Vol. 134, No. 2
JOURNAL OF BACTERIOLOGY, May 1978, p. 476-482
0021-9193/78/0134-0476$02.00/0 Copyright © 1978 American Society for Microbiology
Printed in U.S.A.
Biosynthesis of Riboflavin in Bacillus subtilis: Function and Genetic Control of the Riboflavin Synthase Complex A. BACHERt* AND B. MAILANDERtt Institut fir Mikrobiologie, Universitat Hohenheim, 7000 Stuttgart 70, German Federal Republic
Received for publication 19 September 1977
Two riboflavin synthase activities (heavy and light) have been observed in earlier studies with Bacillus subtilis. The heavy enzyme is a complex of one molecule of light enzyme (consisting of three a subunits) and approximately 60 fi subunits (A. Bacher, R. Baur, U. Eggers, H. Harders, and H. Schnepple, p. 729-732, in T. P. Singer (ed.), Flavins and Flavoproteins, Elsevier, Amsterdam, 1976). The formation of a and f8 subunits is coordinately controlled. Mutants apparently deficient in ,B subunits were isolated as riboflavin requires after mutagenesis of B. subtilis with ICR 191. The mutants could grow with diacetyl instead of riboflavin. Growth with diacetyl was associated with the accumulation of substantial amounts of the riboflavin precursor, 6,7-dimethyl-8-(Dribityl)lumazine. It follows that the mutants are deficient in an enzyme activity required for the formation of the lumazine from the pyrimidine precursor. We conclude that heavy riboflavin synthase is a bifunctional enzyme. The riboflavin synthase activity is mediated by the a subunits, whereas the ,B subunits are necessary for an earlier biosynthetic step.
Riboflavin synthase catalyzes the formation of one molecule each of riboflavin and 5-amino2,6-dihydroxy-4-(D-ribitylamino) pyrimidine from two molecules of 6,7-dimethyl-8-(D-ribityl)lumazine. The enzyme has been found in various microorganisms and plants (for review, see references 13, 25, and 29). Riboflavin synthase from yeast has been highly purified, and the substrate specificity and the stereospecific mode of action have been studied in considerable detail. On the basis of studies with isotopically labeled substrates, a mechanism of reaction has been proposed (25, 26). Little is known about the physical properties and the structure of the yeast enzyme. The concentration of riboflavin synthase in cell extracts of most microorganisms studied is rather low (25). However, high enzyme levels have been found in flavinogenic mutants of Bacillus subtilis (4, 10). It has been shown that the biosynthesis of the enzyme is controlled by repression in this microorganism. Studies with cell extracts from a genetically derepressed mutant of B. subtilis showed the presence of two riboflavin synthase activities of greatly different sizes and molecular weights (3). The light enzyme, which accounts for more than 80% of the total activity, is a trimer of identical a subunits. The heavy enzyme is a complex of one molecule t Present address: Fachbereich Biologie, Universitit Frankfurt, 6000 Frankfurt/Main, German Federal Republic. tt Present address: Pfizer Laboratories, 75 Karlsruhe, German Federal Republic.
