Functional Role of Fatty Acyl-Coenzyme A Synthetase in the Transmembrane Movement and Activation of Exogenous Long-chain Fatty Acids

Fatty acyl-CoA synthetase (FACS, fatty acid:CoA ligase, AMP forming; EC 6.2.1.3) plays a central role in intermediary metabolism by catalyzing the formation of fatty acyl-CoA. InEscherichia coli this enzyme, encoded by thefadD gene, is required for the coupled import and activation of exogenous long-chain fatty acids. The E. coli FACS (FadD) contains two sequence elements, which comprise the ATP/AMP signature motif (213YTGGTTGVAKGA224and 356GYGLTE361) placing it in the superfamily of adenylate-forming enzymes. A series of site-directed mutations were generated in the fadD gene within the ATP/AMP signature motif site to evaluate the role of this conserved region to enzyme function and to fatty acid transport. This approach revealed two major classes of fadD mutants with depressed enzyme activity: 1) those with 25–45% wild type activity (fadD G216A, fadD T217A,fadD G219A, andfadD K222A) and 2) those with 10% or less wild-type activity (fadD Y213A,fadD T214A, andfadD E361A). Using anti-FadD sera, Western blots demonstrated the different mutant forms of FadD that were present and had localization patterns equivalent to the wild type. The defect in the first class was attributed to a reduced catalytic efficiency although several mutant forms also had a reduced affinity for ATP. The mutations resulting in these biochemical phenotypes reduced or essentially eliminated the transport of exogenous long-chain fatty acids. These data support the hypothesis that the FACS FadD functions in the vectorial movement of exogenous fatty acids across the plasma membrane by acting as a metabolic trap, which results in the formation of acyl-CoA esters.

Fatty acyl-CoA synthetase (FACS, 1 fatty acid:CoA ligase, AMP forming; EC 6.2.1.3) plays a central role in intermediary metabolism by catalyzing the formation of fatty acyl-CoA. These bioactive fatty acid metabolites are involved in protein transport (1,2), enzyme activation (3,4), protein acylation (5-7), cell signaling (8 -10), and transcriptional control (11)(12)(13)(14)(15) in addition to serving as substrates for ␤-oxidation and phospholipid biosynthesis. In Escherichia coli the FACS FadD is hypothesized to catalyze the activation of long-chain fatty acid to CoA thioesters concomitant with transport by a process termed vectorial esterification (16,17). The concerted activities of FadD and the outer membrane fatty acid transport protein FadL essentially render the process of fatty acid transport unidirectional. The seminal work of Overath and colleagues (18) showed that FACS activity was found both within the membrane and soluble fractions suggesting that this enzyme moves between the cytosol and plasma membrane to facilitate the vectorial esterification of exogenous fatty acids. Indeed, subsequent studies have suggested that the enzyme is recruited to the plasma membrane, but no information has been amassed on the underlying mechanism (19). The membrane association of FadD may be similar to that of CTP:phosphocholine cytidylyltransferase, an enzyme required for phospholipid biosynthesis that is catalytically active in the membrane and inactive when soluble (20).
FACS catalyzes the formation of fatty acyl-CoA by a two-step process that proceeds through the hydrolysis of ATP to yield pyrophosphate. One of the key features of this catalysis is the formation of an adenylated intermediate (21). This activation step involves the linking of the carboxyl group of the fatty acid through an acyl bond to the phosphoryl group of AMP. Subsequently a transfer of the fatty acyl group to the sulfhydryl group of coenzyme A occurs releasing AMP. By analogy with acetyl-CoA synthetase, it is likely this reaction precedes via a Bi-Uni, Uni-Bi Ter-molecular ping-pong mechanism with fatty acid, ATP, and CoA all serving as substrates (22,23). The formation of an enzyme-bound adenylated intermediate is a common mechanism used by a number of enzymes to activate their substrates. Sequence comparisons of adenylateforming enzymes have identified two highly conserved sequence elements (YTSGTTGXPKGV and GYGXTE) that comprise the ATP/AMP signature motif (24). The first sequence is generally 125-130 residues upstream from the second. A third less-conserved element that is thought to contribute to ATP/ AMP binding overlaps the FACS signature (70 -75 residues downstream from the second), which we have shown is involved in both catalysis and specificity of the fatty acid substrate (25).
