Structure of the vesicular stomatitis virus nucleocapsidin complex with the nucleocapsid-binding domain ofthe small polymerase cofactor, PTodd J. Green and Ming Luo1

Department of Microbiology, University of Alabama School of Medicine, 1025 18th Street South, Birmingham, AL 35294

Edited by Robert A. Lamb, Northwestern University, Evanston, IL, and approved May 1, 2009 (received for review March 23, 2009)

The negative-strand RNA viruses (NSRVs) are unique because theirnucleocapsid, not the naked RNA, is the active template fortranscription and replication. The viral polymerase of nonseg-mented NSRVs contains a large polymerase catalytic subunit (L)and a nonenzymatic cofactor, the phosphoprotein (P). Insight intohow P delivers the polymerase complex to the nucleocapsid haslong been pursued by reverse genetics and biochemical ap-proaches. Here, we present the X-ray crystal structure of theC-terminal domain of P of vesicular stomatitis virus, a prototypicnonsegmented NSRV, bound to nucleocapsid-like particles. P bindsprimarily to the C-terminal lobe of 2 adjacent N proteins within thenucleocapsid. This binding mode is exclusive to the nucleocapsid,not the nucleocapsid (N) protein in other existing forms. Localiza-tion of phosphorylation sites within P and their proximity to theRNA cavity give insight into how the L protein might be orientedto access the RNA template.

negative-strand RNA virus � replication � template � transcription �phosphorylation

Vesicular stomatitis virus (VSV) belongs to the family Rhab-doviridae, which also includes rabies virus. The rhabdoviruses

are part of the broad group of negative-strand RNA viruses(NSRVs), which contain many medically relevant viruses, includingavian influenza, measles, and Ebola. VSV has long served as aprototypic nonsegmented negative-strand RNA virus (NNSRV),partly because of the small number of genes that are encoded by its11-kb genome (1). These genes include the nucleocapsid protein(N), a phosphoprotein (P), a matrix protein (M), a glycoprotein(G), and a large polymerase protein (L). Each of these proteins hasmultiple functions and, as a result, has multiple binding partners,including but not limited to each other.

The NNSRVs are characterized by the unique fact that in theentire replication cycle, their genomes do not exist as naked RNA,but rather are encapsidated by their nucleocapsid proteins. Thenucleocapsid is the active template for transcription and replication(2, 3). Structures of the nucleocapsid-like particles (NLPs) of VSVand rabies virus have recently been solved (4, 5), showing that theN protein has 2 lobes angled together to form a cavity forencapsidation of the genomic RNA. Each N monomer accommo-dates 9 bases of RNA. The structures also revealed that eachmonomer of N interacts with 3 neighboring N molecules across thenucleocapsid. The contacts involve the elongated N terminus andan extended loop (C loop) within the C-terminal lobe, and they arerequired for RNA encapsidation (6). Residues within this loop havealso been implicated in binding to the P (7, 8). Recently, moreinsight into capsid formation was gained through crystallographicstudies of an N protein with a serine-to-tryptophan mutation atresidue 290 (called N290; ref. 6). N290 has lost the ability toencapsidate RNA because of the bulky side chain of tryptophan inthe RNA cavity, yet the capsid assembly functions of the proteinremain intact. Thus, the N protein alone contains all of theinformation for the assembly of a capsid structure.

The multidomain P protein (Indiana strain) of VSV is the 265-aanonenzymatic cofactor of the viral polymerase. The N-terminal

domain comprises the first 106 residues of P and contains 3 residues(Ser-60, Thr-62, and Ser-64) that may be phosphorylated by hostcasein kinase II (9–11). These residues are indispensable becausethey are required for transcription (12–14). The central domain,residues 107–177, is the site of P dimerization and subsequenttetramer formation (15, 16). Previous experiments have shown thatmultimerization beyond the dimeric state is essential to P’s role inreplication (11, 17). The majority of the remaining 89 residues formthe C-terminal domain (PCTD). PCTD contains 2 additional phos-phorylation sites (Ser-226 and Ser-227). The L protein has beenshown previously to interact with these residues (13), and althoughphosphorylation is not required for L binding, phosphorylation ofeither residue is responsible for regulating levels of replication. Thisdomain of P is also the major site of N protein association (18–23).The structure of the PCTD was determined recently by NMR andshown to form a single compact unit comprising an antiparallel�-turn and 5 �-helices (24).

Upon translation and before polymerization of N onto thegenome, N forms an initial complex with P, known as No-P(25–27). This RNA-free, encapsidation-competent complex isdelivered to the active replication site, possibly via the secondaryinteraction of P with the viral polymerase. Here, N comes intocontact with the newly synthesized genomic RNA, and theprocess of encapsidation occurs. In addition to binding N andforming the encapsidation precursor, P binds to the nucleocapsidduring the viral replication cycle that involves polynucleotidesynthesis. The L polymerase subunit cannot recognize thenucleocapsid alone. The processes of transcription and replica-tion require the association of the P as a component of the viralpolymerase complex (2, 28). Thus, P is the determining factor fortemplate recognition for the viral polymerase.

