Virus-Induced Gene Silencing in Chrysanthemum seticuspe Using the Tomato Aspermy Virus Vector

Chrysanthemum is one of the most economically important flowers globally due to its high ornamental value. In recent years, a large percentage of the Chrysanthemum seticuspe genome has been determined, making this species useful as a model chrysanthemum plant. To fully utilize the genome’s information, efficient and rapid gene functional analysis methods are needed. In this study, we optimized the tomato aspermy virus (TAV) vector for virus-induced gene silencing (VIGS) in C. seticuspe. Conventional plant virus inoculation methods, such as the mechanical inoculation of viral RNA transcripts and agroinoculation into leaves, did not achieve successful TAV infections in C. seticuspe, but vacuum infiltration into sprouts was successful without symptoms. The TAV vector harboring 100 nucleotides of the phytoene desaturase (PDS) gene caused photobleaching phenotypes and a reduction in CsPDS expression in C. seticuspe. To our knowledge, this is the first report of VIGS in chrysanthemums.


Introduction
Chrysanthemum belongs to the Asteraceae family, which is the largest family of angiosperms [1]. The cultivated chrysanthemum, Chrysanthemum morifolium Ramat, is one of the most economically important flowers globally due to its high ornamental value [2]. However, cultivated chrysanthemum has a large genome and high polyploidy [3,4]; therefore, the genetic transformation of chrysanthemum is difficult and inefficient compared with that of other major model plants. Chrysanthemum seticuspe (Maxim.) Hand.-Mazz. (hereafter referred to as C. seticuspe) is a wild diploid chrysanthemum closely related to the cultivated chrysanthemum. These properties are preferable for a model plant; therefore, 89-97% of C. seticuspe genomes have been determined by next-generation sequencing [5,6].
To fully utilize this genome information, efficient gene functional analysis methods are needed instead of laboriously inefficient conventional plant transformations.
As a useful tool for gene functional analysis, various virus-induced gene silencing (VIGS) vectors have been developed in recent years [7]. VIGS depends on posttranscriptional gene silencing (PTGS) machinery in a sequence-specific manner. Briefly, double-stranded RNAs (dsRNAs), such as replication intermediates of plant RNA viruses and highly structured genomic RNA, are processed by Dicer-like proteins (DCLs) and generate small interfering RNAs (siRNAs). An RNA-induced silencing complex (RISC) incorporating siRNAs cleaves complementary single-stranded RNA (ssRNA) [8]. When a plant virus vector with a plant gene sequence replicates in plant cells, the complementary plant mRNA is also degraded through the PTGS machinery. VIGS does not require transformations of target plants; therefore, VIGS has an advantage over other laborious conventional methods for gene functional analyses [9].
In this report, we aimed to optimize the TAV VIGS vector for C. seticuspe to establish an efficient VIGS vector system for chrysanthemums. Vacuum infiltration into sprouts was successful for the inoculation of the TAV VIGS vector in C. seticuspe, whereas the mechanical inoculation of viral RNA transcripts and agroinoculation into leaves did not work. Although efficient TAV VIGS in N. benthamiana was achieved by the attenuation of the viral suppressor of RNA silencing (VSR) activity of the 2b protein [13], we demonstrated that the TAV vector with the wild-type 2b protein harboring the partial C. seticuspe phytoene desaturase (CsPDS) sequence caused photobleaching phenotypes in C. seticuspe. The TAV vector will be a useful rapid functional genomics tool in chrysanthemums.

Plant Material
Gojo-0, which is a self-compatible C. seticuspe pure line, was used as the chrysanthemum [18]. Five-or six-leaf-stage C. seticuspe plants were used for conventional inoculation methods, such as the mechanical inoculation of TAV RNA transcripts or agroinoculation into leaves, and germinated seeds were used for sprout vacuum infiltration.

