Selenophosphate Synthetase

Matt D. Wolfe

[SECTION EDITOR, JOHN INGRAHAM]

Posted July 6, 2004

Laboratory of Biochemistry, National Heart, Lung, and Blood Institute, National Institutes of Health, Building 50, Room 2122, Bethesda, MD 20892-8012

Phone: 301-435-8369; Fax: 301-496-0599; E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

Selenophosphate synthetase, the selD gene product from Escherichia coli, is one of the enzymes required for the synthesis and specific insertion of selenocysteine into proteins directed by the TGA codon (16). In addition to its role as selenium donor for protein synthesis, selenophosphate (SePO₃) also is the source of selenium for conversion of 2-thiouridine residues in some tRNAs to 2-selenouridine (18). More specifically, these reactions are catalyzed by selenocysteine synthase and 2-selenouridine synthase for synthesis of selenocysteine-charged tRNAs and 5-methylaminomethyl-2-selenouridine tRNAs, respectively. Selenophosphate synthetases have been isolated from or are thought to be present in most organisms (5); however, the best characterized selenophosphate synthetase is from E. coli, in which both in vivo and in vitro studies have been performed. Leinfelder et al. (13) showed that an E. coli mutant lacking an intact selD gene fails to incorporate Se into both the selenocysteine-containing enzyme formate dehydrogenase (FDH) and tRNA species that normally contain 2-selenouridine residues at the wobble position. Thus, this study strongly implicated selenophosphate as playing a major role in E. coli selenium metabolic pathways.

As shown in Fig. 1, the selenophosphate synthetase reaction requires some form of reduced selenium such as hydrogen selenide (HSe⁻) and ATP as substrates to generate a stoichiometric amount of SePO₃, AMP, and orthophosphate. The 37-kDa E. coli enzyme is thought to carry out catalysis as a monomer with a k _cat of ~1.3 min^-1. While the actual substrate form of selenium that serves as a substrate in vivo is not clear (see below), the standard in vitro selenophosphate synthetase assay utilizes HSe^– produced by chemical reduction of selenite (17). Addition of selenite alone does not lead to production of SePO₃ in vitro. The high K_m value determined for selenium by this assay (10 to 40 μM, toxic concentrations in vivo) strongly suggests that either a different substrate form of selenium exists, or HSe^– is delivered directly to selenophosphate synthetase in vivo. As described in more detail elsewhere in this module, studies of selenocysteine lyase enzymes (NifS-like proteins) have shown that they are able to efficiently provide selenium from selenocysteine to selenophosphate synthetase in vitro, suggesting the lyase enzymes may be selenium delivery proteins in vivo (9, 10).

Fig. 1The selenophosphate synthetase reaction.

Like many ATP-utilizing enzymes, selenophosphate synthetase requires Mg²⁺ for activity (17). Other divalent cations such as Mn²⁺, Fe²⁺, Cu²⁺, Co²⁺, Ca²⁺, and Zn²⁺ cannot substitute for Mg²⁺ in the selenophosphate synthetase reaction. However, binding studies conducted by Kim et al. (8) have shown that both Mn²⁺ and Zn²⁺ bind tightly to selenophosphate synthetase and that Zn²⁺ acts as a potent inhibitor of the reaction at low micromolar concentrations. In addition to Mg²⁺, the selenophosphate synthetase reaction requires monovalent cation, with K⁺ yielding the highest enzyme activity (6, 17). The smaller ions Li⁺ and Na⁺ do not stimulate enzyme activity, and when present with K⁺, Na⁺ inhibits the reaction. Addition of larger cations to the reaction, such as NH₄ ⁺, produces lower levels of selenophosphate synthetase activity than with K⁺. K⁺ is also required for the binding of Mn²⁺ but not Zn²⁺, indicating that the two divalent cations interact with selenophosphate synthetase differently.

