Plagiarizing proteins: enhancing efficiency in asymmetric hydrogenbonding catalysis through positive cooperativity

Full text of DOI:10.1002/anie.200903063

Angewandte
Chemie
DOI: 10.1002/anie.200903063
Cooperative Catalysis
Plagiarizing Proteins: Enhancing Efficiency in Asymmetric Hydrogen-
Bonding Catalysis through Positive Cooperativity**
Christopher R. Jones, G. Dan Pantos¸, Angus J. Morrison, and Martin D. Smith*
Nature achieves nearly catalytic perfection in enzymes
through the juxtaposition of specific functional groups in
space.[1] Although the effects that contribute to the remark-
able rate acceleration observed in enzyme-catalyzed reactions
are manifold, it has been proposed that mutually reinforcing,
or cooperative noncovalent interactions[2] within a receptor
can play a significant role.[3] This tenet has been postulated to
be a factor in the extraordinary avidity of binding in systems
such as streptavidin–biotin, and has been elegantly applied in
the generation of synthetic receptors with enhanced anion-
binding properties.[4] In a catalyst system with positively
cooperative binding, the enthalpic contribution to the tran-
sition-state binding energy may be increased through a
reduction in dynamic behavior, resulting in stronger non-
covalent interactions within the catalyst structure. This can
outweigh the entropic cost of the associated reduction in
motion, and hence lead to tighter transition-state binding, and
a subsequent increase in catalytic efficiency. We considered
whether this effect—which is implicated in the remarkable
efficiency exhibited by some enzymes—could be exploited in
the development of more efficient asymmetric catalysts that
operate by hydrogen bonding.[5] This could lead to the
development of new transformations through the discovery
of more active catalysts with lower loading, shorter reaction
times, and wider substrate scope. In designing these new
catalysts, we reasoned that by analogy with ligand–protein
binding, a series of noncovalent interactions within the
catalyst structure could: 1) direct folding toward population
of an ensemble of structured conformations, preorganizing
the catalyst and minimizing the entropic cost of transition-
state (TS) binding,[3,6] and 2) result in stronger intramolecular
noncovalent interactions and cooperative ligand binding and
hence greater stabilization of charged intermediates and
transition states.[7–9]
Herein we generate conformationally well-defined but
flexible thiourea catalysts[10,11] that benefit from cooperative
ligand binding, and we demonstrate the utility of this
phenomenon in catalytic asymmetric synthesis. We have
constructed a simple and effective turn mimetic that pop-
ulates a well-defined hairpin conformation in solution and in
the solid state stabilized by intramolecular hydrogen bonds.[12]
A major aim of the design and generation of well-defined
unnatural folded materials[13] is the evolution of function
similar to that observed in natural biopolymers. It has been
demonstrated through a series of elegant studies that folded
peptidic materials can be highly effective asymmetric cata-
lysts for a range of synthetic transformations.[14] Our catalyst
scaffold, which can be easily and efficiently prepared, is
depicted in Figure 1.[15]
Figure 1. Rationale behind the catalyst design.
The conformation of materials incorporating this con-
struct have been extensively probed in solution and in the
solid state.[9] NMR studies of model compound 1 (in an
aprotic solvent) confirmed the presence of a series of
intramolecular hydrogen bonds consonant with a turn con-
formation and indicated that the trans,trans thiourea con-
formation is populated;[16,17] this is consistent with the X-ray
structure of 1 (Figure 1).[18,19] In general, for these scaffolds to
populate turn conformations where R3 is relatively large, we
have found that a small R2 group (in the case of 1, an ethyl
group) is optimal. The turn construct is insensitive to the size
of the R1 group, but the absolute configuration at this center
can have important implications when other chiral substitu-
ents are appended to this scaffold.
[*] Dr. M. D. Smith
Chemistry Research Laboratory, University of Oxford
12 Mansfield Road, Oxford OX1 3TA (UK)
E-mail: martin.smith@chem.ox.ac.uk
Homepage: http://msmith.chem.ox.ac.uk
C. R. Jones, Dr. G. Dan Panto¸s
Department of Chemistry, University of Cambridge
Lensfield Road, Cambridge CB2 1EW (UK)
Dr. A. J. Morrison
Schering Plough Corporation
With a reliable turn-forming design in hand, we turned
our attention to investigating the potential utility of the
cooperativity concept in asymmetric catalysis. Our initial
investigations focused on the ability of these materials to
mediate reactions involving N-tert-butoxycarbonyl (N-Boc)
aldimines as substrates, and hence we first examined a model
Mannich reaction with a silyl ketene acetal nucleophile.[20,21]
Newhouse, Lanarkshire, Scotland ML1 5SH (UK)
[**] We acknowledge funding from the Royal Society (M.D.S.), Schering-
Plough/EPSRC (C.R.J.), Pembroke College, Cambridge University
(G.D.P.), and AstraZeneca. We thank Dr. J. Davies for help with X-ray
crystallography.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200903063.
