Evidence from Raman spectroscopy that InhA, the mycobacterial enoyl reductase, modulates the conformation of the NADH cofactor to promote catalysis

Article abstract of DOI:10.1021/ja068219m

InhA, the enoyl reductase from Mycobacterium tuberculosis, catalyzes the NADH-dependent reduction of trans-2-enoyl-ACPs. In the present work, Raman spectroscopy has been used to identify catalytically relevant changes in the conformation of the nicotinamide ring that occur when NADH binds to InhA. For 4(S)-NADD, there is an 11 cm^-1 decrease in the wavenumber of the C4-D stretching band (ν_C-D) and a 50% decrease in the width of this band upon binding to InhA. While a similar reduction in line width is observed for the corresponding band arising from 4(R)-NADD, ν_C-D for this isomer increases 34 cm^-1 upon binding to InhA. These changes in ν_C-D indicate that the nicotinamide ring adopts a bound conformation in which the 4(S)C-D bond is in a pseudoaxial orientation. Mutagenesis of F149, a conserved active site residue close to the cofactor, demonstrates that this enzyme-induced modulation in cofactor structure is directly linked to catalysis. In contrast to the wild-type enzyme, Raman spectra of NADD bound to F149A InhA resemble those of NADD in solution. Consequently, F149A is no longer able to optimally position the cofactor for hydride transfer, which correlates with the 30-fold decrease in feat and 2-fold increase in ^D(V/K_NADH) caused by this mutation. These studies thus substantiate the proposal that hydride transfer is promoted by pseudoaxial positioning of the NADH pro-4S bond, and indicate that catalysis of substrate reduction by InhA results, in part, from correct orientation of the cofactor in the ground state.

Full text of DOI:10.1021/ja068219m

Published on Web 05/02/2007
Evidence from Raman Spectroscopy That InhA, the
Mycobacterial Enoyl Reductase, Modulates the Conformation
of the NADH Cofactor to Promote Catalysis
†
‡
Alasdair F. Bell, Christopher F. Stratton, Xujie Zhang, Polina Novichenok,
§
Andrew A. Jaye, Pravin A. Nair, Sapan Parikh, Richa Rawat, and Peter J. Tonge*
Contribution from the Department of Chemistry, Stony Brook UniVersity,
Stony Brook, New York 11794-3400
Received November 22, 2006; E-mail: peter.tonge@sunysb.edu
Abstract: InhA, the enoyl reductase from Mycobacterium tuberculosis, catalyzes the NADH-dependent
reduction of trans-2-enoyl-ACPs. In the present work, Raman spectroscopy has been used to identify
catalytically relevant changes in the conformation of the nicotinamide ring that occur when NADH binds to
InhA. For 4(S)-NADD, there is an 11 cm-1 decrease in the wavenumber of the C4-D stretching band
(νC-D) and a 50% decrease in the width of this band upon binding to InhA. While a similar reduction in line
width is observed for the corresponding band arising from 4(R)-NADD, νC-D for this isomer increases 34
-
1
cm upon binding to InhA. These changes in νC-D indicate that the nicotinamide ring adopts a bound
conformation in which the 4(S)C-D bond is in a pseudoaxial orientation. Mutagenesis of F149, a conserved
active site residue close to the cofactor, demonstrates that this enzyme-induced modulation in cofactor
structure is directly linked to catalysis. In contrast to the wild-type enzyme, Raman spectra of NADD bound
to F149A InhA resemble those of NADD in solution. Consequently, F149A is no longer able to optimally
position the cofactor for hydride transfer, which correlates with the 30-fold decrease in kcat and 2-fold increase
D
in (V/KNADH) caused by this mutation. These studies thus substantiate the proposal that hydride transfer
is promoted by pseudoaxial positioning of the NADH pro-4S bond, and indicate that catalysis of substrate
reduction by InhA results, in part, from correct orientation of the cofactor in the ground state.
Introduction
Scheme 1
InhA, the enoyl reductase from Mycobacterium tuberculosis
and a member of the short-chain dehydrogenase/reductase (SDR)
1
-3
family,
catalyzes the NADH-dependent reduction of long-
chain trans-2-enoyl-ACP fatty acids in the type II fatty acid
biosynthesis pathway of M. tuberculosis (Scheme 1). This
enzyme is a target for isoniazid, a front-line antituberculosis
drug which inhibits the biosynthesis of mycolic acids, thereby
compromising cell wall integrity.4 Experiments utilizing a
temperature-sensitive mutant of InhA have validated this enzyme
-7
as an excellent target for anti-TB drug discovery, and conse-
quently a detailed understanding of the catalytic mechanism is
important for directing the design of potent InhA inhibitors.8
All members of the SDR family contain a conserved catalytic
triad which includes a tyrosine and a lysine residue. In InhA,
,9
†
Present address: Medarex Inc., Sunnyvale, CA 94041.
Present address: Department of Chemistry and Chemical Biology,
‡
Cornell University, Ithaca, NY 14853.
§
Present address: Memorial Sloan Kettering Cancer Center, New York,
2
NY 10021.
these residues are Y158 and K165 (Figure 1). Similar to the
(
(
(
1) Nakajin, S.; Takase, N.; Ohno, S.; Toyoshima, S.; Baker, M. E. Biochem.
dehydrogenases, structural and kinetic studies have revealed that
K165 in InhA functions primarily in cofactor binding, forming
hydrogen bonds with the nicotinamide ribose 2′ and 3′ hydroxyl
J. 1998, 334, 553-557.
2) Rozwarski, D. A.; Vilcheze, C.; Sugantino, M.; Bittman, R.; Sacchettini,
J. C. J. Biol. Chem. 1999, 274, 15582-15589.
3) Oppermann, U.; Filling, C.; Hult, M.; Shafqat, N.; Wu, X.; Lindh, M.;
Shafqat, J.; Nordling, E.; Kallberg, Y.; Persson, B.; Jornvall, H. Chem.
