Purification and Characterization of MAR1

A relatively thermostable 22-kDa endoribonuclease (MAR1) was purified more than 10,000-fold from a mitochondrial extract of Leishmania tarentolae and the gene cloned. The purified nuclease has a K m of 100–145 ± 33 nm and a V max of 1.8–2.9 ± 2 nmol/min, depending on the RNA substrate, and yields a 3′-OH and a 5′-phosphate. Cleavage was limited to several specific sites in the substrate RNAs tested, but cleavage of pre-edited RNAs was generally independent of the addition of cognate guide RNA. The MAR1gene was expressed in Escherichia coli or in L. tarentolae cells, and the recombinant protein was affinity-purified. The cleavage specificity of the recombinant enzyme from L. tarentolae was identical to that of the native enzyme. The single copy MAR1 gene maps to an 820-kilobase pair chromosome and contains an open reading frame of 579 nucleotides. The 18-amino acid N-terminal sequence shows characteristics of an uncleaved mitochondrial targeting sequence. Data base searching revealed two homologues of MAR1 corresponding to unidentified open reading frames in Caenorhabditis elegans(GenBankTM accession number Z69637) andArchaeoglobus fulgidus (GenBankTM accession number AE000943). The function of MAR1 in mitochondrial RNA metabolism in L. tarentolae remains to be determined.

The processing of mitochondrial RNAs shows great variation between species. In mammalian mitochondria, both DNA strands are completely transcribed, and the transcripts are processed by the excision of interspersed tRNAs (1,2). In Saccharomyces cerevisiae mitochondria, 5Ј end maturation of tRNAs is carried out by a mitochondrial RNase P (3), and RNA degradation may involve a 3Ј-5Ј exonuclease activity (4). The 3Ј end processing of the cytochrome b (Cyb) 1 mRNA in yeast is mediated by the product of the nuclear gene CBT1 (5). In plant mitochondria, inverted repeats at the 3Ј ends of protein-coding genes serve as processing signals for 3Ј end maturation (6), and RNase Z has been identified as the nuclease responsible for the specific processing of the 3Ј ends of tRNAs (7).
In the kinetoplast mitochondrion of the trypanosomatid protozoa, Leishmania tarentolae and Trypanosoma brucei, the maxicircle DNA encodes two rRNA genes and 18 potential protein-coding genes, 12 of which are cryptogenes whose transcripts are edited by the insertion and deletion of uridine residues usually within coding regions (8 -11). RNA editing reactions appear to be initiated by specific cleavages of the preedited mRNAs, mediated by base pairing with specific cognate guide RNAs (gRNAs) (12,13). Little is known about the processing and turnover of mitochondrial RNAs in these cells, either in terms of specific cis-acting signals or enzymatic activities. Three different endoribonuclease activities, separable by sedimentation or anion exchange chromatography, have been identified in a mitochondrial extract from T. brucei (14). One of these activities, which sedimented at 20 S in glycerol gradients, exhibited a gRNA-dependent cleavage at the first mismatch upstream of a duplex RNA region (14 -16), precisely as predicted by the enzyme cascade model for RNA editing (12). Another activity, which sedimented at 15 S and was independent of added gRNA for cleavage, might correspond to an endoribonuclease activity that has been described previously in crude mitochondrial extracts from both L. tarentolae and T. brucei (17,18). The endoribonuclease activities in the crude extracts were both shown to cleave pre-edited Cyb mRNAs two nucleotides upstream of the first editing site. However, the T. brucei activity had specificity for the pre-edited Cyb sequence and did not to cleave the mature edited sequence.
In addition to a gRNA-dependent editing endoribonuclease, additional nucleases are presumably necessary for kinetoplastid mitochondrial mRNA maturation. There is some evidence for polycistronic transcription of the maxicircle (16, 19 -22). Primary transcripts are then subjected to 5Ј processing and 3Ј end cleavage, followed by 3Ј polyadenylation and polyuridylylation. In addition, many of the mitochondrial protein-coding genes have overlapping 5Ј and 3Ј ends, and the maturation of such transcripts could represent an additional level of gene regulation (19). Clearly, multiple specific ribonucleases must be required for the processing and turnover of rRNAs, mRNAs, gRNAs, and tRNAs in the mitochondria of these organisms. However, to date, the only ribonuclease purified to homogeneity from kinetoplastid mitochondria is RNase H (23,24).
In this paper we describe the isolation and characterization of MAR1 (for Mitochondrial Associated Ribonuclease) from a mitochondrial extract of L. tarentolae. MAR1 was purified to homogeneity and the MAR1 gene cloned and expressed. Further biochemical characterization and genetic analysis should help elucidate the role of this nuclease in the processing of RNAs within the mitochondrion of Leishmania. * This work was supported in part by Grant AI09102 from the National Institutes of Health (to L. S.). 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF083881.

