Small RNA profiling in Chlamydomonas: insights into chloroplast RNA metabolism

Abstract

In Chlamydomonas reinhardtii, regulation of chloroplast gene expression is mainly post-transcriptional. It requires nucleus-encoded trans-acting protein factors for maturation/stabilization (M factors) or translation (T factors) of specific target mRNAs. We used long- and small-RNA sequencing to generate a detailed map of the transcriptome. Clusters of sRNAs marked the 5' end of all mature mRNAs. Their absence in M-factor mutants reflects the protection of transcript 5' end by the cognate factor. Enzymatic removal of 5'-triphosphates allowed identifying those cosRNA that mark a transcription start site. We detected another class of sRNAs derived from low abundance transcripts, antisense to mRNAs. The formation of antisense sRNAs required the presence of the complementary mRNA and was stimulated when translation was inhibited by chloramphenicol or lincomycin. We propose that they derive from degradation of double-stranded RNAs generated by pairing of antisense and sense transcripts, a process normally hindered by the traveling of the ribosomes. In addition, chloramphenicol treatment, by freezing ribosomes on the mRNA, caused the accumulation of 32-34 nt ribosome-protected fragments. Using this 'in vivo ribosome footprinting', we identified the function and molecular target of two candidate trans-acting factors.

Figure 1.

Transcriptional profile over petB and the psbF-psbL-petG polycistronic unit. Coverage (log scale) of (A) pooled bi-directional and directional WTSS; (B) pooled sRNA-Seq; (C) mock- (blue) versus RPP-treated (red) WT sRNA-Seq libraries. Red vertical lines indicate the position of the mature 5′ ends. Vertical arrows point to cosRNAs marking a TSS or PTP. In A and C, coverage is normalized as RPM.

Figure 2.

5′-cosRNA are footprints of M factors. Distribution of sRNAs in WT (black) and mutant strains (red) over their target genes. Vertical arrows indicate the position of the mature 5′end in the WT. Coverage is expressed in RPM and averaged over two biological replicates.

Figure 3.

Size distribution of sRNAs in control, CAP- and Linc-treated samples, over 5′UTRs (A), CDS (B) and 3′UTRs (C). For each experiment, the two replicates are shown. psbA was excluded from this representation.

Figure 4.

Characteristics of RPFs (32–34 nt) in control and CAP-treated samples. (A) Position of the 5′ end of 32–34-nt reads respective to the reading frame (all CDS). (B) Profiles of 5′-end positions of 32–34 nt sRNAs sequences (average of two replicates) in CDS (horizontals arrow) and untranslated regions (marked by vertical arrows) of atpA, petA and rbcL. (C) Accumulation of chloroplast transcripts upon CAP treatments, with the nuclear Cβlp2 gene as a loading control.

Figure 5.

Gene expression in mixotrophic and phototrophic conditions (log₂ transformed RPKM values). RPKM values were computed on CDS, on the three exons of psaA and on the WendyB and tscA genes. The value for psbA is the average of the RPKM computed independently for the five exons. Differentially expressed genes are in grey.

Figure 6.

Cp transcripts display different stability between mixotrophic and phototrophic growth. (A) MA plot reporting the log₂FC (Rif-treated over control) against log₂RPM of the RNA levels, distinguishing genes involved in photosynthesis (circle) or other functions (cross) in mixotrophic (black) and phototrophic (blue) condition. (B) Difference between log₂FC in the two conditions. (C) Heatmap of the log₂FC. Genes with significant changes at FDR ≤ 0.05 are marked with an asterisk.

Figure 7.

Size distribution of small RNAs in Rif-treated samples. Abundance over 5′ UTRs (A), CDS (B) and 3′ UTRs (C). For each experiment, the two replicates are shown. psbA was excluded from this representation.

Figure 8.

Characteristics of RPFs (32–34 nt) after Rif-treatment in phototrophically grown cells. (A) log₂FC (Rif-treated over control) of RPF (red) and non-RPF sRNAs (black) for each chloroplast CDS, following order in the genome from petA to WendyA. (B) Profiles of the 5′-end positions of 32–34-nt sequences on psbC after CAP or Rif treatment compared to the controls. (C) Position of the 5′ end of 32–34-nt reads respective to the reading frame (all CDS).

Figure 9.

Antisense transcription in the Cp genome. (A) Comparison of sense (red) and antisense (blue) coverage over each chloroplast region as averaged RPKMs from four directional WTSS datasets and 4 sRNA-Seq datasets derived from the same RNA samples. (B) Coverage in RPM of antisense sRNAs in WT (black) and mutant strains (red). The orientation of the horizontal arrow indicate transcription direction of the mRNA. (C) Profile of s-sRNAs (red) and as-sRNAs (blue) at the atpA, atpB, atpI and petA loci. F1,F2 and F3 primers were used for reverse transcription, then combined with primers R1 for PCR. (D) Strand-specific RT-PCR in the presence (left panel) and absence of the reverse transcriptase (RT) (right panel). (E) Expression level of atpB mRNA (red) and antisense-atpB (in blue) in the mdb1 mutant relative to the WT by qPCR. The values are the average of two independent qPCR assays ± SD.

Figure 10.

The effects of transcription and translation inhibition on the production of as-sRNAs. (A) Coverage of sense (upper panels) and antisense (lower panels) sRNA over the petA gene in (top to bottom) WT, mca1, WT-pG and mca1-pG). (B) log₂FC of the averaged RPMs of drug-treated samples over the control per Cp region. (C) Coverage of antisense sRNAs along the atpA gene, following Rif, Linc or CAP treatment and in mutant tda1.

Figure 11.

Snapshot of the IGV browser showing the alignment of sense (top, red parentheses) and antisense (bottom, blue parentheses) sRNA from CAP-treated WT and mutant strains. (A) opr56 over psbC; (B) ppr1 over petB. For each panel, the upper track displays the coverage, the lower track the reads. Arrows and dashed lines represent the CDS and UTRs and orientation on the genome; vertical arrows point to the 5′ ends of the transcripts. (C) Accumulation of petB mRNA in WT and ppr1 (with psbF as a loading control) and psbC mRNA in WT, ppr1 and opr56 (with psbA as loading control).

Figure 12.

Model for Cp mRNA degradation and generation of small RNAs. (A) Transcription occurs on both strands of a gene locus and generates abundant mRNA (red) and rare antisense transcripts (blue). M factors stabilize the mRNA and T factors activate its translation. After translation, the mRNA can be delivered to degradation Pathway 1, starting with endoribonucleolytic cleavage followed by exoribonucleolytic degradation. Transcripts in excess, namely mRNAs not bound by an M factor or not activated for translation, are mostly directed toward Pathway 1, but they can also base-pair with antisense RNA: the generated dsRNA is substrate for a dsRNA endonuclease (Pathway 2). A block in translation (T factor mutant, Linc or CAP) will exacerbate Pathway 2. In addition, a block in translation with CAP induces the degradation of the mRNA engaged with ribosomes through endoribonucleolytic cleavage between the stalled ribosomes. (B) By-products of RNA degradation comprise mononucleotides (not shown), M factor footprints, ribosome-protected fragments and s-sRNA (derived from Pathways 1 and 3) as well as as-sRNAs derived from Pathway 2.