INTRODUCTION

Selenium, a homolog of sulfur, is an essential trace element for mammals and other organisms. It occurs in some selenoproteins as a selenocysteine residue (1-3), which is incorporated co-translationally into the proteins as directed by the unique codon, UGA (4, 5). Other selenoproteins contain selenium in a dissociable form, coordinated with molybdenum (6, 7). Selenium is metabolized by enzymes including selenophosphate synthetase (8), selenocysteine synthase (9), selenocysteine lyase (10, 11), and selenocysteine methyltransferase (12). Some enzymes participating in sulfur metabolism also act on the selenium analogs of the substrates.

We found selenocysteine lyase in mammals (10) and bacteria (13), and purified the enzyme from pig liver (10) and Citrobacter freundii (11). The enzyme specifically decomposes L-selenocysteine into L-alanine and elemental selenium; L-cysteine is inert as a substrate. Zheng et al. (14) recently demonstrated the function of NIFS protein, which is required for the efficient construction of the Fe-S clusters of nitrogenase in a diazotrophic bacterium Azotobacter vinelandii. NIFS catalyzes the same type of reaction as selenocysteine lyase, but acts on both L-cysteine and L-selenocysteine indiscriminately. The enzyme was named cysteine desulfurase, based on its inherent physiological role. Genes with a sequence similarity to that of nifS have been found not only in diazotrophs but also in non-diazotrophic microorganisms. It has been reported that the nifS-like genes of Bacillus subtilis and Saccharomyces cerevisiae are involved in NAD⁺ biosynthesis (15) and tRNA processing (16), respectively.

The nucleotide sequence of the whole Escherichia coli genome has been determined (17), and the bacterium appears to contain three nifS-like genes (18, 19). One of the genes located at 57.3 min (18) in the chromosome presumably encodes the NIFS-like protein purified by Flint (20). Not only the amino acid sequence but also the catalytic properties of the enzyme resemble those of A. vinelandii NIFS. We have found that the N-terminal amino acid sequence of pig liver selenocysteine lyase is similar to that of A. vinelandii NIFS (21).¹ If we assume that E. coli contains selenocysteine lyase and that the enzyme resembles NIFS, one or both of the other two nifS-like genes may encode selenocysteine lyase(s). Alternatively, the genes may encode new enzymes participating in an unknown metabolism of sulfur or selenium amino acids. We have cloned the nifS-like gene mapped at 63.4 min (18), and found that the gene product is a novel PLP²-dependent enzyme decomposing L-selenocysteine, L-selenocystine, L-cysteine, and L-cystine. L-Cysteine sulfinic acid is also decomposed to L-alanine as the best substrate of the enzyme. We have tentatively named it cysteine sulfinate desulfinase. We describe here the characteristics of the enzyme and compare it with other related enzymes such as selenocysteine lyase and NIFS.

EXPERIMENTAL PROCEDURES

Molecular weight markers for SDS-PAGE and gel filtration were purchased from Pharmacia Biotech Inc. and Oriental Yeast, respectively; restriction enzymes and other DNA modifying enzymes from Takara Shuzo; synthetic oligonucleotides from Japan Bio Service and Biologica. L-Selenocystine was synthesized from L--chloroalanine, which was kindly provided by Showa Denko, and from disodium diselenide, as described previously (22). L-Selenocysteine was prepared from L-selenocystine according to the previous method (10). All other chemicals were of analytical grade.

The DNA fragment containing the ORF for cysteine sulfinate desulfinase was amplified with a Perkin-Elmer Thermal Cycler 480. The reaction mixture (50 µl) contained: 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 10% Me₂SO, 400 µM each dNTP, 0.2 µM each primer (5-GGAATTCATCAAGCCGAGGAGTAC-CATGAACG-3 and 5-AACTGCAGCGGCGAATTGCGGGTTTGTCATTAA-3; underlines indicate EcoRI and PstI sites, and boldface letters indicate a putative ribosome binding sequence), 2.5 units of Ex Taq DNA polymerase (Takara Shuzo), and 100 ng of template DNA from E. coli JM109 isolated by the method reported (23). The conditions were as follows: first cycle, 2 min (94 °C), 5 min (55 °C), and 10 min (72 °C); subsequent 24 cycles, 1 min (94 °C), 1 min (58 °C), then 3 min (72 °C); last cycle, 1 min (94 °C), 1 min (58 °C), 15 min (72 °C). The EcoRI-PstI fragment was ligated into pUC118 to give a plasmid pCSD1.

