A tomato bushy stunt virus–based vector for simultaneous editing and sensing to survey the host antiviral RNA silencing machinery

Abstract

A tomato bushy stunt virus (TBSV)–derived vector system was applied for the delivery of CRISPR/Cas9 gene editing materials, to facilitate rapid, transient assays of host–virus interactions involved in the RNA silencing pathway. Toward this, single guide RNAs designed to target key components of the virus-induced host RNA silencing pathway (AGO2, DCL2, HEN1) were inserted into TBSV-based GFP-expressing viral vectors TBSV-GFP (TG) and its P19 defective mutant TGΔP19. This produced rapid, efficient, and specific gene editing in planta. Targeting AGO2, DCL2, or HEN1 partially rescued the lack of GFP accumulation otherwise associated with TGΔP19. Since the rescue phenotypes are normally only observed in the presence of the P19 silencing suppressor, the results support that the DCL2, HEN1, and AGO2 proteins are involved in anti-TBSV RNA silencing. Additionally, we show that knockdown of the RNA silencing machinery increases cargo expression from a nonviral binary Cas9 vector. The TBSV-based gene editing technology described in this study can be adapted for transient heterologous expression, rapid gene function screens, and molecular interaction studies in many plant species considering the wide host range of TBSV. In summary, we demonstrate that a plant virus can be used to establish gene editing while simultaneously serving as an accumulation sensor for successful targeting of its homologous antiviral silencing machinery components.

Introduction

RNA silencing is an evolutionarily conserved mechanism of post-transcriptional gene silencing (PTGS) in eukaryotes and is used to regulate gene expression and combat invasive nucleic acids such as viruses (1, 2). In the event of a viral infection, double-stranded RNAs (dsRNAs) can accumulate in host cells as a result of viral replication or transcription. This activates the virus-induced RNA silencing pathway, which specifically recognizes and degrades viral RNA (2, 3). Upon pathway activation, the dsRNAs are recognized by a Dicer-like (DCL) protein associated with a dsRNA-binding (DRB) protein and cleaved into small interfering RNA (siRNA) duplexes. Cleaved siRNAs are stabilized via methylation by an HUA enhancer (HEN) methyltransferase before incorporation into the RNA-induced silencing complex (RISC). Members of the Argonaute (AGO) protein family form the key catalytic unit of RISC and are programmed by the siRNAs loaded into the complex. These RISC-associated siRNAs are used as search-and-strike modules by the RISC–AGO complex to survey for complementary viral RNA, which is then targeted for degradation.

To combat this antiviral host defense response, plant viruses have coevolved viral suppressors of RNA silencing (VSRs), which function to prevent viral RNA or messenger RNA (mRNA) degradation or suppress RNA silencing through a variety of mechanisms (2, 3). For example, upon infection of host plants with tomato bushy stunt virus (TBSV), 21-nucleotide (nt) siRNAs accumulate, which without any interference, are normally loaded into RISC (4–6). The method of suppression employed by the TBSV VSR protein P19 involves sequestration of the 21-nt siRNAs in a nonsequence-specific manner, ultimately preventing their incorporation into RISC (5, 7). This inhibits the endonucleolytic cleavage of targeted viral RNA by RISC, thereby allowing systemic viral invasion of the host to occur (6–8).

RNA silencing is functional in highly diverse host roles, and as a modular system, can rely on both redundant and flexible machinery, depending on variable circumstances and cellular environments. For example, which of the essential components such as DCLs, RDRs, HENs, and AGOs are predominantly involved in a particular situation remains largely speculative. Various components of RNA silencing have been identified, mostly in the model plant Arabidopsis thaliana. However, cumulative reports indicate that the specific molecular machinery that is utilized can change or overlap based on host type, age, environment, and protein availability, as well as the occurrence and timing of pathogen invasion (2, 9–13). This is particularly relevant for the defense against specific viruses in non-Arabidopsis plant species, which often have several orthologs and/or paralogs of conserved genes as a result of polyploidy.

