Formate Dehydrogenase from Clostridium thermoaceticum - Journal of

JOURNAL OF BACTERIOLOGY, Oct. 1974, p. 6-14 Copyright 0 1974 American Society for Microbiology

Vol. 120, No. 1 Printed in U.S.A.

Nicotinamide Adenine Dinucleotide Phosphate-Dependent Formate Dehydrogenase from Clostridium thermoaceticum: Purification and Properties JAN R. ANDREESEN1 AND LARS G. LJUNGDAHL

Department of Biochemistry, University of Georgia, Athens, Georgia 30602 Received for publication 30 May 1974

The nicotinamide adenine dinucleotide phosphate (NADP)-dependent formate dehydrogenase in Clostridium thermoaceticum used, in addition to its

natural electron acceptor, methyl and benzyl viologen. The enzyme was purified to a specific activity of 34 (micromoles per minute per milligram of protein) with NADP as electron acceptor. Disc gel electrophoresis of the purified enzyme yielded two major and two minor protein bands, and during centrifugation in sucrose gradients two components of apparent molecular weights of 270,000 and 320,000 were obtained, both having formate dehydrogenase activity. The enzyme preparation catalyzed the reduction of riboflavine 5'-phosphate flavine adenine dinucleotide and methyl viologen by using reduced NADP as a source of electrons. It also had reduced NADP oxidase activity. The enzyme was strongly inhibited by cyanide and ethylenediaminetetraacetic acid. It was also inhibited by hypophosphite, an inhibition that was reversed by formate. Sulfite inhibited the activity with NADP but not with methyl viologen as acceptor. The apparent Km at 55 C and pH 7.5 for formate was 2.27 x 10-4 M with NADP and 0.83 x 10-I with methyl viologen as acceptor. The apparent Km for NADP was 1.09 x 10-' M and for methyl viologen was 2.35 x 10- 3M. NADP showed substrate inhibition at 5 x 10-3 M and higher concentrations. With NADP as electron acceptor, the enzyme had a broad pH optimum between 7 and 9.5. The apparent temperature optimum was 85 C. In the absence of substrates, the enzyme was stable at 70 C but was rapidly inactivated at temperatures above 73 C. The enzyme was very sensitive to oxygen but was stabilized by thiol-iron complexes and formate.

Oi the homoacetate fermentation of glucose or fructose by Clostridium thermoaceticum and Clostridium formicoaceticum, CO2 is used as an electron acceptor and is reduced to acetate (26, 34). The synthesis of the methyl group of acetate occurs via formate, tetrahydrofolate derivatives, and a methyl corrinoid (3, 20, 22). The first evidence that formate was an intermediate was obtained by Lentz and Wood (17). They observed that resting cells of C. thermoaceticum catalyze a rapid isotope exchange between 14CO2 and formate and that radioactive formate in the presence of a large pool of unlabeled CO2 is preferentially incorporated into the methyl group of acetate. Later it was found with cell-free extracts that addition of a pool of formate lowers the incorporation of 14CO2 into the methyl group of acetate (19). In C. thermoaceticum the reduction of CO2 to formate most likely is catalyzed by a nicotina-

mide adenine dinucleotide phosphate (NADP)dependent formate dehydrogenase (18). This enzyme catalyzes an NADP-dependent exchange between "4CO2 and formate (18) and also a net formation of formate from CO2 with either reduced NADP (NADPH) or reduced methyl viologen as electron donor (29; J. R. Andreesen and L. G. Ljungdahl, Bacteriol. Proc., p. 257, 1971). The reduction of CO2 to formate with NADPH is thermodynamically very unfavorable (COJformate E'0 = -0.42 V and NADP+/ NADPH El, = -0.324 V), and therefore the mechanism of the enzyme-mediated electron transfer is of considerable interest. Recently it was observed that the level of formate &hydrogenase in C. thermoaceticum is greatly increased by the addition of tungstate or molybdate and selenite to the growth medium (2). Evidence was also obtained showing that selenium is incorporated into the enzyme. That molybdate and selenite stimulate the formation

'Present address: Institut fur Mikrobiologie der UniversitSit, 3400 Gottingen, Grisebachstrasse 8, West Germany.

