Longitudinal Analysis Reveals Early Development of Three MPER-Directed Neutralizing Antibody Lineages from an HIV-1-Infected Individual

SUMMARY

Lineage-based vaccine design is an attractive approach for eliciting broadly neutralizing antibodies (bNAbs) against HIV-1. However, most bNAb lineages studied to date have features indicative of unusual recombination and/or development. From an individual in the prospective RV217 cohort, we identified three lineages of bNAbs targeting the membrane-proximal external region (MPER) of the HIV-1 envelope. Antibodies RV217-VRC42.01, VRC43.01 and VRC46.01 used distinct modes of recognition and neutralized 96%, 62%, and 30%, respectively, of a 208-strain virus panel. All three lineages had modest levels of somatic hypermutation, normal antibody-loop lengths, and were initiated by the founder virus MPER. The broadest lineage, VRC42, was similar to the known bNAb 4E10. A multimeric immunogen based on the founder MPER activated B cells bearing the unmutated common ancestor of VRC42, with modest maturation of early VRC42 intermediates imparting neutralization breadth. These features suggest that VRC42 may be a promising template for lineage-based vaccine design.

Graphical Abstract

eTOC Blurb

Despite extensive structural work, little is known about how MPER-specific HIV antibodies develop. Krebs et al describe the initiation and early development of three MPER-directed HIV broadly neutralizing antibody lineages from a single donor of the RV217 cohort. One of the antibody lineages appears particularly promising for vaccine design, as it achieved high neutralizing breadth with low mutation from germline.

INTRODUCTION

Neutralizing antibodies (NAbs) against HIV-1 are likely to be a major component of the protective immune response elicited by an effective vaccine (Haynes et al., 2012a; Wagh et al., 2018). Cross-reactive NAbs commonly arise during HIV-1 infection, though only a small subset of infected patients produce NAbs with high breadth and potency (Hraber et al., 2014; Simek et al., 2009). In contrast, vaccine immunogens derived from the HIV-1 envelope glycoprotein (Env) have failed to elicit cross-reactive neutralizing antibodies (Excler et al., 2015). Studying the structure, epitope recognition, and development of broadly neutralizing antibodies (bNAbs) in infected individuals is likely to provide important clues for vaccine design (Klein et al., 2013; Mascola and Haynes, 2013). Current vaccine strategies under investigation include structure-guided immunogen design, with a focus on epitopes of known bNAbs (Xu et al., 2018; Kulp et al., 2017); and lineage-based design (Doria-Rose and Joyce, 2015; Haynes et al., 2012b; Kwong and Mascola, 2018). The latter strategy requires a detailed understanding of the co-evolution of virus and B cell lineages in HIV-infected individuals who developed bNAbs.

Studies of the longitudinal development of bNAbs that include samples from early lineage development have been published for three of the major bNAb-targeted regions on Env, called supersites or sites of vulnerability: CD4 binding site (CD4bs) (Bonsignori et al., 2016; Gao et al., 2014; Liao et al., 2013), V1V2 apex (Doria-Rose et al., 2014; Landais et al., 2017), and glycan-V3 (Bonsignori et al., 2017; MacLeod et al., 2016). In addition to several shared insights, each of these studies revealed aspects of antibody development that were unique to the targeted site. Thus, understanding B cell development in additional individuals continues to be valuable, particularly for sites for which little is currently known.

The membrane-proximal external region of Env (MPER) is a major target for neutralizing antibodies in infected individuals (Doria-Rose et al., 2017), and the target of several known bNAbs, including 2F5, Z13e1, 10E8, DH511, and 4E10 (Buchacher et al., 1994; Huang et al., 2012; Williams et al., 2017; Zwick et al., 2001), which were isolated from chronically infected donors. 4E10, DH511 and 10E8 target the C-terminal region of MPER and are some of the broadest NAbs described, neutralizing over 90% of viral strains on multi-clade panels. 10E8 and DH511 form a class of bNAbs, where a reproducible or multidonor class is defined as a set of antibodies isolated from different individuals that share the same mode of recognition of an epitope and have genetic similarities such as germline VH and VL usage (Kwong and Mascola, 2012; Zhou et al., 2013). 4E10 shares similar genetics to the weak neutralizer CH12, using VH1-69 and Vκ3-20 genes (Morris et al., 2011). Despite extensive structural work on antibody-MPER interactions, little is known about how MPER-specific antibodies develop. Samples from chronic infection have yielded some insight using next-generation sequencing (Soto et al., 2016; Williams et al., 2017). However, the lack of longitudinal and early samples in infection has limited the utility of such studies to uncover the earliest events in the development of these antibodies.

Here, we study longitudinal samples from an HIV-1 infected individual who developed MPER-directed bNAbs. We isolated MPER bNAbs from 3 different lineages and traced their early origins. The most potent lineage, VRC42, is a member of the same bNAb class as 4E10, implying that the lessons learned from this study should be applicable to the design of vaccines aimed at eliciting antibodies of this class.

RESULTS

Participant 40512 from the RV217 Cohort Achieves Broad HIV Neutralization through Three MPER-Directed Lineages

To identify individuals who developed bNAbs, longitudinal plasma samples collected prior to ART initiation from individuals enrolled in RV217.(Robb et al., 2016) were assessed for neutralization of a diverse panel of 34 pseudoviruses. RV217 was a prospective study, which sought to capture individuals in the Fiebig I infection stage prior to seroconversion (RNA⁺, p24Ag⁻, Ab⁻). Following the first RNA-positive sample (pRNA⁺), intensive sampling was taken during acute and chronic infection timepoints. Donor RV217.40512 was identified in Fiebig stage I, and plasma taken at 646 days pRNA⁺ neutralized 85% of the viruses on the panel (Figure 1A and Table S1). Longitudinal timepoints were then assessed for neutralization and viral load (Figure S1A). Single genome amplification of longitudinal samples revealed that this individual became superinfected between 330 and 401 days pRNA⁺ (Figures S1A-B), and prior to neutralization breadth (Figure 1A). Neutralization fingerprinting (Georgiev et al., 2013) predicted the specificity of the bNAbs at day 646 to target MPER (Figure 1B). To confirm this specificity, HIV-2/HIV-1 MPER chimeras (Davis et al., 2009) were assessed for sensitivity to plasma sampled at longitudinal timepoints starting at day 29. MPER chimera neutralization was first observed at day 154 and increased over time (Figures 1A and S1C). Concurrently, autologous plasma was found to potently neutralize the founder virus starting at day 154, while the superinfecting virus was not neutralized by the autologous plasma until after superinfection occurred (Figure 1A).

Figure 1. Donor 40512 from the RV217 cohort achieves broad HIV neutralization through three MPER-directed lineages.

(A) Donor RV217.40512 plasma samples were assessed for neutralization against the founder (red line) and superinfecting (gray line) viruses, HIV-2/HIV1-MPER C1 chimera (dotted line), and heterologous viruses (blue line and fill). Gray bar, window of superinfection. (B) Neutralization-fingerprint analysis using a panel of 34 Env-pseudoviruses. Time point (days) following the first positive RNA test after infection. Values indicate proportion of plasma neutralization associated with the indicated antibody. Colors are defined by epitope (top row) with darker shades indicating higher values. (C) Genetic characterization, isotype, and neutralization breadth of mAbs recovered from B cells at day 646. Percent identity is calculated based on nucleotide sequence. CDR3 length is based on amino acid sequence using Kabat numbering. (D) Neutralization breadth and potency against 208 Env-pseudotyped viruses. Dendrograms show neighbor-joining trees of Env sequences, with branches color-coded by sensitivity to the indicated antibody. See also Table S1 and Figure S1.

To isolate antibodies from this donor, we used an MPER peptide conjugated to fluorescently labeled streptavidin (Williams et al., 2017). The peptide competed neutralization of both autologous (Figure S1D) and heterologous viruses (Figure S1E), suggesting that the broadly neutralizing fraction bound to this peptide. IgG⁺ B cells that bound to the peptide tetramer were sorted from peripheral blood mononuclear cells (PBMCs) from day 646 (Figure S1F-G). Three lineages of neutralizing antibodies were recovered that targeted MPER: RV217-VRC42.01-VRC42.05, RV217-VRC43.01-VRC43.03, and RV217-VRC46.01 (Figure 1C). Genetic analysis showed that all had modest levels of somatic hypermutation, with 9-13% mutation from the corresponding heavy chain V genes and 5-9% from the light chain V genes at the nucleotide level (Figure 1C). These values were similar to the MPER bNAb 4E10 (12% VH), lower than 10E8 and DH511 (22% and 16%, respectively), and much lower than the values for many CD4bs-targeting antibodies (18-39%) (Zhou et al., 2015). Notably, the VRC42 lineage used the same VH and Vκ genes as the bNAb 4E10 and CH12. VRC46 used the same VH gene as VRC42 but a different light chain gene, and VRC43 V genes were different from both VRC42 and VRC46 as well as from all previously published MPER antibodies. While all of our antibodies were subcloned and expressed as IgG1, the original isotypes of VRC42.01-05 were all IgG1, as were VRC43.02 and VRC43.03. In contrast, VRC43.01 and VRC46.01 were both IgG3. (Figure 1C).

VRC46.01, the three VRC43 antibodies, and four of the VRC42 antibodies neutralized the founder virus and multiple heterologous viruses on a panel of 29 pseudotyped viruses (Figure 1C and Table S1). The antibodies collectively, and VRC42.01 alone, neutralized 28 of 29 viruses that were sensitive to the plasma at Day 646 (Table S1). On a multi-subtype panel of 208 Env-pseudoviruses, VRC42.01 neutralized 96% of the viruses with a median IC50 of 4.7 μg/ml against sensitive viruses (Figure 1C-D). VRC43.01 was also very broad and slightly more potent, with breadth of 63% and median IC50 of 1.7 μg/ml against sensitive viruses. VRC46.01 was less broad, but still had activity against 30% of viruses of diverse subtypes and a median IC50 of 11 μg/ml (Figure 1C-D and Table S2).

Three Lineages Target Distinct Epitopes within the MPER via Different Modes of Recognition

To verify that the VRC42, VRC43, and VRC46 antibodies targeted MPER, we tested their ability to neutralize a series of HIV-2/HIV-1 MPER chimeras (Figure 2A). VRC42.01, VRC43.01, and VRC46.01 neutralized the full clade B MPER chimera, but not wild type HIV-2. VRC42.01 targeted the smallest segment of the HIV-1 MPER, with modest activity against a chimera containing the minimal 4E10 epitope at the C-terminus of MPER. VRC43 and VRC46 required longer C-terminal segments of HIV-1 MPER (Figure 2A).

Figure 2. The Three Lineages Target Distinct Epitopes within the MPER via Different Modes of Recognition.

(A) HIV-2/HIV-1 MPER chimeric viruses were used to determine MPER-specific neutralization. The panel of chimeric viruses is shown with the modified MPER peptide highlighted in red for each chimera (left), and the IC50 values (μg/ml) for each mAb (right). Averages from two experiments are shown. (B) Fold change in IC50 between wild-type founder Env or clade B Env BG1168 viruses containing single ala/gly point mutations. Dotted line, three-fold change. Each assay was performed once. (C) Crystal structures of VRC42.04 Fab and VRC46.01 Fab in complex with the founder MPER peptide, and VRC43.01 Fab alone. (D,E) Close-up views of VRC42.04-MPER (D) and VRC46.01-MPER complex (E). MPER residues that conferred substantial IC50 changes when changed to ala (Figure 2B) are highlighted in magenta. Heavy and light chain residues that are in contact with the highlighted MPER residues were shown in sticks and dots. The Cα atoms of VRC46.01 residues that are in contact with MPER were shown in spheres. The dotted red lines represent hydrogen bonds between two atoms. See also Figures S2-S4 and Tables S3-S5.

To understand the binding properties of the antibodies, we utilized multiple assays. Each mAb bound robustly to the full-length Env of the founder virus in context of cell surface expression (Figure S2A). In contrast, while VRC42.01 and VRC46.01 bound strongly to soluble founder Env gp140, VRC43.01 did not bind (Figure S2B). Only VRC42.01 bound to free MPER peptide in ELISA, although VRC42.01, VRC43.01 and VRC46.01 were bound to multimerized founder MPER peptide conjugated to keyhole limpet hemocyanin (KLH) (Figure S2C-D). In addition, MPER-directed bNAbs, in particular 4E10, are known to interact both with Env and lipid. We therefore tested VRC42.01, VRC43.01, and VRC46.01 with panels of sphingolipids and phospholipids, and found that all of the antibodies bound in a manner similar to 4E10 (Figure S2E). We also tested these mAbs for binding to liposomes in the presence or absence of membrane-anchored MPER peptide (Figure S2F). VRC42.01 and VRC42.02 bound more strongly to peptide-bearing liposomes compared to liposomes lacking MPER peptide, and VRC46.01 bound equally to both. Little binding of VRC43.01 was detected to liposomes with or without MPER peptide (Figure S2F).

