JOURNAL
OF
Vol. 127, No. 3 Printed in U.S.A.
BACTERIOLOGY, Sept. 1976, p. 1136-1140
Copyright © 1976 American Society for Microbiology
Synthesis of co-Alicyclic Fatty Acids from Cyclic Precursors in Bacillus subtilis RENATE DREHER, KARL PORALLA,* AND WILFRIED A. KONIG Institut fur Biologie II, Lehrbereich Mikrobiologie I, Universitat Tubingen, D-74 Tubingen, Germany, and Institut fur Organische Chemie und Biochemie der Universitdt Hamburg, D-2 Hamburg 13, Germany Received for publication 16 March 1976
A mutant of Bacillus subtilis synthesizes a variety of c-alicyclic fatty acids when fed with the respective alicyclic carboxylic acids. These fatty acids are: wcyclopropane, c-cyclobutane, co-cyclopentane, o.)-cyclohexane, and co-cyclohexene fatty acids. These unusual fatty acids did not lead to an inhibition of growth at 37°C and pH 7. The selective advantage of these fatty acids under extreme conditions was studied in comparison with the acidophilic, thermophilic bacterium B. acidocaldarius, which normally contains a high proportion of co-cyclohexane fatty acids.
Two systems for the manipulation ofthe fatty acid pattern in membranes of bacteria are well known. These systems comprise fatty acid auxotrophs of Escherichia coli and acheloplasmas, some of which are naturally fatty acid dependent (11, 15). We have exploited a third system, a mutant of Bacillus subtilis, auxotrophic for short branched-chain fatty acids (16) and alicyclic carboxylic acids. These carboxylic acids are incorporated into the distal part of the synthesized fatty acids. Using this mutant of B. subtilis one can ask two questions. First, to what extent can the distal part of the fatty acids, and, thus, the inner part of the cytoplasmic membrane, be altered without inhibiting growth? The variation of the structure of the fatty acids presumably gives hints for their structural requirements in Bacillus species. Second, the bacterium B. acidocaldarius, which possesses a high percentage of co-cyclohexane fatty acids (2, 3), grows optimally at high temperatures (50 to 70°C) and at low pH (2 to 5). Therefore, does the possession of w-cyclohexane fatty acids also for B. subtilis provide a selective advantage for growth under special conditions?
Interference from revertants was prevented by doubling the acetate concentration in the growth media from 20 to 40 mg/liter and cloning the strain before all experiments. In addition controls were done after all experiments to test for the presence of revertants. Media. Cells were grown in a medium containing (per liter of deionized water): K2HPO4 3H20, 4.8 g; KH2PO4, 1.5 g; Na-citrate 2H20, 0.5 g; MgSO4 7 H20, 0.2 g; 113 mg of glutamic acid; 40 mg of Kacetate; 50 mg of L-tryptophan; 1 g of Casamino Acids (Difco); CaCl2 2H20, 2 mg; ZnSO4 7H20, 0.4 mg; MnSO4 H20, 0.4 mg; FeCl3 6H20, 0.2 mg. Where necessary, pH was adjusted with H2SO4 before autoclaving. After autoclaving, glucose was added (2 g/liter) together with 0.1 mM either filtersterilized sodium salts of each of the three short branched-chain fatty acids, DL-2-methylbutyric acid, isovaleric acid, and isobutyric acid or 0.2 mM where only a single fatty acid was added. Growth conditions. All cells were grown in 500 ml of indented Erlenmeyer flasks containing 100 ml of medium and incubated at 37°C with shaking. For fatty acid incorporation experiments, cells of an overnight culture of B. subtilis bfm 49 were washed 3 times with a salt solution (containing 9 g of NaCl and 246 mg of MgSO4-7H20 per liter), suspended in growth medium containing one of the precursors, and incubated for 7 h. For growth measurements, a bio-photorecorder (Temperature Gradient Bio-Photorecorder TN112M, Toyo Kagaku Sangyo Co., Ltd., Tokyo, Japan) was used. The cuvettes were filled with 18 ml of medium and inoculated with washed cells to an extinction of 0.1 at 546 nm (1-cm light path). The cuvettes were agitated at a frequency of 30 cycles/ min. Growth was followed by measuring the change in optical density. Antagonism test. In this test the organisms were seeded into agar medium lacking fatty acid precursors. Then two filter paper strips were placed perpendicular to each other on the agar. One of the
MATERIALS AND METHODS Strains. B. subtilis 168 Marburg strain was obtained from R. Schweyen and B. subtilis bfm 49 (met- trp- brafa- acet-) was obtained from K. Willecke. B. subtilis bfm 49 is a short branched-chain fatty acid-requiring (brafa-) mutant of strain 61141. Presumably this mutant has a similar defect in the branched-chain a-keto acid dehydrogenase (see reaction 1 in Fig. 1) as the mutant B. subtilis 626 described by Willecke and Pardee (16). Such a mutant can utilize short-chain fatty acids (see reaction 2 in Fig. 1) in a way described by Lennarz (10). 1136
co-ALICYCLIC FATTY ACIDS IN B. SUBTILWS
VOL. 127, 1976
Iso-C1
or
iso-C16 fatty acid CoA
Acetyl-CoA Isobutyryl-CoA
co2 2-Ketoisovaleric acid
Isobutyric acid
FIG. 1. Scheme for the synthesis of the primer molecule in branched-chain fatty acid biosynthesis in Bacillus species.
