The NIFS Protein Can Function as a Selenide Delivery Protein in the Biosynthesis of Selenophosphate*

The NIFS protein from Azobacter vinelandii is a pyridoxal phosphate-containing homodimer that catalyzes the formation of equimolar amounts of elemental sulfur andl-alanine from the substrate l-cysteine (Zheng, L., White, R. H., Cash, V. L., Jack, R. F., and Dean, D. R. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 2754–2758). A sulfur transfer role of NIFS in which the enzyme donates sulfur for iron sulfur center formation in nitrogenase was suggested. The fact that NIFS also can catalyze the decomposition ofl-selenocysteine to elemental selenium andl-alanine suggested the possibility that this enzyme might serve as a selenide delivery protein for the in vitrobiosynthesis of selenophosphate. In agreement with this hypothesis, we have shown that replacement of selenide with NIFS andl-selenocysteine in the in vitroselenophosphate synthetase assay results in an increased rate of formation of selenophosphate. These results thus support the view that a selenocysteine-specific enzyme similar to NIFS may be involved as anin vivo selenide delivery protein for selenophosphate biosynthesis. A kinetic characterization of the two NIFS catalyzed reactions carried out in the present study indicates that the enzyme favors l-cysteine as a substrate compared with its selenium analog. A specific activity for l-cysteine of 142 nmol/min/mg compared with 55 nmol/min/mg forl-selenocysteine was determined. This level of enzyme activity on the selenoamino acid substrate is adequate to deliver selenium to selenophosphate synthetase in the in vitroassay system described.

The insertion of selenium into Se-dependent enzymes and Se-tRNAs requires the formation of the highly reactive, reduced selenium donor compound monoselenophosphate (1). The Escherichia coli selD gene product, selenophosphate synthetase, catalyzes the synthesis of monoselenophosphate from selenide and ATP (Scheme 1) (2). ATP ϩ HSe Ϫ ϩ H 2 0 3 H 2 SePO 3 Ϫ ϩ P i ϩ AMP SCHEME 1 Selenophosphate synthetase from E. coli has been characterized previously (2). The enzyme has a determined specific activity of 83 nmol/min/mg (k cat ϭ 3 min Ϫ1 ). The initial K m values reported for each substrate, ATP and selenide, were 0.9 mM and 46 M, respectively. Selenide is extremely oxygen labile, so the kinetics were repeated using more stringent anaerobic condi-tions. Under these conditions, the K m for selenide was calculated as 7.3 M, which is still above the range of 0.1-1 M which is the optimal concentration of selenite used in most growth media. Despite being below the K m value for one essential enzyme in the pathway, this concentration of selenide is sufficient for the biosynthesis of selenium-containing proteins and tRNAs.
The E. coli selenophosphate synthetase contains an essential cysteine, Cys 17 , within the glycine-rich sequence -Gly-Ala-Gly-Cys 17 -Gly-Cys-Lys-Ile-. Replacement of Cys 17 with serine results in complete loss of activity with selenide and ATP as substrates (3). Additionally, both a human form of the enzyme that contains threonine (4) in place of the essential Cys 17 in the E. coli enzyme and a Drosophila enzyme that has an arginine replacement (5) are inactive in the in vitro selenophosphate synthetase assay with selenide and ATP as substrates. The observed low specific activity of the E. coli selenophosphate synthetase and the inactivity of human and Drosophila homologs suggests that an essential unidentified component is required for the formation of selenophosphate.
In an early attempt to understand the mechanism of the incorporation of selenocysteine into proteins, 3 H-, 14 C-, and 75 Se-labeled L-selenocysteine samples were prepared. These labeled compounds were added to media of Clostridium sticklandii cultures to determine whether they were capable of being directly incorporated into selenoprotein A of the glycine reductase complex. 75 Se derived from [ 75 Se]selenocysteine was incorporated more efficiently into selenoprotein A than H 2 75 Se (generated by reduction of 75 SeO 3 2Ϫ by excess thiols in the incubation medium). There was no detectable incorporation of 14 C or 3 H into the selenocysteine amino acid of selenoprotein A (6). These results suggest an activated selenium derived from L-selenocysteine may be the optimal substrate for selenophosphate synthetase. Additionally, this suggests the involvement of a selenocysteine lyase protein in substrate formation.
