Structure

Article Mechanism of Dengue Virus Broad Cross-Neutralization by a Monoclonal Antibody Joseph J.B. Cockburn,1,4,8 M. Erika Navarro Sanchez,1,4 Nickolas Fretes,1,4,9 Agathe Urvoas,2,5,10 Isabelle Staropoli,3,6 Carlos M. Kikuti,1,4,11 Lark L. Coffey,4,7,12 Fernando Arenzana Seisdedos,3,6 Hugues Bedouelle,2,5,* and Felix A. Rey1,4,* 1Unite ´

de Virologie Structurale, De´partement de Virologie de Recherche Pre´vention et The´rapie Mole´culaires des Maladies Humaines, De´partement d’Infection et Epide´miologie 3Unite ´ de Pathoge´nie Virale, De´partement de Virologie Institut Pasteur, 75724 Paris Cedex 15, France 4CNRS, URA3015 5CNRS, URA3012 6INSERM, U819 75724 Paris Cedex 15, France 7De ´ partement de Virologie, Groupe a` 5 ans Populations Virales et Pathogene`se, 75724 Paris Cedex 15, France 8Present address: London Research Institute, Cancer Research UK, London WC2A 3LY, UK 9Present address: University of Michigan Medical School, Ann Arbor, MI 48109, USA 10Present address: Universite ´ Paris 11, Institut de Biochimie Biophysique Mole´culaire et Cellulaire IBBMC, CNRS UMR 8619, 91405 Orsay, France 11Present address: Institut Curie, CNRS UMR144, 75005 Paris, France 12Present address: Blood Systems Research Institute, University of California, San Francisco, San Francisco, CA 94118, USA *Correspondence: [email protected] (H.B.), [email protected] (F.A.R.) DOI 10.1016/j.str.2012.01.001 2Unite ´

SUMMARY

The dengue virus (DENV) complex is composed of four distinct but serologically related flaviviruses, which together cause the present-day most important emerging viral disease. Although DENV infection induces lifelong immunity against viruses of the same serotype, the antibodies raised appear to contribute to severe disease in cases of heterotypic infections. Understanding the mechanisms of DENV neutralization by antibodies is, therefore, crucial for the design of vaccines that simultaneously protect against all four viruses. Here, we report a comparative, high-resolution crystallographic analysis of an ‘‘A-strand’’ murine monoclonal antibody, Mab 4E11, in complex with its target domain of the envelope protein from the four DENVs. Mab 4E11 is capable of neutralizing all four serotypes, and our study reveals the determinants of this cross-reactivity. The structures also highlight the mechanism by which A-strand Mabs disrupt the architecture of the mature virion, inducing premature fusion loop exposure and concomitant particle inactivation.

INTRODUCTION The four serotypes of dengue virus (DENV) constitute the largest vector-borne viral disease burden in the developing world (Monath, 1994), with 50–100 million cases reported yearly. There

is neither approved vaccine nor specific therapy to fight these infections. Dengue fever (DF), the most common form of dengue disease, is an acute febrile illness and is generally nonlethal. However, about 1% of DF cases evolve into severe forms, such as life-threatening dengue hemorrhagic fever (DHF) or dengue shock syndrome (DSS) (Mackenzie et al., 2004). Infection by DENV raises lifelong immunity against the infecting serotype (Sabin, 1952), but only transient protection against the other serotypes. In the long term, subsequent infections by viruses from a different DENV serotype are associated with a greater risk for DHF/DSS (Sangkawibha et al., 1984). A likely explanation is that weakly neutralizing, serotype cross-reactive antibodies present in such patients can mediate viral uptake by—and enhanced infection of—cells bearing Fcg receptors (Halstead and O’Rourke, 1977), a phenomenon termed antibody-dependent enhancement (ADE) of infection. DENV is a member of the flavivirus genus of the Flaviviridae family, which includes several other important human pathogens such as Japanese encephalitis (JE), West Nile (WN), yellow fever (YF), and tick-borne encephalitis (TBE) viruses (Lindenbach et al., 2007). The virus particles are enveloped by a lipid bilayer that is enclosed within an icosahedral scaffold formed by the envelope glycoprotein E (Kuhn et al., 2002) (Figure 1). The virions enter cells by receptor-mediated endocytosis followed by low pH-induced fusion of the viral and endosomal membranes (van der Schaar et al., 2008). The E glycoprotein is both responsible for receptor binding and for inducing membrane fusion in the endosome, and is thus the main player in the flavivirus entry process. Flavivirus E is about 500 amino acids long and is anchored in the viral membrane by two C-terminal transmembrane (TM) helices. A soluble fragment of E, containing roughly the 400 N-terminal amino acids of E, has been crystallized

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 303

Structure Determinants of Dengue Virus Cross-Neutralization

Figure 1. Overview of the DENV Envelope Protein and Its Complexes with scFv 4E11 (A) Crystal structure of the prefusion DENV-2 sE dimer (PDB accession number 1OKE). One E subunit is highlighted colored by domains (DI, red; DII, yellow; DIII, blue). The fusion loop (residues 100–108) is colored green. The arrows indicate the view zoomed in (B). (B) Close-up showing DIII from one subunit in the sE dimer and its orientation with respect to the viral membrane in the mature virion. Note the proximity of the fusion loop of the partner subunit (in the forefront, yellow and green). (C) The icosahedral glycoprotein shell of the mature DENV-2 virion formed by 90 E dimers, docked into a 9.5 A˚ resolution cryo-EM reconstruction (PDB accession number 1THD). Each E subunit is colored according to domains, as in (A). View down an icosahedral 3-fold axis (central gray triangle). Icosahedral 5- and 2-fold axes of the particle are marked with pentagons and ovals, respectively. (D) Amino acid sequence alignment of DIII from the four DENV strains used in this study (listed in Table 1). Residue numbers correspond to the full-length DENV-1 E protein (the numbering for DENV-3 DIII is shifted by two residues). Strictly and semiconserved residues are highlighted in red and yellow, respectively. b Strands are labeled above the sequence with blue arrows. Epitope residues are marked with asterisks. (E) Overview of the structure of DENV-1 DIII (dark blue) in complex with scFv 4E11. The heavy-/light-chain framework regions are dark/light gray, respectively, with the CDR loops color coded and labeled H1-H3 and L1-L3. (F and G) Superposition of the scFvs (F) and DIIIs (G) from the four complexes. The scFv 4E11 molecules were superposed using the framework region residues. Residues at the antibody-antigen interface are labeled according to the Kabat scheme (Kabat et al., 1991) and shown in ball and stick (oxygen, red; nitrogen, dark

