Index References

In most eukaryotes, small RNAs of 18–24 nucleotides (nt), including microRNA (miRNA) and small interfering RNA (siRNA), regulate gene expression by mRNA degradation, translational repression or chromatin remodeling (Bartel, 2004). The miRNAs are processed by Dicer ribonucleases from partially folded stem-loop precursor RNAs, which are transcribed as single-strand non-coding RNA. A number of miRNAs in the eukaryote genome down-regulate their target genes accurately, leading to normal development. By similar mechanisms, siRNAs are processed from longer double-stranded (ds) RNAs and induce gene silencing. In plants, the majority of siRNAs are derived from heterochromatin, transposons, repetitive sequences and transgenes. Plant endogenous siRNAs are classified into three categories with characteristics of their functions and derivation, natural-antisense transcript-derived siRNAs (nat-siRNAs), trans-acting siRNAs (ta-siRNAs), and heterochromatic siRNA (hc-siRNAs) (Borsani et al., 2005; Allen et al., 2005; Lippman et al., 2004).

Recent considerable research of small RNAs in plants have revealed that small RNA-mediated gene silencing in plants involves several pathways, which are regulated by many proteins such as ARGONAUTE (AGO1 and AGO7), DICER-LIKE (DCL1, DCL2, DCL3, and DCL4), HYPONASTIC LEAVES1 (HYL1), dsRNA-BINDING PROTEIN4 (DRB4), HUA ENHANCER1 (HEN1), HASTY (HST), RNA-DEPENDENT RNA POLYMERASE (RDR2, RDR6), SILENCING DEFFECTIVE3 (SDE3), and SUPRESSOR OF GENE SILENCING3 (SGS3) proteins (Vaucheret, 2006). Mutant analyses in Arabidopsis thaliana showed that plant development is influenced by proteins involved in small RNA-mediated gene silencing mechanism (Bohmert et al., 1998; Lu and Fedoroff, 2000; Chen et al., 2002; Telfer and Poethig, 1998; Jacobsen et al., 1999). For example, completely non-functional dcl1 mutant showed an embryo-lethal phenotype (Schauer et al., 2002). The ago1, hen1, hyl1, hst mutants and a leaky dcl1 allele exhibited overall developmental defects in whole tissues, which might be due to reduction of miRNA amount (Park et al., 2002; Han et al., 2004; Vazquez et al., 2004; Park et al., 2005; Mallory and Vaucheret, 2006). Various developmental abnormalities of these mutants indicated that small RNA-mediated gene silencing possesses important functions necessary for sustaining the life cycle of plants. To date, several kinds of miRNAs and siRNA have been cloned and are characterized in Arabidopsis and rice, two major model plants (Llave et al., 2002; Wang et al., 2004; Sunker et al., 2005b).

The male reproductive organ is one of the most drastically changing sites in plant development. In anthers, haploid microspores are produced through meiosis and developed into mature pollen. The pollen development, which is traditionally classified into 4 stages (uninuclear microspore, bicellular pollen, tricellular pollen, mature pollen grain), is regulated by a complex gene expression in both gametophytic and sporophytic tissues (Yang and Sundaresan, 2000; Kapoor et al., 2002; Durbarry et al., 2005). Pina et al. (2005) conducted pollen transcriptome in A. thaliana, and clarified an expression profile of gene families involved in the small RNA-mediated gene silencing pathway, including AGO1, AGO7, DCL1, DCL2, DCL3, HEN1, RDR2, RDR6, and SDE3. They showed that all the genes except for AGO7 are expressed in uninuclear microspore and bicellular pollen (the early developmental stage). Also all the genes, except for AGO1, were not expressed in tricellular microspore and mature pollen grain (the late developmental stage), and AGO1 transcripts were identified in the tricellular pollen, suggesting that small RNAs might not function in the late stage of pollen development. Because mature pollen grains have many important biological roles, such as pollen germination, pollen-stigma interaction, pollen-tube penetration, and double fertilization (Hulskamp et al., 1995; Jiang et al., 2005; Dresselhaus, 2006), it is necessary to determine whether small RNAs exist and function in the late developmental stage of the male organ. In this study, in order to clarify that small RNAs exist during pollen maturation, we performed molecular cloning and expression analysis of small RNAs in the tricellular stage of rice anther.

Total RNA was isolated from anthers with tricellular pollen of rice (Oryza sativa spp. japonica cv. Nipponbare). We constructed a small RNA library using the total RNAs that range in size from 18 to 28 nt, with DynaExpress miRNA Cloning Kit (BioDynamics Laboratory, Tokyo, Japan) and TOPO TA Cloning Kit with pCR2.1 vector (Invitrogen, Carlsbad, CA, USA). Subsequently, 864 randomly-selected clones were sequenced using ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) and were characterized by BLAST ver. 2.2.15 (http://blast.ddbj.nig.ac.jp/top-j.html; Altschul et al., 1997) and RAP (Rice Annotation Project)-BLAST searching (http://rapdb.dna.affrc.go.jp/tools/blast; Ohyanagi et al., 2006). Out of the 864 selected clones, 169 were able to be analyzed as data with high quality and perfectly matched the rice genome. While 113 clones of the 169 clones corresponded to degraded products of rRNA and tRNA, most of the remaining 56 clones (48/56, 85.7%) were judged as small RNAs because the 48 clones ranged in length from 18 to 24 nt as small RNAs (Fig. 1). The most frequent lengths were 21 and 24 nt (35.7% and 17.9%, respectively), being consistent with the size distribution of miRNAs reported in rice previously (Sunker et al., 2005a). We categorized these 56 clones with mfold program (http://frontend.bioinfo.rpi.edu/applications/mfold/cgi-bin/rna-form1.cgi; Zuker, 2003) and the results were as follows: two clones were known miRNAs (miR166 and miR167) (Reinhart et al., 2002; Sunker et al., 2005a), ten were putative siRNAs from retrotransposon which might be categorized in hc-siRNAs, eight were estimated as potential miRNAs by stem-loop structures. Also none of nat-siRNAs were matched to natural cis-antisense transcripts contained in two dataset, the RAP database and the Arabidopsis nat-siRNA dataset (Jin et al., 2008). The characteristics of the 56 clones suggest the existence of small RNAs in the late developmental stage of rice anthers. Nucleotide sequence comparison among the 56 small RNAs in this study showed that all of them were unique.

