Discovery and characterization of a Coenzyme A disulfide reductase from Pyrococcus horikoshii Implications for the disulfide metabolism of anaerobic hyperthermophiles Dennis R. Harris*, Donald E. Ward 1 , Jeremy M. Feasel 2 , Kyle M. Lancaster 2 , Ryan D. Murphy 2 , T. Conn Mallet 3 and Edward J. Crane III 2 1 Genencor International, Palo Alto, CA, USA 2 Department of Chemistry, Pomona College, Claremont, CA, USA 3 Center for Structural Biology, Wake Forest University School of Medicine, Winston-Salem, NC, USA While surveying the genomes of hyperthermophilic and thermophilic Archaea for homologues of the flavoprotein disulfide reductases, many homologues with a high degree of identity to the branch of this family represented by glutathione reductase were found [1]. Most of the homologues appear to belong to the subfamily that depend on a redox-active single cysteine, analogous to the NADH oxidase and per- oxidase of Enterococcus and the coenzyme A disulfide reductase (CoADR; EC 1.8.1.14) of Staphylococcus Correspondence E. J. Crane III, Department of Chemistry, Pomona College, 645 North College Avenue, Claremont, CA 91711, USA Fax: +1 909 607 7726 Tel: +1 909 607 9634 E-mail: ej.crane@pomona.edu Website: http://www.userwebs. pomona.edu/ejc14747/EJ_web.htm *Present address Department of Biochemistry, University of Wisconsin-Madison, Madison, WI, USA (Received 20 July 2004, revised 7 December 2004, accepted 4 January 2005) doi:10.1111/j.1742-4658.2005.04555.x We have cloned NADH oxidase homologues from Pyrococcus horikoshii and P. furiosus, and purified the recombinant form of the P. horikoshii enzyme to homogeneity from Escherichia coli. Both enzymes (previously referred to as NOX2) have been shown to act as a coenzyme A disulfide reductases (CoADR: CoA-S-S-CoA + NAD(P)H + H + fi 2CoA-SH + NAD(P) + ). The P. horikoshii enzyme shows a k cat app of 7.2 s )1 with NADPH at 75 °C. While the enzyme shows a preference for NADPH, it is able to use both NADPH and NADH efficiently, with both giving roughly equal k cat s, while the K m for NADPH is roughly eightfold lower than that for NADH. The enzyme is specific for the CoA disulfide, and does not show significant reductase activity with other disulfides, including dephos- pho-CoA. Anaerobic reductive titration of the enzyme with NAD(P)H pro- ceeds in two stages, with an apparent initial reduction of a nonflavin redox center with the first reduction resulting in what appears to be an EH 2 form of the enzyme. Addition of a second of NADPH results in the formation of an apparent FAD-NAD(P)H complex. The behavior of this enzyme is quite different from the mesophilic staphylococcal version of the enzyme. This is only the second enzyme with this activity discovered, and the first from a strict anaerobe, an Archaea, or hyperthermophilic source. P. furio- sus cells were assayed for small molecular mass thiols and found to contain 0.64 lmol CoAÆg dry weight )1 (corresponding to 210 lm CoA in the cell) consistent with CoA acting as a pool of disulfide reducing equivalents. Abbreviations CoADR, coenzyme A disulfide reductase (EC# 1.8.1.14); pfCoADR, P. furiosus coenzyme A disulfide reductase; phCoADR, P. horikoshii coenzyme A disulfide reductase; DTNB, 5,5¢ dithiobis(2-nitrobenzoic acid); EH 2 , two-electron reduced enzyme; EH 4 , four-electron reduced enzyme; HEPPS, N-(2-hydroxyethyl)piperazine-N¢-3-propanesulfonic acid; NOX, NADH oxidase; NPX, NADH peroxidase; TCA, trichloroacetic acid. FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS 1189 aureus [2,3]. These enzymes are proposed to be involved in the robust oxygen-defense systems of aerobic and facultatively anaerobic organisms [4,5] and would not be expected to be present in the mostly strictly anaerobic hyperthermophiles. Evidence is mounting, however, for the presence of a vigorous oxidative stress response in Pyrococcus, including the discovery of both a novel peroxide-producing super- oxide reductase [6] and an NADH oxidase [1]. Addi- tionally, an oxidative stress response has been characterized in the strictly anaerobic bacterium Clos- tridium perfringens [7]. Microorganisms in the genus Pyrococcus are strictly anaerobic hyperthermophiles (T opt ¼ 100 °C) isolated from marine hydrothermal vents [8–10]. The genomes of P. horikoshii, P. furiosus and P. abyssii each contain at least two NADH oxidase homologues. Characteriza- tion of one of these homologues (NOX1) from P. furiosus has been described previously [1]. NOX1 shows a novel H 2 O and H 2 O 2 producing NADH oxid- ase activity in both the absence and presence of exo- genous FAD. The second NOX homologue examined (from P. horikoshii, previously referred to as NOX2) showed a slow NAD(P)H oxidase activity in the pres- ence of high concentrations of substrate-level FAD, i.e., in addition to the enzyme bound FAD. The results described here demonstrate that this enzyme is not likely to act as an NADH oxidase in vivo, instead act- ing as a CoADR. This is only the second demonstra- ted CoA reductase activity, and the first appearance of this activity in both the Archaea and in a strict anaer- obe. While the best known small molecular mass thiol is probably glutathione, a number of novel thiols such as mycothiol, c-glutamyl cysteine, and trypanithione have been found in microorganisms [11]. The function of these thiols appears to be the maintenance of a reducing intracellular environment. Due to the pres- ence of a CoADR homologue in all three pyrococcal genomes, P. furiosus cells were assayed for the pres- ence of small molecular mass thiols in order to better understand the role of this enzyme and thiol ⁄ disulfide systems in pyrococcal metabolism. The