of light enzyme (i.e., three a subunits) and approximately 60 identical ,B subunits (3; A. Bacher, M. K. Otto, and H. Schnepple, unpublished data). Isolated ,B subunits had no riboflavin synthase activity, and the low specific activity of the heavy enzyme as compared to that of the light enzyme supports the hypothesis that only the a subunits are catalytically involved in the conversion of 6,7-dimethyl-8-(Dribityl)lumazine to riboflavin. The early steps of riboflavin biosynthesis are incompletely understood. The pathway starts at the level of guanosine or a respective nucleotide, as shown by isotope incorporation studies (7, 9, 20). The ribose moiety of the purine precursor is directly converted to the ribityl moiety of the vitamin. It has been suggested that the first committed step of the biosynthesis is catalyzed by GTP cyclohydrolase II in Escherichia coli. The enzyme catalyzes the simultaneous release of carbon-8 and of pyrophosphate from GTP (14). Several pyrimidine-type intermediates have been isolated in studies with riboflavindeficient mutants of Saccharomyces cerevisiae (Fig. 1) (5, 6, 19). The conversion of 5-amino-2,6dihydroxy-4-(D-ribitylamino)pyrimidine to 6,7dimethyl-8-(D-ribityl)lumazine requires the addition of a four-carbon moiety of unknown structure. Several studies suggested the involvement of acetoin or a biogenetically related compound (12, 16, 18). The hypothesis has been criticized by other authors (1, 15). Recent studies suggested the involvement of a pentose or tetrose 476
BIOSYNTHESIS OF RIBOFLAVIN
VOL. 134, 1978
OH
OH
OH
H2N
rib 1
477
N
H2N
HN
NH2
HN
N-NNH2
N
NH2
C H2
(P3)OH2C
I P) OH2CLo 0
H-C-OH H -C-OH
H
v
H-C-OH
OH OH
OH OH
C H20H
I
II
OH
rib 2
HN
N
N
H3C
N
H3C
N
/
N
rib 4
OH
NAO
CH2
C H2
rib 5
\/,/N
~H
H 3C
H3C
N
NAO
CH2
H-C-OH
H-C-OH
H-C-OH
H-C-OH
H-C-OH H-C-OH
H-C-OH H-C-OH
H-C-OH
CHH2OH FIG.
0
0 rib 3
H 2N
CH20H
CH20H
VI v IV 1. Biosynthesis of riboflavin (14,20). Genes involved in S. cerevisiae are indicated according to studies
by Oltnanns et al. (22,23). I, guanosine triphosphate; II, 2,5-diamino-6-hydroxy-4-(D-ribosylamino)pyrimidine 5'-phosphate; III, 2,5-diamino-6-hydroxy-4-(D-ribitylamino)pyrimidine; IV, 5-amino-2,6-dihydroxy-4-(D-ribitylamino)pyrimidine; V, 6,7-dimethyl-8-(D-ribityl)lumazine; VI, riboflavin.
(1, 2). The involvement of 6-methyl-7-(1',2'-dihydroxyethyl)-8-ribityliumazine as a precursor of 6,7-dimethyl-8-(D-ribityl)lumazine has been proposed on the basis of studies with mutants of B. subtilis (11). Several biosynthetic reactions in the pathway of riboflavin biosynthesis are not yet accessible to direct enzymatic studies in spite of considerable efforts. This paper presents evidence that the heavy riboflavin synthase of B. subtilis is a bifunctional enzyme. Whereas the a subunits mediate the known riboflavin synthase activity of the protein, the ,B subunits appear to be necessary for an earlier reaction in the biosynthesis of the vitamin. MATERIALS AND METHODS Bacterial strains. B. subtilis 168M trp2C
TABLE 1. Mutants of B. subtilis used0 Phenotype Origin RibH78 NGb H94 Flavinogenicc (4) H175 RibNG H322 Flavinogenic (4) RibCR2 ICRd RibICR CR6 RibICR CR9 'All mutants were derived from strain 168M and require tryptophan. b N-Methyl-N'-nitro-N-nitrosoguanidine mutagenesis. 'Genetically derepressed mutant accumulating riboflavin. d ICR 191 mutagenesis. Strain
was
pg/ml) ovemight (21). A portion was transferred to
kindly provided by C. Anagnostopoulos, Centre de Recherches Scientifiques, Gif-sur-Yvette, France. The other mutants used are listed in Table 1. Media. The basic medium was Spizizen minimal medium (30) supplemented with tryptophan (50 mg/liter). Vitamin-free Cmino Acids (Difco), riboflavin, and diacetyl were added as required. Chemicals. 6,7-Dimethyl4-8(D-ribityl)lumazine was prepared by published procedures (28). ICR 191 was a gift of Badische Anilin- und Soda-Fabrik, Ludwigshafen. Riboflavin was purchased from Merck AG, Darmstadt, W. Germany, and diacetyl was from Fluka AG, Buchs, Switzerland. Isolation of mutants. B. subtilis 168M was grown in complete medium supplemented with ICR 191 (20
fresh complete medium, and the culture was incubated overnight. Riboflavin-deficient mutants were isolated as described (4). Only one mutant of each respective phenotype was collected from each mutagenized culture. Growth of bacteria and preparation of cell extracts. Bacteria were grown in 0.5-liter batches of medium supplemented with vitamin-free Casamino Acids (5 g/liter). Riboflavin was added as required. The cultures were incubated with shaking overnight. The cells were harvested by centrifugation and stored at -20°C. For the preparation of cell extracts, frozen bacterial cells were thawed in buffer containing 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid, and 10 mM sodium sulfite. The suspension