strips was presoaked in a solution of isobutyric acid (0.2 mM) and the other one was presoaked in a solution of the relevant fatty acid precursor (0.2 mM). When seen, growth inhibition was detectable where the two strips crossed each other and both precursors were available. Chemicals. Cycloheptanecarboxylic acid (99%) was purchased from EGA-Chemie, Steinheim, Germany, and cyclohexaneacetic acid (>98%) and 3cyclohexenecarboxylic acid (97%) were obtained from Fluka AG, Buchs, Switzerland. Cyclopropanecarboxylic acid (98.9%), cyclobutanecarboxylic acid (98%), cyclohexanecarboxylic acid (98%), and all other chemicals were obtained from E. Merck, Darmstadt, Germany. Cyclohexaneglyoxylic acid was prepared by catalytic hydrogenation of D-a-phenylaminoacetic acid and oxidation by 1)-amino acid oxidase. 3-Cyclohexeneglyoxylic acid was isolated as the antibiotic ketomycin from Streptomyces antibioticus (9). For identification of methylated fatty acids by gas chromatography, we used standard mixtures from Applied Science Laboratories Inc., State College, Pa. (branched-chain fatty acid methyl ester BC Mix-L and straight-chain fatty acids LA-206). Fatty acid extraction procedure and analysis. Cells were washed 3 times with the salt solution. Fatty acids were extracted according to a procedure slightly modified from that of Kaneda (7). The fatty acid extracts were evaporated to dryness, and the residues were methylated with diazomethane (12). The methyl esters were determined by gas-liquid chromatography using a Varian Aerograph 1440 with a hydrogen flame ionization detector (Varian, Walnut Creek, Calif.) on OV-17 columns (3% silicone on Chromosorb Q, 6 feet (ca. 183 cm), furnished by Varian) at 180°C, or Silar-lOC columns (3% Silar on Chromosorb Q, 6 feet (ca. 183 cm) Applied Science Laboratories Inc., State College, Pa., at 170°C. The carrier gas flow rate was 25 ml/min. The cyclic fatty acid methyl esters were identified by combined gas-liquid chromatography/mass spec-
1137
trometry. For complete separation of the components, glass capillaries, prepared according to the general procedure described by Schomburg et al. (13), were used. The stationary phases were SE 52 and SE 30. A Carlo-Erba 2101 model gas chromatograph was used. The mass spectra were obtained on a LKB 2091 GC-MS instrument equipped with a two-stage jet molecule separator and a 25-m glass capillary. The percentage of fatty acid methyl esters was determined by the weighing method (8).
RESULTS Growth experiments with alicyclic compounds. Various compounds were tested for their ability to support growth of B. subtilis bfm 49 (Table 1). The following compounds supported growth: cyclopropanecarboxylic acid (which supported slow growth using 0.2 mM, but nornal growth using 0.6 mM), cyclobutanecarboxylic acid, cyclopentanecarboxylic acid, cyclohexanecarboxylic acid, and 3-cyclohexenecarboxylic acid. However, a number of the compounds tested were unable to support growth, these being 2-cyclopenteneacetic acid, cyclopentaneacetic acid, cyclohexaneacetic acid, cyclooctanecarboxylic acid, and 4-cyclooctenecarboxylic acid. Cycloheptanecarboxylic acid permitted very slow growth of the mutant, although combined gas chromatography and mass spectrometry dil not reveal the presence of o-cycloheptane fatty acids. To show that these latter compounds at the concentrations used did not inhibit growth, a wild-type strain was grown under the same conditions as the mutant. With one exception, there was no inhibition of growth of the wild-type strain (Table 1). A control for growth inhibition by the cyclic TABLE 1. Growth ofB. subtilis bfm 49 and 168 Mat 37°C on different fatty acid precursors (0.2 mM) Growth rate of B. subtilis Precursor added
Cyclopropanecarboxylic acid Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid Cyclohexanecarboxylic acid 3-Cyclohexenecarboxylic acid Cycloheptanecarboxylic acid Cyclooctanecarboxylic acid 4-Cyclooctenecarboxylic acid Cyclopentaneacetic acid 2-Cyclopenteneacetic acid Cyclohexaneacetic acid Cyclohexaneglyoxylic acid
3-Cyclohexeneglyoxylic acid Isobutyric acid a Concentration 0.6 mM. b -, No growth. cNT, Not tested.