By using various isotopically labeled ATPs and Se, the products of the selenophosphate synthetase reaction were readily identified. In experiments conducted with α-, β-, or [γ-³²P]ATP, Ehrenreich et al. showed that the phosphoryl group of SePO₃ derives from the γ-phosphate of ATP, while the β-phosphate of ATP is released as orthophosphate during the reaction cycle (2). By ³¹P nuclear magnetic resonance spectroscopy, the product of the selenophosphate synthetase reaction was identified and characterized by comparison to the chemically synthesized SePO₃ (4, 18). The ³¹P resonance of SePO₃ is shifted downfield relative to orthophosphate and the nucleotide substrate and products due to the unique chemical properties of selenium. By following slight changes in the chemical shift of SePO₃, Glass and coworkers determined several pK_a values of the molecule (pK₂, 4.6; pK₃, 8.8) (4). Among the various naturally occurring isotopes of selenium, radioactive ⁷⁵Se has proven to be quite useful for studying selenophosphate synthetase in the laboratory. In a study using a combination of [³²P]ATP substrates and [⁷⁵Se]selenide, Liu and Stadtman discovered that under steady-state reaction conditions, selenophosphate synthetase has approximately 0.6 equivalent of both ⁷⁵Se and ³²P from [γ-³²P]ATP bound to the enzyme (14). Furthermore, a similar amount of [8-¹⁴C]AMP was shown to be bound to selenophosphate synthetase under identical reaction conditions. However, only negligible amounts of [³²P]orthophosphate were associated with the enzyme, indicating that during selenophosphate synthetase turnover, orthophosphate is the first product released, with release of SePO₃ and/or AMP being at least partially rate-limiting in catalysis.

Studies of selenophosphate inhibition have provided further insight into the mechanism of selenophosphate synthetase. At high concentrations of SePO₃ (>1 mM) or AMP (K_i, ~170 μM), product inhibition is observed, suggesting that release of these products is reversibly associated with the rate-limiting step in catalysis (17). No inhibition of selenophosphate synthetase activity was observed in the presence of orthophosphate at concentrations up to 20 mM. These data are consistent with the observation that SePO₃ and AMP are bound to the enzyme during steady-state turnover of the enzyme. Further studies showed that ATP analogs with nonhydrolyzable phosphoryl groups (α,β-CH₂-ATP, β,γ-CH₂-ATP, ATPγS) inhibit selenophosphate synthetase (17). The type of inhibition produced by the ATP analogs was not determined; however, based on the structural similarity of the analogs to ATP, competitive inhibition with this substrate is likely.

An assay by which [¹⁴C]AMP formation is measured in the absence of selenide showed that selenophosphate synthetase catalyzes hydrolysis of ATP to AMP and two orthophosphates in an uncoupled reaction (17). While the rate of the uncoupled reaction is about 200 times slower than that of the fully coupled reaction, studies aimed at revealing the mechanism of ATP hydrolysis and Se–P bond formation were conducted using this system. Using positional isotope exchange (PIX) methodology, Mullins et al. showed that, in contrast to mechanisms for enzymes such as pyruvate phosphate dikinase (Fig. 2A) in which a pyrophosphoryl-enzyme intermediate is generated (3), selenophosphate synthetase likely forms a phosphoryl-enzyme intermediate (Fig. 2B) (15). This mechanistic proposal was further supported in a study by Walker et al. wherein isotope exchange between ATP and [8-¹⁴C]ADP, but not [8-¹⁴C]AMP, was observed in the uncoupled reaction and ¹⁸O derived from [¹⁸O]H₂O was incorporated only into orthophosphate (19). While no phosphorylated enzyme has been trapped or characterized, all of these observations are consistent with the mechanism shown in Fig. 2B, in which a phosphoryl-enzyme intermediate is generated and subsequently attacked by selenide to generate SePO₃.

Fig. 2Illustrations of the pyrophosphoryl-enzyme intermediate of pyruvate phosphate dikinase (A) and the proposed phosphoryl-enzyme intermediate of selenophosphate synthetase (B). An unidentified nucleophilic residue (X) on selenophosphate synthetase would react with the γ-phosphoryl group of ATP. Nucleophilic attack by selenide to form SePO ₃and hydrolysis of ADP would follow to complete the reaction.

The sequencing of selenophosphate synthetase genes from various organisms reveals several conserved regions in the gene product. Figure 3 shows a comparison of sequences from several species showing that two nucleotide binding regions, Walker motifs A and B, are highly conserved, as would be expected for an ATP-dependent enzyme. Within motif A, at the N terminus known as the glycine-rich region, a pair of Cys residues are highly conserved. Site-directed mutants of selenophosphate synthetase were generated to determine the possible roles these and other amino acids in motif A play in catalysis. Table 1 summarizes the functional differences between the mutant enzymes and clearly shows the importance of Lys20 (E. coli numbering) and Cys17 in substrate binding and catalysis, respectively (7, 8).