The N, P, and L proteins are unique in that they form distinctcomplexes at different stages of the replication cycle. As de-scribed above, L and P form the RNA polymerase that associateswith the N-enwrapped template. Additional experimentationhas shown that some P mutants maintain the ability to bind Nand form No-P but are inactive in transcription. These complexeswere, however, capable of supporting replication (22). There-fore, it was surmised that the replicase and transcriptase were 2distinct complexes. Subsequently, an active replicase complexwas isolated, and in fact it was shown to contain the N, P, andL proteins (29). This tripartite complex was distinct from thetranscriptase, a complex that does not contain the N protein.These experiments confer the complexity of the interactions ofN, P, and L, and although structures of N and 2 individual

Author contributions: T.J.G. and M.L. designed research, performed research, analyzeddata, and wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

The atomic coordinates for the N/RNA–PCTD and N290–PCTD complex structures have beendeposited in the Protein Data Bank, www.rcsb.org (PDB ID codes 3HHZ and 3HHW,respectively).

1To whom correspondence should be addressed. E-mail: mingluo@uab.edu.

www.pnas.org�cgi�doi�10.1073�pnas.0903228106 PNAS � July 14, 2009 � vol. 106 � no. 28 � 11713–11718

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

June

12,

202

0

domains of P have been published, the precise nature of thedirect interaction of N and P together is unknown. Likewise, thenature of the tripartite complex between N, P, and L is not wellunderstood. Ultimately, because these complexes play key rolesin the synthetic processes of transcription, replication, andencapsidation of progeny genomes, atomic-level snapshots oftheir interactions would be invaluable in aiding interpretation ofpreviously published works.

To date, a large body of work has been performed to aide inunderstanding how the P interacts with N. These studies have reliedheavily on mutagenesis and biochemical analysis. However, aconcrete model of the direct interaction of N and P was still elusive.The work presented here gives the atomic look at the complexformed between the N protein in NLP and the C-terminal nucle-ocapsid-binding domain of P. Two adjacent monomers within thenucleocapsid form a unique binding site for P, which is a binding sitethat can only be found in the nucleocapsid. Binding occurs solely tothe C lobes of the 2 N proteins. This explains how the P delivers theviral polymerase specifically to the nucleocapsid for RNA synthesis.New insights are found in the mechanism by which the P allows thepolymerase complex to recognize the nucleocapsid as the templatefor viral RNA synthesis.

ResultsN/PCTD Structure. The N protein is present in the nucleocapsid asa linear polymer. The P protein must recognize the polymeric Nprotein template and, subsequently, P must remain associatedwith the nucleocapsid as the polymerase complex moves downthe template during RNA synthetic events. These 2 events areessential to the function of the viral polymerases of the NNSRV.To date, it has been unclear exactly how the N protein and Passociate. The structure of the middle multimerization domainof P (residues 107–177) was reported previously as a dimer (15).It was also reported recently that the structure of the uncom-plexed C-terminal 71 aa of P (residues 195–265) contains a singlecompact domain comprising an antiparallel �-turn and 5 �-he-lices (24). Here, we present the crystal structure of PCTD incomplex with NLP in the presence and absence of RNA. Thereare no significant differences between the 2 structures. Theoverall rmsd between the RNA-encapsidated PCTD complex andthe RNA-deficient N290 PCTD complex is 1.41 Å. When super-imposed, the main noticeable difference is the orientation of theN lobe with respect to the center of the 10-member ring. Thesedifferences were also observed in NLP structures in the absenceof PCTD (4, 6). In the NLP–PCTD complex, N exists as a decamericring with all intramolecular interactions intact, as in the uncom-plexed NLP (4, 6), whereas each monomeric PCTD interacts with2 adjacent N monomers (Fig. 1). This is an important featurebecause this unique binding site for P could be present only inthe nucleocapsid. A total of 18 residues from PCTD contribute tothe interaction with the dimeric N-binding site. N donates 17 aato the interaction: 11 and 6 residues from each N of the bindingsite, respectively. The residues from N are from a contiguousstretch, between 354 and 386, that encompasses helix �13 and theC loop, which rises above the upper surface of the C-terminallobe. Twenty-four hydrogen bonds are formed between P andthe 2 N monomers. The complete list of bonding partners isprovided in Table 1. Two residues of P, Arg-260 and Lys-262,which have been implicated previously in N protein associationand mutation of these residues, have also been shown to affecttranscription (13). These basic residues are bonded to 2 acidicresidues, Asp-358 and Glu-377, which are found on differentmolecules of the N protein-binding site. Interestingly, Tyr-256and Asn-257 of P are unique, because each interacts with bothmolecules of N. The total surface area buried by the interactionsbetween the PCTD and the 2 adjacent N molecules is 956 Å2,whereas the total surface area of the PCTD is 5,100 Å2. Thus,

�19% of the available surface area of PCTD is in complex withthe N protein(s).

