Investigation of the enzymatic and nonenzymatic Cope rearrangement of carbaprephenate to carbachorismate

Article abstract of DOI:10.1021/jo026096s

The dimethyl esters of carbaprephenate and 4-epi-carbaprephenate were prepared by modification of published procedures. In methanol these compounds are converted quantitatively to isomeric 6-hydroxytricyclo[3.3.1.0^2.7]non-3-en-1,3-dimethyl esters via a two-step sequence involving an initial Cope rearrangement, followed by intramolecular Diels-Alder reaction of the dimethyl carbachorismate or 4-epi-carbachorismate intermediates. Carbaprephenate and its epimer were obtained by alkaline hydrolysis of the corresponding dimethyl esters. These compounds, in contrast to their ester precursors, undergo spontaneous acid-catalyzed decarboxylation in aqueous solution. Only at high pH does the Cope rearrangement compete with decarboxylation. At pH 12 and 90°C, carbaprephenate slowly rearranges to carbachorismate, which rapidly loses water to give 3-(2-carboxyallyl)benzoic acid as the major product. A small amount of the intramolecular Diels-Alder adduct derived from carbachorismate is also observed by NMR as a minor product. Carbaprephenate is not a substrate for the enzyme chorismate mutase from Bacillus subtilis (BsCM), nor does carbaprephenate inhibit the normal chorismate mutase activity of this enzyme, even when present in 200-fold excess over chorismate. Its low affinity for the enzyme-active site is presumably a consequence of placing a methylene group rather than an oxygen atom proximal to the essential cationic residue Arg90. Nevertheless, BsCM variants that lack this cation (R90G and R90A) do not accelerate the Cope rearrangement of carbaprephenate either, and a catalytic antibody 1F7, which exhibits modest chorismate mutase activity, is similarly inactive. Poor substrate binding and the relatively high barrier for the Cope compared to the Claisen rearrangement presumably account for the lack of detectable catalysis. Acceleration of this sigmatropic rearrangement apparently requires more than an active site that is complementary in shape to the reactive substrate conformer.

Full text of DOI:10.1021/jo026096s

In vestiga tion of th e En zym a tic a n d Non en zym a tic Cop e
Rea r r a n gem en t of Ca r ba p r ep h en a te to Ca r ba ch or ism a te
Andreas Aemissegger, Bernhard J aun, and Donald Hilvert*
Laboratorium fu¨r Organische Chemie, Swiss Federal Institute of Technology, ETH-Ho¨nggerberg,
CH-8093 Zu¨rich, Switzerland
hilvert@org.chem.ethz.ch
Received J une 24, 2002
The dimethyl esters of carbaprephenate and 4-epi-carbaprephenate were prepared by modification
of published procedures. In methanol these compounds are converted quantitatively to isomeric
6-hydroxytricyclo[3.3.1.02,7]non-3-en-1,3-dimethyl esters via a two-step sequence involving an initial
Cope rearrangement, followed by intramolecular Diels-Alder reaction of the dimethyl carba-
chorismate or 4-epi-carbachorismate intermediates. Carbaprephenate and its epimer were obtained
by alkaline hydrolysis of the corresponding dimethyl esters. These compounds, in contrast to their
ester precursors, undergo spontaneous acid-catalyzed decarboxylation in aqueous solution. Only
at high pH does the Cope rearrangement compete with decarboxylation. At pH 12 and 90 °C,
carbaprephenate slowly rearranges to carbachorismate, which rapidly loses water to give 3-(2-
carboxyallyl)benzoic acid as the major product. A small amount of the intramolecular Diels-Alder
adduct derived from carbachorismate is also observed by NMR as a minor product. Carbaprephenate
is not a substrate for the enzyme chorismate mutase from Bacillus subtilis (BsCM), nor does
carbaprephenate inhibit the normal chorismate mutase activity of this enzyme, even when present
in 200-fold excess over chorismate. Its low affinity for the enzyme-active site is presumably a
consequence of placing a methylene group rather than an oxygen atom proximal to the essential
cationic residue Arg90. Nevertheless, BsCM variants that lack this cation (R90G and R90A) do
not accelerate the Cope rearrangement of carbaprephenate either, and a catalytic antibody 1F7,
which exhibits modest chorismate mutase activity, is similarly inactive. Poor substrate binding
and the relatively high barrier for the Cope compared to the Claisen rearrangement presumably
account for the lack of detectable catalysis. Acceleration of this sigmatropic rearrangement
apparently requires more than an active site that is complementary in shape to the reactive
substrate conformer.
Chorismate mutases (EC 5.4.99.5) accelerate the peri-
cyclic Claisen rearrangement of chorismate 1 to prephen-
ate 2, the first committed step in the biosynthesis of the
aromatic amino acids phenylalanine and tyrosine,1 by a
factor of 106-107. The uncatalyzed2 as well as the
enzymatic3 reactions proceed through chairlike transition
states (Figure 1A). Kinetic isotope effects show that C-O
bond cleavage precedes C-C bond formation in both
cases,4,5 indicating a concerted but asynchronous process.
