Nine ART naïve HIV-1 subtype C chronically infected pediatric cross-neutralizers (AIIMS_329, AIIMS_330, AIIMS_341, AIIMS_346, AIIMS_355, AIIMS_357, AIIMS_505, AIIMS_509 and AIIMS_510) were recruited for this study from the Pediatric OPD, Department of Pediatrics, AIIMS, New Delhi. 5 ml of blood was drawn by venipuncture under aseptic conditions from all the study subjects. Blood was used for PBMC isolation after separating the plasma by centrifugation at 600 × g. The plasma was heat inactivated at 56°C for 1 h and stored at −80°C for further experiments.

The plasmids for envelope glycoprotein expression and the purified proteins used in this study were obtained from various sources; RSC3 core wild-type (WT) and its mutant Δ371I/P363N expression plasmids and proteins (kindly provided by Dr. John R. Mascola, NIAID, VRC, USA), HXBc2 WT and its D368R mutant gp120 plasmids (kindly provided by Dr. Joseph Sodroski, Dana Farber Cancer Institute, USA). The anti-HIV-1 bnAbs 2G12, VRC01 were obtained from the NIH AIDS Reagent Program (NIH ARP). Antibody 1418, specific for the capsid of human parvovirus B19, was kindly provided by Dr. Susan Zolla Pazner, NYU, USA. The plasmids for CAP256 and its N332A mutant were kindly provided by Dr. Lynn Morris, National Institute for Communicable Diseases, South Africa. The pAK100 and pAK400 plasmids were kindly gifted by Dr. Andreas Pluckthun, Department of Biochemistry, University of Zurich, Switzerland.

All antigens were coated on ELISA plates at a concentration of 2 µg/ml and incubated overnight at 4°C. Plates were blocked with 300 μl/well of 15% FBS in RPMI for 1.5 h, followed by addition of 100 μl/well of heat inactivated plasma at different dilutions (1:30, 1:100, 1:300, 1:1,000, 1:3,000, and 1:10,000) or anti-HIV-1 monoclonal antibodies (VRC01, b12, 447-52D and 2F5) and anti-Parvovirus mAb 1418, each at 10 µg/ml concentration and incubated for 1 h at 37°C. Each of the above steps was followed by washing with 1× PBS (Phosphate Buffered Saline) with 0.1% Tween 20. Next, 100 µl of alkaline phosphatase (AP)-conjugated anti-human IgG Fc antibody [1:2,000 diluted in 1× PBS with 0.05% Tween 20 (Southern Biotech)] was added to the plates and the immune complexes were reacted with AP substrate in 10% DAE buffer (1 mg/ml). Reaction was stopped by adding 6 N NaOH and plates were read at 410 nm. The reported Max50 value is the 50% maximal binding of plasma antibodies that showed saturation. Max50 values were calculated as described previously (45).

A pediatric anti-HIV-1 subtype C scFv recombinant phage library was constructed as described previously (43, 44, 46, 47) with a few modifications. Briefly, one million PBMCs from each of the nine select pediatric cross-neutralizers were pooled and total RNA was isolated by Trizol reagent (Sigma) and then reverse transcribed to cDNA, using the Reverse aid M-MuLV reverse transcriptase (Thermo). For construction of scFv, the heavy chain and light chain variable region genes were amplified using specific primers (IDT) (Table S2 in Supplementary Material) (46) and hot start Taq DNA polymerase (Fermentas, USA). An equimolar mixture of pooled heavy and light chain DNA was used in the second round of assembly PCR using Pfu DNA polymerase. Full length scFvs were amplified by pull-through PCR reaction using Hot start Taq DNA polymerase and forward primer PTFw 5′ CCT TTC TAT GCG GCC CAG CCG GCC ATG GCC 3′ and reverse primers PAK kappa Sfi 5′ TCA GCA TGG CCC CCG AGG CCG CAC GTT TRA T 3′, and PAK lambda Sfi 5′ TCA GCA TGG CCC CCG AGG CCG CAC CTA RRA C 3′ (R = G and A). The scFvs were resolved on agarose gel and purified using gel extraction kit (Qiagen). The library was constructed by ligating the scFv into pAK100 phagemid vector by using T4 DNA ligase (New England Biolabs) followed by transformation into TG1 electrocompetent cells by electroporation (current 25°F, resistance 200 Ohms, voltage 2,500 V) (BioRad). The transformed cells were plated on to 2XYT medium agar plates containing chloramphenicol (30 µg/ml) and incubated overnight at 37°C. A glycerol stock of the recombinant scFv library was made and aliquots stored at −80°C.

