Crystal Structure of the Oligomerization Domain of
http://www.100md.com
病菌学杂志 2006年第6期
Department of Microbiology, School of Medicine, University of Alabama at Birmingham, 1025 18th Street South, Birmingham, Alabama 35294
ABSTRACT
In the replication cycle of nonsegmented negative-strand RNA viruses, the viral RNA-dependent RNA polymerase (L) recognizes a nucleoprotein (N)-enwrapped RNA template during the RNA polymerase reaction. The viral phosphoprotein (P) is a polymerase cofactor essential for this recognition. We report here the 2.3--resolution crystal structure of the central domain (residues 107 to 177) of P from vesicular stomatitis virus. The fold of this domain consists of a hairpin, an helix, and another hairpin. The helix provides the stabilizing force for forming a homodimer, while the two hairpins add additional stabilization by forming a four-stranded sheet through domain swapping between two molecules. This central dimer positions the N- and C-terminal domains of P to interact with the N and L proteins, allowing the L protein to specifically recognize the nucleocapsid-RNA template and to progress along the template while concomitantly assembling N with nascent RNA. The interdimer interactions observed in the noncrystallographic packing may offer insight into the mechanism of the RNA polymerase processive reaction along the viral nucleocapsid-RNA template.
INTRODUCTION
Vesicular stomatitis virus (VSV) is an enveloped virus belonging to the family Rhabdoviridae in the order Mononegavirales. VSV contains a single, negative-sense RNA genome (2). As the prototype of nonsegmented negative-strand viruses, the 11,161-nucleotide-long VSV genome encodes five viral proteins: the glycoprotein (G), the matrix protein (M), the nucleoprotein (N), the RNA-dependent RNA polymerase large protein (L), and the phosphoprotein (P) (23). The G protein is anchored on the viral membrane surface and is responsible for virus entry into host cells (25). M is involved in VSV budding and assembly in the host cell (21). The nucleocapsid protein (N) enwraps the viral genome that is associated with L and P to form the ribonucleoprotein core complex (N-RNA) (3, 4).
The N-RNA forms a supercoil structure beneath the viral envelope. Upon entering the host cell, the N-RNA is released and serves as the template for transcription and replication. The N-RNA template is recognized by the viral polymerase, L, and by the P protein that tethers L to the N-RNA template (13, 14). During the early stage of the virus replication cycle, the L and P proteins, which are carried into the host by the virion, initiate transcription of mRNAs of each viral gene. The relative abundance is highest for the N protein, followed by the P protein, simply because they are located near the 3' end of the viral genome, while the L protein, which is nearest the 5' end of the viral genome, is produced in the least abundance (38). When sufficient amounts of the N and P proteins are produced, the polymerase, L, switches from transcribing mRNAs to replicating the viral genome. In genomic replication, a complementary positive-strand genome is first synthesized from the negative-strand genomic template and is then used as the template for producing more copies of the negative-strand genome (16). The newly synthesized genome is then transferred to the budding site to assemble progeny virions. During the entire process, the viral genome, both negative and positive, is always assembled with the N protein, and the L/P polymerase does not bind the viral genomic RNA unless it is in the form of an N-RNA complex.
The P protein plays an essential role in many steps during the replication cycle. The P protein binds both the N and L proteins specifically, which makes it the determining factor in template recognition (15, 26). The P protein also binds the RNA-free form of N, called N0, and brings it to the replication complex so the newly synthesized genomic RNA can be enwrapped by N as genomic replication progresses (5, 27, 32). P also appears to be required for productive virion assembly (11). The P protein can be divided into three domains: the N-terminal domain, the central domain, and the C-terminal domain (12) (Fig. 1). The N-terminal domain contains the residues (Ser60, Thr62, and Ser64) that are phosphorylated by host enzymes and is indispensable for the transcriptional activity of the P protein (8, 30). The C-terminal domain also contains two phosphorylation sites (Ser226 and Ser227) and is necessary for optimal replication (8, 24). The C-terminal domain is the main association site for the N and L proteins (31). The C-terminal domain can still bind N even after the first two domains are deleted (22). Mutations in this domain abolish the replication activity of the P proteins (11). The central domain seems to be the oligomerization domain (12). Experimental data have shown that the ability to form a tetramer is required for the replication activity of the P protein (18, 19). Deletion mutations within the central domain do not reduce the production of the P protein but lead to loss of replication activities (11).
However, prior to this study it was not known how this domain in the VSV P protein oligomerizes. The only known oligomerization domain structure for a P protein is that of the Sendai virus (SeV) (35). In that structure, this domain forms a homotetrameric coiled coil. The main feature is a paralleled coiled coil of a long, 65-residue helix. This structure strategically places the L and N binding domains on opposite ends. The VSV P protein does not have any significant sequence homology with SeV P, despite their functional similarities. Therefore, we began our efforts to determine the crystal structure of this domain. Our proteolysis study showed that the central domain is comprised of residues 107 to 177 and is readily crystallized (12). Facilitated by mutation of Leu139 to methionine, the crystal structure of the P central domain was solved by the single-wavelength anomalous diffraction (SAD) method in this study. The core biological unit of the central domain is a homodimer, tightly associated by interactions between a pair of central helices in each molecule, and one four-stranded sheet, which is formed by two strands from each monomer as a result of domain swapping, on each side of the paired helices. The intermolecular interactions are extensive, burying 4,514 2, and contain packing of both hydrophobic residues and hydrogen bonds. In the crystallographic asymmetric unit, there are three independent dimers. The sheets in the neighboring dimers join together to form an extended eight-stranded sheet. This interdimer interaction suggests a unique tetramer that can be constructed with two dimers side by side. In the tetramer, the twofold axis of each dimer is 120° apart. The mutagenesis data on the RNA synthesis activity in this P region can be explained by the dimer structure. Suggestions for the mechanism of viral RNA polymerization, based on the structure of the P central domain, are discussed below.
MATERIALS AND METHODS
Production and crystallization of the selenium-incorporated protein. The details of the cloning, expression, purification, and crystallization of the P protein fragment encoding amino acids 107 to 177 (named P3) were described in our previous work (12). Since the P3 crystals were grown from a comparatively high concentration of ammonium sulfate, which usually retards binding of heavy atoms to protein molecules, a leucine residue within the P3 sequence was chosen to be mutated into methionine to facilitate structure determination (Fig. 1) (20). The secondary-structure prediction of the P3 protein was performed with Jpred (10), which indicated that the sequence (residues 134 to 148) of P3 may form an helix. Leu139, predicted to be located in the hydrophobic part of the helix, was the residue most suited for mutation into methionine for generating a mutated P3 protein (referred to as P3mut139), since methionine is likely to maintain the hydrophobic interaction of leucine if there is any. The mutation was completed by use of the expression clone P3 and the QuikChange XL site-directed mutagenesis kit (Stratagene). The expression and purification of the P3mut139 protein proceeded according to the protocol established for the wild-type protein as described in reference 12, with changes according to the protocol for seleno-methionine protein production. The resultant recombinant protein has the amino acid sequence shown in Fig. 1, which allowed us to prepare a seleno-methionine-substituted protein for crystallization. The elution profile of the recombinant P3mut139, purified by the gel filtration column Superdex-75 (Amersham Biosciences, Sweden), suggested that it may form an oligomer in the buffer (20 mM HEPES, 150 mM sodium chloride [pH 8.0]). Confirmation of selenium incorporation by the P3mut139 protein was performed by electrospray-time-of-flight mass spectrometry at the UAB Comprehensive Cancer Center Mass Spectrometry Shared Facility. Crystals of P3mut139 were obtained at 291 K with the hanging-drop method by mixing equal volumes of reservoir buffer (2.65 to 2.73 M ammonium sulfate, 100 mM ammonium nitrate, 7 to 8% ethylene glycol) and 25 mg/ml protein.
