Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Journal of General Virology (2009), 90, 3057–3065 DOI 10.1099/vir.0.013425-0 Mutational analysis of eggplant latent viroid RNA processing in Chlamydomonas reinhardtii chloroplast Fernando Martı́nez,1 Jorge Marqués,1 Marı́a L. Salvador2 and José-Antonio Daròs1 1 Correspondence Instituto de Biologı́a Molecular y Celular de Plantas (CSIC-UPV), Avenida de los Naranjos, 46022 Valencia, Spain José-Antonio Daròs [email protected] 2 Departamento de Bioquı́mica y Biologı́a Molecular, Universitat de València, Avenida Dr Moliner, 46100 Burjassot, Spain Received 19 May 2009 Accepted 4 August 2009 Viroids of the family Avsunviroidae, such as eggplant latent viroid (ELVd), contain hammerhead ribozymes and replicate in the chloroplasts of the host plant through an RNA-based symmetrical rolling-circle mechanism in which oligomeric RNAs of both polarity are processed to monomeric linear RNAs (by cleavage) and to monomeric circular RNAs (by ligation). Using an experimental system consisting of transplastomic lines of the alga Chlamydomonas reinhardtii, a mutational analysis of sequence and structural elements in the ELVd molecule that are involved in transcript processing in vivo in a chloroplastic context was carried out. A collection of six insertion and three deletion ELVd mutants was created and expressed in C. reinhardtii chloroplast. All mutants cleaved efficiently except for the control with an insertion inside the hammerhead ribozyme domain, supporting the prediction that this domain is necessary and sufficient to mediate transcript cleavage in vivo. However, two deletion mutants that cleaved efficiently showed ligation defects, indicating that during RNA circularization, other parts of the molecule are involved in addition to the hammerhead ribozyme domain. This is probably a quasi double-stranded structure present in the central part of the molecule which contains the ligation site in an internal loop. However, the mutations prevented the viroid from infecting its natural host, eggplant, indicating that they affected other essential functions in ELVd infectious cycle. The insertion in the terminal loop of the right upper hairpin of ELVd did not have this effect; it was tolerated and partially maintained in the progeny. INTRODUCTION Viroids are small circular single-stranded non-coding RNAs (246–401 nt) that infect plants (Flores et al., 2005; Daròs et al., 2006; Ding, 2009). Their infectious cycle depends completely on functional elements in the viroid RNA molecule and interacting host factors. Eggplant latent viroid (ELVd), which produces asymptomatic infections in eggplants, has a branched and highly base-paired secondary structure of minimum free energy (Fig. 1) and exhibits elevated sequence variability in vivo (Fadda et al., 2003a). ELVd belongs to the family Avsunviroidae (Flores et al., 2000) which also contains Avocado sunblotch viroid, Peach latent mosaic viroid (PLMVd) and Chrysanthemum chlorotic mottle viroid. This classification is based on chloroplastic replication, the presence of hammerhead ribozymes and the lack of the central conserved region Two supplementary figures are available with the online version of this paper. 013425 G 2009 SGM characteristic of nuclear viroids (family Pospiviroidae). Viroids replicate through a rolling-circle mechanism with RNA intermediates characterized by the reiterative transcription of a circular RNA template to produce an oligomeric RNA of complementary polarity (Branch & Robertson, 1984). In viroids, positive (+) polarity is arbitrarily assigned to the most abundant strand in vivo and negative (2) polarity to the complementary strand. The viroids of the family Avsunviroidae follow a symmetrical pathway within this general mechanism, in which both oligomeric (+) and (2) RNAs are processed by cleavage to monomeric linear (ml) RNAs and by subsequent ligation to monomeric circular (mc) RNAs. Both mc (+) and (2) RNAs serve as templates to synthesize oligomeric RNAs of complementary polarity (Hutchins et al., 1985; Daròs et al., 1994; Navarro et al., 1999; Bussière et al., 1999). ELVd and the other viroids in its family are highly host specific. They only infect the species in which they were discovered or another closely related species (Flores et al., Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 Printed in Great Britain 3057 F. Martı́nez and others Fig. 1. Library of ELVd insertion and deletion mutants. (a) Scheme of a dimeric ELVd with arrows indicating the positions where the GCGGCCGC octanucleotide was inserted (mutants I1–I6) and double-headed arrows marking the extent of the deletions (mutants D7–D9). (b) Scheme of the cassettes for dimeric ELVd transcription in vitro and in vivo in C. reinhardtii chloroplast. Arrowheads indicate hammerhead ribozyme self-cleavage sites. The T3 RNA polymerase promoter (grey arrow) allows ELVd transcription of DNA linearized with XbaI in vitro. The rbcL promoter (grey arrow) and psaB terminator (grey rectangle) direct ELVd transcription in vivo in the C. reinhardtii chloroplast. White and black rectangles represent ELVd and non-ELVd (cloning vector) sequences, respectively. Numbering refers to ELVd sequence variant with GenBank accession no. AJ536613. Not to scale. (c) ELVd minimum free energy structure (sequence variant AJ536613). The positions of octanucleotide insertions (mutants I1 to I6) and the extent of deletions (mutants D7 to D9) are indicated with straight and bent arrows, respectively. Flags, boxes and arrowheads indicate the domains, conserved nucleotides and self-cleavage sites of both hammerhead ribozymes. The ribozyme of the (+) RNA (represented) and the (”) RNA (complementary to the represented) are indicated by black and white symbols (boxes/arrowheads), respectively. 