Cloning, Sequencing, and Expression of the fudD Gene of Escherichia coli Encoding Acyl Coenzyme A Synthetase*

In the enteric bacterium, Escherichia coli, acyl coen- zyme A synthetase (fatty acid:CoA ligase (AMP-form-ing) EC 6.2.1.3) activates exogenous long-chain fatty acids concomitant with their transport across the inner membrane into metabolically active CoA thioesters. These compounds serve as substrates for acyl-CoA de-hydrogenase in the first step in the process of B-oxi- dation. The acyl-CoA synthetase structural gene, fadD, has been identified on clone 6D1 of the Kohara E. coli gene library and by a process of subcloning and complementation analyses shown to be contained on a 2.2- kilobase NcoI-CZaI fragment of genomic DNA. The polypeptide encoded within this DNA fragment was identified following T7 RNA polymerase-dependent induction and estimated to be M, = 62,000 using SDS- polyacrylamide gel electrophoresis. The N-terminal amino acid sequence of acyl-CoA synthetase was deter- mined by automated sequencing to be Met-Lys-Lys-Val-Trp-Leu-Asn-Arg-Tyr-Pro. Sequence analysis of the 2.2-kilobase NcoI-CZoI fragment revealed a single open reading frame encoding these amino acids as the first 10 residues of a protein with a molecular weight of 62,028. The initiation codon for methionine was TTG. Primer extension of total in vivo mRNA from two fudD-specific oligonucleotides

polyacrylamide gel electrophoresis. The N-terminal amino acid sequence of acyl-CoA synthetase was determined by automated sequencing to be Met-Lys-Lys-Val-Trp-Leu-Asn-Arg-Tyr-Pro. Sequence analysis of the 2.2-kilobase NcoI-CZoI fragment revealed a single open reading frame encoding these amino acids as the first 10 residues of a protein with a molecular weight of 62,028. The initiation codon for methionine was TTG. Primer extension of total in vivo mRNA from two fudD-specific oligonucleotides defined the transcriptional start at an adenine residue 60 base pairs upstream from the predicted translational start site. Two FadR operator sites of the f d D gene were identified at positions -13 to -29 (OD,) and positions -99 to -1 15 (ODz) by DNase I footprinting. Comparisons of the predicted amino acid sequence of the E. coli acyl-CoA synthetase to the deduced amino acid sequences of the rat and yeast acyl-CoA synthetases and the firefly luciferase demonstrated that these enzymes shared a significant degree of similarity. Based on the similar reaction mechanisms of these four enzymes, this similarity may define a region required for the same function.
Exogenous long-chain fatty acids (CI2-Cls) represent an * This work was supported by National Science Foundation Grant 9104646 and by American Heart Association Grant 901063 (to P. N. B.) and by National Institutes of Health Grant GM38104 (to C. C . D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the CenBankTM/EMBL Data Bank with accession number(s) L02649.
$To whom correspondence and reprint requests should be addressed Dept. of Biochemistry, Suite G01, Molecular Sciences Bldg., 858 Madison Ave., University of Tennessee College of Medicine, Memphis,  important class of hydrophobic compounds that can serve as a sole carbon and energy source to support the growth of the enteric bacterium Escherichia coli. The acquisition of these nutrients from the environment prior to metabolic utilization by cyclic @-oxidation in E, coli occurs by an energy-dependent, protein-mediated process. For long-chain fatty acids destined for @-oxidation, this process minimally requires the products of the fudL and fudD genes. The fudL gene encodes an outer membrane-bound protein (FadL) that binds exogenous longchain fatty acids with a relatively high affinity and by some unknown mechanism transfers these compounds across the outer membrane (1)(2)(3)(4). The fudD gene encodes the inner membrane-associated acyl coenzyme A synthetase (acyl-CoA synthetase (fatty acidcoenzyme A ligase (AMP-forming), EC 6.2.1.3)) (5). This enzyme catalyzes the esterification of fatty acids into metabolically active CoA thioesters concomitant with transport. The mechanisms that govern the transfer of long-chain fatty acids from FadL across the periplasmic space and the inner membrane to the acyl-CoA synthetase remain largely undefined. There is some evidence for an oleic acid binding protein in the inner membrane that has been postulated to be a H+/long-chain fatty acid co-transporter (6,7).
