Nucleotide sequence and genomic organization of apple chlorotic leaf spot closterovirus

VIROLOGY

179,

104-112

(1990)

Nucleotide Sequence and Genomic Organization of Apple Chlorotic Leaf Spot Closterovirus’ S. GERMAN, T. CANDRESSE,2 M. LANNEAU, J. C. HUET,* J. C. PERNOLLET,* ANDJ. DUNEZ Station de Pathologie

&g&ale. Centre de Recherches de Bordeaux, INRA, B.P. 8 I., 33883 Villenave D’Ornon Cedex, France; and l Laboratoire d’Etudes des Protkines, INRA. 78026 Versailles Cedex, France Received January 9, 1990; accepted June 78, 1990

The nucleotide sequence of Apple chlorotic leaf spot closterovirus (ACLSV) genomic RNA has been determined from cDNA clones. It is 7555 nucleotides in length excluding the 3’ terminal poly(A) tail and contains three putative open reading frames capable of encoding proteins of 216.5, 50, and 26 kDa. ACLSV RNA has untranslated regions of 151 and 190 nucleotides at its 5’ and 3’ termini, respectively. The 216.5-kDa ORF encodes a protein which contains the conserved “signature” sequences and has significant homology with the proteins suspected to be involved in viral RNA replication of members of the “Sindbis-like” supergroup of viruses. On the basis of distant homologies with viral movement proteins (M proteins), the BO-kDa ORF is suspected to encode a protein responsible for virus cell-to-cell spread. The 26-kDa ORF contains, in frame, a smaller 21.5-kDa ORF encoding the coat protein of ACLSV. These results show that ACLSV and probably at least the subgroup A of closteroviruses should be regarded as members of the “Sindbis-like” supergroup of RNA viruses. o 199OAcademic Press, IIIC.

INTRODUCTION

ACLSV has an elongated, very flexuous particle of 720 X 12 nm, encapsidating a single-stranded positive-sense genomic RNA of 2.5 X 1O6 Da (Lister and Bar Joseph, 198 1). The particles contain a single coat protein species with a molecular weight of about 22 kDa (Yoshikawa and Takahashi, 1988). In this report, we present the complete nucleotide sequence of the genomic RNA of the P863 strain of ACLSV. A comparison of the proteins encoded by ACLSV RNA with those of other RNA viruses demonstrates that ACLSV belongs to the “Sindbis-like” supergroup of viruses.

Apple chlorotic leaf spot closterovirus (ACLSV) is known to infect most fruit tree species. Although it is more or less symptomless in pome fruits, it is responsible for serious diseases in stone fruits, including peach dark green sunken mottle, false plum pox, and plum bark split (Dunez and Delbos, 1988). The economic importance of ACLSV is largely due to its worldwide distribution and to its capacity to induce severe graft incompatibilities in some Prunus combinations, causing important problems in nurseries. Its impact is, however, limited by the apparent absence of a natural vector for this virus. ACLSV belongs to the closterovirus group, which also contains citrus tristeza virus (CTV), beet yellows virus (BYV), carnation necrotic fleck virus (CNFV), and other viruses with similar flexuous, elongated particles (Bar Joseph and Murant, 1982). The group seems heterogeneous, however, and differences in particle and genome length, vector transmission, and cytopathic inclusions have led to its division into three subgroups: A (e.g., ACLSV), B (e.g.# BYV), and C (e.g., CTV) (Bar Joseph and Murant, 1982). At the moment, very little is known about the molecular organization of these viruses.

MATERIALS AND METHODS Virus and viral RNA preparation Apple chlorotic leaf spot virus (P863 strain, isolated from Prunus domestica and responsible for the bark split disease) was routinely propagated in the herbaceous host Chenopodium quinoa Wild. Virus and viral RNA were purified as previously described by Dunez et al. (1973). Cloning of ACLSV cDNA All molecular biology techniques were performed as described by Maniatis eta/. (1982) unless stated otherwise. Synthesis of the first strand of the cDNA was done with oligo(dT) as primer (pdT12-18, Pharmacia) and the reverse transcriptase of avian myeloblastosis

’ Sequence data from this article have been deposited with the EMBUGenBank Data Libraries under Accession No. M31714. ’ To whom reprint requests should be addressed.

0042-6822190

$3.00

Copyright 8 1990 by Academic Press, Inc. All rights of reproduction in any form resewed.

