A Rare Mutation in an Infant-Derived HIV-1 Envelope Glycoprotein Alters Interprotomer Stability and Susceptibility to Broadly Neutralizing Antibodies Targeting the Trimer Apex

The design of HIV-1 envelope-based immunogens capable of eliciting broadly neutralizing antibodies (bnAbs) is currently under active research. Some of the most potent bnAbs target the quaternary epitope at the V2 apex of the HIV-1 Env trimer. By studying naturally circulating viruses from a perinatally HIV-1-infected infant with plasma neutralizing antibodies targeted to the V2 apex, we identified a rare leucine-to-phenylalanine substitution, in two out of six functional viral clones, that destabilized the trimer apex. This single-amino-acid alteration impaired the interprotomeric interactions that stabilize the trimer apex, resulting in an open trimer conformation and escape from broadly neutralizing autologous plasma antibodies and known V2 apex-directed bnAbs, thereby favoring viral evasion of the early bnAb response of the infected host. Defining the mechanisms by which naturally occurring viral mutations influence the sensitivity of HIV-1 to bnAbs will provide information for the development of vaccines and bnAbs as anti-HIV-1 reagents.

ABSTRACT

The envelope glycoprotein (Env) of human immunodeficiency virus type 1 (HIV-1) is the sole target of broadly neutralizing antibodies (bnAbs). Several mechanisms, such as the acquisition of mutations, variability of the loop length, and alterations in the glycan pattern, are employed by the virus to shield neutralizing epitopes on Env to sustain survival and infectivity within the host. The identification of mutations that lead to viral evasion of the host immune response is essential for the optimization and engineering of Env-based trimeric immunogens. Here, we report a rare leucine-to-phenylalanine escape mutation (L184F) at the base of hypervariable loop 2 (population frequency of 0.0045%) in a 9-month-old perinatally HIV-1-infected infant broad neutralizer. The L184F mutation altered the trimer conformation by modulating intramolecular interactions stabilizing the trimer apex and led to viral escape from autologous plasma bnAbs and known N160 glycan-targeted bnAbs. The L184F amino acid change led to the acquisition of a relatively open trimeric conformation, often associated with tier 1 HIV-1 isolates and increased susceptibility to neutralization by polyclonal plasma antibodies of weak neutralizers. While there was no impact of the L184F mutation on free virus transmission, a reduction in cell-to-cell transmission was observed. In conclusion, we report a naturally selected viral mutation, L184F, that influenced a change in the conformation of the Env trimer apex as a mechanism of escape from contemporaneous plasma V2 apex-targeted nAbs. Further studies should be undertaken to define viral mutations acquired during natural infection, to escape selection pressure exerted by bnAbs, to inform vaccine design and bnAb-based therapeutic strategies.

IMPORTANCE The design of HIV-1 envelope-based immunogens capable of eliciting broadly neutralizing antibodies (bnAbs) is currently under active research. Some of the most potent bnAbs target the quaternary epitope at the V2 apex of the HIV-1 Env trimer. By studying naturally circulating viruses from a perinatally HIV-1-infected infant with plasma neutralizing antibodies targeted to the V2 apex, we identified a rare leucine-to-phenylalanine substitution, in two out of six functional viral clones, that destabilized the trimer apex. This single-amino-acid alteration impaired the interprotomeric interactions that stabilize the trimer apex, resulting in an open trimer conformation and escape from broadly neutralizing autologous plasma antibodies and known V2 apex-directed bnAbs, thereby favoring viral evasion of the early bnAb response of the infected host. Defining the mechanisms by which naturally occurring viral mutations influence the sensitivity of HIV-1 to bnAbs will provide information for the development of vaccines and bnAbs as anti-HIV-1 reagents.

INTRODUCTION

Elicitation of antibodies capable of neutralizing globally circulating human immunodeficiency virus type 1 (HIV-1) viral variants is one of the vital goals of HIV-1 vaccine research (1, 2). The HIV-1 envelope glycoprotein (Env) is a trimer of noncovalently linked heterodimers [(gp120-gp41)₃] and is the primary target of broadly neutralizing antibodies (bnAbs). Persistent antigenic stimulation and viral diversification under immune selection pressure are typically associated with the development of bnAbs, although infected infants have been reported to develop bnAbs as early as 1 year after infection (3–5). The bnAbs targeting viral Env are grouped by epitope class: variable loop 2 and the N160 glycan (V2 apex), the third variable loop and the N332 glycan (V3/N332 glycan supersite, or the high-mannose patch), the CD4 binding site (CD4bs), the gp120-gp41 interface region, the silent face center, and the membrane-proximal external region (MPER) of gp41 (6, 7). These bnAbs are capable of neutralizing diverse circulating variants of HIV-1 and have evolved in rare subsets of infected individuals. The passive administration of such bnAbs in animal models has shown protection from HIV-1 infection (8–10), and studies conducted in humans have shown that passive immunotherapy with bnAbs is effective in the suppression of viremia (11–16).

No vaccination approach has been successful in inducing bnAbs in humans or standard animal models, although vaccination with soluble gp140 SOSIP-stabilized native-like trimers has thus far induced strain-specific and cross-subtype-specific nAbs (17–21). To overcome the high level of genetic diversity of HIV-1 envelope genes, strategies to induce antibodies that cross-react with multiple strains of HIV-1 are required. Considerable interest exists in the field relevant to viral features associated with induction and escape mechanisms responsible for V2 apex bnAbs as these bnAbs have been reported to emerge early (22–24), are elicited frequently (25–28), possess relatively low to moderate levels of somatic hypermutations compared to other bnAbs (22–25, 27–29), and show cross-group-neutralizing activity with Envs of HIV-1 groups M, N, O, and P (30, 31), thereby identifying the V2 apex as one of the promising Env epitopes for vaccine design. The extraordinary ability of HIV-1 to evade host immunity represents a major obstacle to the development of a protective vaccine. Thus, elucidating the mechanisms employed by HIV-1 to protect its external Env, which is the sole target of virus-neutralizing antibodies, is an essential step toward developing rational strategies for optimizing Env-based immunogens.