Acyl-CoA synthetases catalyze the formation of fatty acyl-CoA by a two-step mechanism that proceeds through the pyrophosphorolysis of ATP (8). Mi E. coli contains a single acyl-CoA synthetase which has been purified to homogeneity (9,lO). This enzyme has broad chainlength specificity giving VmaX values ranging from 2632 nmol/ min/mg of protein for lauric acid (C12) to 135 nmol/min/mg of protein for hexanoate (C,) (9). Maximal activities associated with this enzyme are found with fatty acids ranging in length between C12 and Clel (9). Overath and colleagues (11) proposed that acyl-CoA synthetase of E. coli was required for long-chain fatty acid transport and coined the term vectorial acylation to describe the role of this enzyme. Although the precise role of this enzyme in fatty acid transport is not well defined, it is clear that it plays a pivotal role in this process by catalyzing the formation of metabolically active CoA thioesters for subsequent degradation or incorporation into phospholipids (5,9,11).
The structural gene for acyl-CoA synthetase (fudD) was identified by Overath et ul. (5) who mapped this locus to the 40-min region of the E. coli chromosome. In this pioneering work, this enzyme was proposed to be partially membraneassociated and was shown to activate both mono-and poly-25513 Sequencing of the fadD Gene of E. coli unsaturated fatty acids. Acyl-CoA synthetase from E. coli has been estimated to have a native molecular weight of 120,000 based on elution profiles of the purified enzyme on a G-200 column (5). Kameda and Nunn (9) estimated a monomeric molecular weight of 47,000 and proposed that the enzyme is a dimer. Acyl-CoA synthetase activity is induced by oleate, but to a lower relative level, when compared to the levels of induction documented for three other enzymes required for long-chain fatty acid degradation (@-hydroxyacyl-CoA dehydrogenase, enoyl-CoA hydratase, and @-ketothiolase) (5). The fadD gene, like the fadBA, fadE, and fadL genes, is part of the fatty acid degradative regulon under the control of the transcriptional regulator FadR (11)(12)(13). The present work describes the cloning, sequencing, and expression of the acyl-CoA synthetase structural gene (fadB) of E. coli. This work stems from our goal to define the underlying biochemical mechanisms that govern long-chain fatty acid transport in enteric bacteria prior to metabolic utilization.

EXPERIMENTAL PROCEDURES
Bacterial Strains-The E. coli strain JM103 [A(lac pro) thi strA endA sbcB hsdR(F' traD36 proAB lacP ZAM15)] was used for the propagation of M13 derivatives. For routine plasmid propagation and the generation of thefadD88 strain PN235, C600 (F-thi-1 leuB6 lacy1 tonAZl supE44) was used. Strains RS3010 (fadR) and LS6928 (fadD88 zea::TnlO fadR) have been described elsewhere (14,15). The fadD88 strain PN235 was generated by P1 transduction of strain C600 using a phage stock grown on strain LS6928. Bacterial cultures were grown at 37 "C in a Lab Line gyratory shaker in 2YT (16), Luria broth (LB; 16), or Tryptone broth (TB; 17). When minimal medium was required, medium E supplemented with vitamin B, (17) was used. Carbon sources, sterilized separately, were added to final concentrations of 25 mM glucose, 25 mM potassium acetate, 5 mM decanoate, or 5 mM oleate. As required, amino acids were added to a final concentration of 0.01%. When required to maintain plasmids, antibiotics were added to 100 pg/ml ampicillin, 40 pg/ml kanamycin, 10 pg/ml tetracycline, and 40 pg/ml chloramphenicol. Growth of bacterial cultures was routinely monitored using a Klett-Summerson colorimeter equipped with a blue filter.
