- Email: [email protected]
Malonyl-CoA Decarboxylase from Streptomyces erythreus: Purification, properties, and possible role in the production of erythromycin
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 229, No. 2, March, pp. 426-439, 1984
Malonyl-CoA Decarboxylase from Streptomyces erythreus: Purification, Properties, and Possible Role in the Production of Erythromycin’ A. R. HUNAITI Institute
AND
P. E. KOLATTUKUDY’
of Biological Chemistry and Biochemistry/Biophysics Washington State University, Pullman, Washington
Received July 13, 1983; and in revised form November
Program, 99164 16, 1983
Malonyl-CoA decarboxylase was purified (800-fold) from an erythromycin-producing strain of Streptomyces erythreus using DEAE-cellulose, Sephadex G-100, SP-Sephadex, and gel filtration with Sephadex G-75. The molecular weight of the native enzyme was 93,000 as determined by gel filtration and the subunit molecular weight was 45,000 as estimated by sodium dodecyl sulfate-polyacrylamide electrophoresis, suggesting an CY~ subunit composition for the native enzyme. Evidence is presented that during the purification procedure and storage a proteolytic cleavage occurred resulting in the formation of 30- and 15-kDa peptides. The enzyme showed a pH optimum of about 5.0 whereas the vertebrate enzyme showed an optimum at alkaline pH. The enzyme decarboxylated malonyl-CoA with a K, of 143 PM and Vof 250 nmol min-’ rng-‘. For the decarboxylation of methylmalonyl-CoA this enzyme showed the opposite stereospecificity to that shown by vertebrate enzyme; the (R) isomer was decarboxylated at 3% of the rate observed with malonyl-CoA while the (S) isomer was not a substrate. Neither avidin nor biotin affected the rate of malonyl-CoA decarboxylation, suggesting that biotin is not involved in catalysis. Acetyl-CoA and free CoA were found to be competitive inhibitors. Propionyl-CoA, butyryl-CoA, succinyl-CoA, and methylmalonyl-Cob showed little inhibition, and neither thiol-directed reagents nor chelating agents inhibited the enzyme. High ionic strength and sulfate ions caused reversible inhibition of the enzymatic activity. Under two different cultural conditions the time course of appearance of malonyl-CoA decarboxylase was determined by measuring the enzyme activity and the level of the enzyme protein by an immunological method using rabbit antibodies prepared against the enzyme. In both cases the increase and decrease in the decarboxylase correlated with the rate of production of erythromycin, suggesting a possible role for this enzyme in the antibiotic production.
Streptomyces erythreus produces the multiple methyl-branched macrocyclic lactone antibiotic erythromycin, which is presumably synthesized from one molecule of propionyl-CoA and six of methylmalonyl-CoA (1). The enzymology of the biosynthesis of erythromycin has not been elucidated. An acyl-CoA carboxylase which 1 Scientific Paper No. 6597, Project 2001, College of Agriculture Research Center, Washington State University, Pullman, Wash. 99164. ‘To whom correspondence should be addressed. 0003-9861/84 $3.00 Copyright 0 1984by AcademicPress,Inc. All rights of reproductionin any form reserved.
carboxylates propionyl-CoA at a lo-fold higher rate than acetyl-CoA has been recently purified, and this enzyme might provide methylmalonyl-CoA for the antibiotic synthesis (2). The biosynthetic reactions involved in the formation of the macrocyclic lactones have been assumed to be analogous to those involved in fatty acid biosynthesis. The demonstration that vertebrate fatty acid synthase has the inherent capacity to generate multiple methylbranched fatty acids from methylmalonylCoA in the absence of malonyl-CoA (3) 426
MALONYL-CoA
DECARBOXYLASE
lends support to this view, although the products generated by 5’. erythreus fatty acid synthase from methylmalonyl-CoA have not been characterized. In the uropygial gland of goose, which produces 2,4,6,8-tetramethyldecanoic acid as a major component, the fatty acid synthase catalyzes the synthesis of this acid from methylmalonyl-CoA while the tissue-specific presence of a very active malonyl-CoA decarboxylase insures that only methylmalonyl-CoA and acetyl-CoA would be available for fatty acid synthesis (4). The presence of a fatty acid synthase-type enzyme which catalyzes the synthesis of multimethyl-branched mycocerosic acids has been also recently demonstrated in a cellfree system from Mycobacterium tuberculosis (5), and a malonyl-CoA decarboxylase which might also play a role in the production of such acids was isolated from this organism (6). In both cases the presence of malonyl-CoA severely interferes with the production of the multiple methylbranched products and the decarboxylase probably prevents such interference. If the multimethyl-branched aglycone of erythromycin is produced by a fatty acid synthase-like enzyme system as previously observed with the multimethyl-branched fatty acids in the vertebrate (4) and bacterial (5) systems noted above, a malonylCoA decarboxylase could play a role similar to that suggested for the other two systems. In this paper we report the discovery of such a decarboxylase in an erythromycin-producing S. erythreus and the purification of this enzyme. We demonstrate that the properties of this enzyme are dramatically different from those reported for the enzyme from vertebrates, and the present enzyme shows opposite stereospecificity to that shown by the vertebrate enzyme. We also present evidence that this decarboxylase could play a role in the production of erythromycin. MATERIALS
AND
FROM
Streptmyces
erythreus
427
munodiffusion, and complete and incomplete adjuvant (Freund’s) were purchased from Difco Laboratories. [3-‘*C]Malonyl-CoA and [3-i4C]methylmalonyl-CoA were prepared enzymatically from acetyl-CoA and propionyl-CoA, respectively, using NaHi4C03 and other components of the CO,-fixation assay. Purified goose uropygial gland carboxylase (7) was used to prepare malonyl-CoA, and purified M. tuberculosis carboxylase (8) was used for the preparation of [3“C]methylmalonyl-CoA. The products of carboxylation were purified as described previously (9). [1,4“C]Succinyl-CoA was prepared from [1,4-‘*C]succinic anhydride and the product was purified by ion-exchange chromatography as described before (10). (R,S)-[methyl-i4C]Methylmalonyl-CoA was synthesized from synthetic (R,S)-[methyl-‘*C]methylmalonic acidpurchasedfromNewEnglandNuclearasindicated elsewhere (5). All other chemicals and materials were purchased from Sigma Chemical Company.
Culture Conditions S. e-rythrem CAM obtained from Abott Laboratories was maintained on nutrient agar (Difco Laboratories). The slant cultures were used to inoculate the vegetative growth medium consisting of the following components in amounts shown per liter and the pH was adjusted to 7.0 with 10% KOH: sucrose (15.0 g), Difco peptone (5.0 g), yeast extract (2.5 g), and Larginine (0.5 g). A number of 50-ml cultures were grown for 3 days in the vegetative medium in 125ml flasks at 32°C in a rotary shaker operating at 250 rpm. Aliquots (10 ml) of the resulting vegetative inoculum were then transferred into 2800-ml Fernbach flasks containing 1 liter of fresh medium of the same composition and incubated at 32°C on a rotary shaker for 4 days.
Preparation
of Crude Extract
Cells were harvested by centrifugation at 12,OOOg for 15 min, washed twice with 50 mM Tris-HCI buffer, pH 7.5, containing 1 mM DTEa and 5 mM EDTA. The washed cells were suspended in the same buffer containing 1 mM PMSF (cell to buffer ratio was 1:2 v/ v). All further steps were carried out at 0-4°C. The cells were ruptured by a French pressure cell operating at 14,000 psi. The broken cells were centrifuged at 3’7,OOOgfor 30 min and the supernatant was the source of the enzyme.
METHODS
Materials NaH”COa, [1,3-‘“Clmalonic acid, [1,3-i4C]malonylCoA, and [1,4-‘4C]succinic anhydride were from New England Nuclear Corporation. Agar (Nobel) for im-
3 Abbreviations used: DTE, dithioerythritol; PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate; EGTA, ethylene glycol bis(@aminoethyl ether)-N,N,N’,N’-tetraacetic acid.
428
HUNAITI
AND
DEAE-cellulose step. The supernatant from the above step was made into a slurry with DEAE-cellulose which had been previously equilibrated with 50 mM Tris-HCl buffer, pH 7.5, containing 1 mM DTE and 5 mM EDTA (25-50 mg of protein/ml settled bed volume). The resulting slurry was stirred for 40 min and then filtered through a Biichner funnel using Whatman No. 1 filter paper. After the DEAE-cellulose on the filter was washed with 2 vol of the same buffer, the enzyme was eluted with 1 column volume of 0.1 M potassium phosphate buffer, pH 7.0, containing 1 mM DTE, 5 mM EDTA, and 1 mM PMSF. The eluted enzyme was concentrated by ultrafiltration using an Amicon PM-10 membrane. Gel jiltration 012 Sephadex G-100. The enzyme solution obtained from the above step was subjected to gel filtration with a Sephadex G-100 column (2.2 X 100 cm) preequilibrated with 50 mM Tris-HCl buffer, pH 7.5, containing 1 mM DTE and 5 mM EDTA. Proteins were eluted with the same buffer, collecting 4-ml fractions/l5 min. The fractions containing malonylCoA decarboxylase activity were pooled and dialyzed overnight against 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. The flocculent precipitate was removed by centrifugation at 20,OOOgfor 15 min. Cation-excharqe chromatography with SP-S~P~KMIJZ The enzyme solution obtained from previous step was applied to a SP-Sephadex column (1.8 X 16 cm) equilibrated previously with 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. After washing the column with the same buffer, the enzyme was eluted with 20 mM citrate-phosphate buffer, pH 6.0, containing 5% glycerol, 1 mM DTE, and 5 mM EDTA. The flow rate was 4 ml/15 min when the enzyme was applied to the column and 1 ml/min during washing and elution. The fractions (-4 ml each) containing malonyl-CoA decarboxylase activity were pooled and dialyzed against 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. Small amounts of insoluble material present were removed by centrifugation at 20,OOOgfor 15 min. GeljZtration 012SepWx G-75. The enzyme solution from the above step was concentrated by ultrafiltration using an Amicon PM-10 membrane and the enzyme solution (0.4 ml containing 0.8 mg protein) was applied to a Sephadex G-75 column (1.2 X 117 cm) preequilibrated with 20 mM citrate-phosphate buffer, pH 5.0, containing 1 mM DTE and 5 mM EDTA. The enzyme was eluted with the same buffer at a flow rate of 0.4 ml/min and 1.5-ml fractions were collected. Fractions representing the enzyme activity peaks at the void volume (-40 ml) and at 49 ml were separately pooled to give two enzyme fractions. The solutions were concentrated by ultrafiltration using an Amicon PM-10 membrane. Purified enzyme refers to the preparation eluting at the void and containing only 45-kDa peptide as indicated in a later section
KOLATTUKUDY
Enx yme Assay The enzyme activity was assayed by measuring the amount of i4C02 liberated from [3-i4C]malonyl-CoA essentially as described previously (11). The reaction mixture contained 10 mM citrate-phosphate buffer, (sp act pH 5.0,l mM DTE, 0.3 mM [3-i4C]malonyl-CoA 0.35 Ci/mol), and enzyme in a total volume of 0.1 ml. Enzyme activity is expressed as nanomoles of COa released per milligram protein per minute. Protein was measured by the method of Lowry et al. (12) using bovine serum albumin as standard. Radioactivity determination. Radioactivity was determined by liquid scintillation spectrometry in a Packard Tri-Carb UGO-CD equipped with dpm converter with 15 ml of a 0.4% solution of Omniflour in toluene containing 30% ethanol as the scintillation fluid. The aliquots of HPLC fractions were assayed with 10 ml of ScintiVerse (Fischer) as the scintillation fluid. All counting was done with efficiency >90% and ~3% standard deviation.
