A secondary β deuterium kinetic isotope effect in the chorismate synthase reaction

Article abstract of DOI:10.1006/bioo.2000.1174

Chorismate synthase (EC 4.6.1.4) is the shikimate pathway enzyme that catalyzes the conversion of 5-enolpyruvylshikimate 3-phosphate (EPSP) to chorismate. The enzyme reaction is unusual because it involves a trans-1,4 elimination of the C-3 phosphate and

Full text of DOI:10.1006/bioo.2000.1174

Bioorganic Chemistry 28, 191–204 (2000)
doi:10.1006/bioo.2000.1174, available online at http://www.idealibrary.com on
A Secondary ␤ Deuterium Kinetic Isotope Effect in the
Chorismate Synthase Reaction
Stephen Bornemann,*,1 Maria-Elena Theoclitou,† Martin Brune,‡
Martin R. Webb,‡ Roger N. F. Thorneley,* and Chris Abell†
*Biological Chemistry Department, John Innes Centre, Norwich Research Park, Colney,
Norwich NR4 7UH, United Kingdom; †Department of Chemistry, University Chemical Laboratory,
University of Cambridge, Lensfield Road, Cambridge CB2 1EW, United Kingdom; and
‡National Institute for Medical Research, Mill Hill, London NW7 1AA, United Kingdom
Received March 17, 2000
Chorismate synthase (EC 4.6.1.4) is the shikimate pathway enzyme that catalyzes the conver-
sion of 5-enolpyruvylshikimate 3-phosphate (EPSP) to chorismate. The enzyme reaction is
unusual because it involves a trans-1,4 elimination of the C-3 phosphate and the C-6 proR
hydrogen and it has an absolute requirement for reduced flavin. Several mechanisms have been
proposed to account for the cofactor requirement and stereochemistry of the reaction, including
a radical mechanism. This paper describes the synthesis of [4-2H]EPSP and the observation of
kinetic isotope effects using this substrate with both Neurospora crassa and Escherichia coli
chorismate synthases. The magnitude of the effects were D(V) ϭ 1.08 Ϯ 0.01 for the N. crassa
enzyme and 1.10 Ϯ 0.02 on phosphate release under single-turnover conditions for the E. coli
enzyme. The effects are best rationalised as substantial secondary ␤ isotope effects. It is most
likely that the C(3)–O bond is cleaved first in a nonconcerted E1 or radical reaction mechanism.
Although this study alone cannot rule out a concerted E2-type mechanism, the C(3)–O bond
would have to be substantially more broken than the proR C(6)–H bond in a transition state
of such a mechanism. Importantly, although the E. coli and N. crassa enzymes have different
rate limiting steps, their catalytic mechanisms are most likely to be chemically identical.
᭧ 2000 Academic Press
Key Words: chorismate synthase; kinetic isotope effect.
INTRODUCTION
Chorismate synthase (EC 4.6.1.4) is the shikimate pathway enzyme that catalyzes
the conversion of 5-enolpyruvylshikimate 3-phosphate (EPSP) to chorismate (1)
(Scheme 1). It is a potentially attractive target for herbicides and antimicrobials
because it is present in bacteria, fungi, and plants, but not in mammals. The enzyme
is unusual because it has an absolute requirement for reduced flavin (2–4). In addition,
1 To whom correspondence and reprint requests should be addressed. Fax: ϩ44 (0)1603 450018.
E-mail: stephen.bornemann@bbsrc.ac.uk.
191
0045-2068/00 $35.00
Copyright ᭧ 2000 by Academic Press
All rights of reproduction in any form reserved.

192
BORNEMANN ET AL.
SCHEME 1
the reaction involves an unusual trans-1,4 elimination of the C-3 phosphate and the
C-6 proR hydrogen (5–7), indicating that the reaction may be nonconcerted. Several
mechanisms have been proposed over a number of years to account for the cofactor
requirement and stereochemistry of the reaction. These have been extensively reviewed
(1, 8, 9).
The strongest evidence for a nonconcerted reaction comes from kinetic isotope
effect studies (10) and transient kinetics with the natural substrate (11). In addition,
studies using the slow substrate analogue (6S)-6-fluoro-EPSP (12) indicate a noncon-
certed mechanism involving either a radical or cationic intermediate (Schemes 1a and
1b, respectively). Furthermore, the formation of a flavin semiquinone radical with
the substrate analogue (6R)-6-fluoro-EPSP (13) leading to phosphate cleavage (14)
and the lack of activity with 5-deaza-FMN (15, 16) strongly support the radical
mechanism. All of this evidence taken together makes the radical mechanism the
most attractive explanation for both unusual aspects of this enzyme. Objections to
the radical mechanism could be raised given that other studies with model chemical
systems have shown that C(6)–H bond-cleavage could precede C(3)–O cleavage
(17). Moreover, electrochemical studies have indicated that the reduction of model
compounds in solution leads to decarboxylation rather than phosphate elimination (18).
A small but significant kinetic isotope effect (1.13 Ϯ 0.03) has previously been
observed using (6R)-[6-2H]EPSP with the Escherichia coli enzyme (19). This effect
can be most easily interpreted as a small primary effect expressed on R C(6)–2H
bond-cleavage. However, with this substrate, it is not possible to distinguish between
this interpretation and an alternative, involving a secondary vinylogous ␤ effect
expressed on C(3)–O cleavage. This paper describes the synthesis of [4-2H]EPSP
and the observation of a kinetic isotope effect using this substrate with both Neurospora
crassa and E. coli chorismate synthases. This effect is best rationalized as a secondary
␤ isotope effect, providing new evidence for the order of bond-cleavage.

