Comprehensive Characteristics of MicroRNA Expression Profile Conferring to Rhizoctonia solani in Rice

CAO Wenlei , CAO Xinxin , ZHAO Jianhua ZHANG Zhaoyang FENG Zhiming ,OUYANG Shouqiang , ZUO Shimin

(1Jiangsu Key Laboratory of Crop Genetics and Physiology / Key Laboratory of Plant Functional Genomics of the Ministry of Education /Jiangsu Key Laboratory of Crop Genomics and Molecular Breeding, Agricultural College, Yangzhou University, Yangzhou 225009,China; 2College of Horticulture and Plant Protection, Yangzhou University, Yangzhou 225009, China; 3Joint International Research Laboratory of Agriculture and Agri-Product Safety, Yangzhou University, Yangzhou 225009, China; 4Jiangsu Co-Innovation Center for Modern Production Technology of Grain Crops, Yangzhou 225009, China; #These authors contributed equally to this work)

Abstract: MicroRNAs (miRNAs) are about 22 nucleotides regulatory non-coding RNAs that play versatile roles in reprogramming plant responses to biotic and abiotic stresses. However, it remains unknown whether miRNAs confer the resistance to necrotrophic fungus Rhizoctonia solani in rice. To investigate whether miRNAs regulate the resistance to R. solani, we constructed 12 small RNA libraries from susceptible and resistant rice cultivars treated with water/pathogen at 5 h post inoculation (hpi), 10 hpi and 20 hpi, respectively. By taking the advantage of next-generation sequencing, we totally collected 400-450 known miRNAs and 450-620 novel miRNAs from the libraries. Expression analysis of miRNAs demonstrated different patterns for known and novel miRNAs upon R. solani challenge. Thirty-four miRNA families were identified to be expressed specifically in rice, and most of them were involved in plant disease resistance. A particular Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis result revealed that a great majority of target genes of regulated miRNAs belonged to the pathway of plant-pathogen interaction. Moreover, miR444b.2, miR531a, mir1861i, novel_miR1956 and novel_miR135 conferred response to R. solani infection confirmed by Northern blot. Our global understanding of miRNA profiling revealed that the regulation of miRNAs may be implicated in the control of rice immunity to R. solani. Analysis of the expression of miRNAs will offer the community with a direction to generate appropriate strategies for controlling rice sheath blight disease.

Key words: rice; microRNA; sheath blight disease; resistance; transcriptomics

Small RNAs (sRNAs), derived from double-stranded RNA or hairpin-structured RNA, are short singlestranded non-coding RNAs of 19-30 nt in length that guide gene silencing in most eukaryotes (Baulcombe,2004; Vaucheret, 2006). Plants produce different types of endogenous sRNA, which are classified into microRNA (miRNA), small interfering RNA (siRNA),piwi-interacting RNA (piRNA), phased siRNAs(phasiRNAs) and heterochromatic siRNAs (Carthew and Sontheimer, 2009; Ghildiyal and Zamore, 2009;Arikit et al, 2013; Axtell, 2013). Among these, miRNAs present as molecules of 20-24 nucleotides in length endogenously transcribed from single-stranded noncoding RNA (Reinhart et al, 2000; Llave et al, 2002a).Plant miRNA was first identified from Arabidopsis thaliana (Park et al, 2002). To date, erupting studies have demonstrated that miRNAs play critical roles in various cellular processes in many organisms, including plant development (Aukerman and Sakai, 2003;Palatnik et al, 2003; Rubio-Somoza et al, 2009),hormone signaling (Bartel, 2004), innate immunity(Padmanabhan et al, 2009; Qiao et al, 2013; Weiberg et al, 2013) and stress response (Sunkar et al, 2006,2007). The expression of miRNAs is regulated by external stimuli, including abiotic (such as salinity,drought and temperature) and biotic (such as fungi,bacteria and viruses) stresses. It is well known that miRNAs regulate gene expression by guiding posttranscriptional gene silencing via sequence-specific cleavage or translational repression of recognized target mRNAs (Llave et al, 2002b; Brodersen et al, 2008).

