Screening of the Closterovirus Genome by Degenerate Primer-Mediated Polymerase Chain Reaction

Journal of General Virology (1994), 75, 1415-1422. Printed in Great Britain 1415

Screening of the closterovirus genome by degenerate primer-mediated polymerase chain reaction

Alexander V. Karasev,t*t Olga V. Nikolaeva, 1 Eugene V. Koonin, 2 David J. Gumpf~ and Steve M. Garnsey 3

1Department of Plant Pathology, University of California, Riverside, California 92521-0122, ~National Center for Biotechnology Information, NLM, NIH, Building 38A, 8600 Rockville Pike, Bethesda, Maryland 20894 and 3 USDA- ARS Horticultural Research Laboratory, 2120 Camden Road, Orlando, Florida 32803, U.S.A.

The genome of beet yellows virus (BYV), the type complementary to the 5' region of the HSP70 genes of representative of the closterovirus group, encodes a the respective viruses were used to estimate the distance homologue of the cellular heat-shock protein (HSP) 70 between polymerase motif IV and the start point of the family. A pair of degenerate primers targeted to motifs A HSP70 gene for BYV (approximately 1-1 kb), CTV and and E, which are highly conserved in HSP70s, was BYSV (around 2.0 kb) by PCR. The amplified genome synthesized. Genomes of several definite and possible regions of CTV (3026 nucleotides) and BYSV (2837 members of the closterovirus group were screened for nucleotides) were cloned and sequenced. CTV and the presence of the HSP70 gene with PCR using these BYSV were found to encode the gene for an additional degenerate primers. BYV, citrus tristeza virus (CTV), 30K (BYSV) or 33K (CTV) protein between the poly- beet yellow stunt virus (BYSV) and carnation necrotic merase and the small hydrophobic protein genes, which fleck virus templates produced 1 kb amplification pro- was absent in BYV. These two 30K proteins displayed ducts, which were shown by sequencing to represent very weak similarity to each other, unlike the highly fragments of the respective HSP70 genes. Further conserved polymerases, hydrophobic proteins and screening was performed with an additional degenerate HSP70s of BYV, CTV and BYSV. Degenerate primer- primer targeted to the motif IV of the putative viral mediated PCR proved to be an efficient tool for rapid polymerase. This degenerate primer and specific primers screening and subsequent cloning of the viral genomes.

Introduction aphid-transmitted in a semi-persistent manner and some of them can be transmitted mechanically (Bar-Joseph et Closteroviruses are a loosely defined group of plant at., 1979; Bar-Joseph & Murant, 1982). viruses that include'10 definite and 12 possible members BYV, citrus tristeza virus (CTV), carnation necrotic (Francki et at., 1991; Coffin & Coutts, 1993; Dolja et al., fleck virus (CNFV) and beet yellow stunt virus (BYSV) 1994). The group consists of single-component fila- are definite closteroviruses and have similar biological mentous viruses with positive-strand ssRNA genomes characteristics. But they differ significantly in their (Bar-Joseph & Murant, 1982; Milne, 1988). Classifi- genome sizes, which are approximately 14.5 kb, 20 kb, cation of a virus as a closterovirus is based on two main 14.5 kb and 16.1 kb respectively (Bar-Joseph & Hull, criteria. These are, firstly, a very flexible thread-like 1974; Carpenter et at., 1977; Bar-Joseph et aI., 1985; particle morphology and secondly, a specific cyto- Dodds & Bar-Joseph, 1983 ; Reed & Falk, 1989). Lettuce pathology including the induction of beet yellows virus infectious yellows virus (LIYV) induces formation of (BYV)-type vesicles in infected cells (for a review see Bar- BYV-type vesicles in the infected cells, but is transmitted Joseph et al., 1979), Definite members of the group are by whiteflies (Duffus et al., 1986). The whitefly- transmitted beet pseudo-yellows virus (BPYV) does not induce formation of BYV-type vesicles and has a genome -~ Present address: University of Florida, Citrus Research and of only about 9 kb (Liu & Duffus, 1990; Coffin & Coutts, Education Center, 700 Experiment Station Road, Lake Alfred, Florida 1992). So far the genetic similarities between these 3385(~2299, U.S.A. viruses are unclear. However dramatic differences in genome length between closteroviruses undoubtedly The sequence data reported here have been deposited in the GenBank data library under the accession numbers L20760 (CTV) and L20761 reflect differences in genome organization. (BYSV). The sequenced 3Cterminal half of the BYV genome

