Rational Engraftment of Quaternary-Interactive Acidic Loops for Anti-HIV-1 Antibody Improvement

ABSTRACT

Broadly neutralizing antibodies (bNAbs) are the focus of increasing interest for human immunodeficiency virus type 1 (HIV-1) prevention and treatment. Although several bNAbs are already under clinical evaluation, the development of antibodies with even greater potency and breadth remains a priority. Recently, we reported a novel strategy for improving bNAbs against the CD4-binding site (CD4bs) of gp120 by engraftment of the elongated framework region 3 (FR3) from VRC03, which confers the ability to establish quaternary interactions with a second gp120 protomer. Here, we applied this strategy to a new series of anti-CD4bs bNAbs (N49 lineage) that already possess high potency and breadth. The resultant chimeric antibodies bound the HIV-1 envelope (Env) trimer with a higher affinity than their parental forms. Likewise, their neutralizing capacity against a global panel of HIV-1 Envs was also increased. The introduction of additional modifications further enhanced the neutralization potency. We also tried engrafting the elongated CDR1 of the heavy chain from bNAb 1-18, another highly potent quaternary-binding antibody, onto several VRC01-class bNAbs, but none of them was improved. These findings point to the highly selective requirements for the establishment of quaternary contact with the HIV-1 Env trimer. The improved anti-CD4bs antibodies reported here may provide a helpful complement to current antibody-based protocols for the therapy and prevention of HIV-1 infection.

IMPORTANCE Monoclonal antibodies represent one of the most important recent innovations in the fight against infectious diseases. Although potent antibodies can be cloned from infected individuals, various strategies can be employed to improve their activity or pharmacological features. Here, we improved a lineage of very potent antibodies that target the receptor-binding site of HIV-1 by engineering chimeric molecules containing a fragment from a different monoclonal antibody. These engineered antibodies are promising candidates for development of therapeutic or preventive approaches against HIV/AIDS.

INTRODUCTION

The global pandemic of human immunodeficiency virus type 1 (HIV-1) infection remains a major challenge for public health (1). Despite the success of combined antiretroviral therapy (cART), the development of potent and long-acting drugs for HIV-1 prevention and treatment is still a priority. New therapies are needed to achieve safe and effective long-term virus suppression, persistent cART-free virus control, or even virus eradication in chronically infected individuals (2, 3). Among the most promising candidates that have become available are broadly neutralizing antibodies (bNAbs) cloned directly from HIV-infected individuals. This field has seen a dramatic expansion over the past decade, with the cloning of an increasing number of bNAbs that target various vulnerable sites on the HIV-1 envelope (Env) glycoprotein, which is the sole target for neutralization (4, 5). A prime target region for antibody-mediated HIV-1 neutralization is the CD4-binding site (CD4bs) of gp120, which is relatively conserved and plays a key role in the viral life cycle. Two main categories of anti-CD4bs antibodies have been identified: VH gene restricted and CDRH3 dominated (6). VH gene-restricted antibodies originate from heavy-chain germ line genes VH1-2 and VH1-46, and the former, also known as VRC01 class, represents the largest class of potent members, including VRC07-523LS, N6, and 3BNC117 (7–9).

Several strategies, either based on structure-guided rational design or derived from empirical alterations, have been developed to further improve existing anti-CD4bs bNAbs (10). For example, a single mutation in the heavy-chain complementarity-determining region-2 (CDRH2) region, G54W, was reported to significantly increase the potency of bNAb NIH45-46 (11). Other substitutions at this position, like G54H, G54Y, and G54F, were used to improve the potent bNAb VRC07, combined with deletion of N-terminal residues within the light chain (8). Modifications of the Fc region were introduced to increase the antibodies’ in vivo half-life (12). Recently, we discovered that the CD4 receptor establishes a quaternary interaction with the HIV-1 Env spike and identified a second site of CD4 interaction with the Env trimer, located in the inner domain of an adjacent gp120 protomer, that we designated CD4-binding site 2 (CD4-BS2) (13). Moreover, we found that selected neutralizing antibodies mimic the quaternary binding mode of CD4, engaging simultaneously two neighboring gp120 protomers. Based on these observations, we developed a new approach for improving anti-CD4bs antibodies using a paratope engraftment strategy (14). Thus, we transplanted the elongated heavy-chain framework region-3 (FR3) loop of VRC03, which establishes quaternary interaction with a neighboring gp120 protomer of the Env trimer, onto other anti-CD4bs bNAbs. The engrafted loop indeed established new quaternary contacts with the adjacent gp120 protomer, as documented for VRC03, resulting in an increased neutralizing capacity associated with a reduced off-rate of binding. Somewhat unexpectedly, we also observed a prolonged in vivo half-life of the chimeric antibody compared to the original form. In addition, autoreactivity was either reduced or completely eliminated after FR3 chimerization (14). These properties made the chimeric antibodies better candidates for clinical use.

