Generation of virus‐resistant potato plants by RNA genome targeting

Summary CRISPR/Cas systems provide bacteria and archaea with molecular immunity against invading phages and foreign plasmids. The class 2 type VI CRISPR/Cas effector Cas13a is an RNA‐targeting CRISPR effector that provides protection against RNA phages. Here we report the repurposing of CRISPR/Cas13a to protect potato plants from a eukaryotic virus, Potato virus Y (PVY). Transgenic potato lines expressing Cas13a/sgRNA (small guide RNA) constructs showed suppressed PVY accumulation and disease symptoms. The levels of viral resistance correlated with the expression levels of the Cas13a/sgRNA construct in the plants. Our data further demonstrate that appropriately designed sgRNAs can specifically interfere with multiple PVY strains, while having no effect on unrelated viruses such as PVA or Potato virus S. Our findings provide a novel and highly efficient strategy for engineering crops with resistances to viral diseases.


Introduction
Potato (Solanum tuberosum L.) is one of the world's most important food crops. With nearly 400 million tons produced per year, it ranks fourth after rice, wheat and corn (Halterman et al., 2015). However, substantial harvest losses occur due to the plant's susceptibility to a wide range of pathogens, among which Potato virus Y (PVY) is one of the most devastating viral pathogens. PVY is a member of the genus Potyvirus in the family Potyviridae. PVY affects both yield and tuber quality, and can cause crop losses of up to 80% (Quenouille et al., 2013). The symptoms induced by PVY include mosaic, mottled and crinkled leaves as well as leaf and vein necrosis (Faurez et al., 2012;Karasev and Gray, 2013;Quenouille et al., 2013).
The genome of PVY is a single-stranded, positive-sense (+) RNA, with a length of~9.7 kb. The RNA genome is translated into a single large polyprotein of approximately 3061 amino acid residues, and subsequently processed by three virus-specific proteases into the following mature proteins (viral factors): P1, HC-Pro, P3, 6K1, CI, 6K2, NIa (VPg plus Pro), NIb (viral replicase) and CP (capsid protein; Quenouille et al., 2013). In addition, a small open reading frame (ORF), PIPO, was recently found to be embedded within the P3 region (Chung et al., 2008) and shown to be required for efficient viral movement . PVY is prone to high mutation and recombination rates, generating a high level of genetic diversity and a large number of different PVY strains (Davie et al., 2017;Karasev and Gray, 2013). Currently, the main PVY strains are PVY O , PVY N , PVY NTN and PVY N:O (Karasev and Gray, 2013).
Control of PVY and other viruses has been attempted by conventional breeding, transgenic approaches and gene silencing strategies (Chaudhary, 2018). However, its high rates of mutation and recombination allow the virus to evade these strategies, leading to rapid breakdown of resistance. For example, PVY isolates that have overcome the resistance gene va of tobacco plants were reported (Janzac et al., 2014;Nicolas et al., 1997;Takakura et al., 2018). Transgenic RNAi against plant viruses has been reported (Pooggin, 2017), but through long-term coevolution, eukaryotic viruses have acquired effective mechanisms of resistance to RNAi. The production of viral suppressors of RNA silencing (VSRs) is a widespread counter-defense strategy employed (Incarbone and Dunoyer, 2013). For example, VSRs inhibit ARGONAUTE function by precluding target RNA binding to pre-assembled RISC (Kenesi et al., 2017). Another undesired property of RNAi approaches is that constitutive siRNA expression may lead to off-target effects (Pooggin, 2017). For all these reasons, new strategies for improved broad-spectrum resistance to multiple PVY strains in potato are urgently needed.
The clustered regularly interspaced short palindromic repeat and associated proteins (CRISPR/Cas) system provides bacteria and archaea with molecular immunity against invading phages and conjugative plasmids (Barrangou and Marraffini, 2014;Marraffini, 2015;Shmakov et al., 2017). The CRISPR/Cas system has been widely employed to combat DNA viruses in eukaryotes (Hadidi et al., 2016;Mahas and Mahfouz, 2018). By designing appropriate guide RNAs (sgRNAs), CRISPR/Cas9 can cleave the double-stranded DNA of viral genomes and, in this way, confer resistance to DNA viruses (Ali et al., 2015;Ji et al., 2015;Zaidi et al., 2016). A related system, CRISPR/Cas13a (formerly known as C2c2), was shown to cleave single-stranded RNA, thus providing defense against invading RNA phages (Abudayyeh et al., 2016). The CRISPR/Cas13a system can be engineered to mediate the specific knock-down of RNA transcripts. This property makes it somewhat comparable to RNA interference, while having the advantage of reduced off-target effects in both mammalian and plant cells .
The potential of the CRISPR/Cas13a system to interfere with viral RNA replication has recently been tested with the Turnip mosaic virus (TuMV) in the model plant Nicotiana benthamiana in transient transformation assays using a tobacco virus for sgRNA delivery. The data suggest that a reduction in the viral load can be achieved (Aman et al., 2018). Here, we have investigated the potential of the CRISPR/Cas13a system to confer stable resistance to an important viral disease in a major crop. By designing sgRNAs against conserved coding regions of three different PVY strains, we demonstrate that the CRISPR/Cas13a system can be engineered to confer broad-spectrum resistance of transgenic potato plants against multiple PVY strains. We also show that the anti-PVY CRISPR/Cas13a specifically targets PVY genomes, but not other viral genomes (PVA and PVS). Importantly, the efficiency of PVY inhibition was positively correlated with the Cas13a/sgRNA expression levels. Our study suggests that the CRISPR/Cas13a system can be used to engineer resistances to RNA viruses into crop plants.

