cDNA cloning, purification, and characterization of mouse liver selenocysteine lyase. Candidate for selenium delivery protein in selenoprotein synthesis.

Selenocysteine lyase (SCL) (EC 4.4.1.16) is a pyridoxal 5'-phosphate-dependent enzyme that specifically catalyzes the decomposition of L-selenocysteine to L-alanine and elemental selenium. The enzyme was proposed to function as a selenium delivery protein to selenophosphate synthetase in selenoprotein biosynthesis (Lacourciere, G. M., and Stadtman, T. C. (1998) J. Biol. Chem. 273, 30921-30926). We purified SCL from pig liver and determined its partial amino acid sequences. Mouse cDNA clones encoding peptides resembling pig SCL were found in the expressed sequence tag data base, and their sequences were used as probes to isolate full-length mouse liver cDNA. The cDNA for mouse SCL (mSCL) was determined to be 2,172 base pairs in length, containing an open reading frame encoding a polypeptide chain of 432 amino acid residues (M(r) 47, 201). We also determined the sequence of the N-terminal region of putative human SCL. These enzymes were shown to be distantly related in primary structure to NifS, which catalyzes the desulfurization of L-cysteine to provide sulfur for iron-sulfur clusters. The recombinant mSCL overproduced in Escherichia coli was a homodimer with the subunit M(r) of 47,000. The enzyme was pyridoxal phosphate-dependent and highly specific to L-selenocysteine (the k(cat)/K(m) value for L-selenocysteine was about 4,200 times higher than that for L-cysteine). Reverse transcriptase-polymerase chain reaction and Western blot analyses revealed that mSCL is cytosolic and predominantly exists in the liver, kidney, and testis, where mouse selenophosphate synthetase is also abundant, supporting the view that mSCL functions in cooperation with selenophosphate synthetase in selenoprotein synthesis. This is the first report of the primary structure of mammalian SCL.

Selenocysteine residue plays an essential catalytic role in various selenoproteins such as glutathione peroxidases from mammals (1) and formate dehydrogenase from Escherichia coli (2). Selenocysteine is co-translationally incorporated into a nascent polypeptide chain as encoded by the UGA-stop codon (2,3). The initial step of selenoprotein biosynthesis involves selenophosphate synthetase (SPS), 1 which catalyzes the formation of selenophosphate from selenide and ATP (4): selenophosphate is a precursor molecule of selenocysteyl-tRNA (5), which decodes the UGA codon.
Selenocysteine lyase (SCL) is a pyridoxal 5Ј-phosphate (PLP) enzyme that catalyzes the elimination of elemental selenium from L-selenocysteine to yield L-alanine. SCL has been purified from pig liver (6) and Citrobacter freundii (7) as the first enzyme that specifically acts on a selenium-containing compound and not on the sulfur analog. The mammalian enzyme is a homodimer with the subunit M r of 48,000, whereas the bacterial enzyme is monomeric with the M r of 64,000. SCL was proposed to cooperate with SPS in selenophosphate biosynthesis based on the following observations. 1) Selenium derived from L-selenocysteine added to Clostridium sticklandii cultures is more efficiently incorporated into selenoprotein A than selenide (8). 2) Replacement of selenide with L-selenocysteine and NifS from Azotobacter vinelandii in the in vitro SPS assay results in an increased rate of formation of selenophosphate (9). NifS catalyzes the removal of selenium from L-selenocysteine in the same manner as SCL does (9), although the physiological role of NifS is to supply sulfur atom for iron-sulfur clusters by catalyzing the desulfurization of L-cysteine (10). The role of SCL has not been examined in vivo mainly due to lack of information on its sequence.
