Photodecarboxylation of xanthone acetic acids: C-C bond heterolysis from the singlet excited state

Article abstract of DOI:10.1021/ol052953d

Irradiation of 2- and 4-xanthone acetic acid in aqueous buffer (pH 7.4) leads to efficient (Φ = 0.67 and 0.64, respectively) photodecarboxylation to give the corresponding methyl products, consistent with an intermediate benzylic carbanion. Fluorescence a

Full text of DOI:10.1021/ol052953d

ORGANIC
LETTERS
2006
Vol. 8, No. 6
1057-1060
Photodecarboxylation of Xanthone
Acetic Acids: C C Bond Heterolysis
−
from the Singlet Excited State
Jessie A. Blake, Eric Gagnon, Matthew Lukeman, and J. C. Scaiano*
Department of Chemistry, UniVersity of Ottawa, Ottawa, Ontario, Canada K1N 6N5
tito@photo.chem.uottawa.ca
Received December 6, 2005
ABSTRACT
Irradiation of 2- and 4-xanthone acetic acid in aqueous buffer (pH 7.4) leads to efficient (Φ ) 0.67 and 0.64, respectively) photodecarboxylation
to give the corresponding methyl products, consistent with an intermediate benzylic carbanion. Fluorescence and laser flash photolysis (LFP)
studies suggest singlet state reactivity, which is unusual for aromatic ketones. 3-Xanthone acetic acid is photoinert under the same conditions.
The results suggest that the reactive xanthone acetic acids are promising precursors for carbanion-mediated photocages.
A number of aryl acetic acids possessing electron-withdraw-
ing substituents on the aromatic ring are known to undergo
efficient photodecarboxylation in neutral aqueous solution
to give benzylic carbanion intermediates, which typically are
rapidly protonated to give the corresponding aryl alkane.1,2
For example, ketoprofen (1), a potent nonsteroidal antiin-
flammatory drug (NSAID), decarboxylates on excitation (in
aqueous solution, pH > pKa) to give carbanion 2, which is
rapidly protonated in approximately 250 ns to give 3-eth-
ylbenzophone (3) (Scheme 1) with an overall quantum yield
of 0.75, a process that is implicated in some of the drug’s
phototoxic side effects.2,3 The photodecarboxylation process
appears to be general for other ketoaromatics, although
Scheme 1. Photodecomposition of Ketoprofen in Basic
Aqueous Solution
modestly higher reactivities are usually encountered for those
with the ketone positioned meta to the acetic acid function
when compared to those with a para arrangement.4 This
phenomenon is an example of the “meta effect” that was
first proposed by Zimmerman5 to explain the photoinduced
hydrolysis of benzyl acetates and can be readily rationalized
with simple MO considerations.
The mechanism of photodecarboxylation from ketoprofen
has been a source of interest for almost two decades. Both
singlet- and triplet-mediated mechanisms have been pro-
posed.2 Our own research has favored a singlet pathway,
(1) (a) Budac, D.; Wan, P. J. Photochem. Photobiol. A: Chem. 1992,
67, 135. (b) Margerum, J. D.; Petrusis, C. T. J. Am. Chem. Soc. 1969, 91,
2467. (c) Stermitz, F. R.; Huang, W. G. J. Am. Chem. Soc. 1971, 93, 3427.
(d) Krogh, E.; Wan, P. J. Am. Chem. Soc. 1992, 114, 705.
(2) (a) Mart´ınez, L. J.; Scaiano, J. C. J. Am. Chem. Soc. 1997, 119, 11066.
(b) Monti, S.; Sortino, S.; De Guidi, G.; Marconi, G. J. Chem. Soc., Faraday
Trans. 1997, 93, 2269. (c) Cosa, G.; Mart´ınez, L. J.; Scaiano, J. C. Phys.
Chem. Chem. Phys. 1999, 1, 3533. (d) Bosca´, F.; Marin, M. L.; Miranda,
M. A. Photochem. Photobiol. 2001, 74, 637.
(3) Borsarelli, C. D.; Braslavsky, S. E.; Sortino, S.; Marconi, G.; Monti,
S. Photochem. Photobiol. 2000, 72, 163.
(4) Xu, M.; Wan, P. Chem. Commun. 2000, 2147-2148.
(5) Zimmerman, H. E. J. Am. Chem. Soc. 1995, 117, 8988.
10.1021/ol052953d CCC: $33.50
© 2006 American Chemical Society
Published on Web 02/21/2006