Table 1. Degenerate and specific primers used for PCR

Sequence Primer Specificity

HSP70 Motif At GTGATGGTNGTNTTYGGNTTNGAYTTYGGNA I GGNWTNGAYYTNGGNACNACNT I] GYNDTNGGNMTNGAYYTNGGNACNACNTWY T CNWGYGT III Motif E + CARCARMCCKGGYARGTAMGARSWYCCACC SA IV Internal primers:[: ggtc tagAMRNAMGTRGGTATGTA V BYV CACGTAGGTATGTACGATGA VI CTV GTGTCGGAGTATAAAAACAGACATGTTGG VII BYSV Polymerase Motif IVt gg tgaGAYCTNTCNAARTTYGAYAARTCNCA VIII * M represents A or C; R, AorG;W, AorT;S, CorG;Y, CorT;K, GorT;D,A, Gor T; N, any base. Non-virus-specific bases are indicated in lower case. ~ Sense-strand primer sequence. :~ Sequence complementary to the sense-strand. encodes eight genes (Agranovsky et al., 1991a). BYV LIYV in sugar beet (Beta vulgaris) and BPYV in melon (Cucumis melo). polymerase belongs to supergroup 3 of positive-strand Virion RNA of the California isolate of BYV, as well as BYV dsRNA isolated from Tetragonia expansa, were kindly provided by Dr RNA virus polymerases and in phylogenetic recon- R. Creamer (University of California, Riverside, Calif., U.S.A.). structions is closely related to the polymerases of Isolation of virus RNA. The dsRNAs were isolated from virus- tobamo-, tobra-, hordei- and tricornaviruses (Agra- infected sowthistle, sugar beet and melon leaves by two cycles of CF- novsky et al., 1991b; Koonin, 1991; Koonin & Dolja, 11 cellulose column chromatography, essentially as described by 1993). The unique feature of the BYV genome is the Valverde et al. (1990). Either ss- or dsRNA (2 to 3 gg) was denatured presence of a gene encoding a 65K protein, which is a with 20 mM-methylmercuric hydroxide and the first-strand cDNA was homologue of the cellular heat-shock protein (HSP) 70 synthesized using Moloney murine leukaemia virus reverse tran- scriptase (USB) in accordance with conventional procedures (Maniatis family and which appears to retain all functionally et al., 1982). important domains of HSP70 proteins (Agranovsky et al., 1991 b; Ting & Lee, 1988). A small hydrophobic 6-4K Primers. Degenerate primers used to amplify the 5'-terminal fragments of the HSP70-1ike genes and the genome fragments between protein is encoded between the polymerase and HSP70 the polymerase and HSP70-1ike genes are listed in Table 1. All genes of BYV (Agranovsky et al., 1991a). degenerate and specific primers were synthesized at the Biotechnology Here we attempted to screen the genomes of CTV, Instrumentation Facilities at the University of California, Riverside. BYSV, CNFV, LIYV and BPYV for the presence of an PCR amplification. About 5 % of the cDNA obtained in the reverse HSP70-related gene using degenerate primer-mediated transcription reaction was used in PCR. The amplification was PCR. This gene was shown to be encoded by the CTV, performed in a final volume of 100 gl with 1 unit of Taq DNA BYSV and CNFV genomes, in addition to BYV. We polymerase (Boehringer Mannheim), 140 pmol of each primer, 0.4 mM- dNTPs and incubation buffer (10 mM-Tris HC1 pH 8.3, 50 mM-KC1, amplified, cloned and sequenced genome fragments 1.5mt~-MgCl2, 0-1 mg/ml gelatine; Boehringer Mannheim). The between the polymerase motif IV and the HSP70 gene amplification protocol involved: one cycle of 94 °C for 2 min; two for CTV and BYSV using the same PCR technique. In cycles of 94 °C for 30 s, 37 °C for 30 s and 72 °C for 2 min; 30 cycles of this region both CTV and BYSV encode an additional 94 °C for 30 s, 50 °C for 30 s and 72 °C for 2 rain; one cycle of 72 °C gene that is absent from the BYV genome. for 10 min. After completion of the reaction, the resulting PCR products were analysed on 1% agarose gels in Tris-acetate-EDTA buffer. Methods Cloning of PCR products. These were first treated with T4 DNA Virus isolates and propagation. The severe California isolate SY568 polymerase (USB) to give blunt cDNA termini, and the blunt-ended of CTV was maintained in Mexican lime (Citrus aurantijblia) under cDNAs were phosphorylated with T4 polynucleotide kinase (USB). greenhouse conditions, and the virus RNA was isolated as described These were then resolved in 1% low-gelling temperature (LGT) previously (Bar-Joseph et al., 1985). The purified dsRNA preparation agarose gels (Gibco-BRL). Individual bands of interest were cut and of the severe Florida isolate T-36 (Rosner et al., 1986) was provided by the cDNAs were isolated from the LGT agarose (Maniatis et al., 1982). Dr R. F. Lee (CREC, Lake Alfred, Fla., U.S.A.). dsRNA preparation The resulting cDNAs were ligated with T4 DNA ligase (New England of the CNFV (Australia isolate KF) was provided by Dr R. Jordan Biolabs) into Smal-cut pBluescript SK (Stratagene). The ligafion (USDA, Beltsville, Md., U.S.A.). BYSV, LIYV and BPYV were kindly products were transformed into competent DH5~ cells (Gibco-BRL) provided by Dr J. E. Duff'us (USDA, Salinas, Calif., U.S.A.) as virus- according to the manufacturer's protocol. Two BYSV-specific recom- infected plant tissues, with BYSV in sowthistle (Sonchus oleraceus), binant plasmids, pBS417 and pBS61, containing inserts of 2.0 and Closterovirus genome screening 1417

