Cooperative catalysis by tertiary amino-thioureas: Mechanism and basis for enantioselectivity of ketone cyanosilylation

Article abstract of DOI:10.1021/ja0735352

The mechanism of the enantioselective cyanosilylation of ketones catalyzed by tertiary amino-thiourea derivatives was investigated using a combination of experimental and theoretical methods. The kinetic analysis is consistent with a cooperative mechanism

Full text of DOI:10.1021/ja0735352

Published on Web 12/05/2007
Cooperative Catalysis by Tertiary Amino-Thioureas:
Mechanism and Basis for Enantioselectivity of Ketone
Cyanosilylation
Stephan J. Zuend and Eric N. Jacobsen*
Contribution from the Department of Chemistry & Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received May 17, 2007; E-mail: jacobsen@chemistry.harvard.edu
Abstract: The mechanism of the enantioselective cyanosilylation of ketones catalyzed by tertiary amino-
thiourea derivatives was investigated using a combination of experimental and theoretical methods. The
kinetic analysis is consistent with a cooperative mechanism in which both the thiourea and the tertiary
amine of the catalyst are involved productively in the rate-limiting cyanide addition step. Density functional
theory calculations were used to distinguish between mechanisms involving thiourea activation of ketone
or of cyanide in the enantioselectivity-determining step. The strong correlation obtained between experimental
and calculated ee’s for a range of substrates and catalysts provides support for the most favorable calculated
transition structures involving amine-bound HCN adding to thiourea-bound ketone. The calculations suggest
that enantioselectivity arises from direct interactions between the ketone substrate and the amino-acid
derived portion of the catalyst. On the basis of this insight, more enantioselective catalysts with broader
substrate scope were prepared and evaluated experimentally.
via enamine formation7 or Lewis acid-Lewis base dative bond-
Introduction
ing8,9) has been possible in several cases. In contrast, mechanistic
Cooperative catalysissthe simultaneous binding and activa-
tion of two reacting partnerssis a fundamental principle in
enzymatic catalysis,1-4 and has emerged as an important strategy
in small molecule asymmetric catalysis as well.5 Indeed, over
the past several years, a wide range of synthetically important
reactions promoted by chiral small molecules have been
identified and proposed to operate by cooperative catalysis.6
Mechanistic characterization of cooperative enantioselective
reactions that involve strong catalyst-substrate interactions (e.g.,
characterization of catalytic asymmetric reactions involving tran-
sient, noncovalent interactions, such as those employed com-
monly by enzymatic systems, has proved more challenging.10
Chiral thioureas represent a versatile and useful class of
enantioselective catalysts that function as hydrogen bond
donors,11 and several bifunctional thiourea derivatives have been
suggested to operate via cooperative mechanisms.12 Particularly
prominent among these are tertiary amino-thioureas, which have
been applied successfully to a wide range of nucleophile-
electrophile addition reactions.13,14 In each case, structure-
(1) General discussions: (a) Zeffern, E.; Hall, P. L. The Study of Enzyme
Mechanisms; Wiley: New York, 1973; Chapter 7. (b) Royer, G. P.
Fundamentals of Enzymology; Wiley: New York, 1982; Chapter 4. (c)
Dugas, H. Bioorganic Chemistry; Springer: New York, 1999; Chapter 4.
(d) Silverman, R. B. The Organic Chemistry of Enzyme-Catalyzed
Reactions; Academic Press: San Diego, CA, 2002; Chapter 1. (e) Gerlt, J.
A.; Gassman, P. G. Biochemistry 1993, 32, 11943-11952.
(2) Selected examples: (a) Breslow, R. J. Mol. Catal. 1994, 91, 161-174. (b)
Zechel, D. L.; Withers, S. G. Acc. Chem. Res. 2000, 33, 11-18. (c)
Hedstrom, L. Chem. ReV. 2002, 102, 4501-4523. (d) Hopmann, K. H.;
Hallberg, B. M.; Himo, F. J. Am. Chem. Soc. 2005, 127, 14339-14347.
(e) Hernick, M.; Gennadios, H. A.; Whittington, D. A.; Rusche, K. M.;
Christianson, D. W.; Fierke, C. A. J. Biol. Chem. 2005, 280, 16969-16978.
(f) Cleland, W. W.; Hengge, A. C. Chem. ReV. 2006, 106, 3252-3278.
See, also ref 1d, Chapter 2, 6, and 11.
(3) Early examples of cooperative mechanisms in small molecules: (a) Swain,
C. G.; Brown, J. F. J. Am. Chem. Soc. 1952, 74, 2534-2537. (b) Swain,
C. G.; Brown, J. F. J. Am. Chem. Soc. 1952, 74, 2538-2543. (c) Morawetz,
H.; Oreskes, J. J. Am. Chem. Soc. 1958, 80, 2591-2592. (d) Kupchan, S.
M.; Eriksen, S. P.; Friedman, M. J. Am. Chem. Soc. 1966, 88, 343-346.
(e) Kupchan, S. M.; Eriksen, S. P.; Shen, Y.-T. J. Am. Chem. Soc. 1966,
88, 347-350. (f) Higuchi, T.; Takechi, H.; Pitman, I. H.; Fung, H. L. J.
Am. Chem. Soc. 1971, 93, 539-540.
(6) (a) Jacobsen, E. N. Acc. Chem. Res. 2000, 33, 421-431. (b) Shibasaki,
M.; Yoshikawa, N. Chem. ReV. 2002, 102, 2187-2209. (c) Sammis, G.
M.; Danjo, H.; Jacobsen, E. N. J. Am. Chem. Soc. 2004, 126, 9928-9929.
(d) Kanai, M.; Kato, N.; Ichikawa, E.; Shibasaki, M. Synlett 2005, 10,
1491-1508. (e) Shibasaki, M.; Matsunaga, S. Chem. Soc. ReV. 2006, 35,
269-279.
(7) For example, as in proline-catalyzed aldol and alkylation reactions: (a)
Bahmanyar, S.; Houk, K. N. J. Am. Chem. Soc. 2001, 123, 12911-12912.
(b) Hoang, L.; Bahmanyar, S.; Houk, K. N.; List, B. J. Am. Chem. Soc.
2003, 125, 16-17. (c) Allemann, C.; Gordillo, R.; Clemente, F. R.; Cheong,
P. H.-Y.; Houk, K. N. Acc. Chem. Res. 2004, 37, 558-569. (d) Fu, A.;
List, B.; Thiel, W. J. Org. Chem. 2006, 71, 320-326. See, also: (e) Yalalov,
D. A.; Tsogoeva, S. B.; Schmatz, S. AdV. Synth. Catal. 2006, 348, 826-
832.
(8) For example, as in the CBS reduction: (a) Corey, E. J.; Azimioara, M.;
Sarshar, S. Tetrahedron Lett. 1992, 33, 3429-3430. (b) Quallich, G. J.;
Blacke, J. F.; Woodall, T. M. J. Am. Chem. Soc. 1994, 116, 8516-8525.
(c) Alagona, G.; Ghio, C.; Persico, M.; Tomasi, S. J. Am. Chem. Soc. 2003,
125, 10027-10039.
(9) For example, as in (salen)metal-catalyzed epoxide ring opening: (a) Hansen,
K. B.; Leighton, J. L.; Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118, 10924-
10925. (b) Nielsen, L. P. C.; Stevenson, C. P.; Blackmond, D. G.; Jacobsen,
E. N. J. Am. Chem. Soc. 2004, 118, 1360-1362.
(10) For examples, see: (a) Fearman, M. B.; O’Leary, D. J.; Steinmetz, W. E.;
Miller, S. J. J. Am. Chem. Soc. 2004, 126, 6967-6971. (b) Yamagiwa, N.;
Qin, H.; Matsunaga, S.; Shibasaki, M. J. Am. Chem. Soc. 2005, 127, 13419-
13427.
(4) Discussions of concerted general acid-base catalysis: (a) Jencks, W. P. J.
Am. Chem. Soc. 1972, 94, 4731-4732. (b) Jencks, W. P. Chem. ReV. 1972,
72, 705-718. (c) Gerlt, J. A.; Kozarich, J. W.; Kenyon, G. L.; Gassman,
P. G. J. Am. Chem. Soc. 1991, 113, 9667-9669. (d) Gerlt, J. A.; Gassman,
P. G. J. Am. Chem. Soc. 1992, 114, 5928-5934.
(5) For a review, see: Ma, J.-A.; Cahard, D. Angew. Chem., Int. Ed. 2004, 43,
4566-4583.
