ANTI-DENGUE VACCINES AND ANTIBODIES

Abstract

A Dengue virus Envelope Dimer Epitope (EDE) wherein the EDE: c) spans the polypeptides of a Dengue virus Envelope polypeptide dimer; and/or d) is presented on a dimer of Envelope proteins; and/or c) is formed from consecutive or nonconsecutive residues of the envelope polypeptide dimer, wherein the dimer is a homodimer or heterodimer of native and/or mutant envelope polypeptides, from any one or two of DENV-1, DENV-2, DENV-3 and DENV-4. The EDE may be a stabilized recombinant dengue virus envelope glycoprotein E ectodomain (sE) dimer, wherein the dimer is: covalently stabilized with at least one disulphide inter-chain bond between the two sE monomers, and/or covalently stabilized with at least one sulfhydrylreactive crosslinker between the two sE monomers, and/or covalently stabilized by linking the two sE monomers through modified sugars; and/or, covalently stabilised by being formed as a single polypeptide chain, optionally with a linker region, optionally a Glycine Serine rich linker region, separating the sE sequences, and/or non-covalently stabilized by substituting at least one amino acid residue in the amino acid sequence of at least one sE monomer with at least one bulky side chain amino acid, at the dimer interface or in domain 1 (D1)/domain 3 (D3) linker of each monomer. A compound, for example an antibody or antibody fragment that can neutralize more than one Dengue virus serotype, for example an antibody that can bind to an EDE of the invention.

Claims

1-155. (canceled)

156. An antibody comprising: a heavy chain comprising CDR regions having amino acid sequences as set out in SEQ ID NOs: 11, 12 and 13, and a light chain comprising CDR regions having amino acid sequences as set out in SEQ ID NOs: 23, 24 and 25, wherein the antibody comprises a Fc domain that is engineered to comprise a modification relative to a wild-type Fc domain.

157. The antibody of claim 156, wherein the heavy chain comprises an amino acid sequence having at least 95% homology to the amino acid sequence set out in SEQ ID NO: 3.

158. The antibody of claim 156, wherein the heavy chain comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 3.

159. The antibody of claim 156, wherein the heavy chain comprises the amino acid sequence set out in SEQ ID NO: 3.

160. The antibody of claim 156, wherein the light chain comprises an amino acid sequence having at least 95% homology to the amino acid sequence set out in SEQ ID NO: 39 or 140.

161. The antibody of claim 156, wherein the light chain comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 39 or 140.

162. The antibody of claim 156, wherein the light chain comprises the amino acid sequence set out in SEQ ID NO: 39 or 140.

163. The antibody of claim 156, wherein the heavy chain comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 3, and wherein the light chain comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 39 or 140.

164. The antibody of claim 156, wherein the antibody is an IgG antibody.

165. An antibody comprising: (a) a heavy chain variable sequence comprising one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 3, and (b) a light chain variable sequence comprising one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 39 or 140, provided that the heavy chain variable sequence comprises CDR regions having amino acid sequences as set out in SEQ ID NOs: 11, 12 and 13, and the light chain variable sequence comprises CDR regions having amino acid sequences as set out in SEQ ID NOs: 23, 24 and 25.

166. The antibody of claim 165, wherein the antibody comprises a Fc domain that is engineered to comprise a modification relative to a wild-type Fc domain.

167. The antibody of claim 165, wherein the antibody is an IgG antibody.

168. A method for producing an antibody, comprising (a) culturing a host cell comprising a nucleic acid encoding the antibody; and (b) recovering the antibody from the culture medium or cultured cells; wherein the antibody comprises a heavy chain comprising CDR regions having amino acid sequences as set out in SEQ ID NOs: 11, 12 and 13, and a light chain comprising CDR regions having amino acid sequences as set out in SEQ ID NOs: 23, 24 and 25.

169. The method of claim 168, wherein the host cell is an insect cell, a C6/36 insect cell, a human cell, a human dendritic cell, a CHO cell, a microorganism, or a Pichia pastoris cell.

170. The method of claim 168, wherein the heavy chain of the antibody comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 3, and wherein the light chain of the antibody comprises no, one or two amino acid substitutions, insertions or deletions compared to the amino acid sequence set out in SEQ ID NO: 39 or 140.

171. The method of claim 168, wherein the antibody comprises a Fc domain that is engineered to comprise a modification relative to a wild-type Fc domain.

172. The method of claim 168, wherein the antibody is an IgG antibody.

Description

FIGURE LEGENDS

[0658] FIGS. 1A-C. Characterization of human anti-DENV monoclonal antibodies. (FIG. 1A) Serotype specificity and reaction to DENV envelope protein by Western Blot (WB) of 145 DENV mAbs. (FIG. 1B) Serotype specificity of WB positive and EDE (WB negative) antibodies. (FIG. 1C) Schematic of epitope mapping using a panel of mutant VLPs. The position of mutations are shown relative to the domain structure of dengue envelope protein with red, yellow and blue representing domain I, II and III, respectively. Positions of mutations marking the fusion loop around W101 and disrupting the N153 glycosylation motif are shown.

[0659] FIG. 2. EDE antibodies are potent and highly crossreactive in neutralization assays.

[0660] Neutralization assays performed on Vero cells for 9 representative mAbs against all four DENV serotypes produced in C6/36 insect cells (3 each of FL, EDE1 and EDE2). The data were from 3 independent experiments.

[0661] FIGS. 3A-B. EDE-specific antibodies have superior neutralizing activities. Neutralization assays using Vero as the target cell were performed on DENV2 generated from C6/36 insect cells C6/36-DENV. (FIG. 3A) or dendritic cells DC-DENV (FIG. 3B). Red, blue and green bars represented FRNT (% reduction) for mAbs at 0.05, 0.5 and 5 g/ml (final concentration), respectively. The EDE mAbs used in the accompanying paper are marked below the histograms. Antibodies are classified into FL and EDE whereas antibodies positive by WB, which failed to map on the VLPs are termed non-FL. Based on the results of VLP mapping five subgroups of the EDE were identified referred to as EDE15. Titration curves for binding, measure by capture ELISA, and neutralization of DC and C6/36 produced viruses with 2 representative antibodies from the FL and EDE1&2 groups. The data were from 2 independent experiments and are representative of the results from 9 each of anti-FL and anti-EDE1 antibodies and 7 anti-EDE2 antibodies.

[0662] FIG. 4A-B. Antibody binding to viral particles in differing states of maturation. (FIG. 4A) Anti-prM and anti-E ELISAs were used to calculate a ratio of prM:E on the various viral particles and compared to virus from LoVo cells which was defined as 100% prM content. (FIG. 4B) Antibody binding to viral particles in differing states of maturation. Binding of representative mAbs to DENV2 produced from C6/36, DC, 293T, furin-transfected 293T, LoVo cells or acid-treated DENV2 were measured by capture ELISA. Two each of FL, VDE1 and VDE2 mAbs are shown together with a VDE4 mAb sensitive to acid treated virus. The data were from 2 independent experiments and are representative of the results from 8 anti-FL, 10 anti-VDE1 and 8 anti-VDE2 antibodies.

[0663] FIG. 5A-C. FL vs. EDE mAbs from individual patients. (FIG. 5A) Distribution of the FL and EDE responses between patients are shown, FL and EDE. The number in the centre indicates the number of Abs from each patient, one copy of three duplicate antibodies (1 EDE1 and 2 EDE2) from patient 752 which have identical amino acid sequences were excluded from this and all other analyses. (FIG. 5B). FL vs. EDE mAbs from individual patients. ADE, U937 cells were infected with DENV2, grown in either C6/36 cells or DC in the presence of titrations of anti-E mAbs reacting to the FL or EDE. The results are expressed as median peak enhancement (fold) from two independent experiments. 752-2C8, 753(3)C10, 747(4)A11 and 747(4)B7 mAbs used in the accompanying paper. (FIG. 5C). FL vs. EDE mAbs from individual patients. NT50 and NT90 titres between the FL and EDE mAbs are compared on C6/36 and DC virus.

[0664] FIG. 6. Schematic of epitope mapping using a panel of mutant VLPs. The position of mutations are shown relative to the domain structure of dengue envelope protein with red, yellow and blue representing domain I, II and III, respectively. Positions of mutations marking the fusion loop around W101 and disrupting the N153 glycosylation motif are shown.

[0665] FIG. 7. Germline analysis of EDE1 and EDE2 anti-EDE antibodies.

[0666] FIG. 8A-B. Recombinant E dimers bind both EDE1 and EDE2 antibodies. (FIG. 8A) SEC/MALS analysis of recombinant sE protein from the four dengue serotypes in complex or not with the BNA (broadly neutralizing antibodies) Fab fragments. MALS showed that the SEC chromatograms of sE protein (green curves) correspond only to the monomer fraction, and the dimer is not detected. The two peaks observed for sE serotypes 1, 3 and 4 correspond to monomers. The most likely explanation is that the dimer affinity is not high enough, and the dimer-monomer equilibrium is disrupted in the gel filtration column, such that the dimer dissociates. The complex with the Fab fragments clearly stabilizes the dimeric form, allowing the elution of a sE dimer in complex with two Fab fragments (red curves). We noted that only the sE monomer fraction eluting late is converted to complex, whereas the other peak remains unchanged (this was most clear with DENV-3 sE, but it also holds for the other serotypes). For DENV-2 sE, there is a single sE peak with a tail towards the small molecular weights, which disappears upon complex formation. Our explanation for this behavior is that the exposed fusion loop of the fraction containing sE competent to form dimers has a tendency to interact with the support, which is why it elutes so late, similar to our previous information with the alphavirus fusion protein El.sup.28. (FIG. 8B). Recombinant E dimers bind both EDE1 and EDE2 antibodies. Real-time SPR profiles corresponding to the interactions of sE with tethered Fab fragments of EDE2 All, EDE2 B7, EDE1 C8, EDE1 C10. The binding of Fab 5H2, which is specific for DENV-4, is shown as a positive control, given that its well-characterized epitope.sup.29 is not at the dimer interface and does not require dimer formation for binding. The Fab fragments were immobilized to similar densities on a Proteon XPR36 chip. 2 M solutions of DENV sE proteins from four serotypes (as indicated), were injected simultaneously over all the Fabs (see the Online Extended Methods section). SPR signal is presented in response unit (RU) as a function of time in seconds (s). Note that the level of binding is in general agreement with the SEC/MALS plots of the corresponding antibodies in panel FIG. 8A.

[0667] FIG. 9. Crystallographic statistics

[0668] FIGS. 10A-B. Crystal structure of the unliganded DENV-2 FGA02 sE dimer. (FIG. 10A) Comparison with the available structure of sE. Of the three available structures of sE in its pre-fusion form (PDB codes 10AN, 10KE, 1TG8), the one with the PDB code 1OAN displayed the smaller root mean square deviation with the unliganded FGA02 sE structure. (FIG. 10B). Crystal structure of the unliganded DENV-2 FGA02 sE dimer. Phylogenetic tree to position the two genotypes of DENV-2 represented by the structures.

[0669] FIGS. 11A-F. DENV2 sE in complex with four EDE anti-EDE antibodies. (FIG. 11A) Complex with the EDE2 A11 Fab fragment. The sE dimer is oriented with the 2-fold molecular axes vertical and with the viral membrane-facing side below. The sE protomers are shown as surfaces according to domains (I, II and III respectively), and the fusion loop (labeled for one protomer in 11b). Foreground and background sE subunits are distinguished in bright and pale colors. The two N-linked glycan chains (not included in the surface) are shown as ball-and-stick according to atom type (carbon, oxygen, nitrogen) and labeled (N67 and N153). The A11 Fab is shown in ribbon representation with heavy and light chains, respectively. (FIG. 11B). DENV2 sE in complex with four EDE anti-EDE antibodies. The unliganded DENV2 FGA02 sE dimer seen down the 2-fold axis (labeled 2 at the center). Ovals (labeled VH and VL) show roughly the contact sites of heavy and light chains, respectively, with the VHs closer to the 2-fold molecular axis. Polypeptide segments and loops relevant to the description of the epitopes are labeled. (FIG. 11C). DENV2 sE in complex with four EDE anti-EDE antibodies. View down the empty arrow shown in FIG. 11 B, highlighting the fusion loop valley encased between two ridges, the b strand on one subunit and the 150 loop on the other. (FIG. 11D-F). DENV2 sE in complex with four EDE anti-EDE antibodies. The same view as in FIG. 11C, showing the complexes with anti-EDE antibodies EDE2 B7 (FIG. 11D), EDE1 C8 (FIG. 11E) and EDE1 C10 (FIG. 11F) (only the variable domains are shown). A star in FIG. 11E and FIG. 11F mark the region of the 150 loop, which is disordered in those complexes. Note that in the B7 and A11 complexes, the light chain is too far from domain III to interact with it, in contrast to the C8 and C10 complexes.

[0670] FIGS. 12A-E. Overall complexes and imprint of the anti-EDE antibodies on the sE dimer.

[0671] Each row corresponds to a different sE/BNA (broadly neutralizing antibody) complex (except for the first one, which shows the unliganded sE dimer) and each column displays the same orientation, as labeled. In the first two columns the sE dimer is depicted as ribbons and the BNA variable domains as surface colored as in FIGS. 11A-F. In the side view (left column) the viral membrane would be underneath, whereas the bottom view (middle column) corresponds to the sE dimer seen from the viral membrane, with the antibodies visible across the sE ribbons. The top view (right column) shows the sE surface as presented to the immune system on the viral particle, showing the imprint of the antibodies (green) with a white depth-cuing fog. For clarity, a white outline delimits the green imprint on the blue surface of domain III. As a guide, in the top-left panel the glycan chains of foreground and background subunits are labeled in red and black, respectively. The fusion loop and if the loop are labeled on the top-middle panel, and can be seen in the other rows in contact with the anti-EDE antibodies. A red star in the left panels of FIGS. 12C-E mark the location of the 150 loop, which is disordered in the complexes with the EDE1 anti-EDE antibodies. This loop bears the N153 glycan recognized by the EDE2 anti-EDE antibodies, as seen in FIG. 12B, left panel (glycan shown as sticks with carbon atoms colored red). In contrast, all the anti-EDE antibodies are seen contacting the N67 glycan, with C8 displaying the most contacts (row c, left panel, N67 glycan as sticks with carbon atoms yellow). A blue star in FIG. 12C shows a disordered loop in domain III. Note that EDE2 C10 (FIGS. 12D and 12 E) inserts deeper into the sE dimer than the other anti-EDE antibodies.

[0672] FIG. 13. Buried surface areas in the various BNA complexes

[0673] FIG. 14. Electrostatic potential of DENV-2 sE complex, epitopes and paratopes. Open book representation of the complexes, with negative and positive potential displayed and colored according to the bar underneath. Because certain regions are disordered in the complexes, the DENV-2 sE dimer model, generated as described in the Online Methods section, was used to calculate the surface electrostatic potential of the sE dimer. Corresponding areas in contact are indicated by same colored ovals.

[0674] FIG. 15A-B. Residues involved in BNA/antigen interactions. (FIG. 15A) Amino acid sequence alignment of sE from the four DENV serotypes, with residues in black or light blue background highlighting identity and similarity, respectively, across serotypes. The secondary structure elements are displayed and labeled underneath the alignment, with the tertiary structure arrangement indicated by colors as in FIG. 11A-F. DENV2 sE residues contacted by the various anti-EDE antibodies are indicated above the alignment, indicating the BNA region in contact by the code provided in the key. Full symbols correspond to contacts in the reference subunit (defined as the one contributing the fusion loop to the epitope), and empty symbols to contacts across the dimer interface. Colored boxes on the sequence highlight the 5 distinct regions of sE making up the epitopes. FIG. 24A provides a histogram with the number of atomic contacts per sE residue in the complex with each BNA. Because the EDE2 B7 and A11 contacts are very similar, only the B7 contacts are shown here. The question mark on the 150 loop indicates that these residues are likely to contact the EDE1 anti-EDE antibodies, but are not visible in the structure because the loop is disordered. (FIG. 15B) Alignment of the four anti-EDE antibodies crystallized, numbered according to the Kabat definition.sup.30 and with the FRW and CDR regions in grey and white background, respectively. The CDRs corresponding to the IMGT convention.sup.31 are marked with a blue line over the sequence and labeled. Somatic mutations are in red with the corresponding residue in the germ-line written in smaller font underneath. Residues arising from the VDJ (or VJ) recombination process are in green. The sE segment contacted (corresponding to the colored boxes in FIG. 15A is indicated above each sequence, coded as indicated in the key. The secondary structure elements of the EDE2 C8 Ig -barrels are indicated above the sequence, as a guide. The histograms of number of contacts per BNA residue for each crystallized complex are provided in FIG. 26A.

[0675] FIGS. 16A-E. Comparison of the various BNA interactions with DENV-2 sE. (FIG. 16A) The structure of the unliganded EDE2 All scFv (red, unbound, 1.7 A resolution) superposed to the variable domain of Fab All in complex with DENV-2 sE (yellow, 3.8 A resolution), to show that the same conformation is retained in the sE/Fab fragment complex. (FIG. 16B) Stereo view showing the superposed B7 (green) and A11 (yellow) variable domains, together with the 150 loop extracted from the structures of the corresponding Fab/DENV-2 sE complexes. Note that the main chain of the 150 loop adopts different conformations in the two complexes, mainly because of the hydroxyl group of the Y99 side chain in the CDR H3 of B7 makes a hydrogen bond with sE T155. All has a phenylalanine at this position, and so lacks the hydroxyl group. The sE protein in the complex with A11 displays the same conformation as the unliganded sE (not shown). (FIG. 16C) Histograms of the atomic contacts of B7 (above the sE sequence) and A11 (below the sequence). (FIG. 16D) As in FIG. 16C, but showing the pattern of contacts made by BNA C10 (above the sequence) and C8 (below) on the sE protein primary structure. (FIG. 16E). As in FIG. 16C but comparing C10 (EDE1) and B7 (EDE2) along the E protein sequence.

[0676] FIG. 17. The H3 loop in the EDE anti-EDE antibodies and in anti-HIV broadly neutralizing antibody PG9. The Fab or scFv fragments are oriented identically, with the light chain in grey and the heavy chain in green, with the H3 loop highlighted in red. For comparison, the Fab fragment of the potent anti HIV-1 BNA PG9.sup.51, which recognizes the glycan chains to a large extent and has a very long H3 loop (30 aa as calculated by IMGT) is displayed in the same way.

[0677] FIGS. 18A-D. Interactions of the BNA CDRs at sE dimer interface. (FIG. 18A) The right panel shows the sE dimer as ribbon, with the epitope area enlarged in the left panel, with main features labeled. FIGS. 18B-D show the sE surface in a semitransparent representation with the ribbons visible through. The glycan residues are displayed as sticks (and were not included in the calculation of the surface). The relevant CDR loops of the anti-EDE antibodies are shown as ribbons with side chains as sticks on top of the sE protein, colored as in FIGS. 11A-F. The orientation of the left panel in FIGS. 18B-D corresponds to the left panel of FIG. 18A, and the right panel is a view along the arrow in FIG. 11B. Hydrogen bonds are displayed as dotted lines.

