Differences in Allelic Frequency and CDRH3 Region Limit the ...

Download PDF
13 downloads 15 Views 3MB Size Report
Comment
Nov 1, 2016 - Limit the Engagement of HIV Env Immunogens by. Putative VRC01 ..... actions, the buried surface area between Tyr98 and gp120 is ..... 30 primer in a 25-mL reaction containing AccuPrime PFX (Life Technologies). Light chain ...
Article

Differences in Allelic Frequency and CDRH3 Region Limit the Engagement of HIV Env Immunogens by Putative VRC01 Neutralizing Antibody Precursors Graphical Abstract

Authors Christina Yacoob, Marie Pancera, Vladimir Vigdorovich, ..., D. Noah Sather, Andrew T. McGuire, Leonidas Stamatatos

Correspondence [email protected] (M.P.), [email protected] (L.S.)

In Brief Yacoob et al. interrogated the B cell receptors expressed in HIV-1 naive humans. They report that all humans are not equally predisposed to develop a particular class of protective HIV antibodies. By elucidating how the germline CDRH3 domains of such antibodies influence HIV recognition, they provide insights that may help to improve immunogen-design efforts.

Highlights d

Diverse methods used to identify and characterize putative germline VRC01 Abs

d

Not all humans are predisposed to produce such anti-HIV protective antibodies

d

The H-CDR3 domains of germline VRC01 Abs influence Env recognition

d

This information will facilitate the design of more optimal immunogens

Yacoob et al., 2016, Cell Reports 17, 1560–1570 November 1, 2016 ª 2016 The Author(s). http://dx.doi.org/10.1016/j.celrep.2016.10.017

Accession Numbers KX443243–KX443325 5TGB 5TFS 5TF1

Cell Reports

Article Differences in Allelic Frequency and CDRH3 Region Limit the Engagement of HIV Env Immunogens by Putative VRC01 Neutralizing Antibody Precursors Christina Yacoob,1,5 Marie Pancera,1,2,5,* Vladimir Vigdorovich,3 Brian G. Oliver,3 Jolene A. Glenn,1 Junli Feng,1 D. Noah Sather,3 Andrew T. McGuire,1 and Leonidas Stamatatos1,4,6,* 1Vaccine

and Infectious Disease Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, WA 98109, USA Research Center, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD 20892, USA 3Center for Infectious Disease Research, 307 Westlake Avenue North #500, Seattle, WA 98109, USA 4Department of Global Health, University of Washington, 1410 Northeast Campus Parkway, Seattle, WA 98195, USA 5Co-first author 6Lead Contact *Correspondence: [email protected] (M.P.), [email protected] (L.S.) http://dx.doi.org/10.1016/j.celrep.2016.10.017 2Vaccine

SUMMARY

Elicitation of broadly neutralizing antibodies remains a long-standing goal of HIV vaccine research. Although such antibodies can arise during HIV-1 infection, gaps in our knowledge of their germline, pre-immune precursor forms, as well as on their interaction with viral Env, limit our ability to elicit them through vaccination. Studies of broadly neutralizing antibodies from the VRC01-class provide insight into progenitor B cell receptors (BCRs) that could develop into this class of antibodies. Here, we employed high-throughput heavy chain variable region (VH)/light chain variable region (VL) deep sequencing, combined with biophysical, structural, and modeling antibody analyses, to interrogate circulating potential VRC01-progenitor BCRs in healthy individuals. Our study reveals that not all humans are equally predisposed to generate VRC01class antibodies, not all predicted progenitor VRC01-expressing B cells can bind to Env, and the CDRH3 region of germline VRC01 antibodies influence their ability to recognize HIV-1. These findings will be critical to the design of optimized immunogens that should consider CDRH3 interactions. INTRODUCTION VRC01-class antibodies target conserved elements of the CD4binding site (CD4-BS) of the HIV envelope glycoprotein (Env) and potently neutralize diverse HIV-1 viruses (broadly neutralizing antibodies [bNAbs]) (Scheid et al., 2011; Wu et al., 2010, 2011). They protect animals from experimental HIV-1- or SHIV-challenge (Balazs et al., 2014; Pietzsch et al., 2012; Shingai et al., 2014) and can transiently reduce plasma viremia during chronic HIV-1 infection (Caskey et al., 2015; Lynch et al., 2015a). Approx-

imately 29 VRC01-class bNAbs have so far been isolated from nine HIV-1-infected subjects (Zhou et al., 2015). They are highly mutated in both the variable and framework regions (Scheid et al., 2011; Wu et al., 2011), but despite a high degree of amino acid (AA) divergence, they adopt similar structures that allow them to engage the CD4-BS in a similar manner (Scharf et al., 2013; Scheid et al., 2011; Wu et al., 2011; Zhou et al., 2010, 2013, 2015). All known VRC01-class bNAbs are derived from a single heavy chain variable region (VH)1-2 allele, VH1-2*02, and from four light chain variable region (VL) genes, k1-33, k3-20, k3-15, and l2-14. The light chains (LCs) associated with VRC01-class antibodies, all have unusually short CDRL3 domains (five amino acids) (Zhou et al., 2015). Although the mutated VRC01-class antibodies recognize a wide range of recombinant Env proteins and potently neutralize diverse isolates (Scheid et al., 2011; Wu et al., 2011), their inferred germline forms do not display such properties (Hoot et al., 2013). Naturally circulating viral Envs that bind the known inferred VRC01-class antibodies have yet to be identified. The majority of the information we currently have on the interaction of the germline precursors of VRC01-class antibodies with the HIV-1 Env is derived from studies employing two specially designed recombinant Envbased proteins (eOD and 426c ‘‘core’’) and the inferred germline forms of these antibodies (Jardine et al., 2013, 2016; McGuire et al., 2013, 2016). Structural comparisons of the unliganded and liganded forms of the mutated (mu) and inferred germline (igl) antibodies revealed that the iglAbs exhibit preformed antigen-binding conformations that are similar to those of the muAbs (Scharf et al., 2016). Three amino acids are critically important for the interaction of VRC01-class antibodies with Env because of the key contacts they make with specific elements of the CD4-BS: Trp50HC contacts the conserved amino acid Asn280 in Loop D, Asn58HC contacts the conserved amino acid Arg456HC in V5, and Arg71HC contacts amino acid Asp368 in the CD4-BS (Scharf et al., 2013, 2016; West et al., 2012; Zhou et al., 2010, 2013). These three amino acids are present in the gene-encoded VH domains and remain unaltered during the extensive somatic hypermutation process that VRC01-class antibodies undergo to

1560 Cell Reports 17, 1560–1570, November 1, 2016 ª 2016 The Author(s). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Figure 1. Relative Frequencies of VH1-2*02Expressing IgM+ B Cells in HIV-1 Humans Next generation deep-sequencing (MiSeq) was used to identify unique HC sequences of circulating IgM+ B cells. (A) Percentages of VH1-2 and VH1-2*02 sequences from each of the nine donors examined are shown. The numbers represent the percent of VH1-2 sequences out of the total VH1 sequences per donor (left column) or the percentage of VH1-2 sequences derived from the *02 allele (right column). (–) indicates that VH1-2*02 sequences were not detected. The data are representative from multiple (2–4) sequencing library runs from each subject. The data shown are representative of these replicate experiments. (B) The amino acid identity of the VH1-2 HC sequences from all donors examined was compared to the published germline sequences for each of the five VH1-2 alleles. The average and the median percent identity of all VH1-2 sequences from all donors are shown. (C) Distribution of CDRH3 amino acid lengths (percent of total HCs per donor) and percent of all CDRH3 sequences expressing Trp100BHC (irrespective of CDRH3 length) for each donor are shown.

