Transcripts encoding the nanovirus master replication ... - CiteSeerX

Journal of General Virology (2008), 89, 583–593

DOI 10.1099/vir.0.83352-0

Transcripts encoding the nanovirus master replication initiator proteins are terminally redundant Ioana Grigoras, Tatiana Timchenko and Bruno Gronenborn Correspondence

Institut des Sciences du Ve´ge´tal, CNRS, 91198 Gif sur Yvette, France

Bruno Gronenborn [email protected]

Received 2 August 2007 Accepted 3 October 2007

The multicomponent single-stranded DNA plant nanoviruses encode unique master replication initiator (Rep) proteins. We have mapped the 59 and 39 termini of the corresponding polyadenylated mRNAs from faba bean necrotic yellows virus (FBNYV) and subterranean clover stunt virus and found that these are terminally redundant by up to about 160 nt. Moreover, the origin of viral DNA replication is transcribed into RNA that is capable of folding into extended secondary structures. Other nanovirus genome components, such as the FBNYV DNA encoding the protein Clink or an FBNYV DNA encoding a non-essential para-Rep protein, are not transcribed in such a unique fashion. Thus, terminally redundant mRNAs and the resulting transcription of the replication origin appear to be restricted to nanovirus master Rep DNAs. We speculate that this may be a way to regulate the expression of the essential master Rep protein.

INTRODUCTION Members of the family Nanoviridae comprise the species Subterranean clover stunt virus (SCSV), Faba bean necrotic yellows virus (FBNYV) and Milk vetch dwarf virus (MDV) in the genus Nanovirus, and Banana bunchy top virus (BBTV) in the genus Babuvirus (Vetten et al., 2005). The former viruses are pathogens of a wide variety of legumes in Australia (SCSV), Japan (MDV), and west Asia, north and east Africa and Spain (FBNYV), whereas BBTV only infects bananas in south Asia, the Pacific region and some African countries. The nanovirus genome consists of six to eight circular, single-stranded DNA (ssDNA) molecules of about 1 kb (Vetten et al., 2005) (Fig. 1). In general, one nanovirus DNA encodes only one protein. Common to all nanoviruses is a unique master replication initiator protein (M-Rep), encoded by DNA-R, that catalyses replication initiation and termination of its cognate and all other genome components of the respective virus (Timchenko et al., 1999, 2000; Horser et al., 2001). Other genome components encode a capsid protein (DNA-S) (Chu et al., 1993; Katul et al., 1997; Wanitchakorn et al., 1997), a cellcycle link protein (Clink) (DNA-C) (Aronson et al., 2000), a movement protein (DNA-M) and a nuclear shuttle protein (DNA-N) (Wanitchakorn et al., 2000). Several DNAs encode proteins of as-yet-unknown function (DNAU1, -U2, -U3 and -U4) (Vetten et al., 2005). Also, the biological significance of a small open reading frame (ORF) embedded completely within the m-rep gene of some BBTV isolates remains obscure (Beetham et al., 1997). A supplementary figure showing transcripts of FBNYV DNA-R, SCSV DNA-R, FBNYV C11 and FBNYV DNA-C and a supplementary table showing oligonucleotides used in this study are available with the online version of this paper.

0008-3352 G 2008 SGM

Printed in Great Britain

In addition to these ‘integral’ genome components, a considerable number of non-essential DNAs, all encoding Rep proteins, were found associated with infections by all four nanoviruses (Wu et al., 1994; Vetten et al., 2005; Hu et al., 2007). These additional and non-essential rep genes can be considered as paralogues (para-rep) of the essential m-rep gene. Their precise role in the epidemiology of the nanovirus with which they associate remains unknown. However, based on the results obtained in an experimental infection system, it was suggested that the para-Repencoding DNAs may play a role in modulating the severity of nanovirus-induced disease (Timchenko et al., 2006). All Rep proteins recognize the origin of replication of viral DNA (ori) and cleave it in a conserved nonamer sequence, which is flanked by inverted repeats (IRs) (Timchenko et al., 1999). This way, the Rep proteins trigger the initiation of viral positive-strand DNA synthesis, which then continues by rolling-circle replication, very similar to their action in other ssDNA viruses (Laufs et al., 1995; Timchenko et al., 1999). Transcripts have only been mapped for BBTV (Beetham et al., 1997, 1999; Herrera-Valencia et al., 2007). Transcript data for the legume-infecting nanoviruses are absent; only transcription initiation and termination signals on their respective DNAs have been recognized (Boevink et al., 1995; Katul et al., 1997; Sano et al., 1998; Shirasawa-Seo et al., 2005). All nanovirus ORFs are preceded by a TATA sequence and are followed by an AATAAA-like polyadenylation [poly(A)] addition signal (Fig. 1). Three different arrangements of these transcription ‘start’ and ‘stop’ signals in the vicinity of the IR sequences at the replication origin can be 583

I. Grigoras, T. Timchenko and B. Gronenborn

Fig. 1. Nanovirus DNA components. The nanovirus DNAs (approx. 1 kb) are grouped (a, b, c) according to the respective arrangement of their transcription initiation and termination signals as described in the text. Component names are according to Vetten et al. (2005). Bold arrows represent the positions and extent of the ORFs; encoded proteins are named according to identified functions. M-Rep, Master replication initiator protein; CP, capsid protein; Clink, cell-cycle link protein; MP, movement protein; NSP, nuclear shuttle protein; U1–5, proteins of as-yet-unknown function; para-Rep, replication initiator proteins of nonessential DNAs. Previously used component designations are given below each DNA. Transcripts were mapped for representative examples (component designations are framed) of each group: (a) DNA-R of FBNYV and SCSV, (b) DNA-C of FBNYV and (c) C11 (para-rep11) of FBNYV.