Since SCL resembles NifS in catalytic function, we tested whether a particular NifS homolog functions as SCL, which specifically acts on L-selenocysteine. We purified an E. coli NifS homolog, CsdB (11), and found that it catalyzes the decomposition of L-selenocysteine to produce selenium and L-alanine. The enzyme is similar to SCL in its high activity toward Lselenocysteine: the activity of the bacterial enzyme toward L-selenocysteine is 290 times higher than that toward L-cysteine. However, the L-cysteine desulfurase activity of CsdB is not negligibly small, and its physiological relevance to selenium metabolism has not been established. In this respect, Patzer and Hantke (12) recently reported that a csdB (which they termed sufS) mutant is unable to use ferrioxamine B as an iron source and the csdB gene expression is regulated by the iron-dependent Fur repressor, suggesting that the csdB gene is involved in a biological process other than selenoprotein synthesis.
In the present study, we cloned and sequenced the cDNA (termed Scly) encoding mouse SCL (mSCL). We also deter-mined the sequence of the N-terminal region of putative human SCL. The primary structures of these enzymes revealed that mammalian SCLs are distantly related to NifS proteins and have characteristic sequences that are not found in other NifSlike proteins hitherto known. This is the first report on the primary structure of mammalian SCL.

EXPERIMENTAL PROCEDURES
Materials-L-Selenocystine was synthesized as described previously (13). Oligonucleotides were provided by Espec Oligo Service (Tsukuba, Japan). Restriction and DNA modification enzymes were purchased from New England Biolabs (Beverly, MA) and Takara Shuzo (Kyoto, Japan). All chemicals were of analytical grade.
Purification of Pig SCL (pSCL)-All steps were carried out at 4°C unless otherwise stated. A potassium phosphate buffer (KPB) (pH 7.4) containing 20 M PLP and 0.01% 2-mercaptoethanol was used as the standard buffer. A pig liver (1.5 kg) was minced with an ice-cold meat mincer and homogenized in a Waring Blender in 7.5 liters of a 50 mM standard buffer. The homogenate was centrifuged, and the supernatant was passed through a nylon mesh. The crude extract was fractionated with ammonium sulfate (25-45% saturation) and dialyzed with a 10 mM standard buffer. The enzyme solution was applied to a DEAE-Toyopearl column (10 ϫ 16 cm) equilibrated with a 10 mM standard buffer and eluted with a 10-liter linear gradient (0 -0.2 M) of KCl in the same buffer. The fractions containing the enzyme were concentrated by ammonium sulfate precipitation (50% saturation). The enzyme was applied to a Phenyl-Toyopearl column (5 ϫ 15 cm) equilibrated with a 10 mM standard buffer containing 0.6 M ammonium sulfate and eluted with 1.5 liters of the same buffer. The enzyme fractions were concentrated by ultrafiltration with a UP-20 membrane (Advantec, Naha, Japan). The enzyme was dialyzed against a 10 mM standard buffer containing 0.7 M ammonium sulfate, applied to a second Phenyl-Toyopearl column (5 ϫ 14 cm) equilibrated with the same buffer, and eluted with 10 mM standard buffer containing 0.6 M ammonium sulfate. The active fractions were pooled, concentrated by ultrafiltration as above, and dialyzed against 20 mM standard buffer. The enzyme was loaded onto a hydroxyapatite column (2.5 ϫ 10 cm) equilibrated with a 20 mM standard buffer, and eluted stepwise with 40 and 80 mM standard buffers. The enzyme fractions were pooled, concentrated by ultrafiltration, and dialyzed against a 20 mM standard buffer. The enzyme was applied to a Superdex 200 column (1 ϫ 30 cm) and fractionated with the same buffer at a flow rate of 0.25 ml/min with an FPLC pump system. The enzyme fractions were collected and dialyzed against 10 mM standard buffer.