For instance, while one paralog of an essential component may be required for antiviral silencing, another could be involved in transposon regulation, and another still in epigenetic modification (10, 13–15). While having multiple or partially redundant RNA silencing processes may confer robustness, they often complicate the genetic elucidation of individual pathways and functional machinery. Identifying the essential components and interactions at play in such a highly redundant yet variable molecular pathway can be a daunting and tedious task, even with many of the currently available molecular tools. Thus far, it has been established that most studied plants possess multiple versions of the proteins that form the core components of the RNA silencing pathway, including four DCL proteins. DCL1–4 appear to be involved in RNA silencing to various extents, depending on the virus in question (16–19). For example, during infection by RNA-based potyviruses, DCL2 and DCL4 proteins are required for antiviral silencing. However, during infection with DNA-based geminiviruses, DCL1–4 are all employed for antiviral RNA silencing (18, 19). The Nicotiana benthamiana genome encodes ten known AGO genes, including NbAGO1a, NbAGO1b, NbAGO2, NbAGO4a, NbAGO4b, NbAGO5, NbAGO6, NbAGO7, NbAGO10a, and NbAGO10b. While all may contribute to optimal defense responses, NbAGO1 and NbAGO2 are recognized as the substantial defenders involved in antiviral silencing (19–21). In N. benthamiana, AGO2 functions as the primary siRNA loading unit of RISC during silencing against certain viruses, including TBSV (21, 22). Five HEN gene products have been identified from genetic screens in Arabidopsis and appear to have highly diverse roles surrounding interactions with small or short RNAs, including both siRNA and microRNAs (20–23). Despite most of the major players in the RNA silencing pathway being modular, HEN1 appears to be the predominate protein functioning at the methylation step. While the HEN1 protein is capable of methylating 21–24-nt siRNAs in vitro, it preferentially and more efficiently methylates 23-nt siRNAs that are involved in epigenetic modification, rather than in viral defense. Additionally, in Arabidopsis HEN1 mutants, residual siRNA and PTGS activity can still be detected (24). Overall, it seems likely that depending on the situation, the absence or activity of many silencing components may be supplemented with other homologs, or possibly in other hosts, may be taken over by other components. Details such as which RNA silencing components are involved in direct and indirect viral defense, including during recognition and systemic spread of antiviral signals, remain largely unknown, especially in non-Arabidopsis plants.

Due to the limited number of viruses capable of infecting Arabidopsis, virologists have traditionally employed N. benthamiana as the predominate model to explore plant-virus interactions. However, due to the comparatively limited genetic resources available for N. benthamiana, transient gene knockdowns are routinely conducted using tobacco rattle virus (TRV) vectors to perform virus-induced gene silencing (VIGS) of specific host mRNAs (23, 25). While there are many advantages to using VIGS, the further development of tools useful for rapid genetic studies could complement and expand upon the molecular toolbox currently available to plant scientists. Additionally, while studying a process such as RNA silencing that is adaptable to different viruses, it can be advantageous to employ tools that avoid the use of additional viruses other than the virus of interest.

Viral gene vectors represent biotechnological delivery tools capable of rapid and efficient expression of heterologous protein and nucleic acid products in a wide range of plant hosts (23, 26, 27). For instance, virus vector technology enables the study of plant–virus interactions in planta by tracking viral movement in living cells, and the nature of virus replication and spread results in high levels of recombinant product delivery in a short time. Additionally, virus-based vectors are useful for the delivery of gene editing materials, including CRISPR/Cas9 (23). While traditional transgenic methods can be laborious, time consuming, and expensive, viral vectors can be easily delivered via Agrobacterium-mediated infiltration (agroinfiltration).

TBSV-based viral vectors are useful due to the high rate of replication and contaminant gene expression combined with the vast host range, allowing transient recombinant product delivery in numerous species (28–30). Furthermore, the native TBSV protein P19 is a comparatively strong suppressor of virus-induced host RNA silencing, allowing higher accumulation of the viral vector and therefore higher expression of heterologous payloads (7). Given that AGO2 is essential for anti-TBSV silencing in N. benthamiana (9, 21, 31, 32) targeting this gene using a TBSV-based viral vector can serve as an excellent control for evaluating the accuracy of novel gene editing methods.

