Full Paper
Scheme 7. Inversion of the C5 stereogenic center of alcohol 25.
Table 2. Inversion of the C5 stereogenic center of alcohol 25.
Scheme 6. VMAR of aldehyde 23 with dienol silyl ether 24.
Table 1. VMAR of aldehyde 23 and dienol silyl ether 24.
Reagents and conditions
Yield 26:25[a]
[%]
1
2
3
4
5
6
NaBH4, MeOH, ꢀ40 to ꢀ208C
81
69
79
18
0
1.2:1
1.9:1
1.2:1
5.6:1
N.A.[b]
Reagents and conditions
Yield[a]
25:26[b]
NaBH4, CeCl3·7H2O, MeOH/THF (1:1), ꢀ788C
NaBH4, CeCl3·7H2O, iPrOH/THF (1:1), ꢀ78 to 08C
LiAlH(OtBu)3, THF, ꢀ788C to RT
[%]
1
2
3
24, BF3·OEt2, CH2Cl2/Et2O (10:1), ꢀ788C, 50 min
24, (S)-TolBINAP, Cu(OTf)2, TBAT, THF, RT, 22 h
24, (R)-27, iPrOH, nBuCN, ꢀ78 to ꢀ408C, 23 h
94[c]
32 (62)
10 (59)
>20:1
1:2.8
>20:1
L-Selectrideꢁ, THF, ꢀ788C to RT
(R)-2-methyl-CBS-oxazaborolidine, BH3·THF, THF, ꢀ408C 97
>20:1
[a] Diastereomer ratio was estimated by 600 MHz 1H NMR spectroscopic
[a] Recovery of 23 is reported in parentheses. [b] Diastereomer ratio was
estimated by 600 MHz 1H NMR spectroscopic analysis. [c] Overall yield
from 20. TolBINAP=2,2’-bis(di-p-tolylphosphino)-1,1’-binaphthyl, TBAT=
tetra-n-butylammonium difluorotriphenylsilicate.
analysis. [b] N.A.=not applicable. CBS=Corey–Bakshi–Shibata.
the a,b-unsaturated ester moiety and reduction of the ester
functionality, and it did not give 25 nor 26 at all (Table 2,
entry 5). Eventually, we resorted to Corey–Bakshi–Shibata (CBS)
reduction[33] using (R)-2-methyl-CBS-oxazaborolidine and
BH3·THF, which afforded 26 in an almost quantitative yield as
a single stereoisomer (d.r.>20:1; Table 2, entry 6).
the assembly of the C1–C5 domain, as summarized in
Scheme 6 and Table 1. In line with the Felkin–Anh model illus-
trated in Figure 2, the strong syn-selectivity[27] was prominent
in the VMAR of 23 with 24 under the standard conditions
(BF3·OEt2, CH2Cl2/Et2O (10:1), ꢀ788C), giving the alcohol 25 in
94% yield with greater than 20:1 diastereoselectivity (Table 1,
entry 1).[23,29] We also performed the VMAR under the influence
of a chiral Lewis acid. VMAR of 23 with 24 catalyzed by in situ
prepared chiral copper species [CuF·(S)-tolBINAP][30] (TolBI-
The completion of the synthesis of the C1–C11 model com-
pound 4 is depicted in Scheme 8. Acylation of 26 with pro-
pionic anhydride gave the propionate 28 (99%). Cleavage of
the TES ether under mild acidic conditions proceeded without
incident and subsequent acetylation of the resultant alcohol
provided the acetate 29 in 88% yield (two steps). Removal of
the MPM group followed by acylation of the derived alcohol
with isovaleric anhydride furnished the C1–C11 model com-
pound 4 in 77% yield (two steps). Importantly, we did not ob-
serve any migration or cleavage of the O-acyl groups during
this four-step sequence from 28 to 4.
NAP=2,2’-bis(di-p-tolylphosphino)-1,1’-binaphthyl)
provided
26 in 32% yield with only moderate diastereoselectivity
(Table 1, entry 2). The reaction had to be performed at room
temperature; it did not proceed at all under low temperature
conditions. Meanwhile, oxazaborolidinone-catalyzed VMAR of
23 with 24 ((R)-27, iPrOH, nBuCN, ꢀ788C)[31] proved to be un-
productive and delivered the undesired alcohol 25 as the
major stereoisomer in only 10% yield (Table 1, entry 3).
Total synthesis of the proposed structure 2 of didemnaketal
B
Given the high Felkin selectivity observed for the aldehydes
5 and 23, we were interested in whether the conversion of the
undesired C5 epimeric alcohol 25 to the desired alcohol 26 by
an oxidation/reduction sequence might be possible (Scheme 7
and Table 2). Oxidation of 25 with DMP[26] provided the corre-
sponding ketone in 96% yield (Scheme 7). Subsequent reduc-
tion with NaBH4 in the absence or presence of CeCl3·7H2O[32]
proceeded with low diastereoselectivity to give a mixture of
25 and 26 (Table 2, entries 1–3). The use of a bulky reductant
LiAlH(OtBu)3 was beneficial for improving the diastereoselectiv-
ity, although the product yield was unacceptably low (Table 2,
entry 4). Meanwhile, l-Selectrideꢁ caused the 1,4-reduction of
Having completed the studies on the synthesis of the C1–C11
model compound 4, we then focused our attention to the
total synthesis of the proposed structure 2 of didemnaketal B.
Our synthesis plan toward 2 is illustrated in Scheme 9. We en-
visioned that the C21–C28 side chain could be introduced at
the final stage of the total synthesis by means of a Nozaki–
Hiyama–Kishi (NHK) reaction[34] of the aldehyde 30 and the
vinyl iodide 31,[10] given the high functional group tolerance of
the process. The construction of the C1–C7 domain of the al-
dehyde 30 could be achieved via the intermediary of the alco-
Chem. Eur. J. 2014, 20, 1848 – 1860
1852
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim