Genetic analysis of a Piezo-like protein suppressing systemic movement of plant viruses in Arabidopsis thaliana

As obligate intracellular phytopathogens, plant viruses must take advantage of hosts plasmodesmata and phloem vasculature for their local and long-distance transports to establish systemic infection in plants. In contrast to well-studied virus local transports, molecular mechanisms and related host genes governing virus systemic trafficking are far from being understood. Here, we performed a forward genetic screening to identify Arabidopsis thaliana mutants with enhanced susceptibility to a 2b-deleted mutant of cucumber mosaic virus (CMV-2aT∆2b). We found that an uncharacterized Piezo protein (AtPiezo), an ortholog of animal Piezo proteins with mechanosensitive (MS) cation channel activities, was required for inhibiting systemic infection of CMV-2aT∆2b and turnip mosaic virus tagged a green fluorescent protein (GFP) (TuMV-GFP). AtPiezo is induced by virus infection, especially in the petioles of rosette leaves. Thus, we for the first time demonstrate the biological function of Piezo proteins in plants, which might represent a common antiviral strategy because many monocot and dicot plant species have a single Piezo ortholog.

The mouse Piezo1 protein functions as a Ca 2+ permeable channel, regulating the volume of blood cell, integration of vascular architecture, differentiation of neural stem cells, and stretch induced cell proliferation [20][21][22][23] . Most plant species harbor a single Piezo protein 24 , which share high similarities with animal piezo proteins [24][25][26][27][28] . However, biological functions of plant Piezo orthologs are still poorly understood. In this study, the M2 population of ethyl methanesulfonate (EMS)-mutagenized Col-0 plants were used to screen enhanced susceptibility to CMV-2aTΔ2b (esc) mutants. We isolated a novel mutant, named esc1, which is more sensitive to the CMV-2aTΔ2b infection and TuMV-GFP than wild-type Col-0 plants. ESC1 encodes a Piezo-like protein, which shows high similarity with mammalian Piezo proteins. Further analysis revealed that ESC1 limited the systemic movement of plant viruses.

Results
Identification and characterization of esc1 mutant from forward genetic screening. To screen enhanced susceptibility to CMV (esc) mutants, we inoculated the M2 population of EMS-mutagenized Col-0 plants with CMV-2aTΔ2b. Consistent with our previous studies 12 , the 2b-deleted mutant CMV-2aTΔ2b did not cause visible symptom on wild-type Col-0 plants at either 14 days post inoculation (dpi) or 49 dpi due to

Plants Resistant Sensitive
Col-0 20 0 esc1-1 0 20 RDR1-and RDR6-dependent RNA silencing amplification (Fig. 1a). In contrast, the esc1-1 mutant developed severe systemic symptoms like curly leaves at 14 dpi and stunted growth at 49 dpi (Fig. 1a). Besides, the rosette leaves of esc1-1 were slightly smaller than that of Col-0 and esc1-1 plants were shorter than Col-0 (Fig. 1a). Western blotting and northern blotting analyses were further performed to assess virus accumulation in inoculated leaves and systemically infected leaves at 7-and 14-dpi, respectively. The esc1-1 mutant plants accumulated similar levels of viral CP in inoculated leaves but higher level of viral CP in systemically infected leaves compared with those of wild type Col-0 plants (Fig. 1b). Likewise, northern blotting hybridization also showed that esc1-1 mutant plants allowed an increased accumulation of viral genomic/subgenomic RNA in systemically infected leaves but maintained similar levels of viral RNAs in inoculated leaves compared to those of Col-0 plants (Fig. 1c).
To determine the relative accumulation of viral genome RNA3 and CP, the hybridization signal intensities were evaluated. The values of RNA3 and CP in virus-infected Col-0 plants were set as one unit. The analysis of systemically infected leaves revealed that the levels of CMV CP (Fig. 1b) and RNA3 (Fig. 1c) in the esc1-1 mutant plants were 7 times more than those of Col-0 plants. While, there is no significant difference in the inoculated leaves of esc1-1 mutant and Col-0 plants after infection of CMV-2aTΔ2b. Collectively, the systemic infection of CMV-2aTΔ2b was significantly enhanced in esc1-1 mutant plants.

Mapping and identification of ESC1.