results presen- ted below provide an insight into the use of a small molecular mass thiol system for the maintenance of the internal redox environment in an anaerobic hyper- thermophile. Results Characterization of the recombinant CoADR The recombinant CoADR from P. horikoshii (phCo- ADR) was purified 15-fold with a yield of 58% and a specific activity of 3.26 UÆmg )1 (oxidase activity) (Table 1) and 8.3 UÆmg )1 (CoADR activity). The phCoADR had a subunit m ¼ 50 k as determined by SDS ⁄ PAGE, and was shown by both HPLC and con- ventional size-exclusion chromatography to be a tetra- meric enzyme of m ¼ 198 k. The enzyme obtained from the overexpression host is approximately 20% holoenzyme. After reconstitution with FAD the enzyme contains 0.92 flavin per subunit based on the ratio of protein concentration to flavin concentration (as determined at 460 nm). blast and tfasta analysis of the phCoADR revealed a significant level of identity to putative NADH oxidases from hyperthermophiles and bacterial NADH oxidases from mesophilic sources (Fig. 1). Of particular interest was the identity to the well characterized NOXs from mesophilic organisms, with the highest levels of identity found with the NADH oxidases from E. faecalis (28%), Streptococcus mutans (26%), and Brachyspira (Serpulina) hyodysente- riae (24%). The CoADR from S. aureus is 26% identi- cal to the P. horikoshii CoADR. As shown in Fig. 1, the P. furiosus NOX1 and the pf and ph CoADRs con- tain a cysteine which corresponds to the single redox active cysteine of the E. faecalis NOX and NPX, as well as considerable identity in areas that have been shown to be important for NADH and FAD binding in these enzymes. Table 1. Purification of the recombinant coenzyme A disulfide reductase from P. horikoshii. For the purposes of this table, purification was monitored by the FAD-dependent NADH oxidase activity of the enzyme, with a unit equal to the amount of enzyme required to oxidize 1 lmol NADH in 1 min in the presence of 100 lm NADH, 100 lm FAD, in 50 mm potassium phosphate, pH 7.50 at 75 °C. The specific activity of the purified enzyme in terms of coenzyme A disulfide reductase activity is 8.3 UÆmg )1 . Fraction Total units Total protein (mg) Specific activity (unitsÆmg )1 ) Purification (fold) Yield (%) Crude extract 197 888 0.221 – – Heat-treated extract 134 77.2 1.73 7.83 68.0 Q-sepharose 120 40.1 2.99 13.5 60.9 Size-exclusion 114 35.0 3.26 14.8 57.8 CoADR from P. horikoshii D. R. Harris et al. 1190 FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS In order to determine the extinction coefficient of the enzyme-bound FAD, the FAD was released from the holoenzyme by trichloroacetic acid (TCA) precipi- tation of the protein. The e 460 of the enzyme-bound FAD was determined to be 10 200 m )1 Æcm )1 .Itis interesting to note that attempts to remove the FAD from the enzyme by treatment in 6.0 m guanidine ⁄ HCl were unsuccessful, even with overnight incubation at 90 °C. This result is consistent with the extreme stabil- ity of this enzyme, and consistent with the observation that thermostable proteins, including the NOX1 from P. furiosus, are frequently stable in organic solvents and in the presence of denaturants [1]. The visible spectrum of the enzyme as purified (Fig. 2) has the same distinct shoulder in the area of 470 nm as the mesophilic staphylcoccal enzyme [12]. Fig. 1. Multiple sequence alignment of P. horikoshii and P. furiosus CoADRs to known NADH oxidases ⁄ NADH peroxidase ⁄ CoADR. The alignment was performed with CLUSTAL W. The GenBank accession numbers for the other enzymes are as follows: Staphylococcus aureus CoADR (AF041467), E. faecalis NOX (P37061), E. faecalis NPX (P37062), P. furiosus NOX (PF_1430634). D. R. Harris et al. CoADR from P. horikoshii FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS 1191 CoADR activity of the recombinant P. furiosus CoADR homologue The CoADR from P. horikoshii has a homolouge in P. furiosus that is 92% identical (Fig. 1), indicating that this gene product was also likely to be a CoADR. The recombinant P. furiosus CoADR (pfCoADR) was expressed in E. coli. Heat-treated FAD-reconstituted extracts of the putative pf CoADR showed a strong CoADR activity corresponding to a specific activity of 11 UÆmg )1 of heat-treated extract. This result com- pares favorably with the specific activity obtained dur- ing the production of the recombinant CoADR from P. horikoshii. Like the phCoADR, the pfCoADR is active with both NADH and NADPH. CoADR activ- ity was also detected in crude extracts of P. furiosus, however, the level of activity was low and difficult to distinguish from background reactions. Steady-state kinetics of the CoADR and oxidase reactions The kinetic constants for the CoADR and oxidase activities of phCoADR are listed in Table 2. The enzyme does not show significant NADH peroxidase activity and is not active in the reduction of glutathi- one, cystine, or 3 ¢ dephospho-CoA disulfide (unlike the staphylococcal enzyme, which has a 40 lm K m for the dephospho substrate). The enzyme also does not show significant 5,5¢ dithiobis(2-nitrobenzoic acid) (DTNB) reductase in a standard aerobic steady-state kinetic assay; however, in the anaerobic assays of free thiols discussed below the enzyme does show DTNB turn- over with NADPH at a very high concentration of enzyme (30–60 lm). While NADPH is the preferred substrate for the CoADR activity based on K m , the two nucleotide substrates have almost identical k cat s. This result is quite distinct from the substrate specifi- city observed with the staphylococcal enzyme, which shows a marked preference for NADPH [12]. While phCoADR shows a low level of NAD(P)H