Specific sequence changes in the 5h-terminal ... - Semantic Scholar

Journal of General Virology (1998), 79, 905–913. Printed in Great Britain ...................................................................................................................................................................................................................................................................................

Specific sequence changes in the 5«-terminal region of the genome of satellite tobacco mosaic virus are required for adaptation to tobacco mosaic virus Martin Ngon A Yassi and J. Allan Dodds Department of Plant Pathology, University of California, Riverside, CA 92521-0122, USA

The genome of satellite tobacco mosaic virus (STMV) adapted to tobacco mosaic virus (TMV), tomato mosaic virus or green tomato atypical mosaic virus consistently had two single base deletions at positions 1 and 61, corresponding to bases A and G, respectively, as compared to the type-strain genome which is naturally adapted to tobacco mild green mosaic virus (TMGMV). Transcript RNAs (STMVTMV) from clone pSTMVTMV which captured the deletions at positions 1 and 61 were infectious when co-inoculated to tobacco plants with either TMV or TMGMV at infection frequencies of " 90 %. Two new STMV variants were created to investigate whether both deletions were essential for adaptation to TMV. These were STMVTMGMV ∆A1, which had the A at position 1 (A1) deleted, and STMVTMGMV ∆G61,

Introduction Satellite viruses or satellite RNAs are subviral entities that depend on a helper virus for their replication. The specificity of satellite replication varies between different satellite RNAs or satellite viruses and their helper viruses. It has been shown that some satellite RNAs or satellite viruses replicate only with the helper virus with which they were originally associated, whereas some others replicate with different strains of the helper virus, and even with different related viruses from the same virus group (reviewed by Roossinck et al., 1992). Satellite tobacco mosaic virus (STMV) is a small 17 nm spherical virus and its replication depends on co-infection with helper rod-shaped tobamoviruses (Dodds, 1991). The STMV particle contains one positive-sense single-stranded RNA of 1059 bases (Mirkov et al., 1989). STMV is found in association with tobacco mild green mosaic virus (TMGMV, also known as TMV-U5) in Nicotiana glauca, a wild perennial tobacco relative found commonly in Southern California (Valverde & Author for correspondence : J. Allan Dodds. Fax : ­1 909 787 4294. e-mail : dodds!ucr.edu

0001-5200 # 1998 SGM

which lacked G61. STMVTMGMV ∆A1 was infectious (75 % frequency) in the presence of either TMV or TMGMV. Virion RNA of STMVTMGMV ∆A1 lost G61 after one infection cycle with TMV. This deletion did not occur in co-infections with TMGMV. STMVTMGMV ∆G61, like the clone STMVTMGMV, was infectious (100 % frequency) with TMGMV but TMV did not support this clone. When Nicotiana benthamiana protoplasts were transfected with STMVTMGMV, STMVTMGMV ∆A1 or STMVTMV, STMV replicated when TMGMV was the helper virus. STMVTMV and STMVTMGMV ∆A1 replicated in the presence of helper TMV, but STMVTMGMV did not, the same result as in whole plants. The deletion of A1 is thus essential for initial STMV adaptation to TMV and the eventual deletion of G61 is a predicted additional change.

Dodds, 1987). Viral RNA isolated from purified type STMV can be supported by at least eight different tobamoviruses in a variety of hosts in greenhouse experiments (Valverde et al., 1991). However, RNA transcripts from independent fulllength STMV cDNA clones obtained from type STMV RNA were not consistently infectious with any tobamovirus other than the natural helper TMGMV (Mirkov et al., 1990 ; Kurath et al., 1993 b), although occasional adaptation to other tobamoviruses was observed. Kurath et al. (1993 b) observed a deletion of one of five Gs between positions 61 and 65 when STMV was grown in the presence of tobacco mosaic virus (TMV). RNA transcripts synthesized from full-length cDNA clones of STMV RNA populations adapted to TMV and containing this deletion were supported by TMGMV but were non-infectious with TMV (Kurath et al., 1993 b). This implies that ∆G is necessary but not solely sufficient for adaptation '" to TMV. Furthermore, RNA transcripts from an STMV chimera containing the 3« 150 bases of TMV (Kurath et al., 1993 b) were not supported by TMV yet were supported by TMGMV. This implies that the ability of STMV clones to replicate with different helper tobamoviruses other than TMGMV is not specified by sequences in the 3« end of the STMV genome

JAF

M. Ngon A Yassi and J. A. Dodds

despite the fact that sequences at the 3« end are important for promotion of negative-sense RNA synthesis (Ishikawa et al., 1988, 1991). In their studies, Mirkov et al. (1989, 1990) and Kurath et al. (1993 b) did not analyse the genetic changes that may occur in the 5« first 12 bases of the STMV RNA. The 5« primers used to generate previously reported STMV infectious clones were based on the sequence of the TMGMV-adapted type-strain and their use may have hidden or masked specific changes that accompanied helper adaptation of STMV to other tobamoviruses. In this report, we describe the changes that occur in the genome of STMV after adaptation to TMV, tomato mosaic virus (ToMV) and green tomato atypical mosaic virus (G-TAMV), and the construction of full-length cDNA clones of the STMV genome that are adapted to, and are infectious in the presence of, TMV. These clones were used to more clearly define the basis of STMV helper TMV specificity, which mapped to the 5« untranslated region (5« UTR).

