Identification and characterization of human mitochondrial tryptophanyl-tRNA synthetase.

A full-length cDNA clone encoding the human mitochondrial tryptophanyl-tRNA synthetase (h(mt)TrpRS) has been identified. The deduced amino acid sequence shows high homology to both the mitochondrial tryptophanyl-tRNA synthetase ((mt)TrpRS) from Saccharomyces cerevisiae and to different eubacterial forms of tryptophanyl-tRNA synthetase (TrpRS). Using the baculovirus expression system, we have expressed and purified the protein with a carboxyl-terminal histidine tag. The purified His-tagged h(mt)TrpRS catalyzes Trp-dependent exchange of PP(i) in the PP(i)-ATP exchange assay. Expression of h(mt)TrpRS in both human and insect cells leads to high levels of h(mt)TrpRS localizing to the mitochondria, and in insect cells the first 18 amino acids constitute the mitochondrial localization signal sequence. Until now the human cytoplasmic tryptophanyl-tRNA synthetase (hTrpRS) was thought to function as the h(mt)TrpRS, possibly in the form of a splice variant. However, no mitochondrial localization signal sequence was ever detected and the present identification of a different (mt)TrpRS almost certainly rules out that possibility. The h(mt)TrpRS shows kinetic properties similar to human mitochondrial phenylalanyl-tRNA synthetase (h(mt)PheRS), and h(mt)TrpRS is not induced by interferon-gamma as is hTrpRS.

The only components of the mitochondria encoded by the mitochondrial genome include 12 S and 16 S rRNAs, 22 tRNAs, and 13 proteins: 3 cytochrome c oxidase subunits, 2 ATPase subunits, cytochrome b, and 7 components of respiratory-chain NADH dehydrogenase (1)(2)(3). Mitochondrial protein-synthesizing machinery is required for synthesis of these 13 proteins. Therefore, proteins like the mitochondrial translational release factors, mt RF1 and mt RRF (4), and the aminoacyl-tRNA synthetases (aaRS) 1 are encoded by the nuclear genome, trans-lated by cytoplasmic ribosomes, and subsequently imported into the mitochondria. Most mitochondrial targeting is mediated by an amino-terminal amino acid sequence of ϳ20 -60 amino acids, which is cleaved off after import (5).
The aaRSs are enzymes catalyzing the covalent attachment of amino acids to their cognate tRNAs. They can be divided into two different classes based on distinct structural motifs. Class I aaRSs display two common consensus sequences, HXGH and KMSKS, indicating the presence of a Rossmann fold that binds ATP (6), and the acylation occurs at the 2Ј-OH position of the ribose of the CCA end of the tRNA. The class II aaRSs are characterized by three different sequence motifs, motifs 1, 2, and 3, and the acylation occurs at the 3Ј-OH position of the ribose (6).
In 1991 we discovered that a highly interferon-␥ (IFN-␥)inducible protein (designated ␥2) identified by two-dimensional gel electrophoresis (7) was identical to human cytoplasmic tryptophanyl-tRNA synthetase (hTrpRS), and subsequently, the cDNA was cloned and characterized (8 -10). The fact that hTrpRS is induced by IFN-␥ is a unique feature of this group of housekeeping enzymes, and no satisfactory explanation for this has yet been put forward.
hTrpRS belongs to the class I aaRSs. It is an ␣ 2 dimer, and is not in a high-molecular-weight complex like most other aaRSs (11). The gene encoding hTrpRS (designated WARS) is localized to chromosome 14 (14q32.2-q32.32) (12). Tolstrup et al. (13) found evidence for alternative splicing of hTrpRS in the 5Ј-untranslated and -translated regions. Studies of subcellular localization of tryptophanyl-tRNA synthetase (TrpRS) in rat pancreas tissue and in a rat fibroblast cell line revealed that TrpRS was situated in the mitochondria and in the nucleus in addition to the cytoplasm (14). A splice variant might therefore localize to the mitochondria, like the histidinyl-tRNA synthetase from Saccharomyces cerevisiae (15). However, because the anticodon in mitochondrial tRNA Trp is UCA and not CCA, as in cytoplasmic tRNA Trp , and because the tRNA anticodon is one of the principal elements in the recognition process (11), it is more likely that a gene different from the one encoding hTrpRS encodes human mitochondrial tryptophanyl-tRNA synthetase (h mt TrpRS).