The conserved hydrophobic core of PCTD was described previ-ously (24). Two adjacent surfaces of the PCTD present interestingconserved hydrophobic cavities. The first hydrophobic cavity ispositioned between the �-turn and 2 helices (�3 and �4), and thesecond is situated between helix �1 and the loop connecting helices�2 and �3. The relevance of each cavity is unknown and has beenpostulated as a potential binding site for the other VSV proteins.The complex structure between N and PCTD showed that the firstcavity is capped by the C loop of the N protein. The loop does notpenetrate the cavity, but it does lie over the entrance. The secondcavity is distal to the N-binding surface of PCTD and is left exposed.This suggests that these hydrophobic cavities most likely form ahydrophobic core that renders stability to this small domain.However, the second cavity is solvent-accessible and could beavailable to interact with the domains of P that are missing from thisstructure, or an alternate viral or host protein.

PCTD Structure Changes. The topology of the PCTD structure wasdescribed in a recent report by Ribeiro et al. (24). Although thisdomain forms a compact unit, the individual secondary structureelements appear to have some flexibility. It is very likely thatPCTD undergoes some conformational changes, which may not beglobal, upon binding to the nucleocapsid. PCTD in the NLP–PCTDcomplex is topologically identical to that described in the pre-vious study. The size of the recombinant PCTD used in this study

Fig. 1. Ribbon representations of the VSV nucleocapsid in complex with theC-terminal N protein-binding domain of the P protein (PCTD). (A) Three assem-bled N molecules are represented in red, green, and blue. PCTD, shown inyellow, binds to the extended loop (C loop) within the C lobe of 2 adjacent Nmolecules (red and green). (B) Scaled-up image of the interaction. Residues ofthe dual N-binding site that are in contact with PCTD are shown. The illustra-tions found in this and the subsequent figures were generated with PyMOL(41).

11714 � www.pnas.org�cgi�doi�10.1073�pnas.0903228106 Green and Luo

Dow

nloa

ded

by g

uest

on

June

12,

202

0

was slightly longer. To show structural changes in PCTD upon Nbinding, the complexed and uncomplexed PCTD structures weresuperimposed. A single composite structure was created fromthe 20 lowest-energy deposited NMR structures with CNS (30).Superposition of the composite structure with the N-bound struc-ture resulted in an overall rmsd of 2.16 Å for all atoms. By contrast,the rmsd between the best representative conformer in the ensem-ble PCTD structure (as defined in ref. 24) and PCTD structure fromthis study was 2.06 Å. In either case, each of the secondary structureelements and the overall topology of the protein are maintainedupon binding the N protein. There are, however, localized shifts ofthe secondary structure elements, as illustrated in Fig. 2. Thesedisplacements correspond to N-binding regions of P, with the mostnotable shifts occurring at helix �2 and the loop connecting withhelix �3 and the �-turn (Fig. 2).

Changes in the N Protein upon P Binding. Upon P binding, the NLPmaintains the intermolecular N interactions described previ-ously (4, 6), with very little discernible difference in their overallstructures. The main dissimilarity is in the C loop, as shown inFig. 3. Because there seem to be structural changes of N sidechains in the NLP–PCTD complex, we used the higher-resolutionstructure of N290 NLP in complex with PCTD to identify anychanges in N. N290 is an N protein with a mutation of serine totryptophan at residue 290 (6). This mutated N has lost the abilityto bind RNA. Superposition of the N290 structure with thePCTD-bound structure determined here revealed that the C loopis shifted, with many of the side chains rearranged to accom-modate binding of the PCTD (Fig. 3). The remainder of the N290structure superimposed quite well with the PCTD-bound struc-

ture. The 5 monomers overlaid as a single rigid body had an rmsdof 1.117 Å. If the loops (residues 256–369) were removed fromthe calculation, the rmsd was lowered to a value of 1.075 Å. Bycomparison, the rmsd for the N RNA structure to its PCTD-boundcounterpart was 1.223 Å, or 1.066 Å with the loops removed.Interestingly, in the absence of PCTD, the extended loops weredisordered in 3 of the 5 N monomers in the previously deter-mined N RNA structure. In the PCTD-bound structure presentedhere, all 5 of the loops were observed in the electron densitymaps and were easily traced. In this case, PCTD promotedordering of the loop. The inherent flexibility of the loop couldaid in snaring P because P delivers the L protein to thenucleocapsid for RNA synthesis.

