Ring current effects in the active site of medium-chain Acyl-CoA dehydrogenase revealed by NMR spectroscopy

Article abstract of DOI:10.1021/ja050083p

Medium-chain acyl-CoA dehydrogenase (MCAD) catalyzes the flavin-dependent oxidation of fatty acyl-CoAs to the corresponding frans-2-enoyl-CoAs. The interaction of hexadienoyl-CoA (HO-CoA), a product analogue, with recombinant pig MCAD (pMCAD) has been studied using ¹³C NMR and ¹H-¹³C HSQC spectroscopy. Upon binding to oxidized pMCAD, the chemical shifts of the C1, C2, and C3 HD carbons are shifted upfield by 12.8, 2.1, and 13.8 ppm, respectively. In addition, the ¹H chemical shift of the C3-H is also shifted upfield by 1.31 ppm while the chemical shift of the C4 HD-CoA carbon is unchanged upon binding. These changes in chemical shift are unexpected given the results of previous Raman studies which revealed that the C3=C2-C1=O HD enone fragment is polarized upon binding to MCAD such that the electron density at the C3 and C1 carbons is reduced, not increased (Pellet et al. Biochemistry 2000, 39, 13982-13992). To investigate the apparent discrepancy between the NMR and Raman data for HD-CoA bound to MCAD, ¹³C NMR spectra have been obtained for HD-CoA bound to enoyl-CoA hydratase, an enzyme system that has also previously been studied using Raman spectroscopy. Significantly, binding to enoyl-CoA hydratase causes the chemical shifts of the C1 and C3 HD carbons to move downfield by 4.8 and 5.6 ppm, respectively, while the C2 resonance moves upfield by 2.2 ppm, in close agreement with the alterations in electron density at these carbons predicted from Raman spectroscopy (Bell, A. F.; Wu, J.; Feng, Y.; Tonge, P. J. Biochemistry 2001, 40, 1725-33). The large increase in shielding experienced by the C1 and C3 HD carbons in the HD-CoA/MCAD complex is proposed to arise from the ring current field from the isoalloxazine portion of the flavin cofactor. The flavin ring current, which is only present when the enzyme is placed in an external magnetic field, also explains the differences in ¹³C NMR chemical shifts for acetoacetyl-CoA when bound as an enolate to MCAD and enoyl-CoA hydratase and is used to rationalize the observation that the line widths of the C1 and C3 resonances are narrower when the ligands are bound to MCAD than when they are free in the protein solution.

Full text of DOI:10.1021/ja050083p

Published on Web 05/21/2005
Ring Current Effects in the Active Site of Medium-Chain
Acyl-CoA Dehydrogenase Revealed by NMR Spectroscopy
†
‡
Jiaquan Wu, Alasdair F. Bell, Andrew A. Jaye, and Peter J. Tonge*
Contribution from the Department of Chemistry, Stony Brook UniVersity,
Stony Brook, New York 11794-3400
Received January 6, 2005; E-mail: peter.tonge@sunysb.edu
Abstract: Medium-chain acyl-CoA dehydrogenase (MCAD) catalyzes the flavin-dependent oxidation of
fatty acyl-CoAs to the corresponding trans-2-enoyl-CoAs. The interaction of hexadienoyl-CoA (HD-CoA),
13
1
13
a product analogue, with recombinant pig MCAD (pMCAD) has been studied using C NMR and H- C
HSQC spectroscopy. Upon binding to oxidized pMCAD, the chemical shifts of the C1, C2, and C3 HD
1
carbons are shifted upfield by 12.8, 2.1, and 13.8 ppm, respectively. In addition, the H chemical shift of
the C3-H is also shifted upfield by 1.31 ppm while the chemical shift of the C4 HD-CoA carbon is unchanged
upon binding. These changes in chemical shift are unexpected given the results of previous Raman studies
which revealed that the C3dC2-C1dO HD enone fragment is polarized upon binding to MCAD such that
the electron density at the C3 and C1 carbons is reduced, not increased (Pellet et al. Biochemistry 2000,
3
9, 13982-13992). To investigate the apparent discrepancy between the NMR and Raman data for HD-
13
CoA bound to MCAD, C NMR spectra have been obtained for HD-CoA bound to enoyl-CoA hydratase,
an enzyme system that has also previously been studied using Raman spectroscopy. Significantly, binding
to enoyl-CoA hydratase causes the chemical shifts of the C1 and C3 HD carbons to move downfield by
4.8 and 5.6 ppm, respectively, while the C2 resonance moves upfield by 2.2 ppm, in close agreement with
the alterations in electron density at these carbons predicted from Raman spectroscopy (Bell, A. F.; Wu,
J.; Feng, Y.; Tonge, P. J. Biochemistry 2001, 40, 1725-33). The large increase in shielding experienced
by the C1 and C3 HD carbons in the HD-CoA/MCAD complex is proposed to arise from the ring current
field from the isoalloxazine portion of the flavin cofactor. The flavin ring current, which is only present when
13
the enzyme is placed in an external magnetic field, also explains the differences in C NMR chemical
shifts for acetoacetyl-CoA when bound as an enolate to MCAD and enoyl-CoA hydratase and is used to
rationalize the observation that the line widths of the C1 and C3 resonances are narrower when the ligands
are bound to MCAD than when they are free in the protein solution.
Introduction
Scheme 1. Oxidation of Octanoyl-CoA by MCAD
Fatty acid â-oxidation is initiated by the conversion of fatty
acyl-CoAs to the corresponding 2-enoyl-CoAs by the flavin-
1,2
dependent acyl-CoA dehydrogenases (ACDs). Several ACDs
with varying substrate chain length specificity exist, and medium
chain acyl-CoA dehydrogenase (MCAD) is the most heavily
3
studied isozyme. The ACD reaction proceeds via the stereospe-
cific abstraction of the substrate’s C2 proton by the active site
base (E376 in MCAD) with concomitant transfer of a hydride
from C3 to the flavin N5 (Scheme 1). Experimental and
4
theoretical studies support either a concerted or stepwise
catalysis.6,7 A large number of studies have been performed in
order to determine how the enzyme overcomes the two
thermodynamic challenges to substrate oxidation, namely the
mismatch in pKa’s between substrate (21) and general base (5)
as well as the difference in redox potential between free enzyme
5
reaction, respectively, and in either case stabilization of negative
charge accumulation on the thioester carbonyl is critical for
†
Present address: Chemistry Department, Brown University, Providence,
RI 02912.
‡
Present address: Medarex inc., 1324 Chesapeake Terrace, Sunnyvale,
8
,9
(-145 mV) and substrate (-38 mV). These barriers to the
CA 94089.
(
1) Beinert, H. In The Enzymes; Boyer, P. D., Lardy, H., Myrback, K., Eds.;
Academic Press: New York, 1963; Vol. 7, pp 447-466.
2) Thorpe, C.; Kim, J. J. FASEB J. 1995, 9, 718-725.
3) Ghisla, S.; Thorpe, C. Eur. J. Biochem. 2004, 271, 494-508.
4) Pohl, B.; Raichle, T.; Ghisla, S. Eur. J. Biochem. 1986, 160, 109-115.
5) Poulsen, T. D.; Garcia-Viloca, M.; Gao, J. L.; Truhlar, D. G. J. Phys. Chem.
B 2003, 107, 9567-9578.
(6) Engst, S.; Vock, P.; Wang, M.; Kim, J. J. P.; Ghisla, S. Biochemistry 1999,
38, 257-267.
(
(
(
(
(7) Wu, J.; Bell, A. F.; Luo, L.; Stephens, A. W.; Stankovich, M. T.; Tonge,
P. J. Biochemistry 2003, 42, 11846-11856.
(8) Lenn, N. D.; Stankovich, M. T.; Liu, H. W. Biochemistry 1990, 29, 3709-
15.
8424
9
J. AM. CHEM. SOC. 2005, 127, 8424-8432
10.1021/ja050083p CCC: $30.25 © 2005 American Chemical Society

Ring Current Effects in the Active Site
A R T I C L E S
Scheme 2. Polarization of HD-CoA by MCAD
ground state of the substrate, since accumulation of partial
negative charge at the R-carbon will facilitate R-proton abstrac-
tion, while the partial positive charge at the â-carbon will favor
elimination of the hydride.