478
J. BACTERIOL.
BACHER AND MAILANDER
was ultrasonically treated and centrifuged. Assay of riboflavin synthase. Assay mixtures contained 0.1 M phosphate (pH 7.4), 10 mM ethylenediaminetetraacetic acid, 10 mM sodium sulfite, and 0.6 mM 6,7-dimethyl-8-(D-ribityl)lumazine. Assays were performed at 37°C as described by Plaut and Harvey (28). One unit of enzyme activity catalyzes the formation of 1 nmol of riboflavin per h. Sucrose gradient centrifugation. Sucrose gradients (5 to 20%) contained 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid and 10 mM sodium sulfite. Samples of 0.5 ml were layered on top of the gradients. They were centrifuged in an SW27.1 rotor (Spinco) at 23,000 rpm and 4°C for 20 h. Fractions were collected and analyzed. Preparation ofantisera. Rabbits were immunized with 0.2- to 0.5-mg samples of light or heavy riboflavin synthase of B. subtilis (3) in 0.5 ml of Freund complete adjuvant (Difco). Serum obtained after immunization with heavy enzyme was made monospecific for ,8 subunits by addition of light riboflavin synthase. Immunodiffusion. Double-diffusion and simple radial diffusion experiments were performed according to published procedures (17, 24). Plates contained 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid, 0.01% sodium azide, and 1.2% agar (Difco). Light and heavy riboflavin synthase, respectively, were used as standards in quantitative measurements of a and ,B subunits. Results were expressed in terms of enzyme units. Isolation of 6,7-dimethyl-8-(D-ribityl)lumazine. B. subtilis mutant CR2 was grown in 1-liter batches of medium supplemented with vitamin-free Casamino Acids (5 g/liter) and diacetyl (50 id/liter). The cultures were incubated at 37°C for 36 h. The cells were removed by centrifugation. 6,7-Dimethyl-8-(D-ribityl)lumazine was isolated from the culture medium as described (4). Miscellaneous methods. Protein concentration was estimated by the biuret method. Thin-layer chromatography was performed with plates coated with cellulose MN 300 (Macherey & Nagel, Diiren, W. Germany). The solvent systems were 1-butanol-acetic acid-water, 50:15:35 (vol/ vol/vol), and 3% ammonium chloride, respectively.
RESULTS Regulation of enzyme subunit biosynthesis. The level of light and heavy riboflavin synthase in cell extracts of B. subtilis mutants was analyzed by sucrose gradient centrifugation and by simple radial immunodiffusion using monospecific antisera (Table 2). In the flavinogenic mutants H94 and H322, the heavy enzyme accounted for approximately 15% of the total riboflavin synthase activity, as shown by sucrose gradient centrifugation. The values obtained by immunochemical determination of ,B subunits are in agreement within the experimental limits. Figure 2 shows the sucrose gradient centrifugation profile of mutant H94. The peak of f8 subunits coincides with the peak of heavy riboflavin synthase activity. Control experiments showed
TABLE 2. Concentration of light and heavy riboflavin synthase in cell extracts of B. subtilis mutants Riboflavin synthase Strain
Riboflavin' (mg/liter)
(U/mg of protein)
Total
Lightb Heavyb Heavy'
H94 0 107 89 18 18 0 80 12 H322 68 15 H175 0.03 59 51 8 5 H175 10 3 NDd ND 0.7 67 7 CR9 0.03 60 9 10 1.4 ND ND CR9 0.2 CR2 0.03 118 118
JOURNAL OF BACTERIOLOGY, May 1978, p. 476-482
0021-9193/78/0134-0476$02.00/0 Copyright © 1978 American Society for Microbiology
Printed in U.S.A.