6

VOL. 120, 1974

FORMATE DEHYDROGENASE FROM C. THERMOACETICUM

of formate dehydrogenase has also been shown with Escherichia coli (8, 25, 27) and more recently with C. formicoaceticum (1). In the latter organism, as with C. thermoaceticum (2), tungstate stimulates the formation of formate dehydrogenase more than molybdate does. The formation of formate dehydrogenase is also stimulated by selenium in Clostridium sticklandii and Methanococcus vannielii (28) and by molybdenum in Clostridium pasteurianum (30). These results may indicate that the electron transfer catalyzed by formate dehydrogenase is mediated by enzyme-bound metals. In this communication we report catalytic and other properties of purified preparations of formate dehydrogenase from C. thermoaceticum. MATERIALS AND METHODS Culture methods. C. thermoaceticum was grown in 20-liter carboys in the medium described by Ghambeer et al. (11) or in medium 2 described by Andreesen et al. (3) to which had been added ferrous ammonium sulfate (5 x 10-4 M), sodium molybdate (5 x 10-4 M), and sodium selenite (10-6 M). Cells grown in the latter medium had about a 100 times higher level of formate dehydrogenase compared with cells grown in the medium of Ghambeer et al. Usually the cells were harvested near the end of the logarithmic growth phase, and with the medium of Andreesen et al. about 15 g of wet cells was obtained per liter. The cells were either used immediately or stored frozen at -20 C. During storage for 5 months the frozen cells lost about 50% of their formate dehydrogenase activity. Spectrophotometric assay of formate dehydrogenase. The formate dehydrogenase from C. thermoaceticum uses NADP, methyl viologen, or benzyl viologen as electron acceptor. The enzyme was assayed by following the reduction of NADP at 340 nm (e = 6.22 x 10'), of methyl viologen at 600 nm (e = 1.13 x 104) (31) or at 340 nm (e = 6 x 103) (16), or of benzyl viologen at 605 nm (, = 1.47 x 104) (6). The reaction mixture contained, in millimolar concentration: triethanolamine (TEA)-maleate, pH 7.5 (100); sodium formate (20); dithiothreitol (1); and NADP (2); or either methyl or benzyl viologen (20). The assay volume was 1 ml. One unit of enzyme is defined as the amount that reduces 1 umol of NADP (twoelectron acceptor) or 2 gmol of methyl or benzyl viologen (one-electron acceptor) per min, and activities are expressed in units per milligram of protein. The enzyme was normally assayed at 55 C and under oxygen-free nitrogen. All solutions were made with oxygen-free water and stored anaerobically under nitrogen. The cuvettes were sealed with serum stoppers and flushed with nitrogen. All transfers were done with syringes. The temperatures of the cuvettes and the rate of the reaction were monitored with a

7

Gilford 2000 spectrophotometer. Several buffers such as the potassium salts of phosphate, pyrophosphate, or maleate were tested instead of TEA-maleate. In all cases the activity was higher with the latter buffer. The rate of the enzyme reaction was proportional to the amount of enzyme, and normally the reaction started without the lag as observed earlier (18). In reactions using flavines, the following molar extinction coefficients were used: riboflavine 5'-phosphate, 12.2 x 10-'; reduced riboflavine 5'-phosphate, 870; flavine adenine dinucleotide, 11.3 x 103; and reduced flavine adenine dinucleotide, 980, all at 450 nm. Analytical methods. Analytical disc gel electrophoresis was performed according to Brewer and Ashworth (5). Ultracentrifugation in sucrose gradients was done according to Martin and Ames (21). Protein determinations were done by precipitation with trichloroacetic acid (final concentration, 1.5 N) and by reading the turbidity at 660 nm. Bovine serum albumin was used as standard. Purification of formate dehydrogenase from C. thermoaceticum. Formate dehydrogenase from C. thermoaceticum was purified essentially as described earlier (2) by using more rigid anaerobic conditions. The 0.1 M TEA-maleate buffer, pH 7.5, containing 0.025 M sodium thioglycolate, 0.01 M sodium formate, 2 mM ferrous ammonium sulfate, and 0.1 mM ethylenediaminetetraacetic acid (EDTA) was used throughout the purification. This buffer has the advantage that oxygen present in the buffer is removed rapidly through the action of formate dehydrogenase itself and by the ferrous ammonium sulfatethioglycolate complex. Furthermore, the buffer is yellow in the absence of oxygen and reddish in its presence. Therefore, by observing the color of tbp buffer one can ascertain whether the system is reduced. The purification involved preparation of crude extract by using a French pressure cell, fractionation with ammonium sulfate, and chromatography on a Bio-Gel A-5M column, a diethylaminoethyl (DEAE)-cellulose column, and a Sephadex G-200 column. Previously (2), the French pressure cell extract was centrifuged at 37,000 x g for 30 min; we now centrifuged it for 60 min at 100,000 x g. rAie supernatant still contained all the formate dehy*ogenase activity. No other modification was done in the purification method except that the enzyme from the Sephadex G-200 column was further purified by using a DEAE-cellulose column and gradient elution. Thus the fractions containing the highest activity from the Sephadex G-200 column were pooled and diluted with de-aerated water (1:1). This solution was applie6 to a DEAE-cellulose column (1 by 10 cm) previously equilibrated with 0.05 M TEA-maleate buffer containing sodium thioglycolate, sodium formate, ferrous ammonium sulfate, and EDTA. The enzyme was t*n eluted by using a gradient from 0.05 to 0.2 M of the TEA-maleate buffer, pH 7.5. Only one protein peak eluted, and the activity completely coincided with the protein. Further attempts to purify the enzyme using ammonium sulfate fractionation, ultrafiltration, or ultracentrifugation were unsuccessful and resulted in loss of enzyme activity.