2000; Fadda et al., 2003a). This makes research on the molecular biology of these pathogens particularity difficult; this would otherwise be very attractive as they are pure RNA pathogens (Daròs et al., 2006) and the only class of pathogen in which it is well established that replication is within the chloroplast (Flores et al., 2005), an organelle of major biotechnological interest (Bock, 2007). Recently, a heterologous experimental system has been developed to study certain aspects of the replication in viroids of the family Avsunviroidae (Molina-Serrano et al., 2007). This system is based on the use of transplastomic lines of the unicellular green alga Chlamydomonas reinhardtii (phylum Chlorophyta) (Harris, 2001) which expresses viroid transcripts in the chloroplast. Although these viroids do not complete their replication cycle in the algal chloroplast, the expressed viroid RNAs are efficiently processed to their corresponding ml and mc forms (Molina-Serrano et al., 2007). In the present work, a library of ELVd insertion and deletion mutants that affect different sequence and structural elements along the molecule was constructed to identify elements in the 3058 ELVd molecule that are involved in transcript processing in vivo. These mutants have been expressed in the C. reinhardtii chloroplast and their processing in vivo has been analysed. The results suggest that ELVd transcript cleavage depends exclusively on the hammerhead ribozyme, but other elements in the molecule take part in ELVd RNA circularization, probably including a central, highly base-paired structure with an internal loop that contains the ligation site. METHODS Construction of an ELVd mutant library. The plasmid pBrmELVd contains the promoter, the 59 untranslated region (59 UTR) and the beginning of the open reading frame (ORF) of C. reinhardtii rbcL gene (from position 270 to +157) followed by a monomeric ELVd cDNA (sequence variant AJ536613, 333 nt, from nt 211 to 210) in pBluescript II KS+ (GenBank accession no. X52327). From this plasmid, six mutants with insertion of the octanucleotide 59GCGGCCGC-39 inserted between ELVd positions 41–42, 71–72, 138–139, 176–177, 245–246 and 305–306 (mutants I1–I6, respectively, Fig. 1) were constructed by PCR by using couples of adjacent and Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 Journal of General Virology 90 ELVd RNA processing in C. reinhardtii chloroplast divergent primers that contained the sequence CCGC at their 59 ends. The amplification (in 50 ml) was carried out with 5 U Tth DNA polymerase (Biotools), 0.2 mM dNTPs, 1 pmol 59-phosphomonoester (59-P) primers ml21 and 100 ng plasmid template for 2 min at 94 uC, 30 cycles of 40 s at 94 uC, 30 s at 55 uC and 4 min at 72 uC, plus a final extension of 10 min at 72 uC. Amplification products were separated by electrophoresis in a 1 % agarose gel in TAE buffer (40 mM Tris/acetic acid, 20 mM sodium acetate, 1 mM EDTA, pH 7.2) and eluted from the gel. Products were treated with 1 U T4 DNA polymerase (Fermentas) for 20 min at 11 uC in the presence of 0.1 mM dNTPs to generate blunt ends, and circularized with T4 DNA ligase (Fermentas). After transformation of Escherichia coli DH5a, selected clones with plasmids containing an additional NotI site in the appropriate position were sequenced. The monomeric mutant ELVd cDNAs were inserted at the appropriate position on a wild-type pBrmELVd to generate the collection of pBrdELVd plasmids with dimeric ELVd cDNAs in which only the viroid central unit contained the mutations (Fig. 1a). Similarly, ELVd mutants with double insertions were created (I2–I4, I4–I6 and I5–I6). From these, clones with sequence deletions between ELVd positions 72–176, 177–305 and 246–305 of the viroid central unit were created by NotI (Fermentas) digestion and religation with T4 DNA ligase (mutants D7, D8 and D9, respectively, Fig. 1). Transcription and self-cleavage in vitro of ELVd mutants. The pBrdELVd plasmids with dimeric ELVd cDNAs (wild-type and mutants) were linearized with XbaI (Fermentas) (Fig. 1b) and 250 ng of each was transcribed in 20 ml with 20 U T3 RNA polymerase (Roche), 0.5 mM NTPs and 20 U RNase inhibitor for 1.5 h at 37 uC. The template DNA was digested with 4 U DNase I (Fermentas) for 15 min at 37 uC and the RNA was separated by denaturing PAGE in a 5 % gel containing 8 M urea and TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) that was stained with ethidium bromide. Transformation of C. reinhardtii chloroplast and analysis of transcripts. Wild-type and mutant dimeric ELVd cDNAs were inserted between the XbaI and EcoRI sites of the vector pCrc+157 (Suay et al., 2005) to transform C. reinhardtii chloroplast by substituting the b-glucuronidase ORF, which is preceded by the promoter, 59 UTR and beginning of the rbcL gene ORF (from –70 to +157), and followed by the 39 UTR of the psaB gene, both from the C. reinhardtii chloroplast (Fig. 1b). Plasmid (5 mg), precipitated on 0.6 mm gold microprojectiles, was used to transform the C. reinhardtii non-photosynthetic strain CC-373 (ac-au-2-21) using the biolistic PDS-1000/He system (Bio-Rad) (Boynton et al., 1988; Blowers et al., 1989). Stable transplastomic clones were selected that recovered photosynthetic ability by growing in HS agar under constant light (Suay et al., 2005). Lines with a high degree of homoplasmicity were selected by PCR analysis. Cells from C. reinhardtii colonies were resuspended and lysed by incubation for 15 