This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 142/1/280    most recent pp.106.084475v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (9) Citing Articles via Google Scholar Google Scholar Articles by Xie, K. Articles by Xiong, L. Search for Related Content PubMed PubMed Citation Articles by Xie, K. Articles by Xiong, L. Agricola Articles by Xie, K. Articles by Xiong, L.

SYSTEMS BIOLOGY, MOLECULAR BIOLOGY, AND GENE REGULATION

Genomic Organization, Differential Expression, and Interaction of SQUAMOSA Promoter-Binding-Like Transcription Factors and microRNA156 in Rice^1,[W]

National Key Laboratory of Crop Genetic Improvement, National Center of Plant Gene Research (Wuhan), Huazhong Agricultural University, Wuhan 430070, China

	ABSTRACT

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 
Transcription factors play essential roles in the developmental processes of plants. Many such factors are regulated by microRNAs (miRNAs). SQUAMOSA (SQUA) promoter-binding-like (SPL) genes encode plant-specific transcription factors, some of which contain complementary sequences of miRNA156. In this study, 19 rice (Oryza sativa) SPL (OsSPL) genes and 12 rice miRNA156 (OsmiR156) precursors were identified in the rice genome. Sequence and experimental analysis suggested that 11 OsSPL genes were putative targets of OsmiR156. Plant SPL proteins were classified into six subgroups based on the phylogenetic analysis of SQUA promoter-binding protein domain. Diverse exon-intron structures and distinct organizations of putative motifs beyond the SQUA promoter-binding protein domains were identified in the OsSPL gene family. Transcript level analysis of OsSPL genes in various rice tissues and organs revealed different tempospatial expression patterns. More than half of the OsSPL genes including most OsmiR156-targeted genes are predominantly expressed in the young panicles, whereas OsmiR156 genes are predominantly expressed in the young shoots and leaves of rice. Overexpression of two OsmiR156 genes (OsmiR156b and OsmiR156h) in rice resulted in severe dwarfism, strongly reduced panicle size, and delayed flowering, suggesting that OsmiR156 and OsSPL target genes are involved in various developmental processes, especially the flower development of rice. Different patterns of transcript changes (decreased or unchanged) of different target genes in same tissue and of same target gene in different tissues detected in the OsmiR156-overexpressing plants suggested diverse interactions between OsmiR156 and OsSPL target genes in a tissue-specific manner.

SQUAMOSA (SQUA) promoter-binding-like (SPL) genes represent a family of plant-specific transcription factors (Klein et al., 1996; Cardon et al., 1999). The common feature of SPL genes is that SPL proteins contain a highly conserved DNA-binding domain (SQUA promoter-binding protein [SBP] domain). This domain features a novel zinc finger motif that contains two zinc-binding sites assembled as Cys-Cys-His-Cys and Cys-Cys-Cys-His, respectively (Yamasaki et al., 2004). SBP1 and SBP2 are the foremost SPL genes isolated from Antirrhinum majus and were named based on the in vitro-binding activity of SBP1 and SBP2 to the cis-element upstream of the gene SQUA (Klein et al., 1996). SQUA is a member of the MIKC group of the MADS-box gene family in Arabidopsis that specifies flower meristem identity (Huijser et al., 1992; Jack, 2004). The expression of SBP1 and SBP2 is developmentally regulated and transcriptional activation of SBP1 and SBP2 precedes that of SQUA (Klein et al., 1996). In Arabidopsis, AtSPL3 was identified as an ortholog of SBP1 and SBP2 (Cardon et al., 1997, 1999). The AtSPL3 recognized a conserved cis-element in the promoter region of APETALA1 (AP1), an ortholog of SQUA in Arabidopsis (Cardon et al., 1997). The sequence of the cis-element to which SPLs bind in vitro was predicted to be TNCGTACAA in Arabidopsis (Cardon et al., 1999). Another SPL protein (BpSPL1) isolated from Betula pendula also showed binding ability to the cis-element of BpMADS5, a close homolog of Arabidopsis FRUITFULL (Lannenpaa et al., 2004). Recently, a SBP domain protein Copper Response Regulator, which regulates nutritional copper signal and recognizes a GTAC core cis-element, was identified in single-cell green plant Chlamydomonas (Kropat et al., 2005).

In the Arabidopsis genome, 16 putative SPL genes that contained the SBP domain were predicted based on sequence analysis, and several AtSPL genes were thought to have roles in the regulation of plant development (Cardon et al., 1999). Constitutive expression of AtSPL3 in Arabidopsis caused early flowering that is similar to the phenotype that resulted from the overexpression of AP1 (Cardon et al., 1997). However, constitutive expression of AtSPL3 in the ap1 mutant showed that AP1 was not required for the early flowering phenotype of the AtSPL3 transgenic plants (Cardon et al., 1997). The analysis of three independent transposon-tagged atspl8 mutants indicated that AtSPL8 was involved in the regulation of microsporogenesis, megasporogenesis, trichome formation on sepals, and stamen filament elongation (Unte et al., 2003). The T-DNA insertion mutant of atspl14 showed elongated petioles and enhanced leaf margin serration (Stone et al., 2005). In maize (Zea mays), the liguleless1 mutant of an SPL gene showed defects in ligule and auricle formation and no blade-sheath boundary (Moreno et al., 1997).

Discovery of microRNAs (miRNAs) has profoundly enriched our understanding of gene regulation in animals and plants. MiRNAs are small RNA molecules (20–24 nucleotides) that can bind to the mRNAs of target genes by imperfect base pairing and result in cleavage of mRNA or repression of translation through a RNA-induced silencing complex (Bartel, 2004). Numerous miRNAs have recently been identified (Reinhart et al., 2002; Bonnet et al., 2004; Sunkar and Zhu, 2004; Sunkar et al., 2005). More than 2,900 miRNA entries have been registered in the miRNA database (miRbase; http://microrna.sanger.ac.uk/, release 7.0), including 117 from Arabidopsis and 173 from rice (Oryza sativa). Increasing evidence shows that plant miRNAs constitute a substantial fraction of the gene regulatory networks for diverse aspects of plant development such as leaf polarity (Kidner and Martienssen, 2004), leaf morphogenesis (Palatnik et al., 2003), flowering development (Chen, 2004; Millar and Gubler, 2005), root cap formation (Wang et al., 2005), auxin signal regulation (Guo et al., 2005; Mallory et al., 2005), and stress response (Jones-Rhoades and Bartel, 2004; Sunkar and Zhu, 2004). About half the target genes of miRNA are transcription factors (Bartel, 2004). Computational analysis indicated that some Arabidopsis SPL genes were also regulated by miRNA genes of the miRNA156 family (Rhoades et al., 2002; Bonnet et al., 2004). The miR156 and its target genes are thought to be involved in some important developmental processes since overexpression of AtmiR156 in Arabidopsis resulted in various phenotypic changes such as increased number of leaves, delayed flowering, and decreased apical dominance (Schwab et al., 2005).

