Animals and sample collection

10 Infant Indian rhesus macaques (Macaca mulatta), born from the breeding colony of the California National Primate Research Center (CNPRC, Davis), were enrolled in the study at birth and were maintained for the study duration of 120 weeks at CNPRC. The animals were negative for simian immunodeficiency virus, type D retrovirus, and simian T-cell lymphotropic virus type 1. The study was approved by the Institutional Animal Care and Use Committee of the University of California, Davis, and the CNPRC is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC). Animals were nursery reared for the first four months and then were moved into regular non-infectious housing for the remainder of the study. All procedures of animal husbandry and sample collections were performed according to established Standard Operating Procedures that were approved by the Institutional Animal Care and Use Committee of the University of California, Davis, and that were in compliance with the 2011 Guide for the Care and Use of Laboratory Animals53. All experimental manipulations were performed under ketamine anesthesia (10 mg/kg body weight) administered by the intramuscular (IM) route. Blood samples were collected via peripheral venipuncture at weeks 0, 8, 26, 48, 62, 74, 98, and at necropsy (week 120). Both plasma samples and Peripheral Blood Mononuclear Cells (PBMCs) were isolated from the blood draws, processed, and cryopreserved. Lymph node biopsies were collected at weeks 26, 62, 98, and 120 from a peripheral lymph node under general sedation, followed by topical anesthesia and 3 days of analgesia (ketoprofen, 2–5 mg/kg once daily). Animals were euthanized at the end of the study with ketamine sedation, followed by an overdose of sodium pentobarbital (120 mg/kg IV); consistent with American Veterinary Medical Association recommendations54.

Vaccines and vaccine regimens

All 10 macaques received the ENGERIX-B® (GlaxoSmithKline, Philadelphia, PA) hepatitis B virus vaccine dose series (10 μg IM) at 0, 6, and 24 weeks post birth, which mirrors the human infant vaccine schedule. The macaques were then randomized into two groups of 5 animals, balanced for sex (4 males group A, 3 males group B). One group received the HIV Envelope (Env)-only vaccine regimen, and the other group received the conjugate HIV Env HBsAg vaccine regimen. The overall study scheme is shown in Fig. 1a.

HIV-1 Env-only vaccine group: At 60 weeks of age, animals in the HIV-1 Env-only vaccine group were primed with an intradermal (ID) administration of 1 × 107 PFU of the MVA-1086C Env gp160 virus (MVA_Env vaccine). This conventional MVA-based vaccine was generated using red-fluorescent protein (RFP) selection and previously described methods55,56. Briefly, the codon-optimized, subtype C isolate 1086 Env gp160 gene (Gene bank: FJ444399.1) and the RFP reporter gene (DsRed, Clontech Laboratories, Mountain View, CA) were inserted into MVA A681 virus56 within an intergenic region between essential viral genes 0I8R and G1L, under the control of a synthetic early/late poxvirus promoter55. The RFP reporter gene used to aid in plaque selection was flanked by wild-type lox sequences and was removed from the final virus by Cre-mediated deletion as previously described56, producing MVA-1086C Env gp160 virus. At 72 and 96 weeks of age animals received booster immunizations of the MVA_Env vaccine (1 × 107 PFU, ID) and alum-adjuvanted gp120 Env gp120 proteins representing subtype C isolate TV1 (GenBank EU855132.1), subtype AE isolate A24457, and subtype B isolate MN (GenBank AF075722.1). These envelopes were selected due to their recent use in poxvirus prime, protein boost, clinical trials1,50 The final Env protein immunogen was prepared with equimolar amounts of each gp120, in 2% aluminum hydroxide wet gel suspension adjuvant (Alhydrogel®, InvivoGen, San Diego, CA), administered at a final dose of 50 μg, IM.

