4986
S. Barkallah et al. / Tetrahedron Letters 46 (2005) 4985–4987
parison to chemical shifts of related compounds.1–3
The fixed geometric orientation of the anisotropic
shielding cone of the nitrile of 6a resulted in pro-
nounced chemical shift differences in the NMR reso-
nances of the C-3 and C-5ring methylenes relative to
the parent ketone that also facilitated discrimination
of the regiochemistry associated with each isomer (see
Supplementary data).
The exocyclic double bond of cyanoolefin 6a was hydro-
genated over sodium borohydride-reduced palladium
chloride to afford the cyanomethyl derivatives 7a and
7b in 90% yield as a mixture of cis and trans diastereo-
mers, with the trans isomer comprising ꢀ80% of the
mixture (62% diastereomeric excess). The stereochemical
assignments for anti isomer 7a and syn isomer 7b were
made based on comparison to literature examples of
L-proline ester derivatives substituted at C-4 and by
NMR chemical shift analyses. Depicted in Figure 2 is
the structure of the global energy minimum conforma-
tion of 7a and 7b as calculated using AM1 semi-empir-
ical methods. For compound 7b, the energy-minimized
structure clearly shows the proximal orientation of the
nitrile to the methyl ester. Due to their syn relationship,
not only are these two functional groups in potential van
der Waals contact as evidenced by the presence of an
ensemble of methyl ester resonances (conformers), but
the resonance of the methyl ester is significantly down-
field shifted due to the deshielding effect of nitrile aniso-
tropy cone. In contrast, an anti stereochemical
configuration for 7a was deduced as the methyl ester res-
onance is neither influenced by the nitrile anisotropy
cone, nor appears as multiple conformers on the
NMR timescale.
˚
Scheme 1. Reagents and conditions: (a) MeOH/HCl, 4 A mol. sieves,
D (Soxhlet); (b) Boc2O, NMM, DMF; (c) NMMO, TPAP (5mol %),
˚
CH2Cl2, 4 A mol. sieves; (d) (EtO)2P(O)CH2CN, LiHMDS, THF,
0 °C; (e) chromatographic separation; (f) NaBH4, PdCl2, H2 (45psi),
MeOH.
sodium periodate. However, we were able to achieve
similar yields (90%) of ketone 5 using the catalytic oxi-
dant tetrapropylammonium perruthenate6 (TPAP) with
the stoichiometric co-oxidant N-methyl-morpholine-N-
˚
oxide (NMMO) in the presence of 4 A molecular sieve
dust (Scheme 1). Use of TPAP simplified workup as
the crude product was freed of oxidant by passing the
reaction mixture through a pad of silica gel. In addition
to advantages with workup and purification, the poten-
tial hazard of performing preparative scale reactions
with volatile RuO4 was avoided.
The conversion of ketone 5 into cyanoolefins 6a and 6b
was sluggish using stabilized Wittig conditions7 with the
reagent cyanomethyltriphenylphosphonium bromide
and the bases sodium hydride, lithium diisopropyl
amide, or lithium hexamethydisilazide. However, olefin-
ation under Horner–Wadsworth–Emmons conditions8
(Scheme 1) proceeded smoothly at 0 °C to afford cyano-
olefin 6 in 79% yield as a 2:1 mixture of (E) 6a and (Z)
1
6b isomers (based on H NMR analysis of the crude
product).
Unfortunately, two-dimensional NMR difference-NOE
experiments could not be used to unequivocally deter-
mine the stereochemistry of compounds in this series
as the protons that would normally exhibit observable
NOE enhancements (H-3a, H-4, H-6) coincidentally
resonated at 2.5ppm. However, 1H NMR and 13C
NMR analyses of compounds 5, 6a, and 6b in CDCl3
revealed that all three compounds exist as a 45:55 mix-
tures of, respectively, cis and trans urethane rotamers
at room temperature. Furthermore, the stereochemistry
assignment of compounds 6a and 6b were made based
on coupling constant analysis of the vinyl proton and
the diastereotopic methylene protons on C-3 and com-
Figure 2. AM1 energy-minimized structure of 7a (bottom) and 7b
(top) depicting the relative orientations of the cyanomethyl and methyl
ester functional groups.