Combinatorial Catalysis of an Elimination Reaction

Full text of DOI:10.1021/jo9812431

7
578
J . Org. Chem. 1998, 63, 7578-7579
Com bin a tor ia l Ca ta lysis of a n Elim in a tion
Rea ction
Sch em e 1. Libr a r y of Am in es Used to F or m
Com bin a tor ia l P olym er s
Fredric M. Menger,* J ulia Ding, and
Veronique Barragan
Department of Chemistry, Emory University,
Atlanta, Georgia 30322
Received J une 26, 1998
We have recently been engaged in the screening of
randomly generated polymers for their catalytic or reactive
potential.1 Among thousands of such screenings (assisted
by automation), a few polymers were found that are remark-
ably effective, and these were further “fine-tuned” to make
them even better catalysts. The new approach was prompted,
in part, by a feeling that the time has come to supplement,
but by no means supplant, the traditional deterministic
strategy in which complex enzyme models are synthesized
often at great expense.
,2
basic groups (to assist departure of the proton) would likely
catalyze the dehydrationsespecially if the acid-base groups
could adopt the proper spatial orientation. It was the
purpose of this work described herein to develop such
polymers.
Polymers were formed by reacting poly(acrylic anhydride)
simultaneously with three or four amines,4 taken from a
library of the 11 amines in Scheme 1, according to eq 2. The
Previous work described a polymer that catalyzes a
phosphate hydrolysis by a factor of 30 000 (exceeding that
1
of a catalytic antibody for the same reaction). In the present
report, we develop polymers that catalyze a totally different
reaction (an elimination), thereby demonstrating that the
hydrolase activity was not simply a rare happenstance. As
will be seen, the elimination catalysts display a slow
induction period as if the polymers “learn” in situ how to
promote the reaction.
An analogy between the combinatorial approach and that
of Nature, although somewhat tongue-in-cheek, serves to
illustrate our philosophical underpinnings: For billions of
years Nature has also been evolving her catalysts, called
enzymes, in a random manner. Thus, DNA molecules (and,
hence, enzymes) were mutated, and once in great while, a
mutation would create an improved enzyme that endowed
the recipient with a survival advantage over those without
it. By this means, new and improved enzymes were retained
(
just as we retain those randomly produced polymers that
possess desirable properties). While Nature has utilized
survival of the fittest”, our method invokes the “survival of
poly(acrylic anhydride), prepared from acrylic anhydride and
AIBN,4 had an average molecular weight of about 4500
“
(MALDI/TOF mass spectrometry). Fourteen combinations
the fastest” (with ourselves being the agent of selection). And
while Nature amplifies successful enzymes via a process
called life, we amplify our successful catalysts by producing
them in large amounts once their formulas are discovered
by the screening process. Among the more regretful failures
of this analogy is the vast difference in time frame available
to Nature and to the mortal chemist.
of amines (e.g., His/Abn/Hex; His/Pyr/Phe/Cap) were exam-
ined in 96 different ratios each for a total of 1344 polymer
variations. Twenty percent or less of the carboxyl groups
in any given poly(acrylic acid) sample were derivatized (e.g.,
5
% His/5% But/5% Nap). Those polymers that were water-
insoluble in pH ) 7.0 buffer were discarded. A few experi-
mental details below will clarify the procedure.
Liquid deliveries were assisted by an ICN QuadFlex
computer-controlled automatic pipetting system. Thus, poly-
A recent objective was to develop a combinatorial polymer
capable of catalyzing the dehydration of a â-hydroxy ketone
(eq 1). This was an attractive reaction because it allowed
(acrylic anhydride) (100 µL, 0.05 M monomer units in
DMSO) was pipetted into 96 test tubes (12 × 75 mm),
whereupon the DMSO was removed at 40 °C by speed
evaporation (J ouan RC 10-10). The following was then
added to each test tube: (a) 500 µL of pH ) 12 buffer and
(b) three or four amine solutions (25-150 µL of 5 mM each)
as dictated by 96 computerized combinations. After bath
sonication for 30 min (sufficient, according to NMR, to
complete the aminolysis), the water was removed by speed
evaporation, leaving a waxy residue that was dissolved in 1
mL of pH ) 7.0 phosphate buffer (0.05 M) to make a solution
easy monitoring of multiple samples by an increase in UV
absorbance. Moreover, related eliminations are found bio-
logically (e.g., citrate to aconitate or malate to fumarate).
As with the corresponding enzymes, polymers that possess
both acidic groups (to assist departure of the hydroxyl) and
3
5
mM in monomer units. The following were then pipetted
into the 96 400 µL wells of a microtiter plate: (a) 50 µL of
the derivatized polymer solution; (b) 125 µL of 2 × 10
(
1) Menger, F. M.; Eliseev, A. V.; Migulin, V. A. J . Org. Chem. 1995, 60,
666-6667.
2) Menger, F. M.; West, C. A.; Ding, J . Chem. Commun. 1997, 633-
34.
3) Rose, I. A.; O’Connell, E. J . Biol. Chem. 1967, 242, 1870-1879.
-4
M
6
6
(
(4) Choi, S.; Mammen, M.; Whitesides, G. M. J . Am. Chem. Soc. 1997,
119, 4103-4111.
(
1
0.1021/jo9812431 CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/07/1998