The structures of two ionizable substrate analogues that form
enolates upon binding to MCAD has also been examined by
1
3
both RR and C NMR spectroscopies. Nishina et al. reported
RR data for labeled and unlabeled acetoacetyl-CoA (AcAc-CoA)
bound to MCAD and concluded that the acetoacetyl enolate
was localized on the C1 carbon while the C3 carbon remained
reaction are dealt with by substantial changes in both the
R-proton acidity and redox potential of the substrate upon
binding to the enzyme. The nature of these binding-induced
changes in substrate reactivity is of central importance in MCAD
enzymology.
Elegant studies utilizing ionizable substrate analogues indicate
that enzyme-substrate interactions stabilize negative charge
2
3
a ketone (CdO). These studies were subsequently extended
using C NMR spectroscopy in which it was determined that
both the C1 and C3 carbons experienced significantly upfield
chemical shifts while C2 shifted downfield. Similar changes
in C1 and C2 chemical shifts were observed for the substrate
analogue 3-thiaoctanoyl-CoA.
1
3
2
4
accumulation on the thioester carbonyl thereby reducing the pKa
of the acyl-CoA R-protons.10,11 This stabilization is accom-
2
5
plished, in part, via hydrogen bonds from the enzyme and
cofactor to the thioester carbonyl oxygen.6
,9,12
While the above NMR data are consistent with the expectation
that the bound AcAc-CoA enolate is localized primarily on the
C1dO, it is curious why C3 also experiences a large upfield
chemical shift. The previous Raman studies on this ligand bound
to MCAD support a decrease in electron density at C3,23 which
would cause a deshielding of the carbon and hence a downfield
change in the C chemical shift. Since AcAc-CoA is an
ionizable probe of the active site, we have now extended our
own studies on the nonionizable product analogue HD-CoA to
include C NMR and C- H HSQC (heteronuclear single
quantum correlation) spectroscopy. While our Raman studies
on HD-CoA bound to MCAD indicate that the bound ligand is
polarized as shown in Scheme 2, such that there is a decrease
in electron density at C1 and C3, the NMR studies reveal
substantial upfield chemical shifts for both the C1 and C3 HD
carbons (-13 to -14 ppm) as well as the C3-H proton (-1.3
ppm), rather than the anticipated downfield change in chemical
shift. To investigate the origin of these effects, we have revisited
These interactions
also serve to shift the redox potential of the flavin positive by
over 100 mV, such that hydride transfer from substrate to flavin
is now thermodynamically favored.8
,13
In addition to these
interactions, desolvation of the active site upon substrate binding
is also thought to be important for raising the pKa of the catalytic
1
3
base,2,10,14,15
while Thorpe and colleagues have speculated that
charge-transfer interactions in the enzyme-substrate complex
1
6,17
may also play a role in transition state stabilization.
1
3
13
1
Further insights into the enzyme substrate interactions in the
MCAD active site have been gained from Raman and NMR
studies. Nishina et al.18 reported resonance Raman (RR) data
for the crotonyl-CoA product bound to MCAD and concluded
that the enzyme polarized the ground state of the ligand. We
have since extended these studies to include the product
analogue hexadienoyl-CoA19 (HD-CoA) as well as the ionizable
20
ligand 4-hydroxycinnamoyl-CoA. Using 752 nm excitation we
obtained Raman spectra of unlabeled and isotopically labeled
HD-CoA bound to oxidized and reduced pig MCAD (pMCAD)
and concluded that the enzyme polarized the C3dC2-C1dO
enone portion of the HD moiety by stabilizing a withdrawal of
negative charge from the â-carbon onto the carbonyl oxygen
1
3
the C NMR studies reported by Anderson and colleagues for
HD-CoA bound to enoyl-CoA hydratase and have also
obtained C NMR spectra of AcAc-CoA bound to enoyl-CoA
2
6
1
3
1
3
hydratase. Significantly, the C NMR data on HD-CoA bound
to enoyl-CoA hydratase are consistent with our previous Raman
studies which show that enoyl-CoA hydratase causes similar
polarization of the hexadienoyl group as described above for
MCAD in which there is a decrease in electron density at C1
and C3 (Scheme 2). We conclude that the substantial differences
in the C chemical shifts for the C3 and C1 carbons of the
ligands bound to the two enzymes arise as a result of the ring
current from the flavin cofactor in the MCAD active site. In
support of this conclusion, preliminary Density Functional
Theory (DFT) calculations on an isolated HD group in proximity
to an isoalloxazine ring analogue reproduce the direction
although not the full magnitude of the experimentally observed
changes in NMR chemical shift.
(Scheme 2).
The importance of hydrogen bonding interactions with the
substrate carbonyl in ligand polarization was subsequently
6
,7
demonstrated using 2′-deoxyFAD-substituted enzyme. The
changes in electronic structure in these analogues report on the
forces in place to stabilize the transition state(s) for substrate
1
3
oxidation.21,22
These interactions also potentially activate the
(
9) Trievel, R. C.; Wang, R.; Anderson, V. E.; Thorpe, C. Biochemistry 1995,
3
4, 8597-8605.
(
(
10) Rudik, I.; Ghisla, S.; Thorpe, C. Biochemistry 1998, 37, 8437-45.
11) Vock, P.; Engst, S.; Eder, M.; Ghisla, S. Biochemistry 1998, 37, 1848-
6
0.
(
(
(
(
12) Kim, J. J. P.; Wang, M.; Paschke, R. Proc. Natl. Acad. Sci. U.S.A. 1993,
0, 7523-7527.
13) Johnson, B. D.; Mancini-Samuelson, G. J.; Stankovich, M. T. Biochemistry
995, 34, 7047-7055.
14) Bach, R. D.; Thorpe, C.; Dmitrenko, O. J. Phys. Chem. B 2002, 106, 4325-
335.
15) Lamm, T. R.; Kohls, T. D.; Stankovich, M. T. Arch. Biochem. Biophys.
9
1
(21) Bell, A. F.; Feng, Y.; Hofstein, H. A.; Parikh, S.; Wu, J.; Rudolph, M. J.;
Kisker, C.; Whitty, A.; Tonge, P. J. Chem. Biol. 2002, 9, 1247-55.
(22) Cheng, H.; Nikolic-Hughes, I.; Wang, J. H.; Deng, H.; O’Brien, P. J.; Wu,
L.; Zhang, Z. Y.; Herschlag, D.; Callender, R. J. Am. Chem. Soc. 2002,
124, 11295-11306.
(23) Nishina, Y.; Sato, K.; Shiga, K.; Fujii, S.; Kuroda, K.; Miura, R. J. Biochem.
1992, 111, 699-706.
(24) Miura, R.; Nishina, Y.; Fujii, S.; Shiga, K. J. Biochem. 1996, 119, 512-
4
2
002, 404, 136-46.
(
16) Rudik, I.; Thorpe, C. Arch. Biochem. Biophys. 2001, 392, 341-8.
(
17) Dmitrenko, O.; Thorpe, C.; Bach, R. D. J. Phys. Chem. B 2003, 107,
1
3229-13236.
(
(
(
18) Nishina, Y.; Sato, K.; Hazekawa, I.; Shiga, K. J. Biochem. 1995, 117, 800-
8
08.
519.
19) 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.
20) Rudik, I.; Bell, A. F.; Tonge, P. J.; Thorpe, C. Biochemistry 2000, 39,
(25) Tamaoki, H.; Nishina, Y.; Shiga, K.; Miura, R. J. Biochem. 1999, 125,
285-296.
(26) D’Ordine, R. L.; Pawlak, J.; Bahnson, B. J.; Anderson, V. E. Biochemistry
2002, 41, 2630-2640.
9
2-101.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8425

A R T I C L E S
Wu et al.
Scheme 3. Synthesis of Isotopically Labeled AcAc-CoA
Enzymes. Recombinant pig kidney MCAD (pMCAD) was expressed
and purified as described previously. The concentration of pMCAD
7
was determined spectrophotometrically using an extinction coefficient
-
1
-1
31
of 15 400 M cm at 446 nm.
