Matheson Gas Products, Inc. All other chemicals from commercial sources ... Sodium dithionite, 1.0 pl, of a 0.4% w/v deoxygenated solution was introduced to ...... Drake. H. L.. Hu. S.-I.. and Wood. H. G. (1980) J. Biol. Chem. 163,1000-1006. 80.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 262, No. 8, ISSUSof March 15, pp. 3706.3712 1987 Printedin Z~S.A.
Carbon Monoxide Dehydrogenasefrom Methanosarcina barkeri DISAGGREGATION, PURIFICATION, AND P H ~ S I C O C H ~ ~ I CPROPERTIES AL OF THE E N ~ Y ~ E * (Received for publication, September 17,1986)
David A, Grahame and ThressaC.Stadtman From the Laborutorv of Biochemitm. National Heart. Lung, and B h d Institute, National Institutesof Health, Bethesda, Marylani 20892 _
I
Carbon monoxide dehydrogenase from acetategrown cells of Methanosarcina burkeriexists in a high molecular weight form (-3 X lo6)under conditions of high ionic strength but is converted toa much smaller form by dialysis. The enzyme was purified by a procedure which exploits isolation of the aggregated form by gel filtration and subsequent dissociation. Following this, the enzyme was purified to within 92% of homogeneity by chromatography on phenyl-Sepharose and finally on hydroxylapatite. Due to the extreme oxygen lability of the enzyme, the entire procedure was carried out within the anaerobic laboratoryat the National Institutes of Health. The enzyme has an c u d 2 oligomeric structure composed of subunits with molecular weights of 19,700 and 84,500. The amino acid compositions of the individual subunits weredetermined. Anaiysis of the metal content by plasma emission spectroscopy indicated 1.3 f 0.3 ( n= 4) nickel and 15.6 2 5.6 ( n = 5)iron permol ofa2&. The enzyme did not contain significantamounts of cobalt or molybdenum. Ferredoxin, FAD, FMN, 2,3,5-triphenyltetrazolium chloride, methyl viologen, and phenazine methosulfate served as electron acceptors; however, the enzyme failed to reduce NAD+, NADP+,or the 8-hydroxy-5-deazaflavin factorF4zo.The optimum pH was between 7 and 9. The apparent Kmfor methyl viologen was 7.1 mM, whereas the value for 2,3,5-triphenyltetrazolium chloride was below 0.5 mM. Strong inhibition was observed by oxygen and cyanide. Inactivation by glyoxaldehyde required enzymatic turnover which suggested that a reactive group was formed, or exposed, on an enzyme intermediate in catalysis. A high degree of thermostability wasnoted. Carbon monoxide, however, rendered the enzyme more susceptible to temperature inactivation.
ide is derived almost exclusively from the carboxyl group of acetate (6,7), yet the enzymatic steps involved in this process are still unidenti~ed. In autotrophic growth of Clostridium thermoaceticum, carbon monoxide dehydrogenase catalyzes the synthesis of acetyl-coA from CO (in an enzyme-bound form derived from reduction of COz)and methyltetrahydrofolate followingtransfer of the methyl group to a corrinoid protein (8, 9). AcetylCoA is an early product of COz assimilation by Methanobacterium t h e r ~ o a u t o t r o ~ h ~(lo), u m and evidence suggests that methyltetrahydromethanopterin, a corrinoid, and CO dehydrogenase are involved in its synthesis (11, 12). It has been postulated that in acetate-adapted M e t ~ n o s a r c i ~barkmi u GO dehydrogenase may participate in methanogenesis by cleavage of acetate (13). This wouldbe analogous to the reverse of the reaction of CO dehydrogenase in acetate biosynthesis. Suggestive evidence for a role in acetate degradation has come from the observation that the levels ofCO dehydrogenase are elevated substantially when methanolgrown species of ~ e t ~ ~ s aarer subsequently c i ~ adapted to growth on acetate (13, 14). Cyanide inhibition of methanogenesis from acetate, but not from methanol or H2/C02, was noticed and interpretedto be consistent with the involvement of CO dehydrogenase in acetate catabolism (15, 16). In addition, immunological studies of Krzycki et al. (17) suggest that CO dehydrogenase may participate in acetate-driven methanogenesis. In order to conduct a detailed investigation on the role of CO dehydrogenase in acetate catabolism large amounts of the purified enzyme are necessary. Therefore, a scheme was developed for large-scale purification of CO dehydrogenase. This purification procedure together with a number of physicochemical and catalytic properties of the highly purified CO dehydrogenase preparations are reported here.
Acetate is an important methanogenic substrate in anaerobic microbial ecosystems (1-3). Members of two genera of methanogens, namely Methanosarcina and Methanothrix, are capable of growth on acetate assole carbon and energy source (4,s). Methanogenesis from acetate proceeds according to the following equation.
