Purification and Properties of the Inducible Coenzyme A-Linked Butyraldehyde Dehydrogenase from Clostridium Acetobutylicum NEIL R

Home , Butyraldehyde

JOURNAL OF BACTERIOLOGY, JUlY 1988, p. 2971-2976 Vol. 170, No. 7 0021-9193/88/072971-06$02.00/0 Copyright © 1988, American Society for Microbiology Purification and Properties of the Inducible Coenzyme A-Linked Butyraldehyde Dehydrogenase from Clostridium acetobutylicum NEIL R. PALOSAARI* AND PALMER ROGERS Department of Microbiology, University of Minnesota, Minneapolis, Minnesota 55455 Received 13 November 1987/Accepted 25 March 1988

The coenzyme A (CoA)-linked butyraldehyde dehydrogenase (BAD) from Clostridium acetobutylicum was characterized and purified to homogeneity. The enzyme was induced over 200-fold, coincident with a shift from an acidogenic to a solventogenic fermentation, during batch culture growth. The increase in enzyme activity was found to require new protein synthesis since induction was blocked by the addition of rifampin and antibody against the purified enzyme showed the appearance of enzyme antigen beginning at the shift of the fermentation and increasing coordinately with the increase in enzyme specific activity. The CoA-linked acetaldehyde dehydrogenase was copurified with BAD during an 89-fold purification, indicating that one enzyme accounts for the synthesis of the two aldehyde intermediates for both butanol and ethanol production. Butanol dehydrogenase activity was clearly separate from the BAD enzyme activity on TEAE cellulose. A molecular weight of 115,000 was determined for the native enzyme, and the enzyme subunit had a molecular weight of 56,000 indicating that the active form is a homodimer. Kinetic constants were determined in both the forward and reverse directions. In the reverse direction both the V,ax and the apparent affinity for butyraldehyde and caproaldehyde were significantly greater than they were for acetaldehyde, while in the forward direction, the Vmax for butyryl-CoA was fivefold that for acetyl-CoA. These and other properties of BAD indicate that this enzyme is distinctly different from other reported CoA-dependent aldehyde dehydrogenases.

Recently, there has been renewed interest in the obligate possible since the assay and stability problems have been anaerobe Clostridium acetobutylicum and its ability to fer- overcome. ment carbohydrates to acetone, butanol, and ethanol (21). We report here the induction of BAD activity and enzyme From an applications point of view, it is important to protein in cells just prior to solvent production and the maximize the efficiency of production of desired products purification and properties of BAD. We conclude that the and to be able to control the ratios of products based on CoA-linked BAD is most likely the branch-point enzyme for demand (2). To this end it is essential to understand the both ethanol and butanol synthesis from acetyl-CoA and regulation of the activity and the amount of the key enzymes butyryl-CoA. Furthermore, the properties of the BAD from that catalyze the fermentation. It has been known for some C. acetobutylicum indicate that it is significantly different time that, during this saccharolytic fermentation, C. aceto- from other reported CoA-linked aldehyde dehydrogenases. butylicum undergoes a shift from producing acetate and (Parts of this study were reported earlier [N. Palosaari and butyrate to forming butanol, ethanol, and acetone (10). The P. Rogers, Fed. Proc. 46:2293, 1987].) biochemical pathways for both modes of fermentation have been defined in a general way, and there has been some work MATERIALS AND METHODS done on elucidating the signals which affect the shift from Materials. Butyryl phosphate was synthesized with buty- acidogenic to solventogenic fermentation (for reviews, see ric anhydride (Eastman Kodak Co., Rochester, N.Y.) and references 17 and 21). Significant progress has been reported K2HPO4 and crystallized to greater than 95% purity by the on the analysis of the enzyme activities involved in the method of Stadtman (reference 27, procedure B) for the acidogenic stage of fermentation, in particular, the acyl preparation of acetyl phosphate. All aldehydes were ob- kinases and phosphoacetyltransferases from C. acetobutyli- tained from Aldrich Chemical Co. (Milwaukee, Wis.); cum and other clostridia that are involved in acetate and acetyl-CoA, reduced CoA (CoASH), NAD, NADH, NADP, (5, 12, 13, 26, 28-30). However, only the butyrate synthesis NADPH, and Reactive Blue-2 affinity gel were from Sigma coenzyme A (CoA)-linked aldehyde dehydrogenases from Chemical Co. Mo.); and Bio-Gel