MATERIALS AND METHODS The sources of the viral RNAs, oligonucleotides, enzymes and nucleotides have been reported [2].

Transcription

1 Eur. J. Biochem. 150, (1985) (C) FEBS 1985 Nucleotide sequence of cucumber mosaic virus RNA 1 Presence of a sequence complementary to part of the viral satellite RNA and homologies with other viral RNAs M. Ali REZAIAN, Rhys H. V. WILLIAMS and Robert H. SYMONS Adelaide niversity Centre for Gene Technology, Department of Biochemistry, niversity of Adelaide (Received February 6/April24, 1985) - EJB The nucleotide sequence of the 3389 residues of RNA 1 (M, 1.15 x lo6) of the Q strain of cucumber mosaic virus (CMV) was determined, completing the primary structure of the CMY genome (8617 nucleotides). CMY RNA 1 was sequenced by the dideoxy-chain-termination method using M13 clones carrying RNA 1 sequences as well as synthetic oligonucleotide primers on RNA 1 as a template. At the 5 end of the RNA there are 97 noncoding residues between the cap structure and the first AG (98 - loo), which is the start of a single long openreading frame. This reading frame encodes a translation product of 991 amino acid residues (Mr ) and stops 319 nucleotide residues from the 3 end of RNA 1. In addition to the conserved 3 region present in e!l CMV RNAs (307 residues in RNA l), RNAs 1 and 2 have highly homologous 5 leader sequences, a 12-nucleotide segment of which is also conserved in the corresponding RNAs of brome mosaic virus (BMV). CMV satellite RNA can form stable base pairs with a region of CMV RNAs 1 and 2 including this 12-nucleotide sequence, implying a regulatory function. This conserved sequence is part of a hairpin structure in RNAs 3 and 2 of CMV and BMV and in CMV satellite RNA. The entire translation products of RNA 1 of CMV and BMV could be aligned with significant homology. Less prominent homologies were found with alfalfa mosaic virus RNA 1 translation product and with tobacco mosaic virus M,-I26000 protein. The genome of cucumber mosaic virus (CMV) consists of three RNAs. RNA 1 is the largest species and has an estimated molecular mass (M,) of 1.35 x lo6 [l]. RNA 2 and RNA 3 are 3035 and 2193 nucleotides long respectively [2, 31. A fourth, subgenomic RNA of 1027 residues is also encapsidated in CMV particles [3]. The genome of CMV can support the replication of CMVspecific satellite RNAs (sat-rnas) (reviewed in [4, 51). Sat- RNA isolates have nucleotides and, like the helper virus RNAs, have a cap structure at their 5 end [4]. nlike the genomic RNAs, sat-rna is not aminoacylatable [4], although it resembles the conserved 3 end of CMV RNAs in having extensive secondary structure [6]. Satellite RNA has no significant sequence homology with the helper virus when tested by liquid hybridization kinetics [7]. The in vitro translation product of CMV RNA 1 has an M, of 95000, as estimated by gel electrophoresis [XI. RNA 2 encodes a translation product 839 amino acids long (M, [2]) and RNA 3 encodes two translation products, a 5 -terminal 3a protein 333 amino acid residues long and a 3 -terminal coat protein 236 amino acid residues long [3]. RNA 4 also encodes the coat protein. Sequence comparisons, using the translation products of RNA 2 of CMV, brome mosaic virus (BMV) and alfalfa mosaic virus (AMV) and of a region of tobacco mosaic virus (TMV) RNA encoding the M, protein, have revealed significant regions of homol- Correspondence to R. H. Symons, Department of Biochemistry, niversity of Adelaide, Adelaide, South Australia, Australia 5000 Abbreviations. AMV. alfalfa mosaic virus; BMV, brome mosaic virus; CMV, cucumber mosaic virus; TMV, tobacco mosaic virus; cdna, complementary DNA; sat-rna, satellite RNA. Enzymes. Reverse transcriptase (EC ); restriction endonucleases Accl (EC ) and TuqI (EC ); DNA polymerase I (EC ). ogy, indicating that these viruses are evolutionally related [2, 91. Since the primary structures of the CMV genome and its satellite RNAs are complete, we have conducted a comparative analysis of the sequences. The results provide further evidence for the evolutionary relationship between the tripartite viruses and TMV, and implicate specific interaction between CMV and its sat-rna. MATERIALS AND METHODS The sources of the viral RNAs, oligonucleotides, enzymes and nucleotides have been reported [2]. cdna cloning and sequencing of nucleic acids cdna clones of CMV RNA 1 were prepared by ligating TnqI restriction fragments of double-stranded cdna prepared from RNA into the AccI site of MI 3mp7 as described [2] and used for dideoxy-chain-termination sequencing. Sequencing reactions were carried out as described previously [2]. When band compressions occurred in DNA and RNA sequencing, the reactions were modified [2] so that dgtp was substituted with ditp at a concentration of 0.5 mm, the dideoxyguanosine 5 -triphosphate (ddgtp) concentration was reduced to 4 pm and reverse transcriptase was used in place of Klenow fragment of DNA polymerase I on DNA templates. Computer anal-vsis Nucleotide and amino acid sequences were compared using a dot matrix program [lo] with a minimum match of 12 nucleotides or 4 amino acids and a minimum homology of 70%.

