Partitioning of the Nuclear and Mitochondrial tRNA 3′-End Processing Activities between Two different Proteins in Schizosaccharomyces pombe *

Background: Most eukaryotes contain only one tRNase Z gene involved in both nuclear and organellar tRNA 3′-end maturation. Results: Schizosaccharomyces pombe has two tRNase Z genes required for nuclear and mitochondrial tRNA 3′-end processing, respectively. Conclusion: The evolution of two tRNase Z genes and their differential expression in fission yeast may avoid toxic off-target effects. Significance: The results advance our understanding of tRNA 3′-end maturation. tRNase Z is an essential endonuclease responsible for tRNA 3′-end maturation. tRNase Z exists in a short form (tRNase ZS) and a long form (tRNase ZL). Prokaryotes have only tRNase ZS, whereas eukaryotes can have both forms of tRNase Z. Most eukaryotes characterized thus far, including Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, and humans, contain only one tRNase ZL gene encoding both nuclear and mitochondrial forms of tRNase ZL. In contrast, Schizosaccharomyces pombe contains two essential tRNase ZL genes (trz1 and trz2) encoding two tRNase ZL proteins, which are targeted to the nucleus and mitochondria, respectively. Trz1 protein levels are notably higher than Trz2 protein levels. Here, using temperature-sensitive mutants of trz1 and trz2, we provide in vivo evidence that trz1 and trz2 are involved in nuclear and mitochondrial tRNA 3′-end processing, respectively. In addition, trz2 is also involved in generation of the 5′-ends of other mitochondrial RNAs, whose 5′-ends coincide with the 3′-end of tRNA. Thus, our results provide a rare example showing partitioning of the nuclear and mitochondrial tRNase ZL activities between two different proteins in S. pombe. The evolution of two tRNase ZL genes and their differential expression in fission yeast may avoid toxic off-target effects.

Almost all RNA molecules are synthesized as precursors that must undergo processing and modifications to form functional RNAs (1). tRNA processing and modification steps include removal of 5Ј-and 3Ј-extensions, splicing of introns (for only a subset of nuclear pre-tRNAs), base modifications, and addition of 3Ј-CCA sequence (2)(3)(4). For most nuclear pre-tRNAs, 5Ј-end processing occurs prior to 3Ј-end processing, and binding by the conserved La protein to pre-tRNAs is required for the orderly processing of tRNA and for the endonucleolytic processing of pre-tRNA 3Ј-ends (5,6). In the absence of the yeast La protein, the 3Ј-end of nuclear tRNA is matured by exonuclease Rex1p (7,8). In contrast, the La protein is not involved in mitochondrial tRNA (mt-tRNA) 2 processing, and the order of mitochondrial pre-tRNA end processing can vary (9). Additionally, unlike nuclear tRNAs, mt-tRNAs are polycistronic. Moreover, mt-tRNAs normally deviate from the classic tRNA cloverleaf secondary structure and L-shaped tertiary structure due to high sequence and size variation in the D-and T-loops.
The metazoan mitochondrial genome (mtDNA) is typically a compact circular molecule (ϳ16 kb). Transcription of metazoan mtDNA is initiated at a single unidirectional or bidirectional promoter on each strand, and mt-tRNAs are produced from large polycistronic RNAs (for reviews, see Refs. 10 and 11). In contrast, the mtDNA of the budding yeast Saccharomyces cerevisiae is large (ϳ75-85 kb in size) and has numerous promoters (Ͼ20) that direct different levels of transcription (12)(13)(14). In the fission yeast Schizosaccharomyces pombe, the mtDNA is compact (ϳ19 kb in size) and contains three promoters (15,16).
A simple model known as the tRNA punctuation model has been proposed for animal mtRNA processing (17,18). According to this model, the secondary structures of tRNAs serve as mtRNA-processing signals. Processing of mt-tRNA liberates not only tRNA but also flanking RNAs from the precursor RNA. Thus, tRNA plays an important role in animal mitochondrial mRNA and rRNA processing. In contrast, mtRNA processing in S. cerevisiae is a complex process (19). After tRNA processing, which liberates the 5Ј-and 3Ј-ends of pre-mRNAs and pre-rRNAs, additional processing steps are required for the generation of mature mRNAs and rRNAs in S. cerevisiae (20).
