Development of a Mild Viral Expression System for Gain-Of-Function Study of Phytoplasma Effector In Planta

PHYL1 and SAP54 are orthologs of pathogenic effectors of Aster yellow witches’-broom (AYWB) phytoplasma and Peanut witches’-broom (PnWB) phytoplasma, respectively. These effectors cause virescence and phyllody symptoms (hereafter leafy flower) in phytoplasma-infected plants. T0 lines of transgenic Arabidopsis expressing the PHYL1 or SAP54 genes (PHYL1 or SAP54 plants) show a leafy flower phenotype and result in seedless, suggesting that PHYL1 and SAP54 interfere with reproduction stage that restrict gain-of-function studies in the next generation of transgenic plants. Turnip mosaic virus (TuMV) mild strain (TuGK) has an Arg182Lys mutation in the helper-component proteinase (HC-ProR182K) that blocks suppression of the miRNA pathway and prevents symptom development in TuGK-infected plants. We exploited TuGK as a viral vector for gain-of-function studies of PHYL1 and SAP54 in Arabidopsis plants. TuGK-PHYL1- and TuGK-SAP54-infected Arabidopsis plants produced identical leafy flower phenotypes and similar gene expression profiles as PHYL1 and SAP54 plants. In addition, the leafy flower formation rate was enhanced in TuGK-PHYL1- or TuGK-SAP54-infected Arabidopsis plants that compared with the T0 lines of PHYL1 plants. These results provide more evidence and novel directions for further studying the mechanism of PHYL1/SAP54-mediated leafy flower development. In addition, the TuGK vector is a good alternative in transgenic plant approaches for rapid gene expression in gain-of-function studies.


Introduction
Aster yellow witches'-broom (AYWB) phytoplasma causes virescence and phyllody symptoms in host plants [1]. Maclean et al. (2011) individually expressed several putative secreted AYWB phytoplasma proteins (SAPs) in Arabidopsis to identify the phytoplasma effector that induces these virescence and phyllody symptoms (hereafter leafy flower). Large-scale screening revealed SAP54 to be the effector causing leafy flower and crinkled silique phenotypes in Arabidopsis [2]. Furthermore, the PHYL1 effector of Onion yellows phytoplasma, an effector orthologous to SAP54, also results in a leafy flower phenotype in Arabidopsis expressing the PHYL1 gene [3], suggesting that SAP54 and PHYL1 play roles in leafy flower formation. Our pervious study indicated that Peanut witches'-broom (PnWB) phytoplasma also occurs leafy flower symptoms in Catharanthus roseus plants [4], and its PHYL1 gene has also been identified [5]. To evaluate the PHYL1 of PnWB function, the gain-of-function study in Arabidopsis plants is an approach to study the mechanism of leafy flower development. However, the transgenic Arabidopsis expressing SAP54 (SAP54 plants) have a seedless problem that restricted the further study in function of effectors. Therefore, an alternative strategy for PHYL1/SAP54 gain-of-function in vivo that is independent from the transgenic approach is needed.
Since the 1990s, plant viral vectors have been developed to carry foreign genes for different purposes. Unlike the transgenic approach, a viral vector has several advantages, such as selfreplication and long-distance movement, allowing the systemic spread of interesting foreign sequences for efficient expression or virus-induced gene silencing (VIGS) in plant [6][7][8][9][10][11][12]. A viral suppressor is a key regulator for counteracting the RNA-silencing defense in plants [13][14][15]. Because of silencing suppression, foreign proteins in viral vector-infected plants are expressed to a higher degree than in transgenic plants [16][17][18]. In addition, the viral suppressor interferes microRNA (miRNA) biogenesis, resulting in severe symptoms that affect the normal growth status of host plants [19,20]. To overcome this issue, a deconstructed vector was applied for over-expression of a foreign gene through the expression of the required viral elements [21,22]. Moreover, this deconstructed vector can also overcome the problems of size limitations and the deletion of foreign genes, which occurred with viral vectors [21].
Tobacco rattle virus (TRV) and Potato virus X (PVX), which has a weak suppressor in silencing suppression, have been widely applied for VIGS in loss-of-function studies [12]. A suitable viral vector, other than a VIGS vector, for rapid gain-of-function studies in infected wildtype or mutant plants is also important; especially when the transgene causes a lethal phenotype in seedlings of the next generation. Many reports have demonstrated that several virus species can express green fluorescent protein (GFP) gene for proving of the concept in viral vector development; however, functional study of gene is lacking [6,23].
The attenuated Turnip mosaic virus (TuMV) strain (TuGK) has an Arg182Lys mutation in helper component-proteinase (HC-Pro R182K ). HC-Pro R182K is a mutant of a gene-silencing suppressor that has lost its ability to suppress the miRNA pathway, results in no symptoms in TuGK-infected Arabidopsis thaliana (hereafter Arabidopsis) and Nicotiana benthamiana plants [19,24]. TuGK has been modified as a viral vector to express GFP for indicating and monitoring the virus location in the host [19]. An infectious, full-length cDNA clone of TuGK has been constructed in the mini binary vector pBD003 to generate pBD-TuGK, which can be directly inoculated into host plants via agro-infiltration to induce an initial infection [24]. TuGK-infected Arabidopsis plants is symptomless [19], making TuGK suitable for the over-expression of foreign genes in Arabidopsis plants, with few or no side effects due to the severe pathogenicity of the virus.
In this study, TuGK was used as a viral vector for rapid expression to study the function of the PHYL1 and SAP54 gene in Arabidopsis plants. Recombinant TuGK viruses could successfully express effectors in Arabidopsis and mimic the leafy flower phenotype of the transgenic plants, demonstrating the use of TuGK as a rapid system for gain-of-function studies in planta.

