7
96
Can. J. Chem. Vol. 76, 1998
Scheme 1.
Scheme 2.
compound known as seudenol (a sex pheromone of the
Douglas fire beetle), was chosen since the stabilization of 5-Li
and 6-Li by their methyl groups was expected to make them
thermodynamically more stable than their isomers, 7-Li and
showed that the elution order of the enantiomers of 6 indeed
is the reverse of that of the enantiomers of 5. Thus (S)-5
isomerize exclusively to (S)-6.
8
-Li (Scheme 3), respectively. As a consequence, further rear-
Besides the stereospecificity, i.e., the configuration of the
stereogenic carbon center remains intact in the rearrangement,
the above results do not give any details of the mechanism of
the 1,3-proton transfer reaction: Is the proton transfer intermo-
lecular or intramolecular? Does it take place suprafacially or
antarafacially? Is the rearrangement selective with respect to
any one of the two protons in the 4-position of 5? Is the proton
delivered to the carbon syn or anti to the oxygen on the ring?
Experiments are in progress to also answer these questions.
The first problem has been solved in the following way.
To investigate whether the 1,3-proton transfer takes place
intra- or intermolecularly the following experiment was car-
ried out. Compound 5-Li was isomerized in DMTHF into 6-Li
in the presence of 2-Li and the N-deuterated amine, 2-ND
rangement to 7-Li and 8-Li is expected to be relatively slow,
making it possible to study the reversible rearrangement of
5-Li to 6-Li without significant interference of further rear-
rangements.
Racemic 5 is commercially available and can be enzymati-
cally resolved with high selectivity. This procedure was util-
ized to make a mixture of enantiomers with (S)-5 in 76% ee as
measured by chiral gas chromatography (14).
The lithium alkoxide of the enantiomerically enriched al-
lylic alcohol, (S)-5 (76% ee) was reacted with 2-Li in DMTHF
at 30.0°C. Quenching of the reaction and extraction of the
mixture of products and remaining reactants followed by chiral
gas chromatographic analysis showed that one of the enan-
tiomers of 6 had been formed in 76% ee. The ee was studied
as a function of time during the approach of the equilibrium.
The equilibrium mixture contained 66% 5 and 34% 6. The ee
of 5 remained constant, within experimental error, at 76% dur-
ing the whole reaction. That is, no significant racemization of
either 5 nor 6 took place during the rearrangement. Obviously
the methyl substitution had the predicted result of avoiding
racemizations. Thus it is concluded that the rearrangement is
(Scheme 4). After ca. 10 half-lives the reaction was quenched
and the mixture of the allylic alcohol 5 and product homoal-
1
2
lylic alcohol 6 was isolated and analyzed by H and H NMR.
Only traces of deuterium were detected in the product and the
starting material, respectively. The absence of incorporation
of deuterium in the starting material and product during the
reaction indicates that the prototropic rearrangement is close
to 100% intramolecular (Scheme 4).
100% stereospecific (15). The results indicate that the racemi-
In our system the 1,3-proton transfer might take place
within a 1:1 complex of lithium alkoxide, e.g., 5-Li and the
lithium amide 2-Li. Recently we carried out a computational
investigation of a model system, i.e., the lithium amide cata-
lyzed rearrangement of lithium allylic alkoxide to lithium ho-
moallylic alkoxide. Pathways have been calculated using
semiempirical (PM3), ab initio (HF, MP2), and DFT (B3LYP)
methods. The rearrangement takes place in heterodimer com-
plexes in which the two lithiums play different roles in differ-
ent pathways via intermediates. In the calculated
proton-transfer transition states the proton is about half trans-
ferred between carbon and nitrogen (20).
zations observed upon the rearrangement of 3 into 4 are caused
by consecutive 1,3-proton transfer reactions.
Assignment of the absolute configuration of the enantiomer
of 6 formed in enantiomeric excess from (S)-5 was investi-
gated in the following way. If (S)-6 is the product formed from
(S)-5, then the chiral chromatogram of the mixture of the en-
antiomers of 5 and 6 (Fig. 1) shows that the elution order of
the enantiomers of 5 and 6 must be unexpectedly different.
Therefore, the absolute configuration of the enantiomer of 6
formed in enantiomeric excess was established in the follow-
ing way. Racemic seudenol, 5, was rearranged to the homoal-
lylic alcohol 6 by amide 2-Li in DMTHF. Chiral
chromatographic analysis showed to our surprise that com-
pound 5 remained racemic during the rearrangement, i.e., both
enantiomers of 5 react with the chiral amide 2-Li with equal
rate constants. As a consequence, the product 6 was also found,
within experimental error, to be racemic.
The isolated mixture of racemic 5 and 6 was used in enzy-
matic resolution of 6 employing the same enzyme that was
used to resolve 5 to yield (S)-5 in enantiomeric excess. The
enantiomer of compound 6 remaining in excess was by anal-
ogy assumed to be (S)-6 (16–19). Chiral chromatography
Experimental section
General
All glassware used for synthesis was, when needed, dried over-
night in a 120°C oven. Glassware and syringes used for
isomerization reactions were dried at 50°C in a vacuum oven
before transfer into a glove box (Mecaplex GB 80 equipped
with a gas purification system that removes oxygen and mois-
ture) containing a nitrogen atmosphere. Typical moisture
©
1998 NRC Canada
Arvidsson, Per I.