distinguished. (i) In all non-Rep DNAs, in DNA-R of BBTV and in the para-Rep1-encoding DNAs of FBNYV and MDV, TATA boxes are found 39 of the IR sequences, thus preceding the respective genes. AATAAA-like poly(A) addition signals are located 39 of the genes and 59 of the IR sequences (Fig. 1b). (ii) In all para-Rep-encoding DNAs (except for the para-Rep1-encoding DNAs of FBNYV and of MDV), both of these signals are located 59 of the IRs (Fig. 1c), with the TATA box 59-preceding the IR sequence by a few nucleotides. (iii) In DNA-R of FBNYV, MDV and SCSV, both the TATA boxes and the poly(A) addition signals are located 39 of the IR sequences (Fig. 1a). Such an arrangement would lead to the transcription of the IRs and to a potential stem–loop structure in the 39-untranslated region (UTR) of the respective mRNAs. Furthermore, the predicted poly(A) addition signals are located between the TATA box and the ATG codon of the respective m-rep genes, implying the synthesis of a terminally redundant transcript for M-Rep expression. Such a peculiar arrangement of transcription initiation and termination signals may not exist by chance. To determine whether this signal arrangement is reflected by the termini of the mRNAs, we analysed polyadenylated RNAs synthesized from the DNA-R components of the nanoviruses FBNYV and SCSV, and determined the physical start and stop positions of their m-rep transcripts experimentally. For comparison, RNAs synthesized from DNA-C and 584

para-rep11 DNA of FBNYV served as examples of transcripts derived from genome components with a more ‘canonical’ arrangement of transcription initiation and termination signals.

METHODS Nanovirus DNAs. Tandemly repeated copies of DNAs (DNA-C and C11, respectively) from the Egyptian isolate EV1-93 of FBNYV in the binary T-DNA vector pBin19 were used as described previously (Timchenko et al., 1999). Tandem repeats of FBNYV DNA-R plus DNA-C within the same T-DNA vector were described by Timchenko et al. (2006). The SCSV (Australian isolate F) DNA-R was described by Timchenko et al. (2000).

For agroinoculation of Nicotiana benthamiana leaf discs, derivatives of pBin19 carrying the respective nanovirus DNAs were transferred by electroporation into Agrobacterium tumefaciens strain LBA4404 (Hoekema et al., 1983) and used as described previously (Timchenko et al., 1999). In the case of agrobacteria carrying FBNYV C11 or SCSV DNA-R, a co-inoculation with agrobacteria containing FBNYV DNA-C was performed. RNA extraction. Total RNA was extracted from either 200 mg frozen tissue of N. benthamiana leaf discs 1 week after agroinoculation or 200 mg fresh non-inoculated N. benthamiana leaves, using an RNeasy Plant Mini kit (Qiagen). Residual amounts of DNA were removed by a RNase-Free DNase set (Qiagen) for on-column DNase I digestion according to the manufacturer’s protocol. RNA concentration was determined spectrophotometrically.

Journal of General Virology 89

Terminally redundant nanovirus transcripts

Northern blotting analysis. Aliquots of total RNA were electrophoresed in 1.8 % formaldehyde/agarose gels, transferred to HybondXL membranes (Amersham Biosciences) and stained with methylene blue as described by Sambrook et al. (1989). Northern hybridization with a 32P-labelled probe was subsequently carried out in Church’s hybridization buffer (Brown et al., 2004). The probe was a 935 bp fragment of FBNYV DNA-R amplified with the oligonucleotides FBIR2b(2) and C2Nhe-for (see Supplementary Table S1, available in JGV Online). RNA molecular mass markers (0.16–1.77 kb) were from Invitrogen. Rapid amplification of cDNA ends (RACE). Rapid amplification of

the cDNA ends of the nanovirus mRNAs was carried out by using a FirstChoice RLM-RACE kit (Ambion) following the manufacturer’s instructions. Reverse-transcription reactions were carried out at 42 or 50 uC. Oligonucleotides used as primers are listed in Supplementary Table S1 (available in JGV Online). PCR amplifications were performed in a total volume of 50 ml, containing 50 pmol genespecific primer, 20 pmol either 39- or 59-RACE inner primer, 67 mM Tris/HCl (pH 8.8), 16 mM (NH4)2SO4, 0.01 % Tween 20, 1.5 mM MgCl2, 1 mM each dNTP, 5 units Eurobiotaq II DNA polymerase (Eurobio) and 0.2–1.0 ml reverse-transcription reaction. DNAs obtained after 35 cycles (95 uC, 30 s; 55 uC, 30 s; 72 uC, 90 s) were inserted into pBluescript II KS+ (Stratagene) and their sequences were compared with those of FBNYV DNA-R, DNA-C and C11 and SCSV DNA-R (GenBank accession numbers AJ132180, AJ132179, AJ005968 and AJ290434, respectively). Synthesis of a full-length cDNA of FBNYV DNA-R transcripts.

Total RNA extracted from leaf discs agroinoculated by the DNA-R construct was treated according to the FirstChoice 59-RLM-RACE protocol (Ambion). RNA ligated to the 59-RACE adaptor was reversetranscribed at 50 uC in the presence of 5 % DMSO and 50 pmol oligonucleotide C2-3cDNA5pb (see Supplementary Table S1, available in JGV Online). The product of the reverse-transcription reaction (0.5 ml) was amplified by PCR in the presence of 50 pmol C2-5cDNA-Xba primer (see Supplementary Table S1) and 50 pmol 39-RACE inner primer. The PCR mixture was cycled 35 times: 95 uC, 30 s; 45–60 uC gradient, 30 s; 72 uC, 90 s. PCR products from six PCRs, in which the annealing temperature varied from 53 to 60 uC, were inserted into pBluescript II KS+.

