Supplemental Material can be found at: http://www.jbc.org/cgi/content/full/M400100200/DC1 THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 279, No. 30, Issue of July 23, pp. 31717–31726, 2004 Printed in U.S.A.

Structural Basis of the Substrate-specific Two-step Catalysis of S Long Chain Fatty Acyl-CoA Synthetase Dimer*□ Received for publication, January 6, 2004, and in revised form, May 13, 2004 Published, JBC Papers in Press, May 15, 2004, DOI 10.1074/jbc.M400100200

Yuko Hisanaga‡§¶, Hideo Ago‡储**, Noriko Nakagawa‡‡, Keisuke Hamada‡, Koh Ida‡§§, Masaki Yamamoto¶¶, Tetsuya Hori‡, Yasuhiro Arii‡, Mitsuaki Sugahara‡储, Seiki Kuramitsu‡‡, Shigeyuki Yokoyama‡‡, and Masashi Miyano‡储 储储

Long chain fatty acyl-CoA synthetases are responsible for fatty acid degradation as well as physiological regulation of cellular functions via the production of long chain fatty acyl-CoA esters. We report the first crystal structures of long chain fatty acyl-CoA synthetase homodimer (LC-FACS) from Thermus thermophilus HB8 (ttLC-FACS), including complexes with the ATP analogue adenosine 5ⴕ-(␤,␥-imido) triphosphate (AMP-PNP) and myristoyl-AMP. ttLC-FACS is a member of the adenylate forming enzyme superfamily that catalyzes the ATP-dependent acylation of fatty acid in a two-step reaction. The first reaction step was shown to propagate in AMP-PNP complex crystals soaked with myristate solution. Myristoyl-AMP was identified as the intermediate. The AMP-PNP and the myristoyl-AMP complex structures show an identical closed conformation of the small C-terminal domains, whereas the uncomplexed form shows a variety of open conformations. Upon ATP binding, the fatty acid-binding tunnel gated by an aromatic residue opens to the ATP-binding site. The gated fatty acid-binding tunnel appears only to allow one-way movement of the fatty acid during overall catalysis. The protein incorporates a hydrophobic branch from the fatty acid-binding tunnel that is responsible for substrate specificity. Based on these high resolution crystal structures, we propose a unidirectional Bi Uni Uni Bi Ping-Pong mechanism for the two-step acylation by ttLC-FACS.

(1– 6). In single cell organisms, long chain fatty acyl-CoA synthetase (LC-FACS)1 also participates in the transport of various xenobiotic fatty acids. The LC-FACSs in Escherichia coli and Saccharomyces cerevisiae, FadD and Faa1p/Faa4p, are involved in the vectorial movement of exogenous fatty acids across the plasma membrane together with the respective fatty acid transport proteins, FadL and Fat1p (7). This movement results in the accumulation of fatty acyl-CoA esters, the first process of ␤-oxidation (7–10). In mammals LC-FACS is involved in the physiological regulation of various cellular functions through the production of long chain fatty acyl-CoA esters, which have been reported to affect protein transport (11, 12), enzyme activation (13), protein acylation (14), cell signaling (15), and transcriptional regulation (16). Three types of FACS have been defined with respect to the length of the aliphatic chain of the substrate: short, medium, and long chain fatty acyl-CoA synthetases (SC-, MC-, and LCFACSs; EC 6.2.1.1, EC 6.2.1.2, and EC 6.2.1.3, respectively) (5, 6). These utilize C2-C4, C4-C12, and C12-C22 fatty acids as substrates, respectively. Recent studies report that Fat1p and its homologues possess very long chain fatty acyl-CoA synthetase activity and belong to the superfamily of the adenylate forming enzymes (7, 17, 18). All FACSs catalyze a magnesiumdependent multisubstrate reaction, resulting in the formation of fatty acyl-CoA (19, 20). The reaction requires ATP, a fatty acid, and CoA with an overall reaction scheme as described in Reaction 1. fatty acid ⫹ CoA ⫹ ATP 3 fatty acyl-CoA ⫹ PPi ⫹ AMP

Long chain fatty acyl-CoA synthetases participate in the first reaction step of long chain fatty acid degradation in various organisms from bacteria to mammals and including plants * This work was supported in part by the National project on protein structural and functional analysis funded by Ministry of Education, Culture, Sports, Science, and Technology of Japan. 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 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains a supplemental figure. The atomic coordinates and structure factors (codes 1ULT, 1V25 and 1V26) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/). ¶ Present address: Dept. of Cardiac Physiology, National Cardiovascular Center Research Institute, Suita, Osaka 565-8565, Japan. ** To whom correspondence may be addressed. Tel.: 81-791-58-2815; Fax: 81-791-58-2816; E-mail: [email protected]. 储储 To whom correspondence may be addressed. Tel.: 81-791-58-2815; Fax: 81-791-58-2816; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

REACTION 1

The FACS family catalyzes the formation of fatty acyl-CoA in two discrete steps: 1) the formation of a fatty acyl-AMP molecule as a stable intermediate (Reaction 2) and 2) the formation of a fatty acyl-CoA molecule as the final product (Reaction 3). fatty acid ⫹ ATP 3 fatty acyl-AMP ⫹ PPi fatty acyl-AMP ⫹ CoA 3 fatty acyl-CoA ⫹ AMP REACTIONS 2

AND

3

1 The abbreviations used are: LC-FACS, long chain fatty acyl-CoA synthetase; ttLC-FACS, long chain fatty acyl CoA synthetase from T. thermophilus HB8; AMP-PNP, adenosine 5⬘-(␤,␥-imido) triphosphate; FACS, fatty acyl-CoA synthetase; SC, short chain; MC, medium chain; LC, long chain; PheA, phenylalanine-activating A domain; DhbE, 2,3dihydroxybenzoate-activating E domain; MES, 4-morpholineethanesulfonic acid; L motif, linker motif; A motif, adenine motif; G motif, gate motif.

31717

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

From the ‡Structural Biophysics Laboratory, the 储Highthroughput Factory, the ‡‡Structurome Research Group, and the ¶¶Coherent X-ray Optics Laboratory, RIKEN Harima Institute at SPring-8, Kouto, Mikazuki, Sayo, Hyogo 679-5148, Japan, the §Institute for Protein Research, Osaka University, Suita, Osaka 565-0871, Japan, and the §§Graduate School of Integrated Science, Yokohama City University, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan

31718

Structural Basis of LC-FACS Catalysis

MATERIALS AND METHODS

Expression and Purification—The ttLC-FACS gene was amplified by PCR using the primers 5⬘-ATATCATATGGAAGGGGAAAGGATGAACGCGTTCCCAA-3⬘ and 5⬘-ATATAGATCTTTATTAGGCGCCTCCGTAGTAGTTCTTGTAC-3⬘ from T. thermophilus HB8 cDNA. The amplified gene fragment was cloned into the pT7Blue (Novagen). After confirmation of the nucleotide sequence, the ttLC-FACS gene was li-