Identification of the fadD Gene in the Kohara Gene Library-The complete miniset of X clones of the Kohara library (18) were graciously provided by Dr. Y. Kohara (DNA Research Center, National Institute of Genetics, Mishima 411, Japan). Eleven clones representing the 40-min region of the E. coli chromosome (5E12, 4B8, 12H7, 3E12,9F2, 7F2,6D1,12B3,15D5,19H3, and 12C7) were propagated on the bacterial strain NM621 as previously described (19). Lysates were used to infect the XCI857 lysogen derived from strain PN235 at a multiplicity of infection of 1 as described by Miller (17). Following absorption, 1 ml of LB was added and the cells were allowed to recover for 30 min at 30 "C. Following recovery, the cells were pelleted by centrifugation, resuspended in the original volume of Medium E, plated on oleate minimal agar plates, and incubated at 30 "C for 72 h. At 72 h, colonies that were able to grow on oleate as a sole carbon and energy source were identified in cells infected with phage DNA from clones 7F2 and 6D1. These two clones were confirmed to restore the ability of the fadD strain PN235 to grow on oleate (Ole') following a second round of lysogenic complementation. Clones 7F2 and 6D1 from the Kohara library were propagated in strain NM621 for DNA isolation on TB agarose plates (24). Plaques giving nearly confluent lysis were visible 12-14 h later at which time the plates were flooded with 3 ml of 50 mM Tris-HC1, pH 7.5, 100 mM NaC1, 10 mM MgSO,, 0.01% gelatin (SM). A-DNA was isolated using the hexadecyltrimethylammonium bromide (CTAB) method (20). Briefly, 20 ml of each phage stock (4 X 10'' plaque-forming units/ml) was incubated with DNase I at 20 pg/ml for 5 min and then clarified by centrifugation tube, and 10 ml of DEAE-cellulose slurry (80% in SM) was added (10,000 X g for 20 min). The supernatants were transferred to a new and incubated at room temperature for 30 min in an angled rotator. The DEAE-cellulose was pelleted by centrifugation, the supernatants were transferred to a new tube, and EDTA was added to 20 mM and Tris-HC1, pH 8.0, was added to 100 mM. Proteinase K was added to the mixture to a final concentration of 50 pg/ml which was then heated at 45 "C for 15 min. Hexadecyltrimethylammonium bromide was added to a final concentration of 0.5% and incubated at 68 "C for 4 min. The samples were cooled on ice and centrifuged 10,000 X g for 30 min. The DNA pellets were dissolved in 6 ml of 1.2 M NaCl and DNA precipitated by the addition of 2 volumes of absolute ethanol. The final DNA samples were resuspended in 200 pl of 10 mM Tris-HC1, pH 7.5, 5 mM EDTA (TE), analyzed by agarose gel electrophoresis, and used in subcloning experiments as described below.
ClaI or HindIII, ligated into the plasmid vector pACYC177 (21), and CZoning and Sequencing-DNA from clone 6D1 was restricted with transformed into the fudD88 strain PN235. Restriction, ligation, plasmid isolation, and transformation procedures have been described previously (22,23). Ole' transformants were identified in both sets of ligation mixtures. Analysis of restriction patterns generated revealed that restriction fragments from the Ole+ transformants from the ClaI ligation mixture and the HindIII ligation mixture overlapped. As the insert from the ClaI digest was shown to be smaller (3.4 kb),' this plasmid, designated pN300, was used for all further study. Restriction fragments from pN300 were subcloned into either pACYC177 or pACYC184 (21) as described under "Results," yielding the plasmids and complementation patterns illustrated in Fig. 1B. The sequencing strategy of the ClaI insert from pN300 is illustrated in Fig. 1C. The series of M13 clones were sequenced using either the lac2-specific upstream primer (5'-GTTTTCCCAGTCACGAC-3'), the M13 universal primer (5'-GTAAAACGACGGCCAGT-3'), or with fadD-specific oligonucleotides by the dideoxy chain-terminating method of Sanger et al. (24) using Sequenase (v 2.0; U. S. Biochemicals). As shown in Fig. lC, pN300 was sequenced across the Sal1 and HindIII restriction sites usingfadD-specific oligonucleotides to ensure proper alignment between these three fragments of DNA. Sequencing reactions were resolved on a standard 8% polyacrylamide gel (3). All oligonucleotides used in this study were synthesized on a Pharmacia LKB Biotechnology Inc. Gene Assembler Plus.
Analysis of Acyl-CoA Synthetase Actiuity-Bacteria (wild-type and fadD strains containing the collection of fadD' and fadD clones) were grown to midlog phase (6 X 10' cells/ml) in TB or TB supplemented with 5 mM oleate and 0.5% Brij 58 (TBO) and with antibiotics as required. Cells were harvested by centrifugation, washed twice with Medium E, resuspended to a density of 1.2 X lo9 cells/ml in 10 mM Tris-HC1, pH 7.5, and lysed by three cycles of sonication at 0 "C. Acyl-CoA synthetase activities were determined in sonicated cell extracts as described by Kameda and Nunn (9). The reaction mixtures contained 200 mM Tris-HC1, pH 7.5, 2.5 mM ATP, 8 mM MgCl,, 2 mM EDTA, 20 mM NaF, 0.1% Triton X-100, 10 p M [3H]oleate, 0.5 mM coenzyme A, and cell extract in a total volume of 0.5 ml. The reactions were initiated with the addition of coenzyme A, incubated at 35 "C for 10 min, and terminated by the addition of 2.5 ml of isopropyl alcohoh-heptane:l M H,SO, (40101). The radioactive oleic acid was removed by organic extraction using n-heptane (9). Oleoyl-CoA formed during the reaction remained in the aqueous fraction and was quantified by scintillation counting. Protein concentrations in the enzyme extracts were determined using the Bradford assay and bovine serum albumin as a standard (25). The values presented represent the average from at least three independent experiments.
Ouerexpression of Acyl-CoA Synthetase-The 3.4-kb ClaI fragment (fadD+) from pN300 was gel-purified and ligated with a ClaI to BamHI linker. Linkers were purified and phosphorylated using bacterial alkaline phosphatase prior to restriction (22). Following ligation of the linkers, the fragment was restricted with BamHI, repurified, and ligated into the BamHI site of the expression plasmid pCDl3O. pCD130 is derived from pT7-5 and contains the fadR gene in the orientation opposite to the T7 promoter to maintain stability OffadD' (26). Both orientations of the fadD gene were obtained yielding plasmid pN321 and pN324. Plasmid pN324 contained fadD+ under the T7 promoter while pN321 had fadD+ in the orientation opposite to the T7 promoter. The plasmids pN321 and pN324 were transformed into strain BL21 (DE3)(plysS) and expressed following induction with isopropyl-1-thio-@-D-galactopyranoside (27). Following induction and labeling with [35S]methionine, cells were harvested, resuspended in SDS sample buffer, boiled, and resolved on a 12% SDS-polyacrylamide gel using the Laemmli buffer system (28). Following electrophoresis. the gels were dried and subjected to autoradiography for 2-24 h. Partial Purification of Acyl-CoA Synthetase and N-terminal Amino Acid Sequencing-Acyl-CoA synthetase was partially purified from a The abbreviation used is: kb, kilobase pair(s).
Cloning and Sequencing of the fadD Gene of E. coli 2.i.i 15 500-ml culture of strain RLZl(pLysS) harboring the fadD' expression plasmid pN324 after induction. Following induction with isopropyl if-D-thiogalactopyranoside, cultures were grown for an additional 2 h and cells were harvested by centrifugation. The cell pellets were washed twice in minimal medium E, resuspended in 50 mM potassium phosphate, pH 8.0, and disrupted by three cycles of sonication a t 0 "C. The sonicated extract was clarified hy centrifugation (12.000 X g for 15 min). The supernatant was centrifuged a t 60.000 X g for 2 h, and the memhrane pellet was discarded. Acyl-CoA synthetase was partially purified from the supernatant hy batch DEAE-cellulose chromatography and by ammonium sulfate fractionation using the conditions described by Overath et ai. (5) and Kameda et ai. (10). The final protein sample was subjected to preparative electrophoresis on a 12% SDS-polyacrylamide gel and electrophoretically transferred to a prewetted polyvinylidene difluoride (Immohilon Transfer, Whatman) in 10 mM 3-(cyclohexylamino)-l-propanesulfonic acid, pH 11.0, 10% methanol (29). Following electrophoretic transfer, the position of the acyl-CoA synthetase was identified hy staining the polyvinylidene difluoride memhrane briefly with Ponceau red. The strip cont.aining acyl-CoA synthetase was excised, extensively washed with high performance liquid chromatography grade water, dried a t room temperature, and stored a t -70 "C. The N-terminal amino acid sequence from this sample was determined using an Applied Rio-Systems 470A gas phase protein sequenator equipped with an Applied HioSystems Model 120A in-line detector for phenylthiohydantoinderived amino acids from each cycle of Edman degradation at the Harvard University Microchemistry Facility.
