Ring opening of cyclic vinylogous acyl triflates using stabilized carbanion nucleophiles: Claisen condensation linked to carbon-carbon bond cleavage

Article abstract of DOI:10.1021/jo100249g

Addition of stabilized carbanionic nucleophiles to cyclic vinylogous acyl triflates (VATs) triggers a ring-opening fragmentation to give acyclic β-keto ester and related products, much like those observed traditionally in the Claisen condensation. Unlike in the classical Claisen condensation, however, the VAT-Claisen reaction described herein is rendered irreversible by C-C bond cleavage, not by deprotonation of the activated methylene product. Full details of this original reaction methodology are disclosed herein, including how subtle differences between the various nucleophiles impact the proper choice of reaction conditions for making 1,3-diketones, β-keto esters, and β-keto phosphonates.

Full text of DOI:10.1021/jo100249g

pubs.acs.org/joc
Ring Opening of Cyclic Vinylogous Acyl Triflates Using Stabilized
Carbanion Nucleophiles: Claisen Condensation Linked to
Carbon-Carbon Bond Cleavage
David M. Jones,† Marilda P. Lisboa, Shin Kamijo,‡ and Gregory B. Dudley*
Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306-4390.
†Current address: Department of Chemistry, University of Pennsylvania, Philadelphia, PA.
‡Current address: Graduate School of Pharmaceutical Sciences, The University of Tokyo, Japan.
gdudley@chem.fsu.edu
Received February 12, 2010
Addition of stabilized carbanionic nucleophiles to cyclic vinylogous acyl triflates (VATs) triggers a
ring-opening fragmentation to give acyclic β-keto ester and related products, much like those
observed traditionally in the Claisen condensation. Unlike in the classical Claisen condensation,
however, the VAT-Claisen reaction described herein is rendered irreversible by C-C bond cleavage,
not by deprotonation of the activated methylene product. Full details of this original reaction
methodology are disclosed herein, including how subtle differences between the various nucleophiles
impact the proper choice of reaction conditions for making 1,3-diketones, β-keto esters, and β-keto
phosphonates.
Introduction
sulfonyl-stabilized carbanions, and similar nucleophiles re-
act with esters in an analogous fashion to provide β-keto
derivatives.1,2
The Claisen condensation has been an important reaction
in synthetic organic chemistry for over a century.1 In the
classic Claisen condensation, an ester enolate nucleophile
reacts with a carboxylic ester electrophile in a reversible
addition and elimination sequence. The resulting β-keto
ester is then subject to irreversible deprotonation at the
R-position, which shifts equilibrium in favor of the condensa-
tion product (Scheme 1).2 Ketone enolates, phosphonyl- and
SCHEME 1. Mechanism of the Classical Claisen Condensation
of Ethyl Acetate
(1) For reviews on the Claisen condensation, see: (a) Davis, B. R.; Garratt,
P. J. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 2, pp 795-863. (b) Hauser, C. R.; Hudson, D. E.
Org. React. 1942, 1, 266. (c) Hauser, C. R.; Swamer, F. W.; Adams, J. T. Org.
React. 1954, 8, 59. For a review on the Dieckman condensation, see: (d) Schaefer,
J. P.; Bloomfield, J. J. Org. React. 1967, 15, 1.
(2) (a) Hauser, C. R. J. Am. Chem. Soc. 1938, 60, 1957–1959.
(b) McElvain, S. M. J. Am. Chem. Soc. 1929, 51, 3124–3130. (c) Roberts,
D. C.; McElvain, S. M. J. Am. Chem. Soc. 1937, 59, 2007–2008. (d)
Nishimura, T.; Sunagawa, M.; Okajima, T.; Fukasawa, Y. Tetrahedron Lett.
1997, 38, 7063–7066.
This report is focused on the use of a Grob-type heterolytic
fragmentation pathway to drive formation of Claisen
condensation-type products: β-keto derivatives of ketones,
3260 J. Org. Chem. 2010, 75, 3260–3267
Published on Web 04/29/2010
DOI: 10.1021/jo100249g
r
2010 American Chemical Society

Jones et al.
