An oxidopyrylium cyclization/ring-opening route to polysubstituted α-hydroxytropolones

Article abstract of DOI:10.1021/ol302892g

α-Hydroxytropolones are a class of molecules with therapeutic potential against several human diseases. However, structure-activity relationship studies on these molecules have been limited due to a scarcity of efficient synthetic methods to access them. It is demonstrated herein that α-hydroxytropolones can be generated through a BCl₃-mediated ring-opening/aromatization/demethylation process on 8-oxabicyclo[3.2.1]octenes. Used in conjunction with an improved method based on established oxidopyrylium dipolar cycloadditions, several polysubstituted α-hydroxytropolones can be accessed in three steps from readily available α-hydroxy-γ-pyrones.

Full text of DOI:10.1021/ol302892g

ORGANIC
LETTERS
2012
Vol. 14, No. 23
5988–5991
An Oxidopyrylium
Cyclization/Ring-Opening Route to
Polysubstituted r‑Hydroxytropolones
Christine Meck,†,‡ Noushad Mohd,† and Ryan P. Murelli*,†,‡
Department of Chemistry, Brooklyn College, The City University of New York,
2900 Bedford Avenue, Brooklyn College, Brooklyn, New York 11210, United States, and
Department of Chemistry, The Graduate Center, The City University of New York,
365 Fifth Avenue, New York, New York 10016, United States
rpmurelli@brooklyn.cuny.edu
Received October 19, 2012
ABSTRACT
R-Hydroxytropolones are a class of molecules with therapeutic potential against several human diseases. However, structureÀactivity
relationship studies on these molecules have been limited due to a scarcity of efficient synthetic methods to access them. It is demonstrated
herein that R-hydroxytropolones can be generated through a BCl3-mediated ring-opening/aromatization/demethylation process on
8-oxabicyclo[3.2.1]octenes. Used in conjunction with an improved method based on established oxidopyrylium dipolar cycloadditions,
several polysubstituted R-hydroxytropolones can be accessed in three steps from readily available R-hydroxy-γ-pyrones.
R-Hydroxytropolones are a class of molecules with a
broad range of biological activity against several diseases,
from HIV1 and malaria2 to bipolar disorder and hyperten-
sion (Figure 1).3 This activity is often attributed to bime-
tallic enzyme inhibition1c,3 but, in many instances, is not
known. To date, R-hydroxytropolones have not found
utility clinically due in part totheir toxicity.4 Thus, directed
structureÀactivity relationship (SAR) studies are needed
to increase selectivity as well as potency. Current synthetic
methods unfortunately have limitations, particularly as it
relates to the introduction of polysubstitution,5 and SAR
studies to date have opted to focus instead on natural hy-
droxytropolones or synthetic modifications of them.6 SAR
studies will benefit from new routes that address these
limitations. Herein we reporta shortand divergentroute to
polysubstituted R-hydroxytropolones.
† Brooklyn College.
‡ The City University of New York.
(1) For lead examples relevant to HIV, see: (a) Budihas, S. R.;
Gorshkova, I.; Gaidamakov, S.; Wamiru, A.; Bona, M. K.; Parniak,
M. A.; Crouch, R. J.; McMahon, J. B.; Beutler, J. A.; Le Grice, S. F. J.
Nucleic Acids Res. 2005, 33, 1249. (b) Semenova, E. A.; Johnson, A. A.;
Marchand, C.; Davis, D. A.; Yarchoan, R.; Pommier, Y. Mol. Pharma-
col. 2006, 69, 1454. (c) Himmel, D. M.; Maegley, K A.; Pauly, T. A.;
Bauman, J. D.; Das, K.; Dharia, C.; Clark, A. D., Jr.; Ryan, K.; Hickey,
M. J.; Love, R. A.; Hughes, S. H.; Bergqvist, S.; Arnold, E. Structure
2009, 17, 1625.
(2) For lead examples relevant to malaria, see: (a) Omura, S.;
Kasuhiko, O.; Iwatsuki, M.; Shiomi, K.; Masuma, R. Jpn. Kokai
Tokkyo Koho, 2011084534 A 20110428, 2011. (b) Iwatsuki, M.; Takada,
S.; Mori, M.; Ishiyama, A.; Namatame, M.; Otoguro, K.; Nishihara-
Tsukashima, A.; Nonaka, K.; Masuma, R.; Otoguro, O.; Shiomi, K.;
Omura, S. J. Antibiot. 2011, 64, 183.
(3) Piettre, S. R.; Ganzhorn, A.; Hoflack, J.; Islam, K.; Hornsperger,
J.-M. J. Am. Chem. Soc. 1997, 119, 3201.
(4) Nakagawa, Y.; Tayama, K. Chem.-Biol. Interact. 1998, 116, 45.
Figure 1. Examples of bioactive natural R-hydroxytropolones.
The key steps in our route are an oxidopyrylium dipolar
cycloaddition and subsequent ring expansion (Scheme 1).
Ring expansions of oxidopyrylium-generated 8-oxabicyclo-
[3.2.1]octanes have been used previously in tropolone
r
10.1021/ol302892g
Published on Web 11/20/2012
2012 American Chemical Society