Biosynthesis of Riboflavin in Bacillus subtilis: Function and Genetic Control of the Riboflavin Synthase Complex A. BACHERt* AND B. MAILANDERtt Institut fir Mikrobiologie, Universitat Hohenheim, 7000 Stuttgart 70, German Federal Republic
Received for publication 19 September 1977
Two riboflavin synthase activities (heavy and light) have been observed in earlier studies with Bacillus subtilis. The heavy enzyme is a complex of one molecule of light enzyme (consisting of three a subunits) and approximately 60 fi subunits (A. Bacher, R. Baur, U. Eggers, H. Harders, and H. Schnepple, p. 729-732, in T. P. Singer (ed.), Flavins and Flavoproteins, Elsevier, Amsterdam, 1976). The formation of a and f8 subunits is coordinately controlled. Mutants apparently deficient in ,B subunits were isolated as riboflavin requires after mutagenesis of B. subtilis with ICR 191. The mutants could grow with diacetyl instead of riboflavin. Growth with diacetyl was associated with the accumulation of substantial amounts of the riboflavin precursor, 6,7-dimethyl-8-(Dribityl)lumazine. It follows that the mutants are deficient in an enzyme activity required for the formation of the lumazine from the pyrimidine precursor. We conclude that heavy riboflavin synthase is a bifunctional enzyme. The riboflavin synthase activity is mediated by the a subunits, whereas the ,B subunits are necessary for an earlier biosynthetic step.
Riboflavin synthase catalyzes the formation of one molecule each of riboflavin and 5-amino2,6-dihydroxy-4-(D-ribitylamino) pyrimidine from two molecules of 6,7-dimethyl-8-(D-ribityl)lumazine. The enzyme has been found in various microorganisms and plants (for review, see references 13, 25, and 29). Riboflavin synthase from yeast has been highly purified, and the substrate specificity and the stereospecific mode of action have been studied in considerable detail. On the basis of studies with isotopically labeled substrates, a mechanism of reaction has been proposed (25, 26). Little is known about the physical properties and the structure of the yeast enzyme. The concentration of riboflavin synthase in cell extracts of most microorganisms studied is rather low (25). However, high enzyme levels have been found in flavinogenic mutants of Bacillus subtilis (4, 10). It has been shown that the biosynthesis of the enzyme is controlled by repression in this microorganism. Studies with cell extracts from a genetically derepressed mutant of B. subtilis showed the presence of two riboflavin synthase activities of greatly different sizes and molecular weights (3). The light enzyme, which accounts for more than 80% of the total activity, is a trimer of identical a subunits. The heavy enzyme is a complex of one molecule t Present address: Fachbereich Biologie, Universitit Frankfurt, 6000 Frankfurt/Main, German Federal Republic. tt Present address: Pfizer Laboratories, 75 Karlsruhe, German Federal Republic.