A new series of anti-CD4bs bNAbs with high potency and near-pan-neutralizing capacity against HIV-1 was recently isolated by Sajadi and colleagues from donor N49, a chronically infected individual with broad serum neutralization capacity (15). Antibodies N49P6, N49P7, and N49P9 are among the most potent members of this lineage. All three antibodies harbor an FR3 with regular length and seemingly bind only to the primary CD4bs on a single gp120 protomer. Thus, they are ideal candidates to test if paratope engraftment can further improve their potency and pharmacokinetics. To verify this hypothesis, we introduced the FR3 loop of VRC03 into all three antibodies and tested their biological properties. Additionally, we also attempted to improve anti-CD4bs bNAbs by engrafting another quaternary contact loop, the CDR1 region from the very broad and potent bNAb 1-18 (16). The results show that the FR3 transplant successfully improved the activity of two of the N49-series bNAbs, while the 1-18 CDR1 loop transplant was not effective, suggesting that paratope engraftment has stringent structural compatibility requirements. These results provide a new set of potent bNAbs for potential clinical use and confirm the broad applicability of FR3 engraftment to improve anti-CD4bs antibodies.

RESULTS

Potent neutralizing antibodies of the N49 lineage do not establish quaternary contact with the HIV-1 Env trimer.

To investigate whether anti-CD4bs antibodies of the N49 lineage naturally establish quaternary contact with the HIV-1 Env trimer, we aligned the structure of monomeric gp120 complexed with N49P7 (PDB entry 6BCK) or VRC03 (PDB entry 3SE8) to one gp120 protomer of an Env trimer structure (BG505 SOSIP.664; PDB entry 4TVP). While VRC03 makes contact with the neighboring gp120 protomer through a long FR3 loop, N49P7 resembles VRC01 and N6 in its binding mode, as it establishes extensive contact with a single gp120 protomer but has limited, if any, interaction with the neighboring protomer (Fig. 1A). Sequence alignment with members of the VRC01 class showed that antibodies from the N49 lineage have a regular-length FR3 region and do not possess the extended acidic loop that allows VRC03 and VRC06 to reach a second gp120 protomer (Fig. 1A and B), which makes them unable to establish quaternary contact. Considering the high potency and breadth of N49-series antibodies, we reasoned that we could further improve their properties by expanding their contact interface with the Env trimer using the same approach that we successfully employed to improve several anti-CD4bs antibodies by introducing the elongated FR3 loop from VRC03 (FR3-03).

FIG 1.

Antibodies of the N49-series are not projected to establish major quaternary contacts with a second gp120 protomer. (A) Structural alignment of gp120-complexed N49P7 (PDB entry 6BCK) and VRC03 (PDB entry 3SE8) with BG505.SOSIP.664 (PDB entry 4TVP). The closeup view shows the elongated FR3 loop of VRC03 (FR3-03). (B) Sequence alignment of the CDR2 and FR3 regions of VRC01-class antibodies.

Engraftment of the VRC03 FR3 increases the neutralization potency of N49P7 and N49P9.

We engrafted the elongated FR3 loop of VRC03 onto three N49 antibodies, namely, N49P6, N49P7, and N49P9, to generate three FR3-03 chimeras: N49P6 FR3-03, N49P7 FR3-03, and N49P9 FR3-03. All antibodies were expressed in HEK 293 free-style (FS) cells and purified by protein A affinity chromatography. Chimeric N49P7 and N49P9 seem to adopt correct folding like their original forms. In contrast, the reduced N49P6 FR3-03 chimera showed the presence of a few unexpected bands on SDS-PAGE, indicating either protein degradation or improper folding. To assess the neutralization capacity of the chimeric antibodies, we initially tested their activity on a small global panel of representative HIV-1 Envs using pseudoviruses in the TZM-bl assay. Both chimeric N49P7 FR3-03 and N49P9 FR3-03 showed an increased neutralization against 9 out of 12 isolates compared to their respective wild-type (WT) forms (Fig. 2A). In contrast, N49P6 FR3-03 did not fold correctly, as evaluated by SDS gel electrophoresis, making the results of neutralization tests unreliable. These results provide evidence that members of the N49 antibody lineage can be functionally improved by introducing quaternary trimer contact through epitope engraftment. In addition, the FR3-03 chimerization of N49P7 and N49P9 did not cause autoreactivity, as evaluated by antinuclear antibody staining of human epithelial (HEp-2) cells (Fig. 2B). The chimeric N49P6 showed a bit higher autoreactivity than the WT, which is likely due to its poor folding.

