Two Families of Env Antibodies Efficiently Engage Fc-Gamma Receptors and Eliminate HIV-1-Infected Cells

The “open” CD4-bound conformation of HIV-1 envelope glycoproteins is the primary target of antibody-dependent cellular cytotoxicity (ADCC)-mediating antibodies present in HIV-positive (HIV+) sera, such as anti-coreceptor binding site and anti-cluster A antibodies. Here we report that the binding of these two families of antibodies is required to engage FcγRIIIa and mediate ADCC. ABSTRACT HIV-1 conceals epitopes of its envelope glycoproteins (Env) recognized by antibody (Ab)-dependent cellular cytotoxicity (ADCC)-mediating antibodies. These Abs, including anti-coreceptor binding site (CoRBS) and anti-cluster A antibodies, preferentially recognize Env in its “open” conformation. The binding of anti-CoRBS Abs has been shown to induce conformational changes that further open Env, allowing interaction of anti-cluster A antibodies. We explored the possibility that CoRBS Abs synergize with anti-cluster A Abs to engage Fc-gamma receptors to mediate ADCC. We found that binding of anti-CoRBS and anti-cluster A Abs to the same gp120 is required for interaction with soluble dimeric FcγRIIIa in enzyme-linked immunosorbent assays (ELISAs). We also found that Fc regions of both Abs are required to optimally engage FcγRIIIa and mediate robust ADCC. Taken together, our results indicate that these two families of Abs act together in a sequential and synergistic fashion to promote FcγRIIIa engagement and ADCC. IMPORTANCE The “open” CD4-bound conformation of HIV-1 envelope glycoproteins is the primary target of antibody-dependent cellular cytotoxicity (ADCC)-mediating antibodies present in HIV-positive (HIV+) sera, such as anti-coreceptor binding site and anti-cluster A antibodies. Here we report that the binding of these two families of antibodies is required to engage FcγRIIIa and mediate ADCC.

There are many factors that can contribute to the ability of Abs to engage effector cells to clear infected cells, such as (i) the strength of IgG interactions with Fc␥Rs; (ii) the density of IgG opsonization; and (iii) the orientation of the Fc portion (11,12). Additionally, the interaction of Fc␥Rs with Abs bound to distinct epitopes may have additive or synergistic effects on host immune responses. Previous reports have shown improved neutralization efficacy with two or more monoclonal Abs (MAbs) targeting the Ebola virus (13), Hepatitis C virus (14) and HIV-1 (15). Using influenza A virus as a model, it has been shown that the combination of MAbs targeting different epitopes enhanced the induction of ADCC in an additive manner (16). These findings exemplify the complex interactions present in a polyclonal Ab response in vivo.
In the context of HIV-1, it has been shown that antibodies targeting the CD4-bound conformation of viral envelope glycoproteins (Env) mediate strong ADCC activity (17)(18)(19)(20). It should be noted that these Abs require infected cells to present Env in an "open" conformation for ADCC to be achievable (18,19,21,22). Several families of Abs can bind epitopes exposed in this conformation of Env, including anti-cluster A, anti-coreceptor binding site (anti-CoRBS), anti-V3, and anti-gp41 Abs (17). To avoid exposure of these CD4-induced epitopes, however, HIV-1 efficiently internalizes Env (76) and prevents the accumulation of the CD4 receptor on the surface of infected cells through Nef and Vpu accessory proteins (20). Moreover, Vpu decreases Env levels at the cell surface by antagonizing the BST-2 host restriction factor which otherwise tethers budding viral particles (23,24) (the impact of these accessory proteins on ADCC responses was recently reviewed in references [25][26][27]. In agreement with the susceptibility of the CD4-bound conformation of Env to ADCC, it has been shown that small-molecule CD4-mimetics (CD4mc) "force" Env to adopt a similar conformation that enhances susceptibility of infected cells to ADCC triggered by antibodies within sera, breast milk, and cervicovaginal fluids from HIV-1infected subjects (21,(28)(29)(30). The mechanism via which CD4mc facilitate anti-HIV-1 ADCC includes a synergistic interaction of Env with both CD4mc and anti-CoRBS Abs present in HIV-1-positive (HIV-1 ϩ ) sera. The sequential binding of CD4mc and anti-CoRBS Abs opens the Env trimer and facilitates recognition by the anti-cluster A family of Abs. The members of the latter family of Abs recognize epitopes located in the gp120 inner domain which are occluded in the "closed" native trimer (12,20,(31)(32)(33)(34) and are known to be responsible for the majority of the ADCC activity present in HIV-1 ϩ sera (17,19,20,31,35). Accordingly, the addition of anti-cluster A Fab fragments has been shown to block most ADCC activity present in HIV-1 ϩ sera (35). Interestingly, addition of Fab fragments from the 17b antibody, which belongs to the anti-CoRBS family of anti-Env antibodies (36), also significantly blocked ADCC activity of HIV-1 ϩ sera and in some cases to the same extent as anti-cluster A Fab fragments (35). Taken together, these studies suggested that there is a high degree of cooperation between anti-cluster A and anti-CoRBS Abs in the induction of effector functions. To mediate ADCC, Abs must bridge their Fab-bound cognate antigen with effector cells capable of recognizing Ab Fc via Fc␥R. While it was demonstrated that anti-CoRBS Abs facilitate the binding of anti-cluster A Abs to Env (29), it is still not known whether these two families of Abs also coordinate for Fc␥R engagement. Here we took advantage of the availability of a recently characterized dimeric recombinant soluble Fc␥RIIIa protein (11,(37)(38)(39)(40) to evaluate whether these two families of anti-Env Abs coordinate for Fc␥R engagement and ADCC responses.
