- Email: [email protected]
Characterization of cucumber mosaic virus I. Molecular heterogeneity mapping of RNA 3 in eight CMV strains
VIROLOGY
166,495-502
(1988)
Characterization I. Molecular
of Cucumber
Heterogeneity
Mapping
Mosaic Virus
of RNA 3 in Eight CMV Strains
JUDITH OWEN AND PETER PALUKAITIS’ Department
of Plant Pathology, Received March
Cornell University,
Ithaca, New York 14853
16, 1988; accepted June 10, 1988
RNAs from 13 strains of cucumber mosaic virus (CMV) were divided into two groups on the basis of their ability to hybridize to cDNA of either Fny-CMV RNA or WL-CMV RNA. The extent of the cross-hybridization within one of these groups was analyzed by an RNA protection assay. A cDNA clone of RNA 3 of the Fny strain of CMV was placed in a transcription vector between bacterial promoters T3 and T7. Labeled, minus-sense RNA transcripts prepared from all or part of the cDNA to RNA 3 of Fny-CMV were annealed to the genomic RNA of each of a number of cucumoviruses and digested with RNases. The patterns of RNA fragments protected from digestion were specific for each CMV strain and revealed the extent and location of heterogeneity among the viruses as well as within the Fny-CMV natural population. This approach will allow the differences in host range and disease processes to be correlated with variations in genomic RNAs. 0 1988 Academic Press, Inc.
INTRODUCTION
was used to analyze RNA 3 of 12 strains of CMV. For 7 of the 8 strains which were in the same subgroup as Fny-CMV, heterogeneity maps of varying complexities could be constructed.
The genome of cucumber mosaic virus (CMV) consists of three single-stranded, positive-sense RNAs (Peden and Symons, 1973). By analogy to other tripartite viruses, RNAs 1 and 2 are believed to encode proteins involved in the replication of CMV (Nassuth and Bol, 1983; Kiberstis et al., 1981) while RNA 3 encodes a 30-35K protein that has been hypothesized to function in promoting the cell-to-cell movement of such viruses (Stussi-Garaud et al., 1987). RNA 3 also encodes the coat protein of CMV (Habili and Francki, 1974). The coat protein, however, is expressed only from a subgenomic RNA, RNA 4, which is colinear with the 3’half of RNA 3 (Schwinghamer and Symons, 1975; Gould and Symons, 1982). A large number of strains of CMV have been isolated which can be differentiated on the basis of biological rather than physical properties (Kaperand Watetvvorth, 1981). On the basis of serological typing (Devergne and Cardin, 1973) peptide mapping of the viral coat protein (Edwards and Gonsalves, 1983) and nucleic acid hybridization (Gonda and Symons, 1978; Piazzolla et al., 1979; Palukaitis, submitted for publication), strains of CMV can be differentiated into two groups (or subgroups). However, strains within a subgroup cannot be further differentiated by the above techniques. In order to further differentiate strains of CMV, to determine the extent of genetic variation between these strains, and to map the genetic variation, an RNA protection assay (Winter et al., 1985)
MATERIALS
AND METHODS
Reagents and enzymes .EcoRI, Seal, Pstl, mung bean nuclease, T4 DNA polymerase, and EcoRl linkers (dGGAAlTCC) were obtained from New England BioLabs. Sall, SamHI, HindIll, T7 polymerase, T4 DNA ligase, Escherichia co/i DNA polymerase I (Klenow), and T4 polynucleotide kinase were obtained from U.S. Biochemicals. Sacl was purchased from Boehringer-Mannheim. RNasin was obtained from Promega BioTec. Aval, DNase, and the DNA 1-kb ladder were obtained from Bethesda Research Laboratories. T3 and T7 polymerases and the Bluescribe plasmid vectors were purchased from Vector Cloning Systems (VCS). [cx-~*P]UTP, [Cy-32P]dATP, and [y-32P]ATP were obtained from Amersham. Pronase E and RNase A were obtained from Sigma, and RNase Tl was purchased from Calbiochem. Construction
of recombinant
plasmids
Double-stranded cDNA of Fny-CMV was prepared using the Amersham cDNA synthesis system, which utilizes the method of Gubler and Hoffman (1983). The primer was synthesized by the Cornell Biotechnology Program and consisted of a decamer complementary to the 3’ten nucleotides of CMV strain B (5’-TGGTCTC-
’ To whom requests for reprints should be addressed. 495
0042-6822/88
$3.00
Copyright 0 1989 by Academic Press. Inc All rights of reproduction I” any form reserved.
496
OWEN AND PALUKAITIS
C-IT-S’). cDNA was made blunt-ended by incubation with DNA polymerase I and was cloned into the Smal site of pUCl8. A recombinant plasmid containing RNA 3 sequences, pUCl8-Fny3, was used for restriction mapping. The cDNA insert was transferred to PBS M13- (the Bluescribe vector of VCS); i.e., the insert between the BarnHI and EcoRl restriction sites within the multiple cloning site of pUCl8-Fny3 was transferred into the corresponding sites of PBS M 13-. The orientation of the cDNA in PBS-Fny3, the recombinant transcription vector, is such that the cDNA representing the 3’ end of the viral sequence is adjacent to the T7 bacterial promoter, and the cDNA representing the 5’ end of the viral sequence is adjacent to the T3 promoter. In the recombinant plasmid pBS3aF, 900 bp of the cDNA insert of PBS-Fny3, from the internal SalI site to the flanking EcoRl site (representing the 3’ half of RNA 3) were removed, and the plasmid was religated with EcoRl linkers. The restriction sites BarnHI, X&l, S&l, and Pstl were removed from the multiple cloning site of the vector in pBS3aF by digestion with BarnHI and Pstl, treatment of linear DNA with mung bean nuclease, and religation. In both PBS-Fny3 and PBS3aF, there are 23 bp between the T7 promoter and the start of the cloned cDNA sequences. Enzymes were used in cloning procedures as recommended by the manufacturers.
Slot blot hybridization Viral RNA samples (1 pg) were diluted into 80 ~1 500 mM sodium acetate, pH 6.0, 10 mM MgCl*, 20% (v/v) ethanol, 3% (w/v) SDS, 1.O M NaCI, and 0.05% (w/v) bromphenol blue (AMESS Buffer) (Palukaitis et al., 1985) and 20-~1 samples were loaded into each slot of an S &S filtration apparatus. Filters were baked, prehybridized, hybridized to random-primed cDNA probes under high stringency, washed, and autoradiographed as previously described (Palukaitis et a/., 1983). Low stringency hybridization was at 50” and the filters were washed at room temperature and at 42”. Randomprimed cDNA was prepared either to Fny-CMV RNA or to WL-CMV RNA as previously described (Palukaitis et al., 1983). ln vitro transcriptions Plasmid DNA templates were linearized with Seal, BarnHI, Aval, HindIll, or Sall, and also made bluntended with DNA polymerase I when the latter four enzymes were used. Radioactively labeled RNA transcripts were prepared as recommended by VCS. DNA (1 pg) was incubated with 40 mNITris-HCI, pH 8.0, 8 m/M MgCIZ, 2 mM spermidine, 50 mM NaCI, 0.4 mA# ATP, CTP, and GTP, 20 r/M UTP, 10 &i [a-32PlUTP, 30
mM DTT, 40 units RNasin, and 10 units of either T3 or T7 polymerase. Reactions were incubated for 30 min at 37”. DNase (0.2 ng) was added and the incubation was continued at 37” for 10 min. RNA transcripts were extracted with phenol and chloroform and were precipitated by the addition of 0.1 NI sodium acetate and 2.5 vol ethanol.
RNA protection assays and gel analysis Hybridizations and RNase digestions were performed essentially by the method of Winter et al. (1985); i.e., 2.0 X 1O4to 3.5 X 1O4cpm of minus-sense probe RNA was combined with viral RNA in 80% formamide, 0.001 M EDTA, 0.3 M Pipes-KOH, pH 6.7, and 0.4 M NaCl such that the final concentration of RNA was 3 1 to 190 pg/ml. Reaction mixtures were heated to 85” and annealed at 50” from 18 to 19 hr. Reactions were then digested with 20 pg/ml RNase A and 1 pg/ml RNase Tl for 30 min in 0.3 M NaCI, 0.01 M Tris-HCI, pH 7.5, and 0.005 M EDTA at 34”. Pronase E (100 pg) and SDS to 0.55% were added, and digestion was continued for 15 min at 37”. RNA was extracted with phenol and chloroform and analyzed on either 5 or 6% polyacrylamide gels containing 7 n/l urea (Sanger and Coulson, 1978) or on 1.5% agarose gels containing formaldehyde (Maniatis et a/. 1982) and exposed to X-ray film at 4”. Markers were a DNA ladder (BRL) end-labeled with [T-~~P]ATP according to standard reaction conditions.