We next mapped the epitopes of VRC42.01, VRC43.01, and VRC46.01. First, we generated a series of point mutants across the MPER of the founder virus by changing each amino acid to ala or gly (Figure 2B, Table S3). VRC42.01 showed decreased activity against viruses mutated at W672 and W680; VRC43.01 was similarly affected by mutations at N671, D674, and W680; and VRC46.01 was affected by mutations at E662, D664, F673, and D674 (Figure 2B). Additional VRC42 and VRC43 family members gave similar results, although with slight variations; and similar results were obtained for a second virus, BG1168 (Table S3).

For VRC42 family members, binding was also assessed on a series of C-terminal MPER peptides, each with a single residue changed to ala or gly (Figure S3A). The binding patterns for VRC42.01-.04 were similar to those of 4E10 and 10E8, with some differences in specificity between the lineage members (Figure S3A).

VRC43.01 binding to free peptide and soluble proteins was low (Figure S2B-C), limiting our ability to epitope map via binding studies. Therefore, we used a bioinformatic analysis of neutralization data from our 208-virus panel (Figure 1D and S3B). Amino acid resistance analysis (Doria-Rose et al., 2012) showed that certain amino acids at positions 667, 671, and 674 were found primarily in resistant strains (Figure S3B), suggesting the importance of those positions for neutralization activity, and in agreement with the neutralization ala scanning data for positions 671 and 674 (Figure 2B and Table S3).

To further define the structural interactions of the antibodies with MPER, we obtained co-crystal structures of VRC42.04 and VRC46.01 in complex with founder MPER peptide at 2.2 and 1.9 Angstrom resolution, respectively (Figure 2C-E and Table S4). To aid in these studies, we constructed a variant of T117v2, a protein scaffold that holds the C-terminal portion of MPER (Env amino acids 671-683) in a helical conformation (Correia et al., 2010; Irimia et al., 2016). This variant, T117-F, contained the sequence of the founder MPER peptide (Table S7). We obtained a crystal structure of VRC42.01 in complex with T117-F (Figure S4A). Of note, binding of VRC43 antibodies to the founder MPER peptide and T117-F were undetectable (Figure S2C-D and data not shown), and this lack of binding suggests a mode of recognition different from VRC42 or VRC46. We obtained structures of Fab alone for VRC43.01 (Figure 2C) and VRC43.03 (Figure S4A).

The structures of VRC42.04 and VRC46.01 showed differences in binding to MPER peptide (Figure 2C-E and S4B-C). Although they both use predominantly hydrophobic residues on CDR H2 loops for MPER binding, as has been reported for most HIV-1 antibodies using the VH1-69 gene (Huang et al., 2004), the angle of approach for MPER binding and the conformation taken by the MPER peptide differed substantially between the structures. VRC42.04-bound peptide was helical from residues 672 to 683, while the VRC46.01-peptide had an extended coil conformation from residues 665 to 677 followed by a single helical turn at the C-terminal (residues 678-682) (Figure 2C-E and S4B-C). Notably, Trp 672 and Trp 680 of gp41 contacted VRC42.04 CDR H2 and CDR H3 loops with buried surfaces of 173 Ų and 105 Ų, respectively, while Asp 674 made H-bonds with Asn 30 and Asn 32 of VRC42.04 light chain (Figure 2D). Contacts between VRC42.01 and the MPER peptide portion of T117-F were quite similar to those in the VRC42.04-peptide complex (Figure S4A and Table S5). While the VRC46.01-peptide complex was substantially different from VRC42.01 or VRC42.04, several commonalities were noted: Phe 673 of MPER made substantial hydrophobic contacts with the CDR H2 of VRC46.01, while Asp 674 made an H-bond with the protonated Tyr 96 of the VRC46.01 light chain (Figure 2D-E). Together these specific hydrophobic and hydrogen bond interactions provide the structural basis of the reduced sensitivity to VRC42.01 and VRC46.01 when these gp41 residues in the founder and BG1168 viruses were mutated to Ala (Figure 2B).

To further quantify the differences in mode of binding between VRC42, VRC46, and the previously published MPER bNAbs, we compared the peptide structures as determined in antibody-peptide co-crystals (Figure S4C). In agreement with other data, this analysis showed that the VRC42.04-engaged MPER conformation was closest to that of 4E10 (Figure S4C).

We also used a bioinformatic approach to understand the similarity of these antibodies to each other and to known MPER bNAbs. The neutralization fingerprints showed that VRC42.01 was most similar to 4E10, 10E8, and DH511, while VRC43.01 and VRC46.01 correlated best with each other and Z13e1 and 2F5, respectively (Figure S4D).

All Three Lineages Were Initiated Between Days 85 and 154 Post-infection

To understand early development of these MPER lineages, we performed next-generation sequencing (NGS) on samples from days 85, 154, 240, 485, and 646 (Figure 3 and S1G), interrogating both IgM and IgG sequences. No transcripts of the VRC42, 43 or 46 lineages were detected at day 85, however all 3 lineages were detected in the IgG compartment starting from day 154 (Figure 3A). The percent divergence from germline precursors increased over time with significant increase between day 154 and day 240 (Figure 3B). Combined, these data suggest that all three lineages were elicited by descendants of the founder virus and prior to superinfection. From the day 154 NGS data, VRC42 lineage transcripts with as few as 5 nucleotide changes in the VH region were detected, and we inferred the Unmutated Common Ancestor (UCA) (Figure S5A) and early intermediates (I1, I2, I3) along the phylogenetic path to VRC42.01 (Figures 3C, and Table S6-7).

Figure 3. NGS of B cell Transcripts Reveals All Three Lineages To Be Initiated Between days 85 and 154 Post-Infection but with Divergent Frequencies over Time.

(A) Timing of origin and frequencies of VRC42, VRC43, and VRC46 lineages. % Reads, percent of high-quality reads (see Methods) that are assigned to each lineage. (B) Divergence of heavy and light chains over time from germline precursor V genes. (C) Maximum likelihood phylogenetic trees, based on nucleotide sequence, of the VRC42 lineage showing a representative sampling of the longitudinal NGS data. Colors indicate the time point from which the transcript was sequenced. Dots indicate sequences from which peptides were found in plasma by proteomic analysis at day 646. An NGS sequence with the same amino acid sequence as VRC42.I3-H (VRC42.I3-aa) is labelled in green. Trees for the VRC43 (D) and VRC46 (E) lineages are presented similarly.

For the VRC43 lineage, the least mutated transcripts were 5.7% divergent from germline VH and we were unable to infer a high-confidence UCA (Figure 3D and Tables S6-7). Therefore, we inferred the most recent common ancestor (VRC43.I1) and a V gene revertant of this intermediate (VRC43.I1gHgL). Data were limited for VRC46, as only 5 total VRC46 lineage heavy chain transcripts were detected across all 5 timepoints by NGS. Therefore, we inferred the most recent common ancestor (VRC46.I1), and its V gene revertant (VRC46.I1gHgL) (Figure 3E and Tables S6-7).

Proteomic Analysis of Plasma Antibodies Identifies a VRC42 Lineage Member that More Closely Resembles bNAb 4E10

The NGS data showed early branching and diverse sequences within the VRC42 and VRC43 lineages (Figure 3). To determine whether this diversity in the memory B cell repertoire was reflected in circulating antibodies, we performed a proteomic analysis (Williams et al., 2017) of MPER-specific plasma antibodies at day 646 and identified multiple lineage members (Figure 4A). Among the VRC42 heavy chain sequences from NGS, we observed three different lengths of CDRH3: 15, 16, and 18 aa (Kabat numbering). Likewise, all 3 CDRH3 lengths were detected in the plasma IgG (Figure 4A), thus the diversity in the memory B cell compartment is reflected in the IgG that was actively secreted at the time. We also observed VRC43 lineage-member peptides in the plasma, but no peptides clonal to VRC46 (Figure 4A).

Figure 4. VRC42 Lineage Member that More Closely Resembles bNAb 4E10 was Identified by Proteomics Analysis.

(A) Proteomic analysis of plasma antibodies detected VRC42 lineage members with CDRH3 lengths of 15 (grey), 16 (pink), and 18 (blue) amino acids at day 646. These included a peptide matching VRC42.02 (16 amino acid CDRH3, pink stripes) and one matching VRC42.N1 (18 amino acid CDRH3, blue stripes). Two peptides from the VRC43 lineage were also detected (brown). Slices are proportional to total peak area, not number of unique peptides. Numbering employs the Kabat convention. (B) Junction and CDRH3 of VRC42.01-.04, NGS-derived sequence VRC42.N1, and 4E10. Residues in bold are identical to 4E10. (C) Neutralization of 208 Env-pseudotyped viruses of diverse clades. Red line, median of neutralized viruses. See also Table S2.

Of particular interest within the VRC42 lineage was a set of sequences with an 18 amino acid long CDR H3, matching the CDR H3 length found in 4E10. This length resulted from duplication of the GW residues at positions 99 and 100, resulting in a GWGW motif in the same position as found in 4E10 (Figure 4B). We reconstructed one of these antibodies, termed VRC42.N1, by pairing one of the 18 aa CDR H3 heavy chains with an NGS-derived kappa from the corresponding part of the phylogenetic tree (Zhu et al., 2013). VRC42.N1 demonstrated a 4-fold higher potency than VRC42.01, approaching the potency of 4E10, 10E8, and DH511; and greater breadth when measured at IC80<50 μg/ml: 90% compared to 65% for VRC42.01 (Figure 4C and Table S2). Thus, the most 4E10-like VRC42 lineage member was the most potent.

VRC42 and 4E10 Are Members of the Same bNAb Class

The VRC42 lineage is derived from the VH gene segment 1-69 and Vκ segment 3-20 (Table S6), the same genes used by 4E10. The CDR H3 amino acid sequence of 4E10 is quite similar to that of VRC42 antibodies (Figure 4B) and the mapping analyses show similarities between 4E10 and the VRC42 antibodies (Figures 4C, S2 and S4D). To definitively show that VRC42 lineage shares a class with 4E10, we compared the crystal structures of VRC42.N1.and 4E10 (PDB ID: 2FX7) bound to MPER peptides or the T117-F scaffold. When the MPER peptide (671-683) and the corresponding T117-F MPER region were aligned, the variable domains of the two structures were superimposable, and each of the 6 CDR loops took on similar conformations (Figure 5A). Furthermore, most MPER contact residues were at the same positions in their amino acid sequences (Figure 5B and Table S5). In addition, the peptide itself was held in a near identical conformation (RMSD of Cα =0.60A). VRC42.01 and VRC42.04 were also extremely similar to VRC42.N1 and 4E10 (Figure 5C). When aligned, the CDR H3s of VRC042.01 and VRC42.04 overlapped with RMSD of 0.6 A and the CDR H3s of VRC42.N1 and 4E10 with RMSD of 1.91A. However, the tryptophans within the GW and GWGW motif permitted different conformations with varying degrees of predicted membrane interaction (Kwon et al., 2018), which may have resulted in the differences in neutralization potency (Figure 5C). Altogether, these data indicate that the VRC42 antibodies and 4E10 are members of the same class of bNAbs.

Figure 5. VRC42 and 4E10 are Members of the Same bNAb Class.

(A) VRC42.N1 and 4E10 have the same structural mode of recognition. MPER peptide residues (magenta) spanning 671 to 683 of the two structures were superposed and their CDR loops and MPER peptides were displayed in the same angle respect to each other. For clarity, only the MPER region of T117-F in complex with VRC42.N1 is shown. (B) Alignments of heavy and light chain sequences. Top line: predicted naive gene segments for VRC42: VH1-69*10, DH3-10, and JH6. 4E10 and CH12 use VH1-69, but different D and J genes. Residues that contact with gp41 peptide are highlighted in yellow. No structure is available for CH12. (C) Close-up views CDRH3s of VRC42.01, VRC42.04, VRC42.N1, and 4E10 with MPER peptide (magenta). Aromatic residues at the tip were highlighted with surface representation. See also Figure S4 and Tables S4-S5.

