2
M. Eilertsen et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
ring derivatives. As a control, the aryl ring with no substituents,
which has been made previously,13,15–17 was re-synthesised in this
work and subjected to the same photochemical stability testing. All
halogen-containing target compounds (5b–5g) are summarised in
Fig. 2.
The first successful OZO synthesis dates back to 1960 when Bac-
chetti and Alemagna chose to study such compounds with a view
to taking advantage of the characteristic mesoionic ring.17 Similar
rationale led Gotthardt to propose a four-step synthesis in the
1970s and 80s,13,18 before Shaffer and Thompson outlined a six-
step synthesis in the 1990s,16 by which time the significance of
NO and the role of OZOs as NO-donors had been realised, leading
to a patent on their potential use as cardiovascular agents.16 In
2007, Wang proposed an additional route in four-steps with over-
all yields ranging from 39 to 51%, which to date, are the highest to
be reported for this compound class.15 In common with all earlier
examples, the work reported here uses the mandelic acids as start-
ing materials to outline a synthetic route, which the authors
believe to be more robust and reliable than previous methodology
(Scheme 1).
From all previous synthetic approaches the Mitsunobu reaction,
as described by Wang,15 was initially seen as the most attractive
and efficient means of replacing the benzylic secondary alcohol
in 6a with an S-acetyl group. This chemistry relies on the formation
of a complex between diisopropyl azodicarboylate (DIAD) and
triphenyl phosphine, which then deprotonates the benzylic alcohol
before thiolacetic acid is added, to provide the S-acetyl derivative.
However, this approach could not be replicated in this work
despite manipulating the order of addition and reaction times, as
well as tinkering with the number of equivalents of each reagent.
The lack of desired product was attributed to the carboxylic acid
being by far the most acidic group and therefore being the pre-
ferred site of de-protonation. With this in mind, the carboxylic acid
was converted into the methyl ester 7a. This step, using sulphuric
acid and methanol, proceeded without the need for further purifi-
cation in a 68% yield, which is consistent with literature values.19
The Mitsunobu reaction involving 7a still proved to be problematic
even when the thiol was replaced with the thiolate in an attempt
to produce a better nucleophile.
Fig. 1. Previously reported S-nitrosothiols (1–5).
(cf. to compound 1) offered by two methyl groups adjacent to the –
SNO moiety to form a tertiary S-nitrosothiol can aid stability by
reducing both the ease of copper complexation and thiyl radical
dimerization.10 Such subtleties that influence NO release have fas-
cinated the authors and co-workers over many years, which ini-
tially led to the synthesis of a large number of carbohydrate
based S-nitrosothiols (including SNAG 4) to help explore the key
chemical features effecting the stability of the –SNO moiety
(Fig. 1).11 The key result of the carbohydrate-based work, which
looked at varying the lipophilicity, sugar type and even the chiral
form, was to confirm that such compounds are intrinsically unsta-
ble, with many decomposing upon the removal of solvent without
the need for any of the aforementioned decomposition conditions.
Rather than fine-tuning the stability, this emphasised the need for
a much coarser handle on controlling RSNO stability, which led to
interest in 4-aryl-1,3,2-oxathiazolylium-5-olates (OZOs) 5a (Fig. 1),
where the –SNO group is locked into a five-membered heterocyclic
ring in an attempt to gain greater control over NO release.12
Previous work involving carbohydrate based RSNOs, such as
SNAG 4, showed by laser Doppler imaging to bring about a signif-
icant dose dependent vasodilator effect (P = 0.001) to forearm
microvessels following transdermal delivery. A five-fold enhance-
ment in peripheral blood flow over baseline was seen using the
highest SNAG 4 concentration (0.75%, w/w) with intra- and inter-
subject variability of 19% and 16%, respectively.11 Such applica-
tions, where the NO-donor is supplied at or close to the site of
action circumvents, to some degree, the need for highly stable S-
nitrosothiols since degradation is ultimately essential for the quick
and local supply of NO. However, in an attempt to synthesise and
test longer acting RSNOs with improved storage capabilities, the
4-aryl-1,3,2-oxathiazolylium-5-olates were studied with particular
interest in the required ring-opening step prior to RSNO degrada-
tion and NO release.
Attention switched to a bromination and S-acetylation as two
separate steps, using HBr to yield 8a20 and 1.1 equivalents of KSAc
to give 9a (see Scheme 1). When repeated on a larger scale it was
found that PBr3 was a preferable route to 8a, which mirrors the
approach taken by Shaffer and Thompson.16 After washing through
a silica pad, brominated product 8a was obtained in 75% yield and
in pure form, whilst S-acetyl 9a gave an orange oil in a yield of 95%
that did not require further purification. Unlike Shaffer and
Decomposition of 4-aryl-1,3,2-oxathiazolylium-5-olates by
photochemical unimolecular reactions were reported over 30 years
ago13 whilst thermal degradation has also been described at tem-
peratures of 80–140 °C.14 More recently this same family of com-
pounds were shown to decompose when exposed to pH values
close to 5, which would match quite nicely to the skin pH range.15
In this latter work, the ring-opening step required prior to NO
release was shown to depend on the type of substituent attached
to the aryl ring. Based on these findings this work attempted to
expand this family of compounds with a focus on chloro and fluoro
derivatives, since previous findings reported that a 40-chloro sub-
stituent added stability.15 The incorporation of halogens onto the
aromatic ring may also serve to improve the overall lipophilicity
of the OZO and thus make them attractive candidates to study as
peripheral vasodilators via the transdermal delivery route. To fur-
ther explore the role of the substituents, placement on the ring at
the 20, 30 and 40 positions was explored as well as di-substituted
Fig. 2. Target halogen-containing OZO compounds (5b–5g).