Active evolution of memory B-cells specific to viral gH/gL/pUL128/130/131 pentameric complex in healthy subjects with silent human cytomegalovirus infection

Human cytomegalovirus (HCMV) can cause life-threatening infection in immunosuppressed patients, and in utero infection that may lead to birth defects. No vaccine is currently available. HCMV infection in healthy subjects is generally asymptomatic, and virus persists as latent infection for life. Host immunity is effective against reactivation and super-infection with another strain. Thus, vaccine candidates able to elicit immune responses similar to those of natural infection may confer protection. Since neutralization is essential for prophylactic vaccines, it is important to understand how antiviral antibodies are developed in natural infection. We hypothesized that the developmental path of antibodies in seropositive subjects could be unveiled by interrogating host B-cell repertoires using unique genetic signature sequences of mAbs. Towards this goal, we isolated 56 mAbs from three healthy donors with different neutralizing titers. Antibodies specific to the gH/gL/pUL128/130/131 pentameric complex were more potent in neutralization than those to gB. Using these mAbs as probes, patterns of extended lineage development for B-cells and evidence of active antibody maturation were revealed in two donors with higher neutralizing titers. Importantly, such patterns were limited to mAbs specific to the pentamer, but none to gB. Thus, memory B-cells with antiviral function such as neutralization were active during latent infection in the two donors, and this activity was responsible for their higher neutralizing titers. Our results indicated that memory B-cells of neutralizing capacity could be frequently mobilized in host, probably responding to silent viral episodes, further suggesting that neutralizing antibodies could play a role in control of recurrent infection.

donor 2 (B) and donor 3 (C) were plotted for titration, with left y-axis (blue) shows HCMV virion binding activity, and right y-axis (red) indicates the neutralizing activity. Binding and neutralizing activities were tested as described in Supplementary Materials and Methods. The curves for binding and neutralizing activities are shown in blue and red, respectively. Correlation analysis of neutralizing and binding properties of 56 monoclonal antibodies from donor 1 (D), donor 2 (E) and donor 3 (F), with each mAb are plotted for its IC 50 (y-axis) and virus-specific EC 50 (x-axis) activities. The mAbs with virus-specific EC 50 of > 20μg/ml are designated as no detectable binding activity and arbitrary values of 10 are assigned for plotting. The mAbs with IC 50 of > 20μg/ml are designated as no detectable neutralizing activity and arbitrary values of 20 are assigned for plotting. The red open circle represents the control HCMV-HIG (CytoGam®). www.impactjournals.com/oncotarget Supplementary Figure 4A: Pipeline processing of HCMV donor 1 (Dn1) heavy chain PGM sequencing data set. Read length distribution (step 1). Of 1,388,527 reads, 1,221,842 (or 88.3%) are over 400bp. Germline gene distribution (step 2). Correlation between number of gaps in V gene alignment and sequence quality improvement measured by the increase of sequence identity to germline gene after correction (step 3). The correlation coefficient is 0.66 and P-value is less than 0.0001. Distribution of sequence quality improvement (step 4), with an average of 20.9%. www.impactjournals.com/oncotarget Supplementary Figure 4B: Pipeline processing of HCMV donor 1 (Dn1) Kappa chain PGM sequencing data set. Read length distribution (step 1). Of 1,456,992 reads, 1,156,480 (or 79.5%) are over 400bp. Germline gene distribution (step 2). Correlation between number of gaps in V gene alignment and sequence quality improvement measured by the increase of sequence identity to germline gene after correction (step 3). The correlation coefficient is 0.68 and P-value is less than 0.0001. Distribution of sequence quality improvement (step 4), with an average of 22.8%. www.impactjournals.com/oncotarget Supplementary Figure 4C: Pipeline processing of HCMV donor 1 (Dn1) lambda chain PGM sequencing data set. Read length distribution (step 1). Of 1,146,414 reads, 985,300 (or 86.0%) are over 400bp. Germline gene distribution (step 2). Correlation between number of gaps in V gene alignment and sequence quality improvement measured by the increase of sequence identity to germline gene after correction (step 3). The correlation coefficient is 0.63 and P-value is less than 0.0001. Distribution of sequence quality improvement (step 4), with an average of 22.3%. as a function of sequence identity to the template antibody chains and sequence divergence from putative germline V genes. Color coding denotes sequence density. In panel #2, sequences with the same germline origin as the template antibody chain are plotted as black contour lines, with those having a CDR3 identity of ≥92% (original dots) and ≥95% (magenta dots) labeled on the 2D plot. In panel #3, heavy and light chain trees derived from the sequences with a CDR3 identity of 95% or greater are constructed for antibodies 2-25 and 2-55. Synthesized sequences are shown in black on the maximum-likelihood (ML) trees.

