Identification of a novel microRNA (miRNA) from rice that targets an alternatively spliced transcript of the Nramp6 (Natural resistance-associated macrophage protein 6) gene involved in pathogen resistance


Author for correspondence:

Blanca San Segundo

Tel: +34 935636600 ext. 3131



  • Plants have evolved efficient defence mechanisms to defend themselves from pathogen attack. Although many studies have focused on the transcriptional regulation of defence responses, less is known about the involvement of microRNAs (miRNAs) as post-transcriptional regulators of gene expression in plant immunity. This work investigates miRNAs that are regulated by elicitors from the blast fungus Magnaporthe oryzae in rice (Oryza sativa).
  • Small RNA libraries were constructed from rice tissues and subjected to high-throughput sequencing for the identification of elicitor-responsive miRNAs. Target gene expression was examined by microarray analysis. Transgenic lines were used for the analysis of miRNA functioning in disease resistance.
  • Elicitor treatment is accompanied by dynamic alterations in the expression of a significant number of miRNAs, including new members of annotated miRNAs. Novel miRNAs from rice are proposed. We report a new rice miRNA, osa-miR7695, which negatively regulates an alternatively spliced transcript of OsNramp6 (Natural resistance-associated macrophage protein 6). This novel miRNA experienced natural and domestication selection events during evolution, and its overexpression in rice confers pathogen resistance.
  • This study highlights an miRNA-mediated regulation of OsNramp6 in disease resistance, whilst illustrating the existence of a novel regulatory network that integrates miRNA function and mRNA processing in plant immunity.


Plants possess an innate immune system that efficiently detects potential microbial pathogens. The initiation of plant defence responses depends on the recognition of pathogen epitopes, known as pathogen-associated molecular patterns (PAMPs), or elicitors, by host-encoded surface receptors (Chisholm et al., 2006; Jones & Dangl, 2006; Boller & He, 2009). This recognition activates a complex process in which different signalling cascades operate, leading to a massive reprogramming of the transcriptome. The regulation of immune response genes has been studied mostly at the transcriptional level. However, several recent studies have indicated that plants additionally use post-transcriptional regulation of defence responses against pathogens and that host endogenous small RNAs appear to be essential in this gene expression reprogramming process (Ruiz-Ferrer & Voinnet, 2009; Katiyar-Agarwal & Jin, 2010; Staiger et al., 2012).

Small RNAs, including microRNAs (miRNAs) and small interfering RNAs (siRNAs), regulate gene expression in a sequence-specific manner (Baulcombe, 2004; Xie et al., 2005; Jones-Rhoades et al., 2006; Voinnet, 2009). Plant miRNAs are produced from precursors with unique stem–loop structures which are sequentially processed by the RNase III DICER-like 1 (DCL1) to give rise to an miRNA–miRNA* duplex. Alternative pathways for miRNA biogenesis involving DCL4 or DCL3 have also been described (Rajagopalan et al., 2006; Vazquez et al., 2008). The miRNA–miRNA* duplex intermediates are translocated to the cytoplasm, where the miRNA guide strand is selectively incorporated into an ARGONAUTE1 (AGO1)-containing RNA-induced silencing complex (RISC) (Baumberger & Baulcombe, 2005; Xie et al., 2005; Jones-Rhoades et al., 2006; Vaucheret, 2008). miRNAs direct post-transcriptional gene silencing by triggering the cleavage or translational repression of the target transcripts (Llave et al., 2002; Brodersen et al., 2008). They have been proven to influence temporal changes in target gene expression in developmental processes, as well as in abiotic stress tolerance and nutrient starvation (Palatnik et al., 2003; Mallory et al., 2004; Sunkar et al., 2008; Jagadeeswaran et al., 2009; Ruiz-Ferrer & Voinnet, 2009; Jeong et al., 2011). Evidence for miRNAs controlling pathogen resistance came first in Arabidopsis plants, where perception of the bacterial flagellin flg22 causes an increase in miR393 accumulation which negatively regulates transcripts for F-box auxin receptors. Repression of auxin signalling results in bacterial resistance (Navarro et al., 2006). More recently, miRNAs that guide the cleavage of disease resistance (R) genes in Solanaceae and Leguminosae species have been described (Zhai et al., 2011; Li et al., 2012; Shivaprasad et al., 2012).

Most of the miRNAs that were discovered in early reports are highly conserved throughout the plant kingdom and have conserved functions in the regulation of developmental processes (Jones-Rhoades et al., 2006). It is also generally assumed that plants express species-specific miRNAs that might play a regulatory role in a time- and/or spatial-restricted manner, or in specific biological processes (Rajagopalan et al., 2006; Fahlgren et al., 2007; Cuperus et al., 2011). One important challenge now is to identify novel species-specific miRNAs, and to elucidate their biological function, in crop species that undergo major environmental stresses.

Rice (Oryza sativa) is a species of evident interest for miRNA analysis, not only because of its worldwide agricultural importance, but also because it represents the model plant for research in monocotyledonous species with a sequenced genome. Moreover, rice has a long history of natural selection and selective breeding, thus providing an excellent system for studies on the molecular evolution and selection of plant miRNAs. In the present study, we sequenced small RNA libraries from rice tissues (leaves, roots) that had been treated with elicitors obtained from the rice blast fungus Magnaporthe oryzae, the causal agent of the rice blast disease (Talbot, 2003). A diverse set of known miRNAs, both conserved and nonconserved, was found to be responsive to elicitor treatment. We also describe novel miRNA candidates. In particular, we report the functional characterization of a novel miRNA from rice, osa-miR7695, which negatively regulates an alternatively spliced transcript of the OsNramp6 (Natural resistance-associated macrophage protein 6) metal transporter gene. Overexpression of the newly identified miRNA results in enhanced resistance to pathogen infection in rice plants. Finally, we show that this novel miRNA experienced natural and domestication selection events during rice evolution.

Materials and Methods

Plant material and elicitor treatment

Rice plants (Oryza sativa L. cv Nipponbare) were grown at 28 ± 2°C with a 16 h : 8 h light : dark cycle. Elicitors from the M. oryzae strain 18.1 were prepared as described by Casacuberta et al. (1992) and used at a final concentration of 300 μg ml−1. In all experiments, mock inoculations were performed. Cultivated rice varieties from different geographical locations, as well as wild rice accessions, were assayed. They were: Oryza sativa (21 accessions, including eight temperate japonica, six tropical japonica and seven indica accessions), Oryza glaberrima (two accessions) and wild rice species (24 accessions) (Supporting Information Table S1). Rice seeds were obtained from the International Rice Research Institute (IRRI,

Small RNA library construction and sequencing

Total RNA was extracted from tissues (leaves, roots; 30 min and 2 h each tissue) of 15-d-old seedlings using TRIzol reagent (Invitrogen). Three biological replicates were analysed. Each library represented a pool of c. 50 rice plants. The construction of small RNA libraries has been reported previously (Donaire et al., 2009). Amplicons were prepared by adaptor ligation in which the 5′ adaptor contained a ‘barcode’ for each sample. The same quantity of DNA amplicon from each library was pooled and sequenced using 454 Life Sciences Technology. A total of 383 397 row reads was produced. Computational analysis of reads containing recognizable adaptor sequences yielded 271 487 reads. The leaf libraries included 155 465 sequences: controls, 40 480 and 45 812 sequences (30 min and 2 h, respectively); elicitor-treated, 31 357 and 37 816 sequences (30 min and 2 h, respectively). The root libraries included 116 022 sequences: controls, 37 477 and 16 138 (30 min and 2 h, respectively); elicitor-treated, 28 677 and 33 730 (30 min and 2 h, respectively). A total of 96 069 unique small RNA sequences that perfectly matched the rice genome was identified in our libraries. The small RNA sequence data have been deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) (

Data mining of the small RNA pool and prediction of new miRNAs

Small RNA sequences were parsed from FASTA-formatted files from 454 sequencing and assigned to each one of the eight specific libraries through identification of the small RNA/adaptor boundaries and barcode analysis. Small RNA sequences were mapped to the rice genome (Oryza sativa, version 5.0; using BLASTn. The unique RNA sequences that perfectly matched the genome were subjected to subsequent analysis. Reads showing identical sequences to known miRNAs from the miRBase database (, release 18.0, November 2011) were collected. The remaining sequences were considered for computational prediction and the identification of novel miRNAs. For this, genomic sequences spanning the putative miRNA, 1500 nucleotides upstream and downstream, were extracted and used for fold-back secondary structure prediction employing the RNAfold from the Vienna package version 2.0.0 with default parameters ( Criteria for the recognition of candidate miRNA precursors have been outlined elsewhere (Meyers et al., 2008). The same procedure was used to search for orthologous sequences in the genome of different plant species. Sequences with up to three mismatches were retrieved and subjected to fold-back secondary structure prediction as described above.

