Identification of nuclear encoded precursor tRNAs within the mitochondrion of Trypanosoma brucei.

RNAs that function in mitochondria are typically encoded by the mitochondrial DNA. However, the mitochondrial tRNAs of Trypanosoma brucei are encoded by the nuclear DNA and therefore must be imported into the mitochondrion. It is becoming evident that RNA import into mitochondria is phylogenetically widespread and is essential for cellular processes, but virtually nothing is known about the mechanism of RNA import. We have identified and characterized mitochondrial precursor tRNAs in T. brucei. The identification of mitochondrially located precursor tRNAs clearly indicates that mitochondrial tRNAs are imported as precursors. The mitochondrial precursor tRNAs hybridize to cloned nuclear tRNA genes, label with [alpha-32P]CTP using yeast tRNA nucleotidyltransferase and in isolated mitochondria via an endogenous nucleotidyltransferase-like activity, and are processed to mature tRNAs by Escherichia coli and yeast mitochondrial RNase P. We show that T. brucei mitochondrial extract contains an RNase P activity capable of processing a prokaryotic tRNA precursor as well as the T. brucei tRNA precursors. Precursors for tRNA(Asn) and tRNA(Leu) were detected on Northern blots of mitochondrial RNA, and the 5' ends of these RNAs were characterized by primer extension analysis. The structure of the precursor tRNAs and the significance of nuclear encoded precursor tRNAs within the mitochondrion are discussed.

The mitochondrial tRNAs of the protozoan parasite Trypanosoma brucei are encoded by the nuclear DNA Mottram et al., 1991a). Until recently, it was accepted that the mitochondrial genome encodes the mitochondrial rRNAs, the mitochondrial tRNAs, and a selected set of the mitochondrial proteins involved in electron transport. However, it is now evident that mitochondrial tRNAs, in a variety of species, are nuclear encoded. Like T. brucei, all of the mitochondrial tRNAs of Leishmania tarentolae, a closely related kinetoplastid, are nuclear encoded . Evidence indicates that two other protozoans, Tetrahymena (Suyama, 1967(Suyama, , 1986 and Paramecium (Pritchard et al., 1990); an algae, Chlamydomonas (Gray and Boer, 1988); and several higher plants, Phaseolus (Marechal-Drouard et al., 1988), Solanum (Marechal-Drouard et al., 1990), and Triticum (Joyce and Gray, 1989) encode ~ * This work was supported by National Institutes of Health Grant A121401 (to S. L. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Burroughs Wellcome Scholar for Molecular Parasitology. To whom correspondence and reprint requests should be sent. Tel.: 205-934-6033;Fax: 205-975-2547. some but not all of their mitochondrial tRNAs within the nuclear genome. While mitochondrial tRNAs are encoded by the mitochondrial genome in mammals, the RNA component of RNase MRP, a mitochondrial RNA processing enzyme, and probably the RNA component of mitochondrial RNase P are nuclear encoded (Clayton, 1991). RNAs that are mitochondrially located but encoded by the nuclear DNA clearly must be transcribed in the nucleus and transported across the mitochondrial membranes. The identification of nuclear encoded mitochondrial RNAs in protozoa, algae, higher plants, and mammals indicates that RNA import is phylogenetically widespread. Considering the role of the imported RNAs in mitochondrial translation and mitochondrial RNA processing, the import of RNA appears to be essential for cellular processes.
While the pathway of protein import into mitochondria is well characterized (Hart1 and Neupert, 1990;Baker and Schatz, 1991), virtually nothing is known about the pathway of RNA import. One study showed that short pieces of singlestranded or double-stranded DNA, when covalently attached to the C terminus of a precursor protein, could be imported into mitochondria through the protein import pathway (Vestweber and Schatz, 1989). Another investigator has suggested that mitochondrial tRNAs are imported as complexes with their cognate synthetases (Suyama and Hamada, 1978). However, there is no evidence to support this hypothesis.
Sixteen T. brucei tRNA genes have been cloned and sequenced. Fourteen of the tRNA genes are arranged in small clusters of two to five separated by 58-102 nt' (Campbell, 1989;Campbell et al., 1989;Mottram et al., 1991b;Hancock and Hajduk, 1992). The other two tRNA genes (Mottram et al., 1991a;Hancock and Hajduk, 1992) appear to occur singly; however, the clones sequenced were small and adjacent tRNAs may have been excluded. All T. brucei tRNA genes identified have consensus box A and box B intragenic polymerase I11 control regions (Geiduschek and Tocchini-Valentini, 1988;Padayatty, 1990). All of the genes also have runs of T residues from 4 to 8 nt in length near the 3' terminus of the tRNA. Runs of T residues are associated with transcription termination of polymerase I11 transcripts (Geiduschek and Tocchini-Valentini, 1988). Transcription of T. brucei tRNA genes exhibits an intermediate level of sensitivity to a-amanitin, which is characteristic of RNA polymerase 111.' Information regarding the primary transcripts from these genes or the processing of the transcripts has not been reported. In all other systems studied, tRNAs are transcribed as precursors which undergo processing at their 5' and 3' termini to generate mature tRNAs (Deutscher, 1984

23963
in T. brucei Mitochondrion cated precursor tRNAs. The identification of precursor tRNAs within the mitochondrion defines one feature of the tRNA import pathway. Import is via a precursor not a mature tRNA. The precursor tRNAs, like the mature tRNAs, label metabolically in isolated mitochondria with [a-32P]CTP in the absence of transcription. Incorporation of CTP posttranscriptionally at the 3' terminus of tRNA via tRNA nucleotidyltransferase is a well characterized activity (Deutscher, 1982(Deutscher, , 1984. We refer to these metabolically labeled precursor tRNAs as C-RNAs (Harris et d., 1990). C-RNAs hybridize to cloned tRNA genes and are substrates for two well characterized tRNA processing enzymes, tRNA nucleotidyltransferase and RNase P. Specific, C-RNA-sized precursors were detected by Northern blot and primer extension analyses using oligonucleotide probes complementary to mature tRNAA"" and tRNAk". The identification of mitochondrial precursor tRNAs is an important first step in establishing the biosynthetic pathway for mitochondrial tRNA9 and in identifying the substrate for mitochondrial import.