The focus of the present work was to identify specific resi-dues within the ATP/AMP signature critical for function and to further investigate the role of this enzyme in the process of long-chain fatty acid transport. A series of alanine substitutions were generated in the region of FadD corresponding to the ATP/AMP signature motif site. Subsequent analyses of the mutant forms of the enzyme permitted classification into three groups. Those with wild-type or nearly wild-type levels of oleoyl-CoA synthetase activity, those with oleoyl-CoA synthetase activities reduced 40 -80% when compared with wild-type, and those with essentially undetectable levels of oleoyl-CoA synthetase activity. Several of the altered forms of the enzyme with detectable oleoyl-CoA synthetase activity had reduced catalytic rates as opposed to a reduction in affinity for ATP attesting to the importance of this highly conserved region for catalysis. Substitutions within the ATP/AMP signature motif of FadD, which resulted in depressed FACS activities also had depressed fatty acid transport levels. These data support the notion the membrane-associated fatty acyl-CoA synthetase FadD plays a central role in the transmembrane movement and metabolic activation of exogenous long-chain fatty acids in E. coli.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Growth Conditions-The bacterial strains used in this study were: BMH 71-18 (thi supE ⌬(lac proAB) mutS::Tn10 LS6928 (fadR fadD88 zea::Tn10), LS1547 (prototrophic wild-type), LS1548 (⌬fadR), LS1908 (⌬fadR fadD::Kan r ), LS6952 (recBC), and BL21 (DE3)/pLysS. Strain LS1908 was constructed by replacing the coding region of fadD with the kanamycin resistance gene cartridge (Kan r ; Amersham Biosciences) using a two-step process. First, the coding region of fadD in pN300 (24) was deleted and replaced by the Kan r cartridge thereby maintaining Ͼ500 bp of flanking DNA to generate pJW201. Second, this plasmid was digested with BamHI to generate a linear DNA molecule, transformed into strain LS6952 and movement of ⌬fadD::Kan r by P1 transduction into strain LS1548 to generate the ⌬fadR ⌬fadD::Kan r strain LS1908. The fadD deletion in LS1908 was confirmed by PCR of chromosomal DNA using primers specific to bacterial DNA flanking the deletion and by a loss of acyl-CoA synthetase activity. Bacterial cultures were grown at 37°C in a gyratory shaker in Luria broth or tryptone broth (TB). When minimal media was required, medium E supplemented with vitamin B 1 (EB 1 ) was used (26). Carbon sources, sterilized separately, were added to final concentrations of 25 mM glucose, 25 mM potassium acetate, 5 mM decanoate, or 5 mM oleate. When oleate or decanoate were used as a carbon source, polyoxyethylene 20 cetyl ether (Brij 58) was added to a final concentration of 0.5%. As required, amino acids were added to a final concentration of 0.01%. When required to maintain plasmids, antibiotics were added to 100 g/ml ampicillin, 40 g/ml kanamycin, 10 g/ml tetracycline, and 40 g/ml chloramphenicol. Growth of bacterial cultures was routinely monitored using a Klett-Summerson TM colorimeter equipped with a blue filter.
Amino Acid Sequence Comparisons-Protein sequence comparisons were performed using MultAlign version 5.3.3 (27) with sequences that were conserved among the fatty acyl-CoA synthetases or members of the adenylate forming superfamily. Sequence comparisons were directed against the SWIS-PROT data base using a 0.01 probability threshold with complexity filtering and blosum62 score table.
Site-directed Mutagenesis-Site-directed mutations within the fadD gene were generated using the Altered Sites TM mutagenesis system from Promega as previously described (28). Specific mutations were confirmed using dideoxysequencing and fadD-specific oligonucleotides (24). Once the mutation was confirmed, BamHI or SacII-SalI fragments containing the different fadD alleles were purified and ligated into pACYC177. The designations of the final constructions are given in Table I. The plasmid constructs were then transformed into the host strain LS1908 (⌬fadR ⌬fadD::Kan r ) for analysis.