Conjugate vaccine group: At 60 weeks of age, animals in the conjugate vaccine group were primed with an intradermal (ID) administration of 1 × 107 PFU modified vaccinia Ankara (MVA) virus engineered to co-express HIV-1 Env protein and HBsAg (MVA_Env–HBsAg). The MVA-based vaccine was generated starting from the MVA-1086C Env gp160 virus described above. To incorporate HBsAg, the gene for HBsAg subtype adw2 WHO International Standard58 (Gene bank KY003230.1) was inserted into the deletion III region of the plaque-purified MVA-1086 Env gp160 virus. HBsAg was expressed under the control of the enhanced H5 promoter59. PCR of plaque purified virus was used to identify virus expressing both the Env and HBsAg genes, and Env and HBsAg protein production was confirmed by immunoblot (Supplementary Fig. 1A) of MVA_Env–HBsAg-infected cells.

At 72 and 96 weeks of age animals received booster immunizations of the MVA_Env–HBsAg vaccine (1 × 107 PFU, ID) and an alum-adjuvanted Env–HBsAg protein conjugate (50 μg, IM). For generation of the protein conjugate vaccine, Env gp120 protein amine groups were covalently linked to HBsAg protein (subtype adw, Novus Biologicals, Centennial CO) reduced sulfhydryls with a 7.3 Å spacer using Sulfo-GMBS (ThermoFisher, Waltham, MA) according to manufacturer’s instructions. Conjugates were purified by high performance liquid chromatography (HPLC), and quality control validation of conjugation was performed using antibody-coated bead pulldown assays. Negative stain electron microscopy (NSEM) (Supplementary Fig. 1B) provided evidence that the majority of the conjugate protein was present in small, likely monomeric forms, while a minor fraction aggregated into larger particles that did not appear to form well-defined virus-like nanoparticles60,61. Conjugate vaccines were formed using the same 3 different Env gp120 proteins used in the Env-only vaccine: subtype C isolate TV1, subtype AE isolate A244 and subtype B isolate MN. The antigenicity of the conjugate vaccine was compared to that of the Env protein only vaccine using ELISA performed with well-characterized HIV-1 specific antibodies targeting the CD4 binding site, mannose-dependent 2G12 epitope, C1-C2 cluster A region, V1/V2 region, and V3 region (Supplementary Fig. 1C). For the final immunogen, the three different Env–HBsAg conjugates were mixed in equimolar ratios in 2% aluminum hydroxide wet gel suspension adjuvant (Alhydrogel®), at a final IM dose of 50 μg total protein.

anti-HBsAg ELISA: Custom ELISA using a commercially available ELISA buffer kit (ThermoFisher) and recombinant HBsAg (Abcam, Cambridge, UK) was performed to measure plasma levels of HBsAg-specific antibodies. High binding 96 well plates were coated overnight with HBsAg protein at 2 μg/mL and 4 °C in the kit’s coating buffer A (pH 7.4). The next morning the plates were washed with Wash Buffer and then blocked for 1 h in Assay Buffer. Plates were then washed, and plasma samples were added after serial dilution in Assay Buffer (1:25 starting dilution, 12 serial 3-fold dilutions for samples collected at weeks 0, 8, and 26, and 1:100 starting dilution, 6 serial 5-fold dilutions for samples collected at 60, 62, 74, and 98 weeks). Plates were incubated with plasma samples at room temperature for 1 h and then washed with Wash Buffer. Plates were then incubated with the secondary detection antibody (mouse anti-monkey IgG HRP Southern Biotech, 1:10,000 dilution in Assay Buffer) for 1 h at room temperature, washed twice with the Wash Buffer, and developed for 15 min using the kit-provided chromogen solution and stopped with an equal volume of the provided Stopping Buffer. Plates were read on a Perkin-Elmer VICTOR® NivoTM reader at 450 nm within 30 min. EC50 values (inverse dilutions) were calculated using non-linear curve fitting in GraphPad Prism Software (Version 10.0).