bound anti-cluster A (A32) and/or anti-CoRBS (17b) Abs to engage soluble rsFc␥RIIIa was determined. Despite detectable binding of both A32 and 17b to the gp120-coated plates (Fig. 1C), we noted poor binding of rsFc␥RIIIa to either Ab alone or to the two Abs in combination (Fig. 1A). Since we recently reported that anti-CoRBS Abs facilitate anti-cluster A Ab binding upon addition of CD4mc BNM-III-170 (29), we evaluated if this small-molecule CD4mc could facilitate rsFc␥RIIIa engagement. Plates were coated with recombinant gp120, and then antibody and rsFc␥RIIIa binding levels were measured by ELISA in the presence of BNM-III-170. Consistent with the CD4-induced (CD4i) nature of anti-cluster A and anti-CoRBS Abs and the capacity of BNM-III-170 to stabilize gp120 in a CD4-bound conformation (41), both Abs bound gp120 slightly better in the presence of BNM-III-170 (Fig. 1C). Interestingly, only the combination of A32 and 17b resulted in efficient engagement between the Abs and rsFc␥RIIIa (Fig. 1A). To confirm that the CD4-bound conformation of gp120 facilitates rsFc␥RIIIa engagement by anti-cluster A and anti-CoRBS Abs, this Env conformation was stabilized by deleting the V1, V2, V3, and V5 variable regions (ΔV1V2V3V5) (42,43). Using this stabilized recombinant gp120, we observed that neither anti-cluster A (A32; N5-i5) nor anti-CoRBS (17b; N12-i2) Abs could independently engage the rsFc␥RIIIa. We noted, however, that utilizing any combination of Abs exhibiting these two specificities resulted in a dramatic increase in rsFc␥RIIIa engagement (Fig. 1B)-despite similar overall levels of binding of the Abs to FIG 1 Anti-CoRBS and anti-cluster A Abs synergize to engage dimeric Fc␥RIIIa. Indirect ELISA was performed using recombinant full-length YU2 gp120 protein (0.4 g/ml) (A and C) or using recombinant YU2 ΔV1V2V3V5 gp120 protein (0.25 g/ml) (B and D). gp120-coated wells were incubated with a total concentration of 1 g/ml of primary antibodies (A and C) with or without BNM-III-170 (25 M). Antibody binding to gp120 was detected using biotin-tagged dimeric rsFc␥RIIIa (0.1 g/ml) followed by the addition of HRP-conjugated streptavidin (A and B) or using HRP-conjugated anti-human secondary antibody (C and D). Data in graphs represent relative light units (R.L.U.) determined from at least 5 independent experiments done in quadruplicate, with the error bars indicating means Ϯ standard errors of the means (SEM). Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001; ns, nonsignificant). the CD4-bound-stabilized gp120 (Fig. 1D). Cumulatively, these data indicate that anti-cluster A and anti-CoRBS Abs synergize to engage with Fc␥RIIIa.