Angew. Chem. Int. Ed. 2009, 48, 7391 –7394
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7391

Communications
We reasoned that exploring a known reaction manifold would
groups or electron-withdrawing groups on the aldimine at a
loading of 5 mol%, generating products with universally high
enantioselectivity (97 to > 99% ee) and yield (Table 2).
enable us to benchmark the effectiveness of our folded
catalysts against materials known to be proficient in this
reaction and enable evaluation of our catalyst design blue-
print. A preliminary screen of thiourea catalysts indicated
that materials bearing a pyrrolyl-1,2-diamine substituent[22]
were most effective, and hence we generated the small library
of compounds 2–6 based on this template (Figure 2). Catalysts
containing two urea functional groups were also prepared and
evaluated, but these were found to be significantly less
effective.[23]
Table 2: Mannich reactions at different catalyst loadings.[a]
0.1 mol%
Loading[b]
Yield
1 mol%
Loading[c]
Yield
5 mol%
Loading[c]
Yield
R1
ee
ee
ee
[%][d]
[%][e]
[%][d]
[%][e]
[%][d]
[%][e]
H
4-Me
1-naphthyl
4-OMe
3-NO2
75
63
81
40
75
99
81
93
84
93
96
79
86
85
74
>99
95
96
94
99
97
97
84
83
73
>99
97
98
98
>99
[a] Reaction conditions: 0.5 mmol imine, 1 mmol ketene acetal, 0.4 mL
toluene. [b] Reaction time: 96 h. [c] Reaction time: 48 h. [d] Yield of
isolated product. [e] Determined by HPLC on a chiral stationary phase.
Figure 2. Catalysts examined in this study.
We also generated a series of control catalysts 7 and 8 that
do not bear the intramolecular hydrogen-bond-donor group
to investigate the effect of this deletion on asymmetric
induction and catalytic activity. The efficacy of materials 2–8
in mediating the asymmetric Mukaiyama–Mannich reaction
at a loading of 5 mol% is outlined in Table 1. Our inves-
tigations demonstrate that the size of the substituents on the
pyrrole is paramount in determining the effectiveness of the
asymmetric induction, with 4 (R1 = Me, R2 = Ph) affording R
product in 97% yield with greater than 99% ee.[24,25]
When the loading of 4 was reduced to 1 mol%, the
excellent levels of enantioselectivity were maintained (94 to
> 99% ee) and the yield of isolated product decreased only
slightly. Remarkably, the catalyst loading of 4 can be reduced
to 0.1 mol% whilst maintaining high enantioselectivity (for
R1 = H, 3-NO2, 1-naphthyl); less reactive imines (R1 = 4-Me,
4-OMe) give lower ee values at this loading.[26,27] These
loadings are significantly lower than the 5–10 mol% generally
required in thiourea-catalyzed transformations, and confirm
that optimum catalyst 4 is both catalytically proficient and
highly enantioselective. The materials 7 and 8, in which the
intramolecular urea hydrogen-bond donor is absent (Table 1),
produced 10 in 72% yield (95% ee) and 71% yield (91% ee),
respectively, at a loading of 5 mol%. This indicates that the
effect of conformational ordering on asymmetric induction
(through the internal bifurcated hydrogen bond) is significant
but relatively small.
Table 1: Catalyst screening with a Mukaiyama–Mannich reaction.[a]
To probe catalyst efficiency rather than asymmetric
induction we need to investigate kinetic parameters—and
hence we instigated a series of catalyst competition experi-
Cat.
R1
R2
Yield [%][b]
ee [%][c]
2
3
4
5
6
7
8
Ph
Ph
Me
Ph
1-naphthyl
2-naphthyl
Ph
70
97
97
81
77
72
71
9
95
>99
>99
91
Me
Me
Me
Me
Me
Me
ments (Scheme 1).