(doublings/h) bfm 49 168 M 1.42 0.71 (1.25)a 1.4 1.25 1.0 1.17
_b -
1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42
NTP
-
-
1.42
1.42
1138
J. BACTERIOL.
DREHER, PORALLA, AND KONIG
in the mutant strain was an antagonism test, in which the precursors were fed simultaneously with isobutyric acid. Isobutyric acid alone can satisfy the growth requirements of the short branched-chain fatty acid-requiring mutant bfm 49. However, in no case did the simultaneous addition of a cyclic precursor cause growth inhibition of this mutant. 3-Cyclohexeneglyoxylic acid and cyclohexaneglyoxylic acid did not support growth because of the apparent absence of 2-keto-acid dehydrogenase activity in this mutant. However, in ketomycin-resistant mutants of the wild-type strain, both 3-cyclohexeneglyoxylic acid and cyclohexaneglyoxylic acid were incorporated into the fatty acids as shown by the appearance of wcyclohexene and w-cyclohexane fatty acids with a chain length of C17 and C19. Incorporation experiments. All cyclic compounds mentioned in Table 1 were used. However, after 7 h of incubation, incorporation into cyclic fatty acids could only be found with cyclopropanecarboxylic acid, cyclobutanecarboxylic acid, cyclopentanecarboxylic acid, cyclohexanecarboxylic acid, and 3-cyclohexenecarboxylic acid as precursors. The fatty acids synthesized from the respective precursors and the percentage of total fatty acid methyl esters found are listed in Table 2. To show that these cyclic fatty acids were in precursors
fact incorporated into lipids, B. subtilis bfm 49 was grown on cyclohexanecarboxylic acid and the lipids were extracted and separated by thinlayer chromatography. The most ninhydrinpositive spot, presumably phosphatidylethanolamine, was extracted and hydrolyzed, and the fatty acids were methylated and identified by gas chromatography. We could show that a high percentage of cyclohexane fatty acids had been incorporated into this lipid (26%). Identification of the cyclic fatty acids. The fatty acid composition was investigated by gasliquid chromatography and the cyclic components were identified by combined gas-liquid chromatography/mass spectrometry. The mass spectra of w-cycloalkyl fatty acid methyl esters have been described by De Rosa and Gambacorta (6) for the cyclobutyl to cycloheptyl derivatives. Molecule ion intensities are found to decrease going from cyclohexyl to cyclopropyl derivatives according to the increasing ring tension. The cyclopropyl derivative does not show a molecular ion (Fig. 2). In this case, M-OCH3 and M-CH30H are the only significant ions in the upper mass range. The relative intensity of both of these ions increases from cyclohexyl to cyclopropyl fatty acid esters. We do not observe the loss of the cyclic portion by a-cleavage as described by De Rosa and Gambacorta (6), which would result in frag-
TABLE 2. Percentage and chain length ofw-alicyclic fatty acids synthesized by B. subtilis bfm 49 when grown on respective precursors
fatty acid methyl esters
% Total
Fatty acid synthesized
Precursor added
Cyclopropanecarboxylic acid Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid 3-Cyclohexenecarboxylic acid Cyclohexanecarboxylic acid a After a 4-h incubation period.
w-Cyclopropane-C,6
34 70 73 62
w-Cyclohexane-C,7 and C19
60a
w-Cyclobutane-C,5 and C,7 cw-Cyclopentane-C15 and C18 o-Cyclohexene-C,7 and C19
j7 >(CH2)1-COOCH3 inz
LI-
74
MW 268
z
41 wi
160
200
240
m/e
FIG. 2. Mass spectrum of*13-cyclopropyltridecanl oic acid
methyl ester.
co-ALICYCLIC FATTY ACIDS IN B. SUBTILlS
VOL. 127, 1976
TABLE
1139
3. Growth rates ofB. subtilis bfm 49 at different temperatures on various fatty acid precursors atpH 7 and pH 5 Growth rate (doublings/h)
30°C
Precursor added
Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid 3-Cyclohexenecarboxylic acid Cyclohexanecarboxylic acid Isobutyric acid
47C
42°C
37°C
pH7
pH5
pH7
pH5
pH7
pH5
pH7
pH5
0.33 0.76 0.95 0.71
-a
1.33 1.25 1.17 1.0 1.42
0.33 -
1.33 1.42 1.42 1.11 1.42
0.83 0.58 0.52 0.55 0.9
0.9 0.76 0.74 0.8 0.95
0.8
-
-
-
-
-
None a ,No growth.
ments M-55, M-69, and M-83, respectively. Instead, we observe intense ions at electronic mass 55, 69, 83, and 97 typical for alicyclic and olefmic compounds. In all cases, a strong McLafferty rearrangement ion is observed at electronic mass 74, accompanied by electronic mass 87. Temperature and pH dependence of growth on different precursors. Darland and Brock (2) first isolated B. acidocaldarius from acidic thermal environments. The same organism was isolated by De Rosa et al. (3). This organism is a thermophilic and acidophilic Bacillus that grows optimally at temperatures between 50 and 700C and at pH 2 to 5. It is also unusual in that it possesses a high proportion of cocyclohexane fatty acids (C07 and C,9) (3). In addition to these unusual fatty acids, typical Bacillus species fatty acids (iso-, anteiso-, and straight-chain fatty acids) were found. To see if the same relationship held true for B. subtilis bfm 49, we varied the fatty acid composition of the membrane by feeding with various cyclic precursors or with isobutyric acid and studied the growth rates under conditions of altered temperature and/or altered medium pH (Table 3). Growth on cyclohexanecarboxylic acid as a precursor altered the fatty acid pattern of B. subtilis bfm 49 to one similar to that of B. acidocaldarius, but did not enhance growth under acidic conditions or at higher temperatures. In fact, with only one exception, we found better growth with isobutyric acid as a precursor than with the various cyclic precursors. Only with cyclobutanecarboxylic acid as a precursor could the mutant grow under acidic conditions (pH 5) at 37°C. DISCUSSION The experiments show that B. subtilis can incorporate a variety of alicyclic precursors into 0)-cyclic fatty acids without growth impairment in comparison to a culture grown on isobutyric acid. These cyclic fatty acids only occur in this