acquire their broad and potent neutralizing activities. These three amino acids are also present in the VH1-2 alleles *03 and *04 (but not in alleles *01 and *05), and thus, in principle, VRC01-class Abs can also be derived from these two alleles (West et al., 2012). Presently, it is unknown why all known VRC01-class bNAbs are derived from the *02 allele but not from the *03 or *04 alleles. In addition to the VH-encoded part of the HC of VRC01-class antibodies, the CDRH3 found in the muVRC01-class Abs contacts both the outer and inner domains of gp120 and expands the overall binding surface areas of both the mutated and inferred germline antibody forms and Env (Diskin et al., 2011; Scharf et al., 2013; Zhou et al., 2010). All known muVRC01-class bNAbs have a Trp at position 100B (Trp100BHC), located precisely five amino acids prior to framework region 4 (FWR4). Trp100BHC hydrogen bonds with Asn279 in Loop D of gp120 and its presence is necessary for the neutralizing activity of VRC01-class antibodies (West et al., 2012). Due to its location, it is presently unknown whether Trp100BHC is the result of V-D-J recombination at the ‘‘germline antibody level’’ or if it is the result of somatic hypermutation and B cell selection. igl VRC01-class Abs express the CDRH3 found in the corresponding mutated antibody forms (because CDRH3 cannot be reverted with certainty because it represents the combined effects of random V-D-J recombination, N additions, deletions, and somatic mutation). Thus, the features of CDRH3s present in germline VRC01-class antibodies and how they influence the interaction of these antibodies with Env are unknown. Possibly, a wide range of diverse CDRH3s is associated with VH1-2*02 sequences, but only a subset is conducive to Env-binding (e.g., those that express Trp100BHC). Defining the true nature of the germline CDRH3 of VRC01 antibodies will assist efforts to design more optimal immunogens to activate naive B cells that express the precursor BCRs of VRC01-class antibodies. To address this topic, we employed diverse, but complementary approaches, including next generation MiSeq VH/VL sequencing of the naive BCR repertoire of

HIV-1 negative humans, coupled to structural and modeling methods. RESULTS Allelic Frequencies of VH1-2 Sequences from Circulating IgM+ B Cells in Humans Next generation Illumina MiSeq analysis of the VH and VL of circulating IgM+ B cells was performed in nine HIV-1-uninfected individuals. Multiple VH/VL libraries were constructed from each individual and replicate (2–4) MiSeq sequencing was performed. A total of 298,368 unique VH1 sequences were identified by bioinformatics analysis. Depending on the subject, VH1-2 represented between 2% and 30% of VH1 sequences (Figure 1A). In seven of nine subjects examined, allele *02 represented between 75% and 100% of the total VH1-2 sequences. Two remaining subjects, LIB55 and LIBA, exclusively expressed allele *05 or allele *04, respectively. As discussed above, allele *04 may be suitable for the generation of VRC01 antibodies, but allele *05 is not. As expected, the vast majority of VH sequences (from the beginning of FW1 until the YYCAR sequence prior to the beginning of the CDRH3 domain) were very similar to the expected germline V gene for each allele (Figure 1B). The CDRH3 Domains The CDRH3s of most VRC01-class bNAbs range from 11–18 amino acids in length (Zhou et al., 2015), although they can be as long as 23 amino acids (Wu et al., 2015). CDRH3s longer than 19 amino acids were the minority in all subjects examined (Figures 1C and S1A). CDRH3s with Trp100BHC in LIB55 were not observed (Figures 1C and S1B). Thus, HCs derived from the VH1-2 *05 allele not only lack one of the three VH-encoded key amino acids required for Env-binding (Trp50) (Diskin et al., 2011; Klein et al., 2013; Scharf et al., 2013), they also lack Trp100BHC in the CDRH3 that plays a key role in the neutralizing activities of VRC01-class antibodies. In the remaining subjects, the frequency of CDRH3s Cell Reports 17, 1560–1570, November 1, 2016 1561

expressing Trp100BHC varied between 1.6% (LIB3) and 29.3% (LIB5) (Figures 1C and S1B). Light Chains: The Presence of Five Amino Acid-Long CDRL3 A defining feature of VRC01-class bNAbs is the presence of LCs with unusually short (five amino acid-long) CDRL3 region (Kwong and Mascola, 2012; Zhou et al., 2013, 2015). The particular angle of approach of VRC01-class bNAbs, which differs from the angle of approach of non-neutralizing CD4-BS Abs (Chen et al., 2009; Lyumkis et al., 2013), requires such short CDRL3. Thus, the short CDRL3 is presently believed to be present in the germline forms of these BCRs (Zhou et al., 2013). In these same nine subjects described above, we performed a parallel deep sequencing analysis of their circulating LC sequences (V-J) to define the frequencies of light chains with five amino acid-long CDRL3s (Figure S2). Our analysis is in agreement with previous reports (He et al., 2014; West et al., 2012) that five amino acid-long CDRL3s are extremely rare in the human B cell repertoire. Our results reveal, however, that five amino acid-long CDRL3s are more abundant in the k than the l locus. This could be the reason why only a very small fraction of known VRC01-class bNAbs express lLC (Scheid et al., 2011; Wu et al., 2011; Zhou et al., 2015). Selection and Expression of Germline VH1-2 Sequences To investigate how sequence variations in CDRH3s affect the Env-recognition properties of germline VRC01-class antibodies, we selected 83 VH1-2 sequences (from all nine subjects): 9 from the *01, 33 from the *02, 11 from the *03, 18 from the*04, and 12 from the *05 alleles. The complete amino acid sequences of these 83 HCs are presented in Figures S3A–S3F. The HCs were co-expressed with the known germline LC of VRC01-class antibodies (k3-20, k1-33, k 3-11, and l 2-14). Similar approaches were previously employed to investigate the Envbinding properties of diverse anti-HIV antibodies, including mature VRC01-class antibodies (Bonsignori et al., 2016; Wu et al., 2015; Zhu et al., 2013). The CDRH3 amino acid sequences of these HCs are shown in Figures S4A–S4D. Our selection criteria for HCs sequences were as follows: wide range of CDRH3 lengths, presence or absence of Trp100BHC, and identical or divergent sequences to iglVRC01 outside the CDRH3 region. We first attempted to co-express each VH with the k3-20 glLC (that is associated with most known VRC01-class antibodies) (Scheid et al., 2011; Wu et al., 2011; Zhou et al., 2015) (Table S1). Sixty-two HCs (74%) were expressed. We then attempted to express the remaining 21 HCs with k3-11, k1-33, or l2-14. An additional six HCs were expressed (four *04-derived HCs were expressed when paired with l2-14, while one *02 and one *03-derived HC were expressed when paired with either k3-11 or l2-14). Overall, out of the 82 HC evaluated, 67 (82%) expressed as IgG. In cases where IgG-expression did not occur, we assume that this was the result of HC/LC incompatibility. The 15 HCs that were unsuccessfully paired with LCs expressing five amino acid-long CDRL3s, most likely cannot be naturally paired with such LCs in vivo and are thus incompatible with the generation of VRC01-class antibodies. 1562 Cell Reports 17, 1560–1570, November 1, 2016