It has been suggested that mtRNA processing is more similar between S. pombe and humans than between either and S. cerevisiae (19). In the mitochondrial genome of S. pombe, the 3Ј-ends of tRNAs are matched to the 5Ј-ends of most other RNA species (16,19). This suggests that the mt-tRNA 3Ј-endprocessing enzyme not only is responsible for tRNA 3Ј-end maturation but also plays a critical role in the generation of the 5Ј-ends of other mtRNA species in S. pombe. Unlike 5Ј-end maturation, the generation of the 3Ј-ends of mRNAs involves additional endonucleolytic cleavage (19).
The endonuclease tRNase Z (also called RNase Z or 3Ј-tRNase) is near-ubiquitous in all three domains of life and is believed to be responsible for the removal of a 3Ј-extension of CCA-less tRNA precursors, which is an essential step for the generation of mature CCA termini at the tRNA 3Ј-end (for reviews, see Refs. [21][22][23][24]. tRNase Zs belong to the ␤-CAPS (metallo-␤-lactamase-associated CISF/Artemis/SNM1/PSO2) subfamily of the metallo-␤-lactamase superfamily with diverse functions (25). Members of the ␤-CAPS subfamily are various nucleic acid-processing and -degrading enzymes, including cleavage and polyadenylation specificity factors CPSF73 and CPSF100, which are required for the 3Ј-end formation of mRNA, and RNase J, which is involved in the 5Ј-end maturation of rRNAs and mRNA stability in bacteria.
tRNase Z can be classified into a short form (tRNase Z S ) and a long form (tRNase Z L ). Sequence analysis suggests that tRNase Z L , which is about twice the size of the short form, arises from gene duplication of tRNase Z S (26,27). tRNase Z L is present only in eukaryotes, whereas tRNase Z S exists primarily in prokaryotes, plants, and vertebrates (27)(28)(29). Most eukaryotes examined thus far, including S. cerevisiae, Caenorhabditis elegans, and Drosophila melanogaster, contain only one tRNase Z L , which is either demonstrated or predicted to be localized in both the nucleus and mitochondria. In contrast, humans contain one tRNase Z L gene (also called ELAC2) and one tRNase Z S gene (also called ELAC1). The human tRNase Z L gene was originally identified as a candidate prostate cancer susceptibility gene by linkage analysis and positional cloning. However, it remains unknown how variations in the human tRNase Z L gene contribute to increased risk of prostate cancer. Human tRNase Z L is localized in both the nucleus and mitochondria (30 -32), whereas human tRNase Z S of unknown function is found primarily in the cytosol (30,31,33). In plants, the situation is more complex. For example, Arabidopsis thaliana has two tRNase Z L s and two tRNase Z S s. The two long forms of tRNase Z are found in both the nucleus and mitochondria and in the mitochondria, respectively, whereas the two short forms are present in the cytoplasm and chloroplasts, respectively (34).
Although different eukaryotic organisms can have different numbers and forms of tRNase Z with different subcellular localizations, it has been suggested that a single tRNase Z L is responsible for the 3Ј-end processing of both nuclear and mitochondrial pre-tRNAs (27,28,35). In support of this hypothesis, it has recently been found that Drosophila tRNase Z L is involved in both nuclear and mitochondrial pre-tRNA 3Ј-end processing (36). In addition, a role for ELAC2 in mitochondrial pre-tRNA 3Ј-end processing has recently been demonstrated in vivo (30,32). Although tRNase Z L s from some species, including S. cerevisiae (37), S. pombe (38), and A. thaliana (34), have been shown to possess tRNA 3Ј-end processing activity in vitro, their nuclear and mitochondrial tRNA 3Ј-end processing activity in vivo remains to be experimentally demonstrated.
Unlike most eukaryotes, each of all four sequenced Schizosaccharomyces species (S. pombe, Schizosaccharomyces octosporus, Schizosaccharomyces japonicus, and Schizosaccharomyces cryophilus) contains two candidate tRNase Z L s encoded by two different essential genes (27). The two proteins from each species share weak sequence similarity. We have previously demonstrated that the two S. pombe tRNase Z L s (Trz1 and Trz2) are localized to the nucleus and mitochondria, respectively, and can endonucleolytically remove the 3Ј-extension from a human nuclear pre-tRNA in vitro (38). However, compelling in vivo evidence that the nuclear and mitochondrial tRNase Z activities are separated into two distinct proteins in S. pombe is still lacking.