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
193
EXPERIMENTAL
Materials and Methods
All chemicals and biochemicals were of the highest grade available and unless
otherwise stated were purchased from Sigma Chemical Company (Poole, Dorset,
UK). Sodium dithionite was purchased from B.D.H. Chemicals. The potassium salt
of protio-EPSP was prepared as described previously (19). All spectrophotometric
and spectrofluorimetric measurements were obtained with a 1.0-cm path length. Deut-
eriated buffers were made up in 99.9% D2O and the pD adjusted with KOD or DCl
using a Russell combined pH/reference cell electrode coupled to an alpha 200 pH
and conductivity meter at 25ЊC. pD values are quoted as pD ϭ meter reading ϩ 0.4.
Enzymes
N. crassa chorismate synthase was purified by the method of White et al. (4) and
recombinant E. coli chorismate synthase was purified as reported previously (19).
The E. coli protein concentration was estimated using ␧ ϭ 21,440 MϪ1 cmϪ1 per
280
subunit (20). Glycerol was added to substrate solutions to match the final glycerol
concentration in enzyme solutions used for stopped-flow spectrophotometry/fluorime-
try experiments to avoid mixing artefacts. All experiments with E. coli chorismate
synthase were performed at 25ЊC and solutions were buffered using 50 mM Mops/
KOH, pH 7.5, unless otherwise stated. One unit of enzyme is defined as the amount
of enzyme required to convert 1 ␮mol of substrate to product in 1 min.
Chorismate Synthase Assays
N. crassa chorismate synthase was assayed either directly by uv spectrophotometry
(21) or using a coupled spectrofluorimetric assay (3). The uv spectrophotometric
assay involved monitoring the formation of chorismate at 275 nm (␧ ϭ 2630
MϪ1cmϪ1). Stock solutions of FMN (1 mM) and NADPH (2 mM) were freshly
prepared on the day of each series of kinetic experiments and were stored in the dark
on ice. Each assay was carried out in a 1-ml quartz cuvette containing NADPH (20
␮M), FMN (10 ␮M), EPSP (50 ␮M), N. crassa chorismate synthase (typically 5
mU), Bis–Tris–HCl or triethanolamine–HCl buffer (50 mM, pH 7.0) and 50 mM
KCl. Assays were initiated by the addition of NADPH. For consistent isotope effect
measurements, stock solutions were prepared containing all the required ingredients
except for NADPH. A total of nine determinations with EPSP and 15 with [4-2H]EPSP
were made. A kinetic isotope effect and its standard error was calculated from the
means of the protio and deuterio data and their standard errors.
The fluorimetric assay of the N. crassa enzyme (3) involved coupling the production
of chorismate with the production of anthranilate using anthranilate synthase. The
formation of anthranilate was monitored using excitation and emission wavelengths
of 313 and 390 nm, respectively. The fluorimeter was calibrated using anthranilate
standards in Bis–Tris–HCl buffer (50 mM, pH 7.0). The detected fluorescence was
not affected by the presence of either NADPH or FMN. Each assay was carried out
in a 1-ml quartz fluorimeter cuvette containing NADPH (200 ␮M), FMN (10 ␮M),
EPSP (20 ␮M), glutamine (10 mM), N. crassa chorismate synthase (0.2 mU), N.
crassa anthranilate synthase (4 mU), Bis–Tris–HCl buffer (50 mM, pH 7.0), KCl

194
BORNEMANN ET AL.
(50 mM), and MgCl2 (5 mM). Ten times the NADPH concentration used in the UV
assay was required to ensure it was saturating throughout the course of the assay
due to the presence of a diaphorase activity in the partially purified anthranilate
synthase preparation.
E. coli chorismate synthase was assayed directly by uv spectrophotometry as
previously described (22) and substrate concentrations were determined using this
method. Anaerobic assay mixtures contained substrate (50 ␮M), FMN (10 ␮M),
potassium oxalate (1 mM), and buffer. The FMN was photoreduced, with oxalate
being the electron source, before the addition of the E. coli enzyme giving a final
volume of 500 ␮l.
3-[2-2H2,4-2H]Dehydroquinic Acid
3-Dehydroquinic acid, synthesized by the method of Grewe and Haendler (23),
(100 mg, 0.53 mmol) was dissolved in deuterium oxide (D2O, 7.5 ml) and the solution
adjusted to pD of 9.5 with 4-(dimethylamino)pyridine (DMAP). The reaction mixture
was left to exchange at room temperature for 7 days. The reaction was terminated
when more than 95% exchange had occurred as seen by monitoring the disappearance
1
of the C-4 hydrogen signal in the H NMR spectrum. The reaction mixture was
repeatedly washed with CH2Cl2 to remove the DMAP. The solution was adjusted to
neutral pH and freeze dried. The product was used without further purification. ␦H
(250 MHz, D2O) 4.3 (residual signal, 4-H), 3.90 (1 H, dd, J 13.2, 5.8 Hz, 5-H), 2.53
(approx. 0.4 H, d, J 14.2 Hz, 2eq-H), 2.30 (2 H, m, 6-H2), other peaks were obscured
by HOD suppression and signals arising from DMAP were 8.00 (2 H, d), 6.80 (2 H,
d), and 3.10 (6 H, s). The relative percentage deuteriation at C-2ax:C-2eq:C-4 was
99:40:85%. ␦D (61.4 MHz, H2O) 4.3 (1 D, broad s, 4-D), 3.1 (1 D, broad s, 2ax-D),
and 2.5 (approx. 0.6 D, broad s, 2eq-D). m/z ES: 191 (MϪ-H) (2eq-H is approx. 40%
deuteriated and therefore C7H8D2O6 requiring 192 is expected to be the most common
ion); CI: 211.1020 (C7H11D3O6N+ requires 211.1006).
3-[4-2H]Dehydroshikimic Acid
3-[2-2H2,4-2H]Dehydroquinate (approx. 110 ␮mol) was dissolved in triply distilled
MilliQ water (1.64 ml) and E. coli dehydroquinase type I (1.64 U) and A. nidulans
dehydroquinase type II (1.97 U) were added to the solution. The reaction mixture
was stirred at room temperature and the reaction monitored by 1H NMR spectroscopy.
Although the conversion to the product was complete within 2 h the reaction was
left a further 5 days to ensure at least 85% removal of both the deuterium atoms at
the C-2. The enzyme was removed by filtration through an Amicon Centricon 10.
The product was used without further work up. ␦H (400 MHz, D2O, pD 7.0) 6.35 (1
H, d, J 3.2 Hz, 2-H), 3.93 (1 H, dd, J 10.0, 5.3 Hz, 5-H), 3.03 (1 H, dd, J 18.3, 5.3
Hz, pro-S 6-H), 2.60 (1 H, ddd, J 18.3, 10.0, 3.2 Hz, pro-R 6-H) and signals arising
from DMAP were 8.00 (2 H, d), 6.80 (2 H, d), and 3.10 (6 H, s). The residual
percentage deuteriation at C-2 was 15 Ϯ 3%.
[4-2H]Shikimic Acid
To a final buffered solution of 3-[4-2H]dehydroshikimate (12 mg, approx. 70 ␮mol)
in triethanolamine (100 mM, pH 6.5) was added NADPH (52 mg, 75 ␮mol) and

