Dependence on a variable residue limits the breadth of an HIV MPER neutralizing antibody, despite convergent evolution with broadly neutralizing antibodies

Broadly neutralizing antibodies (bNAbs) that target the membrane-proximal external region (MPER) of HIV gp41 envelope, such as 4E10, VRC42.01 and PGZL1, can neutralize >80% of viruses. These three MPER-directed monoclonal antibodies share germline antibody genes (IGHV1-69 and IGKV3-20) and form a bNAb epitope class. Furthermore, convergent evolution within these two lineages towards a 111.2GW111.3 motif in the CDRH3 is known to enhance neutralization potency. We have previously isolated an MPER neutralizing antibody, CAP206-CH12, that uses these same germline heavy and light chain genes but lacks breadth (neutralizing only 6% of heterologous viruses). Longitudinal sequencing of the CAP206-CH12 lineage over three years revealed similar convergent evolution towards 111.2GW111.3 among some lineage members. Mutagenesis of CAP206-CH12 from 111.2GL111.3 to 111.2GW111.3 and the introduction of the double GWGW motif into CAP206-CH12 modestly improved neutralization potency (2.5–3-fold) but did not reach the levels of potency of VRC42.01, 4E10 or PGZL1. To explore the lack of potency/breadth, viral mutagenesis was performed to map the CAP206-CH12 epitope. This indicated that CAP206-CH12 is dependent on D674, a highly variable residue at the solvent-exposed elbow of MPER. In contrast, VRC42.01, PGZL1 and 4E10 were dependent on highly conserved residues (W672, F673, T676, and W680) facing the hydrophobic patch of the MPER. Therefore, while CAP206-CH12, VRC42.01, PGZL1 and 4E10 share germline genes and show some evidence of convergent evolution, their dependence on different amino acids, which impacts orientation of binding to the MPER, result in differences in breadth and potency. These data have implications for the design of HIV vaccines directed at the MPER epitope.


Introduction
The pursuit of an effective vaccine against HIV is an ongoing priority. It is generally accepted that an effective vaccine will require the elicitation of broadly neutralizing antibodies (bNAbs), capable of neutralizing multiple subtypes of HIV [1]. Recent results from the antibody-mediated prevention (AMP) trials demonstrated that passive infusion of the VRC01 bNAb prevented infection by viruses sensitive to this antibody re-invigorating the search for bNAbinducing vaccines [2]. However, eliciting bNAbs by vaccination has proven to be challenging because they develop only in~25% of infected donors even after many years and tend to have unusual features such as high levels of somatic hypermutation (SHM), long heavy chain or short light chain third complementarity determining regions (CDRH3s/CDRL3s) [3]. The identification of bNAb classes, which share common germline antibody genes and target the same region on the HIV envelope, has resulted in several germline-targeting vaccine strategies that aim to trigger unmutated common ancestors (UCAs) of bNAbs [4][5][6][7][8][9]. Studies defining bNAb/virus co-evolution during HIV infection have been invaluable in revealing the characteristics of early precursors/unmutated common ancestors (UCA), antibody intermediates, and the viral variants that engage and drive these lineages [10][11][12][13][14][15][16][17][18][19].
Clonally related sequences were detected at all time-points sequenced for both heavy (n = 75) and light chains (n = 53). Most of the clonally related reads were obtained at 10 wpi for both heavy and light chains, with fewer sequences detected at later time-points ( Fig 1B). Heavy and light chain sequences with no SHM, representing the UCA (light green) of this lineage, were detected at 10 wpi. Somatic hypermutation increased with the duration of infection, with the heavy chain being more mutated than the light chain, reaching an average of 15.8% and 10.3%, respectively ( Fig 1B).

Neutralization is largely mediated by CAP206-CH12 heavy chain
Though the CAP206-CH12 lineage light chain showed limited evolution, we identified a number of light chain sequences that were more mutated than CAP206-CH12. Two of these were selected for downstream analysis, K1 (observed at 120wpi) and K2 (observed at 198wpi) (Fig  2A and 2B). To test the role of the light chain in CAP206-CH12 neutralization, chimeras of the UCA and CAP206-CH12 heavy and light chains were tested against autologous viruses isolated within the first year of infection and compared to the wild-type UCA and CAP206-CH12 ( Fig 2C). Both the UCA and the UCA heavy chain paired with the CAP206-CH12 light chain showed no neutralization against any of the autologous viruses. In contrast, the CAP206-CH12 heavy chain paired with the UCA light chain showed some neutralization of autologous viruses, though with less potency than CAP206-CH12.
In an attempt to improve potency of CAP206-CH12 we paired K1 (6.4% SHM, light purple) and K2 (10.3% SHM, dark purple) with the CAP206-CH12 heavy chain. Neutralization of eight heterologous subtype B and C viruses (CAP85, COT6, Du156, Du422, QHO692, TRO.11, ZM135, and ZM197) was compared to the CAP206-CH12 wild-type ( Fig 2D). For the majority of these viruses, very little or no difference was seen between the CAP206-CH12 heavy chain paired with different light chains. For two of the viruses (Du156

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth and TRO.11), the most mutated light chain K2 showed reduced neutralization compared to CAP206-CH12.