Env-Binding Recognition of Expressed Antibodies Despite the broad binding and neutralizing activities of the mutated forms of VRC01-class antibodies (i.e., as isolated from chronically infected subjects), their inferred germline forms do not display measurable Env-binding and recombinant Envs fail to activate B cells engineered to express the iglVRC01-class B cell receptors (BCRs) (Hoot et al., 2013; Jardine et al., 2013; McGuire et al., 2013). We previously discussed the modification of a clade C-derived Env (426c) that binds known iglVRC01 class-antibodies (McGuire et al., 2016). Briefly, this Env consists of the gp120 subunit with deletions of N and C termini, the first, second, and third variable domains and eliminations of three N-linked glycosylation sites: one in Loop D (N276) and two in V5 (N460 and N463). For brevity, we refer to this construct as ‘‘426c degly-3 core gp120’’ here. We first evaluated the antibodies for binding to the soluble monomeric form of 426c degly-3 core gp120 (Figure 2A). iglVRC01 was used as positive control and it was the only antibody that displayed reproducible binding. We next examined binding of these antibodies to two multimeric forms of 426c degly-3 core gp120: a Ferritin particle-based 24-meric form (Figure 2B) and a C4b-based 7-meric form (Figure 2C) (McGuire et al., 2016). Ten antibodies (seven *02, one *03, and two *04) bound the Ferritin-based Env and nine (six *02, one *03, and two *04) bound the C4b-based Env (Table S2). The apparent affinities of these antibodies varied, with similar kon rates as compared to iglVRC01, but faster koff rates (Table S3). None of the *01- and *05-derived antibodies bound these multimeric Env forms (Figure 2D), confirming that the design of 426c degly-3 core gp120 is specific for the VH1-2 alleles that express the key amino acids of VRC01-class antibodies. When the V1, V2, and V3 domains of gp120 were re-introduced, thus recreating the full gp120, but still lacking the three NLGS in Loop D and V5), two Abs (02-K and 02-O) bound the Ferritin-based Env (426c degly-3 gp120 Ferritin), and only one Ab (02-K) bound the C4b-based particles (426c degly-3 gp120 C4b) (Table S2A). Thus, in most cases, the presence of the variable domains of gp120 hinders the binding of these potential germline VRC01-class antibodies. However, iglVRC01 did bind even in the presence of V1, V2, and V3 (as long as the three NLGS in Loop D and V5 were absent). Overall, the 426c degly-3 core gp120-binding properties of the potential germline antibodies examined here resemble that of the igl form of the VRC01-class antibody 3BNC60 (McGuire et al., 2016). Differences between the Germline VRC01-Class Antibodies that Bind Env and Those that Do Not With the exception of two antibodies, all antibodies that bound 426c degly-3 core gp120 had a 12 amino acid-long CDRH3 (Table S2B). Antibody 02-K had eight amino acid-long CDRH3 while 02-O had 11 amino acid-long CDRH3. However, not every antibody with 8–12 amino acid-long CDRH3 bound 426c modified core gp120 (Figures 2 and S4A–S4D; Table S2C). With the exception of antibody 03-F, the antibodies that bound the 426c modified core gp120 express a related CDRH3 sequence (DxGxxSxW) (Table S2B) and had a Trp100BHC. Non-binding Abs did not express this exact sequence motif, although some expressed the Ser and Trp, but lacked the Asp

Figure 2. Binding of Expressed Germline Antibodies Derived from All Five VH1-2 Alleles to the 426c degly-3 Core gp120 The complete amino acid sequences of each HC expressed as IgG are shown in Figure S3. Each antibody was tested for binding to the 426c degly-3 gp120 core: (A) monomeric form, (B) Ferritin particles (24-mer), and (C) C4b particles (7-mer). The igl VRC01 IgG was included. Binding and non-binding antibodies are indicated. For clarity, the absence of interaction of antibodies derived from alleles *01 and *05 are shown separately (D). Vertical dotted lines represent the start of the dissociation phase.

and Gly (Table S2C). None of the known muVRC01class antibodies express this motif (Scheid et al., 2011; Wu et al., 2011). The Amino Acid Sequence of Germline CDRH3s Alters the Env-Recognition Properties of glVRC01 Antibodies To investigate directly whether and how differences in CDRH3 affect Env-recognition properties of germline VRC01-class antibodies, we focused on key sequences. First, we focused on 02-CB, because it has an identical amino acid sequence outside the CDRH3 as iglVRC01, it has the same length (12 amino acids) but different CDRH3 amino acid sequence than iglVRC01 (Figure 3A), and yet iglVRC01 binds better to 426c degly-3 core than 02-CB does (Figures 3B, 3C, and S5A). We surmised that the different Env-binding properties of these two antibodies are due to differences in the amino acid sequences in the CDRH3 domains. Second, we focused on 02-K and 02-J, because these two antibodies have identical sequences outside their CDRH3 domains and while their CDRH3s have the same length (eight amino acids long), their sequences differ (Figure 4A) and only 02-K binds 426c degly-3 core (Figure S5B). Third, we focused on 02-CC, because its CDRH3 is two amino acids shorter than that of 02-CB (discussed above) but the remaining portion is identical, and the two amino acids are not expected to make contact with Env (based on the available structural information), in contrast to 02-CB it does not bind the 426c modified core gp120 (Figures 5 and S5C). The apparent binding affinities of the antibodies and their mutated forms discussed below are presented in Tables S3 and S4. 02-CB There are ten amino acid differences between the iglVRC01 and 02-CB CDRH3 sequences (Figure 3A). We first focused on the ‘‘Asp-Tyr-Asn’’ (DYN) tripeptide present in iglVRC01, as it has

been shown to interact, along with Trp100BHC, with Env (both with the inner and outer gp120 domains) (Zhou et al., 2010). The corresponding sequence in *02-CB is ‘‘Ser-Ser-Gly’’ (SSG). We determined the structure of the unliganded 02-CB Fab (Figure 6) at 2.75 A˚ resolution in space group C2 (Table S5), which we then superimposed onto the known structure of iglVRC01 bound to eOD-GT6 (Jardine et al., 2013) (PDB: 4PKJ) and to the structure of NIH45-46 bound to 426c degly-3 core gp120 (Scharf et al., 2016) (PDB: 5IGX). This comparison revealed that the presence of SSG would result in the loss of one salt bridge and a loss of hydrophobic interactions. Overall the total buried surface (BSA) area between the CDRH3 of VRC01 and 02-CB are different (110 A2 for igl VRC01 to 78 A2 for 02-CB). To formally examine this point, we mutated SSG to DYN in the 02-CB background (Figures 3B and 3C). This modification improved the binding ability of 02-CB, almost to the same level as iglVRC01. An additional difference between the two CDRH3 sequences is the presence of residues ‘‘Asn97-Cys98’’ (NC) immediately preceding the DYN tripeptide in iglVRC01; that sequence is ‘‘Trp97-Tyr98’’ (WY) in 02-CB. These amino acids do not interact directly with gp120 (Zhou et al., 2010), but in the modeled interactions, the buried surface area between Tyr98 and gp120 is 85 A2 (accounting for more than half the total BSA). When the ‘‘WY’’ was replaced by ‘‘NC’’ on 02-CB, it resulted in loss (in the case of the C4b 7meric) or a reduction (in the case of the Ferritin 24meric) in binding, indicating that Tyr98 most likely plays a role in the interaction with gp120. Our unliganded model suggests a potential interaction of Tyr98 with Arg476 and/or Arg480, but this needs direct confirmation once the structure of the 02-CB/426c degly-3 core gp120 complex becomes available. The superposition of the unliganded 02-CB structure to that of iglVRC01 bound to eOD-GT6 (Figure 6) indicates a possible Cell Reports 17, 1560–1570, November 1, 2016 1563