oxidase activity (in the absence of CoA disulfide) in the absence of substrate level FAD (i.e., FAD added in addition to that present in the enzyme-bound form), a significant amount of oxidase activity can be observed in the presence of additional substrate-level FAD (Table 2). The k cat app obtained in the presence of 100 lm NADH and FAD and 115 lm O 2 is 8.2 s )1 , which correlates well with the catalytic constant observed for the CoADR reaction. Thermostability and thermoactivity of the CoADR phCoADR is stable for months at both )80 °C and )20 °C, and has half-lives of > 100 and 39 h at 85° and 95 °C, respectively. Figure 3 shows the dependence of the oxidase and phCoADR activities on tempera- ture. While both activities show the preference for high temperature expected of an enzyme from a hyper- thermophile, at temperatures above 75 °C both activit- ies appear to plateau slightly, rather than increasing all the way to the optimal growth temperature for Pyro- coccus. Anaerobic reduction with NAD(P)H and redox state of the proposed cysteine nonflavin redox center As shown in Fig. 2, when phCoADR is titrated anaero- bically at 60 °C with NADPH the titration shows two main phases, each corresponding to the addition of 300 0.6 0.4 0.2 0 400 500 Wavelength, nm 0 NADPH, eq A 600.720 0.040 0.080 123 Absorbance 600 700 800 Fig. 2. Anaerobic reductive titration with NADPH at 60 °C. 56 lM phCoADR in 50 mM potassium phosphate at pH 7.50 was titrated with NADPH. Spectra are shown at 0 (—), 1.1 (- - -) and 2.0 (ÆÆÆÆ) NADPH. Inset, Change in absorbance at 600 (s) and 720 (d)nm during titration. Table 2. Michaelis constants for the P. horikoshii CoADR, deter- mined at 75 °Cin50m M potassium phosphate, pH 7.50. Activity-substrate K m app (lM) k cat app (s )1 ) CoADR-NADH a 73 8.1 CoADR-NADPH a < 9.0 7.2 CoADR-CoANaCl ⁄ CitoA b 30 7.1 Oxidase-NADH c 73 8.2 Oxidase-NADPH c 13 2.0 Oxidase-FAD d 22 5.9 a Determined at 200 lm CoA-S-S-CoA, b determined at 100 l m NADPH, c determined at 100 lm FAD, d determined at 100 lM NADH. CoADR from P. horikoshii D. R. Harris et al. 1192 FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS roughly one equivalent of NAD(P)H. At 720 nm the majority of the change occurs during the addition of the first equivalent of the reducing agent as seen in Fig. 2 (inset). Little change is observed at 361 nm dur- ing addition of the first equivalent of NADPH, while an increase is observed during the addition of the sec- ond and subsequent equivalents of NADPH. With the exception of the expected increase at 340 nm from the addition of free NADPH, little additional change is seen in the spectra when NADPH is added to a total of six equivalents. No reduction of the FAD is observed during addition of excess NAD(P)H. Anaerobic titra- tion of phCoADR with NADH shows very similar spectral results to those obtained with NADPH (results not shown). During anaerobic reduction with dithionite (data not shown) the FAD becomes reduced upon the addition of a second equivalent of dithionite, consistent with initial reduction of a nonflavin redox center, fol- lowed by reduction of the FAD by the strongly redu- cing dithionite. While the spectral changes upon addition of NADPH appear to correlate well with enzyme states corresponding to the addition of roughly 1 and 2 equivalents of NADPH, it is difficult to determine directly from the spectrum the fate of the reduced pyr- idine nucleotide. During the addition of the first equiv- alent of NADPH, there is a blue shift in the 380 nm peak and subsequent small increase in absorbance at 340 nm, although the increase is much less than that expected for the amount of NADPH added (the increase corresponds to the addition of 15 lm NADPH, when 50 lm NADPH had been added). When the titration is performed at room temperature, no additional absorbance is observed at 340 nm during the addition of the first equivalent of NADPH con- firming that the pyridine nucleotide is consumed during the addition of the first equivalent (data not shown). Determination of free thiol content In order to characterize the redox state of the pro- posed active site cysteine on the enzyme as purified and after addition of 0.8 equivalent of NADPH, as well as any small molecular mass thiols released by the enzyme, we used the thiol specific reagent DTNB. Because the enzyme contains only one cysteine residue, there is only one possible reactive thiol on the enzyme, in addition to any small molecular mass thiol trapped in the form of a mixed-disulfide with the cysteine (Cys- S-S-R). The results of these experiments are shown in Table 3. The enzyme as purified contains 0.00 equiva- lents of DTNB reactive thiol. Following anaerobic reduction with NADPH, less than 0.01 equivalent of small molecular mass thiol was detected, indicating that little if any of the enzyme is purified in the mixed disulfide form. If the NADPH reduced enzyme is kept anaerobic and assayed for thiol content, 0.85 equival- ent of thiol is detected, indicating that reduction by NADPH produces a reactive thiol. If the enzyme is exposed to air following reduction, this thiol becomes unreactive, suggesting that it is rapidly oxidized to the sulfenic acid. Determination of CoA levels and relative stability in P. furiosus To determine whether CoA might play a role as a pool of reducing equivalents in Pyrococcus, as suggested by the presence of a CoADR homologue in all three pyrococcal genomes, cells of P. furiosus were assayed for small molecular thiols. CoA was present at a con- centration of 0.64 lmol CoAÆg dry weight )1 , which corresponds roughly to 210 lm CoA in the cell (assu- ming a wet ⁄ dry weight ratio of 3 : 1 [13]). Glutathione