Anti-CoRBS and anti-cluster A Abs must recognize the same gp120 monomer in order to engage with dimeric Fc␥RIIIa. In theory, due to their nonoverlapping epitopes, anti-cluster A and anti-CoRBS Abs could bind the same gp120-coated monomer or two different adjacent gp120-coated monomers to facilitate rsFc␥RIIIa engagement. To distinguish between these two possibilities, we coated equivalent amounts of CD4-bound stabilized (ΔV1V2V3V5) gp120 mutants that were competent to engage with anti-cluster A but had decreased binding to anti-CoRBS (R419D) Abs or that were competent to bind anti-CoRBS but had decreased binding to anti-cluster A (W69A) Abs (Fig. 2B). Strikingly, A32 and 17b were able to engage rsFc␥RIIIa only when bound to the CD4-bound stabilized gp120 wild type (WT) (having intact epitopes for anti-cluster and CoRBS Abs) ( Fig. 2A). Indeed, despite similar overall levels of binding of A32 and 17b to wells coated with an equimolar ratio of R419D and W69A CD4-bound stabilized recombinant gp120s, poor rsFc␥RIIIa engagement was observed. These data suggest that A32 and 17b must bind the same gp120 to efficiently engage with Fc␥RIIIa by ELISA.
Anti-CoRBS and anti-cluster A Abs present in HIV-1 ؉ sera are required for efficient engagement of Fc␥RIIIa. Since the two families of Abs tested in this study are easily elicited following HIV-1 infection (19,29,31,44), we evaluated their contribution to the capacity of sera from 15 chronically HIV-1-infected individuals to bind to the rsFc␥RIIIa by ELISA using the gp120 stabilized in the CD4-bound conformation (ΔV1V2V3V5). Compared to sera from 5 healthy HIV-1-uninfected donors, rsFc␥RIIIa was efficiently engaged by Abs present in most HIV-1 ϩ sera tested (Fig. 3A). Interestingly, the addition of A32 or 17b Fab fragments to the HIV-1 ϩ sera significantly blocked the binding of rsFc␥RIIIa (Fig. 3A). Addition of A32 and 17b fragments in combination significantly diminished rsFc␥RIIIa engagement to the level seen with sera from healthy uninfected (HIV Ϫ ) individuals (Fig. 3A). Similarly, gp120 mutants with an impaired FIG 2 Anti-CoRBS and anti-cluster A Abs must engage with the same gp120 subunit in order to interact with dimeric Fc␥RIIIa. Indirect ELISA was performed using recombinant YU2 ΔV1V2V3V5 gp120 protein (WT), its mutated counterpart (W69A or R419D), or an equimolar mixture of both mutated gp120 proteins (W69A and R419D). gp120-coated wells were incubated with a total concentration of 1 g/ml of primary antibodies. Antibody binding to gp120 was detected using (A) biotin-tagged dimeric rsFc␥RIIIa (0.1 g/ml) followed by the addition of HRP-conjugated streptavidin or using (B) HRP-conjugated anti-human secondary antibody. Graphs represent relative light units (R.L.U.) determined from 5 independent experiments done in quadruplicate, with the error bars indicating means Ϯ SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; ****, P Ͻ 0.0001; ns, nonsignificant). ability to engage anti-cluster A (W69A) and/or anti-CoRBS (R419D) Abs failed to facilitate rsFc␥RIIIa engagement with bound Abs (Fig. 3B). As such, the data suggest that anti-cluster A and anti-CoRBS Abs represent the majority of Abs able to recruit Fc␥RIIIa in HIV-1 ϩ sera toward the gp120 stabilized in the CD4-bound conformation.