DiscussionPolynucleotide synthesis is an essential part of the viral replica-tion cycle. The NSRVs have evolved to perform this reactionwith their own specialized polymerase proteins. For the nonseg-mented NSRVs, including VSV, this requires a concerted set ofevents involving the N protein-enwrapped genome, the L pro-tein, and the nonenzymatic P. One of the roles of P is to deliverL to the active template. The L protein is unable to bind thetemplate directly but, rather, binds to the P and is then deliveredto the active template as P binds to N. The structure presentedhere of an NLP bound to the PCTD shows that the N proteinforms a unique binding site for P that involves 2 adjacentmonomers within the nucleocapsid. This binding site is formedby a contiguous stretch of residues (354–386) that includes helix�13 and the extended loop of the C lobe, the C loop. The loopfrom each N protein snares P by clamping onto it from opposite

Table 1. The binding interactions between the residues of the PCTD and the residues of the dimeric N protein binding site

Interaction P residues Atoms N residues (subunits 1 and 2) Atoms

Hydrophobic LEU214 Side chain THR361(1) Side chainH bond GLN215 NE2 ASP359(1) OD1H bond GLN215 OE1 SER360(1) NHydrophobic LEU217 Side chain LEU364(1) Side chainHydrophobic ILE219 Side chain LEU364(1) Side chainH bond SER233 O THR365(2) OG1H bond VAL234 O LEU364(2) NHydrophobic GLY235 CA LEU364(2) Side chainH bond ARG251 NH1, NH2 THR365(2) OGH bond LYS253 NZ ASN367(1) ND2H bond Lys254 NZ ASP384(1) OD2H bond Lys254 NZ ASP384(1) OH bond Lys254 O ASN386(1) ND2Hydrophobic LEU255 Side chain THR366(1) Side chainH bond TYR256 OH ASP358(2) OH bond TYR256 N ASP384(1) OD1H bond ASN257 ND2 ASP359(2) OD1H bond ASN257 ND2 ASP359(2) OD2H bond ASN257 ND2 LYS354(2) NZH bond ASN257 ND2 GLU383(1) OE2H bond ASN257 OD1 ASP359(2) OD1H bond ASN257 OD1 ASP359(2) OD2H bond ASN257 OD1 SER360(2) OGH bond ASN257 OD1 ASP384(1) OD1H bond ASN257 N ASP384(1) OD1H bond GLN258 N ASP384(1) OD1H bond ARG260 NH1 ASP358(2) OD1Hydrophobic VAL261 Side chain VAL367(1) Side chainH bond LYS262 NZ GLU377(1) OE2H bond TYR263 OH GLY362(1) N

The numbers 1 and 2 shown in parentheses adjacent to N protein residue names and numbers refer to molecule 1 or 2, respectively, of the N dimer. The listof residues was compiled with the aid of the Protein Interfaces, Surfaces, and Assemblies (PISA) server (42) and visual inspection with COOT (36).

Green and Luo PNAS � July 14, 2009 � vol. 106 � no. 28 � 11715

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

June

12,

202

0

sides. Such a binding site is exclusive to the assembled nucleo-capsid and is in accordance with the fact that the active templatefor transcription and replication is the N-enwrapped genome.

Comparison with the unbound NLP structure shows that N islargely unchanged upon binding P, with the most significantvariation occurring in the C loops that bind to PCTD (Fig. 3).Likewise, topologically, P is unchanged upon binding to thenucleocapsid. There is, however, a shift of the secondary struc-ture elements in relation to one another—the 2 most notablemovements are helix �2 and the loop connecting to helix �3 andthe �-turn (Fig. 2). Each of these elements is involved inN-binding. It is possible that the induced conformationalchanges in N and P allow the two to be associated more tightlywhen P binds to the nucleocapsid.

The structure of the NLP shows that the RNA is encapsidatedin a cavity located between the 2 lobes of the N protein. Duringtranscription and replication, the L protein must be positionedat the mouth of this cavity to gain access to the genome. The Lprotein has been shown previously to interact with 2 essentialphosphorylation sites (Ser-226 and Ser-227) on P (13). Theseresidues reside in the C-terminal domain of P and are observedin our structure, but they are not phosphorylated. L binding toP is not dependent on their phosphorylation, but phosphoryla-tion of either residue is responsible for regulating levels ofreplication. Ser-226 and Ser-227 are found on the loop thatconnects 310-helix 1 and helix �2 (Fig. 4). The loop is positionedsuch that these residues face the interior of the N protein ring.Neither serine makes contact with the N protein but, rather, eachsits exposed �50 Å directly above the entrance of the RNAencapsidation cavity and aligned with the interior face of the Clobe of the N protein. The interaction of L with the 2 phosphor-ylated serine residues would not prevent PCTD in the P–Lcomplex from docking on the nucleocapsid. At the same time,PCTD could bring L in close contact with the RNA, because the2 serine residues may be viewed as the boundary of contactbetween L and N. Because of the way that the RNA is seques-tered while encapsidated, the N protein must undergo some

Fig. 2. Structural overlay of the VSV PCTD in the N protein-bound state(yellow) and the unbound state (cyan; PDB accession ID: 2k47). Residues of PCTD

that make contact with the N protein have been shaded magenta on theN-bound representation. (A) Shown is the structure of the PCTD, with second-ary structure elements labeled. (B) The overlay of the 2 structures is shown.

