Satellite RNA Is Essential for Encapsidation of Groundnut Rosette Umbravirus RNA by Groundnut Rosette Assistor Luteovirus Coat Protein

Virology 254, 105–114 (1999) Article ID viro.1998.9527, available online at http://www.idealibrary.com on

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Satellite RNA Is Essential for Encapsidation of Groundnut Rosette Umbravirus RNA by Groundnut Rosette Assistor Luteovirus Coat Protein

D. J. Robinson, E. V. Ryabov, S. K. Raj,1 I. M. Roberts, and M. E. Taliansky2

Virology Department, Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, United Kingdom Received September 11, 1998; returned to author for revision September 28, 1998; accepted November 17, 1998

Groundnut rosette disease is caused by a complex of agents comprising groundnut rosette umbravirus (GRV), GRV satellite RNA (sat-RNA) and groundnut rosette assistor luteovirus (GRAV). Both GRAV and GRV sat-RNA are needed for GRV to be aphid transmissible. To understand the role of GRAV and GRV sat-RNA in the aphid transmission of GRV, encapsidation of GRV genomic and satellite RNAs has been studied using transgenic Nicotiana benthamiana plants expressing GRAV coat protein (CP). GRAV CP expressed from a transgene was shown to package GRV genomic and satellite RNAs efficiently, giving a high yield of transcapsidated virus particles. GRV sat-RNA was absolutely essential for this process. GRV genomic RNA was not encapsidated by GRAV CP in the absence of the sat-RNA. Using different mutants of GRV sat-RNA, it was found that some property of full-length satellite RNA molecules, such as size or specific conformation rather than potential open reading frames, was required for the production of virus particles. A correlation between the ability of sat-RNA to stimulate encapsidation of GRV RNA by GRAV CP and its capacity to promote aphid transmission of GRV was observed. © 1999 Academic Press

INTRODUCTION tor virus (GRAV) (Casper et al., 1983; Reddy et al., 1985; Rajeshwari et al., 1987), but neither GRAV nor GRV indi- Rosette disease of groundnut occurs in Africa, south of vidually induce the symptoms of groundnut rosette dis- the Sahara, and causes severe crop damage. The disease (Hull and Adams, 1968; Reddy et al., 1985). The ease is caused by a complex of agents comprising two major cause of the symptoms of the disease is GRV viruses and a satellite RNA (sat-RNA). Groundnut rosette sat-RNA (Murant et al., 1988; Murant and Kumar, 1990). virus (GRV) is a member of the genus Umbravirus (Mu- This is a single-stranded RNA of 895–903 nucleotides rant et al., 1995). The plant viruses in this group have that relies on GRV for its replication and that, more single-stranded RNA genomes but do not produce con- unusually, is needed (together with GRAV) for GRV to be ventional virus particles. They are mechanically trans- aphid-transmissible (Murant, 1990). Thus this RNA is missible, but each depends on a helper virus from the essential for the survival of GRV in nature. However, for family Luteoviridae for transmission by aphids (Murant et the sake of simplicity, we refer to it in this paper as GRV al., 1995; Taliansky et al., 1996). The entire nucleotide sat-RNA. Although different GRV sat-RNA variants con- sequence of GRV comprises 4019 nucleotides and contain up to five potential ORFs in either positive or nega- tains four open reading frames (ORFs) (Taliansky et al., tive sense (Blok et al., 1994), none of the ORFs is essen- 1996). Two ORFs at the 5Ј end of the RNA are expressed tial for replication of the sat-RNA or its spread in infected by a Ϫ1 frameshift to give a single protein, which ap- plants (Taliansky and Robinson, 1997). Likewise, the pro- pears to be an RNA-dependent RNA polymerase. The duction of symptoms in infected Nicotiana benthamiana other two ORFs overlap each other in different reading plants does not require any of the potential translation frames. The 27-kDa ORF3 protein seems to have a role in products but instead involves two untranslated elements long-distance virus movement in the infected plant (E. V. in the sat-RNA, which can act in trans (Taliansky and Ryabov, D. J. Robinson and M. E. Taliansky, manuscript in Robinson, 1997). preparation), but it is apparently not a coat protein (CP) The mechanism of transmission of the entire virus because GRV does not form virus particles. The 28-kDa complex responsible for rosette disease by the aphid ORF4 protein is a cell-to-cell movement protein (Talian- Aphis craccivora as well as the role of GRAV and GRV sky et al., 1996; Ryabov et al., 1998). For aphid transmis- sat-RNA in this process remains unclear. However, it is sion of GRV, the helper virus is groundnut rosette assis- suggested that genomic and satellite RNA molecules of GRV may be encapsidated by GRAV CP in a manner analogous to the encapsidation of carrot mottle umbra- 1 Present address: Plant Virus Laboratory, National Botanical Re- search Institute, Rana Pratap Marg, P.B.No 436 Lucknow-226001, India. virus RNA by the CP of its helper, carrot red leaf virus 2 To whom reprint requests should be addressed. Fax: 44–1382- (Waterhouse and Murant, 1983). The process of encap- 562426. E-mail address: [email protected]. sidation is therefore of considerable interest. However, it

0042-6822/99 $30.00 105 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved. 106 ROBINSON ET AL. is very difficult to study this process in groundnut plants, toms at the same time as nontransformed plants. For line the natural host of the virus complex, because of the low GRAV5, most plants developed more severe symptoms: yield of nucleoprotein particles in preparations purified in addition to showing the yellow blotch mosaic, the from this host plant. On the other hand, N. benthamiana, systemically infected leaves also became rugose and an experimental host for GRV and GRV sat-RNA, cannot deformed (Fig. 1). Immunosorbent electron microscopy be infected with GRAV. Indeed, no convenient experi- (ISEM) of the extracts from systemically infected leaves mental host for the entire complex is known. As an on grids coated with polyclonal antiserum against GRAV alternative approach, we generated transgenic N. revealed the presence of spherical luteovirus-like parti- benthamiana plants expressing the CP gene of GRAV. cles in all transgenic lines (see below), although the Although in natural infections luteovirus particles contain number of particles varied significantly between lines small amounts of a larger protein produced by occa- and was highest for line GRAV5. Virus-like particles were sional read-through of the major CP termination codon never detected in noninoculated transgenic plants or and required for aphid transmission (Gray and Banerjee, nontransformed N. benthamiana plants infected with the 1998), our transgenic plants did not contain the se- YB isolate of GRV. More importantly, virus-like particles quences coding for the read-through domain. Inoculation were not detected in plants of any of the transgenic lines of these plants with GRV isolates containing or lacking when they were infected with MC1, a satellite RNA-free sat-RNA showed that the sat-RNA is absolutely essential isolate of GRV, although MC1 RNA was encapsidated in for encapsidation of GRV RNA by GRAV CP and the the presence of an appropriate sat-RNA (see below). read-through domain is not required for this process. These results suggest that the GRAV CP expressed in transgenic plants is able to transcapsidate heterologous RESULTS (GRV genomic and satellite) viral but not cellular RNAs to Formation of virus particles in GRAV-CP transgenic give virus particles (hereafter referred to as transcapsi- plants inoculated with the YB isolate of GRV dated virus particles, TVP) and that sat-RNA might be essential for this process. Moreover, the read-through The gene encoding GRAV CP was inserted into trans- domain is not essential for such encapsidation. formation vector pROK2, a pBIN19 derivative (Bevan, 1984) between the cauliflower mosaic virus (CaMV) 35S Isolation, composition and properties of virus-like promoter and the nopaline synthase terminator, to give particles accumulated in GRAV CP transgenic plants the plasmid pROK2-GRAV CP. The resulting vector was infected with GRV-YB used for transformation of N. benthamiana plants. Twelve individual transgenic plants (primary transformants; T0 For further analysis of TVP formed in GRAV CP trans- generation), regenerated in the presence of kanamycin, genic plants inoculated with GRV-YB, line GRAV5 was se- were obtained. All of them were normal in appearance lected. In contrast to the phloem-limited character of luteo- and grew and developed like nontransformed plants. viral infections, it was expected that transgenic expression Total leaf RNA extracted from these plants was assayed of the GRAV CP gene might take place in different plant for GRAV CP gene transcripts by reverse transcription– tissues including mesophyll cells. Therefore for purification polymerase chain reaction (RT–PCR) using primers spe- of TVP from infected GRAV5 transgenic plants, a standard cific to the termini of the inserted CP gene sequences. All method of purification of luteoviruses (Takanami and Kubo, 12 primary transformants gave a product of the expected 1979) was employed but without the enzyme treatment size (ϳ600 bp) for the CP gene sequence (data not steps used for digestion of cell walls in vascular tissues. shown). However, neither ELISA nor Western blot (immu- Yields of TVP were unexpectedly high, 2–4 mg/100 g of noblot) analysis with polyclonal antiserum or different fresh leaf material. A time course showing the yield of TVP monoclonal antibodies (Scott et al., 1996) prepared to in a typical experiment is presented in Fig. 2A. Electron GRAV detected the protein product of the CP transgene microscopy showed that all the TVP were essentially sim- in plant extracts. ilar and typical of members of the family Luteoviridae and Despite the negative results of the immunochemical did not differ in appearance from normal GRAV virions assays, self-fertilized T1 progeny plants of all 12 indepen- isolated from naturally infected groundnuts (Figs. 3A and dently transformed lines were inoculated with the YB 3B). The TVP were stable in 1 or 2% aqueous uranyl acetate isolate of GRV. This isolate contains a sat-RNA variant negative stain but less so in either ammonium molybdate or (YB3) that induces brilliant yellow blotch mosaic symp- sodium phosphotungstate. There was no evidence of toms in N. benthamiana (Kumar et al., 1991). Each plant empty or stain-penetrated particles. Some particles had a received a standard, highly concentrated portion of the distinct hexagonal outline, indicative of icosahedral sym- YB inoculum, the infectivity of which was checked by metry, and the proportion of such particles was greater in applying it to the local lesion host Chenopodium ama- preparations fixed with 1% glutaraldehyde. In purified prep- ranticolor. In all transformed plant lines except one arations, the mean diameter of 100 particles was ϳ31 nm (GRAV5), plants developed typical yellow blotch symp- for isolated particles and ϳ28 nm for particles in small sat-RNA IN ENCAPSIDATION OF GRV RNA BY GRAV COAT PROTEIN 107

FIG 1. Transgenic (line GRAV5; A) and nontransgenic (B) N. benthamiana plants infected with GRV YB containing sat-RNA variant YB3. Transgenic N. benthamiana plants (line GRAV5), uninoculated (C), and inoculated with sat-RNA-free GRV MC1 (D). arrays, indicating that some collapse of the isolated parti- yielded a single UV-absorbing peak (Fig. 4B), confirming cles on the support film occurred during drying and that the the electron microscope observations of homogeneity of true diameter of the particles was ϳ28 nm. the TVP. The fraction with the highest UV absorbance TVP purified from GRAV5 transgenic plants infected had a density of 1.40 g/ml, corresponding to a nucleic with GRV-YB formed a single UV-absorbing zone in su- acid content of 29.5% (Sehgal et al., 1970). crose density gradients (Fig. 4A). Unfixed particles were N. benthamiana plants inoculated with TVP prepara- unstable in solutions of cesium chloride in phosphate tions showed typical brilliant yellow blotch symptoms, buffer, but after fixation with 1% glutaraldehyde, centrif- indicating that the particles contained both GRV genomic ugation to equilibrium in discontinuous CsCl gradients and satellite RNAs. RNA extracted from purified TVP was separated by electrophoresis in a 1.5% agarose gel and was stained with ethidium bromide. As shown in Fig. 5 (lane 2), the particles contained two RNA species of ϳ4000 and 900 nucleotides (nt), respectively, corresponding in length to GRV genomic (4019 nt) and YB3 sat-RNA (902 nt). The presence of these RNA species in the particles was confirmed by Northern blot hybridization (Fig. 5, lanes 3–6). No RNA band corresponding to GRV subgenomic RNA (ϳ1300 nt; Taliansky et al., 1996) was revealed in RNA samples from TVP (Fig. 5, lane 2).