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth numbering), known to interact with the MPER in 4E10 [27], to a lysine (K), resulted in a significant increase in potency (p = 0.04). Interestingly, replacing the 4E10 CDRL1 with that of CAP206-CH12 CDRL1 also resulted in a significant increase in potency compared to 4E10 wild-type (p = 0.03). Taken together this data suggests that, while contributing to neutralization through interactions mediated via the K 38 residue, the light chain does not play a dominant role in neutralization by CAP206-CH12.

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth " 111. 2

GW 111.3 " motif in the CDRH3 improves neutralization breadth and potency in CAP206-CH12 lineage
We, therefore, focused on the heavy chain maturation and specifically the CDRH3 since VRC42.01, PGZL1 and 4E10 all showed convergent evolution towards a 111.2 GW 111.3 motif, which is crucial for neutralization [14,20,21,27]. We observed convergent evolution within the CAP206-CH12 lineage towards the same motif, with the sequences seen in the later timepoints having mutated away from the ML motif seen in the UCA (grey) to either a 111.2 GL 111.3 (orange) or 111.2 GW 111.3 (red) (Fig 3A and 3B). Notably, none of the CAP206-CH12 lineage members contained a double 111 GWGW 111.3 motif as seen in 4E10, a 111 GAGW 111.3 as seen in VRC42.01 or a 111 GEGW 111.3 seen in PGZL1 (motif highlighted in grey, Fig 3B). In addition to the CDRH3 111.2 GW 111.3 motif, we observed maturation within other parts of the heavy chain, particularly within the CDRH1, FR2, and CDRH2, with CAP206-CH12.2 sharing the 25 SGGSFS 30 motif, crucial for binding for VRC42.01 and 4E10, though most CAP206-CH12 intermediates maintained the germline-encoded T 28 (Fig 3B).
Despite this, none of the lineage members showed the same degree of neutralization breadth or potency as 4E10, VRC42.01 or PGZL1 (GMT: 4.1, 2.8 and 6.9 ug/mL, respectively). To assess the role of mutations in the CDRHs, CDRH swaps were made between 4E10 and CAP206-CH12. However, these resulted in reduced or no neutralization of the chimeric antibodies compared to the wild-type 4E10 or CAP206-CH12 (Fig 4D), which may represent structural incompatibility between these regions. We noted that the CAP206-CH12 lineage reached a plateau in neutralization at 52% with most mAbs only able to neutralize 48% of viruses sensitive to the CAP206 plasma, from whom this lineage was derived. This suggests additional antibody lineages exist within this donor that account for the remaining plasma neutralization breadth.

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth

Multiple MPER lineages and specificities detected in donor CAP206 plasma
We performed a longitudinal assessment of CAP206 plasma neutralization of all 23 heterologous viruses neutralized by this donor and partitioned the viruses into those sensitive (green) and those resistant (black) to the CAP206-CH12 lineage (Fig 4A). Neutralization of lineage-sensitive viruses was detected from 15 wpi and persisted until 159 wpi. This is consistent with the timing of the CAP206-CH12 lineage, with the UCA detected from 10 wpi, isolation of mAbs  Magnetic beads coated with the MPR.03 peptide were used to adsorb MPER neutralizing antibodies from CAP206 plasma at 159 wpi (~3 years post-infection). We tested the ability of the adsorbed and unadsorbed plasma to neutralize both lineage-sensitive and lineage-resistant viruses (Fig 4B). Within the CAP206-CH12 lineage-sensitive viruses (top panel), the majority (7/11) showed reduced neutralization following depletion (green), confirming MPER specificity [24]. However, four of the viruses that are neutralized by mAbs in this lineage (UG024, 92RW, QH0962, and Q23) showed no reduction of neutralization following MPER-depletion, suggesting other, non-MPER, specificities dominate neutralization of these viruses.
Within the CAP206-CH12 lineage-resistant viruses, five (CAP63, WITO, CAP84, CAP244, and AC10) showed no reduction in neutralization following MPER-depletion, but four (CAP45, CAP88, CAP172, and Du151) did show reduced neutralization titres (Fig 4B, bottom  panel). This data indicates that at least two antibody specificities, one of which is MPERdirected and the other not, account for the neutralization of lineage-resistant viruses. Overall, this suggests that, in addition to the CAP206-CH12 MPER-directed lineage, donor CAP206 also has another MPER lineage as well as another bNAb with unknown specificity.
An alignment of the MPER sequences between the CAP206-CH12 lineage-sensitive (top panel) and lineage-resistant (bottom) viruses highlights a single amino acid at position 674 (highlighted in grey) that distinguishes these two groups (Fig 4C). Where lineage-sensitive viruses have a negatively charged D 674 compared to the uncharged S 674 or N 674 seen in the lineage-resistant viruses. This confirms [24] that the CAP206-CH12 lineage is dependent on D 674 while the additional MPER lineage within donor CAP206 is not.