104

NUCLEOTIDE

SEQUENCE

virus (AMV RTase, Genofit). Second strand synthesis was accomplished with Escherichia co/i DNA Polymerase I in the presence of RNase H (Gubler and Hoffman, 1983). The ends of the double-stranded cDNA were blunted by Sl nuclease treatment followed by repair of the ends using T4 DNA polymerase. The cDNA molecules were then size selected by ultracentrifugation in lo409/osucrose gradients. The heaviest fragments were recovered by ethanol precipitation and were finally inserted into the Smal site of Smal-digested, dephosphorylated plasmid Bluescript M 13+ (Stratagene). The ligation mixture was used to transform competent E. co/i DH5a cells (BRL). Plasmid DNA was prepared from the transformants by the alkaline lysis method and characterized by restriction enzyme analysis. Subcloning of cDNA fragments Selected restriction fragments (EcoRI, Ml, HindIll, Hpall, Sau3Al) were subcloned in plasmid Bluescript (Stratagene). In addition, overlapping deletions in two large independent cDNA clones were produced, using random deoxyribonuclease I digestion (Lin et al., 1985). Determination of the nucleotide sequence Most of the sequence of ACLSV was determined by the dideoxynucleotide chain termination method of Sanger et a/. (1977, 1980) using the Sequenase kit (USB), thio-[a-35S]dATP and supercoiled plasmid DNA as the template. To determine the sequence of the 5’ terminal region, a synthetic oligonucleotide primer complementary to nucleotides 38-55 (5’ AGGCGTTACGTCAATCTG 3’) was labeled with [y-3’P]ATP by the action of T4 polynucleotide kinase, hybridized to genomic RNA, extended by the action of AMV RTase in the presence of deoxynucleotides, gel purified, and finally sequenced by the procedure of Maxam and Gilbert (1980). Amino acid sequence determination Preliminary analysis revealed that the amino terminal end of ACLSV coat protein was blocked and could not be sequenced by Edman degradation. In order to deduce some amino acid sequence information to allow unambiguous assignment of the coat protein gene to one of the ORFs detected on the genome, the following strategy was used. Purified ACLSV coat protein (in the form of purified virus particles) was treated under partial proteolysis conditions with Staphylococcus aureus V8 protease (Cleveland et al., 1977). The peptides ob-

OF ACLSV

105

tained were separated by denaturing electrophoresis in polyacrylamide-SDS-Tricine gels and transfered to lmmobilon membranes (Sallantin et al., 1989). After staining, the zone of the membrane corresponding to the major cleavage product under the treatment conditions used (molecular weight approx 13 kDa) was excised and subjected to automated Edman degradation (Hewick et al., 1981) using an Applied Biosystems (AB) 475A sequencer linked with an AB 120A Pth-amino acid analyzer. Computer sequence analysis Compilation and analysis of the nucleotide sequence was performed using the Cornell DNA sequence analysis package (Fristensky et a/., 1982) run on a Micra1 (Bull) microcomputer. Alignments of conserved amino acid “signature” sequences were obtained with the series of programs CLUSTAL (Higgins and Sharp, 1988). Dot-matrix analysis comparisons were performed using the Microgenie (Beckman) commercial package. Databank searches for homologies to ACLSV-encoded proteins were performed using the FASTA programs (Pearson and Lipman, 1988) running on the BIONET FASTA-MAIL facility. The statistical significance of the homologies found during databank searches was evaluated using the RDF program from the NIH sequence analysis package. RESULTS AND DISCUSSION Nucleotide sequence of ACLSV genomic RNA A cDNA bank to the genomic RNA of ACLSV was obtained and analysis of the clones allowed us to construct a restriction map of the genome. This information was in turn used to select cDNA inserts forming an overlapping set of recombinant DNA molecules covering approximately 99.5% of the ACLSV genome. With the exception of the 5’ terminal 37 nucleotides, each base was sequenced at least once on each strand of the cDNA. Although a single product was obtained during the determination of the 5’ end sequence by primer extension, we cannot guarantee that the reverse transcriptase reached the end of the viral RNA template and thus extra 5’ bases may have been missed during this analysis. No specific attempts were made to determine the presence of any 5’blocking groups. The complete nucleotide sequence of ACLSV genomic RNA is shown on Fig. 1, together with the predicted amino acid sequences of the three potential open reading frames which were detected (see below). The sequence is 7555 nucleotides in length, excluding the 3’poly(A) tail, and has a base composition of 31.5%

