The rapid and highly parallel identification of antibodies with defined … – Nature.com

Phage library construction and expansion

Methods for scFv phage display library construction and expansion were described previously28. Briefly, SS320 bacteria (Lucigen, 60512-2) grown to mid-log phase were supplemented with 2% glucose and infected with M13K07 Helper phage (Antibody Design Labs, PH010L) and library phage. After 1h, cells were resuspended in 2xYT medium with carbenicillin (100g/mL), kanamycin (50g/mL), and IPTG (50M) and grown overnight at 30C. The following morning, the bacterial culture was pelleted at high speed (12,000g) to obtain clarified supernatant. The phage-laden supernatant was precipitated on ice for 40min with a 20% PEG-8000/2.5M NaCl solution at a 1:4 ratio of PEG/NaCl:supernatent. After precipitation, phage was centrifuged at 12,000g for 40min and resuspended in 1X TBS, 2mM EDTA. Phage from multiple tubes were pooled, re-precipitated, and resuspended to an average titer of 110cfu/mL. The precipitated phage represents ~250-fold coverage of the library.

For round 1 of cell-based panning, 100106 CFPAC parental cells (ATCC, CRL-1918) were harvested and washed in PBS. The cells were directly resuspended in the pooled phage library and incubated at 4C for 1h with rotation. For the A3 clone spike-in condition, phage clone A3-Clone 20 was initially added to the pooled library for a final amount of 1:100000 phage. After pelleting, the cells were washed 5X in PBS. To recover the phage for reinfection, 1mL of Elution Buffer (0.2M glycine at pH 2.2) was used to resuspend cells for 10min at 20C. The solution was neutralized with 150L of Neutralization Buffer (1M Tris HCl at pH 9.0). The neutralized phage were then used to infect a 10mL SS320 culture at mid-log phase with the addition of M13K07 helper phage (MOI of 4) and 2% glucose. As previously described above, bacteria were grown overnight and phage particles in conditioned medium were precipitated with PEG/NaCl. Starting in round 2 and for the remainder of biopanning, 15106 CFPAC HLA A3 knockout cells (custom in-house line generated from parental) were washed in PBS and used for negative biopanning. Like positive biopanning, the phage was incubated for 1h at 4C with rotation before cells were pelleted and the supernatant containing unbound phage was used for positive biopanning. The amount of CFPAC parental cells used for positive biopanning decreased with each round: 15106 cells for round 2, 10106 cells for round 3, and 5106 cells for round 4. At the end of each round, cells were pelleted and phage particles were eluted for infection of overnight SS320 cultures.

His-tagged SARS-CoV-2 and MERS antigens (RBD, S1, and FL) were conjugated to HisPurTM Ni-NTA magnetic beads (Thermo Fisher, 88832) at a concentration of 16g protein/mg beads following manufacturers protocol (Supplementary Data9). In order to remove any phage recognizing a portion of the magnetic beads itself, the precipitated library was applied to naked 500L washed Ni-NTA magnetic beads overnight at 4C with rotation. After negative biopanning, the supernatant containing unbound phage was applied to beads conjugated with 10g of each SARS-CoV-2 protein (RBD, S1, FL) for a total of 30g protein for 1h at 4C. The beads were washed 3X with TBST (TBS+0.05% Tween-20) with the last wash overnight at 4C. Phage were eluted, neutralized, and used to infect bacteria as previously described above. Beginning in round 2, we separated the precipitated round 1 phage into three different biopanning strategies corresponding to RBD, S1, or FL. We applied an overnight negative biopanning of 30g of the corresponding MERS protein-conjugated beads (i.e. for SARS-CoV-2 RBD panning, use MERS RBD) prior to positive biopanning. This remained constant for rounds 3 and 4. However, the amount of SARS-CoV-2 protein used for positive biopanning decreased from 4g in round 2 to 1g in round 3 to 0.5g in round 4. After each negative biopanning, the sample was applied to a magnet to isolate the supernatant for positive biopanning. After each round, phage were eluted and used to infect bacteria for overnight cultures.

To identify each clone by Sanger sequencing, 1L of bacteria was mixed following manufacturers protocol with Q5 Hot Start High-Fidelity 2X Master Mix (New England Biolabs, M0494X) and forward and reverse primers (phiS2: 5-ATGAAATACCTATTGCCTACGG and psiR2: 5-CGTTAGTAAATGAATTTTCTGTATGAGG). The cycling conditions used are as described:

PCR products were submitted to Genewiz (South Plainfield, NJ) for Sanger Sequencing. Entire sequence of scFv was identified using SnapGene (San Diego, CA).