gated into the expression vector pET-11a (Novagen) at the NdeI/BamHI sites. The ttLC-FACS expression plasmid was transformed into E. coli strain BL21 (DE3) (Novagen) for overexpression. The cells were cultured at 37 °C in the presence of 100 ␮g/ml of ampicillin in LB medium for 20 h and harvested by centrifugation. The centrifuged pellet was resuspended in 20 mM Tris-HCl (pH 8.0) and heated to 70 °C for 11.5 min. All subsequent steps were performed at 4 °C. After centrifugation, ammonium sulfate was added to the supernatant, and the 30 – 60% (w/v) fraction was applied sequentially to HiTrap-Q HP and HiTrapBlue HP columns (Amersham Biosciences) in the presence of buffer containing 20 mM Tris-HCl (pH 8.0), 1 mM EDTA, 1 mM dithiothreitol, and 10 ␮M phenylmethanesulfonyl fluoride. The purity of samples was verified using SDS-PAGE stained by Coomassie Brilliant Blue, which confirmed the presence of a single band of about 60 kDa eluted from the HiTrap-Blue HP chromatography (supplemental figure). Assay of ttLC-FACS Activity—The activity of acyl-CoA synthetase was assayed at 25 °C by an enzyme coupled spectrophotometric method (34). The assay measures the rate of AMP formation by coupling the reaction of acyl-CoA synthetase with those of adenylate kinase, pyruvate kinase, and lactase dehydrogenase and then detects the oxidation of NADH at 334 nm with a spectrophotometer (SHIMAZU UV-2100PC). The standard reaction mixture for this assay contains 0.1 M Tris-HCl (pH 7.4), 5 mM dithiothreitol, 1.6 mM Triton X-100, 7.5 mM ATP, 10 mM MgCl2, 1 mM CoA, 0.2 mM potassium phosphoenolpyruvate, 0.15 mM NADH, 20 ␮g/ml adenylate kinase, 30 ␮g/ml pyruvate kinase, 30 ␮g/ml lactate dehydrogenase, and 5 ␮g/ml ttLC-FACS. The magnesium-free assay was performed in the presence of 10 mM EDTA instead of MgCl2. Crystallization—The crystals of the uncomplexed ttLC-FACS were obtained by the hanging drop vapor diffusion method at 20 °C. The crystallization drops were prepared by mixing 3 ␮l of ttLC-FACS solution (9.2 mg/ml purified ttLC-FACS, 20 mM Tris-HCl, pH 8.0, 10 ␮M phenylmethanesulfonyl fluoride) with 3 ␮l of reservoir solution (0.1 M sodium citrate, pH 5.5, 0.2 M ammonium sulfate, 21% (w/v) polyethylene glycol 4000, and 0.1 M guanidine hydrochloride). The crystals grew to dimensions 0.3 ⫻ 0.15 ⫻ 0.1 mm within 1 month. The substitution of 10 mM CoA for guanidine hydrochloride improved both the reproducibility and quality of the crystal. These improved crystals were used to prepare heavy atom derivatives by soaking with a solution containing 21% (w/v) polyethylene glycol 4000, 0.1 M sodium citrate (pH 5.5), 0.2 M ammonium sulfate, 10 mM CoA, and 1 mM thimerosal for 5 days. Crystals of the AMP-PNP complex were obtained by hanging drop vapor diffusion at 20 °C by mixing 3 ␮l of ttLC-FACS solution (14 mg/ml ttLC-FACS, 10 mM AMP-PNP, 10 mM CoA, 20 mM Tris-HCl, pH 8.0, 10 ␮M phenylmethanesulfonyl fluoride) and 3 ␮l of reservoir solution (50 mM MES/NaOH, pH 6.5, 0.1 M ammonium sulfate, 15% (w/v) polyethylene glycol mono methyl ester 5000). The myristoyl-AMP complex crystals were prepared by soaking AMP-PNP complex crystals into a sodium myristate solution (125 ␮M) containing 10 mM MgCl2 and 10 mM AMP-PNP for 24 h. X-ray Data Collection and Processing—The diffraction data were collected using beam lines (BL26B1, BL41XU, and BL45XU) (35) at SPring-8 at 100 K. Before flash cooling, the crystals were washed with the reservoir solution containing 20% (v/v) glycerol to avoid formation of ice. The mercury derivative, thimerosal crystal was measured at four different wavelengths to obtain a data set for the multiwavelength anomalous diffraction method. All of the images were processed by HKL2000 (36). Structure Determination and Refinement—Initial phases were calculated by the program SOLVE (37) using multiwavelength anomalous diffraction data from the mercury derivative up to 2.0 Å resolution (Tables I and II). These phases were improved by the program RESOLVE (38), and initial model building was performed by the program ARP/wARP (39). The crystal structures of the uncomplexed ttLC-FACS, the AMP-PNP complex, and the myristoyl-AMP complex were determined by the molecular replacement method using AMoRe (40) and Molrep (41) and the atomic coordinates of the mercury derivative ttLCFACS. Manual model building and the subsequent iterative crystallographic refinement were performed using the programs O (42), CNS (43), and REFMAC5 (44). The dictionaries of AMP-PNP and myristoylAMP used for the restrained crystallographic refinement were prepared by QUANTA/CHARMm (Accelrys) and REFMAC5 (44). Although the reflections for Rfree validation (5% for apo data set and 10% for ligand complexes) were independently selected for each diffraction data set, the molecular replacement method and the simulated annealing method starting from over 3000 K were performed to remove the model bias at the initial refinement cycle for each structure (43). The crystal structures of apo ttLC-FACS at 2.55 Å, AMP-PNP complex at 2.3 Å, and

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

The esterification of fatty acids by LC-FACS has been proposed to proceed via a Bi Uni Uni Bi Ping-Pong mechanism (21) based on extensive kinetic studies of the rat enzyme (22). However, to date the fatty acyl-AMP intermediate has not been isolated nor utilized experimentally as substrate for the second step (Reaction 3) or as product for the first step (Reaction 2) in reverse catalysis (23, 24) in contrast to the acetyl-AMP or butyryl-AMP for SC- or MC-FACSs (25–27). The crystal structures of four adenylate forming enzymes have so far been reported: luciferase (28), phenylalanine-activating A domain (PheA) (29), 2,3-dihydroxybenzoate-activating E domain (DhbE) (30), and acetyl CoA synthetase (SC-FACS) (31, 32). The members of this superfamily show a 20 –30% amino acid sequence identity with several highly conserved regions, and all catalyze the formation of acyl-AMP from ATP and a carboxylated molecule including a fatty acid and an amino acid. All four enzymes consist of a large N-terminal and a small C-terminal domain, with the catalytic site formed at the junction between the two domains. The relative positions of the C- and N-terminal domains may change upon substrate binding. In the absence of substrate, the C-terminal domain of luciferase was shown to be in an open conformation. Upon substrate binding, a closed conformation is adopted where the C- and N-terminal domains approach one another, thus reducing the accessibility of the active site to solvent as reported in the crystal structures of substrate-bound PheA (29), DhbE (30) and SC-FACS (31, 32). However, there does appear to be a certain amount of variability in the extent of these conformational changes because the uncomplexed form of DhbE showed a closed conformation (30). Furthermore the structure of SCFACS with bound substrate analogue for the second half-reaction revealed the existence of the reverse closed conformation of the C-terminal domain compared with those of PheA and DhbE, suggesting that another large structural rearrangement is needed between the first and second reactions as supported by the recent crystal studies of SC-FACS with AMP (31, 32). LC-FACS has been presumed to catalyze acylation in the same manner because of the sequence similarity and the results of extensive mutation studies based on homology modeling using the known crystal structures (8, 10). However, the structurefunction relationship of LC-FACS was still unclear particularly with respect to the formation and subsequent processing of the acyl-AMP intermediate. We have determined the first three-dimensional structure of a LC-FACS domain swap homodimer from an extreme thermophile, Thermus thermophilus HB8 (ttLC-FACS) (33) overexpressed in E. coli. We also determined the structure of the enzyme complexed to AMP-PNP, a nonhydrolyzable ATP analogue. Furthermore, we identified the acyl adenylate intermediate as myristoyl-AMP in the complex crystal structure using the crystals of the AMP-PNP complex acylated by soaking in myristate solution. Based on these high resolution structures, we propose a two-step catalytic mechanism for ttLC-FACS involving a single closed conformation induced by ATP binding (Reaction 2). ttLC-FACS possesses a gated fatty acid-binding tunnel with a dead end branch in each monomer. The unidirectional movement of fatty acid is proposed as a unidirectional Bi Uni Uni Bi Ping-Pong mechanism based on these crystal structures.