2 ~ 332 Residue number from B -end I I I I I 8 I I I I, I I I I I ~ 1 clones I r I H e!! I I Fig. I, Schematic representation qf the strategy used,for sequencing Q-CMV RNA 1. Primers used for RNA sequencing wcre either synthetic oligonucleotides (41, or cloned DNA fragments (m). (0) Partial enzymatic sequencing of the 5 terminus. The RNA sequence at the 3 terkinus, identified by *. has been published [I 31 A weight matrix scanning program (I. B. Dodd, niversity of Adelaide, unpublished) similar to that of Staden [Ill was used to search for 5 conserved regions in the nucleotide sequences of tripartite and other viruses. The weight matrix table was constructed from sequences of 12 nucleotides in the 5 non-coding regions of CMV and BMV RNAs 1 and 2 (two boxed regions in Fig. 4A and B and the single residue in between). The program calculates a mathematical score for the degree of match between each position of the viral sequence and the sequence in the table. Those matches with a high score are clearly evident as homologous regions. To enable estimation of statistical significance of matches, computer-generated random sequences of the same base composition as viral RNAs were scanned with the same weight matrix table to determine the number of random matches expected with scores equal to those observed in viral sequences. Hydrophobicity plots of the translation products wcre obtained by a program (A. H. Reisner, unpublished) using the hydrophobicity values of Nozakii and Tanford [12]. The value plotted at each point corresponds to the average hydrophobicity for the seventeen residues centred at that position. RESLTS Nucleotide sequencing qf Q-CMV RNA I The nucleotide sequence of CMV RNA 1 was determined by a combination of the following methods (Fig. 1): (a) RNA dideoxy-chain-termination sequencing using reverse transcriptase and synthetic oligonucleotide primers or primers prepared from MI 3 clones, (b) dideoxy-chain-termination sequencing of MI3 clones prepared from CMV RNA 1 cdna, and (c) partial enzymatic cleavage of decapped RNA, 5 -labelled with 32P. The same methods were used for the sequence determination of Q-CMV RNA 2 and are described in detail in [2]. The complete sequence of the 3389 nucleotides of CMV RNA 1 (M, 1.15 x lo6) is presented in Fig. 2. The single long open-reuding.frarne of CA4V RNA I CMV RNA 1 has only one long open-reading frame, which starts at the first AG 98 nucleotides from the 5 end and extends 2973 nucleotides before a GA stop codon. This reading frame encodes a translation product of 991 amino acids (M, ) and is considered to correspond to the in vitro translation product of CMV RNA I, which has a M, of as estimated by gel electrophoresis [S]. In the same reading frame there are 30 other in-phase AG codons, 22 of Table 1. Codon usage in the long open-reading,frarne of Q-CMV RNA 1 (nucleotide residues ) Stop codons are indicated with an asterisk. Figures give the numbcr of times each codon is used First Second position Third. position - position C A G Phe 28 Ser 29 Tyr 15 Cys 17 Phe22 Ser 19 Tyr 14 Cys 10 C Leu 16 Scr 14 *0 * I A Leu 20 Ser 7 * O Trp 12 G C Leu 17 Leu 13 Leu 9 Leu 14 A He 27 Ile 13 Ile 10 Met 31 G Val 33 Val 15 Val 9 Val 19 Pro 13 Pro 8 Pro 5 Pro I Thr 28 Thr 15 Thr 13 Thr 12 Ala 39 Ala 16 Ala 9 Ala 8 His 11 His 14 Gln 18 Gln 11 Asn 18 Asn 11 Lys 34 Lys 30 Asp 38 Asp 26 Glu 30 Glu 23 Arg 20 Arg 7 Arg 6 Arg 4 Ser 8 Ser 3 Arg 10 Arg 5 Gly 25 Gly 12 Gly 13 Gly 8 c A CJ C A G C A G which can potentially initiate translation products longer than 536 amino acid residues. The next largest AG-initiated reading frame on the positive strand is 210 nucleotides long while the largest open-reading frame on the negative strand is 20 I nucleotides long. Table 1 summarizes the codon usage in the long opcnreading frame of RNA 1. The distribution of degeneratc codons in RNA 1 is evidently non-random. All quartet codonj (Ala, Gly, Pro, Thr, Val [14]) show a preference for in thc: third position. Overall 44.9% of these codons end with and the remainder with the other three nucleotides. Th: pyrimidine-ending duets (Phe, Cys, Tyr, Asp, His, Asn), except for His, also have a preference for in the third position and the C-G dinucleotide occurs 0.56 times the frequency expected at random. This pattern of codon usage in the reading frame of RNA 1 is very similar to that of CMV RNA 2 [2] and conforms with codon usage in eukaryotic mrnas [14]. The sizes of the translation products of CMV RNA 1 (Mr ) and RNA 2 (M, [2]), as deduced from the nucleotide sequences, are close to those of BMV RNA 1 (Mr ) and RNA 2 (M, 94000) [15], although they differ from the sizes of the in vitro translation products of CMV RNA 1 (M, 95000) and RNA 2 (M, ) as estimated by gel electrophoresis 181. The reason for the discrepancics