EXPERIMENTAL PROCEDURES
Strains and Culture Conditions-L. tarentolae UC strain cells were grown in brain heart infusion (Difco) supplemented with hemin (1.6 M) (Sigma). For small-scale mitochondrial preparations (1-2 liters of culture or less), cells were grown with rotation in a Cell Production Roller Apparatus (Bellco) at 27°C. For large-scale mitochondrial preparations (15 liters), cells were grown in a BioFlo IV fermentor (New Brunswick Scientific). The cells were harvested with a Masterflex tangential filter apparatus (Millipore). For purification of mitochondria, cells (1.5-1.8 ϫ 10 8 cells/ml) were harvested by centrifugation at 5000 ϫ g for 10 min at 4°C. Cells were lysed in hypotonic Tris-EDTA, and the mitochondria were purified by isopycnic flotation in Renografin gradients as described previously (25). The mitochondrial fractions had less than 5% cytosolic contamination. Mitochondria were stored at Ϫ70°C in mitochondrial storage-breakage buffer (MSB) containing 50 mM HEPES (pH 7.5), 50 mM KCl, and 10% glycerol. Protein assays were performed using the BCA assay (Pierce).
Extract Preparation-Renografin-purified mitochondria (1.0 g/ml, wet weight) were thawed on ice and solubilized by adding MSB buffer containing CHAPS and ammonium acetate (10 and 500 mM final concentration, respectively). A protease inhibitor mixture (0.5 g/ml leupeptin, 1 g/ml pepstatin, 0.01 g/ml chymostatin, 0.1 mM benzamidine, 0.1 mM phenylmethylsulfonyl fluoride) was also added to the mitochondrial extract, which was kept at 4°C for 30 min and then clarified at 12,000 ϫ g for 20 min. The extract was then heated at 55°C for 20 min. This heating step causes a large percentage of the proteins to coagulate while MAR1 remains in solution. The heated extract was further clarified by centrifugation at 100,000 ϫ g in a Beckman TLA-100.4 fixed angle rotor for 1 h in a Beckman Optima tabletop centrifuge. The supernatant (S-100) was saved and the pellet discarded. The S-100 (100 ml) was dialyzed against 4 liters of MSB buffer and either stored at Ϫ70°C or immediately used for MAR1 purification and analysis. Storage at Ϫ70°C for long periods of time (1-3 months) produced no loss of nuclease activity.
MAR1 Purification-The S-100 supernatant (in MSB buffer) was loaded onto a Mono Q HR5/5 (anion exchange) column (Amersham Pharmacia Biotech) at a flow rate of 0.5 ml/min at 4°C. After loading, the column was extensively washed with 30 ml of MSB containing 1 mM CHAPS (QA buffer) until no more detectable protein eluted. The column was developed with a linear gradient of 20 ml of QB buffer (QA buffer containing 1 M KCl). Individual fractions (0.5 ml) were assayed for endoribonuclease activity. Fractions containing the peak of nuclease activity were pooled, concentrated with an Ultra-Free 15 concentrator (Millipore), and loaded onto a Superose 12 column (Amersham Pharmacia Biotech) (equilibrated with 0.1 M triethanolamine (pH 8.0)). The Superose 12 column was developed with 30 ml of 0.1 M triethanolamine (pH 8.0) at a flow rate of 0.3 ml/min. Fractions containing the peak of nuclease activity were pooled and loaded onto a Mono P HR5/5 column (Amersham Pharmacia Biotech) at a flow rate of 0.5 ml/min at 25°C. Proteins bound to the Mono P column were eluted with 20 ml of Polybuffer 74 (Amersham Pharmacia Biotech) (pH 5.0), allowing a pH gradient to form in the column. Individual fractions (0.5 ml) were assayed for endoribonuclease activity and peak fractions pooled and concentrated with an Ultra-Free 15 concentrator. The Mono P-pure MAR1 was gravity-loaded onto a PD-10 column (Amersham Pharmacia Biotech) equilibrated with MSB buffer and then eluted in MSB buffer containing 10% glycerol and stored at Ϫ70°C. Peak fractions from the different chromatographic steps were electrophoretically separated in 10% Tricine/SDS gels and silver-stained (Bio-Rad).
In Vitro Transcription and Labeling of Synthetic RNA-Cyb (fully edited and pre-edited) and ND7 (pre-edited) mRNAs were transcribed using T7 RNA polymerase and plasmids containing the templates for the various transcripts under the control of a T7 promoter. Plasmids were linearized by restriction digestion, the linearized plasmids were incubated in an in vitro transcription mixture containing T7 RNA polymerase, nucleotides, and the appropriate buffer (24). The resulting synthetic RNAs (CybNE, CybFE, and ND7) were 3Ј end-labeled by ligation of [␣-32 P]pCp using T4 RNA ligase (Life Technologies, Inc.). Alternatively, the mRNAs were dephosphorylated with calf intestinal alkaline phosphatase in calf intestinal alkaline phosphatase buffer as described by the manufacturer (Life Technologies, Inc.) and 5Ј endlabeled by incubation with [␥-32 P]ATP and T4 polynucleotide kinase for 1 h at 37°C, in T4 polynucleotide kinase buffer (Life Technologies, Inc.). Once labeled, the RNAs were gel-purified by electrophoresis in 7 M urea/8% acrylamide, and the bands were visualized by exposure of the gel to x-ray film. The RNA was recovered by elution for 10 h at 4°C in 400 l of 300 mM sodium acetate, 1 mM EDTA buffer. The labeled RNA was precipitated with 1.2 ml of ethanol at Ϫ20°C for 15 min. The pellets were washed twice with 70% ethanol, dried under vacuum, and suspended in TE buffer (10 mM Tris, 1 mM EDTA (pH 8.0)).