Reagents-Commercially available argon and nitrogen of high grade quality were used without further purification. Research grade carbon monoxide, minimum purity 99.99%,was obtained from Matheson Gas Products, Inc. All other chemicals from commercial sources were of reagent grade quality or better. Coenzyme F*m (8hydroxy-5-deazaflavin) was a generous gift from Dr. Edward DeMoll who purified it from ~ e t vnnn~li~. ~ ~ ~ c ~ Cell Culture-M. burkeri was cultivated under strictly anaerobic conditions a t 30 “C using a 10-20% inoculum in a medium similar to that described by Stadtman and Barker (18). The medium which contained, per liter, NH4Cl, 1.0 g; MgC12.6Hz0, 10 mg; CaC12.2H20, 10 mg; MnSO, .H20, 0.75 mg; Na2Ma04.2H20,0.75 mg; plus CaClp, 50 pM; Na2W04,1.0 pM; NiC12,2.0 pM; ZnS04, 1.0 pM; cuso4, 0.1 PM; and NazSe08, 1.0 p ~ was , autoclaved, cooled, combined with sterile, anaerobic stock solutions of potassium phosphate, pH 6.0 (final concentration, 10 mM) and Na.CH3CO2.3HzO/K-CH&0~ (final concentration, 2.5 g/liter each acetate salt) and made anaerobic by sparging with nitrogen. FeS0,.7Hz0 was added, 10 mg/liter. The
MATERIALS AND METHODS
CHSCOOH -+ CH4 + ‘20%
It has been known since the late 1940s that methane originates predominantly from the methyl group and carbon diox”
”
______.._
* This work was supported by The Gas Research Institute, 8600 West Bryn Mawr Ave., Chicago, IL 60631 (GRI Contract No. 5083260-0967). The casts of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.
3706
Carbon Monoxide Dehydrogenase pH was adjusted to 5.9 with acetic acid and reducing agents were introduced cysteine.HC1,l.O mM and Na,S.9H20, 0.0188 gbiter. As described by Sowers et al. (19), growth on acetatewas poor unless the pH of the culture medium was constantly controlled. Therefore, the pH was maintained between 5.5 and 6.5 by regular additions of glacial acetic acid. Cells were harvested by continuous flow centri~gation in late log phase (as judged by the rate of acetic acid consumption) 3035 days after inoculation. Yields of wet-packed cell paste were in the range of 3-4 g/liter, and overall acid consumption was about 6-8 ml/ g of cells. Cells were frozen in liquid N2 and stored a t -70 "C. Preparation of Crude Cell Homogenate-Frozen cells, 100 g, were thawed in the National Institutes of Health Anaerobic Laboratory (20) in 300 ml of potassium phosphate, 0.40 M; sodium thioglycolate, 4.0 mM; and sodium dithionite, 2.0 mM; pH 7.0; and disrupted using a French pressure cell. After incubation with DNase cellular debris was removed by centrifugation for 20 min a t 16,000 X g. The supernatant was decanted and filtered through a column of glass wool. CO Dehydrogenase-The filLarge-scale I s o ~ of ~ Aggregated ~ n tered crude supernatant, 317 ml, was applied to a column (7.5 X 100 cm) of Sepharose CL-6B-200 equilibrated with potassium phosphate, 0.25 M; sodium thioglycolate, 4.0mM; Na2S204,2.0 mM; pH 6.7. Elution a t 5.0 ml/min was carried out a t room temperature with the same buffer, and fractions, 21.5 mi, constituting the major portion of the high molecular weight CO dehydrogenase peak were pooled. All of the above procedures were carried out in the Anaerobic Laboratory. Discaggregation of CO Dehydrogenase by Dialysis-The high molecular weight CO dehydrogenase fraction from gel filtration on a small scale was dialyzed at room temperature for 43 h in a stoppered flask against 4200 mlof oxygen-free Tris. HCl, 10 mM, and Na2S204, 2 mM, pH 7.9. Large-scale dialysis used in purification of the pooled enzyme from gel filtration (750 ml) was carried out with 40 liters of the same solution. Dialysis understrictly anaerobic conditions was accompanied consistently by an increase in CO dehydrogenase activity in the range of 120-150%. Assay of CO Dehydrogenase-Two methods for determination of CO dehydrogenase have been employed and are described below. ( a ) The standard assay based on methyl viologen reduction was performed under strictly anaerobic conditions at 25 "C in cuvettes containing a solution, 1.0 mf, of methyl viologen, 40 mM, and potassium phosphate, 50 mM, at pH7.0. The mixture was made anaerobic and saturated with carbon monoxide by slow bubbling for 5 min. Sodium dithionite, 1.0 pl, of a 0.4% w/v deoxygenated solution was introduced to obtain a slight degree of methyl viologen reduction. The reaction was started by addition of enzyme (generally 1-5 pl), and absorbance at 578 nm was recorded a t 1-min intervals. A value of 1.16 X lo4 M" cm" (21) was