ANDREESEN AND LJUNGDAHL

8

J . BACTrERIOL

combined as described in Materials and Methods. In maleic acid adjusted to pH 7.5 with Extraction of formate dehydrogenase from either TEA, tris(hydroxymethyl)aminomethC. thermoaceticum. Formate dehydrogenases ane, or imidazole, the formate dehydrogenase is have been reported to be either soluble or more stable than in potassium maleate also at particulate enzymes in various microorga- pH 7.5. The stability is somewhat lower at pH nisms. Several methods were tested to find 7.0, and at pH 8.0 activity is rapidly lost. In the best condition for the extraction of the buffers with phosphate replacing maleate the enzyme from C. thermoaceticum (Table 1). C. enzyme is less stable, and in buffers containing thermoaceticum is gram positive (9), and for- pyrophosphate or citrate enzyme activity is mate dehydrogenase is released by treatment rapidly lost. Sulfhydryl compounds like cyswith lysozyme and deoxyribonuclease. How- teine, glutathione, dithiothreitol, mercaptoever, the best results were obtained either by ethanol, and sodium thioglycolate increase the combining the lysozyme and deoxyribonuclease stability of the enzyme. The stability of formate treatment with the French pressure cell used at dehydrogenase is enhanced also by 0.1 mM 105 kg/cm2 or by using the French pressure cell EDTA and 33% glycerol. Purified preparation of alone at 490 kg/cm2. There was no formate the enzyme were always less stable than crude dehydrogenase activity in the pellet after cen- preparations, and normally most of the activity trifugation at 100,000 x g for 60 min, and the was lost within a few days. Activation of an enzyme that lost activity was addition of the pellet to the supernatant did not increase the activity. We conclude from these observed sometimes, and crude extracts with results that formate dehydrogenase in C. ther- little activity stored overnight at 4 C anaerobimoaceticum is a soluble eiizyme. cally in the TEA-maleate buffer containing Stability of formate dehydrogenase from C. thioglycolate, formate, EDTA, and ferrous amthermoaceticum. Formate dehydrogenase from monium sulfate often increased in specific acC. thermoaceticum is very sensitive to oxygen tivity. The specific activity of purified prepara(18). This instability was observed with the tions was often also increased by adding 5 x 10-2 M cysteine or reduced glutathione and 106 enzyme prepared from cells grown in medium containing little or no iron, molybdate, or sele- M vitamin B12 to the solution. This had been nite (medium of Ghambeer et al. [11]) or in a observed earlier (18) and is likely due to the medium containing these metals. Although the removal of traces of oxygen catalyzed by the B12 specific activity of formate dehydrogenase was (23). However, reactivation of an enzyme to full much higher in extracts from cells grown in the original activity was never achieved. Purification of formate dehydrogenase. A presence of the metals (the increase observed was from 0.002 to about 2 units per mg of summary of a purification of formate dehyprotein [2, 18D, the metals did not effect the drogenase from 30 g of wet cells of C. sensitivity of the enzyme to oxygen. The en- thermoaceticum is given in Table 2. In step IV zyme is stabilized by adding sulfhydryl comthe enzyme was concentrated by adsorption on pounds, ferrous ammonium sulfate, and the a small DEAE-cellulose column and activity substrate formate to the buffer. The best results was lost normally, but the step was needed since were obtained when these compounds were it removed inert proteins and concentrated the

RESULTS

TABLE 1. Solubilization of formate dehydrogenase activity from C. thermoaceticuma Treatment

1. 2. 3. 4. 5. 6.

Freezing and thawing ................................ .................... Lysozyme plus DNase ............ Lysozyme plus DNase and EDTA ..................... French pressure cell (105 kg/cm2) ..................... As 2 in combination with 4 ........................... French pressure cell (490 kg/cm2) .....................

Protein released

Formate dehydrogenase

3.0 7.0 9.5 11.4 23.5 26.0

0.086 0.235 0.342 0.274 0.363 0.342

(mg/ml)

activity (U/mg)

a Two grams of wet cells was suspended in 6 ml of 0.1 M TEA-maleate buffer, pH 7.5, containing 10 mM mercaptoethanol and 5 mM formate. Lysozyme (5 mg), deoxyribonuclease (DNase) (0.5 mg), and 1 mM EDTA were added as indicated. The mixture was incubated anaerobically for 30 min at 37 C, and 2 mM Fe(NH4) 2(SO4)2 was added after the incubation when EDTA was present. The treated cell suspension was centrifuged at 100,000 x g for 1 h. The supernatant was assayed for formate dehydrogenase activity by using NADP as electron acceptor.

VOL. 120, 1974


9

TABLE 2. Purification of formate dehydrogenase from 30 g of wet cells of C. thermoaceticum Step

I. II. III. IV. V. VI.

French pressure extract ....... ...... Ammonium sulfate, 0-42% fraction ... Bio-Gel A-5M column .............. DEAE-cellulose column X Sephadex G-200 column DEAE-cellulose gradient Combined fractions .............. Best fraction .....................

Sp act (U/mg)

Total protein (mg)

NADP

MVa

NADP

MV

Ratio NADP/MV activity

2,720 925 67

3,570 2,370 1,105

7,160 6,100 3,010

1.31 2.56 16.5

2.63 6.6 45

1:2 1:2.6 1:2.7

15

237

660

15.8

44

1:2.8

76 17

235 56

28 34

87 113

1:3.1 1:3.3

2.7 0.5

Total units

MV, Methyl viologen. ° The DEAE-cellulose column, step IV, was used to concentrate the enzyme, and figures refer to the enzyme obtained from the Sephadex G-200 column. a