In this study, we developed an original application of a TBSV-derived vector system for the delivery of CRISPR/Cas9 gene editing materials against RNA silencing pathway components. Toward this, single guide RNAs (gRNAs) for AGO2, DCL2, or HEN1 were inserted into the TBSV-based viral vectors TBSV-GFP (TG) and its P19 defective mutant TGΔP19. The aim was to implement rapid, efficient, and specific gene editing in planta that could be directly and simultaneously visualized by virus accumulation (as inferred by GFP expression) in tissues where this would otherwise be restricted through silencing (Fig. 1). The ultimate goal of this study was to create an efficient platform for rapid genetic screening and functional studies, further expanding and complementing the gene editing toolkit available to researchers (Fig. 2). The TBSV-based tools presented in this study offer a rapid, transient, and easily adaptable platform for the delivery of recombinant materials in high quantities, using technology based on a virus with a wide host range. In addition to circumventing more time-consuming and laborious transgenic methods, these tools allow the ability to easily study the TBSV-induced host RNA silencing pathway without the introduction of erroneous delivery vectors and the ability to target essential proteins that could interfere with normal host function if knocked out completely and permanently. This represents a new approach resulting in a collection of original findings for virus-silencing system interactions in plants.

Results

TBSV-based viral vectors induce rapid editing of AGO2 in planta to rescue TGΔP19

Previously, it was demonstrated that in N. benthamiana, AGO2 encoded the primary AGO protein utilized for antiviral silencing against TBSV infection (21, 32). As a consequence, TGΔP19, which lacks the ability to express the P19 VSR, is unable to replicate in wild-type plants expressing functional levels of AGO2 for antiviral silencing. P19 has been used in heterologous systems to increase recombinant product delivery, including in plant, animal, and prokaryotic systems (33–36). Additionally, P19 supplementation via a nonviral binary vector increases Cas9 expression from the binary Cas9 expression vector pHcoCas9 (Fig. 3C (37). Since P19 functions by interfering with the host RNA silencing system, we surmised that the editing of silencing components in itself could further increase not only GFP but also that of pHcoCas9, thereby providing further evidence that the observed effect is at the level of silencing.

Taking the above into consideration, for the initial development of our TBSV-based gene editing tool, we designed both the TG and TGΔP19 viral vectors to carry CRISPR gRNAs capable of targeting NbAGO2 (TGgAGO2 and TGΔP19gAGO2; Fig. 3). Together with the previously described Cas9 binary delivery vector, pHcoCas9 (Fig. 3C (38)), TGgAGO2 and TGΔP19gago2 were agroinfiltrated into N. benthamiana. The gRNAs targeting AGO2 were designed to target a region of the gene of interest containing the restriction enzyme site for BceAI (Fig. 2). This allowed for the quantification of indel occurrences (29, 38) after incubation of AGO2 PCR amplicons with BceAI. Prior to tissue collection for indel analysis, the leaves were visualized under ultraviolet (UV) light for the observation of GFP reporter gene expression from TGΔP19 (not expressing P19) and TG (expressing P19), allowing virus replication and movement to be tracked. Restriction enzyme–resistant PCR products were present for each of the TGΔP19gAGO2 and TGgAGO2 delivery treatments when combined with pHcoCas9, indicating indel accumulation, while negative control treatments were completely digested (Fig. 4).

Quantification of the restriction enzyme–resistant amplicons permitted the percentage of the successfully edited DNA sample compared to the fully digested nonedited AGO2 DNA sample to be calculated. As early as 3 days post infiltration (dpi), pHcoCas9 coinfiltrated with either TGΔP19gAGO2 or TGgAGO2 produced indel frequencies of 3 and 27%, respectively (Fig. 4). At 7 dpi, indel frequencies exhibited a drastic increase, yielding 13% in tissue coinfiltrated with pHcoCas9 and TGΔP19gAGO2, and 44% in tissue coinfiltrated with pHcoCas9 and TGgAGO2.