Virus Inoculation
For the inoculation of TAV transcripts, after the linearization of the plasmid containing TAV cDNA (pTOPOT1, pTOPOT2, and pTOPOT3) by SacI, the 3 overhang was converted to a blunt end using T4 DNA polymerase (Applied Biological Materials, British Columbia, Canada). Infectious viral RNAs were synthesized using the RiboMAX large-scale RNA production system (Promega). Synthesized RNAs mixed in 50 mM Na 2 HPO 4 , pH 9.2, were mechanically inoculated into two fully expanded leaves of five-or six-leaf-stage C. seticuspe plants.
The Rhizobium radiobacter strain EHA105 was used for agroinoculation. Agroinoculation into leaves was performed as described previously with little modification [13]. Bacterial cultures were resuspended to a cell concentration of OD600 = 1.5.
For sprout vacuum infiltration, bacterial cultures were resuspended in infiltration buffer (10 mM MgCl 2 , 10 mM MES pH 5.6 and 200 µM acetosyringone) to a cell concentration of OD600 = 1.5. Equal volumes of bacterial cultures for RNA1, RNA2, and RNA3 were mixed and incubated at room temperature for 3 h. C. seticuspe seeds were germinated on filter paper presoaked with sterile water in a Petri dish for four days. The germinated C. seticuspe seeds were submerged in the bacterial culture mixture with 0.5% Tween 20 and infiltrated twice using a vacuum pump under 8 hPa pressure for 5 min. Infiltrated sprouts were transplanted into pots and covered with a chamber for 3-4 days to maintain humid conditions. After 3-4 days, the chamber was removed, and the plants were maintained at 24 • C with a 16 h light/8 h dark cycle.

Detection of Viral CP by a Dot Immunobinding Assay (DIBA)
The leaf tissues were ground with a tenfold volume of phosphate-buffered saline (PBST) containing 0.05% Tween 20. The homogenates were spotted onto a BioTrace NT Membrane (Pall Corporation, Port Washington, NY, USA) and blocked with 3% skim milk-Tris-buffered saline (TTBS) containing 0.05% Tween 20. The membrane was washed for 10 min with PBST, and the primary antibody against TAV (1:5000 dilution in PBST) was added and incubated at 4 • C overnight. After being washed twice, the membrane was incubated with alkaline phosphatase-conjugated secondary antibody (1:10,000 dilution in 3% skim milk-TTBS) for 2 h at room temperature. Staining was performed using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium.

Detection of Viral RNA by RT-PCR and Gene Expression Analysis by RT-qPCR
Total RNA was isolated from leaf tissues using ISOGEN II (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. The detection of viral RNA by RT-PCR was performed as described previously [13].
To analyze gene expression levels in plants showing silencing phenotypes, RT-qPCR was performed. Total RNA was treated with a gDNA Removal Kit (Jena Bioscience, Jena, Germany), and 250 ng of gDNA-free total RNA was used as a template. The RT reaction was performed using ReverTra Ace (Toyobo) with an oligo (dT) 15 primer. The resulting five-fold-diluted RT product was then used as a template for qPCR. Real-time PCR was performed using a KAPA SYBR FAST qPCR Master Mix kit (Kapa Biosystems, Wilmington, MA, USA) with QuantStudio 3 (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. For the analysis of CsPDS expression levels, the primer pair CsPDS-qFw (5 -TCC TGA AGA AAC CGG TTT AG-3 ) and CsPDS-qRv (5 -CTG AGG TGT TTA TTG CCA TG-3 ) was used. For the analysis of CsEF1α as an internal control, a primer pair (5 -CTT GTT GCT TGA TGA CTG TGG-3 ) and (5 -ACC ATT CAA GCG ACA GAC TC-3 ) was used [19]. A standard curve method based on the serial dilution of the TAV empty vector was used to calculate the relative expression level of each gene. Each RT-qPCR analysis was performed with five to six biological replicates and analyzed statistically by Student's t-test.

TAV Inoculation Method for C. seticuspe
We previously developed infectious clones of TAV-ChJ isolated from chrysanthemums and performed VIGS in N. benthamiana but not in C. seticuspe [13]. Therefore, we first confirmed the infectivity of TAV cDNA clones to C. seticuspe. The mechanical inoculation of TAV transcripts and the agroinoculation of TAV cDNA clones into leaves were not successful ( Table 1), suggesting that these conventional inoculation methods are not suitable for C. seticuspe. It has been reported that the inoculation of strawberry vein banding virus (SVBV) by vacuum infiltration was more effective than that by agroinoculation using syringe infiltration [20]. Therefore, as an alternative TAV inoculation method, we performed sprout vacuum infiltration (SVI) [21]. C. seticuspe seeds were germinated in a Petri dish, and the germinated sprouts were vacuum-infiltrated with bacterial cultures containing TAV cDNA clones. TAV was detected in uninoculated upper leaves from ten out of twenty of the vacuum-infiltrated C. seticuspe by RT-PCR or DIBA at 28 days post-inoculation (dpi) ( Table 1 and Figure 1a). Importantly, all TAV-detected plants showed no symptoms (Figure 1b). These results suggest that SVI is a suitable inoculation method for C. seticuspe.