X-ray structures6-8 of chorismate mutases from different
sources complexed to a transition state analogue show
that the active site is configured to constrain the flexible
substrate molecule in a reactive conformation in which
the trans C-4 hydroxyl group and C-3 enol pyruvate
substituents are pseudodiaxial (e.g., Figure 1A,B). Mu-
tagenesis studies9-13 and computation14-16 suggest that
a positively charged active site residue (either Lys or Arg)
located proximal to the ether oxygen of bound substrate6,7
(Figure 1B) is essential for catalysis, presumably because
it is able to stabilize the polarized transition state
* To whom correspondence should be addressed. Phone: +41-1-632-
3176. Fax: +41-1-632-1486.
(8) Stra¨ter, N.; Schnappauf, G.; Braus, G.; Lipscomb, W. N. Structure
1997, 5, 1437-1452.
(1) Haslam, E. Shikimic Acid Metabolism and Metabolites; J ohn
Wiley & Sons: New York, 1993.
(9) Kast, P.; Grisostomi, C.; Chen, I. A.; Li, S.; Krengel, U.; Xue, Y.;
Hilvert, D. J . Biol. Chem. 2000, 275, 36832-36838.
(10) Kast, P.; Asif-Ullah, M.; J iang, N.; Hilvert, D. Proc. Natl. Acad.
Sci. U.S.A. 1996, 93, 5043-5048.
(2) Copley, S. D.; Knowles, J . R. J . Am. Chem. Soc. 1985, 107, 5306-
5308.
(3) Sogo, S. G.; Widlanski, T. S.; Hoare, J . H.; Grimshaw, C. E.;
Berchtold, G. A.; Knowles, J . R. J . Am. Chem. Soc. 1984, 106, 2701-
2703.
(4) Addadi, L.; J affe, E. K.; Knowles, J . R. Biochemistry 1983, 22,
4494-4501.
(5) Gustin, D. J .; Mattei, P.; Kast, P.; Wiest, O.; Lee, L.; Cleland,
W. W.; Hilvert, D. J . Am. Chem. Soc. 1999, 121, 1756-1757.
(6) Chook, Y. M.; Gray, J . V.; Ke, H.; Lipscomb, W. N. J . Mol. Biol.
1994, 240, 476-500.
(7) Lee, A. Y.; Karplus, P. A.; Ganem, B.; Clardy, J . J . Am. Chem.
Soc. 1995, 117, 3627-3628.
(11) Liu, D. R.; Cload, S. T.; Pastor, R. M.; Schultz, P. G. J . Am.
Chem. Soc. 1996, 118, 1789-1790.
(12) Cload, S. T.; Liu, D. R.; Pastor, R. M.; Schultz, P. G. J . Am.
Chem. Soc. 1996, 118, 1787-1788.
(13) Zhang, S.; Kongsaere, P.; Clardy, J .; Wilson, D. B.; Ganem, B.
Bioorg. Med. Chem. 1996, 4, 1015-1020.
(14) Wiest, O.; Houk, K. N. J . Org. Chem. 1994, 59, 7582-7584.
(15) Lyne, P. D.; Mulholland, A. J .; Richards, W. G. J . Am. Chem.
Soc. 1995, 117, 11345-11350.
(16) Davidson, M. M.; Hillier, I. H. J . Chem. Soc., Perkin Trans. 2
1994, 1415-1417.
10.1021/jo026096s CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/24/2002
J . Org. Chem. 2002, 67, 6725-6730
6725

Aemissegger et al.
SCHEME 1a
F IGURE 1. (A) Rearrangement of chorismate to prephenate.
(B) Active site of BsCM with a bound transition state analogue
(3, bold).6 Inhibitor 3 binds chorismate mutases with Ki values
in the range 0.1 to 4 µM.21,23
electrostatically. Interestingly, an antibody with weak
chorismate mutase activity (1F7), which was raised in
response to transition state analogue 3,17 preferentially
binds the pseudodiaxial substrate conformer18 but lacks
this electrostatic feature.19
a
Conditions: (i) LiN(iPr)(cHx), BrCH2C(CO2CH3)dCH2; (ii)
AgO2CCF3, PhSeCl; (iii) (A) H2O2, (B) ∆; (iv) NaOH.
ity in the presence and absence of various proteins with
chorismate mutase activity, including the natural enzyme
from Bacillus subtilis (BsCM),23 the R90G and R90A
BsCM variants which lack the essential active site
cation,9 and the catalytic antibody 1F7.
The Cope rearrangement of “carbachorismate”, in
which the ether oxygen of chorismate is replaced with a
methylene group, to give “carbaprephenate”, in which the
carbonyl group of prephenate is replaced with an olefin,
would be expected to have a substantially less polarized
transition state than the Claisen rearrangement of
chorismate to prephenate. This aspect makes this reac-
tion potentially interesting as a probe of the relative
importance of electrostatic9 versus conformational20 ef-
fects in chorismate mutase catalysis. The “carba” ana-
logue of inhibitor 3 binds only 250 times less tightly to a
bifunctional chorismate mutase from E. coli,21 raising the
possibility that the enzymes would also recognize carba-
chorismate and carbaprephenate and promote their in-
terconversion. The catalytic antibody or variants of the
natural enzyme containing an uncharged amino acid in
place of the critical cationic residue might be even better
catalysts since they combine an active site of appropriate
shape for constraining the substrate in a reactive geom-
etry with a more suitable electrostatic environment. To
test these ideas, we have prepared carbaprephenate from
the corresponding dimethyl ester, which has been re-
ported in the literature,22 and have examined its reactiv-
Resu lts a n d Discu ssion
Syn th esis. Unlike the Claisen rearrangement of cho-
rismate to prephenate, which is highly exothermic,24 the
Cope rearrangement of the corresponding carba ana-
logues should have an equilibrium constant closer to 1
and hence be reversible. For this reason, synthetically
more accessible carbaprephenate was chosen in prefer-
ence to carbachorismate for study with the enzymes.