TFH activation-induced markers (AIM) assay

TFH AIM assays were performed as described62. In brief, week 60 and week 74 PBMCs or week 98 lymph node mononuclear cells were thawed in R10 (10% FBS in RPMI with L-glutamine, Pen-strep and gentamycin), viable cells were counted, and then incubated in AIM media (10% Human AB serum in AIM-V serum free media) for three hours at 37 °C. Cells were then recounted, and plated in a 96 well v-bottom plate at 4 million viable cells per mL using the following conditions: Unstimulated: AIM media with 0.1% DMSO; Staphylococcal enterotoxin B (SEB) stimulated: AIM media with 0.5 μg/mL SEB; HBV stimulated: AIM media with 5 μg/mL HBsAg protein (Abcam, Cambridge, UK) + 0.5 μg/mL HBV large envelope protein peptide mix (JPT, Berline, DE); and HIV Env stimulated: AIM media with 5 μg/mL 1086cK160Nd11 and 0.5 μg/mL HIV Env Potential T cell Epitopes peptide pool (NIH AIDS Reagent program, Bethesda, MD). Cells were incubated 18 h at 37 °C and then washed with PBS 1% BSA and incubated for 30 mins room temperature with commercially available antibodies to cell-surface proteins as follows: PD-1 BV421 (Biolegend cat. 329920, clone EH12.2H7, volume 5 μL per 100), CD8a BV570 (Biolegend cat. 301038, clone RPA-T8 volume 0.62 μL per 100), CCR7 BV605 (BD Biosciences cat 563711, clone 3D12 volume 1.25 μL per 100), CD25 BV650 (Biolegend cat 302634, clone BC96 volume 1.25 μL per 100), CD4 BV711 (Biolegend 317440, clone OKT4 volume 1.25 μL per 100), CD45RA BV785 (BD Biosciences cat 741010, clone 5H9 volume 0.62 μL per 100), ICOS BB515 (BD Biosciences cat 565880, clone C398.4A volume 0.31 μL per 100), CCR6 BB700 (BD Biosciences cat 566477, clone 11A9 volume 1.25 μL per 100), OX40 PE (BD Biosciences cat 340420, clone L106 volume 5 μL per 100), CXCR3 PE-CF594 (BD Biosciences cat 562451, clone 1C6 volume 2.5 μL per 100), CD69 PE-Cy5 (Biolegend cat 310908, clone FN50 volume 5 μL 100), CXCR5 Biotin (eBioscience cat 13–9185–82, San Diego, CA, clone MU5UBEE volume 0.62 μL per 100), CD137 AF647 (Biolegend cat 309824, clone 4B4-1 volume 1.25 μL per 100), CD3 APC-Cy7 (BD Biosciences cat 557757, clone SP34-2 volume 1.25 μL per 100). Cells were then washed twice with PBS and stained with a viability dye (LIVE/DEAD Fixable Aqua Dead Cell Stain 1:800 dilution, Invitrogen cat L34957) and PeCy7-Streptavidin (Biolegend 1:800 dilution) for 20 min at room temperature. Cells were then washed with 1% PBS in BSA, resuspended in 1% PFA in PBS and acquired immediately on a LSR II Fortessa and analyzed using FlowJo Version 10.8.0 (BD Biosciences). Gates were drawn using concatenated sample files that included 5000 events from each individual sample. HBsAg and Env specific TFH frequencies were reported as the percentage of TFH cells expressing OX40 and CD137 after background subtraction of the unstimulated condition from the HBsAg and Env conditions.

CXCL13 ELISA

Human CXCL13 ELISA kit was purchased from R&D Systems (Minneapolis, MN) and performed according to manufacturer specifications and as previously described23. Data are reported as concentrations of CXCL13 in peripheral blood sera collected from immunized rhesus macaques, as determined by interpolation using a standard curve.