Identification of the FadR Binding Site by DNase I Footprinting- T h e 353-base pair Sau3A fragment containing thefadD promoter was gel-purified and ligated into the RarnHI site of Ml3mpl9, M13mpl8, and pUC18 thereby generat.ing clones MD21 (to sequence the top strand of the fad11 promoter-containing fragment), MD20 (to sequence the bottom strand of the fadD promoter-containing fragment), and pN330, respectively. Plasmid pN330 was purified and used as a source of DNA for gel shifts and DNase I footprinting assays. Protein-DNA gel retention assays (gel shifts) and DNase I footprinting were carried out essentially as descrihed by DiRusso et a/. (13) using purified FadR. The concentrations of FadR used in these experiments are given in the appropriate figure legends. For gel shifts, the 414base pair RcoRI-Hind111 fragment from pN330 was gel-purified and laheled with [n-:'2P]dATP using the Klenow fragment of DNA polymerase (22). The binding of FadIi to an [a-:"P]dATP-Iabeled fragment containing the fadH operator was used as an internal control as these parameters were previously well defined (13). FadR hinding was estimated as the conversion of the fast mobility complex (unbound DNA) to the slow mobility complex (FadH-bound). Quantitation was carried out with a Riolmage computer-assisted analysis system (MilliCen/Riosearch). For DNase I footprinting, pN330 was restricted with EcoRI (top strand) or Hind111 (hottom strand) and PuulI (cleaves only within the vector), and the promoter-containing fragments were gel-purified. Following purification, the Hindlll-PuulI and EcoRI-1'uuII fragments were 5' end-laheled with [ y -T ] ATP using polynucleotide kinase and restricted with RcoRI and HindIII, respectively. The appropriate'"P-labeled EcoRI-Hind111 (top strand of the fad11 promoter labeled) or Hindlll-EcoRI (bottom strand of the fadl1 promoter labeled) fragments were gel-purified and used for DNase I footprinting using the conditions described by DiRusso et ai. (13). The concentrations of FadR used in these experiments are given in the appropriate figure legends. Clones MD21 and MD20 were sequenced using the EcocoHI-specific primer and the HindIlI-specific primer, respectively, and sequencing reactions were run adjacent to the DNase I footprint reactions on a 6% standard sequencing gel to accurately position the FadR hinding site(s) within thefndD promoter (17).
The analyses of the DNA sequence of the fadD gene and the amino acid sequences of acyl-CoA synthetase. yeast acyl-CoA synthetase, rat acyl-CoA synthetase, and firefly luciferase were done using the Wisconsin Genetics Computer Group programs (3.5) and D N A Inspector I1 (TextCo Inc., West Lehanon. N H ) .
Materials-Reagents and enzymes used for sequencing, transcription mapping, and restriction were purchased from L' . S. I3iochemicals. Rethesda Research Laboratories. and New England Hiolahs. Reagents used for oligonucleotide synthesis were purchased from ARN/Riogenex and Pharmacia. [n-:"'S]dATP, [ m -"I']dATl'. [ y -"PI ATP, [:"'S]methionine, and ["Hloleate were ohtained from Du Pont-New England Nuclear. Antihiotics and other supplements for hacterial growth were purchased from Difco and Sigma. All other chemicals were ohtained from standard suppliers and were of reagent grade.

RESULTS
Cloning thP fadD GenP-The fadD gene was mapped by Overath et al. (5) to the 40-min region of the E . coli chromosome (5). Phage from 11 clones (5E12.4R8, 12H7,.7E12,9F2, 7F2,6D1, 12R.7, 15D5, 19H.7, and 20H4) of the X gene library generated by Kohara ~t al. (18) were transduced into the XCI857 lysogen of the fadD strain PN2.75 (Fig. 1A) for complementation analysis. DNA from clones 7F2 and 6D1 was able to complement the fadDRR defect. X-DNA was purified from clone 6D1 and restricted with Chi, and fragments were ligated into ClaI-restricted pACYClT. Plasmid DNA from one of the transformants that complemented the fadDRR mutation (acquired the ability to grow on the long-chain fatty acid oleate (Ole')) was purified and designated pNROO ( Arrorrs represent MI3 suhclones of the fadl1 gene including those isolates generated using I;xoIII. These clones were sequenced using either the universal primer or the lac%-specific primer. Arrorr,s preceded hy an astprisl: (*) indicate the direction and extent of sequencing usinp either M13-derived clones or pN130 sequenced usinp /odl)-specific. oligonucleotides. The shaded region (h'c.oI-('lnl) represents t hr sequence presented in Fig. 4. nomic DNA. A series of subclones of pN300 were constructed t o delineate the end points of the fadD gene for further studies using DNA sequencing (Fig. 1R). One of these subclones, pN308, contained a 2.7-kb HindIII-ClaI fragment which complemented fadDRR. When a small (500-base pair) NcoI fragment was removed from pN308 to generate pN309, the smallest subclone complementing the fadD88 defect was generated.