CRISPR/Cas13a expression in potato plants and target sites selection
Potato virus Y is prone to high mutation rates and recombination, generating a high level of genetic diversity and a multitude of PVY strains (Davie et al., 2017). To target multiple PVY strains, we analysed the genomes of three major PVY strains, PVY O , PVY N and the recombinant PVY N:O strain. Fourteen conserved sequences with >28 nucleotide identity (the minimum size of a guide RNA for mRNA targeting by CRISPR/Cas13a) were found. Four of them, targeting the P3, CI, NIb and CP regions, respectively, were selected and used for sgRNA design (Figure 1a,b; Table S1). The P3 protein is the potyviral membrane protein involved in virus replication, systemic infection, pathogenicity and movement (Cui et al., 2010;Johansen et al., 2001;Quenouille et al., 2013). The CI protein forms the laminate cytoplasmic inclusion bodies (Edwardson, 1992) involved in virus movement and infection (Carrington et al., 1998;Quenouille et al., 2013;Wei et al., 2010). The NIb protein is the RNAdependent RNA polymerase (RdRp) that participates in the replication of the viral RNA (Hong and Hunt, 1996;Quenouille et al., 2013). The coat protein (CP) is required for virion assembly, cell-to-cell and systemic movement, and aphid transmission (Andersen and Johansen, 1998;Atreya et al., 1990;Quenouille et al., 2013;Rojas et al., 1997).
Because PVY replicates in the cytoplasm (Kopek et al., 2007;Otulak and Garbaczewska, 2013), we constructed binary vectors harbouring a synthetic LshCas13a gene driven by the UBQ10 (Arabidopsis ubiquitin-10) promoter, and sgRNAs with oligo (A)rich tails (serving as signal for nuclear export) (Stewart, 2010) driven by the AtU6 promoter. We next introduced the LshCas13a/sgRNA constructs into potato plants by stable Agrobacterium-mediated transformation (Figure 1b). Transgenic lines were selected on plant regeneration medium with kanamycin as the selection agent. We obtained over 25 independent transgenic lines for each of the four constructs. Total RNA from~10 transgenic lines each was extracted, and the expression of LshCas13a and the sgRNAs was determined by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR). The data revealed that transcript levels of LshCas13a correlated well with those of the sgRNAs (Figures 2 and S1). In addition, by fusing an HA tag to Cas13a, a positive correlation between mRNA accumulation and protein accumulation of Cas13a could be demonstrated in transient expression assays ( Figure S2).

