Most microRNAs in the single-cell alga Chlamydomonas reinhardtii are produced by Dicer-like 3-mediated cleavage of introns and untranslated regions of coding RNAs

Abstract

We describe here a forward genetic screen to investigate the biogenesis, mode of action, and biological function of miRNA-mediated RNA silencing in the model algal species,Chlamydomonas reinhardtii Among the mutants from this screen, there were three at Dicer-like 3 that failed to produce both miRNAs and siRNAs and others affecting diverse post-biogenesis stages of miRNA-mediated silencing. The DCL3-dependent siRNAs fell into several classes including transposon- and repeat-derived siRNAs as in higher plants. The DCL3-dependent miRNAs differ from those of higher plants, however, in that many of them are derived from mRNAs or from the introns of pre-mRNAs. Transcriptome analysis of the wild-type and dcl3 mutant strains revealed a further difference from higher plants in that the sRNAs are rarely negative switches of mRNA accumulation. The few transcripts that were more abundant in dcl3 mutant strains than in wild-type cells were not due to sRNA-targeted RNA degradation but to direct DCL3 cleavage of miRNA and siRNA precursor structures embedded in the untranslated (and translated) regions of the mRNAs. Our analysis reveals that the miRNA-mediated RNA silencing in C. reinhardtii differs from that of higher plants and informs about the evolution and function of this pathway in eukaryotes.

Figure 1.

Screening and isolation of mutants affected in miRNA-mediated RNA silencing. (A) Schematic representation of the artificial miRNA construct used to transform the wild-type strain. Transgenic lines carrying this cassette were further screened by random insertional mutagenesis: (P_RBSC2) RuBisCO small subunit (RBSC)2 promoter; (Paromomycin^R) Streptomyces rimosus AphVIII coding gene; (T₁) RBSC2 transcription terminator; (P_Nit1) nitrate reductase promoter; (PSY_amiRNA) modified version of cre-miR1157 that carries a miRNA against the phytoene synthase; (T₂) RLP12 transcription terminator. (B) Selective cell death of transgenic lines expressing the PSY amiRNA in the presence of nitrate, but not ammonium, as the sole nitrogen source. Transgenic lines carrying the empty amiRNA vector (EV) were used as control. (C) Growth in high light conditions of mutagenized (Spect^R) and nonmutagenized (-) reporter lines (PSY_amiRNA) in solid media containing either nitrate or ammonium as sole nitrogen source. Transgenic lines carrying the empty amiRNA vector (EV) and further transformed with the spectinomycin resistance cassette were used as control. (D) Detection by Northern blot of diverse small RNAs in total RNA samples from the indicated mutants and controls. These mutants were obtained by random insertional mutagenesis of either spectinomycin or hygromycin resistance cassettes. The mutants were grouped (I–IV) (Supplemental Table S1) based on the molecular phenotype. The two displayed mutants belonging to the group II correspond to the characterized mutant 47 (dcl3-2) and mutant 51 (dcl3-3).

Figure 2.

Mapping and complementation of group II mutant 51. (A) Location of the mutagenic hygromycin resistance cassette in mutant 51. (B) Phenotype of the indicated parental line and both complemented and noncomplemented lines (biological triplicates) in the presence of either nitrate or ammonium under high light conditions. (C) Detection by Northern blot of the indicated miRNAs in total RNA samples from the C. reinhardtii strains analyzed in B.

Figure 3.

Effect of dcl3 mutation on C. reinhardtii small RNA population. (A) Size-distribution histograms of sRNAs from the parental line A4-1 and its derivative dcl3-1 mutant expressed as the number of counted reads of a given size per million (CPM) of reads matching the C. reinhardtii genome. The percentage of 21-nt sRNAs with their 5′ nucleotide identities is also shown. (B) Size-distribution histograms of nonredundant sRNAs from the parental line A4-1 and its derivative dcl3-1 mutant expressed as CPM of reads matching the C. reinhardtii genome. Two additional replicates per sample, as well as three replicates from the E9-3 parental and dcl3-3 lines, showed the same result.