His6 FadD Overexpression and Purification-Strain BL21 (DE3) (pLysS) was transformed with plasmid pN3576 encoding the bacterial fatty acyl-CoA synthetase FadD containing a hexameric histidine tag (25). The expression and purification of His6 FadD has been previously described (25). Briefly, cells were induced with isopropyl-1-thio-␤-Dgalactopyranoside (1 mM) for 90 min and the enzyme was purified from a clarified cell extract using Ni 2ϩ -NTA-agarose affinity chromatography (Qiagen). The His-tagged enzyme elutes as a single band between 250 and 300 mM imidazole. Antisera were prepared against purified FadD using a commercial vendor (BioWorld, Dublin, OH). Selected fadD alleles were subcloned as cassettes into pN3576 for expression and purification of mutant forms of the enzyme as detailed under "Results." Western Blotting-Western blotting of cell extracts from wild-type and ⌬fadD strains transformed with the plasmid constructs harboring the different fadD alleles was performed as previously described (30). Briefly total cellular protein or isolated membrane and cytosolic fractions were subjected to SDS-polyacrylamide gel electrophoresis and electrophoretically transferred onto nitrocellulose membranes (pore size, 0.45 m) (Micron Separations Inc.). Following transfer to nitrocellulose, membranes were pretreated for 1 h at room temperature with gentle shaking in 5% powdered milk ϩ 0.1% Tween 20 in Tris-buffered saline (TTBS; 20 mM Tris, 0.5 M NaCl, pH 7.5). Following this blocking step, anti-FadD sera (1:25,000 dilution) was added directly and incubated an additional hour. Membranes were washed with TTBS three times and subsequently incubated with goat anti-rabbit IgG horseradish peroxidase conjugate (1:7,500) in 5% powdered milk ϩ TTBS for 1 h at room temperature. Membranes were washed in TTBS as detailed above and developed using enhanced chemiluminescence (ECL; Amersham Biosciences) or by the addition of the peroxidase substrate tetramethylbenzidene as specified by the vendor (Promega).
The soluble and particulate fractions of cells containing the different fadD alleles were separated and probed for FadD and mutant forms of FadD using Western blots and anti-FadD sera. Briefly, cells were grown as detailed above and lysed using a French pressure cell at 12,000 p.s.i. The cell lysates were clarified to remove unbroken cells and the supernatants were subjected to ultracentrifugation (60,000 ϫ g for 30 min; TLA-100 rotor) to separate the membrane (particulate) fraction from the cytosolic fraction. The membranes were resuspended and centrifugation repeated. The protein concentrations of the membrane and cytosolic fractions were adjusted to 8 -10 g/10 l and analyzed by Western blots as detailed above. The distribution between the membrane and cytosolic fractions of the mutant forms of FadD were compared with those from wild-type FadD.
Measurement of Fatty Acyl-CoA Synthetase Activity-FACS activity was monitored using a modification of a protocol defined by Kameda and Nunn (29). Cell extracts were prepared from wild-type fadD strains by sonication on ice as previously described (25). Cell extracts were added to reaction mixtures containing 200 mM Tris-HCl, pH 7.5, 2.5 mM ATP, 8 mM MgCl 2 , 2 mM EDTA, 20 mM NaF, 0.1% Triton X-100, 10 M [ 3 H]oleate, 0.5 mM coenzyme A and incubated 10 min at 37°C. For the kinetic experiments the concentration of ATP was varied between 0.05 and 2.5 mM. Reactions were initiated by the addition of CoA and terminated by the addition of isopropyl alcohol, n-heptane, 1 M H 2 SO 4 (40:10:1). The aqueous phase, containing acyl-CoA formed during the reaction, was extracted 3 times with 2.5 ml of n-heptane and subjected to scintillation counting.
Fatty Acid Transport-Oleate transport was measured as previously described (30). Following growth to mid-log phase (5 ϫ 10 8 cells/ml) in TB, cells were harvested, washed once with minimal media EB 1 , and resuspended in EB 1 containing 0.5% Brij 58 and 200 g/ml chloramphenicol. The cells were starved for an exogenous energy source with aeration for 30 min at 30°C. Following this starvation period, 1 ml of cells was added to 1 ml of an assay mixture containing 200 M Enzymes for routine DNA manipulations were obtained from Promega, Invitrogen, or New England Biolabs. Antibiotics and other supplements for bacterial growth were obtained from Difco and Sigma. All other chemicals were obtained from standard suppliers and were of reagent grade.