As one of the most important crops worldwide, rice has become a model plant of monocot species for functional genomics studies. Systematic analysis of SPL and miR156 genes in rice will certainly improve our understanding of the complex regulatory networks in monocot species. Here we report on the analyses of SPL and miR156 gene families in rice genome for their genomic organization, gene structures, motif composition, and expression levels in various tissues and organs of rice. Moreover, two OsmiR156 precursors were overexpressed in rice to study the functional relationship of SPL and miR156 genes.

	RESULTS

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 

Identification of OsSPL and OsmiR156 Genes in the Rice Genome

We used BLAST to search the GeneBank database with reported SPL protein sequences as queries, and only 15 putative SPL genes were identified in the rice genome. To identify all putative rice SPL genes, we searched the annotation database of rice (data downloaded from The Institute for Genomic Research [TIGR], Beijing Genomic Institute, and Knowledge-Based Oryza Molecular Biological Encyclopedia [KOME]) with a profile Hidden Molkov model (pHMM) of the SBP domain, since pHMM was considered to be a more efficient approach than pairwise comparison (Eddy, 1998). By removing the redundant sequences, we identified 22 putative protein sequences by pHMM searches with an E value less than 1E-5. However, two of them (AK10915 and 11681.m02353 [gene model number of TIGR rice pseudomolecular 3.0]) had no zinc finger motif that features the SBP domain and were excluded from further analysis. Two other gene models (11681.m02775 and 11681.m02776 that are closely linked but interrupted by a 27.6 kb of genomic sequence) can form one complete SBP domain only when the two gene models are fused. Therefore, at least 19 putative SPL genes existed in the rice genome. For convenience of description, a systemic nomenclature of OsSPLxvn (x for the serial number of SPL genes based on their order on chromosomes, vn for variant number of differentially spliced transcripts) was adopted for rice SPL genes (Table I ). To date, 11 OsSPL genes are supported with full-length cDNAs in the KOME database (Table I). In this study, the complete open reading frames of all the OsSPL genes except OsSPL19 were isolated from indica rice Minghui 63 with gene-specific primers (Supplemental Table I). The OsSPL19 was excluded from the following analysis unless specially pointed out, since it is likely a pseudogene.

A comparison of the OsmiR156 mature sequence to the OsSPL sequences showed that 11 OsSPLs contained sequences that are complementary to the OsmiR156 mature sequence, with one mismatch at the 14th nucleotide (for miR156a–j and miR156l), or one mismatch at the first (for OsmiR156l) or last nucleotide (Fig. 1A ). Another search of all rice genomic sequences with the mature sequence of OsmiR156 resulted in no additional significant matches except the 11 OsSPL and OsmiR156 genes, which suggests that OsmiR156 may specially target OsSPL genes in rice. The target sites of OsmiR156 are located in the coding regions (downstream of the SBP domain) except OsSPL4 and OsSPL13 that contain the target sites in the 3'-untranslated region (UTR). Interestingly, the amino acid residues encoded by the miR156 targeting sites are exactly the same or highly conserved (for one residue) for the nine target OsSPL genes (Fig. 1A) and 10 target AtSPL genes in Arabidopsis (Rhoades et al., 2002).

View larger version (39K): [in this window] [in a new window]   Figure 1. Sequence analysis of OsmiR156 genes. A, Sequence alignment of OsmiR156 mature sequences with complementary sequences of OsSPL genes. The conserved amino acid sequence encoded by the target sequences is shown at the bottom. The dots between miR156 and targeted OsSPL sequences indicate mismatches. The complementary sequences of OsmiR156 in OsSPL4 and OsSPL13 (indicated by an asterisk) are not in the coding regions. B, Distribution of OsSPL and OsmiR156 genes in rice genome. Locations of closely linked genes are magnified. OsmiR156b and OsmiR156c are mapped to the same transcript (corresponding to the cDNA AK110791). OsmiR156j and OsmiR156h are also mapped to same transcript (corresponding to the cDNA AK103769). OsSPL4 and OsSPL5 are closely linked with OsmiR156d, and OsSPL17 and OsSPL18 are closely linked with OsmiR156k. The putative target genes of OsmiR156 are underlined.

Phylogenetic Analysis of the OsSPL and OsmiR156 Gene Families

To study the evolutional relationships of SPL genes in plants, we collected a data set of 48 putative SPL protein sequences, including 18 from rice, 16 from Arabidopsis, and 14 from other flowering plants (maize, A. majus, and B. pendula) for phylogenetic analysis (gene names and accession numbers are listed in Supplemental Table II). Alignment of the full-length protein sequences showed no consensus sequences when the SBP domains were masked. Therefore, only the protein sequences of SBP domain (Supplemental Table II) were used for phylogenetic analysis.

The phylogenetic tree suggested that the plant SPL family was evolutionally diversified (Fig. 2 ). The 48 plant SPLs were classified into six subgroups (S1–S6) according to the unrooted phylogenetic tree. Generally, SPLs from Arabidopsis and rice are almost evenly distributed in the six subgroups (Fig. 2). Within specific subgroups (such as S5), however, rice SPLs had closer relationships to SPLs of maize than to SPLs of Arabidopsis. These results suggest that plant SPL genes may be derived from common ancestors, but some of which may have been differentiated separately in monocotyledon and dicotyledon plants. Interestingly, the miR156-targeted SPLs (only SPLs in rice and Arabidopsis were analyzed) were distributed in four subgroups (S1, S2, S3, and S6) but not in the other two subgroups (S4 and S5).

View larger version (41K): [in this window] [in a new window]   Figure 2. Unrooted tree of SPL family based on the protein sequences of SBP domains. The numbers at branching sites indicated the posterior probability values for nodal support. The SPL genes from Arabidopsis and rice that contained complementary sequence of miR156 were underlined. The names of Arabidopsis SPL genes were downloaded from the Arabidopsis Functional Genomics Network (http://web.uni-frankfurt.de/fb15/botanik/mcb/AFGN/Huijser.htm); other plant SPL genes are from SwissProt. The SBP domain sequences of all these genes and the accession numbers of genes from plants other than rice are listed in Supplemental Table I. The accession numbers of SPL genes of rice are listed in Table I. Am, A. majus; At, Arabidopsis; Zm, maize; Bp, Betu psy; Os, rice.

