Parallel kinetic resolution

Full text of DOI:10.1021/ja963666v

2584
J. Am. Chem. Soc. 1997, 119, 2584-2585
catalyzed by two moderately selective enzymes (Straathof et
Parallel Kinetic Resolution
al.).5a The latter workers concluded that the time evolution of
ee should improve in the parallel experiment, but they did not
include the case where the products of each of the competing
reactions are distinct. This situation has special advantages that
have not been discussed previously. As shown below, compet-
ing parallel reactions that produce two different chiral products
can give substantially improved ee values up to the theoretical
yield limit (50% of each enantiomer). We have called this
process parallel kinetic resolution (PKR).
Consider two chiral reagents Z1 and Z2 having similar but
opposite selectivity toward the enantiomers (R)-E and (S)-E in
a reaction of the racemic mixture to afford enantiomeric products
(R)-P1 and (S)-P1 from Z1 and a second set of enantiomeric
products (R)-P2 and (S)-P2 from Z2. Assume that the larger
rate constant k1(R) (for reaction of (R)-E with Z1) is equal to
k2(S) (for reaction of (S)-E with Z2) and that the smaller rate
constant k1(S) is equal to k2(R). In this ideal case, the overall
rates of conversion of (R)-E and (S)-E must be identical, and
the 1:1 ratio of starting enantiomer concentrations [(R)-E] and
[(S)-E] must be maintained throughout. Therefore, the ratios
of all four possible products (enantiomer pairs (R)-P1 and (S)-
P1; (R)-P2 and (S)-P2) will remain constant from >0% to 100%
conversion. The enantiomeric purity of the products P1 and P2
can be predicted directly from the corresponding selectivities
s1 (equal to k1(R)/k1(S)) and s2 (equal to k2(S)/k2(R)) at all times.
Thus, a PKR experiment using two simultaneous reactions of
complementary enantioselectivity with s1 ) s2 ) 49 (100% con-
version) would be equivalent to a simple kinetic resolution with
s ) 200 at 50% conversion!1 Theoretically, both experiments
would allow total recovery of each enantiomer with 96% ee.
Edwin Vedejs* and Xinhai Chen
Chemistry Department, UniVersity of Wisconsin
Madison, Wisconsin 53706
ReceiVed October 23, 1996
In a competition experiment where a 1:1 mixture of two
substrates reacts simultaneously with a single reagent, the initial
product ratio corresponds to the inherent selectivity s (the ratio
of competing rate constants). However, product ratio decreases
with percent conversion because there is a continuous increase
in the relative concentration (and therefore, the relative rate of
reaction) of the less reactive substrate as the faster-reacting
substrate is consumed. The kinetic consequences are well-
known for the special case where the substrates are two
enantiomers competing for an enantioselective chiral reagent
(simple kinetic resolution).1-3 In principle, the slower reacting
enantiomer can be recovered with very high ee, but only if
conversion is sufficiently high to consume essentially all of the
more reactive enantiomer. Depending on selectivity, this can
drastically reduce the yield of the purified unreacted enantiomer.
Furthermore, increased conversion can only decrease the ee of
the product derived from the more reactive enantiomer, ap-
proaching 0% ee as conversion approaches 100%. Exceptional
selectivity is required to obtain both slow and fast reacting
enantiomers (recovered substrate and product, respectively) with
high ee and the theoretical maximum 50% yield for each (for
example, s ) 200, 96% ee; s ) 500, 98% ee).1 Such selecti-
vities are currently beyond reach for most nonenzymatic kinetic
resolutions and for many of the lipase-esterase experiments.2a,b
We report a method to maximize ee as well as percent
conversion using a simple technique that maintains the optimum
1:1 substrate ratio throughout a competition experiment. This
conceptual variation requires the use of two selective reagents
in parallel. Parallel reactions have been encountered previously
using catalysts that may contain more than one enzyme4a,b or
catalysts that convert each enantiomer to a different product.4b-d
Thus, Brooks et al. reported that baker’s yeast reduces one of
the enantiomers of a â-keto ester to a chiral alcohol and induces
decarboxylation of the other enantiomer to afford an achiral
ketone.4a In this case, enzymatic selectivity in the pathway
leading to the chiral alcohol is already high, and the occurrence
of a parallel reaction from the other enantiomer leading to the
achiral product has no special advantage for enantiomeric purity.