1-0 kbp respectively, were selected for sequencing. The former encompassed the coding region between polymerase and HSP70, and the (a) latter encompassed the Y-terminal half of the HSP70 gene. A CTV- M 1 2 3 4 5 6 specific plasmid, pCT13 with the 2.1 kbp insert, contained the region between the polymerase and HSP70. The Y-terminal half of the CTV HSP70 locus was similarly cloned with the use of the same degenerate kbp primers. The sequence of the complete CTV HSP70 has been reported elsewhere (Pappu et al., 1994). Sequencing. Cloned cDNAs were sequenced directly in double- stranded plasmids using a Sequenase 2.0 kit (USB) according to the manufacturer's instructions, using universal T3 and T7 primers and virus-specific primers. In some cases Y-terminal regions of the HSP70- specific products were sequenced directly after 5' end-labelling of the primers I and IV (Table l), with the use of anfmol PCR sequencing kit (Promega). Both strands were sequenced at least twice. 10-0 -- Analysis of sequences. The sequences were translated by computer with the program PROTMAKE. This program includes an assessment of the efficiency of expression of potential open reading frames (ORFs) 5.0 -- using the algorithm of Trifonov (1987), based on the difference in guanine distribution in coding and non-coding triplets. Putative translation products were compared with the non-redundant sequence database at the National Center for Biotechnology Information (NIH) 3.0 -- using programs based on the BLAST algorithm (Altschul et al, 1990). The program BLASTP was used to search the amino acid sequence database and the program TBLASTN was used to search the 2.0 -- conceptually translated nucleotide sequence database. Pairwise comparisons were performed using the ALIGN-2 program utilizing the DOTHELIX algorithm (Brodsky et al., 1990). Multiple sequence alignments were generated by the MULTALIN program (Corpet, 1988). PROTMAKE and ALIGN-2 are components of the GENEBEE software package (Brodsky et al., 1990). (b) M 1 2 3 4 5 6 C Results Profiles of dsRNA kbp Electrophoretic analysis of the dsRNA isolated from infected plant tissues revealed obvious similarities in the sizes of the largest dsRNA species for BYV and CNFV on one hand, and CTV and BYSV on the other hand (Fig. 1 a). No difference was evident in the electrophoretic mobilities of the largest dsRNA bands within these two pairs. However, the size of the largest dsRNA band, 2.0 m which presumably represents the replicative form of the virus genomic RNA, is greater in CTV and BYSV than in BYV and CNFV (Fig. la). This may suggest a substantially longer genome for BYSV than was reported 0.8 m previously (Reed & Falk, 1989). We estimated the BYSV genome to be approximately the same length as the CTV 0.5-- genome (about 20 kb; Bar-Joseph et aL, 1985). Never- 0.3-- theless, the set of smaller BYSV dsRNA bands seems to be similar to BYV/CNFV dsRNA patterns. In contrast, LIYV and BPYV dsRNA profiles were quite different from the BYV/CNFV and CTV/BYSV patterns as well as from each other (Fig. 1 a). The largest band of BPYV Fig. 1. Electrophoretic analysis in nort-denaturing agarose gets of (a) corresponded to a fragment of approximately 9 kbp, in dsRNA preparations and (b) PCR products amplified using primers I and IV on the corresponding cDNA template. Lanes 1, BYV; lanes 2, agreement with previously published estimates (Coffin & CNFV; lanes 3, CTV; lanes 4, BYSV; lanes 5, BPYV; lanes 6, LIYV; Coutts, 1992), and the largest LIYV dsRNA was lanes M, phage lambda DNA digested with PstI as a marker. Lane C estimated to be around 12 kb (Fig. 1). At least seven in (b) is the PCR control, with no template added. 