Phage display approaches for the isolation of monoclonal antibodies against dengue virus envelope domain III from human and mouse derived libraries

Abstract

Domain III of the dengue virus envelope protein (EDIII, aa295-395) has an immunoglobulin fold and is the proposed receptor-binding domain of the virus. Previous studies have shown that monoclonal antibodies against EDIII can be neutralizing and have therapeutic potential. Here, cloned Fab-phage libraries of human and mouse origin were screened for DENV specific antibodies. Firstly, bacterially expressed EDIII or whole virus particles were used as bait in biopanning against a large naïve human Fab-phage library (>10 billion independent clones). Multiple panning strategies were employed, and in excess of 1000 clones were screened, but all of the antibodies identified bound the envelope in regions outside EDIII suggesting EDIII antibodies are virtually absent from the naïve human repertoire. Next, a chimeric Fab-phage library was constructed from a panel of EDIII specific mouse hybridomas by pooling the VH and VL chain sequences from the hybridomas and cloning these into the pComb3X phagemid vector with human CH and CL encoding sequences. Biopanning against EDIII identified a unique antibody (C9) that cross-reacts with EDIII from DENV1-3 and, in the IgG format, binds and neutralizes DENV2 in cell-based assays. Sequence analysis and saturation mutagenesis of complementary determining regions (CDR) in the C9 light chain suggest an antigen recognition model in which the LCDR3 is a key determinant of EDIII specificity, while modifications in LCDR1 and LCDR2 affect DENV serotype cross-reactivity. Overall, this study supports the current prevailing opinion that neutralizing anti-EDIII monoclonal antibodies can be readily generated in murine systems, but in humans the anti-DENV immune response is directed away from domain III.

Keywords: E protein; Fab; dengue; domain III; hybridoma; phage display.

Figure 1

(A) Structure of dengue virus 2 (DENV2) E protein domain III (aa295-395) in ribbon representation (PDB 1OAN). Residues comprising the 9F12 epitope are marked in orange including conserved lysines in the A-strand (K305, K307, K310) and G330 on the BC loop. Residues recognized by 3H5 in the FG loop are marked in cyan; (B) size exclusion chromatography (SEC) profile for refolded DENV1 EDIII protein using a S75 10/300 column. The major elution peak corresponds to monomeric EDIII (marked by arrow); (C) Far-UV circular dichroism spectra of DENV3 EDIII.

Figure 2

Biopanning the HX02 naïve library with DENV1 whole virus particles and characterization of the anti-DENV Fabs identified. (A) Sequence analysis and variable gene usage of anti-DENV Fab. Shown are the complementary determining region 3 (CDR3) and variable gene family designations, assigned using IMGT/V Quest [24] (B) Phage ELISA of the six unique Fab-phage clones against DENV1 virus (small check), DENV2 virus (large check), DENV1 EDIII (horizontal line) and BSA (vertical line); (C) An ELISA of purified Fab against DENV1 virus (closed symbols) and DENV2 virus (open symbols); (D) A Western blot of selected Fab with whole virus and Esol from DENV1. 4G2 and 2H2 are control mouse mAbs, which bind E and prM proteins, respectively.

Figure 3

(A) Amino acid sequence alignment of mouse variable domains of chimeric mouse/human Fab clones C9 and F1. Shown are the framework regions (FR) and complementary determining regions (CDR) of the Vκ1 and VH1 using Kabat numbering [26]. Notable sequences differences are highlighted in grey and G91 and W96 in LCDR3 are marked (*); (B) Binding of C9 and F1 Fab and C9 IgG to DENV1-4 EDIII and DENV2 and 3 whole virus (D2WV, D3WV); (C) Elisa for C9 IgG (30 nmol/L) against DENV2 whole virus particles competed with increasing concentrations of mouse 9F12 and 3H5 IgG. C9 IgG binding was detected with an anti-human horseradish peroxidase (HRP) conjugate.

Figure 4

(A) Immunofluorescence of uninfected BHK-21 cells (panel I and III) and cells infected with DENV2 (panel II and IV). Cells were incubated with C9 IgG (panel I and II) or 4G2 (3 nmol/L, panel III and IV) followed by goat anti-human or anti-mouse AF594 (Invitrogen), and mounted with ProLong Gold containing DAPI; (B) Plaque reduction neutralization test performed with DENV2 in BHK21 cells. Virus was incubated with C9 IgG or 4G2 (1 μM) prior to infection. Statistical significance (p < 0.001) is marked (*).