Kevin Wiehe

Co-evolution of a broadly neutralizing HIV-1 antibody and founder virus

Current human immunodeficiency virus-1 (HIV-1) vaccines elicit strain-specific neutralizing antibodies. However, cross-reactive neutralizing antibodies arise in approximately 20% of HIV-1-infected individuals, and details of their generation could provide a blueprint for effective vaccination. Here we report the isolation, evolution and structure of a broadly neutralizing antibody from an African donor followed from the time of infection. The mature antibody, CH103, neutralized approximately 55% of HIV-1 isolates, and its co-crystal structure with the HIV-1 envelope protein gp120 revealed a new loop-based mechanism of CD4-binding-site recognition. Virus and antibody gene sequencing revealed concomitant virus evolution and antibody maturation. Notably, the unmutated common ancestor of the CH103 lineage avidly bound the transmitted/founder HIV-1 envelope glycoprotein, and evolution of antibody neutralization breadth was preceded by extensive viral diversification in and near the CH103 epitope. These data determine the viral and antibody evolution leading to induction of a lineage of HIV-1 broadly neutralizing antibodies, and provide insights into strategies to elicit similar antibodies by vaccination.

ZDOCK server: interactive docking prediction of protein-protein complexes and symmetric multimers.

SUMMARY Protein-protein interactions are essential to cellular and immune function, and in many cases, because of the absence of an experimentally determined structure of the complex, these interactions must be modeled to obtain an understanding of their molecular basis. We present a user-friendly protein docking server, based on the rigid-body docking programs ZDOCK and M-ZDOCK, to predict structures of protein-protein complexes and symmetric multimers. With a goal of providing an accessible and intuitive interface, we provide options for users to guide the scoring and the selection of output models, in addition to dynamic visualization of input structures and output docking models. This server enables the research community to easily and quickly produce structural models of protein-protein complexes and symmetric multimers for their own analysis. AVAILABILITY The ZDOCK server is freely available to all academic and non-profit users at: http://zdock.umassmed.edu. No registration is required.

Protein–protein docking benchmark 2.0: An update

We present a new version of the Protein–Protein Docking Benchmark, reconstructed from the bottom up to include more complexes, particularly focusing on more unbound–unbound test cases. SCOP (Structural Classification of Proteins) was used to assess redundancy between the complexes in this version. The new benchmark consists of 72 unbound–unbound cases, with 52 rigid‐body cases, 13 medium‐difficulty cases, and 7 high‐difficulty cases with substantial conformational change. In addition, we retained 12 antibody–antigen test cases with the antibody structure in the bound form. The new benchmark provides a platform for evaluating the progress of docking methods on a wide variety of targets. The new version of the benchmark is available to the public at http://zlab.bu.edu/benchmark2. Proteins 2005;60:214–216.

Integrating statistical pair potentials into protein complex prediction

The biophysical study of protein–protein interactions and docking has important implications in our understanding of most complex cellular signaling processes. Most computational approaches to protein docking involve a tradeoff between the level of detail incorporated into the model and computational power required to properly handle that level of detail. In this work, we seek to optimize that balance by showing that we can reduce the complexity of model representation and thus make the computation tractable with minimal loss of predictive performance. We also introduce a pair‐wise statistical potential suitable for docking that builds on previous work and show that this potential can be incorporated into our fast fourier transform‐based docking algorithm ZDOCK. We use the Protein Docking Benchmark to illustrate the improved performance of this potential compared with less detailed other scoring functions. Furthermore, we show that the new potential performs well on antibody‐antigen complexes, with most predictions clustering around the Complementarity Determining Regions of antibodies without any manual intervention. Proteins 2007.