Here we report the identification of the human cDNA encoding h mt TrpRS. The amino acid sequence contains motifs characteristic of class I aminoacyl-tRNA synthetases. Recombinant h mt TrpRS is shown to be active in the PP i -ATP exchange assay, and mitochondrial localization is confirmed by immunoblotting subcellular fractionations of insect cells that express Histagged h mt TrpRS as well as by fluorescence microscopy of human 293 cells that express h mt TrpRS tagged with green fluorescence protein (GFP).  1 The abbreviations used are: aaRS, aminoacyl-tRNA synthetase; TrpRS, tryptophanyl-tRNA synthetase; hTrpRS, human cytoplasmic tryptophanyl-tRNA synthetase; mt TrpRS, mitochondrial tryptophanyl-tRNA synthetase; h mt TrpRS, human mt TrpRS; mt PheRS, mitochondrial phenylalanyl-tRNA synthetase; GFP, green fluorescence protein; IFN-␥, interferon-␥; PBS, phosphate-buffered saline; EST, expressed sequence tag; PAGE, polyacrylamide gel electrophoresis; bp, base pairs. Biotechnology Information using the sequence encoding the mt TrpRS from S. cerevisiae (accession no. M12081). The IMAGE clone was fully sequenced by the primer walking approach (16). Sequencing was performed on an Applied Biosystems 373A sequencer using a sequencing kit (ABI Prism Dye Terminator Cycle, Perkin Elmer). Each strand was sequenced at least twice.
Cloning the Full-length h mt TrpRS cDNA and Expression of Histidine-tagged h mt TrpRS in the Baculovirus Expression System-Using the sequencing results from RACE, we were able to design primers and to obtain a full-length cDNA clone encoding h mt TrpRS. The forward primer containing an EcoRI site and the missing nucleotides revealed by RACE was 5Ј-CGGAATTCATGGCGCTGCACTCAATGCGGAAAGC-GCGTGAGCGCTGGAGC-3Ј, and the reverse primer containing a 6ϫ His tag, a stop codon, and an XhoI site was 5Ј-CCGCTCGAGTTAATG-ATGATGATGATGATGTAGAAAACCCACCAATTTC-3Ј. These primers were used to amplify the IMAGE 667474 clone using a Pfu proofreading polymerase (Stratagene) thus obtaining a full-length h mt TrpRS cDNA, which was cloned into the transfer vector pFastBac1 (Life Technologies, Inc.). Preparation of recombinant bacmid DNA was performed following the instruction manual for the Bac-to-Bac™ Baculovirus Expression system (Life Technologies, Inc.). This bacmid DNA was then used to transfect Sf9 insect cells, and the h mt TrpRS recombinant baculovirus was isolated (18). The hTrpRS cDNA (8) was cloned into the EcoRI/SmaI sites of the baculovirus transfer vector pVL1392, and the recombinant hTrpRS baculovirus was isolated using the BaculoGold™ transfection method (PharMingen) following the manufacturer's instructions.