min at 65 uC in 100 mM Tris/HCl, pH 8.0, 10 mM EDTA, 0.5 M NaCl, 10 mM 2-mercaptoethanol, 10 % SDS. Chloroplastic DNA was purified from the lysate using silica gel columns (Wizard plus SV minipreps DNA purification system, Promega). The ELVd cDNA was amplified from normalized amounts of DNA (10 ng) using primers P3 (59-TTTAAAGTCCCGACC-39, complementary to ELVd nt 57–71) and P4 (59TTCGGAGGATTCGTC-39, homologous to ELVd nt 72–86) (Supplementary Fig. S1, available in JGV Online) and Tth DNA polymerase (see above). The products were separated by PAGE in a 5 % gel in TAE buffer and stained with ethidium bromide; this was taken as an approximate estimation of the homoplasmicity degree of each clone. Selected transplastomic lines were grown for approximately 1 week to exponential phase in liquid HS media with constant bubbling of 2 % (v/v) CO2 in air at 32 uC with a 12 h (light), 12 h (dark) photoperiod. RNA was purified from 50 ml saturated cultures recovered during the dark period. Cells were lysed in the presence of SDS, total RNA was purified http://vir.sgmjournals.org with a mix of buffer and phenol/chloroform, and the viroid RNA fraction was enriched by non-ionic cellulose CF11 (Whatman) chromatography (Molina-Serrano et al., 2007). Aliquots of the RNA preparations were separated by double PAGE, first in a 5 % gel in TAE buffer that was stained with ethidium bromide, next a segment of the gel between the appropriate size markers was cut (300–400 nt for wild-type ELVd and insertion mutants and 150–300 nt for deletion mutants) and applied on top of a second denaturing gel (5 % polyacrylamide, 8 M urea, 0.256 TBE) to continue separation. RNA from the second gel was electroblotted to a positively charged nylon membrane (Nytran SPC, Whatman), crosslinked with 0.12 J UV light cm22 and hybridized overnight at 70 uC (50 % formamide, 0.1 % Ficoll, 0.1 % polyvinylpyrrolidone, 100 ng salmon sperm DNA ml21, 1 % SDS, 0.75 M NaCl, 75 mM sodium citrate, pH 7.0) with approximately 1 million c.p.m. ELVd (–) 32Plabelled RNA obtained by transcription in vitro in the presence of [a-32P]UTP (400 Ci mmol21). The membrane was washed three times for 10 min with 26 SSC, 0.1 % SDS at room temperature and once for 15 min at 55 uC with 0.16 SSC (150 mM NaCl, 15 mM sodium citrate, pH 7.0), 0.1 % SDS and visualized by autoradiography. Radioactive signals were quantified with an image analyser (Fujifilm FLA-5100). Eggplant inoculation with ELVd mutants and progeny analysis. Wild-type and mutant ELVd ml RNAs (except for mutant I1) produced by transcription and self-cleavage in vitro, were eluted from a denaturing gel and used to inoculate eggplant (Solanum melongena L. cv. Black beauty). For each sample, three plants were inoculated in one leaf with 2 pmol RNA in 10 ml 5 % Carborundum in 50 mM K2HPO4. Inoculated plants were maintained for 1 month at 25 uC with a photoperiod of 16 h (light) and 8 h (dark). Total RNA was purified from systemic leaves of these plants with a mix of buffer and phenol, and enriched in viroid RNAs by non-ionic cellulose CF11 chromatography (Daròs & Flores, 2004). Aliquots of the RNA preparations were separated by denaturing PAGE in a 5 % gel containing 8 M urea and TBE buffer. The RNAs were blotted to a nylon membrane and hybridized as described above. For the infected plants, the viroid progeny was amplified by RT-PCR with two different primer pairs: P1 (59-GTGGCACACACCACCCTATGG-39, complementary to ELVd nt 329–16) and P2 (59-CCCTGATGAGACCGAAAGGTC-39, homologous to ELVd sequence 17–37) or P3 and P4 (Supplementary Fig. S1). The RNA was reverse transcribed with 50 U M-MuLV RT (Revertaid, Fermentas), 10 U RNase inhibitor, 0.1 mM dNTPs and 5 pmol primer P1 or P3 (in 10 ml) for 45 min at 42 uC, 10 min at 50 uC and 5 min at 60 uC. An aliquot (2 ml) of this reaction was used for PCR amplification in 20 ml with 0.4 U Phusion DNA polymerase (Finnzymes), 0.2 mM dNTPs, 0.5 pmol ml21 59-P primers P1 and P2 or P3 and P4 for 30 s at 98 uC, 30 cycles of 10 s at 98 uC, 30 s at 55 uC and 30 s at 72 uC, and a final extension of 10 min at 72 uC. Amplification products were separated by electrophoresis in a 2 % agarose gel in TAE buffer, stained with ethidium bromide and eluted. Next, they were ligated to plasmid pUC18 (GenBank accession no. L08752), linearized with SmaI (Fermentas) and transformed into E. coli. For each sample (wild-type and mutant I5) 10 clones were sequenced, using each pair of primers to amplify five clones. Sequence analyses. ELVd RNA sequences were aligned using the CLUSTAL W program and minimum free energy conformations were calculated using Mfold (Zuker, 2003). RESULTS Construction of a library of ELVd insertion and deletion mutants To identify sequence and structure requirements in the processing (cleavage and ligation) of oligomeric ELVd Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 3059 F. Martı́nez and others RNA in vivo in a chloroplastic context, a library of insertion and deletion mutants of this viroid was constructed (Fig. 1a). In the plasmid pBrdELVd, which contains a dimeric ELVd cDNA (nt 211–210) under the control of the T3 RNA polymerase promoter (Fig. 1b), the octanucleotide GCGGCCGC was introduced, by PCR sitedirected mutagenesis, to six positions in the central viroid unit, which is flanked by two ribozyme self-cleavage sites (Fig. 1a, mutants I1–I6). In the ELVd (+) RNA molecule, these mutations affect a conserved box of the hammerhead ribozyme, an internal loop in the left terminal hairpin, an internal loop in the left upper hairpin, a double-stranded region in the central part of the viroid, the terminal loop of the right upper