A Multimeric Immunogen Based on the Founder MPER Engages UCA and Early Intermediates of the VRC42 Lineage

We next sought to understand the origins and early development of these bNAb lineages, with special focus on the VRC42 lineage. While all 3 bNAb lineages were detected at day 154, no amino acid changes were detected in the MPER from the founder virus until day 401 (Figure 3A). Therefore, we hypothesized that the founder MPER sequence initiated these bNAb lineages. We expressed the inferred UCA and intermediates I1-I3 of the VRC42 lineage (Figure 3 and Tables S6-7) and tested them for binding to the founder MPER peptide (Figure 6). Weak binding of the VRC42.UCA was observed to cell-surface expressed founder gp160 (Figure S6A), but not to soluble founder gp140 as measured by BLI (Figure S6C), MPER peptide or the MPER founder peptide conjugated to KLH as measured by ELISA (Figure 6A). No binding was observed to subtype B gp140 Envs by either the VRC42.UCA or the VRC42. altUCA (Figure S6C). In contrast, VRC42.UCA bound modestly to the monomeric T117-F scaffold (Figure 6B). This result was confirmed using Bio-Layer Interferometry (BLI) with a Fab k_D determined to be 488 nM (Figure 6C and S5B). Using BLI, weak binding of VRC42.UCA to MPER-KLH was also observed (Figure 6D). VRC42.UCA bound very well to T117-F presented on a 60-mer nanoparticle (T117-F-LS 60mer) (Figure 6B and S5C-D). Collectively, these data are consistent with the notion that naïve B cell of the lineage was able to engage the MPER from the founder virus. The early intermediates VRC42.I1, VRC42.I2, and VRC43.I3 also bound to these forms of founder Env and MPER, with increased binding as the maturation increased (Figure 6 and S5B).

Figure 6. A Multimeric Immunogen Based on Founder MPER Engages VRC42.UCA and Early Intermediates.

(A-B) Binding of VRC42 lineage antibodies and 10E8 as measured by ELISA. Plates were coated with (A) biotinylated founder MPER peptide or MPER-KLH, (B) T117-F monomer or T117-F-LS 60mer. Data are representative of 2-4 repeat assays. (C) Affinity of VRC42 lineage antibodies to T117-F monomer was measured by Bio-Layer Interferometry (BLI). (D) Binding to MPER-KLH or T117-F-LS 60mer as measured by BLI. (E) Cell signalling through the B cell receptor measured by calcium flux. Cells loaded with the calcium-sensitive dye Fura Red were stimulated with antigen and interrogated by flow cytometry. The ratio of signals in the V655 and B710 channels indicates the level of intracellular calcium-bound Fura Red. Data are representative of 4 repeat assays. See also Figure S5 and S7.

We constructed a Ramos cell line that expresses VRC42.UCA as surface IgM. T117-F-LS 60mer nanoparticles activated signaling by VRC42.UCA BCR, while monomeric T117-F and MPER-KLH did not (Figure 6E and S5E). Thus, the naïve precursor of the VRC42 lineage was competent to respond to an antigen containing the founder MPER.

We also reconstructed the most recent common ancestor of the other two lineages (VRC43.I1 and VRC46.I1), as well as germline reversions (VRC43.I1gHgL and VRC46.I1gHgL). All of these bound to cell-surface expressed founder gp160 (Figure S6B-C). In addition, VRC43.I1 and VRC46.I1 neutralized heterologous viruses (Figure S6D). Collectively these data show that the early precursors of diverse MPER bNAbs could be activated by the founder MPER.

Many MPER bNAbs are known to be autoreactive or polyreactive (Haynes et al., 2005; Williams et al., 2017). Therefore, we tested the VRC42, VRC43, and VRC46 lineage antibodies for autoreactivity. VRC42 and VRC43 antibodies were negative or mildly reactive in staining HEp-2 cells (Figure S6E). Several VRC42 and VRC43 lineage members showed binding to cardiolipin, at variable levels and in most cases less than 4E10 (Figure S6F). There was no reactivity of VRC42.UCA on HEp-2 cells or to cardiolipin. These antibodies also reacted to several RNA and DNA binding proteins in the ATHENA panel, suggesting polyreactivity (Figure S6G). VRC46.01 was highly reactive in all assays (Figure S6E-G).

Broad Neutralization by VRC42 Lineage Intermediates with Low Somatic Mutation

We next investigated the neutralizing activity of VRC42 lineage members. VRC42.UCA did not neutralize the founder virus, nor a panel of heterologous viruses (Figure 7A-B). VRC42.I1 had modest activity against the founder virus, and increasing potency was observed with increasing maturation (Figure 7A). The first intermediate lineage member to show activity against heterologous viruses was VRC42.I2 (Figure 7B). VRC42.I3 neutralized 51% of heterologous viruses on panels of 34 (Figure 7B) and 208 (Table S2) viruses, yet was a mere 2.2% mutated across the VH and VL compared to the inferred UCA, with a total of 13 amino acids different from UCA (Figure 7B). While there are only three amino acid changes between VRC42.I2 and VRC42.I3, two heavy chain (Q61P and V100dL) and one light chain (S93N), VRC42.I3 demonstrated increased binding to founder MPER peptide in ELISA, and its neutralization activity was 4-fold more potent than VRC42.I2 and substantially broader (Figure 7B). To narrow down which of the three changes between VRC42.I2 and VRC42.I3 were required to achieve this increase in breadth and potency, we paired the VRC42.I2 heavy chain with VRC42.I3 light chain and vice versa. In both cases, the binding and neutralization matched the antibody from which the heavy chain was taken (Figure 7A-B), suggesting that the change in the light chain did not affect activity. When the two heavy chain amino acid changes were tested individually, the resulting antibodies expressed poorly. In summary, two changes in heavy chain were sufficient to impart the improved binding and neutralization.

Figure 7. VRC42 Intermediates with Low Somatic Mutation have Broad Neutralization Activity.

(A) Neutralization of founder virus by VRC42 lineage intermediates. Data are averages of 6 repeats. (B) Neutralization breadth of VRC42 lineage intermediates to a panel of 29 heterologous pseudoviruses. % mutation is the nucleotide change in heavy and light chains compared to UCA. Data are average of 2 repeat assays. (C) Sequences of intermediates and mature VRC42.01. MPER contact residues are highlighted in yellow. (D) Mapping the somatic mutations on the models (VRC42.I1, I2 and I3) and structure (VRC42.01 mature). Somatic mutations inherited from VRC42.I1, I2, and I3 were shown in green, blue, and orange spheres, respectively. Somatic mutations found in mature VRC42.01 were shown in gray and pale green spheres. See also Figure S6.

To understand the structural basis of the improved binding and neutralization, we generated homology models of VRC42.I2 and VRC42.I3 using the crystal structure of the VRC42.01-T117-F scaffold complex (Figures 7C-D and S7A). The models showed that the two heavy chain residues did not make direct contacts with MPER. However, Gln/Pro 61_HC were found at the heavy and light chain interface where they could affect the conformation of CDR L3, which contacts gp41 peptide, by potentially interacting with Pro 95_LC (Figure S7A-B). Interestingly, Pro 61_HC made a favorable hydrophobic-hydrophobic interaction with Pro 95_LC and this interaction was further optimized in mature VRC42.01 with Pro 95_LC changed to Phe, while Val/Leu 100d_HC pointed away from the MPER, suggesting that the improved binding and potency of VRC42.I3 may be attributed to the Pro 61_HC mutation (Figures 7C-D and S7A-B).

The VRC42.I3 antibody has a total of just 9 heavy chain and 4 light chain amino acid changes from the UCA, and was able to neutralize 50% of heterologous viruses (Figure 7B-C). Furthermore, an NGS sequence with an exact match to VRC42.I3-H at the amino acid level was observed at day 240 (Figure 3C), while sequences differing from I3-K at a single position were present at day 154. VRC42.I3 thus represents antibodies present in the memory B cell repertoire as early as day 240. In summary, these data provide evidence that by engaging the founder MPER peptide, neutralization breadth arose in the VRC42 lineage within the first 9 months of infection with just 2% somatic mutation from the UCA.

DISCUSSION

Studies of bNAb development in HIV-infected individuals have yielded valuable insights into vaccine design. Here, we elucidated early steps in the development of bNAbs against the MPER. We isolated MPER-targeting bNAbs from three lineages in a single donor, one of which, VRC42 belongs to the 4E10 class of bNAbs. This antibody arose from a precursor that could be triggered by an immunogen bearing the autologous MPER of the founder Env, and required less than 2% somatic hypermutation to achieve 50% neutralization breadth.

Donor RV217.40512 developed three highly distinct lineages that targeted MPER. Notably, other donors have been shown to have multiple lineages targeting MPER (Williams et al., 2017) as well as the glycan-V3 supersite (Bonsignori et al., 2017; Longo et al., 2016) and CD4 binding site (Gao et al., 2014). The factors that predispose individuals for multiple lineages are unclear. The VRC42, VRC43, and VRC46 lineages all arose between days 85-154, at which time the MPER sequence was identical to founder Env. This suggests that the founder virus Env may be a favorable sequence for immunogen design. Characteristics of the founder virus that may have allowed the elicitation of MPER-directed bNAbs, including its MPER sequence and the accessibility of its MPER to antibodies, are the subject of continuing study.

The VRC42.UCA bound to MPER with the sequence from founder Env when presented in a variety of formats. Importantly, a multimeric protein containing founder MPER was able to activate BCR signaling by cells bearing the VRC42.UCA. These data suggest that proteins bearing founder MPER, including those described here, might serve as immunogens to stimulate naive B cells to initiate 4E10-class antibodies. In addition, early VRC42 lineage members were capable of broad and potent neutralization. Antigens that bind to early intermediates might serve as boosting immunogens to guide maturation of the lineage, and the level of mutation required for VRC42.I3-like activity is well within the range of mutation elicited by protein immunogens in other studies (Moody et al., 2012; Scherer et al., 2014). Ongoing studies of the co-evolution of VRC42, as well as VRC43 and VRC46, with both virus and the other autologous bNAb lineages are still necessary to fully understand how these early mutations were selected in this individual.

One potential obstacle when considering the elicitation of bNAbs to MPER as a vaccine strategy is autoreactivity. HIV bNAbs are frequently autoreactive/polyreactive (Liu et al., 2015), and there is molecular mimicry between some retroviral components and autoantigens (Sekigawa 2002). It has been shown that chronically infected individuals developing neutralization breadth have higher frequencies of autoantibodies compared to individuals who do not develop breadth (Moody et al., 2016). While VRC42 and VRC43 were not particularly autoreactive, VRC46 intermediate and mature antibodies bound strongly to HEp2 cells and cardiolipin. In contrast, VRC42.UCA had no detectable autoreactivity in HEp2 and cardiolipin assays, suggesting that VRC42 B cell development could be initiated by vaccination strategies without tolerance mechanisms that would block the development of these antibodies.

In immunogen design, it is advantageous to target an antibody modality that has been elicited by multiple independent viruses. The genetic and structural similarity of VRC42 and 4E10 are consistent with the notion that the donors of these antibodies found the same solution to the problem of neutralizing HIV. Notably, 4E10 and VRC42 arose independently, one in a clade B infected European patient in the 1980-90s, the other in the 2010s in a CRF01-AE infected Thai patient. The observation that this class has arisen multiple times gives encouragement to vaccine efforts targeting the C-terminal MPER epitope.

We have identified MPER-directed bNAb lineages from natural infection, with implications for vaccine design that expand the arsenal of modes of recognition that are able to target MPER. By elucidating some of the early events in development of MPER-targeted antibodies, and identifying a candidate priming immunogen to elicit them, this work has yielded insights that may guide efforts to elicit this important class of bNAbs.

STAR+METHODS

Detailed methods are provided in the online version of this paper and include the following:

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by Nicole Doria-Rose (nicole.doriarose@nih.gov).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human Subjects

Peripheral blood mononuclear cells (PBMCs) and sera were obtained from a male HIV-infected donor RV217.40512, approximate age 20, who signed informed consent and participated in MHRP protocols approved by Thai and Walter Reed Army Institute of Research (WRAIR) Institute Review Boards (Robb, et al, 2016). RV217 was an Africa and Thailand cohort that prospectively enrolled high-risk participants with negative results on an enzyme immunoassay for antibodies during the surveillance phase. The general study design has been previously described (Robb et al., 2016). Briefly, participants underwent small-volume blood collections by fingerstick measurements twice weekly and large-volume blood collections every 6 months. Small-volume blood samples were tested for HIV-1 RNA within 24 to 48 hours after collection. Following positive HIV-1 RNA tests, large-volume blood samples were obtained twice weekly for the first 4 weeks after infection after which volunteers were enrolled in long-term follow-up. 72 of the infected individuals that were enrolled in the study for at least 1 year in the absence of antiretroviral therapy and their plasma was tested for neutralization capabilities at different time points throughout infection for the ability to neutralize a large panel of HIV pseudotyped viruses. Donor RV217.40512 was identified in Fiebig stage I in that the donor had a nonreactive RNA test 4 days prior to the first positive RNA, and thus was identified within the first week of detectable viral load, where peak viral load occurred at 15 days and nadir was reached 29 days pRNA⁺. RV217.40512 was identified as having neutralization breadth of (85%) 646 days following the first RNA positive test (pRNA⁺).