1418 A. V. Karasev and others' bands of smaller sized products were revealed among M 1 2 3 dsRNAs from LIYV-infected sowthistle plants (Fig. 1a). kbp HSP70 gene-targeted degenerate primer-mediated PCR Three degenerate primers specific for HSP70 motif A were synthesized. The primers differed in length, degree of degeneracy, and position within motif A (Table 1). The preliminary selection of the best motif A-specific primer and the optimum amplification profile were 3.0 m obtained using two specific T-terminal primers complementary to the 3' termini of the BYV and CTV HSP70 2.0-- genes (Agranovsky et al., 1991b; Pappu et al., 1994). Two initial cycles of PCR with a low annealing temperature were found to be essential for amplification 1.0-- with degenerate primers. Primer I gave the best yield under the optimum conditions and was chosen for further work. Primer IV (complementary to HSP70 0-5-- motif E), when combined with primer I, gave rise to a prominent amplification product of about 1 kbp for BYV, CNFV, CTV and BYSV templates (Fig. 1 b). The Fig. 2. Electrophoreticanalysis in a non-denaturingagarose gel of PCR LIYV template yielded a smaller sized product around products amplified on the respective templates. Lane 1, CTV with 0.7 kbp, but no amplification products were observed primers VIII and VI; lane 2, BYV with primers VIII and V; lane 3, with the BPYV template (Fig. l b). Similarly sized BYSV with primers VIII and VII; lane M, phage lambda DNA digested with PstI as a marker. amplification products were not detected when the dsRNA preparations were used directly as PCR templates, suggesting the absence of contaminating HSP70- primer V was used to prime BYV, CTV and BYSV related DNA sequences of host origin (data not shown). templates. However, this primer appeared to be BYV- Preparations of dsRNA used for the CTV first-strand specific, suggesting that even minor variations in se- cDNA synthesis consistently gave better amplifications quence may significantly influence the priming efficiency with degenerate primers than did ssRNAs, as the number of degenerate primers. Primer V will be referred to and intensities of additional, presumably non-specific, hereafter as the BYV-specific primer (Table 1). Two bands were substantially reduced (data not shown). other CTV- and BYSV-specific primers complementary Individual 1 kbp CTV and BYSV-specific bands were to the region about 100 nucleotides downstream of the isolated, cloned and sequenced as described in Methods. HSP70 gene AUG codon were used with the VIII primer. The amplified DNA fragments were confirmed to The BYV polymerase motif IV has not been sequenced represent the 5'-terminal half of the respective CTV and but the 1.1 kbp product amplified on the BYV reported BYSV HSP70 genes, except the extreme 5' and 3' so far (Fig. 2) was of the expected size. This was based on sequences that represented the artificial primers. The the apparent similarity of BYV polymerase to poly- extreme 5' sequence included the most conserved motif, merases of tricornaviruses and on the reported BYV A, characteristic of all HSP70s, so we attempted to genome sequence between the polymerase and the HSP70 obtain the actual sequence by extending the degenerate gene (Agranovsky et al., 1991 a, b). In contrast the CTV primer-mediated PCR to the polymerase gene. and BYSV templates yielded substantially larger PCR products of about 2-1 and 1.9 kbp, respectively with the same VIII primer (Fig. 2). This suggests that CTV and Degenerate primer-mediated PCR targeted to RNA- BYSV may encode additional gene(s) between the dependent RNA polymerase motif IV polymerase and HSP70 genes. Degenerate primer VIII (Table 1) was based upon the consensus sequence of motif IV in the RNA-dependent Genome organization of BYSV and CTV between RNA polymerases of tricornaviruses (Koonin, 1991) and polymerase and HSP70 genes was used to amplify genome fragments of BYV, CTV and BYSV. These fragments were expected to encompass The cDNAs corresponding to the genome regions located the 3'-terminal two-thirds of the polymerase gene and the between polymerase motif IV and HSP70 motif E for 5'-terminal portion of the HSP70 gene. The degenerate BYSV (2837 nucleotides) and CTV (3026 nucleotides) Closterovirus genome screening 1419