P200, Cellex T, Clostridium kluyveri (19, 25) and Escherichia coli (8, 9, 23, (St. Louis, and protein dye reagent were from Bio-Rad Laboratories 24) have been purified and characterized extensively, and (Richmond, Calif.). they are quite different than the enzyme butyraldehyde Preparation of butyryl-CoA. Butyryl-CoA was prepared dehydrogenase (BAD) from C. acetobutylicum that we re- enzymatically from butyryl phosphate by using the phos- port here. Preliminary studies of BAD activity changes phate butyryltransferase (PBT; butyryl-CoA:orthophos- during fermentation by C. acetobutylicum have been re- phate butyryltransferase [EC 2.3.1.19]) from C. acetobutyli- ported (1, 13), although it was apparent that there were cum. The equilibrium of this reaction favors butyryl-CoA difficulties with the assay conditions. Similarly, studies on formation (Keq = 74 [26]). The partially purified PBT was the alcohol dehydrogenases of Clostridium beijerinkii (Clo- isolated from the same Cellex T column that was used to stridium butylicum) (16) and C. acetobutylicum (22) are now purify BAD. The PBT-active fractions were pooled, concentrated, and applied to a P200 column as described below for BAD. The PBT-active fractions from the P200 column were * Corresponding author. pooled and used for the preparation of butyryl-CoA. The 2971 2972 PALOSAARI AND ROGERS J. BACTERIOL. preparative reaction (volume, 1.0 ml) contained potassium cell concentrator (Amicon) by using a PM10 tnembrane. The N-(morpholinopropionyl)sulfonate (MOPS; pH 7.0; 50 pool was then applied to a column (2.5 by 45 cm) of Bio-Gel ,umol), CoASH (4 ,umol), butyryl phosphate (6 imol), and P200 equilibrated with KPM buffer, and the column was partially purified PBT (5 U; see below). The reaction was eluted with the same buffer. incubated for about 5 min at room temperature and moni- (iv) Affinity gel chromatography. The active fractions from tored spectrophotometrically at 233 nm to determine com- the gel filtration column were pooled and made to 40 mM pletion of the reaction. The butyryl-CoA was separated from KCI by adding the solid salt. The pool was then applied PBT by dialysis in a centrifugal dialysis apparatus (Centricon slowly (about 10 ml/h) to a column (1.5 by 7.5 cm) of 30; Amicon Corp., Lexington, Mass.). The effluent with the Reactive Blue-2 affinity gel which had been equilibrated with butyryl-CoA was acidified with 6 N HCI to hydrolyze the 40 mM KCI in KPM buffer. The column was then washed remaining butyryl phosphate (2 min at pH between 2 and 4) stepwise with 15 ml each of KPM buffer containing 50 mM and then neutralized with 1 M MOPS to pH 7.0. The yield of KCI, 100 mM KCI, and 150 mM KCI. The enzyme activity butyryl-CoA was quantitated by ion-paired, reversed-phase was then eluted from the column with 20 ml of KPM buffer high-performance liquid chromatography (3). High-perfor- containing 200 mM KCl and 5 mM NAD. mance liquid chromatography was performed on a C18 Storage. The active fractions from the affinity column were column (4.6 mm by 25 cm; 10-,um-diameter particles), with pooled, and glycerol was added to 25% (vol/vol). The pool 10 mM tetrabutylammonium phosphate as the ion-pairing was then divided into 0.5-ml samples, quick-frozen on dry agent, in aqueous methanol (45% [vol/vol] methanol) at pH ice, and stored at -80°C. The activity of purified BAI) 6.0. The column was eluted isocratically at 1.0 ml/min. remained stable for several months under these conditions. Conversion of CoA to butyryl-CoA was consistently 94%, as Polyacrylamide gel electrophoresis. Nondenaturing poly- determined by high-performance liquid chromatography. acrylamide gel electrophoresis was performed in 7.5% acryl- Recovery after dialysis and acidification was typically 70 to amide gels with 15 mM potassium barbital buffer (pH 8.5) 80%. containing 5 mM 2-mercaptoethanol and 0.1 M sucrose at Organism and growth conditions. C. acetobutylicum B643 4°C for 3 to 3.5 h and 150 V. Sodium dodecyl sulfate- was obtained from L. K. Nakamura (Northern Regional polyacrylamide gel electrophoresis was done by the proce- Research Center, Peoria, Ill.). Cells were grown in yeast dure of Laemmli (18). Silver staining was done by the extract medium (YEM), which consisted of the glucose- procedure of Guevara et al. (11). minimal medium of O'Brien and Morris (20), without biotin Enzyme assays. Aldehyde dehydrogenase activity was or p-aminobenzoic acid, and supplemented with yeast ex- measured in the reverse direction (aldehyde-oxidizing direc- tract (8 g), casein hydrolysate (2.2 g), asparagine (1.0 g), and tion; see Table 3) in 1.0-ml reaction mixtures