Structural and Biochemical Studies of Substrate Selectivity in Ascaris

Article Cite This: Biochemistry XXXX, XXX, XXX−XXX

pubs.acs.org/biochemistry

Structural and Biochemical Studies of Substrate Selectivity in Ascaris suum Thiolases Michael R. Blaisse,† Beverly Fu,† and Michelle C. Y. Chang*,†,‡ †

Department of Chemistry and ‡Department of Molecular and Cell Biology, University of California, Berkeley, California 94720-1460, United States S Supporting Information *

ABSTRACT: Thiolases are a class of carbon−carbon bond forming enzymes with important applications in biotechnology and metabolic engineering as they provide a general method for the condensation of two acyl coenzyme A (CoA) substrates. As such, developing a greater understanding of their substrate selectivity would expand our ability to engineer the enzymatic or microbial production of a broad range of small-molecule targets. Here, we report the crystal structures and biochemical characterization of Acat2 and Acat5, two biosynthetic thiolases from Ascaris suum with varying selectivity toward branched compared to linear compounds. The structure of the Acat2-C91S mutant bound to propionyl-CoA shows that the terminal methyl group of the substrate, representing the α-branch point, is directed toward the conserved Phe 288 and Met 158 residues. In Acat5, the Phe ring is rotated to accommodate a hydroxyl−π interaction with an adjacent Thr side chain, decreasing space in the binding pocket and possibly accounting for its strong preference for linear substrates compared to Acat2. Comparison of the different Acat thiolase structures shows that Met 158 is flexible, adopting alternate conformations with the side chain rotated toward or away from a covering loop at the back of the active site. Mutagenesis of residues in the covering loop in Acat5 with the corresponding residues from Acat2 allows for highly increased accommodation of branched substrates, whereas the converse mutations do not significantly affect Acat2 substrate selectivity. Our results suggest an important contribution of second-shell residues to thiolase substrate selectivity and offer insights into engineering this enzyme class.

C

chain thiolases that synthesize or degrade 3-oxobutyryl-CoA (Thiolase II, EC:2.3.1.9) (Scheme 1A). Their structure and mechanism have been investigated extensively from the initial studies of porcine heart thiolase13−15 to those of bacterial16−24 and human25−28 enzymes. In the biosynthetic direction, C−C bond formation has been determined to require two catalytic Cys residues16,18,19 and two oxyanion holes (OAHs)21,22,26 that participate in a two-step process involving acyl−enzyme formation with the first acyl-CoA substrate followed by Claisen condensation with the second acyl-CoA substrate (Scheme 1B). In the degradative direction of thiolysis, these steps occur in reverse.17,19 The complexity of the reaction requires coordinated motions of two substrates during the catalytic cycle22 as well as the stabilization of different types of oxyanions within a single OAH.23 Understanding selectivity with respect to both acyl-CoA substrates thus poses a challenging problem. Studies of different subclasses of thiolases have yielded insight into factors that impact substrate selectivity or promiscuity.20,22,26−29 One important feature is their ability to

arbon−carbon (C−C) bonds form the foundation of organic molecules. Their formation and degradation is thus fundamental to understanding the structural diversity observed in metabolites in nature. Thiolases are a prominent member of a larger enzyme superfamily, which broadly catalyzes the reversible formation of C−C bonds through Claisen and aldol condensation and supports key metabolic functions in the biosynthesis of fatty acids, amino acids, steroids, isoprenoids, polyketides, and carbon-storage polyesters.1−4 Given the wide array of possible substrates accessible, thiolases have also become important tools for the design of engineered metabolic pathways for the production of advanced fuels,5−8 polymers,9 and chemicals.10−12 Despite the large number of sequences available, the relatively small scope of characterized thiolase substrate selectivities remains limiting for synthetic applications. Therefore, efforts to increase our insight into the structural and biochemical basis of substrate selectivity in thiolases would help further expand their application in the enzymatic and microbial production of small-molecule targets. The thiolase class specifically catalyzes the Claisen condensation between two coenzyme A (CoA) thioester substrates with members falling roughly into two groups. The first group consists of thiolases with broad activity over a range of 3-oxoacyl-CoA carbon chain lengths (Thiolase I, EC:2.3.1.16), whereas the second group comprises short© XXXX American Chemical Society

Special Issue: Current Topics in Mechanistic Enzymology Received: November 6, 2017 Revised: December 22, 2017

A

DOI: 10.1021/acs.biochem.7b01123 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry Scheme 1. Reversible Claisen Condensation Catalyzed by Thiolase Enzymesa

a

(A) Biosynthetic thiolases operate in the direction of C−C bond formation, whereas degradative thiolases carry out the reverse reaction of thiolysis. (B) In the first stage of catalysis, binding of the first acyl-CoA substrate initiates formation of an acyl−enzyme intermediate with a conserved Cys residue with tetrahedral intermediate stabilization occurring at OAH1. The acyl−enzyme adduct then rotates to shift the carbonyl from OAH1 to OAH2 for the subsequent Claisen condensation upon binding of the second acyl-CoA substrate.22 In the second stage of catalysis, Cys 376 deprotonates the acyl-CoA, forming an enolate stabilized by OAH1, which then attacks the acyl-Cys 91 thioester to form a tetrahedral intermediate stabilized by OAH2. This resolves to release Cys 91 and form the 3-oxoacyl-CoA product. Residue numbers are for Acat2. Mechanism of thiolysis follows the same steps in reverse.

Oligonucleotides and gBlocks were purchased from Integrated DNA Technologies (Coralville, IA). DNA purification kits, NiNTA agarose, and NeXtal DWBlock crystallography suites were purchased from Qiagen (Valencia, CA). PD-10 desalting columns were purchased from GE Healthcare (Pittsburgh, PA). Amicon Ultra centrifugal concentrators were purchased from Merck Millipore (Cork, Ireland). LB broth Miller, LB agar Miller, 2× YT broth, and glycerol were purchased from EMD Biosciences (Darmstadt, Germany). Tris(2-carboxyethyl)phosphine (TCEP) was purchased from Biosynth, Inc. (Itasca, IL). Imidazole was purchased from Acros Organics (Morris Plains, NJ). Isopropyl β-D-1-thiogalactopyranoside (IPTG) was purchased from Santa Cruz Biotechnology (Dallas, TX). Carbenicillin sodium salt (Cb), sodium chloride, potassium chloride, potassium phosphate monobasic, potassium phosphate dibasic, 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), and benzonase nuclease were purchased from Fisher Scientific (Pittsburgh, PA). Coenzyme A sodium salt hydrate (CoA), acetoacetyl-CoA, β-mercaptoethanol (BME), and lysozyme from chicken egg white were purchased from Sigma-Aldrich (St. Louis, MO). cOmplete EDTA-free protease inhibitor was purchased from Roche Applied Science (Penzberg, Germany). TEV protease was purchased from the QB3MacroLab at UC Berkeley. Reservoir strips and ViewDrop II covers were purchased from TTP Labtech (Melbourn, United Kingdom). 2-Methyl-2,4-pentanediol (MDP), Paratone N, and Intelli-plate 96 sitting drop

accept substrate and product branching, which greatly increases the structural diversity of small molecule targets beyond substituents to the terminal position. However, it can often be difficult to directly compare phylogenetically diverse members of a superfamily, as sequence relationships are often governed by phylogeny rather than substrate selectivity. In this context, we have recently reported the identification of a set of thiolases (Acat1−5) involved in the fermentation of α-branched organic acids in the roundworm Ascaris suum.12 These thiolases exhibited marked differences in selectivity toward the formation and degradation of linear compared to branched products despite their sequence similarity.12 In this work, we report the free (wild type) and substrate-bound (C91S mutant) crystal structures of Acat2, an α-branch permissive thiolase, along with the structure of Acat5, a linear-selective thiolase. Comparison of these structures highlighted the potential role of a “covering” loop at the back of the active site in α-branch permissiveness. Biochemical characterization of loop mutants indicates that the covering loop serves as a strong determinant of linear vs branched substrate preferences and furthers our understanding of substrate selectivity in thiolase enzymes.

