Cell lines.

Cells were HEK-derived and included 293T, 293F, and Expi293 cells. TZM-bl cells are a He-La derived cell line that express CD4 receptor and CXCR4 and CCR5 chemokine co-receptors and luciferase and β-galactosidase genes under the control of the HIV-1 promoter. Both 293T and TZM-bl cell lines were maintained and used in complete Dulbecco’s Modified Eagle Medium (complete DMEM) which is comprised of high-glucose Dulbecco’s Modified Eagle Medium, 2 mM L-glutamine, 10% fetal bovine serum, and 1x Penicillin-Streptomycin. The 293T and TZM-Bl cell lines were maintained at 37°C and 5% CO₂. 293F and Expi293 cells were maintained in Gibco Freestyle 293 Expression Medium (Thermo Fisher) and Gibco Expi293 Expression Medium (Thermo Fisher) at 37°C and 10% CO₂ with shaking at 120 RPM.

Cow immunizations.

Antigen-specific B cell responses were primed and expanded by immunization of Bos taurus calves. The calves were bled from the jugular vein as often as once a week from the beginning of the immunizations and bi/tri-monthly during the last three months of the immunizations. Blood for sera and for isolation of PBMCs were collected at the same time, and terminal bleeds and splenocytes were collected at study termination. The calves and samples were not randomized or blinded. The calves were immunized with the appropriate immunogen at a dose of 200 μg per calf (formulated in Montanide ISA 201 adjuvant: Seppic, France) for all priming and boosting steps except the final boost. The immunogens were inoculated intradermally (200μl per site) on both sides of the neck region. The final boost was done with a SOSIP cocktail containing BG505.664, CRF250, WITO, and ZmZm containing 100 μg of each SOSIP formulated as above. Group 1 with cow-485 and cow-16157 were immunized with MT145K SOSIP.664 on days 0, 37, and 79, then with C108 SOSIP.664 on days 142 and 226. The final immunization on day 336 was done with a SOSIP.664 cocktail consisting of BG505, CRF250, WITO, and ZmZm. Group 2 with cow-488 and cow-491 were immunized with C108 SOSIP.664 on days 0, 37, 79, 142, and 226. The final immunization on day 336 was done with a SOSIP.664 cocktail consisting of BG505, CRF250, WITO, and ZmZm.

IgG purified from sera.

Sera samples were heat-inactivated in a 56°C water bath for 30 minutes and spun at maximum speed in a table-top centrifuge for 20 minutes. Sera were subsequently incubated with an equal volume of dry Protein G Sepharose beads (GE) and twice the volume of PBS. This was then incubated overnight at 4°C on a nutator. The next day, the beads were washed with 10x the volume of PBS, eluted with IgG elution buffer (Pierce), and neutralized with 2M Tris pH 9.0. The eluted fraction was buffer exchanged into PBS and concentrated into the original sera volume used for the purification. Samples were filtered through a 0.45 μm filter.

Pseudovirus neutralization assays and pseudovirus production.

Replication incompetent HIV pseudovirus was produced in HEK293T cells. They were co-transfected with plasmids containing HIV Env and PSG3ΔEnv backbone using a 1:2 ratio. FuGENE (Promega) was used as a transfection reagent in Opti-MEM (Thermo Fisher) or Transfectagro (Corning). Supernatants from cell cultures were harvested 48–72 hr post-transfection and sterile filtered with a 0.22 mM filter. Pseudovirus was either used neat or concentrated as needed using 50k amicon concentrators. Pseudovirus was either frozen or used fresh. Frozen pseudovirus was titrated to determine the concentration needed for use. Neutralization assays were performed by incubating 25 μl of monoclonal antibodies or IgG purified from sera with 25 μl of pseudovirus for 1 hr at 37°C in full-area 96-well tissue culture polystyrene microplates (Corning). After 1 hr 20 μl of TZM-bl target cells were added at a concentration of 0.5 million cells/mL with 40 μg/mL of dextran final [70]. Cells were grown in a humidified incubator for 24 hr, then 130 μl of media was added to all wells. 48–72 hr after cells were added, plates were read on a luminometer (Biotek) by adding lysis buffer combined with Bright-Glo (Promega). IgG purified from serum neutralization assays are reported as ID₅₀ (1/dilution). Monoclonal antibody neutralization titers are reported as IC₅₀ or IC₈₀ (μg/mL) titers. If the antibody had a IC₅₀ value of less than 0.005, the IC₅₀ of 0.005 was used. Percent breadth was calculated as the % breadth = (number of viruses neutralized)/(total viruses tested). nAb titer data panels are shown as geometric mean titers. ZM233 was used in place of ZMZM for neutralization assays. Analysis and graphing were done in Prism software.

