Boron-mediated double aldol reaction of carboxylic esters [10]

Full text of DOI:10.1021/ja9914184

7168
J. Am. Chem. Soc. 1999, 121, 7168-7169
Boron-Mediated Double Aldol Reaction of
Carboxylic Esters
Atsushi Abiko,*,† Ji-Feng Liu,‡ Dana C. Buske,‡
Satoshi Moriyama,† and Satoru Masamune*,‡
Figure 1. Carboxylic ester reagents.
Venture Laboratory, Kyoto Institute of Technology
Matsugasaki, Sakyo-ku
Kyoto 606-8585, Japan
Scheme 1
Department of Chemistry
Massachusetts Institute of Technology
Cambridge, Massachusetts, 02139
ReceiVed April 30, 1999
Scheme 2a
Recently, we have found that carboxylic esters undergo boron-
mediated aldol reactions.1 Indeed, chiral propionate esters 1 and
2 achieve high stereoselectivities in the construction of anti- and
syn-â-hydroxy-R-methyl carbonyl systems.2,3 An extension of this
investigation would naturally include an acetate aldol reaction
utilizing a chiral ester. In our investigation, we have unexpectedly
found that (i) an aldol reaction with an acetate ester, such as ester
3 (Figure 1), proceeds to provide bis-aldol products via a double
aldol reaction and (ii) these products serve as the starting materials
for the synthesis of chiral triols of C3 symmetry, compounds which
may prove highly valuable in asymmetric synthesis. Herein we
summarize these unprecedented findings.
From our previous work with the propionate esters 1 and 2,
we had empirically determined that the reaction conditions
necessary to provide mono-aldol product in high yield are the
use of 2 equiv of dialkylboron triflate and 2.4 equiv of trialkyl-
amine. However, the reaction of ester 3 under the same reaction
conditions proceeded in an unusual manner to afford a substantial
amount of unexpected bis-aldol products 5, along with the
expected mono-aldol product 4 (Scheme 1).
Efforts to optimize a variety of reaction parameters led to the
establishment of conditions for almost exclusive formation of 5
(4:5 e 5:95). Treatment of 3, under the conditions specified in
Scheme 2 (i and ii), provided bis-aldol products in over 95% yield
with a diastereomer ratio of 5a-A:5a-B:5a-C ) 90:8:2. The fourth
possible isomer was not detected. Bis-aldol product 5a-A, obtained
as a pure crystalline compound, has been fully characterized by
NMR as well as X-ray crystallography.4 The structure of 5a-B,
isolated as a mixture with 5a-A, was proved by reduction of the
mixture of acetonides 6a-A and 6a-B (91:9) to produce a single
alcohol 7a-A with 82% ee. The third isomer was determined to
be 5a-C by NMR analysis of the corresponding acetonide 6a-C
(H5: δ 2.56, t, J ) 2.7 Hz).
a Conditions: (i) c-Hex2BOTf (2.5 equiv), Et3N (3.0 equiv), -78 °C,
15 min; (ii) i-PrCHO (3.0 equiv), -78 °C, 10 min, then room temperature,
30 min, 95%. Note for nomenclature: 1, 2, 3, ... refers to structures;
a, b, c, ... refers to aldehydes; A, B, C, ... refers to stereoisomers.
Scheme 3a
a Conditions: (i) c-Hex2BOTf (2.5 equiv), Et3N (3 equiv), -78 °C,
15 min; (ii) i-PrCHO (1.0 equiv) -78-0 °C, 10 min; (iii) PhCHO (1.5
equiv), 0 °C, room temperature, 30 min, 63%; (iv) Me2C(OMe)2, PPTS,
100%.
The double aldol reaction should proceed in a stepwise manner
(Scheme 2) since mono-aldol 4a-A was the major product when
the reaction was maintained at -78 °C (4a-A:4a-B ) 88:12).
The auxiliary-controlled first aldol reaction produced two dia-
stereomeric mono-aldolates whose stereochemistry corresponds
to 4a-A and 4a-B at position 3. Since 4a-A and 5a-A (and also
4a-B and 5a-B) have the same configuration at position 3, each
pair of compounds is derived from the intermediate mono-aldolate
with the same configuration.5 However, the mono-aldolates’
structures (not the same as the dicyclohexylboron derivative of
4a, see 14 in Scheme 5) are still unknown.