FIG 2.

(A) Neutralization of a small global panel of HIV-1 isolates by chimeric N49 antibodies. A TZM-bl assay was used to generate a dose-response curve for each pseudovirus, and the IC₅₀ (μg/ml) values were calculated. The fold change of the chimera’s neutralizing capacity relative to its respective WT is denoted by different red shades for increased neutralization and green shades for decreased neutralization. All neutralization assays were performed in duplicate wells. (B) Autoreactivity of chimeric N49 antibodies as evaluated by antinuclear antibody staining of human epithelial (HEp-2) cells. Three antibodies, VRC01, 4E10, and VRC07-523LS, were used as controls.

FR3-03 chimerization increases the potency of N49-series antibodies against a large panel of global HIV isolates of different clades.

To confirm the neutralization potency enhancement of the chimeric antibodies, we tested their neutralizing activity against a large panel of 119 global HIV-1 Envs from different genotypic subgroups (clades) and geographic origins. Chimeric N49P7 neutralized with a median half-maximal inhibitory concentration (IC₅₀) of 0.046 μg/ml, i.e., ∼3-fold lower than that of the WT (Fig. 3A and B). The improvement was even more dramatic for N49P9, with a median IC₅₀ for the chimeric form ∼3.5-fold lower than that of the WT. Remarkably, chimeric N49P7 could neutralize 5 isolates that were resistant to the WT N49P7 antibody, thereby also expanding the neutralization breadth (Fig. 3A).

FIG 3.

Increased neutralizing capacity of chimeric N49-series antibodies as evaluated on a large panel of global HIV-1 isolates. (A) Neutralization potency (IC₅₀, in μg/ml) of WT and FR3-03 chimeric N49 antibodies against a large global panel of 119 HIV-1 strains of different clades and circulating recombinant forms (CRF). The median IC₅₀ value against antibody-sensitive strains is indicated by the black horizontal line. The percentage of sensitive strains (IC₅₀, <50 μg/ml) is indicated at the top for each antibody. (B) Neutralization capacity of the antibodies denoted by IC₈₀.

Chimeric N49-series antibodies show increased binding to soluble HIV-1 Env trimers.

To explore the mechanism for antibody improvement upon FR3-03 chimerization, we tested the binding capacity of the N49 chimeras to soluble HIV-1 Env trimer in enzyme-linked immunosorbent assays (ELISA). Plastic plates were coated with Galanthus nivalis lectin to capture the BG505 SOSIP.664 trimer, and the antibody binding was detected using ELISA. Higher binding to trimer was observed with the chimeric form than with the WT for both N49P7 and N49P9 (Fig. 4A, upper). However, there was only a minimal increase of binding to another soluble trimer, JRFL SOSIP.664 (Fig. 4A, lower). We previously observed this phenomenon with other FR3-03 chimeras, such as VRC01 FR3-03 and N6 FR3-03 (unpublished data). This finding is in line with the relatively modest increase in neutralization of chimeric antibodies versus the WT against the JRFL Env (14). We also evaluated the ability of the chimeras to bind to monomeric gp120. Interestingly, for both BG505 and BaL gp120, the chimeric antibody showed a lower affinity than the WT (Fig. 4B). These results provide a mechanism to interpret the increased neutralization potency of chimeric antibodies and suggest that, by establishing quaternary contact with the Env trimer, chimeric antibodies adopt a trimer-preferring binding mode, which is less optimized for interaction with the gp120 protomer.

FIG 4.

