The Active Form of the Norovirus Rna-dependent Rna Polymerase Is a Homodimer with Cooperative Activity

Norovirus (NV) is a leading cause of gastroenteritis worldwide and a major public health concern. So far, the replication strategy of NV remains poorly understood, mainly because of the lack of a cell system to cultivate the virus. In this study, the function and the structure of a key viral enzyme of replication, the RNA-dependent RNA polymerase (RdRp, NS7), was examined. The overall structure of the NV NS7 RdRp was determined by X-ray crystallography to a 2.3 A ˚ (0.23 nm) resolution (PDB ID 2B43), displaying a right-hand fold typical of the template-dependent polynucleotide polymerases. Biochemical analysis evidenced that NV NS7 RdRp is active as a homodimer, with an apparent K d of 0.649 mM and a positive cooperativity (Hill coefficient n H 51.86). Crystals of the NV NS7 homodimer displayed lattices containing dimeric arrangements with high shape complementarity statistics. This experimental data on the structure and function of the NV RdRp may set the cornerstone for the development of polymerase inhibitors to control the infection with NV, a medically relevant pathogen.


INTRODUCTION
Norovirus (NV) is one of the major agents of viral gastroenteritis and a leading cause of food-borne gastroenteritis worldwide (Lopman et al., 2003(Lopman et al., , 2008)).NV belongs to the family Caliciviridae, a virus family including human and non-human pathogenic strains.The genus Norovirus is divided into five genogroups.Strains belonging to genogroups 1, 2 and 4 infect humans, with the genogroup 2 strains being predominantly detected over the years.Animal NV strains, i.e. murine or bovine NV strains cluster mainly within genogroup 5 or 3, respectively (Zheng et al., 2006).The murine NV so far remains the only strain of the NV species that can be propagated in cultured cells.
NV is a non-enveloped RNA virus with a single-stranded positive-oriented genome ranging from 7.3 to 8.5 kb (Green, 2007).The viral genome is organized into three open reading frames (ORF): ORF-1 encodes the nonstructural proteins, ORF-2 encodes the capsid protein or virion protein 1 (VP1) and ORF-3 encodes the virion protein 2 (VP2) (Fig. 1a).Downstream from ORF-2, an untranslated region (UTR) is present (Green, 2007), which is followed by a poly(A) tail of variable length.ORF-1 is predicted to encode a single polyprotein that is cotranslationally processed, resulting in the generation of the viral enzymes of replication.Among those, the RNAdependent RNA polymerase (RdRp) non-structural protein 7 (NS7) plays a key role in the synthesis and amplification of genomic RNA.
So far, the structural and functional characterization of the interaction of NV replicative enzymes remain unaddressed.Recently, Zamyatkin et al. (2008) have described the interaction between the NV RdRp and its RNA template.However, it remains unclear as to whether RNA synthesis by the NV RdRp is regulated by the interaction of two monomeric forms, resulting in positive cooperativity.
In this study, the structural and functional features of NV NS7 homodimers have been investigated.The structure of the protein in a new crystal form has been resolved by Xray crystallography to 2.3 A ˚(0.23 nm) resolution, displaying a dimer-of-dimers arrangement in the crystal lattice.The possible three-dimensional structural arrangement of NV NS7 homodimers was analysed using the five available crystal forms of the protein.Furthermore, we present the first experimental evidence for positive cooperativity of NV NS7 monomers, showing that RNA synthesis by NV NS7 is increased by dimerization of the monomers.

