INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Streptomycetes are known to produce primarily branched-chain fatty acids with only a minor proportion of straight-chain fatty acids (21). Considerable efforts have been made to study the initiation of fatty acid biosynthesis in streptomycetes and the precursors involved. Previous in vivo analysis of branched-chain fatty acid biosynthesis using stable isotope-labeled precursors showed that branched-chain fatty acids are produced by use of the amino acid catabolites isobutyryl coenzyme A (isobutyryl-CoA), 2-methylbutyryl-CoA, and 3-methylbutyryl-CoA as biosynthetic starter units (19, 26). In vitro fatty acid synthase (FAS) assays containing perdeuterated isobutyryl-CoA have confirmed that this functions as a starter unit for the biosynthesis of the branched-chain fatty acid isopalmitate (27). In vivo and in vitro experiments have also indicated that both acetyl-CoA and butyryl-CoA can be utilized as starter units in the synthesis of minor straight-chain fatty acids, palmitate and myristate (19, 26). Thiolactomycin, a known type II FAS inhibitor, appeared to inhibit branched-chain fatty acid biosynthesis and straight-chain fatty acid biosynthesis by using butyryl-CoA as a starter unit (27). Conversely, addition of the antibiotic appeared to stimulate production of straight-chain fatty acids with acetyl-CoA as a starter unit (27). Based on this observation, it has been suggested that streptomycetes have differential pathways for initiation of straight-chain and branched-chain fatty acid biosyntheses. However, none of the enzymes involved in either of these processes have been studied.

Preliminary studies have suggested that fatty acids in streptomycetes are synthesized by type I FAS (3, 20). Type I FASs are large multifunctional enzyme complexes generally found in yeast and mammalian systems. Recent studies, however, have clearly indicated that streptomycetes have a type II FAS, a multienzyme complex commonly found in bacteria and plants (16, 17, 23). A fabD gene encoding a malonyl-CoA:acyl carrier protein (malonyl-CoA:ACP) transacylase (MAT) has been identified in both Streptomyces glaucescens and Streptomyces coelicolor. The MAT appears to be responsible for catalyzing the production of malonyl-ACP that is required for each successive elongation step in fatty acid biosynthesis. The fabD gene is clustered with three other genes, fabH, fabC, and fabB, that encode proteins with high sequence similarity to the following components of the Escherichia coli type II FAS: the FabH -ketoacyl-ACP synthase III (KASIII), the FabC ACP, and the FabB KASI. In E. coli, FabH catalyzes the first step in fatty acid biosynthesis, which involves condensation of the acetyl-CoA starter unit with the first extender unit, malonyl-ACP, to form acetoacetyl-ACP. E. coli FabH also has acetyl-CoA:ACP transacylase (ACAT) activity, which catalyzes the formation of acetyl-ACP from acetyl-CoA and ACP (24). The specific activity of E. coli FabH ACAT activity is 0.5% of the KAS activity (24).

The role of FabH in fatty acid biosynthesis in streptomycetes, which produces a wide range of straight- and branched-chain fatty acids, has not been determined previously. We report herein the purification and characterization of the thiolactomycin-sensitive S. glaucescens FabH. Our results showed that Streptomyces FabH is able to utilize different acyl-CoA substrates, with a significantly lower K_m for the four-carbon acyl-CoA substrates butyryl-CoA and isobutyryl-CoA than for acetyl-CoA. These studies indicate that in Streptomyces, FabH is the enzyme responsible for initiating both branched- and straight-chain fatty acid biosyntheses. Results from a series of in vivo directed biosynthetic experiments using leucine and valine with the antibiotic thiolactomycin are shown to be consistent with this proposed role and to implicate the presence of a FabH-independent pathway for straight-chain fatty acid biosynthesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. The following chemicals were purchased from Sigma (St. Louis, Mo.): [1-¹⁴C]acetyl-CoA (specific activity, 46.3 mCi/mmol), E. coli FAS ACP, malonyl-CoA, CoA, acetylphosphate, phosphotransacetylase, S-acetyl-CoA synthetase, and ATP. [1-¹⁴C]butyryl-CoA (specific activity, 4 mCi/mmol) was from Moravek Biochemicals Inc. [1-¹⁴C]isobutyric acid (specific activity, 56 mCi/mol) was from ICN Radiochemicals. [2-¹⁴C]malonyl-CoA (specific activity, 53.0 mCi/mmol) was purchased from Amersham (Arlington Heights, Ill.). Thiolactomycin was provided by Pfizer Inc. Oligonucleotides were synthesized at the Biopolymer Lab at the University of Maryland at Baltimore. Restriction endonucleases and other enzymes were purchased from Bethesda Research Laboratories (Bethesda, Md.), New England Biolabs (Beverly, Mass.), and Perkin-Elmer (Branchburg, N.J.). Lysozyme was from Boehringer Mannheim (Bedford, Mass.).