[0678] FIGS. 19A-E. Interactions with the glycan chains. (FIG. 19A) The EDE1 C8/DENV-2 sE complex shown as ribbons with selected side chains as sticks, with the sE surface in semi-transparent representation, highlighting the interactions with the N67 glycan. A few hydrogen bonds are displayed as dotted lines. The 150 loop is disordered (labeled at the density break). In both FIG. 19 A and FIG. 19 B, the right panel is a view down the arrow of the left panel, through the glycan chain. (FIG. 19B) The EDE2 B7 BNA/DENV-2 sE complex shown with B7 in the same orientation as C8 in FIG. 19A to highlight that EDE1 and EDE2 anti-EDE antibodies bind in a similar way. EDE2 B7 inserts its long CDR H3 into the fusion loop valley, while its sides contact the two glycan chains, as seen in the left panel. The H3 -helix packs against the N153 glycan. Also shown are a number of hydrogen bonds between anti-EDE antibodies and sE, which are listed in FIG. 21. (FIG. 19C) to the sugar connectivity and nomenclature used in the text and in FIGS. 19 A and 19 B. (FIG. 19D) Contacts of the sugar residues with the antibodies, coded according to the key (same as in FIG. 24A-E). (FIG. 19E) Dengue variants lacking the N153 glycan are more readily neutralized by EDE1 anti-EDE antibodies, whereas they are more resistant to neutralization by EDE2 anti-EDE antibodies. Mean and s.e.m. values were estimated from three independent experiments.

[0679] FIGS. 20A-C. Experimental electron density for the glycan chains. Ribbon representation of (FIG. 20A) the EDE2 All Fab and (FIG. 20B) the EDE2 C8 EDE1 Fab in complex with DENV2 sE, colored as in FIG. 11A-F. The simulated annealing omit maps contoured at 1 sigma (cyan) or 0.6 sigma (gold) show clear density for the N153 (in FIG. 20A) and N67 glycans (in FIGS. 20A and 20B) (red and yellow arrowheads, respectively). To create an unbiased map, all glycan atoms were removed from the structures, all B factors were reset to 20 A.sup.2 and the structures were re-refined using torsion dynamics simulated annealing. Note that the antibody footprint spans the two glycans across the dimer interface (as also shown in FIG. 11A-F). (FIG. 20C) Zoom of complex viewed down the red (left panel) and yellow (right panel) arrowheads of FIGS. 20A and 20B respectively for DENV-2/B7 (left panel) and DENV-2/C8 (right panel). Heavy and light chains are shown as surfaces and the glycans (together with the corresponding asparagine) are labeled with the average B factor for each residue. Also, the glycans are ramp-colored from blue (cold) to red (hot). Note that, in order to show the omit map for the glycans in the three complexes, we have displayed in FIG. 20C the electron density for DENV-2/B7 instead of DENV-2/A11 as in FIG. 20A

[0680] FIG. 21. Polar interactions between antibody and antigen.

[0681] FIGS. 22A-B. C10 BNA imprints on sE dimers of dengue serotypes 2 and 4. (FIG. 22A)Surface representation of DENV-2 sE as viewed from outside the virion, with exposed main chain atoms orange (top panel) or with main chain atoms+strictly conserved side chains in orange, and highly conserved side chains in yellow (bottom panel). The epitopes of EDE1 BNA C10 (black outline) and a EDE2 BNA B7 (green outline) are indicated. (FIG. 22B) The surface of the DENV-2 (top panel) and DENV-4 (bottom) sE dimer extracted from the complex with C10, color-coded by domains as in FIG. 11A-F. The C10 footprint is shown in each case. Note the asymmetry in conformation. The bigger hole on the right hand side of the dimer (in both panels) is due to the ij loop being disordered, and the smaller one on the left (bottom panel, DENV-4 sE/C10) is due to the kl loop disordered.

[0682] FIG. 23A-BComparison of the C10 interactions with DENV-2 and DENV-4 sE. FIG. 23A) The structures of DENV-2 (red) and DENV-4 (grey) sE dimers in complex with C10 were superposed on the C10 moiety. The axes of the sE dimers are drawn at the center, colored accordingly. For clarity, only the C10 scFv on which the superposition was made is displayed per complex. Upon superimposing the antibodies, the sE dimers match only locally, resulting in slightly different orientations of the dimer axes, as drawn. The sE dimers become rotated with respect to each other by about 6 degrees about the axes drawn in blue (labeled with a grey/red curved arrow), which have strikingly different orientations when the superposition is done on the scFv on the left (bound to epitope A) than on the one on the right (bound to epitope B), highlighting the asymmetry of C10 binding to the sE dimer. FIG. 23B) The C10 contacts plotted on an alignment of DENV-2 (above) and DENV-4 sE (below), showing that there is a very similar pattern of contacts in the two complexes. The background of the sequence corresponds to that of FIG. 15A in the main text, showing the conservation on the four serotypes.

[0683] FIGS. 24A-E. Detected asymmetry of BNA binding to the sE dimer. These Figures provide the histograms of number of contacts per sE residue in the structures of all of the independent complexes analyzed here, in FIGS. 24A-E: (FIG. 24A) EDE1 C8/DENV-2 sE; (FIG. 24B) EDE1 C10/DENV-2 sE; (FIG. 24C) EDE2 A11/DENV-2 sE; (FIG. 24D) EDE2 B7/DENV-2 sE; (FIG. 24E) EDE1 C10/DENV-4 sE. Each panel is divided in two portions: part (I) displays the immune complex viewed down the 2-fold axis of the sE dimer on the left (for clarity, the constant domain of the Fab fragments was removed), and on the right with the antibody removed altogether, to show the epitope. This part also defines epitopes A and B used in II. Part (II) shows the histogram of contacts corresponding to the A and B epitopes in the dimer, with the histogram bars color-coded as indicated in the key, to map the antibody region involved in the contact (in parenthesis is the symbol used in FIG. 15 A to mark the corresponding contact). Note that the contacts pattern remains the same, but the number of contacts is not identical on the two epitopes of the sE dimer. The crystals of DENV-2 sE/EDE2 C10 had two complexes in the asymmetric unit (i.e., two sE dimers, each with two C10 scFv), so that in total there are 4 independent views of the epitope, labeled A-D, described in FIG. 24 B.

[0684] FIGS. 25A-C. EDE1 C10 residue Y100 (CDR H3) is likely to interact with F279 on mature dengue virions. (FIG. 25A) One of the 90 E/M heterotetramers that compose the mature DENV-2 particle was extracted from the 3.5 cryo-EM reconstruction.sup.9, and is displayed in side view (2-fold axis vertical, drawn as a white rod labeled 2) with the two E subunits in grey and yellow and the two M subunits in red and salmon. The two black arrows indicate the connection between the E ectodomain (which corresponds to sE) with the -helical membrane interacting region (the horizontal stem -helices and the vertical TM helices). The N-terminal segment of M is seen interacting underneath the E dimer (pink arrow). (FIG. 25B) View down the 2-fold axis, with the region magnified in FIG. 25C framed. (FIG. 25C) The view has been slightly tilted, for clarity, with respect to the view in FIG. 25 B, with the structures of both C10 complexes (DENV-2 and -4 sE) superposed onto virion E. The labels match the color of the corresponding structures (DENV-2 sE/C10 green/mustard; DENV-4 sE/C10 blue/beige, and virion E as in FIGS. 25A and 25B. The M protein is shown as a salmon surface (labeled in white). It lies underneath the E dimer, where it buttresses the base of the kl hairpin (comprised between the arrows, labeled)) and also the ij hairpin across the dimer interface, inducing a different conformation of the kl hairpin, such that F279 (labeled) points away from the hydrophobic core of the E protomer (dark grey sticks), whereas in the sE protein structures (including unliganded sE, not shown) it is part of the hydrophobic core (green and blue sticks, labeled). The side chain of Y100 (labeled) in CDR H3 of C10 has alternative conformations because it doesn't find its partner in sE (Y100 is also illustrated in FIG. 18 D, left panel). The CDR H3 loop is flexible enough so that the Y100 could make a stacking interaction with F279 (which is conserved across serotypes, see FIG. 15 A) in the conformation observed on the virion.

[0685] FIGS. 26A-E. Histogram of contacts on the antibody residues. This Figure mirrors FIG. 24A-E, this time showing the contacts on the antibody side. (FIG. 26A) C8; (FIG. 26B) C10 (from the complex with DENV-2 sE) (FIG. 26C) All; (FIG. 26 D), B7; (FIG. 26 E) C10 (from complex with DENV-4 sE). In each panel, Part I shows the BNA variable domain extracted from the corresponding complex, colored grey (VH dark grey, VL light grey) with somatic mutations in red and junction residues arising from recombination in green. Side chains involved in contacts are displayed in ball and stick and labeled. Part II shows the histogram of the number of atomic contacts per residue, color-coded according to the key to indicate the region of sE that is contacted (in parenthesis, the symbol used in FIG. 15B to mark the corresponding contact). The sequence numbering and the background corresponds to Kabat convention (as in FIG. 15B). The CDRs corresponding to the IMGT convention are displayed as dotted orange lines above the sequences. Somatic mutations are in red, residues arising from VDJ (or VI) recombination are in green.

[0686] FIG. 27. Comparison with the binding properties of potent anti-EDE antibodies targeting other viruses.

[0687] FIG. 28. Sequences of the Dengue envelope protein from serotypes 1 to 4

[0688] FIG. 29. Sequences of the EDE1 and EDE2 type antibodies identified in Example 1.

[0689] FIGS. 30A-D. Methods of performing a neutralisation test.

[0690] FIG. 31. Contact residues in the envelope protein derived from crystal structures.

[0691] FIG. 32. Covalently cross-linked DV2 E dimers.

[0692] FIG. 33. Binding of EDE1 to rE DENV2 WT vs MT A259C

[0693] FIG. 34. Binding of EDE2 to rE DENV2 WT vs MT A259C

[0694] FIG. 35. Binding of FLE to rE DENV2 WT vs MT A259C

[0695] FIG. 36. Binding of Non-FLE to rE DENV2 WT vs MT A259C

[0696] FIG. 37. Binding of EDE1 to rE DENV2 WT vs MT L107C, A313C

[0697] FIG. 38. Binding of EDE2 to rE DENV2 WT vs MT L107C, A313C

[0698] FIG. 39. Binding of FLE to rE DENV2 WT vs MT L107C, A313C

[0699] FIG. 40. Binding of non-FLE to rE DENV2 WT vs MT L107C, A313C

[0700] FIG. 41. Antibody titration on C6/36 DENV2 [0701] Group 1 monomer/monomer [0702] prime and boost with E WT. E WT [0703] Group 2 dimer/dimer [0704] prime and boost with E A259C mutant [0705] Group 3 VLP/VLP [0706] prime and boost with prM/E viral like particle (VLP) [0707] Group 4 dimer/VLP [0708] prime with E A259C mutant followed by boosting with VLP [0709] Group 5 VLP/dimer [0710] prime with VLP followed by boosting with E A259C mutant [0711] Group 6 mock [0712] non immunisation

[0713] FIG. 42. Cross reactivity: Binding to live virus (pooled serum) [0714] Group 1 monomer/monomer [0715] prime and boost with E WT. E WT [0716] Group 2 dimer/dimer [0717] prime and boost with E A259C mutant [0718] Group 3 VLP/VLP [0719] prime and boost with prM/E viral like particle (VLP) [0720] Group 4 dimer/VLP [0721] prime with E A259C mutant followed by boosting with VLP [0722] Group 5 VLP/dimer [0723] prime with VLP followed by boosting with E A259C mutant [0724] Group 6 mock [0725] non immunisation

[0726] FIG. 43. Neutralisation of mouse serum: C6/36 DENV2 [0727] Group 1 monomer/monomer [0728] prime and boost with E WT [0729] Group 2 dimer/dimer [0730] prime and boost with E A259C mutant [0731] Group 3 VLP/VLP [0732] prime and boost with prM/E viral like particle (VLP) [0733] Group 4 dimer/VLP [0734] prime with E A259C mutant followed by boosting with VLP [0735] Group 5 VLP/dimer [0736] prime with VLP followed by boosting with E A259C mutant [0737] Group 6 mock [0738] non immunisation

[0739] FIG. 44. Neutralisation of mouse serum: DC DENV2 [0740] Group 1 monomer/monomer [0741] prime and boost with E WT. E WT [0742] Group 2 monomer/monomer [0743] prime and boost with E A259C mutant (dimer/dimer) [0744] Group 3 VLP/VLP [0745] prime and boost with prM/E viral like particle (VLP) [0746] Group 4 dimer/VLP [0747] prime with E A259C mutant followed by boosting with VLP [0748] Group 5 VLP/dimer [0749] prime with VLP followed by boosting with E A259C mutant [0750] Group 6 mock [0751] non immunisation

[0752] FIG. 45. ADE: Pooled mouse serum: U937 [0753] Group 1 monomer/monomer [0754] prime and boost with E WT. E WT [0755] Group 2 monomer/monomer [0756] prime and boost with E A259C mutant (dimer/dimer) [0757] Group 3 VLP/VLP [0758] prime and boost with prM/E viral like particle (VLP) [0759] Group 4 dimer/VLP [0760] prime with E A259C mutant followed by boosting with VLP [0761] Group 5 VLP/dimer [0762] prime with VLP followed by boosting with E A259C mutant [0763] Group 6 mock [0764] non immunisation

EXAMPLES

Example 1Human DENV Antibodies Form Two Distinct Groups Based on their Ability to Bind to Dengue Envelope Protein on a Western Blot

[0765] Samples from 7 patients (Table 1) were used to produce 145 human monoclonal antibodies reacting to the DENV envelope protein.sup.32,33. Plasmablasts (CD3.sup., CD20.sup.lo/, CD19.sup.+, CD27.sup.hi, CD38.sup.hi) were sorted from peripheral blood; Elispot demonstrated 5090% of these cells secreted anti-DENV antibodies, consistent with frequencies reported by others.sup.34,35. 84% of these antibodies reacted against all four DENV serotypes, 13% reacted to 2 or 3 serotypes and only 3% reacted to a single serotype (FIG. 1a).

TABLE-US-00001 TABLE 1 Summary of DENV-infected patients enrolled in the study Frequency of Frequency of DENV-specific plasmablasts B cells vs. total vs. total IgG + IgM No. of Patient Serotype of Day of CD19+ cells secreting cells anti-E id Severity infection Serology illness (%) (%) Abs / 747 DHF DENV2 Secondary 6 64.9 76.9 18 7/11 749 DF DENV1 Secondary 4 56.7 62.4 11 1/10 750 DHF DENV1 Secondary 5 67.9 71.5 17 5/12 751 DF DENV1 Secondary 4 32.7 75.6 15 8/7 752 DHF Unknown Primary 4 68.3 47.0 32 31/1 753 DHF DENV1 Secondary 5 74 89.9 35 17/18 758 DHF Unknown Secondary 5 ND ND 17 6/11

[0766] The initial antibody screen was performed by ELISA using captured whole virions, rather than recombinant protein or fixed cells, to make sure we obtained a fully representative panel of antibodies. Only 57% of the antibodies reacted to DENV envelope by Western Blot (FIG. 1a), the WB negative mAbs also failed to react to recombinant E by ELISA. This allowed us to group the antibodies into two broad groups, WB reactive and those which only recognize an epitope present on the intact virion; from hereon we refer to these antibodies as reactive to virion dependent epitopes or Envelope Dimer Epitope (VDE or EDE). Most of the WB positive antibodies were fully crossreactive between the four virus serotypes whilst for the EDEEDE mAbs 41/62 were fully crossreactive with a further 17/62 reacting against DENV-1, 2&3 (FIG. 1b). Neutralization assays on virus produced in C6/36 insect cells for three antibodies from each group are shown in FIG. 2. The fusion loop and EDEEDE antibodies were broadly neutralising against all four virus serotypes. For EDE2 747(4)A11 and 747(4)B7 there was lower activity to Den4, which relates to the lack of N-linked glycan at position N153 in the Den4 strain H241.

Methods Relevant to this and Other Examples

[0767] Samples. Blood samples were collected from inpatients following written informed consent. The study protocol was approved by the Scientific and Ethical Committee of the Hospital for Tropical Diseases, the Oxford Tropical Research Ethical Committee and the Riverside Ethics Committee in the UK. Laboratory confirmation of dengue infection was determined by RT-PCR detection of DENV nucleic acid (which also confirmed the infecting serotype), NS1 antigen lateral flow test or seroconversion in an IgM ELISA test. Disease severity was classified according to 1997 World Health Organization criteria. Of the patients enrolled in the study, 2 patients were classified as mild symptom Dengue Fever (DF) and 5 patients were classified as severe symptom with plasma leakage and bleeding Dengue Heamorrhagic Fever (DHF) (Table 1). Secondary infection were defined based on the ratio of dengue specific IgM to IgG less than 1.8 7. Blood samples for B cell sorting were collected during the hospitalization period at time points where the blood plasmablast population was apparent. PBMCs were isolated from whole blood by Ficoll-Hypaque density gradient centrifugation and resuspended in 10% FCS/RPMI for immediate use.

[0768] Cells and antibodies. The C6/36 cell line, derived from the mosquito Aedes albopictus, was cultured in Leibovitz L-15 at 28 C. Vero, U937 and 293T or furin-transfected 293T cells were grown at 37 C. in MEM, RPMI 1640 and DMEM respectively. All media was supplemented with 10% heat-inactivated foetal bovine serum (FBS), 100 units/ml penicillin, 100 g/ml streptomycin and 2 mM L-Glutamine. The furin-deficient LoVo cell line was purchased from ATCC and maintained in F-12 as recommended. Monocytederived dendritic cells (DC) were prepared as previously described.sup.10.

[0769] Antibodies against human CD3, CD19, CD20, CD27 and CD38 (BD Pharmingen), AntiHuman IgG-ALP (Sigma) and anti-Human or mouse IgG-HRP (DAKO) were used in the experiments. anti-DENV envelope, 4G2, and anti-DENV prM, 1H10, murine monoclonal Abs were gifts from Dr C. Puttikhunt and Dr W. Kasinrerk (Puttikhunt, 2003). anti-DENV NS3, E1D8 was a gift from Prof. Eva Harris.

[0770] Virus stock. Dengue virus serotype 1 (Hawaii), serotype 2 (16681), serotype 3 (H87) and serotype 4 (H241) were grown in C6/36 cells. In addition, DENV2 was propagated in DC, LoVo, 293T and Furin-transfected 293T and cell-free supernatants were collected and stored at 80 C. Viral titres were determined by a focus-forming assay on Vero cells and expressed as focus-forming units (FFU) per ml.sup.26

[0771] Generation of DENV-specific human monoclonal Abs. DENV-specific human mAbs were generated from activated B cells/plasmablasts.sup.32,33. Briefly, PBMC were stained with anti-CD3, CD19, CD20, CD27 and CD38. Activated antibody secreting cells (ASCs) were then gated as CD19.sup.+, CD3.sup., CD2.sup.lo/, CD27.sup.high, CD38.sup.high. Single ASCs were sorted into each well of 96 well PCR plates containing RNase inhibitor (Promega). Plates were centrifuged briefly and frozen on dry ice before storage at 80 C. RT-PCR and nested PCR were then performed to amplify Gamma, Lambda and Kappa genes using cocktails of primers specific for IgG. PCR products of heavy and light chains were then digested with the appropriate restriction endonuclease and cloned into IgG1, Ig or Ig expression Vectors; gifts from Dr Hedda Wardemann. To express antibodies, heavy and light chain plasmids were co-transfected into the 293T cell line by Polyethylenimine method and antibody supernatant was harvested for further characterization.