FIG. 2. Typical time course of the yield of transcapsidated virus The dissociated protein from the TVP migrated in poly- particles from transgenic N. benthamiana plants infected with GRV YB acrylamide gel electrophoresis as a single band of mo- (A) and GRV MC1 (B). lecular mass ϳ25 kDa (Fig. 6, lane 1), in reasonable 108 ROBINSON ET AL.

FIG. 3. Electron micrograph of purified transcapsidated virus particles formed by GRAV CP and GRV genomic and satellite RNAs (A) and normal GRAV virions isolated from naturally infected groundnuts (B), negatively stained with 1% uranyl acetate. The bar represents 200 nm. agreement with the molecular mass of GRAV CP calcu- this is consistent with each particle containing one mol- lated from the deduced amino acid sequence (22350 Da; ecule of GRV genomic RNA and one molecule of sat- Scott et al., 1996). In Western blot (immunoblot) analysis, RNA, together comprising ϳ4900 nt. this band reacted with monoclonal antibody SCR 114 As mentioned above, ISEM did not detect any virus- (Fig. 6, lane 2), prepared against GRAV, and corre- like particles in the transgenic plants infected with the sponded in position to that of GRAV CP accumulated in sat-RNA-free GRV isolate MC1. All attempts to isolate groundnut plants infected with GRAV (Fig. 6, lane 4). particles from these plants were unsuccessful, suggest- These results clearly show that GRAV CP produced in ing that GRAV CP was not able to encapsidate GRV RNA transgenic plants is able to transcapsidate GRV genomic in the absence of sat-RNA (Fig. 2B). However, the pos- and satellite RNA. ‘‘Normal’’ luteoviruses have a T ϭ 3 sibility that particles containing GRAV CP and GRV RNA structure with 180 protein subunits, and the TVP probably can be formed, but are unstable, cannot be completely have the same structure. In this case, the total mass of ruled out; such particles might not be detectable by the the particle implies that each particle contains ϳ4800 nt methods we have used. GRV genomic RNA accumulated of RNA. Taking into account the homogeneity of the TVP, to similar levels in infected N. benthamiana plants re- gardless of the presence or absence of sat-RNA (Talian- sky and Robinson, 1997); therefore the role of the sat- RNA is apparently in the encapsidation of GRV RNA by GRAV CP. It should be noted that TVP production is not a specific function of the YB3 variant of the sat-RNA. Infec- tion of GRAV5 transgenic plants with GRV MC1 RNA together with another GRV sat-RNA (MC3a) also resulted in TVP formation (data not shown).

The full-length sat-RNA molecule rather than the FIG. 4. Sedimentation profiles of purified transcapsidated virus par- potential sat-RNA encoded ORFs are required for ticles formed by GRAV CP and GRV genomic and satellite RNAs. encapsidation of GRV RNA with GRAV CP Absorbance pattern (254 nm) of purified particles following centrifugation in a 10–40% linear sucrose gradient (A) or equilibrium centrifuga- Examination of the nucleotide sequences of different tion in a CsCl gradient (B). variants of GRV sat-RNA revealed the presence of up to sat-RNA IN ENCAPSIDATION OF GRV RNA BY GRAV COAT PROTEIN 109 five ORFs in both positive- and negative-sense RNA strands. Sat-RNA YB3b (a clone of YB3 satellite RNA) contained four of these ORFs (ORFs I, III, IV and V) (Blok et al., 1994). Site-directed mutagenesis was used earlier to replace the initiation codon of each ORF with another triplet in such a way as not to change the coding capacity of the opposite RNA strand (Taliansky and Robinson, 1997). Transcripts from the mutated plasmids were co- FIG. 6. Analysis of the protein isolated from TVP. (Lane 1) Electro- inoculated with GRV MC1 RNA to transgenic plants of phoresis of the protein in SDS–polyacrylamide gel, stained by Coomas- line GRAV5. All four mutants (YB-ORFIϪ, YB-ORFIIIϪ, YB- sie blue. Molecular masses and positions of the following proteins Ϫ Ϫ expressed from E. coli and used as markers are indicated on the right: ORFIV and YB-ORFV ) stimulated formation of TVP, GRV ORF4-encoded protein (27 kDa), N-terminal part of potato leaf roll giving a yield comparable to that obtained with YB3b virus (PLRV) ORF5-encoded protein (22 kDa) and PLRV ORF4-encoded sat-RNA (Fig. 7). These results show that no single po- protein (17 kDa). (Lanes 2–4) Immunoblot analysis of the TVP (lane 2), tential sat-RNA-encoded ORF is required for transcapsi- plant extracts obtained from transgenic N. benthamiana plants (line GRAV5) infected with sat-RNA-free GRV MC1 (lane 3) and from ground- dation of GRV genomic and satellite RNAs with GRAV CP. nuts infected with GRAV (lane 4, positive control). The protein blot was The ability of two deletion mutants of YB3b sat-RNA to reacted first with monoclonal antibody SCR 114, followed by reaction mediate the formation of TVP was also tested. YB⌬1 and with rabbit anti-mouse antibody conjugated with alkaline phosphatase. YB⌬3 have deletions of nucleotides 282–470 or 629–845, respectively (Taliansky and Robinson, 1997). When coinoculated with GRV MC1 RNA, both mutants replicated Ability of sat-RNA deletion mutants to mediate aphid and spread in infected plants to levels comparable to transmission those achieved by YB3b sat-RNA (Taliansky and Robin- Murant (1990) showed that groundnut plants must con- son, 1997; Fig. 5, lanes 7–9), but neither stimulated for- tain sat-RNA, as well as GRV and GRAV, for the GRV mation of TVP in GRAV5 transgenic plants either sepa- component to be transmissible by A. craccivora. The rately or as a mixture (Fig. 7). These results suggest that results presented above suggest that this is because some property of full-length sat-RNA molecule, for exam- sat-RNA is needed for the encapsidation of GRV RNA ple, its size or/and specific secondary structure, is es- into the TVP, which are its aphid transmissible form. To sential for production of TVP. confirm that mutant sat-RNAs that are unable to mediate transcapsidation are also unable to mediate vector transmission, aphid transmission tests were done from source groundnut plants infected with GRAV and GRV MC1 and containing both sat-RNAs YB⌬1 and YB⌬3. Of 15 test groundnut plants to which aphids were transferred, 10 became infected with