Similarities between NS2 - Science Direct

Virology 280, 221–231 (2001) doi:10.1006/viro.2000.0764, available online at http://www.idealibrary.com on

Multimers of the Bluetongue Virus Nonstructural Protein, NS2, Possess Nucleotidyl Phosphatase Activity: Similarities between NS2 and Rotavirus NSP2 Zenobia F. Taraporewala, Dayue Chen, and John T. Patton Laboratory of Infectious Diseases, National Institutes of Allergy and Infectious Diseases, National Institutes of Health, 7 Center Drive, MSC 0720, Room 117, Bethesda, Maryland 20892 Received August 23, 2000; returned to author for revision October 27, 2000; accepted November 7, 2000 The nonstructural protein, NS2, of bluetongue virus is a nonspecific single- stranded RNA-binding protein that forms large homomultimers and accumulates in viral inclusion bodies of infected cells. NS2 shares these features with the nonstructural protein, NSP2, of rotavirus, which like BTV is a member of the family Reoviridae. Recently, NSP2 was shown to have an NTPase activity and an autokinase activity that catalyzed its phosphorylation in vitro. To examine NS2 for similar enzymatic activities, the protein was expressed in bacteria with a C-terminal His-tag and purified to homogeneity. Recombinant (r)NS2 possessed nonspecific RNA-binding activity and formed 8–10S homomultimers of the same approximate size as rNSP2 homomultimers. Notably, enzymatic assays performed with rNS2 showed that the protein hydrolyzed the ␣, ␤, and ␥ phosphodiester bonds of all four NTPs. Therefore, rNS2 possesses a nucleotidyl phosphatase activity instead of the NTPase activity of NSP2, which only hydrolyzes the ␥ phosphodiester bonds of NTPs. NS2 did not exhibit any autokinase activity in vitro, unlike NSP2. However, both NS2 and NSP2 were phosphorylated in vitro by cellular kinases. Although the nature of the enzymatic activities differs significantly, the fact that both NS2 and NSP2 hydrolyze NTPs, undergo phosphorylation, bind RNA, and assemble into multimers consisting of 6 ⫾ 2 subunits suggests that they are functional homologs.

type 10. It is the only viral protein reported to undergo phosphorylation during BTV replication (Huismans et al., 1987; Devaney et al., 1988). The phosphorylation of NS2 occurs at serine residues (Thomas et al., 1990). NS2, derived from BTV-infected cells and from recombinant baculovirus-infected cells, exists as a ⬃7S multimeric complex and has high affinity for single-stranded (ss)RNA (Uitenweerde et al., 1995). The results of several studies have indicated that the ssRNA-binding activity of NS2 is nonspecific (Huismans et al., 1987; Thomas et al., 1990; Uitenweerde et al., 1995). However, recent studies by Theron and Nel (1997) have suggested that NS2 has specific affinity, albeit weak, for the 3⬘ terminus of BTV transcripts. The RNA-binding domain has been mapped to the N terminus of NS2 (Zhao et al., 1994). While NS2 is believed to play an important role in genome replication and/or virus assembly, its function in these events is not known. Indeed, the functions of most of the nonstructural proteins encoded by members of the Reoviridae, including rotavirus and reovirus, are not known. Because the replication strategies of BTV, rotavirus, and reovirus are similar, it is reasonable to assume that these viruses encode proteins with homologous functions. The presumed homologs of BTV NS2 are the rotavirus and the reovirus nonstructural proteins NSP2 (35 kDa) and ␴NS (41 kDa), respectively (Zhao et al., 1994, Huismans et al., 1987). Like NS2, NSP2 and ␴NS have nonspecific affinity for ssRNA and form large homomultimeric complexes (Kattoura et al., 1992, 1994;

INTRODUCTION Bluetongue virus (BTV) is the prototype member of the orbivirus genus in the family Reoviridae. The virus has a genome of 10 segments of double-stranded (ds) RNA that codes for seven structural proteins (VP1–VP7) and four nonstructural proteins (NS1–NS3 and NS3A) (Mertens et al., 1984; Van Dijk and Huismans, 1988). The outer layer of the icosahedral BTV virion is made up of VP2 and VP5 (Mertens et al., 1984; Grimes et al., 1997). The inner layer is composed of the major proteins VP3 and VP7 and the minor proteins VP1, VP4, and VP6 (Verwoerd et al., 1972; Van Djik and Huismans, 1988; Martin and Zweerink, 1972; Mertens et al., 1984). The proteins of the inner layer in association with the genome make up the core of the virion. In BTV-infected cells, VP5, VP7, NS1, and NS2 accumulate in large electron-dense structures called viral inclusion bodies (VIBs) (Brooks et al., 1993, Eaton et al., 1990). The presence of cores and morphogenic intermediates in and around the periphery of the VIBs have led to the proposal that BTV replication and assembly occur at such sites (Brooks et al., 1993). The localization of NS1 and NS2 to VIBs suggests that these proteins participate in BTV replication and/or assembly. NS2, is the 41-kDa product of segment 8 of BTV sero-