was not detected. CoA was present almost entirely 25 50 100 75 50 25 0 75 100 Temperature % Activity Fig. 3. NADH oxidase (d) and CoA disulfide reductase (j) activity of phCoADR at varying temperatures. Activities at 81.3 (oxidase) and 85.0 °C (disulfide reductase) were set at 100%. Table 3. Equivalents of free thiol, as detected by DTNB. Equivalents of free thiol phCoADR, as purified 0.00 NADPH reduced phCoADR, anaerobic 0.85 Small molecular mass thiol released upon NADPH reduction 0.01 NADPH reduced, exposed to air 0.00 D. R. Harris et al. CoADR from P. horikoshii FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS 1193 in the thiol form, as there was < 5% detected in the disulfide form, as determined by the concentration of (thiol + disulfide)–(thiol). CoA has previously been shown to be four times more stable than glutathione and 180 times more sta- ble than cysteine to the Cu 2+ -catalyzed oxidation to the disulfide form in the presence of oxygen [14]. Because CoA is significantly more complex than either cysteine or glutathione, we were interested in determin- ing whether the increased stability to overoxidation extends to higher temperatures. For this reason, the oxidation of thiols at a high temperature was assayed. It was found that CoA oxidizes approximately 4.4 times slower than glutathione at 88 °Cin50mm imi- dazole buffer, 1.0 lm Cu 2+ pH 7.50, as shown in Fig. 4 (under these conditions cysteine oxidizes too quickly to be measured using conventional techniques). Under conditions in which the copper and any trace contaminating metals were chelated with EDTA and phosphate (50 mm potassium phosphate, 0.50 mm EDTA, pH 7.50), no significant oxidation of CoA was observed over 100 min at 88 °C. Discussion The phCoADR is only the second CoADR character- ized to date, and is the first from a hyperthermophile and Archaeon. Moreover, the enzyme displays distinct differences from the S. aureus CoADR. The enzyme is quite distinct from the staphylococcal enzyme in its ability to use both NADPH and NADH as substrates. With the S. aureus enzyme, substitution of NADH as the reducing nucleotide gives approximately 17% of the rate observed with NADPH [12], while the pyro- coccal enzyme gives almost identical k cat values with either reduced nucleotide substrate. The K m app of < 9 and 73 lm for NADPH and NADH, respectively, do differ significantly, with NADPH appearing to be the more efficient substrate. The reduced nucleotide concentrations have been determined for P. furiosus [15], and while the total concentration of NAD(P)(H) was found to be about half the value seen in the mesophilic Salmonella typhimurium, the pyrococcal NADP(H) ⁄ NAD(H) ratio was found to be more than twice that found in S. typhimurium. This result is not surprising, given the presence of several unique NADP + -dependent cata- bolic enzymes present in Pyrococcus [15]. The k cat of 7.2 s )1 determined for the phCoADR with NADPH as the reducing substrate is similar to the k cat of 27 s )1 observed for the staphylococcal enzyme, although it is lower [12]. Stabilization of enzyme intermediates Different enzymes within the disulfide reductase family stabilize different intermediates upon anaerobic reduc- tion with their nucleotide substrates. The H 2 O-produ- cing NADH oxidases are reduced by their nucleotide substrates by four electrons to an EH 4 or EH 4 ÆNAD + form, as shown in Scheme 1 [16,17]. In this case the first reducing equivalent reduces the active site cysteine residue, which serves as a nonflavin redox center, from the oxidized sulfenic acid to a reduced thiol species. This 2e – reduced form of the enzyme is referred to as the EH 2 species [17]. Addition of a second equivalent of reduced nucleotide substrate results in reduction of the enzyme-bound FAD, forming the EH 4 or EH 4 ÆNAD(P) + species. In the case of a true oxidase, it makes sense for the enzyme to stabilize the EH 4 form, since the reduced flavin is reactive with O 2 . Other enzymes in the family, such as the NADH peroxidase or glutathione reductase, tend to stabilize the enzyme-reduced nucleotide complex referred to as EH 2 ÆNAD(P)H (Scheme 1) [18,19]. On the EH 2 Æ NADH complex of the NADH peroxidase from E. faecalis [20] (Npx: NADH + H + +H 2 O 2 fi NAD + +2H 2 O), the reducing equivalents of NADH are held in the nicotinamide ring on the re-side of the flavin isoalloxazine ring with the C4 position of the nicotina- mide ring located 3.49 A ˚ from N5 of the isoalloxazine ring. The active site cysteine thiolate is 3.48 A ˚ from N5 on the si-side of the flavin, which allows for oxida- tion of this residue by peroxide, followed by reduction by NADH via the flavin. This configuration ensures 0 0.4 0.6 0.8 1 100 200 300 400 500 Time (min) Fraction thiol remaining Fig. 4. Autoxidation of glutathione (d) and CoA (s) at high tem- perature (88 °C) in the presence of Cu 2+ . The observed rates of oxi- dation correspond to pseudo first order rate constants of 5.7 · 10 )3 and 1.3 · 10 )3 min )1 for glutathione and CoA, respectively. CoADR from P. horikoshii D. R. Harris et al. 1194 FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS that NADH is available to be used ‘on demand’ when a substrate appears while at the same time avoiding the stabilization of a reduced flavin intermediate that could react undesirably with O 2 . In Fig. 2, the redox state of the enzyme-bound FAD is indicated by the peaks at 380 and 460 nm, with for- mal reduction and ⁄ or charge transfer to the isoalloxa- zine ring resulting in a loss of absorbance at these wavelengths. During the addition of the first equivalent of NADPH there is a decrease in absorbance at 460 nm with a concomitant