Leaves exhibited strong green fluorescence beginning at 3 dpi when coinfiltrated with the controls TG (expressing P19) and pHcoCas9 (Fig. 5). However, green fluorescence was significantly reduced for TGΔP19 and pHcoCas9. Upon coinfiltration of either TGΔP19gAGO2 and pHcoCas9 or TGgAGO2 and pHcoCas9, both gRNA-expressing viral vectors were capable of replication, as was expected for TGgAGO2 (Fig. 5B). However, most notably, compared with the control in Fig. 5A, TGΔP19gAGO2 also exhibited increased green fluorescence starting at 3 dpi, which was prolonged and intensified throughout the time course until 7 dpi (Fig. 5B). Therefore, targeted editing of AGO2 using TGΔP19gAGO2 restored the ability of the virus to accumulate and express its heterologous payload. In addition to supporting our previous findings that NbAGO2 encodes the predominant anti-TBSV AGO protein (21, 32) these results provide proof-of-concept evidence that our TBSV-based system can be used as a means to rapidly screen proteins involved in RNA silencing.

Effects of DCL2 targeting on virus accumulation

The recruitment of silencing components involved can change based on the plant species and infecting virus, and there are at least four DCLs in N. benthamiana. To start probing some of the other host RNA silencing pathway components, we targeted several for editing as above for AGO2.

The presence of indels, measured by either restriction enzyme digest or site-specific sequencing, indicated successful targeted gene editing of DCL2 in all experimental samples derived from tissue infiltrated with pHcoCas9 and either TGΔP19gRNA or TGgRNA. Combined with pHcoCas9, TGΔP19gDCL2 produced 25.4% indels and TGgDCL2 produced 40.3% indels (Fig. 6). The percentages are not stunning and may be partially responsible for the observation that the visual effect of targeting DCL2 is subtle (Fig. 7) but noticeable at 5–7 dpi with the appearance of a slight but uniform green fluorescence for pHcoCas9 + TGΔP19gDCL2 (as discussed later). Importantly, similar experiments with other DCL homologs did not reveal accumulation-restorative effects, suggesting a prominent role for DCL2 against TBSV.

Effects of HEN1 targeting on virus accumulation

Western blot using anti-GFP antibody was performed to visualize GFP accumulation in plants after the delivery of TGΔP19gHEN1 and TGgHEN1 (Fig. 8A). Upon coinfiltration with pHcoCas9 and TG, GFP showed strong expression, which was slightly increased by the addition of gHEN1. Upon coinfiltration with pHcoCas9 and TGΔP19, GFP expression was visible and notably increased by the addition of gHEN1. Likewise, compared with leaves infiltrated with pHcoCas9 alone, leaves coinfiltrated with pHcoCas9 and TG (expressing P19) contained increased levels of Cas9 protein (Fig. 8B). Furthermore, in contrast to leaves infiltrated with pHcoCas9 alone, leaves infiltrated with pHcoCas9 and TGΔP19 (not expressing P19) exhibited decreased Cas9 protein. Comparatively, in leaves coinfiltrated with pHcoCas9 and TGΔP19gHEN1 (for which successful HEN1 editing was shown via sequencing, Table 1), the levels of Cas9 protein increased. In conclusion, the effects we observed can be explained by interferences with the silencing pathway.

The visual effects of targeting HEN1 on GFP expression (i.e. virus accumulation) are shown in Fig. 9. Again, the results show that a diffuse green fluorescence appears (as in Fig. 7 for DCL2), illustrating a rescue effect of HEN1 editing in the absence of P19. This rather faint expression also suggests that the untargeted HEN1 genes in many cells in the infiltrated leaves are probably sufficiently expressed to still result in a fair overall silencing response. Site-specific sequencing confirmed the successful targeted editing of HEN1 by both TGgHEN1 and TGΔP19gHEN1 (Table 1).

Together, the results support the notion that disruption of the silencing pathway with either the presence of P19 or by the editing of AGO2, DCL2, or HEN1 (Fig. 1) can interfere with defenses and lead to increased viral accumulation and expression.