specific for a given nanovirus DNA each time. Control reactions omitting the reverse-transcription step yielded no products and demonstrated that the RNA preparations were essentially DNA-free and suitable for transcript analyses. The 59 termini of the respective nanovirus transcripts were determined by RNA ligase-mediated RACE (RLM-RACE), which allows cDNA amplification from capped mRNA only. The presence of the 59-RACE adaptor preceding any viral sequence in the cloned 59-RLM-RACE products led us to conclude that presumably capped 59 ends of the mRNAs were determined. The 39-RACE technique, used for 39-end analyses of the transcripts, led only to amplification of products from polyadenylated RNAs. The FBNYV DNA-R (m-rep) transcript The integrity of FBNYV DNA-R transcripts was assessed by Northern blotting. The FBNYV DNA-R-specific probe hybridized only to RNAs extracted from agroinoculated N. benthamiana leaf discs (Fig. 2). Two major transcripts of approximately 1.1 and 0.8 kb were detected by the DNA-R probe. The initiation site of the FBNYV DNA-R transcripts was determined by 59-RLM-RACE, using the four gene-specific primers C2-Nhe970-as, C2Nhe352rev, C2Nhe388rev and Rep2Spe(2) (see Supplementary Table S1, available in JGV Online). In each case, PCR products with expected sizes of approximately 900, 300, 350 and 600 bp were observed

RESULTS Identification of nanovirus transcripts in N. benthamiana It was shown previously that the nanovirus Rep proteins are functional in N. benthamiana and that the M-Rep protein triggers the replication of the genome components when assayed in leaf discs (Timchenko et al., 1999, 2000). We used this leaf-disc replication system to study the transcripts produced from DNA-R of FBNYV and SCSV, as well as those of C11 (para-rep11) and DNA-C of FBNYV. Total RNA was extracted from agroinoculated N. benthamiana leaf discs as described in Methods. The selective-binding properties of Qiagen’s silica gel-based matrix efficiently removed most of the DNA. Potentially remaining trace amounts of (replicative) DNAs were removed by DNase I digestion. Total RNA was reversetranscribed to produce cDNA, which was then amplified by PCR using different primer pairs, with one primer being http://vir.sgmjournals.org

Fig. 2. Northern blotting analysis of FBNYV DNA-R transcripts. Total RNA (10 mg, lanes 1 and 2; 20 mg, lanes 3 and 4; 40 mg, lanes 5 and 6) extracted from N. benthamiana leaf discs 1 week after agroinoculation (lanes 1, 3 and 5) or from non-inoculated N. benthamiana leaves (lanes 2, 4 and 6) was fractionated by gel electrophoresis and transferred to a Hybond-XL membrane. Right panel, membrane stained with methylene blue; left panel, membrane hybridized to the FBNYV DNA-R-specific probe. MM, Molecular mass marker. The positions of the two major bands of approximately 1.1 and 0.8 kb are indicated by arrows. 585

I. Grigoras, T. Timchenko and B. Gronenborn

586

Journal of General Virology 89

Terminally redundant nanovirus transcripts

Fig. 3. Transcripts of FBNYV DNA-R (a), SCSV DNA-R (b), FBNYV C11 (c) and FBNYV DNA-C (d). Nucleotide sequences of the four nanovirus DNAs are shown in a linear form as a scaled line drawing, starting with the first nucleotide of the IR sequence. Major AATAAA-like poly(A) addition signals are indicated by bold vertical dashes; minor ones are shown as thin vertical dashes. Nucleotide positions corresponding to the 59 and 39 ends of the FBNYV m-rep cDNAs are indicated by an angled arrow (position 105) and a pin (position 158) in (a).

upon 1 % agarose gel electrophoresis. They were cloned and 29 individual PCR products were sequenced. By sequence comparison of the cloned cDNAs with FBNYV DNA-R, the 59 ends of the transcripts were mapped (Fig. 3a; Supplementary Fig. S1a, available in JGV Online). In 25 cases, the 59 end of the mRNA mapped downstream of the predicted TATA sequence (nt 75–82) and upstream of the ATG codon of the m-rep gene: in 10 cases at adenine 105 and in 15 cases at adenine 109. In four additional cases (all obtained when using primer C2-Nhe970-as), the 59 ends mapped to nt 454, 553 and 860, for which no corresponding TATA sequence was evident. The 39-end analysis of the FBNYV DNA-R transcript was performed by using the 39-RACE inner primer and three gene-specific primers: C2Nhe-for, C2HIND+ and REP2K187A (see Supplementary Table S1, available in JGV Online). PCR products with expected sizes of approximately 1000, 300 and 400 bp were obtained and cloned, and 42 inserts were sequenced. The mapped poly(A) sites of the M-Rep mRNA are shown in Fig. 3(a) and Supplementary Fig. S1(a) (available in JGV Online). The only canonical AATAAA sequence (Loke et al., 2005; Dong et al., 2007; and references therein) of FBNYV DNAR (nt 113–118) is located after the TGA stop codon (position 979) of the m-rep gene (Katul et al., 1997). In 21 cases (50 %) of the 39-RACE products sequenced, the poly(A) tail of the mRNA was added 14–33 nt downstream of this poly(A) signal (nt 113–118). Some other AATAAA-like sequences occurring on DNA-R were also found to be used. For instance, in 13 cases, the beginning of the poly(A) tract was mapped downstream of the AATAAT signal at nt 157–162 or the AATGAA signal at position 192–197. Two further AATAAA-like sequences, located closer to the end of the ORF (AATTAA at nt 839– 844 and 882–887, respectively), were used rarely, as we found only three transcripts that ended between nt 890 and 915. A few other 39 ends were mapped to positions for which no corresponding AATAAA-like signals were identified (positions 76, 109, 455, 669 and 702). Thus, in 80 % of cases (34 of 42), the poly(A) signal used was 39 and close to the transcription start. This implies that, initially, transcription bypasses the poly(A) signal, proceeds through the gene and the IR at the origin of replication and terminates after passing the poly(A) signal the second time. The initial bypass of the poly(A) signal did not always occur, as we found, among 34 cloned PCR products, six short ones where polyadenylation had apparently occurred upon the first encounter of the poly(A) signal by the transcription complex. Another five http://vir.sgmjournals.org