3 m7g GAA CAAGAGCGA CGGCAACC CCGCCCC CGAAAAC ACCCGAA CCC CGA CAGAA AACC met ala thr set ser phe asn ile asn glu leu Val ala ser his gly asp 98 AG GCA ACG CC CA C AAC AC AA GAA CG GA GCC CC CAC GGC GA his glu gln leu glu glu gln leu gln his gln arg arg gly leu lys Val 188 CA GAA CAG CG GAG GAG CIA C CAA CA CAA CG AGA GGC C AAG GC ile arg thr arg tyr gly gly lys tyr asp leu his leu ala gln gln glu leu ala pro his gly leu ala gly ala leu arg leu cys 278 A CGG AC CGG A GG GG AAG AC GAC CC CA CC GCC CAG CAG GAA A GC CCC CA GGC CC GC GG GCC CC CGC G G glu thr leu asp cys leu asp phe phe pro arg ser gly leu arg gln asp leu Val leu asp phe gly gly ser trp val thr his tyr 368 GAA AC CC GA G CA GAC C CC CG CA GG CG CGG CAG GAC CC GC A GA C GGA GGA AG GG GC ACA CAC A leu arg gly his asn Val his cys cys ser pro cys leu gly ile arg asp lys met arg his thr glu arg leu met set met arg lys 458 C CGC GGA CAC AAC GA CA GC GC CC CCA G G GG A CG GAC AAG AG CG CAC AC GAA AGG G AG AGC AG CGC AAG Val ile leu asn asp pro gln gln phe asp gly arg gln pro asp phe cys 548 GC A A AAC GA CCA CAA CAG GA GGC CGC CAG CCG GAC GC lys gly leu leu ala thr ala leu Val asp lys thr ala AAA GGA CA C GCG ACA GCC CC G GA AAG ACA GC tyr ile arg asn Val leu asp val lys asp ser glu val AC AC CG AA G G GA GA AAG GAC CC GAA GC thr 1YS ser AC AAG ucu ala ala g1u CYS 1YS GCG GC GAA GC AAA va 1 gln ala his G CAA GCC CAC ala ile ser ile his gly gly tyr asp met gly phe arg gly leu cys glu ala met asn ala his gly thr thr ile leu lys gly thr 638 GC A C AA CAC GGA GGA A GA AG GGC AGA GG A G GAG GCA AG AAC GC CAC GGA AC ACG A G AAA GGG ACG met met phe asp gly ala met met phe asp asp gln gly phe ile pro glu leu 728 AG AG C GAC GG GC AG AG GAC GAC CAA GGC A CC GAR cuu 1YS AAA CYS gln tre arg 1YS ile 1YS ser G CAG GG CGA AAA A AAA ucc ala ehe GCC uuu set ucu glu glu glu asp ala thr cys ser ala ala lys leu asn ser ser Val phe ser arg val arg asn gly lys thr leu ile ala phe asp 818 GAG GAG GAG GA GCC AC G CA GCA GC AAA CC AA C AG G CA CGC GG CGA AA GGG AAA ACC A AC GCA GAC 908 ehe uuc Val g1u 9 l se r G GAG GAA ucc thr met ser ACG AG ucu tyr val his asp tre asp asn ile 1YS ser A G CAC GA GG GA AA AA AAA CG met thr asp gln thr tyr ser RG ACA GA CAA ACG AC ucu phe uuc asn gly met AA GG AG thr tyr gly ile glu arg cys Val ile tyr ala gly Val met thf tyr lys ile val gly Val pro gly met cys pro pro glu leu ile 998 ACC A GGA A GAG CG G G A AC GC GGC GG AG AC AC AAG A G GGC GG CC GG AG G CCG CCC GAA CC A 1088 arg his CYS ile tw phe CGA CA G AC GG uuc ccc Pro see ucu met 1YS asp tyr AG AAG GAC A va 1 gly leu lys i le G GG C AAG A pro ala ser asp asp leu Val lys trp lys thr val arg CCC GCG C GA GAC G G AAA GG AAA ACA GC CG ile leu leu set thr leu arg glu thr glu glu ile ala met arg cys tyr asn asp lys lys asn trp met asp leu phe lys ile ile 1178 A A CG CA ACA A CG GAG AC GAA GAA AA GC AG CG G A AAC GAC AAG AAG AA GG AG GA CA C AAG A A leu gly Val leu set ser lys ser set thr ile val ile asn gly met ser met gln ser gly glu arg ile asp leu asn asp tyr his 1268 CC GG G A CA CG AAG CC C-C ACG AC G AC AA GG RG C AG CAA CC GG GAA CG A GA CC AA GA A CA tyr ile gly phe ala ile leu leu his thr lys leu lys tyr glu gln leu gly lys met tyr asp met trp asn ala ser phe ile trp 1358 AC AC GG GC A C CC CAC ACG AAA A AAA AC GAA CAA C GGA AAA AG AC GA AG GG AA GC CC C A GG lys trp phe ala ser met ser acg pro phe arg val phe phe ser thr val val lys thr leu phe pro thr leu arg pro arg glu glu 1448 AAG GG GCG C AG C AGA CCA C CG G C CC AC G G AAG AC G CCG AC G AGA CCG CGC GAG GAA lys glu phe leu Val lys leu ser thr phe val thr phe asn glu glu cys ser 1538 AAG GAG G GC AAA C CC AC C GC ACC AAC GAG GAG GC ucu asp 9lY gly 1YS glu trp asp val ile ser ser GAC GGA GGG AAA GAA GG GAC GG AA CA CA ala ala phe Val ala thr gln ala Val ala asp gly thr ile leu ala glu glu lys ala lys lys leu ala asp arg leu ala val pro 1628 GCG GC C GA GCC AC CAG GC G GCA GA GGC AC A CG GCC GAG GAG AAA GC AAG AAA A GC GA CG CG GCC GG CC 1718 val 9l 91 va 1 thr ala i le pro G GAA GAA G AC GC A ccu 91 val ser GAG GG ucu Pro ccu thr Pro ACA ccu Val asp gln 9lY thr a la CYS 3lY leu 9lu thl G GA CAG GGC AC GC G GGA CG GAA ACA thr se r ACA CG 3l leu GAA C G asp ser leu set ala gln thr arg ser pro ile ala arg ile ala glu arg ala thr ala met leu glu tyr set ala tyr glu lys gln 1808 GA C CG C GCC CAA ACA CG CC CCC AC GCA CGG AC GC GAA AGA GCG ACC GC AG C GAA A CA GC A GAG AAA CAA 1898 leu his asp thr thr val ser asn leu G CAC GA ACC ACC G CA AA cuu gln arg ile tfe CYS met ala 9lY 9lY asp asn 1YS arg asn set CAA CGA A GG GC AG GCA GG GGC GAC AAC AAG AGA AAC ucu leu glu set asn leu 1YS A GAG AG AA G AAA phe V a l phe asp thr tyr phe ser Val asp ala leu Val asn Val his phe 1988 GG GAC AC A C G GAC GCC CA GG AA G CAC val tyr set val gly tyr asn glu lys gly leu gly pro lys leu asp set 2078 GA AC C G GG A AA GAG AAG GG C GG CC AAA C GA AG pro thr gly arg trp met his pro Val pro glu gly Val CC AC GGG AGA GG AG CAC CC GG