[0772] ELISPOT assay. Elispot plates (Millipore) were coated with either anti-human Ig (Invitrogen) or UV inactivated DENV1-4. Plates were washed with RPMI and blocked with 1% BSA/RPMI for 1 hour. Sorted ASCs were added at 500 cells to the anti-Ig and the DENV coated wells and incubated overnight at 37 C. in 5% C02. Plates were washed and incubated with biotinylated anti-human IgG and IgM (Sigma) for 2 hrs at room temperature, followed by Streptavidin-ALP (Sigma). The reaction was developed and spots were counted using an AID Elispot plate reader.

[0773] Detection of DENV-specificity and serotype cross-reactivity by ELISA. DENV1-4 and mock uninfected supernatant were captured separately onto a MAXISORP immunoplate (NUNC) coated anti-E Abs (4G2). DENV captured wells were then incubated with 1 g/ml of human mAbs followed by ALP-conjugated anti-human IgG. The reaction was visualized by the addition of PNPP substrate and stopped with NaOH. The absorbance was measured at 405 nm.

[0774] Recombinant soluble DENV envelope protein ELISA. Plates were directly coated with 150 ng recombinant soluble E and bovine serum albumin (BSA) was used as negative control antigen. Protein coated wells were then incubated with 1 g/ml of human monoclonal Abs followed by ALP-conjugated anti-human IgG. PNPP substrate was finally added and the reaction was measured at 405 nM.

[0775] Western blot analysis. For western blot analysis, DENV supernatant from C6/36 cells was prepared in non-heated and non-reducing conditions and run on 12% SDS polyacryramide gels and electroblotted onto nitrocellulose membranes (Amersham). The membranes were then blocked with 5% skimmed milk and probed with DENV-specific human mAbs followed by HRP-conjugated anti-human IgG Abs, membranes were developed with enhanced chemiluminescence substrate (Amersham).

Example 2Mutational Analysis Reveals that the EDE Antibodies and the WB Reactive Antibodies Bind Distinct Epitopes

[0776] To gain more insight into the epitopes recognized by the mAbs, we created 65 virus like particles (VLP's) containing alanine substitutions at solvent exposed residues predicted to be on the virion surface. These were taken from the 3D structure of the mature virus particles.sup.4, 7, 8. These mutant VLP's were screened against the 145 monoclonal antibodies by ELISA.sup.22, 36. Mutations that resulted in >80% reduction of antibody binding were deemed significant. Using this panel 112 of the 145 mAbs were assigned an epitope on the envelope protein. Thirty three antibodies, all of which react to E by WB, remained unmapped using the mutant VLP panel. The epitope mapping results are shown in FIGS. 1c and 1n more detail in FIG. 6. These epitopes can be broadly clustered into two groups:

[0777] Group 1: Fusion Loop; a restricted set of residues in and around 101W defining the previously described or classical fusion loop epitope (FL). 46 of the 83 antibodies, which bound to envelope on WB, were sensitive to mutation at position 101W, which has been previously shown to be a key residue for the binding of a number of anti-DENV mAbs.sup.37, 38. Of the FL specific mAbs, 40 of the 46 were sensitive to the W101 mutation only whilst other epitopes contained combinations of the residues W101, G106, and L107. The crystal structure of FL mAb E53 bound to the envelope protein from West Nile virus showed contacts with residues 104-110, but not 101W and with resides 74-79 in the bcloop.sup.39. Only two of the FL specific mAbs were sensitive to changes in the bc-loop where binding was lost when amino acids 76-79 were changed to alanine similar to the 1C19 mAb.sup.40.

[0778] Group 2: The EDE antibodies; these could be subdivided into five distinct subgroups based upon the pattern of reactivity to the VLP mutants (FIG. 1c and FIG. 6). The majority of the EDE antibodies were sensitive to changes in fusion loop residue 101, but are not to be confused with the classical Fusion Loop specific antibodies described above as the epitope is much more extensive, with additional determinants on domains I, II and III. The majority of EDE antibodies can be divided between two distinct subgroups, EDE1 and EDE2, differentiated by sensitivity to changes at residues 153 and 155 in EDE2, which will disrupt an N-linked glycosylation site. Twelve antibodies constituted three further subgroups with different epitopes and function; EDE3 mAbs were similar to EDE1 mAbs but also sensitive to changes at 107L and 295K. EDE4 were not sensitive to changes at 101W and reacted best to acidified virus particles (FIG. 4B). Finally, EDE5 mAbs constitutes a group of 5 mAbs, which bind to the fusion loop 101W only, but the epitope is only recapitulated on intact virions, EDE3, 4 and 5 are all poorly neutralizing antibodies FIGS. 3A and 3B.

[0779] The VLP mutagenesis experiments suggest the EDE is a complex quaternary epitope encompassing more than one envelope protomer.

Methods

[0780] Antibody epitope mapping using virus-like particle (VLP) mutants. Full length prM/E of DENV1 was cloned into the expression vector pHLsec to generate VLP (constructed by Dr Aleksandra Flanagan).sup.55. VLP mutants were generated by PCR-based site-directed mutagenesis.sup.62. Mutagenic PCR was performed to substitute selected amino acid residues in the E protein with alanine using Pfx DNA polymerase (Invitrogen), if already alanine, mutation was made to glycine. After DpnI (NEB) treatment, PCR products were transformed into E. coli. All mutations were confirmed by sequencing. Plasmids were transfected into the 293T cell lines by Polyethylenimine method and culture supernatants were harvested for epitope mapping.

[0781] To identify the epitope-specific Ab, WT and mutant VLPs were captured with mouse anti-prM (1H10). DENV-specific human anti-E Abs were then added at 1-5 g/ml followed by anti-human IgG-ALP. Finally, PNPP substrate was added and the reaction was stopped with NaOH and absorbance measured at 405 nm. The relative recognition index was calculated as [absorbance of mutant VLP/absorbance of WT VLP] (recognized by the test mAb)/[absorbance of mutant VLP/Absorbance of WT VLP] (recognized by a group of 4 mixed mAbs).

Example 3The WB Reactive Antibodies are Incapable of Fully Neutralising Virus Made in Human Dendritic Cells, Unlike the EDE Antibodies

[0782] During a DENV infection, the host is presented with two forms of virus; the initial exposure is to virus produced in insect cells, whilst virus produced in human cells drives subsequent rounds of infection and represents the vast bulk of virus encountered during infection. To look at these two different viral forms we compared neutralization of DENV-2 virus produced in C6/36 insect cells (C6/36-DENV) or in monocyte derived dendritic cells (DC-DENV), which are thought to be infected following injection of virus into the skin from the mosquito bite and to be a site of virus replication in the infected human host.sup.20.

[0783] Of the 83 WB positive mAbs, 46 were mapped to the FL and 37 recognised an as yet unmapped binding site. Surprisingly, all 83 WB positive antibodies were incapable of fully neutralizing DC-DENV, even at high concentration, with only one neutralizing to >80% at 5 g/ml (FIG. 3a&b). On the other hand, most of the EDE1&2 antibodies were able to neutralize DC-DENV to >80% with a number reaching 100% neutralization. Full binding and neutralization curves for representative mAbs, demonstrate that anti-FL mAbs have reduced binding by ELISA to DC-DENV and fail to fully neutralize DC virus infection, whereas the neutralization and binding curves for the EDE mAbs are more closely opposed for C6/36 and DC-DENV.

Methods

[0784] Neutralization and enhancement assays. The neutralization potential of mAbs was determined using the Focus Reduction Neutralization Test (FRNT), where the reduction in the number of the infected foci is compared to control (no antibody).sup.22. Briefly, serially-diluted Ab was mixed with virus and incubated for 1 hr at 37 C. The mixtures were then transferred to Vero cells and incubated for 3 days. The focus-forming assay was then performed using anti-E mAb (4G2) followed by rabbit anti-mouse IgG, conjugated with HRP. The reaction was visualized by the addition of DAB substrate. The percentage focus reduction was calculated for each antibody dilution. 50% FRNT values were determined from graphs of percentage reduction versus concentration of Abs using the probit program from the SPSS package.

Example 4Anti-EDE Antibodies Cannot Bind to Virus with a High Proportion of prM or where the Envelope Protein has Adopted the Trimer Conformation

[0785] To represent these different virus forms, we compared antibody binding to 6 DENV-2 preparations. To assess the degree of prM cleavage, we measured the ratio of prM:E by ELISA and normalized this to DENV produced in LoVo cells, which lack furin activity and produce almost completely non-infectious mature virus particles with a full complement of prM.sup.17 (FIG. 4A). The virus preparations were as follows: 1) C6/36-DENV which has a relatively high prM content of 56%, 2) DC-DENV which has a prM content of 13%, 3) virus produced in furin deficient LoVo cells (LoVo-DENV) which have a prM content approaching 100%.sup.22, 4) Virus produced in 293T cells overexpressing furin (Furin-293T-DENV) which have a prM content of 5%, 5) virus produced in native 293T cells having 60% prM (293T-DENV) and 6) Virus incubated at pH 5.5 which irreversibly adopts the E trimer conformation (acid-DENV).sup.42.

[0786] The EDE1&2 mAbs could not bind to acid-DENV, presumably because trimerization destroys the conformational epitope, or bind to LoVo-DENV likely because a full complement of prM supports prM/E spikes, which may again disrupt the mature envelope dimer epitope or may sterically interfere with access to the EDE (FIG. 4B). The antiFL mAbs showed reduced binding to the low-prM content viruses; binding curves for DC-DENV and 293T-Furin-DENV were shifted 1.5-2 logs to the right of C6/36 produced DENV. Additionally, binding to LoVo-DENV was even more efficient than C6/36DENV, underscoring the importance of prM for the exposure and efficient binding of fusion loop antibodies 39-43 The four EDE4 antibodies, which were isolated from three separate individuals, bound most efficiently to acid-treated virus, but showed negligible neutralization (FIG. 4B).

Methods

[0787] DENV binding ELISA. To determine the binding affinity of Ab to DENV generated from different cell types, Mock, DENV2 produced from C6/36, DC, 293T, furintransfected 293T or LoVo cells and acid-treated C6/36 DENV2 was captured onto plates coated with 4G2 and then incubated with serial dilutions of DENV-specific human monoclonal Abs followed by ALP-conjugated anti-human IgG. The reaction was developed by the addition of PNPP substrate and stopped with NaOH. The absorbance was measured at 405 nm. Antigen loading of the different viral forms and inter-day variation in OD readings between experiments was normalised by a control ELISA using a humanised version of the well described 3H5 mAb, which is specific to Domain III of DENV2.

Example 5the Antibodies within a Particular Patient Show Immunodominance

[0788] The anti-DENV mAbs described here are a complex ensemble of overlapping specificities where the EDE overlaps with the more restricted epitope of FL antibodies. When we compared these antibody groups (FL vs. EDE) within the individual patients we found skewed repertoires showing a preference to pick either the FL or EDE epitopes (FIG. 5A). This immunodominance of recognition within an individual was surprising. As the epitopes are overlapping, it is possible that the most avid antibody would compete off other antibodies, affinity mature and hence dominate the response leading to a stochastic choice between FL and EDE. However, the responses to the EDE or FL are polyclonal (different VDJ recombinations) within individuals, making this less likely as an explanation.

Example 6Anti-EDE Antibodies Cause a Reduced Level of Antibody Dependent Enhancement of Infection

[0789] We tested the ability of the antibodies to enhance DENV infection in Fc receptor expressing U937 cells. All antibodies tested caused ADE, however it was around 4-8 fold less in the EDE group when compared to FL group; the median peak enhancement for the FL vs. the EDE groups were 3745:545 on C6/36-DENV and 2070:480 on DC-DENV (FIG. 5B).

Methods

[0790] or the ADE assay, serially-diluted Ab was pre-incubated with virus for 1 hr at 37 C., then transferred to U937 cells (Fc receptor-bearing human monocyte cell lines) and incubated for 4 days. Supernatants were harvested and titrated on Vero cells by a focus forming assay. The titres of virus were expressed as focus-forming units (FFU) per ml and the fold increment was calculated by comparing the viral titre in the absence of antibody.

Example 7the Anti-EDE Antibodies Bind Recombinant sE Dimer

[0791] For the structural studies we selected four of the most potent anti-EDE antibodies identified: 747(4) A11 and 747 B7 (EDE2) and 752-2 C8 and 753(3) C10 (EDE1), from hereon referred to as A11, B7, C8 and C10. Both EDE2 anti-EDE antibodies were isolated from the same patient (who had a secondary infection with DENV-2), and are somatic variants of the same IgG clone, derived from the IGHV3-74 and IGLV2-23 germ lines. The heavy chain has a very long (26 amino acids) complementarity-determining region 3 (CDR H3). The EDE1 anti-EDE antibodies were isolated from different patients and the corresponding germ lines derive from VH and VL genes IGHV3-64 and IGKV3-11, (EDE1 C8, the patient had a primary infection of undetermined serotype) and IGHV1-3* and IGLV2-14 (EDE1 C10, from a patient with secondary DENV-1 infection). The analysis of the genes for these antibodies is summarized in FIG. 7.

[0792] Recombinant sE protein (the 400 amino terminal residues of the ectodomain of Envelope protein, termed sE for soluble E) and the antigen binding portions (Fab) as well as single-chain variable domains (scFv) of the anti-EDE antibodies were produced in Drosophila S2 cells.sup.44, 45. Because the anti-EDE antibodies did not react with recombinant sE protein in standard ELISA assays, we tested the interaction of the antibody fragments with purified recombinant DENV sE in solution at high concentrations to favour dimer formation. Size exclusion chromatography (SEC) combined with multi-angle light scattering (MALS) experiments showed that the dimer/monomer equilibrium of recombinant DENV-1, -2, -3 and -4 sE was shifted to dimer by the antibody fragments, eluting as a complex corresponding an sE dimer with two antibody fragments in most cases, in spite of the size-exclusion induced dissociation effect upon separation of the various species, as shown in FIG. 8A. This was further confirmed by surface plasmon resonance (SPR) analysis (FIG. 8B).

Example 8Crystal Structures

[0793] We determined in total 7 crystal structures, including the DENV-2 sE dimer in complex with fragments of the four selected anti-EDE antibodies and DENV-4 sE in complex with EDE1 C10 in order to confirm the determinants of cross-reactivity. Because the DENV-2 sE dimer used belongs to a different strain from the one for which structures are already available, we also crystallized the unliganded sE dimer to detect possible changes in conformation induced by the antibodies. In addition, we determined the structure of the unliganded A11 scFv, because it was not clear whether its long CDR H3 maintained the same conformation in the absence of antigen. The crystallization procedures are described Example 15 and the crystallographic statistics are listed in FIG. 9.

DENV-2 sE Strain FGA02, Genotype II

[0794] We did most of the structural studies with recombinant sE from DENV-2 field strain FGA02 (isolated in 2002 in French Guiana), which belongs to the Asian/American genotype III.sup.11 within serotype 2. FGA02 sE displays 13 amino acid differences compared to the previously crystallized DENV-2 sE.sup.5,7, scattered over the 394 residues (3%) of the ectodomain. As expected, the 3 resolution structure of FGA02 sE shows only small differences with the already available structure of DENV-2 sE in its prefusion form (FIGS. 10A and 10B), fitting within the range of conformations observed in the various structures deposited in the PDB (accessions 1OAN, 1OKE, 1TG8). The structure of unliganded FGA02 sE was useful in assessing regions in which antibody binding induces disorderin particular, the 150 loop (see below)by showing that it is not due to the specific amino acid sequence of the E protein of this strain.

Example 9the Envelope Dimer Epitope

The Anti-EDE Antibodies Bind at the sE Dimer Interface

[0795] The crystal structures of the FGA02 sE immune complexes show that the four anti-EDE antibodies bind in a similar way (FIGS. 11A-F), interacting with both subunits of the dimer, (see also ED FIGS. 12A-E, which provides the imprints of each anti-EDE antibody on the sE dimer). The heavy chain binds closer to the 2-fold axis (i.e., the center of the dimer) while the light chain is positioned peripherally. The epitopes largely overlap with the imprint of the prM protein on the E dimer in immature DENV particles exposed to low pH.sup.46. They are centered in a valley lined by the b strand on the domain II side, and the 150 loop on the domain I side (across the dimer interface, FIG. 11C). The 150 loop spans residues 148159, connecting -strands E.sub.0 and F.sub.0 of domain I, and carries the N153 glycan, which lies above the fusion loop of the partner subunit in the dimer. The heavy chains span the distance between the two glycan chains, N67 and N153, across the dimer interface (FIG. 11C-F). The total buried surface per epitope ranges between 1050 .sup.2 and 1400 .sup.2, and the surface complementarity coefficient.sup.13 is between 0.67 and 0.74, which are values typical for antibody/antigen complexes (FIG. 13). The surface electrostatic potentials of epitope and paratopes are mildly charged, with a relatively complementary charge distribution (FIG. 14).

Conserved Residues Make Up the Epitopes

[0796] The anti-EDE antibody contacts cluster essentially on highly conserved residues across the four serotypes (FIGS. 15A and 15B), explaining their cross-reactivity. The 26-residue long CDR H3 of B7 and A11 accounts for the vast majority of the EDE2 anti-EDE antibody contacts on both sE subunits forming the epitope. The H3 loop makes a protrusion in the paratope, adopting a convex shape complementary to the concave surface of the antigen (FIG. 11D). The H3 protrusion is pre-formed in the antibody, as shown by the 1.7 resolution structure of the unliganded EDE2 A11 scFv (FIGS. 9, 16A-E and 17), indicating that there is no entropic cost for binding. On the reference subunit, defined as the one contributing the fusion loop to the epitope, both EDE1 and EDE2 anti-EDE antibodies target the same serotype invariant residues, which cluster in three main polypeptide segments of domain II (boxed in FIG. 15A): the b strand (residues 67-74, bearing the N67 glycan), the fusion loop and residues immediately upstream (aa 97-106), and the ij loop (aa 246-249). Whereas both light and heavy chains of the EDE1 anti-EDE antibodies interact with the reference subunit via all three CDRs, the EDE2 anti-EDE antibodies interact essentially with the heavy chain, with only a few light chain contacts from CDR L3 (FIGS. 18A-D). On the opposite subunit, across the interface, the sE segments targeted are different for the two EDE groups. The EDE2 anti-EDE antibodies interact with the 150 loop and the N153 glycan chain (see below), whereas EDE1 anti-EDE antibodies induce disorder of the 150 loop upon binding. This allows the light chain in EDE1 anti-EDE antibodies to come closer to sE and to interact with domain III in the region of the A strand epitope, which has been structurally characterized previously for murine DENV cross-reactive antibodies.sup.47, 48. These domain III contacts are centred on the conserved sE residue K310, the side chain of which makes a lid covering the indole ring of W101 of the fusion loop, in an important stabilizing sE dimer contact (FIGS. 18A-D). Although the light chains derive from different VL genes (FIG. 7), both EDE1 C10 and C8 use CDR L1 and L2 residues to contact domain III (FIGS. 15A-B and FIGS. 18A-D). In domain I, EDE1 C10 inserts its relatively long CDR H3 (21 aa, FIG. 7) such that it interacts with conserved residues underneath the 150 loop (scattered in the N-terminal 50 amino acids of the E protein, see FIG. 15Aalso circled in FIG. 18D, left panel), whereas the shorter H3 loop of EDE1 C8 cannot reach this region.

