Properties of Purified Carbon Monoxide Dehydrogenase from Clostridium thermoaceticum, a Nickel, Iron-Sulfur Protein*

Carbon monoxide dehydrogenase from Clostridium

' G. Zeikus, personal communication. versatile than what previously has been believed.
The pathway of acetate synthesis from C02 was recently reviewed (7). Evidence indicates that the methyl group of acetate is formed from CO, by reduction via formate and onecarbon derivatives of H4folate,' whereas the carboxyl is derived via a "transcarboxylation" reaction from the C-l of pyruvate (8). In this pathway, CO may function as an electron donor and a source of COS or it may be directly involved as a precursor of acetate carbons.
The enzyme responsible for the initial step of CO metabolism in acetogenic bacteria apparently is CO dehydrogenase, which catalyzes the reaction, CO + H20 -+ CO, + 2H' + 2e-.
High activities of the enzyme have been found in C. thermoaceticum (9), Clostridium formicoaceticum (9), C. thermoautotrophicum (lo), and Acetobacterium woodii (10). Diekert and Thauer (9) have proposed that CO dehydrogenase is involved in the acetate biosynthetic pathway of the acetogenic bacteria and they have demonstrated that CO supplies electrons for the reduction of CO? to acetate in C. thermoaceticum and C. formicoaceticum. The CO may function by either reducing an electron carrier involved in the reduction of CO, to acetate or it may directly reduce a prosthetic group of an enzyme involved in acetate synthesis (9).
A more direct role for CO in acetate synthesis is implied by the demonstration by Drake et al. (11) and Hu et al. (12) that a combination of protein fractions from C. thermoaceticum catalyzes the synthesis of acetate from CO and methyl-H,folate and that one of these protein fractions (F;,) catalyzes an isotope exchange between [ l-'4C]acetyl-CoA and CO. They postulated that CO is a precursor of a formate oxidation level C, intermediate, which may be identical with the previously postulated intermediate formed from C-1 of pyruvate (8).
Fraction F3 is a composite of several proteins including CO dehydrogenase.
Nickel, when added to the growth medium, stimulates the formation of CO dehydrogenase in Clostridiumpasteurianum (13), C. formicoaceticum, and C, thermoaceticum (14). With the use of ""Ni, it has been found that nickel is incorporated into a partially purified CO dehydrogenase from C. thermoaceticum (15) and that ""Ni and enzyme activity coincide on gel electrophoresis (15). Recently, it was demonstrated that oxidation of CO dehydrogenase results in inactivation of the enzyme with concomitant dissociation of nickel from the protein (16). The finding that CO dehydrogenase in C. thermoaceticum is a nickel enzyme places it apart from CO dehydrogenases of the aerobic carboxydobacteria. The enzyme has been purified to homogeneity from Pseudomonas carboxydouorans (1 7, 18) and Pseudomonas carboxydohydrogena (19). The enzyme from the two aerobes contains FAD and the P. carboxydouorans enzyme (18)

Enzyme Assays
Carbon monoxide dehydrogenase, hydrogenase, and formate dehydrogenase activities were all determined a t 50 "C, with 10 mM MV as electron acceptor (IO) using c = 13.9 X IO' mol cm" a t 604 nm (26). In all cases, one unit of enzyme activity is defined as 2 pmol MV reduced min and specific activity as units mg". Protein assays were performed according to Elliott and Brewer (27) using ovalbumin as standard.

Reaction Assays Involving Formate and CO Dehydrogenase
The ability of CO dehydrogenase to either utilize or produce formate was checked in assays performed under a NP, argon or 100% CO atmosphere using anaerobically prepared reagent solutions.
Formation of formate from CO was tested by coupling CO dehydrogenase (130 pg, 88 units) with IO-formyltetrahydrofolate synthetase (0.9 pg, 0.4 units) using the regular assay for the synthetase (22, 291, but with the modification that the gas phase was 100% CO in the place of formate. Oxidation of formate by CO dehydrogenase was tested using MV as electron acceptor in the standard assay method of the enzyme. Sodium formate (IO mM) was substituted for CO in the assay.