Anti-CoRBS and anti-cluster A Abs are required for efficient binding of Fc␥RIIIa to infected cells. The work described above and prior studies used plate-bound Env proteins to measure engagement of Fc␥RIIIa. We therefore developed a more physiologically appropriate system to measure this interaction, using HIV-1-infected cells. Primary CD4 ϩ T cells were infected with primary transmitted/founder (T/F) virus CH58 (CH58 T/F) that was either wild type (WT) or defective for Nef and Vpu expression (N Ϫ U Ϫ ). The N Ϫ U Ϫ virus was used as a means to present Env in the CD4-bound Indirect ELISA using recombinant YU2 ΔV1V2V3V5 gp120 protein (0.25 g/ml). gp120-coated wells were incubated with sera from 15 chronically HIV-1-infected individuals with or without A32 Fab and/or 17b Fab fragments (1 g/ml). (B) Indirect ELISA using recombinant YU2 ΔV1V2V3V5 gp120 protein (WT) or its mutated counterparts (W69A or R419D or the double mutant W69A/R419D). Serum binding to gp120 was detected using biotin-tagged dimeric rsFc␥RIIIa (0.1 g/ml) followed by the addition of HRPconjugated streptavidin. Sera from 5 healthy individuals (HIV Ϫ ) were used as a negative control. Graphs represent relative light units (R.L.U.) determined from 3 independent experiments done in quadruplicate, with the error bars indicating means Ϯ SEM. Statistical significance was tested using a paired t test or a Wilcoxon matched-pair signed-rank test (based on statistical normality) to compare HIV ϩ serum data sets and a Mann-Whitney U test to compare HIV ϩ sera with HIV Ϫ sera (****, P Ͻ 0.0001; ns, nonsignificant). conformation at the surface of infected cells (19,20). Infected cells were identified by intracellular p24 staining. As previously shown (29), cells infected with the wild-type virus and therefore exposing Env in its untriggered closed conformation were poorly recognized by A32 or 17b in the absence of the CD4mc BNM-III-170. The presence of BNM-III-170 enhanced recognition of the infected cells by 17b but not A32 (Fig. 4C). In agreement with the ability of 17b to facilitate A32 binding to Env in the presence of CD4mc (29) (Fig. 5), recognition of infected cells in the presence of BNM-III-170 was significantly higher for the combination of 17b and A32 than for 17b alone. Cobinding of 17b and A32 to CD4mc-bound Env translated to efficient and significant engagement of rsFc␥RIIIa ( Fig. 4A and B). As expected, cells infected with the N Ϫ U Ϫ virus presented Env in its CD4-bound conformation. As such, N Ϫ U Ϫ virus-infected cells were efficiently recognized by 17b and by A32 and by the combination of the two Abs ( Fig.  4G and H). Interestingly, however, only the combination of A32 and 17b was able to engage dimeric rsFc␥RIIIa ( Fig. 4E and F). Thus, anti-cluster A and anti-CoRBS Abs are required to efficiently bind rsFc␥RIIIa to infected cells.
Anti-CoRBS and anti-cluster A Abs are required for efficient ADCC-mediated killing of HIV-1-infected cells. To evaluate whether engagement of rsFc␥RIIIa translates to ADCC, primary CD4 ϩ T cells from at least five healthy uninfected (HIV Ϫ ) donors were infected with CH58 T/F WT and N Ϫ U Ϫ viruses and evaluated for their susceptibility to ADCC mediated by effector cells present in autologous peripheral blood mononuclear cells (PBMCs). ADCC was assessed using an assay measuring the elimination of infected (p24 ϩ ) cells (19-21, 30, 45). In agreement with the lack of rsFc␥RIIIa binding to wild-type-infected cells treated with A32, 17b, or A32/17b in the absence of BNM-III-170 CD4mc (Fig. 4A), only low ADCC activity was observed for A32, 17b, or A32/17b in the absence of BNM-III-170 ( Fig. 6A and B). Significant ADCC activity was observed only upon addition of the small-molecule CD4mc BNM-III-170 with the combination of A32 and 17b Abs as predicted by rsFc␥RIIIa binding. As previously reported, cells exposing Env in its CD4-bound conformation due to Nef and Vpu deletions were susceptible to ADCC mediated by A32 (19,20) but not 17b (17). However, the combination of A32 and 17b further enhanced ADCC ( Fig. 6C and D). Taken together, these data suggest that anti-cluster A and anti-CoRBS Abs cooperate for ADCC activity against infected cells by better engaging Fc␥RIIIa.
