Pyridylalanine (Pal)-peptide catalyzed enantioselective allenoate additions to N-acyl imines proceed via an atypical aza-Morita-Baylis-Hillman mechanism

Article abstract of DOI:10.1021/ol101947s

Mechanistic experiments, including kinetics and hydrogen/deuterium kinetic isotope effects, reveal an atypical rate-determining step in a pyridylalanine-peptide catalyzed enantioselective coupling of allenoates and N-acyl imines. Typically, acrylates participate in both the aldehyde-based Morita-Baylis-Hillman (MBH) reaction and the imine-based variant (the aza-MBH ) through similar mechanisms, with proton transfer/catalyst regeneration often rate-determining. In contrast, the title reaction exhibits kinetics wherein proton transfer is kinetically silent.

Full text of DOI:10.1021/ol101947s

ORGANIC
LETTERS
2010
Vol. 12, No. 21
4800-4803
Pyridylalanine (Pal)-Peptide Catalyzed
Enantioselective Allenoate Additions to
N-Acyl Imines Proceed via an Atypical
“aza-Morita-Baylis-Hillman”
Mechanism
Lindsey B. Saunders, Bryan J. Cowen, and Scott J. Miller*
Department of Chemistry, Yale UniVersity, P.O. Box 208107, New HaVen,
Connecticut 06520, United States
scott.miller@yale.edu
Received August 17, 2010
ABSTRACT
Mechanistic experiments, including kinetics and hydrogen/deuterium kinetic isotope effects, reveal an “atypical” rate-determining step
in a pyridylalanine-peptide catalyzed enantioselective coupling of allenoates and N-acyl imines. Typically, acrylates participate in both
the aldehyde-based “Morita-Baylis-Hillman (MBH)” reaction and the imine-based variant (the “aza-MBH”) through similar mechanisms,
with proton transfer/catalyst regeneration often rate-determining. In contrast, the title reaction exhibits kinetics wherein proton transfer
is kinetically silent.
We recently reported a peptide-catalyzed enantioselective
coupling of N-acyl imines 1 with allenic esters 2 (eq 1).1
The reaction is catalyzed by peptide 3, containing a Lewis
basic pyridylalanine residue, and affords allene-substituted
amides such as 4 in high yield and enantiomeric ratio (er).
This transformation may be considered an allenoate aza-
Morita-Baylis-Hillman (aza-MBH) reaction.2
The MBH reaction is a process that couples activated
(1) (a) Cowen, B. J.; Saunders, L. B.; Miller, S. J. J. Am. Chem. Soc.
2009, 131, 6105. (b) For a report on the DABCO catalyzed coupling of
allenoates with N-Boc imines, see: Guan, X.-Y.; Wei, Y.; Shi, M. J. Org.
olefins with aldehydes, often using a Lewis basic catalyst,3
typically an amine or phosphine.4 It has attracted considerable
Chem. 2009, 74, 6343.
attention, especially as a platform for asymmetric catalysis,
and many examples have been reported.5 Insight into the
mechanism of the MBH reaction has been gleaned through
detailed experimental and computational work.6 For example,
(2) For reviews on the aza-MBH reaction, see: (a) Declerck, V.;
Martinez, J.; Lamaty, F. Chem. ReV 2009, 109, 1. (b) Shi, Y.-L.; Shi, M.
Eur. J. Org. Chem. 2007, 2905.
(3) Denmark, S. E.; Beutner, G. L. Angew. Chem., Int. Ed. 2008, 47,
1560.
10.1021/ol101947s  2010 American Chemical Society
Published on Web 09/28/2010