Selenocysteine lyase enzymes have been isolated from both bacteria (7) and pig liver (8). This family of enzymes catalyzes the exclusive decomposition of L-selenocysteine to L-alanine and elemental selenium. They are very specific for their selenoamino acid substrate and do not utilize the sulfur analog Lcysteine as a substrate. Recently, Mihara et al. (9) have found that the N-terminal amino acid sequence of the pig liver selenocysteine lyase is similar to the amino acid sequence of the cysteine desulfurase protein NIFS.
The nifS gene product (NIFS) from the diazotrophic bacterium Azotobacter vinelandii has been previously isolated and characterized. The protein has been identified as a pyridoxal phosphate-containing homodimer that catalyzes the formation of equimolar amounts of elemental sulfur and L-alanine from the substrate, L-cysteine (10). It was reported that NIFS is also able to catalyze the removal of selenium from selenocysteine; a mechanism similar to the L-cysteine reaction was postulated (11). * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  In this work, we report the characterization of the NIFS catalyzed reaction with L-selenocysteine as substrate. We also report the use of NIFS and L-selenocysteine as a selenide delivery system in the selenophosphate synthetase assay.

Materials
A. vinelandii NIFS protein and pDB551 (nifS gene under control of the T7 promotor) were gifts from Dennis Dean, Virginia Polytechnic Institute. All buffers and reagents were prepared from the highest grade chemicals available.

Methods
Overexpression and Purification of NIFS-A. vinelandii NIFS was overexpressed in E. coli cells and purified by a procedure as described previously, with slight modifications (10). BL21 cells transformed with pDB551 were grown in a 10-liter fermentor containing ampicillin (50 mg/liter) at 37°C to an A 600 of 0.6. Cells were induced by the addition of lactose to a concentration of 0.4%. After 2.5 h, cells were harvested by centrifugation, frozen in liquid N 2 , and stored at Ϫ80°C until needed.
Preparation of Cell Extract-20 grams of E. coli BL21 cells containing overexpressed NIFS were thawed in 50 ml of 25 mM Tris⅐HCl (pH 7.4). Extract was prepared by passing the cell suspension once through a French pressure cell. Cell debris was removed by centrifugation at 15,000 rpm for 20 min. Streptomycin sulfate was added to the supernatant to a final concentration of 15%. After 15 min, the precipitate was sedimented by centrifugation at 15,000 rpm for 20 min. Ammonium sulfate was added to the supernatant to 45% of saturation, and the protein precipitate was resuspended in 10 ml of 25 mM Tris⅐HCl (pH 7.4). Phenyl-Sepharose Chromatography-Freshly dialyzed NIFS was applied to a phenyl-Sepharose CL-4B column equilibrated with the same buffer. Proteins were eluted with a linear gradient of 0.5-0 M ammonium sulfate in 25 mM Tris⅐HCl (pH 7.4) (500 ml). NIFS was eluted in fractions at the end of the gradient. These fractions were pooled, dialyzed, concentrated, and stored at Ϫ80°C.
Assays for NIFS Activity-All NIFS assays were performed anaerobically in a reaction mixture containing 50 mM Tricine⅐KOH 1 (pH 8.0), 4 mM DTT, 8 mM MgCl 2 , 50 mM KCl, and 0.1 mM Mgtitriplex (E. Merck, Darmstadt, Germany). The enzyme activity with L-selenocysteine was measured by determination of H 2 Se with lead acetate as described previously (8). The standard reaction was performed with 0.5 M NIFS using a molar turbidity coefficient of colloidal PbSe at 400 nm of 2.36 ϫ The cysteine desulfurase activity was determined by monitoring sulfide production from cysteine (13). Sulfide was measured by its conversion to methylene blue in a 1:1 mixture of 0.02 M N,N-dimethylp-phenylenediamine sulfate in 7.2 M HCl and 0.03 M FeCl 3 in 1.2 M HCl. After color development, the absorbance at 650 nm was measured, and sulfide concentration was determined based on a standard Na 2 S curve.