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
195
shikimate dehydrogenase (12 U). The reaction mixture was stirred at room temperature
for 24 h and the formation of shikimic acid was monitored by 1H NMR spectroscopy.
Ϫ
The reaction mixture was loaded directly onto a DEAE-Sephacel column (HCO3
form: 12 ϫ 2 cm) at 1 ml minϪ1, which was washed with 5 mM triethanolamine (20
ml) and eluted with a linear gradient of 5 to 500 mM triethanolamine (100 ϩ 100
ml), collecting 3.5-ml fractions. Fractions were analyzed for shikimic acid by HPLC
using a preparative Organic Acids column with an eluent of 50 mM HCO2H. Fractions
containing the product were pooled and lyophilized to yield a white solid. This was
further purified by HPLC, to give [4-2H]shikimic acid (9 mg, 51 ␮mol). ␦H (400
MHz, D2O) 6.70 (1 H, dd, J 4.0, 1.6 Hz, 2-H), 4.40 (1 H, d, J 4.0 Hz, 3-H), 4.03
(1 H, dd, J 6.6, 5.3 Hz, 5-H), 2.75 (1 H, dd, J 18.0, 5.3 Hz, pro-S 6-H), 2.20 (1 H,
1
ddd, J 18.0, 6.6, 2.0 Hz, pro-R 6-H). Extent of deuteriation by H NMR approx.
90%. m/z ES: 174 (MϪ-H) (C7H9O5D requires 175); CI: 193.0935 (C7H13O5DN+
requires 193.0934), extent of deuteriation 85%.
5-Enolpyruvyl-[4-2H]shikimate-3-phosphate ([4-2H]EPSP)
A solution was prepared (1.08 ml) containing [4-2H]shikimic acid (9 mg, 51 ␮mol),
ATP (K+ salt, 60 mM), phosphoenolpyruvate (PEP; K+ salt, 60 mM), MgCl2 (50
mM), and Tris–HCl buffer (300 mM, pH 7.5). Purified E. coli shikimate kinase (0.16
U) and EPSP synthase (0.08 U) were added, and the reaction mixture was incubated
at 30ЊC for 24 h. Apyrase (Sigma grade VII, 1.6 U) and NaCl (ϳ0.5 mg) were added,
and the reaction mixture was incubated overnight at 30ЊC. After this treatment with
apyrase to hydrolyze the ATP and ADP to AMP, the solution was loaded onto a Mono
Q column and eluted with a linear gradient of 5 to 750 mM triethanolamine. Fractions
were collected with continuous monitoring at 240 nm. EPSP eluted at 600 mM
triethanolamine, and pooled fractions were lyophilized to give a highly deliquescent
white solid, which, while still containing triethanolamine, was shown by chorismate
synthase assay to contain 40 ␮mol EPSP (78%). HPLC retention time 9.0 min
(semipreparative Organic Acids column 50 mM H2SO4, 0.6 ml minϪ1). ␦H (D2O, K+
salt) 6.52 (1 H, broad s, 2-H), 5.22 (1 H, d, J 2.6 Hz, pro-E 8-H), 4.73 (1 H, d, J
2.6 Hz, pro-Z 8-H), 4.46 (1 H, dd, J 7.8, 4.4 Hz, 5-H), 3.02 (1 H, dd, J 17.8, 4.4
Hz, pro-S 6-H), 2.18 (1 H, dd, J 17.8, 7.8 Hz, pro-R 6-H), and the 3-H signal was
obscured by HOD. Extent of C-4 deuteriation approx. 90%. ␦D (61.4 MHz, H2O) 4.1
(1 D, s, 4-D). ␦P [101 MHz, D2O, pD 8.0, P(OMe)3] Ϫ136.3. m/z ES: 326 (M+ϩ H).
Stopped-Flow Spectrophotometry
Single-turnover experiments were carried out using a Hi-Tech Scientific SF-61
DX2 stopped-flow spectrophotometer (Salisbury, UK) installed and operated in an
anaerobic glove box with thermostatting of the observation cell Ϯ0.1ЊC. The dead
time of the instrument was measured to be 1.7 ms and only data after this time are
shown with zero time, indicating the moment of mixing. Absorbance was monitored
at 400 nm using a Xe lamp light source. Anaerobic substrate solution was mixed
with anaerobic E. coli enzyme solution containing FMN that was reduced using
dithionite. After mixing, the reaction mixtures contained substrate (20 ␮M), E. coli
enzyme (30 ␮M), reduced FMN (30 ␮M), dithionite (1 mM) and buffer. Data from
1.7 to 945 ms after mixing were collected and analyzed. Kinetic data were fitted

