A viroid-derived system to produce large amounts of recombinant RNA in Escherichia coli

Viruses have been engineered into useful biotechnological tools for gene therapy or to induce the synthesis of products of interest, such as therapeutic proteins and vaccines, in animal and fungal cells, bacteria or plants. Viroids are a particular class of infectious agents of higher plants that exclusively consist of a small non-protein-coding circular RNA molecule. In the same way as viruses have been transformed into useful biotechnological devices, can viroids be converted into beneficial tools? We show herein that, by expressing Eggplant latent viroid (ELVd) derived RNAs in Escherichia coli together with the eggplant tRNA ligase, this being the enzyme involved in viroid circularization in the infected plant, RNAs of interest like aptamers, extended hairpins, or other structured RNAs are produced in amounts of tens of milligrams per liter of culture. Although ELVd fails to replicate in E. coli, ELVd precursors self-cleave through the embedded hammerhead ribozymes and the resulting monomers are, in part, circularized by the co-expressed enzyme. The mature viroid forms and the protein likely form a ribonucleoprotein complex that transitorily accumulates in E. coli cells at extraordinarily amounts.


Results
The co-expression of ELVd RNA and eggplant tRNA ligase in E. coli induces accumulation of notable amounts of circular viroid monomers. To investigate the domains and residues in both the ELVd RNA (Fig. 1) and tRNA ligase involved in viroid circularization, we set up an experimental system that consisted of recombinant E. coli clones transformed with plasmids to co-express both molecules (pLELVd and p15tRnlSm, Supplementary Fig. S1, see Supplementary Information). Longer-than-unit ELVd RNA (from C327 to G46, GenBank accession number AJ536613; note that ELVd is circular and A333 is followed by G1), which included the repetition of the plus-strand hammerhead ribozyme domain (Fig. 1), was constitutively expressed under the control of the E. coli murein lipoprotein promoter and the 5S rRNA (rrnC) terminator. A recombinant version of eggplant tRNA ligase, which included the amino-terminal transit peptide that mediates translocation to chloroplasts 16 , but also a carboxy-terminal hexahistidine tag 9 , was expressed in a compatible plasmid of p15A replication origin under the control of the T7 bacteriophage promoter and terminator. The tRNA ligase expression was induced in cultures of E. coli strain BL21(DE3) by adding isopropyl β-D-1-thiogalactopyranoside (IPTG), which activated the expression of a copy of bacteriophage T7 RNA polymerase set under the control of the lacUV5 promoter. E. coli cultures were grown at 25 °C to an optic density at 600 nm (OD 600 ) of 0.6 and the tRNA ligase expression was induced by adding IPTG to 0.4 mM. Cells were harvested at 10 h post-induction and total RNA was recovered by phenol:chloroform extraction. Aliquots of the RNA preparations were separated by denaturing (8 M urea) polyacrylamide gel electrophoresis (PAGE). Gels were first stained with ethidium bromide, and then RNA was blotted to membranes and analyzed by northern blot hybridization using an ELVd 32 P-labeled RNA probe of complementary polarity. This analysis showed that the bacteria co-transformed with plasmids to express ELVd RNA and eggplant tRNA ligase accumulated substantial amounts of the viroid monomeric circular and linear forms, which were clearly detected in both the stained gel (Fig. 2a, lane 3) and the hybridized membrane (Fig. 2b, lane 3). These particular viroid species hardly accumulate in the bacteria transformed with a plasmid to express ELVd RNA alone (Fig. 2b, lane 2). In contrast, species that migrated faster than the monomeric linear ELVd RNA and that most probably correspond to ELVd degradation products were detected in both samples (Fig. 2b, lanes 2 and 3). No ELVd was detected in control bacteria transformed with a plasmid to express the tRNA ligase alone (Fig. 2b, lane 4). The circular nature of the slow-migrating species was confirmed by separating RNA by two-dimension PAGE ( Fig. 2c and d). The very minor amounts of monomeric circular ELVd RNA detected in the absence of tRNA ligase (Fig. 2b, lane 2) most probably result from the 2′,5′ self-ligation reaction previously described in this viroid family 17,18 .