For each cell line, CFPAC parental and CFPAC A3 KO, 10106 cells were trypsinized and washed 3X with PBS. Cell pellets were resuspended in PBS at a concentration of 10106 cells/mL. Next, 10L of polyclonal phage (1013 titer) was applied to both samples and incubated for 1h at 20C with continuous rotation. Samples were pelleted and washed 3X with PBS. After washing, the samples were resuspended in QuickExtract DNA Extraction Solution (Lucigen, QE09050) according to manufacturers protocol. Samples were passed through the QIAshredder (Qiagen, 79656) and used directly for PCR amplification of CDRH3 region.

His Tag antibody plates (Genscript, L00440C) were individually coated with the spike proteins desired for SLISY comparisons. The spike proteins were diluted in PBS to a concentration of 0.5g/mL and incubated at 4C for 12h. Plates were vigorously washed 6X times with 1X TBST. Next, 100L of polyclonal phage (1013 titer) was applied to each of the wells for comparison and incubated for 1h at 4C. Next, the plates were washed again 6X times with TBST. After washing, 20L of Elution Buffer (0.2M glycine at pH 2.2) was applied to wells for 20min at 20C. Without washing, 3L of Neutralization Buffer (1M Tris HCl at pH 9.0) was directly added to wells and total volume in well (~23L) was recovered for PCR amplification.

Eluted phage was amplified using the following primers (Forward: GGATACCGCTGTCTACTACTGTAGCCG, Reverse: CTGCTCACCGTCACCAATGTGCC) which flank the CDR-H3 region. The sequences at the 5-ends of these primers incorporated molecular barcodes to facilitate unambiguous enumeration of distinct phage sequences. The protocols for PCR-amplification and sequencing are described in Kinde et al. (2011)27. Sequences were demultiplexed and processed to extract and translate CDRH3 regions using custom python software (https://doi.org/10.5281/zenodo.7154344) and analyzed using SQL databases (MSSQL) and Microsoft Excel. The SLISY Binding Ratio was calculated by comparing the phage eluted from the target well/cells to the phage eluted from the non-target (negative) well/cells.

$${{{{{rm{SLISY}}}}}}; {{{{{rm{Binding}}}}}}; {{{{{rm{Ratio}}}}}},({{{{{rm{SBR}}}}}})=frac{{{{{{rm{Clone}}}}}}; {{{{{rm{UID}}}}}}; {{{{{rm{in}}}}}}; {{{{{rm{Target}}}}}}; {{{{{rm{Pool}}}}}}}{{{{{{rm{Clone}}}}}}; {{{{{rm{UID}}}}}}; {{{{{rm{in}}}}}}; {{{{{rm{Negative}}}}}}; {{{{{rm{Pool}}}}}}}$$

Clones that have a UID of zero are given a value of 0.5. Because wells were loaded with equivalent amounts of phage and we were using molecular barcodes to count molecules, we did not normalize for total reads. For comparison, we also calculated a Panning Enrichment Ratio (PER) by dividing the fraction of the phage expressing the specific CDR-H3 in the population after growth of the bound phage by the fraction present in the input material and then converting the value to log base 2.

SLISY initially uses sequencing of just the CDR-H3 region of the scFv to determine whether any promising clones exist in the library following biopanning. If so, we used more extensive sequencing to determine the sequence of the entire scFv. In order to obtain the complete variable regions of the heavy and light chains of the scFv, long-read sequencing utilized a custom protocol with three reads (Supplementary Fig.1). To efficiently multiplex samples, 96 forward primers were designed with unique well barcodes to serve as sample indexes and a Unique IDentifier (UID) of 14 random nucleotides to serve as molecular barcodes (Supplementary Data10). Thus, each primer contained both a well barcode and the means to identify up to 414

unique molecules per well. Each 25l reaction consisted of 12.5l of Q5 High-Fidelity 2X Master Mix (New England Biolabs, Ipswich, MA, cat #0491L), 1l of 10M forward primer (IDT, Coralville, IA), 1l of 10M reverse primer (IDT), 1l of phage, and 9l nuclease-free water. Amplification conditions were as follows: 1 denaturation cycle of 98C for 3min; 10 amplification cycles of 98C for 10s, 61C for 2min, and 72C for 2min; followed by an infinite hold at 4C. Amplicons were then sequenced on a MiSeq using the following custom primers designed to sequence all variable regions of the scFv: read 1 primer 5-ACTGGCCGTCGTTTTACGTCG-3, read 2 primer 5-TGCGCTAATGGTAAAGCGACCTTTCACGCTAT-3, and index primer 5-CTGCAGATGAATAGTCTGCGTGCAGAGGATACAGC-3. Run parameters were as follows: read 1 220 cycles, read 2 135 cycles, and index 145 cycles. Reads passing Illumina chastity filters were included in subsequent analysis.

