Enoyl-acyl carrier protein reductase (fabI) plays a determinant role in completing cycles of fatty acid elongation in Escherichia coli.

The role of enoyl-acyl carrier protein (ACP) reductase (E.C. 1.3.1.9), the product of the fabI gene, was investigated in the type II, dissociated, fatty acid synthase system of Escherichia coli. All of the proteins required to catalyze one cycle of fatty acid synthesis from acetyl-CoA plus malonyl-CoA to butyryl-ACP in vitro were purified. These proteins were malonyl-CoA:ACP transacylase (fabD), beta-ketoacyl-ACP synthase III (fabH), beta-ketoacyl-ACP reductase (fabG), beta-hydroxydecanoyl-ACP dehydrase (fabA), and enoyl-ACP reductase (fabI). Unlike the other enzymes in the cycle, FabA did not efficiently convert its substrate beta-hydroxybutyryl-ACP to crotonyl-ACP, but rather the equilibrium favored formation of beta-hydroxybutyryl-ACP over crotonyl-ACP by a ratio of 9:1. The amount of butyryl-ACP formed depended on the amount of FabI protein added to the assay. Extracts from fabI(Ts) mutants accumulated beta-hydroxybutyryl-ACP, and the addition of FabI protein to the fabI(Ts) extract restored both butyryl-ACP and long-chain acyl-ACP synthesis. FabI was verified to be the only enoyl-ACP reductase required for the synthesis of fatty acids by demonstrating that purified FabI was required for the elongation of both long-chain saturated and unsaturated fatty acids. These results were corroborated by analysis of the intracellular ACP pool composition in fabI(Ts) mutants that showed beta-hydroxybutyryl-ACP and crotonyl-ACP accumulated at the nonpermissive temperature in the same ratio found in the fabI(Ts) extracts and in the in vitro reconstruction experiments that lacked FabI. We conclude that FabI is the only enoyl-ACP reductase involved in fatty acid synthesis in E. coli and that the activity of this enzyme plays a determinant role in completing cycles of fatty acid biosynthesis.

The first step is condensation of malonyl-ACP 1 with either acetyl-CoA to initiate fatty acid synthesis or the growing acyl chain to continue cycles of elongation. The first cycle of elongation is initiated by the condensation of acetyl-CoA with malonyl-ACP catalyzed by ␤-ketoacyl-ACP synthase III (FabH), the product of the fabH gene. Subsequent cycles of elongation are catalyzed by condensing enzyme I (FabB) or II (FabF). Recently, the existence of a fourth condensing enzyme was proposed (2). However, the map position and sequence of this putative new enzyme are the same as that of the fabF gene (3); therefore there is little concrete evidence for the presence of a fourth condensing enzyme. The ␤-ketoacyl-ACP is reduced by a NADPH-dependent ␤-ketoacyl-ACP reductase (FabG). Only a single enzyme is known to be responsible for this step, although the existence of other isozymes cannot be ruled out. There are two ␤-hydroxyacyl-ACP dehydrases (FabA and FabZ) capable of forming trans-2-enoyl-ACP. The product of the fabA gene is specifically involved in the introduction of a cis-double bond into the growing acyl chain at the ␤-hydroxydecanoyl-ACP step. The spectrum of reactions catalyzed by the FabZ dehydrase is unknown, but the fact that mutations in the fabZ gene were isolated based on their ability to suppress the temperature-sensitive growth phenotype in lpxA2(Ts) (␤-hydroxymyristoyl-ACP:UDP-N-acetylglucosamine acyltransferase) mutants indicates that it plays an important role in long-chain saturated fatty acid formation (4). The last reaction in each elongation cycle is catalyzed by enoyl-ACP reductase (FabI). E. coli is thought to possess two enoyl-ACP reductases, one NADH-dependent and the other NADPH-dependent (5), suggesting that more than one isozyme catalyzes this step of fatty acid biosynthesis.