Twenty scFv clones were randomly picked from the library to check the presence of scFv inserts and diversity of the scFv phage library. Colony PCR was done by using forward primer; PTfw 5′ CCT TTC TAT GCG GCC CAG CCG GCC ATG GCC 3′ and reverse primer; (pAK Sfi1) 5′ TCA GCA TGG CCC CCG AGG CCG CAC GTT TRAT 3′, PAK lambda Sfi 5′ TCA GCA TGG CCC CCG AGG CCG CAC CTA RRA C 3′ (R = G and A). Plasmid DNA of colony PCR positive scFv clones was isolated, sequenced commercially by Macrogen (South Korea), and the sequences were analyzed by online IMGT/V-Quest software provided by the international ImMunoGeneTics database (IMGT) (http://www.imgt.org/IMGT_vquest/share/textes/).

The rescue of phage library was done with M13KO7 helper phage (Stratagene) as mentioned previously (44, 48). The phages were then subjected to five rounds of enrichment by bio-panning as described earlier (43). Briefly, RSC3 core protein was coupled to magnetic beads (MyOne Tosyl activated Dynalbeads, Invitrogen) according to the manufacturer’s instructions, and the antigenic integrity of protein after bead coupling was verified by flow cytometry, using bnAb VRC01. Phages were transferred to an eppendorf tube containing 60 µl RSC3 core protein coated magnetic beads and incubated for 1 h at RT with gentle shaking. The unbound phage was removed by washing 10–15 times with 1× PBS with 0.1% Tween 20. The bound phage was eluted with 0.2 M glycine pH 2.2 for 10 min at RT. The eluted phage were neutralized with 1 M Tris-HCl pH 9.2 and immediately added to TG1/HB2151 cells (OD = 0.5) for infection at 37°C without shaking for 30 min and with shaking at 37°C for 30 min. Cells were spun down and plated on 2XTY agar containing chloramphenicol (30 µg/ml). Finally, the individual colonies were picked and grown in 96 well sterile culture plates (Corning) and a glycerol stock of each colony was stored at −70°C. This procedure was again repeated four times to complete five rounds of biopanning.

Soluble phage ELISA was performed as described previously (43, 49, 50), with few modifications. The ELISA plates were coated with 100 µl of RSC3 core and its mutant Δ371I/P363N at 2 µg/ml in 0.1 M NaHCO₃ (pH 8.6) and incubated overnight at 4°C. The anti-CD4bs bnAbs VRC01 and b12 was used as positive controls; anti-CD4bs non-neutralizing antibody (non-nAb) b6, anti-V3 mAb 447-52D, anti-Parvovirus mAb 1418, and scFv against hepatitis virus HepB scFv were used as negative controls at 1 µg/ml concentration in these assays. The phage clones were taken as positive by ELISA, if the absorbance of RSC3 core coated wells was at least two times higher than RSC3 core mutant delta 37I/P363N and at least three times higher than binding of phage clones to unrelated antigens (unrelated peptide pool, BSA).

Sequencing analysis of positive scFv clones of the soluble phage ELISA was performed as described previously (43, 44). Briefly, the plasmid of positive scFv clones from the soluble phage ELISA was isolated by using plasmid purification Mini kit (Qiagen, Germany) and sequenced commercially by Macrogen (South Korea). The sequences were analyzed using IMGT V Quest software.