Example 10Antibody Recognition of the Glycan Chains

[0797] The anti-EDE antibodies make extensive contact with the glycan chains, both at positions N67 and N153 of E (FIGS. 19A-E and FIGS. 20A-C). All four anti-EDE antibodies interact with the N67 glycan via CDR H2 contacts, and will therefore interfere with binding to the DC-SIGN receptor of dendritic cells, which was shown to interact specifically with the N67 glycan.sup.49. The DENV-2 sE/EDE1 C8 complex displays the highest ordered N67 glycan structure, with the distant mannose residues contacting the framework region 3 of the heavy chain (FRW H3, FIG. 15B and FIG. 18C). Except for EDE1 C10 (which is very close to its germ line, FIG. 7), a number of the FRW H3 residues have undergone changes (FIG. 15B), suggesting affinity maturation to recognize the sugars. It is possible that if more glycan residues were visible in the structures, they would be seen interacting with the same FRW H3 residues of the other anti-EDE antibodies as well.

[0798] Although the N150 loop and N153 glycan are disordered in the EDE1 complexes, the limited space between the antibody and the remainder of domain I (FIG. 19A, left panel) suggests that this glycopeptide segment does make contacts with the antibody (as indicated by the question mark over the 150 loop in FIG. 15A), but adopting variable local conformations in each complex such that it averages out to no resolved electron density in the crystal. If the 150 loop remained in place, the CDR H3 loop of the EDE1 anti-EDE antibodies would collide with the N153 glycan, e.g. with the first GlcNAc residue in DENV-2 sE/EDE1 C10 (sugar 1 in FIG. 19C).

[0799] The electron density is clear for the core 6 sugar residues of the N153 glycan of sE in the crystals of the complexes with the EDE2 anti-EDE antibodies (including in omit maps, as shown in FIGS. 20A-C). The CDR H3 of EDE2 anti-EDE antibodies makes a 2-turn -helix (termed H3 helix, FIG. 11D), with one of the carbonyl groups at its C-terminal end capped by a hydrogen bond donated by the N2 atom of the first N153 GlcNAc residue (FIG. 21). The H3 helix projects laterally the aromatic side chains of Y99 (F99 in EDE2 A11) and Y100 to pack against the sugar residues 1, 3 and 4 of the N153 glycan. The most distant residues of the glycan, mannoses 4, 5 and 6, are in contact with the light chain, via residues from CDR L2, including several hydrogen bonds (FIGS. 19A-E and FIG. 21).

[0800] The different type of interactions that EDE1 and EDE2 anti-EDE antibodies make with the 150 loop and N153 glycan is reflected in the contrasting effects of the absence of glycan in their neutralization potency. For instance, a DENV-4 isolate having isoleucine at position 155 (i.e., a natural glycosylation mutant, lacking the 153-NDT-155 glycosylation motif), is more sensitive to neutralization by EDE1 anti-EDE antibodies, as there no collision of CDR H3 with the glycan chain. In contrast, this variant is more resistant to neutralization by the EDE2 anti-EDE antibodies (FIG. 19E), highlighting the importance of the observed specific recognition of the N153 glycan.

Example 11the Main Chain Conformation of the Fusion Loop as Binding Determinant

[0801] In the fusion loop, residues 101-WGNG-104 make a distorted -helical turn that projects the W101 side chain towards domain III across the dimer interface. In the complexes with EDE2 anti-EDE antibodies the helical turn of the fusion loop is under the H3 helix, such that the carbonyl groups at the C-terminal sides of the two helices face each other. Furthermore, S100C of the CDR H3 caps the helical turn by making main chain and side chain hydrogen bonds to the carbonyl group of G102 in the fusion loop. In the complexes with EDE1 anti-EDE antibodies, the fusion loop lies right underneath the VH/VL interface, with the side chains of several aromatic residues of both heavy and light chains packing against it. In particular, the VL main chain runs very close by, donating a hydrogen bond to the main chain carbonyl group of G104. In EDE1 C8, the main chain amide proton donor belongs to N93 from CDR L3, while in EDE1 C10 it belongs to N31 from CDR Li. Residue D50 in the CDR L2 of both C10 and C8 makes a salt bridge with K310 (FIG. 19A and FIG. 18D), which is part of an extensive network of polar interactions in this area (listed in FIG. 21).

[0802] The conformation of the glycine rich fusion loop in the E dimer is such that it essentially exposes the main chain, while the side chains are mostly buried. Together with the main chain of the ij loop, main chain atoms make a large surface patch that is augmented by one edge of the b strand, resulting in an invariant exposed surface recognized by the anti-EDE antibodies. The invariant side chains in this region, together with the exposed main chain atoms at the E dimer surface (FIG. 22A, lower panel), result in a core region of the epitope that is serotype invariant, with non-conserved residues essentially at the periphery. The reason of this conservation is likely to be related to the interaction with prM during particle maturation.sup.46. The least conserved region is the surface of domain III within the EDE1 epitope.

Example 12Structure of DENV-4 sE in Complex with EDE1 C10

[0803] To understand in detail how the anti-EDE antibodies can efficiently recognize multiple viral serotypes, we turned to DENV-4, since it differs most from the other dengue serotypes in amino acid sequence (FIG. 15A), and is also potently neutralized by the EDE anti-EDE antibodies (FIG. 19E). The 2.7 resolution crystal structure of DENV-4 sE in complex with the C10 scFv confirmed the general pattern observed in the complex with DENV-2 sE (FIG. 22B and FIGS. 23A and B), with the 150 loop disordered. As expected, the anti-EDE antibody displays the same interactions with the main chain and with the conserved side chains of the epitope. In the more variable lateral region of the EDE1 epitope, on domain III (FIG. 22A), the contact site includes the side chain of residue 309, which is aspartic acid in DENV-4 but valine in DENV-2 (FIG. 15A). In the latter complex there is Van der Waals packing between the side chains V309 and T52 of CDR L2, while in the former there is a polar interaction, with D309 accepting a hydrogen bond from the T52 side chain (FIG. 23 and FIG. 21). The other contacts with domain III are also maintained, in particular the one at position 362, which involves a hydrogen bond to the main chain carbonyl (FIG. 21).

[0804] EDE1 C10 clearly induces disorder of the 150 loop in DENV-2 sE, but in the case of DENV-4 sE this loop appears to display an intrinsic higher mobility, as suggested by its crystal structure in complex with the Fab fragment of an unrelated chimpanzee antibody termed 5H2 (ref..sup.17). Indeed, although the 5H2 epitope is also in domain I, it is at the side of the sE dimer and does not overlap with the anti-EDE antibody epitopes described here, yet the 150 loop was disordered in that structure. In addition, the structure of the DENV-4 sE/EDE1 C10 complex highlighted a non-negligible degree of asymmetry in the contacts of the anti-EDE antibodies with the two epitopes of the dimer (FIGS. 23A and B, FIG. 13 and FIG. 21). This asymmetry was also detectable in the complexes with DENV-2 sE, as displayed in FIGS. 24A-E. It is likely that stochastically, the binding of the first antibody fragment induces an asymmetric conformational adjustment of the sE dimer, which affects the second site such that when the second one binds it accommodates to the available conformation of the second epitope. Taken together, these observations strongly suggest that the binding determinants of the EDE1 anti-EDE antibodies lie at the conserved core of the epitope, in the region shared with the EDE2 anti-EDE antibodies, and that the contacts at either edge adapt to the particular side chains present in each serotype, which do not compromise binding. This observed plasticity of the E dimer is in line with reports of conformational breathing of the E dimers on virions, exposing normally hidden epitopes for interaction with antibodies.

Example 13Putative Additional EDE1 C10/E Dimer Interactions on Mature Virions

[0805] A close examination of the structure shows that the tip of the CDR H3 of EDE1 C10 reaches the bottom of the sE dimer (circled in FIG. 18D, left panel; see also FIGS. 12D-E), a region which, in the context of the intact virion, is buttressed by protein M underneath (FIGS. 25A-C). The 3.5 resolution cryo-EM structure of the intact mature particle of DENV-2 (ref..sup.9) shows that the interaction with M results in conserved residue F279, at the base of the kl hairpin of E (FIG. 15A), to be exposed at the dimer interface instead of being buried in the hydrophobic core of domain II (FIG. 25C). Superposition with DENV-2 sE/C10 structure shows that, when in the context of the virion, the exposed F279 side chain could interact with Y100 in the H3 loop. Y100 is seen making different interactions with DENV-2 sE compared to DENV-4 (FIG. 25C; compare also panels b and e in ED FIGS. 26A-E), suggesting that it does not find its right partner. Thus the EDE1 C10 binding site on the E dimer in mature virions appears not to be completely recapitulated by the recombinant sE dimer. This observation likely explains the apparent discrepancy between the weak binding of EDE1 C10 to the sE dimer (FIGS. 8A and B) and its potent binding and neutralization of viruses from the four dengue serotypes (NT 50 in the low nM range, see accompanying manuscript). Importantly, we note that the conformation of F279 on the mature virion is similar to that observed in sE bound to a hydrophobic ligand.sup.4, suggesting that it is possible to induce the right conformation of this region of the recombinant sE dimer as immunogen.

Example 14

[0806] We have provided snapshots of anti-EDE antibodies interacting with a major new epitope targeted by human monoclonal antibodies elicited in dengue infected patients. These antibodies appear to have converged toward the same specificity via totally different evolutionary pathways: acquiring a heavy chain with a very long CDR H3 that makes most of the interactions, as in the EDE2 examples, or a fine-tuned combination of light and heavy chains, with the light chain making main-chain contacts to the fusion loop and to domain III for the EDE1 anti-EDE antibodies analyzed here. EDE1 and EDE2 anti-EDE antibodies comprise nearly one third of the antibodies isolated from dengue patients in the accompanying study, and constitute the vast majority of those that recognize conformation-specific quaternary epitopes at the virion surface. Their common signature from the alanine scanning experiments (accompanying manuscript) strongly indicates that they all target the same quaternary epitopes described here.

[0807] Importantly, the binding determinants of the EDE anti-EDE antibodies are totally circumscribed to the E dimer, and do not depend on a higher order arrangement of dimers at the virion surface, as recently suggested for the quaternary epitopes on the DENV particle.sup.50 based on studies on a different flavivirus, the West Nile virus.sup.51. Recent cryo-EM analyses of DENV-2 particles suggest that the herringbone pattern may be disrupted at physiological temperatures in humans, with the dimers reorienting with respect to each other, loosing the symmetric arrangement and/or presenting a different surface pattern.sup.52, 53. The epitopes described here will therefore be accessible in the E dimers independent of swelling or not of the particles and may be the favored target for next generation vaccines. As a corollary, our results indicate that it is feasible to design potent immunogens by stabilizing the dimer contacts in such a way that only E dimers are presented to the immune system, as proposed recently for the respiratory syncytial virus (RSV).sup.54, thus avoiding eliciting antibodies against poorly immunogenic regions that are normally not accessible at the surface of an infectious virion.

[0808] The principal binding determinant of the EDE anti-EDE antibodies appears to be the conformation of the main chain of the fusion loop and its immediate neighbors in the context of an intact E dimer. This is in stark contrast with the other major class of antibodies isolated from humans in the accompanying manuscript, which recognize the fusion loop sequence in a context independent of the quaternary organization. The latter antibodies are cross reactive but poorly neutralizing and have a strong infection enhancing potential.sup.56. A notable feature of the epitopes described here is the number of exposed main chain atoms, which accounts for approximately 30% of the total surface area buried in the complex in the case of EDE1 and 20% for the EDE2 anti-EDE antibodies (this lower EDE2 percentage is largely compensated with 40% of invariant glycan composition) (FIG. 13), whereas in general main chain atoms contribute between 5% to 15% for most immune complexes that we have analyzed. We note that certain very potent neutralizing antibodies also recognize a high percentage (around 30%) of main chain atoms in the antigen (FIG. 27), such as the D25 antibody against the respiratory syncytial virus (RSV), which binds to the antigenic site 0 present exclusively in the pre-fusion form of the RSV fusion protein, stabilizing it in that conformation.sup.57 as do the EDE anti-EDE antibodies. A similar pattern is found with BNA CH65, which neutralizes a broad range of H1 influenza virus isolates by binding to the receptor binding pocket of hemagglutinin (HA).sup.58 or CR8020, a potent group 2 reactive anti influenza human BNA with neutralization activity against H3, H7 and H10 isolates, by binding to the base of the stem of H1.sup.59. CR8020 also recognizes the main chain conformation of the fusion peptide within the context of the quaternary structure, similar to the EDE anti-EDE antibodies described here, in the pre-fusion trimer conformation. Finally, two of the very broad anti HIV-1 anti-EDE antibodies, B12 (ref..sup.60) and VRC01 (ref..sup.61), which recognize the CD4 binding site in the envelope (ENV) protein, display 36% and 33% of main chain atoms in the epitope (FIG. 27), suggesting that recognition of the main chain conformation is an important aspect shared by many (but not all) anti-EDE antibodies. These anti-HIV-1 broadly neutralising antibodies, which also require the correct quaternary structure of the ENV trimer for efficient binding.sup.29, undergo a long affinity maturation process, displaying more than 20% divergence from the germ line, whereas the EDE anti-EDE antibodies against dengue are at most 9% divergent from the germ line (FIG. 27), indicating that they are relatively easy to develop within individuals if an appropriate immunogen is used for vaccination.

[0809] In conclusion, we described a highly conserved binding site for potent highly cross-reactive antibodies against dengue viruses. The poor efficacy of a recent live attenuated polyvalent dengue vaccine has created a pressing need to better understand protective responses in humans and to design a next generation of efficacious vaccines. Serotype specific immunity has often been the goal of dengue vaccines mandating their tetravalent formulation. Our results suggest that a subunit vaccine comprising a stabilized E dimer should be evaluated, that a single optimized universal immunogen may be possible and that the elicitation of anti-EDE antibodies should be considered as a realizable goal for a successful vaccine.

Example 15Additional Methods

[0810] The recombinant sE proteins from DENV serotypes 1 through 4, as well as Fab and scFv BNA fragments, were produced in Drosophila melanogaster Schneider 2 using previously described protocols.sup.44,45,29. The binding of the BNA fragments to the sE proteins was monitored by SEC/MALS and by SPR (FIGS. 8A and B). Crystals of the sE/BNA complexes were obtained by isolating the complex from a mixture by SEC, or by mixing the two in a 1:2 sE:antibody stoichiometric ratio in the case of EDE1 C10. Crystallization trials were made using a robotized facility. Diffraction data were collected at the synchrotron sources SOLEIL and ESRF, and the structures were determined by molecular replacement using the search models listed in FIG. 9, which also provides the relevant crystallographic statistics. The neutralization tests on the DENV-4 variants were carried out using the same procedures outlined in the accompanying paper.

Recombinant sE Protein Production

[0811] Recombinant DENV-1 FGA/89 sE (1-395), DENV-2 FGA02 sE (1-395) and DENV-3 PAH881 sE (1-393) were produced in Drosophila S2 cells essentially as described earlier for DENV-4 sE (Den4_Burma/63632/1976).sup.29, with some modifications. Briefly, sE expression was driven by the metallothionein promoter and was induced by 5 M of CdCl.sub.2 in Insect-XPRESS medium (Lonza). The constructs had a Drosophila BiP signal sequence fused at the N-terminal end of a prM-sE construct for efficient translocation into the ER of the transfected S2 cells. prM was present N-terminal to sE, as in the DENV polyprotein precursor, with the N-termini of prM and sE generated by signalase cleavage in the ER, where prM (which remains membrane-anchored) plays a chaperone role by masking the fusion loop of sE. The prM/sE complex is transported across the acidic compartments, where prM is cleaved by furin into pr (N-terminal half, bound to sE) and M (membrane-anchored C-terminal half). Upon reaching the external milieu, sE and pr dissociate, and the sE component is purified by affinity chromatography from the cells' supernatant fluid. While the DENV-3 and -4 sE constructs had C-terminal fusion with a twin-strep-tag (IBA, www.iba-lifesciences.com/twin-strep-tag.html), DENV-1 and 2 sE had a 6His C-terminal tag. Clarified cell supernatants were concentrated 20-fold using Vivaflow tangential filtration cassettes (Sartorius, cut-off 10 kDa) and adjusted to 0.5M NaCl before purification in an AKTA FPLC system with either StrepTactin affinity purification or HisTrap-HP chromatography after buffer exchange to remove divalent ions, depending on the construct. The His-tagged proteins (DENV-1 and -2 sE) were desalted after elution of the HisTrap column and further purified by ion exchange chromatography on MonoQ. A final purification SEC step using a Superdex 200 10/300 GL column equilibrated in 50 mM Tris pH8, 500 mM NaCl was done with all constructs.

Production of Fabs and ScFvs

[0812] The BNA fragments were cloned into plasmids for expression as Fab.sup.62 and scFv.sup.63 in Drosophila S2 cells. The constructs contain a twin strep tag fused at the C-terminus (only of the heavy chain in the case of the Fab) for affinity purification. The purification protocol included the same steps described above for the strep tagged sE proteins, and the same buffers were used.

Immune Complex Formation and Isolation

[0813] The purified DENV sE proteins were mixed with Fabs or ScFvs (in 2-fold molar excess) in standard buffer (500 mM NaCl, Tris 50 mM pH 8.0 buffer). The volume was brought to 0.2 ml by centrifugation in a Vivaspin 10 kDa cutoff, after 30 min incubation at 4 C., the complex was separated from excess Fab or scFv by SEC, except when a clear peak for the complex was not obtained (as with BNA C10, see FIGS. 8A and B). In this case, a molar ratio 1:2 antigen:antibody mixture (i.e., with an excess of antibody) was directly used for crystallization. In all cases, the buffer was exchanged to 150 mM NaCl, 15 mM Tris, pH 8 for crystallization trials. The protein concentrations used for crystallization, determined by measuring the optical density at 280 nm and using an extinction coefficient estimated from the amino acid sequences, are listed in FIG. 9.

MALS Analysis

[0814] 150 g of purified DENV-1, -2, -3 and -4 sE were mixed with 300 g of A11, B7, C8 and C10 Fab fragments and adjusted to a total volume of 100 l. The individual proteins (DENV sE or Fabs) were also run separately as controls at the same concentration. Samples were incubated for 15 min at RT, and analyzed by MALS as they eluted from an SDX200 10/300 GL gel filtration column run at a flow rate 0.4 ml/min. The elution was followed by refractometry and MALS detection with a DAWN Heleos-Optilab T-rEX setup (Wyatt Technology).