106

GERMAN

A, 17.7% C, 23.8% G, and 279/o U. The calculated molecular weight of the RNA is 2.58 X lo6 Da, in good agreement with the estimations obtained by denaturing gel electrophoresis (Lister and Bar Joseph, 198 1). The presence of a poly(A) sequence had recently been inferred from the behavior of the genomic RNA of a Japanese isolate of ACLSV upon chromatography on oligo(dT) cellulose (Yoshikawa and Takahashi, 1988). Recent results suggest that the genomic RNAs of two other closteroviruses, beet yellows virus (BYV, belonging to subgroup B) and citrus tristeza virus (CTV, subgroup C) probably do not possess such poly(A) tails (Karasev eta/., 1989; M. Bar Joseph, personal communication). Thus, there is at least one major structural difference between ACLSV genomic RNA and those of other known closteroviruses. In this respect, it would be interesting to know if heracleum latent virus, another member of the subgroup A, also possesses a 3’ poly(A) tail. At this moment, nothing is known about the 5’ end structure of the genomic RNA of ACLSV. However, experiments showing the positive effects of virazole (an antiviral compound acting as a cap analog) on the elimination of ACLSV from C. quinoa (Hansen, 1979) suggest that a cap structure might be present at the 5’ end of ACLSV RNA. Work is currently in progress to see if, as for BYV (Karasev et al., 1989) addition of a cap analog to an in vitro translation reaction inhibits the translation of ACLSV genomic RNA. Coding capacity of ACLSV RNA The sequence was searched for potential coding regions in all three reading frames of both strands. No ORF larger than 100 amino acids could be found in the minus orientation strand. Figure 2 shows the position in the genome of the three potential ORFs which have been identified on the plus strand. The genomic organization of ACLSV RNA is very compact since these three ORFs overlap each other so that there is no intergenie region and since only 341 of the 7555 nucleotides are in noncoding regions. Thus, regulatory sequences, required for instance as signals for replication, are either quite short or not restricted to noncoding regions. The AUG codon at nucleotide 152 (the number of the first nucleotide of the triplet is given) is the start codon

ET AL.

of a large open reading frame (ORF 1 in Fig. 2) which ends at the UGA stop codon at position 5804. This open reading frame codes potentially for a protein of 1884 amino acids with a calculated molecular weight of 216.5 kDa. This size is in reasonable agreement with the size of the biggest in vitro translation product of ACLSV RNA in a rabbit reticulocyte lysate system (approximately 190 kDa, Candresse et al., manuscript in preparation). The AUG codon at position 152 is not the first initiation codon encountered from the 5’ end but the other AUG codon (at position 83) is in a very poor context for translation initiation (Kozak, 1987; Lutcke et al., 1987) and is immediately followed in phase by a UAA stop codon at position 95. Thus it seems that, as for several other viruses, the first AUG codon from the 5’end is not the initiation site of the first ORF, initiation on the second AUG being most probably the result of “leaky” scanning by the ribosomes (Kozak, 1987). Partially overlapping with the 216.5-kDa ORF, the next ORF (ORF 2 in Fig. 2) starts at nucleotide 5718 and ends at nucleotide 7098, encoding potentially a 50-kDa protein. The AUG codon at position 5718 occurs in a favorable context for translation initiation, (G in position +4 and A in position -3) (Kozak, 1987; Luteke era/., 1987). The possible readthrough of the termination codon would yield a protein terminating at position 7215 with an estimated molecular weight of 55 kDa. However, we have no evidence for the occurrence of such a readthrough phenomenon, in vivo or in vitro. The third ORF (ORF 3) at the 3’ end of the genome starts from the AUG codon at position 6613 and terminates at the UAG stop codon at nucleotide 7363. It encodes a 250-residue polypeptide of calculated molecular weight 28.3 kDa. Unambiguous assignment of the coat protein coding capacity to this open reading frame was obtained through automated Edman degradation of a V8 protease partial cleavage product of purified ACLSV coat protein. The sequence VKSMEDQS corresponding to positions 120-l 28 of ORF 3 (amino acids underlined in ORF 3 in Fig. 1) was determined as the amino terminal sequence of one of the major cleavage products (estimated molecular weight 13 kDa). In the ORF 3 sequence deduced from nucleic acid data, this sequence is immediately preceded by a glutamic acid,