Bovine liver enoyl-CoA hydratase was expressed in BL21(DE3)-
pLysS cells following induction with 1 mM IPTG. The protein was
3
2
purified using ethanol precipitation as described previously. In most
cases an additional precipitation procedure was performed in order to
remove trace amounts of contaminating proteins. The concentration of
enoyl-CoA hydratase was determined by active site titration with
4
4
-dimethylaminocinnamoyl-CoA using an extinction coefficient of 48
-
1
-1
33
8 000 M cm at 496 nm.
Preparation of NMR Samples. For the 13C NMR experiments,
enzymes were prepared in 50 mM potassium phosphate, 0.3 mM EDTA,
pH 7.6 buffer containing 50% D O. Enzyme samples containing labeled
Experimental Procedures
Chemicals. Hexadienoic acid, 1,1′-carbonyldiimidazole, acetalde-
1
3
1
2
hyde, 1- C-acetyl chloride, coenzyme A (lithium salt; CoA ), L-3-
hydroxyacyl-CoA dehydrogenase, lactate dehydrogenase and pyruvate
HD-CoA were prepared by dissolving lyophilized HD-CoA directly in
the enzyme solution. Initially, one equiv of ligand was added followed
13
were purchased from Sigma Chemical Co. [1,3- C
C), and [2- C] malonic acid (99% 13C) were purchased from
2
] Malonic acid (99%
1
3
1
13
13
by subsequent additions of up to 6 equiv of ligand. For the C- H
HSQC NMR experiments, the pMCAD enzyme was exchanged into
Cambridge Isotope Labs. A plasmid carrying the cDNA for bovine
liver enoyl-CoA hydratase was a generous gift from Professor Hung-
wen Liu.
5
0 mM potassium phosphate buffer prepared from P
2 5
O dissolved in
3
4
9
9.7% D O and adjusted to pD 7.6 with anhydrous K
2
2
CO
3
.
The
concentration of pMCAD in all NMR experiments was in the range of
00 to 600 µM and that of enoyl-CoA hydratase was 1 to 2 mM. A
Isotopically Labeled Hexadienoyl-CoA and Acetoacetyl-CoA. The
5
1
3
synthesis and purification of C1, C2, C3, and C4 C-labeled hexadi-
total volume of 0.5 mL of enzyme was used in each sample.
13C NMR Spectroscopy. The 13C NMR spectra were obtained in 5
mm NMR tubes at the SUNY at Stony Brook NMR center on a Varian
Inova 500 MHz instrument operating at 125.711 MHz using a Nalorac
2
7
enoyl-CoA (HD-CoA) has been described elsewhere. C1, C2, and
1
3
C3 C-labeled acetoacetyl-CoA (AcAc-CoA) were synthesized as
1
3
shown in Scheme 3. Crotonic acid labeled with C at the C1, C2, or
2
7
C3 position was synthesized as described in Bell et al. (Scheme 3A
and B) and then converted to the corresponding 3C-labeled crotonyl-
1
gradient dual band probe. The T
carbonyl carbon of HD-CoA was determined to be 3.1 s using a standard
two-pulse sequence in an inversion-recovery T experiment. All spectra
1
(spin-lattice relaxation time) of the
2
8
CoA as described (Scheme 3C and D). The labeled crotonyl-CoAs
were then converted enzymatically to the respective AcAc-CoAs using
enoyl-CoA hydratase and L-3-hydroxylacyl-CoA dehydrogenase
1
of either pure enzymes or enzyme-ligand complexes, consisting of
25 000-30 000 scans for pMCAD or 15 000-20 000 scans for enoyl-
2
9,30
(
(
1
HAD).
HB-CoA) by enoyl-CoA hydratase (Scheme 3E). In a typical reaction,
0 µL of 0.1 mM enoyl-CoA hydratase was added to a 3 mM crotonyl-
Crotonyl-CoA was first converted to 3-hydroxybutyl-CoA
1
CoA hydratase, were acquired with broad band H decoupling (decou-
pler power 45 W), a relaxation delay of 1.0 s, and an acquisition time
of 1.3 s. These spectra were processed with 25 Hz of line broadening.
A flip angle of 20° was used in each case, except for the complex of
enzyme with 13C1-labeled ligand where a flip angle of 10° was used
to compensate for the relatively short relaxation delay of the carbonyl
carbon. Spectra of pMCAD or enoyl-CoA hydratase and their com-
plexes were obtained at 25 °C. The spectrum of pure enzyme was
obtained prior to the addition of ligand. Spectra of 13C-labeled ligands
in the same buffer as that used for the enzyme samples were recorded
separately using 500-600 transients, processed with 5 Hz of line
broadening and calibrated using TSP. Spectra of the enzyme-ligand
complexes were calibrated using chemical shift values of the unbound
ligand.
CoA solution in 5 mL of 50 mM potassium phosphate buffer, pH 7.6.
The conversion was followed by monitoring the decrease in absorbance
at 260 nm. No discernible absorbance change was observed after ca.
3
0 min, indicating that the majority of the crotonyl-CoA had been
hydrated to HB-CoA. The HB-CoA was then oxidized to AcAc-CoA
+
in situ by adding HAD and NAD to final concentrations of 0.1 µM
and 60 µM, respectively (Scheme 3F). The reaction was driven to
completion by the addition of lactate dehydrogenase (LDH) (70 unit)
and pyruvate (60 mM). After overnight incubation at RT, the resulting
AcAc-CoA was purified by reverse phase HPLC. Conditions were the
2
8
same as that used in the purification of crotonyl-CoA except that the
gradient of buffer B was from 0 to 20% in 60 min. AcAc-CoA eluted
1
13
1
H- C HSQC Spectroscopy. In addition to the direct detection
at 45 min. H NMR (500 MHz, D2O): δ 0.78 (s, 3H, 11′ CH3), 0.91
s, 3H, 10′ CH3), 2.32 (s, 3H, CH3Ac), 2.44 (t, 2H, H6′′), 2.65 (s, 2H,
of carbon chemical shift changes upon binding to pMCAD, indirect
(
1
3
detection was also used to measure the chemical shift of both the carbon
CH2Ac, split into 2 singlets at 2.50 and 2.80 ppm upon C labeling at
the C2 position), 3.05 (t, 2H, H9′′), 3.37 (t, 2H, H8′′), 3.46 (t, 2H,
H5′′), 3.58 (q, 1H, H1′′B), 3.86 (q, 1H, H1′′A), 4.04 (s, 1H, H3′′),
1
13
and the attached proton using H- C HSQC spectroscopy. Sensitivity-
enhanced gradient two-dimensional 1H- C HSQC spectra were
recorded at 25 °C. Spectral widths for the proton and carbon dimensions
were 8000 and 25 000 Hz, respectively. A total of 256 scans were
collected for each t increment and a total of 128 t increments were
13
35
4
.25 (s, 2H, H5′), 4.60 (s, 1H, H4′), 6.15 (d, 1H, H1′), 8.24 (s, 1H,
H2), 8.54 (s, 1H, H8). The CoA H2′ and H3′ resonances are obscured
13
by the solvent resonance. C NMR (125 MHz, D2O): δ 21.13 (C11′′),
1
1
acquired, each with 2048 points. Spectra were referenced to the
chemical shifts of free ligand in both carbon and proton dimensions.
The spectrum of pure enzyme was obtained prior to the addition of a
ligand. Subsequently, spectra were obtained following each addition
of one equiv of the lyophilized ligand. Spectra of 13C-labeled HD-
2
3.74 (C10′′), 30.76 (C9′′), 38.30 (C6′′), 41.15 (C2′′), 41.26 (C5′′),
41.60 (C8′′), 57.70 (C2Ac), 68.44 (C5′), 74.71 (C1′′), 76.79 (C2′), 77.03
(C3′ + C3′′), 86.64 (C4′), 89.34 (C1′), 121.51 (C5), 142.68 (C8), 152.18
(C4), 155.72 (C2), 158.46 (C6), 176.87 (C7′′), 177.61 (C4′′), 196.00
(C1Ac), 201.40 (C3Ac).