used as the extinction coefficient of reduced methyl viologen a t 600 nm. Under assay conditions the absorbance at 578 nm was 82.7% of the value at 600 nm; therefore, 9.59 X lo3 M" cm" was taken as the extinction coefficient at 578 nm. One unit ofCO dehydrogenase activity is defined under these conditions as that amountrequired to catalyze the reduction of 1.0 pmol/min methyl viologen. (6) The standardassay based on 2~3,5-triphenyltetrazoliumchloride (TTC)' reduction was conducted anaerobically at 25 "C in a mixture, 1.0 ml, of TTC, 2 mM, and potassium phosphate, 0.10 M, a t pH 7.0. Anaerobiosis and saturation with carbon monoxide was achieved in the same manner as described for the methyl viologenbased reaction; however, only 1.0 min of bubbling with COwas employed when the mixture was already anaerobic (e.g. where assays were conducted within the anaerobic room). Reactions were terminated by addition of aerobic acetone, 1.0 ml, and absorbance a t 500 nm was measured. Empirically, a value of4.4 X lo3 M" cm" was determined as the extinction coefficient of the formazan product underthese conditions. Unit activity is defined as 1.0 pmol/min reduction of TTC. Incomparisons made of the two methods account has been taken of the fact that methyl viologen and tetrazolium reductions are le- and 2e- processes, respectively. Potassium phosphate concentration in the range of 10-250 mM had no effect on the rate of methyl viologen reduction. ETPLCAnalysis-Stoppered samples of the undialyzed and dialyzed pools from gel filtration were removedfrom the Anaerobic Laboratory and aliquots (200 pl each) were analyzed on a Pharmacia Superose" 6 fast protein liquid chromatography column using an HP 1090 HPLC. Elution was carried out a t 0.3 ml/min with helium-sparged - - _ _ _ _ _ - ~ __~___________ The abbreviations used are: TTC, 2,3,5-triphenyltetrazoiium chloride; HPLC, high performance liquid chromatography; SDS, sodium dodecyl sulfate.
from M . barkeri
3707
potassium phosphate, 0.20 M,pH 7.0. CO dehydrogenase activity was determined by a m ~ f i c a t i o nof the standard assay employing TTC in fractions, 0.3 ml, collected manually under argon. Analysis of Protein-Except where otherwise noted all protein determinations were performed by the Lowry method (22) on samples which had been exposed to air at0-4 "C for 18-24 h. Bovine serum albumin was used as a standard. A value of approximately 1.55 times more protein was estimated by the Lowry procedure than by the method of Bradford (23) in analyses of purified samples of CO dehy~genase. Amino Acid Analysis-Subunits of CO dehydrogenase were separated by HPLC gel filtration in the presence of SDS. The purified enzyme, 5.04 ml (1.5 mg of protein), was concentrated to a volume of 240 pl using two Centricon 30 (Amicon Corp.) ultrafiltration units. A solution (25 pl) of SDS, 10% w/v, was added, and the mixture was incubated a t 30 "C for 60 min during which the brown color of the enzyme slowly faded to become nearly colorless. The stoppered mixture was then heated for 10 min at 100 "C. Chromatography was carried out on an HP 1090 HPLC using a Pharmacia Superose'" 6 fast protein liquid chromatography column equilibrated and eluted at 0.40 ml/min with lithium phosphate, 0.10 M, SDS, 0.1% w/v, pH 7.0. Fractions, 0.4 ml, were collected and pooled to obtain the major portion of each of the two subunit peaks. The pooled fractions were concentrated by ultrafiltration and rechromatographed. SDS-gel electrophoretic analysis of the final preparations indicated no detectable contamination of either subunit with the other. Dialysis against a solution, 2 liters, of 2-propanol, 25% v/v, trifluoroacetic acid, 0.1% v/v, wasused to remove SDS. The dialyzed precipitated samples were dried and resuspended in 200plof trifluoroacetic acid, 0.1% v/v. Aliquots (20 pl taken to dryness) were hydrolyzed in u a c w in 6 N HCl, 100 pl (Pierce Chemical Co.), at 155 "C for 45 min (24). Samples were also taken and dried for determination of cysteine as cysteic acid following performic acid oxidation (25). After drying and redissolving in guanidine HCl, 6 M,potassium phosphate, 20 mM, pH 6.5, tryptophan was determined by multicomponent analysis of the second derivative ultraviolet spectrum (26). This method also confirmed the values for phenylalanine and tyrosine. Analyses were carried out on a Dionex amino acid analyzer using a microbore column of Dionex resin DC-SA. Amino acid derivatives formed in postcolumn reactions with o-phthalaldehyde were detected fluorometrically, and peak areas were quantitated by a Shimadzu C-R1A integrating recorder. Proline was determined in separate experiments using methods similar to those described by Heinrikson and Meredith (27) for reversed phase HPLC separation of the phenylthiocarbamylated amino acid products from reactions with phenylisothiocyanate. RESULTS AND DISCUSSION