Surface Plasmon Resonance

[0815] Real-time SPR measurements of the binding of sE dimers to captured Fab fragments of the anti-EDE antibodies were performed using a ProteOn XPR36 instrument (BioRad).

[0816] The Fab fragment of the DENV-4 specific neutralizing antibody 5H2 was used as control. Biotinylated anti-human CH1 specific antibody (Life Technologies) was immobilized on a Neutravidin ProteOn NLC sensor chip, and used to capture similar densities (400-500 RU) of the different Fab fragments. This anti-CH1 antibody recognizes all IgG subclasses (1, 2, 3 and 4) independently of the light chain subclass (Kappa/Lambda). We found that this anti-CHI antibody also cross reacts with 5H2, a chimpanzee antibody, although with a lower affinity. The Fab fragment of an anti-HCV E2 antibody was used as a negative control. The chip was rotated 90 following Fab capture, and sE of the four DENV serotypes was injected at a concentration of 2 M. Blank injections with running buffer (50 mM Tris pH8, 500 mM NaCl, 0.01% Tween20) were used for double referencing. SPR signals were normalized to the amount of Fab captured. A control injection of the ectodomain of Rubella virus E1 glycoprotein at a similar concentration over all the Fabs showed no apparent binding (data not shown).

Neutralization Assays with DENV-4 Glycosylation Variants

[0817] The neutralization potential of the anti-EDE antibodies was determined using the Focus Reduction Neutralization Test (FRNT).sup.22, where the reduction in the number of infected foci is compared to control (no antibody). DENV-4 strains H241 (with Ile at position 155), 1-0093 and 1-0554 (both with Thr at position 155thus restoring glycosylation at Asn153) were grown in C6/36 cells. Viral titres were determined by a focus-forming assay on Vero cells.sup.64. Briefly, serially-diluted anti-EDE antibodies were mixed with virus and incubated for 1 hr at 37 C. The mixtures were then transferred to Vero cells and incubated for 3 days. The focus-forming assay was then performed using the murine monoclonal 4G2 antibody (which cross-reacts with E protein from all flaviviruses) followed by rabbit anti-mouse IgG, conjugated with horse radish peroxidase. The reaction was visualized by the addition of diaminobenzidine substrate. The percentage foci reduction was calculated for each antibody dilution. 50% FRNT were determined from graphs of percentage reduction versus concentration of Abs using probit (www.statisticalassociates.com/probitregression.htm) with the statistical package SPSS.

Crystallization and 3D Structure Determinations

[0818] Crystallization trials were carried out in sitting drops of 400 nl. Drops were formed by mixing equal volumes of the protein and reservoir solution in the format of 96 Greiner plates, using a Mosquito robot, and monitored by a Rock-Imager. Crystals were optimized with a robotized Matrix Maker and Mosquito setups on 400 nl sitting drops, or manually in 24 well plates using 2-3 l hanging drops (FIG. 9). The crystallization and cryo-cooling conditions for diffraction data collection are listed in FIG. 9.

[0819] X-ray diffraction data were collected at beam lines PROXIMA-1 and PROXIMA-2 at the SOLEIL synchrotron (St Aubin, France), and ID23-2 and ID29 at the European Synchrotron Radiation Facility (Grenoble, France) (FIG. 9). Diffraction data were processed using the XDS package.sup.65 and scaled with SCALA or AIMLESS.sup.66 in conjunction with other programs of the CCP4 suite.sup.67. The structures were determined by molecular replacement with PHASER.sup.68 and/or AMoRe.sup.69 using the search models listed in FIG. 9.

[0820] Subsequently, careful model building with COOT.sup.70, alternating with cycles of crystallographic refinement with program BUSTER/TNT.sup.71, led to a final model. Refinement was constrained to respect non crystallographic symmetry, and also used target restraints (with high resolution structures of parts of the complexes) and TLS refinement.sup.72 depending on the resolution of the crystal (see FIG. 9). Final omit maps were calculated using Phenix.Refine.sup.73.

Analysis of the Atomic Models and Illustrations

[0821] Each complex was analyzed with the CCP4 suite of programs.sup.67. For intermolecular interactions, the maximal cutoff distance used for the interactions was 4.75 . Then the contacts of each residue of the Fab/ScFv or of DENV sE proteins were counted and plotted as a proportional bar above the corresponding residue.

[0822] The Ab sequences were analyzed by Abysis (www.bioinf.org.uk/software) and IMGT (www.imgt.org).sup.31 websites for mapping CDR/FWR regions according to Kabat.sup.30 and IMGT.sup.31 conventions, respectively. The analysis of the putative germline and somatic maturation events was done with the IMGT website (www.imgt.org).

[0823] Multiple sequence alignments and phylogenetic trees were calculated using ClustalW (ClustalW and ClustalX version 2 (ref..sup.74) on the EBI server.sup.75. The tree was calculated using amino acid sequences of sE proteins used in this study: DENV-1 FGA/89 (1-395), DENV-2 FGA02, DENV-3 PAH881 (1-393) and DENV-4 (DEN_Burma/63632/1976). For mapping DENV-2 genotypes, the database from .sup.76 was used to extract amino acid sequences of sE ectodomains and extended to include DENV2 FGA02 sE and DENV-2 10AN sE. For simplicity of representation sub-roots were collapsed to the level of individual genotype for DENV-2. The tree was then rooted with the DENV-4 sE sequence and drawn to scale using the MEGA5 software package.sup.77.

[0824] For FIG. 22A and FIG. 14 and for analysis purposes, a model of DENV-2 sE dimer without gaps in the sequence was used. The model was built using the complete protomer A of the DENV-2 sEB7 complex.

[0825] Figures were prepared using Program ESPript.sup.78 and the PyMOL Molecular Graphics System, Version 1.5.0.4 Schrodinger, LLC. (pymol.sourceforge.net) with APBS79 and PDB2PQR tools.sup.80.

[0826] Finally, current vaccine strategies employ a tetravalent formulation with the aim of raising a balanced type specific response against all four serotypes. The description here of such potent and crossreactive antibodies points the way for subunit vaccines containing the desired epitope and possibly heterologous prime boost strategies to recapitulate responses seen in natural sequential infections.

Example 16: sequence Information

TABLE-US-00002 SEQIDNO's SEQIDNO:1 FullseqofantibodyC8Heavychain EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSMHWVRQAPGKGLEYVSA ITGEGDSAFYADSVKGRFTISRDNSKNTLYFEMNSLRPEDTAVYYCVGGY SNFYYYYTMDVWGQGTTVTV SEQIDNO:2 FullseqofantibodyC10Heavychain EVQLVESGAEVKKPGASVKVSCKASGYTFTSYAMHWVRQAPGQRLEWMGW INAGNGNTKYSQKFQDRVTITRDTSASTAYMELSSLRSEDTAIYYCARDK VDDYGDYWFPTLWYFDYWGQGTLVTV SEQIDNO:3 FullseqofantibodyA11Heavy EVQLVESGGGLVRPGGSLRLSCAASGFSYSNHWMHWVRQAPGKGLVWVSR INSDGSTRNYADFVKGRFTISRDNAENTLYLEMNSLTADDTAVYYCVRDG VRFYYDSTGYYPDSFFKYGMDVWGQGTTVTV SEQIDNO:4 FullseqofantibodyB7Heavychain EVQLVESGGGLVQPGGSLKLSCAASGFTFSSHWMHWVRQAPGKGLVWVSR TNSDGSSTSYADSVKGRFMISRDNSKNTVYLHMNGLRAEDTAVYFCARDG VRYYYDSTGYYPDNFFQYGLDVWGQGTT SEQIDNO:5 C8CDRH1 TYSMH SEQIDNO:6 C8CDRH2 AITGEGDSAFYADSVKG SEQIDNO:7 C8CDRH3 GYSNFYYY SEQIDNO:8 C10CDRH1 SYAMH SEQIDNO:9 C10CDRH2 WINAGNGNTKYSQKFQD SEQIDNO:10 C10CDRH3 DKVDDYGDYWFPTLW SEQIDNO:11 A11CDRH1 NHWMH SEQIDNO:12 A11CDRH2 RINSDGSTRNYADFVKG SEQIDNO:13 A11CDRH3 DGVRFYYDSTGYYPDSFFKY SEQIDNO:14 B7CDRH1 SHWMH SEQIDNO:15 B7CDRH2 RTNSDGSSTSYADSVKG SEQIDNO:16 B7CDRH3 DGVRYYYDSTGYYPDNFFQY SEQIDNO:17 C8-CDRLl RASQSISTFLA SEQIDNO:18 C8CDRL2 DASTRAT SEQIDNO:19 C8CDRL3 QQRYNWPPYT SEQIDNO:20 C10CDRLl TGTSSDVGGFNYVS SEQIDNO:21 C10CDRL2 DVTSRPS SEQIDNO:22 SSHTSRGTWVF SEQIDNO:23 A11CDRL1 TGTSSNADTYNLVS SEQIDNO:24 A11CDRL2 EGTKRPS SEQIDNO:25 A11CDRL3 CSYATSRTLVF SEQIDNO:26 B7CDRL1 TGISSDVETYNLVS SEQIDNO:27 B7CDRL2 EASKRPS SEQIDNO:28 B7CDRL3 CSYAGGKSLV SEQIDNO:29 FulllengthenvelopeproteinsequenceDENV1 >DENV1strainHawaii MRCVGIGNRDFVEGLSGGTWVDVVLEHGSCVTTMAKDKPTLDIELLKTEV TNPAVLRKLCIEAKISNTTTDSRCPTQGEATLVEEQDANFVCRRTFVDRG WGNGCGLFGKGSLITCAKFKCVTKLEGKIVQYENLKYSVIVTVHTGDQHQ VGNETTEHGTIATITPQAPTSEIQLTDYGALTLDCSPRTGLDFNEMVLLT MKEKSWLVHKQWFLDLPLPWTSGASTPQETWNREDLLVTFKTAHAKKQEV VVLGSQEGAMHTALTGATEIQTSGTTKIFAGHLKCRLKMDKLTLKGMSYV MCTGSFKLEKEVAETQHGTVLVQVKYEGTDAPCKIPFSTQDEKGVTQNGR LITANPIVTDKEKPVNIEAEPPFGESYIVVGAGEKALKLSWFKKGSSIGK MLEATARGARRMAILGDTAWDFGSIGGVFTSVGKLVHQIFGTAYGVLFSG VSWTMKIGIGILLTWLGLNSRSTSLSMTCIAVGMVTLYLGVMVQA SEQIDNO:30 fulllengthenvelopenucleotidesequenceDENV1 SEQIDNO:31 fulllengthenvelopeproteinsequenceDENV2 >DENV2strain16681 MRCIGMSNRDFVEGVSGGSWVDIVLEHGSCVTTMAKNKPTLDFELIKTEA KQPATLRKYCIEAKLTNTTTESRCPTQGEPSLNEEQDKRFVCKHSMVDRG WGNGCGLFGKGGIVTCAMFRCKKNMEGKVVQPENLEYTIVITPHSGEEHA VGNDTGKHGKEIKITPQSSITEAELTGYGTVTMECSPRTGLDFNEMVLLQ MENKAWLVHRQWFLDLPLPWLPGADTQGSNWIQKETLVTFKNPHAKKQDV VVLGSQEGAMHTALTGATEIQMSSGNLLFTGHLKCRLRMDKLQLKGMSYS MCTGKFKVVKEIAETQHGTIVIRVQYEGDGSPCKIPFEIMDLEKRHVLGR LITVNPIVTEKDSPVNIEAEPPFGDSYIIIGVEPGQLKLNWFKKGSSIGQ MFETTMRGAKRMAILGDTAWDFGSLGGVFTSIGKALHQVFGAIYGAAFSG VSWTMKILIGVIITWIGMNSRSTSLSVTLVLVGIVTLYLGVMVQA SEQIDNO:32 fulllengthenvelopenucleotidesequenceDENV2 SEQIDNO:33 fulllengthenvelopeproteinsequenceDENV3 >DENV3stainH87 MRCVGVGNRDFVEGLSGATWVDVVLEHGGCVTTMAKNKPTLDIELQKTEA TQLATLRKLCIEGKITNITTDSRCPTQGEAILPEEQDQNYVCKHTYVDRG WGNGCGLFGKGSLVTCAKFQCLESIEGKVVQHENLKYTVIITVHTGDQHQ VGNETQGVTAEITSQASTAEAILPGYGTLGLECSPRTGLDFNEMILLTMK NKAWMVHRQWFFDLPLPWTSGATTETPTWNRRELLVTFKNAHAKKQEVVV LGSQEGAMHTALTGATEIQTSGGTSIFAGHLKCRLKMDKLELKGMSYAMC LNTFVLKKEVSETQHGTILIKVEYKGEDAPCKIPFSTEDGQGKAHNGRLI TANPVVTKKEEPVNIEAEPPFGESNIVIGIGDKALKINWYRKGSSIGKMF EATARGARRMAILGDTAWDFGSVGGVLNSLGKMVHQIFGSAYTALFSGVS WIMKIGIGVLLTWIGLNSKNTSMSFSCIAIGIITLYLGVVVQA SEQIDNO:34 fulllengthenvelopenucleotidesequenceDENV3 SEQIDNO:35 fulllengthenvelopeproteinsequenceDENV4 >DENV4strain241 MRCVGVGNRDFVEGVSGGAWVDLVLEHGGCVTTMAQGKPTLDFELIKTTA KEVALLRTYCIEASISNITTATRCPTQGEPYLKEEQDQQYICRRDVVDRG WGNGCGLFGKGGVVTCAKFSCSGKITGNLVQIENLEYTVVVTVHNGDTHA VGNDIPNHGVTATITPRSPSVEVKLPDYGELTLDCEPRSGIDFNEMILMK MKKKTWLVHKQWFLDLPLPWAAGADTSEVHWNYKERMVTFKVPHAKRQDV IVLGSQEGAMHSALTGATEVDSGDGNHMFAGHLKCKVRMEKLRIKGMSYT MCSGKFSIDKEMAETQHGTTVVKVKYEGAGAPCKVPIEIRDVNKEKVVGR IISSTPFAEYTNSVTNIELEPPFGDSYIVIGVGDSALTLHWFRKGSSIGK MLESTYRGVKRMAILGETAWDFGSVGGLFTSLGKAVHQVFGSVYTTMFGG VSWMVRILIGFLVLWIGTNSRNTSMAMTCIAVGGITLFLGFTVHA SEQIDNO:36-fulllengthenvelopenucleotide sequenceDENV4 SEQIDNO:37 C8lightchain-Seetablebelow SEQIDNO:38 FullseqofantibodyC10lightchain-Seetable below SEQIDNO:39 FullseqofantibodyA11lightchain-Seetable below SEQIDNO:40 B7lightchainSeetablebelow SEQIDNO:37--131antibodylightandheavychain sequencesfromthetablebelow