9
15872
J. AM. CHEM. SOC. 2007, 129, 15872-15883
10.1021/ja0735352 CCC: $37.00 © 2007 American Chemical Society

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
Scheme 1. Thiourea-Catalyzed Enantioselective Cyanosilylation
activity relationship studies reveal a clear requirement for the
tertiary amine, but the respective roles of the thiourea and amine
have not been elucidated. Although thiourea activation of the
electrophile by hydrogen bonding is often invoked based largely
on spectroscopic data of detectable intermediates,15,16 the well-
established anion-binding properties of thioureas17 raise the
possibility that nucleophile binding and activation may be
involved instead. Indeed, a detailed theoretical study has shown
that the nucleophile-activation mechanism is energetically
accessible in some cases, raising questions about the viability
of electrophile-activation mechanisms in related reactions.18
Without a basic understanding of the catalytic mechanism, only
speculation about transition structures and the basis for enan-
tioselectivity is possible.
recently discovered ketone cyanosilylation19,20 promoted by
tertiary amino-thiourea catalyst 1a13c as a prototypical reaction
for mechanistic analysis (Scheme 1).21,22 The resulting experi-
mental and theoretical investigation described in this paper has
led to the elucidation of a cooperative mechanism, characteriza-
tion of likely transition structures with insight into the basis
for enantioselectivity, and the development of more enantiose-
lective and broadly applicable catalysts.
To obtain better insight into the nature and mechanism of
cooperative catalysis with thiourea catalysts, we selected the
(11) For reviews, see: (a) Takemoto, Y. Org. Biomol. Chem. 2005, 3, 4299-
4306. (b) Taylor, M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45,
1520-1543. (c) Connon, S. J. Chem. Eur. J. 2006, 12, 5418-5427.
(12) Primary amino-thioureas: (a) Lalonde, M. P.; Chen, Y.; Jacobsen, E. N.
Angew. Chem., Int. Ed. 2006, 45, 6366-6370 and refs cited therein.
Secondary amido-thioureas: (b) Yoon, T. P.; Jacobsen, E. N. Angew. Chem.,
Int. Ed. 2005, 44, 466-468. Sulfinamido thioureas: (c) Tan, K. L.;
Jacobsen, E. N. Angew. Chem., Int. Ed. 2007, 46, 1315-1317. Hydroxyl
thioureas: (d) Herrera, R. P.; Sgarzani, V.; Bernardi, L.; Ricci, A. Angew.
Chem., Int. Ed. 2005, 44, 6576-6579. Amino-hydroxyl thioureas: (e)
Yamaoka, Y.; Miyabe, H.; Takemoto, Y. J. Am. Chem. Soc. 2007, 129,
6686-6687. Guanidino thioureas: (f) Sohtome, Y.; Hashimoto, Y.;
Nagasawa, K. AdV. Synth. Catal. 2005, 347, 1643-1648.
(13) Tertiary amino-thiourea catalysts derived from chiral 1,2-diamines: (a)
Okino, T.; Hoashi, Y.; Takemoto, Y. J. Am. Chem. Soc. 2003, 125, 12672-
12673. (b) Okino, T.; Hoashi, Y.; Furukawa, T.; Xu, X. N.; Takemoto, Y.
J. Am. Chem. Soc. 2005, 127, 119-125. (c) Fuerst, D. E.; Jacobsen, E. N.
J. Am. Chem. Soc. 2005, 127, 8964-8965. (d) Hoashi, Y.; Okino, T.;
Takemoto, Y. Angew. Chem., Int. Ed. 2005, 44, 4032-4035. (e) Berkessel,
A.; Cleemann, F.; Mukherjee, S. Angew. Chem., Int. Ed. 2005, 44, 7466-
7469. (f) Li, H.; Zu, L. S.; Wang, J.; Wang, W. Tetrahedron Lett. 2006,
47, 3145-3148. (g) Xu, X.; Furukawa, T.; Okino, T.; Miyabe, H.;
Takemoto, Y. Chem. Eur. J. 2006, 12, 466-467. (h) Zu, L. S.; Xie, H. X.;
Li, H.; Wang, H.; Jiang, W.; Wang, W. AdV. Synth. Catal. 2007, 349, 1882-
1886.
Results and Discussion
A. Kinetic Analysis. We carried out a detailed kinetic
analysis of the ketone cyanation reaction to elucidate the basic
composition of the rate-limiting transition structure and identify
catalyst resting states.23 Kinetic studies were performed using
representative substrate 6b and catalyst 1a at synthetically
relevant concentrations using a combination of GC analysis and
in situ IR spectroscopy.24 HCN was generated in situ from
TMSCN and CF3CH2OH,25 and quantitative formation of HCN
(18) (a) Hamza, A.; Schubert, G.; Soo´s, T.; Pa´pai, I. J. Am. Chem. Soc. 2006,
128, 13151-13160. (b) Rokob, T. A.; Hamza, H. A.; Pa´pai, I. Org. Lett.
2007, 9, 4279-4282. See, also: (c) Zhu, Y.; Drueckhammer, D. G. J. Org.
Chem. 2005, 70, 7755-7760. (d) Kotke, M.; Schreiner, P. R. Synthesis
2007, 779-790.
(19) For other examples of asymmetric ketone cyanosilylation, see: (a)
Hamashima, Y.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2000, 122,
7412-7413. (b) Tian, S. K.; Deng, L. J. Am. Chem. Soc. 2001, 123, 6195-
6196. (c) Deng, H. B.; Isler, M. R.; Snapper, M. L. Hoyveda, A. H. Angew.
Chem., Int. Ed. 2002, 41, 1009-1012. (d) Ryu, D. H.; Corey, E. J. J. Am.
Chem. Soc. 2005, 127, 5384-5387. (e) Liu, X. H.; Qin, B.; Zhou, X.; He,
B.; Feng, X. M. J. Am. Chem. Soc. 2005, 127, 12224-12225. (f) Qin, B.;
Liu, X. H.; Shi, J.; Zheng, K.; Zhao, H. T.; Feng, X. M. J. Org. Chem.
2007, 72, 2374-2378 and refs cited therein.
(20) Reviews: (a) Gregory, R. J. H. Chem. ReV. 1999, 99, 3649-3682. (b)
Brunel, J.-M.; Holmes, I. P. Angew. Chem., Int. Ed. 2004, 43, 2752-2778.
(21) Discussions on the biological mechanism for cyanohydrin formation: (a)
Wajant, H.; Pfizenmaier, K. J. Biol. Chem. 1996, 271, 25830-25834. (b)
Dreveny, I.; Kratky, C.; Gruber, K. Protein Sci. 2002, 11, 292-300. (c)
Stranzl, G. R.; Gruber, K.; Steinkellner, G.; Zangger, K.; Schwab, H.;
Kratky, C. J. Biol. Chem. 2004, 279, 3699-3707. (d) Gruber, K.; Gartler,
G.; Krammer, B.; Schwab, H.; Kratky, C. J. Biol. Chem. 2004, 279, 20501-
20510.
(22) Mechanistic investigations of the addition of cyanide to carbonyls not
promoted by chiral catalysts: (a) Lapworth, A. J. Chem. Soc. 1903, 83,
995-1005. (b) Lapworth, A. J. Chem. Soc. 1904, 85, 1206-1214. (c)
Lapworth, A.; Helmuth, R.; Manske, F. J. Chem. Soc. 1928, 2533-2549.
(d) Lapworth, A.; Manske, R. H. F. J. Chem. Soc. 1930, 1976-1981. (e)
Svirbely, W. J.; Roth, J. F. J. Am. Chem. Soc. 1953, 75, 3106-3111. (f)
Svirbely, W. J.; Brock, F. H. J. Am. Chem. Soc. 1955, 77, 5789-5792. (g)
Ching, W. M.; Kallen, R. G. J. Am. Chem. Soc. 1978, 100, 6119-6124.
(h) Corcoran, R. C.; Ma, J. J. Am. Chem. Soc. 1992, 114, 4536-4542. (i)
Denmark, S. E.; Chung, W.-J. J. Org. Chem. 2006, 71, 4002-4005.
(23) Kinetic analyses of catalytic, asymmetric additions of cyanide to CdO or
CdN: (a) Takamura, M.; Hamashima, Y.; Usuda, H.; Kanai, M.; Shibasaki,
M. Angew. Chem., Int. Ed. 2000, 39, 1650-1652. (b) Belokon, Y. N.;
Green, B.; Ikonnikov, N. S.; Larichev, V. S.; Lokshin, B. V.; Moscalenko,
M. A.; North, M.; Orizu, C.; Peregudov, A. S.; Timofeeva, G. I. Eur. J.
Org. Chem. 2000, 2655-2661. (c) Josephsohn, N. S.; Kuntz, K. W.;
Snapper, M. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2001, 123, 11594-
11599. (d) Belokon, Y. N.; Blacker, A. J.; Carta, P.; Clutterbuck, L. A.;
North, M. Tetrahedron 2004, 60, 10433-10477. (e) Yamagiwa, N.; Tian,
J.; Matsunaga, S.; Shibasaki, M. J. Am. Chem. Soc. 2005, 127, 3413-
3422.
(14) Cinchona alkaloid-derived thioureas: (a) Vakulya, B.; Varga, S.; Csampai,
A.; Soo´s, T. Org. Lett. 2005, 7, 1967-1969. (b) McCooey, S. H.; Connon,
S. J. Angew. Chem., Int. Ed. 2005, 44, 6367-6370. (c) Ye, J. X.; Dixon,
D. J.; Hynes, P. S. Chem. Commun. 2005, 4481-4483. (d) Bernardi, L.;
Fini, F.; Herrera, R. P.; Ricci, A.; Sgarzani, V. Tetrahedron 2006, 62, 375-
380. (e) Marcelli, T.; van der Haas, R. N. S.; van Maarseveen, J. H.;
Hiemstra, H. Angew. Chem., Int. Ed. 2006, 45, 929-931. (f) Tillman, A.