However, Brooks et al. recognized the unique feature that
distinct products are formed in the competing reactions.
Mathematical treatments of relevant parallel reactions have
appeared,5 including kinetic models for hypothetical reactions
One simple variation of PKR involves competing reactions
where the reagents Z1 and Z2 are related as the quasienantiomers,
defined as two molecules Z1 and Z2 containing similar stereo-
genic carbons C[a][b][c][d] for Z1 and C[a][b][e][c] (opposite
configuration) for Z2 such that Z1 and Z2 would be true
enantiomers if the substituents [e] and [d] were identical.3b In
this case, the requirement for similar reaction rates and
complementary enantioselectivities is relatively easy to satisfy.
The chiral DMAP-derived salt 3 (from 1 and trichloro-tert-
butylchloroformate) has been shown to discriminate between
enantiomers of 1-(1-naphthyl)ethanol with s ) 42 for the (S)-
enantiomer in an acyl transfer process.6 A quasienantiomer of
1 was easily made by benzylation of the precursor (S)-alcohol
with benzyl bromide to give ether 2. It was more difficult to
find a chloroformate that would react with 2 to give an acyl
transfer agent having high enantioselectivity and also the ability
to form easily separated products. Hindered chloroformates
gave the most promising results, and fenchyl chloroformate
afforded an N-alkoxycarbonylpyridinium salt 4 that was shown
to have the desired properties (s ) 41 for (R)-alcohol). Fenchyl
chloroformate was chosen strictly for reasons of bulk and cost
(fenchyl alcohol, $0.10/g). The fact that the alkyl group is chiral
is probably irrelevant to selectivity.7 This feature did simplify
(1) Kagan, H. B.; Fiaud, J. C. Topics Stereochem. 1988, 18, 249.
(2) (a) Sih, C. J.; Wu, S.-H. Topics Stereochem. 1989, 19, 63. Chen,
C.-S.; Sih, C. J. Angew. Chem., Int. Ed., Engl. 1989, 28, 695. (b) Klibanov,
A. M. Acc. Chem. Res. 1990, 23, 114. Ward, S. C. Chem. ReV. 1990, 90,
1. Drueck-hammer, D. G.; Hennen, W. J.; Pederson, R. L.; Barbas, C. F.,
III; Gautheron, C. M.; Krach, T.; Wong, C.-H. Synthesis 1991, 499. Roberts,
S. M. Chimia 1993, 47, 85. (c) Double kinetic resolution is an option for
increasing efficiency: Brown, S. M.; Davies, S. G.; de Sousa, J. A. A.
Tetrahedron Asymmetry 1991, 2, 511.
(3) (a) Jacques, J.; Collet, A.; Wilen, S. H. In Enantiomers, Racemates,
and Resolutions; Krieger Publishing Co.: Malabar, FL, 1994. (b) Eliel, E.
L.; Wilen, S. H. In Stereochemistry of Organic Compounds; John Wiley &
Sons: New York, 1994; Chapter 7, see pp1205-1206 for relevant
definitions.
(4) (a) Brooks, D. W.; Wilson, M.; Webb, M. J. Org. Chem. 1987, 52,
2244. (b) Konigsberger, K.; Alphand, V.; Furstoss, R.; Griengl, H.
Tetrahedron Lett. 1991, 32, 499. Petit, F.; Furstoss, R. Tetrahedron
Asymmetry 1993, 4, 1341. Mischitz, M.; Faber, K. Tetrahedron Lett. 1994,
35, 81. (c) Bolm, C.; Schlinghoff, G. J. Chem. Soc., Chem. Commun. 1995,
1247. (d) Martin, S. F.; Spaller, M. R.; Liras, S.; Hartmann, B. J. Am. Chem.
Soc. 1994, 116, 4493. Doyle, M. P.; Dyatkin, A. B.; Kalinin, A. V.; Ruppar,
D. A.; Martin, S. F.; Spaller, M. R.; Liras, S. J. Am. Chem. Soc. 1995, 117,
11021.
1
the preliminary product assay (de estimated by H NMR), but
a more precise hplc assay was required to determine the s values.
