- Email: [email protected]
Rose spring dwarf-associated virus has RNA structural and gene-expression features like those of Barley yellow dwarf virus
Available online at www.sciencedirect.com
Virology 375 (2008) 354 – 360 www.elsevier.com/locate/yviro
Rose spring dwarf-associated virus has RNA structural and gene-expression features like those of Barley yellow dwarf virus Nida' M. Salem a , W. Allen Miller b,c , Adib Rowhani a , Deborah A. Golino a , Anne-Laure Moyne a , Bryce W. Falk a,⁎ a
c
Department of Plant Pathology, One Shields Ave., University of California, Davis, CA 95616, USA b Plant Pathology Department, Iowa State University, Ames, Iowa 50011, USA Biochemistry, Biophysics and Molecular Biology Department, Iowa State University, Ames, Iowa 50011, USA Received 21 September 2007; returned to author for revision 29 November 2007; accepted 27 January 2008 Available online 7 March 2008
Abstract We determined the complete nucleotide sequence of the Rose spring dwarf-associated virus (RSDaV) genomic RNA (GenBank accession no. EU024678) and compared its predicted RNA structural characteristics affecting gene expression. A cDNA library was derived from RSDaV double-stranded RNAs (dsRNAs) purified from infected tissue. Nucleotide sequence analysis of the cloned cDNAs, plus for clones generated by 5′- and 3′-RACE showed the RSDaV genomic RNA to be 5808 nucleotides. The genomic RNA contains five major open reading frames (ORFs), and three small ORFs in the 3′-terminal 800 nucleotides, typical for viruses of genus Luteovirus in the family Luteoviridae. Northern blot hybridization analysis revealed the genomic RNA and two prominent subgenomic RNAs of approximately 3 kb and 1 kb. Putative 5′ ends of the sgRNAs were predicted by identification of conserved sequences and secondary structures which resembled the Barley yellow dwarf virus (BYDV) genomic RNA 5′ end and subgenomic RNA promoter sequences. Secondary structures of the BYDV-like ribosomal frameshift elements and cap-independent translation elements, including long-distance base pairing spanning four kb were identified. These contain similarities but also informative differences with the BYDV structures, including a strikingly different structure predicted for the 3′ cap-independent translation element. These analyses of the RSDaV genomic RNA show more complexity for the RNA structural elements for members of the Luteoviridae. © 2008 Elsevier Inc. All rights reserved. Keywords: Rose spring dwarf; Luteoviridae; Luteovirus; BYDV
Introduction Viruses in the family Luteoviridae are among the most ecologically successful and economically important of the plant viruses (Harrison, 1999). All members of the family have isometric virions containing single-stranded positive-sense ssRNA genomes, infections are phloem-limited in plant hosts and the viruses are aphid transmitted in a circulative, non-propagative manner (D'Arcy and Domier, 2005). The viruses of the family Luteoviridae are divided into three genera, Enamovirus, Luteovirus and Polerovirus, based on genome organization, sequence similarities and gene-expression strategies (D'Arcy and Domier, 2005). The luteovirid genome contains five to six major open reading frames ⁎ Corresponding author. Fax: +1 530 752 5674. E-mail address: [email protected] (B.W. Falk). 0042-6822/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.virol.2008.01.035
(ORFs) designated ORF 0 through ORF 6. Some luteovirids possess one or two additional ORFs (Ashoub et al., 1998; Miller et al., 2002). ORF 0 is limited to the Enamovirus and Polerovirus genera and the protein it encodes, P0, exhibits silencing suppressor activities (Pazhouhandeh et al., 2006). ORFs 1 and 2 encode the replication-related proteins. In all three genera, ORF 2 is expressed via a-1 translational frameshift from ORF 1, thereby giving an ORF 1/2 fusion protein. ORF 1 overlaps ORF 2 by less than 20 nt in luteoviruses, but by more than 400 nt in enamo- and poleroviruses. ORFs 3 and 5 encode the coat and readthrough proteins, respectively. ORF 4, which is lacking in the enamoviruses, encodes a putative movement protein (Liu et al., 2005). Some luteoviruses exhibit properties of more than one genus and phylogenetic analyses suggest that recombination has played an important role in the generation of some species within Luteoviridae (Miller et al., 1995). For example, Bean leafroll virus
N.'M. Salem et al. / Virology 375 (2008) 354–360
(BLRV) and Soybean dwarf virus (SbDV) have been shown to be recombinants between viruses from the genera Polerovirus and Luteovirus, because they contain polerovirus-like capsid genes but also luteovirus-like polymerase genes (Domier et al., 2002; Rathjen et al., 1994). Similarly, Sugarcane yellow leaf virus (ScYLV) is also a recombinant virus, but with a luteovirus-like capsid and a polerovirus-like polymerase gene (Maia et al., 2000; Moonan et al., 2000; Smith et al., 2000). In addition to their characteristic ORF organization, some luteovirus genomic (and subgenomic) RNAs have been shown to contain nucleotide sequence elements that specifically facilitate some of the unique gene-expression and genomereplication strategies used by these viruses (Miller and White, 2006). Luteovirus genomic RNAs lack a 5′ cap. Instead, longdistance RNA–RNA interactions between the 3′ BTE (Barley yellow dwarf virus [BYDV]-like cap-independent translation element) and 5′ BCL (BTE complementary loop) regions have been shown to mediate cap-independent translation (Guo et al., 2001). Characteristic nucleotide sequence elements also facilitate the luteovirus-1 ribosomal frameshift and ORF 3–5 readthrough events that are typical of the translational strategy used by luteoviruses (Dreher and Miller, 2006). We recently discovered a previously undescribed luteovirus associated with the Rose spring dwarf disease (Salem et al., in press). In light of the diverse properties exhibited by viruses comprising the family Luteoviridae, we proceeded to determine how this virus, Rose spring dwarf-associated virus (RSDaV) compared to other luteoviruses. Here we report the genomic organization and expression strategy, and identification of replication/translation nucleotide sequence elements of RSDaV. This monocot and dicot-infecting virus is most closely related to BYDV, and exhibits new genomic diversity among viruses of the Luteoviridae.
355
identity with SbDV, BLRV and BYDV-PAV. ORF 2 was more conserved with 66%, 65% and 62% identity to SbDV, BLRV and BYDV-PAV, respectively. There were no significant alignments with members of the Enamovirus or Polerovirus genera. ORF 2 encodes the active site of the viral RNA-dependent RNA polymerase (RdRp) as it contains the core RNA polymerase motif GXXXTXXXN(X25–40) GDD, where X represents any amino acid (Kamer and Agros, 1984; Koonin, 1991). A non-coding intergenic sequence of 117 nucleotides separates ORF 2 from the second block of coding sequence. The characteristic arrangement of the 3′ block of the three ORFs (ORF 3, 4 and ORF 5) common to the Luteovirus and Polerovirus genera (Miller et al., 2002) is conserved in RSDaV genome. ORF 3 (nts 3006–3665), predicted to encode the 24 kDa RSDaV coat protein (CP), is the first ORF to initiate after the non-coding sequence. ORF 4 is contained completely within ORF 3, in a different reading frame (nts 3055–3612), and encodes a 20 kDa protein. Finally, ORF 5 (nts 3006–5045), predicted to encode the RSDaV CP–readthrough protein, is positioned directly downstream of, and contiguous with, ORF 3. The CP–readthrough fusion protein (ORF 3 + ORF 5 product) has a molecular weight of 75 kDa. The predicted RSDaV CP amino acid sequence is most similar to that of BYDV-MAV (44% identity) followed by ScYLV (42% identity) and BYDV-GAV (39% identity). The ORF 4-encoded protein is most similar to BYDV-PAV, BYDVMAV and BYDV-GAV, with amino acid sequence identities of 42%, 42% and 40%, respectively. The ORF 5-encoded protein has 34 to 42% amino acid sequence identity, with similar proteins encoded by BYDV-PAV, BYDV-MAV and BYDV-GAV. In addition to the large ORFs, three smaller ORFs (ORFs 6, 7 and 8) capable of encoding proteins with 4.5 kDa, 3.5 kDa and 4.4 kDa, respectively, are present in the 3′ region RSDaV genome (Fig. 1). Subgenomic RNAs
Results and discussion RSDaV RNA nucleotide sequence analysis and genome organization We chose to use RSDaV-specific double-stranded RNAs (dsRNAs) as templates for cDNA cloning. We showed previously that three dsRNAs were consistently isolated from RSDaV-infected rose plants, but not healthy plants (Salem et al., in press), and here these proved to be good templates for cDNA synthesis. Approximately seventy plasmids from the cDNA and RT-PCR libraries were used to determine the entire sequence of the RSDaV genomic RNA, with at least two plasmids covering each region. The nucleotide sequence of the RSDaV genomic RNA (GenBank accession no. EU024678) was 5808 nt and showed five large ORFs arranged in two groups (Fig. 1). Three, putative small ORFs also were identified and are discussed below. The RSDaV genomic RNA has a non-coding leader sequence of 186 nt. ORF 1 (nts 187–1308) and ORF 2 (nts 1308– 2888) were predicted to encode proteins of 41 kDa and 58 kDa, respectively. Their predicted amino acid sequences were most similar to those of SbDV, followed by BLRV and BYDV-PAV. BLASTX comparisons with RSDaV ORF 1 revealed 36%
The RSDaV genome organization predicts that subgenomic RNAs are generated during infection in order to express downstream ORFs. Therefore, to detect subgenomic RNAs in infected plants, northern blot hybridization analysis was performed using total RNAs and dsRNAs extracted from RSDaV-infected rose plants, with probes complementary to three different regions of the RSDaV genome. When the ORF 2specific probe (transcribed from pRSDaV5′, Fig. 1) was used, only the genomic-length RNA was detected (Fig. 2, lane 1). The CP-specific probe (transcribed from pRSDaVCP, Fig. 1) hybridized with the genomic RNA and with a smaller RNA, designated as sgRNA1, of about 3.0 kb (Fig. 2, lane 2). The 3′ end-specific probe (transcribed from pRSDaV3′, Fig. 1) hybridized with both the genomic RNA and sgRNA1, but also with a third RNA of approximately 1.0 kb, designated sgRNA2 (Fig. 2, lane 3). Identical hybridization patterns were obtained when using total (data not shown) or dsRNAs from RSDaV-infected rose plants using the same three probes. No hybridization signals were observed in the healthy plant RNAs. It is likely that the 3.0 kb RNA is sgRNA1, which is the mRNA for ORFs 3, 4 and 5 of all Luteoviridae, and that the ∼1.0 kb RNA corresponds to the smaller 3′ co-terminal
356
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 1. Schematic representation of the RSDaV genome organization. Numbered boxes represent open reading frames, with functions of RNA-dependent RNA polymerase (pol) and coat protein (CP) indicated. Gray bars above the genome map represent the positions of the indicated RNA probes used in the northern blot hybridizations (Fig. 2). Bold black lines indicate genomic (gRNA) and subgenomic RNAs (sgRNA), with 5′ ends of sgRNAs predicted as discussed in the text. Black boxes on genome indicate predicted cis-acting structures (left to right): BTE complementary loop of genomic RNA (gBCL, Fig. 3), shifty heptanucleotide and bulged stem–loop at frameshift site (FS, Fig. 4), BTE complementary loop at predicted 5′ end of sgRNA1 (sg1BCL, Fig. 3), BYDV-like cap-independent translation element (BTE, Fig. 5) and long-distance frameshift element that interacts with the frameshift site (LDFE, Fig. 4).