Figure 3. The Role of the CDRH3 Sequence on the Binding of Germline 02-CB Antibody to Env (A) The CDRH3 sequences of iglVRC01 and 02-CB antibodies, as well as the reverting mutations (highlighted in orange) introduced in the 02-CB CDRH3 region are shown. Trp100BHC is highlighted in yellow. Residues that make direct contact with Env are indicated by asterisks. (B and C) Binding to the C4b particles (7-mer) (B) and Ferritin particles (24-mer) (C) expressing 426c degly-3 core gp120. Vertical dotted lines indicate the start of the dissociation phases.

clash between residue Tyr98 of 02-CB and Arg46 of eODGT6 (that is absent from the 426c degly-3 core gp120). This suggests that germline VRC01-class antibodies will interact with these two proteins slightly differently. The orientation of this Tyr in the structure of free 02-CC (see below), which has a similar sequence as 02-CB, although a shorter CDRH3, is different. In the 02-CB case, there is a symmetry interaction in the crystal that may stabilize Tyr98 in that particular orientation. When all five positions were mutated (the WYSSG sequence was replaced by NCDYN on the background of 02-CB), the antibody binding further increased (closer to iglVRC01) (Figures 3B and 3C), indicating that likely the mutations are compensatory. These results highlight the important involvement of the germline CDRH3s of VRC01-class antibodies in Env-binding. 02-K and 02-J The structure of the free 02-K Fab (Figure 6B) was solved at 2.3 A˚ in space group P22121 (Table S5) and shows the orientation of the shorter CDRH3. Its superposition with the known structure of the liganded igl VRC01 antibody (Figures 6B and 6C) reveals that the main interaction of its CDRH3 domain with Env will likely be through Trp100BHC (Figure 6). This is reminiscent to what was reported recently for VRC01c-HuGL2, a germline VRC01-class antibody isolated from an HIV-1- human using the eOD-GT8 protein as ‘‘bait’’ (Jardine et al., 2016). This antibody lacks Trp100BHC, but has a Tyr99 that is responsible for all contacts from the CDRH3 with gp120 (Figures 6B and 6C). Interestingly, 02-J has an eight amino acid-long CDRH3, which is the same length as 02-K and also expresses a 1564 Cell Reports 17, 1560–1570, November 1, 2016

Trp100BHC, but it does not bind 426c degly-3 core gp120 (Figures 4B, 4C, and S5B). Its sequence outside of CDR H3 is identical to iglVRC01. A key difference between the CDRH3s of 02-J and 02-K is the presence of a Pro in 02-J (that corresponds to a Phe in 02-K and iglVRC01) (Figure 4A). When this Pro was replaced by Phe on the 02-J background, binding was observed (Figures 4B, 4C, and S5B). Conversely, when the Phe was replaced by a Pro on the background of either 02-K or iglVRC01, binding was lost (Figure S6). These results not only confirm that the germline CDRH3 sequence influences the binding affinities of germline VRC01class antibodies with Env, but also indicate that not every circulating B cells that expresses a ‘‘VRC01-class’’ BCR (i.e., a VH1-2*02 HC paired with a LC with five amino acid-long CDRL3) will be able to bind to and become activated by Env. 02-CC 02-CC and 02-CB differ in five amino acids in the VH1-2 domain, (outside the CDRH3 domains), two of which are in the CDRH2 domain (Figure 5A), which has been shown to interact with Env (Zhou et al., 2010): 02-CC has a Gly-Gly-Asp at position (54– 56) while 02-CB (and iglVRC01) has a Ser-Gly-Gly. Asp at position 56 is expected to clash with gp120 residue 472 using the eOD-GT6 as model for superposition (PDB: 4JPK) (Jardine et al., 2013). The loop formed by residues 471–473 of gp120 adopt a different conformation in the structure with mature VRC01 (PDB: 3NGB) (Zhou et al., 2010). Because we wanted to investigate the role of CDRH3, we decided to replace GlyGly-Asp by Ser-Gly-Gly in 02-CC to match glVRC01 sequence in that region of interactions with gp120. This change resulted in 02-CC binding to 426c degly-3 core gp120 (Figures 5B, 5C, and S5C), confirming that VH1-2-encoded contacts with gp120 are important (Zhou et al., 2010). To understand the contribution of the CDRH3, the structure of 02-CC Fab was solved (Figure 6B) to 1.8 A˚ in space group P21 (Table S5). Interestingly, Trp100BHC side chain adopts a different orientation than that of glVRC01, 02-CB or 02-K in its unliganded conformation that could be due to the CDRH3 length. Superposition to the structure of iglVRC01 bound to eOD-GT6

Figure 4. The Role of the CDRH3 Sequence on the Binding of Germline 02-J Antibody to Env (A) The CDRH3 sequences of iglVRC01, 02-J, and 02-K antibodies. The reverting mutations are highlighted in orange and Trp100BHC is highlighted in yellow. (B and C) Binding to C4b particles (7-mer) (B) and Ferritin particles (24-mer) (C) expressing 426c degly-3 core gp120 are shown. Vertical dotted lines indicate the start of the dissociation phases.

revealed that the modeled interactions with gp120 will be similar to those of 02-CB (Figure 6C). However, because of the side chain rotation of Trp100BHC, the hydrogen bonding with Asn279 on Env is lost in that particular conformation, but the hydrophobic interactions appear to be unaffected. The different side chain orientation of Trp100BHC does not appear to be responsible for 02-CC lack of binding (due to residues in the CDRH2 mentioned above). It is also possible that the side chain rotamer will differ in the structure of the bound complex. DISCUSSION Broad HIV-1 neutralizing serum antibody activities become detectable 1–3 years following infection and only in a subset of infected subjects (Doria-Rose et al., 2010; Goo et al., 2014; Gray et al., 2011; Hraber et al., 2014; Mikell et al., 2011; Piantadosi et al., 2009; Sather et al., 2009; Simek et al., 2009; van Gils et al., 2009). Elucidating the underlying mechanisms of bNAb development during infection will guide more appropriate immunogen design efforts and immunization schedules (Dosenovic et al., 2015). In three well-documented cases, the initiation of bNAb-production required the emergence of viral clones ex-

