Which is the actual catalyst: Chiral phosphoric acid or chiral calcium phosphate?

Article abstract of DOI:10.1002/anie.201000824

(Figure Presented) Both catalysts work: A highly enantioselective direct Mannich-type reaction of NBoc-protected aldimines with 1,3-dicarbonyl compounds has been developed with the use of a chiral phosphoric acid in the presence or absence of Ca . The absolute stereoselectivity of the phosphoric acid catalysis was found to be opposite to that of the calcium phosphate catalysis (see scheme; Boc = tert-butoxycarbonyl).

Full text of DOI:10.1002/anie.201000824

Angewandte
Chemie
DOI: 10.1002/anie.201000824
Asymmetric Catalysis
Which Is the Actual Catalyst: Chiral Phosphoric Acid or Chiral
Calcium Phosphate?**
Manabu Hatano, Katsuhiko Moriyama, Toshikatsu Maki, and Kazuaki Ishihara*
Since Akiyama et al.[1] and Uraguchi and Terada[2] independ-
Table 1: Screening of catalysts.
ently reported chiral phosphoric acid catalysts derived from
3,3’-disubstituted 1,1’-bi(2-naphthol) (BINOL), these have
become recognized as some of the most useful organocata-
lysts.[3] However, BINOL-derived phosphoric acids are read-
ily neutralized to adventitious metal salts such as alkali or
alkaline-earth metal salts by purification on silica gel. Recent
research has been focused on the possibility of metal
contaminants in phosphoric acids. Ding and co-workers
reported that phosphoric acid washed with HCl improved
the catalytic activity in the Baeyer–Villiger reaction.[4] More-
over, Rueping et al. excluded calcium phosphate as the
potential active catalyst in their organocatalytic carbonyl-
ene reaction.[5] Herein, we report the significance of the acidic
purification of chiral phosphoric acids. Not only metal-free
chiral phosphoric acid (H[1b]) but also chiral calcium
phosphate (Ca[1a]2) catalyze enantioselective direct Man-
nich-type reactions of aldimines with 1,3-dicarbonyl com-
pounds (see Table 1). In particular, Ca[1a]2 was effective for
the enantioselective Mannich-type reaction with less-acidic
1,3-dicarbonyl compounds, including b-ketothioesters and
thiomalonates.
The enantioselective direct Mannich-type reaction of
aldimine 2a with acetylacetone (3a) catalyzed by H[1a] was
first developed by Terada and co-workers.[2,6,7] We envisioned
that alkali or alkaline-earth metal phosphates might activate
1,3-dicarbonyl compounds more effectively than the corre-
sponding phosphoric acids because of their stronger Brønsted
basicity.[8,9] As preliminary experiments, the Mannich-type
reaction of 2a with 3a was examined using alkali or alkaline-
earth metal salts of 1a (Table 1). Although LiI, NaI, MgII, and
SrII salts showed disappointing results (entries 1–3 and 5), the
CaII salt (Ca[1a]2) catalyzed the reaction, and (R)-4a was
Entry Catalyst (mol%)[a]
Solvent
T
Yield ee [%][c]
[8C] [%][b] (config.)
1
Li[1a] (5)
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
RT
RT
99 11 (S)
88 9 (S)
2
Na[1a] (5)
3
4
5
Mg[1a]2 (2.5)
Ca[1a]2 (2.5)
Sr[1a]2 (2.5)
RT >99 43 (R)
RT >99 92 (R)
RT >99 59 (R)
6
H[1a] purified on silica gel (2) CH2Cl2
RT
RT
86[d] 92 (R)[d]
88 27 (S)
7
8
H[1a] washed with HCl (2)
H[1b] washed with HCl (5)
H[1b] washed with HCl (5)
Ca[1b]2 (2.5)
CH2Cl2
CH2Cl2
RT >99 49 (S)
9[e]
10
toluene À30 >99 93 (S)
CH2Cl2 RT 93 30 (S)
[a] See the Supporting Information for details on preparation of the
catalysts. [b] Yields of isolated product. [c] Determined by HPLC on a
chiral stationary phase. [d] Compare to the original data by Terada[2]: 99%
yield and 95% ee (R). [e] Used 2a (1.2 equiv) and 3a (1 equiv). Reaction
time was 12 hours. Boc=tert-butoxycarbonyl, Naph=naphthyl.
obtained in > 99% yield with 92% ee in CH2Cl2 at room
temperature (entry 4).[10] Interestingly, this result obtained
with Ca[1a]2 was comparable to Teradaꢀs result using H[1a]
purified by silica gel (entry 6). However, with H[1a] washed
with HCl, poor and opposite enantioselectivity (27% ee,
S configuration) was observed for 4a (entry 7).[11] Moreover,
during optimization, in the absence of CaII, we found that the
sterically demanding H[1b], in place of H[1a], improved the
enantioselectivity to give (S)-4a with 49% ee (entry 8).