(
(
(
(
27) Bell, A. F.; Wu, J.; Feng, Y.; Tonge, P. J. Biochemistry 2001, 40, 1725-
3.
28) Hofstein, H. A.; Feng, Y.; Anderson, V. E.; Tonge, P. J. Biochemistry 1999,
(31) Thorpe, C.; Matthews, R. G.; Williams, C. H. J. Biochemistry 1979, 18,
331-7.
3
(32) Bahnson, B. J.; Anderson, V. E. Biochemistry 1989, 28, 4173-4181.
(33) D’Ordine, R. L.; Tonge, P. J.; Carey, P. R.; Anderson, V. E. Biochemistry
1994, 33, 12635-12643.
3
8, 9508-16.
29) Wu, W. J.; Feng, Y.; He, X.; Hofstein, H. S.; Raleigh, D. P.; Tonge, P. J.
J. Am. Chem. Soc. 2000, 122, 3987-3994.
(34) Ghisla, S.; Thorpe, C.; Massey, V. Biochemistry 1984, 23, 3154-61.
(35) Marion, D.; Ikura, M.; Tschudin, R.; Bax, A. J. Magn. Reson. 1989, 85,
393-399.
30) Feng, Y.; Hofstein, H. A.; Zwahlen, J.; Tonge, P. J. Biochemistry 2002,
4
1, 12883-90.
8426 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Ring Current Effects in the Active Site
A R T I C L E S
CoAs were recorded in the same buffer as that used for the enzyme
samples using 32 scans for each t increment and a total of 32 t
1 1
increments, each with 2048 points. The free ligand spectra were
calibrated with TSP and spectra of the enzyme-HD-CoA complexes
were calibrated using the chemical shift values observed for the free
ligand. Data were processed with Felix 98 (Insight II, Accelrys).
Density Functional Theory Calculations. C and 1H NMR
13
chemical shifts were calculated on a model system using Density
3
6
Functional Theory (DFT) and Gaussian 03. This method of gauging
the effect of ring currents on chemical shifts directly from DFT
3
7
calculations has been applied previously. We used the B3LYP
functional with a 6-31G(d,p) basis set for optimization and chemical
3
8
shift calculation and found along with others that the ‘Continuous
1
3
Set of Gauge Transformations’ (CSGT) produced C chemical shifts
more in agreement with experiment than ‘Gauge-Including Atomic
orbitals’ (GIAO). Initial calculations involved the HD-CoA model
compound trans,trans-2,4-hexadienoyl ethyl thioester (HET) and
utilized a benzene ring as a flavin analogue. The benzene was
constrained to be parallel to the HET molecule at a distance of 3 Å
26
from the C3 HD carbon. As is usual in active site calculations, the
geometry was optimized before calculating the C and 1H NMR
13
chemical shifts.
Results
1
3
C NMR Titration of pMCAD with 13C-Labeled HD-
CoAs. Figure 1 contains the 13C NMR spectra of pMCAD
titrated with C1, C2, C3, and C4 C-labeled HD-CoAs. The
3C NMR spectrum of pMCAD prior to the addition of ligand
13
1
is shown for each spectral region (i) and reveals signals that
arise from the protein. For example, in the region from 160 to
2
00 ppm (Ai), the predominant signals arise from the carboxy-
Figure 1. 13C NMR titration of pMCAD with 13C-labeled HD-CoAs. (A)
39-41
late and carbonyl resonances of the amino acids,
signal at 160.4 ppm can be assigned to the resonance of either
while the
1
3
Titration of pMCAD (540 µM) with (i-iv) 0, 2, 4 and 6 equiv of C1-
13
HD-CoA, (v) 0.2 mM C1-HD-CoA. (B) Titration of pMCAD (560 µM)
4
2
13
13
the arginine guanidino carbons or of the tyrosine Cê. In
addition, the strong signal around 130 ppm (B-D, i) arises from
the aromatic amino acids in the protein.42 As labeled HD-CoA
is titrated into the protein (spectra ii-iv), new signals can be
seen to grow in that can be assigned to the enzyme bound
population of HD-CoA (Table 1). These signals are at 181.4,
with (i-iv) 0, 1, 2, and 4 equiv of C2-HD-CoA, (v) 1 mM C2-HD-
CoA. (C) Titration of pMCAD (540 µM) with (i-iv) 0, 1, 2, and 4 equiv
13
13
of C3-HD-CoA, (v) 0.2 mM C3-HD-CoA. (D) Titration of pMCAD
540 µM) with (i-iv) 0, 1, 2, and 4 equiv of C4-HD-CoA, (v) 2 mM
13
(
13
C4-HD-CoA in buffer.
13
1
Table 1.
C and H Chemical Shift Values for HD-CoA Free and
Bound to pMCAD
24.5, and 129.5 ppm for the 13C1-, C2-, and C3-labeled
13
13
1
freea
bound
HD-CoAs, respectively. Furthermore, as the titration continues,
a second resonance can be observed in each case that has the
same chemical shift as the unbound ligand (spectrum v). The
signals for unbound HD-CoA appear at 193.9 (A: C1), 126.8
c
δ
(ppm)
∆ν
(Hz)
1
/
2
δ
(ppm)
∆ν
(Hz)
1
/
2
∆δ, bound − free
b
HD-CoA
(ppm)
1
3
3
3
3
3
3
C1d
193.9 (186.3)
126.8 (127.3)
143.3 (139.5)
133.4 (131.0)
6.08 (5.84)
29 (2)
30 (5)
40 (5)
15 (5)
181.4
124.5
129.5
133.4
6.18
16
38
21
15
-12.5
-2.3
-13.8
0.0
0.1
-1.3
-0.06
1
1
1
1
1
d
C2
C3
d
(
36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa,
J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.;
Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo,
J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;
Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels,
A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A., Gaussian 03,
ReVision C.02. Gaussian Inc.: Wallingford CT, 2004.
d
C4
C2-He
ND
ND
f
ND
ND
f
e
f
f
C3-H
7.14 (7.35)
6.18 (6.55)
5.83
6.12
13C4-He
NDf
NDf
a
Values are for unbound ligand in the presence of pMCAD. b Values in
parentheses are calculated chemical shifts based on substrate analogue;
13
trans,trans-2,4-hexadienoyl ethyl thioester. C calculated using CSGT and
1
c
H using GIAO (see Experimental Procedures)d. Values in parentheses are
13
for the free ligand in the absence of protein.
from the direct detection experiments. H NMR chemical shifts from the
HSQC experiments. ND, not determined.
C NMR chemical shifts
e
1
f
(
37) Boyd, J.; Skrynnikov, N. R. J. Am. Chem. Soc. 2002, 124, 1832-1833.
(B: C2) and 143.3 (C: C3) ppm, and thus binding to the enzyme
has caused a large upfield shift in the resonances arising from
the HD-CoA C1 (12.5 ppm) and C3 (13.8 ppm) carbons, and a
smaller upfield shift (2.3 ppm) in the C2 HD-CoA resonance
(Table 1). In contrast to the C1-C3 carbons, the chemical shift
of the C4 HD-CoA carbon (133.4 ppm) remains unperturbed
upon binding of the ligand to pMCAD (Figure 1D).
(
38) Kim, D. H.; Eun, H. M.; Choi, H. Bull. Korean. Chem. Soc. 2000, 21,
1
48-150.
(
39) Richarz, R.; Wuthrich, K. Biochemistry 1978, 17, 2263-2269.
(
40) Shindo, H.; Cohen, J. S. Proc. Nat. Acad. Sci. U.S.A. 1976, 73, 1979-
1
983.
(
41) Shindo, H.; Cohen, J. S.; Rupley, J. A. Biochemistry 1977, 16, 3879-
882.
42) Allerhand, A.; Norton, R. S.; Childers, R. F. J. Biol. Chem. 1977, 252,
786-1794.
3
(
1
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8427

A R T I C L E S
Wu et al.