Formation of CO and/or Coy from [I4C]formate was checked by following the disappearance of radioactive formate. The assays were performed in I-ml anaerobic vials containing 0.4 ml of 0.1 M Tris/ maleate, p H 8.6, s~dium['~C] formate, 0.07 pmol (261,000 cpm) and one of the following: reduced MV (9 pmol in 0.3 ml of 3.2 mM dithionite), oxidized MV (3 pmol in 0.3 ml of 5 mM dithiothreitol), or dithiothreitol (0.3 ml, 5 mM). CO dehydrogenease (26 pg, 18 units) was added and the reaction was allowed to proceed for 10 min a t 50 "C. The reaction mix was then bubbled with CO for 10 min and a 0.05-ml aliquot was counted. The reaction was acidified with 0.5 ml of 1 M formic acid and bubbled 10 min with CO; another 0.05-ml aliquot was then counted.
Exchange between I4CO a n d C-1 of Acetyl-CoA This reaction was performed essentially as described by Hu et al. (12) except that I4CO was used instead of [l-14C]acetyl-CoA. The assay mix (0.495 ml) in a I-ml anaerobic vial contained acetyl-coA, 0.132 pmol; ATP, 2 pmol; Fe(NH,),(SOt)y, 0.6 pmol; dithiothreitol, 9.2 pmol; and KP,, pH 6, 50 pmol. The solution was bubbled with CO for 5 min, then 200 p1 of CO was replaced with 200 p1 of '"CO (1.4 pmol, 4.7 X IO" cpm). CO dehydrogenase (115 pg, 46 units) was added and the mix was incubated for 3 min at 55 "C. The reaction was initiated with the addition of acetyl-coA and allowed to proceed for 15 min at 55 "C. It was stopped with the addition of 0.04 ml of 4 M acetic acid which dropped the pH to 3.5. The solution was then made alkaline with the addition of 0.2 ml of 2 N NaOH to hydrolyze the acetyl-coA over a period of 6 h at room temperature. Carrier acetic acid (0.32 nmol) was added and the acetate was then isolated using chromatography on Celite (28) before being assayed for I4C.

Gel Electrophoresis and Other Methods
Standard Tris/glycine alkaline electrophoresis was carried out by the method of Davis (30) except that thioglycolic acid (2 mM) and dithiothreitol (2 mM) were used in the buffer system for the prerun and protein run, respectively. SDS-gel electrophoresis was carried out using a standard procedure (31). Tube gels (0.4 X 15 cm) were loaded with 20 pg of protein in all the electrophoreses. The gels were scanned at 595 nm after staining for protein using the Coomassie blue staining technique (32). Activity staining was done with MV as described earlier (16) except that 50 mM Tris/HCI, pH 7.6, was used and in addition, after the MV band appeared, 1 mg.ml-triphenyl tetrazolium chloride was added to the gel staining solution to form a permanent activity band. Molecular weights were determined using gel filtration (33) and pore limit gel electrophoresis (34). The latter was run at pH 9.1 according to Barrett et al. (35). For both pore limit gel electrophoresis and gel filtration, standard proteins were thyroglobulin (M, = 669,000), ferritin ( M , = 440,000), catalase (M, = 232,000), aldolase (M, = 158,000), and ovalbumin (M, = 45,000).
Metals were quantitated using plasma emission spectroscopy (36). Iron was also quantitated using the total iron method of Doeg and Ziegler (37). Acid-labile sulfide was assayed using the procedure outlined by Rabinowitz (38).

Enzyme Purification
Purification of CO Dehydrogenase from C. thermoaceticum-All steps except gel filtrations were carried out inside a Coy anaerobic chamber under an atmosphere of N2:H2(95:5). Gel filtration columns were anaerobically run outside the chamber. They were connected with buffer reservoirs and fraction collectors inside the chamber using stainless steel tubing. Centrifugations were performed under Ny in capped stainless steel centrifuge tubes using a Type 35 rotor and a Beckman L2-65B centrifuge. The basic buffer (Buffer A) used in the purification contained Tris/HCI, 50 mM, pH 7 5 , and sodium dithionite, 2 mM. MV, 0.2 mM, was included in this buffer when it was used in gel fdtration columns. The purification was done a t 10 "C, but it can be done also a t 25 "C without loss of activity. A summary of a purification is given in Table I.
Step 1. Crude Extract; and Step 2. Heat Treatment-These steps, involving a treatment of a cell suspension in a French pressure cell, centrifugation, and subsequent heat treatment at 67 "C for 30 rnin of the supernatant solution, were identical with the steps used for the purification of formate dehydrogenase (39).
Step 3. Ammonium Sulfate Fractionation-Ammonium sulfate was added to 45% saturation to the supernatant from the heat treatment step. The precipitate containing formate dehydrogenase was removed by centrifugation for 30 min a t 31,000 X g. Additional ammonium sulfate was added to the supernatant solution to obtain 60% saturation. The precipitate recovered by centrifugation was dissolved in Buffer A.