304 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved

Structure Determinants of Dengue Virus Cross-Neutralization

(Modis et al., 2003, 2005; Rey et al., 1995; Zhang et al., 2004). Its structure features three distinct domains (Figure 1A). DI, which contains the N terminus, is at the center, with DII and DIII at either side. DII displays a hydrophobic fusion loop—the amino acid sequence of which is conserved across flaviviruses—at its distal-most tip. The fusion loop inserts into the endosomal membrane during an acid-pH-driven membrane-fusogenic conformational change. This process is necessary to release the viral nucleic acid within the cytoplasm to initiate infection. DIII, the C-terminal portion of the ectodomain, is an eightstranded b sandwich (strands labeled A–G, Figure 1B) with an immunoglobulin superfamily fold. DIII is thought to be involved in receptor binding (Crill and Roehrig, 2001), although the nature of the relevant receptor(s) at the surface of susceptible cells is not known despite intensive studies (Lindenbach et al., 2007). Infectious dengue virions derive from proteolytic activation of noninfectious immature particles, the surface of which is composed of heterodimers of the viral glycoproteins prM and E (Lindenbach et al., 2007). Cleavage by furin or other proteases eliminates the N-terminal half of prM (the ‘‘pr’’ domain’’). This maturation step is often incomplete, resulting in a significant fraction of only partially mature virions released from infected cells (Junjhon et al., 2010). These virons are infectious and have been shown to display both mature- and immature-like patches on the same virion (Plevka et al., 2011), displaying antigenic characteristics of both types (Nelson et al., 2008). The human humoral immune response to DENV is dominated by cross-reactive antibodies, many of which are non-neutralizing and target prM (Crill et al., 2009; Dejnirattisai et al., 2010; Lai et al., 2008; Oliphant et al., 2007). The resulting antigenic heterogeneity of DENV appears to increase the potential for ADE (Cherrier et al., 2009). Identifying the epitopes on the E protein that are targeted by potent, broadly neutralizing antibodies, ideally neutralizing simultaneously all four dengue serotypes, is therefore important for vaccine design. A recent study on memory B cells recovered from DENVinfected individuals revealed the presence of clones encoding Mabs binding to DIII, which potently neutralize multiple serotypes (Beltramello et al., 2010). Cross-neutralizing Mabs targeting DIII are also a well-characterized component of the murine immune response to DENV. Biochemical studies showed their epitopes to be clustered on the DIII A strand (Gromowski et al., 2008, 2010; Lisova et al., 2007; Matsui et al., 2009; Shrestha et al., 2010; Sukupolvi-Petty et al., 2007; Thullier et al., 2001). These ‘‘A-strand’’ Mabs are believed to neutralize mainly by interfering with attachment to host cells (Crill and Roehrig, 2001; Thullier et al., 2001). The crystal structure of the Fab fragment of an A-strand antibody, Mab 1A1D-2, has been determined in complex with recombinant DIII to 3 A˚ resolution, and the structure was docked into a low-resolution cryo-EM reconstruction of the DENV-2 particle in complex with the same Fab (Lok et al., 2008). This study indicated that binding of 1A1D-2 disrupts the organization of the particle. Binding of Fab 1A1D-2

to the mature DENV-2 virion was shown to be temperature sensitive, reflecting the partially occluded nature of the epitope, and its transient exposure through a ‘‘breathing’’ motion of the particle. Mab 1A1D-2 was raised against DENV-2 strain NGC and neutralizes DENV-2, DENV-1, and DENV-3 (in order of decreasing efficacy) but does not neutralize DENV-4 (Roehrig et al., 1998). To our knowledge, the determinants of its crossreactivity have, however, not been investigated. We approach here this issue by analyzing the murine Mab 4E11 (Megret et al., 1992), which also binds to the DIII A strand (Lisova et al., 2007) and neutralizes all four DENV serotypes (Thullier et al., 1999). We report the crystal structures of a recombinant single-chain variable fragment (scFv) of 4E11 in complex with DIII from all four serotypes to high resolution (1.6–2.1 A˚). Our study reveals the basis for serotype-dependent variations in binding affinity and neutralization efficacy, and provides insights into the immunodominance of the DIII A strand in mice. In particular, our study, together with the studies on Mab 1A1D-2 mentioned above, supports a neutralization mechanism for A-strand antibodies that explains how 4E11 can neutralize all four serotypes with IC50 varying between 1 and 100 nM, despite about more than five orders of magnitude difference in Fab/DIII binding affinities across serotypes. RESULTS AND DISCUSSION Characterization of Mab 4E11 Competition ELISA experiments were carried out to determine the affinity, at equilibrium and in solution, between recombinant Fab 4E11 and isolated DIII from each of the clinical isolates of DENV listed in Table 1, spanning all four serotypes. The results show that Fab 4E11 binds with nanomolar affinity to DIII from DENV-1, DENV-2, and DENV-3 (in this order), and with only micromolar affinity to DENV-4 DIII. The ratio of affinities, in the order DENV-1:DENV-2, DENV-2:DENV-3, and DENV-3:DENV-4, is 1:5, 1:20, and 1:500, respectively, resulting in a 1:50,000 ratio between DENV-1 and DENV-4. Table 1 also reports the Mab 4E11 concentration corresponding to 50% neutralization (IC50) of each of the corresponding DENV isolates for infection of Huh7.5 cells (see Experimental Procedures). The trend in IC50 values of the full-length, bivalent Mab roughly reflects the affinity trend between the monovalent Fab and isolated DIIIs from each of the isolates but varies over a much narrower range (1–100 nM), with IC50 ratios of 1:1, 1:50, and 1:2 among DENV-1: DENV-2, DENV-2:DENV-3, and DENV-3:DENV-4, respectively. As expected, neutralization by Mab 4E11 was weakest with DENV-4, yet it was specific because an isotype-matched Mab control at the same concentration had no effect (data not shown). The small difference in neutralization IC50 between DENV-3 and DENV-4 was unexpected, given the 500-fold ratio in affinities between the corresponding recombinant DIII and Fab 4E11.

blue) with carbon atoms following the main-chain color scheme. The relevant DIII b strands A, B, and G, as well as the D-E loop, are labeled. The AG patch (see text) is highlighted in yellow. The views correspond approximately to an ‘‘open book’’ view of the complexes (see also Figure S1), with (G) corresponding roughly to the orientation in (B). See also Table S1.

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 305

Structure Determinants of Dengue Virus Cross-Neutralization

Table 1. Binding and Neutralization Data Fab:DIII Binding Affinity Serotype

Strain

GenBank Accession No.

KD (nM)a

DG (kcal/mol)

DENV-1

Guiana/FGA89/1989

Q9II01

0.082 ± 0.005

13.76 ± 0.04

1.1 (0.88–1.3)

DENV-2

Jamaica/1409/1983

P07564

0.43 ± 0.04

12.79 ± 0.05

0.85 (0.59–1.2)

0.13

DENV-3

Thailand/PaH881/1988

Q98UK6

8.0 ± 1.0

11.07 ± 0.07

54 (26–110)

8.0

DENV-4

Burma/63632/1976

Unpublished

7.36 ± 0.1

100 (65–150)

15.0

4,100 ± 700

IC50 (nM)b

IC50 (mg/ml)b 0.16

KD values between the recombinant DIII and 4E11 Fab fragment at equilibrium in solution at 25
C, measured by competition ELISA. b Concentration of Mab 4E11 corresponding to 50% neutralization of infection as measured by flow cytometry on huh7.5 cells. Values in parentheses correspond to the 95% confidence range. a

Structure Determination For crystallization we used constructs directing periplasmic secretion in E. coli (see Experimental Procedures) of scFv 4E11 and the recombinant DIII from each of the DENV isolates characterized above (Figure 1D). The crystallization and structure determination of all four complexes are described in the Experimental Procedures, and the resulting statistics are listed in Table 2. The crystals of the DENV-1 DIII/scFv complex diffracted better than the others (1.6 A˚ resolution versus 2.0–2.1 A˚, Table 2). We did not obtain enough DIII from the DENV-3 isolate from the E. Coli construct, and therefore, we made cocrystallization trials using

the dimeric DENV-3 sE ectodomain (residues 1–393, or DI-DIIDIII) produced in Drosophila S2 cells. Unexpectedly, the resulting crystals contained only the scFv in complex with DENV-3 DIII, lacking the rest of the ectodomain. SDS-PAGE analysis of the material used for crystallization confirmed the presence, in addition to the scFv, of a protein at the expected molecular weight for DIII (Figures S1A and S1B available online), indicating that DIII had been cleaved from the sE dimer, presumably by a protease contaminant. Such cleavage was not observed in samples of sE not mixed with the scFv preparation, even when left for a long time at 4
C. Indeed, the DENV-3 sE dimer was stable and led