Carbaprephenate and 4-epi-carbaprephenate were syn-
thesized as shown in Scheme 1. The corresponding
dimethyl esters 6a and 6b are known compounds22 and
were prepared in racemic form by minor modification of
published protocols. In the conversion of methyl 3,6-
dihydrobenzoate to 4, the cyclic urea DMPU was em-
ployed as a cosolvent in place of HMPT because of its
lower toxicity.25 With 2 equiv of DMPU, only a modest
decrease in the yield of 4 (64%) was observed compared
to the published value (71%). Treatment of 4 with PhSeCl
afforded 5 as a 3:1 mixture of diastereoisomers, as
previously described, but comparison of its 1H NMR
spectrum with the published data22 showed that the
major and minor isomers were inverted. Subsequent
selenoxide elimination gave the isomeric esters 6a and
(17) Hilvert, D.; Carpenter, S. H.; Nared, K. D.; Auditor, M.-T. M.
Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4953-4955.
(18) Campbell, A. P.; Tarasow, T. M.; Massefski, W.; Wright, P. E.;
Hilvert, D. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 8663-8667.
(19) Haynes, M. R.; Stura, E. A.; Hilvert, D.; Wilson, I. A. Science
1994, 263, 646-652.
(20) Khanjin, N. A.; Snyder, J . P.; Menger, F. M. J . Am. Chem. Soc.
1999, 121, 11831-11846.
(23) Gray, J . V.; Eren, D.; Knowles, J . R. Biochemistry 1990, 29,
8872-8878.
(24) Kast, P.; Tewari, Y. B.; Wiest, O.; Hilvert, D.; Houk, K. N.;
Goldberg, R. N. J . Phys. Chem. B 1997, 101, 10976-10982.
(25) Chimia 1985, 39, 147-148.
(21) Bartlett, P. A.; Nakagawa, Y.; J ohnson, C. R.; Reich, S. H.; Luis,
A. J . Org. Chem. 1988, 53, 3195-3210.
(22) Gajewski, J . J .; J urayj, J .; Kimbrough, D. R.; Gande, M. E.;
Ganem, B.; Carpenter, B. K. J . Am. Chem. Soc. 1987, 109, 1170-1186.
6726 J . Org. Chem., Vol. 67, No. 19, 2002

Cope Rearrangement of Carbaprephenate to Carbachorismate
F IGURE 2. Time course for the rearrangement of dimethyl carbaprephenate (6a , A) and dimethyl 4-epi-carbaprephenate (6b,
B) in methanol at 60 °C. Relative amounts of starting material (b), product of the Cope rearrangement (9), and product of the
Diels-Alder rearrangement (2). Percent conversion was determined by averaging NMR peak areas; error bars indicate deviation
from the average value.
SCHEME 2
6b, which were separated by normal-phase preparative
HPLC. The latter procedure proved more effective in our
hands than flash chromatography.
Esters 6a and 6b were separately hydrolyzed to
carboxylates 7a and 7b with NaOH in a 1:3 mixture of
water and THF. Excess base was used to suppress acid-
catalyzed decarboxylation of the products. The reactions,
which were monitored by NMR spectroscopy, proceeded
quantitatively. Because 7a and 7b decarboxylate readily,
even at neutral pH, they were used directly for kinetic
measurements without further purification.
Assign m en t of 6a a n d 6b. The major and minor
products of the selenoxide elimination of 5 have very
similar 1H NMR spectra, making unambiguous assign-
ment of 6a and 6b difficult. Spectral data for dimethyl
carbaprephenate, allegedly the minor product of the
reaction, have been published,22 although how this
mediate to a single product having the same molecular
mass as the starting material but only one double bond.
The major isomer is quantitatively converted to a closely
related compound, but it reacts about 7 times slower than
its epimer and only trace amounts of the intermediate
are observed during the reaction. Although isolation of
the transient intermediate was not attempted in either
case, the peaks observed in the NMR spectra correspond
to those expected for dimethyl (4-epi)-carbachorismate.
For example, a resonance at 6.9-7 ppm is characteristic
of the hydrogen at C-2 in the conjugated cyclohexadiene
ring system and compares well with the values observed
for dimethyl chorismate and dimethyl 4-epi-chorismate.26
The structure of the final rearrangement product of the
minor isomer was established unambiguously as racemic
10a (Scheme 3) by a complete 1D difference NOE and
2D NMR spectroscopic analysis (see Supporting Informa-
tion for details). The rearrangement product of the major
isomer was assigned to structure 10b on the basis of the
close similarity of its spectra and that of 10a and the
fact that 6a and 6b are epimers at C-4.
structure was assigned was not described. Because the
reported coupling constants for the vinylic ring protons
correspond more closely to those of our major isomer, and
also to the related values for dimethyl 4-epi-prephenate,26
additional support for this assignment was sought.