In a recently reported cohort of HIV-1-infected infants with an early plasma bnAb response targeting the envelope glycoprotein, we identified a 9-month-old infant, AIIMS731, whose plasma bnAbs showed maximum dependence on the V2 apex, with 75% breadth at a geometric mean titer (GMT) of 130 against the standardized 12-virus global panel representing global viral diversity (32). In order to understand the virus-antibody dynamics in the context of neutralizing determinants within the V2 apex and early induction of V2 apex-targeting plasma bnAbs, here, we studied the viral features associated with escape from plasma bnAbs in an infant broad neutralizer, AIIMS731. A rare leucine-to-phenylalanine mutation (L184F) that impaired the interprotomer interactions and consequently led to an open Env trimer conformation was identified. Of note, this rare mutation provided escape from autologous plasma nAbs as well as several known bnAbs targeting the V2 apex and increased viral susceptibility to the V3- and CD4-induced (CD4i)-directed non-nAbs. Understanding the impact of such rare escape mutations on the phenotypic properties of HIV-1 Env will aid in vaccine design.

RESULTS

A rare mutation at the base of hypervariable loop 2 of viral Env confers resistance to autologous plasma bnAbs in an HIV-1-infected infant broad neutralizer, AIIMS731.

On the basis of plasma neutralization data against the difficult-to-neutralize (tier 2/3) global panel of HIV-1 isolates and epitope mapping done using single-base mutants in 25710_2_43, 16055_2_3, CAP45_G3, and BG505.W6M.C2 pseudoviral backbones (32), an HIV-1-infected infant, AIIMS731, was previously categorized as a broad neutralizer with plasma bnAbs targeting the N160 glycan in the V2 apex of HIV-1 Env (Fig. 1A and B). Pseudoviruses grown in the presence of the glycosidase inhibitors kifunensine and swainsonine were resistant to neutralization by AIIMS731 plasma nAbs, highlighting the inability of these plasma nAbs to tolerate an increased density of both high-mannose as well as complex glycans on viral Env.

FIG 1.

Plasma bnAbs from infant broad neutralizer AIIMS731 target the V2 apex of HIV-1 Env. (A) Heat map representing HIV-1-specific neutralization titers (inverse plasma dilution) of plasma bnAbs from infant broad neutralizer AIIMS731 against the 12-virus global panel. ID₅₀ values are color-coded per the key given, with darker colors implying higher ID₅₀ titers. MuLV, murine leukemia virus; NA, not applicable. (B) Epitope mapping of AIIMS731 plasma bnAbs showing ID₅₀ fold changes against the 25710_2_43, 16055_2_3, CAP45_G3, and BG505_W6M_C2 wild-type pseudoviruses and their N160A mutants as well as pseudoviruses grown in the presence of the glycosidase inhibitors kifunensine and swainsonine.

In order to evaluate the viral population dynamics associated with the presence of V2 apex plasma bnAbs in this infant broad neutralizer, we first cloned functional Env genes from contemporaneous plasma RNA at 9 months postinfection (Fiebig stage VI) via single-genome amplification (SGA) and assessed their susceptibility to autologous plasma bnAbs. A total of 40 Env gene sequences (clade C) from infant AIIMS731 were available (32), and the depth of SGA sequencing gave us a 90% confidence interval of identifying circulating variants present at a population frequency of 5%. Based on sequence identity, the SGA sequences represented the six dominant R5-tropic circulating strains in AIIMS731 plasma and were highly homogeneous, with sequence variability between the clones ranging from 0.1 to 0.4% (Fig. 2A to C). Viral variants 73105h and 73106f within the cluster were the dominant circulating strains, while 73105b, 73105c, 73105e, and 73105d represented the remaining circulating strains (with a population frequency of >5%). Each of the six clusters on the phylogenetic tree consisted of identical sequences, and from each cluster, a single viral variant was cloned into the pcDNA3.1(+) mammalian expression vector and pseudotyped for neutralization assays.

FIG 2.

Limited diversity in the circulating viral variants of infant AIIMS731. (A and B) Highlighter plots with maximum likelihood trees of 40 SGA env sequences from infant AIIMS731 showing limited variability in the circulating viral variants and the presence of 6 strains circulating at a population frequency of >5%. In the highlighter plot, mutations compared to the consensus sequence are represented by green for adenine, blue for cytosine, orange for guanine, and red for thymine. In the case of 73105h, both of the adenine mutations were silent (green bars). (C) Donut plot of the distribution of 40 SGA Env amplicons showing viral variants 73106f and 73105h to be the dominant strains.

Despite the high degree of similarity between the viral variants, a consistent hierarchy of neutralization sensitivity to contemporaneous autologous plasma bnAbs was observed (Fig. 3A). Viral clone 73105b showed nearly complete neutralization (maximum percent neutralization [MPN] of 87% ± 4%) by autologous plasma bnAbs, while clones 73105h and 73106f showed a significant abrogation of neutralization sensitivity (maximum percent neutralization of 24% ± 3%) to autologous plasma bnAbs (1:50 was the lowest dilution tested). Despite neutralizing autologous circulating variants with 50% inhibitory dilution (ID₅₀) titers (Fig. 3B) roughly 3-fold higher than the median titers against the multiclade panel of HIV-1 isolates (median ID₅₀ of 362 versus 127), none of the autologous viruses were completely neutralized by plasma bnAbs, with MPN values for sensitive viruses ranging from 83 to 91%. Both clones 73105h and 73106f had MPN values in the range of 21 to 27%. Of note, plasma bnAb-resistant viral variants 73105h and 73106f were the dominant circulating strains (13 and 15 of the 40 SGA sequences, respectively) (Fig. 2C).

FIG 3.

A rare mutation provided neutralization escape from autologous plasma bnAbs in infant AIIMS731. (A) The neutralization susceptibility of circulating viral variants from infant AIIMS731 to contemporaneous autologous plasma bnAbs was assessed via neutralization assays based on TZM-bl cells. Even though four of the circulating viral variants were susceptible to plasma bnAbs, the maximum percent neutralization ranged from 83 to 91%. Viral variants 73105h and 73106f were resistant to autologous plasma bnAbs (maximum percent neutralization values of 24% and 25%, respectively). (B) AIIMS731 viral strains were arranged based on neutralization sensitivity to contemporaneous plasma nAbs (ID₅₀ values). V2 loop sequences (positions 156 to 199 [HXB2 numbering]) of viral variants resistant to neutralization by autologous plasma nAbs showed a leucine-to-phenylalanine mutation at position 184. (C) Complete amino acid sequence comparison among all six viral variants showing a single mutation (L184F) in 73106f (relative to sensitive strain 73105b) that led to neutralization escape from autologous plasma bnAbs. While 73105c also had a rare N229Y mutation, no difference in susceptibility to autologous plasma bnAbs was visible. (D) Amino acid frequency plot at position 184 in all reported HIV-1 Env sequences (7,094 sequences) showing an abundance of leucine or isoleucine, whereas phenylalanine at position 184 occurred at a population frequency of 0.0045% (32/7,094 sequences).