(a) (c) V VI A BYV DNQRKSGASNTWIGNSIETLGILSMFYYTNRFKALFVSGDDSLIFSESPIRNSADAMCTE BYV MVVFGLDFGTTFSSVCAYVGEELYLFKQRDSAYIPTYVFLHSDTQEVAFGYDAEVLSNDL BYSV DCQRKSGASNTWIGNSVVTLGILAMYYDVSKFQALFVSGDDSLIFSADEIANYAEDiCLE BYSV MVVFGVDFGTTFSSVCVFNSGRLHVFKQQNSAYIPTCLFLYSDTMAMSFGYDAETASLDP CTV DGQRRSGGSNTWIGNSLVTLGILSLYYDVSKFDtLLVS6DDSLIYSSEKISNFSSEICLE CTV MVCLGCDFGTTFSTVAMATPSELViLKQFNS½YIPTCLLLHAEPNSVSYGYDAEYLAAS- C D QRkSG SNTWIGNSu TLGIL u@Y F L&VSGDDSLI@S I N uC E C MVu&GuDFGTTFS V L u&KQ S YIPT u&L u @GYDAE

VII VIII BYV LGFETKFLTPSVPYFCSKFFVMIGHDVFFVPDPYKLLVKLGASKDEVDDEFLFEVFTSFR BYV SVRG6FYRDLKRWIGCDEENYRDYLEKLKPHYKTELLKVAQSSKSTVKLDCYSGIVPQNA BYSV LGFETKFLTPSVPYFCSKFLVFTGDKCVFVPDPYKLLVKL6ASGRRLSDEELFEVFVSFR BYSV NVKGGFFRDLKRWVGCDETNIEEYKSKLKPHYSVTLSNFGKGSRKIPTLGSYSGSVQMSG CTV TGFETKFMSPSVPYFCSKFVVQTGNKTCFVPDPYKLLVKLGAPQNKLTDVELFELFTSFK CTV GESGSF{EDLKRWVGCTAKNYQTYLHKLSPSYKVIVKEFGTKSVPVPYLSPLNNDEGLSV C GFETKFu PSVPYFCSKF@V TG FVPDPYKLLVKLGA u DE LFEuF SF C G F@ DLKRWuGC N& Y KL PHY u & S P L & u

B BYV DLTKDLVDERVIELLTHLVHSKYGYESGDTYAALCAIHCIRSNFSSFKKLYPKVKGWVVH BYV TLPGLIATFVKALISTASEAFKCQCTGVICSVPANYNCLQRSFTESCVNkSGYPCVYMVN BYSV DLTKEFGDERVLNTLSELVHLKYEFEWGNIKLALSTIHCLRSNFLSFSKLFVKRTGWKVV BYSV SLSGLIALFIQALVKSAAIEFKCECTELIVSVPANCDCMQRLFTENCVNLSGFTCVHMMN CTV DMTQDFGDQVVL~KLKLEVEAKYGFASGTTMPALCAIHCVRSNFLSFERLFPFIRGW-YV CTV ALPSLIAgYAKSILSDAERVFNVSCTGVICSVPAGYMTLQRAFTQQSISMSGYSCVYIIN C DuT & D Vu L LV KY @ G T AL IHCuRSNF SF L@ 6W & C L LIA @ uu A F CT ul SVPA uQR FT u uSG@ CV u N

BYV YGKLKFV-LRKFAN--CFREKFDT ..... AFGERTFLLTTKLETVL C D BYV EPSAAALSACSRIKGATSPVLVYDFGGGTFDVSVISALNNTFVVRASGGDMNLGGRDIDK BYSV YGI~KYI-LKKFLG--YNIEPITT ..... TFGDAWFVYKE CTV VDALKLRQLRKLTNLICERVVYDNRVSYFSYFDNPFTKPDANDDNVDDLGQAGELATG BYSV EPSA~LSTCGRTDMSARNLLVYDFGGGTFDVSVLSSLNQTFTVRASG6DMNLGGRDVDR C & K& L K& & @ F CTV EPSAAAYSTLPKLNSADKYLAVYDFGGGTFDVSIVSVRLPTFAVRSSSGDMNLGGRDIDK (b) C EPSAAA&S u VYDFGGGTFDVSuuS TF VR S GDMNLGGRDuD BYSV MDCILRAFLPFGFALVI-CFFIAVAAYFFAFFVKNT--HSQDTDVDIRQEDLAG BYV MDCVLRSYLLLAFGFLI-CLFLFCLVVFIWFVYKQILFRTTAQSNEARHNHSTVV BYV AFVEHLYNKAQLPVNYKIDISFLKESLSKKVSFLNFPVVSEQGVRVDVLVNVSELAEVAA CTV MDCVIQGFLTFLVGIAVFCAFAGLIIIVITI-YRCTIKPVRSASP~GTHATV BYSV A-$SKIYQMANLPFDEEADISSLKESESKIDYPITYTVKTKDGESKTVVVSRGtLAEVIV C MDCuu @L & & & u C F &&& & & CTV KLSDKIYEMADFVPQKELNVSSLKEALSLQTDPVKYTV-NHYGMSETVSIDQTVLREIAS

Fig. 3. (a) Amino acid sequence alignment of the C-terminal fragments C uY A & uS LKE LS u @V G V u L Eu of the putative RNA-dependent RNA polymerases of BYV, BYSV and CTV. The motifs V to VIII, conserved in RNA polymerases of positive- strand RNA viruses, are indicated above the sequence alignments. (b) E Amino acid sequence alignment of the small hydrophobic proteins of B~ PFVERTIKIVKEVYEKYCSSMRLEP-NVKAKLLMVGGSSYLPGLL CTV, BYSV and BYV. (e) Amino acid sequence aligmnent of the N- BYSV PFVDRTIKVMKRVFELYVK~MNLKAQDAKAALVLVGGSSYLPGLK terminal fragments of the HSP70 proteins of BYV, BYSV and CTV. CTV VFiNRTIDiLTQV--KVKSSM---PESQSLKLVVVGGSSYLPGLL The conserved motifs (A to E) characteristic of the HSP70 family are indicated above the alignment. Asterisks show identical residues and C Fu RTI uu V & M LuuVGGSSYLPGL colons indicate similar residues. C is the consensus sequence, with residue designations: U for I, L, V or M (bulky aliphatic); @ for F, Y or W (aromatic); & for I, L, V, M, F, Y or W (bulky hydrophobic). conserved. The alignment scores (AS; see Koonin, 1991) exceeded 25s.0. (an AS of over 10s.o. is highly significant). The reported amino acid sequence of the were cloned and sequenced. The sequence of the 5'- BYV polymerase, which starts within the polymerase terminal half of the CTV HSP70 gene (positions 2000 to motif V (Agranovsky et al., 1991 b), is also closely related 3026) has already been presented elsewhere (Pappu et al., to the polymerases of BYSV and CTV (Fig. 3 a). The AS 1994). Both BYSV and CTV appeared to have very calculated for BYV using the DOTHELIX program for similar gene arrangements in this region, with two ORFs pairwise comparisons exceeded 21 s.o. The sequenced located between the polymerase and the HSP70 genes. portions of the BYSV and CTV polymerases cover five Based on the criteria of Trifonov (1987) these ORFs were of the eight conserved motifs in the polymerases of estimated to be efficiently expressed. positive-strand RNA viruses (Koonin, 1991).

