Function identification of miR482b, a negative regulator during tomato resistance to Phytophthora infestans

Tomato is an important horticultural and economic crop cultivated worldwide. As Phytophthora infestans becomes a huge threat to tomato production, it is necessary to study the resistance mechanisms of tomato against P. infestans. Our previous research has found that miR482 might be involved in tomato–P. infestans interaction. In this study, miR482b precursor was cloned from Solanum pimpinellifolium “L3708” and miR482b was shown to decrease in abundance in tomato following P. infestans infection. Compared to wild-type tomato plants, tomato plants that overexpressed miR482b displayed more serious disease symptoms after P. infestans infection, with more necrotic cells, longer lesion diameters, and increased P. infestans abundance. Meanwhile, silencing of miR482b was performed by short tandem target mimic (STTM), resulting in enhancement of tomato resistance to P. infestans. Using miRNA and degradome data sets, NBS–LRR disease-resistance genes targeted by miR482b were validated. Negative correlation between the expression of miR482b and its target genes was found in all miR482b-overexpressing and -silencing tomato plants. Our results provide insight into tomato miR482b involved in the response to P. infestans infection, and demonstrate that miR482b–NBS–LRR is an important component in the network of tomato–P. infestans interaction.


Introduction
Tomato, as an important horticultural and economic crop cultivated worldwide, always suffers from the aggression of various pathogens [1][2][3] . Phytophthora infestans, the agent that causes late blight, has become one of the most devastated pathogens 4 and can cause great losses in tomato production 5 . For example, in Inner Mongolia of China and the USA, late blight caused the loss of tomato production in early 2000s 6,7 . Up to now, the molecular mechanisms of tomato-P. infestans interaction are beyond understanding and the method for controlling tomato late blight is still not effective enough. It is necessary to study tomato resistance mechanisms against P. infestans and identify the key resistance genes which are used in disease-resistance breeding of tomato transgene.
MicroRNA (miRNA), a class of about 22 nt endogenous small noncoding RNA, have been discovered to play an important role in regulating endogenous gene expressions in plants 8,9 . MiRNAs attack their targets by sequencespecific binding to the targeting sites on the gene transcripts, leading to either target mRNA degradation or translational repression mediated by the miRNAassociated proteins 10 . Up to now, the major methods to study miRNA functions are to generate and analyze the transgenic lines overexpressing miRNAs 11,12 . For example, in our previous studies, overexpression of miR172 in tomato enhanced its susceptibility to P. infestans 13 . Meanwhile, the tomato resistance to P. infestans was reduced by the overexpression of miR1918 14 . However, these approaches are insufficient to fully understand miRNA functions. Short tandem target mimic (STTM), which was invented based on an endogenous mechanism that modulates miR399 activity in Arabidopsis thaliana 15 , is a newly developed approach to explore miRNA functions by blocking miRNAs in plants 16 . In previous studies, STTM had been used to silence miR165/166, miR160, and miR159 in A. thaliana, cotton, soybean, tobacco and these miRNAs were found functioning as regulators of plant growth and development [17][18][19][20][21] . In tomato, blocking miR858 by overexpressing STTM858 resulted in high accumulation of anthocyanins, which demonstrated miR858 could negatively control the anthocyanin accumulation of tomato 22 .
Many miRNAs have been demonstrated to play important roles in stress responses. A number of plant miRNAs, such as miR159, miR160, miR166, miR169, miR172, and miR396, are involved in the response to drought, water deficit, and salt stresses 23,24 . More and more evidence have indicated that plant miRNAs also play an important role in plant-pathogen interaction 25,26 . For example, miR393 and miR472 can contribute to pathogen-associated molecular pattern-triggered immunity (PTI) in plants 27,28 . Other miRNAs, such as miR482, miR6019, miR6020, miR6022, miR5300, miR1507, and others, are thought to suppress nucleotide-binding site and leucine-rich repeat (NBS-LRR) defense genes, which play critical roles in effector-triggered immunity (ETI), also a crucial component of plant immune system [29][30][31][32][33] .
Using high-throughput sequencing and homologybased computational research, we have previously identified a number of tomato miRNAs involved in tomato-P. infestans interaction including miR482, miR172, miR6024, miR6026, miR6027, etc. 13,25,47,48 . Our previous work showed that only miR482b was identified in tomato, Solanum pimpinellifolium "L3708" (L3708) inoculated with P. infestans and its transcripts per million clean tags was lower than wild-type (WT) tomato after analysis of the miRNA-Seq data. However, whether miR482b affected tomato resistance to P. infestans has not been determined. In this study, we analyzed the expression patterns of miR482b in tomato L3708 after P. infestans treatment. Then the function of miR482b was explored by overexpressing and silencing approach and its target genes were also identified. These studies will benefit not only understanding of the molecular mechanisms of tomato-P. infestans interaction but also future molecular breeding.