CC GAG GGC GG glu leu tyr ile Val asn gly asp cys Val ile ser asn GAG G AC AA G AA GG GA G GG A CG AAC set his asp leu phe ser ile thr lys ser leu leu ala pro thr gly thr ile set gln Val asp gly Val ala gly cys gly lys thr 2168 AG CA GA G CA AC AC AAA C G A GC CCC ACC GGA ACC AC AGC CAA GC GA GG GA GC GGG GC GGG AAA ACC thr ala ile lys ser met phe asn pro ser thr asp ile ile Val thr ala asn lys lys ser ala gln asp val arg tyr ala leu phe 2258 AC GC AA AAA CC AG AA CC CC ACA GA AA A GC ACA GCC AAC AAG AAA C GC CAA GA GG CG A GCG CG 1YS ser 2348 AAA ucu thr asp ser AC GAC ucc 1YS 91 a la CYS ala phe AAA GAA GC G GC uuu va 1 G arg AGG thr a la asp ser ile leu leu asn asp cys pro thr Val ser arg val leu ACC GC GA ucc AA A CC AA GA GC CC ACG GG C CGA GG C Val asp glu Val Val leu leu his phe gly gln leu cys ala Val met ser lys leu his ala Val arg ala leu cys phe gly asp ser 2438 GG GA GAA G G G G CAC GG CAG G GC GC GC AG CG AAA C CA GC GC AGA GC G G C GGA GAC CC l gln ile ala GAA CAG AA GCC ser ucc ser ucu arg asp ala ser phe CG GAC GC CG uuc asp met arg phe GAC AG CG uuc ser ucu lys leu ile pro asp glu thr ser asp ala asp thr thr AAG C A CCG GA GAG ACC AG GA GCG GAC ACA AC phe arg ser pro gln asp Val val pro leu val arg leu met ala thr lys ala leu pro lys gly thr arq thr lys tyr ser asp gly 2618 C CG AGC CCA CAA GA GA GA CCA C GG CG G AG GC ACG AAG GC CA CCG AAA GGG ACC CG ACG AAA AC CA GA GG 2708 ala gln ser GCC CAR ucu 1YS Val arq 1YS ser AAA GG AGG AAG ucu va 1 G thr AC ser arg ala val ala ser va 1 ser CG CG GC G GC AG GA ucu leu CA met thr gln ala asp lys ala ser leu ile thr arg ala lys glu leu asn 2798 AG ACG CAA GC GA AAA GCC CA CA AC AC AGA GC AAG GAA CG AAC his glu ser gln gly ile ser glu asp his Val thr leu Val arg leu lys 2888 CA GAA C CAA GG A CA GAA GA CA GG ACC CG G AGG G AAC va 1 9 l leu asp Pro G GAA CG GAC ccc thr arg phe ACC AGA uuc tyr ile thr A A ACG leu pro lys ala phe tyr thr asp arq ile lys thr val CA CCC AAG GCA C AC AC GA AGG AA AAG AC GC ser thr lys cys asp leu phe lys lys phe ser tyr cys AG AC AAA G GAC CG C AAG AAA C AC GC leu Val ala Val thr arg his lys Val thr phe arg tyr glu tyr cys gly Val leu gly gly asp leu rle ala asn cys ile pro leu 2978 A G GCA G AC CGA CAC AAG GC ACC C CGC A GAG AC G GG GG A GG GGG GAC CA AC GC AA G A CCG A val *** 3068 GC GA CGCGGA AGGCCGAAG ACGAAAC ACGCCC AGCGAG GCGAGGG AGGC AAACAC GAAGCGC AAACCAGA GG 3175 GCGA ACGGGGC CACCAGC ACGGCAAAA GGCAGA GCCCCAAAGG CAGCCGACA CCACAAGG GCGAGC ACCCGAAA CACCCA GA 3283 CCG GAAGGGCC GGAGAAGC CGGCACGG AAACACGA AACCAAG RGGCGGGA CGCCGGG CCACAG GCCCAA AGGAGACCA 3389 Fig. 2. Nucleotide sequence oj Q-CMV RNA 1 and deduced amino acid sequence of its long open-reading frame. The stop codon of this reading frame is identified as ***

4 334 cnv RNA 1 m G CKV RNA 2 m G CKV RNA 1 CKV RNA 2 p 2? 3P 4p 5p PO GAL +-c*+++*c+++*rr++++~*******~**+r+++****g***~*~***~----* p VJ no? CCCCGACAG?.AAACC~GCAACGCCCACAACA G+**A**GACAC~~C*AG*CCGx*hx++*rA**-**~***CAC+C~** Fig. 3. Sequence Izomology in the 5 non-coding regions qjq-cmv RNAs 1 and 2. The sequence at the 5 end of CMV RNA 2 is aligned with the first 120 nucleotide residues of RNA 1. (*) Homology between RNAs 1 and 2. The initiating AG codons are underlined. Sequences between brackets are also conserved in RNAs 1 and 2 or BMV. ( -) Gaps included to maximize alignment 25 - CMV BMV CMV Satellite RNA - G,.G-C - G - C A - u G - C c :-C C - G -A- C - - A +G u - A - A m7g -GW)u u u A A I 47 A G - C - G A - c m7t I G 25-A-- A c -A L A G G - C G G - C G c-g c C A G-- G A C - G C - G u - A A C G G A G A G A A G G,A - u I G - 41 dg-g A (C)u G c u I 47 B Fig. 4. Proposed secondary structure nt the 5 end qf Q-CMV (A) and BMV (B) RNAs I and 2 I151 and of Q-CMV.satellite RNA (C) (61. The structures shown in (A) and (B) are for RNA 1 and arrows indicate changes from this sequencc in RNA 2. The residues in parentheses are present only in RNA 1. The boxed sequences are conserved in CMV and BMV and the residues in bold type in satellite RNA can base pair with CMV RNA 1. The outcr letters in (C) arc the sequence or RNA 1 as shown in (A). The arrows indicate 5 to 3 direction c between the calculated and the estimated sizes of the CMV translation products is unknown. Non-coding regions of CMV RNA I The 5 non-coding region of CMV RNA 1 starts with a cap structure [16] and contains 97 other nucleotide residues before the first AG codon. The sequence surrounding this initiation codon (CCACG) contains only two of the residues of the concensus sequence (CCAIGCCACG) near the translation initiation sites of eukaryotic RNAs [17]. The nucleotide sequences surrounding the initiating AG of CMV RNAs 1 and 2 are vesy similar, including 11 residues downstream from AG (98-100) of RNA 1 (Fig. 3). Lack of a purine at position -3 places CMV RNA 1 AG (98-100) in the group of initiation codons which rarely occur in eukaryotes [17]. The 5 non-coding regions of CMV RNA 3 (94 residues [3]) and RNA 2 (92 residues [2]) are of similar length to CMV RNA 1. Only the 5 non-coding region of RNA 2 has an AG followed by an in-phase stop codon prior to the initiating AG of its long open-reading frame [2]. The 5 non-coding region of CMV RNA 1 has significant homology with the 5 non-coding region of CMV RNA 2 (Fig. 3). In this region, extended homology is confined to the first 55 nucleotides and the residues sourrounding the initiating AG. A eonserved structure in the RNAs of CMV, BMV and CMV satellite A region of the homologous sequence near the 5 end of CMV RNAs 1 and 2 was found to be conserved in BMV RNAs 1 and 2 (bracketed in Fig. 3). To examine how fre- quently this conserved sequence occurred, the genomes of CMV and BMV and RNAs from five other plant viruses were searched using a weight matrix table [ll]. High scores, indicating close homologies, were not found in any of these sequences except for those regions of CMV and BMV RNAs bracketed in Fig. 3. The sequence was not found in a computer-generated random sequence of nucleotides (equivalent to 12.5 times the CMV genome) with the same base composition as CMV RNA 1. The conserved sequence was not present in any of the RNAs of AMV or TMV; however, a sequence with 83% homology (10 residues out of 12) was found at position 1100 in the intercistronic region of RNA 3 of both CMV [3] and BMV [18] (not shown). The region surrounding the conserved 5 sequence of CMV and BMV RNAs 1 and 2 can be folded into a hairpin loop structure (Fig. 4). In this secondary structure most of the conserved residues are located in the loop-out region of the hairpin. When the nucleotide sequences of the non-coding regions of the three CMV RNAs were compared with the sequence of sat-rna from Q-CMV [6] it was found that the conserved sequence of 12 nucleotides at the 5 end of CMV RNAs 1 and 2 plus six preceding residues could form stable base pairs with the 5 end of CMV sat-rna. This region of sat-rna can form a hairpin loop structure similar to that of CMV and BMV RNAs (Fig. 4C). The satellites of two other strains of CMV (D-CMV [19] and Y-CMV [20]) can also form such base pairs with the genomic RNAs 1 and 2 (not shown). A sequence similar to the conserved nucleotides of the 5 end of CMV RNAs is therefore present at the 3 end of complementary sat-rna. The CMV satellites did not show any extended nucleotide sequence homology with any other parts of the CMV genome.

5 A A AC A C A c c G- -G -A -A -G G- A C I1 - G I-c G- -AA -A G- A C G-C -A C-G G- - A G- A-G G- 11 C G G G G C C A C C C A C -G C -G G -C -A A - - G G -C A C c-e -Ac G-C ca G- A -G C-G G- G- C -G A- -GC Table 2. P~rcmtage honwlogy hetiwen CMV RNA I {runslution produet and the translation product.7 of BMV RNA I. AMV RNA 1 and thz TMV M, protein The N-terminal and C-terminal regions of AMV and TMV translation products are those sections aligned with CMV RNA 1 translation product in Fig. 7. The central region of the BMV translation product is amino acid residues in Fig. 7. AMV and TMV do not show homology in the central region and therefore percentage homology was not calculated (n.c.) for this region RNA Homology in region of translation product N-terminal central C-terminal BMVRNA I AMVRNA n.c TMV RNA 20.6 n.c Fig. 5. Proposed secondary structure in the 3 non-coding region qf CMV RNAs. The sequence shown is that of CMV RNA 1 numbered from the 3 end The 3 non-coding region of CMV RNA 1 is 316 nucleotide residues long (excluding the stop codon). The homology reported for the 3 termini of RNAs 1, 2 and 3 extends beyond 270 residues [13] to 307 residues from the 3 end of RNA 1 and stops abruptly at that point. The sequence of RNA 1 between residues 270 and 307 from the 3 end is identical to the corresponding sequence of RNA 2 and differs from RNA 3 by four nucleotide mismatches and two single-residue gaps. The homologous 3 -terminal region of all four CMV RNAs, corresponding to residues of RNA 1, can be folded into two very stable hairpin loop structures (Fig. 5) additional to the trna-like secondary structure proposed for the 3 termini of CMV RNAs [13]. The hairpins shown in Fig. 5 do not exist in BMV and AMV RNAs, whose 3 conserved sequences are only 219 and nucleotides long respectively [15, 21, 221. Apart from the homologous non-coding regions, no other extended nucleotide sequence homology could be detected between CMV RNAs 1, 2 and 3 by dot matrix analysis. Comparison of the nucleotide sequence of Q-CMV RNA I und its translation product with those qf other viral RNAs The nucleotide sequence of CMV RNA 1 was compared with the sequences of other tripartite viruses and of TMV. Results of comparisons by dot matrix analysis, presented in Fig. 6a, b and c, show the regions of homology as diagonal lines. Strong homology between CMV RNA 1 and BMV RNA 1 is clearly