Purification of Recombinant Human Histidine-tagged mt TrpRS-HighFive™ insect cells (10 8 cells) infected with h mt TrpRS recombinant baculovirus were lysed in 1 ml of 1% (v/v) Triton X-100; 10 mM Tris-HCl, pH 7.5; 130 mM NaCl; 10 mM NaF; 10 mM sodium phosphate, pH 7.5; 10 mM sodium pyrophosphate; and 1 ϫ Protease inhibitor mixture without EDTA (Roche Molecular Biochemicals). The lysate was clarified by centrifugation for 30 min at 40.000 ϫ g at 4°C to pellet the nuclei and cellular debris. The supernatant was mixed with 1 ml of resuspended Ni-NTA agarose (Qiagen) and gently shaken for 1 h at 4°C. The mixture was then applied to a column (Bio-Rad) allowed to settle, and the supernatant was drained from the column. The column was washed four times with 10 ml of 6ϫ His wash buffer (50 mM sodium phosphate, pH 8.0; 300 mM NaCl; 10% (v/v) glycerol). 3 ml of 6ϫ His elution buffer (50 mM sodium phosphate, pH 8.0; 300 mM NaCl; 10% (v/v) glycerol; 0.3 M imidazole) was used to elute the recombinant protein in 500-l fractions. 10-l aliquots of washes and eluate, and a molecular mass marker from Sigma (pyruvate kinase, 63 kDa; fumarase, 52.5 kDa; lactic dehydrogenase, 35 kDa; triosephosphate isomerase, 32 kDa) were analyzed using 10% SDS-polyacrylamide gel electrophoresis (PAGE).
Amino Acid Sequencing-Purified recombinant His-tagged h mt -TrpRS was applied to the 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (ProBlott™) using 10 mM 3-(cyclohexylamino)propanesulfonic acid⅐NaOH, pH 8.0, with 10% methanol. Proteins were visualized using Coomassie Brilliant Blue staining, marked, destained in 50% methanol in water, and cut out of the membrane. The membrane was applied to a protein sequencer (477A, Applied Biosystems) according to the instructions of the manufacturer. A BLOTT-1-modified cycle for protein sequencing was used. The first seven amino acids of the amino-terminal end of the purified protein were sequenced.
Succinate dehydrogenase from mitochondria was measured according to Pennington (19) using succinate as substrate and p-iodonitrotetrazolium violet as electron acceptor. The reaction contained 0-to 30-l fraction, 50 mM sodium succinate, 50 mM potassium phosphate buffer (pH 7.4), 0.1% p-iodonitrotetrazolium, and 25 mM sucrose in a total volume of 100 l. After incubation for 15 min at 37°C, 100 l of 10% trichloroacetic acid was added, the colored product was extracted with 4 volumes of ethyl acetate, and the absorbance was measured at 490 nm. The time and amount of enzyme used for the assay was within the linear range. Units of the succinate dehydrogenase were calculated as dilution ϫ (A 490(sample) Ϫ A 490(control) )/15 min.
Immunoblotting-Cell fractions and the molecular mass marker from Sigma were applied to the 10% SDS-PAGE, and the gel was blotted onto the polyvinylidene difluoride membrane (Immobilon-P™, Millipore) (20). For detecting the His-tagged h mt TrpRS protein, a monoclonal mouse anti-His tag antibody (Qiagen) at 0.1 g/ml was used as the primary antibody. As a secondary antibody horseradish peroxidaseconjugated monoclonal goat-anti-mouse (Dako) was used at 0.5 g/ml. For detecting hTrpRS, a polyclonal rabbit antibody raised against hTrpRS purified from AMA cells (21) was added in a 1:20,000 dilution. Horseradish peroxidase-conjugated monoclonal goat-anti-rabbit (Dako) was used as secondary antibody at 0.06 g/ml. The blots were visualized using the enhanced chemiluminescence method (ECL; Amersham Pharmacia Biotech).
Cloning of h mt TrpRS cDNA with a GFP Tag-h mt TrpRS cDNA was cloned into the pEGFP-N2 (CLONTECH) vector using the same approach as with the pFastBac 1 vector. The same forward primer containing an EcoRI site was used, and a reverse primer containing an XmaI site, 5Ј-GCGACTCCCGGGTAGAAAACCCACCAATTTCTTC-3Ј, was designed for in-frame fusion of h mt TrpRS to GFP. Using the endonuclease restriction sites in the primers, the PCR product of h mt TrpRS cDNA was cloned into the same restriction sites in the pEGFP-N2 vector. Competent XL-1 Blue cells (Stratagene) were transformed by electroporation, and a positive clone was isolated and sequenced.