Biol. Interact. 2003, 143-144, 247-253.
5,10
groups. In addition, crystallographic data suggests that Y158
facilitates catalysis by stabilizing the enolate intermediate which
(
4) Banerjee, A.; Dubnau, E.; Quemard, A.; Balasubramanian, V.; Um, K. S.;
Wilson, T.; Collins, D.; de Lisle, G.; Jacobs, W. R., Jr. Science 1994, 263,
2
27-230.
(8) Vilcheze, C.; Morbidoni, H. R.; Weisbrod, T. R.; Iwamoto, H.; Kuo, M.;
Sacchettini, J. C.; Jacobs, W. R., Jr. J. Bacteriol. 2000, 182, 4059-4067.
(9) Sullivan, T. J.; Truglio, J. J.; Boyne, M. E.; Novichenok, P.; Zhang, X.;
Stratton, C. F.; Li, H.-J.; Kaur, T.; Amin, A.; Johnson, F.; Slayden, R. A.;
Kisker, C.; Tonge, P. J. ACS Chem. Biol. 2006, 1, 43-53.
(10) Parikh, S.; Moynihan, D. P.; Xiao, G.; Tonge, P. J. Biochemistry 1999, 38,
13623-13634.
(
(
(
5) Dessen, A.; Quemard, A.; Blanchard, J. S.; Jacobs, W. R., Jr.; Sacchettini,
J. C. Science 1995, 267, 1638-1641.
6) Quemard, A.; Sacchettini, J. C.; Dessen, A.; Vilcheze, C.; Bittman, R.;
Jacobs, W. R., Jr.; Blanchard, J. S. Biochemistry 1995, 34, 8235-8241.
7) Basso, L. A.; Zheng, R.; Musser, J. M.; Jacobs, W. R., Jr.; Blanchard, J.
S. J. Infect. Dis. 1998, 178, 769-775.
10.1021/ja068219m CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 6425-6431
9
6425

A R T I C L E S
Bell et al.
dehydrogenase (MCAD), provide a detailed framework for
interpreting data on the binding of the diene substrate analog
1
8-21
to InhA.
In the present work we demonstrate that InhA binds NADD
in a conformation that is optimized for hydride transfer. In
addition, the Raman studies, coupled with kinetic isotope effect
data, demonstrate that F149 plays a key role in modulating both
the structure and reactivity of NADH(D) when bound to InhA.
Materials and Methods
Materials. Coenzyme A (CoA) lithium salt, â-NADH, â-NAD+,
equine liver alcohol dehydrogenase (EC 1.1.1.1), and glucose-6-
phosphate dehydrogenase type XXIV from L. mesenteroides were
purchased from Sigma Chemical company. [1-D]-Glucose (98% D)
was from Cambridge Isotopes Labs (Andover, MA). trans,trans-2,4-
Decadienal was purchased from Aldrich Chemical Co. Sephadex G-25
(fine) was purchased from Pharmacia Biotech (Uppsala, Sweden). His-
Bind resin was from Novagen (Madison, WI). Oligonucleotides were
purchased from IDT, Inc. (Coralville, IA). Spin columns were purchased
from Princeton Separations Inc. (Adelphia, NJ). DNA purification kits
Figure 1. (A) Position of F149, Y158, and K165 in a binary complex of
InhA with NADH taken from Dessen et al. (ref 5) (PDB code: 1ENY).
+
(
B) Ternary complex with a C16 substrate and NAD . Although the only
+
available ternary complexes of InhA involve NAD and not NADH, a
ternary complex structure is presented in order to show the rotation of Y158
that occurs upon substrate binding (PDB code: 1BVR) (ref 2). In both
structures the NAD(H) is bound in the anti conformation such that the amide
group forms a hydrogen bond with an NAD(H) phosphate oxygen. The
figure was made with pymol (ref 36).
6
were from Qiagen Inc. (Valencia, OH). Ethanol-d (99% D), all other
buffers and salts (reagent grade and better), solvents (HPLC grade or
better), and chemicals were purchased from Fisher Scientific Co.
2
forms during hydride transfer. However, while replacement of
Y158 with Phe reduces kcat 24-fold, the Y158S mutant has wild-
(Pittsburgh, PA).
type activity, bringing into question the ultimate importance of
this residue in catalysis.10,11
Synthesis of Substrates and Substrate Analogs. trans,trans-2,4-
Decadienoic acid was synthesized from the aldehyde by oxidation with
In the dehydrogenases the third residue in the catalytic triad
is a Ser or Thr, while in the enoyl reductases this residue is
either a Phe or a Tyr (F149 in InhA). In UDP-galactose-4-
epimerase the conserved Ser (S124) functions together with
Y149 to reversibly protonate and deprotonate the ketone oxygen
of the substrate.12,13 However, the role of the conserved Phe/
Tyr in the reductases is less clear. In the crystal structure of the
NADH-InhA complex, the cofactor is observed to bind at the
bottom of a large open cavity with the side chain of F149 lying
2,23
AgNO
3
in the presence of 10% NaOH.2
Briefly, 0.93 g of silver
nitrate in 3.5 mL of water was added to a solution of 0.38 g (0.0025
mol) of trans,trans-2,4-decadienal in 100 mL of ethanol at room
temperature followed by the dropwise addition of 22 mL of 0.5 M
NaOH. After stirring overnight at room temperature, the insoluble
material was removed by filtration, and the solution was concentrated
in vacuo. The alkaline solution was extracted with diethyl ether and
then acidified with dilute HCl to precipitate the product which was
1
obtained by filtration. H NMR (500 MHz, CDCl3): δ 11.76 (s, 1H),
5
7.35 (m, 1H), 6.19 (m, 2H), 5.79 (d, 1H Jtrans ) 16.0 Hz 3J ) 3.0
Hz), 2.19 (q, 2H 3J ) 6.6 Hz), 1.43 (m, 2H), 1.29 (m, 4H), 0.88 (t, 3H
3J ) 6.6 Hz).