EXPERIMENTAL PROCEDURES
Isolation of Mitochondria and Mitochondrial RNA-The procyclic form of Trypanosoma brucei brucei (TREU 667) was grown in a semidefined medium containing 10% heat-inactivated fetal bovine serum and 10 pg/ml gentamicin sulfate (Cunningham, 1977). Mitochondria were isolated from 4 liters of culture at a cell density of 1-2 X lo7 as described by Harris et al. (1990). Isolated mitochondrial vesicles were treated with micrococcal nuclease (Pharmacia LKB Biotechnology Inc.) as described by  except that 0.5 pl of 10% glycerol, 10 mM Tris-HC1, pH 8.0, 1 mM CaC12 buffer was used for each 3.5 X lo6 cells. Mitochondria were prepared under conditions in which minimal contamination with cytoplasmic and nuclear RNAs was detected . The mitochondrial vesicles were solubilized and the mitochondrial RNAs extracted and examined . To gel-purify, mitochondrial RNA was electrophoresed in a 6% acrylamide, 8 M urea gel. The mitochondrial tRNAs were visualized by UV shadowing, and the region of the gel containing C-RNAs was determined by comigration of metabolically labeled C-RNAs as visualized by autoradiography. The RNAs were excised, eluted in 0.5 M ammonium acetate, 0.1% SDS, 1.0 mM EDTA, pH 8.0, overnight at 37 "C, then precipitated with isopropyl alcohol.
Metabolic Labeling of Mitochondrial RNA with [a-32P]CTP-Mitochondrial vesicles were suspended in transcription buffer containing 5 mM Hepes, pH 7.6, 3 mM potassium phosphate, pH 7.7,125 mM sucrose, 10 mM magnesium acetate, 1 mM EDTA, pH 8.0, 5 mM 2mercaptoethanol, and 1 mM ATP. The vesicles were incubated for 30 min at room temperature to deplete the endogenous nucleotide pools. Mitochondrial RNAs were labeled in transcription buffer with 500 pCi/ml [(u-~'P]CTP (3000 Ci/mmol) and 500 pM ATP for 30 min at room temperature. All radioisotopes were purchased from Du Pont-New England Nuclear. Mitochondrial vesicles from 4 liters of cells were labeled in a total volume of 4 ml. Vesicles were recovered in the pellet following centrifugation in a microcentrifuge for 30 s at room temperature. The vesicles were solubilized and the mitochondrial RNA extracted . Metabolically labeled C-RNAs and mitochondrial tRNAs were gel-purified as described above.
Digestion of C-RNAs with RNase P-Metabolically labeled C-RNAs were gel-purified on a 6% acrylamide, 8 M urea gel. The C-RNAs were recovered in two populations, an upper and a lower band. Samples of these two bands were run on a 12% acrylamide, 8 M urea gel and were found to correspond to the two populations of C-RNAs shown in Fig. 1 (lane 5 ) . The lower band was essentially free of contamination by the upper band, but the upper band contained a small amount of the lower band (data not shown). Both the upper and lower bands were used as substrates for Escherichia coli RNase P, yeast mitochondrial RNase P, and a crude 2' . brucei mitochondrial extract.
pDW98, containing the E. coli RNase P RNA template, was a gift from Drs. Norman Pace and Drew Smith (Indiana University). To obtain RNase P RNA, pDW98 was linearized with SnaB I (Stratagene) and transcribed with T7 RNA polymerase (Bethesda Research Laboratories), using a protocol from Promega for the synthesis of large amounts of unlabeled RNA. E. coli RNase P in this paper refers to the RNA rather than to the holoenzyme. For each assay, 830 ng of E. coli RNase P was used. S. cereuisiae mitochondrial RNase P was a gift from Dr. Nancy C. Martin. For each assay, 10 p1 of partially purified yeast mitochondrial RNase P (the active fraction from a Mono S column) was used. The T. brucei mitochondrial extract was prepared by resuspending 1.5 X 10" cell equivalents of mitochondrial vesicles in 1 ml of MX buffer (50 mM Tris-HC1, pH 8.0, 25 mM KC1, 2.5 mM magnesium acetate, 1 mM EDTA, pH 8.0,0.5 mM ditbiothreitol, 0.5 mM ATP, and 10% glycerol). Triton X-100 was added to a final concentration of 0.5%, and the mixture was centrifuged in a microcentrifuge for 10 min at 4 "C. The supernatant was recovered, and 5 pl of this mitochondrial extract was used for each assay.