Data collection and processing. Crystals were cryoprotected in reservoir buffer in which ethylene glycol was added up to a final concentration of 15% before flash-freezing of the crystals in liquid nitrogen. Crystals belong to space group P41212 with unit cell dimensions a = b = 74.39 and c = 157.31 , with six molecules per asymmetric unit. Diffraction data were collected at beamline 22-ID (or 22-BM) in the facilities of the South East Regional Collaborative Access Team at the Advanced Photon Source, Argonne National Laboratory, using a MAR 300 CCD (or MAR 225 CCD) area detector. Beams with series of energy from 12,661.47 to 12,663.00 eV were chosen to collect single-wavelength anomalous diffraction data. Integration, scaling, and merging of data were performed with the HKL2000 package (29) (Table 1).
Structure determination. The crystal structure of P3mut139 (residues 107 to 177) was determined by the SAD method (Table 1). Seven Se sites were determined with SOLVE (36), and solvent flattening was performed with RESOLVE (37). The initial data set for phasing was Se-Met1. Iterative model building and density modification were performed with the PHENIX package (1). The resulting electron density maps were sufficient to trace the core of each of the six molecules in the asymmetric unit.
Refinement of the atomic coordinates was carried out with program CNS (6) and data set Se-Met2. With this initial refined model, the atom update and refinement procedure of Arp/Warp was used to further improve the electron density maps (33). Up to seven residues were added at the C terminus in the improved map. The final refined structure resulted in the statistics presented in Table 2. The model was refined against the more complete data set Se-Met2. The refined model contains the core residues 109 to 170 in each of the six molecules; residue –1 and 107 to 108 additional ordered residues at the N terminus (where residue –1 is the last residue of the histidine tag); and 171 to 177 additional ordered residues at the C terminus in different molecules. These additional ordered residues were stabilized by crystal packing or interdimer interactions. Eighty-four water molecules were also included in the final refined model. The protein data bank identification number for the coordinates is 2FQM.
RESULTS
Structure of the central domain of P. The central domain of the P protein of VSV (serotype Indiana) is comprised of amino acids 107 to 177. The recombinant protein that had been crystallized includes four additional amino acids at the N terminus of the central domain as a result of tag removal by thrombin (12). The secondary-structure elements in each domain include 1 (residues 112 to 116), 2 (residues 119 to 123), the helix (residues 131 to 151), and a one-turn 310 helix (residues 154 to 156) followed by 3 (residues 158 to 160 and 162) and 4 (residues 165 to 167 and 169) (Fig. 2). 1/2 and 3/4 can be described as typical -hairpin motifs. Five residues in 1 form main-chain hydrogen bonds with five residues in the antiparallel 2, with four residues (116 to 119) at the tip of the hairpin forming a type I turn. Similarly, five residues in 3 form main-chain hydrogen bonds with the antiparallel 4, with a loop at the tip of the hairpin formed by residues 162 to 165. The tips in each hairpin appear to be more flexible than the main body of the structure unless stabilized by crystal packing contacts (see below).
Two central-domain molecules form a dimer in which the helices from each molecule interact with each other in a parallel orientation. The major dimer interactions are located at the N-terminal end of the helix. Trp138 is at the core of this hydrophobic interaction. Side chains of Leu126, Pro127, and Leu130 on the loop leading to the helix are packed with the side chain of Trp138. This cluster of hydrophobic side chains keeps the side chain of Trp138 in a conformation that is perpendicular to the side chain of Trp138 in the other helix of the dimer. This causes the side chain of Trp138 in the second helix to have a different conformation, resulting in the formation of a hydrogen bond between the N7 of the Trp138 side chain in the second helix and the hydroxyl group of the Thr141 side chain in the first helix. The Trp138 side chain in the second helix has a slightly different packing arrangement, with the side chains of Leu126, Pro127, and Leu130 in its own molecule. The interactions of side chains are therefore asymmetric in this region even though the two helices are oriented in parallel. Farther down the helix, the side chains of Ile142 and Val145 have more-symmetric interactions between the two helices.
Interactions between the strands provide additional force to stabilize the dimer. 2 in one molecule forms main-chain hydrogen bonds with 4 in the other molecule in an antiparallel fashion. A total of seven hydrogen bonds are formed by this interaction. 1/2 in one molecule and 3/4 in the other molecule therefore constitute a four-stranded antiparallel sheet. This hydrogen bond network in the sheet may provide forces to stabilize the dimer that are as substantial as the hydrophobic interactions between the helices. Furthermore, the side chains of the four-stranded sheet also offer hydrophobic interactions on each side of the helices. The side chains of Leu113, Lys120, Leu122, and Leu124 from 1/2 interact with the side chains of Leu140, Ile142, Ala144, and Val145 in the second helix, whereas the side chains of Phe159, Ala161, Val166, and Ile168 in 3/4 interact with the side chains of Trp138, Met139 (Leu139 in the wild type), Ile142, Lys143, and Val146 in the second helix. A number of hydrophobic residues, including Trp138, Ile142, and Val145, participate in both helix-helix and sheet-helix interactions. The side chains of Trp108, Pro111, Pro127, Trp152, and Leu154 on loops are also packed with the hydrophobic side chains involved in the sheet-helix interactions.
In addition, there are three hydrogen bonds formed between the side chain of Glu128 and the side chain of Gln137 in the second molecule, between the side chain of Gln134 and the main-chain amide nitrogen, and between the side chain of Gln137 and the carbonyl oxygen of Glu128 in the second molecule. There is one intermolecular salt bridge between the side chain of Glu164 and the side chain of Lys135 in the second molecule.
Within this compact dimer, the intermolecular interactions are extensive. There are hydrophobic interactions between the helices, as well as between the helix and the sheets on both sides of the helices. Hydrogen bonds are formed between the hairpins, to form a continuous sheet between the two molecules. In addition, there are hydrophobic interactions, hydrogen bonds, and charge-charge interactions contributed by residues on the loops. The total surface area buried by the interactions between the two molecules is 4,514 2 when the dimer is formed. For the protein molecule of 78 residues, the extensive interactions may provide strong forces to make this dimer exceptionally stable, which may be the essential requirement for the functions of the central domain within the P protein.
Interdimer interaction. In the asymmetric unit of the crystal, there are three individual dimers (Fig. 3). The 1 strand in one dimer forms main-chain hydrogen bonds with the 1 strand in the neighboring dimer in an antiparallel fashion, resulting in an eight-stranded antiparallel sheet between two neighboring dimers. There are four hydrogen bonds formed between residues 113 and 115 in the first 1 strand and residues 113 and 115 in the second 1 strand. This -sheet interaction extends throughout the crystal so that long strings of the dimers are constructed by interdimer interactions involving the twofold crystallographic symmetry. In addition, the main-chain amide nitrogen and carbonyl oxygen of Gln110 in the dimer-related molecule form two hydrogen bonds with the side chain of Gln147 from the neighboring helix, while the side-chain nitrogen of the same Gln110 forms a hydrogen bond with the carbonyl oxygen of residue 140, also from the neighboring helix. These additional hydrogen bonds further stabilize the interdimer interactions. The twofold axis in each dimer is approximately perpendicular to the central axis of the dimer string. The dimer axes are rotated 120° from each other about the central axis. Three dimers, therefore, make one complete rotation in the dimer string. Both the N and C termini are quite flexible in each dimer. The first ordered residue ranges from –1 (one leftover tag residue prior to residue 107) to 109 at the N terminus. Residues prior to residue 109 become ordered only when they are stabilized by interactions with residues in a neighboring dimer. On the other hand, the C-terminal residues become ordered by extending the last residue (Val177) into a hydrophobic pocket in the adjacent dimer strings. The hydrophobic pocket is formed by the side chains of Leu113 and Lys120 in one molecule and of Trp152 in the other dimer-related molecule. In full-length P, the regions corresponding to the termini in the central domain fragment are likely to be very flexible, to allow a wide range of movements by the N-terminal and C-terminal domains as required by their functions (see Discussion). The ordered termini observed in this crystal structure are the result of crystal packing.