Structural Basis of LC-FACS Catalysis

31719

TABLE I Statistics of diffraction data The beam line/detector was BL26B1/Jupiter 210; the space group was P212121; the cell dimensions were a ⫽ 55.7 Å, b ⫽ 125.3 Å, and c ⫽ 211.0 Å; and the resolution was 50.0 –2.0 Å.

Wavelength (Å) Number of reflections Measured Unique Completeness (all/highest shell; %) I/␴(I) (all/highest shell) Rmerge (all/highest shell; %)

Mercury peak

Mercury edge

Mercury remote 1

Mercury remote 2

1.0081

1.0088

1.0000

1.0126

547,659 95,617 95.0/88.3 10.8/7.7 6.9/30.7

546,426 95,545 95.0/88.1 10.7/7.7 6.8/30.8

560,590 96,115 95.5/89.2 9.8/7.3 7.1/33.2

524,024 95,275 94.5/86.9 9.8/7.3 6.7/29.9

TABLE II Statistics of refinement AMP-PNP

myristoyl-AMP

BL41XU/MarCCD 0.7090 50–2.2 P212121 a ⫽ 56.0, b ⫽ 124.7, c ⫽ 212.5

BL45XU/RAXIS-V 0.9840 50–2.3 P212121 a ⫽ 64.5, b ⫽ 101.2, c ⫽ 176.5

BL45XU/RAXIS-V 1.0000 50–2.5 P212121 a ⫽ 64.5, b ⫽ 101.4, c ⫽ 176.9

300,383 48,374 97.3/90.7

223,397 51,688 99.3/96.2

299,868 41,062 99.8/97.5

18.5/9.8 3.6/14.4

11.3/4.8 10.2/30.3

9.7/7.5 12.0/32.3

50–2.55 0.21/0.28 0.25/0.31

48.8–2.3 0.19/0.20 0.24/0.26

46.1–2.5 0.18/0.18 0.24/0.29

0.006 1.22

0.009 1.43

0.010 1.33

Rmerge ⫽ ⌺ 兩 I ⫺ 具I典 兩⌺ I. R and Rfree ⫽ ⌺ 兩Fo ⫺ Fc兩 / ⌺ Fo, where the free reflections (5% for Apo and 10% for liganded in the total used) were held aside for Rfree throughout refinement. a b

myristoyl-AMP complex at 2.5 Å were refined with Rfree values of 0.25, 0.24, and 0.24, respectively. The geometrical quality was checked using PROCHECK (45). The current atomic coordinates have been deposited to the Protein Data Bank (46) with accession codes 1ULT, 1V25, and 1V26. The statistics of data collection and refinement on ttLC-FACS are summarized in Tables I and II. Amino Acid Sequence Analysis—The amino acid sequence of ttLCFACS was aligned with other LC-FACSs and members of other adenylate forming families in the Swiss-Prot data base (47) using PSI-BLAST (48) and T-coffee (49). After the sequence alignment using T-coffee (49), phylogenetic analysis was performed using PHYLIP (50). The amino acid sequence of four structurally characterized family members luciferase (28), PheA (29), DhbE (30), and SC-FACS (31) were aligned structurally with ttLC-FACS using QUANTA (Accerylys) with manual modification. RESULTS

Enzyme Activity and Sequence Analysis—The overexpressed and purified ttLC-FACS protein was shown to catalyze the esterification of a number of long chain fatty acids with CoA in the presence of Triton X-100. No activity was detected in the absence of detergent or Mg2⫹ ions (3, 19, 22, 27) (Fig. 1). Myristate (C14) is the most efficiently processed fatty acid at 25 °C, followed by palmitate (C16). The esterification of stearate (C18) and laurate (C12) was also catalyzed but at lower efficiency. In contrast, ttLC-FACS did not catalyze the esterification of the unsaturated fatty acids mysteroleic and palmitoleic acids. The amino acid sequence of ttLC-FACS was aligned with other LC-FACSs (Fig. 2A). Although the overall sequence homology is low, about 20% sequence identity or less to other LC-FACSs, there are conserved regions corresponding to the linker (L), adenine (A), and gate (G) motifs as well as the P-loop (Thr184-Thr-Gly-Thr-Thr-Gly-Leu-Pro-Lys192), the phosphate-

FIG. 1. Acyl chain length specificity of ttLC-FACS activity. Acyl chain length specificity of ttLC-FACS was measured with an enzymecoupled assay at 25 °C (see text for details) (34). Saturated fatty acids of a variety of chain lengths: C8 (caproate, 8), C10 (decanoate, 10), C12 (laurate, 12), C14 (myristate, 14), C16 (palmitate, 16), C18 (stearate, 18), and C20 (arachidate, 20) were assessed as well as the C14 (myristoleic acid, 14u) and C16 (palmitoleic acid, 16u) unsaturated fatty acids at concentrations of 250, 500, and 750 ␮M. The mean values ⫾ standard deviations from n ⫽ 3 experiments are shown on the graph.

binding site (Fig. 2A) (7, 8, 10, 51). The motifs were designated based on the ttLC-FACS structures presented in this paper. The L motif (Asp432-Arg-Leu-Lys-Asp-Leu437) contains the peptide that acts as a linker between the N- and C-terminal domains, the A motif (Gly323-Tyr-Gly-Lue-Thr-Glu-Thr329) contains the adenine-binding residue Tyr324, whereas the G motif (Val226-Pro-Met-Phe-His-Val-Asn-Ala-Trp234) contains the gate residue Trp234 and the surrounding mobile peptide as described below.

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

Data set Beam line/detector Wavelength (Å) Resolution (Å) Space group Cell dimensions (Å) No. of reflections Measured Unique Completeness (all/highest shell; %) I/␴(I) (all/highest shell) Rmerge (all/highest shells; %)a Refinement Resolution (Å) R (all/ highest shell)b Rfree (all/ highest shell) b Root mean square deviation In bond distances (Å) In bond angles (°)

Apo

31720

Structural Basis of LC-FACS Catalysis

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 2. Amino acid sequence alignments of LC-FACSs and adenylate forming enzymes. A, multisequence alignment of LC-FACS from T. thermophilus (ttLC-FACS, this work), human (LCFA_HUMAN, P41215), yeast (LCF1_YEAST, P30624), and E. coli (LCFA_ECOLI, P29212). The regions marked with thin bars and shaded bars correspond to ␤-strand and ␣-helix, respectively. The boxed areas denoted with bold letters correspond to conserved motifs of LC-FACSs: G, A, and L motifs as well as the P-loop. Filled squares, open circles, filled circles, and filled triangles indicate residues believed to be involved in dimer formation, fatty acid binding, magnesium ion binding, and adenylate binding, respectively. B, sequence alignment based on the known structures of adenylate forming enzymes: DhbE (Protein Data Bank code 1mdb) (30), PheA (Protein Data Bank code 1amu) (29), SCFA (SC-FACS; Protein Data Bank code 1pg3) (31), Luci (luciferase; Protein Data Bank code 1lci) (28), and LCFA (ttLC-FACS; this work). The shaded and underlined letters correspond to ␣-helical and ␤-strand regions, respectively. The residues conserved among all the sequences are indicated by the boxed regions. The bar above the sequence corresponds to the linker regions. G, A, and L motifs in LC-FACSs and luciferase are indicated by the gray boxes, and the filled triangle under the sequence denotes the gating residue of Trp234 of ttLC-FACS.

