responsible for this triaxial conformational preference in the
solid state, since the diol forms asymmetric dimers held
together by H-bonding between the two oxygens of one
molecule and the two hydroxyl protons of the other molecule.
However, 1H NMR of both 6 and its diacetate show that the
proton on the ring hydroxyl-bearing carbon (C2) does not
have any diaxial coupling. Thus, the same conformation with
three groups disposed axially is adopted both in solution and
in the solid state.
With reduction routes available to convert different keto
alcohols to chiral 1,3-diols and monoprotected derivatives
whose absolute structures are now defined, this chemistry
should allow ready provision of further diol analogues as
well as to 1,3-diol derivatives (e.g., phosphinites) as novel
chiral ligand candidates. We have adapted this chemistry to
the synthesis of diol libraries of this type showing that a
range of funtionalized aldols can be prepared, reduced, and
screened as ligands in parallel to identify new enantioselec-
tive chiral ligands.6 There is one related example of a
menthone-derived 1,3-diol of this sort, an aluminum complex
of which catalyzes a Diels-Alder reaction with up to 82%
ee.7
derived system without the side chain attached.8 This
illustrates that this chemistry provides a route into new
families of more functionalized chiral ligand candidates.
We also wished to differentiate the two functional groups
of the keto alcohols to demonstrate entry to various other
new bifunctional and also trifunctional ligand types. Our
route for introduction of nitrogen was via preparation of the
azide derivative of (+)-isomenthone-derived aldol products.
Attempts to introduce the azido group into 4b, 4c, or 2/3
(R1 ) Ph) through mesylation, tosylation, or triflation were
unsuccessful due to competing side reactions. The azide was
successfully introduced through a Mitsunobu-type reaction,
using 2,4,4,6-tetrabromo-2,5-cyclohexadienone, triphenylphos-
pine, and the diazobis(pyridine) zinc.9 Thus, keto alcohol 2
[R1 ) Ph] was converted into azido ketone 10 in good yield
(Scheme 4). Replacement of OH by the azido group excludes
Scheme 4
Our second target was the addition of nonhydride nucleo-
philes to the ring carbonyl to introduce a new C-C bond
and adding new functionality along with a new hydroxylated
quaternary center.
Two illustrative examples are presented here of reactions
of protected aldol 4b with Grignard and organolithium
reagent. Allylmagnesium bromide reacts with 4b to give a
single diastereomeric product 8, while 2-lithiopyridine
reacted with 4b also giving, after deprotection, a single
diastereomeric product 9 (Scheme 3). In both cases, the
effects of H-bonding by the hydroxyl in driving conforma-
tional preference in the crystal structure (for the axial side
chain). However, an X-ray structure of azide 10 shows the
same ring conformation as in all keto alcohols, so that
H-bonding from the aldol hydroxyl does not appear to be
determining in conformational preferences in the solid state
for (+)-isomenthone derivatives. The same azidation chem-
istry was successfully applied to 3 [R1 ) CHdCHPh].
This entry to 1,3-azidoketones provides a valuable hub,
since both the azide and carbonyl groups could be elaborated
to other functionality. There are recent examples of terpenoid-
derived amino alcohols acting as effective catalysts.10 In this
context, we reduced the azide to an amino ketone by
hydrogenation (unoptimized) to provide 11, and this was then
further elaborated to the amido phosphine 12 (Scheme 4).
The combination of P, N, O functionality, frequently
including an amide, is heterofunctionality common to several
recently reported chiral ligands catalyzing various pro-
cesses.11
Scheme 3
nucleophile has added with the same facial sense as with
hydride additions. This strongly suggests a general preference
for any additions to this carbonyl of this substrate type. This
predictivity is valuable for synthesis of further defined
diastereomeric targets from various aldol intermediates. The
pyridyl system 9 has an interesting trifunctional ligand
architecture. The most related prior example is a menthyl-
(8) Herrmann, W. A.; Haider, J. J.; Fridgen, J.; Lobmaier, G. M.; Spiegler,
M. J. Organomet. Chem. 2000, 603, 69-79.
(9) (a) Saito, A.; Saito, K.; Tanaka, A.; Oritani, T. Tetrahedron Lett.
1997, 38, 3955. (b) Viaud, M. C.; Rollin, P. Synthesis, 1990, 130.
(10) (a) Kitamura, M.; Suga, S.; Oka, H.; Noyori, R. J. Am. Chem. Soc.
1998, 120, 9800-9809. (b) Panev, S.; Linden, A.; Dimitrov, V. Tetrahedron:
Asymmetry 2001, 12, 1313-1321. (c) Dimitrov, V.; Dobrikov, G.; Genov,
M. Tetrahedron: Asymmetry 2001, 12, 1323-1329.
(6) Gardiner, J. M.; Crewe, P. D.; Smith, G. E.; Veal, K. T. Unpublished
results, 2002. Ligands identified with up to 90% ee.
(7) Naraku, G.; Hori, K.; Ito, Y. N.; Katsuki., T. Tetrahedron Lett. 1997,
38, 8231-8232.
Org. Lett., Vol. 5, No. 4, 2003
469