To identify residues conferring resistance to contemporaneous autologous plasma bnAbs, we conducted a comparative sequence analysis. Examination of the core V2 apex bnAb epitope revealed no sequence change despite varying neutralization sensitivities between the viral clones (Fig. 3B). Of particular note, all the circulating viral variants retained the key epitope-defining patterns of specific amino acid residues and N-linked glycosylation sites in the V2 apex bnAb epitope. Mutations were mapped relative to variant 73105b as it represented the consensus amino acid sequence and showed the highest susceptibility to autologous plasma bnAbs (Fig. 3B and C). In the case of clone 73105h, in addition to L184F, an additional mutation, I255V (small, nonpolar side chain mutated to another small, nonpolar side chain), within the C2 region was observed. For clone 73105c, N229Y was observed, while for clones 73105d (D135E and S143T) and 73105e (S143T), mutations within hypervariable loop 1 (positions 132 to 152 [HXB2 numbering]) were observed. A single change of L184F was observed in clone 73106f (neutralization resistant) compared to clone 73105b (most susceptible to neutralization), suggesting that a mutation outside that of the bnAb-targeting epitopes in the V2 region may have led to the viral escape of clone 73106f from the V2 apex-targeting autologous plasma bnAbs (Fig. 3B and C). D135E, S143T, and N229Y had no impact on neutralization by autologous plasma bnAbs and were therefore excluded from further analysis.

To assess the frequency of the L184F mutation that led to the replacement of a small nonpolar side chain with a bulky nonpolar side chain, we examined the variability (amino acid changes) at position 184 (HXB2 numbering) in all reported HIV-1 Env sequences. Analysis of 7,094 Env sequences from the Los Alamos National Laboratory (LANL) HIV-1 sequence database revealed the extreme rarity of the L184F mutation. The presence of phenylalanine at position 184 was found in 32 of the 7,094 (population frequency of 0.0045%) reported viral sequences available in the HIV database, with 17 instances reported for clade C (Table 1). Position 184 most commonly contains either isoleucine or leucine, at population frequencies of 59.3 and 28.6%, respectively (Table 1 and Fig. 3D). Overall, these results suggested the acquisition of a rare mutation by both the 73105h and 73106f viral clones that may have led to viral escape from plasma bnAbs in infected infant AIIMS731.

TABLE 1.

Amino acid frequencies at position 184^a

Variant	No. of Env sequences	Frequency
I	4,206	0.5929
L	2,029	0.2860
M	454	0.0640
T	138	0.0195
V	100	0.0141
— (gap)	50	0.0070
F	32	0.0045
S	20	0.0028
N	16	0.0023
D	11	0.0016
A	9	0.0013
P	6	0.0008
E	4	0.0006
R	4	0.0006
* (stop)	4	0.0006
Y	3	0.0004
H	2	0.0003
K	2	0.0003
Q	2	0.0003
C	1	0.0001
G	1	0.0001
W	0	0.0000

^aThe population frequency for position 184 was calculated using AnalyzeAlign (see Materials and Methods) for the Web alignment of all reported HIV-1 Env sequences (7,094) available in the LANL HIV database.

L184F provides neutralization escape from V2 apex-targeting bnAbs and contributes to the open configuration of the trimeric form.

Next, we evaluated the neutralization susceptibilities of all six viral variants from infant AIIMS731 to assess if the mutations acquired by these viral variants altered neutralization to known V2 apex bnAbs. For all V2 apex bnAbs tested (PG9, PG16, PGT145, PGDM1400, CAP256.25, and CH01), as observed with autologous plasma bnAbs, AIIMS731 viral variants segregated into neutralization-sensitive (73105b, 73105c, 73105d, and 73105e) and -resistant (73105h and 73106f) clusters (Fig. 4). For resistant variants, we observed marked increases in the 50% inhibitory concentrations (IC₅₀s) and reductions in the MPN values, with the most significant reduction being observed for PG9 and CAP256.25. Except for PGDM1400 and CAP256.25, two of the most potent V2 apex bnAbs known, none of the V2 apex bnAbs could reach 100% neutralization, even at higher concentrations for 73105b (most sensitive clone). For L184F mutant clones 73105h and 73106f, none of the bnAbs reached 100% neutralization, and they showed shallow dose-response curves. The slopes of the neutralization curves for the V2 apex bnAbs were steeper and had the expected sigmoidal curve for variants in the neutralization-sensitive cluster compared to L184F mutants 73105h and 73106f (Fig. 4), suggesting that the viral Envs in the sensitive cluster were plausibly homogeneously trimeric and uniformly recognized via their trimeric conformation, as the V2 apex-targeting bnAbs are reported to be trimer preferring and to target the HIV-1 Env trimer in a closed conformation (33–36).

FIG 4.

Neutralization curves of AIIMS731 viral variants against V2 apex bnAbs. The neutralization susceptibilities of all six viral variants to the V2 apex bnAbs (PG9, PG16, PGT145, PGDM1400, CAP256.25, and CH01) were assessed via neutralization assays based on TZM-bl cells. Of note, except for PGDM1400 and CAP256.25, none of the V2 apex bnAbs reached 100% neutralization for the AIIMS731 autologous plasma bnAb-sensitive viral cluster (73105b, 73105c, 73105d, and 73105e), while with the L184F mutant clones 73105h and 73106f (autologous plasma bnAb-resistant cluster), all V2 apex bnAbs showed markedly lower neutralization efficiencies. For 73106f, maximum neutralizations of 57% and 53% were reached with PG16 and CAP256.25, respectively. Neutralization assays were repeated three times, and curves were drawn with means ± standard deviations.

Next, we used an exhaustive panel of bnAbs targeting other known epitopes on Env in order to assess the neutralization efficiencies of bnAbs other than those targeting the V2 region by comparing the neutralization curves for all six viral variants. The bnAb panel consisted of V3/N332 glycan supersite bnAbs (10-1074, BG18, AIIMS-P01, PGT121, PGT128, and PGT135), CD4bs bnAbs (VRC01, VRC03, VRC07-523LS, N6, 3BNC117, and NIH45-46 G54W), a silent face-targeting bnAb (PG05), fusion peptide and gp120-gp41 interface bnAbs (PGT151, 35O22, and N123-VRC34.01), and MPER bnAbs (10E8, 4E10, and 2F5). Neutralization assays showed similar neutralization phenotypes of these bnAbs for all six variants regardless of their susceptibility to V2 apex bnAbs (Table 2), suggesting that the L184F mutation was specific for viral escape from susceptibility to neutralization by V2 apex bnAbs and had a negligible effect on neutralization by other classes of bnAbs.