Fig. 3. Structural similarity of the N protein in the PCTD-bound and unboundstates. Two molecules of the PCTD-bound N290 molecules are colored red andgreen. Two additional molecules (colored gray) of the previously determinedstructure N290 (PDB accession ID: 2qvj) were superimposed upon the currentcomplex structure. The main dissimilarity between the structures is observedin the conformation of the C loops (acknowledged with arrows). This corre-sponds to the site of PCTD binding. The models are rotated 180° with respect toFig. 1; in this case, looking from the exterior of the protein rather than the RNAcavity face of N.

Fig. 4. Phosphorylation of the P protein. Cartoon models of 3 molecules ofthe N protein (colored as in Fig. 1) are shown with encapsidated RNA (shownin stick representation). A semitransparent surface covering the 3 molecules ofthe N protein is shown in a shade of light blue, with the RNA cavity colored adarker shade of orange. The bound PCTD is shown in yellow. Residues that arephosphorylated are shown as space-filling spheres and are labeled accord-ingly. The perspective is the same as that in Fig. 1.

11716 � www.pnas.org�cgi�doi�10.1073�pnas.0903228106 Green and Luo

Dow

nloa

ded

by g

uest

on

June

12,

202

0

conformational adjustments in order for L to gain access to thegenome. How this occurs is unknown at this point. One likelyscenario may involve the N-terminal 22 residues of the N protein.These residues form an arm that extends to the C lobe of aneighboring N within the nucleocapsid and holds the 2 lobes of Nin the proper orientation for the formation of the RNA-encapsidation cavity. Previous mutational studies showed that if theN-terminal arm is removed, the ability to bind P remains intact, butthe ability to retain RNA is lost (6). Interestingly, the integrity ofthe assembled nucleocapsid is not completely broken. This featureis important because it dictates that the N-terminal arm candissociate to expose the RNA locally without affecting the globalassociation of the nucleocapsid. The L protein or the L–P complexcould cause this local dissociation, exposing the RNA in the courseof polynucleotide synthesis. P binds to the extended loop of the Clobe of N, and the opposite face of the loop is in contact with theN-terminal arm of N. Upon the P–L complex association with thenucleocapsid, it could be possible that the loop–arm interaction isdestabilized, causing N to open. Subsequently, after transcription orreplication has occurred, the arm is repositioned, and the RNA isagain stored in the cavity.

Here, we propose a model for the process of polynucleotideprocessivity based on the structure described in this work. For Lto encounter the viral genome–nucleocapsid template, thedimeric P is required. This dimer is associated with both L andN. Within the dimer, 1 PCTD associates with the nucleocapsid. Toprocess through the genome, N is forced to open temporarily andexpose the RNA as the polymerase passes along. The mechanismfor forcing the conformational change resulting in the openingof N is due to a local destabilization event caused by P–L binding.Once the binding occurs, N is opened, exposing the RNA;however, upon exit of the active bubble of polynucleotidesynthesis, N returns to the closed state (illustrated in Fig. 5). ThePCTD, as shown in the previous figures, binds to the C lobe of 2adjacent N monomers within the nucleocapsid. The associationshould be fleeting as the polymerase complex moves along thetemplate. Thus, PCTD is expected to repetitively bind and releasethe template as the polymerase complex moves down the tem-plate. In this scenario, the dimer is always associated with L andintermittently associated with N.

There must be differences between the processes of transcrip-tion and replication. For some time, the general thought hasbeen that the difference is the presence of a sufficient quantityof No-P (25–27). As amounts of this encapsidation precursorcomplex increase, the switch to replication is initiated. However,a different hypothesis was proposed, stating that the presence ofNo-P promotes the formation of a tripartite replicase thatswitches to replication (29). Multimerization of P seems to benecessary, because 2 previous studies have shown that both thedeletion of residues 191–210 of P or addition of an artificialpeptide with this sequence affects multimerization and, subse-quently, transcription and replication (16, 31). This region of Pis observed in our structure and is not in contact with the Nprotein. In our model a single C-terminal domain of P is requiredfor recognition of the nucleocapsid. Multimerization of P may bethrough the central domain region, as shown by our previousstructure of the central domain of P, which can form a dimer ortetramer (15). Further structural studies of the complete poly-

Fig. 5. Model of the interactions of P and L with the encapsidated template.The N, P, and L proteins are labeled. Polarity of the template and progeny RNA(both shown as a yellow ribbon) are also labeled with 5� and 3�. The P dimerin the viral polymerase delivers the L subunit to the nucleocapsid template.The nucleocapsid-bound viral polymerase forms an RNA synthesis chamber inwhich the genomic RNA is temporarily unencapsidated to allow the catalyticregion of the polymerase to use it as the template for RNA synthesis. As theviral polymerase moves along the nucleocapsid, the previous unveiled regionof the RNA genome is reencapsidated, and the next segment becomes avail-able for RNA synthesis. The P dimer maintains the association of the viralpolymerase with the nucleocapsid throughout the process of RNA synthesis.Three successive steps are labeled A, B, and C.