O) Polymerase (ii) 30K and 33K proteins Beginning with polymerase motif IV, the C-terminal amino acid sequences of BYSV and CTV polymerases ORFs downstream of the polymerase gene potentially (266 and 291 amino acids respectively) are highly encode proteins of M r 30382 (BYSV, hereafter desig- 1420 A. V. Karasev and others nated 30K) and 33396 (CTV, 33K) proteins. Direct unusually diverse closterovirus group as it might over- comparison of the 30K and 33K proteins using the come difficulties in characterization caused by the low DOTHELIX program revealed only very weak similarity virus concentration in some infected plants and the lack between these similarly encoded proteins (data not of efficient purification protocols. shown). The BLAST search through the database did The sequences of the genomic regions of BYSV and not indicate any reliable relationships. CTV presented here each encode four genes (Fig. 4). Comparison of gene organization and phylogenetic analysis of the RNA-dependent RNA polymerases (iii) Small hydrophobic proteins (Koonin, 1991; Koonin & Dolja, 1993) and of the Downstream of the 30K protein gene of BYSV and the HSP70s clearly indicate that CTV, BYSV and BYV form 33K protein gene of CTV, a short ORF was found which a compact group within the Sindbis-like supergroup of potentially encodes a small hydrophobic protein of 51 positive-strand RNA viruses, which probably includes a amino acids in both (hereafter called the 6K protein). number of other closteroviruses. The presence of the These two proteins appeared to be similar to each other HSP70 gene is a unique hallmark of these viruses. On the and to the small 6.4K protein of BYV (Fig. 3 b), with the other hand, significant differences in genome organ- AS in pairwise comparisons exceeding 6 s.D. (modestly ization have been observed between different members of significant). The similarity between these proteins was the closterovirus group. These differences are illustrated found to be most extensive in the N-terminal portions by the presence of an additional gene coding for a 30K and to gradually decrease towards the C terminus. protein in both CTV and BYSV, but not in BYV (Fig. 4). The role of these proteins in virus replication is not known, but the lack of sequence similarity between them (iv) HSP70 proteins is particularly intriguing. A 250K protein, synthesized by in vitro translation of The incomplete ORF downstream of the BYSV 6K the BYV genomic RNA, was proposed to include the protein gene encoded the N-terminal 345 amino acids of virus polymerase (Karasev et al., 1989). It seems a BYSV HSP70 protein. This conclusion is supported by reasonable to suggest a similar placement of the BYSV the alignment of BYSV HSP70 (Fig. 3c) with the and CTV polymerases within large proteins encoded at respective N-terminal fragments of the HSP70s of BYV the 5'-proximal half of the respective genomes. Several (Agranovsky et al., 1991 b) and CTV (Pappu et al., 1994). subgenomic (sg) RNAs have been detected in plants Six of the eight most conserved motifs that are infected with BYV (Dolja et al., 1990) and CTV (M. E. characteristic of the HSP70 family could be identified in the N-terminal half of the HSP70 of BYSV. Moreover, the similarity of viral HSP70s extends beyond these most (a) conserved, functionally important motifs with the AS in pairwise comparisons above 24 S.D., suggesting close Polymerase HSP70 evolutionary relationships between these three proteins. lm 30K 6K Discussion (b) The technique of degenerate primer-mediated PCR has Polymerase 33K HSP70 been widely used for the screening of homologous genes. In most cases however, this approach is used to amplify lliiiiiiiiiii !iii!i!ii ii ! I relatively short gene fragments for subsequent use as 6K hybridization probes to screen large genomic or cDNA (c) libraries (Chang et al., 1990; Kalman et al., 1991 ; Gupta & Singh, 1992). In this study we demonstrated the Polymerase HSP70 general applicability of degenerate primer-mediated PCR for direct, step-by-step cloning of extended fragments of viral genomes. The key element of the experimental I I 6K 500 design seems to be the optimum choice of the conserved nucleotides protein sequence motif. After the proper target has been Fig. 4. A comparison of the genome organizations of BYSV (a), CTV, found, the technique could be applied for taxonomic and (b) and BYV (c) in the sequenced regions. Rectangles represent ORFs. classification purposes as well (Fig. 1 and 2). This The same shading is used to denote ORFs encoding proteins displaying approach appears to be particularly important for the a statistically significant similarity to each other. Closterovirus genome screening 1421