Cell lines

Human embryonic kidney (HEK)-derived 293T and HEK293S N-acetylglucosaminyltransferase I-negative (GnTI-) cells were obtained from the American Type Culture Collection (ATCC), and HeLa-derived TZM-bl reporter cells were acquired through the NIH AIDS Reagent Program (ARP). The sex of these cell lines is unknown. HEK293T, HEK293S GnTI-, and TZM-bl cells were maintained in complete Dulbecco’s Modified Eagle Medium (herein referred to as cDMEM) containing high glucose Dulbecco’s Modified Eagle Medium (DMEM, Thermo Fisher), 1X Penicillin-Streptomycin (Pen Strep, Thermo Fisher) and 10% fetal bovine serum (FBS, Gemini Bio Products) at 37°C/5% CO₂. Expi293F cells (Thermo Fisher), were maintained in Expi293 Expression Medium (Thermo Fisher), at 37°C/10% CO₂ with shaking at 120 RPM.

METHODS DETAILS

Viral load and SGA

Serological testing of RV217 patients was described previously (Robb et al., 2016). Briefly, plasma HIV-1 RNA levels were measured in batches with the use of RealTime HIV-1 Assay (m2000 RealTime System, Abbott Molecular). HIV serologic testing with the use of standard diagnostic methods was performed at screening, every 6 months, and immediately following a positive HIV result. HIV-1 near full-length genomes or two overlapping half genomes were sequenced from longitudinal plasma RNA using a single genome amplification strategy (Salazar-Gonzalez et al., 2008). HIV-1 subtype was assigned using the HIV-1 Genotyping Tool at the National Center for Biotechnology Information and confirmed by Maximum Likelihood phylogenetic analysis and jumping profile hidden Markov model (jpHMM) (Rozanov, et al 2004; Schultz,et al 2009).

Production of pseudoviruses

Pseudoviruses for use in TZM-bl neutralization assays were produced in 293T cells by cotransfection of a pSG3ΔEnv backbone plasmid and a full HIV-1 Env gp160-encoding plasmid (Li et al., 2005). Briefly, 2X10⁶ cells in 20ml cDMEM were seeded in T75 flasks the day prior to cotransfection. For transfection, 40μl of FuGene 6 reagent (Promega) was diluted into 800μl of room-temperature Opti-MEM I reduced serum medium (Thermo Fisher), followed by addition of 10μg of pSG3ΔEnv backbone plasmid. 3.3 μg of HIV Env plasmid was then added to the mixture, mixed, and incubated for 30 minutes at room temperature. Transfection mixture was then added to media of previously seeded 293T cells in the T75 flask and then distributed evenly on cells. The following day, media was replaced with 20ml fresh cDMEM. Virus was harvested the following day by filtering cell supernatants with 0.45μm Steriflip units (EMD Millipore) and aliquotted.

Neutralization of longitudinal plasma samples

Neutralizing activity in longitudinal RV217 samples was assessed with a recombinant virus assay that included a reference panel of full-length env pseudo-typed viruses previously selected to assess neutralization breadth and to map epitope specificity of plasma antibodies by the breadth of their neutralizing antibody response (Georgiev et al., 2013). HIV env DNA was obtained from the NIH ARP, James Binley (Torrey Pines Institute), David Montefiori (Duke University Medical Center) John Mascola (VRC, NIH/NIAID), and Dana Gabuzda (Dana Farber Research Center). 178 samples from 72 HIV-1 individuals who were infected for at least a year were evaluated for bNAbs starting 30 days post-infection and up to 6 years prior to the initiation of ART. Using a microneutralization assay, samples were analyzed against a panel of 34 viruses from multiple subtypes. Neutralization was measured by reduction of luciferase gene expression as previously described (Montefiori, 2005). Briefly, indicator virus was incubated with serial 4-fold dilutions of plasma samples in duplicate before being added to TZM-bl cells. After a 48-hour incubation, luciferase activities were measured using Britelite Plus Reporter Gene Assay System (Perkin Elmer) substrate solution. Neutralization activity was expressed as the reciprocal plasma dilution that resulted in 50% reduction (ID₅₀ ) of RLU. Luminescence was read using the Molecular Devices Paradigm LUM384 module and data was uploaded into MHRP Labkey analysis server to obtain ID₅₀ values. Positive neutralization was defined as 50% inhibition of infection of an HIV strain at ≥1:40 plasma dilution and less than 50% inhibition of infection of murine leukemia virus (MuLV) was detected. Breadth was calculated as the percent of the 34 virus panel neutralized by ≥1:40 plasma dilution.

mAb Neutralization

10μl of five-fold serially diluted mAbs in cDMEM was incubated with 40ul of diluted HIV-1 Env-pseudotyped virus and incubated for 30 minutes at 37°C in a 96-well CulturPlate (Perkin Elmer). 20 μl of TZM-bl cells (10,000 cells/well) with or without 70μg/ml DEAE-Dextran was then added and incubated overnight at 37°C. Each experiment plate also had a column of cells only (no Ab or virus) and a column of virus only (no Ab) as controls for background TZM-bl luciferase activity and maximal viral entry, respectively. Serial dilutions were performed with a change of tips at each dilution step to prevent carryover. The following day, all wells received 100μl of fresh cDMEM and were incubated overnight at 37°C. The following day, 50μl of Steadylite Plus Reporter Gene Assay System (PerkinElmer) was added to all wells, and plates were shaken at 600RPM for 15 minutes. Luminometry was then performed on a SpectraMax L (Molecular Devices) luminometer. Percent neutralization is determined by calculating the difference in average RLU between virus only wells (cells + virus column) and test wells (cells + plasma/Ab sample + virus), dividing this result by the average RLU of virus only wells (cell + virus column) and multiplying by 100. Background is subtracted from all test wells using the average RLU from the uninfected control wells (cells only column) before calculating the percent neutralization. Neutralizing plasma antibody titers are expressed as the antibody concentration required to achieve 50% neutralization and calculated using a dose-response curve fit with a 5-parameter nonlinear function.

mAb Neutralization against 208 viruses

Select monoclonal antibodies were assessed on a panel of 208 geographically and genetically diverse Env pseudoviruses representing the major subtypes and circulating recombinant forms (Kong et al., 2016). Assays were performed by microneutralization in an optimized and qualified automated 384-well format (Sarzotti-Kelsoe et al., 2014).

Neutralization Fingerprinting Analysis

Published neutralization data for a set of ~200 donor plasma samples (Hraber et al., 2014) was analyzed by using a next-generation neutralization fingerprinting algorithm (Doria-Rose et al., 2017). For a given sample, the approach compares a polyclonal neutralization pattern of a set of diverse viral strains to the neutralization patterns (or fingerprints) of a reference set of broadly neutralizing mAb specificities, to obtain an estimate of the contribution of each of the reference specificities to polyclonal neutralization (Doria-Rose et al., 2017; Georgiev et al., 2013). Reference broadly neutralizing antibodies were used as previously described (Pancera et al., 2014), with the VRC38 fingerprint as a separate category. We applied computational quality control metrics for filtering out plasma samples for which the predictions were deemed unlikely to be accurate. Specifically, these included metrics for: (i) predicting the presence of dominant novel specificities and (ii) computing a confidence score associated with the computational predictions for each given sample (Doria-Rose et al., 2017). Through this process, the initial set of samples was reduced to 80 samples, which were used for the analysis of Ab specificity frequency. Samples were predicted to have between 1 and 3 specificities from the reference set, and the overall frequency of observing the different reference specificities or pairs of reference specificities were analyzed.

Bioinformatic analysis of antibody resistance

Logo plots of VRC43-resistant (IC50>50) and sensitive (IC50<50) sequences were aligned and submitted to weblogo (https://weblogo.berkeley.edu/logo.cgi). The height of each letter is proportional to the frequency of occurrence at a given position. For amino acid frequency analysis, 208 sequences were analyzed. The resistance score for each possible amino acid at a given residue position was defined as: the ratio of its number of occurrences in VRC43.01-resistant sequences to its overall number of occurrences. A higher score indicates that the amino acid was preferentially found among resistant sequences, with a score of 1 indicating that the amino was found only among resistant sequences. Amino acids that occurred at least 3 times at the given position are shown.

B cell sorting for MPER-specific mAbs

Frozen PBMCs (~1X10⁷ cells) were thawed in cDMEM containing 50 U/ml of benzonase (Novagen), then washed in PBS, resuspended in LIVE/DEAD® Fixable Aqua Live/Dead Cell Stain (Thermo Fisher) and incubated for 30 minutes at 4°C. Cells were washed once in PBS and multicolor staining was performed using a panel of fluorophore-labeled mAbs directed to CD3 (APC-Cy7, clone SK7), CD8 (Brilliant Violet 711, clone RPA-T8), CD14 (Brilliant Violet 605, clone M5E2), CD19 (PE-Cy7, clone HIB19) and human IgG (fluorescein isothiocyanate [FITC], clone G18-145). All mAbs were obtained from BD Biosciences except for those directed to CD8 and CD14 (Biolegend) and the anti-IgG-Alexa 680 Ab (custom-conjugated at the Vaccine Research Center, National Institutes of Health, Bethesda, MD). Cells were also stained at the same time with fluorophore-tagged MPR.03 peptide with two different fluorophores, PE and Brilliant Violet 421. All staining was performed at 4°C for 30 minutes, followed by two PBS/10% FBS washes and resuspension in PBS. Cells labeled as CD3⁻/CD8⁻/CD14⁻/CD19⁺/IgG⁺⁻ memory B cells were gated for positive MPER peptide staining. Gated cells were singly index-sorted on a FACS Aria sorter (BD Biosciences) into 96-well PCR plates (Denville Scientific, Holliston, MA) containing 20μl/well lysis buffer consisting of 1U/μl RNAse OUT (Thermo Fisher), 0.3125% Igepal CA-630, 1X SuperScript III First-Strand Buffer and 6.25 mM dithiothreitol (DTT) provided with the Superscript III Reverse Transcriptase kit (Thermo Fisher). Data was collected using FACSDiva software (BD Biosciences). Indexed sort data was analyzed using FlowJo software (FlowJo, LLC, Ashland, OR).

Single Cell RT-PCR of Ig Genes

RT-PCR amplification of IgG heavy and light chain genes was performed as described previously (Doria-Rose et al., 2015). 96-well plates containing single sorted cells were initially frozen at −80°C and thawed to maximize liberation of cellular mRNA. Total mRNA in each well of the 96-well plates was reverse transcribed using 200U/well Superscript III reverse transcriptase (Thermo Fisher), 2 μL of 10mM dNTP mix (Bioline) and 3 μL of 150 ng/ml random hexamers (Gene Link) under the following cycling parameters: 42C 10 min, 25C 10 min, 50C 60 min, 94C 5 min, and 4C hold. First strand cDNA was amplified in a 2-step multiplex nested PCR in which the first step used either mixed Ig gamma (γ)-, kappa (κ)-, or lambda (λ)- chain-specific primers (listed in (Doria-Rose et al., 2015). 50 μL κ- chain amplification reactions were performed using the QIAGEN HotStarTaq Plus DNA Polymerase Kit as follows (all reagents from HotStarTaq Plus kit unless otherwise noted): 1X QIAGEN PCR Buffer, 200 mM dNTP mix (Bioline), 500 mM MgCl2, 1 μL of 50 mM 1st round forward primer mix (5'L-Vκ mix), 1 μL of 25 mM 1st round reverse primer (30°Ck 543), 2U HotStarTaq Plus, and 5 μL of first-strand cDNA, under the following cycling parameters: 95° C 5 min; 50 cycles of 95° C 30 s, 58°C 30 s, 72°C 1min; 72°C 7 min; and 4°C hold. 40 μL gamma-chain and lambda-chain amplification reactions were performed similarly as follows: 1X QIAGEN PCR Buffer, 125 mM dNTP mix (Bioline), 1.4mM MgCl2, 0.2 μL of 50 mM 1st round forward primer mix (G1, G2, or G3 mix for gamma- chain and 5'L-VL-RL mix for lambda-chain), 0.2 mL of 25 mM 1st round reverse primer (3'CgCH1 for gamma- chain and 3’Cl for lambda-chain), 2U HotStarTaq Plus, and 3 μL of first-strand cDNA. Gamma-chains were amplified under the following cycling parameters: 94 °C 5 min; 50 cycles of 94 °C 30 s, 58 °C 30 s, 72 °C 55 s; 72 °C 10 min; and 4 °C hold, where × = 54 °C for G1 mix, 48 °C for G2 mix, and 52°C for G3 mix. Lambda-chains were amplified under the following cycling parameters: 95 °C 5 min; 50 cycles of 95°C 30 s, 50 °C 30 s, 72 °C 1min; 72°C 7 min; and 4 °C hold. The second multiplex PCR step used mixed chain-specific primers that were complementary to regions slightly upstream and downstream of the first step forward and reverse primers, respectively. 50 μL gamma (γ)-, kappa (κ)- and lambda (λ)- chain amplification reactions were performed using the QIAGEN HotStarTaq Plus DNA Polymerase Kit as follows (all reagents from HotStarTaq Plus kit unless otherwise noted): 1X CoralLoad PCR Buffer, 200 mM dNTP mix (Bioline), 1XQSolution, 1 μL of 50mM 2nd round forward primer mix (50L-VH mix + 5xwl-VH mix for γ- chain, 50L-Vk-MS mix for κ- chain, and 50L Vl mix for λ- chain), 1 μL of 25 mM 2nd round reverse primer (3'IgGint for γ- chain, 3'Ck494 for κ- chain, and 3'XhoI Cl for λ- chain), 2U HotStarTaq Plus, and 3.5 μL 1st round PCR product. Gamma-chains were amplified under the following cycling parameters: 95 °C 5 min; 50 cycles of 95 °C 30 s, 58 °C 30 s, 72 °C 1min; 72 °C 7 min; and 4 °C hold. Kappa-chains were amplified under the following cycling parameters: 95 °C 5 min; 50 cycles of 95 °C 30 s, 52 °C 30 s, 72 °C 1min; 72 °C 7 min; and 4 °C hold. Lambda-chains were amplified under the following cycling parameters: 95 °C 5 min; 50 cycles of 95 °C 30 s, 60 °C 30 s, 72 °C 1min; 72 °C 7 min; and 4 °C hold.. 5 μL of each second round PCR product was loaded into 1% ethidium bromide-stained pre-cast gels (Embi Tec, San Diego, CA) and run at 120 V. Gels were visualized under ultraviolet light, and second round PCR products from wells where PCR-amplified bands were obtained were selected for sequencing by ACGT, Inc (Rockville, MD). Alignments of sequences to germline V-, (D-), and J-genes and junctional analyses were performed using IMGT/V-QUEST (www.imgt.org). Clones of interest were identified by several criteria, including germline mutations > 8%) indicating a history of affinity selection, CDRH3 loops of at least 17 amino acids, and by repeated recovery of related sequences in separate wells. Selected PCR products were cloned into pVRC8400 to construct heavy chain- and light chain-expressing plasmids for mAb expression and purification. All antibodies were expressed as IgG1; however, we were able to assess the original isotypes by examining the sequence at the 3’ end of the PCR amplicons, which contained 137 nucleotides of CH sequences.

