C O M M U N I C A T I O N S
ꢀ-hydroxyl amide produced. The use of ammonium acetate led to slow
hydrogenation, although still with >99% ee (entry 4). The use of 2,2,2-
trifluoroethanol (TFE) as the solvent led to only slightly lower ee but a
lower yield due to an elevated level of the ꢀ-hydroxyl amide (entry 5).
We also examined the use of Rh(I)-t-Bu-Josiphos, and contrary to
the anticipation that the reductive amination would suffer from
competitive hydrogenation of 1,3c we observed excellent chemose-
lectivity to the ꢀ-amino amide with an ee as high as that observed for
hydrogenation of preformed 3 (entry 6).4 In this case, however,
ammonium acetate was a better N source than ammonium salicylate,
as the latter led to lower chemoselectivity due to the competitive keto
hydrogenation (entry 6 vs 7).
A study of the effect of excess ammonium salicylate showed that
the ammonium salicylate plays a dual role. In addition to its role in
shifting the equilibrium toward 3 (Scheme 2), it also suppresses dimer
formation and/or breaks up the dimer, releasing the product 2 along
with 3 for hydrogenation. This becomes evident from the concentration
profiles for hydrogenation of 3. As shown in Figure 1, in the presence
of salicylic acid only, the dimer builds up to significant levels and
slowly hydrogenates to form 2 (Figure 1a). The additional 3 equiv of
ammonium salicylate reduces the dimer concentration by a factor of
3 at its peak value and to zero at the end of the reaction (Figure 1b).
The extra 3 equiv of ammonium salicylate lowers neither the rate nor
the ee of the hydrogenation, demonstrating a remarkable tolerance of
the Ru catalytic system toward high concentrations of ammonium ion.
The beneficial effect of the ammonium salicylate in breaking up
the dimer was also observed for the asymmetric reductive amination
of 1 to 2 (Figure 1c), leading to the formation of 2 with high yield
and high ee (Table 2, entry 1). No dimer was observed at the end of
the reaction (Figure 1c). Figure 1c also shows that the ꢀ-keto amide
1 coexists with the enamine 3 during the hydrogenation (e.g., t ) 1-5
h) in a ratio of ∼1:3.4. Nevertheless, the keto hydrogenation of 1 does
not proceed to an appreciable degree (<3%). The remarkable tolerance
to the high concentrations of ammonium ion, the high chemoselectivity
(enamine vs keto hydrogenation), and the high enantioselectivity
(99.5% ee) of the Ru catalyst system underlie the highly efficient
asymmetric reductive amination of 1 to 2. The quasi equilibrium
between 1 and 3 during the hydrogenation suggests that the formation
of 3 is not the rate-limiting step. The rate constant for direct reductive
amination estimated from Figure 1c is ∼50% that of the hydrogenation
of preformed 3 estimated from Figure 1b (assuming a first-order rate
dependence on the pressure), presumably because of the presence of
the greater excess of ammonium ion in the case of direct reductive
amination.
The scope of the Ru-catalyzed reductive amination was also explored.
As shown in Table 3, a variety of alkyl- and aryl-substituted ꢀ-keto amides
5 were converted to the ꢀ-amino amides 6 in high yields and >94% ee’s.
In summary, we have developed a new, high-yield, highly enanti-
oselective reductive amination of ꢀ-keto amides to ꢀ-amino amides.
The atom- and step economical methodology has a broad substrate
scope and has been used to produce sitagliptin in 91% yield with
unprecedented levels of asymmetric induction. The excellent perfor-
mance of the methodology is attributable to the properties of the Ru
catalyst (high chemoselectivity and nearly perfect enantioselectivity)
and its remarkable tolerance to high concentrations of ammonium ion.
We anticipate that direct reductive amination with simple ammonium
salts will continue to be widely exploited in the synthesis of free amine
derivatives.
Acknowledgment. We thank Dr. Yi Hsiao for screening with the
Rh(I)-t-Bu-Josiphos system at an early stage of this study.
Supporting Information Available: Complete refs 4d and 5,
experimental procedures, and product characterization data. This material
References
(1) (a) Cardillo, G.; Tomasini, C. Chem. Soc. ReV. 1996, 117. (b) EnantioselectiVe
Synthesis of ꢀ-Amino Acids; Juaristi, E., Soloshonok, V. A., Eds.; Wiley-
VCH: New York, 2005 and references therein.