Observation of an NOE between the side chain meth-
ylene group at C1 and the hydrogen atom at C4 would
have provided convincing evidence for structure 6a , but
magnetization transfer was not detected with either
isomer, suggesting that the centers of interest are too
far apart. For this reason, the Cope rearrangement of
6a and 6b was examined. In analogy with chorismate
and 4-epi-chorismate,26 carbachorismate and 4-epi-car-
bachorismate should be relatively easy to distinguish.
However, in previous work22 dimethyl carbaprephenate
was found to epimerize and to rearrange to 8 in aqueous
methanol. Both reactions are strongly acid-catalyzed,
which was rationalized by an ionic mechanism involving
carbocation formation (Scheme 2). To suppress this
undesired pathway, the reactions of 6a and 6b were
carried out in CD3OD in base-washed NMR tubes.
Rearrangement of the two epimers was monitored by
1H NMR at 60 °C (Figure 2). Under these conditions, the
minor isomer is slowly converted via a transient inter-
These observations can be rationalized by a two-step
process in which dimethyl (4-epi)-carbaprephenate under-
goes an initial Cope rearrangement to racemic dimethyl
(4-epi)-carbachorismate, followed by an intramolecular
Diels-Alder reaction to give either 10a or 10b (Scheme
3). Because the Cope rearrangement and Diels-Alder
(26) Hoare, J . H.; Policastro, P. P.; Berchtold, G. A. J . Am. Chem.
Soc. 1983, 105, 6264-6267.
J . Org. Chem, Vol. 67, No. 19, 2002 6727

Aemissegger et al.
F IGURE 3. (A) Arrhenius plot for the decarboxylation of carbaprephenate. Reactions were monitored in deuterated phosphate
buffer (ca. 70 mM, pD 10). ∆Hq ) 22.0 kcal mol-1; ∆Sq ) 18.4 cal mol-1 K-1. (B) pH dependence of carbaprephenate decarboxylation
at 30 °C. The rate constant for the value at pH 3 was extrapolated from data obtained at 20 °C with the Arrhenius equation.
SCHEME 3
stable in CDCl3 at 30 °C, whereas dimethyl 4-epi-
chorismate rearranges to dimethyl 4-epi-prephenate with
a half-life of 2.3 h under these conditions. Hydrogen
bonding between the trans C-3 and C-4 substituents in
dimethyl chorismate, which would stabilize the unreac-
tive pseudodiequatorial substrate conformer (13), has
been invoked to explain this difference in reactivity.26 It
is also conceivable that the polarized transition state for
Claisen rearrangement of dimethyl 4-epi-chorismate
would be directly stabilized by a hydrogen bond between
the C-4 hydroxyl group and the ether oxygen (rather than
the ester group) of the enol pyruvate substituent (14),
leading to faster rearrangement of this isomer.27
cycloaddition occur in a stereochemically defined manner,
the structure of the epimeric tricyclononene derivatives
allows unambiguous assignment of the minor and major
products of the selenoxide elimination to structures 6a
(dimethyl carbaprephenate) and 6b (dimethyl 4-epi-
carbaprephenate), respectively.
Fitting the NMR data to a two-step kinetic model (A
f B f C) gives rate constants of 3.4 × 10-6 and 1.0 ×
10-5 s-1 for the Cope rearrangement of dimethyl carba-
prephenate and the subsequent intramolecular Diels-
Alder cycloaddition, respectively. The corresponding val-
ues for dimethyl 4-epi-carbaprephenate are 4.6 × 10-7
and 1.3 × 10-5 s-1. Thus, while the intramolecular Diels-
Alder reactions occur at comparable rates, the Cope
rearrangement of dimethyl carbaprephenate is roughly
an order of magnitude faster than that of dimethyl 4-epi-
carbaprephenate.
The slower rearrangement of dimethyl 4-epi-carba-
prephenate compared to dimethyl carbaprephenate may
reflect, in part, misalignment of the reacting π-system
as a consequence of intramolecular hydrogen bonding or
steric interactions between the C-4 hydroxyl group and
the C-1 methacrylate ester substituent (e.g., 12). Such
an interaction is not possible in dimethyl carbaprephen-
ate, where these groups are trans to one another and
oriented pseudodiaxially in the reactive conformation.