just above the nicotinamide ring. Based on structural data,
Rozwarski et al. have proposed that F149 protects the substrate
from adventitious attack by water.2
trans,trans-2,4-Decadienoyl-CoA (decadienoyl-CoA) was synthe-
In order to further investigate the role of F149 in catalysis,
we have used Raman spectroscopy to probe the structure and
dynamics of 4(R)- and 4(S)-NADD and a substrate analog,
trans,trans-2,4-decadienoyl-CoA (Scheme 1), bound to the wild-
type enzyme and the F149A mutant. An advantage of this
approach is that the Raman effect is instantaneous, and thus
Raman spectra are not susceptible to conformational averaging
but contain a snapshot of all species present at any instant.
Callender and co-workers have demonstrated that the C4-D
stretching vibration (νC-D) of stereospecifically deuterium-
labeled NADD is an excellent probe of the influence of the
environment on the cofactor, as this is an isolated vibration with
sized from the respective acid using the method described previously
for hexadienoyl-CoA1
9,24
and had an A265/A296 ratio of 1.13. trans-2-
Dodecenoyl-CoA (DD-CoA) was synthesized from the respective acid
6,10
using the mixed anhydride method, and DD-CoA was then used for
the synthesis of trans-2-dodecenoyl-acyl carrier protein (DD-ACP) as
described.25
Synthesis of 4(S)- and 4(R)-NADD. 4(S)-NADD was synthesized
+
enzymatically by the reduction of NAD with L. mesenteroides glucose-
2
6,27
6-phosphate dehydrogenase as described previously.
4(R)-NADD
+
was synthesized from NAD and ethanol-d
6
in the presence of equine
liver alcohol dehydrogenase as described previously (31). Both reduced
nucleotides were purified as described previously10 except that the
14-17
little or no vibrational coupling to other normal coordinates.
In addition, our numerous studies on two â-oxidation enzymes,
enoyl-CoA hydratase (ECH) and medium-chain acyl-CoA
(
(
(
18) Pellett, J. D.; Sabaj, K. M.; Stephens, A. W.; Bell, A. F.; Wu, J.; Tonge,
P. J.; Stankovich, M. T. Biochemistry 2000, 39, 13982-13992.
19) Bell, A. F.; Wu, J.; Feng, Y.; Tonge, P. J. Biochemistry 2001, 40, 1725-
1733.
20) Bell, A. F.; Feng, Y.; Hofstein, H. A.; Parikh, S.; Wu, J.; Rudolph, M. J.;
(
11) Fillgrove, K. L.; Anderson, V. E. Biochemistry 2001, 40, 12412-12421.
12) Liu, Y. J.; Thoden, J. B.; Kim, J.; Berger, E.; Gulick, A. M.; Ruzicka, F.
J.; Holden, H. M.; Frey, P. A. Biochemistry 1997, 36, 10675-10684.
13) Thoden, J. B.; Gulick, A. M.; Holden, H. M. Biochemistry 1997, 36, 10685-
Kisker, C.; Whitty, A.; Tonge, P. J. Chem. Biol. 2002, 9, 1247-1255.
(
(21) Wu, J.; Bell, A. F.; Luo, L.; Stephens, A. W.; Stankovich, M. T.; Tonge,
P. J. Biochemistry 2003, 42, 11846-11856.
(
(
(
(
(
(22) Sheehan, J. C.; Robinson, C. A. J. Am. Chem. Soc. 1951, 73, 1207-1210.
(23) Mandai, T.; Gotoh, J.; Otera, J.; Kawada, M. Chem. Lett. 1980, 313-314.
(24) Tonge, P. J.; Anderson, V. E.; Fausto, R.; Kim, M.; PusztaiCarey, M.;
Carey, P. R. Biospectroscopy 1995, 1, 387-394.
1
0695.
14) Deng, H.; Zheng, J.; Sloan, D.; Burgner, J.; Callender, R. Biochemistry
992, 31, 5085-5092.
15) Deng, H.; Burgner, J.; Callender, R. J. Am. Chem. Soc. 1992, 114, 7997-
003.
1
(25) Rafi, S.; Novichenok, P.; Kolappan, S.; Zhang, X.; Stratton, C. F.; Rawat,
R.; Kisker, C.; Simmerling, C.; Tonge, P. J. J. Biol. Chem. 2006, 281,
39285-39293.
8
16) Chen, Y. Q.; van Beek, J.; Deng, H.; Burgner, J.; Callender, R. J. Phys.
Chem. B 2002, 106, 10733-10740.
17) van Beek, J.; Deng, H.; Callender, R.; Burgner, J. J. Raman Spectrosc.
(26) Viola, R. E.; Cook, P. F.; Cleland, W. W. Anal. Biochem. 1979, 96, 334-
340.