In this work, we analyzed the processing of tRNAs in two strains carrying the temperature-sensitive mutant allele of trz1 (trz1-1) or trz2 (trz2-1). We demonstrate that Trz1 and Trz2 are responsible for 3Ј-end processing of nuclear and mitochondrial pre-tRNAs, respectively. We also show that Trz2 plays a crucial role in 5Ј-end maturation of other types of RNAs in mitochondria.

EXPERIMENTAL PROCEDURES
S. pombe Strains, Media, and Genetic Procedures-The S. pombe strains used in this study were YHL6381 (wild-type; h ϩ leu1-32 his3-D1 ura4-D18 ade6-210), YAS56 (wild-type; h ϩ leu1-32 ura4-D18), and YZZ1 (h ϩ trz1-1 leu1-32 his3-D1 ura4-D18 ade6-210) carrying the temperature-sensitive trz1-1 allele (39) and YHD1 (h ϩ trz2-1 leu1-32 his3-D1 ura4-D18 ade6-210) carrying the temperature-sensitive trz2-1 allele. YHD1 was originally derived from DI13 (40), which displays a temperature-sensitive phenotype resulting from mutation of Ala to Val at position 623 of trz2 (which was called trz1 in that work). However, compared with the wild-type strain, DI13 grows extremely slowly in Edinburgh minimal medium even under permissive conditions, making subsequent analysis very hard. We thus constructed the same temperature-sensitive allele of trz2 (which we termed trz2-1) in the YAS56 strain background as follows. A targeting vector containing the 1.3-kb C-terminal coding sequence of trz2, its 293-bp terminator sequence, 1.5 kb of KanMX6, and the 730-bp 3Ј-flanking sequence of trz2 was linearized with SalI and used to transform YAS56. Transformants selected on yeast extract supplement (YES) medium (0.5% yeast extract, 3% (w/v) glucose, and 225 mg/liter each adenine, histidine, leucine, and uracil) containing 100 mg/liter G418 were patched onto YES medium and grown at 26°C. Patches of cells were then replica-plated onto the same plates and grown at 37°C. The trz2-1 mutant was identified by screening for colonies that grew at 26°C but not at 37°C, and the trz2-1 allele was sequenced to check for the A623V mutation. The trz2-1 cells grew faster than DI13 but still grew much more slowly than wild-type cells. Standard media and protocols for tRNA 3-End Processing genetic manipulation of fission yeast were used as described previously (41).
Northern Analysis-The trz1-1 and trz2-1 mutants and their isogenic wild-type strains were grown overnight in YES medium at 26°C. Subsequently, the cells were diluted with fresh YES medium to an A 600 of 0.2. When the cells reached an A 600 of 0.6 -0.8, they were shifted to 37°C. At the indicated times, a sample was withdrawn, and total RNA was isolated. RNA purification and Northern blot analysis were performed as described (39). Briefly, total RNA was isolated using the acid guanidinium/thiocyanate/phenol/chloroform method, separated on a 6% polyacrylamide-urea gel, and transferred to a nylon membrane. The blots were probed with 32 P-labeled oligonucleotide probes complementary to precursors or mature forms of mt-tRNAs. The sequences of oligonucleotide probes used for Northern blots are listed in Table 1.
Quantitative Real-time PCR (qRT-PCR) Measurement-Total RNA from S. pombe cells was isolated using an E.Z.N.A. TM yeast RNA kit (Omega) according to the manufacturer's instructions, followed by on-column DNase I treatment as described in the TransGen Biotech manual. The absence of genomic DNA was verified by PCR using primers specific for ain1 (␣-actinin). RNA quality was verified on a 1.2% agarose gel and spectrophotometric analysis using a NanoDrop 1000 instrument (Thermo Scientific). 1-2 g of total RNA was reverse-transcribed using random hexamers and the Transcript First-Strand cDNA Synthesis SuperMix kit (TransGen Biotech) in a 20-l total volume. qRT-PCR was performed in triplicates on a StepOne real-time PCR system (Applied Biosystems). The primers listed in Table 2 were designed at the adjacent sides of the junctions of mitochondrial tRNA-mRNA and tRNA-rRNA. The PCRs were carried out in a total volume of 20 l containing 200 nM primers, 2.5-5 ng of product from reverse transcription, 1ϫ TransStart TM Eco Green qPCR SuperMix, and 1ϫ Passive Reference Dye I. The mean threshold cycle number (C t ) of triplicate reactions was determined using StepOne v2.2.2 software. Relative -fold changes in the levels of precursor RNAs in temperature-sensitive trz1 or trz2 mutant cells relative to wild-type cells were calculated using the 2 Ϫ⌬⌬Ct method, with normalization to act1 mRNA levels. All graphs were also plotted using Origin version 8.0 (OriginLab).

