4-Aryl-1,3,2-oxathiazolylium-5-olates as pH-controlled NO-donors: The next generation of S-nitrosothiols

Article abstract of DOI:10.1021/ja0682226

S-Nitrosothiols (RSNOs) are important exogenous and endogenous sources of nitric oxide (NO) in biological systems. A series of 4-aryl-1,3,2- oxathiazolylium-5-olates derivatives with varying aryl para-substituents (-CF₃, -H, -Cl, and -OCH₃) were synthesized. These compounds were found to release NO under acidic condition (pH = 5). The decomposition pathway of the aryloxathiazolyliumolates proceeded via an acid-catalyzed ring-opening mechanism after which NO was released and an S-centered radical was generated. Electron paramagnetic resonance (EPR) spin trapping studies were performed to detect NO and the S-centered radical using the spin traps of iron(II) N-methyl-D-glucamine dithiocarbamate [(MGD) ₂-Fe^II] and 5,5-dimethyl-1-pyrroline N-oxide (DMPO). Also, EPR spin trapping and UV-vis spectrophotometry were used to analyze the effect of aryl para substitution on the NO-releasing property of aryloxathiazolyliumolates. The results showed that the presence of an electron-withdrawing substituent such as -CF₃ enhanced the NO-releasing capability of the aryloxathiazolyliumolates, whereas an electron-donating substituent like methoxy (-OCH₃) diminished it. Computational studies using density functional theory (DFT) at the PCM/B3LYP/6-31+G*7/B3LYP/6-31G* level were used to rationalize the experimental observations. The aryloxathiazolyliumolates diminished susceptibility to reduction by ascorbate or gluthathione, and their capacity to cause vasodilation as compared to other S-nitrosothiols suggests potential application in biological systems.

Full text of DOI:10.1021/ja0682226

Published on Web 04/10/2007
4-Aryl-1,3,2-oxathiazolylium-5-olates as pH-Controlled
NO-Donors: The Next Generation of S-Nitrosothiols
Dongning Lu,† Janos Nadas,† Guisheng Zhang,§ Wesley Johnson,‡ Jay L. Zweier,‡
Arturo J. Cardounel,‡ Frederick A. Villamena,*,‡ and Peng George Wang*,†
Contribution from the Departments of Biochemistry and Chemistry, and The DaVis Heart and
Lung Research Institute, and DiVision of CardioVascular Medicine, Department of Internal
Medicine, College of Medicine, The Ohio State UniVersity, Columbus, Ohio 43210, and College
of Chemistry and EnVironmental Science, Henan Normal UniVersity,
Henan Xinxiang 453007, People’s Republic of China
Received November 16, 2006; E-mail: wang.892@osu.edu; frederick.villamena@osumc.edu
Abstract: S-Nitrosothiols (RSNOs) are important exogenous and endogenous sources of nitric oxide (NO)
in biological systems. A series of 4-aryl-1,3,2-oxathiazolylium-5-olates derivatives with varying aryl para-
substituents (-CF3, -H, -Cl, and -OCH3) were synthesized. These compounds were found to release
NO under acidic condition (pH ) 5). The decomposition pathway of the aryloxathiazolyliumolates proceeded
via an acid-catalyzed ring-opening mechanism after which NO was released and an S-centered radical
was generated. Electron paramagnetic resonance (EPR) spin trapping studies were performed to detect
NO and the S-centered radical using the spin traps of iron(II) N-methyl-D-glucamine dithiocarbamate
[(MGD)2-FeII] and 5,5-dimethyl-1-pyrroline N-oxide (DMPO). Also, EPR spin trapping and UV-vis
spectrophotometry were used to analyze the effect of aryl para substitution on the NO-releasing property
of aryloxathiazolyliumolates. The results showed that the presence of an electron-withdrawing substituent
such as -CF3 enhanced the NO-releasing capability of the aryloxathiazolyliumolates, whereas an electron-
donating substituent like methoxy (-OCH3) diminished it. Computational studies using density functional
theory (DFT) at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level were used to rationalize the experimental
observations. The aryloxathiazolyliumolates diminished susceptibility to reduction by ascorbate or
gluthathione, and their capacity to cause vasodilation as compared to other S-nitrosothiols suggests potential
application in biological systems.