Aggregated Behauior of CO Dehydrogenase in Crude Cell Extracts The gel filtration procedure described by Krzycki and Zeithe of CO dehydrogenase kus (28) as a first step in purification proved to be unsatisfactory since CO dehydrogenase eluted as a broad poorly defined zoneof activity (Fig. 1A). Major protein contaminants eluted along with the trailing edge of this peak, and a low degree of purification resulted. When cell disruption and chromatographywere carried out in buffers of increasing ionic strength, two peaks of CO dehydrogenase were observed (Fig. 1, B-13). Under conditions of high ionic strength (Fig. 1D) about 85% of the total activitywas present in the higher apparent molecularweightform. This major peak of CO dehydrogenase eluted prior to the bulk of protein, and a substantial degree of purificationresulted which was not observed at low ionic strength. To characterize this form of the enzyme the high molecular weight peak fractions of CO dehydrogenase activity from gel filtration (run under conditions given in Fig. 1D)were pooled, dialyzed under strictly anaerobic conditions, andanalyzed by HPLC gel filtration as shown in Fig. 2. A sample of the untreated CO dehydrogenase eluted as a single distinct peak with an apparent molecular weight of approximately 3 X loti (Fig. %if, whereas the dialyzed enzyme was converted completely to aform that emerged as a single peak of much lower molecular weight
3708
Carbon Monoxide Dehydrogenase from M. barkeri 0 A I
t
k 2
&
< I n
8
I
,
1
1
1
1
t
I
B
1.6-
3 4
E
1.4-
1.2-
0.80.61.0
100 150 200 250 300 350 400 450 500
ELUTION VOLUME (mL)
FIG. 1. Anaerobic gel filtration chromatography of GO dehydrogenase (COD€€)on Sepharose CL-GB. Acetate-grown cells of M.barkeri, 5.0 g, weresuspended in 15 ml ofthe followinganaerobic buffers, pH 7.0, which also contained Na2S204, 2 mM, and sodium thioglycolate, 4 mM: A, 10 mM Tris. HCl; B, 30 mM Tris. HC1, 50 mM potassium phosphate; C, 30 mM Tris. HCl, 100 m M potassium phosphate; and D, 250 mM potassium phosphate. Crude cell homogenates were prepared as indicated under “Materialsand Methods” and were applied to a 2.5 X 85-cm column of Sepharose CL-GB equilibrated with the indicated anaerobic buffer, Elution was carried out a t 0.60 ml/min, and 6.6-ml fractions were collected and assayed for CO dehydrogenase using the standard assay with methyl viologen as electron acceptor.
(-161,000) (Fig. 2B). These results support the conclusion that CO dehydrogenase is isolated as an aggregate which is stable at high ionic strength but dissociates under conditions of low ionic strength. Dialysis was found to be a good method for gentle and efficient dissociation of the aggregate. Rapid reassociation does not occur upon restoration of high ionic strength since the HPLC analysis of the dialyzed dissociated enzyme was conducted in 0.2 M potassium phosphate and no aggregated form was detected (Fig. 2B). This suggests that under theseconditions the process of dissociation is not freely reversible.
0.2 o’oO
10
20
30
40
50
60
70
80
90
100
TIME (Min)
FIG. 2. HPLC gel filtration analysis of GO dehydrogenase (CODH).Analysis was made on samples of CO dehydrogenase obtained followinggel filtration of crude extracts according to the conditions described in Fig. ID.Within the anaerobic laboratory an aliquot, 7.0 ml, of the pooled CO dehydrogenase (fractions from the major activity peak) was stoppered and setaside while the remaining material, 52.5 ml, was dialyzed as described under “Materials and Methods.” HPLC analysis of ( A )CO dehydrogenase pool not dialyzed and ( B ) the dialyzed pool was performed outside the anaerobic room as described under “Materials and Methods.”
TABLEI Enzyme purification
in~ of Purification of CO Dehydrogen~ef ~ ~ l o wDissociation the Isolated Protein Aggregate
Crude supernatant 5.23 36.4 11,539 2,209 435 Dialyzed Sepharose 6B 10.4 18.0 7,844 PO01 52.0 75.2 39.1 Phenyl-Sepharose pool 3,914 19.5 HTP Bio-Gel 43.1 133.6 2,607 hydroxylapatite pool “The standard assay with TTC was used as described under “Materials and Methods.” * Protein was measured by the Lowry (22) procedure.