Enzyme Assays-Either 5Ј or 3Ј-labeled RNAs were incubated at 27°C for 30 min in 50 mM HEPES (pH 7.5), 1 mM DTT, 20 mM KCl, 10 mM MgCl 2 , 0.2 mM EDTA buffer (HKM buffer), in the presence or absence of enzyme. Samples (100 l) were extracted with 100 l of Tris/EDTA-saturated phenol/chloroform (pH 8.0). The RNA in the aqueous phase was precipitated with 1/10 sample volume (10 l) of 3 M sodium acetate (pH 5.2) and 3 volumes (300 l) of ethanol, with 1 g of glycogen as carrier. Samples were pelleted by centrifugation in an Eppendorf microcentrifuge at 14,000 ϫ g at 4°C for 30 min. The pellets were washed with 1 ml of 70% ethanol, dried under vacuum, and resuspended in denaturing buffer (8 M urea/0.5% SDS, 0.3% xylene cyanol, 0.3% bromphenol blue, 2 mM EDTA, 10 mM Tris-HCl (pH 8)). The samples were heated at 95°C for 5 min and electrophoretically separated in 7 M urea/6% (or 10%) acrylamide at 500 V for 2 h. Gels were transferred to 3MM blotting paper (Whatman), dried under vacuum at 85°C for 45 min, and then exposed to a PhosphorImager screen (Molecular Dynamics). For analysis of the initial velocity kinetic constants, increasing concentrations of 5Ј-labeled synthetic mRNA substrates were incubated as above with constant concentration of enzyme (10 ng). Reactions were carried out at 27°C for 30 min, and samples were processed as above. The velocity of conversion of full-length mRNA into cleaved products was calculated and plotted as a function of substrate concentration. These data were plotted as double-reciprocal plots, allowing for calculation of the first order rate constants (K m and V max ) for the various substrates.
Oligodeoxynucleotides-Oligodeoxynucleotide primers for PCR amplification, RT-PCR, hybridization, and primer extension assays were synthesized by standard phosphoramidite methods (Life Technologies, Inc.) and purified by electrophoresis in an 8 M urea/8% polyacrylamide gel at 250 V. Peptide Isolation and N-terminal Sequencing-The purified protein was electrophoretically separated in a 10% acrylamide/Tricine/SDS gel (Bio-Rad). The purified protein was electroblotted onto a polyvinylidene difluoride membrane and visualized by staining for 1 min with 0.1% Coomassie Blue R-250 in 50% methanol, followed by 2-min destaining with 50% methanol. The portion of the membrane containing the stained band was excised and subjected to N-terminal sequencing on an Applied Biosystem 477A automated sequencer. The following sequence was obtained: MPRLMPHYSTSKTAFLGVDLQCAG.
Nested RT-PCR Reactions-Based on the 24 N-terminal amino acids sequence obtained from the purified MAR1 protein, the degenerate oligonucleotides S-2323 (GACQLDV) and S-2255 (QLDVGLFATK), which were complementary to MAR1 mRNA, were used in a nested RT-PCR reaction to obtain the 5Ј sequence. S-2323 was used as primer for the synthesis of the first-strand cDNA in a reaction containing 1 g of poly(A) ϩ RNA from L. tarentolae; 120 pmol of S-2323 and 200 units of Superscript RT-II (Life Technologies, Inc.) in a 20-l reaction volume, according to the manufacturer's instructions. A 2-l aliquot of the cDNA reaction was used as template in a PCR reaction containing 100 nM S-2323 and the L. tarentolae spliced leader RNA (slRNA)-specific primer, S-2273. Amplification of the MAR1 5Ј sequence was performed in a GeneAmp 9000 thermocycler (Perkin-Elmer) in a reaction mixture containing 20 mM Tris-HCl (pH 8.0), 50 mM KCl, 1.5 mM MgCl 2 , 200 mM dNTPs, and 5 units of Taq polymerase, in a final volume of 100 l. The sample was denatured for 2 min at 94°C, followed by 30 cycles of denaturation at 94°C for 0.5 min, annealing at 40°C for 0.5 min, and extension at 72°C for 1 min. An aliquot of the first PCR reaction (5 l) was used as a template for a second PCR reaction containing the degenerate primer S-2255 and the slRNA-specific primer, S-2273, as described above. The PCR products were separated in a 2% agarose-TAE (0.04 M Tris acetate, 0.001 M EDTA (pH 8.0)) gel containing 0.5 g/ml ethidium bromide and visualized by UV light. A 130-bp DNA fragment was cloned into the pGEM-T Easy vector (Promega), according to the manufacturer's instructions, and transformed into Escherichia coli DH5␣ competent cells. Insert-containing plasmids from several E. coli clones were sequenced.