196
BORNEMANN ET AL.
using the Hi-Tech Scientific KinetAsyst 2.1 software. On one occasion, a total of
five determinations with EPSP followed by five with [4-2H]EPSP were made within
6 min. Kinetic isotope effects and their standard errors on the formation and decay
of the flavin intermediate were calculated from the means of the protio and deuterio
data and their standard errors. This was repeated on another occasion. The two
independent determinations were combined to give the quoted values.
Stopped-Flow Spectrofluorimetry
Inorganic phosphate dissociation was measured in the presence of a fluorescent,
N-[2-(1-maleimidyl)ethyl]-7-(diethylamino)coumarin-3-carboxamide-modified
A197C mutant phosphate binding protein (MDCC-PBP) in real-time using stopped-
flow spectrofluorimetry (24, 25). Fluorescence associated with MDCC-PBP was moni-
tored using a Hg light source, an excitation wavelength of 438 nm, and an emission
filter with a 455-nm cut-off. Anaerobic substrate solution was mixed with anaerobic
E. coli enzyme solution containing FMN that was reduced using dithionite. All
solutions contained MDCC-PBP as well as a “phosphate mop,” to remove adventitious
inorganic phosphate prior to each experiment. The “phosphate mop” consisted of
purine nucleoside phosphorylase and 7-methylguanosine. After mixing, the reaction
mixtures contained substrate (4 ␮M), E. coli enzyme (6 ␮M), reduced FMN (6 ␮M),
dithionite (500 ␮M), MDCC-PBP (12 ␮M), purine nucleoside phosphorylase (1 U
mlϪ1), 7-methylguanosine (1 mM), and buffer. Data from 1.7 to 466 ms after mixing
were collected and analyzed. Kinetic isotope effects were determined and calculated
as above.
RESULTS
Synthesis of [4-2H]EPSP
The strategy used to synthesize EPSP labeled with deuterium specifically at C-4
is shown in Scheme 2. This involved the initial preparation of [4-2H]shikimate,
SCHEME 2

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
197
which in turn was made from 3-dehydroquinate. Unlabeled 3-dehydroquinic acid was
synthesized by the method of Grewe and Haendler (23). Base-catalyzed exchange of
3-dehydroquinate in D2O resulted in introduction of deuterium not only at C-4, but
also at both the axial pro-S and equatorial pro-R C-2 positions. A number of exchange
conditions were explored, and in each case the axial C-2 position exchanged most
rapidly, followed by the C-4 hydrogen. The best conditions found were at pD 9.5
using 4-(dimethylamino)pyridine (DMAP). The reaction was monitored by following
1
the disappearance of the C-4 hydrogen signal in the H NMR spectrum. Exchange
took place slowly at room temperature over several days. The reaction was stopped
when approximately 90% exchange had occurred. At this stage the extent of deuteria-
tion was approximately 99, 40, and 85% at C-2ax, C-2eq, and C-4, respectively.
In order to study the effect of deuteriation at C-4 on the chorismate synthase
reaction, it was essential that deuterium was only present at this position, and not at
C-2. In order to remove the deuterium specifically from C-2 of dehydroquinate, the
regiospecificity of dehydroquinase was exploited. At first sight, this does not appear
a very promising strategy, since incubation of [2-2H2,4-2H]dehydroquinate with type
I dehydroquinase in water would lead to formation of an equilibrium mixture of
(6S)-[2-2H,4-2H]dehydroquinate and [2-2H,4-2H]dehydroshikimate (1:15); the residual
deuterium at C-2 being due to the stereospecific loss of deuterium from only the
pro-R C-2 position in the type I dehydroquinase reaction. However, the type II
dehydroquinase catalyzes the same conversion of 3-dehydroquinate to 3-dehydroshiki-
mate, but by a different mechanism (26) and with stereospecific loss of the pro-S C-
2 hydrogen (27). By incubating [2-2H2,4-2H]dehydroquinate simultaneously with type
I and type II dehydroquinase in water, all the deuterium at C-2 was exchanged. The
reaction was followed by monitoring the increase in intensity of the C-2 hydrogen
in the [4-2H]dehydroshikimate produced. This product is not very stable and so was
immediately reduced to [4-2H]shikimate, using shikimate dehydrogenase and NADPH.
The [4-2H]shikimate was purified using two chromatographic steps. The deuterium
1
was shown to be exclusively at C-4 by H NMR spectroscopy, and the extent of
deuteriation was determined to be 85% by mass spectrometry.
The [4-2H]shikimate was then converted into [4-2H]shikimate 3-phosphate using
shikimate kinase and ATP. This was then converted into [4-2H]EPSP using EPSP
synthase and phosphoenolpyruvate (PEP). The presence of deuterium at C-4 in the
final product was confirmed by 1H and 2H NMR spectroscopy and mass spectrometry.
This synthetic approach using 6 steps from quinic acid to [4-2H]EPSP compares
favorably to an alternative from arabinose that would have required at least 12 (28).
4D(V) Kinetic Isotope Effects with N. crassa and E. coli Chorismate Synthases
The chorismate synthases from E. coli and N. crassa have been studied most
extensively. The E. coli enzyme has the advantage that it has been over-expressed
as a recombinant enzyme (19), providing sufficient material for rapid kinetic and
spectroscopic studies. On the other hand, the N. crassa enzyme has the advantage
that it has an additional NADPH-dependent flavin reductase activity (3, 4), enabling
it to be assayed without the need for chemical reductants under strictly anaerobic
conditions. More detailed comparisons between the two enzymes have been made
elsewhere (19).