C-RNAs were incubated with E. coli RNase P in 1.0 M NHdCl, 25 mM MgC12, 50 mM Tris-HC1, pH 8.0, 0.1% SDS, and 0.05% Nonidet P-40 for 40 min at 37 "C. C-RNAs were incubated with yeast mitochondrial RNase P in 50 mM Tris-HC1, pH 8.0, 10 mM MgC12, 30 mM NaCl, 200 pg/ml RNase-free bovine serum albumin, and 0.25% Tween 20 for 40 min at room temperature. They were incubated with T. brucei mitochondrial extract in MX buffer for 40 min at room temperature. As a control, the substrates were mock-treated in MX buffer for 40 min at room temperature. The total volume of each reaction was 50 pl. The reactions were stopped by extracting twice with pheno1:chloroform (l:l), once with chloroform, followed by ethanol precipitation. The products were analyzed on a 4% acrylamide, 8 M urea gel.
Assay for RNase PActiuity"pDW153, containing the template for E. subtilis pre-tRNAAsP, was a gift from Drs. Norman Pace and Drew Smith. pDW153 was linearized with MuaI (Boehringer Mannheim) and transcribed using T7 RNA polymerase according to protocols from Promega. To obtain uniformly labeled pre-tRNAASP, [a-32P]CTP or [(u-~~PIUTP was included in the transcription reaction. To 5'-endlabel, unlabeled pre-tRNAAsp was treated with bacterial alkaline phophatase (Bethesda Research Laboratories) then labeled with [y-"P] ATP using T4 kinase (Bethesda Research Laboratories) according to tRNAAsP was labeled with [32P]pCp using T4 RNA ligase (Bethesda the manufacturer's recommendations. To 3'-end-label, unlabeled pre-Research Laboratories) according to the manufacturer's recommendations. E. coli RNase P (510 ng), 5 p1 of yeast mitochondrial RNase P, and 2.5 pl of T. brucei mitochondrial extract were reacted with equal counts of the radiolabeled pre-tRNAAnp substrates. All reactions were for 30 min under the conditions given above. The reaction products were analyzed on a 6% acrylamide, 8 M urea gel.
Southern Blot Analysis-T. brucei nuclear and mitochondrial DNA was isolated as described by Fairlamb et al. (1978) and . T. brucei tRNA gene clones are described by Hancock and H a j d~k .~ Nuclear DNA (5 pg), mitochondrial DNA (1 pg), and plasmids containing the cloned tRNA genes (3 pg each) were digested with restriction endonucleases purchased from Bethesda Research Laboratories. The digests were electrophoresed in a 1% agarose gel in 89 mM Tris, 89 mM boric acid, 2 mM EDTA (1 X TBE) with 0.5 pg/ml ethidium bromide. The DNA was visualized by UV transillumination, photographed, then transferred to nitrocellulose membrane (Schleicher & Schuell) via capillary transfer (Sambrook et al., 1989).
Metabolically labeled and gel-purified C-RNAs at 4 X 10' cpm/ml were used as a probe. C-RNAs and gel-purified mitochondrial tRNAs were cleaved with S1 nuclease. S1 nuclease cleaves tRNA at the anticodon loop and yields two half-molecules (Harada and Dahlberg, 1975). Metabolically labeled and gel-purified C-RNAs were digested with S1 nuclease, in the presence of 1 pg of wheat germ tRNA, as described by . The mitochondrial tRNA half-molecules were 3'-end-labeled with [32P]pCp as described above.
The Sl-cut C-RNAs at 2 X lo5 cpm/ml and the S1-cut mitochondrial K. Hancock and S. L. Hajduk, unpublished observations. t,RNAs a t 3 X lo6 cpm/ml were used as hybridization probes. Prehybridization and hybridization was performed as described by . Uncut and S1-cut C-RNA probes were washed to 1 X SSC, 0.1% SDS at 55 "C. S1-cut mitochondrial tRNA probes were washed to 0.1 X SSC, 0.1% SDS at 55 "C (1 X SSC is 150 mM NaCI, 15 mM sodium citrate, pH 7.0). The C-RNA hybridizations were stable when washed to 0.1 X SSC, 0.1% SDS at 55 "C, hut the hybridization signals with clones pTbt2 and pTbt3 were very weak (data not shown).
Oligonucleotides-Oligonucleotides complementary to the D region and to the anticodon (AC) region of T. brucei tRNA asparagine (tRNAA""):' and 7'. brucei tRNA leucine (tRNA*"")  were purchased from Oligos Etc. Inc. The oligonucleotides and their sequences are shown in Fig. 5A.
Northern Blot Analysis-Mitochondrial RNA (10 pg) and cytosolic RNA (20 pg) were electrophoresed in a 6% polyacrylamide, 8 M urea gel, then electrotransferred onto Genescreen Plus (Du Pont-New England Nuclear) in 40 mM Tris, 20 mM acetic acid, 1 mM EDTA (1 X TAE). Following transfer, the RNAs were UV-cross-linked to the membrane using the automatic cross-link mode of Stratagene's Stratalinker. Chloroform-extracted Asn D, Asn AC, Leu D, and Leu AC oligonucleotides were labeled with [-y-'*P]ATP (6000 Ci/mmol) a t a 1:3 molar ratio using T4 polynucleotide kinase (Bethesda Research Laboratories). The labeled oligonucleotides were gel-purified and used t o probe the Northern blots. Prehybridization was in 1 M NaCI, 10% dextran sulfate, 1% SDS at 37 "C. For hybridization, denatured salmon sperm DNA, a t a final concentration of 100 pg/ml, and the labeled oligonucleotide, a t 2-5 X lo6 cpm/ml, were added to fresh prehybridization solution. The blot was hybridized a t 37 "C. The Asn D-probed Northern blot shown was washed to 0.1 X SSC, 1% SDS; the Asn AC to 0.5 X SSC, 1% SDS; the Leu D to 0.1 X SSC, 1% SDS; and the Leu AC to 1 X SSC, 1% SDS. The Asn AC and the Leu AC hybridizations were stable when washed to 0.1 X SSC, 1% SDS (data not shown). All washes were a t 37 "C.