Figure 4.

The cre-miR1157 is an intron-derived miRNA. (A) Schematic representation of constructs carrying the cre-miR1157 intron inserted into the spectinomycin resistance gene coding sequence. The cre-miR1157 intron was modified to either lack the miRNA stem–loop or carry an artificial miRNA against Maa7 in spect/intron and spect/intron(mi) plasmids, respectively: (P) Hybrid RBSC2/HSP70A promoter; (Spectinomycin^R) recoded Escherichia coli-derived aadA coding gene; (T) RBSC2 transcription terminator; (SpeI) unique cleavage site for SpeI restriction enzyme; (Maa7 amiRNA) modified version of cre-miR1157 that carries a miRNA against Maa7. (B, top) Growth of the indicated transgenic lines in solid media carrying spectinomycin with/without 5-Fluorindole (5-Fl). (Bottom) Detection by Northern blot of the artificial miRNA against Maa7 in total RNA samples from the indicated lines (three independent lines per construct). (C) Schematic representation of constructs used to test the requirement of splicing for the expression of id-miRNA. The GT × AT point mutations in the exon/intron junction are indicated. These plasmids also carry the Paromomycin^R cassette (equivalent to the cassette showed in Fig. 1A) to allow the primary selection of transgenic lines in paromomycin. (D) Growth of lines transformed with the indicated plasmids in solid media containing either paromomycin (test for plasmid integration), spectinomycin (test for splicing events), or 5-Fl (test for amiRNA production).

Figure 5.

The effect of miRNA on mRNA accumulation. (A) Steady-state accumulation levels of previously reported miRNA targets (Molnár et al. 2007; Zhao et al. 2007) assessed as the number of normalized reads (y-axis) in RNA-seq data. Error bars for three independent samples are shown. The target genes with their corresponding miRNA are indicated. These miRNAs were predicted as either high confidence miRNAs (cre-miR1162, cre-miR1151a/b) or medium confidence miRNA (miR-C82) (see Table 1), with the exception of cre-miR909 that is a hairpin-derived siRNA also depleted in the dcl3 mutant background. (B) 5′RACE to test the specific cleavage of CPLD52 (Cre13.g608000) mediated by cre-miR1162. The asterisk indicates an unspecific PCR product. The PCR products were sequenced, and the right panel shows the 5′ terminus of these cleavage products aligned to the 5′ to 3′ mRNA sequence and the 3′ to 5′ miRNA. G:U base pairs are indicated by a circle. (C) 5′RACE to test the specific cleavage of OMT2 (Cre17.g713200) mediated by the miR-C82 (Zhao et al. 2007) with the 5′ terminus of these cleavage products aligned to the 5′ to 3′ mRNA sequence and the 3′ to 5′ miRNA as in B.

Figure 6.

The effect of DCL3 on mRNAs with miRNA hairpin-like structures in the 3′ UTR. Cre16.g694950 (serine/threonine kinase) (A) and Cre24.g755697 (aminoglycoside 3′-phosphotransferase) (B) have the respective cre-miR1169 and cre-miR1172 precursors in their 3′ UTR. A schematic representation of both genes is shown with their exons (gray boxes) and introns (black solid lines) at the bottom of each panel. Light gray (A4-1 parental line) and black (dcl3-1) hills represent sRNA and mRNA read counts. Both panels show the results for one replicate of A4-1 parental line and its dcl3-1 derivative knock out mutant. Overaccumulation (X-fold) of the indicated mRNA in dcl3-1 regarding the A4-1 parental line is indicated. Two additional samples from this parental and mutant line combination and a biological triplicate from E9-3 parental and dcl3-3 derivative lines showed the same trend in both miRNA and mRNA accumulation (Table 1; Supplemental Table S3).