L.; Ye, J. X.; Dixon, D. J. Chem. Commun. 2006, 1191-1193. (g) Mattson,
A. E.; Zuhl, A. M.; Reynolds, T. E.; Scheidt, K. A. J. Am. Chem. Soc.
2006, 128, 4932-4933. (h) Song, J.; Wang, Y.; Deng, L. J. Am. Chem.
Soc. 2006, 128, 6048-6049. (i) Wang, J.; Li, H.; Zu, L. S.; Jiang, W.;
Xie, H. X.; Duan, W. H.; Wang, W. J. Am. Chem. Soc. 2006, 128, 12652-
12653. (j) Bode, C. M.; Ting, A.; Schaus, S. E. Tetrahedron 2006, 62,
11499-11505. (k) Wang, B.; Wu, F.; Wang, Y.; Liu, X. J. Am. Chem.
Soc. 2007, 129, 768-769. (l) Zu, L.; Wang, J.; Li, H.; Hexin, X.; Jiang,
W.; Wang, W. J. Am. Chem. Soc. 2007, 129, 1036-1037. (m) Biddle, M.
M.; Lin, M.; Scheidt, K. A. J. Am. Chem. Soc. 2007, 129, 3830-3831. (n)
Wang, Y.; Li, H.; Wang, Q.-W.; Liu, Y.; Foxman, B. M.; Deng, L. J. Am.
Chem. Soc. 2007, 129, 6364-6365. (o) Amere, M.; Lasne, C. M.; Rouden,
J. Org. Lett. 2007, 9, 2621-2624. (p) Liu, T. Y.; Cui, H. L.; Chai, Q.;
Long, J.; Li, B. J.; Wu, Y.; Ding, L. S.; Chen, Y. C. Chem. Commun.
2007, 2228-2230. (q) Wang, J.; Zu, L. S.; Li, H.; Xie, H. X.; Wang, W.
Synthesis 2007, 2576-2580. (r) Dine´r, P.; Nielsen, M.; Bertelsen, S.; Niess,
B.; Jørgensen, K. A. Chem. Commun. 2007, 3646-3648. (s) Gu, C. L.;
Liu, L.; Sui, Y.; Zhao, J. L.; Wang, D.; Chen, Y. J. Tetrahedron: Asymmetry
2007, 18, 455-463.
(15) (a) Vachal, P.; Jacobsen, E. N. J. Am. Chem. Soc. 2002, 124, 10012-
10014. (b) Berkessel, A.; Cleemann, F.; Mukherjee, S; Mu¨ller, T. N.; Lex,
J. Angew. Chem., Int. Ed. 2005, 44, 807-811. (c) Inokuma, T.; Hoashi,
Y.; Takemoto, Y. J. Am. Chem. Soc. 2006, 128, 9413-9419.
(16) For studies of reactions using achiral thioureas consistent with carbonyl
activation, see: (a) Schreiner, P. R.; Wittkopp, A. Org. Lett. 2002, 4, 217-
220. (b) Schreiner, P. R. Chem. Soc. ReV. 2003, 32, 289-296. (c) Fu, A.
P.; Thiel, W. THEOCHEM 2006, 765, 45-72. (d) Kirsten, M.; Rehbein,
J.; Hiersemann, M.; Strassner, T. J. Org. Chem. 2007, 72, 4001-4011.
(17) For a review, see: (a) Schmidtchen, F. P.; Berger, M. Chem. ReV. 1997,
97, 1609-1646. See, also: (b) Antonisse, M. M. G.; Reinhoudt, D. N.
Chem. Commun. 1998, 443-448. (c) Jagessar, R. C.; Shang, M.; Scheidt,
W. R.; Burns, D. H. J. Am. Chem. Soc. 1998, 120, 11684-11692. (d)
Hisaki, I.; Sasaki, S.-I.; Hirose, K.; Tobe, Y. Eur. J. Org. Chem. 2007,
607-615.
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15873

A R T I C L E S
Zuend and Jacobsen
Figure 3. Rate dependence on [1a]. Plot of the rate of cyanosilylation of
6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50 M) catalyzed by
HCN (0.33 M) and 1a at different [1a] and at different conversions of 6b.
The curves represent least-squares fits to eq 2; kinetic parameters are given
in Scheme 2.
Figure 1. Rate dependence on [6b]. Plot of the rate of cyanosilylation of
6b with TMSCN ([TMSCN]i ) 0.50 M) catalyzed by HCN (0.33 M) and
1a (0.025 M) at different [6b]i. Each set of points represents the results
from a single in situ IR experiment: (red) [6b]i ) 0.33 M; (blue) [6b]i )
0.165 M; (black) [6b]i ) 0.083 M.
Scheme 2. Kinetic Parameters for Ketone Cyanation
of an alkoxide or alcohol intermediate to form 7b and regenerate
HCN.
Figure 2. Rate dependence on [HCN]. Plot of the rate of cyanosilylation
of 6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50 M) catalyzed by
HCN and 1a (0.025 M) at different [HCN] and at different conversions of
6b. The curves represent least-squares fits to eq 2; kinetic parameters are
given in Scheme 2.
rate ) d[7b]/dt ) k[1a]1[6b]1[HCN]1[TMSCN]0 (1)
The deviations from simple first-order kinetic behavior at
higher concentrations of catalyst 1a and HCN point to more
complicated and potentially informative nonproductive interac-
tions under certain conditions. To gain insight into these
interactions, we used data from 37 individual kinetics experi-
ments to identify an empirical rate law that was valid over a
wide range of concentrations.27 The rate law shown in eq 2 was
derived on the basis of the mechanism outlined in Scheme 2,28
and it is the simplest rate law that provides an adequate fit to
the kinetic data (Figures 2 and 3, black curves). The data are
consistent with a 1:1:1 complex, 1a‚HCN‚6b, as the active
species. This ternary complex represents the resting state of
catalyst only at high [6b]. Uncomplexed catalyst is dominant
at low concentrations of all species, where the rate law described
by eq 2 simplifies to eq 1 (i.e., at low [6b] and [HCN] the
denominator approaches a value of 1). The observed inhibition
at high HCN concentrations is explained through formation of
a catalytically inactive 1:2 complex, 1a‚(HCN)2. The deviations
from integral kinetics in catalyst concentration are explained
by a 2:1 complex, (1a)2‚HCN, that becomes dominant at high
[1a]. All but one of the equilibrium and rate constants could be
determined directly from the kinetic data (Scheme 2). Although-
formation of a 1:1 complex, 1a‚HCN is a likely intermediate
step in the formation of the 1:2 complex, 1a‚(HCN)2, its
was confirmed by 13C NMR spectroscopy.26 The kinetic data
revealed the following: (1) Whereas no reaction occurs without
HCN, 15 mol % is sufficient to effect >90% substrate
conversion. Thus HCN is a cocatalyst for the ketone cyanation.
(2) The reaction rate is independent of excess [TMSCN]. (3)
Experiments using different initial concentrations of 6b dem-
onstrated that no appreciable catalyst decomposition or product
inhibition occurs over the entire course of the reaction (Figure
1).24a (4) Rate inhibition is observed at high [HCN] (Figure 2).
(5) The reaction rate displays a less than first-order dependence
on [1a] at elevated catalyst concentrations (Figure 3). (6) At
low concentrations (approximately 0.10 M [HCN], 0.20 M [6b],
and 0.015 M [1a]), the reaction rate displays a straightforward
first-order dependence each on [1a], [6b], and [HCN] (eq 1).
These data are consistent with a rate-limiting addition of HCN
to 6b catalyzed by 1a, followed by a post rate-limiting silylation
(24) Kinetic analysis of catalytic reactions under synthetically relevant condi-
tions: (a) Blackmond, D. G. Angew. Chem., Int. Ed. 2005, 44, 4302-
4320. (b) Mathew, J. S.; Klussman, M.; Iwamura, H.; Valera, F.; Futran,
A.; Emanuelsson, E. A. C.; Blackmond, D. G. J. Org. Chem. 2006, 71,
4711-4722.
(25) Similar rates and identical enantioselectivities were observed when HCN
was generated from CH3OH. Addition of independently generated TMSOCH2-
CF3 had no effect on reaction rate.
(26) Substoichiometric amounts of HCN were generated by adding controlled
amounts of CF3CH2OH to solutions of TMSCN and 1a. In the absence of
1a, quantitative formation of HCN from CF3CH2OH or CH3OH and
TMSCN in CDCl3 requires >30 min at room temp; the reaction appears to
be nearly instantaneous when 1a is present.
(27) [HCN] ) 0.05-1.00 M; [7a] ) 0.03-0.26 M; [1a] ) 0.015-0.055 M;
[TMSCN]i ) 0.50 M. Nonlinear least-squares regression was performed
using Sigma Plot version 7.0. Details and kinetic data in tabular format
are provided in the Supporting Information.