Comparison with the oligomerization domain of the SeV phosphoprotein. The phosphoprotein of SeV has functions very similar to that of VSV. The SeV P is 568 amino acids, twice the size of VSV P (265 amino acids [serotype Indiana]). The N-terminal and C-terminal domains of SeV P bind N, while the central domain (residues 320 to 446) is responsible for oligomerization of a tetramer, as shown in the crystal structure (35). A 96--long helix, formed by 65 residues from each monomer contributes to the central tetrameric coiled coil in the tetramer. The coiled coil covers most of the tetramer, stabilizing interactions, while two short helices at the N terminus of the long helix have a small number of interactions with the neighboring monomers. The oligomerization domain of VSV P forms a stable dimer, using a similar, but shorter, helix. The intermolecular -sheet interactions by the hairpins do not have any similarity with the tetramer of the SeV P oligomerization domain. It has been suggested that the SeV P tetramer cartwheels along the N-RNA template while the L polymerase reads the nucleotide. P steps on one N monomer at a time, dictating that the polymerase complex moves 6 nucleotides each time. It has also been shown that VSV P must be capable of forming a tetramer in order to perform its RNA polymerase-related activities (18, 19). The tetramer unit in this crystal structure looks like a twisted sheet instead of a coil wheel (Fig. 3). Despite the structural differences between the two proteins, the essential N-RNA binding domains (the C-terminal domains) of both P proteins are at similar locations near the C terminus of the oligomerization domain. We suggest that the P protein associates with L as a dimer, as a functional unit. During transcription, the L/P unit recruits host factors and gets loaded onto the N-RNA template to transcribe mRNAs, whereas during replication, a P dimer-N0 complex replaces the host factors and forms a 4P:1N:1L replicase unit. Two P dimers may form a tetramer involving the interdimer sheet. The C-terminal domains of the first P dimer act as two feet to walk along the N-RNA template as RNA synthesis progresses.
Functions of the central domain in RNA polymerization. The C-terminal domain of VSV P plays an essential role in binding the N-RNA template and in viral replication (31). The N-terminal domain mainly influences transcription (8, 30). The central domain positions these two functional domains at the opposite ends and brings multiple copies of P (such as a tetramer) in close proximity through oligomerization. It is not clear whether the central domain interacts directly with any other viral proteins. In a recent study, a number of insertion and deletion mutations were introduced into the central domain and their effects on P activities were studied (11). Four Tn5-based insertion mutations of 16 amino acids were introduced at residues 136, 148, 159, and 167, respectively. The insertions at residues 136, 148, and 159 greatly reduced the replication activity, with the insertion at residue 136 being the most severe (about 90% reduction), while the insertion at residue 167 reduced the replication activity by about 40%. Only the insertion at residue 136 significantly reduced the transcription activity. Residue 167 is 3 amino acids away from the last ordered residue at the C terminus of the central domain. The inserted amino acids should have minimum effects on the integrity of the dimer structure of the central domain, probably simply extending the flexible region linking the C-terminal domain to the central domain. On the other hand, residue 136 is at the center of the helix-helix interaction in the dimer. Insertions should be strongly destructive to the P structure, resulting in a significant loss of both replication and transcription activity. Similarly, a deletion from residues 130 to 151 (two-thirds of the helix) significantly reduced replication and transcription activity, while various deletions between residues 140 and 201 had only limited reduction effects on replication and transcription activity. Residues 107 to 139 were able to retain substantial intermolecular interactions through the hydrophobic core around residue Trp138, as well as the potential tetramer interactions by the 1 strands (Fig. 4). This stripped-down dimer may still preserve the basic oligomerization function of the central domain. It is therefore plausible to postulate that the essential role of the central domain is to maintain the dimeric structure to properly position the N-terminal and C-terminal domains for their functional activities. The central domain probably would not have significant specific, direct interactions with any other viral proteins. The fact that insertions and deletions that disrupted most of the helix-helix interactions, as well as intermolecular sheets, reduced but did not completely diminish replication and transcription activity supports the suggestion that the central domain has the isolated function of oligomerization.
DISCUSSION
The crystal structure of the P central domain clearly defined the content of this domain and its architecture. The extensive intermolecular helix-helix interaction and the domain swapping between the two hairpins strongly suggest that the basic functional unit of the active P protein is always in the dimeric form. A trimer form of P suggested in an earlier report (17) could not exist because an oligomer of P could contain only multiples of two copies of P. This notion is particularly consistent with the observation that the P-N0 complex contains at least a P dimer in rabies virus (28). Taking into account the extensive intermolecular interactions in the assembled N-RNA genomic complex (unpublished data), it is likely that the P-N0 complex being recruited into the replication site for encapsidation of the nascent RNA is a 2:1 P dimer/N0 heterotrimer. The C-terminal domain of P primarily functions as a carrier to bind N0 in this complex. N0 can be added only one at a time in the form of a P-N0 complex, and it undergoes a significant conformational change once it becomes associated with RNA. The assembled N-RNA is hard to disrupt when it is formed (22).
The P central domain not only forms a stable dimer but also shows a potential tetrameric form, as observed in the noncrystallographic interaction. The ability to form a tetramer has been shown to be required for VSV replication activity (18, 19). The initiation site for viral replication has been shown to be separate from that for transcription (34, 39), and the replicase was shown to be formed by L/P/N with unspecified stoichiometry while the transcriptase is composed of L/P and host factors (34). The P molecules in the replicase must interact with both the N-RNA template and the additional N protein in the replicase complex. Our previous results showed that the N proteins are aligned in parallel when assembled with RNA (9). Since P also forms a parallel dimer, its twofold relationship prohibits the two C-terminal domains from binding N molecules that are not related by a twofold axis. However, the C-terminal domain of P is linked to the central domain by a very flexible linker. There is a possibility that one C-terminal domain is rearranged in the dimer in order to bind N in an orientation parallel to the first C-terminal domain. This type of interaction would introduce significant constraints on the P structure, although the possibility cannot be ruled out. The P dimer in the replicase complex may be essential for the processivity of the replicase along the N-RNA template. The two C-terminal domains in the P dimer may act as two feet to allow the complex to walk along the N-RNA template through an unclear mechanism. Since the formation of a P tetramer is required for replication activity, the tetramer structure observed in the noncrystallographic asymmetric unit may exist in the replicase complex. There are at least four C-terminal domains of P available in the replicase complex to bind both the N-RNA template and the additional N protein.
At the early stage of virus infection, the L and P proteins are carried into the cytoplasm by the ribonucleoprotein core. Since the virion genome does not carry any N0, there is no P dimer-N0 complex present to form the active replicase complex. It is likely that the L protein is associated with P proteins and host factors to form the transcriptase. The first isolation of the transcriptase complex, by Qanungo et al. (34), showed that a number of host factors, such as translation elongation factor 1, heat shock protein 60, and guanylyltransferase, are present but that the N protein is not. The dimeric P proteins in the transcriptase are not required to recruit N0. Instead, the P proteins mainly function to tether the transcriptase to the N-RNA template and to play a role in processive transcription.
ACKNOWLEDGMENTS
We thank the staff of the South East Regional Collaborative Access Team (SER-CAT) at the Advanced Photon Source for assistance in data collection.
Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under contract no. W-31-109-Eng-38. This work is supported in part by a grant to M.L. (NIH AI050066).