A competition experiment between
1 mol% 4 (known to generate S-configured material) and
1 mol% ent-7 (known to generate R-configured material)
afforded material 10 with 93% ee (S) and 97% yield,
indicating that catalyst 4 significantly outcompetes catalyst
ent-7 for substrate in this specific reaction.[28, 29] We attribute
this dramatic rate enhancement to tighter ligand–receptor
binding in the transition state as a result of strengthened
noncovalent interactions within the catalyst structure in 4 that
are absent in 7. To support this premise, we compared the
anion-binding ability[30] of 3 and 8 to establish that the
noncovalent interactions within the catalyst structure are
mutually reinforcing.[31] According to the cooperative model,
intrareceptor interactions could make a significant contribu-
tion to complex stability. These studies were carried out in
95
91
Me
[a] Reaction conditions: 0.25 mmol 9, 0.5 mmol ketene acetal, 0.2 mL
toluene. [b] Yield of isolated product. [c] Measured by HPLC on a chiral
stationary phase. TBS=tert-butyldimethylsilyl.
With an optimal catalyst for this reaction, we investigated
the generality of this method by preparing a series of
substituted b-aryl b-amino acids using catalyst 4 at different
loadings. This demonstrated that catalyst 4 is uniformly
effective in mediating reactions with either electron-donating
7392
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7391 –7394

Angewandte
Chemie
binding energy derived from them acting together is greater than
the sum of them acting separately”.
[3] a) D. H. Williams, E. Stephens, D. P. OꢀBrien, M. Zhou, Angew.
Chem. 2004, 116, 6760; Angew. Chem. Int. Ed. 2004, 43, 6596;
b) D. H. Williams, E. Stephens, M. Zhou, J. Mol. Biol. 2003, 329,
389.
[4] a) Z. Rodriguez-Docampo, S. I. Pascu, S. Kubik, S. Otto, J. Am.
Chem. Soc. 2006, 128, 11206; b) S. Rashdan, M. E. Light, J. D.
Kilburn, Chem. Commun. 2006, 4578.
[5] M. T. Robak, M. Trincado, J. A. Ellman, J. Am. Chem. Soc. 2007,
129, 15110.
[6] Stabilizing specific ground-state conformations is not necessarily
similar to natural protein–ligand binding; sufficient conforma-
tional flexibility to permit effective transition-state binding is
required.
[7] H. Guo, D. R. Salahub, Angew. Chem. 1998, 110, 3155; Angew.
Chem. Int. Ed. 1998, 37, 2985.
[8] For Brønsted acid catalyzed reactions that may be mediated by a
cooperative hydrogen-bonding effect see: a) Y. Huang, A. K.
Unni, A. N. Thadani, V. H. Rawal, Nature 2003, 424, 146; b) A.
Wittkopp, P. R. Schreiner, Chem. Eur. J. 2003, 9, 407; c) M.
Ganesh, D. Seidel, J. Am. Chem. Soc. 2008, 130, 16464.
[9] For an example of a system where a preorganized dimeric
receptor leads to more efficient ligand binding see: D. H.
Williams, A. J. Maguire, W. Tsuzuki, M. S. Westwell, Science
1998, 280, 711.
[10] For seminal work on asymmetric thiourea-catalyzed reactions
see: a) M. S. Sigman, E. N. Jacobsen, J. Am. Chem. Soc. 1998,
120, 4901; b) M. S. Sigman, P. Vachal, E. N. Jacobsen, Angew.
Chem. 2000, 112, 1336; Angew. Chem. Int. Ed. 2000, 39, 1279;
c) P. Vachal, E. N. Jacobsen, J. Am. Chem. Soc. 2002, 124, 10012.
[11] For recent reviews see: a) A. Dondoni, A. Massi, Angew. Chem.
2008, 120, 4716; Angew. Chem. Int. Ed. 2008, 47, 4638; b) A. G.
Doyle, E. N. Jacobsen, Chem. Rev. 2007, 107, 5713; c) M. J.
Gaunt, C. C. C. Johansson, A. McNally, N. T. Vo, Drug Discov-
ery Today 2007, 12, 8.
Scheme 1. Reaction conditions: 0.25 mmol 9, 0.5 mmol ketene acetal,
0.2 mL toluene.
CDCl3 at À158C using tetra-N-butylammonium chloride as
the anion source, and are consistent with ligand–receptor
stoichiometry of 1:1 for both 3 and 8. Conformational studies
on the anion-bound catalyst 3 indicate population of the
trans,trans thiourea conformer, similar to what is observed for
the ground-state conformation of 1. This may indicate that the
turn scaffold affects the conformational propensities of the
thiourea portion of the catalyst.[17] Catalyst 3 was found to
have Ka = (550 Æ 85)mÀ1 towards chloride anion, while model
compound 8, containing a single urea functionality, has a Ka
less than 10mÀ1. Hence 3 has a binding constant for chloride
anion approximately two orders of magnitude larger than that
for 8, entirely consistent with the tenet of cooperative ligand
binding outlined earlier, and reflected in the significantly
increased catalytic efficiency of this material.[32]
In this study we have outlined the synthesis of conforma-
tionally well-defined thiourea catalysts designed to mimic the
positively cooperative action of enzymes. We have demon-
strated that noncovalent interactions within a b-turn catalyst
structure lead to significantly higher levels of catalytic
efficiency in a model Mukaiyama–Mannich reaction than
noncooperative materials. This increased cooperativity-led
efficiency is reflected in turnover rates and catalyst loading,
and is coupled with extremely high enantioselectivities. We
believe that this nature-inspired cooperative design may have
utility in the development of highly efficient catalysts for
demanding asymmetric transformations.