Two mainly layered innate immune systems have been explored to against pathogens in plants so far.Upon to pathogen invasion, recognition of microbeassociated molecular patterns (MAMPs) by plant pattern-recognition receptors (PRR) leads to a general defense response referred to as PAMP-triggered immunity(PTI), which includes alternation of hormone and metabolite levels, production of reactive oxygen species(ROS), deposition of calluses, and accumulation of pathogenesis related proteins (PRs) (Jones and Dangl,2006; Baldrich et al, 2015). To counteract this innate defense of host, pathogens produce effectors to suppress PTI. In return, many plants have evolved another layer of immunity namely effector-triggered immunity (ETI), which is activated by plant resistance proteins after recognizing pathogen-secreted specific effectors (Jones and Dangl, 2006).

Rice (Oryza sativa) is the most important and worldwide consumed food for the human population.During past decades, a significant progress has been achieved in identification and functional analysis of miRNAs from rice (Liu et al, 2014). In rice, both ETI and PTI confer resistance to bacterial and fungal pathogens (Navarro et al, 2006). The involvement of miRNAs in PTI responses and pathogen resistance is first demonstrated in Arabidopsis, and miR393 is the first identified miRNA to mediate the repression of auxin signaling resulting in bacterial pathogen resistance (Ouyang et al, 2014). Moreover, miRNAmediated regulation of resistance gene (R gene)expression during ETI responses has been demonstrated in both the transcription and post-transcription levels(Campo et al, 2013; Boccara et al, 2014). Several substantial fractions of rice miRNA transcriptomes have been reported to be pathogen-responsive (Ito et al,2014; Baldrich et al, 2015; Yang et al, 2016; Wu et al,2017). However, how these pathogen-regulated miRNAs involve in rice immunity remains largely elusive. Only certain miRNAs, such as miR7695(Baldrich et al, 2015), miR160a and miR398b (Ito et al,2014), have been functionally characterized in the interaction of rice plants with rice blast fungus Magnaporthe oryzae or during infection with viral pathogens. Recently, miR528 was revealed to negatively regulate viral resistance by cleaving L-ascorbate oxidase(AO) messenger RNA leading to reduce AO-mediated accumulation of ROS in rice (Boller and Felix, 2009).

Rice sheath blight disease, caused by Rhizoctonia solani kühn, a basidiomycete necrotrophic fungal pathogen, is one of the most devastating fungal diseases of rice causing serious grain yield reduction and low grain quality. In the fields, the pathogen generally initially infects rice bottom sheath at the late tillering stage and then mainly develops upward along the sheath. Due to lack of stable sources with high resistance in rice germplasm, breeding for R. solani resistant rice has been progressed slowly. By taking the advantage of the rapid development of molecular marker technology, more than 50 R. solani resistance quantitative trait loci (QTLs) have been mapped to all the 12 rice chromosomes (Zuo et al, 2008, 2013;Taguchi-Shiobara et al, 2013; Yadav et al, 2015).Additionally, some members of WRKY (OsWRKY30,OsWRKY80, OsWRKY4, OsWRKY12 and GhWRKY27a)family acting as positive transcript factor confer R.solani resistance in rice (Peng et al, 2012, 2016; Pooja et al, 2015; Wang H H et al, 2015; Yan et al, 2015;Jiang et al, 2018). Over-expressions of OsPGIP1 and OsOSM1 immunize resistance to R. solani in Zhonghua 11 and Xudao 3, respectively (Wang R et al, 2015;Chen et al, 2016; Xue et al, 2016). Green tissuespecific simultaneous over-expressions of two pathogenrelated (PR) genes (OsCHI11 and OsOXO4) enhance resistance against R. solani without damaging important agronomical traits in rice (Karmakar et al, 2016).Recently, by using deep sequencing technology,miRNAs regulating the pathogenesis in R solani AG1 IA were reported (Lin et al, 2016). Based on hyphal small RNA libraries from six different infection periods of the rice leaf, totally 177 miRNA-like small RNAs (milRNAs) were identified including 15 candidate pathogenic novel milRNAs (Lin et al, 2016).However, it remains unknown that how the miRNAs from rice host confer the resistance to R. solani.

In this study, we recruited two rice cultivars,susceptible Xudao 3 and resistant YSBR1, to present different resistance to R. solani strain YN-7. The main objective of this study was to explore the potential miRNAs from sheath tissue elucidated by R. solani in different rice cultivars. Our results indicated that a complex and diverse miRNA population existed in the tested rice cultivars. We also identified and validated a few miRNAs regulated by the infection of R. solani,thereby providing a meaningful database to discover the possible molecular basis of the plant-pathogen interactions in rice.