synthesis and provided ample precedence for this ap-
proach.7 Adapting this strategy toward R-hydroxytropo-
lone synthesis requires an R-hydroxy-γ-pyrone-based
base. Despite this, oxidopyrylium cyclization products
formed over time, suggesting reversibility of the dimer-
ization. Assuming that heat would speed the interconver-
sion as well as the cyclization, we isolated dimer 7 and
heated it in the presence of DMAD. It was found that, at
100 °C in the microwave, bicyclooctene 5a was formed in
only 5 min (Scheme 2, eq 2).
Our second discovery was the formation of salt 9,
presumably through demethylation of ylide 8 (Scheme 2,
eq 3), as was evidenced by a signature 1H NMR pattern of
the γ-position of the pyrone at 6.25 ppm and of the 9
methyl protons of the trimethylammonium counterion at
3.78 ppm. Heating salt 9 in the presence of DMAD did not
lead to product within 5 min, and longer reaction times led
to complex mixtures. After surveying several bases, the use
of the bulkier N,N-diisopropylaniline was found to dra-
matically suppress demethylation.
oxidopyrylium cyclization developed by the Wender and
8
Mascarenas laboratories, and modified for intermolecular
~
cyclizations through N,N-dimethylaniline activation of tri-
flate salt 3 (Scheme 2, eq 1).9a We focused our efforts on this
intermolecular cycloaddition process with SAR studies in
mind because it allows for late-stage diversification with
readily available alkynes.
Scheme 1. General Route toward Polysubstituted R-Hydroxy-
tropolones from Commercially Available Kojic Acid
Scheme 2. Established Oxidopyrylium Cyclization and Relevant
Observations Used for Reaction Optimization
At the onset of our studies, only a single example of this
intermolecular oxidopyrylium cyclization with an alkyne
was known, and it was with the highly dipolarophilic di-
methyl acetylenedicarboxylate (DMAD, eq 1).9a After a
brief survey of other alkynes, we found that the method as
described suffered a few key disadvantages including
sluggish reaction times and low yields. Fortunately, two
key observations were made during these studies.
The first discovery was that compound 7, a known oxi-
dopyrylium dimer whose formation had been previously
optimized against,9a formed instantly upon addition of a
(5) (a) Banwell, M. G.; Collis, M. P.; Crisp, G. T.; Lambert, J. T.;
Reum, M. E.; Scoble, J. A. J. Chem. Soc., Chem. Commun. 1989, 10, 616.
(b) Zinser, H.; Sonja, H.; Foehlisch, B. Eur. J. Org. Chem. 2004, 6, 1344.
With these observations in mind, all reactions were
carried out in the presence of N,N-diisopropylaniline to
eliminate demethylation and with microwave irradiation
to shorten reaction times. Indeed, both yields and times
were improved dramatically when compared to our at-
tempts with previously described conditions (Scheme 3,
conditions A vs conditions B). For example, the known
coupling of 3a with DMAD increased in yield from 74% to
89%, and the reaction was completed in only 5 min
compared to 17 h. Coupling of 3a with ethyl propiolate
or 3-butyn-2-one using conditions A yielded 44% of 5d or
5e. Newly developed conditions B increased these yields
dramatically to 88% and 97% respectively. The internal
alkyne, ethyl but-2-ynoate, is both more sterically con-
jested and electronically rich and would not react at all
using conditions A. Gratifyingly, compound 5g could be
generated in 32% yield within 1 h using conditions B.
Phenylacetylene, which yields 5h in 35% yield after one
week when conditions A are used, provides the same
molecule in 57% yield in 0.5 h when conditions B are used.
Theregioselectivityin thisinstancesuggeststhatthephenyl
ꢀ
(6) For recent examples, see: (a) Didierjean, J.; Isel, C.; Querre, F.;
Mouscadet, J.-F.; Aubertin, A.-M.; Valnot, J.-V.; Piettre, S. R.;
Marquest, R. Antimicrob. Agents Chemother. 2005, 49, 4884. (b) Chung,
S.; Himmel, D. M.; Jiang, J.-K.; Wojtak, K.; Bauman, J. D.; Rausch,
J. W.; Wilson, J. A.; Beutler, J. A.; Thomas, C. J.; Arnold, E.; Le Grice,
S. F. J. J. Med. Chem. 2011, 54, 4462.
(7) (a) Baldwin, J. E.; Mayweg, A. V. W.; Neuman, K.; Prichard,
G. J. Org. Lett. 1999, 1, 1933. (b) Adlington, R. M.; Baldwin, J. E.;
Mayweg, A. V. W.; Pritchard, G. J. Org. Lett. 2002, 4, 3009. (c) Baldwin,
J. E.; Mayweg, A. V. W.; Pritchard, G. J.; Adlington, R. M. Tetrahedron
Lett. 2003, 44, 4543. (d) Graening, T.; Bette, V.; Neudorfl, J.; Lex, J.;
Schmalz, H.-G. Org. Lett. 2005, 7, 4317.
(8) (a) Wender, P. A.; McDonald, F. E. J. Am. Chem. Soc. 1990, 112,
~
~
4956. (b) Rumbo, A.; Mourino, A.; Castedo, L.; Mascarenas, J. L.
~
ꢀ
J. Org. Chem. 1996, 61, 6114. (c) Mascarenas, J. L.; Perez, I.; Rumbo, A.;
Castedo, L. Synlett. 1997, 81. (d) Rodrıguez, J. R.; Rumbo, A.; Castedo,
´
~
L.; Mascarenas, J. L. J. Org. Chem. 1999, 64, 4560. (e) Wender, P. A.; D’
Angelo, N.; Elitzin, V. I.; Ernst, M.; Jackson-Ugueto, E. E.; Kowalski,
J. A.; McKendry, S.; Rehfeuter, M.; Sun, R.; Voigtlaender, D. Org. Lett.
2007, 9, 1829. For some recent reviews covering oxidopyrylium
cycloadditions, see: (f) Advances in Cycloaddition; Harmata, M., Ed.; Jai
Press: Stamford, CT, 1999; Vol. 6. (g) Singh, V.; Krishna, U. M.; Vikrant;
Trivedi, G. K. Tetrahedron 2008, 64, 3405. (h) Pellissier, H. Adv. Synth.
Catal. 2011, 353, 189.
~
(9) (a) Wender, P. A.; Mascarenas, J. L. Tetrahedron Lett. 1992, 33,
~
2115. (b) For related work, see: Wender, P. A.; Mascarenas, J. L. J. Org.
Chem. 1991, 56, 6267.
Org. Lett., Vol. 14, No. 23, 2012
5989