Primer Extension Analysis-Primer extension analysis of mitochondrial RNA (9 pg), cytosolic RNA (9 pg), gel-purified mitochondrial tRNAs (1 pg), and gel-purified C-RNAs (0.3 pg) was performed essentially as described by Sambrook et al. (1989). Oligonucleotides were labeled and gel-purified as described for the Northern blots. Asn D oligonucleotide (2 X 10' cpm ) and Leu D oligonucleotide (5 X lo6 cpm) were hybridized with each RNA. A control reaction without RNA was also included. The hybridization mixtures were incubated at 85 "C for 15 min, then allowed to slow cool overnight to 30 "C. The extension reactions were done a t 40 "C for 1 h, then shifted to 42 "C for 1 h in pH 8.3 reverse transcriptase buffer using 37.5 units of avian myeloblastosis virus reverse transcriptase (Promega). The extension products were analyzed on 8 and 12% acrylamide, 8 M urea gels.

RESULTS
Characterization of Mitochondrial C-RNAs-RNA in mitochondrial vesicles from the procyclic form of T. brucei can be post-transcriptionally labeled with [cY-~~PICTP (Harris et al., 1990). Transcription is arrested by a 30-min incubation during which the endogenous nucleotide pools are depleted. The addition of [a-:"P]CTP results in the incorporation of label into two populations of RNA (Fig. 1, lane 2). One population, migrating between 70 and 86 nt, comigrates with the mitochondrial tRNAs (lane I ) seen by 3'-end-labeling of total mitochondrial RNA . These RNAs are mitochondrial tRNAs that have been labeled by turnover of their 3' end sequence, CCA, a post-transcriptional process (Deutscher, 1982(Deutscher, , 1984. The other is a population of larger RNAs that we refer to as C-RNAs, since they also label with CTP in the absence of transcription. C-RNAs migrate between approximately 120-180 nt on a 6% acrylamide, 8 M urea gel. On a 12% acrylamide, 8 M urea gel, the tRNAs shift up and run between 75 and 95 nt and the C-RNAs split into two populations. One population migrates between 165 and 190 nt, the other between 500 and 735 nt. The reason for the shift of a portion of the C-RNAs from less than 200 to greater than 500 nt is unknown. However, a large apparent increase in size with increasing polyacrylamide concentration is characteristic of RNAs with an unusual structure such as circles or lariats (White and Borst, 1987; van der Veen et al., 1986). The faint bands (lane 2) that comigrate with the mitochondrial 1150-nt 12 S rRNA and 611-nt 9 S rRNA (lane I ) (Sloof et al., 1985) are most likely the result of a small amount of residual transcription.
It is striking that C-RNAs, which are metabolically labeled so prominently (Fig. 1, lane 2), are not detected when total mitochondrial RNA is 3'-end-labeled (lane I ) . It is possible that the 3' ends of C-RNAs are not a suitable substrate for [32P]pCp addition by T4 RNA ligase. However, when total mitochondrial RNA is 5'-end-labeled with [y3'P]ATP using T4 polynucleotide kinase, C-RNAs are still not detectable (data not shown). This suggests that our inability to detect C-RNAs in the total mitochondrial RNA population is not due to labeling inefficiency. Thus, the absence of pCp labeled C-RNAs may reflect their low abundance in the total mitochondrial RNA population. This apparent low abundance and the post-transcriptional labeling of C-RNAs under the same conditions that label tRNAs, through turnover of their CCA tail, suggest that C-RNAs may be tRNA precursors characterized by a 5' extension and a mature 3' end.
In order to test the hypothesis that C-RNAs have a tRNA like mature 3' end, unlabeled C-RNAs were tested for their ability to serve as substrates for purified yeast ATP(CTP)-tRNA nucleotidyltransferase (CCA enzyme) (Chen et al., 1990) in the presence of [cY-~*P]CTP (Fig. 2). Nucleotidyltransferase, in the presence of CTP, catalyzes the sequential addition of 2 C residues to the 3' end of tRNAs lacking all or a portion of the 3'-terminal sequence CCA (Deutscher, 1982). An A residue will be added after the C-C addition if ATP is included in the reaction. Using yeast nucleotidyltransferase, the mitochondrial tRNAs in total mitochondrial RNA (Fig.   2, lane I ) and in gel-purified mitochondrial tRNA (lane 2) are intensely labeled. The nucleotidyltransferase-labeled tRNAs comigrate with metabolically labeled tRNAs (lane 6 ) . A population of RNAs that comigrates with metabolically labeled C-RNAs (lane 5 ) is barely visible in lune 1. With a longer exposure (lune 3), these RNAs are easily seen. They are labeled when total mitochondrial RNA is the substrate (lunes 1 and 3) but not when gel-purified mitochondrial tRNA is the substrate (lanes 2 and 4 ) . There is some difference between the pattern of the larger RNA labeling in lane 3 and the pattern of C-RNA labeling in lane 5. Most likely this is due to the fact that in lune 3 the labeling was done in vitro using a heterologous enzyme, while in lane 5 the labeling was done in isolated mitochondria with the homologous enzyme. We were unable to detect C-RNA size RNAs when cytoplasmic RNA was used as a substrate for the nucleotidyltransferase enzyme (data not shown). This indicated that the mitochondrial cRNAs are not contaminants of abundant cytoplasmic pre-tRNAs in transit to the mitochondria.