GRAV, but only 1 with GRV. When an extract from this one plant was inoculated to N. benthamiana, it induced yellow blotch symptoms typical of GRV infections containing either sat-RNA YB3 or both sat-RNAs YB⌬1 and YB⌬3 (Taliansky and Robin- son, 1997). A Northern blot of RNA extracted from this N. benthamiana plant, probed with a sat-RNA specific probe, detected an RNA of ϳ900 nt, but no molecules of FIG. 5. RNA composition of transcapsidated virus particles. (Lanes 1 ϳ700 nt corresponding to the deletion mutants of the and 2) Electrophoresis in 1.5% agarose gel of RNAs extracted from sat-RNA (Fig. 5A, lane 10). This full-length sat-RNA had purified TVP (lane 2) and RNA species of cucumber mosaic virus (CMV) presumably been formed by recombination in the pri- used as markers (lane 1). The size of CMV RNAs 1, 2, 3 and 4 in ⌬ ⌬ nucleotides is indicated on the left. (Lane 2) TVP RNAs with approxi- mary source plant between YB 1 and YB 3. However, it mate sizes in nucleotides indicated on the right. (Lanes 3–6) Northern can not be completely ruled out that this RNA was a blot analysis with probes for GRV genomic (lanes 3 and 4) and satellite contaminant, although precautions were taken to prevent (lanes 5 and 6) RNAs of RNA isolated from TVP (lanes 3 and 5) and from such an occurrence. nontransgenic N. benthamiana plants infected with GRV YB (lanes 4 and 6). Positions of genomic (g) and satellite (sat) RNAs are indicated. Aphid transmission of TVPs Electrophoresis was in 1.5% agarose gel. (Lanes 7–10) Northern blot analysis of GRV sat-RNA isolated from transgenic N. benthamiana The ability of aphids to transmit TVPs directly from plants (line GRAV5) (lanes 7–9) infected with GRV MC1 together with YB transgenic N. benthamiana plants could not be tested 3b sat-RNA (lane 7), YB⌬3 sat-RNA (lane 8) or YB⌬1 sat-RNA (lane 9) and from a nontransgenic N. benthamiana plant inoculated with an because A. craccivora will not feed on this species. extract from a groundnut plant infected by aphid transmission (lane 10, Instead, aphids were allowed to feed on suspensions of see text). Electrophoresis was in 2% agarose gel. purified TVP through Nescofilm membranes and subse- 110 ROBINSON ET AL.

lated viruses. Unrelated viral RNAs and CPs may be prevented from forming particles by the highly specific character of RNA-protein interactions during virus assembly, described for example for tobacco mosaic virus (Atabekova et al., 1975; Taliansky et al., 1977). When phenotypic mixing does take place between unrelated viruses, the efficiency of heterologous encapsidation is rather low (Dodds and Hamilton, 1974; Hamilton and Nichols, 1977). Taxonomically, umbraviruses and luteoviruses appear to be unrelated viruses. However, in nature, umbraviruses often form biologically stable complexes with luteoviruses (Murant, 1993), and the groundnut rosette disease complex is one such example (Murant, 1990). Thus umbraviral (GRV) and luteoviral (GRAV) compo- nents of the complex are functionally related. The third component of the complex is the sat-RNA. Here we show that GRAV CP expressed from a transgene is able to package GRV genomic and satellite RNAs efficiently, giving a high yield of TVP (Յ4 mg/100 g of leaf tissue in plants of line GRAV5). The high yield of TVP is probably the cause of the significant exacerbation of symptoms induced by GRV-YB infection in the GRAV5 plants. How- ever, it remains unclear why routine immunochemical assays (ELISA and Western blot analysis) did not detect any GRAV CP in uninfected plants of any of the GRAV CP-transgenic lines, including GRAV5. One possible ex- planation of this contradiction is that the GRAV CP expressed in the transgenic plants may be unstable and FIG. 7. Production of transcapsidated virus particles in GRAV CP- rapidly degraded. Packaging in the form of virus-like transgenic plants coinoculated with GRV MC1 RNA and with either particles might stabilize it. RNA size may be a critical wild-type or mutated GRV sat-RNAs. The genome organization of YB3b factor for stable particle assembly (see, for example, Qu sat-RNA, and the positions and types of mutations described earlier (Taliansky and Robinson, 1997) are indicated. Boxes represent poten- and Morris, 1997). tial open reading frames. Deleted sequences in deletion mutants are We show here that the GRV genomic RNA cannot be shown by broken lines. Virus yields are indicated on the right. packaged by GRAV CP in the absence of sat-RNA. It seems that the two RNA molecules are encapsidated together. A possible mechanism for such dependence quently transferred to test plants. However, although the may be that molecular interactions between the two RNA TVP preparation was infective when inoculated mechan- molecules are necessary to make them competent for ically to N. benthamiana (see above), none of 11 test transcapsidation. Since potential proteins encoded by plants became infected, suggesting that TVP are aphid the sat-RNA ORFs are not required for packaging, the transmissible at low efficiency, if at all. sat-RNA molecule itself seems to be involved in these interactions. A functional, physical interaction between DISCUSSION two viral RNA species (RNA 1 and RNA 2 of red clover necrotic mosaic dianthovirus) has been recently demon- Phenotypic mixing, or transcapsidation, describes the strated by Sit et al. (1998). Moreover, sat-RNA molecules coating of the RNA of one virus or isolate either partially with deletions of ϳ200 nt are unable to mediate pack- or completely with the CP of another virus (Rochow, aging, showing that some property of the full-length sat- 1970). The phenomenon of phenotypic mixing was ob- RNA molecule, for example its size, is the determining served in mixed infections (Sarkar, 1969; Atabekov et al., factor. However, the possibility that some specific con- 1970; Rochow, 1970; Kassanis and Bastow, 1971; Bourdin formation formed by the full-length sat-RNA molecule is and Lecoq, 1991) and in transgenic plants expressing the important cannot be ruled out. Sat-RNAs YB⌬1 and YB⌬3 CP of one plant virus and infected with another virus seem also unable to mediate aphid transmission of GRV, (Osbourn et al., 1990; Farinelli et al., 1992; Lecoq