(NOE) but the reaction was slow and gave mixtures of
products, and the yield was disappointing. Generally, in-
tramolecular reactions have advantages over the equivalent
intermolecular reactions due to gains in reaction rates. We
reasoned that, by having a double tether, the cyclization
process would be further facilitated. We therefore decided
to first prepare the 13-membered compound 15 and then look
at the “doubly tethered intramolecular Michael addition”
(DTIMA) to install the γ-lactone and the decalactone systems
at the same time.
Scheme 3. Synthesis of RCM Precursor
As can be seen (Scheme 2), RCM on the divinyl com-
pound 12 produced the required macrolactone 15 and on
treatment with NaH/THF the DTIMA proceeded cleanly and
in 1 h (rather than overnight) to give the required product
14 in an excellent 94% yield. The RCM-DTIMA sequence
shown therefore showed great promise as a rapid and efficient
entry to the tricyclic macrolide core of the dictyosphaeric
acids.
A short study to determine the optimum conditions for
the RCM reaction was then undertaken. As shown in Table
1, performing the reaction at high temperature caused
[R]22 -109.6 (c 1.45, CH2Cl2)}.6,7 Enone 17 was then
D
R-iodinated using the procedure developed by Krafft et al.10
to afford iodo-enone 18 in excellent yield.
Stille coupling of 18 with allyltributylstannane gave
allylated adduct 19 (83%), which was subsequently depro-
tected with TBAF and coupled with acid 2011 to give RCM
precursor 21 in excellent yield using T3P12 (conventional
coupling agents, such as DCC, EDC, and HATU, were less
efficient).
Table 1. RCM of 12 To Form 15
catalyst
(mol %)
additive
(mol %)
solvent,
temp (°C)
yield
(%)
entry
1
2
3
4
5
6
Grubbs II (10)
Grubbs II (10)
Grubbs II (10)
Grubbs II (10)
Grubbs II (20)
Hoveyda-Grubbs (10)
PhMe, 110 65a,b
CH2Cl2, 40 30c
With substrate 21 in hand, we proceeded to study the key
RCM step for formation of the 13-membered ring in bicycle
22 (Scheme 4). The optimum conditons involved performing
Ti(OiPr)4 (30) CH2Cl2, 40 51
Ti(OiPr)4 (30) PhMe, 55
Ti(OiPr)4 (30) PhMe, 55
69
75
CH2Cl2, 40 69
a Unwanted isomerization of product occurred, with the double bond
formed through RCM partially moving into conjugtion with the enone
system. b Analogous conditions with Grubbs I gave only 50% yield, although
without significant isomerization. c Incomplete conversion of starting mate-
rial.
Scheme 4. Synthesis of Fully Saturated Tricyclic
Intermediates
unwanted isomerization of the product (entry 1). Lowering
the temperature by changing solvent led to incomplete
conversion of the starting material and poor yield (entry 2).
However, the addition of Ti(OiPr)4 to the reaction, to break
up any chelation between the starting material and Ru-
catalyst,9 resulted in improved yields, with toluene proving
to be the best solvent for the reaction (entries 3, 4, and 5).
Ultimately though, utilizing the Hoveyda-Grubbs catalyst
(10 mol %) in refluxing dichloromethane gave the best and
most consistent results (entry 6).
We then went on to apply this RCM-DTIMA sequence to
prepare the complete carbon skeleton of the dictyosphaeric
acids in enantiomerically pure form; the preparation of the
RCM precursor 21 is shown in Scheme 3. The (S)-TBS-
protected γ-hydroxy enone 17 was readily prepared from
commercially available 1-methoxycyclohexa-1,4-diene (16)
the reaction with 10 mol % of the Hoveyda-Grubbs catalyst
in refluxing dichloromethane, as described in Table 1, giving
via literature methods {[R]21 -106.5 (c 1.43, CH2Cl2; lit.7
(10) Krafft, M. E.; Cran, J. W. Synlett 2005, 8, 1263-1266.
(11) Acid 20 was prepared in 99% yield from the DCC coupling of the
commercially available (S)-(+)-4-pentene-2-ol and malonic acid.
(12) Wissmann, H.; Kleiner, H. J. Angew. Chem., Int. Ed. 1980, 19, 133-
134.
D
(9) Fu¨rstner, A.; Langemann, K. J. Am. Chem. Soc. 1997, 119, 9130-
9136.
Org. Lett., Vol. 10, No. 2, 2008
355