evident from Fig. 6 a, while AMV RNA 1 and TMV RNA show less homology with CMV RNA 1 (Fig. 6b and c respectively). When the translation products of these viral RNAs are compared by a similar dot matrix analysis, homology becomes more evident (Fig. 6d - 0. The translation products of BMV and AMV RNA 1 and the TMV M,-l26000 protein show decreasing levels of homology with the translation product of CMV RNA 1. Although homologies between CMV RNA 1 translation product and TMV M,-l26000 protein were not obvious by dot matrix analysis (Fig. 60, they were contained in blocks of amino acids conserved in all four translation products when they were aligned on the basis of homologies revealed by dot matrix analyses (Fig. 7). Since the homology between CMV RNA 1 and BMV RNA 1 translation products was strongest, their entire length could be aligned with an overall homology of 43.8%. The alignment of the RNA 1 translation product of CMV with the translation products of AMV RNA 1 and TMV Mr protein was only possible for their C-terminal and N-terminal regions (Fig. 7). The percentage amino acid homology between the translation products of CMV and the other viruses in these regions is summarized in Table 2. In contrast to RNA 1, the translation product of CMV RNA 2 is only homologous to the central regions of the corresponding translation products of BMV, AMV and the TMV MI readthrough protein [2]. The distribution of acidic and basic amino acid residues in the translation products of RNA 1 of the three tripartite viruses and TMV MI protein does not show any distinct pattern. This is in contrast to the RNA 2 translation products, which have acidic N termini and basic C termini PI. Hydrophobicity plots of the translation products of CMV RNA 1, BMV RNA 1, AMV RNA 1 and TMV M, protein showed many distinct hydrophobic and hydrophilic regions in each translation product. The hydrophobicity plots of CMV and BMV RNA 1 translation products could be aligned (Fig. 8) so that a number of peaks and troughs occur at similar positions in the two translation products. Plots of AMV and TMV translation products could not be aligned unambiguously. DISCSSION The nucleotide sequence of Q-CMV RNA 1 completes the primary structure of the Q-CMV genome and provides further evidence for a close relationship between all the tripartite viruses. The homology found between the leader sequences of CMV RNAs 1 and 2 is similar to that found between RNAs 1 and 2 of BMV [15]. In both viruses, RNA 3 lacks extended 5 sequence homology with RNAs 1 and 2. CMV and BMV RNAs 1 and 2, as well as three CMV sat- RNAs, were found to form a stable secondary structure close to their 5 ends (Fig. 4). This hairpin structure is analogous to that reported [25] for the 5 end of AMV RNAs 1 and 2. The presence of this structure in Q-CMV sat-rna [6] and AMV RNAs 1 and 2 [25] has been substantiated by sitespecific enzymatic cleavage. The presence of a hairpin loop at

6 336,5-5 3 RNA 1 CMV N -C TRANSLATION PRODCT 3 BMV RNA 1 AMV RNA 1 TMV RNA Fig. 6. Do[ matrices coiripariiig C W RNA I and its trimslution product wilh llze corresponding RNAs and translation products of BAN, A M and TMV. The complete sequences of the RNAs and translation products of CMV, BMV [I 51 and AMV [23] and only that region of the TMV sequence [24] encoding the M, protein (nucleotide residues ) were used for comparisons. Lines plotted in the matrix correspond to matches of at least 12 nucleotides (a-c) or 4 amino acids (d-9 with a minimum homology of 70%. The 5 to 3 direction for the RNAs is shown in (a) and the N-terminal to C-terminal direction for the proteins is shown in (d) the 5 end of RNAs 1 and 2 therefore appears as another common fcature of the tripartite viruses. It is interesting that the conserved nucleotide sequence at the 5 end of CMV and BMV RNAs 1 and 2 and CMV sat- RNA (Fig. 4) is essentially contained in the looped-out region of the hairpin structure, which implies that both the secondary structure and conserved primary sequence have functional significance. It is also intriguing that the conserved 5 sequence appears in CMV sat-rna in the complementary sense, which suggests the possibility of its binding to the conserved sequence at the 5 end of RNAs 1 and 2 for regulating viral and sat-rna replication. Such interaction is consistent with the finding that the efficiency of CMV sat-rna replication depends on the strain of helper virus used [26]. In fact, experi- ments with some CMV strains and their pseudorecombinants indicate that the rate of CMV replication is controlled by CMV RNA 1 and/or 2 [4]. CMV sat-rna decreases total viral yield and proportionally decreases the level of synthesis of viral RNAs 1 and 2 [26, 271. Plant tripartite viruses are evolutionally related to sindbis virus, an animal alpha virus [9]. Sindbis viral RNA has a secondary structure at its 5 end. which has been suggested as the recognition site on the minus strand for RNA replicase [28]. In BMV the viral replicase binds to a region of the trnalike structure at the 3 end of plus-strand RNA [29, 301. The 5 conserved structure reported here, which is also present at the 3 end of negative-strand tripartite RNAs, may have a similar function in the synthesis of positive strand RNA. A