METHODS
Expression and purification of NV NS7.The NV NS7 as well as its active-site mutant mNS7 were expressed and purified as described previously (Rohayem et al., 2006a, b).Protein concentration was determined with the BCA Protein assay kit (Pierce) based on the Biuret reaction.The fraction containing the recombinant protein was diluted in glycerol (50 % v/v) and stored at -20 uC.
Crystallization of NV NS7 protein.The purified NV NS7 was concentrated to 9 mg ml 21 .The concentrated sample was centrifuged at 10 000 g for 5 min to remove aggregates arising from the concentration procedure.Subsequent crystallization trials were carried out at room temperature and 4 uC leading to optimized conditions of 1 M tri-sodium citrate, 125 mM NaCl and 0.1 M sodium cacodylate, pH 6.5 at 20 uC.
The diffraction data (Table 1) were collected by using the beamline ID23-1 (European Synchotron Radiation Facility).The diffraction data were measured to 2.3 A ˚resolution and the data were processed and scaled using the programs XDS and XSCALE.The phases were obtained by molecular replacement using the program MOLREP of the CCP4 program package and NV RdRp (PDB ID 1SH0) as the model (Ng et al., 2004).The crystals were primitive orthorhombic in space group P2 1 2 1 2 1 with unit cell dimensions of a590.5 (9.05 nm), b5125.6 (12.56 nm) and c5218.1A ˚(21.81 nm), and contain four molecules in the asymmetric unit.The model was built using consecutive rounds of ARP/wARP (Lamzin et al., 1999), manual building in O (Jones et al., 1991) and refinement with REFMAC5 (Murshudov et al., 1997).The coordinates and structure factors have been deposited in the PDB with ID code 2B43.A summary of the model quality is given in Table 1.
Investigation of the interaction of NS7 monomers using native PAGE.Discontinuous native protein gel electrophoresis was used as described by others with slight modifications (Niepmann & Zheng, 2006).Briefly, the purified proteins were mixed with 26 native PAGE sample buffer (100 mM Tris/HCl, pH 9.0, 40 % glycerol (v/v), 0.5 % Coomassie brilliant blue G) and incubated for 10 min at room temperature.The samples were loaded onto a 6-12 % gradient gel.Cathode (100 mM Tris-histidine, pH 8.0, 0.002 % Coomassie brilliant blue G) and anode buffer (100 mM Tris/HCl, pH 8.8) were applied to a vertical electrophoresis chamber (Hoefer) and Coomassie brilliant blue G (Serva) was added to a final concentration of 0.002 % (w/v) to the cathode buffer.Electrophoresis was carried out at 15 V cm 21 for 14-16 h at 4 uC.The gel was then washed with stripping solution (25 % methanol, 10 % acetic acid).For verification of the correct size of the proteins, the protein band was excised from the native PAGE and incubated for 20 min in denaturing solution (350 mM Tris/HCl, pH 6.8, 10.0 % SDS, 36.0 % glycerol, 5.0 % b- mercaptoethanol, 0.012 % bromphenol blue).The gel stripes were then embedded in the stacking gel of a vertical electrophoresis chamber (Hoefer).Electrophoresis for separation of the proteins was carried out at 80 V for the stacking gel (5 %) and at 120 V for the separation gel (15 %).Proteins were subsequently visualized by Western blot using an anti-NS7 polyclonal antibody as described previously (Rohayem et al., 2006b).Investigation of the equilibrium conditions of RNA synthesis by NV NS7.The equilibrium conditions were characterized by measuring the amount of RNA synthesized through de novo initiation by NV NS7 in vitro using homopolymeric poly(C) 20 -RNA templates (GE Healthcare) in the presence of GTP as described previously with slight modifications (Rohayem et al., 2006a, b).Briefly, the reaction mixture consisted of 5 mg synthetic poly(C) 20 -RNA template (20 nt in length, yielding a final concentration of 14.7 mM), 50 mM HEPES (pH 7.0), 3 mM magnesium acetate, 4 mM DTT, 6 mM ZnCl 2 , 50 U RNase inhibitor (RNAsin; Promega), 0.5 mM rGTP and RNase-DNase-free water to a final volume of 50 ml.In each reaction, 1.25 mM purified NV NS7 was added and the reaction carried out at 37 uC.It was stopped by adding 50 ml stop solution (4 M ammonium acetate, 100 mM EDTA) and purified with the Microcon Ultracel JM-10 columns (Millipore) according to manufacturer's instructions.The amount of RNA synthesized was measured in a time-course experiment at 5, 10, 15, 20 and 25 min after initiation of the reaction.For this purpose, 3 ml of the reaction was used in the Quant-iT RiboGreen RNA assay kit (Invitrogen) according to manufacturer's instructions.Fluorescence was measured on an Infinite F200 reader (Tecan).
Characterization of the K d of homodimerization of NS7 monomers.To determine the affinity constants for NS7 dimerization, the RdRp enzymic reaction was used as a tool to measure an apparent K d for the interaction of the NS7 monomers.Therefore, the enzymic activity of NS7 was measured in the presence of increasing concentrations of the active-site mutant mNS7.The K d measured here depends on the values P i and P max , where P i is the amount of RNA synthesized in a set period of time, here 15 min, and P max is the maximal amount of RNA synthesized in a set time period in the presence of saturating concentrations of NS7 or mNS7.The K d as well as the P i and P max were determined using a non-linear regression curve-fitting program with the Prism 4 software (GraphPad software).
Assessment of the cooperativity between NS7 monomers.The cooperativity between two NS7 monomers was investigated using the Hill plots as described by others (Goodrich & Kugel, 2007).Therefore, NS7 was titrated and the amount of synthesized RNA was determined at each concentration of NS7.Next, the ratio P5P i / P max was calculated.P i relates to the amount of RNA synthesized for a given concentration, and P max to the maximal amount of RNA synthesized.To determine the Hill coefficient, the value log (P/12P) was plotted against the logarithmic value of the NS7 concentration.Importantly, only log (P/12P) values ranging between 20.9 and 0.9 were used, resulting in a linear regression.
The correlation coefficient as well as the slope of the line were measured, giving the Hill coefficient n H .An n H .1 indicates positive cooperativity between NS7 monomers.

