(A) MPER-grafted gp120 antigens. The antigens were designed using JR-FL gp120 as a template. Amino acid sequences of the modified V1/2 and V3 regions are shown, with the MPER sequence divided into the 2F5 (blue) and 4E10 (orange, with the key epitope residue W672 shown in magenta) regions. The three residues (LWN) deleted sequentially are labeled in green. Dashes indicate deleted residues. The two flanking cysteines of V1/2 shown are expected to form a disulfide bridge (35). In the Graft2 antigens, the remaining cysteine in the N-terminal part of V1 is replaced by alanine (red). Flanking residues from CMP are shown in italics. (B) Model of MPER-grafted gp120. The predicted helical MPER (red) is joined to gp120 (gray) by amino acid residues (gray dotted line) of the gp120 V1/2 loop. The model was built using the CD4-bound gp120 structure (PDB entry 2B4C) (31) and the G2A sequence. (C) Side and top views of the various Graft2 helices. In the Graft2 antigens, the 4E10 epitope (orange, with the key epitope residue W672 shown in magenta) is flanked by amino acid sequences of the coiled-coil domain of CMP (gray), which should promote helicity of the epitope region. The MPER residues LWN (green) were removed sequentially to create different versions of Graft2 antigens, namely, G2A, G2B, G2C, and G2D. Theoretically, these deletions will rotate the helical epitope through 300 degrees and expose the different faces of the 4E10 epitope region on the grafted gp120. Such rotation is better identified in the top view of the helix, where the side chains of W672, W678, and W680, residues located C-terminal to the deletion region (L669, W670, and N671), change position in reference to the postulated gp120 interface (blue box) and K662, a residue located N-terminal to the deletion. Similar manipulation was performed for the Graft1 antigens, although the epitope is not flanked by a CMP sequence and the deletion occurs sequentially within the interface of the 2F5 and 4E10 epitopes.

Binding of MAbs 2F5 and 4E10 to Graft1 and Graft2 antigens. Saturating levels of recombinant proteins expressed by 293 cells were captured by the D7324 anti-gp120 antibody and detected with serially diluted antibodies. Directly coated gp41 (10 ng/well) was used as a positive control. The binding of FDA2 antibody (A), MAb 2F5 (B), and MAb 4E10 (C) to the antigens is illustrated. The apparent affinities of G2B, G2C, and G2D relative to that of G2A are increased 4.4-, 4.8-, and 1.2-fold, respectively. Data shown are the means of two repeats and are representative of at least two experiments. OD₄₅₀, optical density at 450 nm.

Graft1 and Graft2 antigens expressed by vaccinia virus. The panel of modified gp120 antigens were expressed using recombinant vaccinia virus and purified by affinity and size-exclusion chromatography, except for G1A, which was affinity-purified G1A because of poor yields in the usual method. The proteins were mixed with Laemmli loading buffer, resolved by 4 to 15% gradient nonreducing SDS-PAGE (top) (5 μg protein/lane), and transferred to a polyvinylidene difluoride membrane for detection by MAb 4E10 (bottom) (0.5 μg protein/lane). Arrowheads indicate the approximate locations of the monomers.

End-point titration of mouse anti-gp120 antibody. Mouse sera over the course of immunization were serially diluted, and their reactivities against wild-type JR-FL gp120 were determined by ELISA. The EPTs were defined by the serum dilutions that gave readings threefold higher than the background. Gray and black arrowheads indicate the times of DNA and protein immunization, respectively. The symbol “†” indicates the time that two mice in group G2D died during the experiment.

Reactivities of mouse antisera to MPER peptides. ELISA plate wells were coated with peptides (2.5 μg/well), gp120 (25 ng/well), gp41 (50 ng/well), or ovalbumin (100 ng/well) overnight at 4°C. Mouse sera (preimmune sera or those obtained after the second protein boost) were tested at a 1:50 dilution with control MAbs 2F5 and 4E10 (0.2 μg/ml). The peptide sequences were as follows: 8836, EINCTRPNNNTRKSI; 8837, TRPNNNTRKSIHIGP; 8838, NNTRKSIHIGPGRAF; 8839, KSIHIGPGRAFYTTG; 8840, IGPGRAFYTTGEIIG; 8841, RAFYTTGEIIGDIRQ; 8842, TTGEIIGDIRQAHCN; 8843, IIGDIRQAHCNISRA; CMP1, TLNAKDKFQTKVENWFDI; CMP2, TKWLWELINTLQTSYRL; 8925, QEKNEQELLELDKWA; 8926, EQELLELDKWASLWN; 8927, LELDKWASLWNWFDI; 8928, KWASLWNWFDITNWL; and 8929, LWNWFDITNWLWYIK. Peptides CMP1 and CMP2 contain the flanking CMP sequences (underlined) of Graft2 antigens. Data shown are representative of two experiments.

End-point titration of rabbit anti-gp120 antibody. The EPTs of anti-gp120 antibody in immunized rabbits were defined by the serum dilutions that gave readings threefold higher than the background, as described in the legend to Fig. 4. Gray and black arrowheads indicate the times of DNA and protein immunization, respectively.

(A) Alignment of G2C antigen and NIH consensus B gp120 (ConB) sequences. Only the differences between the sequences are shown. Dashes are deleted residues in G2C, and the additional peptides corresponding to the CMP, 4E10 epitope, V4, and V5 sequences are shown in bold. The five variable regions of gp120 are boxed. (B) Reactivities of rabbit antisera to overlapping peptides. The experiments were performed as described in the legend to Fig. 5, using the NIH consensus B Env peptide library and the CMP1, CMP2, V4a, and V5a peptides to cover the entire G2C sequence, as shown in panel A. Antisera collected after the third protein boost, for rabbits immunized with DNA and protein, or after the fifth protein-only immunization were tested at a 1:200 dilution. Data presented are background-subtracted optical density (OD) readings. The background is defined as twice the average reading for the five least reactive peptides in each rabbit sample.

(A) Anti-V3 antibody in virus neutralization. The rabbit antisera were preincubated with the V3 peptide pool for 30 min before being mixed with the pseudotyped virus JR-FL, SF162, or JR-CSF for 1 h at 37°C, and residual virus infectivity was titrated as described in Materials and Methods. The V3 peptides have one and three different residues in JF-CSF (underlined) (TRPSNNTRKSIHIGPGRAFYTTGEIIGDIRQAH; GenBank accession number AAB03749) and SF162 (TRPNNNTRKSITIGPGRAFYATGDIIGDIRQAH; GenBank accession number P19550), respectively. The viruses were neutralized at a final serum dilution of 1:10. For JR-CSF virus, only antiserum 7672 was tested because the other antisera did not neutralize JR-CSF. N.D., not determined. (B) Anti-CD4bs antibody. The ability of the rabbit antisera to compete with sCD4 binding to monomeric gp120 was studied by ELISA. gp120-coated ELISA wells were preincubated with serially diluted rabbit sera for 30 min before the addition of an equal volume of biotinylated sCD4. Bound sCD4 was detected with peroxidase-conjugated streptavidin. The level of inhibition was defined as the % reduction in optical density readings relative to that without rabbit sera. MAbs b6 and b12 and the broadly neutralizing human antiserum FDA2 were positive controls for anti-CD4bs activity. The 50% effective concentration (EC₅₀) values for the antibodies in inhibiting sCD4 binding to gp120 were calculated by nonlinear curve fitting using Prism 4 software.