Supplementary Figure 8: Identity/divergence analysis and phylogenetic analysis of the selected genetic variant clones isolated from donor 1 (A, B) and donor 2 (C, D).
Repertoire sequences from donor 1 (A) and donor 2 (C) are plotted as a function of sequence identity to the template antibody chains and sequence divergence from putative germline V genes. Color coding denotes sequence density. Sequences with the same germline origin as the template antibody chain are plotted as black contour lines, with those having a CDR3 identity of ≥92% (original dots) and ≥95% (magenta dots) labeled on the 2D plot. Maximum-likelihood (ML) tree were constructed using heavy chain sequences of selected somatic variants from donor 1 (B) and donor 2 (D) with a CDR3 identity of 95% or greater to the template antibody. Synthesized sequences and templated antibody are shown respectively in black and red on the maximum-likelihood (ML) trees.

Memory B-Cell Culture
A total of 80 mL of whole blood from each donor was collected and processed to peripheral blood mononuclear cells (PBMC) isolation. Briefly, 20 mL of whole blood were added to each 50 mL Accuspin tube (Sigma) pre-filled with 10 mL Hank's solution and centrifuged at 800 x g for 15 minutes at room temperature. The upper plasma portion was removed by aspiration and the mononuclear cell band was transferred to a new 50 mL conical tube. The tubes were filled with Hank's solution and centrifuge at 500 x g for 10 minutes at room temperature. The supernatants were discarded and red blood cells in the pellets were lysed by incubation with 5 mL ACK lyse buffer (Gibco Invitrogen) at room temperature for 5 minutes. Afterwards, cell pellets were washed once with 50 ml Hank's solution and re-suspended in 15 mL R10 medium (RPMI 1640 medium with 20% FBS).
IgG + memory B-cells were enriched using the customized EasySep™ Human Memory B-Cell Isolation Kit containing anti-human IgM and IgD antibodies (StemCell). Briefly, PBMC was re-suspended in RoboSep buffer to 5x10 7 cells/mL in a 5 mL polystyrene round bottom tube. EasySep human memory B-cell enrichment cocktail added at 50µl/mL cells. Following 10 minute incubation at room temperature, EasySep D magnetic particles were added at 75 µl/mL. After additional 5 minute incubation at room temperature, RoboSep buffer was added to the cell suspension up to a total volume of 5 mL. Cells were gently mixed in the tube. The tube was placed in the magnet and set aside for 10 min. Supernatant was poured into a new 5 mL polystyrene tube and placed in the magnet for 5 minutes. The procedure was repeated one more time. Supernatant was then poured into a 15 mL conical tube filled with R10 media up to 15 mL. After centrifugation at 200 x g at room temperature for 10 minutes, supernatant was removed and cells were re-suspended in 0.5 mL of R10 medium. Enriched memory B-cells were then seeded at an average of 1.4 B-cells/well in 96-well U-bottom plates in the presence of 40,000 cells/well of gamma-irradiated 293 cells expressing human CD40 ligand and 50 ng/ml of recombinant human IL-21 (Sino Biological). A total of 200 plates were set-up for each donor and cells were cultured for 14 days in the incubator with 5% CO 2 and 90% humidity. Plates were then centrifuged at 2000 rpm for 15 min and supernatants were carefully transferred into new 96-well U-bottom plates without disturbing the cell pellets. Supernatant containing plates were kept at -80 o C till screened. The cell pellets were preserved in 50 µL RLT (Qiagen) supplemented with 1% β-mercaptoethanol (Sigma) and stored at -80 o C for future RNA isolation.

