Knockout of SlDCL2b attenuates the resistance of tomato to potato spindle tuber viroid infection

Abstract

RNA interference, or RNA silencing, is an important defence mechanism against viroid infection in plants. Plants encode multiple DICER‐LIKE (DCL) proteins that are key components of the RNA silencing pathway. However, the roles of different DCLs in defence responses against viroid infection remain unclear. Here, we determined the function of tomato DCL2b (SlDCL2b) in defence responses against potato spindle tuber viroid (PSTVd) infection using SlDCL2b loss‐of‐function tomato mutant plants. Compared with wild‐type plants, mutant plants were more susceptible to PSTVd infection, developing more severe symptoms earlier and accumulating higher levels of PSTVd RNAs. Moreover, we verified the feedback mechanism for the regulation of SlDCL2b expression by miR6026. Functional blocking of tomato miR6026, by expressing its target mimics, can enhance resistance to PSTVd infection in tomato plants. These findings deepen the current understanding of RNAi‐based resistance against viroid infection and provide a potentially new strategy for viroid control.

Tomato DCL2b (SlDCL2b) has a defence function against potato spindle tuber viroid (PSTVd) infection in tomato plants.

RNA interference (RNAi) is a powerful molecular mechanism that acts to regulate gene expression in a sequence‐specific manner (Ghildiyal & Zamore, 2009). It is activated when double‐stranded RNA (dsRNA) or highly structured single‐stranded RNA molecules are cleaved into small (21–25 nucleotide [nt]) duplexed RNA molecules, referred to as microRNAs (miRNA) and small interfering RNAs (siRNA), by RNase III isozymes termed DICER‐LIKE (DCL) ribonucleases. After cleavage, one strand of the small duplexed RNAs is loaded into an ARGONAUTE (AGO) protein—a key component of the RNA‐induced silencing complex (RISC), activating it. The activated RISC is then guided to a complementary target RNA molecule that is then subsequently degraded by the RISC, preventing its ability to be expressed by being translated into a functional protein (Ghildiyal & Zamore, 2009; Lopez‐Gomollon & Baulcombe, 2022).

In plants, RNAi not only serves as a means to regulate endogenous gene expression but can also serve as an effective defence mechanism against viral infection (Guo et al., 2019; Lopez‐Gomollon & Baulcombe, 2022). The key enzymes and their functions involved in antiviral defences in plants have been identified for both RNA and DNA viruses, mainly in the model plant Arabidopsis thaliana (Guo et al., 2019; Yang & Li, 2018). These proteins include DCL2–4, AGO1, AGO4, AGO5 and RNA‐dependent RNA polymerases (RDRs) RDR1, RDR2 and RDR6. RNAi can likewise defend plants against viroid infection (Di Serio et al., 2023). However, the functions of key enzymes may behave differently in response to viroid infection than they would in response to virus infection because viroids, unlike plant viruses, do not encode any proteins (Katsarou et al., 2022) and therefore do not produce suppressors of RNA silencing (VSRs) (Csorba et al., 2015) to counteract the plant RNAi response.

Higher plants encode multiple DCLs that function variably in diverse RNA silencing pathways. A. thaliana encodes four DCLs, DCL1–4 (Fukudome & Fukuhara, 2017). Although all of these are involved in the biogenesis of viral siRNAs, only DCL2 and DCL4 are known to function primarily in antiviral defence in a hierarchical or redundant way (Andika et al., 2015; Blevins et al., 2006, 2011; Bouche et al., 2006; Deleris et al., 2006; Fusaro et al., 2006; Garcia‐Ruiz et al., 2010; Xie et al., 2004). DCL4 is the main protein found in the canonical 21‐nt siRNAs associated with potent antiviral activity. When DCL4 is absent or inhibited by VSRs, DCL2 produces 22‐nt viral siRNAs to compensate for this loss (Bouche et al., 2006; Deleris et al., 2006). It should be noted that DCL2 functions not only as a backup to DCL4 but also appears to function downstream of DCL4 to amplify the RNAi silencing signal (Zhang et al., 2015).