Table 2. Crystallographic Data Processing and Structure Refinement DENV 1

DENV 2

DENV 3

DENV 4

Space group

P3121

P41212

P21

C2221

Copies of complex per asymmetric unit

1

1

2

1

110.71, 110.71, 58.32

60.46, 60.46, 207.46

54.06, 74.67, 86.89

106.60, 155.88, 55.91

90, 90, 120

90, 90, 90

90, 104.49, 90

90, 90, 90

Crystal Features

Unit cell parameters a, b, c (A˚) a, b, g (
) Data Quality Resolution (A˚)a

47.95–1.60 (1.69–1.60)

45.50–2.10 (2.21–2.10)

42.86–2.04 (2.15–2.04)

47.19–2.00 (2.11–2.00)

Number of observations

150,734

186,284

143,216

211,889

Number of unique reflections

54,205

22,893

40,602

31,933

Completeness (%)

99.6 (99.5)

98.3

94.8 (76.0)

99.8 (98.7)

Redundancy

2.8 (2.7)

8.1 (8.2)

3.5 (2.5)

6.6 (6.7)

I/s (I)

13.9 (1.9)

23.5 (3.3)

10.8 (2.0)

18.5 (4.2)

Rsym (%)a,b

4.5 (50.7)

4.6 (56.4)

8.4 (44.4)

6.8 (44.1)

Model Quality Rwork/Rfree (%)c

17.7/18.8

20.6/21.4

21.8/23.4

18.1/20.8

Number of protein atoms

2,561

2653

4,860

2,505

Number of solvent atoms (of which water)

354 (304)

115 (112)

342 (277)

326 (277)

Average atomic B factor for main chains/ side chains and waters (A˚2) Rmsd in bond lengths/angles (A˚,
)

30.8/38.6

57.4/64.4

33.2/41.0

34.1/43.8

0.007/0.95

0.007/0.97

0.007/0.96

0.007/0.99

Ramachandran plot Favored (%)

98.1

96.6

97.5

97.5

Allowed (%)

1.6

3.1

2.5

2.5

Disallowed (%)

0.3

0.3

0.0

0.0

a

Values in parentheses correspond to the highest resolution shell. P P P P b Rsym = 100 3 [ h ijIi(h)  < I(h) > j]/ h iIi(h), where < I(h) > is the mean of all observations {Ii(h)} of reflection h. P P c R = hjFo(h)  Fc(h)j/ h Fo(h), where Fo(h) and Fc(h) are the observed and calculated structure factor amplitudes of reflection h.

306 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved

Structure Determinants of Dengue Virus Cross-Neutralization

Figure 2. The Common Core of the 4E11 Epitopes (A) View down the AG patch, with the molecules colored as in Figure 1E. (B) Same as (A) but rotated by 90
about a vertical axis in the plane of the paper. The boxes in both top panels indicate the regions depicted in more detail in the lower panels, in which the main chain is in a ‘‘worm’’ representation. Shown are the coordinates of the complex for DENV-1. Hydrogen bonds and salt bridges are represented as black dashes (see also Figure S2.).

to crystals of the intact dimeric ectodomain in the absence of scFv, the structure of which was determined in parallel (unpublished data). Furthermore, the DENV-3 sE dimer was resistant to limited proteolysis in the absence of the antibody, indicating that binding of antibody to DIII destabilizes the dimer, likely by leading to dimer dissociation (see below) and exposing a segment of polypeptide N-terminal to DIII (in the DI-DIII linker), which becomes accessible to contaminating proteases. Overall Structure of the Immune Complexes The complexes of scFv 4E11 with DIII from each of the four DENV isolates have, as expected, a very similar overall structure. The 4E11 variable domains are oriented with their complementarity-determining regions (CDRs) facing the A and G b strands, which make the edge of the DIII b sandwich. The VL domain lies above the G strand, and the VH domain rides atop the A strand (Figure 1E). The residues in both paratope and epitope have very similar conformations in all four complexes (Figures 1F and 1G), showing that 4E11 cross-reactivity does not involve a significant induced fit. There are, however, differences in scFv positioning between the complexes, especially for serotypes 3 and 4 when compared to the DENV-1 complex (Figure S1C). The DENV-1, DENV-2, and DENV-4 DIIIs bury 900–950 A˚2 of accessible surface area on binding the scFv, with surface complementarity coefficients in the range 0.7–0.75. The apparent buried area is smaller in the complex with DENV-3 DIII (780 A˚2)

because DIII residues 362–364 (DE loop) are disordered (and thus absent from the calculation), whereas these residues are see contributing to the interface in the other three complexes. The Mab 4E11 variable domains of both the heavy and light chains share 85% amino acid sequence identity with the counterparts from Mab 1A1D-2. Accordingly, the complexes of DENV-2 DIII with scFv 4E11 and Fab 1A1D-2 (PDB accession number 2R29; Lok et al., 2008) can be superposed with a root-mean-square deviation (rmsd) of 1.2 A˚ over 330 Ca atoms, confirming that the epitopes of these two antibodies are superimposable. Determinants of 4E11 Binding to and Affinity for the Four DENV Serotypes The list of all the intermolecular contacts in the structures of the immune complexes of each serotype is provided in Tables S1A and S1B. The four different 4E11 epitopes analyzed here share a common core, located around a hydrophobic patch between the A and G b strands, which we term the ‘‘AG patch’’ (Figures 1G and 2). This patch is part of the hydrophobic core of DIII and consists of the side chains of residues 308, 312 (strand A) and 387, 389, and 391 (strand G). These side chains are directed toward the internal hydrophobic core of the DIII b sandwich but are accessible because the A and G strands form the edge of the b sandwich (highlighted in yellow in Figure 1G). Residues Leu387 and Trp391 are strictly conserved between DENV serotypes

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 307

Structure Determinants of Dengue Virus Cross-Neutralization

Figure 3. Serotype-Dependent Antibody-Antigen Interactions (A–D) Close-ups of the 4E11 paratope (left panel) and epitope (right panel) in the complexes with DENV serotypes 1–4 as indicated. Views correspond approximately to the boxed areas in the key in the top part of the figure, which contain the serotype-specific residues. Semitransparent surfaces of the scFv and DIII are colored according to their electrostatic potentials, calculated separately for each component in the complex. The surface electrostatic potential is represented in a color gradient from blue (basic) through white (neutral) to red (acidic), between values of ±2kBT (where kB is the Boltzmann constant). Salt bridges between residues in the antibody and the antigen are shown as black dashes in the right-hand panels. (E) The corresponding views of the interface between Fab 1A1D-2 and the DENV-2 DIII (PDB accession number 2R29).