Interestingly, the order of reactivity for the Claisen
rearrangement of dimethyl chorismate and dimethyl
4-epi-chorismate is reversed: dimethyl chorismate is
Non en zym atic Reaction of 7a an d 7b. Like prephen-
ate,28 carbaprephenate and 4-epi-carbaprephenate un-
dergo spontaneous acid-catalyzed decarboxylation in
aqueous solution (Figure 3). At pH 3 and 20 °C, the half-
life for prephenate is 18 min, whereas 7a and 7b have
half-lives of 4.6 and 5.5 min, respectively. Even raising
the pH does not completely suppress this undesired side
reaction, although the half-life of carbaprephenate in-
creases to >800 h at pH 10 (30 °C). NMR analysis of
product mixtures obtained at pH 10 at different temper-
atures reveals <5% 11, the product of Cope rearrange-
ment and aromatization (Scheme 2).22 However, when 7a
is heated at 90 °C for 35 h in phosphate buffer (50 mM)
at pH 12 or in 1 M NaOH, compound 11 is the dominant
product. A small amount of the Diels-Alder adduct (ca.
10%) is also observed in the NMR spectra of the crude
product. Under these conditions, rearrangement followed
(27) Carlson, H. A.; J orgensen, W. L. J . Am. Chem. Soc. 1996, 118,
8475-8484.
(28) Zamir, L. O.; Tiberio, R.; J ensen, R. A. Tetrahedron Lett. 1983,
24, 2815-2818.
6728 J . Org. Chem., Vol. 67, No. 19, 2002

Cope Rearrangement of Carbaprephenate to Carbachorismate
methoxide, THF from potassium, and CH2Cl2 from phosphorus
pentoxide. 1,3-Dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone
(DMPU) was distilled from calcium hydride at reduced pres-
sure, sealed in septum-capped vials, and stored at 4 °C. All
other commercially available reagents and solvents were used
as received. Butyllithium was titrated against diphenyl acetic
acid in THF prior to use.30 Organic extracts were dried over
magnesium sulfate, filtered, and concentrated with use of a
rotary evaporator. Nonvolatile oils and solids were dried under
vacuum at 0.01 Torr. Reactions were monitored by thin-layer
chromatography on precoated E. Merck silica gel 60 F254 or
on C18 reverse-phase plates (0.25 mm) and visualized either
by short-wave UV or by staining with 10% ethanolic phospho-
molybdic acid. Flash chromatography31 was performed on E.
Merck silica gel 60. NMR tubes used for kinetic measurements
were soaked in concentrated NH4OH solution for several hours
and dried at 110 °C overnight. Chemical shifts are expressed
in ppm relative to tetramethylsilane or the solvent signal. All
melting points are uncorrected. Wild-type B. subtilis choris-
mate mutase, the R90G and R90A BsCM variants, and the
catalytic antibody 1F7 were produced and purified as previ-
ously described.9,17
1-Meth oxycar bon yl-1-[2-(m eth oxycar bon yl)allyl]cyclo-
h exa -2,4-d ien e (4). N-Butyllithium (1.6 M in hexane, 1.87
mL, 3 mmol) was added to a -78 °C solution of N-cyclohexyl-
isopropylamine (0.5 mL, 3 mmol) in THF (15 mL). After the
solution was stirred for 15 min, methyl 3,6-dihydrobenzoate
(0.4 mL, 3.1 mmol) in DMPU (0.8 mL) was added and stirring
was continued for another 3 min. The dark red solution was
transferred by cannula to a stirred solution of methyl 2-(bromo-
methyl)acrylate (0.48 mL, 3.9 mmol) in THF (20 mL) at -78
°C. After 1 min, the reaction was quenched with 1 mL of acetic
acid. The solution was washed with water, and the aqueous
layer was extracted with diethyl ether. The combined organic
phases were washed twice with 3.2 M HCl and once with
saturated NaHCO3. The product was purified by flash chro-
matography (ethyl acetate/hexane, 1:9) to give 4 (470 mg, 64%)
as a colorless oil. 300-MHz 1H NMR (CDCl3) δ 6.18 (d, J )
1.55 Hz, 1 H), 5.95-5.71 (m, 4 H), 5.46 (d, J ) 1.24 Hz, 1 H),
3.68 (s, 3 H), 3.62 (s, 3 H), 2.74-2.54 (m, 3 H), 2.32 (dd, J )
5.4, 17.9 Hz, 1 H). 75-MHz 13C NMR (CDCl3) δ 174.96, 167.31,
136.09, 128.31, 128.20, 125.20, 124.18, 123.21, 51.78, 51.74,
45.58, 37.35, 30.23.
4-Hyd r oxy-1-m eth oxyca r bon yl-1-[2-(m eth oxyca r bon -
yl)a llyl]-tr a n s-3-p h en ylselen ocycloh exa -5-en e (5). A mix-
ture of silver trifluoroacetate (318 mg, 1.44 mmol) and
phenylselenyl chloride (220 mg, 1.15 mmol) in CH2Cl2 (15 mL)
was stirred at room temperature for 20 min and then cooled
to -78 °C. The mixture was added rapidly to a solution of 4
(260 mg, 1.1 mmol) in CH2Cl2 (5 mL), and the solution was
warmed to room temperature over 5 h. After adding a
saturated solution of K2CO3 in 4:1 H2O/MeOH (25 mL),
followed by concentrated NH4OH (10 mL), the bright yellow
organic layer was separated, and the aqueous phase was
extracted three times with dichloromethane. The crude product
was purified by flash chromatography (acetone/hexane, 1:4)
to give 5 (380 mg, 84%) as a 3:1 mixture of diastereoisomers.