■

MATERIALS AND METHODS

Commercial Materials. Restriction enzymes, Q5 DNA polymerase, and enzymes for Gibson assembly were purchased from New England BioLabs (Ipswich, MA). Deoxynucleotides (dNTPs) were purchased from Invitrogen (Carlsbad, CA). B


Article

Biochemistry

Preparation of 3-Oxoacyl-CoA Substrates. Synthetic substrates were prepared and quantified as previously reported.12 Crystallization and Structure Determination. Crystals of Acat2 were obtained by sitting drop vapor diffusion by combining 200 nL each of protein solution (10 mg/mL Acat2, 5 mM 2-methyl-3-oxopentanoyl-CoA) and reservoir solution (0.2 M tripotassium citrate, 20% w/v PEG 3350). Crystals were observed within 2 days, and a single crystal was obtained by adding 400 nL of cryogenic solution (50% v/v MDP solution containing 5 mM 2-methyl-3-oxopentanoyl-CoA) directly to the crystallization droplet, removing a single crystal within 1 min and freezing in liquid nitrogen. The electron density map of the solved structure (PDB ID 6BJ9) did not contain density for a bound substrate or for acylation of Cys 91. Crystals of Acat5 were obtained by hanging drop vapor diffusion by combining 800 nL each of protein solution (12 mg/mL Acat5, 2 mM 2-oxobutyryl-CoA) and reservoir solution (2 M sodium malonate, pH 6.5). Crystals were observed within 24 h, and a single crystal was obtained by adding Paratone N oil to the droplet, transferring a single crystal to the oil layer, and removing residual water around the crystal before freezing in liquid nitrogen. The electron density map of the solved structure (PDB ID 6BJA) contained density in the binding pocket and at the Cys 92 residue that was best modeled by free CoA and oxidized Cys 92. To obtain crystals of Acat2-C91S, protein was incubated with 2-methyl-3-oxopentanoyl-CoA in attempt to acylate Ser 91. Molecular weight determination by LC-TOF suggested conversion to the acylated form before the protein was desalted with a PD-10 column to remove the acyl-CoA and concentrated to 10 mg/mL for crystallography. Crystals were obtained by hanging drop vapor diffusion by combining 200 nL of protein solution (10 mg/mL Acat2-C91S, 4 mM propionylCoA) and reservoir solution (0.2 M ammonium nitrate, 20% w/v PEG 3350). Crystals were observed within 2 days, and a single crystal was transferred to a droplet of 1:1 reservoir solution and cryogenic solution (50% v/v MDP solution with 2 mM propionyl-CoA) for less than 1 min before freezing in liquid nitrogen. The electron density map of the solved structure (PDB ID 6BJB) contained density for propionyl-CoA and density near Ser 91, but the latter density was disconnected from Ser 91 and best modeled as a nitrate ion. Data were collected at Beamline 8.3.1 at the Advanced Light Source (Lawrence Berkeley National Laboratory). Data were processed with XDS32 and scaled and merged with Aimless.33 Molecular replacement using Phaser34 was used for phase determination. For Acat2 and Acat5, the molecular replacement model was one monomer of T2 thiolase (PDB 2IB8) with side chains pruned using Sculptor35 based on sequence alignment with Acat2 or Acat5. The phase_and_build utility of the Phenix software package36 was used for phase improvement and chain building, and any missing parts of the model were manually built using Coot.37 For Acat2-C91S, the solved structure of wild-type Acat2 was used for molecular replacement, and the resulting model was used directly for subsequent refinement with Phenix.refine. LigandFit38 was used to model CoA and propionyl-CoA ligands when sufficient electron density was observed to create a model. Potassium ions and nitrate and acetate ligands were modeled with Coot, where anomalous unmodeled density was observed. Iterative cycles using Phenix.refine and manual refinement in Coot were used to generate the final model, with Phenix.refine strategies based on

reservoir plates were purchased from Hampton Research (Aliso Viejo, CA). Bacterial Strains. E. coli DH10B-T1R was used for DNA construction, and BL21(de3)-T1R was used for heterologous production of proteins for purification. Site-Directed Mutagenesis. PCR amplifications were carried out with Q5 DNA polymerase, following manufacturer instructions. Acat2-C91S was constructed by amplifying Acat2 gene in two segments, using primers AsAcat2 23a F1 and Acat2-C91S RP for the first segment and primers Acat2-C91S FP and AsAcat2 23a R1 for the second segment (Table S1). The two segments were inserted into pET23a digested with SfoI and BamHI, using the Gibson protocol.30 Other Acat2 and Acat5 mutant constructs were assembled using a similar strategy. Plasmids were verified by sequencing (Quintara Biosciences; Berkeley, CA). Expression and Purification of Acat2 and Acat5 Variants. Acat2 (UniProt F1KYX0) and Acat5 (UniProt F1L3N8) used for X-ray crystallography were expressed and purified as described previously.12 The Acat2-C91S mutant was purified using the same protocol. His6-tagged thiolases used for enzyme kinetics were expressed and purified as follows: 2× YT broth (1 L) containing Cb (50 μg/mL) in a 2.5 L Ultra Yield flask (Thomas Instrument Company, Oceanside, CA) was inoculated to OD600 = 0.05 with an overnight culture of freshly transformed E. coli containing the overexpression plasmid with the variant of interest. The cultures were grown at 37 °C at 250 rpm to OD600 = 0.7 to 1.0, at which point cultures were cooled on ice for 15 min, followed by induction of protein expression with IPTG (0.4 mM) and overnight growth at 16 °C. Cell pellets were harvested by centrifugation at 6000g for 7 min and then resuspended in 5 mL per gram of wet cell pellet with buffer A (50 mM potassium phosphate, 300 mM NaCl, 5% v/v glycerol, 10 mM imidazole, pH 8.0) supplemented with 10 mM BME, 1 mg/mL lysozyme, and 0.04 μL/mL benzonase at 4 °C. The resuspended pellet was stored at −80 °C until purification. Frozen cell pellets were thawed in tepid water. cOmplete EDTA-free protease inhibitor was added, and the cell suspension was homogenized by 10 passes with a glass-Teflon homogenizer and lysed by passage through a French pressure cell at 14 000 psi. An additional 0.04 μL/mL benzonase was added, and the lysate was cleared by centrifugation for 30 min at 14 000g to remove insoluble debris. The supernatant was passed over Ni-NTA resin pre-equilibrated with buffer A and washed with buffer A containing 5 mM BME, followed by buffer B (50 mM potassium phosphate, 300 mM NaCl, 5% v/v glycerol, 34 mM imidazole, pH 8.0) containing 5 mM BME. Protein was eluted with buffer C (50 mM potassium phosphate, 300 mM NaCl, 5% v/v glycerol, 250 mM imidazole, pH 8.0), and fractions containing protein were pooled and concentrated using an Amicon Ultra spin concentrator (10 kDa MWCO), diluted with buffer D (30 mM HEPES, 150 mM NaCl, 15% v/v glycerol, pH 6.8, 0.4 mM TCEP), and concentrated again. Aliquots were frozen in liquid nitrogen and stored at −80 °C. Final protein concentrations were estimated using the ε280nm calculated by ExPASY ProtParam31 as follows: His6-Acat2 (ε280nm = 22 920 M−1 cm−1), His6-Acat5 (ε280nm = 25 900 M−1 cm−1), His6-Acat2-Loop5 (ε280nm = 22 920 M−1 cm−1), His6Acat2-Y153I (ε280nm = 21 430 M−1 cm−1), His6-Acat5-Loop2 (ε280nm = 25 900 M−1 cm−1), and His6-Acat5-I154Y (ε280nm = 27 390 M−1 cm−1). C


