Isotope exchange studies on the Escherichia coli selenophosphate synthetase mechanism

Abstract

Selenophosphate synthetase, the Escherichia coli selD gene product, is a 37-kDa protein that catalyzes the synthesis of selenophosphate from ATP and selenide. In the absence of selenide, ATP is converted quantitatively to AMP and two orthophosphates in a very slow partial reaction. A monophosphorylated enzyme derivative containing the γ-phosphoryl group of ATP has been implicated as an intermediate from the results of positional isotope exchange studies. Conservation of the phosphate bond energy in the final selenophosphate product is indicated by its ability to phosphorylate alcohols and amines to form O-phosphoryl- and N-phosphoryl-derivatives. To further probe the mechanism of action of selenophosphate synthetase, isotope exchange studies with [8-¹⁴C]ADP or [8-¹⁴C]AMP and unlabeled ATP were carried out, and ³¹P NMR analysis of reaction mixtures enriched in H₂¹⁸O was performed. A slow enzyme-catalyzed exchange of ADP with ATP observed in the absence of selenide implies the existence of a phosphorylated enzyme and further supports an intermediary role of ADP in the reaction. Under these conditions ADP is slowly converted to AMP. Incorporation of ¹⁸O from H₂¹⁸O exclusively into orthophosphate in the overall selenide-dependent reaction indicates that the β-phosphoryl group of the enzyme-bound ADP is attacked by water with liberation of orthophosphate and formation of AMP. Based on these results and the failure of the enzyme to catalyze an exchange of labeled AMP with ATP, the existence of a pyrophosphorylated enzyme intermediate that was postulated earlier can be excluded.

Selenium occurs as an integral component, usually as a selenocysteine moiety, of a growing list of proteins in both prokaryotes and eukaryotes. In prokaryotes, specific incorporation of selenocysteine depends on the products of four genes, selA, selB, selC, and selD (1–3). The selC gene product is a selenocysteyl-tRNA^SEC, anticodon UCA, that cotranslationally delivers selenocysteine at certain in-frame UGA codons (4). Initially this tRNA is charged with serine, which then is converted to aminoacrylyl-tRNA by the selA gene product, selenocysteine synthase (5). Addition of a reactive form of selenium across the 2,3 double bond forms selenocysteyl-tRNA (6). The reactive selenium compound in this reaction is selenophosphate (7), which is synthesized by the selD gene product, selenophosphate synthetase (8). Analogous selD gene products have been identified in eukaryotes (9–12), and in one of these (12) a selenocysteine residue is encoded in place of the essential Cys-17 in the Escherichia coli enzyme.

In prokaryotes, selenophosphate also serves as the reactive selenium donor in a substitution reaction in which the sulfur of a 2-thiouridine residue in the anticodons of certain tRNAs is replaced by selenium, forming a 2-selenouridine residue (13, 14).

Selenophosphate synthetase from E. coli catalyzes the synthesis of selenophosphate, AMP, and orthophosphate in a 1:1:1 ratio from selenide and ATP (8). The oxygen-labile selenophosphate product is derived from the γ-phosphoryl group of ATP, and the β-phosphoryl group is released as orthophosphate (8, 15). The enzyme has a K_m for ATP of 0.9 mM, an apparent K_m for selenide of 7.3 μM, and a V_max of 83 nmol/min per mg protein. The only product of the reaction that acts as an inhibitor is AMP, with a K_i of 170 μM. In the presence of 1.0 mM ATP, concentrations of ADP up to 2.0 mM and P_i up to 20 mM do not inhibit enzyme activity. In the absence of selenide, selenophosphate synthetase quantitatively converts ATP to AMP and two orthophosphates (8). Based on the ability of the enzyme to catalyze this partial reaction and the marked differences in inhibitory properties of the products of the overall reaction, it was postulated that the catalytic mechanism of selenophosphate synthetase involved a multi-step reaction pathway with an enzyme-pyrophosphoryl intermediate. However, attempts to trap or detect a ³²P-labeled enzyme-pyrophosphoryl intermediate by using [γ-³²P]ATP and [β-³²P]ATP were unsuccessful (16). Under enzyme turnover conditions, enzyme could be recovered that was labeled with ³²P from [γ-³²P]ATP but no ³²P from [β-³²P]ATP remained with the enzyme, excluding the presence of a detectable enzyme-pyrophosphoryl intermediate. Furthermore, no new resonance peak corresponding to an enzyme-phosphoryl intermediate could be detected by using ³¹P NMR (8, 16), and no back reaction of labeled AMP with ATP was detected (8).