Electrophoresis Polyacrylamide gel electrophoresis in the presence of 0.1% SDS was performed according to the procedure of Laemmli (13). Electrophoresis was done in a slab gel (8 X 8 cm) 1.0 mm thick, with a 7.5% resolving gel and a 2.5% stacking gel. Usually 10 pg of the enzyme was heated in a boiling-water bath for 5 min with 50 ~1 of Laemmli’s SDS buffer (13) and applied to the gel. Electrophoresis was performed at room temperature at constant voltage over a period of 1.52 h. Protein bands were fixed with a 10% aqueous solution of trichloroacetic acid for 10 min and stained for 15 min with 0.05% Coomassie brilliant blue R-250 dissolved in methanol/acetic acid/water (40/30/50, v/v/v). The gel was destained with 7.5% acetic acid and 5% methanol. Analytical gel electrophoresis of the enzyme (1520 pg) under nondenaturing conditions was carried out at 4°C in 7.5% polyacrylamide gel with 0.375 M Tris-glycine buffer, pH 8.3, at 2 mA per tube for 3 h. One gel was cut into small slices (2 mm in length) and an identical gel was stained with Coomassie brilliant blue for 15 h. The slices were macerated in 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA (200 ~1 per slice), kept overnight at 4°C and then assayed for enzymatic activity as described above. Molecular-weight determination. The molecular weight of the native enzyme was determined by gelpermeation chromatography on a calibrated Sephadex G-150 column (1.4 X 110 cm). The protein standards were y-globulin (150,000), bovine serum albumin (68,000), ovalbumin (45,000), and carbonic anhydrase (30,000). The molecular weight of the subunit was determined by SDS-gel electrophoresis in a slab gel as described
MALONYL-CoA
DECARBOXYLASE
by Laemmli (13) with the following protein standards: phosphorylase b (92,000), bovine serum albumin (45,000), carbonic anhydrase (6S,OW, ovalbumin (30,000), soybean trypsin inhibitor (20,100), and cylactalbumin (14,400).
Preparation
of Antisera
Since a method for separation of the intact protein from nicked protein was not available at the time, we purified the intact subunits by SDS-gel electrophoresis. The upper protein band (45 kDa) was cut out, macerated, and placed in an electrophoresis tube containing4% polyacrylamide in the bottom. A small dialysis bag was attached to the bottom of the tube and the protein was eluted from the gel by applying a constant current of 3 mA per tube for 15 h. The purified subunit (175 fig) thus obtained was thoroughly mixed with Freund’s complete adjuvant (1:l v/v), and the resulting emulsion (0.8 ml) was injected into the hind foot pads of a white New Zealand rabbit. Two weeks after the first injection, 125 pg of protein emulsified with incomplete Freund’s adjuvant was injected into the same rabbit. Two weeks after the second injection, the rabbit was bled by heart puncture and the antiserum was collected. The titers of rabbit antidecarboxylase were determined by immunotitration. Aliquots (10 pg) of the enzyme were incubated with varying concentrations of antiserum adjusted to 15 ~1 with control rabbit serum in a total volume of 100 ~1. After 2 h at 4”C, the incubation mixtures were centrifuged at 25009 for 10 min and the enzymatic activity in the supernatant was assayed as described above.
Im,munodiffusion Ouchterlony double diffusion was performed on microscope slides with 2% Nobel agar in 100 mM Verona1 buffer, pH 8.5, containing 0.01% thimerosal (14). After 18-24 h the nonagglutinated proteins were removed and the immunoprecipitin lines were visualized as before (15). The level of decarboxylase was measured by a radial immunodiffusion method as described before (15).
Amino Acid Analysis Lyophilized aliquots of purified decarboxylase were hydrolyzed in 6 N HCI (Pierce Chemical Company) under vacuum for 20 h at 110°C. Cysteine and methionine were analyzed after oxidation with performic acid (16) prior to hydrolysis. Hydrolysates were analyzed with a Beckman Model 121 MB automatic amino acid analyzer.
IdentiJication Products
of Decarboxylation
Enzyme assays were done in a total volume of 0.1 ml with 2.5 pg of the purified enzyme (SP-Sephadex
FROM
Streptomyces
erythreus
429
enzyme) under conditions indicated above for 1 h, except that the labeled substrates were [1,3‘%]malonyl-CoA (sp act 40.8 Ci/mol) or [methyl‘%]methylmalonyl-CoA (sp act 56 Ci/mol). The reaction was stopped by adding 10 ~1 of perchloric acid and the precipitated protein was removed by centrifugation. The supernatants were subjected to descending paper chromatography with isobutyric acid/ (57/35/8, v/v/v) as de1 N ammonium acetate/water scribed elsewhere (9). [1-‘%JAcetyl-CoA was identified as the product of decarboxylation by cochromatography with authentic unlabeled acetyl-CoA. The product of decarboxylation of [methyl-‘4C]methylmalonyl-CoA was identified by descending paper chromatography as described elsewhere (17) with nbutanol/acetone/acetic acid/l N ammonium acetate (g/3/2/6, v/v/v/v).
Stereospeci@ity Preparation of (R)-[methyl-‘4C]methylmalonyl-CoA. (R)-[methyl-“C]Methylmalonyl-CoA was obtained by the removal of the (S) isomer from synthetic (R,S)[methyZ-“C]methy1ma1ony1-CoA by repeated treatment with purified malonyl-CoA decarboxylase from the uropygial gland of goose (17). The reaction mixture contained 100 pM (R,S)-[methyl-‘“C]methylmalonylCoA (56 Ci/mol), 100 pM Tris-HCl, pH 7.0, and 0.5 mM DTE in a total volume of 50 ~1. After incubating the mixture at 30°C for 3 h, the reaction was terminated by the addition of 10 ~1 of perchloric acid, the precipitated protein was removed by centrifugation, and the supernatant was subjected to HPLC analysis using a polyanion Sl HR 5/5 column (Pharmacia Fine Chemicals) which was equilibrated with 0.1 M potassium phosphate, pH 4.5. (R)-[methyZ‘%]Methylmalonyl-CoA was eluted with a 0.1-0.4 M gradient of potassium phosphate, pH 4.5, at a flow rate of 1.5 ml/min. The fractions containing (R)[methyl-“C]methylmalonyl-CoA were pooled, lyophilized, and desalted with a Bio-Gel P-2 column (1.4 X 90 cm) with 0.35 mM HCl as the solvent, and the product was kept frozen in a minimum volume of 0.35 mM HCI. Preparation of (S)-methylmalonyl-CoA. (S)-[3-14C]Methylmalonyl-CoA was enzymatically synthesized from propionyl-CoA using NaH14C03 and purified carboxylase from M. tuberculosis var. bovis BCG as described (8). (S)-[methyl-‘4C]Methylmalonyl-CoA was prepared by the removal of the (R) isomer from synthetic (R,S)-[methy1-‘4C]methylmalonyl-CoA by repeated treatment with S. w-ythreus malonyl-CoA decarboxylase. The reaction mixture contained 10 pM (R,S)-[methyI-‘4C]methylmalonyl-CoA and 100 fiM citrate-phosphate buffer, pH 6.0, in a total volume of 50 ~1. After incubating the mixture at 30°C for 3 h, the reaction was terminated and the product was isolated as described above. The purity of the (S)-[1-“C]methylmalonyl-CoA
430
HUNAITI
AND
preparation was examined by HPLC with 50 ItIM tetrabutylammonium phosphate, pH 5.5, in 50% methanol (18) using a p-Bondapak Cl8 column (0.39 x 30 cm, Water Associates). Decarboxylation of labeled methylmakmyl-CoA by the decarboxylases from S. evythreus and goose w-e pygial glands. Two nanomoles of (R)-[methyli4C]methylmalonyl-CoA (sp act 56 Ci/mol), 100 nmol of citrate-phosphate buffer, pH 6.0, and 28 fig of purified decarboxylase from S. erythreus in a total volume of 100 ~1 were incubated at 30°C for 1 h and the incubation was continued for a total of 3 h with hourly additions of fresh enzyme (28 yg). The reaction was terminated by addition of 10 ~1 of perchloric acid, the precipitated proteins were removed, and the product was analyzed by HPLC as above. Three nanomoles of (S)-[methyl-“C]methylmalonylCoA, 2 pmoi of Tris-HCl, pH 7.5, 25 nmol of DTE, and 5 fig of decarboxylase from goose in a total volume of 100 ~1 were incubated at 30°C for 1 h and incubation was continued for 3 h with hourly additions of fresh enzyme (5 bug). The reaction products were analyzed as described above. of malonyl-CoA decarTime course of appearance boxylose and erythromycin production. The time course was determined in two different media, the vegetative medium in which growth occurred and in 1.5% sucrose containing 0.2% L-alanine, a medium which allowed a high level of antibiotic production without growth. In the former case l-liter cultures were used as described above, and each day 10 ml of culture was centrifuged in a graduated tube at 1200g for 20 min and the packed cell volume was used as a measure of growth. In the latter case 50 ml of 3-day-old cultures in the vegetative medium were transferred to 1 liter of the sucrose-alanine medium. At desired time intervals erythromycin was extracted from the broth with ethyl acetate after removal of the mycelium by centrifugation and adjusting the pH to 9.5, and the antibiotic was measured calorimetrically as described (19). The activity of malonyl-CoA decarboxylase was measured in crude extracts as described above. Malonyl-CoA decarboxylase production was measured by a radial immunodiffusion method as described elsewhere (15). Standard curves were prepared by using serial dilution of the purified decarboxylase. RESULTS
Pur$cation of Malonyl-CoA Decarboxylase Crude extract of S. erythreus showed malonyl-CoA decarboxylase activity. Preliminary studies on the time course of appearance of the decarboxylase activity showed maximal levels at late-stationary phase of growth (4th day). Therefore we
KOLATTUKUDY
used 4-day-old cultures of S. erythreus for all of the subsequent experiments. The specific activity of malonyl-CoA decarboxylase in the crude extract ranged between 0.2 and 0.8 nmol decarboxylated min-’ rng-‘. Purification of the enzyme by some of the classical protein-fractionation methods proved to be difficult because of the sensitivity of the enzyme to high ionic strength and inhibition by sulfate ions as described in a later section. However, the crude extract could be fractionated with DEAE-cellulose which completely retained the enzymatic activity at pH 7.5, and the enzyme could be eluted by 0.1 M potassium phosphate buffer, pH 7.0, with 60-70% recovery and a 5- to 6-fold increase in specific activity. Gel filtration of this enzyme preparation on Sephadex G-100 revealed that the enzyme activity was contained in the second major protein peak (Fig. 1). This step resulted in over 24-fold purification with a 62% recovery of the activity. Dialysis of this enzyme preparation against 10 mM citrate-phosphate buffer, pH 4.5, and the removal of the flocculent precipitate resulted in 50% recovery of the activity with a 40-fold increase in specific activity. SP-Sephadex at pH 4.5 retained all the decarboxylase activity while the bulk of the protein was not retained (Fig. 1). The enzyme could be eluted by the 10 mM buffer at pH 6.0, but this procedure caused precipitation and denaturation of the enzyme, resulting in only 3-4% recovery of the enzymatic activity. However, with the use of 20 mM buffer containing 5% glycerol, the recovery of enzymatic activity increased up to nearly 20% with about a 270-fold purification. Electrophoresis of the resulting protein showed three major bands at 45, 30, and 15 kDa. The relative intensities of these bands varied from preparation to preparation with a decrease in the 45k band accompanied by increases in the other two bands or vice versa. Upon gel filtration of this enzyme preparation on Sephadex G-75, two peaks of enzymatic activities were observed. SDS-electrophoresis of the first peak of protein showed only a single band with a molecular weight of 45k, whereas electrophoresis of the second peak of enzymatic
MALONYL-CoA
DECARBOXYLASE
FROM
Streptmyces
A
G-100 20-
431
erythreus
PROTEIN DECARSOXYLASE -8
l.0
-4
E c
,-\
2
.\------,’
’
N 2 w
\ \ \ -_
0
*0
40
BO
I
SO
DECARSOXYLASE
FRACTION
NUMBER
FIG. 1. Purification of malonyl-CoA decarboxylase from S. erythreus. (A) Sephadex G-100 gel filtration of the enzyme obtained from the DEAE-cellulose step; (B) SP-Sephadex chromatography of the decarboxylase. The application of 20 mM citrate-phosphate, pH 6.0, containing 5% glycerol is indicated. The other experimental details are given in the text.
activity showed two bands at 30k and 15k. (A minor band at 38k was also sometimes observed.) About two-thirds of the applied activity was found in the first peak and the rest in the second peak. These results strongly suggested that the purification procedure resulted in a proteolytic clip in TABLE
this enzyme. All our attempts to prevent proteolysis by the inclusion of protease inhibitors such as PMSF, EDTA, and soybean trypsin inhibitor failed. In any case these purification procedures resulted in about 800-fold purification with an overall recovery of about 12% (Table I). The enI
PURIFICATION OF MALONYL-COA DECARBOXYLASE FROM S. erythreus
Step Crude extract DEAE-Cellulose Sephadex G-100 Dialyzed pH 4.5 SP-Sephadex Sephadex G-75 Note. Experimental
Protein (mg)
Total activity (nmol/min)
2400 300 61 35 1.5 0.35 details
1790 1239 1098 910 300 210
are in the text.
