key: cord-312275-plqturzi authors: Nielsen, Sandra C.A.; Yang, Fan; Jackson, Katherine J.L.; Hoh, Ramona A.; Röltgen, Katharina; Jean, Grace H.; Stevens, Bryan A.; Lee, Ji-Yeun; Rustagi, Arjun; Rogers, Angela J.; Powell, Abigail E.; Hunter, Molly; Najeeb, Javaria; Otrelo-Cardoso, Ana R.; Yost, Kathryn E.; Daniel, Bence; Nadeau, Kari C.; Chang, Howard Y.; Satpathy, Ansuman T.; Jardetzky, Theodore S.; Kim, Peter S.; Wang, Taia T.; Pinsky, Benjamin A.; Blish, Catherine A.; Boyd, Scott D. title: Human B cell clonal expansion and convergent antibody responses to SARS-CoV-2 date: 2020-09-03 journal: Cell Host Microbe DOI: 10.1016/j.chom.2020.09.002 sha: doc_id: 312275 cord_uid: plqturzi B cells are critical for the production of antibodies and protective immunity to viruses. Here we show that patients infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) who develop coronavirus disease 2019 (COVID-19) display early recruitment of B cells expressing a limited subset of IGHV genes, progressing to a highly polyclonal response of B cells with broader IGHV gene usage and extensive class switching to IgG and IgA subclasses with limited somatic hypermutation in the initial weeks of infection. We identify convergence of antibody sequences across SARS-CoV-2 infected patients, highlighting stereotyped naïve responses to this virus. Notably, sequence-based detection in COVID-19 patients of convergent B cell clonotypes previously reported in SARS-CoV infection predicts the presence of SARS-CoV/SARS-CoV-2 cross-reactive antibody titers specific for the receptor-binding domain (RBD). These findings offer molecular insights into shared features of human B cell responses to SARS-CoV-2 and other zoonotic spillover coronaviruses. The novel human severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the 47 overrepresented in the total IGH repertoires in the COVID-19 patients during acute infection. 115 Preferential selection of B cells using particular IGHV is reported in other contexts, such as 116 increased IGHV1-69 in response to some influenza virus antigens (Avnir et al., 2016). Highly 117 utilized IGHV genes in IgG-seroconverted COVID-19 patients show low median IgG1 SHM 118 (range 2.4-7.5%), compared to higher median IgA1 SHM (range 5.8-9.5%) ( Figure 2C ). 119 In COVID-19 patients, expanded clones detected in two or more replicate IGH sequence libraries 120 identify putative SARS-CoV-2-specific antibody signatures, we analyzed clones with shared 142 IGHV, IGHJ, and CDR-H3 region length, and clustered the CDR-H3 sequences at 85% amino 143 acid identity to find clusters spanning two or more COVID-19 patients and absent from the 114 144 HHC individuals. 1,236 convergent clusters met these criteria and showed SHM frequencies 145 averaging 1.7% (range 0.5% to 5.5%). An average of 196 convergent clusters were found per 146 patient, ranging from 69 clusters in patient 7485 to 477 clusters in patient 7455. 1,171 clusters 147 were shared pairwise between two patients, 53 clusters spanned three patients, nine clusters 148 spanned four patients, and three clusters spanned five patients ( Figure 3A ). To assess the 149 significance of these convergent clones in COVID-19 patients, we undertook repeated 150 resampling of 13 randomly selected HHC samples with the same parameters for 100 iterations. 151 The number of convergent clones shared by COVID-19 patients greatly exceeded the mean 152 convergent clone counts from the HHC resampling ( Figure 3B ), consistent with antigen-driven 153 shared selection of the convergent clones identified in COVID-19 patients. To directly test the 154 antigen specificity of these convergent clones, we expressed human IgG1 monoclonal antibodies 155 (mAbs 2A and 4A, Table S2 ) from two COVID-19 convergence groups in two patients 156 following paired immunoglobulin heavy and light chain sequencing of single B cells with the immunosorbent assay (ELISA) testing, both mAbs bound SARS-CoV-2 spike and S1 domain, 160 but not the RBD