SOSIP trimer purification.

MT145K, C108, CRF250, BG505.664dHis, WITO, 25710, CRF250, and ZM233 SOSIP were expressed in HEK293F cells and purified as described previously described [52,71]. Briefly, SOSIP was transfected with 300 μg of trimer DNA, 150 μg of Furin DNA, and 1.5 mL of PEI MAX 40000 (Polysciences) in Transfectagro. Five to six days after transfection supernatants were harvested and purified using a Galanthus nivalis lectin (Vector Labs) column or PGT145 affinity columns which are made by coupling CnBr-activated Sepharose 4B beads (GE Healthcare) to PGT145 bnAb antibody as described previously [71]. Purified proteins were further purified with size exclusion chromatography columns Superdex 200 10/300 GL column (Cytvia) in TBS or PBS. Trimers were frozen for storage at -80°C and then thawed and tested with ELISA or BLI for antigenicity with a range of known non-neutralizing and neutralizing HIV specific antibodies. SOSIP used as sorting baits were biotinylated using the Pierce Biotinylation Kit using the manufacturer’s protocol and tested for biotinylation via ELISA.

Single cell sorting.

Cow PBMCs and splenocytes were sorted as previously described with study specific modifications [56]. Cow cells were thawed in a 37°C water bath until only a small pellet of ice remained. Cells were resuspended with 10 mL of pre-warmed buffer consisting of 50% RPMI (Gibco) and 50% FBS. Cells were washed by centrifuging for 5 min at 400g. Supernatant was discarded and the cell pellet was gently resuspended in 5 mL of cold FACS Buffer (1x DPBS with 1% FBS, 1mM EDTA, and 25 mM HEPES). Cells were counted and washed, the FACS Buffer was then removed. FACS antibody mastermix was added per 10 million cells. FACS antibody master mix per 10 million cells consisted of 3.75 μg of FITC labeled goat-anti cow antibodies (Abcam) and SOSIP baits diluted in FACS Buffer. Baits for the first three sorts used 55.5 pM of each biotinylated and stained SOSIP which was coupled with fluorophore conjugated streptavidin in a 4:1 molar ratio for 1 hr on ice in the dark. The fourth sort used 200 nM of biotinylated SOSIP coupled in a 1:2 ratio with a streptavidin fluorophore. Additionally, one of these SOSIPS was combined with PGT145 Fab after staining to form a SOSIP-Fab complex. The cell and master mix combination was incubated on ice in the dark for 30 minutes. After 30 minutes 1 mL of 1:300 diluted FVS510 Live/Dead stain (BD Biosciences) was added per 10 million cells and incubated in the dark for 15 min. Cells were washed in 10 mL of FACS buffer and resuspended in 500 μl per 10 million cells of FACS buffer. The cells were filtered into a 5 mL round bottom FACS tube with a cell strainer cap (Falcon). The fourth sort used MT145K SOSIP coupled with streptavidin-PE (Invitrogen) and MT145K SOSIP complexed with Fab and coupled to streptavidin AF-647 (Invitrogen). The first and second sort selected for cells that had double positive binding to two of the three SOSIP baits which were 25710 (BV421, BD Biosciences), CRF250 (PE, Invitrogen), and MT145K (APC, Invitrogen). The third sort selected for cells that demonstrated double positive binding for the same SOSIP (MT145K or 25710) coupled individually to PE (Invitrogen) or AF-647 (Invitrogen). Cells from the first three sorts were sorted into 96-well plates containing 20 μl/well of lysis buffer on FACS ARIA III BD FACS sorter and immediately frozen on dry ice. Lysis buffer consists of 2.5 mM RNaseOUT (Invitrogen), 1.25x SuperScript IV Reverse Transcriptase Buffer (Invitrogen), 6.25 mM DTT, and 1% Igepal CA-630 (Sigma-Aldrich). The fourth sort selected cells positive to binding only MT145K SOSIP (PE) and did not have affinity for MT145K-PGT145 Fab Complex (AF-647). Cells were sorted into dry 96-well plates that were immediately frozen on dry ice. Data was analyzed using FlowJo.