Heterologous expression and purification of recombinant NV NS7
A fusion protein bearing a (His) 6 -tag at its C terminus was overexpressed in Escherichia coli and purified by Ni-NTA affinity chromatography.A soluble NV protein of about 57 kDa was obtained (Fig. 1b).Analogously, the active-site mutant of NV NS7 (mNS7) was expressed and purified.The active-site mutant displayed the same characteristics in terms of solubility and migration in SDS-PAGE (Fig. 1b).

Structure of NV NS7 protein
The structure of the NS7 protein from NV strain Hu/NLV/ Dresden174/1997/GE has been determined to 2.3 A resolution (PDB ID 2B43).The model was refined to an R-factor of 19.2 % and free R-factor of 25.7 % and contains 501 aa (aa 6-506) and 1122 water molecules.The overall structure (Fig. 1c) is, as expected, very similar to the previously determined structure of NV NS7 (Ng et al., 2004).A comparison of the structural similarities of different RdRps of the families Caliciviridae and Pircornaviridae is shown in Table 2.The structural differences observed between the present structure and the previously determined structure of NV NS7 (Ng et al., 2004) and the recent structure of NV polymerase in complex with primer-template RNA (Zamyatkin et al., 2008) is most likely due to structural flexibility, since the sequence identity is very high 98-99 % in the structured regions.The main differences are observed in the loop region made up of residues 369-380 (Fig. 1d) and in the C-terminal region.In the uncomplexed structures, this region is located in the active-site cleft (Fig. 1e).In the previously published NV structure (PDB ID 1SH0), the C-terminal residues interact mainly with residues 218-220 preceding the a7 helix, while in the structure published in this study (PDB ID 2B43), the C terminus interacts mainly with residues 441-443 of the a13 helix in the active-site cleft.These interactions also lead to slight changes in the backbone structure of the a13 helix and the N-terminal part of the a7 helix.In the RNA

NV NS7 forms homodimers in a concentrationdependent manner
In order to investigate the interaction of NV NS7 monomers, native PAGE was run using increasing concentrations of the purified protein.As shown in Fig. 2(a), a protein of a molecular mass of about 120 kDa was observed starting at a concentration of about 0.500 mM.The intensity of the protein stained on the native gel increased with increased concentration of NS7.For further characterization, the protein at a concentration of 2.000 mM was excised from the native gel, denatured and analysed by SDS-PAGE.Western blot analysis revealed that the protein migrated at about 57 kDa, indicating that the excised protein of about 120 kDa consisted of two monomers of 57 kDa each (Fig. 2b).

Determination of the K d of NV NS7 homodimers
In order to measure the K d of homodimeritzation, an enzymic assay was developed.This assay measures the amount of RNA synthesized by the NV NS7 in vitro on homopolymeric poly(C) 20 -RNA templates.Therefore and in a first step, we have determined the equilibrium conditions, such as the substrate concentration never decreases to a point where it may affect the amount of RNA synthesized over the time course of the reaction.As observed in the time-course experiment in Fig. 2(c), the amount of RNA synthesized until 25 min displayed a linear correlation (r 2 50.98) with the time of reaction, indicating that the active enzyme complex synthesizes RNA at a constant rate in the presence of saturating amounts of RNA substrate (14.7 mM).Based on this data, the time point chosen where the reaction is at equilibrium was 15 min.
In a further step, the K d of homodimerization was determined in a series of reactions in which the concentration of NS7 is kept constant (1.25 mM), and the concentration of mNS7 is varied over a range extending from 10-fold below the putative K d to 10-fold above the putative K d .By measuring the amount of RNA synthesized at 15 min, the K d as well as the P i and P max were determined using a non-linear regression curve-fitting program (Fig. 2d).The apparent calculated K d ±SEM was 0.649±0.109mM.

