Background
Over the past 25 years, our ability to discover and characterize viral agents has steadily improved, leading to a constant flow of discovery of novel plant viruses as testified by the literature and by the constant increase in the number of viral species described in successive reports of the International Committee for the Taxonomy of Viruses [
1]. The development of next generation sequencing (NGS) techniques promises to increase the rate at which novel plant viruses will be discovered in coming years [
2,
3]. At the same time, our ability to unambiguously establish the contribution of newly characterized viral agents to particular plant diseases has not improved. The fulfilling of Koch's postulates has been modified by L. Bos to be adapted to viruses, and represents a fundamental point in plant virology [
4]. With the application of these postulates, the role of many viruses in diseases has been deciphered. But for many other plant viruses, technical problems in the identification of alternative herbaceous hosts, in purification or in experimental transmission have prevented the analysis of their contribution to particular diseases [
4]. This is especially true for viruses affecting vegetatively propagated crops [
5‐
7], which often have the added disadvantage of being frequently mixed infections [
8]. Thus, for many viruses, the demonstration of their involvement in a given disease has not been completed, but has only been postulated on the basis of an association with symptomatic plants [see for example [
9,
10]].
One strategy to bypass the problems encountered with fulfilling of Koch's postulates involves the use of full-length cDNAs clones (FL-cDNAs) (or DNA clones in the case of DNA viruses) from which infectious RNA transcripts can be obtained
in vitro or
in vivo[
11]. However, the construction of infectious FL-cDNAs is still often complicated and time-consuming for many reasons: the difficulty to optimize a standardized protocol for all viruses, the necessity of a perfect junction of the promoter and 5' end of the viral sequence, the difficulty to clone large cDNA molecules and the frequent instability of such clones [
11].
These difficulties have largely limited the use of FL-cDNAs to studies on reverse genetics of well characterized viruses, which have provided access to valuable information on the expression of viral genomes, their replication and mechanisms involved in the infection cycle. They also provided further insight on the functions of different viral proteins or the mechanisms of interaction between viruses and their host plant(s) or vector(s).
However, despite their potential to address such questions, the use of infectious FL-cDNAs to confirm or refute etiological hypotheses has been rather limited [
12‐
15]. In a recent example, the construction of an agroinfiltrable FL-cDNA clone of
Citrus leaf blotch virus (CLBV) allowed the demonstration that CLBV is the causal agent of the Dweet mottle disease and that in single infections it does not cause the bud union crease disease [
16]. An example of the widespread use of infectious constructs for etiology studies is in the
Geminiviridae family, for which efficient techniques exist for the development of both cloned or uncloned infectious DNA constructs [
17,
18]. However, there are additional technical difficulties when working with RNA viruses that are responsible for the limited use of FL-cDNAs in etiology studies of RNA plant viruses.
Simplified strategies for the easier and faster development of infectious FL-cDNA for etiology studies of plant viruses should have a number of desirable properties. First, is the ability to use total nucleic acids (TNA) extracts from infected plants as template for cDNA synthesis [
12,
19], rather than purified viral RNA as this would bypass the need to propagate and purify the virus. Second, is the ability to use long distance PCR [
20] to amplify the viral genomes as single, large PCR fragments, a technique that has been used rarely for genomes longer than 7 kb [
12,
19,
21‐
23]. In a number of situations, cloning of the infectious FL-cDNA may not be necessary to validate an etiology hypothesis, so that the ability to infect plants using uncloned PCR products is also of potential interest [
24‐
26]. Lastly, when cloning of FL-cDNAs is used, techniques that facilitate the cloning of long PCR fragments or the one-step assembly of complex constructions would be of great interest. One little used strategy with such a potential is the use of the efficient homologous recombination machinery of the yeast
Saccharomyces cerevisiae. Until recently, the application of this system has been limited to yeast genetics and to the construction of plasmids and yeast artificial chromosomes (YACs) [
27]. The full power of this approach has been demonstrated recently by the assembly of 25 overlapping DNA fragments to generate a synthetic mycoplasma genome in a single step [
28]. In virology, the application of this strategy has been used as an alternative to difficult classical cloning in
Escherichia coli, as in the case of Dengue virus type 2, where three cDNA fragments of the virus were assembled by homologous recombination in yeast to generate an infectious FL-cDNA [
29]. The fact that recombination is very efficient, even with short, 20-30 nucleotides-long overlap regions between fragments created using PCR primers has facilitated the construction of recombinant viral genomes [
30,
31].