although the research leading to this conclusion was per-
formed in mixtures of organic solvents with water, in contrast
with the pure aqueous environment in which most of the
ketoprofen reactions have been examined.
Derivatives 4-6 were synthesized in moderate yield via
initial Ulmann coupling between o-iodobenzoic acid and para-,
meta-, or ortho-hydroxyphenylacetic acid, followed by acid-
catalyzed intramolecular acylation.8 Methyl derivatives 7 and
8 were prepared in the same fashion from 2- or 4-cresol.
Absorption spectra of 4-6 in aqueous buffer (pH 7.4) ex-
hibit a desirable improvement over ketoprofen in absorption
above 300 nm, as seen in Figure 1 for 6. The spectra are very
A singlet mechanism represents unusual photochemistry
for benzophenones, which normally is dominated by triplet
processes because of very rapid intersystem crossing (al-
though other examples of singlet-state reactivity are emerg-
ing6). A major difficulty in unequivocally establishing the
multiplicity of the photodecarboxylation of 1 results from
the fact that derivatives of benzophenone are not fluorescent,
which makes it difficult to gather information on the singlet
excited state. To provide insight into the photodecarboxy-
lation mechanism, we decided to study the acetic acids of
xanthone, which is structurally very similar to benzophenone
but is fluorescent in aqueous solutions (due in part to the
much higher oscillator strength of the S0 r S1 transition).
To this end, we decided to prepare and study 2-, 3-, and
4-xanthone acetic acids (4-6, respectively). The differing
substitution of the three isomers is expected to provide
mechanistic insights: if the mechanism proceeded via direct
bond heterolysis (i.e., singlet mechanism), then the reaction
efficiency would be expected to be highly dependent on the
excited-state electronics at the benzylic position, and deriva-
tives with acetic acids substituted meta to the ketone (4 and
6) would be expected to show higher reactivity than at the
5 position. Alternatively, if reaction proceeded by electron
transfer as has been proposed for a triplet-mediated mech-
anism, the substitution should have little effect on the reaction
efficiency because both the distance between and the
oxidation potentials of the donor and acceptor are similar
for the three derivatives.
Figure 1. Equimolar (2.0 × 10-5 M) absorption spectra of
ketoprofen (dashed line) and 6 (solid line).
similar for the three derivatives, and all show the character-
istic π* r π absorption band with a maximum near 340 nm.
The molar absorptivities at their band maxima for 4, 5, and
6 are 5600, 7900, and 8100 M-1 cm-1, respectively, repre-
senting an approximate 40-fold absorption enhancement over
1 (vide infra).
We are also interested in 4-6 as possible new photocage
precursors. Recently, we have developed a new type of
photoremovable protecting group, or “photocage”, based on
the ketoprofen structure that can rapidly release a protected
moiety (such as a carboxylic acid or alcohol) on irradiation
in neutral aqueous solution.7 The photoprocesses of 4 and 6
reported here suggest that these moieties may have improved
spectral properties in relation to their use in photocage design.
To identify the products resulting from irradiation of 4-6,
solutions of each (5 mM in either 0.1 M KOH-H2O or
phosphate buffer, pH 7.4, argon bubbled) were irradiated for
2 min (300 nm) in a merry-go-round apparatus. Each sample,
after workup, was characterized by MS and 1H and 13C NMR.
UVA irradiation of 4 for 2 min gave 2-methylxanthone (7)
in 30% yield as the exclusive photoproduct (Scheme 2),
consistent with a mechanism of reaction involving initial
photodecarboxylation to give a benzylic carbanion that is
subsequently protonated by water to give the observed
photoproduct. With exhaustive irradiation, conversion to 7
was complete with no other photoproducts observed. Irradia-
tion of 4 in deoxygenated 0.1 M NaOD-D2O solution
yielded 7 with one of the arylmethyl hydrogen atoms
replaced with deuterium (7-D, Scheme 2), indicating that
the requisite carbanion intermediate receives a proton (or
deuteron) from the aqueous solvent and that 7 is not formed
via hydrogen abstraction from a radical intermediate, as water
is an extremely poor hydrogen atom donor.
We report here the synthesis and photochemistry of 2-,
3-, and 4-xanthone acetic acids (4, 5, and 6). Very efficient
photodecarboxylation is observed for 4 and 6, and 5 is
essentially photoinert (Scheme 2).
Scheme 2. Photodecomposition of Xanthone Acetic Acids 4
and 6
Similar photochemistry was observed for 6; irradiation in de-
oxygenated phosphate buffer at pH 7.4 (or 0.1 M KOH-H2O)
(6) Yamaji, M.; Susumu, I.; Nakajima, S.; Akiyama, K.; Tobita, S.;
Marciniak, B. J. Phys. Chem. A 2005, 109, 3843.
(7) Lukeman, M.; Scaiano, J. C. J. Am. Chem. Soc. 2005, 127, 7698.
(8) Rewcastle, G. W.; Atwell, G. J.; Baguley, B. C.; Calveley, S. B.;
Denny, W. A. J. Med. Chem. 1989, 32, 793.
1058
Org. Lett., Vol. 8, No. 6, 2006

