Biochemical and structural characterization of the trans-enoyl-coa reductase from treponema denticola

Article abstract of DOI:10.1021/bi300879n

The production of fatty acids is an important cellular pathway for both cellular function and the development of engineered pathways for the synthesis of advanced biofuels. Despite the conserved reaction chemistry of various fatty acid synthase systems, the individual isozymes that catalyze these steps are quite diverse in their structural and biochemical features and are important for controlling differences at the cellular level. One of the key steps in the fatty acid elongation cycle is the enoyl-ACP (CoA) reductase function that drives the equilibrium forward toward chain extension. In this work, we report the structural and biochemical characterization of the trans-enoyl-CoA reductase from Treponema denticola (tdTer), which has been utilized for the engineering of synthetic biofuel pathways with an order of magnitude increase in product titers compared to those of pathways constructed with other enoyl-CoA reductase components. The crystal structure of tdTer was determined to 2.00 A resolution and shows that the Ter enzymes are distinct from members of the FabI, FabK, and FabL families but are highly similar to members of the FabV family. Further biochemical studies show that tdTer uses an ordered bi-bi mechanism initiated by binding of the NADH redox cofactor, which is consistent with the behavior of other enoyl-ACP (CoA) reductases. Mutagenesis of the substrate binding loop, characterization of enzyme activity with respect to crotonyl-CoA, hexenoyl-CoA, and dodecenoyl-CoA substrates, and product inhibition by lauroyl-CoA suggest that this region is important for controlling chain length specificity, with the major portal playing a more important role for longer chain length substrates.

Full text of DOI:10.1021/bi300879n

Article
pubs.acs.org/biochemistry
Biochemical and Structural Characterization of the trans-Enoyl-CoA
Reductase from Treponema denticola
Brooks B. Bond-Watts,† Amy M. Weeks,† 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: The production of fatty acids is an important
cellular pathway for both cellular function and the develop-
ment of engineered pathways for the synthesis of advanced
biofuels. Despite the conserved reaction chemistry of various
fatty acid synthase systems, the individual isozymes that
catalyze these steps are quite diverse in their structural and
biochemical features and are important for controlling
differences at the cellular level. One of the key steps in the
fatty acid elongation cycle is the enoyl-ACP (CoA) reductase
function that drives the equilibrium forward toward chain
extension. In this work, we report the structural and biochemical characterization of the trans-enoyl-CoA reductase from
Treponema denticola (tdTer), which has been utilized for the engineering of synthetic biofuel pathways with an order of
magnitude increase in product titers compared to those of pathways constructed with other enoyl-CoA reductase components.
The crystal structure of tdTer was determined to 2.00 Å resolution and shows that the Ter enzymes are distinct from members of
the FabI, FabK, and FabL families but are highly similar to members of the FabV family. Further biochemical studies show that
tdTer uses an ordered bi-bi mechanism initiated by binding of the NADH redox cofactor, which is consistent with the behavior of
other enoyl-ACP (CoA) reductases. Mutagenesis of the substrate binding loop, characterization of enzyme activity with respect
to crotonyl-CoA, hexenoyl-CoA, and dodecenoyl-CoA substrates, and product inhibition by lauroyl-CoA suggest that this region
is important for controlling chain length specificity, with the major portal playing a more important role for longer chain length
substrates.
he construction of hydrocarbon backbones through the
fatty acid metabolic machinery plays a key role in cellular
serve as the enoyl-CoA reductase in this system.20,21 In terms of
phylogeny, Ters are distinct from members of the FabI family,
which represents the major enoyl-ACP reductase used in type
II FAS systems, as well as from the FabK and FabL
isozymes.9,10 However, they are quite similar to FabV enzymes,
which is a relatively recently discovered class of enoyl-ACP
reductases that was discovered in Vibrio cholerae.22 On the basis
of the ability of Ters to amplify fuel titers in engineered CoA-
linked pathways, we set out to further structurally and
biochemically characterize tdTer. We now report the crystal
structure of tdTer as well as mutagenesis studies to examine the
function of the substrate binding loop in determining chain
length specificity.
T
maintenance1,2 as well as in the development of engineered
pathways for the production of advanced biofuels.3−8 Although
the chemistry of fatty acid synthesis is highly conserved, there is
a range of diversity in the enzyme systems that catalyze these
reactions that is important for their physiological function.9−12
Indeed, the choice between individual FAS components with
different biochemical behavior is important at the cellular level
for controlling product titers from engineered biofuel pathways.
For example, the switch from the native flavin-dependent
enoyl-CoA reductase used in the production of n-butanol, a key
second-generation biofuel, to a flavin-independent trans-enoyl-
CoA reductase from Treponema denticola (tdTer)13 leads to an
order of magnitude increase in product yield in engineered
Escherichia coli.14,15 Interestingly in this case, the Ter appears to
allow the pathway flux to be driven forward without the use of
malonyl-CoA and ATP by introducing a kinetic trap at the
enoyl-CoA reduction step.14
MATERIALS AND METHODS
■
Commercial Materials. Terrific Broth (TB), LB Broth
Miller (LB), LB Agar Miller, reagent grade triethylamine
(TEA), and glycerol were purchased from EMD Biosciences
(Darmstadt, Germany). Isopropyl β-D-1-thiogalactopyranoside
(IPTG), D-glucose, dithiothreitol (DTT), phenylmethanesul-
The Ter family of reductases was first identified from Euglena
gracilis (egTer), a photosynthetic algae with five different FAS
systems, including a type II malonyl-CoA-independent pathway
in the mitochondria that uses CoA rather than ACP as the acyl
carrier.16−19 The egTer enzyme was isolated and characterized
from E. gracilis mitochondrial extracts and was proposed to
Received: June 30, 2012
Revised: August 7, 2012
Published: August 20, 2012
© 2012 American Chemical Society
6827
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
fonyl fluoride (PMSF), sodium phosphate, sodium chloride,
streptomycin sulfate, Tris-HCl, Tris base, hydrogen chloride,
sodium hydroxide, high-performance liquid chromatography
(HPLC) grade methylene chloride, HPLC grade acetonitrile,
and carbenicillin (Cb) were purchased from Fisher Scientific
(Pittsburgh, PA). Crotonyl-CoA, butyryl-CoA, hexanoyl-CoA,
lauroyl-CoA, trans-2-hexenoic acid, coenzyme A hydrate, N,N-
dimethylformamide (DMF), ethyl chloroformate, N,N,N,N-
tetramethylethylenediamine (TEMED), NADH, NADPH,
NAD+, and NADP+ were purchased from Sigma-Aldrich (St.
Louis, MO). trans-2-Dodecenoic acid was purchased from
Oakwood Products, Inc. (West Columbia, SC). Imidazole and
formic acid were purchased from Acros Organics (Geel,
Belgium). Potassium bicarbonate was purchased from Mal-
linckrodt (St. Louis, MO). Polyacrylamide (30%, 37.5:1),
electrophoresis grade sodium dodecyl sulfate (SDS), and
ammonium persulfate were purchased from Bio-Rad Labo-
rabories (Hercules, CA). PageRuler Plus Prestained Protein
Ladder and DNase were purchased from Fermentas (Glen
Burnie, MD). Deoxynucleotides (dNTPs) and Platinum Taq
High-Fidelity polymerase (Pt Taq HF) were purchased from
Invitrogen (Carlsbad, CA). Complete Protease Inhibitor tablets
were purchased from Roche Applied Science (Indianapolis,
IN). All restriction enzymes, antarctic phosphatase, and T4
DNA ligase were purchased from New England Biolabs
(Ipswich, MA). Phusion polymerase was purchased from
Finnzymes (Lafayette, CO). Ni-NTA agarose resin was
purchased from Qiagen (Valencia, CA). pET16b was purchased
from Novagen (San Diego, CA). Oligonucleotides were
purchased from Integrated DNA Technologies (Coralville,
IA) and resuspended at a stock concentration of 100 μM in 10
mM Tris-HCl (pH 8.5). DNA was isolated using the QIAprep
Spin Miniprep Kit, QIAquick PCR Purification Kit, and
QIAquick Gel Extraction Kit (Qiagen) as appropriate. All
absorbance readings were taken on a DU-800 spectrometer
(Beckman-Coulter, Fullerton, CA) or a SpectraMax M2 plate
reader (Molecular Devices, Toronto, ON). RP-HPLC
purifications were performed on an Agilent 1200 series
HPLC system coupled to a diode-array detector. Liquid
chromatography−mass spectrometry data were collected on
an Agilent 1290 series instrument. High-resolution mass
spectral analyses were conducted at the College of Chemistry
Mass Spectrometry Facility.
