Plant Molecular Biology 11:463-471 © Kluwer Academic Publishers, Dordrecht - Printed in the Netherlands
463
Anti-sense R N A s of cucumber mosaic virus in transgenic plants assessed for control of the virus M. Ali Rezaian, Ken G. M. Skene and Jeff G. Ellis 1
CSIRO Division of Horticulture, P.O. Box 350, Adelaide, S.A. 5001, Australia; 1CSIRO Division of Plant Industry, P.O. Box 1600, Canberra, A.C.T. 2601, Australia Received 19 April 1988; accepted in revised form 6 July 1988
Key words: anti-sense RNA, cucumber mosaic virus, transgenic plants, virus inhibition
Abstract
Three synthetic genes for the production of anti-sense RNA to different regions o f the cucumber mosaic virus (CMV) genome were constructed using virus-derived double-stranded cDNA coupled to a promoter sequence from cauliflower mosaic virus. The genes were used to transform tobacco plants by a Ti plasmid vector. Transgenic plants obtained with the three constructs produced anti-sense RNA at different levels. Plants expressing each of the three anti-sense RNAs were inoculated with CMV and their sensitivity to the virus infection was compared with the non-transformed plants. Only one plant line which expressed relatively low levels of one of the anti-sense RNAs showed resistance to CMV but other plants expressing the same or the other two antisense RNAs had similar sensitivity to CMV infection as the non-transformed plants.
Introduction
In recent years anti-sense RNA (RNA complementary to mRNA) has been shown to inhibit gene expression in a variety of prokaryotic and eukaryotic ceils [21, 39, reviewed in 11] including plants [6]. The mechanisms by which anti-sense RNA inhibits gene expression are not well understood and different modes of inhibition have been described. In eukaryotic systems the formation of double-stranded RNA in the nucleus may block mRNA transport to the cytoplasm [21], or the RNA-RNA hybrids formed in the nucleus may be rapidly degraded [26]. Anti-sense RNA may inhibit m R N A translation bound to polysomes [26, 33]. The anti-sense sequences required for the inhibition of gene expression can be relatively short because a 52 residue RNA has been found effective in inhibiting the thymidine kinase gene [17]. Even injection of anti-
sense oligonucleotides (18-23 residues) into Xenopus oocytes can inhibit translation of the target RNA [19]. Because the genomes of plus strand RNA viruses have structural similarities to mRNA and function as mRNA during virus replication [36], it has been suggested that anti-sense RNA may inhibit virus replication [341. Coleman et al. [3] have examined this idea with anti-sense RNA to an RNA phage and reported that induction of anti-sense RNA (referred to as micRNA) directed against the coat protein and/or replicase genes ofE. coli RNA phage SP prevents the phage multiplication. The use of anti-sense RNA against a plant virus has not been examined before. Loesch-Fries et al. [23] reported the synthesis of anti-sense RNA of alfalfa mosaic virus (AMV) in transgenic plants but did not report any anti-viral activity. We have tested the possibility of anti-sense RNA inhibition of cu-
464 cumber mosaic virus (CMV) which is structurally similar to AMV [31, 32] and an important pathogen with wide host range. Our results indicate that, except for one tobacco line, CMV replication was not significantly affected by the expression of anti-sense RNA in transgenic plants.
Materials and methods
Construction of C M V anti-sense genes Double-stranded DNA fragments corresponding to three regions of the CMV genome (Fig. 1) were excised from cloned CMV sequences contained in the replicative form of M13 phage DNA. Fragment 1 from CMV RNA 2 and fragment 20 from CMV RNA 1 (Fig. 1) were from CMV M13mp7 clones prepared previously [31, 32] and fragment 7 (Fig. 1) was from an M13mp9 clone of RNA 3 kindly provided by C. Davies of Adelaide University. An Eco RIBam HI fragment of clone 7 was inserted in Eco RIBarn HI cut plasmid p35S-Nos (provided by Dr D. LleweUyn o f CSIRO) and Eco RI fragments of
clones 1 and 20 were ligated into Eco RI cut p35SNos. This plant expression vector contains the promoter from the 35S protein gene of cauliflower mosaic virus and a 1 kb polyadenylation signal from the nopaline synthase gene. The resulting plasmid construct containing fragment 7 had an insert orientation required for anti-sense RNA synthesis, but constructs 1 and 20 in the anti-sense orientations were identified by restriction enzyme analysis. The procedure for construction of Ti plasmids was the same for the three CMV sequences and this is shown in Fig. 2 for the construct 7 as an example. Each of the three plasmids was cut at a single site and inserted in the binary Ti vector Bin 19 [2]. The Ti plasmid constructs were mobilized from E. coli HB101 to Agrobacterium tumefaciens LBA4404 by a triparental mating procedure using the mob functions of plasmid RK2013 [5]. The presence of CMV gene constructs in the Ti plasmids was confirmed by restriction enzyme analysis of the plasmid DNAs purified from A. tumefaciens. Plant transformations were carried out by co-cultivation of tobacco leaf discs [16] with Agrobacterium carrying the binary vector [2].
1
RNA
3389
b-----
1 --~ o 70
4 ----Antisense 355
R N A 20
1 RNA
3035
2 Antisense RNA 1---~a 2478 1
RNA
3
4 2900 2193
--I .... CP
3A A n t i s e n s e R N A 7 -'-~31I
I
236
RNA
4
1 --~
1027
CP Fig. 1. Genome organization of CMV. Locations of the three regions used for construction of anti-sense RNA genes are shown below CMV RNAs 1, 2 and 3. The size of each RNA and positions of the anti-sense RNAs relative to CMV RNAs are indicated by residue numbers. The open boxes depict the reading frames and the solid boxes show the 3' -end conserved regions of the viral genome. Data for RNA 1 and 2 are from refs. 32 and 31 and for RNA 3 are from ref. 10.
465
~~~H.j;~=:~ #NOS
promoter ~r.jE
[
B
"/
CMV insert
E
promoter ,,f CMV insert
[ :~-E polylinker { p35SCMV 7 )l] LH ~N,~ ~/'NOS LB ~
~ B i
H 19~ ~
~
LB
/
H
~"'~--""-"4~
~ C M V
/
(
n
promoter
\\\
insert
NOS
Bin CMV 7
Schematicrepresentation of CMV anti-sensegene construction.Only one of the three CMV-derivedDNA fragments(fragment 7) is shownas an example.The fragementwas isolated from M13 RF DNA and ligatedinto the expressionvectorP35S-Nos. The resulting plasmid P35S CMV 7 was linearizedand subclonedinto the Ti plasmid Bin 19 to produce Bin CMV 7. Abbreviations:B, Bam HI; E, Eco RI, H, Hind 11I; LB, RB, left and right T-DNA border repeats; Nos, nopaline synthasepolyadenylationsite. Fig. 2.
Plants Transgenic tobacco plants (Nicotiana tabacum cv. Samsun) were maintained in vitro on Murashige and Skoog (MS) medium [28] containing 500/~g/ml cefotaxime and 100 # g / m l kanamycin. Nontransformed plants were grown on MS alone. Plants were maintained in a growth r o o m under illumination of 50/zE m - 2 s -1 on a 15 h light (26°C), 9 h dark (20°C) cycle. Shoot tips (about 2 cm) were transferred to fresh medium every 3 - 4 weeks for sub-culturing or propagation. Cefotaxime was omitted from the medium after several rounds of subculturing. For infectivity tests, plants approximately 5 cm tall were potted in soil and covered with plastic bags. The plants were maintained in a growth r o o m at 23 ___ 2 ° C and illuminated 18 hours per day at 200 #E m -2 s -l. Plastic bags were opened gradually to harden the plants between the third and eighth days when they were inoculated. Cucumber (Cucumis sativus cv. Supermarket)
seedlings were grown in the glasshouse and inoculated with CMV at the cotyledon stage. The inoculated plants were maintained in the growth room as specified above.
