Mechanistic rationalization of organocatalyzed conjugate addition of linear aldehydes to nitro-olefins

Article abstract of DOI:10.1021/ja203660r

Kinetic studies of the conjugate addition of propanal to nitrostyrene catalyzed by diarylprolinol ethers reveal that formation of the product iminium species is rate-determining and is promoted by both the reaction product and acid additives. The beneficial role of a dominant cyclobutane intermediate in maintaining high stereoselectivity is highlighted. This mechanistic understanding led to the design of highly productive reaction protocols.

Full text of DOI:10.1021/ja203660r

COMMUNICATION
pubs.acs.org/JACS
Mechanistic Rationalization of Organocatalyzed Conjugate Addition
of Linear Aldehydes to Nitro-olefins
Jordi Burꢀes,† Alan Armstrong,‡ and Donna G. Blackmond*,†
†Department of Chemistry, The Scripps Research Institute, La Jolla, California 92037, United States
‡Department of Chemistry, Imperial College, London SW72AZ, U.K.
S Supporting Information
b
Scheme 1. Conjugate Addition Catalyzed by Diarylprolinol
Ethers
ABSTRACT: Kinetic studies of the conjugate addition of
propanal to nitrostyrene catalyzed by diarylprolinol ethers
reveal that formation of the product iminium species is rate-
determining and is promoted by both the reaction product
and acid additives. The beneficial role of a dominant
cyclobutane intermediate in maintaining high stereoselec-
tivity is highlighted. This mechanistic understanding led to
the design of highly productive reaction protocols.
regime is followed by a much slower but accelerating rate (Figure 1,
right inset). This behavior is reminiscent of the product-induced
rates we observed in proline-catalyzed aminoxylation and R-amina-
tion reactions, where we have shown that the reaction product as
well as protic additives can accelerate the rate.8 Further kinetic
experiments by NMR spectroscopy confirm that rate is accelerated
by addition of either the reaction product 3 or acetic acid (Figure 2,
presented as product concentration vs time). In the absence of
additives, the accelerating rate, signified by the sigmoidal shape of
the product concentration curves of Figure 2, is unaffected by
changes in the initial concentration of either substrate 1 or 2,
indicating that the intrinsic kinetics of the cycle is zero order in both
substrate concentrations.
Figure 2 shows that for reactions with added acid or reaction
product, temporal conversion profiles are no longer sigmoidal but
become linear. The reaction rate is directly proportional to the
concentration of added acid up to 1 equiv compared to catalyst 4,
after which it exhibits saturation behavior, as shown in Figure 3.
Product enantioselectivity (98% ee) is constant over the course of
the reaction and is unaffected by changes in substrate concentrations
or by the addition of acid or reaction product.7 The syn/anti ratio
remains constant at >95:5 over the course of the reaction until very
high conversion, in both the presence and absence of added acid
(Figure 4a). However, this ratio deteriorates when the reaction
product remains in the presence of the catalyst after completion of
the reaction, equilibrating to 60:40 syn/anti overnight. Acid accel-
erates this post-reaction erosion of diastereoselectivity (Figure 4b).
The observation of zero-order kinetics in both [1] and [2]
implies that the rate-determining step occurs in the cycle after
addition of both substrates and that the resting state of the catalyst
contains both substrates. The correlation between the initial
spike in reaction heat flow and catalyst concentration suggests an
initial rapid buildup to a steady-state catalyst resting state that
occupies the major part of the total catalyst concentration.
Consideration of the generally accepted mechanism of enam-
ine-based organocatalysis1 shows that these observations rule out
tereoselective organocatalyzed conjugate addition has be-
S
come a benchmark reaction for assessing the performance
of new catalysts and in the development of cascade reaction
sequences for the synthesis of complex molecules from simple
building blocks.1 The reaction of preformed enamines in stoi-
chiometric Michael additions to nitro-oelfins was studied ex-
haustively by Seebach2 in the 1980s, and that work inspired the
development of amine-catalyzed versions of stereoselective con-
jugate addition (Scheme 1). Hayashi reported the Michael
addition of aldehydes to nitroalkenes catalyzed by diarylprolinol
ethers such as 4 with high enantio- and diastereoselectivity and
good substrate scope.3,4 Since then a number of reports have
appeared of stereoselective conjugate additions using a variety
of catalysts and substrates.5,6 The mechanism of the catalytic
reaction has not been well-studied, and most groups have relied
on analogy to the stoichiometric reactions and on information
gained from other examples of enamine catalysis, in particular the
proline-mediated aldol reaction, to design organocatalyzed con-
jugate addition reactions. For example, acid additives are widely
reported to influence rate and diastereoselectivity, being sug-
gested variously to enhance the enamine formation step or the
product iminium hydrolysis.