Materials and methods
Plant material collection and P. infestans inoculation Tomato L3708 was grown in a greenhouse at 25 ± 3°C under 16 h of light per day. P. infestans strain "P12103" was cultured in oat medium in the dark at 20°C. Tomato plants at the five to six leaf stage were inoculated with P. infestans spores suspension (10 6 zoospores mL −1 ) and placed at 20 ± 1°C in a 100% relative humidity environment without light. Tomato plants treated with sterile water were used as controls, and they were kept under the same conditions. Leaf samples were collected at 0, 1, 2, 3, and 4 days post inoculation (dpi). All samples were quickly frozen in liquid nitrogen and stored at −80°C until DNA and RNA isolation.

Cloning of pre-miR482b
MiR482b precursor (pre-miR482b) was cloned using tomato DNA as a template. The primers (miR482bF and miR482bR; Table S2) were designed according to the tomato genome sequences using the Primer Premier 5 software. The PCR product was cloned into T-Vector pMD 19 (Takara, Dalian, China), and the identity of the sequence was verified by DNA sequencing. Multiple nucleic acid sequence alignments were performed with ClustalX2.

Construction of silencing and overexpressing plasmid of miR482b and their transient overexpression in tomato
Silencing of miR482b with STTM was performed. The structure of STTM482b was design based on the description of Yan et al. 17 . It contained two copies of imperfect miR482b binding sites (22 nucleotides) linked with a 48 nt linker. A bulge contain three additional nucleotides (CTA) was designed around positions 10 to 11 of the miRNA to trap miR482b without being cleaved by it. The pre-miR482b and STTM482b inserts were then subcloned into BamHI-SacI cut pBI121 to generate the recombinant overexpressing plasmids pBI121-miR482b and silencing plasmid pBI121-STTM482b, in which the expression of miR482b and STTM482b are controlled by the strong constitutive CaMV35S promoter. Freeze-thaw method was used to introduce pBI121-miR482b and pBI121-STTM482b into Agrobacterium tumefaciens strain "GV3101".
A sample from A. tumefaciens culture was centrifuged at 4000 × g for 10 min and the pellet was resuspended in infiltration medium (10 mM MgCl 2 , 10 mM MES and 20 μM acetosyringone; OD 600 = 1.0) 49 . A. tumefaciens was introduced into tomato leaves by infiltration. A. tumefaciens with empty vector was used as a control. After 3 days, each infiltrated leaf region was inoculated with 20 μL of P. infestans (10 6 zoospores mL −1 ). The lesions were observed at the 5th day and the length of the lesion diameter was calculated.

Transformation and identification of transgenic tomato
Tomato transformation was performed based on the method of Li et al. 50 . The seeds of tomato S. lycopersicum "Zaofen No. 2" were cultured in 1/2 Murashige and Skoog (MS) medium at 25 ± 3°C under 16 h of light per day after surface sterilized with 75% (v/v) ethanol and 2.5% (v/v) sodium hypochlorite. Cotyledon excised from 1-week-old seedlings were used as explants. Transgenic tomato plants were generated by A. tumefaciens-mediated leaf disk method. Putative transgenic plants were selected on 1/2 MS medium with 200 mg L −1 carbenicillin and 50 mg L −1 kanamycin. After obtaining T 0 kanamycin-resistant plants, the presence of the transgene in the regenerating plantlets was further confirmed using PCR with a pair of specific primers of neomycin phosphoryltransferase II (nptII) ( Table S2). The abundance of miR482b in the selected positive overexpressing and silencing lines were examined by quantitative reverse transcriptionpolymerase chain reaction (qRT-PCR).

