LAUREATES: AWARDS AND HONORS, SCS FALL MEETING 2006
144
CHIMIA 2007, 61, No. 4
Unfortunately, our first generation syn-
thesis of 1 and related analogs was relative-
ly inefficient and tedious,[6d] which led us
to investigate alternative synthetic routes to
this class of compounds. At the same time,
we have also investigated the effects of the
replacement of the natural epothilone side
chain in 1 with a dimethylbenzimidazole
moiety, leading to analog 2 as an addi-
tional target structure for synthesis (Fig.).
For the latter modification we have previ-
ously shown that it produces a significant
enhancement in antiproliferative activity
in combination with the natural epothilone
macrocycle.[6a,b,11]
Scheme 1. a) 4, LDA, –78 °C, 5 h, then addition of 3, –90 °C, 75 min, 76%,
dr = 8:1; b) PPTS, MeOH, RT, 20 h, 86%; c) (i) TBSOTf, 2,6-lutidine, –78 °C
é RT, 1.5 h; (ii) flash chromatography, 76%; d) (i) H2/Pd-C, MeOH, RT, 20
h; (ii) TPAP, NMO, 4-Å MS, CH2Cl2, RT, 1 h; (iii) MePPh3Br, LiHMDS, THF,
0 °C, 1.5 h, 79% (three steps); e) CSA (1.0 equiv.), CH2Cl2/MeOH 1:1, 0 °C,
1 h, 87%; f) PDC (11 equiv.), DMF, RT, 64 h, 85%.
2. Results and Discussion
Macrocycle formation through ring-
closing olefin metathesis (RCM) featured
as a particularly attractive approach to the
target azathilones 1 and 2, as it would also
provide specific unsaturated analogs (as
the immediate cyclization products), which
could be interesting new antiproliferative
agents in their own right.[12] As shown in
Schemes 1 and 2, our RCM-based synthesis
of target structure 1 and its 9,10-didehydro
derivative 10 involved three key strategic
steps, namely
i) the stereoselective aldol reaction be-
tween aldehyde 3[13] and ketone 4[14] (dr
= 8:1) (Scheme 1),
ii) esterification of carboxylic acid 7 with
the unsaturated alcohol 8,[15] (Scheme 2)
and
iii) RCM with diene 9.
Initial attempts to cyclize 9 employing
the first-generation Grubbs catalyst[16] met
with complete failure and no conversion
was observed. In contrast, the use of the
dihydroimidazol-2-ylidene-based second-
generation catalyst[16] produced the cyclic
olefin in good yield (68%) and with high
E selectivity (>20:1). Similar observa-
tions were made in the cyclization of diene
12, which gave macrolactone 13 in excel-
lent yield and with exclusive E selectivity
(Scheme 3). No trace of the corresponding
Z product could be isolated. Unfortunately,
the efficiency of the cyclization reaction
in both cases was thwarted by serious dif-
ficulties with the subsequent reduction of
the 9,10-double bond, which proved to be
extremely sluggish under all experimental
conditions investigated (thus leading to low
yields and also side reactions such as re-
ductive ester cleavage with H2/Pd-C with-
out reduction of the double bond in the
Scheme 2. a) 8 (1.2 equiv.), DCC (1.3 equiv.), DMAP (0.3 equiv.), CH2Cl2,
0 °C, 30 min, RT, 6 h, 77%; b) 2nd generation Grubbs catalyst (0.09
equiv., incremental addition), CH2Cl2, refl., 24 h, 68% (pure E isomer); c)
HF•pyridine, pyridine, THF, RT, 4.5 h, 65%; d) KO2C-N=N-CO2K (excess),
AcOH, CH2Cl2, refl., 15% (30% based on recovered starting material); pure
1 obtained through purification by preparative HPLC.
sion and after purification by preparative bond formation rather than RCM. As il-
lustrated in Scheme 4, this approach em-
ployed the reductive amination of aldehyde
HPLC. Macrocycle 2 was obtained in 31%
yield from 13 after HPLC purification.
16 with amine 15 (obtained in three steps
The above RCM approach provided
from the known protected tetrol 14[14a]) to
cal testing, but it was clear that more ex- assemble the hetero-aliphatic skeleton of
sufficient amounts of 2 for initial biologi-
tensive profiling of this compound (includ-
ing, perhaps, in vivo studies) would require
the development of an alternative synthetic
2. As the highly polar reductive amina-
tion product was difficult to purify, it was
directly converted to the corresponding N-
case of 13). The most effective method for
route, in order to facilitate the preparation
of larger quantities of material. In light of
highly promising early biological data (vide
infra), we thus embarked on the elaboration
of an alternative route to 2 that would be
based on macrocyclization through ester
tert-butoxycarbonyl derivative (60% yield,
based on amine 15). Selective cleavage of
transforming 10 and 13 into 1 and 2, respec-
tively, involved the use of in situ generated
diimide[17] (Schemes 2 and 3). However,
even under these conditions 1 was obtained
in only 15% isolated yield (30% based on
recovered starting material) at 40% conver-
the primary TBS-ether with CSA, oxida-
tion of the resulting free alcohol with PDC,
and removal of the TBS-protecting group
from C15-O with TBAF furnished seco