Hilf, personal communication). Two of the CTV-specific CARPENTER, J. M., KASSANIS, B. & WHITE, R. F. (1977). The protein and nucleic acid of beet yellows virus. Virology 77, 101-109. sgRNAs correspond in size to putative sgRNAs for the CHANG, T.-H., ARENAS, J. & ABELSOY, J. (1990). Identification of five 33K protein-encoding gene and the HSP70 gene (M. E. putative yeast RNA helicase genes. Proceedings of the National Hilf, personal communication), suggesting that these Academy of Sciences, U.S.A. 87, 1571 1575. COFFIN, R.S. & COUTTS, R. H.A. (1992). DsRNA cloning and genes are expressed during infection. Recently, it was diagnosis of beet pseudo-yellows virus by PCR and nucleic acid demonstrated that the 6"4K protein and the HSP70 hybridization. Intervirology 33, 197-203. protein of BYV are simultaneously expressed in a cell- COFFIN, R.S. & COUTTS, R.H.A. (1993). The closteroviruses, capilloviruses and other similar viruses: a short review. Journal of free transcription-translation system, independently of General Virology 74, 1475-1483. each other and from the same RNA transcript (Karasev CORPET, F. (1988). Multiple sequence alignment with hierarchical et al., 1992; A. V. Karasev, unpublished results). Similar clustering. Nucleic Acids Research 16, 10881 10890. DODDS, J. A. & BAR-JOSEPH, M. (1983). Double-stranded RNA from arrangements of the HSP70 gene and the preceding small plants infected with closteroviruses. Phytopathology 73, 419-423. ORF (Fig. 4) may suggest similar expression strategies DOLJA, V. V., KARASEV, A. W. & AGRANOVSKY, A. A. (1990). Organ- for CTV and BYSV. ization of the beet yellows closterovirus genome. In New Aspects of Positive-Strand RATA Viruses, pp. 31 35. Edited by M. Brinton & The data obtained indicate that BYSV is more closely F. Heinz. Washington, D.C. : American Society for Microbiology. related to CTV than to BYV, in contrast with previous DOLJA, V.V., KARASEV, A.V. & KOONIN, E.V. (1994). Molecular suggestions (Duffus, 1979; Reed & Falk, 1989). Our data biology and evolution of closteroviruses: sophisticated build up of the large RNA genomes. Annual Review of Phytopathology 32 (in suggest the division of closteroviruses into at least two press). groups with CTV and BYV as type representatives, and DUFFUS, J, E. (1979). Beet yellow stunt virus. CM1/AAB Descriptions with BYSV to be tentatively included in the CTV group. of Plant Viruses, no. 207. DUFFUS, J. E., LARSEN, R. C. & LIU, H. Y. (1986). Lettuce infectious yellows virus a new type of whitefly-transmitted virus. Phyto- Note added in proof. After the manuscript was accepted for pathology 76, 97 100. publication, the complete sequence of the BYV polymerase gene was FRANCKI, R. I. B., FAUQUET, C.M., KNUDSON, D. L. & BROWN, F. reported [Agranovsky et al., (1994) Virology 198, 311 324], allowing (1991). Classification and nomenclature of viruses. Fifth Report of the precise calculation of the distance between the polymerase motif IV the International Committee on Taxonomy of Viruses. Archives of (primer VIII) and the position of primer V. The actual distance Virology Supplementum 2. appeared to be the same as experimentally observed (see Fig. 2). GUPTA, R. S. & SINGH, B. (1992). Cloning of the HSP70 gene from Halobacterium marismortui: relatedness of archaebacterial HSP70 to We are grateful to all our colleagues who provided us with its eubacterial homologs and a model for the evolution of the HSP70 closteroviruses and clostero-like viruses, and to Dr H. R. Pappu for gene. Journal of Bacteriology 174, 459~4605. communicating sequence data to us prior to publication. We are KALMAN, M., MURPHY, H. & CASHEL, M. (1991). RhlB, a new indebted to Dr M. E. Hilf for numerous helpful discussions, to Dr Escherichia coli K-12 gene with an RNA helicase-like protein W. O. Dawson for encouragement and support, and to both for a sequence motif, one of at least five such possible genes in a critical reading of the manuscript. This work was supported in part by prokaryote. New Biologist 3, 886-895. the USDA-ARS