CRISPR/Cas13a-mediated interference with PVY
PVY O is a common PVY strain. To test whether our CRISPR/ Cas13a constructs interfere with the accumulation of PVY O , we selected two transgenic lines with high transgene expression levels for each construct and performed PVY resistance assays. To this end, transgenic and wild-type (WT) plants were challenged with PVY O , and the appearance of disease symptoms was monitored. While the typical PVY mosaic symptoms were observed in leaves of the infected WT plants at 25 days postinoculation (dpi), no disease symptoms were observed in any of the transgenic plants ( Figure 3a). Resistance of transgenic plants to PVY O was further confirmed by enzyme-linked immunosorbent assays (ELISA) and qRT-PCR assays which showed strongly reduced PVY accumulation in systemic leaves of transgenic plants at 20 dpi ( Figure S3 To further assess the CRISPR/Cas13a-mediated interference with the replication of the RNA genome of PVY, total RNA from systemic leaves of PVY-infected plants was isolated to detect PVY O genome accumulation by northern blotting. Consistent with the results of the ELISA and qRT-PCR analyses, the data showed that transgenic lines harbouring higher levels of Cas13a/ sgRNA exhibit stronger interference with viral replication than lines with low level of LshCas13a/sgRNA expression (Figure 4d). Together, these data strongly indicate that the efficiency of PVY O inhibition is positively correlated with the LshCas13a/sgRNA expression level.

CRISPR/Cas13a specifically targets PVY genomes
To test if the CRISPR/Cas13a constructs in our transgenic potato plants also confer resistance to PVY N and the recombinant strain These results indicate that the sgRNA sequence can be designed with sufficient specificity to target the genome of a single group of viruses (Figures 1a and 5g).

