JBC Papers in Press. Published on February 16, 2011 as Manuscript M

JBC Papers in Press. Published on February 16, 2011 as Manuscript M111.224527 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.m111.224527 GENETIC AND NEUTRALIZATION SENSITIVITY OF DIVERSE HIV-1 ENV CLONES FROM CHRONICALLY INFECTED PATIENTS IN CHINA Hong Shang, Xiaoxu Han, Xuanling Shi, Teng Zuo, Mark Goldin, Dan Chen, Bing Han, Wei Sun, Hao Wu, Xinquan Wang, Linqi Zhang From the Key Laboratory of AIDS Immunology of Ministry of Health, Department of Laboratory Medicine, No. 1 Hospital of China Medical University, Shenyang 110001, China Comprehensive AIDS Research Center, School of Medicine, Tsinghua University, Beijing, 100084, China. Department of Infectious Diseases, YouAn Hospital, Capital University of Medical Sciences, Beijing, 100054, China. Center for Structural Biology, School of Life Sciences, Tsinghua University, Beijing 100084, China. Running head: Neutralization sensitivity of HIV-1 env clones from China Corresponding author: Linqi Zhang, Comprehensive AIDS Research Center, School of Medicine, Tsinghua University, Beijing, 100084, China. Tel.: 010-6278-8131; Fax: 010-6279-7732; E-mail: zhanglinqi@tsinghua.edu.cn, or Hong Shang, AIDS Research Center, No. 1 Hospital of China Medical University, Shenyang, Liaoning 110001, China. Tel: 86-24-83282634; E-mail:hongshang100@hotmail.com As HIV-1 continues to spread in China from traditional high-risk populations to the general public, its genetic makeup has become increasingly complex. However, the impact of these genetic changes on the biological and neutralization sensitivity of the virus is unknown. The current study aims to characterize the genetic, biological and neutralization sensitivity of HIV-1 identified in China between 2004 and 2007. Based on a total of 107 full-length envelope genes obtained directly from the infected patients, we found those viruses fell into three major genetic groups CRF01_AE, subtype B, and subtype C/07/08/B C. Pseudotyped viruses built upon the viable env genes have demonstrated their substantial variability in mediating viral entry, and in sensitivity to neutralization by subtype-specific plasma pools and broadly neutralizing monoclonal antibodies (bnmab). Many viruses are resistant to one or more bnmab including those known to have high potency against diverse viruses from outside China. Sequence and structural analysis has revealed several mechanisms by which these resistant viruses escape recognition from bnmab. We believe that these results will help us to better understand the impact of genetic diversity on the neutralizing sensitivity of the viruses, and to facilitate design of immunogens capable of eliciting antibodies with similar potency and breadth as bnmab. China is facing a continued upsurge in the prevalence of human immunodeficiency virus type 1 (HIV-1). Historically, HIV-1 infection has been largely confined to high-risk populations such as injecting drug 1 Copyright 2011 by The American Society for Biochemistry and Molecular Biology, Inc.

users (IDUs) and formal commercial blood and plasma donors (FBDs) in geographically separate rural areas. However, HIV-1 prevalence has recently increased rapidly in urban settings through heterosexual and homosexual transmission, raising serious concerns about further spread among the general public (1,2). With the dramatic change in disease distribution, the genetic makeup of HIV-1 in China has become increasingly complex. We and others have demonstrated that HIV-1 strains circulating in China can be broadly classified into three major groups: those clustering closely with circulating recombinant form (CRF) 01_AE, subtype B, or subtype C/CRF07_BC/CRF08_BC/B C (C/07/08/B C for abbreviation) (3-5). Although the proportion of each group among all infected cases remains unknown, it is clear that C/07/08/B C viruses are strongly associated with intravenous drug use, and subtype B viruses with commercial blood and plasma donation, while CRF01_AE viruses are associated with sexual transmission (3,5). Furthermore, most recent studies suggest a rising predominance of CRF01_AE, particularly among those infected through homosexual contact (3). It is unknown whether high prevalence of particular subtypes in a population is emblematic of unique biologic features, or largely a founder effect. But nevertheless, as the number of different subtypes of viruses increases and circulates within the same geographic locations, more recombinant viruses will be expected to emerge and spread. Increased genetic complexity has posed greater challenges to human immune system and vaccine development in China. Some reports suggest that neutralizing activity of sera from HIV-1 infected patients may be independent of genetic subtypes (6-8), others present evidence of possible subtype-specific humoral immune responses (9-12). In either case, we need to better understand the genetic and neutralization sensitivity of viruses. In particular, as several bnmab have recently been isolated from infected individuals with exceptionally broad serum neutralizing activities, it would be critical to investigate whether the epitopes or vulnerable sites identified by these bnmab are also present on the envelope of HIV-1 identified in China. These conserved epitopes include the CD4 binding site on envelope gp120 recognized by bnmab IgG1b12 and VRC01(13-15), the membrane proximal external region (MPER) of gp41 recognized by bnmab 2F5 and 4E10 (16), the glycan moiety on gp120 recognized by bnmab 2G12 (17-19), and epitopes that reside in the V1/V2 and V3 regions on gp120 recognized by bnmab PG9, PG16 (20). Characterization of these epitopes and their impact on neutralization sensitivity will provide crucial molecular bases for the design of immunogens capable of eliciting antibodies with similar potency and breadth as these bnmab (21,22). At present, our knowledge on HIV-1 in China is largely derived from genetic characterization of relatively short viral gene fragments (3-5) or limited number of full-length envelope clones (23,24). While these studies provide important references on phylogenetic as well as epidemiologic patterns of viral spread, they offer limited insight into the functional features of the viruses. Our study aims to systematically characterize genetic, biologic and 2