transcripts were ‘aberrant’, in that they lacked large parts of the m-rep sequence. As no splice-donor or -acceptor sites were found that could account for the sizes of these products, we consider them artefacts of the reversetranscription reaction performed at 42 uC. Indeed, these shorter PCR products were not obtained when reverse transcription was carried out at 50 uC. Consequently, in all other mapping experiments, the reverse-transcription reaction was carried out at 50 uC. The results of mapping the 59 and 39 termini of the FBNYV DNA-R transcripts imply that the origin of DNA-R replication is transcribed into RNA and that the mRNA encoding the master Rep protein is terminally redundant by 5–161 nt, with a mean length of the terminal redundancy of 28–57 nt. The SCSV DNA-R (m-rep) transcript To investigate whether a similar type of terminally redundant transcript also occurred in another distantly related nanovirus, we mapped the DNA-R transcript of SCSV. The 59 ends of SCSV m-rep transcripts were determined following 59-RLM-RACE PCR using three different genespecific primers: C2HIND32, C2Nhe352rev and Rep2Spe(2) (see Supplementary Table S1, available in JGV Online). The PCR products had sizes of approximately 750, 300 and 600 bp. Following cloning, 39 cDNAs were sequenced. In 28 cases, transcription was initiated downstream of the predicted DNA-R TATA sequence (nt 70–77) at adenines 100 (seven cases), 104 (15 cases) and 107 (two cases) (Fig. 3b; Supplementary Fig. S1b, available in JGV Online). In four cases, adenine 112 of the m-rep AUG codon was identified as the 59 end. In 11 cases, the 59 end of the mRNA mapped to positions for which no corresponding TATA box was identified (positions 203, 250 and 352; Fig. 3b; Supplementary Fig. S1b). In 39-RACE experiments, the gene-specific primers C2HIND+ and Rep2Spe(+) (Supplementary Table S1) were used to determine the 39 ends of the SCSV m-rep transcript. PCR products with expected sizes of approximately 250 and 400 bp were cloned and 32 cDNAs were sequenced. Nearly 50 % of the mRNAs (15 clones) ended 1–40 nt after the AATAAA signal (nt 104–109), the major poly(A) addition site being at nt 127 (Fig. 3b; Supplementary Fig. S1b). In 10 other cases, the poly(A) tail was added further upstream, following the sequence AATATA occurring at nt 68–73; a potential poly(A) signal explains these 39 ends. Six other PCR products were shorter 587

I. Grigoras, T. Timchenko and B. Gronenborn

than expected and their 39 ends were mapped to nt 17, 26, 909, 947, 958 and 1004. The 39 end of a longer-thanexpected product (approx. 650 versus 250 bp) mapped to position 534. No poly(A) signals corresponding to any of these polyadenylation positions could be identified. The FBNYV C11 (para-rep11) transcript A 59-RLM-RACE analysis of the FBNYV C11 transcript was performed with oligonucleotide C11EcoRVrev (see Supplementary Table S1, available in JGV Online) as genespecific primer. The PCR product was approximately 500 bp. Sequence analysis of 24 clones (Fig. 3c; Supplementary Fig. S1c) revealed that the majority of the mRNAs (17 cases) mapped between 24 and 32 nt downstream of the sequence TATATATA located at nt 998–1. Three other transcription initiation sites mapped within the rep11 ORF at nt 117, 127 and 129, corresponding to the TATA sequence TATAAAAC at positions 96–103. If translated, such a transcript could lead only to expression of a polypeptide initiated at an ATG codon at position 231, thus lacking 56 aa at its N terminus. Three other PCR products corresponded to transcription initiation at nt 180, 288 and 298, for which no TATA box was identified. A 39-RACE analysis of the FBNYV C11 transcript was carried out using oligonucleotide C11EcoRVdir (see Supplementary Table S1) as gene-specific primer. The PCR product was approximately 500 bp. Sequence analysis of eight clones mapped the 39 end of the para-rep11 transcript to either guanine 935 (two cases) or adenine 939 (six cases) downstream of the poly(A) signal AATAAA at nt 911–916 (Fig. 3c; Supplementary Fig. S1c). The FBNYV DNA-C (clink) transcript As a representative example of the major group of nanovirus DNAs with respect to their transcription signals (Fig. 1b), we mapped the transcript of FBNYV DNA-C encoding Clink. DNA-C was always included in transfection experiments along with the Rep-encoding DNAs to boost their replication (Aronson et al., 2000). Therefore, each RNA preparation also contained clink mRNA and was used for its mapping. The initiation site of the clink mRNA was determined by using oligonucleotides C10HINC22 and C10LSRRA as gene-specific primers (Supplementary Table S1). The PCR products were approximately 450 and 400 bp. Sequence analysis of 20 cloned cDNAs revealed a major transcription initiation site at adenine 305, downstream of the DNA-C sequence TATAAAAA at nt 273–280 (Fig. 3d; Supplementary Fig. S1d). In one case, the 59 end of the clink mRNA mapped to adenine 307, downstream of the same TATA sequence. Four other plasmids contained inserts that were either longer (three cases) or shorter (one case) than expected. The 59 ends of all of them mapped to positions (204, 209 and 353) for which no corresponding TATA sequences were identified. 588