C M V inoculation The Q strain of CMV [7] was grown [18] and the virus was purified [30]. A virus preparation at a concentration of 10 m g / m l in 5 m M sodium borate, 0.5 m M EDTA, p H 9, was mixed with an equal volume of glycerol, divided into 50 #1 aliquots, snapfrozen in liquid nitrogen and maintained at - 70 ° C. The virus was diluted to the required concentrations with the borate-EDTA buffer just before use. Virus inoculations were performed mechanically with the aid of c a r b o r u n d u m dust by applying 20/zl of virus suspension onto each leaf. Four small leaves of each tobacco plant were inoculated and these were not harvested at the time of sampling for virus assays.
466 In all experiments, virus infections were also assayed by inoculating tobacco tissue extracts onto cucumber cotyledons before doing dot blot assays. Tobacco leaf samples were ground in 20 mM phosphate buffer pH 7 containing 1 mM EDTA (0.5 ml per g tissue) and inoculated onto 8 cucumber cotyledons. Under the condition specified for plant growth, infected cotyledons produced distinct lesions about 48 h after inoculation. This procedure did not produce quantitative data but was reliable for testing the presence or absence of the virus.
RNA extraction Total nucleic acids were extracted by grinding tobacco leaf samples in 4 ml of extraction buffer (0.1 M Tris pH 9, 0.1 M NaC1, 10 mM EDTA, 1070SDS and 0.1070 2-mercaptoethanol), 2 ml of phenol and 2 ml of chloroform per g tissue. The aqueous phase recovered by centrifugation was re-extracted twice with phenol. Of each extract 200/~1 was mixed with 200/~1 of 4 M ammonium acetate and 800/xl of ethanol. The RNA was recovered by centrifugation, dried and suspended in 25/zl of 10 mM Tris pH 7.5 and 0.1 mM EDTA.
was diluted to four-fold with H20, applied onto nitrocellulose membrane and hybridized [38] with minus-sense 32p-DNA probes.
Results
Regions of C M V RNA used for anti-sense RNA gene construction The locations of the three regions of CMV genome used for anti-sense gene construction are shown in Fig. 1. Fragments 20 and 7 cover parts of the leader sequences and reading frames of the viral RNA 1 and 3 respectively. Fragments 20 and 7 also overlap the reading frames of CMV-encoded proteins 1 and 3A (Fig. 1) which have been implicated, on the basis of sequence homologies with other viral proteins [29, 32] to be involved in RNA synthesis [13, 32] and virus translocation from cell to cell [4, 13], respectively. Fragment 1 (Fig. 1) also contains a regulatory sequence conserved in all CMV RNAs and a part of it is common to the closely related brome mosaic virus (BMV) RNA [1]. The function of the 3' conserved region of CMV has not been determined but the corresponding region in BMV RNA is the site of minus-strand RNA initiation [27].
Northern blot analysis and dot blot hybridization Total tobacco RNA derived from 30 mg fresh weight of each tissue sample was incubated with formaldehyde and electrophoresed through 1.5 °70 agarose gels containing formaldehyde as described [24]. The Northern transfers and subsequent handling of the filters were as described by Thomas [38]. Hybridization probes were prepared by transcribing DNA [24] from appropriate M13 clones using the Klenow fragment of DNA polymerase and the universal 17 mer M13 primer. The DNA probes had a specific activity of approximately 1.8 × 109 cpm/#g and were used at a concentration of 106 cpm/ml o f hybridization solution. After washing the filters, they were exposed to X-ray films with Du Pont hi-plus intensifying screens for a period of 24 h (constructs 7), 6 days (constructs 1) or 18 days (constructs 20). For dot blot detection of CMV, 1/zl of each RNA
Expression of C M V anti-sense RNA in the transgenic plants Between 12 and 20 plants transformed with each of the three gene constructs were selected on kanamycin and grown separately. As an additional selection step all plants were regenerated once more from leaf pieces on MS medium containing kanamycin and cefotaxime. The transformed plants had normal appearance. RNA was isolated from leaf samples and examined for the presence of CMV anti-sense RNA by Northern blot analysis. The blots were probed with plus-sense M13 transcripts corresponding to CMV sequences in the anti-sense gene constructs. Results of Northern blots of RNA from a number of plants transformed with each construct are shown in Fig. 3. Anti-sense CMV sequences were detected in all the three groups of plants as single RNA bands.
467
Fig. 3. Northern blot analysis of RNA from transgenic plants. Total RNA preparations from tobacco plants transformed with anti-sense gene constructs 1, 7 and 20 were analysed in a 1.5% agarose gel and blotted onto nitrocellulose. The membrane was cut into three pieces, each containing RNA from six different plants (lanes 1- 6) transformed with one of the three constructs and hybridized to their respective 32p-probes. The probes were plus sense M13 transcripts of regions 1, 7 and 20 shown in Fig. 1. RNA samples from the nontransformed parental line was also used as a control (P). Subsequent to autoradiography, staining of the filters [24] showed that all tracks contained similar quantities of total leaf nucleic acid (not shown).
The levels o f anti-sense R N A varied a m o n g the plants within each g r o u p (Fig. 3) and was undetectable in 11 out o f 43 kanamycin-resistant plants (examples are shown in Fig. 3, panel 1, lane 1, panel 7, lane 3; and panel 20, lane 4). N o n e o f the probes hybridized to R N A f r o m n o n - t r a n s f o r m e d plants (Fig. 3). The electrophoretic mobility o f the three antisense R N A s was slightly different (Fig. 3) and the difference correlated with the relative sizes o f antisense regions used (Fig. 1). The sizes o f anti-sense R N A s (about 1300 bases) indicated that they contained the additional sequence derived f r o m the polyadenylation region in the gene constructs. F r o m the exposure times required to observe the three anti-sense R N A bands and the b a n d intensities it was estimated that the levels o f anti-sense R N A in the best expressing plants f r o m the groups 20, 1 and 7 were in the ratio o f 1:15:500 respectively, and these ratios approximate the relative expression o f each group. The reasons for such variability o f R N A expression is not known. Since each construct had the same regulatory sequences, and differed only in antisense sequences, our N o r t h e r n data (Fig. 3) suggest that one factor affecting the level o f R N A expression is the structure o f the sequence inserted in the plant genome.
C M V infectivity tests using the transgenic tobacco plants The Q strain o f C M V does not induce readily discernable s y m p t o m s in S a m s u n t o b a c c o but it replicates in the inoculated plants in a systemic manner. In a preliminary test n o n - t r a n s f o r m e d plants were inoculated with C M V (5/zg/ml) and extracts o f the expanding, non-inoculated leaves were examined for the presence o f the virus by inoculation o n t o cucumber cotyledons. The virus was readily detectable in t o b a c c o 6 days after its infection. In subsequent experiments systemic C M V infections in S a m s u n t o b a c c o were assayed 8 days after inoculation, under the conditions specified, which favour rapid virus growth. For infectivity tests with the transgenic plants, a plant from each o f the three groups expressing high levels o f the respective anti-sense R N A , was selected as well as a non-expressing and a n o n - t r a n s f o r m e d plant. These five plants were propagated by subculturing and each o f the 5 plant lines produced was inoculated with C M V at concentrations o f 0.5, 5, and 50 #g/ml. The plants were maintained in a constant temperature r o o m and samples were taken f r o m non-inoculated expanding leaves after 8 days for dot-blot assays with a C M V probe. Results
468
Fig. 4. Detection of CMV R N A in inoculated transgenic and control tobacco plants by dot blot hybridization. The tobacco lines 1, 7, and 20-9 were obtained by propagating an individual plant regenerant transformed with construct 1, 7 and 20 respectively. A transformed kanamycin resistant line which did not express any anti-sense RNA (C) and the non-transformed parental line (P) were also used as controls. CMV inoculum concentrations are indicated in A. In B the experiment shown in A was repeated but a single inoculum concentration of 5 / z g / m l was used. RNA extracts of non-inoculated plants were also used (H). The minus-sense 32p-probe was the M13 transcript of residue 629 to 773 of CMV RNA 2. Similar results were obtained when a probe to R N A 1 (residue 736 to 1088) or to the conserved 3' -region (residue 2478 to 2900) was used. The vertical rows in each group correspond to tests of separate plants.