immunologic properties of diverse HIV-1 in China. We have generated and characterized 107 full-length envelope (env) molecular clones from multiple risk groups in various geographic locations in China. These env clones are further characterized for their biological and immunologic properties in the context of pseudotyped viruses. In particular, we focus on studying sensitivity of viruses to neutralization by subtype-specific plasma pools generated from chronically infected patients in China, as well as by abovementioned bnmab. Our results have demonstrated that HIV-1 in China is genetically diverse, resulted in their significant differences in mediating viral entry, and in sensitivity to neutralization by subtype-specific plasma pools and bnmab. The implications of our findings for evaluating neutralizing antibody responses in infected individuals and in participants of vaccine clinical trials as well as for future vaccine development are discussed. EXPERIMENTAL PROCEDURES Study subjects A total of 56 HIV-1-positive blood samples were collected between 2004 and 2007 from a disease progression cohort consisting of infected individuals from Yunnan, Xinjiang, Henan, Liaoning, and Jilin provinces of China. The demographic, virologic and immunologic characteristics of these subjects are summarized in Table 1. At the time of blood sampling, all patients were treatment-naïve and some showed symptoms of AIDS with opportunistic infections. Informed consent was written and provided by all study participants and the study was approved by the institution s ethical committee at the No. 1 Hospital of China Medical University in Shenyang, Liaoning Province. Plasma samples, mabs and soluble CD4 Individual plasma samples were obtained from 90 chronically infected patients. Among them, 30 fell into each of three groups patients infected with HIV-1 related to CRF 01_AE, subtype B, or subtype C/07/08/B C based on genetic analysis of gag, pol and env genes as previously described(3). For each group, a pooled plasma sample was generated by mixing equal proportions of the 30 corresponding individual plasma samples. All plasma samples were heat inactivated for 1 h at 56ºC and stored at -80ºC until use. None of these patients had been on any antiretroviral therapy at the time of sampling. The bnmab PG9 and PG16 were kindly provided by Wayne Koff at International AIDS Vaccine Initiative and VRC01 was kindly provided by John Mascola at National Institute of Health. The rest of the bnmab were obtained from NIH reference and reagents program. IgG1 b12, first isolated as an Fab fragment through phage display technologies, recognizes a conserved CD4-binding domain on gp120 and is sensitive to substitutions in multiple regions in gp120 (13,25,26). 4E10 and 2F5 recognize adjacent liner epitopes NWFDIT and ELDKWA respectively, which are located in the membrane-proximal external region (MPER) of HIV-1 gp41 (27-29). The 4E10 epitope NWFDIT, located immediately N-terminal of the transmembrane domain of gp41, is partially inserted into the lipid bilayer (29). 2G12 recognizes an epitope in the C2-V4 region of gp120 involving carbohydrate moieties of N-glycosylated residues (17-19). PG9 and PG16, initially isolated from a subtype A-infected African donor, recognize 3

conserved regions of the V2 and V3 loops of gp120, although PG16 is more sensitive to V3 loop substitutions than PG9(20). VRC01 recognizes the outer domain-contact site for CD4 of gp120, one of the vulnerable sties identified previously (14,15). Soluble CD4 (scd4) was provided by Progenics Pharmaceuticals, Inc. (Tarrytown, NY). Cloning of full-length envelope genes and production of pseudotyped viruses Full-length envelope genes were amplified by PCR directly from proviral DNA extracted from patients unclutured PBMC. The PCR was conducted with initial denaturation at 94ºC for 2min, followed by 35 cycles of 94ºC for 15 seconds, 55ºC for 30 seconds and 68ºC for 4 minutes, followed by a final extension at 68ºC for 10 minutes. Subtype-specific primer sequences (listed in the supplemental materials) were designed to be as conserved as possible based on the published sequences of geographical variants. PCR-amplified fragments were cloned directly into the pcdna 3.1 expression vector (Invitrogen, Carlsbad, CA, USA) and verified by direct sequencing. Env-bearing pseudotyped viruses were generated by co-transfection of env-expressing plasmid together with backbone construct pnl43r-e-luciferase into the 293 cells. A control plasmid expressing envelope glycoprotein of HIV-1 HXB2, SF162, or JRFL and of amphotropic murine leukemia virus (AMLV) was also included. Forty-eight hours post transfection, culture supernatant was collected and tested for luciferase activity to standardize viral input in the subsequent functional analysis. Single-genome amplification (SGA) of full-length HIV-1 genome for recombinant analysis Viral RNA was recovered from the plasma and then reverse-transcribed into cdna using SuperScript III (Invitrogen Life Technologies, Carlsbad, CA) followed by PCR amplification. The primer sequences are listed in the supplemental materials. The same reaction conditions were used for the first and the second rounds of PCR: 94ºC for 2 min followed by 30 cycles of 94ºC for 15 s, 60ºC for 30 s, and 68ºC for 4 min, with a final extension of 68ºC for 10 min. The sequencing of purified PCR products was conducted without cloning (Applied Biosystems, USA). Determination of viral co-receptor usage Pseudotyped viruses were co-cultured with U87MG.CD4/X4 and U87MG.CD4/R5 or GHOST.CD4/X4/R5 cells stably transfected to express HIV-1 receptor CD4 and co-receptor CCR5 or CXCR4 as previously described (30). Luciferase activity in infected cells was measured 48 h post infection. Neutralizing activity of plasma and bnmab against pseudotyped virus Neutralizing activities of bnmab, scd4 and pooled plasma samples were jointly analyzed in our laboratory and by Monogram Biosciences (CA, USA). All plasma samples were heat-inactivated at 56ºC for 1h before testing. In brief, 100 TCID 50 of pseudotyped viruses was incubated with serially diluted patients plasma, scd4 or bnmab in a 96-well plate in triplicate for 1 h at 37ºC. Approximately 2x10 4 TZM-bl cells were then added and the cultures were maintained for an additional 48h at 37ºC. Neutralizing activity was measured by the reduction in luciferase activity compared with controls. The log 10 ID 50 titers were calculated based on the standard algorithm published previously (31). The heatmap program on the Los Alamos database 4