CAP206-CH12 lineage is dependent on a highly variable MPER residue
In order to compare epitopes, we introduced mutations at position 674 and other sites within the MPER known to affect sensitivity to CAP206-CH12, PGZL1, VRC42.01, and 4E10 [14,20,24,27,28] into the COT6 virus which is sensitive to all four antibodies (Fig 5A). We compared the ability of these antibodies (including CAP206-CH12-GW) to neutralize the different mutant viruses compared to the wild-type (COT6).
Strikingly, CAP206-CH12 had distinct epitope dependencies (highlighted in green) compared to VRC42.01, PGZL1 and 4E10 (highlighted in black) (Fig 5A and 5B). Both CAP206-CH12 and CAP206-CH12-GW were highly dependent on the negatively charged aspartic acid (D) at position 674, with mutation of this residue to either an A 674 , S 674 , or N 674 resulting in complete knock-out of neutralization. N 674 had no effect on neutralization for VRC42.01, PGZL1 and 4E10. However, the S 674 or A 674 mutations resulted in reduced neutralization, with VRC42.01 and 4E10 most sensitive to S 674 (10-and 16-fold, respectively), and PGZL1 highly sensitive to A 674 (54-fold). Furthermore, VRC42.01, PGZL1 and 4E10 showed a more than 50-fold reduction or complete loss of neutralization when W 672 , F 673 , T 676 , or W 680 were mutated, whereas these mutations resulted in modest effects on CAP206-CH12 and CAP206-CH12-GW.
When comparing the frequencies of these residues among 7,510 MPER sequences, including all HIV group M subtype sequences, from the Los Alamos Database (www.hiv.lanl.gov), we noted that residues 672, 673, and 680, upon which VRC42.01, PGZL1 and 4E10 are dependent are highly conserved. In contrast, amino acid 674 upon which CAP206-CH12 and CAP206-CH12-GW show dependence is highly variable, with only 53% of global viruses containing an aspartic acid (Fig 5A and 5C). Furthermore, modeling of the residues which form part of the two MPER sub-epitopes recognized by 4E10/VRC42.01/PGZL1 versus CAP206-CH12 suggests that these antibodies likely target different aspects of the MPER. Although CAP206-CH12, VRC42.01, PGZL1 and 4E10 all target the MPER within the turn of the C-terminal helix [14,24,29], VRC42.01, PGZL1 and 4E10 target residues facing the hydrophobic core whereas CAP206-CH12 targets residues on the outer solvent-exposed surface of the helix (Fig 5D). Therefore, CAP206-CH12, VRC42.01, PGZL1 and 4E10 depend on different amino acids, and have different modes of binding to the MPER, likely accounting for the observed differences in breadth and potency. , and 4E10 dependent residues are shown in black) based on 7,510 MPER sequences, representing all HIV group M subtype sequences from the Los Alamos Database (www.hiv.lanl.gov). (D) Schematic representation of the MPER "turn" and "kink" structure [29] with key residues for the CAP206-CH12 lineage shown in green and those for VRC42.01 and 4E10 shown in black. https://doi.org/10.1371/journal.ppat.1010450.g005

