Structure, Vol. 13, 17–28, January, 2005, ©2005 Elsevier Ltd All rights reserved. DOI 10.1016/j.str.2004.10.017

The Structure and Function of the Outer CoatProtein VP9 of Banna Virus

Fauziah Mohd Jaafar,1,4 Houssam Attoui,1,4,*Mohammad W. Bahar,2,4 Christian Siebold,2

Geoffrey Sutton,2 Peter P.C. Mertens,3

Philippe De Micco,1 David I. Stuart,2

Jonathan M. Grimes,2,* and Xavier De Lamballerie1

1Unité des Virus Emergents EA3292EFS Alpes-Méditerranée and Faculté de MédecineUniversité de la Méditerranée27 Bd Jean Moulin13005 MarseilleFrance2Division of Structural BiologyThe Henry Wellcome Building for Genomic MedicineOxford UniversityRoosevelt DriveOxford, OX3 7BNUnited Kingdom3Institute for Animal HealthPirbright LaboratoryAsh RoadPirbright, WokingSurrey, GU24 0NFUnited Kingdom

Summary

Banna virus (BAV: genus Seadornavirus, family Reo-viridae) has a double-shelled morphology similar torotavirus and bluetongue virus. The structure of BAVouter-capsid protein VP9 was determined by X-ray crys-tallography at 2.6 Å resolution, revealing a trimericmolecule, held together by an N-terminal helical bundle,reminiscent of coiled-coil structures found in fusion-active proteins such as HIV gp41. The major domainof VP9 contains stacked � sheets with marked struc-tural similarities to the receptor binding protein VP8of rotavirus. Anti-VP9 antibodies neutralize viral in-fectivity, and, remarkably, pretreatment of cells withtrimeric VP9 increased viral infectivity, indicating thatVP9 is involved in virus attachment to cell surfaceand subsequent internalization. Sequence similaritieswere also detected between BAV VP10 and VP5 por-tion of rotavirus VP4, suggesting that the receptorbinding and internalization apparatus, which is a sin-gle gene product activated by proteoloysis in rotavi-rus, is the product of two separate genome segmentsin BAV.

Introduction

Banna virus (BAV) was initially isolated in China fromthe cerebrospinal fluid and serum of a human patientsuffering from encephalitis (Chen and Tao, 1996; Li,1992; Xu et al., 1990), and it is therefore classified as a

*Correspondence: h-attoui-ets-ap@gulliver.fr; jonathan@strubi.ox.ac.uk

4 These authors contributed equally to this work.

potentially dangerous “BSL3” arboviral agent. BAV hasa genome composed of 12 segments of dsRNA and isthe prototype species of a new genus, “Seadornavi-rus,” within the family Reoviridae, which was first iden-tified during phylogenetic studies of the coltiviruses(Attoui et al., 2000). Two virus species from SoutheastAsia (Banna virus [BAV] and Kadipiro virus [KDV]) weresubsequently reassigned as seadornaviruses, based oncomparisons of their genome sequences (Attoui et al.,2000). Banna virus has recently been further subdividedinto genotype A (which includes the original Chineseisolate and a related Indonesian isolate) and genotypeB (which includes only Indonesian isolates). The sead-ornaviruses appear to be endemic in Southeast Asia,particularly Indonesia and China, and are thought to betransmitted by mosquitoes (Brown et al., 1993; Chenand Tao, 1996).

Virus particles of members of the family Reoviridae(the reoviruses) have icosahedral capsids, which areusually nonenveloped with a diameter of 60–85 nm (ex-cluding the extended fiber proteins that project fromthe surfaces of some virus species) (Attoui et al., 2002;Estes, 2001; Mertens et al., 2004; Mertens et al., 2000;Murphy et al., 1968; Suzuki et al., 1993). Their capsidspossess up to three concentric protein layers: the sub-core, inner capsid (or core), and outer capsid, respec-tively (Baker et al., 1999; Mertens et al., 2000). The reovi-ruses can be divided into two groups based on capsidstructure. These are the “spiked” or “turreted” viruses(e.g., orthoreoviruses and cypoviruses), which have tur-ret-like structures at the 5-fold axes of the innermostcapsid layer (Baker et al., 1999; Hill et al., 1999; Nibertand Schiff, 2001), while the “nonspiked” viruses (e.g.,rotaviruses and orbiviruses) have a relatively smooth orbristled inner capsid appearance (Grimes et al., 1998;Mertens et al., 2000). A striking structural feature com-mon to all reoviruses is the innermost capsid layer,composed of 120 copies of a single viral protein ar-ranged with icosahedral symmetry, that has been de-scribed as pseudo T = 2 (Grimes et al., 1998). Cores ofthe nonturreted viruses are strengthened by an outerlattice of 780 copies of another protein. This lattice isarranged with T = 13l icosahedral symmetry, leading toa symmetry mismatch between different capsid layers(Grimes et al., 1998).

BAV particles are composed of seven structural pro-teins (VP1, VP2, VP3, VP4, VP8, VP9, and VP10), andhave a nonspiked morphology composed of three dis-tinct protein layers (subcore, core, and outer capsid)that is reminiscent of the orbiviruses and rotaviruses(Table 1) (Mohd Jaafar et al., submitted). During the ini-tial stages of infection, the entire BAV core is translo-cated into the cytoplasm of the target cell by a mecha-nism that remains largely uncharacterized (like manyother nonenveloped viruses). However, like the otherreoviruses, it is thought to involve components of theouter capsid (BAV proteins VP4 and VP9) in cell attach-ment and membrane penetration (Ruggeri and Green-berg, 1991; Charpilienne et al., 1997; Hassan and Roy,1999; Denisova et al., 1999; Dowling et al., 2000; Zárate

Structure18

Table 1. Proteins of the Rotavirus A and Banna Virus: Their Functions and Locations within the Virion

Location Rotavirus Function BAV Function

Core

5-fold axes VP1(Pol) RNA polymerase VP1(Pol) RNA polymerase5-fold axes VP3(Cap) Guanylyl and methyl transferase VP3(Cap) Guanylyl transferaseInnermost layer VP2(T2) Binds RNA VP2(T2) Unknown

T = 2Outer layer of VP6(T13) Major virion protein, with group and VP8(T13) Major virion protein

core subgroup antigenic determinantsCore surface VP10 May form a stalk-base for VP9

Outer Coat

VP7 Neutralization, possible cell attachment VP4 Unknownprotein

VP4 Cleaved to Surface spike: neutralization, VP9 Type-specific antigen, neutralization,VP5* and VP8* hemagglutinin, cell attachment cell attachment and penetration, (VP9

and VP10 may have a collective rolesimilar to that of rotavirus VP4)

et al., 2000; Hassan et al., 2001; Chandran et al., 2001; ktDormitzer et al., 2003; Forzan et al., 2004). The BAVaouter layer is stripped away during cell entry, leavingHthe two protein layers of the core. These contain majoruproteins VP2 and VP8, which are thought to be equiva-tlent to components of rotavirus (VP2 and VP6), or BTVtcores (VP3 and VP7), respectively. The BAV core encap-

sidates the viral transcription complexes (VP1 and VP3)tand shields the virus genome, preventing its recogni-dtion by dsRNA-activated host defenses. However, un-Vlike BTV and rotavirus, BAV cores also contain a thirdPmajor structural protein, VP10, located at the surfaceeof the core.SWe report the bacterial expression of BAV outer cap-asid protein VP9 and its functional and structural analy-isis. Recombinant VP9 was used to generate genotype-pspecific, neutralizing antibodies, which identified twoadistinct serotypes, implicating VP9 as the BAV cell-Vattachment protein. Pretreatment of susceptible cellsawith recombinant VP9 also increased BAV-Ch (humansChinese isolate) infectivity significantly, suggesting acrole in the process of membrane penetration. Analysistof recombinant VP9 by X-ray crystallography has re-bvealed a trimeric molecule with structural similarities to(a portion of rotavirus outer capsid protein VP4 (VP8).