Overall Structure—ttLC-FACS forms a domain swapped dimer (52). The monomers of the dimer interact at their Nterminal domains with a contact surface area of 3600 Å2 for

each monomer (Fig. 3). At the back of the domain swapping surface, there is a large electrostatically positive concave in the central valley of the homodimer (Fig. 3B). Each ttLC-FACS

Structural Basis of LC-FACS Catalysis

31721

monomer swaps the eight residues from Ala8 to Glu16 into a ridge on the surface of the other N-terminal domain (Fig. 3C). In the dimer interactions, Asp15 forms an intermolecular salt bridge with Arg176. The main chain carbonyl group of Glu16 forms an inter-molecular hydrogen bond with the side chain of Arg199, whereas Glu175 and Arg199 form an intermolecular salt bridge at the interface (Fig. 3C). The multisequence alignment of LC-FACS (Fig. 2A) reveals that residues corresponding to Asp15, Glu16, Arg176, and Arg199 are highly conserved among the LC-FACS family but not conserved in the other related enzyme families (Fig 2B). Therefore, this type of domain swapped homodimer may be a characteristic feature in LCFACS but not common in other adenylate forming enzyme families (Fig. 2) (28 –32). Each monomer of ttLC-FACS is composed of a large Nterminal domain (residues 1– 431) and a small C-terminal domain (residues 438 –541) that are connected by a six-amino acid peptide linker, the L motif (residues 432– 437) (Figs. 2 and 3D). The core region of the large N-terminal domain exhibits a fold similar to that seen in other adenylate-forming enzymes (28 –32). The N-terminal domain can be further divided into two subdomains: a distorted antiparallel ␤-barrel and two ␤-sheets that are flanked on both sides by ␣-helices forming an ␣␤␣␤␣ sandwich. The small C-terminal globular domain is comprised of a two-stranded ␤-sheet and a three-stranded antiparallel ␤-sheet that is surrounded by three ␣-helices. Fixation of the C-terminal Domain with ATP Binding—In the ttLC-FACS structures, the C-terminal domain adopts open and closed conformations depending on the presence of ligand. In the AMP-PNP complex structure of ttLC-FACS, the C-terminal domains are in the closed conformation with direct interactions formed between the C- and N-terminal domains (Fig. 4A, panel 3). The closed conformation of the C-terminal domain is also maintained in the complex structure with myristoylAMP (Fig. 4A, panel 4). Both of these complex structures with AMP-PNP and myristoyl-AMP are almost the same closed conformation, and superimposition of the four N- and C-terminal domains of the AMP-PNP and myristoyl-AMP complex structures yields average root mean square deviations of 0.34 Å (486 C␣ atoms) and 0.57 Å (57 C␣ atoms), respectively. In contrast in the uncomplexed structure, the C-terminal domains show two types of open conformations (Fig. 4A, panels 1 and 2). The

C-terminal domain is rotated by 54° relative to the N-terminal domain, when the structure of a monomer is compared with that of the other monomer in dimer. There are no direct interactions between the C- and N-terminal domains for both monomers of the dimer. In the closed conformation of ttLC-FACS three residues, Glu443, Glu475, and Lys527 in the C-terminal domain, stabilize the closed conformation by forming noncovalent interactions with residues of the L motif and the N-terminal domain. The side chains of Glu475 and Lys527 form salt bridges with the side chains of Arg433 and Asp436 in the L motif, respectively. In addition, the carboxyl oxygen of Glu443 forms a salt bridge with the side chain of Arg105 and a hydrogen bond with the side chain of His230 in the N-terminal domain. His230 and Glu443 are conserved among LC-FACS enzymes. Hydrogen bond formation between Glu443 and His230 is important because it mediates a conformational change in the G motif upon ATP binding as described below. Furthermore, the AMP moiety of the bound ATP molecule holds the C-terminal domain and the N-terminal domain together via an extensive hydrogen bond network. Fatty Acid-binding Tunnel with Open Gate upon ATP Binding—Compared with SC-FACS (31, 32), LC-FACS should incorporate a larger fatty acid-binding site to accommodate bulkier long chain fatty acids. A fatty acid-binding tunnel was identified in the N-terminal domain of each monomer (Fig. 4, B and C), which extends from the concave cavity in the central valley to the ATP-binding site and is surrounded by a large ␤-sheet including ␤12, ␤13, ␤14, ␤15, and an ␣-helix cluster including ␣g and ␣h (Figs. 3D and 4B). In the complex structure this tunnel is composed of a large central pathway that is divided into two distinct paths, the “ATP path” and the “center path, ” by the indole ring of Trp234 in the G motif (Fig. 4C). In addition, there is another pocket that branches from the central pathway named the “dead end branch. ” All three paths are separated by the indole ring of Trp234 in the uncomplexed structure. The fatty acid-binding tunnel closed by the indole ring of Trp234 in the uncomplexed structure opens upon AMP-PNP binding via hydrogen bond formation between ␤-phosphate and the ring nitrogen of Trp234 (Fig. 4C). At the same time the mobile C-terminal domain adopts the closed conformation as seen in the adenylate binding structures (Fig. 4A, panel 3). The

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 3. ttLC-FACS crystal structure. Ribbon representations of the ttLC-FACS dimer are shown (A). In the panel, the secondary structure of the C-terminal domain is colored in green. In the N-terminal domain, ␣-helix and ␤-sheet are colored in cyan and red, respectively, with the N-terminal domain-swapping peptide colored in yellow. The electrostatic potential surface map of ttLC-FACS dimer in the same orientation as the representation in A. Red represents negatively charged regions, and blue represents positively charged regions (B). Close-up view of the N-terminal peptide involved in domain swapping in the reverse orientation view to A (C). Residues with carbons colored in pink against a cyan surface of one monomer interacts with the concave surface of the other monomer colored in yellow. There are salt bridges at the domain swapping region. The monomer of ttLCFACS with each secondary structure feature is labeled according to the scheme given in Fig. 2A (D).

31722

Structural Basis of LC-FACS Catalysis

ATP-binding site. In the AMP-PNP complex structure, the bound AMPPNP molecule blocks the exit of the ATP path. 2) The center path opens at the central valley of the 2-fold dimer and is the entrance pathway for the fatty acid (B). 3) The dead end branch extends to ␣-helix h. The three paths join near Trp234 (W234). In the absence of ligand, the indole ring of Trp234 closes the junction (green), whereas in the presence of ligand the indole ring (red) opens the fatty acid-binding tunnel.