shown in Fig. 4A demonstrated that only low levels o f CMV were detectable in plants inoculated with 0.5/~g CMV/ml but all plants inoculated with 5 and 50/zg CMV/ml contained higher virus levels except for two of the plants inoculated with 5/~g/ml CMV in the plant line which expressed the anti-sense RNA 20 (Fig. 4A). The location of this RNA related to CMV genome is shown in Fig. 1. In a subsequent experiment (Fig. 4B) the infectivity test with 5 mg/ml of CMV inoculum was repeated and it was observed that there was no distinct difference between the levels of virus in the control plants and plants expressing anti-sense RNA 1 and 7. Again, plants in the 20-9 group contained a noticeably lower virus concentration than the controls and the other transgenic lines (Fig. 4B). In addition to dot blotting we tested tobacco infections by inoculating their extracts onto cucumber cotyledons as described in Materials and methods. Results of the biological test in cucumber confirmed those of dot blots. As an example, the results of cucumber infections with the tobacco extracts used in Fig. 4A are given in Table 1. The cucumber infectivity results also indicate that the infected tobacco plants contained encapsidated CMV RNA because
Table 1. Infectivity tests of CMV-inoculated tobacco extracts on cucumber cotyledons. CMV inoculum (#g/ml)
Tobacco line 1
7
20.9
C
P
0.5 1
-
-
-
+ + +
-
+ + +
+ + +
+
+
5.5
+
+
+ +
+ +
+ +
+ +
+ +
+ + + 50
Individual tobacco plants used in Fig. 4A were tested on 8 cucumber cotyledons. - denotes no lesion, + means an average of more than 10 lesion per inoculated cotyledon. The average lesion numbers between 0 and 10 are given.
the naked viral RNA would rapidly lose infectivity in the crude tobacco extracts. These results indicate that presence of anti-sense RNA 1 and RNA 7, whose potential binding loca-
469 tions to CMV genome are shown in Fig. l, did not induce significant inhibition of virus synthesis. However, the tobacco line 20-9 which expressed 500fold lower levels o f anti-sense RNA compared to plant number 7 showed resistance to an inoculum of 5/~g/ml CMV, a level which infected plants from lines 1 and 7. We then examined whether the observed CMV resistance in the plant line 20-9 would correlate with the presence of anti-sense RNA 20 in other independently transformed lines. From the Northern blot results, three other lines were identified which expressed the same anti-sense RNA at similar or higher levels and used in an infectivity test. Results o f this test (Fig. 5) showed that only plants in line 20-9 were less sensitive to CMV infection. Non-transformed plants and the other lines expressing the anti-sense RNA 20 were similar to the parental line in supporting CMV replication. These results suggested that the occurrence of resistance to CMV in plant 20-9 was not correlated in a straight-forward manner with the presence o f anti-sense RNA 20. A characteristic of the resistant tobacco line 20-9 was that root initiation growth on MS medium supplemented with kanamycin was approximately 3 times slower than that observed for the other trans-
Fig. 5. C o m p a r i s o n o f C M V sensitivity o f different tobacco lines transformed with the gene construct 20 by dot blot hybridization. The transgenic tobacco line 20-9 used in Fig. 4 was compared with three other independently transformed tobacco lines which also expressed the anti-sense R N A 20 and with the non-transformed parental line (P). Probe hybridization was done as in Fig. 4. The vertical rows in each group correspond to tests of separate plants. The levels of anti-sense R N A expression of the lines 20.1, 20.3, 20.9 and 20.11 are shown in Fig. 3, panel 20, lanes 1, 3, 2 and 6 respectively.
genic plants. However, in the absence of kanamycin and when transferred to soil the growth rate of 20-9 was similar to the other transgenic plants. The possibility was considered that the level of anti-sense RNA 20 in the tobacco line 20-9 could be significantly higher under the conditions o f growth in soil, as used for infectivity tests, and the enhanced RNA level would produce resistance. Northern blotting of RNA samples from plants of 20-9 grown in soil to the stage when inoculation would be done and plants grown on kanamycin revealed that the levels of antisense RNA in the plants were similar, irrespective of the different growth conditions.
Discussion
We have found that anti-sense RNA to specific regions of CMV genome was expressed in transgenic Samsun tobacco but did not routinely confer resistance to the virus. In our experiments one line of transgenic plants transformed with construct 20 showed a distinct resistance to CMV infection when an inoculum concentration of 5/~g/ml was used. This line (20-9) produced relatively low levels of antisense RNA, and because the occurrence of resistance did not correlate with the presence of this RNA in other tobacco lines containing the same anti-sense RNA as detected by Northern hybridization, we cannot conclude that the observed resistance is due to the expression of anti-sense RNA 20-9. Reasons for the observed differences in resistance can only be speculated. For example, variability in the site(s) of DNA insertion into the tobacco genome during the transformation event, somaclonal variation (22) or other unknown factors could be involved. In most cases reported to date, anti-sense RNAs have been directed against specific mRNAs and have resulted in inhibition of gene expression [11, 26, 33]. The genome of some plus-strand RNA viruses, such as CMV RNA, have a number of characteristics in c o m m o n with eukaryotic m R N A [10, 31, 35, 36, 37]. However, there are differences between the modes of synthesis o f viral RNA and mRNA and there may be some characteristics of the plant viral replicative process which actively discourages stables doublestranded RNA formation involving anti-sense RNA. Unlike mRNA, viral RNA synthesis occurs by
470 replication in a process which is linked with uncoating, translation and encapsidation [36]. In the case o f the better studied tobacco mosaic virus, whose R N A has end structures similar to CMV R N A [9, 32], it has been proposed that in vivo [35] virus disassembly begins by the exposure of 5 ' - e n d of R N A which binds to the ribosomes in a process which has been termed "cotranslational disassembly". Plusstrand viral R N A also forms complexes with the viral replicase components [36]. Presence of such translation and replication complexes may render the target viral R N A inaccessible to the anti-sense RNA, such as those tested here. It is also possible that CMV synthesis and antisense R N A synthesis may occur in different cellular compartments. Replication of CMV is thought to occur in the plant cytoplasm [14] and anti-sense R N A synthesized in the nucleus would need to be translocated to the cytoplasm to be effective [21]. Rearrangements of the inserted sequences during transformation or the presence of secondary structures in the RNAs which contain the additional sequence derived from the polyadenylation signal may also account for the observed lack of CMV antisense R N A activity. However, the results of Northern blotting (Fig. 3) indicate that the anti-sense RNAs are capable of hybridizing to the plus-sense RNAs. A requirement for effective binding of anti-sense R N A is that it should be present at concentrations in excess of the target sequence. In our experiments, the leaves above inoculation site which lacked any CMV at the time of inoculation contained a varying level of the three anti-sense RNAs. The level of antisense R N A 7 was relatively high and was comparable to the level of the corresponding CMV R N A in systemically infected leaves 8 days post infection. Both these RNAs were detectable by hybridization with similarly made probes after 24 h exposure of the hybridized filters. This observation indicates that an excess of the anti-sense R N A was present in the systemically infected leaves during initiation of infection. Recently, transgenic plants expressing anti-sense coat protein gene of CMV [12] and potato virus [15] have been found to inhibit the replication of their respective viruses at low inoculum levels. The anti-
sense CMV coat protein gene used in the first study above contained the entire 3'-conserved region which occurs in all CMV RNAs [32]. As has been pointed out by Guozzo et al. [12], although the observed inhibition could involve anti-sense RNA binding with CMV RNA, there is a possibility that CMV anti-sense coat protein gene can compete for the viral or host factors involved in the replication of CMV R N A from the minus-strand templates. The potential of anti-sense R N A as means of plant virus control has been discussed by a number of authors [reviews 8, 20, 25, 34, 40]. Our results with the three gene constructs used do not encourage the use of anti-sense R N A as an approach to CMV control. However, the three R N A regions used in this study together constitute about 12% of the CMV genome and it is possible that anti-sense R N A to other regions of CMV RNA may be active, although studies with eukaryotic systems to not reveal a correlation between the location from which anti-sense R N A is derived and its activity [11].