Discussion

Silencing and editing

While RNA silencing has been the focus of extensive research, many facets of its induction by different plant pathogens have yet to be elucidated. As intracellular parasites that encode few genes, viruses interact on an intimate molecular level with the host cell throughout their lifecycle. Additionally, host components have been demonstrated to possess multiple functions that can vary between species, such as in the case of AGO in Arabidopsis compared with N. benthamiana. Furthermore, the majority of RNA silencing studies have been conducted in Arabidopsis, a diploid with a characteristically small genome that is genetically different from N. benthamiana, a common model used by virologists, and from most other crop species (39). While RNA silencing is the most well-known and substantial “nonprotein-based” antiviral defense, significant gaps in knowledge exist regarding the specific components of different viruses that interact in unique and sometimes multiple locations throughout the pathway, their impact on the systemic spread of RNA silencing signals, the cellular locations of these interactions, and alternative functions of both viral components and host RNA silencing machinery.

The usefulness of Arabidopsis as a model when studying plant–virus interactions is limited due to the small number of viruses capable of effectively infecting this species (40, 41). Alternatively, virologists have traditionally employed N. benthamiana as a model. However, due to the comparatively limited genetic resources available for N. benthamiana such as gene knockout libraries and fully annotated genome assemblies, transient gene knockdowns are routinely performed using TRV vectors to perform VIGS of specific host mRNAs (23, 25). There are many advantages to using VIGS, and it represents one of the most powerful tools currently available to plant scientists in their molecular toolbox. However, the present study illustrates a complementary concept tool capable of rapid, robust, and transient gene editing that can be adapted for use in multiple plant species and is easily reprogrammable for different targets.

The TBSV-based editing system

We hypothesized that our TBSV-based biotechnological tools could be implemented to rapidly screen for key components of the virus-induced RNA silencing pathway specifically involved in TBSV defense (Fig. 1). Our experimental goals involved implementing TBSV vectors capable of gRNA delivery for CRISPR/Cas9 gene editing. The aim was to establish the system as a rapid, transient screening method for virus–host interactions by first determining if it could be used to verify the importance of NbAGO2, then subsequently to identify the NbDCL component and examine properties of NbHEN1.

Toward this, we explored the potential of the TBSV-based delivery vectors TGΔP19 and TG to deliver functional gRNAs as fusions with the coding region for GFP expressed through the action of the coat protein (CP) subgenomic promotor (Fig. 3). The results showed that efficient editing occurred after just 7 days in every targeted gene, evaluated using a restriction enzyme digest assay or site-specific sequencing. At this juncture, it also serves to emphasize that in the absence of processing elements (i.e. ribozymes), the modification of TBSV-expressed pre-gRNAs is performed by endogenous enzymes resulting in an active gRNA for Cas9 recruitment, which we previously showed for TMV (38). This leads further credence to the notion that plants have, possibly coincidentally, acquired an ability to properly process gRNAs.

In essence, the collective results illustrate the first use of TBSV-based viral vectors for CRISPR/Cas9-induced gene editing. Additionally, we present these tools for use as a rapid in vivo screening assay for molecular host–virus interactions in the RNA silencing pathway. The TBSV-based expression vectors TG and TGΔP19 are TBSV CP substitution mutants that express GFP (21, 29). While TG expresses a wild-type P19 VSR, P19 expression was eliminated from TGΔP19 via the introduction of two premature stop codons. Assisted by the native P19 protein, TG yields rapid and high levels of recombinant product delivery and GFP, while a lack of functional P19 from TGΔP19 works as a sensor to visualize the effects of the virus-induced RNA silencing response.

Characterizing the anti-TBSV silencing system

Using TGΔP19 and TG expressing single gRNAs designed to target DCL2, HEN1, and AGO2, we probed the virus-induced RNA silencing pathway for host molecular machinery functional in the event of cellular infection by TBSV. Successful targeted gene editing was observed by both TGΔP19 and TG carrying gRNAs when codelivered with Cas9 nuclease. An important note is that the P19-mediated suppression of silencing did not interfere with gene editing, which affirms that silencing and editing are two entirely separate molecular events despite their commonality in using short RNAs.