1 To whom reprint requests should be addressed. Fax: (301) 4968312. E-mail: [email protected].

221 0042-6822/01

222

TARAPOREWALA, CHEN, AND PATTON

Gillian and Nibert, 1998). NSP2 and ␴NS also accumulate in inclusion bodies that form in virus-infected cells (Petrie et al., 1984; Mbisa et al., 2000). In a recent study, it was demonstrated that rotavirus NSP2 has nonspecific nucleoside triphosphatase (NTPase) activity and an autokinase activity that results in its phosphorylation in vitro (Taraporewala et al., 1999). In contrast, activities similar to those of NSP2 were not found for reovirus ␴NS (Gillian et al., 2000). To further define the similarities among the potential homologs, NS2, NSP2, and ␴NS, we expressed BTV NS2 with a C-terminal His-tag in Escherichia coli, purified the recombinant protein (rNS2) to homogeneity, and examined the properties and enzymatic activities of the protein. As expected, rNS2 self-assembled into large homomultimers with non-specific RNA-binding activity. Notably, the protein was like rNSP2 in that it possessed enzymatic activity that catalyzed the hydrolysis of all four NTPs. However, while rNSP2 only cleaved the ␥-phosphodiester bond of NTPs, rNS2 had a strong phosphatase activity that cleaved all three phosphodiester bonds (␣, ␤, and ␥). The BTV protein also lacked any detectable autokinase activity. While NS2, NSP2, and ␴NS share several non-enzymatic features, our results indicate that these three proteins have no single common enzymatic activity. However, the interaction of these proteins with other viral proteins may produce complexes that have shared enzymatic characteristics. RESULTS Expression and purification of rNS2 A cDNA of segment 8 of BTV serotype 10 (US strain 8) was prepared by reverse transcription (RT) and polymerase chain reaction (PCR). The open reading frame (ORF) coding for NS2 in the cDNA had the same nucleotide and amino acid sequence as that previously (GenBank Accession No. D00500). The NS2 ORF was inserted into the IPTG-inducible bacterial expression vector, pQE60, such that the expressed rNS2 protein had a C terminus tag of 6 His residues. Following expression in E. coli, rNS2 was purified from the soluble fraction of the bacterial lysate under non-denaturing conditions using Ni-affinity chromatography. Homogenous preparations of rNS2 were obtained as evaluated by electrophoresis on a polyacrylamide gel containing sodium dodecyl sulfate (SDS– PAGE) and by staining with Coomassie blue (Fig. 1A). The identity of the protein was confirmed by Western blot analysis with an anti-His antibody (Fig. 1B). Although the protein has a theoretical MW of 41 kDa, rNS2 migrated with an apparent MW of ⬃48 kDa (Fig. 1). This unexpectedly high MW for NS2 has been reported previously (Theron and Nel, 1997). The MW of rNS2 was not reduced by treatment with phosphatase (data not shown), indicating that its higher MW (48 kDa) does not stem from the presence of phosphate groups. To further verify

FIG. 1. Expression and purification of rNS2. Proteins were resolved by SDS–PAGE and stained with Coomassie blue (A) or blotted onto nitrocellulose and probed with anti-His antibody (B). Lanes 1 and 4, protein standards; lane 2, bacterial lysate prior to induction; lane 3, bacterial lysate after induction with 1 mM IPTG; lane 5, His-tagged rNSP2 eluted from Ni-affinity column.

that the purified protein was indeed BTV NS2, the purified protein was resolved by SDS–PAGE, eluted, and then subjected to N-terminal micro-sequencing. The analysis showed that the first 20 amino acids of the protein were MEQKQRXFTKNIFVLDVTAK. Using the sprot database (http://prowl1.rockefeller.edu), the sequence of the purified protein was determined to be identical to that of NS2 of BTV serotype 10 (US strain 8). Sedimentation analysis of rNS2 NS2 expressed in infected mammalian cells has been reported to form a ⬃22S complex with RNA (Huismans et al., 1987). When purified from recombinant baculovirusinfected cells and treated with RNase A, NS2 is detected as homomultimers of ⬃7S (Uitenweerde et al., 1995). To examine the status of the bacterial expressed rNS2, the purified protein and the protein size markers, ␥-globulin (7S), catalase (11.3S), and thyroglobulin (19S), were sedimented through linear 5–20% sucrose gradients. Electrophoretic analysis of the gradient fractions revealed that rNS2 formed large complexes sedimenting from 18 to 22S (Fig. 2A). However, treatment of rNS2 with RNase A prior to sedimentation caused the protein to sediment from 8 to 10S (Fig. 2B). The effect of RNase A treatment on the sedimentation indicated that in the preparations of purified rNS2 most of the protein is complexed with bacterial RNA. The monomeric form of rNS2 was expected to be recovered in fractions 1–3 of the sucrose gradients. Because no protein was detected in these fractions, even after RNase treatment, only the multimeric form of the protein must exist within the rNS2 preparation. Moreover, because only the multimeric form is present, the RNA-binding activity observed for the protein is likely associated with rNS2 multimers. Based on its MW and S value, the 8–10 S multimer is estimated to have a size between 140 and 250 kDa and to consist of 6 ⫾ 2 subunits of rNS2.

PHOSPHATASE ACTIVITY OF BLUETONGUE VIRUS NS2 PROTEIN

FIG. 2. Sedimentation analysis of rNS2. Purified rNS2 (50 ␮g), untreated (A) or treated with 5 ␮g of RNase A for 10 min prior to sedimentation (B), was centrifuged through 5–20% sucrose gradients. Fractions (1 ml) from the gradients were analyzed by SDS–PAGE and Coomassie Blue staining. One microgram of rNS2 was co-electrophoresed as the marker (M). The pellet (P) is contained within fraction 12. S values were calculated by co-sedimentation of the following protein size markers, thyroglobulin (650 kDa, 19S), catalase (250 kDa, 11.3S), and ␥-globulin (156 kDa, 7S).

RNA-binding activity of rNS2 To determine whether rNS2 possessed RNA-binding activity, 33P-labeled rotavirus gene 8 ssRNA and purified

223

rNSP2 were mixed together, exposed to UV light to cross-link RNA–protein complexes, and then treated with RNase T1 (Figs. 3A and 3B). SDS–PAGE and autoradiography demonstrated that UV cross-linking resulted in the radiolabeling of a protein co-migrating with rNS2 (Figs. 3A and 3B, lane 1). This result indicated that rNS2 has affinity for ssRNA. The ⬃30-kDa band detected in Fig. 3B, lane 1, is correlated with the addition of RNase T1 to the sample and therefore likely represents an RNA fragment of the rotavirus probe. The absence of any other novel radiolabeled band in lane 1, other than those of rNS2 and the 30-kDa material, ruled out the possibility that other contaminating proteins were present in the rNS2 preparation that possessed RNA-binding activity. Co-analysis of rNSP2 showed that it too became radiolabeled when exposed to UV light in the presence of the rotavirus 33 P-RNA probe (Figs. 3A and 3B), indicating as reported earlier that it is also an RNA-binding protein (Kattoura et al., 1992). To further analyze its RNA-binding activity, rNS2 was incubated with either of two 33P-labeled non-BTV RNA

FIG. 3. Nonspecific ssRNA binding by rNS2. (A and B) Reaction mixtures containing 1. 5 ␮g of purified rNS2 or rNSP2, 33P-labeled rotavirus gene 8 mRNA, and RNase- inhibitor were incubated at room temperature. Some of the samples were then exposed to UV light and digested with RNase T1. After electrophoresis of the samples and protein standards on SDS–polyacrylamide gels, rNS2 and rNSP2 were located in the gels by staining with Coomassie blue (A). Radiolabeled rNS2 and rNSP2 (arrowheads) were detected in the gels by autoradiography (B). The ⬃30-kDa band (*) in lanes 1, 3, 5, 6, and 8 correlates with the addition of RNase T1 and therefore probably represents a digestion product of the gene 8 probe. (C and D) Either 0.3 pmol of 32P-labeled rotavirus gene 8 mRNA (C) or 1 pmol of a 32P-labeled 0.2-kB luciferase RNA (D) was incubated in the absence or presence of differing amounts of rNS2 for 30 min at room temperature. One reaction (C, lane 4) was subsequently treated with 40 ␮g of proteinase K for 15 min at room temperature. The probe–protein complexes in the samples were detected by electrophoresis on either a non-denaturing 1% agarose gel (C) or a non-denaturing 6% polyacrylamide gel (D), followed by autoradiography.