enzyme. The loss of activity in step IV was minimized by avoiding exposure of the enzyme solution to air and by performing the step as fast as possible. The enzyme from the best fraction of step VI was investigated by analytical disc gel electrophoresis (5). The amount of protein applied was about 100 Ag. Several preparations were investigated and four bands were always observed, with two smaller bands between two major ones. Unfortunately we were unable to perform activity stains due to the instability of the enzyme. The enzyme was also investigated by ultracentrifugation in a sucrose gradient. Formyltetrahydrofolate synthetase (molecular weight, 245,000) and catalase (molecular weight, 240,000) were used as standards. In three experiments using methyl viologen as an acceptor, two peaks were obtained with about equal formate dehydrogenase activity. Peak I moved with an apparent molecular weight of 320,000, and peak II moved with a molecular weight of 270,000. The elution profiles (2) from Bio-Gel A-5M and Sephadex G-200 columns indicate a molecular weight for the enzyme of around 300,000. On the Bio-Gel A-5M column, the formate dehydrogenase activity elutes in an unsymmetrical peak that has one or two shoulders, indicating that there are at least two active species of formate dehydrogenase and the result is, thus, in agreement with the ultracentrifuge studies. The Sephadex G-200 column gives a rather symmetrical peak, which is not surprising since a molecular weight of 300,000 is close to the upper limit of resolution on Sephadex G-200. The absorption spectrum of the purified enzyme preparation did not reveal any exceptional characteristic between 310 and 600 nm except a shoulder of a plateau between 370 and 410 nm. However, there was a gradual

increase in the absorption going from 500 to 415 nm. During the purification of the enzyme, the activity was measured with both NADP and methyl viologen as acceptors. Separation of the two activities did not occur during the purification procedure (Table 2). A small increase was observed in the methyl viologen activity over the NADP activity. This was due to a somewhat faster inactivation of the NADP activity, which was also seen during storage or heating of the enzyme. Substrate specificity and reactions catalyzed by formate dehydrogenase. Besides

NADP, the following naturally occurring electron acceptors were tested with formate as

electron donor: ferredoxin from C. thermoaceticum, NAD, riboflavine 5'-phosphate, flavine adenine dinucleotide, cytochrome cs from Desulfovibrio gigas (a gift from J. LeGall), and cytochrome c from Torula yeast. Activity was observed only with NADP, and the formate dehydrogenase from C. thermoaceticum is therefore unique among known formate dehydrogenases in its specificity for NADP. Of other commonly used electron acceptors, activity was observed only with methyl and benzyl viologen, which were equally efficient. Methylene blue, ferricyanide, and phenazine methosulfate did not serve as electron acceptors. In addition to the reversible oxidation of formate with NADP or methyl viologen as electron acceptors, the preparation of formate dehydrogenase from C. thermoaceticum also catalyzed the reactions listed in Table 3 and a reversible reduction of methyl viologen with NADPH. The enzyme preparation catalyzed a

very efficient reduction of flavines (Table 3). The reaction rate was three to five times that of the formate dehydrogenase activity. Likewise

10


TABLE 3. Reactions catalyzed by purified formate dehydrogenase preparations from C. thermoaceticum Reactiona Formate oxidation: Formate- + NADP+ NAD-PH oxidation:

Rate (U/mg of protein)

CO2 + NADPH ........... 0.086

NADPH + H+ + 02- NADP+ + H202,

... .0.935

Flavine reduction: NAPPH + H+ + FMN NADP+ + FMNH2. 0.415 NADPH + H+ + FAD NADP+ + FADH2. 0.317 Formate oxidation coupled with flavine reduction: Formate + FMN N CO2 + FMNH2 ......... 0.077 Formate + FAD _NiDP4 CO2 + FADH2............ 0.070 a Reactions were performed in anaerobic cuvettes containing 1 ml of the regular assay buffer. Substrate (micromoles) was added as indicated: sodium formate, 20; NADP, 2; NADPH, 1; riboflavine 5'-phosphate (FMN) or flavine adenine dinucleotide (FAD), 0.15; and 0.085 mg of formate dehydrogenase. The NADPH-oxidase activity was measured aerobically. The enzyme was passed through a Sephadex G-25 column to remove formate present in the buffer just before use. This procedure was required to remove formate which was present as a stabilizer; however, the enzyme lost substantial activity during the chromatography.

the rate of NADPH-oxidase activity was very high. Formate could serve as reductant of flavines but only in the presence of NADP. The reaction rate was then about the same as for the formate dehydrogenase activity, clearly showing that NADPH is an intermediate in the reduction of the flavines by formate. Formate dehydrogenase catalyzed the reduction of methyl viologen with NADPH. The reaction was very fast and an equilibrium was quickly reached. The equilibrium was, as expected, far toward methyl viologen and NADPH. In a reaction mixture similar to those of Table 3 containing in 1 ml 10 umol of methyl viologen and 1.25 Amol of NADPH but no formate, 93 nmol of methyl viologen was reduced by the NADPH. That formate dehydrogenase is able to reduce methyl viologen with NADPH is in agreement with its capacity to reduce CO2 with NADPH. The oxidation-reduction potential (E'o) for methyl viologen is -0.44 V, which is somewhat lower than for COJformate (-0.42 V). Some assays were performed by using formate as electron donor with both methyl viologen and NADP present as electron acceptors. Both acceptors were reduced, but most of the electrons were accepted by NADP. This was likely a reflection of the equilibrium between methyl viologen and NADPH. It is also possible that NADP inhibits the reaction with methyl viologen as electron acceptor. The com-