In order to map esc1-1 locus by a positional cloning strategy, the esc1-1 mutant plants were crossed with wild-type Landsberg erecta (Ler) plants and the F2 population were developed. The resulting F2 population were further inoculated with CMV-2aTΔ2b and the segregation ratio was about 3:1 (resistant versus susceptible, Table 1). The F2 plants presented sensitive symptoms ( Supplementary  Fig. S1) were subjected to mapping-based cloning analysis with 34 simple sequence length polymorphisms (SSLPs). The low-resolution mapping showed that the mutation responsible for the esc1-1 mutant was mapped to the position between 19. 28 Fig. S1) were collected to a bulked DNA pool for deep sequencing by Illumina. In the region between 19.61 and 19.7 Mb of chromosome 2, there is only a G to A point mutation leading to a premature stop codon at W2248 in the 18th exon of At2g48060 ( Fig. 2b and Supplementary Table S1).
To verify whether the premature stop mutation of At2g48060 is responsible for the esc1-1 phenotypes, a knockout mutant (Salk_003005, named esc1-2) with a T-DNA insertion in the sixth intron of At2g48060 was obtained from Arabidopsis Biological Resource Center (ABRC) (Fig. 2b). The esc1-2 mutant, like the esc1-1 mutant, did not show abnormal developmental phenotypes (Fig. 2c). However, both esc1-1 and esc1-2 plants developed more severe disease symptoms after the CMV-2aTΔ2b infection compared with those of wild-type plants at 14-and 49-dpi (Fig. 2c). The F1 plants of a cross between esc1-1 and esc1-2 mutants exhibited similarly severe disease symptoms with those single mutants after the CMV-2aTΔ2b infection (Fig. 2c). Subsequently, northern blotting hybridization exhibited an increased virus accumulation in systemically infected leaves of the esc1-1, esc1-2, as well as their F1 progeny plants compared with Col-0 plants (Fig. 2d). Additionally, the single mutant of esc1-1 and esc1-2, as well as their F1 plants allowed higher accumulations of viral CP than Col-0 plants (Fig. 2e). These results indicate that esc1-1 and esc1-2 are indeed allelic in the gene of At2g48060.
We further examined whether the esc mutants had enhanced susceptibility to wild-type CMV. To this end, Col-0 plants, esc1-2, and esc1-3 mutants were challenged by CMV. As expected, the CMV-infected esc1-2 and esc1-3 mutants exhibited more severe symptoms than those of Col-0 (Fig. 3a). Western blotting analysis showed that similar accumulations of CMV CP were detected in inoculated leaves of Col-0 plants, esc1-2, and esc1-3 mutants at 4 dpi (Fig. 3b), whereas systemically infected leaves of esc1-2 and esc1-3 mutants allowed significantly higher levels of CP than those of Col-0 plants (Fig. 3c). In consistence, quantitative real-time PCR analysis showed that CMV RNA accumulated to similar levels in all the inoculated leaves, but to a higher levels in systemically infected leaves of esc1-2 and esc1-3 mutants than those of Col-0 plants (Fig. 3d,e).
Taken together, all the esc mutants exhibited enhanced susceptibility to wild type CMV and the CMV-2aTΔ2b mutant. In addition, At2g48060 is responsible for the ESC1-dependent viral immunity. and viral CP (h) in the systemically infected leaves of the plant samples shown in (f) at 14 dpi. Methylene bluestained rRNA were used as loading controls for total RNAs. In the d and g, the bottom values represent relative accumulation (RA) of the CMV RNA3. In the e and h, the bottom values represent RA of CMV CP. All the band intensities were evaluated by ImageJ and the RA values of Col-0 were set as 1. The RbcL was stained by Coomassie blue as protein loading controls. (2019) 9:3187 | https://doi.org/10.1038/s41598-019-39436-3 www.nature.com/scientificreports www.nature.com/scientificreports/ ESC1 is a putative plant Piezo ortholog. We cloned the cDNA of At2g48060 that encoded a protein of 2485 amino acid (aa) residues ( Supplementary Fig. S3). There is a 23 amino acid residues insertion behind the residue K403 compared with the predicted protein sequence of At2g48060 from the available TAIR annotation ( Supplementary Fig. S3, underlined and in bold). ESC1 contains 31 predicted transmembrane (TM) domains, and the point mutant of esc1-1 mutant plants is located on the outside domain between the 30th and 31th transmembrane domain (Fig. 4a). Blast searching showed that ESC1 shares a high similarity with mammalian Piezo proteins (Fig. 4b). Many plant species, including monocot and dicot plants, have a single Piezo ortholog (Fig. 4b). Therefore, ESC1 is renamed AtPiezo, which is predicted in the previous reports 24,29,30 .
