Tyrosine Sulfation in the Second Variable Loop (V2) of HIV-1 Gp120 Stabilizes V2–V3 Interaction and Modulates Neutralization Sensitivity

Home , Sulfation, Tyrosine sulfation

Tyrosine sulfation in the second variable loop (V2) of HIV-1 gp120 stabilizes V2–V3 interaction and modulates neutralization sensitivity

Raffaello Cimbroa,1, Thomas R. Gallanta, Michael A. Dolanb, Christina Guzzoa, Peng Zhanga, Yin Lina, Huiyi Miaoa, Donald Van Ryka, James Arthosa, Inna Gorshkovac, Patrick H. Brownc, Darrell E. Hurtb, and Paolo Lussoa,2

aLaboratory of Immunoregulation and bBioinformatics and Computational Biosciences Branch, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892; and cBiomedical Engineering and Physical Science Shared Resource, National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health, Bethesda, MD 20892

Edited by Dennis R. Burton, The Scripps Research Institute, La Jolla, CA, and Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, and accepted by the Editorial Board January 13, 2014 (received for review August 5, 2013) Elicitation of broadly neutralizing antibodies is essential for the receptor- and coreceptor-binding sites is believed to be a primary development of a protective vaccine against HIV-1. However, the mechanism of immune evasion by HIV-1 (4). native HIV-1 envelope adopts a protected conformation that con- The inherent conformational flexibility of gp120, along with ceals highly conserved sites of vulnerability from antibody recognition. the extensive N-linked glycosylation that covers most of the ex- Although high-definition structures of the monomeric core of the en- posed surface of the glycoprotein, has severely hampered attempts velope glycoprotein subunit gp120 and, more recently, of a stabilized to elucidate the native structure of the HIV-1 envelope spike. As soluble gp140 trimer have been solved, fundamental aspects re- a consequence, most of the available high-definition structures of lated to the conformation and function of the native envelope gp120 have been obtained with deglycosylated, variable loop- remain unresolved. Here, we show that the conserved central re- truncated core monomers in complex with stabilizing ligands such gion of the second variable loop (V2) of gp120 contains sulfated as soluble CD4 (sCD4) (6–10). Important information regarding tyrosines (Tys173 and Tys177) that in the CD4-unbound prefusion the overall conformation and ligand interactions of the trimeric

state mediate intramolecular interaction between V2 and the con- spike at intermediate resolution has emerged from the use of in- MICROBIOLOGY served base of the third variable loop (V3), functionally mimicking creasingly refined cryo-electron microscopy (cryo-EM) technologies sulfated tyrosines in CCR5 and anti–coreceptor-binding-site anti- (11–17). Moreover, the crystal structure of a stabilized, soluble, bodies such as 412d. Recombinant gp120 expressed in continuous cleaved gp140 trimer (BG505 SOSIP.664) at 4.7-Å resolution was cell lines displays low constitutive levels of V2 tyrosine sulfation, reported recently (18). However, despite these advances, many which can be enhanced markedly by overexpression of the tyrosyl critical aspects related to the structural mechanisms of HIV-1 im- sulfotransferase TPST2. In contrast, virion-associated gp120 pro- mune vulnerability and evasion remain unresolved. In particular, + duced by primary CD4 T cells is inherently highly sulfated. Con- the fine molecular details of the interaction between the second and sistent with a functional role of the V2 sulfotyrosines, enhancement third variable loops (V2 and V3, respectively) of gp120, which are of tyrosine sulfation decreased binding and neutralization of HIV-1 believed to play a critical role in stabilizing the prefusion envelope BaL by monomeric soluble CD4, 412d, and anti-V3 antibodies and structure (19–21), are elucidated only partially. Functionally, V2 increased recognition by the trimer-preferring antibodies PG9, PG16, and V3 cooperate in the formation of quaternary epitopes targeted CH01, and PGT145. Conversely, inhibition of tyrosine sulfation increased sensitivity to soluble CD4, 412d, and anti-V3 antibodies Significance and diminished recognition by trimer-preferring antibodies. These results identify the sulfotyrosine-mediated V2–V3 interaction as a Despite intensive efforts, the structure of the native HIV-1 critical constraint that stabilizes the native HIV-1 envelope trimer envelope trimer—the sole relevant target for vaccine design— and modulates its sensitivity to neutralization. has remained elusive. Our work identifies a key structural constraint that stabilizes the native envelope conformation and he development of a protective vaccine remains a high pri- modulates its sensitivity to neutralization. We show that this Tority for the global control of the HIV/AIDS epidemic (1). constraint is established by previously unrecognized sulfated However, the unique biological features of HIV-1 make this task tyrosines within the second variable loop (V2) of the envelope extremely challenging. The main obstacles include the ability of glycoprotein subunit gp120, which mediate intramolecular the virus to integrate into the host chromosomes, a remarkable interaction with the base of the third variable loop, V3. Strik- degree of genetic variability, and the cryptic, antibody-shielded ingly, the V2 sulfotyrosines