cooperative agreement 58-5310-1-205 and a grant KARASEV, A.V., AGRANOVSKY, A.A., ROGOv, V.V., MIROSHNI- from the California Citrus Advisory Board (both to D.J.G.). CHENKO, N.A., DOLJA, V.V. & ATABEKOV, J.G. (1989). Virion RNA of beet yellows closterovirus: cell-free translation and some properties. Journal of General Virology 70, 241 245. KARASEV, A.V., KASHINA, A.S., GELFAND, V.I. & DOLJA, V.V. References (1992). HSP70-related 65-kDa protein of beet yellows closterovirus is a microtubule-binding protein. FEBS Letters 304, 12-14. AGRANOVSKY, A. A., BOYKO, V. P., KARASEV, A. V., LUNINA, N. A., KOONIN, E.V. (1991). The phylogeny of RNA-dependent RNA KOONIN, E. V. & DOLJA, V. V. (1991 a). Nucleotide sequence of the polymerases of positive-strand RNA viruses. Journal of General T-terminal half of beet yellows closterovirus RNA genome: unique Virology 72, 2197~206. arrangement of eight virus genes. Journal of General Virology 72, KOONIN, E.V. & DOLJA, V.V. (1993). Evolution and taxonomy of 15-23. positive-strand RNA viruses: implications of comparative analysis AGRANOVSKY, A. A., BOYKO, V. P., KARASEV, A. V., KOONIN, E. V. & of amino acid sequences. CRC Critical Reviews in Biochemistry and DOLJA, V.V. (1991b). Putative 65-kDa protein of beet yellows Molecular Biology 28, 375-430. closterovirus is a homologue of HSP70 heat shock proteins. Journal L~u, H.Y. & DUFEUS, J.E. (1990). Beet pseudo-yellows virus: of Molecular Biology 217, 603-610. purification and serology. Phytopathology 80, 866-869. ALTSCHUL, S. F., GISH, W., MILLER, W., MYERS, E. W. & LIPMAN, D. J. MANIATIS, T., FRITSCH, E.F. & SAMBROOK, J. (1982). Molecular (1990). Basic local alignment search tool. Journal of Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Biology 215, 403-410. Laboratory. BAR-JosEPH, M. & HULL, R. (1974). Purification and partial character- MILNE, R. G. (1988). Taxonomy of rod-shaped filamentous viruses. In ization of sugar beet yellows virus. Virology 62, 552-562. The Plant Viruses, vol. 4, The Filamentous Plant Viruses, pp. 3-50. BAR-JOSEPH, M. & MURANT, A.F. (1982). Closterovirus group. Edited by R. G. Milne. New York: Plenum Press. CMI/AAB Descriptions of Plant Viruses, no. 260. PAPPU, H. R., KARASEV, A. V., ANDERSON, E. J., PAPPU, S. S., HILF, BAR-JOSEPH, M., GARNSEY, S.M. & GONSALVES, D. (1979). The M. E., FEBRES, V., ECKLOFF, R. M. G., MCCAFFERY, M., BOYKO, V., closteroviruses: a distinct group of elongated plant viruses. Advances GOWDA, S., DOLJA, V. V., KOONIN, E. V., GUMPF, D. J., CLINE, K., in Virus Research 25, 93-168. GARNSEY, S.M., DAWSON, W.O., LEE, R.F. & NIBLETT, C.L. BAR-JOSEPH, M., GUMPF, D. J., DODDS, J. A., ROSNER, A. & GINZBERG, (1994). Nucleotide sequence and organization of eight 3' open I. (1985). A simple purification method for citrus tristeza virus and reading frames of the citrus tristeza closterovirus genome. Virology estimation of its genome size. Phytopathology 75, 195-198. 199, 35-46. BRODSKY, L. I., DRACHEV, A. L., TATUZOV, R. L. & CHUMAKOV, K. M. REED, R. R. & FALK, B. W. (1989). Purification and partial character- (1990). GENBEE: a package of computer programs for biopolymer ization of beet yellow stunt virus. Plant Disease 73, 358 362. sequence analysis. Biopolimery i Kletka 7, 10-14. ROSNER, A., LEE, R.F. & BAR-JOsEPH, M. (1986). Detection of a 1422 A. V. Karasev and others

specific Florida isolate of citrus tristeza virus by differential rRNA nucleotide sequences. Journal of Molecular Biology 194, hybridization with cloned cDNA sequences. Phytopathology 76, 643 652. 82(~824. VALWRDE, R. A., NAMETH, S. T. & JORDAN, R. L. (1990). Analysis of TING, J. & LEE, A. S. (1988). Human gene encoding the 78,000-dalton double-stranded RNA for plant virus diagnosis. Plant Disease 74, glucose-regulated protein and its pseudogene: structure, conser- 255 258. vation, and regulation. DNA 7, 275-286. TRIFONOV, E.N. (1987). Translation framing code and frame-moni- (Received 30 September 1993; Accepted4 January 1994) toring mechanism as suggested by the analysis of mRNA and 16S