Introduction
One of the biggest discoveries in the past decade was the
identification of nitric oxide (NO) as a signaling molecule in
cells and tissues.1 Since then, NO has been shown to participate
in a variety of biological functions,2,3 including the normal
physiological control of vessel dilatation, neurotransmission,
macrophage-induced cytostatics, and cytotoxicity.4
S-Nitrosothiol (RSNOs) species are NO-donor compounds
that are wildly distributed in vivo and have been shown to store,
transport, and release nitric oxide within the mammalian body.
For example, in the blood, S-nitrosoalbumin (SNO-albumin) and
S-nitrosohemoglobin (SNO-Hb) have been reported to constitute
the major conduits for circulating NO bioactivity,5 and the low-
molecular-weight RSNOs, such as S-nitrosoglutathione (GSNO)
(Figure 1) and S-nitrosocysteine (CysNO), have been proposed
Figure 1. Structures of S-nitrosoglutathione (GSNO), S-nitroso-N-acetyl-
penicillamine (SNAP), and aryloxathiazolyliumolates.
as mediators of paracrine protein S-nitrosylation.6 Furthermore,
S-nitrosylation can regulate protein function, as has been
described for numerous proteins.7-11 Therefore, the creation of
synthetic donors, mimicking the endogenous low-molecular-
weight RSNOs for research and therapeutic applications, has
become extremely enticing.12-20
(6) Zhang, Y.; Hogg, N. Am. J. Physiol. 2004, 287, L467.
(7) Lipton, S. A.; Choi, Y. B.; Pan, Z. H.; Lei, S. Z.; Chen, H. S.; Sucher, N.
J.; Loscalzo, J.; Singel, D. J.; Stamler, J. S. Nature 1993, 364, 626.
(8) Lander, H. M.; Ogiste, J. S.; Pearce, S. F.; Levi, R.; Novogrodsky, A. J.
Biol. Chem. 1995, 270, 7017.
(9) Caselli, A.; Camici, G.; Manao, G.; Moneti, G.; Pazzagli, L.; Cappugi, G.;
Ramponi, G. J. Biol. Chem. 1994, 269, 24878.
(10) Gopalakrishna, R.; Chen, Z. H.; Gundimeda, U. J. Biol. Chem. 1993, 268,
27180.
† Departments of Biochemistry and Chemistry, The Ohio State University.
‡ The Davis Heart and Lung Research Institute, and Department of
Internal Medicine, The Ohio State University.
§ Henan Normal University.
(1) Ignarro, L. J. Nitric Oxide 1996, 111.
(11) Mohr, S.; Stamler, J. S.; Brune, B. J. Biol. Chem. 1996, 271, 4209.
(12) Boschi, D.; Tron, G. C.; Di Stilo, A.; Fruttero, R.; Gasco, A.; Poggesi, E.;
Motta, G.; Leonardi, A. J. Med. Chem. 2003, 46, 3762.
(2) Ignarro, L. J., Murad, F., Eds. Nitric Oxide: Biochemistry, Molecular
Biology, and Therapeutic Implications. AdV. Pharmacol. 1995, 34, 530.
(3) Lancaster, J., Jr., Ed. Nitric Oxide: Principles and Actions; 1996; p 355.
(4) Ignarro, L. J., Ed. Nitric Oxide: Biology and Pathobiology; 2000; p 1003.
(5) Foster, M. W.; Pawloski, J. R.; Singel, D. J.; Stamler, J. S. Hypertension
2004, 45, 15.
(13) Cai, T. B.; Lu, D.; Landerholm, M.; Wang, P. G. Org. Lett. 2004, 6, 4203.
(14) Cai, T. B.; Lu, D.; Tang, X.; Zhang, Y.; Landerholm, M.; Wang, P. G. J.
Org. Chem. 2005, 70, 3518.
(15) Huang, J.; Kim-Shapiro, D. B.; King, S. B. J. Med. Chem. 2004, 47, 3495.
9
10.1021/ja0682226 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 5503-5514
5503

A R T I C L E S
Lu et al.