In the present study,
Apple chlorotic leaf spot virus (ACLSV), the type species of the genus
Trichovirus within the family
Betaflexiviridae[
1,
32,
33], was used as a model system for the development of approaches that fulfill some of the above criteria for improved preparation of infectious FL-cDNAs. The genomic RNA of ACLSV is about 7.55 kb in length [
34,
35] and an infectious FL-cDNA for a Japanese isolate from apple (P-205) under the control of the CaMV 35S promoter has been constructed [
36]. We report on the long distance PCR amplification from TNA extracts of infectious FL-cDNAs under the control of the T7 promoter. We also show that the yeast homologous recombination system permitted the efficient cloning of such large FL-cDNAs and the simultaneous one-step tailoring of a ternary yeast-
E. coli-Agrobacterium tumefaciens shuttle vectors allowing efficient infection of plants by agroinfiltration.
Discussion
Long Distance (LD) RT-PCR amplification had been used successfully to generate infectious uncloned PCR products [
24,
26] but starting with purified viral RNA preparations obtained from purified viral particles. The need for such purified viral RNA preparations constitutes a severe limitation, in particular for viruses infecting woody hosts and for which alternative herbaceous hosts and/or purification techniques may not be available. Although the results presented here were obtained with total extracts from
C. quinoa, recently the same technique was used successfully to amplify the complete genome of several isolates of
Apricot latent virus (ApLV, 9.2-9.3 kb) from total RNA extracts obtained from infected GF305 peach seedlings, which indicates that the protocol developed is not limited to herbaceous hosts extracts or to the 7.5 kb length of the ACLSV genome (F. Youssef
et al., unpublished). On the other hand, it was not possible to obtain similar results when using viral double-stranded RNA (dsRNAs) preparations or immunocaptured virions. Given the infectivity of the pYEA-ACLSV agroinoculable construct to GF305 peach seedlings, all the steps seem today at hand for going from TNA extracts obtained from woody hosts to successful inoculation of such woody hosts using uncloned or cloned constructs. The efficiency of such a process remains, however, to be directly evaluated. Similarly, whether the presence in the original host of several viral strains in mixed infection (as often observed in woody hosts) will reduce the efficiency of production of infectious constructs remains to be evaluated.
All three evaluated enzymes (Advantage
® GC Genomic LA Polymerase Mix, Expand High Fidelity PCR System and Phusion
® High Fidelity DNA Polymerase) allowed, although with variable efficiency and reproducibility, the amplification of the T7-FL-cDNA. The results obtained in the present work indicate that the inoculation efficiency may vary considerably depending on the inoculated host species. While such an observation is not overly remarkable when comparing the inoculation of herbaceous hosts with that of the notoriously more difficult woody hosts, the large difference observed when comparing mechanical (Table
2) or biolistic (results not shown) inoculation of
in vitro transcripts on
C. quinoa and on
N. occidentalis 37B is more surprising. Several hypotheses can be proposed to explain this observation, including an intrinsic difference in the susceptibility of these two host plants, possibly linked with a more rapid degradation of inoculated transcripts in
N. occidentalis or, on the contrary, the need for host-specific adaptative mutations in ACLSV in order to infect this host. The fact that, contrary to the situation in
C. quinoa, a significant proportion of
N. occidentalis plants sometimes escape infection upon mechanical inoculation with crude plant homogenates, even when the viral isolate used had been passaged in this host argues against the second hypothesis. However, further experiments are required to clarify this point that could have important practical implications in view of the fact that
N. occidentalis is the only herbaceous host known for many fruit tree-infecting
Betaflexiviridae. In this respect, conflicting results were reported previously, since Saldarelli
et al.[
26] were able to infect this host with transcripts of the
Grapevine virus A and
Grapevine virus B vitiviruses while Vives
et al.[
16] could only obtain
Citrus leaf blotch virus infection when using agroinoculation of a full-length construct.