BAV VP10 shares significant sequence homology withthe other portion of rotavirus VP4 (VP5), suggesting an Vassociation between VP9 and VP10 to form the BAV cell Tentry apparatus. u

Pr

Results and Discussion ge

Oligomeric State of VP9 in Solution tRecombinant GST-VP9 fusion proteins (from both BAV- gCh and BAV-In6969) were overexpressed, purified, and lcleaved by PreScission protease (Amersham Biosci- iences, France) (see the Experimental Procedures). The rpurified VP9 was crosslinked with glutaraldehyde and aanalyzed by size exclusion chromatography and SDS- oPAGE. Three protein species, corresponding to a trimer b

t(w95 kDa, and the most abundant species), dimer (w65

Da), and monomer (w31 kDa) (Figure 1A) were de-ected. This suggests that VP9 naturally forms trimersnd may be present in this form in the virus capsid.owever, no trimers of the expressed or native butncrosslinked VP9 (either boiled or unboiled) were de-ected after SDS-PAGE, suggesting that they are rela-ively fragile and can be disrupted by SDS.

Trimeric proteins have previously been described inhe outer coat of several viruses of the family Reoviri-ae, including σ1 and �1 of orthoreoviruses, VP2 andP5 of orbiviruses, VP5 and VP7 of rotaviruses, and the8 of phytoreoviruses (Dormitzer et al., 2000; Dormitzert al., 2003, Hassan and Roy, 1999; Hassan et al., 2001;choehn et al., 1997; Zhu et al., 1997). VP4, the cellttachment protein of rotavirus has previously been

dentified as a dimer in the outer capsid of native virusarticles by electronmicroscopy (Prasad et al., 1990),lthough the atomic structure of the VP5 domain (fromP4) determined by X-ray crystallography has revealedtrimeric structure (Dormitzer et al., 2003). A significant

tructural rearrangement of VP5 may occur after itsleavage from VP4, which may not only be involved inhe enhanced virus infectivity observed but also maye necessary for its membrane penetration activity

Golantsova et al., 2004).

P9 Can Recoat Viral Coreshe recoating of BAV cores by VP9 was examined bysing a recombinant, expressed His-tagged protein.urified cores were incubated in the presence of the

ecombinant VP9, then submitted to iterative Percollradient purifications. The resulting virus particles werexamined by electron microscopy after sequential reac-ions with anti-VP9 antibodies and protein-A colloidalold (Figures 1B and 1C). The particles were efficiently

abeled with 10 and 20 nm gold particles, demonstrat-ng the presence of VP9 molecules and at least partialecoating of the cores. Anti-His tag antibodies werelso used to identify recombinant VP9, in a preparationf recoated and repurified core particles, by Westernlot analysis; results from this analysis demonstratedhe ability of this protein to bind to the surface of the

Outer Capsid Protein VP9 of Banna Virus19

Figure 1. The Oligomeric State of VP9 andthe Recoating Analysis of BAV Core

(A) Crosslinking analysis of recombinant VP9proteins from BAV-Ch and BAV-In6969.Lanes 1 and 3: uncrosslinked VP9 BAV-Chand VP9 BAV-In6969, respectively, showingonly monomeric VP9; lanes 2 and 4: cross-linked VP9 BAV-Ch and VP9 BAV-In6969,respectively, showing mainly trimers (w95kDa) and a minority of dimers (w60 kDa) andmonomers(w30 kDa). Positions are indi-cated by arrows.(B) Electron micrographs of purified BAV-Chcores recoated with VP9, labeled with pro-tein A gold (10 nm colloidal gold particles)after incubation of recoated cores with anti-VP9 BAV-Ch antibodies.(C) Electron micrographs of purified BAV-Chcores recoated with VP9, labeled with pro-tein A gold (20 nm colloidal gold particles)after incubation of recoated cores with anti-VP9 BAV-Ch antibodies.(D) Western blot of BAV-Ch cores recoatedwith recombinant VP9 detected with anti-Histag antibodies. Lane 1: cores recoated withrecombinant His-tagged VP9; lane 2: recom-binant His-tagged VP9 used as a control.

BAV core. Furthermore, recoating increased the infec-tivity of the cores, although the particles remained lessinfectious than native virus particles (data not shown).However, the recoating efficiency was low, as shown bythe intensity of the recombinant VP9 band in Figure 2C.Recoating experiments have only previously been re-ported for the orthoreoviruses (i.e., spiked viruses). Inthe orthoreovirus system, three studies have describedthe successful recoating by recombinant σ3 protein ofinfectious subviral particles (i.e., particles in which theσ3 protein is missing) and of cores (i.e., particles inwhich the σ3, σ1, and �1 proteins are missing) by a mixof recombinant σ3 and �1 proteins or by a mix of σ3,σ1, and �1 (Chandran et al., 1999; Chandran et al.,2001; Jane Valbuena et al., 1999). Our experimentalwork provides the first description, to our knowledge,of a recoating mechanism for a “nonspiked” reovirus,using only one of the two outer coat components.

The BAV cores appear to be unique amongst the“nonspiked” Reoviridae in possessing a third majorcore protein. Based on the recoating results, we pro-pose that there is a direct interaction between VP9 andVP10, which favors a loose attachment of VP9 to viralcores.

A Role for VP9 in Cell AttachmentIn order to investigate whether VP9 is involved in virusattachment to cells, we studied the effect of anti-VP9antibodies on the infection of C6/36 cells by BAV-Ch.Homologous (anti-recombinant BAV-Ch VP9) and heter-ologous (anti-recombinant BAV-In6969 VP9) antibodieswere used in virus neutralization assays. These homol-ogous and heterologous antibodies define the two

serotypes of Banna virus. The homologous antibodiesprevented the infection of 200 pfu of the virus, at anantibody dilution of up to 1 in 600 (Figure 2A), while theheterologous antibody did not inhibit the infection. Thissuggests that VP9 is involved in cell attachment andcarries serotype-specific, virus-neutralization epitopes.The BAV genotypes A and B have therefore also beenidentified as distinct serotypes A and B.