7 w.i CMV 1 BMV 1 AMV 19 TMV 26 MATSSFN I NELVASH-GDKGLLATALVDK+AHEQLEEQLiHQR---RGLbY I RNVLDVKDSEV I RTRYGGKYDLHLAQQELAPHGLAGALRLCETLDCLDFFPRSGLRQDLVLDFGGSW -MS**IDLLK*I*EK-*ADSQS*QDI**NQVAQ*LSA*IEYAK---*SK*INV**K*SIEEADAF*D****AF**N*T**YH***S******VA*HY****S**P----E*P*I****** SHA*IQEMLRR*VEKQAADDTT*IGK*FSE*-GRAYA*DALPSDKGEV**ISFS--**ATQQNIL*ANF-PGRRTVFSNSSSSS*CF*A*H**L**DFVYRC*GN---TV*SII*L**NF AYTQTATTSA*LDTVR*NNS*VNDLAKRRLYDTAV**FNARD*---*P-**NFSK*IS-EEQTL*A**AYPEFQITFYNTQN*V*S***G**SL*-*EY*MMQIP---YGS*TY*I**NF CMV 117 BMV 112 AMV 132 TMV 113 CMV 218 BMV 207 AMV 227 TMV 231 CMV 329 BMV 299 CMV 449 BMV 419 CMV 560 BMV 538 CMV 680 BMV 652 AMV 803 TMV 800 CMV 790 BMV 763 AMV 917 TMV 914 FDGAMMFDDQGF-IPELKCQWRKIKSAFSEEEDATCSAAKLNSSVFSRVRNGKTLIAFDFVEESTMSYVHDWDNIKSFMTDQTYSFNG---MTYGIERCVIYAGVMTYKIVGVPGMCPPE *****L*iRE*i-L+L*+iHIQRDG"GAD, V*K***EN***L**I*G*QDLG**F*ESVHCID*---T**LL**E~LKCNI*****IATNLR**R* V*AD*LIHNE*E-**NFNVR*EIDRKKD L*H***ID*PNLG*S*RFSLL*HYL---**NAVDLGHAA*R***KQDFG***VIDLTYSL*FV*KM *SENLLLE*SYVNLD*INACFSRDGD KLT*S*AS***LN*C*SYS** PASDDLVKWK~VR I LLSTLRETEE I AMRCYNDKKNWMDLFKI I LGVLSSK~T I v I NGMSMQSGER I DLNDYHY I GFAI L~HTKLKYEQL~KMYDMWNAS~ L I RHC I WFP~MKDYVGLKI TL***V**EDISK***VS**EDWS*NR**C**VAKT*V**V****F**FKES*E*TENM*AVASI**A****VI***QAIMA***L*IE***LVA**LT*NLYQ***K*TALR*GMEWKG I WKWFASMS~PFRVFFSTV~KTLFPTLRPREEKEFLVKLSTFVTFNEEC~FDGGKEWDV i SSAAFVATQAV---ADGT I i AEEKAKKLADRLAVPVEEVTA I PEV----L-SPTPVDQGT WCHH*KTRFWWGGDSSRAK*GW*R-**AS*FPLLR*DSYADSFK*LTRL*NVEEF*,Q*SVPISRLRTFWTEEDLF*RLEHEVQT**TKRSKKKAK*PPAAE**QEEFHDAPESSSPESVS ACGLETETSELDSLSAQTRSPIARIAERATAMLEYSAYEKQLHDTTVSNLQRIWCMAGGDNKRNSLESNLKFVFDTYFSVDALVNVHFPTGRWMHPVPEGVVYSVGYNEKGLGPKLDSEL DDVKPVTDVVP*AEVSVEVPTDP*GIS*HG**K*FVR*C*R**NNSE***RHL*DIS**RGS----*IAN*SI*E**HRI*DM****LAN*N*LY*--KKYD*T*****H*****HAD*T YIVNGDCVISNSHD--LFSITKSLLAPTGTISQVDGVAGCGKTTAIKSMFNPST---DIIVTANKKSAQDVRYALFKSTDSKEACAF-VRTADSILLNDCPTVS---RVL-VDEVVLLHF ***DKT*AC*+LR+--IAEASAKVSVx*CD,+nx*w*+++*******DA*RMGE---*L*****R***E***M***PD*YNSKVALDV******AIMHGV*SCH---*L*-***AG***Y VF*DQS**FASAEAIIPSLEKALG*EAHFSVTI***********N**QIARS*GRDV*L*L*S*RS**DELKETIDC*PLT*LHYI---**C**Y*MSASAVKA---QR*IF**CF*Q*A SV*YS*MAKLRTLR--RLLRNGEPHVSSAKVVL****P*****KE*L*RV*FDE---*L*LVPG*QA*EMI*RRA-N*SGIIV*TKDN*K*V**FMM*FGKSTRCQFKR*FI**GLM**T GQLCAVMSKiHAVRALCFGDSEQ I AFSSRDASFDMRFS-IKL IPDETSDADTTFRSPQD;VPLVRLMATKALPKGTRTK;-SDGAQSKV~KSVTS---R~VASVSLVEL~PTRFY I TMTa ***LV*AALSKCSQV*A***T***S*K****G*KLLHG--N*QY*RRDVVHK*Y*C****--IAAVNLL*RKCGNRD***Q*WTSE***SR*L*K---*RIT*GLQ*TI**N*T*L**** *LVY*AATLAGCSEVIG***T***P*V**NP**VF*HH--**TG-KVERKLI*W***A*A *YC*E*YFYKNKKPVKTN*R*LR*IEV---VPIN*PVS**RNTNAL*LCH** *CVNFLVAMSLCEI*YVY**TQ**PYIN*VSG*PYPAHFA**EV**VETRR**L*C*A** *HY*NRRYEGFV---MST*S*K***SQEMVGGA*VINPISKPLHGKIL*F** CMV 904 BHV 876 AMV 1023 TMV 1023 Fig.1. Alignment of the translation product of CMV RNA I with the translation products of BMV RNA I and TMV M, protein. Amino acid homologies; (----) gaps included to maximize alignment. The entire amino acid sequences of CMV and BMV and the terminal regions of AMV and TMV are shown. The numbers refer to the positions of amino acid residues from the N terminus

8 In CMV-infected plant extracts, an M,-l00000 proteir copurifies with the extensively purified virus-induced RNA., dependent RNA polymerase activity [37,38]. Peptide maps 01 this M, protein and the CMV RNA 1 translatior product have bcen compared [8] and shown to contain some fragments of similar size, although other fragments were no1 common. These results were interpreted to indicate that the M, protein is not the translation product of CM\/ RNA 1 [8]. In view of the strong amino acid homology between the translation products of RNA 1 of CMV and BMV this needs to be re-examined. Sequencing of at least part 01 the CMV-induced Mr-lOOOOO protein will provide a definitive answer as to whether or not it is virus-coded. N 4 C Fig. 8. ~~ydrophohicit~i plots c?f transl~ition products of CMV und BMV RNA 1. Each division of the scale corresponds to 50 residues. See Materials and Methods for furthcr details similar structure is also present in the negative strand of sat- RNA, although its nucleotide sequence is complementary to that of the viral RNAs. It is worth noting that sequences present in the intercistronic region of CMV and BMV RNA 3, which are homologous to the 12 conserved residues at the 5 end of RNAs 1 and 2, are not part of a secondary structure similar to that in RNAs 1 and 2. In BMV RNA 3 this homologous sequence, starting