Partial Reactions Catalyzed by Protein Components of the Acetyl-CoA Decarbonylase Synthase Enzyme Complex from Methanosarcina barkeri *

In methanogens, the acetyl-CoA decarbonylase syn- thase (ACDS) complex, which has five different subunits, catalyzes synthesis and cleavage of acetyl-CoA according to the reaction: CO 2 (cid:49) 2H (cid:49) (cid:49) 2e (cid:50) (cid:49) CH 3 -H 4 SPt (cid:49) CoA ª acetyl-CoA (cid:49) H 4 SPt (cid:49) H 2 O, where H 4 SPt and CH 3 -H 4 SPt are tetrahydrosarcinapterin and N 5 -methyl- tetrahydrosarcinapterin, respectively. We have dissociated the ACDS complex into three protein components by limited proteolytic digestion. Catalysis of acetyl-CoA synthesis was lost in parallel with the loss of the intact (cid:98) subunit; however, no decrease in activity was detected in any of three partial reactions found to be catalyzed by distinct protein components of the proteolyzed ACDS complex: ( a ) CO dehydrogenase, catalyzed by the (cid:97)(cid:101) component, ( b ) CH 3 -H 4 pteridine:cob(I)amide-protein methyltransferase, catalyzed by the intact (cid:103) subunit and fragments of the (cid:100) subunit, and ( c ) acetyltransferase, catalyzed by a truncated form of the (cid:98) subunit. The results indicated that the (cid:98) subunit is responsible for binding CoA separate methyltransferase enzyme, analogous to that required for methylation of the clostridial 88-kDa corrinoid/iron-sulfur N 5 (13, in methyl group transfer from CH 3 -H 4 SPt. The purpose of the present investigation was to investigate fundamental structural and functional properties of subunit/ subprotein interactions in studies on the quaternary structure of the ACDS enzyme complex. The results include development of a method for quantitative dissociation of the enzyme complex along with the isolation and characterization of a truncated form of one of the subunits and two other protein components with characteristic subunit compositions. A new, nonradioac-tive method was developed for analysis of acetyltransferase activity. Identification was made of the protein subcomponents involved in catalysis of two additional partial reactions, that of Co- (cid:98) -methylcob(III)amide:tetrahydropteridine methyltransferase (reaction 3), and acetyltransferase (reaction 4). allowed to proceed to more than approximately 30% completion. The rate of acetyl-CoA hydrolysis was found to be negligible in control reactions carried out in the absence of 3 (cid:57) -dephospho-CoA. Formation of S -acetyl-3 (cid:57) -dephospho-CoA was not detected in reaction mixtures that lacked Ti 3 (cid:49) -EDTA. complex. Analysis of the reaction of acetyl group transfer from acetyl-CoA to 3 (cid:57) -dephospho-CoA catalyzed by the intact ACDS complex was carried out by the method described for assay of acetyltransferase activity in resolved fractions of the proteolyzed enzyme, with the following modifications. Preincubation of the enzyme complex (57 (cid:109) g) was carried out at room temperature ( (cid:59) 27 °C) for 1.5 min under strictly anaerobic conditions in a mixture (50 (cid:109) l total volume) containing approximately 0.8 nmol of ferredoxin, 0.2 (cid:109) mol Ti 3 (cid:49) -EDTA and 50 m M MOPS buffer, pH 7.2. An aliquot (15 (cid:109) l) of the incubation mixture was then transferred to a solution (550 (cid:109) l), which contained, except for dephospho-CoA, all other components of the reaction (90 nmol of acetyl-CoA, 3.2 (cid:109) mol of Ti 3 (cid:49) -EDTA, 0.8 nmol of ferredoxin, and 30 (cid:109) mol of MOPS-Na, pH 7.2). This mixture was incubated for an additional 1.5 min, and the reaction (final volume 600 (cid:109) l) was initiated by the addition of 3 (cid:57) -dephospho-CoA (35 (cid:109) l of a 2.58 m M stock solution). The reaction was allowed to proceed at 27 °C. Samples (60 (cid:109) l) were removed periodically, mixed with an equal volume of 20 m M sodium 2-mercaptoethanesulfonate, 0.5 M sodium citrate, pH 4.0, to stop the reaction, and frozen for subsequent HPLC analysis. Based on the initial rate of the reaction, the specific activity of acetyltransferase was found to be 1.2 (cid:109) mol of acetyl transferred per min/mg of enzyme complex.