Diverse Exon-Intron Structure and Motif Composition in OsSPL Family

Generally, the number of exons in the coding region and the length of the coding sequences are similar for the OsSPL genes within phylogenetic subgroups but quite different between phylogenetic subgroups (Supplemental Fig. 4). However, the exon-intron structures of genes in S1 and S6 are less conserved than in the other four subgroups. An intron existed in the SBP domains of all OsSPL genes except OsSPL4v1 and OsSPL2. The position of this intron is extremely conserved in rice (located in the codon of the 49th amino acid of SBP domain) and in Arabidopsis (Cardon et al., 1999). Nevertheless, the length of this intron (from none to 27 kb) differs sharply among OsSPL genes (Supplemental Fig. 4). Moreover, repetitive sequences were found in the intron for some OsSPL genes (data not shown), which may indicate that the different length of this intron may be partially resulted from the expansion of repetitive sequences. Based on the full-length cDNAs of OsSPL genes from the KOME database, at least eight OsSPLs (OsSPL1, OsSPL4, OsSPL5, OsSPL6, OsSPL10, OsSPL12, OsSPL13, and OsSPL17) have alternatively spliced forms (Supplemental Fig. 4). Except OsSPL4, all other three OsSPL genes have their alternative splicing sites in the UTR regions.

According to the solution three-dimensional structure of AtSPL4 and AtSPL7 proteins (Yamasaki et al., 2004), the SBP domain of OsSPLs can be divided into four motifs (Fig. 3A ): zinc finger 1 (Zn1), zinc finger 2 (Zn2), joint peptide (JP) of Zn1 and Zn2, and nuclear location signal (NLS). Zn1 (CX₄CX_10–11HX₅C) is a CCHC-type zinc finger and divergent in the SBP domain, especially for the S1 and S6 subgroups (Supplemental Fig. 5); Zn2 (CX₂CX₁₅CX₆H) is a CCCH-type zinc finger. JP is less conserved than the two zinc fingers (Fig. 3A) and was thought to have a crucial role in the characteristic packing of SBP domains and thereby modifies the protein-DNA interaction process (Yamasaki et al., 2004). The bipartite NLS motif is highly conserved not only in the SPL family but also in other proteins (Dingwall and Laskey, 1991). Except for SBP domain, no analysis of the remaining region of SPL proteins for conserved domains or motifs has been reported. By using the software Multiple EM for Motif Elicitation (MEME) version 3.0, 10 putative motifs were identified in OsSPL proteins with E values less than 1.00E-30 (Table II ; Fig. 3B). A search of Inter-Pro database by Inter-ProScan (Mulder et al., 2005) with these putative motifs from MEME showed that most motifs (except 4, 5, 7, and 8) have matches in the Inter-Pro database (Table II). Occasionally, one motif (such as the JP in OsSPL14 and motif 4 in OsSPL1) appeared twice in single protein sequence (Fig. 3B). The MEME program considered the SBP domain as three motifs, Zn1, Zn2-NLS, and JP. Motif 7 is encoded by a conserved sequence that is complementary to the mature sequence of miR156 (Fig. 1). Although OsSPL6 contained a predicted motif 7 (Fig. 3B), it cannot be considered the target gene of miR156 since the DNA sequence of the predicted motif 7 has six mismatches with the mature sequence of miR156. Notably, motif 6 is an ankyrin repeat (Interpro accession no. IPR002110), which is one of the most common motifs for protein-protein interaction and has been found in proteins with diverse functions such as transcriptional initiators, cell-cycle regulators, cytoskeleton, ion transporters, and signal transducers (Lux et al., 1990; Gorina and Pavletich, 1996; Luh et al., 1997; Batchelor et al., 1998; Mosavi et al., 2002). The putative functions for the other motifs (4, 5, 8, 9, and 10) generated from MEME are currently unknown.

View larger version (39K): [in this window] [in a new window]   Figure 3. Organization of putative motifs in OsSPLs. A, Sequence LOGO view of the consensus SBP domain sequences based on the 48 plant SPLs in Figure 2 (pHMM logo view of each SPL subgroup is shown in Supplemental Fig. 1). The height of the letter (amino acid) at each position represents the degree of conservation. The rectangles at the top of the plot indicate the motifs of the SBP domain. B, Putative motifs in OsSPLs identified by MEME. Numbered color boxes represent different putative motifs (annotations are listed in Table II). The expected values calculated by MEME are shown after gene names. S1 to S6 indicates the phylogenetic subgroups from Figure 2. The putative target genes of OsmiR156 are underlined.

Diverse gene structures of OsSPL and OsmiR156 genes prompted us to investigate their expression profiles in various tissues or organs of rice. Preliminary experiments by RNA gel-blot analysis suggested that most of the OsSPL genes had very low expression levels in most rice tissues and organs (data not shown). Therefore, semiquantitative reverse transcription (RT)-PCR was used to detect the expression of OsSPL genes in various rice tissues including panicles at different developmental stages since some SPL genes of dicotyledon plants were thought to be involved in inflorescence development (Klein et al., 1996; Cardon et al., 1997, 1999; Unte et al., 2003).

Based on the RT-PCR of the non-target-site regions in 3' UTR, the expression patterns of OsSPL genes can be classified into three types according to their expression patterns (Fig. 4A ). The first type of genes (including OsSPL7, OsSPL12, OsSPL14, OsSPL16, OsSPL17, and OsSPL18) expressed relatively stronger in young panicles than other tissues investigated. The second type of genes (OsSPL2, OsSPL4, OsSPL8, OsSPL10, OsSPL11, and OsSPL15) expressed in most of the tissues investigated, but had higher expression levels in stem, leaf sheath, and young panicles than in other tissues. The third type of genes (OsSPL1, OsSPL3, OsSPL5, OsSPL6, OsSPL9, and OsSPL13) expressed in all the tissues investigated without obvious difference of expression level. Interestingly, Most of the target genes showed higher expression levels in young panicle than in most other tissues. In addition, differentially spliced transcripts were detected in different tissues for several OsSPL genes such as OsSPL4, OsSPL5, and OsSPL13 (Fig. 4A).