MATERIALS AND METHODS

Plant materials and growth conditions

Rice cultivar YSBR1 displays high resistant level to rice sheath blight disease in both field and greenhouse conditions, which is derived from the hybrid progeny of japonica/indica (Zuo et al, 2009). Rice cultivar Xudao 3, a japonica rice cultivar released in Jiangsu Province, China, is susceptible to rice sheath blight disease, which has been used in our previous studies as susceptible control (Zuo et al, 2013, 2014b). Rice plants grew in the rice field of experimental farm located in Yangzhou University, China till tillering stage, and before this study, the plants were sprayed with fungicide and insecticide for keeping healthy.After tillering stage, the plants were transferred to greenhouse and inoculated with R. solani at 7 d later.

Pathogen inoculation and disease scoring

Rhizoctonia solani strain YN-7 (the original name was RH-9), with moderate pathogenicity, was used to inoculate rice plants. Autoclaved thin matchsticks with length of 0.8 to 1.0 cm were incubated with YN-7 strain on potato dextrose broth (PDB) medium for 3-4 d in darkness at 26 °C before applied as the inoculum. Rice plants at the late tillering stage were artificially inoculated with YN-7 using the method as described previously (Zuo et al, 2013, 2014a). Immediately after inoculation, all plants were placed in a plastic moist chamber with humidity of 92%-100% and temperature of (28 ± 2) °C. The sheaths were harvested at 5, 10 and 20 h post inoculation (hpi), respectively. In order to minimize experimental variations, all sheath samples were consisted of five seedlings for each treatment.All experiments were repeated two times.

The upper sections (2-3 cm) of sheaths infected with YN-7 were stained by lacto-phenol cotton blue followed by directly microscopic examination to observe pathogen growth. The bottom sections (2-3 cm) were frozen in liquid nitrogen and stored at -80 °C for total RNA extraction. Disease severity was rated by measuring lesion length at indicated time.

Total RNA extraction, libraries construction and RNA-seq

Total RNA was isolated from sheaths using TRIzol Reagent (Life Technologies, CA, USA) according to the manufacturer’s recommendations. For each sample,all sheaths from two biological repeats were pooled together. After the total RNA extraction and DNase I treatment, small RNA libraries for sequencing were constructed using 1 μg of total RNA, according to the Illumina Hiseq4000 Small RNA library preparation protocol (Illumina, USA). In brief, total RNA samples were separated using the polyacrylamide gel electrophoresis gel, and cut out between 18 and 30 nt stripe to recover small RNA. After 3- and 5-adapter ligation at both ends, the ligated products were reverse transcribed with Superscript II Reverse transcriptase using adapter-specific RT-primers. PCR products were then gel purified to enrich special fragments. During the quality control (QC) steps, Agilent 2100 Bioanaylzer and ABI StepOnePlus Real-Time PCR System were used in quantification and qualification of the libraries.The purified high-quality cDNA library was sequenced using Illumina Genome Hiseq4000.

Normalization of sequence reads and identification of differentially expressed miRNA

Primary sequencing data that produced by Illumina Genome Hiseq4000, called as raw reads, were subjected to QC. After QC, raw reads were filtered into clean reads (18-30 nt small RNAs), which were aligned to the reference sequences as described by previous report (Trapnell et al, 2013). All sequence reads were trimmed to remove the low-quality sequences.The sequence data were subsequently processed using in-house software tool SeqQC V2.2. House-keeping small RNA including rRNAs, tRNAs, snRNAs and snoRNAs were removed by blasting in GenBank(http://www.ncbi.nih.gov/Genbank) servers. The trimmed reads were then aligned to the rice reference genome downloaded from http://www.mirbase.org (miRBase 21.0) and http://www.ncbi.nlm.nih.gov using Bowtie v0.12.5 and TopHat v2.0.0 (Ruepp et al, 2004; Trapnell and Salzberg, 2009) with default settings. To identify known miRNAs, the remaining unique small RNA sequences were then aligned against the miRBase 21.0 allowing maximum one mismatch. After assigning the known miRNA sequences into their respective groups or families, the rest of the sequences were checked for novel miRNAs.