Identification of an ATP/AMP-binding Signature
Motif within the Fatty Acyl-CoA Synthetase FadD-One of the key features of the catalytic mechanism of FACS is the formation of an adenylated intermediate (21). This activation step involves the linking of the carboxyl group of the fatty acid to the phosphoryl group of AMP and subsequent transfer of the acyl chain to the acceptor molecule coenzyme A. The formation of an enzyme-bound adenylated intermediate was a common mechanism used by a number of enzymes to activate their substrates. Sequence comparisons of enzymes, which share this catalytic property, identified two highly conserved sequence elements that comprise the ATP/AMP-binding signature motif (Fig. 1). Within the family of adenylate forming enzymes there was a third sequence element of this signature that was less well conserved and partially overlaps the FACS signature motif (25). In the E. coli FACS FadD, the two regions comprising the ATP/AMP signature have been identified on the basis of sequence similarities as 213 YTGGTTGVAKGA 224 and 356 GYGLTE 361 .
Construction of Site-directed Substitutions within the ATP/ AMP-binding Signature of FACS-The homologies noted above served to guide studies using site-directed mutagenesis to define whether this region of FACS was essential for enzyme activity. Based on data from other adenylate-forming enzymes, the two highly conserved sequence elements contribute to a region of the enzyme hypothesized to be required for ATP binding and catalysis. A number of specific amino acid residues within the ATP/AMP signature motif were substituted with alanine to assess their contribution to enzymatic activity ( Fig.  1, Table I). Western blots using anti-FadD sera demonstrated the protein levels of wild-type FadD and the mutant forms of FadD were equivalent (Fig. 2). These data support the notion that the mutant forms of FadD were expressed at levels ap-proximating the wild type and there was no indication the proteins had reduced stability compared with the wild type. FadD was found in both the soluble and particulate fractions of the bacterial cell (18,29). Thus we monitored the levels of FadD and mutant forms of FadD in both the soluble and particulate fractions using Western blots and found no discernable changes in patterns indicating that amino acid substitutions within the ATP/AMP signature did not result in an enzyme with altered cellular localization (Table II). As the specific amino acids targeted for substitution were postulated to reside within the FIG. 2. Western blots of FadD and mutant forms of FadD using anti-FadD sera. Total cell extracts were prepared and total protein (9 -10 g/lane) was analyzed as detailed under "Experimental Procedures." 1, LS1908 (⌬fadD::Kan r ⌬fadR); 2, LS1908/pACYC177 (vector control); 3, LS1908/pN300 (FadD ϩ ); 4, LS1908/FadD Y213A ; 5, LS1908/ FadD T214A ; 6, LS1908/FadD G216A ; 7, LS1908/FadD T217A ; 8, LS1908/ FadD G219A ; 9, LS1908/FadD K222A ; 10, LS1908/FadD E361A . The arrow on the right indicates the band corresponding to FadD or mutant forms of FadD. The asterisks denote nonspecific proteins, which do not change when fadD is deleted.
FIG. 1. ATP/AMP signature motif common to adenylate forming enzymes. The alignment represents only a small subset of members in the AMPbinding protein family, which share this signature. The asterisks in the last line denote the amino acids substituted with alanine in the different fadD alleles constructed and analyzed in this work. catalytic center of the enzyme, it seems likely that subtle conformational changes will result that cannot be detected using these methods. Whereas the anti-FadD sera was prepared using highly purified His6 FadD, two additional proteins were recognized, which are likely to be other adenylate-forming enzymes that share epitopes in common with FadD (Fig. 2). One of these was likely to be the undefined open reading frame designated YDID, which encodes an acyl-CoA synthetase-like protein that was 30% identical to FadD. YDID does not, however, contribute long-chain acyl-CoA synthetase activity. 2 Fatty Acyl-CoA Synthetase Activities of Whole Cell Extracts from fadD Mutants Containing Substitutions within the Putative ATP/AMP-binding Signature Motif-This work was driven by the hypothesis that the ATP/AMP signature sequence elements were required for ATP binding and the formation of the adenylated intermediate. FACS activities were monitored in whole cell extracts using oleate (C 18:1 ) and decanoate (C 10:0 ) as substrates in the ⌬fadD strain LS1908 transformed with pN300 (wild-type fadD), the plasmid vector pACYC177 and plasmids encoding the collection of fadD alleles (Table II). The alleles fadD Y213A , fadD T214A , fadD G216A , fadD T217A , fadD G219A , and fadD K222A (within the first sequence element of the signature) and fadD E361A (within the second element) had markedly reduced enzymatic activities using oleate as the fatty acid substrate. As with our previous studies describing the FACS signature motif (25), the present studies also revealed two major classes of mutants with depressed enzyme activity: 1) those with 25-45% wild type activity (fadD G216A , fadD T217A , fadD G219A , and fadD K222A ) and 2) those with 10% or less wildtype activity (fadD Y213A , fadD T214A , and fadD E361A ) ( Table II). Given that these mutant forms of FadD were expressed at wild-type levels and had the same localization patterns, these data would argue that these residues were crucial for catalytic activity.