2
002, 33, 397-403.
(27) Orr, G. A.; Blanchard, J. S. Anal. Biochem. 1984, 142, 232-234.
6426 J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007

Structure of NADH Bound to InhA
A R T I C L E S
stability of the NADD during purification was improved by performing
the anion-exchange chromatography (FPLC, MONO Q HR10/10) at
pH 9.0 using 10 mM triethanolamine, rather than pH 7.8.
Scheme 2. Conformations of the Model Nicotinamide Compounds
used for the Calculations
Enzyme Preparation and Characterization. The F149A mutation
was introduced using the QuikChange mutagenesis kit (Stratagene).
After verification of the sequence by ABI DNA sequencing, the plasmid
was transformed into BL21(DE3)pLysS cells (Novagen) for protein
expression. Both wild-type InhA and the F149A mutant were expressed
1
0
and purified as described previously. Protein concentration was
calculated using an ꢀ280 of 37.3 mM cm-1 for the wild-type and an
-
1
280 of 35.9 mM cm-1 for the F149A mutant.
-1
ꢀ
Steady-State Kinetics for Wild-Type and F149A InhA Proteins.
All experiments were performed at 25 °C in 30 mM PIPES and 150
10
mM NaCl (pH 6.8) as described. Kinetic parameters were determined
by varying the concentration of DD-ACP (0-24 µM) at a fixed
concentration of NADH (250 µM) and by varying the concentration
of NADH (0-250 µM) at a fixed, saturating concentration of DD-
ACP (24 µM). Kinetic parameters were determined by fitting the initial
velocity data to eq 1 using GraFit 4 (Erithacus Software Ltd.). Equation
1
allows V/K to be evaluated as a single parameter (VK), thus enabling
2
8
the standard error for V/K to be directly determined.
Vi ) (V[S]VK)/(V + [S]VK)
(1)
D
Kinetic Isotope Effects. Primary kinetic isotope effects on V ( V)
D
and V/K ( (V/K)) were determined at 25 °C in 30 mM PIPES and 150
mM NaCl (pH 6.8) at a fixed, saturating concentration of NADH or
4
(S)-NADD (250 µM) and by varying the concentration of DD-ACP
(0-24 µM). Alternatively, the kinetic isotope effects were determined
at a fixed, saturating concentration of DD-ACP (24 µM) and by varying
the concentration of NADH or 4(S)-NADD (0-250 µM). Kinetic
isotope effects were calculated by fitting the initial velocity data to eq
calibrated against cyclohexanone and deuterated chloroform and are
-1
-1
accurate to (2 cm . The resolution of the Raman instrument is 8 cm
.
The software used to acquire Raman spectra was WinSpec (Princeton
Instrument), and spectral manipulations were performed using
Win-IR.
1
0
2
using Grafit 4.
V ) V[A]/(K (1 + fi(E )) + [A](1 + fi(E )))
(2)
i
A
V/K
V
Density Functional Theory Calculations. Quantum chemical
calculations were carried out using Gaussian 2003 with the Becke-3-
Lee-Yang-Parr (B3LYP) DFT functional and the 6-31G** basis
set.2 This approach was used to both optimize the structure of the
model compound N-methyl-1,4-dihydro-nicotinamide, (with deuterons
at either the 4(R) or 4(S) positions) and calculate vibrational spectra.
Calculations were performed for several conformations of the nicoti-
namide ring (Scheme 2). Both conformers I and II have planar
nicotinamide rings with the carboxamide group either syn (I) or anti
In eq 2, V is Vmax, [A] is the concentration of the varied substrate,
fi is the fraction of deuterium in NADD, and EV/K and E are the isotope
V
9,30
effects minus 1 on V/K and V, respectively. A value of 0.95 was used
for fi.
Fluorescence Titration Experiments. Equilibrium fluorescence
titrations were performed with a model FL3-21 Fluorolog-3 spectrof-
luorimeter (Edison, NJ). All measurements were carried out at 25 °C
in 100 mM PIPES (pH 6.8). The excitation wavelength was 360 nm (5
nm slit width), and the emission wavelength was 440 nm (1 nm slit
width). Briefly, 1 µL aliquots of 3 mM NADH were added to a 3 mL
of solution of 1 µM protein. The fluorescence intensity was recorded
as an average of 60 readings (2 s). For wild-type InhA data were fit to
a quadratic equation as described previously.7,10 For the F149A enzyme
NADH binding curves were not hyperbolic, but sigmoidal. These data
(II) to the ring nitrogen (N1). In conformer III, the C4 carbon is distorted
1
5° out of the plane of the ring and the carboxamide is in the anti
conformation.
Results
Steady-State Kinetics Analysis of Wild-Type and F149A
InhA Proteins. InhA and other enoyl reductases are routinely
assayed using CoA-based substrates such as DD-CoA instead
of substrates based on acyl carrier protein (ACP) due to the
relative ease of synthesizing the enoyl-CoAs. However, recently
we have discovered that enoyl-ACPs do not give rise to substrate
inhibition in contrast to the corresponding CoA substrates.
Consequently, DD-ACP was used in the present work in order
7
,10
were analyzed by the Hill equation.
