ABSTRACT
The plum pox potyvirus (PPV) protein CI is an RNA helicase whose function in the viral life cycle is still unknown. The CI protein contains seven conserved sequence motifs typical of RNA helicases of the superfamily SF2. We have introduced several individual point mutations into the region coding for motif V of the PPV CI protein and expressed these proteins in Escherichia coli as maltose binding protein fusions. Mutations that abolished RNA helicase activity also disturbed NTP hydrolysis. No mutations affected the RNA binding capacity of the CI protein. These mutations were also introduced in the PPV genome making use of a full-length cDNA clone. Mutant viruses carrying CI proteins with reduced RNA helicase activity replicated very poorly in protoplasts and were unable to infect whole plants without rapid pseudoreversion to wild-type. These results indicate that motif V is involved in the NTP hydrolysis step required for potyvirus RNA helicase activity, and that this activity plays an essential role in virus RNA replication inside the infected cell.
The unwinding of duplex RNA using NTP hydrolysis as an energy source is catalyzed by RNA helicases (for reviews, see 1 -4 ). This reaction is required for several different biological processes such as gene transcription, RNA processing and translation, and often contributes to the regulation of cell growth and development.
In addition to the well characterized RNA helicases, many proteins are considered as putative RNA helicases on the basis of amino acid sequence analysis (2 ). Two sequence signatures termed motifs I and II, that correspond to the `Walker box' purine NTP-binding sequence (5 ), are shared by all helicases and a wide variety of other NTP-utilizing proteins. Five other sequence motifs (Ia, III, IV, V and VI) are conserved in RNA and DNA helicases of the superfamilies SF1 and SF2 (2 ).
Both experimental evidence and sequence data analysis indicate the existence of sequences potentially encoding helicases in a large number of virus genomes (4 ). In particular, most of the plus strand RNA virus genomes encode at least one putative helicase protein. The NTPase activity expected for RNA helicases has been demonstrated for viral proteins belonging to each of the three superfamilies defined by Gorbalenya et al. (2 ), but RNA helicase activity has only been demonstrated for RNA virus proteins belonging to the SF2 superfamily (4 ). RNA helicase activity in an RNA virus protein was first demonstrated in the plum pox potyvirus (PPV) CI protein (6 ). This activity has also been found to be associated with the tamarillo mosaic potyvirus (TaMV) CI protein (7 ) and with the hepatitis C virus (8 ,9 ) and bovine diarrhea pestivirus (10 ) NS3 proteins, which also belong to the SF2 superfamily.
The genus Potyvirus is the largest group of plant viruses. The CI protein forms the cylindrical inclusion bodies typical of potyvirus infections and is synthesized as part of the single polyprotein encoded by the potyviral genome (11 ,12 ). The N-terminal half of the potyviral CI protein contains all of the typical domains of helicases of the superfamily SF2 (13 ). The CI C-terminal half shows no homology with RNA helicases, but small deletions in this region of a maltose binding protein (MBP)-PPV CI fusion synthesized in Escherichia coli abolished its RNA helicase activity (14 ).
Some data on the assignment of biochemical functions to the helicase conserved domains have been obtained in some SF2 superfamily proteins, such as the eukaryotic translation initiation factor 4A (eIF4A) and in the NTP phosphohydrolase II (NPH-II) of vaccinia virus (15 -22 ). Much less information is available on the subgroup SF2 RNA helicases of viral RNA origin. Deletion mutagenesis has allowed us to localize two RNA binding domains in the PPV CI protein, which might be related to the helicase motifs Ia and VI (14 ,23 ). In this paper we report the use of point mutations to characterize the role of the motif V in the RNA helicase activity of the PPV CI protein using a variety of enzymatic assays in vitro. We also describe the effect of mutations in conserved residues of the PPV CI motif V on the viruses ability to replicate in isolated cells and to systemically infect plants.
