Protein-mediated protection as the predominant mechanism for defining processed mRNA termini in land plant chloroplasts

Abstract

Most chloroplast mRNAs are processed from larger precursors. Several mechanisms have been proposed to mediate these processing events, including site-specific cleavage and the stalling of exonucleases by RNA structures. A protein barrier mechanism was proposed based on analysis of the pentatricopeptide repeat (PPR) protein PPR10: PPR10 binds two intercistronic regions and impedes 5'- and 3'-exonucleases, resulting in processed RNAs with PPR10 bound at the 5'- or 3'-end. In this study, we provide evidence that protein barriers are the predominant means for defining processed mRNA termini in chloroplasts. First, we map additional RNA termini whose arrangement suggests biogenesis via a PPR10-like mechanism. Second, we show that the PPR protein HCF152 binds to the immediate 5'- or 3'-termini of transcripts that require HCF152 for their accumulation, providing evidence that HCF152 defines RNA termini by blocking exonucleases. Finally, we build on the observation that the PPR10 and HCF152 binding sites accumulate as small chloroplast RNAs to infer binding sites of other PPR proteins. We show that most processed mRNA termini are represented by small RNAs whose sequences are highly conserved. We suggest that each such small RNA is the footprint of a PPR-like protein that protects the adjacent RNA from degradation.

Figure 1.

Mapping RNA termini in the maize and barley rps12–clpP intergenic region. (A) Primer extension analysis of the rps12 5′-end in maize. The ddC and ddA sequencing ladders identify the positions of G and U residues in the RNA template. Two RNA samples were analyzed (WT-1 and WT-2). (B) RNA gel blot hybridizations, using maize seedling leaf RNA and the three oligonucleotide probes diagramed in (C). The panels came from adjacent lanes of the same gel. Kb-RNA size markers. (C) Maize RNA sequence annotated with the 5′- and 3′-termini determined by primer extension and cRT-PCR, respectively. A multiple sequence alignment of the same region is annotated with the position of a small chloroplast RNA. (D) Histogram of barley transcriptome sequence reads mapping to the rps12 5′-region. TEX+ data are derived from a library that had been treated with Terminator Exonuclease, which will degrade processed 5′-termini. TEX- data were derived from an untreated library, and represent both processed and unprocessed 5′-ends. The plateau illustrates an sRNA that matches the sequence at the overlapping 5′- and 3′-termini in the clpP–rps12 intergenic region, and that matches a region of high conservation shown in panel (C). 3′-ends identified in barley by 3′-RACE are marked with arrows and annotated with the number of clones corresponding to each position.

Figure 2.

Mapping the binding site of recombinant HCF152. (A) Elution of affinity-purified MBP-AtHCF152 from a gel filtration column. Aliquots of consecutive fractions were analyzed by SDS–PAGE and staining with Coomassie Blue. The bracketed fractions were pooled and used for RNA binding assays. The elution positions of globular size standards are shown below. The elution profiles of MBP-OsHCF152 and MBP-ZmHCF152 were similar to that for MBP-AtHCF152 (data not shown). (B) Gel mobility shift assays demonstrating sequence-specificity of MBP-HCF152. The proteins indicated were used in RNA binding assays with the radiolabeled RNA oligonucleotides shown below. RNA 4 corresponds to the sequence proposed previously to bind HCF152 (23). Protein concentrations were 0, 35 and 75 nM (left panel), or 0, 12, 25 and 75 nM (right panel). Bound (B) and unbound (U) RNAs were separated by native gel electrophoresis. (C) Genomic context and evolutionary conservation of the HCF152 binding site. The positions of the HCF152-dependent 5′- and 3′-termini (7,8) are marked. The HCF152 binding site is represented by a small RNA in barley chloroplasts, as shown by the histogram of sequence reads below.

Figure 3.

Barley chloroplast transcriptome data documenting sRNAs matching known binding sites of PPR proteins. The binding sites of PPR10 and PGR3, and the RNA termini that are stabilized by these proteins are marked (7,10,14,29). The positions on the atpH RNA at which PPR10 blocks exonucleases in vitro (10) correspond with the borders of the sRNA, providing evidence that the sRNA is PPR10's in vivo footprint.

Figure 4.

RNA gel blots demonstrating PPR-dependent accumulation of two sRNAs. Total leaf RNA (15 µg) of the indicated genotypes was fractionated in denaturing polyacrylamide gels and electrophoretically transferred to charged nylon membrane. Wild-type (WT) samples came from phenotypically normal siblings grown in parallel. The ppr10 and crp1 mutants were described previously and were shown to be null alleles (6,7,10). Duplicate blots were hybridized with oligodeoxynucleotide probes that are diagrammed in the context of the barley transcriptome data to the right. Synthetic RNA oligonucleotides that mimic each sRNA were included in adjacent lanes, and are also diagrammed. The maize and barley sequences are identical in the regions encoding these sRNAs. A portion of the ethidium bromide (EtBr) stain of one of the gels is shown below to illustrate equal sample loading.

Figure 5.

Examples of sRNAs that we suggest to be footprints of uncharacterized PPR-like proteins. Histograms of barley sequence reads are annotated with the positions of mapped mRNA termini. The most stable structure predicted for the sRNA by MFold (37) is shown to the right. The predicted structures are very unstable in comparison with those that are known to stabilize 3′-termini in chloroplasts, which average approximately −25 kcal/mol (22). Each sequence alignment ends with the start codon of the downstream gene. (A) Example of an sRNA corresponding with a processed 5′-end. This sRNA was not detected in Arabidopsis, but its sequence is highly conserved. (B) Example of an sRNA corresponding with a processed 3′-end. The atpF 3′-end has been mapped in maize (7) but not in barley. (C) Example of an sRNA mapping near a transcription start site (TSS). The TSS is inferred from the position of the TEX-resistant 5′-end. The 3′-half of the plateau of sequence reads is TEX-sensitive, corresponds with a conserved sequence element, and is not predicted to form a stable structure. cRT-PCR data place an ndhA 3′-end at the 3′-end of this sRNA (Supplementary Table S1).