Article

Biochemistry

Figure 1. Sequence alignment of Acat2, Acat5, and human T2 thiolases. Catalytic, unmodeled, and other emphasized residues are indicated based on the key. Residues involved in polar interactions at the dimer interface are also highlighted. The residue numbering systems are not identical for the two isozymes due to the presence of gaps in the Acat2 sequence when aligned with Acat5 and differences in sizes of their N-terminal mitochondrial signal peptide. We thus adopt the convention of listing the Acat2 residue first when referring to conserved or corresponding residues in both enzymes.

XYZ coordinates, real-space refinement, atomic occupancies, and individual B-factors. Since the resolutions exceeded 1.6 Å for all three structures, the ReadySet utility was used to place riding hydrogen atoms in a late stage of refinement to improve the model. Chimera39 was used for visualization of final structures. Kinetic Measurements. Thiolysis activity was measured by monitoring the initial formation of new thioester functional groups at 232 nm. Assays were performed at 25 °C in a UVcompatible 96-well plate in a total volume of 200 μL containing 50 mM potassium phosphate, pH 8.0, coenzyme A (100 μM), thiolase enzyme, and 10, 20, 40, 80, 160, 320, or 600 μM 3oxoacyl-CoA. The reaction was initiated by addition of enzyme and mixing rapidly by pipet. Assays with Acat5 also included 100 mM KCl. Data were also collected for previously characterized wt enzyme/substrate pairs12 in order to validate the 96-well plate assay format and make a direct comparison to data collected for mutants under these conditions. kcat and KM were determined by fitting to initial rate data with Origin (OriginLab, Northampton, MA) using the equation v0 =

where v0 is the initial rate and [S] is the substrate concentration, or using the substrate inhibition equation ν0 =

kcat[S] KM + [S] +

[S]2 KI,substrate

if the calculated KI,substrate was less than 100 times the calculated KM.

■

RESULTS Sequence Alignments. Phylogenetic analysis of thiolases indicates that they consist of six clusters, with related thiolases in eukaryotes sharing similar subcellular localizations and functions.40 Prokaryotic thiolases cluster among their eukaryotic relatives rather than forming a separate clade,40 so both groups serve to provide useful information for considering thiolase function and selectivity more broadly. In particular, the human mitochondrial thiolase T2 participates in the degradation of the branched amino acid isoleucine and has been shown to be permissive toward the α-branched substrate, 2-methyl-3oxobutyryl-CoA.27 Sequence alignments and docking studies suggested that a conserved Phe residue is a key determinant for this behavior based on comparison to other human thiolases that prefer linear substrates (mitochondrial thiolase T128 and

kcat[S] KM + [S] D


Article

Biochemistry Table 1. Data Collection and Refinement Statistics for Acat2, Acat5, and Acat2-C91S Acat2 (PDB entry 6BJ9) Data Collection space group cell dimensions a, b, c (Å) wavelength (Å) resolution (Å)a Rmerge (%)a completeness (%)a ⟨I/σI⟩a no. of unique reflectionsa redundancya B-factor from Wilson plot (Å2) Refinement Statistics resolution (Å) no. of reflections in the working set in the test set Rwork/Rfree no. of non-hydrogen atoms protein ligands and ions water rmsd bond lengths (Å) bond angles (deg) average B-factor protein ligands and ions water Ramachandran plot favored (%) outliers (%) a

Acat5 (PDB entry 6BJA)

Acat2-C91S (PDB entry 6BJB)

P22121

I222

P22121

71.57, 88.12, 142.26 1.1159 75−1.53 (1.585−1.53) 6.7 (38.7) 99.6 (96.2) 11.50 (2.49) 135 412 (12 914) 4.0 (3.0) 12.27

71.66, 111.78, 112.76 1.1159 79−1.60 (1.657−1.60) 8.0 (28.6) 99.6 (96.4) 15.67 (3.08) 59 736 (5698) 10.1 (4.2) 17.55

71.89, 88.62, 141.51 1.1159 75−1.50 (1.554−1.50) 7.9 (42.6) 99.9 (100) 9.03 (2.60) 144 700 (14 328) 4.0 (4.0) 12.10

75−1.53

79−1.60

75−1.50

128 707 6703 0.139/0.160 6526 5656 2 868

56 623 3071 0.146/0.173 3269 2842 52 375

137 598 7086 0.149/0.173 6641 5642 120 879

0.009 1.29 16.10 14.40 10.20 27.70

0.010 1.33 24.40 22.70 54.80 33.10

0.010 1.37 16.90 15.10 18.60 28.20

98 0

98 0.53

99 0

Values in parentheses are for the highest-resolution shell.