Virus strains and viral RNA preparation CMV strains were from New York State (Fny-, Sny-, B-, WL-, and L,-CMV), Florida (Pf-CMV), Wisconsin (W,-CMV), France (D-CMV), South Africa (S-CMV), and Australia (T-CMV). Pseudorecombinants M, ,M2,G3CMV, and U,,Kz,KB-CMV were obtained from Dr. R. Francki. G-CMV is a U.S. strain (originally provided by Dr. R. Grogan) and M-CMV is an aphid nontransmissible mutant of Price’s No. 6 strain of CMV, from the UK. U-CMV is a banana strain from Australia and K-CMV is a corn strain from China. Pf-CMV, Fny-CMV, and SnyCMV were from Dr. T. Zitter; T-CMV, U, ,K2,K3-CMV (Rao and Francki, 1982) and M, ,M2,G3-CMV were from Dr. R. Francki. D-CMV, S-CMV, and W,-CMV were from Dr. J. Kaper; and B-CMV (Prowidenti, 1976) L,-CMV (formerly LS-CMV, Prowidenti et al., 1980), and WL-CMV (Gonsalves et al., 1982) were from Dr. D. Gonsalves. BMV was obtained from Promega BioTech. CPMV, TAV, and TMV were obtained from the Cornell virus collection. The origin of the other CMV strains will be described elsewhere (Palukaitis, submitted for publication). CMV strains were propagated in N. tabacum cv. Xanthi nc, squash, or N. clevelandi and the viral
497
CMV STRAIN HETEROGENEITY -Strinasncv Fny-CMV
WL-CMV
Fny-CMV
WL-CMV
BDFllyMl,MZ
SUBGROUP
,.G3-
I
PfSny TwTBMV
by sequence data (Rizzo and Palukaitis, 1988; unpublished data). Conversely, labeled cDNA of WL-CMV hybridized strongly at either stringency to L,-, Q-, and S-CMV, but hardly at all to B-, D-, Fny-, MI ,M2,G3-, Pf-, Sny-, T-, or W,-CMV (Fig. 1, lanes 2 and 4). No hybridization signal is evident against the control viral RNAs of BMV, CPMV, or TMV (Fig. 1, lanes 1 and 3). These CMV strains thus follow two patterns of complementarity and are designated subgroups I and II. The cucumovirus, TAV, showed little complementarity to either probe.
C:PMV
Heterogeneity
TAV TMV L2SUBGROUP
Q-
II
SWL-
FIG. 1. Autoradiogram of a slot blot hybridization. Labeled cDNA probes made from the RNAs of Fny-CMV or WL-CMV (as designated above columns) were hybridized to viral RNAs listed on the left. Hybridizations were done at either high or low stringency, as shown (see Materials and Methods). The viral RNAs are arranged in two groups according to their hybridization behavior.
RNA was purified and extracted as previously scribed (Palukaitis and Zaitlin, 1984).
de-
RESULTS Cross-hybridization
among the RNA 3s of CMV strains
In order to examine the extent and nature of the cross-reactivity seen among the CMV strains of subgroup I, cloned cDNA was prepared from the RNAs of Fny-CMV. The cloned cDNA of Fny-CMV RNA 3 was selected and compared with the RNA 3 sequences of the other CMV strains of subgroup I. Cloned cDNA of RNA 3 of Fny-CMV was oriented in the transcription vector, PBS-Fny3, such that minus-sense RNA could be synthesized in vitro from the bacterial T7 promoter. DNA templates for transcription were linearized, and radioactively labeled, run-off transcripts of differing lengths were generated (Fig. 2). These minus-sense probes were annealed to the total RNA of a number of cucumoviruses and the RNA that was resistant to digestion with RNase was separated on denaturing agarose or polyactylamide gels. Results obtained using a probe containing the entire length of the cloned cDNA are shown in Fig. 3; probes specific for the coat protein gene region of the genome are shown in Fig.
of cucumoviruses
Labeled cDNA produced from the total RNA of FnyCMV was hybridized to the filter-immobilized RNAs of 12 strains of CMV under conditions of either high or low stringency(Fig. 1). D-, M, ,M2,G3-, Pf-, Sny-, T-, and W&MV hybridized strongly to the Fny-CMV probe RNA under both conditions of hybridization, while a much weaker signal from B-CMV was obtained under high stringency conditions. The lower intensity of the signal from B-CMV is not quantitatively significant, since in other experiments, the hybridization signals obtained with Fny-CMV cDNA are comparable among all the subgroup I CMV strains (data not shown). FnyCMV cDNA did not hybridize appreciably to L,-CMV, Q-CMV, S-CMV, and WL-CMV at high stringency, although there is a low level of homology to L2-, Q-, and S-CMV at the low stringency. This is not unexpected, as short stretches of sequence homology have been noted between at least Fny- and Q-CMV, as indicated
Hlmwl (1200n”sleotlds*)
ef-
S0.I wo n”sleotld*s,
FIG. 2. Schematic diagram of the minus-sense probes of CMV RNA 3. The insert of plasmid pBSFny3, containing cDNA of RNA 3 of Fny-CMV (heavy line), is flanked by PBS M 13- vector sequences (1 1 1 1j I). The cDNA corresponding to the 3’ end of the viral genome is adjacent to the T7 promoter. Minus-sense transcripts are a, b, c, d, e, and f. The restriction site used to linearize the cDNA template is shown on the right of each transcript, and the length of the transcript is in parentheses. Transcripts a, b, c, and d were synthesized from pBSFny3. Transcripts e and f were synthesized from PBS-BaF, in which the sequences from Sal1 to EcoRl have been deleted from pBSFny3.
498
OWEN AND PALUKAITIS
subgroup I viruses, less-than-full-length synthesized from the cloned cDNA.
probes were
Heterogeneity within the coat protein gene and the 3’ nontranslated region
FIG. 3. Autoradiogram of a 1.5% formaldehyde-agarose gel containing labeled, minus-sense probe-a, transcribed from pBSFny3 and protected from RNase digestion after annealing to the following viral RNAs: Pf-CMV, lane 2; T-CMV, lane 3; U, ,K1,K3-CMV, lane 4; B-CMV, lane 5; M, ,M,,G,-CMV, lane 6; D-CMV, lane 7; Sny-CMV, lane 8; Fny-CMV, lane 9; and yeast RNA, lane 10. End-labeled FnyCMV RNA 4 is in lane 1, Sizes of DNA markers (lane M) are indicated in nucleotides.
4 and probes specific for the 5’ half of the RNA 3 are presented in Fig. 5. RNA synthesized from the plasmid pBSFny3 linearized at the BarnHI site of the vector (probe-a; Fig. 2) extended across the entire cDNA insert. This probe was used for the initial analysis of RNA 3 of the subgroup I CMV strains. The pseudorecombinants U, ,Kp ,K,-CMV, and M, ,Mp ,G,-CMV were used for the analysis of RNA 3 of subgroup I strains K- and G-CMV, respectively. The resulting pattern of protected RNA fragments from the CMV strains was complex (Fig. 3). When the minus-sense probe was annealed to the natural population of Fny-CMV, the largest fragment resistant to RNase digestion was about 1800 nucleotides (Fig. 3, lane 9). Some smaller RNA species were also present, indicating that some RNA 3 molecules in the population exhibit heterogeneity at internal sites. Thus, the cloned cDNA of RNA 3 of Fny-CMV represents the major component of the virus population, but the population is not homogeneous. One of these smaller bands, approximately 1000 nucleotides, probably represents annealing of probe-a to RNA 4. RNA 3 of FnyCMV appears to be closely related to the RNA 3 of Pf-, T-, D-, and Sny-CMV (Fig. 3, lanes 2, 3, 7, and 8) as evidenced by the protection of large fragments of the probe. By the same criterion, Fny-CMV RNA 3 was less related to K-, B-, and G-CMV (Fig. 3, lanes 4, 5, and 6) as less protection of the probe sequences was apparent. To determine the location of the sites of heterogeneity between the RNA 3 of Fny-CMV and the other
Labeled, minus-sense RNA, about 900 nucleotides in length, was transcribed from plasmid PBS-Fny3, extending from the T7 promoter to the Sal1 site in the cloned cDNA of RNA 3. This probe (probe-b; Fig. 2) represents the 3’half of RNA 3, containing the coat protein gene of the virus along with the 3’ noncoding sequences. RNA probe-b was annealed to total RNA of CMV strains Fny-, Sny-, T-, D-, Pf-, B-, G-, K-, Q-, S-, L2-, and WL-, and also to either TAV or yeast RNA, and was analyzed as above (Fig. 4A). When only yeast RNA was included with probe-b in the annealing reaction, the RNA probe was completely degraded by RNase (Fig. 4A, lane 2). CMV strains of subgroup II protected little or none of the minus-sense probe (Fig. 4A, lanes 1 l15). Subgroup I CMV strains each demonstrated significant homology to probe-b and produced a pattern of protected RNA fragments characteristic for each virus (Fig. 4A, lanes 3-l 0). When probe-b was annealed to the natural Fny-CMV population (Fig. 4A, lane 3) much of the probe was completely protected. The remainder of the probe was cleaved by RNase at one location, resulting in two RNA fragments of about 200 and 700 nucleotides in length. Thus some of the molecules in the population exhibit sequence variance at this location. Fny-, Sny-, and T-CMV showed RNA bands of identical size, but differing in intensity. The major bands in the RNA protection patterns of both Sny- and T-CMV resembled the variant RNA species visible in the FnyCMV population (Fig. 4A, lanes 4 and 5, cf. Fig. 4A, lane 3). CMV strains D-, Pf-, B-, G-, and K- showed progressively less sequence homology with the 3’ 900 nucleotides of Fny-CMV RNA 3 (Fig. 4A, lanes 6, 7, 8, 9, and 10). Two shorter minus-sense probes were synthesized from the T7 promoter of PBS-Fny3 in order to position the site of heterogeneity. Probes-c (650 nucleotides) and -d (400 nucleotides) extended from the 3’ end of the cDNA to the HindIll and Aval sites, respectively (Fig. 2). Regardless of whether probe-b, -c, or -d was used in the reaction, each CMV strain of subgroup I protected a 200-nucleotide RNA fragment at the 3’end of the viral genome (Figs. 4A, B, and C). The length of the other major protected fragment in Sny-, T-, D-, Pf-, G-, and B-CMV increased as the minus-sense probe length increased (Figs. 4A-C, lanes 4-9). In all but RNA 3 of K-CMV (lane lo), these two fragments constitute a major proportion of the total RNA annealed to the probe. Thus, the coat protein sequences share considerable nucleotide sequence homology.