224


FIG. 4. Absence of an autokinase activity for rNS2. One microgram of rNS2 (lanes 1 and 2) or rNSP2 (lanes 3 and 4) was incubated with (lanes 2 and 4) or without (lanes 1 and 3) 10 ␮Ci of [␥- 32P]ATP in kinase buffer for 1 h at 37°C. The proteins were detected by SDS–PAGE and staining with Coomassie blue (A). The 32P-labeled proteins were identified by autoradiography (B).

probes: 1-kB rotavirus gene 8 RNA or 0.2-kB luciferase RNA. The reaction mixtures were then analyzed for the presence of RNA–protein complexes by electrophoresis on a non-denaturing 1% agarose gel (Fig. 3C) or on a non-denaturing 6% polyacrylamide gel (Fig. 3D). The gel mobility shift assays showed that incubation of rNS2 with the probes resulted in the formation of 32P-labeled RNA– protein complexes. The presence of protein in the complexes was verified by treatment with proteinase K, which caused a loss of the complex and the appearance of unbound radiolabeled RNA (Fig. 3C, lane 4). In reaction mixtures where 24 pmol of rNSP2 was incubated with 0.3 pmol of RNA, unbound RNA was not detected (Fig. 3C, lane 3). The absence of free RNA under these conditions, where the ratio of rNS2 to RNA was 80:1 (24 pmol of protein:0.3 pmol of RNA), indicated that multiple subunits of rNS2, and likely rNS2 multimers (Fig. 2), were binding to the RNA molecules in the reaction mixture. The broad range of RNA–protein complexes detected on the polyacrylamide gel probably stemmed from the superior capacity of this electrophoretic system, as opposed to the agarose gel system, to resolve complexes that differed in the number of rNS2 subunits bound to each molecule of RNA (Fig. 3D).

The possibility that cellular kinases could phosphorylate rNS2 was also examined. In this experiment, rNS2 and rNSP2 were first immuno-adsorbed to protein A–Sepharose beads using an anti-His antibody. Following stringent washing of the immuno-adsorbed proteins with buffer containing 0.1% SDS, the bound proteins were then incubated in kinase buffer containing [␥- 32P]ATP, MgCl 2, and, as a source of cellular kinases, a soluble extract prepared from uninfected MA104 cells. Following incubation at 37°C for 1 h, recombinant proteins bound to the beads were resolved by SDS–PAGE and were visualized by staining with Coomassie blue (Fig. 5A). Autoradiography was then used to determine whether rNS2 or rNSP2 had undergone phosphorylation due to incubation with cellular kinases (Fig. 5B). The results showed that both, rNS2 and rNSP2, were phosphorylated but only when incubated with the cellular extract (Fig. 5).

Phosphorylation of rNS2 To determine whether rNS2 has an associated kinase activity that can cause its autophosphorylation, BTV rNS2 and rotavirus rNSP2, as a control, were incubated separately in kinase buffer containing [␥- 32P]ATP and MgCl 2. Afterward, the recombinant proteins in the reaction mixtures were resolved by SDS–PAGE and stained with Coomassie blue (Fig. 4A), and the presence of 32 P-labeled rNS2 and rNSP2 in the gel was detected by autoradiography (Fig. 4B). The results showed that under these reaction conditions, rNSP2 but not rNS2 was phosphorylated (lane 2 vs 4). Therefore, unlike rotavirus rNSP2 (Taraporewala et al., 1999), BTV rNS2 has no autokinase activity in vitro.

FIG. 5. rNS2 and rNSP2 are substrates for cellular kinases. rNS2 (lanes 2 and 3) or rNSP2 (lanes 4 and 5) were immuno-adsorbed onto protein A–Sepharose beads using anti-His antibody. Immuno-adsorbed proteins were incubated with [␥- 32P]ATP in kinase buffer in the presence (lanes 3 and 5) or absence (lanes 2 and 4) of an extract prepared from uninfected MA104 cells. As a control, anti-His antibody bound to protein A–Sepharose beads (with no immuno-adsorbed protein) was incubated by itself in kinase buffer containing [␥- 32P]ATP (lane 1). Cellular extract incubated in kinase buffer containing [␥- 32P]ATP, in the absence of beads, was also analyzed (lane 6). The products of the labeling reactions were analyzed by SDS–PAGE followed by Coomassie blue staining of the gels (A) and autoradiography (B). Lanes with protein MW markers are labeled as “M.”


225

FIG. 6. Nucleotidyl phosphatase activity of rNS2. (A) Reaction mixtures with no added protein (lane 1) or with 0.5 ␮g of either rNS2 or rNSP2 (lane 2) and 10 ␮Ci of each of the four [␣- 32P]NTPs were incubated for 1 h at 37°C. The products of the reaction mixtures were resolved by TLC and detected by autoradiography. The positions of NTP, NDP, NMP, and P i, were determined by co-chromatography of markers (M).

Thus, rNS2 and rNSP2 are substrates for phosphorylation by cellular kinases. The lack of phosphorylation of immuno-adsorbed rNSP2 when incubated without cellular extracts (Fig, 5B, lane 4) suggested that the autokinase activity of NSP2, previously demonstrated in Fig. 4, lane 4, was inactivated in this experiment due to its exposure to SDS or its immuno-absorption with anti-His antibody.

were analyzed by TLC. As shown in Fig. 7, the hydrolysis of [␣- 32P]ATP by the rNS2 preparation increased up to 55°C and was totally inhibited at 65°C. Since bacterial phosphatases are characteristically thermoresistant at 65°C (Tomazic-Allen, 1991), it is unlikely that the phosphatase activity originates from anything other than rNS2.