Analysis of transgenic tomato resistance against P. infestans
The detached leaves of WT and transgenic tomato plants were inoculated with 10 μL of 10 6 zoospore mL −1 zoospore suspension of P. infestans according to the method of Li et al. 50 . In whole-plant assays, a suspension of P. infestans zoospores (10 6 zoospores mL −1 ) was also used to spray plants to run off. The inoculated leaves or whole plants were maintained at 20 ± 1°C under high humidity condition for 24 h without light, and then placed in a greenhouse at 20 ± 1°C under 16 h of light per day. Necrotic regions and disease indices were recorded at 5 dpi. UV photographs were also taken at 5 dpi. Images of the leaves were taken and the diameters of lesions were measured. Trypan blue staining assays used to detect dead cells of harvested leaves was also based on the method described by Li et al. 50 . In addition, the transcript level of P. infestans Actin gene was used to indicate P. infestans growth in plant by qRT-PCR 51 . Categorization of disease grades (DG) was performed as described by Li et al. 50 , and the disease index (DI) was calculated according to the following formula: where DG i is the value of DG, n i is the number of plants in each DG, and n is the total number of plants. Each experiment was carried out at least three times.
The leaves were also collected to measure the content of malonaldehyde (MDA) according to previously described method 50 .

Identification of target genes
The mature-miR482b sequence was input into psRNATarget (http://plantgrn.noble.org/psRNAtarget/) and WMD3 (http://wmd3.weigelworld.org/cgi-bin/ webapp.cgi) to predict its target genes in tomato. Tomato transcripts (cDNA library, version 2.4/2.3) were used as data sets. The degradome sequencing library of tomato inoculated with P. infestans was used to verify the targets. CleaveLand software was used to find the cleavage site of the miR482b targets 52,53 . T-plot figures were produced based on R language 54 .

qRT-PCR analysis
The abundance of miR482b in tomato was examined by qRT-PCR. The specific forward primer for miR482b was designed following the method described by Varkonyi-Gasic et al. 55 . Quantitative RT-PCR was performed using the SYBR PrimeScript miRNA RT-PCR kit (Takara, Dalian, China) with Corbett Rotor Gene 3000 qRT-PCR machine (Corbett Research, Mortlake, Vic., Australia). The primers for the target genes were designed following the manufacturer's instructions for the SYBR Premix Ex Taq II kit (Takara, Dalian, China). The miR482b cleavage sites of target genes were ensured between the forward and reverse primers binding sites. Information on all primers are shown in Table S2. Actin was used as a reference control gene for the miRNA and target gene qRT-PCR analysis. Of the nine leaves sampled in each experiment, three leaves were pooled into one biological replicate, resulting in three biological replicates.

Statistical analysis
All statistical analysis of data was performed with Origin 8.5, and all data were expressed as means ± SEs from three independent experiments. Based on http://abacus.bates. edu/~ganderso/biology/resources/writing/HTW_Guide_ Reporting_Statistics_3-7-2011.pdf, the capital letters were used over the error bars. Letters shared in common between or among the groups would indicate no significant difference.

Results
Cloning, tissue-specific expression and expression patterns of tomato miR482b under P. infestans infection The pre-miR482b was cloned from tomato L3708 based on tomato sRNAs available in the miRBase and tomato genome. The result of multiple nucleic acid sequence alignments showed that the sequence of the pre-miR482b and mature-miR482b were identical to the miR482b from cultivated tomato S. lycopersicum in miRbase ( Figure S1).
To analyze the distribution of miR482b within tomato L3708, qRT-PCR was used to measure the levels of miR482b transcript in the leaf, stem, and root of the plant. MiR482b was found in high abundance in leaf, whereas relatively low abundance was detected in root (Fig. 1a). To verify whether miR482b was involved in the response to P. infestans infection, the expression of miR482b in P. infestans-treated leaves at different dpi was measured. The result showed that after P. infestans treatment, the expression level of miR482b was lowest one day after the infection, but increased to a moderate level four days after the infection (Fig. 1b).