et al., supporting the view that transmissibility is a conse- 1993; Maiss et al., 1994; Hammond and Dienelt, 1997). quence of transcapsidation. The single example of trans- However, phenotypic mixing usually occurs between re- mission that was observed involved a full-length sat-RNA sat-RNA IN ENCAPSIDATION OF GRV RNA BY GRAV COAT PROTEIN 111 molecule, presumably generated by recombination be- universal requirement for the transcapsidation of umbra- tween YB⌬1 and YB⌬3. Interestingly, Taliansky and Rob- virus RNA. inson (1997) failed to detect such a recombinant in N. Incorporation of virus CP genes into plant genomes benthamiana plants in which both YB⌬1 and YB⌬3 were has been successful in providing significant degrees of actively replicating. Its detection after aphid transmission resistance to virus infections. However, a perceived haz- suggests that it was revealed as a result of selection ard of this strategy involves transcapsidation of viral because, unlike YB⌬1 and YB⌬3, the recombinant was RNAs by plant-produced CP (Hull, 1990) and a conse- able to mediate transcapsidation of GRV RNA and con- quent change in the vector specificity of a naturally sequently its aphid transmission. Selection of this mole- occurring virus (Rochow, 1970). The results presented cule by aphid transmission further supports the idea that here show that for luteovirus CP transgenes, the first part only the full-length sat-RNA molecules are competent to of this scenario can occur. However, luteovirus particles mediate transcapsidation and aphid transmission. The contain, in addition to the predominant coat protein, observation that the deleted sat-RNAs were not cotrans- small amounts of a larger protein produced by transla- mitted with the recombinant sat-RNA implies that the tional read-through of the coat protein gene stop codon, function of sat-RNA in transmission cannot act in trans. and particles lacking this protein are not transmissible This is consistent with the idea that sat-RNA is physically by aphids (Gray and Banerjee, 1998). In natural infections involved in the assembly of TVPs and that each particle of groundnuts, TVP formed by encapsidating GRV contains only a single sat-RNA molecule. The ability of genomic and satellite RNA with GRAV CP apparently the GRV genomic and satellite RNAs to be coated by the contain such a protein with the read-through portion and taxonomically unrelated GRAV CP has probably been therefore are aphid-transmissible. In contrast, our trans- acquired in the process of evolution, enabling them to be genic plants, like most of those designed to provide transmitted from plant to plant by aphids. However, it is resistance to luteovirus infections, do not contain the unclear how a precursor of GRV might have survived in sequences coding for the read-through domain and therefore particles constructed from the plant produced nature before it acquired this functional adaptation. CP would not be expected to be aphid transmissible. Because the CP of luteoviruses and other virus groups Indeed, attempts to demonstrate transmission of the are involved in vector specificity (Harrison and Robinson, TVPs by feeding them to aphids through membranes 1988), transcapsidation of GRV RNA with GRAV CP and were unsuccessful. If it is confirmed that such particles the dependence of this process on the GRV sat-RNA are nontransmissible, then the phenomenon of transcap- clarifies the biological roles of GRAV and GRV sat-RNA in sidation in luteovirus CP transgenic plants does not pose the aphid transmission of GRV. GRV is probably trans- a risk to the environment. mitted by A. craccivora in the form of TVP, and sat-RNA is essential for TVP formation. Thus the sat-RNA is seen to MATERIALS AND METHODS be a very important factor in both the development and spread of groundnut rosette disease. It is the sat-RNA Virus cultures and plant inoculation that is primarily responsible for induction of disease The MC1 sat-RNA-free GRV culture was derived by Mu- symptoms, and it is required for transcapsidation of GRV rant and Kumar (1990) from GRV isolate MC, which was RNA by GRAV CP and hence for the aphid transmissibility obtained from a Malawian groundnut plant showing symp- of GRV. In both of these processes (symptom induction toms of chlorotic rosette. The satellite-containing YB isolate and transcapsidation), GRV sat-RNA operates as an un- of GRV from Malawi was described by Kumar et al. (1991). translated RNA molecule directly involved in particular Both isolates were propagated in N. benthamiana by man- biological functions rather than through translation prod- ual inoculation. GRV isolates were maintained under li- ucts as a messenger RNA. cense from the Scottish Office Agriculture, Environment and The sat-RNA of another umbravirus, pea enation mo- Fisheries Department. Plasmid pYB3b, from which biologi- saic virus 2 (PEMV2), is not essential for transcapsida- cally active YB3b GRV sat-RNA can be obtained, was de- tion of PEMV2 with the CP of its helper, pea enation scribed by Demler et al. (1996b). Plasmids corresponding to Ϫ Ϫ Ϫ Ϫ mosaic virus 1 (PEMV1), and because of that, PEMV2 mutants YB-ORFI , YB-ORFIII , YB-ORFIV or YB-ORFV , sat-RNA is not required for aphid transmission of this as well as deletion mutants YB⌬1 and YB⌬3 were de- virus complex. It has been shown that PEMV2 RNA is scribed by Taliansky and Robinson (1997). All sat-RNA plas- encapsidated by PEMV1 CP during mixed infection re- mids were linearized with SpeI and used as templates for in gardless of the presence or absence of PEMV2 sat-RNA, vitro transcription as described by Demler et al. (1996b). although the sat-RNA may stabilize the particles (Demler Transcripts were left uncapped. Transgenic or nontrans- et al., 1996a). Another umbravirus, carrot mottle virus, genic N. benthamiana plants were manually inoculated has no sat-RNA associated with it, but its RNA is encap- either with total RNA samples containing MC1 or YB GRV sidated by the CP of its helper, carrot red leaf virus RNA or with a mixture of sat-RNA transcripts with total RNA (Waterhouse and Murant, 1983). Thus sat-RNA is not a extracted from GRV MC1-infected plants. 112 ROBINSON ET AL.