RNA analyses

For northern blot analysis of rice miRNAs, the low-molecular-weight fraction was obtained from total RNA by PEG8000/NaCl precipitation. RNAs were fractionated in a 17.5% denaturing polyacrylamide gel containing 8 M urea, transferred to nylon membranes and probed with [γ32P]ATP-labelled oligonucleotides (Table S2). Hybridization signals were detected using a Phosphorimager (Bio-Rad). Synthetic RNA oligonucleotides were loaded as size markers.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed in optical 96-well plates (Roche Light Cycler® 480; Roche Diagnostics, Mannheim, Germany) using SYBR Green I dye and the primers listed in Table S2. Data were normalized with OsUbi1 (AK121590) as internal control. Three independent biological samples and three technical replicates per sample were analysed. When appropriate, Student's t-test/ANOVA was used to determine the statistical significance of the differential expression patterns (P ≤ 0.05). Semi-quantitative RT-PCR of miR156 precursors and amplification of the osa-miR7695 precursor transcript are detailed in Methods S1.

Transient expression assays in Nicotiana benthamiana leaves and rice transformation

The DNA fragment containing the osa-MIR7695 precursor was amplified by PCR from rice (O. sativa cv Nipponbare) genomic DNA, and used in transient expression assays in N. benthamiana leaves. To avoid the transgene-derived production of siRNAs, the rdr6IR-N. benthamiana line was used in these studies (Schwach et al., 2005). Details on construct preparation, agroinfiltration and expression analysis of osa-miR7695-related small RNAs are indicated in Methods S1. Transgenic rice plants were produced by Agrobacterium-mediated transformation of embryogenic calli derived from mature embryos (Methods S1).

Target prediction

Target prediction for rice miRNAs was performed using the psRNATarget program ( with default parameters. Target sequences were searched for matching in the O. sativa cDNA set provided by The Institute for Genomic Research (TIGR, Rice Annotation Release 5.0). All predicted target genes were evaluated by scoring, and the criteria used were as follows: each G : U wobble pairing was assigned 0.5 points, each indel was assigned 2.0 points and all other noncanonical Watson–Crick pairings were assigned 1.0 point each. A penalty score of ≤ 3.0 points was considered in our analysis.

Microarray analysis

The GeneChip® rice genome array (Affymetrix, Santa Clara, CA, USA) was used for transcript profiling. Expression studies were performed in leaf tissues, treated or not with elicitors of M. oryzae (30 min and 2 h of treatment). In each case, total RNA was isolated from three different biological replicates, and each replicate was independently hybridized to the rice microarray. The statistical analysis of microarray data is presented in Methods S1. Genes with P ≤ 0.05 and fold changes ≥ 1.2 in all replicates were considered. Microarray data have been deposited in the NCBI GEO database (GSE30583).

Blast resistance assays

Resistance to infection by the rice blast fungus M. oryzae strain Guy11 (courtesy of Ane Sesma; Sesma & Osbourn, 2004) was determined using the detached leaf assay, as described previously (Coca et al., 2004). Further experimental details can be found in Methods S1.


Deep sequencing of small RNA populations from rice leaves and roots

To obtain a genome-wide comprehensive survey of miRNAs in rice showing responsiveness to fungal elicitors, we constructed small RNA libraries from leaves and root tissues that had been treated, or not, with elicitors prepared from the rice blast fungus M. oryzae (30 min or 2 h of elicitor treatment). Libraries containing unique barcodes were combined and subjected to pyrosequencing (454 Life Sciences™; Roche). After trimming the adaptor sequences, distinct sequences perfectly matching the rice genome were identified. The size distribution of small RNAs was determined in each tissue on the basis of both total abundances and unique sequences. In terms of total sequence abundance, the 24-nucleotide class of small RNAs was the most abundant class in leaves, whereas the 21-nucleotide class was predominant in roots (Fig. 1a, left panel). When the size distribution of unique small RNA sequences was determined, we found that the 24-nucleotide sequences prevailed in both tissues (Fig. 1a, right panel).

Figure 1.

Abundance of small RNA sequences, expression profiling of known microRNAs (miRNAs) from rice (Oryza sativa) and novel members of known miRNA families identified in this work. (a) Total abundance of small RNA sequences for each size class (left panel) and size distribution of unique small RNAs (right panel). Asterisks denote the predominant small RNA class. nt, nucleotide. (b) Expression profiling of miRNAs. Reads retrieved from the 454 sequencing dataset for each miRNA were normalized against the total count of reads obtained in the corresponding library. Only miRNAs showing 20 or more reads are presented. In (a, b), grey bars represent leaf tissue and black bars denote root tissue. (c) Precursor fold-back structures of novel members of known miRNA families. Grey bars denote the annotated sequences in the miRBase (miRBase version 18.0; November 2011). Black bars indicate the newly identified miRNAs in either leaf (thick bars) or root (thin bars) tissues. The nucleotide sequence of these precursor structures is shown in Supporting Information Fig. S2.

A sequence similarity search of the sequencing data against the central miRBase registry allowed us to identify up to 114 known miRNAs or miRNAs* representing 63 miRNA families in our rice libraries, which included both conserved and nonconserved miRNAs (Table 1). By sorting the reads according to the barcode added in 454 primers, we determined the accumulation of each miRNA in each tissue (roots, leaves) (Fig. 1b and Table S3). miR168a and miR167defghij were the most abundant miRNAs in rice leaves (Fig. 1b). A good correlation occurred between the frequency observed for miRNAs in the 454 dataset and their expression level, as determined by northern blot analysis of selected miRNAs in rice leaves (Fig. S1). However, caution should be taken when interpreting the expression levels for those miRNAs with a low number of reads in the 454 sequencing dataset.

Table 1. Known microRNAs (miRNAs) present in small RNA libraries from rice (Oryza sativa) leaves and roots
Zm Sb Ta Hv Mt Pt At Gm Vv
  1. Annotated miRNAs or miRNAs* whose sequences were represented at least three times in one or more of the rice libraries are included. miRNA sequences found in root, but not in leaf, tissues are shown in bold. Conservation among different plant species is shown (+ and −, identical and nonconserved sequences, respectively). Zm, Zea mays; Sb, Sorghum bicolor; Ta, Triticum aestivum; Hv, Hordeum vulgare; Mt, Medicago truncatula; Pt, Populus trichocarpa; At, Arabidopsis thaliana; Gm, Glycine max; Vv, Vitis vinifera.

  2. a

    New member of the indicated family.

  3. b

    This sequence was annotated as miR529* in miRBase, but was proposed to be the functional miRNA (Zhu et al., 2008).