A 39-RACE analysis was performed by using C10FboxPLdir and C10BAM as gene-specific primers (see Supplementary Table S1). PCR products of approximately 550 and 600 bp were cloned and 14 cDNAs were sequenced. In eight cases, the 39 end of the clink transcript mapped 2–38 nt downstream of the poly(A) sequence AATTAA at nt 832–837 (Fig. 3d; Supplementary Fig. S1d). In two other cases, the transcripts ended upstream of the clink TGA stop codon (position 821), probably due to occurrence of a potential poly(A) signal (AATAAT) at position 776–781. Four additional 39 ends were mapped to positions 499, 513, 538 and 584, for which no corresponding poly(A) signals were identified. Synthesis of a full-length cDNA of FBNYV DNA-R transcripts To confirm that the FBNYV m-rep transcript was indeed terminally redundant and hence longer than its encoding genome component DNA-R, we isolated and sequenced full-length cDNA clones of the m-rep transcript. Based on the 39- and 59-RACE results (Fig. 3a; Supplementary Fig. S1a), we designed the oligonucleotides C2-3cDNA5pb and C2-5cDNA-Xba (see Supplementary Table S1). The first one was used in a reverse-transcription reaction to synthesize cDNA from the polyadenylated RNA. The second primer (C2-5cDNA-Xba) has 26 nt identical to the m-rep gene and 16 nt identical to the 59RACE adaptor. Oligonucleotide C2-5cDNA-Xba was used along with the 39-RACE inner primer in a PCR that allowed the amplification of the full-length m-rep cDNA from capped mRNA. A PCR product with a size of approximately 1.1 kb was obtained and cloned. Three independently derived plasmids were isolated and sequenced. All contained 1157 nt of m-rep cDNA, with a 59 end corresponding to nt 105 and a 39 end at nt 158. They comprised the entire DNA-R (1003 nt) and an additional 54 nt of terminal redundancy. This confirmed the mapping results of the 59 and 39 termini of the m-rep messenger and demonstrated independently that the origin of DNA replication, including the IR sequences, was transcribed into RNA. No deletions indicative of splicing events were observed.

DISCUSSION TATA sequences and AATAAA-like poly(A) addition signals of the m-rep genes were revealed by inspection of their corresponding DNA-R sequences of the nanoviruses FBNYV, MDV and SCSV (Katul et al., 1997; Timchenko et al., 2000). The arrangement of these transcription signals on DNA-R of the three viruses differs from that of the DNA-R of BBTV and of all other genomic or para-repencoding DNAs (Fig. 1): the poly(A) addition signal is located between the TATA box and the coding sequence. If used for transcription initiation and termination, either a Journal of General Virology 89

Terminally redundant nanovirus transcripts

very short, polyadenylated RNA not encoding any protein or, as DNA-R is circular, an mRNA longer than DNA-R encoding the M-Rep protein would be synthesized. Such an mRNA would be terminally redundant and include the replication origin sequence in its 39-UTR. To characterize the nature of the DNA-R transcripts experimentally, we determined the 59 and 39 ends of polyadenylated transcripts from the most distantly related members of the genus Nanovirus, FBNYV and SCSV (Vetten et al., 2005). Mapping of the transcript ends of DNA-R from FBNYV and SCSV indeed revealed a terminal redundancy (Fig. 3; Supplementary Fig. S1, available in JGV Online). In the case of FBNYV, the corresponding transcripts were terminally redundant by 5–161 nt, depending on the respective combination of their 59 and 39 ends. In the case of SCSV DNA-R, the redundancy was up to 49 nt. As MDV DNA-R is 92 % identical in sequence to that of FBNYV and as its transcription initiation and termination signals are at positions equivalent to those of FBNYV, we inferred that similarly redundant transcripts will probably also occur in the case of this legume nanovirus and did not include it in the analyses. Further, independent evidence for the existence of such terminally redundant mRNAs was the isolation of fulllength cDNAs corresponding to the FBNYV m-rep transcript. These contained the entire m-rep ORF and DNA-R’s origin of replication, and were terminally redundant by 54 nt (Fig. 3a; Supplementary Fig. S1a). The size of these cDNAs is in agreement with the size of the approximately 1.1 kb RNAs detected by Northern blotting (Fig. 2). The shorter, approximately 0.8 kb RNAs in the Northern blot might represent truncated, non-polyadenylated transcripts not detected by the oligo(dT) selection. They could be explained by a premature RNA polymerase arrest at the extended secondary structure of the replication origin. Another peculiar feature of these m-rep transcripts is that, among all predicted nanovirus mRNAs or mapped babuvirus transcripts (Beetham et al., 1997, 1999; Herrera-Valencia et al., 2007), they have the longest 39UTR sequences, which also comprise the IR sequences at the origin of replication. As ssDNA, the IR sequences potentially form a stem–loop structure. Transcription of these sequences leads to the presence of the same putative secondary structure in the mRNA’s 39-UTR region. Secondary structures predicted by MFOLD using default parameters (Zuker, 2003) of the FBNYV m-rep mRNA including or lacking a tail of 50 adenylate residues at its 39 end (Fig. 4a, b, d) suggest that this may indeed be the case. The most stable secondary structures generated (Fig. 4a, b, d) had a free energy (DG) of 2293.37 kcal mol21, and the IR sequences form an extended stem–loop. Comparison of these predictions (Fig. 4a, b, d) with that for an artificially trimmed transcript without the terminal redundancy of the last 54 nt (Fig. 4c, e) (DG52279.84 kcal mol21) reveals that the terminal redundancy of the mRNA may play a role http://vir.sgmjournals.org

