Using a combination of experimental and computational approaches, we have shown that H-transfer reactions can occur by ‘deep’ tunnelling and the reaction can be enhanced by local-ized dy
Trang 1Structural and mechanistic aspects of flavoproteins:
probes of hydrogen tunnelling
Sam Hay, Christopher R Pudney and Nigel S Scrutton
Manchester Interdisciplinary Biocentre and Faculty of Life Science, University of Manchester, UK
Introduction
There is now fairly widespread recognition that
enzyme-catalysed C–H bond cleavage reactions can
occur by quantum mechanical tunnelling [1–5] The
role of protein dynamics in these reactions is still hotly
debated and it has been proposed that promoting
vibrations, nonequilibrated fast (sub-ps) dynamics,
could modify the reaction barrier and profoundly
influence the reaction rate [4,6–12] In recent years, we
have investigated H-transfer reactions in a number of
enzymes, primarily quinoprotein [4,13,14] and
flavo-protein [8,15–20] systems Using a combination of
experimental and computational approaches, we have
shown that H-transfer reactions can occur by ‘deep’
tunnelling and the reaction can be enhanced by
local-ized dynamics in the enzyme active site – putative pro-moting vibrations Although it is fairly well established that enzymatic H transfers often involve tunnelling, the role of promoting vibrations remains contentious [21] In this minireview, we describe experimental methods we have recently employed and developed to probe the role of environmental coupling⁄ promoting vibrations in H-transfer reactions in the Old Yellow Enzyme (OYE) family of flavoproteins
Hydrogen tunnelling
Because of wave⁄ particle duality, electrons and light atoms have appreciable (de Broglie) wavelengths The
Keywords
high pressure; H-tunneling; kinetic isotope
effect; kinetic isotope fractionation; multiple
reactive conformations; Old Yellow Enzyme;
promoting vibration; protein dynamics;
quantum mechanics; stopped-flow kinetics
Correspondence
N S Scrutton, Manchester Interdisciplinary
Biocentre and Faculty of Life Science,
University of Manchester, 131 Princess
Street, Manchester M1 7ND, UK
Fax: +44 161 306 8918
Tel: +44 161 306 5152
E-mail: nigel.scrutton@manchester.ac.uk
(Received 23 December 2008, revised 28
April 2009, accepted 1 May 2009)
doi:10.1111/j.1742-4658.2009.07121.x
At least half of all enzyme-catalysed reactions are thought to involve a hydrogen transfer In the last 10 years, it has become apparent that many
of these reactions will occur, in part, or in full, by quantum mechanical tunnelling We are particularly interested in the role of promoting vibra-tions on H transfer, and the Old Yellow Enzyme family of flavoproteins has proven to be an excellent model system with which to examine such reactions In this minireview, we describe new and established experimental methods used to study H-tunnelling in these enzymes and we consider some practical issues important to such studies The application of these methods has provided strong evidence linking protein dynamics and H-tun-nelling in biological systems
Abbreviations
DHFR, dihydrofolate reductase; EIE, equilibrium isotope effect; ET, electron transfer; GO, glucose oxidase; KIE, kinetic isotope effect; MR, morphinone reductase; OYE, Old Yellow Enzyme; PETNR, pentaerythritol tetranitrate reductase; RHR, reductive half-reaction.
Trang 2wavelength of H (used here to denote H+, H•and H))
is 1 A˚ and thus similar to a typical bond length,
whereas the wavelength of deuterium is shorter by a
factor of ffiffiffi
2
p
As a consequence, the position of H
(and to a lesser extent, D) is somewhat diffuse, and
H transfer may involve an appreciable degree of
quan-tum mechanical tunneling, in which H or D transfer
occurs by ‘tunnelling’ through part of the reaction
barrier rather than by passing over the barrier as is the
case in a classical transition-state reaction [22] It is
accepted that long-range electron transfer (ET)
reac-tions occur by tunnelling [23,24] and we now have
nearly 20 years of both experimental and
computa-tional evidence demonstrating that H-tunnelling
reac-tions can also occur during enzyme-catalysed reacreac-tions
[1–5] It is possible to computationally estimate the
extent to which tunnelling plays a role during an
H-transfer reaction In dihydrofolate reductase,
hydride transfer occurs by tunnelling 50% of the
time [25,26], with the remainder of the H transfer
occurring by an over-the-barrier mechanism
Con-versely, in soybean lipoxygenase-1 [27], aromatic amine
dehydrogenase [4] and the flavoprotein morphinone
reductase (MR) [28], calculations have shown that at
least 99% of H transfer occurs by tunnelling These
reactions can be thought of as ‘deep’ tunnelling
reac-tions because the H tunnels a relatively long way
below the top of the reaction barrier
The rate of a nonadiabatic (deep tunnelling)
H transfer can be described using modified Marcus
theory [23]:
ktun¼2p
h
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 4pkkBT
r
V
j j2ðF:C:Þ exp DGz
kBT
! ð1Þ
where V is the electronic coupling, F.C is the Frank–
Condon nuclear wave function overlap (related to the
de Broglie wavelength of the H or D) and DG is the
Marcus activation energy The activation energy is
described by the driving force, DG0, and reorganization
energy in the standard way:
DGz ¼ DG 0þ k2
The driving force dependence of H transfer in the
flavoprotein glucose oxidase (GO) was investigated by
Brinkley & Roth [29] The endogenous FAD was