All recombinant DNA procedures were carried out by standard methods (24 ). Escherichia coli strains JM109 and DH5[alpha] were used for the cloning of the plasmids. pcNCI harbors the PPV CI coding sequence fused to the maltose binding protein gene of pMal-c (New England Biolabs) and was reported previously (14 ). G to A [nt 3696; numbering of nucleotides corresponds to the full-length sequence of PPV RNA (25 )] and CT to TC (nt 5534, 5535) point mutations that create new ClaI and XbaI restriction sites in the CI cistron were produced by site-directed mutagenesis (26 ) of SacI (vector)-SphI (nt 4056) and SmaI (nt 4238)-PstI (vector) fragments of pcNCI subcloned into a M13 vector. The mutated fragments were introduced into pcNCI to create pcNCIcx.
The point mutations in the PPV CI motif V were created by site-directed mutagenesis (26 ) in the SmaI (nt 4238)-PstI (vector) fragment of pcNCIcx subcloned in a M13 vector using the following mutator oligodeoxinucleotides: 5'-GTTGCAATCACGAAATG-3' (for V303I; the number corresponds to the amino acid sequence of PPV CI), 5'-TTGTTGCTGCCACGAAATG-3' (for V303A), 5'-GTATTTGATGCAACC-3' (for T305S), 5'-GTGACTGA/TATTTTCAATG-3' (for G311A and G311S), 5'-GTCCAATGAGACTCC-3' (for T313S) and 5'-GTCCAATGCGACTCC-3' (for T313A). The plasmids pcNCIcxV303I, pcNCIcxT305S, pcNCIcxG311A, pcNCIcxG311S and pcNCIcxT313S were obtained by replacing the Eco88I (nt 4238)-PstI (vector) fragment from pcNCIcx with the corresponding ones from the recombinant M13 replicative forms that contained the different mutations. For the construction of pcNCIcxV303A and pcNCIcxT313A, the exchanged fragment was Eco88I (nt 4238)-ApaLI (nt 4875).
pGPPV carries a full-length cDNA copy of the PPV genomic RNA cloned downstream of a T7 RNA polymerase promoter (27 ). pGPPVx is a derivative of pGPPV that carries the CT to TC substitution (nt 5534 and 5535) that creates a XbaI restriction site (P.S and J.A.G., unpublished results). pGPPVxV303I was obtained by a triple ligation of the following restriction fragments: NcoI (nt 7678)-Eco88I (nt 4238) from pGPPV, XbaI (nt 5535)-NcoI (nt 7678) from pGPPVx and Eco88I (nt 4238)-XbaI (nt 5535) from pcNCIcxV303I. pGPPVxV303A, pGPPVxT305S and pGPPVxT313S were obtained by replacing the Cfr9I (nt 4238)-XbaI (nt 5535) of pGPPVxV303I by the corresponding ones of pcNCIcxV303A, pcNCIcxT305S and pcNCIcxT313S, respectively. In order to reconstruct pGPPVx from pGPPVxT313S, the Cfr9I (nt 4238)-XbaI (nt 5535) fragment from pGPPVxT313S was replaced by the corresponding one of pGPPVx.
The expression of the recombinant plasmids and the partial purification of the corresponding MBP-CI fusion proteins were carried out essentially as previously described (14 ). After growing at 30oC in LB medium containing ampicillin (100 µg/ml) and induction with 50 µM IPTG, transformed JM109 cells were collected by centrifugation, and lysed by grinding with alumina. The crude extract was loaded onto an amylose resin column (New England Biolabs) equilibrated in 10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 1 M NaCl. The non-retained proteins were successively washed with the same buffer containing 1, 0.5 and 0.2 M NaCl, and without NaCl, whereas the products specifically retained were eluted with buffer containing 10 mM maltose and no NaCl.