Methods + Host range, virus isolates and RNA preparation. The host plants used in this study included Nicotiana tabacum cv. Xanthi (nn), N. benthamiana and N. sylvestris. These plants are hosts for TMV and STMV, or TMGMV and STMV (Valverde et al., 1991). TMV was a subculture from cDNA clone pTMV204 (Dawson et al., 1986) that was maintained in N. sylvestris plants. TMGMV was a California isolate described in previous studies as TMV-U5 (Valverde & Dodds, 1987). Type STMV was maintained and propagated in N. tabacum cv. Xanthi (nn) with TMGMV as previously described (Valverde & Dodds, 1987). The STMV infectious clone used in this study has been previously characterized (Kurath et al., 1992 ; Routh et al., 1997). This clone will be referred to as pSTMVTMGMV. ToMV and G-TAMV were previously described by Valverde et al. (1991). Viruses and viral RNA were purified as previously described (Valverde & Dodds, 1987 ; Kurath et al., 1992). + Adaptation of the STMV genome to ToMV, G-TAMV and TMV. RNA from STMV was gel-purified as previously described (Kurath

et al., 1993 b). This helper-free STMV RNA, which was non-infectious alone, was co-inoculated with TMV to N. sylvestris plants and maintained in a growth chamber for 2 weeks. Leaves that were systemically infected with TMV and STMV were harvested, and virions and virion RNA were purified as indicated above. STMV RNAs adapted to ToMV and GTAMV were purified from frozen tissues following adaptation as described by Kurath et al. (1993 b). + cDNA synthesis and sequence analysis. The 5« RACE system (Gibco BRL) for rapid amplification of cDNA ends was used for cDNA synthesis. To prime the synthesis of the first-strand cDNA from STMV genomic RNA, primer STMV3«HIII, complementary to STMV 3« sequence from nucleotides 1047 to 1059 (Table 1) and containing a HindIII restriction site, was used. Purification and dA- or dC-tailing of the first-strand cDNAs, as well as synthesis of double-stranded cDNA by PCR amplification were performed according to the manufacturer’s instructions. The primers used in different PCRs are described in Table 1. After PCR amplification, the dC-tailed products were restricted with SpeI and HindIII, and inserted into similarly treated plasmid pBluescript KS(­}®) (Stratagene). Clones were screened for insert size and those with the expected size (E 1000 bp) were sequenced using the dideoxynucleotide method (Sanger et al., 1977) with Sequenase version 2.0 (USB) and deoxyadenosine 5«-[α-$&S]thiotriphosphate (Amersham) as the labelled nucleotide. Gel-purified PCR products were sequenced using Taq DNA polymerase and the fmol DNA Sequencing System (Promega) following the manufacturer’s instructions. Primer STMV120R (Table 1), which is complementary to the STMV genome from position 98 to 117, and the T7 RNA polymerase promoter primer were used to determine the sequences in the 5« UTR of helper-adapted STMV genomes from both orientations. The program FOLD of the University of Wisconsin Genetics Computer Group (GCG, Devereux et al., 1984) using the algorithm of Zuker & Stiegler (1981) and the Star 3.0 program, which utilizes the algorithm of Abrahams et al. (1990), were used to analyse putative secondary structures of the 5« UTR. + Construction of TMV-adapted STMV infectious clones and site-directed mutagenesis of STMV infectious clones. The sequence obtained at the 5« end of the TMV-adapted STMV genome (Fig. 2) was used to design a sense-strand primer T7}STMV5«EI (Table 1)

Table 1. DNA sequences of the synthetic oligonucleotide primers used for cDNA synthesis Primers 5« RACE Anchor and 5« RACE UAP, and primers NotI (dT) and NotI adapter were used to "( synthesize double-stranded cDNA from either dC-tailed or dA-tailed first-strand STMV cDNA, respectively, in the touch-down PCR amplification. The sequences in bold represent STMV sequences, in the plus-sense orientation (primers T7}STMV5«EI and STMV5«end) or in the minus-sense orientation (primers STMV3«HIII and STMV120R). Primer

Length

STMV3«HIII T7}STMV5«EI STMV5«end STMV120R 5« RACE Anchor 5« RACE UAP NotI (dT) "( NotI adapter

32-mer 46-mer 19-mer 20-mer 36-mer 20-mer 40-mer 24-mer

JAG

Sequence (5« to 3«) agaattcaagcttgggccgcttacccgcggtt cactggaattctaatacgactcactatagtaaaacttaccaatcaa gatctagagtaaaacttac atggcgactgaaggctctt ggccacgcgtcgactagtacgggiigggiigggiig ggccacgcgtcgactagtac actgaattcaagcttgcggccgcttttttttttttttttt actgaattcaagcttgcggccgct