(http://www.hiv.lanl.gov/content/sequence/ HEATMAP/heatmap.html) was used to analyze the clustering patterns for viruses, plasma pools and bnmab (32,33). Sequence and statistic analysis Sequences were aligned using the Clustal W program together with selected subtypes/crfs of geographical importance. Phylogenetic analysis was conducted using the neighbor-joining method (34). The reliability of the branching orders was tested by bootstrap analysis of 1,000 replicates (35). Bootscanning, similarity plots and informative site analysis were performed using SimPlot v3.5 (36) to define the recombination structure. The Genbank accession numbers of nucleotide sequences reported here are HM215397-HM215428, and HQ699949-HQ700004. Structure modeling of loop D and V5 region in gp120 resistant to VRC01 We selected the gp120 structure (Clade A/E 93TH057) in complex with antibody VRC01 (PDB code: 3NGB) as the template for the modeling. Loop D and V5 region in the gp120 structure were replaced by the respective sequences from HIV strains in China that are resistant to antibody VRC01. Modeling of loop D and V5 region was carried out using ModLoop server (37). The gp120 structures with modeled loop D and V5 region were then docked onto VRC01 with 3NGB as the template. Interactions closer than 2.5 Å between the loop D and V5 regions of gp120 and the antibody VRC01 were identified with the program CONTACT in the CCP4 program suite (38). The structure figure was made with program PyMol (DeLano Scientific, San Carlos, CA, USA). RESULTS Genetic characterization of full-length envelope molecular clones A total of 107 full-length envelope molecular clones were successfully obtained directly from PBMC of 56 chronically infected individuals without in vitro culture. Phylogenetic analysis shows that all sequences fell into three major groups: those clustering closely with CRF 01_AE, subtype B, or subtype C/07/08/B C except for CNE85 which grouped with subtype A (Figure 1). There are six clones (CNE87, CNE15, CNE16, CNE18, CNE7, and CNE67) in C/07/08/B C group that failed to group with prototype reference sequence 07.BC.CN54 or 08.BC.CN.GX6F. Recombination analysis using the SimPlot program indicates the novel recombinant forms of these clones (Figure 2, right panel). Furthermore, sequences in the CRF01_AE and subtype B groups are closely related to reference strains 90TH.CM240 and 99TH.MU2079, respectively, originally identified in Thailand, whereas those in the C/07/08/B C group cluster tightly with reference strain 95IN.21068 from India and prototypes 07.BC.CN54 and 08.BC.CN.GX6F from China (Figure 1). These results suggest that the HIV-1 epidemic in China is the result of multiple introductions of various HIV-1 strains from neighboring countries. The distribution and relative proportion of each HIV-1 subtypes and CRFs in each province is shown (Figure S1). Lastly, as a large fraction of the patients (35/56) came from Yunnan, the observed genetic complexity and distribution pattern are in complete agreement with our previous report (1,5). However, identification of multiple subtypes and CRFs outside of Yunnan has reaffirmed the concerns that HIV-1 continues to spread beyond high risk 5

populations in severely affected area (3,4). To further confirm new recombinant variants, we amplified the full-length genome sequences of HIV-1 from 2 (CNE 15 and CNE18) of the 6 patients indicated above. Single-genome amplification (SGA) and recombination analysis using SimPlot showed both genomes had recombination patterns different from the prototype 07.BC.CN54 and 08.BC.CN.GX6F (Figure 2). CNE15, for instance, has multiple recombination points between subtype B and C across the entire genome, while CNE18 is predominantly comprised of the subtype B sequence with two subtype C fragments in the 3 -half of the genome (Figure 2). This finding clearly shows that CNE15 and CNE18 are novel recombinants. Functional domain analysis of full-length envelope molecular clones We performed functional domain analysis by aligning the novel HIV-1 sequences from China against prototype HIV-1 sequences HXB2, NL43, JRFL, JFCSF, and SF162 (see supplemental materials and Figure S2). We further analyzed the degree of length polymorphism in regions V1/V2, V3, V4, and V5 based on subtype or CRF classification of 71 viable sequences (Figure 3a). The length variation is mostly identified in V1/V2, V4 and V5 whereas the V3 loop remains constant at 35 residues in length. The V1/V2 region (70±7.1) contains the largest number of residues and the highest degree of length variation compared to V4 (31±4.5) and V5 (10±1.7) (Figure 3a). In addition, a high degree of variation in the number and location of potential N-linked glycosylation sites was found among sequences in different subtypes and CRFs. Figure 3b shows that the number of N-linked glycosylation sites in gp120 is more than that in gp41, with a significantly higher average for C/07/08/B C (26 ± 3.0) than for CRF01_AE (24 ± 1.7) (P<0.01). Most of these glycosylation sites are either located within the V1/V2, V4, and V5 regions or in the flanking regions of the V3 loop (Figure S1). In gp41, a large majority of the glycosylation sites are located in the ectodomain outside the transmembrane (Figure 3b). Neutralization sensitivity of diverse HIV-1 to subtype-specific plasma pools from infected patients We generated 88 pseudotyped viruses and found only 71 were able to mediate viral entry while the remaining 17 clones, despite having viable sequences at the nucleotide level, failed to do so. In addition, among the 71 full-length envelope clones, substantial variation was found in mediating viral entry. We have chosen the top 31 clones with the highest entry capacity to study their neutralization sensitivities to pooled plasma and bnmab. Table 2 (left half) summarizes the basic clinical and biological features associated with these 31 clones. As shown, 28 out of 31 (90%) viruses used co-receptor CCR5 for entry while the remaining 3 (10%) used CXCR4. We first studied neutralization sensitivity to the subtype-specific plasma pools B, C/07/08/B C and CRF_01AE from chronically infected patients in China. Each pool is constructed with equal proportions of 30 individual plasma samples with known HIV-1 subtypes or CRFs and with reasonably good neutralization activity (data not shown). Figure 4 shows that many of the viruses share similar levels of sensitivity to neutralization, but there are two extreme subsets at either end of the 6

curve that are either highly sensitive or highly resistant. This finding is consistent with that previously reported by Seaman and colleagues where variable degrees of sensitivity to neutralization were also found among a panel of 109 viruses against 7 different subtype-specific plasma pools (33). The three most sensitive viruses (SF162, CNE20, and CNE40) were sensitive to all three plasma pools, with average reciprocal log 10 ID 50 titers ranging from 3.0 to 4.0, whereas that for the two most resistant viruses (CNE23, and CNE30) was below 1.2. The remaining viruses showed values ranging from 1.4 to 2.9. In addition, for each particular virus, there were differences in neutralization sensitivity to the three different plasma pools, reflected by the differences in the log 10 ID 50 titers along the Y-axis (Figure 4). However, the neutralization sensitivity of these viruses in general follows the same trend towards all three plasma pools. In other words, when a particular virus is among the most sensitive to one plasma pool, it would most likely be among the most sensitive to the other pools (Figure 4). This finding suggests that neutralization sensitivity is largely virus-dependent, although neutralizing activity against a matched subtype of virus in a given plasma pool is indeed more potent than that of a cross-subtype or CRF (Figure 4). Neutralization sensitivity of diverse HIV-1 strains to scd4 and bnmab We further studied the sensitivity of these viruses to neutralization by scd4 and several well-known bnmab such as 2F5, 4E10, 2G12, PG9, PG16, IgG1b12, and VRC01 (14,16-20). All these bnmab were isolated from infected individuals with potent serum neutralization activity against a large panel of viruses, and demonstrated strong protection against infection in passively immunized macaques (39-42). Some of these bnmab have also been used in humans and demonstrated good safety and efficacy profiles in temporally inhibiting viral replication (43). As shown in Table 2 (right panel), 2F5 and 2G12 failed to neutralize almost all viruses in the C/07 /08/B C group at a concentration of 50µg/ml. 2F5 is potent in neutralizing viruses in subtype B and CRF01_AE. 2G12, on the other hand, can only neutralize a fraction (6/9) of subtype B viruses and none of the CRF01_AE viruses. In contrast, 4E10 and scd4 are active against all viruses tested with average IC 50 values of 2.95±3.40 and 3.36±3.53 µg/ml, respectively. Furthermore, PG9 and PG16, recently isolated from a subtype A-infected African donor demonstrated reasonably good neutralization potency as well as breadth. PG9 is able to neutralize 81% (25/31) while PG16 can neutralize 71% (22/31) of the viruses tested, with average IC 50 values of 0.32±0.71 and 1.74±3.90 µg/ml, respectively. It is interesting to note that viruses insensitive to PG9 are all equally insensitive to PG16 but not the other way around (Table 2), suggesting PG9 can tolerate more viral glycoprotein amino acid substitutions than PG16. However, both PG9 and PG16 demonstrated particularly poor neutralization activity against viruses in the subtype B group, with only 5 out of 9 (44.4%) viruses inhibited (Table 2). The remaining two bnmab, IgG1b12 and VRC01, demonstrated significant differences in neutralization potency and breadth, despite both of them recognize the critical CD4-binding domain (14,15,20,44). IgG1b12 can only neutralized 45% (14/31) while VRC01 neutralized about 81% (25/31) of the viruses tested (Table 2). IgG1b12 had an average IC 50 of 13.33±17.56 µg/ml 7