FIG. 1. Nucleotide sequence of the ACLSV genomic RNA. The predicted amino acid sequences of the three putative open reading frames is shown below the nucleotide sequence. Heterogeneities observed during the sequencing of independent cDNA molecules and the corresponding changes in coding properties are shown in lowercase letters. Underlined amino acids in the first ORF correspond to the two conserved “signature” sequences, “nucleotide binding site” (amino acids 1059 to 1066) and “polymerase active site” (amino acids 1639 to 1727). The sequence determined by Edman degradation of a partial V8 protease cleavage product of ACLSV coat protein is underlined in ORF3 (VKSMEDQS, positions 120-l 28).

NUCLEO~IDE SEQUENCE OF ACLSV

GERMAN

ET AL

FIG. l-Continued

NUCLEOTIDE 5’

1 I

216.5

2 1 ORF 1 kDa

3

4 I

*

5 I

6 I

n

7 ’

SEQUENCE

AAA

ORF2 AAA ORF3

FIG. 2. PosItions of the three putative open reading frames of the genomic RNA of ACLSV. The molecular weights of the proteins potentially encoded by each ORF are shown. The dashed line in the ORF3 corresponds to the second initiation codon of this open readIng frame, from which coat protein synthesis IS initiated. The asterisk indicates the position of the conserved “nucleotide binding site” and the dark square that of the “polymerase active site.”

in good agreement with the known specificity of S. aureus V8 protease. The calculated molecular weight of the product encoded by ORF 3 (28.3 kDa) is however larger than the 22-kDa size estimated by SDS-PAGE for the purified viral coat protein (Yoshikawa and Takahashi, 1988). Translation starting from the second initiation codon of this ORF (at position 6784) would yield a 193-residue polypeptide with a calculated molecular weight of 21.5 kDa, in better agreement with the experimentally determined value. In vitro translation of in vitro transcribed RNAs derived from partial cDNA clones indicates that the AUG at position 6784 can indeed be used as an initiation codon. The protein obtained in this case comigrates with purified ACLSV coat protein and is immunoprecipitated by anti-ACLSV immunoglobulins (Candresse et a/., manuscript in preparation). Therefore, it is likely that the AUG codon at position 6784 represents the true start codon of the open reading frame encoding the coat protein of ACLSV. This identification is confirmed by the observation that the amino acid sequence of the 21.5-kDa protein is devoid of ttyptophan, in agreement with previous amino acid analysis results showing that the capsid protein of ACLSV does not contain this amino acid (BarJoseph and Murant, 1982). The second AUG actually is in a better context for initiation of translation in plants than the first (Lutcke et al., 1987), having a G in position +4 and an A in position -3. Thus, the coat protein could be produced as a result of “leaky” scanning of ribosomes along the RNA molecule (Kozak, 1987). Alternatively, we cannot discount, at the moment, two other possible mechanisms for production of ACLSV coat protein, internal initiation directly at the second methionine codon of the ORF or production of a subgenomic RNA with a 5’ end located downstream of the first AUG, which would yield a messenger with the coat-initiating AUG as the first start codon. Work is currently in progress to search for possi-

109

OF ACLSV

ble subgenomic messenger RNAs and for the corresponding dsRNAs in ACLSV-infected plants. A similar situation (coat protein gene located in frame inside a larger open reading frame) has recently been described for potato virus S (Mackenzie et al., 1989), a carlavirus having substantial homology with ACLSV (see below). Sequence

heterogeneities

About 50% of the sequence was determined on at least two independent overlapping cDNA clones. Sequence heterogeneities were observed in the 3’ noncoding sequence at two positions (7372 and 7549) and at 14 positions in the coding sequence. These heterogeneities are indicated in Fig. 1, together with the amino acid changes they would induce in the coded proteins. The alteration at position 7026 in the 50-kDa ORF would change a glutamine into an ochre termination codon and would lead to a protein 23 amino acids shorter. This change was only observed in one of three independent cDNA molecules and we do not know whether it reflects true sequence heterogeneity in the starting RNA preparation or misincorporation during cloning. The other heterogeneities observed do not modify the size of the various ORFs and are either silent or induce conservative amino acid replacements. Homologies between ACLSV proteins those of other RNA viruses