View larger version (33K): [in this window] [in a new window]   Figure 4. Tissue-specific expression patterns of OsSPL genes and OsmiR156. A, Semiquantitative RT-PCR of OsSPLs and OsmiR156 precursors in 13 different tissues or organs of rice. The rice Actin1 gene was used as the internal control. The miR156-targeted OsSPL genes were underlined and two pairs of primers (Supplemental Table I) were used for each gene, one located downstream of the miR156 complementary site and the other (labeled with an asterisk) amplifying the region covering the targeting site. PCR was failed for OsSPL17 with the target site-spanning primers. B, Real-time quantitative PCR analysis of four OsSPL genes. The relative expression levels were calibrated based on sample 6 (0.5–1 cm panicle). C, PAGE RNA gel-blot analysis of mature miR156 using an end-labeled complementary sequence (5'-GTGCTCACTCTCTTCTGTCA-3'). The blot was stripped and rehybridized with oligonucleotide probe complementary to U6 RNA as a loading control. Amount of RNA loaded in each lane was indicated by ethidium bromide staining. Tissues or organs: 1, Root at tillering stage; 2, stems at tillering stage; 3, leaf sheath at 4-leaf stage; 4, leaf lamina at 4-leaf stage; 5, young panicle shorter than 0.2 cm; 6, 0.5 to 1 cm young panicle; 7, 3 to 5 cm young panicle; 8, longer than 10 cm young panicle; 9, stamen; 10, pistil; 11, shoot at 1 d after germination; 12, 3-week-old shoot; 13, 3-week-old etiolated shoot.

The reliability of semiquantitative RT-PCR was checked by real-time quantitative PCR analysis of four OsSPL genes (Fig. 4B). The results showed that the band intensity in semiquantitative RT-PCR analysis generally agreed well with the relative expression level by real-time quantitative analysis, which was especially true for the samples with more than 2-fold difference of relative expression levels.

To evaluate the transcript level of miR156 precursors, gene-specific primers downstream of the hairpin structure of miR156 were designed for RT-PCR with the purpose of minimizing the inhibitive effect of secondary structure on PCR reaction. Because of limited length of the regions for primer designation, only two premature transcripts (OsmiR156d and OsmiR156h) could be amplified predominantly in young shoot, etiolated shoot, and seedling leaves (Fig. 4A). PAGE RNA gel-blot analysis also showed that the mature OsmiR156 had strong expression in young shoot, etiolated shoot, and seedling leaves, weak expression in root, stamen, and pistil, and undetectable expression level in stem and young panicles (Fig. 4C; Supplemental Fig. 6). The PAGE RNA gel-blot analysis suggested that the expression profile of mature OsmiR156 was essentially the same as of precursors. It was intriguing to notice that the mature OsmiR156 had very low level, if any, in young panicles whereas most of the miR156-targeted OsSPL genes had much stronger expression in young panicles than other tissues. Such complementary expression patterns between OsmiR156 and target genes suggested that the target OsSPL genes might be tempospatially regulated by OsmiR156.

Overexpression of OsmiR156 in Rice Resulted in Abnormal Growth and Development

The OsmiR156 family has at least five expressed members with the same matured miRNA sequence, suggesting difficulties in obtaining the loss-of-function mutant of individual OsmiRNA156 genes. Toward identification of the functions of OmiR156 and its putative target genes, two transcriptionally active (supported by cDNA or expressed sequence tag sequences) OsmiR156 precursors (OsmiR156b and OsmiR156h) were transformed into rice under the control of maize ubiquitin promoter (Fig. 5A ). PAGE RNA gel-blot analysis showed that the matured transcript of OsmiR156 in the leaves of the transgenic plants was much higher than that in the wild type, suggesting that OsmiR156b (Mb) and OsmiR156h (Mh) were overexpressed (Fig. 5B). All the OsmiR156-overexpressing plants (T₁ generation) showed dramatic morphological changes, including significantly (P < 0.001) increased number of tillers (10–15 times more than the wild type) and dwarfism (Fig. 5C), late flowering (7–10 d delay of flowering), significantly (P < 0.01) reduced number of spikelets and grains per panicle (Fig. 5D), and secondary branches of panicle (Fig. 5D). Even though the panicle size (length and number of spikelets) of OsmiR156-overexpressing plants was severely affected, the fertility was not significantly different from the wild type. All the phenotypes of OsmiR156-overexpressing plants were stably inherited in T₂ generation (Supplemental Fig. 7). Considering the fact that most of the OsmiR156-targeted OsSPLs, but not the OsmiR156, were predominantly expressed in panicles, such severe morphological change related to panicles in all the OsmiR156-overexpressing plants suggested that some, if not all, of the OsmiR156-targeted genes might be involved in the panicle development in rice.

View larger version (51K): [in this window] [in a new window]   Figure 5. Overexpression of OsmiR156 in rice. A, Schematic diagram of OsmiR156 overexpression construct. UBI, Maize ubiquitin gene promoter; Hpt, hygromycin resistance gene. B, PAGE RNA gel-blot analysis of miR156 in the leaves of transgenic plants at tillering stage. The exposure time of blot was optimal for the signal of overexpressed transgene but not long enough to show the signal in wild type (WT). Mb, OsmiR156b overexpression; Mh, OsmiR156h overexpression. C, Reduced plant height and increased number of tillers in Mb and Mh transgenic plants. The scale bar is 10 cm. D, Number of spikelets and grains per panicle in the transgenic plants and number of secondary branches per panicle in the transgenic plants.

We further checked the transcript levels of nine putative target genes of OsmiR156 in the panicles and leaves of OsmiR156-overexpression transgenic plants by semiquantitative RT-PCR (Fig. 6A ). The result showed that, compared to the wild type, three genes (OsSPL2, OsSPL12, and OsSPL13) had decreased mRNA levels in the flag leaves of transgenic plants, two genes (OsSPL16 and OSPL18) had obviously decreased mRNA levels in the panicles, one gene (OsSPL14) had decreased mRNA levels in both flag leaves and panicles, and the other three genes (OsSPL3, OsSPL7, and OsSPL11) had no change of expression level in either tissue. Real-time PCR analysis of four genes (OsSPL3, OsSPL7, OsSPL11, and OsSPL12) suggested that, except OsSPL3, the results of other three genes agreed well with the result by semiquantitative RT-PCR (Fig. 6B). OsSPL3 had slightly decreased transcript level in the flag leaves of transgenic plants and such a minor change may not have been detected by semiquantitative RT-PCR analysis.

View larger version (106K): [in this window] [in a new window]   Figure 6. Transcript levels of OsmiR156 and target genes in the transgenic plants. A, Semiquantitative RT-PCR of nine putative target genes of OsmiR156 in flag leaves and young panicles of the OsmiR156b (Mb)- and OsmiR156h (Mh)-overexpressing plants. The primers with amplicons spanning the complementary sites were used. The rice Actin1 gene was amplified as the internal control. B, Real-time PCR analysis of four putative target genes of OsmiR156. The relative expression levels were calibrated based on the leaf sample of wild type. C, Transcript levels of two OsmiR156-targeted genes in the young leaves of transgenic plants overexpressing the target genes. DNA fragments downstream of the targeting sites were used as probes. The target gene-overexpressed plants are indicated by diamonds. The uncleaved and cleaved transcripts (bands) are indicated by arrows.