Specific activity (nmol/min/mg) 0.75 4.13 18 26.2 200 600
Yield (%I 100 69 61 51 18 12
Purification (fold) 1 5 24 35 267 800
432
HUNAITI
AND
zyme was relatively unstable; it lost about 50% of its activity upon freezing and thawing. However, it was relatively stable if stored at 4°C in 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. With increasing periods of storage the 45k band intensity decreased with a corresponding increase in the intensities of the bands at 30k and 15k, suggesting that a proteolytic clip similar to that occurring during purification also took place during storage of the purified enzyme. Unless otherwise specified, the purified enzyme containing only 45-kDa peptides was used for characterization of the enzyme. Molecular Weight and Subunit Composition Polyacrylamide gel electrophoresis (‘7.5%) of the enzyme recovered from the SP-Sephadex step showed a single band and all of the decarboxylase activity was contained in this band (Fig. 2), strongly suggesting that the enzyme was near homogeneity. Gel filtration on a Sephadex G150 column with proteins of known mo-
4 DISTANCE
KOLATTUKUDY
lecular weight showed that the native decarboxylase had a molecular weight of 93,000 (Fig. 3). Polyacrylamide disc gel electrophoresis of the enzyme after treatment with SDS and 2-mercaptoethanol showed a single band (Fig. 4). The molecular weight of the protomer, as determined by a linear plot of log molecular weight vs mobility of standard proteins, was 45,000. This observation, together with the molecular weight of the native enzyme, strongly suggests that the present decarboxylase has an (Yesubunit composition. Amino Acid Composition Amino acid composition of the present enzyme is shown in Table II, and for comparison the composition of malonyl-CoA decarboxylases from avian sources is included. Immunological
Properties
Ouchterlony double-diffusion analysis showed that rabbit antibody prepared against the 45k protein gave a single precipitin line (Fig. 5, inset) when purified enzyme or the SP-Sephadex enzyme was used as the antigen. The antibody also gave a single precipitin line with both the 30k and 15k proteins isolated by SDS-electrophoresis. Complete fusion of the lines was observed with the 30k protein but not with the 15k protein. These results suggest that all three peptides contained common antigenic determinants, with the larger two being immunologically very similar, if not identical. These immunological relationships are consistent with the conclusion that the 30k and 15k proteins were derived from the 45k protein by proteolytic cleavage. The enzymatic activity was progressively severely inhibited by increasing amounts of the antiserum, and finally virtually complete inhibition of the enzyme activity was observed (Fig. 5).
8 (cm)
FIG. 2. Polyacrylamide gel electrophoresis of the decarboxylase (25 pg) obtained from the SP-Sephadex step on ‘7.5% gel and the distribution of malonyl-CoA decarboxylase activity in the gel. The experimental details are given in the text.
Catalytic
Properties
The activity of the purified enzyme preparation was linear with protein concentration at least up to 100 pg/ml, and the reaction rate was linear up to at least 20
MALONYL-CoA
DECARBOXYLASE
FROM
TABLE AMINO
ACID
433
erqthreus
Streptmyces
II
COMPOSITION
OF MALONYL-COA
DECARBOXYLASE
Amino acid
(mol%)
Avian” (mol%)
Asx Thr Ser Glx Pro GIY Ala Val Ile Leu Tw Phe LYS His Arg CYS Met
8.1 4.3 16.4 13.6 3.4 17.9 8.3 3.6 1.9 3.6 0.8 1.3 2.9 2.4 2.8 1.5 8.8
6.6 4.1 7.0 14.0 4.4 6.8 7.6 6.4 4.0 12.7 2.1 3.6 5.8 2.4 7.3 1.8 1.5
S. erythrews
45k
30k
15k
FIG. 3. SDS-electrophoresis of malonyl-CoA decarboxylase. The experimental details are given in the text. (A) The enzyme after SP-Sephadex column chromatography; (B) the protein from the first peak of enzymatic activity obtained by Sephadex G-75 gel filtration shown in Table I.
Decarboxylase
(93K)
a The values obtained for the enzyme isolated from three avian sources (7, 20) were averaged.
min of incubation. Therefore, all subsequent experiments were done within such linear ranges for both protein and time. The rate of decarboxylation increased sharply as the pH increased from 3 to 5 and further increases in pH caused a decline in activity, resulting in a rather sharp pH optimum at around 5.0 (Fig. 6). For routine assays 10 mM citrate-phosphate, pH 5.0, was used. Chromatofocusing indicated that the pI of malonyl-CoA decarboxylase was 4.4. Eflect of Substrate Concentrations
x
FRACTION
NUMBER 25
50
Ve-Votml) FIG. 4. Determination of the molecular weight of the malonyl-CoA decarboxylase (from the Sephadex G-75 step) by gel filtration on calibrated Sephadex G-150 column. Inset, the chromatography of the purified decarboxylase on G-150 column. The protein standards used are (A) y-globulin (150k); (B) bovine serum albumin (68k); (C) ovalbumin (45k); and (D) carbonic anhydrase (30k).
A typical Michaelis-Menten-type substrate saturation pattern was observed with malonyl-CoA as the substrate (Fig. 7). From linear double-reciprocal plots the apparent Km and Vvalues were calculated to be 143 pM and 250 nmol mini’ mg-‘, respectively. Inhibitors
Malonyl-CoA decarboxylase from S. erby both acetyl-CoA
ythreus was inhibited
HUNAITI
AND
KOLATTUKUDY
0
FIG. 5. The effect of rabbit antiserum prepared against purified malonyl-CoA decarboxylase on the enzymatic activity. The inset shows Ouchterlony double-diffusion analysis. The center well contained 10 gl of rabbit antiserum and the outer wells, starting from top and going clockwise, contained 0.5, 1.0, 1.5, and 2 pg of purified decarboxylase. Other experimental details are given in the text.
0 A
A
I o-
’ 3
4
5
6
7
6
6
PH FIG. 6. Effect of pH on the rate of decarboxylation of malonyl-CoA by the purified enzyme. The reaction mixtures contained 10 wg of the purified enzyme, and the assay conditions were as described under Materials and Methods.
and free CoA. Double-reciprocal plots of the results obtained with different concentrations of acetyl-CoA and free CoA clearly showed that this inhibitor was competitive with malonyl-CoA (Fig. 8). Replot of the slope vs concentration of the inhibitors (insets) showed that the inhibition was parabolic. Other short-chain acyl-CoAs (Table III) showed less inhibition. Neither avidin (1 mg/ml) nor biotin (1 mM) had any effect on the decarboxylation rate, and, therefore, biotin is probably not involved in catalysis. In contrast to the previously examined malonyl-CoA decarboxylases, the present decarboxylase showed no inhibition by thiol-directed reagents such as p-hydroxymercuribenzoate (up to 0.5 mM) and N-ethylmaleimide (up to 1.0 mM). Neither iodoacetamide nor phenylmethylsulfonyl fluoride (each up to 5 mM) showed any inhibition. Chelating
MALONYL-CoA
l/[Sl 80 [~~AL~NYL-COA]
DECARBOXYLASE
x10*
160
240
FROM
Streptmyces
435
erythreus
decarboxylate (S)-methylmalonyl-CoA whereas the (R) isomer could be nearly completely decarboxylated with this enzyme (Table IV). Upon prolonged incubadecarboxylase with tion of S. erythreus (R,S)-methylmalonyl-CoA only 50% of the methylmalonyl-CoA was decarboxylated. That the remaining methylmalonyl-CoA was exclusively the (S) isomer was shown by the observation that the (S)-specific decarboxylase from the goose uropygial gland
(~tdl)
FIG. 7. Effect of substrate concentration on the rate of decarboxylation and double-reciprocal plot for malonyl-CoA. The reaction mixture contained 10 fig of the purified enzyme, and assay conditions were as described under Materials and Methods.
A 1.0 ACETYLCOA LkM 75 .
E 0 10.5 v)
agents such as EDTA, EGTA (1.0 mM), and bipyridyl (0.5 mM) showed no inhibition, suggesting that heavy metals are not involved in catalysis. Effect of Salts on Decarboxylase
LlIL 0
25
50
75 /
Activity
The effect of several salts on the enzyme activity was investigated, and from the results obtained (Fig. 9) it appeared that the enzyme was inhibited by high ionic strength. SO: ions severely inhibited the enzyme even at low concentration (5 mM). Removal of salts by dialysis or gel filtration resulted in full recovery of the enzymatic activity. Stereospeci&ity It is known that the malonyl-CoA decarboxylase from the uropygial gland of goose is specific for the decarboxylation of (S)-methylmalonyl-CoA, and shows no activity with (R)-methylmalonyl-CoA (17). Therefore, we used this enzyme as a stereospecific reagent to remove the (S) isomer from the (R,S) mixture and thus generated stereochemically pure (R)-methylmalonyl-CoA. It is also known that the acyl-CoA carboxylase of AI. tuberculosis generates specifically (S)-methylmalonylCoA (8). Therefore, we used this enzyme to generate (S)-methylmalonyl-CoA. It was found that the present malonyl-CoA decarboxylase from S. erythreus could not
50
&
0
25 0
.
0
2
11 [MALONYL-COAI
4
6 x
I o2 PM
FIG. 8. Double-reciprocal plots showing competitive inhibition of purified malonyl-CoA decarboxylase by acetyl-CoA (A) and free CoA (B). The concentrations of the inhibitors are indicated and the effects of the concentration of the inhibitor on the slopes are shown in the insets.