(2) (a) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley: New York,
1994 and references therein. For the use of aniline or benzylamine derivatives
in the asymmetric reductive amination of ketone derivatives, see, for example: (b)
Li, C.; Marcos-Villa, B.; Xiao, J. J. Am. Chem. Soc. 2009, 131, 6967. (c)
Kadyrov, R.; Riermeier, T. H.; Dingerdissen, U.; Tararov, V.; Bo¨rner, A. J.
Org. Chem. 2003, 68, 4067.
(3) For aryl ketones: (a) Kadyrov, R.; Riermeier, T. H. Angew. Chem., Int. Ed.
2003, 42, 5472. For ꢀ-keto esters: (b) Bunlaksananusorn, T.; Rampf, F.
Synlett 2005, 2682. (c) Matsumura, K.; Saito, T. U.S. Patent 142,443 A1,
2007. (d) Shimizu, H.; Nagasaki, I.; Matsumura, K.; Sayo, N.; Saito, T.
Acc. Chem. Res. 2007, 40, 1385.
Figure 1. Concentration profiles for (a, b) asymmetric hydrogenation of
ꢀ-enamine amide 3 and (c) reductive amination of ꢀ-keto amide 1. Additives:
(a) salicylic acid (1 equiv); (b) salicylic acid (1 equiv) + NH4SA (3 equiv); (c)
NH4SA (5 equiv). Conditions: 0.25 M substrate in MeOH; Ru(OAc)2((R)-dm-
segphos) (S/C ) 100); (a, b) H2 (290 psi), 75 °C; (c) H2 (435 psi), 70 °C, set
t ) 0 when temperature reached 75 °C (a, b) or 70 °C (c).
Table 3. Asymmetric Reductive Amination of ꢀ-Keto Amidesa
(4) (a) Hsiao, Y.; Rivera, N. R.; Rosner, T.; Krska, S. W.; Njolito, E.; Wang,
F.; Sun, Y.-K.; Armstrong, J. D.; Grabowski, E. J. J.; Tillyer, R. D.; Spindler,
F.; Malan, C. J. Am. Chem. Soc. 2004, 126, 9918. (b) Clausen, A. M.;
Dziadul, B.; Cappuccio, K. L.; Kaba, M.; Starbuck, C.; Hsiao, Y.; Dowling,
T. M. Org. Process Res. DeV. 2006, 10, 723. (c) Shultz, C. S.; Krska, S. W.
Acc. Chem. Res. 2007, 40, 1320. (d) Hansen, K. B.; et al. J. Am. Chem.
Soc. 2009, 131, 8798.
entry
substrate
time (h)
% yieldb
91
% eec
configuration
(5) Kim, D.; et al. J. Med. Chem. 2005, 48, 141.
(6) (a) Matsumura, K.; Zhang, X.; Saito, T. U.S. Patent 7,015,348 B2, 2004.
(b) Matsumura, K.; Hori, K.; Kakizawa, T.; Saito, T. In Proceedings of the
Summer Symposium of the Japanese Society for Process Chemistry, Tokyo,
July 2005; p 146. (c) Saito, T.; Zhang, X.; Matsumura, K.; Yokozawa, T.;
Shimizu, H. Presented at the 19th North American Catalysis Society Meeting,
abstracts/O_244.pdf (accessed July 20, 2009).
1
2
3
4
5
6
7
8
5a
5b
5c
5d
5e
5f
15
15
15
15
15
15
15
24
94.7
98.1
98.9
99.2
99.6
99.4
97.9
99.6
(-)
(-)
(-)
(S)
(+)
(+)
(-)
(+)
92
89
96 (91 )
d
95
96
81
82
5g
5h
(7) The use of ammonium benzoate and formate gave lower yields (19 and 42%,
respectively).
(8) High yields and ee’s were also observed using several other bisphosphine chiral
ligands, such as BINAP (85% yield, 99.1% ee), xyl-binap (90% yield, 98.0%
ee), and SEGPHOS (80% yield, 99.8% ee).
JA905143M
a Conditions: 0.25-0.5 M substrate in MeOH, S/C ) 100, H2 (435 psi),
NH4SA (5 equiv with respect to 5), 80 °C. b Assay yield of 6 by HPLC.
c Assayed by chiral HPLC. d Isolated yield of the free ꢀ-amino amide.
9
J. AM. CHEM. SOC. VOL. 131, NO. 32, 2009 11317