	DISCUSSION

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 

Evolution of SPL and OsmiR156 Gene Families

SPL genes encode proteins that contain SBP domains. So far, no SPL homologous sequence has been found in animals, humans, or bacteria in the public databases, which may suggest that SPL genes appeared after the divergence of plants and animals and function specifically in plants. In this study, all the currently available SPL sequences (including all SPLs from rice and Arabidopsis) in the database were collected and phylogenetically classified into six subgroups (Fig. 2). Our results suggested that SPLs might be derived from several common ancestors before the monocot and dicot plants diverged and that the evolution of SPLs from same ancestor may be independent in monocots (such as rice and maize) and dicots (Arabidopsis and A. majus). Different subgroups diverged quite differently. For example, S5 had relatively lower divergence than other subgroups, which is also supported by the fact that OsSPLs in S5 showed more conserved exon-intron structure (Supplemental Fig. 4) and motif composition (Fig. 3) than did the SPLs in other subgroups. Most OsSPLs in the same phylogenetic subgroup have similar exon-intron structures and motif compositions, which suggests that the evolution of SBP domain may be closely related to the diversification of gene structure. Segmental duplication of rice chromosomes has been reported (Vandepoele et al., 2003). At least four pairs of OsSPL genes (OsSPL1 and 6, OsSPL3 and 12, OsSPL14 and 17, and OsSPL16 and 18) were located within such segmental duplication regions, which implies that duplication of chromosomal segments has also contributed to the expansion of OsSPL gene family.

Most target genes of miRNAs identified in plants are transcription factors. In this study, the miR156 family in the rice genome was analyzed and the putative target genes of OsmiR156 were 11 OsSPL genes that belong to the plant-specific SBP transcription factor family. Although more and more miRNAs have been cloned, the origin of the miRNA remains as a puzzle. A mechanism of miRNA origination and evolution has been proposed by analysis of two Arabidopsis miR161 and miR163 families (Allen et al., 2004). In this model, miR161 and miR163 evolved relatively recently by duplication events and then adapted to the miRNA apparatus. The precursor sequences of miR161 and miR163 have similarity with their target genes beyond the complementary region of mature miRNA. For the OsmiR156 in this study, however, no similarity was identified out of the mature miRNA region compared with the target genes. Nevertheless, most of the OsmiR156 and targeted OsSPL genes were located in clusters on rice chromosomes, which may suggest that OsmiR156 is evolutionarily related to the targeted OsSPL genes. In addition, there is a distinct phylogenetic boundary between miR156-targeted SPLs (S1–S3 and S6 subgroups) and non-miR156-targeted SPLs (S4 and S5 subgroups), suggesting that miR156 has been coexisted with a few ancestors of SPLs with or without miR156 target site. The evidence of miR156-mediated cleavage of SPL genes was reported in moss (Arazi et al., 2005). Interestingly, SPL genes were found in single-cell plant Chlamydomonas reinhardtii (Kropat et al., 2005). We also searched the draft genome sequence of C. reinhardtii (release 2.0, http://genome.jgi-psf.org/chlre2/chlre2.home.html) and no miR156 sequence was identified, which may suggest that miR156 evolved after the appearance of multicell plants.

Gene Structure Diversity of the OsSPL Family

By analyzing the entire complement of transcription factors in model organisms, researchers have proposed that novel transcription factors may have been generated by the shuffling of DNA-binding domains (Morgenstern and Atchley, 1999; Riechmann et al., 2000). Shuffling of the zinc finger may also be an efficient approach to generate novel transcription factors with DNA-binding activity (Bae et al., 2003). In plants, zinc finger-contained transcription factors have been diversified in terms of the different constitutions of zinc finger motifs such as the GATA, Dof (C-C-C-C type), and WRKY families (C-C-H-H type; Takatsuji, 1998). The SBP domain consists of two zinc fingers (Zn1 and Zn2) linked by a short JP and flanked by an NLS peptide. Such a combination of motifs features the SPL transcription factor family and may provide flexibility in generating diverse SBP domains of SPLs with different functions.

Variable protein sequences beyond the SBP domain have largely contributed to the diversification of the SPL gene family. Besides the SBP domain, seven other putative motifs, including the site with a sequence complementary to miR156 were detected in OsSPLs, although most of these motifs await experimental data for their functions. Different OsSPLs have distinct constitutions and organizations of these motifs. Some motifs have duplicated in several SPL genes, which further contributes to the diversity of the SPL gene family. For example, OsSPL9 contains another Zn2 and NLS motif at the C terminus in the reverse direction and motif 4 (unknown function) was duplicated in OsSPL1 in the forward direction. The SPL gene family can be phylogenetically classified into six subgroups based on the protein sequences of SBP domains, suggesting that the SBP domain may also contribute to the diversification of plant SPL genes.

The SBP domains of various SPL genes have conserved exon-intron structures, but intron length varies significantly. Intron length is reported to be negatively correlated with the divergence and recombination rate in Drosophila melanogaster (Comeron and Kreitman, 2000; Haddrill et al., 2005), and intron-specific selective constraints have been maintained after the gene duplication that precedes the divergence of gene functions (Parsch, 2003). The high level of divergence in the intron sequences between OsSPL genes may also indicate that some of these introns have been involved in the evolution and diversification of SPL proteins.

Diverse Tempospatial Expression Patterns of OsSPL Genes

Very low sequence similarity and different composition of various motifs suggest that OsSPL genes might have different expression patterns. To support this assumption, we investigated the transcript levels of all OsSPL genes in 13 different tissues of rice. Our data clearly suggest that OsSPL genes exhibit distinct expression patterns in terms of specificity and expression level (Fig. 4). Even though the OsSPL genes showed sequence diversity and different expression profiles, generally there is no obvious association between gene structures and expression patterns.

Transcript levels of all OsSPL genes in the 13 different rice tissues or organs revealed that most of the putative OsmiR156-targeted genes expressed predominantly in the young panicles. More than half of the OsSPL genes expressed predominantly in the young panicles of rice, suggesting that some of these OsSPL genes might be involved in the development of panicles in rice. Homologous SBP proteins in A. majus (AmSBP1 and AmSBP2) and Arabidopsis (AtSPL3) have been reported with a binding ability to the cis-element in the promoter of floral organ identity genes SQUA and AP1, respectively (Klein et al., 1996; Cardon et al., 1997). However, phylogenetic analysis suggested that, except for OsSPL13, all the OsSPL genes with specific or predominant expression in the young panicles were not in the subgroup (S1) that contained AmSBP1, AmSBP2, and AtSPL3 (Fig. 2).