(Figure 1D), whereas the other locations display conservative substitutions that preserve the short, aliphatic character of the side chains in the AG patch, as well as its shape (defined by the abutting A and G strands). Indeed, the short aliphatic side chains in the AG patch are responsible for the specificity of Mab 4E11 for the DENV group, especially at position 389, which in other flaviviruses is either His or Tyr. The available structures of DIII from TBE (PDB number 1SVB), WN (1ZTX), JE (1PJW), YF (2JV6), and Langat (2GG1) viruses show that the bulky side chains at position 389 protrude from the patch in a way that would prohibit 4E11 binding via steric clashes with TyrL28 (Figures 2A and S2). Strand A of DIII presents a prominent central bulge (conserved in all flaviviruses), located to one side of the AG patch, featuring the conserved residues Lys310 and Glu311 (Figures 1D and 1G). The conserved ‘‘core’’ epitope is recognized by Mab 4E11 in the same way in all four serotypes, as depicted in Figure 2: packing against the hydrophobic side chains of the AG patch, recognition of its shape via hydrogen bonding to the A- and G-strand main-chain groups defining the periphery, and making salt bridges with Lys310 and Glu311 from the central bulge. A number of the core epitope residues identified here were previously shown, through biochemical studies, to make important contributions to the affinity of Fab 4E11 for the DENV-1 DIII. In particular, Lys310 was found critical for binding (Lisova et al., 2007). Likewise, TrpH96 and GluH97 in the CDR H3 loop, which interact with the AG patch and A-strand main-chain groups at the center of the antibody-antigen interface, were also shown to be essential for Fab 4E11 binding to DENV-1 DIII (Bedouelle et al., 2006). Serotype-Specific Contacts The contacts of the antibody with residues in the 4E11 epitope that are not conserved across DENV serotypes also contribute to binding. These E residues are located on the upper portion of DIII, which is exposed in the virion (Figure 3). The principal antibody residues involved in sensing these nonconserved E residues are ArgH94 (third heavy-chain framework region) and GluL55 (CDR L2 loop). Additionally, residues AspH31-ThrH32, TyrH102, and AsnL53 (Figure 1F) engage in a small number of interactions with DIII residues at the periphery of the antibodyantigen interfaces that vary between serotypes (Tables S1A and S1B). ArgH94 and GluL55 create adjacent patches of positive and negative electrostatic potential in the paratope. In the complex with DENV-1 DIII (the serotype against which Mab 4E11 was raised), these residues of the paratope form salt bridges with DIII serotype-specific residues Glu309 and Lys307, respectively (Figures 1D and 3A). Binding to DIII from the other serotypes is facilitated by conformational variability of the ArgH94 side chain and changes in overall antibody positioning, allowing the partial conservation of charge complementarity despite amino acid sequence variations between serotypes over this region (Figures 1D and 3B–3D). This is particularly striking for the complex with DIII from the DENV-3 isolate. Here, the antibody shifts down such that GluL55 and ArgH94 face the DIII B strand (which is antiparallel to strand A; see Figures 1B and 1G) and interact with Lys327 and Glu325, mimicking the interactions with the DENV-1 A-strand residues Lys307 and Glu309 (Figure 3C). Finally, the absence of a basic residue at positions 307 or 327 in DENV-4 DIII results in

308 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved

Structure Determinants of Dengue Virus Cross-Neutralization

a much lower degree of charge complementarity across the antibody-antigen interface than in the complexes with the other serotypes (Figure 3D), reflecting the much lower affinity of Fab 4E11 for DENV-4 DIII (Table 1). Previous mutagenesis studies on the DENV-1 DIII A and G strands showed that mutation of residue Lys307 to alanine reduced the free energy of dissociation (DGd) of Fab 4E11 from 13.8 to 9.7 kcal/mol (Lisova et al., 2007), which was the second-greatest reduction in affinity after mutation of the conserved Lys310. This suggests that the effect of truncating the Lys307 side chain in the DENV-1 DIII accounts for the majority of the difference in DGd of Fab 4E11 binding to DENV-1 and DENV-4 DIIIs (Table 1). Likewise, the DGd value of 12.8 kcal/mol for the DENV-2 DIII (Table 1), compared with 12.5 kcal/mol for the DENV-1 DIII Glu309Ala mutant (Lisova et al., 2007), shows that elimination of the negative charge at position 309 appears to account for the affinity difference of Fab 4E11 for the DENV-2 DIII versus that of DENV-1. Residue changes at positions 309 and 307 are thus the main determinants of the affinities of Fab 4E11 for the DENV-2 and DENV-4 DIIIs compared with that of DENV-1. The ability of the B-strand residues Glu325 and Lys327 of DENV-3 to mimic these interactions, despite having unfavorable A-strand residues at these locations, is likely to be crucial in maintaining high-affinity binding to this serotype. Previous biochemical studies have shown that cross-reactive murine Mabs binding to A-strand residues on the DENV-2 DIII commonly have epitope determinants on the B strand of the DENV-3 DIII (Gromowski et al., 2008; Matsui et al., 2009). These observations are a strong indication that the strand-switching mechanism of 4E11 cross-reactivity with DENV-3 that we describe here is a general feature of all A-strand Mabs. The contacts observed here between scFv 4E11 and the DENV-2 DIII are mostly conserved in the corresponding complex of Fab 1A1D-2. However, an important difference is that 1A1D-2 has an aspartic acid at position H95 in its CDR H3 loop that forms an additional salt bridge with Lys307, whereas the corresponding residue in 4E11 is glycine. As a result, the 1A1D-2 paratope is significantly more negatively charged than that of 4E11 (Figure 3E). This presumably reflects the fact that 4E11 and 1A1D-2 were raised against DENV-1 and DENV-2 strains, respectively, which have different charge distributions along the A strand (because the residue at position 309 is glutamate in DENV-1 and valine in DENV-2). As an additional consequence, the 1A1D-2 paratope displays an even lower charge complementarity with DENV-4 DIII than that of 4E11 (Figures 3D and 3E). These observations explain the differing serotype dependence of neutralization between these two antibodies, and in particular why Mab 4E11 neutralizes DENV-4 strains at moderate antibody titers, whereas Mab 1A1D-2 does not. Molecular Basis for Immunodominance of the DENV DIII A Strand in Mice Analysis of the murine germline shows that the 4E11 and 1A1D-2 variable domains are both derived from the same heavy- and light-chain variable (V) and junction (J) gene segments, but different diversity (D) segments (see Experimental Procedures). In both antibodies, almost all residues that contact DIII are present in the germline sequences of the relevant gene segments

(Figures 4A and 4B). In both cases, VDJ recombination has resulted in CDR H3 loops with identical lengths (six residues), with a glutamate at H97 and an aromatic side chain at position H96 (Trp in 4E11 and Tyr in 1A1D-2) that plug the center of the antibody-antigen interface (Figures 2, 4C,and 4D). The only significant difference is the residue at position H95 (glycine in 4E11 and aspartic acid in 1A1D-2), which dictates the differing serotype responses of these antibodies as discussed above. Of the 14 known murine D segments, 6 would be capable of forming 1A1D-2/4E11-like H3 loops (Figure 4E). This, in conjunction with the inherent capacity of the V gene segments utilized by 4E11 and 1A1D-2 to form high-affinity interactions with DIII residues that are conserved among all four serotypes, may explain the immunodominance of A-strand epitopes in mice. Amino acid sequence variations at position H95, arising from N regions or somatic hypermutations, could then generate a family of cross-reactive Mabs with distinct avidity and neutralization profiles by serotype. In summary this analysis shows that the murine germline harbors gene segments with an inherent capacity for high-affinity binding to DIII from all four DENV serotypes. This is in line with the observation that most strongly neutralizing murine Mabs against DENV bind to DIII, whereas humans do not appear to make such antibodies as readily (Crill et al., 2009; Oliphant et al., 2007; Wahala et al., 2009). However, the memory B cell pool of DENV-infected humans does contain clones that encode anti-DIII antibodies capable of neutralizing multiple DENV serotypes (Beltramello et al., 2010), and a recent experimentally validated computational docking study indicates that these antibodies bind to A-strand epitopes that include residue Lys310 (Simonelli et al., 2010). Insights into the Neutralization Mechanism of Murine A-Strand Mabs The crystal structures of the prefusion DENV sE dimers (Modis et al., 2003, 2005; Zhang et al., 2004) show that the DIII A strand forms one side of the pocket that accommodates residue Trp101 from the fusion loop of the partner subunit in the dimer. The roof of this pocket is formed by the side chain of DIII residue Lys310, which stacks against the side chain of Trp101 of the opposite subunit (Figure 5A). Lys310 is thus involved in intradimer contacts and is also a critical epitope residue for Mabs 4E11, 1A1D-2, 9D12, and every other known A-strand Mab for which mutagenesis studies of the antigen have been made (Gromowski et al., 2008, 2010; Lisova et al., 2007; Matsui et al., 2009; Shrestha et al., 2010; Sukupolvi-Petty et al., 2007; Thullier et al., 2001). Not surprisingly, docking our atomic model for the DIII/4E11 complex onto the structures of the corresponding sE dimers shows that there is a partial clash between the 4E11 CDR H2 loop and the apposed fusion loop in the dimer, as illustrated in Figure 5B. This incompatibility between dimer contacts and binding of A-strand antibodies is in keeping with our observations that 4E11 scFv led to DENV-3 sE dimer instability and cleavage at the DI-DIII linker. This cleavage is likely to follow dimer dissociation because experiments of limited proteolysis showed that the sE dimer is relatively resistant to degradation, as mentioned above. Because the E dimers are the building blocks of the mature flavivirus particles, dimer dissociation upon binding of A-strand antibodies will induce a major rearrangement of the virion