These results suggested that accumulation of monomeric circular and linear ELVd RNAs in E. coli depended on the activity of the co-expressed eggplant tRNA ligase. To gain insight into this hypothesis, we constructed a new plasmid (p15mCherry, Supplementary Fig. S1) to express fluorescent protein mCherry 19 , herein taken as an easy-to-track heterologous control. Once again, only the cells that co-expressed ELVd RNA and tRNA ligase accumulated large amounts of monomeric circular and linear ELVd RNA (Fig. 2e, lanes 5 to 7). It is remarkable to note that, according to the intensity of the bands in the polyacrylamide gels, extraordinary levels of these viroid species accumulated, which were higher than, or at least similar to, those of E. coli 23S and 16S rRNAs ( Fig. 2a and e). However, oligomeric ELVd RNAs ( Fig. 2b and d) and viroid strands of opposite polarity (Fig. 3), typical intermediates of the rolling-circle replication mechanism, were never detected in the hybridization blots, which indicated that ELVd does not replicate in recombinant E. coli cells. Remarkable amounts of the RNAs of interest inserted into the ELVd molecule also accumulate in E. coli when co-expressed with tRNA ligase. The outstanding accumulation of the monomeric forms of ELVd in the E. coli cells, in which eggplant tRNA ligase was co-expressed, led us to consider developing a viroid-derived system to overproduce recombinant RNA. We hypothesized that since ELVd accumulation did not depend on viroid replication, RNAs of interest grafted onto the ELVd molecule would also be efficiently produced in E. coli. According to this aim, we first analyzed the culture condition that maximized ELVd production in E. coli. We compared ELVd accumulation in cultures grown at 25 °C and 37 °C. Although ELVd accumulation was transitory and the viroid almost disappeared with prolonged culture times, the cells grown at 37 °C accumulated more ELVd monomers than those grown at 25 °C ( Supplementary Fig. S2). Within the range of 0.1 to 1.6 mM IPTG, the best ELVd production was observed in the cultures in which the tRNA ligase expression was induced at the lowest IPTG concentration ( Supplementary Fig. S3). However, some IPTG was needed since the control with no added IPTG did not accumulate significant amounts of ELVd. This result reinforces the role of tRNA ligase in the ELVd accumulation process. An analysis of two time-course ELVd expressions, in which IPTG to 0.1 mM was added at cell OD 600 of 0.6 or 0.1, showed better production in the second case ( Supplementary Fig. S4). Remarkably, a comparison of the classic Luria-Bertani (LB) with richer Terrific Broth (TB) media to grow bacteria revealed that ELVd accumulated to extremely high levels in the E. coli cells grown in TB medium (Supplementary Comparing with an RNA standard, we estimated a total combined accumulation of 150 mg of monomeric circular and linear ELVd RNAs per liter of bacterial culture with a circular:linear ratio of 3:1 ( Supplementary Fig. S7).
Next, we investigated the possibility of reducing the size of the ELVd moiety in the final recombinant product. To this end, we constructed a series of pLELVd-derived plasmids in which different viroid cDNA fragments were deleted ( Supplementary Fig. S8). The procedure for the deletions consisted in sequentially removing different elements of the viroid secondary structure without affecting the hammerhead ribozyme domain, which mediates the self-cleavage of the primary transcript ( Supplementary Fig. S9). We analyzed how single ( Supplementary  Fig. S10) and double ( Supplementary Fig. S11) deletions affected the accumulation of the resulting ELVd forms in E. coli. We found that ELVd deleted forms L1R3, L3R1 and L2R2, which respectively consisted of 175, 215 and 246 nt, still accumulated efficiently in E. coli when tRNA ligase was co-expressed ( Supplementary Fig. S11). Moreover, we reasoned that the need to induce the tRNA ligase expression by adding IPTG at a particular cell density can be a serious inconvenience for eventually scaling up the system to