300-MHz 1H NMR (CDCl3) δ 7.65-7.58 (m, 2 H), 7.35-7.25
(m, 3 H), 6.20 (d, J ) 1.25 Hz, 0.25 H), 6.15 (d, J ) 1.24 Hz,
0.75 H), 5.83-5.74 (m, 2 H), 5.46 (d, J ) 0.93 Hz, 0.25 H),
5.34 (d, J ) 0.93 Hz, 0.75 H), 4.02 (d, J ) 9.33 Hz, 1 H), 3.70
(s, 3 H), 3.62 (s, 3 H), 3.27-3.18 (m, 1 H), 2.72-2.60 (m, 3 H),
2.35 (m, 0.5 H), 2.00 (m, 0.5 H).
by dehydration or intramolecular Diels-Alder reaction
thus occurs in preference to decarboxylation. Assuming
that the Cope rearrangement of 7a is pH independent,
we estimate that it is roughly 10-100-times slower than
decarboxylation in the pH range 8 to 9.
En zym a tic Assa ys. Carbaprephenate 7a was tested
as a substrate for Bacillus subtilis chorismate mutase
(BsCM), the R90G and R90A variants of this enzyme,
and the catalytic antibody 1F7. The latter proteins, which
are either inactive or relatively poor chorismate mu-
tases,9,19 have active sites that are complementary in
shape to the constrained substrate conformer needed for
reaction. However, they lack the strategically placed
cationic group that interacts with the ether oxygen of
chorismate in wild-type BsCM.6,9
All reactions were monitored by HPLC with p-ethoxy-
benzoic acid as an internal standard. High concentrations
of enzyme (ca. 100 µM) and antibody (22 µM) were
employed to maximize the chances of detecting even
modest levels of catalysis. Because the catalysts are
nearly pH independent in the range 5 to 9,23,29 the assays
were performed at pH 8.6 (50 mM Tris‚HCl buffer, 30
°C) to minimize the competing spontaneous decarboxy-
lation of substrate. Under these conditions, the half-life
for decarboxylation is approximately 50 h. Assuming a
half-life of 500 to 5000 h for the uncatalyzed Cope
rearrangement of 7a and a Km value much larger than
the substrate concentration used in the assays, a rate
acceleration as small as 10-fold would have been detect-
able. In the event, even after several days no significant
difference in product yield or distribution was observed
between the enzyme solutions and the blanks.
Not unexpectedly, BsCM has a relatively low affinity
for carbaprephenate. However, even at 1 mM concentra-
tions, this compound does not detectably inhibit the
BsCM-catalyzed rearrangement of 50 µM chorismate.
Replacement of the carbonyl oxygen in prephenate with
a methylene group thus appears to be substantially more
destabilizing than replacement of the ether oxygen in
inhibitor 3 with a methylene group, which results in a
250-fold decrease in affinity.21 If carbaprephenate were
to bind to BsCM in an orientation analogous to that of
prephenate6 (which binds to the enzyme with a similar
affinity as chorismate23), the guanidinium cation of Arg90
would be juxtaposed with an apolar methylene group
rather than an oxygen atom capable of accepting a
hydrogen bond.
Nevertheless, removal of this destabilizing interaction
does not promote catalysis. BsCM derivatives in which
Arg90 is replaced with a glycine or an alanine and the
catalytic antibody 1F7, which lacks an analogous cationic
group, do not accelerate the Cope rearrangement of
carbaprephenate. Apparently, an active site configured
simply to constrain the substrate in a conformation
favorable for reaction, but lacking specific interactions
capable of stabilizing the pericyclic transition state, is
insufficient to overcome the energy barrier of the Cope
reaction.
Dim eth yl Ca r ba p r ep h en a te a n d Dim eth yl 4-epi-Ca r -
ba p r ep h en a te (6a , 6b). Compound 5 (370 mg, 0.9 mmol) in
CH2Cl2 (40 mL) was cooled to 0 °C, and a solution of 30% H2O2
(1 mL) in THF (10 mL) was added. The mixture was stirred
for 3.5 h and, after warming to room temperature, washed
Exp er im en ta l Section
Gen er a l Meth od s. All reactions were carried out in dried
glassware. Methanol was freshly distilled from magnesium
(30) Kofron, W. G.; Baclawski, L. M. J . Org. Chem. 1976, 41, 1879-
1880.
(29) Shin, J . A.; Hilvert, D. Bioorg. Med. Chem. Lett. 1994, 4, 2945-
2948.
(31) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.
J . Org. Chem, Vol. 67, No. 19, 2002 6729

Aemissegger et al.
twice with a saturated solution of K2CO3. The aqueous layers
were subsequently extracted twice with CH2Cl2. The combined
organic fractions yielded crude 4-hydroxy-1-methoxycarbonyl-
1-[2-(methoxycarbonyl)allyl]-trans-3-phenylseleninylcyclohexa-
5-ene (388 mg) as an amorphous, pale yellow solid that was
used without further purification. The solid was dissolved in
CH2Cl2 (10 mL), and after addition of diethylamine (220 µL,
2.1 mmol), the solution was transferred to hexane (40 mL) at
reflux. Removal of solvent after 2.5 h and flash chromatogra-
phy (ethyl acetate/hexane, 1:4) afforded a mixture of dimethyl
carbaprephenate and dimethyl 4-epi-carbaprephenate (205
mg). The two isomers were separated by preparative HPLC
on a column (20 × 250 mm) packed with Kromasil 5 µm silica
gel and eluted with ethyl acetate/hexane (1:2) at a flow rate
of 15 mL/min. Dimethyl carbaprephenate 6a (25 mg, 11%) and
dimethyl 4-epi-carbaprephenate 6b (66 mg, 29%) had retention
times of 28.6 and 23.7 min, respectively.