Nucleotidyl phosphatase activity of rNS2

To further analyze the NTP hydrolysis activities of rNS2 and rNSP2, varying amounts of each protein were separately incubated at 37°C for 1 h in reaction mixtures containing the same amount of [␣- 32P]ATP (12.5 pmol;

To determine whether or not rNS2 functions as an NTPase, the recombinant protein was incubated with each of the four [␣- 32P]NTPs, and the products of the reaction mixtures were analyzed by thin-layer chromatography (TLC) and autoradiography. As a control, rNSP2 was also incubated with [␣- 32P]ATP. The results showed that rNS2 generated 32P i from each of the ␣-labeled NTPs, thus revealing that the protein hydrolyzed the ␣ and ␤ phosphodiester bonds of each (Fig. 6). As illustrated more clearly in Fig. 7 (lane 2), incubation of rNS2 with [␣- 32P]ATPs also generated [␣- 32P]ADPs, indicating that the protein was also able to cleave the ␥ phosphodiester bond. Hence, rNS2 has associated nucleotidyl phosphatase activity that can hydrolyze the ␣, ␤, and ␥phosphodiester bonds of NTPs. This finding contrasts with those found for rNSP2 that showed that this protein could only hydrolyze the ␥ phosphodiester bond of NTPs (Fig. 6, Taraporewala et al., 1999). Based on the quantitation of the amount of 32P i generated from each NTP (data not shown), the nucleotidyl phosphatase activity of rNS2 exhibits a preference for the NTP substrates in the order of ATP ⬎ GTP ⬎ UTP/CTP. To address the possibility that a bacterial protein copurifying with rNS2 contributed to the observed phosphatase activity, rNS2 was incubated at 37, 45, 55, and 65°C with [␣- 32P]ATP, and the products of the reactions

ATP hydrolysis by rNS2

FIG. 7. Effect of temperature on the nucleotidyl phosphatase activity of rNS2. Reaction mixtures with no added protein (lanes 1, 3, 5, and 7) or 0.5 ␮g of rNS2 (lanes 2, 4, 6, and 8) were incubated with 10 ␮Ci of [␣-32P]ATP and incubated for 1 h at 37°C (lanes 1 and 2), 45°C (lanes 3 and 4), 55°C (lanes 5 and 6), or 65°C (lanes 7 and 8). The products of the reactions were resolved by TLC and detected by autoradiography. The positions of NTP, NDP, NMP, and Pi, were determined by co-chromatography of markers (marked in each panel as “M”), prepared by partial digestion of [␣- 32P]NTPs with tobacco acid pyrophosphatase and calf intestinal alkaline phosphatase on the same PEI-cellulose sheet.

226


FIG. 8. Characterization of the NTP hydrolysis activity of rNS2 and rNSP2. (A) Reaction mixtures with no protein (lane 1) or from 6 to 96 pmol of either rNSP2 (lanes 2–6) or rNS2 (lanes 7–11) were incubated with 12.5 pmol of [␣- 32P]ATP for 1 h at 37°C. The product were resolved by TLC, detected by autoradiography, and quantified with a phosphorimager. (B) The percentage of [␣- 32P]ATP hydrolyzed by rNS2 (Œ) and rNSP2 (䡲) in the reaction mixtures was plotted as a function of enzyme concentration and calculated as follows: [quantity of 32P-labeled (ADP ⫹ AMP⫹ P i)/quantity of 32 P-labeled (ATP ⫹ ADP⫹ AMP⫹ P i)] ⫻ 100. (C) To analyze the effect of excess competitor substrate on hydrolysis, reaction mixtures were prepared that contained no protein or 30 pmol of either rNS2 or rNSP2 and 12 pmol of [␣- 32P]ATP. To these mixtures was added a 0-, 8-, 16-, 32-, 64-, 128- or 256-fold mass excess of unlabeled ATP. After incubation for 1 h, the products of all the reactions were resolved by TLC, detected by autoradiography, and quantified with a phosphorimager. The markers were generated by partial digestion of [␣- 32P]ATP with tobacco acid phosphatase and calf intestinal alkaline phosphatase. (D) The percentage of [␣- 32P]ATP hydrolysis by rNS2 (Œ) and rNSP2 (䡲) in the reaction mixtures was plotted as a function of the fold excess of cold competitor ATP and was calculated as follows: [quantity of 32P-labeled (ADP ⫹ AMP⫹ P i)/quantity of 32P-labeled (ATP ⫹ ADP⫹ AMP⫹ P i)] ⫻ 100.

0.625 ␮M). The products of the reaction mixtures were analyzed by TLC, and the percentage of ATP hydrolysis in each reaction was calculated and plotted as a function of enzyme (i.e., rNS2 and rNSP2) concentration (Figs. 8A and 8B). The results showed that in the linear range, rNS2 hydrolyzed three to four times more of the substrate, [␣- 32P]ATP, than rNSP2, per molar equivalent amount of protein. Thus, the specific activity of rNS2 for NTP hydrolysis is significantly higher than that of rNSP2. The NTP hydrolysis activities of the two proteins was also examined by incubating 30 pmol of rNSP2 and rNS2, separately, at 37°C for 1 h with 12.5 pmol of [␣- 32P]ATP in

the presence of increasing amounts of unlabeled competitor ATP. The products of the reaction mixtures were then analyzed by TLC and autoradiography. The analysis showed that in the absence of unlabeled competitor ATP, molar equivalent amounts of rNS2 and rNSP2 hydrolyzed 47 and 15% of the [␣- 32P]ATP, respectively (Figs. 8C and 8D). The percentage of [␣- 32P]ATP hydrolysis by rNSP2 remained at ⬃15%, even in the presence of 16-fold excess of competitor ATP. Sharp decreases in the percentage of [␣- 32P]ATP hydrolyzed by rNSP2 were not observed with further increases in competitor substrate to an excess of 256-fold. Thus, the NTPase activity of rNSP2