STMV adaptation to TMV containing an EcoRI restriction site, the T7 RNA polymerase promoter sequence and the first 18 nucleotides corresponding to the 5« end sequence of the TMV-adapted STMV genome. Double-stranded cDNA was generated by PCR using primers STMV3«HIII and T7}STMV5«EI, and a template consisting of first-strand STMV cDNA synthesized from TMV-adapted STMV RNA. The double-stranded cDNA was digested with EcoRI and HindIII, and inserted into a similarly treated plasmid pUC19. Two variant clones of type STMV were created. The first STMV variant (STMVTMGMV∆A ) lacked the first base of the genome and was " constructed using primers T7}STMV5«EI and STMV3«HIII from template DNA pSTMVTMGMV by PCR amplification. The double-stranded cDNAs obtained from this PCR were prepared for ligation with pUC19. The second STMV variant (STMVTMGMV∆G ) contained a single '" deletion of G and the 5« sequence was similar to that of STMVTMGMV. '" Primer STMV5«end and primer STMV3«HIII (Table 1) were used to generate double-stranded cDNA using pSTMVTMV (defined in Results section) as template DNA in a PCR amplification. cDNAs were digested with suitable enzymes and inserted into similarly treated pBluescribe KS(­}®). The 5« termini of the STMV cDNA clones with expected insert size were sequenced prior to in vitro transcription to provide RNA inoculum. + Infectivity experiments and transfection of tobacco protoplasts. RNA was synthesized from HindIII-linearized plasmid DNAs as described above, using either the T7 RNA polymerase (Promega) or the Maxiscript T7 Kit (Ambion) as specified by the manufacturers. Products were diluted in inoculation buffer (20 mM potassium phosphate, pH 7±2) and used to inoculate tobacco protoplasts or whole plants. Infectivity studies were performed in N. tabacum cv. Xanthi (nn) plants as previously described (Kurath et al., 1992) using TMGMV and TMV as helper viruses. Plants were maintained in a greenhouse for 10–20 days during which time the symptoms of helper virus infection became visible. Plants were tested for the presence of STMV by immuno-diffusion assays in agarose gel (Valverde & Dodds, 1987 ; Valverde et al., 1991) and by double antibody sandwich ELISA using polyclonal antibodies raised against purified satellite virions (Valverde & Dodds, 1987). Protoplasts were prepared from leaves of N. benthamiana as described by Routh et al. (1997) and inoculated immediately with purified TMGMV or TMV RNA and RNA transcripts from an STMV infectious clone by the polyethylene glycol-mediated procedure (Rao et al., 1994). Two µg of each helper virus RNA were co-inoculated with 2 µg of STMV transcript RNA into 2±5¬10& isolated protoplasts. Samples were incubated at 25 °C for 24 h under diffuse fluorescent light. Total nucleic acids were purified from transfected protoplasts as described by Rao et al. (1994), and products were subjected to Northern blot analysis using STMV riboprobes labelled with [α-$#P]UTP (Routh et al., 1997) + Analysis of STMV progeny. Serial passages of infected plant sap were performed as described by Kurath & Dodds (1995). Satellite virions were purified after each passage, viral RNA was extracted and doublestranded cDNA was synthesized using the 5« RACE system (Gibco BRL) as described above. cDNA products were cloned into pBluescript KS(­}®) and their 5« termini were sequenced. In some cases, the first 100 nucleotides of the PCR products that had been dC-tailed were sequenced to confirm the sequence of the 5« UTR base of the progeny genome, including the first base.

Results In an attempt to avoid confusion with terminology when describing variants of STMV, the order of the helper virus and satellite virus has been deliberately reversed for isolates (helper

virus first) and for clones (helper virus second) of STMV. For example TMV}STMV is an isolate of STMV that has adapted and replicates with the named helper tobamovirus, in this case TMV. pSTMVTMV is a clone, or transcript (STMVTMV) from that clone, customized to replicate with the named helper tobamovirus. Sequence analysis of STMV cDNAs adapted to TMV, ToMV and G-TAMV

Previous analysis of STMV helper virus dependency showed that the nucleotides near the 5« terminus of STMV RNA were involved in TMV specificity for STMV (Kurath et al., 1993 b). A new preparation of STMV was biologically adapted to TMV (TMV}STMV), ToMV (ToMV}STMV) or G-TAMV (G-TAMV}STMV) by co-infection of tobacco plants with TMV and helper-virus-free type STMV RNA that had been gel-purified and was non-infectious when inoculated alone. STMV was purified from systemically infected leaves and purified RNA was used for cDNA synthesis and sequence analysis. PCR products from dC- or dA-tailed STMV cDNAs that had been adapted to new helper viruses were sequenced to obtain data on the possible changes that may have occurred in the 5«-terminal region of the STMV helper virus adapted genome. Two single nucleotide deletions were detected in the 5« UTR of the STMV helper-adapted genome. The first one, previously described by Kurath et al. (1993 b) and repeated here, was a deletion of one of the Gs between position 61 and 65 of the STMV genome (∆G ) and the second one was the '" deletion of the first base (∆A ) with reference to the type " sequence (Mirkov et al., 1989). The absence of A in numerous " sequencing gels was very clear (Fig. 1). Additional sequence variability was detected within the first 20 bases of the helper virus-adapted STMV genomes, resulting in a sequence shifting on gels (Fig. 1). From observation of the sequences of the cDNAs of genomes of either TMV}STMV, ToMV}STMV or G-TAMV}STMV, an insertion of one base was suspected at position 18 with reference to the type sequence (Mirkov et al., 1989). The genome population was presumed to be made up of two variants, one with four Us between positions 18 and 21 as in the type sequence and the other with five Us between positions 18 and 22 (Fig. 2). To obtain exact sequence information and assess the variability around this 5« region, independent cDNA clones were subjected to sequence analysis. Eight independent clones were sequenced from TMV}STMV, two from ToMV}STMV and three from G-TAMV}STMV. Three out of eight clones from TMV}STMV contained five Ts at positions 17–20 and the others had four Ts in that position. Both clones sequenced from ToMV}STMV had five Ts at these positions. Two clones from G-TAMV}STMV had four Ts between positions 17 and 20 and one had five Ts. Similar variability was also detected in the parental TMGMV}STMV genome population. When primer STMV5«end (Table 1), which contains only the first 12