Function and Interaction of OsmiR156 and OsSPLs

To further prove that OsSPL genes are regulated by OsmiR156, precursors of OsmiR156b and OsmiR156h (two representative OsmiR156 genes supported by expressed sequences) were independently overexpressed in rice. Transgenic plants of both OsmiR156 genes exhibited similar phenotypic changes: a large number of tillers, dwarfism, small panicles, and delayed flowering. These results may suggest that the expression of endogenous OsmiR156 genes should be under strict control to ensure the normal growth and development of rice. A similar phenotype was reported in the miR156-overexpressing Arabidopsis plants that had a moderate delay of flowering time under long days, a large number of leaves, decrease of apical dominant, and flowers from side shoots (Schwab et al., 2005).

To prove that such morphological changes resulted from suppression or loss of function of OsmiR156-targeted genes, we searched our Rice Mutant Database (Zhang et al., 2006) to find mutants of OsSPL genes. We have thus far found two mutants for OsSPL12 and OsSPL14, respectively. The osspl12 mutant (ID no. 03Z11AI07) exhibited delayed flowering time and the osspl14 mutant (ID no. 03Z11BJ86) showed dwarfism (Z. Chen, Q. Zhang, and C. Wu, unpublished data). Both OsSPL12 and OsSPL14 are putative targets of OsmiR156 but differ in the mutant phenotypes. However, the phenotypes of both mutants are included in the phenotypes of OsmiR156-overexpressed transgenic plants, suggesting that OsSPL12 and OsSPL14 might be the targets of OsmiR156 and have different functions.

Most miRNAs regulate more than one target gene, and the target genes are often from one gene family. OsmiR156 control 11 target genes, and it might be one of the largest number of genes by one miRNA. The study of miR164 (interacting with five genes of NAC family) in Arabidopsis shows that different members of miR164 function slightly differently (Baker et al., 2005). In this study, overexpression of two OsmiR156 members resulted in the same phenotype, but some of the target genes (such as OsSPL13 in leaves and OsSPL18 in panicles) showed differential cleavage between the two OsmiR156 overexpressors (Fig. 6A), suggesting differential interactions of a specific target gene with different OsmiR156 members. Different changes of different target genes were also reported in the miR172-overexpressing Arabidopsis plant in which the transcript level of the miR172 target gene TOE3 substantially increased; the transcript level of another target gene TOE2 substantially decreased (Schwab et al., 2005).

To date, most of the identified plant miRNAs such as miR164 (Guo et al., 2005), miR-JAW (Palatnik et al., 2003), and miR160 (Mallory et al., 2005) were reported for their regulation of target genes at transcription level by affecting the stability of target mRNA and thus degrading the target mRNA, whereas some plant miRNAs such as miR172 (Chen, 2004) were proposed to regulate the target gene by translational repression. Recently, the effects on steady-state levels of target transcripts were also suggested to be obscured by a strong feedback regulation (Schwab et al., 2005). In this study, six target genes (OsSPL2, OsSPL12, OsSPL13, OsSPL14, OsSPL16, and OsSPL18) showed decreased transcript level in the leaves or the panicles of the OsmiR156-overexpressing rice plants compared to the wild-type plant. However, the other three target genes (OsSPL3, OsSPL7, and OsSPL11) showed no change of transcript level in either tissue. Besides the possible mechanisms suggested previously, it cannot be excluded that these three genes are regulated by cleavage in other tissues not included in this study.

	MATERIALS AND METHODS

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 

Database Mining for OsSPL Genes in the Rice Genome

We established a local rice sequence database by downloading japonica genomic sequence and annotation data from TIGR (http://rice.tigr.org; Yuan et al., 2005), indica genomic sequence generated by whole genome shotgun sequencing (Yu et al., 2002) from Beijing Genomic Institute, and japonica full-length cDNA from KOME (Kikuchi et al., 2003). We performed a BLAST search (Altschul et al., 1997) first in the local rice (Oryza sativa) sequence database with the known SPL genes as queries. pHMM was used to identify new SPL genes in the rice genome. The HMM profile of the SBP domain (accession no. PF03110.17) was downloaded from the Pfam (http://www.sanger.ac.uk/Software/Pfam/, release 17.0; Bateman et al., 2004), and the HMMER package version 2.1 (Eddy, 1998) was used to search the local rice protein database. All hits with expected values less than 1.0 were collected. To exclude the redundant sequence of OsSPL genes, we aligned all sequences with CLUASTALX (Thompson et al., 1997). We used a BLASTN search to determine the chromosome locations of OsSPL genes by in silico mapping their sequences to the rice BAC/PAC physical map. All nonredundant protein sequences of putative OsSPL genes were manually checked for the SBP domain. Arabidopsis (Arabidopsis thaliana) SPL genes were downloaded from Munich Information Center for Protein Sequences Arabidopsis database (http://mips.gsf.de/proj/thal/db/) and SPL genes from other species were downloaded from Swiss-Prot (http://us.expasy.org/sprot/).

Bioinformatics Analysis of SPL and miR156 Gene Families

Multiple alignments of SPL protein sequences were performed with CLUSTALX (Thompson et al., 1997) and refined manually. Bayes phylogenetic trees were reconstructed with the program MrBayes version 3.0 (Ronquist and Huelsenbeck, 2003) under the JTT-f model of amino acid substitution and 4- category model. A total of 400,001 generations were performed with four Markov chains with default heating values and tree sampling every 100 generations. The Markov chain converged after 9,000 generations and the first 100 sampled trees were discarded. The major consensus tree was deduced from the remaining 901 sampled trees. The tree was edited with TreeView 1.5 (Page, 1996). We used the BLAST program to determine the intron-exon structures of OsSPL genes by mapping cDNAs to genomic sequence. The intron-exon structures of the OsSPL gene were illustrated proportionally to the lengths of introns and exons. The HMM profile of the SBP domain was reconstructed with the HMMbuild program in the HMMER package version 2.1 (Eddy, 1998), and alignment was viewed with sequence LOGO (Crooks et al., 2004). MEME program (Bailey and Elkan, 1994) was used to predict the potential motifs. All motifs discovered by MEME with expected values lower than 2E-30 were searched in the Inter-Pro database with Inter-ProScan (Mulder et al., 2005).