499
CMV STRAIN HETEROGENEITY
M1
2
34567
69101112M131415
M t
2
3
4
5
6
7 6
9 10 M
FIG. 4. Autoradiogram of a denaturing polyacrylamide gel of labeled, minus-sense RNA probes-b (A), -c (5). and -d (C), transcribed from PBSFny3 and protected from RNase digestion after annealing to cucumoviral RNAs. Lane 1: probe alone, no treatment. Lanes 2-9: Protected RNA from reaction mixtures containing probe RNA and either yeast RNA, lane 2; Fny-CMV, lane 3; Sny-CMV, lane 4; T-CMV, lane 5; D-CMV, lane 6; Pf-CMV, lane 7; B-CMV, lane 8; M, ,M,,G,-CMV, lane 9; U,,KP,K&MV, lane 10; TAV, lane 11; 0CMV, lane 12; S-CMV, lane 13; L,-CMV, lane 14; or WL-CMV, lane 15. Sizes (in nucleotides) of DNA markers (lane M) are indicated.
Heterogeneity region
within the 3a gene and the intergenic
In order to assess the genetic variation among the cucumoviruses in the 5’half of RNA 3, a restriction fragment containing 900 bp of the cDNA proximal to the T7 promoter was removed from pBSFny3. The resulting plasmid, pBS3aF, was used to direct the synthesis of two, overlapping, minus-sense probes of 1200 nucleotides and 650 nucleotides (probe-e and probe-f, Fig. 2). Assuming that the genome organization is analogous to Q-CMV (Gould and Symons, 1977), these probes should cover both the intercistronic region of RNA 3 and structural sequences of the 3a gene. In Figs. 5A and B, the two labeled RNA probes were annealed to the viral RNAs and treated as above. With probe-e, a band was present in the control reaction when no plussense viral RNA was added (Fig. 5A, lane 2). This band results from a cloning aberration at the extreme 5’ end of the cDNA template. Approximately 300 bp of internal sequences of RNA 3 are located at the 3’ end of the probe in the opposite orientation to the rest of the template (unpublished sequence data). These sequences do not interfere with the annealing reactions; however, intramolecular self-annealing of probe sequences occurs when competing RNA is absent. This band was also present at lower levels with probe-f (Fig. 5, lane 2) with some, but not all, preparations of probe-f, and
probably arose from a low level of transcription through the Seal site (results not shown). This band was also apparent in reactions containing the subgroup II strains Q-, S-, L2-, and WL-CMV (Fig. 5A, lanes 12-15) and TAV (Fig. 5A, lane 1 l), which otherwise showed little or no nucleotide sequence complementarity to the probe. Aside from this 300-nucleotide probe fragment, the protection patterns obtained using both probes-e and -f indicate that Fny-, Sny-, D-, and Pf- CMVs are highly homologous in this region of the genome. When probe-f was used, a protected RNA species of 550 nucleotides was present in lanes containing Fny-, Sny-, D-, and Pf-CMV (Fig. 5B, lanes 3, 4, 6, and 7). The longer probe-e corroborated the similarity among these four strains (Fig. 5A, lanes 3, 4, 6, and 7). T-, B-, G-, and K-CMVs contained divergent sequences at a number of locations (Figs. 5A and B, lanes 5, 8, 9, and 10). DISCUSSION Thirteen CMV strains have been divided into two subgroups, depending on the ability of the total viral RNA to hybridize with cDNAs to RNAs of either FnyCMV or WL-CMV. Within a subgroup, each of the genomic RNAs is homologous to the corresponding genomic RNA of other strains (Palukaitis, submitted for publication). None of the strains shown here repre-
500
OWEN AND PALUKAITIS
FIG. 5. Autoradiogram of a denaturing polyacrylamide gel of labeled, minus-sense RNA probes-e (A) and -f(B) protected after annealing to cucumoviral RNAs. Samples, lane numbers, and markers are as in Fig. 4.
sented a natural pseudorecombinant formed between RNAs belonging to the two subgroups, although many such pseudorecombinants are indeed viable (Rao and Francki, 1982; Edwards et a/., 1983) indicating a strict conservation of genetic function and protein sequences involved in viral protein-protein and viral RNA-protein interactions. Although many of our subgroup I CMV strains seem more virulent than the subgroup II strains on several hosts (unpublished data), the sample size of subgroup II strains is too small to substantiate this generalization. Variation in RNA 3 of the strains of subgroup I has been analyzed further in this report. RNA 3 contains the information for the coat protein gene in the 3’ half of the RNA, and the information for the 3a gene in the 5’half. It was of interest to determine the extent of heterogeneity among various strains of CMV (that differ in host range) in one or both of these genes. A full-length minussense transcript of RNA 3 of Fny-CMV was used as a probe for heterogeneity in an RNA protection assay, in which the probe was protected against digestion by RNase after annealing to various viral RNAs. A complex pattern of protected fragments was produced for a number of CMV strains, and the pattern was different for each strain. Although two of the New York State strains (Fny- and Sny-CMV) were the most closely related, there was generally no correlation between geographical proximity (to New York State) and similarity of sequence. Using probes covering various regions of RNA 3, it was possible to approximate the positions of many of
from RNase digestion
the cleavage sites and hence the loci of the heterogeneities between a number of the CMV strains (Fig. 6). For one of the strains (K-CMV), the number of cleavage products was too numerous to construct a heterogeneity map. It must be remembered that these maps all relate to Fny-CMV RNA 3: i.e., a unidirectional heterogeneity analysis. Moreover, it should be noted that not every mismatch in the RNA duplex will be recognized and cleaved to completion by RNase A and Tl (Winter et a/., 1985). A site of great heterogeneity between strains may result from a mismatch recognized frequently, or a small stretch of mismatches which together result in complete digestion. The frequency of recognition of a site is, however, completely reproducible, as is the resulting pattern of protected fragments. None of the CMV strains analyzed had a 3’ noncoding region completely homologous to Fny-CMV RNA 3 although strains Sny-, T-, D-, B-, and Pf-CMV showed only one site of heterogeneity within this region. FnyCMV itself showed heterogeneity at both the 3’end and the 5’ end: each of the three 3’-coterminal probes (-b, -c, and -d) showed the presence of a subpopulation with sequence variation at approximately 200 nucleotides from the 3’ end (Figs. 4A, B, and C). Our results show that the heterogeneity in the 3’ end of the FnyCMV RNA 3 subpopulation also appears to be present in Sny- and T-CMV RNA 3. In Sny-CMV, complete hybridization to the RNA species represented by the probe is present only at very low levels, and in T-CMV it is virtually absent (Fig. 4). Comparison of the results in Figs. 4 and 5 show that among many of the strains,
CMV STRAIN HETEROGENEITY
\\coat protein\\
\\\\\\3r\\\\\\\
_---_-
II
1
Fny-CMV
I
_____-
N
I
Sny-CMV
I
D-CMV
I
T-CMV
I
Pf-CMV
I
,,,
B-CMV
I
1.1
G-CMV
I
__--__
N
I
-_
, I,,
___-__
__---_
I
I
IL-
I,
IJJ
,,I
I
111
I
,I/
I
501
or symptom development to be established. Among the strains used in this study, some such differences have been observed and localized to RNA 2 by pseudorecombination (Rao and Francki, 1982; Edwards eta/., 1983). Thus, it would be of interest to see what degree of heterogeneity exists among RNA 2 of the subgroup I CMV strains. The latter analysis should be facilitated by the recent completion of the nucleotide sequence of Fny-CMV RNA 2 (Rizzo and Palukaitis, 1988). Finally, the assay described above should also be useful for identifying sites of recombination between virus strains, since each strain exhibits a unique digestion pattern for which a heterogeneity map can be constructed.