sgRNA2 found in BYDV-infected cells (Kelly et al., 1994; Koev et al., 1999; Shen and Miller, 2004a). To predict the 5′ ends of the sgRNAs, we searched for sequences that resembled the 5′ end of RSDaV genomic RNA (positive strand replication initiation sites) and the 5′ ends of the known genomic and subgenomic RNAs of BYDV. The BYDV genomic RNA begins with AGUGAAG which resembles the GUGAAG sequence at the 5′ ends of sgRNAs 1 and 2. The 5′ terminal sequence of RSDaV RNA, AGUAAAG, differs from BYDVonly at the G to A transition at the fourth position. Aligning this with sequences around the likely 5′ ends of RSDaV sgRNAs 1 and 2, we found candidate initiation sites for those RNAs at nt 2815 or 2817 for sgRNA1, and nt 4953 for sgRNA2 (Fig. 3A). The 5′ ends of sgRNAs predicted by this alignment result in sgRNA1 being 2993 or 2995 nt, and an 857 nt sgRNA2, which is in good agreement with the sizes of the sgRNAs estimated by northern blot hybridization (Fig. 2). In addition to their nucleotide sequence conservation, the conserved elements in the 5′ ends of BYDV genomic RNA and sgRNAs 1 and 2 are in stem–loop structures (Koev and Miller, 2000). Our analyses showed also that the predicted 5′ ends of all three RSDaV RNAs are in
Fig. 2. Northern hybridization analysis of double-stranded RSDaV RNAs from infected plant tissue. Three probes specific for various regions of the genome were derived from pRSDaV5′ (complementary to bases 1938–2574), lane 1; pRSDaVCP (bases 3201–3694), lane 2; and pRSDaV3′ (complementary to bases 5075–5799), lane 3. Mobilities of genomic RNA (gRNA) (∼ 6 kb), subgenomic RNA1 (sgRNA1) (∼3 kb) and subgenomic RNA2 (sgRNA2) (∼ 1 kb) are indicated. Also see Fig. 1 for probe positions. Sizes of the RNAs were estimated from stained RNA standards in the same gel (not shown).
stem–loop structures (Fig. 3B). Importantly, the loops at the 5′ end of genomic RNA and the predicted end of sgRNA1 contain a sequence, UGACA (bold italics, Fig. 3B), complementary to a loop in the putative BTE in the RSDaV 3′ UTR (see below). It is noteworthy that the RSDaV genomic RNA and those of members of genus Luteovirus contain two or three similarlysized and positioned small putative ORFs in the 3′ region downstream of ORF 5. RSDaV sgRNA2 maps to this region and is of the size for the mRNA for ORF 6. sgRNAs similar in size have been identified for BYDV-PAV and some other members of the Luteoviridae (Ashoub et al., 1998; Shen and Miller, 2004a). BYDV-PAV sgRNA2 serves as a trans-acting a riboregulator, negatively affecting translation of the genomic RNA but not sgRNA1, and likely promoting the switch to genomic RNA replication (Shen and Miller, 2004a; Shen et al., 2006). sgRNA2 of BYDV appears to be untranslatable in vivo (Shen et al., 2006). Whether or not RSDaV sgRNA2 performs a similar function is not known, but seems likely based on the conserved sequences and structures seen here. Translational control sequences As with other members of the Luteoviridae, RSDaV ORF 2, which encodes the catalytic domain of the RNA-dependent RNA polymerase, is predicted to be translated via a-1 ribosomal frameshifting during translation of ORF 1 in the region where the two ORFs overlap. This would generate ORF 1/2 fusion protein. Two key nucleotide sequence signals characteristic of a-1 frameshift were identified in the RSDaV sequence. First, a shifty heptanucleotide (GGUUUUU) was found at the end of the 82 nt tract where ORFs 1 and 2 overlap (nts 1302–1308). This fits the consensus sequence XXXYYYZ where X is any base, Y is usually U or A, and Z is any base except G (Brierley, 1995; Brierley and Pennell, 2001). The sequence is identical to the shifty heptanucleotide of Red clover necrotic mosaic virus (RCNMV) (Kim and Lommel, 1998). Secondly, the nucleotide sequence immediately following the shifty heptanucleotide is predicted to form large bulged stem–loop structure similar to that at the frameshift site of BYDV-PAV RNA (Fig. 4). This includes a bulge loop that is capable of base pairing to a stem– loop located 4 kb downstream in the 3′ UTR. An analogous
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 3. Predicted 5′ ends of RSDaV sgRNA1 and sgRNA2. A. Alignment of 5′ end of the RSDaV genomic RNA with sequences located at sites consistent with the 5′ ends of sgRNAs. The conserved sequence at the 5′ ends of BYDV genomic and sgRNAs 1 and 2 is shown below the RSDaV sequence. Bases in gray do not fit the consensus. There are two sites, nts 2815, and 2817 that fit consensus start sites for sgRNA1. B. Predicted secondary structures around the known (genomic RNA) and predicted (sgRNAs) 5′ ends of RSDaV RNAs. Bent arrows, bases in bold indicate the predicted 5′ ends of sgRNAs. Bases in loops, in bold italic are predicted to base pair with the 3′ BTE (see Fig. 5).
357
long-distance interaction (Fig. 4) is essential for frameshifting on BYDV RNA (Barry and Miller, 2002). Viruses in the Luteovirus, Necrovirus and Dianthovirus genera harbor a BTE in the 3′ UTR (Guo et al., 2001; Mizumoto et al., 2003; Shen and Miller, 2004b). The BTE consists of a 17 nt consensus sequence, GGAUCCUGGGAAACAGG, and a nearby stable stem–loop (SL-III) in which the loop base pairs to a stem–loop in the 5′ UTR, forming a kissing stem–loop interaction that is required for cap-independent translation (Guo et al., 2000). We found these features in the 3′ UTR of RSDaV RNA (Fig. 5). Bases 5190–5206 fit the 17 nt consensus with a G → U change at base 10 (in the loop of SL-I; compare Fig. 5A and B). This difference is found in BTEs in the necroviruses (Shen and Miller, 2007) and does not deviate from the GNRNA loop motif (N = any base, R = purine) that is thought to be required in SL-I (Treder et al., 2008). Bases 5257–5275 are predicted to form a GC-rich stem–loop (Fig. 5). Five bases in the loop, UGUCA, are complementary to a loop sequence (UGACA) in the 5′ ends of genomic RNA and subgenomic RNA1 (Fig. 3). These are the same complementary sequences found in the kissing stem–loops of BYDV RNA (Fig. 5) (Guo et al., 2000). Interestingly, the kissing bases in other viruses that harbor a BTE are not always identical to those of BYDV and RSDaV (Miller and White, 2006). The predicted secondary structures for the RSDaV BTE (Fig. 5A) differ markedly from other known BTEs (Miller and White, 2006; Miller et al., 2007). Other BTEs resemble that of BYDV which has three stem–loops connected to the genomic RNA by a fourth helical region, stem-IV (S-IV, Fig. 5B). Some BTEs lack a structural homolog to SL-II, others have five stem– loops (Kneller et al., 2006). In all cases, the stem–loops radiate from a central core. In contrast, the predicted RSDaV BTE
Fig. 4. Predicted secondary structures of frameshift sites of RSDaV and BYDV-PAV RNAs. Predicted long-distance base pairing between a stem–loop 4 kb downstream and a bulged loop adjacent to the frameshift site is indicated at the left. Similar long-distance stem–loop–bulge loop interaction is known to be required for BYDV-PAV frameshifting, (right). The shifty heptanucleotide is in italics. Amino acid sequences of the end of ORF 1 and the overlapping portion of ORF 2 are shown below the shifty site.