The coupled selenophosphate synthetase-NIFS protein assay was performed in the same reaction mixture described above with the following modifications: 2 mM ATP, 0.2 Ci [8-14 C]ATP,10 M selenophosphate synthetase, 5 M NIFS were additionally added, and reaction mixtures were incubated at 37°C. After 30 min, the reactions were terminated by the addition of 1.2 N HClO 4 followed by neutralization with KOH. The nucleotides in the supernatant solutions were separated chromatographically on cellulose-polyethyleneimine thin layer sheets developed in 1.0 M formic acid and 0.5 M LiCl. Nucleotide spots identified by UV quenching were cut out, and radioactivity was measured by liquid scintillation spectroscopy. 31 P NMR Spectroscopy-To detect reaction products of the coupled reaction between NIFS and selenophosphate synthetase, a coupled reaction was performed in an NMR tube under nitrogen for 3 h at 37°C. The reaction mixture (1 ml) contained 50 mM Tricine⅐KOH (pH 8.0), 4 mM DTT, 8 mM MgCl 2 , 2 mM ATP, 2 mM L-selenocysteine, 10 M selenophosphate synthetase, and 10 mM NIFS. The NMR spectra were obtained on a Bruker 500 MHz at 37°C as described previously (1).

RESULTS
Catalytic Properties and Substrate Specificity-As previously reported, NIFS is a pyridoxal phosphate-containing enzyme that catalyzes the formation of L-alanine and sulfur from L-cysteine as substrate (10). To confirm that NIFS can also use L-selenocysteine as substrate and to compare enzyme activities on the two amino acid substrates, two types of assays were carried out. NIFS cysteine desulfurase activity was determined by monitoring the production of H 2 S in the methylene blue assay. Cysteine desulfurase assays were performed in the presence of DTT to keep the sulfur product reduced to H 2 S. The kinetic constants observed for this reaction are reported in Table I. A specific activity of 142 nmol/min/mg was determined for the NIFS reaction with L-cysteine, as well as a k cat of 6 min Ϫ1 and a V max of 2.4 nmol/min. A K m for the substrate L-cysteine could not be determined because, at substrate concentrations as low as 10 M, the velocity was still maximal (Fig.  1).
In our assays, NIFS catalyzed the elimination of selenium from L-selenocysteine as monitored by the reaction of H 2 Se with lead acetate (8). The velocity of the NIFS catalyzed reaction with L-selenocysteine as the substrate exhibited a sub-  Reactions with each substrate were performed anaerobically at ambient temperature. Each reaction mixture contained 50 mM Tricine ⅐ KOH (pH 8.0), 4 mM DTT, 8 mM MgCl 2 , 50 mM KCl, 0.1 mM Mgtitriplex, and the appropriate concentration of substrate. The concentration of NIFS used in assays with L-cysteine was 0.5 M and with L-selenocysteine was 2 M. ND, not determined. strate dependence with a K m of 130 M (Fig. 2). The specific activity of 55 nmol/min/mg and turnover number of 2.4 min Ϫ1 for L-selenocysteine are both one-third lower than the activity with L-cysteine.
Inhibition of NIFS Cysteine Desulfurase Activity with L-Selenocysteine and Selenide-To determine whether L-selenocysteine binds to the same active site as L-cysteine, cysteine desulfurase assays were carried out in the presence of varying concentrations of L-selenocysteine. As shown in Fig. 3, cysteine desulfurase activity can be inhibited in a concentration-dependent manner by the addition of L-selenocysteine to the assay mixture. The addition of selenide (20 and 30 nmol) to the cysteine desulfurase assay mixture did not inhibit the cysteine desulfurase activity (data not shown).
Selenophosphate Synthetase and NIFS Coupled Assays-To test the ability of the NIFS protein to serve as a selenide delivery protein for selenophosphate synthetase, assays were performed in the presence of both NIFS and L-selenocysteine. As previously reported, selenophosphate synthetase catalyzes the synthesis of monoselenophosphate from selenide and ATP (2), additional reaction products include AMP and orthophosphate.