Construction of transgene and plant transformation cellulose nitrate tubes. One milliliter of purified TVP sus- pension containing 0.06 mg of virus was layered onto A cDNA fragment containing the entire GRAV CP gene each gradient and centrifuged to equilibrium at 33,000 was recloned from the pBluescript clone described by rpm for 20 h at 4°C. The gradients were scanned and Scott et al. (1996) between the XbaI and KpnI sites of fractionated. Densities of fractions were determined from pROK2, a binary plant transformation vector based on measurements of refractive index, made using an Atago pBIN19 (Bevan, 1984). The resulting plasmid was mobi- Rx-1 refractometer. lized from Escherichia coli DH5␣ into Agrobacterium tumefaciens LBA 4404 by triparental mating with RNA and protein analysis pRK2013 as described by Armitage et al. (1988). Pieces of N. benthamiana stem tissue were transformed as de- RNAs extracted from purified TVP were separated in scribed by Benvenuto et al. (1991). Transgenic shoots 1.5% agarose gel and stained with ethidium bromide. For were regenerated on a selection medium containing Northern blot analysis, total RNA preparations were de- kanamycin (100 ␮g/ml). Rooted plantlets were trans- natured with formaldehyde and formamide. Electro- ferred to sterilized compost and, after an adaptation phoresis was in 1.5 or 2% agarose gel as outlined in Sambrook et al. (1989). RNA was transferred to Hy- period in a climate room at a humidity of 70%, were ϫ maintained in a glasshouse. bond-N membrane by the capillary method with 20 SSC (3 M sodium chloride and 0.3 M sodium citrate, pH 7.0) and immobilized by UV-cross-linking. Hybridization Purification of transcapsidated virus particles was done as described by Sambrook et al. (1989) with 32 The method was essentially that described by Taka- probes labeled with [ P] dATP using a Random Primers nami and Kubo (1979) for purification of potato leaf roll DNA Labelling kit (Life Technologies). GRV-specific virus (PLRV) but without Celluclast treatment. Freshly probes were prepared from the inserts of clone gr21 picked leaf tissue was triturated in 0.1 M sodium citrate (Taliansky et al., 1996), which represents sequences near buffer, pH 6.0, containing 0.5% ␤-mercaptoethanol and 5Ј end of the GRV genome. sat-RNA-specific probes then emulsified with 0.67 vol chloroform/butanol (1:1 v/v) were prepared from the complete insert of pYB3b (Dem- at room temperature. To the aqueous phase obtained ler et al., 1996b). For protein analysis, samples of purified after centrifugation, polyethylene glycol (MW 6000) and TVP were incubated with 3% ␤-mercaptoethanol and 3% sodium chloride were added to 8% (w/v) and 0.2 M, sodium dodecyl sulfate (SDS) for 5 min at 95°C and then respectively. The liquid was stirred for1hat4°Cand were analyzed by 12.5% SDS–polyacrylamide gel elec- then incubated at room temperature for 2–3 h. The pre- trophoresis (Sambrook et al., 1989). For immunoblot anal- cipitate was resuspended in 0.02 M sodium phosphate ysis, blots were treated with monoclonal antibody SCR buffer, pH 7.5, containing 1% (v/v) Triton X-100. The virus 114, prepared against GRAV (Scott et al., 1996), and then was further purified by two cycles of high-speed centrif- with rabbit anti-mouse antibody conjugated with alkaline ugation through cushions of 20% sucrose. The virus yield phosphatase (Sigma). from each experiment was determined assuming an ex- 0.1% ϭ Aphid transmission experiments tinction coefficient A1cm, 260 8.6 (based on the value calculated for PLRV by Takanami and Kubo, 1979). Aphid transmission tests from groundnut plants infected with GRAV, GRV and sat-RNAs were done as Electron microscopy and sedimentation analysis of described by Demler et al. (1996b). Membrane feeding TVP experiments were done essentially as described by Duf- fus and Gold (1965). Groups of 25 apterous adult A. Electron microscope observations were made with a craccivora were allowed to feed through Nescofilm mem- Philips CM 10 electron microscope. ISEM was done branes on a liquid containing 50 ␮g/ml TVPs and 20% essentially as described previously (Roberts, 1986), us- (w/v) sucrose in 10 mM phosphate buffer, pH 7.0. After ing a polyclonal antiserum prepared to GRAV. Centrifu- 24 h, virtually all the aphids remained alive, and they gation in sucrose density gradients was done in a Beck- were transferred to seedlings of groundnut, 10–12 aphids/ man SW 41 rotor. Virus (0.06 mg) was layered onto linear plant. After 72 h feeding on these plants, the aphids were (10–40% w/v) sucrose gradients and centrifuged at killed by fumigation with nicotine, and the test plants 38,000 rpm for 2 h. The gradients were fractionated and were maintained in the glasshouse for ϳ4 weeks before analyzed using an ISCO model 640 density gradient infectivity testing on Chenopodium amaranticolor or fractionator fitted with a model UA-6 UV analyzer. The N. benthamiana for the presence of GRV (Reddy et al., buoyant density of TVP was determined by equilibrium 1985). banding in a Beckman SW 38 rotor. Cesium chloride was dissolved in 0.01 M Tris–HCl buffer, pH 7.2, and discon- ACKNOWLEDGMENTS tinuous gradients were made by layering 1 ml each of SCRI is grant-aided by the Scottish Office Agriculture, Environment solutions of densities 1.50, 1.40, 1.35, and 1.30 in 5-ml and Fisheries Department. E.V. Ryabov was in receipt of an EMBO sat-RNA IN ENCAPSIDATION OF GRV RNA BY GRAV COAT PROTEIN 113 fellowship, and S.K. Raj was supported by a fellowship from the Indian resistance in transgenic plants. Mol. Plant-Microbe Interact. 