Determination of V max , K m , and k cat Values for Fatty Acyl-CoA Synthetase Activities from fadD Mutants as a Function of ATP Concentration-The data presented above indicated that specific residues within the ATP/AMP signature were important for enzymatic activity. To investigate this further, kinetic studies were undertaken to specifically evaluate the role of ATP in catalysis (Table III, Fig. 3). We suspected that several of the fadD mutants with substitutions within the ATP/AMP signature would be defective for ATP binding and thus would have a significantly higher K m for ATP when compared with the wild-type enzyme. As shown in Fig. 3, the fadD Y213A and fadD E361A alleles resulted in enzymes with no measurable activity over a range of ATP concentrations. FadD Y213A may be defective in ATP binding as is the case for the comparable residue in the mammalian fatty acid transport protein FATP1, which also was a member of the adenylate-forming family of enzymes (31,32). By comparison with the phenylalanine activating subunit (PheA) of gramicidin synthetase in Bacillus brevis, Glu 361 was essential for the catalytic activity of the enzyme. The homologous amino acid in PheA was a glutamate, which appears to function in concert with a neighboring tyrosine to coordinate the binding of Mg 2ϩ -AMP (33). In this regard, Glu 361 may specifically contribute to ATP binding.
Three different fadD alleles (fadD T214A , fadD G216A and fadD G219A ) resulted in enzyme activities that had k cat values 2-15-fold lower than wild-type, whereas the K m values for ATP were relatively unchanged. This suggests that the primary role of these residues were for catalysis as opposed to ATP binding. Two other alleles (fadD T217A and fadD K222A ) affected both K m and k cat . Both mutants had k cat values that were reduced nearly 4-fold and K m values that were increased when compared with the wild type. These results were consistent with the notion that these two residues contribute to ATP binding. On the basis of previous work on PheA, the positive charge on the comparable lysine residue was thought to function by directing ATP to the binding site (33).
Analysis of Long-chain Fatty Acid Transport-In view of the compromised acyl-CoA synthetase activities in the fadD alleles described in the preceding sections, we postulated this would lead to reduced levels of long-chain fatty acid import. Therefore, we selected a subset of these fadD mutants and defined the patterns of fatty acid transport using [ 3 H]oleate as the substrate as detailed under "Experimental Procedures." One of the fadD alleles examined (fadD Y213A ) was severely compromised by oleate transport when compared with the wild type, two (fadD T217A and fadD E361A ) had activities that were reduced considerably, but detectable, and one (fadD G219A ) resulted in fatty acid transport levels that were only modestly reduced when compared with wild type (Fig. 4). When fatty acid transport was measured using different concentrations of oleate in a selected subset of mutants (fadD Y213A , fadD T217A , and fadD E361A ), the same patterns of activity were found (Fig. 5) indicating the changes in transport rate were the direct result of changes in the catalytic efficiency of FadD. Kinetic constants extracted from these experiments indicated that the K m of oleate transport resulting from FadD T217A was essentially equivalent to the wild type, whereas FadD E361A resulted in a 3-fold reduction (Table IV). The changes in V max were more a Detection of FadD and mutant proteins using Western blots and anti-FadD sera; ϩ, levels detected comparable to wild type in whole cell extracts and soluble and particulate fractions; Ϫ, FadD not detected (see Fig. 2).
b Protein concentrations in the reaction mixtures ranged between 36 and 50 g/ml. S.E., standard error of the mean; n ϭ 4. dramatic indicating that a reduction in the catalytic activity of FadD resulted in a commensurate reduction in oleate transport. The addition of exogenous ATP to the fatty acid transport assay mixture had no effect on the measured fatty acid transport rates of strains containing the wild-type FadD or mutant forms of FadD (data not shown), which suggested that the intracellular pools of ATP were required and sufficient for enzymatic activity as previously suggested (45). These data, in conjunction with those presented above, argue that there was specific coupling between fatty acid import and activation of exogenous long-chain fatty acids in E. coli. Lipid Activation of Fatty Acyl-CoA Synthetase-The E. coli FACS FadD appears to be an essential component of the longchain fatty acid transport apparatus, perhaps by acting to metabolically trap the fatty acid as a CoA thioester. A central question that remains to be answered is how this enzyme was specifically involved in this process. Indeed, there is increasing evidence that in higher eukaryotic systems, this enzyme also plays a role in long-chain fatty acid transport (46). The first studies on the subcellular localization of a long-chain FACS were carried out on mammalian small intestinal mucosa where the enzyme was predominantly localized in the microsomal fraction (34,35). A mouse isoform of FACS was found within the plasma membrane and was thought to function in concert with a fatty acid transport protein (36). Previous work on FadD has shown that it was present in both the particulate and soluble fractions of the cell (18,19,29). These observations suggest that FadD may partition into the membrane during its catalytic cycle and suggests this enzyme functions to abstract fatty acids from the membrane concomitant with esterification with CoA.