Plant materials and growth conditions
Arabidopsis seeds were surface sterilized and chilled at 4°C for 2 days before being sown on Murashige and Skoog (MS) medium with/without suitable antibiotics for selection. One-weekold seedlings that had germinated on the MS plates were transferred to soil. N. benthamiana seeds were sown in soil. All of the seedlings and plants were maintained in either a growth chamber or greenhouse (16 hr light/8 hr dark, 20 to 25°C).
The infectious viral clones were transformed into the Agrobacterium tumefaciens C58C1 strain, and virus infection was performed in N. benthamiana plants using the agro-infiltration procedure [24]. Four days after infiltration, the infiltrated leaves of N. benthamiana plants were collected to analyze viral infectivity and were used as the primary inoculum to mechanically inoculate Arabidopsis plants. Recombinant TuGK infection was performed in 2.5-weekold Col-0 or dcl2-4/dcl4-1 double-mutant (dcl2/4) plants. Flower tissues from plants at 20 days post-inoculation (dpi) were used for gene profile evaluation.

Longitudinal sectioning of shoot apical meristem and confocal microscopy
The shoot apical meristem (SAM) of 4 dpi TuGK-or TuGK-PHYL1-infected Arabidopsis and N. benthamiana plants were embedded in 5% agar and sectioned at 200-μm thickness using a microslicer (DTK-1000; Dosaka EM, Japan). The GFP fluorescence in the SAM was monitored using a Leica TCS SP5 II confocal laser-scanning microscope (Joint Center for Instruments and Researches, College of Bioresources and Agriculture, National Taiwan University) that was equipped with a multiline argon laser with a filter set for GFP fluorescence [excitation filter Acousto-optic Tunable filter 488, emission bandwidth 502 to 572 nm, PMT2 offset (-1.0)/grain (895.3)] and a filter set for chlorophyll fluorescence [excitation filter Acousto-optic Tunable filter 488, emission bandwidth 608 to 677 nm, PMT3 offset (0.0)/gain (855)]. All images were graphically arranged using Adobe Photoshop CS3 software (Adobe Systems Inc., Mountain View, CA, U.S.A).
E. coli cells transformed with either pET-PHYL1 or pET-SAP54 were individually cultured in 500 ml of LB medium at 37°C until the absorbance at 600 nm reached 0.5. After that, the 0.5 mM isopropyl-D-1-thiogalactopyranoside (IPTG) was added in the culture and grown at 20°C for 6 hr. Recombinant protein purification and antiserum production were performed using modified protocols published by Chiu et al [27]. Approximately 1 mg of either recombinant PHYL1 or SAP54 protein was mixed with Freund's adjuvant (1:1 v/v) and injected into a New Zealand white rabbit once per week; the procedure was repeated 4 times. For the first injection, the "complete" adjuvant was used, and the "incomplete" adjuvant was for the subsequent injections. The titer of the antiserum was analyzed using western blotting. Institutional Animal Care and Use Committee (IACUC) of National Taiwan University (NTU) approved the antiserum production in this study. The feeding and care of the experimental rabbits were bred at animal room of Institute of Biotechnology of NTU. All antiserum processing, including animal welfare, ameliorates suffering, and sacrifice, were performed according to institutional guidelines under the regulation of the IACUC of NTU and to laws of Taiwan on animal protection.