group is serving as an electron-accepting group rather than
one that is electron-donating. Indeed, the electronically
poor 4-nitrophenylacetylene can also be used, generating
5i with slightly improved yields.
We next moved toward the desired ring expansion. In an
attempt to perform ring opening and demethylation in a
single pot, boron trihalides were tested. We began with
BBr3, which after 20 min led to a complex mixture of com-
pounds from which a 33% isolated yield of R-methoxytro-
polone 10a was obtained (Scheme 4). Attempts at optimiz-
ing the conditions with BBr3 were not fruitful. Gratifyingly,
however, switching to an excess of BCl3 led cleanly to
hydroxytropolone 6a, which was confirmed by X-ray crys-
tallographic analysis.
Scheme 3. Substrate Studies for Oxidopyrylium Cyclization
Scheme 4. Select Results Comparing BBr3 and BCl3 along with
Crystal Structure of R-Hydroxytropolone 6a
The ring expansion is successful with most substrates,
although a significant substrate scope trend was ob-
served with respect to formation of R-methoxytropolone
vs R-hydroxytropolone (Scheme 5, 6aÀi vs 10aÀi). With
all DMAD-derived oxabicyclo[3.2.1]octenes, selective for-
mation of the hydroxytropolone was observed upon treat-
ment with BCl3 (6aÀ6c), although higher equivalents were
necessary for 6b. Methyl ketone 5e also converted selec-
tively to hydroxytropolone 6e. The ethyl propiolate de-
rived bicycle5dalsoledselectively tohydroxytropolone6d,
but with slightly more methoxytropolone (7:1 ratio). In
this case, the mass recovery was significantly lower, which
we suspect may be due to hydrolysis of a more exposed
ester upon aqueous workup. Surprisingly, phenyl ketone
containing bicycle 5f led to a 1:1 ratio of the two tropolones
(6f and 10f).
An even more dramatic change in reactivity was seen
with compound 5g, which possesses an added electron-
donating methyl group at R3. In this instance, clean
conversion to methoxytropolone 10g takes place, and no
hydroxytropolone is observed at all. Phenyl substrate 5h
led to a complex and inseparable mixture of products
(result not shown). The nitroaryl-containing bicycle 5i,
however, led to a mixture of methoxytropolone and hydro-
xytropolone in a 1:1 ratio. In most cases in which mixtures
arose, the major compounds could be isolated using
reversed-phase chromatography. Alternatively, the reac-
tion mixture can be homogenized to solely hydroxytropo-
lone by heating in HBr and in acetic acid.5 This method
was particularly useful for nitroaryl tropolone 6i, which
was unstable to chromatographic conditions (Scheme 6).
The simplest explanation for the onset to the R-methoxy-
tropolone vs R-hydroxytropolone observation was that
some substrates were more prone to demethylation by BCl3
than others, and that methoxytropolones were in fact inter-
mediates toward hydroxytropolones. However, if this were
a Conditions A: 2 equiv of N,N-dimethylaniline, CH2Cl2, rt. Condi-
tions B: 1.2 equiv of N,N-diisopropylaniline, CHCl3, 100 °C, microwave.
Unless otherwise indicated, 20 equiv of alkyne were used. b5 equiv of
alkyne were used. cReaction run without solvent.
The reaction also works with alternative oxidopyrylium
species, as we have demonstrated by synthesizing chloro5b
and methoxy 5c congeners.10 The chloro substrate is par-
ticularly valuable, as it provides a versatile synthon for
further functionalization. In these instances longer reaction
times were needed, and the yields were lower. This trend is
currently being investigated, although our current hypothe-
sis is that the electron-withdrawing nature of the groups
may destabilize the partial positive charge on the adjacent
oxocarbenium carbon resulting in lower concentrations of
the oxidopyrylium ylide in solution.
One of the current limitations to this method as it stands
now is the amount of alkyne needed. Our optimizations,
which had been carried out with phenylacetylene, sug-
gested that 20 equiv maximized yields. However, when the
reaction was run neat, we were able to use 5 equiv and obtain a
42% yield. Given the low costs associated with most alkynes
and our desire not to limit ourselves to liquid alkynes, 20 equiv
were used. However, 1-phenylprop-2-yn-1-one represented a
solid alkyne in which 20 equiv were cost prohibitive. In this
instance 5 equiv of alkyne led to 5f in excellent yields in only
10 min. Moreover, 70% of the unreacted alkyne was recovered
providing a net alkyne consumption of only 2.2 equiv.
(10) For details on the synthesis of salts 3b and 3c, see the Supporting
Information.
5990
Org. Lett., Vol. 14, No. 23, 2012