REFERENCES
Adams, P. D., K. Gopal, R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger, A. J. McCoy, N. W Moriarty, R. K. Pai, R. J. Read, T. D. Romo, J. C. Sacchettini, N. K. Sauter, L. C. Storoni, and T. C. Terwilliger. 2004. Recent developments in the PHENIX software for automated crystallographic structure determination. J. Synchrotron Radiat. 11:53-55.
Baltimore, D., A. S. Huang, and M. Stampfer. 1970. Ribonucleic acid synthesis of vesicular stomatitis virus, II. An RNA polymerase in the virion. Proc. Natl. Acad. Sci. USA 66:572-576.
Banerjee, A. K. 1987. Transcription and replication of rhabdoviruses. Microbiol. Rev. 51:66-87.
Banerjee, A. K., and S. Barik. 1992. Gene expression of vesicular stomatitis virus genome RNA. Virology 188:417-428.
Bell, J. C., E. G. Brown, D. Takayesu, and L. Prevec. 1984. Protein kinase activity associated with immunoprecipitates of the vesicular stomatitis virus phosphoprotein NS. Virology 132:229-238.
Brunger, A. T., P. D. Adams, and L. M. Rice. 1997. New applications of simulated annealing in X-ray crystallography and solution NMR. Structure 5:325-336.
Carson, M. 1997. Ribbons. Methods Enzymol. 277:493-505.
Chen, J. L., T. Das, and A. K. Banerjee. 1997. Phosphorylated states of vesicular stomatitis virus P protein in vitro and in vivo. Virology 228:200-212.
Chen, Z., T. J. Green, M. Luo, and H. Li. 2004. Visualizing the RNA molecule in the bacterially expressed vesicular stomatitis virus nucleoprotein-RNA complex. Structure 12:227-235.
Cuff, J. A., and G. J. Barton. 1999. Evaluation and improvement of multiple sequence methods for protein secondary structure prediction. Proteins 34:508-519.
Das, S. C., and A. K. Pattnaik. 2005. Role of the hypervariable hinge region of phosphoprotein P of vesicular stomatitis virus in viral RNA synthesis and assembly of infectious virus particles. J. Virol. 79:8101-8112.
Ding, H., T. J. Green, and M. Luo. 2004. Crystallization and preliminary X-ray analysis of a proteinase-K-resistant domain within the phosphoprotein of vesicular stomatitis virus (Indiana). Acta Crystallogr. D 60:2087-2090.
Emerson, S. U., and R. R. Wagner. 1972. Dissociation and reconstitution of the transcriptase and template activities of vesicular stomatitis B and T virions. J. Virol. 10:297-309.
Emerson, S. U., and R. R. Wagner. 1973. L protein requirement for in vitro RNA synthesis by vesicular stomatitis virus. J. Virol. 12:1325-1335.
Emerson, S. U., and Y. Yu. 1975. Both NS and L proteins are required for in vitro RNA synthesis by vesicular stomatitis virus. J. Virol. 15:1348-1356.
Ermine, A., and A. Flamand. 1977. RNA syntheses in BHK21 cells infected by rabies virus. Ann. Microbiol. (Paris) 128A:477-488.
Gao, Y., N. J. Greenfield, D. Z. Cleverley, and J. Lenard. 1996. The transcriptional form of the phosphoprotein of vesicular stomatitis virus is a trimer: structure and stability. Biochemistry 35:14569-14573.
Gao, Y., and J. Lenard. 1995. Multimerization and transcriptional activation of the phosphoprotein (P) of vesicular stomatitis virus by casein kinase-II. EMBO J. 14:1240-1247.
Gao, Y., and J. Lenard. 1995. Cooperative binding of multimeric phosphoprotein (P) of vesicular stomatitis virus to polymerase (L) and template: pathways of assembly. J. Virol. 69:7718-7723.
Gassner, N. C., and B. W. Matthews. 1999. Use of differentially substituted selenomethionine proteins in X-ray structure determination. Acta Crystallogr. D 55:1967-1970.
Gaudier, M., Y. Gaudin, and M. Knossow. 2002. Crystal structure of vesicular stomatitis virus matrix protein. EMBO J. 21:2886-2892.
Green, T. J., S. Macpherson, S. Qiu, J. Lebowitz, G. W. Wertz, and M. Luo. 2000. Study of the assembly of vesicular stomatitis virus N protein: role of the P protein. J. Virol. 74:9515-9524.
Huang, A. S., and R. R. Wagner. 1966. Comparative sedimentation coefficients of RNA extracted from plaque-forming and defective particles of vesicular stomatitis virus. J. Mol. Biol. 22:381-384.
Hwang, L. N., N. Englund, T. Das, A. K. Banerjee, and A. K. Pattnaik. 1999. Optimal replication activity of vesicular stomatitis virus RNA polymerase requires phosphorylation of a residue(s) at carboxy-terminal domain II of its accessory subunit, phosphoprotein P. J. Virol. 73:5613-5620.
Jeetendra, E., C. S. Robison, L. M. Albritton, and A. Whitt. 2002. The membrane-proximal domain of vesicular stomatitis virus G protein functions as a membrane fusion potentiator and can induce hemifusion. J. Virol. 76:12300-12311.
Masters, P. S., and A. K. Banerjee. 1986. Phosphoprotein NS of vesicular stomatitis virus: phosphorylated states and transcriptional activities of intracellular and virion forms. Virology 154:259-270.
Masters, P. S., and A. K. Banerjee. 1988. Resolution of multiple complexes of phosphoprotein NS with nucleocapsid protein N of vesicular stomatitis virus. J. Virol. 62:2651-2657.
Mavrakis, M., A. A. McCarthy, S. Roche, D. Blondel, and R. W. Ruigrok. 2004. Structure and function of the C-terminal domain of the polymerase cofactor of rabies virus. J. Mol. Biol. 343:819-831.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276:307-326.
Pattnaik, A. K., L. Hwang, T. Li, N. Englund, M. Mathur, T. Das, and A. K. Banerjee. 1997. Phosphorylation within the amino-terminal acidic domain I of the phosphoprotein of vesicular stomatitis virus is required for transcription but not for replication. J. Virol. 71:8167-8175.
Paul, P. R., D. Chattopadhyay, and A. K. Banerjee. 1988. The functional domains of the phosphoprotein (NS) of vesicular stomatitis virus (Indiana serotype). Virology 166:350-357.
Peluso, R. W. 1988. Kinetic, quantitative, and functional analysis of multiple forms of the vesicular stomatitis virus nucleocapsid protein in infected cells. J. Virol. 62:2799-2807.
Perrakis, A., M. Harkiolaki, K. S. Wilson, and V. S. Lamzin. 2001. ARP/wARP and molecular replacement. Acta Crystallogr. D 57:1445-1450.
Qanungo, K. R., D. Shaji, M. Mathur, and A. K. Banerjee. 2004. Two RNA polymerase complexes from vesicular stomatitis virus-infected cells that carry out transcription and replication of genome RNA. Proc. Natl. Acad. Sci. USA 101:5952-5957.
Tarbouriech, N., J. Curran, R. W. Ruigrok, and W. P. Burmeister. 2000. Tetrameric coiled coil domain of Sendai virus phosphoprotein. Nat. Struct. Biol. 7:777-781.
Terwilliger, T. C. 2000. Maximum likelihood density modification. Acta Crystallogr. D 56:965-972.
Terwilliger, T. C., and J. Berendzen. 1999. Automated MAD and MIR structure solution. Acta Crystallogr. D 55:849-861.
Villarreal, L. P., M. Breindl, and J. J. Holland. 1976. Determination of molar ratios of vesicular stomatitis virus induced RNA species in BHK21 cells. Biochemistry 15:1663-1667.