increase in long wave- length (> 510 nm) absorbance. The observed spectral changes are consistent with the reduction of a nonfla- vin redox center and the formation of an EH 2 species, with charge transfer possibly developing between the FAD and the conserved active site cysteine thiol or thiolate. During the anaerobic reduction of the S. aureus enzyme there is loss of roughly 50% of the absorbance at 452 nm. In that case, the loss of absorbance is attributed to an asymmetric reduction of the dimeric enzyme, with the net result of reduction by the first equivalent being an oxidized flavin ⁄ oxidized cys resi- due in the form of a mixed disulfide on one subunit and a reduced flavin ⁄ reduced cys in the form of a thiol on the other subunit [12]. The decrease of 460 nm absorbance during titration of the phCoADR corres- ponds to at most a 24% loss, which is inconsistent with the scheme proposed for the S. aureus CoADR. It is interesting to note, however, that the overall shape of the titration curve at 452 nm for the sta- phylococcal enzyme and 460 nm for the phCoADR is nearly identical, with most of the decrease in absorb- ance occurring during addition of the first reducing equivalent. Addition of a second equivalent of NADPH to the EH 2 form of phCoADR results in an increase in absorbance in the regions between 507 and 700 nm and 400 and 500 nm. There is also an increase and blue shift in the peak which corresponds to the 380 and 360 nm peaks of the E and EH 2 forms, respect- ively. This result is inconsistent with reduction of the FAD and consistent with the formation of an EH 2 Æ NADPH complex, although further characterization is currently underway to determine definitively the nature of this enzyme species. It is apparent, however, that lit- tle if any of the enzyme is stabilized in the FADH 2 form. It was this finding that led us to originally con- clude that this enzyme was not likely to act as an NAD(P)H oxidase in vivo. The difference in the reductive half reaction, inclu- ding the apparent lack of subunit asymmetry, and the difference in quaternary structure (the S. aureus enzyme is a dimer) suggest that the mesophilic and thermophilic enzymes, which operate at very different temperatures, may use different mechanisms. These differences are currently being investigated. Redox state of the single cysteine Based on homology to the single cysteine members of the disulfide reductase family, it seems likely that the CoADR uses its single cysteine to catalyze its redox chemistry. As purified, the CoADR cysteine is not reactive with DTNB. The observation that the cysteine becomes reactive following reduction with NADPH, along with the absence of small molecular mass thiols released upon reduction, leads to the conclusion that the enzyme is likely to be purified in the sulfenic acid form (Scheme 1). This result is not surprising, given that most of the enzyme is obtained in the apo form from the overexpression host. It is also not particularly surprising that this enzyme, whose physiological func- tion appears to be to serve as a CoADR, is easily oxidized to the sulfenic acid form, when the anaerobic nature of Pyrococcus is taken into account. While Pyrococcus may encounter some oxidative stress, it seems unlikely that it would be regularly subjected to Scheme 1. Possible redox states for single- cysteine containing disulfide reductases. During reduction from ‘E’ to ‘EH 2 ’ the hydride passes from the NAD(P)H through the FAD to the nonflavin redox center. An additional possible species, EH 2 -NAD(P) + ,is not shown. D. R. Harris et al. CoADR from P. horikoshii FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS 1195 the levels of O 2 present in ambient air. Even if it were, however, this apparent sulfenic acid state is freely reducible, so that its production would not seem to put the enzyme at a disadvantage. Physiological role of CoADR and NAD(P)H in Pyrococcus The maintenance of low intracellular levels of cysteine in organisms has been attributed to the avoidance of the hydrogen peroxide produced during the rapid O 2 - dependent oxidation to cystine, necessitating the use of other small molecular mass thiols such as glutathione for the maintenance of internal redox levels. The results presented above are consistent with a role for CoA in maintaining a reducing environment or serving as a pool of reducing equivalents at the very high tem- peratures and high concentrations of metals found in the natural environment of Pyrococcus. The observed CoA concentration of 0.64 lmol CoAÆg dry weight )1 compares favorably with the result of 1.1 lmolÆg )1 dry weight obtained for S. aureus [14], and the correspond- ing intracellular concentration of 210 lm CoA is well above the K m app of 30 lm for the CoA disulfide. The pfCoADR (nox A-2 in the annotation of the P. furio- sus genome) was found to be upregulated more than fivefold by growth on elemental sulfur (S 0 ) [21], a result which suggests a role for this enzyme in the eventual transfer of electrons to S 0 . Because CoA lev- els were determined in cells which were grown in the absence of S 0 , future studies will determine whether CoA levels increase during growth in the presence of S 0 . It is also worth noting that the low CoADR activ- ity noted in crude extracts of P. furiosus was obtained from cells grown in the absence of elemental sulfur, so it seems likely that greatly increased activity will be seen in extracts of cells grown in the presence of S 0 . The NAD(P)H dependence of enzymes in the disul- fide reductase family can be divided into two classes: (a) enzymes whose function appears to be the main- tenance of a high R-SH ⁄ R-S-S-R ratio, such