TdTer F1 and TdTer R101 primers and insertion into the
NdeI-XhoI restriction sites of pET16b. The Y240F mutant was
constructed by splicing by overlap extension polymerase chain
reaction using standard methods with TdTer F1/TdTer SOE
R1 (5′-end) and TdTer SOE F1/TdTer R101 (3′-end) and
inserted into the NdeI−XhoI restriction sites of pET16b (Table
S1 of the Supporting Information). All other mutants were
generated by site-directed mutagenesis using Quikchange
(Table S1 of the Supporting Information). pET23a-His10-
Tev-Ter was constructed by amplification of the synthetic ter
gene with the TdTer F102 and TdTer R101 primers and
insertion into the SfoI−XhoI restriction sites of a pET23a
derivative, which was modified to encode a His10 tag and TEV
protease cleavage site (Macrolab, University of California,
Berkeley, CA). All plasmids were verified by sequencing
following construction (Quintara Biosciences, Albany, CA).
Expression of Ter Variants. TB containing carbenicillin
(50 μg/mL) was inoculated to an OD600 of 0.05 with an
overnight TB culture of E. coli BL21(de3)-T1R, freshly
transformed with the expression plasmid. The cultures were
grown at 37 °C and 200 rpm to an OD600 of 0.6−0.8 before
induction with IPTG (1 mM) and a decrease in the
temperature to 30 °C. Cells were harvested 4 h after IPTG
addition by centrifugation at 12300g for 7 min at 4 °C and
stored at −80 °C.
Expression of Selenomethionine His10-Tev-Ter. LB
containing carbenicillin (50 mL, 50 μg/mL) was inoculated
to an OD600 of 0.05 with an overnight LB culture of E. coli
BL21(de3)-T1R, freshly transformed with pET23a-His10-Tev-
Ter. The cultures were grown at 37 °C and 200 rpm to an
OD600 of 0.6−0.8. The culture was spun down (12300g for 7
min) and washed in 50 mL of M9-MOPS medium [1× M9
salts, 100 mM MOPS (pH 7.4), 1% D-glucose, 2 mM MgSO4,
0.1 mM CaCl2, 0.1 μg/mL thiamine, 10 μM ferrous sulfate, 0.1
μM zinc sulfate, 0.8 μM manganese chloride, 0.15 μM cupric
sulfate, 0.3 μM cobalt chloride, 4 μM boric acid, and 30 nM
ammonium molybdate] containing carbenicillin (50 μg/mL).
The cells were centrifuged (12300g for 7 min) and resuspended
in M9-MOPS Cb medium (50 mL). One liter of prewarmed
(37 °C) M9-MOPS Cb medium was then inoculated with the
cell suspension (10 mL). The culture was grown at 37 °C and
200 rpm to an OD600 of 0.4−0.8 and supplemented with a
filter-sterilized amino acid mixture (leucine, isoleucine, and
valine, final concentration of 50 mg/L; phenylalanine, lysine,
and threonine, final concentration of 100 mg/L; selenomethio-
nine, final concentration of 75 mg/L). After an additional
incubation at 15 min, the cells were cooled on ice while being
shaken for 30 min before induction with IPTG (1 mM). The
culture was then grown overnight (12−16 h) at 16 °C and 200
rpm. Cells were harvested by centrifugation at 12300g for 7 min
at 4 °C and stored at −80 °C.
Bacterial Strains. E. coli DH10B-T1R (Invitrogen) and
BL21(de3)-T1R (New England Biolabs) were used for the
construction of plasmids and protein production, respectively.
Phylogenetic Analysis. Using sequences of characterized
Ter (T. denticola13 and E. gracilis21) and FabV enzymes (from
V. cholerae,22 Burkholderia mallei,23 Pseudomonas aeruginosa,24
Xanthomonas orzyae,25 and Yersinia pestis26) as well as FabI
from E. coli, the UniRef50 database27,28 was searched to yield
sequences that are less than 50% identical. For each homology
search, the top 10 scoring entries were selected. All identified
sequences were compiled, redundant sequences were removed,
and the sequences were aligned with MEGA 529 using the
MUSCLE algorithm.30 The alignment output was analyzed
over 394 positions using the maximum likelihood method in
MEGA with a nearest-neighbor-interchange strategy, while
allowing for deletion of gaps that exist in less than 50% of the
sequences and 500 bootstrap replicates to evaluate the
confidence.
Purification of Ter Variants. The cell pellet was thawed
and resuspended at a level of 5 mL/g of cell pellet in buffer A
[50 mM sodium phosphate, 300 mM sodium chloride, and 20
mM imidazole (pH 8.0) at 4 °C] containing PMSF (500 μM)
and DNase (0.7 unit/g of cell pellet). The cells were
homogenized prior to lysis by being passed through a French
pressure cell (Thermo Scientific) at 14000 psi. The total lysate
was centrifuged at 12300g for 45 min at 4 °C to separate the
soluble and insoluble fractions. The DNA was precipitated from
the cleared cell lysate with streptomycin sulfate [20% (w/v)]
added dropwise over 10 min at 4 °C to a final concentration of
1%. The precipitated DNA was removed by centrifugation at
Construction of Plasmids. pET16b-His10-Ter was con-
structed by amplification of the synthetic ter gene with the
6828
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
12300g for 30 min at 4 °C. The supernatant was loaded onto a
Ni-NTA agarose column (3−5 mL) using an AKTA Purifier
(Lawrence Berkeley National Laboratory, Berkeley, CA). Data
sets for native crystals were collected at a wavelength of 1.116
Å, while selenomethionine-modified Ter crystals were analyzed
with an inflection/peak wavelength of 0.979 Å and remote
wavelength of 0.957 Å. Data sets were processed and merged
with XDS and XSCALE. Experimental phases were determined
from the location of selenium atoms using Phenix AutoSol31
and used to build an initial model for selenomethionine-
modified Ter with Phenix AutoBuild. This model was then used
to determine the structure of the native protein by molecular
replacement using Phenix AutoMR and AutoBuild to build a
near-complete chain trace of each crystal. Iterative cycles of
Phenix AutoRefine and manual refinement in Coot32 were used
to generate the final model. Cocrystallization with stoichio-
metric amounts of substrates and/or products (NADH, NAD+,
NAD+/crotonyl-CoA, and NAD+/butyryl-CoA) was tested, but
insufficient electron density was observed to create a model for
any bound ligand.
Modeling Substrates. The docking model was generated
using Glide SP33−35 in the Maestro molecular visualization
environment. Each substrate was first optimized and minimized
without taking into account nonpolar hydrogens using the
Maestro Preparation Wizard. The NADH was docked first as a
flexible ligand in the tdTer chain A active site in 20 poses by
generating a scoring grid (size, 20 Å × 20 Å × 20 Å; ligand
search range, 14 Å × 14 Å × 14 Å) encompassing the entire
active site and correlated with the location of the bound NADH
in the Y. pestis FabV crystal structure [Protein Data Bank
(PDB) entry 3ZU3].26 The top five poses were analyzed, and
the docking model most closely resembling the Y. pestis FabV
crystal structure was used to generate a receptor grid (size, 20 Å
× 20 Å × 20 Å; ligand search range, 14 Å × 14 Å × 14 Å)
centered around Y240 to dock the crotonyl-CoA substrate
using the same approach.
Synthesis of Acyl-CoAs. Acyl-CoAs were prepared using
the mixed anhydride method.36 Briefly, trans-2-hexenoic acid or
trans-2-dodecenoic acid (52 μmol) was preincubated with
triethylamine (1 equiv, 52 μmol, 7.3 μL) for 30 min at room
temperature in CH2Cl2 (1 mL). The reaction mixture was then
cooled to 4 °C before the addition of ethyl chloroformate (1
equiv, 52 μmol, 5.0 μL). After incubation for 2 h, CH2Cl2 was
evaporated from the reaction mixture with N2 at 4 °C. The
residue was resuspended in DMF (1.5 mL), transferred into a
stirred solution of coenzyme A hydrate (0.5 equiv, 26 μmol, 20
mg) in 0.4 M potassium bicarbonate (pH 8.4, 1.5 mL), and
incubated at room temperature for 10 min. The reaction
mixture was then acidified to pH 3−4 with formic acid, diluted
to 50 mL with water, flash-frozen with liquid N2, and
lyophilized. After resuspension in water (1 mL), the acyl-CoA
was purified by RP-HPLC on an Eclipse XDB C-18 column (5
μm, 9.4 mm × 250 mm, Agilent) using a linear gradient from 0
to 50% acetonitrile over 18 min with water as the aqueous
mobile phase (3 mL/min). The fractions containing product
were pooled and lyophilized.