Single cell PCR amplification and cloning.

cDNA was generated from single cells sorted into plates using RT-PCR. RT-PCRs were done two ways depending on if the plate was dry or had lysis buffer. The plates with lysis buffer had 9 μl of RT-PCR buffer added to them consisting of 100 U of SuperScript IV Reverse Transcriptase Enzyme (Invitrogen), 1.1 mM dNTPs(New England Biolabs), 0.44x SSIV buffer(Invitrogen), and 0.375 μg of random hexamers (Gene Link). Each well in the dry plates received 15 μl of 3.33 mM RNaseOut (Invitrogen), 1x SuperScript IV Reverse Transcriptase Buffer (Invitrogen), 100 U of SuperScript IV Reverse Transcriptase Enzyme (Invitrogen), 0.225 μg of random hexamers (Gene Link), 0.33 mM dNTPs (New England Biolabs), 4.3 mM DTT, and the rest supplemented by RNAse Free Water.. RT PCR used the following PCR program: 10min at 42°C, 10 min at 25°C, 10 min at 50°C, followed by 5 mins at 94°C. Nested PCRs were done to amplify heavy chain and lambda light chain variable regions using multiple primers. PCR1 used primers CowVHfwd1: CCCTCCTCTTTGTGCTSTCAGCCC, CowIgGrev1: GTCACCATGCTGCTGAGAGA, and CowIgGrev2: CTTTCGGGGCTGTGGTGGAGGC. PCR1s were done in 20 μl reactions per well/single cell using 3 μl of RT PCR product, 1x HotStarTaq Mastermix (New England Biolabs), 0.2 μM of forward primers total, and 0.2 μM of reverse primers total, with the rest supplemented by RNAse Free Water. The following PCR program was used: 15s at 95°C, 49 cycles of 30s at 94°C, 30s at 55°C, and 60s at 72°C followed by an extension for 10 min at 72°C. PCR2 was done using heavy chain primers CowVHPCR2:CATCCTTTTTCTAGTAGCAACTGCAACCGGTGTACATTCCMAGGTGCAGCTGCRGGAGTC and CowIgGRevPCR2: GGAAGACCGATGGGCCCTTGGTCGACGCTGAGGAGACGGTGACCAGGAGTCCTTGGCC. Light chains were amplified with primers L Leader 2 F: CACCATGGCCTGGTCCCCTCTG, L Leader 15F: GGAACCTTTCCTGCAGCTC, L Leader 16 F: GCTTGCTTATGGCTCAGGTC, L Leader 35F: GACCCCAGACTCACCATCTC, L Leader 45 F: AGGGCTGCGGGCTCAGAAGGCAGC, L Leader 55 F: CTGCCCCTCCTCACTCTCTGC, and Cow_LC_rev1: AAGTCGCTGATGAGACACACC. PCR2 was done for the light chain using primers Fwd-PCR2-LC: CATCCTTTTTCTAGTAGCAACTGCAACCGGTGTACACCAGGCTGTGCTGACTCAG and LC_REV_const-PCR2: GTTGGCTTGAAGCTCCTCACTCGAGGGYGGGAACAGAGTG. The PCR 2 was used to add tags for Gibson assembly. PCR2 introduced homology to the cut ends of the variable regions for later cloning and were done in 25 μl reactions per well/single cell using 2 μl of PCR product from PCR1, 0.5 U of Phusion Enzyme (Thermo Fisher), 0.2 mM dNTPs at 10 mM each (New England Biolabs), 1 μM of forward primers total, 1 μM of reverse primers total,1.5 mM MgCl₂, 1x HF Buffer (Thermo Fisher), and the rest supplemented with RNAse Free Water using the following PCR program: 30s at 98°C then 34 cycles of 10s at 98°C, 40s at 72°C, followed by an extension for 5 min at 72°C. PCR products were resolved with E-gel 96 Agarose Gels 2% (Invitrogen) to determine if PCR chains were properly amplified. PCR products were purified using SPRIselect beads (Beckman Coulter) and amplified variable regions were sanger sequenced and results were analyzed using the International ImMunoGeneTics (IMGT) Information System (www.imgt.org) V-quest [72]. All isolated variable regions from PCR products were constructed into human antibody expression vectors with appropriate IgG1 or Ig lambda constant domains using high throughput NEBuilder HiFi DNA assembly system. 10 ng of restriction enzyme digested backbone was combined with 20 ng of SPRI cleaned PCR product followed by the appropriate volume of 2x HiFi master mix (New England Biolabs). Gibson assembly reactions were transformed into DH5 Alpha Competent Cells (Biopioneer) and single cultures were grown up and purified using QIAprep Spin Miniprep Kit (Qiagen).