ACKNOWLEDGMENTS __--__
I
I
111
I
I
FIG. 6. Schematic diagram of regions of heterogeneity along the length of the RNA 3 of seven CMV strains. Vertical lines depict a location where the CMV RNA 3 shown diverges from the cloned cDNA of Fny-CMV RNA 3. The height of the vertical line denotes the extent of the heterogeneity as determined from the abundance of the probe protected in a particular region of RNA 3. The measure to the right of each strain gives the maximum variation possible, indicating complete cleavage at the extreme height. Slashed lines (//) indicate a region of heterogeneity not precisely located by analysis of the resulting fragments. A slashed region does not necessarily indicate that no homology exists between strains, as there may be smaller regions of uninterrupted homology. Dashes (---) represent a cloning aberration, described in the text, which does not affect the results. Bracketed vertical lines represent cleavage sites which produce two fragments the orientation of which could not be determined.
the coat protein gene was more conserved than the 3a gene. This is represented schematically in Fig. 6. The largest protected RNA species using probe-a, which represents the entire cDNA, was about 1800 nucleotides (Fig. 3, lane 9): approximately 200 nucleotides less than the actual size of Fny-CMV RNA 3. Sny-CMV was the only strain tested in which at least a portion of its population annealed along the complete length of the Fny-CMV minus-sense RNA probe. The degenerate nature of the genetic code could permit nucleotide sequence variation without affecting strict amino acid conservation for the coat protein function. These results suggest that the RNA sequence itself, or its concomitant secondary structure, must be conserved for various RNA-protein interactions. On the other hand, the considerable difference in nucleotide sequence observed in the coat protein genes of K-CMV, TAV, and the subgroup II CMV strains, suggests that extensive sequence variations in this region can occur without affecting function. Overall, our results show that heterogeneity between strains of CMV can be identified and localized without extensive sequence information. This procedure should permit correlation of specific sequence variations and differences in either host range
This work was supported by Grant DE-FG02-86ER13505 from the Department of Energy. 1.0. was supported by a grant from the Cornell Biotechnology Program which is sponsored by the New York State Science and Technology Foundation, a consortium of industries, and the U.S. Army Research Office.
REFERENCES DEVERGNE,J. C., and CARDIN, L. (1973). Contribution a I’etude du virus de la mosaique du concombre (CMV). IV. Essai de classification de plusieurs isolats sur la base de leur structure antigenique. Ann. fhytopathol. 5,409-430. EDWARDS,M. C., and GONSALVES.D. (1983). Grouping of seven biologically defined isolates of cucumber mosaic virus by peptide mapping. fhytoparhology 73, 1 1 17-l 120. EDWARDS,M. C.. GONSALVES,D., and PROWIDENTI, R. (1983). Genetic analysis of cucumber mosaic virus in relation to host resistance: Location of determinants for pathogenicity to certain legumes and Lactuca saligna. Phytopathology 73, 269-273. GONDA, T. J., and SYMONS, R. H. (1978). The use of hybridization analysis with complementary DNA to determine the RNA sequence homology between strains of plant viruses: Its application to several strains of cucumoviruses. Virology 88, 361-370. GONSALVES,D., PROWIDENTI, R., and EDWARDS,M. C. (1982) Tomato white leaf: The relation of an apparent satellite RNA and cucumber mosaic virus. Phyropathology 72, 1533-l 538. GOULD, A. R., and SYMONS, R. H. (1977). Determination of the sequence homology between the four RNA species of cucumber mosaic virus by hybridization analysis with complementary DNA, Nucleic Acids Res. 4, 3787-3802. GOULD, A. R., and SYMONS, R. H. (1982). Cucumber mosaic virus RNA 3: Determination of the nucleotide sequence provides the amino acid sequences of protein 3A and viral coat protein. Eur. J. Biothem. 126,2 17-226. GUBLER, U., and HOFFMAN, B. J. (1983). A simple and very efficient method for generating cDNA libraries. Gene 25, 263-269. HABILI, N., and FRANCKI, R. I. B. (1974). Comparative studies on tomato aspermy and cucumber mosaic viruses. III. Further studies on relationship and construction of a virus from parts of the two viral genomes. Virology 61, 443-449. KAPER, 1. M., and WATERWORTH, H. E. (1981). Cucumoviruses. In “Plant Virus Infections and Comparative Diagnosis” (E. Kurstak. Ed.), pp. 257-332. Elsevier/Holland Biomedical, New York, KIBERSTIS,P. A., LOESCH-FRIES,L. S., and HALL, T. C. (1981). Viral protein synthesis in barley protoplasts inoculated with native and fractionated brome mosaic virus RNA. virology 112, 804-808.
502
OWEN AND PALUKAITIS
MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory. Cold Spring Harbor, NY. NASSUTH, A., and BOL, J. F. (1983). Altered balance of the synthesis of plus- and minus-strand RNAs induced by RNAs 1 and 2 of alfalfa mosaic virus in the absence of RNA 3. \/irology 124, 75-85. PALUKAITIS, P. Characterization of cucumber mosaic virus. II. Subgenomic RNAs 4a and 5 are 3’.coterminal with CMV RNA 3 and are CMV subgroup specific. Submitted for publication. PALUKAITIS,P., Corrs, S., and ZAITLIN. M. (1985). Detection and identification of viroids and viral nucleic acids by “dot-blot” hybridization. Acta Horticult. 164, 109-l 18. PALUKAITIS,P., GARCIA-ARENAL, F., SULZINSKI, M. A., and ZAITLIN, M. (1983). Replication of tobacco mosaic virus. VII. Furthercharacterization of single- and double-stranded virus-related RNAs from TMV infected plants. Virology 131, 533-545. PALUKAITIS,P., and ZAITLIN. M. (1984). Satellite RNAs of cucumber mosaic virus: Characterization of two new satellites. Virology 132, 426-435. PEDEN, K. W. C., and SYMONS, R. H. (1973). Cucumber mosaic virus contains a functionally divided genome. Virology 53,487-492. PIAZZOLLA, P., DIM-RUIZ. J. R., and KAPER,J. M. (1979). Nucleic acid homologies of eighteen cucumber mosaic virus isolates determined by competition hybridization. /. Gen. Viral. 45, 361-369.
PROVVIDENTI,R. (1976). Reaction of Phaseolus and Macroptilium species to a strain of cucumber mosaic virus. PlantDis. Rep. 60,289293. PROVVIDENTI,R., ROBINSON, R. W., and SHAIL, J. W. (1980). A source of resistance to cucumber mosaic virus in Lactuca saligna L. for lettuce (L. sativa L.). Hortscience 15, 528-529. RAO, A. L. N., and FRANCKI, R. I. B. (1982). Distribution of determinants for symptom production and host range on the three RNA components of cucumber mosaic virus. 1. Gen. Viral. 61, 197-205. Rizzo, T., and PALUKAITIS,P. (1988). Nucleotide sequence and evolutionary relationships of cucumber mosaic virus strains (CMV): CMV RNA 2.1. Gen. Viral. 69, in press. SANGER, F., and COULSON, A. R. (1978). The use of thin acrylamide gels for DNA sequencing. FEBS Lett. 87,107-l 10. SCHWINGHAMER,M. W., and SYMONS, R. H. (1975). Fractionation of cucumber mosaic virus RNA and its translation in a wheat embtyo cell-free system. Virology63, 252-262. STUSSCGARAUD,C., GARAUD, J.-C., BERNA, A., and GODEFROY-COLBURN,T. (1987). ln situ location of an alfalfa mosaic virus non-structural protein in plant cell walls: Correlation with virus transport. J. Gen. Virol. 68, 1779-l 784. WINTER, E., YAMAMOTO, F., ALMOGUERA, C., and PERUCHO,M. (1985). A method to detect and characterize point mutations in transcribed genes: Amplification and overexpression of the mutant cKi-ras allele in human tumor cells. Proc. Nat/. Acad. Sci. USA 82, 7575-7579.