Recent experiments using positional isotope exchange (PIX) methods have provided clues to the catalytic mechanism of selenophosphate synthetase (17). In these experiments all four oxygens of the γ-phosphoryl group of ATP were labeled with oxygen-18. After incubation of the labeled ATP with the enzyme in the absence of selenide, scrambling of the oxygen-18 label at the β,γ-bridge position with the two oxygen-16 molecules at the β-nonbridge positions was observed. The rate of this exchange was catalytically competent. These results are consistent with an initial cleavage by the enzyme of the γ-phosphoryl group of ATP to form an enzyme-phosphoryl intermediate. Rotation of the β-phosphoryl group followed by reversal of the cleavage reaction accounts for the isotope exchange. In similar experiments in which the oxygens of the β-phosphoryl group were labeled with ¹⁸O, no isotope exchange was detected. In the case of an enzyme-pyrophosphoryl intermediate, exchange of β-phosphoryl group oxygens with the α-phosphoryl group is expected.

To obtain further evidence for an initial phosphorylated enzyme intermediate, isotope exchange studies have been carried out to determine if exchange of [8-¹⁴C]ADP with ATP could be detected. Furthermore, to investigate the role of water in the catalytic mechanism, the incorporation of ¹⁸O from H₂¹⁸O into products was studied by taking advantage of the measurable isotope resonance shift of the ³¹P NMR spectra that is observed when ¹⁸O is bound to a phosphorous atom. These studies have been utilized to indicate whether the ¹⁸O is incorporated into the AMP or orthophosphate.

MATERIALS AND METHODS

Materials.

The following chemicals were purchased from the indicated sources: [8-¹⁴C]ATP (49 mCi/mmol; 1 Ci = 37 GBq) was from ICN, [8-¹⁴C]ADP (56.7 mCi/mmol) was from DuPont/NEN, and [8-¹⁴C]AMP was from DuPont/NEN (584.5 mCi/mmol) or American Radiolabeled Chemicals (48 mCi/mmol). Na₂ADP, KADP, and Na₂ATP were from Boehringer Mannheim, deuterium oxide was from Isotec, H₂¹⁸O was from either Isotec or Cambridge Isotope Laboratories, MgTitriplex (MgEDTA) was from Merck, and tetrabutylammonium phosphate was from Sigma. Ion exchange and molecular sieve reagents for chromatography were from Pharmacia LKB. HPLC columns were from Tosohaas (DEAE; Montgomeryville, PA) or Jones Chromatography (C¹⁸; Lakewood, CO).

Purification of E. coli Selenophosphate Synthetase.

The 37-kDa selenophosphate synthetase was purified as described with some modifications (8). The second DEAE-Sepharose column was replaced with a butyl-Sepharose column (2.5 × 10 cm) equilibrated in 50 mM N-tris(hydroxymethyl)methylglycine (Tricine)⋅KOH (pH 7.2), 2.0 mM DTT, 0.1 mM MgTitriplex, and 1.0 M (NH₄)₂SO₄. Enzyme-containing fractions from the DEAE-Sepharose column were pooled, and the solution (about 150 ml), which contained about 180 mg of protein, was made 1.0 M in (NH₄)₂SO₄ before loading onto the butyl-Sepharose column. Proteins were eluted with a linear gradient of 400 ml equilibration buffer and 400 ml buffer B (50 mM Tricine⋅KOH, pH 7.2/2.0 mM DTT/0.1 mM MgTitriplex). This was followed by size separation on the molecular sieve column, and for the final purification step a Tosohaas DEAE-5PW column (21.5 mm i.d. × 15 cm) was used. The enzyme was eluted from the DEAE HPLC column with a linear gradient from 0.1 M Tricine⋅KOH (pH 7.2) with 2.0 mM DTT and 0.1 mM MgTitriplex (buffer A) to buffer A containing 0.3 M potassium chloride. A HP1050 HPLC at a flow rate of 3.0 ml/min was used. The protein concentration was determined by Bio-Rad protein assay.

