Grignard reactions of 4-substituted-2-keto-1,3-dioxanes: Highly diastereoselective additions controlled by a remote alkyl group

Article abstract of DOI:10.1021/ol015914k

(matrix presented) The reactions of Grignard reagents with a representative series of simple cis-2-keto-4-substituted-1,3-dioxanes have been investigated. The stereochemical outcome of these highly diastereoselective additions (dr > 90:10) is consonant wi

Full text of DOI:10.1021/ol015914k

ORGANIC
LETTERS
2001
Vol. 3, No. 12
1865-1868
Grignard Reactions of
4-Substituted-2-keto-1,3-dioxanes:
Highly Diastereoselective Additions
Controlled by a Remote Alkyl Group
William F. Bailey,* David P. Reed, Daniel R. Clark, and Gabriel N. Kapur
Department of Chemistry, UniVersity of Connecticut, Storrs, Connecticut 06269-3060
bailey@uconn.edu
Received March 30, 2001
ABSTRACT
The reactions of Grignard reagents with a representative series of simple cis-2-keto-4-substituted-1,3-dioxanes have been investigated. The
stereochemical outcome of these highly diastereoselective additions (dr > 90:10) is consonant with Cram’s chelate model on the assumption
that RMgX coordinates preferentially with the ring oxygen remote from the C(4) substituent.
Since the formulation of what has come to be known as
“Cram’s chelate rule” 1 describing the stereochemical out-
come of nucleophilic addition to a chiral aldehyde or ketone
bearing an R-polar substituent capable of chelation with an
organometallic reagent,2 considerable effort has been devoted
to development of chiral auxiliaries incorporating such polar
groups to control facial selectivity in the addition of an
organometallic reagent to a carbonyl group.3 Among the
more successful approaches to chelation-controlled asym-
metric induction in addition reactions of organometallic
reagents are those employing various 1,3-oxathianes,4 1,3-
oxazines,5 or 1,3-dioxolanes derived from C2-symmetric
diols6 as the chiral adjuvant. Somewhat surprisingly, there
appear to be no reports of the use of a simple 4-substituted
1,3-dioxane as the control element in organometallic addition
to a carbonyl group. Herein we disclose the results of an
exploratory investigation of the reaction of Grignard reagents
with cis-4-substituted-2-keto-1,3-dioxanes demonstrating that
such addition reactions are highly diastereoselective.
Although relatively few 2-keto-1,3-dioxanes have been
reported in the literature, the substrates used in the present
study (Figure 1) were easily prepared by one of the two
Figure 1. 2-Keto-1,3-dioxanes.
(1) Cram, D. J.; Kopecky, K. R. J. Am. Chem. Soc. 1959, 81, 2748.
(2) (a) Eliel, E. L. In Asymmetric Synthesis; Morrison, J. D., Ed.;
Academic Press: New York, 1983; Vol. 2. (b) Chen, X.; Hortelano, E. R.;
Eliel, E. L.; Frye, S. V. J. Am. Chem. Soc. 1990, 112, 6130. (c) Reetz, M.
T. Acc. Chem. Res. 1993, 26, 462.
(3) For a recent review, see: Mengel, A.; Reiser, O. Chem. ReV. 1999,
99, 1191.
methods depicted in Scheme 1. While transacetalization,7
illustrated for the preparation of 1, is the more direct route,
the reaction invariably affords a mixture of isomeric 2-keto-
10.1021/ol015914k CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/18/2001

