Three gRNAs were designed against the CP gene of GVA, using the software cas13design (https://cas13design.nygenome.org/, accessed on 15/01/2020) (Wessels et al., 2020; Guo et al., 2021). gRNAs were selected based on their rank, guide score, quartile, and off-target hits. N. benthamiana off-targets were assessed against the N. benthamiana transcriptome (https://sefapps02.qut.edu.au/blast/blast_link2.cgi, accessed on 15/01/2020). The secondary structure of the CasRx scaffold sequence and gRNA target [gaacccctaccaactggtcggggtttgaaacG (22-28nt target)] were assessed using the programs mFold (http://www.unafold.org/mfold/applications/rna-folding-form.php, accessed on 20/01/2020) and RNAfold (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi, accessed on 20/01/2020). The secondary structure of the target RNA was assessed with the same software, to ensure that the gRNA, specifically the seed region, was designed to target-accessible (single-strand or looped) regions (Bandaru et al., 2020). The gRNAs were designed to contain BbsI overhangs on their 5’-ends, for subsequent cloning with Golden Gate assembly. For cloning into the TRV2 vector, the gRNAs were designed to contain an XbaI overhang and CasRx scaffold sequence on their 5’-end and an XhoI overhang on their 3’-end ( Supplementary Table 1 ).

Gibson Assembly (NEBuilder HiFi DNA Assembly Cloning Kit; NEB, USA) was used to modify two intermediate vectors: pJJB308, (Addgene; plasmid #107699) and pJJB296, (Addgene; plasmid #107691) to contain the inserts CasRx-gRNA promoter with scaffold sequence (~300bp), and the CasRx gene (~3kb), respectively, following the manufacturer’s protocol. Vector pJJB308 was digested with AarI, and the 2017bp vector fragment was recovered. Vector pJJB296 was digested with BamHI and HindIII, and the ~3.1kb vector fragment was recovered. PCR with a high-fidelity DNA polymerase (Phusion^® High-Fidelity DNA Polymerase; NEB, USA) was performed to amplify the desired insert fragments from pXR003:CasRx-gRNA-cloning-backbone, (Addgene plasmid #109053) and pXR001:EF1a-CasRx-2A-EGFP, (Addgene plasmid #109049) using specific primers ( Supplementary Table 2 ). The purified insert and vector fragments were assembled according to the manufacturer’s instructions. The targets were ordered from IDT (Integrated DNA Technologies, USA) and synthesized as single strand DNA (ssDNA) oligonucleotides containing a BbsI (NEB, USA) overhang on the 5’-end ( Supplementary Table 1 ). Adapter preparation consisted of annealing the ssDNA oligos in a reaction containing 1.5µL forward oligo (100µM), 1.5µL reverse oligo (100µM), 5µL 10X NEBuffer 2.1, and dH2O to a final volume of 50µL. The reactions were incubated in a thermocycler at 95°C for 4 minutes, followed by 70°C for 10 minutes. Reactions were transferred to a beaker containing 1L of H₂O at 70°C and left to cool to room temperature. The pJJB308-CasRx-backbone was linearized with restriction enzyme BbsI (NEB, USA) and the ~2.3kb fragment was recovered. A ligation reaction containing the annealed oligos and linearized vector was set up using 1µL annealed oligos and 20ng linearized vector. The annealed gRNAs were cloned under the hU6 promoter, and upstream of the CasRx-gRNA scaffold. The vector maps of the Gibson-assembled intermediate vectors are shown in Supplementary Figure 1 .

The final binary vector pCasRx : CP was constructed using a one-step Golden Gate assembly, with the following vectors: the Gibson-assembled vectors pJJB308-CasRx-(gRNA) and pJJB296-CasRx, and intermediate module vector pMOD_C0000 (Addgene; plasmid #91081) and transformation backbone vector pTRANS_220d (Addgene; plasmid #91114). For the assembly of the vector without a gRNA (pCasRx-EMPTY), the modified pJJB308-CasRx-backbone plasmid was used in the Golden Gate assembly reaction. The reaction was set up as described in the paper by Čermák and colleagues (Čermák et al., 2017).