industrial fermenters, in which cells could be grown continuously. So in order to further optimize the system, we constructed a new plasmid (p15LtRnlSm) in which the T7 bacteriophage RNA polymerase promoter present in p15tRnlSm was replaced with the constitutive E. coli murein lipoprotein promoter ( Supplementary Fig. S12). E. coli BL21(DE3) was co-transformed with pLELVd, and with either the original p15tRnlSm or new p15LtRnlSm. Recombinant clones were grown to OD 600 0.1 and were induced, or not, with IPTG depending on the harbored tRNA ligase plasmid. A time-course analysis of the RNA produced in both cultures showed no substantial difference in circular and linear ELVd RNA production ( Supplementary Fig. S13). Then, we assayed the production of RNAs of interest inserted into the ELVd molecule. We chose the terminal loop of the upper-right hairpin present in the hypothetical ELVd conformation of minimum free energy as the insertion site (positions U245-U246 of ELVd-AJ536613, Fig. 1). A previous mutational analysis of this viroid has demonstrated that an insertion of eight nucleotides into this particular position, unlike other assayed positions, does not abolish viroid infectivity in eggplants 20 . As an RNA of interest, we chose the RNA aptamer Spinach 13 , a 98-nt-long RNA that emits green fluorescence, which is comparable in brightness to that of fluorescent proteins, when binding to the 3′5-difluoro-4-hydroxybenzylidene (DFHBI) fluorophore. We constructed plasmid pLELVd-Spinach ( Supplementary Fig. S14) in which a cDNA coding for Spinach was inserted between positions T245-T246 of the cDNA coding for ELVd. E. coli BL21(DE3) was co-transformed with pLELVd-Spinach and p15tRnlSm. A recombinant clone was grown and the tRNA ligase expression was induced. Aliquots of the culture were taken at different time points after induction. Interestingly, chimeric ELVd-Spinach RNA also accumulated in the recombinant E. coli cells in vast amounts (Fig. 4a, lane 7). We estimated an accumulation of approximately 75 mg of circular and linear ELVd-Spinach RNA per liter of bacterial culture at 12 h post-induction ( Supplementary Fig. S15). The circular:linear ratio was again 3:1. We observed intense green fluorescence when adding DFHBI to either an aliquot of intact cells or an RNA preparation, which demonstrated that a properly folded Spinach aptamer was produced (Fig. 4b). Next, we analyzed how the partial deletion of the ELVd moiety affected the production of the RNA of interest. We chose the deleted form ELVd L3R1 that conserved the U245-U246 insertion site, and prepared plasmid pLELVdL3R1-Spinach, in which the Spinach cDNA was inserted into the deleted viroid cDNA (Supplementary Fig. S14). Despite the production of the corresponding chimeric RNA being lower than that achieved with full-length ELVd, it still obtained the notable amount of 30 mg of circular and linear ELVd L3R1-Spinach RNA per liter of bacterial culture (Supplementary Fig. S16). In this particular experiment circular:linear ratio was 9:1. We also tested whether the chimeric ELVd-Spinach RNA was infectious in eggplants. Results were negative, in contrast to what occurred with the wild-type ELVd control ( Supplementary  Fig. S17).
We wondered whether this system would also serve to produce recombinant RNAs with extended double-stranded motifs, like those that induce RNA silencing 14 . To answer this question, we constructed plasmids to express RNA hairpin structures of different lengths, which ranged from 40 to 100 nt (pLELVd-hairpin40, pLELVd-hairpin60, pLELVd-hairpin80 and pLELVd-hairpin100; Supplementary Fig. S18). Initial attempts to overproduce these chimeric RNAs in E. coli BL21(DE3) failed ( Supplementary Fig. S19), but hairpin RNAs were efficiently produced (Fig. 5a) when we used the E. coli strain HT115(DE3) that lacked RNase III 21 . Finally, we also produced a 65-nt-long structured RNA moiety consisting of the precursor of a crRNA with two repeated hairpins, like those that guide a Cpf1 nuclease for genome editing 15 (Fig. 5b and Supplementary Fig. S20).