containing cysteine (0.5 g). The pH was adjusted to 7.0. Cultures were potassium 2-(N-cyclohexylamino)ethane sulfonate (CHES; grown and maintained under anaerobic conditions as de- pH 9.0; 50 ,umol), NAD (1.0 ,umol), dithiothreitol (10 ,umol), scribed previously (22). Larger cultures (12 liters of YEM CoA (reduced form; 0.2 ,umol), protein (2.5 to 250 jig), and with 5% glucose) were inoculated with 120 ml of exponen- butyraldehyde or another aldehyde (10 ,umol). The assay tially growing cells and grown in a 16-liter fermentor (New solutions were preincubated for 10 min at room temperature Brunswick Scientific Co., Inc., Edison, N.J.), which was before the reaction was initiated by the addition of aldehyde. stirred continuously at 100 rpm under an anaerobic atmo- Enzyme activity was measured by NADH formation, as sphere (5% hydrogen, 10% carbon dioxide, and 85% nitro- determined by measuring the increase in the A340 (14) on a gen) for 20 h. The cells were concentrated in a concentrator spectrophotometer (Response II; Gilford Instrument Labo- (DC1OL with an HSMP01-43 cartridge; Amicon) and washed ratories, Inc., Oberlin; Ohio) with a kinetic program. The twice with 2 liters of ice-cold 25 mM potassium phosphate aldehyde dehydrogenase was assayed in the forw'ard direc- buffer with 10 mM 2-mercaptoethanol (pH 7.4). The washed tion in 1.0-ml reaction mixttures containing MOPS (ph 7.0; 50 cells were centrifuged for 5 min at 1,500 x g. The cell pellets ,umol), NADH (0.32 ,umol), acyl-CoA (0.2 ptmol), and af- were quick-frozen on dry ice and stored at -80°C. About 12 finity gel-purified BAD (3.0 to 12 ,ug). The reaction was to 15 g of cells per liter were recovered from these larger- initiated by the addition of the acyl-CoA (butyryl-CoA or scale fertnentations. acetyl-CoA), and the activity was determined by measuring Enzyme purification. (i) Preparation of crude extracts. the rate of NADH oxidation, as described above for the Frozen cells (15 g) were suspended in 75 ml of 25 mM assay in the reverse direction. In both assay directions, 1 U potassium phosphate buffer with 10 mM 2-mercaptoethanol of enzyme activity was defined as the amount of enzyme (pH 7.4), and cells were broken by three 1-min pulses of required to form or oxidize 1 ,umol of NAD(P)H per min. ultrasonic treatment at 1 to 4°C. The cell extract was The NADP-dependent alcohol dehydrogertase (alcohol: centrifuged for 30 min at 27,000 x g and 4°C. The clear NADP+ oxidoreductase [EC 1.1.1.2]) was assayed in the supernatant (crude extract) was pipetted off and either used reverse direction (alcohol-oxidizing direction), with 50 mM immediately for chromatography or quick-frozen on dry ice n-butanol used as the substrate (butanol dehydrogenase), as and stored at -80°C. described previously (22). One unit of activity was the same (ii) Anion-exchange chromatography. All steps in the puri- as that defined above for the aldehyde dehydrogenase. fication of enzymes were performed at 4°C. Crude extract PBT was assayed in the reverse direction, as described (60 ml; about 720 mg of protein) was applied to a column of previously (22). One unit of activity was defined as the Cellex T (TEAE cellulose; 6 by 18 cm; Bio-Rad) that was amount of enzyme required to form 1 ,umol of acyl-CoA per equilibrated with 10 mM potassium phosphate buffer with 10 min. mM 2-mercaptoethanol (KPM buffer; pH 7.4). The column In situ activity assay. Immediately after nondenaturing was eluted at 75 mI/h with a linear gradient of 0.0 to 0.6 M electrophoresis gels were run, they were placed in 20 ml of KCI in column buffer (total volume, 1,500 ml), and 8-ml the aldehyde dehydrogenase assay mixture described above fractions were collected. for the reverse direction. The gels were preincubated for 5 (iii) Gel filtration. The activity peak from the anion- min before the aldehyde was added. The gels were then exchange column was pooled and concentrated in a stirred incubated for 15 to 30 min at room temperature in a covered VOL. 170, 1988 BUTYRALDEHYDE DEHYDROGENASE OF C. ACETOBUTYLICUM 2973

40 16- 14F E 30 E .' 12 +Rif 0 C7 w 0 i -0- I E E 'IA CD U) El 10 20 O E .0c 0 *0 + Rif 0)(A c 0) 0. .8 IE :3 10 w i,