J. BACTERIOL.

bined rate of the reaction with NADP and

methyl viologen equaled that of NADP alone. The reaction rate with methyl viologen alone was somewhat higher. Apparent Km values. The apparent Km values for purified formate dehydrogenase from C. thermoaceticum for formate were 2.27 x 10-I M with NADP and 0.83 x 10-4 M with methyl viologen as electron acceptor. The apparent Km for methyl viologen was 2.35 x 10-s M and for NADP was 1.09 x 10' M. All Km values were obtained at pH 7.5 and at 55 C. These values are in agreement with values previously published with less purified enzyme (18). Concentrations of NADP over 5 x 10- 3 M show substrate inhibition. Effect of pH. The effect of pH on the activity of the purified formate dehydrogenase was studied by using 0.1 M TEA-chloride buffer from pH 6.0 to 8.5 and 0.1 M glycine-potassium hydroxide buffer from pH 9.0 to 10.5. With NADP as electron acceptor, the enzyme was active from pH 6.0 to 10.5, with a broad maximum between pH 7.0 and 9.5. With methyl viologen as acceptor, the anzyme was inactive below pH 7.0. The activity then rapidly increased to pH 9. A formate-independent reduction of methyl viologen was observed at pH values above 8. This reduction was partially dependent on dithiothreitol, and over pH 9.5 this reduction was faster than the formate-dependent reaction. Methyl viologen is reduced in alkaline solutions (6). Requirement for sulfhydryl compounds and effect of metals on purified formate dehydrogenase. The formate dehydrogenase from C. thermoaceticum has an absolute requirement for a sulfhydryl compound for activity. Dithiothreitol, cysteine, glutathione, or thioglycolate at concentrations of 10-6 M satisfied this requirement. Mercaptoethanol, on the other hand, was inhibitory, especially with NADP as electron acceptor. Addition of dithiothreitol or the other active sulfhydryl compounds up to 10-2 M did not further stimulate the reaction, except that less active enzyme preparations sometimes were slowly activated by the high concentrations of the sulfhydryl compounds. Since ferrous ions greatly increase the cell yield when added to the growth medium, and since molybdate and selenite increase the formate dehydrogenase in cells of C. thermoaceticum, the effect of these metals on the purified enzyme was investigated by the regular assay except that the dithiothreitol concentration was 10- 5 M. There was no activation by any of these compounds. Ferrous ammonium sulfate

VOL. 120,1974


in concentrations between 10-5 to 10-2 M were slightly inhibitory (up to 20%) with both NADP and methyl viologen as acceptors. Ferric chloride (10-i M) did not inhibit the reaction, but a precipitate formed at higher concentrations. Sodium molybdate (10-2 M) had no effect on the reaction. Sodium selenite was not inhibitory at 10-8 M with NADP as electron acceptor, but at 10-2 M it completely-inhibited the reaction. With methyl viologen, selenite was inhibitory at 10-3 M. Sodium sulfite had no effect on the reaction with methyl viologen as acceptor but inhibited the NADP-dependent reaction. Thus with NADP the enzyme, after incubation for 5 min at 55 C, lost 20, 40, and 90% of its activity with 10-i3 10-2, and 10- I M sodium sulfite, respectively. Inhibition of formate dehydrogenase by hypophosphite, cyanide, and EDTA. Hypophosphite is a structural analogue to formate and is known to inhibit formate dehydrogenase (24). Sodium hypophosphite also inhibited formate dehydrogenase from C. thermoaceticum (Table 4). This inhibition was probably competitive and was completely reversed by formate. EDTA and sodium cyanide inhibited the formate dehydrogenase (Fig. 1). The inhibition with cyanide was observed at 10-6 M and was almost complete at 10-3 M. EDTA was only slightly inhibitory at 10-3 M but strongly inhibitory at concentrations over 10-2 M. The inhibition by cyanide and EDTA indicates a possible role of metals in the catalysis by formate dehydrogenase and is in agreement with our earlier data (2, 3). Effects of temperature. The activity of the formate dehydrogenase was measured by the standard assay with NADP and methyl viologen as acceptors at different temperatures (Fig. 2). With both acceptors an increase of the activity with temperature occurred up to 70 C. However, the increase was not smooth, as is evident by the breaks in the curves in Fig. 2. Breaks were seen both with NADP and methyl viologen as acceptors and both with purified enzyme and crude extract. The activity with NADP was somewhat more sensitive to heat inactivation then with methyl viologen. With the latter acceptor at 90 C the reaction rate was linear for the first 30 s, after which the reaction rate slowly decreased, indicating denaturation of the enzyme. In the absence of substrate the enzyme did not lose activity when incubated at 70 C for 10 min in 0.1 M TEA-maleate buffer, pH 7.0, at a protein concentration of 12 to 15 mg/ml. However, when the enzyme was incubated at 73 C it lost 66% of the activity during 10 min,

11

and at 76 C after 10 min the activity was only 10% of the original.

DISCUSSION Formate dehydrogenases have been reported that use ferredoxin, NAD, NADP, ubiquinones, cytochromes, and oxygen as natural acceptors and methyl viologen, benzyl viologen, tetrazolium salts, phenazine methosulfate, ferricyanide, methylene blue, dichlorophenolindophenol, and also oxygen as artificial acceptors. The formate dehydrogenase from C. thermoaceticum uses NADP as the natural electron acceptor, and it is the only known NADP-dependent formate dehydrogenase. It couples also with flavines but only in the TABLE 4. Inhibition of formate dehydrogenase by hypophosphite with NADP and methyl viologen as electron acceptorsa Hypophosphite

NADP

Methyl viologen

(M)

(nmol/min)

(nmol/min)

None 10-4 5 x 10-'

17.3 7.7 0.9

19.6 3.9

10-3

0

0 0

a Standard assay conditions were used except that sodium formate was 1 mM and sodium hypophosphite was added as indicated. In the reaction with NADP, 68 ug of protein was used, and with methyl viologen 14 /Ig of protein was used.