hairpin and a double-stranded region within the right terminal hairpin (Fig. 1c, mutants I1–I6, respectively). The sequence of the inserted octanucleotide matches the recognition site of NotI and was chosen to facilitate mutant preparation. Moreover, three deletion mutants were also constructed in which the sequences between ELVd positions 72–176 (105 nt), 246–305 (60 nt) and 177–305 (129 nt) were deleted and substituted with the GCGGCCGC octanucleotide (Fig. 1a, mutants D7–D9). The deletions correspond to, approximately, the left third of the viroid molecule, including the two left hairpins; the right fifth of the molecule, including half of the right upper hairpin and the whole terminal right hairpin; and the right third of the molecule, including both right hairpins (Fig. 1c, mutants D7–D9, respectively). As a control experiment, the in vitro self-cleavage ability of the collection of ELVd mutant transcripts was analysed (Supplementary Fig. S2, available in JGV Online). Dimeric ELVd (+) transcript self-cleaves efficiently in vitro through the hammerhead ribozyme (Fadda et al., 2003a) to produce a set of five RNAs which are products of partial and double self-cleavage, and include the 333 nt ml ELVd (+) RNA (Supplementary Fig. S2a). Equal amounts of the plasmids were transcribed in vitro with T3 RNA polymerase; the products of transcription and subsequent in vitro self-cleavage were analysed by denaturing PAGE. In all cases, the five expected selfcleavage products, as well as the uncleaved primary transcript, were detected (Supplementary Fig. S2b). As expected, mutant I1, which contains the insertion between conserved nucleotides of the first hammerhead ribozyme, showed defective processing in the first selfcleavage site. The sizes of the ml RNA products of double self-cleavage for the different mutants were consistent with those expected according to the introduced mutations: 341 nt for insertion mutants I1–I6, and 236, 281 and 212 nt for deletion mutants D7, D8 and D9, respectively (Supplementary Fig. S2b). It was noticeable that the ml ELVd 212 nt deletion mutant migrated faster than the 200 nt RNA employed as a marker in the present work (Supplementary Fig. S2b, compare lanes 1 and 11). This is possibly due to the highly compact structure that this mutant can adopt (see below) that might partly resist denaturation in the urea gel. 3060 Processing of mutant ELVd transcripts in C. reinhardtii chloroplast. The whole set of ELVd mutants was transferred to a plasmid vector that allows transformation of the C. reinhardtii chloroplast genome by homologous recombination generating an expression cassette under the control of the rbcL promoter (Rubisco large subunit) and the psaB terminator (photosystem I subunit B), both from C. reinhardtii (Fig. 1b). Transcription in vivo in the chloroplasts produces a dimeric transcript preceded by a fragment of rbcL mRNA and followed by sequences of the psaB terminator (MolinaSerrano et al., 2007). After transforming chloroplasts, C. reinhardtii clones in which the photosynthetic ability is restored were selected and two or three lines with a high degree of homoplasmicity were chosen for each mutant by detection of viroid cDNA by PCR. Subsequent analysis of independent clones corresponding to the same mutant class showed complete reproducibility with respect to viroid RNA processing. Consequently, the clone expressing more viroid RNA transcript for each mutant class was chosen for comparative analysis. Viroid-enriched RNA was purified by cellulose CF11 chromatography from liquid cultures of the selected clones. ELVd RNAs were analysed by Northern blot hybridization with an ELVd (2) RNA probe after separation by double PAGE, first in a native gel and next in a denaturing one, which allows a better resolution of circular forms than single denaturing PAGE. Hybridization signals corresponding to mc and ml RNA forms were detected for wild-type ELVd and most of the mutants, according to their positions and comparison with controls and RNA markers (Fig. 2a). An exception was the negative control mutant I1, with the insertion in a conserved box of the hammerhead ribozyme, in which no in vivo processing product was detected (Fig. 2a, lane 4). A similar analysis, in which the RNA was separated by single denaturing PAGE, proved the correct expression of this primary mutant transcript in the selected transplastomic C. reinhardtii line (data not shown). For the remaining octanucleotide insertion mutants (I2– I6), both ml and mc RNA forms were detected in a ratio similar to the wild-type control (Fig. 2a, compare lane 3 with lanes 5–9). The mc RNA was always more abundant than the ml RNA, similar to the situation in the natural host eggplant (Fig. 2a, lane 1). A tentative quantification of the ligation ratio [mc/(mc+ml)] based on these hybridization signals confirmed this observation (Fig. 2b). In the deletion mutant D8, the corresponding 281 nt ml and mc RNAs were also detected (Fig. 2a, lane 11); again, the ratio of mc : ml was similar to wild-type (Fig. 2b). In the deletion mutant D7, both 236 nt ml and mc RNA forms were detected (Fig. 2a, lane 10), but in this case, the band corresponding to mc RNA was not significantly more intense than that of ml RNA (Fig. 2a, lane 10 and Fig. 2b). Finally, in the deletion mutant D9, only one band corresponding to the 212 nt ml RNA was detected (Fig. 2a, lane 12). These results indicate that none of the mutations (insertion or deletion) studied in this work, apart from that which directly disturbs the ribozyme, affect ELVd transcript cleavage in vivo in a Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 