̈
FPLC system (GE Healthcare) at a flow rate of 0.5 mL/min.
The column was washed at a flow rate of 2 mL/min with 20
column volumes (cv) of buffer A followed by 20 cv of 92%
buffer A and 8% buffer B [50 mM sodium phosphate, 300 mM
sodium chloride, and 250 mM imidazole (pH 8.0) at 4 °C].
The column was eluted with buffer B. Protein-containing
fractions were identified and pooled by their absorbance at 280
nm and used for downstream purification as described below.
The concentration of purified protein was estimated using the
extinction coefficient at 280 nm calculated by the ExPASy
Peptide Properties Calculator for His10-Ter (42560 M−1 cm−1).
His10-Ter Variants. The pooled fractions from the Ni-NTA
agarose column were concentrated to <5 mL using an Amicon
stirred ultrafiltration cell with an Ultracel 5000 molecular
weight cutoff (MWCO) filter membrane (Millipore), loaded
onto a G-25 column (100 mL), and eluted with buffer C [20
mM Tris-HCl and 50 mM sodium chloride (pH 7.5) at 4 °C].
The protein-containing fractions were identified and pooled by
their absorbance at 280 nm. After concentration to ∼5 mg/mL,
glycerol (60% stock solution) was added to a final
concentration of 5% (v/v). Proteins were then flash-frozen in
liquid nitrogen and stored at −80 °C.
GTGA-Ter. The pooled fractions from the Ni-NTA agarose
column were concentrated to <3 mL as previously described.
TEV protease (1 mg of TEV/50 mg of Ter) and DTT (1 mM)
were added, and the mixture was dialyzed at 4 °C for 16 h
against TEV cleavage buffer [3.5 L; 50 mM sodium phosphate,
300 mM sodium chloride, and 1 mM DTT (pH 8.0)] using
3500−5000 MWCO cellulose ester tubing. To remove the
His10-tagged TEV protease and the His10 peptide, the dialysate
was passed over the washed (20 cv of buffer A) Ni-NTA
agarose column. The column flow-through and an additional
wash (1 cv of buffer A) were collected and concentrated to 2
mL using a 3000 MWCO Amicon Ultra-15 apparatus. The
concentrated protein was loaded onto a HiLoad 16/60
Superdex 75 prep grade column (GE Healthcare) using an
̈
AKTA Purifier FPLC system (1 mL/min). GTGA-Ter was
eluted with buffer C and concentrated to 30−60 mg/mL for
immediate use in crystallization. For purification of selenome-
thionine GTGA-Ter, all solutions were supplemented with
TCEP (0.5 mM).
Size-Exclusion Chromatography. Ter variants (∼0.5 mg)
were analyzed on a HiLoad 16/60 Superdex 200 prep grade
̈
column (GE Healthcare) using an AKTA Purifier FPLC system
(1 mL of buffer C/min, 500 μL sample loop). The elution
volume:void volume ratio was compared to those of the
standard proteins (Bio-Rad). Protein molecular weight was
estimated by fitting the data using Excel (Microsoft, Redmond,
WA) to the equation log10 MW = m × Ve/Vo + b, where MW is
the molecular weight, Ve is the elution volume, and Vo is the
void volume.
Crystallization and Structure Determination of GTGA-
Ter. Protein crystals were obtained using the hanging drop
vapor diffusion method by combining equal volumes of a
protein solution (diluted to 20 mg/mL) and a reservoir
solution [0.1 M sodium citrate (pH 7.5), 25% polyethylene
glycol 6000, and 2.5% glycerol]. Crystals grew within 2 days
and were cryoprotected by being briefly soaked in a solution
containing 75% reservoir solution and 25% ethylene glycol
followed by flash-freezing in liquid nitrogen. Data were
collected at Beamline 8.3.1 at the Advanced Light Source
trans-2-Hexenoyl-CoA. The >85% pure fractions were
pooled and repurified by RP-HPLC to yield a >95% pure
1
product: H NMR (600 MHz, D2O, 25 °C) δ (alkyl chain
numbered as CX* from carbonyl) 8.52 (s, 1H, H8), 8.22 (s, 1H,
H2), 6.89 (m, 1H, C3*), 6.13 (m, 2H, H1′ and C2*), 4.57 (m,
1H, H4′), 4.21 (m, 2H, H5′), 4.00 (s, 1H, H3″), 3.82 (m, 1H,
H1a″), 3.53 (m, 1H, H1b″), 3.41 (t, J = 6.6 Hz, 2H, H5″), 3.31 (t,
J = 6.3 Hz, 2H, H8″), 2.99 (t, J = 6.3 Hz, 2H, H9″), 2.39 (t, J =
6.6 Hz, 2H, H6″), 2.13 (q, J = 7.2 Hz, 2H, C4*), 1.41 (sex, J =
6829
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
7.3 Hz, 2H, C5*), 0.85 (m, 6H, H10″ and C6*), 0.73 (s, 3H,
H
dodecenoyl-CoA (5, 10, or 15 μM) with 1.5, 2.5, 5, 10, 25, 37.5,
50, 100, or 200 μM NADH.
11″); 13C NMR (600 MHz, D2O, 25 °C) δ (alkyl chain
Product Inhibition. Assays were performed as described
for steady-state kinetics. The specific conditions used for each
tdTer variant are available in the Supporting Information. Data
(n ≥ 3) were analyzed by nonlinear curve fitting to the
numbered as CX* from carbonyl) 193.86 (C1*), 174.73 (C7″),
174.00 (C4″), 155.67 (C6), 152.60 (C2), 149.32 (C4), 148.59
(C8), 139.89 (C3*), 127.81 (C2*), 118.63 (C5), 86.34 (C1′),
83.52 (C4′), 74.23 (C2′), 74.20 (C3′), 74.09 (C3″), 73.80 (C1″),
71.86, 65.35 (C5″), 38.63 (C8″), 38.31 (C2″), 38.26, 35.38 (C5″),
35.31 (C6″), 33.67 (C4*), 27.74 (C9″), 20.83 (C10″), 20.50
(C5*), 18.11 (C11″), 12.86 (C6*); HR-ESIMS (M − 2H)2− calcd
for C27H42O17N7P3S m/z 430.5791, found m/z 430.5787.
trans-2-Dodecoyl-CoA. The >95% pure fractions were
Michaelis−Menten equation to obtain Vmax,app and KM,app
.
RESULTS AND DISCUSSION
■
Phylogeny of trans-Enoyl-CoA Reductases. To examine
the relationship between the Ter and functionally related FabV
and FabI families, a phylogenetic tree was constructed using a
maximum likelihood approach with 500 bootstrap replicates
from a sequence alignment of the top 10 homologues from the
UniRef50 database of two characterized Ter enzymes (T.
denticola13 and E. gracilis21), five characterized FabVs (V.
cholerae,22 B. mallei,23 P. aeruginosa,24 X. orzyae,25 and Y.
pestis26), as well as the FabI from E. coli9,10 (Figure 1 and Figure
S1 of the Supporting Information). Like the FabVs, Ters share
little sequence identity with and are significantly larger in size
than FabIs, which serve as the major enoyl-ACP reductase of
canonical type II fatty acid synthase systems.9,10 The Ter
sequences were consequently found to cluster with FabVs on a
separate branch compared to FabIs. In this analysis, the
prokaryotic Ter enzyme from T. denticola (tdTer) lies on a
different branch compared to that of the characterized FabVs
(46−52% identical sequences); however, the eukaryotic Ter
from E. gracilis (egTer) was found to cluster with the FabV
branch (52−57% identical sequences with the functionally
encoded protein). These observations suggest that Ter
enzymes may potentially diverge from FabVs, but that the
two families of reductases may overlap or be difficult to
distinguish by sequence analysis or annotation alone.