was affected by the products of hydrolysis through uncompetitive or noncompetitive feedback inhibition as was reported previously (Taraporewala et al., 1999). In contrast, the hydrolysis of [␣- 32P]ATP by rNS2 was reduced from 47% in the absence of competitor ATP to less than half, 22%, in the presence of 8-fold excess of cold competitor substrate and to 17% in the presence of 16fold excess cold ATP. Additional increases in competitor ATP to an excess of 256-fold further reduced the percentage of [␣- 32P]ATP hydrolysis by rNS2 to just 7%. These results suggest that rNS2 catalyzes several rounds of ATP hydrolysis without reduction in activity and that the phosphatase activity of rNS2 is subject to competitive inhibition. DISCUSSION The mode of genome packaging and core assembly in the VIBs of BTV-, rotavirus-, and reovirus-infected cells is poorly defined (Brookes et al., 1993; Petrie et al., 1994). In an effort to gain insight into these processes, the structure and function of the nonstructural proteins contained within VIBs was studied. Two such proteins, NS2 of BTV and NSP2 of rotavirus, are expressed at high concentrations in infected cells and are believed to participate in genome packaging and core assembly (Theron and Nel, 1997; Helmberger-Jones and Patton, 1986; Gallegos and Patton, 1989; Chen et al., 1990). Recently, it was shown that rotavirus NSP2, apart from binding RNA nonspecifically and forming homomultimers, also possesses NTPase and autokinase activity (Taraporewala et al., 1999). Since BTV NS2 is considered to be the functional homolog of rotavirus NSP2, the BTV protein was expressed in E. coli, purified to homogeneity, and analyzed for similar enzymatic activities. The rNS2 exhibited non-specific affinity for ssRNA as was reported earlier for the authentic protein derived from BTV-infected cells (Huismans et al., 1987). Sedimentation analysis showed that the bacterial-expressed NS2 assembled into 8–10S homomultimers and thus had a quaternary configuration similar to that of NS2 expressed in BTV-infected cells or in insect cells infected with a recombinant baculovirus encoding this protein (Uitenweerde et al., 1995). Based on similarities in the S values of the homomultimers of rNS2 and rNSP2, it appears that both consist of the same number of subunits. From its S value and MW, the rNS2 homomultimer may be predicted to consist of about six protein subunits, consistent with the results of Uitenweerde et al. (1995). Since only the 8–10S multimeric form of the protein was identified in the RNase-treated rNS2 preparations, the RNA-binding activity detected for the protein is probably associated with the multimers. Uitenweerde et al. (1995) have also demonstrated that sucrose gradient purified multimers of NS2 show poly(U)–Sepharose binding. In this respect, NS2 is like NSP2 in that the RNA-binding

227

activity of the NSP2 is associated with multimers of this protein (Taraporewala et al., 1999). BTV rNS2 possessed a nucleotidyl phosphatase activity that enabled the protein to hydrolyze the ␣, ␤, and ␥ phosphodiester bonds of all four NTPs. As substrates, the phosphatase activity displayed a preference for purine over pyrimidine NTPs. While this result shows that rNS2, like rNSP2, can hydrolyze NTPs, the nature of the activity for rNS2 is quite different from that of rNSP2 as the latter protein possesses an NTPase activity that can only hydrolyze the ␥ phosphodiester bond of NTPs. rNS2 showed a higher rate of NTP hydrolysis than rNSP2. In addition, hydrolysis of radiolabeled ATP by rNS2 was competitively inhibited in vitro by the presence of cold ATP, while under the same reaction conditions, hydrolysis of radiolabeled ATP by rNSP2 was noncompetitively or uncompetitively inhibited. These findings serve to emphasize the fundamental differences in the NTP hydrolysis activities of NS2 and NSP2. The function of the enzyme activities of these proteins in the biology of BTV and rotavirus are unknown, but the hydrolysis of NTPs by NS2 and NSP2 may be important for generating energy that is used for transport, sorting or packaging of viral plus-strand RNAs. NS2 expressed in BTV-infected cells was shown earlier to be phosphorylated at serine residues (Devaney et al., 1998; Thomas et al., 1990). However, analysis of rNS2 indicated that it lacks autokinase activity in vitro, and, therefore, the protein may undergo phosphorylation via the kinase activity of cellular proteins or of another viral protein. This conclusion is in agreement with that of Theron et al., (1994), who also found that bacterial-expressed rNS2 lacked autokinase activity in vitro but was phosphorylated when incubated with lysates prepared from uninfected insect cells. These authors proposed from their results that a ubiquitous cellular kinase was responsible for the phosphorylation of the protein in infected cells. In contrast to rNS2, when rNSP2 is incubated by itself in vitro, the protein exhibits an autokinase activity that results in its phosphorylation. Also, NSP2 is phosphorylated during transient expression in mammalian cells (Taraporewala et al., 1999) and when incubated with uninfected cell lysates. Based on present information, it seems that both, NS2 and NSP2 are acted upon by cellular kinases and that NSP2, additionally, can phosphorylate itself. A distinctive feature of these proteins is that in infected cells, phosphorylated forms of NS2 can be detected while phosphorylated forms of NSP2 cannot. The absence of phosphorylated NSP2 in the infected cell is an enigma, considering the significant autokinase activity observed for the protein in vitro. Perhaps in the infected cell, the phosphorylated form of the NSP2 is extremely short lived due to it interaction with an accessory protein that mediates the removal of phosphate groups from the protein, which otherwise may negatively affect