pressing Env with particular features that allowed them to engage the pre-immune BCRs of particular bNAbs. A cascade of viral Env evolutionary and BCR maturation steps ensued that lead to the production of specifically mutated antibodies capable of neutralizing not just the autologous virus, but heterologous viruses as well. In one case, the bNAb (CAP256VRC26) targeted the apex of the Env spike (Doria-Rose et al., 2014). In the second and third cases, two inter-dependent CD4-BS lineages (CH103 and CH235) evolved in the same subject (Bonsignori et al., 2016; Gao et al., 2014; Liao et al., 2013). bNAbs like CAP256, CH103, and CH235 contact their respective epitopes primarily through their CDRH3 domains (Kwong and Mascola, 2012), which is the more common way by which antibodies recognize their targets. Overall, however, the underlying mechanisms associated with the development of most other types of HIV-1 bNAbs are poorly understood. This is the case for VRC01-class antibodies, which bind the CD4-BS of Env in a manner that is different from that of CH103-like antibodies (Zhou et al., 2015), display very broad neutralizing activities (Scheid et al., 2011; Wu et al., 2011), and exhibit impressive protective potential under experimental animal viral-challenge settings (Shingai et al., 2014). Samples from the early stages of infection from relevant HIV-1+ subjects are unavailable for analysis (Lynch et al., 2015b; Wu et al., 2012, 2015), and, as a result, the features of the viral Env clones that initiated the development of VRC01-class antibodies in these subjects remain unknown. In contrast to the CAP256 and CH103 cases mentioned above, a natural viral clone that can bind the inferred VRC01-class antibodies is presently unavailable. That and the fact that the inferred germline forms of VRC01-class antibodies do not bind a wide range of Envs, suggests that viral clonal variants with specific Env features initiated the process of VRC01-class antibody-production during infection in a subset of HIV-1-infected subjects. The 426c degly-3 core gp120 is one of two optimized recombinant Env-derived proteins capable of binding and activating the inferred forms of several known VRC01-class antibodies Cell Reports 17, 1560–1570, November 1, 2016 1565

Figure 5. The Role of the CDRH2 Sequence on the Binding of Germline 02-CC Antibody to Env (A) The CDRH2 (top) and CDRH3 (bottom) sequences of iglVRC01, 02-CC, and 02-CB antibodies are shown. The reverting mutations in the CDRH2 region of 02-CC at position 54 and 56 are highlighted in orange. Trp100BHC is highlighted in yellow. (B and C) Binding to C4b particles (7-mer) (B) and Ferritin particles (24-mer) (C) expressing 426c degly-3 core gp120 are shown. Vertical dotted lines indicate the start of the dissociation phases.

(Jardine et al., 2013, 2016; McGuire et al., 2013, 2016). Here, we show that 426c degly-3 core gp120 binds to several germline VH1-2*02 HCs paired with the germline LC of known VRC01 antibodies. Our study highlights the previously underappreciated, role the CDRH3 domains of glVRC01-class antibodies play during the initial step in the development of VRC01-class broadly neutralizing antibody responses. The CDRH3 may be less important in the context of the already mutated antibody, most likely because global changes resulting from affinity maturation allow the antibody to optimally engage Env. We have analyzed over 298,368 unique VH1-2*02 sequences and the closest match to the known (mature) CDRH3 of VRC01 antibodies had 58% identity. Based on our results and those published previously (DeKosky et al., 2015; He et al., 2014; Jardine et al., 2016; Lin et al., 2016), we expect that a small fraction of those sequences (0.01%–1.3%) will be paired with LCs with five amino acidlong CDRL3s to generate putative germline VRC01 class antibodies (between 60 and 7,750 depending on the subject). 1566 Cell Reports 17, 1560–1570, November 1, 2016

The results shown in this study suggest that the modified Env examined in this study will engage 8%–10% of those antibodies. Improving the ability of immunogens to engage a broader range of putative VRC01 precursor B cells, will require extensive and unbiased (i.e., not based on recognition to presently available recombinant Env proteins) isolation and characterization of numerous germline VRC01-class antibodies (i.e., VH12*02,*03, or *04 HCs naturally paired with LCs expressing five amino acidlong CDRL3s), combined with computational and modeling analyses of thousands of germline CDRH3s to identify specific ‘‘CDRH3 structural clusters,’’ around which one can redesign the presently available germline-binding immunogens. Most likely, each ‘‘CDRH3 structural cluster’’ will require a distinct set of immunogen modifications. Ninety percent of the potential germline VRC01-class antibodies studied here that bound 426c degly-3 core gp120 expressed Trp100BHC. In contrast, 30% of bona fide germline VRC01-class antibodies isolated using eOD-GT8 as ‘‘bait’’ expressed Trp100BHC (Jardine et al., 2016). The CDRH3s of all known VRC01-class bNAbs have a Trp100BHC, which hydrogen bonds with Asn/Asp279 in Loop D of gp120, but it also interacts with the LC to stabilize the antibody structure (West et al., 2012). B cells that express germline BCRs containing Trp100BHC may have advantages when encountering Env, over those that do not. Such BCRs may more rapidly develop into broadly neutralizing antibodies than the BCRs that lack Trp100BHC (that will require additional maturation steps). The presence of the inner gp120 domain on 426c degly-3 core gp120 (that is absent from eOD-GT8) potentially selects for germline BCRs that already express Trp100BHC. IGHJ2*01 is the only known human J segment that encodes Trp100BHC, and 42% of the CDRH3 sequences with a Trp100BHC were derived from that allele (50% were derived from the J4*02 allele). J5*01 is the only other J allele that encodes a Trp, but not at position 100BHC. The presence of Trp100BHC in all the VH1-2

Figure 6. Structural Analysis of CDRH3 of Germline VRC01-Class Antibodies (A) CDRH3 sequence alignment from matured VRC01 (muVRC01), chimeric antibodies 02-CB, 02-K, 02-CC, and recently isolated germline VRC01c-HuGL2 (Jardine et al., 2016). The residues that are known to interact with gp120 from previously-reported crystal structures (Jardine et al., 2013, 2016; Zhou et al., 2010) are shown in red, while the residues that interact with gp120 based on modeling analysis performed here are underlined. (B) Structures of unbound germline VRC01-class Fabs. Structures of each single chain variable fragment (Fv) are shown in ribbon representation with the CDRH3 residues shown in ‘‘sticks.’’ The superposition of five Fabs is shown on the far left panel. (C) Structures of germline VRC01-class antibodies bound to eOD-GT6 or eOD-GT8 (in the case of VRC01c-HuGL2). Superposition of the bound antibodies is shown in the far left panel. Ribbon representations of each germline antibody bound to eOD are shown separately with the residues in CDRH3 shown as sticks and residues that interact with gp120 are labeled. The structures of 02-CB, 02-K, and 02-CC antibodies bound to eOD are modeled (shown in lighter shade of gray). (D) Summary of residues in the CDRH3 of the germline antibodies that interact with gp120. Buried surface area (BSA) over 5 A2, and hydrogen bonds (H) or salt bridges (S) are indicated. Analysis for modeled interactions is italicized.

sequences analyzed here was significantly higher in the *02 allele than the *03 or *04 alleles (all of which express the remaining three key amino acids that make contact with the CD4-BS: Trp50HC, Asn58HC, and Arg71HC). This observation may explain why all known VRC01-class bNAbs are derived from the *02 than the *03 or *04 alleles (West et al., 2012). The binding of most (but not all) germline antibodies examined here was abrogated by the variable regions 1–3 on gp120. The known, mutated VRC01-class antibodies efficiently bind fulllength Env in a way that is unaffected by the presence of these variable regions (Lyumkis et al., 2013; Stewart-Jones et al., 2016). The weaker binding affinities of germline antibodies, and differences in the angles of approach of the mutated and germline VRC01 antibodies (Scharf et al., 2016), combined with the flexibility of the variable regions 1–3 on gp120, appears therefore to more drastically impact the binding of germline than mutated