The gene encoding the NADH-dependent enoyl-ACP reductase (fabI) was discovered as an outgrowth of research on the mechanism of action of diazaborines, a class of heterocyclic antibiotics. Biosynthesis of fatty acids and phospholipids is inhibited when E. coli or Salmonella typhimurium are treated with diazaborines, and resistance to these antibiotics is associated with an allelic form of the envM gene product (6). The envM gene was first identified as a temperature-sensitive mutation with an osmotically-repairable membrane defect (7). Because the EnvM protein, purified from an overproducing strain, possesses NADH-dependent enoyl-ACP reductase activity and binds radiolabeled diazaborine, the gene was renamed fabI (8). The fabI (envM) gene has been sequenced, and the diazaborine resistance is associated with a point mutation in this locus (9). Similarly, the product of the fabI analog in Mycobacterium tuberculosis, the inhA gene product, is the target for the antibiotics isoniazid and ethionamide, and a single point mutation * This work was supported by National Institutes of Health Grant GM34496, Cancer Center (CORE) Support Grant CA 21765, and funds from the American and Lebanese Syrian Associated Charities. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  in the inhA gene confers resistance to both antibiotics in M. tuberculosis (10). InhA has recently been purified, crystallized, and shown to be an NADH-dependent enoyl-ACP reductase (11). The focus of this study is to determine the number of enoyl-ACP reductase enzymes in E. coli and to define the role of enoyl-ACP reductase in regulating cycles of fatty acid elongation.

EXPERIMENTAL PROCEDURES
Materials-␤-[3-3 H]Alanine (specific activity, 56 Ci/mmol) and En 3 Hance were purchased from DuPont NEN. [2-14 C]Malonyl-CoA (specific activity, 57 mCi/mmol) was purchased from Moravek Biochemicals Inc. Cerulenin, tetradecanoic acid, and ACP were purchased from Sigma. The cis-7-tetradecenoic acid was the generous gift of Dr. John E. Cronan, Jr. Fatty acids were enzymatically coupled to ACP (12,13) and then concentrated, and the salt was removed by centrifugal filtration in a Centricon-3 concentrator (Amicon). Protein amounts were assayed by the Bradford method (14), and purity was estimated by electrophoresis in a 2.5 M urea, 15% acrylamide gel (15). Acetyl-CoA and malonyl-CoA were from Pharmacia Biotech Inc. All other chemicals were of reagent grade or better.
In Vivo Labeling of the ACP Pool-Strain RJH13 (panD fabI(Ts)) (16) was grown at 30°C to a density of 5 ϫ 10 8 cells/ml in M9 medium (17) supplemented with 0.4% glucose and 0.5 M ␤-[ 3 H]alanine (specific activity, 56 Ci/mmol), at which time an aliquot of the culture was removed and placed into a flask prewarmed to 42°C. Samples (0.5 ml) were taken after 10 min of incubation, and cerulenin (1 mg/ml) was added to the remaining portions and incubated for a further 10 min. Samples were immediately pelleted by centrifugation in a microfuge (15,000 rpm, 4°C, 3 min) and then lysed by a procedure using successive treatments with sucrose, lysozyme, EDTA, and Triton X-100 (18). Products were analyzed by conformationally sensitive electrophoresis in 13% polyacrylamide gels containing 0.5 M urea (15,16,19), and the bands were visualized by fluorography. Strain SJ16 (panD) (20) was grown at 37°C to a density of 5 ϫ 10 8 cells/ml, and the culture was split; one half was treated with 1 mg/ml cerulenin for 10 min, and the other half was harvested as a control.
Preparation of Cell-free Extracts-The crude fatty acid synthase preparation and the fabI(Ts) extract were obtained as described (21). Briefly, a 500-ml culture of either E. coli strain UB1005 or RJH13 (fabI(Ts)) was grown to late log phase in LB medium, and the cells were harvested by centrifugation. Cells were resuspended in 5 ml of lysis buffer (0.1 M sodium phosphate, pH 7.0, 5 mM ␤-mercaptoethanol, 1 mM EDTA) and disrupted by passage through a French pressure cell at 18,000 psi. The lysate was centrifuged in a JA-20 rotor at 20,000 rpm at 4°C for 1 h to remove cell debris, and protein in the 45-80% ammonium sulfate cut of the supernatant was collected. This pellet was resuspended in 2 ml of lysis buffer and dialyzed overnight at 4°C against 1 liter of lysis buffer. Protein was determined by the Bradford method (14).