Salts 3 and 4 (1.1 mol equiv) were generated in separate flasks
(CH2Cl2 solution), combined, and treated with MgBr2 (2.25 mol
(5) (a) Straathof, A. J. J.; Rakels, J. L. L.; Heijnen, J. J. Biocatalysis
1990, 4, 89. (b) Parallel reactions in mutual kinetic resolution: Brandt, J.;
Jochum, C.; Ugi, I.; Jochum, P. Tetrahedron 1977, 33, 1353. (c) Regio-
isomeric products from a racemic mixture and chiral reagents: Kagan, H.
Croat. Chem. Acta 1996, 69, 669.
(6) Vedejs, E.; Chen, X. J. Am. Chem. Soc. 1996, 118, 1809.
(7) (a) Both enantiomeric menthyl chloroformates have comparable
selectivity s ) ca. 12 toward the bromomagnesium alkoxide derived from
7a. A more specific test would require the enantiomer of fenchyl alcohol.
(b) Evidence for magnesium alkoxide: Vedejs, E.; Daugulis, O. J. Org.
Chem. 1996, 61, 5702.
S0002-7863(96)03666-9 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 10, 1997 2585
Table 1. PKR Experiments Using 3 and 4 to Resolve 7a
deviation from the 1:1 ratio and a correspondingly larger
difference in product ee values (entry 3), probably due in part
to undesired competition from the “mismatched” alcohol
enantiomer reacting with 3.
entry
Ar
5:6b
yield (5) ee (5)d yield (6) ee (6)d
1
2
3
1-naphthyl 1.0:1.0 46% (49%)c 88% 49% (49%)c 95%
2-naphthyl 1.0:1.0 49% (49%)c 86% 43% (49%)c 93%
Identical reactivities (k1(R) ) k2(S), etc.) in the competing
reactions are desirable, but they are probably not obligatory. In
principle, it should be sufficient to adjust the relative concentra-
tions of [Z1] and [Z2] so that the parallel reactions occur with
similar rates over most of the reaction coordinate. Furthermore,
there is no need for the parallel reactions to involve similar
functional group conversions, as in the examples illustrated here.
All that is necessary is that the two competing reactions (1)
occur without mutual interference, (2) have similar rates, (3)
have complementary enantioselectivity, and (4) afford distinct
products. This means that any two selective derivatizations of
a chiral racemic substrate can be used as the competing reactions
in the PKR experiment.10
In principle, one or both of the competing reactions can be
catalytic, provided that each catalyst is selective for the reagent
as well as for the substrate enantiomer. Enzymatic catalysts
could be used, but this variation would be worthwhile only if
the selectivity is below s ) ca. 150. Otherwise, efficient
enantiomer recovery is already possible without PKR. The PKR
process would not be preferred in systems where dynamic
kinetic resolution is possible or where the less reactive enan-
tiomer can be recycled in situ.11,12 These methods provide an
alternative means for maintaining the optimal 1:1 ratio of
enantiomers where necessary, and they have the advantage that
both enantiomers are converted into a single product.
The W-tube “resolving machine” concept of Cram et al.
(enantioselective transport) also depends on maintaining a 1:1
enantiomer ratio.12 In this case, dilution-driven transport
corresponds to the chemical derivatization component of PKR.
Another PKR application is inherent in reactions where two
enantiomers of a racemic substrate are selectively converted into
separable, chiral products (regioisomers, diastereomers, etc.) by
a single chiral reagent or catalyst. Some of the enantioselective
Bayer-Villiger oxidations4b,c and diazoketone insertions4d
reported previously belong to this category, and increased effi-
ciency is expected under ideal PKR conditions (equal or similar
rates; complementary enantiomer-dependent regioselectivity).
PKR is not restricted to the derivatization of enantiomers
under stoichiometric conditions. However, this is one of the
easier variations, and it provides sufficient advantages in
selectivity and efficiency that the use of stoichiometric reagents
in kinetic resolutions may need to be reevaluated in certain cases.