166,495-502
(1988)
Characterization I. Molecular
of Cucumber
Heterogeneity
Mapping
Mosaic Virus
of RNA 3 in Eight CMV Strains
JUDITH OWEN AND PETER PALUKAITIS’ Department
of Plant Pathology, Received March
Cornell University,
Ithaca, New York 14853
16, 1988; accepted June 10, 1988
RNAs from 13 strains of cucumber mosaic virus (CMV) were divided into two groups on the basis of their ability to hybridize to cDNA of either Fny-CMV RNA or WL-CMV RNA. The extent of the cross-hybridization within one of these groups was analyzed by an RNA protection assay. A cDNA clone of RNA 3 of the Fny strain of CMV was placed in a transcription vector between bacterial promoters T3 and T7. Labeled, minus-sense RNA transcripts prepared from all or part of the cDNA to RNA 3 of Fny-CMV were annealed to the genomic RNA of each of a number of cucumoviruses and digested with RNases. The patterns of RNA fragments protected from digestion were specific for each CMV strain and revealed the extent and location of heterogeneity among the viruses as well as within the Fny-CMV natural population. This approach will allow the differences in host range and disease processes to be correlated with variations in genomic RNAs. 0 1988 Academic Press, Inc.
INTRODUCTION
was used to analyze RNA 3 of 12 strains of CMV. For 7 of the 8 strains which were in the same subgroup as Fny-CMV, heterogeneity maps of varying complexities could be constructed.
The genome of cucumber mosaic virus (CMV) consists of three single-stranded, positive-sense RNAs (Peden and Symons, 1973). By analogy to other tripartite viruses, RNAs 1 and 2 are believed to encode proteins involved in the replication of CMV (Nassuth and Bol, 1983; Kiberstis et al., 1981) while RNA 3 encodes a 30-35K protein that has been hypothesized to function in promoting the cell-to-cell movement of such viruses (Stussi-Garaud et al., 1987). RNA 3 also encodes the coat protein of CMV (Habili and Francki, 1974). The coat protein, however, is expressed only from a subgenomic RNA, RNA 4, which is colinear with the 3’half of RNA 3 (Schwinghamer and Symons, 1975; Gould and Symons, 1982). A large number of strains of CMV have been isolated which can be differentiated on the basis of biological rather than physical properties (Kaperand Watetvvorth, 1981). On the basis of serological typing (Devergne and Cardin, 1973) peptide mapping of the viral coat protein (Edwards and Gonsalves, 1983) and nucleic acid hybridization (Gonda and Symons, 1978; Piazzolla et al., 1979; Palukaitis, submitted for publication), strains of CMV can be differentiated into two groups (or subgroups). However, strains within a subgroup cannot be further differentiated by the above techniques. In order to further differentiate strains of CMV, to determine the extent of genetic variation between these strains, and to map the genetic variation, an RNA protection assay (Winter et al., 1985)
MATERIALS
AND METHODS
Reagents and enzymes .EcoRI, Seal, Pstl, mung bean nuclease, T4 DNA polymerase, and EcoRl linkers (dGGAAlTCC) were obtained from New England BioLabs. Sall, SamHI, HindIll, T7 polymerase, T4 DNA ligase, Escherichia co/i DNA polymerase I (Klenow), and T4 polynucleotide kinase were obtained from U.S. Biochemicals. Sacl was purchased from Boehringer-Mannheim. RNasin was obtained from Promega BioTec. Aval, DNase, and the DNA 1-kb ladder were obtained from Bethesda Research Laboratories. T3 and T7 polymerases and the Bluescribe plasmid vectors were purchased from Vector Cloning Systems (VCS). [cx-~*P]UTP, [Cy-32P]dATP, and [y-32P]ATP were obtained from Amersham. Pronase E and RNase A were obtained from Sigma, and RNase Tl was purchased from Calbiochem. Construction
of recombinant
plasmids
Double-stranded cDNA of Fny-CMV was prepared using the Amersham cDNA synthesis system, which utilizes the method of Gubler and Hoffman (1983). The primer was synthesized by the Cornell Biotechnology Program and consisted of a decamer complementary to the 3’ten nucleotides of CMV strain B (5’-TGGTCTC-
’ To whom requests for reprints should be addressed. 495
0042-6822/88
$3.00
Copyright 0 1989 by Academic Press. Inc All rights of reproduction I” any form reserved.
496
OWEN AND PALUKAITIS
C-IT-S’). cDNA was made blunt-ended by incubation with DNA polymerase I and was cloned into the Smal site of pUCl8. A recombinant plasmid containing RNA 3 sequences, pUCl8-Fny3, was used for restriction mapping. The cDNA insert was transferred to PBS M13- (the Bluescribe vector of VCS); i.e., the insert between the BarnHI and EcoRl restriction sites within the multiple cloning site of pUCl8-Fny3 was transferred into the corresponding sites of PBS M 13-. The orientation of the cDNA in PBS-Fny3, the recombinant transcription vector, is such that the cDNA representing the 3’ end of the viral sequence is adjacent to the T7 bacterial promoter, and the cDNA representing the 5’ end of the viral sequence is adjacent to the T3 promoter. In the recombinant plasmid pBS3aF, 900 bp of the cDNA insert of PBS-Fny3, from the internal SalI site to the flanking EcoRl site (representing the 3’ half of RNA 3) were removed, and the plasmid was religated with EcoRl linkers. The restriction sites BarnHI, X&l, S&l, and Pstl were removed from the multiple cloning site of the vector in pBS3aF by digestion with BarnHI and Pstl, treatment of linear DNA with mung bean nuclease, and religation. In both PBS-Fny3 and PBS3aF, there are 23 bp between the T7 promoter and the start of the cloned cDNA sequences. Enzymes were used in cloning procedures as recommended by the manufacturers.
Slot blot hybridization Viral RNA samples (1 pg) were diluted into 80 ~1 500 mM sodium acetate, pH 6.0, 10 mM MgCl*, 20% (v/v) ethanol, 3% (w/v) SDS, 1.O M NaCI, and 0.05% (w/v) bromphenol blue (AMESS Buffer) (Palukaitis et al., 1985) and 20-~1 samples were loaded into each slot of an S &S filtration apparatus. Filters were baked, prehybridized, hybridized to random-primed cDNA probes under high stringency, washed, and autoradiographed as previously described (Palukaitis et a/., 1983). Low stringency hybridization was at 50” and the filters were washed at room temperature and at 42”. Randomprimed cDNA was prepared either to Fny-CMV RNA or to WL-CMV RNA as previously described (Palukaitis et al., 1983). ln vitro transcriptions Plasmid DNA templates were linearized with Seal, BarnHI, Aval, HindIll, or Sall, and also made bluntended with DNA polymerase I when the latter four enzymes were used. Radioactively labeled RNA transcripts were prepared as recommended by VCS. DNA (1 pg) was incubated with 40 mNITris-HCI, pH 8.0, 8 m/M MgCIZ, 2 mM spermidine, 50 mM NaCI, 0.4 mA# ATP, CTP, and GTP, 20 r/M UTP, 10 &i [a-32PlUTP, 30
mM DTT, 40 units RNasin, and 10 units of either T3 or T7 polymerase. Reactions were incubated for 30 min at 37”. DNase (0.2 ng) was added and the incubation was continued at 37” for 10 min. RNA transcripts were extracted with phenol and chloroform and were precipitated by the addition of 0.1 NI sodium acetate and 2.5 vol ethanol.
RNA protection assays and gel analysis Hybridizations and RNase digestions were performed essentially by the method of Winter et al. (1985); i.e., 2.0 X 1O4to 3.5 X 1O4cpm of minus-sense probe RNA was combined with viral RNA in 80% formamide, 0.001 M EDTA, 0.3 M Pipes-KOH, pH 6.7, and 0.4 M NaCl such that the final concentration of RNA was 3 1 to 190 pg/ml. Reaction mixtures were heated to 85” and annealed at 50” from 18 to 19 hr. Reactions were then digested with 20 pg/ml RNase A and 1 pg/ml RNase Tl for 30 min in 0.3 M NaCI, 0.01 M Tris-HCI, pH 7.5, and 0.005 M EDTA at 34”. Pronase E (100 pg) and SDS to 0.55% were added, and digestion was continued for 15 min at 37”. RNA was extracted with phenol and chloroform and analyzed on either 5 or 6% polyacrylamide gels containing 7 n/l urea (Sanger and Coulson, 1978) or on 1.5% agarose gels containing formaldehyde (Maniatis et a/. 1982) and exposed to X-ray film at 4”. Markers were a DNA ladder (BRL) end-labeled with [T-~~P]ATP according to standard reaction conditions.