Enhanced binding of chimeric N49 antibodies to soluble Env trimers. (A) Each pair of WT and FR3-03 chimeric antibodies was tested for binding to soluble BG505 SOSIP.664 and JRFL SOSIP.664 trimers by ELISA. (B) Binding of WT and chimeric N49P7 to BG505 and BaL gp120 monomers.

The engrafted FR3-03 loop interacts with the V3 base of a neighboring gp120 protomer.

Through structural alignment, the FR3-contacting region of VRC03 was mapped to the V3 base and part of CD4-BS2 of a neighboring protomer. We postulated that the engrafted FR3-03 on N49P7 and N49P9 should establish quaternary contact with the Env trimer in a similar manner. Thus, we created a series of Env mutants with single or double amino acid substitutions in V3 and CD4-BS2 and tested their sensitivity to the WT and chimeric antibodies. Both alanine mutants and charge inversions were generated. For WT N49P7, the neutralizing capacity was hardly altered by mutations in CD4-BS2, with IC₅₀ changes below 2-fold (Fig. 5). In contrast, some V3 mutants, like R304E, R308E, and Y318E, became less sensitive to the N49P7 FR3-03 chimera compared to the WT Env, with IC₅₀ increases between 3- and 10-fold, indicating that the V3 base is involved in contact with the engrafted FR3-03 loop. The effect of the alanine mutations was consistently less dramatic than charge inversions. The most consequential mutations were R304E and Y318E. For chimeric N49P9, V3 mutations had only a limited impact on neutralization sensitivity (Fig. 5), suggesting a weaker quaternary interaction between the FR3-03 loop and the neighboring gp120 protomer in the BG505 Env, as also illustrated by the limited improvement in neutralization capacity against BG505 pseudovirus (∼2-fold).

FIG 5.

Mapping of quaternary FR3-03-contact region on the Env trimer. The neutralization sensitivity of the WT and chimeric antibodies was tested against pseudoviruses bearing a series of BG505 Env mutants. The IC₅₀ fold change was calculated relative to the WT Env (set to 1). All neutralization assays were performed in duplicate wells.

Additional modifications further improve FR3-03 chimeric N49P7.

It has previously been demonstrated that a bulky residue, like tryptophan (W), histidine (H), or tyrosine (Y), at site 54 of the heavy chain of VRC01-class bNAbs could fill the Phe43 cavity on the HIV Env trimer and increase the neutralization potency. In N49P7, there is a glycine at position 54, while in N49P9 it is a tyrosine (Fig. 6A). N49P9 is slightly more potent than N49P7, although its effect is less broad. We then explored if N49P7 could be further improved by introducing a bulky residue at site 54. Thus, we mutated G54 to W, H, or Y and produced the three variants in HEK293FS cells. The folding of all 3 mutants was correct. When tested against the small global panel of 12 HIV-1 isolates, the three variants showed an enhanced neutralization against most of the isolates, with a median IC₅₀ ∼2-fold lower than that of the original N49P7 chimera. G54W was slightly more potent than G54H and G54Y (Fig. 6B and C). We also assessed the autoreactivity of the G54 mutants. Both G54H and G54Y are not autoreactive, like their parental N49P7 FR3-03. G54W showed a slightly increased autoreactivity but still remains at a very low level (Fig. 2B). Taken together, these results show that by introducing mutations at site 54, we obtained chimeric N49P7 variants with further enhanced neutralization potency.

FIG 6.

G54 mutations further increase the neutralization potency of chimeric N49P7 FR3-03 antibody. (A) Sequence alignment of the CDR2 and FR3 regions of VRC01-class antibodies. The residue at site 54 of N49-series antibodies are in red. (B) Neutralization of the G54 mutants derived from N49P7 FR3-03 against a small global panel of HIV-1 isolates. The fold change of the G54 mutants’ neutralizing capacity relative to the original N49P7 FR3-03 is denoted by different red shades for increased neutralization and green shades for decreased neutralization. All neutralization assays were performed in duplicate wells. (C) IC₅₀ and IC₈₀ values of N49P7 FR3-03 G54 mutants were plotted, and the median value was indicated with a black horizontal line.

Chimerization with the CDR1 loop of antibody 1-18 fails to improve other anti-CD4bs antibodies.