-
-
0.83 -
bacterium when such a precursor is provided. With B. acidocaldarius, which normally contains a high proportion of w-cyclohexane fatty acids, De Rosa et al. (4) did similar incorporation experiments without measuring the effect on growth. In direct contrast to our results with B. subtilis, these authors found that cyclopropanecarboxylic and cyclobutanecarboxylic acid were not incorporated, whereas cyclopentaneacetic and cyclohexaneacetic acid were incorporated into cellular lipids (4). Possible explanations for this difference are that the transport system, the activating enzyme, or the fatty acid-synthesizing system of B. acidocaldarius prefers bulkier precursors. When the distal part of the fatty acids was considerably altered, growth was not inhibited. Thus, the iso- and anteiso-branching of the normal fatty acids in B. subtilis is not the only possible structure in the membrane lipids of the cells grown under normal growth conditions (370C, pH 7). From measurements of spin-label signals, one can deduce that the degree of order of the acyl chain in a double membrane decreases with the distance from the polar head group (14). So it is perhaps understandable that bulky groups at the end ofthe acyl chain cannot disturb the order of a double membrane. When temperature and pH were altered simultaneously two observations were made. First, cyclobutanecarboxylic acid was an unfavourable precursor at 300C and pH 7 (Table 3). Second, the same precursor at 370C and pH 5 fulfilled the requirements for growth better than isobutyric acid. These results suggest that in one case cyclic fatty acids may provide a selective advantage to B . subtilis at low pH and a relatively higher temperature. However, the alteration of the fatty acid composition of B. subtilis to resemble B. acidocaldarius does not allow B. subtilis to grow at pH 5 and 470C, even with a high proportion of co-alicyclic fatty acids in its membrane. The nature of lipids must also be considered and it is interesting that B. aci-
1140
DREHER, PORALLA, AND KONIG
docaldarius possesses a 5-time-higher proportion of glycolipids (T. A. Langworthy, Abstr. Annu. Meet. Am. Soc. Microbiol. 1975, K64, p. 157) as compared with B. subtilis (1). Measurements of the phase-transition points of cyclohexane fatty acid-containing lipids obtained from B. subtilis and B. acidocaldarius are in progress to determine possible differences in membrane properties. ACKNOWLEDGMENTS We are indebted to J. Eyem, LKB Produkter AB, Bromma, Sweden for running samples on the GC/MS combination. We thank R. Hancock for reading the manuscript and K. Willecke for sending the mutant bfm 49. This work was supported by Deutsche Forschungsgemeinschaft (SFB 76). LITERATURE CITED 1. Bishop, D. G., L. Rutberg, and B. Samuelson. 1967. The chemical composition of the cytoplasmic membrane ofBacillus subtilis. Eur. J. Biochem. 2:448-453. 2. Darland, G., and T. D. Brock. 1971. Bacillus acidocaldarius sp. nov., an acidophilic, thermophilic sporeforming bacterium. J. Gen. Microbiol. 67:9-15. 3. De Rosa, M., A. Gambacorta, L. Minale, and J. D. Bu'Lock. 1971. Cyclohexane fatty acids from a thermophilic bacterium. Chem. Commun. p. 1334. 4. De Rosa, M., A. Gambacorta, and J. D. Bu'Lock. 1974. Specificity effects in the biosynthesis of fatty acids in BaciUus acidocaldarius. Phytochemistry 13:905-910. 5. De Rosa, M., A. Gambacorta, and J. D. Bu'Lock. 1974. Effects of pH and temperature on the fatty acid composition of Bacillus acidocaldarius. J. Bacteriol. 117:212-214.
J. BACTERIOL. 6. De Rosa, M., and A. Gambacorta. 1975. Identification of natural and semisynthetic c-cycloalkyl fatty acids. Phytochemistry 14:209-210. 7. Kaneda, T. 1966. Biosynthesis of branched-chain fatty acids. IV. Factors affecting relative abundance of fatty acids produced by Bacillus subtilis. Can. J. Microbiol. 12:501-514. 8. Kates, M. 1972. Techniques of lipidology. North-Holland Publishing Co., Amsterdam. 9. Keller-Schierlein, W., K. Poralla, and H. Zahner. 1969. Stoffwechselprodukte von Mikroorganismen. 78. Isolierung, Identifizierung und Wirkungsweise von Ketomycin [(R)-3-Cyclohexenylglyoxyls&ure] und dessen Umwandlungsprodukt 3-Cyclohexenylglycin. Arch. Mikrobiol. 67:339-356. 10. Lennarz, W. J. 1963. A long-chain fatty acid acyl-CoA synthetase in Bacillus megaterium. Biochim. Biophys. Acta 73:335-337. 11. Machtiger, N. A., and C. F. Fox. 1973. Biochemistry of bacterial membranes. Annu. Rev. Biochem. 42:575600. 12. Schlenk, H., and J. L. Gellerman. 1960. Esterification of fatty acids with diazomethane on a small scale. Anal. Chem. 32:1412-1414. 13. Schomburg, G., H. Husman, and F. Weeke. 1974. Preparation, performance and special applications of glass capillary columns. J. Chromatogr. 99:63-79. 14. Seelig, J. 1972. Motion of spin labelled fatty acids in membrane structures, p. 267-279. In F. Kreuzer and J. F. G. Slegers (ed.), Biomembranes, vol. 3. Plenum Press, New York. 15. Silbert, D. F. 1975. Genetic modification of membrane lipid. Annu. Rev. Biochem. 44:315-339. 16. Willecke, K., and A. B. Pardee. 1971. Fatty acid-requiring mutant of Bacillus subtilis defective in branchedchain aketo acid dehydrogenase. J. Biol. Chem. 246:5264-5272.