Fc regions of anti-cluster A and anti-CoRBS Abs are required to engage Fc␥RIIIa and mediate ADCC. To evaluate how the combination of anti-cluster A and anti-CoRBS Abs facilitates engagement of rsFc␥RIIIa, we introduced double leucine-to-alanine substitutions at positions 234 and 235 (LALA) of the Fc region of anti-CoRBS (N12i2) and anti-cluster A (N5i5) Abs. The LALA mutation significantly decreases Ab engagement with Fc␥RIIIa (46)(47)(48). First, we evaluated the ability of the unmutated Abs and their LALA counterparts to engage rsFc␥RIIIa by ELISA. As shown in Fig. 7A, only the combination of Fc-competent anti-cluster A and anti-CoRBS Abs engaged rsFc␥RIIIa, despite similar levels of Ab binding to coated gp120 under conditions employing single Abs or combinations of LALA mutant Abs (Fig. 7B). These results were corroborated by experiments assessing rsFc␥RIIIa engagement with Abs bound to cells infected with wild-type and N Ϫ U Ϫ viruses. In cells infected with the wild-type virus, anti-cluster A and anti-CoRBS Abs were able to efficiently engage the rsFc␥RIIIa upon small-molecule CD4mc BNM-III-170 addition, but only if Fc-competent Abs were used (Fig. 7C). Similarly, only combinations of Fc-competent anti-cluster A and anti-CoRBS Abs were able to engage rsFc␥RIIIa to Nef Ϫ Vpu Ϫ virus-infected cells (Fig. 7D). Importantly, rsFc␥RIIIa engagement translated into ADCC activity ( Fig. 7E and F). Likewise, even though Fab=2 fragments of anti-CoRBS Abs sufficiently expose anti-cluster A epitopes in the presence of CD4mc (29) (Fig. 5), only the combination of anti-cluster A Abs with full IgG anti-CoRBS Abs, and not Fab'2 fragments, optimally recruited rsFc␥RIIIa and mediated potent ADCC in the presence of CD4mc (Fig. 8 and 9). Thus, the data suggest that a threshold of Fc␥RIIIa engagement is required to mediate ADCC. Moreover, our results indicate that both Fc regions of these two families of Abs are required to engage Fc␥RIIIa to mediate ADCC.

DISCUSSION
HIV-1 has evolved several mechanisms to evade the humoral response elicited against its envelope glycoproteins, including conformational masking of vulnerable epitopes, high sequence variability in the variable loops, and extensive glycosylation (49)(50)(51). The conformational dynamics of Env also constitute a major barrier for Env inactivation by Abs. The functional Env trimer mainly exists in an untriggered closed conformation (state 1) that cannot be seen by nonneutralizing Abs (nnAbs). However, Env has intrinsic access to the open CD4-bound conformation (state 3) through one necessary conformational intermediate (state 2) in which only one protomer is competent for CD4 binding (52,53). Env from primary tier 2 viruses are more closed, with elevated state 1 occupancy, explaining why they are resistant to binding and neutralization by nnAbs (53). CD4 and CD4mc interaction modifies the conformational landscape toward state 3 by lowering the energy barrier to open states 2 and 3, rendering Env vulnerable to nnAb attack (18,(52)(53)(54). The majority of the antibodies elicited during natural infection are strain-specific neutralizing or nnAbs (55) and are thought to play a minimal role in controlling viral replication. Recent studies, however, showed that these antibodies can exert constant selection pressure and alter the course of HIV-1-infection in vivo (56,57). Nonneutralizing Abs with Fc-effector function are present in HIV-1 ϩ sera and have been shown to mediate potent ADCC activity against cells presenting Env in its open CD4-bound conformation (17-21, 26, 30, 58-60). Among the different nnAbs tested for their ADCC activity, the members of the anticluster A family of Abs were identified as being the most potent against target cells expressing Env in this open conformation (17,31,33,35). Of note, these Abs can be readily elicited in nonhuman primates with monomeric gp120 immunogens (22). Additional nnAbs such as anti-CoRBS Abs, which are unable to mediate potent ADCC activity on their own (17), were shown to facilitate engagement of anti-cluster A Abs and potentiate their ADCC activity (29). Here we provide evidence that anti-cluster A and anti-CoRBS Abs collaborate in engaging the dimeric form of the Fc␥RIIIa receptor when bound to the cognate epitopes within the same monomeric gp120. We utilized a novel probe, a covalent Fc␥RIIIa dimer, as a surrogate of Fc␥RIIIa present on effector cells (11,37,39,40,61) and showed that sequential binding of antibodies specific for these two CD4i epitope targets is required for effective receptor engagement and potent ADCC. Recent structural studies defined the Env epitopes recognized by anticluster A and anti-CoRBS Abs at the atomic level and mapped the cluster A epitope region to the C1-C2 regions of the gp120 inner domain in its CD4-bound conformation in the infected (p24 ϩ ) population determined from 5 independent experiments, with the error bars indicating means Ϯ SEM. Statistical significance was tested using an unpaired t test (*, P Ͻ 0.05; ****, P Ͻ 0.0001; ns, nonsignificant). (12,32,33,62) and CoRBS to the bridging sheet and the base of the V3 loop of the CD4-triggered gp120 (63-65). These two epitopes are not overlapping. In order to validate our findings of collaboration in effective engagement of the dimeric Fc-gamma receptor by antibodies specific for these two epitope targets, we prepared a model of the putative immune complex that needs to be formed on the gp120 antigen in order to engage a covalent Fc␥RIIIa dimer (Fig. 10). We used the crystallographic dimer of Fc␥RIIa-HR (PDB code 3RY5), a high-responder polymorphism Fc␥RIIa (66), as a template to generate the putative Fc␥RIIIa dimer. As predicted, the anti-cluster A N5-i5 and anti-CoRBS 17b antibodies bound to nonoverlapping sites on the same gp120 protomer such that their Fc domains were brought in close proximity (Fig. 10). Simple adjustment of Fc positions, which could be mediated by the IgG1 hinge region, oriented the Fcs to permit interactions with the Fc␥RIIIa dimer (Fig. 10). Taken together, the in silico modeling data clearly support the possibility of the formation of complexes in which the anti-cluster A and anti-CoRBS Abs are bound to the same gp120 antigen and their Fc domains engaged in an interaction with the dimeric Fc gamma receptor. However, whether the same configuration is maintained on trimeric Env remains to be shown.