DCL2 and DCL4 also function in the defence against viroid infection, but in a somewhat different manner. In Nicotiana benthamiana, instead of restricting potato spindle tuber viroid (PSTVd) accumulation, DCL4 appears to stimulate it because the down‐regulation of DCL4 reduced PSTVd accumulation levels (Dadami et al., 2013). Simultaneous down‐regulation of DCL2 and DCL4 had a similar result. By contrast, simultaneous down‐regulation of DCL2 and DCL3 increased PSTVd accumulation levels, indicating combined activity of DCL2 and DCL3 proteins in the defence against viroid infection (Katsarou et al., 2016). In tomato (Solanum lycopersicum), simultaneous down‐regulation of DCL2 and DCL4 increased PSTVd accumulation levels in the early stages of infection (Suzuki et al., 2019), rather than increasing observed in N. benthamiana (Dadami et al., 2013). This difference is probably related to the different experimental hosts used in different studies. An alternative explanation may be the difference in the important gene in the RNA silencing pathway, RNA‐dependent RNA polymerase 1 (RDR1), between N. benthamiana and tomato. RDR1 is normally expressed and functional in tomatoes but not in N. benthamiana (Yang et al., 2004). Together, the required DCLs and their roles in plant defences against viroid infection remain unclear. Here, we used SlDCL2b loss‐of‐function mutant (dcl2b) tomato plants (cv. Ailsa Craig) (T. Wang et al., 2018) to determine the function of SlDCL2b in the viroid defence response. We discovered a feedback mechanism controlled by SlDCL2b‐dependent miR6026 that acts to regulate SlDCL2b expression.

In tomato, the SlDCL2 gene family consists of SlDCL2a, SlDCL2b, SlDCL2c and SlCL2d (Wang et al., 2016), among which SlDCL2a and SlDCL2b are the most abundantly expressed genes (Z. M. Wang et al., 2018). The tomato mutant dcl2b was generated using the CRISPR/Cas9 gene‐editing system without observable developmental defects (T. Wang et al., 2018). Compared with wild‐type (WT) tomato plants, dcl2b plants were more susceptible to viral infection and showed more severe symptom development, despite SlDCL4 remaining active. Similarly, compared with WT tomato plants of another cultivar (M82), the SlDCL2b loss‐of‐function mutant plants were more susceptible to infection by both tobacco mosaic virus (TMV) and potato virus X (PVX) (Z. M. Wang et al., 2018). Thus, SlDCL2b appears to play a critical role in antiviral defences in tomato.

To verify whether SlDCL2b also functions in the defence against viroid infection, mutant tomato (cv. Ailsa Craig) (dcl2b) plants, kindly provided by Professor Hongliang Zhu at China Agricultural University, and WT plants were mechanically inoculated by in‐vitro RNA transcripts of dimeric PSTVd (MK303581) cDNA and infiltrated by Agrobacterium tumefaciens GV1301 transformed with the binary plasmid pCAM1300 carrying dimeric PSTVd cDNA. Six to 10 plants were used for each treatment. Inoculated plants were grown in an insect‐free climate‐controlled chamber kept at 25–28°C under a 16:8 h light/dark cycle. The experiment was repeated three times.

Mutant (dcl2b) plants expressed stunting and leaf curling earlier than WT plants and these symptoms were more severe in dcl2b plants (Figure 1a,b). The average height (n = 10) of dcl2b plants was reduced by about 30%. These results showed that functional loss of SlDCL2b increases tomato susceptibility to PSTVd infection. This observation is partially in line with the findings that down‐regulation of both SlDCL2 and SlDCL4 changes the response of tomato cv. Moneymaker to PSTVd infection from tolerance to lethal systemic necrosis (Suzuki et al., 2019). The attenuated resistance to PSTVd infection in dcl2b plants was correlated with higher accumulation levels, based on the northern blotting analysis of PSTVd at 14, 21 and 28 days post‐inoculation (dpi) (Figure 1c).

FIGURE 1.

Loss‐of‐function mutation of the SlDCL2b gene in tomato attenuates the RNAi defence response against potato spindle tuber viroid (PSTVd) infection. (a). PSTVd‐agroinfiltrated wild‐type (WT) and mutant (dcl2b) tomato (cv. Alisa Craig) plants at 14, 21 and 28 days post‐inoculation (dpi). (b) The top of PSTVd‐agroinfiltrated WT and dcl2b plants at 21 dpi. (c) Northern blotting analysis of PSTVd for WT and dcl2b plants at 14, 21 and 28 dpi.

Because sRNAs help guide the RISC to degrade RNAs, the correlation between PSTVd RNA and PSTVd‐derived sRNA accumulation levels was identified by sRNA sequencing of mechanically inoculated WT and dcl2b plants at 21 dpi. The uppermost, young, fully opened leaves were collected for total RNA extraction using TRIzol (Transgen) according to the manufacturer's instructions, followed by sRNA separation, library construction and sequencing (Li et al., 2021). The sequencing was performed on an Illumina Hiseq 2500 platform (Lc‐Bio Technologies; Hangzhou, China) to obtain 50 bp single‐end reads. Two replicates were analysed for each treatment.