substituted with other chemically modified flavins with
differing redox potentials, and these authors showed
that the apparent rate of H transfer obeys Eqns (1)
and (2) and the reorganization energy of this reaction
appears to be large (280 kJÆmol)1) We have since
shown a driving force dependence during the reduction
of the quinoprotein aromatic amine dehydrogenase with p-substituted phenylethylamine substrates [30] and estimated the reorganization energy for the reac-tion of this enzyme with tryptamine to be 250–300 kJÆ mol)1 [9] From the little experimental evidence cur-rently available, it appears that it is appropriate to describe H-tunnelling reactions using Marcus theory, and that a general feature of these reactions may be a large reorganization energy
The Frank–Condon term in Eqn (1) has been described by Kuznetsov & Ulstrup:
F:C:0;0¼
Z r 0
0 explixiDr2=2h
expðEX=kBTÞdX
ð3Þ where l and x are the mass and frequency of the transferred H or D and Ex is the environmental energy
or promoting vibration, which reduces the H-transfer distance from an equilibrium distance, r0, by Dr = (r0) rX) [9,31,32] The kinetic isotope effect (KIE =
kH⁄ kD) arises because of differences in the mass, frequency and consequently the transfer distance of H and D Experimentally, the identification of promoting vibrations is extremely challenging and, as yet, there is
no method to directly visualize such vibrations because they occur in the relatively inaccessible THz region The first experimental evidence for a role of environ-mental coupling during H-tunnelling reactions in enzymes [5,13] was inferred from observations of KIEs with aberrant temperature dependencies that do not conform to the predictions of semiclassical transition state theory or Bell-type quantum correction models [33] However, we have recently shown that the KIE temperature dependence (DDH, see below) is not a reliable diagnostic of environmental coupling [9] and other experiments are required to corroborate predic-tions based solely on DDHvalues
Measurement of H-transfer reactions
in OYEs
The OYE family of flavoproteins comprise a large group of FMN-containing NAD(P)H-dependent oxid-ases We have concentrated our studies on two homol-ogous OYEs: MR isolated from Pseudomonas putida M10 and pentaerythritol tetranitrate reductase (PET-NR) from Enterobacter cloacae For reference, in Table 1, we summarize the flavoproteins for which a specific H-tunnelling study has been performed Gener-ally, the KIEs of H transfer in flavoproteins are fairly modest (< 10) but the temperature dependencies of these KIEs are quite varied (Table 1)
Trang 3The reductive half-reaction (RHR) of MR and
PET-NR occurs in three steps:
Eoxþ NADðPÞH !binding½Eox NADðPÞHCT
!
reduction=
Htransfer
Ered NADðPÞþ !
product release
Eredþ NADðPÞþ
ð4Þ
MR reacts only with NADH, whereas PETNR has a
preference for NADPH It is sometimes possible to
trap the binary CT complex in OYEs by substituting
NAD(P)H with the nonreactive analogue
1,4,5,6-tetra-hydroNAD(P)H (NAD(P)H4) [19] MR binds NADH4
with Kd= 0.35 mm [34] and we have recently solved
the X-ray crystallographic structure of MR in complex
with NADH4 to a resolution of 1.3 A˚ [19] The
struc-ture of the active site in MR is shown in Fig 1, as is
the proposed mechanism of hydride transfer⁄ FMN
reduction In MR and PETNR, the H transfer is
ste-reospecific because the NAD(P)H nicotinamide moiety
can only bind in one orientation within the active site,
which places the pro-R (transferred) primary hydrogen
(Hp) directly over the FMN N5 acceptor atom
(labelled in Fig 1)
The reduced OYEs can be reoxidized by molecular
oxygen (k 1 s)1) [35] or with various oxidative
sub-strates – one generic substrate being cylcohexen-1-one
With many of the oxidative substrates tested, the
oxi-dative half-reactions of MR and PETNR are fully rate
limiting during steady-state turnover Consequently,
the steady-state KIE on the RHR hydride transfer is
unity and steady-state analysis is clearly not appropri-ate to study these reactions However, in MR, we have measured a KIE of 3.5 ± 0.2 on kcat for the hydride transfer from the reduced FMN to cylcohexen-1-one
Table 1 Kinetic isotope effects observed in selected flavoproteins CO, choline oxidase; DD, human class 2 dihydroorotate dehydrogenase; FDTS, flavin dependent thymidylate synthase; GO, glucose oxidase; MAO A ⁄ B, monoamine oxidase A ⁄ B; MR, morphinone reductase; PAO,
L -phenylalanine oxidase; PETNR, pentaerythritol tetranitrate reductase; TMADH, trimethylamine dehydrogenase; TSOX, heterotetrameric sar-cosine oxidase; nd, not determined or reported The KIEs are for pre-steady-state flavin reduction by the denoted substrate unless otherwise stated Isotope effects are H ⁄ D unless otherwise stated.
a Revised from previously reported, manuscript in preparation b Data from kcat⁄ K m measurements c Data not corrected for the calculated commitment to catalysis.dData from H ⁄ T isotope effect e
No error given.fData for the H172Q mutant.gCalculated from the KIE and DDH values.
A
B
Fig 1 (A) A model of the active site of NADH-bound morphinone reductase based on pdb 2R14 [19] Residues which form hydrogen bonds (dotted lines) to the bound NADH are shown, as are the NADH nicotinamide pro-R (Hp, the transferred H) and pro-S (Hs) hydrogens and the FMN N5 (acceptor) (B) The proposed reaction mechanism of hydride transfer ⁄ FMN reduction during the reductive half reaction in old yellow enzymes.