156miR156abcdefghij, miR156dmiR156i*, miR156hj*(+)(+)(+)(+)(+)(+)(+)(+)(+)
159miR159ab, miR159fmiR159a.1*(+)(+)(+)(+)(+)(+)(+)(+)(+)
160miR160abcd,miR160e, miR160f (+)(+)(+)(−)(+)(+)(+)(+)(+)
162miR162a, miR162b (+)(+)(−)(−)(+)(+)(+)(+)(+)
164miR164abf, miR164c, miR164d, miR164e (+)(+)(+)(−)(+)(+)(+)(+)(+)
166miR166abcdfn, miR166gh, miR166mmiR166c*, miR166n* miR166g*(+)(+)(−)(+)(+)(+)(+)(+)(+)
167miR167abc, miR167defghij, miR167e, miR167f (+)(+)(+)(−)(+)(+)(+)(+)(+)
168miR168a (+)(+)(−)(+)(+)(+)(+)(+)(+)
169miR169a, miR169bc, miR169 fg, miR169hijklm, miR169no (+)(+)(−)(+)(+)(+)(+)(+)(+)
miR169i.2 (−)(−)(−)(−)(−)(−)(−)(−)(−)
171miR171bcdef, miR171f, miR171h, miR171imiR171f*(+)(+)(+)(+)(+)(+)(+)(+)(+)
172miR172ad (+)(+)(−)(−)(+)(+)(+)(+)(+)
319 miR319a.2 a  (−)(−)(−)(−)(−)(−)(−)(−)(−)
393miR393, miR393bmiR393b*(+)(+)(−)(−)(+)(+)(+)(+)(+)
394miR394 (+)(+)(−)(−)(−)(+)(+)(+)(+)
396miR396ab, miR396c, miR396de (+)(+)(−)(−)(+)(+)(+)(+)(+)
397miR397a, miR397bmiR397b*(+)(+)(−)(+)(−)(+)(+)(−)(+)
399miR399d (+)(+)(+)(−)(+)(+)(+)(−)(+)
444miR444a2d2e, miR444b1c1, miR444b2c2, miR444d3 (−)(−)(+)(+)(−)(−)(−)(−)(−)
528miR528 (+)(+)(−)(−)(−)(−)(−)(−)(−)
529miR529b (+)(+)(−)(−)(−)(−)(−)(−)(−)
535miR535 (−)(−)(−)(−)(−)(+)(−)(−)(+)
810miR810b1, miR810b2, miR810bamiR810b1*(−)(−)(−)(−)(−)(−)(−)(−)(−)
812miR812ghij (−)(−)(−)(−)(−)(−)(−)(−)(−)
827miR827ab (+)(−)(−)(−)(−)(+)(+)(−)(−)
1317miR1317-5p, miR1317-3p (−)(−)(−)(−)(−)(−)(−)(−)(−)
1318miR1318 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1423miR1423, miR1423b (−)(−)(−)(−)(−)(−)(−)(−)(−)
1425miR1425 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1427miR1427 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1429miR1429-3p (−)(−)(−)(−)(−)(−)(−)(−)(−)
1430miR1430 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1437 miR1437*(−)(−)(−)(−)(−)(−)(−)(−)(−)
1849 miR1849*(−)(−)(−)(−)(−)(−)(−)(−)(−)
1850miR1850.1 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1851miR1851.2a (−)(−)(−)(−)(−)(−)(−)(−)(−)
1856miR1856 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1861 miR1861hj  (−)(−)(−)(−)(−)(−)(−)(−)(−)
1862miR1862d, miR1862emiR1862d*(−)(−)(−)(−)(−)(−)(−)(−)(−)
1865miR1865-5p (−)(−)(−)(−)(−)(−)(−)(−)(−)
1870 miR1870  (−)(−)(−)(−)(−)(−)(−)(−)(−)
1871miR1871 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1873miR1873 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1876miR1876 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1878miR1878 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1879miR1879 (−)(−)(−)(−)(−)(−)(−)(−)(−)
1884 miR1884b  (−)(−)(−)(−)(−)(−)(−)(−)(−)
2863miR2863c (−)(−)(−)(−)(−)(−)(−)(−)(−)
2869 miR2869*(−)(−)(−)(−)(−)(−)(−)(−)(−)
2871 miR2871a*(−)(−)(−)(−)(−)(−)(−)(−)(−)
2873miR2873 (−)(−)(−)(−)(−)(−)(−)(−)(−)
2877miR2877 (−)(−)(−)(−)(−)(−)(−)(−)(−)
3979 miR3979-3p  (−)(−)(−)(−)(−)(−)(−)(−)(−)
3982miR3982-3p (−)(−)(−)(−)(−)(−)(−)(−)(−)
5508miR5508 (−)(−)(−)(−)(−)(−)(−)(−)(−)

This study also revealed the presence of miRNA sequences representing new members of known miRNA families, namely miR169r, miR1851.2, miR1437b, miR810b and miR3982 (Figs 1c, S2). Concerning the miR810 and miR3982 precursor structures, the annotated sequence for each precursor is extended relative to the annotated sequence in the miRBase.

Computational identification and experimental validation of novel miRNAs

Criteria for the annotation of novel plant miRNAs are based on both computational and experimental data. They include the excision from a stem–loop precursor structure, dcl dependence and conservation among species of both the stem–loop secondary structure and the mature miRNA sequence (Meyers et al., 2008). In the absence of genetic tools (i.e. dcl mutants), sequencing of both miRNA and miRNA* is recommended. Nonetheless, some rice miRNAs are registered solely on the basis of computational prediction or with little evidence of expression, and their authenticity is not certain.

To identify novel miRNAs from rice, we scanned the rice genome for stem–loop hairpin structures comprising the small RNA sequences identified in our libraries. For each small RNA sequence that had a perfect match in the rice genome, we determined the ability of the surrounding genomic sequences to fold into stem–loop hairpin structures. By using a maximum length of 3 kb, 219 loci that fulfilled the hairpin structure criterion for miRNA precursors were identified (Figs S3, S4; Table S4). The stem region for most of these precursor structures showed a high degree of complementarity, a feature characteristic of young recently evolved MIR genes (Vazquez et al., 2008). In those cases in which several small RNAs mapped at different positions along the same fold-back structure, the various small RNAs were consolidated into a single prospective MIRNA locus. In this respect, the production of two or more miRNAs from an miRNA precursor has been described already in rice and Arabidopsis (Zhu et al., 2008; Zhang et al., 2010). However, although we provide bioinformatics evidence for the 219 miRNA precursor structures supported by small RNA profiling, these miRNA precursor candidates should still be considered as miRNA candidates (accordingly, the names of the stem–loop precursors identified in this work are hyphenated to distinguish them from annotated miRNAs).

A search in the genome of different plant species revealed that 20 of the 219 miRNA candidates from rice have orthologue sequences in the genome of at least one of the other plant species tested (Table 2). In each plant species, the genomic regions surrounding the small RNA sequence also possessed intramolecular folding capacities, thus indicating that these hairpin-forming precursors might well represent previously uncharacterized miRNAs from rice.