To elucidate the pathway of the second aldol reaction, the
enolate of 3 was treated with two different aldehydes in a stepwise
manner (Scheme 3). The stereochemistry of positions 2 and 1′ of
the major product 8, determined as its acetonide 9, was proved
to be the same as that of anti-aldol product 10 (obtained from an
E(O)-enolate of propionate 1) at positions 2 and 3. Thus, 8 (and
also 5a-A) must have been formed from an E(O)-enolate of the
mono-aldolate, rather than a Z(O)-enolate.5
The utility of the bis-aldol products 5 has been demonstrated
by the synthesis of chiral triols of C3 symmetry (Scheme 4).
Compounds of C3 symmetry have attracted much attention as
valuable ligands for asymmetric catalysts.6 Among the class of
† Kyoto Institute of Technology.
‡ Massachusetts Institute of Technology.
(1) Abiko, A.; Liu, J.-F.; Masamune, S. J. Org. Chem. 1996, 61, 2590.
(2) Abiko, A.; Liu, J.-F.; Masamune, S. J. Am. Chem. Soc. 1997, 119, 2586.
(3) Liu, J.-F.; Abiko, A.; Pei, Z.; Buske, D. C.; Masamune, S. Tetrahedron
Lett. 1998, 39, 1873.
(4) X-ray data in Supporting Information.
(5) 5a-C is possibly produced from both diastereomers of the mono-aldolate.
For a detailed discussion of the stereochemistry of the double aldol reaction
and information to support a cyclic transition state involving an E(O)-enolate,
see the Supporting Information.
(6) (a) Moberg, C. Angew. Chem., Int. Ed. 1998, 37, 248. (b) Nugent, W.
A. J. Am. Chem. Soc. 1998, 120, 7139 and references therein.
10.1021/ja9914184 CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/14/1999

Communications to the Editor
J. Am. Chem. Soc., Vol. 121, No. 30, 1999 7169
Scheme 4a
Scheme 5
a a: R ) i-Pr; b: R ) Et; c: R ) Ph. Conditions: (i) LiAlH4, THF;
(ii) PDC, CH2Cl2; (iii) RMgX, THF; (iv) TFA, MeOH. a: (i) 99%, (ii)
95%, (iii) 65%, 97:3 (7a-A, 35%), (iv) 96%; b: (i) 95%, (ii) 95%, (iii)
98%, >99:1, (iv) 93%; c: (i) 94%, (ii) 95%, (iii) 92%, 95:5, (iv) 99%.
implies that 14, the dicyclohexylboron derivative of 4′ (and also
that of 4a), is not a precursor of the double aldol reaction and
thus is clearly distinct from the mono-aldolate species in Scheme
2.12
It is now a distinct possibility that two kinds of boron species
are involved in the boron-mediated aldol reaction of carbonyl
compounds. The first boron species exists with ketones and
thioesters and leads to the single aldol reaction, since boron can
only chelate with the carbonyl group to form a 1:1 complex. The
second boron species13 exists with acetate esters, in which
coordination to the extra oxygen atom forms a 2:1 complex which
leads to the double aldol reaction. We are currently investigating
the structure of this boron-ester complex.14
In conclusion, we have discovered the unprecedented double
aldol reaction of acetate esters and have identified unique features
of the enolization of carboxylic esters with boron triflate and
amine. The bis-aldol products obtained were utilized in the
synthesis of chiral C3 symmetric triols. Further studies on the
application of these C3 triols as chiral ligands are in progress in
our laboratories.
chiral C3 symmetric ligands, chiral triols have not been studied
adequately, presumably because of the limited availability of these
materials.7 Starting with the two preexisting stereocenters (posi-
tions 3 and 1′) in 5-A, introduction of a third carbinol at position
1 produces the chiral C3 symmetric triols (13-As). Thus, alcohol
7-A8 was oxidized to aldehyde 11-A by PDC.9 Stereoselective
introduction of the third substituent via a Grignard reaction gave
alcohol 12-A, with >95:5 selectivity.10 Acidic hydrolysis of the
acetonide group in 12-A provided the chiral C3 symmetric triols
13-A. The existence of C3 symmetry in 13-A was easily
1
ascertained by H and 13C NMR analyses.