82°) and Val 91 (φ = 45°; ψ = −129°), which precedes the catalytic Cys 92. These residues were well-defined in the electron density map, and both have been found with unfavorable φ and ψ values in other reported thiolase structures as well.22,26 The Acat2 and Acat5 monomers show high similarity when superimposed (Cα average rmsd = 0.92 Å). Due to a small difference in the angle of the two subunits with respect to the tetramer center (56.9° for Acat2 vs 58.8° for Acat5) and in the orientation of opposing dimer pairs with respect to the long axis (42.1° for Acat2 vs 40.7° for Acat5), the rmsd for the second subunit within the same dimer increases to 1.89 Å when the tetramer is aligned with a single monomer. The greatest differences between single subunits is found in a motif consisting of the Nα3 helix, the small loop leading to Cβ1, a highly sequence-variable active-site “covering” loop, and the Lα2 helix (Figure 2B). The Nα3 helix is at the end of the complex and is distal to the substrate-binding pocket and intersubunit contacts. However, the covering loop and Lα2, across from the pantetheine loop and catalytic residues, form a part of the substrate-binding cavity and shield the active site from solvent. In contrast to the opposing dimer pairs, the lateral interface between subunits of a dimer contain a number of polar interactions (Table S2). Overall, Acat5 has more polar interactions between dimer subunits than Acat2 and many

cytosolic thiolase CT26). In aligning the sequences of Acat2 and Acat5 with T2, we find that both Acat2 and Acat5 contain the conserved Phe (Acat2, Phe 288; Acat5, Phe 292) despite the large differences in their preferences for linear compared to branched substrates (Figure 1). We thus set out to gain insight into the structural factors that may influence α-branch selectivity in Acat2 and Acat5 given their close sequence relationship. Overall Architecture of the Acat2 and Acat5 Crystal Structures. Crystal structures for Acat2 (PDB ID 6BJ9) and Acat5 (PDB ID 6BJA) were solved by molecular replacement to ≤1.6 Å using a modified search model of human T2 thiolase (PDB ID 2IB8) (Table 1). Acat2 contained two monomers per asymmetric unit, which were derived from adjacent tetramer assemblies in the crystal, whereas only a single monomer was contained in the asymmetric unit of Acat5. These models show the expected dimer of dimers architecture and are also consistent with their estimated oligomeric state in solution.12 Both Acat2 and Acat5 contain characteristic homotetramer assembly loops, as observed among thiolases of T1, CT, and T2 clusters, and higher-order models could be constructed from copies of the adjacent unit cells (Figure 2A). As in most thiolases, the opposing dimer pairs of Acat2 and Acat5 contain few contacts other than those within the tetramerization loops. The Acat5 structure also contained two residues with unfavorable Ramachandran statistics, Asn 68 (φ = 31°; ψ = E


Article

Biochemistry

Figure 2. Comparison of unliganded structures of Acat2 and Acat5. (A) Tetramer models of Acat2 (dark slate blue) and Acat5 (gray), aligned by the upper left subunit. (B) Single subunit of Acat2 (slate blue) and Acat5 (gray), with regions of highest rmsd highlighted in red for Acat2 and pink for Acat5. (C) Acat5 dimer with hydrogen bonds and ionic contacts between subunits indicated by solid lines. Black lines indicate contacts found in both Acat2 and Acat5. Red lines indicate contacts particular to Acat5. (D) Hydrogen bonding between the covering loop and opposing subunit in Acat5. Hydrogen bonds are indicated by solid red lines.

fill the density observed for the adenosine moiety (Figure 3A). We also solved the structure of a catalytically inactive mutant of Acat2 with the active-site Cys 91 mutated to serine bound to the propionyl-CoA substrate (PDB ID 6BJB). The propionylCoA ligand is well-defined by the electron density map (Figure 3B), and superposition of Acat2-C91S with wt Acat2 shows essentially no change in the position of Cα atoms, with rmsd < 0.25 Å. High B-factors in the central diphosphate and pantoic acid sections of CoA in Acat5 indicated low occupancy and multiple orientations of CoA within the crystal lattice (Figure 3C), but the adenine moiety is well-defined in both propionylCoA and CoA ligands, consistent with the observation that T2type thiolases have additional features that provide tight binding of the ring,27 including hydrogen bonds with Asn 224 (Ile 226 in Acat5), the hydroxyl of Tyr 184/185, and a tightly coordinated water (Figures 3D,E and S1). The CoA portion of the propionyl-CoA substrate in Acat2C91S aligns extremely well with the free CoA bound to T2 (PDB ID 2IBW) (Figure 3E).27 The mode of binding is substantially similar, with Ser 247/251 providing direct hydrogen bonds with the N1 and N2 atoms of the pantetheine arm and all other polar interactions with pantetheine mediated by structured water molecules also observed in the apo

other thiolases, with an additional five hydrogen bonds per monomer. These additional interactions complete a collection of hydrogen bonds that line the entire upper half of the lateral interface (Figure 2C). Interestingly, it has been suggested that increased rigidity of a canonical acetoacetyl-CoA thiolase from intersubunit hydrogen bonds may contribute to both its high efficiency and more limited substrate scope.22 Of particular interest is a hydrogen bond in Acat5 between OG(Ser 153) of the covering loop and N(Gln 67) of the opposing subunit (Figure 2D). Immediately following Gln 67 is Asn 68, one of the Ramachandran disfavored residues, which bends upward to form two hydrogen bonds with Cys 85 and Thr 87 of the Nβ3 strand of the same monomer (Figure 2D). Although the hydrogen bond between NE2(Gln 67) and O(Gly 149) of Lα1 is conserved, the additional bond with Ser 153 and the interactions with Asn 68 are unique to Acat5 and could help draw the opposing subunit even closer to the covering loop and active-site cavity. Mode of CoA Binding. The wt Acat2 and Acat5 proteins were crystallized in the presence of 2-methyl-3-oxovaleryl-CoA and 3-oxobutyryl-CoA, respectively. However, the wt Acat2 structure did not contain electron density in the substratebinding pocket, and Acat5 was best modeled with free CoA to F


Article

Biochemistry

Figure 3. Ligand-binding modes in Acat2 and Acat5. (A) Acat5 (gray) bound to CoA. The mFobs − DFcalc omit map (green) for the ligand is contoured at 3.5σ. Electron density for CoA in Acat5 supported occupancy primarily for the adenine and ribose moieties. (B) Acat2-C91S (blue) bound to propionyl-CoA and nitrate ion. The mFobs − DFcalc omit map (green) for the ligands is contoured at 4σ. (C) Alignment of propionyl-CoA from Acat2-C91S and CoA from Acat5 colored by B-factor. (D) Diagram of interactions between Acat2-C91S and propionyl-CoA. Dashed lines indicate hydrogen bond contacts. (E) Overlay of the Acat2-C91S/propionyl-CoA crystal structure (blue) with the CoA ligand (green) from T2 (PDB ID 2IBW) (green). Dashed lines show hydrogen bond contacts. Residues for Acat2 are labeled. The K+ ion in T2 is replaced with water in Acat2-C91S, as K+ was not present in the crystallization solution.