Discussion
Potato virus Y causes one of the most devastating viral diseases in potato plants (Karasev and Gray, 2013;Quenouille et al., 2013). Attempts to control the disease with conventional approaches such as breeding and RNA interference have met with limited success (Chaudhary, 2018). Here we demonstrate that the CRISPR/Cas13a immune system can be engineered to generate virus-resistant potato plants for PVY control.
By generating stable transgenic potato plants, we first aimed to express LshCas13a/sgRNAs targeting the P3, CI, NIb or CP region   Figure 3). In transient assays, CRISPR RNAs targeting HC-Pro and GFP sequences exhibited better interference with Turnip mosaic virus (TuMV) replication than those targeting CP sequence (Aman et al., 2018). It remains to be investigated whether this is a specific phenomenon of TuMV and/or the transient sgRNA delivery system used in this study.
By directly comparing the interference efficiency in transgenic lines with high and low expression levels of LshCas13a/sgRNAs that target the same CI region in the PVY O genome, the efficiency of PVY inhibition was found to be positively correlated with the LshCas13a/sgRNA expression level (Figure 4), suggesting that the resistance of engineered plants against PVY O depends on the expression level of the LshCas13a/sgRNA construct. Our findings further indicate that a sgRNA designed against a conserved sequence is capable of conferring broadspectrum viral resistance by targeting multiple PVY strains. We successfully targeted PVY O , PVY N and PVY N:O simultaneously with a sgRNA (Figures 3 and 5a-d), while unrelated viruses (PVA and PVS) that have low sequence similarity with PVY remain unaffected (Figure 5e,f).
Potato virus Y replicates in the cytosol. To trigger sgRNA export out of the nucleus, an oligo(A) sequence (of seven adenosines; Data S4) was included immediately adjacent to the spacer sequence. In vitro biochemical and crystal structural data have shown that the secondary structure and sequence of the Scaffold region of the sgRNA are critical for interaction with LshCas13a and ssRNA cleavage (Abudayyeh et al., 2016;Liu et al., 2017a). The Cas13a-sgRNA complex undergoes a significant conformational change upon target RNA binding. A 20-bp guide-target RNA duplex is sufficient to activate Cas13a for efficient cleavage of the target RNA (Liu et al., 2017b). A few of unpaired (extra) nucleotides at the 3 0 end of the sgRNA do not affect cleavage activity (Abudayyeh et al., 2016). Since the extra oligo(A) tail at the 3 0 end of sgRNA in our design does not change any of the essential sequence elements of the sgRNA, and is also unlikely disturb the stem-loop structure of scaffold sequence located at the 5 0 end of the sgRNA, targeting of the RNA virus is unlikely to be affected by the oligo(A) sequence. PVY resistance of our transgenic plants supports this conclusion. If presence of the oligo (A) tail would be of any concern, it can be avoided by employing more precise RNA processing tools, such as the Csy4 RNase (Nissim et al., 2014), a self-cleaving ribozyme (Gao and Zhao, 2014;Gao et al., 2018;Tang et al., 2017) or endogenous tRNA processing enzymes (Xie et al., 2015).
Recent work demonstrated the feasibility of simultaneously expressing several sgRNAs by employing the endogenous tRNAprocessing system of plant cells, thus guiding Cas9 to edit multiple chromosomal targets (Xie et al., 2015). In nature, plants are exposed to a variety of different viruses. Thus, it would be worthwhile to investigate whether the CRISPR/Cas13a system can be used to express multiple sgRNAs against several RNA viruses to produce multiresistant plants. Multiplexing is also a suitable strategy to prevent (or at least substantially delay) the development of resistance of the virus to cleavage by the CRISPR/Cas13a system. Mutation of the target site of the sgRNA in the viral genome can potentially lead to resistance. This can be counteracted by (i) selecting target sites that have little leeway to mutate and/or (ii) targeting several sites in the genome with two or more sgRNAs.
Besides Cas13a, other RNA-targeting Cas proteins have been reported. Recently, the CRISPR-Cas9 system from Francisella novicida was reprogrammed to confer molecular immunity against RNA viruses in N. benthamiana and Arabidopsis thaliana . The VI-D Cas13d from Ruminococcus flavefaciens was identified as a new RNA-targeting CRISPR effector (Konermann et al., 2018;Yan et al., 2018), which is prominently smaller than other Cas13 subtypes and exhibits favourable RNA knock-down efficiency and specificity compared to RNAi (Konermann et al., 2018). All these findings should provide ample opportunities for the development of efficient antiviral strategies based on CRISPR/Cas systems.
Crop plants with virus resistances engineered via the CRISPR/ Cas13 system are transgenic plants and, thus, will fall under the current GMO legislation. This means that, similar to insect-resistant plants expressing Bt toxins, regulatory approval will be required before such varieties can be commercialized. Since potato plants are vegetatively propagated, obtaining regulatory approval may be somewhat less demanding than with sexually propagated plants (that potentially transmit their transgenes via pollen).
In conclusion, we engineered high-level PVY resistance into potato plants by introducing the CRISPR/Cas13a prokaryotic immune system. Importantly, the CRISPR/Cas13a system, while allowing the specific targeting of viral genomes, can be used to interfere with multiple PVY strains. Together with previously reported transient transformation assays in tobacco (Aman et al., 2018), our findings obtained with stably transformed potato plants suggest a new promising strategy to engineer crop plants with resistances to RNA viruses.

Plants, viruses and viral strains
Potato (S. tuberosum cultivar Desiree) plants were grown under standard greenhouse conditions as described previously . The viral strains and isolates used in this study, including PVY O -FL (accession number HM367075, Data S1; Nie et al., 2011), PVY N -Jg (accession number AY166867; Nie and Singh, 2003), PVY N:O -Mb112 (accession number AY745491; Nie et al., 2004), PVS (accession number KU896946; Wang et al., 2016) and PVA (accession number KF977085; He et al., 2014) were maintained in tobacco or potato plants as hosts. Strain identity and purity were verified by reverse-transcription polymerase chain reaction (RT-PCR) and ELISA as described previously (Wang et al., 2016).

Design and construction of LshCas13a/sgRNA cassettes for the expression in Solanum tuberosum
The Leptotrichia shahii Cas13a (LshCas13a) gene sequence was designed with codon optimization for S. tuberosum (Data S2 and S3). The LshCas13a gene and the 32 bp sgRNA-encoding sequence containing two BsaI restriction sites were synthesized by Genecreate (Wuhan, China). The LshCas13a gene was linked to the UBQ10 promoter and the nos terminator. The sgRNAs (Data S4) were designed as primer dimers (Table S2) with overhangs, and were cloned under the Arabidopsis U6 promoter using the restriction enzymes SacII and SbfI. The final vectors (including LshCas13a and the sgRNA expression cassettes) were named pP3, pCI, pNIb and pCP. All cloned sequences were confirmed by resequencing.