We also investigated the influence of the solublemonomeric and trimeric forms of recombinant VP9 onvirus infection. When C6/36 cells were preincubatedwith uncrosslinked recombinant BAV-Ch or BAV-In6969VP9s prior to infection with BAV-Ch, no significant ef-fect on virus infectivity was observed by immunofluo-rescence or Western blot assays (Figure 2B). However,when the cells were treated with either recombinantBAV-Ch or BAV-In6969 VP9 that had been crosslinkedwith glutaraldehyde (to create mainly trimers; Figure 1),the infectivity of BAV-Ch was markedly increased (Fig-ure 2B). In contrast, the infectivity of whole avian ortho-reovirus particles is reduced by soluble recombinant σ3attachment protein (analogous to mammalian orthoreo-virus σ1) (Grande et al., 2000). This may indicate thatwhereas in orthoreoviruses, attachment and penetra-tion are two functions carried by separate proteins (σ1and �1 in mammalian reovirus), stabilized VP9 trimersof BAV not only bind receptor but also initiate endocy-tosis, perhaps by receptor oligomerization. The initia-tion of endocytosis facilitates penetration by virus par-ticles present at or near the cell surface and therebyincreases their infectivity. The difference in the behaviorof the virus in the presence of soluble monomeric ortrimeric VP9 suggests that the functional activity of VP9

Structure20

Figure 2. Banna Virus Neutralization with Antibodies against VP9 Proteins and Assay of Infectivity in the Presence of VP9 Proteins

(A) The effect of anti-recombinant BAV-Ch VP9 antibodies on virus infectivity. Cells were inoculated with 200 pfu BAV-Ch that had beenpretreated with the ascitic fluid dilutions as described in the text. Cells were counterstained with Evans blue stain (red coloration). Fluores-cence due to viral protein production (productive infection) is detected only in cells infected with 200 pfu BAV preincubated with a dilution of1/800 of anti-BAV-Ch ascitic fluid.(B) The effect of pretreatment of cells with VP9 on the level of infection. Cells were pretreated with either VP9 or crosslinked VP9 and werethen inoculated with 200 pfu BAV-Ch. Cells pretreated with trimeric VP9 show a marked increase in infection (panel 3) compared to untreatedcells or cells treated with monomeric VP9 (panel 1 and panel 2). Cells inoculated with 2500 pfu virus show a similar level of infection to cellspretreated with trimeric VP9 and infected with 200 pfu virus (panel 4).

requires its trimerization; this is in line with both type I Rand type II viral fusion machines, which require trimeric wassociations of the protein to drive cell entry (Bressa- lnelli et al., 2004, Skehel and Wiley, 2000). T

βcStructure Determination and Subunit Structure1Selenomethionated BAV-Ch VP9 was expressed iniE. coli, purified and crystallized in nanolitre drops (seedthe Experimental Procedures). X-ray data were col-(lected to a resolution of 2.6 Å in a three-wavelengthcanomalous dispersion (MAD) experiment performed on8a single crystal at the UK MAD beamline, BM14, ESRF,TGrenoble. The crystal belonged to space group C2 and3contained three subunits of VP9 in the crystallographic

asymmetric unit. The structure has been refined to an t

factor of 18.3% including all data to 2.6 Å resolution,ith good stereochemistry (rmsd from ideal bond

engths 0.008 Å, see the Experimental Procedures andable 2 for details). The VP9 subunit is composed of 11strands (β1−β11), 6 α helices (α1–α6), and 4 310 heli-

es (η1–η4), as assessed by DSSP (Kabsch and Sander,983). These secondary structural elements are defined

n Figure 3. The subunit is arranged into two structurallyistinct domains, an N-terminal extended helical “stalk”

residues 32–88) some 50 Å in length, which supports aompact, predominantly β, C-terminal “head” (residues9–283, approximate dimensions of 30 × 30 × 50 Å3).he overall structure of the subunit is shown in Figure. The first 26–30 amino acids (the number varies be-ween the three noncrystallographically related sub-

Outer Capsid Protein VP9 of Banna Virus21

Table 2. Crystallographic Statistics

Peak Inflection Remote

Data Collection

Resolution range (Å) 30.0–3.0 (3.11–3.00) 30.0–2.55 (2.64–2.55) 30.0–3.0 (3.11–3.00)Space group C2Cell dimensions (Å) a = 126.4, b = 73.8, c = 95.8Cell angles (°) α = γ = 90, β = 97.2Wavelength (Å) 0.97880 0.979055 0.8731f#,f$ −6.9, 4.3 −9.7, 3.1 −1.8, 2.6Unique reflections 16,805 25,233 17,041Completeness (%)a 96.9 (88.6) 89.6 (91.2) 98.8 (94.2)Rmerge (%)a,b 9.2 (17.1) 9.7 (45.5) 9.4 (23.9)I/σI 22.6 (5.8) 9.0 (1.7) 11.6 (3.4)Average redundancy 7.5 2.5 3.0

Phasing

Rano (%)c 6.9 6.7 6.9Rdisp (%)d 4.5 5.7FOM (solve/resolve) 0.51/0.68

Refinement

Resolution range (Å) 95–2.6Number of reflections 26,540R factor (%)e 18.3Rfree (%)f 24.8rmsd bonds (Å) 0.008rmsd angles (°) 1.1rmsd main chain bond B (Å2) 3.3rmsd side chain bond B (Å2) 4.2Number of protein atoms per asymmetric unit 5,809 (200)

(waters)Average protein B factors (waters) (Å2) 19.1 (18.7)

FOM, figure of merit; rmsd, root-mean-square deviation from ideal geometry.a The numbers in parentheses refer to the appropriate outer shell.b Rmerge = ΣhklΣi|I(hkl;i) − <I(hkl)>|/ΣhklΣiI(hkl;i), where I(hkl;i) is the intensity of an individual measurement, and <I(hkl)> is the average intensityfrom multiple observations.c Rano = Σi|I(+) − I(−)|/Σ<I>| for anomalous differences, where <I> is the average of Friedel amplitudes.d Rdisp = Σi|Iλ1− Iλ2|/Σ<I>| for dispersive differences, where <I> is the average amplitude at two wavelengths, λ1 and λ2.e Rfactor = Σhkl||Fobs| − k|Fcalc||/Σhkl|Fobs|.f Rfree equals the R factor against 5% of the data removed prior to refinement.

units) of VP9 were not visible in the electron densitymap; the structure is otherwise complete, with the ex-ception of the His6 tags at the C termini, which are com-pletely disordered. The N-terminal stalk domain is com-posed of four α helices. The two N-terminal helicesprotrude downward, extending some 20 Å below thehead domain, which is constructed from a distorted βsandwich dominated by a central 5-stranded antiparal-lel sheet (β5–7, β10–11) overlaid on a smaller, also anti-parallel, sheet (β8–9) flanked by an α helix (α5). Integralto the head is an additional two-layer β-pleated subdo-main composed of two β hairpins (β1–2 and β3–4,respectively) stacked at an angle to each other.

The VP9 TrimerAs discussed above, the physiological state of associa-tion of native VP9 is likely to be trimeric. The crystalstructure supports this, since the molecular unit pre-sent in the crystallographic asymmetric unit is a homo-trimer. The three independent subunits within the crys-tallographic asymmetric unit are related by a 3-foldrotation axis and are very similar to each other (rmsd ofmain chain atoms 0.1 Å). The molecule resembles a

three-lobed floret (of approximate dimensions 60 ×70 × 70 Å3) resting on a trimeric stalk (Figure 4). Sub-units associate through the N-terminal stalk and thestacked β hairpins of the head. The N-terminal domainsform a three-membered, left-handed super-helix, withhelices α1, α2, and α4 associating directly with theirsister helices to form a stalk of diameter 20 Å andlength 50 Å, while helix α3 loops out to form an orthog-onal buttress supporting the C-terminal head domainof the preceding monomer (Figure 4). Each subunit isfurther fastened onto the preceding monomer throughdirect interactions of the β1-β2 hairpin with the neigh-boring C-terminal helix. Approximately 3600 Å2 of surfacearea are buried per monomer upon trimer formation (ARE-AIMOL) (CCP4, 1994). Of this area, the contribution bythe stalk region (residues 32–88) is approximately 3200Å2. Conventional wisdom would suggest that thiswould render the trimer stable (Janin et al., 1988);hence, the evidence for monomers and dimers in solu-tion might suggest that the stalk stabilization is un-usually small, as would be expected if the molecule hasa functional requirement for conformational change. Inthe absence of more detailed structural data for the in-

