Biosci. Biotechnol. Biochem., 75 (6), 1184–1187, 2011
Note
Selenite Reduction by the Thioredoxin System: Kinetics and Identification of Protein-Bound Selenide Takashi T AMURA,1; y Kumi S ATO,1 Kentaro K OMORI,1 Takeshi IMAI,2 Mitsuhiko K UWAHARA,1 Takahiro O KUGOCHI,1 Hisaaki M IHARA,2 Nobuyoshi E SAKI,2 and Kenji I NAGAKI1 1
Department of Bioresources Chemistry, Graduate School of Natural Science and Technology, Okayama University, 1-1-1 Tsushima-naka, Kita-ku, Okayama 700-8530, Japan 2 Laboratory of Molecular Microbial Science, Institute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan Received November 30, 2010; Accepted February 23, 2011; Online Publication, June 13, 2011 [doi:10.1271/bbb.100847]
Selenite (SeO3 2 ) assimilation into a bacterial selenoprotein depends on thioredoxin (trx) reductase in Esherichia coli, but the molecular mechanism has not been elucidated. The mineral-oil overlay method made it possible to carry out anaerobic enzyme assay, which demonstrated an initial lag-phase followed by timedependent steady NADPH consumption with a positive cooperativity toward selenite and trx. SDS-PAGE/autoradiography using 75 Se-labeled selenite as substrate revealed the formation of trx-bound selenium in the reaction mixture. The protein-bound selenium has metabolic significance in being stabilized in the divalent state, and it also produced the selenopersulfide (-S-SeH) form by the catalysis of E. coli trx reductase (TrxB). Key words:
thioredoxin reductase; selenite; Escherichia coli; anaerobic enzyme assay; selenopersulfide
Formate dehydrogenase H (FDH-H) is a bacterial selenoprotein that constitutes the anaerobic nitrate respiration system in E. coli.1) It has a catalytically essential selenocysteine (Sec) residue in its active site, which is incorporated into the growing polypeptide chains at in-frame opal codon UGA.2,3) Although the mechanism of co-translational Sec insertion differs in several respects in the three domains of life forms, bacteria, eukaryotes, and archaea, it is universally conserved in that Sec is directed by the UGA codon and is synthesized on a specific tRNAUGA that is initially charged with L-serine.4,5) It is also common to the three domains that the synthesis of Sec-tRNAUGA requires a highly reactive selenium donor compound, monoselenophosphate, which is synthesized from ATP and selenide by selenophosphate synthetase (SPS), the selD product in the case of E. coli.6) Uptake and facile assimilation of extracellular inorganic selenite (SeO3 2 ) into FDH-H served as a model system for studying the selenium uptake mechanism in the past decade. Because the concentration of selenium is normally very low in bacterial cells, SeO3 2 must be incorporated into selenoproteins through a specific pathway.7,8) The metabolic pathway by which inorganic
tetravalent selenium (IV) is reduced to divalent selenide in vivo is perhaps distinct from the sulfate-reducing pathway.9) An initial attempt to elucidate the reductive selenite assimilation pathway was reported by Painter et al.,10) who stated that the chemical reaction of SeO3 2 with several organic thiols tended to produce selenotrisulfides (R-S-Se-S-R0 ). It was initially thought that a main pathway for selenite assimilation is initiated by reduction by GSH, the most abundant thiol in the cytoplasm.9,11–13) However, at physiological pH levels, the GS-Se-SG intermediate can be very unstable, and it is rapidly decomposed to produce elemental selenium as end product. Our preceding study indicated the physiological significance of the thioredoxin-dependent reducing system in bacterial selenite assimilation. The wild-type MC4100 strain and a glutathione reductase knockout (gor::tet) mutant produced FDH-H activity and incorporated 75 Se-labeled selenite into the polypeptide of FDH-H, whereas the thioredoxin reductase knockout (trxB::kan) strain failed to incorporate it into the bacterial selenoprotein.14) This suggests that reduction of selenite in vivo depends on the thioredoxin-reducing system, but the mechanism of trxB-dependent selenite assimilation is not clearly understood. Kumar et al. first reported that sodium selenite is a good substrate for the E. coli trx system, causing oxygen-dependent, non-stoichiometric NADPH consumption under aerobic conditions and stoichiometric oxidation of 3 mol NADPH/mol selenite under anaerobic conditions.15) It was also reported that the initial rates under aerobic and anaerobic conditions were nearly identical, suggesting that the non-stoichiometric NADPH-consumption resulted from a continued reaction by the reduction product, presumably selenide, which is highly sensitive to auto-oxidation. Although it was reported in several following studies that selenite can be bound to a pair of cysteine residues such as protein disulfide isomerase,16,17) Kumar et al. concluded that the product was a free selenide, which mediates the reoxidation of trx by dissolved oxygen in the solution. Here we report the kinetics of E. coli trx-mediated selenite reduction with a novel positive cooperativity
