Derailing Dehydroquinate Synthase by Introducing a Stabilizing Stereoelectronic Effect in a Reaction Intermediate

Full text of DOI:10.1021/jo9713060

8
582
J . Org. Chem. 1997, 62, 8582-8585
Der a ilin g Deh yd r oqu in a te Syn th a se by
In tr od u cin g a Sta bilizin g Ster eoelectr on ic
Effect in a Rea ction In ter m ed ia te
Sch em e 1. Mech a n ism for th e Deh yd r oqu in a te
a
Syn th a se Ca ta lyzed Rea ction
†
‡
,†
Emily J . Parker, J ohn R. Coggins, and Chris Abell*
University Chemical Laboratory, Lensfield Road,
Cambridge CB2 1EW, England, and Department of
Biochemistry, University of Glasgow,
Glasgow G12 8QQ, Scotland
Received J uly 17, 1997
3
-Dehydroquinate synthase (E.C. 4.6.1.3) catalyses the
conversion of 2-deoxy-D-arabino-heptulosonic acid 7-phos-
phate (DAHP) (1) into 3-dehydroquinate 6. This is the
second step on the shikimate pathway to the aromatic
amino acids.1 The mechanism of this transformation
2
(
Scheme 1) was first proposed by Sprinson and substan-
3
tially confirmed by Knowles.
The complexity of the dehydroquinate synthase mech-
anism raised the question of how the oxidation and
reduction, phosphate cleavage, and aldol ring closure
reactions could be catalyzed by a small monomeric
enzyme. A series of elegant studies suggested that the
inherent reactivity of the various enzymic intermediates
reduced the role of the enzyme to little more than that
a
When DAHP 1 is the substrate only dehydroquinate 6 is
formed, and the intermediate enolpyranose 3 is cyclized on the
enzyme. When 3 is generated by deprotection of 4, a mixture of 6
4
of an “internal cycling dehydrogenase”. For example,
and 7 is formed nonenzymatically.7 The solid arrows indicate
evidence was presented to show that the base responsible
for the deprotonation of the intermediate 2 at C-6 was
the substrate phosphate dianion.5
enzymatic reactions, and the dotted arrows are nonenzymatic
transformations.
properties.9 Using a mixture of (3R)- and (3S)-3-fluo-
The involvement of the enzyme in the conversion of
the enol pyranose 3 to dehydroquinate 6 has also been
questioned. Photochemical removal of the o-nitrobenzyl
protecting group from 4 was initially reported to result
in complete conversion of 3 to dehydroquinate in the
absence of enzyme.6 Furthermore, the cyclization was
shown to proceed through the same chairlike transition
state as had been established for the enzyme-catalyzed
aldol step. This led to the suggestion that the enolpyra-
nose 3 was the product of the enzyme reaction. Upon
subsequent reinvestigation of this nonenzymatic reaction,
it was found that the reaction of 3 was not entirely
stereospecific with 2.5-4% of 1-epi-dehydroquinate 7 also
10
roDAHP (8 and 11) generated by DAHP synthase, we
observed that the (3R)-isomer was very rapidly converted
to (6R)-6-fluorodehydroquinate (10) before the (3S)-3-
fluoroDAHP was slowly converted to (6S)-6-fluorodehy-
droquinate (13) and an unknown product. We now report
on a detailed study of this reaction and identify the
unknown product as (6S)-1-epi-6-fluorodehydroquinate
(14). The formation of 13 and 14 appears to be from a
common intermediate 12 after its release from the
enzyme (Scheme 2).