Expression of GFP-tagged h mt TrpRS in Human 293 Cells and Fluorescence Microscopy-One day prior to transfection, coverslips were incubated for 2 h with 10 g/ml poly-lysine in PBS (Sigma) and dried. Human 293 cells were seeded on coverslips in 32-mm dishes (Nunc, Denmark) in 2 ml of medium, and at 50% confluence they were transfected with 0.3 g of pEGFP-h mt TrpRS plasmid and Superfect (Qiagen), according to the manufacturer's instructions, and incubated at 37°C for 24 h. In vivo fluorescent red staining of mitochondria was achieved by incubating cells for 5 min with 25 nM MitoTracker™ Red CMXRos (Molecular Probes) in growth medium. Cells were washed with PBS, and the coverslips were mounted directly on the glass slides. Fluorescence microscopy was done using a Leica DMRB fluorescence microscope equipped with a cooled charge-coupled device camera (Sensys, Photometrics) and an automatic filter wheel (Ludl). Images were captured and analyzed using SmartCaptureVP (Digital Scientific) and IPLab software (Scanalytics).
PP i -ATP Exchange Activity Assay-The aaRS activity was tested with a modification of a Trp-dependent PP i -ATP exchange assay using [␥-32 P]ATP instead of [ 32 P]pyrophosphate: Trp ϩ ATP ϩ EnzymeT rp-adenylate⅐enzyme ϩ PP i (22). The reaction contained 4 mM ATP; 0.02 Ci of [␥-32 P]ATP; 50 mM Tris-HCl, pH 9.0; 1 mM KF; 0.02% gelatin; 10 mM MgCl 2 ; 0.1 mM L-Trp; 4 mM PP i ; and 10 l of purified enzyme in a total volume of 20 l. The last components to be added were PP i and then the purified enzyme (final concentration, 0.14 M). The samples were incubated for 30 min at room temperature and stopped by rapid cooling to 0°C. 3 l was spotted on a polyethyleneimine-cellulose chromatogram (Polygram®, Macherey-Nagel), which was developed with 1 M KH 2 PO 4 , dried, and visualized using a PhosphorImager™ (Molecular Dynamics). His-tagged hTrpRS purified from Escherichia coli, kindly provided by Bente Vestergaard, Aarhus University, Denmark, was used as a positive control. Two negative controls were included. One with in vitro translated luciferase from the control DNA in the TNT®T7/T3 Coupled Reticulocyte Lysate System (Promega), and one in which 0.1 mM Leu was added instead of Trp. In reactions with varying ATP concentrations, Trp was held constant at 2 mM. When the concentration of Trp was varied, the ATP was kept at 4 mM. Exchange reactions were measured at 1, 5, 10, 15, and 20 min, and initial velocities were calculated. Each assay was repeated four times. These data were used to make plots of the reaction velocity (V) as a function of the substrate concentration [S], and to calculate the K m and V max values for ATP and Trp. Nonlinear regression was made by fitting the data to the Michaelis-Menten equation.
RT-PCR for Determination of IFN Induction-Reverse transcription (RT)-PCR was performed using nontreated cells or AMA cells treated with IFN-␣ (500 U/ml) or IFN-␥ (100 U/ml). Total RNA was purified using the acid guanidine thiocyanate/phenol/chloroform extraction method (23). 2 g of total RNA was reverse-transcribed using the first strand cDNA synthesis kit from Amersham Pharmacia Biotech. The h mt TrpRS primers used in the PCR reactions were 5Ј-CATTCAACCTA-CAGGAATC-3Ј and 5Ј-CTTCTCTAAATGGTCCTTGTCCA-3Ј. The PCR setting was 25 cycles with denaturation at 95°C for 40 s, annealing at 56°C for 40 s, and extension at 72°C for 2 min, resulting in an 880-base pair (bp) PCR product. To compare it with IFN-␥-induced hTrpRS, the PCR was carried out with the same cDNA using the hTrpRS primers 5Ј-GCTCAAGAAGGCACTCATAGAGGTTCTGC-3Ј and 5Ј-TATTGAT-AACATACACAGGCTTACAGAGG-3Ј. This PCR was carried out with 25 cycles of denaturation at 95°C for 40 s, annealing at 61°C for 40 s, and extension at 72°C for 1 min, resulting in a 277-bp PCR product.