Table 2. Nucleotide sequences and chromosomal locations of previously uncharacterized miRNAs from rice (Oryza sativa) that are conserved in other plant species
  1. The sequences given represent the small RNA sequences identified in the 454 dataset obtained from the leaf and root libraries whose precursor sequences have the capacity to adopt hairpin structures in rice as well as in the other plant species, monocots or dicots. A search for miRNA sequence homology was performed by BLASTN against National Center for Biotechnology Information (NCBI) genomes by allowing zero to three nucleotide substitutions. For those sequences mapping in the genome of any other species, the surrounding genomic sequences were analysed to confirm their ability to form fold-back structures. Their precursor structures are shown in Supporting Information Figs S3 and S4 (boxed in blue colour). At, Arabidopsis thaliana; Mt, Medicago truncatula; Pt, Populus trichocarpa; Sb, Sorghum bicolor; Vv, Vitis vinifera; Zm, Zea mays).

miR-2 CGGAGCCGGUGGUGGCGGUGG Sb  Chr122014352201515(+)
miR-35 AGAUAAAUGGUCAAACAUAUGAGA   Vv Chr33413489234134650(−)
miR-38 AUGACACCGUUGACUUCUUGACCA Zm Mt Chr4435880435638(−)
miR-40 ACUUUUGGAUAUGAUGUUUGACCA   Pt Chr41647530616475525(+)
miR-58 AAGACAAGUGGUCAAAUAUUGCAA   Pt Chr61011813810117881(−)
miR-65 AGUAGGUAGCAUAUAAGUAUGAGA   Mt Chr62219567422195395(−)
miR-68 AGAAUAAGACGAAUGGUCAAACG Sb, Zm Chr62725234627252589(+)
miR-73 GACUUACAUGUUUGACCGUUCGUC Sb  Chr72274006722739835(−)
miR-82 AAGACAGAUGGUCAAACGUUGGAA Sb  Chr81706148617061639(+)
miR-83 AUACGAAUGGUCAAACAUGUAAGA Sb Vv Chr81854256118542802(+)
miR-87 AUAAGACGGGUGAUCAAAGUUGGG Sb, Zm Chr91065076610650539(−)
miR-117 ACGGGUUUUGAUAGUUGAGGGAUC Sb  Chr172613427261710(+)
miR-122 AAGACAUGUGUAUAUGAUAGGUGA Zm Mt Chr13256937532569766(+)
miR-157 CCCUUGGCUGUGGAGAGAGAGA Sb Pt Chr514102351410880(+)
miR-163 GGAGUCUGACAUGCGUGCGAGUC  At, MtChr683655638366436(+)
miR-167 AGAGACCGGGAUGACACAUGCGAA Sb  Chr62754736927546632(−)
miR-171 ACAGUCGAACAAGUAUGAGGACCU Sb  Chr73001951030020256(+)
miR-182 UGUAGUCUGCAAGGAGAAGGC   Mt Chr91500830315009268(+)
miR-207 GGGAAAUCACGUGAAAGUUAUGAG   Pt Chr22360292123603940(+)
miR-218 UGGCGCGGAGGCCGCGGCGGUG Sb, Zm Chr859904325988987(−)

In this work, six of the 219 hairpin-forming structures were selected for experimental validation. Northern blot analysis demonstrated that small RNAs mapping to all six selected precursor structures accumulated in one or another rice tissue, some of these precursor structures generating various small RNAs (Fig. 2). The names assigned in the miRBase registry for the novel miRNAs identified in this work were osa-miR7692, osa-miR7693, osa-miR7694 and osa-miR7695. Three of the new miRNAs (osa-miR7692, osa-miR7694 and osa-miR7695) were found in leaf libraries, whereas osa-miR7693 appeared in leaf and root libraries. As for the two other experimentally validated miRNAs (osa-miR5150 and osa-miR5144), Chen et al. (2011) identified these particular miRNAs in a population of small RNAs from in vitro-cultured rice embryogenic calli. These two miRNAs were also present in our small RNA libraries from vegetative rice tissues (leaves, roots), and their accumulation has been validated experimentally in this work (Fig. 2).

Figure 2.

Predicted hairpin structures and experimental validation of novel microRNA (miRNA) candidates from rice (Oryza sativa). Small RNA sequences recovered from 454 sequencing data mapping into these structures are represented by black bars (thick bars, leaf libraries; thin bars, root libraries). Following miRNA nomenclature (Meyers et al., 2008), we named the small RNA sequences within the hairpin structure as miR-x.y, where x denotes a number for a particular miRNA precursor and y specifies the position of each sequenced small RNA along the precursor starting from the 5′ end (the suffix -5p or -3p was used to refer to the mapping arm within the stem–loop). Northern blot analysis of small RNAs, and corresponding ethidium bromide staining, is shown on the right side. The small RNA fraction obtained from 100 to 350 μg of total RNA, depending on the miRNA, was probed with synthetic oligonucleotide sequences complementary to the indicated sequences. The experimental validation of two miRNAs previously identified in small RNA libraries from in vitro-cultured rice embryogenic calli (Chen et al., 2011), osa-miR5144 and osa-miR5150, is also shown.

Elicitor responsiveness of rice miRNAs

Further analysis of the sequencing data revealed that treatment with fungal elicitors is accompanied by alterations in the accumulation of a repertoire of rice miRNAs, including miRNAs that are recognized as major regulators of gene expression in developmental processes (Fig. 3a,b). In some cases, the elicitor responsiveness of a miRNA was in the same direction in the two tissues (e.g. miR164abf and miR168a), whereas, in other cases, a different response to elicitors occurred depending on the tissue (e.g. miR156a-j and miR160abcd). Moreover, a different response could be observed at one or another time of elicitor treatment for particular miRNAs (e.g. miR159ab, Fig. 3a). Of note, the expression of miR168 was found to be up-regulated by fungal elicitors in the two rice tissues (elicitor responsiveness of miR168 was further confirmed by northern blot analysis; Fig. S5). Knowing that miR168 controls AGO1 homeostasis (Vaucheret et al., 2006), this finding anticipates an important role of miR168 functioning in the rice response to fungal elicitors.

Figure 3.

Expression and elicitor responsiveness of known microRNAs (miRNAs) from rice (Oryza sativa). (a, b) Fold change of miRNAs in elicitor-treated leaves (a) and roots (b) relative to nontreated tissues at 30 min or 2 h of treatment (light and dark bars, respectively). Fold change was calculated on the basis of normalized reads + 1 (treated vs untreated tissues). Only miRNAs represented by 20 or more reads in the datasets are shown. (c) Semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) of MIR156 precursors generating the same mature miRNA sequence (miR156abcdefghij) (c, control; e, elicitor-treated). Numbers in the lower part indicate the normalized relative abundance in the 454 sequencing data in each tissue and condition.

This analysis also revealed different expression patterns among members of a particular miRNA family (Fig. 3). In those cases in which several members of a family share the same mature miRNA sequences (i.e. miR156abcdefghij, miR160abcd or miR171bcdej), the 454 sequencing data do not provide information on which specific member is under tissue-specific control and/or exhibits elicitor responsiveness. In this work, we examined the expression of miRNA precursors for individual members of the MIR156 family. Thus, MIR156 has nine paralogous genomic loci that generate an identical mature miRNA sequence (miR156abcdefghij, miR156h and miR156j derive from the same locus). Semi-quantitative RT-PCR showed that the osa-MIR156b precursor is only expressed in roots, whereas osa-MIR156d is preferentially expressed in leaves (Fig. 3c). Although osa-MIR156b does not respond to elicitor treatment in roots, a significant reduction in osa-MIR156d accumulation occurred in elicitor-treated leaves relative to control leaves. Likewise, osa-MIR156hj exhibited a transient increase in elicitor-treated tissues (30 min of treatment) (Fig. 3c). Precursors for the other members of this family either accumulated at similar levels or were barely detected (data not shown). Reasonably, the reduction in miR156-related sequencing reads which was observed in elicitor-treated leaves relative to nontreated leaves might reflect the down-regulation of osa-MIR156d in this tissue. Likewise, the increase in the number of reads found in elicitor-treated roots might result from variations in the expression of osa-MIR156hj. Overall, these findings revealed a differential and dynamic regulation of the accumulation of an important number of rice miRNAs in response to treatment with fungal elicitors, which could also be observed among members of a particular miRNA family.