358
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 5. Predicted secondary structures of the RSDaV BTE compared with the known structure of the BYDV BTE (boxed in B). The three most stable suboptimal structures of the RSDaV sequence are shown. The 17 nt consensus sequence in all BTEs is in green, and the bases that form a kissing interaction with a stem–loop in the 5′ UTR of genomic RNA and sgRNA1 are in blue bold italics. These two features allow identification of SL-I and SL-III respectively, as in the BYDV BTE. The stem–loop adjacent to SL-III (SL-II) and the helical region upstream of SL-I (S-IV) are present in all 31 MFOLD-predicted structures that have stability (ΔG) within 20% of the most stable.
structures contain no such radiating stem–loops but have large tracts of unpaired or weakly paired bases (Fig. 5A). Because of this, computer predictions revealed many alternative suboptimal structures with very similar predicted minimum free energies (31 with a ΔG within 20% of the most stable). The three most stable structures have virtually identical predicted stabilities (Fig. 5A). Despite this structural “indecision” all predicted structures for the RSDaV BTE contain SL-I, SL-III (with a loop complementary to the 5′ UTR), SL-II immediately upstream of SL-III, and a stem– loop in the approximate position of S-IV, relative to SL-I. Other regions in the sequence show no consistent structure and long single-stranded tracts. We predict that structure (i) in Fig. 5A is the functional BTE because its SL-I and S-IV regions most closely resemble those of BYDV (Fig. 5B). However, the sequence below SL-I, as drawn (Fig. 5A), and between SL-I and the SL-II–SL-III regions is unpredictable and may simply be unstructured (single stranded). Assuming this structure is functional, it reveals the remarkable structural variation tolerated by these cap-independent translation elements. The RSDaV BTE also differs from others in its genomic position relative to the small ORFs in the 3′ end. In BYDV RNA, the BTE is located immediately upstream of the first small ORF (ORF 6) on sgRNA2. In contrast, the BTE of RSDaV overlaps with the 3′ end of ORF 6 and the 5′ end of ORF 7. This supports
the notion that the small ORFs on sgRNA2 may not function as protein coding genes (Shen et al., 2006). Phylogenetic analysis Phylogenetic comparison of the complete genome nucleotide sequence of RSDaV with homologous nucleotide sequences from representatives of the three Luteoviridae genera (Enamovirus, Luteovirus and Polerovirus) revealed that RSDaV is most closely related to members of genus Luteovirus (Fig. 6). Similar results were obtained when only the RSDaV RdRp and CP were compared with the above viruses in Salem et al. (in press). Our data show that the RSDaV genomic RNA has the typical genomic organization and predicted cis-acting control sequences of members of the genus Luteovirus, family Luteoviridae. This is further supported by phylogenetic analysis of RSDaV ORFs 1/ 2 and CP predicted amino acid sequences with other viruses from family Luteoviridae, which positioned RSDaV in the genus Luteovirus. In fact, the RSDaV genomic RNA sequence is most similar to that of BYDV-PAV, including sharing sequence elements controlling cap-independent translation, and a-1 frameshifting at the ORF 1–2 overlap and 3′ UTR sequence. RSDaV is a new member of the Luteoviridae, and adds to the currently recognized genomic diversity for the viruses in this
N.'M. Salem et al. / Virology 375 (2008) 354–360
359
mercuric hydroxide and heated at 94 °C during 5 min. The SuperScript Choice System for cDNA Synthesis (Invitrogen Corp., Carlsbad, CA) was used to construct a cDNA library from viral RNA following the manufacturer's instructions. The resulting dsDNAs were ligated to pGEM-T Easy (Promega, Madison, WI) and plasmids were transformed into Escherichia coli 10G supreme electrocompetent cells (Lucigen Corp., Middleton, WI). Recombinant colonies were screened and their plasmids were isolated using the FastPlasmid Mini kit (Eppendorf, Westbury, NY), digested with EcoR1 (New England Biolabs, Beverly, MA) to determine insert size and analyzed using agarose gel electrophoresis. Nucleotide sequencing and analysis
Fig. 6. Phylogenetic relationship between the RSDaV complete genome nucleotide sequence compared with other members of the family Luteoviridae. The tree was constructed by using the Minimum evolution algorithm provided in the MEGA 2 software package (Kumar et al., 2001). The numbers at each node indicate the bootstrap values resulting from 1000 replicates. Barley yellow dwarf virus-GAV (BYDV-GAV), Barley yellow dwarf virus-MAV (BYDVMAV), Barley yellow dwarf virus-PAS (BYDV-PAS), Barley yellow dwarf virus-PAV (BYDV-PAV), Bean leafroll virus (BLRV), Beet chlorosis virus (BChV), Beet mild yellowing virus (BMYV), Beet western yellows virus (BWYV), Cereal yellow dwarf virus-RPS (CYDV-RPS), Cereal yellow dwarf virus-RPV (CYDV-RPV), Cucurbit aphid-borne yellows virus (CABYV), Pea enation mosaic virus-1 (PEMV-1), Potato leafroll virus (PLRV), Soybean dwarf virus (SbDV), Sugarcane yellow leaf virus (ScYLV) and Turnip yellows virus (TuYV).
important family. While RSDaV is most similar at the genomic level to BYDV-PAV, it shares some biological features as well. One of the aphid vectors for RSDaV is Metapolophium dirhodum, the rose-grass aphid, also is a vector for BYDV-PAV. RSDaV has a host range including oats, an important host for BYDV-PAV, but RSDaV-infected oats are symptomless (Salem et al., in press). Materials and methods Virus source and dsRNA isolation The RSDaV-infected rose (Rose of Freedom) plants used in this study were obtained from the virus-indexed rose block at Foundation Plant Services (FPS), Davis, CA. Thirty grams of bark scrapings were collected and powdered in liquid nitrogen. dsRNA was extracted and purified using two rounds of CF-11 column chromatography, followed by treatment with Ribonuclease A, Proteinase K and Deoxyribonuclease I–RNase-free to eliminate contaminating single-stranded (ss) RNA, ssDNA and dsDNA (Valverde et al., 1990). The isolated dsRNAs were analyzed by 6% polyacrylamide gel electrophoresis (PAGE) and detected by ethidium bromide staining. cDNA synthesis and cloning cDNAs were synthesized using 200 ng dsRNAs as templates. Initially, dsRNAs were denatured with 20 mM methyl
Nucleotide sequencing was performed on both cloned cDNA strands by using an AB1373 Automated Sequencer at the CAES Genomics Facility (CGF), University of California (Davis, CA, USA). Sequences were analyzed using sequence analysis and data management software from Invitrogen, Vector NTI Advance™ 10 (InforMax, Nort Bethesda, MD). The assembly was done with ContigExpress. Genome structure, identification of open reading frames (ORFs) and conserved domains, as well as translated protein sequences were obtained using the BLASTN, BLASTX and ORF finder available at http://www. ncbi.nlm.nih.gov/. Missing sequence gaps were generated by using RT-PCR and primers designed based on the internal sequences. The 5′ and 3′ terminal sequences were determined by using 5′/3′ RACE Kit (Invitrogen). All PCR products were recovered and purified using Zymoclean Gel DNA Recovery Kit (Zymo Research Corp., Orange, CA) and then cloned and sequenced as above. Northern hybridization Three RNA probes corresponding to different RSDaV genomic regions were used in northern hybridization to identify RSDaV genomic and subgenomic (sg) RNAs. Total RNAs extracted with the RNeasy Plant Mini Kit (Qiagen, Valencia, CA), or dsRNAs, were denatured with glyoxal and dimethylsulfoxide (Sambrook and Russell, 2001), separated on a 1% agarose gel, and transferred to Hybond-N+ membranes (Amersham Biosciences, Piscataway, NJ). [α-32P]-UTP-labeled RNA probes were generated from each cloned fragment using the Sp6/T7 MAXIscript in vitro Transcription kit (Ambion, Austin, Texas). Prehybridization and hybridization steps were carried out at 65 °C using commercial PerfectHyb Plus™ buffer from Sigma. Phylogenetic analysis RSDaV phylogenetic relationships were inferred by comparing the complete genome nucleotide sequence of RSDaV to other members of the Luteoviridae. Sequences were aligned with ClustalX (Thompson et al., 1997). Phylogenetic trees were constructed using MEGA2 (Kumar et al., 2001). Sequences were obtained from the GenBank database under the following GenBank accession nos: BYDV-GAV, NC_004666; BYDV-