Although PCR-amplified viral FL-cDNA molecules have been cloned with success in
E. coli vectors [for example [
12,
13,
16,
19,
21‐
23]], such cloning frequently is difficult, in particular for large FL-cDNAs. We therefore evaluated the potential usefulness of the yeast homologous recombination system for the one-step cloning of the ACLSV FL-cDNAs. As demonstrated by the successful simultaneous assembly of the ternary shuttle vector and cloning of the ACLSV FL-cDNA, this system is indeed versatile and efficient. Given that it has been used for the simultaneous assembly of numerous large DNA fragments [
28], it should not prove particularly sensitive to the size of the cloned viral genome. However, in the experiments reported here, infectious plasmids only represented a fraction of all plasmids that showed the expected pattern upon restriction mapping (from 2.5% to 6.5% depending on experiment). A similarly low rate of recovery of infectious cDNA clones was reported for another
Betaflexiviridae, CLBV [
16]. Spetz
et al.[
19] reported a very low recovery of infectious constructs and the only recovered plasmid showed different biological properties than the input virus. In the experiments reported here, the rate of recovery of infectious constructs was dramatically improved when a FL-cDNA amplified from an infectious cDNA clone was used instead of one amplified from total RNAs from infected plants. The low rate of recovery of infectious constructs might be due to either a low rate of infectious molecules in the starting viral RNA population and/or the introduction of detrimental mutations by the reverse transcription step but seem to exclude an impact of the PCR amplification step. In any case, the pooling strategy reported here allowed the efficient and rapid identification of the few infectious clones. It could probably be further improved by using a two or three dimensional pooling approach, so that individual infectious molecules could be identified in a single step. The potential to assemble multiple DNA fragments [
28] could also be exploited to join in a single step partial cDNA fragments spanning the very long genomes [
29] of viruses such as members of the
Closteroviridae family.
Conclusions
We report here for the first time a set of techniques and amplification conditions that allow the use of the crude total RNA extracts from ACLSV infected C. quinoa plants as templates for the Long Distance (LD) RT-PCR amplification of infectious full-length viral cDNAs. As far as we can ascertain, this represents the largest RNA plant virus genome amplified by PCR from TNA extracts. We also demonstrate that homologous recombination in yeast allows for the fast and efficient cloning of infectious FL-cDNAs and/or one-step tailoring of complex constructs.
Overall the strategies reported here allow for the rapid generation of uncloned or cloned infectious FL-cDNA from total RNA preparations from infected hosts and should prove useful in a range of studies, and, in particular, for the validation of etiological hypotheses involving difficult to manipulate plant viral agents.
Materials and methods
Virus source and RNA extraction
The P-205 Japanese strain of ACLSV, [
34], was provided by Dr. N. Yoshikawa (Iwate University, Japan) as an infectious cDNA clone under the control of the CaMV 35S promoter, pCLSF [
36]. The viral isolate obtained upon biolistic inoculation of pCLSF to
C. quinoa was propagated in this host by mechanical inoculation and used as source of virus or of viral nucleic acids. Total nucleic acids were extracted with the SV Total RNA Isolation System (Promega, Lyon, France) from 30 mg of leaf tissue from symptomatic
C. quinoa plants, following the manufacturer's instructions. The yield was approximately 15 μg of RNAs in 60 μl of sterile water.