Step 4. DE32-cellulose-The enzyme solution of Step 3 was desalted using an Amicon PM-10 ultrafiltration membrane and then applied to a DE32-cellulose column (5.5 X 13 cm) that had been preequilibrated with Buffer A. After washing with several column volumes of Buffer A, the enzyme was eluted using a Niter linear gradient from 0.15 to 0.40 M Tris/HCl. The enzyme began eluting at 0.29 M Tris in a total volume of 340 ml.
Step. 5. Bio-Gel HTP-Fractions with a specific activity greater than 145 were combined and applied to a hydroxylapatite column (200 mi) which had been previously equilibrated with Buffer A. After washing with Buffer A, a linear gradient consisting of 750 ml of Buffer A and 750 ml of 200 mM potassium phosphate, 2 mM sodium dithionite eluted the enzyme which appeared in fractions containing about 50 mM phosphate.
Step 6. Bio-GelA-l.5m/Vltrogel AcA 22, Gel Filtration-Fractions with a specific activity greater than 420 were combined and concentrated from 175 to 5 ml using an Amicon PM-30 ultrafitration membrane. The concentrated solution was applied to a tandem Bio-Gel A-1.5m/Ultrogel AcA 22 (both 2.5 X 85 cm) column setup. The gel filtration was carried out at a constant flow rate of 18 ml h" using Buffer A. Fractions with a specific activity greater than 550 were combined. These fractions had a relatively constant specific activity across the peak.
Step 7. DEAE-Sephacel-The combined fractions from Step 6 were applied to a DEAE-Sephacel column (1.5 X 15 cm). The enzyme obtained from this column, which was run similarly to the DE32cellulose column of Step 4, was of apparent homogeneity.
Step 7 was not always included since pure protein was often obtained after the gel Titration in Step 6.
G-25 Sephadex Column to Remove Dithionite and/or MV from Pure Enzyme-The purified enzyme from Step 6 or Step 7 was applied to a Sephadex G-25 column equilibrated with oxygen-free 50 mM Tris/HCI, pH 7.6. The enzyme was eluted with the same buffer. The enzyme was active under these conditions for at least 2 days.

RESULTS
Purification of CO Dehydrogenase-A representative purification is summarized in Table I. Occasionally, Step 7 involving a DEAE-Sephacel column was necessary. The enzyme was also successfully chromatographed using a Phenyl-Sepharose column. With some preparations, the high specific activity of 675 was never obtained. Since these preparations contained a lower content of zinc, we postulate that zinc might have dissociated from the enzyme. The heat treatment step which has been designed for the purification of formate dehydrogenase (39) involves heating the enzyme for 30 min at 67 "C. While apparent loss of CO dehydrogenase occurs during this heating step, this loss was regained during ammonium sulfate fractionation. Heating the extract for only a few min at 70 "C instead of for 30 min "activates" the CO dehydrogenase and, if the aim is to purify only CO dehydrogenase, the heating step perhaps should be modified.
The amount of CO dehydrogenase is high in C. thermoaceticum. Assuming a specific activity of 675 pmol min" mg", and 12% recovery, approximately 2% of the soluble cell protein exists as carbon monoxide dehydrogenase.
Stability of CO Dehydrogenase-The purified enzyme can be stored in several ways provided it is in an oxygen-free atmosphere. The enzyme can be frozen in buffer A or mixed with an equal volume of glycerol containing 2 mM dithionite and stored at -20 or 5 "C. In no case was the enzyme stable for more than one month.
Additions of the enzyme to an aerobic buffer (50 mM Tris/HCl, pH 7.6) resulted in a 98% loss of activity within 15 min, whereas it is stable for a t least 4 h when diluted with Buffer A to a concentration of 7 pg m1-l. The presence of CO had no apparent effect on the stability of the enzyme.
Gel Electrophoresis a n d Molecular Weight-Gel electrophoresis using anaerobic or standard aerobic techniques demonstrated that the enzyme with a specific activity of 675 pmol min" mg" is at least 99% pure (Fig. 1A ). Electrophoresis in the presence of SDS (Fig. 1B) gave two protein bands of equal intensity, indicating that the enzyme consists of two different subunits of equal concentration. The M, of the subunits as obtained by SDS-gel electrophoresis are 77,600 and 70,900.