Antibody and Fab production and purification.

Monoclonal antibodies were transiently expressed in Expi293 and 293F cells. Monoclonal cow antibodies were expressed with bovine V_L, V_H, and human C_L, C_H regions. In Expi293 cells heavy chain and light chain pairs were co-transfected at a ratio of 1:2.5 with FectoPRO (Polyplus) in Opti-MEM (Thermo Fisher) or Transfectagro (Corning). After 22–24 hr cultures were supplemented with 300 mM of sodium valproic acid solution (Sigma-Aldrich) and 45% D-(+)- glucose solution (Sigma-Aldrich). In 293F cells, heavy chain and light chain DNA were added in a 1:1 ratio with PEI as a transfection reagent in Opti-MEM (Thermo Fisher) or Transfectagro (Corning). The supernatant was collected 4–5 days after transfection and sterile filtered through an 0.22 or 0.45 mm filter. Sterile supernatants were used for screening or whole IgGs were purified using Protein A Sepharose (GE Healthcare) or Praesto AP (Purolite).

Fabs were produced either with papain digestion or recombinantly. For digestion from monoclonal antibodies, Fabs were prepared with the Pierce Fab Preparation Kit (Thermo Fisher) following the manufacturer’s protocol for human Fab digest. Fab from polyclonal IgG purified from sera were prepared using Pierce Fab Preparation Kit with each digest at 0.5 mg/mL and a digest time of 16 hr.

Fabs with bovine V_L, V_H, and human C_L, C_H1 regions were expressed by transient transfection in Expi293F cells (Thermo Fisher) following the manufacturer’s standard protocol. Fabs were purified from the media using CaptureSelect CH1-XL resin (Thermo Fisher) and further purified by size exclusion chromatography on a Superdex 200 16/600 column (Cytiva) with running buffer 20mM Tris, 150mM NaCl, pH 8.

High throughput antibody screening.

Antibodies were transiently expressed in 3 mL cell cultures of Expi293 cells using 24 deep-well plates (Thermo Fisher). After 4 days, supernatants were harvested and filtered with 0.22 or 0.45 μm membrane filters and were used neat for initial screening of expression, SOSIP binding, and neutralization. We used an antibody detection ELISA to determine if antibodies were expressed and able to bind to trimer (see ELISA section). If antibodies were expressed, neat supernatant was tested for neutralization. Neutralization assays were done as previously described using autologous CRF250 and C108 viruses and 12-virus panel virus CNE55. If any neutralization was detected, antibodies were expressed in 30 to 50 mL cultures and purified using Protein A Sepharose (GE Healthcare) or Praesto AP (Purolite). Purified monoclonal antibodies were tested for neutralization and subsequent analysis. IgHV1-7 derived heavy chains were paired with universal V30 light chains and native light chains if they were recovered. Others were only screened with native light chain when recovered [25,44]. The neutralization was not different between purified monoclonals tested with native and universal light chain and all large-scale production of Bess monoclonals were expressed with V30 light chains. ElsE antibodies were all expressed and further tested with their native light chains.

BioLayer interferometry (BLI) assays.