The gRNAs for pTRV2 (SPDK3876 (TRV2-pPEBV-MCS), Addgene plasmid #149275) were designed to contain the CasRx-scaffold sequence (5’-AACCCCTACCAACTGGTCGGGGTTTGAAACG-3’) at the 5’-end, and gRNA sequence with poly-T-tail [5’-(gRNA sequence)TTTTTTTTT-3’] at the 3’-end, as well as XbaI and XhoI overhangs on the 5’- and 3’-end respectively ( Supplementary Table 1 ). The targets which were to be cloned into TRV were ordered and synthesized as ssDNA oligos from IDT (Integrated DNA Technologies, USA). The oligos were phosphorylated using T4 Polynucleotide Kinase (Promega, USA), following the manufacturer’s protocol. The phosphorylated oligos were annealed as previously described. The TRV2 vector was digested with XbaI and XhoI and ligated with the phosphorylated and annealed gRNAs.

Overnight-grown cultures of A. tumefaciens strain EHA105 containing the construct pCasRx-EMPTY were centrifuged and suspended in sterile MS3 broth (4.4g/L Murashige & Skoog (Murashige and Skoog, 1962) (MS) medium, 30g/L sucrose, 100µg/mL acetosyringone, pH 5.8) to an optical density (OD)₆₀₀ of 0.8-1. Leaves of wild-type N. benthamiana grown in vitro were dissected into approximately 50 leaf disc explants of 1cm² each, and used for Agrobacterium-mediated stable transformation, as described by Clemente (Clemente, 2006). Shoot organogenesis was induced on leaves cultured on selective medium (4.4g/L MS, 30g/L sucrose, 3.3g/L PhytoAgar, 1mL 6-Benzylaminopurine (BAP) (1µM), pH 5.8), incubated at 25°C with a 16hr/8hr light:dark, and explants on fresh selective medium every 14 days. Regenerated shoots were excised from calli and placed on hormone-free rooting medium (half-strength MS, 15g/L sucrose, 3.3g/L PhytoAgar (Duchefa Biochemie, NL), pH 5.8) containing 100µg/mL kanamycin and 400µg/mL carbenicillin, and maintained in Magenta™ pots.

Leaves were harvested from in vitro-grown plantlets, and the standard cetyltrimethylammonium bromide (CTAB) method (Murray and Thompson, 1980) was used for genomic DNA extraction. Putative transgenic plants were screened with an end-point PCR, using the primers CaMV-35S_F2/CasRx-screening_R1 ( Supplementary Table 2 ). Positive lines were propagated until the desired number of plants for that line were achieved.

A. tumefaciens strain EHA105 harbouring the binary pCasRx vectors containing the respective CP gRNAs, named pCasRx : CP_T1, pCasRx : CP_T2, pCasRx : CP_T3 were grown overnight with 50µg/mL kanamycin and 30µg/mL rifampicin. Cultures were centrifuged at 1000 RPM for 15 minutes at room temperature, and resuspended in infiltration buffer (10mM MES [pH 5.6], 10mM MgCl₂, 200µM acetosyringone) to a final OD₆₀₀ of 0.8-1.0. The resuspended cultures were incubated in the dark at ambient room temperature for 2 hours. Mixed cultures at a 1:1 ratio of the pCasRx constructs and the GVA infectious clone were infiltrated in three fully expanded four-week-old wild-type N. benthamiana leaves grown under long-day conditions (16-hour light, 8-hour dark at 25°C), using a needleless 2mL syringe, for transient-expression experiments ( Figure 1 left).

Cultures of A. tumefaciens strain GV3101 containing the unmodified pTRV1 (pYL192, Addgene plasmid #148968) vector, and the modified pTRV2 vectors harbouring the CP gRNAs (pTRV2:CP_T1, pTRV2:CP_T2, pTRV2:CP_T3) were grown overnight, containing 50µg/mL kanamycin, 10µg/mL rifampicin and 30µg/mL gentamicin. Cultures were centrifuged at 1000 RPM for 15 minutes at room temperature, and resuspended in infiltration buffer to OD₆₀₀ 1. The culture of the GVA clone (OD_{600 =} 0.05) was mixed with the TRV mixture (containing either pTRV2:CP_T1, pTRV2:CP_T2, pTRV2:CP_T3) and infiltrated into three leaves of CasRx-EMPTY plants, for single-gRNA experiments. For multi-gRNA experiments, the desired combinations of pTRV2:CP vectors were mixed in an equal ratio, which was then subsequently mixed in a 1:1 ratio with pTRV-RNA1. The TRV mixture was infiltrated two days before the GVA clone. The mixtures were infiltrated as previously described, into either wild-type N. benthamiana or CasRx-EMPTY plants ( Figure 1 ).