Cysteine sulfinate desulfinase was assayed by determination of H₂Se formed from L-selenocysteine with lead acetate as described previously (10). A standard reaction mixture containing 5 mM L-selenocysteine, 50 mM dithiothreitol, 0.02 mM PLP, 120 mM Tricine-NaOH buffer (pH 7.5), and enzyme (0.006-0.032 units) in a final volume of 0.1 ml was incubated at 37 °C. A molar turbidity coefficient of PbSe at 400 nm, 1.18 × 10⁴ M¹·cm¹, was used. One unit of enzyme was defined as the amount of enzyme that catalyzes the formation of 1 µmol of the product (alanine or elemental selenium)/min. Specific activity was expressed as units/mg of protein.

Cysteine desulfurase activity was measured by the determination of H₂S formed from L-cysteine with lead acetate in the manner described above. An apparent molar turbidity coefficient of colloidal PbS at 360 nm, 1.31 × 10⁴ M¹·cm¹, was used (24). Pyruvate was determined with lactate dehydrogenase (Sigma) at 37 °C in a reaction mixture (1 ml) containing 5 mM L--chloroalanine, lactate dehydrogenase (22 units), 0.02 mM PLP, 0.15 mM NADH, 120 mM Tricine-NaOH buffer (pH 8.5), and enzyme. Sulfite produced from L-cysteine sulfinic acid was determined with fuchsin (25). Protein was determined with a Bio-Rad protein assay kit with bovine serum albumin as a standard.

The buffer used throughout the purification was 10 mM potassium phosphate (pH 7.4) and was supplemented with salts when required for chromatographies. All operations (Table I) were done at 4 °C. Elution patterns of protein were estimated by absorption at 280 nm. E. coli XL1-Blue (Stratagene) carrying pCSD1 was cultured aerobically in Luria-Bertani broth (3 liters) supplemented with ampicillin (200 µg/ml) at 37 °C for 11 h, and then isopropyl-1-thio--D-galactopyranoside was added to the culture at a final concentration of 1 mM. The cells were cultured for another 2 h, and harvested by centrifugation. The final preparation of the enzyme was stored frozen at 30 °C in the buffer supplemented with 0.02 mM PLP until use.

Table I. Purification of cysteine sulfinate desulfinase Step Total protein Total activitya Specific activity Yield mg units units/mg % Crude extractb 1900 1100 0.58 100 Ammonium sulfatec 720 920 1.3 84 1st DEAE-Toyopearld 420 790 1.9 72 2nd DEAE-Toyopearle 90 360 4.0 33 1st Phenyl-Toyopearlf 29 190 6.6 17 2nd Phenyl-Toyopearlg 28 230 8.2 21 a Determined with L-selenocysteine as a substrate. b Cells (13.8 g) suspended in 50 ml of buffer were sonicated. c The fraction between 25 and 50% saturation was collected, then dialyzed. d Column size, 3.0 × 18 cm; elution, with 0.15 M NaCl. The active fractions were concentrated with 55% saturation of ammonium sulfate, then dialyzed. e Column size, 3.0 × 18 cm; elution, by 1000-ml linear gradient of 0-0.2 M NaCl. The active fractions were concentrated with an UP-20 membrane (Advantec), then dialyzed against 0.4 M ammonium sulfate. f Column size, 0.7 × 6 cm; equilibration, with 0.4 M ammonium sulfate; elution, with 0.4 M ammonium sulfate. The active fractions were concentrated with an UP-20 membrane, then dialyzed against 0.65 M ammonium sulfate. g Column size, 0.7 × 6 cm; equilibration, with 0.65 M ammonium sulfate; elution, by 100-ml linear gradient of 0.65-0.3 M ammonium sulfate. The active fractions were concentrated with an UP-20 membrane, then dialyzed against the buffer containing 0.02 mM PLP.

The molecular mass of the enzyme was determined with a Vision 2000 reflector-type time-of-flight mass spectrometer (Thermoquest, Tokyo, Japan) equipped with a nitrogen laser (337 nm, pulse length 10 ns). The enzyme solution (1.0 × 10⁵ ~ 2.5 × 10⁶ M) was mixed with the same volume of 1% (w/v) 2,5-dihydroxybenzoic acid solution in 10% acetonitrile containing 0.1% (v/v) trifluoroacetic acid. Spectral measurements were repeated 100 times, and the average of their sum was recorded. Bovine serum albumin was used as a standard protein.