DNA Ligation and Colony Hybridization-DNA ligations were performed as described (Invitrogen) using the pCR 2.1-TOPO cloning kit (Invitrogen). Colony hybridization experiments were performed as described (26).
Northern Analysis-A Northern blot of 6 g of total RNA and 2 g of poly(A) ϩ RNA isolated from L. tarentolae cells was performed in a 18% formaldehyde/1.5% agarose gel (16 h, 1.7 V/cm). The RNA was transferred to a Zeta-Probe membrane (Bio-Rad), which was hybridized with either end-labeled oligonucleotide primers or the random-primed MAR1 probe ( Fig. 7A) (Prime-IT II kit, Stratagene).
DNA Sequencing and Generation of Nested Deletions-DNA sequencing of cloned inserts was performed manually by the dideoxy-termination reaction with 10 pmol of plasmid DNA template and Sequenase V.2.0 (Amersham Pharmacia Biotech), according to the manufacturer's instructions. The sequencing products were then separated in an 8 M urea/6% polyacrylamide gel and visualized by autoradiography. The L. tarentolae 3 kilobase pair genomic fragment containing the MAR1 gene sequence was cloned into the EcoRI-BamHI sites of the pBluescript II SK(ϩ) vector (Stratagene), generating pB1-MAR1. The pB1-MAR1 plasmid was linearized by digestion with KpnI and EcoRI, to generate an Exo-III-resistant restriction site and an Exo-III-sensitive site. Ten 400-bp nested deletion fragments were generated with the Erase-A-Base kit (Promega), according to the manufacturer's instructions. Five clones from each deletion time point were sequenced by the Applied Biosystems-Perkin-Elmer automated sequencer (Applied Biosystems-Perkin-Elmer). This approach ensured the sequencing of each nucleotide from at least three independent clones in both directions.
Contour-clamped Homogeneous Field Electrophoresis (CHEF)-L. tarentolae genomic DNA blocks were prepared according to Van der Ploeg (27) and separated in 1% agarose-1/2 TBE in a CHEF apparatus (28) at 200 V for 15 h with 1-min pulses, followed by electrophoresis for 9 h using 90-s pulses. The chromosomal bands were visualized by ethidium bromide staining and the gel photographed using a C-80 Epi-illumination UV Darkroom (Ultraviolet Products). The DNA was transferred to a Zeta-Probe membrane (Bio-Rad), according to the manufacturer's instructions and hybridized to the MAR1 gene randomlabeled using the Prime-IT kit (Stratagene).
Mapping the 5Ј and 3Ј Terminus of the Mature MAR1 mRNA-The 5Ј terminus of the MAR1 mRNA was defined by RT-PCR in a reaction containing the oligonucleotides S-2474, which is antisense to the MAR1 mRNA, and S-2273, which corresponds to a part of the spliced leader sequence. The PCR products were separated in 2% agarose-TAE and cloned into pCR 2.1-TOPO (Invitrogen), and transformed into competent E. coli DH5␣ cells.
The mapping of the 3Ј-untranslated sequence of the MAR1 transcript was performed by a modification of the procedure described above. cDNA synthesis from poly(A) ϩ RNA was performed using an oligo(dT) primer, S-2271, and Superscript II RT (Life Technologies, Inc.). The cDNA mixture was incubated for 30 min at 25°C and then for 1 h at 45°C. An aliquot of 5 l of the cDNA synthesis reaction was PCRamplified with the primers, S-2272 and S-2356, in a standard PCR reaction mixture (Promega). The sample was subjected to 2-min denaturation at 94°C followed by 30 cycles of denaturation at 94°C for 0.5 min, annealing at 60°C for 0.5 min, and extension at 72°C for 2 min and a final step of 72°C for 10 min. A 5-l aliquot of the first PCR reaction was subjected to a second PCR amplification in a standard PCR reaction mixture (Promega) with the primers S-2272 and S-2630, using the same conditions as described for the first PCR amplification. The PCR products were separated in 1% agarose/TAE gel and the amplified DNA was cloned into the pCR 2.1-TOPO vector (Invitrogen).
RNA Mapping and Sequencing-5Ј-Labeled in vitro synthesized RNAs (Cyb or ND7) were incubated with sequence-specific nucleases (T1, U2, Phy M, and Bacillus cereus), as described in the RNA Sequencing Kit (Amersham Pharmacia Biotech). These reactions were used as sequencing ladders in 10% acrylamide/7 M urea gels together with MAR1 cleavage reactions with the same mRNAs. To assay for gRNA dependence of MAR1 cleavage, reactions were performed in the presence and absence of gRNAs and separated as above.