Maturation Pathway from Germline to Broad HIV-1 Neutralizer of a CD4-Mimic Antibody

Antibodies with ontogenies from VH1-2 or VH1-46-germline genes dominate the broadly neutralizing response against the CD4-binding site (CD4bs) on HIV-1. Here, we define with longitudinal sampling from time-of-infection the development of a VH1-46-derived antibody lineage that matured to neutralize 90% of HIV-1 isolates. Structures of lineage antibodies CH235 (week 41 from time-of-infection, 18% breadth), CH235.9 (week 152, 77%), and CH235.12 (week 323, 90%) demonstrated the maturing epitope to focus on the conformationally invariant portion of the CD4bs. Similarities between CH235 lineage and five unrelated CD4bs lineages in epitope focusing, length-of-time to develop breadth, and extraordinary level of somatic hypermutation suggested commonalities in maturation among all CD4bs antibodies. Fortunately, the required CH235-lineage hypermutation appeared substantially guided by the intrinsic mutability of the VH1-46 gene, which closely resembled VH1-2. We integrated our CH235-lineage findings with a second broadly neutralizing lineage and HIV-1 co-evolution to suggest a vaccination strategy for inducing both lineages.

ZDOCK and RDOCK performance in CAPRI rounds 3, 4, and 5

We present an evaluation of the results of our ZDOCK and RDOCK algorithms in Rounds 3, 4, and 5 of the protein docking challenge CAPRI. ZDOCK is a Fast Fourier Transform (FFT)‐based, initial‐stage rigid‐body docking algorithm, and RDOCK is an energy minimization algorithm for refining and reranking ZDOCK results. Of the 9 targets for which we submitted predictions, we attained at least acceptable accuracy for 7, at least medium accuracy for 6, and high accuracy for 3. These results are evidence that ZDOCK in combination with RDOCK is capable of making accurate predictions on a diverse set of protein complexes. Proteins 2005;60:207–213.

Diversion of HIV-1 vaccine–induced immunity by gp41-microbiota cross-reactive antibodies