Construction of Expression Vectors and Purification of the His-Tag
Proteins-The genes were amplified from genomic DNA obtained from strain UB1005. Primers created novel restriction sites for NdeI at the N-terminal methionine and BamHI downstream of the stop codon. Polymerase chain reaction was performed with Taq DNA polymerase (Promega), and the fragments were ligated into the TA cloning vector pCRII (Invitrogen) and transformed into E. coli OneShot cells (Invitrogen). Following overnight growth, plasmid was isolated and digested with NdeI and BamHI, and the appropriate fragments were isolated by GeneClean (Bio 101, Inc.) and ligated into NdeI and BamHI-digested pET-15b (Novagen). These mixtures were used to transform strain BL21(DE3) (Novagen) to ampicillin resistance, and resultant colonies were screened for their ability to overexpress proteins of the correct size following isopropyl-1-thio-␤-D-galactopyranoside induction. One such clone was chosen for each gene and grown in 100 ml of M9 medium (17) supplemented with 1% casamino acids (Difco), 0.4% glucose, and 1 mM MgCl 2 to a density of approximately 5 ϫ 10 8 cells/ml. Isopropyl-1-thio-␤-D-galactopyranoside was then added to a final concentration of 1 mM, and incubation was continued for a further 3 h at 37°C. Cells were collected by centrifugation in a JA-20 rotor (8,000 rpm, 4°C, 10 min), and stored at Ϫ20°C overnight. Lysis was affected according to Ref. 22, with the addition of lysozyme to 0.1 mg/ml after resuspension. Soluble protein was applied to a Ni 2ϩ -agarose column (QIAGEN) and washed with 40 mM imidazole in a metal-chelation affinity chromatography buffer (20 mM Tris-Cl, pH 7.4, 0.5 M NaCl, 1 mM ␤-mercaptoethanol). His-tagged proteins were eluted with 200 mM imidazole in the same buffer, and the purified proteins were stored at Ϫ20°C.
Enzyme Assays-To assay for reduction of short chain enoyl-ACPs, reactions (40 l) contained 50 M acetyl CoA, 25 M ACP, 0.1 M sodium phosphate, pH 7.0, 1 mM ␤-mercaptoethanol, 100 M NADH, 100 M NADPH, 0.1 M LiCl, 30 g of fatty acid synthase extract from either strain UB1005 or RJH13 (fabI(Ts)) and, when added, 1 g of purified His-tag FabI. Cerulenin, when present, was at 100 M. Reactions were preincubated for 5 min at room temperature to allow the cerulenin to inactivate the FabB and FabF condensing enzymes and then initiated by the addition of 25 M [2-14 C]malonyl-CoA (specific activity, 57 Ci/mmol). After incubation at 37°C for 10 min, reactions were stopped by placing the tubes in an ice slurry. Gel loading buffer (10 l) was then added, and 40 l of the incubation mixture was fractionated by conformationally sensitive gel electrophoresis in 13% acrylamide gels containing 0.5 M urea (15). Electrophoresis was performed at 25°C, 32 mA/gel. Fixed gels were soaked in En 3 Hance (DuPont NEN) and then subjected to fluorography at Ϫ80°C.
To investigate the reduction of long-chain saturated and unsaturated acyl-ACPs, reactions were performed as above, except that the acetyl-CoA was replaced with tetradecanoyl-ACP or cis-7-tetradecenoyl-ACP, respectively. Reactions were performed with and without 100 M NADH and 1 g of His-tag FabI. The products were analyzed by conformationally sensitive gel electrophoresis in 2.5 M urea, 15% acrylamide gels to separate the long-chain products.
Reconstruction of Fatty Acid Synthesis in Vitro-A single cycle of fatty acid elongation was reconstructed in vitro by sequentially adding the individual enzymes that catalyze each reaction in the cycle. The reaction mix included, in a volume of 40 l, 50 M ACP, 1 mM ␤-mercaptoethanol, 0.1 M Na 2 PO 4 , pH 7.0, 100 M NADH, 100 M NADPH, 100 M acetyl-CoA, and 50 M [ 14 C]malonyl-CoA (specific activity, 57 mCi/mmol). Purified, His-tagged enzymes were added at 0.2-0.6 g/ assay with FabD being added last to initiate the reaction. Reactions were incubated at 37°C for 20 min and then stopped by placing into an ice slush. Analysis of the ACP thioesters was performed by conformationally sensitive gel electrophoresis of the products in 13% polyacrylamide gel containing 0.5 M urea, and the bands were visualized by fluorography.