When selenide is replaced with L-selenocysteine, selenophosphate is not formed, as determined by the lack of ATP hydrolysis (Table II). However, replacement of selenide in the assay with L-selenocysteine in the presence of added NIFS protein results in the formation of AMP (Fig. 4). The amount of AMP formed is dependent on the concentration of NIFS added. In the coupled assay, when concentrations of NIFS are lower than the selenophosphate synthetase concentration, AMP production is lower as compared with the assay using selenide. However, if the concentration of NIFS is equal to or double the concentration of selenophosphate synthetase, the production of AMP is increased to almost twice the amount formed in the selenidedependent assay.
In addition to monitoring the production of AMP in the coupled assay, we also determined both the amount of H 2 Se produced and its rate of formation by NIFS in the coupled enzyme system. When selenophosphate synthetase is present, the amount of H 2 Se produced is increased by 35%. The rate of elimination of selenium from L-selenocysteine increases, as well, from 6.1 to 9.5 nmol/min (Fig. 5). The presence of selenophosphate synthetase in the assay may activate NIFS toward L-selenocysteine. In contrast, in standard selenophosphate synthetase assays performed, with the addition of NIFS, using H 2 Se as substrate, no increase in either   the rate or the amount of AMP product formed was observed (data not shown).
Analysis of Reaction Products by 31 P NMR-To directly observe the products generated by selenophosphate synthetase in the presence of NIFS and L-selenocysteine, reaction products were analyzed by 31 P-NMR. As shown in Fig. 6, products formed in the coupled selenophosphate synthetase and NIFS assay from ATP and L-selenocysteine produced 31 P chemical resonances expected for AMP, inorganic phosphate, and selenophosphate. Additional spectra were obtained for NIFS ϩ ATP ϩ L-selenocysteine as well as selenophosphate synthetase ϩ L-selenocysteine ϩ ATP. Both reactions did not produce the expected 31 P chemical resonance for selenophosphate (data not shown). DISCUSSION Because of the chemical similarity between selenium and sulfur, enzymes that metabolize sulfur compounds frequently demonstrate similar activities with the selenium analogs of their natural substrates (9). It has been shown in recent studies with cysteine sulfinate desulfinase isolated from E. coli that the enzyme exhibits both selenocysteine lyase and cysteine desulfurase activities (9). Moreover, previous work on the A. vinelandii NIFS protein demonstrated that NIFS can act on both L-cysteine and L-selenocysteine (10). In the present paper, we characterized in more detail the enzymatic decomposition of L-selenocysteine by the NIFS protein.
Our studies confirm the work by Zheng et al., 1993 showing that the protein can utilize both L-cysteine and L-selenocysteine as substrates (10). Our kinetic values indicate the enzyme favors L-cysteine as a substrate over its selenium analog (Table  I). This is not surprising because NIFS has been shown to have a biologically significant role in the biosynthesis of Fe-S clusters (11). A specific activity for L-cysteine of 142 nmol/min/mg, compared with 55 nmol/min/mg for L-selenocysteine was deter- mined. An interesting aspect of our characterization of NIFS was that even at L-cysteine concentrations as low as 10 M, the rate of the cysteine desulfurase activity remains at the V max . The independence of the determined velocity of the cysteine desulfurase activity on substrate concentration, indicating a very high desulfurase rate, makes the determination of K m more difficult. Our determined specific activity as well as the inability to determine a K m for L-cysteine differs from the previous results (10,11) in which a specific activity of 70 nmol/min/mg and a K m of 75 M for the substrate L-cysteine were reported. These investigators suggest NIFS activity is highly sensitive to both pH and temperature, which affects the determined K m and specific activity values (11).