Expression of elicitor-regulated miRNAs from rice correlates inversely with the expression of their target genes

Global transcript profiling was carried out to investigate whether changes in the accumulation of elicitor-responsive miRNAs are accompanied by inverse trends in target gene expression. This microarray analysis identified genes whose expression was affected significantly (up- and down-regulation) in response to elicitors (P ≤ 0.05; fold changes ≥ 1.2-fold). Moreover, the examination of expression profiles of microarray data and 454-sequencing datasets revealed that a substantial number of validated target genes for known miRNAs exhibited an opposite response to elicitor treatment with respect to that of their corresponding regulatory miRNA (Table 3). Among them were miRNAs that target genes involved in protection against oxidative stress (e.g. miR528 and miR1879, targeting laccase and catalase, respectively) as well as genes that control plant development and hormone signalling (i.e. miR156, miR160, miR169 and miR393). This observation supports a functional interaction between these miRNAs and their corresponding target genes during the rice response to elicitor treatment. Consistent with the observed up-regulation of miR168 (Figs 3a, S5), microarray analysis confirmed the down-regulation of AGO1 genes in response to elicitor treatment (Table 3). Presumably, a pathogen-regulated adjustment of miR168 levels would contribute to the maintenance of the appropriate levels of AGO1, and accordingly of miRNA functioning, during the plant response to fungal elicitors.

Table 3. Elicitor responsiveness of rice (Oryza sativa) microRNAs (miRNAs) and their target genes
 miRNADirection of miRNA expressionTarget gene Fold change
30 min2 h
  1. Expression of known rice miRNAs in response to elicitor treatment, as determined by 454 sequencing (–, no alteration). The elicitor responsiveness of target genes was determined by microarray analysis. The fold change (elicitor-treated vs control nontreated tissue, P ≤ 0.05) for the target gene(s) for each miRNA family is shown.

  2. a

    New member of the indicated family.

  3. b

    This sequence was annotated as miR529* in miRBase, but was proposed to be the functional miRNA (Zhu et al., 2008).

  4. c,d,eThe target genes for these miRNAs have been described in the literature: cLi et al. (2010); dWu et al. (2009); eZhou et al. (2010). The remaining target genes are predicted but not experimentally validated. TF, transcription factor.

156miR156abcdefghijDownDownSBP domain-containing protein (OsSPL4) (SBP TF)c,d,eLOC_Os02g07780+ 1.65 (30′)
160miR160abcd, miR160eDownDownAuxin response factor (ARF10) (ARF TF)c,d,eLOC_Os06g47150+ 1.20 (30′)
 miR160fUpAuxin response factor (ARF16) (ARF TF)c,d,e LOC_Os10g33940−1.37 (2 h)
164miR164abf, miR164c, miR164dDownDownNAC domain-containing protein (NAC TF)c,d,eLOC_Os06g23650+ 1.33 (2 h)
    NAC domain-containing protein (NAC TF)c,d,eLOC_Os12g41680+ 1.25 (30′)
166miR166mDownSTART domain-containing protein (HB TF)c,d,eLOC_Os03g01890+ 1.33 (30′)
167miR167abc, miR167defghijDownDownRetinol dehydrogenase 14 cLOC_Os06g03830+ 1.31 (30′)
168miR168aUpUpARGONAUTE1 protein (AGO1)c,eLOC_Os04g47870−1.44 (2 h)
    ARGONAUTE1 protein (AGO1)c,eLOC_Os02g58490−1.32 (2 h)
169miR169aUpUpNuclear transcription factor Y subunit (NF-YA)c,d,eLOC_Os02g53620−1.24 (30′)
 miR169bc, miR169 fg, miR169hijklmDownDown Nuclear transcription factor Y subunit (NF-YA)c,d,eLOC_Os12g42400+ 1.26 (2 h)
 miR169no, miR169r aDownDownNuclear transcription factor Y subunit (NF-YA)c,d,eLOC_Os12g42400+ 1.26 (2 h)
171miR171iUpUpSCARECROW gene regulator (SCL) (GRAS TF)c,d,eLOC_Os06g01620−1.23 (2 h)
390miR390DownDownSTRUBBELIG-RECEPTOR FAMILY 6LOC_Os03g51040 + 1.25 (2 h)
    Wall-associated receptor kinase-like 10LOC_Os04g30060+ 1.91 (2 h)
    Leucine-rich repeat (LRR) family proteinLOC_Os04g45170+ 1.46 (30′)
393miR393bDownDownTransport inhibitor response 1 protein (TIR1)c,d,eLOC_Os05g05800+ 1.36 (30′)
394miR394UpRNA polymerase sigma factor rpoDeLOC_Os05g51150−1.31 (2 h)
396miR396cDownDownGrowth-regulating factor (GRF TF)c,d,eLOC_Os06g10310 + 1.25 (30′)
    Growth-regulating factor (GRF TF)c,d,eLOC_Os03g51970+ 1.20 (2 h)
    DeaminasecLOC_Os06g29430+ 1.17 (30′)
444miR444b1c1, miR444b2c2DownDownMADS-box transcription factor (MADS TF)c,d,eLOC_Os02g49840+ 1.28 (30′)
528miR528UpDownLaccaseLOC_Os01g62600 −1.20 (30′)
    Copper ion binding proteinLOC_Os01g03620−1.26 (30′)
    Copper ion binding proteinLOC_Os01g03640+ 1.20 (2 h)
529miR529bDownDownSBP-box gene family memberLOC_Os02g07780+ 1.65 (30′)
820miR820abcUpUpCellulose synthase like C12LOC_Os11g13650−1.22 (30′)
827miR827abUpDownOsWAK receptor-like protein kinaseLOC_Os02g56370+ 1.72 (2 h)
1425miR1425DownDownProtein kinasecLOC_Os01g49614+ 1.33 (2 h)
1430miR1430DownUpASYMMETRIC LEAVES 2 (MYB TF)LOC_Os05g34450 −1.23 (2 h)
    myb/SANT domain proteinLOC_Os03g13790+ 1.20 (30′)
1850miR1850.1UpDownPectinesterase inhibitor domainLOC_Os08g04650+ 1.20 (2 h)
1865miR1865-5pDownAspartate aminotransferaseLOC_Os02g14110+ 1.27 (30′)
1876miR1876UpDownEsterase/lipase/thioesteraseLOC_Os02g31200+ 1.48 (2 h)
1879miR1879UpUpCatalase isozyme BLOC_Os06g51150−1.60 (2 h)

The target genes for four of the experimentally validated miRNAs (osa-miR7692, osa-miR7693, osa-miR7694 and osa-miR7695) were predicted using the psRNATarget program ( Several predicted mRNA targets possessed a function coherent with plant response to pathogen infection, that is, genes involved in protection against oxidative stress and detoxification, disease resistance and receptor protein kinase genes (Table S5). Moreover, a search in the Genevestigator database ( revealed a pathogen-associated expression for most predicted targets, including infection by M. oryzae (Table S5). It is also true that, although microarray analysis demonstrated opposite trends in the expression of known, highly conserved miRNAs relative to their corresponding target transcripts, we could not observe such opposite expression patterns between novel miRNA candidates and their predicted target genes. However, the comparative analysis of data generated by transcriptome analysis and small RNA profiling excludes target genes that are under translational repression by their regulatory miRNA or target genes that exhibit a dynamic response to elicitor treatment (transcription profiles were determined at only two time points of elicitor treatment).

osa-miR7695, a novel DCL4-processed miRNA from rice

In this work, osa-miR7695 was characterized in more detail. Initially, we demonstrated that the entire osa-MIR7695 precursor structure is transcribed as a single transcriptional unit that comprises all five sequenced small RNAs mapping to this precursor (Fig. 4a,b). Moreover, osa-MIR7695-related small RNAs were consistently detected in leaves, but not in roots, of rice plants at different developmental stages (14-, 28- and 90-d-old rice plants; Fig. 4c).

Figure 4.