As a control, a PCR was performed using the same cDNA with primers in glyceraldehyde-3-phosphate dehydrogenase, 5Ј-GGTCGGAGTCAACG-GATTT-3Ј and 5Ј-CCAGCATCGCCCCACTTGA-3Ј. The condition for this PCR was as for hTrpRS except annealing was at 55°C.

RESULTS
The Nucleotide Sequence of h mt TrpRS-We searched the human EST data base for clones with similarity to the gene encoding the mt TrpRS from S. cerevisiae as described by Myers and Tzagoloff (24) and to TrpRS in different eubacteria. Several human EST clones belonging to the same tentative human consensus group (THC) appeared. One of these clones (IMAGE 667474) had been partly sequenced from the 3Ј-end and included the poly(A) signal and tail. The entire clone was sequenced, and to obtain a full-length cDNA sequence we performed RACE using total RNA from HeLa cells. The h mt TrpRS cDNA sequence could thus be extended by 25 nucleotides at the 5Ј-end. This extension resulted in two in-frame Met codons, the first at position 10 and the second at position 25. Using the Met codon at position 10 as the translation start, the open reading frame encoded a polypeptide of 360 amino acids, with a predicted molecular mass of 40.1 kDa. The complete cDNA sequence has been submitted to GenBank™ (accession no. AJ242739).
Alignment of h mt TrpRS with Different Forms TrpRSs-Using the ClustalW program, an alignment was made with h mt -TrpRS, mt TrpRS from S. cerevisiae, the putative mt TrpRS from C. elegans, and TrpRSs from the eubacteria E. coli, B. stearothermophilus, B. subtilis, and H. influenza. The sequence identity of h mt TrpRS to mt TrpRS in S. cerevisiae and C. elegans is 37% and 38%, respectively. As for mt TrpRS from S. cerevisiae and C. elegans, the h mt TrpRS also has a very high homology to the bacterial forms of TrpRS (Fig. 1). The h mt TrpRS has between 37% and 40% identical residues compared with the four eubacterial TrpRSs. Some of the well-conserved domains are the consensus sequences HXGH (HLGN for mitochondria and TIGN for eubacteria) and KMSKS responsible for the Rossmann fold.
Both the human and S. cerevisiae mt TrpRSs have aminoterminal extensions of approximately 32 amino acid residues, which are not represented in the eubacterial variants. These extensions are rich in Lys and Arg residues and thus most likely constitute the mitochondrial localization signal.
An alignment was made with h mt TrpRS and hTrpRS, which showed that h mt TrpRS aligns to the carboxyl-terminal part of hTrpRS and that the sequence identity is only 11%. The computer program MitoProt II predicts a probability of 73.5% for the human mt TrpRS to be localized to the mitochondria. The predictions are done according to the theory of Claros et al. (25), taking into account the net charge, hydrophobicity, and the ability to form ␣-helical amphiphilic structures. The MitoProt II program was run with mt TrpRS from human, S. cerevisiae, and C. elegans and hTrpRS. In comparison to h mt TrpRS, S. cerevisiae mt TrpRS is given a probability of 98.8% of being mitochondrial, whereas the mt TrpRS from C. elegans is only given a probability of 22.5%. As expected, the hTrpRS is only given a probability of 8.4%. The cleavage sites in human, S. cerevisiae, and C. elegans mt TrpRS are predicted to be between the amino acids 46 and 47, 24 and 25, and 22 and 23, respectively, whereas the cleavage site for hTrpRS is not predictable.