[12] C. R. Jones, M. K. N. Qureshi, F. R. Truscott, S. T. D. Hsu, A. J.
Morrison, M. D. Smith, Angew. Chem. 2008, 120, 7207; Angew.
Chem. Int. Ed. 2008, 47, 7099.
[13] a) S. H. Gellman, Acc. Chem. Res. 1998, 31, 173; b) R. P. Cheng,
S. H. Gellman, W. F. DeGrado, Chem. Rev. 2001, 101, 3219; c) D.
Seebach, D. F. Hook, A. Glꢁttli, Biopolymers 2005, 84, 23;
d) D. J. Hill, M. J. Mio, R. B. Prince, T. S. Hughes, J. S. Moore,
Chem. Rev. 2001, 101, 3893; e) C. M. Goodman, S. Choi, S.
Shandler, W. F. DeGrado, Nat. Chem. Biol. 2007, 3, 252.
[14] a) G. T. Copeland, S. J. Miller, J. Am. Chem. Soc. 2001, 123, 6496;
b) J. T. Blank, S. J. Miller, Biopolymers 2006, 84, 38; c) D. R.
Kelly, S. M. Roberts, Biopolymers 2006, 84, 74; d) A. Berkessel,
B. Koch, C. Toniolo, M. Rainaldi, Q. B. Broxterman, B. Kaptei,
Biopolymers 2006, 84, 90.
[15] For other bis(thio)urea catalysts see: a) Y. Sohtome, A. Tana-
tani, Y. Hashimoto, K. Nagasawa, Tetrahedron Lett. 2004, 45,
5589; b) A. Berkessel, K. Roland, J. K. Neudꢂfl, Org. Lett. 2006,
8, 4195; c) E. M. Fleming, T. McCabe, S. J. Connon, Tetrahedron
Lett. 2006, 47, 7037; d) M. Shi, X.-G. Liu, Org. Lett. 2008, 10,
1043; e) C. Rampalakos, W. D. Wulff, Adv. Synth. Catal. 2008,
350, 1785. These catalysts do not benefit from the cooperative
hydrogen-bonding phenomenon described herein.
Received: June 7, 2009
Revised: July 15, 2009
Published online: September 2, 2009
Keywords: asymmetric catalysis · cooperativity ·
.
hydrogen bonds · thioureas
[16] DMSO titration data, IR data, and NOE data are all consistent
with population of a folded conformation for compound 1. See
the Supporting Information for full details.
[17] For insightful studies on the conformational and mechanistic
aspects of thiourea catalysis see: a) A. Hamza, G. Schubert, T.
Soꢃs, I. Pꢄpai, J. Am. Chem. Soc. 2006, 128, 13151; b) G.
Tꢄrkꢄnyi, P. Kirꢄly, S. Varga, B. Vakulya, T. Soꢃs, Chem. Eur. J.
2008, 14, 6078.
[1] a) W. P. Jencks, Catalysis in Chemistry and Enzymology, Dover,
New York, 1987; b) M. Garcia-Viloca, J. Gao, M. Karplus, D. G.
Truhlar, Science 2004, 303, 186.
[2] We define this term as outlined in reference [3a]: “Non-covalent
interactions act in a positively cooperative manner when the
Angew. Chem. Int. Ed. 2009, 48, 7391 –7394
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7393

Communications
[18] The X-ray structure of compound 1 indicates a folded con-
[23] Urea analogues of catalysts 2, 3, and 4 afforded 10 in 27% ee
(40% yield), 81% ee (45% yield), and 72% ee (45% yield),
respectively.
[24] The absolute configuration of the products was confirmed
through comparison with literature data.
formation where both inter- and intramolecular hydrogen bonds
can be observed. This must obviously be considered in terms of
both short- and longer-range effects: J. D. Dunitz, A. Gavezzotti,
Helv. Chim. Acta 2002, 85, 3949.