McQuade and Aggarwal independently showed that, in the
absence of protic additives, certain MBH reactions are first
order in an amine catalyst and in acrylate, but second order
in aldehyde.7 It is proposed that the second equivalent of
aldehyde assists in the proton transfer step that follows
carbon-carbon bond formation (Figure 1). It has also been
Figure 2. Proposed catalytic cycle for the aza-MBH reaction as
reported by Jacobsen. The quoted KIE, with xylenes as the reaction
solvent, is for the acrylate R-proton (bold).
reaction with methyl acrylate and N-nosyl imines displays a
large primary KIE with kH/kD ) 3.81, even in the relatively
nonpolar solvent of xylenes.8a
Because of the complexity and lack of mechanistic
generality of such reactions, we were interested to determine
the kinetic parameters of the allenoate-imine coupling
reaction. We report here that the allenoate aza-MBH reactions
we have studied exhibit unique kinetic parameters and a
mechanistic pathway that is distinct from several other aza-
MBH reactions that have been previously studied.
Figure 1. Proposed catalytic cycle for the MBH reaction in the
absence of protic additives as reported by McQuade and Aggarwal.
Quoted KIEs are for the acrylate R-proton (bold).
We first carried out a series of rate studies employing
allenoate 2a, N-acyl imine 1 (R ) H), and catalyst 3. The
kinetic order of each component was determined by con-
structing plots of kobs versus concentration (Figures 3, 4).9
The allenoate and catalyst plots each show a good linear
correlation between the initial rate constant and the substrate
or catalyst concentration, respectively (Figure 3). In addition,
when the data are fit to a power curve, the exponent in each
case is approximately equal to 1 (Table 1). We therefore
conclude that the reaction is first order in allenoate and first
order in peptide under the conditions we examined.
On the other hand, the plot of kobs versus imine concentra-
tion shows essentially a constant reaction rate for all of the
imine concentrations evaluated (0.01-0.15 M).10 This
implies that the reaction could be zero order in imine. The
kinetic order of the imine electrophile in this reaction is
unusual in comparison to that observed for most other aza-
MBH reactions8 and in comparison to the aldehyde-based
classical MBH.6 Only recently have examples of aza-MBH
processes been documented in which the imine is not found
to factor into the rate-determining step.11 Our observations
suggest that the peptide-catalyzed allenoate variant follows
this “atypical” or at least more recently documented pathway.
One explanation for the apparent zero order dependency
of the reaction on imine concentration is the possibility of
established that MBH reactions tend to show primary kinetic
isotope effects (KIEs) for the R-proton of the acrylate or
activated olefin, which supports proton transfer as being the
rate-determining step in the reaction mechanism. However,
the magnitude of the KIE may depend on the exact nature
of the substrates, in addition to the polarity of the solvent
(kH/kD ) 1.0-5.2).6
Kinetic studies on the aza-MBH reaction by Jacobsen8a and
by Leitner8b have also raised intriguing mechanistic proposals.
These reactions were found to be first order in catalyst and in
acrylate, which is in strong analogy to the classical MBH
reaction. However, the reactions involved imines that either
show rate saturation8a or exhibit a kinetic order of 0.5.8b
Nevertheless, for these aza-MBH reactions, the proton transfer
step has also been shown to be unambiguously rate-determining,
in analogy to many MBH reactions involving aldehydes (Figure
2). A key piece of evidence in support of this finding is that
the diazabicyclo[2.2.2]octane (DABCO) catalyzed aza-MBH
(4) For reviews on the MBH reaction, see: (a) Basavaiah, D.; Rao, K. V.;
Reddy, R. J. Chem. Soc. ReV. 2007, 36, 1581. (b) Basavaiah, D.; Rao, A. J.;
Satyanarayana, T. Chem. ReV. 2003, 103, 811.
(5) For a review, see: Masson, G.; Housseman, C.; Zhu, J. Angew. Chem.,
Int. Ed. 2007, 46, 4614.
(6) (a) Hill, J. S.; Isaacs, N. S. J. Phys. Org. Chem. 1990, 3, 285. (b)
Price, K. E.; Broadwater, S. J.; Walker, B. J.; McQuade, D. T. J. Org.
Chem. 2005, 70, 3980. (c) Price, K. E.; Broadwater, S. J.; Jung, H. M.;
McQuade, D. T. Org. Lett. 2005, 7, 147. (d) Aggarwal, V. K.; Fulford,
S. Y.; Lloyd-Jones, G. C. Angew. Chem., Int. Ed. 2005, 44, 1706. (e)
Robiette, R.; Aggarwal, V. K.; Harvey, J. N. J. Am. Chem. Soc. 2007, 129,
15513.
(9) For reaction rate plots at the various concentrations, see Supporting
Information.
(10) The imine concentration was not raised above 0.15 M due to its
insolubility in toluene at higher concentrations. The slightly lower value of
the rate constant at 0.15 M may be due to solubility issues.
(11) The following case exhibits similar kinetics, in the presence of protic
additives. See: Yukawa, T.; Seelig, B.; Morimoto, H.; Matsunaga, S.;
Berkessel, A.; Shibasaki, M. J. Am. Chem. Soc. 2010, 132, 11988.
(7) In the presence of protic additives, the reaction becomes first order
in aldehyde. See refs 6d and 6e.
(8) (a) Raheem, I. T.; Jacobsen, E. N. AdV. Synth. Catal. 2005, 347,
1701. (b) Buskens, P.; Klankermayer, J.; Leitner, W. J. Am. Chem. Soc.
2005, 127, 16762.
Org. Lett., Vol. 12, No. 21, 2010
4801