or nucleocapsid ( Figure 3C ). We further identified ten sequence clusters shared 161 with independent external COVID-19 patient data sets and reported to be specific for SARS-162 With IGH from two COVID-19 patients (7455 and 7480), we expressed four mAbs (CoV.1-168 CoV.4) from a cluster that is convergent with BD-494 (Yuan et al., 2020) , pairing these with the 169 germline-reverted BD-494 light chain and reported CDR-L3. All four mAbs bound SARS-CoV-170 2 spike protein, S1, and RBD, but not nucleocapsid ( Figure 3C ). BD-494 belongs to a cluster of 171 SARS-CoV-2 RBD-specific antibodies (BD-500, BD-506, and BD-507) using IGHV3-53 and 172 IGHJ6 with CDR-H3 length of 11 amino acid residues ( Figure 3D ). IGHV3-53 antibodies are 173 reported to contact RBD with several germline-encoded residues in CDR-H1 and CDR-H2, as 174 well as residues in CDR-H3 (Yuan et al., 2020) . Contact residues in CDR-H1 and CDR-H2 had 175 the germline IGHV-encoded sequence in CoV.2-CoV.4, while CoV.1 had a T/I change in CDR-176 H1, and overall IGHV SHM frequency of 4.4% ( Figure 3E ). CoV.1-CoV.4 each differed from 177 BD-494 by one amino acid residue in CDR-H3. 178 We hypothesized that in addition to sharing common RBD-binding antibody types, some 179 COVID-19 patients might also demonstrate breadth in their antibody responses and recognize 180 antigens from the distinct but related sarbecovirus, SARS-CoV that was responsible for SARS. Figure S4B and C). These SARS-CoV-specific convergent IGH 184 sequences were not detected in the 114 HHC samples. To evaluate whether such in silico IGH 185 sequence comparisons could predict the serological responses of patients, we tested the plasma 186 samples from COVID-19 patients in SARS-CoV RBD ELISAs and detected cross-reactivity in 187 five of the 13 patients ( Figure 3F) or RBD antigens will also stimulate B cells expressing these common antibody types in a 214 significant fraction of the human population. A subset of patients also had B cell clones 215 expressing convergent IGH to previously described SARS-CoV RBD antibodies; strikingly, the 216 patients with these SARS-CoV-2/SARS-CoV clonotypes also had the highest SARS-CoV RBD 217 binding serum antibody IgG levels. This association suggests that it may become possible to 218 predict the fine specificity of human serological responses from IGH sequence data, as the 219 number of documented antigen-specific clonotypes in public databases increases. This example 220 also highlights the possibility that common modes of human antibody response may enable some histogram shows the distribution of convergent clones shared by HHC samples over these trials. 316 The black dashed line is the mean value, and the wide red dashed line is the number of 317 convergent clones shared among the 13 COVID-19 participants. (C) ELISA results for mAb 318 binding to SARS-CoV-2 spike protein (Spike), S1 domain (S1), RBD, and nucleocapsid. Purified 319 mAbs (mAb2A, mAb4A, B8, C11, D6, E7, F6, F10, mAb CR3022) or HEK 293 cell 320 supernatants (CoV.1-CoV.4) were tested. mAbs generated from COVID-19 patient convergent 321 clones (red) or clones convergent with BD-494 (green) were compared to negative control mAbs 322 B8, C11, D6, E7, F6, and F10 (black). The SARS-CoV-2 RBD-binding mAb CR3022 (purple) 323 and pooled SARS-CoV-2 patient plasma (Plasma1.pos and Plasma3.pos) were positive controls. 324 Purified mAbs were tested at 50 µg/mL except for mAb CR3022, which was tested at 0.506 325 µg/mL. (D) CDR-H3 AA sequence logos from anti-SARS-CoV-2 convergent IGH. mean OD values are shown. See also Figure S4 and Clones with <1% mutation were defined as unmutated and clones with ≥ 1% were defined as 463 being mutated. Subclass fractions were determined for each subject by dividing the number of 464 clones for a given subclass by the total number of clones for that isotype category. Expanded 465 clones within each sample were defined as clones that were present in two or more of the gDNA 466 replicate