JOCArticle
esters, phosphonates, phosphine oxides, and sulfones. The
classic addition and elimination of esters (Claisen con-
densation) is replaced by the addition and fragmentation
of vinylogous esters, specifically, cyclic vinylogous acyl tri-
flates (VATs). Our lab has been exploring the synthetic
utility of triggering Grob fragmentation reactions by nucleo-
philic addition to VATs.3 A broad range of nucleophiles
react with cyclic VATs in an addition/bond cleavage process
to provide acyclic alkynyl ketones (eq 1). This reaction has
been applied and extended to the synthesis of a moth
pheromone, homopropargyl alcohols, and indane building
blocks.4
SCHEME 2. VAT-Claisen Reaction
high yield from symmetrical diones (eq 4). The two-step conver-
sion of cyclic diones to tethered alkynyl ketones has been shown
to be general, affording a wide variety of differentially functio-
nalized substrates.3c
Recent contributions from other laboratories complement
our ongoing methodology and add to the growing arsenal of
C-C bond cleavage reactions for production of valuable
synthetic building blocks. The Williams lab recently ex-
panded the utility of VATs to the synthesis of allenes,
showing through elegant competition experiments and cal-
culations that loss of triflate and formation of the strong
carbonyl bond can drive rapid formation of the higher
energy cumulated allene π-system.8 Brewer and co-workers
employed an analogous yet distinct addition and C-C bond
cleavage process to prepare tethered alkynyl aldehydes;9
note that aldehydes are difficult to access using our method
of hydride-triggered ring opening of VATs.10
VAT-Claisen Reaction
Enolate nucleophiles trigger fragmentation of VATs in a
process we call the VAT-Claisen reaction (Scheme 2).3b The
VAT-Claisen reaction provides 1,3-dicarbonyl-type com-
pounds through an addition and C-C bond cleavage pro-
cess, as opposed to the traditional Claisen condensation,
which delivers 1,3-dicarbonyl-type compounds through
addition and alkoxide elimination reaction of simple esters.
The success of the VAT-Claisen reaction is noteworthy for
several reasons. Enolate addition to ketones (cf. 1a f 8) is
reversible, and in many cases the equilibrium lies on the side
of reactants. Moreover, R,β-unsaturated ketones react with
enolates typically via Michael 1,4-addition, not 1,2-addition.
Nonetheless, we observed products derived from 1,2-addition
of enolates to VAT 1, followed upon warming by fragmenta-
tion to generate 1,3-dicarbonyl compounds. Preliminary
observations were reported previously, and a more detailed
study is the subject of this report.
The pivotal fragmentation process is related to the Eschenmoser-
Tanabe reaction5,6 (3 f 4, eq 2), but it is approached by
nucleophilic addition to cyclic vinylogous esters by analogy to
a related reaction that is known to provide cyclic enones (5 f 6,
eq 3).7 Consequently, ring opening of VATs provides products
that are not accessible using the Eschenmoser-Tanabe frag-
mentation, such as amides and homopropargyl alcohols, as
detailed in previous reports.3c,4b VATs are readily available in
Results and Discussion
(3) (a) Kamijo, S.; Dudley, G. B. J. Am. Chem. Soc. 2005, 127, 5028–5029.
(b) Kamijo, S.; Dudley, G. B. Org. Lett. 2006, 8, 175–177. (c) Kamijo, S.;
Dudley, G. B. J. Am. Chem. Soc. 2006, 128, 6499–6507.
(4) (a) Jones, D. M.; Kamijo, S.; Dudley, G. B. Synlett 2005, 936–938. (b)
Tummatorn, J.; Dudley, G. B. J. Am. Chem. Soc. 2008, 130, 5050–5051. (c)
Jones, D. M.; Dudley, G. B. Tetrahedron 2010, in press; DOI: 10.1016/j.
tet.2010.03.014.
The VAT-Claisen reaction is a key part of a two-step
strategy for converting symmetric cyclic 1,3-diones (e.g., 7,11
eq 4) into value-added acyclic building blocks comprising
(5) (a) Eschenmoser, A.; Felix, D.; Ohloff, G. Helv. Chim. Acta 1967, 50,
708–713. (b) Felix, D.; Shreiber, J.; Ohloff, G.; Eschenmoser, A. Helv. Chim.
Acta 1971, 54, 2896–2912. (c) Tanabe, M.; Crowe, D. F.; Dehn, R. L.; Detre,
G. Tetrahedron Lett. 1967, 40, 3943–3946. (d) Tanabe, M.; Crowe, D. F.;
Dehn, R. L.; Detre, G. Tetrahedron Lett. 1967, 38, 3739–3743.
(6) (a) Weyerstahl, P.; Marschall, H. Fragmentation Reactions. In Compre-
hensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press:
Elmsford, NY, 1991; Vol. 6, pp 1041-1070. (b) Grob, C. A.; Schiess, P. W.
Angew. Chem., Int. Ed. Engl. 1967, 6, 1–15. (c) Grob, C. A. Angew. Chem., Int.
Ed. Engl. 1969, 8, 535–546.
(8) Kolakowski, R. V.; Manpadi, M.; Zhang, Y.; Emge, T. J.; Williams,
L. J. J. Am. Chem. Soc. 2009, 131, 12910–12911.
(9) (a) Draghici, C.; Brewer, M. J. Am. Chem. Soc. 2008, 130, 3766–3767.
(b) Bayir, A.; Draghici, C.; Brewer, M. J. Org. Chem. 2010, 75, 296–302.
(10) Kamijo, S.; Dudley, G. B. Tetrahedron Lett. 2006, 47, 5629–5632.
(11) Preparation of triflate 1a is essentially a quantitative process (yields
typically >95%) when freshly recrystallized 2-methyl-1,3-cyclohexanedione
(cf. 7) is employed. However, 2-methyl-1,3-cyclohexanedione is unstable to
prolonged storage. Routinely, we elect to make use of the crude dione as
received from commercial sources, then purify and store the more stable
VAT 1a. For example, a ca. 80% pure sample of 2-methyl-1,3-cyclohexane-
dione can be converted in ca. 80% yield to VAT 1a, which can then be stored
in the freezer for several months without noticeable decomposition.