View Details

Fig. 1 Size distribution of 56 small RNAs cloned from rice anthers at the tricellular pollen stage.

We further focused on validating existence of the small RNAs identified with our sequence analysis though small-RNA cloning was not saturated by the present small-scale sequencing using the conventional cloning method. In the case of miRNA, Lu et al. (2008) suggested that high-throughput deep-sequencing is essential for effective and confident annotating of non-conserved miRNA. They sequenced more than four million small RNAs from rice by using massively parallel sequencing, and found that most of the reported rice-specific miRNA candidates, except conserved miRNAs in Arabidopsis, were mis-annotated as miRNA even though they were predicted to form hairpin structures and confirmed by RNA gel blot analyses. Therefore, the miR166 and miR167, which were validated as conserved miRNAs with the RAP-DB, are available markers at present to ask if small RNAs occur in the late developmental stage of rice anthers. In our RNA gel blot analysis, both miR166 and miR167 were observed in anthers from the uninuclear microspore stage to the tricellular pollen stage (Fig. 2). Their accumulation was more abundant in anthers of the tricellular pollen stage than in leaves, indicating that these miRNAs might be derived from pollens, not from vegetative tissues of the anthers, such as epidermal cell layer. A target gene of miR166 encodes a homeodomain-leucine zipper (HD-ZIPIII) transcription factor, which regulates radial polarity of the leaves and stems of the shoot in Arabidopsis (Rhoades et al., 2002; Emery et al., 2003) and shoot apical meristem (SAM) formation in rice (Nagasaki et al., 2007). Identification of miR166 in the maturing pollen suggests that negative regulation of HD-ZIPIII is needed in the pollen development and that an accurate polarity of the pollen grains, which might affect an asymmetric division of microspore mitosis following organelle arrangement, would be maintained by the expression of the HD-ZIPIII regulated by miR166. miR167 is known to be highly expressed in floral organs of Arabidopsis, and target genes of miR167 are auxin response factor genes, ARF6 and ARF8 (Reinhart et al., 2002; Rhoades et al., 2002). In Arabidopsis, miR167 regulates anther development, and overexpression of miR167 causes a male sterile phenotype (Ru et al., 2006; Wu et al., 2006). These observations, suggesting the accurate regulation of maturing male gametes by miR167, are consistent with our data in which miR167 was highly accumulated during early developmental stage of anthers and until the tricellular pollen stage.

View Details

Fig. 2 RNA gel blots of two known small RNAs, miR166 and miR167. Anther stage classification, which was divided into three stages by the number of cells in microspore, was followed by Tsuchiya et al. (1992). Ten micrograms of total RNA from roots (R), leaves (L), and anthers containing uninuclear microspores (1), bicellular pollens (2) and tricellular pollens (3) were resolved in a denaturing 10% polyacrylamide gel, transferred to a nylon membrane, and hybridized with a 32P end-labeled oligonucleotide complementary to miR166 or miR167. rRNA stained with ethidium bromide was used as loading controls.

We conclude that small RNAs would have functions in the late developmental stage of the male organ in rice; they might be processed in the early stage of anthers and maintained until the later stage. Honys and Twell (2004) performed comprehensive transcriptome analysis of male gametophyte development in Arabidopsis, and found large-scale repression of various genes during transition from bicellular to tricellular pollen. This characteristic down-regulation of genes during pollen maturation, which increases the proportion of male gametophyte-specific transcripts, might be strictly regulated by small RNAs. Further analysis of small RNAs by massively parallel sequencing will lead to identification of novel gene networks regulating the haploid male gametophyte development in plants.

This work was supported in part by Grants-in-Aid for Special Research on Priority Areas (Nos. 18075003), a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Integrated research project for plant, insect, and animal using genome technology IPG-0019) and Grants-in-Aid for the 21st Century Center of Excellence Program from the Japan Society for Promotion of Science (JSPS) to MW, a grant for the Ministry of Agriculture, Forestry and Fisheries of Japan (Integrated research project for plant, insect, and animal using genome technology IPG-0018) to MKK, and a grant for the Ministry of Agriculture, Forestry and Fisheries of Japan (Integrated research project for plant, insect, and animal using genome technology GPN-0007) to AM. K.S. is the recipient of the Research Fellowship for Postdoctoral Fellowships for Young Researchers of JSPS. The authors are grateful to Ayako Chiba (Iwate University), Hiromi Masuko, Kosuke Matsumoto, Masumi Miyano, Hiromi Shoji, Yuta Tsunaga and Ayumi Yamakawa (Tohoku University) for their technical assistance.