Transcriptional mapping and RNA processing of the Plasmodium falciparum mitochondrial mRNAs.

The mitochondrial genome of Plasmodium falciparum encodes three protein coding genes and highly fragmented rRNAs. The genome is polycistronically transcribed and, since gene-size transcripts are much more abundant than the polycistronic transcripts, the latter are presumably cleaved to produce the smaller, mature mRNAs and rRNAs. Mapping the transcripts of the P. falciparum mitochondrial protein coding genes shows that the 3' end of each gene directly abuts the 5' end of the gene located immediately downstream. The 5' ends of the protein coding genes are also closely apposed to adjacent genes, with one directly abutting a gene on the same DNA strand and two others separated by just 13 nt from an rDNA fragment encoded on the opposite strand. These mapping data are consistent with production of the mRNAs by cleavage from a polycistronic precursor transcript. Further processing of the mRNAs comes from addition of oligo(A) tails. Unexpectedly, the presence and length of such tails varies in a gene-specific fashion. In this regard, polyadenylation of the P. falciparum mitochondrial mRNAs is more similar to that seen for the P. falciparum mitochondrial rRNAs than that of mitochondrial mRNAs in other organisms.

The fragmented mitochondrial ribosomal RNAs of Plasmodium falciparum have short A tails.

The mitochondrial genome of Plasmodium falciparum encodes highly fragmented rRNAs. Twenty small RNAs which are putative rRNA fragments have been found and 15 of them have been identified as corresponding to specific regions of rRNA sequence. To investigate the possible interactions between the fragmented rRNAs in the ribosome, we have mapped the ends of many of the small transcripts using primer extension and RNase protection analysis. Results obtained from these studies revealed that some of the rRNA transcripts were longer than the sequences which encode them. To investigate these size discrepancies, we performed 3' RACE PCR analysis and RNase H mapping. These analyses revealed non-encoded oligo(A) tails on some but not all of these small rRNAs. The approximate length of the oligo(A) tail appears to be transcript-specific, with some rRNAs consistently showing longer oligo(A) tails than others. The oligoadenylation of the rRNAs may provide a buffer zone against 3' exonucleolytic attack, thereby preserving the encoded sequences necessary for secondary structure interactions in the ribosome.

Homeless is required for RNA localization in Drosophila oogenesis and encodes a new member of the DE-H family of RNA-dependent ATPases.

The homeless (hls) gene of Drosophila is required for anteroposterior and dorsoventral axis formation during oogenesis. At a low frequency, females homozygous for mutations in hls generate early egg chambers in which the oocyte is positioned incorrectly within the cyst. At a high frequency, late-stage egg chambers exhibit a ventralized chorion. Sequence analysis of the hls cDNA predicts a protein with amino-terminal homology to members of the DE-H family of RNA-dependent ATPases and putative helicases. Similarity of 51% in the amino-terminal third of the protein was found to two yeast splicing factors, PRP2 and PRP16, and to Drosophila Maleless, which is required for dosage compensation. To analyze Hls function, RNA localization patterns were determined for seven different transcripts in hls mutant ovaries. Previtellogenic transport to the oocyte was unaffected for all transcripts examined. Transport and localization of bicoid and oskar messages during vitellogenic stages were strongly disrupted, and the distribution and/or quantity of gurken, orb, and fs(1)K10 mRNAs were also affected, but to a lesser degree. In contrast, hu-li tai shao and Bicaudal-D transcripts were transported and localized normally in hls mutants. In addition, Kinesin heavy chain:beta-Galactosidase fusion protein failed to localize correctly to the posterior of the oocyte in vitellogenic egg chambers. Examination of the microtubule structure with anti-alpha-Tubulin antibodies revealed aberrant microtubule organizing center movement and an abnormally dense cytoplasmic microtubule meshwork. We discuss potential roles for Hls in organizing a cytoskeletal framework essential for localizing specific RNAs.

Endonucleolytic cleavage of a long 3'-trailer sequence in a nuclear yeast suppressor tRNA.

Transcripts of Saccharomyces cerevisiae nuclear tRNA genes are normally terminated within a few nucleotides of the tRNA coding region, in contrast to mitochondrially encoded tRNAs, which are contained within polycistronic transcripts and thus require 3'-processing by mitochondrial endonucleases. We show that 3'-processing activities capable of removing artificially extended 3'-trailer sequences from some tRNA substrates are also present in the yeast nucleus. Correct 3'-processing in vivo resulted in the formation of functional suppressor tRNA. The 3'-processing activities were also identified in vitro through analysis of transcription-processing products in cell-free yeast S-100 extracts. Comparison of several pre-tRNA substrates showed that the tRNA structure played a major role in determining the processability of a substrate but that the nature of the 3'-trailer sequence also modulated the rate of 3'-processing. Pre-tRNA containing mitochondrial tRNA(Val) sequence was a good substrate for in vitro processing, independent of its 3'-trailer. A 200-nt-long pre-tRNA, encoding the nuclear SUP4 tRNA gene and a mitochondrial 3'-trailer, was processed in yeast S-100 extract in a multistep pathway into mature-sized tRNA(Tyr). Part of the 3'-processing was due to an endonuclease which cleaved near or precisely at the 3'-end of the coding region of the tRNA. A short sequence around this endonucleolytic 3'-cleavage site was crucial for the formation of active suppressor tRNA in vivo. A 9-nt-long sequence motif derived from the mitochondrial 3'-trailer allowed processing, while sequences derived from lacZ or pBR322 DNA were processed neither in vitro nor in vivo.