Replication of Avocado Sunblotch Viroid in the Cyanobacterium Nostoc Sp. PCC 7120

1Aix Marseille University, CNRS, UMR Chemistry Laboratory Bacterial 7283. 31 Chemin Joseph Aiguier 13009 Marseille Cedex 20, France 2Institute of Systematics, Evolution, Biodiversity ( ISyEB ), CNRS, MNHN , UPMC, EPHE, UPMC Sorbonne University, 57 rue Cuvier, PO Box 50, F-75005 Paris, France 3Team Enzymology of RNA, UR6, UPMC Sorbonne University, 75252 Paris Cedex 05 France 4Laboratory of Plant Biotechnology, Faculty of Sciences of Sfax, University of Sfax, BP 1171, 3000 Sfax, Tunisia


Introduction
Viroids are the smallest pathogens known to exist on Earth. They were found in the early 70s to be the causal agents responsible for the potato tubercle disease, which had been blamed so far on bacteria or viruses [1]. This pioneering study led to the discovery of a highly intriguing non coding RNA, which is able to infect a large panel of crop plants [2,3]. Unlike viruses, viroids are composed of free RNA without any envelope or capsid. They can occur in infected plant cells in two forms: in the positive polarity form, which is the most abundant, and the minus polarity one, which corresponds to a replication intermediate. They are circular, structured RNAs consisting of about 250 to 400 nucleotides that replicate via an RNA-RNA copying process according to a "rolling-circle" mechanism [4]. Two families of viroids are known to exist so far, the Pospiviroidae, which include five genera (pospiviroids, hostuviroids, cocadviroids, apscaviroids and coleviroids) and the Avsunviroidae, which include three genera (avsunviroids, pelamoviroids and elaviroids). While the viroids belonging to the Pospiviroidae family replicate in the cell nucleus, replication of the members of the Avsunviroidae takes place in the chloroplasts of infected plants via a process involving a hammerhead ribozyme (HHR) on which the processing of the transcription products depends. During the replication of Avsunviroidae, oligomers are formed and cleaved to monomers, which are then ligated and generate the circular forms. Replication of viroids requires the presence of host factors. The DNAdependent RNA polymerase II has been found to replicate members of the Pospiviroidae family [5], while replication of members of the Avsunviroidae family is thought to depend on the nuclear-encoded DNA-dependent RNA polymerase present in the chloroplasts [6].
Several environmental factors are known to affect the replication of viroids in plants. Heat treatment has been found to increase the accumulation of mutations during the replication cycle of the Hop latent viroid, thus increasing the variability of these infectious RNAs [7]. The activity of the HHR of the Chrysanthemum chlorotic mottle viroid has been found to resist drastic environmental conditions such as high pressure [8]. In addition, self-cleavage-catalyzed reactions have been reported to occur in the HHR of the Avocado sunblotch viroid (ASBVd) exposed to high temperatures (of up to 60°C) [9]. The fact that potato spindle tuber viroid has been detected in potato seeds stored for 21 years shows that this molecule is endowed with remarkable longevity [10].
The structure of viroids, the catalytic activity of some of them, and their long-term persistence have been presented as arguments supporting the idea that they were involved at the beginning of the so-called "RNA world" [11]. We recently established that ASBVd replicates in the yeast Saccharomyces cerevisiae, which indicates that the ribozyme is able to sustain the cleavage/ligation reaction during the rolling circle replication process. Since ASBVd replication has continued for 25 generations of the yeast, the viroid replication process can be said to be surprisingly adaptable in this unconventional host [12].
Are Avsunviroidae able to replicate in a prokaryotic cell? The aim of the present study was to answer this question. It was therefore proposed to analyze the replication of ASBVd, a typical member of the Avsunviroidae family, in the cyanobacterium Nostoc PCC 7120. Since ASBVd replicates in chloroplasts, we predicted that cyanobacteria,