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 309

Structure Determinants of Dengue Virus Cross-Neutralization

Figure 4. Immunodominance of the DIII A Strand (A and B) Genetic origin of the residues in the 4E11 (A) and 1A1D-2 (B) variable domains (heavy-chain and light-chain colored black and white, respectively) that contact DIII in their respective complexes. The side chains contacting DIII are shown as sticks with carbon atoms colored yellow and magenta for conserved and nonconserved residues within the murine germline, respectively. (C and D) Models of VDJ recombination giving rise to the H3 loops of 4E11 (C) and 1A1D-2 (D). In (C) the nucleotide sequences of the 30 and 50 ends of the VH and JH gene segments at either side of the junction, and the entire D segment, are aligned against the 4E11 nucleotide sequence (Bedouelle et al., 2006). Nucleotides conserved with 4E11 are written in bold capital letters. N/SH shows nucleotides added by the terminal deoxynucleotidyl transferase during VDJ joining, or generated by subsequent somatic hypermutations—these are written in lowercase bold letters at the appropriate positions along the sequence. The amino acid sequence of the 4E11 H3 loop is shown in single-letter code at the bottom. Residues arising from N regions or somatic hypermutations are underlined, and are additionally written in magenta if they contact DIII. In (D), because the nucleotide sequence of Mab 1A1D-2 is not published, our analysis of VDJ recombination is based on amino acid sequence comparisons, and the nucleotide sequence shown in the figure is hypothetical. Several VDJ recombination schemes are possible as a result of the degeneracy within the murine SP2 family of D segments (see E)—shown here is one example. In both (C) and (D), the TGGGA and TA(T/C)GA nucleotide sequences from the D segments that give rise to the Trp/Tyr-Glu sequences in the 4E11 and 1A1D-2 H3 loops (in conjunction with the addition of a G nucleotide at the 30 end) are written in green. (E) Nucleotide sequences of the 14 known murine D segments (taken from the IMGT website). The TGGGA and TA(T/C)GA sequences are written in bold green uppercase letters.

surface, as shown earlier by cryo-EM studies on Fab 1A1D-2 bound to the DENV-2 virion (Lok et al., 2008). This is also in line with binding competition assays with DENV virions showing that prior binding of the A-strand Mab 9D12 (Sukupolvi-Petty et al., 2007) promotes the binding of antibodies against the fusion loop (Henchal et al., 1985). Indeed, the epitopes of the latter antibodies are poorly accessible on mature virions because they are partially buried at the dimer interface (Stiasny et al., 2006). These observations suggest that disruption of the mature virion architecture and premature exposure of the fusion loop is a general feature of antibodies binding to the A-strand epitope. Indeed, binding of A-strand Mabs to mature particles was proposed to disrupt the virion architecture by capturing thermally deformed states of the glycoprotein shell and preventing their relaxation back to the equilibrium ‘‘herringbone’’ arrangement of E dimers of the unbound virus (Lok et al.,

2008). Disruption of the mature virion architecture is likely to be irreversible, even upon dissociation of the antibody from the virus. If so, then the neutralization potency of A-strand Mabs such as 4E11 would be determined not by their avidities for the virion but rather their rates of association (kon) and the incubation time. Also, as suggested by Lok et al. (2008), the tight packing of E dimers on the virions is such that binding to one of the 180 E subunits may propagate the conformational change in the particle, exposing the fusion loops and inactivating the virions by aggregation via the exposed fusion loops. In such a model, once the thermal motion allowed the binding to one of the 180 E subunits at the virion surface, the particle distortion would propagate to induce disruption of additional dimers, allowing more antibodies to bind. The observation that Fab binding was marginal at 4
C—temperature at which the majority of particles had no bound Fab, but the few that were bound were saturated

310 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved

Structure Determinants of Dengue Virus Cross-Neutralization

Figure 5. Role of 4E11 Epitope Residues in sE Dimer Contacts (A) The side-chain Trp101 (green) at the tip of the fusion loop (in green) of one E subunit in the dimer, which is the building block of the icosahedral glycoprotein shell, is buried at the DI/DIII interface (red/blue) of its dimer mate (PDB accession number 1OKE; Modis et al., 2003). The molecular surface of the 4E11 epitope is semitransparent. Note the location of the Lys101 side chain, stabilizing this intersubunit interaction with Trp101. (B) Docking of scFv 411 (color coded as in the previous figures) from its complex with the DENV-2 DIII onto the crystal structure of the DENV-2 sE dimer. Note the clash between the tip of the H2 loop and the fusion loop of the partner E subunit in the dimer.

with Fab (Lok et al., 2008)—is in support of this hypothesis. The relatively small difference in neutralization IC50 of Mab 4E11 across DENV serotypes when compared to the difference in the respective binding affinity of its Fab fragment for the exposed epitope in isolated DIII also supports this view (Table 1). Recent data have shown that antibodies against flaviviruses—WN virus and DENV—increase their neutralization capacity when left to interact for longer periods of time (Dowd et al., 2011). This suggests that even antibodies that bind poorly (such as 4E11 binding to DENV-4 particles) will lead to inactivation of the virion if left to interact for long enough. Our data, however, suggest a threshold level in affinity from which the antibody jumps from efficient neutralization to relatively poor neutralization. This is exemplified in Table 1 by the fact that, despite having 500-fold higher affinity for DENV-3 than for DENV-4 DIII, 4E11 displays only a 2-fold increase in neutralization efficiency (IC50). In contrast the 20-fold increase in affinity between DENV-2 and DENV-3 DIII and 4E11 leads to a 50-fold increase in neutralization efficiency (i.e., an IC50 50 times lower). These values would suggest that the threshold is likely to lie in the range between 0.5 and 5 nM Fab/DIII binding affinity. However, this estimate does not take into account possible differences in the breathing rate of DENV particles from different strains, which may strongly affect the neutralization kinetics, further illustrating the complexity of this process. Biological Implications Are there implications of the presence of the exposed, crossreactive A-strand epitope for identifying the region of DIII postulated to be responsible for receptor binding? The identification of the DENV receptors is difficult likely because it may involve low-affinity interactions at the cell surface. However, if all four DENV serotypes recognize the same receptor, then it is possible that certain features of the DIII surface are maintained, which make recognition by 4E11 possible, while still being restricted to the DENV group. The A-strand epitope is, therefore, a plausible candidate surface for interaction with receptor. If this is confirmed, a new question is: Does interaction with receptor have the same effect on the particle as do A-strand antibodies? Such an effect would mean that the particle exposes the fusion loop already at the plasma membrane of the infected cell, making the interaction irreversible, and explaining why the initial interactions may require only low affinity. A similar process was sug-