A crude supernatant prepared from 100 g of acetate-grown M . burkeri cells was subjected to gel filtration, and thepooled high molecular weight fraction containing CO dehydrogenase was disaggregated by dialysis (see “Materials and Methods”). The dialyzed sample contained7840 units of CO dehydrogenase at a specific activity which was approximately 3.4 times greater than the crude supernatant (Table I). (i) Phenyl-Sepharose Chromatography-Hydrophobic interaction chromato~aphyof the dialyzed CO dehydrogenase preparation was carried out as the second step inpurification. CO dehydrogenase was among the first proteins to emerge from the column (Fig. 3); thus itappears to be one of the least hydrophobic proteins present in the original aggregate. The pooled fractions comprised about 74% of the total activity recovered. Chromatography on phenyl-Sepharose resulted in
a further increase in specific activity by a factor of about 4.2 (Table I). (ii) Chromatography on Hydroxylapatite-Hydroxylapatite chromatography was used as the final purification step. Two peaks of protein were observed (Fig. 4).The second broad but symmetrical peak contained CO dehydrogenase a t relatively constant specific activity and accounted for 94% of the enzyme appiied. The pooled enzyme was dialyzed overnight against 13 liters of potassium phosphate, 2 mM, sodium dithionite, 2 mM, pH 7.0, and frozen by dripping into liquid nitrogen. Samples thawed under anaerobic conditions exhibited full activity even after 6 months of storage at -70 “C. Freezing in higher ionic strength buffers almost completely inactivated CO dehydrogenase.
Carbon MonoxideDehydrogenase from M.barkeri 3.0
1 .oo
60
t -a -
D
E
F
G
E
>E
3
v
l n -
2
6
0
a
z
= 10Z '
45,000
0.25
_"_
" " " "
0 0
20
60
BO
100
120
140
L
FIG. 3. Chromatography of CO dehydrogenase (CODH) on phenyl-Sepharose.Under strictly anaerobic conditions the dialyzed disaggregated CO dehydrogenase from the Sepharose CL-GB gel filtration step, 750 ml, was adjusted to contain(NH4),S04, 1.5 M, potassium phosphate, 0.10 M, and Na2S204,2 mM, at a final pH of 6.5. The resulting solution (950 ml) was applied a t 3.0 ml/min to a column (5 X 12 cm) of phenyl-Sepharose CL-4B equilibrated with 1.5 M (NH4),S04in the same buffer, pH 6.5. Colloidal material was not bound and was effectively removed at this point. After washing the column with 160 ml of the equilibration buffer, the adsorbed proteins were eluted with a negative linear gradient of (NH4)$04, 1.5 M, and potassium phosphate, 0.1 M, in Na2SZO4,2 mM, 500 ml, and water containing Na2S204,2 mM, 500 ml. Continued elution was carried out with 2 mM Na2S204in H 2 0 after the gradient. Fractions, 11.1 ml, were collected and assayed for CO dehydrogenase activity by the standardTTC reduction method. Protein was measured by the method of Bradford (23). Fractionshaving the highest specific activity, 93-101, were pooled for subsequent chromatography on hydroxylapatite. I 0.70
701
20
40
29,000 24,000
n nn 160
FRACTION NUMBER
0
36,000
"_ ".."~~~.. ,.
40
60
FRACTION NUMBER
FIG. 4. Hydroxylapatite chromatography of CO dehydrogenase (CODH).The pooled enzyme, from phenyl-Sepharose chromatography, 99 ml, was applied directly a t 1.5 ml/min to a hydroxylapatite column, 2.5-cm diameter, containing 25 g of Bio-Gel HTP which had been equilibrated under strictly anaerobic conditions with potassium phosphate, 10 mM, Na2S204,2 mM, pH 7.0. After washing with approximately 25 ml of the same solution, a linear gradient of potassium phosphate, 10 mM, Na2SZO4,2 mM, pH 6.5, 250 ml, was applied at 1.5 ml/min. Fractions, 7.0 ml, were collected and assayed for CO dehydrogenase activity and protein as described in the legend of Fig. 3. The fractions of maximal and constantspecific activity (3542) were pooled.
