O-Nucleophilic Amino Alcohol Acyl-Transfer Catalysts: The Effect of Acidity of the Hydroxyl Group on the Activity of the Catalyst

Article abstract of DOI:10.1021/ol035510n

(Equation presented) Amino alcohol-derived acyl-transfer catalysts are shown to operate by an O-nucleophilic mechanism, and catalysts bearing electron-withdrawing groups in proximity to the hydroxyl group are found to be more active. This is attributed to

Full text of DOI:10.1021/ol035510n

ORGANIC
LETTERS
2003
Vol. 5, No. 22
4105-4108
O-Nucleophilic Amino Alcohol
Acyl-Transfer Catalysts: the Effect of
Acidity of the Hydroxyl Group on the
Activity of the Catalyst
Kjirsten A. Wayman† and Tarek Sammakia*
Department of Chemistry and Biochemistry, UniVersity of Colorado,
Boulder, Colorado 80309-0215
sammakia@colorado.edu
Received August 11, 2003
ABSTRACT
Amino alcohol-derived acyl-transfer catalysts are shown to operate by an O-nucleophilic mechanism, and catalysts bearing electron-withdrawing
groups in proximity to the hydroxyl group are found to be more active. This is attributed to an increase in the acidity of the hydroxyl group
of the catalyst.
The serine proteases are a class of enzymes which function
by a mechanism in which the hydroxyl group of a serine
residue at the active site attacks the carbonyl of an amide,
thereby cleaving the amide and forming an acyl-enzyme
intermediate.1 This intermediate then undergoes hydrolysis
to complete the catalytic cycle and provide the hydrolyzed
amide. Both of these steps are catalyzed by basic residues
at the active site, and since the elucidation of this mechanism,
numerous workers have synthesized molecules bearing
hydroxyl groups in proximity to basic residues in an attempt
to mimic this mechanism of action.2 In this paper, we explore
the effect of acidity of the hydroxyl group on the activity of
these catalysts.
ketone or an aldehyde at the 2-position of a 4-aminopyridine.
They operate by a mechanism in which the hydroxyl group
of the hydroxy ester substrate (2) attacks the carbonyl of
the catalyst, thereby forming a covalent complex (hemiacetal
or hemiketal 3). The hydroxyl group of the hemiacetal/
hemiketal then attacks the carbonyl of the substrate and
produces dioxolanone 4 in a step that is catalyzed by the
neighboring pyridine. The dioxolanone then undergoes
methanolysis in a step that is also catalyzed by the pyridine
to produce the methyl ester of the product bound to the
catalyst as the hemiacetal/hemiketal (5). Breakdown of the
(2) For leading references of serine protease mimics and related
compounds, see: (a) Hine, J.; Khan, M. N. J. Am. Chem. Soc. 1977, 99,
3847. (b) Madder, A.; De Clercq, P. J.; Declercq, J.-P. J. Org. Chem. 1998,
63, 2548. (c) Skorey, K. I.; Somayaji, V.; Brown, R. S. J. Am. Chem. Soc.
1989, 111, 1445. (d) Cram, D. J.; Katz, H. E. J. Am. Chem. Soc. 1983,
105, 135. (e) Ghosh, M.; Conroy, J. L.; Seto, C. T. Angew. Chem., Int. Ed.
1999, 38, 514. (f) Menger, F. M.; Whitesell, L. G. J. Am. Chem. Soc. 1985,
107, 707-708. (g) Breslow, R.; Dong, S. D. Chem. ReV. 1998, 98, 1997-
2011. (h) Diederich, F.; Schu¨rmann, G.; Chao, I. J. Org. Chem. 1988, 53,
2744-2757.
We recently described a series of pyridine-derived acyl-
transfer catalysts which are effective for the selective
methanolysis of R-hydroxy esters over ordinary esters (1,
Scheme 1).3 These catalysts are bifunctional and contain a
† Current address, Department of Chemistry, Humboldt State University,
Arcata, CA 95521.
(1) Hedstrom, L. Chem. ReV. 2002, 102, 4501.
(3) (a) Sammakia, T.; Hurley, T. B. J. Am. Chem. Soc. 1996, 118, 8967.
(b) Sammakia, T.; Hurley, T. B. Org. Chem. 1999, 64, 4652.
10.1021/ol035510n CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/04/2003