TABLE 2.

Neutralization of 73105b and 73106f by known bnAbs^a

bnAb epitope	bnAb name	Mean IC50 for viral variant (μg/ml) ± SD
		73105b	73105c	73105d	73105e	73105h	73106f
V2 apex	PG9	0.062 ± 0.021	0.063 ± 0.018	0.041 ± 0.011	0.077 ± 0.046	9.362 ± 0.458	>10
	PG16	0.077 ± 0.025	0.102 ± 0.013	0.092 ± 0.024	0.124 ± 0.032	3.562 ± 0.321	4.246 ± 0.516
	PGT145	3.964 ± 0.112	2.256 ± 0.194	3.025 ± 0.183	2.056 ± 0.365	>10	>10
	PGDM1400	0.005 ± 0.002	0.002 ± 0.001	0.004 ± 0.001	0.001 ± 0.001	4.254 ± 0.187	5.031 ± 0.365
	CAP256.25	0.051 ± 0.016	0.036 ± 0.012	0.042 ± 0.024	0.041 ± 0.016	>10	>10
	CH01	2.454 ± 0.256	1.256 ± 0.178	2.065 ± 0.191	2.036 ± 0.324	>10	>10
V3 glycan	10-1074	3.654 ± 0.278	4.526 ± 0.325	3.256 ± 0.421	3.065 ± 0.364	2.451 ± 0.285	4.125 ± 0.427
	BG18	>10	9.632 ± 0.542	>10	>10	>10	>10
	AIIMS-P01	>10	>10	10	>10	>10	9.851 ± 0.425
	PGT121	3.654 ± 0.428	3.026 ± 0.387	2.063 ± 0.336	2.857 ± 0.374	1.023 ± 0.248	2.462 ± 0.363
	PGT128	>10	>10	>10	>10	>10	>10
	PGT135	>10	>10	>10	>10	>10	>10
CD4bs	VRC01	3.274 ± 0.412	2.658 ± 0.311	3.625 ± 0.305	3.524 ± 0.427	3.256 ± 0.523	2.664 ± 0.485
	VRC03	6.495 ± 0.421	5.236 ± 0.357	6.321 ± 0.325	4.256 ± 0.456	2.056 ± 0.322	4.125 ± 0.369
	VRC07-523LS	0.984 ± 0.098	0.742 ± 0.115	0.685 ± 0.086	0.954 ± 0.069	0.745 ± 0.125	0.847 ± 0.081
	N6	0.003 ± 0.001	0.003 ± 0.001	0.002 ± 0.001	0.004 ± 0.001	0.002 ± 0.001	0.003 ± 0.001
	3BNC117	1.259 ± 0.163	1.026 ± 0.125	1.625 ± 0.146	0.985 ± 0.163	2.056 ± 0.127	3.624 ± 0.164
	NIH45-46 G54W	1.026 ± 0.265	1.365 ± 0.243	1.025 ± 0.194	0.958 ± 0.117	0.635 ± 0.131	0.958 ± 0.125
Silent face	PG05	>10	>10	>10	>10	>10	>10
gp120-gp41 interface	PGT151	4.532 ± 0.343	3.652 ± 0.321	4.254 ± 0.354	5.256 ± 0.327	5.256 ± 0.412	6.412 ± 0.333
	35O22	>10	>10	>10	>10	>10	>10
	N123-VRC34.01	2.036 ± 0.288	2.214 ± 0.257	1.026 ± 0.225	1.856 ± 0.341	1.356 ± 0.197	1.241 ± 0.263
MPER	10E8	0.971 ± 0.163	0.748 ± 0.118	0.985 ± 0.107	1.255 ± 0.252	1.635 ± 0.225	1.654 ± 0.186
	4E10	4.632 ± 0.296	3.652 ± 0.342	3.658 ± 0.306	4.251 ± 0.287	2.364 ± 0.168	2.023 ± 0.246
	2F5	>10	>10	>10	>10	>10	>10

^aNeutralization susceptibilities of AIIMS731 viral variants were assessed by utilizing a broad panel of bnAbs targeting all major antigenic sites on HIV-1 Env. IC₅₀ (50% inhibitory concentration) values for all tested bnAbs are shown and grouped according to the antigenic sites (V2 apex, V3/N332 glycan supersite, CD4bs, silent face, gp120-gp41 interface, and MPER). Neutralization assays were performed with TZM-bl cells and repeated three times. For bnAbs that neutralized AIIMS731 viruses, IC₅₀ values (micrograms per milliliter) are reported as means ± standard deviations. For bnAbs that failed to neutralize AIIMS731 viruses at 10 μg/ml, values are reported as >10.

In neutralization assays performed with nonneutralizing antibodies (non-nAbs) targeting the V3 loop (447-52d and 19b) and CD4-induced epitopes (17b, A32, 48d, and b6), L184F mutants 73105h and 73106f showed neutralization by the V3 loop non-nAbs (MPN values of 46% ± 8% and 48% ± 4% for 447-52D and MPN values of 63% ± 7% and 61% ± 8% for 19b at 50 μg/ml, respectively) and CD4i non-nAbs (MPN values of 61% ± 6% and 66% ± 5% for 17b and MPN values of 46% ± 6% and 48% ± 8% for 48d at 50 μg/ml, respectively) (Fig. 5A and B). Of note, 17b binds preferentially to the CD4-induced CCR5 coreceptor binding site epitope on Env (37). Thus, the L184F mutation resulted in increased susceptibility to neutralization by antibodies known to target the relatively more open conformation of Env on tier 1A/B viruses, suggesting that the rare L184F mutation allowed Env to sample more open states resembling the CD4-bound conformation where the CCR5 binding site is exposed, although this observation can be accurately validated only by undertaking in-depth structural studies.

FIG 5.

Neutralization of AIIMS731 viral variants by non-nAbs targeting the V3 loop and CD4-induced epitopes. The neutralization susceptibilities of all six viral variants to the V3 loop-targeting non-nAbs (447-52D and 19b) and CD4-induced non-nAbs (17b, A32, 48d, and b6) were assessed via neutralization assays based on TZM-bl cells. Viral variants 73105h and 73106f showed moderate neutralization by 447-52D, 19b, 17b, and 48d. Neutralization assays were repeated three times, curves were drawn with means ± standard deviations, and MPN values were calculated based on average neutralization.