functionally mimic those present in conformation adopted by the viral envelope in the native spikes the N terminus of the CCR5 coreceptor, which bind to the same that protrude from the virion surface (2). These spikes are com- V3 region. Our results shed light on the mechanisms adopted posed of homotrimers of heterodimers of the envelope gly- by HIV-1 to elude immunologic control and open new per- coprotein subunits gp120 and gp41 maintained in an energetically spectives for vaccine design. unfavorable, metastable conformation (3, 4). Upon binding to CD4 and a coreceptor such as CCR5 or CXCR4, gp120 undergoes Author contributions: R.C. and P.L. designed research; R.C., T.R.G., M.A.D., C.G., P.Z., Y.L., H.M., D.V.R., I.G., P.H.B., D.E.H., and P.L. performed research; R.C., M.A.D., J.A., D.E.H., and dramatic conformational changes that lead to a low-energy state, P.L. analyzed data; and R.C. and P.L. wrote the paper. creating permissive conditions for activation of the gp41 fusogenic The authors declare no conflict of interest. mechanism (3). In the prefusion conformation, gp120 effectively This article is a PNAS Direct Submission. D.R.B. is a guest editor invited by the Editorial conceals its highly conserved receptor- and coreceptor-binding Board. sites from antibody recognition, imposing a high-entropy penalty 1Present address: Division of Rheumatology, The Johns Hopkins School of Medicine, Baltimore, for interaction with CD4 or antibodies to the coreceptor-binding MD 21224. site such as 17b; in contrast, in the open, low-energy conformation, 2To whom correspondence should be addressed. E-mail: [email protected]. gp120 interacts with CD4 and 17b with minimal thermodynamic This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. changes (4, 5). This conformational masking of the vulnerable 1073/pnas.1314718111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1314718111 PNAS Early Edition | 1of6 Downloaded by guest on October 2, 2021 by some of the most potent and broadly neutralizing mAbs hitherto identified (22, 23), and V2 effectively masks neutralization epitopes AB in V3 (24–27). Cryo-EM studies have provided evidence that in the prefusion conformation V2 and V3 are spatially contiguous and account for most of the density at the apex of the trimeric envelope spike (12–17). Although various fragments of V2 and V3 were crystallized separately using antibody-complexed synthetic peptides (28, 29), scaffolded chimeric constructs of the first and second variable loops (V1V2) (30, 31), or a V3-containing gp120 core monomer (8), the only study in which the two loops were visualized simultaneously is the recent report of the BG505 SOSIP.664 trimer crystal structure (18). In this artificially stabilized trimer, which CD displays several antigenic features of the native envelope (32), V2 and V3 appear to interact directly at the trimer apex with the V3 β-hairpin extensively buried under the V1V2 four-stranded Greek- key β-sheet (18). In the present study, we provide evidence that the conserved central region of the gp120 V2 loop contains previously unrecognized sulfated tyrosines that, in the CD4-unbound pre- Fig. 1. The V2 domain of HIV-1 gp120 contains sulfated tyrosines. (A)Se- fusion state, mediate intramolecular interaction between V2 and quence alignment of the CCR5 N-terminal domain and the conserved central the CCR5-binding site at the base of V3. Our results suggest that region of the V2 domain of HIV-1 gp120 (consensus sequence for subtype B). the sulfotyrosine-bolstered interaction between V2 and V3 is Two colinear conserved tyrosine residues present in CCR5 and gp120 V2 are a key structural constraint that stabilizes the native conformation highlighted in green. (B) Detection of sulfated tyrosines by Western blot in gp120 purified from HeLa cells expressing WT HIV-1 BaL gp160 or a partially of the HIV-1 envelope trimer. V2-deleted mutant (Δ164–190) lacking Tyr173 and Tyr177 by vaccinia technology. gp120 was immunoprecipitated from the cell surface and analyzed Results by Western blot using a specific mAb for sulfated tyrosines (1C-A2); a murine The Conserved Central Region of the V2 Loop of HIV-1 gp120 Contains anti-gp120 mAb (b13) was tested in parallel as a loading control (gp120). (C) Sulfated Tyrosines. Despite their definition as variable loops, the Detection of sulfated tyrosines by autoradiography in metabolically labeled V2 and V3 regions of gp120 contain highly conserved domains HEK293 cells expressing HIV-1 BaL WT or Δ164–190 gp160 by vaccinia (33). Because several lines of evidence suggest that V3 estab- technology. The cells were labeled with free [35S]sulfate or [35S]cysteine/[35S] – methionine by overnight culture in sulfate-free or cysteine/methionine-free lishes direct interaction with V2 (12 21), and the conserved base 35 of V3 is the binding site for the N-terminal region of the CCR5 medium, respectively. The films were exposed for 48 h for [ S]sulfate labeling and for 18 h for [35S]cysteine/[35S]methionine labeling. (D) Detection coreceptor (34), we looked for potential structural homology bet- of sulfated tyrosines by Western blot in HEK293 cells transfected with ween V2 and CCR5. We recognized that the conserved central a plasmid encoding WT HIV-1 BaL gp160 or the phenylalanine-substituted region of V2 contains two tyrosine residues (Tyr173 and Tyr177) mutants BaL Y173F, BaL Y177F, and BaL