Localization of h mt TrpRS-To investigate the localization of h mt TrpRS we fractionated lysates of insect cells expressing
His-tagged h mt TrpRS into three different crude fractions. One fraction contained the nuclei, one contained the cytoplasm, and one the mitochondria. Immunoblots with anti-His-tag antibodies revealed that h mt TrpRS is highly expressed in the mitochondrial fraction and to a much weaker extent in the nuclear fraction ( Fig. 2A), whereas this band was completely absent from the cytoplasmic fraction. The His-tagged h mt TrpRS migrates with a molecular mass of 35 kDa, which is less than the predicted 40.9 kDa of the noncleaved precursor of h mt TrpRS with the 6 ϫ His tag. However, this correlates with the loss of the signal peptide during mitochondrial localization. The faint band of ϳ60 kDa in the cytoplasmic fraction of both h mt TrpRS and hTrpRS cells most likely corresponds to unspecific binding of the antibody.
To compare the localization of h mt TrpRS with hTrpRS, a similar immunoblot was carried out using polyclonal rabbit hTrpRS antibody. The hTrpRS antibody recognized a distinct band of ϳ53 kDa corresponding to hTrpRS in both the cytoplasmic and nuclear fractions, whereas only a weak band appeared in the mitochondrial fraction of the hTrpRS-expressing cells (Fig. 2B). No bands were detected in any of the fractions from the h mt TrpRS-expressing insect cells (Fig. 2B), which indicates that there is no cross-reactivity between the hTrpRS antibody and h mt TrpRS. Furthermore, no endogenous insect cell TrpRS could be detected by the hTrpRS antibody. To confirm the quality of the mitochondrial fractionation, we performed an assay for the activity of the succinate dehydrogenase. This enzyme catalyzes the oxidization of succinate to fumarate in the citric acid cycle in mitochondria. We found the highest values for succinate dehydrogenase activity in the mitochondrial fraction, although the fraction containing the nuclei also showed some activity (Table I). These data correlate with the weak h mt TrpRS band in the nuclear fraction of the h mt TrpRS-expressing insect cells.
Mitochondrial localization of h mt TrpRS in human cells was confirmed by fluorescence microscopy of human 293 cells trans-fected with the pEGFP-h mt TrpRS vector. As can be seen on Fig.  3A the expressed GFP-tagged h mt TrpRS is localized to subcellular compartments surrounding the nucleus. Fig. 3B shows the same cells stained red by MitoTracker dye, and it can be seen that the mitochondria co-localize with the GFP-tagged h mt TrpRS. It is noteworthy that the GFP staining is confined to the periphery of the mitochondria.
The Mitochondrial Import Signal Peptide-Recombinant carboxyl-terminally His-tagged h mt TrpRS from baculovirus was purified on a Ni-NTA column. The gel shows a single band of the expected size of ϳ35 kDa in the eluate (Fig. 4A).
The cleavage site for the signal peptide of His-tagged h mt -TrpRS was determined by amino-terminal sequencing of the recombinant protein purified from insect cells. The first seven amino-terminal amino acids were LHKGSAA. This suggests that 18 amino acids were cleaved off during the mitochondrial localization.
Activity Assay of h mt TrpRS-To test the activity of the purified His-tagged h mt TrpRS, we used a PP i -ATP exchange assay with [␥-32 P]ATP, which has the advantage that mitochondrial tRNA Trp is not needed in the reaction. The activity is demonstrated by autoradiography of a polyethyleneimine-cellulose chromatograph (Fig. 4B) of an exchange assay containing 0.14 M of the purified enzyme. Recombinant hTrpRS was used as a positive control, whereas Leu and luciferase was used as neg-ative controls. Formation of radioactive PP i is Trp-dependent both for h mt TrpRS and hTrpRS.