228


the function of NSP2. The protein suspected to affect the phosphorylation of NSP2 in the infected cell is NSP5, the phosphoprotein and putative kinase of rotavirus (Blackhall et al., 1997, Poncet et al., 1997). NSP2, when coexpressed with NSP5, forms VIB-like structures and enhances the extent of NSP5 phosphorylation (Fabbretti et al., 1999; Afrikanova et al., 1998). In conjunction, these studies suggest that the functions of NSP2 and NSP5 in rotavirus replication are coupled. The reovirus nonstructural protein ␴NS has characteristics similar to those of NS2 and NSP2: these includes its MW (41 kDa), its nonspecific affinity for ssRNA, its intrinsic ability to assemble into large homomultimers that bind RNA, and its accumulation in VIBs of infected cells (Gillian and Nibert, 1998; Gillian et al., 2000; Mbisa et al., 2000). However, unlike NS2 and NSP2, the reovirus ␴NS protein does not have any detectable activity that catalyzes the hydrolysis of any of the phosphodiester bonds of NTPs (Gillian et al., 2000). Also, there is no evidence that the protein possesses an autokinase activity or exists in a phosphorylated form. Hence, it is not clear whether ␴NS2 is a functional homolog of NS2 and NSP2. It may be that ␴NS interacts with another viral protein, and in combination, these proteins have enzymatic activities reminiscent of either NS2 or NSP2. Indeed, just like rotavirus NSP2, which requires the presence of NSP5 to accumulate in VIBs, ␴NS is dependent on another protein, ␮2, to accumulate in VIBs (Mbisa et al., 2000). Remarkably, ␮2 has affinity for ssRNA and possesses NTPase activity (Brentano et al., 1998; Noble and Nibert, 1997). Therefore, ␴NS and ␮2 together would appear to have enzymatic activity similar to BTV NS2 and to rotavirus NSP2-NSP5, and, as a result, these reovirus proteins operating in combination may function as homologs of NS2 and NSP2–NSP5. Despite the evidence that NS2, NSP2, and ␴NS may be playing related roles in virus replication, these proteins display no sequence homology. MATERIALS AND METHODS Construction of the NS2 expression vector Total dsRNA of BTV serotype 10 (US strain 8), kindly provided to us by Siba Samal (University of Maryland), was denatured by resuspending in 90% dimethylsulfoxide followed by heating to 95°C for 2 min. The NS2 ORF in genome segment 8 was amplified from the denatured dsRNA using the Expand High Fidelity RT (Boehringer Mannheim) and the plus-sense primer CCGAAACCatggagcaaaagcaac (NcoI site is underlined) and the minus-sense primer CGGAGATCTaacgccgaccggcaatatg (BglII site is underlined). Virus-specific sequences in the primers are in lowercase. The amplified product was gel purified, digested with NcoI and BglII, and ligated into the IPTGinducible expression vector pQE60 (Qiagen), similarly digested with NcoI and BglII. Following transformation into E. coli DH5␣, bacteria with the appropriate plasmid (pQE60-

NS2) were identified based on antibiotic resistance, plasmid size, and restriction enzyme digestion. The plasmid pQE60-NS2 was amplified and the sequence accuracy of the NS2 ORF was confirmed by automated sequencing with an ABI PRISM 310 Genetic Analyzer (PE Applied Biosystems). pQE60-NS2 was then electroporated into E. coli M15 carrying the pREP4 repressor plasmid. Appropriate transformants were identified based on antibiotic resistance, restriction enzyme digestion, and expression of rNS2. In pQE60-NS2, the ORF for NS2 is situated immediately upstream from six in-frame codons for His. Thus, rNS2 expressed from pQE60-NS2 is tagged at its C terminus with six His residues. The expression vector for rotavirus rNSP2 was constructed as described previously (Taraporewala et al., 1999). Expression and purification of rNS2 M15[pREP4] bacteria containing pQE60-NS2 were grown to an OD 600 of 0.5 in Terrific Broth (Quality Biologics) at 30°C, and the expression of rNS2 was induced by adding IPTG to a final concentration of 250 ␮M. After incubation for 12 h at 30°C, the bacteria were recovered by centrifugation at 4000 g for 30 min, and the 6xHistagged rNS2 was purified under native conditions on a Ni-nitrilotriacetic acid (NTA) agarose column according to the protocol of the manufacturer (Qiagen). The final eluate was dialyzed against low-salt buffer [LSB: 2 mM Tris–HCl (pH 7.2), 0.5 mM EDTA, 0.5 mM dithiothreitol (DTT)] for 48 h at 4°C. The concentration of the purified protein was determined by Bradford assay using bovine serum albumin (BSA) as the protein standard and by comparison with known amounts of BSA co-electrophoresed on SDS–polyacrylamide gels and stained with Coomassie blue R-250. Purified rNS2 was adjusted to a concentration of 0.1–0.25 mg/ml and stored at 4°C. The same protocol was used to express and purify His-tagged rNSP2 except that rNSP2 was induced by adding IPTG to a final concentration of 1 mM, followed by incubation at 37°C for 4–5 h (Taraporewala et al., 1999). Protein sequencing To conduct direct N-terminal sequence analysis, purified rNS2 was resolved by SDS–PAGE with tricine running buffer (Novex) and electro-transferred to a ProBlott membrane (Applied Biosystems). Protein bands on the membrane were visualized by staining with 0.1% Coomassie blue R-250 in 40% methanol and 1% acetic acid, followed by destaining in 50% methanol. The band of interest was excised and subjected to N-terminal sequence analysis with a model 477A protein sequencer coupled to a model 120A phenylthiohydantoin (PTH) analyzer (Applied Biosystems) with the program NORMAL-1.


229

Rate zonal centrifugation

Gel mobility shift assay

Recombinant NS2 and protein standards were layered separately onto 12 ml 5–20% (w/vol) sucrose gradients in LSB and centrifuged at 200,000 g for 16 h in a Beckman SW40Ti rotor at 4°C. The following three proteins were used as size markers in the gradients: thyroglobulin (650 kDa, 19S), catalase (250 kDa, 11.3S), and ␥-globulin (156 kDa, 7S). One milliliter fractions were collected from the gradients and analyzed for protein content by electrophoresis on 12% polyacrylamide gels containing SDS (SDS–12% PAGE) (Laemmli, 1970) and staining with Coomassie blue R-250.

The procedures used to study the rNS2-RNA interaction was a modified version of that described by Thomas et al. (1990). Rotavirus 32P-labeled gene 8 RNA was first denatured by heating to 95°C for 2 min. The denatured RNA probe (0.3 pmol) was incubated with 12–24 pmol of rNS2 in LSB containing 1 mM NaCl and 11 units of ANTI-RNase (Ambion) for 30 min at room temperature. The mixtures were analyzed by electrophoresis on a non-denaturing 1% agarose gel containing 50 mM Trisglycine and 0.1% deoxycholate (Konarska and Sharp, 1987). The probe-protein complexes were detected in the gel by autoradiography.