Figure 4. Structural analysis of propionyl-CoA binding in Acat2-C91S. (A) Binding of propionyl-CoA (blue) in the catalytic active site of Acat2C91S. The thioester carbonyl of propionyl-CoA occupies OAH1 formed by His 348 and coordinated water. Nitrate ion (red) occupies OAH2 formed by amide nitrogens of Ser 91 and Gly 378. (B) Side view of Acat2-C91S active site. Red lines indicate geometry of carbonyl and α-carbons of propionyl-CoA (blue) with respect to Ser 91 and Cys 376. The geometry between Ser 91 and propionyl-CoA is unsuitable for nucleophilic attack of the thioester, but Cys 376 is suitably located for α-deprotonation. Conserved residues of the long chain acyl-CoA binding pocket28 are highlighted in orange.

structure but is turned inward in Acat2-C91S, Acat5, and T227 to form polar and ionic interactions with the ribose 3′-

structure (Figure 3D,E). The major exception is Lys 227/229, which rotates out toward the solvent in the unbound Acat2 G


Article

Biochemistry

CoA-bound Acat2-C91S structure therefore seems to be most consistent with the intermediate preceding Claisen condensation, with propionyl-CoA positioned as the nucleophile and the nitrate ion occupying OAH2 acting as a substitute for acyl− enzyme (Figure S2). Furthermore, the resulting 2S stereochemistry for the α-methyl substituent of the product in this case would be consistent with the stereochemistry expected for the product of the ketoreductase from A. suum that carries out the downstream reaction to form the 2-methyl-3-oxoacylCoA.12 Substrate Selectivity. Like the CT and T2 short-chain thiolases, Acat2 and Acat5 work poorly with acyl-CoA substrates of four or more carbons (to give products with main chain length of six or more). On the basis of the longchain T1 thiolase, the acyl chain for the initial substrate is thought to point downward toward Met 90/Val91, Gln 66/67, and the end of Lα1 and continue on to a binding pocket formed by Ile 350/Leu 354 and Val 59/60 before terminating at Glu 119/120 (Figure 4B).28 However, this cavity is blocked off in Acat2 and 5, primarily by the increased length of the Lα1 helix from an additional Asp residue, found in CT and T2 thiolases but not in T1. The different Lα1 helix lengths are also thought to play a role in the selectivity of BktB, a T1-like bacterial enzyme.29 On the basis of the sequence variability of the Lα1 region, the size of the Lα1 helix is difficult to determine by sequence alignment and appears to be the same length in T1, T2, and CT-like thiolases. The ability to carry out structure-guided alignment thus helps to clarify its role in longchain permissiveness (Figure S3). The terminal methyl group of propionyl-CoA in the bound Acat2-C91S structure fits precisely between the conserved residues Phe 288/292 and Met 158/159 (Figure 5A). This is consistent with computational docking studies that predict the 2S-methyl-3-oxobutyryl-CoA substrate would bind with the αmethyl branch pointed toward the Phe residue ring face in T2.27 Interestingly, mutation of the corresponding Met to Ala

phosphate and 2′- and 3′-hydroxyl groups. Hydrophobic residues including Leu 149/150, Phe 235/237, Leu 249/Ile 253, Phe 319/323, and Ile 350/Leu 354 line the pantetheine binding tunnel, as observed in other thiolases. A few differences between Acat2 and T2 are observed at the interface of the pantetheine arm and the top of the binding cavity. For example, the O3 atom of propionyl-CoA in Acat2C91S forms a water-mediated hydrogen bond with the amide nitrogen of Gly 159 (Figure 3D,E), whereas it has no notable interactions in T2. Furthermore, the O2 atom in Acat2-C91S is connected to this network through an additional water, whereas in T2 this atom forms water-mediated hydrogen bonds to His 157 and to a Tyr residue in the tetramerization loop of the opposing dimer.27 In Acat2 and Acat5, the His is replaced by Leu 157/Pro 158 and the Tyr by Phe 135/136, adding even greater hydrophobicity to the substrate-binding cavity. Active Site Structure. Catalytic features are highly conserved in the superfamily3 and include the catalytic Cys (Cys 91/92) that forms an acyl−enzyme covalent intermediate with the thioester substrate,17 as well as two oxyanion holes (OAH1 and OAH2) that stabilize various enolate or tetrahedral intermediates formed in the catalytic cycle.23 The catalytic Cys (Cys 91/92) is found at the N-terminal end of the Nα3 helical dipole and is proposed to be deprotonated by an adjacent His residue (His 348/352) upon substrate binding.22,23,26 OAH1 is formed by conserved His (His 348/352) or Asn (Asn 316/320) residues or coordinated waters at the Cβ2−Cα2 and Cβ3−Cα3 hairpins.22,23 OAH2 is formed by the backbone amide nitrogens of both the catalytic Cys (Cys 91/92) and Gly 378/382 at the Nβ4−Nβ5 hairpin (Figure 4A).22 In the crystal structures of Acat2 and Acat5, the positions and oxidation states of Cys 91/ 92 (Ser 91 in Acat2-C91S) are analogous to what was reported for human T2.27 Specifically, a slight rotation of Cys 91 (Ser 91) toward His 348 is seen between apo Acat2 compared to either liganded Acat2-C91S or Acat5. In Acat5, Cys 92 was best modeled as an oxidized cysteine sulfoxide, with the oxygen seated in OAH2. The propionyl-CoA bound structure of Acat2-C91S shows that the carbonyl group of the substrate is found in OAH1 formed by His 348 and a water that is hydrogen bonded to Asn 319 (Figure 4A). OAH1 is involved both in the initial formation of the acyl−enzyme intermediate as well as in enolate stabilization in the Claisen condensation step. As such, the positioning of propionyl-CoA in the structure could be consistent with either intermediate. In the first case, propionylCoA would be the electrophile and the terminal methyl group would contribute to extending the linear chain length of the 3oxoacyl-CoA product, but in the second case, propionyl-CoA would be the nucleophile, and the terminal methyl group installs an α-methyl substituent in the product (Figure S2). The positioning of the second catalytic Cys (Cys 376) that serves as a general base to form the enolate nucleophile18,19,21 appears to be consistent with the second possibility (Figure 4B). In the Acat2-C91S structure, Cys 376 is well positioned to abstract the pro-S α-proton of propionyl-CoA. Furthermore, the plane of the resulting enolate is directly above a nitrate ion bound in OAH2, suggesting that it is poised for nucleophilic attack upon an electrophile. In contrast, the geometry of Ser 91 with respect to the plane of the propionyl-CoA thioester is not suitable for nucleophilic attack to form an acyl−enzyme intermediate since the acyl-CoA would be expected to be rotated so that the plane is perpendicular to OG(Ser 91) (Figure 4B). The propionyl-

Figure 5. Structural features affecting substrate selectivity in Acat2 and Acat5. An overlay of Acat2-C91S (blue) and Acat5 (gray) is shown. The terminal methyl group of the propionyl-CoA ligand from the Acat2-C91S structure is circled in red. Residues differing between Acat2 and Acat5 proposed to affect branch chain accommodation are labeled and colored according to protein. Residues conserved in the active site are shown with Acat2 numbering and colored in red. (A) The terminal methyl of propionyl-CoA in Acat2-C91S is directed toward Met 158/159 and Phe 288/292. Acat5-F292 is rotated toward Thr 293. Met 158 undergoes a conformation change upon binding. The Met 158 side chain of apo-Acat2 is also shown in this overlay in dark slate blue. (B) View of the binding pocket from above Lα2 and the covering loop. Tyr 153 is in close proximity to Nα2. H