Overexpression and silencing of miR482b by agrobacterium infiltration
STTM482b sequence was constructed to silence miR482b, which contained two copies of imperfect miR482b binding sites with a 48 nt linker and each cope had a cleavage-preventive bulge contained three additional nucleotides (CTA) (Fig. 2a). Then the overexpressing and silencing plasmids, pBI121-miR482b and pBI121-STTM482b were constructed (Fig. 2b) and introduced into tomato leaf by agrobacterium infiltration. The abundance of miR482b was significantly increased in tomato plants that transiently overexpressed miR482b (TO482b) and significantly decreased in tomato plants transiently silenced miR482b (TS482b) (Fig. 2c). Compared to control plants, TO482b tomatoes displayed more severe disease condition following infection of P. infestans, with significantly larger lesions in the leaves, while TS482b tomato showed less severe disease condition after P. infestans infection (Fig. 2d, e). These results suggested that miR482b may have weakened the resistance of tomato against P. infestans.

Identification of transgenic plants with high and low miR482b expression levels
The transgenic tomatoes were constructed using pBI121-miR482b and pBI121-STTM482, respectively. Four-week-old tomato seedlings (WT and representative positive transgenic lines) were transferred to pots and grown for 2 weeks before they were subjected to qRT-PCR analysis to estimate the abundance of miR482b in their leaf tissue. MiR482b was more abundant in all overexpressing lines and less abundant in all silencing lines compared to WT lines (Fig. 3a, b) with the expression in three overexpressing lines (OE1, OE2, and OE3) being~8-fold, 7.4-fold, and 7.8-fold, respectively, the expression in WT, while in three silencing lines (ST1, ST2, and ST3) the expression were 0.32-fold, 0.28-fold, 0.36-fold, respectively, the expression in WT. In addition, after genomic DNA-PCR detection based on a pair of specific primers of nptII, transgenic lines exhibited the expected transgene-specific band (395 bp), WT was not found ( Figure S2).

MiR482b enhances tomato susceptibility to P. infestans
After the inoculation of P. infestans, the detached leaves were gradually darkened and the disease symptoms a Tissue-specific expression of miR482b in tomato root, stem, and leaf. b Expression pattern of miR482b in tomato leaves at the indicated time after being sprayed with P. infestans. Data are the means ± SEs of three independent experiments. Letters indicate significant difference among samples, and letters shared in common between or among the groups indicate no significant difference at the P < 0.05 level gradually appeared. Compared to WT, the disease symptoms were more severe on the leaves of miR482boverexpressing lines and less severe on the leaves of miR482b-silencing lines (Fig. 3c). The dead cells of the treated leaves were examined by trypan blue staining. The darker blue marks in the leaves of overexpressing plants indicated that a greater number of dead cells occurred in miR482b-overexpressing leaves than WT, while the dark blue marks were reduced after miR482b silenced which demonstrated fewer dead cells were existed in miR482bsilencing leaves compared to WT (Fig. 3c). Furthermore, lesion diameters of miR482b-overexpressing and miR482b-silencing leaves were significantly longer and shorter than WT leaves, respectively (Fig. 3d).
In the whole-plant inoculation assay, the DI was found to be higher in miR482b-overexpressing lines than WT, which indicated that blight lesions obviously occupied more leaf area in tomato plants overexpressed miR482b. Conversely, compared to WT, the DI of silencing lines was significantly reduced (Fig. 3e). The transcript levels of P. infestans Actin gene measured by qRT-PCR were used to indicate P. infestans growth in plant and the result showed a significant increase and decrease in abundance of P. infestans in overexpressing and silencing lines compared to WT, respectively (Fig. 3f). In plant-pathogen interaction, plant cellular membranes were damaged because of late massive ROS generation 56 . MDA, which is taken as an indicator of membrane damage 50 , was detected in the overexpressing, silencing and WT plants inoculated with P. infestans. A significantly higher MDA level was found in overexpressing lines, suggesting that these plants suffered more severe membrane damage than WT (Fig. 3g). In silencing lines, the lower MDA level suggested that they suffer less severe membrane damage compared to WT plants (Fig. 3g). To some extent, these results also suggested that miR482b had led to the enhancement of tomato susceptibility to P. infestans.