10, National Science Academy and the Royal Society under an exchange 1023–1027. agreement. We thank Angelika Ziegler for preparing the GRAV CP Harrison, B. D., and Robinson, D. J. (1988). Molecular variation in transformation vector, and Tony Murant for help with the aphid trans- vector-borne plant viruses: epidemiological significance. Philos. mission experiments. We are also greatful to Peter Palukaitis and Trans. R. Soc. Lond. B Biol. Sci. 321, 447–462. Michael Wilson for helpful discussions and for valuable comments on Hull, R. (1990). The use and misuse of viruses in cloning and expres- the manuscript. sion in plants. In ‘‘Recognition and Response in Plant Virus Interac- tions’’ (R. S. S. Fraser, Ed.), pp. 443–457. Springer-Verlag, Berlin. REFERENCES Hull, R., and Adams, A. N. (1968). Groundnut rosette and its assistor virus. Ann. Appl. Biol. 62, 139–145. Atabekov, J. G., Schaskolskaya, N. D., Atabekova, T. I., and Sakha- Kassanis, B., and Bastow, C. (1971). Phenotypic mixing between strains rovskaya, G. N. (1970). Reproduction of temperature-sensitive strains of tobacco mosaic virus. J. Gen. Virol. 11, 171–176. of TMV under restrictive conditions in the presence of temperature- Kumar, I. K., Murant, A. F., and Robinson, D., J. (1991). A variant of the resistant helper strain. Virology 41, 397–407. satellite RNA of groundnut rosette virus that induces brilliant yellow Atabekova, T. I., Taliansky, M. E., and Atabekov, J. G. (1975). Specificity blotch mosaic symptoms in Nicotiana benthamiana. Ann. Appl. Biol. of protein-RNA and protein-protein interactions upon assembly of 118, 555–564. TMV in vivo and in vitro. Virology 67, 1–13. Lecoq, H., Ravelonandro, M., Wipf-Scheibel, C., Monsion, M., Raccah, Armitage, P., Walden., R., and Draper, J. (1988). Vectors for transforma- B., and Dunez, J. (1993). Aphid transmission of a non-aphid-transmis- tion of plant cells using Agrobacterium. In ‘‘Plant Genetic Transfor- sible strain of zucchini yellow mosaic potyvirus from transgenic mation and Gene Expression–A Laboratory Manual’’ (S. Draper, R. plants expressing the capsid protein of plum pox potyvirus. Mol. Scott, P. Armitage, and R. Walden, Eds.), pp. 1–67. Blackwell Scientific Plant-Microbe Interact. 6, 403–406. Publications, Oxford. Maiss, E., Koenig, R., and Lesemann, D.-E. (1994). Heterologous en- Benvenuto, E., Ordas, R. J., Tavazza, R., Ancora, G., Biocca, S., Cattaneo, capsidation of viruses in transgenic plants and in mixed infections. In A., and Galeffi, P. (1991). ‘‘Phytoantibodies’’: A general vector for the ‘‘Proceedings of the International Symposium on Biosafety Results of expression of immunoglobulin domains in transgenic plants. Plant Field Tests of Genetically Modified Plants and Microorganisms’’ Mol. Biol. 17, 865–874. (D. D. Jones, Ed.), 3rd Ed. pp. 129–139. University of California, Bevan, M. (1984). Binary Agrobacterium vectors for plant transforma- Division of Agriculture and Natural Resourses, Oakland. tion. Nucleic Acids Res. 12, 8711–8721. Murant, A. F. (1990). Dependence of groundnut rosette virus on its Blok, V. C., Ziegler, A., Robinson, D. J., and Murant, A. F. (1994). Se- satellite RNA as well as on groundnut assistor luteovirus for trans- quences of 10 variants of the satellite-like RNA3 of groundnut rosette mission by Aphis craccivora. J. Gen. Virol. 71, 2163–2166. virus. Virology 202, 25–32. Murant, A. F. (1993). Complexes of transmission-dependent and helper Bourdin, D., and Lecoq, H. (1991). Evidence that heteroencapsidation viruses. In ‘‘Diagnosis of Plant Virus Diseases’’ (R. E. F. Matthews, between two potyviruses is involved in aphid transmission of non- Ed.), pp. 333–357. CRC Press, Boca Raton, FL. aphid-transmissible isolate from mixed infections. Phytopathology Murant, A. F., and Kumar, I. K. (1990). Different variants of the satellite 81, 1459–1464. RNA of groundnut rosette virus are responsible for the chlorotic and Casper, R., Meyer, S., Lesemann, D.-E., Reddy, D. V. R., Rajeshwari, R., green forms of groundnut rosette disease. Ann. Appl. Biol. 117, Misari, S. M., and Subbarayadu, S. S. (1983). Detection of a luteovirus 85–92. in groundnut rosette diseased groundnuts (Arachis hypogaea)by Murant, A. F., Rajeshwari, R., Robinson, D. J., and Rascke, J. H. (1988). enzyme-linked immunosorbent assay and immunoelectron micros- A satellite RNA of groundnut rosette virus is largely responsible for copy. Phytopathol. Z. 108, 12–17. symptoms of groundnut rosette disease. J. Gen. Virol. 