Further optimization of the reaction conditions with H[1b]
gave (S)-4a with 93% ee in toluene at À308C (entry 9),
although Ca[1b]2 also gave (S)-4a, albeit with low enantio-
selectivity (30% ee; entry 10).
Based on the unexpected preliminary results with the
organocatalysts shown in Table 1, we thoroughly investigated
the metal-free phosphoric acid catalysis with H[1b] washed
with HCl (Scheme 1). Acyclic a-substituted b-diketones
smoothly gave the Mannich adducts (4b–d) with high
enantioselectivities (92–95% ee). Cyclic b-diketone and b-
ketoesters also gave the corresponding products (4e–i) with
both high anti diastereoselectivities and high enantioselectiv-
ities (syn/anti = < 10: > 90, 95–98% ee). The anti diastereom-
[*] Dr. M. Hatano, Dr. K. Moriyama, Dr. T. Maki, Prof. Dr. K. Ishihara
Graduate School of Engineering
Nagoya University
Furo-cho, Chikusa, Nagoya 464-8603 (Japan)
Fax: (+81)52-789-3222
E-mail: ishihara@cc.nagoya-u.ac.jp
Homepage: http://www.nubio.nagoya-u.ac.jp/nubio4/index.htm
Prof. Dr. K. Ishihara
Japan Science and Technology Agency (JST), CREST
Furo-cho, Chikusa, Nagoya 464-8603 (Japan)
[**] Financial support for this project was provided by the JSPS, Grants-
in-Aid for Scientific Research (KAKENHI; 20245022), the Ministry of
Education, Culture, Sport, Science and Technology (MEXT)
(KAKENHI; 21750094, 21200033), and the Global COE Program of
the MEXT.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201000824.
Angew. Chem. Int. Ed. 2010, 49, 3823 –3826
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3823

Communications
bulky tBu moiety in the Boc group, which should show
significant repulsion with the 1,3-dicarbonyl compound,
would then be directed to the inside of the catalyst cavity.
However, this repulsion can be partially relieved by rotation
of the ester substituent (R’) on the aldimine in favored TS-1,
while the ketone moiety cannot, inherently, turn away in
disfavored TS-2. In TS-1, a pronucleophile would be activated
=
by coordination with the Brønsted basic P O moiety, and
subsequent attack on the Brønsted acid (i.e. proton) coordi-
nated aldimine on the si face to give the anti product.
Next, we investigated the scope of the enantioselective
direct Mannich-type reaction with 1,3-dicarbonyl compounds
under the optimized reaction conditions for Ca[1a]2
(Scheme 2). Unfortunately, the Mannich-type reaction of 2
Scheme 1. anti-Selective catalysis with chiral phosphoric acid, H-
[1b]. [a] Used 2 (1.5 equiv). [b] Diastereoselectivities for 4j and 4k
are given in the Supporting Information. Yields of isolated product.
The ee values were determined by HPLC on a chiral stationary
phase.
ers obtained here are valuable because preceding catalysts
gave the syn diastereomers as major products.[7,12] Further-
more, when S-aryl thioacetoacetate was used, the adducts
(4j and 4k) were obtained in > 99% yield with 82–86% ee.
Unfortunately, however, reactions with thiomalonates did
not proceed, and compounds 4l and 4m were obtained
only in trace quantities.
The catalysis by the unprecedented, highly anti-selec-
tive H[1b] is thought to proceed through cyclic transition
states, as the catalyst can activate both the aldimine and
the 1,3-dicarbonyl compound in a synclinal conformation.