Enzyme Assays.

Selenophosphate synthetase was assayed as described (8) with one modification. Instead of separating the adenosine nucleotides on polyethylenimine-cellulose thin layer sheets, they were separated by ion-pair HPLC chromatography as reported (18). After stopping the reaction by addition of HClO₄, 50 μl aliquots of the neutralized supernatant solutions were injected onto an Apex octadecyl 5-μm column (4.6 mm × 15 cm) for separation of the nucleotides. Radioactivity in each nucleotide fraction was determined by liquid scintillation spectroscopy, and nucleotide concentrations were determined by absorbancy measurements at 260 nm. Isotope exchange experiments and ADP or ATP hydrolysis experiments were carried out in the same manner in the absence of selenide.

Assay to Detect Adenylate Kinase Contamination.

Because isotope exchange experiments involved high enzyme concentrations and long incubation times, it was necessary to ensure that any traces of contaminating adenylate kinase had been removed. In addition to determining selenophosphate synthetase activity with the standard assay, enzyme preparations also were assayed for adenylate kinase activity by using a coupled assay with lactate dehydrogenase and pyruvate kinase (19).

Isotope Exchange Assays.

For measuring exchange between [8-¹⁴C]AMP and ATP, reaction mixtures (100 μl) contained 0.1 M Tricine⋅KOH (pH 7.2), 2.0 mM DTT, 3.0 mM MgCl₂, 1.0 mM [8-¹⁴C]AMP (2 μCi), 0.5 mM ATP, and 37 μM selenophosphate synthetase. Reaction mixtures were incubated at 37°C from 1 to 24 h and analyzed as in the standard enzymatic assay.

To detect enzyme catalyzed exchange of ADP with ATP, reaction mixtures contained 0.1 M Tricine⋅KOH (pH 7.2), 2.0 mM DTT, 3.0 mM MgCl₂, 1.0 mM [8-¹⁴C]ADP (2 μCi), 0.5 mM ATP, and 37 μM selenophosphate synthetase. Reaction mixtures were incubated at 37°C for 1–24 h.

Hydrolysis of ADP or ATP in the Absence of Selenide.

To measure the decomposition of ATP, reaction mixtures containing 0.1 M Tricine⋅KOH (pH 7.2) 2.0 mM DTT, 3.0 mM MgCl₂, 1.5 mM [8-¹⁴C]ATP (1 μCi), and either 160 μM or 37 μM selenophosphate synthetase were incubated at 37°C for 24 h. For the hydrolysis of ADP, reaction mixtures containing 0.1 M Tricine⋅KOH (pH 7.2), 2.0 mM DTT, 3.0 mM MgCl₂, 1.0 mM [8-¹⁴C]ADP (2 μCi), and 37 μM selenophosphate synthetase were incubated at 37°C for 4–24 h, and the rate of AMP formation was determined. The rate of AMP production at a higher ADP concentration was determined by using reactions mixtures containing 0.1 M Tricine⋅KOH (pH 7.2), 2.0 mM DTT, 10 mM KADP, 15 or 30 mM MgCl₂, and 30 or 37 μM selenophosphate synthetase. The AMP production was measured after 1, 3, 5, 7, and 24 h at 37°C.

³¹P NMR Spectroscopy.