while that for VRC01 was about 1.14±1.80 µg/ml. IgG1b12 insensitive viruses were found in all three subtype or CRF groups, while those insensitive to VRC01 were mainly from subtype B and C/07/08/B C (Table 2). Lastly, eight viruses (CNE1, CNE66, CEN23, CNE30, CNE46, CNE47, CNE53 and CNE107) were resistant to 4 of the 7 bnmab tested, and two of which (CNE1 and CNE23) were resistant to both VRC01 and PG16. However, no virus identified so far can overcome the neutralization activity of all these bnmab, indicating that there exist multiple vulnerable sites on the viral envelope that can be targeted by antibody responses, and could potentially be used for vaccine design. Neutralization sensitivity toward positive control serum sample Z23 was also found to be highly variable among the viruses tested (Table 2). Lack of association between neutralization sensitivity to plasma pools and bnmab We further analyzed whether neutralization sensitivity to bnmab was predictive of that to the plasma pools or vice versa. We used two-dimensional clustering heatmap approach to group viruses with similar neutralization sensitivity to bnmab or the plasma pool. As shown in Figure 5, significant differences were found in the number and pattern of virus groups based on the two methods. For example, 3 groups of viruses were identified based on their sensitivity to bnmab and scd4 (Figure 5a) while the same set of viruses separated to at least 5 groups based on their sensitivity to plasma pools (Figure 5b). Furthermore, no direct association between the groups identified by bnmab and plasma pools was identified. For instance, viruses in one group identified by bnmab such as CNE15, CNE 16, CNE5, CNE58, CNE49, CNE19, CNE68, CNE65, and CNE17 are scattered to many groups identified by the plasma pool. The most sensitive viruses to the plasma pools such as CNE40, CNE20, and CNE14 failed to have similar sensitivity profile to bnmab (Figure 5). These results suggest that the number and patterns of virus groups identified by the plasma pools are more complex than that revealed by the individual bnmab. Polyclonal plasma samples are more suitable for evaluating overall sensitivity of virus while bnmab are particularly useful for identifying and grouping viruses based on defined number of epitopes. Analysis of epitope sequences recognized by bnmab One major mechanism for neutralization insensitivity may be due to mutations in the bnmab epitope. We therefore compared epitope sequences between sensitive and insensitive viruses tested. Table 3 summarizes the key residues known to be important for epitope formation of listed bnmab. As shown earlier, 2F5 recognizes the epitope centered on peptide sequence ELDKWA, and mutagenesis studies have revealed that LDKW is the critical core epitope and the major determinant for 2F5 neutralization (16,45). In particular, amino acid substitutions at K665 appeared to be the key determinants of resistance (45). In this study, all 2F5-resistant viruses were in the C/07/08/B C group and had lysine (K) to serine (S) substitution at position 665, consistent with their phenotypic resistance (Table 3). 4E10 binds an epitope consisting of core sequence NWFDIT, located at the C-terminus of the 2F5 epitope. Residues beyond these six residues have also been found to be critical for binding (16,27). 8

Mutagenesis studies have identified residues W672, F673 and W680 as critical for binding as well as neutralization activity (27). Analysis of epitope sequences of the viruses tested has revealed complete conservation of these three key residues, providing good explanation for their sensitive phenotype. Observed substitutions at position 671, 674, 675, and 676 therefore have minimal effect on viral sensitivity to 4E10 (Table 3). 2G12 recognized a complex mannose cluster on gp120 involving N-linked glycans at positions 295, 332, 339, 392, 386, 397, and 448 (17-19). Sequence analysis of 2G12-resistant viruses has revealed that 23/24 (96%) of the viruses lack the glycan at position 295, or 332, or both (Table 3), confirming these two sites are critical for 2G12 binding. CNE57, on the other hand, is highly resistant but possesses all the glycans except at position 397. However, the same glycan distribution pattern was also found for three sensitive viruses CEN9, CNE10 and CNE64, suggesting that mutations apart from that of the glycan at position 397 play an important role in conferring observed resistance. In addition, a marginally resistant CNE49 lacks glycans at position 339, 397 and 448, which may indicate that combinational changes can also contribute to the resistant phenotype. PG9 and PG16 bind to overlapping but distinct gp120 eptiopes involving V2, V3, the V1/V2 stem, and perhaps a portion of the co-receptor binding site (20,46). Alanine scanning mutagenesis has shown that both antibodies are equally sensitive to substitutions within V1 and V2, although PG16 is more sensitive to V3 loop substitutions than PG9 (20). Sequence analysis of PG9/PG16-resistant viruses has indicated complex mutation patterns for residues critical for PG9/PG16 binding, some of which are more obvious than others. For resistant viruses in subtype B group, for instance, three (CNE57, CNE14, and CNE9) have residue substitutions N160D, S162N, or F176L in the V2 domain (Table 3), known to dramatically reduce the recognition of PG9 and PG16 (20). The double resistant phenotype of CNE4, on the other hand, could be due to the combination effect of three substitutions, I307M, I309L and F317W, each shown to have a moderate effect on PG9 and PG16 resistance (20). PG16-resistant CNE1 is the likely result of two substitutions, I309L and F317W (Table 3). For resistant viruses in C/07/08/B C group, however, the potential resistance mechanism is less clear. PG16-resistant viruses such as CNE23, CNE46 and CNE53 have residue sequences identical to sensitive ones, such as CNE19, CNE29, CNE40 and so on, with the exception of CNE46 which has a lysine (K) to arginine (R) substitution at position 305. This finding suggests that the resistant phenotype of these three viruses is likely due to substitutions elsewhere. In addition, the double resistant phenotype of CNE30 could be due to combinational substitutions Y173H, K305R, and V318F (Table 3). Lastly, for CNE107 in the CRF01_AE group, substitutions S158T, I309M and F317Y may jointly contribute to its resistant phenotype, since substitution S158T alone, although critical for binding of both PG9 and PG16 in the context of JRCSF (20), did not result in resistance for CNE55. It is therefore possible the substitutions as well as the context where these substitutions occurred jointly determined the resistance phenotype. IgG1 b12, first isolated as an Fab fragment through phage display library derived from a clade B HIV-1-infected 9