libraries. Clonal expansion in the isotype data was inferred from the gDNA data. 467 Analyses were conducted in R (Team, 2017) using base packages for statistical analysis and the 468 ggplot2 package for graphics (Wickham, 2016) . ELISA conditions for mAbs were as described for COVID-19 plasma samples with the 548 following modifications. Purified mAbs were tested at 50 µg/mL for intra-COVID-19 549 convergent antibodies or peanut-specific negative mAb controls, or at 0.506 µg/mL for mAb 550 CR3022; antibody supernatants were diluted 1:1 with PBST plus 1% milk; plates were coated 551 overnight with RBD (0.1 µg per well), S1 (0.5 µg per well), spike protein (0.1 µg per well), or 552 nucleocapsid (0.1 µg per well); and bound mAbs were detected with rabbit anti-human IgG 553 gamma chain-specific/HRP (Agilent: cat. P0214, 1:15,000 dilution). 554 Statistical tests were performed in R (Team, 2017) using base packages for statistical analysis 557 and the ggplot2 package was used for graphics (Wickham, 2016) . Box-whisker plots show 558 median (horizontal line), interquartile range (box), and 1.5 times the interquartile range 559 (whiskers). Linear regression lines were plotted with 95% confidence intervals calculated using 560 the function stat_smooth. In all analyses where statistical significance was tested, significance 561 was defined as: ***p-value < 0.001; **p-value < 0.01; *p-value ≤ 0.05; not significant (NS): p-562 value > 0.05. Differences between the seropositive group and HHC (Figure 2A ) was tested using 563 two-sided Wilcoxon-Mann-Whitney (for patients with more than one sample, the mean value of 564 these was used). Patient IGHV gene SHM for seroconverted and non-seroconverted samples 565 were compared to HHC ( Figure 2D ) using paired Wilcoxon tests with Bonferroni correction for 566 multiple hypothesis testing. Differences between the expanded/non-expanded groups and HHC 567 ( Figure 2F ) were tested using one-way ANOVA with Tukey's HSD test. IGHM IGHD IGHG3 IGHG1 IGHA1 IGHG2 IGHG4 IGHA2 Fraction of unmutated clones 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 IGHM IGHG3 IGHG1 gDNA IGHG1 IGHD gDNA IGHM IGHD IGHG3 IGHG1 IGHA1 IGHG2 IGHA2 gDNA IGHM IGHD IGHG3 IGHG1 IGHG2 gDNA IGHG3 IGHG1 IGHG2 IGHM IGHG1 IGHD IGHG3 IGHG1 IGHG2 gDNA IGHM IGHD IGHG3 IGHG1 IGHG2 gDNA IGHM IGHG3 IGHG1 IGHM IGHG3 IGHG1 IGHG2 gDNA IGHM IGHG1 IGHM IGHG1 IGHM IGHG3 IGHG1 IGHG2 gDNA IGHM IGHG3 IGHG1 IGHA1 IGHG2 IGHA2 IGHG1 IGHG2 IGHM IGHG3 IGHG1 IGHG2 gDNA IGHM IGHG1 IGHM IGHM IGHG1 IGHG3 IGHG1 IGHG2 gDNA IGHM IGHG1 IGHG2 IGHG3 Gene utilization frequencies for each isotype subclass for each IGHV gene were compared 41 between COVID-19 and HHC (paired Wilcoxon tests with Bonferroni correction for multiple 42 hypothesis testing). Where gene usage differed significantly between the cohorts and the gene 43 was utilized at a higher median frequency among the COVID-19 patients the adjusted 44 significance is shown in blue, whereas genes that differed significantly but with lower median 45 usage for the COVID-19 patients relative to the HHC are plotted in red. Instances with 46 insufficient data for test (only 1 or no data points) are in white. The plot y-axes were chosen to 47 show the box-whiskers on a readable scale; rare outlier points with extreme values are not shown 48 but were included in all analyses. 49 Isotype IGHM IGHD IGHG3 IGHG1 IGHA1 IGHM IGHD IGHG3 IGHG1 IGHM IGHD IGHG3 IGHG1 IGHA1 IGHG2 IGHG4 IGHA2 0 5 10 15 20 Mean IGHV SHM (%) *** *** *** *** *** *** *** *** *** *** *** *** * IGHM IGHD IGHG3 IGHG1 IGHA1 IGHG2 IGHG4 IGHA2 0 10 20 30 40 IGHM IGHD IGHG3 IGHG1 IGHA1 IGHG2 IGHG4 IGHs. For each set of convergent IGH the sequence logo and alignment for the reported antigen-79 specific CDR-H3 is shown at the top, and sequence logos for clones from each patient are 80 aligned below (colored black where they match a conserved residue in the reported CDR-H3, 81 colored for non-conserved as depicted in the alignment, or gray if no match). The