Sequence Analysis of MAR1-The available sequence data bases (GenBank TM nonredundant data base and Swissprot) were searched using the predicted polypeptide deduced from the MAR1 open reading frame. The search programs employed were BLASTP (National Center for Biotechnology Information at http://www.ncbi.nlm.nih.gov/index.html), using a Blosum62 data matrix, and BLITZ (European Bioinformatics Institute at http://www.ebi.ac.uk/searches/searches.html). The program CLUSTALX was used to align the amino acid sequence of L. tarentolae MAR1, C. elegans, Bacillus subtilis, and Archaeoglobus fulgidus. Manual refinements were performed to optimize the alignments, taking into consideration the predicted secondary structure obtained from the individual polypeptides. Predict Protein (PHDsec) (http://www.ebi.ac.uk/searches/searches.html) was used to perform the secondary structure predictions. The Monte Carlo algorithm implemented by the RDF2 program (GCG package) on a VAX 4000 computer was used to evaluate the statistical significance of the alignments. The result is shown as S.D. values (standard deviations from the mean of randomized sequences) of 200 rounds of randomization with a ktuple of 2. The isoelectric point of MAR1 was determined with the programs ISOELECTRIC (GCG package) and MultiIdent (http://expasy.hcuge.ch/ www/expasy-top.html). The N terminus amphipathic nature of MAR1 and C. elegans sequences was determined with the program HELICAL-WHEEL (GCG package).
Expression and Purification of the Recombinant MAR1 Protein and Generation of Polyclonal Antiserum-The MAR1 gene was used to express a histidine-tagged (His-tagged) version of the MAR1 protein in E. coli cells or an epitope-tagged version in L. tarentolae cells. For E. coli expression, the coding sequence of the MAR1 gene was cloned into the pQE-31 expression vector (Qiagen), which places the gene under the control of an isopropyl-␤-D-thiogalactopyranoside-responsive promoter. E. coli BL21 cells were grown in 2 ϫ YT medium, and recombinant protein expression was induced by the addition of isopropyl-␤-D-thiogalactopyranoside (2 mM final concentration) as described in the Qiagen manual. The His-tagged recombinant MAR1 protein was purified by metal-chelate chromatography (Qiagen).
Recombinant MAR1 protein expressed in E. coli was used for the production of polyclonal antiserum in rabbits (Animal Pharm Services, Inc.). A 1:10,000 dilution of serum, from the seventh cycle of immunization, was used for Western analysis.
For expression in L. tarentolae, the MAR1 gene was ligated into the pX vector (29). Prior to ligation, a PCR reaction was carried out with oligonucleotides S-2545 and S-2546, which are complementary to the 5Ј and 3Ј end, respectively, of the MAR1 coding region. These oligonucleotides were designed to insert a BamHI site at the 5Ј end and an XbaI site at the 3Ј end of the MAR1 coding sequence. In addition, the 3Ј oligomer, S-2545, carries the sequence coding for the FLAG-epitope (N-Asp Tyr Lys Asp Asp Asp Asp Lys-C) (Eastman Kodak Co.). The PCR product generated by amplification of MAR1 with oligonucleotides, S-2545 and S-2546, was digested with BamHI and XbaI and cloned between the BamHI and XbaI sites of pX, generating the plasmid pXMARFLAG. The pXMARFLAG plasmid was used to transfect L. tarentolae cells. Transformants were selected by the G418-resistant phenotype provided by the plasmid-derived marker (29). Two liters of pXMARFLAG-transformed L. tarentolae cells were grown in brain heart infusion medium containing 200 g/ml of G418 for 14 h at 27°C and harvested as described above. The pellet was suspended in 50 ml of MSB and sonicated at medium power. The extract was centrifuged at 100,000 ϫ g and loaded onto a 1-ml FLAG-affinity column (Kodak). The cleavage activity and specificity of the recombinant MAR1 were determined as described above.

RESULTS
Optimum Temperature and Divalent Cation Requirements-In crude mitochondrial extracts the optimal temperature for the MAR1 reaction is 27°C (data not shown). However, the activity was relatively resistant to heating at 55°C for 20 min, and the thermal stability was independent of the presence of RNA substrate. As shown in Fig. 1A, 70% of the activity in the crude extract remained in the supernatant after the 55°C heating step. A similar stability against thermal denaturation was observed with purified MAR1 (data not shown), indicating that this is an intrinsic characteristic of the enzyme and not due to some extraneous factor.
MAR1 has an absolute requirement for divalent cations (Fig.  1B). When the chloride salts of Mg 2ϩ , Mn 2ϩ , Zn 2ϩ , and Ca 2ϩ were used in the cleavage reaction, Mg 2ϩ (up to 10 mM) was the preferred cation, but Mn 2ϩ also could satisfy this requirement, while Zn 2ϩ and Ca 2ϩ worked to a much lesser extent.
Purification of MAR1-Due to the relative thermal stability of MAR1, heating of a 12,000 ϫ g for 20 min clarified CHAPS lysate at 55°C was used as the initial step in purification (Table I), followed by centrifugation at 100,000 ϫ g for 1 h. Chromatography of the resulting S-100 supernatant through Superose 12, Mono Q ( Fig. 2A), and Mono P (Fig. 2B) yielded a 16,000-fold enrichment of specific activity ( Table I). Analysis of the peak fractions by SDS-acrylamide gel electrophoresis and silver staining is shown in Fig. 2C. The Mono P peak showed a major 22-kDa band that represented 97% of the protein in this fraction. No changes in substrate cleavage patterns were observed between the enzyme activity in the crude extract or the Mono P fraction (data not shown).