OF
Vol. 127, No. 3 Printed in U.S.A.
BACTERIOLOGY, Sept. 1976, p. 1136-1140
Copyright © 1976 American Society for Microbiology
Synthesis of co-Alicyclic Fatty Acids from Cyclic Precursors in Bacillus subtilis RENATE DREHER, KARL PORALLA,* AND WILFRIED A. KONIG Institut fur Biologie II, Lehrbereich Mikrobiologie I, Universitat Tubingen, D-74 Tubingen, Germany, and Institut fur Organische Chemie und Biochemie der Universitdt Hamburg, D-2 Hamburg 13, Germany Received for publication 16 March 1976
A mutant of Bacillus subtilis synthesizes a variety of c-alicyclic fatty acids when fed with the respective alicyclic carboxylic acids. These fatty acids are: wcyclopropane, c-cyclobutane, co-cyclopentane, o.)-cyclohexane, and co-cyclohexene fatty acids. These unusual fatty acids did not lead to an inhibition of growth at 37°C and pH 7. The selective advantage of these fatty acids under extreme conditions was studied in comparison with the acidophilic, thermophilic bacterium B. acidocaldarius, which normally contains a high proportion of co-cyclohexane fatty acids.
Two systems for the manipulation ofthe fatty acid pattern in membranes of bacteria are well known. These systems comprise fatty acid auxotrophs of Escherichia coli and acheloplasmas, some of which are naturally fatty acid dependent (11, 15). We have exploited a third system, a mutant of Bacillus subtilis, auxotrophic for short branched-chain fatty acids (16) and alicyclic carboxylic acids. These carboxylic acids are incorporated into the distal part of the synthesized fatty acids. Using this mutant of B. subtilis one can ask two questions. First, to what extent can the distal part of the fatty acids, and, thus, the inner part of the cytoplasmic membrane, be altered without inhibiting growth? The variation of the structure of the fatty acids presumably gives hints for their structural requirements in Bacillus species. Second, the bacterium B. acidocaldarius, which possesses a high percentage of co-cyclohexane fatty acids (2, 3), grows optimally at high temperatures (50 to 70°C) and at low pH (2 to 5). Therefore, does the possession of w-cyclohexane fatty acids also for B. subtilis provide a selective advantage for growth under special conditions?
Interference from revertants was prevented by doubling the acetate concentration in the growth media from 20 to 40 mg/liter and cloning the strain before all experiments. In addition controls were done after all experiments to test for the presence of revertants. Media. Cells were grown in a medium containing (per liter of deionized water): K2HPO4 3H20, 4.8 g; KH2PO4, 1.5 g; Na-citrate 2H20, 0.5 g; MgSO4 7 H20, 0.2 g; 113 mg of glutamic acid; 40 mg of Kacetate; 50 mg of L-tryptophan; 1 g of Casamino Acids (Difco); CaCl2 2H20, 2 mg; ZnSO4 7H20, 0.4 mg; MnSO4 H20, 0.4 mg; FeCl3 6H20, 0.2 mg. Where necessary, pH was adjusted with H2SO4 before autoclaving. After autoclaving, glucose was added (2 g/liter) together with 0.1 mM either filtersterilized sodium salts of each of the three short branched-chain fatty acids, DL-2-methylbutyric acid, isovaleric acid, and isobutyric acid or 0.2 mM where only a single fatty acid was added. Growth conditions. All cells were grown in 500 ml of indented Erlenmeyer flasks containing 100 ml of medium and incubated at 37°C with shaking. For fatty acid incorporation experiments, cells of an overnight culture of B. subtilis bfm 49 were washed 3 times with a salt solution (containing 9 g of NaCl and 246 mg of MgSO4-7H20 per liter), suspended in growth medium containing one of the precursors, and incubated for 7 h. For growth measurements, a bio-photorecorder (Temperature Gradient Bio-Photorecorder TN112M, Toyo Kagaku Sangyo Co., Ltd., Tokyo, Japan) was used. The cuvettes were filled with 18 ml of medium and inoculated with washed cells to an extinction of 0.1 at 546 nm (1-cm light path). The cuvettes were agitated at a frequency of 30 cycles/ min. Growth was followed by measuring the change in optical density. Antagonism test. In this test the organisms were seeded into agar medium lacking fatty acid precursors. Then two filter paper strips were placed perpendicular to each other on the agar. One of the
MATERIALS AND METHODS Strains. B. subtilis 168 Marburg strain was obtained from R. Schweyen and B. subtilis bfm 49 (met- trp- brafa- acet-) was obtained from K. Willecke. B. subtilis bfm 49 is a short branched-chain fatty acid-requiring (brafa-) mutant of strain 61141. Presumably this mutant has a similar defect in the branched-chain a-keto acid dehydrogenase (see reaction 1 in Fig. 1) as the mutant B. subtilis 626 described by Willecke and Pardee (16). Such a mutant can utilize short-chain fatty acids (see reaction 2 in Fig. 1) in a way described by Lennarz (10). 1136
co-ALICYCLIC FATTY ACIDS IN B. SUBTILWS
VOL. 127, 1976
Iso-C1
or
iso-C16 fatty acid CoA
Acetyl-CoA Isobutyryl-CoA
co2 2-Ketoisovaleric acid
Isobutyric acid
FIG. 1. Scheme for the synthesis of the primer molecule in branched-chain fatty acid biosynthesis in Bacillus species.