y To whom correspondence should be addressed. Tel: +81-86-251-8293; Fax: +81-86-251-8388; E-mail:
[email protected] Abbreviations: trx, thioredoxin; TrxB, E. coli trx reductase; Sec, selenocysteine; FDH-H, formate dehydrogenase H
Enzymatic Reduction of SeO3 2 to Se2 by the Trx System
Abs 340 nm
A
0.3
0.2
0.1
0.0
0
100
200
300
Time (s)
B Relative velocity
1.0 0.8 0.6 0.4 0.2 0.0 0
0.2
0.4
0.6
0.8
1.0 1.2
Trx (µM)
C 1.0
Relative velocity
with respect to trx and selenite. We also succeeded in inidentifying stable selenium-bound trx in the reaction mixture. The formation of Se-bound trx requires catalytic turnover driven by NADPH and thioredoxin reductase (TrxB). Our findings strongly suggest that bacterial selenite assimilation involves a Se-bound trx, which functions as a specific Se-delivery protein. For our in vitro kinetic study, approximately 6 mg of C-terminally His-tagged trx was purified from a 100-mL culture of E. coli BL21 (DE3) (Novagen, Madison, WI) harboring pET32a-trx-6xHis, which was constructed by introducing a TAA codon at the end of trx-tag in pET32a plasmid (Novagen). The genes of E. coli trxB were expressed as an N-terminally His-tagged recombinant protein using a pTrc-His-TOPO expression vector (Invitrogen, Carlsbad, CA). The final yield of TrxB from the 100-mL culture was 5.0 mg. These recombinant proteins were homogeneous on SDS–PAGE. Steady state kinetic studies using C-terminally 6xHis-tagged trx and N-terminally 6xHis-tagged TrxB by DTNB-dependent spectrophotometric assay18) revealed that Km and kcat values for the recombinant trx were 1:26 0:11 mM and 1,520 min1 , close to the reported values of 1.25 mM and 1,320 min1 for the DTNB-dependent E. coli trx assay,19) suggesting that the 6xHis tags on the recombinant proteins did not disturb their catalytic functions significantly. A convenient, effective technique was developed to measure the rate of selenite reduction under anaerobic conditions simply by overlaying 50 mL of mineral oil (ABI, Tokyo) over the enzyme reaction mixture (1 mL) in the quartz cuvette. The components of solutions such as the buffer and substrates were thoroughly purged with argon gas before they were mixed, and the solutions were mixed beside a UV spectrometer in the open air. Additional purging may be required to obtain a reproducible NADPH consumption rate after several measurements due to gradual air-oxidation of the reserved solutions. Accordingly, selenite reduction was performed in 1 mL of reaction mixture containing 40 mM HEPES pH 7.2, 20 mM KCl, 3 mM MgCl2 , 0.1 mM EDTA, 0.2 mM NADPH, 10 units of SOD, and 10 units of catalase. NADPH-dependent selenite reduction was carried out using the recombinant E. coli TrxB at a concentration of 1.2 mM in the presence of various concentrations of recombinant trx in a range of 0– 1.2 mM. The concentrations of sodium selenite were 50, 125, and 250 mM. Selenite served as an efficient oxidant for the recombinant trx system, as has been reported for native proteins,15) and the absorption at 340 nm did not decrease when a component of the thioredoxin system, trx, TrxB, or NADPH, was omitted from the reaction mixture. A redox-inert trx protein bearing a Cys32Ser point mutation did not show selenite-dependent NADPH consumption, suggesting that selenite reduction occurs at the active site of trx. Catalysis was initiated by the addition of trx, the time course of the enzyme reaction showed a clear lag in the initial stage of the enzyme reaction, and the steady-state rate was obtained from the linear decrease in absorption at 340 nm after the lag phase (Fig. 1A). The initial reaction rate reached maximum velocity, Vmax , when the concentration of trx was increased to 1.2 mM (Fig. 1B), and Vmax varied
1185
0.8 0.6 0.4 0.2 0.0 0
100
200
300
Na2SeO3 (µM) Fig. 1. Selenite-Dependent Oxidation of NADPH Catalyzed by the E. coli trx System. A, A cuvette containing 1 mL of enzyme reaction mixture was sealed from air with 50 mL of mineral oil on top of the solution. The reaction was started by addition of trx (1.2 mM), and a steady decrease in absorbance at 340 nm started after a lag-phase of 150 s. The broken line represents blank without trx. B, Effects of the trx concentration on the NADPH oxidation rate in the presence of 250 ( ), 125 ( ), and 50 mM ( ) sodium selenite. C, Effects of sodium selenite concentrations on the NADPH oxidation rate in the presence of 0.04 ( ), 0.12 ( ), 0.48 ( ), and 1.2 mM ( ) trx. The Vmax values for various trx concentrations are summarized in Table 1.