RESULTS
Inactivation of trz1 Affects 3Ј-End Processing of Nuclear tRNAs-We first wanted to examine whether trz1 is directly involved in 3Ј-end processing of nuclear tRNAs. Because trz1 is essential, we used a strain carrying a temperature-sensitive allele of trz1 (trz1-1) (39) to evaluate the role of trz1 in nuclear tRNA 3Ј-end processing. We examined the processing of two well characterized nuclear tRNAs (tRNA AAC Val and tRNA CUU Lys ) in the trz1-1 and wild-type strains by Northern blotting using the intron-specific and mature oligonucleotide probes as described previously (5,39,(42)(43)(44)(45). These two tRNAs are encoded by the most abundant intron-containing tRNA genes in S. pombe, and thus, their processing intermediates can be easily detected. The intron-containing processing intermediates comprise a precursor tRNA with 5Ј-and 3Ј-extensions, a precursor with a mature 5Ј-end and 3Ј-extension, and a precursor with mature 5Ј-and 3Ј-ends. Because wild-type cells grew much more rapidly than trz1-1 cells even at permissive temperature, wild-type and mutant cells were cultured at the nonpermissive temperature of 37°C for different periods of time. The A 600 of wild-type cells at the 6-h time point was comparable to that of trz1-1 cells at the 10ϳ12-h time point. trz1-1 samples from the later time points (after 10 h) could not be analyzed because trz1-1 cells showed a very substantial loss of viability after a longer incubation at nonpermissive temperature (39).
Comparison of the processing of tRNA Val and tRNA Lys in the wild-type and trz1-1 cells grown at 37°C revealed an increased accumulation of a pre-tRNA species containing only a 3Ј-extension in the trz1-1 cells (Fig. 1), indicating that trz1 inactivation compromises nuclear tRNA 3Ј-end processing. These results parallel our previous findings (39) that overexpression of trz1 results in a reduced amount of a pre-tRNA species containing a 3Ј-extension. In contrast, inactivation of trz2 did not cause an increased accumulation of pre-tRNA species containing a mature 5Ј-end and a 3Ј-extension (data not shown). To determine whether the levels of mature tRNA Val and tRNA Lys were affected in the trz1-1 mutant, we reprobed the blots with oligonucleotides that detect the mature tRNA species. It appears that the levels of the mature species of these two tRNAs did not change appreciably in the trz1-1 mutant. We also examined the processing of intron-containing tRNA Leu and obtained similar results (data not shown).
Inactivation of trz2 Affects 3Ј-End Processing of mt-tRNAs-We next investigated whether trz2 is involved in mt-tRNA 3Ј-end processing. For this purpose, we used a temperaturesensitive mutant of trz2 (trz2-1) that permits conditional inactivation of trz2. This mutant strain harbors an Ala-to-Val substitution at position 623 of trz2 and was constructed as described under "Experimental Procedures." trz2-1 cells grew very poorly in Edinburgh minimal medium at permissive temperature relative to the isogenic wild-type cells and did not grow at nonpermissive temperature. As shown in Fig. 2, the growth defect of trz2-1 cells could be complemented by wildtype trz2 expressed from plasmid pREP81X under the control of a weaker version of the thiamine-repressible nmt1 promoter (46). Full suppression of the temperature sensitivity could be achieved only when trz2 was expressed under repressed conditions. In contrast, trz2 expressed from the pREP4X vector under the control of a strong thiamine-repressible nmt1 promoter could not rescue the temperature sensitivity of trz2-1 (data not shown). These results are consistent with our previous findings that high level expression of trz2 is toxic to the cells  (38). These results also suggest that a low level of Trz2 activity may be sufficient to function in mt-tRNA maturation.