(28) A derivation of the rate law is provided in the Supporting Information.
9
15874 J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
2K2kcat[HCN][6b][1a]tot
rate )
(2)
1 + K1[HCN] + K2[HCN][6b] + K3[HCN]2 + (1 + K [HCN] + K [HCN][6b] + K [HCN]2)2 + 8K [HCN][1a]
x
1
2
3
3
tot
formation constant K1 could not be determined accurately from
the kinetic data: values of K1 ) 0-3 M-1 all afford good fits
to the experimental data and have only a small effect on the
other constants (<20% change).29
is not straightforward. Changes in the 1H NMR chemical shifts
of the thiourea N-H protons have been used to probe these
interactions,13b,15b,31,32 but such experiments cannot always
distinguish between direct interactions and medium effects.
Indeed, the chemical shifts of all heteroatom-bound protons in
thiourea catalysts can change upon addition of Lewis basic
substrates.15a,33 To try to distinguish changes in proton reso-
nances due to specific binding interactions from those due to
To independently corroborate the individual terms in the rate
law and to gain a better understanding of how the thiourea
catalyst interacts with and activates HCN and is also susceptible
to inhibition, we analyzed the effects of unreactive substrate
and catalyst analogues and determined the effect of isotopic
substitution of HCN on the reaction rate and enantioselectivity.
(1) Effect of Added CH3CN and Pyridine Derivatives. To
determine whether the inhibitory effect observed at high
concentrations of HCN may be due to the Lewis basic properties
of this substrate, we tested the effect of added CH3CN on the
rate of the thiourea-catalyzed cyanosilylation of ketones (Figure
4). Rate inhibition was observed without any observable change
1
changes in the medium, we examined the H NMR chemical
shifts of both thiourea protons and the secondary amide proton
of 1a as a function of substituted-pyridine concentration (Figure
6).34 As expected, the chemical shifts of each of the three H
1
resonances changed upon titration of each of the three substituted
Figure 5. Rate dependence on [substituted pyridine]. Plot of the rate of
cyanosilylation of 6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50
M) catalyzed by HCN (0.33 M) and 1a (0.025 M) in the presence of
substituted pyridines: pyridine (b, red), 2,6-lutidine (1, blue), and 2,6-di-
tert-butylpyridine (9, green). The curves represent least-squares fits to eq
Figure 4. Rate dependence on [CH3CN]. Plot of the rate of cyanosilylation
of 6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50 M) catalyzed by
HCN (0.33 M) and 1a (0.025 M) in the presence of CH3CN. The curves
represent least-squares fits to eq 3 with In ) CH3CN; kinetic parameters
are given in Scheme 4. Data from [CH3CN] ) 0.0-1.5 M were used in the
least-squares fit.
3; Indpyridine, K5 ) 31 ( 2 M-1; Ind2,6-lutidine, K5 ) 8.8 ( 0.6 M-1
;
Ind2,6-di-tert-butylpyridine, K5 ) 0.8 ( 0.2 M-1
.
in enantioselectivity (97% ee under all conditions). In principle,
this result can be ascribed either to a specific binding interaction
or to a general medium effect. Specific interactions are expected
to be sensitive to both the electronic and steric properties of
the Lewis base, whereas medium effects are expected to be
largely insensitive to the steric properties of the Lewis base.
We therefore examined the effect of added substituted pyridines
on the rate of ketone cyanation (Figure 5).30 The data reveal
that pyridine is a potent inhibitor of ketone cyanosilylation,
whereas 2,6-lutidine and 2,6-di-tert-butylpyridine are succes-
sively weaker inhibitors. In all cases, cyanosilylation product
7b was isolated in 97% ee. The data provide strong support for
a direct interaction between Lewis basic compounds and 1a
rather than a generalized medium effect.
Figure 6. Plot of the 1H chemical shift of the amide and thiourea protons
of 1a (500 MHz; CD2Cl2; [1a]i ) 0.026 M, 22 °C) upon titration of various
amounts of substituted pyridines as determined by 1H NMR spectroscopy.
Substituted pyridines: pyridine (b, red), 2,6-lutidine (1, blue), and 2,6-
di-tert-butylpyridine (9, green).
Although catalysis by thiourea derivatives is generally
proposed to occur by hydrogen bonding of the thiourea to Lewis
basic substrates,11 direct characterization of such interactions
(31) (a) Kelly, T. R.; Kim, M. H. J. Am. Chem. Soc. 1994, 116, 7072-7080.
(b) Scheerder, J.; Engbersen, J. F. J.; Casnati, A.; Ungaro, R.; Reinhoudt,
D. N. J. Org. Chem. 1995, 60, 6446-6454. (c) Wu, J.-L.; He, Y.-B.; Zeng,
Z.-Y.; Wei, L.-H.; Meng, L.-Z.; Yang, T. X. Tetrahedron 2004, 60, 4309-
4314. (d) Brooks, S. J.; Gale, P. A.; Light, M. E. Chem. Commun. 2005,
4696-4698. (e) Brooks, S. J.; Edwards, P. R.; Gale, P. A.; Light, M. E.
New J. Chem. 2006, 65-70.
(32) For a review, see: Conners, K. A. Binding Constants; John Wiley &
Sons: New York, 1987.
(33) Vachal, P. Ph.D. Thesis, Harvard University, Cambridge, MA, June, 2003.
(29) If K1 is unfixed, its value is -0.3 ( 0.2 M-1
.
(30) Substituted tetrahydrofuran derivatives have been used in an analogous
manner to probe the direct involvement of Lewis basic solvents in reaction
mechanisms: (a) Wax, M. J.; Bergman, R. G. J. Am. Chem. Soc. 1981,
103, 7028-7030. (b) Cotton, J. D.; Markwell, R. D. Organometallics 1985,
4, 937-939. (c) Galiano-Roth, A. S.; Collum, D. B. J. Am. Chem. Soc.
1989, 111, 6772-6778. (d) Church, T. L.; Getzler, Y. D. Y. L.; Coates, G.
W. J. Am. Chem. Soc. 2006, 128, 10125-10133.
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15875

A R T I C L E S
Zuend and Jacobsen
pyridines. However, less sterically demanding pyridines caused
significantly greater changes in chemical shift of both the amide
and thiourea protons.35 The fact that the most downfield thiourea
proton shows the smallest chemical shift change may be
attributed to its internal hydrogen bond to the tertiary amine;
DFT calculations suggest that this bond is maintained upon
pyridine binding. We attribute these results to a combination
of nonspecific, medium effects and specific, pyridine-thiourea
interactions that are sensitive to the steric properties of the
pyridine derivative.36
The kinetic and spectroscopic data demonstrate unambigu-
ously that H-bond acceptors such as pyridine bind to the thiourea
of 1a, and that this binding inhibits the ketone cyanation of 6b
without affecting the enantiomeric excess of the product 7b.
From the similarities in the kinetic behavior of CH3CN and
pyridine, we infer that CH3CN binds to 1a in an analogous
manner under the reaction conditions, and that HCN also binds
to the thiourea moiety of 1a through its Lewis basic nitrogen
lone pair. The observation of constant enantioselectivity in the
presence or absence of CH3CN suggests that the mechanism of
the productive pathway remains unchanged; that is, no reaction
occurs when the inhibitor is bound to the catalyst, because such
a mechanism would almost certainly induce different enanti-
oselectivity. It can be concluded that the thiourea functionality
is involved directly in catalysis of cyanosilylation; however,
the data do not reveal how the thiourea is involved in the
catalytic mechanism. Thus, although we conclude that HCN can
bind to the thiourea as a Lewis base, we cannot determine
whether this binding is productive (i.e., on the catalytic cycle)
or inhibitory (off the catalytic cycle): inferring a transition
structure from an observable substrate-catalyst interaction could
lead to invalid conclusions.37
Figure 7. Rate dependence on [trialkylamine]. Plot of the rate of
cyanosilylation of 6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50
M) catalyzed by HCN (0.33 M) and 1a (0.025 M) in the presence of
trialkylamines: i-Pr2NEt (b, red), Et3N (9, blue), and Me2NEt (1, green).
Rate data were obtained at 20% conversion of 6b.
Figure 8. Effect of trialkylamines on enantioselectivity. Plot of the
enantiomeric excess (%) of 7b obtained in the cyanosilylation of 6b ([6b]i
) 0.33 M) with TMSCN catalyzed by HCN (0.33 M) and 1a (0.025 M) in
the presence of trialkylamines: i-Pr2NEt (b, red), Et3N (9, blue), and Me2-
NEt (1, green).
(2) Effect of Added Trialkylamines. To establish whether
the trialkylamine portion of 1a is involved in either productive
(34) The 1H NMR spectrum of 1a becomes uninterpretable below ca. -10 °C,
presumably because bond rotation occurs on NMR time scales at those
temperatures. We were thus limited to NMR experiments at or near room
temperature.
(35) The binding constants of pyridine and substituted pyridines with different
metal ions have been determined: Kapinos, L. E.; Sigel, H. Inorg. Chim.