Trang 4[15] There is also a solvent KIE of 2.3 ± 0.3 and a
double isotope effect (measured with deuterated FMN
in D2O) of 8.2 ± 1.4 The rule of geometric mean [36–
38] states that multiple KIEs should be described by:
KIEa;b¼ KIEa KIEb ð5Þ and in the case of the oxidative half-reaction of MR,
the observed double KIE should be: 3.5· 2.3 = 8.0,
which is in agreement with the observed value of
8.2 ± 1.4 [15] Although it has been argued that
vio-lation of the rule of geometric mean may be used as
evidence for H-tunnelling [36–38], the oxidative KIEs
in MR are not measurably temperature dependent – a
diagnostic of ground state H-tunnelling [15]
Using a stopped-flow spectrometer, it is possible to
determine most of the rate constants for the steps in the
OYE RHR reaction (Eqn 4) above However, care must
be made to keep the samples anaerobic by either using
an anaerobic glove box or by adding glucose⁄ GO The
binary complex has a characteristic p-p charge transfer
(CT) absorbance and NAD(P)H binding and
dissocia-tion can be measured by following the formadissocia-tion of this
CT absorbance at, for example, 555 nm while
perform-ing a concentration dependence [35,39]:
kobs¼ koffþ kon½NADðPÞH ð6Þ
Similarly, the rate of hydride transfer can be
deter-mined because H transfer is concomitant with FMN
reduction By following the bleaching of FMN
absor-bance at 465 nm while performing a concentration
dependence [15,18,35,39], it is possible to characterize
the RHR:
kobs¼ koxþ kred
½NADðPÞH
koff=kon
ð Þ þ ½NADðPÞH ð7Þ
In both MR and PETNR, at room temperature, kon,
the apparent rate of coenzyme binding is106m)1Æs)1
The rate of NAD(P)H dissociation, koff, is 102s)1
and the apparent rate of H transfer, kred, is 56 and
33 s)1 in MR and PETNR, respectively [18,35,39]
Product (NAD(P)+) inhibition of MR and PETNR is
very weak suggesting that NAD(P)+ rapidly
dissoci-ates from the active site once FMN reduction occurs
We have been unable to measure the reverse rate of
hydride transfer, kox, in either enzyme and it appears
to be close to zero [18,40] We have also determined
the driving force for hydride transfer during the RHR
of MR with NADH to be 60 kJÆmol)1 [40], which
is also consistent with an effectively irreversible
H transfer
We have mutated various amino acid residues within the active site of MR and PETNR [19,41,42] In the wild-type enzymes, FMN reduction occurs as a mono-exponential process (Fig 2) – greatly simplifying the stopped-flow analysis In the H186A and N189A active-site mutants in MR (Fig 1A), FMN reduction kinetics become more complex – i.e multi-exponential [41] As an extreme example, we have measured at least four kinetically resolved components of FMN reduction in the N189A mutant of MR, each with a significant KIE (Fig 2) [19], and each able to intercon-vert [42a] We have attributed this complexity to the formation of multiple reactive configurations in the mutant enzyme because of improper binding of the NADH nicotinamide moiety within the active site Of concern is that, had we performed a steady-state analy-sis of this mutant, we would not have observed this heterogeneity and the mutant enzyme would have appeared to be quite similar to the wild-type enzyme Caution is therefore needed when using steady-state approaches to analyse tunnelling, especially with mutant enzymes
During the RHR of MR and PETNR, when pro-R deuterated NAD(P)H (denoted (R)-[4-2H]-NAD(P)H)
is used in place of the protiated coenzyme, a significant KIE is observed on hydride transfer (kred) but not on
koff⁄ kon (Table 1) [15,18] Because the H transfer is effectively irreversible and kinetically resolved from coenzyme binding (no KIE on koff⁄ kon), the observed KIE is essentially the intrinsic KIE Using stopped-flow methods, we have measured the temperature dependence of the rate of H transfer in both MR and PETNR For convenience, we tend to measure kobs(in
0.20
0.15
0.10
0.05
0.00
0.01
0.0 0.1 0.00
0.05 0.10 0.15 0.20
20 40
0.1
wt
N189A
t·s–1
Fig 2 Multiple reactive conformations during an H-transfer reac-tion Stopped-flow traces of the observed FMN reduction during the reaction of the wild-type (wt) and N189A mutant of MR with NADH The wt trace is fit to a single exponential and the N189A trace to a 4-exponential function – see the main text for more details (Inset) The same data on a split-axis linear time scale Adapted from Pudney et al [19].