Virus strains and viral RNA preparation CMV strains were from New York State (Fny-, Sny-, B-, WL-, and L,-CMV), Florida (Pf-CMV), Wisconsin (W,-CMV), France (D-CMV), South Africa (S-CMV), and Australia (T-CMV). Pseudorecombinants M, ,M2,G3CMV, and U,,Kz,KB-CMV were obtained from Dr. R. Francki. G-CMV is a U.S. strain (originally provided by Dr. R. Grogan) and M-CMV is an aphid nontransmissible mutant of Price’s No. 6 strain of CMV, from the UK. U-CMV is a banana strain from Australia and K-CMV is a corn strain from China. Pf-CMV, Fny-CMV, and SnyCMV were from Dr. T. Zitter; T-CMV, U, ,K2,K3-CMV (Rao and Francki, 1982) and M, ,M2,G3-CMV were from Dr. R. Francki. D-CMV, S-CMV, and W,-CMV were from Dr. J. Kaper; and B-CMV (Prowidenti, 1976) L,-CMV (formerly LS-CMV, Prowidenti et al., 1980), and WL-CMV (Gonsalves et al., 1982) were from Dr. D. Gonsalves. BMV was obtained from Promega BioTech. CPMV, TAV, and TMV were obtained from the Cornell virus collection. The origin of the other CMV strains will be described elsewhere (Palukaitis, submitted for publication). CMV strains were propagated in N. tabacum cv. Xanthi nc, squash, or N. clevelandi and the viral
497
CMV STRAIN HETEROGENEITY -Strinasncv Fny-CMV
WL-CMV
Fny-CMV
WL-CMV
BDFllyMl,MZ
SUBGROUP
,.G3-
I
PfSny TwTBMV
by sequence data (Rizzo and Palukaitis, 1988; unpublished data). Conversely, labeled cDNA of WL-CMV hybridized strongly at either stringency to L,-, Q-, and S-CMV, but hardly at all to B-, D-, Fny-, MI ,M2,G3-, Pf-, Sny-, T-, or W,-CMV (Fig. 1, lanes 2 and 4). No hybridization signal is evident against the control viral RNAs of BMV, CPMV, or TMV (Fig. 1, lanes 1 and 3). These CMV strains thus follow two patterns of complementarity and are designated subgroups I and II. The cucumovirus, TAV, showed little complementarity to either probe.
C:PMV
Heterogeneity
TAV TMV L2SUBGROUP
Q-
II
SWL-
FIG. 1. Autoradiogram of a slot blot hybridization. Labeled cDNA probes made from the RNAs of Fny-CMV or WL-CMV (as designated above columns) were hybridized to viral RNAs listed on the left. Hybridizations were done at either high or low stringency, as shown (see Materials and Methods). The viral RNAs are arranged in two groups according to their hybridization behavior.
RNA was purified and extracted as previously scribed (Palukaitis and Zaitlin, 1984).
de-
RESULTS Cross-hybridization
among the RNA 3s of CMV strains
In order to examine the extent and nature of the cross-reactivity seen among the CMV strains of subgroup I, cloned cDNA was prepared from the RNAs of Fny-CMV. The cloned cDNA of Fny-CMV RNA 3 was selected and compared with the RNA 3 sequences of the other CMV strains of subgroup I. Cloned cDNA of RNA 3 of Fny-CMV was oriented in the transcription vector, PBS-Fny3, such that minus-sense RNA could be synthesized in vitro from the bacterial T7 promoter. DNA templates for transcription were linearized, and radioactively labeled, run-off transcripts of differing lengths were generated (Fig. 2). These minus-sense probes were annealed to the total RNA of a number of cucumoviruses and the RNA that was resistant to digestion with RNase was separated on denaturing agarose or polyactylamide gels. Results obtained using a probe containing the entire length of the cloned cDNA are shown in Fig. 3; probes specific for the coat protein gene region of the genome are shown in Fig.
of cucumoviruses
Labeled cDNA produced from the total RNA of FnyCMV was hybridized to the filter-immobilized RNAs of 12 strains of CMV under conditions of either high or low stringency(Fig. 1). D-, M, ,M2,G3-, Pf-, Sny-, T-, and W&MV hybridized strongly to the Fny-CMV probe RNA under both conditions of hybridization, while a much weaker signal from B-CMV was obtained under high stringency conditions. The lower intensity of the signal from B-CMV is not quantitatively significant, since in other experiments, the hybridization signals obtained with Fny-CMV cDNA are comparable among all the subgroup I CMV strains (data not shown). FnyCMV cDNA did not hybridize appreciably to L,-CMV, Q-CMV, S-CMV, and WL-CMV at high stringency, although there is a low level of homology to L2-, Q-, and S-CMV at the low stringency. This is not unexpected, as short stretches of sequence homology have been noted between at least Fny- and Q-CMV, as indicated
Hlmwl (1200n”sleotlds*)
ef-
S0.I wo n”sleotld*s,
FIG. 2. Schematic diagram of the minus-sense probes of CMV RNA 3. The insert of plasmid pBSFny3, containing cDNA of RNA 3 of Fny-CMV (heavy line), is flanked by PBS M 13- vector sequences (1 1 1 1j I). The cDNA corresponding to the 3’ end of the viral genome is adjacent to the T7 promoter. Minus-sense transcripts are a, b, c, d, e, and f. The restriction site used to linearize the cDNA template is shown on the right of each transcript, and the length of the transcript is in parentheses. Transcripts a, b, c, and d were synthesized from pBSFny3. Transcripts e and f were synthesized from PBS-BaF, in which the sequences from Sal1 to EcoRl have been deleted from pBSFny3.
498
OWEN AND PALUKAITIS
subgroup I viruses, less-than-full-length synthesized from the cloned cDNA.
probes were
Heterogeneity within the coat protein gene and the 3’ nontranslated region
FIG. 3. Autoradiogram of a 1.5% formaldehyde-agarose gel containing labeled, minus-sense probe-a, transcribed from pBSFny3 and protected from RNase digestion after annealing to the following viral RNAs: Pf-CMV, lane 2; T-CMV, lane 3; U, ,K1,K3-CMV, lane 4; B-CMV, lane 5; M, ,M,,G,-CMV, lane 6; D-CMV, lane 7; Sny-CMV, lane 8; Fny-CMV, lane 9; and yeast RNA, lane 10. End-labeled FnyCMV RNA 4 is in lane 1, Sizes of DNA markers (lane M) are indicated in nucleotides.
4 and probes specific for the 5’ half of the RNA 3 are presented in Fig. 5. RNA synthesized from the plasmid pBSFny3 linearized at the BarnHI site of the vector (probe-a; Fig. 2) extended across the entire cDNA insert. This probe was used for the initial analysis of RNA 3 of the subgroup I CMV strains. The pseudorecombinants U, ,Kp ,K,-CMV, and M, ,Mp ,G,-CMV were used for the analysis of RNA 3 of subgroup I strains K- and G-CMV, respectively. The resulting pattern of protected RNA fragments from the CMV strains was complex (Fig. 3). When the minus-sense probe was annealed to the natural population of Fny-CMV, the largest fragment resistant to RNase digestion was about 1800 nucleotides (Fig. 3, lane 9). Some smaller RNA species were also present, indicating that some RNA 3 molecules in the population exhibit heterogeneity at internal sites. Thus, the cloned cDNA of RNA 3 of Fny-CMV represents the major component of the virus population, but the population is not homogeneous. One of these smaller bands, approximately 1000 nucleotides, probably represents annealing of probe-a to RNA 4. RNA 3 of FnyCMV appears to be closely related to the RNA 3 of Pf-, T-, D-, and Sny-CMV (Fig. 3, lanes 2, 3, 7, and 8) as evidenced by the protection of large fragments of the probe. By the same criterion, Fny-CMV RNA 3 was less related to K-, B-, and G-CMV (Fig. 3, lanes 4, 5, and 6) as less protection of the probe sequences was apparent. To determine the location of the sites of heterogeneity between the RNA 3 of Fny-CMV and the other
Labeled, minus-sense RNA, about 900 nucleotides in length, was transcribed from plasmid PBS-Fny3, extending from the T7 promoter to the Sal1 site in the cloned cDNA of RNA 3. This probe (probe-b; Fig. 2) represents the 3’half of RNA 3, containing the coat protein gene of the virus along with the 3’ noncoding sequences. RNA probe-b was annealed to total RNA of CMV strains Fny-, Sny-, T-, D-, Pf-, B-, G-, K-, Q-, S-, L2-, and WL-, and also to either TAV or yeast RNA, and was analyzed as above (Fig. 4A). When only yeast RNA was included with probe-b in the annealing reaction, the RNA probe was completely degraded by RNase (Fig. 4A, lane 2). CMV strains of subgroup II protected little or none of the minus-sense probe (Fig. 4A, lanes 1 l15). Subgroup I CMV strains each demonstrated significant homology to probe-b and produced a pattern of protected RNA fragments characteristic for each virus (Fig. 4A, lanes 3-l 0). When probe-b was annealed to the natural Fny-CMV population (Fig. 4A, lane 3) much of the probe was completely protected. The remainder of the probe was cleaved by RNase at one location, resulting in two RNA fragments of about 200 and 700 nucleotides in length. Thus some of the molecules in the population exhibit sequence variance at this location. Fny-, Sny-, and T-CMV showed RNA bands of identical size, but differing in intensity. The major bands in the RNA protection patterns of both Sny- and T-CMV resembled the variant RNA species visible in the FnyCMV population (Fig. 4A, lanes 4 and 5, cf. Fig. 4A, lane 3). CMV strains D-, Pf-, B-, G-, and K- showed progressively less sequence homology with the 3’ 900 nucleotides of Fny-CMV RNA 3 (Fig. 4A, lanes 6, 7, 8, 9, and 10). Two shorter minus-sense probes were synthesized from the T7 promoter of PBS-Fny3 in order to position the site of heterogeneity. Probes-c (650 nucleotides) and -d (400 nucleotides) extended from the 3’ end of the cDNA to the HindIll and Aval sites, respectively (Fig. 2). Regardless of whether probe-b, -c, or -d was used in the reaction, each CMV strain of subgroup I protected a 200-nucleotide RNA fragment at the 3’end of the viral genome (Figs. 4A, B, and C). The length of the other major protected fragment in Sny-, T-, D-, Pf-, G-, and B-CMV increased as the minus-sense probe length increased (Figs. 4A-C, lanes 4-9). In all but RNA 3 of K-CMV (lane lo), these two fragments constitute a major proportion of the total RNA annealed to the probe. Thus, the coat protein sequences share considerable nucleotide sequence homology.