TABLE-US-00003 SEQ SEQ Sequence ID ID SequenceAA ID epitope NO: SequenceAA(Hchain) NO: (Lchain) 747(4) EDE 40 QVQLQESGPGLMKPSETLSLTCSVSGVSISTHYW 86 QTVVTQEPSLTVSPGG B3 1 SWIRQPPGKGLEWIGFIYNSGGTHYNPSLKSRVTI TVTLTCGSNTGPVTN SADTSKNQFALTLSSVTAADTAVYYCARGRRAY GHYPYWFQQKSGQAP DSSGYVKYYYFYGVDVWGQGTTVTVSS RTLIYDTTNRQSWTPV RFSGSLLGGKAALTLS GAQPEDEADYHCLLS YSDGLVFGGGTKLTVL 747 EDE 41 EVQLVESGSELKKPGASVKVSCRASGFTFTSYTF 87 CMTPAPSTLAVTPGEP A12 1 NVVVRQAPGQGLEWMGWIDTKSGRPTYAQGFTG ASISCRSTQSLLHSDG RFVLSLDTSVSTAYLQINSLKVEDTAMYYCARV YNYLDWYLQKPGQSP HTGGYPPELRYYYYGMDVWGQGTTVTVSS HLLIYLGSHRASGVPD RFSGSGSDTDFTLKIS RVEAEDVGVYYCMQ PLRTPPTFGQGTKLEIK 752 EDE 42 EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 88 EIVLTQSPATLSLSAG B10 1 HVVVRQAPGKGLEYVSAITTDGNSAFYADSVKGR DRATLSCRASQDISSF FTISRDNSKNTMYFHMNSLRPEDTAVYYCVGGY LAWYQQKPGQAPRLL SSFYYYYTMDVWGQGTTVTVSS MYDTSNRATGVPARF SGSRSGTDFTLTISTLE PEDVAVYYCQHRYN WPPYTFGQGTKVEIK 752 EDE 43 QVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 89 EIVLTQSPATLSLSPGE B11 1 HVVVRQAPGKGLEYVSAITTDGDSAFYADSVKGR RATLSCRASQSISSFLA FTISRDNSKNTMFFHMSNLRPEDTAVYYCVGGY WYQQKPGQAPRLLIY SSFYYYYTLDVWGQGTTVTVSS DASNRVTGVPARFSG SRSGTDFTLTISTLEPE DFAVYYCQHRYNWP PYTFGQGTKVEIK 752 EDE 44 EVQLVESEGGLVQPGGSLRLSCSASGFTFSTYSM 90 EIVLTQSPATLSLSPGE C9 1 HVVVRQAPGKGLEYVSAITTNGDSTFYADSVKGR RATLSCRASQSISTYL FTISRDNSKNTLYFQMSSLRAEDTGVYYCVGGY AWYQQKPGQAPRLLI SSFYYYYTMDVWGQGTTVTVSS YDASNRATGVPARFS GSRSGTDFTLTISTLEP EDFAVYYCQQRYNW PPYTFGQGTKVEIK 752(2) EDE 45 EVQLVQSGPEMRKPGASVKVSCKASGYTFTSHG 91 DIQMTQSPSSLSASVG A2 1 INWVRQVPGQGPEWMGWSSSYTDNTNYAQKFK DRVTITCRASQTISGSL GRVTMTTDPSTSTAYMELRSLRSDDTAIYFCARG SWYQHKPGKAPKLLI FYSGSYYPTAPFDIWGQGTLVTVSS YAASSLQSGVPSRFSG SGSGTDFTLTISSLQPE DFATFYCQQSYSTPYT FGQGTKVEIK 752(2) EDE 46 EVQLVQSGAEVKKPGASVKVSCKASGYTFTTYG 92 DIQMTQSPSSLSASIG A5 1 LSWVRQAPGQGLEWMGWCSSYNDNTNYAQKF DRVTITCRASESISSQL KGRVTMTTDTSTNTAYMELRSLRSDDTAVYYC HWYQQKPGKAPRLLI ARVFYSGSYYPNSPFDYWGQGTLVTVSS YAASSLQGGVPSRFSG SGSGTDFTLTISGLQPE DFATYCCQQSFTTPYT FGQGTKVEIK 752(2) EDE 47 QVQLQESGPGLVKPSQTLSLTCTVSGDSISSNNY 93 EIVMTQSPATLSASPG A7 1 QWNWIRQPAGKGLEWLGRIDTTGSTNYNPSLKS ERATLSCRASQDVSTF RISISIDTSKKQFSLRLNSVTAADTAVYYCARSLW VAWFQQNPGQAPRLL SGELWGGPLGYWGQGTLVTVSS IYDASTRAPGIPARFS GSRSGTEFTLTINSLQS EDFAVYYCQQYYNW PPWTFGQGTKVEIK 752(2) EDE 48 EVQLVESGAEVKNPGASVKVSCKASGYTFIGYYI 94 DIQMTQSPSSVSASVG A8 1 HWVRQAPGQGLEWMGWINPNSGATYSAQKFQ DRVTISCRASQDISAS GRVTLTGDASPSTVYMELSSLRSDDTAIYYCAGR LGWYQQKPGKAPKLL SYNWNDVFYYYYMDVWGQGTTVTVSS IYRASNLEGGVPSRFR GSGSGTDFTLTISSLQP EDFATYYCLQANSFPL TFGGGTKVEIK 752(2) EDE 49 EVQLVESGPGLVKPSETLSLTCTISGVSISDYYWT 95 DIQMTQSPSSLSASVG B10 1 WIRQPPGKGLEWIGNIYNTGSTNYNPSLKSRVAI DSVTVACRASQPIYRN WMDTSKNKFSLRLTSVTSADTAVYYCARVEGGP LNWYQQKPGKAPKLL KYYFGSGDFYNLWGRGSLVTVSS IYDASTLQSGVPARFS GSGSGTDFTLTISSLQ AEDFATYYCQQSYSS PRTFGQGTKVEIK 752(2) EDE 50 SQVQLVQSGAELKKPGASVKVSCKTSGYTFSYYI 96 DIQMTQSPSTLSASVG C2 1 HWVRQAPGQGLEWMAMINPTSGSTSYAQRFQG DRVTITCRASQSISTYL RVTMTRDTPTNTVYMEVRSLRSDDTAVYFCASR AWYQQKPGKAPKLLI GYNWNDVQYYYTMDVWGQGTTVTVSS YKASSLEIGVPSRFSG SGSGTEFTLTISSLQPD DFAIYYCQQYNNYSP PVTFGGGTKVEIK 752(2) EDE 51 SEVQLVQSGAELKKPGASVKVSCKASGYTFSYYI 97 DIQMTQSPSTLSASVG D4 1 HWVRQAPGQGLEWMAIINPTSGSTSYAQRFQGR DRVTITCRASQSISTYL VTMTRDTSTNTVYMELSSLISEDTAVYYCASRG AWYQQKPGKAPKLLI YNWNDVHYYYTMDVWGQGTTVTVSS YKASTLESGVPLRFSG SGSGTEFTLTISSLQPD DFAIYYCQQYNNYSP PVTFGGGTKVEIK 752(2) EDE 52 QVQLVESGAEVKKPGSSVKVSCKASGYTFTTYG 98 DIQMTQSPSSLSASVG B11 1 LSWVRQAPGQGLEWMGWCSSYEDNTNYAPRFK DAVSITCRASESVSRQ GRVTMTTDTSTNTAYMELRSLRFDDTAVYYCAR LNWYQQKPGKAPNLL VFYSGSYYPNSPFDSW IYAASSLQGGVPSRFS GSGSGTDFTLTISGLQ PEDFATYYCQQGYST PYSFGQGTKVEIK 752-2 EDE 53 QVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 99 EIVLTQSPATLSLSAG A2 1 HWVRQAPGKGLEYISAITTDGDSAFYADSVKGR ERATLSCRASQSISSY FTISRDNSKNTMYFHMNSLRPEDTAVYYCVGGY LAWYQQKPGQAPRLL SSFYYYYTMDVWGQGTTVTVSS IYDASNRATGVPARFS GSQSGTDFTLTISTLEP EDFAVYYCQLRYNWP PYTFGQGTKVEIK 752-2 EDE 54 EVQLVESGAEVKKPGASVKVSCKASGYTFTSYGI 100 DIQMTQSPSPLSASVG A4 1 NWVRQAPGQGLEWMGWISSDSGHTNYARKLK DRVTITCRASQSISSHL GRVTMTTDTSTTTAYMELRSLRSDDTAVYYCAR NWYQQKSGKVPKLLI GLYSVSYYPTSPFDYWGQGSTVTVSS YAASSLQSGVPSRFSG SGSGTDFTLTITSLQPE DFATYYCQQSDTTPY TFGQGTKVEIK 752-2 EDE 55 QVQLVESGAEVKKPGSSVKVSCRASGYTFTTYG 101 DIQMTQSPSSLSASVG A5 1 LSWVRQAPGQGLEWMGWCSSYNDNTNYAQKF DAVSITCRASESIARQ KGRVTMTTDTSTNTAYMELRSLRSDDTAVYYC LNWYQQKPGKAPNLL ARVFYSGSYYPNSPFDSWGQGTLVTVSS IYAASSLQGGVPSRFS GSGSGADFTLTISGLQ PEDFATYYCQQGYST PYTFGQGTKVEIK 752-2 EDE 56 EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 102 EIVLTQSPATLSLSAG A9 1 HWVRQAPGKGLEYVSAITTDGDSAFYADSVKGR ERATLSCRASQDISTF FTISRDNSKNTMYFHMNSVRPEDTAVYYCVGGY LAWYQQKPGQAPRLL SSFYYYYTMDVWGQGTTVTVSS IYDTSTRATGVPARFS GSRSGTDFTLTITTLEP EDFAVYYCQHRYNW PPYTFGQGTKVEIK 752-2 EDE 57 EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 103 EIVLTQSPATLSLSAG B2 1 HWVRQAPGKGLEYVSAITTDGDSAFYADSVKGR ERATLSCRASQSISSY FTISRDNSKNTMYFHMNSLRPEDTAVYYCVGGY LAWYQQKPGQAPRLL SSFYYYYTMDVWGQGTTVTVSS IYDASNRATGVPARFS GSRSGTDFTLTISTLEP EDFAVYYCQHRYNW PPYTFGQGTKVEIK 752-2 EDE 58 EVQLLESGGGLVQPGGSLRLSCSASGFTFSTYSM 104 EIVLTQSPATLSLSPGE B3 1 HWVRQAPGKGLEYVSAISTDGDSAFYADSVKGR RATLSCRASHSISTFL FTISRDNSKNTLYFHMSSLRAEDTAVYYCLGGYS AWYQQKPGQAPRLLI TFYYYYTMDVWGQGTTVTVSS YDTSTRATGVPARFS GSRSGTDFTLTINTLEP EDFAVYYCQQRYNW PPYTFGQGTKVEIK 752-2 EDE 59 QVQLVESGGGLVQPGGSLRLSCSASGFPFSTYSM 105 EIVLTQSPATLSLSPGE B4 1 HWVRQAPGKGLEYVSAITTNGDSTFYADSVKGR RATLSCRASQSISSFLA FTISRDNSKNTVYFQLSSLRAEDTAVYYCVGGYS WYQQKPGQAPRLLIY SFYFYYTMDVW DTSNRATGVPARFSGS RSGTDFTLTISTLEPED FAIYYCQHRYNWPPY TFGQGTKVEIK 752-2 EDE 60 EVQLVQSGAEVKKPGASVKVSCKASGYTYTNY 106 DIQMTQSPSSLSASVG B7 1 GLSWVRQAPGQGLEWMGWMSSYNDNTNYSQK DRVTITCRASQSISRSL FKGRVTMTTDPSTTTAYMELRSLRSDDTAVYYC NWYQQKPGKAPKLLI ARGLYSGSHYPTSPLDYWGQGTLVTVSS YAASTLQSGVPSRFSG SGSGTDFALTISSLQPE DFATYSCQQSDRTPY TFGQGTKVEIK 752-2 EDE 61 EVQLVESGGGLVQPGGSLRLSCSASGFTFTTYSL 107 EIVLTQSPATLSLSPGE B11 1 HWVRQTPGKGLEYVSAITTDGDSAFYADSVKGR RATLSCRASQSISTYL FTISRDNSKNTMYFHMSSLRPEDTAVYYCVGGY VWYQQKPGQAPRLLI SSFYYFYTVDVWGQGTTVTVSF YDASTRATGVPARFS GSRSGTDFTLTISTLEP EDFAVYYCQHRYNW PPYTFGRGTKVEIK 752-2 EDE 62 SQVQLVESGAELKKPGASVKVSCKASGYTFSYY 108 DIQMTQSPSTLSASVG C4 1 MHWVRQAPGQGLEWMAIINPTSGSTTYAQRFQ DRVTITCRASQSISTYL GRVTMTRDTSTSTVYMELSSLRSEDTAVYYCAS AWYQQKVGKAPKLLI RGYNWNDVHYYYTMDVWGQGTTVTVSS YKASTLEGGVPSRFSG SGSGTEFTLTISSLQPE DFAIYYCQQYNNYSP PVTFGGGTKVEIK 752-2 EDE 1 EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSM 37 EIVLTQSPATLSLSPGE C8 1 HWVRQAPGKGLEYVSAITGEGDSAFYADSVKGR RATLSCRASQSISTFL FTISRDNSKNTLYFEMNSLRPEDTAVYYCVGGYS AWYQHKPGQAPRLLI NFYYYYTMDVWGQGTTVTVSS YDASTRATGVPARFS GSRSGTDFTLTISTLEP EDFAVYYCQQRYNW PPYTFGQGTKVEIK 753(3) EDE 2 EVQLVESGAEVKKPGASVKVSCKASGYTFTSYA 38 QSALTQPASVSGSPGQ C10 1 MHWVRQAPGQRLEWMGWINAGNGNTKYSQKF SITISCTGTSSDVGGFN QDRVTITRDTSASTAYMELSSLRSEDTAIYYCAR YVSWFQQHPGKAPKL DKVDDYGDYWFPTLWYFDYWGQGTLVTVSS MLYDVTSRPSGVSSRF SGSKSGNTASLTISGL QAEDEADYYCSSHTS RGTWVFGGGTKLTVL 753(3) EDE 63 EVQLVESGPEVKKPGASVKVSCKTSGYTFINYYI 109 DIVMTQSPLSLSVTPG B10 1 HWVRQAPGQGLEWLGLINPRGGNTNYAEKFED EPASISCRSSQSLVYSD RVTMTRDTSTSTVNMELSSLTSEDTAVYYCARP GNKYLDWYVQKPGQ LAHTYDFWSGYHRATGYGMDVWGQGTTVTVS SPQLLIYLTSTRASGV S PDRFSGSASGTDFTLK ISRVEAEDVGLYYCM QALQTPFTFGPGTKV DIK 758 EDE 64 EVQLVESGGGLVQPGGSLRLSCAAFGFTFVNYA 110 EIVMTQSPATLSVSPG P6A1 1 MNWVRQAPGKGPEWVAVIYAAGDGANYGDSV ERATLTCRASQTISTF KGRFTISRDNSRNTLYLQMNSLRAEDTAIYYCAK LAWYQQKPGQPPRLL PAHYDDSGYPYMAYFDSWGQGTLVTVSS IYDTSTRATGIPGRFSG SRSGTEFTLTISSLQSE DVAVYYCQHYYNWP PWTFGQGTKVEIK 758 EDE 65 QVQLVQSGAEVKKPGSSVKVSCKASGGFFSSYAI 111 QSALTQPPSASGSPGQ P6A3 1 TWVRQAPGQGLEWMGGIIPDYDSAKYAQKFQG SVTISCTGSSSDIGGNE RVTITADESTSTAYLELRSLRSEDTAVYYCARRH YVSWYQLQPGKAPKL CSSTSCSDPWTFFPSWGQGTLVTSPQ MIYEVTKRPSGVPNRF SGSKSGNTASLTVSGL QSEDEGDYYCSSYAD NSVLFGGGTTLTVL 758 EDE 66 EVQLVESGAEMKKPGSSVKVSCKASGATFTSFA 112 QSVLTQPPSASGSPGQ P6A1 1 MYWVRQAPGQGLEWMGRIIPMFASAEYAQKFQ SVTISCTGTSSDVGAY 2 GRLTMTADESTTTAYMELSSLRSDDTAVYYCAG YYVSWYQQHPGKAP RYCSSTSCSDPWTYFPHWGQGTLVTVSS KLIIYEVNKRPSGVPA RFSGSKSGNTASLTVS GLQGEDEADYYCTSY AGSNTVIFGGGTKLTVL 758 EDE 67 EVQLVQSGATVRKPGASVTISCKTSGYTFTDYAL 113 EIVLTQSPVTLSLSPGE P6B4 1 HWVRQAPGQRLEWMGWLIPGSGYTKFAENFQG RATLSCRASQTVDST RVTITRATSAHTAYMELSNLRSEDTAVYYCARW YLAWYQQKPGRAPRL GGDCNAGSCYGPYQYRGLDAWGQGTTVTVSS LIYGASNRAIGVPSRF TGSGSGTDFTLTISRLE PEDFALYYCQQSDGS LFTFGPGTKVDIK 758 EDE 68 EVQLVQSGAEVKKPGASVKVSCKASGYSFIGYY 114 DIQMTQSPASVSASVG P6B5 1 LHWVRQAPGQGLEWMGRINPNSGGIDYGQTFQ DRVTISCRASQGIASW GRVTMTRDMSSSTVYLELTRLRSDDTARYYCAG LAWYQQKPGKAPRLL RSDNWNDVYYNYALDVWGQGTTVTVSS IYGASSLQSGVPSRFR GSGSGTDFTLTISSLQP EDFATYYCQQANSFP FTFGPGTKVDIK 758 EDE 69 EVQLLESGGGVVQPGRSLKLSCAASGFTFSGYA 115 QSALTQPASVSGSPGQ P6B1 1 MHWVRQAPGKGLEWLAVISYDATTTYYTPSVK SITISCTGTSSDVGRYN 1 GRFTISRDNSKNTLYLQINSLRAEDAAVYYCAKE VVSWYQQHPGKAPK ISYCGGDCQNFFFYYNMDVWGQGTTVTVSS LIIYGSTKRPSGVSYRF SASKSGNTASLTISGL QAEDEAEYHCCSYAS GSVWVFGGGTKLTVL 758 EDE 70 QVQLVQSGAEVKKPGASVKVSCKASGYTFTAY 116 QSALTQPPSASGSPGQ P6C4 1 YIHWVRQAPGQGLEWMGSINPNNGGTNYAQGF SVTISCTGTSSDVGGY QGRVTMTRDTSIRTVYMELSKLRSDDTALYYCA NYVSWYQHHPGKAP RDLGAMGYYLCSAGNCPFDYWGQGTLVTVSS KLIIYEVSKRPSGVPH RFSGSKSGNTASLTVS GLQAEDEAEYYCSSY AGSNTFTFGGGTKLT VL 747 EDE 71 QVQLVESGGALVKPGGSLRLSCAASGFTFRSHW 117 QSALTQTASVSGSPGQ B8 2 MHWVRQAPGKGLVWVSRINSDGSSTNYADFVK SITISCTGTSSDAEIYN GRFTTSRDNAENTLYLEMNSLTADDTAVYYCVR LVSWYQQHPGKAPKL DGVRYYYDSSGYYPDSFFKYGMDVWGQGTTVT IIYEGSKRPSGVSNRFS VSS ASKSAGAASLRISGLQ PEDEADYYCCSYATS KTLVFGGGTKLTVV 747 EDE 72 EVQLVESGGGLVQPGGSLRLSCAASGFTFRSSAM 118 DVVMTQSPLSLPVTL C2 2 YWVRQAPGKGLEFVSCIRSNGVTHYADSVKGRF GQPASISCRSSRSLLNS TISRDNSKNTLHLQMGGLRPDDMAVYYCTRDD DGNTYLNWFHQRPG GPYSGYDWPWASSMDVWGQGTTVTVSS QSPRRLIFKLSNRDSG VPDRFSGSGSGTDFTL KISRVEAEDVGIYYC MQGTHWPVTFGGGT KVEIK 747 EDE 73 EVQLVESGGGLVQPGGSLRLSCAASGFIFSNHW 119 QSALTQPASVSGSPGQ D8 2 MHWVRQAPGKGLVWVSRTNSDGSSTSYADFVK SITISCTGTSSGVGSYN GRFTISRDNAKNTLHLQINSLRADDTAVYYCAR LVSWYQQHPGKAPKF DGVRYYYDSTGYYPDSYYEYGLDVWGQGTTVT IIYEGSKRPSGVSNRFS VSS GSNSGNTASLTISGLQ AEDEADYYCCSYAGS KTLVFGGGTKVTVL 747(4) EDE 74 EVQLVESGGGLVQPGGSLRLSCAASGFIFNRHW 120 QSVLTQPASVSGSPGQ A3 2 MHWVRQGPGKGLVWVSRINSDGSSTSYADSVK SITISCTGTSSDVGSYN GRFTISRDNAKNTLHLQINSLRAEDTAVYYCARD LVSWYQQHPGKAPKF GVRYYYDSTGYYPDSYYEYGMDVWGQGTTVT IIYEGSKRPSGVSNRFS VSS GSNSGNTASLTISGLQ AEDEADYYCCSYAGS KTLVFGGGTKVTVL 747(4) EDE 75 QVQLVQSGGALVKPGGSLRLSCVASGFTFGSHW 121 QSALTQPASVSGSPGQ A10 2 MHWVRQAPGKGLVWVSRVNSDGSSTNYADFV SITISCTGTSSDIGIYNL KGRFTTSRDNAENTLYLEMNSLTADDTAVYYCV VSWYQQHPGKAPKLII RDGVRYYYDSSGYYPDSFFKYGMDVWGQGTTV YEGSKRPSGVSNRFSA TVSS SKSAGAASLTISGLQP EDEADYYCCSYATSK TLVFGGGTKLTVV 747(4) EDE 3 EVQLVESGGGLVRPGGSLRLSCAASGFSYSNHW 39 QSVLTQPASVSGSPGQ A11 2 MHWVRQAPGKGLVWVSRINSDGSTRNYADFVK SITISCTGTSSNADTYN GRFTISRDNAENTLYLEMNSLTADDTAVYYCVR LVSWYQQRPGKAPKL DGVRFYYDSTGYYPDSFFKYGMDVWGQGTTVT MIYEGTKRPSGVSNRF VSS SASKSATAASLTISGL QPEDEADYYCCSYAT SRTLVFGGGTKLTVV 747(4) EDE 76 QVQLQESGPGLVRPSETLSLTCTVSGLSVSTYYW 122 EIVMTQSPATLSVSPG B4 2 SWIRQPPGKGLEWIAYVYSRGGTNYNPSLESRVT ERATLSCRASQSVKSN ISVDTATNQFSLRLRSVTAADTAVYFCARATNYF LAWYQQKPGQAPRLL DSSGYFFAPWFDPWGQGILVTVSS MYGASTRVVTIPARFS GSGSGTEFTLTISSLQS EDFAVYYCQQYNKW PLTFGGGTKVEIK 747(4) EDE 77 QVQLVQSGAEVKKPGSSVKVSCKASGGTRSSYA 123 QSALTQPASVSGSPGQ B6 2 ISWVRRAPGRGLEWMGVIIPFFGTANYAQIFQGR SITISCTGTSSDIGGFN LTITADESTSIANMELTSLTPEDTAIYYCASGGGG YVSWYQQHPGKAPK YAGYNWFDPWGQGTLVTVSS VMIFDVSNRPSGVSNR FSGSKSGNTASLTISG LQAEDEADYYCSSYT TRTTYVFGTGTKVTVL 747(4) EDE 4 EVQLVESGGGLVQPGGSLKLSCAASGFTFSSHW 40 QSALTQPASVSGSPGQ B7 2 MHWVRQAPGKGLVWVSRTNSDGSSTSYADSVK SITISCTGISSDVETYN GRFMISRDNSKNTVYLHMNGLRAEDTAVYFCAR LVSWYEQHPGKAPKL DGVRYYYDSTGYYPDNFFQYGLDVWGQGTTVT IIYEASKRPSGVSNRFS VSS GSKSGNTASLAISGLQ AEDEADYYCCSYAGG KSLVFGGGTRLTVL 747(4) EDE 78 EVQLVQSGGGLIQPGGSLKLSCAASGFSFRNHW 124 QSALTQPASVSGSPGQ D6 2 MHWVRQAPGKGLVWVSRVNSDGYSTSYADSV SITISCSGFSSDVGGDK KGRFTISRDNAKNTLYLQMNSLRPEDTAVYFCA VVSWYEQHPGKVPKL RDGVRFYSDSTGYYPDNYFPYGMDVWGQGTTV IIYEGSKRPSGVSNRFS TVSS GSKSGNTASLTISGLQ AEDEADYYCCSYAGP KTLVFGGGTKVTVL 747 EDE 79 EVQLVESGGGLVQPGGSLRLSCKVSGFTFKAYW 125 NSPLSLSASVGDRVTI B2 2 MHWVRQAPGKGLVWVSRINGLGSSRDYADSVR TCRASRTIDNFLHWY GRFTISRDDAENTVYLQMNSLTAEDTAMYYCAR QQKPGKAPNLLIYAA DVXFHDSSGYYRXGFXAPWG SSLQSGVPSRFRGSGS GTDFTLTINSVQPEDF ATYYCQQSYTIPPTFG GGTKVEIR 747 EDE 80 EVQLVESGGGLVQPGGSLRLSCAASGFAFSNHW 126 QSALTQPASVSGSLGQ C4 2 MHWVRQAPGKGLVWVSRINSDGSSTTYADSVK SITISYTGTAIDVGSYN GRFTISRDNAKNTLSLELNSLRAEDTAIYYCARD LVSWYQQHPGKVPKL GVRFYYDSTGYYPDPYFQYGLDVWGQGTTVTV MIYEGSKRPSGVSNRF SS FGSKSGNTASLTISGL QSEDEAEYYCCSYGG SRTLLFGGGTKLTVL 747 EDE 81 EVQLVESGGGLVQPGASLRVSCAASGFTFSTYN 127 DIVMTQSPLSLPVTLG C7 2 MNWVRQAPGKGLEWVSYISSRSSTIYYADSVQG EPASISCRSSRSLLHSN RFTISRDNAKNSLYLQMNSLRAEDTAVYYCARD GYNYLDWYLQKPGQ IGHYYDSSGYFHYSFGMDVWGQGTTVTVSS SPQLLIYLGSNRASGV PDRFSGSGSGTDFTLK ISRVEAEDVGVYYCM QARQTPVTFGGGTKV EIK 747 EDE 82 EVQLVESGGGLVQPGGSLRLSCAASGFIFRNYW 128 GPFTLSASVGDRVTIT D5 2 MHWVRQAPGKGLVWVSRINGLGSTTTYADSVE CRASRSINTFLNWYQ GRFTITRDDAKNTIFLQMNSLRAEDTAVYYCAR QKTGSAPKLLIYGAST DVNFYDSSGYYREGWFDSWGPGTTVTVSS LQSGVPSRFSGSGSGT DFALTITSLQPDDFAA YYCQQSYTTPLTFGG GTRVEIK 747 EDE 83 EVQLLESGAEVKKPGSSVKISCKASGGTFSNYAIS 129 SYELTQPPSVSVAPGK D11 2 WVRQAPGRGLEWLGGIIPIFGTPNYAQRFQGRVT TATITCGGDNIGSKTV ITADESTSTAYMELNSLTSDDTAIYYCARDHPTVI HWYQQKPGQAPLLVI NPTFVGSWFDPWGQGTLVTVSS YYNGDRPPGIPERFSG SNSGNTATLTITRVEA GDEADYCCQIWDSRS SHPVFGGGTKLTVL 752 EDE 84 QVQLVESGAEVKKPGASVKVSCKASGFTFTSYYI 130 DIVMTQSPLSLPVTPG B6 2 HWVRQAPGQGLEWMGVINPSGGTTIYARNLQG EPASISCRSSQSLLHTN RVTMTRDTSTTTVYMELSSLKSEDTAVYYCARA GYNFLDWYVQKPGQ HSGNYDFWSGSNYHYYYGMDVWGQGTTVTVS SPQLLIYLGSSRASGV S PDRFSGSGSGTDFTLK ISRVEAEDVGLYYCM QALHTPRTFGQGTKV EIK 752(2) EDE 85 EVQLVESGAEVKKPGASVKVSCKASGFTFTSYYI 131 DIVMTQSPLSLPVTPG D2 2 HWVRQAPGQGLEWMGVINPSGGTTIYAQNFQG EPASISCRSSQSLLHTN RVTMTRDTSTTTVYMELSSLKSEDTAVYYCARA GYNFLDWYVQKPGQ HSGNYDFWSGSNYHYYYGMDVWGQGTTVTVS SPQLLIYLGSSRASGV S PDRFSGSGSGTDFTLK ISRVEAEDVGLYYCM QALQTPRTFGQGTKV EIK