The kinetic parameters V max , K m , k cat , and k cat /K m for the interaction of h mt TrpRS with Trp and ATP were determined by measuring the initial rate of the PP i -ATP exchange followed by nonlinear regression fitting to the Michaelis-Menten equation. With a constant saturating concentration of 4 mM ATP, the concentration of Trp was varied between 10 and 800 M (Fig.  5A). The K m value of h mt TrpRS is comparable to the K m value for E. coli TrpRS (26), whereas the value for B. subtilis is much higher (27) and the value for hTrpRS is much lower (28) (Table  II). In addition, the k cat value is almost 100-fold lower than the k cat for E. coli. A similar pattern is seen with the k cat values for E. coli phenylalanyl-tRNA synthetase (PheRS) and h mt PheRS (29,30).
At a constant Trp concentration of 2 mM, the ATP concentration was varied between 0.04 and 4 mM (Fig. 5B). This K m value for ATP is more than twice as high as for the K m values of E. coli and B. subtilis TrpRS, as well as the value for hTrpRS (Table II). Again, the k cat value for h mt TrpRS is almost 100-fold lower than the k cat for E. coli, and this pattern is also seen with k cat values for h mt PheRS and E. coli PheRS. There seems to be a tendency for higher K m values for ATP in mitochondrial aaRS than for eubacteria and cytoplasmic aaRS, which might be explained by the higher concentration of ATP in mitochondria.
The Effect of Interferon on Expression of h mt TrpRS-To get an indication of whether the h mt TrpRS, like the cytoplasmic hTrpRS, is induced by interferons, RT-PCR with primers in the h mt TrpRS gene was carried out using cDNA from AMA cells treated with IFN-␣ or IFN-␥ compared with nontreated cells. For comparison, PCR with primers in the hTrpRS gene was performed on the same cDNA, and as a control for RNA integrity and comparable RT reactions, PCR with primers in the   glyceraldehyde-3-phosphate dehydrogenase gene was also performed on the same cDNA. It can be seen in Fig. 6 that neither IFN-␣ nor IFN-␥ induced the transcription of the h mt TrpRS gene. This is in contrast to the hTrpRS gene, which is highly induced by IFN-␥ as expected. DISCUSSION We have identified a human cDNA encoding the h mt TrpRS, which is different from the human gene encoding hTrpRS. A BLAST search with mt TrpRS from S. cerevisiae against the human EST data base identified several different EST clones with a very high homology to the gene encoding mt TrpRS from S. cerevisiae. The high homology of several yet unidentified human EST clones to the mt TrpRS gene from S. cerevisiae gave us reason to believe that they could encode the h mt TrpRS. Complete sequencing of one of these clones (IMAGE 667474) and performing 5Ј RACE resulted in an open reading frame of 1080 bp. The entire cDNA was cloned and expressed with a carboxyl-terminal His-tag in the baculovirus expression system. Activity assays, together with kinetic studies and localization experiments using both GFP-tagged h mt TrpRS in the human cell line 293 and subcellular fractionations of h mt -TrpRS-expressing insect cells, identified with certainty that our cDNA clone was human mitochondrial TrpRS.
In addition, we have recently used radiation hybrids to locate the gene encoding the h mt TrpRS (designated WARS2) to chromosome 1 (1p13.3-p13.1) (31), thus confirming the human origin of the IMAGE 667474 EST clone.
The existence of h mt TrpRS makes it unlikely that a hTrpRS splice variant should be localized to the mitochondria to provide tryptophanyl-tRNA synthetase activity to the mitochondria. Using a monoclonal antibody labeled with gold particles against a conserved epitope of TrpRS, Popenko et al. (14) showed that rodent TrpRS localized not only to the cytoplasm but also to the mitochondria and to the chromatin region in the nucleus. In addition to immunostaining of ultrathin sections of rat pancreas tissue and rat fibroblasts, E. coli and Methanococcus halophilus cells were also stained. The distribution of TrpRS in E. coli and M. halophilus was mainly concentrated in the cytoplasm, but some was also detected adjacent to the nucleoid border. This detection was feasible because the antibody could cross-react with TrpRS from prokaryotic, archaebacterial, and eukaryotic cells (32). This particular antibody most likely also recognizes mt TrpRS in eukaryotes, taking into account its high homology to prokaryotic TrpRS. However, this does not explain why TrpRS also was found localized to the  nucleus. In addition, our immunoblots, using antibodies against hTrpRS not cross-reacting with h mt TrpRS, showed hTrpRS mainly in nucleic and cytoplasmic fractions of hTrpRSexpressing insect cells.