198
BORNEMANN ET AL.
Using the uv spectrophotometric assay, where the formation of chorismate was
monitored directly, a 4D(V) kinetic isotope effect of 1.08 Ϯ 0.01 was observed with
the N. crassa enzyme. This effect was small but significant. The fluorimetric coupled
assay, where the formation of anthranilate was monitored in the presence of anthranilate
synthase, was not precise enough to reveal the isotope effect observed using the direct
assay. Similarly, an isotope effect was not detectable with the E. coli enzyme (1.00
Ϯ 0.05) using a modified direct uv spectrophotometric assay, because the added
requirement for strict anaerobicity with this enzyme led to (V) determinations with
relatively large errors.
4Dk Kinetic Isotope Effects with the E. coli Enzyme during Single-
Turnover Reactions
Transient kinetic studies with the E. coli enzyme have shown that partially rate
limiting steps occur after the potentially isotope-sensitive chemical steps in the cata-
lytic cycle (11). Therefore, any intrinsic kinetic isotope effect would be suppressed
in (V), providing one explanation why 4D(V) was undetectable with the enzyme from
this source. The presence of an isotope effect on the chorismate synthase reaction
was therefore determined using single-turnover stopped-flow experiments. There are
two processes that can be monitored with sufficient potential accuracy and precision
to allow the determination of these small effects: inorganic phosphate dissociation
from the enzyme and the transient formation of a flavin intermediate.
The release of inorganic phosphate from the E. coli enzyme during a single-
turnmover experiment with EPSP and [4-2H]EPSP, as monitored using the fluorescent
probe MDCC-PBP (24, 25), is shown in Fig. 1. It would appear that there is an
isotope effect on phosphate release. The process with EPSP can be fitted using two
single exponentials describing a small lag of 210 Ϯ 20 sϪ1, with a small negative
amplitude, and a main phase of 63 Ϯ 2 sϪ1, with a large positive amplitude (11).
Rapid chemical quench experiments with single turnover reactions have previously
shown that the rate of the main phase mirrors that of EPSP consumption, and therefore
a chemical step. The lag is difficult to discern because of its small amplitude and the
first 1.7 ms being obscured by the dead time of the equipment. However, the two
exponentials can be resolved because their amplitudes have opposite signs. This is
illustrated by the residuals shown in Fig. 1. Those associated with the best fit lines
of the EPSP (Fig. 1a) and [4-2H]EPSP (Fig. 1c) data using two exponentials indicate
good fits, while those using a single exponential (Figs. 1b and 1d) show a clear
deterioration of the fits.
Fitting of the individual time courses using two exponentials gave isotope effects
of 1.07 Ϯ 0.10 on the lag and 1.10 Ϯ 0.02 on the main phase. In order to illustrate
the reproducibility of the quantification of the effect, the quoted value for the main
phase is the mean of very similar values obtained on separate occasions; 1.09 Ϯ 0.02
(n ϭ 5) and 1.11 Ϯ 0.02 (n ϭ 5). There is clearly a substantial ␤-effect on the main
phase. However, there is a large error in the measured isotope effect on the lag,
mainly due to the small amplitude of this phase. By constraining the fits to give an
unchanged rate constant for either the lag or the main phase, maximum isotope effects
of either 1.5 Ϯ 0.1 on the lag or 1.11 Ϯ 0.01 on the main phase are obtained. It is

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
199
FIG. 1. A kinetic isotope effect on phosphate release under single-turnover conditions. A small but
significant isotope effect is observed on the release of phosphate with [4-2H]EPSP during a single-turnover
stopped-flow experiment. After mixing, the anaerobic reaction mixtures contained substrate (4 ␮M), E.
coli enzyme (6 ␮M), reduced FMN (6 ␮M), dithionite (500 ␮M), MDCC-PBP (12 ␮M), purine nucleoside
phosphorylase (1 U mlϪ1), 7-methylguanosine (1 mM) and Mops buffer, pH 7.5, at 25ЊC. The curves
shown are averages of all determinations and the observed data are normalized from 0 to 1. The best fit
lines using two exponentials (smooth lines) are essentially coincident with the data. The residuals are
associated with: (a) the best fit of the EPSP data; (b) a fit of the EPSP data using a single exponential,
showing a deterioration of the fit; (c) the best fit of the [4-2H]EPSP data; (d) a fit of the [4-2H]EPSP
data using a single exponential, showing a deterioration of the fit.
not possible to satisfactorily distinguish between the statistical quality of these con-
strained fits. Therefore, it is still not clear whether the effect was also expressed in
the lag. However, an effect of 1.5 Ϯ 0.1 is an unfeasibly large secondary ␤ effect of
this type, while one of 1.11 Ϯ 0.01 is within the normal range of up to 1.15 (29).
This strongly points to an effect on the main phase of 1.10 Ϯ 0.02.
A flavin intermediate, which has an increased absorbance at 400 nm, is transiently
formed during the reaction (30). This intermediate is produced by the protonation of
the reduced anionic enzyme-bound FMN (31) after binding of the substrate and before
any bonds are cleaved (11). After the bond-cleavage steps and phosphate release, the
reduced FMN decays back to the anionic form. Given the observation of an isotope
effect on phosphate release, an effect should be observed on the decay of the flavin
intermediate, but not on its formation. The magnitude of the effect would also be
expected to be smaller than that on phosphate release due to the increased suppression
of intrinsic isotope effects manifested in steps later in the catalytic cycle.
Figure 2 shows the formation and decay of the flavin intermediate during single-
turnmover experiments using the E. coli enzyme with EPSP and [4-2H]EPSP. It has