Multimeric arrangement of NV NS7 RdRp
In the present structure, the NV NS7 crystallizes in space group P2 1 2 1 2 1 with four monomers in the asymmetric unit.The protein arranges as a dimer-of-dimers tetramer (Fig. 3a).Given the biochemical data supporting higher activity of the dimeric protein, an interesting question is whether the active dimer is represented in the crystal packing of this crystal, or the other crystal forms available for the protein.To investigate this issue, we analysed the crystal packing of the five available crystal forms for the NV NS7 protein represented by structures NV NS7 PDB ID 1SH0, 1SH2, 1SH3, 2B43 and 3BSO (Table 3).The crystal packing was examined to define plausible C2-symmetrical dimeric arrangements.The resulting dimers were analysed for interacting surface area and surface shape complementarity (SC) as defined by others (Lawrence & Colman, 1993).SC values and interaction surface areas were calculated using the CCP4 programs: SC and Areaimol, respectively (Lawrence & Colman, 1993;Lee & Richards, 1971).
The first observation is that all but one of the crystal forms produces C2 dimers in the crystal lattice and that all these have different dimer interactions, even though the dimers formed in the NV NS7 PDB ID 1SH2 and 1SH3 crystal forms are essentially the same (Fig. 3b).The interface areas (sum of the interacting area of both subunits) in the different dimeric arrangements ranges from 1528 to 2271 A ˚2.The buried surfaces are thus in all cases of average size for protein-protein interactions, or even somewhat larger (Lo Conte et al., 1999).The SC statistic on the other hand clearly puts the dimeric arrangements into two groups.The NV NS7 PDB ID 1SH2 and 1SH3 structures produce, as mentioned before, very similar dimers that also have very poor SC values, well below what is to be expected for a protein-protein complex.In contrast, the NV NS7 PDB ID 1SH0 and 2B43 crystal lattices contain dimeric arrangements with very good SC statistics.It should be mentioned that these dimeric interactions, unlike NV NS7 PDB ID 1SH2 and 1SH3, are fundamentally different.

NV NS7 monomers display positive cooperativity
The interdependence between RNA synthesis by NV NS7 and dimerization was investigated using the enzymic assay described previously.When increasing the concentration of NS7 in the reaction, the amount of RNA synthesized increases until a plateau phase (Fig. 4a), indicating an equilibrium shift from the monomeric form of the enzyme into the dimeric form.In order to determine whether a positive cooperativity between the monomers occurs, the Hill plots were used.As shown in Fig. 4(b), the amount of RNA synthesized correlated linearly with the NS7 concentration (r 2 50.99).The Hill coefficient n H ±SEM was 1.86±0.13,indicating positive cooperativity between the NS7 monomers.