The stem-loop sequences of the plant miR156 family were downloaded from the miRbase (Griffiths-Jones, 2004) to reconstruct the phylogenetic tree. The sequences of stem regions (mature miRNA sequences and the complementary sequences) were aligned first by CLUSTALX. The sequences beyond the stem regions were subsequently aligned with manual modification. The secondary structures of OsmiR156 RNAs were generated with the RNAalifold program (Hofacker et al., 2002), and sequences beyond the consensus secondary structure were masked for phylogenetic analysis. The Bayes tree was constructed with a mixed nucleotide substitution model of REV and RNA7D described in the PHASE program, which was specifically designed for RNA phylogenetic inference according to the secondary structures (Hudelot et al., 2003). The secondary structures of OsmiR156 RNAs and consensus secondary structure of AtmiR156 and OsmiR156 were generated with the Vienna RNA package version 1.4 (Hofacker, 2003). The target genes of OsmiR156 were predicted by using the mature sequences to search the rice genomic sequences. According to the characterized miRNAs in plants (Rhoades et al., 2002; Chen, 2004; Mallory et al., 2004; Guo et al., 2005; Mallory et al., 2005), rice genes (except OsmiR156 precursors) that contained sequences with fewer than three mismatches to mature miRNA156 sequence are considered to be putative targets of OsmiR156.

Isolation of OsSPL and OsmiR156 Genes and Rice Transformation

For each OsSPL gene, a pair of primers (Supplemental Table I) was designed to amplify the predicted full-length cDNA with cDNA templates prepared from different tissues of Minghui 63 (rice L. subsp. indica), a parental line for elite hybrid rice in China. ExTaq DNA polymerase (Takara) was used to amplify the OsSPL genes with the following cycling profile: 94°C for 4 min; 25 to 30 cycles of 94°C for 40 s, 55°C for 40 s, and 72°C for 2 or 3 min; and extension at 72°C for 10 min. The amplified products were cloned into pGEM-T easy vector (Promega) and sequenced from both ends by using BigDye Terminator Sequencing Ready kit (version 2.0 or 3.0) in an ABI PRISM 377 or 3730 sequencer.

Gene-specific primers (Supplemental Table I) were used to amplify the OsmiR156b and OsmiR156h precursors from the rice genome. The genomic fragments were cloned to pGEM-T Easy vector (Promega) and sequenced. Then the fragments of two OsmiR156 precursors were cut by KpnI and BamHI and ligated into the transformation vector pCAMBIA1301U under the control of a maize (Zea mays) ubiquitin gene promoter. The full-length cDNA of OsSPL12 and OsSPL14 were cloned into pCAMBIA1301 (provided by CAMBIA) under the control of Cauliflower mosaic virus 35S promoter.

The Agrobacterium-mediated transformation method was used to introduce constructs into Agrobacterium tumefaciens strain EHA105 by electroporation and transformed into rice Zhonghua11 (rice L. subsp. japonica; Hiei et al., 1994).

RT-PCR Analysis

The TRIZol reagent (Invitrogen) was used according the manufacturer's instructions to extract total RNAs of various tissues or organs from a life cycle of rice Minghui63 (rice L. subsp. indica). Before RT, total RNA was treated with amplification-grade DNase I (Invitrogen) for 15 min to degrade possibly contaminated residual genomic DNA. SuperScriptII reverse transcriptase (Invitrogen) was used according to the manufacturer's instructions to synthesize first-strand cDNA from the DNase I-treated total RNA. About 1/20 of the first-strand cDNA generated from 1 µg total RNA was used as template for PCR in a reaction volume of 50 µL with the rTaq DNA polymerase (Takara). PCR was performed in an ABI 9700 Thermocycler (Applied System) with the following cycling profile: 94°C for 3 min; 25 to 40 cycles at 94°C for 40 s, 55°C or 60°C for 40 s, and 72°C for 1 min. Fifteen microliters of the PCR product was separated in a 1.2% agarose gel and stained with ethidium bromide for visualization. We used a pair of primers specific to rice Actin1 gene (accession no. AK060893) for RT-PCR as internal control to compare the band intensity between samples. For each OsSPL gene, a pair of primers with a 400 to 600 bp amplicon was used for RT-PCR (Supplemental Table I) with 25, 30, 35, and 40 cycles, depending on the expression levels of different genes. All RT-PCRs were repeated three times with independently reverse-transcribed templates.

Real-Time Quantitative RT-PCR

Relative quantification of gene expression by real-time PCR was performed on an ABI PRISM 7500 instrument (Applied Biosystems). The primers for real-time PCR were designed by Primer Express Version 2.0 (Applied Biosystems; Supplemental Table III). Rice Actin1 gene was used as endogenous control. Real-time PCR was performed in an optical 96-well plate, including 12.5 µL 2x SYBR Green Master mix reagent (Applied Biosystems), 1 µL cDNA samples, and 0.2 µM of each gene-specific primers, in final volume of 25 µL, using the thermal cycles as follows: 50°C for 2 min, 95°C for 10 min; 40 or 45 cycles of 95°C for 30 s; 60°C for 30 s; and 72°C for 1 min. Disassociation curve analysis was performed as follows: 95°C for 15 s; 60°C for 20 s; 95°C for 15 min. The relative expression levels were determined as described previously (Liang et al., 2006).

RNA Gel-Blot Analysis of miRNA and SPL Genes

Mini-PROTEAN III system (Bio-Rad) was used to separate low-mass RNA by electrophoresis in 20% urea-denatured polyacrylamide gel and blot it onto nylon membranes. DNA Oligo (5'-GTGCTCACTCTCTTCTGTCA-3') was synthesized as a probe to detect the OsmiR156 level. The probes were labeled and hybridized essentially according to Bartel's description (Bartel, 2004). The blot was washed four times (5, 10, 40, and 40 min, respectively) at 50°C with washing buffer (3x SSC, 25 mM NaH₂PO₄, pH 7.5, 5% SDS, and 10x Denhardt's solution). The FUJIFILM system (FJL5100, Fuji) was used to quantify signal intensity. All the probes were stripped and then rehybridized with DNA probe (5'-TGTATCGTTCCAATTTTATCGGATGT-3') complementary to U6 RNA.

Sequence data of microRNA156 in rice from this article can be found in the miRBase database under accession numbers MI0000653 to MI0000662, MI0001090, and MI0001091.

	ACKNOWLEDGMENTS

 
We thank Yinglong Cao and Meng Cai for kindly constructing and providing the p1301U vector.

Received June 1, 2006; accepted July 13, 2006.

	FOOTNOTES

 
¹ This work was supported by grants from the National Program on High Technology Development, the National Program on the Development of Basic Research, and the National Natural Science Foundation, China.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: Lizhong Xiong (lizhongx{at}mail.hzau.edu.cn).