Acta 2002, 337, 131-142.
or unproductive interactions with HCN, we examined the effect
of exogeneous trialkylamines (dimethylethylamine, Me2NEt;
triethylamine, Et3N; diisopropylethylamine, i-Pr2NEt) on the rate
and enantioselectivity of ketone cyanosilylation (Figures 7 and
8). The data reveal that the degree of rate inhibition and the
observed decreases in enantioselectivity correlate with the size
of the trialkylamine. It is possible that trialkylamines bind to
the thiourea of 1a directly. However, titration of 1a with
trialkylamines revealed the same dependence on steric properties
as observed with substituted pyridines. Addition of up to 0.50
M i-Pr2NEt to a 0.026 M solution of 1a in CD2Cl2 resulted in
no significant changes to the chemical shifts of either the amide
or thiourea protons (<0.05 ppm); addition of Et3N and Me2-
NEt resulted in small but significant changes (up to 0.2 and 0.5
ppm, respectively). Because the inhibition trends determined
kinetically and the binding trends determined spectroscopically
show opposite dependence on the steric properties of trialky-
lamines, we conclude that direct binding cannot be solely
responsible for rate inhibition. Instead, the inhibition effects may
be explained by the fact that the Brønsted basicity of trialky-
lamines also correlates with their size.38 It is likely that
(36) The correlation between the steric properties of the Lewis base and the
binding affinity to the thiourea might suggest that catalysts such as 1a and
1b might be highly sensitive to substrate structure and therefore limited in
scope. However, this is not the case in the ketone cyanation reaction. Despite
compelling evidence that cyanosilylation proceeds via ketone binding to
the thiourea protons (vide infra), the reaction rate and enantioselectivity
1
are remarkably insensitive to the steric properties of the ketone. H NMR
studies of acetophenone derivatives with differing steric properties (ac-
etophenone, isobutyrophenone, and 2,2-dimethylpropiophenone) also re-
vealed relatively small chemical shift changes (0.1-0.3 ppm) that were
insensitive to the steric properties of the substrate. These observations may
be ascribed simply to the fact that the ketone substituents are remote from
the catalyst compared to the substituents on 2,6-disubstituted pyridines.
(37) For the classic example in the context of asymmetric catalysis, see: Halpern,
J. Science 1982, 217, 401-407.
(38) For a compilation of the gas-phase basicities (proton affinities) of sterically
differentiated amines, see: Drago, R. S.; Cundari, T. R.; Ferris, D. C. J.
Org. Chem. 1989, 54, 1042-1047. Proton affinity is strongly correlated
with the number and size of alkyl groups on the amine. For example, PA-
(NH3) ) 204.0, PA(NMe3) ) 225.1, PA(NEt3) ) 232.3.
(39) n-Bu4NCN was also found to be a potent inhibitor of ketone cyanosilylation.
Unfortunately, we were unable to achieve conditions under which kinetic
data with this inhibitor were reproducible, perhaps because of the
hygroscopic nature of this reagent.
(40) In the absence of 1a, i-Pr2NEt mediates the cyanosilylation of 6b at a rate
sufficient to account for the reduced ee observed under the catalytic
conditions (4% conversion after 4 h with 0.50 M i-Pr2NEt). In contrast,
Me2NEt does not mediate the cyanosilylation of 6b at -78 °C (< 0.5%
conversion after 4 h with 0.50 M Me2NEt).
(42) Details are provided in the Supporting Information. Because the inhibition
effect by DCN is greater than with HCN, this KIE represents an upper
limit for the intrinsic KIE (i.e., the true intrinsic KIE, kH/kD, is e 0.64 (
0.04). Although it is possible, in principle, to determine the KIE more
accurately from experiments at lower concentrations of HCN and DCN,
such experiments are prone to both random and systematic error because
trace amounts of water can generate HCN from TMSCN.
(41) K2 ) 20 ( 1 M-2, K3 ) 25 ( 1 M-2, K4 ) 96 ( 30 M-2, K2kcat ) 0.099
( 0.005 s-1 M-2. As with HCN, an accurate value for K1 could not be
determined, and K1 ) 0 was used for the analysis.
9
15876 J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
Scheme 3. Possible Modes of Rate Inhibition and Enantioselectivity Suppression Induced by Trialkylamines
trialkylamines engage in a hydrogen-bond interaction with HCN
under the reaction conditions, and the resulting complex may
inhibit cyanosilylation by binding to 1a, in a manner analogous
to the much weaker binding of neutral HCN or CH3CN (Scheme
3).39 Trialkylamines also catalyze a racemic reaction, leading
to depressed ee’s at high concentrations, and more basic
trialkylamines promote such pathways more effectively.40 The
trends that are observed with exogenous trialkylamines are also
pronounced with tertiary amino-thiourea catalysts: di-n-propy-
lamino catalyst 1a catalyzes the cyanosilylation of 6b 20 times
faster than dimethylamino 1b, and with slightly higher enan-
tiomeric excess (97% ee versus 95% ee).
slightly stronger hydrogen bond has been observed in the gas
phase between DCN and NMe3 than betweeen HCN and
NMe3.44,45 The greater acidity of DCN in this reaction is also
borne out by DFT calculations described below.46
B. Mechanistic Possibilities. The kinetic analysis using
Lewis and Brønsted basic additives has revealed several modes
of inhibition, as predicted by eq 2. To test whether the effects
of exogenous inhibitors could themselves be modeled quanti-
tatively, we derived an analogous rate law that accounts for
direct inhibition via hydrogen bonding to 1a (eq 3, Scheme 4).
Scheme 4. Kinetic Parameters for Ketone Cyanation in the
Presence of CH3CN
(3) Isotope Effects: Comparison of HCN and DCN. The
rates of both HCN- and DCN-promoted ketone cyanosilylation
were measured over a 20-fold range of [HCN] and [DCN]
(Figure 9). Kinetic modeling over the entire range of DCN
concentrations reveals excellent agreement with the rate law
given by eq 2 and demonstrates that both the intrinsic rate (at
low concentrations) and the inhibitory effect (at high concentra-
tions) are greater with DCN than with HCN.41 Comparison of
the rates at low concentrations of HCN and DCN where the
inhibitory effect is small (e0.10 M) allows us to estimate an
intrinsic kinetic isotope effect (kH/kD) of 0.64 ( 0.05.42 Both
effects may be associated with the greater acidity of DCN if its
partial or complete deprotonation plays a role in both the
productive and inhibitory pathways. Although a priori predic-
tions of relative acidities of isotopically labeled compounds can
Rate data in the presence of inhibitors most structurally
analogous to HCN and 1a were used at concentrations similarto
those in the catalytic reaction (e.g., [CH3CN] ) 0.3-1.5 M).
Lewis basic pyridine derivatives were shown to bind to
thethiourea of 1a, and both CH3CN and HCN appear to be
sufficiently basic to do so as well. Thus the binding constant
of CH3CN (K5 ) 1.65 ( 0.07 M-1) provides an estimate for
the binding constant of HCN as a Lewis base (Scheme 4). The
binding constants of substituted pyridines were determined in
an analogous manner (Figure 5).
be challenging and depend on the degree of proton transfer,43
a
It is striking that the factors that contribute to effective
catalysis are the same factors that lead to depressed rates at
(43) (a) Scheiner, S.; EÅ uma, M. J. Am. Chem. Soc. 1996, 118, 1511-1521. (b)
Scheiner, S. Hydrogen Bonding: A Theoretical PerspectiVe, Oxford
University Press: New York, 1997. (c) Scheiner, S. Biochem. Biophys.
Acta 2000, 28-42.
(44) Rego, C. A.; Batten, R. C.; Legon, A. C. J. Chem. Phys. 1998, 89, 696-
702.
(45) HCN/DCN isotope effects: (a) Hout, R. F., Jr.; Wolfsberg, M.; Hehre, W.
J. Am. Chem. Soc. 1980, 102, 3296-3298. (b) Reenstra, W. W.; Abeles,
R. H.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 1016-1024. (c)
Cybulksi, S. M.; Scheiner, S. J. Am. Chem. Soc. 1987, 109, 4199-4206.
(d) Zabarnick, S.; Fleming, J. W.; Lin, M. C. Chem. Phys. 1991, 150, 109-
115. (e) McDowell, S. A. C.; Forde, T. S. J. Chem. Phys. 2002, 117, 6032-
6037. See also ref 33.
(46) It is interesting that the complex kinetic dependence on HCN (1st order at
low concentrations, negative order at high concentrations) leads to a
situation where there is almost no KIE at the concentrations at which the
reaction is normally run (0.30-0.40 M).
Figure 9. Rate dependence on [HCN] and [DCN]. Plot of the rate of
cyanosilylation of 6b ([6b]i ) 0.33 M) with TMSCN ([TMSCN]i ) 0.50
M) catalyzed by HCN and 1a (0.025 M) at different [HCN] or [DCN] at
50% conversion of 6b. The curves represent least-squares fits to eq 2.
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15877

A R T I C L E S
Zuend and Jacobsen
Scheme 5. Mechanistic Possibilitiesa
a Both mechanisms are consistent with the experimentally determined rate law and are kinetically indistinguishable. The structures of these intermediates
are based on DFT calculations (B3LYP/6-31G(d)) using model catalyst 1b.