Raman Spectroscopy. Raman spectra were acquired as described
previously using 700 mW of 752 nm laser excitation.1 Initially, 70
µL of enzyme (ca. 500 µM) was added to a quartz cell and 480 × 1 s
scans were accumulated. Subsequently, 1 equiv of the NADD cofactor
was added to form a binary complex, and the Raman spectrum
measured. Then, 1 equiv of dienoyl-CoA (typically between 1 and 2
µL of about 20 mM concentration) was introduced into the same cell,
and again the Raman spectrum was recorded. The Raman difference
spectra of the bound cofactor were obtained by subtracting the spectrum
of the enzyme from the spectrum of the binary complex. Similarly,
the spectrum of the substrate analog was produced by subtracting the
spectrum of the binary complex from that of the ternary complex. An
appropriate scaling factor was applied to each subtraction to eliminate
any residual protein signals. The difference spectra were wavenumber
9
D
to facilitate evaluation of V. Steady-state kinetic parameters
obtained using DD-ACP synthesized from E. coli ACP are given
in Table 1, where it can be seen that F149A InhA mutant had
a kcat value that was reduced 30-fold compared to that of wild-
type InhA. The F149A mutation had only a small affect on the
Km values for the two substrates, and the kcat/KDDACP and kcat/
(
29) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
(28) Northrop, D. B. Anal. Biochem. 1983, 132, 457-461.
(30) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007 6427

A R T I C L E S
Bell et al.
Table 1. Kinetic Parameters for Wild-Type and F149A InhA
k
cat
k
cat/KDDCOA
1
k
cat/KNADH
1
K
(
DDCoA
K
NADH
K
d NADH
enzyme
(min-1)
(min-
µ
M-1)
(min-
µ
M-1)
µM)
(µ
M)
(µM)
wild-type
F149A
190 ( 3
6.55 ( 0.25
34.9 ( 3.2
1.51 ( 0.25
5.4 ( 0.6
0.43 ( 0.03
5.4 ( 0.6
4.33 ( 0.88
35.2 ( 4.5
15.2 ( 1.64
0.60 ( 0.09
10.9 ( 1.7
ratio
29
23
13
1.2
2.3
0.055
Table 2. Vibrational Data for NADD Free in Solution and Bound
to Wild-Type and F149A InhA
4
(S)-NADD
D stretch
cm-1)
4(R)-NADD
D stretch
(cm-1)
2111
2145
C(4)
−
fwhm
C(4)
−
fwhm
(cm-1)
(
(cm-1)
2
115
59
26
free, in solution
52
29
2
104
bound to wild-type InhA
(binary complex)
2
106
111
23
52
bound to wild-type InhA
NDb
NDb
47
a
(
ternary complex )
bound to F149A
binary complex)
2
2114
(
a
Ternary complex with trans,trans-2,4-decadienoyl-CoA. b ND, not
determined.
Table 3. Primary Kinetic Isotope Effects for Wild-Type and F149A
InhA
enzyme
DV
D(V/KNADH
)
D(V/KDDCoA
)
wild-type
F149A
1.87 ( 0.36
1.54 ( 0.25
2.16 ( 0.82
4.74 ( 0.8
2.23 ( 0.21
2.1 ( 0.87
KNADH values were reduced 23- and 13-fold, respectively,
compared to those of wild-type InhA.
Figure 2. Raman difference spectra of 4(S)-NADD (A) in aqueous solution
and (B) in a binary complex with wild-type InhA. The difference spectrum
of NADD in solution was obtained by subtracting a Raman spectrum of
the buffer.
Kinetic Isotope Effects. To further investigate the role of
F149 in substrate reduction, we measured the primary kinetic
isotope effects for wild-type and F149A InhA (Table 3). While
D
D
the V and (V/KDDACP) values for wild-type and F149A InhA
were the same, within experimental error, D(V/KNADH) was
increased 2-fold for the F149A mutant.
Equilibrium Binding of NADH. Dissociation constants (Kd)
for the interaction of NADH with wild-type InhA and F149A
7
InhA were determined by fluorescence spectroscopy (Table 1).
For wild-type InhA, NADH binding curves were hyperbolic and
Kd value obtained for wild-type InhA was 0.60 ( 0.09 µM,
which is identical within experimental error to that reported
7
earlier (0.57 ( 0.04 µM). In contrast, NADH binding to the
F149A InhA enzyme was not hyperbolic but sigmoidal, indicat-
ing that the cofactor binds cooperatively to the enzyme.
Inspection of the binding curve indicated that NADH binding
was 50% complete at 10.9 ( 1.7 µM NADH. Fits of the data
7
to the Hill equation gave a Hill coefficient of 2.6 ( 0.2.
Raman Spectroscopy. The Raman difference spectra of
Figure 3. Raman difference spectra in the C-D stretching region of 2000-
deuterium-labeled 4(S)-NADD in solution and bound to wild-
-1
2
200 cm : (A) 4(S)-NADD in aqueous solution, (B) a binary complex of
-
1
type InhA are shown in Figure 2. Between 600 and 1700 cm
4(S)-NADD and wild-type InhA, (C) a binary complex of 4(S)-NADD with
F149A-InhA, (D) 4(R)-NADD in aqueous solution, (E) a binary complex
of 4(R)-NADD and wild-type InhA, and (F) a binary complex of 4(R)-
NADD with F149A-InhA. The difference spectra of NADD in solution
were obtained by subtracting Raman spectra of the buffer.
many vibrational bands can be observed which arise from
different portions of the NADD molecule. The spectra closely
resemble data previously presented and discussed by Callender
and colleagues1
4-17
and are only presented here to show the
overall spectral quality and relative intensity of the C-D
stretching vibration (νC-D), which can be observed at 2115 and
NADD isomer. The νC-D spectra are shown in Figure 3, whereas
the band positions are band widths are summarized in Table 2.