1
pooled and used for enzyme assays: H NMR (600 MHz,
D2O, 25 °C) δ (alkyl chain numbered as CX* from carbonyl)
8.54 (s, 1H, H8), 8.24 (s, 1H, H2), 6.92 (m, 1H, C3*), 6.14 (m,
2H, H1′ and C2*), 4.58 (m, 1H, H4′), 4.23 (m, 2H, H5′), 4.02,
(s, 1H, H3″), 3.84 (m, 1H, H1a″), 3.76 (m, 1H), 3.63 (m, 1H,
H1b″), 3.54 (m, 2H), 3.43 (t, J = 6.6 Hz, 2H, H5″), 3.35 (t, J =
6.6 Hz, 2H, H8″), 3.03 (t, J = 6.3 Hz, 2H, H9″), 2.41 (t, J = 6.3
Hz, 2H, H6″), 2.15 (q, J = 7.2 Hz, 2H, C4*), 1.37 (sex, J = 7.3
Hz, 2H, C5*), 1.18 (m, 12H, C6*, C7*, C8*, C9*, C10*, C11*), 0.88
(s, 3H, H10″), 0.82 (t, J = 7.2 Hz, 3H, C12*), 0.75 (s, 3H, H11″);
13C NMR (900 MHz, D2O, 25 °C) δ (alkyl chain numbered as
CX* from carbonyl) 196.65 (C1*), 184.63, 177.46 (C7″), 176.74
(C4″), 158.33 (C6), 155.56 (C2), 152.10 (C4), 151.82 (C8),
142.60 (C3*), 130.54 (C2*), 121.42 (C5), 89.10 (C1′), 86.31
(C4′), 77.00 (C2′), 76.88 (C3′), 76.71 (C3″), 74.83 (C1″), 74.63,
68.11, 65.26 (C5″), 41.42 (C8″), 41.11 (C2″), 38.23 (C5″), 38.18
(C6″), 34.37 (C9″), 33.89 (C10″), 31.32 (C4*), 31.14 (C10*),
31.11 (C6*), 30.89 (C7*), 30.58 (C5*), 29.73 (C8*), 26.47 (C9*),
24.80 (C11*), 23.66 (C10″), 20.96 (C11″), 16.19 (C12*); HR-
ESIMS (M − 2H)2− calcd for C33H54O17N7P3S m/z 472.6260,
found m/z 472.6251.
Steady-State Kinetic Measurements. Ter activity was
measured by monitoring the decrease in absorbance of
NAD(P)H at 340 nm using a modified literature protocol.14
The specific conditions used for each tdTer variant are available
in the Supporting Information. Data were analyzed by
nonlinear curve fitting to the Michaelis−Menten equation
with data reported as mean the standard error (n ≥ 3). The
error in kcat/KM was calculated by error propagation from the
individual kinetic constants.
Ternary Complex Formation. Assays were performed as
described for steady-state kinetics. Data (n ≥ 3) were plotted
using a double-reciprocal plot and analyzed using least-squares
regression linear fitting. Ten nanograms of His10-Ter was used
per reaction (250 μL). Crotonyl-CoA was used at concen-
trations of 5, 10, 25, 50, and 100 μM and NADH at
concentrations of 3, 4, 5, 7.5, and 10 μM. All crotonyl-CoA
concentrations were run at all NADH concentrations.
Order of Binding. Assays were performed as described for
steady-state kinetics. The specific conditions for each reaction
are available in the Supporting Information. Data (n ≥ 3) were
plotted using a double-reciprocal plot and analyzed using least-
squares regression linear fitting to analyze product inhibition
patterns by NAD+ and butyryl-CoA. Twenty nanograms of
His10-Ter was used per reaction (250 μL).
Crystal Structure of the trans-Enoyl-CoA Reductase
from T. denticola. On the basis of the distinct phylogeny of
tdTer and the absence of any Ter crystal structures deposited in
the Protein Data Bank, we set out to structurally characterize
tdTer and to study its function in more detail. For structural
studies, the synthetic gene encoding tdTer was heterologously
expressed in E. coli with an N-terminal His10 tag and an
intervening TEV protease site, which was removed to yield
enzyme with a residual GTGA linker at the N-terminus (Figure
S2 of the Supporting Information). tdTer crystallized in an
asymmetric unit containing four identical copies of the Ter
monomer, which is in agreement with the apparent molecular
weight measured in solution (Figure S3 of the Supporting
Information). Until the recent deposit of the X. orzyae25 and Y.
pestis26 FabV structures, there were no structural homologues
with a sufficiently high degree of similarity to tdTer to use as
search models for molecular replacement. Therefore, the initial
model for tdTer was built with phasing determined from a two-
wavelength MAD data set of selenomethionine-substituted
protein, with the most complete monomer serving as the
molecular replacement model for the additional tdTer chains in
the asymmetric unit [root-mean-square deviation (rmsd),
0.170−0.256 Å between monomers]. Electron density was
identified for all 401 native residues in each of the monomers
but was not observed for the N-terminal GTGA linker, which
was not included in the model. The crystal structure of native
tdTer was then determined to 2.00 Å resolution by molecular
replacement using the selenomethionine-modified structure
and found to be in close agreement (rmsd, 0.246 Å for a single
monomer) (Table 1).
Substrate Inhibition. Assays were performed as described
for steady-state kinetics. Data (n ≥ 3) were analyzed by
nonlinear curve fitting to the Michaelis−Menten equation to
obtain Vmax,app and KM,app, which were used to calculate Ki and
Ki′. The following substrate concentrations were used for His10-
Ter (20 ng, 250 μL): hexenoyl-CoA (10, 100, or 200 μM) or
6830
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
Figure 1. Phylogenetic tree of Ter, FabV, and FabI homologues. Maximum likelihood analysis of characterized Ter, FabV, and FabI enzymes (red)
and homologues with <50% identical sequences from the UniRef50 database. Bootstrap values are indicated at branch points.
Table 1. Data Collection and Refinement Statistics for GTGA-Ter Structures
a
selenomethionine tdTer
tdTer
Data Collection
space group
unit cell parameters
a, b, c (Å)
P1
P1
P1
62.7, 87.6, 92.4
106.4, 109.6, 98.3
0.979
62.4, 87.6, 91.6
106.3, 109.9, 98.4
0.957 (Se remote)
19.85−2.05 (2.12−22.05)
16.9 (72.8)
62.05, 87.05, 91.46
106.14, 109.72, 98.52
1.116
α, β, γ (deg)
wavelength (Å)
resolution (Å)
Rmerge (%)
19.85−2.05 (2.12−2.05)
15.1 (59.9)
19.68−2.00 (2.071−2.00)
13.5 (69.9)
⟨I/σ(I)⟩
20.67 (4.38)
13.29 (2.77)
15.61 (2.60)
total no. of reflections
no. of unique reflections
completeness (%)
multiplicity
438763 (70586)
114220 (17931)
97.1 (96.1)
472313 (75927)
122645 (19279)
97.4 (96.0)
222904 (14852)
105852 (7127)
94.04 (88.17)
3.8 (3.9)
3.9 (3.9)
2.1 (2.1)
Refinement
Rwork (%)
18.22 (23.92)
20.62 (30.66)
23.47 (35.02)
12957
Rfree (%)
21.45 (27.29)
12861
12316
545
no. of atoms
nonsolvent
12316
solvent
641
B factor
39.90
37.70
protein
39.90
37.70
water
41.20
37.90
rmsd
bond lengths (Å)
bond angles (deg)
Ramachandran plot (%)
most favored
disallowed
0.004
0.78
0.003
0.66
98
0
97
0.25
PDB entry
4GGP
4GGO
a
Selenomethionine tdTer data sets were merged to build the model.
The tdTer structure contains 15 α-helices and 10 β-strands
comprising one six-stranded parallel β-sheet (strands 1−4, 9,
and 10) and two two-stranded antiparallel β-sheets (strands 5
and 6 and strands 7 and 8) (Figure 2A). Like other members of
the SDR family, tdTer possesses a typical Rossman fold for
nucleotide binding based on the presence of the G69-X-G71-
XXXX-G76 consensus motif. In the crystal structure, the
Rossman fold is formed by helices 3 and 5 along with the six-
stranded β-sheet and contains E80 in the cofactor binding
pocket (Figure 2A and Figure S4 of the Supporting
Information), predicting specificity for NADH over
NADPH,37−39 which is consistent with previous biochemical
studies of tdTer.13 The active site can be located by the putative
catalytic tyrosine, Y240 (Figure 2B); however, several
reductases along the tdTer branch, including tdTer, contain
both a Y-X8-K and a Y-X6-K consensus sequence that are the
respective signatures of FabVs23 and the other more common
families of enoyl-ACP reductases, such as FabI, FabK, and
6831
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
Figure 2. Crystal structure of tdTer. (A) Cartoon representation of tdTer looking down at the active site (Y240 and K249) and Rossman fold (E80)
as well as a view rotated 165° around the y-axis. The major (helices 3, 4, 6, 8−10, and 15 and strands 7 and 8) and minor (helices 1, 3, 12, and 15)
portals and the substrate binding loop (residues T278−V286, including helix 10) are indicated. (B) View of the putative active site showing Y240
and K249 as well as E80 and the NADH binding pocket formed by the Rossman fold. (C) Overlay of the tdTer crystal structure (blue) with the
FabV from Y. pestis (gray). The differences in the main chain trace are highlighted in red (tdTer) or magenta (FabV).