JAH

M. Ngon A Yassi and J. A. Dodds (a)

(b)

(c)

(d)

Fig. 1. Autoradiographs representing sequence analysis of the 5«-terminal region of PCR products (a, b) or cloned cDNA (c, d) from helper-adapted STMV RNA. The cDNAs were sequenced either from the interior of the molecule using primer complementary to positive-sense STMV RNA from positions 98 to 117, or from the exterior of the molecule using the T7 RNA polymerase promoter primer when the cDNA was cloned into the plasmid pBluescript KS(­/®). The autoradiographs in (a) and (b) have been inverted for convenience. Interpretations of the terminal bases of the sense strand ssRNA are shown. (­1) corresponds to the first base of the TMV-adapted STMV which lacks a terminal A (arrow) compared to the sequence of the type genome (Mirkov et al., 1989). T-tail or G-tail correspond to poly(T) (a and b) or poly(G) (c and d) sequences respectively, resulting from dA or dC tailing of the first-strand cDNA. (a) Sequence of the 5« first 27 (28) nucleotides of a PCR product of a TMV-adapted STMV cDNA. The written sequences on the right represent the nucleotide sequence of STMV. As a result of a mixture of molecules in this population, sequence shifting from base (­21, ­20) to the 5« base (G­1) has occurred. The sequence in upper case represents the molecule containing five As between positions 17 and 21 and the sequence in lower case represents the molecules containing four As in this region. An asterisk indicates a possibility of a G or a T. (b) Nucleotide sequence of a dA-tailed first-strand molecule of a ToMV-adapted STMV. This population contains five As between positions 17 and 21. (c) Nucleotide sequence of a cloned cDNA from a TMV-adapted STMV that contains four As between positions 17 and 20. (d) Nucleotide sequence of a cloned cDNA from a TMV-adapted STMV that contains five As between positions 18 and 22.

Fig. 2. Summary of the nucleotide sequence of the 5« termini of eight TMV helper-adapted STMV sequences, two ToMVadapted STMV sequences and three G-TAMV-adapted STMV sequences, in comparison with previously reported sequences of STMV. (Gn) is the number of G residues that were added to the 3« end of the first-strand cDNA during synthesis. Asterisks represent nucleotides that are present in the type sequence as reported by Mirkov et al. (1989), but are deleted in the helperadapted molecules. ∆ represents the deletions in the genome at positions 1 and 61. ­ represents the region of variability where an insertion of an A can occur in the population. In contrast to Fig. 1, ­1 corresponds to the first base of the type sequence. N5 and N39 represent regions of the genome in the 5« UTR where the nucleotide sequence is identical to the typestrain sequence and include 5 and 39 bases, respectively.

bases of the STMV genome, and primer STMV3«HIII were used to obtain double-stranded cDNA from TMGMV-adapted STMV genomes, insertion of one T at position 18, previously

JAI

undetected, was observed in some individual cDNA clones. The variability observed between positions 18 and 21 or 22 of the STMV genome occurred whether it was replicating in the

STMV adaptation to TMV

Fig. 3. Schematic representation of the STMV variants used in this study. The genome of STMV is represented by straight lines and the regions in the 5« termini of the clones where there are changes in the nucleotide sequence are identified. Transcripts STMVTMGMV and STMVTMGMV∆G61 have 35 non-viral nucleotides upstream of the first base of the satellite genome. Transcripts STMVTMV and STMVTMGMV∆A1 sequences start with the second base (G2) of the type-strain sequence as described by Mirkov et al. (1989). 1, 61 and 65 represent positions in the genome.

presence of its natural helper TMGMV or with the other helper viruses. The most consistent changes in all the cDNA clones analysed in this study following adaptation to TMV, ToMV and G-TAMV were ∆A and ∆G with reference to the " '" published type sequence (Mirkov et al., 1989). Aided by this new information, we decided to study in more detail the molecular basis of STMV helper adaptation using TMGMV, TMV and customized STMV clones. Construction of STMV infectious clones adapted to TMV and site-directed mutagenesis