2K2kcat [HCN][6b][1a]tot
rate )
(3)
1 + K1[HCN] + K2[HCN][6b] + K3[HCN]2 + K5[In] + (1 + K [HCN] + K [HCN][6b] + K [HCN]2 + K [In])2 + 8K [HCN][1a]
x
1
2
3
5
3
tot
high concentrations: (1) more basic amines are both better
catalysts (1a versus 1b) and stronger inhibitors (i-Pr2NEt versus
Me2NEt) and (2) DCN is both more reactive at low concentra-
tions and a better inhibitor at high concentrations than HCN.
Scheme 5 depicts a plausible set of equilibria that accounts for
the observations gleaned from the kinetic analysis.47 One mode
of catalyst inhibition at high concentrations of HCN is via
formation of a 1:2 complex between 1a and HCN (1a‚2HCN,
Scheme 5). In addition, trialkylamines including 1a are basic
enough to complex or deprotonate HCN and DCN, and general
or specific base activation of HCN/DCN is important for both
catalysis and inhibition. Inhibition by trialkylamines likely
occurs through the formation of cyanide anion, which is more
Lewis basic than HCN and so binds to thioureas at lower
concentrations. The lower-than-first-order kinetic dependence
on 1a may be ascribed to the reversible formation of inactive
catalyst aggregates, either spontaneously or through bridging
cyanide ions. The latter possibility is supported by the observed
inhibitory effect of simple tertiary amine additives.48
However the kinetic analysis also demonstrates that multiple
modes exist for catalyst association to HCN, and the data do
not allow us to distinguish productive and unproductive binding
modes. Two mechanisms that are consistent with the available
data are depicted in Scheme 5: mechanism A involves addition
of cyanide to a thiourea-bound ketone, whereas mechanism B
involves addition of thiourea-bound cyanide to a ketone
activated by the protonated amine. Because the species depicted
in Scheme 5 are in rapid equilibrium,49 a kinetic analysis alone
cannot distinguish between these mechanisms: both have
transition structures with identical compositions and both involve
HCN deprotonation, and therefore have identical elementary rate
laws and similar predicted isotope effects. Assessing which of
these mechanisms is operative therefore requires an alternative
approach.
C. Theoretical Analysis of the Selectivity-Determining
Step. Despite the complexities observed in the kinetic analysis,
the enantioselectivity of ketone cyanosilylation is unchanged
over a range of concentrations. The lack of dependence of
enantioselectivity on reaction conditions suggests that the
transition structures leading to the major and minor enantiomer
have the same stoichiometry, and the selectivity-determining
The key findings of the kinetic analysissthat tertiary amine
activation of HCN is important for catalysis and unproductive
binding to the thiourea of 1a inhibits catalysissprovide compel-
ling evidence that both the thiourea and the trialkylamine
functions of 1a are involved directly in the catalytic pathway.
(48) 1H NMR experiments show that the chemical shifts of the amide and
thiourea protons of 1a are concentration-dependent in the absence of other
substrates; however, the addition of 2,6-di-tert-butylpyridine results in small
but significant 1H NMR chemical shift changes without inhibiting the rate
of ketone cyanosilylation. Thus a concentration-dependent 1H NMR
spectrum does not constitute evidence for catalyst aggregation. IR spectra
of 1a (0.02-0.10 M) show good adherence to Beer’s law, indicating that
significant direct aggregation does not occur. The possibility of small
amounts of direct aggregation cannot be excluded, but this would not change
our overall conclusions. Details are included in the Supporting Information.
(47) The structures of these intermediates are based on DFT calculations
(B3LYP/6-31G(d)) using model catalyst 1b. The calculations suggest that
the secondary amide lone pair may play a role in HCN binding, perhaps
providing an explanation for the large inhibitory effect observed kinetically.
However, evaluating the importance of such an interaction will require
higher level calculations and inclusion of solvent effects.
9
15878 J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
Figure 10. Variation of C-C-C-O dihedral angles between CdO and
CdC (shown in red) in the transition structures for cyanide ion addition to
(A) 3-penten-2-one and (B) acetophenone in the absence of catalyst.
step is the same under all conditions examined. Encouraged that
this system is mechanistically well behaved and therefore may
be amenable to theoretical analysis, we turned to density
functional theory (DFT) calculations to distinguish between the
pathways involving HCN-amine interactions depicted in Scheme
5.50,51 Although DFT methods have been used successfully to
probe the mechanisms of catalytic asymmetric reactions, limita-
tions have also been noted;52 however, the wealth of available
enantioselectivity data in the ketone cyanation reaction might
allow correlation between experiment and theory and serve to
validate the computational results.
Our analysis began with a thorough conformational study of
1b, which catalyzes the cyanosilylation of acetophenone (6a)
with 95% ee, using the B3LYP level of theory53 and several
basis sets.54 We then examined the transition structure for the
gas-phase addition of anionic cyanide to 3-penten-2-one and
acetophenone in the absence of catalyst (Figure 10). In both
cases, transition structures were identified in which the CdO
bond was nearly coplanar with the CdC bond (in 3-penten-2-
one) or the aromatic ring (in acetophenone). The energetic cost
of accessing conformations in which these bonds were not
coplanar was assessed by systematically varying the C-C-
C-O dihedral angle between the CdO and CdC bonds for the
transition structures shown in Figure 10.55 The data reveal a
distinct preference of the CdO of the electrophile to be coplanar
with the CdC bond in the transition structure for both substrates
(Figure 11). Two minimum energy conformations were identi-
fied for 3-penten-2-one (A, s-cis, I; and a distorted s-trans, III),
but the s-cis transition structure is significantly more stable;
because of its rotational symmetry, acetophenone has only one
minimum energy conformation (i.e., B, I ) III).
Figure 11. Torsional energy profiles for the addition of cyanide ion to
(A) 3-penten-2-one and (B) acetophenone. Plot of relative energy of cyanide
addition transition structure at the B3LYP/6-31+G(d,p) level versus C-C-
C-O dihedral angle. Selected structures are depicted in Figure 10.
Transition-state calculations at the B3LYP/6-31G(d) level56,57
on the cyanide addition to acetophenone promoted by amino-
thiourea catalysts reveal geometries that are remarkably similar
to the calculated transition-structure geometries in the absence
of catalyst (Supporting Information). Most significant, calcula-
tions on model catalyst 5 indicate that both mechanisms A and
(55) Conformational analysis was performed by first locating the lowest energy
transition structure for the addition of cyanide to each substrate. The length
of partially formed C-C bond between cyanide and the carbonyl was found
to be 2.030 Å for acetophenone and 1.945 Å for 3-penten-2-one. This C-C
bond distance was then fixed, and the C-C-C-O dihedral (shown in red)
was varied in 15° increments. A geometry optimization was performed at
each point with all other coordinates fully relaxed. The transition structures
in which the C-C-C-O dihedral angle is varied will likely have somewhat
different optimal C-C bond lengths in the true transition structures for
those conformations, so this procedure provides a lower limit for the
energetic preference for the aromatic ring or vinyl group and the reacting
CdO bond to be coplanar in the transition structure.
(56) Calculations at the B3LYP level of theory using the 6-31G(d) basis set
have been shown previously to reproduce relative energies of diastereomeric
transition states that involve hydrogen bonding or hydrogen atom transfer.
For examples, see: (a) Bahmanyar, S.; Houk, K. N.; Martin, H. J.; List, B.
J. Am. Chem. Soc. 2003, 125, 2475-2479. (b) Bahmanyar, S.; Houk, K.
N. Org. Lett. 2003, 5, 1249-1251. (c) Cheong, P. H.-Y.; Houk, K. J. Am.
Chem. Soc. 2004, 126, 13912-13913.
(57) For a discussion on differences between the 6-31G(d) and 6-31+G(d,p)
basis sets in catalytic asymmetric reactions involving hydrogen atom transfer
see: Clemente, F. R.; Houk, K. N. J. Am. Chem. Soc. 2005, 127, 11294-
11302.
(49) 1H NMR spectra reveal only two thiourea resonances under all conditions
examined, consistent with rapid exchange. Unlike most other carbon acids,
HCN undergoes rapid proton exchange with oxygen and nitrogen bases:
(a) Bednar, P. A.; Jencks, W. P. J. Am. Chem. Soc. 1985, 107, 7117-
7126. (b) Bednar, P. A.; Jencks, W. P. J. Am. Chem. Soc. 1985, 107, 7126-
7134.
(50) DFT calculations were carried out using Gaussian 98 or Gaussian 03:
Frisch, M. J.; et. al. Gaussian 03; Gaussian, Inc.: Pittsburgh, PA, 2003.
(51) For theoretical studies on the addition of cyanide to imines, see: (a) Arnaud,
R.; Adamo, C.; Cossi, M.; Milet, A.; Valle´e, Y.; Barone, V. J. Am. Chem.
Soc. 2000, 122, 324-330. (b) Li, J.; Jiang, W.-Y.; Han, K. L.; He, G.-Z.;
Li, C. J. Org. Chem. 2003, 68, 8786-8789. (c) Su, Z. S.; Hu, C. W.; Qin,
S.; Feng, X. M. Tetrahedron 2006, 62, 4071-4080.