Microbiota can mislead antibodies Unlike the response to many viral infections, most people do not produce antibodies capable of clearing HIV-1. Non-neutralizing antibodies that target HIV-1s envelope glycoprotein (Env) typically dominate the response, which is generated by B cells that cross-react with Env and the intestinal microbiota. Williams et al. analyzed samples from individuals who had received a vaccine containing the Env protein, including the gp41 subunit. Most of the antibodies were non-neutralizing and targeted gp41. The antibodies also reacted to intestinal microbiota, suggesting that preexisting immunity to microbial communities skews vaccineinduced immune responses toward an unproductive target. Science, this issue 10.1126/science.aab1253. The antibody response to an HIV-1 vaccine is dominated by preexisting immunity to microbiota. INTRODUCTION Inducing protective antibodies is a key goal in HIV-1 vaccine development. In acute HIV-1 infection, the dominant initial plasma antibody response is to the gp41 subunit of the envelope (Env) glycoprotein of the virus. These antibodies derive from polyreactive B cells that cross-react with Env and intestinal microbiota (IM) and are unable to neutralize HIV-1. However, whether a similar gp41-IM cross-reactive antibody response would occur in the setting of HIV-1 Env vaccination is unknown. RATIONALE We studied antibody responses in individuals who received a DNA prime vaccine, with a recombinant adenovirus serotype 5 (rAd5) boost (DNA prime–rAd5 boost), a vaccine that included HIV-1 gag, pol, and nef genes, as well as a trivalent mixture of clade A, B, and C env gp140 genes containing both gp120 and gp41 components. This vaccine showed no efficacy. Thus, study of these vaccinees provided an opportunity to determine whether the Env-reactive antibody response in the setting of Env vaccination was dominated by gp41-reactive antibodies derived from Env-IM cross-reactive B cells. RESULTS We found that vaccine-induced antibodies to HIV-1 Env dominantly focused on gp41 compared with gp120 by both serologic analysis and by vaccine-Env memory B cells sorted by flow cytometry (see the figure). Remarkably, the majority of HIV-1 Env-reactive memory B cells induced by the vaccine produced gp41-reactive antibodies, and the majority of gp41-targeted antibodies used restricted immunoglobulin heavy chain variable genes. Functionally, none of the gp41-reactive antibodies could neutralize HIV, and the majority could not mediate antibody-dependent cellular cytotoxicity. Most of the vaccine-induced gp41-reactive antibodies cross-reacted with host and IM antigens. Two of the candidate gp41-intestinal cross-reactive antigens were bacterial RNA polymerase and pyruvate-flavodoxin oxidoreductase, which shared sequence similarities with the heptad repeat 1 region of HIV gp41. Next-generation sequencing of vaccinee B cells demonstrated a prevaccination antibody that was reactive to both IM and the vaccine–Env gp140, which demonstrated the presence of a preexisting pool of gp41-IM cross-reactive B cells from which the vaccine gp41-reactive antibody response was derived. CONCLUSION In this study, we found that the DNA prime–rAd5 boost HIV-1 vaccine induced a gp41-reactive antibody response that was mainly non-neutralizing and derived from an IM-gp41 cross-reactive B cell pool. These findings have important implications for HIV-1 vaccine design. Because IM antigens shape the B cell repertoire from birth, our data raise the hypothesis that neonatal immunization with HIV-1 envelope may be able to imprint the B cell repertoire to respond to envelope antigenic sites that may otherwise be subdominant or disfavored, such as Env broadly neutralizing antibody epitopes. Our data also suggest that deleting or modifying amino acids in the gp41 heptad repeat 1 region of Env-containing vaccine immunogens may avoid IM-gp41 cross-reactivity. Thus, an obstacle that may need to be overcome for development of a successful HIV vaccine is diversion of potentially protective HIV-1 antibody responses by preexisting envelope-IM cross-reactive pools of B cells. Diversion of HIV-1 vaccine–induced immunity by Env gp41–microbiota cross-reactive antibodies. Immunization of humans with a vaccine containing HIV-1 Env gp120 and gp41 components, including the membrane-proximal external region (MPER) of Env, induced a dominant B cell response primarily from a preexisting pool of gp41-IM cross-reactive B cells. This response diverted the vaccine-stimulated antibody response away from smaller subdominant B cell pools capable of reacting with potentially protective epitopes on HIV-1 Env. An HIV-1 DNA prime vaccine, with a recombinant adenovirus type 5 (rAd5) boost, failed to protect from HIV-1 acquisition. We studied the nature of the vaccine-induced antibody (Ab) response to HIV-1 envelope (Env). HIV-1–reactive plasma Ab titers were higher to Env gp41 than to gp120, and repertoire analysis demonstrated that 93% of HIV-1–reactive Abs from memory B cells responded to Env gp41. Vaccine-induced gp41-reactive monoclonal antibodies were non-neutralizing and frequently polyreactive with host and environmental antigens, including intestinal microbiota (IM). Next-generation sequencing of an immunoglobulin heavy chain variable region repertoire before vaccination revealed an Env-IM cross-reactive Ab that was clonally related to a subsequent vaccine-induced gp41-reactive Ab. Thus, HIV-1 Env DNA-rAd5 vaccine induced a dominant IM-polyreactive, non-neutralizing gp41-reactive Ab repertoire response that was associated with no vaccine efficacy.