SDS-Gel Electrophoretic Analysis of Enzyme Preparations The purity of enzyme preparations at each step of the isolation procedure is shown in Fig. 5. Many of the proteins
P
66,000
0.50
a
I
C
J
2.0
0.0-
B
- 0.75
n
5
A
3709
a
20,100
FIG. 5. Analysis of purity by SDS-polyacrylamidegel electrophoresis. The method of Laemmli (29) was used for analysis of the following fractions on a 10% acrylamide slab gel: A, standards: bovine serum albumin, ovalbumin, glyceraldehyde-3-phosphate dehydrogenase, carbonic anhydrase, trypsinogen, and soybean trypsin inhibitor; B, crude supernatant, 48 pg; C, dialyzed pool from Sepharose CL-GB, 28 pg; D,pooled fractions from phenyl-Sepharose, 6.7 pg; E, hydroxylapatite fraction 31, 3.7 pg; F and G, pooled fractions from hydroxylapatite, 3.7 and 7.6 pg, respectively.
present in the crude homogenate (Fig. 5 B ) were separated from the high molecular weight aggregate fraction that contained CO dehydrogenase activity (Fig. 5C) by chromatography on Sepharose CL-GB. After dissociation of the aggregate and chromatography onphenyl-Sepharose, the pooled enzyme fractions contained CO dehydrogenase as the major component (Fig. 5 0 ) . The remaining major contaminant of the enzyme preparation (Fig. 5 E ) was eluted in the first protein peak (Fig. 4) that emerged from the hydroxylapatite column. The second protein peak from this column contained highly purified CO dehydrogenase (Fig. 5, F and G).A contaminant ( M , 29,000) still present in this preparation amounted to only 3% of the total protein as judged by densitometric analysis. By this criterion the enzyme was 92% pure. In earlier experimentspure samples ofCO dehydrogenase were obtained when chromatography on hydroxylapatite preceded phenylSepharose chromatography. However, the present method is preferable since a greater proportion of the enzyme is recovered in pure state. With slight modifications the method has been used three times and found to be reproducible. A s shown in Table I, the procedure provided a reasonably large amount of purified CO dehydrogenase, 19.5mg, with an overall recovery of 23%. Based on purification of 25-fold, CO dehydrogenase is about 4% of the total soluble protein in acetate-grown M. barkeri.
Compositional Studies (i) Subunit Stoichiometry and Molecular Weight Determination of CO Dehydrogenase-As estimated by SDS-gel electrophoretic analysis (29) (average of two determinations) CO dehydrogenase is composed of two dissimilar subunits of molecular weights 84,500 and 19,700 (Fig. 5 , F and G ) . The conclusion that these two subunits are present in equimolar amounts is based on the following evidence. 1)Densitometric analysis of SDS gels indicated a 1.21:l.O (small to large) molar ratio of subunits. 2) Protein determinations by the Lowry method (22) indicated a molar ratio of1.16:l.OO (small to
Carbon Monoxide Dehydrogenasefrom M. barkeri
3710
TABLE I1 large) after the subunitsof SDS-denatured enzyme were sepAmino acid analysis of CO dehydrogenase subunits arated by gel filtration in the presence of SDS. 3) the amino A minimum molecular weight calculated from the amino acid acid compositions of the individual subunits when added together in a 1:l ratio matched closely the amino acid com- composition of the 0-subunit is 84,700 and for the a-subunitis 19,500. @ Subunit (M. = 84,500) 01 Subunit (M. = 19,700) position found from analysis of the intactenzyme. Amino Estimation of the native molecular weight of disaggregated Number Of Nearest Nearest acid residues/ integer CO dehydrogenase as 161,000 was made by HPLC gel filtrainteger chain chain tion analysis (Fig. 2B). However, this value does not correAsx 70.6 71 22.0 22 spond to ana@or ana2P2combination of 84,500- and 19,70031.7 Thr 32 13.0 13 dalton subunits. To address further the question of the oliSer 32.3 32 8.1 8 gomeric structure ofCO dehydrogenase two other types of 85.4 85 Glx 12.1 12 72 72.0 13.9 14 experiments were carried out. Purified CO dehydrogenase Gb 81.7 Ala 82 18.8 19 when subjected to gradient gel pore size exclusion electropho39.4 Val 39 10.4 10 resis under nondissociating conditions appeared as a single 21.0 21 Met 2.2 2 discrete band on the stained gel. A value of 205,000 was 50.7 51 Ile 13.1 13 65.6 66 Leu estimated for the molecular weight relative to standard pro18 17.9 29.3 29 7.5 8 TYr teins (data not shown). This value suggests an a2P2subunit 15.5 15 Phe 6 6.0 composition. Further evidence in favor of an a2P2structure 12.5 His 13 3.4 3 was obtained from chemical cross-linking experiments. SDS63.1 63 17.2 17 LYS gel electrophoretic analysisof cross-linked CO dehydrogenase 34.9 35 4.9 5 Arg showed the presence of two additional proteinbands (Fig. 6B) Pro" 39.1 39 7.1 7 as contrasted to the unreacted control (Fig. 6A). The molec1.3 23.1 