DISCUSSION
Replication of the NV genome depends on viral enzymes of replication.Among those, the RdRp is predicted to play an essential role being involved in the synthesis of antiviral genomic RNA, but also subgenomic RNA.Recently, the crystal structure of prototype strains of NV, sapovirus and lagovirus RdRp (Fullerton et al., 2007;Ng et al., 2004) were determined.The NV NS7 RdRp resembles the right-hand configuration of many DNA and RNA polymerases (Biswal et al., 2005;Choi et al., 2004;Lesburg et al., 1999;Ng et al., 2002Ng et al., , 2004;;Thompson & Peersen, 2004).Moreover, the functional features of the NV NS7 have been dissected in vitro (Rohayem et al., 2006a, b).However, it remains unclear whether the active form of the NV NS7 RdRp is a homodimer displaying cooperative activity.This question is of relevance, in particular in the light of innovative therapeutic strategies aiming at disassembling the replication complex of NV through inhibition of viral enzymes interaction.
In this study, the structure and functional properties of NV NS7 homodimers have been determined for the first time.
According to our structural data, NV NS7 monomers interact to form homodimers. Our results also show that homodimerization of NV NS7 monomers is concentrationdependent, and that the apparent K d of homodimerization lies within the nM range.Moreover, the monomers display a positive cooperativity in vitro.
In order to investigate the structural and functional features of the NV NS7 RdRp, a domain encompassing the predicted NV RNA polymerase motifs A-E was designed.This domain displayed homologies to other RdRps of calicivirus, i.e. the lagovirus, sapovirus and vesivirus.After amplification of the encoding DNA and subsequent cloning in a prokaryotic expression vector, expression and purification yielded the soluble recombinant protein.In addition, an active-site mutant of NV NS7 was generated.
The structure of NV NS7 RdRp from strain Hu/NLV/ Dresden174/1997/GE was resolved to 2.3 A ˚resolution.Although the sequence is highly similar to the previously determined structures of NV NS7 RdRp (PDB ID 1SH0), some differences in the structure are observed (Fig. 1d and  e).Because of the high sequence identity, the observed differences are most likely indicative of structural flexibility in these parts.It is interesting to note that the C terminus Although missing from the electron density map, it is interesting to note that this drastic change in physicochemical properties of the C terminus does not prevent binding to the active-site cleft.
In the next step, we investigated whether the NV NS7 RdRp is active in a monomeric or dimeric form.Therefore, dimerization of NS7 was examined on native gels with increasing concentrations of the protein.Discontinuous native gel electrophoresis that is suitable for analysis of oligomeric proteins under native conditions was used (Niepmann & Zheng, 2006).The difference between the discontinuous native gel electrophoresis and the classical native Laemmli Tris-glycine buffer system is the use of Tris-histidine gel buffer and a polyacrylamide gradient, allowing a reliable characterization of the physical properties of a protein consistent with its size, oligomeric state and shape (Niepmann & Zheng, 2006).As observed in Fig. 2(a), NV NS7 dimerizes in a concentration-dependent manner.This observation is in accordance with previous studies in other viral RdRps such as in vesivirus, hepatitis C virus and influenza virus (Jorba et al., 2008;Kaiser et al., 2006;Wang et al., 2002).For those proteins, dimerization but also oligomerization of monomeric subunits of the polymerase complex have been evidenced.
In order to investigate further the dimerization of NV NS7 in solution, size exclusion chromatography (SEC) was performed.With this assay, the NS7 protein eluted as a single peak corresponding in size to the monomer (data not shown).The reasons for such a discrepancy are various.One possible explanation could lie in the lower resolution of SEC in comparison to discontinuous native gel electrophoresis (Niepmann & Zheng, 2006), the dimers not being eluted, as such, out of the column in SEC.Furthermore, discrepancies in the systems used for characterization of the oligomerization status of a protein may be related to the physical separation principles of the  systems, such as gravity flow in the case of SEC versus voltage in the case of gel electrophoresis.This is the case for the polypyrimidine tract-binding protein that binds to the picornavirus RNA, where discontinuous native protein gel electrophoresis yields different products than those observed by SEC (Perez et al., 1997;Song et al., 2005), but also for hepatitis C virus NS3 protein (Dumont et al., 2006;Levin & Patel, 1999;Sikora et al., 2008).Moreover, and as observed in other viral RdRps (Wang et al., 2002), the NS7 monomers may not obligatorily interact in a stable complex during RNA synthesis, but rather enhance a certain point of the reaction, such as initiation or release of the product.This may explain the lack of dimerization under SEC conditions, in contrast to RNA synthesis conditions where positive cooperativity was evidenced.
The apparent K d of the NV NS7-NS7 homodimer was measured using an enzymic assay as described by others (Goodrich & Kugel, 2007).The enzymic assay was designed in such a way that the concentration of the substrate is sufficiently high and the reaction time limited enough to ensure that the amount of RNA synthesized increases linearly over time.This was done by assembling a reaction in the presence of saturating amounts of substrate, and stopping the reaction at 5, 10, 15, 20 and 25 min, measuring the amount of RNA synthesized and plotting it versus time.As observed in Fig. 2(c), the plot is linear through the time points used until 25 min, indicating a non-limiting rate of the reaction.Based on those parameters, a series of reactions was performed in which the concentration of NS7 is kept constant (1.25 mM) and the concentration of mNS7 is varied over a range extending from 10-fold below the putative K d to 10-fold above the putative K d .In the case of the addition of mNS7, the mNS7 competes with the NS7 monomer for formation of the NS7-mNS7 heterodimer.If a decrease is observed in the presence of increased concentrations of the mNS7, this will indicate that only active NS7-NS7 homodimers can convert substrate to product, whereas NS7-mNS7 heterodimers will not.Previous reports on the K d of the RdRp of hepatitis C virus have reported a K d for homodimerization in the nanomolar range (22 nM; Wang et al., 2002).The concentrations of mNS7 chosen were therefore chosen to range from 0 to 2 mM.The amount of RNA synthesized at 15 min was measured and the K d as well as the P i and P max determined using a non-linear regression curve-fitting program (Fig. 2d).Here, P i is the amount of RNA synthesized in a set time period by the NS7 RdRp in the absence of mNS7, whereas P max is the amount of RNA synthesized in a set period of time when NS7 is saturated with mNS7.The apparent calculated K d ±SEM was 0.649±0.109mM.
To investigate further the structural features of NS7 RdRp dimers, crystals of the NS7 protein were grown and the structure resolved at 2. suggest that none of the lattice interactions is expected to form a stable dimer in solution.Since the catalytically relevant multimeric state is a key factor in understanding the protein function, an in-depth investigation of the structural features of multimerization was performed.In order to adopt defined dimeric structures, the protein dimers need to be C2 symmetrical, or at least possess a pseudo C2 symmetry.Indeed, dimers can only form if the same interacting surface is employed in both monomers, otherwise a polymerization or higher oligomerization would occur.We have analysed the molecular packing in the five different crystal forms to evaluate the plausibility of observed dimerization.Four of the five crystal packages produce C2-symmetrical dimers (Table 3 and Fig. 3b).Interestingly, all four dimeric arrangements are different, although the dimers observed in PDB ID 1SH2 and 1SH3 are essentially the same in terms of surface and geometry of interaction.In all cases, the size of the interacting surfaces are as expected for a protein-protein complex of this size, although two of the dimers display a very poor surface SC.The other two dimers, on the other hand, have a very good SC and are thus more probable candidates for a biologically relevant dimer.In order to assess the plausibility of the dimeric arrangements during the different steps of the catalytic cycle, models of the RNA-bound forms of the dimers were constructed based on RNA binding observed in the 3BSO structure.From these models, it is obvious that three of the dimers are not compatible with the full polymerization reaction of the enzyme.The dimers of the 1SH2 or 1SH3 structures would prevent template entrance.In contrast, the NV 2B43 dimer would prevent exit of the product RNA.However, the dimer observed in the 1SH0 crystal lattice seems compatible with the full reaction (Fig. 5).
In conclusion, assuming a defined dimeric arrangement, there is only one available crystal form of the protein that shows a dimer in the crystal lattice that seems plausible in terms of the interaction area, SC and compatibility with the entire catalytic cycle.However, as observed in other viral RdRp (Wang et al., 2002), the NS7 monomers may not obligatorily interact in a stable complex during RNA synthesis, but rather enhance a certain point of the reaction, such as initiation or release of product.If this is the case, the NV NS7 PDB ID 1SH0 dimer may not be the only possible form of dimerization.If the dimeric interaction enhances the initiation step, the packing observed in the present (2B43) structure may also represent a plausible arrangement considering its high SC.Based on the interaction and modelling studied performed here we are unable to tell whether the biologically relevant dimerization is represented in the available crystal packing or not; however, the analysis has picked out the most probable candidates and may serve as a good foundation to investigate this further, e.g. by mutational studies.Importantly, the NS7 proteins displayed a cooperative activity (n H ±SEM51.86±0.13),indicating that the interaction of the monomers increases the activity of the protein.Based on the structural analysis it seems most likely that the dimers observed in vitro may reflect a transient state within the replication complex of NV.

Fig. 1 .
Fig. 1.Domain design, expression and structure of the NV RdRp.(a) Schematic representation of the organization of NV clone pUS-NorII (GenBank accession no.AY741811).The complete NV clone pUS-NorII (7555 bp in length), as well as the ORF-1 (5099 bp in length), ORF-2 (1619 bp in length) and ORF-3 (806 bp in length) are shown.The putative cleavage sites at the interface NS6/NS7 (E 1189 /G) and the C terminus of the ORF-1 (*) are indicated.The domain NV-AY741811-NS7 corresponding to the active RdRp of NV is shown.The active-site mutant of the NV RdRp displays a mutation at the Asp 343 Asp 344 in the catalytic site of the enzyme.The predicted size of the non-structural proteins (NS) is indicated.(b) Expression and purification of NV NS7 proteins in E. coli.SDS-PAGE analysis of purified recombinant NV NS7 expressed in E. coli.NS7, Wild-type NV NS7.mNS7, Active site NV NS7 mutant (YGD 343G D 344G ).M, Molecular mass marker (kDa).(c) Overall structure of the NV NS7 monomer coloured from the N terminus (blue) to the C terminus (red).The three major domains of thumb, fingers and palm are visible with the (blue) N-terminal extension reaching across from the top of the fingers to the thumb domain (red).(d) Conformational differences in the 369-380 loop domain, NV NS7 PDB ID 2B43 in blue, NV NS7 PDB ID 1SH0 in red and the NV NS7 PDB ID 3BSO RNA complex structure in green.(e) Conformational differences in the C terminus of NV NS7 PDB ID 2B43 in blue, NS7 PDB ID 1SH0 in red.Note the difference in interaction with the NV NS7 PDB ID 1SH0.C-terminal residues interact mainly with residues 218-220 preceding the a7 helix.The NV NS7 PDB ID 2B43 C terminus interacts mainly with residues 441-443 of the a13 helix.