Table 2. Data collection and refinement statistics

Complex N290-Pctd N/RNA-Pctd

Space group P21212 P21212Unit cell, Å

a 170.6 166.5b 234.5 235.2c 95.0 96.1

Resolution 30.0–2.70 30.0–3.50High-resolution bin 2.80–2.70 3.63–3.50Reflections (unique/total) 77,814/228,906 48,185/316,345Completeness, % 73.1 (44.9) 98.3 (88.4)I/�I 20.79 (2.10) 21.48 (4.31)Rmerge 0.073 (0.419) 0.116 (0.434)Rcryst 0.263 0.258Rfree 0.296 0.323Mean B value, Å2 74.58 28.09Model

No. of atoms 19,495 19,455No. of residues 2,460 2,460No. of nucleic acid bases 0 45

rmsdBonds, Å 0.006 0.008Angles, ° 0.957 1.179

RamachandranFavored, % 98.6 96.0Allowed, % 100.0 100.0

Values in parentheses represent values within high-resolution shells.

Green and Luo PNAS � July 14, 2009 � vol. 106 � no. 28 � 11717

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

June

12,

202

0

merase complex are required to fully understand the directinteraction of L with the P-bound template.

MethodsWild-type VSV N/RNA, N (Ser-2903Trp) mutant (N290), and the PCTD residues183–265 were expressed in Escherichia coli and purified as described previously(refs. 6, 15, and 18, respectively). Purified protein samples of N/RNA, N290, andPCTD [or selenomethionine-substituted PCTD (Se-PCTD)] were concentrated to 11, 9,and 12 mg/mL. N (or N290) was mixed with PCTD (or Se-PCTD) at a 1:1.2 molar ratio.N290/Se-PCTD crystals were grown by the hanging drop vapor diffusion method in24-well VDX plates (Hampton Research) at 22 °C (32).

Crystals of the N/RNA–PCTD complex were grown by the hanging dropmethod at 4 °C. Crystallization drops were formed by mixing equal volumes ofprotein with reservoir solution containing 7% PEG 4000, 250 mM NaCl, and100 mM citrate buffer, pH 5.6. This crystal form was cryoprotected withreservoir solution containing a final concentration of 20% PEG-4000 (Hamp-ton Research) and 20% glycerol. This crystal form belonged to the orthor-hombic space group P21212. Maximal crystal growth occurred within 2 weeks.Data were collected at the Stanford Synchrotron Radiation Lightsource (SSRL)beamline 11-1 at a wavelength of 1.0 Å, with an oscillation angle of 0.3° andcrystal-to-detector distance of 450 mm on a Mar 325 CCD (Marresearch)detector at cryotemperature.

The N290–Se-PCTD protein complex was mixed with a 1:1 volume equivalent ofreservoir solution consisting of 0.8–1.0 M K/Na tartrate, 200 mM NaCl, and 100mM imidazole buffer, pH 8.0. Orthorhombic crystals grew within 1 to 2 weeks.Before data collection, crystals were cryoprotected stepwise to a final solutioncontaining reservoir solution plus 20% glycerol and were flash frozen in liquidnitrogen. Data were collected at the South East Regional Collaborative AccessTeam (SER-CAT) BL22-ID at Advanced Photon Source (APS) at a wavelength of0.94 Å, with an oscillation angle of 0.3° and crystal-to-detector distance of 450mm on a MAR 325 CCD (Marresearch) detector at cryotemperature.

In all cases, raw intensity images were processed with the HKL2000 package(33), and structure factors were calculated with TRUNCATE (34) through theCCP4 program suite. Location of the N protein/RNA decamers in the N/RNA–PCTD complex was achieved by molecular replacement with COMO (35) using

the previously determined VSV N/RNA structure [Protein Data Bank (PDB) IDcode: 2gic] as the search model. An initial model of the PCTD was built withCOOT (36) in 2Fo � Fc maps of the orthorhombic data. The N290/Se-PCTD