Identification of miR482b targets
Identification of target genes is essential to show the regulatory networks of miRNA. Prediction of miR482b target genes using psRNATarget and WMD3 showed 30 NBS-LRR disease-resistance genes (Table S1). Four of these genes, Solyc02g036270.2, Solyc04g009070.1, Solyc12g016220.2, and Solyc05g008070.2, were found to have been cleaved at the miR482b-specific cleavage site according to the available tomato degradome sequencing data, indicating that they were the target genes of miR482b (Fig. 4a). Overexpression of miR482b led to significant reduction in the expression of its four target genes (Fig. 4b). On the contrary, these four target genes was upregulatedly expressed in miR482b-silencing tomatoes (Fig. 4c).
Upon infection by P. infestans, the expression levels of Solyc02g036270.2, Solyc04g009070.1, and Solyc12g016220.2 were initially increased, but then decreased (Fig. 5). Their expression trend was in contrast to that of miR482b, which further verified the relationship between miR482b and the three genes. However, the expression level of Solyc05g008070.2 was decreased at the first day after infection by P. infestans, but remained constant for the next two days, and then increased (Fig. 5), suggesting a lack of correlation with the expression trend of miR482b. According to the tomato degradome sequencing data, Solyc05g008070.2 was also found to have been cleaved by miR6024, which might be the reason why there was no correlation between the expression of miR482b and Solyc05g008070.2. These results therefore f Transcript accumulation of P. infestans Actin gene in these inoculated plants. g MDA content of the leaves from WT, and overexpressing and silencing tomato lines at 5 days after inoculation with P. infestans. Data are the means ± SEs from three independent experiments. Letters indicate significant difference among samples, and letters shared in common between or among the groups indicate no significant difference (P < 0.05) indicated that miR482b might play an important role in the regulation of NBS-LRR genes during tomato-P. infestans interaction.