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 4. Conformational changes induced by ATP binding in ttLC-FACS. A, the open and closed conformations of the C-terminal domain in the uncomplexed and two complex ttLC-FACS structures. Panels 1 and 2 show the open conformation in the uncomplexed structure. The C-terminal domain conformation is slightly different for the two monomers of the dimer. Panels 3 and 4 show the identical closed conformation of the C-terminal domain for both the AMP-PNP and myristoyl-AMP structures. The backbone structures of the large Nterminal and small C-terminal domains are presented in cartoon and wire models colored in silver and yellow green, respectively. The bound ligands of AMP-PNP and myristoyl-AMP are shown as the space-filling models of Corey, Pauling, and Kolten. To facilitate the orientation of the C-terminal domain, Lys439 is shown as a stick model, and the ␤-strand 22 is shown as a cartoon model. B, the C␣ trace of the AMP-PNP complex ttLC-FACS dimer. The N-terminal domains are colored in cyan and blue, and the C-terminal domains in orange and yellow, respectively. The bound AMP-PNP and the fatty acid-binding tunnel are represented by a red space-filling model and a solid magenta surface, respectively. The entrance of the fatty acid-binding tunnel opens in the central valley at the dimer interface. The fatty acid-binding tunnel extends through the N-terminal domain from the central concave region to the ATP-binding site. The funnel cavity formed between the N- and C-terminal domains, and the central valley at the dimer interface is shown in yellow green and orange, respectively. C, the fatty acid-binding tunnel consists of three paths: 1) The ATP path connects with the

indole ring of Trp234 rotates by 29° and ⫺81° around the angles of ␹1 and ␹2, respectively in the complexed structures compared with the uncomplexed structure (Fig. 4C). The ring nitrogen of Trp234, which is free in the open conformation forms a hydrogen bond with the ␤-phosphate in the AMP-PNP complex structure. Furthermore, in the closed structures of ttLCFACS, the dead end branch is wider than that of the uncomplexed forms because of a shift in the flexible loop of the G motif, His230-Val-Asn-Ala-Trp-Cys-Leu236 by 0.9 Å (Figs. 2A and 4C). The average temperature factors of this loop in the uncomplexed and complexed structures of ttLC-FACSs are 45.4 and 13.1 Å2, respectively. Adenylate Binding—The crystal structures with the bound AMP-PNP allowed for a detailed characterization of adenylate binding in the closed conformation (Fig. 5). In the AMP-PNP complex structure, an AMP-PNP is bound to each monomer in a crevasse at the interface between the N- and C-terminal domains (Fig. 4A, panel 3). The AMP moiety is oriented by a combination of hydrophobic interactions as well as hydrogen bonds with Arg433 and Lys435 from the L motif, Lys439 and Trp444 from the C-terminal domain, and a number of residues from the N-terminal domain (Fig. 5A). After soaking the complex crystals of AMP-PNP in the presence of myristate, the adenylate intermediate myristoyl-AMP was clearly identified as the acylation product of Reaction 2 in the crystals (Fig. 5B). The ␣-phosphate mediates interactions between the N- and C-terminal domains (Fig. 5). The observed hydrogen bond network between the ␣-phosphate and the Nterminal domain is essentially the same in both the AMP-PNP and the myristoyl-AMP complex structures. The putative acidic oxygen of the ␣-phosphate (O2A) forms hydrogen bonds with the backbone nitrogen and the side chain hydroxyl group of Thr327 (Fig. 5C). Furthermore, O2A is coordinated to the essential magnesium ion (Mg2⫹). The Mg2⫹ is further bonded to the hydroxyl group of Thr184 and the carboxyl oxygen of Glu328 (Fig. 5D), and this is consistent with the mutation studies of FadD (8). In contrast to the N-terminal domain, the hydrogen bond network between the ␣-phosphate and the C-terminal domain differs in the AMP-PNP and myristoyl-AMP complex structures that may reflect the two distinct acylation steps of Reactions 2 and 3 (Fig. 5D). The hydrogen bond between the side chain amino group of Lys439 and the acidic oxygen (O1A) of the ␣-phosphate is only formed in the myristoyl-AMP complex, and the hydrogen bond distance between the nitrogen of the indole ring of Trp444 and O1A of the ␣-phosphate increases from 2.8 Å in the AMP-PNP complex to 3.4 Å in the myristoyl-AMP complex. There is rotation around the C5⬘-O5⬘ and the O5⬘-␣phosphorus bonds in the AMP moiety of 10° and 30° between the AMP-PNP and myristoyl-AMP complex structures, respectively (Fig. 5, C and D). In the pyrophosphate moiety of the AMP-PNP complex structure, the ␤-phosphate interacts with both the N- and C-terminal domains (Fig. 5A). O1B and O2B of the ␤-phosphate form hydrogen bonds to the amino group of Lys439 and the nitrogen of the indole ring of Trp234, respectively (Figs. 4C and 5D). The O2B forms additional interactions with the backbone nitrogen of Gly325 in the A motif via a tightly bound water molecule. Weaker electron density with higher temperature factors indi-

Structural Basis of LC-FACS Catalysis

cated that the ␥-phosphate interacts far less than either the ␣or ␤-phosphate moieties. Fatty Acid Binding—In the fatty acid-binding tunnel, the ATP path is a hydrophobic channel connected to the ATPbinding site (Figs. 4 and 5). The center path is the entrance site for the fatty acid and extends from the dimer interface along ␤-strand 13 to the ATP path (Fig. 4B). In the absence of ATP, the indole ring of Trp234 blocks the connection between the two paths as described (Fig. 4C). The center path is filled with ordered water molecules that can be seen in both the AMP-PNP and the myristoyl-AMP complex structures. These ordered water molecules connect to the bulk solvent region through the entrance of the center path. The entrance of the center path is

located at the concave surface in the central valley of the dimer interface (Fig. 4B), which generates a positive electrostatic potential because of the basic residues (Lys219, Arg296, Arg297, Arg321, Lys350, and Lys354) from each monomer (Fig. 3B). The third path, the dead end branch, extending from the fatty acid-binding tunnel to ␣-helix h (residues 235–243) is also gated by Trp234 in the uncomplexed structure (Fig. 4C). The bottom of the dead end branch has a hydrophilic environment generated from water molecules and polar side chains (His204, Ser209, and Thr214) (Fig. 5). The aliphatic hydrocarbon chain of the myristoyl-AMP inserts into the dead end branch via the ATP path (Fig. 5, B and C). The amino acid residues within a 4.5 Å radius of the myristoyl moiety are His204, Ala208, Thr214, Val231, Trp234, Cys235, Leu236, Ala239, Val299, Gly301, Gly323, Tyr324, Gly325, Leu326, Thr327, Pro331, Val332, and Gln335 of the N-terminal domain and Lys439 of the C-terminal domain. The aliphatic chain of the myristoyl-AMP is in an extended conformation with a kink between C9 and C10, resulting in a torsion angle of 59° (Fig. 5B). This gauche conformation results in the C10 to C14 portion of the aliphatic chain being positioned in the dead end branch (Fig. 5C). Structural Comparison of Adenylate-forming Enzymes—A comparison was made of the closed conformations of the structurally characterized adenylate forming enzymes including DhbE, PheA, and SC-FACS as well as ttLC-FACS by the superimposition of the bound adenine rings in the complex structures (Fig. 6) (29 –32). There are two-type of closed C-terminal domain conformations of SC-FACS complexed to the ligands, AMP (32) and propyl-AMP (31); one is the DhbE (30) and PheA (29) type, and the other is the ttLC-FACS type. These two conformations show almost reverse orientations of C-terminal domain relative to the N-terminal domain. The position of the C-terminal domain differs by 180° in the two types. The protein backbones are oriented in opposite directions after the linker region, resulting in the putative catalytic residue differences. (Fig. 6) (31). DISCUSSION