360
N.'M. Salem et al. / Virology 375 (2008) 354–360
MAV, NC_003680; BYDV-PAS, NC_002160; BYDV-PAV, NC_004750; BLRV, NC_003369; Beet chlorosis virus (BChV), NC_002766; Beet mild yellowing virus (BMYV), NC_003491; Beet western yellows virus (BWYV), NC_004756; Cereal yellow dwarf virus-RPS (CYDV-RPS), NC_002198, CYDV-RPV, NC_004751; Cucurbit aphid-borne yellows virus (CABYV), NC_003688; Pea enation mosaic virus-1 (PEMV1), NC_003629; PLRV, NC_001747; SbDV, NC_003056; ScYLV, NC_000874 and Turnip yellows virus (TuYV), NC_003743. Acknowledgments N. Salem is grateful to FULBRIGHT for a postdoctoral fellowship. NMS was supported by funding provided by the Garden Rose Council and the University of California, Davis. WAM was funded by NIH grant GM067104, and USDA NRI grant 2004-35600-14227. References Ashoub, A., Rohde, W., PrÜfer, D., 1998. In planta transcription of a second subgenomic RNA increases the complexity of the subgroup 2 luteovirus genome. Nucleic Acids Res. 26, 420–426. Barry, J.K., Miller, W.A., 2002. A programmed-1 ribosomal frameshift that requires base-pairing across four kilobases suggests a novel mechanism for controlling ribosome and replicase traffic on a viral RNA. Proc. Natl. Acad. Sci. U. S. A. 99, 11133–11138. Brierley, I., 1995. Ribosomal frameshifting on viral RNAs. J. Gen. Virol. 76, 1885–1892. Brierley, I., Pennell, S., 2001. Structure and function of the stimulatory RNAs involved in programmed eukaryotic-1 ribosomal frameshifting. The Ribosome. Cold Spring Harbor Symposia on Quantitative Biology, LXVI, vol. 66. Cold Spring Harbor Laboratory Press, New York, pp. 233–248. D'Arcy, C.J., Domier, L.L., 2005. Family Luteoviridae. In: Fauquet, C.M., Mayo, M.A., Maniloff, J., Desselberger, U., Ball, L.A. (Eds.), Virus Taxonomy: Eighth Report of the International Committee on Taxonomy of Viruses. Academic Press, pp. 343–352. Domier, L.L., Mccoppin, N.K., Larsen, R.C., D'Arcy, C.J., 2002. Nucleotide sequence shows that Bean leafroll virus has a luteovirus-like genome organization. J. Gen. Virol. 83, 1791–1798. Dreher, T.W., Miller, W.A., 2006. Translational control in positive strand RNA plant viruses. Virology 344, 185–197. Guo, L., Allen, E., Miller, W.A., 2000. Structure and function of a capindependent translation element that functions in either the 3' or the 5' untranslated region. RNA 6, 1808–1820. Guo, L., Allen, E., Miller, W.A., 2001. Base-pairing between untranslated regions facilitates translation of uncapped, nonpolyadenylated viral RNA. Mol. Cell 7, 1103–1109. Harrison, B.D., 1999. Steps in the development of Luteovirology. In: Smith, H., Backer, H. (Eds.), The Luteoviridae. CABI Publishing, Wallingford, pp. 1–14. Kamer, G., Agros, P., 1984. Primary structural comparison of RNA-dependent polymerases from plant, animal and bacterial viruses. Nucleic Acids Res. 12, 7269–7282. Kelly, L., Gerlach, W.L., Waterhouse, P.M., 1994. Characterisation of the subgenomic RNAs of an Australian isolate of barley yellow dwarf luteovirus. Virology 202, 565–573. Kim, K.H., Lommel, S.A., 1998. Sequence element required for efficient-1 ribosomal frameshifting in red clover necrotic mosaic dianthovirus. Virology 250, 50–59. Kneller, E.L., Rakotondrafara, A.M., Miller, W.A., 2006. Cap-independent translation of plant viral RNAs. Virus Res. 119, 63–75. Koev, G., Miller, W.A., 2000. A positive strand RNA virus with three very different subgenomic RNA promoters. J. Virol. 74, 5988–5996.
Koev, G., Mohan, B.R., Miller, W.A., 1999. Primary and secondary structural elements required for synthesis of barley yellow dwarf virus subgenomic RNA1. J. Virol. 73, 2876–2885. Koonin, E.V., 1991. The phylogeny of RNA-dependent RNA polymerases of positive-strand RNA viruses. J. Gen. Virol. 72, 2197–2206. Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17, 1244–1245. Liu, K., Xia, Z., Zhang, Y., Wen, Y., Wang, D., Brandenburq, K., Harris, F., Phoenix, D.A., 2005. Interaction between the movement protein of barley yellow dwarf virus and the cell nuclear envelope: role of a putative amphiphilic alpha-helix at the N-terminus of the movement protein. Biopolymers 79, 86–96. Maia, I.G., Gonclaves, M.C., Arruda, P., Vega, J., 2000. Molecular evidence that sugarcane yellow leaf virus (ScYLV) is a member of the Luteoviridae family. Arch. Virol. 145, 1009–1019. Miller, W.A., Dineshkumar, S.P., Paul, C.P., 1995. Luteovirus gene expression. Crit. Rev. Plant Sci. 14, 179–211. Miller, W.A., Liu, S., Beckett, R., 2002. Barley yellow dwarf virus: Luteoviridae or Tombusviridae? Mol. Plant Pathol. 3, 177–183. Miller, W.A., White, K.A., 2006. Long-distance RNA–RNA interactions in plant virus gene expression and replication. Annu. Rev. Phytopathol. 44, 447–467. Miller, W.A., Treder, K., Wang, Z., 2007. The amazing diversity of capindependent translation elements in the 3' untranslated regions of plant viral RNAs. Bioch. Soc. Trans. 36, 1629–1633. Mizumoto, H., Tatsuta, M., Kaido, M., Mise, K., Okuno, T., 2003. Capindependent translational enhancement by the 3' untranslated region of red clover necrotic mosaic virus RNA1. J. Virol. 77, 12113–12121. Moonan, F., Molina, J., Mirkov, T.E., 2000. Sugarcane yellow leaf virus: an emerging virus that has evolved by recombination luteoviral and poleroviral ancestors. Virology 269, 156–171. Pazhouhandeh, M., Dieterle, M., Marrocco, K., Lechner, E., Berry, B., Brault, V., Hemmer, O., Kretsch, T., Richards, K., Genschik, P., Ziegler-Graff, V., 2006. F-box-like domain in the polerovirus protein P0 is required for silencing suppressor function. PANS 103, 1994–1999. Rathjen, J.P., Karageorgos, L.E., Habili, N., Waterhouse, P.M., Symons, R.H., 1994. Soybean dwarf luteovirus contains the third variant genome type in the luteovirus group. Virology 198, 671–679. Salem, N., Golino, D.A., Falk, B.W., Rowhani, A., in press. Identification and partial characterization of a new luteovirus associated with rose spring dwarf disease. Plant Dis. Sambrook, J., Russell, D., 2001. Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Shen, R., Miller, W.A., 2004a. Subgenomic RNA as a riboregulator: negative regulation of RNA replication by Barley yellow dwarf virus subgenomic RNA 2. Virology 327, 196–205. Shen, R., Miller, W.A., 2004b. The 3' untranslated region of tobacco necrosis virus RNA contains a barley yellow dwarf virus-like cap-independent translation element. J. Virol. 78, 4655–4664. Shen, R., Miller, W.A., 2007. Structures required for poly(A) tail-independent translation overlap with, but are distinct from, cap-independent translation and RNA replication signals at the 3' end of Tobacco necrosis virus RNA. Virology 358, 448–458. Shen, R., Rakotondrafara, A.M., Miller, W.A., 2006. trans regulation of capindependent translation by a viral subgenomic RNA. J. Virol. 80, 10045–10054. Smith, G.R., Borg, Z., Lockhart, B.E.L., Braithwaite, K.S., Gibbs, M.J., 2000. Sugarcane yellow leaf virus: a novel member of the Luteoviridae that probably arose by inter-species recombination. J. Gen. Virol. 81, 1865–1869. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The CLUSTALX windows interface; flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882. Treder, K., Pettit Kneller, E., Allen, E.M., Wang, Z., Browning, K.S., Miller, W.A., 2008. The 3' cap-independent translation element of Barley yellow dwarf virus binds eIF4F via the eIF4G subunit to initiate translation. RNA 14, 134–147. Valverde, R.A., Nameth, S.T., Jordan, R.L., 1990. Analysis of double-stranded RNA for plant virus diagnosis. Plant Dis. 74 (3), 255–258.