Trang 5the presence of saturating [NAD(P)H]) rather than kred
(Eqn 7), although they give equivalent results
[8,15,18,19,35] We define saturating coenzyme
concen-trations as [NAD(P)H] > 10· KS, where KS= koff⁄
kon Care must be taken because KScan be quite
tem-perature dependent [15,19,35] We typically analyse
these data in terms of Eyring (transition state) theory:
ln kobs
T
¼ ln kB h
þDSz
T
DHz
with KIEobs¼ kH
obs=kD
obsand the temperature dependence
on the KIE:
DDHz ¼ DHzD DHzH¼ DEa ð9Þ
In wild-type MR and PETNR it is generally possible
to determine observed rate constants with an accuracy
of1% (measured with different samples on different
days) It is then possible to determine KIEs to an
accuracy of 2-5% and DDH with an error of 1-2 kJÆ
mol)1[18]
The use of KIE analyses relies heavily on the ability
to obtain isotopically pure substrates or coenzymes
One of the reasons OYEs are particularly attractive to
study is the ability to create stereospecifically labelled
isotopologues of NAD(P)H (described in the next
sec-tion) This allows the measurement of 1 KIEs, 2 KIEs
and double KIEs – where both Hpand Hsare deuterated
[18] We have been able to use stopped-flow methods to
measure quite accurately both the magnitude and
tem-perature dependence of a-2 KIEs during the RHR in
MR and PETNR [18,40], and for hydride transfer in the
thermophilic dihydrofolate reductase (DHFR) from
Thermotoga maritima [43] The equilibrium isotope
effect (EIE) on NAD(P)H oxidation was measured by
Cook & Cleland to be 1.13 [44] The observation of a-2
KIEs values larger than the EIE was rationalized by
Huskey & Schowen [36] because of coupling of the
motion between the 2 hydrogen (labelled in Fig 1A)
and the 1 (transferred) hydrogen during an
H-tunnel-ling reaction We have measured identical a-2 KIE
val-ues of1.2 in MR and PETNR, which are significantly
larger than the EIE [18,40] We have also measured the
double KIE in MR [18] and shown that in this reaction,
the rule of geometric mean (Eqn 5) is most likely
vio-lated [39] We have shown computationally that the
H transfer in MR occurs by deep tunnelling [28] so
Hus-key & Schowen’s [36] interpretation of inflated 2 KIEs
would seem plausible However, we have measured a
normal (KIE£ EIE) and temperature-independent a-2
KIE in TmDHFR, yet this reaction proceeds by 50-80%
tunneling, depending on the temperature [25,43] A
simi-lar observation has been observed in the Escherichia coli DHFR [45] Consequently, it would appear that inflated a-2 KIEs may be indicative of a tunnelling contribution
to the H transfer, but normal KIEs do not rule out tun-nelling [43] A similar argument can be made for 1 KIEs – although KIEs£ 7 can be explained using transi-tion state theory, the KIE arising from a full tunnelling reaction (described by Eqns 1–3) can be any value [10,46]
Preparation of coenzymes
The methods of coenzyme deuteration are well described [47–50], but are typically for microscale syn-theses, 1 mg This can be an issue for stopped-flow experiments, substrate-binding titrations and crystallo-graphic studies when a very large amount of the sub-strate may be required; typical NAD(P)H saturation constants for OYEs are 0.1-1 mm and as an example,
a typical measurement, by stopped-flow, of the temper-ature dependence of the 1 KIE of an OYE may require 100 mg of isotopically pure substrate For reference, we briefly describe our preferred methods of synthesis for large-scale (1 g) preparations of all three deuterated NADH and NADPH isotopologues: (R)-[4-2H]-NAD(P)H, (S)-[4-2H]-NAD(P)H and (R,S)-[4,4-2H2]-NAD(P)H These syntheses typically yield > 95% isotopologue purity (based on 1H NMR spectra, see Pudney et al [18] for examples), with the corre-sponding impurity being the protiated coenzyme Kohen [49] recently developed syntheses for extremely high-purity NADPH isotopologues and the method of McCracken et al [49] has been reported to yield > 99% isotopologue purity – a purity necessary when performing competitive isotope experiments We also describe our synthesis of 1,4,5,6-tetrahydroNAD(P)H
We typically prepare (R)-[4-2H]-NADH by the ste-reospecific reduction of NAD+(500 mg) with 1-[2H6 ]-ethanol (1 g) using yeast alcohol dehydrogenase (200 U) and aldehyde dehydrogenase (100 U) in
20 mm Taps pH 8.5 (20 mL) at room temperature This method is a slight modification of the procedure reported in Viola et al [48] (R)-[4-2H]-NADPH is pre-pared through a stereospecific reduction of NADP+ (300 mg) with 1-[2H6]-isopropanol (1 g) using NADP+-dependent alcohol dehydrogenase from Ther-moanaerobacter brokii (100 U) in 20 mm Taps pH 8.5 (50 mL) at 42 C These reactions are deemed com-plete when A340stopped increasing and A260:A340< 3, typically after 1 h (S)-[4-2H]-NADH and (S)-[4-2 H]-NADPH are prepared through stereospecific reduction
of NAD+ (500 mg) and NADP+ (500 mg), respec-tively, with 1-[2H]-glucose (150 mg) using glucose
Trang 6dehydrogenase (150 U) in Taps pH 8.5 (10 mL) at
room temperature This method is a slight
modifica-tion of the procedures reported in Ottolina et al [50]
and the reactions typically take 3 h (R,S)-[4,4-2H2
]-NADH is prepared by stereospecific oxidation of
(S)-[4-2H]-NADH (300 mg) with 100 mm
cyclohexen-1-one catalysed by 10 lm MR in Taps pH 8.5
(10 mL) The deuterated NAD+is purified in the same
manner as for (R)-[4-2H]-NADH (R,S)-[4,4-2H2
]-NADH is then prepared through a further
stereospe-cific reduction of [4-2H]-NAD+ with 1-[2H]-glucose
(100 mg) and glucose dehydrogenase (100 U) in Taps
pH 8.5 (10 mL) (R,S)-[4,4-2H2]-NADPH can be
pre-pared in the same manner as (R,S)-[4,4-2H2]-NADH
However, an NADPH-specific enzyme such as PETNR
must be used in place of MR NADH4 and NADPH4
are prepared by maintaining a slight pressure
(1.2 bar) of hydrogen (> 99%) over a solution of
NAD(P)H (500 mg) and palladium-activated charcoal
(30 mg) in Tris⁄ Cl pH 9.0 (5 mL) stirred on ice [19]
The reaction is stopped when no absorbance at
340 nm is observed and typical A266⁄ A288 ratios of
1.06 are obtained
We purify the coenzymes using anion-exchange
(Q-Sepharose) chromatography, eluting NADH and
NADPH isotopologues (including the tetrahydro forms)
in200 and 500 mm ammonium bicarbonate,
respec-tively [18] All of the enzymes used in these syntheses
(excluding MR) are available from Sigma-Aldrich (St
Louis, MO, USA) and the coenzymes are available
from Melford Laboratories (Chelsworth, UK) We use
extinction coefficients of 6.22 mm)1Æcm)1at 340 nm for
NAD(P)H isotopologues and 16.8 mm)1Æcm)1 at 289
nm for NAD(P)H4[34] Usually the enzymatic synthesis