Abstract
Viroids are small infectious RNA molecules that replicate in plants via RNA-RNA replication processes. The molecular mechanism responsible for this replication has attracted great interest, and studies on this topic have yielded interesting biological findings on the processes in which RNA is involved. Viroids belonging to the Avsunviroidae family replicate in the chloroplasts of infected hosts. It has by now been established that chloroplasts and cyanobacteria share a common have ancestor. In view of this phylogenetic relationship, we investigated whether a member of the Avsunviroidae family could be replicated in a cyanobacterium. The results obtained here show that Avocado Sunblotch Viroid (ASBVd) RNA is able to replicate in the filamentous cyanobacterium Nostoc PCC 7120. Indeed, Northern blot hybridization showed that linear forms of "minus" polarity were detected in RNA extracted from Nostoc cells expressing ASBVd dimers of positive "polarity", and that linear forms of "negative" polarity were detected in RNA extracted from Nostoc cells expressing ASBVd dimers of "positive" polarity. ASBVd replication does not impair the growth of Nostoc. These results provide the first evidence that a prokaryotic cell possesses all the machinery required to sustain the process of RNA-RNA replication. The data obtained here are of great importance, since they might shed light on the evolution of the cellular factors on which RNA replication processes depend.
which are share their progenitors with chloroplasts, would be excellent candidates for this study.
Cyanobacteria are among the most diverse and widely distributed phyla of bacteria. In photosynthetic prokaryotes, they have the unique ability to perform oxygenic photosynthesis. They play key roles in the carbon and nitrogen cycles and constitute highly suitable organisms for investigating the evolution of photosynthetic eukaryotes, since the ancestry of the chloroplasts of all photosynthetic organisms can be traced back to cyanobacteria [13]. The authors of taxonomic studies have classified the cyanobacteria in five subsections based on their level of morphological complexity. Organisms in subsections IV (Nostocales) and V (Stigonematales) are diazotrophic bacteria that can differentiate specific cells called heterocysts, which are involved in nitrogen fixation processes [14]. Nostoc PCC 7120 (which will be referred to below as Nostoc) belongs to the Nostocales subsection. This strain was used here as an experimental model to determine whether ASBVd can replicate in cyanobacteria. This is the first time a viroid RNA replication process has been reported to occur in a prokaryote.

Strains and growth conditions
Nostoc sp. PCC 7120 (Pasteur Institute collection, France) was grown in BG11 medium at a temperature of 30°C in air under continuous illumination (40 µE m -2 s -1 ). Cyanobacterial growth was monitored by measuring the absorbance at 750 nm (OD 750 ). Each experiment was performed in triplicate. Nostoc filaments were observed using the Optical microscope Nikon Eclipse E800.

Oligonucleotides
The oligonucleotides used in this study were obtained from Eurogentec (France). Their sequences are listed in Table 1.

Construction of Nostoc/ASBVd recombinant strains
The ASBVd dimer was extracted from the pCRII-TOPOdASBVd plasmid expressing the ASBVd DNA dimer [12], using the EcoRI enzyme. The resulting fragment was cloned under the control of the petE promoter of the pSKpetE plasmid [15]. Plasmids expressing dimeric (+) or (-) ASBVd were selected after the sequencing procedure (Millegen, France). The petEASBVd fragments were then subcloned into the pRL25 plasmid [15] after performing a linearization step using the NotI and KpnI enzymes. The resulting plasmids were sequenced in order to further confirm the (+) or (-) orientation of the viroid. The recombinant pRLASBVd(+) and pRLASBVd(-) obtained were conjugated into Nostoc, and exoconjugants were selected with 50 µg/ mL neomycin.

RNA extraction
RNA was extracted using TRizol reagent (Invitrogen, France) in line with the manufacturer's instructions. Chromosomal DNA was removed by treating RNA preparations (50 µL) with 1 µL of DNase (Ambion at 2 U/µL) for 1 hour at 37°C. DNase treatment was checked by performing RT-PCR (see below), omitting the reverse transcription step. Only DNA-free RNAs were used in all our experiments. The concentration of the RNA was determined spectrophotometrically.

Strand-specific RT-PCR
One µg of total RNA was subjected to reverse transcription with ThermoScript reverse transcriptase (Invitrogen) using 2 pmol of LKASBV-O1 and LKASBV-O2 as primers to synthesize (+) and (-) cDNAs respectively. One-tenth of these reactions were used as templates for PCR. Amplifications were performed using the high fidelity Taq polymerase (Jena Biosciences, Germany) in line with the manufacturer's instructions. ASBV-O1 and ASBV-O2 primers (10 pmol) were used in each reaction. The standard program was: 5 min at 94°C, followed by 35 cycles of 40 s at 94°C, 45 s at 50°C and 45 s at 72°C, followed by a final 5 min at 72°C. The products (247-nt corresponding to the ASBVd monomers) were separated on a 1.2% agarose gel.
RNA marker (Thermofisher, France) co-migrated with total RNAs samples on the 6% polyacrylamide-urea gels and stained with ethidium bromide before northern blotting.  strains bearing either the empty pRL25 plasmid or pRLASBVd(+) or pRLASBVd(-) were grown up to the late exponential phase (OD 750 : 1). RNAs were extracted and the replication of ASBVd strands was tested using the strand-specific RT-PCR assay, as described in Figure 1. Briefly, the reverse transcription reaction induced using the LKASBV-O1 primer results in the synthesis of the (+) cDNA and that of the LKASBV-O2 primer results in the synthesis of the (-) cDNA. An LK sequence was added to the 5' end of each of these reverse-strand primers. The LK sequence alone was used with the strand-specific primer for the PCR (Figure 1 and Table 1). This procedure ensured that only DNA specifically synthesized during the RT step could be amplified during the PCR [17], thus making it possible after the PCR reaction to detect the RNA-RNA replication of the ASBVd (-) strand. The amplification of the rnpB gene encoding the RNase P subunit B was used to check the amount of RNA used in each assay ( Figure  2B, Lanes 1 and 2). The data presented in Figure 2A