Sequencing of pSCL-For N-terminal peptide sequencing, the enzyme (10 g) separated by SDS-PAGE was electroblotted to a polyvinylidene difluoride membrane, stained with 0.1% Ponceau S, excised with a razor blade, and subjected to sequence analysis with an automated protein sequencer PPSQ-10 (Shimadzu, Kyoto, Japan). In order to determine the internal sequences of pSCL, in-gel digestion was performed as follows. The enzyme (0.3 mg) separated by SDS-PAGE was stained with Coomassie Brilliant Blue, and the stained band corresponding to the enzyme was cut with a razor blade. The gel band was washed twice with 150 l of 50% acetonitrile in 0.1 M Tris-HCl (pH 9.0), at 30°C for 20 min, left in the air for 10 min at room temperature until semi-dry, and then partially rehydrated with 70 l of 0.02% Tween 20 in 0.1 M Tris-HCl (pH 9.0). Lysyl endopeptidase (0.5 g in 30 l of 0.1 M Tris-HCl, pH 9.0) was added to the sample, and the mixture was incubated for 16 h at 37°C. The peptides generated in the gel were recovered with 500 l of 60% acetonitrile containing 0.1% trifluoroacetate at 37°C for 20 min by vortexing gently. The gel was crushed, and a second extraction was performed with the same solution by shaking for 1 h. The crushed gel was removed with an Ultrafree C3GV filter (Millipore, Bedford, MA). The extracts were combined and concentrated to 20 l with a vacuum evaporator. The peptides were separated with an Asahipak ODP-50 column (6 ϫ 150 mm) connected to an high performance liquid chromatography system (Tohso, Yamaguchi, Japan) with a 50-ml linear gradient (24 -64%) of acetonitrile in 0.05% (v/v) trifluoroacetate at a flow rate of 0.5 ml/min. The separated peptides were subjected to sequence analysis.
Isolation of cDNA Encoding mSCL-BLAST analysis of the EST data bases using the pSCL sequences resulted in the identification of highly homologous mouse and human cDNA sequences (Fig. 1). Mp1 (5Ј-CCGACAGTGCGCTCCCTTCAA-3Ј) and Mp2 (5Ј-GTGAACCATGTAT-CCCTTCAG-3Ј) were used to amplify a 340-bp fragment of AA107712, and Mp3 (5Ј-CAGGATCGGTGCTCTGTATGT-3Ј) and Mp4 (5Ј-GGCTG-TTCAAATGGATTCTCT-3Ј) were used to amplify a 250-bp fragment of MUS94C09 (Fig. 2). The PCR products were gel-purified, labeled with digoxigenin, and used as probes to obtain an mSCL cDNA clone from a mouse liver ZAP cDNA library constructed with ZAP-cDNA Synthesis Kit (Stratagene, La Jolla, CA). Immunodetection was performed with Anti-DIG, Fab fragment AP Conjugate (Roche Diagnostics, Basel, Switzerland), nitro blue tetrazolium, and 5-bromo-4-chloro-3-indolylphosphate-p-toluidine salt. Positive clones were isolated, and their cDNA inserts were excised in the form of pBluescript plasmid and analyzed by sequencing. Sequencing of an isolated clone, Scly1, revealed that it is comprised of 2,137 bp including a poly(A) ϩ tail in the 3Ј end region. However, the clone lacked 4 bp, TAGG, at position 165, which corresponds to position 196 -199 of the EST clone AA107712, and no in-frame ATG codon was found near the 5Ј end of this sequence (Fig. 2). To determine the sequence of the 5Ј region of the transcript, 5Ј-rapid amplification of cDNA ends (14) and CapFinder (CLONTECH, Palo Alto, CA) techniques were employed. The 4 bp, TAGG, were found in the PCR products from the fresh preparation of a transcript. A possible initiation ATG codon was found in a position adjacent to the 5Ј end of Scly1 (Fig. 2). The initiation codon, ATG, and four bases, TAGG, were introduced into the incomplete cDNA clone Scly1 ( cDNA Cloning of Mouse Selenocysteine Lyase GCGCG-3Ј) and Mp6 (5Ј-CACCTTGGCCTTCCTACCTGACACATAGG-AGC-3Ј) were used for amplification of the DNA fragment encoding the N-terminal part of mSCL. Another set of primers, Mp7 (5Ј-CAGGTAG-GAAGGCCAAGGACATTATAAATG-3Ј) and Mp8 (5Ј-CCCCAAGCTTG-AGCCGCCCTTCCAGTTGGGCC-3Ј), was used to amplify the DNA fragment coding for the C-terminal fragment of mSCL. The initiation codon to be introduced is double underlined, and the underlined nucleotides indicate the 4 bases to be inserted. The restriction enzyme sites for NdeI and HindIII are shown in italics. After the 2nd round of PCR, the final 1.3-kbp product was purified, digested with NdeI and HindIII, and subcloned into the same sites of pET21a(ϩ) to form Scly2 (Fig. 2). Scly2 provides the fusion protein (mSCL2) with a C-terminal His 6 tag. The fusion protein was produced in E. coli BL21(DE3) as an inclusion body, purified with a His⅐Bind column (Novagen, Madison, WI), and used for the generation of polyclonal antibodies in rabbits.