Acyl-CoA Synthetase Expression-Acyl-CoA synthetase activities were monitored in the wild-type strain K12, the fadR strain RS3010, and the fadR fadD strain LS6928 harboring the fadD' and fad11 plasm?ls shown in Fig. 1R. As shown in Fig. 2 A , acyl-CoA synthetase activity was inducible %fold in the presence of the long-chain fatty acid oleate in the prototrophic strain K12. Acyl-CoA synthetase activities in the fadR strain RS3010 grown under both conditions were comparable to the levels found in E. coli K-12 grown in the presence of oleate. As expected, no acyl-CoA synthetase activity was observed in the fadD fadH strain LS6928. The fadR fadD strain LS6928 harboring the fadD and fadD' plasmids pN300, pN104, pN305, pN306, pN307, pN.108, or pN.109 had acyl-CoA synthetase activities that reflected their complementation patterns (Fig. 2R).
Identification of the fadD Gene Product and N-terminal Amino Acid Sequence Analvsis of Acyl-CoA Synthetaye-The 3.4-kb ClaI fragment from pN300, containing the entire fadD gene, was isolated, ligated with a ClaI to RamHI linker, and cloned into the T7 expression plasmid pCD130 (13) to yield the plasmids pN321 and pN324 (Fig. 3A). A protein with an M , of 62,000 was identified by SDS-polyacrylamide gel electrophoresis in extracts of cells harboring pN324 following induction which was presumed to be acyl-CoA synthetase (Fig. 3 R ) . A second protein that was poorly produced relative t o acyl-CoA synthetase with an M , of 22,000 was also identified in these extracts which, based on the sequence data described below, was presumed to be distinct from the fadl) gene. Neither protein was produced in cells harboring pN321 (fadD+ in the reverse orientation to the T7 promoter) or pCD130 (plasmid vector)(data not shown). A polypeptide with a n M , of 15,000 was produced from both constructs as well as  The M , = 62,000 pol-ypeptide assumed to he acyl-CoA synthetase was partially purified from HI,21(plvsS)(pN324) following induction as described under "Experimental Procedures" and subjected to automated N-terminal amino acid sequence analysis as described under "Experimental Procedures." The N-terminal amino acid sequence WAS shown to he

Met-Lys-Lys-Val-Trp-Leu-Asn-Arg-Tvr-Pro.
Sequence of the fadD Gene-The entire 3.4-kh Clal fragment from pN300 was sequenced as shown in Fig. 1C. The fadD gene was shown by complementation to be localized on a 2.2-kb NcoI-ClaI fragment. Sequence analvsis of this fragment of DNA as shown in Fig. 4 (2230 base pairs) revealed a single open reading frame beginning with AT(; at nucleotide 241 encoding a polypeptide consisting of 580 amino acid residues with a molecular weight of 64,406. This open reading frame did not encode the N-terminal amino acid sequence defined from the purified acyl-CoA synthetase (see above). Furthermore, this ATG was five nucleotides upstream from the adenine residue defined as the transcriptional initiation site by primer extension (see below). Careful analysis of the reading frame revealed that the amino acid sequence Leu-Lys-Lys-Val-Trp-Leu-Asn-Arg-Tyr-Pro, starting at nucleotide 307, had 9 of the 10 amino acid residues defined for the purified protein. The notable difference was in the N-terminal amino acid. The DNA sequence predicted a leucine (TTG) while the protein sequence defined a methionine (ATG). We propose that the TTG beginning at nucleotide 307 represented the initiation codon and encodes a methionine residue (thus giving the predicted N-terminal amino acid sequence Met-

Lys-Lys-Val-Trp-Leu-Asn-Arg-Tyr-Pro). This proposal was
based on three lines of evidence. First, the data obtained from the N-terminal amino acid sequence of the acyl-CoA synthetase clearly indicated the presence of a methionine at this position. Second, the protein sequence beginning at nucleotide 240 did not result in an amino acid sequence that would indicate that this protein was post-translationally modified (i.e. cleavage of a signal peptide and/or modificntion of the N-terminal amino acid). Third, UUG can act as an alternative initiation codon (30). Assuming that the TTG at nucleotide 307 encodes the initiation methionine, the predicted size of the acyl-CoA synthetase from our expression of pN324 (62,000) was in close agreement with that predicted from the DNA sequence (62,028). The coding sequence for the M , =

Cloning and
Sequencing of the fadD Gene of E. coli region WAS ligated into M13mp18, M13mp19, and pUC18 as described under "Experimental Procedures." Using DNAprotein gel retention assays, this fragment was shown to bind FadR with high affinity indicating it contained a FadR binding site (Fig. 6A). The apparent KEQ of this fragment for FadR was estimated to be 1 X IO-" M. DNase I footprinting defined two FadR operator sites within the fadD promoter-containing fragment in pN330 (Fig. 6, R and C). The first FadR operator site, designated Onl, covered 17 base pairs and included the sequence defined above a t position -29 to -13. The second site at position -115 to -98 has been designated Onn. Within Om2 was the sequence CTGGT which was also found in OD, as well as the fadR and fadL operator sites. Overall, On2 had 9 of 17 nucleotides in common with t.he proposed FadR consensus binding site (13).2 The region protected by FadR from DNase I digestion at On:! was identified a t concentrations of FadR that were 10-50-fold higher that that used to identify Onl indicating this site had a low binding affinity for FadR compared to On and Onl.