gested for alphaviruses, with the exposure of ‘‘transitional’’ epitopes in glycoprotein E2 described for Sindbis virus, which become accessible only upon incubation of the virion at 37
C with susceptible cells (Meyer and Johnston, 1993). The crystal structure of the alphavirus E2/E1 heterodimer (Voss et al., 2010) showed that the fusion loop of E1, which is homologous to flavivirus E, lies in a groove of E2. The transitional epitope maps to this groove, indicating that exposure of this epitope can only take place upon exposure of the fusion loop at the cell surface, which makes binding to cells irreversible. Understanding whether receptor binding by flaviviruses and alphaviruses is merely a passive event, allowing particle uptake to be disassembled in the endosome upon exposure to low pH, or whether the receptor has a more active role, inducing partial particle disassembly, remains an important issue to be further investigated. EXPERIMENTAL PROCEDURES Cells, Viruses, and Monoclonal Antibodies Mosquito AP61 cells and confluent human hepatocyte (Huh 7.5) cells were prepared as described in the Supplemental Experimental Procedures. The virus strains used throughout this work are listed in Table 1. Virus stocks were grown in AP61 cells, purified from supernatants by ultracentrifugation, and titered as described in the Supplemental Experimental Procedures. Mab 4E11 (IgG2a/k) was produced and purified from hybridoma cells as previously described (Thullier et al., 1999, 2001). Mab 2H2, a murine antibody binding to protein prM of all four dengue serotypes, was affinity purified from hybridoma cells obtained from ATCC (reference HB-114). Neutralization Assays Each virus strain (at 4.5 3 105 pfu/ml for DENV-1, DENV-3, and DENV-4; 2.25 3 106 pfu/ml for DENV-2) was incubated with 10-fold serial dilutions of Mab 4E11 at 37
C for 2 hr in medium (DMEM containing 2% FCS, 1% penicillin, and 1% streptomycin), and added to 80% confluent Huh 7.5 cells at moi values of 0.1 for DENV-1, 5.0 for DENV-2, and 0.5 for DENV-3/DENV-4. Each concentration was tested in duplicate. An isotype-matched control murine Mab was used as well. Infection was allowed to proceed for 2 hr at 37
C before the virus/Mab mixtures were removed and the cells washed. The infected cell cultures were made up to 1 ml vol in medium and left at 37
C for 24 hr. They were transferred to 96-well plates and fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences). Cells were stained using 10 mg/ml Mab 2H2 for 1 hr at 4
C, and a phycoerythrin-conjugated goat anti-mouse secondary antibody (Santa Cruz Biotechnology; RO480) at 1:500 dilution for 30 min at 4
C. Cells were analyzed using a BD FACScalibur instrument. The percentage of infected cells was determined using the FlowJo software package. IC50 values were determined by fitting dose-response curves to these data using the Prism software package.

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 311

Structure Determinants of Dengue Virus Cross-Neutralization

Binding-Affinity Measurements The dissociation constant (KD) values at equilibrium in solution at 25
C between recombinant Fab 4E11 and DIIIs were measured by competition ELISA as described previously (Lisova et al., 2007). Protein Expression, Purification, and Crystallization Recombinant, His-tagged Fab 4E11, scFv 4E11, and DIII from serotypes 1, 2, and 4 were expressed and purified from E. coli periplasmic extracts using nickel-ion affinity chromatography as described previously (Bedouelle et al., 2006; Lisova et al., 2007; Renard et al., 2003). Protein expression was optimized in microfermentors and scaled up to a 4 liter fermentor in high-density medium (Frachon et al., 2006) at the Protein Production Facility of Institut Pasteur. The protein samples were further purified by size-exclusion chromatography using a Superdex 75 10/30 column (GE Healthcare) in 150 mM NaCl and 20 mM Tris-HCl (pH 8.0). Pure fractions were pooled and concentrated to 15 mg/ml in 50 mM NaCl and 5 mM Tris-HCl (pH 8.0). The DENV-3 sE ectodomain (strain PaH881/88) was produced in Drosophila cells, and purified by affinity and size-exclusion chromatography (see Supplemental Experimental Procedures). Purified DENV-3 sE was concentrated to 12 mg/ml in 50 mM NaCl and 25 mM Tris-HCl (pH 8.0). Concentrated scFv and DIII preparations were mixed to give a 1:1 stoichiometry, except for serotype 4 (1:1.5), and total protein concentrations of 14–16 mg/ml. Concentrated scFv and DENV-3 sE were mixed to give a 1:1 stoichiometry and a total protein concentration of 13 mg/ml. In all cases these mixtures were incubated at 4
C overnight and used for crystallization trials without further purification the following day. Crystallization conditions were screened by robot at 19
C in 96-well sitting drop format (0.2 ml of protein with an equal volume of precipitant per drop). The crystals were cryoprotected prior to cryo-cooling in liquid nitrogen. Table S2 details the crystallization and cryo-cooling conditions for structure determination. Structure Determination and Refinement Data were collected at the Swiss Light Source beamline PX1 using the Pilatus 6M detector. Data were processed with XDS (Kabsch, 1988) and scaled and reduced with SCALA (Evans, 1993) and other CCP4 programs (Collaborative Computational Project, Number 4, 1994). The X-ray data were phased by molecular replacement with Phaser (McCoy et al., 2007), using the crystal structure of scFv 1F9 (PDB accession number 1DZB), DIII from previously published DENV-2 and DENV-3 sE crystal structures (1OAN and 1UZG, respectively), and the solution NMR structure of the DENV-4 DIII (2H0P) as search models. Structures were visualized and rebuilt using Coot (Emsley and Cowtan, 2004) and refined with BUSTER (Bricogne, 1993) with the heavy chain, light chain, and DIII defined as individual TLS groups. Water molecules were added automatically during the later stages of refinement and verified manually. Structure Analysis Structural superpositions, buried surface area and surface complementarity coefficient calculations, and intermolecular contacts were calculated using LSQKAB, AREAIMOL, SC, and NCONT from the CCP4 suite. Figures were prepared in PyMOL (http://pymol.sourceforge.net). Electrostatic potentials were calculated using the APBS suite (Baker et al., 2001). Structures were validated using the MolProbity web server (http://molprobity.biochem.duke.edu/). Sequence Analysis of Mabs 4E11 and 1A1D-2 The murine germline gene segments from which Mab 4E11 derives have already been reported (Bedouelle et al., 2006). The VH and VL germline segments of Mab 1A1D-2 were identified with the IMGT/DomainGapAlign tool of the IMGT website (http://www.imgt.org). Its JH and JL germline segments were identified by aligning its sequence with the translations of the 10 IGKJ and 8 IGHJ existing germline segments. Finally, its D region, which contains at most four residues (DYEG), was compared with the translations of the existing IGHD alleles. ACCESSION NUMBERS The coordinates and structure factors of the 4E11/DIII complexes for DENV-1, DENV-2, DENV-3, and DENV-4 were deposited in the PDB with accession numbers 3UZQ, 3UZV, 3UZE, and 3UYP, respectively.