Antibody expression and production (UCA, intermediates, and mature forms)

mAbs were expressed by cotransfection of Expi293F or Freestyle 293F cells (Thermo Fisher) with heavy chain- and light chain-expressing plasmids according to the cell manufacturer’s directions. Following a 6-day incubation, transfection mixtures were pelleted by centrifugation and filtered through 0.22 μm Stericup filter units (EMD Millipore). Filtered supernatant was applied to a column containing a 1ml bed of Protein A Sepharose Fast Flow (GE Healthcare, Chicago, IL) equilibrated with Pierce Protein A IgG Binding Buffer (Thermo Fisher). The column was washed with Protein A IgG Binding Buffer, and mAb was eluted with Pierce IgG Elution Buffer (Thermo Scientific) and collected in a 1:10 volume of 1M Tris pH 8 solution. Antibodies were buffer-exchanged in PBS using 10,000 MWCO Amicon Ultra-15 centrifugal filter units (EMD Millipore) over three rounds of spinning.

For Fab production, the cleavage site (GLEVLFQGP) for the 3C protease of human rhinovirus (HRV3C) was inserted into the hinge region of the Ab heavy chain expression plasmid. Following expression and purification of mAb, the protein was subjected to HRV3C protease. Products were purified by gel filtration, and Fab fragments were collected for use in assays.

Antibody binding to cell-surface expressed Env

Cell-surface Env binding was measured on transfected 293T cells. Plasmids were codon-optimized Env constructs generated by Blue Heron (Bothell, WA); gp160ΔCT are truncated after amino acid 709 to delete the cytoplasmic tail (CT). On day 1, 8 million 293T cells were cultured in 20 mL of complete DMEM, in 2×150 mm² Nunclon Delta treated tissue culture dishes. They were then incubated overnight at 37°C. Day 2, 50 μg each of Furin was diluted and the desired Env plasmids in 1 ml of RT prewarmed OptiMEM (Fisher). 120 μl of Trufect Max (United Biosystem) transfection reagent was added to a separate 1 ml aliquot of RT OptiMEM. 2 ml of the Trufect-OptiMEM mixture was added to each plasmid solution, mixed gently, and incubated at RT for 20 minutes. Afterward, a p1000 pipet was used to slowly add Trufect-plasmid in droplets, then rocked the plates back and forth to distribute the plasmid. These cells were incubated for two more days at 37°C. On day 4, the cells were harvested. The cells were then washed twice, stained with a ViViD live/dead stain (Life Technologies) for 30 min, and distributed to a 96-well V-bottom plate; two transfected plates were used for each 96-well plate. Cells were stained with five-fold serial dilutions of various antibodies starting at 100 μg/ml. The cells and antibodies were co-incubated for 30 minutes at RT, and then washed three times. After staining with α-Human PE secondary antibody (Jackson ImmunoResearch) diluted 1:200, the plates were incubated again for 30 minutes at RT, washed three times, and fixed in 1% PFA. For the antibody incubations and washes, a 2% FBS solution in PBS was used. After incubation overnight at 4°C, the cells were read using a 96-well HTS apparatus attached to an X-50 cytometer, and the data analyzed in FlowJo.

ELISA Binding Assays

Env ELISAs

ELISAs using gp140 trimers and gp120 monomers were performed on Reacti-Bind 96-well polystyrene plates (Pierce). Briefly, plates were coated with 2μg/ml of monomer/trimer in PBS overnight at 4 deg C. Following 6 washes with PBS-T and blocking for 1 hour at 37°C with B3T buffer: 8.8 g/liter NaCl, 7.87 g/liter Tris-HCl, 334.7 mg/liter EDTA, 20 g BSA Fraction V, 33.3 ml/liter fetal calf serum, 666 μl/liter Tween-20, and 0.02% Thimerosal, pH 7.4), 5-fold serially diluted sera or mAbs were added in B3T buffer. After a 1-hour incubation at 37°C and 6 washes with PBS-T, plates were incubated with HRP-conjugated anti-human IgG, FcGamma fragment-specific antibody (Jackson ImmunoResearch) diluted 1:10,000 in B3T buffer for 1 hour at 37°C. After 6 washes with PBS-T (1xPBS, 0.05% Tween-20), SureBlue TMB Substrate (KPL) was added, incubated for 10 minutes, and the reaction was stopped with 1N H₂SO₄ before measuring binding at 450nm.

Epitope mapping by peptide ELISA

Serial dilutions of each antibody were added to ELISA plates coated with wild type peptide CNWFDITNWLWYIRKKK, or peptides with single amino acids changed to ala. The A450 values were plotted and the area under the curve (AUC) was calculated in Prism software. Data are presented as fold change in AUC: the AUC for wild type divided by the AUC for each mutant. The results from three independent experiments are shown.

Binding to Founder peptide

Binding to peptide with sequence matched to RV217.40512 MPER was performed using biotinylated peptide KKKNEKELLELDKWASLWNWFDITKWLWYIKKK{Lys(Biotin)} (Genscript). Peptide was added to streptavidin plates (Streptavidin Coated High Capacity Plates) at 0.23 mM in B3T buffer in 100 ul per well and incubated for 2 hours at 37°C. Plates were washed and the assays proceeded as above.

Lipid Binding

Assays assessing the binding of mAbs to phospholipids or glycolipids were performed as previously described (Matyas et al., 2009). 1,1’,2,2’-Tetramyristoyl cardiolipin (Heart, Bovine), 1,2-Dimyristoyl-sn-glycero-3-phsophate (DMPA), 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-Dimyristoyl-sn-glycero-3-phosphoethanolamine (DMPE), 1,2-Dimyristoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (DMPG), 1,2-dimyristoyl-sn-glycero-3-phospho-L-serine (DMPS), L-α-Phosphatidylcholine extracted from chicken egg (Egg PC), L-α-Phosphatidylinositol extracted from soy (PI), L-α-Phosphatidylinositol-4-phosphate extracted from porcine brain (PIP), Sphingomyelin extracted from porcine brain, C18 Ceramide, Galactosyl Ceramide, Glucosyl Ceramide, GM3 ganglioside extracted from bovine milk, and Lactosyl Ceramide, and Sulfatides extracted from porcine brain were purchased from Avanti Polar Lipids. Lipids were dissolved in chloroform, methanol, or a chloroform:methanol mixture in order to prepare a 10 mM working stock. Lipids were further diluted in ethanol or methanol to a final concentration of 10 μM or for cholesterol 100 μM and 0.1 mL of the diluted lipid was added to each well. An ELISA was performed for each lipid as described (Matyas et al., 2009). Briefly, following overnight incubation at room temperature to evaporate the methanol or ethanol, the plates were blocked for 2 hours with 1% BSA Tris buffered saline, pH 7.4 or Casein PBS, pH 7.4 for cholesterol. mAbs were diluted to 2 μg/mL in the appropriate blocking buffer and were diluted 2-fold on the plate for 1 hour. Plates were then washed with Tris Buffered Saline and 0.1 mL of HRP-conjugated human secondary antibody (The Binding Site, Birmingham, UK) diluted 1:1000 was added to each well and allowed to incubate for an additional hour. Plates were then washed and 0.1 mL/well of ABTS substrate (KPL) was added, plates were read at A405 after an hour.

Preparation of MPER peptide liposomes

MPER peptide liposomes were prepared as described previously (Alam et al., 2007; Alam et al., 2009; Dennison et al., 2009; Dennison et al., 2011). POPC (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine), POPE (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine), DMPA (1,2-dimyristoyl-sn-glycero-3-phosphate) and cholesterol (molar ratio 45:25:20:1.33) in chloroform purchased from Avanti Polar lipids, were mixed at room temperature with and without MPER peptide (MPER656-GTH1 peptide, QQEKNEQELLELDKWASLWNWFNITNWLYIKYKRWIILGLNKIVRMYS), dried under nitrogen, and placed in a vacuum overnight. The peptide-liposome film was reconstituted in PBS, pH7.4, so that the final peptide concentration was 0.25 mg/mL. The liposomes without the peptide (blank liposome) were also reconstituted in PBS, pH7.4. The liposomes were incubated at 50-60°C for 1-2 hours, vortexed and then sonicated prior to extrusion. After sonication, liposomes were extruded through a 0.4 μm filter and then a 0.2 μm filter using the Lipex Thermobarrel Extruder. Liposomes were used within 1week of extrusion.

Liposome binding by Bio-layer Interferometry (BLI)

VRC antibodies were tested for MPER656-GTH1 liposome and blank liposome binding by BLI using the ForteBio OcteRed96. Liposomes were prepared at a 1:20 dilution in PBS and captured on APS (aminopropysilane) sensor tips to a level of approximately 1.0-1.5nm.The duration of liposome capture was 600s. The liposome loaded sensor tips were then coated with 0.01% BSA for 600s to block any non-specific interaction of the antibody with the sensor tips. The sensor tips were then washed with PBS for 120s. After washing with 1xPBS, the sensor tips were dipped into the antibodies diluted down to 20 μg/mL and 1 well of the MPER mAb 13H11(Alam et al., 2007) also diluted to 20 μg/mL for an association length of 600s. 13H11 was used as a negative control mAb for MPER peptide liposomes since it does not bind to MPER peptide in the context of lipids or to peptide-free liposomes (Alam et al. 2007). After antibody binding, the sensors were placed back into PBS for a dissociation length of 600s. Antibody binding analysis was performed using the ForteBio Data Analysis 10.0 software. The Y-axis was aligned to the baseline from 175s to 179.8s and the inter-step correction was aligned to dissociation. 13H11 mAb (Alam et al., 2007) binding was subtracted from the binding of each antibody to exclude background and signal drift.

Cardiolipin ELISA

MAb binding to cardiolipin was tested by ELISA per the manufacturer’s protocol (Inova Diagnostics, San Diego, CA). Starting at 100μg/ml, mAbs were tested in a 3-fold series. Assays were validated using positive and negative controls and standards provided in the kit. OD values were converted to IgG anti-phospholipid (GPL) units by linear regression. 4E10 and VRC01 were used as additional positive and negative controls. Cardiolipin binding was scored as follows: no binding for <15 GPL, indeterminate for 15-20 GPL, low positive for 20-80 GPL and high positive for >80 GPL.

