Mechanistic insights from reaction progress kinetic analysis of the polypeptide-catalyzed epoxidation of chalcone

Article abstract of DOI:10.1021/ol051708r

(Chemical Equation Presented) Reaction progress kinetic analysis of the poly(L)-leucine (PLL)-catalyzed epoxidation of substituted chalcones 1a-1c helps to refine an earlier mechanistic proposal by demonstrating that the reaction proceeds via reversible a

Full text of DOI:10.1021/ol051708r

ORGANIC
LETTERS
2005
Vol. 7, No. 22
4847-4850
Mechanistic Insights from Reaction
Progress Kinetic Analysis of the
Polypeptide-Catalyzed Epoxidation of
Chalcone
Suju P. Mathew,† Suthahari Gunathilagan,‡ Stanley M. Roberts,§ and
Donna G. Blackmond*,†
Department of Chemistry, Imperial College, London SW7 2AZ, United Kingdom,
Department of Chemistry, UniVersity of Hull, Hull HU6 7RX, United Kingdom, and
School of Chemistry, UniVersity of Manchester, M60 1QD, United Kingdom
d.blackmond@imperial.ac.uk
Received July 20, 2005
ABSTRACT
Reaction progress kinetic analysis of the poly(
L
)-leucine (PLL)-catalyzed epoxidation of substituted chalcones 1a
−1c helps to refine an earlier
mechanistic proposal by demonstrating that the reaction proceeds via reversible addition of chalcone to a PLL-bound hydroperoxide, forming
a fleeting hydroperoxy enolate species. Observation of an induction period offers an alternate rationalization for effects formerly attributed to
substrate inhibition. Previous clues about the origin of enantioselectivity in this system are supported by this work.
Reaction protocols for the peptide-catalyzed asymmetric
epoxidation of R,â-unsaturated enones (Julia´-Colonna reac-
tion,1 Scheme 1)2 have advanced into an efficient synthetic
method for oxidation of electron-deficient enones that has
enabled exploration of the intriguing catalytic role of the
polypeptide chain in imparting enantioselectivity.3 Berkessel4
has shown that a minimum of five L-Leu residues is sufficient
for efficient and selective catalysis, and that the helicity of
the peptide determines the configuration of the epoxide. A
recent theoretical study identified a critical role for the three
N-terminal amidic N-H groups of the R-helical polyleucine.5
It has been proposed that the reaction proceeds through the
steady-state random bireactant system shown in Scheme 2.6
Here we report detailed reaction progress kinetic analysis
Scheme 1. Poly(L)-leucine-Catalyzed Epoxidation of Chalcones
† Imperial College.
‡ University of Hull.
§ University of Manchester.
(1) (a) Julia´, S.; Masana, J.; Vega, J. C. Angew. Chem. 1980, 92, 968;
Angew. Chem., Int. Ed. Engl. 1980, 19, 929. (b) Julia´, S.; Guixer, J.; Masana,
J.; Rocas, J.; Colonna, S.; Annuziata, R.; Molinari, H. J. Chem. Soc., Perkin
Trans. 1 1982, 1317. (c) Banfi, S.; Colonna, S.; Molinari, H.; Julia´, S.;
Guixer, J. Tetrahedron 1984, 40, 5207.
(2) (a) Weitz, E.; Scheffer, A. Ber. 1921, 54, 2327. (b) Apeloig, Y.;
Karni, M.; Rappoport, Z. J. Am Chem. Soc. 1983, 105, 2784.
10.1021/ol051708r CCC: $30.25
© 2005 American Chemical Society
Published on Web 09/29/2005