The binding clones from the soluble phage ELISA were selected for soluble scFv expression as described earlier (43). The pAK100 vector carrying scFv DNA and pAK400 vector were digested with SfiI restriction enzyme (New England Biolabs). The digested scFv DNA fragments and pAK400 vector were gel purified using gel extraction kit (Qiagen). The scFv DNA was ligated into pAK400 vector at a 3:1 M ratio using T4 DNA ligase (New England Biolabs) in a reaction volume of 20 µl followed by transformation into HB2151 cells using calcium chloride method. The HB2151 cells carrying scFv cloned pAK400 plasmid were grown and induced by 1 mM IPTG for 4 at 24°C (48). Next, scFv was purified from the periplasmic extract by Ni-NTA (Qiagen) affinity chromatography as described earlier (44). The protein was eluted with 300 mM imidazole followed by extensive dialysis against ice cold 1× PBS (pH 7.4), and protein samples were concentrated using ultrafiltration columns (Ambion) over a 10 kDa cut off and filtered by 0.2 µm sterile syringe filter. The purity of the purified proteins was assessed by SDS-PAGE and Western blot. Briefly, resolving the scFv on 12% SDS-PAGE followed by Western blotting, the scFv was detected using anti-His-tag antibody raised in mouse (Sigma) at a 1:1,000 dilution in 1× PBS with 0.05% Tween 20. Secondary anti-mouse HRP (1:3,000 dilutions) antibody was added and incubated at RT for 2 h, and color was developed using 3,3′-diaminobenzidine tetrahydrochloride (DAB) as substrate. Proteins with >95% purity were stored at −80°C immediately after snap freezing in liquid nitrogen.

ELISA of purified scFvs was done by using HIV-1 monomeric and trimeric envelope glycoproteins (HIV-1 96ZM651 gp120 monomer subtype C, BG505 SOSIP.664 D7324 trimeric antigen and HxBc2 gp120 and HxBc2 D368R mutant gp120 monomers) as described previously (2, 43). Briefly, all antigens were coated on ELISA plates at 2 µg/ml concentration, blocked with 15% FBS, 2% BSA in a RPMI media; threefold serial dilutions of scFvs and mAb controls (VRC01, b12, 447-52D, 2G12, 2F5 and 1418) were added followed by detection with 1:1,000 dilution of primary antibody (Anti-His tag for scFvs and HRP conjugated anti-human Fc antibody for mAb controls) in 2% milk phosphate-buffer saline (MPBS). For scFvs, 1:2,000 diluted anti-mouse HRP-conjugated secondary antibody in 2% MPBS was added. Each of the above steps was followed by washing the plate with 1× PBS with 0.1% Tween 20. Immediately after washing, 100 µl of TMB substrate (BioLegend) was added and incubated at RT till the color developed. Reaction was stopped by adding 8 N H₂SO₄ and the absorbance was read at 450 nm.

Affinity analysis of the scFvs was done by ELISA as described previously (43). Briefly, different dilutions of BG505:SOSIP.664 gp140 (10–800 nM) were incubated with fixed concentration of the scFvs (3.125 nM) for 16 h at 4°C. Next day, 100 µl of the equilibrated solution was incubated with BG505:SOSIP.664 gp140 coated ELISA plates (500 ng/well) for 20 min at RT to capture the free scFvs. Bound scFv was detected with primary anti-His tag antibody (raised in mouse, Sigma) at 1:1,000 dilution and with secondary anti-mouse HRP conjugated antibody (Sigma) at 1:2,000 dilution as described before. Saturation binding curve was plotted using non-linear regression method and KD values were calculated. Competition ELISA was performed as described in Ref. (13, 43) by using biotinylated VRC01 and BG505:SOSIP.664 gp140 antigen. All the ELISAs were repeated twice and data was analyzed by using GraphPad Prism software 5.0.

The neutralization efficiency of the purified scFv monoclonals was tested with HIV-1 pseudoviruses and primary isolates as described previously (43, 44, 51). Briefly, HIV-1 envelope expression plasmids from a standard panel for subtypes A, B, C, AD, CD, and CRF2 AG were obtained from the NIH AIDS Reagent Program (NIH ARP). Pseudoviruses were generated by co-transfecting HEK293T cells with a plasmid expressing envelope and envelope deficient HIV-1 backbone vector pSG3Δenv as described earlier (51), and primary isolates were derived from PBMCs of HIV-1 infected Indian donors as reported earlier (33, 45, 52). Next, starting with 100 µg/ml of scFv concentration, three-fold serial dilutions were made in 10% DMEM [Dulbecco Modified Eagles’s Medium, HyClone, GE Healthcare supplemented with 10% FBS (Fetal Bovine Serum) and 1% PenStrep (Penicillin and Streptomycin) antibiotics] and incubated with 200 TCID of the pseudoviruses/primary isolates for 1 h at 37°C. Then, 10⁴ cells/well TZM-bl cells were added to the scFv and virus complex along with 25 µg/ml DEAE Dextran (Sigma) (Indinavir (1 mM) was additionally added with primary isolates). After 48 h, 150 µl volume of culture medium was removed and 50 µl of Bright-Glo Luciferase assay reagent (Promega) was added to TZM-bl cells followed by 2-min incubation at room temperature to allow complete cell lysis. Then, luciferase activity was read in 96-well black plate (Costar) using Synergy 2 Multi-Mode Reader (BioTek). The IC50 values for the scFvs were calculated by a dose-response curve fit with non-linear function, with the help of Graph Pad prism software 5.0. Each experiment was repeated twice and performed in duplicates and IC50 titers were calculated.