The identification of a population of RNAs that are substrates for nucleotidyltransferase and are larger than mature tRNAs implies that these RNAs are precursor tRNAs possessing mature 3' ends. The fact that these RNAs comigrate with C-RNAs strongly suggests that C-RNAs are mitochon-brucei Mitochondrion drial precursor tRNAs with a 5' end extension. RNase P is an enzyme whose function is to form the mature 5' end of tRNAs (Altman, 1989(Altman, , 1990Pace and Smith, 1990). If C-RNAs are precursor tRNAs with an immature 5' end, then C-RNAs should be substrates for RNase P and T. brucei mitochondria should have an RNase P activity.
Metabolically labeled C-RNAs were gel-purified on a 6% acrylamide, 8 M urea gel into two populations, an upper band (Fig. 3A, lane 1) and a lower band (lane 7). The upper band C-RNAs (lunes 4-6) and the lower band C-RNAs (lanes 10-12) were treated with E. coli RNase P, yeast mitochondrial RNase P, and a T. brucei mitochondrial extract. With the two RNase P enzymes and with the mitochondrial extract, 25-75% of the lower band C-RNAs were converted to mature tRNA size products (compare with lanes 2 and 8 ) . Under identical conditions, virtually none of the upper band C-RNAs were converted. The small amount of conversion from the upper band is most likely due to the contamination of the upper band by the lower band. Upper band C-RNAs are not a substrate for RNase P. It may be that upper band C-RNAs contain sequences that must be removed prior to RNase P cleavage or they may have an unusual structure that is not a substrate for RNase P. C-RNAs that are mock-treated (lanes 3 and 9 ) do not convert to mature tRNAs, demonstrating that the appearance of the mature tRNA is dependent upon the addition of enzyme.
Nonspecific nucleases can generate mature size tRNAs from precursor tRNAs by degrading the relatively unstructured portions of the precursor. The mature tRNA portion of the precursor is resistant to degradation by virtue of its secondary and tertiary structure. In order to determine whether the mitochondrial extract contains an RNase P activity, it is necessary to detect both the 5' and 3' cleavage products produced when a precursor tRNA is the substrate. A 112-nt B. subtilis pre-tRNA""" was used as the defined substrate (Fig. 3B) for E. coli RNase P (lanes 1-3), yeast mitochondrial RNase P (lunes 4-6), and T. brucei mitochondrial extract (lanes 7-9). For the two RNase Ps, cleavage of the uniformly labeled pre-tRNA yielded two products, a 35and a 77-nt RNA (lanes 1 and 4 ) . 5'-end-labeled pre-tRNA was cleaved to a 35-nt RNA (lanes 2 and 5 ) , the predicted 5' cleavage product. 3'-end-labeled pre-tRNA was cleaved to a 77-nt RNA (lunes 3 and 6 ) , the predicted 3' cleavage product.
When uniformly labeled pre-tRNAAs" was incubated with the 7'. brucei mitochondrial extract (Fig. 3B, lane 7), two cleavage products were again seen. One is the 77-nt mature tRNA and the other is the 35-nt 5' extension. This result is confirmed by the presence of the 35-nt RNA when 5'-endlabeled pre-tRNA was used as the substrate (lune 8) and by the presence of the 77-nt RNA when 3'-end-labeled pre-tRNA was used as the substrate (lane 9 ) . The possibility that the RNAse P activity measured was due to nuclear contamination was addressed by electron microscopy of the Percoll gradient fractions. No nuclei were seen in fractions containing mitochondria (data not shown). The contaminants in these fractions were flagella and glycosomes. These results show that T. brucei contain a mitochondrial RNase P activity and provide support for the presence of precursor tRNAs in the mitochondrion. In addition to the cleavage products, there was a product larger than 112 nt when kinase-labeled substrate was treated with mitochondrial extract (lane 8). The appearance of this band is probably due to the endogenous RNA ligase activity in trypanosome mitochondria, which use RNAs with 5'-monophosphates as substrates (Bakalara et ul., 1989).4 The substrate RNA in lanes 7 and 9 have 5"triphos-  and 3-6) and a lower band (lanes 7 and 9-12), were digested with E. coli RNase P (lanes 4 and IO), yeast mitochondrial RNase P (lanes 5 and Z I ) , and 7'. brucei mitochondrial extract (lanes 6 and 12). Mocktreated C-RNAs are shown in lanes 3 and 9, untreated C-RNAs in lanes 1 and 7, and metabolically labeled mitochondrial tRNAs in lanes 2 and 8. Lane M contains radiolabeled 4x174 DNA, digested with HaeIII, as size markers. R, RNase P activity in 7'. brucei mitochondrial extract. R. subtilis pre-tRNAA'" was transcribed in uitro and uniformly labeled (lanes I , 4, and 7), 5"end-labeled (lanes 2, 5 , and 8 ) , and 3"end- labeled (lanes 3 , 6 , and 9). Pre-tRNA'"'' in lanes 1-3 was digested with E. coli RNase P, in lanes 4-6 with yeast mitochondrial RNase P, and in lanes 7-9 with 7'. brucei mitochondrial extract. Pre-tRNAA"I' is 112 nt, the 5' cleavage product is 35 nt, and the 3' cleavage product is 77 nt. These RNAs are indicated by arrows. Lane 10 contains radiolabeled 6x174 DNA, digested with HaeIII, as size markers.
phates and are not substrates for RNA ligase. The conversion of lower band C-RNAs to mature tRNAs by two defined RNase P enzymes clearly shows that a subpopulation of C-RNAs is a substrate for RNase P and strongly suggests that C-RNAs are precursor tRNAs with 5' extensions. The presence of a RNase P activity in the T. brucei mitochondrial extract that also converts lower band C-RNAs to mature tRNAs indicates that processing of C-RNAs in vivo occurs through an RNase P activity in the mitochondrion.