Screening of Memory B-cell Culture Supernatants
B-cell culture supernatants were first tested for HCMV virion binding and neutralization activities. Samples from the positive hit wells, if there are still supernatants left, were then tested for binding to recombinant HCMV gB and pentameric gH complex proteins, as described previously [1,2]. Briefly, for ELISA assays, antigens were immobilized at 1-2μg/mL in PBS on 96-well Nunc MaxiSorp Immuno plates (ThermoFisher Scientific) at 4°C overnight. Plates were blocked with 3% (vol/vol) nonfat milk in PBS/0.05% Tween-20 and then incubated with undiluted culture supernatants in 40-50 µL volume. Plates were washed afterwards and then HRP-conjugated goat anti-human IgG was added (Southern Biotech, Birmingham, AL) at a dilution of 1:2000. Plates were washed and developed with a HRP substrate, TMB (3,3',5,5'-tetramethylbenzidine) solution (Virolabs) at 100 µL per well for 3 to 5 min. The reaction was stopped by adding 100 µL per well of 1M H 2 SO 4 (Sigma). Plates were read at OD 450nm on VICTOR III Multilabel Counter (PerkinElmer). A sample with OD450 reading equal or greater than 0.1 was scored as positive for that well. For neutralization assay, 50µl of supernatants were mixed with 25 µL of AD169rev (~200 pfu) and incubated at room temperature for 1h. The mixture was then transferred into 96-well flat plates which were pre-seeded with ARPE19 cells at 15,000 cells per well one day before. Plates were incubated at 37°C, 5% CO 2 with 90% humidity for 24h and then stained for the expression of immediate early proteins (IE). The total fluorescence intensity was acquired on Li-Cor Odyssey (LI-COR Biosciences). The wells with >80% reduction in IE expression were scored as neutralization positive hits.
To estimate the concentration of IgG in B cell culture supernatants ( Figure S1B), supernatants from 8-10 plates were randomly chosen for testing in human IgG ELISA assay. Briefly, Nunc Maxisorp Immuno Plates were coated with 50 µL per well of goat anti-human IgG (Fc) antibody (ThermoFisher Scientific) at 4 µg/mL in PBS at 4°C overnight. Plates were washed in PBS/0.05% Tween 20 and then blocked in 3% non-fat milk/PBS at room temperature for 60 min. Blocking buffer was removed and 50 µL per well of B cell culture supernatants was added. Pates were incubated at room temperature for 90 min and washed in PBS/0.05% Tween 20. Afterwards, 50 µL per well of HRP-conjugated goat anti-Human IgG (Southern Biotech) at 1:2,000 dilution was added. Plates were incubated at room temperature for 60 min and washed. TMB solution was then added (100 µL per well). Reaction was stopped 5 min later with 100 µL per well of 1M H 2 SO 4 . Plates were read at 450nm on VICTOR III Multilabel Counter.

Cloning and Analysis of the VH and VL Encoding Genes from Memory B-Cells
Cloning of the V H and V L encoding genes from single human memory B cells was based on a protocol established in our group for cloning of V H and V L encoding genes from single monkey plasma B cells with modifications [3]. Total RNA was isolated from lysed B-cell cultures using the RNeasy Mirco kit (Qiagen). Reverse transcription was carried out using the SuperScript III First-Strand Synthesis SuperMix kit (Invitrogen). Genes encoding IgG heavy and light chain variable regions were amplified by PCR using V H , V k , V λ family-leader region specific primers ( Figure S2), using 3.5 μL cDNA template. All PCR reactions were performed in a total volume of 25 μL containing a mixture of 20nM for each primer, 300nM dNTP, 5μL 5X buffer and 1U Primerstar GXL (Clontech) with 5 min incubation at 94°C for the first cycle and followed by 40 cycles of 98°C, 10s; 55°C, 30s; and 68°C, 30s; and a final elongation step at 68°C for 5 min before cooling to 4°C. One microliter of the PCR product was then used as DNA template in a second round of PCR for the generation of DNA fragment for in-frame cloning of the V H and V L encoding genes in the IgG expression vectors by infusion cloning. The second round PCR products were run on 2% agarose gels and the 400bp V H and V L encoding gene bands were cut and purified for infusion PCR. Infusion PCR was carried out using the In-Fusion® HD Cloning kit (Clontech). Briefly, 50ng of purified PCR products was added to the In-Fusion PCR reaction containing 2 μL In-Fusion HD enzyme premix and 100ng linearized vector with water added up to a total volume of 10 μL. The reaction was carried out by incubation at 50°C for 15 minutes and then stopped on ice. 4 μL of the In-Fusion PCR product was then used for TOP10 competent E. coli transformation.