The alanine mutants for cysteine residues were prepared by the Kunkel method (26). Single-stranded DNA was obtained with another plasmid named pCSD2, which was constructed in the same manner as pCSD1, except that the following primer was used for the upstream region: 5-GGGAATTCCATATGAACGTTTTTAATCCCGCGCAGTTTCG-3 (the EcoRI site is underlined). The single-stranded DNA was prepared from E. coli BW313 carrying pCSD2, by infection with helper phage VCSM13 (Stratagene). The following mutagenic primers were used: 5-ACGCGCATAGGCTTGTGCC-AC-3 (Cys-100), 5-CAGATCCGGGGCACCGCCAGT-3 (Cys-176), 5-GGAATCCTGGGCGCGGAATGA-3 (Cys-323), 5-GGCTGAGCGGCATGCTGCCC-3 (Cys-358) (mutagenized nucleotides are underlined). The nucleotide sequences were confirmed with a Dye Terminator sequencing kit and an Applied Biosystems 370A DNA sequencer. Mutant enzymes were prepared with E. coli JM109 as a host.

RESULTS

We determined the N-terminal amino acid sequence of selenocysteine lyase purified from pig liver, and found that 8 out of 14 residues at the N-terminal sequence coincide with the residues at the corresponding positions in the A. vinelandii NIFS sequence (21). Thus, selenocysteine lyase, if it occurs in E. coli, probably has a primary structure similar to that of A. vinelandii NIFS. The amino acid sequence deduced from ORF o401 located at 63.4 min (18) in the E. coli K-12 genome was found to have about 20% homology with that of A. vinelandii NIFS. We cloned the gene by polymerase chain reaction, with the E. coli JM109 chromosomal DNA as a template and the synthetic primers shown above. The DNA sequence of the gene thus cloned agreed completely with that registered in GenBank^TM (accession number U295810).

The molecular weight of the homogeneous preparation of the gene product estimated by SDS-PAGE (27) (about 43,000) agreed with the value calculated from the deduced amino acid sequence (43,238). The N-terminal amino acid sequence of the purified protein agreed with that deduced from the nucleotide sequence (Fig. 1). Although we did not determine the C-terminal amino acid sequence of the protein, the molecular mass of the protein determined by mass spectrometry was essentially identical with the predicted value: m/z 43279.6 corresponding to (M + H)¹⁺. The molecular weight of the purified protein in the native form was estimated to be about 97,000 by gel filtration with a Superose 12 (1 × 30 cm) column.

An extract of the cloned cells showed both selenocysteine lyase and cysteine desulfurase activity. Because the selenocysteine lyase activity was higher than that of cysteine desulfurase, we routinely used L-selenocysteine as a substrate of the enzyme throughout the purification (Table I).

The specific activity of the homogeneous preparation of the enzyme toward L-selenocysteine (8.1 units/mg) was comparable to that of selenocysteine lyase from C. freundii (6.47 units/mg) (11). However, these values were about 5 times lower than that of selenocysteine lyase from pig liver (37 units/mg) (10).

The cysteine desulfurase activity of our new enzyme (3.4 units/mg) was also much higher than that of A. vinelandii NIFS (14) and the E. coli NIFS-like enzyme (20). However, the concentration of L-cysteine we used (80 mM) was much higher than that used for the other enzymes (A. vinelandii NIFS, 0.5 mM; E. coli NIFS-like protein, 2.5 mM). The specific activity of our enzyme decreased at lower L-cysteine concentrations (0.050 unit/mg at 0.5 mM; 0.23 unit/mg at 2.5 mM). These values were comparable to those reported for A. vinelandii NIFS (0.089 unit/mg at 0.5 mM) (28) and the E. coli NIFS-like protein (0.078 unit/mg at 2.5 mM) (20).

The optimal pH values for the removal of selenium and sulfur atoms from L-selenocysteine and L-cysteine were around pH 7.0 and 7.5, respectively, in Tricine-NaOH. The enzyme kept essentially the same activity at 4 °C for at least 2 weeks, as well as at 30 °C for more than 6 months.