of (R)-[4-2H]-NAD(P)H does not proceed to
comple-tion We have observed that freezing or freeze-drying
the reaction volume before purification usually leads to
the formation of a significant impurity of undeuterated
NAD(P)H Consequently, on this scale, it is important
to purify the reaction volume as quickly as possible after
synthesis has ceased Also, one must take care to
main-tain the pH at 8.5 over the course of the enzymatic
synthesis, because acid catalysed decomposition of
NAD(P)H may be a significant contributor to substrate
(in)activity [51]
As an aside, we have recently investigated the effect
of incomplete coenzyme⁄ substrate deuteration on the
observed KIE measured using stopped-flow methods
[39] We found that, if there is a reversible chemical
step preceding H transfer and the reverse rate of this
step (koff in Eqn 4) is comparable with the rate of
H transfer, then kinetic isotope fractionation can
occur, leading to the formation of more protiated than
deuterated product This fractionation also leads to an overestimation of kD and consequently an underesti-mation of the KIE It appears that it is possible to cor-rect for incomplete deuteration using a simple linear relationship in an analogous fashion to that used to correct steady-state data:
kD obs¼ kHð1 fDÞ þ kDfD ð10Þ where fDis the fraction of substrate deuteration, which can usually be determined quite accurately by
1H NMR [18,48,52] or possibly MS [18,53] In Fig 3,
we use Eqn (10) to model the effect of partial deutera-tion on the observed rate and KIE of an H transfer reaction If kD is underestimated then so too will be DDHand the effect of fDon the apparent temperature dependence of the KIE is also shown in Fig 3 We determined Eqn (10) empirically and this relationship
is quite approximate Nevertheless, we have been able
to correct the RHRs of MR and PETNR and also the RHR of aromatic amine dehydrogenase with benzyl-amine using Eqn (10) [39] However, further studies are required to confirm the general validity of this correction method That fractionation can occur emphasizes the need for: (a) care in preparing high-purity coenzymes, and (b) correction for small isotope impurities in the analysis of tunnelling kinetics using stopped-flow single turnover measurements
Hydrostatic pressure
Hydrostatic pressure offers an alternative or comple-mentary method to temperature with which to study
6
A
B
5
4
3
2
1
4
2
0
10
Kobs
8
6
4
2
0 0.0 0.2 0.4 0.6 0.8 1.0 Fraction deuteration
‡ / kJ·mol
Fig 3 The effect of substrate isotopic purity on the observed rate
of deuterium transfer (filled squares) and the corresponding KIE (open circles) (A), and on the temperature dependence (B) of a modelled H-transfer reaction The data are modelled using Eqn (10) with kH= 5 s)1, a KIE of 5 and various values of DDH
Trang 7enzymatic reactions Semiclassical transition-state
the-ory states that pressure effects on isotope effects arise
because of differences in vibrational frequencies [54–56]
and stretching vibrations are insensitive to pressures of a
few kbar [54] Consequently, the KIEs of purely
transi-tion state reactransi-tions are expected to be insensitive to
pres-sure Conversely, several chemical systems with inflated
KIEs indicative of a tunnelling contribution to the
H transfer have been shown to exhibit a significant
pres-sure-dependence of both rate and KIE [57] Thus, in
principle, the pressure-dependence of an isotope effect
provides an excellent method for distinguishing between
transition state and tunnelling reactions
The use of pressure to study enzymatic H-tunnelling
reactions was pioneered by Northrop, who, 10 years
ago, developed a model [58] for the
pressure-depen-dence of H-transfer reactions based on the Bell
correc-tion [33,58] This model was then used quite
successfully to model the pressure dependence of
steady-state H-transfer reactions in alcohol
dehydroge-nase and aldehyde dehydrogedehydroge-nase [57–59] We recently
performed a high-pressure stopped-flow study of the
hydride transfer during the RHR of MR with NADH
The apparent rate of hydride transfer increased by
approximately twofold per kbar increase in pressure
and the 1 KIE also showed a small but significant
increase in magnitude with pressure (Fig 4C)
Together, these observations could not be explained
using Northrop’s model [58], nor with a simple
non-adiabatic H-tunnelling model (e.g Eqns 1–3) when
pressure simply causes a compression of the reaction
barrier [60] However, we found that we could
qualita-tively model the data by invoking a promoting
vibra-tion that changes frequency with pressure [8] We have
since refined this analysis and recently described a
sim-ple nonadiabatic H-tunnelling model which explicitly
includes pressure as a variable [61]:
kH=kD exp l½ DxD lHxH r0þ Dr:p2
=2h
exp lf ½ DxD lHxHkBT=h jð 0þ Dj:pÞg
ð11Þ
where r0 is the average H-transfer distance, Dr is the
change in this distance with pressure, j0 is the force
constant describing the promoting vibration and Dj is
the change in this force constant with pressure
Equa-tion (11) can be used as a fitting funcEqua-tion with four
adjustable parameters and the KIE can either increase
or decrease with increasing pressure when Dr and⁄ or
Dj become significant (Fig 4) Although this model is
oversimplistic, it is possible to use Eqn (11) to describe
a reaction in which both the apparent rate and KIE
increase (or decrease) with pressure [61] The model
10
A
B
C
8
6
4
2
10
8
6
4
2
10
8
6
4
2 2.0 1.5 1.0 0.5 0.0 3.2 3.3 3.4 3.5 3.6
1.0 0.5 0.0 –0.5 –1.0 2.0 1.5 1.0 0.5 0.0
0.02 0.01 dr/Å kbar –1
·kbar –1
/T·K –1
0.00 –0.01 –0.02 2.0
1.5
Pressure/kbar
Pressure/kbar
Press ure/kbar
1.0 0.5 0.0
Fig 4 A variable pressure H-tunnelling model (Eqn 3) [61] The KIE pressure dependence is modeled when (A) the frequency of the promoting motion or (B) the H-transfer distance changes with pres-sure Positive values of Dj and Dr reflect increases in frequency and distance with pressure, respectively The data are modeled with j = 5 JÆm)2, r0= 0.52 A ˚ (KIE0= 5) and only one parameter in each plot is varied It is possible for both Dj and Dr to vary with pressure (as we have modelled in MR) [61], causing curvature in the KIE versus pressure plots We have also plotted (C) the com-bined pressure and temperature dependence of the observed KIE
on hydride transfer during the reductive half reaction of morphinone reductase The data are taken from Hay et al [8] We have not plot-ted error bars for clarity but the average error in the KIE for this data set is ±5% and the minimum and maximum error is 1% and 18%, respectively.