436
HUNAITI TABLE
AND
KOLATTUKUDY
III
EFFECT OF INHIBITORS ON MALONYL-COA DECARBOXYLASE FROM S. erythreus”
Concentration (mM)
Inhibitor
Relative rate (% of control)
Coenzyme A
0.5 1.0
36 30
Acetyl-CoA
0.5 1.0
38 31
Propionyl-CoA
0.5 1.0
60 60
Butyryl-CoA
0.5 1.0
56 47
Methylmalonyl-CoA
0.5 1.0
61 48
Succinyl-CoA
0.5 1.0
91 41
SALT
( mbl
)
FIG. 9. Effect of salts on the enzymatic activity of the purified malonyl-CoA decarboxylase from S. erythreus. Ammonium sulfate gave results identical to those shown for sodium sulfate. The indicated concentrations of salts were present in the reaction mixtures
served with malonyl-CoA when both substrates were used at 0.3 mM. Succinyl-CoA was not decarboxylated, neither was maionic acid.
a Enzyme (10 pg) in 90 gl of 10 mM citrate-phosphate buffer, pH 5.0, was preincubated with inhibitors at the indicated concentration at room temperature for 15 min, and reactions were initiated by the addition of 10 pl of substrate.
Product IdentiJication completely decarboxylated it. These results show that the present decarboxylase, unlike the goose decarboxylase, shows absolute stereospecificity for (R)-methylmalonyl-CoA.
and Stoichimetry
To be certain that the release of CO2 from malonyl-CoA involved decarboxylation of the substrate with the formation of acetylCoA, rather than some other type of reactions involving release of COa, the products generated from malonyl-CoA were identified and quantitated. HPLC and paper chromatographic analyses of the products showed that the amount of acetyl-CoA generated from [1,3-‘4C]malonyl-CoA was
Substrate Speci$city Malonyl-CoA was highly preferred over (R)-methylmalonyl-CoA, the latter being decarboxylated at about 3% of the rate obTABLE
IV
STEREOSPECIFICITY IN THE DECARBOXYLATION OF METH~MALON~-COA BY S. ergthrews MALONYL-COA DECARBOX~LASE
Source of decarboxylase
Substrate (S)-[l-‘4C]Methylmalonyl-CoA (R,S)-[methyl-i4C]Methylmalonyl-CoA (R)-[meth&‘4C]Methylmalonyl-CoA (S)-[methyl-14C]Methylmalonyl-CoA Note. Experimental
details
are in the text.
S. e-rythreus S. erythreua S. eqthrevs Goose uropygial
gland
Percentage decarboxylation 0 50 90 88
MALONYL-CoA
DECARBOXYLASE
equivalent to the amount of CO2 released. The product generated from methylmalonyl-CoA was identified as propionyl-CoA by paper chromatography and HPLC. Time Course of Appearance of MalonylCoA Decarboxylase and Erythrmycin Production The specific activity of malonyl-CoA decarboxylase was dependent on age of the culture. The specific activity was at a maximum just prior to the onset of erythromycin production (Fig. 10). To test whether the increase in specific activity of the decarboxylase reflected an increase in the level of the enzyme protein or some type of activation, the decarboxylase level was measured immunologically using rabbit antiserum prepared against the purified decarboxylase. The area of the precipitant rings generated in the radial immunodiffusion assay was directly proportional to the amount of the antigen. The results obtained by this method were in agreement with those obtained by measurement of the enzyme activity, and showed that the amount of decarboxylase increased about lo-fold just prior to the onset of erythromycin production. Both methods showed
FROM
2
3
4
5
erythreus
437
that the decarboxylase level decreased as the antibiotic production leveled off. Such a correlation between the decarboxylase level and antibiotic production was also found in a sucrose-alanine medium in which production of erythromycin proceeds without significant growth (Fig. 10). DISCUSSION
The purification of malonyl-CoA decarboxylase from a microorganism, first reported in the present paper, enabled us to compare this bacterial enzyme with that isolated from avian and mammalian sources (20, 21, 22). The subunit composition of the 5’. erythreus decarboxylase is in sharp contrast to the decarboxylases isolated from avian and mammalian tissues; the present enzyme is a dimer of 45kDa protomers whereas the mammalian and avian enzymes are tetramers of 50kDa peptides. Since the mature protomer from the animal mitochondrial enzyme was a 47-kDa peptide, the size of the protomers of the decarboxylases can be considered similar. The amino acid composition of the S. erythreus decarboxylase is also quite different from that of the previously examined decarboxylases (20). The present de-
00
1
Streptomyces
i
6
DAYS
FIG. 10. Time course of the appearance of malonyl-CoA decarboxylase and production of erythromycin by S. erythreus in a vegetative growth medium (A) and in a sucrose-alanine medium (B). The experimental details are given in the text.
438
HUNAITI
AND
carboxylase contained considerably lower amounts of hydrophobic and basic amino acids than those found in the avian enzymes, whereas the contents of the acidic residues are quite similar. Antibodies raised against S. erythreus decarboxylase neither inhibited nor cross-reacted with the avian enzyme, and the antibodies prepared against the decarboxylase from the uropygial gland of goose neither affected the activity nor cross-reacted with the S. erythreus enzyme, suggesting that the present enzyme is immunologically quite different from the avian enzyme. Catalytic properties of S. erythreus decarboxylase were quite different from those of the avian and mammalian decarboxylases (20-22). The present enzyme showed an acidic pH optimum (5.0) compared to the basis pH optimum (9.0) observed with the avian and mammalian enzymes. The present enzyme also showed a higher Km and lower V than the previously examined decarboxylases. Effects of the various inhibitors on the S. erythreus enzyme were significantly different from those on the mammalian and avian malonyl-CoA decarboxylases. The present enzyme was not significantly inhibited by p-hydroxymercuribenzoate or other thiol-directed reagents whereas the other decarboxylases were severely inhibited by such reagents. CoA and acetyl-CoA inhibited the enzyme from 5’. erythreus whereas they had little effect on the avian or mammalian enzyme. The properties of the present enzyme are different from those of a partially purified decarboxylase previously obtained from hf. tuberculosis and Pseudomonas jluorescens, although some similarities were also apparent. S. erythreus decarboxylase is a dimer of 45,000 while those of M. tuberculosis and P. JluorescerzS were found to be monomeric with molecular weights of 45,000 and 56,000, respectively. Effect of the various inhibitors also showed significant differences among the three enzymes. p-Hydroxymercuribenzoate showed little inhibition with the enzyme from S. erythreus or P. fluorescens whereas it severely inhibited the enzyme from M. tuberculosis. Succinyl-CoA inhibited the enzyme from M. tuberculosis whereas it showed little in-
KOLATTUKUDY
hibition with the enzyme from S. erythreus and P. jlwrescens. Acetyl-CoA and CoA inhibited the enzyme from M. tuberculosis and S. erythreus whereas little inhibition was noted with enzyme from P.JEuorescens. Finally, the high sensitivity of the enzyme from S. erythreus to high ionic strength and sulfate ions revealed that it was quite different from the previously examined decarboxylases. The present enzyme shows a novel stereospecificity. It decarboxylates (R)methylmalonyl-CoA, but not (S)-methylmalonyl-CoA. This stereospecificity is opposite to that observed with goose malonylCoA decarboxylase (17) which accepts only (S)-methylmalonyl-CoA as the substrate. The decarboxylase preparation from M. tuberculosis also decarboxylates (S)-methylmalonyl-CoA (unpublished) whereas even the crude extract of S. erythreus decarboxylates only the (R) isomer, indicating the absence of an (S)-specific decarboxylase and a methylmalonyl-CoA racemase in this organism. The possible functional significance of this novel stereospecificity is not clear. It is possible that the present enzyme is involved in producing propionyl-CoA from the (R)-methylmalonyl-CoA generated from succinylCoA, which would be derived from the carbohydrate provided as a major carbon source in the medium. Another possibility is that the observed stereospecificity has no real physiological significance and that the enzyme in vivo decarboxylates only malonyl-CoA. The relationship between the time course of appearance of the decarboxylase and that of erythromycin synthesis suggested that the enzyme might play a role in the antibiotic production. In the vegetative medium the cells would require malonylCoA for the synthesis of the n-fatty acids needed for membrane synthesis during the period of rapid growth of the organism. As the growth was nearly complete the decarboxylase level increased. Also, in the sucrose-alanine medium the decarboxylase level dramatically increased just prior to the rapid production of the antibiotic. In both cases the role of the decarboxylase might be to insure that the antibiotic syn-
MALONYL-CoA
DECARBOXYLASE
thase would not be adversely affected by the presence of malonyl-CoA. Such a role had previously been suggested in the case of production of multiple methyl-branched fatty acids by the uropygial gland and M. tuberculosis. However, it is not known whether the antibiotic synthase is adversely affected by the presence of malonylCoA. In any case, the observed relationship between the time course of appearance of the decarboxylase and that of erythromycin production is consistent with the conclusion that this enzyme plays a role in the antibiotic production.
FROM
D. L., AND KOLATTUKUDY, P. E. (1983) J. Biol. Chem. 258, 2979-2985. 6. KIM, Y. S., AND KOLATNKUDY, P. E. (1979) Arch. Biochem. Biophys. 196,543-551. 7. BUCKNER, J. S., KOLA~KUDY, P. E., AND POULOSE, Biophys. 177,539A. J. (1976) Arch B&hem 551. 8. RAINWATER, 9. BUCKNER,
(Corcoran, J. W., ed.), Vol. 4, pp. 133-174, Springer-Verlag, Berlin/Heidelberg/New York. 2. HUNAITI, A. R., AND KOLA’ITUKUDY, P. E. (1982) Arch. B&hem. Biophys. 216,362-371. 3. BUCKNER, J. S., KOLATTUKUDY, P. E., ANDROGERS, L. (1978) Arch. Biochem. Biophys. 186,152-163. 4. BUCKNER,
J. S., AND KOLATTUKUDY,
P. E. (1976)
in Chemistry and Biochemistry of Natural Waxes (Kolattukudy, P. E., ed.), pp. 147200, Elsevier, Amsterdam/New York.
J. S., AND KOLA~UKUDY,
10. GREGOLIN,
P. E. (1975)
14,1768-1773.
C., RYDER,
J. Biol
P. E. (1982)
151, 905-911.
Biochemistry
E., ANDLANE,
M. D. (1968)
Chem. 243,4227-4235.
P. E. (1982) 14, 609-614. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., AND RANDALL, R. J. (1951) J. BioL Chem. 193, 265-275. LAEMMLI, V. K. (1970) Nature (ikndolz) 277, 680685. OUCHTERLONY, 0. (1953) Acta Pathol. Microbial Scam! 32,231-240. LIN, T. S., AND KOLATTUKUDY, P. E. (1978) J. Bacteriol. 133,942-951. MOORE, S. (1962) J. Biol. Chem 238,235-237. KIM, Y. S., AND KOLATTUKUDY, P. E. (1980) J. Biol. Chem. 255, 686-689.
11. RAINWATER,
D. L., AND KOLAITUKUDY,
Int. J Biochem
12.
13. 14.
15.
17.
1. CORCORAN, J. W. (1981) in Antibiotics
D. L., AND KOLATXJKUDY,
J. Bacterial.
16.
REFERENCES
439
erythreus
5. RAINWATER,
ACKNOWLEDGMENTS Dr. Andrew Bognar conducted some of the preliminary experiments. We thank Dr. A. J. Poulose and Dr. D. Rainwater for helpful discussion. This work was supported by Grant GM18278 from the Institute of General Medical Sciences of the U. S. Public Health Service.
Streptomyces
18. RAINWATER, D. L., AND KOLATTUKUDY, P. E. (1981) Arch. Biochem. Biophys. 213,372-383. 19. FORD, J. M., PRESCOIT, G. C., HINMAN, J. W., AND CAROIV, E. L. (1953) Anal. Chem 25,1195-1197. 20. KIM, Y. S., AND KOLAIXIKUDY, P. E. (1978) Arch. B&hem. Biophys. 190,585-597. 21. KIM, Y. S., AND KOLA~UKUDY, P. E. (1978) Arch. B&hem. Biophys. 190,234-246. 22. KIM,
Y.
S.,
Biochim.
AND
KOLATTUKUDY,
P.