Y173F/Y177F. Western blot analysis with spacing identical to that of two tyrosines (Tyr10 and Tyr14) was performed as in B. present in the N-terminal region of CCR5 (Fig. 1A). Tyr177 is highly conserved across all HIV-1 subtypes (>99%), whereas Tyr173 is conserved in subtypes A, B, and C but often is replaced All three mutants showed a loss of Tys signal by Western blot by histidine in subtypes D, E, and F (∼66% cross-subtype con- (Fig. 1D), confirming that the central V2 tyrosines selectively servation). Because Tyr10 and Tyr14 in CCR5 were shown to be account for gp120 sulfation. The absence of signal in both posttranslationally modified by O-sulfation (35) and to play a crit- positions with individual phenylalanine mutants is in line with ical role in the interaction of CCR5 with V3 (36), we investigated observations made with single phenylalanine mutants of the whether the V2 tyrosines also can be sulfated. Western blot anal- CCR5 N terminus (35), suggesting that tyrosine sulfation is ysis using a highly specific anti-sulfotyrosine (Tys) mAb (Fig. S1) hampered by the presence of neighboring phenylalanines. documented the presence of sulfated tyrosines in gp120 immunoprecipitated from the surface of HeLa cells expressing full-length Overexpression of Tyrosyl Protein Sulfotransferase 2 Enhances the cleavable gp160 from the subtype-B HIV-1 isolate BaL (Fig. 1B). Low Constitutive Levels of gp120 Tyrosine Sulfation in Continuous Because gp120 contains additional tyrosine residues both within Cell Lines. Because continuous cell lines are generally inefficient and outside V2, we produced a partial V2-deletion mutant that in posttranslationally modifying tyrosines by O-sulfation (37), we selectively excludes Tyr173 and Tyr177 (BaL Δ164–190). Despite evaluated by Western blot the con-stitutive levels of gp120 ty- comparable levels of gp120 expression, the Tys signal was abro- rosine sulfation in CHO cells and the effects of overexpression of gated in the deletion mutant (Fig. 1B), indicating that gp120 sul- tyrosyl protein sulfotransferase 2 (TPST2) or treatment with the fation is specific and occurs selectively at Tyr173 and Tyr177. sulfotransferase inhibitor sodium chlorate (NaClO3) (35). To To confirm the presence of sulfated tyrosines in gp120 and their visualize low levels of tyrosine sulfation, high amounts of gp120 selective localization in the V2 loop using a different methodol- were loaded (fivefold dilutions starting at 1,250 ng per lane). Fig. ogy, we performed metabolic labeling with free [35S]sulfate in cells 2A shows that the constitutive levels of tyrosine sulfation in expressing HIV-1 BaL WT or Δ164–190. Gp120 was immuno- CHO-expressed recombinant gp120 (BaL) were low, but they precipitated from metabolically labeled cells, deglycosylated to could be enhanced markedly by overexpression of TPST2, 35 exclude [ S]sulfate incorporated into glycans, and analyzed by whereas treatment with 30 mM NaClO3 totally abrogated the autoradiography. Although both WT and Δ164–190 gp120 were signal (Fig. 2A), even though it had no adverse effects on cell labeled with comparable amounts of [35S]cysteine and [35S]me- viability (Fig. S2). thionine, only WT gp120 incorporated free [35S]sulfate (Fig. 1C), confirming the presence of sulfated tyrosines in gp120 and their Virion-Associated gp120 Produced by Infected Primary CD4+ T Cells Is specific localization within the central region of V2. Highly Sulfated. Next, we tested the levels of V2 tyrosine sulfation The identity of the gp120 sulfotyrosines was investigated fur- in gp120 purified from HIV-1 virions produced by primary hu- + ther by mutating Tyr173 and Tyr177 to phenylalanine either man CD4 T lymphocytes, which are physiologically relevant individually (Y173F, Y177F) or in combination (Y173F/177F). target cells for HIV-1 infection. Relative quantification of sulfation

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1314718111 Cimbro et al. Downloaded by guest on October 2, 2021 (R-gp120). These results indicate that a major fraction of virion- A NaClO Untreated TPST2 + 3 associated gp120 produced by primary CD4 T cells contains ng: 50 250 1250 50 250 1250 50 250 1250 Sulfated sulfated tyrosines. tyrosines To confirm the presence and efficiency of tyrosine sulfation in ng: 20 20 20 40 80 40 80 40 80 different HIV-1 envelopes, we tested gp120 purified from whole + gp120 virions of six unrelated viral isolates produced in primary CD4 T cells, including both laboratory-expanded and primary isolates displaying different coreceptor-use phenotypes: three isolates were CCR5-tropic (BaL, JR-FL, ADA); two were dual-tropic mAb 412d V-gp120 R-gp120 B (92US077, 92HT599); and one was CXCR4-tropic (IIIB). Fig. ng: 10 20 40 10 20 40 10 20 40 2C shows that sulfotyrosine signals were detected in all cases Sulfated 160 kD with 10 ng of gp120, indicating a high degree of sulfation effi- tyrosines 120 kD ciency. Of note, 92HT599, a rare isolate that lacks the highly ng: 10 20 40 10 20 40 10 20 40 conserved Tyr177, displayed a reduced sulfotyrosine signal. Taken HuIgG together, these data demonstrate that in different HIV-1 isolates gp120 grown in physiologically relevant target cells the V2 tyrosines are sulfated efficiently, in sharp contrast with the inefficiency documented in continuous cell lines. C A Tyrosine-Sulfated V2 Peptide Mimetic Interacts with the CCR5- Binding Site at the Base of V3.