Acknowledgements We thank Dr M. Bevan for providing the Bin 19 Binary construct, Dr D. Llewellyn for the 35S promoter construct, Dr N. Scott and Dr R. Francki for discussions and Susan Johnson and Les Krake for technical assistance.
References 1. Ahlquist P, Dasgupta R, Kaesberg P: Near identity of 3' R N A secondary structure in bromoviruses and cucumber mosaic virus. Cell 2 3 : 1 8 3 - 1 8 9 (1981). 2. Bevan M: Binary Agrobacterium vectors for plant transformation. Nucl Acids Res. 12." 8711-8721 (1984). 3. Coleman J, Hirashima A, Inokuchi Y, Green PJ, Inouye M: A novel i m m u n e system against bacteriophage infection using complementary R N A (micRNA). Nature 3 1 5 : 6 0 1 - 6 0 3 (1985). 4. Deom CM, Oliver M J, Beachy RN: The 30-kilodalton gene product of tobacco mosaic virus potentiates virus movement. Science 2 3 7 : 3 8 9 - 3 9 4 (1984). 5. Ditta G, Stanfield S, Corbin D, Helinski DR: Broad host range D N A cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium melitoti. Proc Natl Acad Sci USA 7 7 : 7 3 4 7 - 7 3 5 1 (1980).
471 6. Ecker JR, Davis RW: Inhibition of gene expression in plant cells by expression of anti-sense RNA. Proc Natl Acad Sci USA 83:5372-5376 (1986). 7. Francki RIB, Randles JW, Chambers TC, Wilson JB: Some properties of purified cucumber mosaic virus. Virology 28: 729-741 (1966). 8. Gibbs A, Skotnicki A: Strategic defence initiatives; plants, viruses and genetic engineering. In: Varma A, Verma JP (eds) Vistas in Plant Pathology, pp 467-480. Malhotra Publications, New Delhi (1986). 9. Goelet P, Lomonossoff GP, Bulter PJG, Akam ME, Gait M J, Karn J: Nucleotide sequence of tobacco mosaic virus RNA. Proc Natl Acad Sci USA 79:5818-5822 (1982). 10. GouldAR, SymonsRH:Cucumber mosaic virus RNA 3. Determination of the nucleotide sequence provides the amino acid sequences of protein 3A and viral coat protein. Eur J Biochem 126:217-226 (1982). 11. Green PJ, Pines O, Inouye M: The role of anti-sense RNA in gene regulation. Ann Rev Biochem 55:569-597 (1986). 12. Guozzo M, O'Connell KM, Kaniewski W, Fang RX, Chua NH, Turner NE: Viral protection in transgenic plants expressing the cucumber mosaic virus coat protein or its antisense RNA. Biotechnology 6:549-557 (1988). 13. Haseloff J, Goelef P, Zimmern D, Ahlquist P, Dasgupta R, Kaesberg P: Striking similarities in amino acid sequence among nonstructural proteins encoded by RNA viruses that have dissimilar genomic organization. Proc Natl Acad Sci USA 81:4358-4362 (1984). 14. Hatta T, Francki RIB" Cytoplasmic structures associated with tonoplasts of plant cells infected with cucumber mosaic and tomato aspermy viruses. J Gen Viro153:343 - 346 (1981). 15. Hemenway C, Fang RX, Kaniewski WK, Chua NH, Tumer NE: Analysis of the mechanism of protection in transgenic plants expressing the potato virus coat protein or its antisense RNA. EMBO J 7 : 1 2 7 3 - 1 2 8 0 (1988). 16. Horsch RB, Fry JE, Hoffmann NL, Eichholtz D, Rogers SG, Fraley RT: A simple and general method for transferring genes into plants. Science 227:1229-1231 (1985). 17. Izant JG, Weintraub H: Constitutive and conditional suppression of exogenous and endogenous genes by anti-sense RNA. Science 229:345-352 (1985). 18. Jaspars EMJ, Gill DS, Symons RH: Viral RNA synthesis by a particulate fraction from cucumber seedlings infected with cucumber mosaic virus. Virology 144:410-425 (1985). 19. Kawasaki ES: Quantitative hybridization arrest of mRNA in Xenopus oocytes using single-strandedcomplementary DNA or oligonucleotide probes. Nucl Acids Res 13: 4991-5004 (1985). 20. Kerr A: The impact o f molecular genetics on plant pathology. Ann Rev Phytopathol 25:87-110 (1987). 21. Kim SK, Wood BJ: Stable reduction of thymidine kinase activity in cells expressing high levels of anti-sense RNA. Cell 42:129-138 (1985). 22. Larkin PL, Scowcroft WR: Somaclonal variation- a novel source of variability from cell cultures for plant improvement. Theor Appl Genet 60:197-214 (1981). 23. Loesch-Fries LS, Halk E, Merlo D, Jarvis N, Nelson S, Krahn K, Burhop L, Arntzen C J, Ryan C, in: Molecular strategies
for crop protection, Expression of alfalfa mosaic virus coat protein gene and anti-sense cDNA in transformed tobacco tissue. UCLA Symp Proc Mol Cell Biol New Set 48: 221- 234 (1987). 24. Maniatis T, Fritsch EF, Sambrook J: Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratories, Cold Spring Harbor, NY (1982). 25. Matthews REF: The changing scene in plant pathology. Ann Rev Phytopathol 25:10-23 (1987). 26. Melton DA: Injected anti-sense RNAs specifically block messenger RNA translation in vivo. Proc Natl Acad Sci USA 82: 144-148 (1985). 27. Miller WA, Bujarski J J, Dreher TW, Hall TC: Minus-strand initiation by brome mosaic virus replicase within the 3' tRNA-like structure of native and modified RNA templates. J Mol Biol 187:537-546 (1986). 28. Murashige T, Skoog F: A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 15: 473-497 (1962). 29. Murthy MRN: Comparison of the nucleotide sequences of cucumber mosaic virus and brome mosaic virus. J Mol Biol 168:469-475 (1983). 30. Peden KWC, Symons RH: Cucumber mosaic virus contains a functionally divided genome. Virology 53:487 - 492 (1973). 31. Rezaian MA, Williams RHV, Gordon KHJ, Gould AR, Symons RH: Nucleotide sequence of cucumber-mosaic-virus RNA 2 reveals a translation product significantly homologous to corresponding proteins of other viruses. Eur J Biochem 143:277-284 (1984). 32. Rezaian MA, Williams RHV, Symons RH: Nucleotide sequence of cucumber mosaic virus RNA 1. Presence of a sequence complementary to part of the viral satellite RNA and homologies with other viral RNAs. Eur J Biochem 150: 331-339 (1985). 33. Rosenberg UB, Preiss A, Seifert E, Jackle H, Knipple DC: Production of phenocopies by Kruppel anti-sense RNA injection into Drosophila embryos. Nature 313:703-706 (1985). 34. Sanford JC, Johnston SA: The concept of parasite-derived resistance - Deriving resistance genes from the parasite's own genome. J Theor Biol 113:395-405 (1985). 35. Shaw JG, Plaskitt KA, Wilson TMA: Evidence that tobacco mosaic virus particles disassemble cotranslationally in vivo. Virology 148:326-336 (1986). 36. Strauss EG, Strauss JH: Replication strategies for single stranded RNA viruses of eukaryotes. Curr Top Microbiol Immunol 105:1-98 (1983). 37. Symons RH: Cucumber mosaic virus RNA contains 7methyl guanosine at the 5' -terminus of all four RNA species. Mol Biol Reports 2:277-285 (1975). 38. Thomas PS: Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Natl Acad Sci USA 77: 5201-5205 (1980). 39. Weintraub H, Izant JG, Harland RM: Anti-sense RNA as a molecular tool for genetic analysis. Trends Genet. January: 22-25 (1985). 40. Zaitlin M, Hull R: Plant virus - host interactions. Ann Rev Plant Physiol 38:291-315 (1987).