The enzyme showed at pH 7.4 an absorption maximum at 420 nm (Fig. 2, curve a) that is characteristic of bound PLP. Reduction with sodium borohydride resulted in the disappearance of the absorption band at 420 nm with a concomitant increase in the absorbance at 335 nm (Fig. 2, curve b). The reduced enzyme was catalytically inactive, and the addition of PLP did not reverse the inactivation. These results show that the enzyme requires PLP as a cofactor.

The enzyme resembles selenocysteine lyase and NIFS in that it removes elemental sulfur or selenium from L-cysteine or L-selenocysteine in the reaction. L-Cysteine sulfinic acid acted as the best substrate of the enzyme, and essentially the same amounts of L-alanine and sulfite were produced in the reaction (data not shown). Maximum activity for the desulfination was found at around pH 8.2 in Tricine-NaOH. We named our new enzyme cysteine sulfinate desulfinase, because the enzyme showed the lowest K_m value and the highest k_cat and k_cat/K_m values for L-cysteine sulfinic acid (Table II).

Table II. Substrate specificity of cysteine sulfinate desulfinase and kinetic constants of the enzyme reactions The following amino acids and derivatives were inert as the substrates, when production of alanine or consumption of the substrates were examined with a Beckman high performance amino acid analyzer 7300: D-cysteine, D-cystine, DL-cysteic acid, DL-serine, S-methyl-L-cysteine, S-benzyl-L-cysteine, DL-homocysteine, DL-homocystine, DL-methionine, DL-homoserine, L-homocysteic acid, cysteamine, cystamine, selenocystamine, L-asparagine, L-aspartic acid, L-kynurenine, DL-lanthionine, L-cystathionine, L-allocystathionine, and DL-djenkolic acid. Km Vmax kcat kcat/Km mM µmol·min1·mg1 s1 mM1·s1 L-Cysteine sulfinic acida 0.24 20 15 63 L-Selenocysteineb 1.0 7.4 5.3 5.3 L-Cysteineb 35 3.4 2.4 0.070 L-Cystinea 3.3 0.017 0.012 0.0036 a Alanine formed, in the reaction system with Tricine-NaOH (pH 8.5), was determined with alanine dehydrogenase with a mixture (1.0 ml) containing 5 mM substrate, 0.02 mM PLP, 2.5 mM NAD+, 0.3 unit of L-alanine dehydrogenase (Unitika), 120 mM Tricine-NaOH buffer (pH 8.5), and enzyme at 37 °C. b The reactions were carried out in Tricine-NaOH (pH 7.5), and H2Se and H2S formed were determined with lead acetate.

The enzyme resembles aspartate -decarboxylase (EC 4.1.1.12) and kynureninase (EC 3.7.1.3) because alanine is produced from the substrate. However, both L-aspartate and L-kynurenine were inert as the substrates (Table II). L-Aspartate was not converted to alanine even at pH 5.0 in an acetate buffer, which are the optimum conditions for the aspartate -decarboxylase reaction (29).

Cysteine sulfinate desulfinase is distinct from selenocysteine lyase and NIFS in that it acts also on L-cystine and L-selenocystine. L-Alanine was produced from either substrate as well as elemental sulfur or selenium. The amount of L-alanine produced was only 1.5 times larger than that of L-cystine consumed in the reaction. We expected that double amounts of alanine would be produced from cystine in the reaction, because S-sulfocysteine is presumably produced from cystine and then converted with the release of elemental sulfur to reproduce cysteine, which would give another molecule of alanine in the second reaction. Therefore, some part of the S-sulfocysteine is probably converted to another unknown compound, which is inert as a substrate.

Action on -Chloroalanine

-Chloroalanine is a mechanism-based inactivator for several PLP-dependent enzymes (30). Cysteine sulfinate desulfinase was incubated with various concentrations of L--chloroalanine (0.5-8.0 mM) for various periods up to 4.5 h, and the remaining activity was determined with L-selenocysteine as a substrate. No inactivation of the enzyme was observed. The enzyme catalyzed only ,-elimination of -chloroalanine to form pyruvate without formation of alanine.