Based on the long-read sequencing, geneblocks corresponding to the entire scFv were ordered from Integrated DNA Technologies and resuspended at a concentration of 0.05M in TE Buffer. The pADL-10b phagemid vector was digested using BglI enzyme (New England Biolabs, R0143S) with Antarctic Phosphatase (New England Biolabs, M0289L) in manufacturer recommended Buffer 3.1 for 12h at 37C. Following PCR purification (Qiagen, 28104), 100ng of linearized pADL-10b plasmid and 1L of each resuspended geneblock was added to NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs, E2621X) for a reaction total volume of 20L that was incubated at 50C for 1h. One L of the ligation product was mixed with 25L of electrocompetent SS320 cells. This mixture was electroporated using a Gene Pulser electroporation system (Bio-Rad, BZA648860) and allowed to recover in Recovery Media (Lucigen, 80026-1) for 1h at 37C with shaking. Next, 100L of 1:500 dilution of transformed cells were plated on 2xYT agar plates supplemented with carbenicillin (100g/mL) and 2% glucose. Cells were grown at 37C for 12h. Individual colonies were inoculated into 200l of 2xYT medium containing 100g/mL carbenicillin and 2% glucose and grown for three hours at 37C. The cells were then infected with 1.6107 M13K07 helper phage and incubated for an additional 1h at 37C. The cells were pelleted, resuspended in 300L of 2xYT medium containing carbenicillin (100g/mL) and kanamycin (50g/mL), and grown overnight at 30C. Clones were submitted for Sanger sequencing and inserts were confirmed using SnapGene. Once verified, cells were pelleted and 100L of phage-laden supernatant was used to infect 9mL of bacteria to produce phage overnight as described above.

CFPAC parental and HLA-A3 KO lines were harvested, washed with PBS, and resuspended in ice-cold flow cytometry staining buffer (PBS, 0.5% BSA, 2mM, EDTA, 0.1% sodium azide) at a concentration of 10106 cells/mL. Next, 10L of precipitated phage (1013 titer) was applied to 100L of both cells and incubated on ice for 15min. After washing 3X with staining buffer, cell pellets were resuspended to same concentration (10106 cells/mL) and stained with rabbit anti-M13 polyclonal antibody (Novus Biologicals, Littleton, CO) at 1:100 final dilution on ice for 15min. After washing 3X again, cell pellets were resuspended and stained with PE donkey anti-rabbit IgG antibody (Biolegend, San Diego, CA) at 1:100 final dilution for 15min on ice. After a final wash 3X, stained CFPAC cells were analyzed using a LSRII flow cytometer (Becton Dickinson, Franklin Lakes, NJ) to measure mean fluorescence intensity. For gating, any PE signal above background in unstained cells was considered positive (Supplementary Fig.6).

His Tag antibody plates were coated with a 0.5g/mL solution of spike protein antigens diluted in PBS at 4C for 12h and washed with 1X TBST.

To test binding and specificity of phage to spike antigens, precipitated phage (1013 titer) was diluted 1:1000 in PBS and 100L was added to each well and incubated for 1h at 4C. After vigorous washing, the bound phage were then incubated with 100L of rabbit anti-M13 antibody diluted 1:5000 in 1X TBST for 1h at 4C. Following 6 washes with 1X TBST, wells were then incubated with 100L of goat anti-rabbit IgG (H+L) antibody HRP (Thermo Fisher, A27036) diluted 1:10,000 in 1X TBST for 1h at 4C.

Each converted full-length IgG antibody clone was diluted to 1.0ug/mL in PBS and 100L was added to each well with 1h incubation at 4C. Wells were then incubated directly with 100L of goat anti-human IgG Fc-HRP antibody (Abcam, ab98624) diluted 1:10,000 in 1X TBST for 1h at 4C. After 6 more 1X TBST washes, 100L of TMB substrate (Biolegend, 421101) was added and allowed to develop. The reaction was quenched with 50L of 2N sulfuric acid. Absorbance at 450nm and 540nm was measured with a Synergy H1 Multi-Mode Reader (BioTek, Winooski, VT). O.D. readings were measured as absorbance at 450nm minus that at 540nm.