Journal of General Virology 90 ELVd RNA processing in C. reinhardtii chloroplast Fig. 2. Processing of ELVd insertion and deletion mutants in C. reinhardtii chloroplasts in vivo. (a) Northern blot hybridization of RNA from C. reinhardtii transplastomic clones separated by double PAGE (native and denaturing) with an ELVd (”) RNA probe. Lanes: 1, infected eggplant control; 2, non-transformed C. reinhardtii control; 3, C. reinhardtii expressing a wild-type ELVd transcript in the chloroplast; 4–12, C. reinhardtii expressing ELVd transcripts with insertions I1–I6 and deletions D7–D9, respectively, in the chloroplast. Linear (left) and circular (right) RNA markers are indicated. (b) Plot of the ligation ratio of the processing products of mutant ELVd transcripts in C. reinhardtii chloroplast in vivo. The ligation ratio was estimated from the radioactive signals of the different samples from Northern blot hybridization (a) and calculated as mc/(mc+ml). chloroplastic context. The insertions, as well as deletion D8, also have no effect on ELVd circularization in vivo. However, deletion D7 causes a ligation defect and deletion D9 abolishes it completely. Infectivity of ELVd mutants in eggplant The infectivity of the ELVd mutant collection was also assayed in the natural host eggplant. For this purpose, ml http://vir.sgmjournals.org ELVd RNAs from the wild-type and the different mutants produced by self-cleavage in vitro were eluted from the gel after separation by denaturing PAGE. Mutant I1, which produces minor amounts of its ml RNA in vitro, was excluded from this analysis. Three plants were inoculated with equal molar amounts of the different RNAs and 1 month later, RNA was purified from systemic leaves. Viroid infection was tested by Northern blot hybridization following separation, by denaturing PAGE, of the RNA that had been pooled from each group of three plants. Only wild-type ELVd and mutant I5 (with an insertion in the terminal loop of the right upper hairpin) were able to infect eggplant (Fig. 3a, lanes 3 and 7). When analysed separately, the three plants belonging to each group in both cases were all infected. The progenies from wild-type and mutant I5 ELVd were analysed 1 month post-infection. Viroid cDNAs were amplified by RT-PCR from the RNA preparations with two different couples of adjacent primers: P1 and P2, and P3 and P4. The amplified cDNAs were cloned and 10 different clones from wild-type ELVd and 10 from mutant I5 were sequenced (five obtained with each pair of primers). Five of the 10 clones corresponding to mutant I5 progeny conserved an octanucleotide in the insertion site between positions 245 and 246. However, in all cases, some mutation of the original GCGGCCGC octanucleotide had been produced (Fig. 3b). In the other five clones, the octanucleotide had been simplified to a GC dinucleotide (three clones), a G mononucleotide (one clone) or completely deleted (one clone) (Fig. 3b). All 10 clones exhibited other point mutations along the molecules, similar to the clones corresponding to wild-type ELVd progeny, and consistent with the fact that this viroid accumulates high sequence variability during infection (Supplementary Fig. S1). These results indicate that, apart from insertion I5, all mutations introduced into ELVd impede the biological viability of the viroid. DISCUSSION In this work, the C. reinhardtii-based experimental system was employed in combination with expression of a library of viroid insertion and deletion mutants to study possible sequence and structural requirements in the processing (cleavage and ligation) of ELVd (+) transcripts in vivo in a chloroplastic context. This experimental approach allows structure–function studies in the biochemical and physiological context of the subcellular compartment where the processing of this viroid takes place, which would not otherwise be possible since almost all assayed mutations are lethal (Fig. 3). Moreover, this study benefits from the lack of viroid replication in C. reinhardtii chloroplast, which allows a more direct interpretation of the effect of mutations in transcript processing (Molina-Serrano et al., 2007). In the ELVd molecule, the six insertions (Fig. 1c) distort the hammerhead ribozyme (I1), the left terminal hairpin (I2), the left upper hairpin (I3), the central double- Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 3061 F. Martı́nez and others part of the molecule, splitting the right upper hairpin and completely deleting the right terminal hairpin (D8), and the right part of the molecule, including the whole of both terminal hairpins (D9). Cleavage of ELVd transcripts during rolling-circle replication Fig. 3. Infectivity of ELVd insertion and deletion mutants. (a) RNAs purified from eggplants inoculated with wild-type and mutant ELVd monomeric linear RNAs were analysed by Northern blot hybridization with a complementary RNA probe after separation by denaturing PAGE. Lanes: 1 and 2, ELVd-infected and mockinoculated eggplant controls; 3, eggplant inoculated with wild-type ELVd; 4–11, eggplants inoculated with ELVd insertion mutants I2– I6 and deletion mutants D7–D9, respectively. RNA markers are indicated on the left. ELVd monomeric circular (mc) and monomeric linear (ml ) RNAs are indicated on the right. (b) Evolution of the insertion in mutant I5 1 month after infection of eggplant. Sequence fragment of ELVd mutant I5 showing the GCGGCCGC insertion (grey box) between positions 245 and 246 and the sequence observed in 10 different clones of the progeny. Mutations in the progeny the same as the original octanucleotide sequence are indicated