Article

Biochemistry

Figure 6. Role of the covering loop in thiolase substrate selectivity. (A) Alignment of eukaryotic and prokaryotic thiolases. Residues of the covering loop targeted for mutagenesis are highlighted in gray. Tyr 153, Met 158, and Phe 288 are shaded in black, and conserved residues in these positions are highlighted in yellow. Human T1 and CT prefer linear substrates, whereas human T2 accept branched substrates. (B) Substrates used in biochemical characterization of thiolases. (C) Steady-state kinetic characterization of Acat2 and Acat5. Data are the mean ± standard error (n = 3) as derived from nonlinear curve fitting to the Michaelis−Menten equation. Data were fit to the substrate inhibition equation when indicated with an entry for KI,substrate. Error in the relative kcat/KM was obtained from propagation. Lower bounds of kcat and KM are given for substrates that did not reach saturation.

stages of catalysis22 also contain the corresponding Met turned up toward the covering loop, as in the bound Acat2 and T2 states. Taken together, the direct contact of the conserved Met (Met 158/159) with the α-substituent and its fluxionality in different structures suggests that it may be important in determining the selectivity of thiolases toward substituents in the α-position. Furthermore, we hypothesize that secondsphere interactions between Met and the sequence-variable active-site covering loop may give rise to differences in activesite structure and flexibility that affect the accommodation of different substrates. Biochemical Studies of the Covering Loop. In order to better identify residues that participate in governing substrate selectivity, we examined a sequence alignment of characterized linear and branched chain-selective thiolases (Figure 6A). On the basis of the substrate-bound crystal structure of Acat2C91S, the closest residues to Met 158 are Asp 151 of the covering loop and Gln 66 of the neighboring subunit (Figure 5B). However, both are highly conserved and may thus serve an important functional role. In contrast, Tyr 153 (Acat2 numbering) is a residue shared by several thiolases permissive of α-methyl substrates that is less common in thiolases selective for linear substrates. Tyr is a bulky residue in close proximity to the Nα2 helix of the opposite subunit of the dimer, but it is not positioned to form any favorable interactions with this helix.

in BktB (M290A) was reported to confer the ability to act on an α-methyl substrate.29 No change in the Phe 288 side chain is observed between the apo-Acat2 and holo-Acat2-C91S structures, but in Acat5 the ring is rotated by 47° toward the substrate. This rotation appears to decrease the space available for an α-substituent and is related to the presence of a bulkier residue (Thr 293) in Acat5 compared to branched-chain permissive thiolases (Acat2, Acat3, T2) where it is replaced with Gly or Pro. The hydroxyl group of Thr 293 is directed toward the ring face of Phe 292 and also possibly forms a weak hydrogen bond interaction (3.5 Å) with the π-system of the ring. In comparing the structures of apo-Acat2 and holo-Acat2C91S, Met 158 undergoes a conformational shift upon binding of propionyl-CoA, where it turns up toward the back of the binding pocket formed by the active-site covering loop (Figure 5). In both structures, the position of Met 158 is well-defined with no evidence of secondary conformations based on the electron density (average B-factors: apo-Acat2, 12.4 Å2; holoAcat2-C91S, 10.2 Å2). While this movement was also observed in T2 upon binding of free CoA, both positions of this Met residue (Met 193) are observed in the apo T2 structure.27 This fluxionality is found with Met 159 in Acat5 as well (atomic Bfactors ranging from 21 to 36 Å2 moving from the CB to CE atoms). Structures of CT- and T1-type thiolases at different I


Biochemistry

Article

■

DISCUSSION In this work, we report structural and biochemical studies on two thiolase enzymes, Acat2 and Acat5, from A. suum, which demonstrate differing substrate selectivity toward branched and linear compounds. This work provides new information about the origin of thiolase selectivity and helps expand our knowledgebase for a key family of enzymes involved in both physiological and biotechnological processes. On the basis of several observations, the propionyl-CoA bound structure of Acat2-C91S, which demonstrates branched substrate accommodation, is hypothesized to represent an important intermediate consistent with the substrate positioned with the terminal methyl group at the α-branch site. Analysis of this structure shows that the methyl branch lies between Met 158/ 159 and Phe 288/292, near the active-site covering loop. Interestingly, this conserved Met demonstrated fluxionality in response to substrate binding. Furthermore, rotation of the Phe 292 ring facilitated by interaction with a bulky residue appears to reduce the size of the binding pocket in Acat5, which strongly prefers linear substrates. Taken together, both of these residues appeared to influence the geometry of the binding pocket, potentially through second-sphere interactions with the adjacent covering loop. Swapping loops or motifs between related proteins has been used successfully as a strategy for peptide or enzyme engineering.41,42 In our mutagenesis and biochemical studies, we found that exchange of the Acat5 covering loop with that of Acat2 was sufficient to increase the relative permissiveness of Acat5 with 2-methyl-3-oxobutyryl-CoA by 40-fold, reaching similar selectivity displayed by Acat2. In particular, a single swap of Ile 154 for Tyr appears to account for a large part of this selectivity change (11-fold). Interestingly, Tyr is frequently excluded at this position from characterized thiolases that do not permit a branched substrate (Figure 6A). Aside from Tyr 153/Ile 154, another important residue in the covering loop might be Ser 153, whose presence in Acat5 allows an additional hydrogen bond interaction at the dimer interface. In contrast to the mutagenesis results with Acat5, the converse mutagenesis of the Acat2 covering loop with that of Acat5 did not have any substantive effect on Acat2 substrate selectivity. Additional factors that may account for Acat5 branched selectivity are the presence of Thr 293 and Asn 68. Thr 293 may induce the rotation of Phe 292 in that enzyme, whereas Asn 68, which was found to be a Ramachandran outlier, may facilitate hydrogen bonding near the covering loop and dimer interface. The subtle factors that affect differences in substrate selectivity in thiolases are complex, especially since these enzymes display promiscuity similar to other acyl-CoA utilizing enzyme families. While contributions from the first sphere of the active site have been previously explored,27−29 these studies offer new insight into the role of second-sphere interactions with the covering loop. Given the sequence variability found in this loop, we hope that these findings can help guide the discovery and engineering of thiolase substrate selectivity.