where Fd stands for ferredoxin, and CH 3 -H 4 SPt and H 4 SPt denote N 5 -methyl-tetrahydrosarcinapterin and tetrahydrosarcinapterin, respectively. The sarcinapterin compounds are used in M. barkeri in place of the corresponding tetrahydrofolate derivatives. The structures of CH 3 -H 4 SPt and H 4 SPt are shown in Fig. 1.
The overall reaction of acetyl-CoA synthesis or cleavage may be divided into several possible partial reactions. One of these is CO:acceptor oxidoreductase (CO dehydrogenase). This reaction may be written as follows, REACTION 2. and is carried out by an ␣ 2 ⑀ 2 component, containing nickel and iron-sulfur centers, as shown in studies (6) that preceded those on the multienzyme complex. This partial reaction is catalyzed also by the intact ACDS complex from the genus Methanosarcina (1,3), from Methanothrix strain CALS 1, 2 and from Methanococcus vannielii, 4 as well as by the ␣ 2 ⑀ 2 component isolated from the multienzyme complexes from M. barkeri (2,6), M. vannielii (7), and Methanosarcina thermophila (8). The ␣ 2 ⑀ 2 protein in Methanothrix sohengenii has been reported to exhibit an oxygen-stable CO dehydrogenase activity and has also been the subject of several detailed studies (9 -12). Interestingly, evidence for a high molecular mass enzyme complex in this organism has not yet been presented.
A discrete, 102 kDa, ␥ 1 ␦ 1 corrinoid/iron-sulfur protein subcomponent of the ACDS complex from M. thermophila has been identified in experiments in which the complex was dissociated by treatment with a cationic detergent (8). In these studies, a portion of the enzyme complex remained undissociated, and recovery of the ␤ subunit was not reported. It was shown that the reduced corrinoid protein became methylated in the presence of methyl iodide. However, methylation by CH 3 -H 4 SPt was not described, and it was unknown whether or not a separate methyltransferase enzyme, analogous to that required for methylation of the clostridial 88-kDa corrinoid/ironsulfur protein by N 5 -methyl-tetrahydrofolate (13,14) was involved in methyl group transfer from CH 3 -H 4 SPt.
The purpose of the present investigation was to investigate fundamental structural and functional properties of subunit/ subprotein interactions in studies on the quaternary structure of the ACDS enzyme complex. The results include development of a method for quantitative dissociation of the enzyme complex along with the isolation and characterization of a truncated form of one of the subunits and two other protein components with characteristic subunit compositions. A new, nonradioactive method was developed for analysis of acetyltransferase activity. Identification was made of the protein subcomponents involved in catalysis of two additional partial reactions, that of Co-␤-methylcob(III)amide:tetrahydropteridine methyltransferase (reaction 3), and acetyltransferase (reaction 4). A strong correlation was shown between the ability of the ACDS complex to catalyze reaction 1 and the level of intact 60-kDa ␤ subunit that remains during proteolytic digestion. Finally, experiments are described in which determination is made of the KЈ eq of reaction 3. The significance of the value of the KЈ eq to methyl group transfer in vivo is discussed.
Subunit-specific Polyclonal Antibodies-Antibodies against each of the five ACDS subunits were raised separately in New Zealand White rabbits. Each of the subunit protein antigens was isolated by extraction from preparative SDS slab gels. Antisera were obtained through Hazleton Research Products Inc. Each antibody preparation was characterized by Western blot analysis of SDS gels containing the entire enzyme complex. The results showed that antibodies raised against four of the subunits (␣, ␤, ␥, and ␦) each detected their respective subunits specifically and did not cross-react with any of the other subunit bands.
Cross-reactivity observed in the antibodies against the ⑀ subunit was markedly reduced by affinity purification. Preimmune sera drawn from the animals did not react with any of the five ACDS subunits.