Biophys. Actu 531, 187-196.
E.
(1978)
Malonyl-CoA Decarboxylase from Streptomyces erythreus: Purification, Properties, and Possible Role in the Production of Erythromycin’ A. R. HUNAITI Institute
AND
P. E. KOLATTUKUDY’
of Biological Chemistry and Biochemistry/Biophysics Washington State University, Pullman, Washington
Received July 13, 1983; and in revised form November
Program, 99164 16, 1983
Malonyl-CoA decarboxylase was purified (800-fold) from an erythromycin-producing strain of Streptomyces erythreus using DEAE-cellulose, Sephadex G-100, SP-Sephadex, and gel filtration with Sephadex G-75. The molecular weight of the native enzyme was 93,000 as determined by gel filtration and the subunit molecular weight was 45,000 as estimated by sodium dodecyl sulfate-polyacrylamide electrophoresis, suggesting an CY~ subunit composition for the native enzyme. Evidence is presented that during the purification procedure and storage a proteolytic cleavage occurred resulting in the formation of 30- and 15-kDa peptides. The enzyme showed a pH optimum of about 5.0 whereas the vertebrate enzyme showed an optimum at alkaline pH. The enzyme decarboxylated malonyl-CoA with a K, of 143 PM and Vof 250 nmol min-’ rng-‘. For the decarboxylation of methylmalonyl-CoA this enzyme showed the opposite stereospecificity to that shown by vertebrate enzyme; the (R) isomer was decarboxylated at 3% of the rate observed with malonyl-CoA while the (S) isomer was not a substrate. Neither avidin nor biotin affected the rate of malonyl-CoA decarboxylation, suggesting that biotin is not involved in catalysis. Acetyl-CoA and free CoA were found to be competitive inhibitors. Propionyl-CoA, butyryl-CoA, succinyl-CoA, and methylmalonyl-Cob showed little inhibition, and neither thiol-directed reagents nor chelating agents inhibited the enzyme. High ionic strength and sulfate ions caused reversible inhibition of the enzymatic activity. Under two different cultural conditions the time course of appearance of malonyl-CoA decarboxylase was determined by measuring the enzyme activity and the level of the enzyme protein by an immunological method using rabbit antibodies prepared against the enzyme. In both cases the increase and decrease in the decarboxylase correlated with the rate of production of erythromycin, suggesting a possible role for this enzyme in the antibiotic production.
Streptomyces erythreus produces the multiple methyl-branched macrocyclic lactone antibiotic erythromycin, which is presumably synthesized from one molecule of propionyl-CoA and six of methylmalonyl-CoA (1). The enzymology of the biosynthesis of erythromycin has not been elucidated. An acyl-CoA carboxylase which 1 Scientific Paper No. 6597, Project 2001, College of Agriculture Research Center, Washington State University, Pullman, Wash. 99164. ‘To whom correspondence should be addressed. 0003-9861/84 $3.00 Copyright 0 1984by AcademicPress,Inc. All rights of reproductionin any form reserved.
carboxylates propionyl-CoA at a lo-fold higher rate than acetyl-CoA has been recently purified, and this enzyme might provide methylmalonyl-CoA for the antibiotic synthesis (2). The biosynthetic reactions involved in the formation of the macrocyclic lactones have been assumed to be analogous to those involved in fatty acid biosynthesis. The demonstration that vertebrate fatty acid synthase has the inherent capacity to generate multiple methylbranched fatty acids from methylmalonylCoA in the absence of malonyl-CoA (3) 426
MALONYL-CoA
DECARBOXYLASE
lends support to this view, although the products generated by 5’. erythreus fatty acid synthase from methylmalonyl-CoA have not been characterized. In the uropygial gland of goose, which produces 2,4,6,8-tetramethyldecanoic acid as a major component, the fatty acid synthase catalyzes the synthesis of this acid from methylmalonyl-CoA while the tissue-specific presence of a very active malonyl-CoA decarboxylase insures that only methylmalonyl-CoA and acetyl-CoA would be available for fatty acid synthesis (4). The presence of a fatty acid synthase-type enzyme which catalyzes the synthesis of multimethyl-branched mycocerosic acids has been also recently demonstrated in a cellfree system from Mycobacterium tuberculosis (5), and a malonyl-CoA decarboxylase which might also play a role in the production of such acids was isolated from this organism (6). In both cases the presence of malonyl-CoA severely interferes with the production of the multiple methylbranched products and the decarboxylase probably prevents such interference. If the multimethyl-branched aglycone of erythromycin is produced by a fatty acid synthase-like enzyme system as previously observed with the multimethyl-branched fatty acids in the vertebrate (4) and bacterial (5) systems noted above, a malonylCoA decarboxylase could play a role similar to that suggested for the other two systems. In this paper we report the discovery of such a decarboxylase in an erythromycin-producing S. erythreus and the purification of this enzyme. We demonstrate that the properties of this enzyme are dramatically different from those reported for the enzyme from vertebrates, and the present enzyme shows opposite stereospecificity to that shown by the vertebrate enzyme. We also present evidence that this decarboxylase could play a role in the production of erythromycin. MATERIALS
AND
FROM
Streptmyces
erythreus
427
munodiffusion, and complete and incomplete adjuvant (Freund’s) were purchased from Difco Laboratories. [3-‘*C]Malonyl-CoA and [3-i4C]methylmalonyl-CoA were prepared enzymatically from acetyl-CoA and propionyl-CoA, respectively, using NaHi4C03 and other components of the CO,-fixation assay. Purified goose uropygial gland carboxylase (7) was used to prepare malonyl-CoA, and purified M. tuberculosis carboxylase (8) was used for the preparation of [3“C]methylmalonyl-CoA. The products of carboxylation were purified as described previously (9). [1,4“C]Succinyl-CoA was prepared from [1,4-‘*C]succinic anhydride and the product was purified by ion-exchange chromatography as described before (10). (R,S)-[methyl-i4C]Methylmalonyl-CoA was synthesized from synthetic (R,S)-[methyl-‘*C]methylmalonic acidpurchasedfromNewEnglandNuclearasindicated elsewhere (5). All other chemicals and materials were purchased from Sigma Chemical Company.
Culture Conditions S. e-rythrem CAM obtained from Abott Laboratories was maintained on nutrient agar (Difco Laboratories). The slant cultures were used to inoculate the vegetative growth medium consisting of the following components in amounts shown per liter and the pH was adjusted to 7.0 with 10% KOH: sucrose (15.0 g), Difco peptone (5.0 g), yeast extract (2.5 g), and Larginine (0.5 g). A number of 50-ml cultures were grown for 3 days in the vegetative medium in 125ml flasks at 32°C in a rotary shaker operating at 250 rpm. Aliquots (10 ml) of the resulting vegetative inoculum were then transferred into 2800-ml Fernbach flasks containing 1 liter of fresh medium of the same composition and incubated at 32°C on a rotary shaker for 4 days.
Preparation
of Crude Extract
Cells were harvested by centrifugation at 12,OOOg for 15 min, washed twice with 50 mM Tris-HCI buffer, pH 7.5, containing 1 mM DTEa and 5 mM EDTA. The washed cells were suspended in the same buffer containing 1 mM PMSF (cell to buffer ratio was 1:2 v/ v). All further steps were carried out at 0-4°C. The cells were ruptured by a French pressure cell operating at 14,000 psi. The broken cells were centrifuged at 3’7,OOOgfor 30 min and the supernatant was the source of the enzyme.
METHODS
Materials NaH”COa, [1,3-‘“Clmalonic acid, [1,3-i4C]malonylCoA, and [1,4-‘4C]succinic anhydride were from New England Nuclear Corporation. Agar (Nobel) for im-
3 Abbreviations used: DTE, dithioerythritol; PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate; EGTA, ethylene glycol bis(@aminoethyl ether)-N,N,N’,N’-tetraacetic acid.
428
HUNAITI
AND
DEAE-cellulose step. The supernatant from the above step was made into a slurry with DEAE-cellulose which had been previously equilibrated with 50 mM Tris-HCl buffer, pH 7.5, containing 1 mM DTE and 5 mM EDTA (25-50 mg of protein/ml settled bed volume). The resulting slurry was stirred for 40 min and then filtered through a Biichner funnel using Whatman No. 1 filter paper. After the DEAE-cellulose on the filter was washed with 2 vol of the same buffer, the enzyme was eluted with 1 column volume of 0.1 M potassium phosphate buffer, pH 7.0, containing 1 mM DTE, 5 mM EDTA, and 1 mM PMSF. The eluted enzyme was concentrated by ultrafiltration using an Amicon PM-10 membrane. Gel jiltration 012 Sephadex G-100. The enzyme solution obtained from the above step was subjected to gel filtration with a Sephadex G-100 column (2.2 X 100 cm) preequilibrated with 50 mM Tris-HCl buffer, pH 7.5, containing 1 mM DTE and 5 mM EDTA. Proteins were eluted with the same buffer, collecting 4-ml fractions/l5 min. The fractions containing malonylCoA decarboxylase activity were pooled and dialyzed overnight against 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. The flocculent precipitate was removed by centrifugation at 20,OOOgfor 15 min. Cation-excharqe chromatography with SP-S~P~KMIJZ The enzyme solution obtained from previous step was applied to a SP-Sephadex column (1.8 X 16 cm) equilibrated previously with 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. After washing the column with the same buffer, the enzyme was eluted with 20 mM citrate-phosphate buffer, pH 6.0, containing 5% glycerol, 1 mM DTE, and 5 mM EDTA. The flow rate was 4 ml/15 min when the enzyme was applied to the column and 1 ml/min during washing and elution. The fractions (-4 ml each) containing malonyl-CoA decarboxylase activity were pooled and dialyzed against 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. Small amounts of insoluble material present were removed by centrifugation at 20,OOOgfor 15 min. GeljZtration 012SepWx G-75. The enzyme solution from the above step was concentrated by ultrafiltration using an Amicon PM-10 membrane and the enzyme solution (0.4 ml containing 0.8 mg protein) was applied to a Sephadex G-75 column (1.2 X 117 cm) preequilibrated with 20 mM citrate-phosphate buffer, pH 5.0, containing 1 mM DTE and 5 mM EDTA. The enzyme was eluted with the same buffer at a flow rate of 0.4 ml/min and 1.5-ml fractions were collected. Fractions representing the enzyme activity peaks at the void volume (-40 ml) and at 49 ml were separately pooled to give two enzyme fractions. The solutions were concentrated by ultrafiltration using an Amicon PM-10 membrane. Purified enzyme refers to the preparation eluting at the void and containing only 45-kDa peptide as indicated in a later section
KOLATTUKUDY
Enx yme Assay The enzyme activity was assayed by measuring the amount of i4C02 liberated from [3-i4C]malonyl-CoA essentially as described previously (11). The reaction mixture contained 10 mM citrate-phosphate buffer, (sp act pH 5.0,l mM DTE, 0.3 mM [3-i4C]malonyl-CoA 0.35 Ci/mol), and enzyme in a total volume of 0.1 ml. Enzyme activity is expressed as nanomoles of COa released per milligram protein per minute. Protein was measured by the method of Lowry et al. (12) using bovine serum albumin as standard. Radioactivity determination. Radioactivity was determined by liquid scintillation spectrometry in a Packard Tri-Carb UGO-CD equipped with dpm converter with 15 ml of a 0.4% solution of Omniflour in toluene containing 30% ethanol as the scintillation fluid. The aliquots of HPLC fractions were assayed with 10 ml of ScintiVerse (Fischer) as the scintillation fluid. All counting was done with efficiency >90% and ~3% standard deviation.