1
3
1
13
1
3
1
13
Figure 3. C- H HSQC spectra of C3-HD-CoA free and bound to
13
Figure 2. C- H HSQC spectra of C2-HD-CoA free and bound to
13
pMCAD. pMCAD (410 µM) with (A-C) 1, 2, and 4 equiv of C3-HD-
13 1
pMCAD. pMCAD (410 µM) with (A-C) 0.8, 1.6, and 3 equiv of C2-
HD-CoA. (D) 2 mM C2-HD-CoA in buffer. Cross-peak I ( H: 6.18 ppm,
C: 124.3 ppm) is from bound ligand while cross-peak II ( H: 6.08 ppm,
C: 126.4 ppm) arises from free ligand.
13
1
CoA. (D) 2 mM C2-HD-CoA in buffer. Cross-peak I ( H: 5.83 ppm,
1
1
3
3
1
1
1
3
3
1
C: 129.5 ppm) is from bound ligand while cross-peak II ( H: 7.14 ppm,
C: 143.3 ppm) arises from free ligand.
The observation of 13C resonances arising from both the free
ligand is added (Figure 2B and C), indicating that it arises from
C2 of the bound ligand. Observation of cross-peak II after the
addition of 1.6 equiv of ligand (Figure 2B) indicates that peak
I is from the bound ligand while peak II arises from the free
ligand. The 2.3 ppm difference between these two cross-peaks
in the carbon dimension is identical to that observed for bound
and bound ligand (Figure 1A-C) indicates that the two forms
4
3
of HD-CoA are in slow exchange on the NMR time scale. In
addition, as expected the resonances of both free and bound
HD-CoA are broader in the presence of protein compared to
the signals arising from ligand in buffer alone (Table 1).
Interestingly, the line width (∆ν1/2, defined as the width of the
peak at half-height in Hz) for the C1 and C3 resonances are
actually narrower (13 and 19 Hz, respectively) when HD-CoA
is bound to pMCAD than when the ligand is free in the protein
solution (Table 1). Finally, for 13C4-labeled HD-CoA (Figure
1
3
and free C2-HD-CoA in the direct detection experiment (see
above, Table 1). In addition, the HD-CoA C2-H proton is
deshielded 0.1 ppm on binding to the enzyme, as revealed by
the change in this resonance from 6.08 ppm in the free ligand
to 6.18 ppm in the bound ligand (Table 1).
Similar to the HSQC experiment with 3C2-HD-CoA, titration
of pMCAD with 1 equiv of 13C3-HD-CoA gave one cross-peak
at 129.5 ppm in the carbon dimension and 5.83 ppm in the
proton dimension arising from the bound ligand (Figure 3A).
Addition of further ligand resulted in the bound ligand peak
intensifying and a second cross-peak appearing at 143.3 ppm
in the carbon dimension and 7.14 ppm in the proton dimension
(Figure 3B-D). The 13.8 ppm upfield shift in the 13C3 ligand
resonance on binding is in agreement with the results of the
direct detection experiment (see above). In addition, the HSQC
1
1
D), where no change in chemical shift is observed upon
binding, the resonance arising from the bound ligand population
is 10 Hz broader than that for the ligand in the absence of protein
(
Table 1). The line broadening for unbound HD-CoA in the
presence of pMCAD results from chemical exchange between
the free and bound HD-CoA populations. As noted by Anderson
and co-workers,26 this difference in line width can be used to
calculate a dissociation rate constant (k-1) using eq 1:
π∆ν1/2,obsd ) π∆ν1/2,free + k-1
(1)
1
3
spectrum of the C3-HD-CoA also reveals a 1.31 ppm upfield
where ∆ν1/2,obsd and ∆ν1/2,free are the line widths of the free
ligand in the presence and absence of protein, respectively. For
1
shift in the H resonance of the C3-H proton upon binding to
pMCAD (Table 1).
Titration of pMCAD with 13C4-HD-CoA was also monitored
1
3
13
13
C1-, C2-, and C3-labeled HD-CoAs, line broadening of
-1
2
7, 25, and 35 Hz provides k-1 estimates of 80-110 s . Using
a Kd value of 3.5 µM, this gives an association rate constant
by HSQC spectroscopy (data not shown). Consistent with the
19
13
13
C NMR direct detection experiment, the C4 chemical shift
of HD-CoA did not change upon binding to the enzyme (Table
). However, careful scrutiny of spectra obtained from pMCAD
7
-1 -1
for HD-CoA of 2.3-3.1 × 10 M s .
1
13
H- C HSQC Spectroscopy. In addition to the direct
1
13
detection C NMR experiments, the HD-CoA/pMCAD complex
has also been characterized using H- C HSQC spectroscopy
samples containing excess ligand revealed two sets of cross-
1
13
1
peaks that differed by about 0.06 ppm in the H dimension (6.18
in order to extend our analysis to the protons attached to the
ppm free; 6.12 ppm bound). Thus, while binding does not
1
3
C-labeled carbons. Figures 2 and 3 contain the NMR data for
13
perturb the C4 chemical shift, the attached proton experiences
1
3
13
the titration of pMCAD with C2- and C3-HD-CoA, respec-
a slight upfield change in chemical shift.
1
3
tively. For C2-HD-CoA cross-peak I arising at low ligand
concentration (Figure 2A) is a doublet with a coupling constant
13C NMR Titration of Enoyl-CoA Hydratase with 13
C-
Labeled HD-CoAs. Recombinant bovine liver enoyl-CoA
(
J) of 14 Hz. The intensity of this cross-peak increases as further
13
13
13
hydratase was titrated with C1-, C2-, and C3-labeled HD-
CoA. The results obtained are in good agreement with the data
(
43) Jardetzky, O.; Roberts, G. C. K. NMR in Molecular Biology; Academic
2
6
Press: New York, 1981.
reported previously, and consequently the NMR spectra are
8428 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Ring Current Effects in the Active Site
A R T I C L E S
Table 3. 13C Chemical Shift Values for AcAc-CoA Free and
Bound to Enoyl-CoA Hydratase
free
bound
AcAc-CoA
δ
(ppm)
∆ν
1
/
2
(Hz)
δ
(ppm)
∆ν
1/
2
(Hz)
∆δ (ppm)
1
1
1
3
3
3
C1
C2
C3
196.0
57.7
201.4
20
ND
20
185.5
92.6
210.4
40
40
15
-10.5
+35.1
+9.0
a
a
ND, not determined. The 13C2 resonance of the unbound ligand overlaps
with the protein aliphatic resonances.
indicates that the two forms of AcAc-CoA are in slow exchange
4
3
13
13
13
on the NMR time scale. For C1, C2-, and C3-labeled
AcAc-CoAs the chemical shift of the bound ligand population
is observed at 185.5, 92.6, and 210.4 ppm, respectively, while
the corresponding resonances for free ligand appear at 196.0,
57.7, and 201.4 ppm, respectively (Figure 4 and Table 3). Thus,
binding to enoyl-CoA hydratase causes the C1 AcAc-CoA
resonance to move upfield by 10.5 ppm, while the C2 and C3
resonances move downfield by 35.1 and 9.0 ppm, respectively.
Density Functional Calculations. In an effort to try and
predict how the ring current of FAD may affect the chemical
1
3
1
shift of the substrate, we calculated C and H NMR chemical
shifts on a model system using Density Functional Theory
(
DFT). The calculated chemical shifts of the isolated model
compound trans,trans-2,4-hexadienoyl ethyl thioester (HET) are
given in Table 1. Small differences between the values
calculated for HET and the experimentally observed chemical
shifts for the free HD-CoA ligand likely arise from interactions
of the HD group with solvent water. As a first approximation
of a ‘bound’ model, we used a benzene ring as a flavin analogue
and constrained it to be parallel with HET at a distance of 3 Å
from the HD C3 carbon. The differences in the chemical shifts
(bound - free) were -0.4, +1.3, -3.8, -1.2, and +3.6 for
C1, C2, C3, H3, and C4, respectively. The direction of these
chemical shift changes are in agreement with experiment (except
for C4, the chemical shift of which is not observed to change
when bound to MCAD). We subsequently found that the degree
and direction of shift was highly dependent on the position of
the substrate relative to the aromatic ring. For example, by
moving the benzene ring 0.5 Å closer to C3, the chemical shift
differences became -2.6, +5.0, and -12.2 for C1, C2, and C3,
respectively. These results suggest that the predicted chemical
shifts are very sensitive to protein environment and that a
geometrically accurate model of the experimental system is
required. In an attempt to improve our model, we replaced the
benzene ring with the more structurally relevant flavin analogue,
methylisoalloxazine, and used the geometry in Figure 5 as initial
coordinates for optimization. However, we were still unable to
predict successfully both the direction and magnitude of all the
observed chemical shifts.