Agro-infiltration of Nicotiana benthamiana leaves
Constructs harbouring the HA-Cas13a clone were transformed into Agrobacterium tumefaciens strain GV3101 by electroporation. Inoculation of leaves with Agrobacterium was performed according to previously described protocols (Ji et al., 2015). Briefly, overnight-grown cultures were harvested by centrifugation and suspended in infiltration medium (10 mM MES, 10 mM MgCl 2 , 150 lM acetosyringone). The suspension was diluted to concentrations of OD 600 1.0, 0.33, 0.25 and 0.2, and then infiltrated into 4-week-old leaves of N. benthamiana with a 1 mL needleless syringe.

Western blot analysis
Cas13a expression was detected by western blot analysis. A commercial anti-HA antibody (ABclonal, Oxfordshire, UK) was used as primary antibody (1 : 3000 dilution). The signals on the membrane were visualized using the enhanced chemiluminescence substrate (SuperSignal West Pico; Pierce, Rockford, IL) following the manufacturer's instructions.

Generation of transgenic Solanum tuberosum plants
Agrobacterium tumefaciens strains harbouring the LshCas13a/ sgRNA constructs were employed to transform S. tuberosum var. D esir ee plants using previously described protocols . Transgenic plants were identified by their resistance to kanamycin (50 lg/mL) and initially tested for the presence of the transgene by PCR assays. Transgene expression was subsequently confirmed by qRT-PCR.

Mechanical inoculation of viruses and ELISA assay
To determine the response of transgenic plants to various viral strains and isolates, wild-type and transgenic plants were grown in the greenhouse at 20-25°C. Five to eight plants of each line at the 6-8 leaf stage were infected with inocula of viral strains or isolates (leaf extract: 1 g leaf tissue homogenized in 0.1 M potassium phosphate buffer, pH 7.6) by mechanical wounding as described previously (Nie et al., 2004). The same number of wildtype plants inoculated with the corresponding virus served as control. Symptom development in the potato plants was recorded at 5-day intervals for 1 month on both the inoculated and the upper uninoculated (systemic) leaves. Double-antibody sandwich enzyme-linked immunosorbent assays (DAS-ELISA) with virusspecific antibodies (Agdia, Elkhart, IN) were conducted to measure virus accumulation in the upper uninoculated systemic leaves at 15, 20 and 25 days post inoculation (dpi) as described previously (Wang et al., 2016).

RNA extraction and qRT-PCR analyses
Transcription of LshCas13a/sgRNA in transgenic plants was detected by qRT-PCR assays. Total RNA of potato plants was extracted with the TRIzol reagent (Invitrogen, Carlsbad, CA). cDNA was synthesized using reverse transcriptase and oligo (dT) primer (Takara, Shiga, Japan). qRT-PCRs were carried out in a CFX96 Touch TM real-time PCR detection system (Bio-Rad, Hercules, CA) using SYBR â Premix Ex Taq TM II (Takara). qRT-PCRs were repeated in three independent experiments. Three technical replicates were performed for each biological replicate. The potato TUBULIN 2 gene was used as reference. Primer sequences for qRT-PCR are listed in Table S2.

Northern blot
Northern blot analyses were performed according to the protocol reported by Zhang et al. (2015). A 480 bp RNA probe for PVY was synthesized and labelled using the DIG Northern Starter Kit (Roche, Basel, Switzerland).

Statistical analyses
The results presented in the figures represent the average of three independent experiments. The significance of the differences between two datasets was assessed by independent-samples ttest.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.

Figure S1
Analysis of the relative expression of sgRNAs in transgenic lines by qRT-PCR. Figure S2 Transgenic potato plants display resistance to PVY O . Figure S3 Determination of Cas13a expression levels in transient assays by qRT-PCR and western blot analyses. Table S1 PVY genome annotation. Table S2 List of oligonucleotides used in this study. Data S1 PVY O full-length sequence and sgRNA target sequences. Data S2 LshCas13a amino acid sequence. Data S3 LshCas13a full-length DNA sequence (codon optimized for expression in the plant nuclear genome). Data S4 Sequences of synthetic genes for expression of sgRNAs.