Mitochondrial C-RNA Hybridize to tRNA Gene Clones-C-RNAs were further evaluated by Southern blot analysis. Nuclear DNA was cloned, and three clones were identified that contain tRNA genes (Hancock and Hajduk, 1992). pTbtl (Fig. 4A, lane 1 ) contains the genes for tRNA"'s and tRNA"". pTbt2 and pTbt3 (lanes 2 and 3 ) both contain an identical 423-bp fragment with the gene for tRNAA". When a Southern blot of the plasmids was probed with C-RNAs, only the DNA fragments containing the tRNA genes gave a hybridization signal (Fig. 4, B and C, lanes 1-3). The hybridization of mitochondrial tRNAs to these gene fragments is confirmed in Fig. 40. Under the hybridization conditions used, it was necessary to cleave the mitochondrial tRNAs with S1 nuclease into two half-molecules in order to obtain a hybridization signal ). The C-RNAs gave a hybridization signal when they were cut with S1 nuclease (Fig. 4B) and when they were uncut (Fig. 4C).
C-RNAs do not hybridize to mitochondrial DNA (Fig. 4B, lanes 6 and 7) but do hybridize weakly to a large PstI fragment of nuclear DNA (lane 4 ) . This fragment appears to be the same as a large PstI fragment that hybridized with mitochondrial tRNAs . No hybridization signal was seen when a HaeIII digest of nuclear DNA was probed (lane 5 ) . The cloned tRNA genes were readily detected, but tRNA genes dispersed within the total nuclear DNA were more difficult to detect. It is possible that the large PstI fragment contains multiple tRNA genes and that the generally smaller HaeIII fragments do not contain enough copies of tRNA genes to be detectable.
Detection of Mitochondrial Precursor tRNA""" and tRNA'"" by Northern Blot Analysis-Characterization of C-RNAs as substrates for nucleotidyltransferase and RNase P and the hybridization of C-RNAs to cloned tRNA genes strongly suggests that these RNAs are precursor tRNAs. In order to confirm the existence of mitochondrial precursor tRNAs, Northern blots of mitochondrial and cytosolic RNA (Fig. 5) were probed with oligonucleotides specific for T. brucei tRNA""" (see footnote 3) and tRNA'"" . Neither the tRNA""" nor tRNAIRU genes has been identified as genes encoding mitochondrially targeted tRNA. However, the similarity between mitochondrial and cytosolic tRNAs  and the similarity that different isoacceptors exhibit (Mottram et al., 1991a) enabled us to detect mitochondrial and cytosolic tRNAs. The oligonucleotides were complementary to the D and anticodon (AC) regions (Fig. 5A). All of the probes detect mature tRNA, indicated by arrowheads. From the gene sequence, the predicted size of tRNAA"" is 76 nt (Hancock and Hajduk, 1992) and the predicted size of tRNA'"" is 85 nt . The smaller size of the band detected by the Asn probes (Fig. 5, B and C), as compared to the band detected by the Leu probes (Fig. 5, D and E), correlates with the predicted size difference between the two tRNAs. The slight discrepancy in size between the mature tRNAs on blots B and C and on blots D and E is due to "smiling" of the gel. The Asn D oligonucleotide (Fig. 5B, lane 1 ) detected a larger mature tRNA that is not detected by the Asn AC oligonucleotide (Fig. 5C, lane 1 ). This tRNA may be an isoacceptor or a cross-reacting tRNA.
In addition to detecting mature tRNA, all of the probes detected larger RNAs, indicated by arrows (Fig. 5, B-E, lane I ) , that comigrate with C-RNAs (Fig. 5F). For  shown. R, a Southern hlot of the above gel hybridized with S1-cut C-RNAs. C, a Southern blot of the tRNA clones hybridized with uncut C-RNAs. D, a Southern blot of the tRNA clones hybridized with S1-cut mitochondrial tRNAs.
for a different region of a mature tRNA, detected the same larger RNAs strongly suggests that the larger RNAs are precursor tRNAs with 5' and/or 3' extensions. Each tRNA probe detected more than one precursor tRNA. Arrows mark the position of the two major bands, but for both tRNAs a minor band can also be detected. The presence of multiple bands suggests two possibilities. One is that the different sizes represent different processing intermediates. The other is that the oligonucleotide probes are detecting isoacceptors of tRNAAs" and tRNA'"".
The precursor tRNAs are unique to the mitochondrial RNA. They were not detected in the cytosolic RNA (Fig. 5, EL", lane 2). Shorter exposures of the Leu-probed Northern blots (Fig. 5 , D and E) do not show any specific bands in the cytosolic RNA other than the mature tRNA.