Whelan, S. P., and G. W. Wertz. 2002. Transcription and replication initiate at separate sites on the vesicular stomatitis virus genome. Proc. Natl. Acad. Sci. USA 99:9178-9183.(Haitao Ding, Todd J. Gree)
ABSTRACT
In the replication cycle of nonsegmented negative-strand RNA viruses, the viral RNA-dependent RNA polymerase (L) recognizes a nucleoprotein (N)-enwrapped RNA template during the RNA polymerase reaction. The viral phosphoprotein (P) is a polymerase cofactor essential for this recognition. We report here the 2.3--resolution crystal structure of the central domain (residues 107 to 177) of P from vesicular stomatitis virus. The fold of this domain consists of a hairpin, an helix, and another hairpin. The helix provides the stabilizing force for forming a homodimer, while the two hairpins add additional stabilization by forming a four-stranded sheet through domain swapping between two molecules. This central dimer positions the N- and C-terminal domains of P to interact with the N and L proteins, allowing the L protein to specifically recognize the nucleocapsid-RNA template and to progress along the template while concomitantly assembling N with nascent RNA. The interdimer interactions observed in the noncrystallographic packing may offer insight into the mechanism of the RNA polymerase processive reaction along the viral nucleocapsid-RNA template.
INTRODUCTION
Vesicular stomatitis virus (VSV) is an enveloped virus belonging to the family Rhabdoviridae in the order Mononegavirales. VSV contains a single, negative-sense RNA genome (2). As the prototype of nonsegmented negative-strand viruses, the 11,161-nucleotide-long VSV genome encodes five viral proteins: the glycoprotein (G), the matrix protein (M), the nucleoprotein (N), the RNA-dependent RNA polymerase large protein (L), and the phosphoprotein (P) (23). The G protein is anchored on the viral membrane surface and is responsible for virus entry into host cells (25). M is involved in VSV budding and assembly in the host cell (21). The nucleocapsid protein (N) enwraps the viral genome that is associated with L and P to form the ribonucleoprotein core complex (N-RNA) (3, 4).
The N-RNA forms a supercoil structure beneath the viral envelope. Upon entering the host cell, the N-RNA is released and serves as the template for transcription and replication. The N-RNA template is recognized by the viral polymerase, L, and by the P protein that tethers L to the N-RNA template (13, 14). During the early stage of the virus replication cycle, the L and P proteins, which are carried into the host by the virion, initiate transcription of mRNAs of each viral gene. The relative abundance is highest for the N protein, followed by the P protein, simply because they are located near the 3' end of the viral genome, while the L protein, which is nearest the 5' end of the viral genome, is produced in the least abundance (38). When sufficient amounts of the N and P proteins are produced, the polymerase, L, switches from transcribing mRNAs to replicating the viral genome. In genomic replication, a complementary positive-strand genome is first synthesized from the negative-strand genomic template and is then used as the template for producing more copies of the negative-strand genome (16). The newly synthesized genome is then transferred to the budding site to assemble progeny virions. During the entire process, the viral genome, both negative and positive, is always assembled with the N protein, and the L/P polymerase does not bind the viral genomic RNA unless it is in the form of an N-RNA complex.
The P protein plays an essential role in many steps during the replication cycle. The P protein binds both the N and L proteins specifically, which makes it the determining factor in template recognition (15, 26). The P protein also binds the RNA-free form of N, called N0, and brings it to the replication complex so the newly synthesized genomic RNA can be enwrapped by N as genomic replication progresses (5, 27, 32). P also appears to be required for productive virion assembly (11). The P protein can be divided into three domains: the N-terminal domain, the central domain, and the C-terminal domain (12) (Fig. 1). The N-terminal domain contains the residues (Ser60, Thr62, and Ser64) that are phosphorylated by host enzymes and is indispensable for the transcriptional activity of the P protein (8, 30). The C-terminal domain also contains two phosphorylation sites (Ser226 and Ser227) and is necessary for optimal replication (8, 24). The C-terminal domain is the main association site for the N and L proteins (31). The C-terminal domain can still bind N even after the first two domains are deleted (22). Mutations in this domain abolish the replication activity of the P proteins (11). The central domain seems to be the oligomerization domain (12). Experimental data have shown that the ability to form a tetramer is required for the replication activity of the P protein (18, 19). Deletion mutations within the central domain do not reduce the production of the P protein but lead to loss of replication activities (11).
However, prior to this study it was not known how this domain in the VSV P protein oligomerizes. The only known oligomerization domain structure for a P protein is that of the Sendai virus (SeV) (35). In that structure, this domain forms a homotetrameric coiled coil. The main feature is a paralleled coiled coil of a long, 65-residue helix. This structure strategically places the L and N binding domains on opposite ends. The VSV P protein does not have any significant sequence homology with SeV P, despite their functional similarities. Therefore, we began our efforts to determine the crystal structure of this domain. Our proteolysis study showed that the central domain is comprised of residues 107 to 177 and is readily crystallized (12). Facilitated by mutation of Leu139 to methionine, the crystal structure of the P central domain was solved by the single-wavelength anomalous diffraction (SAD) method in this study. The core biological unit of the central domain is a homodimer, tightly associated by interactions between a pair of central helices in each molecule, and one four-stranded sheet, which is formed by two strands from each monomer as a result of domain swapping, on each side of the paired helices. The intermolecular interactions are extensive, burying 4,514 2, and contain packing of both hydrophobic residues and hydrogen bonds. In the crystallographic asymmetric unit, there are three independent dimers. The sheets in the neighboring dimers join together to form an extended eight-stranded sheet. This interdimer interaction suggests a unique tetramer that can be constructed with two dimers side by side. In the tetramer, the twofold axis of each dimer is 120° apart. The mutagenesis data on the RNA synthesis activity in this P region can be explained by the dimer structure. Suggestions for the mechanism of viral RNA polymerization, based on the structure of the P central domain, are discussed below.
MATERIALS AND METHODS
Production and crystallization of the selenium-incorporated protein. The details of the cloning, expression, purification, and crystallization of the P protein fragment encoding amino acids 107 to 177 (named P3) were described in our previous work (12). Since the P3 crystals were grown from a comparatively high concentration of ammonium sulfate, which usually retards binding of heavy atoms to protein molecules, a leucine residue within the P3 sequence was chosen to be mutated into methionine to facilitate structure determination (Fig. 1) (20). The secondary-structure prediction of the P3 protein was performed with Jpred (10), which indicated that the sequence (residues 134 to 148) of P3 may form an helix. Leu139, predicted to be located in the hydrophobic part of the helix, was the residue most suited for mutation into methionine for generating a mutated P3 protein (referred to as P3mut139), since methionine is likely to maintain the hydrophobic interaction of leucine if there is any. The mutation was completed by use of the expression clone P3 and the QuikChange XL site-directed mutagenesis kit (Stratagene). The expression and purification of the P3mut139 protein proceeded according to the protocol established for the wild-type protein as described in reference 12, with changes according to the protocol for seleno-methionine protein production. The resultant recombinant protein has the amino acid sequence shown in Fig. 1, which allowed us to prepare a seleno-methionine-substituted protein for crystallization. The elution profile of the recombinant P3mut139, purified by the gel filtration column Superdex-75 (Amersham Biosciences, Sweden), suggested that it may form an oligomer in the buffer (20 mM HEPES, 150 mM sodium chloride [pH 8.0]). Confirmation of selenium incorporation by the P3mut139 protein was performed by electrospray-time-of-flight mass spectrometry at the UAB Comprehensive Cancer Center Mass Spectrometry Shared Facility. Crystals of P3mut139 were obtained at 291 K with the hanging-drop method by mixing equal volumes of reservoir buffer (2.65 to 2.73 M ammonium sulfate, 100 mM ammonium nitrate, 7 to 8% ethylene glycol) and 25 mg/ml protein.