To examine whether conditional inactivation of trz2 affects 3Ј-end processing of mt-tRNAs, we performed Northern analysis of total RNA isolated from the trz2-1 mutant and its isogenic wild-type strain before and after a temperature shift from the permissive to nonpermissive temperature. The mtDNA of S. pombe encodes 25 tRNA species, most of which are organized in six clusters interspersed between the protein-coding and rRNA genes. The intergenic spacers between mt-tRNA genes often vary in size and nucleotide sequence. We chose to examine the processing of mt-tRNA UUU Lys , mt-tRNA UCG Arg , and mt-tRNA GUG His because their processing intermediates significantly differ in length and thus could be easily distinguished by PAGE. We tested multiple probe sequences for each of these mt-tRNA precursors and mature tRNAs and found that the majority of oligonucleotide probes failed to give a detectable hybridization signal. The success of hybridization was highly dependent upon the position of the probe in the tRNA molecule. Wild-type and trz2-1 mutant cells were cultured at the nonpermissive temperature of 37°C for different periods of time because the mutant strain had a much lower growth rate than the wild-type strain even at permissive temperatures (Fig. 2).
We first examined the processing of the mt-tRNA Lys precursor containing a long 3Ј-spacer (38 nucleotides (nt)). Unlike the

tRNA 3-End Processing
other two mt-tRNA genes examined here, an rRNA gene (rnl), but not a tRNA gene, is located near the 5Ј-end of the mt-tRNA Lys gene. The probe P Lys , which was designed to detect the mt-tRNA Lys precursor containing the 3Ј-spacer sequence, is complementary to the 3Ј-end of mature tRNA Lys and the 5Ј-end of the 3Ј-spacer sequence (Fig. 3). We found that conditional inactivation of trz2 caused marked accumulation of a 111-nt band corresponding to the monomeric mt-tRNA Lys precursor containing the mature mt-tRNA Lys and the 3Ј-spacer sequence (Fig. 3). The absence of the 5Ј-spacer sequence was confirmed by reprobing the blot with a probe specific for the 5Ј-spacer sequence (data not shown). In addition to the monomeric precursor, accumulation of a dimeric precursor containing mt-tRNA Lys was also observed in the trz2-1 mutant strain, but not in the wild-type strain. A band corresponding to the 3Ј-spacer sequence removed from the monomeric mt-tRNA Lys precursor could not be detected by P Lys because the probe is only partially complementary to the spacer sequence.
When using the M Lys probe, which is complementary to the 5Ј-region of mature mt-tRNA Lys , we detected a 73-nt band that corresponds to the expected size of the mature form of mt-tRNA Lys in both the trz2-1 mutant and wild-type strains. How-ever, the level of mature mt-tRNA Lys was not significantly affected by trz2-1 inactivation.
We next examined whether inactivation of trz2 also affect 3Ј-end processing of the mt-tRNA Arg precursor with a relatively short 3Ј-spacer (9 nt). The lengths of the 5Ј-and 3Ј-spacer sequences of mt-tRNA Arg are 111 and 9 nt, respectively. This size difference allows us to distinguish between processing intermediates with the 5Ј-or 3Ј-spacer sequence using gel electrophoresis. Using the P Arg probe, which is specific for the mt-tRNA Arg precursor with the 3Ј-spacer sequence, we found that accumulation of the monomeric mt-tRNA Arg precursor containing the 3Ј-spacer sequence was increased in the wild-type strain but decreased in the trz2 mutant strain (Fig. 4). In contrast, trz2 inactivation resulted in a marked accumulation of two major bands, which potentially correspond in size to the 175-nt dimeric precursor of tRNA Arg -tRNA Ile containing the 3Ј-spacer sequence and the 285-nt dimeric precursor of tRNA Tyr -tRNA Arg containing the 3Ј-spacer sequence, respectively (Fig. 4). It is also notable that the length of the processing intermediates accumulated in the trz2 mutant tended to become progressively longer over time. As expected, the level of the mature mt-tRNA Arg was greatly reduced when trz2 was inactivated.