The coordination of the aldimine to H[1b] is sterically
controlled, as illustrated by TS-1 and TS-2 (Figure 1)
because of the steric repulsion between the 9-anthryl
Scheme 2. Catalysis with chiral calcium phosphate, Ca[1a]2. [a] Diastereose-
lectivities for 4j, 4k, and 4n–s are given in the Supporting Information.
moiety of H[1b] and the aryl moiety of the aldimine. The Yields of isolated product. Xyl=xylyl.
with a-substituted b-ketoesters only gave moderate
enantio- and diastereoselectivities (see for results for
4 f).[13] Instead, a-nonsubstituted b-ketothioesters[14]
were more suitable pronucleophiles. The Mannich-
type reaction of 2a with S-phenyl thioacetoacetate
proceeded smoothly and afforded the desired product
(4j) in > 99% yield with 90% ee. The S-2,6-xylyl
thioacetoacetate also provided the corresponding prod-
ucts (4k and 4n–q) with good enantioselectivities (90–
98% ee) in the reaction of aldimines with Me, MeO, Cl,
and Br substituents with 0.5–2.5 mol% of Ca[1a]2. The
2-naphthyl and 3-thienyl substrates also gave the
Figure 1. Proposed transition states for H[1b]-based catalysis.
3824
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3823 –3826

Angewandte
Chemie
desired products (4r and 4s) with high enantioselectivities
(96–97% ee). To our great delight, S-2,6-xylyl thiomalonate
could be used in the presence of Ca[1a]2, and the correspond-
ing adducts (4m, 4t, and 4u) were obtained in high yields with
high enantioselectivities (91–95% ee). The catalysis with
thiomalonates by Ca[1a]2 gave results that were in sharp
contrast to those by H[1b], as the Brønsted basicity of H[1b]
was not sufficient to promote the reactions through the
activation of less-acidic pronucleophiles. To the best of our
knowledge, this is the first practical example of the direct
Mannich-type reaction of aldimine with b-ketothioesters[7l]
and thiomalonate. Moreover, a change in the absolute
configuration of the amino stereocenter of the products was
observed with H[1b] (Scheme 1) versus Ca[1a]2 (Scheme 2).
It is therefore noteworthy that these catalyses give enantio-
divergent methodology while using catalysts with the same
absolute configurations.
[Eq. (5)]. Moreover, b-N-Boc-protected thioester 9 was
obtained from 4u in 74% yield with 95% ee by decarbox-
ylation in DMSO/H2O at 1108C in the presence of NaCl
[Eq. (6); DMSO = dimethyl sulfoxide].[18]
b-Ketothioester 4k and thiomalonate 4m were readily
transformed into b-ketoester 5 and malonate 6, respectively,
without racemization by treatment with CF3CO2Ag/MeOH[15]
in THF at 508C [Eqs. (1) and (2); THF = tetrahydrofuran].
Although further investigation of the CaII-based catalysis
is necessary,[10,19] the Lewis acidic calcium bis(phosphate) salt
Ca[1a]2 might play a role. This salt was generated in situ and
identified in the FAB-HRMS analysis (m/z for [Ca[1a]2]+;
calcd: 1543.3736, found: 1543.3749). The 31P NMR analysis of
Ca[1a]2 in CD2Cl2 showed a broad peak at 0.05 ppm, which
suggests an oligomeric structure. However, when 2a (1 equiv)
and 3a (1.1 equiv) were added to the solution of Ca[1a]2
(2.5 mol%), a new, sharp singlet was observed at 4.55 ppm at
1 minute (just beginning) to 24 hours, which suggests a
monomeric structure of Ca[1a]2. Therefore, other possible
candidates, such as calcium mono(phosphate) salt, CaII-free
organocatalyst (i.e. H[1a] washed with HCl, 2.02 ppm in the
31P NMR spectrum), or CaII-enolate, are unlikely. The CaII
center would be highly sterically hindered by the four 4-(b-
naphthyl)-C6H4 moieties, and a half-pipe-like chiral groove
would be formed around the CaII center (Figure 2). There-
fore, the less sterically hindered cyclic transition state (TS-3)
would be favored in Ca[1a]2-based catalysis, and pronucleo-
Decarboxylations of b-ketoesters and malonates are often
difficult as basic conditions are often required. In sharp
contrast, compound 4k was converted into b-amino ketone
7—which is difficult to obtain directly from 2a and ace-
tone[16]—by PdII/HCO2H catalysis[17] [Eq. (3); DMF = N,N-
dimethylformamide, dppf = 1,1’-bis(diphenylphosphanyl)fer-
rocene]. However, for 5, decarboxylation with LiOH gave a
complex mixture (7, 13% yield and 69% ee) [Eq. (4)].
=
philes activated by the Brønsted basic P O moiety would
attack the aldimine activated by the Lewis acidic CaII center
on the re face to give the R products.
The reduction of 4m with Raney Ni under H2 (1 atm) gave
the chiral b-amino alcohol 8 in 90% yield with 95% ee
Figure 2. Proposed transition state for Ca[1a]2-based catalysis.