B cell sorting and phenotyping

Week 98 cryopreserved PBMC samples were thawed, counted, and CD3 depleted using anti-CD3 NHP microbeads (Miltenyi Biotech, Gaithersburg, MD). Cells were then washed with DPBS and stained with a viability dye (LIVE/DEAD Fixable Aqua Dead Cell Stain, Invitrogen) for 20 min, washed twice in buffer (PBS + 1%FBS), and then surface stained for 30 min at room temperature using commercially available antibodies and 1086cK160Nd11 gp120 hooks conjugated to either AF647 or BV42124,25,26. Cells were then washed and resuspended in 0.04% BSA in PBS and immediately acquired on a upgraded FACS Aria II. HIV specific B cells were identified by sorting on Live Cells, CD14/CD19/CD3 negative cells, CD20 Positive cells, then gp120-BV421, gp120-AF647 dual positive cells. Fluorescently conjugated Abs used for cell surface staining were: IgM BV605 (BD Biosciences cat 562977, Clone G20-127 volume 0.62 μL per 100), CD3 BV650 (BD Biosciences cat 563916, clone SP34-2 volume 1.25 μL 100), CD16 BV650 (BD Biosciences cat 563692, clone 3G8 volume 1.25 μL per 100), CD14 BV650 (BD Biosciences cat 563419, clone M5E2 volume 1.25 μL per 100), CD27 BV711 (Biolegend cat 302834, San Diego, CA, clone 0323 volume 0.62 μL per 100), CD20 BV785 (Biolegend cat 302356, clone 2H7 volume 0.62 μL per 100), CD21 Pe-Cy7 (BD Biosciences cat 561374, clone B-ly4 volume 0.62 μL per 100), and IgG APC-H7 (BD Biosciences cat 561297, clone B56 volume 1.25 μL per 100).

Antigen-specific B cell single-cell RNA sequencing (scRNA-seq)

Week 98 HIV Env-specific B cells present in peripheral blood were flow-sorted as described above, and gene expression and B cell receptor (BCR) sequencing was performed using scRNA-seq with the 10x Genomics Platform. Briefly, sorted cells were resuspended in PBS + 0.04% BSA. To improve efficiency of single-cell partitioning and minimize cell loss we added 5000 irrelevant cells (murine EL4 cell line, ATCC TIB-39) to each cell suspension. Transcripts from these filler cells are easily removed during initial sequencing data quality control analysis due to species and cell-type mismatch. Single cell suspensions were loaded onto a GemCode Single-Cell instrument (10× Genomics, Pleasanton, CA) and single-cell RNA-seq libraries were then prepared from the single-cell bead emulsions using a 5’ GEX library and VDJ library (10x Genomics) followed by sequencing using 150 cycle high output NextSeq (Illumina, San Diego, CA). Mean reads per cell exceeded 20 thousand for GEX and 5 thousand for VDJ.

scRNA-seq mapping and analysis

Transcriptome and immune profiling analysis were performed using the Cell Ranger Single Cell Software Suite (version 4.0.0, 10x Genomics) as previously described63. Briefly, for transcriptomic analysis, samples were demultiplexed, assembled, filtered, and aligned to the Mmul10 reference (accession NCBI:GCA_003339765.3) followed by unique molecular identifier (UMI) counting. For the immune profiling analysis, samples were demultiplexed, assembled, filtered, and aligned to a custom rhesus macaque VDJ reference. The custom VDJ reference was created by compiling three reference libraries: the default macaque Ig gene library from the software Cloanalyst (https://www.bu.edu/computationalimmunology/research/software/), macaque reference Ig gene segments from IMGT, and macaque reference Ig gene segments from macaque genome sequencing64 and formatted for compatibility with Cell Ranger. A text file specifying the sequences of the rhesus macaque inner enrichment primers was also used as described in the documentation for running customized libraries in Cell Ranger that is provided by 10X Genomics. Cell ranger sequences and chain annotations were then used as input into the Cloanalyst software package. Immunogenetics information of rhesus macaque antibody sequences were assigned using the default rhesus macaque Cloanalyst library and sequences were grouped into clones with Cloanalyst. Graph-based cell clustering, dimensionality reduction, and data visualization were performed using the Seurat R package (version 4.0.0). Cells that exhibited >5% mitochondrial transcripts, expressed less than 200 unique transcripts, or more than 2500 unique transcripts were excluded. Additionally, only cells with a single copy of functional light and heavy chain (1:1 functional) in the BCR sequencing are included in the transcriptomic analysis. After applying the above QC criteria we identified a total of 6130 cells (3205 cells from Env immunized, and 2925 cells from Env–HB immunized macaques) to be included in the analysis. Differentially expressed transcripts were determined using the Likelihood-ratio test for single-cell gene expression statistical tests in the Seurat R package. Graphics were generated using Seurat and ggplot2 R packages.