(7) (a) Woods, G. F.; Tucker, I. W. J. Am. Chem. Soc. 1948, 70, 2174–
2177. (b) Zimmerman, H. E.; Nesterov, E. E. J. Am. Chem. Soc. 2003, 125,
5422–5430.
J. Org. Chem. Vol. 75, No. 10, 2010 3261

Jones et al.
JOCArticle
SCHEME 3. VAT-Claisen Reaction of 1a and 10, the Lithium
Enolate of Acetophenone
SCHEME 4. Possible Mechanistic Pathways and Deuterium-
Labeling Experiment
alkyne and carbonyl-activated methylene functionality (e.g.,
9, Scheme 2). Such building blocks are valuable for a range of
synthetic applications, as the alkyne and activated methylene
can be processed in parallel by taking advantage of their
complementary reactivity.12 These compounds are also va-
luable tools for synthetic methodology, especially for the
development of intramolecular reactions for the synthesis of
five-membered rings.13
Preliminary Identification and Analysis of the VAT-Claisen
Reaction. In our examination of the Claisen-type condensa-
tion reactions of VATs, we first studied the reaction of the
lithium enolate of acetophenone.3b The stoichiometry of this
reaction played a pivotal role in the ability of the reaction to
proceed to completion. In analogy to the traditional Claisen
condensation, which requires excess base, at least 2 equiv of
enolate nucleophile is necessary to drive the VAT-Claisen
fragmentation of 1 to completion (Scheme 3). The initial
addition is conducted cold, with subsequent warming of the
reaction mixture to promote fragmentation. Reaction of 2.2
equiv of the lithium enolate of acetophenone with VAT 1a
gave rise to 1,3-diketone 9a in 85% yield, whereas with only
1.2 equiv of enolate the yield of 9a dropped to 56% and
unreacted 1a was recovered. If the enolate (1.2 equiv) is
supplemented with additional base (1.0 equiv of LiHMDS),
then complete consumption of VAT 1a is observed and
β-diketone 9a is isolated in 78% yield.
Two mechanistic alternatives were initially considered to
account for the need for 2 equiv of enolate and/or base.
Deuterium labeling helped differentiate between the two
(Scheme 4). The first (path 1, cf. Scheme 3), which we
ultimately came to favor, parallels the addition and frag-
mentation pathway for the Grignard-triggered ring opening
of VATs: direct 1,2-addition of the ketone enolate to VAT 1a
provides tetrahedral intermediate aldolate D, which under-
goes fragmentation upon warming to diketone E. The acidic
methylene of diketone E is metalated as it forms in an
acid/base reaction with excess enolate, which explains the
need for 2 equiv of enolate.
Specific concerns regarding the feasibility of this pathway
prompted us to consider, evaluate, and ultimately discard an
alternative mechanism (path 2). The first concern with path 1
was that ketone enolates typically add to R,β-unsaturated
ketones in a Michael 1,4 fashion, whereas the mechanism
proposed in path 1 involves a direct 1,2-addition. Moreover,
ketone aldolates typically undergo retro-aldol cleavage upon
warming, and (unproductive) proton transfer between the
enolate and ketone often out-competes ketone aldol addition
processes. In other words, we were initially skeptical as to the
feasibility of path 1. Although general reactivity trends do
not necessarily translate to this specific system, we consid-
ered an alternative mechanism initiated by productive proton
transfer between the enolate and VAT 1a (path 2).
(12) Jones, D. M.; Dudley, G. B. Synlett 2010, 02, 223–226.
(13) Kitagawa, O.; Suzuki, T.; Inoue, T.; Watanabe, Y.; Taguchi, T. J.
Org. Chem. 1998, 63, 9470–9475. Nevado, C.; Cardenas, D. J.; Echavarren,
A. M. Chem.;Eur. J. 2003, 9, 2627–2635. Hashimi, A. S. K.; Schwarz, L.;
Choi, J. -H.; Frost, T. M. Angew. Chem., Int. Ed. 2000, 39, 2285–2288. Itoh,
Y.; Tsuji, H.; Yamagata, H. -I.; Endo, K.; Tanaka, I.; Nakamura, M.;
Nakamura, E. J. Am. Chem. Soc. 2008, 130, 17161–17167.
3262 J. Org. Chem. Vol. 75, No. 10, 2010

Jones et al.