(3R)-3-FluoroDAHP (8) and (3S)-3-fluoroDAHP (11)
were separately synthesized enzymatically from eryth-
rose 4-phosphate and (E)-3-fluoroPEP and (Z)-3-fluo-
7
being formed. As this product is not observed in the
11
roPEP, respectively, using DAHP synthase. The fluo-
enzymatic reaction, the role of the enzyme in guiding the
conversion of 3 to 6 was reinstated. We now show that
a modification of the enol pyranose intermediate by
putting an axial fluorine next to the hemiketal center
roDAHPs were then used in kinetic studies as substrate
analogues for dehydroquinate synthase. The kinetic
constants were determined in a coupled spectrophoto-
metric assay by using an excess of E. coli dehydro-
(e.g., in 12) promotes the dissociation of the enol pyranose
quinase to convert the 6-fluorodehydroquinates to the
intermediate from the enzyme, so that it cyclizes in
solution.
12,13
corresponding 6-fluorodehydroshikimates,
after first
confirming that (6R)- and (6S)-6-fluorodehydroquinate
are good substrates for dehydroquinase.
We have previously developed an enzymatic synthesis
8
of 6-fluoroshikimates as part of a study of their antibiotic
(
8) Duggan, P. J .; Parker, E.; Coggins, J .; Abell, C. Bioorg. Med.
*
To whom correspondence should be addressed. Tel.: +44-1223-
Chem. Lett. 1995, 5, 2347-2352.
3
36405. FAX: +44-1223-336362. E-mail: ca26@cam.ac.uk.
(9) Davies, G. M.; Barrettbee, K. J .; J ude, D. A.; Lehan, M.; Nichols,
W. W.; Pinder, P. E.; Thain, J . L.; Watkins, W. J .; Wilson, R. G.
Antimicrob. Agents Chemother. 1994, 38, 403-406.
(10) Pilch, P. F.; Somerville, R. L. Biochemistry 1976, 15, 5315-
5320.
†
University Chemical Laboratory.
University of Glasgow.
‡
(
1) Bentley, R. CRC Crit. Rev. Biochem. 1990, 25, 307-384.
(2) Sprinson, D. B.; Rothschild, M.; Sprecher, M. J . Biol. Chem. 1963,
2
38, 3170-3175.
(11) Details of the preparation and purification of 8 and 11, the
1
9
(
3) Knowles, J . Aldrichim. Acta 1989, 22, 59-66.
kinetic data for the enzymatic conversions, and F NMR spectra of
10, 13, and 14 are available as Supporting Information.
(12) Dehydroquinate synthase was purified according to the method
of: Frost, J . W.; Bender, J . L.; Kadonaga, J . T.; Knowles, J . R.
Biochemistry 1984, 23, 4470-4475.
(4) Bender, S. L.; Mehdi, S.; Knowles, J . R. Biochemistry 1989, 28,
7
1
1
5
555-7560.
5) Widlanski, T.; Bender, S. L.; Knowles, J . R. J . Am. Chem. Soc.
989, 111, 2299-2300.
6) Bartlett, P. A.; Satake, K. J . Am. Chem. Soc. 1988, 110, 1628-
630.
(
(
(13) However, these findings are at variance with the work of Le
app
Marechal who reported that the K
M
for 8 and 11 was in the range
(7) Bartlett, P. A.; Mclaren, K. L.; Marx, M. A. J . Org. Chem. 1994,
40-60 µM and the Vmax values were in the order 1 >11 > 8. Le
Marechal, P.; Froussios, C.; Azerad, R. Biochimie 1986, 68, 1211-1215.
9, 2082-2085.
S0022-3263(97)01306-6 CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 24, 1997 8583
Sch em e 2. F or m a tion of
(
6S)-6-F lu or od eh yd r oqu in a te (10) w h en
(
3R)-3-F lu or oDAHP 8 Is Used a s a Su bstr a te for
a
Deh yd r oqu in a te Syn th a se
a
However, when (3S)-3-fluoroDAHP (11) is the substrate a 2:1
mixture of (6S)-6-fluorodehydroquinate (13) and (6S)-1-epi-6-
fluorodehydroquinate (14) is formed. The solid arrows indicate
enzymatic reactions, and the dotted arrows are possible nonen-
zymatic steps.