read count per 82 patient that contributed to the sequence logo is plotted for each isotype, with the SHM frequency 83 shown beside the bar. 84 CD27dull and CD27bright B Cells Ensures the Flexibility, Stability, and Resilience of Human B 603 Cell Memory Sample Pooling as a Strategy to Detect 605 Community Transmission of SARS-CoV-2 Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. The 608 Lancet Human responses to influenza vaccination show 611 seroconversion signatures and convergent antibody rearrangements Human neutralizing antibodies elicited by SARS-CoV-2 infection FLASH: fast length adjustment of short reads to improve 616 genome assemblies Shaping of infant B cell receptor repertoires 619 by environmental factors and infectious disease Tiger team: a panel 622 of human neutralizing mAbs targeting SARS-CoV-2 spike at multiple epitopes. bioRxiv Convergent antibody signatures in human 626 dengue Coronavirus as a possible cause of severe acute 629 respiratory syndrome Profiling the 631 TRB and IGH repertoire of patients with H5N6 Avian Influenza Virus Infection by high-632 throughput sequencing Cross-neutralization of SARS-CoV-2 by a 635 human monoclonal SARS-CoV antibody Convergent antibody responses to SARS-638 CoV-2 in convalescent individuals A human neutralizing antibody targets the receptor binding site of SARS-CoV-2 SARS-CoV-2 Seroconversion in 645 Humans: A Detailed Protocol for a Serological Assay, Antigen Production, and Test Setup Genetic Recombination, and Pathogenesis of Coronaviruses Rapid generation of neutralizing 652 antibody responses in COVID-19 patients R: A Language and Environment for Statistical Computing Antibody Combination against SARS Coronavirus: Synergy and Coverage of Escape Mutants Design and standardization 660 of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene 661 recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action 662 BMH4-CT98-3936 ggplot2: Elegant Graphics for Data Analysis Neutralizing antibody responses to SARS-CoV-2 in a COVID-19 recovered patient 666 cohort and their implications. medRxiv A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding 669 to its receptor ACE2 IgBLAST: an immunoglobulin variable 671 domain sequence analysis tool Structural basis of a shared antibody response to SARS-CoV-2. Science Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia Cutting Edge: Ig H Chains Are Sufficient to Determine 679 Most B Cell Clonal Relationships A Novel Coronavirus from Patients with Pneumonia in China Related to Figure 2. (A) Mean IGHV SHM of 53 expanded and non-expanded clones in COVID-19 patients compared to HHC. SHM frequency 54 for each isotype in each sample was summarized as the median SHM of reads expressed as the 55 Points are jittered on the x-axis to decrease over-plotting of samples with the same value (y-57 axis). p-values were calculated by two-sided Wilcoxon-Mann-Whitney tests Points indicate seronegative (red) or seropositive (blue) for the isotype (IgM 61 serology for IgM; IgG serology for IgG subclasses; and IgA serology for IgA subclasses), with 62 gray for HHC. IgD is indicated in purple as serology was not tested (NT). (B) Fraction of 63 mutated expanded clones (y-axis) by seroconversion status (x-axis). Points are colored by 64 participant. p-values were calculated by Fisher's exact tests. (C) Mean CDR-H3 charge and 65 aromaticity of expanded clones in COVID-19 patients. Percent aromaticity (left panel) and mean 66 charge (right panel) for CDR-H3 amino acid residues. p-values were calculated by one-way 67 ANOVA (D) Odds ratio (OR) of IGHV gene usage in expanded clones compared to 69 non-expanded clones in COVID-19 patients. Each dot is the OR value for each IGHV gene, the 70 bars represent confidence interval. The plot y-axis was chosen to show the points on a readable 71 scale; extreme values are not shown but were included in all analyses. Each point was colored by 72 p-value tested by Fisher's exact test J o u r n a l P r e -p r o o f