Taken together, these results suggest that the L184F mutation conferred resistance to neutralization via trimer-preferring V2 apex bnAbs and allowed the Env trimer to transition toward a more open configuration that partially exposed the occluded non-nAb epitopes within the V3 loop and CD4bs.

Preferential recognition of the closed Env trimer by potent plasma antibodies from pediatric neutralizers.

Destabilization of the trimer apex has been shown to alter the neutralization susceptibility of HIV-1 Env to antibodies present in the plasma of infected individuals. As the L184F mutation resulted in a more open trimer configuration, we next evaluated the sensitivity of the L184F mutant to a panel of HIV-1 clade C-infected pediatric patient plasma samples with varied neutralization potencies (weak versus strong neutralizers) against the global panel of representative HIV-1 isolates (38–40, 55). Patient plasma antibodies neutralizing more than half of the global panel were considered strong neutralizers, while those neutralizing less than half of the panel were considered weak to moderate neutralizers, depending on their breadth and potency. The neutralization susceptibility profile of L184F mutants 73105h and 73106f to plasma antibodies of well-characterized HIV-1 clade C-infected pediatric donors (whose plasma antibodies showed varied neutralization activities against the 12-virus global panel) showed comparable ID₅₀ values between weak and strong neutralizers (Fig. 6A to C), confirming that viral variants belonging to the autologous plasma bnAb neutralization-sensitive cluster (73105b, 73105c, 73105d, and 73105e) showed a tier 2 phenotype (closed trimeric conformation), while viral variants 73105h and 73106f had a tier 1B neutralization phenotype (open trimeric conformation).

FIG 6.

Viral variant 73106f is highly susceptible to subtype-matched heterologous plasma antibodies. (A and B) Violin plot and heat map representing the neutralization susceptibilities of AIIMS731 circulating viral variants against plasma antibodies from HIV-1-infected pediatric individuals in chronic stages of disease. Distinct neutralization profiles were seen for AIIMS731 autologous plasma bnAb-sensitive (73105b, 73105c, 73105d, and 73105e) and -resistant (73105h and 73106f) viral clusters. The plasma panel contained well-characterized HIV-1 clade C-infected pediatric donors whose plasma antibodies showed varied neutralization activities against the 12-virus global panel. Plasma samples were categorized as strong or weak based on their breadth and potency against the 12-virus global panel (see Materials and Methods). A comparison is shown for 73105b and 73106f, although similar patterns were observed by comparing sensitive versus resistant clusters. (C) Viruses belonging to the sensitive cluster were primarily neutralized by plasma samples that were categorized as strong, while viruses belonging to the resistant cluster showed considerable neutralization by plasma samples categorized as weak. P values are given by asterisks, where ** implies a P value of <0.01, *** implies a P value of <0.001, and **** implies a P value of <0.0001.

The 73105b (most sensitive to autologous plasma bnAbs) virus clone was highly susceptible to the plasma of strong neutralizers, with ID₅₀ values ranging from 1:248 to 1:1,965, while its susceptibility to weak neutralizers ranged from 1:50 to 1:188 (1:50 was the lowest dilution tested). For the L184F mutant 73106f, the ID₅₀ titers of strong neutralizers (range, 1:388 to 1:2,756) versus weak neutralizers (range, 1:176 to 1:1,246) had similar profiles, suggesting that regardless of their ability to generate antibodies capable of targeting closed Env trimers, weak neutralizers develop high titers of antibodies targeting the open configuration of the Env trimer.

The L184F escape mutation results in reduced entry kinetics in cell-to-cell transmission.

As V1V2 stabilizes the Env spike forming the trimer apex, we next performed a stability-of-function assay, called the T90 assay, which determines Env stabilization (viral infectivity) as a function of temperature (41, 42), to evaluate the effect of the L184F mutation on Env stability. A slight increase in the T90 value, the temperature at which viral infectivity decreased by 90% in 1 h, from 43.27°C to 43.69°C (P = 0.19), was observed between viral strain 73105b and the L184F mutant 73106f, although the impact of L184F on thermal stability did not appear markedly noticeable (Fig. 7A and B). We further tested AIIMS731 viral variants for infectivity decay at 37°C and did not observe any change in the half-life of infectivity decay (Fig. 7C). Overall, the impact of the L184F mutation on thermal stability and functional infectivity decay (Env decay) suggested that the rare mutation did not alter trimer stability.

FIG 7.

The L184F mutation has no impact on Env trimer stability. (A to C) AIIMS731 viral variants were tested for stability by assessing decay in infectivity as a function of temperature (thermostability) and time (spontaneous decay). Similar infectivity decay curves were observed for all viral variants, and no significant change in the T90 value (temperature at which 90% of infectivity was lost) or half-life of Env decay was observed. Infectivity decay assays were repeated three times in triplicates, and the curves were drawn using means ± standard deviations. For thermostability, infectivity at 37°C was taken as 100% infectivity. For spontaneous decay, infectivity at 0 h was taken as 100% infectivity.

Changes at the trimer apex have been shown to alter virus sensitivity and often come with a fitness cost (42–45). In order to investigate the effect of the acquisition of the rare L184F viral immunotype on the functional stability of Env, we assessed the impact of the L184F mutation on viral infectivity in free virus and cell-to-cell transmission. The relative infectivity of all six viral variants was assessed by titration curves after normalizing pseudoviral infectivity by using the viral stock dilution that gave a relative luminescence unit (RLU) value of 150,000 in TZM-bl cells. No change in the infectivity of the L184F mutant was observed in the case of infection with free virus, although we detected substantial variability in the entry kinetics of L184F mutants 73105h and 73106f in cell-to-cell transmission (Fig. 8A and B), suggesting a plausible fitness cost associated with escape from plasma bnAbs via the acquisition of the rare viral L184F immunotype, which needs to be further confirmed.

FIG 8.

The L184F mutation leads to reduced cell-to-cell transmission. (A) Pseudoviruses were titrated after normalization, and replicate titration curves were used to calculate the area under the curve (AUC) values. Each experiment was repeated three times in triplicates, providing a total of 9 reference values. (B) Fusogenicity in cocultures of Tat/Env-cotransfected 293T and TZM-bl cells was used as a measure of cell-to-cell transmission ability. The fusion of AIIMS731 Env in relation to the fusion observed with the well-characterized Env of HIV-1 isolate MW965.26 was calculated. Each experiment was repeated three times in triplicates, providing a total of 9 reference values. Two-tailed Student’s t test was used for comparison (*** implies a P value of <0.001). A comparison is shown for 73105b and 73106f, although similar patterns were observed by comparing sensitive versus resistant clusters.