JOCArticle
SCHEME 5. Reported Claisen-Type Condensation Reactions
of VAT 1a
SCHEME 7. Proposed Fragmentation Reaction Pathway
Between 1a and 18
SCHEME 6. Prospective Application to Palmerolide A
Claisen-type condensation reaction had to be reoptimized
for the synthesis of olefinating reagents similar to 9e. Chan-
ging the nucleophile from the lithium anion of dimethyl
methylphosphonate (15) to the lithium anion of methyl-
diphenylphosphine oxide (17) afforded Horner-Wittig reagent
9f in the 70% yield range. Upon further optimization,
we found that only 1.1 equiv of the phosphine oxide
nucleophile was necessary to convert VAT 1a effectively
to the corresponding product 9f (eq 6). This optimization
culminated in the synthesis of the C1-C15 fragment of
palmerolide A.12
The alternative mechanistic pathway starts with abstrac-
tion of the acidic R-proton of triflate 1a to form enolate G.
Heterolytic fragmentation then would generate ketene inter-
mediate H, analogous to the allene formation reported by
Williams.8 Addition of another molecule of the lithium
enolate of 14 and proton transfer would yield the most
stable lithium enolate of the 1,3-diketone J through the
intermediate I. Aqueous workup and keto-enol equilibrium
in the presence of proton source would furnish the final
product 9a.
A deuterium-labeling experiment (eq 5, Scheme 4) provided
insight into the mechanistic pathway for the present Claisen-
type condensation. We conducted the reaction using VAT 1a
and acetophenone-methyl-d3 (14-d3). The resultant products
were the corresponding 1,3-diketone 9a without noticeable
deuteration and recovery of 14-d3. These products are con-
sistent with expectations based on path 1, and they are incon-
sistent with material that would be formed via path 2.
Accordingly, we favored the mechanism outlined in path 1.
Preliminary Scope. With this mechanistic model in mind,
we carried out the remainder of our initial experiments under
the same protocol. Many of the nucleophiles examined
gave satisfactory results (Scheme 5),3b but not the anion of
dimethyl methylphosphonate (15); Horner-Wadsworth-
Emmons reagent 9e was isolated in only 21% yield. This
limitation was not inconsequential, as 9e promised to be an
ideal starting material for a total synthesis of palmerolide A
(Scheme 6).14
Refined Mechanistic Analysis. We reopened our investiga-
tion into the VAT-Claisen mechanism in order to gain a
better understanding of the subtle balance of factors in-
volved in determining optimal conditions. To reiterate our
previous observations, the VAT-Claisen reaction of 1a with
the lithium enolate of acetophenone requires 2 equiv of
enolate (Scheme 3), whereas the similar reaction involving
the lithium anion of methyldiphenylphosphine oxide is best
accomplished with 1 equiv of the stabilized nucleophile
(eq 6).
The next logical experiment was reaction of VAT 1a
with 1.1 equiv of 18, the lithium enolate of ethyl acetate
(Scheme 7). Ethyl acetate was one of the best prenucleophiles
in our earlier study (1a f 9c, 88% yield, Scheme 5),3b and
esters are intermediate in acidity between ketones and
phosphine oxides.15
We became interested in the synthesis of palmerolide A, in
part because 9e maps conveniently onto the “northeast”
(as drawn) region of the target molecule. However, the
(14) Diyabalanage, T.; Amsler, C. D.; McClintock, J. B.; Baker, B. J. J.
Am. Chem. Soc. 2006, 128, 5630–5631.
(15) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456–463.
J. Org. Chem. Vol. 75, No. 10, 2010 3263

Jones et al.
SCHEME 8. Proposed Mechanism of the Reaction between
VAT 1a and 17
JOCArticle
The enolate of ethyl acetate provided valuable data. When
1.1 equiv of 18 was added to VAT 1a, the desired fragmenta-
tion product (9c) was obtained in 56% yield (Scheme 7). This
result was in line with our previous observation using 1.2
equivoftheacetophenoneenolate (10) (56%yield, Scheme3).3b
In this case, however, a previously unobserved byproduct was
isolated, which we have tentatively identified as alcohol 21
1
(ca. 26% yield) on the basis of its H NMR spectrum. This
byproduct proved to be unstable to storage, but when immedi-
ately dissolved in THF, treated with excess NaH (approximately
3 equiv), and heated to 60 °C for 30 min, this byproduct gave rise
to the fragmentation product 9c in 78% yield, which provides
support for our proposed structure (21).
A revised mechanistic hypothesis is needed to account for
the formation of β-hydroxy ester 21. On the basis of identi-
fication of aldol 21 after aqueous workup, we now consider
aldolate 19 to be a persistent intermediate. In other words,
ester enolate addition to 1a is not freely reversible. Aldolate
19 begins to undergo fragmentation upon warming, provid-
ing β-ketoester 9c. When only 1 equiv of enolate 18 is
employed, the enolate is essentially consumed following
addition to 1a and prior to fragmentation. Subsequent
deprotonation of the β-ketoester by the alkoxide, not the
enolate, occurs. Aldol 21 does not undergo fragmentation, so
full conversion to β-keto ester 9c is not realized. Thus, 2 equiv
of base is still required for compete conversion of VAT 1a to
9c. Although the isolation of byproduct 21 provides evidence
for the proposed reaction pathway, a competing deprotona-
tion of the β-ketoester by the enolate resulting from a retro-
addition cannot be ruled out.