HEp-2 Cell Staining

Autoreactivity staining assays were performed on HEp-2 cells per the manufacturer recommendations (Zeus Scientific, Branchburg, NJ). MAbs were diluted to 50 and 25 μg/ml using SAVe Diluent. 20μl of the appropriate dilution was coated onto cells fixed on the slide and incubated for 30 minutes at room temperature in a humidified chamber. Slides were rinsed in 1X PBS, washed twice in 1X PBS in Coplin jars for 3-5 minutes and then stained with 20μl of FITC-conjugated secondary antibody for 30 minutes in a humidified chamber. Slides were rinsed in 1X PBS, washed twice in 1X PBS in Coplin jars for 3-5 minutes and mounted with 15μl of mounting media per well and a cover glass (Thermo Scientific). Slides were imaged on a Nikon Eclipse E800 microscope at 20X in the RGB mode for 2 seconds using SPOT 5.0 software (SPOT Imaging, Sterling Heights, MI). VRC01, 4E10, VRC07-523 and VRC07-G54W were used as control mAbs for staining and given a score of 0, 1, 2 and 3 respectively based on their staining intensity. Test antibodies were assigned scores based on visual comparisons of staining intensity to the control antibodies.

Peptide synthesis and peptide-carrier protein conjugate preparation

The HIV-1 membrane proximal external region peptide S₆₆₈LWNWFDITKWLWYIK₆₈₃GGC (MPER) matching the RV217.40512 founder MPER sequence was synthesized (GenScript) and conjugated to the carrier protein keyhole limpet hemocyanin (KLH) (Thermo-Scientific) using m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) as previously described (Xu et al., 2018).

Preparation of oligomeric T117-F

T117-F was genetically fused to the N-terminus of SpyCatcher (Zakeri et al., 2012) through a double gly linker, and a signal peptide “MGWSCIILFLVATATGVHS” was added to the N-terminus of T117-F for secretion to the culture supernatant; a thrombin site, 8xHis-tag and Strep-tag were added to the C-terminus of SpyCatcher for purification. An expression plasmid of this construct was transiently transfected in HEK293 Expi293F cells (Thermo Fisher) with 1 mg of plasmid diluted in 50ml Opti-MEM medium (Thermo Fisher) per 1L of cell culture. 3ml of Turbo293 transfection reagent (Speed BioSystems) was diluted into 50ml Opti-MEM medium and incubated for 5 minutes. Diluted transfection reagent was then added into diluted DNA and incubated for 15 minutes. The transfection reagent and DNA complex were added into 800ml of Expi293 cells at 2.5 million/ml. On the second day of transfection, 25ml of CelBooster and 25ml of ProBooster (ABI Scientific) together with EXPI293 expression medium were added to make up to 1L of cell cultures. Flasks were returned to a shaker incubator for an additional 5 days at 120rpm, 37°C and 9% CO2. Culture supernatant were affinity purified by cOmplete His-Tag Purification Resin (Roche) and eluted in 500mM of imidazole in 1xPBS. The purification tag was removed by adding 1.32U of thrombin (Novagen) per 1mg of T117-F-SpyCatcher fusion protein and incubated at 4°C for overnight before subjected to size-exclusion chromatography. SpyTag-Lumazine Synthase (SpyTag-LS) fusion protein was constructed by adding a signal peptide and SpyTag sequences to the N-terminus of lumazine synthase (LS) from Aquifex aeolicus. To facilitate LS expression in mammalian cells, we added a N-glycosylation site and mutated a N-glycosylation site. SpyTag-LS was expressed and purified with the same procedure as T117-F-SpyCatcher. Excess amount of the tag-free T117-F-SpyCatcher fusion protein was mixed with SpyTag-LS and incubated at room temperature overnight before purification of the conjugation product by size-exclusion chromatography. The resulting complex is formally named T117-F-SC-ST-lumazine synthase and called T117-F LS 60mer throughout the manuscript.

Antigenic characteristics of MPER peptide immunogens using Biolayer Interferometry (BLI)

Antigenic characteristics of KLH-coupled MPER peptide and MPER scaffolded nanoparticle (T117-F LS 60-mer) to 10E8 and VRC42.UCA antibodies were assessed with Biolayer Interferometry on an Octet RED384 (ForteBio) instrument. A series of scouting experiments using different buffer conditions and concentration ranges (1μM to 0.95nM for TF-MPER-KLH and 500nM to 0.8nM for T117-F LS 60-mer) were performed to determine suitable experimental conditions. Experiments were performed at 30°C in tilted black 384-well plates (Geiger Bio-One) with agitation set to 1,000 rpm in HBS-EP+ or 1x kinetic buffer obtained from ForteBio and a well volume of 65 μl. Variability in loading within a row of eight tips did not exceed 0.1 nm for each of these steps. The two IgGs at 30 μg/ml in PBS were immobilized for 300s on anti-human IgG Fc capture (AHC) biosensors tips and equilibrated for 60s in HBS-EP+ or 1x kinetic buffer prior to measuring association with TF-MPER-KLH (1.25nM to 40nM) or T117-F LS 60-mer (0.8nM to 50nM) in solution for 300s; the multimeric antigens were then allowed to dissociate for 600 s. Parallel correction to subtract systematic baseline drift was carried out by subtracting the measurements recorded for a loaded sensor incubated in HBS-EP+ or 1x Kinetic buffer. Data analysis and curve fitting were carried out using BLI software, version 9.0. The datasets were plotted using PRISM (PRISM 7 GraphPad Software). Of note, the off-rates for MPER-KLH and T117-F LS 60-mer were extremely slow, possibly due to multivalent interactions, such that kDs could not be calculated accurately.

T117-F Scaffold was coated onto a His1K Octet biosensor surface (ForteBio, Menlo Park, CA) at 10μg/ml for 60 seconds, followed by a baseline for 60 seconds and an association step with saturating concentrations of various mAbs (50 μg/ml of IgG for 30 minutes or 50 μg/ml Fab fragments for 45 minutes). Response values from the last 30 seconds of the association step were averaged to obtain the saturation binding. Mean binding response and standard deviations from 3 independent experiments were plotted against the known number of binding sites occupied by each mAb or mAb cocktail by regression analysis to generate a standard curve for trimer occupancy.

Generation of B cell lines

B cell lines expressing germline-reverted monoclonal antibodies were generated by transduction of surface IgM-negative Ramos cells as previously described (Weaver et al., 2016). Briefly, DNA encoding for VDJ (heavy chain) and VJ (light chain) immunoglobulin regions was synthesized by GenScript and cloned, along with the consensus human IgM C region, into the pLVX-ZsGreen and pLVX-mCherry expression vectors (both from Clontech) respectively. Each of these vectors was then co-transfected with the lentivirus packaging plasmid psPAX2 and with the VSV-G envelope expressing plasmid pMD2.G (both from Addgene) into HEK293 T cells for formation of lentiviral particles using Lipofectamine 2000 (Thermofisher). Supernatant was harvested 3 days after transfection and cleared by centrifugation before IgM-negative Ramos cells were co-transduced for both heavy and light chains. After 5 days, transduced Ramos cells were enriched by cell sorting for expression of ZsGreen and mCherry using a FACS AriaII interfaced to the FacsDiva software (BD Biosciences). This procedure was repeated until a pure double positive population could be selected. Cell lines were further enriched for high IgM surface expression by using a fluorescently labeled anti-human IgM monoclonal antibody.

Calcium Flux

Ramos cells were stained with FuraRed calcium indicator dye (Invitrogen) at 0.35 μl dye per 1 million cells at 10 million/ml in serum-free DMEM, for 30 minutes at room temperature in the dark (Weaver et al., 2016). Cells were washed in 10x volumes and resuspended at 1 million cells/300 microliters, both in serum-free DMEM, aliquoted to 4 ml tubes at 300 μl per tube, and kept dark and at room temperature until use. To trigger the BCR, one tube at a time was warmed to 37°C for 2 minutes in a water bath for 2 minutes, and then loaded on a BD Fortessa X-50 flow cytometer. After ensuring the threshold rate was stable at around 500 events/second, cells were acquired for 20 seconds. Then, while still acquiring, the tube was removed to add stimulus, vortexed and put back on the cytometer, with acquisition continuing for a total of 3 minutes. Stimuli included: 10 ug MPER-KLH, 600 ng T117-F-Spycatcher (T117-F-SC) monomer, 600 ng lumazine synthase-Spytag 60mer (LS 60mer), and 600 ng T117-F-LS 60mer. As a positive control, 12 μl of anti-IgM (Anti-Human unlabeled Mouse F(ab’)2 Anti IgM, SouthernBiotech, Cat#: 9023-01) was added. To ensure IgM expression, each Ramos cell line was also stained with anti-IgM (Brilliant Violet anti-human IgM Antibody, BioLegend). Unstimulated tubes were removed from the cytometer during acquisition, vortexed, and replaced without adding any substances. Data was analyzed in FlowJo, with gating of live cells by forward scatter-side scatter, exclusion of doublets via FSC-H by FSC-A, and measurement of calcium-flux signal in V655 and B710 detectors (Weaver et al., 2016). Expression of VRC42.UCA IgM was also confirmed via the B515 and G610 channels, for heavy and light chain, respectively. Change in ratio of bound to un-bound calcium over time was assessed using the Kinetic platform in FlowJo.

Immunoglobulin Repertoire Library Preparation and Next Generation Sequencing

RNA was isolated using Qiagen RNeasy kits (Valencia, CA). Reverse transcription (RT) was performed using Clontech SMARTer cDNA template switching: 5’ CDS oligo(dT) (12 μM) was added to RNA and incubated at at 72°C for 3 minutes and 4°C for at least 1 minute. The RT mastermix (5x RT Buffer (250 mM Tris-HCl (pH 8.3), 375 mM KCl, 30 mM MgCl₂), Dithiothreitol, (DTT 20 mM), dNTP Mix (10 mM), RNAse Out (40U/μL), SMARTer II A Oligo (12 μM), Superscript II RT (200U/gL)) was added to the reaction and incubated at 42°C for 90 minutes and 70°C for 10 minutes. First-strand cDNA was purified using AMPure XP beads (Beckman Coulter catalog# A63882). Following RT, two PCR rounds were carried out to generate immunoglobulin amplicon libraries compatible with Illumina sequencing. All oligos were ordered from Integrated DNA Technologies. Separate reactions were carried out for IgM and IgG heavy chains; the primer sets did not allow us to distinguish between the subtypes of IgG. The first PCR amplification was carried out using KAPA Real-Time Library Amplification Kit (catalog# KK2702). cDNA was combined with master mix (2X KAPA PCR Master Mix, 12 μM μL 5PIIA and 5 μL IgM/IgG/IgK/IgL Constant Primer (2 μM)). The amplification was monitored using real-time PCR and was stopped during the exponential phase. The amplified products were again purified using AMPure XP beads. A second round of PCR amplification was carried out for addition of barcodes and Illumina adapter sequences: master mix (2X KAPA PCR Master Mix 2x, Nuclease-free water), 10 μM of P5_Seq BC_XX 5PIIA, 10 μM of P7_i7_XX IgM/IgG/IgK/IgL and were combined with amplified Immunglobulin from the first round PCR and amplified using real-time PCR monitoring. The P5_Seq BC_XX 5PIIA primers contain a randomized stretch of four to eight nucleotides to increase diversity. This was followed by purification with AMPure XP beads. A final PCR step was performed for addition of remaining Illumina adapters by mixing master mix (2X KAPA PCR Master Mix, 10 μM P5_Graft P5_seq, Nuclease-free water), 10 μM of P7_ i7_XX IgM/IgG/IgK/IgL oligo and amplified products from the previous PCR step followed by purification with AMPure XP beads. The quality of library was assessed using Agilent Bioanalyzer. The amplicon libraries were pooled and sequenced on either an Illumina HiSeq 2500 in rapid run mode or an Illumina MiSeq as 2×300 paired-end runs.

Analysis of NGS data, inference of ancestral sequences, and identification of VRC42.N1

Paired end reads were merged and filtered for quality using USEARCH. Germline V and J genes were assigned using BLAST+ in SONAR (Schramm et al., 2016), discarding reads with missing assignments, out-of-frame junctions, and/or stop codons. To account for possible PCR and sequencing error, only V(D)J sequences supported by 5 identical reads in each technical replicate were used. For samples from days 85 and 646, which were not sequenced in replicate, V(D)J sequences were clustered at 99% nucleotide sequence identity and only clusters of 3 or more sequences were retained. Sequences related to the VRC42, VRC43, or VRC46 lineages were identified using identity-divergence plots as previously described (Doria-Rose et al., 2014; Schramm et al., 2016)and manually validated. Maximum likelihood phylogenetic trees were calculated and ancestral sequences reconstructed using DNAML in SONAR. Divergence from germline was calculated using MUSCLE in SONAR. While we were able to confidently assign the V, D, and J segments of the lineage, one nucleotide remained ambiguous, leading to a single amino acid change in the D-J junction; we therefore synthesized both versions: the most likely recombination event, termed VRC42.UCA, and an alternate version, termed alt-UCA (Figure S5A). The alternate UCA was also functional, with binding observed to T117-F (Figure S5B). One clade of the VRC42 heavy chain phylogenetic tree comprised of 8 sequences with an 18 amino acid CDR H3 and the 4e10-like GWGW motif. We selected 4 of these based on their representation in the raw data and combinatorially paired them with 3 VRC42-lineage light chain NGS sequences selected by phylogenetic pairing (Zhu et al., 2013) and VRC42.02-K. All 16 antibodies expressed and neutralized the autologous founder virus with IC50s ranging from 0.111-3.45 μg/mL. The most potent one was designated as VRC42.N1.

