Figure S1). The relaxed S1 potential energy surface (PES)
scan along the coordinate of the C9ÀO bond length
starting from the S1 energy minimum of both forms
led to an intermediate similar to 4 that was, however,
>40 kcal molÀ1 higher in energy (Figure S2). The calcu-
lations did not indicate involvement of a conical intersec-
tion along the scanned coordinate, and we were unable to
locate any local minimum for 4 on the ground state PES
using various methods (Supporting Information). It is
in agreement with the fact that R-lactone formation is
favored only in systems that possess strong electron-
withdrawing groups.14 In addition, a ground-state transi-
tion state that connects 1 to both 3 and CO was found
(Figure S3), but its high energy (>50 kcal molÀ1) pre-
vents a spontaneous decarbonylation of 1 at 20 °C. We
also could not find a transition state for the CO release on
the triplet hypersurface. Although our DFT calculations
did not provide any evidence that the process involves 4,
its intermediacy should not be ruled out. Additional
experiments and theoretical multiconfigurational models
must be employed to fully understand the CO photore-
lease mechanism from 1.
Scheme 3. Photochemistry of 1a or b at pH 5.7À7.4 (isotopically
labeled 18O is shown in red; the presence of 18O in CO is only a
presumption)
the decomposition quantum yield of the individual form1b
was estimated to be higher by approximately 1 order of
magnitude compared to that of 1a, provided that Φ for
each of the species is not affected by pH in this pH range.
The reaction efficiency at pH = 7.4 was not affected by
the presence of oxygen. Thus either a triplet state was not
involved or its lifetime was too short.
In contrast, an undetermined product with a λmax
of 430 nm (Figure S21) was formed in an aq solution at
pH = 9.5 (the dianion 1a was present exclusively) prob-
ably via a new concomitant (photo)reaction at such high
hydroxide ion concentrations. Practically no photochem-
istry was observed at pH = 4.5, at which 1c was the major
absorbing species. 1 precipitated at pH = 2.5; thus the
quantum yield could not be determined.
Scheme 4. Formation of a Putative Intermediate 4
As a result, we conclude that both 1a and 1b are the only
reactive species which undergo the phototransformation
shown in Scheme 3 in the pH range 5.7À7.4.
Formation of two plausible gaseous side photopro-
ducts, carbon monoxide and carbon dioxide, was con-
sidered. Irradiation of isotopically labeled 1 (ÀC18O2H in
the C9-position), prepared photochemically from 2 in
D218O (Scheme 1),11 in H216O-based buffer (pH = 7.4;
Scheme 3) gave 3 possessing the Cd18O group (Figure S23).
No isotopic incorporation to 3 occurred when 1 with
the ÀC16O2H group was irradiated in D218O (Figure S22).
These experiments thus ruled out the direct involvement of
the solvent in the phototransformation and suggested
that carbon monoxide is the second photoproduct, most
probably containing the oxygen atom from the parent
carboxylic moiety (thermal decomposition of 1 leads to
decarboxylation; Scheme S1).
Photorelease Mechanism. Based on the results of our
isotopic labeling experiments, we hypothesized that the
R-lactone 4, which would further decarbonylate to form
3 (Scheme 4), might be formed as a primary product. It is
known that R-lactones (oxiranones) are short-lived
intermediates12 that decompose efficiently by decarbony-
lation; the most stable known R-lactone has a half-life
of ∼8 h at 24 °C.13 Our TD-DFT calculations showed that
vertical excitation of 1 at the wavelengths of irradiation
used (∼500 nm) populates the lowest singlet excited state
(S1) for both the 1a and 1b forms (Tables S3ÀS5 and
CO Trapping with Hemoglobin. A fast and sensitive
method for determination of CO present in blood15 or
photoreleased from CORMs16 often involves its com-
plexation with hemoglobin (Hb) to form carboxyhemo-
globin (COHb). In this work, an aqueous solution of
uncomplexed Hb (FeII) was prepared by reduction of
bovine methemoglobin (MetHb, FeIII; c = 2.3 Â 10À5 M)
by sodium dithionite.17 It was subsequently mixed with a
solution of 1 (c = 1.3 Â 10À4 M in 0.1 M aq phosphate
buffer, pH = 7.4, purged with N2), and 1 was irradiated at
503 ( 15 nm until complete conversion of Hb to COHb
was observed. Formation of COHb was followed by
absorption spectroscopy (Figure 2), although specific
fluorescence signals of both 1 and 3 also allowed mon-
itoring the course of the reaction. The distinct absorption
characteristics of all species involved, 1a (λmax = 488 nm),
Hb (λmax = 405 nm), and COHb (λmax = 419 nm) (Figure
S17), therefore provide unique advantages for simulta-
neous observation of the CO complexation by using a
(14) Showalter, B. M.; Toscano, J. R. J. Phys. Org. Chem. 2004, 17,
743.
(15) Widdop, B. Ann. Clin. Biochem. 2002, 39, 378.
(16) (a) Pfeiffer, H.; Rojas, A.; Niesel, J.; Schatzschneider, U. Dalton
Trans. 2009, 4292. (b) Zijlstra, W. G.; Buursma, A. Comp. Biochem.
Phys. B 1997, 118, 743.
ꢀ
(12) L’Abbe, G. Angew. Chem., Int. Ed. 1980, 19, 276.
(13) (a) Adam, W.; Liu, J.-C.; Rodriguez, O. J. Org. Chem. 1973, 38,
2269. (b) Coe, P. L.; Sellars, A.; Tatlow, J. C.; Whittaker, G.; Fielding,
H. C. J. Chem. Soc., Chem. Commun. 1982, 0, 362.
(17) Rodkey, F. L.; Hill, T. A.; Pitts, L. L.; Robertson, R. F. Clin.
Chem. 1979, 25, 1388.
Org. Lett., Vol. XX, No. XX, XXXX
C
Antony, Lovely Angel Panamparambil