VRC01-class antibodies. When soluble but stabilized trimeric Envs (Sanders et al., 2015) that can bind germline VRC01-class antibodies become available, it would be informative to determine whether such constructs are recognized by the antibodies examined here. The identification of naturally occurring, or the specific design of, Env proteins that can activate the germline precursors of VRC01-class antibodies are central to our efforts to elicit such antibodies by vaccination. Our study reveals, however, that not all humans are equally predisposed to generate VRC01-class bnAbs, either during infection or by immunization. This should be considered when designing clinical trials to evaluate the potential of immunogens to elicit broadly neutralizing antibody responses. In cases where the *02 allele is not expressed, immunogens that target non-VRC01 class antibodies should be evaluated. Cell Reports 17, 1560–1570, November 1, 2016 1567

In summary, our study improves our understanding of an important class of broadly neutralizing and protective antibodies against HIV-1. We expect this information to positively contribute to the eventual elicitation of such antibodies by vaccination.

ACCESSION NUMBERS The accession numbers for all heavy chain immunoglobulin sequences from human donors reported in this paper are NCBI GenBank: KX443243– KX443325. The accession numbers for the coordinates and structure factors for 02-CB, 02-K, and 02-CC Fabs reported in this paper are PDB: 5TGB, 5TFS, and 5TF1, respectively.

EXPERIMENTAL PROCEDURES SUPPLEMENTAL INFORMATION Sample Acquisition Whole blood samples were obtained by venipuncture from nine human subjects. All sample collection and research involving human subjects was conducted under human subjects research protocol HS029 at the Center for Infectious Disease Research. The research and methods outlined in HS029 were reviewed and approved by the Western Institutional Review Board prior to the initiation of any research. Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood by Ficoll density gradient separation. After two washes in RPMI containing 10% fetal bovine serum (FBS), PBMCs were slowly frozen in FBS (Sigma) supplemented with 10% DMSO in Mr. Frosty (Thermo Fisher Scientific) cryo-containers. Cell Preparation PBMCs from HIV donors were briefly thawed at 37 C, diluted 10-fold in RPMI, and centrifuged for 10 min at 400 3 g. Cells were washed once in PBS, centrifuged again, resuspended in the fluorescence-activated cell sorting (FACS) buffer (PBS with 2% FBS) and enumerated using a Countess II FL cell counter (Thermo Fisher Scientific). Cells were then resuspended in Buffer RLT (QIAGEN) supplemented with BME and stored at 80 C until RNA was extracted. Library Preparation for Illumina MiSeq B cell libraries for Illumina MiSeq sequencing were prepared as follows. Total RNA was extracted from 5–10 million PBMCs using the All Prep DNA/RNA Mini Kit (QIAGEN). cDNA synthesis was subsequently initiated using AB Biosystems High Capacity cDNA Reverse Transcription Kit according to the manufacturer’s protocol. For PBMC cDNA synthesis, 0.25–0.5 mg of total RNA was used. Multiplex PCR was performed with 50 ng of cDNA as a template. A 10 mM cocktail of primers corresponding to VH1, VH2, VH3, VH4, VH5, VH6, and VH7 (PMID: 21764753) (Scheid et al., 2011) was used with a 10-mM IgM 30 primer in a 25-mL reaction containing AccuPrime PFX (Life Technologies). Light chain sequences were amplified with previously described forward and reverse primer sets (Wardemann et al., 2003). PCR conditions were performed according to the manufacturer’s suggested protocol employing the following reaction conditions: 95 C for 5 min followed by 25–40 cycles with 95 C for 20 s, 55 C for 35 s, and 68 C for 35 s. Subsequent to primary cycling, a 68 C, 5 min final elongation was performed, followed by a holding cycle at 10 C. Amplicons were visualized on 1.2% Flash Gels (Lonza) in order to assess quality. Libraries of sufficient cleanliness and quality were then gel purified using 1.2% Lonza Flash Gel purification gels. Approximately 50 ng of purified library was used as a template for indexing PCR reactions, using 0.3 mL of Nextera indexing primers (Illumina) and Accumprime PFX polymerase (Life Technologies) in a volume of 25 ml. Cycling conditions are identical to those above with the following exceptions: reactions were performed with an annealing temperature of 52 C and ten cycles were performed. Indexed libraries were then precipitated, concentrated, and purified on 1.2% Flash Gels (Lonza). Purified libraries were quantitated by Qubit (Life Technologies) prior to being combined and loaded onto the MiSeq (Illumina). Statistical Analysis Samples were analyzed with non-parametric, two-way ANOVA using GraphPad Prism version 7 when indicated. Data are presented as the average of triplicates with error bars representing mean values ± SD and are representative of multiple experiments. Any additional statistical analysis methods are indicated in the corresponding methods.

1568 Cell Reports 17, 1560–1570, November 1, 2016

Supplemental Information includes Supplemental Experimental Procedures, six figures, and five tables and can be found with this article online at http:// dx.doi.org/10.1016/j.celrep.2016.10.017. AUTHOR CONTRIBUTIONS C.Y. and M.P. designed and performed experiments and analyzed the data. L.S. conceived the study and participated in the experimental design and in the analysis and the interpretation of data. J.A.G. and J.F. performed experiments. A.T.M. contributed to the design of experiments and the analysis of data. D.N.S., B.G.O., and V.V. performed MiSeq experiments and analyzed those data. L.S., C.Y., and M.P. wrote the manuscript, with contributions of all other authors. ACKNOWLEDGMENTS This study was supported by NIH grants R01 AI104384 and R01 AI081625 (L.S.). We acknowledge many helpful discussions with Pamela Bjorkman and Anthony West on the topics discussed here. We also acknowledge the support by the J.B. Pendleton Charitable Trust. Received: July 13, 2016 Revised: September 14, 2016 Accepted: October 5, 2016 Published: November 1, 2016 REFERENCES Balazs, A.B., Ouyang, Y., Hong, C.M., Chen, J., Nguyen, S.M., Rao, D.S., An, D.S., and Baltimore, D. (2014). Vectored immunoprophylaxis protects humanized mice from mucosal HIV transmission. Nat. Med. 20, 296–300. Bonsignori, M., Zhou, T., Sheng, Z., Chen, L., Gao, F., Joyce, M.G., Ozorowski, G., Chuang, G.Y., Schramm, C.A., Wiehe, K., et al.; NISC Comparative Sequencing Program (2016). Maturation pathway from germline to broad HIV-1 neutralizer of a CD4-mimic antibody. Cell 165, 449–463. Caskey, M., Klein, F., Lorenzi, J.C., Seaman, M.S., West, A.P., Jr., Buckley, N., Kremer, G., Nogueira, L., Braunschweig, M., Scheid, J.F., et al. (2015). Viraemia suppressed in HIV-1-infected humans by broadly neutralizing antibody 3BNC117. Nature 522, 487–491. Chen, L., Kwon, Y.D., Zhou, T., Wu, X., O’Dell, S., Cavacini, L., Hessell, A.J., Pancera, M., Tang, M., Xu, L., et al. (2009). Structural basis of immune evasion at the site of CD4 attachment on HIV-1 gp120. Science 326, 1123–1127. DeKosky, B.J., Kojima, T., Rodin, A., Charab, W., Ippolito, G.C., Ellington, A.D., and Georgiou, G. (2015). In-depth determination and analysis of the human paired heavy- and light-chain antibody repertoire. Nat. Med. 21, 86–91. Diskin, R., Scheid, J.F., Marcovecchio, P.M., West, A.P., Jr., Klein, F., Gao, H., Gnanapragasam, P.N., Abadir, A., Seaman, M.S., Nussenzweig, M.C., and Bjorkman, P.J. (2011). Increasing the potency and breadth of an HIV antibody by using structure-based rational design. Science 334, 1289–1293. Doria-Rose, N.A., Klein, R.M., Daniels, M.G., O’Dell, S., Nason, M., Lapedes, A., Bhattacharya, T., Migueles, S.A., Wyatt, R.T., Korber, B.T., et al. (2010). Breadth of human immunodeficiency virus-specific neutralizing activity in sera: clustering analysis and association with clinical variables. J. Virol. 84, 1631–1636.