A novel anti-CD4bs bNAb with very high potency and breadth, 1-18, has recently been reported and shown to establish quaternary contact with the HIV-1 Env. However, unlike VRC03, which reaches the second gp120 protomer through its elongated FR3 loop, 1-18 contains a regular-length FR3 but can contact an adjacent gp120 protomer through a prolonged heavy-chain CDR1 loop (16). Interestingly, both the VRC03 FR3 loop and the 1-18 CDR1 loop interact with a similar region on the second gp120 protomer, comprising residue K207 and the base of the V3 loop. In addition, the sequence of both loops is enriched in the same amino acid, aspartic acid (Fig. 7A). This led us to hypothesize that the 1-18 CDR1 loop could be engrafted onto other anti-CD4bs bNAbs that lack quaternary contact. Thus, we introduced the 1-18 CDR1 loop onto N6, VRC07-523LS, and CH31, creating three chimeras, N6-CDR1-18, VRC07-523LS-CDR1-18, and CH31-CDR1-18, respectively, and tested their neutralization against the small global panel of HIV-1 Envs. All chimeras showed a reduced neutralizing capacity compared to their respective original forms, with median IC₅₀ values increasing by more than 5-fold compared to their respective WT form (Fig. 7B). This disappointing result suggested that 1-18, being a VH1-46 class antibody, adopts a different binding angle from that adopted by VRC01-class antibodies. Indeed, when we overlaid the BG505 DS-SOSIP.664 Env trimer in complex with VRC03 (PDB entry 6CDI) onto the BG505 SOSIP.664 Env trimer in complex with 1-18 (PDB entry 6UDJ), the axis that traverses FR3 of VRC03 was tilted by 18 degrees with respect to that of 1-18 FR3 (Fig. 7C). As a result, FR3 of VRC03 overlaps CDRH1 of 1-18 and CDRH1 of VRC03 shifts toward the neighboring protomer, suggesting that the transplanted CDRH1 loop on the chimeras induces steric clashes with the quaternary binding site on the adjacent gp120 protomer. However, when gp120 in complex with antibody 8ANC131 (PDB entry 4RWY), a member of the VH1-46 family, was overlaid onto the BG505 SOSIP.664 Env trimer in complex with 1-18, the overall folds of the heavy-chain-variable domains of the two antibodies were superimposable with space between CDRH1 of 8ANC131 and the neighboring protomer, which can be filled in by engrafting CDRH1 of 1-18 to enhance the interprotomer contacts (Fig. 7D). Thus, these observations suggest that the use of paratope engraftment to enhance the neutralizing activity of antibodies should be V-gene family specific.

FIG 7.

Engraftment of the CDR1 loop from bNAb 1-18 does not increase the neutralizing potency of other anti-CD4bs bNAbs. (A) Sequence alignment of CDR1 region of 1-18 and other CD4bs antibodies. (B) The neutralization of three 1-18 CDR1-chimeric bNAbs was tested against the small global panel of HIV-1 isolates. The fold change of the chimeric antibodies relative to its respective WT is denoted by different red shades for increased neutralization and green shades for decreased neutralization. All neutralization assays were performed in duplicate wells. (C) Structural alignment of the BG505 DS-SOSIP.664 Env trimer in complex with VRC03 (PDB entry 6CDI) and the BG505 SOSIP.664 Env trimer in complex with 1-18 (PDB entry 6UDJ). The tilt angle of the VRC03 FR3 axis with respect to that of the 1-18 FR3 is shown in magenta lines. (D) Structural alignment of gp120 in complex with 8ANC131 (PDB entry 4RWY) and the BG505 SOSIP.664 Env trimer in complex with 1-18 (PDB entry 6UDJ).

DISCUSSION

The dramatic progress achieved in recent years in large-scale recombinant antibody production has opened new perspectives for the use of monoclonal antibodies for a wide variety of clinical applications. More than 100 monoclonal antibodies are currently approved for therapeutic use in human autoimmune, neoplastic, and infectious diseases (https://www.antibodysociety.org/resources/approved-antibodies/). In the field of infectious diseases, antibodies are becoming viable alternatives to small-molecule inhibitors for both prevention and treatment, with a large number of clinical products being developed against viral agents responsible for global pandemics, such as HIV-1 and SARS-CoV-2 (17, 18). To qualify for potential clinical applicability, antiviral antibodies should possess potent neutralizing activity, a broad spectrum of action, and favorable in vivo pharmacokinetics. Although in some cases native antibodies that are directly cloned from seropositive individuals display optimal biological features, various strategies for engineering antibodies with enhanced properties are being evaluated, particularly with regard to increased potency and prolonged in vivo half-life. Our recent success in enhancing potent bNAbs against the HIV-1 CD4bs by engrafting the extended FR3 loop of VRC03, a VRC01-class bNAb with limited potency and breadth, has provided a novel approach to structure-guided antibody improvement (14). Encouraged by these initial results, in the present study we have extended the scope of our paratope engraftment strategy to additional bNAbs specific for the CD4bs.