Since an M, of 436,000 is obtained on gel filtration, the native enzyme apparently has an ( a p )~ structure. Pore limit gels yielded an M, of 161,000. This result indicates that the enzyme can dissociate and apparently form an ap structure with an average M, of about 155,000. Sedimentation equilibrium centrifugation also revealed that the CO dehydrogenase has a tendency to dissociate. The dissociation phenomenon has not been investigated however, it may be pH dependent. The electrophoresis runs were made at a higher pH (9.1) than the gel filtration (7.6).
Metals and Acid-labile Sulfide Content-Plasma emission spectroscopy analyses of four different preparations revealed that homogeneous CO dehydrogenase contains per mol of the dimeric enzyme (Mr = 155,000) 1.7 0.2 nickel, 10.8 5 1.8  Enzyme, 20 yg, was run in gels (4 X 150 mm). The spikes at 13.5 cm (A) and 14 cm (B) represent positions of the tracking dye bromphenol blue. A, electrophoresis in Tris/glycine buffer, pH 8.0 (30), containing 2 mM dithiothreitol under anaerobic conditions. The gel was prerun with buffer containing 2 mM thioglycolate for 2 h. The buffer was deoxygenated before addition of reducing agents and the electrophoresis chamber was in an atmosphere of argon. Activity stain (not shown) corresponded with the protein band. B, SDS-gel electrophoresis. For molecular weight determination (not shown) the following standards were used: Phosphorylase b,92,500, bovine serum albumin, 66,200; ovalbumin, 45,000, carbonic anhydrase, 31,000; soybean trypsin inhibitor, 21,500; and lysozyme, 17,440. iron, and 1.3 f 1.1 zinc. Wide variation in the concentrations of sodium and calcium has been found with no correlation to the final specific activity of the enzyme. The highest specific activity enzyme has been found to contain as much as 2-3 mol of zinc/mol of dimeric enzyme. In addition, it has 14 acidlabile sulfurs. Plasma emission spectroscopy also quantitates the following elements: Al, B, Cd, Co, Cr, Cu, K, Mg, Mn, P, Pb, Si, Mo, Sr, and Ba. None of these was found in an amount over 0.1 mol/mol of enzyme. The presence of selenium was also ruled out since no radioactive selenium was incorporated into the enzyme when cells were grown on ["Selselenite. When the enzyme, purified from cells which had been cultivated on media supplemented with "NiC12 (16), was exposed to 0.1 N HC1O4 and the precipitated protein removed by centrifugation, virtually all of the nickel was recovered in the supernatant fluid. Subsequent passage of the dissociated "Nicontaining extract through a Sephadex G-25 (superfine) column equilibrated with 1 mM HCl demonstrated that the nickel eluted quantitatively at a molecular weight slightly larger than that of cyanocobalamin (M, = 1357), and was well resolved from the eluting position of a "NiC12 standard. We have concluded that CO dehydrogenase contains a nickel factor which can be dissociated from the enzyme and which apparently is stable under mild acid conditions. Absorption Spectrum-Spectra recorded of active enzyme preparations freed of dithionite and MV are shown in Fig. 2.
The enzyme under anaerobic conditions in 0.05 M Tris/HCl, pH 7.6, has a peak at 277 nm and a shoulder at 390 nm. In the presence of COz the 390 shoulder is increased and the 277 nm peak is decreased and slightly blue-shifted. CO caused a decrease in the 390 shoulder (Fig. 2, inset). CO also decreases the 277 nm absorbance but does not shift the position of the peak. In the presence of oxygen, the enzyme's spectrum is identical to the COz-treated enzyme. Exposure of the native  Expressed as pmol of CO oxidized min" mg" at 50 PM concentration of electron acceptor. enzyme to hydrogen, nitrogen, or cyanide did not affect its spectrum.
Catalytic Properties-MV, benzylviologen, methylene blue, ferredoxins I and I1 from C. therrnoaceticurn (40, 41), and rubredoxin, flavodoxin, and ferredoxin from C. formicoaceticum accept electrons from carbon monoxide at rates shown in Table 11. Rubredoxin is by far the best electron acceptor of the artificial and proteinaceous electron carriers, that have been tested. These results will be more detailed in a separate communication." With hydrogen or formate replacing CO as electron donor to MV, there was absolutely no reaction. This indicates that the purified CO dehydrogenase was free of hydrogenase (42) and formate dehydrogenase activity (39). The latter precipitates with a 45% saturation of ammonium sulfate, whereas the CO dehydrogenase activity is then found in the supernatant.