and

The large 216.5-kDa ORF of ACLSV has all the marks of a replication-associated protein as it contains the two well-conserved “signature” sequences (“nucleotide binding site” and “polymerase active site”) found in such proteins (Goldbach, 1986, 1987). The position of these two sequences in the 216.5-kDa protein is shown in Fig. 2. The first “signature,” located between residues 3321 and 4041 in the central part of the protein (Figs. 1 and 2), contains the extensively conserved sequence GxxGxGKS/T thought to be associated with a nucleoside triphosphate (NTP) binding site or with a helicase activity (Hodgman, 1988; Gorbalenya eta/., 1988). Gorbalenya et a/. (1989) have recently shown that three families of proteins possessing the core “signature” can be distinguished on the basis of the presence of additional blocks of homology. As shown in Fig. 3A, the sequence of the ACLSV protein contains the six conserved blocks of homology typical of the family found in the “Sindbis-like” supergroup of viruses. The second “signature” is located in the C-terminal region of the 216.5-kDa protein and contains three main conserved blocks (DxxxxD, GxxxTxxxN, and GDD flanked by hydrophobic residues). This “signature” is

110

GERMAN ET AL.

A ACLS” QVX x3 Tmv-Type m-v Alnv TH” ""lgarc Sindbis

OPAp

126X 830 nrQ2 723

Y V

(lOaa> VmT t11aw so

DKVTL~

t13as>

YC

DR”Y?CR

tl9aa>

MI

KQVYA’$I


LNNv_

1298 957 1199 940 1087 1078 961

B ACLSV PVS TYHV type PVX x3 BMV AlMV Tt4V vulgare Sindbis

216~ 206K 165K 2a

2a 183K nsP4

t43aaN <41aa> <43aa> t43aa>

1636 * 1575

1239 462

<18aa> d8aa> tl7aa> <18aa> tl8am

527 1383 2269

1731 * 1669 1334 563 630 1484 2372

<20aa> tl8aa> t20aa>

FIG. 3. Alignment of the conserved “signature” sequences of ACLSV 2t6.5kDa protein with those of RNA viruses belonging to the “Sindbislike” supergroup. The viral protein in which the “signature” is found as well as the strain of the virus used are indicated. The numbering refers to the amino acids immediately outside of the sequence shown. Highly conserved amino acids are boxed; amino acids identical with the sequence of ACLSV are underlined. (A) Alignment of the six conserved blocks of homology constituting the “nucleotide binding site.” (B) Alignment of the three conserved blocks of homology constituting the “polymerase active site.”

characteristic of the proteins thought to be associated with the replication of RNA viruses and is regarded as the probable active and/or recognition site of an RNAdependent RNA poiymerase {Kamer and Argos, 7984). Again, as shown on Fig. 38, this region of the 216.5 kDa protein of ACLSV shows homology with the equivalent sequence of a number of members of the “Sindbis-like” supergroup of viruses. As shown in Fig. 4, in addition to the presence of these highly conserved “signatures,” the 2 16.5-kDa protein shows extensive overall homology with the putative replication-associated proteins of potexviruses, ~moviruses, and potato virus S (PVS, a carlavirus). The highest homology (49% overall amino acid identi~) is with the partial sequence of the PVS protein (tvlacken-

~~~

zie et a/., 1989). It is interesting to note the fact that the two “signatures” are located on a single protein is a property shared only, so far, by tymoviruses and potexviruses. Also, in common with ACLSV and PVS, potexviruses are the only plant viruses of the “Sindbis-like” supergroup to possess a 3’ poly(A) tail. Despite these structural homologies with PVS and potexviruses, ACLSV shows a quite different genomic organization in the 3’ terminal third of its genome: in these two groups, three small (usually about 25, 12, and 8 kDa) overlapping ORFs occupy the position equivalent to that of ACLSV 50-kDa ORF. There is good homology between the corresponding ORFs of PVS and potexviruses but no detectable homology between any of these small ORFs with the 50-kDa ORF of

~~~~

100 200 300

PVS

so0

PVX

-:-’ 1000

500

1000

1500

TYMV

FIG. 4. Dot matrix comparison of ACLSV 2 16.5-kDa protein with the partial sequence of the potato virus S replication-associated protein (PVS), the l65K protein of potato virus X strain X3 (PVX), and the 206K protein of turnip yellow mosaic virus type strain (TYMV). The numbering refers to the amino acid sequence. The comparison was performed using the Microgenie (Beckman) package with a window of 20 amino acids and a cut-off value of 50% homology.