individual, binds to the CD4-binding domain on gp120 and is sensitive to substitutions in multiple regions on gp120 (13,25,26,44). It has an unusual structure with long complementarity determining loops that allow it to access the CD4-binding site on some primary HIV-1 isolates(44). Based on the a combination of sequence analysis, computational modeling, and site-directed mutagenesis, it was found that several substitutions within the dominant IgG1 b12 contact surface, called the CD4-binding loop, mediated IgG1 b12 resistance(25). Other reports also shown that substitutions in both the V2 and C3 regions and perhaps in gp41 confer IgG1 b12 resistance (26,47). Sequence analysis of IgG1 b12-resistant viruses has failed to reveal any obvious amino acid substitutions or signatures associated with their resistant phenotype. It is highly possible that observed resistant phenotype is not only dependent on the residues at the major contact sites between IgG1 B12 and gp120 but also on the context in which the major contacts sites exist. Further studies would be required to define the genetic bases for observed resistant phenotype identified in this study. Structure analysis of VRC01-resistant envelope glycoprotein VRC01 recognizes the outer domain-contact site for CD4 of gp120 and is able to neutralized over 90% of circulating HIV-1 isolated identified outside China (14,15). Previous structural studies have indicated that residues with bulky side chain in the loop D and V5 regions of gp120 may contribute to the natural resistance to VRC01 by some strains (15). Structural modeling proposed that the steric clashes brought by these bulky residues would interfere with the gp120-vrc01 interaction (15). To analyze potential resistance mechanism, we conducted sequence and structural analysis of VRC01-resistant viruses indentified in our study. As shown in Figure 6, sequence alignments also revealed residues with bulky side chain in the loop D and V5 regions of gp120 from resistant isolates in China. Structural modeling and docking analysis of these resistant strains support the potential roles of these residues in the natural resistance to VRC01, either through conformational masking and glycan shielding. Future studies will be required to dissect out the actual mechanism for their resistant phenotype. DISCUSSION We have describes here the first extensive characterization of the genetic and neutralization sensitivity of diverse HIV-1 strains from China. Phylogenic analysis based on full-length env clones has revealed that these viruses belong largely within CRF01_AE, subtype B, and subtype C/07/08/B C, with several clones bearing novel recombinant features. Genetically, these strains are clearly distinct from those circulating in other parts of the world. While subtype B has been dominating in the United States and Western Europe, multiple subtypes and CRFs have been identified in Africa (48). In Asia, as the epidemic continues to spread from the high risk populations to the general public, the genetic makeup of HIV-1 has become increasingly complex. For instance, while CRF01-AE is dominant in Thailand, Vietnam, Cambodia, and Myanmar, subtype C is the most prevalent in India (49). At the current stage, it is uncertain whether the current epidemic can be explained by how founder viruses spread among various risks populations or by genetic and biologic 10

differences among various human populations in different continents. In addition, we conducted biologic and immunologic characterization of full-length env clones in the context of pseudotyped viruses. Our study has identified a spectrum of viral differences in co-receptor usage, in mediating viral entry, and in sensitivity to neutralization by subtype-specific plasma pools, scd4 and bnmab. In particular, while a large proportion of viruses exhibited a similar sensitivity, a small subset was found to be either highly sensitive or resistant to neutralization by the three plasma pools. This finding is consistent with previous reports in which a three-tier or rank-order approach was undertaken to classify viral neutralization sensitivity to individual or pooled plasma (33). Furthermore, viral sensitivity to plasma neutralization seems dependent more on viral characteristics than on the type of plasma pools used. This hypothesis is supported by the observation that a virus sensitive to one plasma pool was often sensitive to the others. However, subtype-matched neutralization was indeed more potent than with non-matched pairs, suggesting viruses within one subtype share more neutralizing epitopes than those between subtypes. But the fact that some degree of cross-neutralization does occur between subtypes argues for the existence of common epitopes shared by all viruses. Identifying these conserved epitopes and ultimately using them to induce broadly neutralizing antibodies are therefore paramount in our vaccine development against HIV-1. through cloning and characterization of a series of bnmab from individuals with potent and broad neutralizing activities (14,20,32,50). One major objective of this study was to evaluate how conserved these epitopes were among HIV-1 strains in China by analyzing their relative sensitivity to neutralization by these bnmab. We found that all viruses tested were sensitive to neutralization by 4E10 and scd4. This is in great contrast to viruses isolated during early infection where vast majority are resistant to scd4 (51). Furthermore, many were resistant to one or more bnmab including those recently identified PG9, PG16 and VRC01, known to possess exceedingly high potency and breadth against diverse viruses from outside China (14,20). The resistant viruses to 2F5 are largely involved in changes to the epitope region critical for antibody binding. Viral resistance to 2G12, PG9, PG16, IgG1b12 and VRC01, however, is perhaps due to changes that ultimately lead to conformational alternations in the binding domain of these antibodies. These changes could affect direct binding, and/or indirectly shield the binding domain through steric hindrance. Changes in glycan patterns as well as in charge and side chain properties of relevant residues are good examples (15,17,19). In any case, identification of HIV-1 in China resistant to one or more bnmab strongly argues for larger scale studies including more diverse HIV-1 strains across the world. Only with such studies will we better understand the relative conservation of these epitopes and their potential use as immunogens capable of eliciting antibodies with similar potency and breadth as bnmab. Recently, great progress has been made in identifying these conserved epitopes, In summary, we have systematically characterized diverse HIV-1 viruses in 11