Although our results revealed that anti-cluster A and anti-CoRBS Abs collaborate to trigger anti-HIV-1 ADCC, note that a collaboration between these Abs is required for ADCC only in the context of CD4mc. Indeed, target cells infected with N Ϫ U Ϫ viruses bind A32 and are susceptible to ADCC through A32 alone. Similarly, gp120-coated CD4 ϩ T cells are susceptible to ADCC triggered by A32 alone (45,67). The exact  (Continued on next page) mechanisms behind the differential requirements for anti-CoRBS Abs for ADCC against target cells infected with wild-type virus in the presence of CD4mc and target cells coated with gp120 or infected with N Ϫ U Ϫ viruses remain undetermined. One possible contributor to this phenomenon is the minimal density and/or proximity of Env on cells infected with wild-type virus, increasing the need for multiple Abs per Env spike (68). Another possible contributor is that the conformational changes in Env induced by cell surface CD4 (present on target cells coated with gp120 or infected with N Ϫ U Ϫ viruses) are sufficient to allow binding to cell surface coreceptors and negate the need for anti-CoRBS Abs to reveal the A32 epitope (12,20). The conformational changes in Env induced by CD4mc, however, are not sufficient to allow coreceptor binding to Env, thus necessitating anti-CoRBS Ab binding to reveal the A32 epitope (29).

FIG 7 Legend (Continued)
g/ml) followed by the addition of Alexa Fluor 647-conjugated streptavidin. Graphs represent mean fluorescence intensities (MFI) in the infected (p24 ϩ ) population determined from 5 independent experiments, with the error bars indicating means Ϯ SEM. (E and F) Primary CD4 ϩ T cells from at least 5 different healthy donors infected with CH58 T/F virus strains that were either (E) WT or (F) defective for Nef and Vpu expression (N-U-) were used as target cells and autologous PBMC as effector cells in a FACS-based ADCC assay (29). (E) Analysis of ADCC with CH58 T/F WT was done in the presence of BNM-III-170 (50 M). Shown in panels E and F are the percentages of ADCC-mediated killing obtained with a total concentration of 2.5 g/ml of antibodies. Data are representative of results from 5 independent experiments, with the error bars indicating means Ϯ SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001; ns, nonsignificant).

FIG 8
The Fc portions from both anti-CoRBS and anti-cluster A Abs are required to engage dimeric Fc␥RIIIa. Cell surface staining of primary CD4 ϩ T cells infected with CH58 T/F WT virus was performed with a total concentration of 2.5 g/ml of primary antibodies with or without BNM-III-170 (50 M). Antibody binding was detected either by (A and B) using biotin-tagged dimeric rsFc␥RIIIa (0.2 g/ml) followed by the addition of Alexa Fluor 647-conjugated streptavidin or by (C and D) using Alexa Fluor 647-conjugated anti-human secondary antibody. Graphs represent the mean fluorescence intensities (MFI) in the infected (p24 ϩ ) population determined from 5 independent experiments, with the error bars indicating means Ϯ SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ns, nonsignificant).
Taken together, our data identified a complex interplay between two families of easy-to-elicit antibodies, present in most HIV-1-infected individuals, with good potential to eliminate HIV-1-infected cells, provided that the Env in the target cell is in a more open conformation. Interestingly, small-molecule CD4-mimetics have been shown to "open up" the Env trimer (21,22,30); therefore, these data suggest that strategies aimed at stabilizing open Env conformations with CD4mc might expedite the design of new strategies aimed at fighting HIV-1 infection.