Discussion
Antibodies CAP206-CH12, VRC42.01, PGZL1 and 4E10 all target the C-terminal helix of MPER of gp41 and use the same variable germline genes in both the heavy and light chains (IGHV1-69 and IGKV3-20) [14,20,24,28]. Despite these shared features, VRC42.01, PGZL1 and 4E10 are amongst the broadest antibodies (>80%), whereas CAP206-CH12 only displays 6% breadth [14]. Here we show, using antibody lineage sequencing, that the CAP206-CH12 lineage undergoes some convergent evolution with VRC42.01, PGZL1 and 4E10, introducing a 111.2 GW 111.3 motif within the CDRH3 region of the heavy chain. As with VRC42.01 and 4E10 [14], this 111.2 GW 111.3 motif improves neutralization potency within the CAP206-CH12 lineage, but not to the same levels as seen in VRC42.01, PGZL1 and 4E10. Within VRC42.01 and 4E10 the GW and GWGW motifs are not part of the antibody paratope but have been shown to enable conformational changes within the antibodies. This may contribute to overall neutralization breadth and potency, and these residues may perform a similar role in CAP206-CH12 [14]. We explored the mechanism for limited breadth and showed that CAP206-CH12 is dependent on a highly variable residue on the solvent-exposed surface on the elbow of the MPER whereas VRC42.01, PGZL1 and 4E10 are dependent on multiple conserved residues facing the hydrophobic core of the MPER. CDRL1 swaps between CAP206-CH12 and 4E10 resulted in improved neutralization in 4E10 and limited improvement for CAP206-CH12 with the K 38 alone significantly improving CAP206-CH12 neutralization potency. However, CDRH1 and CDRH3 swaps, known to be important in lipid viral membrane interactions in 4E10 [21], between 4E10 and CAP206-CH12 resulted in drastic reductions or no neutralization compared to CAP206-CH12. Thus, despite using the same germline genes and demonstrating convergent evolution, the dependence of these lineages on different residues within the same epitope and orientation of binding to the MPER likely determine neutralization breadth and potency.
The CAP206-CH12 lineage contains a string of hydrophobic tyrosine residues next to the 111.2 GW 111.3 motif which is encoded by IGHJ6. In contrast VRC42.01 and 4E10 use IGHJ4 which is much shorter and doesn't contain the string of hydrophobic tyrosine residues. The hydrophobic residues in the CAP206-CH12 lineage may cause a steric clash between the antibody heavy chain and the hydrophobic core of the MPER, perhaps driving this lineage to favor the external solvent exposed region. Thus, the pairing of the IGHV1-69 with IGHJ6 instead of IGHJ4 may have resulted in the CAP206-CH12 lineage initially engaging the MPER in a different orientation than that of VRC42.01 and 4E10 and thereafter maturation dependence on D 674 , leading to narrower neutralization.
We show that donor CAP206 contained multiple bNAb specificities. In addition to the CAP206-CH12 lineage, we have provided evidence for an additional MPER lineage as well as another bNAb specificity to an undefined epitope. We and others have previously described the development of multiple bNAb specificities within a single donor, which contribute towards the overall breadth of these donors [14,30,39,[31][32][33][34][35][36][37][38]. An effective vaccine would likely need to induce responses to multiple epitopes [2,40]. Therefore, studying donors that develop multiple bNAb lineages during infection may provide the insight needed for immunogen design for such strategies.
Neutralizing antibody responses to the MPER tend to be rarer than those targeting other epitopes on the HIV envelope [41]. It is therefore intriguing that both CAP206-CH12 and VRC42.01 were identified within donors that had multiple lineages targeting the MPER [14]. This suggests that there may be certain properties within the viral variants that infected the donors CAP206 and RV217.40512 (the donor from whom the VRC42 lineage was isolated) that gave rise to multiple MPER responses. Due to the highly conserved nature of the MPER,

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth bNAbs targeting this region tend to be among the broadest antibodies isolated to date [42]. Sequential immunogens designed based on variants identified in CAP206 and/or RV217.40512 may enable the development of such responses in a vaccine setting.
Overall, this data helps to explain why the CAP206-CH12 lineage, despite convergent evolution of variable heavy and light chain genes and similar epitope targeting, fails to develop the same level of breadth and potency as PGZL1, VRC42.01 and 4E10. Thus, while shared germline genes and convergent evolution within bNAb classes are crucial aspects of germline-targeting, they may not be sufficient to drive neutralization breadth.

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth selected from Illumina Indexing Kit V2 Set B were added to the pooled MiSeq amplicon libraries. All products were checked on an Agilent bioanalyzer High Sensitivity DNA kit (Diagnostech) and Qubit dsDNA HS assay (ThermoFisher Scientific) and cleaned-up using 0.75X Ampure Beads (Beckman-Coulter). A final concentration of 4.5 pM denatured DNA library with 10% PhiX control (Illumina) was run on the Illumina MiSeq, using the MiSeq reagent kit (version 3) with 2 x 300 paired-end reads.