Trang 8should only be used qualitatively, but is useful to
esti-mate whether: (a) the tunnelling distance changes with
pressure, and (b) to confirm that there is
environmen-tal coupling and to determine whether the frequency of
this vibration is likely to change with pressure
Although it seems intuitive that hydrostatic pressure
will ‘squeeze’ the enzyme and thus compress the
H-transfer distance (achieved by increasing the
popula-tion of enzyme–substrate conformers with shorter
H-transfer distances in an equilibrium distribution of
conformational states), until recently, this assumption
remained untested Ewald [62] has shown that increasing
pressure causes a progressive shortening of the CT bond
in synthetic p-p complexes with an accompanying shift
to red wavelengths and increase in absorbance We have
recently shown that increasing pressure also causes a
shortening of the CT bond (increase in CT absorbance)
in NADH4-bound MR, which we have interpreted as
pressure-induced barrier compression [34] Using
vari-able pressure molecular dynamics simulations of
NADH-bound MR, we were able to corroborate this
finding [34] We found that the heavy atom transfer
dis-tance has an approximately Gaussian distribution that
both narrows and shifts to shorter distances at elevated
pressures It appears, at least in MR, that pressure does
not physically squeeze the microscopic reaction barrier,
but rather reduces the average barrier width, the
macro-scopic barrier, by restricting the conformational space
available to the NADH and FMN moieties within the
active site Further studies are required to determine
whether this is a general phenomenon
Other experimental probes
In addition to temperature and hydrostatic pressure, it
is possible to experimentally probe enzymatic
H-tun-nelling reactions using additional experimental
parame-ters and we briefly discuss the use of varying the
solvent composition to probe the effect of solvent
dielectric and viscosity on H transfer chemistry
It is predicted from von Smoluchowski’s theory [63]
that the rate of a diffusion-controlled (bimolecular)
reaction will be inversely proportional to the bulk
solution viscosity The effect of viscosity on a
unimo-lecular reaction is more complicated but can be
described in combination with the Eyring equation
according to Ansari et al [64]:
kobs¼kBT
h
1þ r
gþ r
exp DSz R
! exp DHz RT
! ð12Þ
where r, in units of viscosity, is the contribution of
the protein friction to the total friction of the system
The activation entropy and enthalpy can be deter-mined independently from the temperature dependence
of the reaction [65] Solution viscosity has been used to probe the role of dynamics in interprotein [65–68] and intraprotein [69] ET reactions and protein rearrange-ment after carbon monoxide dissociation from myoglo-bin [64] In general, the rates of ET reactions that are conformationally gated decrease upon an increase in solvent viscosity
The viscosity dependence of several enzymatic H-transfer reactions has now been investigated Protein dynamics can be affected by surface glycosylation and this approach has been used by Klinman and cowork-ers to study the viscosity dependence of the rate of hydride transfer in GO [70,71] These authors studied various glycoforms of the enzyme (varying in the extent of glycosylation) [70] and also replaced the native polysaccharide with different polymeric forms
of polyethylene glycol [71] A decrease in the ‘fitness’
of GO was observed when the apparent surface viscos-ity increased or decreased relative to the wild-type enzyme Fitness was defined as a reduction (away from unity) in the Arrhenius pre-exponential ratio (AD: AT) [70,71] In a more conventional study, we found that the magnitude and temperature dependence of the pre-steady-state rate and KIE for proton tunnelling during the RHR of the quinoprotein methylamine dehydroge-nase are unchanged following the addition of 30% glycerol – an increase in solvent viscosity of approxi-mately two- to threefold [13] Conversely, a decrease in KIE and increase in apparent enthalpy for the RHR
of l-phenylalanine oxidase upon the addition of 30% glycerol has been reported [72] In a more systematic study, we recently showed that the rate of coenzyme capture decreases, whereas the rate and KIE of hydride transfer during the RHR in MR are invariant over a 10-fold increase in solution viscosity [20] We found it was possible to use a conventional stopped-flow to make these measurements by varying the viscosity between 0.9 and 9 cP at 25 C with the addition of 0–60% w⁄ w glycerol Addition of > 60% glycerol leads to mixing artefacts that precluded further measurements The addition of glycerol to the solvent will also reduce the solvent dielectric We inde-pendently probed the role of solvent dielectric on the RHR of MR by measuring the temperature depen-dence in this reaction in the presence of ethanol Neither the rate nor enthalpy significantly changed upon the addition of 20% v⁄ v ethanol – a change in dielectric from 80 to 65, but NADH binding was significantly compromised with an increase in KSfrom 0.2 to 2.7 mm observed Unfortunately, no clear trends have emerged as to the effect of viscosity (or dielectric)