499
CMV STRAIN HETEROGENEITY
M1
2
34567
69101112M131415
M t
2
3
4
5
6
7 6
9 10 M
FIG. 4. Autoradiogram of a denaturing polyacrylamide gel of labeled, minus-sense RNA probes-b (A), -c (5). and -d (C), transcribed from PBSFny3 and protected from RNase digestion after annealing to cucumoviral RNAs. Lane 1: probe alone, no treatment. Lanes 2-9: Protected RNA from reaction mixtures containing probe RNA and either yeast RNA, lane 2; Fny-CMV, lane 3; Sny-CMV, lane 4; T-CMV, lane 5; D-CMV, lane 6; Pf-CMV, lane 7; B-CMV, lane 8; M, ,M,,G,-CMV, lane 9; U,,KP,K&MV, lane 10; TAV, lane 11; 0CMV, lane 12; S-CMV, lane 13; L,-CMV, lane 14; or WL-CMV, lane 15. Sizes (in nucleotides) of DNA markers (lane M) are indicated.
Heterogeneity region
within the 3a gene and the intergenic
In order to assess the genetic variation among the cucumoviruses in the 5’half of RNA 3, a restriction fragment containing 900 bp of the cDNA proximal to the T7 promoter was removed from pBSFny3. The resulting plasmid, pBS3aF, was used to direct the synthesis of two, overlapping, minus-sense probes of 1200 nucleotides and 650 nucleotides (probe-e and probe-f, Fig. 2). Assuming that the genome organization is analogous to Q-CMV (Gould and Symons, 1977), these probes should cover both the intercistronic region of RNA 3 and structural sequences of the 3a gene. In Figs. 5A and B, the two labeled RNA probes were annealed to the viral RNAs and treated as above. With probe-e, a band was present in the control reaction when no plussense viral RNA was added (Fig. 5A, lane 2). This band results from a cloning aberration at the extreme 5’ end of the cDNA template. Approximately 300 bp of internal sequences of RNA 3 are located at the 3’ end of the probe in the opposite orientation to the rest of the template (unpublished sequence data). These sequences do not interfere with the annealing reactions; however, intramolecular self-annealing of probe sequences occurs when competing RNA is absent. This band was also present at lower levels with probe-f (Fig. 5, lane 2) with some, but not all, preparations of probe-f, and
probably arose from a low level of transcription through the Seal site (results not shown). This band was also apparent in reactions containing the subgroup II strains Q-, S-, L2-, and WL-CMV (Fig. 5A, lanes 12-15) and TAV (Fig. 5A, lane 1 l), which otherwise showed little or no nucleotide sequence complementarity to the probe. Aside from this 300-nucleotide probe fragment, the protection patterns obtained using both probes-e and -f indicate that Fny-, Sny-, D-, and Pf- CMVs are highly homologous in this region of the genome. When probe-f was used, a protected RNA species of 550 nucleotides was present in lanes containing Fny-, Sny-, D-, and Pf-CMV (Fig. 5B, lanes 3, 4, 6, and 7). The longer probe-e corroborated the similarity among these four strains (Fig. 5A, lanes 3, 4, 6, and 7). T-, B-, G-, and K-CMVs contained divergent sequences at a number of locations (Figs. 5A and B, lanes 5, 8, 9, and 10). DISCUSSION Thirteen CMV strains have been divided into two subgroups, depending on the ability of the total viral RNA to hybridize with cDNAs to RNAs of either FnyCMV or WL-CMV. Within a subgroup, each of the genomic RNAs is homologous to the corresponding genomic RNA of other strains (Palukaitis, submitted for publication). None of the strains shown here repre-
500
OWEN AND PALUKAITIS
FIG. 5. Autoradiogram of a denaturing polyacrylamide gel of labeled, minus-sense RNA probes-e (A) and -f(B) protected after annealing to cucumoviral RNAs. Samples, lane numbers, and markers are as in Fig. 4.
sented a natural pseudorecombinant formed between RNAs belonging to the two subgroups, although many such pseudorecombinants are indeed viable (Rao and Francki, 1982; Edwards et a/., 1983) indicating a strict conservation of genetic function and protein sequences involved in viral protein-protein and viral RNA-protein interactions. Although many of our subgroup I CMV strains seem more virulent than the subgroup II strains on several hosts (unpublished data), the sample size of subgroup II strains is too small to substantiate this generalization. Variation in RNA 3 of the strains of subgroup I has been analyzed further in this report. RNA 3 contains the information for the coat protein gene in the 3’ half of the RNA, and the information for the 3a gene in the 5’half. It was of interest to determine the extent of heterogeneity among various strains of CMV (that differ in host range) in one or both of these genes. A full-length minussense transcript of RNA 3 of Fny-CMV was used as a probe for heterogeneity in an RNA protection assay, in which the probe was protected against digestion by RNase after annealing to various viral RNAs. A complex pattern of protected fragments was produced for a number of CMV strains, and the pattern was different for each strain. Although two of the New York State strains (Fny- and Sny-CMV) were the most closely related, there was generally no correlation between geographical proximity (to New York State) and similarity of sequence. Using probes covering various regions of RNA 3, it was possible to approximate the positions of many of
from RNase digestion
the cleavage sites and hence the loci of the heterogeneities between a number of the CMV strains (Fig. 6). For one of the strains (K-CMV), the number of cleavage products was too numerous to construct a heterogeneity map. It must be remembered that these maps all relate to Fny-CMV RNA 3: i.e., a unidirectional heterogeneity analysis. Moreover, it should be noted that not every mismatch in the RNA duplex will be recognized and cleaved to completion by RNase A and Tl (Winter et a/., 1985). A site of great heterogeneity between strains may result from a mismatch recognized frequently, or a small stretch of mismatches which together result in complete digestion. The frequency of recognition of a site is, however, completely reproducible, as is the resulting pattern of protected fragments. None of the CMV strains analyzed had a 3’ noncoding region completely homologous to Fny-CMV RNA 3 although strains Sny-, T-, D-, B-, and Pf-CMV showed only one site of heterogeneity within this region. FnyCMV itself showed heterogeneity at both the 3’end and the 5’ end: each of the three 3’-coterminal probes (-b, -c, and -d) showed the presence of a subpopulation with sequence variation at approximately 200 nucleotides from the 3’ end (Figs. 4A, B, and C). Our results show that the heterogeneity in the 3’ end of the FnyCMV RNA 3 subpopulation also appears to be present in Sny- and T-CMV RNA 3. In Sny-CMV, complete hybridization to the RNA species represented by the probe is present only at very low levels, and in T-CMV it is virtually absent (Fig. 4). Comparison of the results in Figs. 4 and 5 show that among many of the strains,
CMV STRAIN HETEROGENEITY
\\coat protein\\
\\\\\\3r\\\\\\\
_---_-
II
1
Fny-CMV
I
_____-
N
I
Sny-CMV
I
D-CMV
I
T-CMV
I
Pf-CMV
I
,,,
B-CMV
I
1.1
G-CMV
I
__--__
N
I
-_
, I,,
___-__
__---_
I
I
IL-
I,
IJJ
,,I
I
111
I
,I/
I
501
or symptom development to be established. Among the strains used in this study, some such differences have been observed and localized to RNA 2 by pseudorecombination (Rao and Francki, 1982; Edwards eta/., 1983). Thus, it would be of interest to see what degree of heterogeneity exists among RNA 2 of the subgroup I CMV strains. The latter analysis should be facilitated by the recent completion of the nucleotide sequence of Fny-CMV RNA 2 (Rizzo and Palukaitis, 1988). Finally, the assay described above should also be useful for identifying sites of recombination between virus strains, since each strain exhibits a unique digestion pattern for which a heterogeneity map can be constructed.
