LETTER
Stereoselective Synthesis of Substituted 1,3-Dienes
2773
OH
O
R2
OH
OH
3
R1
10b R1 = Me, R2 = H
TMS O
TMS
O
R2
OH
R2
OH
TBAF
O
3
O
THF
R1
O
on standing
100%
3
R1
80–90%
R2 OH
R1
11
9
3
a) R1 = R2 = H
b) R1 = Me, R2 = H
c) R1 = H, R2 = Me
12a R1 = R2 = H
12c R1 = H, R2 = Me
Scheme 2
It is possible that the epoxide opening under acidic condi- With the precursor 19 in hand, our optimised conditions
tions competes to generate a tertiary carbocation as well for the Ireland–Claisen rearrangement were tested.21 To
as the allylic carbocation species.
our delight the isolated product 20 had undergone a rear-
rangement/silicon-mediated fragmentation cascade and
existed as a single diastereomer, resulting from a chair
transition state in the rearrangement step. Although the
acid 20 was isolated in modest yield, we believe that the
cascade reaction observed will be of significant synthetic
use. Aziridine opening by chloride anion after rearrange-
ment has depressed the yield of the final product and work
is now in progress to prevent this undesired reaction.
The synthesis of the E,Z-dienes 12 was also exploited as
the lactonised product 11 was obtained upon standing the
hydroxy acids 9 at low temperature. The stereodefined
E,Z-dienes 12 were isolated in excellent yield by fragmen-
tation of the lactones 11 with tetrabutylammonium fluo-
ride. The basic character of TBAF seemed to be
responsible for the isomerisation. In the case of the tri-
substituted alkene 11b, the isomerisation did not occur,
therefore the E,E-diene 10b was obtained. Exposure of the In conclusion, we have extended our existing methodolo-
lactones 11 to acid afforded a mixture of dienes 10 and 12 gy to synthesise substituted dienes. The reaction sequence
(Scheme 2).
is efficient and simple enough to prepare the stereodefined
dienes, E,E or E,Z and its application towards natural
products will be reported in due course.
To conclude our studies on the generalised stereoselective
synthesis of substituted dienes we had to address an un-
answered question concerning the poor stereochemical
outcome of the Ireland–Claisen rearrangement. It was
reasoned that the tight chelation between the silicon of the
silyl enol ether and the epoxide oxygen would inhibit the
chair-like transition state as some non-chair transition
state models have been suggested in certain bicyclic sys-
tems.12 In order to test this plausible explanation, it was
decided to prepare a three-membered ring model system,
General Procedure for the Ireland–Claisen Rearrangement.
To a stirred solution of the propionate ester (6.4 mmol) in THF (50
mL) at –78 °C was added KHMDS (31.9 mmol in toluene). The yel-
low solution was stirred at –78 °C for 20 min and then a mixture of
TMSCl–Et3N (64.0 mmol/31.9 mmol) in THF (10 mL) was added.
After stirring for 10 min at –78 °C the reaction mixture was allowed
to warm up to r.t. and was stirred for a further 2 h. The reaction mix-
whose lone pair electrons were no longer available for ture was quenched with sat. NH4Cl solution and extracted with Et2O
(3 × 25 mL). The combined organic extract was dried (MgSO4), fil-
tered and concentrated to yield the required acids, which were puri-
fied by flash column chromatography (Et2O–petroleum ether
40:60).
chelation. The known epoxy alcohol 1a was protected as
its TBS silyl ether and treated with sodium azide to give
the two chromatographically separable azido alcohols 13
and 14 in good overall yield. The resulting azido alcohols
13 and 14 were converted into the aziridine 15 by the
Staudinger reduction protocol in excellent yield.18 Protec-
tion of the aziridine 15 as its t-butyl carbonate and subse-
quent deprotection of the silyl ether gave the alcohol 16.
Parikh–Doering oxidation of 16 followed by the addition
of the lithiated alkenylsilane of 4 to the resulting aldehyde
17 gave the Cram products 18 and its diastereomer with
expected selectivity (anti:syn = 1.6:1.0). However, to our
surprise, the subsequent esterification turned out to be
troublesome as the aziridine ring was opened by chloride
anion.19 Ring closure of 19a was partially successful
under basic conditions20 and the ring-opening in the ester-
ification step was avoided by using propionic anhydride
instead of propionyl chloride (Scheme 3).
(3E,5E)-(2S,7S)-7-tert-Butoxycorbonylamino-2-methyltrideca-
3,5-dienoic Acid (20)
TLC (50% Et2O–50% petroleum ether): Rf = 0.19; [a]D +3.1 (c
25
0.65 in CH3Cl). IR (thin film): nmax = 3341 (br m, NH), 2930 and
2858 (s, C-H), 1710 (s, carboxylic C=O), 1513 (m) cm–1. 1H NMR:
(500 MHz, CDCl3): d = 8.20 [1 H, br s, COOH], 6.13 (2 H, m), 5.74
(1 H, m), 5.57 (1 H, m), 4.55 (1 H, br s), 4.10 (1 H, br s), 3.20 (1 H,
p, 7.1 Hz), 1.50–1.28 (19 H, m) and 0.89 (3 H, t, 7.1 Hz). 13C NMR
(125 MHz, CDCl3): d = rotamer: 180.7, 180.2 (C=O), 155.6, 157.8
(C), 134.8, 134.5 (CH), 131.6, 131.7 (CH), 131.4, 131.2 (CH),
129.5, 129.3 (CH), 81.1, 79.7 (Boc-CMe3), 52.2, 53.5 (CH), 43.0
(CH), 35.7, 34.7 (CH2), 32.1 (CH2), 29.4, 29.3 (CH2), 28.7, 28.6
(CH2), 26.0 (CH3), 22.9 (CH2), 17.3 (CH3) and 14.5 (CH3). HRMS:
679.4887, ([MMH]+, C38H67N2O8 requires 679.4892).
Synlett 2004, No. 15, 2771–2775 © Thieme Stuttgart · New York