Binding antigen multiplex assay (BAMA)

Custom BAMA was performed with heat-inactivated plasma samples as previously described65. HIV Env-specific antigens were conjugated to polystyrene beads (Bio-Rad Laboratories, Inc., Hercules, CA), and the coupled beads were used to approximate the binding responses of the vaccine-induced antibodies to a panel of envelope glycoproteins. The following panel of antigens was tested: gp70 B case A2 V1V2, 1086.C V1V2, gp70 ConC_V3tags, Con6 gp120/B (106), and 1086.C K160N gp120. The following Env peptides were used: Biotin (Bio) V2.1086 C (Bio-KKKTELKDKKHKVHALFYKLDVVP), Bio V3.C (Bio-KKKNNTRKSIRIGPGQTFYATGDIIGDIRQAHC), Bio V2.B (Bio-KKKTSIRDKVQKEYALFYKLDVVP), C1 Biotin (Bio-KKKMQEDVISLWDQSLKPCVKLTPLCV), and Bio_RV144_ C5.2.C (Bio-KKKSELYKYKVVEIKPLGIAPTKAKRRVVEREKRAV). RIVIG (rhesus immunodeficiency virus immune globulin) was used as positive control. The antigen-conjugated beads were incubated with plasma at a 1:100 dilution. IgG binding in plasma was detected with PE-conjugated mouse anti-monkey IgG antibody (Southern Biotech, Birmingham, AL) at 4 μg/ml. Assays were read on a Bio-Plex 200 instrument (Bio-Rad Laboratories), and binding results were expressed in MFI units. The MFI of beads in wells that did not contain sample (blank wells) was subtracted from sample-specific MFIs within respective assay plates and the blank beads or beads coated with MulV gp70 His6 were used for background. An HIV Env-specific antibody response was considered positive if it had MFI values above the lower detection limit of 100 MFI after background correction. To ensure consistency between assays, 50% effective concentration and maximum MFI values of the positive controls were tracked by Levey-Jennings charts.

ADCC GTL assay

The ADCC-GTL assay was performed as previously described66. Briefly, gp120 coated CEM.NKRCCR5 cells (NIH Reagent Program, from A. Trokla) were labeled with TFL4 and the viability marker NFL1 (both from OncoImmunin, Gaithersburg, MD, each at final dilution of 1:1000). After counting and washing, 5000 target cells per well were added to 96-well V-bottom plates and incubated with the Granzyme B (GzB) substrate (OncoImmunin) and Human PBMC effector cells (30:1 effector to target ratio) for 5 min at room temperature. Serial dilutions of plasma samples were then added and the plate was incubated an additional 15 min at room temperature, centrifuged for 1 min at 300 x g, then incubated for 1 h at 37 °C 5% CO2. Well contents were then washed and re-suspended in 150 µl 1%FBS in PBS and acquired directly with BD Fortessa flow cytometer (BD Biosciences, San Jose, CA) within 4 h using the High Throughput Sampler (HTS, BD Biosciences). Prior to acquisition with FACSDiva software (BD Biosciences), the area scaling factor was adjusted to ensure that cell singlets (CEM.NKRCCR5 CD4+ T cells) indicated equivalent ratios of FSC-H and FSC-A. Flow cytometry data analysis was performed using FlowJo 10.8.0 software (BD Biosciences).