E 4 N 0 C A Time (Hours) 21 FIG. 1. Induction of BAD activity during fermentation by C. acetobutylicum. A batch fermentation by C. acetobutylicuim was carried out in YEM with 5% glucose. The medium was inoculated 12 14 16 18 with 2.5% (vol/vol) of a freshly heat-shocked spore culture in the Time (hr) same medium and grown at 37°C. At the indicated times, samples were removed, cells were harvested, crude extracts were prepared, FIG. 2. Effect of rifampin on the induction of BAD. A batch and BAD specific activity (O) was determined as described in the culture of C. acetobutylicum was grown in YEM with 1% glucose. text. The concentrations (in millimolar) of butanol (A) and butyrate The medium was inoculated with 1% (vol/vol) of a spore culture in (A) in the medium were determined by gas chromatography. Culture the same medium and grown at 37°C. Culture samples were with- turbidity (0) was measured at an A6. in a Klett-Summerson drawn at the indicated times, and crude extracts were prepared from colorimeter, in which 0.2 optical density units = 2.5 x 108 CFU/ml. washed cells. Milliunits of BAD specific activity (nanomoles per minute per milligram of protein) were determined as described in the text. One culture was left untreated (0) and subcultures were glass dish. Activity was detected by observing the fluores- removed during the induction to which rifampin (+Rif) was added to cence of NADH under long-wave UV light. 2 ,ug/ml at 15 h (-) and 16 h (A) (arrows). Protein determination. Protein concentrations were determined by the dye-binding method of Bradford (6), with bovine serum albumin used as the standard. neously with BAD (data not shown). The induction of BAD Antibody production. Purified BAD (50 to 100 ,ug) in was blocked by the addition of rifampin (2 lLg/ml of culture; Freund incomplete adjuvant was administered subcutane- Fig. 2). Similarly, the induction of BAD was blocked by ously to New Zealand White rabbits. Four injections were chloramphenicol (100 ,ug/ml). We noted that not only was the administered at 2-week intervals. Rabbits were bled 7 days rise in specific activity blocked by the addition of rifampin after the fourth injection. Antisera were prepared and were but following the addition of these inhibitors there was also assayed for anti-BAD antibody by double radial immunodif- a fairly rapid drop in the specific activity of BAD. We did not fusion by standard procedures (15). The antisera gave a study this further. In contrast to BAD, the specific activities precipitin band when assayed with 90 ng of the purified BAD of PBT and butyrate kinase (the enzymes required for enzyme. butyrate formation) remained relatively constant over the Determination of fermentation products. The concentra- entire course of the fermentation, with only a two- to tions of butyrate and butanol were determined on extracted fourfold variation in specific activity (data not shown) (5). samples of the culture supernatants after cells were removed Further evidence for the induction of new enzyme protein by gas-liquid chromatography with a thermal conductivity came from results of the double radial immunodiffusion detector by previously described methods (22). The concen- assay (15) (Table 1). When cell samples removed during the trations of butyrate and butanol were calculated by compar- fermentation (Fig. 1) were reacted with anti-BAD antiserum, ison with standards that were extracted as described above precipitin bands were revealed that correlated with the rise for culture samples and were expressed as millimolar in the in the specific activity of BAD. We reported previously (22) original culture medium. that extracts of a nonsporulating, nonsolventogenic mutant of C. acetobutylicum (strain 6A) contained no detectable RESULTS BAD activity. Extracts of strain 6A from various stages of fermentation at any dilution were found to be devoid of BAD Induction of CoA-linked BAD. The CoA-linked BAD was antigen when the same antiserum was used. found to be an inducible enzyme. The approximately 200- Purification of CoA-linked BAD from C. acetobutylicum. fold rise in specific activity for BAD correlated with the The BAD enzyme was purified 89-fold from crude extracts of transition period or switch of the fermentation between acid C. acetobiutylicum by the procedures described above (Table and solvent production (Fig. 1). During this time, the cells 2). The final product was greater than 95% pure, as indicated completed exponential growth but continued a linear mass by silver staining of sodium dodecyl sulfate-polyacrylamide increase, halted the net synthesis of butyrate, and com- gels after electrophoresis (Fig. 3B). Since the fermentation menced butanol synthesis immediately following the rise in by C. acetobutylicum produces ethanol as well as butanol, it BAD specific activity (Fig. 1). The NADP-specific butanol was possible that this bacterium might produce two separate dehydrogenase activity was found to be induced simulta- aldehyde dehydrogenase enzymes. However, the activity for 2974 PALOSAARI AND ROGERS J. BACTERIOL.