structure was solved by molecular replacement with the previously deter-mined N290 model (PDB ID code: 2qvj). Crude placement of the PCTD domainswas performed with the aid of an intermediate VSV N/RNA–PCTD model usingthe superpositioning protocols in O (37). The asymmetric unit of each crystal ofthe N/RNA–PCTD or the N290–PCTD complexes contains one-half of the decamericnucleocapsid-like particle bound to 5 PCTD monomers. The stoichiometry of N:P issuch that each available N-binding site (10:10) is occupied when averaged overthe entire crystal. However, there is a reduced occupancy of the PCTD in certainpositions. The highest substitution occurs in alternating rather than adjacentN-binding positions. The lack of complete P substitution is not surprising, becauseN and P do not exist in a 1:1 ratio in mature virions (38). Rigid body refinement ofthe individual domains and extensions was carried out with REFMAC5 (39).Manual model building and real-space refinement were performed with COOT.Selenium sites were used to aid in confirmation of the constructed sequences.Restrained and TLS refinement were performed with REFMAC5. Refinementstatistics are shown in Table 2. No outliers were found in the Ramachandran plot.All superimpositions for determining the differences between bound and un-bound complexes as discussed in the text were carried out by using the secondarystructure mapping procedure in COOT (40).

ACKNOWLEDGMENTS. We thank the staff of the SER-CAT at the APS, ArgonneNational Laboratory, for assistance in data collection. We thank the generosityof the staff at the SSRL. Portions of this research were carried out at the SSRL,a national user facility operated by Stanford University on behalf of the U.S.Department of Energy, Office of Basic Energy Sciences. Use of the APS wassupported by the U.S. Department of Energy, Office of Science, Office of BasicEnergy Sciences, under Contract W-31-109-Eng-38. SER-CAT supporting insti-tutions may be found at www.ser-cat.org/members.html. The SSRL StructuralMolecular Biology Program is supported by the Department of Energy, Officeof Biological and Environmental Research, and by the National Institutes ofHealth (NIH), National Center for Research Resources, Biomedical TechnologyProgram, and the National Institute of General Medical Sciences. This workwas supported in part by NIH Grant AI050066.

1. Huang AS, Wagner RR (1966) Comparative sedimentation coefficients of RNA ex-tracted from plaque-forming and defective particles of vesicular stomatitis virus. J MolBiol 22:381–384.

2. Emerson SU, Wagner RR (1972) Dissociation and reconstitution of the transcriptase andtemplate activities of vesicular stomatitis B and T virions. J Virol 10:297–309.

3. Naito S, Ishihama A (1976) Function and structure of RNA polymerase from vesicularstomatitis virus. J Biol Chem 251:4307–4314.

4. Green TJ, Zhang X, Wertz GW, Luo M (2006) Structure of the vesicular stomatitis virusnucleoprotein-RNA complex. Science 313:357–360.

5. Albertini AA, et al. (2006) Crystal structure of the rabies virus nucleoprotein-RNA complex.Science 313:360–363.

6. Zhang X, Green TJ, Tsao J, Qiu S, Luo M (2008) Role of intermolecular interactions ofvesicular stomatitis virus nucleoprotein in RNA encapsidation. J Virol 82:674–682.

7. Dietzschold B, et al. (1987) Localization and immunological characterization of antigenicdomains of the rabies virus internal N and NS proteins. Virus Res 8:103–125.

8. Toriumi H, Kawai A (2005) Structural difference recognized by a monoclonal antibody#404–11 between the rabies virus nucleocapsid (NC) produced in virus infected cells andthe NC-like structures produced in the nucleoprotein (N) cDNA-transfected cells. MicrobiolImmunol 49:757–770.

9. Barik S, Banerjee AK (1992) Phosphorylation by cellular casein kinase II is essential fortranscriptional activity of vesicular stomatitis virus phosphoprotein P. Proc Natl Acad SciUSA 89:6570–6574.

10. Das T, et al. (1995) Role of cellular casein kinase II in the function of the phosphoprotein(P) subunit of RNA polymerase of vesicular stomatitis virus. J Biol Chem 270:24100–24107.

11. Gao Y, Lenard J (1995) Multimerization and transcriptional activation of the phosphopro-tein (P) of vesicular stomatitis virus by casein kinase-II. EMBO J 14:1240–1247.

12. Takacs AM, Barik S, Das T, Banerjee AK (1992) Phosphorylation of specific serine residueswithin the acidic domain of the phosphoprotein of vesicular stomatitis virus regulatestranscription in vitro. J Virol 66:5842–5848.

13. Chen JL, Das T, Banerjee AK (1997) Phosphorylated states of vesicular stomatitis virus Pprotein in vitro and in vivo. Virology 228:200–212.

14. PattnaikAK,etal. (1997)Phosphorylationwithintheamino-terminalacidicdomain Iofthephosphoprotein of vesicular stomatitis virus is required for transcription but not forreplication. J Virol 71:8167–8175.

15. Ding HT, Green TJ, Lu S, Luo M (2005) Crystal structure of the oligomerization domain ofthe phosphoprotein of vesicular stomatitis virus. J Virol 80:2808–2814.

16. Das SC, Pattnaik AK (2005) Role of the hypervariable hinge region of phosphoprotein P ofvesicular stomatitis virus in viral RNA synthesis and assembly of infectious virus particles.J Virol 79:8101–8112.