The sequence determined at the 5« end of the TMVadapted STMV genome was used to design a sense-strand primer (T7}STMV5«EI) that matched nucleotides 1–18 of the new sequence (Fig. 2) and included the T7 RNA polymerase promoter. Two of the resulting full-length clones, pSTMVTMV- with five Ts between positions 18 and 22, and ( pSTMVTMV- with four Ts at that position were used for * infectivity experiments. Four plants (N. sylvestris) were each inoculated with transcript RNA from each of the two individual clones and with TMV and all tested positive for STMV in systemically infected leaves. There was no evidence for STMV in plants inoculated with TMV alone, or with transcript STMV RNA alone. Clone pSTMVTMV- (Fig. 3) was selected for * further investigation because it contained sequence similar to that of pSTMVTMGMV (Kurath et al., 1992) at positions 18–21 (four Ts instead of five). This clone will be referred to from now on as pSTMVTMV. Sequence analysis of the entire STMV genome in pSTMVTMV revealed two extra mutations, substitution of a C to a U and of a C to an A at positions 94 and 753, respectively, with reference to the type sequence (Mirkov et al., 1989). The first mutation is located in the 5« UTR of the TMV}STMV genome and the second one is in the 3« UTR. Results from the previous section indicated that the first nucleotide of the type genome sequence (Mirkov et al., 1989) was absent from the TMV}STMV genome. Site-directed mutagenesis was carried out to delete this nucleotide from the type sequence. PCR amplification was performed using primers T7}STMV5«EI and STMV3«HIII, and pSTMVTMGMV (Kurath

Table 2. Summary of infectivity assays comparing four STMV RNA transcripts, STMV virion RNA and two helper viruses Values are the number of plants infected with STMV out of the total number of tobacco plants inoculated, with percentages in parentheses. STMV RNA inoculum

TMV helper

TMGMV helper

STMVTMGMV STMVTMGMV∆G '" STMVTMGMV∆A " STMVTMV Type STMV virion RNA

0}10 (0 %) 0}10 (0 %) 15}20 (75 %) 18}20 (90 %) 10}10 (100 %)

10}10 (100 %) 10}10 (100 %) 20}20 (100 %) 20}20 (100 %) 10}10 (100 %)

et al., 1992) as template DNA. Ligation of this PCR amplification product into pUC19 resulted in cDNA clone pSTMVTMGMV∆A (Fig. 3). This clone is different from the " type clone because A is deleted from the 5« terminus and is " different from pSTMVTMV because it carries five G bases at positions 61–65 instead of four. The 5« terminus of pSTMVTMGMV∆A was sequenced and the deletion of the first " nucleotide was confirmed. In order to investigate the independent significance of the deletion of G to helper adaptation, a second mutagenesis was '" carried out to create clone pSTMVTMGMV∆G , which contains '" 35 nucleotides upstream of the STMV genome and includes the first base A of the type sequence but has a deletion of G . '" Infectivity experiments

Transcript RNAs from pSTMVTMGMV, pSTMVTMGMV , pSTMVTMGMV∆A and pSTMVTMV (STMVTMGMV, '" " STMVTMGMV∆G , STMVTMGMV∆A and STMVTMV, re'" " spectively ; Fig. 3) were inoculated to N. tabacum cv. Xanthi (nn) with TMGMV and to N. sylvestris with TMV. Systemic tissue samples of inoculated plants were tested for the presence of STMV 10 days post-inoculation by immuno-diffusion and ELISA. The results of co-infections of these STMV transcript RNAs and STMV virion RNA with TMGMV or TMV are summarized in Table 2. Of a total of 10 plants (N. tabacum) ∆G

JAJ

M. Ngon A Yassi and J. A. Dodds

Of a total of 20 plants of N. tabacum inoculated with STMVTMGMV∆A , 20 became infected with STMV when " TMGMV was used as helper virus. Fifteen out of twenty plants of N. tabacum (75 %) became infected with STMV in a co-inoculation of TMV and STMVTMGMV∆A . STMVTMGMV " ∆A is thus also well adapted to replicate in mixed infections " with TMV. A serial passage was conducted by inoculating plant sap from a pool of TMV and STMVTMGMV∆A " systemically infected leaves to 15 N. tabacum plants. All 15 inoculated plants (100 %) became infected with STMV. Of a total of 10 plants of N. tabacum inoculated with STMVTMGMV∆G , 10 became infected with STMV when '" TMGMV was used as the helper virus. None was infected with STMV in a co-inoculation of STMVTMGMV∆G and TMV. '" Fig. 4. Replicative competence of STMV variants in tobacco protoplasts in the presence of TMV or TMGMV. Protoplasts were transfected with a combination of STMV RNA transcripts and helper virus RNAs and incubated for 24 h. Total nucleic acids were extracted and submitted to Northern blot analysis using 32P-labelled minus-sense STMV RNA representing the STMV genome from bases 104 to 791. RNAs were electrophoresed in 1±4 % agarose gels under denaturing conditions with formaldehyde, transferred to nylon membranes and submitted to hybridization with the radioactive probe. The helper viruses were used as purified virion RNA and the STMV variants were transcript RNA from fulllength cDNA clones.

inoculated with TMGMV and STMVTMGMV (type clone), 10 became infected with STMV and TMGMV. Another set of 10 N. tabacum plants were co-inoculated with TMV and STMVTMGMV. Ten plants became systemically infected with TMV, but none of the plants became infected with STMV, confirming earlier results (Kurath et al., 1993 b ; Table 2). Of a total of 20 plants inoculated with STMVTMV, 20 became infected with STMV when TMGMV was used as the helper virus. When another set of 20 plants were co-inoculated with TMV and STMVTMV, 18 became infected with STMV. With this high frequency of infection, it was determined that STMVTMV is adapted to replicate efficiently in mixed infections with TMV.