Trang 9on enzymatic H-transfer reactions and more systematic
studies are probably required to determine whether
this is a useful probe of H transfer dynamics
Future perspectives
It is now fairly well established that many enzyme
H-transfer reactions involve a degree of quantum
mechanical H-tunnelling The role of promoting
vibra-tions, which couple protein dynamics to the H transfer
reaction coordinate, remains contentious Although
there is now a growing body of compelling experimental
and computational evidence for such vibrations, the
experimental evidence is all by inference A combined
temperature and pressure study seems to be the best
experimental probe of environmental coupling to
H-transfer chemistry [8] Computational studies are also
invaluable because these can determine the extent of
H-tunnelling and also visualize promoting vibrations [9]
The challenge for the future remains the direct
measure-ment of such vibrations If they are found to exist then
a further challenge is to exploit them for practical gain
Acknowledgement
This work was funded by the UK Biotechnology and
Biological Sciences Research Council NSS is a
BBSRC Professorial Fellow
References
1 Garcia-Viloca M, Gao J, Karplus M & Truhlar DG
(2004) How enzymes work: analysis by modern rate
theory and computer simulations Science 303, 186–
195
2 Cha Y, Murray CJ & Klinman JP (1989) Hydrogen
tunneling in enzyme-reactions Science 243, 1325–1330
3 Benkovic SJ & Hammes-Schiffer S (2003) A
perspec-tive on enzyme catalysis Science 301, 1196–1202
4 Masgrau L, Roujeinikova A, Johannissen LO, Hothi
P, Basran J, Ranaghan KE, Mulholland AJ, Sutcliffe
MJ, Scrutton NS & Leys D (2006) Atomic description
of an enzyme reaction dominated by proton tunneling
Science 312, 237–241
5 Kohen A, Cannio R, Bartolucci S & Klinman JP (1999)
Enzyme dynamics and hydrogen tunnelling in a
thermo-philic alcohol dehydrogenase Nature 399, 496–499
6 Antoniou D & Schwartz SD (1997) Large kinetic
iso-tope effects in enzymatic proton transfer and the role
of substrate oscillations Proc Natl Acad Sci USA 94,
12360–12365
7 Caratzoulas S, Mincer JS & Schwartz SD (2002)
Iden-tification of a protein-promoting vibration in the
reac-tion catalyzed by horse liver alcohol dehydrogenase
J Am Chem Soc 124, 3270–3276
8 Hay S, Sutcliffe MJ & Scrutton NS (2007) Promoting motions in enzyme catalysis probed by pressure studies
of kinetic isotope effects Proc Natl Acad Sci USA 104, 507–512
9 Johannissen LO, Hay S, Scrutton NS & Sutcliffe MJ (2007) Proton tunneling in aromatic amine dehydroge-nase is driven by a short-range sub-picosecond pro-moting vibration: consistency of simulation and theory with experiment J Phys Chem B 111, 2631–2638
10 Knapp MJ & Klinman JP (2002) Environmentally cou-pled hydrogen tunneling – linking catalysis to dynam-ics Eur J Biochem 269, 3113–3121
11 Antoniou D & Schwartz SD (2001) Internal enzyme motions as a source of catalytic activity: rate-promot-ing vibrations and hydrogen tunnelrate-promot-ing J Phys Chem B
105, 5553–5558
12 Bruno WJ & Bialek W (1992) Vibrationally enhanced tunneling as a mechanism for enzymatic hydrogen transfer Biophys J 63, 689–699
13 Basran J, Sutcliffe MJ & Scrutton NS (1999) Enzy-matic H-transfer requires vibration-driven extreme tun-neling Biochemistry 38, 3218–3222
14 Basran J, Patel S, Sutcliffe MJ & Scrutton NS (2001) Importance of barrier shape in enzyme-catalyzed reac-tions – vibrationally assisted hydrogen tunneling in tryptophan tryptophylquinone-dependent amine dehy-drogenases J Biol Chem 276, 6234–6242
15 Basran J, Harris RJ, Sutcliffe MJ & Scrutton NS (2003) H-tunneling in the multiple H-transfers of the catalytic cycle of morphinone reductase and in the reductive half-reaction of the homologous pentaerythr-itol tetranitrate reductase J Biol Chem 278, 43973– 43982
16 Basran J, Sutcliffe MJ & Scrutton NS (2001) Deute-rium isotope effects during carbon–hydrogen bond cleavage by trimethylamine dehydrogenase – implica-tions for mechanism and vibrationally assisted hydro-gen tunneling in wild-type and mutant enzymes J Biol Chem 276, 24581–24587
17 Harris RJ, Meskys R, Sutcliffe MJ & Scrutton NS (2000) Kinetic studies of the mechanism of carbon– hydrogen bond breakage by the heterotetrameric sar-cosine oxidase of Arthrobacter sp 1-IN Biochemistry
39, 1189–1198
18 Pudney CR, Hay S, Sutcliffe MJ & Scrutton NS (2006) Alpha-secondary isotope effects as probes of
‘tunneling-ready’ configurations in enzymatic H-tunnel-ing: insight from environmentally coupled tunneling models J Am Chem Soc 128, 14053–14058
19 Pudney CR, Hay S, Pang JY, Costello C, Leys D, Sutcliffe MJ & Scrutton NS (2007) Mutagenesis of morphinone reductase induces multiple reactive