The GFP-tagged h mt TrpRS seems to localize to the mitochondrial periphery. Similar localization is observed when the GFP derivative yellow fluorescence protein is anchored in the outer mitochondrial membrane using the signal peptide from the Tom70 protein from yeast (33,34). However, GFP fusion proteins can also be targeted to the mitochondrial matrix, as has been demonstrated by the presequence from citrate synthase (35). We can thus conclude that the h mt TrpRS-GFP protein is located in the mitochondrial periphery and not in the matrix.
Suggestions have been made that the amino-terminal part of the cytoplasmic TrpRS could possess a secondary unknown function (36). Such a secondary function could be related to the mitochondrial and nuclear distribution, together with the IFN induction and the alternatively spliced 5Ј-end. Although several theories have been put forward about the involvement of TrpRS in the antiviral effect of IFN (37)(38)(39)(40), it is at present not certain why the human TrpRS is among the most strongly IFN-␥-induced proteins.
The signal peptide of 18 amino acids deviates from the predicted 46 amino acids using the MitoProtII program, which most likely is due to a wrong prediction by the program. However, another explanation for this discrepancy might be that the h mt TrpRS was expressed in insect cells and not in human cells. It is possible that there is a difference in the mitochondrial localization process between the two cell types, although earlier results from expression of mitochondrial proteins in the baculovirus expression system have indicated a tendency for correct processing (41)(42)(43). Therefore, it is likely that the mitochondrial import mechanisms in human and insect cells are closely related and that the signal peptide of h mt TrpRS expressed in human cells is identical to the signal peptide observed in insect cells.
Like the human cytoplasmic and the bacterial TrpRSs, the h mt TrpRS also belong to class I aaRSs harboring sequences similar to HXGH and KMSKS. Some variation of these motifs has been observed between species. The bacterial and mitochondrial HXGH sequence in TrpRS deviates by having an Asn instead of the last His (HLGN). In addition, bacterial forms have a Thr instead of the first His (T(I/L)GN). Instead of KMSKS, the hTrpRS has the sequence KMSAS, which is a change from a basic hydrophilic amino acid (Lys) to an aliphatic hydrophobic one (Ala). These changes are interesting because the imidazole side chain of both His residues in HXGH and the second Lys in KMSKS have been shown to be closely involved in binding of ATP and formation of the aminoacyladenylate in tyrosyl-and glutaminyl-tRNA synthetases from B. stearothermophilus and E. coli, respectively (44,45). The high homology of h mt TrpRS to B. subtilis TrpRS, which has been shown to be an ␣ 2 dimer (27), makes it likely that h mt -TrpRS also functions as an ␣ 2 dimer. This correlates well with cytoplasmic hTrpRS, which also has a homodimeric structure (22). However, as illustrated by the recent description of h mt -PheRS (29), a correlation between cytoplasmic and mitochondrial aaRS structures does not always apply. Cytoplasmic and prokaryotic PheRS belong to class II of aaRS with an ␣ 2 ␤ 2 tetramer structure, whereas the h mt PheRS is shown to be a monomer containing the three sequence motifs normally present in the ␣ subunit as well as a well-conserved sequence from the ␤ subunit.
The present data describe h mt TrpRS as a mitochondrial enzyme from the translational system, which will need much further investigation to clarify the increasing number of functions attributed to mitochondria in health and disease.