200
BORNEMANN ET AL.
FIG. 2. A kinetic isotope effect on the decay of the flavin intermediate under single-turnover condi-
tions. A small kinetic isotope effect with [4-2H]EPSP is observed on the decay of the protonated form
of reduced flavin during a single-turnover stopped-flow experiment. After mixing, the anaerobic reaction
mixtures contained substrate (20 ␮M), E. coli enzyme (30 ␮M), reduced FMN (30 ␮M), dithionite (1
mM), and Mops buffer, pH 7.5, at 25ЊC. The curves shown are averages of all determinations. The
residuals are associated with: (a) a best fit of the EPSP data; (b) a fit where the second exponent was
constrained to be that for [4-2H]EPSP, showing a deterioration of the fit; (c) a best fit of the [4-2H]EPSP
data; (d) a fit where the second exponent was constrained to be that for normal EPSP, showing a
deterioration of the fit.
previously been shown that with EPSP, the kinetics can be described by two single
exponentials, kobs ϭ 160 Ϯ 10 and 52 Ϯ 2 sϪ1 describing its formation and decay,
respectively (19). [4-2H]EPSP causes a perturbation in the kinetic profile which, as
expected, appears to be entirely on the decay (Fig. 2). A further indication that this
is the case comes from comparing free and constrained fits of the data. The residuals
associated with the best fit lines for the EPSP (Fig. 2a) and [4-2H]EPSP (Fig. 2c)
data indicate reasonably good fits. The residuals associated with constrained fits,
where the second exponent was forced to be that for the other respective substrate,
show a small deterioration of the fits (Figs. 2b and 2d). This is consistent with the
effect being on the formation alone.
In order to quantify the effect, each individual time course was fitted using two
exponentials. Confidence in the fits comes from the fact that the exponentials for the
formation and decay had essentially equal and opposite amplitudes in all cases. The
kinetic isotope effects were 0.94 Ϯ 0.09 and 1.06 Ϯ 0.03 for the formation and decay,
respectively. The large error in the measured effect on the formation is due to the
fact that most of this rapid process is obscured by the dead time of the equipment

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
201
(1.7 ms). Although there is a good indication of an isotope effect on the decay, as
expected, its magnitude is smaller than that on phosphate release.
DISCUSSION
A small but significant kinetic isotope effect has been detected using [4-2H]EPSP
with the chorismate synthase reaction. Evidence comes from the phosphate release
experiments with the E. coli enzyme with a (V) effect having also been detected with
the N. crassa enzyme. Since the C(4)–2H bond is not broken during the reaction, the
effects are best explained by hyperconjugative secondary ␤ effects. This would result
from a build up of substantial electron deficient carbocation or radical (32) character
at C-3 in a transition state associated with one of the rate limiting chemical steps of
the catalytic cycle.
To our knowledge, there are no published values of deuterium kinetic isotope
effects ␤ to the leaving group in an elimination reaction, whether it be initiated by
radical chemistry or not. However, such isotope effects are well understood for normal,
nonradical substitution reactions. This is helpful since the first step, and therefore
also the first transition state, is the same for SN1 and E1 reactions. One of the best
criteria for an SN1 reaction mechanism is the observation of an ␣ or ␤ secondary
deuterium kinetic isotope effect of Ն 1.07, with values up to 1.15 being possible
(29). This criterion can therefore be used for E1 reactions. By contrast, values of Յ
1.05 are associated with SN2 reactions. Although the transition state for a concerted
E2 reaction is not the same as that for SN2 reactions, similar values would be expected
because they both share a small extent of electron deficiency at the carbon bearing
the leaving group. Finally, the smallest effects would be expected with an E1cB
mechanism involving the initial rate limiting cleavage of the C–H bond followed by
the rapid loss of the leaving group from an anionic intermediate. Since kinetic isotope
effects at C-4 of 1.10 Ϯ 0.02 and 1.08 Ϯ 0.01 were determined for phosphate release
with the E. coli enzyme and for (V) with the N. crassa enzyme, respectively, an
E1, or possibly E2, mechanism would be the most likely if no radical chemistry
were involved.
The magnitude of the isotope effect for the preferred radical mechanism would be
expected to be somewhat smaller than that for a normal E1 mechanism (1.07–1.15).
This is due to the smaller contribution of hyperconjugation in the stabilization of
radicals compared with cations. However, if the reaction proceeded with the radical
mechanism, the isotope effect would be strongly expressed on C(3)–O bond-cleavage
because it would be expected be the first, discrete, and rate limiting bond breaking
step. It is therefore difficult to distinguish between the radical and cation mechanisms
from the magnitudes of the observed isotope effects alone.
The magnitude of a secondary ␤ effect is determined by equation (1) (33):
ln(kH/kD) ϭ cos2␪.ln(kH/kD)max ϩ ln(kH/kD)i
(1)
The first hyperconjugative term is geometry-dependent where the effect is maximal
when the electron deficient p orbital is eclipsed with the C–2H bond and increases
with increasing carbocation or radical character on the adjacent carbon. The second,
much smaller inductive term is geometry independent and in the opposite sense to