FACS activity from E. coli was routinely monitored in the presence of low concentrations of nonionic detergent. When detergent was removed, FACS activity either cannot be measured or was extremely low indicating that either the substrate requires delivery in a micellar form and/or that FACS prefers a lipophilic environment for activity. Total E. coli lipid can functionally replace the nonionic detergent in restoring enzyme activity (Fig. 6) implying that the enzyme requires a lipophilic environment to be active. In this regard, FadD may indeed be lipid activated as it associates with the membrane.

DISCUSSION
The present studies were initiated to identify residues within fatty acyl-CoA synthetase required for catalysis and the vectorial import of exogenous long-chain fatty acids. Protein sequence homology searches have identified a number of amino acid residues as active site candidates, and in particular, identified residues common to the more limited family of fatty acyl-CoA synthetases and to the larger family of ATP/AMPbinding proteins. Site-directed mutagenesis of the fadD gene was undertaken to specifically evaluate the contribution of amino acid residues identified in the region called the ATP/ AMP signature motif in the E. coli FACS, FadD. The crystal- lographic structures of two members of the ATP/AMP-binding family (firefly luciferase and the phenylalanine activating subunit of gramicidin S synthetase (PheA)) have been solved and thus serve as a platform for discussions of conserved residues that may participate in ligand binding and enzyme catalysis (33,37). The experimental data obtained using the site-directed fadD mutants detailed from this work and from our previous work (25) coupled with the predicted structure of FadD (38) provide a foundation of information required to more fully understand the role of this enzyme in the fatty acid activation/transport process.
The conservation of residues within the two sequence elements of the ATP/AMP signature motif by all adenylate family members strongly suggests that these regions contribute to the binding of ATP and/or to the formation of the enzyme adenylated reaction intermediate, because ATP is the one substrate common to all members. The crystal structure of firefly luciferase reveals that the N-terminal domain comprises the major portion of the molecule and consists of a distorted antiparallel ␤-barrel and two ␤-sheets, which were flanked on either side by ␣-helices. Based on the predicted three-dimensional model of FadD from E. coli, the majority of these residues are clustered in a cleft separating two domains of the enzyme (38). In addition, the FACS signature sequence motif is predicted to lie in the same cleft providing compelling evidence that the ATP/ AMP and FACS signature sequences contribute to the catalytic core of the enzyme (25,38).
The first sequence element of the ATP/AMP signature motif is the most highly conserved sequence of amino acids in the superfamily of adenylate-forming enzymes (39,40). Although this sequence (YTSG(ST)TGXPKGV) has poor homology to the Walker A type motif (GXXXXGK(TS)), it is rich in glycine and contains a lysine. The Walker A motif forms a phosphatebinding loop found in a large number of nucleotide-binding proteins (41). The lysyl residue in this motif is thought to be responsible for binding the ␥-phosphate of the nucleoside triphosphate. A comparable Lys has been identified in the ATP-binding site of SecA, a protein involved in protein transport across the plasma membrane of E. coli (42). In B. brevis the comparable Lys is critical for ATP-PPi exchange in the nonribosomal peptide synthetase tyrocidine synthetase I (43). Substitution of this homologous residue to methionine in 4-chlorobenzoate:CoA ligase results in a nearly 3-fold increase in the K m for ATP and a 4-fold drop in catalytic activity indicating the importance of this residue in ATP binding (44). In the present work, steady state kinetic results from whole cells expressing the fadD K222A allele demonstrated a 3-fold increase in K m for ATP and a 4-fold decrease in catalysis, which supports the hypothesis that Lys 222 of the E. coli FACS was involved in ATP binding. The predicted three-dimensional model of FadD places Lys 222 in a disordered loop connecting two antiparallel ␤-strands that faces the cleft hypothesized to represent the catalytic core of the enzyme (38).