Bacterial strains and plasmids. The bacterial strains and plasmids used in this study are listed in Table 1. E. coli BL21(DE3) (Novagen, Madison, Wis.) was grown in Luria-Bertani (LB) medium, at 37°C, supplemented with ampicillin (100 µg/ml) when necessary.

Expression of the S. glaucescens fabH gene in E. coli. The forward primer 5'-GCCGACCGACATATGTCGAAGATCAAGCC-3' and the reverse primer 5'-TTCTATCCAGATCTTGTGGCGGTGGG-3' were used to amplify the fabH gene with plasmid pWHM194 as a template (23). The primers created restriction sites NdeI at the codon for the N-terminal methionine and BglII downstream of the fabH stop codon. PCR was performed by using the GeneAmp XL-PCR kit from Perkin-Elmer. The PCR product was eluted from agarose gel by using Qiax (Qiagen, Chatsworth, Calif.) and cloned into pET15b, which had been double-restricted with NdeI and BamHI to create pLH14. pLH14 was used to transform E. coli BL21(DE3) to ampicillin resistance, and individual colonies were picked and grown in LB medium at 37°C for their ability to overexpress FabH upon isopropyl--D-thiogalactopyranoside (IPTG) induction. IPTG was added to a final concentration of 0.4 mM when the optical density at 595 nm of the culture reached 0.4 to 0.7, and the culture was incubated at 37°C for an additional 3 h. Cells were collected by centrifugation (10,000 × g, 4°C, 10 min). For analysis of expression, cells were suspended in 100 µl of a buffer that contained 25 mM Tris-HCl (pH 8), 25 mM EDTA, and 200 mM glucose and 100 µl of Laemmli loading buffer. The mixture was boiled for 5 min, and 10 µl of the mixture was loaded onto a sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS-12% PAGE) gel. Following electrophoresis, the gel was stained with Coomassie blue. For protein purification, the cells were lysed and purified by metal chelate chromatography by following the recommended procedures provided by Novagen.

Expression of the S. glaucescens fabD gene in E. coli. Two oligonucleotides, 5'-CCCAAGAGAGCATATGAGAGTGCTCGTACTCGTCGC-3' (forward primer) and 5'-CTTGGCGGGCTGGATCCTCGACATGGGTCGGTCGGC-3' (reverse primer), were used to amplify the fabD gene with pWHM19 as a template (23). Two unique restriction sites, NdeI at the 5' end and BamHI downstream of the fabD stop codon, were created. The PCR product was digested with NdeI and BamHI and ligated into pET15b to create pLH16. pLH16 was introduced into E. coli BL21(DE3) by transformation and selection for ampicillin resistance. Expression of fabD was induced by the addition of IPTG to a final concentration of 0.4 mM. Cells were collected by centrifugation (10,000 rpm, 4°C, 10 min) at 3 h post-IPTG induction and stored at 80°C. Analysis of protein expression and subsequent purification of FabD were carried out in a manner analogous to that described for FabH.

Preparation of malonyl-ACP. Malonyl-ACP was prepared by using purified histidine-tagged MAT (His tag-FabD) by a modification of a previously published protocol (4, 8, 10). A reaction mixture containing 350 µM malonyl-CoA, 110 µM ACP, 100 µg of His tag-FabD, 5 mM acetylphosphate, 10 U of phosphotransacetylase, and 50 mM potassium phosphate (pH 6.8) in a final volume of 1 ml was incubated at 37°C for 30 min (CoA generated in the MAT reaction is converted to acetyl-CoA by using the phosphotransacetylase and acetyl phosphate). The mixture was then applied to a Mono Q 5/5 column (Pharmacia Biotech, Alameda, Calif.) equilibrated with 0.2 M NaCl in 50 mM bis-Tris (pH 6.5). Malonyl-ACP was eluted with a linear gradient from 0.2 M NaCl in 50 mM bis-Tris (pH 6.5) to 100% 0.5 M NaCl in 50 mM bis-Tris (pH 6.5) over 60 min at a flow rate of 0.3 ml/min. The malonyl-ACP-containing fractions were pooled and subsequently concentrated and desalted by using a Microcon 3 microconcentrator (Amicon).

Preparation of [1-¹⁴C]isobutyryl-CoA. S-Acetyl-CoA synthetase (Sigma) was used to prepare [1-¹⁴C]isobutyryl-CoA. A reaction mixture containing 0.089 µM [1-¹⁴C]isobutyric acid, 0.134 µM CoA, 9.1 mM ATP, 4 mM MgCl₂, 70 mM potassium phosphate (pH 7.5), and 0.3 to 0.8 U of S-acetyl-CoA synthetase in a final volume of 150 µl was incubated at 37°C overnight. After incubation, the reaction mixture was acidified with HCl and extracted with diethyl ether. The water layer was removed, adjusted to pH 4, dried by vacuum centrifugation, and redissolved in water. The sample was then applied to a C₁₈ reverse-phase column preequilibrated with 50 mM phosphate buffer (pH 4) in 30% methanol. [1-¹⁴C]isobutyryl-CoA was eluted with the same buffer and dried by vacuum centrifugation.

Desalting of acyl-CoA substrates. [1-¹⁴C]acetyl-CoA, [1-¹⁴C]butyryl-CoA, and [1-¹⁴C]isobutyryl-CoA were desalted by applying each of the substrates to a G-10 Sephadex column (Sigma). The acyl-CoA substrates were eluted with distilled water. Fractions containing acyl-CoA compounds were pooled and dried by vacuum centrifugation. The pellet was dissolved in a small amount of distilled water.

ACAT and KAS assays. The ACAT transferase assay was carried out as described by Tsay et al. (24) and Gulliver and Slabas (4). The standard assay contained 15 µM [1-¹⁴C]acetyl-CoA (or [1-¹⁴C]butyryl-CoA or [1-¹⁴C]isobutyryl-CoA), 10 µM ACP, 100 mM imidazole-HCl (pH 7.0), and 0.8 µg of FabH in a final volume of 25 µl. The reaction mixture was incubated at 30°C for 10 min and terminated by the addition of 100 µl of ice-cold 10% trichloroacetic acid (TCA). Five microliters of a 10-mg/ml bovine serum albumin solution was added to the mixture as a carrier. Precipitation was completed by incubation on ice for 10 min, and the precipitate was collected by centrifugation at 10,000 rpm for 10 min. The pellet was washed with 10% TCA and finally resuspended in 100 µl of 2% SDS in 20 mM NaOH. The suspension was combined with 5 ml of scintillation cocktail and analyzed with a scintillation counter. The KAS assay was essentially identical to the ACAT assay except that ACP was replaced by malonyl-ACP (10 µM).

MAT assay. The MAT assay was carried out as previously described (4). Briefly, a reaction mixture containing 7.5 µM [2-¹⁴C]malonyl-CoA, 10 µM ACP, 100 mM imidazole-HCl (pH 7.0), and 5 µg of FabD in a final volume of 50 µl was incubated at 30°C for 10 min and the reaction was terminated by the addition of 6 µl of ice-cold 100% TCA. After 10 min of incubation on ice, the precipitate was collected by centrifugation at 10,000 rpm for 10 min. The pellet was washed first with 100 mM sodium citrate-HCl (pH 4.0) and then with water. The pellet was then resuspended in 100 µl of 100 mM imidazole-HCl (pH 7.4). The quantity of radioactive malonyl-ACP present in the pellet was determined with a scintillation counter, as described above.

Kinetic determinations. The K_m values of FabH for acetyl-CoA, butyryl-CoA, and isobutyryl-CoA were determined under the standard conditions with 5 µM malonyl-ACP and variable concentrations of the CoA substrates. Reactions were terminated after 2 and 4 min, and the rate was determined by plotting the extent of the reaction at 0, 2, and 4 min. The K_m of FabH for malonyl-ACP was determined by performing the KAS assay with 5 µM [1-¹⁴C]butyryl-CoA and variable concentrations of malonyl-ACP. Reactions were similarly terminated at 2 and 4 min.