Discussion
To gain insight into the molecular mechanisms responsible for the late blight resistance, we previously used high-throughput sequencing to identify P. infestansinduced miRNAs in tomato, including miR169, miR398, miR482, miR6024, miR6026, and miR6027 25,47,48 . Among these miRNA candidates, miR482 is one of the most important miRNAs involved in plant-pathogen interaction. In this study, we showed that the abundance of miR482b in tomato was reduced following P. infestans infection (Fig. 1b). Similar results have also been reported by other investigators, who found that decreased abundance of miR482 occurs in Solanaceae upon infection with different pathogens, including V. dahlia, Fusarium oxysporum, etc. 30,31 . Recently, it was found that RNA silencing suppressors from pathogens might suppress the accumulation of miRNA in plant-pathogen interaction 57 . For example, two effectors form the soybean pathogen P. sojae, named PsPSR1 and PsPSR2 suppressed transgene silencing by inhibiting the accumulation of plant sRNAs 58,59 . In addition, the accumulation of miRNA also was regulated by lncRNAs and transcription factors. For example, in tomato-TYLCV interaction, a tomato lncRNA, slylnc0195 acted as a target mimic to suppress the expression of miR166a 60 . Meanwhile, ripening inhibitor (RIN), a vital transcription factor acted though sequence-specific interactions with cis-regulatory DNA elements in the promoter of MIR172a to regulate its accumulation 61 .
Tomato L3708 is a resistant tomato to P. infestans. But, the miR482b sequences in tomato L3708 and cultivated tomato showed 100% sequence identity ( Figure S1). To Various letters indicate significant difference among samples, and letters shared in common between or among the groups indicate no significant difference (P < 0.05) study the function of miR482b in tomato-P. infestans interaction, we introduced miR482b into cultivated tomato to constructed transgenic tomato. Three independent transgenic tomato lines were generated, which increased transcript amounts of miR482b. After the infection of P. infestans, more serious disease symptoms occurred in tomatoes overexpressed miR482b than WT, which suggested miR482b could enhance the susceptibility of tomato to P. infestans infection (Fig. 3). Similar results have also been reported in potato. The overexpression of miR482e in potato led plants to be sensitive to V. dahliae infection 31 . These results implicate a role for miR482b in the defense response of tomato against P. infestans infection.
As overexpression is not enough to fully understand miRNA functions and some consequence from overexpression studies may be caused by temporal or spatial misexpression 62 , miRNA-silencing technology should be used to explore functions of miRNAs. STTM, an effective tools to restrain miRNA activity, make it technically possible to silence miRNAs in various plants. In rice, transgenic STTM lines silencing 35 miRNA families (miR398, miR172, miR156, etc.) were produced as resources for functional studies and crop improvement 62 . STTM was also be used in tomato to silence miR396, which resulted in tomato flowers, sepals and fruits all obviously becoming bigger 63 . Meanwhile, effectively inhibiting the expression of miR858 by STTM induced anthocyanin accumulation in tomato 22 . In present study, we used STTM to silence miR482b to validate its function. In miR482-silencing lines, with the abundance of miR482b reduced significantly, plant resistance to P. infestans enhanced observably, which further proved miR482b could enhance the susceptibility of tomato to P. infestans.
In this study, four NBS-LRR genes were identified as the target genes of miR482b by degradome sequencing and qRT-PCR (Fig. 4). All these target genes showed more than 95% sequence identity and their binding regions of miR482b are completely identical between tomato L3708 and cultivated tomato. For example, the sequence identity of the target gene, Solyc02g036270.2 is 99% in these two tomato source (unpublished). That is to say, miR482b cleaves the same miR482b-specific cleavage site in tomato L3708 and cultivated tomato. NBS-LRR proteins have been shown to function as immune receptors that recognize specific pathogen effector proteins during plant-pathogen interaction, leading to the initiation of plant ETI 25,28,64 . The simplest model for its functioning is that NBS-LRR proteins recognize their corresponding effectors from various pathogens, and their interaction would lead to the activation of a series of defense responses 65 . Before the attack of a pathogen, the expression level of NBS-LRR genes is very low, but would rapidly increase at the stage of infection [66][67][68] . For example, VaRGA1, a grapevine NBS-LRR gene, is strongly induced by Plasmopara viticola at 12 hpi 65 . The expression of a novel peanut NBS-LRR gene, AhRRS5, is also increased in response to R. solanacearum infection, both in resistant and susceptible peanut cultivars 69 . Similar to these previous studies, the expression of Solyc02g036270.2, Solyc04g009070.1, and Solyc12g016220.2, three of the four NBS-LRRs genes identified as target genes of miR482b, were also elevated after P. infestans infection, which clearly demonstrated that these genes could respond to the infection of P. infestans in tomato (Fig. 5). In addition, previous studies also showed that overexpression of the NBS-LRR gene VaRGA1 in Nicotiana benthamiana would enhance its resistance to P. parasitica 65 , whereas silencing of a CC-NBS-LRR gene through VIGS in cotton dramatically increased its susceptibility to V. dahliae 70 . Similar results were obtained in this study. The suppression of NBS-LRRs, which was caused by the overexpression of miR482b (Fig. 4b), led to more disease symptoms in miR482b-overexpressing tomato lines (Fig. 3). In miR482b-silencing lines, the expression levels of NBS-LRRs was significantly upregulated (Fig. 4c) and the resistance of tomato plants to P. infestans enhanced (Fig. 3). Thus, these results provided additional evidence that these NBS-LRR genes were targeted by miR482b and they might play a role in the activation of tomato defense responses.
Based on these results, we propose a working model to explain how miR482b affects tomato resistance to P. infestans (Fig. 6). Following our results for degradome analysis, miR482b cleaves its target genes, the members of NBS-LRR family to suppress the abundance of NBS-LRRs (Fig. 4). When miR482b is overexpressed in tomato, the abundance of NBS-LRRs is decreased and more serious disease symptoms are shown after P. infestans infection. On the contrary, in miR482b-silencing tomato plants, more abundance of NBS-LRRs results in enhancement of tomato resistance to P. infestans. Our results provide insight into tomato miR482b-NBS-LRR module involved in the response to P. infestans infection and guidance for molecular breeding to improve P. infestans stress tolerance in the future. Fig. 6 Model of miR482b-NBS-LRR regulation network involved in the tomato-P. infestans interaction. Overexpression of miR482b in tomato negatively modulated P. infestans defense response by decreasing the abundance of NBS-LRRs, while the resistance was enhanced after miR482b silencing