Doria-Rose, N.A., Schramm, C.A., Gorman, J., Moore, P.L., Bhiman, J.N., DeKosky, B.J., Ernandes, M.J., Georgiev, I.S., Kim, H.J., Pancera, M., et al.; NISC Comparative Sequencing Program (2014). Developmental pathway for potent V1V2-directed HIV-neutralizing antibodies. Nature 509, 55–62.

McGuire, A.T., Hoot, S., Dreyer, A.M., Lippy, A., Stuart, A., Cohen, K.W., Jardine, J., Menis, S., Scheid, J.F., West, A.P., et al. (2013). Engineering HIV envelope protein to activate germline B cell receptors of broadly neutralizing anti-CD4 binding site antibodies. J. Exp. Med. 210, 655–663.

Dosenovic, P., von Boehmer, L., Escolano, A., Jardine, J., Freund, N.T., Gitlin, A.D., McGuire, A.T., Kulp, D.W., Oliveira, T., Scharf, L., et al. (2015). Immunization for HIV-1 broadly neutralizing antibodies in human Ig knockin mice. Cell 161, 1505–1515.

McGuire, A.T., Gray, M.D., Dosenovic, P., Gitlin, A.D., Freund, N.T., Petersen, J., Correnti, C., Johnsen, W., Kegel, R., Stuart, A.B., et al. (2016). Specifically modified Env immunogens activate B-cell precursors of broadly neutralizing HIV-1 antibodies in transgenic mice. Nat. Commun. 7, 10618.

Gao, F., Bonsignori, M., Liao, H.X., Kumar, A., Xia, S.M., Lu, X., Cai, F., Hwang, K.K., Song, H., Zhou, T., et al. (2014). Cooperation of B cell lineages in induction of HIV-1-broadly neutralizing antibodies. Cell 158, 481–491.

Mikell, I., Sather, D.N., Kalams, S.A., Altfeld, M., Alter, G., and Stamatatos, L. (2011). Characteristics of the earliest cross-neutralizing antibody response to HIV-1. PLoS Pathog. 7, e1001251.

Goo, L., Chohan, V., Nduati, R., and Overbaugh, J. (2014). Early development of broadly neutralizing antibodies in HIV-1-infected infants. Nat. Med. 20, 655–658.

Piantadosi, A., Panteleeff, D., Blish, C.A., Baeten, J.M., Jaoko, W., McClelland, R.S., and Overbaugh, J. (2009). Breadth of neutralizing antibody response to human immunodeficiency virus type 1 is affected by factors early in infection but does not influence disease progression. J. Virol. 83, 10269–10274.

Gray, E.S., Madiga, M.C., Hermanus, T., Moore, P.L., Wibmer, C.K., Tumba, N.L., Werner, L., Mlisana, K., Sibeko, S., Williamson, C., et al.; CAPRISA002 Study Team (2011). The neutralization breadth of HIV-1 develops incrementally over four years and is associated with CD4+ T cell decline and high viral load during acute infection. J. Virol. 85, 4828–4840. He, L., Sok, D., Azadnia, P., Hsueh, J., Landais, E., Simek, M., Koff, W.C., Poignard, P., Burton, D.R., and Zhu, J. (2014). Toward a more accurate view of human B-cell repertoire by next-generation sequencing, unbiased repertoire capture and single-molecule barcoding. Sci. Rep. 4, 6778. Hoot, S., McGuire, A.T., Cohen, K.W., Strong, R.K., Hangartner, L., Klein, F., Diskin, R., Scheid, J.F., Sather, D.N., Burton, D.R., and Stamatatos, L. (2013). Recombinant HIV envelope proteins fail to engage germline versions of anti-CD4bs bNAbs. PLoS Pathog. 9, e1003106. Hraber, P., Seaman, M.S., Bailer, R.T., Mascola, J.R., Montefiori, D.C., and Korber, B.T. (2014). Prevalence of broadly neutralizing antibody responses during chronic HIV-1 infection. AIDS 28, 163–169. Jardine, J., Julien, J.P., Menis, S., Ota, T., Kalyuzhniy, O., McGuire, A., Sok, D., Huang, P.S., MacPherson, S., Jones, M., et al. (2013). Rational HIV immunogen design to target specific germline B cell receptors. Science 340, 711–716. Jardine, J.G., Kulp, D.W., Havenar-Daughton, C., Sarkar, A., Briney, B., Sok, D., Sesterhenn, F., Eren˜o-Orbea, J., Kalyuzhniy, O., Deresa, I., et al. (2016). HIV-1 broadly neutralizing antibody precursor B cells revealed by germline-targeting immunogen. Science 351, 1458–1463. Klein, F., Diskin, R., Scheid, J.F., Gaebler, C., Mouquet, H., Georgiev, I.S., Pancera, M., Zhou, T., Incesu, R.B., Fu, B.Z., et al. (2013). Somatic mutations of the immunoglobulin framework are generally required for broad and potent HIV-1 neutralization. Cell 153, 126–138. Kwong, P.D., and Mascola, J.R. (2012). Human antibodies that neutralize HIV-1: identification, structures, and B cell ontogenies. Immunity 37, 412–425. Liao, H.X., Lynch, R., Zhou, T., Gao, F., Alam, S.M., Boyd, S.D., Fire, A.Z., Roskin, K.M., Schramm, C.A., Zhang, Z., et al.; NISC Comparative Sequencing Program (2013). Co-evolution of a broadly neutralizing HIV-1 antibody and founder virus. Nature 496, 469–476. Lin, S.G., Ba, Z., Du, Z., Zhang, Y., Hu, J., and Alt, F.W. (2016). Highly sensitive and unbiased approach for elucidating antibody repertoires. Proc. Natl. Acad. Sci. USA 113, 7846–7851. Lynch, R.M., Boritz, E., Coates, E.E., DeZure, A., Madden, P., Costner, P., Enama, M.E., Plummer, S., Holman, L., Hendel, C.S., et al.; VRC 601 Study Team (2015a). Virologic effects of broadly neutralizing antibody VRC01 administration during chronic HIV-1 infection. Sci. Transl. Med. 7, 319ra206.