Processing of the osa-MIR7695 precursor and the production of transgenic rice (Oryza sativa) lines. (a) Structure of the osa-MIR7695 precursor. Arrows indicate the primers used for PCR amplification in (b). (b) Detection of the osa-MIR7695 precursor by nested reverse transcription polymerase chain reaction (RT-PCR). Sequencing of the 487-bp DNA fragment confirmed the specific amplification of the entire osa-MIR7695 precursor, whilst revealing that osa-MIR7695 is transcribed as a single transcriptional unit that comprises all five sequenced small RNAs. Control reactions without addition of the reverse transcriptase enzyme were included (RT–). (c) Accumulation of osa-MIR7695-related small RNAs (osa-miR7695.3-3p and osa-miR7695.5-3p) in leaves (L) and roots (R) of rice plants at different developmental stages (14, 28 and 90 d). Corresponding ethidium bromide (EtBr)-stained gels served as loading controls. (d) Northern blot analysis of osa-MIR7695-derived small RNAs in dcl1 and dcl4 rice mutants. The production of osa-MIR7695-related small RNAs is impaired in the dcl4 mutant. (e) Functional analysis of the osa-MIR7695 precursor in Nicotiana benthamiana leaves (rdr6IR line). The small RNA sequences produced from this precursor were detected by agroinfiltration of N. benthamiana leaves, followed by Northern blot analysis using oligonucleotides complementary to the five sequenced small RNAs produced by the osa-MIR7695 precursor. No signals were detected in control leaves transformed with the empty pCAMBIA vector for any of the probes (control). (f) Accumulation of osa-miR7695.5-3p in leaves of independently generated rice lines overexpressing the osa-MIR7695 precursor (ubi::MIR7695::nos) and control plants expressing the empty vector (pC). Results obtained for representative transgenic lines are presented. The same RNAs stained with EtBr are shown in the lower panel.

Next, we examined the accumulation of osa-MIR7695-related species in the loss-of-function Osdcl1 and Osdcl4 genetic backgrounds (Liu B et al., 2005, 2007). Although rice dcl2, dcl3 and rdr knockdown rice mutants have been reported (Urayama et al., 2010), this material was unavailable for our study. It is generally assumed that canonical 21-nucleotide miRNAs are generated by DCL1, and that, during evolution of MIR genes, a progressive shift in DCL usage from young to old MIR genes occurs, namely from DCL4/DCL3 to DCL1 (Voinnet, 2009). Of interest, the accumulation of osa-miR7695 species was found to be severely compromised in the dcl4 mutant, but remained unaffected in the dcl1 mutant (Fig. 4d). These findings, together with the observation that osa-MIR7695 shows a high degree of complementarity in the stem–loop precursor structure, supports the notion that osa-MIR7695 represents a novel, recently evolved miRNA-generating locus that is processed by DCL4 to produce multiple unique small RNAs (possible miRNAs and/or miRNA-like RNAs). Similar results have been reported previously for young miRNAs in other plant species (Rajagopalan et al., 2006; Ben Amor et al., 2009).

osa-miR7695 down-regulates the expression of an alternatively spliced transcript of the Nramp6 gene

In order to identify the target gene(s) for osa-miR7695, we generated transgenic rice lines overexpressing the osa-MIR7695 precursor. For this, the DNA fragment containing the osa-MIR7695 fold-back structure was PCR amplified from genomic DNA and cloned into a plant expression vector. Before rice transformation, we confirmed that the cloned osa-MIR7695 sequence was actually a functional source of osa-MIR7695-derived small RNAs. Indeed, transient expression assays in N. benthamiana leaves confirmed the processing of the osa-MIR7695 precursor and the production of all five expected small RNAs (Fig. 4e; details on transient expression assays in N. benthamiana are given in Methods S1).

Transgenic rice was produced by Agrobacterium-mediated transformation. As controls, transgenic rice plants were transformed with the empty vector (pCAMBIA 1300). Northern blot confirmed that the transgenic lines accumulated higher levels of osa-MIR7695-related small RNAs (Fig. 4f). T2 homozygous progeny plants were obtained which were phenotypically indistinguishable from wild-type plants.

Based on target gene prediction, the osa-miR7695.3-3p and osa-miR7695.5-3p small RNAs showed extensive sequence complementarity with two rice genes: Nramp6 (Natural resistance-associated macrophage protein 6) gene (Os01g31870) and a lectin-like receptor kinase gene (Os08g03002) (see Table S5). Concerning OsNramp6, cDNA data indicated that eight transcript variants were produced from this gene by alternative splicing (Fig. 5a). Among the various OsNramp6 splice variants, only the shortest transcript variant Os01g31870.8 contained complementary sites for osa-MIR7695-derived small RNAs (osa-miR7695.5-3p and osa-miR7695.3-3p), which were located at the 3′ untranslated region of this transcript variant (Fig. 5b). Interestingly, the accumulation of short transcripts of OsNramp6 (Os1g31870.8 splice variant) was found to be drastically reduced in transgenic rice lines overexpressing the osa-MIR7695 precursor, which was indicative of an osa-miR7695-mediated regulation of this transcript variant (Fig. 5c). By contrast, no significant alterations were observed in the level of nontarget transcripts of OsNramp6 (Os1g31870.4 splice variant) between transgenic and control plants (Fig. 5d). The accumulation of other OsNramp6 transcript variants could not be accurately determined because of their extremely low level of expression in rice leaves. In addition, no significant alterations were observed in the expression of other genes that were predicted for osa-MIR7695-related small RNAs, such as the lectin-like receptor kinase (Os08g03002, predicted target gene for osa-miR7695.3-3p and osa-miR7695.5-3p), nucleotide-binding site–leucine-rich repeat (NBS-LRR) resistance (Os12g18360, predicted target gene for osa-miR7695.2-5p) and lipase (Os12g01030, predicted target gene for osa-miR7695.2-5p) genes, in any of the osa-MIR7695 transgenic lines relative to control plants (data not shown).

Figure 5.

osa-miR7695 targets an alternatively spliced transcript of the OsNramp6 (Natural resistance-associated macrophage protein 6) gene. (a) Alternative splicing transcript variants of OsNramp6. The Os01g31870.1 transcript variant encodes the full-length protein and was taken as reference for intron/exon numbering. Only the short variant contains the target sites for osa-miR7695 (black bars at the 3′ untranslated region (UTR) of Os01g31870.8). Arrows in the 3′ region of Os01g31870.8 and Os01g31870.4 denote the primers used for expression analyses. (b) Complementarity of osa-miR7695.3p-related small RNAs with the 3′ UTR region of Os01g31870.8 transcripts. (c) Accumulation of short OsNramp6 transcripts (Os01g31870.8) in leaves of rice plants overexpressing the osa-MIR7695 precursor (ubi::MIR7695::nos) and vector control (pC) lines. Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was carried out using the ubiquitin 1 (Ubi1) gene as the internal control. (d) Accumulation of Os01g31870.4 OsNramp6 transcripts in transgenic rice lines overexpressing the osa-MIR7695 precursor. (e) Accumulation of short OsNramp6 transcripts (Os01g31870.8) in dcl4 and dcl1 mutants. (f) Accumulation of Os01g31870.8 OsNramp6 transcripts in leaves and roots of wild-type rice (cv Nipponbare) plants. (g) Accumulation of short OsNramp6 transcripts (Os01g31870.8) in control and elicitor-treated leaves of wild-type rice (cv Nipponbare) plants (black and grey bars, respectively). Each RNA was prepared from a pool of leaves from 50 rice plants. Differences in the accumulation levels were statistically significant (**,  0.001; *,  0.05). (h) Resistance of rice plants overexpressing the osa-MIR7695 precursor to infection by the rice blast fungus Magnaporthe oryzae. Five independent osa-miR7695 transgenic lines, four independent vector control plants as well as Nipponbare (WT) plants were assayed with similar results. Leaves were locally inoculated with a M. oryzae spore suspension (105 spores ml−1). Disease symptoms of leaves at 4 d post-infection are shown. Representative results from one of three independent experiments that produced similar results are presented. Error bars in (c–g) represent ± SD.