In vitro synthesis of RNA The DNA template for synthesis of the Luc200 RNA probe was produced by amplifying a 0.2-kB portion of the luciferase gene in the plasmid pGL2 (Promega) with Taq DNA polymerase (Life Technologies) and the positivesense primer, taatacgactcactataccatggaagacgccaaaaacataaagaaagg (the T7 promoter sequence is underlined) and the negative-sense primer, catttcgaagtactcagc. 33P-labeled Luc200 RNA was transcribed from the amplified DNA in the presence of 10 ␮Ci of [␣- 33P]UTP per 20-␮l reaction mixture using the Ambion MEGAshortscript system (Taraporewala et al., 1999). The radiolabeled RNA was purified by gel electrophoresis on and elution from 8% polyacrylamide gels containing 7 M urea (Patton, 1996). To produce the template for synthesis of the 1-kB gene 8 RNA, the plasmid pSP65g8(⫹) was linearized with SacII, and blunt-ended by treatment with T4 DNA polymerase. 32P-labeled gene 8 RNA was synthesized from the linearized DNA using the Ambion MAXIscript system and [␣- 32P]UTP (Patton, 1996). To produce 33P-labeled gene 8 RNA, transcription was carried out in the presence of 10 ␮Ci of [␣- 33P]ATP instead of [␣- 32P]UTP. The RNA product was purified by phenol-chloroform extraction and isopropanol precipitation. The quality of the gene 8 RNAs was assessed by electrophoresis on 5% polyacrylamide gels containing 7 M urea (Patton, 1996). RNA concentrations were calculated from optical densities at 260 nm. UV cross-linking assay Approximately 1.5 ␮g of rNS2 or rNSP2 was incubated 1 pmol of 33P-labeled gene 8 RNA and 5 units of AntiRNase (Ambion) in LSB at room temperature for 30 min. Afterward, some samples were exposed to UV light (254 nm) for 10 min at a distance of 4 cm and then treated with 33 units of RNase T1 (Life Technologies) for 30 min at 37°C. The samples were then analyzed by electrophoresis on 10% polyacrylamide gels containing SDS. The location of rNS2 and rNSP2 were determined by staining the gels with Coomassie blue and radiolabeled proteins were detected by autoradiography.

NTPase/nucleotidyl phosphatase assay Typically, reaction mixtures for the NTPase/phosphatase assay contained 0.5–2 ␮g of rNSP2 or rNS2, 50 mM Tris–HCl (pH 7.5), 5 mM MgCl 2 and 10 ␮Ci of [␣- 32P]ATP, -GTP CTP, or -UTP (3000 Ci/mmol, Amersham) in a final volume of 20 ␮l. After incubation at 37°C for 1 h, 1 ␮l of 0.5 mM EDTA was added to the reaction mixtures, and the samples were then deproteinized by phenol:chloroform extraction. One-microliter aliquots of each were spotted onto PEI-cellulose F sheets (EM Science) along with the markers, 32P-labeled NTP, NDP, NMP, and P i, and were resolved by ascending TLC in 1.2 M LiCl. Radiolabeled spots on the sheets were detected by autoradiography and quantified with a phosphorimager. The NDP and NMP markers were made by limited hydrolysis of [␣- 32P]NTP with tobacco acid pyrophosphatase (Epicentre) and the P i marker was made by hydrolysis of the radiolabeled NTP with calf intestinal phosphatase (New England Biolabs). In vitro phosphorylation assay The standard reaction mixture for in vitro phosphorylation contained 1 ␮g of rNS2 or rNSP2, 10 ␮Ci of [␥- 32P]ATP (3000 Ci/mmol), 50 mM Tris–HCl (pH 7.5), and 5 mM MgCl 2 in a final volume of 10 ␮l and was incubated at 37°C for 1 h. The phosphorylation assays were terminated by adding SDS sample buffer and heating to 100°C for 2 min. The phosphorylated proteins were evaluated by SDS–12% PAGE and autoradiography. Kinase assay with uninfected mammalian cell extracts To determine whether cellular kinases could phosphorylate rNS2 and rNSP2, fetal rhesus monkey kidney (MA104) cells maintained in medium 199 (Gibco) containing 5% fetal bovine serum were harvested and washed in 20 mM Tris–HCl, pH 7.5, and 150 mM NaCl. The cells were lysed in buffer containing 50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 30 mM KCl, 1% Triton X-100, 1% sodium deoxycholate, 20 mM EDTA, and 1 mg/ml of aprotinin.

230


After sonication on ice, the clarified cell extract was recovered by centrifugation a 10,000g for 30 min. For the kinase assay, 10 ␮g of rNS2 and rNSP2 was immuno-adsorbed onto protein A–Sepharose beads using mouse anti-Penta-His monoclonal antibody (Qiagen) at a dilution of 1:100 in RIPA buffer (50 mM Tris–HCl, pH 8, 150 mM NaCl, 1% Triton X-100, 0.6% deoxycholate, and 0.1% SDS). The immuno-adsorbed rNS2 and rNSP2 were incubated separately with 10 ␮Ci of [␥- 32P]ATP in kinase buffer (50 mM Tris–HCl, pH 7.5, and 5 mM MgCl 2) with or without the MA104 cell extracts, for 1 h at 37°C. After the kinase reaction, the protein A–Sepharose beads were washed three times with RIPA buffer, then mixed with SDS sample buffer and analyzed by SDS–PAGE and autoradiography (Laemmli, 1970). Western blot analysis Proteins were resolved by SDS–12% PAGE and then electroblotted onto nitrocellulose sheets (Millipore). The blots were blocked by soaking in phosphate-buffered saline containing 5% skim milk suspension. Subsequently, the blots were incubated with anti-His antibody (Qiagen) at a dilution of 1:500. Goat anti-mouse HRPconjugated antibodies were used as secondary antibodies at a dilution of 1:5000. The blots were developed by the Sigma Fast system. ACKNOWLEDGMENTS We are grateful to Siba Samal for the gift of BTV RNA and to Karen Kearney for critical review of the manuscript. We acknowledge the technical assistance of Melinda Jones in sequencing of the NS2 ORF and of Mark Garfield in micro-sequencing of the N-terminal amino acids of NS2.

REFERENCES Afrikanova, I., Fabretti, E., Miozzo, M. C., and Burrone, O. R. (1998). Rotavirus NSP5 phosphorylation is up-regulated by interaction with NSP2. J. Gen. Virol. 79, 2679–2686. Blackhall, J., Fuentes, A., Hansen, K., and Magnusson, G. (1997). Serine protein kinase activity associated with rotavirus phosphoprotein NSP5. J. Virol. 71, 138–144. Brentano, L., Noah, D. L., Brown E. G., and Sherry, B. (1998). The reovirus protein mu2, encoded by the M1 gene, is an RNA-binding protein. J. Virol. 72, 8354–8357. Brookes, S. M., Hyatt, A. D., and Eaton, B. T. (1993). Characterization of virus inclusion bodies in bluetongue virus-infected cells. J. Gen. Virol. 74, 525–530. Chen, D., Gombold, J. L., and Ramig, R. F. (1990). Intracellular RNA synthesis directed by temperature-sensitive mutants of simian rotavirus SA11. Virology 178, 143–151. Devaney, M. A., Kendall, J., and Grubman, M. J. (1988). Characterization of a nonstructural phosphoprotein of two orbiviruses. Virus Res. 11, 151–164. Eaton, B. T., Hyatt, A. D., and Brookes, S. M. (1990). The replication of bluetongue virus. Curr Top Microbiol Immunol 162, 89–118. Fabbretti, E., Afrikanova, I., Vascotto, F., and Burrone, O. R. (1999). Two non-structural rotavirus proteins, NSP2 and NSP5, form viroplasmlike structures in vivo. J. Gen. Virol. 80, 333–339.