SUPPLEMENTAL INFORMATION Supplemental Information includes two figures, two tables, and Supplemental Experimental Procedures and can be found with this article online at doi:10.1016/j.str.2012.01.001. ACKNOWLEDGMENTS The project was initiated and realized thanks to a grant from the Pediatric Dengue Vaccine Initiative (PDVI) to F.A.R. and F.A.S. We thank Scott Halstead and Susie Kliks from the PDVI for support; Genevie`ve Milon and Arnaud Blondel for support through a ‘‘Programme Transversal de Recherche’’ (PTR) from I. Pasteur; Philippe Despre`s for providing the cDNAs for the DIII constructs; Jacques Bellalou, Ahmed Haouz, and Patrick Weber from the protein production and crystallization facilities for their help; Elodie Brient-Litzler and Olesia Lisova for technical assistance; and the staff of the PX1 beamline at the Swiss Light Source for beamline support. F.A.R. also acknowledges funding by Merck SERONO and by an ANR grant (‘‘DENtry’’). J.J.B.C. was supported by an EMBO Long-Term Fellowship (ALTF-194-2005), a Marie Curie IntraEuropean Fellowship (EIF-25456-DENLIG), and a Pasteur PTR fellowship. H.B. was funded by the French Ministry of Defense (DGA Nos 04.34.025 and 01.34.062) and the European Union (FP6-2003-INCO-DEV2 No 517711 [DENFRAME]). The authors declare no competing interests. Received: October 31, 2011 Revised: December 9, 2011 Accepted: January 2, 2012 Published online: January 26, 2012 REFERENCES Baker, N.A., Sept, D., Joseph, S., Holst, M.J., and McCammon, J.A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl. Acad. Sci. USA 98, 10037–10041. Bedouelle, H., Belkadi, L., England, P., Guijarro, J.I., Lisova, O., Urvoas, A., Delepierre, M., and Thullier, P. (2006). Diversity and junction residues as hotspots of binding energy in an antibody neutralizing the dengue virus. FEBS J. 273, 34–46. Beltramello, M., Williams, K.L., Simmons, C.P., Macagno, A., Simonelli, L., Quyen, N.T., Sukupolvi-Petty, S., Navarro-Sanchez, E., Young, P.R., de Silva, A.M., et al. (2010). The human immune response to Dengue virus is dominated by highly cross-reactive antibodies endowed with neutralizing and enhancing activity. Cell Host Microbe 8, 271–283. Bricogne, G. (1993). Direct phase determination by entropy maximization and likelihood ranking: status report and perspectives. Acta Crystallogr. D Biol. Crystallogr. 49, 37–60. Cherrier, M.V., Kaufmann, B., Nybakken, G.E., Lok, S.M., Warren, J.T., Chen, B.R., Nelson, C.A., Kostyuchenko, V.A., Holdaway, H.A., Chipman, P.R., et al. (2009). Structural basis for the preferential recognition of immature flaviviruses by a fusion-loop antibody. EMBO J. 28, 3269–3276. Collaborative Computational Project, Number 4. (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763. Crill, W.D., and Roehrig, J.T. (2001). Monoclonal antibodies that bind to domain III of dengue virus E glycoprotein are the most efficient blockers of virus adsorption to Vero cells. J. Virol. 75, 7769–7773. Crill, W.D., Hughes, H.R., Delorey, M.J., and Chang, G.J. (2009). Humoral immune responses of dengue fever patients using epitope-specific serotype-2 virus-like particle antigens. PLoS One 4, e4991. Dejnirattisai, W., Jumnainsong, A., Onsirisakul, N., Fitton, P., Vasanawathana, S., Limpitikul, W., Puttikhunt, C., Edwards, C., Duangchinda, T., Supasa, S., et al. (2010). Cross-reacting antibodies enhance dengue virus infection in humans. Science 328, 745–748. Dowd, K.A., Jost, C.A., Durbin, A.P., Whitehead, S.S., and Pierson, T.C. (2011). A dynamic landscape for antibody binding modulates antibody-mediated neutralization of West Nile virus. PLoS Pathog. 7, e1002111.

312 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved

Structure Determinants of Dengue Virus Cross-Neutralization

Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.

to define linear and discontinuous antigenic sites on the dengue virus envelope glycoprotein. Virology 187, 480–491.

Evans, P.R. (1993). Data collection and processing. L. Sawyer, N. Isaacs, and S. Bailey, eds. Proceedings of the CCP4 Study Weekend, 114–122.

Meyer, W.J., and Johnston, R.E. (1993). Structural rearrangement of infecting Sindbis virions at the cell surface: mapping of newly accessible epitopes. J. Virol. 67, 5117–5125.

Frachon, E., Bondet, V., Munier-Lehmann, H., and Bellalou, J. (2006). Multiple microfermentor battery: a versatile tool for use with automated parallel cultures of microorganisms producing recombinant proteins and for optimization of cultivation protocols. Appl. Environ. Microbiol. 72, 5225–5231.

Modis, Y., Ogata, S., Clements, D., and Harrison, S.C. (2003). A ligand-binding pocket in the dengue virus envelope glycoprotein. Proc. Natl. Acad. Sci. USA 100, 6986–6991.

Gromowski, G.D., Barrett, N.D., and Barrett, A.D. (2008). Characterization of dengue complex-specific neutralizing epitopes on the envelope protein domain III of dengue 2 virus. J. Virol. 82, 8828–8837.

Modis, Y., Ogata, S., Clements, D., and Harrison, S.C. (2005). Variable surface epitopes in the crystal structure of dengue virus type 3 envelope glycoprotein. J. Virol. 79, 1223–1231.

Gromowski, G.D., Roehrig, J.T., Diamond, M.S., Lee, J.C., Pitcher, T.J., and Barrett, A.D. (2010). Mutations of an antibody binding energy hot spot on domain III of the dengue 2 envelope glycoprotein exploited for neutralization escape. Virology 407, 237–246.

Monath, T.P. (1994). Dengue: the risk to developed and developing countries. Proc. Natl. Acad. Sci. USA 91, 2395–2400.