Ta ble 1. Stea d y-Sta te Kin etic P a r a m eter s for DAHP ,
F igu r e 1. 19F NMR spectra (235 MHz) showing the conversion
of (3S)-3-fluoroDAHP (11) into a 2:1 mixture of (6S)-6-
fluorodehydroquinate (13) and (6S)-1-epi-6-fluorodehydro-
quinate (14) (a) before addition of dehydroquinate synthase,
(
3R)-3-F lu or oDAHP (8), a n d (3S)-3-F lu or oDAHP (11)
w ith DHQ Syn th a se
KMapp
kcat.
kcat./KM
app
-
1
(s-1
M
-1) × 106
(
µM)
5.7 ( 0.2 50 ( 3
1.6 ( 0.3 7.7 ( 0.6
3S)-3-fluoroDAHP (11) 21 ( 2 0.9 ( 0.1
(s
)
(
b) t ) 20 min, (c) t ) 80 min, (d) t ) 180 min. Spectra were
DAHP (1)
3R)-3-fluoroDAHP (8)
8.8
4.8
0.04
run in H O containing 10-20% D O (total volume 550 µL) and
were referenced to CCl
)
3
2
2
(
(
3
F at 0 ppm. [Dehydroquinate synthase]
+
10 µM (18 units), [(3S)-3-fluoroDAHP] ) 45 mM, [NAD ] )
0 µM, pH )7.0, Co2 50 µM, T ) 26 °C. The weak signal at
+
Incubation of (3R)-3-fluoroDAHP (8) with dehydro-
-215 ppm is due to the C-1 epimer of 11.
quinate synthase results in the rapid and quantitative
formation of (6R)-6-fluorodehydroquinate (10). Although
spectra, and NOEs observed with both compounds. In
app
the kcat. is reduced approx 7-fold relative to DAHP, K
is also smaller and the specificity constant (kcat./K
M
app)
addition, the HRMS of the sodium salts of 10, 13, and
11
14 prove that they are all isomeric.
M
for (3R)-3-fluoroDAHP is similar to that for the natural
substrate (Table 1).
By contrast, incubation of dehydroquinate synthase
with (3S)-3-fluoroDAHP (11) results in the much slower
formation of two products (6S)-6-fluorodehydroquinate
The significant differences between 13 and 14 are in
the chemical shift of the fluorine signal and the stability
of the two compounds. The chemical shift of the fluorine
in 14 comes at -209 ppm, compared to -202 ppm for 13
and -201 ppm for 10. Significantly, the chemical shift
for the fluorine in (3S)-3-fluoroDAHP is -205, while that
of its C-1 anomer, where the carboxyl is also antiperipla-
nar to the fluorine, comes at -215 ppm. Like 1-epi-
(13) and (6S)-1-epi-6-fluorodehydroquinate (14) in a 2:1
ratio (Figure 1). The specificity constant is 100 times
lower than for (3R)-3-fluoroDAHP (8), consistent with our
observation that (3R)-3-fluoroDAHP is converted into
7
dehydroquinate 7, (6S)-1-epi-6-fluorodehydroquinate 14
8
,13
product first when a mixture of isomers is present.
is not a substrate or inhibitor for E. coli dehydroquinase.
The conversion of (3S)-3-fluoroDAHP (11) to a mixture
of 13 and 14 goes to completion when sufficient enzyme
is used for the reaction to take place in 20-30 min. If
less enzyme is used, so that the reaction proceeds over
several hours, the reaction becomes strongly inhibited
and addition of more enzyme has no observable effect.
This inhibition has been ascribed to decomposition
(6S)-1-epi-6-Fluorodehydroquinate (14) decomposes over
several hours at room temperature but can be stored at
-20 °C. The decomposition was first seen as a steady
1
reduction in the signal-to-noise of the H NMR spectrum
2
of a sample in D O, with no other signals appearing. The
decomposition product is 1,2,4-trihydroxybenzene, which
becomes fully deuterated and hence is not detected
+
1
products of NAD that reversibly inhibit dehydroquinate
directly in the H NMR spectrum. The mechanism for
synthase.4
formation of this product is presumed to be by anti-
periplanar elimination of carbon dioxide and fluoride
followed by dehydration of the resulting diketo diol.