Data collection and processing. Crystals were cryoprotected in reservoir buffer in which ethylene glycol was added up to a final concentration of 15% before flash-freezing of the crystals in liquid nitrogen. Crystals belong to space group P41212 with unit cell dimensions a = b = 74.39 and c = 157.31 , with six molecules per asymmetric unit. Diffraction data were collected at beamline 22-ID (or 22-BM) in the facilities of the South East Regional Collaborative Access Team at the Advanced Photon Source, Argonne National Laboratory, using a MAR 300 CCD (or MAR 225 CCD) area detector. Beams with series of energy from 12,661.47 to 12,663.00 eV were chosen to collect single-wavelength anomalous diffraction data. Integration, scaling, and merging of data were performed with the HKL2000 package (29) (Table 1).
Structure determination. The crystal structure of P3mut139 (residues 107 to 177) was determined by the SAD method (Table 1). Seven Se sites were determined with SOLVE (36), and solvent flattening was performed with RESOLVE (37). The initial data set for phasing was Se-Met1. Iterative model building and density modification were performed with the PHENIX package (1). The resulting electron density maps were sufficient to trace the core of each of the six molecules in the asymmetric unit.
Refinement of the atomic coordinates was carried out with program CNS (6) and data set Se-Met2. With this initial refined model, the atom update and refinement procedure of Arp/Warp was used to further improve the electron density maps (33). Up to seven residues were added at the C terminus in the improved map. The final refined structure resulted in the statistics presented in Table 2. The model was refined against the more complete data set Se-Met2. The refined model contains the core residues 109 to 170 in each of the six molecules; residue –1 and 107 to 108 additional ordered residues at the N terminus (where residue –1 is the last residue of the histidine tag); and 171 to 177 additional ordered residues at the C terminus in different molecules. These additional ordered residues were stabilized by crystal packing or interdimer interactions. Eighty-four water molecules were also included in the final refined model. The protein data bank identification number for the coordinates is 2FQM.
RESULTS
Structure of the central domain of P. The central domain of the P protein of VSV (serotype Indiana) is comprised of amino acids 107 to 177. The recombinant protein that had been crystallized includes four additional amino acids at the N terminus of the central domain as a result of tag removal by thrombin (12). The secondary-structure elements in each domain include 1 (residues 112 to 116), 2 (residues 119 to 123), the helix (residues 131 to 151), and a one-turn 310 helix (residues 154 to 156) followed by 3 (residues 158 to 160 and 162) and 4 (residues 165 to 167 and 169) (Fig. 2). 1/2 and 3/4 can be described as typical -hairpin motifs. Five residues in 1 form main-chain hydrogen bonds with five residues in the antiparallel 2, with four residues (116 to 119) at the tip of the hairpin forming a type I turn. Similarly, five residues in 3 form main-chain hydrogen bonds with the antiparallel 4, with a loop at the tip of the hairpin formed by residues 162 to 165. The tips in each hairpin appear to be more flexible than the main body of the structure unless stabilized by crystal packing contacts (see below).
Two central-domain molecules form a dimer in which the helices from each molecule interact with each other in a parallel orientation. The major dimer interactions are located at the N-terminal end of the helix. Trp138 is at the core of this hydrophobic interaction. Side chains of Leu126, Pro127, and Leu130 on the loop leading to the helix are packed with the side chain of Trp138. This cluster of hydrophobic side chains keeps the side chain of Trp138 in a conformation that is perpendicular to the side chain of Trp138 in the other helix of the dimer. This causes the side chain of Trp138 in the second helix to have a different conformation, resulting in the formation of a hydrogen bond between the N7 of the Trp138 side chain in the second helix and the hydroxyl group of the Thr141 side chain in the first helix. The Trp138 side chain in the second helix has a slightly different packing arrangement, with the side chains of Leu126, Pro127, and Leu130 in its own molecule. The interactions of side chains are therefore asymmetric in this region even though the two helices are oriented in parallel. Farther down the helix, the side chains of Ile142 and Val145 have more-symmetric interactions between the two helices.
Interactions between the strands provide additional force to stabilize the dimer. 2 in one molecule forms main-chain hydrogen bonds with 4 in the other molecule in an antiparallel fashion. A total of seven hydrogen bonds are formed by this interaction. 1/2 in one molecule and 3/4 in the other molecule therefore constitute a four-stranded antiparallel sheet. This hydrogen bond network in the sheet may provide forces to stabilize the dimer that are as substantial as the hydrophobic interactions between the helices. Furthermore, the side chains of the four-stranded sheet also offer hydrophobic interactions on each side of the helices. The side chains of Leu113, Lys120, Leu122, and Leu124 from 1/2 interact with the side chains of Leu140, Ile142, Ala144, and Val145 in the second helix, whereas the side chains of Phe159, Ala161, Val166, and Ile168 in 3/4 interact with the side chains of Trp138, Met139 (Leu139 in the wild type), Ile142, Lys143, and Val146 in the second helix. A number of hydrophobic residues, including Trp138, Ile142, and Val145, participate in both helix-helix and sheet-helix interactions. The side chains of Trp108, Pro111, Pro127, Trp152, and Leu154 on loops are also packed with the hydrophobic side chains involved in the sheet-helix interactions.
In addition, there are three hydrogen bonds formed between the side chain of Glu128 and the side chain of Gln137 in the second molecule, between the side chain of Gln134 and the main-chain amide nitrogen, and between the side chain of Gln137 and the carbonyl oxygen of Glu128 in the second molecule. There is one intermolecular salt bridge between the side chain of Glu164 and the side chain of Lys135 in the second molecule.
Within this compact dimer, the intermolecular interactions are extensive. There are hydrophobic interactions between the helices, as well as between the helix and the sheets on both sides of the helices. Hydrogen bonds are formed between the hairpins, to form a continuous sheet between the two molecules. In addition, there are hydrophobic interactions, hydrogen bonds, and charge-charge interactions contributed by residues on the loops. The total surface area buried by the interactions between the two molecules is 4,514 2 when the dimer is formed. For the protein molecule of 78 residues, the extensive interactions may provide strong forces to make this dimer exceptionally stable, which may be the essential requirement for the functions of the central domain within the P protein.
Interdimer interaction. In the asymmetric unit of the crystal, there are three individual dimers (Fig. 3). The 1 strand in one dimer forms main-chain hydrogen bonds with the 1 strand in the neighboring dimer in an antiparallel fashion, resulting in an eight-stranded antiparallel sheet between two neighboring dimers. There are four hydrogen bonds formed between residues 113 and 115 in the first 1 strand and residues 113 and 115 in the second 1 strand. This -sheet interaction extends throughout the crystal so that long strings of the dimers are constructed by interdimer interactions involving the twofold crystallographic symmetry. In addition, the main-chain amide nitrogen and carbonyl oxygen of Gln110 in the dimer-related molecule form two hydrogen bonds with the side chain of Gln147 from the neighboring helix, while the side-chain nitrogen of the same Gln110 forms a hydrogen bond with the carbonyl oxygen of residue 140, also from the neighboring helix. These additional hydrogen bonds further stabilize the interdimer interactions. The twofold axis in each dimer is approximately perpendicular to the central axis of the dimer string. The dimer axes are rotated 120° from each other about the central axis. Three dimers, therefore, make one complete rotation in the dimer string. Both the N and C termini are quite flexible in each dimer. The first ordered residue ranges from –1 (one leftover tag residue prior to residue 107) to 109 at the N terminus. Residues prior to residue 109 become ordered only when they are stabilized by interactions with residues in a neighboring dimer. On the other hand, the C-terminal residues become ordered by extending the last residue (Val177) into a hydrophobic pocket in the adjacent dimer strings. The hydrophobic pocket is formed by the side chains of Leu113 and Lys120 in one molecule and of Trp152 in the other dimer-related molecule. In full-length P, the regions corresponding to the termini in the central domain fragment are likely to be very flexible, to allow a wide range of movements by the N-terminal and C-terminal domains as required by their functions (see Discussion). The ordered termini observed in this crystal structure are the result of crystal packing.