Kinetic Analysis of MAR1 and Cleavage Specificity-To make an initial analysis of the cleavage specificity of MAR1 and also to examine the possibility that MAR1 is involved in   enzyme and 5Ј or 3Ј end-labeled substrates. As shown in Fig.  3A, two major cleavage sites were located at editing site 2 (ES2) and editing site 3 (ES3) of pre-edited ND7 mRNA. Another major cleavage site was located just upstream of editing site 7 (ES7), and five additional cleavage sites were located downstream of the pre-edited region (PER). The locations of the major cleavage sites provided no evidence for sequence specificity, but it is also apparent that the digestion is not random. Digestion with a large excess of enzyme for extended periods led to complete cleavage of the substrate to short oligomers (8 -10 nucleotides) (data not shown).
Also shown in Fig. 3A, the addition of excess cognate gRNA for block I of ND7 did not induce a cleavage at ES1 as would be predicted for a nuclease involved in the initial step of gRNAmediated editing. Instead, the cleavages at ES2 and ES3 were inhibited and the major cleavage upstream of ES7 was unaffected, as were the downstream cleavages.
In the case of the pre-edited Cyb substrate, the precise cleavage sites were not determined. However, major cleavages located at or near ES1 and at several sites within the 3Ј region of the RNA downstream of the ER were observed (Fig. 3B). A 5Ј end-labeled substrate yielded identical results as the 3Ј endlabeled substrate shown (data not shown). Use of a fully edited Cyb substrate led to the disappearance of one of the 3Ј cleav-ages, but there was no effect on the ES1 cleavage. To test the possibility that there could still be a substrate preference for pre-edited mRNAs, a kinetic analysis was carried out under initial velocity conditions. With all of the substrates tested, MAR1 cleavage activity followed Michaelis-Menten kinetics with a K m of 120 Ϯ 15 nM and a V max of 2.8 Ϯ 2 nmol/min for the pre-edited Cyb mRNA (Fig. 4B) and a K m of 100 Ϯ 10 nM and V max of 2.9 Ϯ 0.5 nmol/min for the fully edited version of the same mRNA (Fig. 4C). Similar values for K m and V max were obtained (Fig. 4A) with the pre-edited ND7 substrate. We conclude that both pre-edited and fully edited RNAs were cleaved by MAR1 with comparable efficiency.
Cloning and Sequencing of the MAR1 Gene-A 24-amino acid N-terminal sequence (MPRLMPHYSTSKTAFLGVDLQCAG) was obtained by microsequence analysis of the purified MAR1 protein. Based on this sequence, the degenerate S-2323 primer was used to synthesize the first-strand cDNA. This cDNA was then subjected to RT-PCR amplification utilizing the S-2323 3Ј primer and the slRNA-specific S-2273 5Ј primer. The product of the first PCR reaction was subsequently re-amplified in a nested PCR reaction with the degenerate S-2255 3Ј primer and the S-2273 5Ј primer. A 130-bp DNA fragment was cloned and sequenced. The sequence of the 130-bp fragment revealed three amino acid differences (MPRLMPHYSTSKTAFLCVDLQEAF) (shown in boldface letters) from that of the N-terminal sequence obtained from microsequencing. These amino acid differences are toward the C terminus of the peptide and represented ambiguous amino acid assignments. However, the internal deduced amino acid sequence upstream of primer S-2255 (STSKTA) confirmed the identity of the cloned fragment (Fig. 5).
The 130-bp PCR product hybridized with a 3-kb EcoRI-BamHI genomic fragment. The 3-kb region of EcoRI-BamHIdigested genomic DNA was cloned into the pGEM-7Zf(ϩ) vector. The MAR1-containing plasmid was identified by colony hybridization with the 5Ј-labeled MAR1-specific S-2356 oligonucleotide and the sequence of the insert DNA determined.
This sequence contains a single open reading frame of 579 nucleotides that encodes for a protein of 192 amino acids with a molecular mass of 21.6 kDa and a predicted isoelectric point of 7.38 (Fig. 5). RT-PCR of the 5Ј region of the transcript indicated the presence of two alternative splice acceptor sites for the trans-spliced 39-nt slRNA (data not shown). One site (SAS-1) is located at position Ϫ 33 from the adenylate residue of the predicted methionine translation initiation codon (boxed in Fig. 6) and is preceded by a 10 nucleotide polypyrimidine track at positions Ϫ57 to Ϫ48. The second splice acceptor site (SAS-2) is located at position Ϫ123, with an upstream polypyrimidine track at positions Ϫ157 to Ϫ149 (Fig. 5). Primer extension assays of total RNA using the 5Ј end-labeled S-2474 primer revealed that SAS-1 is used for splicing of the slRNA in 60% of the transcripts, and SAS-2 is used in the remaining 40% (data not shown). RT-PCR analysis of the 3Ј end of MAR1 revealed two polyadenylation sites, one at position ϩ1003 and another at position ϩ1143, 424 and 564 nucleotides, respectively, from the second adenylate residue of the predicted TAA termination codon. These results would predict a transcript ranging from a maximum of 1305 nucleotides to a minimum of 1075 nucleotides (not including the length of the poly(A) tail).