in enhancing the stability of the secondary structure assumed by the origin of replication as a part of the transcript. The difference in free energy of 13.53 kcal mol21 is in the same range as those found between wild-type Kunjin virus and two cyclization-sequence mutants, where free-energy differences of 11.5 or 14.4 kcal mol21 abolish RNA replication (Khromykh et al., 2001). When random RNAs of the same length and nucleotide composition as the 1157 nt m-rep mRNA, generated by using DISHUFFLE (http://clavius.bc.edu/~clotelab/RNAdinucleotideShuffle/ dinucleotideShuffle.html) (Altschul & Erickson, 1985), were folded by using the same parameters, all secondary structures obtained had a higher free energy than the least stable structure of the m-rep mRNA. This may suggest that the formation of a stem–loop in the 39-UTR of the m-rep mRNA is an important feature for the virus cycle. Such an assumption is consistent with predictions indicating that the RNAs for which a secondary structure is functionally important have a lower free energy of folding than random RNAs of the same nucleotide composition (Clote et al., 2005). As shown in Fig. 4, only the terminal redundancy of the m-rep transcripts allows the formation of an extended stretch of double-stranded RNA, due to pairing of 59- and 39-terminal sequences. Whether this end pairing alone, or enhanced by binding of a protein, interferes with or stimulates the circularization of mRNA by protein members of the translation complexes (Mazumder et al., 2003; Gebauer & Hentze, 2004) remains open. Either way, it may provide a way to regulate M-Rep expression. Transcription through the origin of replication of plant ssDNA viruses other than nanoviruses has only been described for two mastreviruses, Digitaria streak virus (Accotto et al., 1989) and wheat dwarf virus (Dekker et al., 1991), where a minor complementary-sense transcript spanning the origin was found. It was suggested that a stable secondary structure in the 59 leader sequence of the mRNA is inhibitory for the translation of the downstream rep genes. However, the biological significance of such an mRNA species remains to be demonstrated. The existence of the whole genomic information of nanovirus DNA-R in a terminally redundant form of RNA is also interesting in that it resembles the replicative intermediate or pregenomic RNA of retro- and pararetroviruses. Most striking is the similarity to the plant and vertebrate pararetroviruses of the families Caulimoviridae (Hull et al., 2005) and Hepadnaviridae (Mason et al., 2005). Members of both families employ terminally redundant RNAs as replicative intermediates, but also as mRNAs (Pfeiffer & Hohn, 1983; Loeb et al., 2002). Whilst retroand pararetroviruses use reverse transcription of their pregenomic RNA and require its terminal redundancy for the template switch by reverse transcriptase to occur (Seeger & Mason, 2000; Wilhelm & Wilhelm, 2001), no comparable activity has been described for the nanoviruses. Hence, the precise role of the terminal redundancy of nanovirus m-rep mRNA remains to be determined. 589

I. Grigoras, T. Timchenko and B. Gronenborn

590

Journal of General Virology 89

Terminally redundant nanovirus transcripts

Fig. 4. Secondary-structure predictions of FBNYV m-rep mRNA. Folding predictions of the full-length terminally redundant 1157 nt m-rep mRNA transcribed from nt 105 to 158 of the DNA-R (a) with or (b) without a tail of 50 adenylate residues at its 39 end. (c) Folding prediction of an artificially trimmed transcript lacking the last 54 nt, i.e. a non-terminally redundant m-rep mRNA corresponding to nt 105–104 of the DNA-R, also lacking a poly(A) tail. (d, e) Magnifications of the secondary structure at the origin of replication (ori) of the foldings in (b) and (c), respectively.

In most retro- and pararetroviruses, the generation of a terminally redundant RNA must be regulated, as the same AATAAA poly(A) signal occurs twice on the mRNA, once at the 59 end and once at the 39 end. Several mechanisms have been proposed to account for the bypass of the poly(A) signal by RNA polymerase at the first time of encounter and its recognition at the second time (reviewed by Rothnie et al., 1994). In FBNYV and SCSV, the AATAAA sequence is located in very close proximity to the majority of the mapped 59 ends of the mRNA. This and the fact that plant poly(A) signals have a modular architecture (Loke et al., 2005; Dong et al., 2007; and references therein) may strongly facilitate an initial bypass, as the complete poly(A) signal is not yet present in the 59 part of the transcript. Only upon transcription of the entire DNA-R do the upstream elements of a modular poly(A) signal become physically present in the 39 part of the same RNA. In FBNYV DNA-R, two other AATAAA-like signals are located near the 59 end of the mRNA (Fig. 3a; Supplementary Fig. S1a, available in JGV Online) and are used, but are not always bypassed, as a few short-stop mRNAs were detected by the 39-end mapping experiments. Comparable ‘short-stop’ transcripts were found in cauliflower mosaic virus-infected plants, and the ratio of shortstop RNA to full-length RNA was found to be higher in resistant than in susceptible host plants (Sanfacon & Wieczorek, 1992). By analogy with lentiviruses, for which an RNA secondary structure has been suggested to play a role in poly(A) site selection, the potential stem–loop structure formed by the IR sequences of the nanovirus m-rep mRNA may be a TAR analogue. In human immunodeficiency virus type 1 (Gilmartin et al., 1992) or equine infectious anemia virus (Graveley & Gilmartin, 1996), such an element has been shown to act by juxtaposing the poly(A) site and an upstream enhancer of 39 processing. Indeed, U- or GU-rich sequences characteristic of the far-upstream elements of plant poly(A) signals are found within the 39 part of the m-rep gene and can be brought closer to the AAUAAA sequence by the formation of an extended stem–loop structure including the origin of replication. Stem–loop structures in the 39-UTRs are often found in the transcripts of various other mRNAs of viral or non-viral origin. They have been shown to be involved in translational control or mRNA stability via interactions with RNA-binding proteins (Casey et al., 1989; Ross, 1995; Mazumder et al., 2005). In nanoviruses, M-Rep expression may be regulated via transcript stability, and the secondary http://vir.sgmjournals.org

structure in the 39-UTR might be implicated in such a regulation. In this context, it should be noted that, in geminiviruses, expression of the Rep proteins is regulated negatively by Rep binding to short, direct repeat sequences in its promoter region (Sunter et al., 1993; Eagle et al., 1994). However, no directly equivalent sequence elements have been described for the legume nanoviruses. Yet, also for nanovirus multiplication, a feedback-regulated expression of their master Rep proteins is easily conceivable, and the formation of an extended secondary structure in the 39 part of their encoding mRNAs may serve that purpose. Given that the three-dimensional structures of Rep protein domains of examples of the geminiviruses (tomato yellow leaf curl Sardinia virus), the circoviruses (porcine circovirus type 2) and the nanoviruses (FBNYV) reveal a striking similarity to the RRM domains of RNA-binding proteins (Campos-Olivas et al., 2002; Vega-Rocha et al., 2007a, b), potential binding of the FBNYV M-Rep protein to (folded structures of) its coding mRNA in order to negatively regulate its own expression emerges as an attractive hypothesis.

ACKNOWLEDGEMENTS We thank H.-J. Vetten for critical reading of the manuscript. I. G. is supported by a CNRS-SDV postdoctoral fellowship.