Structure22

Figure 3. The VP9 Subunit

(A and B) The secondary and tertiary structures of VP9 aligned with amino acid sequences are shown. (A) is a conventional cartoon represen-tation of the subunit, colored from blue at the N terminus through green and yellow to red at the C terminus. α helices and β strands arelabeled. The 3-fold axis of the trimer runs vertically, in the plane of the paper. (B) is an alignment of the amino acid sequences of all availableBAV VP9 sequences drawn with ESPRIPT (Gouet et al., 1999). The secondary structural elements are marked above the sequence alignment.The secondary structural elements of VP9 (as defined with DSSP, Kabsch and Sander, 1983) comprise β strands: residues 95–96, 103–105,111–113, 128–129, 142–151, 156–169, 180–188, 197–200, 228–231, 239–240, and 257–264; α helices: α1–α6, residues 36–42, 46–51, 63–68,74–84, 206–215, and 270–280; and 310 helices: residues 189–191, 194–196, 220–222, and 267–269.

tact virion (only negative stain electron microscopy re- rcsults are available [F. Mohd Jaafar et al., submitted]),

we cannot be certain of the arrangement of VP9 on theecapsid; however, the antigenic data suggest that the

molecule is exposed on the capsid surface, and we t(propose that the molecule is trimeric in this context.

It seems plausible that the disordered N-terminal w30 c

esidues make stabilizing interactions with anotherapsid protein.Electrostatic potential calculations (GRASP) (Nicholls

t al., 1991) reveal that while much of the surface of therimer is relatively uncharged and rather hydrophobicFigure 4), there is a markedly basic depression at theenter of the extensive upper surface. It is conceivable

Outer Capsid Protein VP9 of Banna Virus23

Figure 4. The VP9 Trimer

A series of three different representations are shown from left to right. The two rows show orthogonal views; in the upper images, themolecular 3-fold axis is vertical and in the plane of the paper (as in Figure 3), while the lower set are drawn with the molecular 3-foldperpendicular to the page. The left-hand representation is a conventional cartoon (colored by subunit). The central images show van derWaals representations, with residues color coded according to the conservation of amino acid sequence (colors are as used, such that redrepresents completely conserved). A conserved patch on the outside of the trimer is identified by a yellow ellipse. The right-hand imagesshow the surface charge distribution (GRASP (Nicholls et al., 1991)). An acidic pocket is identified by a yellow circle (the view is tilted by 10°about the horizontal to reveal this pocket).

that this might be a site of interaction with an acidiccellular receptor such as heparin sulfate (by analogywith, for instance, foot-and-mouth disease virus, Fry etal., 1999); however, the residues forming this depres-sion are not conserved across the Banna viruses (seeFigure 3), suggesting either a switch in the primary re-ceptor or that this site binds a secondary receptor(since VP9 is antigenic, we would expect functionallyunconstrained surface residues to change under im-mune pressure from the host). In contrast, there is aconserved patch on the outer surface of the trimer (Fig-ure 4) that may be a site for protein-protein interaction.Finally, there is a small, negatively charged pocket be-neath the β sandwich domain (Figure 4).

The Role of VP9 in Viral Attachment and PenetrationAccording to the DALI server (http://www.ebi.ac.uk/dali/) the VP9 subunit is unlike any fold observed pre-viously. Close inspection reveals a convincing struc-tural homology with rotavirus VP8 (a cleavage productof the spike protein VP4 [Dormitzer et al., 2002]) (pro-gram SHP [Stuart et al., 1979] aligns 79 out of 283 resi-dues with an rmsd in Cαs of 3.5 Å). The topology ofthe cores of the VP9 head domain and rotavirus VP8 isidentical (Figure 5). This core comprises seven β strandsand the C-terminal α helix. The differences between the

two molecules correspond to a single insertion into thismodule of a series of further secondary structural units,at different places for the two molecules. For VP9, thisaddition comprises a β-α-β structure inserted towardthe C terminus of the core, whereas rotavirus VP8 hasan additional pair of β hairpins toward the core N termi-nus (Figure 5). The result is that VP8 has an extra layerof β structure, which it shares with the galectins, whileVP9 has structural elaborations on the “outer” portionof the trimeric structure. The fold of rotavirus VP8 ismore similar to the carbohydrate binding galectin familyof proteins than to BAV VP9. This is in line with its roleas the viral hemagglutinin, since the VP8 sialic acidbinding site is not conserved in BAV VP9 and sialidasetreatment of the C6/36 cells does not inhibit BAV-Chinfectivity. These findings suggest that sialic acid doesnot bind to VP9 and does not play a key role in BAVadsorption to the cell surface. On the basis of the struc-tural similarity between rotavirus VP8 and the galectins,Dormitzer et al. (2002) proposed that an ancestral ro-tavirus VP4 acquired a host-derived, galectin-like car-bohydrate binding domain as a cell attachment do-main. However, an argument against this is that sugarbinds to different surfaces for VP8 and the galectins.The structural similarity between BAV VP9 and rotavirusVP8 probably reflects the evolutionary relationship be-tween these two members of the Reoviridae family, and

Structure24

Figure 5. Structural Relationship between VP9 and Rotavirus VP8

Comparison of VP9 with rotavirus VP8 (Dormitzer et al., 2002). The two molecules are drawn separately as stereo images for clarity (top:residues 125–283 of VP9; bottom: residues 108–224 of VP8 bottom). The relative orientation is that determined by SHP (Stuart et al., 1979),which matches 79 Cα atoms with an rms deviation of 3.5 Å. The topologically similar core region is drawn in green and represented as asolid object. The insertion in VP9 relative to the conserved core is drawn in semitransparent red, while the insertion in VP8 is drawn insemitransparent blue. For clarity, a topological diagram of each molecule is also drawn, colored as drawn in the stereo cartoons, with theconserved core in green and insertions in red and blue for VP9 and VP8, respectively.

the structural divergence is likely to reflect a diver- aqgence in cellular receptors. More surprising is the ap-

parent variation in oligomeric state and size between (rthese two outer coat components, since cryo-EM

analysis of rotavirus suggests that VP4 forms a long, irextended dimeric molecule in the virus (approximately

100 Å in length), with VP8 located at the outer tips (Dor- timitzer et al., 2002). However, recent results suggest

that this may be misleading since VP5 is trimeric (Dor- Vomitzer et al., 2003). VP5 and VP8 are the two products

of proteolytic cleavage of VP4 (a cleavage that pro- rpmotes viral infectivity). VP5 forms a stalk that presents

VP8 and is involved in membrane permeabilization, ippossessing a putative fusion peptide (which shows se-

quence similarity to those of viruses using the type II d1fusion mechanism, such as flaviviruses) toward its N

terminus (Dowling et al., 2000). The implication is that htrotavirus VP4 could be trimeric but one of the three VP8

domains is simply disordered in the EM analysis. The t

mino acid sequence of BAV VP10 shows 26% se-uence identity with the C-terminal half of rotavirus VP5

VP10 possesses 249 residues and VP5 possesses 535esidues) (F. Mohd Jaafar et al., submitted). This sim-larity is greater than that observed between VP9 andotavirus VP8 (7 sequence identities out of the 79 struc-urally equivalent residues) and is presumably reflectedn structural similarity (sequence similarity betweenP9 and VP8 may have been eroded by a combinationf change in receptor usage and antigenic variation). Inotavirus, VP5 and VP8 are expressed as a single generoduct (VP4), and activation of the virus for cell entry

s achieved by proteolysis to form VP5 and VP8. It ap-ears that in BAV, the equivalent proteins are found onifferent genome segments (seadornaviruses possess2 genome segments, in contrast to rotaviruses, whichave 11). We suggest that VP10 helps anchor VP9 inhe virion (using the w30 residue flexible N-terminal ex-ension of VP9), enabling VP10-bearing cores to reab-