We next analyzed the processing of the mt-tRNA His precursor, which is flanked by spacers of 20 and 38 nt at the 5Ј-and 3Ј-ends, respectively (Fig. 5). The P His probe was designed to detect the precursor of mt-tRNA His containing the 3Ј-spacer sequence. Using this probe, we found that, at the nonpermissive temperature, wild-type cells accumulated a weak band corre- FIGURE 3. Northern blot analysis of mt-tRNA Lys from the WT and trz2-1 mutant strains. A, schematic representation of a mt-tRNA gene cluster containing the gene for mt-tRNA Lys . The sizes of the genes within the cluster and the intergenic spacers separating the two genes are shown in base pairs. The relative positions of the probes used for Northern blot analysis are indicated. Total RNA was extracted from the wild-type and trz2-1 mutant strains after a shift to various times (0 h represents the time just before temperature shift). The RNA blot was probed for mt-tRNA Lys precursors containing a 3Ј-spacer using the 32 P-labeled probe mitP Arg (B), followed by stripping and reprobing for mature mt-tRNA Lys using the 32 P-labeled probe mitM Lys (C). U1 snRNA was probed as a loading control (D). The identity of each tRNA species is shown on the right. M, single-stranded DNA markers in nucleotides. Black boxes indicate mt-tRNAs, and thick lines denote intergenic spacers. The positions of the probes are indicated by horizontal bars. Asterisks denote bands whose identities could not be determined unambiguously. These bands may be identified by, for example, RNA deep sequencing. . Northern blot analysis of mt-tRNA Arg from wild-type and trz2-1 mutant strains. A, schematic diagram showing a mt-tRNA gene cluster containing the mt-tRNA Arg gene. The RNA blot was probed sequentially with the 32 P-labeled probe mitP Arg , which is specific for mt-tRNA Arg precursors containing a 3Ј-spacer (B), and with the 32 P-labeled probe mitM Arg , which is specific for mature mt-tRNA Arg (C). The same blot was also probed for U1 snRNA as a loading control (D). See the legend to Fig. 3 for a more detailed description. M, single-stranded DNA markers in nucleotides. SEPTEMBER 20, 2013 • VOLUME 288 • NUMBER 38 sponding to the 111-nt monomeric mt-tRNA His precursor containing a 73-nt tRNA His and a 38-nt 3Ј-spacer sequence, and this band was more prominent at the late stationary phase time point (6 h). In contrast, the trz2-1 mutant displayed increased accumulation of the monomeric mt-tRNA His precursor and the dimeric tRNA His -tRNA Pro precursor containing the 3Ј-spacer sequence after shifting to the nonpermissive temperature. With the M His probe, which detects mature mt-tRNA His , we found that the level of mature mt-tRNA His was dramatically reduced in the trz2-1 mutant strain after a shift to the nonpermissive temperature (Fig. 5). In contrast, inactivation of trz1 had no effect on the processing of mt-tRNA Lys , mt-tRNA Arg , and mt-tRNA His (data not shown).

tRNA 3-End Processing
The trz2-1 Mutation Causes Increased Accumulation of Other Unprocessed mtRNAs beyond tRNAs-To examine whether the trz2-1 mutation affects 5Ј-end processing of other mtRNAs, we measured the abundance of dicistronic precursors containing upstream tRNAs by qRT-PCR (Fig. 6). In the S. pombe mitochondria, the 3Ј-ends of mt-tRNAs map precisely to the 5Ј termini of two noncoding RNAs (rns and the RNA component of mitochondrial RNase P (rnpB)) and seven mRNAs coding for cytochrome c oxidase subunits 1 and 2 (cox1 and cox2), cob, the ATPase subunits (Atp6, Atp8, Atp9), and the ribosomal protein (Rps3). We found that the levels of all mitochondrial pre-RNAs (other than pre-tRNAs) that are flanked on the 5Ј-side by a tRNA were increased in the trz2-1 mutant strain, but not in the trz1-1 mutant strain, grown at the nonpermissive temperature relative to the wild-type strain grown at the same temperature (Fig. 6). The largest increases were observed for atp8 mRNA and rns. The levels of these two RNAs increased by 18 -20-fold. These results suggest that trz2 is also involved in the maturation of the 5Ј-ends of mitochondrial mRNAs and noncoding RNAs except tRNAs, thus supporting the tRNA punctuation model for RNA processing in S. pombe mitochondria.

DISCUSSION
Previous studies from our laboratory have shown that S. pombe contains two tRNase Z L s, one located in the nucleus and the other in the mitochondria (38,39). These two proteins can remove the 3Ј-extension from a T7 RNA polymerase-synthesized nuclear pre-tRNA in vitro (38). However, it is unclear that Trz2 can process mitochondrial pre-tRNAs because mt-tRNAs are very different from nuclear tRNAs in sequence and structure. Furthermore, there is no direct in vivo evidence that Trz1 and Trz2 are involved in nuclear and mitochondrial tRNA 3Ј-end processing, respectively.