Comparison of the E. coli Acyl-CoA Synthetase with the Rat Acyl-CoA Synthetase, Yeast Acyl-CoA Synthetase, and Firefly Luciferase-Using the Genetics Computer Group programs BESTFIT and GAP, we compared the deduced amino acid sequence from the E. coli acyl-CoA synthetase to that deduced for the rat (32) and the yeast enzymes (33) and firefly luciferase (34). These four enzymes were found to have extensive similarities along their entire lengths (48-51% sequence similarity when conservative amino acid substitutions are considered). Overall these four enzymes were 24-27% identical. Further analysis of these data indicated that two regions of these enzymes were highly conserved (Fig. 7). In the first region (residues 200-273 of the E . coli acyl-CoA synthetase). there was a 32-35'36 sequence identity which was extended to 5 3 4 7 % similarity when conserved residues were included to residues 255-327 of the yeast enzyme and residues 262-334 of the rat enzyme. These was no apparent similarity observed between the firefly luciferase and the three acyl-CoA synthetases in this region. In the second region (amino acid residues 353-455 of the E. coli enzyme), 34-440; of the amino acid residues were identical and 6 0 4 5 % were similar for all four enzymes (Fig. 7 H ) . Suzuki et al. (32) proposed that for the rat enzyme, this second region may represent the ATP binding site. Part of our current research efforts are being directed at addressing whether or not this region of the E . coli acyl-CoA synthetase actually represents a nucleotide binding region.

DISCUSSION
In the present paper, we report the cloning, sequencing, and expression of the fadD gene of E. coli encoding acyl coenzvme A synthetase. The fadD gene was identified in clone 6D1 from the Kohara gene library and subsequently shown to be encoded within a 2.2-kb NcoI-ClaI fragment of genomic DNA by complementation analysis. The expression of the fudD gene was monitored both by following acyl-CoA synthetase activities in the collection of fadD and fadD' plasmids and by following induction of the fadD+ gene using T7 RNA polymerase. Acyl-CoA synthetase levels were only 2-fold inducible in the presence of the long-chain fatty acid oleate which differed from the levels of induction observed for other fad gene products (acyl-CoA dehydrogenase, enoyl-CoA hydratase, @-hydroxyacyl-CoA dehydrogenase, and @-ketothiolase) (5,26). The DNA sequence of the fadD gene predicted a protein with 558 amino acid residues and a molecular weight of 62,028 starting with UUG as the translational initiation codon. This alignment was confirmed by N-terminal amino acid sequence analysis of purified acyl-CoA synthetase. No evidence was obtained which indicated the acyl-CoA synthetase was post-translationally processed (i.e. the presence of a signal sequence and/or N-terminal amino acid modification). The transcriptional initiation site of the fadD gene was determined to be an adenine residue 60 nucleotides upstream from the initiation site of translation using primer extension of two different fudD-specific oligonucleotides. The T7 RNA polymerase experiments estimated the size of the E. coli acyl-CoA synthetase to be 62,000 which was in agreement with that deduced from the DNA sequence and that defined from earlier reports.
The fadD gene contained two operator sites for the binding of FadR. The first (OD1) was slightly upstream (-29 to -13) from the transcriptional start and had a relatively high affinity for FadR (Keq -1 X lo-' M ) . The estimated affinity of the fadD operator 0~~ toward FadR (-1 X M) was nearly an order of magnitude lower than that defined for the fadB promoter (3 x 10"' M) (13). This operator site was appropriately positioned to block transcription when filled by FadR as it overlapped the presumptive -10 region. The second operator site (OD2) was found 114 base pairs upstream from the transcriptional start (-115 to -99). This site had considerably less affinity toward FadR (Keq 2 1 x lo-' M ) as estimated using DNase I footprinting. Studies are being conducted to determine the precise contribution of this site as well as the contribution of OD1 in the expression of the fadD gene.