The cryo-electron microscopy structure of the mouse Piezo1 and chimer assays showed that its C-terminal region was responsible for ion-permeation properties [31][32][33] . Multiple alignments of the AtPiezo protein and its orthologs in C. sativus, G. max, O. sativa and M. musculus show that those C-terminal regions contain evolutionarily conserved motifs, including Anchor, Outer helix, Cap, Inner Helix, and CTD domains (Fig. 4c). Furthermore, amino acid residues P2129-F2130(X2)-E2133(X6)-W2140 of the mouse Piezo1 are also conserved in plant Piezo orthologs (Fig. 4c). The premature stop codon at W2248 in esc1-1 plants results in deletion in the region Cap, Inner Helix, and CTD domains those are essential for ion-permeation properties [31][32][33] . Therefore, the AtPiezo of esc1-1 plants is probably a loss-of-function mutation of ion channel pore properties. tissue expression pattern of Atpiezo. To investigate the spatial and temporal activities of the AtPiezo promoter, we analyzed β-glucuronidase (GUS) staining in tissues of transgenic Col-0 plants with the AtPiezo native promoter fused with a GUS reporter gene (Piezo Pro ::GUS). GUS activity was detected in root, leaf petioles, flower petals and silique stalks, especially in vascular system of leaf petioles (Fig. 5a-e). In the mock (buffer)-treated plants, Piezo Pro ::GUS was expressed mainly in the shoot meristem and the upper part of hypocotyl with a very weak level (Fig. 5f, left panel). In the CMV-2aTΔ2b-infected plants, Piezo Pro ::GUS was expressed in a similar pattern but in a more expanded regions, especially in petioles of rosette leaves (Fig. 5f, right panel). In addition, the induction of AtPiezo in petioles of inoculated leaves by CMV-2aTΔ2b infection was revealed by quantitative real-time PCR (Fig. 5g). Collectively, our findings demonstrate that AtPiezo is transcriptionally induced by virus infection. To monitor the TuMV infection, we used an engineered strain with a GFP insertion (TuMV-GFP) and measured GFP fluorescence of infected plants at different time points after inoculation as described previously 19 . At 6 dpi, the infected esc1-2 and esc1-3 mutant plants exhibited obvious GFP fluorescence in the systemic full-expanded leaves, whereas only inoculated leaves rather than systemic leaves of Col-0 plants accumulated GFP fluorescence (Fig. 6a,  left panels). Subsequently, the infected esc1-2 and esc1-3 mutant plants exhibited higher fluorescence intensity than wild-type plants at 7-, 8-, and 9-dpi (Fig. 6a). Of infected wild-type plants (n = 30), 29.7% were systemically infected by 7 dpi. By contrast, 72.2% of esc1-2 and 71.3% of esc1-3 mutant plants were systemically infected by TuMV-GFP by 7 dpi (Fig. 6b). By 9 dpi, almost all the infected plants exhibited GFP fluorescence in systemic leaves (Fig. 6b). However, the numbers of fully infected leaves of esc1-2 and esc1-3 mutant plants were greater than those of wild-type Col-0 plants at 9 dpi (Fig. 6a).
Western blotting was further performed to demonstrate that the GFP accumulation was increased in esc1-2 and esc1-3 mutant plants compared with Col-0 plants at 7 dpi (Fig. 6c). Moreover, quantitative real-time PCR analysis was conducted to examine accumulation of TuMV-GFP genomic RNAs in the upper systemically infected leaves, which showed that a significant increase of TuMV-GFP mRNA in esc1-2 and esc1-3 mutant plants than in wild-type Col-0 plants at 7 dpi (Fig. 6d). Similarly, esc1-1 mutant plants also exhibited enhanced www.nature.com/scientificreports www.nature.com/scientificreports/ susceptibility to infection of TuMV-GFP ( Supplementary Fig. S5). Meanwhile, TuMV-GFP RNA accumulations in the protoplasts isolated from esc1-1 and esc1-2 mutant plants were comparable to that in those of Col-0 plants (Fig. 6e). Collectively, these results suggested that AtPiezo plays a negative role in the systemic infection but not in the intracellular replication of TuMV-GFP.
The MCA protein family have MS channel activities to mediate mechanical signal-induced Ca 2+ transport in regulation of cell growth and development 29,30 . To determine whether the MCA protein family have antiviral activities, the mca1, mca2 single mutant plants, as well as mac1 mac2 double mutant plants were confirmed to be null alleles and inoculated with TuMV-GFP ( Supplementary Fig. S6). At 7 dpi, GFP fluorescence in the mca1, mca2 single mutant plants, as well as mca1 mca2 double mutant plants was comparable to wild-type Col-0 plants, while esc1-2 plants exhibited enhanced susceptibility to infection of TuMV-GFP (Supplementary Fig. S6). These results indicate that MCA1 and MCA2, unlike AtPiezo protein, have no observable effects on systemic infections of TuMV-GFP.