significantly as a function of the selenite concentration (Table 1). The maximum velocity for NADPH consumption increased as the concentration of Na2 SeO3 increased from 50 to 250 mM (Fig. 1C). The plot of the trx concentration versus the NADPH consumption rate was hyperbolic when the selenite concentration was 50 mM, and the best-fit values of the Michaelis constant (Km ) and the maximum velocity (Vmax ) were 0.44 mM and 2.97 mM/min respectively. In the presence of 125– 250 mM selenite, however, a clear positive cooperativity was observed with Hill coefficients of 1.7 and 1.8 respectively. A Hill coefficient less than 2 is consistent with the fact that E. coli TrxB has a homodimeric quaternary structure.20) The kcat value, 6.1 min1 (Vmax 14.65 mM/min) attained in the presence of 1.2 mM trx and 250 mM selenite was much smaller than the kcat value (1,520 min1 ) observed with DTNB as oxidant.
1186
T. T AMURA et al.
Table 1. Vmax Values for Selenite-Dependent trx Reductase Activity Trx conc. (mM)
Na2 SeO3 conc. (mM)
Vmax (mM/min)a
Fig. 1B
1.20 1.20 1.20
50 125 250
2.97 7.65 14.65
Fig. 1C
0.04 0.12 0.48 1.20
250 125b 125b 250
0.80 1.01 2.41 8.04
A
1
2
3
4
5
1
2
3
4
5
TrxR Trx
B TrxR Trx
a
Vmax values for the steady NADPH consumption rate, normalized in Fig. 1B and C. Figure 1B and 1C, were determined using preparations of TrxB and trx proteins, and thus the maximum velocity not exactly the same under the same concentrations of trx and Na2 SeO3 . b Vmax values were obtained when Na2 SeO3 was 125 mM, as shown in Fig. 1C.
Our kinetic studies using a recombinant trx system indicated for the first time that enzymatic selenite reduction shows remarkable sigmoid kinetics with respect to selenite and trx concentrations. Because TrxB did not show such cooperativity when insulin or GSSG was used as the terminal oxidant for trx turnover, the substrate in the present NADPH consumption is not likely to be oxidized trx. To identify the substrate compound generated in the reaction mixture, 75 Selabeled selenite was added to the reaction mixture, and the proteins in the solution were developed on SDS– PAGE in the absence of -mercaptoethanol. 75 SeLabeled proteins subsequently visualized by autoradiography were identified by their molecular masses after Coomassie Blue staining. Most strikingly, trx was efficiently labeled with radioactive selenium in the complete mixture (Fig. 2, lane 1). The solution in which TrxB was omitted also allowed trx to be labeled with 75 Se, but much less efficiently (lane 2). The seleniumbound trx accumulated without TrxB presumably indicated the formation of trx-S-SeO2 H adduct form. A very small amount of selenium was also bound to TrxB when trx or NADPH was absent from the solution (lanes 3–4). In the case of the absence of trx, TrxB was labeled with a very small amount of selenium during incubation. In contrast, the absence of reducing equivalent NADPH, resulted in poor labeling on trx and TrxB, suggesting that the formation of 75 Se-labeled trx and TrxB to lesser extent depends on the enzymatic reaction. Sulfur transferases and sulfur-specific delivery proteins have been identified and characterized in detail in sulfur metabolic pathways.21,22) On this analogy, the involvement of specific selenium-delivery proteins was proposed based on two major biochemical factors. First, in vitro assay of bacterial SPS normally requires high levels of selenide and dithiothreitol, and yet the apparent Km value for the selenide anion is reportedly as high as 20 mM, far above the optimal concentrations of selenium for the growth of various bacterial species.23) In fact, levels above 10 mM selenium are highly toxic to many bacterial species. Secondly, effective selenium assimilation requires a specific gateway mechanism that discriminates selenium from sulfur, which is 1,000-fold more abundant in cells. Also, a hypothetical seleniumtransport protein must fulfill conflicting requirements: it must produce a stable complex with selenium so as to discriminate selenium from sulfur, but at the same time
Fig. 2. 75 Se-Labeled Proteins Produced in trx System-Dependent Selenite Reduction. Coomassie Brilliant Blue (R250) staining of SDS/12% PAGE gel in the absence of -mercaptoethanol (A) and PhosphorImager analysis (B) of the enzyme reaction mixture with 75 Se-labeled selenite. The complete reaction mixture (lane 1), and the control without TrxB (lane 2), without trx (lane 3), without NADPH (lane 4), and without 75 Se-labeled selenite (lane 5).