Acetyl-CoA Decarbonylase Synthase Complex-The ACDS complex was isolated from acetate-grown cells of M. barkeri by anaerobic gel filtration on Sepharose CL-6B, as described previously (3). Preparations of the enzyme complex were stored in liquid N 2 . Protein was assayed by the method of Bradford (16), with the dye reagent supplied by Bio-Rad and with bovine ␥-globulin as standard.
Proteolysis and Dissociation of the ACDS Complex into Subcomponent Proteins-The enzyme complex, 5.3 mg, was treated under anaerobic conditions with 0.2 mg of chymotrypsin for 60 min at 24°C in a reaction mixture (1.0 ml) that contained 50 mM Tris⅐HCl buffer, pH 7.5. This procedure resulted in complete loss of the ability to catalyze acetyl-CoA synthesis. However, no decrease was observed in activities of CO dehydrogenase, Co-␤-methylcob(III)amide:tetrahydropteridine methyltransferase, or acetyltransferase. After the incubation with chymotrypsin, the mixture was applied to an anion-exchange column (Bio-Rad Econo-Pac Q cartridge). Elution was carried out at 1 ml/min with a linear gradient (80 ml) of 0 -1 M NaCl in 50 mM Tris⅐HCl, pH 7.5. Fractions were collected at 1-min intervals. The time chosen for proteolysis (60 min) was not critical because samples incubated for either 20 or 60 min produced ion-exchange elution profiles that were indistinguishable. Therefore, no special treatment was found necessary to terminate the reactions prior to chromatography.
Assay of Acetyltransferase Activity in Resolved Fractions of the Proteolyzed ACDS Complex-Development of a new assay for acetyltransferase was based on the analogous reaction of 3Ј-[ 32 P]CoA/acetyl-CoA exchange, first described for carbon monoxide dehydrogenase from C. thermoautotrophicum (17,18). The exchange reaction has been shown to be strictly dependent on redox potential and does not proceed to a significant extent in the absence of a strong reductant (18). The method developed herein also uses an HPLC separation step; however, radioactively labeled coenzyme A is not required.
Prior to assay for acetyltransferase activity, samples from the fractions were reduced by incubation at one-tenth of their original concentration for 10 min in the presence of 2 mM Ti 3ϩ -EDTA in 50 mM MOPS buffer, pH 7.2, in a total volume of 100 l. Aliquots (5-10 l) of the incubation mixtures were then added to a solution (97-92 l) containing 18 nmol of acetyl-CoA, 0.8 nmol of ferredoxin, 460 nmol of Ti 3ϩ -EDTA, and 6 mol of MOPS buffer, pH 7.2. The acetyl transfer reaction was initiated by the addition of 18 l of 1 mM 3Ј-dephosphocoenzyme A. The reaction mixtures (final volume of 120 l) were maintained at 23°C for 20 min. Thereafter, the reactions were stopped by addition of 120 l of a solution containing 20 mM sodium 2-mercaptoethanesulfonate and 0.5 M sodium citrate, pH 4.0. The final mixtures were stored frozen in liquid nitrogen prior to analysis. Quantitative determination of the products coenzyme A and S-acetyl-3Ј-dephospho-CoA was carried out on samples, 100 l, analyzed by reversed phase HPLC under anaerobic conditions. HPLC analysis was carried out as described previously (3,5) with modification of the solvent gradient to allow separation of the four derivatives: coenzyme A, 3Ј-dephospho-CoA, acetyl-CoA, and S-acetyl-3Ј-dephospho-CoA. In the modified procedure, the column (APEX octadecyl, 250 ϫ 4.6 mm diameter, from Jones Chromatography, Inc.) was equilibrated in 50 mM tetramethylammonium phosphate, pH 4.7, and a linear gradient of 0 -20% acetonitrile in the same solution was applied at 1 ml/min over a period of 35 min. As shown in Fig. 2, the rate of product formation declines steadily as the reaction proceeds. Therefore, the reaction time and amount of enzyme employed for single time point assays of fractions was chosen so that none of the reactions were

Partial Reactions of the ACDS Enzyme Complex 8353
allowed to proceed to more than approximately 30% completion. The rate of acetyl-CoA hydrolysis was found to be negligible in control reactions carried out in the absence of 3Ј-dephospho-CoA. Formation of S-acetyl-3Ј-dephospho-CoA was not detected in reaction mixtures that lacked Ti 3ϩ -EDTA.