00 0

a.

-6 -5 -4 -3 -2 -l

Log1

Molarity

FIG. 1. Inhibition of formate dehydrogenase of C. thermoaceticum by cyanide and EDTA. Standard assay conditions were used with NADP (solid symbols) and methyl viologen (open symbols) as electron acceptors. Squares show activity with sodium cyanide, and circles show activity with EDTA. With NADP 15 Mg of protein was used, and with methyl viologen 3 Mg of protein was used.

120 E

E

J. BACTERIOL.


12

so

40

40

10

30

70

50

Temperature

(C

90

.

thetacctivityof ord

act FIG. 2. Effect of temperature on the formate dehydrogenase from C. thern moaceticum with NADP (solid circles) and methyl vi( cles) as electron acceptors. Standa assay conditions were used except the tempera ture was varied. With NADP 40 ug of protein was used, and with methyl viologen 8 ug of protein was used.

of NADP, and it is un able to couple with ferredoxin (18 and this relport). Thus, it differs from formate dehydrogen Lases present in other Clostridia. The enzyrmes from C. pasteurianum and C. acidiurici are ferredoxin dependent (14, 15). The latter al: so couples with NAD but only in the presence of iferredoxin. The natural electron acceptor for th( e enzyme of C. formicoaceticum has not been foound. There are no indications that it is ferredoxcin, flavodoxin, or rubredoxin (1), nor is it NADIP (22). It reacts with NAD, but the activity is lo w and is only a fraction of the activity observe d with methyl viologen as acceptor (1). The main problems encounterred in studying formate dehydrogenases, especieally those from bacterial sources, are instability and sensitivity toward oxygen; thus, few attemipts have been made to purify the enzyme fr om Clostridia. Kearny and Sagers (15) pu rified the C. acidiurici formate dehydrogenas4e 28-fold. They defined a unit of activity as tthe amount of protein required to produce Ean absorbancy change at 555 nm of 0.1 per mLin with benzyl viologen as acceptor in a volumeL of 3 ml. Their best preparation had an activit3 y of 901, which seems to correspond to an activ ity of 8.5 umol per min per mg. Li et al. (18), in a previous attempt to purify the C. thermtoaceticum enzyme, obtained an activity of 0 .0177 ,umol per min per mg using NADP as acceptor. This activity is very low and much lc wer than what now is found in crude extr-acts from C. thermoaceticum cells grown wit h selenite, mopresence

lybdate, and ferrous ammonium sulfate in the medium. In our purification of formate dehydrogenase, we obtained a specific activity of 34 with NADP as acceptor, which corresponds to an increase in activity of about 2,000 compared with earlier results (18). The enzyme from C. thermoaceticum with a specific activity of 34 is not homogeneous. Disc gel electrophoresis revealed two major and two minor protein bands, and during centrifugation in a sucrose gradient two active species of formate dehydrogenase were observed. It is interesting to note that similar results were obtained with formate dehydrogenase from Pseudomonas oxalaticus (13). This enzyme separates into two active species of different molecular weights (300,000 and 200,000), and during into two major disc electrophoresis it and three minor bands. The P. oxalaticus enzyme is an iron-sulfur flavoprotein (13). There is no evidence that the C. thermoaceticum enzyme contains a flavine, although this is not excluded. The spectrum does not show a pronounced absorption in the 450-nm region, which is typical for flavoproteins, but neither do purified preparations of formate dehydrogenase from P. oxalaticus in the presence of formate. The P. oxalaticus enzyme preparation has diaphorase and NADH oxidase activities (13), and similarly the C. thermoaceticum enzyme preparation exhibits diaphorase and NADPH oxidase activities. Such activities are also catalyzed by several other pyridine nucleotidedependent enzymes, including the nonheme iron flavoprotein nicotinic acid hydroxylase from Clostridium barkeri (12), the flavoproteins NADP-cytochrome f reductase from spinach chloroplasts (36), the ferredoxin-NADP reductases from Bacillus polymyxa (35) and the alga Bumilleriopsis filiformis (4), and also the NADrubredoxin reductase from Pseudomonas oleovorans (33). During the purification of formate dehydrogenase from C. thermoaceticum, the activities dependent on NADP and methyl viologen closely followed each other, and both activities were inhibited by EDTA and cyanide, strongly indicating that they are catalyzed by the same enzyme. However, the two activities respond somewhat differently to other inhibitors and heat treatment of the enzyme. Thus, sulfite inhibits the NADP but not the methyl viologendependent activity. The latter activity apears to be less sensitive to heat inactivation than is the NADP activity. The NADP activity has a broad pH optimum between 7 and 9.5, whereas the methyl viologen activity increases with pH from