Halstead, S.B., and O’Rourke, E.J. (1977). Dengue viruses and mononuclear phagocytes. I. Infection enhancement by non-neutralizing antibody. J. Exp. Med. 146, 201–217. Henchal, E.A., McCown, J.M., Burke, D.S., Seguin, M.C., and Brandt, W.E. (1985). Epitopic analysis of antigenic determinants on the surface of dengue-2 virions using monoclonal antibodies. Am. J. Trop. Med. Hyg. 34, 162–169. Junjhon, J., Edwards, T.J., Utaipat, U., Bowman, V.D., Holdaway, H.A., Zhang, W., Keelapang, P., Puttikhunt, C., Perera, R., Chipman, P.R., et al. (2010). Influence of pr-M cleavage on the heterogeneity of extracellular dengue virus particles. J. Virol. 84, 8353–8358. Kabat, E.A., Wu, T.T., Perry, H.M., Gottesman, K.S., and Foeller, C. (1991). Sequences of Proteins of Immunological Interest (Washington, DC: US Department of Health and Human Services). Kabsch, W. (1988). Automatic indexing of rotation diffraction patterns. J. Appl. Crystallogr. 21, 67–71. Kuhn, R.J., Zhang, W., Rossmann, M.G., Pletnev, S.V., Corver, J., Lenches, E., Jones, C.T., Mukhopadhyay, S., Chipman, P.R., Strauss, E.G., et al. (2002). Structure of dengue virus: implications for flavivirus organization, maturation, and fusion. Cell 108, 717–725. Lai, C.Y., Tsai, W.Y., Lin, S.R., Kao, C.L., Hu, H.P., King, C.C., Wu, H.C., Chang, G.J., and Wang, W.K. (2008). Antibodies to envelope glycoprotein of dengue virus during the natural course of infection are predominantly crossreactive and recognize epitopes containing highly conserved residues at the fusion loop of domain II. J. Virol. 82, 6631–6643. Lindenbach, B.D., Thiel, H.J., and Rice, C.M. (2007). Flaviviridae: the viruses and their replication. In Fields Virology, Fifth Edition, D.M. Knipe, P.M. Howley, D.E. Griffin, R.A. Lamb, M.A. Martin, B. Roizman, and S.E. Straus, eds. (Philadelphia: Lippincott Williams & Wilkins), pp. 1101–1152. Lisova, O., Hardy, F., Petit, V., and Bedouelle, H. (2007). Mapping to completeness and transplantation of a group-specific, discontinuous, neutralizing epitope in the envelope protein of dengue virus. J. Gen. Virol. 88, 2387–2397. Lok, S.M., Kostyuchenko, V., Nybakken, G.E., Holdaway, H.A., Battisti, A.J., Sukupolvi-Petty, S., Sedlak, D., Fremont, D.H., Chipman, P.R., Roehrig, J.T., et al. (2008). Binding of a neutralizing antibody to dengue virus alters the arrangement of surface glycoproteins. Nat. Struct. Mol. Biol. 15, 312–317. Mackenzie, J.S., Gubler, D.J., and Petersen, L.R. (2004). Emerging flaviviruses: the spread and resurgence of Japanese encephalitis, West Nile and dengue viruses. Nat. Med. 10 (12, Suppl), S98–S109. Matsui, K., Gromowski, G.D., Li, L., Schuh, A.J., Lee, J.C., and Barrett, A.D. (2009). Characterization of dengue complex-reactive epitopes on dengue 3 virus envelope protein domain III. Virology 384, 16–20. McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C., and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40, 658–674. Megret, F., Hugnot, J.P., Falconar, A., Gentry, M.K., Morens, D.M., Murray, J.M., Schlesinger, J.J., Wright, P.J., Young, P., Van Regenmortel, M.H., et al. (1992). Use of recombinant fusion proteins and monoclonal antibodies

Nelson, S., Jost, C.A., Xu, Q., Ess, J., Martin, J.E., Oliphant, T., Whitehead, S.S., Durbin, A.P., Graham, B.S., Diamond, M.S., and Pierson, T.C. (2008). Maturation of West Nile virus modulates sensitivity to antibody-mediated neutralization. PLoS Pathog. 4, e1000060. Oliphant, T., Nybakken, G.E., Austin, S.K., Xu, Q., Bramson, J., Loeb, M., Throsby, M., Fremont, D.H., Pierson, T.C., and Diamond, M.S. (2007). Induction of epitope-specific neutralizing antibodies against West Nile virus. J. Virol. 81, 11828–11839. Plevka, P., Battisti, A.J., Junjhon, J., Winkler, D.C., Holdaway, H.A., Keelapang, P., Sittisombut, N., Kuhn, R.J., Steven, A.C., and Rossmann, M.G. (2011). Maturation of flaviviruses starts from one or more icosahedrally independent nucleation centres. EMBO Rep. 12, 602–606. Renard, M., Belkadi, L., and Bedouelle, H. (2003). Deriving topological constraints from functional data for the design of reagentless fluorescent immunosensors. J. Mol. Biol. 326, 167–175. Rey, F.A., Heinz, F.X., Mandl, C., Kunz, C., and Harrison, S.C. (1995). The envelope glycoprotein from tick-borne encephalitis virus at 2 A resolution. Nature 375, 291–298. Roehrig, J.T., Bolin, R.A., and Kelly, R.G. (1998). Monoclonal antibody mapping of the envelope glycoprotein of the dengue 2 virus, Jamaica. Virology 246, 317–328. Sabin, A.B. (1952). Research on dengue during World War II. Am. J. Trop. Med. Hyg. 1, 30–50. Sangkawibha, N., Rojanasuphot, S., Ahandrik, S., Viriyapongse, S., Jatanasen, S., Salitul, V., Phanthumachinda, B., and Halstead, S.B. (1984). Risk factors in dengue shock syndrome: a prospective epidemiologic study in Rayong, Thailand. I. The 1980 outbreak. Am. J. Epidemiol. 120, 653–669. Shrestha, B., Brien, J.D., Sukupolvi-Petty, S., Austin, S.K., Edeling, M.A., Kim, T., O’Brien, K.M., Nelson, C.A., Johnson, S., Fremont, D.H., and Diamond, M.S. (2010). The development of therapeutic antibodies that neutralize homologous and heterologous genotypes of dengue virus type 1. PLoS Pathog. 6, e1000823. Simonelli, L., Beltramello, M., Yudina, Z., Macagno, A., Calzolai, L., and Varani, L. (2010). Rapid structural characterization of human antibody-antigen complexes through experimentally validated computational docking. J. Mol. Biol. 396, 1491–1507. Stiasny, K., Kiermayr, S., Holzmann, H., and Heinz, F.X. (2006). Cryptic properties of a cluster of dominant flavivirus cross-reactive antigenic sites. J. Virol. 80, 9557–9568. Sukupolvi-Petty, S., Austin, S.K., Purtha, W.E., Oliphant, T., Nybakken, G.E., Schlesinger, J.J., Roehrig, J.T., Gromowski, G.D., Barrett, A.D., Fremont, D.H., and Diamond, M.S. (2007). Type- and subcomplex-specific neutralizing antibodies against domain III of dengue virus type 2 envelope protein recognize adjacent epitopes. J. Virol. 81, 12816–12826. Thullier, P., Lafaye, P., Me´gret, F., Deubel, V., Jouan, A., and Mazie´, J.C. (1999). A recombinant Fab neutralizes dengue virus in vitro. J. Biotechnol. 69, 183–190. Thullier, P., Demangel, C., Bedouelle, H., Me´gret, F., Jouan, A., Deubel, V., Mazie´, J.C., and Lafaye, P. (2001). Mapping of a dengue virus neutralizing

Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved 313

Structure Determinants of Dengue Virus Cross-Neutralization

epitope critical for the infectivity of all serotypes: insight into the neutralization mechanism. J. Gen. Virol. 82, 1885–1892.

tion of Chikungunya virus particles revealed by X-ray crystallography. Nature 468, 709–712.

van der Schaar, H.M., Rust, M.J., Chen, C., van der Ende-Metselaar, H., Wilschut, J., Zhuang, X., and Smit, J.M. (2008). Dissecting the cell entry pathway of dengue virus by single-particle tracking in living cells. PLoS Pathog. 4, e1000244.

Wahala, W.M., Kraus, A.A., Haymore, L.B., Accavitti-Loper, M.A., and de Silva, A.M. (2009). Dengue virus neutralization by human immune sera: role of envelope protein domain III-reactive antibody. Virology 392, 103–113.

Voss, J.E., Vaney, M.C., Duquerroy, S., Vonrhein, C., Girard-Blanc, C., Crublet, E., Thompson, A., Bricogne, G., and Rey, F.A. (2010). Glycoprotein organiza-

Zhang, Y., Zhang, W., Ogata, S., Clements, D., Strauss, J.H., Baker, T.S., Kuhn, R.J., and Rossmann, M.G. (2004). Conformational changes of the flavivirus E glycoprotein. Structure 12, 1607–1618.

314 Structure 20, 303–314, February 8, 2012 ª2012 Elsevier Ltd All rights reserved