Discussion
Unlike other central players of biological processes like DNA or proteins, the large amounts of recombinant RNA molecules required for research and biotechnological applications are not easily produced in vitro or in vivo, possibly due to the intrinsically short half-life of RNA. However in recent years, some successful strategies, based mainly on using very stable RNA scaffolds 22 or on producing stable ribonucleoprotein complexes 23 , have been developed to produce recombinant RNA in E. coli cultures. We herein present a new system that incorporates both concepts and which, to our understanding, produces much larger amounts of recombinant RNAs than those previously reported. The viroid molecule is an extremely stable RNA scaffold. Indeed, viroids have been shaped by evolution to survive as naked RNAs in the hostile milieu of infected plant cells. The co-expressed eggplant tRNA ligase recognizes the ELVd moiety of recombinant RNA, and most probably participates in the formation of a stable ribonucleoprotein complex that accumulates in vast amounts in E. coli cells. This large accumulation occurs despite the absence of viroid replication intermediates supports no ELVd replication in E. coli. tRNA ligase activity also contributes to the production of recombinant RNAs as circular forms, which must exhibit further stability in vivo due to resistance to exonuclease attack. However, our experimental results also indicate that ELVd and derived RNAs are finally degraded in E. coli at late culture phases and an optimum time must be determined to harvest the recombinant RNA. Another advantage of circularity is that the absence of endogenous circular RNAs in E. coli allows the easy purification of recombinant RNA to virtual homogeneity ( Supplementary Fig. S21). It is true that when using this system recombinant RNAs are produced as chimeric molecules, which include a viroid moiety. Although we also show that the viroid component can be partially reduced, methods have been described to excise RNA of interest from these kinds of hybrid molecules 24 . Concerns about our methodology may arise from using an infectious agent like ELVd. However, ELVd has never been transmitted to other plants, apart from its natural host, the eggplant, where it induces latent infections with no apparent symptoms 12 . Moreover, ELVd chimeric forms with inserted RNAs of interest and, obviously, those that included extended viroid deletions, were not infectious at all.
Viroids are a class of infectious agents of higher plants that exclusively consist of small non-protein-coding circular RNA molecules that occupy the bottom step in the scale of life 25 . Despite this simplicity, they are able to direct their own replication in infected cells and movement through infected plants, frequently inducing disease 26 . Here we show that, like viruses, they can also be transformed into useful biotechnological tools to produce large amounts of recombinant RNAs in bacterial cells, such as aptamers or other structured RNAs. Most importantly, by using the viroid-derived system we accomplished the amazing production of 75 mg of ELVd-Spinach RNA per liter of bacterial culture under standard laboratory conditions.

Methods
Plasmid construction. All plasmids were constructed using standard molecular cloning techniques. DNAs were amplified by the polymerase chain reaction (PCR) using the Phusion High-Fidelity DNA polymerase (Thermo Scientific). DNAs were digested with the type-IIS restriction enzyme Bpi I (Thermo Scientific) and assembled 27 into plasmid by ligation with T4 DNA ligase (Thermo Scientific). The sequences of the resulting plasmids were experimentally confirmed (3130xl Genetic Analyzer, Life Technologies).
Escherichia coli culture. The strains BL21(DE3) (Novagen) and HT115(DE3) 21 of E. coli were electroporated or co-electroporated (ECM 399, BTX) with the different plasmids and recombinant clones selected at 37 °C in plates of LB solid medium (10 g/l tryptone, 5 g/l yeast extract, 10 g/l NaCl and 1.5% agar) that included the appropriate antibiotics (50 µg/ml ampicillin, 34 µg/ml chloramphenicol or both). E. coli was grown in liquid cultures in LB or TB medium (12 g/l tryptone, 24 g/l yeast extract, 0.4% glycerol, 0.17 M KH 2 PO 4 and 0.72 M K 2 HPO 4 ), containing the appropriate antibiotics (see above), at the indicated temperature (in general 37 °C) with vigorous shaking (225 revolutions per min -rpm-). Cell densities were measured at 600 nm with a colorimeter (CO8000, WPA). When needed, induction of eggplant tRNA ligase expression was carried out by adding the appropriate amount of 0.25 M IPTG to the culture.
RNA extraction and analysis. At the desired time points, 2-ml aliquots of the liquid cultures were taken and cells were sedimented by centrifuging at 13,000 rpm for 2 min. In general, cells were resuspended in 50 µl of TE buffer (10 mM Tris-HCl, pH 8.0 and 1 mM EDTA) by vortexing (40-fold concentration). One volume (50 µl) of a 1:1 (v/v) mix of phenol (saturated with water and equilibrated at pH 8.0 with Tris-HCl, pH 8.0) and chloroform was added and the cells broken by vigorous vortexing. The aqueous and organic phases were separated by centrifugation for 5 min at 13,000 rpm. The aqueous phases were recovered and re-extracted with one volume (50 µl) of chloroform. The aqueous phases containing total bacterial nucleic acids were finally recovered by pipetting and either subjected directly to further analyses or stored frozen at −20 °C.