NUCLEOTIDE

SEQUENCE

OF

111

ACLSV

This study provides complete nucleotide sequence information as well as preliminary data on the genomic organization of a member of the closterovirus group, apple chlorotic leaf spot virus. These results demonstrate that this virus belongs to the “Sindbis-like” supergroup of plant RNA viruses, its closest affinities being with a member of the carlaviruses. Given the wide variation in properties found within the seemingly heterogeneous closterovirus group, caution should however be used before extending these conclusions to other members of the group.

ACKNOWLEDGMENTS Databank searches were performed on the FASTA-MAIL facility of BIONET (NIH Grant P41 RR01 685). We are grateful to Kathryn MayoCandresse for help with the English of the manuscript.

REFERENCES ACLSV. In particular, the 50-kDa ORF protein does not contain the additional “nucleotide binding” “signature” contained in the 25-kDa protein of potexviruses and PVS. On the other hand, as shown in Fig. 5, the 50-kDa ORF of ACLSV shares limited homology with the protein encoded by the ORF I of cauliflower mosaic virus. This homology extends to the ORF I product of the other caulimoviruses and was first detected using the FASTA databank searching program (Pearson and Lipman, 1988). Since the homology is quite weak, its statistical significance was assessed using the RDF program. The score obtained is more than 9 standard deviations above the mean score obtained for matches between proteins of similar composition and size (generated by scrambling the sequence of the original proteins), suggesting that the homology found is highly significant. The ORF l-encoded protein of CaMV has tentatively been identified as mediating the cell to cell movement of CaMV (Albrecht et a/., 1988; Linstead et a/., 1988). The ORF I protein of CaMV has also been shown to share some homology with the 30-kDa protein of TMV, known unambiguously to have such a cell to cell movement function (Deom et al., 1987). However, since in each case the regions of homology are different, no significant homologies could be found between the 50kDa protein of ACLSV and the 30.kDa protein of TMV. Based on this limited homology with the ORF I of CaMV, we suggest that the 50-kDa protein of ACLSV might be involved in the cell to cell spread of the virus. No significant homologies could be found between ORF 3 of ACLSV and any of the entries in the most recent release of the PIR and Swiss-Prot databases.

ALERECHT, H., GELDREICH, A., MENISSIER DE MURCIA, J., KIRCHHERR, D., MESNARD, J. M., and LEBEURIER, G. (1988). Cauliflowermosaicvirus gene I product detected In a cell-wall enriched fraction. virology 163,503-508. BAR JOSEPH, M., and MURANT. A. F. (1982). Closterovirus group. “C.M.I./A.A.B. Description of Plant Viruses,” No. 260. CLEVELAND, D. W., FISCHER, S. G., KIRSCHNER. M. W., and LAEMMLI, U. K. (1977). Peptide mapping by limited proteolysls in sodium dodecyl sulfate and analysis by gel electrophoresis. /. Biol. Chem. 252,1102-1106. DEOM, C. M., OLIVER, M. J., and BEACHY, R. N. (1987). The 30 kilodalton gene product of Tobacco mosaic virus potentlates virus movement. Science 237, 389-394. DUNEZ, J., DELBOS, R., and BERTRANET, P. (1973). Le rdle de differents facteurs de stabilisation appllqu& 6 la purification du virus du Chlorotic Leaf spot. Ann. Phyroparhol. 5, 255.-265. DUNEZ, J., and DELBOS, R. (1988). Closterovlruses. In “European Handbook of Plant Diseases” (I. M. Smith, i. Dunez, R. A. Lelliott, D. H. PhIllIps, and S. A. Archer, Eds.) pp. 5-7. Blackwell, Oxford. FRISTENSKY, B., LIS, J., and Wu, R. (1982). Portable mlcrocomputer software for nucleotlde sequence analysis. Nucleic Acids Res. 10, 6451-6463. GOLDBACH, R. W. (1986). Molecular evolution of plant RNA viruses. Annu. Rev. Phyroparhol. 24, 289-310. GOLDBACH, R. W. (1987). Genome simllaritles between plant and anlmal RNA viruses. Mlcrobiol. Sci. 4, 197-202. GORBALENYA, A. E., KOONIN, E. V., DONCHENKO, A. P., and BLINOV, V. M. (1988). A conserved NTP-motif in putative hellcases. Nature 333,22. GORBALENYA, A. E., BLINOV, V. M.. DONCHENKO, A. P., and KOONIN, E. V. (1989). An NTP-binding motif is the most conserved sequence In a highly diverged monophyletlc group of proteins involved in posltlve strand RNA viral replication. 1. Mol. Evol. 28, 256-268. GUBLER, U., and HOFFMAN, B. J. (1983). A simple and very efficient method for generating cDNA libraries. Gene 25, 263-269. HANSEN, A. J. (1979). Inhibition of apple chlorotic leaf spot virus in Chenopodium quinoa by ribavlrin. Plant Dis. Rep. 63, 17-20. HEWICK, R. M., HUNKAPILLER, M. W., HOOD, L. E., and DREYER, W. I. (1981). A gas-liquid solld phase peptide and protein sequenator. J. B/o/. Chem. 256, 7990-7997.