RESULTS
The ability of limited proteolytic digestion to cause dissociation of the M. barkeri ACDS complex was investigated by reaction of the enzyme complex with chymotrypsin followed by anion-exchange chromatography. As shown in Fig. 3, this procedure resulted in resolution of three major peaks of protein.
The three protein components emerged from the column following a peak of unbound material that contained peptides too small to be resolved on a 12% acrylamide SDS gel. Each of the three protein peaks possessed characteristic subunit compositions, as shown by SDS-PAGE analysis (Fig. 3). Unaltered subunit bands as well as modified subunits were identified by Western blot analyses using subunit-specific antibodies (data not shown). As shown in Fig. 3, the first protein peak contained the ␥ subunit and four other polypeptides corresponding to partially degraded forms of the ␦ subunit (␦*). The second protein component contained the ␣ and ⑀ subunits. The final peak contained a truncated form of the ␤ subunit (␤*). Evidence for the presence of iron-sulfur centers in each of the three peaks was found both by direct determinations of iron and by measurement of the UV-visible absorption spectra of the three protein components. No fraction was recovered that contained detectable quantities of the unresolved enzyme complex. Therefore, the overall dissociation of the complex was judged to be efficient and quantitative.
The extent of proteolytic digestion was varied by incubating samples of the enzyme complex at a fixed concentration with different concentrations of chymotrypsin. SDS-PAGE was used to monitor the progress of digestion of each of the subunits. The band corresponding to the ␤ subunit was found to be highly susceptible to proteolytic attack (Fig. 4). As the concentration of chymotrypsin was increased, the ␤ subunit band intensity decreased markedly. Complete loss of the intact ␤ subunit band occurred under conditions in which intense bands were still found for all other subunits. Densitometric analyses showed that loss of the ␤ band occurred with concomitant loss of the ability to carry out overall synthesis of acetyl-CoA, as shown in Fig. 4. Further digestion at the levels greater than those shown in Fig. 4 then resulted in selective loss of the band corresponding to the intact ␦ subunit. Samples that contained residual amounts of the intact ␦ band were found to be only partially resolved by subsequent ion-exchange chromatography. These findings indicated that activity is lost prior to extensive dissociation of the complex.
Although overall acetyl-CoA synthesis activity was abolished by proteolytic treatment of the enzyme complex, no decrease in activity was detected for CO dehydrogenase, CH 3 -H 4 pteridine: cob(I)amide-protein methyltransferase, or acetyltransferase. Assays for total cobamide content were carried out on fractions obtained in resolution of the protein subcomponents, as shown in Fig. 5A. The major peak of cobamide was closely associated with the first protein peak, as shown in Fig. 5A. Peak 1 contained 74% of the total cobamide. The remaining 26% eluted over a broad region of the gradient and did not coincide with any of the three major protein peaks.
Since Co-␤-methylcob(III)amide:tetrahydropteridine transmethylation had not been reported with isolated components of the ACDS complex, tests were performed to determine whether or not the reduced peak 1 corrinoid protein was capable of undergoing direct reaction with CH 3 -H 4 SPt. The Co(I) form of the corrinoid protein was first generated by reaction with 1 mM Ti 3ϩ -EDTA. As described previously for detection of the reduced cobamide in the intact complex (4), reduction to the Co(I) level was indicated by the development of absorption spectra that showed a prominent peak at 394 nm. The reduction reached 50% completion in approximately 35 s, as shown in Fig. 6. In contrast, under otherwise similar conditions, the undigested enzyme complex required tens of minutes to become reduced (5). Addition of CH 3 -H 4 SPt to the reduced corrinoid protein resulted in immediate loss of the 394-nm absorbance peak, and concomitant increase in absorbance around 456 nm. The spectral changes were characteristic of formation of the base-off Co 3ϩ -methyl cobamide, as observed previously in the intact complex (4). This finding indicated that methyl group transfer from CH 3 -H 4 SPt does not require additional subunit proteins (␣, ␤, or ⑀) and that methyltransferase activity is an intrinsic property of the ␥␦ corrinoid protein subcomponent.