Even a modest selectivity of s ) 20 in the derivatization step
could provide chiral products in 90% ee with near-total material
recovery using PKR conditions.
o-tolyl
1.13:1.0 46% (53%)c 83% 46% (47%)c 94%
a 0.56 mol equiv each of 3 and 4, 2.25 mol equiv of MgBr2, 3 mol
equiv of Et3N, 1 mol equiv if racemic ArCH(OH)CH3 in CH2Cl2 at
room temperature, 36 h. b NMR ratio of crude products. c Yield vs
internal standard, NMR assay. d hplc assay; see Supporting Information.
equiv), Et3N (3 mol equiv), and racemic 1-(1-naphthyl)ethanol
(1 mol equiv). The reaction was allowed to proceed to >98%
conversion (NMR detection limit), and the product mixture of
5a and 6a was then treated with Zn/HOAc to selectively cleave
the trichlorobutyl protecting group.8 The resulting mixture of
alcohol (S)-7a and the mixed fenchyl carbonate 6a was easily
separated. If desired, the reusable quasienantiomers 1 and 2
could be recovered (ca. 90% combined yield after chromato-
graphic separation).9
According to hplc assay using a chiral stationary phase (csp),
(S)-7a was shown to have 88% ee (49% yield based on internal
standard NMR assay; 46% isolated). The mixed carbonate 6a
was assayed using hplc and was determined to have 95% de
(49% isolated yield). These findings are summarized in Table
1 (entry 1), along with similar results from two other racemic
secondary alcohols. To place the results into perspective, we
note that a simple kinetic resolution would have to operate at s
>125 to allow 49% recovery of one enantiomer with 95% ee
(or de). The 95% de value observed for 6a is identical to that
expected in the ideal PKR experiment with s ) 41-42 and
identical rates for both reagents. However, the 88% ee for 5a
indicates that in this case there is some interference from
components of the competing parallel reaction. Otherwise, this
value should also be ca. 95% ee. One possible explanation is
that a small amount (ca. 3%) of “leakage” occurs from 4 to 8
in the course of the PKR experiment, but our assay was not
sufficiently precise to rule out other possible explanations.
Acknowledgment. This work was supported by the National
Science Foundation. The authors also wish to thank Profs. H. Kagan
and C. Bolm for helpful comments and for pointing out relevant
literature analogies.
Supporting Information Available: Preparation of 2, fenchyl
chloroformate, characterization of 6a,b,c, and representative procedures
for PKR (5 pages). See any current masthead page for ordering and
Internet access instructions.
JA963666V
(10) In principle, mixtures of two achiral components (diastereomers;
regioisomers; etc.) could also be derivatized with improved efficiency using
two complementary reagents in parallel if they react selectively with each
isomer to afford a distinct product.
(11) (a) Review: Ward, R. S. Tetrahedron Asymmetry 1995, 6, 1475.
(b) Kitamura, M.; Tokunaga, M.; Noyori, R J. Am. Chem. Soc. 1993, 115,
144.
(12) (a) Biade, A.-E.; Bourdillon, C.; Laval, J.-M.; Mairesse, G.; Moiroux,
J. J. Am. Chem. Soc. 1992, 114, 893. (b) Tan, D. S.; Gunter, M. M.;
Drueckhammer, D. G. J. Am. Chem. Soc. 1995, 117, 9093. (c) Nakamura,
K.; Inoue, Y.; Matsuda, T.; Ohno, A. Tetrahedron Lett. 1995, 36, 6263.
(d) van der Deen, H.; Cuiper, A. D.; Hof, R. P.; van Oeveren, A.; Feringa,
B. L.; Kellogg, R. M. J. Am. Chem. Soc. 1996, 118, 3801.
(13) Newcomb, M.; Toner, J. L.; Helgeson, R. C.; Cram, D. J. J. Am.
Chem. Soc. 1979, 101, 4941.
The above data confirm the feasibility of PKR. Products 5a
and 6a and 5b and 6b were formed in a 1:1 ratio, indicating
that nearly ideal relative rate conditions were achieved. This
was desired to maintain the optimum 1:1 relative ratio of
enantiomers throughout the conversion from the racemate to
products. In the case of 5c and 6c, there was a measurable
(8) Eckert, H.; Listl, M.; Ugi, I. Angew. Chem. 1978, 90, 388.
(9) The true enantiomer of 1 could have been used in place of 2, but
this PKR experiment would racemize the chiral DMAP derivatives while
it resolves the substrate enantiomers.