Figure 3. Comparison of the ACP binding site of FabV from Y. pestis compared to tdTer. (A) Sequence alignments of the N-terminal region of
FabVs and Ters involved in ACP recognition. (B) View and electrostatic map of the ACP binding surface of FabV. The electrostatic map was
generated by APBS in PyMOL with the scale indicating the calculated potential in solution. (C) View and electrostatic map of the same region of
tdTer.
FabL.40 From the relative orientations of K247 and K249
(Figure S5 of the Supporting Information), tdTer appears to
share the FabV consensus sequence as the side chain of K249 is
directed into rather than away from the active site. On the basis
of the similarity in active site architecture, tdTer also may also
share the resistance to triclosan inhibition as observed for
FabVs.22−24,26 Surrounding the active site, tdTer also has the
major (helices 3, 4, 6, 8−10, and 15 and strands 7 and 8) and
minor (helices 1, 3, 12, and 15) portals into the active site and
substrate binding loop (residues T278−V286, including helix
10) that are shared with structurally characterized FabVs and
FabIs.
Overall, the tdTer structure aligns well at the structural level
with the recently reported FabV structures from X. orzyae
(rmsd, 0.630 Å) and Y. pestis (rmsd, 0.691 Å) (Figure 2C). One
of the main differences in the main chain trace of tdTer is the
6832
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
α-helix (helix 2) that is missing from the FabV structures.
Although this loop is distal to the substrate binding pocket, it
connects helix 1 to the six-stranded β-sheet and may play a role
in the movement of one of the α-helices (helix 1) of the minor
portal. In addition to these differences, there is a 3.3 Å shift in
the substrate binding loop away from the minor portal that
increases the size of the binding pocket in the tdTer structure
compared to those of the FabV structures.
One possible functional difference between Ters and FabVs
is that the Ter enzymes are proposed to be part of a FAS
system that relies on CoA rather than ACPs,17,21 which serve as
the acyl carrier in canonical type II FAS systems. The
interaction of ACPs with the enzymes involved in fatty acid
biosynthesis is mediated by the N-terminal α-helix of the ACP
that contains negatively charged residues surrounded by a
hydrophobic region and associates with the corresponding
positively charged patch on its protein partners that is also
surrounded by hydrophobic residues.41−43 In Y. pestis FabV, the
ACP recognition domain is suggested to consist of basic
residues K4, R6, and R8 and hydrophobic residues G9, F10,
I11, V13, and A15 at the N-terminus.26 In comparison,
reductases in the tdTer branch lack the full complement of
basic residues, and several of the proposed hydrophobic areas
are replaced with charged or polar residues instead (Figure 3A).
For tdTer specifically, K4 and R8 overlay well with the
corresponding residues in FabV but the exchange of R6 for M6
results in the formation of a hydrophobic center in this charged
patch and a significant decrease in the positive charge at the
surface (Figure 3B,C). In addition, G9 and F10 in the
hydrophobic region are changed to polar amino acids, N9
and N10, respectively, thereby reducing the hydrophobicity of
the N-terminal region of tdTer. Taken together, these changes
Figure 4. Docking model of tdTer bound to NADH and crotonyl-
CoA. (A) Initial model of tdTer with the docked NADH redox
cofactor (magenta). Dotted lines represent distances of <4 Å. (B)
Docking model of the tdTer ternary complex with NADH (magenta)
and crotonyl-CoA (blue). (C) View of the terminal carbon of the
crotonyl-CoA acyl group (blue) with respect to the minor portal.
absence of the final β-strand at the C-terminus found in the
FabVs that forms part of the β-sheet in the Rossman fold
(Figure 2C). These residues are situated directly next to the
binding pocket for the adenine moiety of NADH but do not
appear to directly interact with the cofactor. The tdTer
structure also contains an elongated loop (T40−K48), with an
Figure 5. Steady-state kinetic characterization of tdTer. (A) Double-reciprocal plots with varying NADH and crotonyl-CoA concentrations indicate
ternary complex formation. (B) Double-reciprocal plots monitoring NAD+ product inhibition with respect to NADH (top) and crotonyl-coA
(bottom). Data are fit to a linear regression and represent means the standard deviation (n ≥ 3). (C) Double-reciprocal plots monitoring butyryl-
CoA product inhibtion with respect to NADH (top) and crotonyl-CoA (bottom). Data are fit to a linear regression and represent means the
standard deviation (n ≥ 3).
6833
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
a
Table 2. Kinetic Constants Measured for Enoyl-CoA Reduction by Wild-Type and Mutant TdTers
crotonyl-CoA
KM (μM)
hexenoyl-CoA
C6:C4
enzyme
kcat (s−1
91
)
kcat/KM (M−1 s−1
)
kcat (s−1
112
)
KM (μM)
kcat/KM (M−1 s−1
(9 2) × 106
)
kcat/KM ratio
wild type
Y240F
4
70 10
(1.3 0.2) × 106
(1.5 0.1) × 104
(4 1) × 104
5
12
2
7
2
b
c
c
c
c
0.73 0.02
7.1 0.8
49
4
nd
15.3 0.6
nd
3.4 0.6
nd
nd
100 40
I287A
160 50
110 30
100 20
150 30
60 10
(4.5 0.8) × 106
(1.3 0.2) × 106
(3.4 0.4) × 106
(2.7 0.8) × 106
(3 1) × 106
L291A
37
73
92
3
4
6
(3 1) × 105
50
64
48
19
2
3
4
3
38
19
18
7
4
2
5
4
5
1
1
2
F295A
(7 1) × 105
Y370A
(6 1) × 105
4
L276A/V277A
L276A/V277A/F295A
15.6 0.2
18.9 0.4
(2.6 0.5) × 105
(8.9 0.7) × 104
3
11
17
5
210 20
22.4 0.6
15
1
(1.5 0.1) × 106
2
dodecenoyl-CoA
enzyme
kcat (s−1
)
KM (μM)
kcat/KM (M−1 s−1
)
wild type
Y240F
90 10
3
1
(3 1) × 107
c
c
c
nd
nd
nd
I287A
6
2
4
2
(2 1) × 106
c
c
c
c
c
c
c
c
c
L291A
nd
nd
nd
nd
nd
nd
nd
nd
nd
c
F295A
nd
c
Y370A
nd
c
L276A/V277A
nd
L276A/V277A/F295A
24
1
1.1 0.3
(2.1 0.5) × 107
a
Data are means the standard error (n ≥ 3) as determined from nonlinear curve fitting. The error in the kcat/KM parameter was obtained from
b
propagation of the error from the individual kinetic terms. Y240F kinetic parameters were measured with 200 mM NADH; all other parameters
were measured with 100 mM NADH. Not determined.
c
a
Table 3. Inhibition of Hexanoyl-CoA and Lauroyl-CoA on Wild-Type and Mutant TdTers
enzyme
inhibitor
[inhibitor] (μM)
Ki (μM)
Ki′ (μM)
wild type
butyryl-CoA
500
1000
50
400 100
400 100
50 10
750 60
4100 400
110 10
hexanoyl-CoA
lauroyl-CoA
lauroyl-CoA
lauroyl-CoA
250
1
80 30
250 30
0.5 0.1
0.4 0.1
0.06 0.02
0.03 0.02
3.1 0.2
1.5
0.025
0.1
5
3.7 0.5
I287A
0.53 0.08
0.23 0.06
L276A/V277A/F295A
10
3
2
1
12
1
10
5.6 0.8
a
Data are means the standard error (n ≥ 3) as determined from nonlinear curve fitting. Errors in Ki and Ki′ were obtained from propagation of the
error from the individual kinetic terms.
could lead to a difference in acyl carrier specificity between the
Ter and FabV families.
ring adjacent to Y240 (Figure 4A and Figure S6 of the
Supporting Information).