CO dehydrogenase was also measured in the fractions derived from anion-exchange chromatography of the digested enzyme complex. As shown in Fig. 5B, CO dehydrogenase was associated exclusively with peak 2. These results constitute the first demonstration that the ␣⑀ carbon monoxide dehydrogenase subcomponent may be released from the enzyme complex by the action of a protease.
Acetyltransferase activity of the intact complex from M. thermophila was previously detected based on the exchange of 3Ј-[ 32 P]CoA with acetyl-CoA (19). However, the subunit location of this activity was not investigated. In order to identify the protein component responsible for acetyltransferase activity, fractions obtained during resolution of the digested enzyme complex were assayed for acetyltransferase (reaction 3), as described under "Materials and Methods." As shown in Fig. 5C, acetyltransferase activity was found to be associated with the protein peak containing the truncated ␤ subunit. The results establish a previously unrecognized activity of the ␤ subunit and indicate that the binding site for CoA and acetyl-CoA reside within a domain that remains intact in the truncated ␤ subunit.
Titration of the reduced corrinoid protein with CH 3 -H 4 SPt was carried out in order to determine the equilibrium constant for methyl transfer in reaction 3. As shown in Fig. 6, each successive addition of CH 3 -H 4 SPt made to the reduced corrinoid protein resulted in an immediate decrease in the absorbance function (A 394 nm -A 456 nm ). The addition of excess H 4 SPt to the methylated corrinoid protein caused an immediate increase in the absorbance function, indicating that the reaction was freely reversible. The ratio of methylated to demethylated  Fig. 7. The KЈ eq value measured at pH 7.5 was 5.5 Ϯ 0.3.

DISCUSSION
It was demonstrated previously that the ACDS complex catalyzes the synthesis and cleavage of acetyl-CoA (reaction 1) (3,5); however, unambiguous assignment of the catalytic roles of the various protein subcomponents of the enzyme complex has not been reported. In order to obtain information on the quaternary structure of the ACDS complex and to identify catalytic properties of individual protein subunits or subcomponents, we developed a new procedure for dissociation of the enzyme complex. Important advantages over the previously described detergent fractionation protocol (8) are (a) that three protein components are resolved instead of two and (b) that dissociation of the enzyme complex is quantitative and does not produce a fraction corresponding to the residual unresolved enzyme complex. The three separate protein components so obtained display distinct subunit compositions and exhibit characteristic catalytic activities.
Component 1, N 5 -Methyl-tetrahydropteridine:cob(I)amide-protein Methyltransferase-The first of the three protein peaks produced by chromatographic resolution of the digested enzyme complex is a corrinoid protein that contains the 50-kDa ␥

FIG. 4. Chymotrypsin digestion of the ACDS complex, correlation of acetyl-CoA synthesis activity with the presence of various subunits detected by SDS-PAGE.
Digestion of the enzyme complex (5 mg/ml) with chymotrypsin was carried out by a procedure similar to that described under "Materials and Methods," with the exception that the reaction time was decreased to 20 min. The weight ratio of chymotrypsin to enzyme complex protein is indicated and was varied over a range of 1:3200 -1:400. Incubations were performed sequentially, and at the end of each incubation time period, an aliquot was removed and immediately assayed for acetyl-CoA synthesis using the substrates CH 3 -H 4 SPt, CoA, and CO 2 (as bicarbonate), in the presence of ferredoxin and Ti 3ϩ -EDTA according to the method described previously (5). At the same time, the remaining digestion mixture was mixed with an equal volume of 2 ϫ concentrated SDS sample buffer (16) and kept for analysis by SDS-PAGE. The rate of acetyl-CoA synthesis observed in the absence of chymotrypsin was taken as 100% activity. Band area (%) corresponds to the densitometrically determined areas of peaks at the positions of the indicated subunits expressed as a percentage of the areas detected for the same subunits in the undigested enzyme complex.