Conversely, in Acat5 this residue is Ile 154, a smaller amino acid that could also more easily engage in hydrophobic interactions with the methylene groups of Arg 71 in the Nα2 helix. Since the position of Nα2 is not substantially different between Acat2 and Acat5 (Figures 2D and 5B), we hypothesized that the covering loop may shift depending on the presence of bulky Tyr compared to smaller Ile, thereby adjusting the volume of the active site pocket. This change in binding pocket size could then tune the permissiveness of the enzyme toward α-methyl substrates. To test this hypothesis, we cloned and purified the Acat2-Y153I and Acat5-I154Y mutants (Figure S4) in which these residues were switched and measured kinetic parameters with 3-oxobutyryl-CoA, 2methyl-3-oxobutyry-CoA, and 2-methyl-3-oxopentanonoylCoA substrates (Figure 6BC, Figure S5). We also characterized the Acat2-Loop5 and Acat5-Loop2 mutants in which the six covering loop residues located between the conserved Asp 151/ 152 and Met 158/159 were exchanged between the two isozymes. Wild-type Acat2 displays relatively low preference for linear (3-oxobutyryl-CoA) vs C5-branched (2-methyl-3-oxobutyrylCoA, kcat/KM(relative) = 0.6 ± 0.1) and C6-branched (2methyl-3-oxopentanoyl-CoA, kcat/KM(relative) = 0.6 ± 0.1) substrates. This result indicates that it accommodates both branched and longer substrates without selecting against the smaller linear substrate. In comparison, wt Acat5 shows defects in catalytic efficiency of 91- and 2900-fold with the C5- and C6branched substrates, respectively. These data indicate that Acat5 demonstrates an ∼102 preference against the α-methyl branch specifically in addition to a significant selection against the longer C6 substrate. Exchanging six residues of the covering loop of Acat5 with those of Acat2 results in a relatively small drop in efficiency (6.4-fold) with the linear 3-oxobutyryl-CoA, but it results in a 40-fold increase in its acceptance of the corresponding branched substrate, 2-methyl-3-oxobutyryl-CoA (kcat/KM(relative) = 0.44 ± 0.06), bringing it close to the preference displayed by Acat2 (kcat/KM(relative) = 0.6 ± 0.1). Characterization of the single-site mutant Acat5-I154Y shows that a substantial portion (11-fold) of this 40-fold increase in relative permissiveness arises from this particular point mutation. In contrast, characterization of Acat2-Loop5 and Acat2-Y153I showed that Acat2 was relatively agnostic to these mutations in terms of their effect on the enzyme’s relative permissiveness toward α-methyl substrates. The Acat5-Loop2 mutant may also reveal additional information about chain length selectivity as one-carbon extension of the substrate main chain (2-methyl-3-oxobutyrylCoA to 2-methyl-3-oxopentanoyl-CoA) confers a greater catalytic defect (60-fold) than that observed with wt Acat5 (32-fold). This result is especially surprising given that main chain extension was found to have a much smaller effect (4-fold for wt Acat5) with the corresponding linear substrates, 3oxobutyryl-CoA and 3-oxopentanoyl-CoA.12 Since the hydrophobic tunnel for binding longer chain acyl-CoA substrates is obstructed in CT- and T2-type thiolases (Figures 4B and S3), the terminal methyl of a propionylated enzyme intermediate might be rotated up toward Met 159, Phe 292, and the covering loop, rather than down toward the tunnel. Given that a subsequent propionyl-CoA nucleophile must bind with the terminal methyl directed toward these same residues, the binding affinity for propionyl-CoA may be lower than usual in propionylated enzyme.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.7b01123. Sequences of oligonucleotides, tabulation of interdimer hydrogen and ionic bonding, additional structure J


Article

Biochemistry

(7) Shen, C. R., Lan, E. I., Dekishima, Y., Baez, A., Cho, K. M., and Liao, J. C. (2011) Driving forces enable high-titer anaerobic 1-butanol synthesis in Escherichia coli. Appl. Environ. Microbiol. 77, 2905−2915. (8) Sheppard, M. J., Kunjapur, A. M., Wenck, S. J., and Prather, K. L. J. (2014) Retro-biosynthetic screening of a modular pathway design achieves selective route for microbial synthesis of 4-methyl-pentanol. Nat. Commun. 5, 5031. (9) Zhuang, Q., Wang, Q., Liang, Q., and Qi, Q. (2014) Synthesis of polyhydroxyalkanoates from glucose that contain medium-chainlength monomers via the reversed fatty acid beta-oxidation cycle in Escherichia coli. Metab. Eng. 24, 78−86. (10) Martin, C. H., Dhamankar, H., Tseng, H. C., Sheppard, M. J., Reisch, C. R., and Prather, K. L. (2013) A platform pathway for production of 3-hydroxyacids provides a biosynthetic route to 3hydroxy-gamma-butyrolactone. Nat. Commun. 4, 1414. (11) Cheong, S., Clomburg, J. M., and Gonzalez, R. (2016) Energyand carbon-efficient synthesis of functionalized small molecules in bacteria using non-decarboxylative Claisen condensation reactions. Nat. Biotechnol. 34, 556−561. (12) Blaisse, M. R., Dong, H., Fu, B., and Chang, M. C. Y. (2017) Discovery and engineering of pathways for production of α-banched organic acids. J. Am. Chem. Soc. 139, 14526−14532. (13) Gehring, U., Riepertinger, C., and Lynen, F. (1968) Purification and crystallization of thiolase; study of its action mechanism. Eur. J. Biochem. 6, 264−280. (14) Harris, J. I., and Gehring, U. (1970) The active site cysteines of thiolase. Eur. J. Biochem. 16, 492−498. (15) Harris, J. I., and Gehring, U. (1970) The subunit structure of thiolase. Eur. J. Biochem. 16, 487−491. (16) Masamune, S., Palmer, M. A. J., Gamboni, R., Thompson, S., Davis, J. T., Williams, S. F., Peoples, O. P., Sinskey, A. J., and Walsh, C. T. (1989) Bio-claisen condensation catalyzed by thiolase from Zoogloea ramigera. Active site cysteine residues. J. Am. Chem. Soc. 111, 1879−1881. (17) Thompson, S., Mayerl, F., Peoples, O. P., Masamune, S., Sinskey, A. J., and Walsh, C. T. (1989) Mechanistic studies on.beta.ketoacyl thiolase from Zoogloea ramigera: Identification of the activesite nucleophile as Cys89, its mutation to Ser89, and kinetic and thermodynamic characterization of wild-type and mutant enzymes. Biochemistry 28, 5735−5742. (18) Palmer, M. A., Differding, E., Gamboni, R., Williams, S. F., Peoples, O. P., Walsh, C. T., Sinskey, A. J., and Masamune, S. (1991) Biosynthetic thiolase from Zoogloea ramigera. Evidence for a mechanism involving Cys-378 as the active site base. J. Biol. Chem. 266, 8369−8375. (19) Williams, S. F., Palmer, M. A., Peoples, O. P., Walsh, C. T., Sinskey, A. J., and Masamune, S. (1992) Biosynthetic thiolase from Zoogloea ramigera. Mutagenesis of the putative active-site base Cys-378 to Ser-378 changes the partitioning of the acetyl S-enzyme intermediate. J. Biol. Chem. 267, 16041−16043. (20) Modis, Y., and Wierenga, R. K. (1999) A biosynthetic thiolase in complex with a reaction intermediate: The crystal structure provides new insights into the catalytic mechanism. Structure 7, 1279−1290. (21) Modis, Y., and Wierenga, R. K. (2000) Crystallographic analysis of the reaction pathway of Zoogloea ramigera biosynthetic thiolase. J. Mol. Biol. 297, 1171−1182. (22) Kursula, P., Ojala, J., Lambeir, A.-M., and Wierenga, R. K. (2002) The catalytic cycle of biosynthetic thiolase: A conformational journey of an acetyl group through four binding modes and two oxyanion holes. Biochemistry 41, 15543−15556. (23) Meriläinen, G., Poikela, V., Kursula, P., and Wierenga, R. K. (2009) The thiolase reaction mechanism: The importance of Asn316 and His348 for stabilizing the enolate intermediate of the Claisen condensation. Biochemistry 48, 11011−11025. (24) Kim, E.-J., and Kim, K.-J. (2014) Crystal structure and biochemical characterization of PhaA from Ralstonia eutropha, a polyhydroxyalkanoate-producing bacterium. Biochem. Biophys. Res. Commun. 452, 124−129.

