found to be reluctant to undergo aldol reaction with oxazole
6, and the trans-oxazoline product 19 was obtained in only
10% yield after several days at elevated temperature (Scheme 3).
Wandless reports that an increased catalyst loading of 10%,
and strictly anhydrous conditions, were all that was required
for the electron rich aldehyde 5 to undergo the aldol reaction
effectively. However, the electron-donating nature and steric
factors introduced by the additional ortho-sulfur substituent
presumably limit the reactivity of 18.
Accordingly, we sought to tune the electronic properties of
the aldol substrate. The sulfide 18 was first converted to the
corresponding sulfoxide in order to switch from an electron
donating to an electron withdrawing group; however, no aldol
adduct was produced at 35 ◦C.
We next sought to convert the methoxy group para to
the aldehyde to a less electron-donating substituent. Selective
demethylation of 18 by treatment with potassium thioacetate in
DMSO, followed by treatment of the resulting phenol 20 with
acetic anhydride, gave the acetate 21 (Scheme 4). A considerable
improvement in the subsequent aldol reaction of 21 with oxazole
6 was observed, with the reaction proceeding at 35 ◦C to give the
desired cis-oxazoline 22 in 39% yield. Despite this improvement,
however, the still-modest yield and difficulties associated with S-
benzyl deprotection led us to further modify the aldol substrate.
alkylation of tertiary amines upon treatment with acylating
agents is a well-known process,12 but very few examples of
the corresponding reactions of sulfides are reported, and only
involving intramolecular acylation.13 We therefore investigated
the feasibility of an S-acylation/dealkylation process for the
removal of the S-Tmob group. Accordingly, the S-Tmob
compound 24 was treated with acetic anhydride at reflux,
and gratifyingly the S-Ac, O-Ac compound 25 was isolated
in 55% yield (Scheme 5). Three transformations are in fact
occurring in this one step: Tmob removal, S-acetylation, and
O-acetylation. Presumably the transformation proceeds through
bis-acylation of S-Tmob compound 23 to give S-acylsulfonium
ion 27, which would lose the stable trimethoxybenzyl cation 28
to give the S-acetyl product 25 (Scheme 6). Note that the lability
of arylthiol esters renders an acylative dealkylation–thioester
hydrolysis process a novel method for the deprotection of S-
Tmob protected arylthiols (vide infra).
Scheme 4 Modified aldol substrate.
Scheme 6 Acylative de-alkylation of Tmob-protected arylthiol.
The 2,4,6-trimethoxybenzyl (Tmob) group has been employed
as a thiol-protecting group and is removed under acidic
conditions.10 Accordingly, we prepared the S-Tmob protected
analogue 23 (Scheme 5) using a similar route to that described
for the S-benzyl compound 20. However, in this case the route
was significantly improved by employing two equivalents of the
Tmob-thiol11 such that the SNAr and demethylation reactions
occurred in the same pot to give 23 in a single step from 17.
Acetylation under mild conditions then furnished the substrate
24. Unfortunately, 24 did not undergo the salen–Al-catalysed
aldol reaction, even at elevated temperature and with increased
catalyst loading. In retrospect, this substrate is significantly more
sterically encumbered than 21.
Compound 25 proved to be the best substrate for the salen–
Al-catalysed aldol reaction to date. Treatment of the aldehyde
25 and oxazole 6 with 25 mol% Al-salen catalyst at 30 ◦C gave
a 52% yield (91% based on recovered starting material) of the
oxazoline 26, as a separable 4 : 1 mixture of cis- and trans-isomers
(Scheme 5). The cis-isomer was produced in 93% e.e., whereas
the trans-isomer was produced in just 17% e.e., indicating the
trans-isomer is produced in poor enantioselectivity directly from
the aldol reaction, rather than by epimerisation of the cis-isomer.
Attempts to drive the reaction to completion resulted in reduced
stereoselectivity.
With the framework of the b-hydroxy-6-mercaptodopa
residue in place, epimerization of the oxazoline, attachment of
the norvaline moiety and asymmetric oxidation of the sulfur
were next investigated. Treatment of aldol adduct cis-26 with
catalytic DBU effected epimerization to the corresponding trans-
oxazoline, while the use of stoichiometric DBU was found to
also effect removal of the S-acetyl group. Ultimately, we were
again able to develop a one-pot procedure to effect several
transformations: treatment of aldol adduct cis-26 with 1.5 equiv.
of DBU in the presence of bromolactone 12 gave the sulfide 30
directly in 69% yield (Scheme 7). Again, three transformations
are conducted in one step: in this case epimerization of the
oxazoline, selective S-deacetylation to give the corresponding
thiophenolate, and subsequent substitution of the bromide 12.
We next investigated the asymmetric oxidation of the sulfide
30 to the corresponding sulfoxide. Our model study (Scheme 2)7
had shown that use of a catalytic amount of (R)-BINOL and
Ti(OiPr)4 in the presence of tBuOOH gave the (S)-sulfoxide 14 as
the only detectable isomer (d.r. >50 : 1). However, when applying
this oxidant system to the more complex substrate 30, very low
stereoselectivity was observed. After much experimentation, the
optimal conditions developed were the use of a stoichiometric
amount of (R)-BINOL and Ti(OiPr)4 and a large excess of
Scheme 5 Optimised aldol substrate.
Replacement of the Tmob group of 24 with an acetyl group
was then investigated, in order to improve both the steric and
electronic nature of the sulfur substituent. While the Tmob
group is normally removed under acidic conditions, we envis-
aged an alternative procedure that would allow direct conversion
of the S-Tmob group to the S-acetyl group. Acylative de-
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 7 3 2 – 7 3 4
7 3 3
Hunter, Luke