Fig. 3Sequence alignment of several selDgenes encoding selenophosphate synthetases from several organisms. Nucleotide binding motifs A and B are boxed, and residues of the E. coliselenophosphate synthetase that have been mutated and the resulting enzymes studied are indicated with (?). U indicates selenocysteine. Organisms listed: D.m., Drosophila melanogaster; H.s., Homo sapiens; M.m., Mus musculus; H.i., Haemophilus influenzae; E.c., Escherichia coli; M.j., Methanococcus jannaschii.

Table 1Some properties of wild-type and various mutants of E. coli selenophosphate synthetase

Curiously, several of the selenophosphate synthetases listed in Table 1 have residues other than Cys at positions analogous to the E. coli Cys17. Similar to several of the eukaryote selenophosphate synthetases, Haemophilus influenzae contains a SeCys at position 17. SeCys is known to be much more reactive than the more common sulfur counterpart and has been shown to be necessary for maximum activity in enzymes such as formate dehydrogenase and human thioredoxin reductase (1, 12). However, following expression and isolation of the H. influenzae selenophosphate synthetase, Lacourciere and Stadtman found no significant difference in the activities of the two enzymes (11). This result suggested that the function of the residue at position 17 is either not catalytic or not rate-limiting. Several selenophosphate synthetase isozymes contain residues other than Cys and SeCys at position 17, but their activities have yet to be adequately determined. In another study, Lacourciere et al. discovered that while no activity is observed with the E. coli Cys17Ser mutant using the standard in vitro assay (see Table 1), selenophosphate synthetase activity was measurable using an assay in which selenocysteine lyase enzymes provide the selenium substrate in some form of elemental Se rather than a more reduced Se^2– (9). While the Cys17Ser mutation may not result in a physiologically functional selenophosphate synthetase, the Se generated by lyase enzymes may be a more efficient substrate for SePO₃ generation. Studies of the interaction between selenophosphate synthetase and selenocysteine lyases continue to be a focus in our laboratory.

Recent investigations into the mechanism of selenophosphate synthetase have revealed a property of selenophosphate synthetase not previously observed (20). In samples of purified selenophosphate synthetase, an unusual optical absorption spectrum is seen (Fig. 4A). In addition to the peak at 280 nm resulting from aromatic amino acids, selenophosphate synthetase has a broad shoulder, "the chromophore," centered at about 315 nm (ε ₃₁₅,~3,000 M^–1 cm^–1). The chromophore appears to be covalently associated with the enzyme and is perturbed when selenophosphate synthetase interacts with catalytically relevant ions and molecules. Both Mg²⁺ and Mn²⁺ cause a small red shift in the spectrum of the chromophore, while Zn²⁺ binding leads to quenching of the chromophore. More interestingly, a large red shift in the optical absorption spectrum is observed when both metal and ATP are present (Fig. 4B). The shift from 315 nm to 340 nm in the presence of ATP requires that either Mg²⁺ or Mn²⁺ be bound and that K⁺ be present. The ATP-dependent shift is not observed in the presence of inhibitor ions such as Zn²⁺ and Na⁺. Furthermore, ATP analogs containing nonhydrolyzable γ-phosphate groups do not induce the red shift in the presence of K⁺ and Mg²⁺ or Mn²⁺. These experiments provide more evidence for the formation of a phosphoryl-enzyme intermediate and lay the foundation for future kinetic experiments addressing the selenophosphate synthetase mechanism.

Fig. 4Optical absorption spectra of E. coliselenophosphate synthetase. (A) UV/visible (UV/Vis) spectrum of purified enzyme. The inset is an enlargement showing the region of the chromophore. (B) UV/Vis spectra of the chromophore in the absence (solid line) and presence (dotted line) of Mg ²⁺/ATP and 20 mM KCl. The inset is a difference spectrum of the Mg ²⁺/ATP-bound and -unbound forms of selenophosphate synthetase.