The L184F escape mutation impaired interprotomer interactions at the trimer apex.

The V2 apex bnAbs target quaternary epitopes formed by interprotomeric interactions at the apex of the HIV-1 Env trimer. The core epitope for V2 apex bnAbs is formed by the N-linked glycan sites N156 and N160 and the lysine-rich region of strand C (positions 156 to 177 [HXB2 numbering]). As the L184F escape mutation did not arise within the core epitope, and as this mutant virus showed a tier 1 neutralization phenotype, we reasoned that the L184F mutation was plausibly responsible for disrupting the interprotomer interactions that stabilize the closed conformation of the Env trimer.

To elucidate the mechanism by which the L184F mutation could have altered the conformation of the Env trimer, we analyzed the L184F mutation using the ligand-free prefusion closed structure of the BG505 SOSIP.664 HIV-1 Env trimer (PDB accession number 4ZMJ). Residues 165 and 184 in BG505 were changed to their counterparts (R165 and L184) in 73105b. In previous reports, I184 of one protomer has been shown to interact with L165 of the neighboring protomer. This interprotomeric interaction has been shown to be critical for quaternary interactions leading to the stabilization of the V1V2 regions of neighboring protomers, and its loss has been shown to render JR-FL, a clade B HIV-1 strain, highly sensitive to V3 monoclonal antibodies (mAbs) (45, 46). On similar lines, we observed L184 of one protomer interacting with R165 of another protomer (via van der Waals interactions between the solvent-accessible surfaces [SASs] of R165 and L184 on neighboring protomers) (Fig. 9A). The side chain of L184 was observed to be outward facing and did not make significant intraprotomeric interactions. Upon mutating L184 to F184, a disruption of the SAS between the bulky side chain of F184 on one protomer and R165 on the neighboring protomer was seen (Fig. 9B). In addition, we generated a homology model based on the sequence of 73105b based on multiple structural templates, and after loop refinement, Man-9 glycan sites were added to potential N-linked glycosylation sites (PNGSs) in silico to produce a nearly fully glycosylated gp160 trimeric model. In the 73105b homology model, similar interprotomeric interactions were seen between L184 of one protomer and R165 of the neighboring protomer, which were lost when L184 was mutated to F184.

FIG 9.

Critical role of L184 in modulating interprotomer interactions at the trimer apex. Shown are interprotomer interactions between R165 (protomer A) and L184 (protomer B). The dot meshes surrounding R165_A (orange) and L184_B and F184_B (yellow) represent the solvent-accessible surfaces (SASs) (van der Waals surfaces expanded by the water molecule radius). In panel A, interprotomer contacts (lipophilic) between R165_A and L184_B can be seen by overlapping SASs. In the case of F184_B, no interprotomer contacts can be seen, evident by the lack of SAS overlap. The HIV-1 Env protomeric backbones are represented by two distinct colors (protomer A in green and protomer B in cornflower blue). The illustration was generated from data under PDB accession number 4ZMJ. L165 and I184 were rotamerized to the respective residues in 73105b (R165 and L184) and 73106f (R165 and F184).

DISCUSSION

During HIV-1 infection, the humoral immune response targets the HIV envelope glycoprotein (Env), which consists of three heavily glycosylated noncovalently linked gp41-gp120 protomers (3–7). While strain-specific antibodies recognize exposed and variable sites, bnAbs target relatively conserved and occluded sites, including the quaternary V1V2 epitope at the trimer apex (V2 apex), the V3/N332 glycan supersite, the CD4 binding site (CD4bs), the gp120-gp41 interface, and the membrane-proximal external region (MPER). Of these, bnAbs targeting the quaternary V1V2 epitope (called V2 apex bnAbs) are elicited frequently and relatively early (22–29). HIV-1 eludes recognition by host bnAbs through a variety of mechanisms, although the most common mechanism includes sequence alterations that can lead to large variations in sensitivity to antibody-mediated neutralization among different circulating viral isolates. Here, we investigated the viral escape mechanisms in a 9-month-old perinatally HIV-1-infected infant with broadly neutralizing plasma antibodies targeting the V2 apex.

Escape from contemporaneous autologous plasma bnAbs occurred by a rare leucine-to-phenylalanine mutation at position 184. Interestingly, a high degree of similarity was observed between circulating viral strains, regardless of their sensitivity to plasma bnAbs, and the L184F mutation, alone, was enough for escape from neutralization by plasma bnAbs. Similar neutralization profiles for sensitive and resistant strains were observed when susceptibility to reported bnAbs targeting diverse epitopes on Env was assessed. While the L184F mutation did not alter the neutralization profile of the viral strains to bnAbs targeting the V3/N332 glycan supersite, the CD4 binding site, the gp120-gp41 interface, or the MPER, a significant reduction in neutralization susceptibility to several V2 apex bnAbs was seen. The most significant reduction observed was for CAP256.25 (also referred to as VRC25.26), a trimer-specific bnAb that recognizes HIV-1 Env trimers via its long protruding loop that interacts with strand C, insertions into the apex hole at the trimer 3-fold axis, as well as electrostatic interactions with cationic V1V2 surface residues (22, 33, 36). Given the complex mode of trimer recognition via CAP256.25, which is a combination of the PG9 and PGT145 classes of bnAbs, and, therefore, its stringent need for a closed conformation of the Env trimer (36), the L184F mutation most likely seemed to alter the conformation of the Env trimer.

As V2 apex bnAbs target the closed conformation of the trimer, based on the neutralization profile observed, we hypothesized that the most plausible reason for the loss of susceptibility to V2 apex bnAbs was the loss of the closed conformation of Env. In its closed conformation, the Env trimer occludes several immunodominant epitopes that are targeted by nonneutralizing antibodies (non-nAbs) (42, 44). When the neutralization profile against several non-nAbs (those targeting CD4-induced epitopes and the V3 loop) was assessed, the L184F mutant virus showed relatively higher neutralization susceptibility to non-nAbs. Of note, four of the six non-nAbs could achieve 50% neutralization below 50 μg/ml (17b, which binds preferentially to the CD4-induced CCR5 coreceptor binding site epitope on Env, achieved a maximum percent neutralization of 66% ± 5%) (47), suggesting that the L184F mutation exposed immunodominant nonneutralizing epitopes that are predominantly concealed in the closed trimeric conformation of Env. In addition, the L184F mutant virus showed high susceptibility to neutralization by polyclonal antibodies present in the plasma of weak neutralizers (HIV-1-infected patients who do not generate a potent response to HIV-1 Env) (38–40). Typically, such weak polyclonal responses consist of antibodies that target epitopes that are primarily exposed on monomeric gp120 or open Env trimers (such as non-nAbs targeting the V3 loop and epitopes exposed after CD4 binding [CD4i]) but are concealed in native closed trimers.