In this study, we examined the roles of Trz1 and Trz2 in the processing of nuclear and mitochondrial pre-tRNAs in vivo. We have demonstrated that inactivation of trz1 and trz2 impairs nuclear and mitochondrial tRNA 3Ј-end processing in vivo, respectively. In addition, we have shown that inactivation of trz2, but not trz1, affects the maturation of the 5Ј-end of mitochondrial mRNAs and rRNAs. Thus, in fission yeast, the nuclear and mitochondrial tRNase Z functions are partitioned between two different proteins.
It is noteworthy that the processing intermediates for different mt-tRNAs accumulated to different levels in the wild-type and tRNase Z L mutant strains. This presumably reflects a complex interplay among transcription, processing, and degradation, which varies for different mt-tRNAs.
Inactivation of tRNase Z L would be expected to cause an increase in the level of precursor species containing a 3Ј-spacer sequence and a concomitant reduction in the level of mature species. Indeed, the levels of mature mt-tRNA Arg and mt-tRNA His were dramatically reduced upon inactivation of trz2. FIGURE 5. Northern blot analysis of mt-tRNA His from wild-type and trz2-1 mutant strains. A, schematic of a cluster of mt-tRNA genes containing the mt-tRNA His gene. The RNA blot was probed with the 32 P-labeled probe mitP His to detect mt-tRNA His precursors containing a 3Ј-spacer (B) and with the 32 Plabeled probe mitM His to detect mature mt-tRNA His (C). U1 snRNA served as a loading control (D). See the legend to Fig. 3 for a more detailed description. M, single-stranded DNA markers in nucleotides. However, although the two nuclear tRNAs examined and mt-tRNA Lys exhibited defects in 3Ј-end processing in response to tRNase Z L inactivation, the production of their mature species was not detectably affected upon tRNase Z L inactivation. There are two possible explanations, not mutually exclusive, for why the levels of these mature tRNAs were unchanged. The first explanation is that tRNase Z activity is not completely abolished at the nonpermissive temperature, and low levels of residual tRNase Z activity may be sufficient for tRNA 3Ј-end processing. The second explanation is that these tRNAs may have slower decay rates. It should be noted that the S. pombe genome encodes 171 nuclear tRNAs. Although the levels of the mature nuclear tRNAs we examined seemed to be unchanged, it is likely that the levels of other mature nuclear tRNAs not examined here could have been reduced. Although these results are unexpected, similar results have been observed in tRNase Z L knockdown Drosophila. When tRNase Z L was knocked down in Drosophila, the levels of all five mature nuclear tRNAs examined and one mature mt-tRNA were not significantly altered even though the levels of the corresponding pre-tRNAs with a 3Ј-extension increased considerably (36).
Eukaryotic organisms have developed many mechanisms for targeting tRNase Z L to different cellular compartments. Metazoans appear to use alternative translation initiation sites. For example, in humans, the mitochondrial and nuclear forms of tRNase Z L can be generated from a single gene by the alternative use of two translation start codons (31,35). Because the first start codon of the human tRNase Z L gene is in a suboptimal context, some ribosomes would bypass it to initiate translation at a downstream AUG codon, thus generating essentially two identical tRNase Z L proteins, one mitochondrial and one nuclear. In contrast, in Caenorhabditis nematodes, the nuclear and mitochondrial forms of tRNase Z L are likely generated by alternative splicing of the tRNase Z L gene (35,47).
Our results provide the first example that S. pombe (and most likely, other fission yeast species) achieves dual localization of tRNase Z L to the nucleus and mitochondria by the use of two genes, which is the simplest and basic mechanism of dual targeting. This mechanism also allows for greater flexibility in the control of differential expression of similar genes. Indeed, unlike Trz1, Trz2 seems to be present in very low abundance in cells (it ranks in the bottom 5% of 3521 proteins) and is ϳ19fold less abundant than Trz1 (48). These data indicate that trz1 and trz2 are differentially expressed and that different amounts of tRNase Z proteins are required for nuclear and mitochondrial tRNA 3Ј-end processing in S. pombe. It seems likely that the differential expression of the nuclear and mitochondrial tRNase Z genes may be required to avoid the off-target effects of trz2 in S. pombe because the increased expression of trz2 leads to lethality in a dose-pendent manner (38). Efforts are under way to elucidate the mechanisms involved in differential regulation of these two tRNase Z genes. Because the nuclear and mitochondrial tRNase Z activities reside in two different genes in fission yeast (unlike in the majority of organisms examined to date), fission yeast may provide opportunities for focused studies that separate the nuclear and mitochondrial functions of tRNase Z.