we must then rationalize why deprotonation of the product
β-ketophosphine oxide does not occur (only 1 equiv of base is
needed). That full conversion was realized with only 1 equiv
of nucleophile suggests that either the β-ketophosphine
oxide product (9f) is not acidic or the presumed17 intermedi-
ate oxy-anion (23) is not basic (Scheme 8). We favor the latter
alternative; the organolithium nucleophile adds irreversibly
at cold temperatures, and the resulting oxy-anion coordi-
nates to the phosphine oxide to provide intermediate 23. This
intermediate is envisioned to resemble an oxaphosphetane
intermediate, much like that formed during a Wittig olefina-
tion reaction.18 Such an intermediate would reduce the oxy-
anion’s ability to deprotonate the β-ketophosphine oxide
product, and it would reduce the possibility of a retro-
addition, thus allowing for the use of 1 equiv of nucleophile
to consume the starting material. When the reaction mixture
is subsequently warmed, the postulated oxaphosphetane-like
intermediate collapses and provides the fragmentation
product 9f, instead of undergoing the classical retro-[2 þ 2]
pyrolysis reaction to provide dienyl triflate 24a (not
observed19).
The results of the condensation reactions of VAT 1a with
the various stabilized anions (cf. 10, 17, and 18) provided a
better understanding of the intricacies of the VAT-Claisen
reaction pathway. It is not solely the reactivity of the
nucleophile that determines stoichiometry requirements,
and the chemical properties of the aldolate-type intermediate
(cf. 11, 19, and 23) must be considered. Armed with new
mechanistic insights, we expanded our methodology to the
synthesis of various β-ketophosphonates. β-Ketophospho-
nates provide reactivity similar to that of β-ketophosphine
oxides (both are olefinating reagents), but the phosphonates
provide some distinct advantages: they are cheaper, more
widely available, and easier to work with than their phosphine
oxide analogues. The next section addresses the conversion of
VATs to novel phosphonate-based olefinating reagents.
Synthesis of β-Ketophosphonates Using the VAT-Claisen
Reaction. The use of phosphonates in organic chemistry
A crossover experiment confirms that ester enolate addi-
tion is reversible to only a limited degree (eq 7). We generated
what we presume to be a 1:1.8 mixture of aldolate 19 and the
lithium enolate of methyl acetate (22) by treating aldol 21
with 2.8 equiv of 22. The isolated fragmentation product
mixture comprised ethyl and methyl esters 9c and 9g in an
88:12 ratio (as estimated by 1H NMR, 78% combined yield),
indicating that retro-aldol reaction is occurring only to a
minor extent. If the aldol addition were rapidly reversible,
then we would expect to see a 35:65 ratio16 of 9c and 9g. If no
retro-aldol reaction were occurring, then only ethyl ester 9c
would be observed. Because both esters are detected with the
ethyl ester as the major product, we conclude that enolate
addition is reversible to a limited extent in this system.
This more detailed understanding of the reaction pathway
enabled us to reconsider the reaction between 1.0 equiv of
phosphine oxide nucleophile 17 and VAT 1a (eq 6). Specifi-
cally, we must discard the assumption of an initial reversible
addition of the phosphorus-stabilized anion to VAT 1a, but
(17) Unlike intermediate 21 from ethyl ester enolate addition, the aldol-
type product derived from aqueous quench (protonation) of 23 has not been
isolated.
(18) (a) Kelly, S. E. Alkene synthesis. In Comprehensive Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 1, pp 729-818. (b)
Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 863–927. (c) Vedejs, E.;
Peterson, M. J. Top. Stereochem. 1994, 21, 1–157.
(16) The ratio of 9c to 9g would be as low at 25:75 if proton transfer
between ethyl acetate, methyl acetate, and their respective enolates were also
to occur.
(19) Small amounts of such byproduct are observed in some cases, as
discussed later in the manuscript.
3264 J. Org. Chem. Vol. 75, No. 10, 2010

Jones et al.
JOCArticle
SCHEME 9. Common Methods for the Preparation of Phos-
phonates
TABLE 1. Reactions of Vinylogous Acyl Triflates with 1.1 equiv of
Phosphonatesa
revolutionized the synthesis of alkenes.20 The ability to
generate E- and Z-alkenes selectively, the mild conditions
required for reaction, and the ease of their synthesis are
distinct advantages of phosphonates as olefination reagents
over their phosphorane (Wittig reagents) or phosphine oxide
(Horner-Wittig) counterparts. Two general strategies are
available for the synthesis of phosphonates (Scheme 9): (1)
the Arbuzov reaction,21 which involves the alkylation of the
corresponding trialkyl phosphite to prepare alkyl-, benzyl-,
and allylphosphonates as well as ester-derived phosphonates
(eq 8) or (2) a Claisen-type condensation between esters and a
dialkyl methylphosphonates to prepare β-ketophosphonates
(eq 9).22 Synthesis of β-ketophosphonates using the Arbuzov
reaction is also known, but one must recognize the potential
for the competing Perkow reaction, which gives rise to enol
phosphates.21a
The synthesis of β-ketophosphonates was of particular
interest to us. Having obtained a poor yield (21%) of phos-
phonate product 9e upon treating VAT 1a with 2.2 equiv of
dimethyl lithiomethylphosphonate (15) under our original
conditions,3b we aimed to determine if the conditions opti-
mized for the lithiomethyldiphenylphosphine oxide nucleo-
phile (17) would provide increased yields of the β-ketopho-
sphonates. The use of lithiomethyldiphenylphosphine oxide as
the nucleophile trigger provided excellent yields (up to 89%,
eq 6). However, the use of phosphonates for alkene synthesis is
aTriflate 1a (0.50 mmol) reacted with nucleophile (0.55 mmol,
generated from 0.6 mmol of phosphonate and 0.55 mmol of n-BuLi).