The two products 13 and 14 were separated by HPLC
1
and their identities confirmed spectroscopically. The H
NMR spectra of (6S)-6-fluorodehydroquinate (13) and
The formation of two products 13 and 14 from the
enzymatic conversion of 11 contrasts with the enzymatic
formation of only 6 from 1 but is reminiscent of the
nonenzymatic formation of a mixture of 6 and 7 from 3.
It is unlikely that dehydroquinate synthase can catalyze
(
(
6S)-1-epi-6-fluorodehydroquinate (14) are very similar
Figure 2) with almost identical J H-H and J H-F coupling
7
,14
constants, consistent with the relative configuration and
conformation of the fluorine and the hydrogens on the
ring being the same. This similarity extends to the COSY
the alternative aldol reactions to form 13 and 14 and

8
584 J . Org. Chem., Vol. 62, No. 24, 1997
Notes
The equatorial fluorine atom at C-3 might have a
similar effect on the partitioning of the enol pyranose
intermediate 9, which could also dissociate from the
enzyme. However, the conformation for cyclization in
solution favored by the Felkin-Ahn rule would lead to
1
0. This is the same cyclization as would be expected
on the enzyme, and so it is impossible to conclude from
the exclusive formation of 10 whether this cyclization has
occurred on or off the enzyme.
The steady-state kinetic parameters determined for the
three substrates give limited insight to what is happening
M
on the enzyme. The kcat./K for reaction of 11 is 200 times
slower than for DAHP (1) and about 100 times slower
than for (3R)-3-fluoroDAHP (8). However, if the elimina-
tion of phosphate from 2 is accompanied by release of
inorganic phosphate (P
should become effectively irreversible. If this is so, then
the differences in kcat./K must be explained by differ-
i
) from the enzyme, then this step
M
ences in the rates of steps before phosphate release and
cannot be explained by differences in the rates of the
processing of the enol pyranose intermediate.
The kcat. for the conversion of 11 is also slower than
for 1 and 8 (by factors of 50 and 7, respectively) and also
slower than the very fast nonenzymatic conversion of 3
7
to a mixture of 6 and 7. This could be due to slowing of
steps either before or after phosphate release. One
possibility is that it reflects release of enolpyranose 12
from the enzyme at a rate that is slower than the rate
for further processing of 3 (and 9) on the enzyme. This
interpretation would imply that the fluorine substitution
in 12 is indeed stabilizing the enol pyranose form as
expected and not simply enhancing its dissociation from
the enzyme.
F igu r e 2. 1H NMR spectra (500 MHz) of (a) (6R)-6-fluorode-
hydroquinate (10), (b) (6S)-6-fluorodehydroquinate (13), and
(
c) (6S)-1-epi-6-fluorodehydroquinate (14). Spectra were re-
It has previously been concluded that dehydroquinate
synthase plays some role in the cyclization of 3 to 6,
perhaps more as a template than as a catalyst. The
corded at 26 °C in D O with presaturation on the HOD signal
and were referenced to HOD at 4.67 ppm.
2
7
evidence for the role of the enzyme was the exclusive
formation of 6. By the simple expedient of introducing
a fluorine to stabilize the enol pyranose intermediate long
enough to allow it to dissociate from the enzyme we have
highlighted how subtle the role of the enzyme must be.