Pietzsch, J., Gruell, H., Bournazos, S., Donovan, B.M., Klein, F., Diskin, R., Seaman, M.S., Bjorkman, P.J., Ravetch, J.V., Ploss, A., and Nussenzweig, M.C. (2012). A mouse model for HIV-1 entry. Proc. Natl. Acad. Sci. USA 109, 15859–15864. Sanders, R.W., van Gils, M.J., Derking, R., Sok, D., Ketas, T.J., Burger, J.A., Ozorowski, G., Cupo, A., Simonich, C., Goo, L., et al. (2015). HIV-1 VACCINES. HIV-1 neutralizing antibodies induced by native-like envelope trimers. Science 349, aac4223. Sather, D.N., Armann, J., Ching, L.K., Mavrantoni, A., Sellhorn, G., Caldwell, Z., Yu, X., Wood, B., Self, S., Kalams, S., and Stamatatos, L. (2009). Factors associated with the development of cross-reactive neutralizing antibodies during human immunodeficiency virus type 1 infection. J. Virol. 83, 757–769. Scharf, L., West, A.P., Jr., Gao, H., Lee, T., Scheid, J.F., Nussenzweig, M.C., Bjorkman, P.J., and Diskin, R. (2013). Structural basis for HIV-1 gp120 recognition by a germ-line version of a broadly neutralizing antibody. Proc. Natl. Acad. Sci. USA 110, 6049–6054. Scharf, L., West, A.P., Sievers, S.A., Chen, C., Jiang, S., Gao, H., Gray, M.D., McGuire, A.T., Scheid, J.F., Nussenzweig, M.C., et al. (2016). Structural basis for germline antibody recognition of HIV-1 immunogens. eLife 5, e13783. Scheid, J.F., Mouquet, H., Ueberheide, B., Diskin, R., Klein, F., Oliveira, T.Y., Pietzsch, J., Fenyo, D., Abadir, A., Velinzon, K., et al. (2011). Sequence and structural convergence of broad and potent HIV antibodies that mimic CD4 binding. Science 333, 1633–1637. Shingai, M., Donau, O.K., Plishka, R.J., Buckler-White, A., Mascola, J.R., Nabel, G.J., Nason, M.C., Montefiori, D., Moldt, B., Poignard, P., et al. (2014). Passive transfer of modest titers of potent and broadly neutralizing anti-HIV monoclonal antibodies block SHIV infection in macaques. J. Exp. Med. 211, 2061–2074. Simek, M.D., Rida, W., Priddy, F.H., Pung, P., Carrow, E., Laufer, D.S., Lehrman, J.K., Boaz, M., Tarragona-Fiol, T., Miiro, G., et al. (2009). Human immunodeficiency virus type 1 elite neutralizers: individuals with broad and potent neutralizing activity identified by using a high-throughput neutralization assay together with an analytical selection algorithm. J. Virol. 83, 7337–7348. Stewart-Jones, G.B., Soto, C., Lemmin, T., Chuang, G.Y., Druz, A., Kong, R., Thomas, P.V., Wagh, K., Zhou, T., Behrens, A.J., et al. (2016). Trimeric HIV-1Env structures define glycan shields from clades A, B, and G. Cell 165, 813–826. van Gils, M.J., Euler, Z., Schweighardt, B., Wrin, T., and Schuitemaker, H. (2009). Prevalence of cross-reactive HIV-1-neutralizing activity in HIV-1-infected patients with rapid or slow disease progression. AIDS 23, 2405–2414.

Lynch, R.M., Wong, P., Tran, L., O’Dell, S., Nason, M.C., Li, Y., Wu, X., and Mascola, J.R. (2015b). HIV-1 fitness cost associated with escape from the VRC01 class of CD4 binding site neutralizing antibodies. J. Virol. 89, 4201– 4213.

Wardemann, H., Yurasov, S., Schaefer, A., Young, J.W., Meffre, E., and Nussenzweig, M.C. (2003). Predominant autoantibody production by early human B cell precursors. Science 301, 1374–1377.

Lyumkis, D., Julien, J.P., de Val, N., Cupo, A., Potter, C.S., Klasse, P.J., Burton, D.R., Sanders, R.W., Moore, J.P., Carragher, B., et al. (2013). Cryo-EM structure of a fully glycosylated soluble cleaved HIV-1 envelope trimer. Science 342, 1484–1490.

West, A.P., Jr., Diskin, R., Nussenzweig, M.C., and Bjorkman, P.J. (2012). Structural basis for germ-line gene usage of a potent class of antibodies targeting the CD4-binding site of HIV-1 gp120. Proc. Natl. Acad. Sci. USA 109, E2083–E2090.

Cell Reports 17, 1560–1570, November 1, 2016 1569

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., et al. (2010). Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1. Science 329, 856–861. Wu, X., Zhou, T., Zhu, J., Zhang, B., Georgiev, I., Wang, C., Chen, X., Longo, N.S., Louder, M., McKee, K., et al.; NISC Comparative Sequencing Program (2011). Focused evolution of HIV-1 neutralizing antibodies revealed by structures and deep sequencing. Science 333, 1593–1602. Wu, X., Wang, C., O’Dell, S., Li, Y., Keele, B.F., Yang, Z., Imamichi, H., DoriaRose, N., Hoxie, J.A., Connors, M., et al. (2012). Selection pressure on HIV-1 envelope by broadly neutralizing antibodies to the conserved CD4-binding site. J. Virol. 86, 5844–5856. Wu, X., Zhang, Z., Schramm, C.A., Joyce, M.G., Kwon, Y.D., Zhou, T., Sheng, Z., Zhang, B., O’Dell, S., McKee, K., et al.; NISC Comparative Sequencing Program (2015). Maturation and diversity of the VRC01-antibody lineage over 15 years of chronic HIV-1 infection. Cell 161, 470–485.

1570 Cell Reports 17, 1560–1570, November 1, 2016

Zhou, T., Georgiev, I., Wu, X., Yang, Z.Y., Dai, K., Finzi, A., Kwon, Y.D., Scheid, J.F., Shi, W., Xu, L., et al. (2010). Structural basis for broad and potent neutralization of HIV-1 by antibody VRC01. Science 329, 811–817. Zhou, T., Zhu, J., Wu, X., Moquin, S., Zhang, B., Acharya, P., Georgiev, I.S., Altae-Tran, H.R., Chuang, G.Y., Joyce, M.G., et al.; NISC Comparative Sequencing Program (2013). Multidonor analysis reveals structural elements, genetic determinants, and maturation pathway for HIV-1 neutralization by VRC01-class antibodies. Immunity 39, 245–258. Zhou, T., Lynch, R.M., Chen, L., Acharya, P., Wu, X., Doria-Rose, N.A., Joyce, M.G., Lingwood, D., Soto, C., Bailer, R.T., et al.; NISC Comparative Sequencing Program (2015). Structural repertoire of HIV-1-neutralizing antibodies targeting the CD4 supersite in 14 donors. Cell 161, 1280–1292. Zhu, J., Wu, X., Zhang, B., McKee, K., O’Dell, S., Soto, C., Zhou, T., Casazza, J.P., Mullikin, J.C., Kwong, P.D., et al.; NISC Comparative Sequencing Program (2013). De novo identification of VRC01 class HIV-1-neutralizing antibodies by next-generation sequencing of B-cell transcripts. Proc. Natl. Acad. Sci. USA 110, E4088–E4097.