The entire coding sequence of mSCL cDNA (Scly) was obtained by PCR with the primers Mp5 and Mp9 (5Ј-CCCCAAGCTTCTAGAGC-CGCCCTTCCAGTTGGGCC-3Ј) and total RNA as a template. An additional 18-bp insertion, 5Ј-AGGTCACTTTTGTCCCAG-3Ј, was found in the region corresponding to that between positions 448 and 449 in the ORF of Scly2 (Fig. 2). The product was subcloned by PCR into the NdeI and HindIII sites of pET21a(ϩ) to yield pESL. This construct provides a functional mSCL without a His 6 tag (Fig. 2).
Purification of mSCL-Purification of mSCL was performed at 0 -4°C, and KPB (pH 7.4) was used as the standard buffer. E. coli BL21(DE3) harboring pESL was grown in 500 ml of LB medium containing 100 g/ml ampicillin at 28°C for 18 h by induction of the gene expression with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. The cells were harvested by centrifugation, suspended in 50 mM KPB containing 2 mM phenylmethylsulfonyl fluoride, 5 mM EDTA, and 2 g/ml pepstatin A, and then disrupted by sonication. The cell debris was removed by centrifugation, and the supernatant solution was fractionated by ammonium sulfate precipitation (1.0 -3.0 M). The enzyme was dissolved in 10 mM KPB containing 1.0 M ammonium sulfate and applied to a Butyl-Toyopearl column (3 ϫ 9.5 cm) equilibrated with the same buffer. After the column was washed with the same buffer, the enzyme was eluted with a 0.8-liter linear gradient of 1.0 -0 M ammonium sulfate in the buffer. The active fractions were pooled and concentrated by 3.0 M ammonium sulfate. The enzyme was collected by centrifugation, resuspended in 10 mM KPB containing 1 mM dithiothreitol, 0.5 mM phenyl-methylsulfonyl fluoride, 1 mM EDTA, and 1 g/ml pepstatin A, and then desalted with a Sephadex G-25 column (2 ϫ 24 cm) equilibrated with 50 mM KPB. The active fractions were collected and applied to a Q-Sepharose column (3 ϫ 10 cm) equilibrated with the same buffer. After the column was washed with the same buffer, the enzyme was eluted with a 0.8-liter linear gradient of 0 -0.2 M NaCl in the buffer, and the active fractions were pooled and concentrated with ammonium sulfate as above. The enzyme collected by centrifugation was resuspended in 10 mM KPB containing 1 mM dithiothreitol, 0.5 mM EDTA, and 20 M PLP and then dialyzed against the same buffer. The final preparation was stored at Ϫ80°C until use.
Preparation of Cell Extracts and Subcellular Fractions from Mouse Liver-The liver from a BALB/c mouse (6 weeks, male) was homogenized in ice-cold 0.25 M sucrose solution containing 3 mM Tris-HCl (pH 7.4), and 0.1 mM EDTA, and centrifuged for 10 min at 700 ϫ g. The supernatant was centrifuged for 10 min at 7,000 ϫ g to obtain crude mitochondrial pellets. The supernatant was centrifuged again for 60 min at 105,000 ϫ g to obtain a microsomal fraction and a cytosolic fraction. A pure nuclear fraction was obtained using the method of Blobel and Potter (15).
Western Blot Analysis with Anti-mSCL2 Antibody-Proteins in various cell homogenates and the subcellular fractions were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. Western blot analysis was performed with the anti-mSCL2 antibodies, and the proteins were detected by chemiluminescence using CDP-Star (Roche Diagnostics, Basel, Switzerland).