After editing of DCL2, HEN1, and AGO2, the TGΔP19 mutant was capable of stronger initial and sustained accumulation, despite the absence of a functional P19 VSR. These P19-mimicking effects confirm previous findings that NbAGO2 is involved in TBSV-induced RNA silencing (32) and for the first time, strongly suggest that NbDCL2 and NbHEN1 are essential components. Furthermore, the successful gene editing of HEN1 resulted in higher Cas9 protein expression independent of P19, supporting a decrease in functionality of the RNA silencing pathway against foreign nucleic acid invasion.

The rescue of GFP expression is incomplete as indicated by a relatively weak GFP expression associated with TGΔP19. This is not surprising when considering that (i) editing percentages are often less than 50%, (ii) the silencing pathway is fully active upon entry of the virus suggesting that it has to overcome this initial hindrance, and (iii) partial redundance for silencing components is known to exist. Nevertheless, this is the first time that a plant virus has been used as an editing tool to probe its homologous antiviral silencing defense pathway.

While this research showed the connection among DCL2, HEN1, AGO2, and TBSV defense in N. benthamiana, the homologs of DCL, HEN, and AGO, as well as other functional components of RNA silencing, such as DRBs and RNA-dependent RNA polymerases, require similar scrutiny to determine their role in TBSV defense, as we performed preliminarily with VIGS (31). The wide host range of TBSV makes it a prime candidate for direct studies in nonmodel plants, including economically significant crops.

Despite the proposed utility, virus-based targeting of silencing components should be viewed within the context that many components are redundant and possess multiple roles. Therefore, one can surmise that while DCL2, HEN1, and AGO2 are functional in TBSV defense in N. benthamiana, upon editing, their absence may be supplemented with other homologs, or in other hosts, their roles may be taken over by other components. For example, in regard to HEN1, this notion may be supported by the fact that residual siRNA and PTGS activity can still be detected in Arabidopsis HEN1 mutants (24). Considering the redundancy and flexibility of RNA silencing machinery, this may suggest the putative existence of components capable of aiding or compensating for HEN1, such as other cytosolic methyltransferases. Therefore, although our results demonstrate and confirm the contribution of certain silencing components, their established roles are not yet absolute.

The tools presented in this study offer multiple complements to the current technology available in the gene editing toolbox. These include: (i) rapid, transient delivery of recombinant materials in high quantities, driven by vector technology based on TBSV, a virus with a wide host range, (ii) transient gene editing useful for targeting essential host components that could interfere with normal host function if knocked out completely and permanently, (iii) the ability to circumvent laborious and difficult traditional transgenic methods, and (iv) the ability to study the TBSV-induced host RNA silencing pathway without the introduction of erroneous delivery vectors, such as the viral vectors used in VIGS.

Conclusion

This report represents an original example of a study where a plant virus that either possesses or lacks its native VSR has been used to deliver gRNAs for gene editing, allowing it to function as a sensor for probing its homologous antiviral RNA silencing mechanism. The present work is based on TBSV, which given its wide host range may have far-reaching implications for plant biology in general. It is therefore tempting to speculate that the system presented in this study may function as an exemplar for the development of tools that can be rapidly employed for functional genomics in a variety of different plant species, including those that are recalcitrant to traditional or current techniques. Regarding new information in the context of molecular antiviral silencing, this study suggests that even though visual effects appear camouflaged, presumably due to gene redundancy, the results affirm a role for AGO2 and point to a newly uncovered involvement of DCL2 and HEN1 in anti-TBSV silencing in plants.

