Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Nov 8;55(11):2149-2167.e9.
doi: 10.1016/j.immuni.2022.09.001. Epub 2022 Sep 29.

Human immunoglobulin repertoire analysis guides design of vaccine priming immunogens targeting HIV V2-apex broadly neutralizing antibody precursors

Jordan R Willis  1 , Zachary T Berndsen  2 , Krystal M Ma  1 , Jon M Steichen  1 , Torben Schiffner  1 , Elise Landais  1 , Alessia Liguori  1 , Oleksandr Kalyuzhniy  1 , Joel D Allen  3 , Sabyasachi Baboo  4 , Oluwarotimi Omorodion  2 , Jolene K Diedrich  3 , Xiaozhen Hu  1 , Erik Georgeson  1 , Nicole Phelps  1 , Saman Eskandarzadeh  1 , Bettina Groschel  1 , Michael Kubitz  1 , Yumiko Adachi  1 , Tina-Marie Mullin  1 , Nushin B Alavi  1 , Samantha Falcone  5 , Sunny Himansu  5 , Andrea Carfi  5 , Ian A Wilson  2 , John R Yates 3rd  4 , James C Paulson  6 , Max Crispin  3 , Andrew B Ward  2 , William R Schief  7
Affiliations

Human immunoglobulin repertoire analysis guides design of vaccine priming immunogens targeting HIV V2-apex broadly neutralizing antibody precursors

Jordan R Willis et al. Immunity. .

Abstract

Broadly neutralizing antibodies (bnAbs) to the HIV envelope (Env) V2-apex region are important leads for HIV vaccine design. Most V2-apex bnAbs engage Env with an uncommonly long heavy-chain complementarity-determining region 3 (HCDR3), suggesting that the rarity of bnAb precursors poses a challenge for vaccine priming. We created precursor sequence definitions for V2-apex HCDR3-dependent bnAbs and searched for related precursors in human antibody heavy-chain ultradeep sequencing data from 14 HIV-unexposed donors. We found potential precursors in a majority of donors for only two long-HCDR3 V2-apex bnAbs, PCT64 and PG9, identifying these bnAbs as priority vaccine targets. We then engineered ApexGT Env trimers that bound inferred germlines for PCT64 and PG9 and had higher affinities for bnAbs, determined cryo-EM structures of ApexGT trimers complexed with inferred-germline and bnAb forms of PCT64 and PG9, and developed an mRNA-encoded cell-surface ApexGT trimer. These methods and immunogens have promise to assist HIV vaccine development.

Keywords: AIDS vaccines; HIV antibodies; germline targeting; immunoinformatics; structural vaccinology.

PubMed Disclaimer

Conflict of interest statement

Declaration of interests J.R.W., K.M.M., J.M.S., and W.R.S. are named inventors on patent applications filed by Scripps and IAVI regarding ApexGT immunogens in this manuscript.