Replication of negative and positive ASBVd strands in Nostoc assessed by Northern blot technique
The replication of the ASBVd in Nostoc cells was further investigated using a Northern blot approach. For this purpose, total RNAs were extracted from Nostoc cells bearing the pRLASBVd(+) or pRLASBVd(-) plasmids, loaded onto a 6% polyacrylamide gel, and transferred to nitrocellulose membranes. The membranes were probed with riboprobes resulting from in vitro transcription of the (+) or (-) ASBVd dimer RNAs extracted from Nostoc harboring the empty pRL25 plasmid were used to run negative controls. In the positive controls, we used total RNA extracted from avocado tree leaves infected with ASBVd. The level of rnpB transcripts served to monitor the loading control.
When RNAs extracted from Nostoc/pRLASBVd (-) were used as a template, linear form were observed when the negative strand was used as a riboprobe ( Figure 3A, Lane 5). This confirms that the (-) strands were actually replicated and yielded the complementary (+) strands. Similar results were obtained when the RNAs extracted from Nostoc/ pRLASBVd (+) were loaded and when the positive strand was used as a riboprobe during Northern hybridization assays ( Figure 3B, Lane 8), although the amount of negative strand was much lower than that of positive strand. In both cases, the form detected had the size expected for an ASBVd linear form. This could be assessed thanks to the use of an RNA molecular marker (Figure 3), and positive controls consisting in a linear ASBVd from synthesized in vitro and also in RNA extracted from avocado tree leaves infected with ASBVd ( Figure 3). This and ASBVd(+) DNA dimmers, respectively, from the petE promoter (grey boxes). HH: hammerhead ribozyme. During the first step, self-cleavage of the ASBVd dimeric form results in the linear monomeric form (lmASBVd) and the circular monomeric form (cmASBVd). During the second step, the RNA-dependent replication process generates linear oligomers (loASBVd), lmASBVd and cmASBVd having the opposite polarity. The primers used during the reverse transcription have a linker sequence at their 5' extremity (LKASBV-O1 and LKASBV-O2). In order to prevent any amplification of DNA from the plasmid, the primers used during the PCR step are LK and ASBV-O2 for the replication of ASBVd(-) and LK and ASBV-O1 for ASBVd(+), respectively, which makes it possible to discriminate between the RNAs resulting from plasmid transcription and those resulting from RNA-RNA replication.   and (+) ASBVd forms in Nostoc using the Northern blot procedure. RNAs were extracted from a Nostoc strain harboring either the empty pRL25 plasmid or pRLASBVd(+) or pRLASBVd(-). They were separated on 6% polyacrylamide-urea gels and analyzed by hybridization using riboprobes corresponding to the ASBVd strands of (-) or (+) polarities. The hybridization using a probe corresponding to the rnpB gene served as RNA loading control. Lane 1 corresponds to synthetic linear in vitro transcript of ASBVd. Molecular markers (in nucleotide) co-migrated on the 6% polyacrylamide-urea gels are stained by Ethidium bromide before autoradiography. The reference (lanes 2, 6, 10) refers to the total RNA extract from natural ASBVd infected plant. Lane 3 and 7: no product was detected by riboprobe (-) when RNAs extracted from Nostoc harboring the pRL25 empty plasmid were loaded. Lane 4: linear transcript form detected by riboprobe (-) when RNAs extracted from Nostoc harboring the pRLASBVd(+) plasmid were loaded. Lane 5: linear replicative form detected by riboprobe (-) when RNAs extracted from Nostoc/pRLASBVd(-) were loaded. Lane 8: linear replicative form detected by riboprobe (+) with RNAs extracted from Nostoc/pRLASBVd(+). Lane 9: linear transcript form to be detected by riboprobe (+) with RNAs extracted from Nostoc/pRLASBVd(-) was above the detection limit. indicates that the cellular machinery of Nostoc is able to replicate the negative and negative ASBVd forms; even though the replication and/ or stability of the strand of the minus seems more efficient than that of the positive form. Interestingly, "plus" strands were found to be present in significantly higher amounts in infected plant cells than "minus" strands, which suggests that a difference in replication efficiency and/ or stability may exist between the two forms of ASBVd having different polarities [18,19].

Impact of the replication of ASBVd on Nostoc physiology
The replication of ASBVd in Nostoc did not impair the growth of the strain. The generation time of the ASBVd-expressing strains was similar to that of the wild type strain ( Table 2). The filaments observed in the wild type strain and the ASBVd-expressing strains were similar ( Figure 4). It was therefore concluded that replication of this viroid in Nostoc did not have any significant physiological effects on the bacterium, at least under the experimental conditions tested here.