Scheme 1
Scheme 2
hemiacetal/hemiketal provides the product (6) and regener-
ates the catalyst (1), ready to repeat the catalytic cycle.
Substrates that lack a hydroxyl group cannot form the
requisite hemiacetal/hemiketal and cannot undergo meth-
anolysis by this mechanism, hence leading to the selectivity
of these catalysts.
During the course of these studies, we found that ketone-
derived catalysts are far more selective than aldehyde-derived
catalysts (Figure 1). For example, aldehyde-derived catalyst
philicity of which is greatly diminished due to the hindrance
of the tertiary hydroxyl group that is produced. Thus, the
background reaction does not occur within the limits of our
detection, and these catalysts are very selective.
Hemiacetals are ∼4-5 pK units more acidic than ordinary
alcohols,5 and we felt that this increase in acidity may
facilitate both steps of the catalytic cycle. In the first step
(nucleophilic attack of the alcohol on the ester), the more
acidic hydroxyl group will be deprotonated to a greater extent
by the pyridine, thereby leading to greater charge on the
oxygen and producing a more nucleophilic species. In the
second step (methanolysis of the acylated catalyst), the more
acidic alcohol is a better leaving group. We have, therefore,
designed a series of catalysts to test this idea wherein the
acidity of the alcohol is varied by the presence of an electron-
withdrawing group. In our design, we sought to keep the
general features of compound 10 but to introduce the ability
to modify the acidity of the alcohol. As such, we prepared
catalysts bearing a 4-aminopyridine nucleus substituted with
a hydroxyalkyl group at the 2-position and a methyl group
at the 6-position (Figure 2). The 4-amino substituent
Figure 1. Aldehyde- and ketone-derived catalysts.
8 displays a rate difference of 96:1 in the methanolysis of 2
vs 7, whereas ketone-derived catalyst 9 displays a rate
difference of over 1700:1 for the same reaction. After
extensive mechanistic studies, we found that this is due to
an undirected background reaction reminiscent of the mech-
anism of action of the serine proteases (Scheme 2).4 This
undirected reaction proceeds by way of hemiacetal interme-
diate 10, which is produced by the addition of methanol to
the aldehyde of catalyst 8. The hydroxyl group of this
hemiacetal can attack the carbonyl of an active ester to
produce the acyl catalyst intermediate 11 in a step that is
catalyzed by the proximal basic pyridine. This intermediate
can then undergo methanolysis, again with catalysis by the
proximal pyridine, to provide the product (12) and regenerate
the catalyst. In the case of the ketone-derived catalysts,
addition of methanol produces a hemiketal (13), the nucleo-
Figure 2. Design elements of new catalysts.
increases the basicity of the pyridine nitrogen, while the
6-methyl group renders it non-nucleophilic. The substituent
R on the hydroxyalkyl group allows us to vary the electronics
(4) Sammakia, T.; Hurley, T. B. J. Org. Chem. 2000, 65, 974.
4106
Org. Lett., Vol. 5, No. 22, 2003

