Characterization of 43 non-protein-coding mRNA genes in Arabidopsis, including the MIR162a-derived transcripts

Abstract

Messenger RNAs that do not contain a long open reading frame (ORF) or non-protein-coding RNAs (npcRNAs) are an emerging novel class of transcripts. Their functions may involve the RNA molecule itself and/or short ORF-encoded peptides. npcRNA genes are difficult to identify using standard gene prediction programs that rely on the presence of relatively long ORFs. Here, we used detailed bioinformatic analyses of expressed sequence tag/cDNA databases to detect a restricted set of npcRNAs in the Arabidopsis (Arabidopsis thaliana) genome and further characterized these transcripts using a combination of bioinformatic and molecular approaches. Compositional analyses revealed strong nucleotide strand asymmetries in the npcRNAs, as well as a biased GC content, suggesting the existence of functional constraints on these RNAs. Thirteen of these transcripts display tissue-specific expression patterns, and three are regulated in conditions affecting root architecture. The npcRNA 78 gene contains the miR162 sequence in an alternative intron and corresponds to the MIR162a locus. Although DICER-LIKE 1 (DCL1) mRNA is known to be regulated by miR162-guided cleavage, its level does not change in a mir162a mutant. Alternative splicing of npcRNA 78 leads to several transcript isoforms, which all accumulate in a dcl1 mutant. This suggests that npcRNA 78 is a genuine substrate of DCL1 and that splicing of this microRNA primary transcript and miR162 processing are competitive nuclear events. Our results provide new insights into Arabidopsis npcRNA biology and the potential roles of these genes.

Figure 1.

Bioinformatic analysis of compositional biases and nucleotide conservation of npcRNA genes. A, GC content in the regions surrounding the 5′- and 3′-npcRNA gene extremities calculated in adjacent windows starting from each gene extremity in both directions. In the abscissa, the distance (in nucleotides) of each 100-bp window to the indicated gene extremity is presented, zero values of the abscissa corresponding to 5′ (left) or to 3′ (right) gene extremities. In the ordinate, the mean values, for all npcRNAs, of the GC content are calculated at the corresponding abscissa in the corresponding windows; vertical bars indicate se. B, Strand asymmetries S_TA = (T − A)/(T + A), where A and T are the numbers of the corresponding nucleotides in the sequence window; abscissa is as in A. C, Conserved nucleotide regions in npcRNAs. Regions conserved in ESTs from other species are indicated for npcRNAs 41 (TAS3) and 375. For npcRNA 41 and npcRNA 375, homologies are present in 14 (e.g. Oryza sativa, Populus tremula, Glycine max, and Picea glauca) and 11 (e.g. Oryza sativa, Citrus sinensis, Antirrhinum majus, and Hordeum vulgare) species, respectively. npcRNA 52 and 72 genes are two paralogous candidate genes. Hatched boxes, Conserved nucleotide fragments; gray boxes, longest ORF. Numbers above boxes indicate the size of the matching regions in nucleotides (in Arabidopsis); the percentage of identity is indicated in parentheses.

Figure 2.

Expression profiling of npcRNAs. Total RNA from roots, rosettes, stems, cauline leaves, inflorescences, and cell suspensions was assayed by semiquantitative RT-PCR. Fourteen npcRNAs (including DVL20) display specific developmental expression patterns. Other npcRNAs are broadly expressed in all tissues analyzed. PLS (POLARIS) mRNA, which accumulates preferentially in roots, was used as a tissue-specific control. β-Tubulin 4 mRNA served as a constitutively expressed control. B, DNA contamination controls for several npcRNAs with highly specific expression patterns. ±, RT-PCR performed with or without reverse transcriptase.

Figure 3.

Regulation of npcRNAs under conditions modifying root architecture. Expression of npcRNAs in roots of plants grown in low (−P) or high (+P) phosphate conditions and in the presence of cytokinin (BA) was examined. Real-time RT-PCR was performed for npcRNAs 34, 43, and 60 and data were normalized with ACTIN2. Values for roots grown in high-phosphate conditions or not treated with 0.1 μm BA are arbitrarily fixed to 1. For each cDNA synthesis, quantifications were made in triplicate and two biological replicates (I and II) were analyzed. Values are means ± sds.

Figure 4.

Analysis of the npcRNA 78/MIR162a transcripts and molecular characterization of the mir162a mutant. A, RT-PCR detection of different npcRNA 78/MIR162a transcripts. RT-PCR was performed using npcRNA 78-specific primers located in exons 2 and 4 (78E2F/78E4R). B, Diagrammatic representation of the differentially spliced transcripts of npcRNA 78 as deduced from A and annotated structure of the miR162a predicted hairpin region. Exonic sequences are in uppercase bold letters, intronic sequences are in lowercase letters, and the miR162 sequence is in bold lowercase letters. Square brackets, 5′ and 3′ splice sites; BP, likely branch point; star, alternative codon. The arrow indicates the position of the SALK_107598 T-DNA insertion. C, Uncleaved DCL1 mRNA levels are similar in mir162a T-DNA insertion mutant and wild-type inflorescences. Real-time RT-PCR was performed using primers flanking the miR162-directed cleavage site in the DCL1 mRNA. Quantifications were normalized with ACTIN2. Values in wild-type inflorescences were arbitrarily fixed to 1. Quantifications were made in triplicate (error bars represent sds). Results for two biological replicates are presented. D, miR162 levels are comparable in wild-type and mir162a plants. Blots of RNA extracted from inflorescences of wild-type and mir162a plants were successively hybridized with miR162 and miR171 LNA-modified probes. Hybridization to a U6 probe served as a loading control.

Figure 5.

npcRNA 78/MIR162a transcripts are stabilized in the dcl1-9 mutant. A, RT-PCR was performed using npcRNA 78-specific primers located in exons 2 and 4 (78E2F/78E4R) on RNA from rosettes of RNAi-related mutants. B, Dilution series of cDNAs from dcl1-9 and wild-type inflorescences shows consistently higher amplification by RT-PCR in the mutant when using the above-mentioned primers. C, Accumulation of miR162a and b primary transcripts in the dcl1-9 mutant. Specific primer pairs used for npcRNA 78/MIR162a amplification were located either in exons 2 and 4 (78E2F/78E4R) or in exon 5 (78E5F/78E5R). Primers pairs used for amplification of pri-miR162b (pri-miR162b-F1/R1 and pri-miR162b-F1/R2) did not reveal any alternative splicing for this transcript. The miR172b (EAT) primary transcript is a known pri-miRNA. npcRNA 21 is not affected by the dcl1 mutation. ±, RT-PCR performed with or without reverse transcriptase. D, Accumulation of miR162a and b primary transcripts in the dcl1-9 mutant measured by real-time RT-PCR. Primers used for npcRNA 78/MIR162a amplification were located in exon 5 (78E5F/78E5R). Primers pri-miR162b-F1 and R1 were used for amplification of pri-miR162b. Data were normalized with ACTIN2. Values in wild-type inflorescences were arbitrarily fixed to 1. Three technical replicates were performed for wild-type and dcl1-9 inflorescences. Values are means ± sds. For all these RT-PCR experiments, analyses were performed on homozygous mutants and their wild-type siblings.