Analysis of STMV RNA in inoculated protoplasts

A total of three separate experiments were performed in which the amount of transcript RNA used as inoculum, the number of protoplasts and the method of inoculation remained the same. Total nucleic acids were extracted from samples of protoplasts that were incubated for 24 h and STMV RNA was analysed by Northern blot. There was no evidence for STMV replication in protoplasts inoculated with water, STMV transcript RNA alone or helper virus RNA alone (Fig. 4). TMGMV supported the replication of STMVTMGMV, STMVTMV and STMVTMGMV∆A (Fig. 4). When TMV RNA " was used as helper, STMVTMV and STMVTMGMV∆A " replicated (Fig. 4). STMVTMGMV did not replicate in the presence of TMV (Fig. 4). These results confirm those obtained from whole-plant infectivity experiments (Table 2). Persistence of TMV-adapted STMV and progeny sequence in tobacco

Under our experimental conditions, helper-adapted STMVs and type STMV could be detected serologically in systemically infected leaves as early as 5 days post-inoculation in coinoculations with either TMV or TMGMV. The yield of virus purified from systemically infected leaves of these tobacco

Table 3. Analysis for the presence or absence of A1 and G61 in TMV-adapted STMV RNA progeny recovered from infected tobacco plants with TMV or TMGMV as tobamovirus helper Values in parentheses indicate the number of passages. ∆A1

STMVTMV STMVTMGMV∆A

JBA

"

∆G61

TMV

TMGMV

TMV

TMGMV

Deletion conserved (2) Deletion conserved (2)

Deletion conserved (2) Deletion conserved (2)

Deletion conserved (1) Deletion detected (1)

Deletion conserved (1) No deletion detected (3)

STMV adaptation to TMV

plants was similar to that obtained from inoculations with type STMV RNA and TMGMV. To identify what kind of changes may have occurred during co-cultivation with TMV, STMV virions were extracted from systemically infected leaves at the first and second passages and progeny RNA was purified. Double-stranded cDNA was synthesized using the 5« RACE system, and the 5« 100 bases of the satellite genome were sequenced from independent cDNA clones. Sequencing of several first and second passage cDNA clones from infections with TMV­STMVTMV revealed that the deletion of G at position 61 was maintained (Table 3). When STMVTMGMV∆A " was used in mixed infections with TMV, sequence analysis of STMV progeny revealed that a deletion of G occurred in '" 100 % of the clones after the first passage and the deletion of A was maintained. When a combination of TMGMV and " STMVTMV was used in mixed infections, the deletions of A " and G were maintained in all the clones after two passages. '"

Discussion Previously, Kurath et al. (1993 b) reported that STMV helper TMV specificity involves a domain near the 5« end of the satellite virus genome. A deletion of G in the STMV '" genome occurred reproducibly during STMV adaptation when infectious transcripts derived from a full-length cDNA clone of the type sequence (five Gs from position 61 to 65) were used with TMV in tobacco plants. The fact that transcripts from fulllength cDNA clones derived from the adapted sequence were not infectious in the presence of TMV and other tobamoviruses used by Kurath et al. (1993 b) raised questions about additional genetic changes within the 5«-terminal 12 nucleotides. Clones of STMV infectious with TMV have now been created by making specific changes in the terminal region of the genome. The changes to be made were discovered by analysis of this region of the genome in populations adapted to TMV. One host used, N. sylvestris, was able to separate TMV from TMGMV that could derive from the satellite virus preparation. The satellite viral RNA preparations were not infectious alone, implying that no TMGMV was present. The STMV adaptation to TMV involves a deletion of A and of G as compared to " '" the type sequence (Mirkov et al., 1989). A deletion of the A residue at position 1 from the type sequence, which still retained the fifth G at positions 61–65, was sufficient to permit replication of STMV in the presence of TMV in either N. sylvestris or N. tabacum. However, the progeny of this infection dispensed with the fifth G at positions 61–65. Since these genetic changes occurred in both N. sylvestris and N. tabacum with clone STMVTMGMV∆A , the mechanism of STMV " adaptation to TMV does not depend on the tobacco host used, and the selective host (N. sylvestris) only plays a role in screening for helper virus TMV. This implies that the sequence variation seen is an adaptation to the helper virus used in experiments. Results from the STMV}helper virus replication study in tobacco protoplasts paralleled those obtained from