Functional categorization of differentially expressed miRNA

Differentially expressed miRNAs were functionally categorized online for all pairwise comparisons according to the Munich Information Center for Protein Sequences(MIPS) functional catalogue (Ruepp et al, 2004). The functional categories and subcategories were regarded as enriched in the genome if an enrichment P- and FDR-value was below 0.05. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed using interface on blast2GO v2.6.0(http://www.blast2go.com/b2ghome) for all differentially expressed miRNA to identify gene enrichment on a specific pathway.

Gene ontology (GO) and pathway enrichment analysis

GO annotation and pathway enrichment were performed for the target genes of the differentially expressed miRNAs to determine and compare gene functions using the Balst2GO and DAVID softwares. The sequences of target genes were mapped to get the GO terms and the results were categorized on biological process, molecular function and cellular component at level 2. Graphs of the top 20 enriched GO terms for each library were generated using the Cytoscape Enrichment Map plugin (Merico et al, 2010; Smoot et al,2011).

Northern blot analysis

For Northern blot analysis, 10 μg total RNA was resolved on urea denaturing polyacrylamide gels(Urea-PAGE). miRNA-specific oligonucleotide probes were end-labeled using γ-32P-ATP (New England Biolabs, Ipswich, UK). The upper section of the blot was used for the loading control using a U6 oligonucleotide probe. All blots were imaged by a PhosphorImager (GE Life Sciences, Pittsburgh, PA,USA). The images were cropped and adjusted with brightness and contrasted in Photoshop CS6 from original digital images.

RESULTS

Post inoculation phenotype in Xudao 3 and YSBR1

Resistant cultivar YSBR1 and susceptible cultivar Xudao 3 were inoculated with the matchsticks colonized by the mycelium of R. solani isolate YN-7 (Fig. 1-A and-B). At 10 hpi, runner hyphae of YN-7 grew along the junctions of epidermal cells in two cultivars (Fig. 1-C and -F). The sheaths of both cultivars were found yellowing and presented water soaked lesions (Fig.1-D and -G) at 20 hpi. These results were observed on at least 20 individual rice seedlings, which indicated that the phenotypes were similar between the two cultivars before 1 d post inoculation (dpi) (Fig. 1-C,-D, -F and -G). However, the lesion length of YSBR1 was shorter than that of Xudao 3 at 2 dpi, and the difference became increasingly obvious after 2 dpi(Fig. 1-I). Our results showed that YSBR1 possessed significantly stronger resistance to R. solani than Xudao 3.

Small RNA library construction and analysis

To identify miRNAs responding to R. solani infection in rice, we generated totally 12 small RNA libraries including water treatment (mock) and R. solani treatment (infection) to susceptible cultivar Xudao 3 and resistant cultivar YSBR1, respectively, at 5, 10 and 20 hpi, named as XC5, XC10 and XC20 for Xudao 3 by water treatment, XT5, XT10 and XT20 for Xudao 3 by R. solani treatment, YC5, YC10 and YC20 for YSBR1 by water treatment, YT5, YT10 and YT20 for YSBR1 by R. solani treatment, respectively(Supplemental Fig. 1).

Using Illumina sequencing, a total of 312 120 151 clean reads were collected. For each library, about 23-29 million clean reads were obtained, and at least 87% reads were singletons (Supplemental Table 1).The composition of sRNA pool was comprehensive and contained a huge portion of other non-coding RNA species including rRNA, repeat, snRNA, snoRNA,tRNA, and unannotation sRNA (Supplemental Table 2). Summarily, the mapping results showed that our sRNA libraries were highly enriched in miRNAs. For read distribution, the majority reads were from chromosomes 2 and 9, and of these reads, over 90%reads were transcripted from sense strain (Fig. 2-A).The lengths of all the sRNAs ranged from 18 to 30 nt,among which 21- and 24-nt sRNA species were the two most abundant classes (Fig. 2-B). Within the miRNA population of sequences, more than 95% of the reads began with a uracil (U) (Fig. 2-C).