the delivery protein must liberate divalent selenide for SPS. Rhodanese-like proteins have been postulated to be selenium-carrier proteins, which perhaps form the selenopersulfide intermediate state at the active site (Protein-S-SeH).24,25) Ambiguity remains, however, about this selenopersulfide intermediate due to the lack of chemical evidence for the selenopersulfide compounds (-S-SeH). It has been found experimentally that protein-bound selenium is unexpectedly stable. It even required a dithiol reductant such as 2 mM reduced lipoic acid, when the bound selenide was released from glycerlaldehyde-3-phosphate dehydrogenase (GAPDH).24,25) Although selenopersulfide is one of the chemical forms most likely to be generated, the proteinbound selenium might yet be a selenotrisulfide form with one equivalent of glutathione retained on the protein-bound selenide, GAPDH-S-Se-SG. In fact, it has been reported that binding of the selenotrisulfide form of penicillamine (Pen-S-Se-S-Pen) to the Cys residue of hemoglobin depends essentially on the thiol exchange reaction.26,27) In contrast to the low efficiency of monoCys type carrier proteins, the trx system has the advantage of being implemented with NADPH-dependent reducing machinery. The non-stoichiometric oxidative burst in the presence of oxygen strongly suggests that a transitory selenopersulfide is generated at the active site of trx. It remains to be determined whether the catalytic formation of selenopersulfide on trx can serve as an efficient substrate for bacterial SPS.
Acknowledgments This work was supported by a Grant-in-aid for Scientific Research 18580122 to TT.
References 1) 2) 3) 4) 5)
Ingledew W and Poole R, Microbiol. Rev., 48, 222–271 (1984). Ganther H, Biochemistry, 10, 4089–4098 (1971). Sandholm M and Sipponen P, Arch. Biochem. Biophys., 155, 120–124 (1973). Chambers I, Frampton J, Goldfarb P, Affara N, McBain W, and Harrison P, EMBO J., 5, 1221–1227 (1986). Zinoni F, Birkmann A, Stadtman TC, and Bo¨ck A, Proc. Natl. Acad. Sci. USA, 83, 4650–4654 (1986).
Enzymatic Reduction of SeO3 2 to Se2 by the Trx System 6) 7) 8) 9) 10) 11) 12) 13) 14)
15) 16)
Leinfelder W, Forchhammer K, Veprek B, Zehelein E, and Bo¨ck A, Proc. Natl. Acad. Sci. USA, 87, 543–547 (1990). Stadtman TC, Davis JN, Zehelein E, and Bo¨ck A, Biofactors, 2, 35–44 (1989). Heider J and Bo¨ck A, Adv. Microbial. Phys., 35, 71–109 (1993). Turner RJ, Weiner JH, and Taylor DE, Biometals, 11, 223–227 (1998). Painter EP, Chem. Rev., 28, 179–213 (1941). Kosower N and Kosower E, Int. Rev. Cytol., 54, 109–160 (1978). Ganther HE, Carcinogenesis, 20, 1657–1666 (1999). Hsieh H and Ganther H, Biochemistry, 14, 1632–1636 (1975). Takahata M, Tamura T, Abe K, Mihara H, Kurokawa S, Yamamoto Y, Nakano R, Esaki N, and Inagaki K, J. Biochem., 143, 467–473 (2008). Kumar S, Bjornstedt M, and Holmgren A, Eur. J. Biochem., 207, 435–439 (1992). Schoenmakers CHH, Pigmans I, and Visser TJ, Biochim. Biophys. Acta, 1121, 160–166 (1992).
17) 18) 19) 20) 21) 22) 23) 24) 25) 26) 27)
1187
Sinha R, Bansal MP, Ganther H, and Medina D, Carcinogenesis, 14, 1895–1900 (1993). Holmgren A, J. Biol. Chem., 252, 4600–4606 (1977). Prongay AJ, Engelke DR, and Williams CH, J. Biol. Chem., 264, 2656–2664 (1989). Lennon BW, Williams CH, and Ludwig ML, Protein Sci., 8, 2366–2379 (1999). Kambampati R and Lauhon CT, J. Biol. Chem., 275, 10727– 10730 (2000). Nagahara N, Ito T, and Minami M, Histol. Histopathol., 14, 1277–1286 (1999). Lacourciere GM and Stadtman TC, Proc. Natl. Acad. Sci. USA, 96, 44–48 (1999). Ogasawara Y, Lacourciere G, and Stadtman TC, Proc. Natl. Acad. Sci. USA, 98, 9494–9498 (2001). Lacourciere GM, Levine RL, and Stadtman TC, Proc. Natl. Acad. Sci. USA, 99, 9150–9153 (2002). Haratake M, Fujimoto K, Ono M, and Nakayama M, Biochim. Biophys. Acta, 1723, 215–220 (2005). Haratake M, Fujimoto K, Hirakawa R, Ono M, and Nakayama M, J. Biol. Inorg. Chem., 13, 471–479 (2008).