Electrophoresis Polyacrylamide gel electrophoresis in the presence of 0.1% SDS was performed according to the procedure of Laemmli (13). Electrophoresis was done in a slab gel (8 X 8 cm) 1.0 mm thick, with a 7.5% resolving gel and a 2.5% stacking gel. Usually 10 pg of the enzyme was heated in a boiling-water bath for 5 min with 50 ~1 of Laemmli’s SDS buffer (13) and applied to the gel. Electrophoresis was performed at room temperature at constant voltage over a period of 1.52 h. Protein bands were fixed with a 10% aqueous solution of trichloroacetic acid for 10 min and stained for 15 min with 0.05% Coomassie brilliant blue R-250 dissolved in methanol/acetic acid/water (40/30/50, v/v/v). The gel was destained with 7.5% acetic acid and 5% methanol. Analytical gel electrophoresis of the enzyme (1520 pg) under nondenaturing conditions was carried out at 4°C in 7.5% polyacrylamide gel with 0.375 M Tris-glycine buffer, pH 8.3, at 2 mA per tube for 3 h. One gel was cut into small slices (2 mm in length) and an identical gel was stained with Coomassie brilliant blue for 15 h. The slices were macerated in 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA (200 ~1 per slice), kept overnight at 4°C and then assayed for enzymatic activity as described above. Molecular-weight determination. The molecular weight of the native enzyme was determined by gelpermeation chromatography on a calibrated Sephadex G-150 column (1.4 X 110 cm). The protein standards were y-globulin (150,000), bovine serum albumin (68,000), ovalbumin (45,000), and carbonic anhydrase (30,000). The molecular weight of the subunit was determined by SDS-gel electrophoresis in a slab gel as described
MALONYL-CoA
DECARBOXYLASE
by Laemmli (13) with the following protein standards: phosphorylase b (92,000), bovine serum albumin (45,000), carbonic anhydrase (6S,OW, ovalbumin (30,000), soybean trypsin inhibitor (20,100), and cylactalbumin (14,400).
Preparation
of Antisera
Since a method for separation of the intact protein from nicked protein was not available at the time, we purified the intact subunits by SDS-gel electrophoresis. The upper protein band (45 kDa) was cut out, macerated, and placed in an electrophoresis tube containing4% polyacrylamide in the bottom. A small dialysis bag was attached to the bottom of the tube and the protein was eluted from the gel by applying a constant current of 3 mA per tube for 15 h. The purified subunit (175 fig) thus obtained was thoroughly mixed with Freund’s complete adjuvant (1:l v/v), and the resulting emulsion (0.8 ml) was injected into the hind foot pads of a white New Zealand rabbit. Two weeks after the first injection, 125 pg of protein emulsified with incomplete Freund’s adjuvant was injected into the same rabbit. Two weeks after the second injection, the rabbit was bled by heart puncture and the antiserum was collected. The titers of rabbit antidecarboxylase were determined by immunotitration. Aliquots (10 pg) of the enzyme were incubated with varying concentrations of antiserum adjusted to 15 ~1 with control rabbit serum in a total volume of 100 ~1. After 2 h at 4”C, the incubation mixtures were centrifuged at 25009 for 10 min and the enzymatic activity in the supernatant was assayed as described above.
Im,munodiffusion Ouchterlony double diffusion was performed on microscope slides with 2% Nobel agar in 100 mM Verona1 buffer, pH 8.5, containing 0.01% thimerosal (14). After 18-24 h the nonagglutinated proteins were removed and the immunoprecipitin lines were visualized as before (15). The level of decarboxylase was measured by a radial immunodiffusion method as described before (15).
Amino Acid Analysis Lyophilized aliquots of purified decarboxylase were hydrolyzed in 6 N HCI (Pierce Chemical Company) under vacuum for 20 h at 110°C. Cysteine and methionine were analyzed after oxidation with performic acid (16) prior to hydrolysis. Hydrolysates were analyzed with a Beckman Model 121 MB automatic amino acid analyzer.
IdentiJication Products
of Decarboxylation
Enzyme assays were done in a total volume of 0.1 ml with 2.5 pg of the purified enzyme (SP-Sephadex
FROM
Streptomyces
erythreus
429
enzyme) under conditions indicated above for 1 h, except that the labeled substrates were [1,3‘%]malonyl-CoA (sp act 40.8 Ci/mol) or [methyl‘%]methylmalonyl-CoA (sp act 56 Ci/mol). The reaction was stopped by adding 10 ~1 of perchloric acid and the precipitated protein was removed by centrifugation. The supernatants were subjected to descending paper chromatography with isobutyric acid/ (57/35/8, v/v/v) as de1 N ammonium acetate/water scribed elsewhere (9). [1-‘%JAcetyl-CoA was identified as the product of decarboxylation by cochromatography with authentic unlabeled acetyl-CoA. The product of decarboxylation of [methyl-‘4C]methylmalonyl-CoA was identified by descending paper chromatography as described elsewhere (17) with nbutanol/acetone/acetic acid/l N ammonium acetate (g/3/2/6, v/v/v/v).
Stereospeci@ity Preparation of (R)-[methyl-‘4C]methylmalonyl-CoA. (R)-[methyl-“C]Methylmalonyl-CoA was obtained by the removal of the (S) isomer from synthetic (R,S)[methyZ-“C]methy1ma1ony1-CoA by repeated treatment with purified malonyl-CoA decarboxylase from the uropygial gland of goose (17). The reaction mixture contained 100 pM (R,S)-[methyl-‘“C]methylmalonylCoA (56 Ci/mol), 100 pM Tris-HCl, pH 7.0, and 0.5 mM DTE in a total volume of 50 ~1. After incubating the mixture at 30°C for 3 h, the reaction was terminated by the addition of 10 ~1 of perchloric acid, the precipitated protein was removed by centrifugation, and the supernatant was subjected to HPLC analysis using a polyanion Sl HR 5/5 column (Pharmacia Fine Chemicals) which was equilibrated with 0.1 M potassium phosphate, pH 4.5. (R)-[methyZ‘%]Methylmalonyl-CoA was eluted with a 0.1-0.4 M gradient of potassium phosphate, pH 4.5, at a flow rate of 1.5 ml/min. The fractions containing (R)[methyl-“C]methylmalonyl-CoA were pooled, lyophilized, and desalted with a Bio-Gel P-2 column (1.4 X 90 cm) with 0.35 mM HCl as the solvent, and the product was kept frozen in a minimum volume of 0.35 mM HCI. Preparation of (S)-methylmalonyl-CoA. (S)-[3-14C]Methylmalonyl-CoA was enzymatically synthesized from propionyl-CoA using NaH14C03 and purified carboxylase from M. tuberculosis var. bovis BCG as described (8). (S)-[methyl-‘4C]Methylmalonyl-CoA was prepared by the removal of the (R) isomer from synthetic (R,S)-[methy1-‘4C]methylmalonyl-CoA by repeated treatment with S. w-ythreus malonyl-CoA decarboxylase. The reaction mixture contained 10 pM (R,S)-[methyI-‘4C]methylmalonyl-CoA and 100 fiM citrate-phosphate buffer, pH 6.0, in a total volume of 50 ~1. After incubating the mixture at 30°C for 3 h, the reaction was terminated and the product was isolated as described above. The purity of the (S)-[1-“C]methylmalonyl-CoA
430
HUNAITI
AND
preparation was examined by HPLC with 50 ItIM tetrabutylammonium phosphate, pH 5.5, in 50% methanol (18) using a p-Bondapak Cl8 column (0.39 x 30 cm, Water Associates). Decarboxylation of labeled methylmakmyl-CoA by the decarboxylases from S. evythreus and goose w-e pygial glands. Two nanomoles of (R)-[methyli4C]methylmalonyl-CoA (sp act 56 Ci/mol), 100 nmol of citrate-phosphate buffer, pH 6.0, and 28 fig of purified decarboxylase from S. erythreus in a total volume of 100 ~1 were incubated at 30°C for 1 h and the incubation was continued for a total of 3 h with hourly additions of fresh enzyme (28 yg). The reaction was terminated by addition of 10 ~1 of perchloric acid, the precipitated proteins were removed, and the product was analyzed by HPLC as above. Three nanomoles of (S)-[methyl-“C]methylmalonylCoA, 2 pmoi of Tris-HCl, pH 7.5, 25 nmol of DTE, and 5 fig of decarboxylase from goose in a total volume of 100 ~1 were incubated at 30°C for 1 h and incubation was continued for 3 h with hourly additions of fresh enzyme (5 bug). The reaction products were analyzed as described above. of malonyl-CoA decarTime course of appearance boxylose and erythromycin production. The time course was determined in two different media, the vegetative medium in which growth occurred and in 1.5% sucrose containing 0.2% L-alanine, a medium which allowed a high level of antibiotic production without growth. In the former case l-liter cultures were used as described above, and each day 10 ml of culture was centrifuged in a graduated tube at 1200g for 20 min and the packed cell volume was used as a measure of growth. In the latter case 50 ml of 3-day-old cultures in the vegetative medium were transferred to 1 liter of the sucrose-alanine medium. At desired time intervals erythromycin was extracted from the broth with ethyl acetate after removal of the mycelium by centrifugation and adjusting the pH to 9.5, and the antibiotic was measured calorimetrically as described (19). The activity of malonyl-CoA decarboxylase was measured in crude extracts as described above. Malonyl-CoA decarboxylase production was measured by a radial immunodiffusion method as described elsewhere (15). Standard curves were prepared by using serial dilution of the purified decarboxylase. RESULTS
Pur$cation of Malonyl-CoA Decarboxylase Crude extract of S. erythreus showed malonyl-CoA decarboxylase activity. Preliminary studies on the time course of appearance of the decarboxylase activity showed maximal levels at late-stationary phase of growth (4th day). Therefore we
KOLATTUKUDY
used 4-day-old cultures of S. erythreus for all of the subsequent experiments. The specific activity of malonyl-CoA decarboxylase in the crude extract ranged between 0.2 and 0.8 nmol decarboxylated min-’ rng-‘. Purification of the enzyme by some of the classical protein-fractionation methods proved to be difficult because of the sensitivity of the enzyme to high ionic strength and inhibition by sulfate ions as described in a later section. However, the crude extract could be fractionated with DEAE-cellulose which completely retained the enzymatic activity at pH 7.5, and the enzyme could be eluted by 0.1 M potassium phosphate buffer, pH 7.0, with 60-70% recovery and a 5- to 6-fold increase in specific activity. Gel filtration of this enzyme preparation on Sephadex G-100 revealed that the enzyme activity was contained in the second major protein peak (Fig. 1). This step resulted in over 24-fold purification with a 62% recovery of the activity. Dialysis of this enzyme preparation against 10 mM citrate-phosphate buffer, pH 4.5, and the removal of the flocculent precipitate resulted in 50% recovery of the activity with a 40-fold increase in specific activity. SP-Sephadex at pH 4.5 retained all the decarboxylase activity while the bulk of the protein was not retained (Fig. 1). The enzyme could be eluted by the 10 mM buffer at pH 6.0, but this procedure caused precipitation and denaturation of the enzyme, resulting in only 3-4% recovery of the enzymatic activity. However, with the use of 20 mM buffer containing 5% glycerol, the recovery of enzymatic activity increased up to nearly 20% with about a 270-fold purification. Electrophoresis of the resulting protein showed three major bands at 45, 30, and 15 kDa. The relative intensities of these bands varied from preparation to preparation with a decrease in the 45k band accompanied by increases in the other two bands or vice versa. Upon gel filtration of this enzyme preparation on Sephadex G-75, two peaks of enzymatic activities were observed. SDS-electrophoresis of the first peak of protein showed only a single band with a molecular weight of 45k, whereas electrophoresis of the second peak of enzymatic
MALONYL-CoA
DECARBOXYLASE
FROM
Streptmyces
A
G-100 20-
431
erythreus
PROTEIN DECARSOXYLASE -8
l.0
-4
E c
,-\
2
.\------,’
’
N 2 w
\ \ \ -_
0
*0
40
BO
I
SO
DECARSOXYLASE
FRACTION
NUMBER
FIG. 1. Purification of malonyl-CoA decarboxylase from S. erythreus. (A) Sephadex G-100 gel filtration of the enzyme obtained from the DEAE-cellulose step; (B) SP-Sephadex chromatography of the decarboxylase. The application of 20 mM citrate-phosphate, pH 6.0, containing 5% glycerol is indicated. The other experimental details are given in the text.
activity showed two bands at 30k and 15k. (A minor band at 38k was also sometimes observed.) About two-thirds of the applied activity was found in the first peak and the rest in the second peak. These results strongly suggested that the purification procedure resulted in a proteolytic clip in TABLE
this enzyme. All our attempts to prevent proteolysis by the inclusion of protease inhibitors such as PMSF, EDTA, and soybean trypsin inhibitor failed. In any case these purification procedures resulted in about 800-fold purification with an overall recovery of about 12% (Table I). The enI
PURIFICATION OF MALONYL-COA DECARBOXYLASE FROM S. erythreus
Step Crude extract DEAE-Cellulose Sephadex G-100 Dialyzed pH 4.5 SP-Sephadex Sephadex G-75 Note. Experimental
Protein (mg)
Total activity (nmol/min)
2400 300 61 35 1.5 0.35 details
1790 1239 1098 910 300 210
are in the text.