BLI assays were performed using polypropylene black 96-well microplate (Greiner) on an Octet RED384 instrument at 30°C, in the octet buffer comprising PBS with 0.05% Tween. All reagents were diluted with octet buffer. Antibody competition assays were done in-tandem to determine the binding epitopes of isolated monoclonal antibodies. 222 nM of biotinylated 25710 SOSIP was captured with Streptavidin (SA) biosensors (Sartorius) for 10 min and transferred to Octet buffer for 1 min to wash off unbound SOSIP. Sensors were then moved into monoclonal Bess and ElsE antibodies at a concentration of 400 nM for 10 min and washed off in octet buffer for 1 min. Biosensors were then moved into known antibodies at a concentration of 200 nM for 5 min. The percent inhibition in binding was calculated with the formula: Percent binding inhibition (%) = 1- (competitor antibody binding response in presence of saturating antibody) / (binding response of the competitor antibody without saturating antibody). BLI assays to determine binding of Bess and ElsE to anti-sulfotyrosine antibodies were performed using polypropylene black 384-well microplate (Greiner) on an Octet RED384 instrument at 30°C in octet buffer. We used anti-human IgG Fc Capture (AHC) biosensors (Sartorius) to capture monoclonal antibodies at a concentration of 50 μg/mL for 5 min and washed with octet buffer. After antibodies were captured, the biosensor was moved into 1:50 diluted mouse Anti-Sulfotyrosine Antibody, Clone Sulfo-1C-A2 (MilliporeSigma) to detect binding for 200 sec. All dilution and incubation steps were performed in 1x PBS with 0.05% Tween. Analysis and graphing were done in Prism software.

ELISA assays.

All ELISA assays were done with half-area 96-well high binding ELISA plates (Corning). All washes were done 3x with PBS containing 0.05% Tween20 and all antibodies, except those used for coating, were diluted in PBS containing 1% BSA and 0.025% Tween20. All reagents added to the plate were in 50 μl volumes per well, except blocking which was done in 150 μl volumes. All steps and incubations were done at room temperature unless otherwise stated. All plates were developed with phosphatase substrate (Sigma-Aldrich) diluted in alkaline phosphatase staining buffer (pH 9.8), according to manufacturer’s instructions. Optical density at 405 nm was read on a microplate reader (Molecular Devices). Sera competition ELISA assay plates were coated with PBS containing 250 ng of anti-6x His tag monoclonal antibody (Invitrogen) and left covered overnight at 4°C. The next day plates were washed and blocked with 3% BSA for 1–2 hr at 37°C. After washing 125 μg of his-tagged BG505 was added to all wells for 1 hr. IgG purified from sera were added using serial dilutions starting at 1:20 with seven 4-fold dilutions and incubated for 1 hr. Biotinylated monoclonal antibodies with known epitopes were then added at a constant concentration, depending on their previously measured EC₇₀ value, for 1 hr. After washing, alkaline phosphatase conjugated to streptavidin at a dilution of 1:1000 (Jackson ImmunoResearch) was added to plates and allowed to incubated for 1 hr. Plates were washed, substrate was added, and plates were read as mentioned previously. Biotinylated antibody EC₇₀’s were measured using the same methods, but without the addition of purified IgG from sera. ELISA assays for gp120 binding were done by coating each plate with PBS containing 250 ng of BG505 gp120 and left covered overnight at 4°C. The next day plates were washed and blocked with 3% BSA for 1–2 hrs. Monoclonal antibodies were added using serial dilutions starting at 20 μg/mL with seven 6-fold dilutions and incubated for 1 hr. After washing, alkaline phosphatase conjugated anti-human F(ab’)₂ secondary antibodies (Jackson ImmunoResearch) at a dilution of 1:1000 were added to plates and incubated for 1 hr. After a final wash, substrate was added, and plates were read as mentioned previously. High throughput antibody expression assays used ELISA for the detection of antibody expression and BG505 SOSIP binding in harvested expi293 transfected supernatant. Plates were coated overnight at 4°C with PBS containing 100 ng of AffiniPure Fragment Goat Anti-Human IgG, F(ab’)₂ (Jackson ImmunoResearch). Next day the plates were washed and blocked with 3% BSA for 2 hr at 37°C. 50 μl of neat supernatant were added to each well for 1 hr. After washing Alkaline phosphatase conjugated anti-human Fc secondary antibodies (Jackson ImmunoResearch) was added at a dilution of 1:1000 for 1 hr. After a final wash, substrate was added, and plates were read as mentioned previously. The trimer binding ELISA assay used plates coated with 100 ng of 6x His tag monoclonal antibodies (Invitrogen) in PBS and left overnight at 4°C. The next day plates were washed then blocked with 3% BSA for 2 hr at 37°C. Plates were washed, then 125 μg of 6x His-tagged BG505 SOSIP were incubated in each well for 1 hr. Plates were washed and neat supernatant from high throughput antibody transfections were added for 1 hr. Plates were washed and alkaline phosphatase conjugated anti-human F(ab’)₂ secondary antibodies (Jackson ImmunoResearch) at a dilution of 1:1000 were added to each well. After a final wash, substrate was added, and plates were read as mentioned previously. All data and graphs were analyzed using Prism software.