RESULTS AND DISCUSSION
Protein Purification and Substrate Synthesis-We purified the five enzymes required to carry out a single cycle of fatty acid synthesis to investigate the role of each enzyme in completing the cycle. Each gene encoding a component of the cycle was cloned into the pET-15b vector, and the proteins were purified by affinity chromatography as described under "Experimental Procedures." In all cases, active enzymes were obtained in a highly purified state (Fig. 2). Acyl-ACPs (14:0 and 14:1⌬7) were prepared using the acyl-ACP synthetase method There are four reactions in each cycle of fatty acid elongation. The first step is catalyzed by ␤-ketoacyl-ACP synthase, which condenses malonyl-ACP with either acetyl-CoA (synthase III, FabH) or acyl-ACP (synthases I and II, FabB, and FabF, respectively). The malonyl-ACP used in the condensing enzyme step is formed by malonyl-CoA:ACP transacylase (FabD). Reduction of the ketoester is catalyzed by the NADPH-dependent ␤-ketoacyl-ACP reductase (FabG). The third step is catalyzed by ␤-hydroxyacyl-ACP dehydrase (either FabA or FabZ). The final step in each cycle is catalyzed by the NADHdependent enoyl-ACP reductase (FabI), which converts trans-2-enoyl-ACP to acyl-ACP. The inner arrows indicate the overall direction of the cycle in fatty acid biosynthesis. The outer thick arrows diagrammatically indicate the equilibrium positions for the enzymatic reactions determined in this study. (12) and were pure (Ͼ95%) as judged by conformationally sensitive gel electrophoresis (not shown).
Reconstruction of Fatty Acid Biosynthesis with Purified Components-A single round of fatty acid synthesis was reconstructed in vitro by the sequential addition of purified components to positively identify the intermediates formed by each of the enzymes in the cycle and to verify the role of FabI in completing rounds of elongation (Fig. 3). The addition of malonyl-CoA:ACP transacylase plus [2-14 C]malonyl-CoA led to the formation of [ 14 C]malonyl-ACP. The addition of ␤-ketoacyl-ACP synthase III (FabH) resulted in the conversion of malonyl-ACP to ␤-ketobutyryl-ACP. The gel electrophoresis technique underestimated the actual amount of ␤-ketobutyryl-ACP formed because the high pH, temperature, and urea used to achieve the separation led to degradation and low recovery of the unstable ␤-ketoesters. The addition of ␤-ketoacyl-ACP reductase (FabG) to the mixture (which also contained NADPH) resulted in the complete conversion of the malonyl-ACP to ␤-hydroxybutyryl-ACP. Significantly, the addition of ␤-hydroxyacyl-ACP dehydrase (FabA) to the mixture resulted in only an 8% conversion of ␤-hydroxybutyryl-ACP to crotonyl-ACP. Increasing the amount of FabA enzyme in the assay mixture did not lead to increased formation of crotonyl-ACP (Fig. 3). The ratio of ␤-hydroxybutyryl-ACP and crotonyl-ACP products was determined in this and several other experiments by densitometric analysis of the fluorograph. Crotonyl-ACP accounted for 8 -11% of the total density in the ␤-hydroxybutyryl-ACP plus crotonyl-ACP bands and was independent of the amount of FabA enzyme added to the assay or the length of the incubation. These data indicated that the FabA reaction rapidly reached an equilibrium that favored ␤-hydroxybutyryl-ACP. The addition of FabI to the reaction mixture yielded butyryl-ACP, and the extent of butyryl-ACP formation increased as the amount of FabI in the assay increased. These data show that the activity of FabI is responsible for pulling cycles of fatty acid elongation.