illustrations and sequence alignments, SDS-PAGE, and initial rate data for Acat2 and Acat5 kinetic characterization (PDF) Accession Codes

Atomic coordinates and structure factors have been deposited in the Protein Data Bank (entries 6BJ9, 6BJA, and 6BJB).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Michelle C. Y. Chang: 0000-0003-3747-7630 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Funding

This work was funded by the Center for Sustainable Polymers, an NSF Center for Chemical Innovation (CHE-1413862). M.R.B. also acknowledges the support of an NSF Graduate Research Fellowship. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. George Meigs and Dr. James Holton for assistance with X-ray data collection. X-ray data were collected at the Advanced Light Source Beamline 8.3.1, which is operated by the University of California Office of the President, Multicampus Research Programs and Initiatives (MR-15328599), the National Institutes of Health (R01 GM124149 and P30 GM124169), Plexxikon Inc., and the Integrated Diffraction Analysis Technologies program of the U.S. Department of Energy Office of Biological and Environmental Research. The Advanced Light Source is a national user facility operated by Lawrence Berkeley National Laboratory on behalf of the U.S. Department of Energy under contract number DEAC02-05CH11231, Office of Basic Energy Sciences.

■ ■

ABBREVIATIONS C−C, carbon−carbon; CoA, coenzyme A; OAH, oxyanion hole; wt, wild type REFERENCES

(1) Masamune, S., Walsh, C. T., Sinskey, A. J., and Peoples, O. P. (1989) Poly-(R)-3-hydroxybutyrate (PHB) biosynthesis: Mechanistic studies on the biological Claisen condensation catalyzed by β-ketoacyl thiolase. Pure Appl. Chem. 61, 303−312. (2) Hiser, L., Basson, M. E., and Rine, J. (1994) ERG10 from Saccharomyces cerevisiae encodes acetoacetyl-CoA thiolase. J. Biol. Chem. 269, 31383−31389. (3) Haapalainen, A. M., Meriläinen, G., and Wierenga, R. K. (2006) The thiolase superfamily: Condensing enzymes with diverse reaction specificities. Trends Biochem. Sci. 31, 64−71. (4) Miziorko, H. M. (2011) Enzymes of the mevalonate pathway of isoprenoid biosynthesis. Arch. Biochem. Biophys. 505, 131−143. (5) Bond-Watts, B. B., Bellerose, R. J., and Chang, M. C. Y. (2011) Enzyme mechanism as a kinetic control element for designing synthetic biofuel pathways. Nat. Chem. Biol. 7, 222−227. (6) Dellomonaco, C., Clomburg, J. M., Miller, E. N., and Gonzalez, R. (2011) Engineered reversal of the β-oxidation cycle for the synthesis of fuels and chemicals. Nature 476, 355−359. K


Article

Biochemistry (25) Middleton, B. (1973) The oxoacyl-coenzyme A thiolases of animal tissues. Biochem. J. 132, 717−730. (26) Kursula, P., Sikkilä, H., Fukao, T., Kondo, N., and Wierenga, R. K. (2005) High resolution crystal structures of human cytosolic thiolase (CT): A comparison of the active sites of human CT, bacterial thiolase, and bacterial KAS I. J. Mol. Biol. 347, 189−201. (27) Haapalainen, A. M., Meriläinen, G., Pirilä, P. L., Kondo, N., Fukao, T., and Wierenga, R. K. (2007) Crystallographic and kinetic Studies of human mitochondrial acetoacetyl-CoA thiolase: The importance of potassium and chloride ions for its structure and function. Biochemistry 46, 4305−4321. (28) Kiema, T.-R., Harijan, R. K., Strozyk, M., Fukao, T., Alexson, S. E. H., and Wierenga, R. K. (2014) The crystal structure of human mitochondrial 3-ketoacyl-CoA thiolase (T1): Insight into the reaction mechanism of its thiolase and thioesterase activities. Acta Crystallogr., Sect. D: Biol. Crystallogr. 70, 3212−3225. (29) Fage, C. D., Meinke, J. L., and Keatinge-Clay, A. T. (2015) Coenzyme A-free activity, crystal structure, and rational engineering of a promiscuous β-ketoacyl thiolase from Ralstonia eutropha. J. Mol. Catal. B: Enzym. 121, 113−121. (30) Gibson, D. G., Young, L., Chuang, R. Y., Venter, J. C., Hutchison, C. A., 3rd, and Smith, H. O. (2009) Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 6, 343− 345. (31) Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M. R., Appel, R. D., and Bairoch, A. (2005) Protein Identification and Analysis Tools on the ExPASy Server, in The Proteomics Protocols Handbook (Walker, J. M., Ed.), pp 571−607, Humana Press. (32) Kabsch, W. (2010) XDS. Acta Crystallogr., Sect. D: Biol. Crystallogr. 66, 125−132. (33) Evans, P. R., and Murshudov, G. N. (2013) How good are my data and what is the resolution? Acta Crystallogr., Sect. D: Biol. Crystallogr. 69, 1204−1214. (34) McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C., and Read, R. J. (2007) Phaser crystallographic software. J. Appl. Crystallogr. 40, 658−674. (35) Bunkoczi, G., and Read, R. J. (2011) Improvement of molecularreplacement models with Sculptor. Acta Crystallogr., Sect. D: Biol. Crystallogr. 67, 303−312. (36) Adams, P. D., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J., Moriarty, N. W., Read, R. J., Sacchettini, J. C., Sauter, N. K., and Terwilliger, T. C. (2002) PHENIX: Building new software for automated crystallographic structure determination. Acta Crystallogr., Sect. D: Biol. Crystallogr. 58, 1948−1954. (37) Emsley, P., and Cowtan, K. (2004) Coot: Model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126−2132. (38) Montes, M., Miteva, M. A., and Villoutreix, B. O. (2007) Structure-based virtual ligand screening with LigandFit: Pose prediction and enrichment of compound collections. Proteins: Struct., Funct., Genet. 68, 712−725. (39) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera–a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605−1612. (40) Peretó, J., López-García, P., and Moreira, D. (2005) Phylogenetic analysis of eukaryotic thiolases suggests multiple proteobacterial origins. J. Mol. Evol. 61, 65−74. (41) Baerga-Ortiz, A., Popovic, B., Siskos, A. P., O’Hare, H. M., Spiteller, D., Williams, M. G., Campillo, N., Spencer, J. B., and Leadlay, P. F. (2006) Directed mutagenesis alters the stereochemistry of catalysis by isolated ketoreductase domains from the erythromycin polyketide synthase. Chem. Biol. 13, 277−285. (42) Nicaise, M., Valerio-Lepiniec, M., Minard, P., and Desmadril, M. (2004) Affinity transfer by CDR grafting on a nonimmunoglobulin scaffold. Protein Sci. 13, 1882−1891.

L


Structural and Biochemical Studies of Substrate Selectivity in Ascaris

Recommend Documents