Total RNA from E. coli was analyzed by denaturing PAGE. In general, 20 µl of RNA preparations were mixed with 1 volume (20 µl) of loading buffer (98% formamide, 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.0025% bromophenol blue and 0.0025% xylene cyanol), heated for 1.5 min at 95 °C and snap cooled on ice. Electrophoresis was run for 2 h and 30 min at 200 V in 5% polyacrylamide gels (37.5:1 acrylamyde:N,N '-methylenebisacrylamide) of 140 × 130 × 2 mm in TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) including 8 M urea. Electrophoresis buffer was TBE without urea. Gels were stained by shaking for 15 min in 200 ml of 1 µg/ml ethidium bromide. After washing three times with water, gels were photographed under UV light (UVIdoc-HD2/20MX, UVITEC Cambridge). Image analyses were performed using the UVIDoc 1D software (UVITEC Cambridge). In some experiments (e.g. Fig. 2c and d), RNAs separated by single denaturing PAGE in TBE buffer were subjected to a second denaturing electrophoresis in lower (0.25 × TBE) ionic strength. After the first dimension, the whole lane from 5% polyacrylamide, 8 M urea, TBE gel was cut and laid transversally on top of a 5% polyacrylamide, 8 M urea, 0.25 × TBE gel of the same dimensions and run for 2.5 h at 350 V. An upper limit of 25 mA was set.
For northern blot hybridization analysis, RNAs separated by electrophoresis were electroblotted to positively charged nylon membranes (Nytran SPC, Whatman) and cross-linked by irradiation with 1.2 J/cm 2 UV light (254 nm, Vilber Lourmat). Hybridization was performed overnight at 70 °C in 50% formamide, 0.1% Ficoll, 0.1% polyvinylpyrrolidone, 100 ng/ml salmon sperm DNA, 1% sodium dodecyl sulfate -SDS-, 0.75 M NaCl, 75 mM sodium citrate, pH 7.0, with approximately 1 million counts per minute of 32 P-labelled ELVd RNA of complementary polarity. Hybridized membranes were washed three times for 10 min with 2 × SSC, 0.1% SDS at room temperature and once for 15 min at 55 °C with 0.1 × SSC, 0.1% SDS (1 × SSC is 150 mM NaCl, 15 mM sodium citrate, pH 7.0). Results were registered by autoradiography using X-ray films (Fujifilm). The radioactive probes were produced by in vitro transcription of a linearized plasmids containing a dimeric ELVd (sequence variant AJ536613) cDNA in the proper orientation. One µg of linearized plasmid (Hind III or Xba I) was transcribed with 50 U of T3 bacteriophage RNA polymerase (Epicentre) in a 20-µl reaction containing 40 mM Tris-HCl, pH 8.0, 6 mM MgCl 2 , 20 mM DTT, 2 mM spermidine, 0.5 mM each of ATP, CTP, and GTP, and 50 μCi of [α-32 P] UTP (800 Ci/mmol), 20 U RNase inhibitor (RiboLock, Thermo Scientific) and 0.1 U yeast inorganic pyrophosphatase (Thermo Scientific). Reactions were incubated for 1 h at 37 °C. After transcription, the DNA template was digested with 20 U DNase I (Thermo Scientific) for 10 min at 37 °C, and the probes were purified by chromatography using a Sephadex G-50 column (Mini Quick Spin Column, Roche Applied Science).
ELVd RNA purification. For preparative purposes ( Supplementary Fig. S21), 1 l of total E. coli culture in TB medium distributed in four 1 l baffled Erlenmeyer flasks was grown at 37 °C with intense shaking (180 rpm). Protein expression was induced at an OD 600 of 0.1 by adding IPTG to 0.1 mM. Cells were recovered at 14 h post-induction by centrifugation at 10,000 rpm for 10 min. Sedimented cells were resuspended in water