the studies conducted in whole plants and lead to the conclusion that STMV helper adaptation is directly involved with virus replication and does not involve virus movement or accumulation in the plant. Helper adaptation of STMV seems to depend on tobamovirus replicase specificity acting on the 5« UTR of the satellite molecule selected. STMVTMGMV∆G was constructed late in the course of '" this study and was not used in protoplast experiments, but there is reason to believe that this clone would behave similarly to STMVTMGMV. Like STMVTMGMV, it moves systemically in plants with TMGMV but not with TMV. Therefore, both are likely to be compromised in replication rather than in movement. Kurath et al. (1993 b) have shown that a similar clone was infectious in the presence of TMV at low infection frequencies (! 15 %) and that the deletion of G was '" conserved in the progeny. What was lacking in their study was the analysis of the 5« proximal sequence of the progeny genome. The deletion of G is not sufficient for initial '" replication, but is required for the eventual adaptation to TMV. What is essential is the deletion of A . It is possible that type " STMV RNA consists of only one functional sequence in the 5« UTR that mutates during helper adaptation. Alternatively, natural genetic diversity in the 5« UTR of the STMV genome may be the source of these mutations (Kurath et al., 1993 a). The virus population may be composed of a pool of RNA molecules with different sequences at the 5« end that have different specificities for different helper viruses. Studies conducted to characterize the genetic diversity in field isolates of STMV showed sequence differences in the 5« UTR, as well as in other parts of the genome of STMV (Kurath et al., 1992, 1993 a ; our unpublished results). Additional evidence for a diverse natural population includes the observation that the STMV genome exists as two variant molecules at positions 18 to 21 or 22, the first one containing four Us and the second containing five Us. The implications for genome plasticity in this region of the 5« UTR are not known, since both types of molecules were infectious with TMGMV or TMV. Accurate determination of the first base of a viral RNA is difficult. Traditional cDNA cloning techniques often generate clones missing the most terminal nucleotide(s). Several methods have been suggested for the determination of the terminal base(s) of an RNA genome (Weng & Xiong, 1995). We added homopolymer tails to the 3« end of first-strand cDNA, followed by PCR amplification to determine the sequence of the 5«-terminal base of an RNA population that had been adapted to three different helper tobamoviruses individually. These homopolymer tailings were performed using dA or dC nucleotides. The absence of first base A was very clear in these populations whether the cDNAs were dAor dC-tailed (Fig. 1). We then constructed full-length cDNA clones that produced infectious transcripts in the presence of TMV or TMGMV and their biology was similar to that of wild-type STMV infections in tobacco plants. The tailing experiments did generate some results which are open to

JBB

M. Ngon A Yassi and J. A. Dodds

different interpretations. Sequencing of the PCR products generated from helper-adapted STMV RNA when the firststrand cDNA was dA-tailed suggested the possibility of an extra base (G or U) at the 5« terminus of the STMV genome. These bases co-migrated in the sequencing gel (Fig. 1 b). We believe that this interpretation is incorrect and there is only one G at the 5« terminus of the helper-adapted STMV genome in part because the extra G runs at the same position as the first T of the homopolymer tail. In addition, when 13 independent cDNA clones that had been dC-tailed were sequenced, none of the sequences showed a T at that position, and therefore there was no competition between a G or a T at this position. Furthermore, when RNA transcripts generated from cDNA clones that included an extra G or U at the 5« end were inoculated to N. tabacum plants, a severe necrosis was induced (unpublished). This is a significantly different biology from type STMV infections and was not seen in this study. Finally, TMV did not support the replication of these clones in plants. From the observations that STMV adaptation to TMV leads to a deletion of bases 1 and 61 and that STMVTMGMV∆A " becomes ∆A and ∆G in infected plants, it could be suggested " '" that the TMV replicase not only requires a specific nucleotide sequence in the 5« UTR of the STMV genome, but also requires a specific secondary structure. The terminal base may also need to be a G since this is the case with TMV. Computer folding of the first 100 bases of the TMV}STMV RNA sequence revealed a stem–loop similar to that described in the homologous regions of brome mosaic virus RNAs in which its plays an important role in replication (Pogue & Hall, 1992). Our unpublished results show an involvement of this structure in the replication of STMV RNA. It may have a similar function for helper-adapted STMV when TMV is used as helper virus. No such structure is predicted in the folding of the genome of type STMV, which is supported by TMGMV. TMGMV replicase seems to have fewer requirements regarding the sequence and secondary structure at the 5« terminus of STMV RNA based on this study and the previous observations that full-length STMV clones that contained two or more extra bases at the 5« end (Mirkov et al., 1990 ; Kurath et al., 1992, 1993 b ; Routh et al., 1997) can be supported by TMGMV. TMV is one of the better studied plant viruses and infectious clones of this virus are being used to express foreign genes in infected plants for a number of purposes including the production of pharmaceutical proteins (Turpen et al., 1994). One of the problems encountered when using this approach is the loss of the foreign gene from the chimeric virus. Additional studies of STMV could lead to the creation of a vector for use in co-infections with TMV to allow the expression of genes in infected plants. It will be important to have STMV clones that can replicate in plants in the presence of helper tobamoviruses for which infectious clones have been created. We have created infectious clones of STMV that are able to replicate with both TMGMV and TMV. This work is an introduction to understanding the phenomenon of satellite virus helper

JBC

specificity and helper-dependent replication. A satellite genome can adapt and replicate efficiently with a new helper virus once specific changes have occurred. It also draws attention to the importance of specific replicase recognition of the 5« UTR of RNA genomes in virus replication. We thank Claude Fauquet for assistance in secondary structure analysis, Geoffrey Routh and Deborah Mathews for help with photography, and Jim Heick for help providing plants. We also thank Bret Cooper, A. L. N. Rao, T. Erik Mirkov and Gael Kurath for helpful comments on the manuscript.