Activity of CO dehydrogenase was tested between pH 5.2 and 8.6 in 0.1 M Tris/maleate containing 3.2 mM dithiothreitol. Within this range, the enzyme was active and had a pH optimum of 8.2 to 8.4. At pH 8. 4 and 50 "C under 100% CO atmosphere, the apparent K , for MV is 3.03 k .01 mM and the apparent V,,, is 750 pmol of CO oxidized min" mg". Pyruvate, 3 mM; acetyl-coA, 78 pM; or ATP, 1 mM, tested as possible inhibitors or activators of the enzyme, had no effect on the apparent K , or V,,, of the reaction with MV.
Cyanide (40 p~) inhibited the enzyme by 60% within 10 s; however, carbon monoxide reversed the inhibition. These findings are consistent with those of Drake et al. (15). Methyl iodide (2.5 mM) was found to irreversibly inhibit the enzyme in a fiist order reaction. Light did not prevent or relieve the inhibition and carbon monoxide seemed to stimulate the inhibition. The data closely fit an exponential decay curve with a half-time of inactivation of 29, 26, 22, and 10 min when the enzyme was incubated with methyl iodide in the presence of nitrogen, nitrogen plus light, carbon monoxide, or carbon monoxide plus light, respectively.
A series of experiments were performed to examine the proposal (11, 12) that carbon monoxide dehydrogenase may be involved in the production of a formate intermediate. ["C] Formate was tested as a substrate for CO dehydrogenase in the presence of either reduced or oxidized MV or in the absence of an electron carrier. Formate was not converted to CO or C02 and appeared inert since all of the I4C remained in the formate. As mentioned earlier, formate does not replace CO in the standard assay. To further test for a formate intermediate, an assay of CO dehydrogenase using a 100% CO gas phase was performed in the presence of 10-formyl-H4folate synthetase to trap any formate. No formate was detected. A control with formate revealed that CO did not affect the activity of the synthetase. Hu et al. (12) proposed that an enzyme-bound formate  [l-14C]acetyl-CoA in a reaction catalyzed by protein fraction F.r (11) from C. thermoaceticum. To test whether or not this exchange is catalyzed by purified CO dehydrogenase, the enzyme was incubated with 14C0 and acetyl-coA. The latter was hydrolyzed and the acetate was isolated using Celite chromatography. No label was detected in the acetate. However, a crude extract and an extract freed of cofactors from C. thermoaceticum under the same conditions catalyzed the exchange between ',CO and acetyl-coA in accordance with the result obtained using protein fraction F3 (12).

DISCUSSION
Previously, aerobic CO dehydrogenases have been purified to homogeneity from P . carboxydohydrogena (19) and P. carboxydovorans (17,18). The enzyme from P. carboxydohydrogena consists of three different subunits (M, = 14,000; 28,000; and 85,000) and has the structure (a,P,y)a. It contains at least 3 mol of FAD. The P. carboxydouorans enzyme has an M , = 230,000 and consists of two assumed identical subunits (17). It apparently contains (per mol) 2 FAD, 2 molybdenum, 8 iron, 8 inorganic sulfur, 3 zinc, and 2 copper (18). In contrast, the anaerobic CO dehydrogenase from C. thermoaceticum, the purification of which is reported in this paper, consists of two nonidentical subunits with M, = 77,600 and 70,900 having the structure (a,/?) or, as it is obtained during isolation, a (a,,& form. Per mol, the dimeric enzyme apparently contains 2 nickel, 1 to 3 zinc, 11 iron, and 14 acid-labile sulfur but no molybdenum or FAD. The main difference in metal content between the aerobic CO dehydrogenases and the anaerobic CO dehydrogenases is clearly that of molybdenum in the former (18) and nickel in the latter (Refs. 13,15,16 and this paper). Similarities involve the presence of zinc and the possible presence of iron-sulfur centers. It can be postulated by comparison with other molybdenum-and nickel-containing enzymes that these two metals are involved in the catalytic processes as are the iron-sulfur centers. However, functions of zinc in both enzyme types and of copper in the aerobic type are still completely unknown.
Similarly, the roles of sodium and calcium in the CO dehydrogenase from C. thermoaceticum are uncertain. Many enzymes depend on or are stimulated by monovalent cations. These enzymes include formyl-H4folate synthetase from C. thermoaceticum (29). Calcium ions have been found to stabilize some enzymes such as thermolysin (43) and a-amylase (44) from thermophiles.