or in 0.1 M NaOD-D2O gave 8 or 8-D, respectively. How-
ever, to our surprise, 5 was photoinert under the same condi-
tions, with complete recovery of starting material realized.
Quantum yields for photodecarboxylation from 4 and 6
in aqueous phosphate buffer (pH 7.4) were measured using
the photodecarboxylation reaction of ketoprofen (Φ ) 0.75
at pH 7.49) as a secondary reference standard. Both deriva-
tives showed high quantum efficiencies of Φ ) 0.67 for 4
and Φ ) 0.64 for 6 (Table 1), indicating that the xanthone
Table 1. Photochemical Parameters for 4-8
compound ΦR
ΦF
τS (ns) kF (ns-1) τT (ns) triplet yield
4
5
6
7
8
0.67 0.0077 0.079
<0.01 0.044 0.73
0.64 0.010 0.10
0.098
0.061
0.097
0.051
0.039
5400
4500
6600
9500
6200
0.4
1.1
0.3
1.0
(1.0)a
Figure 2. Fluorescence increase with irradiation of 4 (0.02 mM)
in pH 7.4 phosphate buffer. Spectra recorded at 0, 2, 4, 6, 8, 10,
15, 20, 25, 35, and 60 s.
0.24
0.15
4.8
3.9
a Used as a reference for relative triplet yields.
The triplet state yields were evaluated using laser flash
photolysis. On a first approximation, we assume that the trip-
let states of 4-8 have similar extinction coefficients, and thus,
the triplet signal intensities for matched samples can be used
as a measure of the triplet yield. These values are shown in the
last column of Table 1; they are consistent with the fluores-
cence data, i.e., molecules 4 and 6, with highly reactive (de-
carboxylating) singlets that produce lower yields of the triplet
state.
chromophore is able to promote carbanion formation with
efficiencies similar to the benzophenone-based derivatives.
Some xanthone derivatives show fluorescence emission
upon electronic excitation; thus, we expected that fluores-
cence spectroscopy could be a versatile tool in the study of
the photodecarboxylation reaction of 4 and 6. The fluores-
cence excitation and emission spectra of each compound
were recorded for samples of 4-6 (absorbance ) 0.1 at λex)
in phosphate buffer (pH 7.4), and each showed a single weak
fluorescence band with λmax ) 410, 390, and 405 nm,
respectively. Brief irradiation of the solution of 4 with a
handheld TLC UV lamp (λex ) 368 nm, 6.2 W m-2, fwhm
) 16 nm) led to a drastic increase in the fluorescence
intensity and a small shift in the band (Figure 2), presumably
resulting from the formation of the more fluorescent pho-
toproduct, 7. Irradiation of this solution for 1 min gave
quantitative conversion to 7, as evidenced by no further
increases in the fluorescence intensity. In all, a ∼32-fold
increase in fluorescence emission was observed for 4.
Derivative 6 showed similar behavior, with 1 min of
irradiation giving quantitative conversion to 8 resulting in a
∼15-fold enhancement in fluorescence emission. No fluo-
rescence enhancement was observed for irradiation of the
solution of 5; instead, a small decrease in fluorescence
intensity was observed after several minutes of irradiation,
probably resulting from inefficient photodegradation of the
starting xanthone.
We have had success in detecting and characterizing
photogenerated carbanion intermediates from ketoprofen and
related derivatives using laser flash photolysis (LFP), so we
applied the same approach in an attempt to characterize the
transients produced in the photochemistry of 4-6. LFP of
nitrogen-saturated flowing solutions of 4-6 (λex ) 355 nm)
in either 0.1 M KOH-H2O or phosphate buffer adjusted to
pH 7.4: acetonitrile (80:20) gave rise to transients with broad
visible absorption bands with λmax ) 580, 600, and 585 nm
that decayed with clean first-order kinetics at all wavelengths
with lifetimes of 5.4, 4.5, and 6.6 µs (Figure 3). These bands
were readily quenched by common triplet quenchers such
as oxygen, naphthalene-methanol, and sorbate ions. Because
of the similarity of these spectra to the known T-T
absorption spectrum of xanthone, as well as the observed
quenching behavior, we assign the transients observed for
these three species to the corresponding triplet excited state.
We hoped to completely supress the T-T absorption for 4
and 6 in an attempt to detect a possible weak signal arising
from the proposed carbanion intermediates that might be
obscured by the strong signal from the triplet absorption.
Although complete supression of the triplet was achieved
for both 4 and 6 by employing a 50 mM sorbate (quencher)
concentration, no residual signal was detected. This suggests
that the lifetimes of the intermediate carbanions are too short
lived (τ < 20 ns) to be observed on our nanosecond system.
Quenching of the triplet excited state of 4 and 6 by sorbate
was extended to steady-state irradiation studies. Two 8 mM
solutions of 6 were prepared in 1:1 pH 7.4 phosphate buffer
Fluorescence quantum yields and lifetimes were estimated
for 4-8, with values reported in Table 1. Reactive acids 4
and 6 have significantly lower fluorescence quantum yields
and significantly shorter fluorescence lifetimes than the
nonreactive xanthones, suggesting that an additional deac-
tivational pathway is available to the excited singlet state of
4 and 6, namely, the decarboxylation reaction.
(9) Costanzo, L. L.; De Guidi, G.; Condorelli, G.; Cambria, A.; Fama,
M. Photochem. Photobiol. 1989, 50, 359.
Org. Lett., Vol. 8, No. 6, 2006
1059