Enzyme Assays-The enzyme was assayed in a 0.12 M Tricine-NaOH buffer at pH 9.0. The enzymatic activity toward L-selenocysteine was measured with lead acetate as described previously (6). Sulfite produced from L-cysteine sulfinic acid was determined with fuchsin (16). Production of alanine from L-cysteine and other substrates was determined with a Beckman high performance amino acid analyzer 7300   . Specific activity was expressed as units/mg of protein, with 1 unit of enzyme defined as the amount that catalyzed the formation of 1 mol of the product in 1 min. Analytical Methods-Protein was quantified by the Bradford method (17) using Protein Assay CBB Solution (Nacalai Tesque, Kyoto, Japan) with bovine serum albumin as a standard. The concentration of the purified enzyme was determined with the value ⑀ M ϭ 1.7 ϫ 10 4 M Ϫ1 ⅐cm Ϫ1 at 280 nm, which was calculated from the content of tyrosine, tryptophan, and cystine (18). The subunit and the native M r of mSCL were determined by SDS-PAGE (19) and gel filtration with Superose 12 (Amersham Pharmacia Biotech, Uppsala, Sweden), respectively.

Determination of the Amino Acid Sequence of pSCL-pSCL
was purified 1,600-fold to about 60% homogeneity with specific activity of 21 units/mg. The preparation provided two bands at 37 and 43 kDa on SDS-PAGE. The 43-kDa band corresponding to pSCL was further purified by SDS-PAGE and digested with lysyl endopeptidase. The peptides were separated by high performance liquid chromatography and subjected to sequence analysis. We determined the amino acid sequences of five peptides, comprising 104 amino acid residues in total (Fig. 1).
Structure of cDNA Encoding Functional mSCL-We have isolated full-length mSCL cDNA, termed Scly, as described under "Experimental Procedures" using primers designed based on the amino acid sequences shown in Fig. 1. Its nucleotide sequence and deduced amino acid sequence are shown in Fig. 3. The total length of Scly was 2,172 bp, containing an open reading frame encoding a polypeptide chain of 432 amino acid residues. The 3Ј-untranslated region contains poly(A) ϩ tail and two potential overlapping polyadenylation signals, AATTAA and ATTAAA, located 13 and 12 bp, respectively, upstream from the poly(A) ϩ tail.
Functional Analysis of Recombinant mSCL-The functional enzyme, mSCL, was produced, and the product amounted to about 20% of the total protein in the crude extract of E. coli BL21(DE3) harboring pESL. Purified mSCL (Table I) provided a 47-kDa single band on SDS-PAGE (Fig. 4). The N-terminal sequence of the purified enzyme, MDAARNGALG, agreed with that deduced from the nucleotide sequence of Scly. The M r of the native enzyme was determined to be 105,000 by gel filtration. Therefore, the enzyme is a dimer composed of two identical subunits. The enzyme was yellow and exhibited its absorption maximum at 420 nm, which is characteristic of PLP enzymes. The requirement of PLP as a cofactor of the enzyme was examined as follows. Reduction with sodium borohydride resulted in irreversible inactivation of the enzyme and disappearance of the absorption band at 420 nm with concomitant increase in absorbance at 330 nm. When the enzyme was incubated with 1 mM hydroxylamine, the activity decreased to 25% of the original activity. Addition of 0.2 mM PLP to the dialyzed enzyme restored 90% of the original activity. Thus, PLP serves as a cofactor of the enzyme and probably binds to the ⑀-amino group of a lysine residue at the active site of the enzyme through an aldimine linkage, as in other PLP enzymes studied so far.
The specific activity of the purified enzyme (29 units/mg) for L-selenocysteine was comparable to that of pSCL (37 units/mg) (6) ( Table I). The enzyme showed maximum reactivity at about pH 9 when measured in Tricine-NaOH (pH 7.0 -9.5) and glycine-NaOH (pH 8.5-11) buffers. The optimum pH of mSCL activity is similar to that of pSCL activity (pH 9.0) (6). In the mSCL reaction, alanine and selenide were formed in a 1:1 stoichiometric ratio from L-selenocysteine in the presence of dithiothreitol (data not shown). However, the selenide product was shown to be produced by the nonenzymatic reduction of selenium by dithiothreitol as was shown in the previous studies on pSCL (20). This indicates that the product of the enzyme reaction is elemental selenium.