Phylogenetic tree building.

Phylogenetic trees were created using Geneious Prime software. The full variable region or the CDRH3 of NC-Cow1, Bess, and ElsE amino acids starting were aligned with Clustal Omega 1.2.3. The alignment was made into a tree using Geneious Tree Builder software. We used Jukes-Cantor for the genetic distance model and the Neighbor-Joining Tree Build Method.

X-ray crystallography.

Recombinant Fabs were screened for crystallization with our robotic Rigaku CrystalMation system and JCSG 1–4 crystal screens at 4° and 20°. Crystallization conditions for crystals used for data collection are included in S6 Table. Data were collected at synchrotron beamlines at SSRL, ALS, and APS (S6 Table) and processed with HKL-2000 [73]. Structures were determined by molecular replacement using model coordinates from Fab BLV1H12 (PDB 4K3D) and Phaser [25,74]. Models were adjusted and rebuilt in Coot and refined with Phenix.refine [75,76]. Final data collection and refinement parameters are listed in S6 Table.

Electron-microscopy polyclonal epitope mapping (EMPEM).

Polyclonal IgG was isolated and digested to Fab using papain as described above (Sera IgG purifications and Antibody and Fab production and purification). 1 mg of polyclonal Fab from either cow-485 or cow-488 was incubated with 15 μg of BG505 SOSIP.664 overnight and purified the next morning using a Cytiva Superdex 200 Increase size-exclusion column. Fractions corresponding to immune complexes were pooled and adsorbed onto 400-mesh carbon-coated copper grids (Electron Microscopy Sciences) at a concentration of ~0.02 mg/mL for 10 s before blotting off the excess liquid. Grids were stained for 45 s using 2% (w/v) uranyl formate (Electron Microscopy Sciences) before blotting. Imaging was performed using an FEI Tecnai TF20 transmission electron microscope operating at 200 kV, equipped with a 4K TVIPS TemCam Tietz F416 camera (62,000x magnification, 1.774 Å pixel size). Automated data collection was performed using the Leginon suite and data processing was performed using Relion 3.0, following standard procedures[77,78].

Cryo-electron microscopy.