Hz, 1 H), 3.89 (m, 1 H), 3.75 (s, 3 H), 3.60 (s, 3 H), 3.40 (mb,
1 H), 2.53 (m, 1 H), 2.05 (m, 1 H), 1.83 (dd, J ) 3.73, 12.73
Hz, 1 H), 1.35-1.32 (m, 1 H), 1.22 (d, J ) 9.70 Hz, 1 H). 125-
MHz 13C NMR δ 176.34, 167.02, 147.45, 132.78, 73.31, 52.41,
52.14, 46.64, 41.00, 40.74, 40.10, 37.36, 32.98. MS (EI) m/z (M
+ H)+ 223.
exo-6-H yd r oxyt r icyclo[3.3.1.02,7]n on -3-en -1,3-d im et h -
yl ester (10b): Viscous colorless oil. 500-MHz 1H NMR
(CD3OD) δ 7.51 (dd, J ) 2.35, 7.42 Hz, 1 H), 3.85 (m, 1 H),
3.73 (s, 3 H), 3.61 (s, 3 H), 3.53 (m, 1 H), 2.93 (mb, 1 H), 2.30
(dd, J ) 3.83, 12.39 Hz, 1 H), 2.22-2.17 (m, 2 H), 2.00 (d, J )
8.51 Hz, 1 H), 1.36 (dd, J ) 2.09, 12.41 Hz, 1 H). 125-MHz 13
C
NMR δ 176.52, 167.16, 149.05, 131.84, 70.33, 52.38, 52.21,
47.32, 42.660, 41.81, 33.77, 31.62, 31.41. MS (EI) m/z (M +
H)+ 223.
Kin etics. NMR tubes for kinetic measurements were
soaked in concentrated ammonia for several hours and dried
overnight at 110 °C. After sample loading, the tubes were
sealed and heated to the desired temperature in a thermo-
stated oil bath. Rearrangement of compounds 6a and 6b in
CD3OD at 60 °C was monitored daily for the first week and
weekly afterward. Decarboxylation of compounds 7a and 7b
(3 mM) was monitored spectrophotometrically at 255 nm at
pH 3 (100 mM citrate buffer), 7 (70 mM phosphate buffer),
8.6 (50 mM Tris-HCl), and 10 (70 mM phosphate buffer). This
reaction was also monitored by NMR in deuterated phosphate
buffer (ca. 70 mM, pD 10) at 30, 45, 60, 75, and 90 °C,
respectively. Rearrangement to 3-(2-carboxyallyl)benzoic acid
(11) was observed under the latter conditions but to a minor
extent (<5%) that could not be quantified accurately by NMR.
Carrying out the reaction in 50 mM phosphate buffer at pH
12 or in 1 M NaOH resulted predominantly in conversion to
11 within 35 h at 90 °C as judged by analytical HPLC (Waters
Nova-Pak C-18, 10% to 90% acetonitrile over 30 min with 0.1%
TFA in water, using p-ethoxybenzoic acid as an internal
standard).
En zym a tic Rea ction s. The enzymatic reactions were
carried out in Tris-HCl buffer (50 mM, pH 8.6) in a total
volume of 0.5 mL. Substrate 7a (880 µM) was incubated with
high concentrations of antibody (22 µM) or enzyme (100 µM)
at 30 °C and the product distribution was analyzed periodically
by analytical HPLC. The reaction mixture (10 µL) was
separated on a Waters Nova-Pak C18 column (30 min gradient
from 10% to 90% acetonitrile in 0.1% TFA, total flow 1 mL/
min). Product distribution was determined at 232 nm by
integrating peak areas for 11, the dehydrated product of the
Cope rearrangement (retention time 10.1 min), and 2-benzyl-
2-propenic acid (retention time 13.4 min), the product of
decarboxylation. p-Ethoxybenzoic acid was used as an internal
standard (retention time 11.9 min).
Dim eth yl ca r ba p r ep h en a te (6a ): Viscous, colorless oil.
1
300-MHz H NMR (CDCl3) δ 6.18 (d, J ) 1.37 Hz, 1 H), 6.01
(dd, J ) 2.8, 10.50 Hz, 2 H), 5.94 (dd, J ) 1.26, 10.50 Hz, 2
H), 5.46 (d, J ) 1.28 Hz, 1 H), 4.37 (s, 1 H), 3.71 (s, 6 H), 2.79
(d, J ) 0.69 Hz, 2 H). 75-MHz 13C NMR (CDCl3) δ 173.38,
167.86, 135.38, 129.93, 129.10, 128.97, 61.69, 52.56, 51.90,
48.91, 39.90. MS (EI) m/z (M+ - H2O) calcd 234.0892, found
234.0886.