We report here detailed kinetic and structural studies revealing
several novel mechanistic features of the conjugate addition
reaction catalyzed by diarylprolinol ethers shown in Scheme 1.
Clarification of the rate-determining step, the catalyst resting
state, and the role of acid in the cycle allows us to develop one of
the most efficient conjugate addition reactions reported to date.
The beneficial role of an apparently “parasitic” catalytic inter-
mediate in maintaining high stereoselectivity is revealed.
Our initial investigations of the reaction in Scheme 1 monitored
by reaction calorimetry7 revealed the unusual kinetic behavior
shown in Figure 1 (presented as reaction heat flow vs time, where
heat flow is directly proportional to reaction rate). The reaction
commences with a rapid rate spike (Figure 1, left inset), correspond-
ing closely to one turnover of the catalytic cycle.7 This initial rate
Received: March 2, 2011
Published: May 15, 2011
r
2011 American Chemical Society
8822
dx.doi.org/10.1021/ja203660r J. Am. Chem. Soc. 2011, 133, 8822–8825
|

Journal of the American Chemical Society
COMMUNICATION
Figure 3. Effect of added acid concentration on reaction rate for the
reaction in Scheme 1 with [1] = 1.2 M, [2] = 1.0 M, and [4] = 0.1 M.
Rate measured from reaction calorimetric profiles at 20% conversion.
Figure 1. Temporal reaction rate in the reaction in Scheme 1 with initial
concentrations [1]0 = 1.7 M, [2]0 = 1.5 M, and [4] = 0.1 M monitored by
reaction calorimetry, showing an initial spike in rate (left inset) and the
subsequent accelerating rate profile (right inset).
Figure 4. Syn/anti selectivity in the Michael reaction in Scheme 1 under
the conditions given in Figure 2 and acetic acid concentrations as shown,
as a function of (a) fraction conversion during reaction and (b) time after
reaction completion.
Figure 2. Temporal concentration of product 3 in the reaction in
Scheme 1 monitored by 1H NMR spectroscopy. Red circles: [1]0 = 1.2 M,
[2]0 = 1.0 M. Blue circles: [1]0 = 1.7 M, [2]0 = 1.0 M. Green circles:
[1]0 = 1.7 M, [2]0 = 1.5 M. Turquoise circles: same as red circles, with
0.5 M 3 added. Pink circles: same as red circles, with 0.1 M CH3COOH
added. Catalyst loading is [4] = 0.1 M in all cases.
Scheme 2. 1H NOESY Identification of Species 5
enamine formation as the rate-determining step but could be
consistent with rate-determining hydrolysis of the product
iminium species. However, experiments adding water to the
reaction show that rate is neither accelerated nor substantially
slowed in the presence of water, meaning that hydrolysis to form
the product is unlikely to be rate-determining.7
We turned to further detailed NMR studies to probe the
nature of the resting state of the catalyst. Carrying out the reaction
with an excess of catalyst and in the presence of molecular sieves
allows us to capture the catalytic species without interference
from the cycle turnover. We identified the major species present
as the cyclobutane 5 shown in Scheme 2 via 2D NMR experi-
ments (HSQC, HMBC, COSY, and NOESY).7
Attempts to isolate 5 were unsuccessful, in agreement with
previous work, where isolation of similar cyclobutanes was possible
only for those with multiple bulky substituents.10 However,
in situ monitoring of reactions by NMR spectroscopy allowed
us to follow the temporal evolution of 5 as well as the free catalyst
4 and a species identified as the product enamine 6. A minor
amount (<5%) of a deactivated catalyst species 7, formed from
conjugate addition of 4 to the nitro-olefin, is also observed.7
Figure 5 shows these temporal concentration profiles for the
reaction in the absence of added acid. Dominance of the cyclobutane
intermediate 5 persists until full consumption of the limiting
reactant, at which time the resting state shifts to the product
enamine 6. Reaction in the presence of acid inverts the ratio of
free catalyst 4 to product enamine 6 at the end of the reaction
from 1:10 to 10:1.7 Under acidic conditions 4 is present in
protonated form after reaction completion.