Western blot analysis
Plant tissue was ground in 5-10 volumes (w/v) of 20 mM phosphate buffer, pH 7.0, and the plant extracts were suspended in 2 volumes (w/v) of 2× protein sample dye (2% SDS, 10% glycerol, 1% ß-mercaptoethanol, 0.005% bromophenol blue, and 50 mM Tris-HCl, pH 6.8), denatured 100°C for 10 min and then cooled on ice for 2 min. The protein samples were separated on 12% polyacrylamide-SDS gels and then transferred to a PVDF membrane (GE Healthcare) with transfer buffer (50 mM Tris-HCl, 40 mM glycine, 1 mM SDS, and 20% methanol). For TuMV coat protein (CP) detection, the CP antiserum was used at a 10,000× dilution . For phytoplasma effector detection, SAP54 or PHYL1 antisera were used at a 10,000× dilution. For GFP detection, a GFP monoclonal antibody (GE Healthcare) was used at an 8000× dilution. HRP-conjugated anti-rabbit or anti-mouse antibodies (GE Healthcare) were used as secondary antibodies at a 10,000× dilution, and the signals were detected using WesternBright ECL (Advansta). The membrane was stained with staining solution (0.6 mM Coomassie blue R-250, 50% methanol, 10% acetic acid) to perform riblose-1,5-bisphosphate carboxylase (RUBISCO) staining as a loading control.

Results
TuGK expresses phytoplasma effectors in planta pBD-TuGK was used to carry the PHYL1 or SAP54 gene (Fig 1A). Both effector genes were fused at the C-terminus of GFP and inserted between the NIb and CP genes with NIa protease cleavage sites (Fig 1A). pBD-TuGK-PHYL1 and pBD-TuGK-SAP54 were produced by the TuGK-PHYL1 and TuGK-SAP54 recombinant viruses in N. benthamiana plants after agro-infiltration and exhibited a GFP signal under fluorescence microscopy ( Fig 1B). Compared with plants infected with a wild-type TuMV that expresses GFP (TuGR), GFP fluorescence was lower in the TuGK-, TuGK-PHYL1-, and TuGK-SAP54-infected plants (Fig 1B). Thus, TuGK can express both effectors at lower expression levels.

Evaluating the efficiency of PHYL1 and SAP54 expression by TuGK
TuGR and TuGK expressed GFP in both inoculated leaves (ILs) and systemic leaves (SLs) of N. benthamiana plants at 4 dpi (Fig 2A). In contrast, only ILs showed TuGK-PHYL1 and TuGK-SAP54 infection at 4 dpi (Fig 2A), suggesting that phytoplasma effectors might interfere with TuGK replication and infection in host plants. Surprisingly, the PHYL1 gene was deleted from the TuGK genome in SLs, leaving only GFP (25 kDa) and truncated fusion proteins in SLs at 8 dpi (Fig 2B). Note, the SAP54 gene was also deleted from TuGK in some instances (data not shown).
TuGK-infected leaves of N. benthamiana plants at 4 dpi were used as the inoculum to inoculate Arabidopsis plants. TuGK-PHYL1-and TuGK-SAP54-infected Col-0 plants showed partial deletion of the GFP-PHYL1 and GFP-SAP54 fusion proteins at 20 dpi; nevertheless, high levels of the fusion proteins persisted (Fig 2C). Moreover, SAP54 and PHYL1 antisera specifically detected GFP-SAP54 and GFP-PHYL1, respectively, in the infected Arabidopsis plants (Fig 2D), suggesting that TuGK expresses either the PHYL1 or SAP54 effector gene, respectively. Although PHYL1 and SAP54 exhibit 60.4% similarity, no cross-reactivity was observed by western blotting, demonstrating the high specificity of the antisera in distinguishing these two effectors.

TuGK was detected in SAM
The SAM is an important position in plants with regard to leafy or flower organ determination. We assume that 35S promoter-driven effectors affect the gene expression in the SAM of PHYL1 and SAP54 plants, resulting in leafy flowers. Next, we evaluated whether TuGK can migrate to the SAM and express the effector in the meristem. The results of longitudinal section showed GFP fluorescence in the SAM regions of TuGK-and TuGK-PHYL1-infected N. benthamiana plants (Fig 3). Moreover, the SAM of Arabidopsis also showed GFP fluorescence of TuGK (S1 Fig). These results indicated that TuGK and TuGK-PHYL1 migrate to the meristem and express GFP-PHYL1 in the SAM region. Therefore, TuGK-delivered PHYL1 might affect gene expressions in the meristem as in PHYL1 plants.