The frequency and bandwidth of the C-D bands for both 4(S)-
and 4(R)-NADD are similar and are consistent with values
previously reported by Callender and co-workers for the free
-1
2104 cm for free and bound 4(S)-NADD, respectively. Raman
spectra were also obtained for 4(S)-NADD bound to F149A
and for the corresponding enzyme complexes with the 4(R)-
6428 J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007

Structure of NADH Bound to InhA
A R T I C L E S
Table 4. Calculated Changes in Vibrational Stretching
Frequencies between Compounds II and IIIa
∆ν
C-D
(cm-1)
method
axial
equatorial
B3LYPb
MINDO/3
AM1c
-17
-14
-5
+26
+18
+11
c
a
b
∆
νC-D is the change in C4-D stretching frequency on binding. Using
c
the 6-31G** basis set. Taken from ref 14.
Table 5. Calculated Changes in Vibrational Stretching
Frequencies between Compounds I and IIIa
Figure 4. Raman spectra of trans,trans-2,4-decadienoyl-CoA in (A)
aqueous solution and (B) bound to InhA in a ternary complex with NADH.
A small contribution from NADH has been removed from the spectra. The
difference spectrum of the ligand in solution was obtained by subtracting
a Raman spectrum of the buffer.
∆ν
C-D
(cm-1)
method
axial
equatorial
B3LYPb
MINDO/3
AM1c
+3
+8
+4
+45
+40
+33
c
cofactors.14-17 However, whereas νC-D decreases 11 cm-1 when
4
(S)-NADD binds to wild-type InhA, Figure 3 shows that νC-D
a
b
∆
νC-D is the change in C4-D stretching frequency on binding. Using
-1
actually increases 34 cm for 4(R)-NADD upon binding (Table
the 6-31G** basis set. c Taken from ref 14.
2
). Finally, formation of the ternary complex of 4(S)-NADD
with InhA in the presence of the substrate analog, decadienoyl-
CoA, results in no further change in νC-D (data not shown).
The effect of enzyme binding on NADD is also illustrated
by the full width at half-maximum (fwhm) of the νC-D bands
of any significant change in the vibrational band positions of
the substrate analog on binding to InhA indicates that either
ground-state substrate distortion is not a significant contributor
to catalysis, as suggested by Fillgrove and Anderson,11 or that
the substrate analog does not sufficiently well duplicate the
structure of the true substrate to reveal enzyme-substrate
interactions important for catalysis.
Density Functional Theory Calculations. In order to assess
the impact of distorting the nicotinamide carbon out of the ring
plane, we used DFT methods to calculate the vibrational
stretching frequency (νC-D) of the pseudoaxial and pseudoequa-
torial C4-D bonds for the three model compounds shown in
Scheme 2. These calculations predict that a 15° distortion of
(Table 2), which yields information on the conformational
flexibility of the cofactor. Overall, binding of 4(S)- and 4(R)-
NADD to wild-type InhA causes the νC-D bands to narrow.
-
1
The fwhm decreases from 59 cm for 4(S)-NADD in solution
-1
-1
to 26 cm in the binary enzyme-NADD complex and 23 cm
in the ternary complex with decadienoyl-CoA. Similar changes
in the fwhm are also observed for 4(R)-NADD, and the changes
in fwhm indicate that the conformational flexibility of the
cofactor is dramatically reduced on binding to the enzyme.
Replacement of the F149 residue in InhA with Ala results in
a 30-fold reduction in kcat and a 13-fold decrease in kcat/KNADH.
To probe the mechanistic basis for these changes in kinetic
constants, we measured the Raman spectrum of 4(S)- and 4(R)-
NADD bound to F149A (Figure 3). In contrast to wild-type
InhA, when 4(S)-NADD is bound to F149A, the C4(S)-D
-
1
the C4 carbon out of the ring plane will result in a 17 cm
-
1
decrease and 26 cm increase in νC-D for the pseudoaxial and
pseudoequatorial bonds, respectively, when the carboxamide
group is in the anti conformation (model compound II to III,
Table 4). In addition, when the out of plane distortion of the
C4 carbon is accompanied by rotation of the nicotinamide
carboxamide group from a syn to an anti conformation, (model
compound I to III), the calculations indicate that νC-D will
increase for both the pseudoaxial and pseudoequatorial bonds
-
1
stretching band is observed at 2111 cm with a fwhm of 52
-
1
cm . Similar data were obtained for 4(R)-NADD, and thus,
both the position and the fwhm of the C4-D stretching bands
for the F149A-cofactor complexes are more similar to the
values observed for NADD in solution than when bound to the
wild-type enzyme (Table 2).
-
1
by 3 and 45 cm , respectively (Table 5). These DFT calcula-
tions are in good agreement with the semiempirical calculations
1
4
originally performed by Deng et al.
In a parallel set of Raman studies we examined the structure
of a substrate analog, decadienoyl-CoA, in ternary complexes
with NADH and InhA as shown in Figure 4. The Raman
spectrum of decadienoyl-CoA in solution has two intense bands
Discussion
NADD Binding to Wild-Type InhA. The Raman spectra of
4(R)- and 4(S)-NADD in solution show very broad νC-D bands
that have similar vibrational frequencies (Table 2). On binding
to InhA, there is a dramatic reduction in the fwhm of the νC-D
bands (Table 2). The narrowing of these bands indicates that
the nicotinamide ring, which exists as a population of rapidly
interconverting conformers in solution in which the C4-D
bonds are both pseudoaxial and pseudoequatorial, is conforma-
tionally constrained upon binding to InhA. In order to determine
the conformation of the bound cofactor, we must consider the
effect of ring puckering and orientation of the carboxamide
-1
at 1593 and 1631 cm . On the basis of extensive Raman studies
with isotopically labeled hexadienoyl-CoA binding to enoyl-
CoA hydratase and medium-chain acyl-CoA dehydrogenase
-
1
(
MCAD), we can assign the bands at 1593 and 1631 cm to
modes involving the in-phase and out-of-phase combinations
of the two CdC stretching coordinates which are also coupled
to the CdO stretching coordinate. For enoyl-CoA hydratase and
MCAD, binding results in a ground-state polarization of
2
0
hexadienoyl-CoA representing activation of the ligand. Sur-
1
4-17
prisingly, on binding to InhA, with or without NADH (Figure
group on νC-D.