FIG. 5. Analysis of cobamide content, CO dehydrogenase, and acetyltransferase activity of resolved ACDS protein subcomponents. The activities of CO dehydrogenase and acetyltransferase, and the concentration of cobamide was measured in fractions obtained following limited proteolytic digestion and anion-exchange chromatographic resolution of the ACDS subcomponent proteins, as described under "Materials and Methods." Panel A, the fractions were assayed for cobamide concentration by a method that employed differential spectrophotometric detection of the reaction with cyanide, as described previously (3). Panel B, CO dehydrogenase activity was measured by CO-dependent reduction of methylviologen as described previously (2). Panel C, individual fractions were assayed for acetyltransferase activity, as described under "Materials and Methods." Activity is expressed as nmol of S-acetyl-3Ј-dephospho-CoA formed per min/ml of fraction. subunit and fragments of the 48-kDa ␦ subunit. A similar ␥␦ subcomponent was dissociated previously by the detergent treatment procedure (8); however, we have extended the characterization of the corrinoid protein to investigate reactivity of this component with the physiological methyl donor substrate CH 3 -H 4 SPt. Although evidence for methyl group transfer from CH 3 -H 4 SPt to the corrinoid cofactor contained in the intact complex has been presented earlier (3), analogy to the acetyl-CoA synthesizing system of C. thermoaceticum might suggest that a separate methyltransferase (possibly one of the subunits other than ␥ or ␦) would be required for methylation of the corrinoid protein by CH 3 -H 4 SPt. However, we now show that in the absence of other proteins the isolated, reduced ␥␦* cob(I)amide protein reacts very rapidly with CH 3 -H 4 SPt. This demonstrates that methyltransferase activity is an intrinsic property of the ␥␦ corrinoid component. Although this differs from the clostridial corrinoid/iron sulfur protein, the intrinsic activity is analogous to the ability of the active form of methionine synthase to catalyze methylation of the bound B 12 cofactor by reaction with N 5 -methyl-H 4 folate.
Equilibrium studies of methyl transfer between the enzymebound corrinoid moiety and the reduced pteridine substrate indicate that the process is freely reversible under physiologically relevant conditions. The equilibrium constant for CH 3 -H 4 SPt:cob(I)amide-protein methyl transfer (Reaction 3) was 5.5 [H ϩ ] Ϫ1 M Ϫ1 (Fig. 7). Therefore, in the process of acetyl-CoA cleavage, methyl group transfer from the enzyme to H 4 SPt is slightly thermodynamically unfavorable under standard state conditions (KЈ eq ϭ 1/5.5, ⌬G 0 Ј ϭ ϩ1.0 kcal/mol). However, very low levels of CH 3 -H 4 SPt are detected during purification of H 4 SPt from cell extracts, and based on the amounts of ACDS complex and H 4 SPt obtained during purification from an equal amount of cell paste, the ratio of H 4 SPt to enzyme cobamide is estimated to be approximately 50:1. These findings indicate that at equilibrium in the presence of physiological concentra-tions of enzyme and CH 3 -H 4 SPt, approximately 91% of the enzyme-bound methyl groups would be transferred to the cellular pool of H 4 SPt. Demethylation of the enzyme corrinoid is also exceedingly rapid. Therefore, it is unlikely that methyl group transfer to H 4 SPt presents either a kinetic or thermodynamic barrier of significance in the overall process of acetyl-CoA cleavage in vivo. Component 2, CO:Acceptor Oxidoreductase-A previous study showed that a portion of the CO dehydrogenase ␣⑀ component could be dissociated from the M. thermophila enzyme complex in the presence of a cationic detergent (8). We now show that complete release of the CO dehydrogenase component occurs as a result of limited proteolytic digestion of the enzyme complex (Fig. 5B). Two different forms of CO dehydrogenase have been noticed previously in extracts of M. thermophila (1), 5 and M. barkeri (2,3). The ability to generate a second form of the enzyme by proteolytic action now suggests an explanation for the presence of the different forms.
