features equal to DH3-15*01 in reading frame two are marked with green and orange symbols, respectively. In panel b, D genes utilized by HCDR3-dominated human V2 apex lineages are marked according to the legend, and the rhesus DH3-15*01 gene expressed in reading frame two is included as a reference to facilitate comparison.

Figure 8.

Cross-species antibody recognition of the HIV-1 V2 apex site of vulnerability and their induction by SHIV. (a) Summary of rhesus and human structural modes of HCDR3-dominated V2 apex recognition. Roman numerals denote the neutralization mapping group (groups I–V) for each lineage as detailed in extended data Fig. 2 D; infecting SHIV strain and antibody class are also delineated as detailed in extended data Table S2 and extended data Fig. 6, respectively. (b–d) The 5-Å Env footprints of the most broadly neutralizing rhesus (top) and human (bottom) lineages with needle-like (b), axe-like (c), and combined (d) modes of recognition are mapped in red onto each respective trimer complex. The remaining gp120 surface is shown in gray, and all other components of the structure are omitted for clarity. Top view of trimer. (e) Epitope and paratope characteristics of V2 apex lineages compared across species and extended class. Each plot is a different structural feature that groups lineages by host species on the left and HCDR3 topology on the right. P values are only listed for statistically significant differences (*, P < 0.05; ***, P < 0.0005; unpaired t test) between two groups on the same side of the plot.

Figure 9.

Highly selected immunogenetic elements can adopt diverse structural and functional roles in epitope recognition. A timeline of the vaccine field’s evolving understanding of antibody development and specificity. It was initially thought that the high diversity of possible antibodies generated from V(D)J recombination and heavy-light chain pairing meant that, for vaccine purposes, each generated antibody would be unique, and antibodies in different individuals would be different. However, from 2004 to 2013, several studies established that virtually identical antibodies could arise in different individuals due to genetic elements being highly selected to recognize the same HIV-1 gp120 or influenza HA epitopes through similar, often identical, structural modes of recognition. Thus, the same antibodies could arise in different individuals with genetic elements selected to have the same mode of recognition. Now, we have discovered that it is possible for a highly selected genetic element to adopt different modes of recognition – with the selected genetic element exhibiting both structural diversity and distinct recognition chemistries.