Polyreactivity and Autoreactivity among HIV-1 Antibodies

ABSTRACT It is generally acknowledged that human broadly neutralizing antibodies (bNAbs) capable of neutralizing multiple HIV-1 clades are often polyreactive or autoreactive. Whereas polyreactivity or autoreactivity has been proposed to be crucial for neutralization breadth, no systematic, quantitative study of self-reactivity among nonneutralizing HIV-1 Abs (nNAbs) has been performed to determine whether poly- or autoreactivity in bNAbs is a consequence of chronic antigen (Ag) exposure and/or inflammation or a fundamental property of neutralization. Here, we use protein microarrays to assess binding to >9,400 human proteins and find that as a class, bNAbs are significantly more poly- and autoreactive than nNAbs. The poly- and autoreactive property is therefore not due to the infection milieu but rather is associated with neutralization. Our observations are consistent with a role of heteroligation for HIV-1 neutralization and/or structural mimicry of host Ags by conserved HIV-1 neutralization sites. Although bNAbs are more mutated than nNAbs as a group, V(D)J mutation per se does not correlate with poly- and autoreactivity. Infrequent poly- or autoreactivity among nNAbs implies that their dominance in humoral responses is due to the absence of negative control by immune regulation. Interestingly, four of nine bNAbs specific for the HIV-1 CD4 binding site (CD4bs) (VRC01, VRC02, CH106, and CH103) bind human ubiquitin ligase E3A (UBE3A), and UBE3A protein competitively inhibits gp120 binding to the VRC01 bNAb. Among these four bNAbs, avidity for UBE3A was correlated with neutralization breadth. Identification of UBE3A as a self-antigen recognized by CD4bs bNAbs offers a mechanism for the rarity of this bNAb class. IMPORTANCE Eliciting bNAbs is key for HIV-1 vaccines; most Abs elicited by HIV-1 infection or immunization, however, are strain specific or nonneutralizing, and unsuited for protection. Here, we compare the specificities of bNAbs and nNAbs to demonstrate that bNAbs are significantly more poly- and autoreactive than nNAbs. The strong association of poly- and autoreactivity with bNAbs, but not nNAbs from infected patients, indicates that the infection milieu, chronic inflammation and Ag exposure, CD4 T-cell depletion, etc., alone does not cause poly- and autoreactivity. Instead, these properties are fundamentally linked to neutralization breadth, either by the requirement for heteroligation or the consequence of host mimicry by HIV-1. Indeed, we show that human UBE3A shares an epitope(s) with HIV-1 envelope recognized by four CD4bs bNAbs. The poly- and autoreactivity of bNAbs surely contribute to the rarity of membrane-proximal external region (MPER) and CD4bs bNAbs and identify a roadblock that must be overcome to induce protective vaccines.

The performance of ZDOCK and ZRANK in rounds 6–11 of CAPRI

We present an evaluation of our protein–protein docking approach using the ZDOCK and ZRANK algorithms, in combination with structural clustering and filtering, utilizing biological data in Rounds 6–11 of the CAPRI docking experiment. We achieved at least one prediction of acceptable accuracy for five of six targets submitted. In addition, two targets resulted in medium‐accuracy predictions. In the new scoring portion of the CAPRI exercise, we were able to attain at least one acceptable prediction for the three targets submitted and achieved three medium‐accuracy predictions for Target 26. Scoring was performed using ZRANK, a new algorithm for reranking initial‐stage docking predictions using a weighted energy function and no structural refinement. Here we outline a practical and successful docking strategy, given limited prior biological knowledge of the complex to be predicted. Proteins 2007.

Reconstructing a B-Cell Clonal Lineage. II. Mutation, Selection, and Affinity Maturation

Affinity maturation of the antibody response is a fundamental process in adaptive immunity during which B-cells activated by infection or vaccination undergo rapid proliferation accompanied by the acquisition of point mutations in their rearranged immunoglobulin (Ig) genes and selection for increased affinity for the eliciting antigen. The rate of somatic hypermutation at any position within an Ig gene is known to depend strongly on the local DNA sequence, and Ig genes have region-specific codon biases that influence the local mutation rate within the gene resulting in increased differential mutability in the regions that encode the antigen-binding domains. We have isolated a set of clonally related natural Ig heavy chain–light chain pairs from an experimentally infected influenza patient, inferred the unmutated ancestral rearrangements and the maturation intermediates, and synthesized all the antibodies using recombinant methods. The lineage exhibits a remarkably uniform rate of improvement of the effective affinity to influenza hemagglutinin (HA) over evolutionary time, increasing 1000-fold overall from the unmutated ancestor to the best of the observed antibodies. Furthermore, analysis of selection reveals that selection and mutation bias were concordant even at the level of maturation to a single antigen. Substantial improvement in affinity to HA occurred along mutationally preferred paths in sequence space and was thus strongly facilitated by the underlying local codon biases.