Virology 375 (2008) 354 – 360 www.elsevier.com/locate/yviro
Rose spring dwarf-associated virus has RNA structural and gene-expression features like those of Barley yellow dwarf virus Nida' M. Salem a , W. Allen Miller b,c , Adib Rowhani a , Deborah A. Golino a , Anne-Laure Moyne a , Bryce W. Falk a,⁎ a
c
Department of Plant Pathology, One Shields Ave., University of California, Davis, CA 95616, USA b Plant Pathology Department, Iowa State University, Ames, Iowa 50011, USA Biochemistry, Biophysics and Molecular Biology Department, Iowa State University, Ames, Iowa 50011, USA Received 21 September 2007; returned to author for revision 29 November 2007; accepted 27 January 2008 Available online 7 March 2008
Abstract We determined the complete nucleotide sequence of the Rose spring dwarf-associated virus (RSDaV) genomic RNA (GenBank accession no. EU024678) and compared its predicted RNA structural characteristics affecting gene expression. A cDNA library was derived from RSDaV double-stranded RNAs (dsRNAs) purified from infected tissue. Nucleotide sequence analysis of the cloned cDNAs, plus for clones generated by 5′- and 3′-RACE showed the RSDaV genomic RNA to be 5808 nucleotides. The genomic RNA contains five major open reading frames (ORFs), and three small ORFs in the 3′-terminal 800 nucleotides, typical for viruses of genus Luteovirus in the family Luteoviridae. Northern blot hybridization analysis revealed the genomic RNA and two prominent subgenomic RNAs of approximately 3 kb and 1 kb. Putative 5′ ends of the sgRNAs were predicted by identification of conserved sequences and secondary structures which resembled the Barley yellow dwarf virus (BYDV) genomic RNA 5′ end and subgenomic RNA promoter sequences. Secondary structures of the BYDV-like ribosomal frameshift elements and cap-independent translation elements, including long-distance base pairing spanning four kb were identified. These contain similarities but also informative differences with the BYDV structures, including a strikingly different structure predicted for the 3′ cap-independent translation element. These analyses of the RSDaV genomic RNA show more complexity for the RNA structural elements for members of the Luteoviridae. © 2008 Elsevier Inc. All rights reserved. Keywords: Rose spring dwarf; Luteoviridae; Luteovirus; BYDV
Introduction Viruses in the family Luteoviridae are among the most ecologically successful and economically important of the plant viruses (Harrison, 1999). All members of the family have isometric virions containing single-stranded positive-sense ssRNA genomes, infections are phloem-limited in plant hosts and the viruses are aphid transmitted in a circulative, non-propagative manner (D'Arcy and Domier, 2005). The viruses of the family Luteoviridae are divided into three genera, Enamovirus, Luteovirus and Polerovirus, based on genome organization, sequence similarities and gene-expression strategies (D'Arcy and Domier, 2005). The luteovirid genome contains five to six major open reading frames ⁎ Corresponding author. Fax: +1 530 752 5674. E-mail address: [email protected] (B.W. Falk). 0042-6822/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.virol.2008.01.035
(ORFs) designated ORF 0 through ORF 6. Some luteovirids possess one or two additional ORFs (Ashoub et al., 1998; Miller et al., 2002). ORF 0 is limited to the Enamovirus and Polerovirus genera and the protein it encodes, P0, exhibits silencing suppressor activities (Pazhouhandeh et al., 2006). ORFs 1 and 2 encode the replication-related proteins. In all three genera, ORF 2 is expressed via a-1 translational frameshift from ORF 1, thereby giving an ORF 1/2 fusion protein. ORF 1 overlaps ORF 2 by less than 20 nt in luteoviruses, but by more than 400 nt in enamo- and poleroviruses. ORFs 3 and 5 encode the coat and readthrough proteins, respectively. ORF 4, which is lacking in the enamoviruses, encodes a putative movement protein (Liu et al., 2005). Some luteoviruses exhibit properties of more than one genus and phylogenetic analyses suggest that recombination has played an important role in the generation of some species within Luteoviridae (Miller et al., 1995). For example, Bean leafroll virus
N.'M. Salem et al. / Virology 375 (2008) 354–360
(BLRV) and Soybean dwarf virus (SbDV) have been shown to be recombinants between viruses from the genera Polerovirus and Luteovirus, because they contain polerovirus-like capsid genes but also luteovirus-like polymerase genes (Domier et al., 2002; Rathjen et al., 1994). Similarly, Sugarcane yellow leaf virus (ScYLV) is also a recombinant virus, but with a luteovirus-like capsid and a polerovirus-like polymerase gene (Maia et al., 2000; Moonan et al., 2000; Smith et al., 2000). In addition to their characteristic ORF organization, some luteovirus genomic (and subgenomic) RNAs have been shown to contain nucleotide sequence elements that specifically facilitate some of the unique gene-expression and genomereplication strategies used by these viruses (Miller and White, 2006). Luteovirus genomic RNAs lack a 5′ cap. Instead, longdistance RNA–RNA interactions between the 3′ BTE (Barley yellow dwarf virus [BYDV]-like cap-independent translation element) and 5′ BCL (BTE complementary loop) regions have been shown to mediate cap-independent translation (Guo et al., 2001). Characteristic nucleotide sequence elements also facilitate the luteovirus-1 ribosomal frameshift and ORF 3–5 readthrough events that are typical of the translational strategy used by luteoviruses (Dreher and Miller, 2006). We recently discovered a previously undescribed luteovirus associated with the Rose spring dwarf disease (Salem et al., in press). In light of the diverse properties exhibited by viruses comprising the family Luteoviridae, we proceeded to determine how this virus, Rose spring dwarf-associated virus (RSDaV) compared to other luteoviruses. Here we report the genomic organization and expression strategy, and identification of replication/translation nucleotide sequence elements of RSDaV. This monocot and dicot-infecting virus is most closely related to BYDV, and exhibits new genomic diversity among viruses of the Luteoviridae.
355
identity with SbDV, BLRV and BYDV-PAV. ORF 2 was more conserved with 66%, 65% and 62% identity to SbDV, BLRV and BYDV-PAV, respectively. There were no significant alignments with members of the Enamovirus or Polerovirus genera. ORF 2 encodes the active site of the viral RNA-dependent RNA polymerase (RdRp) as it contains the core RNA polymerase motif GXXXTXXXN(X25–40) GDD, where X represents any amino acid (Kamer and Agros, 1984; Koonin, 1991). A non-coding intergenic sequence of 117 nucleotides separates ORF 2 from the second block of coding sequence. The characteristic arrangement of the 3′ block of the three ORFs (ORF 3, 4 and ORF 5) common to the Luteovirus and Polerovirus genera (Miller et al., 2002) is conserved in RSDaV genome. ORF 3 (nts 3006–3665), predicted to encode the 24 kDa RSDaV coat protein (CP), is the first ORF to initiate after the non-coding sequence. ORF 4 is contained completely within ORF 3, in a different reading frame (nts 3055–3612), and encodes a 20 kDa protein. Finally, ORF 5 (nts 3006–5045), predicted to encode the RSDaV CP–readthrough protein, is positioned directly downstream of, and contiguous with, ORF 3. The CP–readthrough fusion protein (ORF 3 + ORF 5 product) has a molecular weight of 75 kDa. The predicted RSDaV CP amino acid sequence is most similar to that of BYDV-MAV (44% identity) followed by ScYLV (42% identity) and BYDV-GAV (39% identity). The ORF 4-encoded protein is most similar to BYDV-PAV, BYDVMAV and BYDV-GAV, with amino acid sequence identities of 42%, 42% and 40%, respectively. The ORF 5-encoded protein has 34 to 42% amino acid sequence identity, with similar proteins encoded by BYDV-PAV, BYDV-MAV and BYDV-GAV. In addition to the large ORFs, three smaller ORFs (ORFs 6, 7 and 8) capable of encoding proteins with 4.5 kDa, 3.5 kDa and 4.4 kDa, respectively, are present in the 3′ region RSDaV genome (Fig. 1). Subgenomic RNAs
Results and discussion RSDaV RNA nucleotide sequence analysis and genome organization We chose to use RSDaV-specific double-stranded RNAs (dsRNAs) as templates for cDNA cloning. We showed previously that three dsRNAs were consistently isolated from RSDaV-infected rose plants, but not healthy plants (Salem et al., in press), and here these proved to be good templates for cDNA synthesis. Approximately seventy plasmids from the cDNA and RT-PCR libraries were used to determine the entire sequence of the RSDaV genomic RNA, with at least two plasmids covering each region. The nucleotide sequence of the RSDaV genomic RNA (GenBank accession no. EU024678) was 5808 nt and showed five large ORFs arranged in two groups (Fig. 1). Three, putative small ORFs also were identified and are discussed below. The RSDaV genomic RNA has a non-coding leader sequence of 186 nt. ORF 1 (nts 187–1308) and ORF 2 (nts 1308– 2888) were predicted to encode proteins of 41 kDa and 58 kDa, respectively. Their predicted amino acid sequences were most similar to those of SbDV, followed by BLRV and BYDV-PAV. BLASTX comparisons with RSDaV ORF 1 revealed 36%
The RSDaV genome organization predicts that subgenomic RNAs are generated during infection in order to express downstream ORFs. Therefore, to detect subgenomic RNAs in infected plants, northern blot hybridization analysis was performed using total RNAs and dsRNAs extracted from RSDaV-infected rose plants, with probes complementary to three different regions of the RSDaV genome. When the ORF 2specific probe (transcribed from pRSDaV5′, Fig. 1) was used, only the genomic-length RNA was detected (Fig. 2, lane 1). The CP-specific probe (transcribed from pRSDaVCP, Fig. 1) hybridized with the genomic RNA and with a smaller RNA, designated as sgRNA1, of about 3.0 kb (Fig. 2, lane 2). The 3′ end-specific probe (transcribed from pRSDaV3′, Fig. 1) hybridized with both the genomic RNA and sgRNA1, but also with a third RNA of approximately 1.0 kb, designated sgRNA2 (Fig. 2, lane 3). Identical hybridization patterns were obtained when using total (data not shown) or dsRNAs from RSDaV-infected rose plants using the same three probes. No hybridization signals were observed in the healthy plant RNAs. It is likely that the 3.0 kb RNA is sgRNA1, which is the mRNA for ORFs 3, 4 and 5 of all Luteoviridae, and that the ∼1.0 kb RNA corresponds to the smaller 3′ co-terminal
356
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 1. Schematic representation of the RSDaV genome organization. Numbered boxes represent open reading frames, with functions of RNA-dependent RNA polymerase (pol) and coat protein (CP) indicated. Gray bars above the genome map represent the positions of the indicated RNA probes used in the northern blot hybridizations (Fig. 2). Bold black lines indicate genomic (gRNA) and subgenomic RNAs (sgRNA), with 5′ ends of sgRNAs predicted as discussed in the text. Black boxes on genome indicate predicted cis-acting structures (left to right): BTE complementary loop of genomic RNA (gBCL, Fig. 3), shifty heptanucleotide and bulged stem–loop at frameshift site (FS, Fig. 4), BTE complementary loop at predicted 5′ end of sgRNA1 (sg1BCL, Fig. 3), BYDV-like cap-independent translation element (BTE, Fig. 5) and long-distance frameshift element that interacts with the frameshift site (LDFE, Fig. 4).