Reactions were carried out in H₂¹⁸O both in the absence and presence of selenide to determine the role of water in the reaction mechanism. Reaction mixtures were incubated at 37°C for 1–24 h and quenched at specific time points by addition of 0.5 M EDTA to a final concentration of 83 mM. Reaction mixtures in the absence of selenide contained 0.1 M Tricine⋅KOH (pH 7.2), 5.0 mM Na₂ATP, 10 mM MgCl₂, 26 μM selenophosphate synthetase, and 10% D₂O. After quenching, the calculated oxygen-18 enrichment was 67%. Control experiments in unenriched water were also carried out in the absence of enzyme. For experiments in the presence of selenide, the same reaction mixture components supplemented with 1.5 mM HSe⁻ and the usual catalytic level, 5 μM, of selenophosphate synthetase were used. All solutions were thoroughly degassed and sealed to maintain anaerobic conditions. The calculated oxygen-18 enrichment after quenching was 71%. A control experiment was also done in the absence of enzyme. One-dimensional ³¹P NMR spectra were run on a Bruker Instruments (Billerica, MA) AMX 600 (measurements conducted at 242.94 MHz). Spectra were acquired at 23°C by using 16K data points and zero-filled to 32K.

RESULTS

PIX studies have indicated that the catalytic mechanism of selenophosphate synthetase involves an enzyme-phosphoryl intermediate in which the γ-phosphoryl group of ATP is bound to the enzyme in ester linkage (17). This reaction can occur in the absence of selenide. To obtain further evidence for this putative intermediate, isotope exchange studies were carried out by using conditions similar to those used in the PIX experiments.

Exchange of [8-¹⁴C]AMP with ATP.

To ensure that adenylate kinase, which equilibrates AMP, ADP, and ATP, was not a significant contaminant in purified selenophosphate synthetase preparations used in the present studies, exchange between [8-¹⁴C]AMP and unlabeled ATP was measured in the absence of selenide. After a 24-h incubation with a high amount of enzyme (30 μM), only about 4% of the radioactivity of the added [8-¹⁴C]AMP was found in ADP and ATP (Table 1, Fig. 1). For various enzyme preparations tested, the specific radioactivity of the ADP produced was less than half that of the AMP. Because the specific radioactivity of ADP resulting from adenylate kinase activity would be one-half that of the AMP due to dilution from unlabeled ATP, the small amount of total radioactivity found in the ADP most likely did result from a trace amount of adenylate kinase contamination (Table 1, Fig. 1). However, this small contribution did not significantly affect the results of the various isotope exchange experiments.

Table 1.

Exchange between [8-¹⁴C]AMP and ATP

Enzyme	AMP	ADP	ATP
Total cpm recovered
0	3,741,600  (99.6)	9,580  (0.3)	4,520  (0.1)
+	3,131,560  (95.5)	77,600  (2.3)	69,440  (2.1)
Specific radioactivity
0	35,660	1,564	68
+	29,289	10,448	1,886

Values are the average of two 24-h incubations. Rows 1 and 2 give total cpm recovered and % cpm (in parenthesis) for each nucleotide. Rows 3 and 4 give specific radioactivity of each nucleotide. Reaction mixtures are described in Materials and Methods. 

Figure 1.

HPLC chromatographic analysis of [8-¹⁴C]AMP exchange with ATP. Reaction mixtures, described in Materials and Methods, were incubated for 24 h: (A) minus enzyme control, and (B) plus 37 μM selenophosphate synthetase.

Exchange of [8-¹⁴C] ADP with ATP.

To determine if there is enzyme-catalyzed exchange of ADP and ATP, which would indicate that ADP is an intermediate in the catalytic mechanism, selenophosphate synthetase was incubated with [8-¹⁴C] ADP and unlabeled ATP in the absence of selenide. After a 24-h incubation, 11% of the recovered radioactivity was found in the ATP fraction (Table 2). No exchange was observed in the absence of enzyme. In addition to the enzyme catalyzed exchange between ADP and ATP, a significant amount of AMP was also produced (Fig. 2). After 24 h, 32% of the recovered radioactivity was found in the AMP fraction. The specific radioactivity of the AMP was less than that of the added free ADP, indicating that it was partially derived from enzyme-bound ADP formed from unlabeled ATP. Under these conditions, the rate of AMP production appeared to be linear over the 24-h incubation period.

Table 2.