ATP-dependent Closure of the C-terminal Domain—Based on the crystal structures of both the uncomplexed and complexed ttLC-FACS, the ATP-dependent closure of the C-terminal domain is concluded to be the reactive conformation. This reactive conformation is stabilized by extensive noncovalent interactions involving residues conserved only among LC-FACSs. In addition this closed conformation should be maintained through the whole catalytic reaction (Figs. 4 and 5), meaning that the fatty acyl-AMP intermediate is unlikely to be released from the enzyme. This is consistent with extensive experimentation that has been unable to isolate such intermediates (23, 24). In contrast to LC-FACS, SC-FACS is suggested to adopt two different closed conformations for the two steps of the reaction (31, 32), whereas the catalytic Bi Uni Uni Bi PingPong mechanism should be the same in all FACSs (22, 25–27). If it is the case, the conformational change via the open form between the two distinct C-terminal forms in the two-step catalysis should take place in SC- or MC-FACSs, because these FACSs were shown to release and utilize the acetyl-AMP or butyryl-AMP as the two-step reaction intermediate, respectively, but not in LC-FACS (23, 24, 26, 27). When the closed conformation is stabilized by the binding of ATP, ttLC-FACS is able to readily catalyze the formation of the fatty acyl-AMP intermediate using a fatty acid entering the active site that opens upon ATP binding (Reaction 2). In fact, the first half-reaction of ttLC-FACS was shown to propagate in the AMP-PNP complex crystals soaked with myristate without crystal damage, resulting in the tightly bound myristoyl-AMP

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 5. The ligand binding site of ttLC-FACS complex structures. A, stereo representation of the ATP-binding site with bound AMP-PNP. The amino acid residues of the ATP-binding site are shown as stick models. The carbon atoms of the residues of the N- and Cterminal domains and the linker peptide are colored in green, magenta, and cyan, respectively, and other atoms are colored in red (oxygen), blue (nitrogen), and yellow (phosphorus). The sigmaA weighted Fo ⫺ Fc electron density map is 3␴ (silver) and 5␴ (orange) (57). B, the bound myristoyl-AMP in the ttLC-FACS complex structure shown is as a ball and stick model (sky blue). The sigmaA weighted Fo ⫺ Fc electron density map of bound myristoyl-AMP (gold) was contoured at 3␴. C, schematic drawing of the polar interactions in ttLC-FACS and bound myristoyl-AMP. The aliphatic chains from C1 to C9 and from C10 to C14 occupy the ATP path and the dead end branch of the fatty acidbinding tunnel, respectively. D, stereo representation of the loop structure around the substrate binding site. The loops (brown) surrounding the bound myristoyl-AMP (carbon atoms in sky blue) are shown with the putative catalytic residues (light green, carbon) and a superimposed bound AMP-PNP molecule. The bound magnesium ion is colored in pink. The regions corresponding to the LC-FACS specific conserved motifs are labeled as well as the P-loop.

31723

31724

Structural Basis of LC-FACS Catalysis

as the reaction intermediate (Fig. 5B). The myristoyl-AMP is the substrate for the second reaction with CoA, which takes place in the closed form without any further conformational rearrangement of the enzyme (Reaction 3). Indeed, the crystal structures of ttLC-FACS complexes with AMP-PNP and myristoyl-AMP showed the same closed conformation both before and after myristoyl-AMP formation; thus, the overall catalysis is concluded to proceed to completion, i.e. the release of fatty acyl CoA and AMP, in a fixed conformation of the C-terminal domain initiated by ATP binding. ATP Binding-dependent Conformational Changes—Prior to the binding of a fatty acid molecule, a molecule of ATP is required to bind into ttLC-FACS, an event that results in adoption of the closed conformation and the opening of the Trp234 gate in the fatty acid-binding tunnel (Fig. 4). In the closed conformation with bound ATP, the fatty acid-binding tunnel penetrates the N-terminal domain from the central valley of the dimer interface to the ATP-binding site connecting to the funnel cavity in both AMP-PNP and myristoyl-AMP complex structures (Figs. 4 and 5). A long chain fatty acid can enter the gate-opened fatty acid-binding tunnel through the center path from the central valley of the dimer (Fig. 4B). The aliphatic carbon tail of the myristoyl-AMP extends into the bottom of the dead end branch via the ATP path in the complex structure (Fig. 5C). In the structure of the uncomplexed enzyme, the indole ring of Trp234 closes the junction between the ATP path and the additional paths in the fatty acid-binding tunnel. In the AMP-PNP complex structure, the Trp234 gate is open and the dead end branch is wider than that of the uncomplexed ttLC-FACS because of a shift of the G motif loop. This shift is the result of the formation of a hydrogen bond between ␤-phosphate and the ring nitrogen of Trp234 (Fig. 4C). The formation of a salt bridge between the side chains of His230 in the loop and Glu443 on the closing C-terminal domain, residues that are conserved in all except the E. coli LC-FACSs, may also be involved in the movement of the G motif (Fig. 5D). Unidirectional Binding and Release of Fatty Acid and Its Product Fatty Acyl-CoA—In the closed conformation of the enzyme, the fatty acid should travel unidirectionally through the gate-opened fatty acid-binding tunnel from the central valley of the dimer interface to the ATP-binding site (Figs. 4 and 5). The positively charged entrance to the tunnel located at the dimer interface should attract the negatively charged carboxyl

group of the long chain fatty acid (Fig. 3B). The final products, fatty acyl-CoA and AMP, should only be released from the side of the ATP-binding site after opening of the C-terminal domain. This is due to the fact that both the fatty acyl-CoA and AMP contain adenylate moieties that are too bulky to pass back through the fatty acid-binding tunnel to the central valley (Fig. 4B). Structural Determinants of Fatty Acid Specificity—The depth of the dead end branch of the fatty acid-binding tunnel of ttLC-FACS determines the chain length and specificity of the fatty acid substrate (Figs. 4C and 5). The in vitro enzyme assay demonstrated that myristate (C14) and palmitate (C16) are good substrates at 25 °C, whereas acylation of laurate (C12) and stearate (C18) proceeds but at a reduced level, which is consistent with previous results (19, 20). The distance between C14 and the bound water at the bottom of the fatty acidbinding tunnel is 3.5 Å in the myristoyl-AMP complex. Palmitate fits well into the tunnel because the extra carbon atoms occupy a further length of 2.6 Å. Thus, the depth of the dead end branch accounts for substrate specificity by its compatibility with the fatty acid chain length. The hydrophilic dead end branch may also contribute to the release of the final product, fatty acyl-CoA. In the AMP-PNP complex structure the dead end branch is lined with polar side chains and two bound water molecules. The hydrophilic environment is surrounded by the hydrophobic aliphatic chain of the bound myristoyl-AMP in the complex structure and would promote release of the fatty acyl-CoA from the fatty acid-binding tunnel enthalpically. The diameter of the opening at the center path (3.5 Å) may be a selectivity filter for distinguishing saturated fatty acids over unsaturated ones. The unsaturated fatty acids myristoleic acid and palmitoleic acid were not substrates for ttLC-FACS under the described assay conditions at 25 °C. The diameter of the tunnel for unsaturated fatty acids would need to be wider than that for the saturated fatty acid because of the rigid and bulky aliphatic chain containing the 9-cis double bond. However, it cannot be excluded that ttLC-FACS catalyzes acylation of these fatty acids to some extent at higher temperatures, closer to physiological for this enzyme, because assays were only carried out at 25 °C. Structural Basis of ttLC-FACS Acylation Catalysis—The molecular mechanism of LC-FACS is proposed to be compatible

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 6. Superimposed structures of the vicinity of linker peptides and bound adenylates in adenylate forming enzyme complexes in stereo. The adenylate complexed enzymes of the known structures, DhbE (Protein Data Bank code 1mdb) (30), PheA (Protein Data Bank code 1amu) (29), SC-FACS (Protein Data Bank code 1pg3) (31), and ttLC-FACS (this work) are superimposed around each bound adenosine moiety. The backbone of the linker region (Lys431-Asp-Arg-Leu-Lys-Asp-Leu437) including the L motif in ttLC-FACS complex structure and the corresponding peptides are presented as wire models (ttLC-FACS, thick violet; SC-FACS, red violet; DhbE, blue; PheA, light green). The bound myristoyl-AMP in the ttLC-FACS is represented as by thick green sticks, and other bound adenylates each shown in thin colored sticks. Arg433 and Lys439 of ttLC-FACS and the corresponding residues are also shown.

