Communications
coupling strategy, which would lead directly to the
intermediate 14 without the requirement of generating
preactivated and stereodefined olefinic coupling part-
ners. In that regard, Micalizio and co-workers
employed a related approach in the synthesis of
callystatin A by using a homopropargylic ether.[7b]
However, in the present case, the coupling of the
propargylic ether 11 with a terminal alkyne did not
afford the desired diene 14 (Scheme 2C, entries 1 and
2); instead, only the regioisomer 12 and allene 13 were
isolated. We envisioned that the in situ generated
internal alkyne–titanium complex int-1[14] was allowed
to undergo carbometallation with the terminal alkyne
presumably through int-2 and int-3 to afford 12 and 13
respectively,[15] as functionalization of the terminal
alkyne substrate generally occurred at the terminal
carbon atom of the alkyne.[16] Notably, instead of
forming the trisubstituted diene after hydrolysis, the
corresponding allene compounds 13 were generated
presumably through b elimination of the OTBS group
in int-3 (Scheme 2).[17] To overcome the influence of
the propargylic ether, we planned to evaluate the
directing ability of the secondary hydroxy group in 5
since there was a possibility that the proximal heter-
oatom (hydroxy group) would coordinate with the
neighboring metal center.[18] The reactions were ini-
tiated by preforming the lithium oxide using nBuLi,[19]
which would undergo ligand exchange with titanium[20]
to produce lithium isopropoxide, and lead to the
presumed metallacyclopropene intermediate int-4.[21]
If the structure of tethered alkoxide int-4 was retained
À
during the C C bond formation, it would preferentially
afford the metallacyclopentadiene intermediate int-6,
since int-5 would encounter significant strain when
forming the bridgehead alkene.[21] Although the coor-
dination of alkoxide with titanium was expected to
exhibit considerable ring strain in the bicyclic metal-
lacyclopropene int-4,[21a] diene 14c was obtained with
excellent regioselectivity and good yield (entry 3), and
without detection of by-products 12c or 13c. The
terminal alkyne 6 also reacted in a regioselective
manner to give the desired diol product 14d (entry 4),
which was accompanied by a small amount of allene
by-product 13d.[22]
The synthesis of the C14–C19 subunit (Scheme 3A)
commenced with hydrolysis of the oxazole ester 18,[23]
and subsequent coupling with d-proline benzyl ester
hydrochloride. A subsequent Lewis acid promoted
cleavage of the benzyl ester provided amide subunit 4
as a 3:1 mixture of rotamers.
Scheme 2. A) Synthesis of C1–C8 fragment: a) Isobutyraldehyde, TiCl4, CH2Cl2,
À208C, 63%, d.r. >20:1, 95% ee; b) AlMe3, propargylamine, 87%. B) Synthesis
of C9–C13 fragment: c) Zn(OTf)2, (À)-N-methylephedrine, propyne, Et3N,
80%, 95% ee; d) TBSOTf, 2,6-lutidine, 96%. C) Alkyne–alkyne reductive cou-
pling: e) Ti(OiPr)3Cl, cyclopentyl magnesium chloride, toluene, À788C to
À308C; f) nBuLi, Ti(OiPr)3Cl, cyclopentyl magnesium chloride, Et2O, À788C to
À308C. The reported yields are those for the product isolated after column
chromtography. The yields reported for entry 4 are based on recovered starting
material. OTf=trifluoromethanesulfonate, TBS=tert-butyldimethylsilyl.
protocol[12] and subsequent silylation with TBSOTf to afford
the propargylic ether 11 (Scheme 2B). Asymmetric addition
In making use of the difference in the steric environment
using Carreiraꢀs protocol proceeded efficiently with propyne,
and without evidence of self-aldol condensation of the
starting aldehyde.
With the two alkyne reaction partners now available, our
next objective was to assemble the conjugated diene fragment
14. Although there were several options available for the
construction of branched conjugated dienes,[13] we were
interested in the use of a titanium-mediated reductive cross-
of the two secondary hydroxy groups in 14d, the C11 hydroxy
group was selectively silylated as a TBDPS ether 3. Subse-
quent fragment coupling with 4 proceeded efficiently under
the Yamaguchi condition[24] to yield 19 as a mixture of 1:1
rotamers (Scheme 3B). The final stages of the synthesis began
with the selective removal of the primary silyl group of 19, and
oxidation of the derived primary alcohol using IBX afforded
the advanced aldehyde 2.[3b] Treatment of the chloroaldehyde
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 6165 –6168