Identification of known and novel miRNAs

In our analysis, all the non-coding RNA species were removed, and the genome sequence of Oryza sativa L.japonica. cv. Nipponbare was used as the reference for miRNA identification. For each library, we collected almost the same number (400-450) of known miRNA.However, the number of novel miRNAs was 450 in the library XC5 and 617 in the library XT20, respectively.For known miRNA, miRNA-5P and miRNA-3P shared similar abundance (Supplemental Table 3). When analyzing miRNA regulation, up-regulated known miRNAs were much abundant than down-regulated ones in both Xudao 3 and YSBR1 at 20 hpi (Supplemental Fig. 2). There is no significant difference between up-regulated and down-regulated novel miRNAs in all the libraries except for XC20 vs XT20. However,novel miRNAs were dramatically down-regulated in Xudao 3 infected with R. solani (Supplemental Fig.2-B). Based on the normalized expression of known/novel miRNAs under R. solani infection, a significant change in expressed miRNA was observed between mock and R. solani-infected rice. Venn diagrams were generated for the miRNAs in 5, 10 and 20 hpi respectively for miRNAs from the two rice cultivars (By Venny 2.1).For known miRNAs, three and five miRNAs were regulated in all the three time points in the two cultivars, respectively (Supplemental Fig. 2-C). For novel miRNAs, 5 and 13 miRNAs were differently expressed in all the three time points in YSBR1 and Xudao 3, respectively (Supplemental Fig. 2-D). Coexpressed known and novel miRNAs from different libraries in the three time points were detailed in Supplemental Table 4.

miRNA expression patterns upon to R. solani infection

Generally, 74 conserved miRNA families were detected in all the libraries. Of these miRNA families, 34 miRNA families were expressed specifically in rice.Several miRNAs were reported both in rice and other species including miR169_4 (in Zea mays), miR821 and miR1435 (in Sorghum_bicolor), and miR1878 and miR5179 (in Brachypodium distachyon) (Supplemental Table 5). Among these specific expressed miRNA families, 12 of them have been reported known function in rice, which are listed in Supplemental Table 6.

To further reveal the miRNA expression patterns,hierarchical clustering analysis was preformed to show the differential regulation between mock and pathogen infection. For known miRNAs, osa-miR398a,osa-miR1881, osa-miR530, osa-miR444, osa-miR812,osa-miR1861, osa-miR3980 and osa-miR531f families were significantly induced in R. solani treated rice but undetectable in mock rice. However, osa-miR171f-5p,osa-miR2863a, osa-miR3979-3p, osa-miR1428e-3p and osa-miR156 families were repressed greatly in R. solani treated rice (Supplemental Fig. 3-A). Among above miRNAs, only osa-miR398a belongs to miR398 family,which was predicted to target mRNAs coding for copper superoxide dismutases and cytochrome C oxidase (Jones-Rhoades and Bartel, 2004). For novel miRNAs, interestingly, a large amount of novel miRNAs were discovered in the R. solani treated rice compared with the control (Supplemental Fig. 3-B).

Target gene prediction and functional analysis

To better understand the function of the target genes of miRNAs, GO terms and KEGG analysis were performed. Using the Cytoscape Enrichment Map plugin software, the annotated coding regions were distributed into about 30 different level-2 GO categories referenced to three main classes: biological processes,molecular function and cellular component. Among the target genes of both known and novel miRNAs,metabolic process was the most represented category in biological process, associated with about 100 and 200 genes in Xudao 3 and YSBR1, respectively. Most of the annotated target genes were found to be associated with the cell and cell part category in cellular component,assigned with nearly 300 and 400 functional genes in Xudao 3 and YSBR1, respectively. For molecular function class, the binding and catalytic activities were associated with over 100 and 200 transcripts in Xudao 3 and YSBR1, respectively (Fig. 3).

To elucidate that the biological processes of target genes may participate in responding to R. solani infection, we analyzed the allotment of these target genes to a particular KEGG pathway. An enrichment analysis based on GO terms was performed to reveal top 20 terms that were significantly regulated among the target genes of known/novel miRNAs in each cultivar (Supplemental Fig. 4). A close inspection of genes belonging to a particular pathway was differentially regulated during the infection. Target genes involved in plant hormone signal transduction were present at 5 and 20 hpi in Xudao 3, at 10 and 20 hpi in YSBRI for known miRNAs. For novel miRNAs, genes involved in plant hormone signal transduction were predicted at 20 hpi in YSBR1 only. We further found that the genes associated with plant-pathogen interaction were recognized at 10 hpi in YSBR1 for novel miRNAs. A total of 145 (11.42%) genes were involved in plantpathogen interaction, which encode WRKY family proteins (6 genes), nucleotide-binding domain containing proteins (42 genes), retrotransposon proteins (32 genes),stripe rust resistance proteins (7 genes), disease resistance proteins (12 genes), resistance genes (10 genes), Mla family proteins (9 genes) and receptor kinases (3 genes) (Supplemental Table 7).