Exchange between [8-¹⁴C]ADP and ATP

Time, h	AMP	ADP	ATP
Total cpm recovered
4	250,920  (5.7)	3,937,820  (91)	139,140  (3.2)
7	414,840  (10)	3,377,060  (85)	210,800  (5.2)
24	1,370,740  (32)	2,461,700  (57)	472,920  (11)
24*	74,340  (1.7)	4,349,100  (97)	43,800  (1.0)
Specific radioactivity
4	18,768	43,980	2,460
7	24,306	42,404	3,879
24	29,847	41,499	8,933
24*	9,373	45,203	776

Rows 1–4 give total cpm recovered and % cpm (in parenthesis) for each nucleotide. Rows 5–8 give specific radioactivity of each nucleotide. Values are the average of two experiments. Reaction mixtures are described in Materials and Methods. 

^*Control experiment in the absence of enzyme. 

Figure 2.

HPLC chromatographic analysis of [8-¹⁴C]ADP exchange with ATP. Reaction mixtures, described in Materials and Methods, were incubated for 24 h: (A) minus enzyme control, and (B) plus 37 μM selenophosphate synthetase.

Previous experiments have shown that in the absence of selenide 1.5 mM [8-¹⁴C]ATP is quantitatively converted to AMP and two orthophosphates by 160 μM selenophosphate synthetase in 22 h (8). These experiments were repeated by using two different enzyme levels, 160 μM and 37 μM, the latter being the concentration used in the exchange experiments. As before, at the high enzyme concentration, 99% of the recovered radioactivity was found in the AMP fraction (Table 3). At the lower selenophosphate synthetase level (37 μM), 9.3% of the recovered radioactivity was found in AMP and 9.3% was found in ADP, indicating that ADP is an intermediate in the enzymatic hydrolysis of ATP to AMP. The remaining radioactivity (81.4%) was in residual ATP.

Table 3.

Hydrolysis of [8-¹⁴C]ATP in the absence of selenide

Enzyme, μM	AMP	ADP	ATP
0	7,420  (0.4)	69,820  (3.3)	1,956,260  (96.3)
37	186,640  (9.3)	185,880  (9.3)	1,627,280  (81.4)
160	1,639,320 (99.2)	2,440  (0.2)	10,520  (0.6)

Reaction mixtures described in Materials and Methods were incubated for 24 h. Data are given as cpm recovered and % total cpm recovered (in parenthesis). It was shown previously (8) that in the absence of enzyme ATP is stable under similar conditions. 

Hydrolysis of ADP.

The hydrolysis of ADP by selenophosphate synthetase in the absence of selenide was studied further by measuring the rate of AMP production with ADP as the sole nucleotide substrate. At 1.4 mM ADP, the rate of AMP production (0.019 mM/h) was similar to that observed in the ADP⋅ATP exchange experiment (0.015 mM/h), in which 1.0 mM ADP was used. These small differences in rate could be caused by the different initial concentrations of ADP, 1.4 mM and 1.0 mM. The presence or absence of ATP may also have an effect on the reaction rate because, in the exchange experiment, either ADP or ATP could bind to the enzyme. When ATP binds, an extra hydrolysis step would be required to synthesize AMP. Although the rate of AMP production in the ADP⋅ATP exchange experiment appeared linear over the 24-h incubation, when 1.4 mM ADP is the sole substrate, the reaction progression appeared slightly nonlinear. When 10 mM ADP is used as the sole substrate, the rate of AMP production shows a nonlinear progression with an apparent lag phase. However, the reaction never appears to approach linearity before the AMP production decreases, most likely caused by the decrease in available ADP (Fig. 3). Attempts were made to overcome the nonlinear reaction progression by adding 2.0 mM AMP to the reaction mixture, but no change was observed (data not shown).

Figure 3.

Time course of production of AMP with 10 mM ADP as substrate. Reaction mixtures described in Materials and Methods contained 30 μM selenophosphate synthetase and 15 mM MgCl₂ (line A), 37 μM selenophosphate synthetase and 30 mM MgCl₂ (line B), or 37 μM selenophosphate synthetase and 15 mM MgCl₂ (line C).

³¹P NMR Spectroscopy.