Cursory inspection of the data in Table 1 demonstrates
that the additions are quite diastereoselective, particularly
those involving reaction of PhMgBr with the methyl ketones
1, 3, or 5 (Table 1, entries 1-3). Moreover, as detailed below,
the less diastereoselective reactions of MeMgBr with the
phenyl ketones (Table 1, entries 4-6) are easily optimized
to give dr values on the order of ∼95:5 by employing a
solvent other than Et2O as the reaction medium. It should
also be noted that an organolithium reagent is much less
selective in reaction with a 2-keto-1,3-dioxane than is the
corresponding Grignard reagent (Table 1, cf. entries 4 and
7).
Scheme 1
A particularly striking feature of these results is the high
diastereoselectivity engendered by a single methyl substituent
at C(4) in the cis-2-keto-4-methyl-1,3-dioxanes (1 and 2).
Indeed, a methyl group is approximately as effective in
directing facial selectivity in the addition reactions as is a
much larger t-Bu group at this position (Table 1; cf. entries
1 and 3, cf. entries 4 and 6). In this connection, it might be
noted that the isomerically homogeneous 2-keto-4,4-di-
methyl-1,3-dioxanes (such as 3 and 4) are particularly
attractive substrates for facially selective Grignard reactions
since they do not require separation of cis-trans isomers8
before use.
To establish the stereochemical outcome of these facially
selective additions, it is necessary to determine the config-
uration of the stereogenic center created in the reaction. To
this end, (4R)-(+)-cis-2-benzoyl-4-methyl-1,3-dioxane [(+)-
2] was prepared from commercially available (-)-1,3-
butanediol following the two-step route depicted in Scheme
1 for the racemic material (2).
4-substituted-1,3-dioxanes8 from which the more polar cis-
isomer must be isolated by chromatography. The two-step
preparation of cis-2-keto-4-substituted-1,3-dioxanes, illus-
trated in Scheme 1 for the synthesis of 2, is often the more
convenient method since it provides isomerically pure ma-
terial and avoids the tedious chromatographic separation.
Exploratory Grignard reactions were performed at -78
°C by addition of an excess (typically 2 equiv)9 of the reagent
to a solution of the 2-keto-1,3-dioxane (Figure 1) in dry
diethyl ether. The results of these experiments are sum-
marized in Table 1; the configuration of the diastereoisomeric
tertiary alcohol products (designated major and minor in
Table 1) was established as described below.
As illustrated in Scheme 2, treatment of (+)-2 with
MeMgBr under conditions identical to those used for the
racemic substrate (Table 1, entry 4) gave an 85:15 mixture
of diastereoisomeric alcohols. While it might seem a simple
matter to reveal the latent R-hydroxy aldehyde by hydrolysis
of the 1,3-dioxane, this approach was avoided. Not only are
Table 1. Grignard Reactions of
4-Substituted-2-keto-1,3-dioxanesa
Scheme 2
entry ketone
R1
CH3 CH3
CH3 CH3 CH3
R2
R3
RMgX
yield,b %
drc
1
2
3
4
5
6
7
1
3
5
2
4
6
2
H
PhMgBr
84
98
78
86
90
80
78
93:7
93:7
91:9
85:15
84:16
90:10
64:36
CH3 t-Bu
H
H
Ph
Ph
Ph
Ph
CH3
MeMgBr
CH3 CH3
t-Bu
CH3
H
H
MeLi
a The Grignard reagent (2 molar equiv) in Et2O was added at -78 °C to
a solution of the keto-1,3-dioxane in Et2O (0.15 M final concentration),
and the resulting solution was stirred at -78 °C for 1 h before being
quenched with MeOH. b Isolated yield of a mixture of diastereoisomers.
c Diastereomeric ratio determined by GC analysis of TMS ethers.
1866
Org. Lett., Vol. 3, No. 12, 2001