Discussion
Only eukaryotic cells have been found so far to be able to copy viroids via an RNA-RNA replication mechanism. The replication of the ASBVd occurring in the yeast S. cerevisiae was an important finding, since it showed that replication of these RNA molecules is not strictly limited to plants and can occur in organisms devoid of chloroplasts [12]. The present study is a further step in this direction, since it shows that a prokaryotic possesses all the machinery required to copy ASBVd; both the (+) and (-) strands were copied in Nostoc recombinant cells bearing the dimeric replicative DNA of the ASBVd (Figures 2 and 3).
In a previous study [12], it was reported that ASBVd transcripts with "plus" and "minus" polarities showed differences in electrophoretic mobility under native conditions and in their thermal denaturation profiles. The models obtained by performing RNA-selective 2'-hydroxyl acylation were analyzed using the primer extension (SHAPE) method [20], and the results showed that the two polarities fold into different structures. Structural differences in the regions neighboring the initiation sites of both ASBVd (−) and (+) may influence their respective interactions with the RNA polymerase and/or transcription factors, which might account for the difference observed in the levels of accumulation observed here between the negative and positive forms ( Figure 3). In the Avsunviroidae family members, the "minus" circular forms have been found to accumulate less extensively than the (+) circular forms at the beginning of the replication process, probably because this form (-) is an intermediate replication form which cannot accumulate in high levels, contrary to the (+) circular form [18]. It was also recently established that in the other viroid family (Pospiviroidae), plant cell viroids are subject to decay [19].
The results obtained in this study on the replication of a viroid belonging to the Avsunviroidae family in a cyanobacterium were also highly informative from the evolutionary point of view. The nuclearencoded DNA-dependent RNA polymerase (NEP) was previously found to fulfill this function [6]. It has been also reported that the RNA polymerase of the enterobacterium Escherichia coli was able to replicate the Peach latent mosaic viroid, which belongs to the same family as ASBVd [20]. The ability of the Nostoc RNA polymerase to replicate ASBVd further supports the idea that bacterial polymerases are able to replicate viroids. The second replication step, which involves the cleavage of the oligomeric RNA intermediates, must be achieved by the HHR, as extensively described in the case of the viroids belonging to the Avsunviroidae family in their natural hosts [21][22][23]. The hammerheadmediated self-cleavage of the multimeric ASBVd transcripts has been found to be facilitated by the chloroplastic RNA chaperones PARBP33 and PARBP35 [24]. These proteins are RNA-binding proteins involved in the editing and stabilization of transcripts [25,26]. A survey of the Nostoc genome has shown the presence of six open reading frames (alr0741, alr2311, alr2087, all2928, alr4683, all277) with approximately 48% amino acid identity with PARBP33. The question as to whether one or several of these proteins interact with the ASBVd transcripts in Nostoc and assist the ribozyme in its cleaving function would be well worth investigating.
The self-cleavage of the oligomeric RNAs is followed by the circularization of the resulting monomeric linear viroid RNAs with (+) and (-) polarities, both harboring 5′-hydroxyl and 2′,3′-cyclic phosphodiester termini [27]. This step has been reported to be performed by the chloroplastic isoform of the plant tRNA ligase, which is involved in pre-tRNA splicing and naturally recognizes 5′-hydroxyl and 2′,3′-cyclic phosphodiester termini [28]. A Blast search using several plant tRNA ligases as seeds did not result in the detection of any significant homologues in the Nostoc genome or in any of the 127 cyanobacterial genomes sequenced so far [14]. It is therefore possible that RNA ligases showing no homology with plants or other eukaryotic enzymes may exist in cyanobacteria. This would explain the circularization of the RNA occurring during the replication of the ASBVd in Nostoc. Discovering these putative enzymes and understanding their natural function is an exciting perspective.
The replication of ASBVd in Nostoc had no physiological effects on this bacterium (Table 2 and Figure 4). Replication of this viroid in S. cerevisiae has also been reported to be asymptomatic [12]. All in all, these results raise the question as to whether other RNAs may exist in the living kingdoms, where they may replicate via an RNA-RNA process. In this scenario, the viroids could be said to be the submerged part of the iceberg; if they were avirulent against their host plant, as occurs in the case of Eggplant latent viroid [29], their replication would have never been discovered. Do many non-infectious RNAs therefore exist and replicate in living organisms? What might their origin be? Answering these questions constitutes an important challenge for improving our knowledge of the non-coding genome. Generation time (hours) WT 25 ± 3 pRLASBVd(+) 25 ± 5 pRLASBVd(-) 26 ± 4