The C4 substrate, crotonyl-CoA, which was previously
reported to be the major substrate for tdTer,13 was then
docked into the tdTer−NADH model using the same
approach. In this case, the full acyl-CoA substrate and the
substrate without the adenine or adenosine moiety were used
for docking studies to control for the lack of adenosine-specific
interactions typical of many CoA binding enzymes.45,46
Regardless of the ligand, all models place the double bond of
the butenoyl functional group between the docked NADH
substrate and Y240 and the terminal carbon of the acyl group at
the entrance to the minor portal near a tightly packed area
formed by a strand containing G278−V280 and helix 11
(Figure 4B,C and Figure S7 of the Supporting Information).
Docking was also attempted with longer enoyl-CoA substrates
but did not yield any models with substrate bound within the
active site. We then used MOLE47 to further examine the
minor portal of tdTer and the FabVs from X. orzyae and Y.
pestis to explore how these substrates might be accommodated.
These studies show that while the minor portal is open in all
Docking Model for NADH and Enoyl-CoA Binding. A
docking model was generated to examine how the enoyl-CoA
substrate and NADH cofactor might be oriented within the
tdTer active site. The structures of the substrates were first
minimized and then docked as flexible ligands using
Glide.33−35,44 The initial model was generated with NADH
alone, as there are a significant number of crystal structures
determined with a nicotinamide cofactor bound to a Rossman
fold that could be used for model validation. The top five
binding models were examined (Figure S6 of the Supporting
Information) with the top four models placing the nicotinamide
in the predicted binding pocket with N6 of the adenine within
hydrogen bonding distance of D116 and the 3′-OH of the
ribose sugar bound to E80. These results agree with the NADH
binding mode found in other structures containing a Rossman
fold, including the Y. pestis FabV. The second solution was
selected for docking of the enoyl-CoA substrate because it was
the most similar to the FabV structure with the nicotinamide
6834
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
three structures, it contains significant bottlenecks of 1.5−2 Å
that would prevent binding of substrates larger than crotonyl-
CoA to a static structure. A conformational change or dynamic
movement in the minor portal would therefore be necessary to
allow tdTer to interact with these substrates if they were to
bind with the acyl moiety in the minor portal of the enzyme.
Biochemical Characterization of Ter. Enzymes that
catalyze reactions similar to those of tdTer, including the
FabV from B. mallei,23 have generally been found to form a
ternary complex using an ordered bi-bi reaction mechanism in
which the redox cofactor is bound first followed by the enoyl-
CoA substrate.48−51The docking model of tdTer with NADH
and crotonyl-CoA appears to support this mechanism, as the
binding site of NADH would be occluded if crotonyl-CoA were
to bind first. Steady-state kinetic experiments with varying
NADH and crotonyl-CoA concentrations confirm the for-
mation of a ternary complex, with the double-reciprocal plots
showing a set of intersecting lines (Figure 5A). The selectivity
of tdTer for NADH [KM = 5.2 0.4 μM; kcat/KM = (1.6 0.1)
× 107 M−1 s−1] over NADPH [KM = 190 20 μM; kcat/KM =
(3.9 0.5) × 105 M−1 s−1] was also confirmed at this time. The
order of binding was then established by examining the effect of
product inhibition on the steady-state kinetics. These studies
show that NAD+ acts as a competitive inhibitor for NADH
binding and a mixed inhibitor with respect to crotonyl-CoA
(Figure 5B), while butyryl-CoA serves as a mixed inhibitor of
both substrates (Figure 5C). This pattern of product inhibition
indicates that tdTer also uses an ordered bi-bi reaction
mechanism with NADH binding first.
We next turned our attention to mutagenesis studies to
elucidate the function of the putative catalytic tyrosine, Y240, as
well as residues that might be involved in substrate selectivity.
Introduction of the Y240F mutation leads to a 5000-fold
decrease in catalytic efficiency with no significant change in KM
(Table 2), which is consistent with a key role in catalysis as
suggested for FabVs23,26 and FabIs.10,49,52 In terms of substrate
specificity, the kcat/KM of tdTer for crotonyl-CoA is similar to
the value measured for the FabV from V. cholerae22 (Table 2).
These data are consistent with the proposal that Ters can
service CoA-dependent FAS pathways; however, their activity
with ACP-dependent substrates has not yet been fully explored.
Although tdTer had previously been reported to exclude the
hexenoyl-CoA substrate,13 it has been used in the construction
of synthetic fuel pathways for hexanol production, which shows
that it is active in vivo on a C6 substrate.53 Indeed, we find that
tdTer is 7- and 25-fold more active in vitro with hexenoyl-CoA
and dodecenoyl-CoA, respectively, as a result of both higher kcat
values and lower KM values. With this result in hand, we
decided to explore the role of the substrate binding loop in
influencing chain length specificity. The I287A, L291A, F295A,
Y370A, L276A/V277A, and L276A/V277A/F295A mutations
were selected because these residues appear to be involved in
mediating the conformation of the binding loop (Figure S8 of
the Supporting Information). Three of the mutants, L291A,
F295A, and Y370A, demonstrated kinetic behavior relatively
similar to that of wild-type tdTer on crotonyl-CoA and
hexenoyl-CoA (Table 2). Because of the significant decreases in
kcat with the remaining three mutants, I287A, L276A/V277A,
and L276A/V277A/F295A, the data were normalized by
calculating the kcat/KM ratios for the C4 versus C6 substrate
for each individual mutant. Using this analysis, we see that both
the I287A single mutant and the L276A/V277A/F295A triple
mutant exhibit larger increases in catalytic efficiency on the
longer hexenoyl-CoA substrate of 100- and 17-fold compared
to that of the wild type (7-fold). These results suggest that
these mutations may increase the accessibility of the longer acyl
chain to the active site pocket through perturbations of either
the minor or major portal, each of which is modulated by the
substrate binding loop.
In addition to demonstrating higher catalytic efficiencies and
lower KM values, the longer chain substrates also cause a
significant level of inhibition of tdTer under saturating
conditions. Neither hexenoyl-CoA nor dodecenoyl-CoA
changes KM,NADH, which shows that the longer chain acyl-
CoA substrates do not compete with NADH binding (Figure
S9 of the Supporting Information). In contrast, we observe
strong product inhibition with the addition of hexanoyl-CoA
and lauroyl-CoA (Table 3 and Figure S10 of the Supporting
Information). The data were fit to a model for mixed inhibition
based on the patterns observed in the double-reciprocal plot
where both kcat and KM are affected. Because of the order of
binding as well as the absence substrate inhibition, which is
consistent with previous studies,23 we interpret these data to
mean that the acyl-CoA product can bind either tdTer-NADH
or tdTer-NAD+. The two mutants with higher C6:C4 ratios of
activity, I287A and L276A/V277A/F295A, were also charac-
terized with respect to product inhibition by lauroyl-CoA
(Table 3 and Figure S10 of the Supporting Information).
Interestingly, Ki and Ki′ are much lower than that of the wild
type for the I287A mutant but not the triple mutant. The
correlation of C12 product inhibition with the increased C6:C4
activity ratio for the I287A mutation, which is located near the
front of the binding loop, seems to suggest that the major
portal is more important than the minor portal for recognition
of longer chain substrates. In comparison, the weak preference
of the L276A/V277A/F295A mutant for the C6 substrate along
with the higher Ki and Ki′ versus wild-type tdTer seems to
imply that relaxing of the minor portal can accommodate
slightly larger substrates without increasing the affinity for an
extended acyl chain.
CONCLUSIONS
■
Enoyl-ACP (CoA) reductases catalyze a key step in the
biosynthesis of reduced hydrocarbons and are important for
driving the pathway equilibrium forward toward chain
elongation. The diversity of isozymes available for this reaction
is interesting with regard to the evolution of different FAS
systems as well as the engineering of pathways to produce
advanced biofuels. In this work, we have biochemically and
structurally characterized a member of the Ter family of
reductases, which is proposed to participate in CoA-dependent
rather than ACP-dependent pathways.21 Although the Ters are
distinct from the FabI, FabK, and FabL enoyl-ACP reductase
isozymes, they are highly related to the FabV family based on
their close phylogeny as well as their structural similarity. One
interesting difference between the tdTer structure and the FabV
structures from X. orzyae and Y. pestis is the difference in the
ACP binding surface, which may lead to differences in acyl
carrier specificity. However, it would not be surprising if Ters
are competent to reduce both CoA- and ACP-linked substrates.