separates

VOL. 120, 1974


7 to 9.5, at which pH non-enzyme reduction of methyl viologen is too fast to allow determinations of the enzyme activity. These differences may indicate that there are different binding sites for methyl viologen and NADP on the enzyme. This is further indicated by the observation that it catalyzes the reduction of methyl viologen by NADPH. Boger (4) found separate binding sites for methyl viologen and NADP on the ferredoxin-NADP reductase isolated from an alga. Most likely the formate dehydrogenase from C. thermoaceticum is a metal enzyme containing selenium, molybdenum, or tungsten, and perhaps also iron (2). The strong inhibition by cyanide as shown in this paper is further evidence for the presence of metals in formate dehydrogenase. Likewise the inhibition by EDTA supports this view. The enzymes from C. acidiurici and C. pasteurianum are also strongly inhibited by cyanide and EDTA, indicating that these enzymes contain a metal or metals (15, 30). Cyanide inhibits the molybdenum-containing xanthine oxidase and related enzymes, and it has been shown that cyanide reacts with the molybdenum (7). In contrast to the enzymes of C. acidiurici (15) and P. oxalaticus (13), the formate dehydrogenase from C. thermoaceticum is not light sensitive. Nitrate reductase from E. coli (10) contains nonheme iron and molybdenum. The purified formate dehydrogenase from C. thermoaceticum has the same nonheme iron-like spectrum as the E. coli nitrate reductase. The formate dehydrogenase is very oxygen sensitive and contains acid-labile selenide, and C. thermoaceticum requires iron for growth. These properties indicate that the formate dehydrogenase from C. thermoaceticum may contain nonheme iron. During purification of the formate dehydrogenase from C. acidiurici, the enzyme lost its capacity to couple with ferredoxin and coupled only with artificial electron acceptors like benzyl viologen (15). The purified enzyme from C. thermoaceticum, on the other hand, couples with the natural acceptor NADP as well as with methyl viologen. This may indicate that the C. thermoaceticum enzyme is more stable. Formate dehydrogenase is a fairly large enzyme (molecular weight about 300,000) and most likely consists of subunits. It is possible that subunits dissociate from the enzyme during purification, yielding a partially active enzyme. This may explain why the C. acidiurici enzyme loses its ability to couple with ferredoxin and also why at least two active species of different

13

molecular weights are seen of the C. thermoaceticum enzyme. The rapid inactivation of formate dehydrogenases may also be a reflection of loss of their metal contents as found for the enzyme from P. oxalaticus (13) and from C. thermoaceticum (2). Irreversible oxidation of essential thiol or other reduced groups may also be the cause of inactivation. A study of the role of metal and of the subunit structure of formate dehydrogenase from C. thermoaceticum should clarify these points. ACKNOWLEDGMENTS This investigation was supported by Public Health Service grant AM-12913 from the National Institute of Arthritis, Metabolism, and Digestive Diseases. A postdoctoral fellowship (AN 63/1) from Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, for J.R.A. is gratefully acknowledged. LITERATURE CITED 1. Andreesen, J. R., E. El Ghazzawi, and G. Gottschalk. 1974. The effect of ferrous ions, tungstate and selenite on the level of formate dehydrogenase in Clostridium formicoaceticum and formate synthesis from CO2 dur2.

3.

4.

5. 6. 7.

8.

9.

10.

11.

ing pyruvate fermentation. Arch. Microbiol. 96:103-118. Andreesen, J. R., and L. G. Ljungdahl. 1973. Formate dehydrogenase of Clostridium thermoaceticum: incorporation of selenium-75, and the effects of selenite, molybdate, and tungstate on the enzyme. J. Bacteriol. 116:867-873. Andreesen, J. R., A. Schaupp, C. Neurauter, A. Brown, and L. G. Ljungdahl. 1973. Fermentation of glucose, fructose and xylose by Clostridium thermoaceticum: effect of metals on growth yield, enzymes, and the synthesis of acetate from C0,. J. Bacteriol. 114:742-751. Boger, P. 1971. Zusammenhang zwischen Transhydrogenase- und Diaphorase Aktivitat der FerredoxinNADP-Reduktase und der photosynthetischen NADPReduktion. Z. Naturforsch. 26b:807-815. Brewer, J. M., and R. B. Ashworth. 1969. Disc electrophoresis. J. Chem. Educ. 46:41-45. Corwin, A. H., R. R. Arellano, and A. B. Chivvis. 1968. Anomalies of viologens in bases and water. Biochim. Biophys. Acta 162:533-538. Coughlan, M. P., K. V. Rajagopalan, and P. Handler. 1969.The role of molybdenum in xanthine oxidase and related enzymes. Reactivity with cyanide, arsenite, and methanol. J. Biol. Chem. 244:2658-2663. Enoch, H. G., and R. L. Lester. 1972. Effects of molybdate, tungstate, and selenium compounds on formate dehydrogenase and other enzyme systems in Escherichia coli. J. Bacteriol. 110:1032-1040. Fontaine, F. E., W. H. Peterson, E. McCoy, M. J. Johnson, and G. J. Ritter. 1942. A new type of glucose fermentation by Clostridium thermoaceticum, N. Sp. J. Bacteriol. 43:701-715. Forget, P. 1974. The bacterial nitrate reductases. Solubilization, purification, and properties of the enzyme A of Escherichia coli K12. Eur. J. Biochem. 42:325-332. Ghambeer, R. K., H. G. Wood, M. Schulman, and L. G. Ljungdahl. 1971. Total synthesis of acetate from CO2. m. Inhibition by alkylhalides of the synthesis from CO,, methyltetrahydrofolate, and methyl-B,2 by Clostridium thermoaceticum. Arch. Biochem. Biophys. 143:471-484.