CAP206-CH12 lineage analysis
Paired-end MiSeq reads were merged into full-length reads for each time point using PEAR [44]. Paired sequences were then de-replicated using USEARCH [45] resulting in unique reads. The SONAR bioinformatics pipeline [46] was used to identify clonally related reads to CAP206-CH12. In brief, germline V and J genes were assigned, the CDR3 regions were identified, and identity to CAP206-CH12 and level of somatic hypermutation for each of the reads was calculated. Clonally related sequences were selected based on germline gene usage and CDR3 identity �80% to CAP206-CH12. To account for sequencing and/or PCR error, all singletons were removed and the remaining sequences were clustered at 99% identity using USEARCH. A representative from clusters with a size of more than 1% of the most abundant read per time-point was selected for downstream analyses. Tools from the immcantation portal were used to reconstruct the antibody lineage and create the phylogenetic tree [47]. Briefly, clonally related heavy chain sequences were submitted to IMGT High V-quest [48,49] for analysis, and the resulting summary files were then parsed into a Change-O database. The UCA of the lineage was used as the germline-defined sequence for the reconstruction. Within R, the Alakazam package [47] was used to read in the Change-O database and reconstruct the lineage creating clones and introducing inferred sequences using buildPhylipLineage. The resulting tree was plotted using igraph [50]. The edge lengths of the tree were further modified to be proportional to the number of mutations between nodes and all nodes with no mutations from the parent node were deleted using Inkscape v 0.91.

CAP206-CH12 lineage antibody production
Selected clonally related heavy sequences were ordered codon-optimized from GenScript (Hong Kong, China) and sub-cloned in the CMV/R expression vector [51,52]. All heavy chain clonally related sequences (H1-H6), were paired with the mature CAP206-CH12 light chain (CH12 L). The UCA heavy chain was paired with the UCA light chain (UCA L). CAP206-CH12 was also paired with two additional light chains (K1 and K2). These heavy and light chain pairs were sub-cloned into an IgG1 backbone, co-transfected, and expressed in HEK293F cells (obtained from the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH) grown in FreeStyle 293 Expression Medium (Gibco, Thermo Scientific, MA, USA) at 37˚C, 5% CO2, 70% humidity and 125 rpm. Cultures were harvested after seven days by centrifugation at 4000 x g, and supernatants were filtered through 0.22 μm before purification by protein A chromatography.

PLOS PATHOGENS
Mode of recognition within MPER antibody class determines neutralization breadth #R0176L and NEBuilder HiFi DNA Assembly Master Mix cat # M552OAA) using primers that were synthesized by Inqaba Biotech. Mutations were confirmed by sequence analysis.

Neutralization assays
Neutralization was measured as a reduction in luciferase reporter gene expression after a single round of infection of TZM-bl cells with Env-pseudotyped viruses as previously described [53]. Briefly, monoclonal antibodies (mAbs) or heat-inactivated plasma/serum were incubated with pseudovirus for 1 hour, before the addition of cells for infectivity. Neutralization titres were calculated as the inhibitory antibody concentration (IC 50 ) or reciprocal plasma/serum dilution (ID 50 ) causing 50% reduction of relative luminescence units (RLU) with respect to the virus control wells (untreated virus). Starting concentrations of mAbs were at 200 μg/mL in order to determine a precise titre.

Anti-MPER activity
Streptavidin-coated magnetic beads (Dynal MyOne Streptavidin C1; Invitrogen) were incubated with the biotinylated peptide MPR.03 (KKKNEQELLELDKWASLWNWFDITNW LWYIRKKK-biotin-NH2) (NMI, Reutlingen, Germany) at a ratio of 1 mg of beads per 20 μg peptide at room temperature for 30 min. Sera were diluted 1:20 in Dulbecco's modified Eagle's medium (DMEM)-10% fetal bovine serum and incubated with the coated beads for one hour at a ratio of 2.5 mg of coated beads per mL of diluted plasma. This was followed by a second adsorption at a ratio of 1.25 mg of coated beads per mL of diluted sample. After each adsorption, the beads were removed with a magnet, followed by centrifugation, and were stored at 4˚C. The separated immunoglobulin G (IgG) was removed and placed into a separate tube, where the pH was adjusted to between 7.0 and 7.4 with 1 M Tris (pH 9.0) buffer. The same beads were acid eluted twice more. The pooled eluates were then diluted in DMEM, washed over a 10-kDa Centricon plus filter, and resuspended in DMEM. The adsorbed sera were then used in neutralization assays. The reduction (>32-fold) in neutralization by the MPERadsorbed plasma (green, top panel Fig 4B) compared to unabsorbed plasma (black) against C1C, an HIV-2/HIV-1 MPER chimera involving an HIV-2 backbone (7312A) with an HIV-1 MPER region, demonstrated the efficient removal of anti-MPER antibodies.