of light enzyme (i.e., three a subunits) and approximately 60 identical ,B subunits (3; A. Bacher, M. K. Otto, and H. Schnepple, unpublished data). Isolated ,B subunits had no riboflavin synthase activity, and the low specific activity of the heavy enzyme as compared to that of the light enzyme supports the hypothesis that only the a subunits are catalytically involved in the conversion of 6,7-dimethyl-8-(Dribityl)lumazine to riboflavin. The early steps of riboflavin biosynthesis are incompletely understood. The pathway starts at the level of guanosine or a respective nucleotide, as shown by isotope incorporation studies (7, 9, 20). The ribose moiety of the purine precursor is directly converted to the ribityl moiety of the vitamin. It has been suggested that the first committed step of the biosynthesis is catalyzed by GTP cyclohydrolase II in Escherichia coli. The enzyme catalyzes the simultaneous release of carbon-8 and of pyrophosphate from GTP (14). Several pyrimidine-type intermediates have been isolated in studies with riboflavindeficient mutants of Saccharomyces cerevisiae (Fig. 1) (5, 6, 19). The conversion of 5-amino-2,6dihydroxy-4-(D-ribitylamino)pyrimidine to 6,7dimethyl-8-(D-ribityl)lumazine requires the addition of a four-carbon moiety of unknown structure. Several studies suggested the involvement of acetoin or a biogenetically related compound (12, 16, 18). The hypothesis has been criticized by other authors (1, 15). Recent studies suggested the involvement of a pentose or tetrose 476
BIOSYNTHESIS OF RIBOFLAVIN
VOL. 134, 1978
OH
OH
OH
H2N
rib 1
477
N
H2N
HN
NH2
HN
N-NNH2
N
NH2
C H2
(P3)OH2C
I P) OH2CLo 0
H-C-OH H -C-OH
H
v
H-C-OH
OH OH
OH OH
C H20H
I
II
OH
rib 2
HN
N
N
H3C
N
H3C
N
/
N
rib 4
OH
NAO
CH2
C H2
rib 5
\/,/N
~H
H 3C
H3C
N
NAO
CH2
H-C-OH
H-C-OH
H-C-OH
H-C-OH
H-C-OH H-C-OH
H-C-OH H-C-OH
H-C-OH
CHH2OH FIG.
0
0 rib 3
H 2N
CH20H
CH20H
VI v IV 1. Biosynthesis of riboflavin (14,20). Genes involved in S. cerevisiae are indicated according to studies
by Oltnanns et al. (22,23). I, guanosine triphosphate; II, 2,5-diamino-6-hydroxy-4-(D-ribosylamino)pyrimidine 5'-phosphate; III, 2,5-diamino-6-hydroxy-4-(D-ribitylamino)pyrimidine; IV, 5-amino-2,6-dihydroxy-4-(D-ribitylamino)pyrimidine; V, 6,7-dimethyl-8-(D-ribityl)lumazine; VI, riboflavin.
(1, 2). The involvement of 6-methyl-7-(1',2'-dihydroxyethyl)-8-ribityliumazine as a precursor of 6,7-dimethyl-8-(D-ribityl)lumazine has been proposed on the basis of studies with mutants of B. subtilis (11). Several biosynthetic reactions in the pathway of riboflavin biosynthesis are not yet accessible to direct enzymatic studies in spite of considerable efforts. This paper presents evidence that the heavy riboflavin synthase of B. subtilis is a bifunctional enzyme. Whereas the a subunits mediate the known riboflavin synthase activity of the protein, the ,B subunits appear to be necessary for an earlier reaction in the biosynthesis of the vitamin. MATERIALS AND METHODS Bacterial strains. B. subtilis 168M trp2C
TABLE 1. Mutants of B. subtilis used0 Phenotype Origin RibH78 NGb H94 Flavinogenicc (4) H175 RibNG H322 Flavinogenic (4) RibCR2 ICRd RibICR CR6 RibICR CR9 'All mutants were derived from strain 168M and require tryptophan. b N-Methyl-N'-nitro-N-nitrosoguanidine mutagenesis. 'Genetically derepressed mutant accumulating riboflavin. d ICR 191 mutagenesis. Strain
was
pg/ml) ovemight (21). A portion was transferred to
kindly provided by C. Anagnostopoulos, Centre de Recherches Scientifiques, Gif-sur-Yvette, France. The other mutants used are listed in Table 1. Media. The basic medium was Spizizen minimal medium (30) supplemented with tryptophan (50 mg/liter). Vitamin-free Cmino Acids (Difco), riboflavin, and diacetyl were added as required. Chemicals. 6,7-Dimethyl4-8(D-ribityl)lumazine was prepared by published procedures (28). ICR 191 was a gift of Badische Anilin- und Soda-Fabrik, Ludwigshafen. Riboflavin was purchased from Merck AG, Darmstadt, W. Germany, and diacetyl was from Fluka AG, Buchs, Switzerland. Isolation of mutants. B. subtilis 168M was grown in complete medium supplemented with ICR 191 (20
fresh complete medium, and the culture was incubated overnight. Riboflavin-deficient mutants were isolated as described (4). Only one mutant of each respective phenotype was collected from each mutagenized culture. Growth of bacteria and preparation of cell extracts. Bacteria were grown in 0.5-liter batches of medium supplemented with vitamin-free Casamino Acids (5 g/liter). Riboflavin was added as required. The cultures were incubated with shaking overnight. The cells were harvested by centrifugation and stored at -20°C. For the preparation of cell extracts, frozen bacterial cells were thawed in buffer containing 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid, and 10 mM sodium sulfite. The suspension