which the rate rises before the steady-state catalytic cycle is
established. This suggests that kinetic studies based on initial
rate data acquired at low conversion may not provide an
accurate description of the relationship between rate and
substrate concentration. Figure 2 confirms that, at conversions
Scheme 2. Proposed Consecutive-Parallel Reaction Network6
from in situ monitoring of this reaction that supports the
major hydroperoxide route and helps to shed light on the
origin of enantioselectivity in this system.
Reactions of chalcone (1a), p-Cl-chalcone (1b), and
p-OCH3-chalcone (1c) with H2O2 were carried out in the
presence of DBU base and soluble PEG-supported poly(L)-
leucine catalyst (PLL)7 in THF. Reaction rate was monitored
as a function of time using reaction calorimetry,8 as shown
in Figure 1. Comparison of fraction conversion determined
Figure 2. Reaction rate versus time for the epoxidation of chalcone
1a (Scheme 1) at three different initial concentrations of 1a. Rate
data obtained from reaction calorimetric profiles at different
conversions of 1a as noted.
below 10%, rates appear to be suppressed at high initial
concentrations of 1a, while above 10% conversion, a well-
behaved first-order relationship between rate and chalcone
concentration is observed. This induction behavior provides
an alternate rationalization for similar rate suppression
observed at high [1a] in the recent initial rate kinetic study
of this reaction by Ottolina and co-workers,6 which they
attributed instead to substrate inhibition at high concentra-
tions of chalcone.
A recent review9 of the methodology of reaction progress
kinetic analysis describes a simple experimental test to probe
for unsteady-state influences on kinetic behavior, such as
substrate or product inhibition or activation, as well as
catalyst activation or deactivation. Reactions are carried out
using different initial concentrations but the same [“excess”],
a variable which is defined as the difference between the
initial concentrations of the oxidant and enone substrates
([“excess”] ) [H2O2]0 - [1]0). Constant catalyst concentra-
tion and steady-state behavior are confirmed when kinetic
profiles for two reactions at the same [“excess”] oWerlay one
another when plotted as rate versus substrate concentration.
Figure 3 shows these plots for reactions of the three
chalcones 1a, 1b, and 1c. At conversions higher than 10-
15%, the kinetic profiles exhibit the “overlay” that confirms
steady-state reaction within the catalytic cycle under these
conditions. Figure 3 thus provides further support for the
suggestion that the initial rate behavior at high chalcone
concentration may be attributed to an induction period that
occurs before the steady-state cycle is established, rather than
to substrate inhibition.
Figure 1. Reaction heat flow and conversion versus time for the
epoxidation of chalcone 1a as in Scheme 1.
by reaction calorimetry and by periodic sampling and HPLC
analysis confirms that the observed heat flow is an accurate
measure of the epoxidation rate.
The reaction progress curve of Figure 1 reveals that a key
feature of the reaction is a brief induction period during
(3) (a) Geller, T.; Roberts, S. M. J. Chem. Soc., Perkin Trans. 1 1999,
1397. (b) Flood, R. W.; Geller, T. P.; Petty, S. A.; Roberts, S. M.; Skidmore,
J.; Volk, M. Org. Lett. 2001, 3, 683. (c) Bentley, P. A.; Kroutil, W.;
Littlechild, J. A.; Roberts, S. M. Chirality 1997, 9, 198. (d) Bentley, P. A.;
Flood, R. W.; Roberts, S. M.; Skidmore, J.; Smith, C. B.; Smith, J. A.
Chem. Commun. 2001, 1616. (e) Tsogeova, S. B.; Wo¨ttinger, J.; Jost, C.;
Reichert, D.; Ku¨hnle, A.; Krimmer, H.-P.; Drauz, K. Synlett 2002, 707.
(4) Berkessel, A.; Gasch, N.; Glaubitz, K.; Koch, C. Org. Lett. 2001, 3,
3839.
(5) Kelly, D. R.; Roberts, S. M. Chem. Commun. 2004, 2018.
(6) (a) Carrea, G. S.; Colonna, S.; Meek, A. D.; Ottolina, G.; Roberts,
S. M. Chem. Commun. 2004, 1412. (b) Carrea, G. S.; Colonna, S.; Meek,
A. D.; Ottolina, G.; Roberts, S. M. Tetrahedron: Asymmetry 2004, 15, 2945.
(7) Kelly, D. R.; Bui, T. T. T.; Caroff, E.; Drake, A. F.; Roberts, S. M.
Tetrahedron Lett. 2004, 45, 3885.
The network shown in Scheme 2 allows for four different
catalytic intermediate species: (i) polyleucine catalyst sites
with no bound substrates (PLL); (ii) PLL bound either by
(8) For examples of kinetic studies using reaction calorimetry, see: (a)
Nielsen, L. P. C. N.; Stevenson, C. P.; Blackmond, D. G.; Jacobsen, E. N.
J. Am. Chem. Soc. 2004, 126, 1360. (b) Rosner, T.; LeBars, J.; Pfaltz, A.;
Blackmond, D. G. J. Am. Chem. Soc. 2001, 123, 1848.
(9) Blackmond, D. G. Angew. Chem., Int. Ed. 2005, 44, 4302.
Org. Lett., Vol. 7, No. 22, 2005
4848