Recombinant MAR1 Shows Identical Specific Activity and Cleavage Specificity as Native MAR1-The MAR1 gene was expressed in E. coli cells as a His-tagged protein and purified by affinity chromatography. The specific activity of the E. coliexpressed MAR1 was 200-fold lower than that of the native enzyme (data not shown). However, expression in E. coli allowed the production of sufficient recombinant MAR1 protein to generate anti-MAR1 polyclonal antibodies in rabbits.
To circumvent the low specific activity of the bacterial expressed MAR1, an epitope-tagged (FLAG tag) version of the MAR1 protein was expressed in L. tarentolae cells (see "Experimental Procedures"). The recombinant protein was purified using an antibody-affinity column (Kodak). This expression approach yielded apparently homogeneous MAR1 protein as determined by SDS-polyacrylamide gel electrophoresis (Fig.  6A), as well as by Western analysis using either anti-MAR1 antiserum or anti-FLAG antibody (Kodak) (Fig. 6B). The anti-MAR1 antiserum recognized in a total cell lysate of L. tarentolae a single band of the identical gel mobility as the biochemically purified MAR1 and the recombinant MAR1 proteins (Fig.  6B, lane 3). The recombinant epitope-tagged protein expressed in L. tarentolae had a specific activity (160 Ϯ 1.7 nmol/min/mg) and a cleavage specificity that was indistinguishable from the native enzyme (Fig. 3C), confirming that the MAR1 gene encodes the MAR1 nuclease.
MAR1 Is a Single Copy Gene That Maps to an 820-kb Chromosome and Has a Transcript of 1.3 kb-The MAR1 gene was analyzed by Southern blot hybridization of digested genomic DNA. L. tarentolae DNA was digested with a variety of restriction enzymes and the blots hybridized with the full-length MAR1 gene (data not shown). The enzymes, EcoRI, BamHI, and HindIII, which do not cleave within MAR1, generate a single DNA fragment that hybridizes with the MAR1 probe. The enzymes, BglII, HinfI, and NcoI, which cleave once within the MAR1 open reading frame, generated two DNA bands with the MAR1 probe. BglII also cleaves at the 5Ј end of the gene leaving a short fragment that was not detected. Double digestions also generated patterns consistent with MAR1 being a single copy gene. The restriction map of MAR1 is shown in Fig. 7A.
The MAR1 gene was localized to a 820-kb chromosome of L. tarentolae by CHEF gel electrophoresis and Southern analysis (Fig. 7C). A transcript of 1.3 kb is shown by the Northern analysis in Fig. 7C.
MAR1 Does Not Contain a Cleavable Mitochondrial Targeting Signal-Analysis of the first 18 amino acids at the Nterminal of MAR1 revealed characteristics compatible with other signal sequences reported in the literature. The presence of Arg, Leu, Ser, and Ala, the presence of at least one Thr and one Lys, and the absence of Glu and Asp residues are all consistent with this serving as a mitochondrial targeting sequence (Fig. 5). Moreover, the N-terminal 18 amino acids of MAR1 show the characteristic amphipathic pattern of known importation signal sequences (data not shown). Three genes were identified as possible homologues of MAR1: C. elegans (GenBank TM accession number Z69637), B. subtilis (GenBank TM accession number P37532), and Archaeoglobus fulgidus (GenBank TM accession number AE000943). These sequences are unidentified open reading frames. The alignments to the MAR1 sequence are shown in Fig. 8, A and B. The amino acid alignments showed a high degree of similarity of the MAR1 sequence with the C. elegans and A. fulgidus sequences (Fig. 8A). Moreover, the presence of an amphipathic helix at the N terminus of C. elegans and its absence in the A. fulgidus is consistent with this N terminus sequence representing a mitochondrial importation signal in C. elegans as well. The statistical significance of the alignments are shown in Fig.  8C; these values suggest that the alignments of the MAR1 sequence with the C. elegans and the A. fulgidus are highly significant. The alignment of the MAR1 sequence with the B. subtilis sequence is limited to the short region shown boxed in Fig. 8B, but the statistical significance of this alignment is high. The B. subtilis sequence is a homologue of the isochorismatase gene family, but a comparison of the MAR1, C. elegans, and A. fulgidus polypeptide sequences with several members of the isochorismatase gene family did not provide evidence that these genes belong to that family (data not shown).