Scheme 1. Photo- and Thermal Decomposition of
1,3,2-Oxathiazolium-5-olates
The most commonly used synthetic RSNO is S-nitroso-N-
acetylpenicillamine (SNAP) (Figure 1), which can induce
apoptosis in a neuronal cell line by the production of different
reactive molecules.21 Other RSNOs have been found to inac-
tivate aconitase and inhibit the uptake of norepinephrine in
sympathetic neurons.22,23 In addition, RSNOs may directly serve
as drug in the treatment of a variety of diseases such as
hypertension,24 atherosclerosis,25 and congestive heart failure.26
RSNOs can also be used as potent antiplatelet agent and
vasodilator. Even though most of these biological functions are
usually attributed to NO release, it has been suggested that
RSNOs can also exhibit direct effects without NO generation
such as S-nitrosation of protein systems.27 An additional
important function of RSNOs is in stress response during GSH
depletion, which may contribute to the well-known oxidant
signaling pathways.28,29 Therefore, the design and synthesis of
novel S-nitrosothiols exhibiting better pharmacokinetic proper-
ties have been the focus in this field for a long time.30
A series of novel sugar-S-nitrosothiols (sugar-SNAPs), de-
veloped by Wang and co-workers,31 have shown better water
solubility, cell penetration, and drug-receptor interaction. Butler
and co-workers32 have also developed a series of novel
S-nitroso-1-thiolsugars, which have both hydrophobic and
hydrophilic groups, allowing them to be delivered transdermally.
A number of S-nitroso peptides33 have also been synthesized
with considerable stability in the presence of copper ion as
compared to SNAP.
of S-nitrosothiols appears to be influenced by the structure of
the organic substituent with the primary and secondary RSNOs
reported to be highly unstable with half-lives of seconds to
minutes,37 whereas tertiary RSNOs such as those derived from
SNAP have been isolated and are indefinitely stable due to the
bulkiness of the alkyl group.38
Because the general instability of S-nitrosothiols has made
them difficult to study, a lot of research has been done in
improving their stability. Electronic and steric effects primarily
determine their stability. For example, evidence has shown that
a small pool of protein S-nitrosothiols are stable for hours in
the intracellular environment due to either electronic static or
steric factors.39 Yet, exogenously, the most efficient synthetic
approaches to making more stable S-nitrosothiols have been
focused on the tertiary RSNOs derivatives. Thus, the exploration
is limited by the chemical structure of the R group, which also
makes it more difficult to fine-tune the stability of these
compounds.
One novel approach to solve this problem is to look for a
more stable pro-drug of S-nitrosothiol that can be converted
into a normal S-nitrosothiol upon activation. A new class of
S-nitrosothiol pro-drugs are the 4-aryl-1,3,2-oxathiazolium-5-
olates. The synthesis and the chemical and photochemical
properties of 4-phenyl-1,3,2-oxathia-zolium-5-olates were first
reported by Gotthardt et al. more than three decades ago.40-43
The structure of 4-aryl-1,3,2-oxathiazolium-5-olates is a cyclic
version of an R-S-nitroso-R-phenyl acetic acid (Figure 1).
Because of this unique cyclic structural feature, aryloxathiaz-
olyliumolates may exhibit better stability as compared to the
linear S-nitrosothiols, and the relative stability of this series of
compounds may be fine-tuned by the different para-substitutions
on the aryl ring. Although the NO-releasing property of
aryloxathiazolyliumolates has not yet been explored, there is
indirect evidence to suggest that NO was generated during
photochemical or thermal decomposition (Scheme 1).40,44 The
final products ArCOCO2Et and hetereocyclic dithioliumolate
were formed via a resonance-stabilized phenyl(oxomethylene)-
thiyl radical.
Despite the enormous biomedical potential of S-nitrosothiols,
they are unstable in solution due to the S-N bond being weak,
sterically hindered, or strongly polarized. The general instability
of the S-N bond leads to low S-NO homolytic/heterolytic bond
dissociation energies.34 The estimated homolytic bond dissocia-
tion energy is between 22 and 32 kcal/mol.35,36 Also, the stability
(16) Huang, J.; Zou, Z.; Kim-Shapiro, D. B.; Ballas, S. K.; King, S. B. J. Med.
Chem. 2003, 46, 3748.
(17) Jia, Q.; Janczuk, A. J.; Cai, T.; Xian, M.; Wen, Z.; Wang, P. G. Expert
Opin. Ther. Pat. 2002, 12, 819.
(18) Lazzarato, L.; Rolando, B.; Lolli, M. L.; Tron, G. C.; Fruttero, R.; Gasco,
A.; Deleide, G.; Guenther, H. L. J. Med. Chem. 2005, 48, 1322.
(19) Sorba, G.; Galli, U.; Cena, C.; Fruttero, R.; Gasco, A.; Morini, G.; Adami,
M.; Coruzzi, G.; Brenciaglia, M. I.; Dubini, F. ChemBioChem 2003, 4,
899.