Antibody Sequence Evaluation
Five colonies for each clone from the TOP10 competent E. coli transformation were picked for sequencing. Vκ, Vλ and V H gene sequences were analyzed using GeneBank IgBLAST to identify germline V(D)J gene segments. Individual Vκ, Vλ and V H genes were mapped to germline genes in the major IGL and IGH locus [4]. Framework and CDR sequences were annotated by following the IMGT (http://www.imgt.org/) nomenclature. Sequence alignment and phylogenetic analysis were performed using CLUSTALX2 (Center for Evolutionary Medicine) [5] and MEGA 5 (Center for Evolutionary Medicine) [6], respectively.

Expression of Human Monoclonal Antibodies
Expression of antibodies in mammalian cells and purification by Protein A/G were described previously [7,8]. Briefly, equal molar amounts of heavy-chain and light-chain plasmids were co-transfected into HEK293F cells for transient expression of antibodies. Supernatants were harvested 7 days later and subjected to testing in ELISA binding and neutralization assays as described above. Those mAbs with confirmed activities were selected for scale-up expression. IgGs were purified with Protein A beads (GE Healthcare) according to the manufacturer's instructions and concentrations were determined using Nanodrop (ThermoFisher Scientific). The purified mAbs were then re-tested in the antigen binding and viral neutralization assays.

Sample Preparation for PGM Sequencing
The 5'-RACE PCR protocol developed for unbiased antibody repertoire capture [9] has been modified to improve the template preparation. Briefly, total RNA was extracted from 20 million PBMCs into 30 μl of water with RNeasy Mini Kit (Qiagen). For unbiased antibody repertoire analysis, 5'-RACE was performed with SMARTer RACE cDNA Amplification Kit (Clontech). The immunoglobulin PCRs were set up with Platinum Taq High-Fidelity DNA Polymerase (Invitrogen Life Technologies, Carlsbad, CA) in a total volume of 50 µl, with 5 μl of cDNA as template, 1 μl of 5'-RACE primer, and 1 μl of 10 µM reverse primer. The 5'-RACE primer contained a PGM P1 adaptor, while the reverse primer contained a PGM A adaptor. 25 cycles of PCRs were performed and the expected PCR products (~600 bp) were gel purified (Qiagen).

Ion Torrent PGM Sequencing of Human Antibody Libraries
We adopted the same PGM sequencing procedure as previously described [9]. Briefly, the antibody heavy (H) and light (κ and λ) chain libraries were quantitated using Qubit® 2.0 Fluorometer with Qubit® dsDNA HS Assay Kit, and then used at a ratio of 1:1:1 for all the PGM sequencing experiments. The dilution factor required for Ion Torrent PGM template preparation was determined such that the final concentration was 50 pM. Template preparation was performed with the isothermal amplification (IA) Kit obtained from the Early Access Program. Prior to PGM sequencing, quality control of the template was determined by the Qubit® 2.0 Fluorometer with the Ion Sphere™ Quality Control Kit. Sequencing was performed on the Ion Torren PGM sequencer with the PGM™ Hi-Q 400 Kit using an Ion 318 v2 chip for a total of 1200 nucleotide flows. Raw data was processed without the 3'-end trimming in base calling in order to extend the read length [9].

Bioinformatics Analysis of Antibody Sequencing Data
The Antibodyomics 1.0 pipeline was used to analyze the HCMV antibody repertoires [9][10][11][12][13][14]. Given an antibody NGS dataset, each sequence was (1) reformatted and labeled with a unique index number; (2) assigned to V, D (for heavy chain only), and J gene families using the current human germline gene database and an in-house implementation of IgBLAST, and sequences with E-value > 10 -3 for V gene assignment were removed from the data set; (3) subjected to a templatebased error-correction procedure, in which insertion and deletion (indel) errors in the V gene segment were detected based on the alignment to their respective germline gene sequences. Note that only indels of less than three nucleotides were corrected; (4) compared to the template antibody sequences at both the nucleotide level and the amino acid level using a global alignment module in CLUSTALW2 [5]; (5) subjected to a multiple sequence alignment (MSA)-based procedure to determine the complementarity determining region 3 (CDR3), which was further compared to the template CDR3 sequences at the nucleotide level, and to determine the sequence boundary of the V(D)J coding region. After full-length variable region sequences were obtained, a bioinformatics filter was applied to detect and remove erroneous sequences that may contain swapped gene segments due to PCR errors. Specifically, a full-length read would be removed if the V-gene alignment was less than 250bp. The processed and annotated antibody sequences were then subjected to a more in-depth repertoire analysis.