China and have demonstrated significant differences in their biologic and immunologic properties. We believe that our results will help us to better understand, at the molecular level, the interplay between diverse viral strains and antibody responses in infected patients in China. The viruses characterized here will also provide a strong foundation for establishing a broader and more representative Chinese virus panel to evaluate the antibody responses in infected and vaccinated individuals, and to contribute to the international virus panel where only a limited number of Chinese viruses are currently available (33). ACKNOWLEDGEMENTS We are grateful to Dr. Wayne Koff at International AIDS Vaccine Initiative (IAVI) and Drs. John Mascola and Tongqing Zhou of Vaccine Research Center at National Institute of Health (NIH) for providing bnmab PG9, PG16 and VRC01, respectively. We thank Drs. David Montefiori and Marcella Sarzotti-Kelsoe at Duke Human Vaccine Institute for their kind suggestions and advice. We would also like to thank all the study participants and research staff at Tsinghua University, China Medical University and YouAn Hospital, Capital University of Medical Sciences. REFERENCES 1. Lu, L., Jia, M., Ma, Y., Yang, L., Chen, Z., Ho, D. D., Jiang, Y., and Zhang, L. (2008) Nature 455, 609-611 2. China, Ministry of Health Report (2010) China 2010 UNGASS Country Progress Report (2008-2009). 3. Han, X., Dai, D., Zhao, B., Liu, J., Ding, H., Zhang, M., Hu, Q., Lu, C., Goldin, M., Takebe, Y., Zhang, L., and Shang, H. (2010) J Acquir Immune Defic Syndr 53 Suppl 1, S27-33 4. Zhang, L., Chen, Z., Cao, Y., Yu, J., Li, G., Yu, W., Yin, N., Mei, S., Li, L., Balfe, P., He, T., Ba, L., Zhang, F., Lin, H. H., Yuen, M. F., Lai, C. L., and Ho, D. D. (2004) J Virol 78, 13591-13599 5. Zhang, Y., Lu, L., Ba, L., Liu, L., Yang, L., Jia, M., Wang, H., Fang, Q., Shi, Y., Yan, W., Chang, G., Zhang, L., Ho, D. D., and Chen, Z. (2006) PLoS Med 3, e443 6. Kostrikis, L. G., Cao, Y., Ngai, H., Moore, J. P., and Ho, D. D. (1996) J Virol 70, 445-458 7. Moore, J. P., Cao, Y., Leu, J., Qin, L., Korber, B., and Ho, D. D. (1996) J Virol 70, 427-444 12

8. Weber, J., Fenyo, E. M., Beddows, S., Kaleebu, P., and Bjorndal, A. (1996) J Virol 70, 7827-7832 9. Bures, R., Morris, L., Williamson, C., Ramjee, G., Deers, M., Fiscus, S. A., Abdool-Karim, S., and Montefiori, D. C. (2002) J Virol 76, 2233-2244 10. Mascola, J. R., Louder, M. K., Surman, S. R., Vancott, T. C., Yu, X. F., Bradac, J., Porter, K. R., Nelson, K. E., Girard, M., McNeil, J. G., McCutchan, F. E., Birx, D. L., and Burke, D. S. (1996) AIDS Res Hum Retroviruses 12, 1319-1328 11. Mascola, J. R., Louwagie, J., McCutchan, F. E., Fischer, C. L., Hegerich, P. A., Wagner, K. F., Fowler, A. K., McNeil, J. G., and Burke, D. S. (1994) J Infect Dis 169, 48-54 12. van Gils, M. J., and Schuitemaker, H. (2010) Curr HIV Res 8, 579-586 13. Barbas, C. F., 3rd, Bjorling, E., Chiodi, F., Dunlop, N., Cababa, D., Jones, T. M., Zebedee, S. L., Persson, M. A., Nara, P. L., Norrby, E., and et al. (1992) Proc Natl Acad Sci U S A 89, 9339-9343 14. Wu, X., Yang, Z. Y., Li, Y., Hogerkorp, C. M., Schief, W. R., Seaman, M. S., Zhou, T., Schmidt, S. D., Wu, L., Xu, L., Longo, N. S., McKee, K., O'Dell, S., Louder, M. K., Wycuff, D. L., Feng, Y., Nason, M., Doria-Rose, N., Connors, M., Kwong, P. D., Roederer, M., Wyatt, R. T., Nabel, G. J., and Mascola, J. R. (2010) Science 329, 856-861 15. Zhou, T., Georgiev, I., Wu, X., Yang, Z. Y., Dai, K., Finzi, A., Kwon, Y. D., Scheid, J. F., Shi, W., Xu, L., Yang, Y., Zhu, J., Nussenzweig, M. C., Sodroski, J., Shapiro, L., Nabel, G. J., Mascola, J. R., and Kwong, P. D. (2010) Science 329, 811-817 16. Stiegler, G., Kunert, R., Purtscher, M., Wolbank, S., Voglauer, R., Steindl, F., and Katinger, H. (2001) AIDS Res Hum Retroviruses 17, 1757-1765 17. Sanders, R. W., Venturi, M., Schiffner, L., Kalyanaraman, R., Katinger, H., Lloyd, K. O., Kwong, P. D., and Moore, J. P. (2002) J Virol 76, 7293-7305 18. Trkola, A., Purtscher, M., Muster, T., Ballaun, C., Buchacher, A., Sullivan, N., Srinivasan, K., Sodroski, J., Moore, J. P., and Katinger, H. (1996) J Virol 70, 1100-1108 19. Scanlan, C. N., Pantophlet, R., Wormald, M. R., Ollmann Saphire, E., Stanfield, R., Wilson, I. A., Katinger, H., Dwek, R. A., Rudd, P. M., and Burton, D. R. (2002) J Virol 76, 7306-7321 20. Walker, L. M., Phogat, S. K., Chan-Hui, P. Y., Wagner, D., Phung, P., Goss, J. L., Wrin, T., Simek, M. D., Fling, S., Mitcham, J. L., Lehrman, J. K., Priddy, F. H., Olsen, O. A., Frey, S. M., Hammond, P. W., Kaminsky, S., Zamb, T., Moyle, M., Koff, W. C., Poignard, P., and Burton, D. R. (2009) Science 326, 285-289 21. Walker, L. M., and Burton, D. R. (2010) Curr Opin Immunol 22, 358-366 22. Hoxie, J. A. (2009) Annu Rev Med 61, 135-152 23. Chong, H., Hong, K., Zhang, C., Nie, J., Song, A., Kong, W., and Wang, Y. (2008) J Acquir Immune Defic Syndr 47, 535-543 24. Nie, J., Zhang, C., Liu, W., Wu, X., Li, F., Wang, S., Liang, F., Song, A., and Wang, Y. J Acquir Immune Defic Syndr 53, 440-450 25. Wu, X., Zhou, T., O'Dell, S., Wyatt, R. T., Kwong, P. D., and Mascola, J. R. (2009) J Virol 83, 10892-10907 13