The substrate specificity and kinetic parameters for mSCL with L-selenocysteine, L-cysteine sulfinate, and L-cysteine are summarized in Table II. The enzyme exhibits an extremely high, although not absolute, specificity to L-selenocysteine. The k cat /K m value for L-selenocysteine is about 100 times and 4,200 times higher than that for L-cysteine sulfinate and L-cysteine, respectively, confirming that this enzyme is the mouse counterpart of pSCL. The production of alanine was not detected when any of the following amino acids were used as substrates (at 5 mM): L-aspartate, DL-kynurenine, L-selenocystine, L-cystine, D-selenocysteine, or D-selenocystine.
FIG. 5. RT-PCR analysis of tissue distribution of mSCL mRNA. RT-PCR products were subjected to 1% agarose gel electrophoresis. Lanes: M1, 100-bp DNA ladder marker; B, brain; H, heart; Lu, lung; St, stomach; Lv, liver; K, kidney; Sp, spleen; T, testis; M2, 1-kbp DNA ladder marker; Gen, a PCR product obtained with the same primers using a mouse genome DNA as a template. cDNA Cloning of Mouse Selenocysteine Lyase Tissue Distribution and Intracellular Localization of mSCL-In order to determine the tissue distribution of the Scly transcript, RT-PCR was performed using a set of primers (Mp10 and Mp11) specific to Scly and total RNAs prepared from various mouse tissues (brain, heart, lung, stomach, liver, kidney, spleen, and testis) as a template. As a control, PCR was performed using the same set of primers with the mouse genomic DNA as a template. As shown in Fig. 5, a 270-bp fragment was amplified in all examined tissues, indicating that the Scly gene is expressed ubiquitously. In the control reaction using genomic DNA, a fragment of about 4.5 kbp was amplified (Fig. 5). The results indicate that this region of the genomic DNA contains at least one intron and that the 270-bp fragments were derived from RNA and not from contaminating genomic DNA.
The tissue distribution of mSCL was examined by Western blotting. Polyclonal antibodies raised against mSCL2 were used to detect mSCL in various tissue homogenates of the mouse. An immunoreactive protein of 47 kDa was detected in all cell homogenates examined (Fig. 6A). Liver, kidney, and testis appeared to have the highest mSCL content.
We determined the intracellular localization of mSCL by Western blot analysis with anti-mSCL2 antibodies. Mouse liver was homogenized and fractionated into nuclear, mitochondrial, microsomal, and cytosolic fractions. The immunoreactive 47-kDa protein band was detected mainly in the cytosolic fraction (Fig. 6B).

DISCUSSION
In this study, we cloned cDNA encoding SCL from mouse liver. The properties of the recombinant mSCL are very similar to those of pSCL in molecular weight, amino acid sequence, optimum pH, subcellular localization, and high specificity toward L-selenocysteine.
Homology searches using the BLAST program against the nonredundant data base revealed that several human EST sequences show strong homology with mSCL. By assembling 11 independent EST sequences and a sequence of cDNA encoding putative human SCL (GenBank accession number AF175767) isolated by the 5Ј-rapid amplification of cDNA ends method, 2 we determined an amino acid sequence encoding the N-terminal region of putative human SCL (hSCL) (Fig. 7). The peptide sequences of pSCL have striking similarity to mSCL and hSCL (Fig. 7), indicating that SCL is highly conserved in mammals.
Overall sequence similarity (ϳ30%) is found between mammalian NifS homologs and SCLs. However, they are clearly classified into two distinct groups: one includes mSCL, hSCL, and pSCL, and the other includes a mouse NifS homolog (m-Nfs1) and a human NifS homolog (hNifS) (Fig. 7). In particular, the regions corresponding to Gln 105 -Gly 121 and Asn 205 -Pro 214 of mSCL are not found in mammalian NifSs.