The TuGK vector enhances the phenotype formation rate
In PHYL1 plants, the PHYL1 gene was constructed under the control of the 35S promoter in the pBA-DC-myc binary vector and was transferred into Arabidopsis for over-expression. The T0 lines of the PHYL1 plants showed a low ratio of leafy flower formation. Of the 774 individual T0 seedlings of PHYL1 plants, which all showed Basta resistance (a selection marker), only 46 showed a leafy flower phenotype (Table 1). In this case, the leafy flower formation rate is 5.9% in this transgenic approach (Table 1). A similar phenomenon was also observed in the T0 seedlings of SAP54 plants (data not shown), suggesting the existence of unclear mechanism(s) influence the effector-mediated leafy flower formation.
The leafy flower rates were 18.6% and 25.4% in TuGK-PHYL1-infected Col-0 and dcl2/4 plants, respectively, 3 to 5 times the ratio compared with the transgenic approach (Table 1). Similarly, leafy flower rates of 15.4 to 33.3% were found for TuGK-SAP54-infected Arabidopsis (Table 1). These data indicated that the PHYL1-and SAP54-mediated leafy flower formation rate could be enhanced through the TuGK expression approach. Moreover, the DCL2/ DCL4-dependent virus defense system might interfere with recombinant virus infection and result in a reduction in the expression of foreign genes [19]. Despite the results of TuGK--SAP54 infection, TuGK-PHYL1 triggered a leafy flower formation rate in dcl2/4 plants of 25.4%, whereas a formation rate of 18.6% was observed for TuGK-PHYL1-infected Col-0 plants ( Table 1), suggesting that TuGK-expressed PHYL1 in dcl2/4 plants can also enhance the phenotype formation.

Flower-related gene expression profiles in TuGK-PHYL1-and TuGK-SAP54-infected leafy flowers
To compare the gene expression profiles of Col-0 and SAP54 plants, we evaluated the expression levels of 3 flower-related genes (AtAP3, AtSVP, and AtFT) in flower tissues from Col-0 and SAP54 plants based on their transcriptome profiles (Fig 5). The AtAP3 and AtFT genes, which promote flowering, were down-regulated in SAP54 plants, whereas the flowering repressor AtSVP was up-regulated (Fig 5). Next, real-time PCR was used to evaluate AtAP3, AtSVP, and AtFT expression levels in TuGK-PHYL1-and TuGK-SAP54-infected Arabidopsis (Fig 5). AtUBQ10 is a housekeeping gene that has been used as a references gene for real-time RT-PCR normalization in many studies [28][29][30]. AtUBQ10 has 6 alternative splicing isoforms (S1 Table), and the transcriptome profile indicated no significant differences of the average RPKM of 6 isoforms (p-value of t-test is 0.314) between Col-0 and SAP54 plants (Fig 5D, left panel). Moreover, the real-time RT-PCR primer set for AtUBQ10 was designed based on the conserved region of the 6 isoforms, and the real-time RT-PCR results indicated that AtUBQ10 expression levels are consistent in virus-infected or uninfected Arabidopsis plants (Fig 5D, right panel). These results indicated that AtUBQ10 expression is stable under the various conditions and treatments in this study and that it can be used as a references gene for normalization of flower-related gene expression.
The expression levels of AtAP3 and AtFT were repressed in TuGK-PHYL1-infected or TuGK-SAP54-infected Arabidopsis and SAP54 plants compared to mock-infected plants, whereas AtSVP was up-regulated in these plants (Fig 5A, 5B and 5C; right panels). These realtime RT-PCR results are consistent with the deep sequencing results, indicating that the TuGK expression system is consistent with the transgenic approach.

Discussion
The lack of symptoms of TuGK-infected plants is helpful for phenotype observation Foreign gene expression by viral vectors in planta has an advantage in shortening the development time compared with transgenic plants. However, most viral vectors are constructed from viral strains that cause severe symptoms and affect the normal growth status of the plant host. Thus, it can be difficult to distinguish whether the phenotypes are caused by the expressed gene or plant immune responses. A deconstructed vector approach utilizes the minimum required viral elements for efficient expression; the missing functions can be provided by nonviral components to avoid symptoms. The deconstructed approach thus provides a solution for the symptom issue; however, the vector delivery system still relies on the agrobacterium-mediated transgenic method.
The HC-Pro R182K of TuGK is a mutant defective in miRNA pathway suppression, resulting in a lack of symptoms in infected plants [19]. Therefore, TuGK can be used to perform unambiguous gain-of-function analyses in infected plants. Currently, the high-throughput next-generation (NGS) technique provides a powerful method to analyze whole-transcriptome profiles. Hundreds to thousands of candidate genes can be identified via high-throughput network analysis [4], and an efficient way to verify the functions of these candidate genes is needed. The TRV vector with the Gateway recombinant system has been developed for efficiency in constructing VIGS sequences of interests [12]. Therefore, the Gateway cassette can be employed with the TuGK vector for future high-throughput gain-of-function analyses.
The stability of a heterologous gene in a viral vector is dependent on the whether the gene interferes with the viral infection pathway. In addition, losing the foreign gene is strongly depended on demographic conditions [31]. For instance, the GFP gene is more stable in the short-term passages of TuMV, whereas the losing GFP gene in the long-term passage [31]. We assume the PHYL1 and SAP54 effectors might cause side effect to interfere with TuMV infection, whereas the harmless GFP gene can be passed to many progenies by viral vectors [10].