Unwinding reaction mixtures contained 30 mM Tris-HCl (pH 7.5), 1.5 mM MgCl2, 15 mM DTT, 30 µg/ml bovine serum albumin, 0.12 U/µl RNasin, 2 mM ATP, partially dsRNA substrate A (0.15 µM), prepared according to Laín et al. (6 ), and the indicated amount of the purified protein fractions. After 20 min of incubation at 25oC, the reactions were stopped by adding 3 µl of 0.5% SDS/40 mM EDTA. Samples were loaded in 8% polyacrylamide gels containing 0.1% SDS and 0.5* TBE buffer.
The dATPase assays were carried out at 25oC in 20 µl reaction mixtures containing 15 mM HEPES-KOH (pH 7.5), 2.5 mM Mg (CH3CO2)2, 1 mM DTT, 0.1 mM [[alpha]-32P]dATP (~125 Ci/mol), 0.2 mM poly A and the indicated amount of the partially purified protein fractions. Reactions were stopped in ice after different incubation periods by adding EDTA to a concentration of 50 mM, and samples were analyzed by polyethylenimine cellulose thin-layer chromatography with 0.2 M potassium phosphate (pH 7.5) as the liquid phase. Hydrolysis activity was calculated by quantitating the radioactivity of the spots with a Storage Phosphor Imaging System (model GS-525, Bio-Rad).
The partially purified proteins were subjected to SDS-PAGE and transferred to nitrocellulose using a Trans-Blot apparatus (BioRad). The membrane was incubated at room temperature for 1 h in a renaturation solution containing 10 mM Tris-HCl (pH 7), 1 mM EDTA, 60 mM NaCl, 0.1% Triton X-100, 1* Denhardt's reagent. This step was repeated four times. The membrane was then incubated in the same buffer containing 6 × 104 c.p.m. of RNA labeled with [[alpha]-32P]UTP by in vitro transcription of plasmid pT4ps2 digested with EcoRI and PvuII, followed by removal of unincorporated nucleotides by spun column centrifugation through Sephadex G-50 (6 ). After three washes (15 min each) in the renaturation buffer, the nitrocellulose membrane was air dried and exposed to X-ray film.
pGPPV and its derivatives were linearized with PvuII and PstI, that cut upstream of the T7 promoter and immediately downstream the poly(A) tail at the 3'-end of the PPV sequence, respectively. Capped full-length transcripts were synthesized from these constructs using T7mMESSAGEmMACHINE (Ambion) or T7 Cap-Scribe (Boehringer) kits. The yield and integrity of the transcripts were analyzed by agarose gel electrophoresis.
Three Nicotiana clevelandii primary leaves were dusted with Carborundum and mechanically inoculated with 2 µl (1 mg/ml of transcript) of the transcription reaction diluted 1:1 in 5 mM sodium phosphate buffer (pH 7.2). Virus infection was assessed by visual inspection of symptoms and by western blot analysis of PPV capsid protein accumulated in the inoculated plants. RT-PCR amplified fragments which included the PPV CI motif V were sequenced to determine if the mutations of the viral transcripts used as inocula were maintained in viral progeny.
Protoplasts were isolated from N.clevelandii leaves as described previously (28 ). Samples of 8 × 105 protoplasts resuspended in 0.8 ml of electroporation buffer were exposed together with 20 µl of in vitro transcription reaction mixture to an electroporation pulse (110 V, 1000 µF, 20-25 ms) in an Electro Cell Manipulator 600 apparatus (BTX). After a 15 min recovery period on ice, the protoplasts were washed, resuspended in 1.2 ml culture medium and incubated under diffuse light at 24oC for 30 h. To assess virus multiplication in the protoplasts, PPV RNA was detected by northern blot analysis and PPV virions by RT-PCR preceded by immunocapture PCR (IC-PCR).
For northern blot analysis, protoplasts (2 × 105) were collected by centrifugation, resuspended in 200 µl of extraction buffer [100 mM Tris-HCl (pH 9), 300 mM NaCl, 20 mM EDTA, 2% SDS, 0.25 mg/ml proteinase K and 0.5% heparin] and incubated at 37oC for 30 min. After phenol extraction, RNA was precipitated in the presence of 2 M LiCl, washed by ethanol precipitation, and resuspended in water. RNA was electrophoresed in 1.2% agarose-formaldehide gels, transferred to Zeta-probe membranes (BioRad) and hybridized with a 32P-labeled riboprobe specific of the PPV NIb cistron, synthesized by in vitro transcription using a MAXIscript kit (Ambion).