ADA ADA 92HT599 IIIB Having established that the V2 JR-FL JR-FL BaL 92US077 Sulfated loop of gp120 contains sulfated tyrosines with spacing identical tyrosines ng: 10 10 10 10 10 10 to that in the CCR5 N terminus, we hypothesized that these modified tyrosines could functionally mimic the CCR5 sulfotyr- gp120 osines and mediate intramolecular interaction of V2 with the Fig. 2. Efficiency of V2 tyrosine sulfation in recombinant gp120 produced in base of V3 in the CD4-unbound, prefusion gp120 state. This continuous cell lines and virion-associated gp120 produced in primary CD4+ interaction is compatible with the recently published crystal

T cells. (A) Detection of sulfated tyrosines in recombinant gp120 produced in structure of a soluble SOSIP gp140 trimer (18), in which the 173– MICROBIOLOGY CHO cells. The effects of overexpression of the sulfotransferase TPST2 or 177 segment is directly juxtaposed to the CCR5-binding region at treatment with the sulfotransferase inhibitor NaClO3 (30 mM) were tested in the base of V3, even though sulfotyrosines were not detected in parallel. Gp120 was expressed by transfection, purified from the cell culture this structure (most likely because the trimer was produced in supernatants by lectin affinity, and quantified. Identical amounts of serially HEK293 cells). Because the sulfated human mAb 412d interacts diluted gp120s were loaded onto the gel, as indicated over each lane. The presence of sulfated tyrosines was evaluated by Western blot with the with the CCR5-binding site in V3 (34), we tested the ability of specific mAb 1C-A2; loading controls were revealed with a murine anti- a tyrosine-sulfated 18-amino acid peptide mimetic derived from gp120 mAb (b24). To visualize low levels of tyrosine sulfation, high amounts the central region of V2 (pV2α-Tys; amino acids 168–185, of gp120 were used (fivefold dilutions starting from 1,250 ng per lane). (B) bearing sulfations on both Tyr173 and Tyr177) to compete with Direct side-by-side comparison of tyrosine sulfation levels in a reference 412d binding to gp120 (BaL) by surface plasmon resonance. sulfated mAb (412d) versus virion-associated gp120 (BaL) produced by + Because the 412d-binding site becomes accessible only after infected primary human CD4 T cells (V-gp120) and recombinant gp120 (BaL) binding to CD4 (37), gp120 was precomplexed with sCD4. Fig. produced in HEK293 cells (R-gp120). Identical amounts of serially diluted 3A shows that peptide pV2α-Tys potently inhibited binding of mAb 412d and purified gp120s were loaded onto the gels under non- surface-bound mAb 412d to CD4-activated gp120, whereas its denaturing conditions, as indicated over each lane. For V-gp120, the protein α was immunoprecipitated directly from the infectious viral stock (BaL), unsulfated counterpart (pV2 ) had a limited effect. No binding quantified, and analyzed by Western blot as described in A. The loading was detected in the absence of sCD4. As an additional control, we control for mAb 412d was revealed using a murine mAb specific for human tested a tyrosine-sulfated peptide derived from the CCR5 N ter- IgG. (C) Detection of sulfated tyrosines in virion-associated gp120 purified minus, pCCR5-Tys (amino acids 1–22, bearing sulfations on both + from different HIV-1 isolates grown in primary human CD4 T cells. Infectious Tyr10 and Tyr14), which, in line with previous observations, viral stocks from six primary and laboratory isolates with different coreceptor- markedly reduced gp120-sCD4 binding to 412d (37), whereas the use phenotypes were used: Three were CCR5-tropic (BaL, JR-FL, ADA), two corresponding unsulfated peptide (pCCR5) had no effect. Speci- were dual-tropic (92US077, 92HT599), and one was CXCR4-tropic (IIIB). ficity was further confirmed by the lack of inhibition of gp120 The sequences of the tyrosine-sulfated segments of V2 (amino acids 170– A 181) from the tested isolates are as follows: JR-FL: QKEYALFYKLDV; BaL: binding to a control mAb, F105 (Fig. 3 ). QKEYALFYELDI; 92US077: QKEDAFFYKSDV; ADA: KKDYALFYRLDV; 92HT599: The interaction of the V2 peptide mimetic with the 412d-binding QKEYALFSKLDV; IIIB: QKEYAFFYKLDI. Although BaL contains an unusual acidic site was confirmed using a virion-capture assay in which the HIV-1 residue (Glu) at position 178, which may favor tyrosine sulfation at position 177, envelope is displayed on the surface of intact virions. To reveal the all the other isolates contained a basic residue (Lys or Arg) at position 178. The 412d-binding site, the viral stock (BaL) was treated with sCD4 gp120 proteins were immunoprecipitated directly from the viral stocks, quanti- before capture by antibody-coated magnetic beads. Fig. 3B shows fied, and analyzed by Western blot as in A. that peptide pV2α-Tys dramatically reduced virion capture by 412d, whereas the unsulfated peptide (pV2α) had a limited effect; likewise, 412d-mediated virion capture was significantly reduced by levels was obtained by direct side-by-side comparison of purified the sulfated CCR5 peptide but not by its unsulfated counterpart. gp120 with a reference sulfated human mAb, 412d, which contains As an additional proof of specificity, none the peptides affected two sulfotyrosines in its CDRH3 domain (37). When identical virion capture by a control mAb, 2G12 (Fig. 3B). Taken together, amounts of serially diluted virion-associated gp120 from HIV-1 these results support the concept that in unliganded gp120 the V2 BaL (V-gp120) and mAb 412d were loaded onto the same gel, sulfotyrosines interact with the CCR5- and 412d-binding site at the both proteins displayed high levels of tyrosine sulfation, as in- base of V3. dicated by the presence of visible bands at protein concentrations as low as 10 ng per lane, which is close to the lower detection limit Tyrosine Sulfation in the V2 Loop Specifically Modulates gp120 Epitope of the method (Fig. 2B). In contrast, no signal was detected with Exposure and Neutralization Sensitivity. To investigate the functional recombinant homologous gp120 (BaL) produced in HEK293 cells role of the V2 sulfotyrosines, we tested the effects of the modulation