of 6 (Table 1) actually implies lower reactivity. The rate
constant for singlet decarboxylation can be estimated as kR
) ΦR/τS (see Table 1) and yield values of kR ) 8.5 × 109
s-1 for 4 and of kR ) 6.4 × 109 s-1 for 6, consistent with
MO expectation of a somewhat lower reactivity for 6.
One drawback associated with the ketoprofenate photocage
on which we recently reported7 is that the absorptivity at
wavelengths required for biological applications (>300 nm)
is fairly low (Figure 1). For this reason, it is desirable to
design analogues which retain or improve on the reactivity
of the ketoprofen scaffold but absorb more strongly in the
UVA region, minimizing the secondary effects on the
biological system induced by light irradiation. Xanthone
analogues of the ketoprofen structure hold significant promise
for these applications because they retain the electron-
withdrawing carbonyl which is believed to be necessary for
the initial photodecarboxylation but absorb much more
strongly in the UVA region and, as shown in this contribu-
tion, undergo decarboxylation as efficiently as ketoprofen.
Further, the photochemistry of 6 might also have important
medical implications because this compound and related
derivatives have shown promise as anticancer drugs.8
In summary, a very efficient photodecarboxylation of 4
and 6 occurs in aqueous solution. LFP, fluorescence, and
quenching studies all point to a singlet-mediated mechanism.
Further, the spectral properties of 4 and 6 are such that they
hold promise in the design of new photocages. The readily
detectable fluorescence of products from 4 and 6 in aqueous
media suggests that photocages based on these moieties may
also provide a tool for photorelease mapping; i.e., it should
be possible to determine when and where decarboxylation
(and thus photorelease) has occurred. This may prove a
valuable tool in both in Vitro and in ViVo studies.
Figure 3. Normalized transient decays for 0.08 mM 5 in deoxy-
genated pH 7.4 phosphate buffer/acetonitrile (80:20) with (4) and
without (O) 1 mM sorbate after 355 nm excitation monitored at
600 nm and fitted with first-order expressions.
and acetonitrile, one solution with 50 mM of sorbate and
1
the other without. Following simultaneous irradiation, H
NMR analysis showed that the yield of decarboxylation in
both cases was 26%, on the basis of 1H NMR quantification.
Because we know from the LFP studies that this sorbate
concentration will quench >95% of the triplet yield, the
absence of quenching in photodecarboxylation argues strongly
in favor of a singlet-mediated mechanism.
An additional point of interest in this work is the large
disparity in the reactivity of 4 and 6 vs isomer 5. This appears
to be a result of the meta effect, first proposed by Zimmer-
man, which states that electron-withdrawing and electron-
donating substituents attached to the meta position on an
aromatic ring have a much stronger influence in the excited
state than those substituents substituted at the para position.5
We propose that the high reactivity of 4 and 6 results from
the enhanced electron-withdrawing ability of the ketone that
is meta to the benzylic site in both of these isomers, which
helps stabilize the carbanion formation. The ketone is
substituted para to the benzylic carbon in 5, so this effect is
absent in this case. Instead, the electron-donating ether
oxygen is meta to the benzylic site in 5 which would actually
destabilize carbanion formation. A similar effect is expected
for ortho substituents on an aromatic ring. For this reason,
it was initially surprising that 4 and 6 exhibit similar quantum
yields despite the electron-donating (deactivating) oxygen
ortho to the benzylic position in 6. Although MO calculations
confirm the electronic effect, the longer singlet-state lifetime
Acknowledgment. We acknowledge the generous finan-
cial support of the Natural Sciences and Engineering
Research Coucil (NSERC). J.A.B. thanks NSERC for a post-
graduate scholarship. E.G. thanks NSERC and RISE for
undergraduate research awards, and M.L. thanks NSERC for
a postdoctoral fellowship.
Supporting Information Available: Synthetic procedures
for 4-8, description of photolysis experiments, absorption
extinction coefficients, 1H and 13C NMR of 4-8 before and
after irradiation. This material is available free of charge via
the Internet at http://pubs.acs.org.
OL052953D
1060
Org. Lett., Vol. 8, No. 6, 2006