bIsolated yields. cObtained byproduct 24b (ca. 4% yield). dObtained
byproduct, proposed to be 24c (ca. 8% yield). eDecomposition of VAT
1a. f35% recovered VAT 1a.
much more common.18a,20e,h Moreover, the use of dimethyl
methylphosphonate has a distinct advantage for large-scale
synthesis, in that it costs far less than methyldiphenylpho-
sphine oxide (ca. 66 mmol/$1 vs 1 mmol/$1, respectively).23
Table 1 summarizes the data resulting from the VAT-
Claisen reactions of various VATs with 1.1 equiv of phos-
phonate-derived organolithium reagents. The reaction be-
tween VAT 1a and 1.1 equiv of 15 (dimethyl lithiomethyl-
phosphonate) proceeded in excellent yield (97%, entry 1).
This result marks nearly a 5-fold increase in the yield of 9e
compared to our previous report, in which 2.2 equiv of
nucleophile was used.3b The vinylogous acyl triflates derived
from dimedone and 1,3-cycloheptanedione (1b and 1c,
(20) (a) Boutagy, J.; Thomas, R. Chem. Rev. 1974, 74, 87–99. (b) Heron,
B. M. Heterocycles 1995, 41, 2357–2386. (c) Ernst, H.; Muenster, P.
Carotenoids 1996, 2, 307–310. (d) Lawrence, N. J. The Wittig reaction and
related methods. In Preparation of Alkenes; Oxford University Press: New York,
1996; pp 19-58. (e) Nicolaou, K. C.; Harter, M. W.; Gunzer, J. L.; Nadin, A.
Liebigs Ann. Rec. 1997, 1283–1301. (f) Motoyoshiya, J. Trends Org. Chem.
1998, 7, 63–73. (g) Minami, T.; Okauchi, T.; Kouni, R. Synthesis 2001, 349–357.
(h) Prunet, J. Angew. Chem., Int. Ed. 2003, 2, 2826–2830. (i) Still, W. C.;
Gennari, C. Tetrahedron Lett. 1983, 24, 4405–4408. (j) Blanchette, M. A.; Choy,
W.; Davis, J. T.; Essenfeld, A. P.; Masamune, S.; Roush, W. R.; Sakai, T.
Tetrahedron Lett. 1984, 25, 2183–2186. (k) Iorga, B.; Eymery, F.; Mouries, V.;
Savignac, P. Tetrahedron 1998, 54, 14637–14677.
(21) (a) Arbuzov, B. A. Pure Appl. Chem. 1964, 9, 307–353. (b) Bhattacharya,
A. K.; Thyagarajan, G. Chem. Rev. 1981, 81, 415–430.
(22) (a) Somu, R. V.; Boshoff, J.; Quao, C.; Bennet, E. M.; Barry, C. E.;
Aldrich, C. C. J. Med. Chem. 2006, 49, 31–34. (b) Palacios, F.; Ochoa de
Retana, A. M.; Alonso, J. M. J. Org. Chem. 2006, 71, 6141–6148. (c)
Wetermann, J.; Schneider, M.; Platzek, J.; Petrov, O. Org. Process Res.
Dev. 2007, 11, 200–205. (d) Maloney, K. M.; Chung, J. Y. J. Org. Chem. 2009,
74, 7574–7576. (e) Milburn, R. R.; McRae, K.; Chan, J.; Tedrow, J.; Larsen,
R.; Faul, M. Tetrahedron Lett. 2009, 50, 870–872.
(23) Prices made available from Sigma-Aldrich: http://www.sigmaal-
drich.com/united-states.html.
J. Org. Chem. Vol. 75, No. 10, 2010 3265

Jones et al.
SCHEME 11. Synthesis of Phosphonate 9l, an Analogue of
Phosphonate 9h
JOCArticle
SCHEME 10. Synthesis of Phosphonate 9i
respectively) both provided their respective phosphonate
products, 9h and 9i, in acceptable yields. Interestingly, in
the case of 1b, an unstable byproduct was isolated (ca. 4%),
whose 1H NMR spectrum is consistent with diene 24b. Such
a byproduct would support our proposed oxaphosphetane-
like intermediate (cf. structure 23, Scheme 8).