His Tag antibody plates were coated with a 0.5g/mL solution of spike protein antigens diluted in PBS at 4C for 12h then washed with 1X TBST. To test blocking of ACE2 binding, 100L of precipitated phage (1013 titer) or 10g/mL of converted antibody was applied to each well and incubated for 1h at 4C. After vigorous washing, the wells were then incubated with 100L of recombinant ACE2-His protein (RayBiotech Life, Inc, Peachtree Corners, GA) for 1h at 4C. After 6X washing, 100L of rabbit anti-6X His tag polyclonal antibody (Abcam, ab9108) diluted 1:1,000 in PBS was added and incubated for 1h at 4C. Following 6 washes with 1X TBST, wells were then incubated with 100L of goat anti-rabbit IgG (H+L) antibody HRP diluted 1:10,000 in 1X TBST for 1h at 4C. Competitive ELISAs were developed and measured identical to binding ELISAs described above. A negative control antibody, SARS-CoV-2 spike mouse mAb (40591-MM42) (Sino Biological, Wayne, PA) that binds to the SARS-CoV-2 spike protein but does not neutralize, and a positive control antibody, SARS-CoV-2 spike mouse mAb (40592-MM57) (Sino Biological) that blocks interaction with ACE2 were used at 10g/mL. An ACE2-His alone condition without phage or antibody served as the baseline signal for spike-ACE2 binding. Blocking potential was measured as a decrease in signal from ACE2-His alone.

Assays were performed following manufacturers protocols for measuring neutralization using the SARS-CoV-2 sVNT kit (Genscript)31. Briefly, 100L of 1:1 and 1:10 phage (1013 titer) diluted in Sample Dilution Buffer was mixed with diluted HRP-RBD solution with a volume ratio of 1:1. Mixtures were incubated at 37C for 30min. For antibodies, 100L of 1g/mL and 0.1g/mL antibody diluted in Sample Dilution Buffer was mixed with HRP-RBD solution. Provided positive and negative controls were prepared similarly following manufacturer protocols. Next, 100L of mixtures were applied to wells pre-coated with recombinant ACE2 protein and incubated at 37C for 15min. After washing plates with 260L 1X Wash Solution four times, 100L of provided TMB solution was added and the reaction developed for 15min at 25C. Reaction was quenched with Stop Solution and plate read at 450nm and 540nm absorbance. According to manufacturers protocol, neutralization was measured as a decrease in signal greater than or equal to 30% of signal from the negative control well.

$${{{{{rm{Inhibition}}}}}}=left(1-frac{{{{{{rm{OD}}}}}}; {{{{{rm{value}}}}}}; {{{{{rm{of}}}}}}; {{{{{rm{Sample}}}}}}}{{{{{{rm{OD}}}}}}; {{{{{rm{value}}}}}}; {{{{{rm{of}}}}}}; {{{{{rm{Negative}}}}}}; {{{{{rm{Control}}}}}}}right) , times , 100%$$

For IgG Antibodies, the light and heavy chain variable sequences of the scFvs were grafted on the chains of 4D5/trastuzumab and cloned into pcDNA3.4 backbone with a mouse IgKVIII leader signal peptide as we have done before in Hsiue et al. (2021)24. Freestyle 293-F cells were transfected with light and heavy chain plasmids at a ratio of 1:1 using PEI at 1:3 at a concentration of 2106 to 2.5106 per ml and incubated for 6 days at 37oC. Transfection and expression of the antibodies were carried out at the Eukaryotic Tissue Culture Core Facility at Johns Hopkins University. The media was harvested by centrifugation, filtered through a 0.22m PES membrane and purified via protein A affinity chromatography on a HiTrap MabSelect SuRe column (Cytiva, Malborough, MA) with the following running buffer: 20mM Sodium Phosphate, 150mM NaCl pH 7.2. The antibodies were eluted with a linear gradient from 0 to 100mM glycine pH 3.0 over 30 column volumes. Fractions were collected in prefilled tubes with 1M TRIS pH 9.0. The fractions were quantified by SDS-PAGE gel electrophoresis and the fractions of pure antibody were pooled and dialyzed into 20mM Sodium Phosphate, 150mM NaCl pH 7.2. Gel filtration chromatography was used for further purification in 20mM Sodium Phosphate, 150mM NaCl pH 7.2 with the Superdex 200 increase, 10/300 GL column (Cytiva). Final fractions were quantified by SDS-PAGE gel electrophoresis and the fractions containing antibody were frozen with liquid nitrogen and stored at 80C.