Cimbro et al. PNAS Early Edition | 3of6 Downloaded by guest on October 2, 2021 A of HIV-1 BaL to neutralization by sCD4, 412d, and anti-V3 antibodies but at the same time made the virus more sensitive to neutralization by PG9, PG16, CH01, and PDT145. On the other hand, inhibition of V2 tyrosine sulfation by NaClO3 treatment increased neutralization by sCD4, 412d, and anti-V3 antibodies and decreased sensitivity to trimer-preferring antibodies. Overexpression of TPST2 had only a slight inhibitory effect on neutralization by 2G12, b12, and VRC01, in accordance with the ability of these antibodies to bind with high affinity to the native trimeric spike. Neutralization by the control anti-gp41 antibody 2F5 was totally unaffected. Con- B sistent with a diminished interaction with CD4, the efficiency of the fusion reaction was reduced slightly in TPST2-expressing cells compared with untreated and NaClO3-treated cells (see the legend for Fig. 4B). As a further proof of specificity, NaClO3 treatment did not affect the neutralization sensitivity of the gp160 Δ164–190 mutant (Fig. S4), which lacks sulfated tyrosines (Fig. 1 B and C). Al- though these results await confirmation on a larger panel of HIV-1 isolates of different genetic subtypes, they suggest that the sulfotyrosine-mediated V2–V3 interaction plays a critical role in stabi- Fig. 3. Competition of a tyrosine-sulfated V2 mimetic peptide with mAb lizing the trimeric envelope conformation and specifically modulates 412d binding to gp120. (A) Effect of a tyrosine-sulfated V2-loop mimetic the sensitivity of HIV-1 to antibody-mediated neutralization. peptide (pV2α-Tys; amino acids 168–185) on mAb 412d binding to gp120 as assessed by surface plasmon resonance. A tyrosine-sulfated peptide derived Discussion from the CCR5 N terminus (pCCR5-Tys; amino acids 1–22) and unsulfated peptides from V2 (pV2α) and CCR5 (pCCR5) were tested in parallel as con- In this study, we provide evidence for the presence of sulfated trols. MAb 412d was immobilized on the sensor surface and tested for tyrosines within the V2 loop of the external HIV-1 envelope binding to gp120 (BaL) pretreated with two-domain sCD4 in the presence or glycoprotein subunit gp120 and show that these modified tyro- absence of the peptides used at 166 μM. A control antibody (F105; no sCD4 sines play an important role in the functional structuring of the pretreatment) was tested in parallel as a control. The data are from a rep- native envelope by stabilizing the intramolecular interaction resentative experiment of three that were performed with similar results. (B) between V2 and V3. Tyrosine sulfation is a posttranslational Effect of tyrosine-sulfated and unsulfated V2-loop and CCR5 N terminus modification that is estimated to occur in ∼7% of mammalian mimetic peptides on HIV-1 virion capture by mAb 412d. Infectious viral proteins and is increasingly recognized as an important modu- stocks from HIV-1 BaL were pretreated with sCD4 (5 μg/mL) in the presence – or absence of the indicated peptides (each at 100 μM) and then were mixed lator of protein protein interaction (43). Tyrosine sulfation of with immunomagnetic beads prearmed with mAb 412d. A control antibody functional relevance was identified previously in several host (2G12; no sCD4 pretreatment) was tested in parallel as a further specificity proteins that bind to gp120, including coreceptors (35), antibodies control. The data presented are mean values (± SD) from three independent to the coreceptor-binding site such as 412d (37), and, more re- experiments. Asterisks denote significant differences from the virion capture cently, antibodies to glycan-dependent quaternary epitopes such in the peptide-untreated control (P < 0.01 by unpaired Student t test). as PG9 and PG16 (44). Unlike these previous examples, however, our data identify tyrosine sulfation within a viral protein, as reported for varicella-zoster virus envelope glycoproteins (45). Fur- of tyrosine sulfation on gp120 epitope accessibility and neutrali- thermore, it is remarkable that HIV-1 appears to have modeled the zation using a panel of antibodies to the major neutralization tyrosine-sulfated region of the V2 loop after the N-terminal domain regions of gp120 and monomeric sCD4. The experiments were of the CCR5 coreceptor, which interacts with the same conserved performed on full-length, cleavable gp160 derived from two un- region at the base of V3. Our results shed light on the mechanisms related HIV-1 isolates, BaL and YU2, expressed on the surface of whereby V2 occludes the coreceptor-binding site in the prefusion HeLa cells. A striking dichotomous effect was observed on the envelope conformation, preventing access to 412d and other sul- antigenic profile of gp120 from both isolates (Fig. 4A and Fig. S3). fated antibodies. Upon binding to CD4, however, V2 unclamps Enhancement of tyrosine sulfation by TPST2 overexpression dra- from V3, unraveling the coreceptor-binding site and allowing the matically reduced binding of monomeric sCD4 (receptor-binding sulfated coreceptor N terminus to replace V2 at the base of V3. site), mAb 412d (coreceptor-binding site) (37), and anti-V3 loop These events promote subsequent conformational changes that mAbs B4e8 (38) and D19 (39), all ligands with a restricted binding eventually lead to the fusion event. capacity to the native trimeric spike. In contrast, TPST2 over- The direct V2–V3 interaction described herein is supported expression markedly increased recognition by the trimer-preferring by biological and structural evidence. Functional and antigenic antibodies PG9, PG16, CH01, and PGT145, which are directed to interactions between V2 and V3 have long been recognized (21– quaternary, glycan-dependent V2 epitopes that are stabilized on the 27), and spatial proximity between these two loops has been native trimer (22, 23, 40). Opposite effects were observed when documented by both cryo-EM studies (12–17) and the recently tyrosine sulfation was inhibited by NaClO3 treatment, with a solved crystal structure of a stabilized soluble SOSIP trimer at marked increase of sCD4, 412d, and