TuGK can deliver effectors to the meristem and trigger leafy flower development
It is generally considered that the meristem of a plant is a cell division-active and virus-free region [32]. However, recent studies have demonstrated that several viruses, including Tobacco ringspot virus, Pepper ringspot virus, PVX, Odontoglossum ringspot virus, and Barley stripe mosaic virus, can infect the meristem region [32][33][34][35][36][37]. The confocal section data indicated that TuMV migrates to the meristem, suggesting that a TuGK-delivered foreign gene can affect host gene regulation in the meristem, resulting in altered morphology. Indeed, TuGK-PHYL1or TuGK-SAP54-infected Arabidopsis showed a leafy flower phenotype identical to PHYL1 and SAP54 plants.
Our pervious study demonstrated that PnWB-mediated leafy flowers in C. roseus plants, which show up-regulation of CrSVP1 and CrSVP2, but down-regulation of CrAP3.1, CrAP3.2, and CrFT [4]. The gene expression profiles in TuGK-PHYL1-and TuGK-SAP54-infected Arabidopsis plants indicated that AtAP1, AtSVP, and AtFT expressions are consistent as transgenic Arabidopsis and PnWB-infected C. roseus plants. Therefore, these data indicated that the TuGK vector can be used for gain-of-function studies. Moreover, the T0 lines of PHYL1 and SAP54 plants showed severe leafy flowers that did not produce seeds. TuGK-PHYL1 and TuGK-SAP54 infection of Arabidopsis provides sufficient experimental marital and can infect various mutant plants for further genetic analysis.
Furthermore, based on these data, we can immediately ask whether the PHYL1 or SAP54 effector acts similarly to the HC-Pro in silencing suppression or compensate the loss-of-miRNA suppression of HC-Pro R182K . Different pathogen effectors have been shown to exhibit synergistic effects upon co-infection [38]. However, TuGK-PHYL1-and TuGK-SAP54-infected Arabidopsis plants did not exhibit the severe yellow mosaic symptoms of TuMV, indicating that PHYL1 and SAP54 did not compensate for the loss-of-function of HC-Pro R182K .
The TuGK vector enhances the phenotype formation rate The leafy flower formation rate of the T0 lines of PHYL1 plants was 5.9% (Table 1) and a similar phenomenon was also observed for SAP54 plants (data not shown). Most of Basta-resistant T0 lines of PHYL1 or SAP54 plants produced normal flowers. We assume that the protein stability or another unclear mechanism affects PHYL1-or SAP54-mediated leafy flower formation. However, the phenotype formation rate of the TuGK-PHYL1-or TuGK-SAP54-infected plants was increased by 2.6 to 5.6 times compared with the PHYL1 or SAP54 plants. These data indicated that the TuGK vector enhances the phenotype formation.

Conclusions
In this study, we used the TuGK mild strain, which lacks the ability to suppress the miRNA pathway, as a vector for a gain-of-function study and found phenotypes and gene expression levels similar to those of transgenic plants. Moreover, the TuGK expression system serves as an additional strategy for effector expression to bypass the issue of lack of seeds with PHYL1 and SAP54 plants. Combination of the Gateway recombinant system with the TuGK vector will facilitate future cloning and efficient screening in gain-of-function studies. The mock-and TuGK-infected SAM sections were evaluated by confocal microscopy. Complete stacks demonstrate that GFP (green), and chlorophyll (red) were present in the SAM region. The SAM tissues (white dashed-line boxes) were corrected at 7 dpi of mock-or TuGKinfected Arabidopsis plants. Bar,200 μm. (TIF) S1 Table. The read counts and RPKM of 6 AtUBQ10 isoforms in Col-0 and SAP54 plants. (DOCX)