202
BORNEMANN ET AL.
FIG. 3. A half-chair conformation of EPSP. The C-3 phosphate and C-6 pro-R hydrogen are pseudo
axial in this half-chair conformation, which is appropriate for the trans-1,4 elimination reaction.
the first term. It is most probable that EPSP is in a half-chair conformation such that
the bonds to be broken are pseudo-axial (Fig. 3). It therefore follows that the transition
state associated with C(3)–O bond-cleavage would essentially be in such a conforma-
tion. Shikimate-3-phosphate and the pyruvoyl ketal of EPSP have been determined
by NMR spectroscopy and MM2 calculations to be in a half-chair conformation in
solution with the phosphate pseudo-axial (34).
If the reaction proceeded with a stepwise E1 or radical mechanism, such as those
shown in Scheme 1, secondary isotope effects would be expected at C-3, -4, and -6,
with the smallest vinylogous effect being at pseudo-equatorial S C-6. This is the case
(Table 1). The moderate to large secondary effects at both C-4 and -3 indicate that
the C(3)–O bond is substantially broken in the transition state. Although no significant
effect at S C-6 was detected above the experimental error, it may not be completely
absent. The effect at R C-6 should have a secondary ␤ component, but an additional
primary component is also possible if its subsequent bond-cleavage were partially
rate limiting. This is certainly the case with the N. crassa enzyme (15, 35).
If, on the other hand, the reaction proceeded with a concerted E2 mechanism,
secondary isotope effects would be expected at C-3, -4, and -6 with a concomitant
TABLE 1
Summary of Kinetic Isotope Effects on the Chorismate Synthase Reaction
Enzyme
N. crassa D(V)
D(V/K)
Isotope effect
C-3 2Њ ␣
C-4a 2Њ ␤
R C-6 1Њ/2Њ ␤
S C-6 2Њ ␤
1.10 Ϯ 0.02b
1.08 Ϯ 0.01a
2.7 Ϯ 0.2c
2.64 Ϯ 0.02b
1.6 Ϯ 0.1c
1.22 Ϯ 0.04b
ndd
ndd
0.97 Ϯ 0.02b
ndd
ndd
ndd
ndd
T(V/K)
1.047 Ϯ 0.012e
E. coli
D(V)
ndd
ndd
ndd
1.00 Ϯ 0.05a
1.06 Ϯ 0.03a
1.10 Ϯ 0.02a
1.13 Ϯ 0.03f
1.17 Ϯ 0.04f
1.27 Ϯ 0.03h
1.01 Ϯ 0.03f
1.03 Ϯ 0.03g
ndd
Dk(FMN)i
Dk(Pi)j
Note. (a) This work; (b) (15); (c) (35); (d) not determined; (e) (10); ( f ) (19); (g) Determined by
Bornemann et al. (19), but value unpublished; (h) (11); (i) Kinetic isotope effect on the decay of the
transient protonated reduced flavin species during a single-turnover; (j) Kinetic isotope effect on the
main phase of the release of inorganic phosphate during a single-turnover. Quoted errors are those as
given in each reference.

ISOTOPE EFFECT WITH CHORISMATE SYNTHASE
203
primary effect at R C-6. In this case, the moderate to large secondary effects at C-4
and -3 and the small to moderate primary effects at R C-6 (Table 1) indicate not only
that the C(3)–O bond is substantially broken in this transition state but also that the
proR C(6)–H bond is not. Normal elimination reaction mechanisms can fall anywhere
within the E1–E2–E1cB spectrum, where the bond cleavages are simultaneous in the
pure E2 reaction. Therefore, if the reaction were E2, it would have some E1 character
and tend toward that shown in Scheme 1b.
The isotope effects observed using [4-2H]EPSP with the enzymes from both sources
suggest that C(3)–O cleavage precedes C(6)–H cleavage. Although it is not possible
to clearly distinguish between the radical and cation mechanisms (Scheme 1) from
these isotope effects alone, they do provide evidence against alternative mechanisms.
Since the overall reaction is effectively irreversible and (V/K) isotope effects are
expressed at both C-3 and C-6 (Table 1), the second step must be the common
irreversible chemical step (10). It therefore follows that it is most likely that C(6)–H
cleavage is irreversible. Both the proR C-6 and C-4 hydrogens are pseudo-axial and
equally capable of leaving the radical or carbocation intermediate to give different
isomeric products. The enzyme must therefore tightly control the deprotonation at
proR C-6 with an appropriately positioned active site base. Postchemical steps, such
as those associated with chorismate release, are likely to be more rapid with the N.
crassa enzyme because the (V) isotope effects are expressed more strongly with this
enzyme. Therefore, although the E. coli and N. crassa enzymes seem to share a
common chemical mechanism, they have different rate limiting steps.
ACKNOWLEDGMENTS
We gratefully acknowledge the financial support of the Biotechnology and Biological Sciences Research
Council, for a postdoctoral award to M.-E.T. and the Competitive Strategic Grant to the JIC, and the
Medical Research Council. We are grateful to Professor John R. Coggins for the over-expressing E. coli
strain, to Dr. Trevor Rayment for advice on statistics, and to Dr. David J. Lowe for helpful discussions.
REFERENCES
1. Macheroux, P., Schmid, J., Amrhein, N., and Schaller, A. (1999) Planta 207, 325–334.
2. Morell, H., Clark, M. J., Knowles, P. F., and Sprinson, D. B. (1967) J. Biol. Chem. 242, 82–90.
3. Welch, G. R., Cole, K. W., and Gaertner, F. H. (1974) Arch. Biochem. Biophys. 165, 505–518.
4. White, P. J., Millar, G., and Coggins, J. R. (1988) Biochem. J. 251, 313–322.
5. Hill, R. K., and Newkome, G. R. (1969) J. Am. Chem. Soc. 91, 5893–5894.
6. Onderka, D. K., and Floss, H. G. (1969) J. Am. Chem. Soc. 91, 5894–5896.
7. Floss, H. G., Onderka, D. K., and Carroll, M. (1972) J. Biol. Chem. 247, 736–744.
8. Bartlett, P. A., McLaren, K. L., Alberg, D. G., Fa¨ssler, A., Nyfelder, R., Lauhon, C. T., and Grissom,
C. B. (1989) in Proceedings of the Society for Chemical Industry Pesticides Group Meeting, Vol.
BCPC Monograph 42 (Copping, L. G., Dalziel, J., and Dodge, A. D., Eds.), pp. 155–170, Lavenham
Press, Suffolk, UK.
9. Bornemann, S., Lowe, D. J., and Thorneley, R. N. F. (1996) Biochem. Soc. Trans. 24, 84–88.
10. Balasubramanian, S., Coggins, J. R., and Abell, C. (1995) Biochemistry 34, 341–348.
11. Bornemann, S., Lowe, D. J., and Thorneley, R. N. F. (1996) Biochemistry 35, 9907–9916.
12. Bornemann, S., Ramjee, M. K., Balasubramanian, S., Abell, C., Coggins, J. R. Lowe, D. J., and
Thorneley, R. N. F. (1995) J. Biol. Chem. 270, 22811–22815.
13. Ramjee, M. N., Balasubramanian, S., Abell, C., Coggins, J. R., Davies, G. M., Hawkes, T. R., Lowe,
D. J., and Thorneley, R. N. F. (1992) J. Am. Chem. Soc. 114, 3151–3153.