strips was presoaked in a solution of isobutyric acid (0.2 mM) and the other one was presoaked in a solution of the relevant fatty acid precursor (0.2 mM). When seen, growth inhibition was detectable where the two strips crossed each other and both precursors were available. Chemicals. Cycloheptanecarboxylic acid (99%) was purchased from EGA-Chemie, Steinheim, Germany, and cyclohexaneacetic acid (>98%) and 3cyclohexenecarboxylic acid (97%) were obtained from Fluka AG, Buchs, Switzerland. Cyclopropanecarboxylic acid (98.9%), cyclobutanecarboxylic acid (98%), cyclohexanecarboxylic acid (98%), and all other chemicals were obtained from E. Merck, Darmstadt, Germany. Cyclohexaneglyoxylic acid was prepared by catalytic hydrogenation of D-a-phenylaminoacetic acid and oxidation by 1)-amino acid oxidase. 3-Cyclohexeneglyoxylic acid was isolated as the antibiotic ketomycin from Streptomyces antibioticus (9). For identification of methylated fatty acids by gas chromatography, we used standard mixtures from Applied Science Laboratories Inc., State College, Pa. (branched-chain fatty acid methyl ester BC Mix-L and straight-chain fatty acids LA-206). Fatty acid extraction procedure and analysis. Cells were washed 3 times with the salt solution. Fatty acids were extracted according to a procedure slightly modified from that of Kaneda (7). The fatty acid extracts were evaporated to dryness, and the residues were methylated with diazomethane (12). The methyl esters were determined by gas-liquid chromatography using a Varian Aerograph 1440 with a hydrogen flame ionization detector (Varian, Walnut Creek, Calif.) on OV-17 columns (3% silicone on Chromosorb Q, 6 feet (ca. 183 cm), furnished by Varian) at 180°C, or Silar-lOC columns (3% Silar on Chromosorb Q, 6 feet (ca. 183 cm) Applied Science Laboratories Inc., State College, Pa., at 170°C. The carrier gas flow rate was 25 ml/min. The cyclic fatty acid methyl esters were identified by combined gas-liquid chromatography/mass spec-
1137
trometry. For complete separation of the components, glass capillaries, prepared according to the general procedure described by Schomburg et al. (13), were used. The stationary phases were SE 52 and SE 30. A Carlo-Erba 2101 model gas chromatograph was used. The mass spectra were obtained on a LKB 2091 GC-MS instrument equipped with a two-stage jet molecule separator and a 25-m glass capillary. The percentage of fatty acid methyl esters was determined by the weighing method (8).
RESULTS Growth experiments with alicyclic compounds. Various compounds were tested for their ability to support growth of B. subtilis bfm 49 (Table 1). The following compounds supported growth: cyclopropanecarboxylic acid (which supported slow growth using 0.2 mM, but nornal growth using 0.6 mM), cyclobutanecarboxylic acid, cyclopentanecarboxylic acid, cyclohexanecarboxylic acid, and 3-cyclohexenecarboxylic acid. However, a number of the compounds tested were unable to support growth, these being 2-cyclopenteneacetic acid, cyclopentaneacetic acid, cyclohexaneacetic acid, cyclooctanecarboxylic acid, and 4-cyclooctenecarboxylic acid. Cycloheptanecarboxylic acid permitted very slow growth of the mutant, although combined gas chromatography and mass spectrometry dil not reveal the presence of o-cycloheptane fatty acids. To show that these latter compounds at the concentrations used did not inhibit growth, a wild-type strain was grown under the same conditions as the mutant. With one exception, there was no inhibition of growth of the wild-type strain (Table 1). A control for growth inhibition by the cyclic TABLE 1. Growth ofB. subtilis bfm 49 and 168 Mat 37°C on different fatty acid precursors (0.2 mM) Growth rate of B. subtilis Precursor added
Cyclopropanecarboxylic acid Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid Cyclohexanecarboxylic acid 3-Cyclohexenecarboxylic acid Cycloheptanecarboxylic acid Cyclooctanecarboxylic acid 4-Cyclooctenecarboxylic acid Cyclopentaneacetic acid 2-Cyclopenteneacetic acid Cyclohexaneacetic acid Cyclohexaneglyoxylic acid
3-Cyclohexeneglyoxylic acid Isobutyric acid a Concentration 0.6 mM. b -, No growth. cNT, Not tested.