Specific activity (nmol/min/mg) 0.75 4.13 18 26.2 200 600
Yield (%I 100 69 61 51 18 12
Purification (fold) 1 5 24 35 267 800
432
HUNAITI
AND
zyme was relatively unstable; it lost about 50% of its activity upon freezing and thawing. However, it was relatively stable if stored at 4°C in 10 mM citrate-phosphate buffer, pH 4.5, containing 1 mM DTE and 5 mM EDTA. With increasing periods of storage the 45k band intensity decreased with a corresponding increase in the intensities of the bands at 30k and 15k, suggesting that a proteolytic clip similar to that occurring during purification also took place during storage of the purified enzyme. Unless otherwise specified, the purified enzyme containing only 45-kDa peptides was used for characterization of the enzyme. Molecular Weight and Subunit Composition Polyacrylamide gel electrophoresis (‘7.5%) of the enzyme recovered from the SP-Sephadex step showed a single band and all of the decarboxylase activity was contained in this band (Fig. 2), strongly suggesting that the enzyme was near homogeneity. Gel filtration on a Sephadex G150 column with proteins of known mo-
4 DISTANCE
KOLATTUKUDY
lecular weight showed that the native decarboxylase had a molecular weight of 93,000 (Fig. 3). Polyacrylamide disc gel electrophoresis of the enzyme after treatment with SDS and 2-mercaptoethanol showed a single band (Fig. 4). The molecular weight of the protomer, as determined by a linear plot of log molecular weight vs mobility of standard proteins, was 45,000. This observation, together with the molecular weight of the native enzyme, strongly suggests that the present decarboxylase has an (Yesubunit composition. Amino Acid Composition Amino acid composition of the present enzyme is shown in Table II, and for comparison the composition of malonyl-CoA decarboxylases from avian sources is included. Immunological
Properties
Ouchterlony double-diffusion analysis showed that rabbit antibody prepared against the 45k protein gave a single precipitin line (Fig. 5, inset) when purified enzyme or the SP-Sephadex enzyme was used as the antigen. The antibody also gave a single precipitin line with both the 30k and 15k proteins isolated by SDS-electrophoresis. Complete fusion of the lines was observed with the 30k protein but not with the 15k protein. These results suggest that all three peptides contained common antigenic determinants, with the larger two being immunologically very similar, if not identical. These immunological relationships are consistent with the conclusion that the 30k and 15k proteins were derived from the 45k protein by proteolytic cleavage. The enzymatic activity was progressively severely inhibited by increasing amounts of the antiserum, and finally virtually complete inhibition of the enzyme activity was observed (Fig. 5).
8 (cm)
FIG. 2. Polyacrylamide gel electrophoresis of the decarboxylase (25 pg) obtained from the SP-Sephadex step on ‘7.5% gel and the distribution of malonyl-CoA decarboxylase activity in the gel. The experimental details are given in the text.
Catalytic
Properties
The activity of the purified enzyme preparation was linear with protein concentration at least up to 100 pg/ml, and the reaction rate was linear up to at least 20
MALONYL-CoA
DECARBOXYLASE
FROM
TABLE AMINO
ACID
433
erqthreus
Streptmyces
II
COMPOSITION
OF MALONYL-COA
DECARBOXYLASE
Amino acid
(mol%)
Avian” (mol%)
Asx Thr Ser Glx Pro GIY Ala Val Ile Leu Tw Phe LYS His Arg CYS Met
8.1 4.3 16.4 13.6 3.4 17.9 8.3 3.6 1.9 3.6 0.8 1.3 2.9 2.4 2.8 1.5 8.8
6.6 4.1 7.0 14.0 4.4 6.8 7.6 6.4 4.0 12.7 2.1 3.6 5.8 2.4 7.3 1.8 1.5
S. erythrews
45k
30k
15k
FIG. 3. SDS-electrophoresis of malonyl-CoA decarboxylase. The experimental details are given in the text. (A) The enzyme after SP-Sephadex column chromatography; (B) the protein from the first peak of enzymatic activity obtained by Sephadex G-75 gel filtration shown in Table I.
Decarboxylase
(93K)
a The values obtained for the enzyme isolated from three avian sources (7, 20) were averaged.
min of incubation. Therefore, all subsequent experiments were done within such linear ranges for both protein and time. The rate of decarboxylation increased sharply as the pH increased from 3 to 5 and further increases in pH caused a decline in activity, resulting in a rather sharp pH optimum at around 5.0 (Fig. 6). For routine assays 10 mM citrate-phosphate, pH 5.0, was used. Chromatofocusing indicated that the pI of malonyl-CoA decarboxylase was 4.4. Eflect of Substrate Concentrations
x
FRACTION
NUMBER 25
50
Ve-Votml) FIG. 4. Determination of the molecular weight of the malonyl-CoA decarboxylase (from the Sephadex G-75 step) by gel filtration on calibrated Sephadex G-150 column. Inset, the chromatography of the purified decarboxylase on G-150 column. The protein standards used are (A) y-globulin (150k); (B) bovine serum albumin (68k); (C) ovalbumin (45k); and (D) carbonic anhydrase (30k).
A typical Michaelis-Menten-type substrate saturation pattern was observed with malonyl-CoA as the substrate (Fig. 7). From linear double-reciprocal plots the apparent Km and Vvalues were calculated to be 143 pM and 250 nmol mini’ mg-‘, respectively. Inhibitors
Malonyl-CoA decarboxylase from S. erby both acetyl-CoA
ythreus was inhibited
HUNAITI
AND
KOLATTUKUDY
0
FIG. 5. The effect of rabbit antiserum prepared against purified malonyl-CoA decarboxylase on the enzymatic activity. The inset shows Ouchterlony double-diffusion analysis. The center well contained 10 gl of rabbit antiserum and the outer wells, starting from top and going clockwise, contained 0.5, 1.0, 1.5, and 2 pg of purified decarboxylase. Other experimental details are given in the text.
0 A
A
I o-
’ 3
4
5
6
7
6
6
PH FIG. 6. Effect of pH on the rate of decarboxylation of malonyl-CoA by the purified enzyme. The reaction mixtures contained 10 wg of the purified enzyme, and the assay conditions were as described under Materials and Methods.
and free CoA. Double-reciprocal plots of the results obtained with different concentrations of acetyl-CoA and free CoA clearly showed that this inhibitor was competitive with malonyl-CoA (Fig. 8). Replot of the slope vs concentration of the inhibitors (insets) showed that the inhibition was parabolic. Other short-chain acyl-CoAs (Table III) showed less inhibition. Neither avidin (1 mg/ml) nor biotin (1 mM) had any effect on the decarboxylation rate, and, therefore, biotin is probably not involved in catalysis. In contrast to the previously examined malonyl-CoA decarboxylases, the present decarboxylase showed no inhibition by thiol-directed reagents such as p-hydroxymercuribenzoate (up to 0.5 mM) and N-ethylmaleimide (up to 1.0 mM). Neither iodoacetamide nor phenylmethylsulfonyl fluoride (each up to 5 mM) showed any inhibition. Chelating
MALONYL-CoA
l/[Sl 80 [~~AL~NYL-COA]
DECARBOXYLASE
x10*
160
240
FROM
Streptmyces
435
erythreus
decarboxylate (S)-methylmalonyl-CoA whereas the (R) isomer could be nearly completely decarboxylated with this enzyme (Table IV). Upon prolonged incubadecarboxylase with tion of S. erythreus (R,S)-methylmalonyl-CoA only 50% of the methylmalonyl-CoA was decarboxylated. That the remaining methylmalonyl-CoA was exclusively the (S) isomer was shown by the observation that the (S)-specific decarboxylase from the goose uropygial gland
(~tdl)
FIG. 7. Effect of substrate concentration on the rate of decarboxylation and double-reciprocal plot for malonyl-CoA. The reaction mixture contained 10 fig of the purified enzyme, and assay conditions were as described under Materials and Methods.
A 1.0 ACETYLCOA LkM 75 .
E 0 10.5 v)
agents such as EDTA, EGTA (1.0 mM), and bipyridyl (0.5 mM) showed no inhibition, suggesting that heavy metals are not involved in catalysis. Effect of Salts on Decarboxylase
LlIL 0
25
50
75 /
Activity
The effect of several salts on the enzyme activity was investigated, and from the results obtained (Fig. 9) it appeared that the enzyme was inhibited by high ionic strength. SO: ions severely inhibited the enzyme even at low concentration (5 mM). Removal of salts by dialysis or gel filtration resulted in full recovery of the enzymatic activity. Stereospeci&ity It is known that the malonyl-CoA decarboxylase from the uropygial gland of goose is specific for the decarboxylation of (S)-methylmalonyl-CoA, and shows no activity with (R)-methylmalonyl-CoA (17). Therefore, we used this enzyme as a stereospecific reagent to remove the (S) isomer from the (R,S) mixture and thus generated stereochemically pure (R)-methylmalonyl-CoA. It is also known that the acyl-CoA carboxylase of AI. tuberculosis generates specifically (S)-methylmalonylCoA (8). Therefore, we used this enzyme to generate (S)-methylmalonyl-CoA. It was found that the present malonyl-CoA decarboxylase from S. erythreus could not
50
&
0
25 0
.
0
2
11 [MALONYL-COAI
4
6 x
I o2 PM
FIG. 8. Double-reciprocal plots showing competitive inhibition of purified malonyl-CoA decarboxylase by acetyl-CoA (A) and free CoA (B). The concentrations of the inhibitors are indicated and the effects of the concentration of the inhibitor on the slopes are shown in the insets.
436
HUNAITI TABLE
AND
KOLATTUKUDY
III
EFFECT OF INHIBITORS ON MALONYL-COA DECARBOXYLASE FROM S. erythreus”
Concentration (mM)
Inhibitor
Relative rate (% of control)
Coenzyme A
0.5 1.0
36 30
Acetyl-CoA
0.5 1.0
38 31
Propionyl-CoA
0.5 1.0
60 60
Butyryl-CoA
0.5 1.0
56 47
Methylmalonyl-CoA
0.5 1.0
61 48
Succinyl-CoA
0.5 1.0
91 41
SALT
( mbl
)
FIG. 9. Effect of salts on the enzymatic activity of the purified malonyl-CoA decarboxylase from S. erythreus. Ammonium sulfate gave results identical to those shown for sodium sulfate. The indicated concentrations of salts were present in the reaction mixtures
served with malonyl-CoA when both substrates were used at 0.3 mM. Succinyl-CoA was not decarboxylated, neither was maionic acid.
a Enzyme (10 pg) in 90 gl of 10 mM citrate-phosphate buffer, pH 5.0, was preincubated with inhibitors at the indicated concentration at room temperature for 15 min, and reactions were initiated by the addition of 10 pl of substrate.