463
Anti-sense R N A s of cucumber mosaic virus in transgenic plants assessed for control of the virus M. Ali Rezaian, Ken G. M. Skene and Jeff G. Ellis 1
CSIRO Division of Horticulture, P.O. Box 350, Adelaide, S.A. 5001, Australia; 1CSIRO Division of Plant Industry, P.O. Box 1600, Canberra, A.C.T. 2601, Australia Received 19 April 1988; accepted in revised form 6 July 1988
Key words: anti-sense RNA, cucumber mosaic virus, transgenic plants, virus inhibition
Abstract
Three synthetic genes for the production of anti-sense RNA to different regions o f the cucumber mosaic virus (CMV) genome were constructed using virus-derived double-stranded cDNA coupled to a promoter sequence from cauliflower mosaic virus. The genes were used to transform tobacco plants by a Ti plasmid vector. Transgenic plants obtained with the three constructs produced anti-sense RNA at different levels. Plants expressing each of the three anti-sense RNAs were inoculated with CMV and their sensitivity to the virus infection was compared with the non-transformed plants. Only one plant line which expressed relatively low levels of one of the anti-sense RNAs showed resistance to CMV but other plants expressing the same or the other two antisense RNAs had similar sensitivity to CMV infection as the non-transformed plants.
Introduction
In recent years anti-sense RNA (RNA complementary to mRNA) has been shown to inhibit gene expression in a variety of prokaryotic and eukaryotic ceils [21, 39, reviewed in 11] including plants [6]. The mechanisms by which anti-sense RNA inhibits gene expression are not well understood and different modes of inhibition have been described. In eukaryotic systems the formation of double-stranded RNA in the nucleus may block mRNA transport to the cytoplasm [21], or the RNA-RNA hybrids formed in the nucleus may be rapidly degraded [26]. Anti-sense RNA may inhibit m R N A translation bound to polysomes [26, 33]. The anti-sense sequences required for the inhibition of gene expression can be relatively short because a 52 residue RNA has been found effective in inhibiting the thymidine kinase gene [17]. Even injection of anti-
sense oligonucleotides (18-23 residues) into Xenopus oocytes can inhibit translation of the target RNA [19]. Because the genomes of plus strand RNA viruses have structural similarities to mRNA and function as mRNA during virus replication [36], it has been suggested that anti-sense RNA may inhibit virus replication [341. Coleman et al. [3] have examined this idea with anti-sense RNA to an RNA phage and reported that induction of anti-sense RNA (referred to as micRNA) directed against the coat protein and/or replicase genes ofE. coli RNA phage SP prevents the phage multiplication. The use of anti-sense RNA against a plant virus has not been examined before. Loesch-Fries et al. [23] reported the synthesis of anti-sense RNA of alfalfa mosaic virus (AMV) in transgenic plants but did not report any anti-viral activity. We have tested the possibility of anti-sense RNA inhibition of cu-
464 cumber mosaic virus (CMV) which is structurally similar to AMV [31, 32] and an important pathogen with wide host range. Our results indicate that, except for one tobacco line, CMV replication was not significantly affected by the expression of anti-sense RNA in transgenic plants.
Materials and methods
Construction of C M V anti-sense genes Double-stranded DNA fragments corresponding to three regions of the CMV genome (Fig. 1) were excised from cloned CMV sequences contained in the replicative form of M13 phage DNA. Fragment 1 from CMV RNA 2 and fragment 20 from CMV RNA 1 (Fig. 1) were from CMV M13mp7 clones prepared previously [31, 32] and fragment 7 (Fig. 1) was from an M13mp9 clone of RNA 3 kindly provided by C. Davies of Adelaide University. An Eco RIBam HI fragment of clone 7 was inserted in Eco RIBarn HI cut plasmid p35S-Nos (provided by Dr D. LleweUyn o f CSIRO) and Eco RI fragments of
clones 1 and 20 were ligated into Eco RI cut p35SNos. This plant expression vector contains the promoter from the 35S protein gene of cauliflower mosaic virus and a 1 kb polyadenylation signal from the nopaline synthase gene. The resulting plasmid construct containing fragment 7 had an insert orientation required for anti-sense RNA synthesis, but constructs 1 and 20 in the anti-sense orientations were identified by restriction enzyme analysis. The procedure for construction of Ti plasmids was the same for the three CMV sequences and this is shown in Fig. 2 for the construct 7 as an example. Each of the three plasmids was cut at a single site and inserted in the binary Ti vector Bin 19 [2]. The Ti plasmid constructs were mobilized from E. coli HB101 to Agrobacterium tumefaciens LBA4404 by a triparental mating procedure using the mob functions of plasmid RK2013 [5]. The presence of CMV gene constructs in the Ti plasmids was confirmed by restriction enzyme analysis of the plasmid DNAs purified from A. tumefaciens. Plant transformations were carried out by co-cultivation of tobacco leaf discs [16] with Agrobacterium carrying the binary vector [2].
1
RNA
3389
b-----
1 --~ o 70
4 ----Antisense 355
R N A 20
1 RNA
3035
2 Antisense RNA 1---~a 2478 1
RNA
3
4 2900 2193
--I .... CP
3A A n t i s e n s e R N A 7 -'-~31I
I
236
RNA
4
1 --~
1027
CP Fig. 1. Genome organization of CMV. Locations of the three regions used for construction of anti-sense RNA genes are shown below CMV RNAs 1, 2 and 3. The size of each RNA and positions of the anti-sense RNAs relative to CMV RNAs are indicated by residue numbers. The open boxes depict the reading frames and the solid boxes show the 3' -end conserved regions of the viral genome. Data for RNA 1 and 2 are from refs. 32 and 31 and for RNA 3 are from ref. 10.
465
~~~H.j;~=:~ #NOS
promoter ~r.jE
[
B
"/
CMV insert
E
promoter ,,f CMV insert
[ :~-E polylinker { p35SCMV 7 )l] LH ~N,~ ~/'NOS LB ~
~ B i
H 19~ ~
~
LB
/
H
~"'~--""-"4~
~ C M V
/
(
n
promoter
\\\
insert
NOS
Bin CMV 7
Schematicrepresentation of CMV anti-sensegene construction.Only one of the three CMV-derivedDNA fragments(fragment 7) is shownas an example.The fragementwas isolated from M13 RF DNA and ligatedinto the expressionvectorP35S-Nos. The resulting plasmid P35S CMV 7 was linearizedand subclonedinto the Ti plasmid Bin 19 to produce Bin CMV 7. Abbreviations:B, Bam HI; E, Eco RI, H, Hind 11I; LB, RB, left and right T-DNA border repeats; Nos, nopaline synthasepolyadenylationsite. Fig. 2.
Plants Transgenic tobacco plants (Nicotiana tabacum cv. Samsun) were maintained in vitro on Murashige and Skoog (MS) medium [28] containing 500/~g/ml cefotaxime and 100 # g / m l kanamycin. Nontransformed plants were grown on MS alone. Plants were maintained in a growth r o o m under illumination of 50/zE m - 2 s -1 on a 15 h light (26°C), 9 h dark (20°C) cycle. Shoot tips (about 2 cm) were transferred to fresh medium every 3 - 4 weeks for sub-culturing or propagation. Cefotaxime was omitted from the medium after several rounds of subculturing. For infectivity tests, plants approximately 5 cm tall were potted in soil and covered with plastic bags. The plants were maintained in a growth r o o m at 23 ___ 2 ° C and illuminated 18 hours per day at 200 #E m -2 s -l. Plastic bags were opened gradually to harden the plants between the third and eighth days when they were inoculated. Cucumber (Cucumis sativus cv. Supermarket)
seedlings were grown in the glasshouse and inoculated with CMV at the cotyledon stage. The inoculated plants were maintained in the growth room as specified above.