Statistical analyses were performed using Graph Pad Prism 5. Non-linear regression curve straight line was plotted using the method of least squares to determine the Max50 and IC50 values. Mean Max50 binding titers were compared using unpaired t test. P values <0.05 were considered significant.

Nine anti-retroviral naïve HIV-1 subtype C infected pediatric cross-neutralizers were recruited for the generation of the human anti-HIV-1 scFv recombinant library, based on the viral cross neutralizing potential exhibited by their plasma antibodies (35, 36, 38). The demographic and clinical profile of the study subjects is provided in Table 1. The infected donor plasma samples were analyzed for the presence of anti-CD4bs antibodies using recombinant gp120 proteins (13, 53) known to bind CD4bs-directed bnAbs and their corresponding mutants; the RSC3 core protein and its mutant Δ371I/P363N, HXBc2 WT gp120 and its mutant D368R. The plasma antibodies of 4 pediatric donors (AIIMS_341, AIIMS_346, AIIMS_357, and AIIMS_361) demonstrated high binding titers with RSC3 core WT protein as compared to its mutant Δ371I/P363N, suggesting the presence of VRC01 like anti-CD4bs antibodies in their plasma (Figure 1A); this was further confirmed with HXBc2 WT gp120 and its mutant D368R (Figure 1B). The results for the binding of positive and negative control antibodies with RSC3 core WT and its mutant Δ371I/P363N are shown in (Figure 1C) and with HXBc2 WT and its mutant D368R in (Figure 1D).

The PBMCs from 9 pediatric cross-neutralizers were pooled and an anti-HIV-1 pediatric scFv phage library was constructed, as described in the methodology section. The variable genes for heavy chain (VH) and light chains kappa and lambda (VLκ and VLλ) were successfully amplified showing desired 400 bp bands, except VLκ3 and VLλ2 (Figures S1A–C in Supplementary Material) followed by successful construction of scFv gene (800 bp) by pull-through PCR (Figure S1D in Supplementary Material). Next, colony PCR of 20 random clones from the unselected phage library (before biopanning) showed two clones with no scFv inserts (Figure S2 in Supplementary Material) thus, confirming 90% ligation efficiency. DNA sequencing of 18 colony PCR positive clones confirmed that sequence of all scFv inserts were distinct (Table S1 in Supplementary Material). On the basis of these results, it is inferred that, diversity of the constructed phage library was 90% with 9 × 10⁸ diverse clones. Biopanning of the library was performed using RSC3 protein coated on magnetic beads. The integrity of coated RSC3 protein on the beads was checked by flow cytometry by using anti-CD4bs bnAb VRC01 and more than 99% beads were found coated with RSC3 core protein (Figure S3 in Supplementary Material). After five rounds of biopanning, 150 scFv clones were randomly selected, their binding specificity was checked and eight scFv clones were identified which showed highest binding with the RSC3 protein and a decrease in binding of at least twofold with RSC3 Δ 371I/P363N protein, demonstrating CD4bs specificity (Figure 2).