Knowing that the osa-MIR7695 precursor was processed in a DCL4-dependent manner, we assayed whether genetic inactivation of DCL4 also had an effect on the accumulation of target transcripts of osa-miR7695. As shown in Fig. 5(e), qRT-PCR analysis revealed that short OsNramp6 transcripts accumulated at significantly higher levels in dcl4 plants relative to wild-type rice plants, indicating that suppression of osa-MIR7695 precursor processing, and the subsequent production of osa-MIR7695-derived small RNAs, is accompanied by a higher accumulation of short OsNramp6 transcripts. In contrast, the accumulation of short OsNramp6 transcripts was not affected in the dcl1 rice mutant relative to wild-type plants, which was consistent with the observation that DCL1 is not required for the processing of the osa-MIR7695 precursor. In agreement with the finding that osa-miR7695 accumulates at high levels in leaves relative to roots in wild-type rice (see Fig. 4c), the short transcripts of OsNramp6 (Os01g31870.8) accumulated at significantly higher levels in roots relative to leaves at the different developmental stages assayed in this work (Fig. 5f). Collectively, these results strongly suggest that osa-miR7695 down-regulates transcript levels of the short transcript variant of OsNramp6 (Os01g31870.8).

Finally, we investigated the effect of elicitor treatment on the accumulation of short OsNramp6 transcripts in rice leaves. Although these transcripts accumulated to almost identical levels in control untreated leaves, elicitor treatment was accompanied by changes in the accumulation of short OsNramp6 transcripts during the period of elicitor treatment assayed in this work (Fig. 5g).

osa-miR7695 overexpression confers resistance to pathogen infection

To obtain further insights into the biological function of osa-miR7695, transgenic plants overexpressing the osa-MIR7695 precursor were tested for resistance to infection with the fungal pathogen M. oryzae. Five independent T2 homozygous lines overexpressing osa-MIR7695 were assayed. All the transgenic lines displayed enhanced disease resistance to pathogen infection relative to control plants (empty vector transgenic lines and wild-type plants) (representative results are presented in Fig. 5h). In agreement with the visual inspection, the inoculated leaves from osa-MIR7695 transgenic lines exhibited a lower percentage of diseased leaf area relative to inoculated leaves from nontransformed plants (Fig. S6). Depending on the line, the leaves from osa-MIR7695 transgenic lines exhibited 1.04–5.53% of their area affected by blast lesions at 4 d after inoculation. Under the same experimental conditions, leaves of control plants were affected in 15.21% of their area (Fig. S6). From this, it is concluded that osa-miR7695 accumulation positively regulates resistance to infection by the fungal pathogen M. oryzae.

osa-miR7695 occurs in japonica, but not indica, subspecies of cultivated rice

To investigate whether osa-miR7695 is conserved in plant species, its expression was analysed in other monocotyledonous species, as well as in dicotyledonous species. As shown in Fig. 6(a), osa-miR7695 could not be detected in any of the other plant species assayed here, suggesting that this miRNA may be specific to rice.

Figure 6.

Northern blot analysis of osa-miR7695 in plant species and rice varieties. (a) Monocotyledonous and dicotyledonous species. (b) Cultivated varieties of the genus Oryza (Oryza sativa and O. glaberrima). (c) Wild species of the genus Oryza. (d) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis of short OsNramp6 transcripts (Os01g31870.8) in cultivated rice varieties. Differences between indica and japonica cultivars were statistically significant ( 0.001). (e) qRT-PCR analysis of short OsNramp6 transcripts (Os01g31870.8) in wild species (O. nivara cultivars). Differences between O. nivara 1, 2, 4 and O. nivara 5, 6, 7) cultivars were statistically significant ( 0.001). Error bars in (d–e) represent ± SD.

The genus Oryza comprises two cultivated and 22 wild species ( The two cultivated species are O. sativa and O. glaberrima. Most of the cultivated rice varieties belong to the O. sativa group, which includes japonica and indica subspecies. Oryza sativa japonica is further differentiated into temperate japonica (japonica) and tropical japonica (javanica) varieties (Garris et al., 2005). In the present study, a collection of cultivated O. sativa and O. glaberrima cultivars was surveyed for osa-miR7695 expression. This collection included temperate japonica (eight accessions), tropical japonica (six accessions) and indica (seven accessions) subspecies from O. sativa, as well as two O. glaberrima cultivars. Interestingly, osa-MIR7695-related small RNAs were detected in all the japonica subspecies of O. sativa, both temperate and tropical japonica subspecies, but remained below northern blot detection limits in all the indica subspecies analysed here (Fig. 6b). In O. glaberrima cultivars, only osa-miR7695.5-3p small RNAs (the 24-nucleotide species) were found to accumulate in the leaves, whereas osa-miR7695.3-3p could not be detected in these cultivars. When examining osa-miR7695 accumulation in wild rice species, and similar to that observed in cultivated O. glaberrima, only the 24-nucleotide species of osa-miR7695.3-3p could be detected in some, but not all, wild rice species, and none accumulated osa-miR7695.5-3p (Fig. 6c).

Finally, the observed accumulation of osa-miR7695 in japonica subspecies (temperate and tropical japonica) correlated well with a lower level of OsNramp6 target transcripts (Os01g31870.8) relative to that observed in indica subspecies of O. sativa (Fig. 6d). Similarly, lower levels of OsNramp6 target transcripts occurred in wild rice species in which osa-miR7695 could be detected relative to those in which osa-miR7695 could not be detected (Fig. 6e).

Altogether, these results suggest that osa-miR7695 might represent a rice-specific miRNA that is produced in japonica, but not in indica, subspecies from the O. sativa genus. Moreover, differences in the processing of the osa-MIR7695 precursor appear to occur between the two groups of cultivated Oryza species, namely O. sativa and O. glaberrima. Processing of the osa-MIR7695 precursor in cultivated O. glaberrima subspecies resembles that of wild rice accessions (rather than cultivated O. sativa japonica subspecies).


In this study, we have shown that treatment with fungal elicitors is accompanied by dynamic alterations in the accumulation of a set of miRNAs from rice, both conserved and nonconserved miRNAs. Opposite expression patterns occur between elicitor-regulated known miRNAs and their corresponding target genes, suggesting that the observed elicitor-induced alterations in this set of miRNAs have an impact in shaping the plant transcriptome. The target genes for elicitor-regulated miRNAs are known to be involved in a variety of biological processes, such as stress responses, hormone regulation and development, or miRNA functioning. In particular, our data indirectly support a role for the miR168/AGO1 pair, and hence of the miRNA pathway, in coordinating the response of rice plants to fungal elicitors. Because miRNAs provide the quantitative regulation of target gene expression, rather than on–off regulations, the observed dynamic responses on miRNA accumulation could provide the fine-tuning of gene expression in different physiological processes. This, in turn, could enhance the plant's ability to escape from, resist or compensate for disease. This work also provides a rich source of expression data for a set of potential novel miRNAs showing elicitor responsiveness. Although we cannot rule out the possibility that some of the predicted miRNAs and target genes represent false-positive predictions, the possibility that some are young, recently evolved miRNAs, or even derive from proto-MIR genes, that exist without actual targets should be considered (Cuperus et al., 2011). Future experimental work will determine whether these miRNA candidates, and their predicted targets, are genuine miRNA/target gene partners.

osa-miR7695 represents a recently evolved miRNA that experienced natural and/or domestication selection during rice evolution

A large proportion of MIR genes appear to be generated by inverted gene duplication events that give rise to new MIR genes. The transcription of such young miRNA genes produces fold-back structures that are processed by DCL4 which, through the accumulation of mutations, lead to a switch to DCL1 processing (Vazquez et al., 2008; Cuperus et al., 2011). In this way, young MIR genes have stem structures with few bulges, are often processed imprecisely and give rise to miRNAs of variable length. By contrast, ancient MIR genes show reduced similarity in the fold-back arms and produce canonical miRNAs. In this work, we have shown that the processing of the osa-miR7695 precursor is largely dependent on DCL4. This finding, together with the long extensive base pairing within the stem region of osa-MIR7695, supports the hypothesis that this miRNA is probably an evolutionarily recent MIR gene. Moreover, osa-MIR7695-related small RNAs were detected in rice and not in any of the other monocotyledonous or dicotyledonous species analysed here, suggesting that osa-miR7695 evolved either after the divergence of the monocotyledonous and dicotyledonous lineages and/or during rice domestication.