26. Utachee, P., Nakamura, S., Isarangkura-Na-Ayuthaya, P., Tokunaga, K., Sawanpanyalert, P., Ikuta, K., Auwanit, W., and Kameoka, M. (2010) J Virol 84, 4311-4320 27. Zwick, M. B., Jensen, R., Church, S., Wang, M., Stiegler, G., Kunert, R., Katinger, H., and Burton, D. R. (2005) J Virol 79, 1252-1261 28. Ofek, G., Tang, M., Sambor, A., Katinger, H., Mascola, J. R., Wyatt, R., and Kwong, P. D. (2004) J Virol 78, 10724-10737 29. Muster, T., Steindl, F., Purtscher, M., Trkola, A., Klima, A., Himmler, G., Ruker, F., and Katinger, H. (1993) J Virol 67, 6642-6647 30. Zhang, L., He, T., Huang, Y., Chen, Z., Guo, Y., Wu, S., Kunstman, K. J., Brown, R. C., Phair, J. P., Neumann, A. U., Ho, D. D., and Wolinsky, S. M. (1998) J Virol 72, 9307-9312 31. Richman, D. D., Wrin, T., Little, S. J., and Petropoulos, C. J. (2003) Proc Natl Acad Sci U S A 100, 4144-4149 32. Corti, D., Langedijk, J. P., Hinz, A., Seaman, M. S., Vanzetta, F., Fernandez-Rodriguez, B. M., Silacci, C., Pinna, D., Jarrossay, D., Balla-Jhagjhoorsingh, S., Willems, B., Zekveld, M. J., Dreja, H., O'Sullivan, E., Pade, C., Orkin, C., Jeffs, S. A., Montefiori, D. C., Davis, D., Weissenhorn, W., McKnight, A., Heeney, J. L., Sallusto, F., Sattentau, Q. J., Weiss, R. A., and Lanzavecchia, A. (2010) PLoS One 5, e8805 33. Seaman, M. S., Janes, H., Hawkins, N., Grandpre, L. E., Devoy, C., Giri, A., Coffey, R. T., Harris, L., Wood, B., Daniels, M. G., Bhattacharya, T., Lapedes, A., Polonis, V. R., McCutchan, F. E., Gilbert, P. B., Self, S. G., Korber, B. T., Montefiori, D. C., and Mascola, J. R. (2009) J Virol 84, 1439-1452 34. Saitou, N., and Nei, M. (1987) Mol Biol Evol 4, 406-425 35. Felsenstein, J. (1988) Annu Rev Genet 22, 521-565 36. Lole, K. S., Bollinger, R. C., Paranjape, R. S., Gadkari, D., Kulkarni, S. S., Novak, N. G., Ingersoll, R., Sheppard, H. W., and Ray, S. C. (1999) J Virol 73, 152-160 37. Fiser, A., and Sali, A. (2003) Bioinformatics 19, 2500-2501 38. (1994) Acta Crystallogr D Biol Crystallogr 50, 760-763 39. Baba, T. W., Liska, V., Hofmann-Lehmann, R., Vlasak, J., Xu, W., Ayehunie, S., Cavacini, L. A., Posner, M. R., Katinger, H., Stiegler, G., Bernacky, B. J., Rizvi, T. A., Schmidt, R., Hill, L. R., Keeling, M. E., Lu, Y., Wright, J. E., Chou, T. C., and Ruprecht, R. M. (2000) Nat Med 6, 200-206 40. Hessell, A. J., Poignard, P., Hunter, M., Hangartner, L., Tehrani, D. M., Bleeker, W. K., Parren, P. W., Marx, P. A., and Burton, D. R. (2009) Nat Med 15, 951-954 41. Mascola, J. R., Stiegler, G., VanCott, T. C., Katinger, H., Carpenter, C. B., Hanson, C. E., Beary, H., Hayes, D., Frankel, S. S., Birx, D. L., and Lewis, M. G. (2000) Nat Med 6, 207-210 42. Veazey, R. S., Shattock, R. J., Pope, M., Kirijan, J. C., Jones, J., Hu, Q., Ketas, T., Marx, P. A., Klasse, P. J., Burton, D. R., and Moore, J. P. (2003) Nat Med 9, 343-346 43. Trkola, A., Kuster, H., Rusert, P., Joos, B., Fischer, M., Leemann, C., Manrique, A., Huber, M., Rehr, M., Oxenius, A., Weber, R., Stiegler, G., Vcelar, B., Katinger, H., Aceto, L., and Gunthard, H. F. (2005) Nat Med 11, 615-622 14