SPR experiments were carried out on a Biacore T200 (Cytiva) at 25C of a CM5 chip. Protein A/G was diluted (1:25 dilution, 1M diluted concentration) in 10mM sodium acetate buffer at pH 4.5 and immobilized on all flow cells (Fcs) of the CM5 chip to a level of ~4100 response units (RU) using standard amine coupling chemistry. HBS-P (10mM Hepes pH 7.4, 150mM NaCl, 0.05% v/v surfactant P20) was used as the immobilization and capture running buffer. Approximately ~70-200 RU of each SLISY-selected antibody was captured onto Fcs 2 through 4. Fc1 was used as reference subtraction. Single-cycle kinetics were performed for the analytes binding to the captured ligands in the presence of HBS-P by increasing concentrations (5, 20, 80, 160, and 320nM, four-fold dilutions) of purified target analytes flowed over Fc 1-4 at a rate of 50L/min. The analytes used were SARS-CoV-2 RBD, SARS-CoV-2 S1, and SARS-CoV-2 FL (Supplementary Data9). The contact and dissociation times were 120s and 600s, respectively. One 20s injection of Glycine pH 2.0 was used for surface regeneration. This regeneration also took away captured ligands. Therefore, ligands were captured in the beginning of every cycle. Binding responses for kinetic analyses were reference and blank subtracted. All curves were fit with a 1:1 kinetic binding model using Biacore Insight evaluation software. All SPR measurements were done in triplicates.

SARS-CoV-2 S-protein pseudotyped replication incompetent lentiviral particles were produced by first transfecting HEK-293T (ATCC, CRL-3216) with GeneJuice transfection reagent (Millipore-Sigma) and SARS-Related Coronavirus 2, Wuhan-Hu-1 Spike-Pseudotyped Lentiviral Kit (Spike-Pseudotyped Lentiviral Kit, NR-52948; BEI Resources Repository, Manassas, VA). Viral supernatant was collected 48h after transfection and filtered through a 0.45M filter. For variant viruses, commercially available pseudotyped luciferase rSARS-CoV-2 spike virus for the Beta, Delta, and Lambda strains were obtained from Creative Biosciences (Shirley, NY). 1.25104 293T cells engineered to express hACE2 receptor (hACE2.293T cells) were plated on day 1 in black 96-well microplates (Corning, Corning, NY)37. Parental 293T cells served as a control for nonspecific cell transduction. On Day 0, S-pseudoviral particles+/ the indicated antibodies were added and the plate centrifuged at 800g for 30 at 32C. Optimal amounts of S-pseudoviral particles for each virus were titrated up to maximal signal-to-noise without antibodies. Cells were then incubated at 37C in 5% CO2 for 48h at which time the viral-containing supernatant was aspirated and fresh media containing 150ug/ml D-Luciferin (Millipore-Sigma) for the original strain or Renilla Luciferase (Promega, Madison, WI) for variant viruses was added. BLI was measured and reactive light units (RLU) determined after subtraction of virus-only background. Inhibition of infectivity by each antibody was calculated as percentage decrease in BLI over baseline (virus without antibodies). The assay was performed in experimental triplicate.

An indirect ELISA format was used to identify whether identified neutralizing antibody clones could compete for the same antigenic epitope. Briefly, antibodies were biotinylated and an initial indirect ELISA was performed to calibrate the appropriate concentration (OD value approaches 1.5) of each antibody for the binning assay against the SARS-CoV-2 RBD-His. Antibodies with low binding activity on the RBD-His were excluded from downstream assays due to difficulty in showing self-competition. Biotinylated antibody is mixed with each free antibody with the calibrated concentration through a volume ratio of 1:1 and applied to an indirect ELISA format to analyze competition within each pair. Antibody pairs are categorized as likely having the same epitope if one of the competition values (ODmAb1-mAb2) is lower than the self-competition value of the other antibody (ODmAb2) and the other competition value (ODmAb2-mAb1) demonstrates an inhibition of greater than 30% than that of the biotinylated antibody alone (ODmAb1). Epitope binning assay performed and analyzed by Genscript.

$${{{{{rm{Inhibition}}}}}}_{{mAb}1-{mAb}2}=left(1-frac{{{OD}}_{{mAb}1-{mAb}2}-{{OD}}_{{mAb}2}}{{{OD}}_{{mAb}1}-{{OD}}_{{mAb}2}}right) , times , 100%$$

Unless otherwise indicated, error bars represent the standard deviation of three biological replicates. No statistical method was used to predetermine sample size. Experiments were performed in replicates of three to generate a standard deviation of means. No data were excluded from the analyses.

Further information on research design is available in theNature Portfolio Reporting Summary linked to this article.

Here is the original post:
The rapid and highly parallel identification of antibodies with defined ... - Nature.com