For IC-PCR assays, the centrifugation pellet of protoplast cultures (2 × 105) was resuspended in 50 µl of immunocapture buffer (PBS, 0.05% Tween-20, 2% polyvinylpyrrolidone K25). This extract was incubated for 2 h in the presence of RNase (2.5 µg) and DNase (0.25 U) in tubes previously coated with rabbit anti-PPV IgGs. The immunoretained material was used as template for RT-PCR amplification of a DNA fragment corresponding to PPV nt 8839-8900. The reaction products were characterized by Southern blot analysis. They were electrophoresed in 1.2% agarose gels, transferred to Zeta-probe membranes (BioRad) and hybridized with a probe consisting of a PCR-synthesized PPV cDNA fragment (nt 8389-8900) 32P-labeled by the random primer technique using a DECAprime kit (Ambion).
The level of conservation of motif V is very high among all the sequenced members of the genus Potyvirus (Fig. 1 ). Indeed, there is a stretch of 11 amino acids (from Val 303 to Thr 313) that is strictly conserved in all cases (Fig. 1 B). Figure 1 C shows the comparison of the potyvirus motif V consensus sequence with the motif V of different members of the DExH and DEAD groups of the SF2 superfamily of helicases. The extent of amino acid similarity is striking not only within the RNA virus subgroup, but also between helicase-like proteins of this subgroup and other members of the DExH group and of the more distantly related DEAD group (Fig. 1 C). The most frequently conserved residue is Thr 305 which is invariant in all of the proteins compared. A hydrophobic residue is always present in positions 302 and 319, while all the DExH proteins have an Asp in position 321, and all the RNA virus proteins have a Gly in position 311 (Fig. 1 C).
In order to elucidate the role that motif V plays in the activity of the PPV CI protein, several individual residues in the segment conserved in all the potyviruses were modified by site directed mutagenesis. Both Thr 305 and Gly 311 which are highly conserved in SF2 RNA helicases, and Val 303 and Thr 313 which seem to accept some sequence variations, were replaced either by a closely related amino acid or by Ala, a neutral residue that usually does not disturb the polypeptide chain.
The mutations were introduced into pcNCIcx, which contains the coding sequence of a modified CI fused to the MBP. This CI sequence contained two mutations, which were introduced to facilitate the cloning procedures. One of the mutations gives rise to a conservative Val to Ile change in position 16 of the protein, and the other is a silent mutation. The enzymatic activities of MBP-CI and MBP-CIcx were indistinguishable (data not shown).
MBP-CIcxV303I and MBP-CIcxT313S showed detectable RNA helicase activity (Fig. 2 ). Whereas the activity of MBP-CIcxV303I was similar to that of MBP-CIcx (lanes 3 and 4), that of MBP-CIcxT313S was clearly lower (lane 5). Less conservative substitutions in the same positions (V303A and T313A) lead to products without detectable helicase activity (lanes 7 and 9). On the other hand, mutations in the most conserved residues, even though they produced very conservative changes (T305S, G311A), caused the complete loss of RNA helicase activity (lanes 6 and 8). The same result was observed with a G311S mutation (results not shown).
In order to ascertain whether the motif V mutations affected the ability of PPV CI to interact with RNA, the MBP-CI mutant proteins were transferred to nitrocellulose and incubated with a 32P-labeled RNA probe (northwestern assay). Although this assay does not allow precise quantitative estimations of the RNA binding, all the motif V mutant CI proteins showed efficient interaction with RNA (Fig. 3 ).