as glutathione reductase, trypanathione reductase and thi- oredoxin reductase, which tend to be NADPH depend- ent; and (b) enzymes that appear to be involved more directly in the reduction of oxidizing compounds and in the regeneration of oxidized nucleotides for glycoly- sis, which show a distinct preference for NADH, such as the NADH oxidase, NADH peroxidase, the NOX1 of P. furiosus, alkyl hydroperoxide reductase, and lipo- amide dehydrogenase. The pyrococcal CoADR described in this work is able to efficiently utilize both NADPH and NADH, a result which is consistent with the unusual utilization of reduced nucleotide coenzymes by Pyrococcus. The central metabolism of this organism uses an unusual NADPH-dependent sulfide dehydrogenase which is capable of both the NADPH-dependent reduction of elemental sulfur and the NADP + -dependent oxida- tion of ferredoxin [22]. A unique NADPH-dependent alcohol dehydrogenase with wide substrate specificity and a strong preference for the reduction of alde- hydes to alcohols is also found in this organism [23]. This unusual mix of NADH and NADPH dependent reactions in catabolic processes may account for the finding that the pyrococcal CoADR, which would be expected to show a strong preference for NADPH, is able to efficiently utilize either reduced nucleotide substrate. Comparison of the phCoADR to the pyrococcal NADH oxidase There are now two members of the disulfide reductase family that have been characterized from Pyrococcus, the NADH oxidase and the CoADR. These two enzymes display unique biochemical and enzymatic properties that are not present in their mesophilic counterparts. The P. horikoshii CoADR and the P. furiosus NADH oxidase (NOX1) are 37% identical. Each appears to play distinct roles in the fermentative metabolism of Pyrococcus and each is regulated differ- ently [1,21]. The difference in their reactivity is reflec- ted in the behavior of these two enzymes in reductive anaerobic titrations. As shown in Fig. 2 and discussed above, upon addition of excess NAD(P)H the phCo- ADR forms what appears to be an EH 2 ÆNAD(P)H complex with little or no FADH 2 character (Scheme 1). This can be contrasted to the behavior of the P. furiosus NOX, which forms what appears to be an EH 4 ÆNAD + complex with a fully reduced FAD (Scheme 1) upon addition of two equivalents of NADH (data not shown). In-depth studies comparing the mechanisms of these two pyrococcal enzymes to each other and to their mesophilic counterparts will be presented in a future communication. Experimental procedures Growth of microorganisms P. furiosus (DSM 3638) was grown in a sea salts medium as previously described [24]. Cellobiose (30 mm), maltose (10 mm), pyruvate (40 mm), or tryptone (5 gÆL )1 ) was inclu- ded as primary carbon source. E. coli JM109(kDE3) and XL-1 were grown at 37 °C in TYP medium [yeast extract (16 gÆL )1 ), tryptone (16 gÆL )1 ), NaCl (5 gÆ L )1 ), and CoADR from P. horikoshii D. R. Harris et al. 1196 FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS K 2 HPO 4 (2.5 gÆL )1 )]. When required, the following anti- biotics were included in the medium: kanamycin (50 lgÆmL )1 ), ampicillin (50 lgÆmL )1 ), and tetracycline (15 lgÆmL )1 ). Cloning and expression of the gene encoding the P. horikoshii and P. furiosus CoADRs The structural gene encoding the CoADR [previously referred to as NOX2 (PH0572)] was amplified from P. hori- koshii or P. furiosus genomic DNA with the oligonucleo- tides TG100 (5¢-GGCCTCATGAAGAAAAAGGTCGTCA TAATT-3¢), and TG101 (5¢-GGCCAAGCTTCTAGAAC TTGAGAACCCTAGC-3¢) (for the P. horikoshii CoADR), and TG104 (5¢-CGCGCCATGGAAAAGAAAAAGGTA GTCATAA-3¢) and TG105 (5¢-CGCGGTCGACCTAGAA CTTCAAAACCCTGGC-3¢) for the P. furiosus CoADR. The N terminus of the CoADRs were based on the pres- ence and proper spacing of the ribosome binding site, anno- tation from the genome sequence, and their good agreement with the NOX homologues from mesophilic sources that had been characterized in detail. PCR amplifi- cation was carried out using Pfu polymerase (Stratagene, La Jolla, CA), and the resulting 1.3 kb PCR product was cloned in the NcoI and HindIII sites of pET-24d. The resulting plasmids, pSMC002 (P. horikoshii) and pSMC004 (P. furiosus) were transformed into E. coli BL21(kDE3). For production of the recombinant enzymes the strain JM109 (kDE3) was used. Production and purification of recombinant phCoADR The phCoADR was overexpressed in E. coli JM109(kDE3). A fermentor growth with 5 L TYP media at 37 °C and kanamycin (50 lgÆmL )1 ) was started with a 2% inoculation of overnight culture. When the culture reached D 600 ¼ 1.0 ( 4.0 h) isopropyl thio-b-d-galactoside was added to a final concentration of 1.0 mm to induce enzyme production. The culture was incubated for an additional 4.0 h and the cells were collected by centrifugation at 8000 g at 4 °C for 15 min. The cells were washed with 50 mm Tris buffer pH 7.5 and collected by centrifugation (8000 g,4°C, 15 min). The resulting cells ( 20 g) were frozen at )80 °C until purification. The thawed cells were resuspended in 50 mm Tris buffer pH 7.5 (70 mL total volume). Eight-milliliter aliquots of cell solution were disrupted by 3 · 1 min sonication treat- ments. Cell debris was removed by centrifugation at 15 000 g and 4 °C for 15 min. The pellet was resuspended and sonication treatments were repeated (20 mL total vol- ume) and cell debris was removed by centrifugation (15 000 g,4°C, 15 min). The supernatant was then incuba- ted at 85 °C for 10 min and placed on ice. Heat denatured protein was removed by centrifugation at 15 000 g and 4 °C for 20 min. Chromatography was performed using an AKTA low- pressure chromatography