(52) Wodrich, M. D.; Corminboeuf, C.; Schreiner, P. R.; Fokin, A. A.; Shleyer,
P. v. R. Org. Lett. 2007, 9, 1851-1854.
(53) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377.
(54) Details of these analyses are provided in the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15879

A R T I C L E S
Zuend and Jacobsen
Table 1. Experimental and Calculated Relative Activation
Energies for the Cyanosilylation of Ketonesa
Figure 12. Calculated transition structures for the addition of HCN to
acetophenone catalyzed by 1b. Transition structures leading to the (A) major
(S) and (B) minor (R) enantiomers are shown.
Consistent with experimental observations, the calculations
predict that both the secondary amide and the tert-leucine
components of the catalyst play an important role in defining
enantioselectivity (entries 1-4). They also correctly identify
enantioselectivity trends among related substrates using catalyst
1b (compare entries 1, 5, and 6; 7 and 8; and 9 and 10).63,64
The calculations also provide an explanation for the 60× higher
reactivity of secondary amide catalysts compared with tertiary
amide catalysts: in the catalyst conformation shown, a tertiary
amide would block the addition of cyanide by placing an alkyl
group rather than a hydrogen in the active site. Other transition
structures with 1b that avoid this have been calculated but are
higher in energy (>0.7 kcal/mol), presumably because the
carbonyl group is pointing directly into the substrate in these
conformations.
a Experimental ee’s were determined by chiral GC from reactions run
on 0.2 mmol scale at -78 °C using 2.2 equiv TMSCN, 1.0 equiv
F3CCH2OH, and 5-10 mol % catalyst in CH2Cl2 (0.4 M). ∆∆Gq values
were estimated with: ∆∆Gq ) -RT ln(kS/kR), T ) 195 K. Calculated ∆∆Eq
values were determined using the B3LYP/6-31G(d) level of density
functional theory and include a zero-point energy correction scaled by 0.9806
(Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513).
Transition structures were fully optimized and shown to have a single
imaginary frequency. Values in parentheses are derived from single-point
energy calculations at the mPW1K/6-31+G(d,p)//B3LYP/6-31G(d) level.
b Experimental selectivities were determined using the dipropylamine
catalysts listed; calculations were conducted with their dimethylamine
analogues (i.e., 1b, 2b, 3b, and 4b). c Four transition structures were located
with the following relative energies (kcal/mol): S, s-cis: 0.0; R, s-cis: 2.2
(1.7); S, s-trans: 2.8 (2.3); R, s-trans: 2.9 (2.4).
B in Scheme 5 are energetically accessible, but that mechanism
A is significantly (major enantiomer: 4.7 kcal/mol) more
favorable. Furthermore, calculations of diastereomeric transition
states within mechanism A using catalyst 1b predict the
experimentally observed enantiomer of product (∆∆Eq ) 0.9
kcal/mol); analogous calculations within mechanism B identify
no basis for enantioselectivity (∆∆Eq ) 0.0 kcal/mol).58 All
transition structures were shown to be legitimate first-order
saddle points by the existence of a single imaginary frequency
corresponding to the addition of cyanide to the CdO bond.
Transition state calculations at the B3LYP/6-31G(d) level of
several substrate-catalyst combinations undergoing reaction by
mechanism A reveal solid agreement with experimental trends
(Table 1). Nearly identical trends were observed with single-
point energy calculations at the mPW1K/6-31+G(d,p) level.59-62
Further support for mechanism A was provided by compari-
son of experimental and computed isotope effects in reactions
with DCN or HCN. The calculations predict an inverse HCN/
DCN kinetic isotope effect of kH/kD ) 0.25 and 0.31 for the
cyanation of 6a using catalyst 1b and 5, respectively, consistent
with the experimentally determined KIE of kH/kD e 0.64
determined with catalyst 1a (vide supra).42,65
(61) For a comparison between B3LYP and mPW1K for transition-state
calculations, see: Ess, D. H.; Houk, K. N. J. Phys. Chem. A 2005, 109,
9542-9553.
(62) Examples of the use of the mPW1K functional in studying the mechanisms
of complex organic reactions: (a) Cohen, R.; Rybtchinski, B.; Gandelman,
M.; Rozenberg, H.; Martin, J. M. L.; Milstein, D. J. Am. Chem. Soc. 2003,
125, 6532-6546. (b) Iron, M. A.; Sundermann, A.; Martin, J. M. L. J.
Am. Chem. Soc. 2003, 125, 11430-11441. (c) Fomine, S.; Tlenkopatchev,
M. A. J. Organomet. Chem. 2006, 691, 5189-5196. (d) Ussing, B. R.;
Hang, C.; Singleton, D. A. J. Am. Chem. Soc. 2006, 128, 7594-7607. (e)
Shelton, G. R.; Hrovat, D. A.; Borden, V. T.; J. Am. Chem. Soc. 2007,
129, 164-168. (f) Henon, E.; Bercier, A.; Plantier-Royon, R.; Harakat,
D.; Portella, C. J. Org. Chem. 2007, 72, 2271-2278.
(58) Larger basis sets (6-31G(d,p) and 6-31+G(d,p)) afford no significant change
in relative energy (<0.3 kcal/mol) or geometry for a model system (see
Supporting Information for details). In addition, single-point calculations
at the mPW1K/6-31+G(d,p)//B3LYP/6-31G(d) level provide similar
results: thiourea-ketone, S ) 0.0 kcal/mol, R ) 0.9 kcal/mol; thiourea-
cyanide, S ) 4.5 kcal/mol, R ) 4.9 kcal/mol.
(59) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem. A 2000,
104, 4811-4815.
(60) The mPW1K functional is more accurate than B3LYP at determining the
energies of nonbonding and hydrogen bonding interactions: Zhao, Y.;
Truhlar, D. G. J. Chem. Theory Comput. 2005, 1, 415-432.
(63) The calculations generally predict higher than observed selectivities for
the most selective cases. For a discussion on differences between predicted
and observed selectivities in catalytic asymmetric reactions, see ref
56a.
(64) In our case, mPW1K/6-31+G(d,p) is slightly more accurate than B3LYP/
6-31G(d) (average absolute errors: 0.56 and 0.60 kcal/mol).
9
15880 J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
Figure 13. Binding geometry of the ketone-thiourea interaction in the lowest energy calculated transition state for the addition of HCN to acetophenone
with catalyst 1b: (A) front view; (B) side view.
Table 2. Cyanosilyation of Dialkyl Ketones Using Catalysts with
Sterically Demanding Amides
(1) Basis for Enantioselectivity. The basis for high enanti-
oselectivities in the ketone cyanation reactions can be traced to
interactions of the electrophile aryl or vinyl groups with the
amino acid-derived portion of the catalyst (Figure 12). As noted
above, conjugated ketones have a strong preference for the
π-system of the substrate to remain coplanar with the breaking
CdO bond in the transition structure. In the pathway leading
to the minor enantiomer, this geometry leads to repulsive van
der Waals interactions between the substrate and the catalyst
that are absent in more flexible systems such as dialkyl ketones
or alanine-derived catalyst 2a. Direct repulsive interactions
between the amide π-system and the substrate π-system may
be important (Table 1, entries 1 and 3).66,67 The low predicted
enantioselectivity for pent-3-en-2-one reacting through an s-trans
conformationsin which the vinyl group is oriented away from
the amino acid-derived portion of the catalystsis a striking
indication that high enantioselectivity is not a simple conse-
quence of the presence of an sp2-hybridized substituent on the
ketone. Indeed, cyclohexenone, which is restricted to reacting
in the s-trans configuration, undergoes cyanosilylation in 10%
ee.
We sought to better understand the precise nature of the
a The ee values were determined by chiral GC from reactions run on
0.2 mmol scale at -78 °C using 2.2 equiv TMSCN and 1.0 equiv
F3CCH2OH in CH2Cl2 (0.4 M).
ketone-thiourea interaction in the cyanation reaction, and more
specifically to address the question of whether one or both
ketone lone pairs are engaged by the thiourea in the transition
state. The computed transition structure for the dominant
pathway in the cyanation of acetophenone with catalyst 1b
(Figure 12A) is reproduced in Figure 13, with all atoms except
those of the thiourea unit and the ketone omitted for clarity.
Inspection of this structure reveals that the partially rehybridized
ketone is positioned to interact through both lone pairs in a
nearly symmetrical geometry reminiscent of an enzymatic
oxyanion hole.68,69 This stands in sharp contrast to Lewis acid-
catalyzed enantioselective additions to ketones that have been
proposed to proceed through binding of a single lone pair;70-72
in these cases, stereochemical induction is thought to arise from
the differences in Lewis basicity of the two lone pairs. The
thiourea-catalyzed ketone cyanation appears to proceed through
a fundamentally different mode of ketone activation that
effectively involves symmetrical binding of both carbonyl lone
pairs in the transition structure.
(68) A detailed analysis of the geometry of thiourea-ketone interactions with
all ten catalyst-substrate combinations listed in Table 1 is provided in the
Supporting Information.