Infectivity of 2b Mutants That Have Various VSR Activities to C. seticuspe
Our recent data suggested that efficient VIGS induction by the TAV-ChJ vector in N. benthamiana was achieved when the VSR activity of the 2b protein was attenuated [13]. Therefore, we examined the effect of 2b VSR activity on TAV infection of C. seticuspe. TAVs containing the ORF for the wild-type or mutated 2b protein retaining various VSR activities (∆C61, ∆C23, R46C, and S4042A) were vacuum infiltrated into C. seticuspe sprouts. The deletion mutants ∆C61 and ∆C23 was not detected at 28 dpi (Table 2). S4042A and R46C mutants were detected with detection percentages of 4.2 and 15%, respectively (Table 2), which were drastically lower than that of the wild type (50%). These results suggest that efficient TAV infection of C. seticuspe requires intact VSR activity of the 2b protein.
Considering the lower detection percentage of 2b mutants, we decided to use wild-type TAV for VIGS in C. seticuspe.

VIGS in C. seticuspe
The 100-nucleotide (nt) fragment from the CsPDS gene, a visible VIGS marker gene, was introduced downstream of the CP ORF in pJL89T3 in a sense (CsPDS100-S) or antisense orientation (CsPDS100-AS) and vacuum-infiltrated with pJL89T1 and pJL89T2 (Figure 2a). C. seticuspe plants inoculated with TAV-CsPDS100-S and TAV-CsPDS100-AS showed photobleaching phenotypes, whereas the TAV empty vector did not induce photobleaching phenotypes (Figure 2b). The silencing frequency was defined as the percent of plants showing silencing phenotypes after inoculation with VIGS vectors [22], and the silencing frequencies of TAV-CsPDS100-S and TAV-CsPDS100-AS were 48.7 ± 13.7% and 24.5 ± 10.9% (8-18 plants were inoculated with three replicates), respectively. The silencing frequency of photobleaching phenotypes induced by TAV-CsPDS100-S was nearly equal to the detection percentage of the TAV 2b wt ( Table 2), suggesting that TAV-CsPDS100-S can induce photobleaching phenotypes at a high rate if they are infected. The photobleaching areas of TAV-CsPDS100-S were larger than those of TAV-CsPDS100-AS (Figure 2b). Consistent with the photobleaching phenotype, RT-qPCR showed that the relative CsPDS expression levels of TAV-CsPDS100-S and TAV-CsPDS100-AS were decreased compared with that of the TAV empty vector (Figure 2d). The presence of the TAV vector harboring the partial CsPDS sequence was confirmed by RT-PCR in the photobleached tissues of C. seticuspe (Figure 2c). These results indicate that the photobleaching phenotypes were induced by TAV vectors containing the partial CsPDS sequence, and VIGS was more effective when the partial CsPDS sequence was inserted in a sense orientation. RT-PCR analysis of the TAV vector harboring the partial CsPDS sequence. RNA samples were extracted from C. seticuspe leaves at 28 dpi. pJL89 was used as a negative control. TAV not harboring the partial CsPDS sequence was used as the empty vector control. The RT-PCR product of the TAV vector harboring the partial CsPDS sequence was approximately 1200 base pairs (bp), whereas that of the TAV empty vector was approximately 1100 bp. (d) RT-qPCR analysis of CsPDS expression levels. RNA samples were extracted from C. seticuspe leaves at 28 dpi. TAV not harboring the partial CsPDS sequence was used as the empty vector control. Each RT-qPCR analysis was performed with five to six biological replicates. Boxes show the interquartile ranges including 25-75% of the values, and whiskers indicate the highest and lowest values of data. Horizontal lines and cross marks in boxes indicate the medians and means, respectively. The value outside 1.5 times the interquartile range between 25% and 75% of each group was considered an outlier and indicated with a dot. Asterisks indicate significant differences by Student's t-test at p < 0.05.