The history of rice domestication remains an issue of debate. Although it is generally assumed that O. glaberrima originated from its wild ancestor O. barthii (Linares, 2002), controversy still exists about the wild ancestor for O. sativa species. It has been proposed that japonica and indica subspecies of O. sativa are the products of separate domestication events from pre-differentiated ancestral O. rufipogon populations (Londo et al., 2006). The detection of osa-miR7695 in japonica, but not in indica, subspecies is consistent with the idea that the two groups, japonica and indica, originated from different wild populations. In line with this, our results show that osa-MIR7695-related small RNAs are detected in certain wild rice varieties, but not in others. In addition, osa-miR7695 appears to occur in both temperate japonica and tropical japonica, which is consistent with the already described genetic relationship between the two japonica subgroups of O. sativa.

Finally, that osa-MIR7695 precursor processing is under molecular evolution is supported by the observation that the accumulation of osa-MIR7695-derived small RNAs differs in the two types of cultivated rice, O. sativa and O. glaberrima. Indeed, osa-MIR7695 is imperfectly processed in O. glaberrima, its processing pattern being more closely related to that occurring in wild rice (instead of that occurring in O. sativa cultivars). Moreover, osa-miR7695.3-3p and osa-miR7695.5-3p small RNAs are produced in japonica subspecies, both temperate and tropical japonica rice, whereas only osa-miR7695.5-3p accumulates in O. glaberrima subspecies, which might indicate that osa-MIR7695 precursor processing evolved during the domestication of japonica rice. In line with this, a role for miRNA genes as one of the driving forces in rice domestication has been proposed recently (Wang et al., 2012). Clearly, the availability of whole genome sequences for an increasing number of wild and cultivated rice species will greatly facilitate the reconstruction of the evolutionary history of newly identified miRNAs from rice.

osa-miR7695, a novel miRNA targeting an alternatively spliced transcript of OsNramp6 that contributes to pathogen resistance

We have reported that osa-miR7695 negatively regulates the accumulation of an alternatively spliced transcript of the Nramp6 gene. Several lines of evidence support this conclusion: first, the overexpression of the osa-MIR7695 precursor in transgenic rice, and the subsequent increase in the accumulation of osa-MIR7695-derived small RNAs, results in a drastic reduction in OsNramp6 transcripts; second, the accumulation of both osa-miR7695 and OsNramp6 target transcripts was affected in the dcl4 rice mutant; and third, the accumulation of short OsNramp6 transcripts is regulated by treatment with fungal elicitors. In addition, a good anti-correlation in osa-miR7695 and short OsNramp6 transcripts is observed in the different rice species and cultivars analysed in this work. Collectively, these results support the existence of a regulatory mechanism that integrates both miRNA function and mRNA processing for the control of OsNramp6 gene expression in rice plants. In this respect, a recent study using annotated gene models and publicly available high-throughput RNA sequencing data led the authors to propose that alternative splicing events might represent a mechanism for the attenuation of miRNA-mediated gene regulation in Arabidopsis plants (Yang et al., 2012). The results presented here fully support this notion and add another layer of complexity to the already known mechanisms in plant immunity based on miRNA- and mRNA processing-based regulation of gene expression. From an evolutionary perspective, a course of osa-miR7695 and OsNramp6 co-evolution can be reasoned in which OsNramp6 expression could escape direct repression by osa-miR7695 through alternative splicing events and the selective production of target and nontarget OsNramp6 transcripts. Interestingly, alternative splicing of R genes has been shown previously to play a role in pathogen defence, the production of these alternative splicing forms being required for full resistance. Some examples are the tobacco N gene conferring resistance to tobacco mosaic virus, and the RPS4 (RESISTANCE TO PSEUDOMONAS SYRINGAE 4) gene conferring resistance to P. syringae expressing AvrRps4 (Dinesh-Kumar & Baker, 2000; Zhang & Gassmann, 2003).

However, a role for distinct miRNAs in plant immunity has been documented. It was first described for miR393, an miRNA functioning in antibacterial defence by the repression of auxin signalling (Navarro et al., 2006; Staiger et al., 2012). More recently, miRNA regulation of innate immune receptors has been reported (Zhai et al., 2011; Li et al., 2012; Shivaprasad et al., 2012). In this study, we have shown that the overexpression of osa-miR7695 contributes to a phenotype of resistance to infection by the rice blast fungus M. oryzae. This finding supports a positive role for osa-miR7695 in disease resistance, most probably by controlling the accumulation of an alternative splicing variant of the OsNramp6 gene.

Concerning the biochemical function of NRAMP proteins, they are known to be involved in the transport of divalent metals and in the maintenance of metal homeostasis in a wide range of organisms, including plants (Cellier et al., 1996). Although several Arabidopsis NRAMP proteins have been shown to function as iron transporters (Curie et al., 2000; Thomine et al., 2000; Gross et al., 2003), very few have been assigned a physiological function. AtNramp3 and, to a lesser extent, AtNramp4 participate in iron mobilization in Arabidopsis and appear to be involved in resistance against the bacterial pathogen Erwinia chrysanthemi (Lanquar et al., 2005; Segond et al., 2009). Iron is an essential element for most living organisms, but, when in excess, iron induces the production of hydroxyl radicals which can cause multiple damage to cellular structures, eventually leading to death. Therefore, iron homeostasis must be tightly regulated in plant cells. Moreover, during pathogen infection, there is a competition between the host and the microorganism for iron. Of interest, a link between iron homeostasis and the expression of defence responses to pathogen attack has been documented in graminaceous species (Liu G et al., 2007; Lemanceau et al., 2009) and, more recently, in Arabidopsis plants (Kieu et al., 2012). Clearly, to be cost-effective, plant defence mechanisms need to be tightly regulated during pathogen infection. Transient alterations in osa-miR7695 accumulation would provide a flexible mechanism by which OsNramp6 could contribute to disease resistance. Further studies are needed to determine whether osa-miR7695/OsNramp6 functioning plays a role in the modulation of iron homeostasis. It would also be of interest to decipher the biological significance of alternative splicing in OsNramp6 in the context of both plant development and innate immunity, an aspect that remains to be investigated.

To conclude, the results presented here on miRNA functioning for the control of gene expression in disease resistance might also have broad implications in rice breeding programmes. Taking into account that rice has been adopted as the model species in cereal genomics, efforts to identify gene regulation networks that integrate miRNA functioning and alternative splicing events at target genes in rice will improve our understanding of the adaptation to pathogen infection in cereal species of agricultural importance.


We thank Dr X. Cao (Institute of Genetics and Developmental Biology, Beijing, China) for providing us with seeds of dcl1 and dcl4 mutants, Dr D. Baulcombe (University of Cambridge, UK) for the N. benthamiana RdR6i line, and Dr A. Sesma (Centre for Plant Biotechnology and Genomics, Universidad Politécnica de Madrid, Spain) for the M. oryzae strain Guy 11. We also thank E. Guiderdoni and the Rice Functional Genomics (REFUGE) international hosting platform established in Montpellier, France and funded by the Agropolis Foundation, D. Mieulet and M. Bundó for their assistance in rice transformation. This work was supported by grants BIO2009-08719/BIO2012-32838 and BIO2009-12004 from Ministerio de Ciencia e Innovación (MICINN) to B.S.S. and C.L., respectively, and the Consolider-Ingenio CSD2007-00036 to the Centre for Research in Agricultural Genomics (CRAG). We also thank the Generalitat de Catalunya (Xarxa de Referencia en Biotecnología and SGR 09626) for substantial support.