0.2 mg of BG505 SOSIP was incubated with 0.15 mg of Bess4 Fab (from this study) and 0.3 mg of base-directed RM20A3 Fab (to increase angular sampling in cryo-EM) overnight at room temperature and purified the following morning using a HiLoad 16/600 Superdex 200 pg (Cytiva) gel filtration column [79]. The complex was then concentrated to 5 mg/mL for application onto cryoEM grids. Cryo grids were prepared using a Vitrobot Mark IV (Thermo Fisher Scientific). The temperature was set to 4°C and humidity was maintained at 100% during the freezing process. The blotting force was set to 1 and wait time was set to 10 s. Blotting time was varied from 3 to 4 s. Detergents lauryl maltose neopentyl glycol (LMNG; Anatrace) or n-Dodecyl-β-D-Maltoside (DDM; Anatrace) at final concentrations of 0.005 or 0.06 mM, respectively, were used for freezing. Quantifoil R 1.2/1.3 (Cu, 300-mesh; Quantifoil Micro Tools GmbH) or UltrAuFoil 1.2/1.3 (Au, 300-mesh; Quantifoil Micro Tools GmbH) grids were used and treated with Ar/O₂ plasma (Solarus plasma cleaner, Gatan) for 8 sec before sample application. 0.5 μL of detergent was mixed with 3.5 μL of samples and 3 μL of the mixture was immediately loaded onto the grid. Following blotting, the grids were plunge-frozen into liquid nitrogen-cooled liquid ethane. Samples were loaded into a Thermo Fisher Scientific Titan Krios operating at 300 kV. Exposure magnification was set to 130,000x with a pixel size at the specimen plane of 1.045 Å. Leginon software was used for automated data collection [77]. Micrograph movie frames were motion corrected and dose weighted using MotionCor2 and imported into cryoSPARC for the remainder of data processing. CTF correction was performed using cryoSPARC Patch CTF [80,81]. Particle picking was performed using blob picker initially followed by template picker. Multiple rounds of 2D classification and 3D ab-initio reconstruction were performed prior to 3D non-uniform refinement with global CTF refinement. Due to substoichiometric binding and high flexibility of the Bess4 antibody with respect to the trimer, particles were symmetry expanded (C3 symmetry), subjected to 3D Variability with a mask over the trimer apex, and 3D Variability clustering analysis. Clusters with visible Fab density were pooled, duplicate particles removed, and a final Non-Uniform Refinement was performed (C1 symmetry) with global resolution estimated by FSC 0.143. Final data collection and processing stats are summarized in S8 Table. The Bess4 Fab x-ray structure (this study) was docked into the EM map along with a model of BG505 SOSIP in complex with RM20A3 (PDB 6x9r) in UCSF Chimera [82]. The missing CDRH3 residues from the x-ray structure were built manually in Coot 0.9.8 and real space refinement using Rosetta and Phenix [83–85]. The final model includes 41 of the 55 CDRH3 residues of Bess4, and all other regions of the Fab were trimmed due to lower map resolution resulting from high flexibility. The final model was validated using MolProbity and EMRinger in the Phenix suite, and statistics are summarized in S8 Table. The map and model have been deposited to the Electron Microscopy Data Bank and Protein Data Bank, respectively with accession codes summarized in S8 Table.

Production of cDNA from PBMCs.

Total RNA was extracted from PBMCs using RNeasy mini kit (Qiagen) from D359. Following RNA isolation, cDNA was made using Superscript IV reverse transcriptase (Thermo Fisher). For the cDNA reaction, 549 ng (3 μL) of RNA was used as template, along with 167 nM (1 μL) of JH Bov rvs primer (TGAGGAGACGGTGACCAGGAGTC), 1 μL dNTP, and 9 μL of DNase and RNase-free water. This mixture was incubated at 65°C for 5 min. 5X buffer was vortexed and spun down, after which 4 μL of 5X buffer was mixed with 1 μL DTT, 1 μL ribonuclease inhibitor, and 1 μL SuperScript IV. All 8 components were then mixed, followed by incubation at 52°C for 10 min and then 80°C for 10 min.

PCR and sequencing.

PCR to amplify cow antibody heavy chains was set up as follows: 2.5 μL of 10 μM forward primer, 2.5 μL of 10 μM reverse primer, 2 μL of cDNA, 25 μL of 2X Q5 Hot-start High-fidelity Master Mix (NEB), 18 μL of DNase and RNase-free water. The cycling parameters were as follows: 98°C for 1 min, 98°C for 10 s, 72°C for 1 min, repeat Steps 2 and 3 34 more times, 72°C for 2 min, hold at 4°C. The CDR1F forward primer, which anneals to VH1-7, has the following sequence: TTGAGCGACAAGGCTGTAGGCTG. The VH4 forward primer, which anneals to all cow VH genes, has the following sequence: CTGGGTCCGCCAGGCTCC. The JH Bov rvs reverse primer (see above), which anneals to cow JH2-4, was used as the reverse primer in all reactions. Following PCR, the products were purified using the QIAquick PCR Purification Kit (Qiagen) according to the manufacturer’s protocol. PCR products were then sent to GeneWiz for Amplicon-EZ sequencing, which was performed on an Illumina platform to generate ~100,000 2 X 250 bp paired-end reads per sample.

Computational analysis of sequences.

Once sequences were obtained from GeneWiz, non-functional sequences were removed using SAS. Sequences were considered non-functional if they did not start with the conserved VRXA motif, where X represents any amino acid, or if they did not contain the conserved WG motif at the end of CDRH3. In addition, any sequence was considered non-functional if it contained a premature stop codon in the VH region or if it was less than 220 bp (73 amino acids) long. Using the R computer programming language, the abundance of each amino acid sequence was determined. Analysis of CDRH3 sequences and ultralong antibodies was also performed in R.