Figure 4. 13C NMR titration of enoyl-CoA hydratase with 13C-labeled
AcAc-CoAs. (A) Titration of enoyl-CoA hydratase (1.3 mM) with (i-iv)
1
3
13
0
, 1, 2, and 4 equiv of C1-AcAc-CoA, (v) 2 mM C1-AcAc-CoA in
D2O buffer. (B) Titration of enoyl-CoA hydratase (1.2 mM) with (i-iii) 0,
1
3
13
1
.2, and 3.6 equiv of C2-AcAc-CoA, (iv) 3 mM C1-AcAc-CoA in
buffer. (C) Titration of enoyl-CoA hydratase (1.2 mM) with (i-v) 0, 1, 2,
1
3
4
, and 6 equiv of C3-AcAc-CoA, (vi) Sample from (v) after removing
enzyme. /, acetate contamination in the AcAc-CoA sample.
Table 2. 13C Chemical Shift Values for HD-CoA Free and Bound
to Enoyl-CoA Hydratase
free
bound
HD-CoA
δ
(ppm)
∆ν
1
/
2
(Hz)
δ
(ppm)
∆ν
1
/
2
(Hz)
∆δ (ppm)
1
1
1
3
3
3
C1
C2
C3
193.9
126.8
142.8
20
40
20
198.7
124.6
148.4
40
40
40
+4.8
-2.2
+5.6
not presented here. In each case, titration of the enzyme with
ligand results in the initial appearance of a broad resonance that
gradually shifts toward the position observed for the unbound
ligand as more HD-CoA is added. For 13C1-, C2-, and C3-
labeled HD-CoAs the chemical shift of the bound ligand
population is observed at 198.7, 124.6, and 148.4 ppm,
respectively (Table 2). Thus, binding to enoyl-CoA hydratase
causes the C1 and C3 HD-CoA resonances to move downfield
by 4.8 and 5.6 ppm, respectively, while the C2 resonance moves
upfield by 2.2 ppm. The chemical shift changes for the C1 and
C2-labeled HD-CoAs are close to the values of 4.3 and -2.2
13
13
Discussion
Spectroscopic Probes of Enzyme Mechanisms. The chemi-
cal transformation of a substrate into a product involves
electronic rearrangement in the substrate during the reaction,
which necessarily involves alterations in charge density on the
reactant’s functional groups. A critical aspect of an enzyme’s
catalytic power stems from the ability of the active site to cause
and stabilize this charge rearrangement. Since the seminal
2
6
ppm reported by D’Ordine et al.
1
3
13
C NMR Titration of Enoyl-CoA Hydratase with C-
Labeled Acetoacetyl-CoAs. Figure 4 contains the 13C NMR
spectra of enoyl-CoA hydratase titrated with C1-, C2-, and
C3-labeled acetoacetyl-CoAs (AcAc-CoAs). The observation
of separate resonances for free and bound ligand in each case
13
13
1
3
44
studies of Knowles and colleagues, vibrational and NMR
spectroscopies have been used extensively to probe the structure
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8429

A R T I C L E S
Wu et al.
analogue for the reaction catalyzed by enoyl-CoA hydratase,
to probe the active sites of both enzyme. Using Raman
spectroscopy coupled with isotope labeling, normal mode
calculations, model studies, and site-directed mutagenesis, we
have developed a detailed picture of the alterations in structure
of the hexadienoyl group on binding to both enzymes.7
,19,21,27
Raman spectra of HD-CoA are dominated by bands associated
with the -CdC-CdC- double bond stretching vibrations of
the hexadienoyl moiety. These modes undergo shifts to lower
wavenumber when HD-CoA binds to MCAD or enoyl-CoA
hydratase due to altered vibrational coupling between the Cd
C stretching coordinates and the thioester CdO stretching mode.
The alteration in coupling occurs primarily due to a decrease
in wavenumber of the CdO mode caused by hydrogen bonding
interactions between the enzyme and the HD carbonyl group.
Both MCAD and enoyl-CoA hydratase provide two hydrogen
bond donors to the substrate/HD CdO that play critical roles
in catalysis by stabilizing the accumulation of negative charge
on the substrate carbonyl in the carbanionic (like) transition state
Figure 5. Active sites of MCAD and enoyl-CoA hydratase. (A) Structure
of MCAD complexed with hexadienoyl-CoA (HD-CoA). The HD-CoA was
modeled into the active site using the structure of 3-thiaoctanoyl-CoA bound
to MCAD (1UDY.pdb).51 The ligand is in close proximity to the isoallox-
azine ring of the flavin (FAD) and is hydrogen bonded to the flavin’s 2′-
ribityl hydroxyl group and the backbone NH of the active site base E376.
(
B) Structure of HD-CoA bound to enoyl-CoA hydratase from 1MJ3.pdb.21
The HD-CoA carbonyl group is shown hydrogen bonded to the backbone
NH groups of A98 and G141. Also shown are the two catalytic glutamates
(
(
E144 and E164) as well as the water molecule that hydrates the substrate
WAT).
1
2,21
for the reactions catalyzed by both enzymes.
In the ground
of ligands bound to enzymes.45,46 In many cases these methods
state, these interactions polarize the ligand such that there is a
small increase in electron density on the CdO oxygen and C2
carbon and a decrease in electron density at C1 and C3 (Scheme
2). The changes in electron density at C2 and C3 will promote
the reaction catalyzed by MCAD by facilitating hydride transfer
from C3 and proton abstraction from C2.
reveal subtle changes in the electronic structure of the ligand
upon binding.4
4,47,48
These perturbations in ground-state structure
are a consequence of placing the ligand in an environment that
2
1,22
has evolved to stabilize the transition state for the reaction.
Consequently, the distortions in structure reveal interactions in
place to stabilize the transition state for the reaction. In addition,
the ground state structural changes may also directly contribute
to the catalytic process by destabilizing the reactant ground
state.2 In this paper, we extend our spectroscopic studies into
the mechanism of the reaction catalyzed by medium chain acyl-
The structure of HD-CoA bound to enoyl-CoA hydratase has
13
also previously been studied using C NMR spectroscopy by
26
Anderson and co-workers. These studies reveal a downfield
1,49
shift of the C1 resonance on binding to enoyl-CoA hydratase
and an upfield shift at C2, consistent with the Raman studies
in which electron density is reduced at C1 and C3, and increased
at C2 (Table 2). We have now supplemented these data by
observing that the C3 HD-CoA resonance is also shifted
downfield on binding, as expected from previous NMR data
1
3
1
CoA dehydrogenase (MCAD) by reporting C NMR and H-
1
3
C HSQC spectroscopy for the product analogue HD-CoA
bound to MCAD.
Previous Studies on the Interaction of HD-CoA with
MCAD and Enoyl-CoA Hydratase. MCAD and enoyl-CoA
hydratase, the first and second enzymes in the fatty acid
â-oxidation pathway, share a number of similarities in their
active site architecture (Figure 5). Both enzymes provide two
hydrogen bonds to the acyl-CoA thioester CdO group in order
to stabilize the carbanionic (like) transition states in each
reaction. These hydrogen bonds are provided by the backbone
amide of E376 and the 2′ ribityl hydroxyl of the flavin in
26
on other R,â-unsaturated acyl-CoAs and from the Raman
studies. On the basis of computational studies, Anderson and
co-workers estimated that the ground-state strain revealed by
26
the ligand polarization contributed 3.2 kcal/mol to catalysis
while our Raman studies suggest that up to 7 kcal/mol may be
27
available for the enoyl-CoA hydratase reaction. Analogous
studies on MCAD suggest similar energetic consequences for
the hydrogen bonding interactions in this enzyme. Specifically,
the use of 2-deoxy FAD-reconstituted enzyme, where one of
the oxyanion hydrogen bond donors has been removed, indicates
that this hydrogen bond may be worth up to 4 kcal/mol in
12
MCAD, and the backbone amides of A98 and G141 in enoyl-
CoA hydratase.50 In addition, both enzymes have an active site
acid/base to catalyze protonation/deprotonation of the substrate’s
R-carbon. In MCAD this residue is E376, in enoyl-CoA
hydratase it is E164. From this point the active sites diverge.
