of our reaction sequence for the synthesis of polypropionate
stereopentads.
Table 1. Mukaiyama Reactions of 10 and 12a
In the case of the 2,3-anti-3,4-anti bis(benzyloxy)aldehyde
9, this approach was successfully demonstrated.10 Central
to the present study are aldehydes 10 and 12.11 Indeed, their
relative 2,3-syn stereochemistry could pose new problems
in the Cram-chelated Mukaiyama reaction. As illustrated in
Figure 1, the Lewis acid chelated intermediates (whether in
Figure 1. Possible Lewis acid chelated transition states in boat
(A) or half-chair (B) conformations.
boat or half-chair conformation) may not be thermodynami-
cally favored because of unfavorable steric interactions
between the 2,3-syn substituents. This may lead to an erosion
of stereocontrol by allowing competing reaction pathways
involving the less hindered monodentate species. Indeed, our
first experiments aimed at probing the Cram-chelate pathway
using the 2,3-syn-3,4-anti aldehyde 10 and the bromo-
enoxysilane 2 (4:1 E:Z mixture) with bidentate Lewis acids
(MgBr2‚OEt2, Et2BOTf, SnCl4, Me2AlCl, etc.) were dis-
appointing. Even TiCl4, which proved to be effective in the
Mukaiyama reaction involving aldehyde 9,10b turned out to
be ineffective in this case (Table 1, entry 1). Extensive
decomposition of the aldehyde and cleavage of the primary
benzyl ether were noted.
a Aldehyde 10 or 12 (0.1 M) in CH2Cl2 was precomplexed at -78 °C
with the appropriate Lewis acid followed by addition of bromoenoxysilane
2 (1.3 equiv). b Ratios were determined by 1H NMR spectroscopy. c Yields
of isolated products. d Degradation of the aldehyde was observed. e Starting
material was recovered. f Aldehyde 10 or 12 (0.1 M) in CH2Cl2 was treated
at -78 °C with BF3‚OEt2 and then with bromoenoxysilane 2 (1.3 equiv).
interesting was our observation that this drawback could be
overcome by increasing the Lewis acid:aldehyde stoichiom-
etry, as indicated by the impressive 3,4-anti stereoselectivity
favoring compound 14 (entries 4 and 5). The 3,4-syn product
13 was observed with excellent diastereomeric ratio using
the monodentate Lewis acid BF3‚OEt2 (entry 6). Similar
results were achieved with aldehyde 12. The Cram-chelate
pathway was favored with 2.5 equiv of (i-PrO)TiCl3,
exclusive formation of product 16 was observed (entry 7).
Conversely, the Mukaiyama adduct 15 was the only observed
product when BF3‚OEt2 was used, indicative of a reaction
under Felkin-Anh control (entry 8). Again, TiCl4 was
ineffective in this case.
The necessity of having to use 2.5 equiv of (i-PrO)TiCl3
to achieve high stereocontrol may suggest the existence of
reactive complexes different from the simple chelates il-
lustrated in Figure 1. Our preliminary NMR studies are
consistent with the presence of an ate complex in solution.15
Further investigations will be required to fully characterize
the structure of the reacting complex.16
Obviously, changing the primary hydroxy-protecting group
would have been a solution to circumvent the latter reaction.
Instead, we decided to evaluate other Lewis acids. Since the
cleavage of a benzyl ether requires activation of the benzylic
oxygen by the Lewis acid, we focused on lowering the
titanium Lewis acidity. This was done by considering
12
(i-PrO)TiCl3 and (i-PrO)2TiCl2,13 which have been found
to be useful in aldol reactions.14 As seen in entry 2, no
reaction (Mukaiyama aldol nor benzyl ether cleavage) was
noted when (i-PrO)2TiCl2 was used. Interestingly, the use
of (i-PrO)TiCl3 provided our first positive result, the aldol
products 13 and 14 being obtained (Table 1, entry 3) in good
yield albeit with modest Cram-chelate selectivity. Even more
The first step of our planned consecutive process having
been completed, we then turned our attention to the free-
(10) In this case, one should have expected 1,2- and 1,3-inductions to
oppose each other, thus potentially eroding stereocontrol in the Cram-chelate
Mukaiyama reaction. However, we and Evans et al. showed 1,2-induction
to be dominant when using hindered enoxysilanes: (a) Evans, D. A.; Dart,
M. J.; Duffy, J. L.; Yang, M. C. J. Am. Chem. Soc. 1996, 118, 4322. (b)
Mochirian, P.; Cardinal-David, B.; Gue´rin, B.; Pre´vost, M.; Guindon, Y.
Tetrahedron Lett. 2002, 43, 7067.
(11) For preparation of aldehydes 10 and 12, see Supporting Information.
(12) Solsona, J. G.; Romea, P. D.; Urpy´, F.; Vilarrasa, J. Org. Lett. 2003,
5, 519.
(13) Mikami, K.; Terada, M.; Nakai, T. J. Am. Chem. Soc. 1990, 112,
3949.
(14) Selected examples: (a) Ishimura, K.; Monda, K.; Yamamoto, Y.;
Akiba, K. Tetrahedron 1998, 54, 727. (b) See ref 12.
(15) Evans et al. noticed a similar preference for the Cram-chelate
controlled aldol reaction as a result of an increase of the Lewis acid (Me2-
AlCl) stoichiometry. An ate complex was invoked; see ref 1e.
(16) Suggestions on the structure of the complexes could be derived from
the work of Gau, who suggested the prevalence of six-coordinate complexes
in which the relative bonding ability of various ligands can be established
as i-PrO- > Cl- > THF > Et2O > PhCHO. (a) Lee, C.-H.; Kuo, C.-C.;
Shao, M.-Y.; Gau, H.-M. Inorg. Chim. Acta 1999, 285, 254. (b) Wu, Y.-
T.; Ho, Y.-C.; Lin, C.-C.; Gau, H.-M. Inorg. Chem. 1996, 35, 5948. (c)
Gau, H.-M.; Lee, C.-S.; Lin, C.-C.; Jiang, M.-K.; Ho, Y.-C.; Kuo, C.-N. J.
Am. Chem. Soc. 1996, 118, 2936.
Org. Lett., Vol. 6, No. 15, 2004
2601