Outer Capsid Protein VP9 of Banna Virus25

sorb VP9 and reconstitute particles of significant infec-tivity. However, BAV VP10 is much smaller than VP5 ofrotavirus, and it lacks the portion of the molecule bear-ing the putative type II fusion peptide. In this context,it is intriguing that the helical bundle at the core of theVP9 trimer is reminiscent of the inner helical bundles ofproteins that form the fusion machinery of viruses thatuse the Type I fusion mechanism (Figure 4). It is an openquestion as to whether the cell entry mechanism ofseadornaviruses has parallels with either of the well-studied class I and class II viral fusion machines, orwhether it uses a different mechanism to achieve mem-brane penetration. However, we note that the analo-gous molecular apparatus responsible for cell attach-ment and cell entry in the flaviviruses acts by dimer/trimer transition triggered by a change in pH, accompa-nied by massive conformational rearrangments (Allisonet al., 1995; Bressanelli et al., 2004). By analogy, it islikely that conformational rearrangements of the helicalstalk of BAV VP9 would destabilize the trimeric arrange-ment of the globular heads, triggering further confor-mational changes.

ConclusionsThe family Reoviridae is unusual in that the genetic re-lationships and host range of its members are diverse;nevertheless, a fundamental similarity in structure be-tween the innermost capsid layers of otherwise dispa-rate family members betrays their common origin(Bamford et al., 2001). Until now, relationships betweenthe proteins in the outermost capsid layers have beenless clear, and gene transfer has been invoked to ex-plain some of the morphological diversity of theseviruses and to provide a mechanism for expanding thehost range (Dormitzer et al., 2002). Our structural andfunctional analyses of VP9 from the newly charac-terized Banna virus reveal a trimeric molecule that pres-ents globular head domains on the virus surface for re-ceptor binding. The structural similarity between Bannavirus VP9 and rotavirus VP8 (taken with the similarity infunction and oligomeric state) now suggests that thereare also underlying similarities in the mechanisms ofcell entry between different genera of the Reoviridae(for instance Rotavirus, Orbivirus, and Seadornavirus).

Experimental Procedures

Expression and Purification of SelenomethionylVP9 Protein of BAV-ChSegment 9 of BAV-Ch was PCR amplified by using PCR primers(5#-3#): VP9S-Pres: CCCAGGAATTCCCCTGGAAGTTCTGTTCCAGGGGCCCATGTTATCGGAGACTGAGTTGAGGGCTT (bold: EcoRI,italics: PreScission protease site, underlined: segment 9-specificsequence) and VP9R-Pres: ACGATGCGGCCGCTCATTAGTGATGGTGATGGTGATGAGGCAAATAACTTAAAGCAT (bold: NotI, italics: 6xHistag, underlined: segment 9-specific sequence).

The pGEX-4T-2 vector and the PCR products were double-digested separately by EcoRI and NotI (Invitrogen), gel purified,and ligated by using T4 DNA ligase. The construct should allow theexpression of a GST fusion protein (GST being located at the Nterminus), permitting the purification of the fusion protein by gluta-thione affinity chromatography. A PreScission protease (AmershamBiosciences, France) cleavage site between GST and the VP9 en-ables the release of the latter from the GST moiety.

The recombinant vector was transfected into BL-Codon-Plus(DE3)-RP-X bacteria. A single colony was grown overnight (ON)

in 10 ml Selenomet Medium (Molecular Dimensions Limited) con-taining L-methionine. Fractions (1 ml) of the overnight culture werecentrifuged at 2000 × g and washed with sterile deionized water.The pellet was used to inoculate ten flasks of 100 ml SelenometMedium containing L-selenomethionine as directed by the manu-facturer.

The culture was grown and induced as described earlier (MohdJaafar et al., 2004). Bacteria were harvested by centrifugation at3000 × g, suspended in PBS (containing Complete cocktail antipro-tease from Roche Applied Sciences, France), and sonicated byusing a microtip of a Vibracell sonicator (20% output for a total of1.5 min). The fusion protein was bound to a glutathione sepharosematrix in a column (Amersham Biosciences, France) as directed bythe manufacturer. The bound protein was cut with 35 U PreScissionprotease (Amersham Biosciences, France) at 8°C ON in 600 µl Pre-Scission protease buffer (50 mM Tris-HCl [pH 8.0], 1 mM EDTA, 1mM DTT, 40 mM NaCl) to cleave the VP9 away from the bound GSTmoiety. Selenomethionyl VP9 was further purified on ion exchangeQ Vivapure membrane columns (Vivascience). Recombinant BAV-In6969 VP9 was produced and purified by using essentially iden-tical procedures.

Functional Analysis of VP9Recoating Core Particles with Recombinant VP9BAV-Ch-infected cells were lysed by using deionized water andwere then treated with Vertrel-XF solvent (F. Mohd Jaafar et al.,submitted). The aqueous phase was layered on top of a discontinu-ous gradient of caesium chloride made of two layers of CsCl (40%and 55% w/v) (Burroughs et al., 1994) in 100 mM Tris-HCl (pH 8.0).The gradients were spun at 35,000 × g in an SW41 rotor. Coresformed a compact layer at the interface of the two caesium layers.Cores were diluted in 100 mM Tris-HCl (pH 8.0) and 10 mM MgCl2and were centrifuged at 100,000 × g in a bench-top ultracentrifugewith a TL100.2 rotor, over a cushion (200 µl) of 66% w/w sucrosein Tris-HCl 100 mM (pH 8.0). Cores were recovered and diluted in100 mM Tris-HCl (pH 8.0), 10 mM MgCl2, and 150 mM NaCl andwere dialyzed overnight against the dilution buffer. The core par-ticles were adjusted to a concentration of 1 mg/ml in 400 µl. 200µg recombinant VP9 in PreScission protease buffer was added. Themixture was incubated at 37°C for 2 hr and layered on top of alinear gradient of percoll in VP buffer (145 mM NaCl, 250 mMsucrose, 1 mM MgCl2, 4 mM CaCl2, 10 mM Tris-HCl [pH 8.0]). Thetubes were centrifuged at 111,000 × g in an SW41 rotor. The treatedcore band that was visible in the lower quarter of the gradient wasrecovered, diluted in VP buffer, and centrifuged again in an iden-tical percoll gradient. The resulting band was centrifuged on a thirdpercoll gradient. The band from the third gradient was diluted in 10mM Tris-HCl (pH 8.0), 10 mM MgCl2 and was centrifuged at111,000 × g in an SW41 rotor for 1 hr to precipitate the silica beadsfrom the percoll. The treated cores were recovered at the surface ofthe precipitated beads, diluted again, and centrifuged at 100,000 × gin a TL100.2 rotor.Western Blot Analysis of Recoated CoresThe treated core was mixed with loading buffer (160 mM Tris-HCl,4 mM EDTA, 3.6% SDS, 60 mM DTT, 0.2% β-mercaptoethanol) andanalyzed by SDS-PAGE by using 10% acrylamide. The proteins inthe gel were blotted on nitrocellulose membranes as describedelsewhere (Mohd Jaafar et al., 2004). Anti-His tag antibodies wereused to detect the presence of the recombinant VP9 that recoatedthe cores.Labeling of the Recoated Cores with Colloidal GoldThe cores were diluted in 10 mM Tris-HCl (pH 7.5) containing 150mM NaCl. The anti-VP9 ascitic fluid was diluted to 1/100 in the coresuspension and incubated at 37°C for 1 hr. The mixture was coatedon formvar carbon grids. Protein A labeled with 10 or 20 nm colloi-dal gold (PAG) was diluted in PBS to OD520 of 0.25. The grid wasincubated in the dilute PAG at 37°C for 1 hr, followed by washingin 100 mM Tris-HCl (pH 7.5). The grids were stained in 2% uranylacetate for 30 s, dried, washed briefly in 10 mM Tris-HCl (pH 7.5),and observed by transmission electron microscopy.Virus Neutralization with Antibodies against VP9 Proteins200 µl of the virus suspension containing 200 pfu was incubatedwith an equal volume of serially diluted anti-VP9 ascitic fluid in L-15.

