Identification and characterization of RNA cis-elements and nucleases required for 5' processing of mitochondrial RNAs in Arabidopsis thaliana
Final Report Abstract
In Arabidopsis mitochondria, at least seven RNA Processing Factors (RFP1 to RPF7) and Restorer of Fertility-Like Protein 2 (RFL2) are required for efficient posttranscriptional generation of 5' ends of several mitochondria-encoded transcripts. These factors are P-type pentatricopeptide repeat proteins that usually do not exhibit any enzymatic activity. Thus, it is speculated that additional proteins are required, which perform the exonucleolytic or most likely endonucleolytic cuts to form the mature 5' termini. We have identified the potential Mitochondrial Nucleases 1 and 2 (MNU1 and 2), suggested to be involved in 5' end maturation of some mitochondrial RNAs. However, the exact roles and molecular functions as ribonucleases of these proteins remained unknown. Likewise, the binding sites of RPFs and their molecular roles were still unclear. In the completed funding period, we used CRISPR-Cas9 to establish RPF3 and RPF6 single and double knockout mutants in Col and C24. Transcript analyses in these mutants unambiguously confirm the clear-cut correlation of RPF3 and RPF6 functions with 5' processing of ccmC RNAs derived from distinct mitochondrial ccmC genotypes with their specific upstream sequence elements. Recombinant RPF3 indeed binds to a predicted binding site in this sequence element in vitro. Binding was also observed upstream but not downstream of the predicted site. RPF3 and RPF6 act completely independent from each other and are irrelevant for processing of other transcripts. The function of both proteins depends on a conserved C-terminal sequence downstream of the PPR motifs. In contrast to RPF6, the addition of protein tags to RPF3 inactivated the protein in vivo, but tagged RPF6 was undetectable in mitochondrial fractions as it was seen for another RPF protein. To facilitate the identification of potential binding sites for PPR proteins, we developed an in silico prediction tool based on data describing the PPR code. In addition to RPF3 and RPF6, we identified and characterized RPF8 and analyzed an additional P-type PPR protein that influences the level of ccmC transcripts probably as a secondary effect to its influence on the abundance of rpl5-cob mRNAs. The in vivo analyses of the MNU2 protein revealed that the mutagenesis or removal of most of the conserved domains of this protein did not influence its function with respect to the formation of nad3-rps12 transcripts in vivo. Surprisingly, also an exchange of conserved amino acids in the NYN nuclease domain has no clear negative effect on nad3-rps12 transcript formation suggesting a function of MNU2 beyond its potential role as a nuclease. In line with this observation, we haven’t found clear evidence for nuclease activity of this protein in vitro. However, our studies do not allow final conclusions in this respect. In contrast to all other domains, the MNU2 Lotus 1 domain is essential for MNU2 function and results of RNA oligonucleotide pull down assays suggest this domain to be required for binding to RNA.