In vitro effect of thiolactomycin on KAS activity of FabH. A thiolactomycin stock solution (10 mg/ml) was prepared as described previously (27). The required amount of thiolactomycin was aliquoted into an Eppendorf tube and allowed to air dry. The dried antibiotic was then resuspended in the assay buffer and incubated with FabH for 15 min. For a control, FabH was suspended in assay buffer for 15 min in the absence of thiolactomycin. KAS activities were then determined as described above.

Native molecular weight determination. Native molecular weights of FabH and FabD were estimated by using a Superose 6 HR 16/50 gel filtration column (Pharmacia Biotech, Alameda, Calif.). The column was preequilibrated with 100 mM potassium phosphate (pH 7.0) containing 5 mM EDTA. The column was calibrated by using dextran blue (M_r 2,000,000), horse heart ferritin (M_r 440,000), rabbit muscle pyruvate kinase (M_r 240,000), pig lactic dehydrogenase (M_r 140,000), bovine albumin (M_r 67,000), and horse heart cytochrome (M_r 12,500). The purified recombinant proteins were loaded onto the column and eluted at a flow rate of 0.4 ml/min. Fractions (0.5 ml) were collected and concentrated with Microcon 10 concentrators (Amicon) and subsequently assayed for KAS activity for FabH or MAT activity for FabD.

Directed biosynthesis of fatty acids by using leucine and valine. Fatty acid profiles of S. glaucescens were obtained as previously described (26, 27). Parallel fermentations were carried out in minimal medium (27) and with minimal media supplemented with either leucine or valine (100 mM). Each of these fermentations was also carried out in the presence of thiolactomycin (480 µM). Fatty acids were extracted from mycelia and analyzed as described previously (26).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Purification of FabH and FabD. To determine the role of FabH in the initiation of fatty acid biosynthesis, pure S. glaucescens FabH protein was required. A His tag-FabH fusion protein expression construct (pLH14) was made, and FabH was overexpressed in E. coli and purified to homogeneity by metal chelate column chromatography (Fig. 1). Since FabD was needed for preparation of the malonyl-ACP substrate for the KAS assay of FabH, a His tag-FabD fusion protein expression plasmid (pLH16) was also constructed. FabD was purified to homogeneity by metal chelate column chromatography (Fig. 1). The M_r (39,000) of the purified His tag-FabH estimated by SDS-PAGE analysis corresponded to the predicted mass of the protein based on the fabH sequence plus the M_r of the histidine tag (23). The apparent M_r of 36,000 of the purified His tag-FabD estimated by SDS-PAGE was slightly bigger than the expected molecular weight of the protein (M_r 34,000). Similar observations were reported for S. coelicolor FabD (17). The M_r estimated by gel filtration chromatography on Superose 6 was 72,000 ± 3,000 for the native FabH, and a M_r of 69,000 ± 3,000 was estimated for the native FabD. Thus, the data suggest homodimeric structures for both S. glaucescens FabH and FabD enzymes. The FabH enzymes of avocado (4) and spinach (2) also appear to be homodimeric.

View larger version (47K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE of the purified His tag-FabD and His tag-FabH.

ACAT and KAS activities of S. glaucescens FabH. The S. glaucescens fabH encodes a protein resembling E. coli FabH, an enzyme with both ACAT and KAS activities. The purified His tag-FabH was therefore assayed for both of these. The two assay conditions were identical except that in the KAS assay, malonyl-ACP (10 µM) was substituted for ACP (10 µM). An ACAT activity that was approximately 12% of the KAS activity was observed with each of the substrates, acetyl-CoA, butyryl-CoA, and isobutyryl-CoA. The ACAT activities of the E. coli and spinach FabH proteins are approximately 0.5 and 1%, respectively, of the KAS activity (2, 24).

Substrate specificity. S. glaucescens FabH exhibited typical saturation kinetics in response to increasing concentrations of acyl-CoA substrates. Apparent K_m values of 3.0 ± 0.4 µM for acetyl-CoA, 0.60 ± 0.05 µM for butyryl-CoA, and 0.40 ± 0.01 µM for isobutyryl-CoA were determined (Table 2). The order of reactivity of FabH with the various acyl-CoA substrates at saturation was butyryl-CoA > acetyl-CoA > isobutyryl-CoA (Table 2). FabH also exhibited saturation kinetics in response to increasing concentrations of malonyl-ACP, with an apparent K_m of 3.66 ± 0.8 µM.