miRNA expression patterns responding to R. solani infection by Northern blot

To further confirm the reliability of our sRNA-seq data and characterize the miRNAs dynamic expression patterns upon R. solani infection, three known miRNAs(miR444b.2, miR531a and miR1861i), and two novel miRNAs (novel_miR1956 and novel_ miR135) were evaluated by Northern blot analysis (detail sequences of miRNAs and probes were presented in Supplemental Table 8). Our data indicated that miR444b.2 was induced in both cultivars by R. solani infection (Fig.4). However, five miRNAs were repressed congruously in both cultivars when treated with R. solani (Fig. 4). More interesting, miR531a was presented only in resistant cultivar YSBR1, and down-regulated significantly at 20 hpi (Fig. 4). These data displayed a credible consistence with the results of sRNA-seq. Meanwhile,these miRNAs may be involved in the response to R.solani infection in rice.

DISCUSSION

Next-generation sequencing is a powerful bioinformatic technology tool to generate a digitalized expression profile of miRNAs due to its high specificity and sensitivity, and has significantly contributed to the discovery of miRNAs in plants. The current miRBase(release 21) holds 604 precursors and 738 mature rice miRNAs (MSU7) (Kozomara and Griffiths-Jones, 2014).During the last years, several studies have reported that miRNAs were regulated in rice responding to pathogen/virus invasion including M. oryzae (causing rice blast disease) (Campo et al, 2013; Ito et al, 2014;Baldrich et al, 2015; Li et al, 2016), Xanthomonas oryzae pv. oryzae (causing rice bacterial blight disease)(Zhao et al, 2015), rice stripe virus (Yang et al, 2016),and rice black streaked dwarf virus (Sun et al, 2015).In this study, we exploited the availability of susceptible and resistant rice cultivars towards R. solani to identify miRNAs important for rice defense. This study attempted to understand the crosstalk between host miRNAs and fungal pathogen underlying the resistance mechanism and to capture miRNA candidates responding to R. solani in rice.

Our results profiled the conserved and nonconserved miRNA families only identified in rice genome, and most of them were related to abiotic or biotic stress response (Nischal et al, 2012; Zhang et al,2014; Cheah et al, 2015; Mutum et al, 2016). miR810(Cheah et al, 2015), miR812 (Li et al, 2011), miR814(Zhang et al, 2014), miR815 (Mutum et al, 2016),miR820 (Sharma et al, 2015), miR1846 (Nischal et al,2012), miR1862 (Peng et al, 2014) and miR2871(Barrera-Figueroa et al, 2012) are conferred to respond to abiotic stress, and miR2863 (Baldrich et al, 2015) is demonstrated to respond to biotic stress in rice.miR531 (Raghuram et al, 2014) and miR1861 (Peng et al, 2014) are involved in signaling pathway in rice.miR1863 (Cao et al, 2013) is required for silencing heterochromatin by methylation in rice (Supplemental Table 6). Among these miRNA families, only miR2863b,a member of miR2863 family, is induced in rice root but accumulated at a relatively low level in leaves when treated with elicitors obtained from the rice blast fungus M. oryzae (Baldrich et al, 2015). Moreover,further analyzing the function of target genes in plantpathogen interaction pathway revealed that a great quantity of disease resistance genes were regulated in responding to R. solani invasion. Intriguingly, retrotransposon genes were also involved in rice-R. solani interaction. Various retrotransposons have been shown to be induced by plant pathogens or elicitors in rice by inserting in disease resistance gene cluster during pathogen infection (Chen et al, 2007; Stetson et al,2008). When compared to PTI, ETI is a stronger defense layer in plant, which is usually mediated by nucleotide-binding/leucine-rich repeat (NB-LRR or NLR) proteins encoded by NLR gene family. Apart from miRNAs mediating in PTI via hormonal signal transduction, numerous miRNAs target transcripts from NLR genes. Moreover, most miRNAs also trigger the generation of phasiRNAs from NLR targets(Fei et al, 2013; Zhai et al, 2015). In Solanaceous species, such as tomato, slymiR482f and slymiR5300 target NLR genes to confer tomato wilt disease caused by Fusarium oxysporum (Ouyang et al, 2014).