Communications
J . Org. Chem., Vol. 63, No. 22, 1998 7579
F igu r e 1. Change in absorbance at 340 nm for eq 1 (pH ) 7.0,
F igu r e 2. Change in absorbance at 340 nm for eq 1 (pH ) 7.0,
23 °C) catalyzed by combinatorial polymer with 5% His/5% But/
5% Nap. Note the striking induction period in which there is no
detectable reaction for the first 24 h.
2
3 °C) catalyzed by polymer with 5% His/5% Oct/5% Phe/5% Cap
(squares) as compared to a polymer with 5% His/7.5% But/2.5%
Nap (circles). These are examples of “fast” and “slow” polymers,
respectively. Less than 1% of the screened polymers are “fast”.
combinatorial polymers have an advantage over antibodies
in that they are readily prepared with no input of biotech-
nology.
One of the most interesting aspects of our investigations
relates to the presence of an induction period evident in
Figures 1 and 2. It is as if the polymers “learn” to catalyze
after, roughly, a 24 h time period of inactivity. To better
define the system, a variety of experiments were carried out
with 5% His/5% Oct/5% Phe/5% Cap, and the results are
now briefly summarized: (a) Rates are twice as fast at pH
substrate 1 in water-acetonitrile (9.5/0.5 v/v); (c) 75 µL of
a pH ) 7.0 phosphate buffer. Thus, the catalysis was
screened with solutions whose final concentrations were 1.2
-
5
×
10 M polymer (assuming a MW ) 6100 for the deriva-
4
-
tized material) and 1.0 × 10 M substrate. Monitoring of
the 96 wells was carried out periodically at 340 nm for the
appearance of the â-aryl-R,â-unsaturated ketone using an
ICN MCC/340 scanner. Reactions run to completion gave
infinity absorbances consistent with a calculated infinity and
with an absorbance measured on a plate with purified
product 2. The presence of 2 in an actual run was also
verified by GC-MS.
)
6.0 as at pH ) 7.0, while rates at pH ) 5.0 are very slow.
(
b) A remarkable temperature sensitivity was observed: The
fast rates at 23 °C are killed at both 0 and 40 °C. Normal
kinetic curves are restored when a substrate/polymer solu-
tion at 40 °C is cooled back to 23 °C. (c) Sonication of
substrate/polymer solutions also destroys activity. (d) Cross-
linking the polymer via addition of 1% 1,7-diaminoheptane
Reactions range from “slow” (∆A ) 96 h absorbance
change ) 0.07-0.15) to “fast” (∆A g 0.4 corresponding to
>
2t1/2’s). Figure 1 gives an example of each in which 5%
His/7.5% But/2.5% Nap is slow, whereas 5% His/5% Oct/5%
Phe/5% Cap is fast. Less than 1% of the 1344 runs can be
categorized as “fast”. The catalytic ability of a polymer
depends not only upon the types of substituents but upon
the ratio of substituents within one particular substituent
set. For example, ∆A ) 0.35 for 5% His/2.5% Cap/7.5% Pyr/
(along with the four substituent amines) greatly retards the
rate. (e) Induction periods persisted even when the polymers
were allowed to remain in solution for 4 days prior to
addition of the substrate.
The above observations are most easily explained in terms
of a substrate-induced transformation into a catalytically
active conformation. At 0 °C, the polymer transformation
is slow, while heating to 40 °C or sonication disrupts the
active conformation. Cross-linking rigidifies the polymer
and impedes its conformational rearrangement necessary to
achieve catalysis. It appears, therefore, that we are observ-
ing a type of nonbiological “induced fit”.6
5
1
% Leu, while ∆A ) 0.14 for 5% His/2.5% Cap/2.5% Pyr/
0% Leu. Deletion of one substituent, but with no other
change, can have a substantial effect upon the rate. For
example, removing Pyr from 5% His/10% Pyr/5% Phe/5%
Hex diminishes ∆A from 0.36 to 0.18. Fast rates were
duplicated independently by two people who prepared the
polymers identically. Preliminary attempts to purify fast
polymers (Sephadex G-75 and Superdex 30 Prep) failed to
give fractions with additional activity. Since control experi-
ments with suitable mixtures of free amines were found to
give no rate enhancement, catalysis is predicated upon
attachment of the functionalities onto the polymer frame-
work (a conclusion that also mitigates any concern that
catalysis arises from traces of unreacted amine).
Combinatorial polymeric catalysis is a new field, and
naturally there remain a host of unanswered questions (e.g.,
catalyst specificity, substrate binding, laboratory-to-labora-
tory reproducibility, scale-up behavior, etc.). Yet our pre-
liminary results affirm our conviction that, at a time when
water is becoming more and more desirable as an industrial
solvent, “fortuitous” catalysis with water-soluble polymers
Antibody-catalyzed dehydration of our substrate at pH )
7
systems is an attractive area for further investigation.
7
.0 (37 °C) affords a catalysis of 1200 above the background
5
reaction. Our best dehydrations are comparable, i.e., a
catalysis of 920 at pH ) 7.0 (23 °C) on the basis of the nearly
linear rate ensuing after 24 h. Thus, combinatorial polymers
Ack n ow led gm en t . This work was supported by the
National Science Foundation.
(or antibodies for that matter) do not measure up to an
J O9812431
enzyme. Yet it should be borne in mind that we have hardly
begun to sample the available polymer combinations includ-
ing those containing phenols, thiols, metals, etc. And
(
6) Koshland, D. E., J r., In The Enzymes; Academic Press: New York,
1959, Vol. 1, Chapter 7.
7) For an entirely different approach to the subject, see: Brocchini, S.;
(
J ames, K.; Tangpasuthadol, V.; Kohn, J . J . Am. Chem. Soc. 1997, 119,
4553-4554.
(5) Uno, T.; Schultz, P. G. J . Am. Chem. Soc. 1992, 114, 6573-6574.