TABLE-US-00004 envelopeectodomainproteinsequenceDENV1 SEQIDNO:132 FHLTTRGGEPHMIVSKQERGKSLLFKTSAGVNMCTLIAMDLGELCEDTMTYKCP RITEAEPDDVDCWCNATDTWVTYGTCSQTGEHRRDKRSVALAPHVGLGLETRT ETWMSSEGAWKQIQKVETWALRHPGFTVIALFLAHAIGTSITQKGIIFILLMLVTP SMAMRCVGIGNRDFVEGLSGATWVDVVLEHGSCVTTMAKNKPTLDIELLKTEV TNPAVLRKLCIEAKISNTTTDSRCPTQGEATLVEEQDANFVCRRTVVDRGWGNG CGLFGKGSLLTCAKFKCVTKLEGKIVQYENLKYSVIVTVHTGDQHQVGNETTEH GTIATITPQAPTSEIQLTDYGTLTLDCSPRTGLDFNEVVLLTMKEKSWLVHKQWF LDLPLPWTSGASTSQETWNRQDLLVTFKTAHAKKQEVVVLGSQEGAMHTALTG ATEIQTSGTTTIFAGHLKCRLKMDKLTLKGMSYVMCTGSFKLEKEVAETQHGTV LVQVKYEGTDAPCKIPFSTQDEKGVTQNGRLITANPIVTDKEKPINIETEPPFGESY IIVGAGEKALKLSWFKKG envelopeectodomainproteinsequenceDENV2 SEQIDNO:133 MRCIGISNRDFVEGVSGGSWVDIVLEHGSCVTTMAKNKPTLDFELIKTEAKQPAT LRKYCIEAKLTNTTTESRCPTQGEPSLNEEQDKRFICKHSMVDRGWGNGCGLFG KGGIVTCAKFTCKKNMEGKIVQPENLEYTIVITPHSGEEHAVGNDTGKHGKEIKIT PQSSTTEAELTGYGTVTMECSPRTGLDFNEMVLLQMEDKAWLVHRQWFLDLPL PWLPGADTQGSNWIQKETLVTFKNPHAKKQDVVVLGSQEGAMHTALTGATEIQ MSSGNLLFTGHLKCRLRMDKLQLKGMSYSMCTGKFKIVKEIAETQHGTIVIRVQ YEGDGSPCKIPFEITDLEKRHVLGRLITVNPIVTEKDSPVNIEAEPPFGDSYIIVGVE PGQLKLNWFKRG envelopeectodomainproteinsequenceDENV3 SEQIDNO:134 FHLTSRDGEPRMIVGKNERGKSLLFKTASGINMCTLIAMDLGEMCDDTVTYKCP HITEVEPEDIDCWCNLTSTWVTYGTCNQAGEHRRDKRSVALAPHVGMGLDTRT QTWMSAEGAWRQVEKVETWALRHPGFTILALFLAHYIGTSLTQKVVIFILLMLV TPSMTMRCVGVGNRDFVEGLSGATWVDVVLEHGGCVTTMAKNKPTLDIELQKT EATQLATLRKLCIEGKITNITTDSRCPTQGEAILPEEQDQNYVCKHTYVDRGWGN GCGLFGKGSLVTCAKFQCLESIEGKVVQHENLKYTVIITVHTGDQHQVGNETQG VTAEITSQASTAEAILPEYGTLGLECSPRTGLDFNEMILLTMKNKAWMVHRQWFF DLPLPWTSGATTKTPTWNRKELLVTFKNAHAKKQEVVVLGSQEGAMHTALTGA TEIQTSGGTSIFAGHLKCRLKMDKLKLKGMSYAMCLNTFVLKKEVSETQHGTILI KVEYKGEDAPCKIPFSTEDGQGKAHNGRLITANPVVTKKEEPVNIEAEPPFGESNI VIGIGDKALKINWYRKG envelopeectodomainproteinsequenceDENV4 SEQIDNO:135 FSLSTRDGEPLMIVAKHERGRPLLFKTTEGINKCTLIAMDLGEMCEDTVTYKCPL LVNTEPEDIDCWCNLTSTWVMYGTCTQSGERRREKRSVALTPHSGMGLETRAET WMSSEGAWKHAQRVESWILRNPGFALLAGFMAYMIGQTGIQRTVFFVLMMLVA PSYGMRCVGVGNRDFVEGVSGGAWVDLVLEHGGCVTTMAQGKPTLDFELTKT TAKEVALLRTYCIEASISNITTATRCPTQGEPYLKEEQDQQYICRRDVVDRGWGN GCGLFGKGGVVTCAKFSCSGKITGNLVQIENLEYTVVVTVHNGDTHAVGNDTSN HGVTAMITPRSPSVEVKLPDYGELTLDCEPRSGIDFNEMILMKMKKKTWLVHKQ WFLDLPLPWTAGADTSEVHWNYKERMVTFKVPHAKRQDVTVLGSQEGAMHSA LAGATEVDSGDGNHMFAGHLKCKVRMEKLRIKGMSYTMCSGKFSIDKEMAETQ HGTTVVKVKYEGAGAPCKVPIEIRDVNKEKVVGRIISSTPLAENTNSVTNIELEPPF GDSYIVIGVGNSALTLHWFRKG envelopeectodomainnucleotidesequenceDENV1 SEQIDNO:136 ttccatttgaccacacgagggggagagccacacatgatagttagtaagcaggaaagagga aagtcactcttgttcaagacctctgcaggtgtcaatatgtgcactctcattgcgatggat ttgggagagttatgtgaggacacaatgacttacaaatgcccccggatcactgaggcggaa ccagatgacgttgactgctggtgcaatgccacagacacatgggtgacctatgggacgtgt tctcaaaccggtgaacaccgacgagacaaacgttccgtggcactggccccacacgtggga cttggtctagaaacaagaaccgaaacatggatgtcctctgaaggcgcctggaaacaaata caaaaagtggagacttgggctttgagacacccaggattcacggtgatagctcttttttta gcacatgccataggaacatccatcactcagaaagggatcattttcattctgctgatgctg gtaacaccatcaatggccatgcgatgcgtgggaataggcaacagagacttcgttgaagga ctgtcaggagcaacgtgggtggacgtggtattggagcatggaagctgcgtcaccaccatg gcaaaaaataaaccaacattggacattgaactcttgaagacggaggtcacgaaccctgcc gtcttgcgcaaattgtgcattgaagctaaaatatcaaacaccaccaccgattcaagatgt ccaacacaaggagaggctacactggtggaagaacaagacgcgaactttgtgtgtcgacga acggttgtggacagaggctggggcaatggctgcggactatttggaaaaggaagcctactg acgtgtgctaagttcaagtgtgtgacaaaactggaaggaaagatagttcaatatgaaaac ttaaaatattcagtgatagtcactgtccacacaggggaccagcaccaggtgggaaacgag actacagaacatggaacaattgcaaccataacacctcaagctcctacgtcggaaatacag ttgacagactacggaacccttacactggactgctcacccagaacagggctggactttaat gaggtggtgctattgacaatgaaagaaaaatcatggcttgtccacaaacaatggtttcta gacttaccactgccttggacttcgggggcttcaacatcccaagagacttggaacagacaa gatttgctggtcacattcaagacagctcatgcaaagaagcaggaagtagtcgtactggga tcacaggaaggagcaatgcacactgcgttgaccggggcgacagaaatccagacgtcagga acgacaacaatctttgcaggacacctgaaatgcagattaaaaatggataaactgacttta aaagggatgtcatatgtgatgtgcacaggctcatttaagctagagaaggaagtggctgag acccagcatggaactgtcctagtgcaggttaaatacgaaggaacagatgcgccatgcaag atccccttttcgacccaagatgagaaaggagtgacccagaatgggagattgataacagcc aatcccatagttactgacaaagaaaaaccaatcaacattgagacagaaccaccttttggt gagagctacatcatagtaggggcaggtgaaaaagctttgaaactaagctggttcaagaaa gga envelopeectodomainnucleotidesequenceDENV2 SEQIDNO:137 ttccatttaaccacacgtaacggagaaccacacatgatcgtcagtagacaagagaaaggg aaaagtcttctgtttaaaacagaggatggtgtgaacatgtgtaccctcatggccatggac cttggtgaattgtgtgaagatacaatcacgtacaagtgtccttttctcaggcagaatgaa ccagaagacatagattgttggtgcaactctacgtccacatgggtaacttatgggacgtgt accaccacaggagaacacagaagagaaaaaagatcagtggcactcgttccacatgtggga atgggactggagacacgaactgaaacatggatgtcatcagaaggggcctggaaacatgcc cagagaattgaaacttggatcttgagacatccaggctttaccataatggcagcaatcctg gcatacaccataggaacgacacatttccaaagagccctgattttcatcttactgacagct gtcgctccttcaatgacaatgcgttgcataggaatatcaaatagagactttgtagaaggg gtttcaggaggaagctgggttgacatagtcttagaacatggaagctgtgtgacgacgatg gcaaaaaacaaaccaacattggattttgaactgataaaaacagaagccaaacaacctgcc actctaaggaagtactgtatagaggcaaagctgaccaacacaacaacagattctcgctgc ccaacacaaggagaacccagcctaaatgaagagcaggacaaaaggttcgtctgcaaacac tccatggtggacagaggatggggaaatggatgtggattatttggaaaaggaggcattgtg acctgtgctatgttcacatgcaaaaagaacatgaaaggaaaagtcgtgcaaccagaaaac ttggaatacaccattgtgataacacctcactcaggggaagagcatgcagtcggaaatgac acaggaaaacatggcaaggaaatcaaaataacaccacagagttccatcacagaagcagag ttgacaggctatggcactgtcacgatggagtgctctccgagaacgggcctcgacttcaat gagatggtgttgctgcaaatggaaaataaagcttggctggtgcacaggcaatggttccta gacctgccgttgccatggctgcccggagcggacacacaaggatcaaattggatacagaaa gagacattggtgactttcaaaaatccccatgcgaagaaacaggatgttgttgttttggga tcccaagaaggggccatgcacacagcactcacaggggccacagaaatccagatgtcatca ggaaacttactgttcacaggacatctcaagtgcaggctgaggatggacaaactacagctc aaaggaatgtcatactctatgtgcacaggaaagtttaaagttgtgaaggaaatagcagaa acacaacatggaacaatagttatcagagtacaatatgaaggggacggttctccatgtaag atcccttttgagataatggatttggaaaaaagacatgttttaggtcgcctgattacagtc aacccaatcgtaacagaaaaagatagcccagtcaacatagaagcagaacctccattcgga gacagctacatcatcataggagtagagccgggacaattgaagctcaactggtttaagaaa gga envelopeectodomainnucleotidesequenceDENV3 SEQIDNO:138 ttccacttaacttcacgagatggagagccgcgcatgattgtggggaagaatgaaagagga aaatccctactttttaagacagcctctggaatcaacatgtgcacactcatagccatggat ttgggagagatgtgtgatgacacggtcacttacaaatgcccccacattaccgaagtggag cctgaagacattgactgttggtgcaaccttacatcgacatgggtgacttatggaacatgc aatcaagctggagagcatagacgcgataagagatcagtggcgttagctccccatgtcggc atgggactggacacacgcactcaaacctggatgtcggctgaaggagcttggagacaagtc gagaaggtagagacatgggcccttaggcacccagggtttaccatactagccctatttctt gcccattacataggcacttccttgacccagaaagtggttatttttatactattaatgctg gttaccccatccatgacaatgagatgtgtgggagtaggaaacagagattttgtggaaggc ctatcgggagctacgtgggttgacgtggtgctcgagcacggtgggtgtgtgactaccatg gctaagaacaagcccacgctggacatagagcttcagaagactgaggccactcagctggcg accctaaggaagctatgcattgagggaaaaattaccaacataacaaccgactcaagatgt cccacccaaggggaagcgattttacctgaggagcaggaccagaactacgtgtgtaagcat acatacgtggacagaggctggggaaacggttgtggtttgtttggcaagggaagcttggtg acatgcgcgaaatttcaatgtttagaatcaatagagggaaaagtggtgcaacatgagaac ctcaaatacaccgtcatcatcacagtgcacacaggagaccaacaccaggtgggaaatgaa acgcagggagttacggctgagataacatcccaggcatcaaccgctgaagccattttacct gaatatggaaccctcgggctagaatgctcaccacggacaggtttggatttcaatgaaatg attttattgacaatgaagaacaaagcatggatggtacatagacaatggttctttgactta cccctaccatggacatcaggagctacaacaaaaacaccaacttggaacaggaaagagctt cttgtgacatttaaaaatgcacatgcaaaaaagcaagaagtagttgtccttggatcacaa gagggagcaatgcatacagcactgacaggagctacagagatccaaacctcaggaggcaca agtatttttgcggggcacttaaaatgtagactcaagatggacaaattgaaactcaagggg atgagctatgcaatgtgcttgaatacctttgtgttgaagaaagaagtctccgaaacgcag catgggacaatactcattaaggttgagtacaaaggggaagatgcaccctgcaagattcct ttctccacggaggatggacaagggaaagctcacaatggcagactgatcacagccaatcca gtggtgaccaagaaggaggagcctgtcaacattgaggctgaacctccttttggggaaagt aatatagtaattggaattggagacaaagccctgaaaatcaactggtacaggaagggaa envelopeectodomainnucleotidesequenceDENV4 SEQIDNO:139 ttttccctcagcacaagagatggcgaacccctcatgatagtggcaaaacatgaaaggggg agacctctcttgtttaagacaacagaggggatcaacaaatgcactctcattgccatggac ttgggtgaaatgtgtgaggacactgtcacgtataaatgccccctactggtcaataccgaa cctgaagacattgattgctggtgcaacctcacgtctacctgggtcatgtatgggacatgc acccagagcggagaacggagacgagagaagcgctcagtagctttaacaccacattcagga atgggattggaaacaagagctgagacatggatgtcatcggaaggggcttggaagcatgct cagagagtagagagctggatactcagaaacccaggattcgcgctcttggcaggatttatg gcttatatgattgggcaaacaggaatccagcgaactgtcttctttgtcctaatgatgctg gtcgccccatcctacggaatgcgatgcgtaggagtaggaaacagagactttgtggaagga gtctcaggtggagcatgggtcgacctggtgctagaacatggaggatgcgtcacaaccatg gcccagggaaaaccaaccttggattttgaactgactaagacaacagccaaggaagtggct ctgttaagaacctattgcattgaagcctcaatatcaaacataactacggcaacaagatgt ccaacgcaaggagagccttatctgaaagaggaacaggaccaacagtacatttgccggaga gatgtggtagacagagggtggggcaatggctgtggcttgtttggaaaaggaggagttgtg acatgtgcgaagttttcatgttcggggaagataacaggcaatttggtccaaattgagaac cttgaatacacagtggttgtaacagtccacaatggagacacccatgcagtaggaaatgac acatccaatcatggagttacagccatgataactcccaggtcaccatcggtggaagtcaaa ttgccggactatggagaactaacactcgattgtgaacccaggtctggaattgactttaat gagatgattctgatgaaaatgaaaaagaaaacatggctcgtgcataagcaatggtttttg gatctgcctcttccatggacagcaggagcagacacatcagaggttcactggaattacaaa gagagaatggtgacatttaaggttcctcatgccaagagacaggatgtgacagtgctggga tctcaggaaggagccatgcattctgccctcgctggagccacagaagtggactccggtgat ggaaatcacatgtttgcaggacatcttaagtgcaaagtccgtatggagaaattgagaatc aagggaatgtcatacacgatgtgttcaggaaagttttcaattgacaaagagatggcagaa acacagcatgggacaacagtggtgaaagtcaagtatgaaggtgctggagctccgtgtaaa gtccccatagagataagagatgtaaacaaggaaaaagtggttgggcgtatcatctcatcc acccctttggctgagaataccaacagtgtaaccaacatagaattagaacccccctttggg gacagctacatagtgataggtgttggaaacagcgcattaacactccattggttcaggaaa ggg A11Lightchain SEQIDNO:140 QSVLTQPVSVSGSPGQSITISCTGTSSNADTYNLVSWYQQRPGKAPKLMIYEGTK RPSGVSNRFSASKSATAASLTISGLQPEDEADYYCCSYATSRTLVFGGGTKLTVV B7Lightchain SEQIDNO:141 RSQSALTQPASVSGSPGQSITISCTGISSDVETYNLVSWYEQHPGKAPKLIIYEASK RPSGVSNRFSGSKSGNTASLAISGLQAEDEADYYCCSYAGGKSLVFGGGTRLTVL GQPKAAPSVTLFPPSSEELQANKATLVCLISDFYPGAVTVAWKADSSPVKAGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTVEKTVAPTECS vHchainofA11 SEQIDNO:142 EVQLVESGGGLVRPGGSLRLSCAASGFSYSNHWMHWVRQAPGKGLVWVSRINS DGSTRNYADFVKGRFTISRDNAENTLYLEMNSLTADDTAVYYCVRDGVRFYYD STGYYPDSFFKYGMDVWGQGTTVTV vHB7 SEQIDNO:143 EVQLVESGGGLVQPGGSLKLSCAASGFTFSSHWMHWVRQAPGKGLVWVSRTNS DGSSTSYADSVKGRFMISRDNSKNTVYLHMNGLRAEDTAVYFCARDGVRYYYD STGYYPDNFFQYGLDVWGQGTTVTV vHC8 SEQIDNO:144 EVQLVESGGGLVQPGGSLRLSCSASGFTFSTYSMHWVRQAPGKGLEYVSAITGE GDSAFYADSVKGRFTISRDNSKNTLYFEMNSLRPEDTAVYYCVGGYSNFYYYYT MDVWGQGTTVTV vLightC8 SEQIDNO:145 EIVLTQSPATLSLSPGERATLSCRASQSISTFLAWYQHKPGQAPRLLIYDASTRATG VPARFSGSRSGTDFTLTISTLEPEDFAVYYCQQRYNWPPYTFGQGTKVEIK vHC10 SEQIDNO:146 EVQLVESGAEVKKPGASVKVSCKASGYTFTSYAMHWVRQAPGQRLEWMGWIN AGNGNTKYSQKFQDRVTITRDTSASTAYMELSSLRSEDTAIYYCARDKVDDYGD YWFPTLWYFDYWGQGTLVTV vLC10 SEQIDNO:147 QSALTQPASVSGSPGQSITISCTGTSSDVGGFNYVSWFQQHPGKAPKLMLYDVTS RPSGVSSRFSGSKSGNTASLTISGLQAEDEADYYCSSHTSRGTWVFGGGTKLTVL 150loopofDenv-1 SEQIDNO:148 QHQVGNETTEHG 150loopofDenv2 SEQIDNO:149 EHAVGNDTGKHG 150loopofDenv3 SEQIDNO:150 QHQVGNETQG 150loopofDenv4 SEQIDNO:151 THAVGNDIPNHG