DISCUSSION
MAR1 is the first mitochondrial endoribonuclease purified from a kinetoplastid protozoan. Isolation of this enzyme was facilitated by its high relative thermal stability, which allowed the use of an initial heating step to remove a large amount of protein prior to chromatography on three columns. The puri- FIG. 8. Alignment of MAR1 amino acid sequence with data base sequences. A, the predicted polypeptide sequence of MAR1 is aligned with the C. elegans (GenBank TM accession number Z69637) and A. fulgidus (GenBank TM accession number AE000943) sequences. The predicted N terminus signal sequences of MAR1 and the C. elegans sequence are underlined. The consensus of the alignment is indicated. "*" indicates amino acid identities; ":" and "." indicate conserved substitutions. The predicted secondary structure fold (above 90% certainty) utilized as a guide to refine the sequence alignment is shown: H, ␣-helices; L, loops, and B, ␤-sheets. B, domain of the multiple sequence alignment, including the isochorismatase sequence from B. subtilis (GenBank TM accession number P37532) (boxed region in A). C, table of S.D. values resulting from a Monte Carlo statistical analysis of the different sequences (Ͼ5 S.D. units is considered a significant match). fied MAR1 has an absolute requirement for divalent cations and migrated in an SDS acrylamide gel as a 22-kDa band. This enzyme may correspond to previously described mitochondrial endoribonucleases from L. tarentolae and T. brucei (17,18). The L. tarentolae nuclease activity had an estimated molecular mass between 10 and 30 kDa, had an absolute divalent cation requirement, and showed a major cleavage site of pre-edited Cyb mRNA 2 nucleotides upstream of editing site 1 in addition to several other cleavages throughout the PER (18). However, the L. tarentolae activity also showed a stimulation by heparin or by digestion of the crude lysate with proteinase K and an inhibition by adenylate nucleotides or GTP. The T. brucei nuclease activity showed a relatively high thermal stability, had a specificity for sites within the PER's of Cyb, cytochrome oxidase II, and cytochrome oxidase III mRNA substrates, and did not cleave within the fully edited regions of the same RNAs (17). Both activities were only characterized from crude mitochondrial lysates.
Purified MAR1 cleaved a synthetic pre-edited Cyb substrate at approximately the same location as the previously described activities, but showed additional cleavages 5Ј and 3Ј of the PER. The MAR1 enzyme also did not distinguish between pre-edited and fully edited Cyb mRNA substrates and was unaffected by the addition of heparin and was sensitive to proteinase K (data not shown). Although these differences could mean that MAR1 and the previously described nucleases are different enzymes, they could also reflect structural differences between the Cyb RNA substrates used in the cleavage reactions and could be due to the fact that the previous nuclease activities were detected and characterized in crude mitochondrial lysates, whereas this study was performed both with a highly purified and a recombinant enzyme.
The demonstration that the affinity-purified recombinant FLAG-tagged MAR1 protein expressed in L. tarentolae exhibited nearly identical specific activity and cleavage specificity to that of the native enzyme provides definitive evidence that the MAR1 gene encodes the MAR1 enzyme.
The cleavage of the Cyb and ND7 pre-edited mRNA substrates by MAR1 occurred in a gRNA-independent fashion. Interestingly, when increasing concentrations of a cognate gRNA were annealed to the ND7 substrate mRNA prior to the cleavage reaction, an inhibition of the cleavage adjacent to the first editing site was observed. The presence of several gRNAindependent endoribonucleases in a mitochondrial extract from T. brucei has been described previously (14,15,31). One of these activities co-sedimented with other editing activities in glycerol gradients, but was separable from the gRNA-dependent nuclease believed to be the editing-specific nuclease (14). This activity required DTT to function in vitro and cleaved pre-edited mRNAs in the PER. MAR1 shows a similar cleavage and fractionation behavior as the DTT-requiring enzyme, but has no requirement for DTT (data not shown).
The lack of specificity for pre-edited mRNAs and the lack of a gRNA-dependent cleavage at ES1 both suggest that MAR1 may not be involved in RNA editing. However, it should be pointed out that additional, yet undetermined, specificity factors could confer gRNA-dependence to MAR1 and this must remain an open question. Also, an in vitro RNA editing-like activity independent of gRNA but dependent on the secondary structure of the mRNA substrate has been described in mitochondrial extracts from L. tarentolae (32,33), and an involvement of MAR1 in this process still remains a possibility.
MAR1 is a mitochondrial protein that lacks a cleavable Nterminal mitochondrial targeting sequence. We presented evi-dence that the 18 N-terminal amino acids represent a noncleaved signal sequence, but this must be confirmed by direct experimentation. Cytochrome c 1 is the only other known example of a trypanosome protein targeted to the mitochondrion without a cleavable presequence (34). The MAR1 protein sequence is fairly conserved in evolution, as homologues were found in a eukaryote and an archaebacterium, and a conserved motif was found in a eubacterial protein. The homologies, however, could not be used to determine a function, as the sequences are unidentified reading frames.
As stated above, the possibility of MAR1 being involved in RNA editing remains open, but it is equally likely that MAR1 could be involved in mitochondrial RNA turnover. The expression of an epitope-tagged MAR1 in E. coli and in L. tarentolae has permitted the generation of an anti-MAR1 antiserum. Use of this immune reagent together with a detailed kinetic characterization of the cleavage reaction should help answer further questions about the specificity, substrate recognition, and mechanism of action of this nuclease.