Infiltrated wild-type N. benthamiana leaf material was harvested 5-days post-infiltration (dpi) for transient experiments, or 5-dpi from GVA infiltration for stable experiments (7-dpi from TRV mixture infiltration), and three infiltrated leaves per plant were pooled together and subsequently frozen in liquid nitrogen. Total RNA was extracted from 100mg aliquots using the Spectrum™ Plant Total RNA Kit (Sigma, USA) following the manufacturer’s instructions. First-strand complementary DNA (cDNA) synthesis was performed using 1µg of DNase-treated total RNA with Random Primers (Promega, USA) and Maxima Reverse Transcriptase (Thermo Fisher Scientific, USA), following the manufacturer’s instructions. cDNA synthesis was confirmed with an end-point PCR by amplifying the housekeeping gene actin.

A QuantStudio™ 3 Real-Time PCR System (Applied Biosystems, USA), along with the reporter PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, USA), were used for expression analysis. The housekeeping gene adenine phosphoribosyltransferase like (APR) (Liu et al., 2012) was chosen as the internal control ( Supplementary Table 3 ), and the desired RT-qPCR primer sets for gene expression analysis are listed in Supplementary Table 3 . The comparative C_T (ΔΔ C_T) method was selected and the cycle parameters were set for Standard cycling mode (Primer T_m≥ 60°C): 50°C/2min + 95°C/2min + 40x (95°C/15sec + 60°C/1min), and the melt curve parameters were set as the default settings. The 2^-ΔΔCT method was used to calculate the gene expression levels, relative to the reference sample, in the online Design and Analysis app (Thermo Fisher Scientific, USA). The error bars serve as an indication between the RQ minimum and RQ maximum for the sample. A confidence interval of 95% was used.

The relative gene expression of GVA was analyzed statistically using a two-tailed unpaired Student’s t-test. Data represents the mean ± standard error of the mean (SEM), and the significance determination was set at p ≤ 0.05. The Pearson correlation coefficient (r) was used to calculate the correlation between the relative gene expression of the respective gRNA compared to the expression of the GVA CP gene. GraphPad Prism (version 9.1.1, GraphPad Software, Boston, Massachusetts USA) software was used to compute the analyses.

N. benthamiana plants stably expressing CasRx are a useful tool for the testing of different gRNAs. The transformation process and results are illustrated in Figure 2 : The leaf discs used for co-cultivation ( Figure 2A ); developed callus and shoots on a selective medium ( Figure 2B ); the shoots were isolated and cultured on a rooting medium ( Figure 2C ); and rooted plants were maintained on a selection medium ( Figure 2D ). Regenerated plantlets were screened to identify positive lines. Two positive lines from the pCasRx-EMPTY construct stable transformation were confirmed and propagated further.

Open reading frame four (ORF4) of pBINSN_GVA118_NbPD encodes the CP gene, and was chosen as the target for CRISPR/CasRx virus interference, as previous literature had observed decreased viral accumulation when the CP was targeted (Aman et al., 2018a; Aman et al., 2018b). Three gRNAs were designed against the CP gene of GVA, in a region which revealed increased accessibility ( Figure 3A ). The cas13design software output ranks the gRNA in number and quartile, where quartile 4 (Q4) are the best scored and quartile 1 (Q1) are the worst scored gRNAs. Accessibility of the seed region (position 11-18 nt) and single-stranded regions were emphasized during the gRNA selection process (Bandaru et al., 2020), as well as the folding of the gRNA with the scaffold sequence ( Supplementary Figure 2 ). The CP gRNAs selected were: Target 1 (T1) (rank 90, Q3), Target 2 (T2) (rank 3, Q4), and Target 3 (T3) (rank 164, Q3) ( Figure 3A ). The sequences of each target’s seed region were searched against the N. benthamiana transcriptome to identify possible off-targets: T1 contained two off-targets which complemented the seed region, but never the entire gRNA; T2 contained the most off-target possibilities (15) for the seed region, but never the entire gRNA sequence; and T3 contained no off-targets complementing the seed region, nor the rest of the gRNA.

As a preliminary analysis, transient assays were performed to assess the efficacy of the individual CP gRNAs to induce GVA interference. The expression of the CasRx gene was driven by the CaMV 35S promoter, while the gRNA was driven by the hU6 promoter ( Figure 3B ). Wild-type N. benthamiana plants were co-infiltrated with the GVA clone and the different binary vectors pCasRx : CP_T1, pCasRx : CP_T2 and pCasRx : CP_T3 (right side of the leaf). As a control, co-infiltrations of the binary vector containing a non-specific (ns) gRNA (pCasRx-ns), along with the GVA clone, were performed (left side of the leaf). For gRNA CP-T1, virus interference was observed in three samples ( Figure 4 ). Indeed, GVA quantification for sample 1.1, 2.1, and 3.2 were ~2-fold, ~4-fold, and ~1.5-fold reduced, respectively. However, when CP-T2 or CP-T3 were used, no samples showed a decrease in GVA, when compared to the control samples infiltrated with a ns gRNA (pCasRx-ns).

