Broadly neutralizing antibodies targeted to the membrane-proximal external region of human immunodeficiency virus type 1 glycoprotein gp41

Abstract

The identification and epitope mapping of broadly neutralizing anti-human immunodeficiency virus type 1 (HIV-1) antibodies (Abs) is important for vaccine design, but, despite much effort, very few such Abs have been forthcoming. Only one broadly neutralizing anti-gp41 monoclonal Ab (MAb), 2F5, has been described. Here we report on two MAbs that recognize a region immediately C-terminal of the 2F5 epitope. Both MAbs were generated from HIV-1-seropositive donors, one (Z13) from an antibody phage display library, and one (4E10) as a hybridoma. Both MAbs recognize a predominantly linear and relatively conserved epitope, compete with each other for binding to synthetic peptide derived from gp41, and bind to HIV-1(MN) virions. By flow cytometry, these MAbs appear to bind relatively weakly to infected cells and this binding is not perturbed by pretreatment of the infected cells with soluble CD4. Despite the apparent linear nature of the epitopes of Z13 and 4E10, denaturation of recombinant envelope protein reduces the binding of these MAbs, suggesting some conformational requirements for full epitope expression. Most significantly, Z13 and 4E10 are able to neutralize selected primary isolates from diverse subtypes of HIV-1 (e.g., subtypes B, C, and E). The results suggest that a rather extensive region of gp41 close to the transmembrane domain is accessible to neutralizing Abs and could form a useful target for vaccine design.

FIG. 1

Deduced amino acid sequences of the variable regions of Fabs selected against MN peptide 2031 from FDA-2 kappa and lambda phage display libraries.

FIG. 2

ELISA binding curve of Fab Z13 (κ) and Fab Z13λ1 (λ) against immobilized 2031 peptide. O.D., optical density.

FIG. 3

Specificities of Fab Z13, IgG 4E10, and IgG 2F5 for a range of immobilized proteins and peptides by ELISA. Fab Z13 (A), IgG 4E10 (B), and IgG 2F5 (C) were titrated against MN peptides 2030, 2031c, and 2032; the SIV mac239 peptide (aa 671 to 690); BSA; ovalbumin; NFDM; or recombinant gp120_JR-FL. O.D., optical density.

FIG. 4

Competition ELISA using Fab Z13, IgG 4E10, and IgG 2F5 against their BIO counterparts for recognition of immobilized MN 2031c peptide. Fab Z13 (BIO) (A), IgG 4E10 (BIO) (B), and IgG 2F5 (BIO) (C) were coincubated with various concentrations of competing MAb and added to wells containing immobilized 2031c. Each data point represents the mean of duplicate samples.

FIG. 5

Competition ELISA of IgG 4E10 and IgG 2F5 against their BIO counterparts for recognition of immobilized gp41(VT)_IIIB. IgG 4E10 (BIO) (A) and IgG 2F5 (BIO) (B) were coincubated with various concentrations of competing MAb and added to wells containing immobilized gp41(VT)_IIIB. No competition was observed for either of the BIO MAbs using an irrelevant MAb, IgG KZ52. Each data point represents the mean of duplicate samples.

FIG. 6

Amino acid sequences of peptides affinity selected from an HxB2 gp160 fragment library using IgG 4E10 or IgG 2F5. The HxBc2 gp160 protein fragment phage display library was subjected to four rounds of panning using either IgG 4E10 (A) or IgG 2F5 (B) as selecting agents. Random clones were selected from the enriched pools of phage and tested for reactivity against IgG 4E10 and/or IgG 2F5 by ELISA, and the positive clones were sequenced in the insert region. The residues that are in bold text and underlined represent the minimal motif that is shared by overlapping peptides.

FIG. 7

Binding by ELISA of Fab b12, IgG 2F5, IgG 4E10, Fab Z13, and IgG KZ52 to various concentrations of HIV-1_MN (MN-1000×) viral particles captured by human anti-HLA MAb L-243. All MAbs were used at 10 μg/ml. O.D., optical density.

FIG. 8

MAb binding to H9 cells infected with HIV-1_MN and HIV-1_HxB2 as measured by flow cytometry. (A) H9 cells infected with HIV-1_MN were probed with IgG 4E10 (100 μg/ml), Fab Z13 (200 μg/ml), IgG 2F5 (100 μg/ml), IgG b12 (50 μg/ml), IgG 50–69 (10 μg/ml), and Fab ELZ510 (200 μg/ml). (B) H9 cells infected with HIV-1_HxB2 were probed with IgG 4E10, IgG 2F5, IgG b12, and IgG KZ52 (100 μg/ml each) or IgG 50–69 (20 μg/ml). Infected cells were incubated in the presence or absence of 25 μg of sCD4 per ml for 1 h at 37°C prior to the addition of MAbs. Uninfected H9 cells were also probed by each MAb at the same concentration and always yielded background signals that were lower than that observed for the negative control MAbs on infected cells.

FIG. 9

Binding of anti-gp41 MAbs to SOS-gp140_JRFL(A), gp41(VT)_IIIB (B), and gp160(PS)_IIIB (C) coated directly on polystyrene plates either untreated (Native) or following treatment with SDS, dithiothreitol, and boiling (Denatured [for more details, see Materials and Methods]), as measured by ELISA. IgG 50–69 and Fab T3 are cluster I (cl.I) and cluster II (cl.II) MAbs, respectively.

FIG. 10

Cartoon model of the HIV-1 putative trimeric envelope spike. Most of the surface of gp41 is believed to be occluded by gp120 and other molecules of gp41. A region of gp41 close to the membrane defined by MAbs 2F5, Z13, and 4E10 is suggested to be exposed to antibody binding. An IgG molecule is shown to scale in proximity to the proposed binding region. The trimer is based on the structure proposed by Kwong et al. (32). The IgG molecule dimensions are taken from those of a human IgG1 molecule (46). Surfaces were calculated using the msms program (59).