Inhibition of the KAS activity of S. glaucescens FabH by thiolactomycin. The effect of thiolactomycin on the KAS activity of FabH was examined by incubating the enzyme with various concentrations of thiolactomycin followed by the addition of [1-¹⁴C]butyryl-CoA and malonyl-ACP at a final concentration of 5 µM each. Approximately 50% inhibition of the enzyme was effected by 20 µM thiolactomycin, and almost complete inhibition was observed at a concentration of 250 µM (Fig. 2).

View larger version (10K): [in this window] [in a new window]   FIG. 2.   Thiolactomycin inhibition of the KAS activity of FabH.

pH dependence. The pH optimum for the KAS reaction was measured for FabH. Assays were conducted with [1-¹⁴C]butyryl-CoA and malonyl-ACP at a final concentration of 5 µM each. Within the pH range of 6.6 to 7.7, no significant increase or decrease in enzyme activity was observed.

Directed biosynthesis of fatty acids by using leucine and valine. The broad specificity of the S. glaucescens FabH for various acyl-CoA substrates suggests that each of these may compete for binding to the enzyme and predicts that changes in the in vivo ratio of these substrates will directly influence the fatty acid profiles. Addition of valine (100 mM) gave rise to a dramatic increase in the ratio of isopalmitate (generated from the valine degradation product isobutyryl-CoA) to fatty acids generated from other acyl-CoA substrates (Fig. 3). Similarly, increasing the in vivo concentrations of 3-methylbutyryl-CoA by the addition of leucine caused a corresponding increase in the ratio of isopentadecanoate to other fatty acids. In each of these experiments, the addition of thiolactomycin did not significantly alter the relative ratios of the various branched-chain fatty acids produced by S. glaucescens but it did cause an increase in the ratio of straight-chain/branched-chain fatty acids (Fig. 3).

View larger version (27K): [in this window] [in a new window]   FIG. 3.   (A) Relative abundance of the major fatty acids produced by S. glaucescens grown in the presence and absence of thiolactomycin. (B and C) Results from parallel experiments using valine (B) and leucine (C) supplementation. Fatty acid abbreviations: iC15, isopentadecanoate; aiC15, anteisopentadecanoate; iC16, isopalmitate; C16, palmitate.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Fatty acid biosynthesis in dissociated (type II) FASs such as that found in E. coli produces straight-chain fatty acids by using acetyl-CoA as a starter unit (9). KASIII (FabH) catalyzes the condensation of acetyl-CoA with malonyl-ACP and initiates this process. This enzyme has a strong preference for acetyl-CoA over butyryl CoA as a substrate (its ability to react with a branched-chain acyl-CoA substrate such as isobutyryl-CoA has not been reported) (7). Recent evidence suggests that streptomycetes that produce both straight- and branched-chain fatty acids also have a type II FAS (16, 17, 23, 27). In these cases, however, acetyl-CoA and butyryl-CoA appear to be used as starter units for straight-chain fatty acid biosynthesis while isobutyryl-CoA and methylbutyryl-CoA are used for the biosynthesis of the branched-chain fatty acid units (Fig. 4) (19, 26, 27). The broad substrate specificity of the S. glaucescens FabH (binding with lower K_m values for butyryl-CoA and isobutyryl-CoA than for acetyl-CoA) are consistent with a role in initiating branched- and straight-chain fatty acid biosyntheses in these systems (Fig. 4). The ability of FabH to utilize methylbutyryl-CoA (the primer for the biosynthesis of branched-chain fatty acids containing an odd number of carbons) could not be determined since a radiolabeled form of this substrate was not readily available.

View larger version (11K): [in this window] [in a new window]   FIG. 4.   Proposed role of the S. glaucescens FabH in initiation of biosynthesis of branched-chain and straight-chain fatty acids containing an even number of carbons. Initiation of the biosynthesis of branched-chain fatty acids containing an odd number of carbons by using 2- and 3-methylbutyryl-CoA is also thought to be catalyzed by FabH. A FabH-independent pathway for initiation of straight-chain fatty acid biosynthesis is not shown.

The E. coli ACP was used in the assays of the S. glaucescens FabH and FabD proteins due to problems in producing sufficient quantities of the purified modified (holo) form of the S. glaucescens ACP (FabC) (18). ACPs are well conserved among type II FAS systems, and it has previously been shown that E. coli ACP is capable of interacting with heterologous FAS components. For instance, the E. coli ACP can be used to replace both the Bacillus subtilis ACP for FAS assays of B. subtilis cell extracts (1) and the Streptomyces ACP for similar assays with Streptomyces collinus cell extracts (27). The E. coli ACP has also been used previously to assay the avocado FabH protein and to prepare malonyl-ACP by using the avocado FabD protein (4). In this study, E. coli ACP has been shown to be an effective substrate for the S. glaucescens FabD and FabH proteins.