(doublings/h) bfm 49 168 M 1.42 0.71 (1.25)a 1.4 1.25 1.0 1.17
_b -
1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42
NTP
-
-
1.42
1.42
1138
J. BACTERIOL.
DREHER, PORALLA, AND KONIG
in the mutant strain was an antagonism test, in which the precursors were fed simultaneously with isobutyric acid. Isobutyric acid alone can satisfy the growth requirements of the short branched-chain fatty acid-requiring mutant bfm 49. However, in no case did the simultaneous addition of a cyclic precursor cause growth inhibition of this mutant. 3-Cyclohexeneglyoxylic acid and cyclohexaneglyoxylic acid did not support growth because of the apparent absence of 2-keto-acid dehydrogenase activity in this mutant. However, in ketomycin-resistant mutants of the wild-type strain, both 3-cyclohexeneglyoxylic acid and cyclohexaneglyoxylic acid were incorporated into the fatty acids as shown by the appearance of wcyclohexene and w-cyclohexane fatty acids with a chain length of C17 and C19. Incorporation experiments. All cyclic compounds mentioned in Table 1 were used. However, after 7 h of incubation, incorporation into cyclic fatty acids could only be found with cyclopropanecarboxylic acid, cyclobutanecarboxylic acid, cyclopentanecarboxylic acid, cyclohexanecarboxylic acid, and 3-cyclohexenecarboxylic acid as precursors. The fatty acids synthesized from the respective precursors and the percentage of total fatty acid methyl esters found are listed in Table 2. To show that these cyclic fatty acids were in precursors
fact incorporated into lipids, B. subtilis bfm 49 was grown on cyclohexanecarboxylic acid and the lipids were extracted and separated by thinlayer chromatography. The most ninhydrinpositive spot, presumably phosphatidylethanolamine, was extracted and hydrolyzed, and the fatty acids were methylated and identified by gas chromatography. We could show that a high percentage of cyclohexane fatty acids had been incorporated into this lipid (26%). Identification of the cyclic fatty acids. The fatty acid composition was investigated by gasliquid chromatography and the cyclic components were identified by combined gas-liquid chromatography/mass spectrometry. The mass spectra of w-cycloalkyl fatty acid methyl esters have been described by De Rosa and Gambacorta (6) for the cyclobutyl to cycloheptyl derivatives. Molecule ion intensities are found to decrease going from cyclohexyl to cyclopropyl derivatives according to the increasing ring tension. The cyclopropyl derivative does not show a molecular ion (Fig. 2). In this case, M-OCH3 and M-CH30H are the only significant ions in the upper mass range. The relative intensity of both of these ions increases from cyclohexyl to cyclopropyl fatty acid esters. We do not observe the loss of the cyclic portion by a-cleavage as described by De Rosa and Gambacorta (6), which would result in frag-
TABLE 2. Percentage and chain length ofw-alicyclic fatty acids synthesized by B. subtilis bfm 49 when grown on respective precursors
fatty acid methyl esters
% Total
Fatty acid synthesized
Precursor added
Cyclopropanecarboxylic acid Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid 3-Cyclohexenecarboxylic acid Cyclohexanecarboxylic acid a After a 4-h incubation period.
w-Cyclopropane-C,6
34 70 73 62
w-Cyclohexane-C,7 and C19
60a
w-Cyclobutane-C,5 and C,7 cw-Cyclopentane-C15 and C18 o-Cyclohexene-C,7 and C19
j7 >(CH2)1-COOCH3 inz
LI-
74
MW 268
z
41 wi
160
200
240
m/e
FIG. 2. Mass spectrum of*13-cyclopropyltridecanl oic acid
methyl ester.
co-ALICYCLIC FATTY ACIDS IN B. SUBTILlS
VOL. 127, 1976
TABLE
1139
3. Growth rates ofB. subtilis bfm 49 at different temperatures on various fatty acid precursors atpH 7 and pH 5 Growth rate (doublings/h)
30°C
Precursor added
Cyclobutanecarboxylic acid Cyclopentanecarboxylic acid 3-Cyclohexenecarboxylic acid Cyclohexanecarboxylic acid Isobutyric acid
47C
42°C
37°C
pH7
pH5
pH7
pH5
pH7
pH5
pH7
pH5
0.33 0.76 0.95 0.71
-a
1.33 1.25 1.17 1.0 1.42
0.33 -
1.33 1.42 1.42 1.11 1.42
0.83 0.58 0.52 0.55 0.9
0.9 0.76 0.74 0.8 0.95
0.8
-
-
-
-
-
None a ,No growth.
ments M-55, M-69, and M-83, respectively. Instead, we observe intense ions at electronic mass 55, 69, 83, and 97 typical for alicyclic and olefmic compounds. In all cases, a strong McLafferty rearrangement ion is observed at electronic mass 74, accompanied by electronic mass 87. Temperature and pH dependence of growth on different precursors. Darland and Brock (2) first isolated B. acidocaldarius from acidic thermal environments. The same organism was isolated by De Rosa et al. (3). This organism is a thermophilic and acidophilic Bacillus that grows optimally at temperatures between 50 and 700C and at pH 2 to 5. It is also unusual in that it possesses a high proportion of cocyclohexane fatty acids (C07 and C,9) (3). In addition to these unusual fatty acids, typical Bacillus species fatty acids (iso-, anteiso-, and straight-chain fatty acids) were found. To see if the same relationship held true for B. subtilis bfm 49, we varied the fatty acid composition of the membrane by feeding with various cyclic precursors or with isobutyric acid and studied the growth rates under conditions of altered temperature and/or altered medium pH (Table 3). Growth on cyclohexanecarboxylic acid as a precursor altered the fatty acid pattern of B. subtilis bfm 49 to one similar to that of B. acidocaldarius, but did not enhance growth under acidic conditions or at higher temperatures. In fact, with only one exception, we found better growth with isobutyric acid as a precursor than with the various cyclic precursors. Only with cyclobutanecarboxylic acid as a precursor could the mutant grow under acidic conditions (pH 5) at 37°C. DISCUSSION The experiments show that B. subtilis can incorporate a variety of alicyclic precursors into 0)-cyclic fatty acids without growth impairment in comparison to a culture grown on isobutyric acid. These cyclic fatty acids only occur in this