REFERENCES Accotto, G. P., Donson, J. & Mullineaux, P. M. (1989). Mapping of

Digitaria streak virus transcripts reveals different RNA species from the same transcription unit. EMBO J 8, 1033–1039. Altschul, S. F. & Erickson, B. W. (1985). Significance of nucleotide

sequence alignments: a method for random sequence permutation that preserves dinucleotide and codon usage. Mol Biol Evol 2, 526–538. Aronson, M. N., Meyer, A. D., Gyorgyey, J., Katul, L., Vetten, H. J., Gronenborn, B. & Timchenko, T. (2000). Clink, a nanovirus-encoded

protein, binds both pRB and SKP1. J Virol 74, 2967–2972. Beetham, P. R., Hafner, G. J., Harding, R. M. & Dale, J. L. (1997). Two

mRNAs are transcribed from banana bunchy top virus DNA-1. J Gen Virol 78, 229–236. Beetham, P. R., Harding, R. M. & Dale, J. L. (1999). Banana bunchy

top virus DNA-2 to 6 are monocistronic. Arch Virol 144, 89–105. Boevink, P., Chu, P. W. & Keese, P. (1995). Sequence of subterranean

clover stunt virus DNA: affinities with the geminiviruses. Virology 207, 354–361. Brown, T., Mackey, K. & Du, T. (2004). Analysis of RNA by Northern

and slot blot hybridization. In Current Protocols in Molecular Biology, pp. 4.9.1–4.9.19. Edited by F. M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith & K. Struhl. New York: Wiley. 591

I. Grigoras, T. Timchenko and B. Gronenborn

Campos-Olivas, R., Louis, J. M., Clerot, D., Gronenborn, B. & Gronenborn, A. M. (2002). The structure of a replication initiator

biochemical characterization of Rep protein function, a review. Biochimie 77, 765–773.

unites diverse aspects of nucleic acid metabolism. Proc Natl Acad Sci U S A 99, 10310–10315.

Loeb, D. D., Mack, A. A. & Tian, R. (2002). A secondary structure that

Casey, J. L., Koeller, D. M., Ramin, V. C., Klausner, R. D. & Harford, J. B. (1989). Iron regulation of transferrin receptor mRNA levels

requires iron-responsive elements and a rapid turnover determinant in the 39 untranslated region of the mRNA. EMBO J 8, 3693–3699. Chu, P. W., Keese, P., Qiu, B. S., Waterhouse, P. M. & Gerlach, W. L. (1993). Putative full-length clones of the genomic DNA segments of

subterranean clover stunt virus and identification of the segment coding for the viral coat protein. Virus Res 27, 161–171. Clote, P., Ferre, F., Kranakis, E. & Krizanc, D. (2005). Structural RNA

has lower folding energy than random RNA of the same dinucleotide frequency. RNA 11, 578–591. Dekker, E. L., Woolston, C. J., Xue, Y. B., Cox, B. & Mullineaux, P. M. (1991). Transcript mapping reveals different expression strategies for

the bicistronic RNAs of the geminivirus wheat dwarf virus. Nucleic Acids Res 19, 4075–4081.

contains the 59 and 39 splice sites suppresses splicing of duck hepatitis B virus pregenomic RNA. J Virol 76, 10195–10202. Loke, J. C., Stahlberg, E. A., Strenski, D. G., Haas, B. J., Wood, P. C. & Li, Q. Q. (2005). Compilation of mRNA polyadenylation signals in

Arabidopsis revealed a new signal element and potential secondary structures. Plant Physiol 138, 1457–1468. Mason, W. S., Burrell, C. J., Casey, J., Gerlich, W. H., Howard, C. R., Kann, M., Lanford, R., Newbold, J., Schaefer, S. & other authors (2005). Family Hepadnaviridae. In Virus Taxonomy: Eighth Report of

the International Committee on Taxonomy of Viruses, pp. 373–384. Edited by C. M. Fauquet, M. A. Mayo, J. Maniloff, U. Desselberger & L. A. Ball. London: Elsevier/Academic Press. Mazumder, B., Seshadri, V. & Fox, P. L. (2003). Translational control

by the 39-UTR: the ends specify the means. Trends Biochem Sci 28, 91–98. Mazumder, B., Sampath, P. & Fox, P. L. (2005). Regulation of

Dong, H., Deng, Y., Chen, J., Wang, S., Peng, S., Dai, C., Fang, Y., Shao, J., Lou, Y. & Li, D. (2007). An exploration of 39-end processing

macrophage ceruloplasmin gene expression: one paradigm of 39UTR-mediated translational control. Mol Cells 20, 167–172.

signals and their tissue distribution in Oryza sativa. Gene 389, 107–113.

Pfeiffer, P. & Hohn, T. (1983). Involvement of reverse transcription in

Eagle, P. A., Orozco, B. M. & Hanley-Bowdoin, L. (1994). A DNA

the replication of cauliflower mosaic virus: a detailed model and test of some aspects. Cell 33, 781–789.

sequence required for geminivirus replication also mediates transcriptional regulation. Plant Cell 6, 1157–1170.

423–450.

Gebauer, F. & Hentze, M. W. (2004). Molecular mechanisms of

Rothnie, H. M., Chapdelaine, Y. & Hohn, T. (1994). Pararetroviruses

translational control. Nat Rev Mol Cell Biol 5, 827–835.

and retroviruses: a comparative review of viral structure and gene expression strategies. Adv Virus Res 44, 1–67.

Gilmartin, G. M., Fleming, E. S. & Oetjen, J. (1992). Activation of

HIV-1 pre-mRNA 39 processing in vitro requires both an upstream element and TAR. EMBO J 11, 4419–4428. Graveley, B. R. & Gilmartin, G. M. (1996). A common mechanism for

the enhancement of mRNA 39 processing by U3 sequences in two distantly related lentiviruses. J Virol 70, 1612–1617. Herrera-Valencia, V. A., Dugdale, B., Harding, R. M. & Dale, J. L. (2007). Mapping the 59 ends of banana bunchy top virus gene

transcripts. Arch Virol 152, 615–620. Hoekema, A., Hirsch, P. R., Hooykaas, P. J. J. & Schilperoort, R. A. (1983). A binary plant vector strategy based on separation of vir- and

T-region of the Agrobacterium tumefaciens Ti-plasmid. Nature 303, 179–180.