Structural Basis of LC-FACS Catalysis

31725

with the Bi Uni Uni Bi Ping-Pong based on the three high resolution structures of ttLC-FACS, which include the uncomplexed form and two liganded complexes (Fig. 7) (22, 53). The reaction scheme is summarized as follows. First, the binding of ATP induces the closed conformation and opening of the Trp234 gate of the fatty acid-binding tunnel (Fig. 7, A and B). Second, the fatty acid molecule binds and the fatty acylAMP intermediate is formed (Fig. 7B). Third, the pyrophosphate molecule leaves (Fig. 7C). Fourth, a CoA molecule binds and the final product fatty acyl-CoA is formed (Fig. 7, C and D). Fifth, the fatty acyl-CoA followed by the AMP leave after opening of the C-terminal domain (Fig. 7E) (21). This scheme is fully consistent with the extended kinetic studies (22). In contrast to SC- and MC-FACS as well as other adenylate forming enzymes, the deeply buried fatty acyl-AMP intermediate is not released in the tightly closed conformation of LC-FACS (25–27, 31, 32). For the first reaction step (Reaction 2), the short hydrogen bond between ␣-phosphorus and Trp444 (2.8 Å) results in an electron deficient ␣-phosphorus atom to form an electrophile in the AMP-PNP complex structure (Figs. 4C and 5D). The positive charge of the bound Mg2⫹ coordinated with O2A can additionally contribute to the electron deficiency of ␣-phosphorus. The fatty acid-binding tunnel connects to the ATP-binding site and opens to a plane formed of O5⬘, O2A, and O3A in the ␣-phosphate (Figs. 4C and 5, A and D). Thus, the carboxyl group of the fatty acid would approach the electron deficient ␣-phosphorus atom in the manner of the adjacent mechanism to form a putative pentacovalent intermediate (54). According to the adjacent mechanism, O1A in the pentacovalent intermediate should be negatively charged, which would be stabilized by the hydrogen bond between O1A and the indole ring of Trp444 in the first reaction step (Figs. 4C and 5D). Multiple hydrogen bonds formed between Lys439 and the adenylate moiety are likely to play an important role in the second step of catalysis (Reaction 3) (Fig. 5, C and D). In the bound myristoyl-AMP structure the side chain amino group of Lys439 forms hydrogen bonds with the carbonyl oxygen atom (2.7 Å) and the O1A of the ␣-phosphate (3.3 Å), generating an electron deficient carbonyl carbon and stabilizing the negative

charge on the AMP leaving group, respectively. Lys439 and Trp444 of the C-terminal domain are involved in both reactions in the two-step catalysis, and are conserved in yeast, rat, and human LC-FACSs but not in E. coli (Fig. 2A). In the multisequence alignment of LC-FACS proteins, there are three conserved motifs, the G, A, and L motifs, specific to LC-FACSs in addition to the P-loop, which is conserved among all classes of adenylate forming enzymes (Figs. 2 and 5D) (7, 8). The L motif may be important as a determinant for the alternative closed conformations of the C-terminal domain in the different classes of adenylate forming proteins as described. The L motif consensus sequence for all LC-FACSs is Asp432Arg-Xaa-Lys435 (Fig. 2A), whereas for the SC- and MC-FACSs as well as DhbE and PheA the consensus motif sequence is Gly-Arg-Xaa-Asp (Fig. 2B) (29 –32). The sequence of luciferase is close to that of LC-FACSs (Fig. 2B) (28). In LC-FACSs a linker missing the Gly residue should be less flexible compared with SC- and MC-FACSs. This extra flexibility conferred by the Gly residue may explain why SC- and MC-FACSs adopt two different closed conformations of the C-terminal domain with the bound AMP (32) and propyl-AMP (31), respectively (Fig. 6). This mobility of the C-terminal domain in SC- and MC-FACSs should allow the release and utilization of the acyl adenylate intermediate (25–27). During the final stages of preparation of this manuscript, new data has been published on a group of acyl-AMP ligases from Mycobacterium tuberculosis that release acyl-AMP as product. In these enzymes there is a conserved Gly residue that is the first residue of the L motif as in SC- and MC-FACSs. This is consistent with the role of the L motif (55). In ttLC-FACS the G motif contains the gating residue Trp234. Phe247 in the G motif of firefly luciferase may be the gating residue of luciferase (Fig. 2B), because luciferase possesses LC-FACS catalytic activity (56), and the Phe is equivalent to Trp234 of ttLC-FACS structurally. Based on the conservation of these motifs, luciferase is anticipated to evolve from an ancestral LC-FACS keeping its LC-FACS activity (56), whereas DhbE (30) and PheA (29) may have evolved from of SC-FACS. Thus, ttLC-FACS may be an archetype both of eubacterial and eukaryotic LC-FACSs, and the proposed acylation of fatty acid by ttLC-FACS may be a model for the ATP-

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

FIG. 7. The schematic mechanism for the unidirectional ordered catalysis by ttLC-FACS. In the proposed overall catalysis of the acylation, the substrate fatty acid is processed, unidirectionally through the fatty acid-binding tunnel in the N-terminal domain and funnel cavity at the interface of N- and Cterminal domains from the central valley of the dimer interface. This schematic is based on Fig. 4B. The binding of ATP to ttLC-FACS is the initial event in the catalysis process (A); it is the trigger for both closing the C-terminal domain and the opening and widening of the gated fatty acid-binding tunnel (B). The fatty acid-binding tunnel conveys the substrate molecule unidirectionally from the positively charged concave in the central valley of the dimer interface to the ATPbinding site (B). The pyrophosphate is released after the formation of a fatty acyl-AMP. A CoA then binds to the fatty acyl-AMP complex (C). The thiol group of the bound CoA attacks the acyl carbon of the fatty acyl-AMP (D). Opening the Nand C-terminal domains again, the fatty acyl-CoA and AMP products are released from the ttLC-FACS (E). Overall catalysis is represented after Cleland’s expression (F) (21).