Based on the studies of Zheng et al. (11) and Flint (14) with the E. coli NIFS protein, it was proposed that in the NIFS catalyzed reaction a sulfhydryl group, from an active site cysteine, acts as a nucleophile in the mobilization of sulfur from the L-cysteine substrate. Flint suggests that after sulfur is transferred to NIFS, creating a persulfide on the active site cysteine, catalysis stops or slows down until this sulfane sulfur is removed by an uncharacterized pathway. This would account for the relatively low turnover number for the enzyme. When assays are performed in the presence of DTT, as in this report, the rate-limiting sulfane sulfur undergoes nucleophilic displacement by DTT forming H 2 S (14). Although the formation of a perselenide (R-S-Se Ϫ ) in the NIFS reaction with L-selenocysteine has not been investigated in the present study, Zheng et al. 1994 (11) have proposed that the mechanism of selenocysteine lysis by NIFS is mechanistically similar to the cysteine desulfurase reaction. This is suggested by the requirement of Cys 325 that was determined to be required for both activities (11) (Scheme 2).
Because of the high reactivity of selenide within a biological system, the overall cellular concentrations must remain relatively low. The availability of selenide delivery proteins could effectively raise the concentration of selenide in a localized region of the cell where it could be rapidly metabolized by a second enzyme. Although concentrations of free selenide may be low, selenium is present in cells at higher concentrations in other less reactive forms such as selenomethionine. For instance, bacteria, plants, and yeast have been found to synthesize the selenium-containing amino acid selenomethionine. The presence of selenium in place of the sulfur of methionine offers no altered biological function. Selenomethionine is abundant in plants, and when consumed organisms can utilize this form as readily as the natural occurring amino acid methionine by incorporating it freely into proteins (15). In addition to becoming incorporated into cellular proteins, selenomethionine can be catabolized to selenocysteine by the transsulfuration pathway (16). Selenocysteine can then be metabolized by the pyridoxal phosphate-dependent selenocysteine lyase proteins to release elemental selenium. In the presence of reducing compounds, elemental selenium is reduced to selenide. This could provide an effective mechanism to maintain concentrations of selenide within the cell.
In addition to the formation of selenide from L-selenocysteine, the nonenzymatic reaction of selenite with thiols, such as glutathione or cysteine, is believed to be a significant pathway for the incorporation of selenium into living cells. The reduction to elemental selenium requires the combination of thiols to selenite in a 4:1 stoichiometric ratio. Further reduction to selenide requires two additional equivalents of thiols (17).
Our results clearly show the use of a selenide delivery protein can increase the turnover rate of an enzyme which utilizes selenide as a substrate (Figs. 4 and 6). The exact mechanism responsible for the increase in rate is not yet known. Possibilities include higher local concentrations of selenide in the vicinity of selenophosphate synthetase. Perhaps both proteins interact in solution, allowing direct transfer of selenide product from NIFS to selenophosphate synthetase. The precise chemical form of selenium produced in the reaction by NIFS may be an important determinant. Selenide can exist in several forms analogous to sulfane sulfur species, depending on the reaction conditions used. Thus, the product formed from NIFS may be a more optimal substrate for selenophosphate synthetase. Future experiments are now being planned to answer these questions.
The essential role of NIFS in the pathway for the biosynthe-SCHEME 2 sis of FeS clusters cannot be disputed as one of its primary biological roles in vivo. In the in vitro coupled assay, NIFS additionally can catalyze the elimination of selenium from Lselenocysteine as well as function as a selenide delivery protein. However, in vivo concentrations of sulfur containing compounds are on the order of a thousand times greater than their selenium analogs (18), and thus proteins such as NIFS will preferentially bind to and metabolize L-cysteine over the corresponding selenium analog L-selenocysteine. Hence, for effective selenide delivery systems to operate, enzymes which are specific for selenide compounds are required. Among such possible enzymes are selenocysteine lyases that have been identified in several organisms. Although kinetic characterization of these enzymes reveals they have relatively high K m values in vitro for their selenium containing substrates, the assay conditions used may not represent true cellular conditions where the K m may actually be lower. These enzymes may play an important role in the regulation of selenium availability in biological systems. Thus their ability in vitro to metabolize selenocysteine to alanine and selenium or selenide suggests that, in vivo, they may function as selenium delivery proteins analogous to sulfur delivery enzymes.