C M V inoculation The Q strain of CMV [7] was grown [18] and the virus was purified [30]. A virus preparation at a concentration of 10 m g / m l in 5 m M sodium borate, 0.5 m M EDTA, p H 9, was mixed with an equal volume of glycerol, divided into 50 #1 aliquots, snapfrozen in liquid nitrogen and maintained at - 70 ° C. The virus was diluted to the required concentrations with the borate-EDTA buffer just before use. Virus inoculations were performed mechanically with the aid of c a r b o r u n d u m dust by applying 20/zl of virus suspension onto each leaf. Four small leaves of each tobacco plant were inoculated and these were not harvested at the time of sampling for virus assays.
466 In all experiments, virus infections were also assayed by inoculating tobacco tissue extracts onto cucumber cotyledons before doing dot blot assays. Tobacco leaf samples were ground in 20 mM phosphate buffer pH 7 containing 1 mM EDTA (0.5 ml per g tissue) and inoculated onto 8 cucumber cotyledons. Under the condition specified for plant growth, infected cotyledons produced distinct lesions about 48 h after inoculation. This procedure did not produce quantitative data but was reliable for testing the presence or absence of the virus.
RNA extraction Total nucleic acids were extracted by grinding tobacco leaf samples in 4 ml of extraction buffer (0.1 M Tris pH 9, 0.1 M NaC1, 10 mM EDTA, 1070SDS and 0.1070 2-mercaptoethanol), 2 ml of phenol and 2 ml of chloroform per g tissue. The aqueous phase recovered by centrifugation was re-extracted twice with phenol. Of each extract 200/~1 was mixed with 200/~1 of 4 M ammonium acetate and 800/xl of ethanol. The RNA was recovered by centrifugation, dried and suspended in 25/zl of 10 mM Tris pH 7.5 and 0.1 mM EDTA.
was diluted to four-fold with H20, applied onto nitrocellulose membrane and hybridized [38] with minus-sense 32p-DNA probes.
Results
Regions of C M V RNA used for anti-sense RNA gene construction The locations of the three regions of CMV genome used for anti-sense gene construction are shown in Fig. 1. Fragments 20 and 7 cover parts of the leader sequences and reading frames of the viral RNA 1 and 3 respectively. Fragments 20 and 7 also overlap the reading frames of CMV-encoded proteins 1 and 3A (Fig. 1) which have been implicated, on the basis of sequence homologies with other viral proteins [29, 32] to be involved in RNA synthesis [13, 32] and virus translocation from cell to cell [4, 13], respectively. Fragment 1 (Fig. 1) also contains a regulatory sequence conserved in all CMV RNAs and a part of it is common to the closely related brome mosaic virus (BMV) RNA [1]. The function of the 3' conserved region of CMV has not been determined but the corresponding region in BMV RNA is the site of minus-strand RNA initiation [27].
Northern blot analysis and dot blot hybridization Total tobacco RNA derived from 30 mg fresh weight of each tissue sample was incubated with formaldehyde and electrophoresed through 1.5 °70 agarose gels containing formaldehyde as described [24]. The Northern transfers and subsequent handling of the filters were as described by Thomas [38]. Hybridization probes were prepared by transcribing DNA [24] from appropriate M13 clones using the Klenow fragment of DNA polymerase and the universal 17 mer M13 primer. The DNA probes had a specific activity of approximately 1.8 × 109 cpm/#g and were used at a concentration of 106 cpm/ml o f hybridization solution. After washing the filters, they were exposed to X-ray films with Du Pont hi-plus intensifying screens for a period of 24 h (constructs 7), 6 days (constructs 1) or 18 days (constructs 20). For dot blot detection of CMV, 1/zl of each RNA
Expression of C M V anti-sense RNA in the transgenic plants Between 12 and 20 plants transformed with each of the three gene constructs were selected on kanamycin and grown separately. As an additional selection step all plants were regenerated once more from leaf pieces on MS medium containing kanamycin and cefotaxime. The transformed plants had normal appearance. RNA was isolated from leaf samples and examined for the presence of CMV anti-sense RNA by Northern blot analysis. The blots were probed with plus-sense M13 transcripts corresponding to CMV sequences in the anti-sense gene constructs. Results of Northern blots of RNA from a number of plants transformed with each construct are shown in Fig. 3. Anti-sense CMV sequences were detected in all the three groups of plants as single RNA bands.
467
Fig. 3. Northern blot analysis of RNA from transgenic plants. Total RNA preparations from tobacco plants transformed with anti-sense gene constructs 1, 7 and 20 were analysed in a 1.5% agarose gel and blotted onto nitrocellulose. The membrane was cut into three pieces, each containing RNA from six different plants (lanes 1- 6) transformed with one of the three constructs and hybridized to their respective 32p-probes. The probes were plus sense M13 transcripts of regions 1, 7 and 20 shown in Fig. 1. RNA samples from the nontransformed parental line was also used as a control (P). Subsequent to autoradiography, staining of the filters [24] showed that all tracks contained similar quantities of total leaf nucleic acid (not shown).
The levels o f anti-sense R N A varied a m o n g the plants within each g r o u p (Fig. 3) and was undetectable in 11 out o f 43 kanamycin-resistant plants (examples are shown in Fig. 3, panel 1, lane 1, panel 7, lane 3; and panel 20, lane 4). N o n e o f the probes hybridized to R N A f r o m n o n - t r a n s f o r m e d plants (Fig. 3). The electrophoretic mobility o f the three antisense R N A s was slightly different (Fig. 3) and the difference correlated with the relative sizes o f antisense regions used (Fig. 1). The sizes o f anti-sense R N A s (about 1300 bases) indicated that they contained the additional sequence derived f r o m the polyadenylation region in the gene constructs. F r o m the exposure times required to observe the three anti-sense R N A bands and the b a n d intensities it was estimated that the levels o f anti-sense R N A in the best expressing plants f r o m the groups 20, 1 and 7 were in the ratio o f 1:15:500 respectively, and these ratios approximate the relative expression o f each group. The reasons for such variability o f R N A expression is not known. Since each construct had the same regulatory sequences, and differed only in antisense sequences, our N o r t h e r n data (Fig. 3) suggest that one factor affecting the level o f R N A expression is the structure o f the sequence inserted in the plant genome.
C M V infectivity tests using the transgenic tobacco plants The Q strain o f C M V does not induce readily discernable s y m p t o m s in S a m s u n t o b a c c o but it replicates in the inoculated plants in a systemic manner. In a preliminary test n o n - t r a n s f o r m e d plants were inoculated with C M V (5/zg/ml) and extracts o f the expanding, non-inoculated leaves were examined for the presence o f the virus by inoculation o n t o cucumber cotyledons. The virus was readily detectable in t o b a c c o 6 days after its infection. In subsequent experiments systemic C M V infections in S a m s u n t o b a c c o were assayed 8 days after inoculation, under the conditions specified, which favour rapid virus growth. For infectivity tests with the transgenic plants, a plant from each o f the three groups expressing high levels o f the respective anti-sense R N A , was selected as well as a non-expressing and a n o n - t r a n s f o r m e d plant. These five plants were propagated by subculturing and each o f the 5 plant lines produced was inoculated with C M V at concentrations o f 0.5, 5, and 50 #g/ml. The plants were maintained in a constant temperature r o o m and samples were taken f r o m non-inoculated expanding leaves after 8 days for dot-blot assays with a C M V probe. Results
468
Fig. 4. Detection of CMV R N A in inoculated transgenic and control tobacco plants by dot blot hybridization. The tobacco lines 1, 7, and 20-9 were obtained by propagating an individual plant regenerant transformed with construct 1, 7 and 20 respectively. A transformed kanamycin resistant line which did not express any anti-sense RNA (C) and the non-transformed parental line (P) were also used as controls. CMV inoculum concentrations are indicated in A. In B the experiment shown in A was repeated but a single inoculum concentration of 5 / z g / m l was used. RNA extracts of non-inoculated plants were also used (H). The minus-sense 32p-probe was the M13 transcript of residue 629 to 773 of CMV RNA 2. Similar results were obtained when a probe to R N A 1 (residue 736 to 1088) or to the conserved 3' -region (residue 2478 to 2900) was used. The vertical rows in each group correspond to tests of separate plants.