The reaction between VAT 1a and the anion of diethyl
ethylphosphonate proceeded cleanly in 94% yield (entry 4).
This result is remarkable. In our previous studies of the
Claisen-type condensation reactions, substituents at the
R-position of the nucleophile were linked to decomposition
of the starting VAT. In the case of the nucleophile derived
from diethyl 2-phenylethylphosphorane (entry 5), the phos-
phonate product 9k was obtained in 70% yield. This reaction
provided an unstable byproduct consistent with an E/Z
mixture of dienes 24c, in a roughly 1:1 ratio (ca. 8% yield).
Again, alkene byproducts are consistent with a postulated
oxaphosphetane-like intermediate (Scheme 8). The anion of
bis(trifluoroethyl)-methylphosphonate (entry 6), which would
give rise to a Still-Gennari-type20i olefination reagent, is
prone to homocondensation,22d thus hampering its viability
in Claisen-type condensation reactions.
SCHEME 12. Proposed Mechanism of the Reaction between
VAT 1a and Phosphonates
corresponding phosphine oxide (9f). Fragmentation of
VAT 1a with lithiomethyldiphenylphosphine oxide provides
9f (eq 6), and carrying out a subsequent KAPA zipper (alkyne
isomerization) reaction provides Horner-Wittig reagent 9m
(eq 11, Scheme 11), an analogue of phosphonate 9i.
The stability of various VATs 1a-c deserves comment.
Vinylogous acyl triflates 1b and 1c, which lack the R-methyl
substituent, are relatively unstable when compared to their
analogue, VAT 1a. Vinylogous acyl triflate 1a can be stored
under an inert atmosphere for several months at -10 °C
without any observable decomposition by 1H NMR spectro-
scopy, whereas VAT 1b begins to discolor after 1-2 days.
VAT 1c is even less stable; it began to decompose upon removal
of solvent and had to be used immediately. In addition, 1,3-
cycloheptanedione, the precursor to VAT 1c, is quite expensive
(1 g/$264.50, approximately 30 μmol/$1).23 For these reasons,
the two-step conversion from 1,3-cycloheptanedione (25) to
β-ketophosphonate 9i (Scheme 10) is less than ideal.
We devised an alternative strategy for accessing olefina-
tion reagents linked to alkynes by homologated tethers (>3
carbons, cf. 9i) using the KAPA (potassium 3-amino-
propylamide) acetylene zipper reaction.24 The KAPA acetyl-
ene zipper reaction rearranges internal alkynes to terminal
alkynes. Rearrangement of phosphonate 9e was not effective
(eq 10, Scheme 11), likely because of competing amidation of
the phosphonate with 1,3-propanediamine. This technical
problem was easily overcome by switching back to the
The proposed mechanism of the VAT-Claisen reaction of
phosphonate-derived nucleophiles is illustrated in Scheme 12.
Addition of 1 equiv of lithioalkylphosphonate to VAT 1 gen-
erates β-alkoxyphosphonate 26a and/or oxaphosphetane deri-
vative 26b, depending on the extent of interaction between
oxygen and phosphorus. One equivalent of phosphonate nu-
cleophile is necessary and sufficient for complete consumption of
VAT 1. Excess phosphonate nucleophile can react further with
phosphonate intermediate 26a and/or 26b via the known phos-
phonate Claisen condensation.22 Therefore, use of excess phos-
phonate nucleophile would be detrimental, as observed in our
earlier report.3b The required stoichiometry stands in contrast
to enolate nucleophiles (cf. Schemes 4 and 7). Upon warming,
β-alkoxyphosphonate 26a undergoes fragmentation to β-keto
phosphonate 9. Alternatively, pyrolysis of 26b can occur to
generate cyclic dienyl triflate 24 in small amounts, especially as
steric congestion increases in the system (Scheme 12).
In summary, we have identified, explored, and refined the
VAT-Claisen reaction for the synthesis of alkynes tethered to
1,3-dicarbonyl-type structures. This methodology further
(24) Yamashita, A.; Brown, C. A. J. Am. Chem. Soc. 1975, 97, 891–892.
3266 J. Org. Chem. Vol. 75, No. 10, 2010

Jones et al.
JOCArticle
enables the convenient two-step strategy for the synthesis of
differentially functionalized acyclic keto alkynes from sym-
metrical cyclic diones. The full details disclosed in this work
provide valuable insight into the mechanism of the VAT-
Claisen reaction. The observance of the suspected alcohol
byproduct 21 in reactions of VAT 1a with the lithium enolate
of ethyl acetate allowed for a better understanding of the
mechanism involving phosphine oxide derived nucleophiles.
As is often the case, this is an example of natural pro-
ducts synthesis driving innovation in organic methodo-
logy: our interest in palmerolide encouraged us to reexamine
a poor substrate and improve the scope of the earlier VAT-
Claisen methodology. Ultimately, the results obtained from
careful examination of various nucleophiles and VAT sub-
strates provided a better understanding of these reactions
and allowed for the expansion of the method to the synthesis
of β-ketophosphonates.