system (FPLC) from Pharmacia Biotech (Piscataway, NJ). The supernatant was applied to a 25 mL Q-Sepharose Hi-Load column (Pharmacia Biotech) equilibrated with 50 mm Tris buffer pH 7.5 at 3.0 mLÆ min )1 . The column was washed with 60 mL 0.1 m KCl 50 mm Tris pH 7.5 to remove excess protein. phCoADR was eluted by a 120-mL linear gradient from 0.10 to 0.45 m KCl, with phCoADR eluting between 0.27 and 0.32 m KCl. Fractions containing phCoADR were pooled and reconsti- tuted with a final concentration of 1 mm FAD by incuba- tion at 85 °C for 10 min. The reconstituted pool was then concentrated and rinsed with 50 mm Tris buffer pH 7.50 by ultrafiltration with 30 k molecular mass limit filtration tubes at 3700 g,4°C. The concentrate (total volume 1.7 mL) was applied to a 120-mL Sephacryl S-200 HR (Pharmacia Biotech) size- exclusion column equilibrated with 50 mm Tris buffer pH 7.5 at 0.5 mLÆmin )1 . Fractions containing phCoADR were pooled (total volume 20 mL). The procedure yields 35 mg of enzyme with little loss between the steps following heat treatment ( 30% of activity is lost during the heat treatment, which is assumed to be due to association of the enzyme with the large pellet of heat denatured protein obtained at this stage). The enzyme shows a single band corresponding to a m ¼ 49 k on SDS ⁄ PAGE (Fig. 5) and Fig. 5. Coomassie blue stained SDS ⁄ PAGE analysis of the purifica- tion of the recombinant P. furiosus CoADR. Lane 1, crude extract (15 lg); lane 2, heat-treated extract (5 lg); lane 3, pooled fractions from the Q-sepharose column (5 lg); lanes 4 and 5, pooled frac- tions from the size-exclusion column (1 and 2 lg, respectively). The left lane contains marker proteins with the indicated subunit molecular mass (top to bottom): myosin, 195 k; b-galactosidase, 112 k; BSA, 59 k; carbonic anhydrase, 30 k. D. R. Harris et al. CoADR from P. horikoshii FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS 1197 gives a single peak by both conventional and HPLC size- exclusion chromatography. A summary of the purification is shown in Table 1. Figure 2 shows the UV ⁄ visible spec- trum of phCoADR, which is typical of a flavoprotein, with a 280 ⁄ 460 ratio of 11.1. Production and assay of the recombinant pfCoADR pfCoADR was overexpressed in E. coli JM109(kDE3). A 2.8-L Fernbach flask with 325 mL TYP media and kana- mycin (50 lgÆmL )1 )at37°C was started with a 2.5% inoculation of overnight culture. Cells were grown, protein production was induced, and cells were harvested as des- cribed above. Cells were disrupted by sonicating for 3 · 30 s on ice, followed by reconstitution with FAD (1.0 mm)at85°C for 10 min. E. coli proteins that precipi- tated during the heat treatment were removed by centrifu- ging in a microcentrifuge for 15 min at 14 000 g. Standard assays Standard assays were conducted at 75 °C (cuvette tempera- ture) in 1.5-mL cuvettes. All substrates, including oxidized CoA and dephosphoCoA, were from Sigma (St. Louis, MO). Enzyme and substrate(s) were added to preheated buffer for a total reaction volume of 840–950 lL. Oxidation of NAD(P)H was monitored at 340 nm. The standard oxid- ase assay used to monitor the purification of phCoADR was conducted with 100 lm NADH and FAD as above. A unit of CoADR activity is defined as the amount of enzyme required to reduce 1 lmol of CoA disulfideÆmin )1 in 50 mm potassium phosphate pH 7.5, 230 lm CoA disulfide and 100 lm NADPH at 75 °C. Protein and enzyme-bound FAD extinction coefficient determinations Protein concentrations were determined using a modified Bradford assay (Bio-Rad Protein Assay, Hercules, CA). The extinction coefficient of the enzyme-bound FAD in 50 mm Tris pH 7.50 was determined using the following equation: e 460 phCoADR ¼ (A 460 phCoADR) ⁄ [(A 450 of FAD released from phCoADR in 15% (v ⁄ v) TCA) ⁄ (e 450 FAD in 15% (v ⁄ v) TCA {9750 m )1 Æcm )1 })]. Quaternary structure determination The 120-mL Sephacryl S-200 HR (Pharmacia Biotech) size- exclusion column was used to determine the quaternary molecular mass of phCoADR. The column was eluted with 50 mm Tris buffer pH 7.5 at 0.5 mLÆ min )1 . The native molecular mass of phCoADR was determined by compar- ison to standards of mass ¼ 670 k (thyroglobulin), 158 k (gamma globulin), 44 k (ovalbumin), and 17 k (myoglobin). This result was confirmed by size exclusion HPLC chroma- tography on a 300 · 7.8 mm Bio-Rad Bio-Sil SEC-400-5 column. Determination of enzyme free and released thiols DTNB was used to detect both protein and small molecular mass thiols. All determinations used a final concentration of DTNB of 200 lm . Small molecular mass thiols that might be in the form of mixed disulfide on the enzyme (Cys-S-S-R) were determined by reducing the enzyme (56 lm) in an anaerobic titration with NADPH, followed by cooling the cuvette on ice, opening the cuvette to air, and immediately centrifuging the enzyme on a 30 k mole- cular mass cut-off filter to separate the enzyme from small molecular mass thiols. The flow-through from the filtration experiment was assayed for thiol content (providing a measure of small molecular mass thiols), while the retentate was rinsed three times with 50 mm potassium phosphate, 0.5 mm EDTA pH 7.5 and also assayed for thiol content (providing a measure of air-stable, accessible protein thiols). In order to detect protein thiols that were not stable in air (such as those that may be converted to sulfenic acids), the enzyme was reduced by 0.8 equivalents of NADPH in an anaerobic titration at room temperature (because of a slow turnover of the enzyme with DTNB, more than one equiv- alent of NADPH would result in a DTNB reductase activ- ity), followed by a tip to