shown in Fig. 4A demonstrated that only low levels o f CMV were detectable in plants inoculated with 0.5/~g CMV/ml but all plants inoculated with 5 and 50/zg CMV/ml contained higher virus levels except for two of the plants inoculated with 5/~g/ml CMV in the plant line which expressed the anti-sense RNA 20 (Fig. 4A). The location of this RNA related to CMV genome is shown in Fig. 1. In a subsequent experiment (Fig. 4B) the infectivity test with 5 mg/ml of CMV inoculum was repeated and it was observed that there was no distinct difference between the levels of virus in the control plants and plants expressing anti-sense RNA 1 and 7. Again, plants in the 20-9 group contained a noticeably lower virus concentration than the controls and the other transgenic lines (Fig. 4B). In addition to dot blotting we tested tobacco infections by inoculating their extracts onto cucumber cotyledons as described in Materials and methods. Results of the biological test in cucumber confirmed those of dot blots. As an example, the results of cucumber infections with the tobacco extracts used in Fig. 4A are given in Table 1. The cucumber infectivity results also indicate that the infected tobacco plants contained encapsidated CMV RNA because
Table 1. Infectivity tests of CMV-inoculated tobacco extracts on cucumber cotyledons. CMV inoculum (#g/ml)
Tobacco line 1
7
20.9
C
P
0.5 1
-
-
-
+ + +
-
+ + +
+ + +
+
+
5.5
+
+
+ +
+ +
+ +
+ +
+ +
+ + + 50
Individual tobacco plants used in Fig. 4A were tested on 8 cucumber cotyledons. - denotes no lesion, + means an average of more than 10 lesion per inoculated cotyledon. The average lesion numbers between 0 and 10 are given.
the naked viral RNA would rapidly lose infectivity in the crude tobacco extracts. These results indicate that presence of anti-sense RNA 1 and RNA 7, whose potential binding loca-
469 tions to CMV genome are shown in Fig. l, did not induce significant inhibition of virus synthesis. However, the tobacco line 20-9 which expressed 500fold lower levels o f anti-sense RNA compared to plant number 7 showed resistance to an inoculum of 5/~g/ml CMV, a level which infected plants from lines 1 and 7. We then examined whether the observed CMV resistance in the plant line 20-9 would correlate with the presence of anti-sense RNA 20 in other independently transformed lines. From the Northern blot results, three other lines were identified which expressed the same anti-sense RNA at similar or higher levels and used in an infectivity test. Results o f this test (Fig. 5) showed that only plants in line 20-9 were less sensitive to CMV infection. Non-transformed plants and the other lines expressing the anti-sense RNA 20 were similar to the parental line in supporting CMV replication. These results suggested that the occurrence of resistance to CMV in plant 20-9 was not correlated in a straight-forward manner with the presence o f anti-sense RNA 20. A characteristic of the resistant tobacco line 20-9 was that root initiation growth on MS medium supplemented with kanamycin was approximately 3 times slower than that observed for the other trans-
Fig. 5. C o m p a r i s o n o f C M V sensitivity o f different tobacco lines transformed with the gene construct 20 by dot blot hybridization. The transgenic tobacco line 20-9 used in Fig. 4 was compared with three other independently transformed tobacco lines which also expressed the anti-sense R N A 20 and with the non-transformed parental line (P). Probe hybridization was done as in Fig. 4. The vertical rows in each group correspond to tests of separate plants. The levels of anti-sense R N A expression of the lines 20.1, 20.3, 20.9 and 20.11 are shown in Fig. 3, panel 20, lanes 1, 3, 2 and 6 respectively.
genic plants. However, in the absence of kanamycin and when transferred to soil the growth rate of 20-9 was similar to the other transgenic plants. The possibility was considered that the level of anti-sense RNA 20 in the tobacco line 20-9 could be significantly higher under the conditions o f growth in soil, as used for infectivity tests, and the enhanced RNA level would produce resistance. Northern blotting of RNA samples from plants of 20-9 grown in soil to the stage when inoculation would be done and plants grown on kanamycin revealed that the levels of antisense RNA in the plants were similar, irrespective of the different growth conditions.
Discussion
We have found that anti-sense RNA to specific regions of CMV genome was expressed in transgenic Samsun tobacco but did not routinely confer resistance to the virus. In our experiments one line of transgenic plants transformed with construct 20 showed a distinct resistance to CMV infection when an inoculum concentration of 5/~g/ml was used. This line (20-9) produced relatively low levels of antisense RNA, and because the occurrence of resistance did not correlate with the presence of this RNA in other tobacco lines containing the same anti-sense RNA as detected by Northern hybridization, we cannot conclude that the observed resistance is due to the expression of anti-sense RNA 20-9. Reasons for the observed differences in resistance can only be speculated. For example, variability in the site(s) of DNA insertion into the tobacco genome during the transformation event, somaclonal variation (22) or other unknown factors could be involved. In most cases reported to date, anti-sense RNAs have been directed against specific mRNAs and have resulted in inhibition of gene expression [11, 26, 33]. The genome of some plus-strand RNA viruses, such as CMV RNA, have a number of characteristics in c o m m o n with eukaryotic m R N A [10, 31, 35, 36, 37]. However, there are differences between the modes of synthesis o f viral RNA and mRNA and there may be some characteristics of the plant viral replicative process which actively discourages stables doublestranded RNA formation involving anti-sense RNA. Unlike mRNA, viral RNA synthesis occurs by
470 replication in a process which is linked with uncoating, translation and encapsidation [36]. In the case o f the better studied tobacco mosaic virus, whose R N A has end structures similar to CMV R N A [9, 32], it has been proposed that in vivo [35] virus disassembly begins by the exposure of 5 ' - e n d of R N A which binds to the ribosomes in a process which has been termed "cotranslational disassembly". Plusstrand viral R N A also forms complexes with the viral replicase components [36]. Presence of such translation and replication complexes may render the target viral R N A inaccessible to the anti-sense RNA, such as those tested here. It is also possible that CMV synthesis and antisense R N A synthesis may occur in different cellular compartments. Replication of CMV is thought to occur in the plant cytoplasm [14] and anti-sense R N A synthesized in the nucleus would need to be translocated to the cytoplasm to be effective [21]. Rearrangements of the inserted sequences during transformation or the presence of secondary structures in the RNAs which contain the additional sequence derived from the polyadenylation signal may also account for the observed lack of CMV antisense R N A activity. However, the results of Northern blotting (Fig. 3) indicate that the anti-sense RNAs are capable of hybridizing to the plus-sense RNAs. A requirement for effective binding of anti-sense R N A is that it should be present at concentrations in excess of the target sequence. In our experiments, the leaves above inoculation site which lacked any CMV at the time of inoculation contained a varying level of the three anti-sense RNAs. The level of antisense R N A 7 was relatively high and was comparable to the level of the corresponding CMV R N A in systemically infected leaves 8 days post infection. Both these RNAs were detectable by hybridization with similarly made probes after 24 h exposure of the hybridized filters. This observation indicates that an excess of the anti-sense R N A was present in the systemically infected leaves during initiation of infection. Recently, transgenic plants expressing anti-sense coat protein gene of CMV [12] and potato virus [15] have been found to inhibit the replication of their respective viruses at low inoculum levels. The anti-
sense CMV coat protein gene used in the first study above contained the entire 3'-conserved region which occurs in all CMV RNAs [32]. As has been pointed out by Guozzo et al. [12], although the observed inhibition could involve anti-sense RNA binding with CMV RNA, there is a possibility that CMV anti-sense coat protein gene can compete for the viral or host factors involved in the replication of CMV R N A from the minus-strand templates. The potential of anti-sense R N A as means of plant virus control has been discussed by a number of authors [reviews 8, 20, 25, 34, 40]. Our results with the three gene constructs used do not encourage the use of anti-sense R N A as an approach to CMV control. However, the three R N A regions used in this study together constitute about 12% of the CMV genome and it is possible that anti-sense R N A to other regions of CMV RNA may be active, although studies with eukaryotic systems to not reveal a correlation between the location from which anti-sense R N A is derived and its activity [11].