23 1 Cysb ular weights of the larger proteins in the treatedsample were 7.0 7 1.0 1 Trp' 106,000 and 206,000, indicative of CUPand a2P2cross-linked 179 Total 775 . . structures.Therefore, the molecular weight obtained from By phenylisothiocyanate analysis of amino acids. HPLC gel filtration of pure CO dehydrogenase appears tobe As cysteic acid after performic acid oxidation. By second derivative ultraviolet spectroscopy. an underestimation of the true value, perhaps due to some degree of interaction of CO dehydrogenase with the column matrix. The same a2P2subunit combination in CO dehydroloo 100 genase from M. barkeri has been suggested by Krzycki and 90 Zeikus (28) based upon gel filtration experiments.Our present
G:f$:y
~
80
A B
~~
I -
70 -
\
-\
60 -
-
50
40 -
a
=
19,700--
FIG. 6. Analysis of chemically cross-linkedCO dehydrogenase by SDS-polyacrylamide gel electrophoresis. Under anaerobic conditions purified CO dehydrogenase, 8.1 pg in 25 pl,was brought to pH 9.0 by addition of a buffered solution, 2.5 pl, of triethanolamine HCl, 0.50 M. To this was added 1.0 pl of 50 mM disuccinimidyl suberate in acetone. Precipitation of a small amount of disuccinimidyl suberate was noted a t this point. After 15 min a t room temperature, the mixture was cooled to 4 "C. Following incubation overnight (20 h) a t 4 "C, 25 pl of a solution containing SDS, 2% w/v; glycerol, 20% v/v; dithiothreitol, 80 mM; EDTA, 2.0 mM; pyronin Y, 10 pg/ml; and Tris.HC1, 20 mM; a t pH 7.8 was added. Excess disuccinimidyl suberate was solubilized at thispoint and the sample was heated for 2 min a t 100 "C. SDS electrophoresis was carried out on 6% acrylamide gels (0.5 X 6.0 cm) according to the method of Fairbanks et al. (30). Gels were stained with Coomassie Brilliant Blue G-250 in 3.5% HClO, (31,32). For gel A unreacted purified CO dehydrogenase, 3.2 pg, was applied, and for gel B the products of reaction of the enzyme with disuccinimidyl suberate. The proteins used as molecular weight standards (Bio-Rad's SDS-polyacrylamide gel electrophoresis high and low molecular weight calibration standard kits) were myosin (200,000), 0-galactosidase (116,250), phosphorylase b (92,500), bovine serum albumin (66,200), ovalbumin (45,000), carbonic anhydrase (31,000), soybean trypsininhibitor (21,500), and lysozyme (14,400).
30
-
20
-
10
1
0
I
I
I
I
I
I
60 80 100 TEMPERATURE (C) FIG. 7. Heat denaturationof pure CO dehydrogenase. Samples (200 pl) of purified CO dehydrogenase (3.6 units/ml) in 2.5 mM potassium phosphate, 0.5 mM Na2S20r, pH7.0, were heated in 15 X 75-mm stoppered test tubes for 10 min a t 24, 39,62,80, and 104 "C. After cooling to room temperature (24 "C),aliquots (5 pl) were withdrawn and assayed by the standardmethod using TTC. A parallel set of samples was heated in the presence of carbon monoxide. The procedure was identical except that carbon monoxide was bubbled slowly through the samples for 1min prior to sealing the tubes. Both procedures were conducted entirely within the Anaerobic Laboratory.
20
40
results provide additional and strong evidence for a native a2P2oligomeric structure. (ii) Metal Ion Content-CO dehydrogenases isolated from methanebacteria (28, 33) and clostridia (34-37) contain nickel and iron-sulfur centers. Plasmaemission spectroscopic analyses showed that our purified preparations contained 1.3 0.3 ( n= 4) nickel and 15.6 5.6 (n = 5 ) atoms of iron/mol
*
*
Monoxide Carbon Dehydrogenase
from M . barkeri
3711
of enzyme ( M , 208,400). The calculated values were based on extremely sensitive to oxygen. When samples were exposed proteindeterminations by the Lowry method (22). These to air, most of the activity was lost in less than 1 min, and complete inactivation. values would be somewhat higher if based on estimation of longer incubationtimesresultedin protein by the procedureof Bradford (23)(see "Materials and Cyanide is a potent inhibitor. Partial protection againstcyaMethods"). The amountof nickel and iron and the enzymaticnide inactivation by carbon monoxide suggests that cyanide activity were not decreased when purified CO dehydrogenase may react at the active site. These findings are similar to was dialyzed anaerobically overnight against buffer contain- those of others for clostridial CO dehydrogenase (34) and ing 1.0 mM EDTA, pH 7.0. But, zinc, which was present a t purified (28) and partially purified CO dehydrogenase (39) 14.0 atoms/mol, was lowered to a value of 1.1.Ragsdale et al. from Methanosarcina. It appears from these observations that (37) have noted variable amounts of zinc in the enzyme from cyanide may react at the active site. Acetobacterium woodii, and the significance of this metal in A variety of compoundshave beentestedaspotential CO dehydrogenases remains t o be established. Although