Raman spectra of 4(R)- and 4(S)-NADD show that upon
+
4
) or NAD (data not shown), there is no significant shift in
the position of the decadienoyl-CoA Raman bands. The lack
binding to wild-type InhA, νC-D for the C4(S)-D bond
J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007 6429

A R T I C L E S
Bell et al.
-
1
decreases 11 cm , whereas νC-D for the C4(R)-D bond
binds at the bottom of a large open cavity and the benzyl side
chain of residue F149 lies just above the nicotinamide ring5
(Figure 1). The proximity of F149 to the NADH nicotinamide
ring suggests that this residue may be involved in controlling
the conformation of the bound cofactor. In support of this
hypothesis, νC-D for 4(R)- and 4(S)-NADD bound to F149A-
InhA closely resemble the corresponding bands for NADD in
solution. Thus, whereas binding to wild-type InhA causes a 50%
reduction in the fwhm values of both C4-D bands, NADD
bound to F149A-Inha shows a reduction of fwhm of only 10%
(Figure 3). Additionally, the significant changes in νC-D seen
with the binding of NADD to wild-type InhA are virtually absent
in the F149A enzyme complexes. In particular, whereas wild-
-
1
increases 34 cm
(Table 2). On the basis of the DFT
calculations (Table 4), these data are thus consistent with the
C4(S)-D and C4(R)-D bonds adopting pseudoaxial and
pseudoequatorial orientations, respectively, on binding to InhA.
The correlation with the experimental data is particularly good
when the DFT calculations on compounds II and III are used,
in which the carboxamide group is maintained in the anti
conformation (model compound II to III).
However, while the carboxamide group is anti to the
nicotinamide ring in crystal structures of NADH bound to InhA,
5
Fischer et al. have proposed that the carboxamide group is syn
in solution.31 If this is so, then the binding of NADH to InhA
must be accompanied by a syn to anti rotation of the nicotina-
mide carboxamide group. While the calculations still predict a
significant binding-induced increase in νC-D for the C4(R)-D
bond (compounds I and II, +45 cm ), inclusion of syn to anti
rotation of the carboxamide predicts a small 3 cm increase in
νC-D for the C4(S)-D bond rather than the observed 11 cm
decrease (Table 5). The discrepancy between theory and
experiment is not large, and we note that in the X-ray structure
the amide is bent out of the ring plane by 20° (Figure 1), which
could have an additional impact on νC-D that is not accounted
for in the calculations.
type InhA causes the ν
C-D
of 4(R)-NADD to increase by 34
-1
-1
cm , F149A-InhA causes only a 3 cm increase in frequency.
Thus, we conclude that F149 is directly involved in controlling
the conformation of the nicotinamide ring.
-
1
-
1
In order to shed more light on the role of F149 in catalysis,
-
1
D
D
we have determined the primary isotope effects on V and -
V/K) using NADD as a substrate. Previous experiments have
(
relied on the use of CoA-based substrates, but here we replaced
DD-CoA with DD-ACP since the latter substrate does not give
D
D
rise to substrate inhibition. For the wild-type enzyme, V, -
D
(
V/KNADH), and (V/KDDACP) have similar values, within ex-
InhA is a B-type reductase and transfers the pro-4(S) hydrogen
to the substrate. Our conclusion that InhA binds NADD in a
conformation in which the C4(S)-D bond is pseudoaxial to the
ring is consistent with separate studies by Nambiar et al. and
Almarsson and Bruice in which it was proposed that hydride
perimental error, consistent with a mechanism in which NADH
and DD-ACP bind randomly to the enzyme. In addition, the
small normal values observed for the latter isotopes effects are
consistent with a stepwise mechanism in which both hydride
transfer and protonation of the enolate intermediate are partially
rate limiting. Interestingly, when F149 is replaced with an Ala
the only significant change in the measured primary isotope
transfer is assisted by placing the appropriate C4-H bond in a
pseudoaxial orientation.3
2,33
Thus, one method InhA employs
to promote catalysis is to the productively bind the cofactor in
a conformation which is optimized for hydride transfer. The
similarity in both νC-D and fwhm for the ternary complex
obtained using trans,trans-2,4-decadienoyl-CoA (Table 2) im-
plies that the enzyme-substrate complex is already fully
organized when only the NADH cofactor is bound.
D
effects is on (V/KNADH), which increases from 2.1 in wild-
type InhA to a value of 4.9 for the mutant (Table 3). The effect
D
on (V/KNADH) can be accounted for by proposing the mutation
has both decreased the external commitment factor for NADH
and destabilized the transition state for hydride transfer, resulting
in an increase in the primary kinetic isotope effect for this step.
Since the Raman data indicate that the F149A mutant is unable
to productively orient the nicotinamide ring in the active site,
The ability of enzymes to bind NADH in the correct geometry
for hydride transfer has been previously observed in other
16,17
dehydrogenases using Raman spectroscopy.