in white on a black background. When the same sequence was found in the progeny more than once, this is indicated (grey text). stranded region (I4), the right upper hairpin (I5) and the right terminal hairpin (I6). These insertions affect doublestranded stretches (I1, I4 and I6) as well as loops (I2, I3 and I5). The three deletions (Fig. 1c) encompass the left part of the molecule, including both left hairpins (D7), the right 3062 ELVd strands of both polarities self-cleave efficiently in vitro through the hammerhead ribozymes (Fadda et al., 2003a), as occurs with the rest of viroids of the family Avsunviroidae (Flores et al., 2000). Evidence exists that hammerhead ribozymes are also functional in vivo; this has been obtained mainly via the characterization of linear RNAs purified from infected plants that are naturally opened at the ribozyme cleavage site (Daròs et al., 1994; Navarro & Flores, 1997) and the analysis of numerous natural sequence variants that tend to maintain the structure and conserved residues of the ribozymes (Rakowski & Symons, 1989; Navarro & Flores, 1997; Ambrós et al., 1998; Fadda et al., 2003a). However, the finding that an RNA-binding protein from avocado chloroplast binds Avocado sunblotch viroid RNA in vivo and facilitates the ribozyme self-cleavage of this viroid transcript in vitro suggests that the activity of hammerhead ribozymes during viroid replication might be helped by host proteins (Daròs & Flores, 2002). All mutants assayed in this work, apart from the negative control which directly affects the hammerhead ribozyme (I1), cleaved as efficiently as wild-type RNA in vivo (Fig. 2a). This suggests that the function of transcript cleavage in vivo resides exclusively in the hammerhead ribozyme domain and that there is no other element in the molecule that participates directly or indirectly (for example by recruiting host auxiliary factors) in the process. Nevertheless, participation of some domain not affected by the mutations produced in this study cannot be disregarded. Ligation of cleavage products of ELVd transcripts during rolling-circle replication Mature ELVd accumulates in the tissues of infected eggplant mainly as circular forms (Fadda et al., 2003a) (Fig. 2a, lane 1); nonetheless, viroid circularization continues to be the least-known step of the viroid replication cycle (Flores et al., 2005). This reaction might be mediated by a host factor such as an RNA ligase or it may be an autocatalytic process inherent to the viroid RNA. The reaction could also be RNA-based, but assisted by a host factor, such as an RNA chaperone, helicase or annealer. Alternatively, the reaction may only occur efficiently in a particular chloroplastic biochemical environment that is difficult to replicate in vitro. Plant tRNA ligase recognizes the 59-hydroxyl (59-OH) and 29,39-cyclic phosphodiester (29,39.P) termini generated by hammerhead ribozymes and efficiently circularizes the ml viroid RNA by producing mc RNA (Kikuchi et al., 1982; Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 Journal of General Virology 90 ELVd RNA processing in C. reinhardtii chloroplast Branch et al., 1982). Moreover, this enzyme has been proposed to localize in the nucleus, cytoplasm and chloroplast of plant cells (Englert et al., 2007). On the other hand, the PLMVd ml RNA product of ribozyme selfcleavage, which contains 59-OH and 29,39.P termini, selfligates in vitro generating a 29,59-phosphodiester bond (Côté & Perreault, 1997). But, although the ml RNAs resulting from ribozyme self-cleavage in most of the other viroids of the family Avsunviroidae also self-ligate in vitro and some PLMVd mc forms closed by a 29,59-phosphodiester bond have been detected in infected tissue (Côte et al., 2001), this reaction is inefficient and does not seem to be responsible for the production of most of the viroid mc RNAs that accumulate in vivo which are closed mainly by a 39,59-phosphodiester bond (D. Molina-Serrano, R. Flores and J.-A. Daròs, unpublished results). It has also been reported that the hammerhead ribozyme can work in reverse by catalysing the ligation of an RNA with 59-OH and 29,39.P termini producing a 39,59-phosphodiester bond (Hertel et al., 1994; Nelson et al., 2005; Canny et al., 2007) in a manner similar to the hairpin ribozyme of plant virus RNA satellites (Fedor, 2000). However, the ELVd ligation product observed in vivo in the host eggplant (Fig. 2a, lane 1) or in C. reinhardtii chloroplast (Fig. 2a, lane 3) is much larger than that produced in vitro, which is negligible (Supplementary Fig. S2). An estimation of the ligation ratio using the hybridization signals from the present work (Fig. 2a) indicates a ratio of about 0.8 in infected eggplant and about 0.9 in C. reinhardtii chloroplast (Fig. 2b), suggesting that a very efficient mechanism operates in vivo, most probably mediated by a chloroplastic RNA ligase or, alternatively, by some host auxiliary factor or some chloroplast biochemical context that potentiates ELVd RNA self-ligation. Moreover, the effect of sequence deletions on ELVd ml RNA ligation in the C. reinhardtii chloroplast observed here involves other sequences, in addition to the hammerhead ribozyme domain, in the ligation function. This is shown by deletion mutant D7, which affects the left part of the molecule, that had a ligation defect (Fig. 2a, lane 10 and Fig. 2b) and deletion mutant D9, which affects the right part of the molecule, that did not ligate at all (Fig. 2a, lane 12). The ligation site in the ELVd (+) minimum free energy conformation is located in an internal loop situated in a quasi double-stranded structure in