To determine the role of water in the catalytic mechanism, reactions were carried out in H₂¹⁸O. Spectra collected from reactions carried out in the absence of selenide demonstrate unambiguously that the ¹⁸O is exclusively incorporated in the orthophosphate product (Fig. 4). The observed isotope shift of 0.023 ppm to higher field for the ¹⁸O-containing orthophosphate relative to the unenriched form is in agreement with previously reported results (20). After a 24-h incubation, the ratio of the integrated peak areas for unenriched to ¹⁸O-enriched orthophosphate is 1.0:1.7, or 63% enrichment. This value corresponds closely to the calculated ¹⁸O enrichment of 67% for the reaction mixture. The area of the AMP peak was approximately half that of the orthophosphate, indicating that in the absence of selenide, the enzyme-phosphoryl intermediate is hydrolyzed by the solvent water, resulting in the production of two enriched orthophosphates from each ATP. The ratio of the integrated peak areas of the orthophosphate to ATP were 1.5:7.8, corresponding to approximately an 8% conversion of ATP to orthophosphate. A small amount of ADP (about 3% of the area of the ATP peak) accumulated in this experiment. Spectra obtained from experiments that included selenide also gave similar results. In this case, after a 7-h incubation, the areas of the AMP and orthophosphate peaks were 10% and 20% of the ATP peaks, respectively, corresponding to a 10% conversion of ATP to AMP and two orthophosphates. Under these conditions, the extremely oxygen-labile selenophosphate that was produced was degraded to elemental selenium and orthophosphate on exposure to air during the quenching of the reaction with EDTA. Reaction of the released phosphoryl group of selenophosphate with solvent water, thus resulted in the production of the additional ¹⁸O-enriched orthophosphate. The results of these experiments indicate that both in the absence and presence of selenide, hydrolysis of ADP probably occurs through a nucleophilic attack by water on the β-phosphoryl group of ADP.

Figure 4.

³¹P NMR spectrum of a reaction mixture incubated 24 h in [¹⁸O]water in the absence of selenide. Experimental conditions are described in Materials and Methods. The two resonances at 2.53 and 2.51 ppm (expanded in the lower Inset) are the ¹⁶O and ¹⁸O orthophosphate, respectively. The upper Inset represents orthophosphate in a 24-h reaction mixture in unenriched water. The resonances at −5.82, −10.47, and −21.08 ppm correspond to the γ-, α-, and β-phosphoryl groups of ATP, respectively. The resonance at 4.0 ppm corresponds to the AMP, and the resonances at −6.07, and −10.14 ppm to the β- and α-phosphoryl groups of ADP, respectively. The peak at −7.00 ppm is an unidentified impurity.

DISCUSSION

Although the products of the selenophosphate synthetase reaction and their inhibition patterns suggest on the one hand, a catalytic mechanism involving an enzyme-pyrophosphoryl intermediate, attempts to trap or detect this intermediate have been unsuccessful. On the other hand, data obtained using PIX methods have provided evidence for a phosphoryl-enzyme intermediate (17). These data indicate that, in the absence of selenide, there is an initial reaction of enzyme with the γ-phosphoryl group of ATP to form an enzyme-phosphoryl intermediate and ADP.

In support of the PIX data, isotope exchange experiments reported here indicate that, in the absence of selenide, selenophosphate synthetase catalyzes an exchange reaction between free [8-¹⁴C]ADP and ATP, with subsequent production of AMP. After a 24-h incubation, 11% of the radiolabel was found in the ATP, whereas the AMP product contained 32% of the radiolabel. The specific radioactivity of the AMP, 28% lower than that of the ADP, indicates that the majority of the AMP was produced by hydrolysis of the labeled ADP, and a small portion was derived from unlabeled ATP. At extremely high enzyme concentrations (160 μM; 5.8 mg/ml), selenophosphate synthetase hydrolyzes ATP completely to AMP, or at lower enzyme concentrations (37 μM), to both ADP and AMP. These results indicate that, in the absence of selenide, ADP is an intermediate in the hydrolysis of ATP to AMP.