anti-V3 antibody binding and 4.7-Å resolution (18). Despite the lack of sulfated tyrosines (pre- a reciprocal decrease in recognition by trimer-preferring antibodies. sumably because the soluble trimer was produced in HEK293 Binding of three broadly neutralizing antibodies which readily rec- cells), the latter structure is consistent with our model because it ognize the native trimer, b12 and VRC01 (CD4-binding site) (9, shows the 173–177 segment of V2 directly juxtaposed to the CCR5- 41), and 2G12 (outer domain) (42), was not affected significantly. binding region at the base of V3, with the most conserved of the Of note, neither TPST2 overexpression nor NaClO3 treatment had two sulfotyrosines, Tyr177, establishing van der Waals contacts with adverse effects on the levels of gp160 expression, as evaluated by Ile420, Leu154, and Leu175, as well as an H-bond with Asn302, reactivity with a control antibody, 2F5, directed against the gp41 all residues that are proximal to the predicted binding pocket membrane-proximal external region (Fig. 4A and Fig. S3). for the homologous sulfotyrosine (Tys14) of CCR5 (34). It is The pattern of epitope exposure observed by flow cytometry likely that the addition of sulfate groups would markedly affect was mirrored precisely by the results of neutralization studies. As the local conformations and interactions, including the establish- illustrated in Fig. 4B, TPST2 overexpression increased the resistance ment of additional H-bonds between Tys177 and Arg298, Asn300,

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1314718111 Cimbro et al. Downloaded by guest on October 2, 2021 and the backbone of Arg327, resulting in a significant increase in the total binding energy. Thus, the absence of sulfated tyrosines A may be one of the reasons why the SOSIP.664 trimer, albeit re- portedly in a “near-native” conformation (32), maintains some phenotypic properties of the open gp120 structure, such as reactivity with anti-V3 mAbs that do not neutralize the homologous virus, intermediate-affinity binding to monomeric sCD4, and an unfavorable ratio between the concentrations of trimer-preferring antibodies required for binding versus neutralization of the homologous virus (32). The divergent effects on gp120 epitope exposure and neutralization that we observed upon modulation of sulfation attest to the biological relevance of tyrosine sulfation in the V2 loop. Although the present results were obtained with only two HIV-1 envelopes and need to be validated on a larger panel of genotypically and B phenotypically diverse isolates, they suggest that the sulfotyrosine- bolstered V2–V3 interaction is a critical mechanism whereby HIV- 1 constrains the envelope trimer in its metastable native conformation. The native trimer was shown to shield specific vulnerable sites—most notably the receptor- and coreceptor-binding sites (1– 4)—effectively from antibody recognition, but at the same time it stabilizes quaternary epitopes recognized by a unique group of trimer-preferring antibodies, such as PG9, CH01, and PGT145, that make direct contact with high-mannose glycan moieties at the trimer apex (22, 23, 40). Thus, HIV-1 must cope with an immunologic oxymoron by which protection of key vulnerable sites is associated with exposure of other vulnerable sites. However, the

generation of trimer-preferring antibodies is a challenging process MICROBIOLOGY for the host immune system that requires selection for long CDRH3 domains and extensive somatic hypermutation (2, 40). Several considerations point to Tys177, analogous to Tys14 in CCR5, as the most important V2 sulfotyrosine for stabilizing the interaction with V3, whereas Tys173, corresponding to Tys10 in CCR5, may play a less critical role. In agreement with this model, Tys177 is extremely conserved across all HIV-1 subtypes, whereas Tys173 shows greater variability, being conserved in subtypes A, B, and C but often replaced by histidine in subtypes D, E, and F. Although we cannot exclude that different viral genotypes may have devised slightly different solutions to stabilize the V2–V3 interaction, genetic and phenotypic diversity is a true hallmark of HIV-1, as reflected by a broad range of neutralization sensitivities Fig. 4. Effect of V2 tyrosine sulfation on gp120 (BaL) epitope exposure and among primary isolates. Thus, variation at position 173 or neigh- neutralization. (A) Modulation of V2 tyrosine sulfation alters gp120 (BaL) boring amino acids that may influence the efficiency of V2 tyrosine epitope accessibility. Flow cytometry was performed on HeLa cells expressing sulfation (43) might be exploited by different HIV-1 isolates to fine- full-length, cleavable HIV-1 BaL gp160 on their surface. V2 tyrosine sulfation tune the tightness of their antibody shield, balancing the need to was enhanced by overexpression of the sulfotransferase TPST2 or inhibited maintain a high replication fitness and the need to elude immu- by treatment with the sulfotransferase inhibitor NaClO3 (30 mM). A panel of human mAbs directed to the indicated domains of gp120 and gp41 or four- nologic recognition of the receptor- and coreceptor-binding sites. domain soluble CD4 (sCD4) (each at 5 μg/mL) was used. BS, binding site; CoR, As seen with other posttranslational modifications such as N- coreceptor; OD, outer domain. The sequence of the tyrosine-sulfated seg- linked glycosylation, tyrosine sulfation varies remarkably in dif- ment of V2 (amino acids 170–181) of HIV-1 BaL is QKEYALFYELDI. The data ferent cell lineages. Indeed, we found that gp120 sulfation in represent mean fluorescence intensity values after subtraction of back- certain continuous cell lines is inefficient, presumably because of ground fluorescence levels. The signals obtained with the reference anti- limited endogenous expression of tyrosyl sulfotransferases. This body 2F5 were used to verify that cell-surface envelope expression levels observation has important practical implications, because cell lines were comparable in the three cultures. The data presented are from a rep- such as CHO and HEK293 are widely used for the production