MATERIALS AND METHODS
Ethics statement. Written informed consent was obtained from all study participants (the Montreal Primary HIV Infection Cohort [69,70] and the Canadian Cohort of HIV Infected Slow Progressors [71,72]), and research adhered to the ethical guidelines of Centre de Recherche du CHUM (CRCHUM) and was reviewed and approved by the CRCHUM institutional review board (Ethics Committee approval number CE16.164-CA). Research adhered to the standards indicated by the Declaration of Helsinki. All participants were adults and provided informed written consent prior to enrollment in accordance with Institutional Review Board approval. All sera were heat inactivated for 30 min at 56°C and stored at 4°C until use in subsequent experiments.
Isolation of primary cells. CD4 T lymphocytes were purified from resting PBMCs by negative selection and activated as previously described (20,21). Briefly, PBMCs were obtained by leukapheresis. CD4 ϩ T lymphocytes were purified using immunomagnetic beads per the instructions of the manufacturer (StemCell Technologies). CD4 ϩ T lymphocytes were activated with phytohemagglutinin-L (10 g/ ml) for 48 h and then maintained in RPMI 1640 complete medium supplemented with recombinant interleukin-2 (rIL-2) (100 U/ml). Viral production and infections. Vesicular stomatitis virus G (VSVG)-pseudotyped viruses were produced and titrated as described previously (19). Viruses were used to infect activated primary CD4 T cells from healthy HIV-1-negative donors by spin infection at 800 ϫ g for 1 h in 96-well plates at 25°C.
CD4-mimetics. The small-molecule CD4-mimetic compound BNM-III-170 was synthesized as described previously (41). The compounds were dissolved in dimethyl sulfoxide (DMSO) at a stock HIV-1-infected primary CD4 ϩ T cells. Additionally, the following anti-Env MAbs were also used for cell surface staining: N12i2 LALA, N5i5 LALA, the Fab fragments, Fab=2 fragments, or full CoRBS MAbs 17b and N12-i2. Goat anti-human Alexa Fluor-647 secondary Ab (Thermo Fisher Scientific) was used to determine overall antibody and serum binding and AquaVivid (Thermo Fisher Scientific) as a viability dye. Alexa Fluor 647-conjugated streptavidin was used to determine levels of biotin-tagged dimeric recombinant soluble Fc␥RIIIa (V 158 ) binding. Sera from HIV-infected and healthy donors were collected, heat inactivated, and conserved as previously described (19).
Protein production of recombinant proteins. FreeStyle 293F cells (Invitrogen) were grown in FreeStyle 293F medium (Invitrogen) to a density of 1 ϫ 10 6 cells/ml at 37°C with 8% CO 2 with regular agitation (150 rpm). Cells were transfected with a codon-optimized plasmid expressing His6-tagged wild-type or mutant HIV-1 YU2 gp120 using ExpiFectamine 293 transfection reagent, as directed by the manufacturer (Invitrogen). At 1 week later, the cells were pelleted and discarded. The supernatants were filtered (Thermo Fisher Scientific) (0.22-m-pore-size filter), and the gp120 glycoproteins were purified by the use of nickel affinity columns, as directed by the manufacturer (Invitrogen). Monomeric gp120 was subsequently purified by fast protein liquid chromatography (FPLC), as reported previously (43). The gp120 preparations were dialyzed against phosphate-buffered saline (PBS) and stored in aliquots at Ϫ80°C. To assess purity, recombinant proteins were loaded on SDS-PAGE polyacrylamide gels in the absence of beta-mercaptoethanol and stained with Coomassie blue.
The biotin-tagged dimeric recombinant soluble Fc␥RIIIa (V 158 ) protein was produced and characterized as described previously (37) with additional purification performed using a Superdex 200 10/300 GL column (GE Life Sciences).