Both A. vinelandii NIFS and the E. coli NIFS-like protein contain cysteinyl residues that are catalytically essential (14, 20). Selenocysteine lyase from C. freundii was completely inactivated by thiol reagents (11). Cysteinyl residues corresponding to Cys-325 of A. vinelandii NIFS, which was shown to be catalytically essential (14), are fully conserved among all NIFS family proteins (Fig. 3). We examined the roles of all cysteinyl residues of cysteine sulfinate desulfinase including Cys-358, which corresponds to the conserved Cys-325 of A. vinelandii NIFS, by site-directed mutagenesis. All the alanine mutants for Cys-100, Cys-176, Cys-323, and Cys-358 were fully active. Thus, cysteine sulfinate desulfinase has no essential cysteine residues and differs markedly from A. vinelandii NIFS and the NIFS-like protein of E. coli.

DISCUSSION

NIFS of A. vinelandii participates in construction of the Fe-S clusters of not only nitrogenase (31), but also other iron-sulfur proteins such as SoxR (32) and FNR (33). The NIFS-like enzyme of E. coli found by Flint (20) also provides apo-dihydroxy-acid dehydratase with a [4Fe-4S] cluster to reconstitute the enzyme in vitro. The N-terminal amino acid sequence of the enzyme was identical with that deduced from another nifS-like gene of E. coli (Eco1 in Table III). This gene, together with a nifU-like gene, forms a unique gene cluster, which is similar to that found for nifS of Anabaena sp. (34-36). Three nifS-like genes have been demonstrated also in the genome of Haemophilus influenzae, and a similar gene cluster occurs around the Hin1 gene (Table III) of the genome. Therefore, Eco1 and Hin1 probably participate in construction of the Fe-S clusters of iron-sulfur proteins, in the same manner as NIFS of A. vinelandii. Synechocystis sp. PCC6803 also contains three nifS-like genes. However, we found no such gene organization around it. Similarly, none of the cysteine sulfinate desulfinase genes and the third nifS-like gene (Eco2 in Table III) of E. coli form such gene clusters. The same is true for the other nifS-like genes of H. influenzae (Hin2 and Hin3). Therefore, these NIFS-like proteins probably have biochemical functions different from those of Eco1, Hin1, and A. vinelandii NIFS.

Table III. NIFSs and NIFS-like proteins Accession indicates the accession numbers. Length is the length of the protein. The corresponding sequence data are indicated in Footnotes d-j. Source Abbreviation Accession Length Ref. A. vinelandii Avi P05341a 402 (21) Azotobacter chroococcum Ach P23120a 396 (48) Azospirillum brasilense Abr U26427b 398 (49) Enterobacter agglomerans Eag X99694b 401 Klebsiella pneumoniae Kpn P053344a 397 (21) Rhodobacter sphaeroides Rsh Q01179a 387 (50) Rhodobacter capsulatus Rca Q07177a 384 (51) Anabaena sp. PCC7120 Asp P12623a 400 (35) Anabaena variabilis Ava2 U49859b 398 (52) Anabaena azollae Aaz L34879b 400 (36) Synechocystis sp. PCC6803 Ssp1 D64004b,d 420 (53) Synechocystis sp. PCC6803 Ssp2 D63999b,e 386 (53) Synechocystis sp. PCC6803 Ssp3 D90899b,f 391 (54) H. influenzae Rd Hin1 HI0378c 406 (55) H. influenzae Rd Hin2 HI1295c 437 (55) H. influenzae Rd Hin3 HI1343c 238 (55) E. coli K-12 Eco1 D90883b,g 404 (18) E. coli K-12 Eco2 D90811b,h 406 (19) E. coli K-12 CSD U295810b,i 401 (18) S. cerevisiae Sce P25374a 497 (15) Lactobacillus delbrueckii Lde P31672a 355 (56) B. subtilis Bsu P38033a 395 (14) Caenorhabditis elegans Cel U23139b,j 328 (57) Mycobacterium leprae Mle U00013b 418 Mycoplasma pneumoniae Mpn AE000034b 408 (58) Mycoplasma genitalium Mge U39716b 408 (59) a SWISS-PROT. b GenBankTM/EMBL. c TIGR microbial data base. d slr0077. e slr0387. f sll0704. g yzz0. h o320#17. i o401. j F13H8.9.

NIFS has been classified into the same group as aminotransferases of class V (37) and subgroup IV (38), which include serine-pyruvate aminotransferase (EC 2.6.1.51) and phosphoserine aminotransferase (EC 2.6.1.52), on the basis of sequence homology analysis. Isopenicillin-N-epimerase belongs to the same group as various PLP-dependent enzymes, other than aminotransferases (37, 38). It has therefore been suggested that NIFS and isopenicillin-N-epimerase evolved from the common ancestral protein for the aminotransferases of these classes (37, 38).