69, 1479–1486. Demler, S. A., de Zoeten, G. A., Adam, G., Harris, K. F. (1996a). Pea Murant, A. F., Robinson, D. J., and Gibbs, M. J. (1995). Genus Umbravi- enation mosaic enamovirus: properties and aphid transmission. In rus. In ‘‘Virus Taxonomy-Classification and Nomenclature of Viruses. ‘‘The Plant Viruses. Polyhedral Virions and Bipartite RNA Genomes’’ (B. D. Harrison and A. F. Murant, Eds.), vol. 5, chapt. 12, pp. 303–334. Sixth Report of the International Committee on the Taxonomy of Plenum Press, New York. Viruses’’ (F. A. Murphy, C. M. Fauquet, D. H. L. Bishop, S. A. Gabrial, Demler, S. A., Rucker, D. G., de Zoeten, G. A., Ziegler, A., Robinson, D. J., A. W. Jarvis, G. P. Martelli, M. A. Mayo, and M. D. Summers, Eds.), pp. and Murant, A. F. (1996b). The satellite RNAs associated with the 388–391. Springer-Verlag, Vienna. groundnut rosette disease complex and pea enation mosaic virus: Osbourn, J. K., Sarkar, S., and Wilson, T. M. A. (1990). Complementation Sequence similarities and ability of each other’s helper virus to of coat protein-defective TMV mutants in transgenic tobacco plants support their replication. J. Gen. Virol. 77, 2847–2855. expressing TMV coat protein. Virology 179, 921–925. Dodds, J. A., and Hamilton, R. I. (1974). Masking of the RNA genome of Qu, F., and Morris, T. J. (1997). Encapsidation of turnip crinkle virus is tobacco mosaic virus by the protein of barley stripe mosaic virus in defined by a specific packaging signal and RNA size. J. Virol. 71, doubly infected barley. Virology 59, 418–427. 1428–1435. Duffus, J. E., and Gold, A. H. (1965). Transmission of beet western Rajeshwari, R., Murant, A. F., and Massalski, P. R. (1987). Use of yellows virus by aphids feeding through a membrane. Virology 27, monoclonal antibody to potato leafroll virus for detecting groundnut 388–390. rosette assistor virus by ELISA. Ann. Appl. Biol. 111, 353–358. Farinelli, L., Malnoe, P., and Collet, G. (1992). Heterologous encapsida- Reddy, D. V. R., Murant, A. F., Duncan, G. H., Ansa, O. A., Demski, J. W., tion of potato virus Y strain O (PVYo) with the transgenic coat protein and Kuhn, C. W. (1985). Viruses associated with chlorotic rosette and of PVY strain N (PVYN)inSolanum tuberosum cv. Bintje. BioTechnol- green rosette diseases of groundnut in Nigeria. Ann. Appl. Biol. 107, ogy 10, 1020–1025. 57–64. Gray, S. M., and Banerjee, N. (In press). Mechanisms of arthropod Roberts, I. M. (1986). Immunoelectron microscopy of extracts of virus- transmission of plant and animal viruses. Microbiol. Mol. Biol. Rev. infected plants. In ‘‘Electron Microscopy of Proteins’’ (J. R. Harris and Hamilton, R. I., and Nichols, C. (1977). The influence of bromegrass R. W. Horne, Eds.), pp. 293–357. Academic Press, San Diego. mosaic virus on the replication of tobacco mosaic virus in Hordeum Rochow, W. F. (1970). Barley yellow dwarf virus: Phenotypic mixing and vulgare. Phytopathology 67, 484–489. vector specificity. Science 167, 875–878. Hammond, J., and Dienelt, M. M. (1997). Encapsidation of potyviral RNA Ryabov, E. V., Oparka, K. J., Santa Cruz, S., Robinson, D. J., and Talian- in various forms of transgene coat protein is not correlated with sky, M. E. (1998). Intracellular location of two groundnut rosette 114 ROBINSON ET AL.

umbravirus proteins delivered by PVX and TMV vectors. Virology 242, Takanami, Y., and Kubo, S. (1979). Enzyme-assisted purification of two 303–313. phloem-limited plant viruses: Tobacco necrotic dwarf and potato Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ‘‘Molecular Cloning: leafroll. J. Gen. Virol. 44, 153–159. A Laboratory Manual,’’ 2nd Ed. Cold Spring Harbor Laboratory Press, Taliansky, M. E., Atabekova, T. I., and Atabekov, J. G. (1977). The Cold Spring Harbor, NY. formation of phenotypically mixed particles upon mixed assembly of Sarkar, S. (1969). Evidence of phenotypic mixing between two strains of some tobacco mosaic virus (TMV) strains. Virology 76, 701–708. tobacco mosaic virus. Mol. Gen. Genet. 105, 87–90. Taliansky, M. E., and Robinson, D. J. (1997). Trans-acting untranslated Sehgal, O. P., Jean, J.-H., Bhalla, R. B., Soong, M. M., and Krause, G. F. elements of groundnut rosette virus satellite RNA are involved in (1970). Correlation between buoyant density and ribonucleic acid symptom production. J. Gen. Virol. 78, 1277–1285. content in viruses. Phytopathology 60, 1778–1784. Taliansky, M. E., Robinson, D. J., and Murant, A. F. (1996). Complete Scott, K. P., Farmer, M.-J., Robinson, D. J., Torrance, L., and Murant A. F. nucleotide sequence and organization of the RNA genome of (1996). Comparison of the coat protein of groundnut rosette assistor groundnut rosette umbravirus. J. Gen. Virol. 77, 2335–2345. virus with those of other luteoviruses. Ann. Appl. Biol. 128, 177–183. Waterhouse, P. M., and Murant, A. F. (1983). Further evidence on the Sit, T. L., Vaewhongs, A. A., and Lommel, S. A. (1998). RNA-mediated nature of the dependence of carrot mottle virus on carrot red leaf trans-activation of transcription from a viral RNA. Science 281, 829–832. virus for transmission by aphids. Ann. Appl. Biol. 103, 455–464.