Comparison with the oligomerization domain of the SeV phosphoprotein. The phosphoprotein of SeV has functions very similar to that of VSV. The SeV P is 568 amino acids, twice the size of VSV P (265 amino acids [serotype Indiana]). The N-terminal and C-terminal domains of SeV P bind N, while the central domain (residues 320 to 446) is responsible for oligomerization of a tetramer, as shown in the crystal structure (35). A 96--long helix, formed by 65 residues from each monomer contributes to the central tetrameric coiled coil in the tetramer. The coiled coil covers most of the tetramer, stabilizing interactions, while two short helices at the N terminus of the long helix have a small number of interactions with the neighboring monomers. The oligomerization domain of VSV P forms a stable dimer, using a similar, but shorter, helix. The intermolecular -sheet interactions by the hairpins do not have any similarity with the tetramer of the SeV P oligomerization domain. It has been suggested that the SeV P tetramer cartwheels along the N-RNA template while the L polymerase reads the nucleotide. P steps on one N monomer at a time, dictating that the polymerase complex moves 6 nucleotides each time. It has also been shown that VSV P must be capable of forming a tetramer in order to perform its RNA polymerase-related activities (18, 19). The tetramer unit in this crystal structure looks like a twisted sheet instead of a coil wheel (Fig. 3). Despite the structural differences between the two proteins, the essential N-RNA binding domains (the C-terminal domains) of both P proteins are at similar locations near the C terminus of the oligomerization domain. We suggest that the P protein associates with L as a dimer, as a functional unit. During transcription, the L/P unit recruits host factors and gets loaded onto the N-RNA template to transcribe mRNAs, whereas during replication, a P dimer-N0 complex replaces the host factors and forms a 4P:1N:1L replicase unit. Two P dimers may form a tetramer involving the interdimer sheet. The C-terminal domains of the first P dimer act as two feet to walk along the N-RNA template as RNA synthesis progresses.
Functions of the central domain in RNA polymerization. The C-terminal domain of VSV P plays an essential role in binding the N-RNA template and in viral replication (31). The N-terminal domain mainly influences transcription (8, 30). The central domain positions these two functional domains at the opposite ends and brings multiple copies of P (such as a tetramer) in close proximity through oligomerization. It is not clear whether the central domain interacts directly with any other viral proteins. In a recent study, a number of insertion and deletion mutations were introduced into the central domain and their effects on P activities were studied (11). Four Tn5-based insertion mutations of 16 amino acids were introduced at residues 136, 148, 159, and 167, respectively. The insertions at residues 136, 148, and 159 greatly reduced the replication activity, with the insertion at residue 136 being the most severe (about 90% reduction), while the insertion at residue 167 reduced the replication activity by about 40%. Only the insertion at residue 136 significantly reduced the transcription activity. Residue 167 is 3 amino acids away from the last ordered residue at the C terminus of the central domain. The inserted amino acids should have minimum effects on the integrity of the dimer structure of the central domain, probably simply extending the flexible region linking the C-terminal domain to the central domain. On the other hand, residue 136 is at the center of the helix-helix interaction in the dimer. Insertions should be strongly destructive to the P structure, resulting in a significant loss of both replication and transcription activity. Similarly, a deletion from residues 130 to 151 (two-thirds of the helix) significantly reduced replication and transcription activity, while various deletions between residues 140 and 201 had only limited reduction effects on replication and transcription activity. Residues 107 to 139 were able to retain substantial intermolecular interactions through the hydrophobic core around residue Trp138, as well as the potential tetramer interactions by the 1 strands (Fig. 4). This stripped-down dimer may still preserve the basic oligomerization function of the central domain. It is therefore plausible to postulate that the essential role of the central domain is to maintain the dimeric structure to properly position the N-terminal and C-terminal domains for their functional activities. The central domain probably would not have significant specific, direct interactions with any other viral proteins. The fact that insertions and deletions that disrupted most of the helix-helix interactions, as well as intermolecular sheets, reduced but did not completely diminish replication and transcription activity supports the suggestion that the central domain has the isolated function of oligomerization.
DISCUSSION
The crystal structure of the P central domain clearly defined the content of this domain and its architecture. The extensive intermolecular helix-helix interaction and the domain swapping between the two hairpins strongly suggest that the basic functional unit of the active P protein is always in the dimeric form. A trimer form of P suggested in an earlier report (17) could not exist because an oligomer of P could contain only multiples of two copies of P. This notion is particularly consistent with the observation that the P-N0 complex contains at least a P dimer in rabies virus (28). Taking into account the extensive intermolecular interactions in the assembled N-RNA genomic complex (unpublished data), it is likely that the P-N0 complex being recruited into the replication site for encapsidation of the nascent RNA is a 2:1 P dimer/N0 heterotrimer. The C-terminal domain of P primarily functions as a carrier to bind N0 in this complex. N0 can be added only one at a time in the form of a P-N0 complex, and it undergoes a significant conformational change once it becomes associated with RNA. The assembled N-RNA is hard to disrupt when it is formed (22).
The P central domain not only forms a stable dimer but also shows a potential tetrameric form, as observed in the noncrystallographic interaction. The ability to form a tetramer has been shown to be required for VSV replication activity (18, 19). The initiation site for viral replication has been shown to be separate from that for transcription (34, 39), and the replicase was shown to be formed by L/P/N with unspecified stoichiometry while the transcriptase is composed of L/P and host factors (34). The P molecules in the replicase must interact with both the N-RNA template and the additional N protein in the replicase complex. Our previous results showed that the N proteins are aligned in parallel when assembled with RNA (9). Since P also forms a parallel dimer, its twofold relationship prohibits the two C-terminal domains from binding N molecules that are not related by a twofold axis. However, the C-terminal domain of P is linked to the central domain by a very flexible linker. There is a possibility that one C-terminal domain is rearranged in the dimer in order to bind N in an orientation parallel to the first C-terminal domain. This type of interaction would introduce significant constraints on the P structure, although the possibility cannot be ruled out. The P dimer in the replicase complex may be essential for the processivity of the replicase along the N-RNA template. The two C-terminal domains in the P dimer may act as two feet to allow the complex to walk along the N-RNA template through an unclear mechanism. Since the formation of a P tetramer is required for replication activity, the tetramer structure observed in the noncrystallographic asymmetric unit may exist in the replicase complex. There are at least four C-terminal domains of P available in the replicase complex to bind both the N-RNA template and the additional N protein.
At the early stage of virus infection, the L and P proteins are carried into the cytoplasm by the ribonucleoprotein core. Since the virion genome does not carry any N0, there is no P dimer-N0 complex present to form the active replicase complex. It is likely that the L protein is associated with P proteins and host factors to form the transcriptase. The first isolation of the transcriptase complex, by Qanungo et al. (34), showed that a number of host factors, such as translation elongation factor 1, heat shock protein 60, and guanylyltransferase, are present but that the N protein is not. The dimeric P proteins in the transcriptase are not required to recruit N0. Instead, the P proteins mainly function to tether the transcriptase to the N-RNA template and to play a role in processive transcription.
ACKNOWLEDGMENTS
We thank the staff of the South East Regional Collaborative Access Team (SER-CAT) at the Advanced Photon Source for assistance in data collection.
Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under contract no. W-31-109-Eng-38. This work is supported in part by a grant to M.L. (NIH AI050066).
REFERENCES
Adams, P. D., K. Gopal, R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger, A. J. McCoy, N. W Moriarty, R. K. Pai, R. J. Read, T. D. Romo, J. C. Sacchettini, N. K. Sauter, L. C. Storoni, and T. C. Terwilliger. 2004. Recent developments in the PHENIX software for automated crystallographic structure determination. J. Synchrotron Radiat. 11:53-55.