TABLE 1. Appearance of BAD antigen in C. acetobutylicuim during a fermentation time course Time (h) Immunoprecipitation reaction at the following dilution": cell sample withdrawn" 1:1 1:2 1:4 1:8 1:16 1:32 1:64 10.0 11.3 - - - r 12.8 + ± - - _ 4 14.3 ++ ++ + + + 15.8 ++ ++ ++ ++ ++ ++ ± _. ;e,,{Wk 17.8 ++ ++ ++ ++ ++ ++ ± 21.8 ++ ++ ++ ++ ++ ++ + 32.0 ++ ++ ++ ++ ++ ++ + Cells samples were from the experiment for which the results are shown in Fig. 1. FIG. 3. Polyacrylamide gel electrophoresis of BAD. (A) Nonde- b The immunoprecipitation reactions were between anti-BAD antiserum naturing polyacrylamide gel electrophoresed for 3 to 3.5 h at 150 V. and crude extracts of cell samples, which were prepared as described in the Lanes 1 and 2, Enzyme purified to the affinity gel step (0.3 ,ug) and text. Portions (25 ,ul) of crude extract (200 to 450 ,g of protein at 1:1 dilution) the Cellex T step (6 pLg) (see Table 2), respectively, and assayed for were serially diluted, as indicated, and placed in wells of Ouchterlony BAD; lanes 3 and 4, enzyme preparations (0.3 and 6 pug, respec- diffusion plates opposite 25 ,u1 ofanti-BAD antiserum, as described in the text. tively) assayed for acetaldehyde dehydrogenase activity. The in situ Visual detection of immunoprecipitin bands was designated as follows: -, no method that was used was described in the text. (B) Sodium dodecyl precipitin band; ±, barely detectable band; +, easily detectable band: + +. sulfate-polyacrylamide gel electrophoresis of the affinity gel prepa- strong precipitin band. ration (0.16 ,g) of BAD (lane 1) and the following molecular size markers (lane 2): P-galactosidase, 116 kDa; bovine serum albumin, acetaldehyde dehydrogenase was found to copurify with 67 kDa; chicken egg albumin, 45 kDa; and carbonic anhydrase, 29 BAD activity, since the proportion of the two activities kDa. The gel was silver stained as described in the text. remained constant throughout the purification procedure (Table 2). Anion-exchange chromatography also revealed a single, fairly symmetrical peak of BAD activity that co- pH 9.0, which is similar to the pHs for other aldehyde eluted with acetaldehyde dehydrogenase activity (Fig. 4). dehydrogenases (23, 25). Different buffers gave slightly Following electrophoresis in a nondenaturing gel, the in situ different activities at the same pH. Of the buffers tested, assay of BAD and acetaldehyde dehydrogenase activities CHES gave the highest activity and was therefore used as showed only a single band of activity at the same position the buffer of choice in the standard assay. The pH optimum (Fig. 3A), indicating that a single enzyme protein carries in the forward direction was near 7.0, which is in agreement both activities. with results presented earlier by Andersch et al. (1). We Previous reports indicate that the CoA-linked aldehyde observed that a Tris hydrochloride buffer could not be used dehydrogenase and alcohol dehydrogenase activities of E. with this enzyme, since cell extracts prepared with Tris coli (23, 24) and C. kluyveri (19) may exist as a single protein. buffer were unstable and activity measured in Tris buffer was In the case of C. acetobutylicum, we found that these two less than 10% of that with CHES or glycylglycine buffer. activities were well separated by anion-exchange chroma- While it was necessary to add high concentrations of tography (Fig. 4), thereby indicating that there are discrete sulfhydryl reagent (2-mercaptoethanol or dithiothreitol) to proteins for each enzyme activity. We also found that the detect activity in the crude extract, it was not required for stabilities of these two activities are very different. The activity assays with the purified enzyme, although it did help alcohol dehydrogenase is much less stable, with a half-life of to maintain enzyme activity. There was no apparent require- 8 h at 0°C in crude extracts, while BAD has a half-life of ment for metals with this enzyme, since there was no greater than 3 days at 0°C. enhancement of activity when Zn2+, Mg2+, or Fe3" was Properties of BAD. Under denaturing conditions the puri- added and no loss when 1 mM EDTA was added. fied enzyme migrated as a single 56-kilodalton (kDa) poly- Kinetic analysis. As demonstrated by the Km values (Table peptide (Fig. 3B), while by gel filtration on a P200 column 3), the apparent affinity of NAD+ and NADH for BAD was (2.5 by 45 cm; see above) the native active enzyme migrated far greater than that of NADP+ and NADPH. It was also as a 115-kDa protein (data not shown), indicating that the found that CoASH inhibited the reaction at concentrations native form of BAD is most likely a homodimer. above 0.3 mM, which has been reported for another clos- The effect of pH on enzyme activity was assayed in the tridial CoA-linked aldehyde dehydrogenase (25). In the reverse direction, and maximum activity occurred at about reverse direction, BAD catalyzed the oxidation of several