The rate of exchange between the oxygen-18 label of the γ-phosphoryl group and the oxygen-16 of the β,γ-bridge position of ATP determined in the PIX experiments was 34 nmol/min per mg compared with a rate of 83 nmol/min per mg for the overall reaction (17). Thus the rate of exchange measured in this manner appears catalytically competent. Although the rate of exchange between [8-¹⁴C]ADP and ATP was shown to be enzyme-dependent, it is extremely slow (0.027 nmol/min per mg) compared with the rate of the overall enzymatic reaction. However, unlike the PIX studies, the isotope exchange studies are dependent on the rate of association and dissociation of free ADP. Because ADP has never been detected in significant amounts as a free intermediate in the overall reaction (in ³¹P NMR studies the ADP peak area was 3% of the ATP peak area after a 7-h incubation) and is not an effective inhibitor of the enzyme (up to 2.0 mM shows no inhibition), poor binding or dissociation could explain the slow rate of exchange seen in these experiments. To determine if the slow exchange rate was due to poor binding of ADP, attempts were made to determine the K_m for ADP, which led to the discovery that AMP production from ADP does not follow a linear progression but resembles an autocatalytic reaction. Addition of one of the products, AMP, to the reaction mixtures had no effect on the reaction progression.

Alternatively, the observed lag phase could result from slow formation of an enzyme–ADP complex, resulting in very slow attainment of steady state conditions. Furthermore, the high concentration of MgCl₂ used in the hydrolysis experiment may have affected enzyme activity. Previous studies showed that inhibition of enzyme activity was observed at a 10:1 ratio of MgCl₂ to ATP at 1.5 mM ATP (21). Although a MgCl₂ to ADP ratio of 1.5:1 was used in the ADP hydrolysis experiments, the concentrations of both substrates were extremely high, which may have affected enzyme activity. It was previously reported that selenophosphate synthetase may have a second metal-binding site. This site has a high affinity for Zn²⁺ which at micromolar concentrations inhibits enzyme activity. Mn²⁺ also binds at this site and at 0.1 mM inhibits enzyme activity by about 60% (22). Thus, at the high MgCl₂ concentrations used in the hydrolysis experiments, binding of Mg²⁺ at this site might inhibit enzyme activity.

The PIX and isotope exchange data indicate that the first step in the selenophosphate synthetase catalyzed reaction is a nucleophilic attack by the enzyme on the γ-phosphoryl group of ATP to form an enzyme-phosphoryl intermediate with ADP still bound at the active site (Scheme S1). Preliminary attempts to demonstrate the putative phosphorylated enzyme intermediate by ³¹P NMR spectroscopy have not been successful even though high enzyme concentrations (500 μM) were used (8, 16). This may be due to a high rate of conversion of the intermediate, preventing accumulation to the extent required for detection by NMR. Based on the ³¹P NMR data from the H₂¹⁸O experiments, hydrolysis of the enzyme-bound ADP involves a nucleophilic attack by a water molecule on the β-phosphoryl group of ADP resulting in the formation of AMP and orthophosphate (Scheme S1). The slow rate of hydrolysis of ADP reported here could partially explain the low catalytic activity for the overall reaction, although the rate of association of free ADP with enzyme could be a significant variable. These data suggest a catalytic mechanism in which an enzyme-phosphoryl intermediate is attacked by selenide to form the selenophosphate product, whereas the ADP intermediate that remains enzyme-bound is hydrolyzed to orthophosphate and AMP by a solvent water molecule. In the absence of selenide, the enzyme-phosphoryl intermediate also is hydrolyzed by solvent water to form orthophosphate, resulting in formation of two enriched orthophosphates for each ATP when the reaction is carried out in H₂¹⁸O.

Scheme 1.

Although many questions remain concerning the catalytic mechanism of selenophosphate synthetase, the data presented in this paper support the PIX data indicating that the mechanism involves initial formation of an enzyme-phosphoryl intermediate with ADP being an intermediate in the reaction pathway. The energy of the β,γ-pyrophosphate bond of ATP, conserved as an enzyme-phosphoryl group, is then used for the synthesis of the energy-rich selenophosphate product. Although the energy of the Se–P bond has not been determined, it has recently been shown (23) that selenophosphate can serve as a phosphate donor in the phosphorylation of alcohols and amines, attesting to the energy rich nature of the compound.

ABBREVIATIONS

PIX

positional isotope exchange

Tricine

N-tris(hydroxymethyl)methylglycine