LC-MS/MS analysis

Plasma IgG was purified from donor RV217.40512 using Protein G Plus agarose affinity chromatography, and F(ab′)2 fragments were generated by digestion with 20 μg IdeS (Genovis) per 1 mg IgG at 37°C for 2 hr. Antigen-specific F(ab′)2 was isolated by affinity chromatography using biotinylated MPER peptide (founder virus) coupled to NeutrAvidin resin and eluted in 100 mM gly (pH 2.7). Each collected fraction was neutralized, and the protein-containing fractions were pooled and prepared for LC-MS/MS as described previously (Boutz et al., 2014). Briefly, the pooled F(ab’)2 fragments were concentrated under vacuum, resuspended in phosphate buffered saline, denatured with 50% (v/v) 2,2,2-trifluoroethanol and 2.5 mM dithiothreitol (DTT), and then incubated at 55°C for 45 min. The reduced samples were alkylated with iodoacetamide at room temperature for 30 min in the dark before quenching the reaction by addition of 20 mM DTT. The samples were diluted to 5% TFE in 40 mM Tris (pH 8.0) and digested with trypsin (trypsin/protein ratio of 1:75 at 37°C for 4 hours). The digestion was quenched by adding formic acid to 1% (v/v), concentrated under vacuum, and resuspended in 5% acetonitrile and 0.1% formic acid. Peptides were desalted using C18 Spin Tips according to the manufacturer’s protocol. Subsequently, the peptides were separated by reversed-phase chromatography (Dionex UltiMate 3000 RSLCnano System with Dionex Acclaim PepMap RSLC C18 column, Thermo Scientific) and analyzed online by nano–electrospray ionization–MS/MS on Orbitrap Velos Pro (Thermo Scientific). MS1 scans were collected in the orbitrap at a resolution of 60,000 Å, and the ions with >+1 charge were fragmented by collision-induced dissociation with up to 20 MS2 spectra collected per MS1.

The above methods were validated prior to testing of serum by testing an equimolar mixture of VRC42.01, VRC43.01, and VRC46.01 wth the full set of methods as described above. All three antibodies were successfully identified by these methods.

Computational interpretation of peptide mass spectra

The patient-specific target protein sequence database was constructed from the full-length VH and VL sequences from visit 20 (≥3 reads), Ensembl human protein-coding sequences, and common contaminants (maxquant.org). The spectra were then searched against this database using SEQUEST (Proteome Discoverer 1.4, Thermo Scientific) with previously described settings (Lavinder et al., 2014). The resulting PSMs were filtered with Percolator (Proteome Discoverer 1.4) at a false discovery rate of <1%; the average mass deviation (AMD) was calculated for all high-confidence PSMs, and only peptides with an AMD of <1.5 parts per million were included in the final data set. Informative peptides, as defined previously (Lavinder et al., 2014), were grouped by their CDR H1, CDR H2, and CDR H3 association, and for each group, the relative abundances of the corresponding peptide matches were determined by the sum of the extracted ion chromatograms of the respective precursor ions.

Crystallization, data collection, structure determination, and refinement

VRC42, 43, and 46 Fabs were incubated in a 1:1.5 molar ratio with T117-F, or in a 1:3 molar ratio with MPER peptides. The Fab-T117-F complexes were purified with size exclusion chromatography (GE Healthcare, HiLoad16/600, Superdex 200 pg) in 5 mM HEPES, 7.4, 150 nM NaCl. Initial crystallization conditions of VRC42, 43, and 46 lineage antibodies in complex with T117-F scaffold or MPER peptide (KKK₆₅₈NEKELLELDKWASLWNWFDIT KWLWYIK₆₈₃KK) were screened robotically by using commercially available Hampton Research, Precipitation Synergy, Wizard, and JCSG core I and II screens. The T117-F scaffold was constructed by replacing ₆₇₇NV₆₇₈ of T117v2 scaffold (Irimia et al., 2017) with KW and W₆₈₁ with Y. The initial conditions were hand optimized and the diffracting quality crystals were grown by hanging-drop vapor-diffusion method using protein solutions (~10 mg/ml) in 5 mM HEPES, pH 7.4, 150 mM NaCl with equal volumes of the reservoir solution at 20 °C. Crystals of VRC42.01 in complex with T117-F scaffold were grown in the reservoir solution containing 10% PEG 400, 2% PEG3350, 0.1M Na Acetate, pH 5.5. Crystals of VRC42.04 Fab in complex with MPER peptide grew in the reservoir solution containing 14% PEG 3350, 10% iso-propanol, 0.1M Tris, pH 8.5. Crystals of VRC42.N1 in complex with T117-F scaffold were grown in the reservoir solution consisting of 5% PEG400, 0.1M Tris, pH 8.5, 0.1M magnesium sulfate, 2M ammonium sulfate. Crystals of VRC43.01 Fab and VRC43.03 Fab were grown in the reservoir solutions containing 34% PEG4000, 0.2M ammonium sulfate, 0.1M sodium citrate, pH 5.6 and the reservoir solution consisting of 40% PEG 400, 8% PEG 8000, 0.1M HEPES, pH 7.4, respectively. Crystals of VRC46.01 Fab in complex with MPER peptide were grown in the reservoir consisting of 20% PEG3350, 0.2M potassium iodide. The crystals were soaked briefly in a cryoprotectant (the same reservoir solution with 30% PEG400 for VRC42.01 and VRC42.N1, with 30% glycerol for VRC43.01, with 15% glycerol, 15% ethylene glycol, 7.5% 2r,3r-butanediol for VRC42.04, VRC43.03 and VRC46.01) and flash frozen in the liquid nitrogen for data collection at SER-CAT ID22 beamline at APS (Argonne National laboratory, Chicago, IL). The data were processed by HKL2000 (Otwinowski and Minor, 1997) and the structures were solved by molecular replacement with PHASER(McCoy et al., 2007). The homology models were generated using SWISS-MODEL (Waterhouse et al., 2018)for molecular replacement. Iterative model buildings were carried out manually using Coot (Emsley et al., 2010). The refinement was performed with PHENIX(Adams et al., 2010). Molecular graphics were generated by PyMOL (PyMOL Molecular Graphics System, Version 2.0 Schrodinger, LLC).

MPER Peptide structural comparison

To quantify the differences in mode of binding between VRC42, VRC46, and the previously published MPER bNAbs, we compared the peptide structures as determined in antibody-peptide co-crystals. In a pairwise analysis, for each pair of MPER peptides we determined the distance between the alpha carbons of each contact residue. MPER peptide residues shown to be in contact with antibody in peptide-antibody co-crystal structures were superimposed in PyMOL software. The C_α - C_α root-mean-square-deviation (RMSD) of their overlap was measured.

QUANTIFICATION AND STATISTICAL ANALYSES

For all mAb/plasma pseudovirus neutralization assays, data were fitted to a 5-parameter asymmetric nonlinear regression model to obtain the IC50, or concentration of mAb/dilution of plasma needed to obtain 50% neutralization against a given pseudovirus. For neutralization assays in which a fold-change in IC50 imparted by a particular virus mutant, virus treatment, or binding stoichiometry (e.g., IgG vs. Fab) was reported, the IC50 obtained for one virus/assay condition was divided by the IC50 obtained for the other virus/assay condition, as indicated in the Figure Legends and y-axes of data graphs. All neutralization assays were repeated at least 2 times, and data shown are from representative experiments.

Multiple binding assays were performed to investigate VRC42.01 binding to HIV-1 Env trimers and monomers. All ELISAs were repeated at least two times, and curves shown are from representative experiments. BLI measurements were taken over 3 independent experiments, and the means and standard deviations of binding values were plotted.