Correction of the Fatty Acid Synthesis Defect in Extracts of fabI(Ts) Mutants with Purified FabI-Extracts isolated from
wild-type and fabI(Ts) defective strains were used to examine in vitro the role of the FabI protein in fatty acid elongation (Fig.  4). As previously reported (20), extracts from the wild-type strain incorporated [ 14 C]malonyl-CoA into butyryl-ACP and long-chain acyl-ACP. The addition of cerulenin blocked the activity of FabB and FabF but not FabH, leading to the accumulation of butyryl-ACP. Extracts from strain RJH13 (fabI(Ts)) were unable to form either butyryl-ACP or long-chain acyl-ACP either in the presence or the absence of cerulenin. Two intermediates accumulated in these reactions. The major intermediate comigrated with ␤-hydroxybutyryl-ACP, and the minor species comigrated with crotonyl-ACP. Densiometric analysis of the fluorograph showed that the crotonyl-ACP band comprised 10% of the total density of the ␤-hydroxybutyryl-ACP plus crotonyl-ACP bands. The ratio of these two products formed in the crude cell extract was virtually identical to the ratio of products observed in the purified system in the absence of FabI (Fig. 3). The addition of purified FabI protein to the extract from the fabI(Ts) strain restored the ability of the extract to synthesize both long-chain acyl-ACP and butyryl-ACP in the presence of cerulenin. These data show that FabI is essential for the first cycle of fatty acid elongation and is the only component missing from the extracts of strain RJH13 (fabI(Ts)).
The data suggested that FabI participated in subsequent elongation cycles, but the possibility remained that another enoyl-ACP reductase (the NADPH-dependent enzyme predicted by Weeks and Wakil (5)) was present in the extracts and contributed to the elongation of saturated or unsaturated fatty acids to produce long-chain acyl-ACP (Fig. 4). To verify that FIG. 2. Purification of FabD, FabH, FabG, FabA, and FabI proteins. The coding sequences for each of these proteins was cloned into the pET-15b His-tag expression vector, expressed to high levels following isopropyl-1-thio-␤-D-galactopyranoside induction, and purified by chromatography on Ni 2ϩ -agarose as described under "Experimental Procedures." The purity of each protein preparations used was examined by SDS gel electrophoresis on 12% polyacrylamide gels.

FIG. 3. Reconstitution of a cycle of fatty acid synthesis with purified enzymes.
A single cycle of fatty acid biosynthesis was reconstructed in vitro by the sequential addition of each purified enzyme to a reaction mixture containing NADPH, NADH, and ACP as cofactors and acetyl-CoA and [2-14 C]malonyl-CoA as substrates as described under "Experimental Procedures." The products were analyzed by conformationally sensitive gel electrophoresis using 13% polyacrylamide gels containing 0.5 M urea.

FIG. 4. Restoration of butyryl-ACP and long-chain acyl-ACP formation in extracts of fabI(Ts) by the addition of purified FabI protein.
Cell-free extracts of strain UB1005 (wild type) and RJH13 (fabI(Ts)) were prepared, and fatty acid synthase reactions containing NADPH, NADH, ACP, acetyl-CoA, and [2-14 C]malonyl-CoA were performed as described under "Experimental Procedures." Cerulenin (100 M) was added to the indicated reactions (ϩ) to block the activity of ␤-ketoacyl-ACP synthases I and II (FabB and FabF). Purified His-tag FabI (1 g) was added to the indicated assays. The products were analyzed by conformationally sensitive gel electrophoresis using 13% polyacrylamide gels containing 0.5 M urea, and the bands were visualized by fluorography.