Bowtie 2 (Langmead & Salzberg, 2012) was used to search for 21–24 nt sRNAs that perfectly matched with the PSTVd genome (MK303581), yielding 1.4–2.7 million PSTVd‐sRNAs for these samples (Table S1). The amount of PSTVd‐sRNAs was lower in dcl2b plants than in WT plants, mainly due to the marked decrease of 22 nt species (Figure 2a). Unlike the observations in PSTVd‐infected N. benthamiana (Dadami et al., 2013) and Moneymaker tomato plants (Suzuki et al., 2019), the amount of 24 nt species seemed quite low. Different plants or varieties used in different experiments could be the main reason. Certain sequencing biases cannot be completely excluded because of the much lower amount of 24 nt species than 21 and 22 nt species. The amount of 22 nt species in dcl2b plants decreased by more than 60%, further confirming the major role of SlDCL2b in the biogenesis of this species (T. Wang et al., 2018). Moreover, the amount of 21 nt species in dcl2b plants increased by nearly 30%. This did not result from increased expression levels of SlDCL4 as the functional loss of SlDCL2b had no effect on SlDCL4 expression (T. Wang et al., 2018); this likewise did not result from changes in the cleavage function SlDCL4 as the distribution profiles of 21 nt sRNAs along the PSTVd genome were almost the same in both the mutant and WT plants (Figures 2b and S1). Thus, the increase in the amount of 21 nt PSTVd‐sRNAs in dcl2b plants could probably be explained by functional compensation by SlDCL4 in response to SlDCL2b inactivation. Together, accumulation levels of PSTVd RNAs were positively related to those of 21 nt PSTVd‐sRNAs but negatively related to those of 22 nt species, suggesting that 22 nt PSTVd‐sRNAs might play important roles in viroid resistance.

FIGURE 2.

Loss‐of‐function mutation of the DCL2b gene affects the biogenesis of PSTVd‐sRNAs. (a) The amount (reads per million) of total PSTVd‐sRNAs and 21–24 nucleotide (nt) PSTVd‐sRNAs in the mutant (dcl2b) and wild‐type (WT) tomato plants. (b) Distribution profiles of PSTVd‐sRNAs against the PSTVd genome.

Given that SlDCL2 expression in tomato is regulated by a feedback loop involving SlDCL2‐dependent miR6026 (Z. M. Wang et al., 2018), we further assessed the presence of this miRNA using the sRNA sequencing data from PSTVd‐infected dcl2b and WT tomato plants. Known miRNAs were identified using the modified software miRdeep2 (Friedlander et al., 2012) based on the miRbase 20.0 database. Expression levels of miRNAs were estimated using RPM (reads per million) as described by Zhou et al. (2010) and differential expression analysis was performed using the DESeq2 R package (v. 1.41.1) with the threshold of a corrected p‐value of 0.05. Compared with PSTVd‐infected WT plants, PSTVd‐infected dcl2b plants had nine differently expressed miRNAs, four of which were down‐regulated and five were up‐regulated (Figure 3a, Table S3). While similar miR6026 expression levels were observed in virus‐infected dcl2b and WT tomato plants previously (T. Wang et al., 2018), markedly lower miR6026 expression levels were observed in PSTVd‐infected dcl2b plants than in WT plants (Figure 3a, Table S3). This result was confirmed by reverse transcription‐quantitative PCR (RT–qPCR) (Figure 3b, left; primers listed in Table S2). Thus, the feedback loop‐mediated regulation of SlDCL2 by miR6026 may also be involved in the susceptibility of dcl2b plants to viroid infection.

FIGURE 3.