We demonstrated that insertion of the extended FR3 loop of VRC03 in bNAbs of the N49 lineage, which already possess a high potency (15), resulted in a further increase in neutralization potency against a large panel of global HIV-1 isolates of different clades and geographic origin, in agreement with our previous results with FR3-chimeric VRC01, VRC07-523LS, and N6. This increase in neutralization capacity correlated with an increased binding affinity for the HIV-1 Env trimer, in line with the acquisition of the new quaternary contacts with a second gp120 protomer. Of note, chimerization further expanded the neutralization breadth of N49P7 to 5 additional HIV-1 isolates. The successful engraftment of the VRC03 FR3 loop in different anti-CD4bs bNAbs was not entirely surprising, because all these antibodies, despite originating from different infected patients, are members of the same antibody class, characterized by VH1-2 usage and a basic pI. Indeed, structural studies have documented a similar angle of approach and mode of interaction with the Env trimer for all these anti-CD4bs antibodies (6, 14, 19, 20). In agreement with these analogies, mutagenesis studies demonstrated that quaternary Env contact by all of these FR3-chimeric bNAbs depends on the same amino acid residues in the second gp120 protomer, i.e., K207 at the base of the V1V2 loop complex, and various residues, mostly basic, at the base of the V3 loop. The consistency of these results raises hope that any new, potent antibody of the VH1-2-derived VRC01 class discovered in the future will be suitable for chimerization by VRC03 FR3-loop insertion, resulting in increased neutralizing activity and, potentially, more favorable pharmacokinetics.

In contrast to the consistent enhancement of neutralizing activity that we achieved by engraftment of the VRC03 FR3 loop, our attempts to enhance antibody potency by mimicking the quaternary mode of interaction of bNAb 1-18 (16) were not successful. 1-18 is a one-of-a-kind bNAb with several unique features, including a reported 97% breadth of neutralization and picomolar activity and a unique quaternary interaction with the HIV-1 Env trimer mediated by an elongated aspartate-enriched CDR1 loop. Of note, the 1-18 CDR1 loop appears to make quaternary contact with the same amino acid residues that interact with VRC03 or FR3-03 chimeras, suggesting that the peculiar angle of engagement of the Env trimer by 1-18 places the CDR1 loop in a position and orientation similar to those of the FR3 loop of VRC03 and its chimeric derivatives. The reasons for our lack of success in increasing the antiviral activity of various bNAbs by insertion of the 1-18 CDR1 loop are unclear, but a similar failure occurred when we attempted to engraft the short FR3 loop of bNAb 3BNC117 or the long CDR1 loop of VRC-CH31 onto other CD4bs-specific bNAbs (14). These disappointing results may be related to subtle differences in the angle of engagement and loop orientation among different antibodies, which have clearly evolved different modes of interaction with the CD4bs, whose discontinuous quaternary configuration makes it a challenging and highly protected target for antibody recognition. Despite the present setback, which well illustrates the challenges associated with structure-based antibody enhancement strategies, it remains possible that both the 1-18, 3BNC117, and VRC-CH31 quaternary-binding loops successfully transferred to other bNAbs by further modification of the engrafted loop or the recipient antibody framework to achieve the optimal orientation for interaction with the second gp120 protomer. Conversely, all the quaternary-interactive bNAbs appear to target the same, or at least a widely overlapping, region in the neighboring gp120 protomer, encompassing K207 at the base of the V1V2 loop complex and several residues at the V3 base. This finding reinforces the key role of this quaternary site as a new target for both vaccine and therapeutic strategies against HIV-1.

In conclusion, our study corroborates the efficacy of our paratope engraftment approach for the development of increasingly effective monoclonal antibodies for use in HIV-1 prevention and treatment. The availability of increasingly effective chimeric antibodies paves the way for their clinical use, especially in combination therapy with other potent antibodies against different sites of HIV-1 vulnerability.

MATERIALS AND METHODS

Mutagenesis.