at residue 1100, is outside the proposed site for internal initiation of RNA 4 synthesis [31]. The striking amino acid sequence homology of the translation products of CMV RNAs 1, 2 [2] and 3 [32], with the translation products of the corresponding BMV RNAs, establishes a close evolutionary relationship between these viruses. Recently the amino acid sequences of translation products of a number of DNA and RNA viruses have been compared [2, 9, 331, but none have shown sequence homologies to the extent found between CMV and BMV. Not only do the two viruses encode similar proteins, but their RNAs also show structural relationships within the 3 noncoding regions [34] and the 5 non-coding regions (Fig. 3). Moreover, the distributions of hydrophobic and hydrophilic residues in the translation products of CMV RNAs 1 and 2 are very similar to BMV RNA 1 and 2 translation products. Conserved sequences in the different viruses imply functional similarities. CMV and BMV would therefore be expected to have the strongest functional similarities. Whether any of the gene functions of either virus could be uscd by the other virus is a matter of interest. As yet, the functions of the non-structural genes of CMV are unknown. In BMV the profile of tryptic peptides of the RNA 1 translation product is very similar to that of the virus-induced M, protein found in particulate RNA polymerase preparations from infected plants. [35]. These results support data obtained from infectivity studies with separate BMV RNAs [36] and imply that RNA 1 encodes a component of the viral RNA replicase. We thank Dr Alex Rcisner (Commonwealth Scientific and Industrial Research Organization, Division of Molecular Biology, Sydney) for dot matrix analyses, hydrophobicity plots and for his gencrous advice, Dr Allan Gould for construction of CMV clones, Dr Derck Skingle, Stephen Rogers and Bruce May for the synthetic oligonucleotides, Dr R. I. B. Francki for glasshouse facilities and Jennifer Cassady and Sharon Freund for assistance. This work was supported by the Australian Research Grants Scheme and by a Commonwealth Government grant to the Adelaide niversity Centre for Genc Technology in the Departincnt of Biochemistry. REFERENCES 1. Peden, K. W. C. & Symons, R. H. (1973) Virologj~53, Rezaian, M. A., Williams, R. H. V.. Gordon, K. H. J., Gould A. K. & Symons, R. H. (1984) Eur. J. Bioclletn. 143, Could, A. R. &Symons, R. H. (1982) Eur. J. Bioclicm Francki, R. 1. B. (1985) Annu. Rev. Microhiol., in thc press. 5. Murant, A. F. & Mayo, M. A. (1982) Annu. Rev. P~iyropcrthol 20, Gordon, K. H. J. & Symons, R. H. (1983) Nucleic A(,ids Res Could, A. R., Palukaitis, P., Symons, R. H. & Mossop. D. W. (1978) Virology 84, Gordon, K. H..I., Gill, D. S. & Symons, R. H. (1982) Virologl 123, Haseloff, J., Goelet, P., Zimmern, D.. Ahlquist, P., Dasgupta, R. & Kaesbcrg, P. (1984) Proc. Nut1 Acud. Sci. SA 81, Reisner, A. H. & Bucholtz, C. A. (1983) EMBO J Staden, R. (1984) Nzicleic Acids Res. 12, Nozakii, Y. & Tanford, C. (1971) J. Biol. Chern Symons, R. H. (1979) Nucleic Acids Res. 7, , 14. Grantham, R., Gautier, C., Gouy, M., Jacobzone, M. & Mcrcier. R. (1981) Nucleic Acids Res. 9, r43-r Ahlquist, P., Dasgupta, R. & Kacsberg, P. (1984) J. Mu/. Bid. 172, Symons, R. H. (1975) Mnl. Biol. Rep. 2, Kozak. M. (1984) Nucleic Aci& Res. 12, Ahlquist, P., Luckow, V. & Kaesberg, P. (1981) J. Moi. Bid. 153, 23 ~ 19. Richards, K. E., Jonard, G., Jacquemond, M. & Lot, H. (1978) Virologj* 89, Hidaka, S., Ishikawa, K., Takanami. Y., Kubo, S. & Miura, K. (1984) FEBS Lf tt. 174, Koper-Zwarthoff, E. C., Brederode, F. Th., Walstra, P. & Bol. J. F. (1979) Nucleic Acids Res. 7, Gunn, M. R. & Symons, R. H. (1 980) FEBS Lett. 109, Cornelissen, B. J. C., Brederode, F. Th., Moormann. R. J. M. & Bol, J. F. (1983) Nucleic Acids Res Goelet, P., Lornonossoff, G. P.. Butler, P. J. G., Akani. M. E., Gait, M. J. & Karn, J. (1982) Proc. Nut1 Acad. Sci. SA 7Y, Ravelonandro, M., Godefroy-Colburn, T. & Piiick. L. (1983) Nucleic Acids Res. 11,

9 Mossop, D. W. & Francki, R. I. B. (1979) Virology 95, Kaper, J. M. & Tousignant, M. E. (1977) Viro1og.y 80, Strauss, E. G. & Strauss, J. H. (1983) Curr. Top. Microbiol. Immunol. 105, 1-98, 29. Drcher, T. W., Bujarski, J. J. & Hall, T. C. (1984) Nature (Lond.) 311, Ahlquist, P., Bujarski, J. J., Kaesberg, P. & Hall, T. C. (1984) Plant Mol. Biol. 3, Miller, W. A,, Drcher, T. W. &Hall, T. C. (1985) Nature (Lond.) 313, Murthy, M. R. N. (1983) J. Mol. Bid. 168, Argos, P., Kamer, G., Nicklin, M. J. H. & Wimnicr, E. (2984) Nucleic Acids Res. 12, Ahlquist, P., Dasgupta, R. & Kaesberg, P. (1981) Cell Bujarski, J..I., Hardy, S. F., Miller. W. A. & Hall, T. C. (1982) Virology 119, Kiberstis, P., Loesch-Fries, L. S. & Hall, T. C. (1981) Virology 33, Kumarasamy, R. & Symons, R. H. (1979) Virology 96, Gill, D. S., Kumarasamy, R. & Symons, R. H. (1981) Viro1og.y 113, 1-8.