In enoyl-CoA hydratase there is a second Glu (E144) that is
involved in addition of an OH group to the substrate’s â-carbon,
while MCAD possesses a flavin to accept a hydride from the
â-carbon. We have used hexadienoyl-CoA (HD-CoA), a product
analogue for the reaction catalyzed by MCAD and a substrate
3
ground-state destabilization or a rate increase of 10 out of a
11
7
total rate increase >10 for MCAD.
13C NMR Spectroscopy of HD-CoA Bound to MCAD. On
the basis of the above similarities in the structure of HD-CoA
bound to both MCAD and enoyl-CoA hydratase revealed by
Raman spectroscopy, we anticipated similar correspondence
13
between the C NMR data for HD-CoA bound to both enzymes.
While the change in chemical shift of the C2 HD carbon (-2.3
ppm) on binding to MCAD is similar to that observed for enoyl-
CoA hydratase (-2.2 ppm), interaction with MCAD causes the
C1 and C3 HD carbon chemical shifts to move upfield by 12-
14 ppm (Table 1) in contrast to the ∼5 ppm downfield shift
observed for the binding of HD-CoA to enoyl-CoA hydratase
(
(
(
(
44) Belasco, J. G.; Knowles, J. R. Biochemistry 1980, 19, 472-477.
45) Carey, P. R.; Tonge, P. J. Acc. Chem. Res. 1995, 28, 8-13.
46) Deng, H.; Callender, R. Methods Enzymol. 1999, 308, 176-201.
47) Austin, J. C.; Kuliopulos, A.; Mildvan, A. S.; Spiro, T. G. Protein Science
1
992, 1, 259-270.
(
48) Dong, J.; Xiang, H.; Luo, L.; Dunaway-Mariano, D.; Carey, P. R.
Biochemistry 1999, 38, 4198-4206.
(
49) Jencks, W. P. AdV. Enzymol. Relat. Areas Mol. Biol. 1975, 43, 219-410.
50) Engel, C. K.; Mathieu, M.; Zeelen, J. P.; Hiltunen, J. K.; Wierenga, R. K.
EMBO J. 1996, 15, 5135-5145.
(
1
(Table 2). In addition, the H chemical shift of the C3-H is also
8430 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Ring Current Effects in the Active Site
A R T I C L E S
Table 4. 13C Chemical Shift Values for AcAc-CoA and
moved upfield by 1.3 ppm on binding to MCAD. If we assume
that HD-CoA binds to MCAD in an analogous fashion to that
observed for octanoyl-CoA12 or 3-thiaoctanoyl-CoA51 (Figure
3-Thiaoctanoyl-CoA Free and Bound to MCAD
free
bound
(ppm) ∆ν
c
c
5
), then the C1 and C3 HD carbons will be in close proximity
ligand
δ
(ppm) ∆ν
1
/
2
δ
1
/
2
∆δ (ppm)
to the isoalloxazine ring. Consequently, we believe that the
changes in 13C and H NMR chemical shifts observed for HD-
CoA in complex with MCAD are a result of the isoalloxazine
ring current. Importantly, these ring current effects only occur
in the presence of an external magnetic field (i.e. during the
NMR experiment but not during Raman data collection).
Ring current effects on H and 13C NMR chemical shifts in
proteins arising from aromatic amino acids have been exten-
AcAc-CoAa
13
13
C1 198.5
C2 59.9
b
181.3
103.4
192.3
163.7
101.2
n
-17.2
+43.5
-16.5
-39.6
+56.9
1
n
b
b
n
b
n
n
b
1
3
C3 208.8
C1 203.3
-thiaoctanoyl-CoAb
13
13
3
C2
44.3
a
Data from Miura et al.24 b Data from Tamaoki et al.25 c The line width
was not reported. However, the line shapes of the free and bound species
were compared and denoted as “b” if the band was relatively broad and
“n” if the band were narrower.
1
5
2-59
sively studied both experimentally and computationally.
Application of an external magnetic field causes the delocalized
electrons in the aromatic ring to circulate which in turn results
in a local magnetic field from the ring. The local field is
antiparallel to the applied field at the center of the ring and
parallel to the applied field outside the ring. Consequently, nuclei
occupying positions above or below the plane of the aromatic
ring will experience a local field opposing the applied field and
hence undergo an upfield change in chemical shift. In contrast,
nuclei in the plane of the aromatic ring where the local field
reinforces the external field will experience a downfield change
in chemical shift.58 In lysozyme, the ring current effect induced
a 3 ppm downfield change in the chemical shift of the M105
â-H due to deshielding of this nucleus by the aromatic amino
acids Y23, W108, and W111.56 Ring current effects due to
enzyme-bound aromatic cofactors such as porphyrins also cause
large effects on the chemical shifts of nearby residues. In the
light harvesting complex 2 from Rhodopseudomonas acidophila,
Hδ and Hγ in the His residues that coordinates the porphyrin
Mg (â-H30 and R-H31) are shifted upfield 3.50 and 3.70 ppm,
respectively.60 Additionally, in cytochrome c2 from Rhodospir-
illum rubrum, M91 is one of the heme axial ligands and is
therefore situated directly under the heme ring. As a result,
resonances of the Câ, Cγ, and Cꢀ of Met-91 are shifted upfield
by 5.7, 3.3, and 1.6 ppm, respectively, from their random coil
values. Of particular relevance to this work, Giessner-Prettre
and Pullman59 calculated the magnitude of the chemical shift
brought about by a flavin ring at a distance of 3.4 Å from an
intermolecular proton. Their work showed that at this distance,
the chemical shift of the proton could be changed by as much
as 1.2 ppm.
Returning to our studies of HD-CoA bound to MCAD, the
structural data (Figure 5) indicate that the HD moiety is below
the flavin ring and thus occupies a position where a ring current
would cause an upfield change in chemical shift, consistent with
the NMR data for the bound ligand. However, the ring current
shifts we observe due to the flavin ring are significantly larger
than any that have so far been reported. Indeed, if we assume
that the C1 and C3 HD carbons would have chemical shifts
similar to those observed for HD-CoA bound to enoyl-CoA
hydratase in the absence of the applied magnetic field, then the
ring current from the isoalloxazine ring causes an upfield shift
of around 18 ppm for the C1 and C3 carbons. These large ring
current effects are undoubtedly a consequence of the close
proximity of the ligand and flavin. Attempts to reproduce the
experimental changes in NMR chemical shift using DFT
calculations on a simple model system established that (1) large
chemical shifts, of the order and direction observed experimen-
tally, can be predicted and that (2), the NMR spectrum depends
highly on the model geometry. However, closer agreement
between theory and experiment will undoubtedly require
methods that allow inclusion of a greater portion of the protein,
such as would be possible using a hybrid QM/MM approach.
Charge transfer bands have been observed in many MCAD-
ligand complexes and it has been argued that these bands are
evidence not only for electronic interaction between flavin and
ligand but also have catalytic consequence by stabilizing the
1
7,51
transition state(s) for the reaction.
This interaction between
ligand and flavin is exquisitely sensitive to subtle perturbations.
For example, Thorpe and co-workers have noted that 5-deaza
flavin-reconstituted MCAD is incapable of catalyzing R-proton
61
16
exchange in a bound substrate. On the basis of this information,
1
3
we can now reexamine previous C NMR studies on ionizable
ligands bound to MCAD.