Product IdentiJication completely decarboxylated it. These results show that the present decarboxylase, unlike the goose decarboxylase, shows absolute stereospecificity for (R)-methylmalonyl-CoA.
and Stoichimetry
To be certain that the release of CO2 from malonyl-CoA involved decarboxylation of the substrate with the formation of acetylCoA, rather than some other type of reactions involving release of COa, the products generated from malonyl-CoA were identified and quantitated. HPLC and paper chromatographic analyses of the products showed that the amount of acetyl-CoA generated from [1,3-‘4C]malonyl-CoA was
Substrate Speci$city Malonyl-CoA was highly preferred over (R)-methylmalonyl-CoA, the latter being decarboxylated at about 3% of the rate obTABLE
IV
STEREOSPECIFICITY IN THE DECARBOXYLATION OF METH~MALON~-COA BY S. ergthrews MALONYL-COA DECARBOX~LASE
Source of decarboxylase
Substrate (S)-[l-‘4C]Methylmalonyl-CoA (R,S)-[methyl-i4C]Methylmalonyl-CoA (R)-[meth&‘4C]Methylmalonyl-CoA (S)-[methyl-14C]Methylmalonyl-CoA Note. Experimental
details
are in the text.
S. e-rythreus S. erythreua S. eqthrevs Goose uropygial
gland
Percentage decarboxylation 0 50 90 88
MALONYL-CoA
DECARBOXYLASE
equivalent to the amount of CO2 released. The product generated from methylmalonyl-CoA was identified as propionyl-CoA by paper chromatography and HPLC. Time Course of Appearance of MalonylCoA Decarboxylase and Erythrmycin Production The specific activity of malonyl-CoA decarboxylase was dependent on age of the culture. The specific activity was at a maximum just prior to the onset of erythromycin production (Fig. 10). To test whether the increase in specific activity of the decarboxylase reflected an increase in the level of the enzyme protein or some type of activation, the decarboxylase level was measured immunologically using rabbit antiserum prepared against the purified decarboxylase. The area of the precipitant rings generated in the radial immunodiffusion assay was directly proportional to the amount of the antigen. The results obtained by this method were in agreement with those obtained by measurement of the enzyme activity, and showed that the amount of decarboxylase increased about lo-fold just prior to the onset of erythromycin production. Both methods showed
FROM
2
3
4
5
erythreus
437
that the decarboxylase level decreased as the antibiotic production leveled off. Such a correlation between the decarboxylase level and antibiotic production was also found in a sucrose-alanine medium in which production of erythromycin proceeds without significant growth (Fig. 10). DISCUSSION
The purification of malonyl-CoA decarboxylase from a microorganism, first reported in the present paper, enabled us to compare this bacterial enzyme with that isolated from avian and mammalian sources (20, 21, 22). The subunit composition of the 5’. erythreus decarboxylase is in sharp contrast to the decarboxylases isolated from avian and mammalian tissues; the present enzyme is a dimer of 45kDa protomers whereas the mammalian and avian enzymes are tetramers of 50kDa peptides. Since the mature protomer from the animal mitochondrial enzyme was a 47-kDa peptide, the size of the protomers of the decarboxylases can be considered similar. The amino acid composition of the S. erythreus decarboxylase is also quite different from that of the previously examined decarboxylases (20). The present de-
00
1
Streptomyces
i
6
DAYS
FIG. 10. Time course of the appearance of malonyl-CoA decarboxylase and production of erythromycin by S. erythreus in a vegetative growth medium (A) and in a sucrose-alanine medium (B). The experimental details are given in the text.
438
HUNAITI
AND
carboxylase contained considerably lower amounts of hydrophobic and basic amino acids than those found in the avian enzymes, whereas the contents of the acidic residues are quite similar. Antibodies raised against S. erythreus decarboxylase neither inhibited nor cross-reacted with the avian enzyme, and the antibodies prepared against the decarboxylase from the uropygial gland of goose neither affected the activity nor cross-reacted with the S. erythreus enzyme, suggesting that the present enzyme is immunologically quite different from the avian enzyme. Catalytic properties of S. erythreus decarboxylase were quite different from those of the avian and mammalian decarboxylases (20-22). The present enzyme showed an acidic pH optimum (5.0) compared to the basis pH optimum (9.0) observed with the avian and mammalian enzymes. The present enzyme also showed a higher Km and lower V than the previously examined decarboxylases. Effects of the various inhibitors on the S. erythreus enzyme were significantly different from those on the mammalian and avian malonyl-CoA decarboxylases. The present enzyme was not significantly inhibited by p-hydroxymercuribenzoate or other thiol-directed reagents whereas the other decarboxylases were severely inhibited by such reagents. CoA and acetyl-CoA inhibited the enzyme from 5’. erythreus whereas they had little effect on the avian or mammalian enzyme. The properties of the present enzyme are different from those of a partially purified decarboxylase previously obtained from hf. tuberculosis and Pseudomonas jluorescens, although some similarities were also apparent. S. erythreus decarboxylase is a dimer of 45,000 while those of M. tuberculosis and P. JluorescerzS were found to be monomeric with molecular weights of 45,000 and 56,000, respectively. Effect of the various inhibitors also showed significant differences among the three enzymes. p-Hydroxymercuribenzoate showed little inhibition with the enzyme from S. erythreus or P. fluorescens whereas it severely inhibited the enzyme from M. tuberculosis. Succinyl-CoA inhibited the enzyme from M. tuberculosis whereas it showed little in-
KOLATTUKUDY
hibition with the enzyme from S. erythreus and P. jlwrescens. Acetyl-CoA and CoA inhibited the enzyme from M. tuberculosis and S. erythreus whereas little inhibition was noted with enzyme from P.JEuorescens. Finally, the high sensitivity of the enzyme from S. erythreus to high ionic strength and sulfate ions revealed that it was quite different from the previously examined decarboxylases. The present enzyme shows a novel stereospecificity. It decarboxylates (R)methylmalonyl-CoA, but not (S)-methylmalonyl-CoA. This stereospecificity is opposite to that observed with goose malonylCoA decarboxylase (17) which accepts only (S)-methylmalonyl-CoA as the substrate. The decarboxylase preparation from M. tuberculosis also decarboxylates (S)-methylmalonyl-CoA (unpublished) whereas even the crude extract of S. erythreus decarboxylates only the (R) isomer, indicating the absence of an (S)-specific decarboxylase and a methylmalonyl-CoA racemase in this organism. The possible functional significance of this novel stereospecificity is not clear. It is possible that the present enzyme is involved in producing propionyl-CoA from the (R)-methylmalonyl-CoA generated from succinylCoA, which would be derived from the carbohydrate provided as a major carbon source in the medium. Another possibility is that the observed stereospecificity has no real physiological significance and that the enzyme in vivo decarboxylates only malonyl-CoA. The relationship between the time course of appearance of the decarboxylase and that of erythromycin synthesis suggested that the enzyme might play a role in the antibiotic production. In the vegetative medium the cells would require malonylCoA for the synthesis of the n-fatty acids needed for membrane synthesis during the period of rapid growth of the organism. As the growth was nearly complete the decarboxylase level increased. Also, in the sucrose-alanine medium the decarboxylase level dramatically increased just prior to the rapid production of the antibiotic. In both cases the role of the decarboxylase might be to insure that the antibiotic syn-
MALONYL-CoA
DECARBOXYLASE
thase would not be adversely affected by the presence of malonyl-CoA. Such a role had previously been suggested in the case of production of multiple methyl-branched fatty acids by the uropygial gland and M. tuberculosis. However, it is not known whether the antibiotic synthase is adversely affected by the presence of malonylCoA. In any case, the observed relationship between the time course of appearance of the decarboxylase and that of erythromycin production is consistent with the conclusion that this enzyme plays a role in the antibiotic production.
FROM
D. L., AND KOLATTUKUDY, P. E. (1983) J. Biol. Chem. 258, 2979-2985. 6. KIM, Y. S., AND KOLATNKUDY, P. E. (1979) Arch. Biochem. Biophys. 196,543-551. 7. BUCKNER, J. S., KOLA~KUDY, P. E., AND POULOSE, Biophys. 177,539A. J. (1976) Arch B&hem 551. 8. RAINWATER, 9. BUCKNER,
(Corcoran, J. W., ed.), Vol. 4, pp. 133-174, Springer-Verlag, Berlin/Heidelberg/New York. 2. HUNAITI, A. R., AND KOLA’ITUKUDY, P. E. (1982) Arch. B&hem. Biophys. 216,362-371. 3. BUCKNER, J. S., KOLATTUKUDY, P. E., ANDROGERS, L. (1978) Arch. Biochem. Biophys. 186,152-163. 4. BUCKNER,
J. S., AND KOLATTUKUDY,
P. E. (1976)
in Chemistry and Biochemistry of Natural Waxes (Kolattukudy, P. E., ed.), pp. 147200, Elsevier, Amsterdam/New York.
J. S., AND KOLA~UKUDY,
10. GREGOLIN,
P. E. (1975)
14,1768-1773.
C., RYDER,
J. Biol
P. E. (1982)
151, 905-911.
Biochemistry
E., ANDLANE,
M. D. (1968)
Chem. 243,4227-4235.
P. E. (1982) 14, 609-614. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., AND RANDALL, R. J. (1951) J. BioL Chem. 193, 265-275. LAEMMLI, V. K. (1970) Nature (ikndolz) 277, 680685. OUCHTERLONY, 0. (1953) Acta Pathol. Microbial Scam! 32,231-240. LIN, T. S., AND KOLATTUKUDY, P. E. (1978) J. Bacteriol. 133,942-951. MOORE, S. (1962) J. Biol. Chem 238,235-237. KIM, Y. S., AND KOLATTUKUDY, P. E. (1980) J. Biol. Chem. 255, 686-689.
11. RAINWATER,
D. L., AND KOLAITUKUDY,
Int. J Biochem
12.
13. 14.
15.
17.
1. CORCORAN, J. W. (1981) in Antibiotics
D. L., AND KOLATXJKUDY,
J. Bacterial.
16.
REFERENCES
439
erythreus
5. RAINWATER,
ACKNOWLEDGMENTS Dr. Andrew Bognar conducted some of the preliminary experiments. We thank Dr. A. J. Poulose and Dr. D. Rainwater for helpful discussion. This work was supported by Grant GM18278 from the Institute of General Medical Sciences of the U. S. Public Health Service.
Streptomyces
18. RAINWATER, D. L., AND KOLATTUKUDY, P. E. (1981) Arch. Biochem. Biophys. 213,372-383. 19. FORD, J. M., PRESCOIT, G. C., HINMAN, J. W., AND CAROIV, E. L. (1953) Anal. Chem 25,1195-1197. 20. KIM, Y. S., AND KOLAIXIKUDY, P. E. (1978) Arch. B&hem. Biophys. 190,585-597. 21. KIM, Y. S., AND KOLA~UKUDY, P. E. (1978) Arch. B&hem. Biophys. 190,234-246. 22. KIM,
Y.
S.,
Biochim.
AND
KOLATTUKUDY,
P.
Biophys. Actu 531, 187-196.
E.
(1978)