Characterization of the 5' End of Pre-tRNAAs" and Pre-tRNA'"" by Primer Extension Analysis-Pre-tRNAA"" and pre-tRNA'"" were further characterized by primer extension analysis to determine the length of the 5' extension. The Asn D and Leu D oligonucleotides were hybridized with mitochondrial and cytosolic RNA and extended to detect both mature tRNA (Fig. 6, A and B ) and precursor tRNA (Fig. 6, C and  D ) . The extension of the hybridized oligonucleotide to the 5' end of mature tRNA confirmed that the assay was giving the expected results. Asn D oligonucleotide hybridized with gelpurified mitochondrial tRNAs (Fig. 6A, lane 1) and total mitochondrial RNA (lane 2 ) yielded 29-and 30-nt extension products. This corresponds to a predicted size of 30 nt for extension to the 5' end of mature tRNAA"". Mature tRNA extension products of 29, 30, 32, and 33 nt were seen with cytosolic RNA (lane 5 ) . Leu D oligonucleotide hybridized with gel-purified mitochondrial tRNAs (Fig. 6B, lane 1 ) and total mitochondrial RNA (lane 2 ) yielded 27and 28-nt extension products. This corresponds to a predicted size of 27 nt for extension to the 5' end of mature tRNAku. The same extension products are seen with cytosolic RNA (lane 5 ) .
In order to distinguish the mature tRNA extension products from the labeled oligonucleotides, it was necessary to run the products on 12% acrylamide, 8 M urea gels and expose the gel to film for a fairly short period of time (Fig. 6, A and B ) . The same set of extension products were run on 8% acrylamide, 8 M urea gels and exposed for a longer period of time to detect extension products from precursor tRNA (Fig. 6, C and D).
Only total mitochondrial RNA and gel-purified mitochondrial C-RNAs were expected to contain mitochondrial precursor tRNAs. Both C-RNAs and total mitochondrial RNA should yield precursor tRNA extension products that are the same size.
The Asn D oligonucleotide yielded an extension product of 132 nt (Fig. 6C) when hybridized with both total mitochondrial RNA (lane 2) and gel-purified C-RNAs (lane 3 ) . Allowing for the 3' portion of the pre-tRNA that was not primerextended, the pre-tRNAA"" with a 3' CCA is 178 nt. Since mature tRNAAs" is 76 nt, the precursor must extend 102 nt upstream of the RNase P cleavage site. The Leu D oligonucleotide yielded a larger extension product of 172 nt (Fig. 6D) when hybridized with both total mitochondrial RNA (lane 2) and gel-purified C-RNAs (lane 3 ) . This extension product is abundant enough to be detected on the short exposure shown in Fig. 6B. Again, allowing for the 3' portion of the pre-tRNA, pre-tRNA'"" with a 3' CCA is 230 nt. Since mature tRNA'"" is 85 nt, the length of the 5' extension must be 145 nt. A size of 178 nt for pre-tRNAA"" is within the size range of C-RNAs. A size of 230 nt for pre-tRNA'"" puts pre-tRNAk" at the upper end of the C-RNA size range. Typically, C-RNAs run in the range to 120-180 nt on a 6% acrylamide, 8 M urea gel (Fig. 1, lane 2) but there is some variability. In Fig. 5F (lane  4 ) , it is clear that C-RNAs are running from 120 nt to greater in T. brucei Mitochondrion 23969  (Fig. 6, A-D, lane 4 ) does not yield any products. For the Leu D oligonucleotide, there are clearly extension products smaller than 172 nt in both total mitochondrial RNA and C-RNAs (Fig. 6D, lanes 2 and 3 ) . The shorter extensions may represent different 5' ends of the same precursor tRNA, 5' ends of isoacceptor pre-tRNA, or premature termination of the extension reaction. Extension of mitochondrial RNA with the Asn D oligonucleotide (Fig. 6C, lane 2 ) and with the Leu D oligonucleotide (Fig. 6D, lane 2 ) yields faint products that are larger than 450 nt. Corresponding products are not found in the C-RNA lane, the sizes of these RNAs do not correlate with the sizes of the RNAs detected on the Northern blot, and these RNAs are clearly not major products. These extension products are probably due to aberrant priming of mitochondrial RNA and do not reflect priming of precursor tRNA. Extension of cytosolic RNA with the Asn D oligonucleotide (Fig. 6C, lane 5 ) and with the Leu D oligonucleotide (Fig. 6D, lane 5)   brucei Mitochondrion primer extension analysis, using oligonucleotides based on mature tRNA gene sequences, detect precursor tRNAs similar in size to C-RNAs. The localization of precursor tRNAs in the mitochondrion has important implications in the transport of RNA into mitochondria and suggests that mature mitochondrial tRNAs are not the import substrates.
To characterize C-RNAs, we used two tRNA-processing enzymes, tRNA nucleotidyltransferase and RNase P, and determined if C-RNAs are suitable substrates for these enzymes in in vitro assays. Nucleotidyltransferase acts on precursor tRNA, whose 3' trailer sequence has been removed by an exonucleolytic or an endonucleolytic activity, and catalyzes the addition and repair of CCA to form the mature 3' terminus (Deutscher, 1982(Deutscher, , 1984. Nucleotidyltransferase activity has been found in all prokaryotic and eukaryotic organisms examined, and in mitochondria and chloroplasts. Purified nucleotidyltransferases function equally well with tRNA substrates from other species and are highly specific for tRNA substrates. Purified yeast nucleotidyltransferase labels tRNAs and C-RNAs in total T. brucei mitochondrial RNA. Labeling of C-RNAs, which are larger in size and much lower in abundance than tRNAs, with nucleotidyltransferase is evidence that C-RNAs are precursor tRNAs with mature 3' ends. The post-transcriptional labeling of tRNAs and C-RNAs in isolated mitochondria is evidence that a nucleotidyltransferase activity is present in the mitochondrion of T. brucei. RNase P is the enzyme responsible for the final 5' end processing of precursor tRNA (Altman, 1989(Altman, , 1990Pace and Smith, 1990). RNase P makes a single endonucleolytic cleavage in precursor tRNA to generate a mature 5' terminus. Like nucleotidyltransferase, RNase P activity has been detected in all organisms examined and RNase P cleaves precursor tRNA across species lines. Substrate recognition by RNase P is highly specific and is dependent upon the recognition of portions of mature tRNA within precursor tRNA. A subpopulation of C-RNAs, lower band C-RNAs, are recognized as precursor tRNAs by E. coli RNase P and yeast mitochondrial RNase P. The ability of lower band C-RNAs to function as substrates for RNase P identifies them as precursor tRNAs with a 5' leader sequence. I n vitro processing of RNA to mature tRNA has been used to identify the RNA substrate as precursor tRNA (Martin et al., 1985). The detection of an RNase P activity in T. brucei mitochondrial extract indicates that the in vitro processing of C-RNAs with RNase P has an in uiuo correlate.