Fig. 2 .
Fig. 2. Evidence of homodimerization of NV NS7 RdRp.(a) Detection of NS7 homodimers by native PAGE.NS7 was loaded in increasing concentrations as indicated.The NS7 dimer [NS7] 2 is observed starting from a concentration of 0.500 mM.The protein at a concentration of 2.000 mM was excised from the native gel as indicated and visualized on SDS-PAGE.M, High Molecular Weight Calibration kit for native gel electrophoresis (Amersham Bioscience).(b) Visualization of the excised protein by Western blot after SDS-PAGE denaturation.The product migrated at a level corresponding to the predicted molecular mass of the monomer [NS7] 1 , which is of 57 kDa.M, Molecular mass marker (kDa).(c) Kinetics of RNA synthesis by NV NS7.RNA synthesis was examined in the presence of wild-type NV RdRp (NS7), a synthetic homopolymeric poly(C) 20 used as template and 0.5 mM rGTP.The reaction was stopped at the indicated time points and the total amount of RNA measured by fluorescence using the Quant-iT RiboGreen RNA assay kit (Invitrogen) on an Infinite F200 reader (Tecan).The correlation coefficient r 2 is shown.The mean±SEM of three independent measures are shown.The 95 % confidence intervals (CI) of the curve are shown as dashed lines.(d) RNA synthesis by NS7 was performed in the presence of increasing concentrations of active-site NS7 mutant (mNS7, YGD 343G D 344G ) as indicated.The apparent calculated K d ±SEM is 0.649±0.109mM.P i and P max values are shown.The mean±SEM of three independent measures are shown.The 95 % CI of the curve are shown as dashed lines.

Fig. 3 .
Fig. 3. Crystal structures of the observed NV NS7 protein homodimers.(a) Tetrameric arrangement in the asymmetric unit of NV NS7 PDB ID 2B43 structure.(b) C2-symmetrical dimers formed in the crystal lattice in the different structures of NV NS7 RdRp, as indicated.

Fig. 4 .
Fig. 4. Assessment of the positive cooperativity of NV NS7.(a) Concentration-dependent synthesis of RNA by NV NS7.RNA synthesis was examined in the presence of wild-type NV RdRp (NS7), a synthetic homopolymeric poly(C) 20 used as template and 0.5 mM rGTP.The total amount of RNA was measured by fluorescence using the Quant-iT RiboGreen RNA assay kit (Invitrogen) on a Infinite F200 reader (Tecan).The 95 % CI of the curve are shown as dashed lines.The mean±SEM of three independent measures are shown.(b) Assessment of positive cooperativity using the Hill plots.The Hill coefficient n H ±SEM as well as the correlation coefficient r 2 are shown.The mean±SEM of three independent measures are shown.The 95 % CI of the curve are shown as dashed lines.
Fig. 5. Model of the RNA complex of the NV NS7 homodimer (PDB ID 1SH0).(a) View from the product exit side.(b) Orientation showing entrance tunnel for the template strand.The template strand is coloured in yellow and the primer strand in green.

Table 2 .
Structural similarity of the present structure to other polymerases

Table 3 .
Comparative analysis of molecular packing in the different crystal forms of NV NS7 protein