Methods

Construct design

The gRNA targeting sequences were determined for DCL2, HEN1, and AGO2 in Arabidopsis and N. benthamiana using standard gene sequence alignment, and the CRISPR design toolset (benchling.com) was used to design full gRNAs based on the Streptococcus pyogenes CRISPR/Cas9 gRNA scaffold sequence, as previously described (25, 37, 38). Q5 polymerase mutagenesis was used to insert gDCL2, gHEN1, and gAGO2 into the construction vectors, and heat shock was used to transform the vectors into the Escherichia coli competent cell line 5α (New England BioLabs). After antibiotic selection, correct vector insertion was confirmed via colony PCR and sequencing. Plasmid DNA was isolated from colonies containing the correct insert and used for Gibson Assembly cloning into TBSV TG and TGΔP19 viral vectors. Heat shock was used to transform the viral vectors into the E. coli competent cell line 5α. After antibiotic selection, colony PCR and sequencing were used to confirm the correct insert size. Plasmid DNA was isolated and used for transformation into the Agrobacterium tumefaciens strain pGV3101 via electroporation.

The binary pHcoCas9 plasmid (addgene plasmid: 42230) expresses a human codon-optimized (Hco) Cas9 protein. Expression in plants was mediated in the pBinPLUS-sel plasmid with a CaMV 35S dual promotor and the tobacco etch virus (TEV) translational enhancer region driving Cas9 gene expression as previously described (42). TBSV vectors TGΔP19 and TG have been previously described (29).

Agroinfiltration of plants

A. tumefaciens cultures harboring the constructs were grown in liquid Luria Broth (LB) media containing 50 mg/L of kanamycin and incubated overnight at 28°C and shaking at 250 rpm. Bacterial cells were harvested by centrifugation at 3,900 × g for 20 min at room temperature, resuspended in infiltration buffer (IB; 10 mM MgCl₂, 10 mM MES pH 5.6, and 200 μM acetosyringone), and incubated at room temperature for 1 h. Cultures were adjusted to a final concentration of OD600 0.5 with IB. The first three true leaves of 4-week-old N. benthamiana plants were infiltrated using a needleless syringe with IB alone (as a negative control), pHcoCas9 alone, coinfiltrated with pHcoCas9 and TGΔP19, coinfiltrated with pHcoCas9 and TG, coinfiltrated with pHcoCas9 and TGΔP19gRNA, or coinfiltrated with pHcoCas9 and TGgRNA. Infiltrated plants were maintained on a light shelf under long-day conditions (16/8 h light/dark cycle). Three to four biological replicates were performed for each experiment, each with three technical replicates (three individually infiltrated leaves) per plant.

GFP visualization and indel analysis

Infiltrated plants were visualized under a UVP Blak-Ray B-100A handheld UV mercury lamp. Leaf tissues were collected and pooled from three infiltrated leaf replicates, totaling 100mg from each plant, and stored at −80°C until processing. Whole genome DNA was extracted using the GeneCatch Plant Genomic DNA Miniprep Kit (Epoch) according to the manufacturer's instructions. The region of interest, including the gene targeted by the gRNA, was PCR amplified. For tissue where DCL2 or AGO2 was targeted, the PCR product was cleaned using the DNA Clean & Concentrator -5 kit (Zymo Research) and resuspended in DNase-free and RNase-free water. DNA targeted for DCL2 was subjected to BspHI restriction enzyme digestion, and DNA targeted for AGO2 was subjected to BceAI restriction enzyme digestion then separated on an agarose gel until bands were sufficiently resolved. The gDCL2 and gAGO2 were designed to target a region of their respective genes containing the restriction enzyme sites for BspHI and BceAI, respectively. If editing in the region of interest has occurred, the restriction enzyme site would be disrupted by the presence of indels induced by the gRNA/Cas9 complex, and the restriction enzyme would be incapable of cleavage. In contrast, samples that have not been gene edited would exhibit complete digestion. The quantification of gene-specific indel percentages was accomplished using Image J analysis software (NCBI). After PCR amplification of the region of interest, DNA targeted for HEN1 was visualized on an ethidium bromide–stained agarose gel. Bands were gel extracted using the Zymoclean Gel DNA Recovery kit (Zymo Research) and cloned into the pGEM-T Easy vector system (Promega) for sequencing.