Discussion
Plants sense and respond to a wide variety of mechanical stimulation by mechanosensitive (MS) ion channels that provide pathways for ions movement from one side to the other side of the channels-targeted membranes. To date, three protein families have been proposed to be MS ion channels in A. thaliana including MSL, MCA, and Piezo families, although the molecular basis for these MS ion channels remain to be revealed. In contrast with MSL and MCA families, studies of which have gained breakthroughs in recent years 29,30,34,35 , plant Piezo proteins and their biological functions are largely unknown except their similarity with the well-studied animal Piezo orthologs 28,36 . Here, we took advantage of forward genetic screening to provide the first insight into the biological functions of plant Piezo orthologs in negatively regulating virus systemic spread (Figs 1 and 2). All the mutants of AtPiezo, including the premature stop mutant (esc1-1), the T-DNA inserted line (esc1-2), as well as two CRISPR/ Cas9 mutants (esc1-3 and esc1-4), exhibited enhanced susceptibility to CMV-2aTΔ2b infection in the systemic leaves but remained unchanged in inoculated leaves (Fig. 1b), indicating AtPiezo limits viral systemic infection (Figs 1, 3 and 6). These results demonstrate that the potential MS ion channel AtPiezo plays a role in restriction of systemic spread of a couple of viruses. www.nature.com/scientificreports www.nature.com/scientificreports/ AtPiezo accumulated to very low level in mock plants (Fig. 5), which is probably responsible for the non-discernable developmental defect in mutant plants (Figs 1 and 2). Therefore, AtPiezo probably plays non-essential roles in plant development, which is the reason why the functions of plant Piezo have not been revealed in the previous studies. Nonetheless, AtPiezo is induced after virus infection suggesting that AtPiezo might be involved in virus immunity pathway. Besides, tissue-specific expression of virus-induced AtPiezo in the plant vasculature demonstrates that the negative effect of AtPiezo on viral systemic movement should occur in the vascular bundles. Indeed, induction of AtPiezo by virus infection in those petioles of rosette leaves probably impedes viral translocation and/or movement in phloem systems.  the (b,d,e), the error bars represent SD from three independent experiments. In each bar chart, letters above bars indicate statistically significant differences according to the Duncan's multiple range test. P < 0.05 was considered significant differences and P < 0.01 was considered extremely significant differences.
www.nature.com/scientificreports www.nature.com/scientificreports/ So far, only a few genes were identified in blocking virus long-distance movement. A cadmium-induced glycine-rich protein (cdiGRP) and a cdiGRP-interacting protein, named GrIP, restricts virus long-distance trafficking through improving callose deposition in phloem cells 3,4 . Genetic characterization of natural Arabidopsis accessions and mutants were employed to identify host genes involved in restricting tobacco etch virus movement (RTM). The results showed that at least five RTM genes were involved in defense against virus systemic infection [37][38][39][40][41] . In addition, the down-regulation of the subunit RPN9 of the 26S proteasome changes vascular development leading to inhibition of virus systemic infection 42 . Here, we used genetic analysis to show that AtPiezo suppresses systemic movement of plant viruses in Arabidopsis thaliana. Unfortunately, because two pairs of repeat sequences usually lead to a deletion from 1326 bp to 3315 bp of AtPiezo cDNA by intramolecular homologous recombination ( Supplementary Fig. S4), we could not obtain the full-length cDNA of AtPiezo for biological and functional analysis.
Methods plant materials and growth conditions. Seeds of A. thaliana wild type in ecotype Columbia-0 (Col-0) and Lersberg erecta (Ler), as well as the T-DNA insertion mutant esc1-2 (Salk_003005) were obtained from the Arabidopsis Biological Resource Center (ABRC, Columbus, OH). The esc1-3 and esc1-4 mutants were generated using a CRISPR/Cas9-based method as described previously 43 . Primers for the generation of CRISPR/Cas9 constructs are listed in Supplementary Table S2. A. thaliana seeds were sterilized by 10% bleach, plated on Murashige and Skoog (MS) medium supplemented with 3% sucrose and 1% agar, and vernalized at 4 °C for 3 days. Then, these plates were incubated in climate-controlled chambers at 22 °C for 7-10 days under a long day condition (16 h light/8 h dark). Seedlings were transferred into soil and grown in a growth room at 22 °C to 24 °C with a 10/14 h in light/dark program. To genotype the T-DNA insertion mutant, genomic DNA and RNA were extracted from the T-DNA insertion mutant leaves then PCR and RT-PCR were performed with gene-specific primers as listed in Supplementary Table S2.