R-hydroxy aldehydes notoriously difficult to manipulate,10
the tertiary, benzylic alcohol function is also prone to
dehydration under the hydrolysis conditions. Although such
considerations would seem to belie the utility of cis-4-
substituted-1,3-dioxanes as chiral auxiliaries, the system
possesses precisely the stereochemistry (an axial C(2)-H
flanked by two oxygen atoms) required for facile insertion
of ozone into the C(2)-H bond under neutral conditions.11
This oxidation, developed by Deslongchamps’ group,11 was
used to convert the addition products to a mixture of esters
from which (-)-atrolactic acid (presumably with er ) dr )
85:15) was isolated in 60% overall yield by saponification
(Scheme 2); a single recrystallization afforded pure (-)-
atrolactic acid.12 Since the levorotatory acid is known to have
the R-configuration,13 the sense of the facial selectivity of
the Grignard additions is established to be that shown in
Table 1.
The stereochemistry of the Grignard addition reactions is
easily rationalized, as shown below, in terms of Cram’s
chelate model1,2 on the reasonable assumption that RMgX
coordinates preferentially to the ring oxygen remote from
the C(4) substituent. In this connection, it might be noted
that similar chelation control exerted by a C(4) group has
been observed in the Grignard-ortho ester reactions of cis-
2-methoxy-4-methyl-1,3-dioxane14 as well as in the highly
regioselective acylative cleavage of 4-substituted-1,3-diox-
anes.15
1,3-dioxanes in Et2O (Table 1) are not necessarily the highest
possible for these addition reactions. As demonstrated by
the data summarized in Table 2, solvent has a pronounced
Table 2. Grignard Reactions of
cis-2-Benzoyl-4-methyl-1,3-dioxane (2)a
entry
RMgX
solvent
yield,b %
drc
1
2
3
4
5
MeMgBr MTBE-Et2O (9:1 by vol)
(i-Pr)2O-Et2O (9:1 by vol)
THF-Et2O (9:1 by vol)
MeMgCl
72
82:18
79:21
95:5
92:8
96:4
23d
90
96
69e
MeMgI
a The Grignard reagent (2 molar equiv) in Et2O was added at -78 °C to
a solution of 2 in the indicated solvent (0.15 M final concentration), and
the resulting solution was stirred at -78 °C for 1 h. b Isolated yield of a
mixture of diastereoisomers. c Diastereomeric ratio determined by GC
analysis of the TMS ethers. d Incomplete reaction after 1.5 h. e Dehydration
of the alcohol products accounts for the lower yield.
effect on the facial selectivity of the addition reaction and
the nature of the halogen in the RMgX reagent has a more
modest influence. Whereas addition of MeMgBr to 2 in Et2O
proceeds with moderate selectivity at -78 °C (viz. 85:15;
Table 1, entry 4), the selectivity is improved significantly
(dr ) 95:5) when an ethereal solution of the Grignard is
added to a solution of 2 in THF (Table 2, entry 3).
Conversely, selectivity is eroded when reactions are run in
less strongly coordinating solvents such as MTBE and (i-
Pr)2O (Table 2, entries 1 and 2). The etiology of these solvent
effects may well be complex: a more poorly coordinating
solvent should favor formation of the putative Cram chelate
between RMgX and the dioxane2,3 but solvent may also affect
the Schlenk equilibrium and the degree of association of the
Grignard reagent.
As noted above, the diastereoselectivities observed for
reaction of Grignard reagents with cis-2-keto-4-substituted-
(4) (a) Eliel, E. L.; Koskimies, J. K.; Lohri, B. J. Am. Chem. Soc. 1978,
100, 1614. (b) Eliel, E. L.; Koskimies, J. K.; Lohri, B.; Frazee, W. J.; Morris-
Natschke, S.; Lynch, J. E.; Soai, K. In Asymmetric Reactions and Processes
in Chemistry; Eliel, E. L., Otsuka, E., Eds.; ACS Symposium Series 185,
American Chemical Society: Washington, DC, 1982; pp 37-53. (c) Frye,
S. V.; Eliel, E. L. J. Am. Chem. Soc. 1988, 110, 484, and references
therein.
(5) (a) He, X.-C.; Eliel, E. L. Tetrahedron 1987, 43, 4979. (b) Eliel, E.
L.; He, X.-C. J. Org. Chem. 1990, 55, 2114. (c) Ko, K.-Y.; Park, J.-Y.
Tetrahedron Lett. 1997, 38, 407.
In summary, the results of this preliminary investigation
demonstrate for the first time that simple 4-substituted 1,3-
dioxanes are effective as chiral auxiliaries for control of facial
selectivity in the addition of a Grignard reagent to a ketone.
The highly diastereoselective additions appear to be the result
of preferential formation of a Cram-type chelate between
the RMgX and the ring oxygen remote from the C(4)sub-
stituent. The ready availability of a variety of 1,3-diols for
the synthesis of enantiopure cis-2-keto-4-substituted-1,3-
dioxanes, coupled with the ease with which the adjuvant may
be removed by Deslongchamps oxidation from the tertiary
(6) (a) Tamura, Y.; Ko, T.; Kondo, H.; Annoura, H.; Fuji, M.; Takeuchi,
R.; Fujioka, H. Tetrahedron Lett. 1986, 27, 2117. (b) Heitz, M. P.; Gellibert,
F.; Mioskowski, C. Tetrahedron Lett. 1986, 27, 3859. (c) Thiam, M.;
Chastrette, F. Bull. Chim. Soc. Fr. 1992, 192, 161. (d) Akhoon, K. M.;
Myles, D. C. J. Org. Chem. 1997, 62, 6041.
(7) Cho, B. T.; Chun, Y. S. J. Chem. Soc., Perkin Trans. 1 1999, 2095.
(8) Although the conformational energy of a 2-acetyl or 2-benzoyl group
in the 1,3-dioxane system has not been determined, the low conformational
free energy of a 2-CO2Et group (-∆G° ) 0.92 kcal/mol) suggests that the
isomeric composition of product mixtures prepared by transacetalization
may reflect the actual equilibrium ratio, see: Tschierske, C.; Ko¨hler, H.;
Zaschke, H.; Kleinpeter, E. Tetrahedron 1989, 45, 6987.
(9) It should be noted that the use of 1 equiv of RMgX gives virtually
the same product composition (dr); however, yields were somewhat lower
in reactions employing only 1 equiv of reagent.
(10) See, for example: Lynch, J. E.; Eliel, E. L. J. Am. Chem. Soc. 1984,
106, 2943.
(13) McKenzie, A.; Clough, G. W. J. Chem. Soc. 1910, 97, 1016.
(14) Bailey, W. F.; Croteau, A. A.; Rivera, A. D. Tetrahedron Lett. 1997,
38, 4047.
(15) (a) Bailey, W. F.; Rivera, A. D. J. Org. Chem. 1984, 49, 4958. (b)
Bailey, W. F.; Zarcone, L. M. J.; Rivera, A. D. J. Org. Chem. 1995, 60,
2532.
(11) (a) Deslongchamps, P.; Moreau, C. Can. J. Chem. 1971, 49, 2565.
(b) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry;
Pergamon: New York, 1983; pp 41-47.
(12) Mp 112-114 °C (lit.13 mp 114-116 °C); [R]24D ) -36.3 [c 3.43,
EtOH] (lit.13 [R]13.8 ) -37.7 [c 3.4, EtOH].
D
Org. Lett., Vol. 3, No. 12, 2001
1867

carbinol center created in the Grignard reaction (Scheme 2),
suggests that this system may be of significant practical value
for the asymmetric synthesis of enantiomerically pure
R-hydroxy acids. We are actively investigating this possibil-
ity.
Research Experience for Undergraduates (REU) Program
grant to G.N.K.
Supporting Information Available: Detailed experi-
mental procedures. This material is available free of charge
via the Internet at http://pubs.acs.org.
Acknowledgment. This work was supported by Procter
& Gamble Pharmaceuticals, Mason, OH, and by an NSF
OL015914K
1868
Org. Lett., Vol. 3, No. 12, 2001