Table 1. Kinetic Order of Allenoate and Peptidea
kinetic
linear fit
exponent in power fit
compound order correlation constant (correlation constant)
2a
3
1
1
0.985
0.989
0.998 (0.982)
0.993 (0.976)
a Linear and power fits as well as correlation constants were calculated
using the trendline feature on Microsoft Excel. See Supporting Information
section for plots with power trendlines.
the reaction conditions. On the other hand, rate competition
experiments with electronically perturbed imines reveal that
a p-bromo-substituted imine is consumed faster than a
p-methoxy-substituted imine (p-Br ∼3 times faster than
p-OMe). This result could be consistent with saturation by
the imine, albeit at a quite low imine concentration.
We also explored isotope effects in the allenoate-imine
coupling reaction. In particular, we probed the allenoate
R-proton in order to establish any KIE. Intriguingly, and in
contrast to the previously studied aza-MBH reactions dis-
cussed above,8 kH/kD for the allenoate R-proton was 1.08
(Table 2).12 This small KIE implies that the proton transfer
step is not likely to be rate-determining.
Figure 3. Kinetic order plots. (Top) Initial observed rate constant
versus concentration of allenoate 2a. Concentration of imine 1 was
maintained at 0.10 M. Concentration of peptide 3 was maintained
at 0.01 M. (Bottom) Initial observed rate constant versus concentra-
tion of peptide 3. Concentration of allenoate 2a was maintained at
0.15 M. Concentration of imine 1 was maintained at 0.10 M. All
values of kobs were measured by plotting the relative product
formation (ratio of 4a to 2,3-dimethylnaphthalene, the internal
standard) with values e0.20 versus time at the specific concentra-
tions of allenoate or peptide, respectively.10 Linear fit equations
and correlation constants are included. The ( error bar for each
value of kobs is equal to the corresponding standard deviation from
the multiple trials.
Table 2. KIEs for Allenoate R-Protona
parameter average value trial 1 trial 2 trial 3
SD
kH (min-1
)
0.0129
0.0119
1.08
0.0119 0.0131 0.0137 0.000917
0.0110 0.0123 0.0125 0.000814
kD (min-1
kH/kD
)
1.08
1.07
1.10
0.0155
a For each trial, the protio- and deutero-allenoate reactions were run
simultaneously to minimize variables such as minor changes in temperature
or humidity. The average value of kH/kD is taken as the overall KIE for the
allenoate R-proton.
On the basis of the results described above, the rate
equation for the allenoate aza-MBH reaction could be
represented simplistically by eq 2 within the concentration
range of our measurements.
d[P]/dt ) kobs[allenoate][peptide]
(2)
A possible mechanism for the allenoate aza-MBH reaction
is shown in Figure 5. Peptide 3 first adds to allenoate 2a to
form zwitterionic intermediate I. Since the kinetic data
support the absence of the imine in the rate-determining step,
the catalyst addition to the allenoate could be rate-determin-
ing, in accord with eq 2. Intermediate I then adds to imine
1 to form a new C-C bond and intermediate II. The next
step involves a fast proton transfer to form zwitterionic
intermediate III, which then eliminates the catalyst to form
coupled product 4a. In the case where there could be
saturation of the catalyst by imine, C-C bond formation
could be rate-determining, with the reaction pseudo-zero
Figure 4. Kinetic order plot. Initial observed rate constant versus
concentration of imine 1. Concentration of allenoate 2a was
maintained at 0.15 M. Concentration of peptide 3 was maintained
at 0.01 M. See Figure 3 legend and Supporting Information for
additional details.
mesauring rates at catalyst saturation by the imine. Our
1
experiments do not exclude this possibility. However, H
NMR spectra of the comixed catalyst and imine reveal
minimal ∆δ (0 to <0.05 ppm; relative to the independent
spectra) for proton resonances at concentrations relevant to
(12) For individual reaction rate plots for each trial and the procedure
for the preparation of allenoate 2a-D, see Supporting Information.
4802
Org. Lett., Vol. 12, No. 21, 2010