FabI was capable of participating in the elongation of these long-chain species and to rule out the presence of another enzyme, we examined the elongation of 14:0-ACP 2 and 14:1⌬7-ACP (Fig. 5). Extracts from strain RJH13 (fabI(Ts)) were unable to convert either 14:0-ACP or 14:1⌬7-ACP to 16:0-ACP or 16:1⌬9-ACP, respectively. These extracts contained FabB (or FabF) that catalyzed the condensation of [ 14 C]malonyl-CoA with the acyl-ACP, FabG, and FabZ (or FabA) that carried out the NADPH-dependent reduction of the ␤-ketoacyl-ACPs and the dehydration of ␤-hydroxyacyl-ACP to the trans-2-enoyl-ACP intermediate. Two intermediates accumulated in each experiment that corresponded in electrophoretic mobility to the ␤-hydroxyacyl and trans-2-enoyl intermediates in the elongation cycles. Like the electrophoretic relationship between ␤-hydroxybutyryl-ACP and crotonyl-ACP observed in Fig. 3, the long-chain ␤-hydroxyacyl intermediates migrated more slowly than the corresponding acyl-ACP, and the long-chain trans-2enoyl intermediates migrated slightly more quickly than the corresponding acyl-ACP. These relative mobilities were the same as reported previously for long-chain ␤-hydroxyacyl and trans-2-enoyl acyl-ACP (4). The ratio of the long-chain ␤-hydroxyacyl-ACP to trans-2-enoyl-ACP was quantitated by densitometry for each experiment (Fig. 5). The average ratio of these two intermediates was 1:1 indicating that the equilibrium for the dehydrase operating on the long-chain ␤-hydroxyacyl-ACPs was more favorable for the formation of the trans-2-enoyl-ACP product than was observed with the shortchain intermediates (Figs. 3 and 4). The addition of purified FabI to these extracts restored the ability of the extracts to elongate 14:0-ACP to 16:0-ACP and 14:1⌬7-ACP to 16:1⌬9-ACP. NADH was required in addition to FabI for the formation of acyl-ACP. These data rule out the existence of a NADPH-dependent enoyl-ACP reductase participating in the elongation of either long-chain saturated or unsaturated fatty acids. These experiments also verified that FabI is capable of reducing the trans-2 double bond in both long-chain saturated and unsaturated enoyl-ACP.
Composition of the ACP Pool in fabI(Ts) Mutants-Previously, we concluded from the analysis of fatty acid synthesis and long-chain acyl-ACP and malonyl-CoA accumulation in fabI(Ts) mutants that FabI was required for the first elongation cycle that forms butyryl-ACP (16). We also noted a change in the composition of the short-chain ACP thioester pool, and we examined this pool more closely to determine if our in vitro experiments with extracts from fabI(Ts) mutants and purified components correlated with the in vivo results. Strain RJH13 (panD2 fabI(Ts)) was labeled with ␤-[ 3 H]alanine to uniformly label the ACP pool, and the composition of this pool was analyzed by conformationally sensitive gel electrophoresis at both the permissive and nonpermissive temperatures using the wild-type strain SJ16 for comparison (Fig. 6). To facilitate the analysis, a portion of each culture was incubated with cerulenin, an irreversible inhibitor of the FabB and FabF condensing enzymes. At the permissive temperature (30°C), the major ACP species in strain RJH13 co-migrated with acetyl-ACP with significant amounts of malonyl-ACP also detected. This pattern of ACP intermediates was essentially the same as in the control strain SJ16. The addition of cerulenin to strain RJH13 grown at 30°C led to the accumulation of butyryl-ACP, which is the same major change in the pool composition found in the wildtype control strain SJ16. In contrast, treatment of strain RJH13 with cerulenin after a temperature shift to 42°C did not lead to the accumulation of butyryl-ACP indicating that the first cycle of fatty acid biosynthesis was blocked in the fabI(Ts) mutants. Shifting strain RJH13 to the nonpermissive temperature (42°C) lead to a significant change in the ACP pool composition. Acetyl-ACP essentially disappeared, malonyl-ACP accumulated, and a new species migrating slightly more quickly than ACP appeared. We also observed the appearance of a new minor species that migrated just ahead of butyryl-ACP. The migration positions of these two ACP intermediates in vivo corresponded to the migration positions of ␤-hydroxybutyryl-ACP and crotonyl-ACP (Figs. 3 and 4). One important conclusion from these experiments was that the ACP pool composition of fabI(Ts) mutants at the nonpermissive temperature ( Fig. 6) was essentially identical to the distribution of products observed in the in vitro system with either purified components (Fig. 3) or extracts from fabI(Ts) mutants (Fig. 4). Densitometric analysis of the fluorograph showed that the band with the migration position of crotonyl-ACP was 13% of the total density in the ␤-hydroxybutyryl-ACP plus crotonyl-ACP bands. Thus, our in vitro reconstitution experiments accurately reflect the properties of the fatty acid synthase in vivo.