(69) Theoretical analyses of transition structures for the enzymatic additions to
carbonyl groups involving hydrogen bonding: (a) Wlodek, S. T.; An-
tosiewicz, J.; Briggs, J. M. J. Am. Chem. Soc. 1997, 119, 8159-8165. (b)
Li, G.-S.; Maigret, B.; Rinaldi, D.; Ruiz-Lo´pez, M. F. J. Comput. Chem.
1998, 19, 1675-1688. (c) Nishihara, J.; Tachikawa, H. J. Theor. Biol. 1999,
196, 513-519. (d) Zhang, Y.; Kua, J.; McCammon, J. A. J. Am. Chem.
Soc. 2002, 124, 10572-10577. (e) Zhan, C.-G.; Zheng, F.; Landry, D. W.
J. Am. Chem. Soc. 2003, 125, 2464-2474. (f) Tachikawa, H.; Igarishi,
M.; Nishihara, J.; Ishabishi, T. J. Photochem. Photobiol. B 2005, 79, 11-
23. (g) Gao, D.; Zhan, C.-G. J. Phys. Chem. B 2005, 109, 23070-23076.
(h) Manojkumar, T. K.; Cui, C.; Kim, K. S. J. Comput. Chem. 2005, 26,
606-611.
(70) For examples, see: (a) Corey, E. J.; Helal, C. J. Tetrahedron Lett. 1995,
36, 9153-9156. (b) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30,
97-102. (c) Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37,
1986-2012. (d) Masumoto, S.; Suzuki, M.; Kanai, M.; Shibasaki, M.
Tetrahedron 2004, 60, 10497-10504. (e) Kim, J. G.; Waltz, K. M.; Garcia,
I. F.; Kwiatkowski, D.; Walsh, P. J. J. Am. Chem. Soc. 2004, 126, 12580-
12585. For a crystal structure of an aldehyde bound to a chiral Lewis acid,
see: (f) Huang, F.; Iwama, T.; Rawal, V. H. J. Am. Chem. Soc. 2002, 124,
5950-5951.
(65) KIEs were calculated from zero-point energies obtained from frequency
calculations of HCN and DCN and of the transition states. The KIEs are
therefore path-independent: any equilibrium isotope effects due to the
formation of HCN-catalyst complexes are inconsequential.
(66) Duan, G.; Smith, V. H.; Weaver, D. F. J. Phys. Chem. A 2000, 104, 4521-
4532.
(67) This observation should be interpreted with caution, given the limitations
of the B3LYP functional in describing π-π interactions: Zhao, Y.; Truhlar,
D. G. Phys. Chem. Chem. Phys. 2005, 2701-2705. Efforts to probe this
interaction using electronically differentiated substrates were inconclusive.
We observe sharp decreases in enantioselectivity when using electron-
deficient acetophenone derivatives. However, these more reactive substrates
may react via an entirely different, less selective mechanism.
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15881

A R T I C L E S
Zuend and Jacobsen
Table 3. Enantioselective Cyanosilylation of Ketones Promoted by
Optimized Catalysts 8 and 10
proved to be the most selective cyanosilylation catalyst for
acetophenone reported to date (Table 3, entry 1); in addition
highly electron-deficient and sterically demanding aromatic
ketonesschallenging substrates with the first generation catalysts
undergo cyanosilylation to provide silylcyanohydrins in excellent
yield and enantiomeric excess (entries 2-5). The more sterically
demanding catalyst 10 proved effective at catalyzing the
cyanosilylation of a range of dialkyl ketones in excellent yields
and useful levels of enantiomeric excess (entries 6-10). The
fact that 2-heptanone undergoes cyanosilylation in lower enan-
tiomeric excess than dialkyl ketones bearing π or lone pair
substituents is consistent with our hypothesis that direct
electronic interactions between substrate and catalyst play an
important role in controlling enantioinduction.
Conclusions
A wide range of experimental and theoretical results support
the proposal that the amino-thiourea catalyzed ketone cyanosi-
lylation proceeds through a mechanism that involves simulta-
neous activation of HCN by the tertiary amine and of the ketone
by the thiourea. Key experiments and findings include the
following:
(1) Kinetic analysis under synthetically relevant conditions
is consistent with ternary transition structure 1a‚HCN‚6b and
revealed mechanistically significant inhibition effects at elevated
concentrations.
(2) Kinetic analysis in the presence of inactive inhibitors and
isotopically substituted substrates provides strong evidence that
both the thiourea and tertiary amine are involved directly in
catalysis.
(3) 1H NMR experiments using substrates with different steric
properties were used to distinguish between medium effects and
thiourea binding, providing a potentially general method for
characterizing these interactions.
(4) Transition state DFT calculations reproduce experimental
trends in enantioselectivity across a range of substrates and
catalysts, thereby providing strong evidence for a mechanism
of catalysis involving ketone-thiourea rather than cyanide-
thiourea interactions.
(5) Analysis of the transition structures reveals a basis for
enantioselectivity that involves direct substrate-catalyst interac-
tions and binding through both ketone lone pairs, suggesting
that the mode of asymmetric induction with chiral hydrogen-
bond donors may be fundamentally different than with chiral
Lewis acids.
The recent emergence of a wide variety of catalytic asym-
metric reactions that are proposed to operate via cooperative
mechanisms suggests that cooperative reactivity may play as
dominant a role in small-molecule asymmetric catalysis as it
does in enzymatic catalysis. However, even when structure-
activity relationships indicate that multiple functional groups
are necessary for catalysis, several different mechanisms are
often possible, and these mechanistic possibilities are often
difficult to distinguish. The present study represents a case in
which it has been possible to fully characterize the rate- and
selectivity-determining step of a synthetically interesting reaction
promoted by a bifunctional catalyst, and it highlights how
classical and modern mechanistic tools can be used successfully
in combination to probe complex transition structures in
asymmetric catalysis. This work has provided necessary insight
a Reactions using catalyst 8 were run on 1.0 mmol scale in CH2Cl2 (0.5
M) using 2.2 equiv TMSCN and 1.0 equiv CF3CH2OH. Reactions using
catalyst 10 were run on 1.0 mmol scale in 3:1 CH2Cl2/toluene (0.4 M).
b Isolated yields after silica gel chromatography. c Determined by chiral GC.
d Reaction run using 2.5 mol % catalyst.
D. Development of More Highly Enantioselective Cata-
lysts. Catalyst 1a promotes the cyanosilylation of dialkyl ketones
with low enantioselectivity (Table 2, entry 1), and high ee’s
are observed only with aryl-alkyl or alkenyl-alkyl ketones.
However, the computational studies described above suggest
that the basis for high enantioinduction results from direct steric
and electronic interactions of the substrate with the amino acid-
derived portion of the catalyst. On this basis, we reasoned that
improved results in the asymmetric cyanosilylation of ketones
may be obtained by accentuating these interactions; in particular,
fine-tuning the steric properties of the amide could potentially
lead to catalysts that promote the highly enantioselective
cyanosilylation of dialkyl ketones.
In accord with this hypothesis, analogues of 1a bearing more
sterically demanding secondary amides yielded addition products
in substantially higher ee’s (Table 2, entries 2-4). Catalyst 8
(71) For discussions and leading references on carbonyl binding to hydrogen-
bond donors, see: (a) Etter, M. C. Acc. Chem. Res. 1990, 23, 120-126.
(b) Phiko, P. M. Angew. Chem., Int. Ed. 2004, 43, 2062-2064. See, also
ref 16d.
(72) For an example of a hydrogen-bond donor proposed to activate carbonyls
by binding through a single lone pair, see: (a) Unni, A. K.; Takenaka, N.;
Yanamoto, H.; Rawal, V. H. J. Am. Chem. Soc. 2005, 127, 1336-1337.
(b) Gordillo, R.; Dudding, T.; Anderson, C. D.; Houk, K. N. Org. Lett.
2007, 9, 501-503 and refs cited therein.
9
15882 J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007

Cooperative Catalysis by Amino-Thioureas
A R T I C L E S
into one class of thiourea-catalyzed reactions, and we expect
that many of the results may be generally applicable. Neverthe-
less, it would be premature to conclude that all thiourea-
catalyzed reactions proceed via analogous mechanisms: we
remain intrigued and often perplexed by the mechanistic
possibilities that other thiourea-catalyzed reactions suggest. The
current work indicates that characterizing this potential mecha-
nistic diversity may be possible.
Dr. Lars P. C. Nielsen, and a referee for helpful suggestions,
and Ye Tao for assistance in the preparation of several catalysts.
Supporting Information Available: Complete ref 50; com-
plete experimental procedures, characterization data, chiral
chromatographic analyses of racemic and enantiomerically
1
enriched products, kinetic data, derivation of eq 2, H NMR
titration data, and geometries and energies of calculated station-
ary points. This material is available free of charge via the
Internet at http://pubs.acs.org.
Acknowledgment. This work was supported by the NIH
(GM-43214). We thank Prof. Charles B. Musgrave (Stanford),
JA0735352
9
J. AM. CHEM. SOC. VOL. 129, NO. 51, 2007 15883