Trang 10configurations and identifies potential ambiguity in
kinetic analysis of enzyme tunneling mechanisms
J Am Chem Soc 129, 13949–13956
20 Hay S, Pudney CR, Sutcliffe MJ & Scrutton NS
(2008) Are environmentally coupled enzymatic
hydrogen tunneling reactions influenced by changes
in solution viscosity? Angew Chem Int Ed Engl 47,
537–540
21 Ball P (2004) Enzymes – by chance, or by design?
Nature 431, 396–397
22 Gladstone S, Laidler KJ & Eyring H (1941) The Theory
of Rate Processes McGraw-Hill, New York, NY
23 Marcus RA & Sutin N (1985) Electron transfers in
chemistry and biology Biochim Biophys Acta 811, 265–
322
24 Devault D, Parkes JH & Chance B (1967) Electron
tunnelling in cytochromes Nature 215, 642–644
25 Pang J, Pu JZ, Gao JL, Truhlar DG & Allemann RK
(2006) Hydride transfer reaction catalyzed by
hyper-thermophilic dihydrofolate reductase is dominated by
quantum mechanical tunneling and is promoted by
both inter- and intramonomeric correlated motions J
Am Chem Soc 128, 8015–8023
26 Garcia-Viloca M, Truhlar DG & Gao JL (2003)
Reac-tion-path energetics and kinetics of the hydride transfer
reaction catalyzed by dihydrofolate reductase
Bio-chemistry 42, 13558–13575
27 Tejero I, Garcia-Viloca M, Gonzalez-Lafont A, Lluch
JM & York DM (2006) Enzyme dynamics and
tunnel-ing enhanced by compression in the hydrogen
abstrac-tion catalyzed by soybean lipoxygenase-1 J Phys
Chem B 110, 24708–24719
28 Pang J, Hay S, Scrutton NS & Sutcliffe MJ (2008)
Deep tunneling dominates the biologically important
hydride transfer reaction from NADH to FMN in
morphinone reductase J Am Chem Soc 130, 7092–
7097
29 Brinkley DW & Roth JP (2005) Determination of a
large reorganization energy barrier for hydride
abstrac-tion by glucose oxidase J Am Chem Soc 127, 15720–
15721
30 Hothi P, Hay S, Roujeinikova A, Sutcliffe MJ, Lee M,
Leys D, Cullis PM & Scrutton NS (2008) Driving force
analysis of proton tunnelling across a reactivity series
for an enzyme–substrate complex ChemBioChem 9,
2839–2845
31 Knapp MJ, Rickert K & Klinman JP (2002)
Tempera-ture-dependent isotope effects in soybean
lipoxygenase-1: correlating hydrogen tunneling with protein
dynam-ics J Am Chem Soc 124, 3865–3874
32 Kuznetsov AM & Ulstrup J (1999) Proton and
hydro-gen atom tunnelling in hydrolytic and redox enzyme
catalysis Can J Chem 77, 1085–1096
33 Bell R (1980) The Tunnel Effect in Chemistry
Chap-man and Hall, London
34 Hay S, Pudney C, McGrory TA, Pan J, Sutcliffe MJ & Scrutton NS (2009) Barrier compression enhances an enzymatic H-transfer reaction Angew Chem 48, 1452– 1454
35 Craig DH, Moody PCE, Bruce NC & Scrutton NS (1998) Reductive and oxidative half-reactions of morphinone reductase from Pseudomonas putida M10:
a kinetic and thermodynamic analysis Biochemistry
37, 7598–7607
36 Huskey WP & Schowen RL (1983) Reaction-coordinate tunneling in hydride-transfer reactions J Am Chem Soc 105, 5704–5706
37 Saunders WH (1985) Calculations of isotope effects in elimination-reactions – new experimental criteria for tunneling in slow proton transfers J Am Chem Soc
107, 164–169
38 Ostovic D, Roberts RMG & Kreevoy MM (1983)
analogs J Am Chem Soc 105, 7629–7631
39 Hay S, Pudney CR, Hothi P & Scrutton NS (2008) Correction of pre-steady state KIEs for isotopic impu-rities and the consequences of kinetic isotope fraction-ation J Phys Chem A 112, 13109–13115
40 Hay S, Pudney CR, Sutcliffe MJ & Scrutton NS (2008) Solvent as a probe of active site motion and chemistry during the hydrogen tunnelling reaction
in morphinone reductase ChemPhysChem 9, 1875– 1881
41 Messiha HL, Munro AW, Bruce NC, Barsukov I & Scrutton NS (2005) Reaction of morphinone reductase with 2-cyclohexen-1-one and 1-nitrocyclohexene J Biol Chem 280, 10695–10709
42 Khan H, Barna T, Bruce NC, Munro AW, Leys D & Scrutton NS (2005) Proton transfer in the oxidative half-reaction of pentaerythritol tetranitrate reductase FEBS J 272, 4660–4671
42a Pudney CR, McGrory T, Lafite P, Pang J, Hay S, Leys D, Sutcliffe MJ & Scrutton NS (2009) Parallel pathways and free-energy landscapes for enzymatic hydride transfer probed by hydrostatic pressure ChemBioChem 10, 1379–1384
43 Hay S, Pang JY, Monaghan PJ, Wang X, Evans RM, Sutcliffe MJ, Allemann RK & Scrutton NS (2008) Secondary kinetic isotope effects as probes of environ-mentally-coupled enzymatic hydrogen tunneling reac-tions ChemPhysChem 9, 1536–1539
44 Cook PF, Blanchard JS & Cleland WW (1980) Primary and secondary deuterium-isotope effects on equilibrium-constants for enzyme-catalyzed reactions Biochemistry 19, 4853–4858
45 Pu JZ, Ma SH, Garcia-Viloca M, Gao JL, Truhlar
DG & Kohen A (2005) Nonperfect synchronization of reaction center rehybridization in the transition state
of the hydride transfer catalyzed by dihydrofolate reductase J Am Chem Soc 127, 14879–14886