Figure 3. Reaction rate versus substrate concentration for two sets
of reactions carried out at the same [“excess”] value ([“excess”] )
[H2O2]0 - [1]0). Rate given as catalyst-normalized turnover
frequency (rate in M/min)/([PLL] in g/L). Note that in these plots
the temporal direction of the reaction is from right to left (from
high to low concentration of substrate).
chalcone 1a, 1b, or 1c (I1a, I1b, or I1c); (iii) hydroperoxide
-
bound by PLL (IOOH ); and (iv) hydroperoxide-chalcone-
enolate-PLL species (I2a, I2b, or I2c). The relative magnitudes
of the equilibrium constants for forming the various species
dictate the relative populations of these species as well as
the relative contributions of Pathways I and II. Thus, the
model depicted in Scheme 2 potentially involves a large
number of rate and equilibrium constants to describe reac-
tions of the three chalcone substrates. Further mechanistic
information may be extracted from reaction progress kinetic
profiles using the portion of the data obtained after establish-
ment of steady-state behavior. We have shown previously
that kinetic profiles of separate reactions employing at least
two different values of [“excess”] define a unique and
quantitative mathematical solution for a network such as that
shown in Scheme 2.8
Figure 4. Reaction rate versus fraction conversion of substrate as
in Scheme 1 for the reaction of (a) 1a; (b) 1b; and (c) 1c. Reaction
conditions given in Supporting Information. Symbols, experimental
data; solid lines, kinetic model fit to eqs (1a-1c). Values for the
kinetic parameters are given in Table 1 in the Supporting Informa-
tion.
The results of fitting the experimental data to this kinetic
model are shown in Figure 4. The data included from these
reactions provide over 2500 [rate, concentration] data pairs,
each of which is equivalent to a separate initial rate
measurement at a different substrate concentration in reac-
tions of 1a, 1b, and 1c. Because of the presence of an
induction period, only data collected between 30 and 90%
conversion are used in modeling. The clear difference
between the model prediction and experimental data at low
conversion in Figure 4 reiterates the importance of decon-
voluting the transient initial behavior from that of the steady-
state kinetics.
This simplified kinetic model reveals important implica-
tions for the proposed mechanism of Scheme 2. Rate profiles
for all three chalcones fit to a single value for a binding
constant for hydroperoxide binding (KOOH ) 24.3 M-1,
standard deviation 1.24%); no binding constants related to
chalcone association are obtained, showing that binding of
any of the chalcones must be at least an order of magnitude
weaker than hydroperoxide binding to PLL. This suggests
that the chalcone-bound Pathway II is kinetically negligible
at practical chalcone concentrations, and that the reaction
proceeds exclusively through the hydroperoxide Pathway I.
The kinetic study of Ottolina and co-workers6 suggested that
Pathway I is kinetically preferred, but their kinetic data were
unable to quantitate the extent of this preference. Our kinetic
model indicates that there is no statistical justification for
invoking even a minor contribution from the chalcone-bound
Most strikingly, these data gave an excellent, statistically
significant fit (R2 ) 0.997, SSE ) 0.0001) to a substantially
simplified form of the kinetic model comprising one rate-
determining step rate constant specific for each chalcone
substrate and one common binding constant, that for PLL
binding to the hydroperoxide according to the eq 1:
kcat[1][H2O2][PLL]total
rate )
(1)
1 + KOOH- [H2O2]
Org. Lett., Vol. 7, No. 22, 2005
4849