-
-
0.83 -
bacterium when such a precursor is provided. With B. acidocaldarius, which normally contains a high proportion of w-cyclohexane fatty acids, De Rosa et al. (4) did similar incorporation experiments without measuring the effect on growth. In direct contrast to our results with B. subtilis, these authors found that cyclopropanecarboxylic and cyclobutanecarboxylic acid were not incorporated, whereas cyclopentaneacetic and cyclohexaneacetic acid were incorporated into cellular lipids (4). Possible explanations for this difference are that the transport system, the activating enzyme, or the fatty acid-synthesizing system of B. acidocaldarius prefers bulkier precursors. When the distal part of the fatty acids was considerably altered, growth was not inhibited. Thus, the iso- and anteiso-branching of the normal fatty acids in B. subtilis is not the only possible structure in the membrane lipids of the cells grown under normal growth conditions (370C, pH 7). From measurements of spin-label signals, one can deduce that the degree of order of the acyl chain in a double membrane decreases with the distance from the polar head group (14). So it is perhaps understandable that bulky groups at the end ofthe acyl chain cannot disturb the order of a double membrane. When temperature and pH were altered simultaneously two observations were made. First, cyclobutanecarboxylic acid was an unfavourable precursor at 300C and pH 7 (Table 3). Second, the same precursor at 370C and pH 5 fulfilled the requirements for growth better than isobutyric acid. These results suggest that in one case cyclic fatty acids may provide a selective advantage to B . subtilis at low pH and a relatively higher temperature. However, the alteration of the fatty acid composition of B. subtilis to resemble B. acidocaldarius does not allow B. subtilis to grow at pH 5 and 470C, even with a high proportion of co-alicyclic fatty acids in its membrane. The nature of lipids must also be considered and it is interesting that B. aci-
1140
DREHER, PORALLA, AND KONIG
docaldarius possesses a 5-time-higher proportion of glycolipids (T. A. Langworthy, Abstr. Annu. Meet. Am. Soc. Microbiol. 1975, K64, p. 157) as compared with B. subtilis (1). Measurements of the phase-transition points of cyclohexane fatty acid-containing lipids obtained from B. subtilis and B. acidocaldarius are in progress to determine possible differences in membrane properties. ACKNOWLEDGMENTS We are indebted to J. Eyem, LKB Produkter AB, Bromma, Sweden for running samples on the GC/MS combination. We thank R. Hancock for reading the manuscript and K. Willecke for sending the mutant bfm 49. This work was supported by Deutsche Forschungsgemeinschaft (SFB 76). LITERATURE CITED 1. Bishop, D. G., L. Rutberg, and B. Samuelson. 1967. The chemical composition of the cytoplasmic membrane ofBacillus subtilis. Eur. J. Biochem. 2:448-453. 2. Darland, G., and T. D. Brock. 1971. Bacillus acidocaldarius sp. nov., an acidophilic, thermophilic sporeforming bacterium. J. Gen. Microbiol. 67:9-15. 3. De Rosa, M., A. Gambacorta, L. Minale, and J. D. Bu'Lock. 1971. Cyclohexane fatty acids from a thermophilic bacterium. Chem. Commun. p. 1334. 4. De Rosa, M., A. Gambacorta, and J. D. Bu'Lock. 1974. Specificity effects in the biosynthesis of fatty acids in BaciUus acidocaldarius. Phytochemistry 13:905-910. 5. De Rosa, M., A. Gambacorta, and J. D. Bu'Lock. 1974. Effects of pH and temperature on the fatty acid composition of Bacillus acidocaldarius. J. Bacteriol. 117:212-214.
J. BACTERIOL. 6. De Rosa, M., and A. Gambacorta. 1975. Identification of natural and semisynthetic c-cycloalkyl fatty acids. Phytochemistry 14:209-210. 7. Kaneda, T. 1966. Biosynthesis of branched-chain fatty acids. IV. Factors affecting relative abundance of fatty acids produced by Bacillus subtilis. Can. J. Microbiol. 12:501-514. 8. Kates, M. 1972. Techniques of lipidology. North-Holland Publishing Co., Amsterdam. 9. Keller-Schierlein, W., K. Poralla, and H. Zahner. 1969. Stoffwechselprodukte von Mikroorganismen. 78. Isolierung, Identifizierung und Wirkungsweise von Ketomycin [(R)-3-Cyclohexenylglyoxyls&ure] und dessen Umwandlungsprodukt 3-Cyclohexenylglycin. Arch. Mikrobiol. 67:339-356. 10. Lennarz, W. J. 1963. A long-chain fatty acid acyl-CoA synthetase in Bacillus megaterium. Biochim. Biophys. Acta 73:335-337. 11. Machtiger, N. A., and C. F. Fox. 1973. Biochemistry of bacterial membranes. Annu. Rev. Biochem. 42:575600. 12. Schlenk, H., and J. L. Gellerman. 1960. Esterification of fatty acids with diazomethane on a small scale. Anal. Chem. 32:1412-1414. 13. Schomburg, G., H. Husman, and F. Weeke. 1974. Preparation, performance and special applications of glass capillary columns. J. Chromatogr. 99:63-79. 14. Seelig, J. 1972. Motion of spin labelled fatty acids in membrane structures, p. 267-279. In F. Kreuzer and J. F. G. Slegers (ed.), Biomembranes, vol. 3. Plenum Press, New York. 15. Silbert, D. F. 1975. Genetic modification of membrane lipid. Annu. Rev. Biochem. 44:315-339. 16. Willecke, K., and A. B. Pardee. 1971. Fatty acid-requiring mutant of Bacillus subtilis defective in branchedchain aketo acid dehydrogenase. J. Biol. Chem. 246:5264-5272.