The neutralization tier phenotypes of HIV-1 isolates can be understood in the context of the dynamic nature of Env trimers on the virus surface (47, 48). These trimers spontaneously transition between closed, open, and at least one intermediate conformation. Open trimers expose more epitopes than closed trimers and are typically reported to have a tier 1 neutralization phenotype, while trimers in the closed state have a tier 2/3 neutralization phenotype. Tier 3 Envs have been shown to exist exclusively in a relatively narrow range of closed conformations. The specific amino acid residues within the V1/V2 region ofgp120 that restrict the Env transition from a metastable closed conformation (state 1) via an intermediate to an open conformation were described previously in extensive mutational analysis studies (49). Among a number of distinct residues identified, which also included L184, the hydrophobic amino acid Leu at position 193, highly conserved across viral clades, was shown to play a key role in maintaining the closed conformation of the trimer (49). Here, by characterizing infant circulating viruses, we showed that the selection of L184 during natural infection led to the acquisition of a relatively open trimeric conformation, which in turn led to escape from autologous plasma nAbs and V2 apex-directed bnAbs. HIV-1 Env evades recognition by antibodies through diverse mechanisms, but one of the most effective mechanisms is the adoption of a closed trimeric configuration (typically associated with tier 2 or 3 Envs). Our observation of the acquisition of a rare mutation that led to an open, yet inaccessible to plasma V2 apex bnAbs, Env state suggests that HIV-1 can lose its trimeric Env form to evade plasma bnAbs.

In silico structural analysis was then utilized to reveal the molecular feature that conferred resistance to V2 apex bnAbs. The replacement of the small side chain of leucine with the bulky nonpolar side chain of phenylalanine led to the disruption of interprotomer interactions that have been observed to be critical for the maintenance of the trimeric apex. Mutations that alter or disrupt interprotomer contacts in the Env trimer have been shown to change its susceptibility to several classes of bnAbs (42, 44–46). In silico structural analyses were performed using the solved structures of soluble gp140 SOSIP-stabilized native-like trimers, and although recent studies have shown that these SOSIP trimers adopt conformations that differ from those of naturally occurring Env trimers, they are the closest mimic of the functional Env spike in its stable, prefusion conformation. Overall, our results provide information on the role of a rare escape mutation, L184F, in the viral envelope that led to resistance to bnAbs targeting the V2 apex. Furthermore, our data suggest a prominent role of L184-mediated intramolecular interactions that are necessary for the maintenance of the trimer apex and add to the information on conformational epitopes on the HIV-1 envelope toward the development of effective vaccines. Evaluation of viruses across different clades for the acquisition of escape mutations during natural infection will help in understanding viral evolutionary dynamics and provide information toward the development of polyvalent vaccines and bnAb-based therapies.

MATERIALS AND METHODS

Study design and participants.

The present study was designed to assess the viral population dynamics in an infant broad neutralizer with plasma bnAbs targeting the V2 apex of HIV-1 Env. At recruitment, infant AIIMS731 was antiretroviral naive and asymptomatic; was 9 months old, with a CD4 count of 1,385 cells/mm³ and a viral load on a log scale of 5.894 RNA copies/ml (785,000 RNA copies/ml); and was recruited from the Pediatric Chest Clinic, Department of Pediatrics, All India Institute of Medical Sciences (AIIMS). After written informed consent was obtained from guardians, blood was drawn into 3-ml EDTA vials, and plasma was aliquoted for plasma neutralization assays, viral RNA isolation, and viral load determinations. The study was approved by the institute ethics committee of the All India Institute of Medical Sciences (approval number IECPG-307/07.09.2017).

Plasmids, viruses, monoclonal antibodies, and cells.

Plasmids carrying HIV-1 env genes representing different clades, monoclonal antibodies, and TZM-bl cells were procured from the NIH AIDS Reagent Program. 10-1074 and BG18 expression plasmids were kindly provided by Michel Nussenzweig, Rockefeller University; VRC07-523LS and N123.VRC34.01 expression plasmids were provided by John Mascola, VRC, NIH; and PG05 expression plasmids were provided by Peter Kwong, VRC, NIH. CAP256.09, CAP256.25, and b6 were procured from the IAVI Neutralizing Antibody Center. 293T cells were purchased from the American Type Culture Collection (ATCC).

HIV-1 envelope sequences and phylogenetic analysis.

HIV-1 envelope genes were PCR amplified from plasma viral RNA by single-genome amplification and directly sequenced commercially. Individual sequence fragments of SGA-amplified amplicons were assembled using Sequencher 5.4 (Gene Codes Corporation). Subtyping for SGA sequences was performed with the REGA HIV subtyping tool (400-bp sliding window with a 200-bp step size). Interclade recombination was examined with RIP 3.0 (Recombinant Identification Program) and jpHMM. Nucleotide sequences were aligned with MUSCLE in MEGA X. Maximum likelihood trees were computed with MEGA X using a general-time-reversible substitution model incorporating a discrete gamma distribution with 5 invariant sites.

Cloning of autologous HIV-1 envelope genes and production of replication-incompetent pseudoviruses.

Autologous replication-incompetent envelope pseudoviruses were generated from infant AIIMS731 as described previously (32, 40). Briefly, viral RNA was isolated from 140 μl of plasma using the QIAamp viral RNA minikit and reverse transcribed, using gene-specific primer OFM19 (5′-GCACTCAAGGCAAGCTTTATTGAGGCTTA-3′) and Superscript III reverse transcriptase, into cDNA, which was used in a two-round nested PCR for the amplification of the envelope gene using Phusion high-fidelity DNA polymerase (New England Biolabs). The envelope amplicons were purified and ligated into the pcDNA3.1D_ITR vector via overlap extension cloning. Pseudoviruses were prepared by cotransfecting 1.25 μg of an HIV-1 envelope-containing plasmid with 2.5 μg of an envelope-deficient HIV-1 backbone (PSG3Δenv) vector at a molar ratio of 1:2 using PEI-Max as a transfection reagent in HEK293T cells seeded in 6-well culture plates. Culture supernatants containing pseudoviruses were harvested at 48 h posttransfection, filtered through a 0.4-μm filter, aliquoted, and stored at −80°C until further use. The 50% tissue culture infective dose (TCID₅₀) was determined by infecting TZM-bl cells with serially diluted pseudoviruses in the presence of DEAE-dextran and lysing the cells at 48 h postinfection. Infectivity titers were determined by measuring luminescence activity in the presence of Bright-Glo reagent (Promega).