Angew. Chem. Int. Ed. 2010, 49, 3823 –3826
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3825

Communications
2010, 16, 779; l) M. Hatano, T. Horibe, K. Ishihara, J. Am. Chem.
Soc. 2010, 132, 56.
In summary, we have developed a highly enantioselective
direct Mannich-type reaction of aldimines with a broad range
of 1,3-dicarbonyl compounds involving unprecedented b-
ketothioesters and thiomalonate conversions with the use of a
chiral phosphoric acid in the presence or absence of CaII. The
presence of small amounts of metal contaminants in “puri-
fied” phosphoric acid may trigger unexpected excellent
results or may invalidate these evaluations, and the further
application of this approach to another asymmetric catalysis
with chiral alkali or alkaline-earth salts is now underway.
[8] Selected reports of asymmetric catalyses using phosphoric acid/
metal salts: a) H. Alper, N. Hamel, J. Am. Chem. Soc. 1990, 112,
2803; b) J. Inanaga, Y. Sugimoto, T. Hanamoto, New J. Chem.
1995, 19, 707; c) G. L. Hamilton, E. J. Kang, M. Mba, F. D. Toste,
Science 2007, 317, 496; d) M. Hatano, T. Ikeno, T. Matsumura, S.
Torii, K. Ishihara, Adv. Synth. Catal. 2008, 350, 1776; e) K. Shen,
X. Liu, Y. Cai, L. Lin, X. Feng, Chem. Eur. J. 2009, 15, 6008.
[9] For reviews on acid–base chemistry, see: a) M. Kanai, N. Kato,
E. Ichikawa, M. Shibasaki, Synlett 2005, 1491; b) K. Ishihara, A.
Sakakura, M. Hatano, Synlett 2007, 686.
[10] During the preparation of this manuscript, Kobayashi and co-
workers reported that the chiral calcium catalyst with a neutral
tridentate pybox ligand was effective for the enantioselective
direct Mannich-type reaction of aldimines with less-acidic
malonates. T. Poisson, T. Tsubogo, Y. Yamashita, S. Kobayashi,
J. Org. Chem. 2010, 75, 963. pybox = 2,6-bis(2-oxazolin-2-yl)pyr-
idine).
Received: February 10, 2010
Revised: March 13, 2010
Published online: April 20, 2010
Keywords: asymmetric catalysis · calcium ·
.
Mannich-type reactions · organocatalysis · phosphoric acids
1
[11] We found that different H NMR spectra, in [D6]DMSO, were
observed for H[1a] purified on silica gel and H[1a] washed with
HCl. LiI, NaI, MgII, CaII, and SrII salts showed the same 1H NMR
spectra as H[1a] purified on silica gel. In the FAB-LRMS
analysis of H[1a] purified on silica gel, two major m/z peaks at
775 and 791 and two minor m/z peaks at 1529 and 1543 were
observed, which indicate [H[1a] + Na]+, [Ca[1a]]+, [2(H[1a])
+ Na]+, and [Ca[1a]2]+, respectively. Therefore, H[1a] purified
on silica gel involved at least NaI and CaII salts, although the ratio
of NaI/CaII was not determined.
[1] T. Akiyama, J. Itoh, K. Yokota, K. Fuchibe, Angew. Chem. 2004,
116, 1592; Angew. Chem. Int. Ed. 2004, 43, 1566.
[2] a) D. Uraguchi, M. Terada, J. Am. Chem. Soc. 2004, 126, 5356;
b) I. D. Gridnev, M. Kouchi, K. Sorimachi, M. Terada, Tetrahe-
dron Lett. 2007, 48, 497.
[3] For reviews on chiral phosphoric acid, see: a) T. Akiyama,
Chem. Rev. 2007, 107, 5744; b) X. Yu, W. Wang, Chem. Asian J.
2008, 3, 516; c) M. Terada, Chem. Commun. 2008, 4097.
[4] S. Xu, Z. Wang, X. Zhang, X. Zhang, K. Ding, Angew. Chem.
2008, 120, 2882; Angew. Chem. Int. Ed. 2008, 47, 2840.
[5] M. Rueping, T. Theissmann, A. Kuenkel, R. M. Koenigs, Angew.
Chem. 2008, 120, 6903; Angew. Chem. Int. Ed. 2008, 47, 6798.
[6] For reviews on asymmetric Mannich-type reactions, see: a) A.
Cꢁrdova, Acc. Chem. Res. 2004, 37, 102; b) G. K. Friestad, A. K.