TABLE 2. Purification of the CoA-linked BAD from C. acetobutylicum

Purification step Protein Total activity Yield Sp act (U/mg) ofb: Purification BAD/AAD' (mg) (U)" ~~~~~~~~BADAAD (od Crude extract 720 90 100 0.12 0.011 1 11.6 Cellex T 95 74 82 0.78 0.068 6.2 11.4 P200 28 52 58 1.86 0.166 15 11.2 Reactive Blue-2 1.8 20 22 11.1 0.99 89 11.2 a Total activity is defined as follows: 1 U = 1 pumol of NADH produced per min of assay in the reverse direction, as described in the text. b Specific activities (units per milligram of protein) were determined; AAD, acetaldehyde dehydrogenase. ' Ratio of specific activities of BAD to AAD. VOL. 170, 1988 BUTYRALDEHYDE DEHYDROGENASE OF C. ACETOBUTYLICUM 2.975

TABLE 3. K,,, and Vmax values for BAD from 120 C. acetobutylicum'

I Direction Substrate (mM) Vmax (nmol/min) 'E 90 , 0 0E E 0 Forward Acetyl-CoA 0.055 3.4 N t_ Butyryl-CoA 0.045 18.1 (+NADH) 60 Xr 0.5 I 24.3 (+NADPH) :- _. NADH 0.003 14.7 'a _ 41 NADPH 0.040 25.6 .30 X' 0.25 or Reverseb Acetaldehyde 7.2 60 am- Propionaldehyde 6.7 660 Butyraldehyde 1.6 875 JO O. Caproaldehyde 0.37 615 Column Fraction CoA 0.045 925 NAD+ 0.16 850 FIG. 4. Elution of BAD from C. acetobutylicum on TEAE NADP+ 2.0 880 cellulose. A freshly prepared crude extract (60 ml, 720 mg of protein) was applied to a column (6 by 18 cm) of Cellex T and eluted "The enzyme preparation used was purified to 11.1 U/mg of protein (Table with a linear gradient of 0.0 to 0.6 M KCI (- -) in KPM buffer, as 2). Activities were assayed as described in the text, and the kinetic values described in the text. The A280 ( ) of individual 8-ml fractions were determined by a double-reciprocal plot of velocity versus substrate was measured over the course of the gradient. Fractions were concentration. ' Vl,,,, values for the reverse direction were determined by secondary plots assayed for acetaldehyde dehydrogenase (DH) (O), BAD ([1), and of y intercepts (from a series of primary double-reciprocal plots) versus the butanol dehydrogenase (BDH) (A) activities as described in the text. reciprocal of substrate concentration, as described previously (25). Enzyme activities are expressed as milliunits per milliliter of fraction (nanomoles per minute per milliliter). exists in two forms. A soluble form has a reported molecular straight-chain aldehydes, although it was most active weight of 290,000 (25), which is more than twice the molec- (greatest Vmax) with butyraldehyde. Also, there was an ular weight we found for BAD, and has negligible alcohol 18-fold drop in the Km of the aldehyde for BAD with dehydrogenase activity, like BAD. Also, there has been increasing chain length from acetaldehyde to caproaldehyde. reported (19) a particulate form from C. kluyveri that is 194 In the forward direction it can be seen that while there was kDa, that is composed of two different subunits (55 and 42 very little difference in the Km values for CoA, acetyl-CoA, kDa), and that contains acetaldehyde and alcohol dehydrog- and butyryl-CoA, the significant difference in the Vmax (in enase activities. The particulate form from C. kluyveri is the forward direction) indicates that the turnover rate for interesting in two respects. First, the larger subunit is very butyryl-CoA by BAD is 5.5-fold higher than that for acetyl- similar in size to the C. acetobutylicum BAD subunits (56 CoA. kDa); and second, both enzymes exhibit a higher apparent affinity for NAD(H), but both can also utilize NADP(H). In DISCUSSION contrast, the soluble enzymes from C. kluyveri and E. coli were specific for NAD+ and inactive with NADP+ (23, 24). The induction of BAD and other solventogenic enzyme The enzyme from C. acetobutylicum had a much lower Km activities by C. acetobutylicum had been reported previ- for NAD(H) than for NADP(H), both in the forward and ously (1, 4). However, for BAD and butanol dehydrogenase, reverse directions (Table 3). However, in the forward direc- the level of increase was not demonstrated clearly (21). By tion the Vmax for NADPH was slightly higher than that for use of an improved