14


J. BACTERIOL.

Methanobacillus omelianskii. J. Bacteriol. 75:356-359. acid metabolism. Ill. Purification and properties of a 25. Pinsent, J. 1954. The need of selenite and molybdate in the formation of formic dehydrogenase by members of nicotinic acid hydroxylase. J. Biol. Chem. 244:1194-1203. the coli-aerogenes group of bacteria. Biochem. J. 13. Hopner, T., and A. Trautwein. 1972. Some properties of 57:10-16. formate dehydrogenase. Z. Naturforsch. 27b: 1075- 26. Schulman, M., D. Parker, L. G. Ljungdahl, and H. G. 1076. Wood. 1972. Total synthesis of acetate from CO2. V. Determination by mass analysis of the different types 14. Jungermann, K., H. Kirchniawy, and R. K. Thauer. 1970. of acetate formed from 13CO2 by heterotrophic bacteFerredoxin dependent CO2 reduction to formate in ria. J. Bacteriol. 109:633-644. Clostridium pasteurianum. Biochem. Biophys. Rps. 27. Schum, A. C., and J. C. Murphy. 1972. Effects of Commun. 41:682-689. selenium compounds on formate metabolism and coin15. Kearny, J. J., and R. D. Sagers. 1972. Formate dehydrocidence of selenium-75 incorporation and formic dehygenase from Clostridium acidiurici. J. Bacteriol. drogenase activity in cell-free preparations of Esche109:152-161. richia coli. J. Bacteriol. 110:447-449. 16. Kosower, E. M., and J. C. Cotter. 1964. Stable free radicals. II The reduction of 1-methyl-4-cyanopyri- 28. Stadtman, T. C. 1974. Selenium biochemistry. Science 183:915-922. dinium ion to methyl viologen cation radical. J. Amer. 29. Thauer, R. K. 1972. C02-reduction to formate by Chem. Soc. 86:5524-5527. NADPH. The initial step in the total synthesis of 17. Lentz, K., and H. G. Wood. 1955. Synthesis of acetate acetate from CO2 in Clostridium thermoaceticum. from formate and carbon dioxide by Clostridium FEBS Lett. 27:111-115. thermoaceticum. J. Biol. Chem. 215:645-654. 18. Li, L.-F., L. Ljungdahl, and H. G. Wood. 1966. Properties 30. Thauer, R. K., G. Fuchs, U. Schnitker, and K. Jungermann. 1973. CO2 reductase from Clostridium of nicotinamide adenine dinucleotide phosphatepasteurianum: molybdenum dependence of synthesis dependent formate dehydrogenase from Clostridium and inactivation of cyanide. FEBS Lett. 38:45-48. thermoaceticum. J. Bacteriol. 92:405-412. 19. Ljungdahl, L., E. Irion, H. G. Wood. 1965. Total synthe- 31. Trudinger, P. A. 1970. On the absorbancy of reduced methyl viologen. Anal. Biochem. 36:22-224. sis of acetate from CO2. I. Co-methylcobyric acid and Co-(methyl)-5-methoxybenzimidazolylcobamide as in- 32. Turner, D. C., and T. C. Stadtman. 1973. Purification of protein components of the Clostridial glycine reductermediates with Clostridium thermoaceticum. Biotase system and characterization of protein A as a chemistry 4:2771-2779. selenoprotein. Arch. Biochem. Biophys. 154:366-381. 20. Ljungdahl, L. G., and H. G. Wood. 1969. Total synthesis of acetate from CO2 by heterotrophic bacteria. Annu. 33. Ueda, T., E. T. Lode, and M. J. Coon. 1972. Enzymatic w-oxidation. VI. Isolation of homogeneous reduced Rev. Microbiol. 23:515-538. diphosphopyridine nucleotide-rubredoxin reductase. 21. Martin, R. G., and B. N. Ames. 1961. A method for J. Biol. Chem. 247:2109-2116. determining the sedimentation behavior of enzymes: application to protein mixtures. J. Biol. Chem. 34. Wood, H. G. 1952. A study of carbon dioxide fixation by mass determination of the types of C'3-acetate. J. Biol. 236:1372-1379. Chem. 194:905-931. 22. O'Brien, W. E., and L. G. Ljungdahl. 1972. Fermentation of fructose and synthesis of acetate from carbon 35. Yoch, D. C. 1973. Purification and characterization of ferredoxin-nicotinamide adenine nucleotide phosphate dioxide by Clostridium formicoaceticum. J. Bacteriol. reductase from a nitrogen-fixing bacterium. J. 109:626-632. Bacteriol. 116:384-391. 23. Peel, J. L. 1963. The catalysis of the autooxidation of 2-mercaptoethanol and other thiols by vitamin B12- 36. Zanetti, G., and G. Forti. 1966. Studies on the triphosphopyridine nucleotide-cytochrome f-reductase of derivatives. Polarographic and other investigations. chloroplasts. J. Biol. Chem. 241:279-285. Biochem. J. 88:296-308. 24. Pine, M. J. 1958. Methane fermentation of formate by 12. Holcenberg, J. S., and E. R. Stadtman. 1969. Nicotinic