478
J. BACTERIOL.
BACHER AND MAILANDER
was ultrasonically treated and centrifuged. Assay of riboflavin synthase. Assay mixtures contained 0.1 M phosphate (pH 7.4), 10 mM ethylenediaminetetraacetic acid, 10 mM sodium sulfite, and 0.6 mM 6,7-dimethyl-8-(D-ribityl)lumazine. Assays were performed at 37°C as described by Plaut and Harvey (28). One unit of enzyme activity catalyzes the formation of 1 nmol of riboflavin per h. Sucrose gradient centrifugation. Sucrose gradients (5 to 20%) contained 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid and 10 mM sodium sulfite. Samples of 0.5 ml were layered on top of the gradients. They were centrifuged in an SW27.1 rotor (Spinco) at 23,000 rpm and 4°C for 20 h. Fractions were collected and analyzed. Preparation ofantisera. Rabbits were immunized with 0.2- to 0.5-mg samples of light or heavy riboflavin synthase of B. subtilis (3) in 0.5 ml of Freund complete adjuvant (Difco). Serum obtained after immunization with heavy enzyme was made monospecific for ,8 subunits by addition of light riboflavin synthase. Immunodiffusion. Double-diffusion and simple radial diffusion experiments were performed according to published procedures (17, 24). Plates contained 0.1 M phosphate (pH 6.9), 10 mM ethylenediaminetetraacetic acid, 0.01% sodium azide, and 1.2% agar (Difco). Light and heavy riboflavin synthase, respectively, were used as standards in quantitative measurements of a and ,B subunits. Results were expressed in terms of enzyme units. Isolation of 6,7-dimethyl-8-(D-ribityl)lumazine. B. subtilis mutant CR2 was grown in 1-liter batches of medium supplemented with vitamin-free Casamino Acids (5 g/liter) and diacetyl (50 id/liter). The cultures were incubated at 37°C for 36 h. The cells were removed by centrifugation. 6,7-Dimethyl-8-(D-ribityl)lumazine was isolated from the culture medium as described (4). Miscellaneous methods. Protein concentration was estimated by the biuret method. Thin-layer chromatography was performed with plates coated with cellulose MN 300 (Macherey & Nagel, Diiren, W. Germany). The solvent systems were 1-butanol-acetic acid-water, 50:15:35 (vol/ vol/vol), and 3% ammonium chloride, respectively.
RESULTS Regulation of enzyme subunit biosynthesis. The level of light and heavy riboflavin synthase in cell extracts of B. subtilis mutants was analyzed by sucrose gradient centrifugation and by simple radial immunodiffusion using monospecific antisera (Table 2). In the flavinogenic mutants H94 and H322, the heavy enzyme accounted for approximately 15% of the total riboflavin synthase activity, as shown by sucrose gradient centrifugation. The values obtained by immunochemical determination of ,B subunits are in agreement within the experimental limits. Figure 2 shows the sucrose gradient centrifugation profile of mutant H94. The peak of f8 subunits coincides with the peak of heavy riboflavin synthase activity. Control experiments showed
TABLE 2. Concentration of light and heavy riboflavin synthase in cell extracts of B. subtilis mutants Riboflavin synthase Strain
Riboflavin' (mg/liter)
(U/mg of protein)
Total
Lightb Heavyb Heavy'
H94 0 107 89 18 18 0 80 12 H322 68 15 H175 0.03 59 51 8 5 H175 10 3 NDd ND 0.7 67 7 CR9 0.03 60 9 10 1.4 ND ND CR9 0.2 CR2 0.03 118 118