Infected cell antibody binding assay (ICABA)

Indirect surface staining was used to measure the ability of plasma antibodies to bind to the surface of HIV-1 infected cells using methods similar to those previously described28,65. Briefly, CEM.NKRCCR5 cells were mock infected or infected with an HIV-1 infectious molecular clone virus expressing the C.1086, C.TV-1, or AE.CM235 Env protein. The cells were then washed and incubated with a 1:100 dilution of plasma samples for 2 h at 37 °C and then stained with Live/Dead Aqua Dead Cell Stain (Thermo Fisher Scientific) to exclude dead cells from analysis. Cells were next washed and then permeabilized with Cytofix/Cytoperm solution (BD Biosciences) prior to staining with RD1-conjugated anti-p24 antibody KC57 (Beckman Coulter, Inc., Indianapolis, IN) and fluorescein isothiocyanate (FITC)-conjugated goat anti-rhesus IgG (H + L) polyclonal antiserum (Southern Biotech). Cells positive for serum antibody binding were defined as viable, p24 positive, and FITC positive. Final results were reported as the change in FITC median fluorescent intensity (MFI) for post-vaccination samples compared to the pre-vaccination sample; gated on the live, infected cell population (p24 positive cells) after subtraction of the MFI observed for staining of cells stained with secondary antibody alone and mock-infected cells.

ADCC-Luc

The ADCC-Luc assay was performed as previously described57,67. Briefly, CEM.NKRCCR5 were infected with replication competent infectious molecular clones (IMCs) of C.1086, C.TV-1, or AE.CM235. These contain C.1086, C.TV-1, or AE.CM235 envelope sequences on NL-LucR.T2A backbone68. The infection of the cells was confirmed with p24 intracellular staining. The infected cells were co-plated with the human NK92 cell line engineered to express the 158I variant of rhesus macaque FcγRIIIa69 at a 10:1 ratio for 6 h in complete media with serially diluted plasma from the vaccinated animals. At the end of the 6-hour incubation, Viviren (Promega, Madison, WI) was added to the plate and the luminesce of each well was measured. ADCC activity, reported as percent specific killing, was calculated from the change in relative light units (RLU) resulting from the loss of intact target cells in wells containing effector and target cells in the presence of serum samples compared to RLU in control wells containing target cells and effector cells alone, and after baseline subtraction using the pre-vaccination time point Week 60.

Neutralization

Neutralization was measured as the ability of serum samples to reduce virus infection of TZM-bl cells as previously described70. Serum was incubated with subtype B tier 1 isolate (B.MN.3), tier 2 isolate C.TV-1, or tier 1 isolate AE.TH023.6 pseudoviruses for 45 min at 37 °C (29, 30), plated with TZM-bl cells, and then allowed to incubate for 48 h. A luciferase reagent (Bright-Glo; Promega) was added, and luminescence was measured. Results were reported as the 50% inhibitory dilution (ID50), which is the dilution of serum resulting in 50% reduction in luminescence compared to virus control wells.

Statistical analysis

Spearman’s rank correlation coefficient was used to calculate correlations between immunologic markers. Mann-Whitney U tests were used to compare these markers between Env and Env-HB groups and the resulting p-values were adjusted for multiple testing to control for the false discovery rate (FDR) with the Benjamini-Hockberg method. Random forest models were developed as a classification method and to identify the immunologic markers that classified each vaccine group. The variable importance measures for each random forest model were calculated using the mean decrease in Gini impurity as a ranking measure which determines how the model’s accuracy decreased when a variable was removed. The out-of-bag estimate error rate was used to assess the accuracy of each random forest. We also used lasso-regularized logistic regression models as an alternative approach to identify the immunologic variables that classified the two vaccine groups, and compared these results to random forest results. The penalty parameter λ was selected at the value with the lowest binomial deviance after 10-fold leave-one-out cross-validation. The penalized model was then fit using the minimum λ value and the non-zero coefficients were extracted.