All eight common nucleoside triphosphates support CI helicase activity (6 ). We assayed NTPase activity by measuring dATP hydrolysis because background dATPase activity is known to be very low in maltose-eluted fractions from cells harboring control plasmids (14 ). As expected from its RNA helicase activity, MBP-CIcxV303I had a dATPase activity in the presence of poly A similar to that of the MBP-CIcx control (Fig. 4 ). Likewise, MBP-CIcxT313S, which showed a reduced helicase activity, had a lower dATPase activity than those of MBP-CIcx and MBP-CIcxV303I (Fig. 4 ). Even though MBP-CIcxV303A and MBP-CIcxT305S did not show RNA helicase activity (Fig. 2 , lanes 5 and 6), they did demonstrate a weak dATPase activity at high protein concentrations (Fig. 4 ). No detectable dATPase activity was found associated with the MBP-CIcxG311A and MBP-CIcxT313A proteins (Fig. 4 ). No significant differences in the rate of stimulation by poly A were observed between the active mutant proteins and the wild-type CI (results not shown).
In order to investigate the potential effects of the enzymatic dysfunctions caused by the CI motif V modifications on the viral replication cycle, the mutations which did not completely abolish the enzymatic activities analyzed in vitro (V303I, V303A, T305S and T313S) were introduced into the full-length PPV clone pGPPVx. Protoplasts and intact plants were inoculated with RNA transcripts synthesized in vitro from the resulting plasmids.
The percentage of N.clevelandii plants infected by inoculation with transcripts from pGPPVxV303I (tGPPVxV303I) was similar to that of wild-type tGPPV (20 plants infected out of 24 inoculated versus 19 out of 28, respectively, in three different experiments). Sequencing of a PPV cDNA RT-PCR fragment amplified from progeny virus of plants infected with tGPPVxV303I demonstrated that the V303I mutation was conserved during virus multiplication. No other mutations were detected in the sequenced region. These results corroborated those obtained in vitro and suggest that the V303I mutation in the CI protein is well tolerated.
In order to determine whether modifications in the CI motif V interfered with virus replication in isolated cells, protoplasts prepared from N.clevelandii leaves were inoculated with transcripts synthesized in vitro from PPV full-length cDNA clones harboring the different mutations.
Northern blot analysis of two samples of protoplasts inoculated with tGPPVxV303I showed levels of virus RNA similar to those of protoplasts inoculated with wild-type tGPPV(Fig. 5 A, lanes 1-3). Viral RNA accumulation could not be clearly detected by northern blot analysis in protoplasts inoculated with tGPPVxT305S (Fig. 5 A, lane 5), tGPPVxT313S (Fig. 5 A lanes 6 and 7) or tGPPVxV303A (Fig. 5 A, lanes 8 and 9), although in the last case very faint hybridizing bands were observed. This result indicates that CI motif V mutations hampered virus replication in isolated cells.
The high level of conservation of the motif V observed among the potyvirus CI proteins (Fig. 1 ) suggests that it might play an important role in the protein function. The results reported in this paper indicate that small modifications in this region, which are not expected to disturb the overall conformation of the protein, have profound effects on the enzymatic activities of the PPV CI protein (Fig. 6 ). In the most invariant positions (305 and 311), even very conservative substitutions (Thr to Ser or Gly to Ala) abolished the RNA helicase activity (Fig. 2 ). The functional requirements at positions 303 and 313 seem to be less strict since MBP-CI fusion products harboring the V303I or T313S mutations still displayed RNA helicase activity. Actually, MBP-CIcxT313S showed low helicase activity relative to that of wild-type CI. More drastic substitutions (V303A and T313A) also reduced RNA helicase activity to undetectable levels in vitro.
This work was supported by grants from CICYT, Comunidad de Madrid and European Union (Biotechnology Programme). A.F., H.S.G., P.S., L.S.-B. and M.G.C were recipients of fellowships from Basque Regional Government (A.F. and M.G.C.), Comunidad de Madrid (H.S.G. and P.S.) and FPI (L.S.-B). The help of Victor Buckwold in improving the English of the manuscript is gratefully acknowledged.
REFERENCES