References Abrahams, J. P., van den Berg, M., van Batenburg, E. & Pleij, C. (1990). Prediction of RNA secondary structure, including pseudo-

knotting, by computer simulation. Nucleic Acids Research 18, 3035–3044. Dawson, W. O., Beck, D. L., Knorr, D. A. & Grantham, G. L. (1986).

cDNA cloning of the complete genome of tobacco mosaic virus and production of infectious transcripts. Proceedings of the National Academy of Sciences, USA 83, 1832–1836. Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Research 12, 387–395. Dodds, J. A. (1991). Structure and function of the genome of satellite tobacco mosaic virus. Canadian Journal of Plant Pathology 13, 192–195. Ishikawa, M., Meshi, T., Watanabe, Y. & Okada, Y. (1988). Replication of chimeric tobacco mosaic viruses which carry heterologous combinations of replicase genes and 3« noncoding regions. Virology 164, 290–293. Ishikawa, M., Kroner, P., Ahlquist, P. & Meshi, T. (1991). Biological activities of hybrid RNAs generated by 3«-end exchanges between tobacco mosaic virus and brome mosaic virus. Journal of Virology 65, 3451–3459. Kurath, G. & Dodds, J. A. (1995). Mutation analyses of molecularly cloned satellite tobacco mosaic virus during serial passage in plants : evidence for hot spots of genetic changes. RNA 1, 491–500. Kurath, G., Rey, M. E. C. & Dodds, J. A. (1992). Analysis of genetic heterogeneity within the type strain of satellite tobacco mosaic virus reveals several variants and a strong bias for A to G substitution mutations. Virology 189, 233–244. Kurath, G., Heick, J. A. & Dodds, J. A. (1993 a). RNase protection analyses show high genetic diversity among field isolates of satellite tobacco mosaic virus. Virology 194, 414–418. Kurath, G., Rey, M. E. C. & Dodds, J. A. (1993 b). Tobamovirus helper specificity of satellite tobacco mosaic virus involves a domain near the 5« end of the satellite genome. Journal of General Virology 74, 1233–1243. Mirkov, T. E., Mathews, D. M., Du Plessis, D. H. & Dodds, J. A. (1989).

Nucleotide sequence and translation of satellite tobacco mosaic virus RNA. Virology 170, 139–146. Mirkov, T. E., Kurath, G., Mathews, D. M., Elliot, K., Dodds, J. A. & Fitzmaurice, L. (1990). Factors affecting efficient infection of tobacco

with in vitro transcripts from cloned cDNA of satellite tobacco mosaic virus. Virology 179, 395–402. Pogue, G. P. & Hall, T. C. (1992). The requirements of a 5« stem–loop structure in brome mosaic virus replication supports a new model for viral positive strand RNA initiation. Journal of Virology 66, 674–684. Rao, A. L. N., Duggal, R., Lasher, F. & Hall, T. C. (1994). Analysis of RNA replication in plant viruses. Molecular virology techniques. Methods in Molecular Genetics 4, 216–236.

STMV adaptation to TMV Roossinck, M. J., Sleat, D. & Palukaitis, P. (1992). Satellite RNAs of plant viruses - structural and biological effect. Microbiological Reviews 56, 265–279. Routh, G., Ngon A Yassi, M., Rao, A. L. N., Mirkov, T. E. & Dodds, J. A. (1997). Replication of wild-type and mutant clones of satellite tobacco

mosaic virus in Nicotiana benthamiana protoplasts. Journal of General Virology 78, 1271–1275. Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences, USA 74, 5463–5467. Turpen, T. H., Reinl, S. J., Charoenvit, Y., Hoffman, S. l., Fallarme, V. & Grill, L. K. (1994). Malarial epitopes expressed on the surface of

recombinant tobacco mosaic virus. Bio}Technology 13, 53–57.

Valverde, R. A. & Dodds, J. A. (1987). Some properties of isometric

virus particles which contain the satellite RNA of tobacco mosaic virus. Journal of General Virology 68, 965–972. Valverde, R. A., Heick, J. A. & Dodds, J. A. (1991). Interactions between satellite tobacco mosaic virus, helper tobamoviruses, and their hosts. Phytopathology 81, 99–104. Weng, Z. & Xiong, Z. (1995). A method for accurate determination of terminal sequences of viral genomic RNA. Genome Research 5, 202–207. Zuker, M. & Stiegler, P. (1981). Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acids Research 9, 133–148. Received 8 September 1997 ; Accepted 4 November 1997

JBD