Full-length ACLSV cDNA amplification using Long Distance (LD) RT-PCR
Synthesis of the first-strand cDNA was primed with oligonucleotide FLAC3: 5' T(30)GTAGTAAAATATTTAAAAGTCTACAG 3', which is complementary to the 3'-terminal nucleotides (positions 7552-7527) of the genomic RNA of ACLSV-P205 (GenBank D14996). The synthesis was performed using about 1 μg of TNA and the PrimeScript™ Reverse Transcriptase (Clontech/TaKaRa Bio Europe, Saint-Germain en Laye, France) or the Expand Reverse Transcriptase (Roche Diagnostics, Meylan, France), following the manufacturer instructions. In order to obtain a FL-cDNA under the control of the T7 promoter, primer FLAC5 (5' TAATACGACTCACTATAG TGATACTGATACAGTGTACACTCACG 3') was used in combination with FLAC3 in a LD-PCR experiment. FLAC5 contains the T7 promoter (in bold and italic) and the first 26 5' nucleotides of the genome of ACLSV-P205. Three commercial DNA polymerases or DNA polymerases mixes were compared for their efficiency in this study: the Advantage® GC Genomic LA Polymerase Mix (Clontech), the Expand High Fidelity PCR System (Roche) and the Phusion® High Fidelity DNA Polymerase (Finnzyme/Fischer Scientific, Illkirch, France). All enzymes were used according to their supplier recommendations, using as template 3 μl from the first strand cDNA product either undiluted or diluted 10-fold. All PCRs were performed in a 25 μl final reaction volume. The PCR cycling parameters recommended by the suppliers were used, with the exception of the annealing temperature which was fixed at 60°C for all experiments. Amplicons were purified using the MSB® Spin PCRapace kit (Invitek, les Ullis, France) and eluted in Diethylpyrocarbonate (DEPC) treated, distilled water. They were then quantified by electrophoresis on non-denaturing agarose gels and image analysis using the Image J 1.42q software (W. Rasband, NIH, USA). As a control in some experiments, LD-PCR amplification was performed directly on the pCLSF infectious plasmid instead of on cDNAs obtained from TNA extracts.
T7 RNA polymerase in vitro transcription of ACLSV FL-cDNAs
Two hundred nanograms of purified ACLSV FL-cDNA amplicons were used in in vitro transcription reactions performed using the phage T7 RNA polymerase and the mMESSAGE mMACHINE kit (Ambion, Courtaboeuf, France) in the presence of the cap analogue m7G(5')ppp(5')G (Ambion). At the end of the synthesis, which was performed as recommended by the kit supplier, transcripts were treated with 1 μl of TURBO DNase (Ambion) and then used either directly or after purification on Macro SpinColumn™G-50 (Harvard Apparatus, Les Ullis, France) following the manufacturer's instructions. Transcripts were quantified as described above for the PCR products and, if necessary, were stored at -80°C until use in plant inoculation assays.
Inoculation of herbaceous or woody ACLSV host plants
The ACLSV FL-RNA transcripts were inoculated on young (four leaf stage) plants of C. quinoa or N. occidentalis cv. 37B by gently rubbing 5 μl of in vitro transcripts (adjusted to the desired concentration with DEPC-treated distilled water) on two celite-dusted leaves of each plant (Celite 545, 0,01-0,04 mm, Merck/VWR, Fontenay-sous-Bois, France). Following inoculation, leaves were rinsed briefly with distilled water. As control, plants were inoculated using DEPC-treated distilled water without RNA transcripts. All plants were grown (16 hours day/8 hours night) and observed weekly for symptoms development.