The CO dehydrogenases from P. carboxydohydrogena (19) and P. carboxydouorans (18) are oxygen stable and can be isolated without the strict anaerobic procedure described in this paper. Electron acceptors which are slightly positive are the best acceptors for the Pseudomonas enzymes as well as for the C. thermoaceticum enzyme. However, the former enzymes do not use electron acceptors with potentials much below -34 mV (17,19), whereas the C. thermoaceticum enzyme reduces viologens and ferredoxins with potentials as low as -440 mV (Table IT and Refs. 9 and 15).
The nature of the nickel in the clostridial CO dehydrogenase is not known. Nickel is a component of the active site of jack bean urease (45-47) and is extractable from the enzyme at low pH by EDTA (46). Hydrogenases of Alcaligenes eutrophus (48), Desulfovibrio gigas (49), and Methanobacterium thermoautotrophicum (50, 51) contain nickel as does factor F4:,,,, a porphinoid of methanogens (52-55). The structure of FA.$,, has recently been elucidated by Pfaltz et al. (56), and Ellefson and Wolfe (57) have proposed that F4:10 may he the prosthetic group of methyl-Coenzyme M reductase. While we have found that nickel dissociates from CO dehydrogenase as a small molecular weight factor, our spectral studies of the native enzyme suggest that it does not contain Fw,. The molecular weight of the factor is somewhat larger than for vitamin B1,. Diekert and Thauer (9) actually suggested that CO dehydrogenase in C. thermoaceticum should be a corrinoid enzyme. They based this suggestion on the observation that alkyl halides in the presence of CO inactivated the enzyme and that this inactivation was reversible by photolysis. As reported in this paper, the purified enzyme is inhibited by methyl iodide in a reaction stimulated by CO. However, the inhibition was not reversed by light. In contrast, Drake et al. (15) found that with a partially purified enzyme alkylating reagents have a minimal effect on the enzyme activity. An explanation of these apparent discrepancies must perhaps await the elucidation of the structure of the Ni cofactor.
Recent EPR studies of the hydrogenases from M . thermoautotrophicum (51) and D. gigas (49) have revealed that the nickel in these enzymes is redox-sensitive and it was proposed that nickel is the binding site of the substrate, Hz. Similarly, in EPR studies, we have recently found that nickel in the CO dehydrogenase from C. thermoaceticum is redox-active (58). In addition, we obtained evidence for iron-sulfur clusters in the enzyme.
Since CO dehydrogenase from C. thermoaceticum and from the carboxydobacteria (17-19) is a major component of the soluble cell protein, it probably plays a major role in the metabolism of these bacteria. The carboxydobacteria can utilize carbon monoxide as a sole carbon and energy source (2-4) and can serve as a sink for CO in the atmosphere (2). Recently, it has been found that Pseudomonas carboxydoflava can use energy derived from the oxidation of CO for the assimilation of organic material when grown under heterotrophic conditions (59). Under these conditions, CO was not utilized as a carbon source.
A number of acetogenic bacteria grow on CO as energy source and CO is incorporated into acetate (5, 6).I Hu et al. (12) have shown that acetate can be synthesized from CO and 5-methyl-H4folate, and they have proposed that CO dehydrogenase supplies an enzyme-bound formyl-level CI intermediate which can serve as the direct source of the carboxyl group of acetate. Furthermore, they postulated that the CI intermediate could couple with the H,folate pathway to yield 5methyl-H4folate and serve also as a potential source of the methyl group of acetate (11,12). To test this possibility, we have coupled CO dehydrogenase with H4folate enzymes purified from C. thermoaceticum and C. formicoaceticum (22)(23)(24). With CO dehydrogenase coupled with 10-formyl-H4folate synthetase and I4CO as substrate, there was no formation of IO-formyl-H,folate. Similarly, with CO dehydrogenase COUpled with 10-formyl-H4folate synthetase, 5,lO-methenyl-H4folate cyclohydrolase, and 5,lO-methylene-H4folate dehydrogenase and I4CO as substrate, there was no formation of 5,10-methylene-H4folate. In control experiments using ["c] formate, the expected H4folate compounds were readily produced. While these findings do not exclude the possibility of an enzyme-bound, formyl-level intermediate for CO dehydrogenase, the purified enzyme apparently cannot couple directly with enzymes of the H4folate pathway. This suggests that, for CO to be incorporated into the methyl group of acetate, CO