method for the detection of BAD NADH, a curious property that has not been reported for activity, we demonstrated that the increase in enzyme activ- other CoA-linked aldehyde dehydrogenases. The enzymes ity that we observed during the fermentation switch corre- from C. kluyveri and C. acetobutylicum are quite different in lated with the synthesis of new protein. The main improve- their activities with different aldehydes. The Km values for ment in detecting enzyme activity was to prepare extracts in acetaldehyde, propionaldehyde, and butyraldehyde are 1.5, potassium phosphate buffer and to assay the reaction in the 4.5, and 11.0 mM, respectively, for the C. kluyveri enzyme reverse direction with CHES or glycylglycine buffer in order (7), while our data for the enzyme from C. acetobutylicum to maintain enzyme activity and to favor catalysis. The Tris showed just the opposite trend for K,n versus chain length of buffer used by other workers (1, 4) destabilized the enzyme the aldehyde (Table 3). The decrease in activity with al- and did not permit the detection of significant activity. In dehydes from acetaldehyde to butyraldehyde has been re- contrast, butanol dehydrogenase activity survived well in ported with the enzyme from E. coli B, although Km values extracts made with Tris buffer and was catalytically active in were not given (23). the same buffer. We do not know the reasons for these Our examination of the Kin values for acetyl-CoA and differences, but these data strengthen the view that for C. butyryl-CoA by assaying BAD in the forward direction acetobutylicum, these dehydrogenases are separate pro- revealed almost identical values (0.055 and 0.045 mM, re- teins. spectively), which were only slightly higher than those for In comparison with two of the well-studied CoA-linked CoA in the reverse direction (0.045 mM) (Table 3). The aldehyde dehydrogenases from C. kluyveri and E. coli, the relative turnover rates (Vmax) for butyryl-CoA to acetyl-CoA BAD from C. acetobutylicum is significantly different. The of 5.5 to 1 that we reported here could potentially produce apparent molecular weight for the enzyme from E. coli is the same ratio for the turnover of butyraldehyde to acetal- about 200,000; this appears to be a double-headed enzyme dehyde. Thus, BAD may ultimately be responsible for the that carries both alcohol dehydrogenase and CoA-linked final observed molar ratio of 6:1 for butanol to ethanol during aldehyde dehydrogenase activities (8, 9, 23, 24). The CoA- the typical AEB (acetone-ethanol-butanol) fermentation (for linked aldehyde dehydrogenase of C. kluyveri apparently a review, see references 17 and 21). 2976 PALOSAARI AND ROGERS J. BACTERIOL.

Recent data from our laboratory for C. acetobuttylicum butyl alcohol fermentation. I. Nutritional and other factors (22) and others for C. beijerinkii (C. butylicum) (16) indicate involved in the preparation of active suspensions of Clostridium that only a single NADPH-dependent alcohol dehydroge- acetobutylicum. Biochem. J. 35:1320-1331. nase is produced by these bacteria. The kinetic properties of 11. Guevara, J., Jr., D. A. Johnston, L. S. Ramagali, B. A. Martin, the alcohol dehydrogenase from C. acetobutylicum and the S. Capetillo, and L. V. Rodriguez. 1982. Quantitative aspects of possible role it may play in the of silver deposition in proteins resolved in complex polyacryl- modulating stoichiometry amide gels. Electrophoresis 3:197-205. alcohol production from the aldehydes produced via BAD 12. Hartmanis, M. G. N. 1987. Butyrate kinase from Clostridium are unknown. a(etobutylicum. J. Biol. Chem. 262:617-621. The stability characteristics, the requirement for thiol 13. Hartmanis, M. G. N., and S. Gatenbeck. 1984. Intermediary reagents, the pH optima, and the K., values for CoA of the metabolism in Clostridium acetobutylicum: levels of enzymes CoA-linked aldehyde dehydrogenases from E. coli (8, 23) involved in the formation of acetate and butyrate. Appl. Envi- and C. kluyveri (25) reported previously are similar to those ron. Microbiol. 47:1277-1283. we have reported here. On the other hand, the kinetic 14. Horecker, B. L., and A. Kornberg. 1948. The extinction coeffi- characteristics, which show a preference for longer-chain cients of the reduced band of pyridine nucleotides. J. Biol. substrates by affinity (reverse direction) or Vmax (forward Chem. 175:385-389. direction), lead to the conclusion that the BAD of C. 15. Hudson, L., and F. C. Hay. 1980. Practical immunology, 2nd ed. acetobutylicum is distinctly different from other CoA-linked Blackwell Scientific Publications, Boston. acetaldehyde dehydrogenases (EC 1.2.1.10) from bacteria 16. Hui, S. F., C.