[25] These reactions were conducted at À408C to ensure the best
enantioselectivity, but catalyst 4 maintains remarkable selectiv-
ity and activity at noncryogenic temperatures: À208C:
> 99% ee, 96% yield; 08C: 99% ee, 97% yield; room temper-
ature: 97% ee, 96% yield.
[26] A loading of 0.01 mol% produced 10 in 30% yield and 61% ee.
[27] For a highly efficient asymmetric hydrogen-bonding catalyst
based upon a squarate scaffold see: J. P. Malerich, K. Hagihara,
V. H. Rawal, J. Am. Chem. Soc. 2008, 130, 14416.
[28] NMR experiments conducted at the (low) concentration used in
the competition experiments do not indicate any significant
interaction between the two catalysts.
[29] A competition experiment between 4 and the benchmark
catalyst, described in reference [20] (both at 1 mol% loading),
afforded 10 in 97% yield and 94% ee (S), indicating that
cooperative catalyst 4 outcompetes the benchmark for substrate
in this specific reaction. See the Supporting Information for full
details.
[30] For an insightful anion-binding model for thiourea catalysis see:
I. T. Raheem, P. S. Thiara, E. A. Peterson, E. N. Jacobsen, J. Am.
Chem. Soc. 2007, 129, 13404; S. E. Reisman, A. G. Doyle, E. N.
Jacobsen, J. Am. Chem. Soc. 2008, 130, 7198.
[31] We conducted these studies on 2,5-dimethylpyrrole materials 3
and 8, rather than 2-methyl-5-phenylpyrrole materials 4 and 7 to
facilitate analysis in the urea proton region of the 1H NMR
spectrum. These experiments are described in detail in the
Supporting Information.
[32] Chloride binding has no direct mechanistic implication in the
Mukaiyama–Mannich reaction (unlike the acyl-Mannich reac-
tion described in reference [22]). Attempts to cocrystallize 3 or 4
with chloride anion were unsuccessful; see the Supporting
Information for full details.
[19] CCDC 734184 (1) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
[20] a) A. G. Wenzel, E. N. Jacobsen, J. Am. Chem. Soc. 2002, 124,
12964; b) A. G. Wenzel, M. P. Lalonde, E. N. Jacobsen, Synlett
2003, 1919.
[21] For Mannich reactions catalyzed by thiourea derivatives see
reference [20] and: a) T. Okino, S. Nakamura, T. Furukawa, Y.
Takemoto, Org. Lett. 2004, 6, 625; b) T. P. Yoon, E. N. Jacobsen,
Angew. Chem. 2005, 117, 470; Angew. Chem. Int. Ed. 2005, 44,
466; c) A. L. Tillman, J. Ye, D. J. Dixon, Chem. Commun. 2006,
1191; d) J. Song, Y. Wang, L. Deng, J. Am. Chem. Soc. 2006, 128,
6048; e) C. Bode, M. A. Ting, S. E. Schaus, Tetrahedron 2006, 62,
11499; f) Y. Yamaoka, H. Miyabe, Y. Yasui, T. Takemoto,
Synthesis 2007, 2571; g) J. Song, H. W. Shih, L. Deng, Org. Lett.
2007, 9, 603; h) T. Y. Liu, H. L. Cui, J. Long, B. J. Li, Y. Wu, L. S.
Ding, Y. C. Chen, J. Am. Chem. Soc. 2007, 129, 1878; i) D. A.
Yalalov, S. B. Tsogoeva, T. E. Shibuna, I. M. Martynova, T.
Clark, Angew. Chem. 2008, 120, 6726; Angew. Chem. Int. Ed.
2008, 47, 6624; j) C. Wang, Z. Zhou, C. Tang, Org. Lett. 2008, 10,
1707; k) C. J. Wang, X. Q. Dong, Z. H. Zhang, Z. Y. Xue, H. L.
Teng, J. Am. Chem. Soc. 2008, 130, 8606; l) C. Rampalakos,
W. D. Wulff, Adv. Synth. Catal. 2008, 350, 1785; m) X. Tian, K.
Jiang, J. Peng, W. Du, Y. C. Chen, Org. Lett. 2008, 10, 3583.
[22] Similar substitution patterns are effective for the acyl-Mannich
reaction: a) M. S. Taylor, E. N. Jacobsen, J. Am. Chem. Soc.
2004, 126, 10558; b) M. S. Taylor, N. Tokunaga, E. N. Jacobsen,
Angew. Chem. 2005, 117, 6858; Angew. Chem. Int. Ed. 2005, 44,
6700.
7394
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7391 –7394