Component 3, Acetyltransferase-Whereas previous investigations demonstrated acetyl transfer/exchange activity of the intact enzyme complex (19), we now show that the site of redox-dependent acetyltransferase is located on the 60-kDa ␤ subunit. As revealed by SDS-PAGE, low levels of chymotrypsin act selectively on the 60-kDa ␤ subunit to produce a truncated subunit (␤*) of about 50 kDa (Fig. 2) that possesses a high specific activity of acetyltransferase (Fig. 5). Similar results were also found with other proteases such as bromelain and trypsin (data not shown). Thus, the data suggest that a region of the ␤ polypeptide is highly susceptible to general proteolytic attack. In experiments in which the extent of digestion is varied, overall acetyl-CoA synthetic activity is lost in direct proportion to the loss of the intact ␤ subunit (Fig. 4) 6. Reduction of the protein-bound cobamide with Ti 3؉ -EDTA and reaction with CH 3 -H 4 SPt. Reduction of the ACDS corrinoid protein subcomponent and titration with CH 3 -H 4 SPt was carried out at 23°C with a solution (800 l) that contained 50 mM Tris⅐HCl, pH 7.5, and the corrinoid protein (3.55 M cobamide) obtained from ionexchange chromatography of the proteolytically digested enzyme complex (peak 1 in Fig. 3). The processes of reduction by Ti 3ϩ and methylation by CH 3 -H 4 SPt were monitored spectrophotometrically. At the indicated time point 12 l of 64 mM Ti 3ϩ -EDTA was added. Addition of the indicated amounts of CH 3 -H 4 SPt was made from a 0.485 mM stock solution. After each addition of CH 3 -H 4 SPt, two 60-l samples were removed from the cuvette, mixed with 60 l each of a solution containing 20 mM 2-thioethanesulfonate and 0.5 M sodium citrate, pH 4.0, and frozen in liquid nitrogen. In each sample, the [H 4 SPt]/[CH 3 -H 4 SPt] concentration ratio was subsequently determined by HPLC analysis according to the procedure described previously (5). Absorbance data have been corrected for dilution resulting from reagent additions. of the complex. Loss of the intact 48-kDa ␦ subunit occurs as the level of digestion is further increased and correlates with the ability to obtain high resolution of the three protein components (and with the absence of residual undissociated enzyme). The results allow us to formulate the hypothesis that integrity of the ␦ subunit may be essential for maintaining the quaternary structure of the enzyme complex. Furthermore, since loss of acetyl-CoA synthetic activity occurs without overall dissociation of the enzyme complex, the involvement of the ␤ subunit apparently extends beyond that of a structural role or that of the ability to carry out acetyl transfer. Further investigations are needed to explain the precise mechanism for the decline in overall ACDS activity coinciding with the loss of the intact ␤ subunit.
In summary, the results demonstrate catalytic roles for the ␥␦ (or possibly ␥ alone), ␣⑀, and ␤ protein subcomponents of the ACDS complex. We may now write reactions 5-7 to include the enzyme subcomponents with a subscript denoting the relevant subunits. The ␤ subunit catalyzes the exchange/transfer of the acetyl group of acetyl-CoA. Consequently a likely role for the ␤ subunit in acetyl-CoA synthesis in the intact ACDS complex would be to catalyze the reversible synthesis of acetyl-CoA from CoA and an acetyl group bound to the ␤ subunit. The role of the ␥ subunit in the intact ACDS complex would be to catalyze the reversible transfer of the methyl group from CH 3 -H 4 SPt to the bound cobamide cofactor, and perhaps ultimately to the site of acetyl synthesis. The exact role of the ␦ subunit is unknown, however, our data suggest that integrity of the ␦ subunit could be involved in maintaining the overall quaternary structure. During the process of acetyl-CoA cleavage, the transfer of the methyl group from the corrinoid protein to the cellular pool of H 4 SPt is rapid and thermodynamically favorable under conditions likely to exist in vivo.