Conclusions-Our results point to enoyl-ACP reductase (the fabI gene product) as a determinant factor in completing rounds of fatty acid elongation (Fig. 1) 6. Composition of the ACP pool in fabI(Ts) mutants. Strain RJH13 (panD fabI(Ts)) was grown at 30°C and strain SJ16 (panD) at 37°C with ␤-[3-3 H]alanine to a density of 5 ϫ 10 8 cells/ml. The strain RJH13 culture was split, and one half was shifted to 42°C. 10 min later, samples were removed, remaining cultures were treated with 1 mg/ml cerulenin for a futher 10 min, and additional 1-ml samples were removed. A portion of the strain SJ16 culture was also treated with cerulenin for 10 min. The ACP pool composition was analyzed by conformationally sensitive gel electrophoresis in 13% polyacrylamide gels containing 0.5 M urea, and the bands were visualized by fluorography as described under "Experimental Procedures." The migration position of acyl-ACP standards are indicated to the right.
ness of the arrows in Fig. 1. The amount of butyryl-ACP formed is dependent on the amount of FabI protein added due to the dehydration reaction reaching a rapid equilibrium that favors the accumulation of ␤-hydroxybutyryl-ACP over crotonyl-ACP by a ratio of 9:1. These data indicate that increasing the rate of initiation of a cycle of fatty acid elongation by a factor of 10 would only double the amount of enoyl-ACP product produced by FabA. Although it is not clear whether FabA, FabZ, or both contribute to the dehydration of ␤-hydroxybutyryl-ACP in vivo, the fact that our in vitro reconstitution experiments reflect the properties of the fatty acid synthase in vivo provides compelling support for the conclusion that FabI activity plays a determinant role in completing rounds of fatty acid elongation.
Our experiments show that enoyl-ACP reductase (the fabI gene product) is the only reductase required to complete both saturated and unsaturated fatty acid synthesis in the type II, dissociated fatty acid synthase system of E. coli. Extracts derived from fabI(Ts) strains were unable to synthesize butyryl-ACP or complete the elongation of long-chain saturated or unsaturated fatty acids without supplementation with exogenous FabI. We found no evidence for the NADPH-dependent isoform postulated by Weeks and Wakil (5). The Mycobacterium tuberculosis FabI analog (inhA) is also specific for NADH (23) and participates in ACP-dependent fatty acid synthesis in a type II system similar to that in M. smegmatis (24). We do not have a complete picture of the substrate specificity of InhA enoyl-ACP reductase, but the enzyme effectively reduces trans-2-octenoyl-ACP as well as long-chain intermediates (24) indicating a broad chain length specificity.
Hypothesis-The fact that a single enoyl-ACP reductase plays a determinant role in completing cycles of chain elongation suggests that FabI may be a key regulatory component in dissociated fatty acid synthase systems. Two aspects of fatty acid synthesis may be influenced by modulation of FabI activity. First, FabI may participate in establishing the basal saturated:unsaturated fatty acid ratio. FabA is the critical branch point because it is the only dehydrase capable of forming both cis-3-decenoyl-ACP and trans-2-decenoyl-ACP (1). FabB is essential for unsaturated fatty acid synthesis presumably because it is the only condensing enzyme capable of utilizing cis-3-decenoyl-ACP. In contrast, FabI reduces trans-2-decanoyl-ACP to decanoyl-ACP to initiate saturated fatty acid biosynthesis. Genetic manipulation of the intracellular levels of FabB and FabA establish that altering the relative ratios of these two proteins has a dramatic impact on the saturated:unsaturated fatty acid ratio (25). One testable hypothesis is that increased levels of FabI protein would increase synthesis of saturated fatty acids by increasing the utilization of trans-2-decenoyl-ACP. Second, FabI may be a target for regulation of the total rate of chain elongation. Acyl-ACPs are emerging as key regulators of the rate of fatty acid biosynthesis, and their destruction in vivo by the expression of acyl-ACP thioesterases of bacterial (26,27) or plant (28 -30) origin leads to the acceleration of fatty acid biosynthesis and the copious secretion of fatty acids into the medium. Furthermore, down-regulation of fatty acid synthesis following the cessation of phospholipid production at the glycerol-phosphate acyltransferase step correlates with the elevation of intracellular acyl-ACP (16,31). The intracellular targets for acyl-ACP regulation are unknown. One tenable hypothesis is that acyl-ACPs attenuate FabI activity by product inhibition, and this mechanism may contribute to the regulation of pathway activity by acyl-ACP.