Mathies, Tetrahedron 2007, 63, 2541; c) A. Ting, S. E. Schaus,
Eur. J. Org. Chem. 2007, 5797; d) J. M. M. Verkade, L. J. C.
van Hemert, P. J. L. M. Quaedflieg, F. P. J. T. Rutjes, Chem. Soc.
Rev. 2008, 37, 29.
[7] Asymmetric direct Mannich-type reactions of aldimines with
1,3-dicarbonyl reagents: a) Y. Hamashima, N. Sasamoto, D.
Hotta, H. Somei, N. Umebayashi, M. Sodeoka, Angew. Chem.
2005, 117, 1549; Angew. Chem. Int. Ed. 2005, 44, 1525; b) A. L.
Tillman, J. Ye, D. J. Dixon, Chem. Commun. 2006, 1191; c) J.
Song, Y. Wang, L. Deng, J. Am. Chem. Soc. 2006, 128, 6048; d) F.
Fini, L. Bernardi, R. P. Herrera, D. Pettersen, A. Ricci, V.
Sgarzani, Adv. Synth. Catal. 2006, 348, 2043; e) J. Song, H.-W.
Shih, L. Deng, Org. Lett. 2007, 9, 603; f) Y. Yamaoka, H. Miyabe,
Y. Yasui, Y. Takemoto, Synthesis 2007, 2571; g) S. Lou, P. Dai,
S. E. Schaus, J. Org. Chem. 2007, 72, 9998; h) Z. Chen, H.
Morimoto, S. Matsunaga, M. Shibasaki, J. Am. Chem. Soc. 2008,
130, 2170; i) M. Hatano, T. Maki, K. Moriyama, M. Arinobe, K.
Ishihara, J. Am. Chem. Soc. 2008, 130, 16858; j) X. Han, J.
Kwiatkowski, F. Xue, K.-W. Huang, Y. Lu, Angew. Chem. 2009,
121, 7740; Angew. Chem. Int. Ed. 2009, 48, 7604; k) Y. Pan, Y.
Zhao, T. Ma, Y. Yang, H. Liu, Z. Jiang, C.-H. Tan, Chem. Eur. J.
[12] Gong and co-workers reported the chiral phosphoric acid
catalyzed, three-component direct Mannich-type reaction.
Their substrates were simple ketones, not 1,3-dicarbonyl com-
pounds, and the corresponding anti products were obtained. Q.-
X. Guo, H. Liu, C. Guo, S.-W. Luo, Y. Gu, L.-Z. Gong, J. Am.
Chem. Soc. 2007, 129, 3790.
[13] Compound 4 f was obtained in 97% yield (syn/anti = 25:75) with
72% ee (anti) using 2.5 mol% of Ca[1a]2.
[14] Diastereoselective direct Mannich-type reactions of thioesters:
N. Utsumi, S. Kitagaki, C. F. Barbas III, Org. Lett. 2008, 10, 3405.
[15] a) S. Masamune, Y. Hayase, W. Schilling, W. K. Chan, G. S.
Bates, J. Am. Chem. Soc. 1977, 99, 6756; b) G. Zhou, D. Lim,
D. M. Coltart, Org. Lett. 2008, 10, 3809.
[16] J. W. Yang, M. Stadler, B. List, Angew. Chem. 2007, 119, 615;
Angew. Chem. Int. Ed. 2007, 46, 609.
[17] We found that CuI (1.7 equiv) was essential to promote the
reaction. Decarboxylation of allyl b-ketocarboxylates with PdII/
HCO2H has been reported. I. Shimizu, J. Tsuji, J. Am. Chem.
Soc. 1982, 104, 5844.
[18] W. Ye, Z. Jiang, Y. Zhao, S. Li, M. Goh, D. Leow, Y.-T. Soh, C.-H.
Tan, Adv. Synth. Catal. 2007, 349, 2454.
[19] Recent chiral CaII-salt catalyses: a) T. Suzuki, N. Yamagiwa, Y.
Matsuo, S. Sakamoto, K. Yamaguchi, M. Shibasaki, R. Noyori,
Tetrahedron Lett. 2001, 42, 4669; b) G. Kumaraswamy, M. N. V.
Sastry, N. Jena, Tetrahedron Lett. 2001, 42, 8515; c) S. Saito, T.
Tsubogo, S. Kobayashi, J. Am. Chem. Soc. 2007, 129, 5364; d) T.
Tsubogo, S. Saito, K. Seki, Y. Yamashita, S. Kobayashi, J. Am.
Chem. Soc. 2008, 130, 13321.
3826
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3823 –3826