Alternatively, the ACLSV FL-RNA transcripts were biolistically inoculated on leaves of young C. quinoa plants using the Helios Gene Gun (Bio-Rad/Marnes-La-Coquette, France). For inoculation, 5 μg of transcripts were used to prepare 10 cartridges as recommended by Bio-Rad, using 1 μm gold particles. Each plant was bombarded on two leaves, using a 200 psi pressure. For the inoculation of woody host plants, young germinating peach (Prunus persicae, cv. GF305) or apple (Malus domestica) seedlings were inoculated when the growing radicle had reached a length of about 0.5-1 cm. Briefly, the envelopes of germinating seeds were gently removed and the cotyledons were then inoculated by biolistic bombardment (1 shot on each of the cotyledons for the GF305 and 3 shots for groups of 10 germinating seeds for M. domestica. After inoculation, seeds were rinsed with distilled water and then placed on wet tissue paper in Petri dishes at 4°C in the dark for 24 hours. The next day, they were treated by a broad spectrum fungicide by soaking for 30 min in a 0.5% solution of Propamocarb-HCl followed by rinsing with water. The germinating embryos were finally transplanted to sterile moist sand and left to develop in the greenhouse up to the 4-leaf stage before being transplanted into pots containing potting soil. As control, plants were biolistically inoculated using the infectious pCLSF plasmid.
ACLSV detection in inoculated plants
ACLSV was detected as described [
37], using double-antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) and immunoglobulins from the P2 polyclonal antiserum raised against the P863 ACLSV isolate (INRA, Bordeaux, France). Alternatively, ACLSV infection was detected using the A52-A53 ACLSV-specific RT-PCR assay described [
38]. Identity of the amplified virus was confirmed by direct sequencing of the amplification products (Beckman Coulter Genomics, Meylan, France).
One step cloning of ACLSV FL-cDNAs by homologous recombination in Saccharomyces cerevisiae
The yeast (
S. cerevisiae) strain YPH501 (
MAT a/
MAT α
ura3-52 lys2-801 amber
ade2-101 ochre
trp1- Δ
63 his3- Δ
200 leu2- Δ
1) was used [
39]. The yeast-
E. coli vector used in these experiments is a pBS70T plasmid [
40] in which a DNA fragment containing a yeast 2μ origin of replication and the yeast
TRP1 gene as a selection marker has been inserted (Yeast-pBS70T). The pBS70T backbone provides an
E. coli ColE1 origin of replication and an ampicillin resistance gene for selection in
E. coli, in addition to signals for transcription
in planta: duplicated CaMV 35S promoter (70SP) and Nopaline synthase gene terminator (NosT) of the cloned FL-cDNA. The complete vector was amplified using the Phusion
® High Fidelity DNA Polymerase (Finnzyme) and either one of the divergent primer pairs shown in Table
1. This strategy produces PCR products that represent a linear copy of the vector with terminal regions of 30 nt overlapping with the ends of the FL-cDNA PCR product to be cloned. All PCR products were purified using the MSB
® Spin PCRapace kit before use.
Yeast cultures were grown at 30°C in YPD medium (1% yeast extract, 2% peptone, 2% glucose) prior to transformation and in SD (Minimal synthetic defined Base) selective medium with the mixture of the amino acid without tryptophan (-Try DO, Dropout) after transformation. Yeast was transformed using a lithium acetate method and denatured carrier DNA as described [
41], following the modifications of [
30]. Approximately 1-2 μg in total of the mixture of the PCR fragments in equimolar amounts were used per transformation. All yeast colonies growing on the selection plates following transformation were collected in 3 ml of SD liquid medium and grown for 12 h at 30°C. Plasmid DNA was isolated from these cultures according to [
42] and used for transformation of
E. coli cells (XL1) by electroporation. Further characterization of the recombinant plasmids and their large-scale purification were carried out using standard protocols [
43]. All junctions in the recombinant plasmids obtained were confirmed by sequencing (Beckman Coulter Genomics).