With these findings we are able to propose the reaction mech-
anism shown in Scheme 3. The rapid initial buildup of intermediate 5
as the resting state was observed by in situ reaction calorimetric
monitoring, shown in Figure 1, reaching the maximum concentration
prior to acquisition of the first NMR spectrum in Figure 5. Cyclo-
butane 5 may exist as an off-cycle reservoir that sequesters the catalyst
under reaction turnover. Alternatively, a route directly from cyclobu-
tane 5 that bypasses 9 might be envisioned.11 The lower stability of
product enamine 6 compared to 5 prevents it from accumulating
until after the limiting substrate has been fully consumed and 5
begins to decay, as shown in Figure 5. Added acid enhances the
reaction rate by promoting the irreversible protonation step, thus
8823
dx.doi.org/10.1021/ja203660r |J. Am. Chem. Soc. 2011, 133, 8822–8825

Journal of the American Chemical Society
COMMUNICATION
Scheme 4. Crossover between Cyclobutane Species: (a)
Upon Addition of Nitro-olefin 2 to Cyclobutane 12 To Give 5
and 11 and (b) Upon Addition of Aldehyde 1 during Reaction
of 14 and 2, Showing a Shift from 13 to 57
Figure 5. Temporal evolution of catalytic species in the reaction in
Scheme 1, monitored by 1H NMR spectroscopy. Data are taken from the
reaction with no added acid, shown in red circles in Figure 2, with [1]0 =
1.2 M, [2]0 = 1.0 M, and [4]0 = 0.1 M. 4 is the free catalyst (open green
diamonds), 5 is the cyclobutane intermediate (open magenta circles),
and 6 is the product enamine (open blue squares).
show that cyclobutane 5 converts with ease to the product enamine
6, even in the absence of added acid, after the limiting substrate is
fully reacted. The near closure of a mass balance on catalyst
intermediates observed in Figure 5 indicates that irreversible
deactivation of the catalyst is minimal in this system.
Scheme 3. Proposed Catalytic Cycle
Confirmation that intermediate 5 is not an irreversible sink for
the catalyst but remains reversibly linked to the catalytic cycle is
given by several different crossover experiments. NMR studies
show that the cyclobutane species 12 formed from p-OMe trans-
nitrostyrene 11 equilibrates with cyclobutane 5 in the presence of
2 (Scheme 4), even as the total concentration of cyclobutane
species remains constant.7 Similarly, upon addition of linear
aldehyde 1 to the reaction mixture of disubstituted aldehyde 14
and 2, a shift from cyclobutane 13 to 5 is observed. Reaction
turnover is significantly slower for R,R-disubstituted aldehydes,
and hence the appearance of 5 can only be accounted for by
reverse reaction of 13. These results demonstrate not only that
cyclobutane formation from catalyst 4 is reversible but also that
both enamine formation and addition of enamines to nitro-
olefins are reversible steps for systems with catalyst 4.
These results highlight an important consequence of species 5
acting as the reversibly formed catalyst resting state: all steps
within the cycle preceding formation of 5 are equilibrated to
accommodate the hold-up in the cycle. Thus, the first irreversible
step within the cycle is the protonation of 9 and is, by definition,
rate-determining for a cycle under steady-state conditions.15 This
conclusion might be unexpected in light of what is known about
other enamine-based reaction mechanisms: for example, in
reactions of 1 with electrophiles where no reversible reservoir
subsequent to the electrophile addition exists, enamine forma-
tion is rate-determining for highly reactive electrophiles such as
azodicarboxylates,8 and both enamine formation and the sub-
sequent electrophile addition are irreversible.
Intermediates that may exist outside—but linked reversibly
to—a catalytic cycle have been characterized as “parasitic”,1a but in
the case of this Michael addition, the reservoir species 5 plays a critical
role beneficial to stereoselection. The dominance of 5 suppresses
erosion of diastereoselectivity during reaction turnover by preventing
an accumulation of iminium 10 and therefore of product enamine 6
during reaction. Erosion of diastereoselectivity is a consequence of
conditions where the catalyst is not fully sequestered as species 5, e.g.,
after reaction turnover is complete, as shown in Figure 5. The
unexplained variability in diastereoselectivities that may be noted in
drawing species 5 out of the resting state. In the absence of added
acid, the initially slow turnover increases as the temporally increasing
concentration of the acidic nitro reaction product promotes its
formation, leading to the temporally increasing rate shown in
Figures 1 and 2. We found that addition of simple nitroalkanes also
accelerates the rate, while added NEt3 decreases the rate.7
Cyclic species similar to 5 and cyclic nitronates have been
reported in stoichiometric reactions with enamines.10,12,13 See-
bach and co-workers carried out extensive studies of stoichio-
metric additions of prolinol methyl ether-derived enamines to
nitro-olefins and alkylidene malonates, including detailed stere-
ochemical investigations.2 Michael addition of the enamine to
the nitroalkene results in a zwitterionic species that may undergo
either O-alkylation or, as in the present case with nitrostyrene,
C-alkylation to form the cyclobutane 5.10b
To our knowledge, the only published proposal of a cyclobutane
species such as 5 in a catalytic conjugate addition system was made
by Jacobsen and co-workers in reactions of R,R-disubstituted
aldehydes catalyzed by primary amine thiourea catalysts, where a
putative cyclobutane species was suggested to form irreversibly off-
cycle and was implicated in catalyst deactivation.14 In the present
case for catalyst 4 and linear aldehydes, however, the data in Figure 5
8824
dx.doi.org/10.1021/ja203660r |J. Am. Chem. Soc. 2011, 133, 8822–8825

Journal of the American Chemical Society
COMMUNICATION
literature reports7 for this reaction could in fact be due to a lack of
precise control of reaction end point time for sampling, and lower
reported values may reflect post-reaction isomerization rather than
intrinsic catalyst selectivity.