Finally, we should emphasize unique features of the boron-
mediated aldol reaction of carboxylic esters and the unusual
behavior of the boron-containing intermediate.
(1) While double aldol reactions proceed with an acetate ester
as the enolate source, this double aldol reaction does not occur
with other carbonyl compounds such as methyl ketones, acetate
thioesters, and propionate esters, exemplified by 1 and 2.
(2) When acetate ester 3 was enolized with less than 2 equiv
of dialkylboron triflate, both bis-aldol products 5 and recovered
starting material were obtained. Therefore, the formation of bis-
aldol products is not solely due to the ratio of the excess
dialkylboron triflate to ester.
(3) When the enolate of ent-3 (enantiomer of 3, 1 equiv),
prepared with c-Hex2BOTf (2 equiv) and Et3N (2.5 equiv), was
mixed with S-phenyl thioacetate (1 equiv), no enolization of
S-phenyl thioacetate occurred. This result is in contrast to the
same experiment with acetophenone in place of ent-3, in which
the aldol product of the thioester was obtained in ∼50% yield.
Thus, the ester ent-3 somehow renders the second equivalent of
c-Hex2BOTf unreactive. Indeed, 2 equiv of boron triflate are
necessary for the ester aldol reaction, since the extent of
enolization of ester ent-3 with 1.0 and 1.5 equiv of boron triflate
was 47 and 73%, respectively. These results demonstrate that the
stoichiometry of the enolization of the carboxylic ester with boron
triflate is 1:2. The “second equivalent of boron triflate” is neither
free boron triflate nor an equilibrium form of boron triflate. These
findings were totally unexpected!
Acknowledgment. A.A. was supported by a grant from the Ministry
of Education, Science, Sports and Culture of Japan (Grant-in-Aid
10640578). S.M. was supported by a grant from NIH (CA48175). D.C.B.
was supported in part by a National Cancer Institute training grant
(CA09112) and a fellowship from Boehringer-Ingelheim Pharmaceuticals,
Inc. We are grateful to Dr. W. M. Davis for assistance in obtaining the
X-ray structure of 5a-A.
Supporting Information Available: Further discussion regarding the
transition state and stereoselectivity of the double aldol reaction,
experimental procedures, and spectral data for all compounds, and X-ray
structural information for 5a-A (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA9914184
(11) Compare with: Luke, G. P.; Morris, J. J. Org. Chem. 1995, 60, 3013.
(12) Only this mono-aldolate species is responsible for the double aldol
reaction. This species seems to be rather labile and readily decomposed, thus
allowing for the formation of mono-aldol products. Indeed, when the aldehyde
was added rapidly, a significant amount of the mono-aldol product 4 was
formed.
(13) The oligomeric form of boron triflate, with a dimeric diboradioxetane
structure, may be responsible for the enolization of carboxylic esters. See,
Ooi, T.; Uraguchi, D.; Maruoka, K. Tetrahedron Lett. 1998, 39, 8105.
(4) Subjection of an equimolar mixture of 3 and 4′ (ds ) 75:
25) under the optimized double aldol reaction conditions afforded
5as in the same ratio as shown in Scheme 2, while 4′ was
recovered unchanged, quantitatively (Scheme 5).11 This result
(7) Lu¨tjens, H.; Wahl, G.; Mo¨ller, F.; Knochel, P.; Sundermeyer, J.
Organometallics 1997, 16, 5869.
(8) The reaction with EtCHO produced 5b-A:5b-B:5b-C ) 89:3:8 in 95%
yield, from which 5b-A of 94% de was obtained by chromatography. In the
subsequent step, diastereomerically pure acetonide 6b-A was isolated by
recrystallization. The reaction with PhCHO produced 5c-A:5c-B:5c-C ) 85:
1:14 in 97% yield, from which 5c-A was isolated by recrystallization. For
characterization and determination of the absolute stereochemistry of 5b-A
and 5c-A, see the Supporting Information.
(14) It is not clear that the boron species described in ref 13 is incorporated
in the aldol reaction of propionate esters in general. The failure of the double
aldol reaction for propionate ester 1 and 2 very likely be attributed to the
steric reason.
(9) For oxidation of 7c-A, pyridinium chlorochromate was used instead of
pyridinium dichromate.
(10) With i-PrMgCl, reduction product 7a-A was obtained in 35% yield.