-X. Zhu, R.-T. Yan, and J.-S. Chen. 1987. Butanol-ethanol dehydrogenase and butanol-ethanol-isopro- which must deal only with the interconversion of acetyl-CoA panol dehydrogenase: different alcohol dehydrogenases in two and ethanol. strains of Clostridium beijerinckii (Clostridium butylicum). Appl. Environ. Microbiol. 53:697-703. ACKNOWLEDGMENTS 17. Jones, D. T., and D. R. Woods. 1986. Acetone-butanol fermen- This study was supported by grant DE-FG02-86ER13512 from the tation revisited. Microbiol. Rev. 50:484-524. Division of Biological Energy Sciences, U.S. Department of En- 18. Laemmli, U. K. 1970. Cleavage of structural proteins during the ergy. assembly of the head of bacteriophage T4. Nature (London) We thank Pamela Reilly Contag for preparation of the rabbit 227:680-685. antiserum and assistance with the immunoprecipitin assay and 19. Lurz, R., F. Mayer, and G. Gottschalk. 1979. Electron micro- Steven Schnell for assistance. scopic study on the quaternary structure of the isolated particulate alcohol-acetaldehyde dehydrogenase complex and on its LITERATURE CITED identity with polygonal bodies of Clostridium kluyveri. Arch. Microbiol. 120:255-262. 1. Andersch, W., H. Bahl, and G. Gottschalk. 1983. Level of 20. O'Brien, R. W., and J. G. Morris. 1971. Oxygen and the growth enzymes involved in acetate, butyrate, acetone and butanol and Clostridium formation by Clostridium acetobutylicum. Eur. J. Appl. Micro- metabolism of acetobutylicum. J. Gen. Micro- biol. Biotechnol. 18:327-332. biol. 68:307-318. 2. Bahl, H., M. Gottwald, A. Kuhn, V. Rale, W. Andersch, and G. 21. Rogers, P. 1986. Genetics and biochemistry of Clostridium Gottschalk. 1986. Nutritional factors affecting the ratio of sol- relevant to development of fermentation processes. Adv. Appl. vents produced by Clostridium acetobutylicum. Appl. Environ. Microbiol. 31:1-60. Microbiol. 52:169-172. 22. Rogers, P., and N. Palosaari. 1987. Clostridium acetobutylicum 3. Baker, F. C., and D. A. Schooley. 1979. Analysis and purification mutants that produce butyraldehyde and altered quantities of of acyl coenzyme A thioesters by reversed-phase ion-paired solvents. Appl. Environ. Microbiol. 53:2761-2766. liquid chromatography. Anal. Biochem. 94:417-424. 23. Rudolph, F. B., D. L. Purich, and H. J. Fromm. 1968. Coen- 4. Ballongue, J., J. Amine, E. Masion, H. Petitdemange, and R. zyme A-linked aldehyde dehydrogenase from Escherichia coli. Gay. 1985. Induction of acetoacetate decarboxylase in Clostrid- J. Biol. Chem. 243:5539-5545. ium acetobutylicum. FEMS Lett. 29:273-277. 24. Schmitt, B. 1975. Aldehyde dehydrogenase activity of a com- 5. Ballongue, J., J. Amine, H. Petitdemange, and R. Gay. 1986. plex particle from E. coli. Biochimie 57:1001-1004. Regulation of acetate kinase and butyrate kinase by acids in 25. Smith, L. T., and N. 0. Kaplan. 1980. Purification, properties, Clostridium acetobutylicum. FEMS Lett. 35:295-301. and kinetic mechanism of coenzyme A-linked aldehyde dehy- 6. Bradford, M. M. 1976. A rapid and sensitive method for the drogenase from Clostridium kluyveri. Arch. Biochem. Biophys. quantitation of microgram quantities of protein utilizing the 203:663-675. principle of protein-dye binding. Anal. Biochem. 72:248-254. 26. Stadtman, E. R. 1955. Phosphotransacetylase from Clostridium 7. Burton, R. M., and E. R. Stadtman. 1953. The oxidation of kluyveri. Methods Enzymol. 1:596-599. acetaldehyde to acetyl-coenzyme A. J. Biol. Chem. 202:873- 27. Stadtman, E. R. 1957. Preparation and assay of acetyl phos- 889. phate. Methods Enzymol. 3:228-231. 8. Clark, D. P., and J. E. Cronan, Jr. 1980. Acetaldehyde coen- 28. Twarog, R., and R. S. Wolfe. 1962. Enzymatic phosphorylation zyme A dehydrogenase of Escherichia coli. J. Bacteriol. of butyrate. J. Biol. Chem. 237:2474-2477. 144:179-184. 29. Twarog, R., and R. S. Wolfe. 1963. Role of butyryl phosphate in 9. Cunningham, P. R., and D. P. Clark. 1986. The use of suicide the energy metabolism of Clostridium tetanomorphum. J. Bac- substrates to select mutants of Escherichia coli lacking enzymes teriol. 86:112-117. of alcohol fermentation. Mol. Gen. Genet. 205:487-493. 30. Valentine, R. C., and R. S. Wolfe. 1960. Purification and role of 10. Davies, R., and M. Stephenson. 1941. Studies on the acetone- phosphotransbutyrylase. J. Biol. Chem. 235:1948-1952.