Structure26

The ascitic fluid dilutions in the final volume of inoculum were 1/10, (q1/100, 1/200, 1/600, 1/800, 1/1,000, 1/10,000, and 1/100,000. 400 µl

aliquots of these mixes were incubated for 2 hr at 37°C and then aaused to inoculate (at 27°C) C6/36 cell monolayers grown in shell

vials containing a coverslide. The medium was removed and re- m(placed with fresh culture medium. The vials were incubated at 27°C

for 48 hr, and the cells were fixed with cold acetone. The slides uCwere washed with PBS and incubated for 1 hr at 37°C with anti-

BAV-Ch ascitic fluid diluted 1/800. After washing in PBS, an FITC- wTconjugated anti-mouse IgG was added on the coverslides, and

slides were incubated for 1 hr at 37°C. The coverslides were Pmmounted on glass slides in the presence of buffered glycerol and

were examined by light fluorescence microscopy. Alternatively, cells lqwere dissolved in denaturating buffer, and Western blot analysis

was performed with anti-BAV-Ch antibodies. toAssays of Infectivity of C6/36 Cells by BAV-Ch in Presence

of Recombinant VP9 from BAV-Ch and BAV-In6969 nC6/36 cells were incubated in L-15 medium for 2 hr at 27°C in thepresence of serial dilutions (containing 200, 10, and 1 µg/ml) of VP9proteins (BAV-Ch or BAV-In6969). 200 pfu of the purified whole virus Awere added to the culture. The cells were allowed to stand in theprotein-virus mixture for an additional 2 hr at 27°C. The excess Wliquid was removed, the cells were washed, and the medium was dreplaced by fresh culture medium. The cells were incubated at 27°C pfor 48 hr and were then analyzed by IF and Western blot with anti- GBAV-Ch ascitic fluid. bAssays of Infectivity in Neuraminidase-Treated C6/36 Cells ERecombinant neuraminidase from Salmonella typhimurium was 0purchased from Sigma (Poole, UK). The activity of the enzyme was ptested by using homozygote human red blood cells (RBCs) of the fMNS type supplied by the blood bank of Marseille (EFS-AM, Mar- aseille, France). The M surface antigen of these RBCs contains sialic sacid as an essential component for recognition by a monoclonal Eanti-M agglutinating antibody (Biotest, France). The untreatedRBCs were agglutinated, while neuraminidase-treated RBCs did

Rnot agglutinate in the presence of the anti-M antibody, confirmingRthe activity of the neuraminidase. Cells were incubated in L-15 me-Adium in the presence of 40 mU/ml neuraminidase for 2 hr at 37°CP(Mendez et al., 1993; Nibert et al., 1995). The cells were washed

twice with L-15 medium, and 200 pfu of BAV-Ch was added. CellsRwere incubated at 27°C for 2 hr, after which the excess inoculum

was removed. Cells were washed, and fresh culture medium wasAadded, followed by incubation at 27°C for 48 hr prior to analysis byHIF and Western blot.a6

Structural Analysis ACrystallization and Data Collection (Prior to crystallization, the selenomethionine-substituted VP9 was Bconcentrated by ultrafiltration to a final protein concentration of 12 gmg/ml in 50 mM Tris-HCl (pH 8.0), 1 mM DTT, 1 mM EDTA, 40 mM

8NaCl. An initial crystallization screen of 400 conditions was carried

Aout by the sitting drop vapor diffusion method with a 200 nl dropPsize (100 nl protein + 100 nl precipitant) by using a Cartesian robot((Brown et al., 2003; Walter et al., 2003). The protein crystallizedOunder several different conditions. Crystals suitable for X-ray struc-Bture determination were found with the Hampton Screen I conditiond41 (0.1 M HEPES-Na [pH 7.5], 20% [w/v] polyethylene glycol 4000,i10% [v/v] iso-propanol). They diffracted to a resolution of 2.6 ÅMafter flash freezing at 105K. Prior to flash freezing, VP9 crystals

were transferred for approximately 1 min to a cryoprotectant solu- Btion of crystallization buffer plus 20% glycerol. A three-wavelength MMAD experiment was performed at the ESRF, Grenoble, France on Sbeamline BM14 using a MAR CCD 133 mm detector. The X-ray data

Bwere processed and scaled with the HKL suite (Otwinowski and

LMinor, 1997) (Table 2).

eStructure Determination and Analysis

cThe VP9 structure was determined by the MAD method. Each sub-

Bunit of VP9 contains 8 methionine residues, and of the 24 methio-Cnines expected in the crystallographic asymmetric unit (the sizeoof the unit cell was consistent with there being 3 subunits in the

crystallographic asymmetric unit), the positions of 21 were readily BBdetermined, and phases computed, with SOLVE (Terwilliger and Be-

rendzen, 1999). Density modification was effected by using RESOLVE S

Terwilliger, 2000). The resulting electron density map was of excellentuality and allowed automatic chain tracing with RESOLVE (69% ofll residues in the asymmetric unit was built, and 49% of the aminocid sequence was placed) (Terwilliger, 2003). An initial proteinodel was built into the electron density by using the program O

Jones et al., 1991). Noncrystallographic symmetry restraints weresed to refine the model to a resolution of 2.6 Å using the programNS and Refmac (CCP4, 1994; Brunger et al., 1998). The modelas refined at 2.6 Å to an R factor of 18.3%, with an Rfree of 24.8%.he structure has reasonable stereochemistry, as assessed byROCHECK (Laskowski et al., 1993); 87% of residues lie within theost favored regions of the Ramachandran plot, and 2 out of 768

ie in disallowed regions. Further phasing, refinement, and modeluality statistics are shown in Table 2. Structural superposi-ions were performed with SHP (Stuart et al., 1979), and unlesstherwise credited, figures were produced with BOBSCRIPT (Es-ouf, 1999).

cknowledgments

e thank R. Esnouf and K. Harlos for computation and in-houseata collection, and Nicolas Aldrovandi for electron microscopyreparations. We thank the staff of the UK MAD beam line, BM14,renoble and, in particular, Martin Walsh. The work was supportedy the Medical Research Council, UK. C.S. is supported by theuropean Commission Integrated Program SPINE (QLG2-CT-2002-0988). J.M.G. is supported by the Royal Society, and D.I.S. is sup-orted by the Medical Research Council, UK. Support also comes

rom European Union Grant “Reo ID” number QLK2-2000-00143,nd the Institut de Recherche pour le Développement and Etablis-ement Français du Sang Alpes-Méditerranée. The “Unité des Virusmergents” is an associated research unit of the IRD.

eceived: June 9, 2004evised: October 21, 2004ccepted: October 21, 2004ublished: January, 11, 2005

eferences

llison, S.L., Schalich, J., Stiasny, K., Mandl, C.W., Kunz, C., andeinz, F.X. (1995). Oligomeric rearrangement of tick-borne enceph-litis virus envelope proteins induced by an acidic pH. J. Virol. 69,95–700.