the centre of the molecule (Fig. 1c and Fig. 4). This structure results from base pairing of the sequences corresponding to both (+) and (2) hammerhead ribozyme domains. The degree of distortion of this central structure produced in the different deletion mutants correlates with the ligation defect observed in vivo (Fig. 2a, lanes 10–12, Figs 2b and 4). While deletion D8, in which the stability of this structure is not affected (Fig. 4), did not influence ligation (Fig. 2a, lane 11), ligation efficiency was diminished in deletion mutant D7 (Fig. 2a, lane 10 and Fig. 2b) in which this central structure is partially affected, without directly affecting the sequence of the loop, the ligation site or the http://vir.sgmjournals.org Fig. 4. Schematic representation of the minimum free energy conformation of wild-type ELVd (sequence variant AJ536613) and the deletion mutants D7, D8 and D9. The position of the viroid processing site is indicated with an arrowhead. The parts of the molecules that are distorted as a consequence of the deletions are boxed in grey. adjacent sequences (Fig. 4). Finally, in deletion mutant D9, in which the upper strand of the quasi double-stranded structure is affected, including the upper part of the loop with the ligation site (Fig. 4), there is no ligation in vivo (Fig. 2a, lane 12). In short, these results suggest that, in addition to the hammerhead ribozyme domain, other ELVd sequences or structural elements are essential for ligation in vivo. It is possible that the correct folding of ELVd ml molecule in a ligation conformation is necessary, including a quasi double-stranded structure in the central part of the molecule that contains the ligation site in a loop. Furthermore, this implies that the motifs in the viroid molecule involved in cleavage and ligation are different. Biological viability of an ELVd mutant with an inserted octanucleotide in the terminal loop of the right upper hairpin In the tissues of viroid-infected plants, numerous sequence variants with point mutations and small deletions or insertions have been reported, in the case of both chloroplast viroids of the family Avsunviroidae and nuclear viroids of the family Pospiviroidae (Flores et al., 2005). Despite this, viroids are very compact molecules and biologically viable mutants with large insertion or deletions have not been described frequently. Exceptions are the sequence variants within block duplications of parts of the rod-like structure in coconut cadang-cadang viroid or citrus exocortis viroid (Haseloff et al., 1982; Semancik et al., 1994; Fadda et al., 2003b) or the PLMVd sequence variants with 12–13 nt insertions, some of which induce the peach calico symptom (Malfitano et al., 2003; Rodio et al., 2006, 2007). It is notable that ELVd insertion mutant I5, with an insertion in the terminal loop of the right upper hairpin, is infectious (Fig. 3a, lane 7) and that this insertion is partly conserved during a 1 month infection process (Fig. 3b). Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 3063 F. Martı́nez and others This insertion resembles, according to size and position, those described in PLMVd variants which also fold into hairpins and are located in the loop capping a long viroid arm (Malfitano et al., 2003). The observation that ELVd tolerates an 8 nt insertion in this particular position of the molecule must contribute to understanding of the structural requirements of this viroid during its infectious cycle. Moreover, this position will be suitable for the introduction of sequence tags in future experiments to study the molecular biology of this viroid and, in particular, the host factors interacting during the infectious cycle. Côte, F., Lévesque, D. & Perreault, J. P. (2001). Natural 29,59phosphodiester bonds found at the ligation sites of peach latent mosaic viroid. J Virol 75, 19–25. Daròs, J. A. & Flores, R. (2002). A chloroplast protein binds a viroid RNA in vivo and facilitates its hammerhead-mediated self-cleavage. EMBO J 21, 749–759. Daròs, J. A. & Flores, R. (2004). Arabidopsis thaliana has the enzymatic machinery for replicating representative viroid species of the family Pospiviroidae. Proc Natl Acad Sci U S A 101, 6792–6797. Daròs, J. A., Marcos, J. F., Hernández, C. & Flores, R. (1994). Replication of avocado sunblotch viroid: evidence for a symmetric pathway with two rolling circles and hammerhead ribozyme processing. Proc Natl Acad Sci U S A 91, 12813–12817. Daròs, J. A., Elena, S. F. & Flores, R. (2006). Viroids: an Ariadne’s thread into the RNA labyrinth. EMBO Rep 7, 593–598. ACKNOWLEDGEMENTS This work was supported by grants AGL2004-06311-C02-01 (to J.-A. D.) and BFU2006-07783 (to M. L. S.) from Ministerio de Educación y Ciencia (MEC) from Spain, and grant BIO2008-01986 (to J.-A. D.) from Ministerio de Ciencia e Innovación from Spain. F. M. and J. M. were recipients of masters and predoctoral fellowships, respectively, from MEC. F. M. was a recipient of a predoctoral fellowship from Universidad Politécnica de Valencia from Spain. We thank Drs Ricardo Flores, Gustavo Gómez and Douglas Grubb for critical reading of the manuscript. Ding, B. (2009). The biology of viroid-host interactions. Annu Rev Phytopathol 47, 105–131. Englert, M., Latz, A., Becker, D., Gimple, O., Beier, H. & Akama, K. (2007). Plant pre-tRNA splicing enzymes are targeted to multiple cellular compartments. Biochimie 89, 1351–1365. Fadda, Z., Daròs, J. A., Fagoaga, C., Flores, R. & Duran-Vila, N. (2003a). Eggplant latent viroid, the candidate type species for a new genus within the family Avsunviroidae (hammerhead viroids). J Virol 77, 