31726

Structural Basis of LC-FACS Catalysis

dependent vectorial transport of fatty acid (Fig. 7). The precise biological function of ttLC-FACS remains to be clarified under physiological conditions. Acknowledgments—We thank C. Kuroishi, Y. Nodake, and Y. Yorinaga for their in preparing ttLC-FACS enzyme and A. Nakagawa, T. Tsukihara, and H. Mori for support on this project. We are grateful to B. Byrne and J. Abramson for critical reading of the manuscript. REFERENCES

Downloaded from www.jbc.org at Cohen Library, City College of New York (ACU3267) on July 22, 2009

1. O’Brien, W. J., and Frerman, F. E. (1977) J. Bacteriol. 132, 532–540 2. Mishina, M., Kamiryo, T., Tashiro, S., Hagihara, T., Tanaka, A., Fukui, S., Osumi, M., and Numa, S. (1978) Eur. J. Biochem. 89, 321–328 3. Krisans, S. K., Mortensen, R. M., and Lazarow, P. B. (1980) J. Biol. Chem. 255, 9599 –9607 4. Fulda, M., Shockey, J., Werber, M., Wolter, F. P., and Heinz, E. (2002) Plant J. 32, 93–103 5. Schomburg, D., and Schomburg, I. (ed.) (2001) in Springer Handbook of Enzymes, 2nd Ed., Vol. 2, pp. 186 –218, Springer-Verlag, Berlin 6. Schomburg, I., Chang, A., Ebeling, C., Gremse, M., Heldt, C., Huhn, G., and Schomburg, D. (2004) Nucleic Acids Res. 32, D431–D433 7. Black, P. N., and DiRusso, C. C. (2003) Microbiol. Mol. Biol. Rev. 67, 454 – 472 8. Weimar, J. D., DiRusso, C. C., Delio, R., and Black, P. N. (2002) J. Biol. Chem. 277, 29369 –29376 9. Zou, Z., Tong, F., Faergeman, N., Beargeman, C., Black, P. N., and DiRusso, C. (2003) J. Biol. Chem. 278, 16414 –16422 10. Black, P. N., DiRusso, C. C., Sherin, D., MacColl, R., Kundsen, J., and Weimar, J. D. (2000) J. Biol. Chem. 275, 38547–38553 11. Glick, B. S., and Rothman, J. E. (1987) Nature 326, 309 –312 12. Pfanner, N., Glick, B. S., Arden, S. R., and Rothman, J. E. (1990) J. Cell Biol. 110, 955–961 13. Lai, J. C., Liang, B. B., Jarvi, E. J., Cooper, A. J., and Lu, D. R. (1993) Res. Commun. Chem. Pathol. Pharmacol. 82, 331–338 14. Li, Z. N., Hongo, S., Sugawara, K., Sugahara, K., Tsuchiya, E., Matsuzaki, Y., and Nakamura, K. (2001) J. Gen. Virol. 82, 1085–1093 15. Murakami, K., Ide, T., Nakazawa, T., Okazaki, T., Mochizuki, T., and Kadowaki, T. (2001) Biochem. J. 353, 231–238 16. van Aalten, D. M., DiRusso, C. C., and Knudsen, J. (2001) EMBO J. 20, 2041–2050 17. Watkins, P. A., Lu, J.-F., Steinberg, S. J., Gould, S. J., Smith, K. D., and Braiterman, L. T. (1998) J. Biol. Chem. 273, 18210 –18219 18. Fraisl, P., Forss-Petter, S., Zigman, M., and Berger, J. (2004) Biochem. J. 377, 85–93 19. Saunders, C., Voigt, J. M., and Weis, M. T. (1996) Biochem. J. 313, 849 – 853 20. Bar-Tana, J., Rose, G., and Shapiro, B. (1971) Biochem. J. 122, 353–362 21. Cleland, W. W. (1963) Biochim. Biophys. Acta 67, 104 –137 22. Bar-Tana, J., Rose, G., Brandes, R., and Shapiro, B. (1973) Biochem. J. 131, 199 –209 23. Bar-Tana, J., Rose, G., and Shapiro, B. (1973) Biochem. J. 135, 411– 416 24. Philipp, D. P., and Parsons, P. (1979) J. Biol. Chem. 254, 10785–10790 25. Berg, P. (1956) J. Biol. Chem. 222, 991–1013 26. Guranowski, A., Gunther Sillero, M. A., and Sillero, A. (1994) J. Bacteriol. 176, 2986 –2990 27. Takao, S., Ito, T., and Tamida, M. (1987) Agric. Biol. Chem. 51, 145–152

28. Conti, E., Franks, N. P., and Brick, P. (1996) Structure 4, 287–298 29. Conti, E., Stachelhaus, T., Marahiel, M. A., and Brick, P. (1997) EMBO J. 16, 4174 – 4183 30. May, J. J., Kessler, N., Marahiel, M. A., and Stubbs, M. T. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 12120 –12125 31. Gulick, A. M., Starai, V. J., Horswill, A. R., Homick, K. M., and EscalanteSemerena, J. C. (2003) Biochemistry 42, 2866 –2873 32. Jogl, G., and Tong, L. (2004) Biochemistry 43, 1425–1431 33. Yokoyama, S., Hirota, H., Kigawa, T., Yabuki, T., Shirouzu, M., Terada, T., Ito, Y., Matsuo, Y., Kuroda, Y., Nishimura, Y., Kyogoku, Y., Miki, K., Masui, R., and Kuramitsu, S. (2000) Nat. Struct. Biol. 7, 943–945 34. Hosaka, K., Mishina, M., Tanaka, T., Kamiryo, T., and Numa, S. (1979) Eur. J. Biochem. 93, 197–203 35. Kumasaka, T., Yamamoto, M., Yamashita, E., Moriyama, H., and Ueki, T. (2002) Structure 10, 1205–1210 36. Otwinowski, Z., and Minor W. (1997) Methods Enzymol. 276, 307–326 37. Terwilliger, T. C., and Berendzen, J. (1999) Acta Crystallogr. Sect. D Biol. Crystallogr. 55, 849 – 861 38. Terwilliger, T. C. (2000) Acta Crystallogr. Sect. D Biol. Crystallogr. 56, 965–972 39. Morris, R. J., Perrakis, A., and Lamzin, V. S. (2002) Acta Crystallogr. Sect. D Biol. Crystallogr. 58, 968 –975 40. Navaza, J. (1994) Acta Crystallogr. Sect. A 50, 157–163 41. Vagin, A., and Teplyakov, A. (1997) J. Appl. Crystallogr. 30, 1022–1025 42. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Crystallogr. Sect. A 47, 110 –119 43. Bru¨ nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., GrosseKunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr. Sect. D Biol. Crystallogr. 54, 905–921 44. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr. Sect. D Biol. Crystallogr. 53, 240 –255 45. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283–291 46. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalow, I. N., and Bourne, P. E. (2000) Nucleic Acids Res. 28, 235–242 47. Boeckmann, B., Bairoch, A., Apweiler, R., Blatter, M. C., Estreicher, A., Gasteiger, E., Martin, M. J., Michoud, K., O’Donovan, C., Phan, I., Pilbout, S., and Schneider, M. (2003) Nucleic Acids Res. 31, 365–370 48. Altschul, S. F., Madden, T. L., Scha¨ ffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389 –3402 49. Notredame, C., Higgins, D., and Heringa, J. (2000) J. Mol. Biol. 302, 205–217 50. Felsenstein, J. (1989) Cladistics 5, 164 –166 51. Saraste, M., Sibbald, P. R., and Wittinghofer, A. (1990) Trends Biochem. Sci. 15, 430 – 434 52. Liu, Y., and Eisenberg, D. (2002) Protein Sci. 11, 1285–1299 53. Pan, Y. H., Epstein, T. M., Jain, M. K., and Bahnson, B. J. (2001) Biochemistry 40, 609 – 617 54. Fersht, A. (1999) in Structure and Mechanism in Protein Science (Julet, M. R., and Hadler, G. L., eds) pp. 245–272, W.H. Freeman and Company, New York 55. Trivedi, O. A., Arora, P., Sridharan, V., Tickoo, R., Mohanty, D., and Gokhale, R. S. (2004) Nature 428, 441– 445 56. Oba, Y., Ojika, M., and Imouye, S. (2003) FEBS Lett. 540, 251–254 57. Collaborative Computational Project, Number 4 (1994) Acta Crystallogr. D Biol. Crystallogr. 50, 760 –763