204
BORNEMANN ET AL.
14. Macheroux, P., Bornemann, S., and Thorneley, R. N. F. (1997) in Flavins and Flavoproteins 1996
(Stevenson, K. J., Massey, V., and Williams Jr, C. H., Eds.), pp. 113–122, Univ. Calgary Press,
Calgary, Canada.
15. Lauhon, C. T., and Bartlett, P. A. (1994) Biochemistry 33, 14100–14108.
16. Bornemann, S., Coggins, J. R., Lowe, D. J., and Thorneley, R. N. F. (1995) in Perspectives on Protein
Engineering and Complementary Technologies 1995 (Geisow, M. J., and Epton, R., Eds.), pp.
134–135, Mayflower Worldwide, Birmingham, UK.
17. Giese, B., and Almstead, N. G. (1994) Tetrahedron Lett. 35, 1677–1680.
18. Theoclitou, M. E., Duggan, P. J., and Abell, C. (1996) Bioorg. Med. Chem. Lett. 6, 1285–1288.
19. Bornemann, S., Balasubramanian, S., Coggins, J. R., Abell, C., Lowe, D. J., and Thorneley, R. N.
F. (1995) Biochem. J. 305, 707–710.
20. Mach, H., Middaugh, C. R., and Lewis, R. V. (1992) Anal. Biochem. 200, 74–80.
21. Gibson, F. (1970) Methods Enzymol. 17, 362–364.
22. Ramjee, M. K., Coggins, J. R., and Thorneley, R. N. F. (1994) Anal. Biochem. 220, 137–141.
23. Grewe, R., and Haendler, H. (1968) Biochem. Prep. 11, 21–26.
24. Brune, M., Hunter, J. L., Corrie, J. E. T., and Webb, M. R. (1994) Biochemistry 33, 8262–8271.
25. Brune, M., Hunter, J. L., Howell, S. A., Martin, S. R., Hazlett, T. L., Corrie, J. E. T., and Webb, M.
R. (1998) Biochemistry 37, 10370–10380.
26. Harris, J. M., GonzalezBello, C., Kleanthous, C., Hawkins, A. R., Coggins, J. R., and Abell, C.
(1996) Biochem. J. 319, 333–336.
27. Harris, J., Kleanthous, C., Coggins, J. R., Hawkins, A. R., and Abell, C. (1993) J. Chem. Soc. Chem.
Commun. 1080–1081.
28. Guilford, W. J., Copley, S. D., and Knowles, J. R. (1987) J. Am. Chem. Soc. 109, 5013–5019.
29. Westaway, K. C. (1987) in Isotopes in Organic Chemistry—Secondary and Solvent Isotope Effects
(Buncel, E., and Lee, C. C., Eds.), Vol. 7 pp. 275–392, Elsevier, Amsterdam.
30. Ramjee, M. N., Coggins, J. R., Hawkes, T. R., Lowe, D. J., and Thorneley, R. N. F. (1991) J. Am.
Chem. Soc. 113, 8566–8567.
31. Macheroux, P., Bornemann, S., Ghisla, S., and Thorneley, R. N. F. (1996) J. Biol. Chem. 271, 25850–
25858.
32. Symons, M. C. R. (1962) Tetrahedron 18, 333–341.
33. Guo, X., and Sinnott, M. L. (1993) Biochem. J. 294, 653–656.
34. Castellino, S., Leo, G. C., Sammons, R. D., and Sikorski, J. A. (1991) J. Org. Chem. 56, 5176–5181.
35. Balasubramanian, S., Abell, C., and Coggins, J. R. (1990) J. Am. Chem. Soc. 112, 8581–8583.