Infectivity and neutralization assays.

Viral infectivity and neutralization assays were carried out using TZM-bl cells, a genetically engineered HeLa cell line that constitutively expresses CD4, CCR5, and CXCR4 and contains luciferase and β-galactosidase genes under the control of the HIV-1 tat promoter, as described previously (40). Viral infectivity was determined after normalizing pseudoviruses to an RLU value of 150,000, followed by titration curve generation to calculate relative infectivity. Neutralization studies included heat-inactivated plasma samples from infant AIIMS731, 30 previously characterized plasma samples from chronically infected children (38–40), 25 bnAbs (PG9, PG16, PGT145, PGDM1400, CAP256.25, CH01, 10-1074, BG18, AIIMS-P01, PGT121, PGT128, PGT135, VRC01, VRC03, VRC07-523LS, N6, 3BNC117, NIH45-46 G54W, PG05, PGT151, 35O22, N123.VRC34.01, 10E8, 4E10, and 2F5), and 6 non-nAbs (447-52D, 19b, 17b, A32, 48d, and b6). Briefly, envelope pseudoviruses were incubated in the presence of serially diluted heat-inactivated plasma samples, bnAbs, or non-nAbs for 1 h. After incubation, freshly trypsinized TZM-bl cells were added with 25 μg/ml DEAE-dextran. The plates were incubated for 48 h at 37°C, cells were lysed in the presence of Bright-Glo reagent, and luminescence was measured. Using the luminescence of serially diluted bnAbs or plasma, a nonlinear regression curve was generated, and titers were calculated as the bnAb concentration or reciprocal dilution of serum that showed a 50% reduction in luminescence compared to the untreated virus control. For epitope mapping, 25710_2_43, 16055_2_3, CAP45_G3, and BG505_W6M_C2 N160K mutants and pseudoviruses grown in the presence of kifunensine and swainsonine were used, and >3-fold reductions in ID₅₀ titers were classified as dependence.

HIV-1 strain tier designation.

Tier designation was assigned based on the neutralization susceptibility of AIIMS731 viral strains against non-nAbs targeting the V3 loop and CD4-induced epitopes. Tier 1 viruses comprise easy-to-neutralize viruses that adopt an open trimeric conformation, while tier 2/3 viruses comprise viruses that are not easily neutralized by non-nAbs. Neutralization assays with non-nAbs were performed at 50 μg/ml (compared to 10 μg/ml for bnAbs), and AIIMS731 viral strains that showed predominant neutralization by these non-nAbs were designated tier 1 viruses, while those that were not neutralized were designated tier 2 viruses.

HIV-1 Env stability-of-function assay.

Thermostability (T90) assays and spontaneous Env decay assays were performed as described previously (41, 42). Briefly, for T90 assays, AIIMS731 pseudotyped viral variants were incubated at temperatures from 37°C to 50°C for 60 min using a temperature gradient on a PCR thermal cycler (Bio-Rad). For spontaneous Env decay, AIIMS731 viral variants were incubated at 37°C for various durations (0, 1, 2, 4, 8, 16, 24, and 48 h). Pseudoviruses were then aliquoted in a 96-well culture plate, followed by the addition of 10,000 TZM-bl cells per well. Infectivity was determined by performing titration curves and plotted as a function of temperature or time. T90 values were interpolated as the temperature at which virus infectivity decreased by 90%. For thermostability, infectivity at 37°C was taken as 100% infectivity. For spontaneous decay, infectivity at 0 h was taken as 100% infectivity.

HIV-1 Env cell-to-cell fusion assay.

HIV-1 Env-mediated cell-to-cell fusion assays were performed as described previously (50, 51). Briefly, 293T cells were cotransfected with pTAT (pcTat.BL43.CC, catalog number 11785; NIH AIDS Reagent Program) and an envelope plasmid encoding either AIIMS731 viral variants or MW965.26 Env. To remove intra- and interassay variability, all technical replicates were cotransfected and expressed on the same day. 293T cells transfected with pTAT only were used as negative controls. At 24 h posttransfection, 10,000 pTAT/Env- or pTAT-transfected 293T cells were mixed with 10,000 TZM-bl cells (1:1 ratio) in 96-well culture plates and incubated for 6 h. Luciferase activity was measured using Bright-Glo reagent, and fusion activity (cell-to-cell transmission) for AIIMS731 pseudotyped viral variants was normalized to MW965.26 Env-mediated fusion activity.

Structural modeling and analysis.

The crystal structure of the ligand-free BG505.SOSIP.664 HIV-1 Env trimer (PDB accession number 4ZMJ) was used to assess interprotomeric interactions due to L184 and F184. Mutations were modeled using the Rotamers tool with the Dunbrack 2010 rotamer library (52). For modeled trimers based upon AIIMS731 Env sequences, crystal structures of the HIV-1 envelope trimer (PDB accession numbers 6P65, 6PWU, 6MZJ, and 6B0N) were used as the templates to generate homology models based on the 73105b amino acid sequence. Homology modeling was carried out using the Modeller 9.22 interface (53) with the UCSF Chimera package (54). High-mannose (Man-9) glycans were added to the modeled trimer using the glycoprotein builder interface available at the Glycam Web server (http://glycam.org/). To limit computational complexity, Man-9 glycans were selected, as computational complexity increases exponentially with complex and/or oligomannose glycans, which have multiple branching topologies.

Statistical analysis.

Two-tailed Student’s t test for paired analysis and a Mann-Whitney U test for unpaired analysis were used. For assessing relative infectivity, areas under the curve were calculated. All statistical analyses were performed using GraphPad Prism 8. A P value of <0.05 was considered significant.

Data availability.

The SGA-amplified HIV-1 envelope sequences used for inference of phylogeny and highlighter plots are available in the GenBank database under accession numbers MT366192 to MT366197. All data required to state the conclusions in the paper are present in the paper. Additional information related to the paper, if required, can be requested from the authors.