To date, the accurate role of most reported pathogenrelated miRNAs still remains elusive in rice innate immunity. Only several certainly miRNAs have been functionally characterized in the interaction of rice and fungal pathogen M. oryzae or during infection with viral pathogens in rice. The molecular mechanisms how rice responses to R. solani invasion remain poor understood either. This study was designed to capture miRNA profiling associated with R. solani by taking advantage of different rice varieties. Here, five miRNAs have been identified to confer to R. solani invasion in rice. Among these miRNAs, miR444 (miR444b.2 in this study), a monocot-specific microRNA, was reported as a positive factor in relaying the antiviral signaling from virus infection. miR444 targets the MIKC(C)-type MADS complexes to repress the expression of OsRDR1, which activates the OsRDR1-dependent antiviral RNA-silencing pathway in rice (Wang et al,2016). Our results showed that miR444b.2 had a similar expression pattern upon R. solani infection at early stage, which implicated that miR444b.2 might be involved in the resistance to R. solani in rice.Furthermore, it was an intriguing discovery that miR531a was found only in the resistant cultivar YSBR1, and was repressed significantly upon R. solani infection,particularly at 20 hpi. This observation made us speculate that miR531a played roles not only in speciesdependent development but also the association with the resistance to R. solani. miR531a also distinguished from other miRNAs by containing 24 nucleotides. As we know, 24-nt sRNAs are the most abundant class of plant sRNAs, and often associate with DNA methylation of matching sequences, particularly to transposable elements and other repetitive genomic regions (Matzke and Mosher, 2014). We further revealed that target genes were regulated during antagonistic interactions with R. solani, including plant hormone signal transduction and plant-pathogen interaction pathway,with a role in disease resistance being prominent.

This study explored unique miRNAs involved in resistance against R. solani by using sRNA sequencing in rice. Totally, 34 miRNA families were presented to be expressed specifically and involved in R. solani resistance in rice. We further confirmed that miR444b.2,miR531a, mir1861i, novel_miR1956 and novel_miR135 responded to R. solani infection. It remains to be determined whether silence/over-expression mature and/or precursor forms of candidate miRNAs can up-regulate/down-regulate target gene expression in accordingly plants, respectively. In this scenario, it can be prospected that the discovery of rice miRNAs involved in host defense responses to R. solani will facilitate research on host resistance.

ACKNOWLEDGEMENTS

We are indebted to Dr. CHEN Xuewei for advice on small RNA library construction and many helpful suggestions over the course of this project. This work was partially supported by the National Key Research and Development Program of China (Grant No.2016YFD0100601), National Natural Science Foundation of China (Grant No. 31672013), FOK YING TUNG Education Foundation (Grant No. 151026), High Talent Project of Yangzhou University and Priority Academic Program Development of Jiangsu Higher Education Institutions.

SUPPLEMENTAL DATA

The following materials are available in the online version of this article at http://www.sciencedirect.com/science/journal/16726308; http://www.ricescience.org.

Supplemental Table 1. Summary of reads collected by RNA-seq from twelve libraries.

Supplemental Table 2. Exploring miRNA regulation in rice responding to R. solani invasion.

Supplemental Table 3. Number of known/novel miRNAs and target genes from each library.

Supplemental Table 4. Details of co-expressed known and novel miRNAs from different libraries in three time points in Venn.

Supplemental Table 5. GO miRNA family analysis.

Supplemental Table 6. Functions of miRNA families specially present in rice.

Supplemental Table 7. Details of target genes involved in the plant-pathogen interaction pathway.

Supplemental Table 8. Primers used for Northern blot analysis in this study.

Supplemental Fig. 1. Library construction and workflow for the prediction of candidate miRNAs.

Supplemental Fig. 2. Properties of differential expressed miRNAs in resistant and susceptible rice cultivars treated with water or the fungal pathogen R. solani.

Supplemental Fig. 3. Expression patterns of miRNAs from all libraries.

Supplemental Fig. 4. KEGG pathway based on GO terms classification of enrichment analysis unigenes differentially expressed between the water and pathogen treatment at different time point.