DiRusso and her colleagues (13) demonstrated that the long-chain fatty acyl-CoA molecule is the inducer of the fatty acid degradative genes by showing that inclusion of these compounds (in nanomolar concentrations) prevented FadR binding to the fadB operator site in DNA-protein gel retention assays while long-chain fatty acids did not (13). Due to the affinity of O D~, it is expected that when this site is filled, transcription of fadD+ is likely to be turned off or maintained at a low basal level. The role of O D 2 is less clear; perhaps this second site regulates a second promoter upstream from the primary promoter identified here. Alternatively, there may be cooperative interaction between proteins bound at OD1 and Om that contribute to enhanced repression of the fadD promoter. When fatty acids are present in the growth media, the acyl-CoA synthetase enzymatically produces an increased intracellular pool of long-chain fatty acyl-CoA molecules that results in the derepression of transcription. The net result is coordinate induction of transcription of the genes involved in fatty acid transport, activation, and degradation, including fadD.
The presence of a UUG translation initiation codon for acyl-CoA synthetase was noted in the course of the present study. This initiation codon is relatively rare (found in 1% of E. coli genes) and in some cases acts to down-regulate the expression of a given protein (30). It is plausible that the production of acyl-CoA synthetase may also be down-regulated. In the case of the rnd gene of E. coli (encoding RNase D), replacement of the native UUG with AUG results in an 11-fold increase in RNase D expression (34). We are presently investigating whether the acyl-CoA synthetase activity is subject to comparable regulation.
Acyl-CoA synthetase is crucial for the uptake of exogenous long-chain fatty acids that are destined to be utilized as a source of carbon and metabolic energy. This enzyme has been proposed to vectorially transport long-chain fatty acids across the inner membrane with a concomitant thioesterification to the CoA derivatives. Acyl-CoA synthetase functions by generating a fatty acid-adenylate intermediate which in turn is converted into a fatty acyl-CoA. In this respect, this enzyme is likely to bind ATP (see "Discussion" below) and thus may be functionally analogous to the ATPase component of bacterial permeases that represents class of transport proteins collectively referred to as "traffic ATPases" (36). There is evidence that suggests a H+/long-chain fatty acid co-transporter is present in the inner membrane (6,7). If this is the case, this postulated component must interact directly with the acyl-CoA synthetase in the vectorial transport of longchain fatty acids.
In the course of our analysis of the fadD gene, we compared the deduced amino acid sequence of the E. coli acyl-CoA synthetase to the rat liver acyl-CoA synthetase, the yeast acyl-CoA synthetase, and firefly luciferase. Although the four proteins shared significant similarity along their entire lengths, a higher degree of similarity in the amino acid sequences from these four proteins was identified toward their carboxyl ends (amino acid residues 353-455 of the E. coli enzyme). Due to the common mechanism of action of these proteins, it seems likely that this region may be of functional importance. Suzuki et al. (32) proposed that this region of the rat acyl-CoA synthetase represented the ATP binding domain. This proposal is, in part, based on the similar enzyme mechanism proposed for firefly luciferase. The firefly luciferase reaction, like the acyl-CoA synthetase reaction, proceeds by a two-step mechanism that results in the formation of an adenylated intermediate. In both the firefly luciferase and the acyl-CoA synthetase enzyme mechanisms, there is a reaction between the carboxyl group of the substrate (lucifern and long-chain fatty acid, respectively) and ATP to form the adenylated intermediate. In both cases, this activation reaction requires the pyrophosphorolysis of ATP. The formation of an adenylated intermediate requires that ATP bind transiently to the enzyme as part of the catalytic cycle (8,32). The second region of the E. coli acyl-CoA synthetase that shared similarity with the yeast and rat enzymes was more centrally located along the linear amino acid sequence of these proteins (amino acid residues 200-273 of the E. coli enzyme). As this homology was not seen with the firefly luciferase, this region may specify a component unique to acyl-CoA synthetases (e.g. fatty acid binding domain and/or the coenzyme A binding domain). The significance of these similarities with respect to the function of the E. coli acyl-CoA synthetase is presently under investigation.