One step construction of an agroinoculable ACLSV FL-cDNA in a ternary yeast-E. coli-A. tumefaciens vector
ACLSV FL-cDNA was cloned in a
de novo constructed ternary yeast-
E. coli-
A. tumefaciens vector by homologous recombination in yeast cells. This was achieved by simultaneous transformation of yeast cells using 4 PCR products having 25-32 bp overlapping ends derived from the PCR primers (Figure
3 and Table
1). Besides the ACLSV FL-cDNA, the three fragments used were: (1) a 10.7 kbp product covering the 35S terminator, the T-DNA right border (RB), the
E. coli ColE1 origin of replication, the
A. tumefaciens OriV origin of replication and finally the kanamycin resistance gene from the pBin61 binary vector [
44], (2) a ~2 kbp fragment containing the yeast 2μ origin of replication and the yeast
TRP1 gene derived from the Yeast-pBS70T vector described above and (3) a ~2 kbp fragment containing the T-DNA left border (LB) and the CaMV 35S promoter of pBin61. The long fragment was amplified using the Advantage
® GC Genomic LA Polymerase Mix kit while the two shorter ones were amplified with the Phusion
® High Fidelity DNA Polymerase. After purification of all fragments using the MSB
® Spin PCRapace kit, yeast cells were transformed with a total of about 3 μg of the four fragments in equimolecular amounts. The various steps following transformation were carried out as described above and the recombinant plasmids harbored by
E. coli colonies finally verified by restriction analysis. Plasmids showing the expected
Eco RI restriction pattern were divided into pools and used to transform by electroporation (with ~100 ng of purified plasmid pools)
A. tumefaciens C58C1 cells carrying the pG90 helper plasmid (provided by S. Vernhettes, INRA-Versailles).
Agrobacterium transformants were selected on LB medium plates supplemented with 50 mg/l rifampicin, 50 mg/l kanamycin and 20 mg/l gentamycin. All transformed bacteria growing on the Petri dishes were collected and grown as pools in LB liquid medium with the same antibiotics and this culture was then used as the starter culture to prepare
Agrobacterium cells for agroinfiltration.
Inoculation of plants using agroinfiltration
For agroinfiltration of plants, A. tumefaciens cells carrying the relevant plasmid(s) were first grown overnight at 28°C in 5 ml of LB medium plus selection antibiotics (see above). These pre-cultures were then used to inoculate 25 ml cultures of induction medium [LB medium supplemented with selection antibiotics, 10 mM of 2N-morpholino-ethane sulfonic acid (MES) and 20 μM of acetosyringone]. Following incubation overnight at 28°C under agitation, bacteria were collected by centrifugation at 6000 g for 15 min at room temperature, re-suspended in infiltration medium (10 mM MgCl2, 10 mM MES, pH 5.6, and 150 μM acetosyringone) and the bacterial suspension adjusted to an optical density of 0.6 at 600 nm. The suspension was then incubated for 3 h at room temperature before being infiltrated in the intercellular spaces of young C. quinoa or N. occidentalis leaves, using a syringe directly placed on the lower leaf surface. Alternatively, young GF305 peach seedlings were inoculated by injections of the Agrobacterium cells suspension in their stems using a syringe and a small gauge needle. Following their inoculation, plants were monitored weekly for symptoms appearance and ACLSV infection was confirmed as described above.
Acknowledgements
FY was supported by a fellowship from the Syrian Ministry of Higher Education. The Apple chlorotic leaf spot virus infectious cDNA clone pCLSF was a generous gift of Dr. N. Yoshikawa (Iwate University, Japan), who also shared with us prior to publication his protocol for the biolistic inoculation of apple seedlings. The Agrobacterium C58C1 strain carrying the pG90 helper plasmid was provided by Dr. S. Vernhettes (INRA Versailles, France) and the YPH501 yeast strain by Dr Benoit Moury (INRA Avignon, France). We also wish to thank M. Roncoroni, T. Mauduit, and A. Bailly for taking excellent care of all the plants used in this work, Dr. Emmanuel Jacquot (UMR Bio3P, Rennes, France) for valuable discussions and suggestions and Anas Abdul-Razzak (UMR GDPP, Bordeaux, France) for sharing with us his skills on cloning by homologous recombination in yeast.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors contributed to the results presented. FY, AM, PG and TC contributed to and edited the manuscript. All authors read and approved the final manuscript.