Scheme 5. Ring-Expansion Substrate Scopea
Scheme 7. Proposed Mechanism
In instances where R3 is a proton, variability exists. For
example, when R2 is a methyl ketone or ester, selectivity for
hydroxytropolone formation exists. These groups may
destabilize the carbocation through inductive withdraw-
ing, favoring hydroxytropolone formation (6d and 6e).
While the phenyl ketone of substrate 5f may also have
similar inductive withdrawing capabilities, it may also
participate in a through-space cationÀπ interaction, sta-
bilizing the carbocation of 11.11 These competing factors
could explain the lower selectivity in this instance. Nitroar-
yl derivative 5i has less inductive withdrawing then the
carbonyls, leading again to a 1:1 ratio.
In conclusion, we have demonstrated that polysubsti-
tuted R-hydroxytropolones can be generated readily
through the use of oxidopyrylium cyclizations followed
by BCl3-mediated ring expansions. This route will benefit
SAR studies devoted to R-hydroxytropolone therapeutic
development.
a Reactions were run with 7 equiv of BCl3 unless otherwise noted.
Reaction yields are reported following aqueous workup with ratios of
b
6aÀi to 10aÀi being calculated by 1H NMR integration. 15 equiv of
BCl3 were used.
the case, longer reaction times would be expected to lead to
more hydroxytropolone, which was not observed. In addi-
tion, subjecting 10a to BCl3 did not lead to demethylation.
Scheme 6. Example of Homogenizing R-Methoxy and R-Hy-
droxytropolone Mixture by HBr/AcOH Demethylation
Acknowledgment. The authors thank Brooklyn Col-
lege, the National Institutes of Health (SC2GM099596),
and PSC CUNY (65420-00-43) for funding. The authors
also thank William W. Brennessel (University of Rochester)
for X-ray crystallographic studies, Connie M. David
(Louisiana State University) for mass spectroscopy studies,
and John A. Beutler (National Cancer Institute at Frederick)
for helpful discussions.
Our current hypothesis is that when R3 is an electron-
donating group, such as the methyl group in 5g, the ring
opening is favored due to the increased stabilization of the
allylic cation, and the reaction proceeds smoothly to R-
methoxytropolones (11 f 10aÀi, Scheme 7). When R3 is
instead an electron-withdrawing group, such as the ethyl
esters in 5aÀc, ring opening is disfavored due to cation
destabilization. Instead, demethylation drives the ring
expansion, and hydroxytropolones are formed (13 f
6aÀi, Scheme 7).
Note Added after ASAP Publication. Scheme 5 contained
an error in version published ASAP November 20, 2012;
the correct version reposted November 21, 2012.
Supporting Information Available. Full experimental
data and characterization for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(11) For lead reference, see: Dougherty, D. A. Science 1996, 271, 163.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 23, 2012
5991