pathway even at high [chalcone]:[OOH-] ratios. The differ-
ence in reactivity between the chalcones arises from a
difference in the rate at which each chalcone successfully
interacts with the bound hydroperoxide species. The kinetic
measure of this rate is given by kcat,1a, kcat,1b, and kcat,1c (values
of 0.54, 1.19, 0.27 L2 mol-1 min-1, standard deviations
<0.85%), reflecting the net forward rate of forming the
hydroperoxy enolate.
the concentration of hydroperoxy enolate never builds up
during the course of the reaction; the forward rate of
formation of this species is rate-limiting because this rate is
much lower than the rates of either of the two possible
choices (dissociation or OH- elimination) for consumption
of this species. Establishment of this preliminary equilibrium
allows the hydroperoxy enolate species to be involved in
the rate-limiting step without representing the resting state
of the catalyst within the cycle.
That formation of the hydroperoxy enolate species is rate-
determining correlates with a linear free energy relationship.
Even with this narrow range of Hammett substitution
Kelly and Roberts5 proposed a structure for a species, such
as I1a, based on theoretical calculations, but its energy relative
to free PLL or to other species in the network was not
reported.5 However, such a bound species is unlikely to be
able to achieve the orientation required to produce the
hydroperoxide-chalcone enolate when attacked with OOH-
without first dissociating. Thus, chalcone binding would
suppress rate by occupying catalyst sites in an inactive state.
Our studies reveal both the absence of substrate inhibition
and the absence of a kinetically meaningful concentration
of bound chalcone. The kinetic picture of our work, sug-
gesting a rapid and reversible chalcone “docking” to bound
hydroperoxide species leading predominantly back to free
chalcone andsless often but with complete stereochemical
specificitysforward to form the epoxychalcone, is more
compatible with the comprehensive body of experimental
data on this system.
constants (σp-Cl ) 0.23, σp-OCH ) -0.27), these relative
3
rate constants give a significant positive reaction constant
(F ) 2.95; R2) 0.992), suggesting that the rate is accelerated
when negative charge is more efficiently removed from the
ring in the rate-determining step.
The model predicts that the only intermediates exhibiting
kinetically meaningful concentrations are the PLL catalyst
-
itself and the hydroperoxide-bound species, IOOH , with this
hydroperoxide species exhibiting ca. 3 times greater con-
centration than that of the unbound PLL catalyst. This
supports previous results showing that the PLL catalyst is
able to sequester significant concentrations of hydroper-
oxide.10 The model also reveals that the hydroperoxy enolate
species exhibits only a fleeting concentration. This result
requires some discussion since it appears to contradict
previous suggestions that rapid, reversible addition of hy-
droperoxide to chalcone forms the enolate, which ultimately
gives the epoxychalcone product via slow, intramolecular
displacement of hydroxide.5 Roberts and co-workers recently
reported that the hydroperoxy enolate forms rapidly via
random facial addition under reaction conditions.10a The
reaction proceeds forward only if the conformation of the
hydroperoxy enolate thus formed is suitable for elimination.
These authors suggested that in a majority of additions failure
to achieve the proper orientation leads to a much faster
backward rate (giving isomerization) than forward rate
(giving epoxidation). This is consistent with a model where
This work demonstrates that reaction progress kinetic
analysis, in conjunction with other characterization tools,
provides significant mechanistic insight, helping to distin-
guish between different mechanistic proposals and offering
clues about the origin of enantioselectivity in this peptide-
catalyzed system.
Acknowledgment. Funding from EPSRC/DTI/LINK
(“From Micrograms to Multikilos”) is gratefully acknowl-
edged.
Supporting Information Available: Experimental and
kinetic modeling procedures (7 pages). This material is
available free of charge via the Internet at http://pubs.acs.org.
(10) (a) Kelly, D. R.; Caroff, E.; Flood, R. W.; Heal, W.; Roberts, S. M.
Chem. Commun. 2004, 2016. (b) Lopez-Pedrosa, J.-M.; Pitts, M. R.; Roberts,
S. M.; Saminathan, S.; Whittall, J. Tetrahedron Lett. 2004, 45, 5073.
OL051708R
4850
Org. Lett., Vol. 7, No. 22, 2005