Protein extraction and western blot analysis

From each experimental plant, 50 mg of infiltrated leaf tissue was collected. Proteins were extracted in 500 μL of 5× cracking buffer (645 mM Tris pH 6.8, 10% w/v SDS, 715 mM β-mercaptoethanol, 40% v/v glycerol, and 0.0005% w/v bromophenol blue), boiled for 5 min, then spun down at 10,000 × g for 2 min. 20 μL of the supernatant was collected for visualization on a 7.5% polyacrylamide-SDS gel. Gels were initially run at 80 V for 20 min, then at 150 V for 80 min in 1× Laemmli running buffer (25 mM Tris, 192 mM glycine, and 0.1% w/v SDS). The separated proteins were transferred to a nitrocellulose membrane (Bio-Rad, CA, USA) in Tris-glycine transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, and pH 7.0) at 270 mAmp for 90 min. The transferred proteins were blocked in TBST buffer (0.2 M NaCl, 50 mM Tris, 0.05% v/v Tween 20, and pH 7.4) containing 5% nonfat milk for 1 h, followed by overnight incubation with mouse IgG anti-CRISPR (Cas9) primary antibody (Biolegend) at 1:5,000 dilution or primary antibodies for GFP (43) at 1:10,000 dilution at 4°C. After incubation, membranes were subjected to three 5 min washes with 1× TBS buffer. A secondary IgG anti-mouse antibody conjugated to alkaline phosphatase (Sigma) was added to the membranes at 1:10,000 dilution and incubated for 1 h at room temperature, and the membrane was rewashed three times.

Colorimetric detection of Cas9 protein was achieved by submerging the membrane in a solution of 33 μL of 5-bromo-4-chloro-3-indolyl phosphate (100 mg/mL), 66 μL of nitro blue tetrazolium (20 mg/mL) to 10 mL of 1× alkaline phosphatase buffer (100 mM Tris, pH 9.5, 1 M NaCl, and 0.5 M MgCl₂). The presence of the Cas9 protein was identified by the characteristic molecular mass of 164 kDa compared with a protein ladder (Thermo Scientific PAGE Ladder Plus Prestained Protein Ladder).

To confirm the equal loading of proteins for each sample, Ponceau-S or Coomassie staining was performed. For staining with Ponceau-S, the nitrocellulose membrane was washed three times in diH₂O for 5 min after transfer. The membrane was submerged in Ponceau-S stain (0.5% w/v Ponceau-S, 1% v/v acetic acid) until bands appeared. The stain was removed with multiple rounds of washing in diH₂O. The rapid staining of an identically loaded and run polyacrylamide-SDS gel in Coomassie Brilliant Blue R-250 was achieved by microwaving the gel twice in water for 2 min, followed by microwaving in stain solution (40% methanol, 10% acetic acid, 50% water, and 0.1% w/v Coomassie Brilliant Blue R-250). Gels were incubated with gentle swirling at room temperature for 2 min. Gels were submerged in destaining solution (40% methanol, 10% acetic acid, and 50% water) with gentle swirling until the 55-kDa RuBisCo band could be clearly observed.

Acknowledgments

The authors thank Will Cody and Kelvin Chiong for input and assistance throughout the project and K.-B. G. Scholthof for helpful critical comments and insightful discussions throughout the study or during the manuscript preparation. The authors also thank Holly Meier and Andrew Lee for various experimental or technical contributions.

Funding

Funding was received from Agriculture and Food Research Initiative (AFRI) awards 2015-67013-22916 and Hatch-1016098, both from the United States Department of Agriculture (USDA) National Institute of Food and Agriculture (NIFA).

Author Contributions

A.D., M.R.M., and H.B.S. together conceived the original project and research plans. A.D. and M.R.M. designed and conducted the experiments, while A.D. also analyzed the data and wrote the manuscript. H.B.S. supervised the experiments and assisted with editing and agreed to serve as the author responsible for contact and communication ([email protected]).

Data Availability

All the data in the manuscript are complete, comprehensive, and original. No other or additional repositories exist, and we will share any information requested.

References

Author notes