Dim eth yl 4-epi-car bapr eph en ate (6b): Amorphous, color-
1
less solid. Mp 37-39 °C. 300-MHz H NMR (CDCl3) δ 6.15 (d,
J ) 1.24 Hz, 1 H), 5.99 (dd, J ) 3.6, 10.45 Hz, 2 H), 5.88 (dd,
J ) 1.14, 10.45 Hz, 2 H), 5.42 (d, J ) 1.15 Hz, 1 H), 4.37 (s, 1
H), 3.72 (s, 6 H), 2.84 (d, J ) 0.73 Hz, 2 H). 75-MHz 13C NMR
(CDCl3) δ 173.15, 168.41, 136.23, 129.36, 129.00, 128.53, 61.36,
52.61, 52.10, 49.05, 39.38. MS (EI) m/z (M+ - H2O) calcd
234.0892, found 234.0896.
Disod iu m Ca r ba p r ep h en a te (7a ). Aqueous NaOH (1 M,
45 µL) was added to a solution of 6a (5 mg, 20 µmol) in 0.5
mL of THF/H2O (3:1) at 0 °C. After 30 min the solution was
warmed to room temperature, stirred for 60 min, and then
1
lyophilized. Hydrolysis was incomplete as judged by H NMR,
so the residue was redissolved in 0.5 mL of THF/H2O (3:1),
cooled to 0 °C and treated with NaOH (1 M, 45 µL). After an
additional 45 min at 0 °C and 90 min at room temperature,
solvent was removed by lyophilization to give 7a as a white
1
solid. 400-MHz H NMR (D2O) δ 5.91 (dd, J ) 1.46, 10.35 Hz,
2 H), 5.83 (dd, J ) 3.23, 10.37 Hz, 2 H), 5.58 (d, J ) 1.87 Hz,
1 H), 5.13 (m, 1 H), 4.45 (m, 1 H), 2.68 (s, 2 H). 100-MHz 13C
NMR (D2O) δ 183.99, 180.16, 145.73, 134.81, 129.88, 124.64,
64.28, 54.44, 43.99. MS (ESI) m/z (M + Na)- calcd 245.0426,
found 245.0427.
Disod iu m 4-epi-Ca r ba p r ep h en a te (7b). Aqueous NaOH
(1 M, 45 µL) was added to a solution of 6b (5 mg, 20 µmol) in
0.5 mL of THF/H2O (3:1) at 0 °C. After being stirred for 30
min, the solution was warmed to room temperature and stirred
for another 60 min. Lyophilization afforded 7b as a pale yellow
En zym e In h ibition . The BsCM-catalyzed rearrangement
of chorismate was performed in 50 mM potassium phosphate
buffer at pH 7.5 by monitoring the decrease in absorption at
274 nm as previously described.9 The concentrations of BsCM
and chorismate were 50 nM and 50 µM, respectively. Inhibitor
7a was present at concentrations of 10, 100, and 1000 µM.
1
solid. 400-MHz H NMR (D2O) δ 5.83 (dd, J ) 1.57, 10.37 Hz,
2 H), 5.74 (dd, J ) 3.13, 10.39 Hz, 2 H), 5.53 (d, J ) 1.75 Hz,
1 H), 5.11 (m, 1 H), 4.37 (m, 1 H), 2.67 (s, 2 H). 100-MHz 13C
NMR (D2O) δ 183.72, 180.49, 145.85, 134.62, 129.21, 124.32,
64.13, 54.36, 43.12. MS (ESI) m/z (M + Na)- calcd 245.0426,
found 245.0426.
Ack n ow led gm en t. This work was supported in part
by the Swiss National Science Foundation and Novartis
Pharma AG. We are grateful to Dr. F. Oesterhelt from
F. Hoffmann-La Roche AG (Basel, Switzerland) for high-
resolution ESI mass spectra and Dr. A. Flohr for helpful
discussions.
6-Hyd r oxytr icyclo[3.3.1.02,7]n on -3-en -1,3-d im eth yl Es-
ter (10a , 10b). Solutions of dimethyl carbaprephenate (5 mg,
20 µmol) and dimethyl 4-epi-carbaprephenate (5.2 mg, 21 µmol)
in CD3OD (0.6 mL) were sealed in NMR tubes and heated to
60 °C in a thermostated oil bath. 1H NMR spectra were
recorded daily for the first two weeks and, subsequently, once
a week. After completion of the reaction, solvent was evapo-
rated to yield the products in quantitative yield.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra of 6a , 6b, 7a , 7b, 10a , and 10b. This material is
available free of charge via the Internet at http://pubs.acs.org.
en d o-6-Hyd r oxytr icyclo[3.3.1.02,7]n on -3-en -1,3-d im eth -
yl est er (10a ): Viscous colorless oil. 500-MHz 1H NMR
(CD3OD) δ 7.44 (dd, J ) 2.11, 6.72 Hz, 1 H), 3.94 (d, J ) 4.23
J O026096S
6730 J . Org. Chem., Vol. 67, No. 19, 2002