AcAc-CoA and 3-Thiaoctanoyl-CoA Bound to MCAD.
1
3
Binding to MCAD causes the C NMR chemical shifts of the
C1 and C3 AcAc-CoA carbons to move upfield by around 17
ppm, while the C2 carbon resonance moves downfield by over
2
4
40 ppm. Similar changes are observed for the binding of
3
5
-thiaoctanoyl-CoA where binding causes changes of -40 and
7 ppm for the C1 and C2 carbons, respectively (Table 4).
(
51) Satoh, A.; Nakajima, Y.; Miyahara, I.; Hirotsu, K.; Tanaka, T.; Nishina,
25
Y.; Shiga, K.; Tamaoki, H.; Setoyama, C.; Miura, R. J. Biochem. 2003,
1
34, 297-304.
These alterations in chemical shift have been ascribed to proton
abstraction from C2 with the formation of an enzyme-stabilized
enolate. However, while enolate formation is undoubtedly
responsible for some of the chemical shift changes observed,
we believe that the flavin ring current is also playing a key
role. Specifically, for AcAc-CoA two enolate species can be
formed (Scheme 4). In both cases the C2 carbon undergoes a
(52) Perkins, S. J.; Johnson, L. N.; Phillips, D. C.; Dwek, R. A. FEBS Lett.
1
977, 82, 17-22.
(
53) Perkins, S. J.; Dower, S. K.; Gettins, P.; Wainhobson, S.; Dwek, R. A.
Biochem. J. 1977, 165, 223-225.
(54) Perkins, S. J.; Wuthrich, K. Biochim. Biophys. Acta 1979, 576, 409-23.
(55) Perkins, S. J.; Wuthrich, K. J. Mol. Biol. 1980, 138, 43-64.
(56) Perkins, S. J.; Dwek, R. A. Biochemistry 1980, 19, 245-258.
(57) Perkins, S. J.; Johnson, L. N.; Phillips, D. C.; Dwek, R. A. Biochem. J.
1
981, 193, 553-572.
(
58) Perkins, S. J. In Biological Magnetic Resonance; Berliner, L. J., Reuben,
J., Eds.; Plenum Press: New York, 1982; Vol. 4, pp 193-336.
59) Giessner-Prettre, C.; Pulman, B. J. Theor. Biol. 1970, 27, 87-95.
60) Alia; Matysik, J.; de Boer, I.; Gast, P.; van Gorkom, H. J.; de Groot, H. J.
J. Biomol. NMR 2004, 28, 157-64.
3
2
change in hybridization from sp to sp , in line with the large
downfield change in C chemical shift for the C2 carbon.
(
(
13
24,62
(
61) Blanchard, L.; Hunter, C. N.; Williamson, M. P. J. Biomol. NMR 1997, 9,
(62) Breitmaier, E.; Voelter, W. Carbon-13 NMR Spectroscopy; VCH: New
3
89-95.
York, 1987; p 232.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8431

A R T I C L E S
Wu et al.
Scheme 4. AcAc-CoA Enolate Tautomers
unbound ligand (Table 1). This binding-induced decrease in line
width is also observed for the H signal arising from the proton
1
attached to C3. Transformation of the first FID block from the
13
HSQC experiment of C3-HD-CoA bound to MCAD gives the
C3-H proton spectrum of HD-CoA bound to MCAD (data not
shown) where it can be seen that the peak arising from the bound
ligand at 5.83 ppm is narrower (12 Hz) than that for the unbound
ligand at 7.14 ppm (24 Hz). This decrease in line width is not
observed for the C2 HD-CoA carbon, which experiences a
similar change in chemical shift in complexes with both MCAD
and enoyl-CoA hydratase (-2 ppm). Similar trends in line
widths are also observed when AcAc-CoA and 3-thiaoctanoyl-
CoA are bound to MCAD (Table 3) and thus it appears that
only those nuclei in the ligand that experience the flavin ring
current in MCAD show a decrease in line width upon binding.
This suggests that an increase in T2 accompanies the upfield
change in chemical shift caused by the flavin’s isoalloxazine
ring.
However, on the basis of the fact that the enzyme active site is
designed to stabilize charge accumulation on the carbonyl
oxygen, we believe that enol form I will be preferentially present
in the AcAc-CoA/MCAD complex.
To provide a reference point for this discussion, we thus
13
examined the C NMR spectrum of AcAc-CoA bound to enoyl-
CoA hydratase. AcAc-CoA is a µM inhibitor and binds to enoyl-
CoA hydratase as the enolate.6 Our NMR data show chemical
shift changes of -10.5, 35.1, and 9 ppm for the AcAc-CoA
C1, C2, and C3 resonances, respectively, upon binding to enoyl-
CoA hydratase (Table 3). Thus, the C1 and C2 AcAc carbons
experience changes in chemical shift that are in the same
direction, albeit smaller, than those observed in MCAD.
However, the C3 AcAc carbon moves downfield 9 ppm upon
binding to enoyl-CoA hydratase in contrast to the -16.5 ppm
change upon binding to MCAD. On the basis of these data, we
believe that both MCAD and enoyl-CoA hydratase bind AcAc-
CoA as enol form I (Scheme 4) and that the upfield change in
the C3 resonance in the MCAD complex is a result of the flavin
ring current. This conclusion is supported by data on ethyl
acetoacetate where keto to enol(ate) tautomerism causes an
upfield change in chemical shift of the carbonyl carbon involved
in the tautomerism, but a downfield change in chemical shift
3,64
Conclusion
13
1
C and H NMR data have been obtained for HD-CoA bound
to MCAD and enoyl-CoA hydratase. Binding causes a large
upfield change of -13 to -14 ppm in the chemical shift for
the C1 and C3 HD carbons upon binding to MCAD whereas
smaller downfield chemical shift changes of 4 to 5 ppm are
observed when HD-CoA is bound to enoyl-CoA hydratase.
These data are in contrast to the picture provided by Raman
spectroscopy which indicates that the electronic structure of the
hexadienoyl group is similar when HD-CoA is bound to both
enzymes. The discrepancy between the NMR and Raman data
for ligands bound to MCAD is rationalized by proposing that
the isoalloxazine ring of the MCAD flavin exerts a strong ring
current effect on the C1 and C3 HD carbons. The local field
produced by the isoalloxazine ring occurs only in the presence
of an external magnetic field and causes the C1 and C3 HD
resonances to move upfield by as much as 18 ppm. DFT
calculations on a simple model system reproduce the direction
though not the full magnitude of the experimentally observed
changes in chemical shift and pave the way for a more detailed
theoretical analysis using hybrid QM/MM methodology. Our
2
4,62
of the second carbonyl carbon.
Finally, the fact that the
change in 13C1 chemical shift is larger in MCAD than in enoyl-
CoA hydratase (-17.2 vs -10.5 ppm) suggests that the flavin
ring current is also partly responsible for the chemical shift of
this carbon in the MCAD complex.
Ring Current Effects on Line Width. From Table 2 we
can see that C1 and C3 HD-CoA resonances are broader when
HD-CoA is bound to enoyl-CoA hydratase than the signals
arising from the unbound ligand, reflecting the decreased
transverse relaxation (T2) delays as a result of the decreased
tumbling rate of the bound ligand. In contrast, when HD-CoA
binds to MCAD, the bands arising from the C1 and C3 HD
carbons are narrower than the corresponding bands from the
1
3
conclusions are substantiated by C NMR data for complexes
of AcAc-CoA with MCAD and enoyl-CoA hydratase, which
show that this ligand is bound as an enolate to both enzymes
with the enol tautomer localized on C1 in each case.
Acknowledgment. This work was supported by NIH grant
GM63121 to P.J.T. P.J.T. is an Alfred P. Sloan Fellow. We
are grateful to Prof. Q. Cui for helpful comments regarding the
DFT calculations.
(
63) Waterson, R. M.; Hill, R. L. J. Biol. Chem. 1972, 247, 5258-5265.
64) Furuta, S.; Miyazawa, S.; Osumi, T.; Hashimoto, T.; Ui, N. J. Biochem.
(
1
980, 88, 1059-1070.
JA050083P
8432 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005