^[W] The online version of this article contains Web-only data.

www.plantphysiol.org/cgi/doi/10.1104/pp.106.084475

^* Corresponding author; e-mail lizhongx{at}mail.hzau.edu.cn; fax 86–27–87287092.

	LITERATURE CITED

TOP ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

 
Allen E, Xie Z, Gustafson AM, Sung GH, Spatafora JW, Carrington JC (2004) Evolution of microRNA genes by inverted duplication of target gene sequences in Arabidopsis thaliana. Nat Genet 36: 1282–1290[CrossRef][ISI][Medline]

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402[Abstract/Free Full Text]

Arazi T, Talmor-Neiman M, Stav R, Riese M, Huijser P, Baulcombe DC (2005) Cloning and characterization of micro-RNAs from moss. Plant J 43: 837–848[CrossRef][ISI][Medline]

Bae KH, Kwon YD, Shin HC, Hwang MS, Ryu EH, Park KS, Yang HY, Lee DK, Lee Y, Park J, et al (2003) Human zinc fingers as building blocks in the construction of artificial transcription factors. Nat Biotechnol 21: 275–280[CrossRef][ISI][Medline]

Bailey TL, Elkan C (1994) Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc Int Conf Intell Syst Mol Biol 2: 28–36[Medline]

Baker CC, Sieber P, Wellmer F, Meyerowitz EM (2005) The early extra petals1 mutant uncovers a role for microRNA miR164c in regulating petal number in Arabidopsis. Curr Biol 15: 303–315[CrossRef][ISI][Medline]

Bartel DP (2004) MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116: 281–297[CrossRef][ISI][Medline]

Batchelor AH, Piper DE, de la Brousse FC, McKnight SL, Wolberger C (1998) The structure of GABPalpha/beta: an ETS domain-ankyrin repeat heterodimer bound to DNA. Science 279: 1037–1041[Abstract/Free Full Text]

Bateman A, Coin L, Durbin R, Finn RD, Hollich V, Griffiths-Jones S, Khanna A, Marshall M, Moxon S, Sonnhammer EL, et al (2004) The Pfam protein families database. Nucleic Acids Res 32: D138–D141[Abstract/Free Full Text]

Bonnet E, Wuyts J, Rouze P, Van de Peer Y (2004) Detection of 91 potential conserved plant microRNAs in Arabidopsis thaliana and Oryza sativa identifies important target genes. Proc Natl Acad Sci USA 101: 11511–11516[Abstract/Free Full Text]

Cardon G, Hohmann S, Klein J, Nettesheim K, Saedler H, Huijser P (1999) Molecular characterisation of the Arabidopsis SBP-box genes. Gene 237: 91–104[CrossRef][ISI][Medline]

Cardon GH, Hohmann S, Nettesheim K, Saedler H, Huijser P (1997) Functional analysis of the Arabidopsis thaliana SBP-box gene SPL3: a novel gene involved in the floral transition. Plant J 12: 367–377[CrossRef][ISI][Medline]

Chen X (2004) A microRNA as a translational repressor of APETALA2 in Arabidopsis flower development. Science 303: 2022–2025[Abstract/Free Full Text]

Comeron JM, Kreitman M (2000) The correlation between intron length and recombination in Drosophila: dynamic equilibrium between mutational and selective forces. Genetics 156: 1175–1190[Abstract/Free Full Text]

Crooks GE, Hon G, Chandonia JM, Brenner SE (2004) WebLogo: a sequence logo generator. Genome Res 14: 1188–1190[Abstract/Free Full Text]

Dingwall C, Laskey RA (1991) Nuclear targeting sequences—a consensus? Trends Biochem Sci 16: 478–481[CrossRef][ISI][Medline]

Eddy SR (1998) Profile hidden Markov models. Bioinformatics 14: 755–763[Abstract/Free Full Text]

Gorina S, Pavletich NP (1996) Structure of the p53 tumor suppressor bound to the ankyrin and SH3 domains of 53BP2. Science 274: 1001–1005[Abstract/Free Full Text]

Griffiths-Jones S (2004) The microRNA registry. Nucleic Acids Res 32: D109–D111[Abstract/Free Full Text]

Guo H-S, Xie Q, Fei J-F, Chua N-H (2005) MicroRNA directs mRNA cleavage of the transcription factor NAC1 to downregulate auxin signals for Arabidopsis lateral root development. Plant Cell 17: 1376–1386[Abstract/Free Full Text]

Haddrill PR, Charlesworth B, Halligan DL, Andolfatto P (2005) Patterns of intron sequence evolution in Drosophila are dependent upon length and GC content. Genome Biol 6: R67

Hiei Y, Ohta S, Komari T, Kumashiro T (1994) Efficient transformation of rice (Oryza sativa L.) mediated by Agrobacterium and sequence analysis of the boundaries of the T-DNA. Plant J 6: 271–282[CrossRef][ISI][Medline]

Hofacker IL (2003) Vienna RNA secondary structure server. Nucleic Acids Res 31: 3429–3431[Abstract/Free Full Text]

Hofacker IL, Fekete M, Stadler PF (2002) Secondary structure prediction for aligned RNA sequences. J Mol Biol 319: 1059–1066[CrossRef][ISI][Medline]

Hudelot C, Gowri-Shankar V, Jow H, Rattray M, Higgs PG (2003) RNA-based phylogenetic methods: application to mammalian mitochondrial RNA sequences. Mol Phylogenet Evol 28: 241–252[CrossRef][ISI][Medline]

Huijser P, Klein J, Lonnig WE, Meijer H, Saedler H, Sommer H (1992) Bracteomania, an inflorescence anomaly, is caused by the loss of function of the MADS-box gene squamosa in Antirrhinum majus. EMBO J 11: 1239–1249[ISI][Medline]

International Rice Genome Sequencing Project (2005) The map-based sequence of the rice genome. Nature 436: 793–800[CrossRef][Medline]

Jack T (2004) Molecular and genetic mechanisms of floral control. Plant Cell 16: S1–S17[Free Full Text]

Jones-Rhoades MW, Bartel DP (2004) Computational identification of plant microRNAs and their targets, including a stress-induced miRNA. Mol Cell 14: 787–799[CrossRef][ISI][Medline]

Kidner CA, Martienssen RA (2004) Spatially restricted microRNA directs leaf polarity through ARGONAUTE1. Nature 428: 81–84[CrossRef][Medline]

Kikuchi S, Satoh K, Nagata T, Kawagashira N, Doi K, Kishimoto N, Yazaki J, Ishikawa M, Yamada H, Ooka H, et al (2003) Collection, mapping, and annotation of over 28,000 cDNA clones from japonica rice. Science 301: 376