Discussion
To utilize most of the advantages of VIGS, the inoculation method of the VIGS vector should be efficient and simple. We first attempted conventional plant virus inoculation methods, such as mechanical inoculation and agroinoculation, but these two methods failed to establish TAV infections in C. seticuspe (Table 1 and Figure 1a). We showed that SVI achieved successful TAV infections in C. seticuspe when the same bacterial inoculums, except for the acetosyringone concentration (100 and 200 µM for leaf infiltration and SVI, respectively), were used. It has been reported that vacuum infiltration of the SVBV cDNA clone was more infectious than that delivered by agroinfiltration into leaves [20]. The reason why the difference in infectivity occurred between these two agroinfiltration methods is unclear, but there are two differences between leaf infiltration and SVI. First, C. seticuspe sprouts were used for SVI soon after germination, whereas five-or six-leaf-stage C. seticuspe were infiltrated in the leaf infiltration. Second, bacterial mixtures were vacuum-infiltrated in whole seedlings, including cotyledons, stems, and roots, for the SVI, whereas bacterial mixtures were infiltrated into only true leaves for leaf infiltration. Namely, the timing and sites of the agrobacterium introduction (and/or viral RNA, which is transcribed from T-DNA) were different between these two agroinfiltration methods. To achieve efficient TAV infections in C. seticuspe, it may be suitable to agroinoculate the whole seedlings when C. seticuspe is extremely young and more susceptible to viral infection. It has been reported that C. seticuspe were infected with chrysanthemum stunt viroid by the injection of viroid transcripts into stems [23], supporting the importance of the pathogen entering the site for successful infection.
The TAV 2b protein is a counterdefense protein against RNA silencing. Some mutations in TAV 2b decreased its VSR activity but did not cause a loss of infectivity of TAV in N. benthamiana [13]. However, in this study, the TAV 2b mutants lost or showed extremely decreased infectivity in C. seticuspe ( Table 2), showing that the intact 2b protein is required for the efficient infection of TAV in C. seticuspe. Previously, we showed that decreased VSR activity of the 2b protein was required for VIGS of the TAV vector in N. benthamiana [13]. Recently, Zhou et al. reported that the percentage of silenced plants induced by the tobacco rattle virus (TRV) vector expressing the TAV 2b protein was only 3% in pepper, whereas that induced by the TRV vector expressing the CMV 2b protein was 93% [24]. Thus, TAV 2b seems to suppress VIGS. However, in this study, the TAV VIGS vector with wild-type 2b induced PDS silencing in C. seticuspe with a higher rate of 25-49% ( Figure 2). Considering these reports and our results, it is suggested that the effect of TAV 2b VSR activity on VIGS is different between Asteraceae and Solanaceae plants or that the PTGS machinery itself is different between these two plant families. N. benthamiana has been used as a model plant for plant virology for a long time due to its high susceptibility to pathogens [25]. In addition, many RNA silencing-based crop protection systems have been reported in Solanaceae plants because this family includes economically important crops such as tomato and potato [26]. For these reasons, much information about the RNA silencing machinery is accumulated in Solanaceae plants. To our knowledge, the RNA silencing mechanism of Asteraceae has not been investigated. A detailed analysis of the RNA silencing mechanism of Asteraceae plants may help to provide a better understanding of the effect of TAV 2b VSR activity on VIGS.
The orientation of the target sequence affects VIGS effectiveness. Some reports showed that antisense orientation was more effective for VIGS [27][28][29], but others showed that there was no difference between the sense and antisense orientation [30][31][32][33][34]. These reports differ in the viral vectors used, the host plants, and the target genes; thus, the effect of gene orientation to induce VIGS may depend on the experimental conditions. In this study, the TAV vector harboring the partial CsPDS sequence in any orientation caused photobleaching phenotypes, but the antisense orientation did not induce more effective VIGS against CsPDS than the sense orientation (Figure 2b,d). It has been reported that the position or sequence of the target gene also affects VIGS [29][30][31]. Therefore, it remains possible that other positions or sequences of the CsPDS gene may result in a different VIGS effect in a sense and antisense orientation.
In summary, we optimized the inoculation method and VSR activity of the 2b protein for the TAV VIGS system in C. seticuspe. The optimized TAV VIGS vector system silenced a C. seticuspe endogenous gene, but the silenced phenotype was in a limited area. Therefore, modifications allowing VIGS in a broader area will provide more useful functional genomics tools in chrysanthemums.