of resentative experiment of three that were performed with similar results. (B) recombinant gp120 for structural, biological, and clinical studies, Modulation of V2 tyrosine sulfation alters HIV-1 BaL neutralization sensi- – tivity. Neutralization experiments were performed using an HIV-1 envelope- including atomic-level structure determination (6 10, 18), stan- mediated fusion assay with the same panel of anti-gp120/gp41 antibodies dardized neutralization assays (46), and experimental vaccine trials and sCD4 as in A, which were tested at the indicated concentrations. The (47, 48). These considerations emphasize the need to use in vitro fusion assay was performed using vaccinia technology as described (39) with expression systems that implement physiologically relevant post- HeLa cells expressing full-length cleavable HIV-1 BaL gp160 as effector cells translational modifications. Importantly, we found that in primary + and NIH 3T3 cells expressing human CD4 and CCR5 as target cells. Tyrosine human CD4 T lymphocytes, which are the main physiological sulfation was enhanced and inhibited as described in A. The fusion values target cells for HIV-1 infection, gp120 from different viral isolates were normalized to the value obtained with untreated controls for each was sulfated efficiently, corroborating the biological relevance of culture condition. The mean absolute levels of envelope-mediated fusion for the experiment shown, after background subtraction, were the following: this modification. Finally, our findings provide a further rationale for consider- untreated, 1.136 optical density units; NaClO3-treated, 1.192 optical density units; TPST2 overexpression, 0.891 optical density units. The data presented ing the conserved central region of V2 as a potential target for are mean values (± SE) of duplicate wells from a representative experiment HIV-1 vaccine development, emphasizing tyrosine sulfation in of three that were performed with similar results. this region not only as a component of the immunogenic epitopes

Cimbro et al. PNAS Early Edition | 5of6 Downloaded by guest on October 2, 2021 but also as a key modulator of their 3D structure. Thus, the pres- given in SI Materials and Methods. This section describes methods for the ence and extent of V2 tyrosine sulfation may be relevant criteria for detection of sulfated tyrosines, modulation of tyrosine sulfation, peptide the evaluation of candidate vaccine immunogens, because a high rate synthesis, surface plasmon resonance, HIV-1 virion capture, flow cytometry, of gp120 sulfation may favor the adoption of a more stable prefusion neutralization of HIV-1 envelope-mediated fusion, and statistical analysis. conformation. Moreover, the stabilizing effect of V2 tyrosine sulfation on the HIV-1 envelope structure may be exploited to promote ACKNOWLEDGMENTS. We thank Edward A. Berger and Anthony S. Fauci for helpful suggestions and critical reading of the manuscript; John R. Mascola, crystallization of gp120 in both monomeric and trimeric forms. George K. Lewis, and James E. Robinson for anti-gp120 mAbs; Michael Farzan for the TPST2-expressing vector; Jocelyn Ray for help in producing recombi- Materials and Methods nant gp120; the IAVI consortium for mAbs PG9, PG16, CH01, and PGT145; and The presence and localization of sulfated tyrosines in HIV-1 gp120 were in- the National Institutes of Health (NIH) AIDS Reagent Program for cell lines vestigated by Western blot using a specific anti-sulfotyrosine antibody and by and reagents. This work was supported by the National Institute for Al- metabolic labeling. Details of materials and methods used in this work are lergy and Infectious Diseases, NIH, Intramural Research Program.

1. Burton DR, et al. (2012) A blueprint for HIV vaccine discovery. Cell Host Microbe 12(4): 26. Liu L, Cimbro R, Lusso P, Berger EA (2011) Intraprotomer masking of third variable 396–407. loop (V3) epitopes by the first and second variable loops (V1V2) within the native HIV- 2. Mascola JR, Haynes BF (2013) HIV-1 neutralizing antibodies: Understanding nature’s 1 envelope glycoprotein trimer. Proc Natl Acad Sci USA 108(50):20148–20153. pathways. Immunol Rev 254(1):225–244. 27. Rusert P, et al. (2011) Interaction of the gp120 V1V2 loop with a neighboring gp120 3. Wyatt R, Sodroski J (1998) The HIV-1 envelope glycoproteins: Fusogens, antigens, and unit shields the HIV envelope trimer against cross-neutralizing antibodies. J Exp Med – immunogens. Science 280(5371):1884 1888. 208(7):1419–1433. 4. Kwong PD, et al. (2002) HIV-1 evades antibody-mediated neutralization through 28. Ghiara JB, Stura EA, Stanfield RL, Profy AT, Wilson IA (1994) Crystal structure of the conformational masking of receptor-binding sites. Nature 420(6916):678–682. principal neutralization site of HIV-1. Science 264(5155):82–85. 5. Myszka DG, et al. (2000) Energetics of the HIV gp120-CD4 binding reaction. Proc Natl 29. Liao HX, et al. (2013) Vaccine induction of antibodies against a structurally hetero- Acad Sci USA 97(16):9026–9031. geneous site of immune pressure within HIV-1 envelope protein variable regions 1 6. Kwong PD, et al. (1998) Structure of an HIV gp120 envelope glycoprotein in complex and 2. Immunity 38(1):176–186. with the CD4 receptor and a neutralizing human antibody. Nature 393(6686): 30. McLellan JS, et al. (2011) Structure of HIV-1 gp120 V1/V2 domain with broadly neu- 648–659. – 7. Kwong PD, et al. (2000) Structures of HIV-1 gp120 envelope glycoproteins from lab- tralizing antibody PG9. Nature 480(7377):336 343. oratory-adapted and primary isolates. Structure 8(12):1329–1339. 31. Pancera M, et al. (2013) Structural basis for diverse N-glycan recognition by HIV-1- – 8. Huang CC, et al. (2005) Structure of a V3-containing HIV-1 gp120 core. Science neutralizing V1-V2-directed antibody PG16. Nat Struct Mol Biol 20(7):804 813. 310(5750):1025–1028. 32. Sanders RW, et al. (2013) A next-generation cleaved, soluble HIV-1 Env Trimer, BG505 9. Chen L, et al. (2009) Structural basis of immune evasion at the site of CD4 attachment SOSIP.664 gp140, expresses multiple epitopes for broadly neutralizing but not non- on HIV-1 gp120. Science 326(5956):1123–1127. neutralizing antibodies. PLoS Pathog 9(9):e1003618. 