ELISA-based rsFc␥RIIIa dimer-binding assay. Nonsaturating concentrations of recombinant gp120 glycoproteins in their full-length form (0.4 g/ml) or lacking the V1, V2, V3, and V5 variable regions (0.25 g/ml) were prepared in PBS as well as in bovine serum albumin (BSA) (2 g/ml) as a negative control and were adsorbed to plates (MaxiSorp; Nunc) overnight at 4°C. Coated wells were subsequently blocked with blocking buffer (Tris-buffered saline [TBS] containing 0.1% Tween 20 and 1% [wt/vol] BSA) for 1 h at room temperature. Anti-HIV-1 Env monoclonal antibodies (1 g/ml) or HIV ϩ sera (1:1,000 dilution) were diluted in blocking buffer and incubated with the gp120-coated wells for 1 h at room temperature. When indicated, 25 M BNM-III-170 (or equivalent amounts of DMSO) was diluted with the antibodies. Plates were washed four times with washing buffer (TBS containing 0.1% Tween 20). This was followed by incubation with 0.1 g/ml purified dimeric rsFc␥RIIIa-biotin in blocking buffer for 1 h at room temperature. Horseradish peroxidase (HRP)-conjugated antibody specific for the Fc region of human IgG (Pierce) or high-sensitivity streptavidin-HRP (Thermo Fisher) (1/1,000 in blocking buffer) was added for 1 h at room temperature, followed by four washes. HRP enzyme activity was determined after the addition of a 1:1 mix of Western Lightning oxidizing and luminol reagents (Perkin Elmer Life Sciences). Light emission was measured with an LB 941 TriStar luminometer (Berthold Technologies).
Flow cytometry analysis. Cell surface staining was performed as previously described (19,21,45). Cells infected with HIV-1 primary isolates were identified by intracellular staining of HIV-1 p24 using a Cytofix/Cytoperm fixation/permeabilization kit (BD Biosciences) and the phycoerythrin (PE)-conjugated anti-p24 MAb, clone KC57 (Beckman Coulter). The percentage of infected cells (p24 ϩ cells) was determined by gating the living cell population based on viability dye staining with AquaVivid (Thermo Fisher Scientific). Samples were analyzed on an LSRII cytometer (BD Biosciences), and data analysis was performed using FlowJo vX.0.7 (Tree Star).
FACS-based ADCC assay. Measurement of ADCC using the fluorescence-activated cell sorter (FACS)-based assay was performed at 48 h postinfection as previously described (20,45,67,73). Briefly, infected primary CD4 ϩ T cells were stained with AquaVivid viability dye and cell proliferation dye (eFluor670; eBioscience) and used as target cells. Autologous PBMC effector cells, stained with another cellular marker (cell proliferation dye eFluor450; eBioscience), were added at an effector/target ratio of 10:1 in 96-well V-bottom plates (Corning). The indicated concentrations of ADCC-mediating MAbs (0.3125, 0.625, 1.25, 2.5, or 5 g/ml) with 50 M BNM-III-170, or equivalent amounts of DMSO, when specified, were added to appropriate wells, and cells were incubated for 15 min at room temperature. The plates were subsequently centrifuged for 1 min at 300 ϫ g and were incubated at 37°C and 5% CO 2 for 5 h before being fixed in a 2% PBS-formaldehyde solution. Samples were acquired on an LSRII cytometer (BD Biosciences), and data analysis was performed using FlowJo vX.0.7 (Tree Star). The percentage of ADCC was calculated by gating on infected live target cells with the following formula: (percent p24 ϩ cells in targets plus effectors) Ϫ (percent p24 ϩ cells in targets plus effectors plus Abs)/(percent p24 ϩ cells in targets).
In silico modeling. Models of the engagement of dimeric rsFc␥RIIIa receptor by anti-CoRBS and anti-cluster A Abs IgGs bound to the target monomeric gp120 immobilized on a plate surface were built based on available crystal structures as follows. The crystal structures of complexes of N5-i5 Fab-gp120 93TH057 core e -d1d2CD4 (PDB code 4H8W) (12) and 17b Fab-gp120 HXBC2 core-d1d2CD4 (PDB code 1RZJ) (63) were superimposed based on gp120 with the IgG for each antibody built by overlaying murine antibody subclass IgG1 (PDB code 1IGY) (74) onto Fabs from the two complexes. The model was assembled assuming monovalent binding of IgG1, and the dimeric rsFc␥RIIIa receptor was built using the crystallographic dimer of Fc␥RIIa-HR (PDB code 3RY5; high-responder polymorphism Fc␥RIIa) (66) as a template. The interaction between N5-i5 and 17b Fcs and dimeric rsFc␥RIIIa was modeled based on the complex of Fc␥RIIIa bound to the Fc of IgG1 (PDB code 1E4K) (75).
Statistical analyses. Statistics were analyzed using GraphPad Prism version 6.01 (GraphPad, San Diego, CA, USA). Every data set was tested for statistical normality, and this information was used to apply