DTNB. The enzyme as purified was assayed for reactive thiols simply by adding DTNB, as above. Thiol stability assays Thiol assays were performed using DTNB as a thiol detec- tion reagent in a modification of a previously published pro- cedure [25]. An aliquot of the thiol to be assayed (generally 100 lL) was removed and added to 900 lL of assay buffer containing (final concentrations) 5 mm DTNB, 0.25 m potassium phosphate, 0.5 mm EDTA pH 7.2. After a 5-min incubation the absorbance was read at 412 nm. Thiol con- tent was determined by comparison to a standard curve determined under the same conditions with an extinction coefficient of 13 180 m )1 Æcm )1 , which is in good agreement with the previous literature values of 13 600 m )1 Æcm )1 [25]. Thermostability and thermoactivity assays Thermostability assays were performed at )80 °C, ) 20 °C, 85 °C and 95 °C over extended periods of time. An aliquot of enzyme was placed in a 2.0-mL microcentrifuge tube and sealed with Parafilm. For lower temperatures, the tube was placed in the respective freezer and standard activity assays were performed after 1 month. At higher temperatures, a CoADR from P. horikoshii D. R. Harris et al. 1198 FEBS Journal 272 (2005) 1189–1200 ª 2005 FEBS [...]... R Harris et al preweighed microcentrifuge tube containing enzyme was set into a covered beaker along with a damp tissue to prevent evaporation Both the mass of the tube and activity of the enzyme were monitored over time using standard NADH oxidase assay methods It was necessary to correct for the mass of the tube due to concentration of the enzyme solution from the small amount of evaporation that... evaporation that took place at high temperature Thermoactivity assays were done at temperatures ranging from 30 to 85 °C Background rates have been subtracted from each assay at the respective temperatures Anaerobic titrations Anaerobic titrations were conducted as described previously [26] The temperature of 60 °C was chosen (rather than higher temperatures) to minimize the background rate of NAD(P)H decomposition... centrifuging in a microcentrifuge for 15 min at 14 000 g, and the supernatant was analyzed by HPLC Monobromobimane derivatives of thiol standards were prepared as described previously [28] To determine disulfide content, cell extracts were treated with dithiothreitol (5 mm) to reduce the disulfides, and assayed as above Thiols were quantified by comparison of peak areas to standard curves of those standards Coenzyme. .. Staphylococcus aureus Purification and characterization of the native enzyme J Biol Chem 273, 5744– 5751 Pan G, Verhagen MF & Adams MW (2001) Characterization of pyridine nucleotide coenzymes in the hyperthermophilic archaeon Pyrococcus furiosus Extremophiles 5, 393–398 Mallett TC, Parsonage D & Claiborne A (1999) Equilibrium analyses of the active-site asymmetry in enterococcal NADH oxidase: role of. .. Coenzyme A was measured as the sum of the areas of the CoA and dephospho CoA peaks, as described previously [14] Acknowledgements C Davis, K Dorschner and C Hummel provided valuable technical support, R Fahey and G Newton provided helpful discussion regarding intracellular thiol detection methods This research was supported by an award from Research Corporation T.C.M was supported by NIH Grant GM35394 to Al... Fahey RC (1998) Mycothiol biosynthesis and metabolism Cellular levels 1199 CoADR from P horikoshii 14 15 16 17 18 19 20 of potential intermediates in the biosynthesis and degradation of mycothiol in Mycobacterium smegamatis J Biol Chem 273, 30391–30397 delCardayre SB, Stock KP, Newton GL, Fahey RC & Davies JE (1998) Coenzyme A disulfide reductase, the primary low molecular weight disulfide reductase from. .. Bacteriol 181, 1163–1170 24 Ward DKS, van der Oost J & de Vos W (2000) Purification and characterization of the alanine aminotransferase from the hyperthermophilic Archaeon Pyrococcus furiosus and its role in alanine production J Bacteriol 182, 2559–2566 25 Jocelyn PC (1987) Spectrophotometric assay of thiols Methods Enzymol 143, 44–67 26 Crane E, Parsonage D, Poole L & Claiborne A (1995) Analysis of. .. & Brandt KG (1980) Kinetic studies of the mechanism of pyridine nucleotide dependent reduction of yeast glutathione reductase Biochemistry 19, 4569– 4575 Yeh JI & Claiborne A (2002) Crystal structures of oxidized and reduced forms of NADH peroxidase Methods Enzymol 353, 44–54 1200 D R Harris et al 21 Schut GJ, Zhou J & Adams MWW (2001) DNA microarray analysis of the hyperthermophilic Archaeon Pyrococcus. .. sp nov., a new hyperthermophilic archaeon isolated from a deep-sea hydrothermal vent Arch Microbiol 160, 338–349 11 Fahey RC (2001) Novel thiols of prokaryotes Annu Rev Microbiol 55, 333–356 12 Luba J, Charrier V & Claiborne A (1999) Coenzyme A- disulfide reductase from Staphylococcus aureus: evidence for asymmetric behavior on interaction with pyridine nucleotides Biochemistry 38, 2725–2737 13 Anderberg... isolated from a hydrothermal vent at the Okinawa Trough Extremophiles 2, 123–130 9 Fiala G & Stetter KO (1986) Pyrococcus furiosus sp nov represents a novel genus of marine heterotrophic archaebacteria growing optimally at 100°C Arch Microbiol 145, 56–61 10 Erauso G, Reysenbach A, Godfroy A, Meunier JR, Crump B, Partensky F, Baross JA, Marteinsson V, Barbier G, Pace NR & Prieur D (1993) Pyrococcus abyssi . CoADR), and TG104 (5¢-CGCGCCATGGAAAAGAAAAAGGTA GTCATAA-3¢) and TG105 (5¢-CGCGGTCGACCTAGAA CTTCAAAACCCTGGC-3¢) for the P. furiosus CoADR. The N terminus of the. Discovery and characterization of a Coenzyme A disulfide reductase from Pyrococcus horikoshii Implications for the disulfide metabolism of anaerobic hyperthermophiles Dennis