Other potential key residues within this loop that may contribute to ATP binding were Tyr 213 and Thr 214 . The data presented here suggests that both Tyr 213 and Thr 214 were important for the catalytic competence of the enzyme. Whereas nearly 70% of the members of the adenylate-forming family of enzymes contain a Tyr at the position comparable with Tyr 213 in FadD, this work was the first to document a potentially important role for this residue in catalytic activity. We suggest that Tyr 213 contributes to the catalytic integrity of the enzyme as the mutant form FadD T312A was devoid of any enzymatic activity even though it was expressed at wild type levels and  bridge between the invariant glutamate of the second sequence element of the ATP/AMP signature (corresponding to Glu 361 of FACS) and the O-1 phosphate of AMP. Substitution of this glutamate to alanine in FadD results in complete loss of enzyme activity that results in an inability to transport longchain fatty acids. The predicted three-dimensional structure for FACS, which we have generated, suggests that this glutamate lies in proximity to Thr 213 and Tyr 214 , which is presumed to be involved in nucleotide binding (33).
An additional goal in the present work was to provide evidence linking the catalytic activity of FadD to long-chain fatty acid transport. Overath and colleagues (18) first showed that FadD was partially membrane associated and required for the conversion of exogenous long-chain fatty acids to CoA thioesters thereby rendering fatty acid transport unidirectional. This enzyme is clearly required for fatty acid import and activation in E. coli as revealed in the present studies showing that the import of [ 3 H]oleate is abolished in a ⌬fadD strain. Additionally, specific mutations corresponding to the ATP/AMP signature motif that reduced or abolished enzyme activity likewise reduced or abolished the import of exogenous long-chain fatty acids. Indeed, these data fully support Overaths hypothesis that vectorial acylation is one underlying biochemical mechanism that governs fatty acid transport.
We have previously shown that the long-chain fatty acid transport system in E. coli: 1) is partially shock sensitive, 2) requires ATP generated by either substrate-level or oxidative phosphorylation, and 3) requires the proton electrochemical gradient across the inner membrane for maximal proficiency (45). More specifically, the transport of oleate requires ATP generated through the substrate level or oxidative phosphorylation, which reflects the ATP requirement of FadD as a component of this transport apparatus. In addition, these previous studies demonstrate that the process of long-chain fatty acid transport is linked to the proton electrochemical gradient across the inner membrane. In other words, the long-chain fatty acid transport system of E. coli requires both intracellular pools of ATP and an energized inner membrane. Fig. 7 shows a model of how we hypothesize FadD functions in the vectorial transport of exogenous fatty acids. FadD is found both within the particulate and soluble fractions of the cell (18,19,29). There is evidence that the enzyme is recruited to the membrane although the signal that promotes this association remains elusive (19). We have shown that an energized plasma membrane is necessary for optimal fatty acid transport and have suggested that uncharged fatty acids simply flip across the inner membrane by diffusion and are then abstracted from the membrane by FadD (45). It is temping to speculate that FadD specifically associates with the membrane in response to exogenous protonated long-chain fatty acids in the membrane. In the present work, we show that the mutant forms of FadD containing substitutions within the ATP/AMP signature motif are expressed at levels equivalent to the wild type and are found within both the particulate and soluble fractions of the cell. These data argue that there must be regions of FadD outside the catalytic core, which are responsible for membrane association. In addition, whereas the signal that promotes membrane association is undefined, it seems plausible that an increase in free fatty acid concentration within the inner membrane was sufficient to promote the association of ATP-bound FadD to the membrane. The formation of the fatty acyl adenylate effectively pulls the free fatty acids from the membrane for subsequent thioesterification to coenzyme A. The net result is the metabolic trapping of fatty acyl-CoA within the cell rendering the process unidirectional. Upon release of fatty acyl-CoA, the enzyme becomes ATP-bound and the cycle repeats.
The present work has established a platform of information, which serves for our current investigations toward understanding the mechanism and specificity of substrate and cofactor binding of this enzyme further. Clearly, this line of investigation is essential to understand the mechanisms that underpin the general role of fatty acyl-CoA synthetases and the particular role of the E. coli FadD in the coupled import and activation of exogenous long-chain fatty acids.