sgRNA2 found in BYDV-infected cells (Kelly et al., 1994; Koev et al., 1999; Shen and Miller, 2004a). To predict the 5′ ends of the sgRNAs, we searched for sequences that resembled the 5′ end of RSDaV genomic RNA (positive strand replication initiation sites) and the 5′ ends of the known genomic and subgenomic RNAs of BYDV. The BYDV genomic RNA begins with AGUGAAG which resembles the GUGAAG sequence at the 5′ ends of sgRNAs 1 and 2. The 5′ terminal sequence of RSDaV RNA, AGUAAAG, differs from BYDVonly at the G to A transition at the fourth position. Aligning this with sequences around the likely 5′ ends of RSDaV sgRNAs 1 and 2, we found candidate initiation sites for those RNAs at nt 2815 or 2817 for sgRNA1, and nt 4953 for sgRNA2 (Fig. 3A). The 5′ ends of sgRNAs predicted by this alignment result in sgRNA1 being 2993 or 2995 nt, and an 857 nt sgRNA2, which is in good agreement with the sizes of the sgRNAs estimated by northern blot hybridization (Fig. 2). In addition to their nucleotide sequence conservation, the conserved elements in the 5′ ends of BYDV genomic RNA and sgRNAs 1 and 2 are in stem–loop structures (Koev and Miller, 2000). Our analyses showed also that the predicted 5′ ends of all three RSDaV RNAs are in
Fig. 2. Northern hybridization analysis of double-stranded RSDaV RNAs from infected plant tissue. Three probes specific for various regions of the genome were derived from pRSDaV5′ (complementary to bases 1938–2574), lane 1; pRSDaVCP (bases 3201–3694), lane 2; and pRSDaV3′ (complementary to bases 5075–5799), lane 3. Mobilities of genomic RNA (gRNA) (∼ 6 kb), subgenomic RNA1 (sgRNA1) (∼3 kb) and subgenomic RNA2 (sgRNA2) (∼ 1 kb) are indicated. Also see Fig. 1 for probe positions. Sizes of the RNAs were estimated from stained RNA standards in the same gel (not shown).
stem–loop structures (Fig. 3B). Importantly, the loops at the 5′ end of genomic RNA and the predicted end of sgRNA1 contain a sequence, UGACA (bold italics, Fig. 3B), complementary to a loop in the putative BTE in the RSDaV 3′ UTR (see below). It is noteworthy that the RSDaV genomic RNA and those of members of genus Luteovirus contain two or three similarlysized and positioned small putative ORFs in the 3′ region downstream of ORF 5. RSDaV sgRNA2 maps to this region and is of the size for the mRNA for ORF 6. sgRNAs similar in size have been identified for BYDV-PAV and some other members of the Luteoviridae (Ashoub et al., 1998; Shen and Miller, 2004a). BYDV-PAV sgRNA2 serves as a trans-acting a riboregulator, negatively affecting translation of the genomic RNA but not sgRNA1, and likely promoting the switch to genomic RNA replication (Shen and Miller, 2004a; Shen et al., 2006). sgRNA2 of BYDV appears to be untranslatable in vivo (Shen et al., 2006). Whether or not RSDaV sgRNA2 performs a similar function is not known, but seems likely based on the conserved sequences and structures seen here. Translational control sequences As with other members of the Luteoviridae, RSDaV ORF 2, which encodes the catalytic domain of the RNA-dependent RNA polymerase, is predicted to be translated via a-1 ribosomal frameshifting during translation of ORF 1 in the region where the two ORFs overlap. This would generate ORF 1/2 fusion protein. Two key nucleotide sequence signals characteristic of a-1 frameshift were identified in the RSDaV sequence. First, a shifty heptanucleotide (GGUUUUU) was found at the end of the 82 nt tract where ORFs 1 and 2 overlap (nts 1302–1308). This fits the consensus sequence XXXYYYZ where X is any base, Y is usually U or A, and Z is any base except G (Brierley, 1995; Brierley and Pennell, 2001). The sequence is identical to the shifty heptanucleotide of Red clover necrotic mosaic virus (RCNMV) (Kim and Lommel, 1998). Secondly, the nucleotide sequence immediately following the shifty heptanucleotide is predicted to form large bulged stem–loop structure similar to that at the frameshift site of BYDV-PAV RNA (Fig. 4). This includes a bulge loop that is capable of base pairing to a stem– loop located 4 kb downstream in the 3′ UTR. An analogous
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 3. Predicted 5′ ends of RSDaV sgRNA1 and sgRNA2. A. Alignment of 5′ end of the RSDaV genomic RNA with sequences located at sites consistent with the 5′ ends of sgRNAs. The conserved sequence at the 5′ ends of BYDV genomic and sgRNAs 1 and 2 is shown below the RSDaV sequence. Bases in gray do not fit the consensus. There are two sites, nts 2815, and 2817 that fit consensus start sites for sgRNA1. B. Predicted secondary structures around the known (genomic RNA) and predicted (sgRNAs) 5′ ends of RSDaV RNAs. Bent arrows, bases in bold indicate the predicted 5′ ends of sgRNAs. Bases in loops, in bold italic are predicted to base pair with the 3′ BTE (see Fig. 5).
357
long-distance interaction (Fig. 4) is essential for frameshifting on BYDV RNA (Barry and Miller, 2002). Viruses in the Luteovirus, Necrovirus and Dianthovirus genera harbor a BTE in the 3′ UTR (Guo et al., 2001; Mizumoto et al., 2003; Shen and Miller, 2004b). The BTE consists of a 17 nt consensus sequence, GGAUCCUGGGAAACAGG, and a nearby stable stem–loop (SL-III) in which the loop base pairs to a stem–loop in the 5′ UTR, forming a kissing stem–loop interaction that is required for cap-independent translation (Guo et al., 2000). We found these features in the 3′ UTR of RSDaV RNA (Fig. 5). Bases 5190–5206 fit the 17 nt consensus with a G → U change at base 10 (in the loop of SL-I; compare Fig. 5A and B). This difference is found in BTEs in the necroviruses (Shen and Miller, 2007) and does not deviate from the GNRNA loop motif (N = any base, R = purine) that is thought to be required in SL-I (Treder et al., 2008). Bases 5257–5275 are predicted to form a GC-rich stem–loop (Fig. 5). Five bases in the loop, UGUCA, are complementary to a loop sequence (UGACA) in the 5′ ends of genomic RNA and subgenomic RNA1 (Fig. 3). These are the same complementary sequences found in the kissing stem–loops of BYDV RNA (Fig. 5) (Guo et al., 2000). Interestingly, the kissing bases in other viruses that harbor a BTE are not always identical to those of BYDV and RSDaV (Miller and White, 2006). The predicted secondary structures for the RSDaV BTE (Fig. 5A) differ markedly from other known BTEs (Miller and White, 2006; Miller et al., 2007). Other BTEs resemble that of BYDV which has three stem–loops connected to the genomic RNA by a fourth helical region, stem-IV (S-IV, Fig. 5B). Some BTEs lack a structural homolog to SL-II, others have five stem– loops (Kneller et al., 2006). In all cases, the stem–loops radiate from a central core. In contrast, the predicted RSDaV BTE
Fig. 4. Predicted secondary structures of frameshift sites of RSDaV and BYDV-PAV RNAs. Predicted long-distance base pairing between a stem–loop 4 kb downstream and a bulged loop adjacent to the frameshift site is indicated at the left. Similar long-distance stem–loop–bulge loop interaction is known to be required for BYDV-PAV frameshifting, (right). The shifty heptanucleotide is in italics. Amino acid sequences of the end of ORF 1 and the overlapping portion of ORF 2 are shown below the shifty site.