DATA AND SOFTWARE AVAILABILITY

The founder env sequence for 40512v2 has been deposited in GenBank under ID code MK116905. The sequences for VRC42.01-VRC42.05, VRC42.UCA, VRC42.UCAalt, VRC42.I1-I3, VRC42.N1, VRC43.01, VRC43.I1, VRC46.01, and VRC46.I1 heavy chains and VRC42.01-VRC42.05, VRC42.UCA, VRC42.I1-I3, VRC42.N1, VRC43.01, VRC43.I1, VRC46.01, and VRC46.I1 light chain have been deposited in GenBank under ID codes MH605107-MH505138 respectively (see Key Resource Table). Raw NGS data, including metadata meeting the MiAIRR standard (Musen et al., 2015; Rubelt et al., 2017) has been deposited in the SRA under Bioproject PRJNA486355. The atomic coordinates and structure factors of VRC42.01:T117-F MPER scaffold, VRC42.04:gp41 peptide, VRC42.N1:T117-F MPER scaffold, VRC43.01, VRC43.03, and VRC46.01:gp41 peptide were deposited in the Protein Data Bank (PDB) under accession codes 6MTO, 6MTP, 6MTQ, 6MTR, 6MTS, and 6MTT, respectively.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Monoclonal anti-HIV Env 4E10	NIH AIDS Reagent Program;www.hiv.lanl.gov	Cat#10091
Monoclonal anti-HIV Env 10E8	NIH AIDS Reagent Program;www.hiv.lanl.gov	Cat#12294
Monoclonal anti-HIV Env 2F5	NIH AIDS Reagent Program;www.hiv.lanl.gov	Cat#1475
Monoclonal anti-HIV Env DH511.2	(Williams et al., 2017)	N/A
Monoclonal anti-HIV Env Z13.e1	(Zwick et al., 2001)	N/A
Monoclonal anti-HIV Env CH12	(Morris et al., 2011)	N/A
Monoclonal anti-HIV Env VRC38	(Kong et al., 2016)	N/A
Monoclonal anti-human CD3-APC-Cy7, clone SK7	BD Biosciences	Cat#557832
Monoclonal anti-human CD8-BV711, clone RPA-T8	Biolegend	Cat#301044
Monoclonal anti-human CD14-BV605, clone M5E2	Biolegend	Cat#301833
Monoclonal anti-human CD19-PE-Cy7, clone HIB19	BD Biosciences	Cat#560728
Monoclonal anti-human IgG-FITC, clone G18-145	BD Biosciences	Cat#555786
Monoclonal anti-human IgG-AlexaFluor 680, clone G18-145	Mario Roederer, NIH	N/A
HRP-conjugated anti-human IgG, Fcγ fragment Ab	Jackson ImmunoResearch	Cat#109-035-098
Purified mouse anti-human IgG Ab	BD Biosciences	Cat#555784
Human IgG (Fc), alkaline phosphatase (ALP)	Accurate Chemical and Scientific Corp.	Cat#JGH055008
Anti-human PE secondary antibody	Jackson ImmunoResearch	Cat#109-116-098
Brilliant Violet anti-human IgM Antibody	BioLegend	Cat#314516
Bacterial and Virus strains
E.coli DH5α
HIV-1 Env-pseudotyped viruses	John R. Mascola, NIH (Kong R et al, 2016)	N/A
RV217.40512 founder pseudotyped virus	This study	N/A
HIV-2 MPER chimeras	G. Shaw	(Davis et al., 2009)
Biological Samples
PBMC from MHRP RV217 donor 40512	(Robb et al 2016)	N/A
Plasma from MHRP RV217 donor 40512	(Robb et al 2016)	N/A
Chemicals, Peptides and Recombinant Proteins
MPR.03 peptide	(Williams et al., 2017)	N/A
MPER founder peptide	This study	N/A
RV217.40512 founder gp140 Env	This study	N/A
MPER-tm688	This study	N/A
MPER-tm694	This study	N/A
MPER-KLH	This study	N/A
T117-F Scaffold	This study	N/A
gp41 peptide 671-683	This study	N/A
RV217.40512 founder gp140 uncleaved trimer	Vincent Dussupt, MHRP (This study)	N/A
RV217.40512 founder gp120 monomer	Vincent Dussupt, MHRP (This study)	N/A
LIVE/DEAD® Fixable Aqua Dead Cell Stain	Thermo Fisher	Cat#L34957
Streptavidin, R-phycoerythrin (SA-PE)	Thermo Fisher	Cat#S866
Streptavidin-allophycocyanin (SA-APC)	Thermo Fisher	Cat#S868
RNAse OUT	Thermo Fisher	Cat#10777019
Random Hexamers	Gene Link	Cat#26-4000-03
10mM dNTP mix	Bioline	Cat#BIO-39053
EZ-Link Sulfo-NHS-Biotin	Thermo Fisher	Cat#21217
SigmaFAST p-nitrophenyl phosphate tablets	Sigma	Cat#N1891-5SET
SureBlue TMB Peroxidase Substrate	KPL	Cat#52-00-03
Streptavidin-alkaline phosphatase	Vector	Cat#SA-5100
Strep-Tactin alkaline phosphatase	IBA Life Sciences	Cat#2-1503-001
CompBead Anti-Mouse Ig, κ Compensation Particles	BD Biosciences	Cat#552843
DEAE-Dextran	Sigma	Cat#D9885-10G
Luciferase Cell Culture Lysis 5X Reagent	Promega	Cat#E1531
Steadylite plus Reporter Gene Assay System	Perkin Elmer	Cat#6066759
Protease Inhibitor Cocktail powder	Sigma	Cat#P2714
SigmaFast BCIP/NBT substrate	Sigma	Cat#B5655-5TAB
Nano-W Stain	Nanoprobes, Inc.	Cat#2018
cOmplete His-Tag Purification Resin	Millipore Sigma	Cat# 5893682001
1,1’,2,2’-Tetramyristoyl cardiolipin	Avanti Polar Lipids	Cat# 750332P
1,2-Dimyristoyl-sn-glycero-3-phsophate (DMPA)	Avanti Polar Lipids	Cat# 830845P
1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC)	Avanti Polar Lipids	Cat# 840345P
1,2-Dimyristoyl-sn-glycero-3-phosphoethanolamine (DMPE)	Avanti Polar Lipids	Cat# 850745P
1,2-Dimyristoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (DMPG)	Avanti Polar Lipids	Cat# 840445P
1,2-dimyristoyl-sn-glycero-3-phospho-L-serine (DMPS)	Avanti Polar Lipids	Cat# 840033P
L-a-Phosphatidylcholine extracted from chicken egg (Egg PC)	Avanti Polar Lipids	Cat# 840051P
L-a-Phosphatidylinositol extracted from soy (PI)	Avanti Polar Lipids	Cat# 840044P
L-a-Phosphatidylinositol-4-phosphate extracted from porcine brain (PIP)	Avanti Polar Lipids	Cat# 840045P
Sphingomyelin extracted from porcine brain	Avanti Polar Lipids	Cat# 860062P
C18 Ceramide	Avanti Polar Lipids	Cat# 860518P
Galactosyl Ceramide	Avanti Polar Lipids	Cat# 860544P
Glucosyl Ceramide	Avanti Polar Lipids	Cat# 860543P
GM3 ganglioside extracted from bovine milk	Avanti Polar Lipids	Cat# 860058P
Lactosyl Ceramide	Avanti Polar Lipids	Cat# 860545P
Sulfatides extracted from porcine brain	Avanti Polar Lipids	Cat# 131305P
Casein powder	Sigma	Cat# C3400-500G
1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE)	Avanti Polar Lipids	Cat# 850757C
1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC)	Avanti Polar Lipids	Cat#850457C
cholesterol	Avanti Polar Lipids	Cat# 700000P
ABTS substrate	KPL	Cat# 5120-0032
Dulbecco’s Modified Eagle Medium (DMEM)	Thermo Fisher	Cat# 11965126
Penicillin-Streptomycin	Thermo Fisher	Cat# 15140122
Fetal Bovine Serum (FBS)	Gemini Bio Products	Cat# 10438018
FuGene 6	Promega	Cat# E2692
Opti-MEM	Thermo Fisher	Cat# 31-985-062
Benzonase	Novagen	Cat# 70664-3
FuraRed Calcium Indicator dye	Invitrogen	F3021
Real-Time Library Amplification Kit	KAPA	Cat# KK2702
AMPure XP beads	Beckman Coulter	Cat# A63882
ANA HEp-2 Test System	Zeus Scientific	Cat# FA2400
Protein A Sepharose Fast Flow	Thermo Fisher	Cat# 97067-896
Protein A IgG Binding Buffer	Thermo Fisher	Cat# PI-21007
Trufect Max	United Biosystem	Cat#TM5501-4
SureBlue TMB substrate	KPL	Cat#5120-0077
Turbo293 transfection Reagent	Speed BioSystems	Cat# PXX1002
CelBooster	ABI Scientific	Cat# 2250
HotStarTaq Plus DNA Polymerase Kit	Qiagen	Cat# 203607
Crystallization reagents
Polyethylene glycol (PEG) 4000	Rigaku	Cat# 1008059
Polyethylene glycol (PEG) 6000	Rigaku	Cat# 1008061
Polyethylene glycol (PEG) 8000	Rigaku	Cat# 1008063
Sodium acetate	Rigaku	Cat# EB-250-NAAT
Ammonium sulfate	Rigaku	Cat# 1008358
MPD	Rigaku	Cat# 1008409
MES (pH 6.5)	Rigaku	Cat# 1008229
Tris buffer (pH 8.5)	Rigaku	Cat# 1008315
Sodium cacodylate trihydrate (pH 6.5)	Rigaku	Cat# 1008146
Sodium citrate (pH 5.6)	Rigaku	Cat# 1008027
Zinc acetate	Rigaku	Cat# 1008321
Calcium acetate hydrate	Rigaku	Cat# 1008142
Glycerol	Rigaku	Cat# 1008077
Cysteine-HCl	ThermoFisher Scientific	Cat# 44889
EDTA	Fisher Scientific	Cat# BP-2482-1
Isopropanol	Rigaku	Cat# 1008425
Deposited Data		Accession Number
40512v02 founder env sequence	(This study)	GenBank# MK116905
VRC42.01-H	(This study)	GenBank# MH605107
VRC42.01-K	(This study)	GenBank# MH605108
VRC42.02-H	(This study)	GenBank# MH605109
VRC42.02-L	(This study)	GenBank# MH605110
VRC42.03-H	(This study)	GenBank# MH605111
VRC42.03-L	(This study)	GenBank# MH605112
VRC42.04-H	(This study)	GenBank# MH605113
VRC42.04-L	(This study)	GenBank# MH605114
VRC42.05-H	(This study)	GenBank# MH605115
VRC42.05-L	(This study)	GenBank# MH605116
VRC42.UCA-H	(This study)	GenBank# MH605117
VRC42.altUCA-H	(This study)	GenBank# MH605118
VRC42.UCA-K	(This study)	GenBank# MH605119
VRC42.I1-H	(This study)	GenBank# MH605120
VRC42.I1-I2-K	(This study)	GenBank# MH605121
VRC42.I2-H	(This study)	GenBank# MH605122
VRC42.I3-H	(This study)	GenBank# MH605123
VRC42.I3-K	(This study)	GenBank# MH605124
VRC42.N1-H	(This study)	GenBank# MH605125
VRC42.N1-K	(This study)	GenBank# MH605126
VRC43.01-H	(This study)	GenBank# MH605127
VRC43.01-L	(This study)	GenBank# MH605128
VRC43.02-H	(This study)	GenBank# MH605129
VRC43.02-L	(This study)	GenBank# MH605130
VRC43.03-H	(This study)	GenBank# MH605131
VRC43.03-L	(This study)	GenBank# MH605132
VRC43.I1-H	(This study)	GenBank# MH605133
VRC43.I1-L	(This study)	GenBank# MH605134
VRC46.01-H	(This study)	GenBank# MH605135
VRC46.01-L	(This study)	GenBank# MH605136
VRC46.I1-H	(This study)	GenBank# MH605137
VRC46.I1-L	(This study)	GenBank# MH605138
NGS of IgM, IgG, Igκ, and Igλ variable region transcripts from 5 time points	(This study)	BioProject# PRJNA486355
Structure of VRC42.01:T117-F MPER scaffold	(This study)	Protein Data Bank (PDB)# 6MTO
Structure of VRC42.04:gp41 peptide	(This study)	PDB#6MTP
Structure of VRC42.N1:T117-F MPER scaffold	(This study)	PDB# 6MTQ
Structure of VRC43.01	(This study)	PDB# 6MTR
Structure of VRC43.03	(This study)	PDB# 6MTS
Structure of VRC46.01:gp41 peptide	(This study)	PDB# 6MTT
Experimental Models: Cell lines
Human: HEK 293T	ATCC	Cat#CRL-3216, RRID:CVCL_0063
Human: HEK 293S GnT1-	ATCC	Cat#CRL-3022, RRID:CVCL_A785
Human: HeLa-derived TZM-bl	NIH AIDS Reagent Program	Cat#8129-442, RRID:CVCL_B478
Human: FreeStyle 293F	Thermo Fisher	Cat#R79007
Human: Expi293F	Thermo Fisher	Cat#A14527
Oligonucleotides
66 human immunoglobulin VH-, Vκ-, and Vλ-specific primers	John R. Mascola, NIH (Doria-Rose et al. 2016)	N/A
IgG1 expression vectors	(Saunders et al., 2015)	N/A
1st round IgM Constant primer	(This paper)	AGGAGACGAGGGGGAAAAGGGTTGGGGCGGATG
1st round IgG Constant primer	(This paper)	GCCAGGGGGAAGACCGATGGGCCCTTGGTGGA
1st round Igκ Constant primer	(This paper)	GCGGGAAGATGAAGACAGATGGTGCAGCCACAG
1st round Igλ Constant primer	(This paper)	GGCCTTGTTGGCTTGAAGCTCCTCAGAGGAGGG
P5_Seq BC_xx 5PIIA	(This paper)	CACGACGCTCTTCCGATCT NNNN NNNNNNNN AAGCAGTGGTATCAACGCAGAGT
P7 i7_xx IgM	(This paper)	CAAGCAGAAGACGGCATACGAGAT NNNNNNNN AGGAGACGAGGGGGAAAAGGGTTGGGGCGGATG
P7 i7_xx IgG	(This paper)	CAAGCAGAAGACGGCATACGAGAT NNNNNNNN GCCAGGGGGAAGACCGATGGGCCCTTGGTGGA
P7 i7_xx IgK	(This paper)	CAAGCAGAAGACGGCATACGAGAT NNNNNNNN GCGGGAAGATGAAGACAGATGGTGCAGCCACAG
P7 i7_xx Igλ	(This paper)	CAAGCAGAAGACGGCATACGAGAT NNNNNNNN GGCCTTGTTGGCTTGAAGCTCCTCAGAGGAGGG
Software and Algorithms
FlowJo v.9.9.4	FlowJo, LLC	www.flowjo.com
IMGT/V-QUEST	International ImMunoGeneTics Information System; Marie-Paule Lefranc (Marie-Paule.Lefranc@igh.cnrs.fr), University of Montpellier, France	www.imgt.org
BLI Data Analysis software (v9.0)	Pall FortéBio	N/A
RCSB validation server		https://validate-rcsb-1.wwpdb.org/
RCSB deposition server		https://deposit-1.wwpdb.org/
Prism (v7.0)	GraphPad Software	www.graphpad.com
HKL2000	(Otwinowski and Minor, 1997)	N/A
Phaser	(McCoy et al., 2007)	N/A
ARP/wARP server	(Langer et al., 2008)	https://arpwarp.embl-hamburg.de/
Diffraction Anisotropy Server	(Strong et al., 2006)	https://services.mbi.ucla.edu/anisoscale/
RAPD@NECAT (Diffraction data processing)		https://rapd.nec.aps.anl.gov/
CCP4i	(Potterton et al., 2003)	N/A
COOT	(Emsley and Cowtan, 2004)	N/A
Pymol	DeLano et al., 2002	http://www.pymol.org
Phenix	(Adams et al., 2002)	N/A
SONAR	(Schramm et al., 2016)	N/A
BLAST+	(Camacho et al., 2009)	N/A
USEARCH (v9.0.2132)	(Edgar, 2010)	N/A
Muscle (v3.8.31)	(Edgar, 2004)	N/A
DNAML	(Felsenstein and Churchill, 1996)	N/A

Supplementary Material

1

Table S1. Neutralization against founder and heterologous Env-pseudotyped viruses by RV2017-40512 mAbs and plasma. Related to Figure 1.

2

Table S2. Neutralization of 208 Env-pseudoviruses by MPER-directed antibodies. Related to Figures 1 and 4.

3

Table S3. Alaine scanning to finely map MPER epitope. Related to Figure 2.

4

Table S4. Crystal data and structure refinement. Related to Figures 2 and 5.

5

Table S5. Contact residues and buried surface area for antibody-MPER peptide crystal structures. Related to Figures 2 and 5.

6

Table S6. Genetics and neutralization breadth of HIV-1 MPER-directed antibodies. Related to Figures 1, 3, 4, and 7.

7

Table S7. Sequences of isolated and inferred antibodies, and MPER protein scaffolds. Related to Figures 1, 3, 4, 6, and 7.

Highlights:

• Multiple MPER-directed bNAb lineages developed in a single individual

• The broadest lineage belongs to the same antibody class as the 4E10 antibody

• Low levels of somatic hypermutation of the RV217-VRC42 lineage can impart breadth

• A multimeric immunogen activates VRC42 precursor B cells