Initial characterization of tdTer indicates that Y240 is the
catalytic tyrosine and that K249 is likely the active site lysine
shared by FabVs and FabIs. Further biochemical studies show
that tdTer prefers NADH over NADPH and uses an ordered
bi-bi mechanism initiated by binding of the redox cofactor.
Although tdTer had been previously reported to prefer C4
6835
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
substrates,13 we observe that the catalytic efficiency is higher on
both C6 and C12 substrates. The strong product inhibition of
wild-type and mutant Ters implies that the hydrocarbon tail of
the substrate is important for binding and recognition.
Preliminary mutagenesis experiments reveal that the substrate
binding loop appears to be involved in determining chain
length specificity through changes in either the minor or the
major portal. We have identified two mutants, I287A and
I276A/V277A/F295A, that demonstrate a stronger preference
for C6 over C4 substrates than the wild type. In this regard, the
strong impact of the I287A mutation, which is located near the
front of the major portal, on interaction with both C6 and C12
acyl-CoAs could be consistent with the binding mode observed
in FabI structures.54
cleotide; NADP(H), nicotinamide adenine dinucleotide
phosphate; Ter, trans-enoyl-CoA reductase; TEV, tobacco
etch virus.
REFERENCES
■
(1) Luckey, M. (2008) Membrane structural biology: With biochemical
and biophysical foundations, Cambridge University Press, Cambridge,
U.K.
(2) Simons, K. (2011) in Cold Spring Harbor Perspectives in Biology,
Cold Spring Harbor Laboratory Press, Woodbury, NY.
(3) Lu, X., Vora, H., and Khosla, C. (2008) Overproduction of free
fatty acids in E. coli: Implications for biodiesel production. Metab. Eng.
10, 333−339.
(4) Steen, E. J., Kang, Y., Bokinsky, G., Hu, Z., Schirmer, A.,
McClure, A., Del Cardayre, S. B., and Keasling, J. D. (2010) Microbial
production of fatty-acid-derived fuels and chemicals from plant
biomass. Nature 463, 559−562.
(5) Schirmer, A., Rude, M. A., Li, X., Popova, E., and del Cardayre, S.
B. (2010) Microbial biosynthesis of alkanes. Science 329, 559−562.
(6) Chisti, Y. (2007) Biodiesel from microalgae. Biotechnol. Adv. 25,
294−306.
(7) Hu, Q., Sommerfeld, M., Jarvis, E., Ghirardi, M., Posewitz, M.,
Seibert, M., and Darzins, A. (2008) Microalgal triacylglycerols as
feedstocks for biofuel production: Perspectives and advances. Plant J.
54, 621−639.
(8) Meng, X., Yang, J., Xu, X., Zhang, L., Nie, Q., and Xian, M.
(2009) Biodiesel production from oleaginous microorganisms. Renew-
able Energy 34, 1−5.
(9) Rock, C. O., and Cronan, J. E. (1996) Escherichia coli as a model
for the regulation of dissociable (type II) fatty acid biosynthesis.
Biochim. Biophys. Acta 1302, 1−16.
(10) White, S. W., Zheng, J., Zhang, Y. M., and Rock, C. O. (2005)
The structural biology of type II fatty acid biosynthesis. Annu. Rev.
Biochem. 74, 791−831.
(11) Schweizer, E., and Hofmann, J. (2004) Microbial type I fatty
acid synthases (FAS): Major players in a network of cellular FAS
systems. Microbiol. Mol. Biol. Rev. 68, 501−517.
(12) Asturias, F. J., Chadick, J. Z., Cheung, I. K., Stark, H., Witkowski,
A., Joshi, A. K., and Smith, S. (2005) Structure and molecular
organization of mammalian fatty acid synthase. Nat. Struct. Mol. Biol.
12, 225−232.
(13) Tucci, S., and Martin, W. (2007) A novel prokaryotic trans-2-
enoyl-CoA reductase from the spirochete Treponema denticola. FEBS
Lett. 581, 1561−1566.
(14) Bond-Watts, B. B., Bellerose, R. J., and Chang, M. C. (2011)
Enzyme mechanism as a kinetic control element for designing
synthetic biofuel pathways. Nat. Chem. Biol. 7, 222−227.
(15) 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.
(16) Inui, H., Miyatake, K., Nakano, Y., and Kitaoka, S. (1982) Wax
ester fermentation in Euglena gracilis. FEBS Lett. 150, 89−93.
(17) Inui, H., Miyatake, K., Nakano, Y., and Kitaoka, S. (1984) Fatty
acid synthesis in mitochondria of Euglena gracilis. Eur. J. Biochem. 142,
121−126.
(18) Inui, H., Miyatake, K., Nakano, Y., and Kitaoka, S. (1985) The
physiological role of oxygen-sensitive pyruvate dehydrogenase in
mitochondrial fatty acid synthesis in Euglena gracilis. Arch. Biochem.
Biophys. 237, 423−429.
(19) Tucci, S., Vacula, R., Krajcovic, J., Proksch, P., and Martin, W.
(2010) Variability of wax ester fermentation in natural and bleached
Euglena gracilis strains in response to oxygen and the elongase inhibitor
flufenacet. J. Eukaryotic Microbiol. 57, 63−69.
ASSOCIATED CONTENT
* Supporting Information
■
S
Conditions for steady-state kinetic measurements, conditions
for order of binding experiments, conditions for product
inhibition experiments, list of oligonucleotides, SDS−PAGE
gels, size-exclusion chromatograms, additional structure figures,
and additional steady-state kinetic data. This material is
available free of charge via the Internet at http://pubs.acs.org.
Accession Codes
The atomic coordinates and structure factors for the structures
reported herein have been deposited in the Protein Data Bank
as entries 4GGO and 4GGP.
AUTHOR INFORMATION
Corresponding Author
*E-mail: mcchang@berkeley.edu. Telephone: (510) 642-8545.
Fax: (510) 642-9863.
■
Author Contributions
B.B.B.-W. conducted all experiments. A.M.W. assisted in the
analysis of data. B.B.B.-W., A.M.W., and M.C.Y.C. designed the
experiments and wrote the manuscript.
Funding
This work was funded by the generous support of a National
Science Foundation CAREER award (CHE-0548245) and by
the Aldo DeBenedictis and Chevron Predoctoral Fellowships to
B.B.B.-W. A.M.W. acknowledges the support of a National
Institutes of Health NRSA Training Grant (1 T32
GMO66698) and a National Science Foundation Graduate
Research Fellowship. The Advanced Light Source is supported
by the U.S. Department of Energy under Contract DEAC02-
05CH11231.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank James Holton and George Meigs (beamline 8.3.1,
Advanced Light Source, Lawrence Berkeley National Labo-
ratory) and Kathy Durkin (University of California, Berkeley,
CA) for assistance with X-ray data collection and molecular
modeling, respectively, as well as Jonathan Winger and Tiago
Barros in the Kuriyan laboratory (University of California,
Berkeley, CA) for helpful discussions.
ABBREVIATIONS
■
(20) Inui, H., Miyatake, K., Nakano, Y., and Kitaoka, S. (1986)
Purification and some properties of short chain-length specific trans-2-
enoyl-CoA reductase in mitochondria of Euglena gracilis. J. Biochem.
100, 995−1000.
ACP, acyl carrier protein; CoA, coenzyme A; cv, column
volumes; FAS, fatty acid synthase; MAD, multiwavelength
anomalous diffraction; NAD(H), nicotinamide adenine dinu-
6836
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837

Biochemistry
Article
(21) Hoffmeister, M., Piotrowski, M., Nowitzki, U., and Martin, W.
(2005) Mitochondrial trans-2-enoyl-CoA reductase of wax ester
fermentation from Euglena gracilis defines a new family of enzymes
involved in lipid synthesis. J. Biol. Chem. 280, 4329−4338.
(22) Massengo-Tiasse, R. P., and Cronan, J. E. (2008) Vibrio cholerae
FabV defines a new class of enoyl-acyl carrier protein reductase. J. Biol.
Chem. 283, 1308−1316.
(23) Lu, H., and Tonge, P. J. (2010) Mechanism and inhibition of
the FabV enoyl-ACP reductase from Burkholderia mallei. Biochemistry
49, 1281−1289.
(24) Zhu, L., Lin, J., Ma, J., Cronan, J. E., and Wang, H. (2010)
Triclosan resistance of Pseudomonas aeruginosa PAO1 is due to FabV, a
triclosan-resistant enoyl-acyl carrier protein reductase. Antimicrob.