358
N.'M. Salem et al. / Virology 375 (2008) 354–360
Fig. 5. Predicted secondary structures of the RSDaV BTE compared with the known structure of the BYDV BTE (boxed in B). The three most stable suboptimal structures of the RSDaV sequence are shown. The 17 nt consensus sequence in all BTEs is in green, and the bases that form a kissing interaction with a stem–loop in the 5′ UTR of genomic RNA and sgRNA1 are in blue bold italics. These two features allow identification of SL-I and SL-III respectively, as in the BYDV BTE. The stem–loop adjacent to SL-III (SL-II) and the helical region upstream of SL-I (S-IV) are present in all 31 MFOLD-predicted structures that have stability (ΔG) within 20% of the most stable.
structures contain no such radiating stem–loops but have large tracts of unpaired or weakly paired bases (Fig. 5A). Because of this, computer predictions revealed many alternative suboptimal structures with very similar predicted minimum free energies (31 with a ΔG within 20% of the most stable). The three most stable structures have virtually identical predicted stabilities (Fig. 5A). Despite this structural “indecision” all predicted structures for the RSDaV BTE contain SL-I, SL-III (with a loop complementary to the 5′ UTR), SL-II immediately upstream of SL-III, and a stem– loop in the approximate position of S-IV, relative to SL-I. Other regions in the sequence show no consistent structure and long single-stranded tracts. We predict that structure (i) in Fig. 5A is the functional BTE because its SL-I and S-IV regions most closely resemble those of BYDV (Fig. 5B). However, the sequence below SL-I, as drawn (Fig. 5A), and between SL-I and the SL-II–SL-III regions is unpredictable and may simply be unstructured (single stranded). Assuming this structure is functional, it reveals the remarkable structural variation tolerated by these cap-independent translation elements. The RSDaV BTE also differs from others in its genomic position relative to the small ORFs in the 3′ end. In BYDV RNA, the BTE is located immediately upstream of the first small ORF (ORF 6) on sgRNA2. In contrast, the BTE of RSDaV overlaps with the 3′ end of ORF 6 and the 5′ end of ORF 7. This supports
the notion that the small ORFs on sgRNA2 may not function as protein coding genes (Shen et al., 2006). Phylogenetic analysis Phylogenetic comparison of the complete genome nucleotide sequence of RSDaV with homologous nucleotide sequences from representatives of the three Luteoviridae genera (Enamovirus, Luteovirus and Polerovirus) revealed that RSDaV is most closely related to members of genus Luteovirus (Fig. 6). Similar results were obtained when only the RSDaV RdRp and CP were compared with the above viruses in Salem et al. (in press). Our data show that the RSDaV genomic RNA has the typical genomic organization and predicted cis-acting control sequences of members of the genus Luteovirus, family Luteoviridae. This is further supported by phylogenetic analysis of RSDaV ORFs 1/ 2 and CP predicted amino acid sequences with other viruses from family Luteoviridae, which positioned RSDaV in the genus Luteovirus. In fact, the RSDaV genomic RNA sequence is most similar to that of BYDV-PAV, including sharing sequence elements controlling cap-independent translation, and a-1 frameshifting at the ORF 1–2 overlap and 3′ UTR sequence. RSDaV is a new member of the Luteoviridae, and adds to the currently recognized genomic diversity for the viruses in this
N.'M. Salem et al. / Virology 375 (2008) 354–360
359
mercuric hydroxide and heated at 94 °C during 5 min. The SuperScript Choice System for cDNA Synthesis (Invitrogen Corp., Carlsbad, CA) was used to construct a cDNA library from viral RNA following the manufacturer's instructions. The resulting dsDNAs were ligated to pGEM-T Easy (Promega, Madison, WI) and plasmids were transformed into Escherichia coli 10G supreme electrocompetent cells (Lucigen Corp., Middleton, WI). Recombinant colonies were screened and their plasmids were isolated using the FastPlasmid Mini kit (Eppendorf, Westbury, NY), digested with EcoR1 (New England Biolabs, Beverly, MA) to determine insert size and analyzed using agarose gel electrophoresis. Nucleotide sequencing and analysis
Fig. 6. Phylogenetic relationship between the RSDaV complete genome nucleotide sequence compared with other members of the family Luteoviridae. The tree was constructed by using the Minimum evolution algorithm provided in the MEGA 2 software package (Kumar et al., 2001). The numbers at each node indicate the bootstrap values resulting from 1000 replicates. Barley yellow dwarf virus-GAV (BYDV-GAV), Barley yellow dwarf virus-MAV (BYDVMAV), Barley yellow dwarf virus-PAS (BYDV-PAS), Barley yellow dwarf virus-PAV (BYDV-PAV), Bean leafroll virus (BLRV), Beet chlorosis virus (BChV), Beet mild yellowing virus (BMYV), Beet western yellows virus (BWYV), Cereal yellow dwarf virus-RPS (CYDV-RPS), Cereal yellow dwarf virus-RPV (CYDV-RPV), Cucurbit aphid-borne yellows virus (CABYV), Pea enation mosaic virus-1 (PEMV-1), Potato leafroll virus (PLRV), Soybean dwarf virus (SbDV), Sugarcane yellow leaf virus (ScYLV) and Turnip yellows virus (TuYV).
important family. While RSDaV is most similar at the genomic level to BYDV-PAV, it shares some biological features as well. One of the aphid vectors for RSDaV is Metapolophium dirhodum, the rose-grass aphid, also is a vector for BYDV-PAV. RSDaV has a host range including oats, an important host for BYDV-PAV, but RSDaV-infected oats are symptomless (Salem et al., in press). Materials and methods Virus source and dsRNA isolation The RSDaV-infected rose (Rose of Freedom) plants used in this study were obtained from the virus-indexed rose block at Foundation Plant Services (FPS), Davis, CA. Thirty grams of bark scrapings were collected and powdered in liquid nitrogen. dsRNA was extracted and purified using two rounds of CF-11 column chromatography, followed by treatment with Ribonuclease A, Proteinase K and Deoxyribonuclease I–RNase-free to eliminate contaminating single-stranded (ss) RNA, ssDNA and dsDNA (Valverde et al., 1990). The isolated dsRNAs were analyzed by 6% polyacrylamide gel electrophoresis (PAGE) and detected by ethidium bromide staining. cDNA synthesis and cloning cDNAs were synthesized using 200 ng dsRNAs as templates. Initially, dsRNAs were denatured with 20 mM methyl
Nucleotide sequencing was performed on both cloned cDNA strands by using an AB1373 Automated Sequencer at the CAES Genomics Facility (CGF), University of California (Davis, CA, USA). Sequences were analyzed using sequence analysis and data management software from Invitrogen, Vector NTI Advance™ 10 (InforMax, Nort Bethesda, MD). The assembly was done with ContigExpress. Genome structure, identification of open reading frames (ORFs) and conserved domains, as well as translated protein sequences were obtained using the BLASTN, BLASTX and ORF finder available at http://www. ncbi.nlm.nih.gov/. Missing sequence gaps were generated by using RT-PCR and primers designed based on the internal sequences. The 5′ and 3′ terminal sequences were determined by using 5′/3′ RACE Kit (Invitrogen). All PCR products were recovered and purified using Zymoclean Gel DNA Recovery Kit (Zymo Research Corp., Orange, CA) and then cloned and sequenced as above. Northern hybridization Three RNA probes corresponding to different RSDaV genomic regions were used in northern hybridization to identify RSDaV genomic and subgenomic (sg) RNAs. Total RNAs extracted with the RNeasy Plant Mini Kit (Qiagen, Valencia, CA), or dsRNAs, were denatured with glyoxal and dimethylsulfoxide (Sambrook and Russell, 2001), separated on a 1% agarose gel, and transferred to Hybond-N+ membranes (Amersham Biosciences, Piscataway, NJ). [α-32P]-UTP-labeled RNA probes were generated from each cloned fragment using the Sp6/T7 MAXIscript in vitro Transcription kit (Ambion, Austin, Texas). Prehybridization and hybridization steps were carried out at 65 °C using commercial PerfectHyb Plus™ buffer from Sigma. Phylogenetic analysis RSDaV phylogenetic relationships were inferred by comparing the complete genome nucleotide sequence of RSDaV to other members of the Luteoviridae. Sequences were aligned with ClustalX (Thompson et al., 1997). Phylogenetic trees were constructed using MEGA2 (Kumar et al., 2001). Sequences were obtained from the GenBank database under the following GenBank accession nos: BYDV-GAV, NC_004666; BYDV-