10. Pancera M, et al. (2010) Structure of HIV-1 gp120 with gp41-interactive region reveals 33. Zolla-Pazner S, Cardozo T (2010) Structure-function relationships of HIV-1 envelope layered envelope architecture and basis of conformational mobility. Proc Natl Acad sequence-variable regions refocus vaccine design. Nat Rev Immunol 10(7):527–535. Sci USA 107(3):1166–1171. 34. Huang CC, et al. (2007) Structures of the CCR5 N terminus and of a tyrosine-sulfated 11. Liu J, Bartesaghi A, Borgnia MJ, Sapiro G, Subramaniam S (2008) Molecular archi- antibody with HIV-1 gp120 and CD4. Science 317(5846):1930–1934. tecture of native HIV-1 gp120 trimers. Nature 455(7209):109–113. 35. Farzan M, et al. (1999) Tyrosine sulfation of the amino terminus of CCR5 facilitates 12. White TA, et al. (2010) Molecular architectures of trimeric SIV and HIV-1 envelope HIV-1 entry. Cell 96(5):667–676. glycoproteins on intact viruses: Strain-dependent variation in quaternary structure. 36. Farzan M, et al. (2000) A tyrosine-sulfated peptide based on the N terminus of CCR5 PLoS Pathog 6(12):e1001249. interacts with a CD4-enhanced epitope of the HIV-1 gp120 envelope glycoprotein and 13. Hu G, Liu J, Taylor KA, Roux KH (2011) Structural comparison of HIV-1 envelope spikes inhibits HIV-1 entry. J Biol Chem 275(43):33516–33521. – with and without the V1/V2 loop. J Virol 85(6):2741 2750. 37. Choe H, et al. (2003) Tyrosine sulfation of human antibodies contributes to recogni- 14. Harris A, et al. (2011) Trimeric HIV-1 glycoprotein gp140 immunogens and native HIV- tion of the CCR5 binding region of HIV-1 gp120. Cell 114(2):161–170. 1 envelope glycoproteins display the same closed and open quaternary molecular 38. Cavacini LA, et al. (2003) Conformational changes in env oligomer induced by an architectures. Proc Natl Acad Sci USA 108(28):11440–11445. antibody dependent on the V3 loop base. AIDS 17(5):685–689. 15. Julien JP, et al. (2013) Asymmetric recognition of the HIV-1 trimer by broadly neu- 39. Lusso P, et al. (2005) Cryptic nature of a conserved, CD4-inducible V3 loop neutrali- tralizing antibody PG9. Proc Natl Acad Sci USA 110(11):4351–4356. zation epitope in the native envelope glycoprotein oligomer of CCR5-restricted, but 16. Bartesaghi A, Merk A, Borgnia MJ, Milne JL, Subramaniam S (2013) Prefusion structure of trimeric HIV-1 envelope glycoprotein determined by cryo-electron microscopy. not CXCR4-using, primary human immunodeficiency virus type 1 strains. J Virol – Nat Struct Mol Biol 20(12):1352–1357. 79(11):6957 6968. 17. Lyumkis D, et al. (2013) Cryo-EM structure of a fully glycosylated soluble cleaved HIV-1 40. Bonsignori M, et al. (2011) Analysis of a clonal lineage of HIV-1 envelope V2/V3 envelope trimer. Science 342(6165):1484–1490. conformational epitope-specific broadly neutralizing antibodies and their inferred 18. Julien JP, et al. (2013) Crystal structure of a soluble cleaved HIV-1 envelope trimer. unmutated common ancestors. J Virol 85(19):9998–10009. Science 342(6165):1477–1483. 41. Zhou T, et al. (2010) Structural basis for broad and potent neutralization of HIV-1 by 19. Sullivan N, Thali M, Furman C, Ho DD, Sodroski J (1993) Effect of amino acid changes antibody VRC01. Science 329(5993):811–817. in the V1/V2 region of the human immunodeficiency virus type 1 gp120 glycoprotein 42. Trkola A, et al. (1996) Human monoclonal antibody 2G12 defines a distinctive neu- on subunit association, syncytium formation, and recognition by a neutralizing an- tralization epitope on the gp120 glycoprotein of human immunodeficiency virus type 1. tibody. J Virol 67(6):3674–3679. JVirol70(2):1100–1108. 20. Xiang SH, et al. (2010) A V3 loop-dependent gp120 element disrupted by CD4 binding 43. Moore KL (2003) The biology and enzymology of protein tyrosine O-sulfation. J Biol stabilizes the human immunodeficiency virus envelope glycoprotein trimer. J Virol Chem 278(27):24243–24246. – 84(7):3147 3161. 44. Pejchal R, et al. (2010) Structure and function of broadly reactive antibody PG16 re- 21. Kwon YD, et al. (2012) Unliganded HIV-1 gp120 core structures assume the CD4- veal an H3 subdomain that mediates potent neutralization of HIV-1. Proc Natl Acad bound conformation with regulation by quaternary interactions and variable loops. Sci USA 107(25):11483–11488. Proc Natl Acad Sci USA 109(15):5663–5668. 45. Edson CM (1993) Tyrosine sulfation of varicella-zoster virus envelope glycoprotein 22. Walker LM, et al.; Protocol G Principal Investigators (2009) Broad and potent neu- gpl. Virology 197(1):159–165. tralizing antibodies from an African donor reveal a new HIV-1 vaccine target. Science 46. Mascola JR, et al. (2005) Recommendations for the design and use of standard virus 326(5950):285–289. panels to assess neutralizing antibody responses elicited by candidate human im- 23. Walker LM, et al.; Protocol G Principal Investigators (2011) Broad neutralization – coverage of HIV by multiple highly potent antibodies. Nature 477(7365):466–470. munodeficiency virus type 1 vaccines. J Virol 79(16):10103 10107. 24. Cao J, et al. (1997) Replication and neutralization of human immunodeficiency virus 47. Pitisuttithum P, et al.; Bangkok Vaccine Evaluation Group (2006) Randomized, dou- type 1 lacking the V1 and V2 variable loops of the gp120 envelope glycoprotein. ble-blind, placebo-controlled efficacy trial of a bivalent recombinant glycoprotein 120 J Virol 71(12):9808–9812. HIV-1 vaccine among injection drug users in Bangkok, Thailand. J Infect Dis 194(12): 25. Pinter A, et al. (2004) The V1/V2 domain of gp120 is a global regulator of the sen- 1661–1671. sitivity of primary human immunodeficiency virus type 1 isolates to neutralization by 48. Rerks-Ngarm S, et al.; MOPH-TAVEG Investigators (2009) Vaccination with ALVAC and antibodies commonly induced upon infection. J Virol 78(10):5205–5215. AIDSVAX to prevent HIV-1 infection in Thailand. N Engl J Med 361(23):2209–2220.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1314718111 Cimbro et al. Downloaded by guest on October 2, 2021