Tomato miR6026 produced by SlDCL2 helps defend against PSTVd infection through the down‐regulation of SlDCL2 expression via a feedback mechanism. (a) Differentially expressed tomato miRNAs in PSTVd‐infected dcl2b mutant tomato (cv. Ailsa Craig) plants and wild‐type (WT) plants, identified by small RNA sequencing. The down‐regulated miR6026 is indicated in black. (b) Reverse transcription‐quantitative PCR (RT‐qPCR) analysis of miR6026 expression in PSTVd‐infected dcl2b plants and WT plants (left) and in tomato (cv. Ailsa Craig) plants agroinfiltrated with a potato virus X (PVX) empty expression vector (PVX:EV; n = 8, **p < 0.01 by Student's t test) and those agroinfiltrated by the vector PVX:TM6 (***p < 0.001) containing target mimics of miR6026 (right). The SlActin was used as the reference gene and U6 was used as an internal control. (c) RT‐qPCR analysis to determine difference in expression of two gene targets of miR6026, SlDCL2a and SlDCL2b, between PVX:EV (*p < 0.05) and PVX:TM6 (**p < 0.01) agroinfiltrated tomato plants. (d) Northern blotting analysis of PSTVd in infected tomato plants that were agroinfiltrated by PVX:EV and PVX:TM6 at 14 days post‐infiltration (dpi). Mock‐infected plants (M) and PSTVd‐infected samples (P) obtained previously were as controls. (e) PSTVd‐infected WT and L6 tomato (cv. M82) plants. (f) Northern blotting analysis of PSTVd in infected WT tomato plants and the transgenic tomato (L6) plants expressing the target mimics of miR6026 at 14, 21 and 28 dpi. Mock‐infected plants (M) and PSTVd‐infected samples (P) were used as controls.

To verify the function of miR6026 in the defence against viroid infection in tomato, a potato virus X (PVX) vector‐based miRNA silencing method (Zhao et al., 2016) was used to express miR6026 target mimics in tomato (cv. Ailsa Craig). The fragment containing two short tandem target mimics (STTM) of miR6026 with a 48 nt spacer sequence (Table S4) in the middle was inserted into the PVX vector to create the recombinant construct PVX:TM6 (Tang et al., 2012). An empty PVX vector (PVX:EV) was used as the negative control. The constructs were transformed into A. tumefaciens GV3101 and infiltrated into tomato; eight plants were used for each treatment and the experiment was repeated twice. RT–qPCR detection at 5 days post‐infiltration revealed a significant decrease of miR6026 expression levels in PVX:TM6‐infiltrated plants compared with the PVX:EV‐infiltrated plants (p < 0.001) (Figure 3b, right). The targets of miR6026 include both SlDCL2a and SlDCL2b (Z. M. Wang et al., 2018). Thus, RT–qPCR was performed to detect the expression levels of these two genes. Correspondingly, they markedly increased in the PVX:TM6‐infiltrated plants (p < 0.05 and p < 0.01, respectively) (Figure 3c).

Next, the PVX:TM6‐ and PVX:EV‐infiltrated plants (eight plants for each) were agroinfiltrated with PSTVd. This experiment was repeated twice. PVX:EV‐infiltrated plants began developing visible, but mild, leaf curling symptoms at 14 dpi, earlier than the 21 dpi timepoint observed for PVX:TM6‐infiltrated plants. Moreover, PVX:EV‐infiltrated plants developed more severe symptoms than the PVX:TM6‐infiltrated plants at 21 dpi (Figure S2a). These results indicate that functional inhibition of miR6026 increases PSTVd resistance in tomato. In addition, the increase in resistance was correlated with delayed and reduced PSTVd accumulation levels (Figures 3d and S2b).

To further verify the involvement of miR6026 in the defence against viroid infection, we agroinfiltrated WT tomato (cv. M82) plants and the L6 transgenic plants (10 for each) (Z. M. Wang et al., 2018) expressing the target mimics of miR6026, which were kindly provided by Professor Zhengming Wang at the Huazhong Agricultural University. PSTVd was detected by northern blotting at 14, 21 and 28 dpi. L6 plants expressed necrosis and leaf curling symptoms only in the first two or three true leaves at 14 dpi (Figure S3). This phenomenon is worth exploring in the future. No marked symptoms were found in other leaves and plant height between inoculated L6 and WT plants (Figures 3e and S3). Northern blotting analysis showed that PSTVd accumulation was delayed in L6 plants (Figure 3f), indicating that L6 plants were more tolerant towards PSTVd infection than WT plants. Thus, the feedback regulation of SlDCL2 by miR6026 (Z. M. Wang et al., 2018) also functions in the defence of tomato against viroid infection. It should be noted that this regulation mechanism appears to function only in the early stage of viroid infection. It means that miR6026 functions in viroid resistance in an indirect way or other complementary or redundant resistance mechanisms exist in tomato.