’ AUTHOR INFORMATION
Corresponding Author
blackmond@scripps.edu
The dominance and reversibility of species 5 also helps to
rationalize the success that the reaction in Scheme 1 using
catalyst 4 has enjoyed in applications involving cascade reactions.
This reaction was first employed by Enders in an elegant triple
organocatalytic (enamine-iminium-enamine) cascade sequence,16
and it has been featured—always as the first cycle—in many
subsequent reports of cascade sequences. Our work suggests that
the dominance of 5 allows the cycle to proceed without significant
competition for the catalyst from the subsequent reactions. For
example, we were unable to detect any interaction between 4 and
cinnamaldehyde, the substrate for the second reaction in En-
ders’s cascade, in the presence of 5. As turnover in this first cycle of
the cascade nears completion and 5 decays, the potential for erosion
of product diastereoselectivity is avoided because the product and
catalyst are irreversibly drawn into subsequent cycles. Because this
catalytic cycle monopolizes the catalyst fully, however, such con-
jugate additions must be placed at the beginning of the cascade, as
indeed is the case in all reported examples. Our work provides the
mechanistic rationale for this empirically derived protocol.
’ ACKNOWLEDGMENT
A.A. and D.G.B. acknowledge research funding from the
EPSRC. J.B. acknowledges a postdoctoral fellowship from the
Education Ministry of Spain (EX2009-0687). The authors thank
D. Seebach and Y. Hayashi for stimulating discussions and for
sharing their unpublished work.
’ REFERENCES
(1) (a) Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Chem. Rev.
2007, 107, 5471–5569. (b) Melchiorre, P.; Marigo, M.; Carlone, A.;
Bartoli, G. Angew. Chem. Int. Ed. 2008, 47, 6138–6171. (c) Bertelsen, S.;
Jørgensen, K. A. Chem. Soc. Rev. 2009, 38, 2178. (d) Palomo, C; Mielgo,
A. Angew. Chem. Int. Ed. 2006, 45, 7876–7880. (e) Sulzer-Mosse, S.;
Alexakis, A. Chem. Commun. 2007, 3123–3135.
(2) (a) Seebach, D.; Golinski, J. Helv. Chim. Acta 1981, 64, 1413–1423.
(b) Blarer, S. J.; Schweizer, W. B.; Seebach, D. Helv. Chim. Acta 1982,
65, 1637–1654.
(3) Hayashi, Y.; Gotoh, H.; Hayashi, T.; Shoji, M. Angew. Chem. Int.
Ed. 2005, 44, 4212–4215.
A critical concern for practical application of organocatalytic reac-
tions is productivity. The development of reaction protocols that
employ high reagent concentrations (low solvent volumes) and low
catalyst loadings is a key challenge. In many enamine-based catalyst/
reaction combinations that exhibit positive order substrate kinetics,
however, highly concentrated reaction mixtures result in higher rates
and a concomitant increase in evolved heat in the early stages of the
reaction, rendering such desired process intensification unfeasible.
In the present case, we reasoned that the observed zero-order kinetic
profiles make operation under highly concentrated conditions
practical. Indeed, we demonstrated that the reaction proceeds
smoothly in the absence of solvent by dissolving solid 2 into neat
1 (1.4 equiv) with 2.5 mol % catalyst 4 and a 1:1 ratio of 4 and acetic
acid, with quantitative yield achieved in 10 min.7 Lower catalyst
loadings have been demonstrated previously in this Michael reaction
using 4 as well as other catalysts; for example, Wennemers6b
obtained high yields using as low as 0.1 mol % of a tripeptide
catalyst, but typical reaction conditions employed in that work
required 48 h reaction time and used ca. 10 volumes of solvent for
each volume of product formed. Our protocol provides an un-
precedented example of practical and scaleable process intensifica-
tion in organocatalysis, derived from mechanistic understanding.
