Structural and Functional Characterization of Anti-A33 Antibodies Reveal a Potent Cross-Species Orthopoxviruses Neutralizer

Abstract

Vaccinia virus A33 is an extracellular enveloped virus (EEV)-specific type II membrane glycoprotein that is essential for efficient EEV formation and long-range viral spread within the host. A33 is a target for neutralizing antibody responses against EEV. In this study, we produced seven murine anti-A33 monoclonal antibodies (MAbs) by immunizing mice with live VACV, followed by boosting with the soluble A33 homodimeric ectodomain. Five A33 specific MAbs were capable of neutralizing EEV in the presence of complement. All MAbs bind to conformational epitopes on A33 but not to linear peptides. To identify the epitopes, we have adetermined the crystal structures of three representative neutralizing MAbs in complex with A33. We have further determined the binding kinetics for each of the three antibodies to wild-type A33, as well as to engineered A33 that contained single alanine substitutions within the epitopes of the three crystallized antibodies. While the Fab of both MAbs A2C7 and A20G2 binds to a single A33 subunit, the Fab from MAb A27D7 binds to both A33 subunits simultaneously. A27D7 binding is resistant to single alanine substitutions within the A33 epitope. A27D7 also demonstrated high-affinity binding with recombinant A33 protein that mimics other orthopoxvirus strains in the A27D7 epitope, such as ectromelia, monkeypox, and cowpox virus, suggesting that A27D7 is a potent cross-neutralizer. Finally, we confirmed that A27D7 protects mice against a lethal challenge with ectromelia virus.

Fig 1. Complement and isotype dependence of anti-A33 MAb neutralization of VACV EEV.

VACV EEV neutralization activity of purified anti-A33 MAbs in the absence (MAbs) or presence (MAbs+10% C´) of complement. Rabbit anti-A33 polyclonal Abs (N628) were used as positive control. Anti-B5 MAbs B126 (IgG2a), and B96 (IgG1) were used as positive and negative neutralization controls, respectively. Human anti-ditrophenol (DNP, IgG1) and VACV EEV (EEV) were used as negative controls. Error bars indicate SEM in each condition. Dashed line indicates the 50% of plaques number of VACV EV in panels A and B. The data are representative of two experiments. Three more experiments were done and they show comparable results.

Fig 2. Protection of Balb/c mice from VACV_WR by anti-A33 mAbs.

Antibody protection against weight loss (A) and mortality (B) caused by VACV_WR challenge. All analyzed anti-A33 MAbs protect both against weight loss and death. One of two independent experiments is shown. Significance for all A33 antibodies (****) was P<0.0001, while the anti-L1 control antibody M12B9 was not significant (ns) with P = 0.1577. The negative control antibody A10 provided no protection as expected, while the positive control anti-B5 antibody B126 conferred protection. (B) Note that slight shifts on the Y-axis were implemented for visualization purposes for all antibodies that confer full protection.

Fig 3. Fab/A33 binding interactions.

(A) Determination of A33/Fab binding stoichiometry by size exclusion chromatography (SEC). Fab/A33 complex formation of five different MAbs (A2C7, A26C7, A25D11, A27D7, and A20G2) is illustrated. Two major peaks corresponding in size to 1 Fab: 1 A33 dimer (red and blue curve) and 2 Fabs: 1 A33 dimer (brown, cyan and grey) complexes are visible. Presence of both Fab and A33 in each peak are confirmed by SDS-PAGE (bottom left corner). Molecular weight markers with sizes in kDa are shown as a reference (thin grey curve). (B) Real-time A33 binding curves to immobilized MAbs as assessed by BLI. Note that A33 dissociates much slower from MAbs A2C7 and A20G2, compared to A27D7, likely to its ability to simultaneous bind to two Fabs.

Fig 4. Structures of A33/A2C7, A20G2, and A27D7-Fab complexes.

Structures of (A) A33/A2C7, (B) A20G2, and (C) A27D7-Fab complexes. Structures are shown as cartoon representations, with light chains in green and heavy chains in orange (top), while the antibody footprints on the molecular surface of A33 are shown below. In (C), A33 subunits (SUs) are shown in different colors because the epitope spreads asymmetrically over both SUs (SU A in aquamarine, SU B in grey). Footprints on A33 is colored according to contacting CDR loops (L1, brown; L3, yellow; H1, green; H2, cyan; H3, blue). Stripe patterns highlight A33 residues that elicit contacts with more than 1 CDR. The continuous line and the asterisk indicate A33 dimer axis. Letters indicate when this axis constitutes a crystallographic axis (X) along with the A33 dimerization axis (P). A33 residues that are part of the orthopox variation profile are labeled in magenta.

Fig 5. Detailed contacts at the Ag:Fab interface.

Interactions between A33 residues and light-chain (L, orange) or heavy chain (H, green) CDRs (1–3) of the MAbs A2C7, A20G2, and A27D7. Residues that elicit actual contacts as defined in S3 Table are colored according to Fig 4. Main chain is shown uniformly for residues eliciting VdW interactions, while individual atoms are colored for residues eliciting electrostatic interactions. Amino acid side chains are shown when relevant. A33 backbone in grey, and contact residues in black. Only main chain atoms are shown when residues are involved in VdW interactions.

Fig 6. Binding kinetics of WT-A33 and mutants to antibodies A2C7, A20G2, and A27D7.

Apparent equilibrium binding constants (K_D) calculated from the binding curves shown in S3 Fig. Note that K_D values can include avidity effects, since both antibodies and A33 are dimeric proteins. Grey shading qualitatively illustrates the extent of binding impairment by the indicated A33 mutation (from white, no impairment to black, full binding abrogation). Lys123Ala (K123A) was required for successful crystallization and is considered as wild-type A33 (WT) since it is located outside any MAb epitope [40].

Fig 7. Treatment with antibody A27D7 protects mice against a lethal ECTV challenge.

Mice were treated with A2C7, A20G2 or A27D7 at T = -1 or T = +1 relative to a 1000 PFU ECTV challenge. (A) Mice were monitored for morbidity, as measured by weight-change. Mice treated with A27D7 at T = +1 were fully protected against morbidity. N = 5 mice/group. Data presented is from 1 of 2 studies—both with similar results. (B) Mice were monitored for mortality until T = +21. Only mice treated with A27D7 were significantly protected against challenge. P = 0.0004 for T = +1 and P = 0.006 for T = -1 (90% protected). N = 10 animals/group. Data is pooled from 2 independent studies of 5 animals/group. Controls include no ECTV challenge (PBS/NT), challenge without treatment (NT), treatment with vehicle only (Veh) or CDV treatment.