Mutagenesis of HIV-1 Env and antibody-expressing genes was performed with the QuikChange II site-directed mutagenesis kit (200524; Agilent Technologies) or Q5 site-directed mutagenesis kit (E0554S; New England BioLabs Inc.) per the manufacturer’s protocols.

Antibody expression and purification.

All antibodies were produced by cotransfecting equal amounts of heavy-chain and light-chain plasmids to HEK293FS cells (human embryonic kidney cells, female; obtained from ThermoFisher) using FreeStyle MAX reagent (16447500; Thermo Fisher Scientific). Transfected cells were incubated at 37°C for 5 days, and cell-free supernatants were collected by centrifugation. After filtering through a 0.22-μm filter, the supernatant was loaded onto a protein A column and washed with phosphate-buffered saline (PBS). IgG elution buffer (21009; Thermo Fisher Scientific) was used to elute proteins, followed by neutralization with 100 mM Tris-HCl, pH 9.0. Finally, the protein was dialyzed to PBS and stored at 1 mg/ml at −80°C.

HIV-1 Env trimer expression and purification.

BG505 and JRFL SOSIP.664 trimers were expressed in HEK293FS cells by cotransfecting an Env-expressing plasmid and a Furin-expressing plasmid. The cell supernatant was collected after a 5-day incubation at 37°C and filtered through a 0.22-μm filter. The filtered supernatant was first purified with a Galanthus nivalis lectin column (Vector Laboratories) and dialyzed to PBS, followed by two rounds of size-exclusion chromatography. Misfolded proteins were further removed by passing through a 447-52D affinity column. Purified proteins were concentrated and stored in PBS supplemented with 10% glycerol at −80°C.

Pseudovirus production and neutralization assays.

Viral pseudoparticles were produced in HEK293T cells by cotransfection of a backbone plasmid, pSG3^ΔENV, and Env-expressing plasmids using TransIT-293 transfection reagent (Mirus). A single-round infection assay in TZM-bl cells (human malignant epithelial cells, female; obtained from the NIH AIDS Reagent Program) was carried out. Serial dilutions of each antibody were incubated with pseudoviruses in a 96-well plate at room temperature (RT) for 30 min, followed by adding 10,000 TZM-bl cells/well. After 48 h of incubation, luminescence signal was detected using a luciferase assay kit (Promega). The antibody concentrations that yielded 50% or 80% neutralization were calculated by Prism 7. The large panel for 119 HIV-1 pseudovirus neutralization testing was examined using a luciferase-based assay in TZM-bl cells, as previously described (15). Briefly, the monoclonal antibodies were serially diluted, and 100 μl was added to duplicate wells (96-well flat-bottom plate). Fifty microliters of 200 50% tissue culture infective doses of pseudovirus was added to each well and incubated for 1 h at 37°C. TZM-bl cells in 10% Dulbecco’s modified Eagle medium (DMEM) containing DEAE-dextran (Sigma, St. Louis, MO) at a final concentration of 11 μg/ml were then added (10,000 cells/well in 100 μl volume). After 48 h of incubation at 37°C, luminescence signal was detected using a Bright-Glo luciferase reagent (Promega).

Autoreactivity analysis.

Autoreactivity was measured by ANA HEp-2 staining analysis (no. FA2400; ZEUS Scientific) per the manufacturer’s protocols. All antibodies were tested at 50 μg/ml. Three antibodies, VRC01, 4E10, and VRC07-523LS, were used as controls.

Enzyme immunoassays.

Ninety-six-well ELISA plates (Corning) were coated with 5 μg/ml G. nivalis lectin (L8275-5MG; Sigma) at 4°C overnight. Three washes were performed with 1× wash buffer (R&D Systems) after each step. The coated plates were first blocked with 0.2% casein in PBS, and the SOSIP.664 trimer or gp120 monomer was added at 1 or 2 μg/ml. After incubation for 1 h at RT, serially diluted antibodies were added to duplicate wells for a 1-h incubation, followed by a horseradish peroxidase-conjugated goat antihuman IgG (A8419 [Sigma] or 109-035-008 [Jackson ImmunoResearch]) at a 1:5,000 dilution. The binding signal was revealed by incubating with substrate reagent (R&D Systems) for 10 to 20 min, and the reaction was stopped by stop solution (R&D Systems). Light absorption at 450 nm was recorded with a luminometer (PerkinElmer).