Figures

None
Graphical abstract
Figure 1
Figure 1
Human antibody repertoire analysis guides V2-apex germline-targeting immunogen design Repertoire analysis, immunogen design, and structure determination (steps with solid arrows) were carried out in this study. Immunization in a knockin mouse model (dashed arrows) was carried out in a companion manuscript (Melzi et al., 2022).
Figure 2
Figure 2
Analysis of precursor frequency and HCDR3 distance to bnAb identifies PCT64 as best and PG9/16 as second-best targets for vaccine design (A) HCDR3 junctions for HCDR3-dominant V2-apex bnAbs aligned with the inferred VH, DH, and JH gene sequences. Amino acids in gray are either located within a junction (non-templated) or mutated from a germline gene. (B) HCDR3 structures for each class of HCDR3-dominant V2-apex antibody: PCT64-35B (PDB:5FEH), CH04 (PDB:3TCL), PGDM1400 (PDB:4RQQ), PGT145 (PDB:5V8L), and CAP256-VRC26.25 (PDB:5DT1). Templated portions of the HCDR3 are colored as in (A). (C) Precursor detection rates among 14 donors, with 95% confidence intervals computed by the Wilson score method (Agresti and Coull, 1998). Positive detection was defined as at least one precursor being found in a donor. (D) Precursor frequencies for each donor, and for each class of HCDR3-dominant V2-apex bnAb, considering HC only (H) or heavy and light chains (H+L). The number of unique sequences used in the precursor search for each donor is indicated. Black lines indicate median precursor frequencies computed over non-zero values, because the zero values (non-responders) are accounted for in (C). (E) The mean number of mutations to a known bnAb in the class for each precursor found in (D). Each symbol represents the mean for one donor. Black lines indicate the median over donors. See also Tables S1–S5.
Figure 3
Figure 3
Multiple factors affect V2-apex bnAb precursor frequency (A) HCDR3 length distribution across all 14 donors, in which symbols represent the mean frequency across all donors, and error bars indicate the standard deviations. V2-apex bnAb precursor lengths are highlighted in colored boxes: PCT64, gray; CH04, red; PG9/PG16, magenta; PGT/PGDM, cyan; CAP256, gold. (B) Average D-gene motif start position within HCDR3 for the PCT64 motif in HCDR3s of length 25; the CH01-CH04 motif in HCDR3s of length 26; and the PG9/PG16 motif in HCDR3s of length 30; across the 14 donors. The highlighted colored bar indicates the start position of the D gene motif for each V2-apex bnAb. PGT/PGDM and CAP256 classes had insufficient matches to compute average D motif start positions. (C) DJ-gene usage frequency heat map among all HCDR3s, long HCDR3s (>20 amino acids), and very long HCDR3s (>24 amino acids) in 14 HIV-unexposed donors. The J gene frequency is shown in a single dimension at the bottom of each DJ heatmap. Points are shown for each DJ gene used for the V2-apex bnAbs. (D) PG9 DJ junctional analysis. The frequency of amino acids at position 100p, for all HCDR3s of length 30 with a PG9-like DJ junction. The most common amino acid at 100p was tyrosine, observed in multiple donors and clonotypes. (E) PG9 and PCT64 iGL variant junction alignments used in germline-targeting immunogen design. (F) PG9 and PCT64 iGL variants binding affinities measured by SPR against the native-like trimers BG505 SOSIP.D664 (Sanders et al., 2013) and BG505 SOSIP MD39 (Steichen et al., 2016). NB, no binding, is the highest concentration tested. See also Table S3.
Figure 4
Figure 4
V2-apex germline-targeting immunogen design produces ApexGT trimers as candidate immunogens to prime PCT64 and PG9/16 responses (A) The immunogen design pathway starting from BG505 SOSIP.D664. The V1/V2 region is shown with remodeled glycans as cyan spheres. The PG9 binding site is beige and two of the three protomers are shown in shades of gray. The mutations determined from each library are shown in red surface patches. Loop2b (residues 181-191 in HXBC2 numbering) is shown as a red tube representation. The germline variant that bound the immunogen is shown below each step. For clarity, glycans are shown on only a single protomer. (B) SPR KDs for ApexGT trimer analytes binding to PCT64 and PG9 variants as IgG ligands with data fit using a 1:1 binding model. NB, no binding, is the highest concentration tested. (C) ELISA antigenic profiles of MD39 and ApexGT trimers. AUC is the area under the curve of the dilution series of the antibody shown on the x-axis. (D) Glycan composition for ApexGT trimers using two methods, SSGA (Allen et al., 2021) and DeGlyPHER (Baboo et al., 2021). High mannose, green; complex, pink; unoccupied, gray; N.D., glycan could not be resolved. (E) DSC melting temperatures of ApexGT trimers. See also Figures S1 and S2.
Figure 5
Figure 5