We have found that NIFS family proteins are classified into two groups, I and II, according to their sequence similarities. The two groups are clearly distinct from each other in the regions named a, b, c, and d (Fig. 3). Average sequence similarities of cysteine sulfinate desulfinase to Group I and II members were 23 and 37%, respectively. The similarity relationship among NIFS family proteins is shown in a phylogenetic tree (Fig. 4), which also indicates that the proteins are classified into two major groups. The proteins Cel, Lde, Bsu, and Ssp3 are far from the others, but are close to the members of Group I than to the Group II proteins.

We have shown that selenocysteine lyase from C. freundii is quite different from the pig liver enzyme in various physicochemical properties (11). The amino acid compositions of pig liver selenocysteine lyase (PIG), cysteine sulfinate desulfinase (CSD), A. vinelandii NIFS (Avi), and E. coli NIFS-like protein (Eco1) resemble each other, but are distinct from that of selenocysteine lyase from C. freundii (CFR) (Fig. 5). Therefore, the latter enzyme probably belongs to a different family of proteins.

The NIFS-like protein from E. coli and NIFS from A. vinelandii have common characteristics; both contain essential cysteinyl residues at the active sites. The thiol group presumably attacks as a nucleophile the sulfur atom of the substrate, cysteine, to form the intermediate, enzyme-bound cysteinyl persulfide (14, 20). By contrast, no cysteinyl residue of cysteine sulfinate desulfinase is essential for catalysis. The cysteine sulfinate desulfinase reaction is assumed to proceed through direct release of elemental selenium or sulfur atom from the substrate, selenocysteine or cysteine. It has been assumed that formation of the enzyme-bound cysteinyl persulfide is crucial to deliver sulfur atoms efficiently to iron-sulfur proteins. If this is the case, cysteine sulfinate desulfinase will not be related metabolically to the formation of Fe-S clusters, although sulfur atoms produced from cysteine by the enzyme are probably incorporated into iron-sulfur proteins with low efficiency in the same manner as observed for O-acetylserine sulfhydrylase A (EC 4.2.99.8), O-acetylserine sulfhydrylase B (EC 4.2.99.8), and -cystathionase (EC 4.4.1.8) (39). The fact that the K_m value of cysteine sulfinate desulfinase for cysteine is high also suggests that cysteine is not the physiological substrate of the enzyme.

The irreversible inactivation of PLP enzymes by -chloroalanine has been shown to proceed through modification of the enzyme-bound PLP with nascent -aminoacrylate formed from -chloroalanine (40-42). Cysteine sulfinate desulfinase catalyzes the same type of reaction as do selenocysteine lyase, aspartate -decarboxylase, and kynureninase. All these enzymes except cysteine sulfinate desulfinase are inactivated by -chloroalanine (10, 43, 44). Nascent -aminoacrylate is probably released from the active site of cysteine sulfinate desulfinase much more quickly than from those of the other enzymes. Alternatively, -aminoacrylate may be hydrolyzed quickly to pyruvate and ammonia, and so the enzyme can escape from modification with -aminoacrylate.

In mammals, cysteine is oxidized by cysteine dioxygenase (EC 1.13.11.20) to form cysteine sulfinic acid, which is decarboxylated to form hypotaurine by cysteine sulfinate decarboxylase (EC 4.1.1.29). cDNAs for cysteine dioxygenase (45) and cysteine sulfinate decarboxylase (46) were cloned and sequenced. We found no sequences similar to those of the cDNAs in the whole genomic sequence of E. coli K-12. If E. coli has a cysteine dioxygenase, it will have little sequence similarity to the mammalian enzyme. Alternatively, if no cysteine dioxygenase occurs in E. coli, the cysteine desulfination may be a side function of the enzyme with no metabolic relevance. Aspartate -decarboxylase and aspartate aminotransferase also use cysteine sulfinate as a good substrate and desulfinate it (29, 43, 47). Whatever the physiological function of cysteine sulfinate desulfinase is, this is the first enzyme in Group II whose catalytic function has been clarified (Fig. 4). Other proteins of this group probably have a similar catalytic function to cysteine sulfinate desulfinase. Cloning and expression of the Eco2 gene, the last nifS-like gene of E. coli mapped at 38.3 min (19) in the chromosome, and characterization of the gene product, are now being studied.