Acknowledgements We thank Dr M. Bevan for providing the Bin 19 Binary construct, Dr D. Llewellyn for the 35S promoter construct, Dr N. Scott and Dr R. Francki for discussions and Susan Johnson and Les Krake for technical assistance.
References 1. Ahlquist P, Dasgupta R, Kaesberg P: Near identity of 3' R N A secondary structure in bromoviruses and cucumber mosaic virus. Cell 2 3 : 1 8 3 - 1 8 9 (1981). 2. Bevan M: Binary Agrobacterium vectors for plant transformation. Nucl Acids Res. 12." 8711-8721 (1984). 3. Coleman J, Hirashima A, Inokuchi Y, Green PJ, Inouye M: A novel i m m u n e system against bacteriophage infection using complementary R N A (micRNA). Nature 3 1 5 : 6 0 1 - 6 0 3 (1985). 4. Deom CM, Oliver M J, Beachy RN: The 30-kilodalton gene product of tobacco mosaic virus potentiates virus movement. Science 2 3 7 : 3 8 9 - 3 9 4 (1984). 5. Ditta G, Stanfield S, Corbin D, Helinski DR: Broad host range D N A cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium melitoti. Proc Natl Acad Sci USA 7 7 : 7 3 4 7 - 7 3 5 1 (1980).
471 6. Ecker JR, Davis RW: Inhibition of gene expression in plant cells by expression of anti-sense RNA. Proc Natl Acad Sci USA 83:5372-5376 (1986). 7. Francki RIB, Randles JW, Chambers TC, Wilson JB: Some properties of purified cucumber mosaic virus. Virology 28: 729-741 (1966). 8. Gibbs A, Skotnicki A: Strategic defence initiatives; plants, viruses and genetic engineering. In: Varma A, Verma JP (eds) Vistas in Plant Pathology, pp 467-480. Malhotra Publications, New Delhi (1986). 9. Goelet P, Lomonossoff GP, Bulter PJG, Akam ME, Gait M J, Karn J: Nucleotide sequence of tobacco mosaic virus RNA. Proc Natl Acad Sci USA 79:5818-5822 (1982). 10. GouldAR, SymonsRH:Cucumber mosaic virus RNA 3. Determination of the nucleotide sequence provides the amino acid sequences of protein 3A and viral coat protein. Eur J Biochem 126:217-226 (1982). 11. Green PJ, Pines O, Inouye M: The role of anti-sense RNA in gene regulation. Ann Rev Biochem 55:569-597 (1986). 12. Guozzo M, O'Connell KM, Kaniewski W, Fang RX, Chua NH, Turner NE: Viral protection in transgenic plants expressing the cucumber mosaic virus coat protein or its antisense RNA. Biotechnology 6:549-557 (1988). 13. Haseloff J, Goelef P, Zimmern D, Ahlquist P, Dasgupta R, Kaesberg P: Striking similarities in amino acid sequence among nonstructural proteins encoded by RNA viruses that have dissimilar genomic organization. Proc Natl Acad Sci USA 81:4358-4362 (1984). 14. Hatta T, Francki RIB" Cytoplasmic structures associated with tonoplasts of plant cells infected with cucumber mosaic and tomato aspermy viruses. J Gen Viro153:343 - 346 (1981). 15. Hemenway C, Fang RX, Kaniewski WK, Chua NH, Tumer NE: Analysis of the mechanism of protection in transgenic plants expressing the potato virus coat protein or its antisense RNA. EMBO J 7 : 1 2 7 3 - 1 2 8 0 (1988). 16. Horsch RB, Fry JE, Hoffmann NL, Eichholtz D, Rogers SG, Fraley RT: A simple and general method for transferring genes into plants. Science 227:1229-1231 (1985). 17. Izant JG, Weintraub H: Constitutive and conditional suppression of exogenous and endogenous genes by anti-sense RNA. Science 229:345-352 (1985). 18. Jaspars EMJ, Gill DS, Symons RH: Viral RNA synthesis by a particulate fraction from cucumber seedlings infected with cucumber mosaic virus. Virology 144:410-425 (1985). 19. Kawasaki ES: Quantitative hybridization arrest of mRNA in Xenopus oocytes using single-strandedcomplementary DNA or oligonucleotide probes. Nucl Acids Res 13: 4991-5004 (1985). 20. Kerr A: The impact o f molecular genetics on plant pathology. Ann Rev Phytopathol 25:87-110 (1987). 21. Kim SK, Wood BJ: Stable reduction of thymidine kinase activity in cells expressing high levels of anti-sense RNA. Cell 42:129-138 (1985). 22. Larkin PL, Scowcroft WR: Somaclonal variation- a novel source of variability from cell cultures for plant improvement. Theor Appl Genet 60:197-214 (1981). 23. Loesch-Fries LS, Halk E, Merlo D, Jarvis N, Nelson S, Krahn K, Burhop L, Arntzen C J, Ryan C, in: Molecular strategies
for crop protection, Expression of alfalfa mosaic virus coat protein gene and anti-sense cDNA in transformed tobacco tissue. UCLA Symp Proc Mol Cell Biol New Set 48: 221- 234 (1987). 24. Maniatis T, Fritsch EF, Sambrook J: Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratories, Cold Spring Harbor, NY (1982). 25. Matthews REF: The changing scene in plant pathology. Ann Rev Phytopathol 25:10-23 (1987). 26. Melton DA: Injected anti-sense RNAs specifically block messenger RNA translation in vivo. Proc Natl Acad Sci USA 82: 144-148 (1985). 27. Miller WA, Bujarski J J, Dreher TW, Hall TC: Minus-strand initiation by brome mosaic virus replicase within the 3' tRNA-like structure of native and modified RNA templates. J Mol Biol 187:537-546 (1986). 28. Murashige T, Skoog F: A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 15: 473-497 (1962). 29. Murthy MRN: Comparison of the nucleotide sequences of cucumber mosaic virus and brome mosaic virus. J Mol Biol 168:469-475 (1983). 30. Peden KWC, Symons RH: Cucumber mosaic virus contains a functionally divided genome. Virology 53:487 - 492 (1973). 31. Rezaian MA, Williams RHV, Gordon KHJ, Gould AR, Symons RH: Nucleotide sequence of cucumber-mosaic-virus RNA 2 reveals a translation product significantly homologous to corresponding proteins of other viruses. Eur J Biochem 143:277-284 (1984). 32. Rezaian MA, Williams RHV, Symons RH: Nucleotide sequence of cucumber mosaic virus RNA 1. Presence of a sequence complementary to part of the viral satellite RNA and homologies with other viral RNAs. Eur J Biochem 150: 331-339 (1985). 33. Rosenberg UB, Preiss A, Seifert E, Jackle H, Knipple DC: Production of phenocopies by Kruppel anti-sense RNA injection into Drosophila embryos. Nature 313:703-706 (1985). 34. Sanford JC, Johnston SA: The concept of parasite-derived resistance - Deriving resistance genes from the parasite's own genome. J Theor Biol 113:395-405 (1985). 35. Shaw JG, Plaskitt KA, Wilson TMA: Evidence that tobacco mosaic virus particles disassemble cotranslationally in vivo. Virology 148:326-336 (1986). 36. Strauss EG, Strauss JH: Replication strategies for single stranded RNA viruses of eukaryotes. Curr Top Microbiol Immunol 105:1-98 (1983). 37. Symons RH: Cucumber mosaic virus RNA contains 7methyl guanosine at the 5' -terminus of all four RNA species. Mol Biol Reports 2:277-285 (1975). 38. Thomas PS: Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Natl Acad Sci USA 77: 5201-5205 (1980). 39. Weintraub H, Izant JG, Harland RM: Anti-sense RNA as a molecular tool for genetic analysis. Trends Genet. January: 22-25 (1985). 40. Zaitlin M, Hull R: Plant virus - host interactions. Ann Rev Plant Physiol 38:291-315 (1987).