ACKNOWLEDGMENTS __--__
I
I
111
I
I
FIG. 6. Schematic diagram of regions of heterogeneity along the length of the RNA 3 of seven CMV strains. Vertical lines depict a location where the CMV RNA 3 shown diverges from the cloned cDNA of Fny-CMV RNA 3. The height of the vertical line denotes the extent of the heterogeneity as determined from the abundance of the probe protected in a particular region of RNA 3. The measure to the right of each strain gives the maximum variation possible, indicating complete cleavage at the extreme height. Slashed lines (//) indicate a region of heterogeneity not precisely located by analysis of the resulting fragments. A slashed region does not necessarily indicate that no homology exists between strains, as there may be smaller regions of uninterrupted homology. Dashes (---) represent a cloning aberration, described in the text, which does not affect the results. Bracketed vertical lines represent cleavage sites which produce two fragments the orientation of which could not be determined.
the coat protein gene was more conserved than the 3a gene. This is represented schematically in Fig. 6. The largest protected RNA species using probe-a, which represents the entire cDNA, was about 1800 nucleotides (Fig. 3, lane 9): approximately 200 nucleotides less than the actual size of Fny-CMV RNA 3. Sny-CMV was the only strain tested in which at least a portion of its population annealed along the complete length of the Fny-CMV minus-sense RNA probe. The degenerate nature of the genetic code could permit nucleotide sequence variation without affecting strict amino acid conservation for the coat protein function. These results suggest that the RNA sequence itself, or its concomitant secondary structure, must be conserved for various RNA-protein interactions. On the other hand, the considerable difference in nucleotide sequence observed in the coat protein genes of K-CMV, TAV, and the subgroup II CMV strains, suggests that extensive sequence variations in this region can occur without affecting function. Overall, our results show that heterogeneity between strains of CMV can be identified and localized without extensive sequence information. This procedure should permit correlation of specific sequence variations and differences in either host range
This work was supported by Grant DE-FG02-86ER13505 from the Department of Energy. 1.0. was supported by a grant from the Cornell Biotechnology Program which is sponsored by the New York State Science and Technology Foundation, a consortium of industries, and the U.S. Army Research Office.
REFERENCES DEVERGNE,J. C., and CARDIN, L. (1973). Contribution a I’etude du virus de la mosaique du concombre (CMV). IV. Essai de classification de plusieurs isolats sur la base de leur structure antigenique. Ann. fhytopathol. 5,409-430. EDWARDS,M. C., and GONSALVES.D. (1983). Grouping of seven biologically defined isolates of cucumber mosaic virus by peptide mapping. fhytoparhology 73, 1 1 17-l 120. EDWARDS,M. C.. GONSALVES,D., and PROWIDENTI, R. (1983). Genetic analysis of cucumber mosaic virus in relation to host resistance: Location of determinants for pathogenicity to certain legumes and Lactuca saligna. Phytopathology 73, 269-273. GONDA, T. J., and SYMONS, R. H. (1978). The use of hybridization analysis with complementary DNA to determine the RNA sequence homology between strains of plant viruses: Its application to several strains of cucumoviruses. Virology 88, 361-370. GONSALVES,D., PROWIDENTI, R., and EDWARDS,M. C. (1982) Tomato white leaf: The relation of an apparent satellite RNA and cucumber mosaic virus. Phyropathology 72, 1533-l 538. GOULD, A. R., and SYMONS, R. H. (1977). Determination of the sequence homology between the four RNA species of cucumber mosaic virus by hybridization analysis with complementary DNA, Nucleic Acids Res. 4, 3787-3802. GOULD, A. R., and SYMONS, R. H. (1982). Cucumber mosaic virus RNA 3: Determination of the nucleotide sequence provides the amino acid sequences of protein 3A and viral coat protein. Eur. J. Biothem. 126,2 17-226. GUBLER, U., and HOFFMAN, B. J. (1983). A simple and very efficient method for generating cDNA libraries. Gene 25, 263-269. HABILI, N., and FRANCKI, R. I. B. (1974). Comparative studies on tomato aspermy and cucumber mosaic viruses. III. Further studies on relationship and construction of a virus from parts of the two viral genomes. Virology 61, 443-449. KAPER, 1. M., and WATERWORTH, H. E. (1981). Cucumoviruses. In “Plant Virus Infections and Comparative Diagnosis” (E. Kurstak. Ed.), pp. 257-332. Elsevier/Holland Biomedical, New York, KIBERSTIS,P. A., LOESCH-FRIES,L. S., and HALL, T. C. (1981). Viral protein synthesis in barley protoplasts inoculated with native and fractionated brome mosaic virus RNA. virology 112, 804-808.
502
OWEN AND PALUKAITIS
MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory. Cold Spring Harbor, NY. NASSUTH, A., and BOL, J. F. (1983). Altered balance of the synthesis of plus- and minus-strand RNAs induced by RNAs 1 and 2 of alfalfa mosaic virus in the absence of RNA 3. \/irology 124, 75-85. PALUKAITIS, P. Characterization of cucumber mosaic virus. II. Subgenomic RNAs 4a and 5 are 3’.coterminal with CMV RNA 3 and are CMV subgroup specific. Submitted for publication. PALUKAITIS,P., Corrs, S., and ZAITLIN. M. (1985). Detection and identification of viroids and viral nucleic acids by “dot-blot” hybridization. Acta Horticult. 164, 109-l 18. PALUKAITIS,P., GARCIA-ARENAL, F., SULZINSKI, M. A., and ZAITLIN, M. (1983). Replication of tobacco mosaic virus. VII. Furthercharacterization of single- and double-stranded virus-related RNAs from TMV infected plants. Virology 131, 533-545. PALUKAITIS,P., and ZAITLIN. M. (1984). Satellite RNAs of cucumber mosaic virus: Characterization of two new satellites. Virology 132, 426-435. PEDEN, K. W. C., and SYMONS, R. H. (1973). Cucumber mosaic virus contains a functionally divided genome. Virology 53,487-492. PIAZZOLLA, P., DIM-RUIZ. J. R., and KAPER,J. M. (1979). Nucleic acid homologies of eighteen cucumber mosaic virus isolates determined by competition hybridization. /. Gen. Viral. 45, 361-369.
PROVVIDENTI,R. (1976). Reaction of Phaseolus and Macroptilium species to a strain of cucumber mosaic virus. PlantDis. Rep. 60,289293. PROVVIDENTI,R., ROBINSON, R. W., and SHAIL, J. W. (1980). A source of resistance to cucumber mosaic virus in Lactuca saligna L. for lettuce (L. sativa L.). Hortscience 15, 528-529. RAO, A. L. N., and FRANCKI, R. I. B. (1982). Distribution of determinants for symptom production and host range on the three RNA components of cucumber mosaic virus. 1. Gen. Viral. 61, 197-205. Rizzo, T., and PALUKAITIS,P. (1988). Nucleotide sequence and evolutionary relationships of cucumber mosaic virus strains (CMV): CMV RNA 2.1. Gen. Viral. 69, in press. SANGER, F., and COULSON, A. R. (1978). The use of thin acrylamide gels for DNA sequencing. FEBS Lett. 87,107-l 10. SCHWINGHAMER,M. W., and SYMONS, R. H. (1975). Fractionation of cucumber mosaic virus RNA and its translation in a wheat embtyo cell-free system. Virology63, 252-262. STUSSCGARAUD,C., GARAUD, J.-C., BERNA, A., and GODEFROY-COLBURN,T. (1987). ln situ location of an alfalfa mosaic virus non-structural protein in plant cell walls: Correlation with virus transport. J. Gen. Virol. 68, 1779-l 784. WINTER, E., YAMAMOTO, F., ALMOGUERA, C., and PERUCHO,M. (1985). A method to detect and characterize point mutations in transcribed genes: Amplification and overexpression of the mutant cKi-ras allele in human tumor cells. Proc. Nat/. Acad. Sci. USA 82, 7575-7579.