Immunoblotting

Monolayers of DF-1 cells (ATCC #CRL-3586) were mock-infected or infected with MVA_Env, MVA_Env-HBsAg, or parental MVA at a multiplicity of infection of 3 for 24 h. Cells were then washed with PBS and lysed by repeat freeze-thaw cycles, sonication, and scraping in the presence of protease inhibitor (HALT, Protease Inhibitor Cocktail, Thermo-Fisher Scientific). The total protein concentration of the lysates was determined using the Bio-Rad Quick StartTM Bradford assay (Bio-Rad Laboratories, Inc). Equal amounts of total protein (2 μg each) and 5 μL protein standards solution (Precision Plus Unstained Protein Standards, Bio-Rad Laboratories, Inc.) were loaded onto 4% to 20% gradient polyacrylamide gels (Mini-PROTEAN® TGX, Bio-Rad Laboratories) in Laemmli buffer with β-mercaptoethanol and separated by electrophoresis on the BioRad Mini-PROTEAN Tetra System. The proteins were transferred to PVDF membranes using the TransBlot® Turbo System (Bio-Rad Laboratories, Inc.) and blocked for 1 h in 2% BSA in tris-buffered saline (TBS) with 0.05% tween (TBST). Primary antibodies (anti-Env, mAb 2G12, DHVI Protein Production Facility; and anti-HBsAg, mAb ab68520, Abcam) were diluted in 1% BSA TBST and incubated with the membrane for 1 h at 4 °C. The membranes were then washed extensively in TBS. Secondary detection reagents (anti-human HRP for anti-Env immunoblot, Sigma Chemical; and Streptactin HRP for anti-HBsAg immunoblot, Bio-Rad Laboratories) were diluted in 1% BSA TBST and incubated with the membranes for 1 h at 4 °C, followed by extensive washing with TBST and one wash with TBS. Band visualization was performed using the ClarityTM Western ECL substrate (Bio-Rad Laboratories, Inc) according to manufactures’ instructions, and images were acquired on the Bio-Rad ChemiDoc imaging system. All blots were derived from the same experiment and were processed in parallel.

Negative stain electron microscopy (NSEM)

NSEM analysis was performed via a Core Service of the Duke Human Vaccine Institute under the direction of Dr. R.J. Edwards. A frozen aliquot of the Env–HB conjugate vaccine from −80 °C was thawed at room temperature in an aluminum block for 5 min. Sample was then diluted to 40 µg/ml with 0.02 g/dl Ruthenium Red in HBS (20 mM HEPES, 150 mM NaCl pH 7.4) buffer. After 10–15 min incubation, the sample was applied to a glow-discharged carbon-coated EM grid for 8–10 s, blotted, consecutively rinsed with 2 drops of 1/20X HBS, and stained with 2 g/dL uranyl formate for 1 min, blotted and air-dried. Grids were examined on a Philips EM420 electron microscope operating at 120 kV and nominal magnification of 49,000x, and 10 images were collected on a 76 Mpix CCD camera at 2.4 Å/pixel.

Anti-HIV-1 Envelope Enzyme-Linked Immunosorbent Assay (ELISA)

Anti-HIV Env ELISA was performed using the same general approach as described for anti-HBsAg ELISA. Briefly, high-binding plates were coated with the Env protein vaccine or the Env-HB protein conjugate vaccine at 2 μg/mL overnight at 4 °C using Coating Buffer provided in the ThemoFisher ELISA Buffer Kit (ThermoFisher Scientific). Plates were then washed in Wash Buffer and blocked for 1 h using Assay Buffer provided in the kit. HIV-1 specific human monoclonal antibodies targeting different regions of the HIV-1 Env (mAbs VRC01, 2G12, A32, CH59, 830 A, PG9, Ab3074, 10-1074) or the anti-influenza control mAb CH65 were serially diluted (6 serial 5-fold dilutions starting at 0.5 μg/mL) in Assay Buffer and added to the plates for 1 h at room temperature. All conditions were tested in duplicate wells. Plates were washed and binding of primary antibodies was detected by incubation with goat anti-Human IgG HRP (Jackson ImmunoResearch, Inc., West Grove, PA), diluted in Assay Buffer, for 1 h at room temperature. Plates were developed using the buffers and chromogen solution provided in the kit according to manufacturer’s recommendations. Plates were read on a Perkin-Elmer VICTOR® NivoTM reader at 450 nm within 30 min.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.