suggests that at least 14 and perhaps both compounds
are formed off the enzyme. This could occur if the effect
of the fluorine atom at C-3 is to increase the rate of
dissociation of the enol pyranose intermediate 12 from
the enzyme or decrease the rate of ring opening of 12 on
the enzyme or both.15 Neighboring fluorine atoms are
known to favor hemiketals and hydrates relative to the
corresponding ketones. Once in solution, ring opening
of 12 and subsequent cyclization could proceed via the
chairlike conformation 15 to form 13. Labeling studies
have shown that the formation of 6 enzymatically from
Exp er im en ta l Section
Kin etic Stu d ies Usin g DAHP a n d (3R)- a n d (3S)-3-
F lu or oDAHP s a s Su bstr a tes for Deh yd r oqu in a te Syn -
th a se. All kinetic parameters were determined in “bis-tris”
+
propane‚HCl buffer (50 mM) containing NAD (35 µM), MnSO
4
(
4
50 µM), and CoSO (20 µM) at pH 7.5 and 25 °C. Solutions
1
6
1
and nonenzymatically from 3 proceeds through the
also contained E. coli dehydroquinase (1.5 units). Reactions
were initiated by adding 10 µL of a solution containing dehyd-
roquinate synthase (0.6 µM for DAHP, 1.4 µM for (3R)-fluo-
roDAHP and 7 µM for (3S)-fluoroDAPH). The absorbance
increase was then monitored spectrophotometrically at 234 nm
for 1 and 230 nm for 8 and 11. The extinction coefficients at
these wavelengths for the corresponding dehydroshikimates
7
intermediate in a chair conformation 5. However, the
Felkin-Ahn model for attack on a carbonyl adjacent to
a chiral center predicts that the attack would be favored
when the electronegative fluorine atom is othogonal to
and behind the ketone group.17 Cyclization via the
chairlike conformation 16 may account for the increased
proportion of the epimeric product 14 relative to 13
4
-1
-1
3
-1
-1
were determined to be 1.2 × 10 M cm , 3.4 × 10 M cm
,
3
-1
-1
and 9.6 × 10
M
cm , respectively. Initial rates were
recorded, and the data were fitted to the Michaelis-Menton
(
compared to the ratio of 7 to 6 formed nonenzymatically
7
from 3).
equation using Enzfitter.
HP LC Con d ition s Used To P u r ify th e 6-F lu or od eh yd -
r oqu in a tes. HPLC was performed on an LKB HPLC system
using a BioRad Aminex ion-exclusion HPX-87H organic acid
semipreparative column (300 × 16 mm). The compounds were
eluted using 50 mM formic acid as eluant at a flow rate of 1.2
mL per min with continuous detection at 277 nm. Under these
conditions, dehydroquinate elutes after 18.1 min.
(
14) It is considered unlikely that the formation of the 14 is due to
dehydroquinate synthase using the C-1 epimer of the (6S)-6-fluo-
roDAHP as a substrate. Less than 4% of this compound is present in
the starting material (see Figure 1).
(
15) Analogous to the increase in stability seen when a fluorine is
introduced next to the anomeric center in 2-deoxy-2-fluorosugars.
Street, I. P.; Rupitz, K.; Withers, S. G. Biochemistry 1989, 28, 1581-
587.
En zym a tic Con ver sion of (3R)-F lu or od eh yd r oqu in a te
in to (6R)-6-F lu or od eh yd r oqu in a te. To an unbuffered aque-
ous solution (600 µL) containing (3R)-3-fluorodehydroquinate (45
1
(
16) Widlanski, T. S.; Bender, S. L.; Knowles, J . R. J . Am. Chem.
Soc. 1987, 109, 1873-1875.
17) Anh, N. T.: Eisenstein, O. Nouv. J . Chem. 1977, 1, 61-70.