Agents Chemother. 54, 689−698.
(25) Li, H., Zhang, X., Bi, L., He, J., and Jiang, T. (2011)
Determination of the crystal structure and active residues of FabV, the
enoyl-ACP reductase from Xanthomonas oryzae. PLoS One 6, e26743.
(26) Hirschbeck, M. W., Kuper, J., Lu, H., Liu, N., Neckles, C., Shah,
S., Wagner, S., Sotriffer, C. A., Tonge, P. J., and Kisker, C. (2012)
Structure of the Yersinia pestis FabV enoyl-ACP reductase and its
interaction with two 2-pyridone inhibitors. Structure 20, 89−100.
(27) Li, W., Jaroszewski, L., and Godzik, A. (2001) Clustering of
highly homologous sequences to reduce the size of large protein
databases. Bioinformatics 17, 282−283.
(28) Consortium, T. U. (2012) Reorganizing the protein space at the
Universal Protein Resource (UniProt). Nucleic Acids Res. 40, D71−
D75.
(29) Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and
Kumar, S. (2011) MEGA5: Molecular evolutionary genetics analysis
using maximum likelihood, evolutionary distance, and maximum
parsimony methods. Mol. Biol. Evol. 28, 2731−2739.
(30) Edgar, R. C. (2004) MUSCLE: Multiple sequence alignment
with high accuracy and high throughput. Nucleic Acids Res. 32, 1792−
1797.
(31) Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I.
W., Echols, N., Headd, J. J., Hung, L.-W., Kapral, G. J., Grosse-
Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R.
J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P.
H. (2010) PHENIX: A comprehensive Python-based system for
macromolecular structure solution. Acta Crystallogr. D66, 213−221.
(32) Emsley, P., and Cowtan, K. (2004) Coot: Model-building tools
for molecular graphics. Acta Crystallogr. D60, 2126−2132.
(33) Friesner, R. A., Banks, J. L., Murphy, R. B., Halgren, T. A., Klicic,
J. J., Mainz, D. T., Repasky, M. P., Knoll, E. H., Shelley, M., Perry, J. K.,
Shaw, D. E., Francis, P., and Shenkin, P. S. (2004) Glide: A new
approach for rapid, accurate docking and scoring. 1. Method and
assessment of docking accuracy. J. Med. Chem. 47, 1739−1749.
(34) Halgren, T. A., Murphy, R. B., Friesner, R. A., Beard, H. S., Frye,
L. L., Pollard, W. T., and Banks, J. L. (2004) Glide: A new approach
for rapid, accurate docking and scoring. 2. Enrichment factors in
database screening. J. Med. Chem. 47, 1750−1759.
(35) Friesner, R. A., Murphy, R. B., Repasky, M. P., Frye, L. L.,
Greenwood, J. R., Halgren, T. A., Sanschagrin, P. C., and Mainz, D. T.
(2006) Extra precision glide: Docking and scoring incorporating a
model of hydrophobic enclosure for protein-ligand complexes. J. Med.
Chem. 49, 6177−6196.
(36) Rasmussen, J. T., Borchers, T., and Knudsen, J. (1990)
Comparison of the binding affinities of acyl-CoA-binding protein and
fatty-acid-binding protein for long-chain acyl-CoA esters. Biochem. J.
265, 849−855.
(40) Kim, K. H., Ha, B. H., Kim, S. J., Hong, S. K., Hwang, K. Y., and
Kim, E. E. (2011) Crystal structures of enoyl-ACP reductases I (FabI)
and III (FabL) from B. subtilis. J. Mol. Biol. 406, 403−415.
(41) Zhang, Y. M., Wu, B., Zheng, J., and Rock, C. O. (2003) Key
residues responsible for acyl carrier protein and β-ketoacyl-acyl carrier
protein reductase (FabG) interaction. J. Biol. Chem. 278, 52935−
52943.
(42) Rafi, S., Novichenok, P., Kolappan, S., Zhang, X., Stratton, C. F.,
Rawat, R., Kisker, C., Simmerling, C., and Tonge, P. J. (2006)
Structure of acyl carrier protein bound to FabI, the FASII enoyl
reductase from Escherichia coli. J. Biol. Chem. 281, 39285−39293.
(43) Chan, D. I., and Vogel, H. J. (2010) Current understanding of
fatty acid biosynthesis and the acyl carrier protein. Biochem. J. 430, 1−
19.
(44) Repasky, M. P., Shelley, M., and Friesner, R. A. (2007) Flexible
ligand docking with Glide. Current Protocols in Bioinformatics 8, Units
8−12.
(45) Barycki, J. J., O’Brien, L. K., Strauss, A. W., and Banaszak, L. J.
(2000) Sequestration of the active site by interdomain shifting.
Crystallographic and spectroscopic evidence for distinct conformations
of L-3-hydroxyacyl-CoA dehydrogenase. J. Biol. Chem. 275, 27186−
27196.
(46) Thoden, J. B., Holden, H. M., Zhuang, Z., and Dunaway-
Mariano, D. (2002) X-ray crystallographic analyses of inhibitor and
substrate complexes of wild-type and mutant 4-hydroxybenzoyl-CoA
thioesterase. J. Biol. Chem. 277, 27468−27476.
(47) Petrek, M., Kosinova, P., Koca, J., and Otyepka, M. (2007)
MOLE: A Voronoi diagram-based explorer of molecular channels,
pores, and tunnels. Structure 15, 1357−1363.
(48) Wratten, C. C., and Cleland, W. W. (1963) Product inhibition
studies on yeast and liver alcohol dehydrogenases. Biochemistry 2,
935−941.
(49) Parikh, S., Moynihan, D. P., Xiao, G., and Tonge, P. J. (1999)
Roles of tyrosine 158 and lysine 165 in the catalytic mechanism of
InhA, the enoyl-ACP reductase from Mycobacterium tuberculosis.
Biochemistry 38, 13623−13634.
(50) Sahni-Arya, B., Flynn, M. J., Bergeron, L., Salyan, M. E.,
Pedicord, D. L., Golla, R., Ma, Z., Wang, H., Seethala, R., Wu, S. C., Li,
J. J., Nayeem, A., Gates, C., Hamann, L. G., Gordon, D. A., and Blat, Y.
(2007) Cofactor-specific modulation of 11β-hydroxysteroid dehydro-
genase 1 inhibitor potency. Biochim. Biophys. Acta 1774, 1184−1191.
(51) Askonas, L. J., Ricigliano, J. W., and Penning, T. M. (1991) The
kinetic mechanism catalysed by homogeneous rat liver 3α-hydrox-
ysteroid dehydrogenase. Evidence for binary and ternary dead-end
complexes containing non-steroidal anti-inflammatory drugs. Biochem.
J. 278, 835−841.
(52) Rafferty, J. B., Simon, J. W., Baldock, C., Artymiuk, P. J., Baker,
P. J., Stuitje, A. R., Slabas, A. R., and Rice, D. W. (1995) Common
themes in redox chemistry emerge from the X-ray structure of oilseed
rape (Brassica napus) enoyl acyl carrier protein reductase. Structure 3,
927−938.
(53) Dekishima, Y., Lan, E. I., Shen, C. R., Cho, K. M., and Liao, J. C.
(2011) Extending carbon chain length of 1-butanol pathway for 1-
hexanol synthesis from glucose by engineered Escherichia coli. J. Am.
Chem. Soc. 133, 11399−11401.
̀
(54) Rozwarski, D. A., Vilcheze, C., Sugantino, M., Bittman, R., and
Sacchettini, J. C. (1999) Crystal structure of the Mycobacterium
tuberculosis enoyl-ACP reductase, InhA, in complex with NAD+ and a
C16 fatty acyl substrate. J. Biol. Chem. 274, 15582−15589.
(37) Scrutton, N. S., Berry, A., and Perham, R. N. (1990) Redesign of
the coenzyme specificity of a dehydrogenase by protein engineering.
Nature 343, 38−43.
(38) Perham, R. N., Scrutton, N. S., and Berry, A. (1991) New
enzymes for old: Redesigning the coenzyme and substrate specificities
of glutathione reductase. BioEssays 13, 515−525.
(39) Bocanegra, J. A., Scrutton, N. S., and Perham, R. N. (1993)
Creation of an NADP-dependent pyruvate dehydrogenase multi-
enzyme complex by protein engineering. Biochemistry 32, 2737−2740.
6837
dx.doi.org/10.1021/bi300879n | Biochemistry 2012, 51, 6827−6837