May and Hoye
JOCNote
SCHEME 3. Competitive Activation of Thioesters
SCHEME 5. Dioxinone Formation and Control Experiments
SCHEME 4. Silver Activation of Phenyl Thioacetoacetate
we would not expect dioxinones to be formed in the presence
of alcohols.15 In testing this hypothesis, we observed that the
addition of silver trifluoroacetate to phenyl thioacetoacetate
(12) in the presence of acetone in CDCl3 at room temperature
resulted in formation of dioxinone 23 (Scheme 5, top).16
Additionally, when isopropyl alcohol was present in the
reaction medium (iPrOH/acetone ∼1:1), none of 23 was
observed; instead, isopropyl acetoacetate (14) was formed.
Both of these results are consistent with the intermediacy of
acetylketene (20).17
Control experiments were run to establish that silver(I) is
playing a definitive role in the activation process (Scheme 5,
bottom): (i) Formation of dioxinone 23 in the absence of
AgO2CCF3 is not acid-catalyzed because phenyl thioacetoa-
cetate (12) is unreactive when treated with acetone and TFA.
(ii) Esterification to produce 14 is also not acid-catalyzed
because 12 is unreactive in the presence of iPrOH and TFA.18
Dioxinone 23 is stable when treated with (iii) iPrOH/TFA or
(iv) iPrOH/AgO2CCF3. Taken collectively, these results
show both that silver is necessary and that the observed
selectivity for reaction with alcohols over ketones is kinetic in
nature. In other words, dioxinone 23 is not an intermediate
en route to 14.
We probed some of the generality of this new dioxinone
forming reaction between these acylketenes and various
ketones. Results are shown in Scheme 6. Good yields were
obtained for products 25a-c; more hindered ketones [(iPr)2CO
or Bn2CO] gave only trace amounts of the corresponding
dioxinone. This is likely because the starting acetoacetate
derivative itself contains a relatively unhindered ketone that
will compete with more hindered ketone partners in trapping
the acylketene intermediate. For this reason, we used a 3:1
ratio of 24 to 12 to obtain the indicated yields.19 Substitution
activate simple thioesters lacking β-keto groups.12 However,
we conjectured that there might be significant differences in
the rates of activation of thioesters having a β-keto group
versus those that do not. We first tested this hypothesis by
exposing a 1:1 mixture of phenyl thioacetoacetate (12) and
phenyl thioacetate (13) to isopropyl alcohol with AgO2CCF3
in CDCl3 and found that 12 was preferentially consumed,
giving a 14:1 ratio of isopropyl acetoacetate (14) and iso-
propyl acetate (15) (at ∼80% conversion, Scheme 3 and
Supporting Information).13
This preferential reaction of 12 led us to the mechanistic
considerations presented in Scheme 4. Silver(I) ion could
complex to the thioacetoacetate 12 or its enol 16. Either of
the resulting cationic intermediates 17 and 18, upon proton
loss, would give the zwitterion 19, which could collapse
directly to acetylketene (20) [Douglas and co-workers have
shown14 that E1cb elimination from the anion of thioace-
toacetates (acetoacetyl CoA) is facile]. While acylium ions 21
and 22 (from 17 or 18, respectively) cannot be ruled out as
intermediates, such a pathway seems unlikely given the
observed lower reactivity of phenyl thioacetate (13).
Since acetylketene (20) is known to be trapped by ketones
to form 1,3-dioxin-4-ones (cf. 2), we reasoned it might be
possible to trap the intermediate ketene if thioacetoacetate
activations are proceeding via these reactive species. Addi-
tionally, acylketenes are known to be trapped faster (ca. 3
orders of magnitude) by alcohols [to generate β-ketoesters
(4)] than by ketones [to generate dioxinones (2)]. Therefore,
(16) While we cannot rule out the intermediacy of a mixed anhydride like
MeCOCH2CO2COCF3 in these reactions, it is noteworthy that use of silver
phosphate or silver triflate was also found to effect this transformation.
(17) We did not observe any species attributable to acetylketene when
monitoring reaction progress by 1H NMR spectroscopy. This was not
unexpected given that (i) the only stable acylketenes are those bearing two
sterically demanding groups (e.g., 1, R1 = R2 = tBu), which lock the species
into the less reactive s-trans conformation (Leung-Toung, R.; Wentrup, C. J.
Org. Chem. 1992, 57, 4850-4858), and (ii) acetylketene dimerizes and
oligomerizes upon being warmed from -77 K by itself and reacts with
alcohol traps between -90 and -50 °C.6
(18) Fischer esterification to form iPrO2CCF3 was observed instead.
(19) Experiments in which hindered ketones or no ketone (or alcohol) at
all were used gave rise to detectable (1H NMR spectroscopy) amounts of the
acetylketene homodimer, dehydroacetic acid (3-acetyl-2-hydroxy-6-methyl-
4H-pyran-4-one), along with other byproducts.
(12) For comparisons of mercury, silver, and copper salts for the activa-
tion of thioesters, see: Masamune, S.; Hayase, Y.; Schilling, W.; Chan, W.;
Bates, G. S. J. Am. Chem. Soc. 1977, 99, 6756–6758.
(13) For the synthesis of 12 (by refluxing 2,2,6-trimethyl-4H-1,3-dioxin-
4-one with thiophenol) see: Clemens, R. J.; Hyatt, J. A. J. Org. Chem. 1985,
50, 2431–2435.
(14) (a) Douglas, K. T.; Yaggi, N. F. J. Chem. Soc., Chem. Commun.
1980, 15, 728–730. (b) Douglas, K. T.; Alborz, M.; Rullo, G. R.; Yaggi, N. F.
J. Chem. Soc., Chem. Commun. 1982, 245–246.
(15) Competitive trapping of acetylketene with a 20-fold excess of cyclo-
hexanone vs pentanol leads to a ∼300:1 ratio of alcohol trapping over ketone
trapping. See: Birney, D. M.; Xu, X. L.; Ham, S.; Huang, X. M. J. Org.
Chem. 1997, 62, 7114–7120.
J. Org. Chem. Vol. 75, No. 17, 2010 6055