112

GERMAN ET AL.

HIGGINS,D. G., and SHARP, P. M. (1988). CLUSTAL: A package for performing multiple sequence alignment on a microcomputer. Gene

73,237-244.

HODGMAN, T. C. (1988). A new superfamily of replicative proteins. Nature

(London)

333,23.

KAMER, G., and ARGOS, P. (1984). Primary structural comparison of RNA-dependent polymerases from plant, animal and bacterial viruses. Nucleic Acids Res. 12,7269-7282. KARASEV,A. V., AGRANOVSKY,A. A., ROGOV,V. V., MIROSHNICHENKO, N. A., DOLJA,V. V., and ATAEIEKOV,J. G. (1989). Virion RNA of beet yellows closterovirus: Cell-free translation and some properties. J. Gen.

Viral. 70,241-245.

KOZAK, M. (1987). Possible role of flanking nucleotides in recognition of the AUG initiation codon by eukaryotic ribosomes. Nucleic Acids Res. 15,8125-8148.

LISTER,R. M., and BARJOSEPH,M. (1981). Closteroviruses. ln “Handbook of Plant Virus Infections and Comparative Diagnosis (E. Kurstak Ed.), pp. 809-844. ElsevierINorth-Holland, Amsterdam. LIN, H. C., LEI, S. P., and WILCOX, G. (1985). An improved DNA sequencing strategy. Anal. Biochem. 147, 1 14-l 19. LINSTEAD, P. J., HILLS, G. J., PLASKITT, K. A., WILSON, I. G., HARKER, C. L., and MAULE, A. J. (1988). The subcellular location of the gene 1 product of cauliflower mosaic virus is consistent with a function associated with virus spread. J. Gen. Viral. 69, 1809- 18 18. LUTCKE, H. A., CHOW, K. C., MICKEL, F. S., Moss, K. A., KERN, H. F., and SCHEELE, G. A. (1987). Selection of AUG initiation codons differs in plants and animals. EMBO J. 6,43-48.

MACKENZIE,D. J.,TREMAINE,J. H., and STACE-SMITH, R. (1989). Organization and inter-viral homologies of the 3’-terminal portion of potato virus S RNA. J. Gen. Viral. 70, 1053-l 063. MANIATIS, T., FRITSH, E. F., and SAMBROOK, 1. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. MAXAM, A. M., and GILBERT,W. (1980). Sequencing end-labeled DNA with base specific chemical cleavages. ln “Methods in Enzymology,” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499-560. Academic Press, New York. PEARSON,W. R., and LIPMAN, D. J. (1988). Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85, 24442448. SALLANTIN, M., HUET, J. C., DEMARTEAU, C., and PERNOLLET,J. C. (1989). Reassessment of commercially available molecular weight standards for peptide sodium dodecyl sulfate polyacrylamide gel electrophoresis using electroblotting and microsequencing. Electrophoresis, in press. SANGER,F., NICKLEN,S., and COULSON,A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Narl. Acad. Sci. USA 74, 5463-5467. SANGER,F., COULSON,A. R., BARRELL,B. G., SMITH, A. J. H., and ROE, B. A. (1980). Cloning in single stranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 143, 161-t 78. YOSHIKAWA,N., and TAKAHASHI,T. (1988). Properties of RNAs and proteins of apple stem grooving and apple chlorotic leaf spot viruses. J. Gen. Virol. 69, 241-245.