mo- inhibitors and/or activators by inclusion in the CO dehydrolybdenumwas reportedin CO dehydrogenase fromPseugenase assay mixture. A typical free radical scavenger, hydomonas carboxydouorans (38) and cobalt was found in the droxyurea (5 mM), caused very little inhibition, buta marked partially purified enzyme from Methurwsarcinu thermophila inactivation was observed by hydroxylamine (1 mM), also a (39), our preparations contained neither of these metals in radicalscavenger. These results provide no evidencefora significant quantities. These results are similar to the findings radical mechanism, but do suggest that a functional group of Ragsdale et al. (36) who noted the absenceof molybdenum existsonthe enzyme which isessential for activityand and cobalt in theenzyme from C. thermoaceticum. susceptible to nucleophilic attack by hydroxylamine. (iii) Amino Acid Composition of CO Dehydrogenase SubNone of the metal ions tested (MgSO,, CoC12, NiC12, and units-Based on amino acid analyses of the individual CO ZnS04)stimulated CO dehydrogenaseactivity. Enzymatic dehydrogenase subunits thatwe separated by gel filtration in activity was not decreased by addition of the nickel chelator, the presence of SDS (Table II), there area total of approxi- dimethylglyoxime, and as noted above, dialysis against 1.0 mately 775 and 179 amino acidresidues inthe (3 and a mM EDTA failed to decrease the iron and nickel content of subunits, respectively. Each a(3 heterodimer contains about the enzyme or its activity. These results are similar to the 136 basic amino acids (taken as the sum of arginine, lysine, behavior of CO dehydrogenasefrom C. thermoaceticum (8) and histidine) and190 residuesof aspartic and glutamic acids and indicate that catalytically important metals are tightly including their amides. At pH 7.0 the overall charge on the bound by the enzyme. protein is negative since it binds DEAE-cellulose to (data not Aldehydes such as formaldehyde and acetaldehyde did not shown).Therefore,thesedata suggest that fewer than 54 inactivate CO dehydrogenase at concentrations as high as 20 asparagineplusglutamine residues existper a(3 unit. The mM. However, the two structurally analogouscompounds finding of a large number of free aspartic and glutamic acid glyoxaldehyde and 2,3-butanedionewere found to be strongly residues isconsistentwiththe relativelyhigh affinity for hydroxylapatite noted earlier. The presence of a small number inhibitory. In contrast, Ragsdale and Wood (8) found that of aromatic residues compared with the aliphatic hydrophobic although 14C0 exchange into the carbonyl group of acetylamino acid content appears to be in accord with the moder- CoA by CO dehydrogenase from C. thermoaceticum was inately low affinity for phenyl-Sepharose. It is of interest to hibited, the rate of CO oxidation was unchanged by this class note that the small subunit containsonly a single residue of of compounds. The M. barkeri CO dehydrogenase was almost tryptophan. Since a fairly large number of glycyl, alanyl, and completely inactivated by 2 mM glyoxaldehyde and totallyby prolyl residues were found, it seemsreasonable that a-helical 2,3-butanedione at 11.5 mM. Preincubation of CO dehydrocontent may be restricted in some regions of the protein. This genase with glyoxaldehyde had no effect on the initial rateof may be related to the high degree of thermostability of the reaction when either methyl viologen or CO was added to enzyme. A ferredoxin-like function of the small subunitdoes start the reaction. Only when both CO and methyl viologen not seem likely in view of its low content of cysteine. This were present was inactivation by glyoxaldehyde observed. suggests thatiron-sulfurcentersare located onthe large This suggests that inactivation is dependent on enzyme turnover and that reaction of glyoxaldehyde occurs with adistinct subunit which contains most of the cysteine residues. enzyme intermediate formedonly during catalysis. Since Catalytic Properties glyoxaldehyde and 2,3-butanedione are classic arginine-mod(i) Electron Acceptors-Several different compounds have ifying reagents, it is tempting to speculate that an essential been tested as electron acceptors in the CO dehydrogenase- arginyl residue may become exposed or have increased reactivity in the intermediateenzyme form. catalyzed oxidation of carbon monoxide. Under saturating (iu) Thermal Stability-Aliquots of purified CO dehydrogenconditionsbothmethyl viologen andtriphenyltetrazolium chloride give nearly equivalent rates of CO oxidation. How- ase were heat treated as described in the legend to Fig. 7. A ever, the apparentK, for methyl viologen is 7.1 mM, whereas high degree of thermostability is exhibited by the enzyme even after heatingat 80 "C T T C exhibited a Km(app) of