Thiolactomycin is known to be selective against type II FASs (5, 6, 14, 15, 22, 25), inhibiting KAS reactions by blocking malonyl-ACP binding (13). The FabH of avocado was completely inhibited by 40 µM thiolactomycin (50% inhibition was observed in the presence of 7.5 µM thiolactomycin) (4). A dose-response curve for the in vitro inhibition of E. coli FabH has been shown to correlate with the in vivo inhibition of fatty acid biosynthesis (approximately 50% inhibition of growth at approximately 75 µM thiolactomycin) (8). These data have been used to support the role of E. coli FabH as the physiologically relevant target of thiolactomycin, due to its position at the beginning of the fatty acid biosynthetic pathway (8). The S. glaucescens FabH likewise is inhibited by thiolactomycin (50% inhibition at a concentration of 20 µM). Surprisingly, in vivo data from this study and previous studies have shown that concentrations of approximately 480 µM did not significantly inhibit growth of either S. collinus or S. glaucescens (concentrations of 2 mM do, however, lead to almost complete inhibition of the growth of S. collinus) (27). The substantial differences in the thiolactomycin concentrations required to inhibit FabH in vitro and fatty acid biosynthesis in vivo might reflect poor transport of the antibiotic into the cell. However, it has previously been shown that fatty acid biosynthesis in a cell extract also requires relatively high concentrations of thiolactomycin (240 µM) to yield a 50% inhibition (27). Thus, one interpretation of these data is that while FabH is a target for thiolactomycin inhibition, an additional FabH-independent pathway (less sensitive to thiolactomycin) must operate for fatty acid biosynthesis to take place in S. glaucescens.

In vivo studies of fatty acid biosynthesis in S. collinus have shown that while thiolactomycin (480 µM) does not significantly inhibit fatty acid biosynthesis, it does produce an inhibition of branched-chain fatty acid biosynthesis and a stimulation of straight-chain fatty acid biosynthesis (27). This phenomenon has also been observed in this study for S. glaucescens and, taken together with the sensitivity of S. glaucescens FabH to thiolactomycin, supports the role of this enzyme as the major factor for controlling the initiation of branched-chain fatty acid biosynthesis (Fig. 4). Varying the in vivo ratio of different branched-chain acyl-CoA substrates by the addition of valine or leucine resulted in a corresponding increase in the relevant fatty acid, consistent with these substrates competing for binding to FabH (although other mechanisms that would lead to the observed shifts in fatty acid composition could also be proposed). The overall inhibition of branched-chain fatty acid biosynthesis by thiolactomycin in each of these experiments did not substantially change their relative ratios. Thus, if additional enzymes are involved in initiating branched-chain fatty acid biosynthesis, they would have to be similarly inhibited by thiolactomycin.

The FabH-independent pathway for straight-chain fatty acid biosynthesis in Streptomyces has not yet been delineated. For E. coli and plants, FabH-independent pathways for initiation of straight-chain fatty acid biosynthesis have been proposed (4, 11). In one pathway, acetyl-CoA is first transacylated to generate acetyl-ACP, which is then condensed with malonyl-ACP by FabB (KASI) to form acetoacetyl-ACP. In plants there appears to be a separate ACAT (in E. coli it appears that the only ACAT activity is that which can be attributed to FabH) (4, 11). A second pathway requires decarboxylation of malonyl-ACP to form acetyl-ACP, which is condensed with malonyl-ACP by FabB (25). Either of these processes may initiate straight-chain fatty acid biosynthesis in Streptomyces. The possibility that straight-chain fatty acid biosynthesis could use a different FAS (such as type I FAS) also cannot be ruled out at this point (16).

It is interesting to note that a cluster of fatty acid biosynthesis genes that has recently been cloned from Bacillus subtilis does not contain a fabH gene (12). B. subtilis, like streptomycetes, produces branched- and straight-chain fatty acids. We have observed that the addition of thiolactomycin to cultures of B. subtilis does not result in a change in the ratio of branched- and straight-chain fatty acids (unpublished results). Thus, it appears that the pathways for initiation of fatty acid biosynthesis in Streptomyces and B. subtilis differ.