ttoui, H., Billoir, F., Biagini, P., de Micco, P., and de Lamballerie, X.2000). Complete sequence determination and genetic analysis ofanna virus and Kadipiro virus: proposal for assignment to a newenus (Seadornavirus) within the family Reoviridae. J. Gen. Virol.1, 1507–1515.

ttoui, H., Mohd Jaafar, F., Biagini, P., Cantaloube, J.F., de Micco,., Murphy, F.A., and de Lamballerie, X. (2002). Genus Coltivirusfamily Reoviridae): genomic and morphologic characterization ofld World and New World viruses. Arch. Virol. 147, 533–561.

aker, T.S., Olson, N.H., and Fuller, S.D. (1999). Adding the thirdimension to virus life cycles: three-dimensional reconstruction of

cosahedral viruses from cryo-electron micrographs. Microbiol.ol. Biol. Rev. 63, 862–922.

amford, D.H., Gilbert, R.J., Grimes, J.M., and Stuart, D.I. (2001).acromolecular assemblies: greater than their parts. Curr. Opin.truct. Biol. 11, 107–113.

ressanelli, S., Stiasny, K., Allison, S.L., Stura, E.A., Duquerroy, S.,escar, J., Heinz, F.X., and Rey, F.A. (2004). Structure of a flavivirusnvelope glycoprotein in its low-pH-induced membrane fusiononformation. EMBO J. 23, 728–738.

rown, S.E., Gorman, M., Tesh, R.B., and Knudson, D.L. (1993).oltiviruses isolated from mosquitoes collected in Indonesia. Virol-gy 196, 363–367.

rown, J., Walter, T.S., Carter, L., Abrescia, N.G.A., Aricescu, A.R.,atuwangala, T.D., Bird, L.E., Brown, N., Chamberlain, P.P., Davis,.J., et al. (2003). A procedure for setting up high-throughput, na-

Outer Capsid Protein VP9 of Banna Virus27

nolitre crystallisation experiments. II. Crystallisation results. J.Appl. Crystallogr. 36, 315–318.

Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P.,Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M.,Pannu, N.S., et al. (1998). Crystallography & NMR system: a newsoftware suite for macromolecular structure determination. ActaCrystallogr. D Biol. Crystallogr. 54, 905–921.

Burroughs, J.N., O’Hara, R.S., Smale, C.J., Hamblin, C., Walton, A.,Armstrong, R., and Mertens, P.P.C. (1994). Purification and proper-ties of virus particles, infectious subviral particles, cores and VP7crystals of African horsesickness virus serotype 9. J. Gen. Virol. 75,1849–1857.

CCP4 (Collaborative Computational Project, Number 4)(1994). TheCCP4 suite: programs for protein crystallography. Acta Crystallogr.D Biol. Crystallogr. 50, 760–763.

Chandran, K., Walker, S.B., Chen, Y., Contreras, C.M., Schiff, L.A.,Baker, T.S., and Nibert, M.L. (1999). In vitro recoating of reoviruscores with baculovirus-expressed outer capsid proteins Mu1 andSigma3. J. Virol. 73, 3941–3950.

Chandran, K., Zhang, X., Olson, N.H., Walker, S.B., Chappell, J.D.,Dermody, T.S., Baker, T.S., and Nibert, M.L. (2001). Complete in vi-tro assembly of the reovirus outer capsid produces highly infec-tious particles suitable for genetic studies of the receptor-bindingprotein. J. Virol. 75, 5335–5342.

Charpilienne, A., Abad, M.J., Michelangeli, F., Alvarado, F., Vasseur,M., Cohen, J., and Ruiz, M.C. (1997). Solubilized and cleaved VP7,the outer glycoprotein of rotavirus, induces permeabilization of cellmembrane vesicles. J. Gen. Virol. 78, 1367–1371.

Chen, B., and Tao, S. (1996). Arbovirus survey in China in recentten years. Chin. Med. J. (Engl.) 109, 13–15.

Denisova, E., Dowling, W., LaMonica, R., Shaw, R., Scarlata, S.,Ruggeri, F., and Mackow, E.R. (1999). Rotavirus capsid protein VP5*permeabilizes membranes. J. Virol. 73, 3147–3153.

Dormitzer, P.R., Greenberg, H.B., and Harrison, S.C. (2000). Purifiedrecombinant rotavirus VP7 forms soluble, calcium-dependent tri-mers. Virology 277, 420–428.

Dormitzer, P.R., Sun, Z.Y., Wagner, G., and Harrison, S.C. (2002).The rhesus rotavirus VP4 sialic acid binding domain has a galectinfold with a novel carbohydrate binding site. EMBO J. 21, 885–897.

Dormitzer, P.R., Nason, E., Prasad, B.V.V., and Harrison, S.C. (2003).Crystal structure of the VP4 membrane interaction domain sug-gests major structural rearrangements during entry and upon prim-ing. Abstract W2.1: Eighth International Symposium on Double-Stranded RNA Viruses.

Dowling, W., Denisova, E., LaMonica, R., and Mackow, E.R. (2000).Selective membrane permeabilization by the rotavirus VP5* proteinis abrogated by mutations in an internal hydrophobic domain. J.Virol. 74, 6368–6376.

Esnouf, R.M. (1999). Further additions to MolScript version 1.4, in-cluding reading and contouring of electron-density maps. ActaCrystallogr. D Biol. Crystallogr. 55, 938–940.

Estes, M.K. (2001). Rotaviruses and their replication. In Fields Virol-ogy, D.M. Knipe and P.M. Howley, eds. (Philadelphia: Lippincott Wil-liams and Wilkins), pp. 1747–1785.

Forzan, M., Wirblich, C., and Roy, P. (2004). A capsid protein ofnonenveloped Bluetongue virus exhibits membrane fusion activity.Proc. Natl. Acad. Sci. USA 101, 2100–2105.

Fry, E.E., Lea, S.M., Jackson, T., Newman, J.W., Ellard, F.M., Blake-more, W.E., Abu-Ghazaleh, R., Samuel, A., King, A.M., and Stuart,D.I. (1999). The structure and function of a foot-and-mouth diseasevirus-oligosaccharide receptor complex. EMBO J. 18, 543–554.

Golantsova, N.E., Gorbunova, E.E., and Mackow, E.R. (2004).Discrete domains within the rotavirus VP5* direct peripheral mem-brane association and membrane permeability. J. Virol. 78, 2037–2044.

Gouet, P., Courcelle, E., Stuart, D.I., and Metoz, F. (1999). ESPript:analysis of multiple sequence alignments in PostScript. Bioinfor-matics 15, 305–308.

Grande, A., Rodriguez, E., Costas, C., Everitt, E., and Benavente,

J. (2000). Oligomerization and cell-binding properties of the avianreovirus cell-attachment protein sigmaC. Virology 274, 367–377.

Grimes, J.M., Burroughs, J.N., Gouet, P., Diprose, J.M., Malby, R.,Zientara, S., Mertens, P.P.C., and Stuart, D.I. (1998). The atomicstructure of the bluetongue virus core. Nature 395, 470–478.

Hassan, S.H., and Roy, P. (1999). Expression and functional charac-terization of bluetongue virus vp2 protein: role in cell entry. J. Virol.75, 8356–8367.

Hassan, S.H., Wirblich, C., Forzan, M., and Roy, P. (2001). Expres-sion and functional characterization of bluetongue virus vp5 pro-tein: role in cellular permeabilization. Virology 75, 8356–8367.