In addition, Beis
D
the increase in (V/KNADH) thus supports a direct link between
et al. have reported the 1.5 Å structure of NADH bound to
dTDP-D-glucose dehydratase, another member of the SDR
the conformation of the NADH ring and the hydride transfer
reaction.
34
superfamily (PDB code: 1OC2). In this structure, the nico-
tinamide ring adopts a boat conformation which orients the C4
carbon and N1 nitrogen atoms 20° and 30° out of the ring plane,
respectively. This places the catalytically active C4-H bond
in a pseudoaxial conformation optimized for hydride transfer
to the substrate.
While the above argument is attractive, we must also account
D
D
for the observation that V and (V/KDDACP) are not affected
D
D
by the mutation. The fact that (V/KNADH) is affected while -
V/KDDACP) remains unchanged is consistent with the cofactor
(
binding experiments which show the affinity of NADH for the
enzyme is significantly reduced in the F149A mutant. For the
wild-type enzyme, NADH and ACP substrate dissociate with
similar rates from the ternary complex, suggesting both the
cofactor and substrate have similar external commitments.
Alternatively, for the mutant, NADH dissociates more rapidly
than the ACP substrate, thereby decreasing the external com-
mitment of the cofactor to catalysis. This explanation suggests
that binding of cofactor to the wild-type enzyme is partially
rate limiting, a result which is consistent with evidence that
NADH binding causes a conformational change of the protein
NADD Binding to F149A-InhA. InhA is a member of the
SDR family.1 All members of this family have a conserved
catalytic triad consisting of a tyrosine and lysine residues. In
the dehydrogenases the third residue is a Ser or Thr, while in
the enoyl reductases this residue is either a Phe, as in InhA, or
a Tyr, as in FabI, the enoyl reductase from E. coli. In the crystal
structure of the wild-type InhA-NADH complex, the cofactor
-3
(
31) Fischer, P.; Fleckenstein, J.; Hones, J. Photochem. Photobiol. 1988, 47,
1
93-199.
(
32) Nambiar, K. P.; Stauffer, D. M.; Kolodziej, P. A.; Benner, S. A. J. Am.
Chem. Soc. 1983, 105, 5886-5890.
D
(unpublished data). Finally, the lack of effect on V suggests
(
33) Almarsson, O.; Bruice, T. C. J. Am. Chem. Soc. 1993, 115, 2125-2138.
34) Beis, K.; Allard, S. T. M.; Hegeman, A. D.; Murshudov, G.; Philp, D.;
Naismith, J. H. J. Am. Chem. Soc. 2003, 125, 11872-11878.
the F149A mutation has not only raised the barrier to hydride
transfer but also must have decreased the rate constant for a
(
6430 J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007

Structure of NADH Bound to InhA
A R T I C L E S
step following the first irreversible step, for example, product
dissociation or protonation of the enolate intermediate, if hydride
transfer is effectively irreversible.11,35
Ternary Complexes of InhA with NADH and a Substrate
Analog. In the second series of experiments we have probed
the effect of InhA on the ground-state structure of a substrate
analog, decadienoyl-CoA. Surprisingly, there is no evidence for
any ground-state activation in this system despite the fact that
the carbonyl oxygen of the substrate oxygen is believed to be
wild-type InhA, is conformationally restricted such that the C4-
(S) bond of the nicotinamide ring adopts a pseudoaxial orienta-
tion. Replacement of F149 with Ala correlates with a 30-fold
decrease in kcat which KIE data suggest results from an increased
energy barrier to hydride transfer and decrease in the external
commitment factor for NADH. In addition, KIE data indicate
that the F149A mutation is also associated with the increase in
energy barrier for a step following hydride transfer, such as
protonation of the enolate intermediate or product dissociation.
Site-directed mutagenesis coupled with primary kinetic isotope
effect data thus indicates that F149, a member of the InhA
catalytic triad, plays a critical role in lowering the energy of
the transition state by correctly orienting the NADH cofactor
to its most active form for hydride transfer.
directly involved in two hydrogen bonds with Y158 and the
2
2
′-hydroxyl of the ribose ring of NADH. By comparison, two
enzymes from the â-oxidation pathway with similar hydrogen
bonding arrangements, enoyl-CoA hydratase and MCAD, were
both found to strongly activate the ground-state structure of a
18-21
diene analog, hexadienoyl-CoA.
However, Fillgrove and
Acknowledgment. This work was supported by Grants from
the National Institutes of Health (GM63121 and AI44639,
P.J.T.). P.J.T. is an Alfred P. Sloan Research Fellow. C.F.S. is
a Howard Hughes Medical Institute Research Fellow. In
addition, this material is based upon work supported in part by
the U.S. Army Research Office under Grant DAAG55-97-1-
Anderson have proposed that enoyl and dienoyl reductases,
where the proton-transfer step is rate limiting, may not require
stabilization of enol(ate) or dienol(ate) intermediates, as this
would be unfavorable for the proton-transfer step.11
Conclusion
0
083. We are grateful to Deborah Stoner-Ma for her assistance
In summary, Raman spectroscopy has demonstrated that the
bound NADH cofactor, in binary and ternary complexes with
with some of the Raman experiments. We also thank the
reviewers for their insightful comments.
(
35) Cook, P. F.; Cleland, W. W. Biochemistry 1981, 20, 1790-1796.
(
36) Delano, W. L. http://www.pymol.org, 2002.
JA068219M
J. AM. CHEM. SOC.
9
VOL. 129, NO. 20, 2007 6431