44. Zhou, T., Xu, L., Dey, B., Hessell, A. J., Van Ryk, D., Xiang, S. H., Yang, X., Zhang, M. Y., Zwick, M. B., Arthos, J., Burton, D. R., Dimitrov, D. S., Sodroski, J., Wyatt, R., Nabel, G. J., and Kwong, P. D. (2007) Nature 445, 732-737 45. Binley, J. M., Wrin, T., Korber, B., Zwick, M. B., Wang, M., Chappey, C., Stiegler, G., Kunert, R., Zolla-Pazner, S., Katinger, H., Petropoulos, C. J., and Burton, D. R. (2004) J Virol 78, 13232-13252 46. Pancera, M., McLellan, J. S., Wu, X., Zhu, J., Changela, A., Schmidt, S. D., Yang, Y., Zhou, T., Phogat, S., Mascola, J. R., and Kwong, P. D. (2010) J Virol 84, 8098-8110 47. Gnanakaran, S., Daniels, M. G., Bhattacharya, T., Lapedes, A. S., Sethi, A., Li, M., Tang, H., Greene, K., Gao, H., Haynes, B. F., Cohen, M. S., Shaw, G. M., Seaman, M. S., Kumar, A., Gao, F., Montefiori, D. C., and Korber, B. PLoS Comput Biol 6, e1000955 48. Kuiken, C., Korber, B., and Shafer, R. W. (2003) AIDS Rev 5, 52-61 49. Rerks-Ngarm, S., Pitisuttithum, P., Ganguly, N., Zhang, L., Tamashiro, H., Cooper, D. A., Vun, M. C., Bela, B., Ditangco, R., Van Kinh, N., Bernstein, A., Osmanov, S., Mathieson, B., Kent, S. J., and Shao, Y. Curr Opin HIV AIDS 5, 435-452 50. Scheid, J. F., Mouquet, H., Feldhahn, N., Seaman, M. S., Velinzon, K., Pietzsch, J., Ott, R. G., Anthony, R. M., Zebroski, H., Hurley, A., Phogat, A., Chakrabarti, B., Li, Y., Connors, M., Pereyra, F., Walker, B. D., Wardemann, H., Ho, D., Wyatt, R. T., Mascola, J. R., Ravetch, J. V., and Nussenzweig, M. C. (2009) Nature 458, 636-640 51. Wei, X., Decker, J. M., Wang, S., Hui, H., Kappes, J. C., Wu, X., Salazar-Gonzalez, J. F., Salazar, M. G., Kilby, J. M., Saag, M. S., Komarova, N. L., Nowak, M. A., Hahn, B. H., Kwong, P. D., and Shaw, G. M. (2003) Nature 422, 307-312 15

FIGURE LEGENDS Figure 1. Unrooted neighbor-joining tree depicting the genetic relationship among the 88 full-length envelope molecular clones. The branch length is drawn to scale so that the relatedness between different sequences can be readily assessed. Individual sequences clustered with CRF01_AE are colored in green, those clustered with subtype B s in red, and those clustered with subtype C/07/07/B C in blue. A number of commonly used reference sequences for classifying HIV-1 subtypes and CRFs were also included and highlighted in each group with a different color. Closed circles represent those clones capable of mediating viral entry, whereas those indicated by open circles failed to do so. Figure 2. Identification of putative novel recombinants by SimPlot analysis. The right panel depicts the recombination patterns for envelope genes while the left panel for the entire HIV-1 genome. The envelope genes analyzed include CNE15, CNE16, CNE18, CNE7 and CNE67 and were compared with prototype 07_BC.CN54 and 08_BC.GX6F as well as previously identified non-typical recombinants BC.YNRL9607 and BC.YNRL9613. CNE87, shown to be a possible novel recombinant by phylogenetic analysis (Figure 1), is not included as it failed to mediate viral entry. Analysis of the full-length HIV-1 genome was conducted for two clones CNE15-YN8 and CNE18-YN22. The y-axis shows the percentage of identity within a sliding window of 400 bp, with a step size between plots of 40 bp. Comparison of these sequences was made against selected reference strains B.CN.RL42, C.95IN21068 and 01_AE.90.TH.CM240. Figure 3. Comparison of the number of amino acid residues in the variable loops (A) and number of potential N-linked glycosylation sites (B) among CRF01_AE (N=12), C/0708/B C (N=47), and subtype B (N=12) sequences. Significant differences are indicated by horizontal lines. Figure 4. Neutralization sensitivity of selected HIV-1 isolates to subtype-specific plasma pools. The top 31 envelope clones capable of mediating viral entry were evaluated for neutralization sensitivity using the three subtype-specific plasma pools from chronically infected individuals. The average reciprocal log 10 ID 50 titer for each plasma pool is indicated by a representative symbol. The black filled circles indicate the average log 10 ID 50 titer across all three plasma pools. Figure 5. Two-dimensional clustering heatmap to group viruses with similar neutralization sensitivity to bnmab (A), to subtype-specific plasma pools (B), and to compare between the two. The top 31 envelope clones capable of mediating viral entry were assessed for neutralization sensitivities using either 7 bnmab and scd4, or the three subtype-specific plasma pools from chronically infected patients. Individual bnmab, scd4 and subtype-specific plasma pools are indicated at the bottom of the heatmap. The magnitude of neutralization (ID 50 titer) is indicated by color, in which lower neutralization values are represented by lighter colors and higher values by darker colors. Viruses are grouped based on their overall neutralization sensitivity patterns and exemplified by the dendrogram 16

alongside of the heatmap. Straight lines connecting the same viruses between the two heatmaps were used to assess whether viruses grouped by bnmab were also grouped by the subtype-specific plasma pools. Clustering of bnmabs, scd4 and subtype-specific plasma pools was also conducted simultaneously based on their similar neutralization profiles, shown by the dendrogram on top of the heatmap. Figure 6. Close-up structural view of the interactions between gp120 of VRC01-resistant viruses and antibody VRC01 (upper panel). The interactions are shown wheat-colored, with loop D in blue and V5 in orange. The heavy chain and light chain of antibody VRC01 are shown in surface with green and cyan colors, respectively. Residue Arg61 on the heavy chain of VRC01 that plays significant roles in binding is colored in yellow. Residues in the loop D and V5 regions of gp120 with bulky side chains predicted to interfere with the binding with VRC01 are colored red. Sequence alignment of VRC01-resistant isolates at the loop D and V5 regions is shown at lower panel. The bulky residues that may interfere with binding of VRC01 are highlighted in red. Dots represent identical residues, while dashes represent gaps introduced to preserve alignment with the reference 93TH057 sequence. 17

B.CN.RL42 C. 95IN21068 % of permuted trees 07BC_CN54 08BC_GX6F 01_AE. 90. TH.CM240 CNE15-YN8 CNE18-YN22 Nucleotide position along the HIV-1 genome Figure. 2

V1/V2 V3 V4 V5 gp120 gp41 gp41ecto No. of amino acid residues a No. of N-glycosylation sites b Figure. 3

Figure. 4

a b Figure. 5

Loop D β23 V5 β24 93TH057 ENLTNNAKT SNITGILLTRDGG----ANNTSNETFRPGGGNIKDN CNE1 K.FM...EV...L...----DDENTT.I...MR.. CNE6 S..SI.H.N.K...L...-----A.E.DLI...T..DMR.. CNE64 S.F.D.L.....L...-KNESDSNTT...MR.. CNE66 K.V.D.G.I...L..V...D----KTSNTT...Q..DMR.. CNE23...D......L..V...TNLT.ELNQT.I...E...MR.. CNE47......L..V...--QTNETNNT...MR.. Figure. 6