360
N.'M. Salem et al. / Virology 375 (2008) 354–360
MAV, NC_003680; BYDV-PAS, NC_002160; BYDV-PAV, NC_004750; BLRV, NC_003369; Beet chlorosis virus (BChV), NC_002766; Beet mild yellowing virus (BMYV), NC_003491; Beet western yellows virus (BWYV), NC_004756; Cereal yellow dwarf virus-RPS (CYDV-RPS), NC_002198, CYDV-RPV, NC_004751; Cucurbit aphid-borne yellows virus (CABYV), NC_003688; Pea enation mosaic virus-1 (PEMV1), NC_003629; PLRV, NC_001747; SbDV, NC_003056; ScYLV, NC_000874 and Turnip yellows virus (TuYV), NC_003743. Acknowledgments N. Salem is grateful to FULBRIGHT for a postdoctoral fellowship. NMS was supported by funding provided by the Garden Rose Council and the University of California, Davis. WAM was funded by NIH grant GM067104, and USDA NRI grant 2004-35600-14227. References Ashoub, A., Rohde, W., PrÜfer, D., 1998. In planta transcription of a second subgenomic RNA increases the complexity of the subgroup 2 luteovirus genome. Nucleic Acids Res. 26, 420–426. Barry, J.K., Miller, W.A., 2002. A programmed-1 ribosomal frameshift that requires base-pairing across four kilobases suggests a novel mechanism for controlling ribosome and replicase traffic on a viral RNA. Proc. Natl. Acad. Sci. U. S. A. 99, 11133–11138. Brierley, I., 1995. Ribosomal frameshifting on viral RNAs. J. Gen. Virol. 76, 1885–1892. Brierley, I., Pennell, S., 2001. Structure and function of the stimulatory RNAs involved in programmed eukaryotic-1 ribosomal frameshifting. The Ribosome. Cold Spring Harbor Symposia on Quantitative Biology, LXVI, vol. 66. Cold Spring Harbor Laboratory Press, New York, pp. 233–248. D'Arcy, C.J., Domier, L.L., 2005. Family Luteoviridae. In: Fauquet, C.M., Mayo, M.A., Maniloff, J., Desselberger, U., Ball, L.A. (Eds.), Virus Taxonomy: Eighth Report of the International Committee on Taxonomy of Viruses. Academic Press, pp. 343–352. Domier, L.L., Mccoppin, N.K., Larsen, R.C., D'Arcy, C.J., 2002. Nucleotide sequence shows that Bean leafroll virus has a luteovirus-like genome organization. J. Gen. Virol. 83, 1791–1798. Dreher, T.W., Miller, W.A., 2006. Translational control in positive strand RNA plant viruses. Virology 344, 185–197. Guo, L., Allen, E., Miller, W.A., 2000. Structure and function of a capindependent translation element that functions in either the 3' or the 5' untranslated region. RNA 6, 1808–1820. Guo, L., Allen, E., Miller, W.A., 2001. Base-pairing between untranslated regions facilitates translation of uncapped, nonpolyadenylated viral RNA. Mol. Cell 7, 1103–1109. Harrison, B.D., 1999. Steps in the development of Luteovirology. In: Smith, H., Backer, H. (Eds.), The Luteoviridae. CABI Publishing, Wallingford, pp. 1–14. Kamer, G., Agros, P., 1984. Primary structural comparison of RNA-dependent polymerases from plant, animal and bacterial viruses. Nucleic Acids Res. 12, 7269–7282. Kelly, L., Gerlach, W.L., Waterhouse, P.M., 1994. Characterisation of the subgenomic RNAs of an Australian isolate of barley yellow dwarf luteovirus. Virology 202, 565–573. Kim, K.H., Lommel, S.A., 1998. Sequence element required for efficient-1 ribosomal frameshifting in red clover necrotic mosaic dianthovirus. Virology 250, 50–59. Kneller, E.L., Rakotondrafara, A.M., Miller, W.A., 2006. Cap-independent translation of plant viral RNAs. Virus Res. 119, 63–75. Koev, G., Miller, W.A., 2000. A positive strand RNA virus with three very different subgenomic RNA promoters. J. Virol. 74, 5988–5996.
Koev, G., Mohan, B.R., Miller, W.A., 1999. Primary and secondary structural elements required for synthesis of barley yellow dwarf virus subgenomic RNA1. J. Virol. 73, 2876–2885. Koonin, E.V., 1991. The phylogeny of RNA-dependent RNA polymerases of positive-strand RNA viruses. J. Gen. Virol. 72, 2197–2206. Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17, 1244–1245. Liu, K., Xia, Z., Zhang, Y., Wen, Y., Wang, D., Brandenburq, K., Harris, F., Phoenix, D.A., 2005. Interaction between the movement protein of barley yellow dwarf virus and the cell nuclear envelope: role of a putative amphiphilic alpha-helix at the N-terminus of the movement protein. Biopolymers 79, 86–96. Maia, I.G., Gonclaves, M.C., Arruda, P., Vega, J., 2000. Molecular evidence that sugarcane yellow leaf virus (ScYLV) is a member of the Luteoviridae family. Arch. Virol. 145, 1009–1019. Miller, W.A., Dineshkumar, S.P., Paul, C.P., 1995. Luteovirus gene expression. Crit. Rev. Plant Sci. 14, 179–211. Miller, W.A., Liu, S., Beckett, R., 2002. Barley yellow dwarf virus: Luteoviridae or Tombusviridae? Mol. Plant Pathol. 3, 177–183. Miller, W.A., White, K.A., 2006. Long-distance RNA–RNA interactions in plant virus gene expression and replication. Annu. Rev. Phytopathol. 44, 447–467. Miller, W.A., Treder, K., Wang, Z., 2007. The amazing diversity of capindependent translation elements in the 3' untranslated regions of plant viral RNAs. Bioch. Soc. Trans. 36, 1629–1633. Mizumoto, H., Tatsuta, M., Kaido, M., Mise, K., Okuno, T., 2003. Capindependent translational enhancement by the 3' untranslated region of red clover necrotic mosaic virus RNA1. J. Virol. 77, 12113–12121. Moonan, F., Molina, J., Mirkov, T.E., 2000. Sugarcane yellow leaf virus: an emerging virus that has evolved by recombination luteoviral and poleroviral ancestors. Virology 269, 156–171. Pazhouhandeh, M., Dieterle, M., Marrocco, K., Lechner, E., Berry, B., Brault, V., Hemmer, O., Kretsch, T., Richards, K., Genschik, P., Ziegler-Graff, V., 2006. F-box-like domain in the polerovirus protein P0 is required for silencing suppressor function. PANS 103, 1994–1999. Rathjen, J.P., Karageorgos, L.E., Habili, N., Waterhouse, P.M., Symons, R.H., 1994. Soybean dwarf luteovirus contains the third variant genome type in the luteovirus group. Virology 198, 671–679. Salem, N., Golino, D.A., Falk, B.W., Rowhani, A., in press. Identification and partial characterization of a new luteovirus associated with rose spring dwarf disease. Plant Dis. Sambrook, J., Russell, D., 2001. Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Shen, R., Miller, W.A., 2004a. Subgenomic RNA as a riboregulator: negative regulation of RNA replication by Barley yellow dwarf virus subgenomic RNA 2. Virology 327, 196–205. Shen, R., Miller, W.A., 2004b. The 3' untranslated region of tobacco necrosis virus RNA contains a barley yellow dwarf virus-like cap-independent translation element. J. Virol. 78, 4655–4664. Shen, R., Miller, W.A., 2007. Structures required for poly(A) tail-independent translation overlap with, but are distinct from, cap-independent translation and RNA replication signals at the 3' end of Tobacco necrosis virus RNA. Virology 358, 448–458. Shen, R., Rakotondrafara, A.M., Miller, W.A., 2006. trans regulation of capindependent translation by a viral subgenomic RNA. J. Virol. 80, 10045–10054. Smith, G.R., Borg, Z., Lockhart, B.E.L., Braithwaite, K.S., Gibbs, M.J., 2000. Sugarcane yellow leaf virus: a novel member of the Luteoviridae that probably arose by inter-species recombination. J. Gen. Virol. 81, 1865–1869. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The CLUSTALX windows interface; flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882. Treder, K., Pettit Kneller, E., Allen, E.M., Wang, Z., Browning, K.S., Miller, W.A., 2008. The 3' cap-independent translation element of Barley yellow dwarf virus binds eIF4F via the eIF4G subunit to initiate translation. RNA 14, 134–147. Valverde, R.A., Nameth, S.T., Jordan, R.L., 1990. Analysis of double-stranded RNA for plant virus diagnosis. Plant Dis. 74 (3), 255–258.