Gallegos, C. O., and Patton, J. T. (1989). Characterization of rotavirus replication intermediates: a model for the assembly of single-shelled particles. Virology 172, 616–627. Gillian A. L., and Nibert, M. L. (1998). Amino terminus of reovirus nonstructural protein ␴NS is important for ssRNA binding and nucleoprotein complex formation. Virology 240, 1–11. Gillian, A. L., Schmechel, S. C., Livny, J., Schiff, L. A., and Nibert, M. L. (2000). Reovirus protein sigmaNS binds in multiple copies to singlestranded RNA, and shares properties with single-stranded DNA binding proteins. J. Virol. 74, 5939–5948 Grimes, J. M., Jakana, J., Ghosh, M., Basak, A. K., Roy, P., Chiu, W., Stuart, D. I., and Prasad, B. V. (1997). An atomic model of the outer layer of the bluetongue virus core derived from X-ray crystallography and cryoelectron microscopy. Structure 5, 885–893. Helmberger-Jones, M., and Patton, J. T. (1986). Characterization of subviral particles in cells infected with simian rotavirus SA11. Virology 15, 655–665. Huismans, H., Van Dijk, A. A., and Bauskin, A. R. (1987). In vitro phosphorylation and purification of a non-structural protein of bluetongue virus with affinity for their single stranded RNA. J. Virol. 61, 3589–3595. Kattoura, M. D., Chen, X., and Patton, J. T. (1994). The rotavirus nonstructural protein, NS35, (NSP2), forms 10S multimers and interacts with the viral RNA polymerase. Virology 202, 802–813. Kattoura, M. D., Clapp, L. L., and Patton, J. T. (1992). The rotavirus non-structural protein, NS35, is a nonspecific RNA-binding protein. Virology 191, 698–708. Konarska, M. M., and Sharp, P. A. (1987). Interactions between small ribonucleoprotein particles in formation of spliceosomes. Cell 49, 763–774. Laemmli, U. K. (1970). Cleavage of the structural proteins during assembly of the head bacteriophage T4. Nature 227, 680–685. Martin, S. A., and Zweerink, H. J. (1972). Isolation and characterization of two types of bluetongue virus particles. Virology 50, 495–506. Mbisa, J. L., Becker, M. M., Zhou, S., Dermody, T. S., and Brown, E. G. (2000). Reovirus ␮2 protein determines strain-specific differences in the rate of viral inclusion formation in L929 cells. Virology 272, 16–26. Mertens, P. P. C., Brown, F., and Sangar, A. V. (1984). Assignment of the genome segments of bluetongue virus type 1 to the proteins which they encode. Virology 135, 455–465. Noble, S., and Nibert M. L. (1997). Characterization of an ATPase activity in reovirus cores and its genetic association with core-shell protein lambda1. J. Virol. 71, 2182–2191. Patton, J. T. (1996). Rotavirus VP1 alone specifically binds to the 3⬘ end of viral mRNA but the interaction is not sufficient to initiate minusstrand synthesis. J. Virol. 70, 7940–7947. Petrie, B. L., Greenberg, H. B., Graham, D. Y., and Estes M. K. (1984). Ultrastructural localization of rotavirus antigens using colloidal gold. Virus Res. 1, 133–152. Poncet, D., Lindenbaum, P., Haridon, R. L., and Cohen, J. (1997). In vivo and in vitro phosphorylation of rotavirus NSP5 correlates with its localization in viroplasms. J. Virol. 71, 34–41. Taraporewala, Z. F., Chen, D., and Patton, J. T. (1999). Multimers formed by the rotavirus nonstructural protein NSP2 bind to RNA and have nucleoside triphosphatase activity. J. Virol. 73, 9934–9943. Theron J., and Nel, L. H. (1997). Stable protein-RNA interaction involves the terminal domains of bluetongue virus mRNA, but not the terminally conserved sequences. Virology 229, 134–142. Theron, J., Uitenweerde, J. M., Huismans, H., and Nel, L. H. (1994). Comparison of the expression and phosphorylation of the nonstructural protein NS2 of three different orbiviruses: Evidence for the involvement of an ubiquitous cellular kinase. J. Gen. Virol. 75, 3401– 3411. Thomas, C. P., Booth, T. F., and Roy, P. (1990). Synthesis of bluetongue virus-encoded phosphoprotein and formation of inclusion bodies by

PHOSPHATASE ACTIVITY OF BLUETONGUE VIRUS NS2 PROTEIN recombinant baculovirus in insect cells: It binds the single-stranded RNA species. J. Gen Virol. 71, 2073–2083. Tomazic-Allen, S. J. (1991). Recombinant bacterial alkaline phosphatase as an immunodiagnostic enzyme. Ann. Biol. Clin. (Paris) 49, 287–290. Uitenweerde, J. M., Theron, J., Stoltz, M. A., and Huismans, H. (1995). The multimeric nonstructural NS2 proteins of bluetongue virus, African horsesickness virus and epizootic hemorrhagic disease virus differ in their single-stranded RNA-binding ability. Virology 209, 624–632.

231

Van Dijk, A. A., and Huismans, H. (1988). In vitro transcription and translation of bluetongue virus mRNA. J. Gen. Virol. 69, 573– 581. Verwoerd, D. W., Els, H. J., De Villiers, E. M., and Huismans, H. (1972). Structure of the bluetongue virus capsid. J. Virol. 50, 783– 794. Zhao, Y., Thomas, C., Bremer, C., and Roy, P. (1994). Deletion and mutational analysis of bluetongue virus NS2 protein indicate that the amino but not the carboxy terminus of the protein is critical for RNA-protein interactions. J. Virol. 68, 2179–2185.