Rationally Targeted Mutations at the V1V2 Domain of the HIV-1 Envelope to Augment Virus Neutralization by Anti-V1V2 Monoclonal Antibodies

Abstract

HIV-1 envelope glycoproteins (Env) are the only viral antigens present on the virus surface and serve as the key targets for virus-neutralizing antibodies. However, HIV-1 deploys multiple strategies to shield the vulnerable sites on its Env from neutralizing antibodies. The V1V2 domain located at the apex of the HIV-1 Env spike is known to encompass highly variable loops, but V1V2 also contains immunogenic conserved elements recognized by cross-reactive antibodies. This study evaluates human monoclonal antibodies (mAbs) against V2 epitopes which overlap with the conserved integrin α4β7-binding LDV/I motif, designated as the V2i (integrin) epitopes. We postulate that the V2i Abs have weak or no neutralizing activities because the V2i epitopes are often occluded from antibody recognition. To gain insights into the mechanisms of the V2i occlusion, we evaluated three elements at the distal end of the V1V2 domain shown in the structure of V2i epitope complexed with mAb 830A to be important for antibody recognition of the V2i epitope. Amino-acid substitutions at position 179 that restore the LDV/I motif had minimal effects on virus sensitivity to neutralization by most V2i mAbs. However, a charge change at position 153 in the V1 region significantly increased sensitivity of subtype C virus ZM109 to most V2i mAbs. Separately, a disulfide bond introduced to stabilize the hypervariable region of V2 loop also enhanced virus neutralization by some V2i mAbs, but the effects varied depending on the virus. These data demonstrate that multiple elements within the V1V2 domain act independently and in a virus-dependent fashion to govern the antibody recognition and accessibility of V2i epitopes, suggesting the need for multi-pronged strategies to counter the escape and the shielding mechanisms obstructing the V2i Abs from neutralizing HIV-1.

Fig 1. Alignment of V1V2 sequences from HIV-1 isolates tested and the HIV-1 Env trimeric structure with the 830A epitope.

(A) Comparison of V1V2 sequences from HIV-1 isolates tested in this study. These five viruses display distinct neutralization sensitivity to V2i mAbs [12]. The key contact residues for mAb 830A are underlined. The integrin α4β7-binding motif ¹⁷⁹LDV/I¹⁸¹ is shown in red, and amino acids that diverge from this motif are highlighted in yellow. The green highlight shows a distinct amino acid at position 153 in the V1 loop of the ZM109 strain. The gray highlight shows the sequence and length of the hypervariable V2 loop. To stabilize the flexible hypervariable V2 loop, disulfide bonds were introduced at positions highlighted in blue (183–191 and 184–190). (B) The surface of V2i mAb 830A epitope is illustrated on the pre-fusion Env trimer represented by the BG505 SOSIP structure (PDB Code 4TVP). The gp120 trimer is shown as ribbons and the V1V2 regions are colored in magenta. The 830A-binding site surface is colored in orange. However, since this surface faces down from this top view, only the majority of the backside (colored in dark grey) is seen. The surface of the PG9 epitope (colored in blue) is shown as a reference. (C) A zoom-in of the 830A epitope surface from an angle of view about 90 degree rotated from that in Fig 1B. For clarity, only one gp120 is shown.

Fig 2. A single mutation creating the LDV motif has minimal effects on neutralization sensitivity of 6535 and REJO pseudoviruses to most V2i mAbs.

(A) Infectivity of 6535 and its mutant T179L in the TZM-bl cells. Virus infectivity was measured based on β-galactosidase activity after 48 hrs. (B) Neutralization of 6535 and its mutant pseudoviruses by V2i, V2q, CD4bs, and irrelevant control mAbs. (C) Titration curves showing significant differences in neutralization of 6535 WT vs. T179L mutant by V2i mAbs 697 and 830A. *, p<0.05 based on two-way ANOVA test of the titration curves. (D) Neutralization of REJO WT and Q179L mutant pseudoviruses by V2i, V2q, and control mAbs. For neutralization assay, virus was pre-incubated with mAb for 1 hr at 37°C, and added to the TZM-bl cells. Inhibition of virus infectivity was measured by β-galactosidase activity. (D) ELISA reactivity of V2i, V3, CD4bs, and control mAbs to gp120 from 6535 WT and T179L mutant. For this assay, gp120s from virus lysates were captured onto 96-well plates by sheep anti-gp120 Abs, and reacted with mAbs tested. MAb binding was detected with alkaline phosphatase-conjugated goat anti-human IgG Fc and p-nitrophenyl phosphate substrate. Means and standard errors from two or more repeat experiments are shown.

Fig 3. A charge change at position 153 in the V1 region of ZM109 increases the virus sensitivity to most V2i mAbs.

(A) Neutralization of ZM109 WT and R153E mutant pseudoviruses by V2i and other mAbs. *, p<0.05 based on a two-way ANOVA test of the mAb titration results. (B) Titration curves showing differences in sensitivity of ZM109 WT vs. R153E to neutralization by five of the seven V2i mAbs. Statistical analyses were performed for neutralization data reaching above 50%. (C) ELISA reactivity of mAbs and CD4-IgG2 to gp120s from virus lysates of ZM109 WT and R153E.

Fig 4. Addition of a disulfide bond to the flexible hypervariable V2 loop of ZM109 significantly increases the virus sensitivity to neutralization by some V2i mAbs.

(A) Neutralization of ZM109 WT by V2i and other gp120-specific mAbs following virus-mAb pre-incubation for 1 hour vs. 24 hours. (B) Stabilization of the hypervariable region of V2 loop by a disulfide bond. The two disulfide bonds introduced individually were shown together in this model. The V1V2 region is rendered as ribbons with the mutated residues colored in magenta. The distances of the C atoms of the two pairs are indicated. The epitope locations of 830A and PG9 are indicated by orange and blue curves, respectively. (C) Infectivity of pseudoviruses with WT or mutant ZM109 Envs in TZM-bl cells. (D) Neutralization of ZM109 WT and the disulfide mutant (183–191) by V2i and other mAbs. (E) Titration curves showing differences in neutralization sensitivity of ZM109 WT and the 183–191 mutant to V2i mAbs 1361, 1393A, and 2158. Statistical analyses were performed for neutralization curves reaching above 50%. *, p<0.05 based on two way ANOVA test of the mAb titration curves. (F) ELISA reactivity of gp120s from ZM109 WT, 183–191, and the Y191C single mutant with V2i and other gp120-specific mAbs. CD4-IgG2 and the irrelevant mAb 1418 were also tested as controls.

Fig 5. Addition of a disulfide bond to the hypervariable V2 loops of SF162 and BaL modulates virus sensitivity to V2i mAbs.

(A) Infectivity of SF162 WT and its disulfide mutants in TZM-bl cells. (B) Neutralization of SF162 WT and its disulfide mutants by V2i and other mAbs. (C) Titration curves showing differences in neutralization sensitivity of SF162 WT and mutants to V2i mAbs 1361, 1393A, and 2158. (D) Infectivity of BaL.01 and its disulfide mutants in TZM-bl cells. (E) Neutralization of BaL.01 and its disulfide mutants by V2i and other mAbs. (F) Titration curves showing differences in neutralization sensitivity of BaL.01 WT and mutants to V2i mAbs 697, 830A, 1361, 1393A, and 2158. Statistical analyses were performed for neutralization curves reaching above 50%. *, p<0.05 based on two-way ANOVA test of the mAb titration data. (G) ELISA reactivity of V2i, V3, and control mAbs with gp120 from WT or S-S mutants of SF162 and BaL.01.