In summary, detailed kinetic and structural analysis allows a
comprehensive picture of the mechanism of Michael additions of
aldehydes to nitroolefins catalyzed by diarylprolinol ethers, identify-
ing the rate-determining step and the role of acid additives in the
cycle. This work reveals the role of a key intermediate in the
maintenance of high stereoselectivity. The concept that a see-
mingly “parasitic” catalytic species could have a beneficial influence
on selectivity is novel, and its generality is currently under in-
vestigation in our laboratories. This mechanistic understanding
allowed design of highly efficient reaction protocols and may
inspire further catalyst and reaction design.
(4) Contemporaneous with ref 3, Jorgensen reported the use of diaryl-
prolinol ether catalysts in the conjugate addition of methyl vinyl ketones to
aldehydes: Franzen, J.; Marigo, M.; Fielenbach, D.; Wabnitz, T. C.; Kjaers-
gaard, A.; Jorgensen, K. A. J. Am. Chem. Soc. 2005, 127, 18296–18304.
(5) Diamine catalysts: (a) Betancort, J. M.; Barbas, C. F., III Org. Lett.
2001, 3, 3737–3740. (b) Alexakis, A.; Andrey, O. Org. Lett. 2002,
4, 3611–3614.
(6) Peptide-based catalysts: (a) Guo, L.; Chi, Y.; Almeida, A. M.;
Guzei, I.; Parker, B. K.; Gellman, S. H. J. Am. Chem. Soc. 2009, 131,
16018–16020. (b) Wiesner, M.; Upert, G.; Angelici, G.; Wennemers, H.
J. Am. Chem. Soc. 2010, 132, 6–7.
(7) See Supporting Information for details.
(8) (a) Mathew, S. P.; Iwamura, H.; Blackmond, D. G. Angew. Chem. Int.
Ed. 2004, 43, 3317. (b) Iwamura, H.; Mathew, S. P.; Blackmond, D. G. J. Am.
Chem. Soc. 2004, 126, 11770–11771. (c) Zotova, N.; Moran, A.; Armstrong,
A.; Blackmond, D. G. Adv. Synth. Catal. 2009, 351, 2765–2769. (d) Mathew,
S. J.; Klussmann, M.; Iwamura, H.; Wells, D. H., Jr.; Armstrong, A.;
Blackmond, D. G. Chem. Commun. 2006, 4291–4294.
(9) During the preparation of this manuscript, we became aware of a
contemporaneous study of the conjugate addition of nitro-olefins to
aldehydes using diarylprolinol ether catalysts, also identifying the
cyclobutane species 5 and exploring the role of acid in the reaction:
Seebach, D.; Hayashi, Y., personal communication.
(10) (a) Risaliti, A.; Forchiassin, M.; Valentin, E. Tetrahedron 1968,
24, 1889. (b) Felluga, F.; Nitti, P.; Pitacco, G.; Valentin, E. Tetrahedron
1989, 45, 5667.
(11) If this route is competitive with protonation of 9, this places 5 on
the cycle and provides an explanation for the much lower reactivity of R,R-
disubstituted aldehydes, for which such a route is unavailable. Experimental
data presently available cannot distinguish between these possibilities.
(12) (a) Brannock, K. C.;Bell, A.;Burpitt,R. D.;Kelly, C. A.J. Org. Chem.
1964, 29, 801. (b) Kuehne, M. E.; Foley, L J. Org. Chem. 1965, 30, 4280. (c)
Risaliti, A.; Forchiassin, M.; Valentin, E. Tetrahedron Lett. 1966, 7, 6331. (d)
Nielsen, A. T.; Archibald, T. G. Tetrahedron 1970, 26, 3475.
(13) Denmark, S. E.; Thorarensen, A. Chem. Rev. 1996, 96, 137.
(14) Lalonde, M. P.; Chen, Y. C.; Jacobsen, E. N. Angew. Chem. Int.
Ed. 2006, 45, 6366.
’ ASSOCIATED CONTENT
(15) Boudart, M. Kinetics of Catalytic Processes; Prentice-Hall:
Englewood Cliffs, NJ, 1968.
(16) Enders, D.; H€uttl, M. R. M.; Raabe, G.; Grondal, C. Nature
S
Supporting Information. Experimental details, kinetic
b
studies, and structural identification of catalytic species. This material
2006, 441, 861–863.
is available free of charge via the Internet at http://pubs.acs.org.
8825
dx.doi.org/10.1021/ja203660r |J. Am. Chem. Soc. 2011, 133, 8822–8825