The plasmid of eight soluble phage ELISA positive scFv clones was isolated and sequenced, and we found 7/8 clones were similar in sequence because of the enrichment of antigen-specific clones after five rounds of biopanning. Therefore, two clones (2B10 and 2E4) out of eight identified scFv clones were distinct and were taken for further analysis. Analysis of the gene usage of the scFvs by IMGT/V-Quest software program revealed that both 2B10 and 2E4 scFvs are variants of the same immunoglobulin heavy chain gene family IGHV1-02*02, but having different D and J region alleles, i.e., IGHD6-19*01 and IGHJ4*02 for 2B10 and IGHD6-13*01 and IGHJ3*01 for 2E4. The CDRH3 and CDRL3 region of 2B10 is shorter (13 amino acids in CDRH3 and 12 amino acids in CDRL3) than CDRH3 and CDRL3 regions of 2E4 (18 amino acids in CDRH3 and 13 amino acids in CDRL3). Also, higher number of somatic hypermutations (SHMs) was present in the immunoglobulin heavy chain and light chain genes of 2B10 than 2E4 (Table 2).

The SDS-PAGE and Western blot analysis of the purified scFv clones 2B10 and 2E4 confirmed that both scFv proteins were more than 95% pure and showed the presence 32 kDa scFv bands (Figure S4 in Supplementary Material). The purified 2B10 and 2E4 scFvs were tested further in ELISA binding assays using HXBc2 gp120 and its mutant D368R. Both the scFvs showed binding to HXBc2 gp120 protein, with no significant binding to its corresponding HXBc2 D368R mutant, as is expected of CD4bs directed Abs (Figure 3A). This was further confirmed by competition ELISA with biotinylated VRC01 (Figure 3B). Moreover, both scFv monoclonals did not show any binding with V3 and MPER peptides (data not shown). In addition, the neutralization activity of both the scFv monoclonals was assessed for their V1V2 and V3-glycan reactivity by using CAP256 WT and its V3-glycan N332A mutant pseudoviruses (Table 3) and 16055 WT and its V1/V2 mutants N156K, N160K, K169A, and K171A pseudoviruses (Table 4). We observed, neither of the two scFvs exhibit V1/V2 or V3-glycan reactivity.

The 2B10 and 2E4 scFvs were next tested for their ability to bind native-like HIV-1 trimeric envelope glycoprotein BG505:SOSIP.664 D-7324, and we observed high binding of both the scFvs with the trimeric antigen, which is known to bind only to bnAbs with high affinity and negligible binding with non-NAbs (Figure 4A). Further, binding analysis of both scFv monoclonals with recombinant HIV-1 96ZM651 subtype C protein confirmed their ability to bind with subtype C viral envelopes (Figure 4B). The affinity dissociation constant (KD) of both 2B10 scFv-BG505:SOSIP.664 gp140 complex and 2E4 scFv-BG505:SOSIP.664 gp140 complex was determined by ELISA and saturation binding curve was plotted through non-linear regression analysis (Figure 5). The Kd value for 2B10 scFv is 9.2 ± 1.4 × 10⁻⁸ M (R2 = 0.9709) and for 2E4 scFv is 1.4 ± 0.2 × 10⁻⁷ M (R2 = 0.9662). In binding analysis experiments, HepB, a scFv directed against hepatitis B antigen (54), was used as negative control and anti-CD4bs bnAb VRC01 was used as positive control.

We assessed the neutralization activity of 2B10 and 2E4 scFvs against a standard panel of 49 pseudoviruses and primary isolates (7 subtype A, 13 subtype B, 24 subtype C, 1 subtype CD, 1 subtype AD, and 3 subtype AG) of different origin. The scFv monoclonals were tested at a starting concentration of 100 µg/ml with three-fold serial dilutions. The Geometric Mean Titers for the viruses neutralized by scFvs were calculated directly using MS-Excel. The 2B10 scFv monoclonal neutralized 38/49 viruses with GMT (Geometric mean titer) 17.9 µg/ml (20/24 subtype C, 9/13 subtype B, 5/7 subtype A, 1/1 subtype CD, 1/1 subtype AD, 2/3 subtype CRF02 AG) whereas, 2E4 neutralized 16/49 viruses with GMT 51.2 µg/ml (8/25 subtype C, 4/14 subtype B, 2/7 subtype A, 0/1 subtype CD, 0/1 subtype AD, and 1/6 subtype CRF02 AG) (Table 5). These results revealed better cross-neutralization breadth of the 2B10 scFv than 2E4; with 2B10 displaying cross neutralization across subtype C viruses from different geographical regions. The murine leukemia virus (MuLV) virus was used as negative control in all neutralization assays.