17. Gao Y, Lenard J (1995) Cooperative binding of multimeric phosphoprotein (P) of vesicularstomatitis virus to polymerase (L) and template: Pathways of assembly. J Virol 69:7718–7723.

18. Green TJ et al. (2000) Study of the assembly of vesicular stomatitis virus N protein: Role ofthe P protein. J Virol 74:9515–9524.

19. Gill DS, Chattopadhyay D, Banerjee AK (1986) Identification of a domain within thephosphoprotein of vesicular stomatitis virus that is essential for transcription in vitro. ProcNatl Acad Sci USA 83:8873–8877.

20. Takacs AM, Das T, Banerjee AK (1993) Mapping of interacting domains between thenucleocapsid protein and the phosphoprotein of vesicular stomatitis virus by using atwo-hybrid system. Proc Natl Acad Sci USA 90:10375–10379.

21. Emerson SU, Schubert M (1987) Location of the binding domains for the RNA polymeraseL and the ribonucleocapsid template within different halves of the NS phosphoprotein ofvesicular stomatitis virus. Proc Natl Acad Sci USA 84:5655–5659.

22. Das T, et al. (1997) Basic amino acid residues at the carboxy-terminal eleven amino acidregion of the phosphoprotein (P) are required for transcription but not for replication ofvesicular stomatitis virus genome RNA. Virology 238:103–114.

23. Paul PR, Chattopadhyay D, Banerjee AK (1988) The functional domains of the phospho-protein (NS) of vesicular stomatitis virus (Indiana serotype). Virology 166:350–357.

24. RibeiroEA, Jr,etal. (2008)Solutionstructureof theC-terminalnucleoprotein-RNAbindingdomain of the vesicular stomatitis virus phosphoprotein. J Mol Biol 382:525–538.

25. Masters PS, Banerjee AK (1988) Complex formation with vesicular stomatitis virus phos-phoprotein NS prevents binding of nucleocapsid protein N to nonspecific RNA. J Virol62:2658–2664.

26. Peluso RW (1988) Kinetic, quantitative, and functional analysis of multiple forms of thevesicular stomatitis virus nucleocapsid protein in infected cells. J Virol 62:2799–2807.

27. Peluso RW, Moyer SA (1988) Viral proteins required for the in vitro replication of vesicularstomatitis virus defective interfering particle genome RNA. Virology 162:369–376.

28. Emerson SU, Yu Y (1975) Both NS and L proteins are required for in vitro RNA synthesis byvesicular stomatitis virus. J Virol 15:1348–1356.

29. Gupta AK, Shaji D, Banerjee AK (2003) Identification of a novel tripartite complex involvedin replication of vesicular stomatitis virus genome RNA. J Virol 77:732–738.

30. Brunger AT, et al. (1998) Crystallography & NMR system: A new software suite formacromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54:905–921.

31. Chen M, Ogino T, Banerjee AK (2006) Mapping and functional role of the self-associationdomain of vesicular stomatitis virus phosphoprotein. J Virol 80:9511–9518.

32. Mcpherson A (1982) Preparation and Analysis of Protein Crystals (John Wiley, New York).33. Otwinowski Z, Minor W (1997) Processsing of x-ray diffraction data collected in oscillation

mode. Methods in Enzymology, eds Carter CW, Sweet RM (Academic, New York), Vol 276,pp 307–326.

34. French S, Wilson K (1978) On the treatment of negative intensity observations. ActaCrystallogr A 34:517–525.

35. JoglG,TaoX,XuY,TongL(2001)COMO:Aprogramforcombinedmolecular replacement.Acta Crystallogr D Biol Crystallogr 57:1127–1134.

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

37. Jones TA, Zou JY, Cowan SW, Kjeldgaard M (1991) Improved methods for building proteinmodels in electron density maps and the location of errors in these models. Acta Crystal-logr A 47:110–119.

38. Thomas D, et al. (1985) Mass and molecular composition of vesicular stomatitis virus: Ascanning transmission electron microscopy analysis. J Virol 54:598–607.

39. Murshudov GN, Vagin AA, Dodson EJ (1997) Refinement of macromolecular structures bythe maximum-likelihood method. Acta Crystallogr D Biol Crystallogr 53:240–255.

40. Krissinel E, Henrick K (2004) Secondary-structure matching (SSM), a new tool for fastprotein structure alignment in three dimensions. Acta Crystallogr D Biol Crystallogr60:2256–2268.

41. DeLano WL (2002) The PyMOL User’s Manual (Delano Scientific, San Carlos, CA).42. Krissinel E, Henrick K (2007) Inference of macromolecular assemblies from crystalline state.

J Mol Biol 372:774–797.

11718 � www.pnas.org�cgi�doi�10.1073�pnas.0903228106 Green and Luo

Dow

nloa

ded

by g

uest

on

June

12,

202

0