Ross, J. (1995). mRNA stability in mammalian cells. Microbiol Rev 59,

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning:

a Laboratory Manual, 2nd edn. Edited by C. Nolan. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Sanfacon, H. & Wieczorek, A. (1992). Analysis of cauliflower mosaic

virus RNAs in Brassica species showing a range of susceptibility to infection. Virology 190, 30–39. Sano, Y., Wada, M., Hashimoto, Y., Matsumoto, T. & Kojima, M. (1998). Sequences of ten circular ssDNA components associated with

the milk vetch dwarf virus genome. J Gen Virol 79, 3111–3118. Seeger, C. & Mason, W. S. (2000). Hepatitis B virus biology.

Microbiol Mol Biol Rev 64, 51–68.

Horser, C., Harding, R. & Dale, J. (2001). Banana bunchy top

Shirasawa-Seo, N., Sano, Y., Nakamura, S., Murakami, T., Seo, S., Ohashi, Y., Hashimoto, Y. & Matsumoto, T. (2005). Characteristics of

nanovirus DNA-1 encodes the ‘master’ replication initiation protein. J Gen Virol 82, 459–464.

the promoters derived from the single-stranded DNA components of Milk vetch dwarf virus in transgenic tobacco. J Gen Virol 86, 1851–1860.

Hu, J. M., Fu, H. C., Lin, C. H., Su, H. J. & Yeh, H. H. (2007).

Sunter, G., Hartitz, M. D. & Bisaro, D. M. (1993). Tomato golden

Reassortment and concerted evolution in banana bunchy top virus genomes. J Virol 81, 1746–1761.

mosaic virus leftward gene expression: autoregulation of geminivirus replication protein. Virology 195, 275–280.

Hull, R., Geering, A., Harper, G., Lockhart, B. E. & Schoelz, J. E. (2005). Family Caulimoviridae. In Virus Taxonomy: Eighth Report of

Timchenko, T., de Kouchkovsky, F., Katul, L., David, C., Vetten, H. J. & Gronenborn, B. (1999). A single Rep protein initiates replication of

the International Committee on Taxonomy of Viruses, pp. 385–396. Edited by C. M. Fauquet, M. A. Mayo, J. Maniloff, U. Desselberger & L. A. Ball. London: Elsevier/Academic Press.

multiple genome components of faba bean necrotic yellows virus, a single-stranded DNA virus of plants. J Virol 73, 10173–10182.

Katul, L., Maiss, E., Morozov, S. Y. & Vetten, H. J. (1997). Analysis of

six DNA components of the faba bean necrotic yellows virus genome and their structural affinity to related plant virus genomes. Virology 233, 247–259.

Timchenko, T., Katul, L., Sano, Y., de Kouchkovsky, F., Vetten, H. J. & Gronenborn, B. (2000). The master Rep concept in nanovirus

replication: identification of missing genome components and potential for natural genetic reassortment. Virology 274, 189–195.

Khromykh, A. A., Meka, H., Guyatt, K. J. & Westaway, E. G. (2001).

Timchenko, T., Katul, L., Aronson, M., Vega-Arreguin, J. C., Ramirez, B. C., Vetten, H. J. & Gronenborn, B. (2006). Infectivity of nanovirus

Essential role of cyclization sequences in flavivirus RNA replication. J Virol 75, 6719–6728.

DNAs: induction of disease by cloned genome components of Faba bean necrotic yellows virus. J Gen Virol 87, 1735–1743.

Laufs, J., Jupin, I., David, C., Schumacher, S., Heyraud-Nitschke, F. & Gronenborn, B. (1995). Geminivirus replication: genetic and

Vega-Rocha, S., Byeon, I. J., Gronenborn, B., Gronenborn, A. M. & Campos-Olivas, R. (2007a). Solution structure, divalent metal and

592

Journal of General Virology 89

Terminally redundant nanovirus transcripts DNA binding of the endonuclease domain from the replication initiation protein from porcine circovirus 2. J Mol Biol 367, 473–487. Vega-Rocha, S., Gronenborn, B., Gronenborn, A. M. & CamposOlivas, R. (2007b). Solution structure of the endonuclease domain

from the master replication initiator protein of the nanovirus faba bean necrotic yellows virus and comparison with the corresponding geminivirus and circovirus structures. Biochemistry 46, 6201–6212. Vetten, H. J., Chu, P. W. G., Dale, J. L., Harding, R., Hu, J., Katul, L., Kojima, M., Randles, J. W., Sano, Y. & Thomas, J. E. (2005). Family

Nanoviridae. In Virus Taxonomy: Eighth Report of the International Committee on Taxonomy of Viruses, pp. 343–352. Edited by C. M. Fauquet, M. A. Mayo, J. Maniloff, U. Desselberger & L. A. Ball. London: Elsevier/Academic Press.

http://vir.sgmjournals.org

Wanitchakorn, R., Harding, R. M. & Dale, J. L. (1997). Banana bunchy

top virus DNA-3 encodes the viral coat protein. Arch Virol 142, 1673–1680. Wanitchakorn, R., Hafner, G. J., Harding, R. M. & Dale, J. L. (2000).

Functional analysis of proteins encoded by banana bunchy top virus DNA-4 to -6. J Gen Virol 81, 299–306. Wilhelm, M. & Wilhelm, F. X. (2001). Reverse transcription of

retroviruses and LTR retrotransposons. Cell Mol Life Sci 58, 1246–1262. Wu, R.-Y., You, L.-R. & Soong, T.-S. (1994). Nucleotide sequences of

two circular single-stranded DNAs associated with banana bunchy top virus. Phytopathology 84, 952–958. Zuker, M. (2003). Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31, 3406–3415.

593