Here, we detail the function of SlDCL2b in the defence response against PSTVd infection and its feedback regulation by SlDCL2‐dependent miR6026 in tomato plants. Although the involvement of DCL2 in the defence response against viroid infection has been previously reported in tomato (Suzuki et al., 2019) and N. benthamiana (Katsarou et al., 2016), this function was achieved by combining with other DCLs such as DCL4 or DCL3. Moreover, the down‐regulation of NbDCL2 did not appear to affect resistance against viroid infection in N. benthamiana (Dadami et al., 2013). In this study, we produced genetic evidence that supports the function of SlDCL2b in the defence response against viroid infection in tomato. Importantly, this finding indicates that the defence function of DCL2 against viroid infection is not fully compensated for by other DCLs, especially DCL4, in tomato, even though DCL2 function can be compensated by DCL4 in N. benthamiana (Katsarou et al., 2016).

The finding that the function of SlDCL2b in suppressing viroid infection is not completely compensated by SlDCL4 raised the question of how viroids counter the SlDCL4‐mediated defence response. Host‐adapted plant viruses encode VSRs that can inhibit the antiviral function of DCL4 (Deleris et al., 2006); however, viroids do not encode any proteins and therefore lack VSRs. Thus, viroids would have had to evolve distinct strategies to counter the defence response function of DCL4. The compact secondary genomic structure of viroids has been shown to be resistant to RISC‐mediated degradation (Itaya et al., 2007). However, this mechanism means that RISCs guided by 21 and 22 nt viroid‐derived siRNAs degrade mature viroid genomic RNAs at different rates, which is yet to be verified. Moreover, the compartmentation of genomic RNAs of different polarities (Qi & Ding, 2003) and even possible phase separation during viroid replication may affect the DCL4‐mediated defence response.

Importantly, the findings presented here could potentially be used to enhance viroid resistance in susceptible plants, by engineering plants with altered miRNAs that regulate the expression of genes involved in the defence response against viroids (Z. M. Wang et al., 2018). PSTVd resistance in tomato was obviously enhanced by inhibiting the function of miR6026 by expressing mimics of its target genes. In recent years, various RNAi‐based strategies have been employed to control viroid infection, with a particular focus on using highly specific artificial sRNA‐based tools such as artificial microRNAs and synthetic trans‐acting siRNAs (Carbonell & Daros, 2017). However, potential off‐target effects of expressed sRNAs hamper the application of these approaches to some extent. The miRNA‐blocking strategy avoids this problem because the expressed mimics do not directly target rapidly evolving viroid genomic RNAs and thus can provide more durable resistance.

In short, based on the genetic evidence, we determined the resistance function of SlDCL2b to PSTVd and found that this function is feedback‐regulated by SlDCL2‐dependent miR6026. These findings deepen our understanding of the mechanism of plant resistance to viroid by RNA silencing and provide a new strategy for viroid control by expressing an miRNA.

Supporting information

Figure S1. Distribution profiles of PSTVd‐sRNAs against PSTVd genome in the repeated samples of wild type (WT‐2) and mutant tomato plants (dcl2b‐2).

Figure S2. (a) Symptom development in PSTVd‐infected tomato (cv. Ailsa Craig) plants that were agroinfiltrated by PVX:EV and PVX:TM6 at 14 and 21 dpi. (b) Northern blotting analysis of PSTVd in infected plants that were agroinfiltrated by PVX:EV and PVX:TM6 at 21 days post‐infiltration.

Figure S3. PSTVd‐infected wild‐type (WT) tomato (cv. M82) plants and the transgenic tomato plants (L6) expressing the target mimics of tomato miR6026 at 14 and 21 days post‐inoculation.

Table S1. The sequences and amount of potato spindle tuber viroid (PSTVd)‐sRNAs (21–24 nucleotides).

Table S2. Primers used for reverse transcription (RT)‐PCR and RT‐quantitative PCR in this study.

Table S3. miRNA analysis in PSTVd‐infected mutant (dcl2b) tomato (cv. Ailsa Craig) plants and PSTVd‐infected wild‐type tomato plants.

Table S4. Nucleotide sequences of the short tandem target mimics (STTM) of tomato miR6026.

Zhang, Y. , Tian, X. , Xu, H. , Zhan, B. , Zhou, C. , Li, S. et al. (2024) Knockout of SlDCL2b attenuates the resistance of tomato to potato spindle tuber viroid infection. Molecular Plant Pathology, 25, e13441. Available from: 10.1111/mpp.13441

DATA AVAILABILITY STATEMENT

Small RNA sequencing data generated in this work has been deposited in GenBank at https://www.ncbi.nlm.nih.gov/genbank/ with accession numbers SAMN39479130–SAMN39479133.