Example 17. Site-Directed Mutagenesis of the DV2 E Protein in Order to Obtain Stable E Dimers

[0827] Based on the 3D structure, we generated 3 different E mutants in order to create disulphide bonds to stabilise the E dimer. The first had A259C, the second S255C and the third had two simultaneous changes: L107C and A313C (FIG. 32). The mutants were expressed in insect S2 cells using the same procedure that we had developed for wild-type DV-2 E protein, as described above. Mutant A259C gave the highest yields of purified dimeric protein, but the other two constructs also yielded reasonable amounts of cross-linked dimers. The mutants were compared to wild type in antibody-binding and in mice immunization experiments.

1. Binding of FLE and EDE mAbs to the Mutants.

[0828] A panel of FLE (fusion loop epitope), EDE (envelope dimer epitope) and other (non FLE) mAbs were tested on the mutant and WT protein by ELISA. FIGS. 33-36 show the binding activity of FLE and EDE mAbs on the A259C mutant and WT.

[0829] FIG. 37-40 shows the binding activity of FLE and EDE mAbs on the L107/A313 mutant and WT

2 Mice Immunisation with the A259C Mutant

[0830] Mice were set into 6 groups and immunised as prime followed by boost as describe below [0831] Group 1 prime and boost with E WT. E WT (monomer/monomer) [0832] Group 2, prime and boost with E A259C mutant (dimer/dimer) [0833] Group 3 prime and boost with prM/E viral like particle (VLP) (VLP/VLP) [0834] Group 4 prime with E A259C mutant followed by boosting with VLP (dimer/VLP) [0835] Group 5 prime with VLP followed by boosting with E A259C mutant (VLP/dimer) [0836] Group 6 control mice (mock)

[0837] FIGS. 41-45 shows the anti E antibody titre, binding to yeast expressing E domain 1 to 3 (all 3 domains), domain 1-2 and domain 3, serotype cross reactivity, neutralisation on insect and DC virus and ADE on insect virus.

Methods

Recombinant Soluble DENV Envelope Protein Binding ELISA

[0838] To determine the binding affinity of human monoclonal Abs to recombinant soluble DENV envelope protein (rE), the Nunc Immobilizer Amino plates (436006, Thermo Scientific) were directly coated with 50 l of 10 g/ml rE DENV2 wild type monomer (WT), mutant dimer (A259C or L107C/A313C) or bovine serum albumin (BSA; negative control) in 50 mM carbonate buffer pH 9.6 (C3041, Sigma). Following overnight incubation at 4 C., plates were washed 3 times with wash buffer (PBS+0.1% Tween-20) and blocked with 200 l blocking buffer (PBS+3% BSA) for 1 hr at the room temperature followed by 50 l of 1-10 g/ml human monoclonal Abs in blocking buffer at 37 C. for 1 hr. Afterwards, Plates were washed again 3 times and further incubated with 50 l of ALP-conjugated anti-human IgG at 1:10,000 dilutions in blocking buffer (A9544, Sigma) for 1 hr at 37 C. Finally, after 3 washing, 100 l of PNPP substrate (N2770, Sigma) was added and left for 1 hr at the room temperature. The reaction was measured at 405 nm.

Mice

[0839] Female C57BL/6 mice were obtained from Harlan UK (Bicester, UK). Mice were used at 6-8 weeks of age. All animal experiments were performed in accordance with United Kingdom governmental regulations (Animal Scientific Procedures Act 1986) and were approved by the United Kingdom Home Office.

Immunization Experiment

[0840] Mice were intra-peritoneally administered with 1% v/v of antigen (5 g) co-adsorbed on 2% alhydrogel (Invivogen). The antigen-alum mix was allowed to stand for about 5 min prior to injection. At 3 weeks post priming, a booster injection was given with 5 g antigen similarly adsorbed on alum. Serum samples were collected at 3 weeks following the boost and tested in various assays. DV2-VLP supernatant was generated by PEI mediated transfection of HEK293T cells with pHLsec-prM-E plasmid DNA. The VLP supernatant collected in UltraDoma protein free medium (Lonza, USA) was concentrated and buffer exchanged to PBS using Centricon (100 KDa cut-off). E-protein was estimated using capture ELISA. Briefly VLP supernatant was captured using mouse anti-FL (4G2) and detected using DENV-specific human antibody to E protein, 30-E2 (from patients); followed by AP-conjugated antibody to human IgG (A9544; Sigma). The colorimetric reaction was developed using PNPP substrate and absorbance measured at 405 nm. E-protein in the VLP supernatant was quantified based on non-linear regression analysis of standard curve generated with purified E-protein monomer. The E-protein equivalent used for immunization was 7 ng/mouse corresponding to a total protein concentration of 5 mg/mouse. The total protein concentration for VLP preparation was performed by Bradford method using BSA as standard.

Measurement of Anti E Antibody Titre on Live Virus

[0841] Virus from supernatants of C636 cells infected with various Dengue serotypes was captured on Maxisorp immunoplate (442404; NUNC) coated with 10 g/ml human anti-prM antibody, 3-147. Wells were then incubated with various dilutions of mouse serum diluted in 1% BSA, followed by 1:2000 dilution of Fe-specific goat anti-mouse IgG-alkaline phosphatase conjugate (A2429, Sigma). Reaction was visualized by the addition of PNPP substrate and read for absorbance at 405 nm after the reaction was stopped with 0.4N NaOH. Data was plotted and analysed using GraphPad prism v6.03.

Neutralization Assay

[0842] The neutralization potential of mouse sera was determined using the Focus Reduction Neutralization Test (FRNT), where the reduction in the number of the infected foci is compared to control (no antibody). For FRNT, Fifty-five microlitres of DENV-derived C6/36 cells (C6/36 DENV) or DENV-derived DC (DC-DENV) were mixed with an equal volume of serial 3-fold dilutions of mouse sera (from 1:50 to 1:36450 and incubated for 1 hr at 37 C. Fifty microlitres of the mixtures were then transferred to Vero cell monolayer in duplicate in 96-well plate and incubated for 3 days at 37 C. The focus-forming assay was then performed by washing the cell monolayer with 200 l of PBS twice. Cells were then fixed with 100 l of 3.7% formaldehyde in PBS for 10 min at the room temperature and then permeabilized with 100 l of 2% TritonX-100 in PBS for 10 min at the room temperature. Following 2 times wash with PBS, 50 l of mouse monoclonal anti-DENV envelope Ab (4G2) was added to each well and incubated for 2 hrs, at 37 C. Cells were washed again with PBS and incubated for 1 hr at 37 C. with 50 l of HRP-conjugated goat anti-mouse IgG (P0447, Dako) at 1:1,000 dilutions in 0.05% tween-20/2% FBS in PBS. The reaction was visualized by the addition of DAB substrate (PBS+0.05 g/ml DAB+0.03% H.sub.2O.sub.2+0.32% NiCl.sub.2). The percentage focus reduction was calculated for each antibody dilution.

Antibody Dependent Enhancement Assay

[0843] Serially diluted heat-inactivated mouse serum or control antibody (anti FL: 4G2) was pre-incubated with DV2-virus for 1 h at 37 C. The virus-antibody complexes were then transferred to U937 cells (Fc receptor-bearing human monocyte cell lines) plated at 110.sup.5 cells/well. Cells were incubated with virus-antibody complexes for 4 days and viral titres determined by titration on Vero cells by a focus-forming assay using anti-FL, 4G2 antibody for detection. The virus titres were read out as focus-forming units per ml and fold enhancement of infection calculated based on the titres observed in the absence of antibody. Data was plotted and analysed using GraphPad prism v6.03.

CONCLUSIONS

[0844] The further data in this Example shows that we can make dimer; that it is correctly folded; binds to the EDE antibodies and is immunogenic.

[0845] FIG. 32 shows the locations of single and double site mutations. The A259C mutant binds to the panel of EDE 1 antibodies (FIG. 33). Likewise EDE2 panel antibodies bind also bind to the A259C mutant (FIG. 34). However (FIG. 35) FLE panel antibodies also bind (which is considered less desirable) probably because the E monomers can pivot around the central cysteine link allowing access to the FLE. The non-FLE panel of antibodies (antibodies which have not been mapped, termed non-fusion loop epitope (non FLE) mAb)) likewise bind to the A259C dimer (FIG. 36).

[0846] The double mutant L107C and A313C likewise forms a stable dimer which binds EDE1 antibodies (FIG. 37) and EDE2 antibodies (FIG. 38). However, this double mutant is locked at both ends and is much less recognised by the FLE antibodies (FIG. 39), which is ideal as one would prefer an immunogen that did not promote the generation of a FLE response. There is also less non-FLE recognition (FIG. 40) which is also good.

[0847] A series of mouse immunisations with different combinations of monomer, dimer and vlp are shown in FIG. 41. Dimer+/VLP generate good serum antibodies recognising DENV2 virus particles (FIG. 41). Combinations with VLP initiate good cross reactive binding responses (FIG. 42), and are expected to provide good neutralisation. A tetravalent or prime boost approach may be required to generate broad neutralisation.

[0848] VLP was used at 5 g E protein equivalent ie amount of E WT, mutant and VLP were 5 g. E WT and mutant were protein and measured concentration based on OD whereas E conc on VLP prep was measured by ELISA and WT E protein was used for setting up a standard curve. Thus, E-protein was estimated using capture ELISA. Briefly VLP supernatant was captured using mouse anti-FL (4G2) and detected using DENV-specific human antibody to E protein, 30-E2 (from patients); followed by AP-conjugated antibody to human IgG (A9544; Sigma). The colorimetric reaction was developed using PNPP substrate and absorbance measured at 405 nm. E-protein in the VLP supernatant was quantified based on non-linear regression analysis of standard curve generated with purified E-protein monomer. 5 g of VLP is considered to correspond to total protein containing about 7 ng of E protein equivalent.

[0849] The VLP may induce anti-prM activity that will not have been induced by monomer or dimer. The anti-prM activity induced by the VLP may contribute to the virion binding, cross reactivity, ADE and neutralisation results.

[0850] Neutralisation results of DENV2 show superior response from the dimer above the monomer on both insect (high prM; FIG. 43) and DC virus (low prM; FIG. 44) viruses. FIG. 45 indicates that antibodies raised to the A259C dimer can still cause ADE (Antibody dependent enhancement of DENV infection). The A259C dimer also reacts with FL-Abs and thus could elicit FL-like Abs which cause strong ADE and displace EDE Abs. It is not yet known whether the L107C/A313C dimer induces ADE.

[0851] One possibility to test which component is important for ADE is to deplete serum from DIII binding Abs and to perform ADE tests again.

[0852] We are also performing other cavity filling approaches to the stable dimer in order to enhance the desired EDE response and minimise the less desirable FLE and nonFLE/nonEDE responses. Extensive mutagenic resurfacing of the dimer is also performed to further reduce the generation of non-EDE suboptimal responses by mutation of residues or addition of glycan (to assist in masking the less desirable FLE and nonFLE/nonEDE epitopes/responses). Modelling and optimisation of the core EDE epitope is also performed to produce an optimal sequence to induce BNA's (broadly neutralising antibodies).

[0853] Priming and boosting with a variety of heterologous techniques may be required to focus in on the EDE.

[0854] A further dimer that is considered to be useful is a A259C/S255C double mutant, which may (similarly to the L107C/T313C double mutant) provide a dimer in which the FLE is less accessible.

[0855] A further mutation that is considered to reshape the kl-loop and to mimick the virion-like conformation is L278F as discussed above. Combinations of such a mutation and one or more mutations to establish cysteine links between monomers to form a dimer may be useful. As noted above, a molecule displaying the EDE, for example a stabilised dimer, may be useful in screening for broadly neutralising antibodies, for example.

Example 18

Further Strategies for Optimising EDE Constructs or Binding Compounds

Protein Folding

[0856] In order to promote proper folding and assembly of stabilised dimer molecules, it may be useful to co-express EDE-binding compounds, for example Fabs or scFv in the same cells as the E protein. This is considered to aid protein folding and may assist in eliminating or reducing protein aggregates.

Reduction in prM Level

[0857] It may be desirable and possible to produce VLPs that lack prM, thereby potentially increasing their immunogenicity.

scFv Optimisation

[0858] A yeast display screen, for example, could be used to screen for optimised scFvs. Already-identified scFvs, for example, can be randomly (or non-randomly) mutated and expressed in yeast. Recombinant stabilised E dimers from the four serotypes can be prepared and each tagged with a different colour. The scFv-expressing yeast can be stained using these tagged proteins, and yeast cells that carry all four colours selected (as the scFvs are able to bind to the stabilised E dimer from each of the four serotypes.

[0859] Yeast staining may be carried based on the following. Yeast cells expressing E-protein domain 1+2 or domain 3 or all 3 domains were washed with PBS. Cells were resuspended in FACS buffer (PBS containing 1% FCS, 0.5% BSA) and aliquoted in 96-well U bottom plates. Mouse serum samples (diluted to 1:300) were added to cells and cells were incubated overnight at 4 degrees. Cells were washed and stained with 1:150 dilution of PE-conjugated (F.sub.ab).sub.2 fragment of rabbit anti-mouse Ig (Dako R0439). Cells were stained for 30 min at 4 degrees, washed well in PBS and fixed using 1% PFA in PBS. Data were acquired using a FACS.sub.VERSE (Becton-Dickinson, Mountain View, CA) and analysed using FlowJo software, (TreeStar, Ashland, OR).