6528–6532. Fadda, Z., Daròs, J. A., Flores, R. & Duran-Vila, N. (2003b). REFERENCES Identification in eggplant of a variant of citrus exocortis viroid (CEVd) with a 96 nucleotide duplication in the right terminal region of the rod-like secondary structure. Virus Res 97, 145–149. Ambrós, S., Hernández, C., Desvignes, J. C. & Flores, R. (1998). Fedor, M. J. (2000). Structure and function of the hairpin ribozyme. J Mol Biol 297, 269–291. Genomic structure of three phenotypically different isolates of peach latent mosaic viroid: implications of the existence of constraints limiting the heterogeneity of viroid quasispecies. J Virol 72, 7397– 7406. Blowers, A. D., Bogorad, L., Shark, K. B. & Sanford, J. C. (1989). Studies on Chlamydomonas chloroplast transformation: foreign DNA can be stably maintained in the chromosome. Plant Cell 1, 123–132. Bock, R. (2007). Plastid biotechnology: prospects for herbicide and insect resistance, metabolic engineering and molecular farming. Curr Opin Biotechnol 18, 100–106. Boynton, J. E., Gillham, N. W., Harris, E. H., Hosler, J. P., Johnson, A. M., Jones, A. R., Randolph-Anderson, B. L., Robertson, D., Klein, T. M. & other authors (1988). Chloroplast transformation in Chlamydomonas with high-velocity microprojectiles. Science 240, 1534–1538. Flores, R., Daròs, J. A. & Hernández, C. (2000). The Avsunviroidae family: viroids containing hammerhead ribozymes. Adv Virus Res 55, 271–323. Flores, R., Hernández, C., Martı́nez de Alba, A. E., Daròs, J. A. & Di Serio, F. (2005). Viroids and viroid-host interactions. Annu Rev Phytopathol 43, 117–139. Harris, E. H. (2001). Chlamydomonas as a model organism. Annu Rev Plant Physiol Plant Mol Biol 52, 363–406. Haseloff, J., Mohamed, N. A. & Symons, R. H. (1982). Viroid RNAs of cadang-cadang disease doi:10.1038/299316a0. of coconuts. Nature 299, 316–321. Hertel, K. J., Herschlag, D. & Uhlenbeck, O. C. (1994). A kinetic and thermodynamic framework for the hammerhead ribozyme reaction. Biochemistry 33, 3374–3385. Branch, A. D. & Robertson, H. D. (1984). A replication cycle for Hutchins, C. J., Keese, P., Visvader, J. E., Rathjen, P. D., McInnes, J. L. & Symons, R. H. (1985). Comparison of multimeric plus and minus viroids and other small infectious RNAs. Science 223, 450–455. forms of viroids and virusoids. Plant Mol Biol 4, 293–304. Branch, A. D., Robertson, H. D., Greer, C., Gegenheimer, P., Peebles, C. & Abelson, J. (1982). Cell-free circularization of viroid Kikuchi, Y., Tyc, K., Filipowicz, W., Sänger, H. L. & Gross, H. J. (1982). progeny RNA by an RNA ligase from wheat-germ. Science 217, 1147– 1149. Circularization of linear viroid RNA via 29-phosphomonoester, 39,59phosphodiester bonds by a novel type of RNA ligase from wheat-germ and Chlamydomonas. Nucleic Acids Res 10, 7521–7529. Bussière, F., Lehoux, J., Thompson, D. A., Skrzeczkowski, L. J. & Perreault, J. P. (1999). Subcellular localization and rolling circle Malfitano, M., Di Serio, F., Covelli, L., Ragozzino, A., Hernández, C. & Flores, R. (2003). Peach latent mosaic viroid variants inducing peach replication of peach latent mosaic viroid: hallmarks of group A viroids. J Virol 73, 6353–6360. calico (extreme chlorosis) contain a characteristic insertion that is responsible for this symptomatology. Virology 313, 492–501. Canny, M. D., Jucker, F. M. & Pardi, A. (2007). Efficient ligation of the Molina-Serrano, D., Suay, L., Salvador, M. L., Flores, R. & Daròs, J. A. (2007). Processing of RNAs of the family Avsunviroidae in Schistosoma hammerhead ribozyme. Biochemistry 46, 3826–3834. Côté, F. & Perreault, J. P. (1997). Peach latent mosaic viroid is locked Chlamydomonas reinhardtii chloroplasts. J Virol 81, 4363–4366. by a 29,59-phosphodiester bond produced by in vitro self-ligation. J Mol Biol 273, 533–543. Navarro, B. & Flores, R. (1997). Chrysanthemum chlorotic mottle 3064 viroid: unusual structural properties of a subgroup of self-cleaving Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 Journal of General Virology 90 ELVd RNA processing in C. reinhardtii chloroplast viroids with hammerhead ribozymes. Proc Natl Acad Sci U S A 94, 11262–11267. in infected plants and requirements of the insertion containing the pathogenicity determinant. J Gen Virol 87, 231–240. Navarro, J. A., Daròs, J. A. & Flores, R. (1999). Complexes containing Rodio, M. E., Delgado, S., De Stradis, A., Gómez, M. D., Flores, R. & Di Serio, F. (2007). A viroid RNA with a specific structural motif both polarity strands of avocado sunblotch viroid: identification in chloroplasts and characterization. Virology 253, 77–85. Nelson, J. A., Shepotinovskaya, I. & Uhlenbeck, O. C. (2005). Hammerheads derived from sTRSV show enhanced cleavage and ligation rate constants. Biochemistry 44, 14577–14585. inhibits chloroplast development. Plant Cell 19, 3610–3626. Semancik, J. S., Szychowski, J. A., Rakowski, A. G. & Symons, R. H. (1994). A stable 463-nucleotide variant of citrus exocortis viroid produced by terminal repeats. J Gen Virol 75, 727–732. Rakowski, A. G. & Symons, R. H. (1989). Comparative sequence Suay, L., Salvador, M. L., Abesha, E. & Klein, U. (2005). Specific roles studies of variants of avocado sunblotch viroid. Virology 173, 352– 356. of 59 RNA secondary structures in stabilizing transcripts in chloroplasts. Nucleic Acids Res 33, 4754–4761. Rodio, M. E., Delgado, S., Flores, R. & Di Serio, F. (2006). Variants of Zuker, M. (2003). Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31, 3406–3415. Peach latent mosaic viroid inducing peach calico: uneven distribution http://vir.sgmjournals.org Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Thu, 03 Aug 2017 21:37:32 3065