Nucleotidyltransferase defines the 3' end of C-RNAs as mature tRNAs. RNase P defines the 5' end of C-RNAs as precursor tRNAs with a 5' leader sequence. Cleavage of C-RNAs by RNase P produces mature tRNAs indicating that the order of processing of T. brucei mitochondrial precursor tRNAs is maturation of the 3' end followed by RNase P cleavage to produce the mature 5' end. From studies in other species, it has been shown that the order of the final cleavage and addition reactions that produce the mature 5' and 3' termini of tRNA is not conserved (Delp et al., 1991;O'Connor and Peebles, 1991;Marchfelder et al., 1990) and may even be unordered within a species (Hanic-Joyce and Gray, 1990).
Northern blot hybridization identified two mitochondrial asparagine precursor tRNAs and two mitochondrial leucine precursor tRNAs. Primer extension analysis identified an asparagine precursor whose predicted size is 178 nt and one in lanes 2 and 3. 11, the extension products from the Leu D oligonucleotide run on an 8% acrylamide, 8 M urea gel. A 5' extension product of 172 nt is shown in lanes 2 and 3. Cartoons showing the predicted structure of the RNA that was the substrate for extension are shown. major leucine precursor whose predicted size is 230 nt. The shorter extensions of pre-tRNAL"" are likely to represent premature termination of the extension reaction or 5' ends of isoacceptor pre-tRNAs. The detection of two precursors by Northern blot hybridization and the detection of one precursor by primer extension suggests that the larger precursor detected on the Northern blot contains additional sequence at the 3' end. If, however, the leucine primer extension data is interpreted as indicating multiple 5' ends, then the difference in size of the two precursors detected on the Northern blot is predicted to be in the length of their 5' ends. The unusual mobility of upper band C-RNAs in acrylamide gels suggests another possible interpretation of the data. The larger precursor detected on the Northern blots may correspond to upper band C-RNAs, the smaller precursor to lower band C-RNAs. The difference between them may therefore reflect a structural difference rather than a difference in size.
The identification of precursor tRNAs within the mitochondrion indicates that the substrate for import is a precursor tRNA. It has been suggested that one possible mechanism for the import of nuclear encoded mitochondrial tRNA is via the corresponding amino acyl tRNA synthetase (Suyama and Hamada, 1978). Synthetases interact with mature tRNAs and catalyze the formation of an amino acyl bond between a tRNA and its cognate amino acid. In all systems studied, the synthetases are nuclear encoded and imported into the mitochondrion (Gopinathan, 1990). It was suggested that they may provide a means of targeting tRNA and/or of transporting tRNA across the mitochondrial membrane. However, our finding that the mitochondrial tRNAs are imported as precursors makes it unlikely that import is via interaction with the synthetases.
Primer extension analysis and RNase P sensitivity indicate that mitochondrial precursor tRNAs have 5' leader sequences. In the case of tRNAAs" and tRNAh", the length of the 5' extension is 102 and 145 nt, respectively. However, the length of the 5' leader detectable in the mitochondrion may not reflect the length of the leader before import. There may be a processing event that occurs concurrently with import analogous to the cleavage of the signal peptide upon import of mitochondrial precursor proteins (Hartl and Neupert, 1990;Baker and Schatz, 1991). In addition, there may be 5'-endtrimming within the mitochondrion that we are not detecting. While it is clear that the mitochondrial precursor tRNAs that are substrates for import have a 5' leader, the nature of their 3' termini is unclear. If the larger of the two precursor tRNA bands detected on the Northern blots is due to a 3' trailer sequence on the precursor tRNAs, then the imported precursor tRNAs must also have a 3' trailer. However, if the larger band is due to a longer 5' extension or to a different structure, then the 3' end of the mitochondrial precursor tRNAs may be mature. However, characterizing the 3' end of the precursor tRNAs within the mitochondrion as mature does not mean that import competent precursors also have mature 3' termini. All of the activities which process tRNA transcripts from nuclear genes have been localized to the nucleus in mammalian systems (Melton et al., 1980;Nishikura and De Robertis, 1981;DeRobertis et al., 1981). The finding that precursor tRNAs exist in the mitochondrion of T. brucei when their genes are nuclear encoded indicates that the transcripts targeted to the mitochondrion have escaped processing in the nucleus and have been exported from the nucleus as precursor tRNAs. In other systems, export of tRNA seems to be dependent upon processing of the precursor to mature tRNA as well as upon certain conserved sequences in the D and T stem-loop regions (Tobian et al., 1985;Zasloff et al., 1982;Zasloff, 1983). Perhaps the targeting of mitochondrial precursor tRNA begins in the nucleus when a protein binds to the pre-tRNA, protects the transcript from processing, and targets it for export from the nucleus and for import into the mitochondrion.