Hill, C.L., Booth, T.F., Prasad, B.V., Grimes, J.M., Mertens, P.P.C.,Sutton, G.C., and Stuart, D.I. (1999). The structure of a cypovirusand the functional organization of dsRNA viruses. Nat. Struct. Biol.6, 565–568.

Jane-Valbuena, J., Nibert, M.L., Spencer, S.M., Walker, S.B., Baker,T.S., Chen, Y., Centonze, V.E., and Schiff, L.A. (1999). Reovirus vi-rion-like particles obtained by recoating infectiuos subvirion par-ticles with baculovirus-expressed sigma3 protein: an approach foranalysing sigma3 functions during virus entry. J. Virol. 73, 2963–2973.

Janin, J., Miller, S., and Chothia, C. (1988). Surface, subunit inter-faces and interior of oligomeric proteins. J. Mol. Biol. 204, 155–164.

Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjeldgaard, M. (1991). Im-proved methods for building protein models in electron densitymaps and the location of errors in these models. Acta Crystallogr.A 47, 110–119.

Kabsch, W., and Sander, C. (1983). Dictionary of protein secondarystructure: pattern recognition of hydrogen-bonded and geometricalfeatures. Biopolymers 22, 2577–2637.

Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M.(1993). PROCHECK: a program to check the stereochemical qualityof protein structures. J. Appl. Crystallogr. 26, 283–291.

Li, Q.P. (1992). First isolation of 8 strains of new orbivirus (Banna)from patients with innominate fever in Xinjiang. Endemic Dis. Bull.7, 77–82.

Mendez, E., Arias, C.F., and Lopez, S. (1993). Binding to sialic acidsis not an essential step for the entry of animal rotaviruses to epithe-lial cells in culture. J. Virol. 67, 5253–5259.

Mertens, P.P.C. (2004). The dsRNA viruses. Virus Res. 101, 3–13.

Mertens, P.P.C., Arella, M., Attoui, H., Belloncik, S., Bergoin, M.,Boccardo, G., Booth, T.F., Chiu, W., Diprose, J.M., Duncan, R., etal. (2000). Reoviridae. In Virus Taxonomy: The Seventh Reportof the International Committee on Taxonomy of Viruses, M.H.V.Van-Regenmortel, C.M. Fauquet, D.H.L. Bishop, E.B. Carsten, M.K.Estes, S.M. Lemon, J. Maniloff, M.A. Mayo, D.J. McGeoch, C.R.Pringle, and R.B. Wickner, eds. (New York: Academic Press), pp.395–480.

Mertens, P.P.C., Attoui, H., Duncan, R., and Dermody, T.S. (2004).Reoviridae. In Virus Taxonomy. Eighth Report of the InternationalCommittee on Taxonomy of Viruses, C.M. Fauquet, M.A. Mayo, J.Maniloff, U. Desselberger, and L.A. Ball, eds. (London: Elsevier/Academic Press), pp. 447–454.

Mohd Jaafar, F., Attoui, H., Gallian, P., Isahak, I., Wong, K.T.,Cheong, S.K., Nadarajah, V.S., Cantaloube, J.F., Biagini, P., DeMicco, P., et al. (2004). Recombinant VP9-based enzyme-linked im-munosorbent assay for detection of immunoglobulin G antibodiesto Banna virus (genus Seadornavirus). J. Virol. Methods 116, 55–61.

Murphy, F.A., Coleman, P.H., Harrison, A.K., and Gary, W.G. (1968).Colorado tick fever virus: an electron microscopy study. Virology35, 28–40.

Nibert, M.L., and Schiff, L.A. (2001). Reoviruses and their replica-tion. In Fields Virology, D.M. Knipe and P.M. Howley, eds. (NewYork: Lippincott Williams and Wilkins), pp. 1679–1728.

Nibert, M.L., Chappell, J.D., and Dermody, T.S. (1995). Infectioussubvirion particles of reovirus type 3 dearing exhibit a loss in infec-tivity and contain a cleaved s1 protein. J. Virol. 69, 5057–5067.

Nicholls, A., Sharp, K.A., and Honig, B. (1991). Protein folding and

Structure28

association: insights from the interfacial and thermodynamic prop-erties of hydrocarbons. Proteins 11, 281–296.

Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffrac-tion data collected in oscillation mode. Methods Enzymol. 276,307–326.

Prasad, B.V., Burns, J.W., Marietta, E., Estes, M.K., and Chiu, W.(1990). Localization of VP4 neutralization sites in rotavirus by three-dimensional cryo-electron microscopy. Nature 343, 476–479.

Ruggeri, F.M., and Greenberg, H.B. (1991). Antibodies to the trypsincleavage peptide VP8 neutralize rotavirus by inhibiting binding ofvirions to target cells in culture. J. Virol. 65, 2211–2219.

Schoehn, G., Moss, S.R., Nuttall, P.A., and Hewat, E.A. (1997).Structure of broadhaven virus by cryoelectron microscopy: correla-tion of structural and antigenic properties of broadhaven virus andbluetongue virus outer capsid proteins. Virology 235, 191–200.

Skehel, J.J., and Wiley, D.C. (2000). Receptor binding and mem-brane fusion in virus entry: the influenza hemagglutinin. Annu. Rev.Biochem. 69, 531–569.

Stuart, D.I., Levine, M., Muirhead, H., and Stammers, D.K. (1979).Crystal structure of cat muscle pyruvate kinase at a resolution of2.6 Å. J. Mol. Biol. 134, 109–142.

Suzuki, H., Konno, T., and Numazaki, Y. (1993). Electron micro-scopic evidence for budding process-independent assembly ofdouble-shelled rotavirus particles during passage through endo-plasmic reticulum membranes. J. Gen. Virol. 74, 2015–2018.

Terwilliger, T.C. (2000). Maximum-likelihood density modification.Acta Crystallogr. D Biol. Crystallogr. 56, 965–972.

Terwilliger, T.C. (2003). Improving macromolecular atomic modelsat moderate resolution by automated iterative model building, sta-tistical density modification and refinement. Acta Crystallogr. DBiol. Crystallogr. 59, 1174–1182.

Terwilliger, T.C., and Berendzen, J. (1999). Automated MAD andMIR structure solution. Acta Crystallogr. D Biol. Crystallogr. 55,849–861.

Walter, T.S., Diprose, J., Brown, J., Pickford, M., Owens, R.J., Stu-art, D.I., and Harlos, K. (2003). A procedure for setting up high-throughput, nanolitre crystallisation experiments. I. Protocol designand validation. J. Appl. Crystallogr. 36, 308–314.

Xu, P., Wang, Y., Zuo, J., Lin, J., and Xu, P. (1990). New orbivirusesisolated from patients with unknown fever and encephalitis in Yun-nan province. Chin. J. Virol. 6, 27–33.

Zárate, S., Espinosa, R., Romero, P., Mendez, E., Arias, C.F., andLopez, S. (2000). The VP5 domain of VP4 can mediate attachmentof rotaviruses to cells. J. Virol. 74, 593–599.

Zhu, P., Hemmings, A.M., Iwasaki, K., Fujiyoshi, Y., Zhong, B., Yan,J., Isogai, M., and Omura, T. (1997). Details of the arrangement ofthe outer capsid of rice dwarf phytoreovirus, as visualized by two-dimensional crystallography. J. Virol. 71, 8899–8901.

Accession Numbers

The coordinates and structure factors have been deposited in theProtein Data Bank with accession code 1W9Z.