CasRx-expressing plants (CasRx-EMPTY) were used to determine the capability of the CRISPR/CasRx system to confer virus interference against GVA, when gRNAs targeting the CP were delivered with TRV. Subsequent to the co-infiltration of the GVA clone and pTRV2 vector harbouring the respective gRNA, the relative fold expression of GVA and the gRNA were analyzed 5-dpi ( Figure 5A ). As the control, two wild-type N. benthamiana plants were infiltrated with GVA and pTRV2:CP-T1.

As shown in Figure 5A , after infiltration with pTRV2:CP-T1, GVA interference was observed in sample 1, 2 and 5, as there was a significant difference between GVA quantification in the sample compared to the two wild-type N. benthamiana control plants (WT1, WT2). Of the three samples, sample 2 indicated the highest GVA reduction when compared to the control samples (~5-fold reduction, p<0.0001). In Figure 5B , three samples (sample 1, 2 and 5) demonstrated GVA interference when pTRV2:CP-T2 was infiltrated in CasRx-EMPTY, compared to the WT controls. The highest GVA interference was an ~8-fold reduction, observed in sample 1 and 2 (p< 0.0005), when compared to WT1. The relative T2 gRNA expression was negatively correlated with virus accumulation, at the P ≤ 0.05 level. The Pearson correlation coefficient was –0.7738. Lastly, when pTRV2:CP-T3 was infiltrated, samples 1, 4 and 5 ( Figure 5C ) had significantly lower GVA quantification compared to WT1 and WT2, with sample 5 demonstrating a ~1.7-fold reduction in GVA (p=0.0007).

To analyze virus interference with a multi-guide system, as well as GIGS, the GVA infectious clone was infiltrated into CasRx-EMPTY plants, as well as into wild-type N. benthamiana plants, two days after the TRV : CP guides were infiltrated. Figure 6 shows the relative quantification of GVA (see GVA_qPCR CP_F1/GVA_qPCR CP_R1 primers in Supplementary Table 3 ) in both CasRx-EMPTY plants and wild-type plants.

When all three CP gRNAs were infiltrated ( Figure 6A ), all samples showed a reduction in GVA accumulation, with ‘sample 3’ having the lowest GVA accumulation, indeed a >3.5-fold reduction compared to ‘control 1 to 3’ (infiltrated with TRV-EMPTY and GVA) was observed. When the GIGS samples were analyzed for GVA reduction ( Figure 6B ), no clear difference in GVA accumulation was observed in the ‘GIGS’ sample when compared to the control samples which did not receive gRNAs. ( Figure 6B ).

In the next experiment, a two-guide system was compared to a three-guide system. In Figure 6C , the ‘two-gRNA’ CasRx-EMPTY plants group presented a ~3-fold reduction in GVA, when compared to the control group, although the reduction was not significant, while GIGS samples showed no reduction in GVA quantification, compared to the controls. When CasRx-EMPTY plants were infiltrated with all three CP gRNAs (three-gRNA), followed by GVA two days later, there was a ~5-fold reduction in GVA accumulation, when compared to the control group ( Figure 6D ). The GIGS group also showed lower GVA quantification (~>1-fold reduction), compared to the control group. Although a reduction in GVA for both GIGS and ‘three-gRNA’ samples was observed, the reduction was not significant when p ≤ 0.05.

In order to assess the efficiency of a TRV vector at delivering and expressing the CP gRNA transiently and systemically, RT-qPCR analysis (see primers CasRx_gRNA-scaf_F/CP_T2_qPCR_R in Supplementary Table 3 ), compared the relative expression of the T2 gRNA, expressed either under the hU6 promoter from the binary pCasRx vector, or under the PEBV promoter, from the TRV2 vector ( Figure 7A ). The relative fold increase in gRNA expression from the TRV2 vector, was >15-fold, when compared to ‘T-DNA 1’. When the samples were grouped, a ~13-fold increase was observed when gRNA T2 was expressed from TRV2, compared to regular T-DNA ( Figure 7B ).