Baltimore, D., A. S. Huang, and M. Stampfer. 1970. Ribonucleic acid synthesis of vesicular stomatitis virus, II. An RNA polymerase in the virion. Proc. Natl. Acad. Sci. USA 66:572-576.
Banerjee, A. K. 1987. Transcription and replication of rhabdoviruses. Microbiol. Rev. 51:66-87.
Banerjee, A. K., and S. Barik. 1992. Gene expression of vesicular stomatitis virus genome RNA. Virology 188:417-428.
Bell, J. C., E. G. Brown, D. Takayesu, and L. Prevec. 1984. Protein kinase activity associated with immunoprecipitates of the vesicular stomatitis virus phosphoprotein NS. Virology 132:229-238.
Brunger, A. T., P. D. Adams, and L. M. Rice. 1997. New applications of simulated annealing in X-ray crystallography and solution NMR. Structure 5:325-336.
Carson, M. 1997. Ribbons. Methods Enzymol. 277:493-505.
Chen, J. L., T. Das, and A. K. Banerjee. 1997. Phosphorylated states of vesicular stomatitis virus P protein in vitro and in vivo. Virology 228:200-212.
Chen, Z., T. J. Green, M. Luo, and H. Li. 2004. Visualizing the RNA molecule in the bacterially expressed vesicular stomatitis virus nucleoprotein-RNA complex. Structure 12:227-235.
Cuff, J. A., and G. J. Barton. 1999. Evaluation and improvement of multiple sequence methods for protein secondary structure prediction. Proteins 34:508-519.
Das, S. C., and A. K. Pattnaik. 2005. Role of the hypervariable hinge region of phosphoprotein P of vesicular stomatitis virus in viral RNA synthesis and assembly of infectious virus particles. J. Virol. 79:8101-8112.
Ding, H., T. J. Green, and M. Luo. 2004. Crystallization and preliminary X-ray analysis of a proteinase-K-resistant domain within the phosphoprotein of vesicular stomatitis virus (Indiana). Acta Crystallogr. D 60:2087-2090.
Emerson, S. U., and R. R. Wagner. 1972. Dissociation and reconstitution of the transcriptase and template activities of vesicular stomatitis B and T virions. J. Virol. 10:297-309.
Emerson, S. U., and R. R. Wagner. 1973. L protein requirement for in vitro RNA synthesis by vesicular stomatitis virus. J. Virol. 12:1325-1335.
Emerson, S. U., and Y. Yu. 1975. Both NS and L proteins are required for in vitro RNA synthesis by vesicular stomatitis virus. J. Virol. 15:1348-1356.
Ermine, A., and A. Flamand. 1977. RNA syntheses in BHK21 cells infected by rabies virus. Ann. Microbiol. (Paris) 128A:477-488.
Gao, Y., N. J. Greenfield, D. Z. Cleverley, and J. Lenard. 1996. The transcriptional form of the phosphoprotein of vesicular stomatitis virus is a trimer: structure and stability. Biochemistry 35:14569-14573.
Gao, Y., and J. Lenard. 1995. Multimerization and transcriptional activation of the phosphoprotein (P) of vesicular stomatitis virus by casein kinase-II. EMBO J. 14:1240-1247.
Gao, Y., and J. Lenard. 1995. Cooperative binding of multimeric phosphoprotein (P) of vesicular stomatitis virus to polymerase (L) and template: pathways of assembly. J. Virol. 69:7718-7723.
Gassner, N. C., and B. W. Matthews. 1999. Use of differentially substituted selenomethionine proteins in X-ray structure determination. Acta Crystallogr. D 55:1967-1970.
Gaudier, M., Y. Gaudin, and M. Knossow. 2002. Crystal structure of vesicular stomatitis virus matrix protein. EMBO J. 21:2886-2892.
Green, T. J., S. Macpherson, S. Qiu, J. Lebowitz, G. W. Wertz, and M. Luo. 2000. Study of the assembly of vesicular stomatitis virus N protein: role of the P protein. J. Virol. 74:9515-9524.
Huang, A. S., and R. R. Wagner. 1966. Comparative sedimentation coefficients of RNA extracted from plaque-forming and defective particles of vesicular stomatitis virus. J. Mol. Biol. 22:381-384.
Hwang, L. N., N. Englund, T. Das, A. K. Banerjee, and A. K. Pattnaik. 1999. Optimal replication activity of vesicular stomatitis virus RNA polymerase requires phosphorylation of a residue(s) at carboxy-terminal domain II of its accessory subunit, phosphoprotein P. J. Virol. 73:5613-5620.
Jeetendra, E., C. S. Robison, L. M. Albritton, and A. Whitt. 2002. The membrane-proximal domain of vesicular stomatitis virus G protein functions as a membrane fusion potentiator and can induce hemifusion. J. Virol. 76:12300-12311.
Masters, P. S., and A. K. Banerjee. 1986. Phosphoprotein NS of vesicular stomatitis virus: phosphorylated states and transcriptional activities of intracellular and virion forms. Virology 154:259-270.
Masters, P. S., and A. K. Banerjee. 1988. Resolution of multiple complexes of phosphoprotein NS with nucleocapsid protein N of vesicular stomatitis virus. J. Virol. 62:2651-2657.
Mavrakis, M., A. A. McCarthy, S. Roche, D. Blondel, and R. W. Ruigrok. 2004. Structure and function of the C-terminal domain of the polymerase cofactor of rabies virus. J. Mol. Biol. 343:819-831.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276:307-326.
Pattnaik, A. K., L. Hwang, T. Li, N. Englund, M. Mathur, T. Das, and A. K. Banerjee. 1997. Phosphorylation within the amino-terminal acidic domain I of the phosphoprotein of vesicular stomatitis virus is required for transcription but not for replication. J. Virol. 71:8167-8175.
Paul, P. R., D. Chattopadhyay, and A. K. Banerjee. 1988. The functional domains of the phosphoprotein (NS) of vesicular stomatitis virus (Indiana serotype). Virology 166:350-357.
Peluso, R. W. 1988. Kinetic, quantitative, and functional analysis of multiple forms of the vesicular stomatitis virus nucleocapsid protein in infected cells. J. Virol. 62:2799-2807.
Perrakis, A., M. Harkiolaki, K. S. Wilson, and V. S. Lamzin. 2001. ARP/wARP and molecular replacement. Acta Crystallogr. D 57:1445-1450.
Qanungo, K. R., D. Shaji, M. Mathur, and A. K. Banerjee. 2004. Two RNA polymerase complexes from vesicular stomatitis virus-infected cells that carry out transcription and replication of genome RNA. Proc. Natl. Acad. Sci. USA 101:5952-5957.
Tarbouriech, N., J. Curran, R. W. Ruigrok, and W. P. Burmeister. 2000. Tetrameric coiled coil domain of Sendai virus phosphoprotein. Nat. Struct. Biol. 7:777-781.
Terwilliger, T. C. 2000. Maximum likelihood density modification. Acta Crystallogr. D 56:965-972.
Terwilliger, T. C., and J. Berendzen. 1999. Automated MAD and MIR structure solution. Acta Crystallogr. D 55:849-861.
Villarreal, L. P., M. Breindl, and J. J. Holland. 1976. Determination of molar ratios of vesicular stomatitis virus induced RNA species in BHK21 cells. Biochemistry 15:1663-1667.
Whelan, S. P., and G. W. Wertz. 2002. Transcription and replication initiate at separate sites on the vesicular stomatitis virus genome. Proc. Natl. Acad. Sci. USA 99:9178-9183.(Haitao Ding, Todd J. Gree)