We previously classified NifS family proteins into two groups 2 7. Comparison of the amino acid sequences of mammalian SCLs and NifSs. The amino acid sequences of mammalian SCLs from mouse (mSCL), GenBank accession number AF175407, and human (hSCL) and those of the mammalian nifS-like gene products from mouse (m-Nfs1), GenBank accession number AJ222660, and human (hNifS), GenBank accession number AF097025, are shown. The sequences of pSCL fragments are also shown. The alignment was generated with the CLASTAL W program (29). The black boxes show the residues conserved between mSCL and other proteins. The gray boxes show the residues conserved only between m-Nfs1 and hNifS. In order to deduce the hSCL sequence, the following sequences in the GenBank/EST data bases were assembled: AF175767 AA312035, AA911739, AI422406, AI424010, H30676, HSCOY-A051, N56305, R13678, R13816, R13983, and R37777. (21): group I includes m-Nfs1, hNifS, A. vinelandii NifS, E. coli IscS, etc., and group II includes E. coli CsdB, E. coli CSD, etc. Sequence alignment (Fig. 7) and the phylogenetic tree (Fig. 8) indicate that mammalian SCLs belong to neither of them. Although E. coli CsdB is similar to mSCL in its high specificity toward L-selenocysteine, it shows only 16% sequence identity with mSCL. Accordingly, mammalian SCLs define a new group of enzymes.
Recent studies on eukaryotic NifS-like proteins from human (22), mouse (23), and yeast (23)(24)(25) showed that these proteins have mitochondrial sorting signals at their N-terminal regions. m-Nfs1 predominantly exists in the mitochondrial matrix (23). Yeast Nfs1 was shown to be sorted mainly to mitochondrial matrix (25). hNifS is produced in two different forms because of the presence of two alternative initiation codons in its gene transcript (22). The larger protein has a mitochondrial targeting signal and is transported into mitochondria. Many ironsulfur proteins such as components of the TCA cycle and the respiratory chain are present in eukaryotic mitochondria. The subcellular localization of mammalian NifSs is consistent with the hypothesis that they are directly involved in the de novo formation of iron-sulfur clusters (22).
In contrast, mSCL exists mainly in the cytosolic fraction. The subcellular localization of mSCL is consistent with the hypothesis that mSCL delivers an active form of selenium to the SPS reaction, which proceeds in cytosol. Recently, Lacourciere and Stadtman (9) found that the replacement of selenide by NifS and L-selenocysteine in an in vitro SPS assay system resulted in an increased rate of formation of selenophosphate, indicating that selenium derived from L-selenocysteine by the action of NifS serves as a better substrate than selenide for SPS. It is reasonable to assume that an enzyme specific toward L-selenocysteine, namely SCL, functions in a physiological selenium delivery system.
RT-PCR and immunoblot analysis showed that the mRNA of mSCL is distributed in brain, heart, lung, stomach, liver, kidney, spleen, and testis (Figs. 5 and 6A). This result is consistent with the wide distribution of this enzyme's activity in various mammalian tissues described previously (6). mSCL was predominantly expressed in liver, kidney, and testis (Fig. 6A). Interestingly, SPS is also highly expressed in liver, kidney, and testis, where selenoproteins are produced (26). This supports the view that mSCL cooperates with SPS in the production of selenophosphate.
Free selenocysteine is formed by degradation of selenoproteins containing selenocysteine residues. It can also be produced from selenomethionine by cystathionine ␤-synthase and cystathionine ␥-lyase in mammalian cells (27). Since a high concentration of free selenocysteine is toxic and its accumula-tion is lethal (28), excess selenocysteine has to be decomposed. SCL, which specifically decomposes selenocysteine into alanine and elemental selenium, may have the role of maintaining the normal concentration of selenocysteine. The result obtained in the present study that the SCL gene is expressed in various tissues is compatible with this hypothesis. The nucleotide sequence of cDNA encoding SCL determined in the present study enables us to examine the physiological role of SCL genetically.