We demonstrated throughout the course of our research
into the tandem nucleophilic addition/C-C bond cleavage
reactions of vinylogous acyl triflates that this class of com-
pounds can give rise to interesting and synthetically useful
compounds. Tethered alkynyl ketones, alkynyl β-ketoesters,
alkynyl β-ketophosphine oxides, and now, through reopti-
mized conditions, alkynyl β-ketophosphonates are available
from these easily prepared VAT substrates. The synthetic
utility of the VAT-Claisen reaction has been demonstrated in
the preparation of the C1-C15 fragment of palmerolide A
and the synthesis of some new and useful β-ketophospho-
nates. Studies on the reactivity and synthetic utility of
vinylogous acyl triflates continue in the Dudley laboratory.
Standard Procedure for the Claisen-Type Condensation of the
Vinylogous Acyl Triflates with Phosphonate Nucleophiles. To a
THF solution (2 mL) of dimethyl methylphosphonate (0.6
mmol) was added n-BuLi (2.5 M solution in hexanes, 0.22 mL,
0.55 mmol) at -78 °C. After 20 min of stirring at -78 °C, a
solution of vinylogous acyl triflate 1a (0.50 mmol) in THF was
added dropwise to the resulting solution. The mixture stirred at
-78 °C for 10 min, at 0 °C for 10 min, at rt for 30 min, and at
60 °C for 30 min; during the course of the reaction the solution
changed from clear to yellow and then from yellow to a reddish
solution. The solution was diluted with 3 mL of Et2O. A half-
saturated aqueous NH4Cl solution was used to quench the
reaction, and the mixture was extracted 3 times with 5 mL
portions of EtOAc. The combined organic layers were washed
with 5 mL of NaHCO3(aq) and 5 mL of saturated aqueous NaCl,
dried with MgSO4, filtered, and concentrated. The residual oil
was purified by silica gel column chromatography (gradient
elution from 10% to 40% EtOAc/hexanes) to afford 112 mg of
β-ketophosphonate 9e, together with its enol tautomer (97%
yield) as a pale yellow oil: 1H NMR (300 MHz, CDCl3) δ 3.78 (d,
J = 11 Hz, 6H), 3.10 (d, J = 22 Hz, 2H), 2.73 (t, J = 7.2 Hz, 2H),
2.16 (tq, J = 6.9, 2.5 Hz, 2H), 1.76 (t, J = 2.5 Hz, 3H), 1.75 (app.
quintet, J = 7.0 Hz, 2H); 13C NMR (75 MHz, CDCl3) δ 201.4,
78.0, 76.3, 52.9 (d, J = 6.5), 42.8, 41.3 (d, J = 128 Hz), 22.6,
17.8, 3.3; IR (thin film) 1712, 1449, 1254, 1025, 810 cm-1
;
HRMS (EIþ) Calcd for C10H17O4Pþ [Mþ] 232.0864, found
232.0860. Spectroscopic data consistent with previous report.3b
Acknowledgment. This research was supported by grants
from the National Science Foundation (NSF CHE 0749918),
the James and Ester King Biomedical Research Program,
Florida Department of Health, the Donors of the American
Chemical Society Petroleum Research Fund, an award from
Research Corporation, and by the FSU Department of
Chemistry and Biochemistry. M.L. is a recipient of the
CAPES-Fulbright Graduate Research Fellowship (2008).
S.K. received a Postdoctoral Fellowship for Research
Abroad (2004) from the Japan Society for the Promotion
of Science (JSPS). We are profoundly grateful to all of these
agencies for their support. We thank Dr. Tom Gedris
for support and administration of the NMR facilities,
Dr. Umesh Goli for providing the mass spectrometry data,
and the Krafft Lab for access to their FT-IR instrument.
Stimulating conversations at the 2009 Gordon Research Con-
ference on Natural Products are gratefully acknowledged.
Experimental Section
Standard Procedure for the Claisen-Type Condensation of the
Triflate (1a) with Acetophenone. To a THF solution (2 mL) of
acetophenone (0.14 mL, 1.2 mmol) was added LiHMDS (1.0 M
solution in THF, 1.1 mL, 1.1 mmol) at -78 °C under Ar
atmosphere. After 30 min of stirring at -78 °C, 2-methyl-
3-(trifluoromethanesulfonyloxy)-2-cyclohexenone (1a) (93 μL,
0.50 mmol) was added to the resultant solution. The mixture was
stirred at -78 °C for 10 min, at 0 °C for 10 min, at rt for 30 min,
and then at 60 °C for 30 min. Saturated aqueous NH4Cl solution
was added to quench the reaction, and the mixture was extracted
with ether. The organic layer was washed with water, dried over
MgSO4, filtered, and concentrated. The residue was purified by
silica gel column chromatography (hexanes/ether =100:1-20:1)
to give 1-phenyl-1,3-dioxo-7-nonyne (9a, together with its enol
tautomer) in 85% yield (96 mg).3b
Supporting Information Available: Characterization data
and NMR spectra for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 10, 2010 3267