Cryo-EM structures of ApexGT2.2MUT bound to PCT64 LMCA Fab, and ApexGT2 bound to PCT64 35S Fab, reveal complex interactions to guide design of PCT64-targeting ApexGT trimers (A) Refined atomic models of both complexes. (B) Isolated structure and domain organization of both Fabs aligned to their HCs, with PCT64 35S shown as partially transparent. (C) Electrostatic potential surfaces of both Fabs and of ApexGT2.2MUT (without glycans). (D) Close-up views of the binding interface showing antibody-gp120 protein interactions for both complexes. The outset shows an additional h-bond between the LMCA HC and gp120 not visible in the close-up. (E) Same as (D) but showing antibody-glycan interactions, with some parts of the LC hidden to enable viewing of specific residue contacts. (F) Table showing the number of h-bonds and total interfacial surface area between the different components of the epitope/paratope of both complexes. (G) Both complexes aligned on gp120A with arrows indicating the change in binding angle from LMCA to 35S. The additional red fragment on the rightmost panel is a model of the native BG505 SOSIP.MD39 loop2B highlighting the potential steric clashes with the LMCA HC. See also Table S6 and Figures S2–S5.
Figure 6
Figure 6
Cryo-EM structures of ApexGT3A bound to PG9 iGL Fab, and ApexGT3A.N130 bound to PG9 Fab, reveal complex interactions to guide design of PG9-targeting ApexGT trimers (A) Refined atomic models of both complexes. (B) Isolated structure and domain organization of both Fabs aligned on their HCs, with PG9 shown as slightly transparent. (C) Electrostatic potential surfaces of both Fabs and of ApexGT3A (without glycans) calculated with APBS. (D) Close-up views of the binding interface showing antibody-gp120 protein interactions for both PCT64 complexes. All gp120 residues within 4Å of the HCDR3 are shown and h-bonds are indicated with dashed blue lines. (E) Same as (D) but antibody-gp120 glycan interactions. (F) Table showing number of h-bonds and total interfacial surface area between the different components of the epitope/paratope of both complexes. (G) Both complexes aligned on gp120A revealing an identical binding angle. The additional red fragment on the rightmost panel is a model of the native BG505 SOSIP.MD39 loop2B highlighting the potential steric clashes with the HC of both Fabs. (H) NSEM 2-D class averages of ApexGT3 in complex with PG9 showing classes with more than one Fab bound (Fabs false-colored blue) along with a segmented 3-D reconstruction of the 2 Fab bound class (highlighted in red). See also Table S6 and Figures S5–S7.
Figure 7
Figure 7
Developing membrane-bound ApexGT trimers produces candidates for nucleic acid delivery (A) Cartoon schematic of cell-surfaced displayed ApexGT5 trimer. Link14, shown in pink, bridges gp41 and gp120. Location of GT mutations is indicated in green. (B) Cell surface antigenic profile for DNA-expressed membrane-anchored trimers binding to IgG for control bnAbs (quaternary, PGT151 and PGT145; CD4bs, 12A12; and V3-glycan, PGT121), non-nAbs (V3, 4025; CD4bs, B6 and F105), V2-apex bnAbs (PG9, PCT64), and V2-apex bnAb precursors (PCT64.LMCA, PCT64.LMCA.JREV, PCT64.iGL). Mean-fluorescence intensity (MFI) via fluorescence-activated cell sorting (FACS) binding was normalized to PGT121 binding with error bars representing standard deviation (n = 2). All trimers are based on the BG505 isolate, and all have a c-terminal truncation at residue 709. gp151 contains no other modifications. MD39 contains stabilizing mutations in BG505 SOSIP MD39 (Steichen et al., 2019). ApexGT trimers contain GT mutations described in the text. (C) Similar to (B) but with membrane-anchored trimers expressed from mRNA.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Abbott R.K., Lee J.H., Menis S., Skog P., Rossi M., Ota T., Kulp D.W., Bhullar D., Kalyuzhniy O., Havenar-Daughton C., et al. Precursor frequency and affinity determine B cell competitive fitness in germinal centers, tested with germline-targeting HIV vaccine immunogens. Immunity. 2018;48:133–146.e6. doi: 10.1016/j.immuni.2017.11.023. - DOI - PMC - PubMed
    1. Agirre J., Iglesias-Fernández J., Rovira C., Davies G.J., Wilson K.S., Cowtan K.D. Privateer: software for the conformational validation of carbohydrate structures. Nat. Struct. Mol. Biol. 2015;22:833–834. doi: 10.1038/nsmb.3115. - DOI - PubMed
    1. Agresti A., Coull B.A. Approximate is better than "Exact" for interval estimation of binomial proportions. Am. Statistician. 1998;52:119–126. doi: 10.2307/2685469. - DOI
    1. Alam S.M., Dennison S.M., Aussedat B., Vohra Y., Park P.K., Fernández-Tejada A., Stewart S., Jaeger F.H., Anasti K., Blinn J.H., et al. Recognition of synthetic glycopeptides by HIV-1 broadly neutralizing antibodies and their unmutated ancestors. Proc. Natl. Acad. Sci. USA. 2013;110:18214–18219. doi: 10.1073/pnas.1317855110. - DOI - PMC - PubMed
    1. Aldon Y., McKay P.F., Allen J., Ozorowski G., Felfödiné Lévai R., Tolazzi M., Rogers P., He L., de Val N., Fábián K., et al. Rational design of DNA-expressed stabilized native-like HIV-1 envelope trimers. Cell Rep. 2018;24:3324–3338.e5. doi: 10.1016/j.celrep.2018.08.051. - DOI - PMC - PubMed

Publication types

Substances