+
(
mM), NAD (30 µM), and cobalt sulfate (50 µM) at pH 7.0 was

Notes
J . Org. Chem., Vol. 62, No. 24, 1997 8585
added dehydroquinate synthase (0.5 units). The reaction was
incubated at 26 °C for 3 h. The enzyme was then removed by
filtration through an Amicon Centricon concentrator and the
filtrate purified by HPCL. (6R)-6-Fluorodehydroquinate (10):
(500 MHz, D
2
O) correlations between H6 and H5, H5 and H4,
+
H2ax and H2eq; found M + Na (+ve ESMS) 231.026 43,
C
7
H
9
O
6
FNa requires 231.028 12; HPLC retention time (organic
acids column, 50 mM formic acid) 16.3 min. (6S)-6-Fluoro-1-
epidehydroquinate (14): δ (500 MHz, D
δ
9
1
H
(500 MHz, D
2
O) 5.15 (dd, J ) 47.7, 9.5 Hz, H6), 4.33 (d, J )
H
2
O) 4.86 (dd, J ) 50.8,
.8 Hz, H4), 3.95 (ddd, J ) 12.8, 9.8, 9.5 Hz, H5), 3.13 (d, J )
2.4 Hz, H6), 4.38 (d, J ) 9.9 Hz, H4), 4.03 (ddd, J ) 30.0, 9.9,
2.4 Hz, H5), 2.95 (d, J ) 14.7 Hz, H2ax), 2.64 (d, J ) 14.7 Hz,
4.8 Hz, H2ax), 2.61 (J ) 14.8, 7.6 Hz, H2eq); δ
F
(235 MHz, D
2
O)
O)
-
200.9 (ddd, J ) 47.7, 12.7, 7.6 Hz); COSY (500 MHz, D
2
H2eq): δ
207 (M - H ); COSY (500 MHz, D
and H5, H5 and H4, H2ax and H2eq; found M + Na (+ve ESMS)
231.027 21, C FNa requires 231.028 12; HPCL retention
F
(235 MHz, D
2
O) -209.0 (dd, J ) 50.8, 30.0 Hz); m/z
-
correlations between H6 and H5, H5 and H4, H2ax and H2eq
;
2
O) correlations between H6
-
-
+
m/z (-ve ESMS) 207 (M - H ), 208 (M - 2H + D ); found (+ve
ESMS) M + Na 231.027 28, C
M
+
7
H
Na
9
O
6
+
FNa requires 231.028 12,
7 9 6
H O
-
H
+
D
+
232.034 87,
C
7
H
8
D-
time (organic acids column, 50 mM formic acid) 14.6 min.
O
6
FNa requires 232.034 39; HPLC retention time (organic acids
column, 50 mM formic acid) 15.5 min.
Ack n ow led gm en t. We thank the Association of
Commonwealth Universities for a studentship to E.J .P.
and Dr. F. J . Leeper for helpful discussions.
En zym a tic Con ver sion of (3S)-3-F lu or od eh yd r oqu in a te
in to (6S)-6-F lu or od eh yd r oqu in a te a n d (6S)-6-F lu or o-1-
ep id eh yd r oqu in a te. To an unbuffered aqueous solution (600
+
µL) containing (3R)-3-fluorodehydroquinate (15 mM), NAD (30
µM), and cobalt sulfate (50 µM) at pH 7.3 was added dehydro-
Su p p or tin g In for m a tion Ava ila ble: Methods for prepa-
ration of DAHP, (3R)- and (3S)-3-fluoroDAHP, data for the
quinate synthase (45 units). The reaction was incubated at 26
°
C for 30 min. The enzyme was then removed by filtration
through an Amicon Centracon concentrator and the filtrate
purified by HPLC. (6S)-6-Fluorodehydroquinate (13): δ (500
MHz, D O) 4.85 (dd, J ) 48.9, 2.5 Hz, H6), 4.51 (d, J ) 10.1 Hz,
H4), 4.0 (ddd, J ) 29.6, 10.2, 2.5 Hz, H5), 3.29 (d, J ) 14.8 Hz,
H2ax), 2.53 (d, J ) 14.8 Hz, H2eq); δ (235 MHz, D O) -201.7
dd, J ) 48.9, 29.6 Hz); m/z (-ve ESMS) 207 (M - H ); COSY
19
enzymatic conversion of 1, 8, and 11, and F NMR spectra of
1
0, 13, and 14 (9 pages). This material is contained in libraries
H
on microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
2
F
2
-
(
J O9713060