Virus inoculation.
Four-week-old A. thaliana plants were mechanically inoculated with CMV-2aTΔ2b or CMV virions at the concentration of 50 μg/mL or 20 μg/mL as described previously 44 . The disease symptoms were recorded and virus CP and RNA were detected. N. benthamiana leaves infiltrated with TuMV-GFP were ground in 10 mM phosphate buffer with 2% PVP-40, pH 7.4, then the extract was used to mechanically inoculate 5-week-old A. thaliana plants as described previously 45 . The GFP fluorescence of plants inoculated with TuMV-GFP was photographed using a digital camera equipped with a yellow filter under a long wave length UV lamp (UVP, Upland, CA, USA) as described previously 46 . Forward genetic screen, map-based cloning and sNp analysis using next-generation sequencing. EMS mutagenesis in Col-0 background was performed as described previously 47 . Briefly, about 50,000 M1 seeds of Col-0 plants were incubated with 100 mM phosphate buffer (pH 7.5) 4 °C for 12 h, followed by mutagenized with 0.6% ethyl methanesulfonate (EMS) in phosphate buffer at 25 °C for 8 h. The M1 seeds were washed with sterilized water for 20 times and were grown in soil to generate the M2 population. The esc1-1 was identified through screening susceptible mutants of the M2 populations after infection of CMV-2aTΔ2b. To identify the ESC1 gene, esc1-1 mutant plants were crossed with wild type Ler plants, and the resulting F1 plants were self-fertilized to generate F2 mapping populations. Genomic DNAs isolated from 310 susceptible F2 plants and 34 SSLP markers were used to narrow down the susceptible locus. Genomic DNA extracted from about 100 susceptible F2 plants generated from backcrossing esc1-1 with Col-0 plants were utilized for SNP analysis by next-generation sequencing with Illumina HiSeq2000 platform. The wild type Col-0 genomic DNA was sequenced as a reference. Alignment of SNP analysis and reference DNA sequence was performed using SAM tools 48 . The SNP sites was visualized and confirmed by the Integrative Genome Viewer 49 . phylogenetic analysis. To identify the coding region sequence of AtPiezo (At2g48060), three parts of the full CDS were amplified and inserted into pMD19T (Takara). The sequences were amplified with high fidelity DNA polymerase and all plasmids were verified by DNA sequencing. Primers used are listed in Supplementary  Table S2. Protein sequences of other Piezo proteins were downloaded from the NCBI database (www.ncbi.nlm. nih.gov). Multiple alignment of amino acid sequences was performed with ClustalW 50 . The phylogenetic tree was generated using MEGA5 software implementing a neighbor joining method 51 . Phylogeny was indicated using the bootstrap method on 1000 replications.

GUs staining.
To clone the promoter of AtPiezo, the genomic region between 2.8 kb upstream and 45 bp downstream of the start codon of AtPiezo was introduced into the pBI101 vector (Clontech). The resulting plasmid containing Piezo Pro ::GUS was transformed into Col-0 plants by the Agrobacterium-mediated floral-dipping method. In vivo GUS staining was carried out as described previously 14 . Two independent lines of Piezo Pro ::GUS transgenic plants were examined for GUS expression at 2 days after treated by buffer or CMV-2aTΔ2b. Plant tissues were incubated with X-Gluc solution (50 mM NaPO4 buffer, pH 7.0, 0.1% Triton X-100, 10 mM potassium ferrocyanide, 10 mM potassium ferricyanide, and 3 mM 5-bromo-4-chloro-3-indolyl-β-D-glucuronide) at 37 °C for 14 h. Stained tissues were washed in an ethanol series.
Quantitative real-time pCR. Total RNA was extracted from CMV or TuMV-GFP-infected systemic leaves of plants with TRIzol reagent (Invitrogen). After treated with RNase-free DNaseI (Takara), total RNA was used for the reverse transcription into cDNAs (MLV, Promega). To estimate the CMV CP mRNA and TuMV-GFP genomic RNA accumulation level, quantitative real-time PCR was performed with CMV CP and TuMV CP specific primers (Supplementary Table S2) using SsoFastTM EvaGreen ® Supermix according to the