Table 3. KIEs for Allenoate R-Proton with Pyridine Catalysta
parameter average value
trial 1
trial 2
SD
kH (min-1
)
0.00099
0.00096
1.03
0.00097 0.00102 0.000035
0.00096 0.00097 0.0000071
kD (min-1
kH/kD
)
1.01
1.05
0.028
a See Table 2 legend and Supporting Information for details.
lack of a primary KIE supports the assertion that a step other
than proton transfer is rate-determining in the pyridine-
catalyzed case. Moreover, the pyridine-catalyzed reaction is
significantly slower than the peptide-catalyzed variant,
indicating that the peptide may be better able to activate the
substrate. This fact also suggests that pyridine and the Pal-
based catalysts may themselves function through different
mechanisms, with different rate-determining steps.
The exclusion of the proton transfer step as rate-determin-
ing brings new insight to some of our previously discovered
empirical observations. For example, a decrease in temper-
ature (from 23 to 0 °C) leads to slower reactions. However,
with allenoate 2b, where the benzyl ester has been changed
to a phenyl ester, the rate is significantly faster such that
high yields of product are obtained in only 1 h at 0 °C.1a
Since 2b may contain a more electrophilic sp-hybridized
carbon than that of benzyl ester 2a, catalyst addition could
be faster with 2b than with 2a. These rate differences could
be consistent with catalyst addition being the rate-determining
step.
In summary, we have determined that the allenoate aza-
MBH follows a divergent mechanism in comparison to
several other MBH and aza-MBH reactions. Our experiments
are consistent with an alternative rate-determining step in
which events prior to proton transfer appear to be the
bottleneck in these reactions. The nature of the zwitterionic
intermediates, with possible electronic delocalization due to
the presence of the allenoate, in addition to other factors,
could contribute to this behavior.
Figure 5. Possible mechanism for the allenoate-imine coupling.
order in imine. Nonetheless, in either case, it appears that
neither proton transfer nor catalyst regeneration are rate-
determining.
The difference in the rate-determining step of the allenoate
aza-MBH versus other acrylate aza-MBH8a reactions is
notable, especially considering that the reactions were
conducted in similar solvents (toluene vs xylenes; no protic
additives). These analogies imply that the nature of the
allenoate, the catalyst, or a combination of factors could cause
the mechanistic dichotomy. One proposal for the relative ease
in proton transfer in the allenoate reaction, as compared to
the acrylate reaction, is that intermediate III (Figure 5),
formed upon proton transfer, contains a resonance stabilized
allylic anion in the formal zwitterionic species.13 The
analogous intermediate in an acrylate-based aza-MBH reac-
tion is not allylic. The allylic nature could stabilize inter-
mediate III relative to intermediate II, lowering the activation
energy associated with proton transfer, and therefore make
this step substantially faster in the allenoate reaction.
We wished to further establish if the alternative mechanism
is due to the nature of the pyridyl moiety of the catalyst
(e.g., in comparison to DABCO), the peptide-based func-
tionality, or the nature of the allenoate. Therefore, we ran
an analogous KIE study with pyridine as the catalyst, rather
than peptide 3. It was found that the kH/kD for the allenoate
R-proton was 1.03 (Table 3), nearly identical to that observed
with peptide-catalyst 3.12 As in the 3-catalyzed reaction, this
Acknowledgment. We are grateful to the National Science
Foundation (CHE-0848224) for support. We thank Professor
Kenneth B. Wiberg (Yale University) and Professor Seth
Herzon (Yale University) for discussions.
Supporting Information Available: Kinetic procedures
and characterization data. This material is available free of
charge via the Internet at http://pubs.acs.org.
(13) For a discussion of related P-initiated zwitterions, see: Khong, S. N.;
Tran, Y. S.; Kwon, O. Tetrahedron 2010, 66, 4760, references therein.
OL101947S
Org. Lett., Vol. 12, No. 21, 2010
4803