and acidity of the hydroxyl group and study this effect on
the rate of the reaction.
We prepared four catalysts with this structure in which
the R substituent is trifluoromethyl (14), methyl (15),
pentafluorophenyl (17), or phenyl (18, Table 1).6,7 These
Table 1. Relative Rates of Reaction Using Pyridine-Derived
Catalysts
entry
R
R′
k(rel) entry
R
R′
k(rel)
1
2
3
H
H
CF3 (14)
CH3 (15)
94
32
1
4
5
H
H
C6F5 (17)
C6H5 (18)
130
60
CH3 CF3 (16)
compounds were studied in the methanolysis of 7. Reactions
were conducted at a concentration of 0.1 M in CDCl3 at a
catalyst loading of 10%, with 10 equiv of deuteriomethanol,
and were monitored by 1H NMR. We find that reactions with
catalysts bearing electron-withdrawing groups are faster
(compare entry 1 vs entry 2, and entry 4 vs entry 5), which
is consistent with our hypothesis. As a control experiment,
we also studied compound 16 in which the hydroxyl group
is blocked as a methyl ether and found this compound
catalyzes the reaction 94 times slower than catalyst 14,
providing good evidence for an O-nucleophilic mechanism.
Because the reactions were monitored by NMR, we were
able to identify several species in solution and measure their
concentrations. Data from a reaction are plotted in Figure 3,
and this plot provides further evidence for an O-nucleophilic
mechanism.8 At the onset of the reaction, the substrate and
catalyst display a burst of reactivity and are rapidly consumed
at the same rate that a new catalyst species, the acylated
catalyst (19), appears. However, there is an induction period
before the appearance of product (12) is observed, and this
induction period correlates with the time required for the
acylated catalyst to be formed. After this initial burst, there
is a slower, steady-state reaction that is observed in which
the product is formed at the same rate that the starting
Figure 3. Kinetics of typical reaction using 20% of catalyst 14.
material is consumed. As the reaction progresses, the
concentration of the acylated catalyst decreases as does the
rate of formation of product, deviating from steady state at
high conversion. These data are consistent with a kinetic
scheme in which the acylation of the catalyst is faster than
the de-acylation until late in the reaction when most of the
starting material has been consumed.
We have also studied catalyst designs in which the basic
residue is further removed from the hydroxyl group (Table
2).6 The methanolysis of 7 with these catalysts was studied
under the same conditions as before, and in this case, there
is a greater difference in the activity of the trifluoromethyl
Table 2. Relative Rates of Reaction Using Benzene-Derived
Catalysts
(5) (a) Bell, R. P. In AdVances in Physical Organic Chemistry; Gold,
V., Ed.; Academic: New York, 1966; Vol 4, pp 12-16. (b) For data on
the acidity of pyridine carboxaldehyde hydrates, see: Owen, T. C. J.
Heterocycl. Chem. 1990, 27, 987.
(6) See the Supporting Information for the synthesis of these compounds.
(7) Trifluoromethyl-substituted alcohols are known to be more acidic
than alkyl-substituted alcohols. For example, the pKa of trifluoroethanol is
12.4, whereas as the pKa of ethanol is 15.9. We therefore estimate the pKa
of the trifluoromethyl-substituted catalyst 14 to be about 3.5 units lower
than the corresponding methyl-substituted catalyst 15. See: Ballinger, P.;
Long, F. A. J. Am. Chem. Soc. 1960, 82, 795-798
entry
R
R′
k(rel) entry
R
R′
k(rel)
1
2
H
H
CF3 (20)
CH3 (21)
930
25
3
4
Me
H
CF3 (22)
H (23)
1
37
(8) This reaction was conducted at a catalyst loading of 20% in order to
better observe the catalytic species by NMR.
Org. Lett., Vol. 5, No. 22, 2003
4107

vs methyl bearing catalysts (20 vs 21, 37:1 rate difference).
The more dramatic increase in rate for the CF3-bearing
catalyst can be attributed to the fact that the basic residue is
well insulated from the CF3 group. In the pyridine-derived
catalysts, the substitution of an alkyl or aryl group for an
electron-withdrawing perfluoroalkyl or perfluoroaryl group
likely has two opposing effects: while it renders the hydroxyl
group more acidic, we suspect that it also renders the pyridine
less basic due to its proximity to the pyridine. In catalyst
20, there is no decrease in the basicity of the pyrrolidine
upon substitution of the methyl for a trifluoromethyl because
it is far removed from the trifluoromethyl group.
Interestingly, the activating effect of a trifluoromethyl
group is great enough to render reactions with catalyst 20
930 times faster than those with blocked catalyst 22 bearing
a methyl ether in place of the hydroxyl group and 25 times
faster than 23, which bears a primary alcohol and is less
hindered. Reactions with catalyst 20 are also 3.9 times faster
than those with catalyst 14 (Table 1).
where the hydroxyl group and basic residue are insulated
from each other, there is a greater benefit to having an
electron-withdrawing group near the hydroxyl group. Cata-
lysts in which the hydroxyl group is blocked as a methyl
ether are significantly less active, providing good evidence
that the reaction proceeds by an O-nucleophilic mechanism.
Furthermore, the kinetic plots show zero-order behavior in
substrate, with the acyl-catalyst intermediate being the
predominant form of the catalyst, suggesting that attack of
methanol on the acylated catalyst is turnover limiting and
that the rate enhancement in the systems bearing electron-
withdrawing groups is due to the enhanced leaving-group
ability of the alcohol. Further studies to prepare more active
catalysts and to study asymmetric synthesis with these
catalysts are in progress.
Acknowledgment. This work was supported by the Na-
tional Institutes of Health (GM48498) and Array Biopharma.
In conclusion, we have described two classes of O-
nucleophilic acyl transfer catalysts and have shown that
alcohols in close proximity to electron-withdrawing groups
provide more active catalysts. We attribute this effect to an
acidification of the hydroxyl group, leading to more facile
deprotonation by the neighboring base, and to the better
leaving group ability of the more acidic alcohol. In catalysts
Supporting Information Available: Experimental pro-
cedures for kinetic measurements and for the synthesis of
all catalysts and spectral data for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL035510N
4108
Org. Lett., Vol. 5, No. 22, 2003