(20) Xian, M.; Fujiwara, N.; Wen, Z.; Cai, T.; Kazuma, S.; Janczuk, A. J.; Tang,
X.; Telyatnikov, V. V.; Zhang, Y.; Chen, X.; Miyamoto, Y.; Taniguchi,
N.; Wang, P. G. Bioorg. Med. Chem. 2002, 10, 3049.
(21) Terwel, D.; Nieland, L. J. M.; Schutte, B.; Reutelingsperger, C. P. M.;
Ramaekers, F. C. S.; Steinbusch, H. W. M. Eur. J. Pharmacol. 2000, 400,
19.
(22) Bouton, C.; Raveau, M.; Drapier, J. C. J. Biol. Chem. 1996, 271, 2300.
(23) Kaye, D. M.; Wiviott, S. D.; Kobzik, L.; Kelly, R. A.; Smith, T. W. Am.
J. Physiol. 1997, 272, H875.
Herein, the synthesis, stability, theoretical and experimental
mechanistic studies of NO release, and potential biological
applications as vasodilating agents of four 4-aryl-1,3,2-oxathia-
zolylium-5-olate derivatives are presented. This is the first report
of the acid-catalyzed decomposition of aryloxathiazolyliumolates
in aqueous solution and in biological systems that demonstrates
the unique NO-releasing property of these compounds.
(24) Luscher, T. F.; Raij, L.; Vanhoutte, P. M. Hypertension 1987, 9, 157.
(25) Lefer, A. M.; Sedar, A. W. Pharmacol. Res. 1991, 23, 1.
(26) Drexler, H.; Hayoz, D.; Munzel, T.; Hornig, B.; Just, H.; Brunner, H. R.;
Zelis, R. Am. J. Cardiol. 1992, 69, 1596.
(27) Bizzozero, O. A.; DeJesus, G.; Howard, T. A. Neurochem. Res. 2004, 29,
1675.
(28) Zhang, Y.; Hogg, N. Biennial Meeting of the Society for Free Radical
Research International, 12th, Buenos Aires, Argentina, May 5-9, 2004; p
167.
(29) Mathews, W. R.; Kerr, S. W. J. Pharmacol. Exp. Ther. 1993, 267, 1529.
(30) Wang, K.; Zhang, W.; Xian, M.; Hou, Y. C.; Chen, X. C.; Cheng, J. P.;
Wang, P. G. Curr. Med. Chem. 2000, 7, 821.
(31) Hou, Y.; Wang, J.; Andreana, P. R.; Caaantauria, G.; Tarasia, S.; Sharp,
L.; Braunschweiger, P. G.; Wang, P. G. Bioorg. Med. Chem. Lett. 1999, 9,
2255.
(36) Stamler, J. S.; Toone, E. J. Curr. Opin. Chem. Biol. 2002, 6, 779.
(37) Roy, B.; du Moulinet d’Hardemare, A.; Fontecave, M. J. Org. Chem. 1994,
59, 7019.
(32) Butler, A. R.; Field, R. A.; Greig, I. R.; Flitney, F. W.; Bisland, S. K.;
Khan, F.; Belch, J. J. F. Nitric Oxide 1997, 1, 211.
(38) Carnahan, G. E.; Lenhert, P. G.; Ravichandran, R. Acta Crystallogr., Sect.
B 1978, B34, 2645.
(33) Butler, A. R.; Al-Sa’doni, H. H.; Megson, I. L.; Flitney, F. W. Nitric Oxide
1998, 2, 193.
(34) Chen, X.; Wen, Z.; Xian, M.; Wang, K.; Ramachandran, N.; Tang, X.;
Schlegel, H. B.; Mutus, B.; Wang, P. G. J. Org. Chem. 2001, 66, 6064.
(35) Bartberger, M. D.; Mannion, J. D.; Powell, S. C.; Stamler, J. S.; Houk, K.
N.; Toone, E. J. J. Am. Chem. Soc. 2001, 123, 8868.
(39) Zhang, Y.; Hogg, N. Free Radical Biol. Med. 2004, 36, 947.
(40) Gotthardt, H. Chem. Ber. 1972, 105, 2008.
(41) Gotthardt, H. Tetrahedron Lett. 1971, 17, 1277.
(42) Gotthardt, H. Chem. Ber. 1972, 105, 196.
(43) Gotthardt, H. Tetrahedron Lett. 1971, 17, 1281.
(44) Gotthardt, H.; Reiter, F.; Kromer, C. Liebigs Ann. Chem. 1981, 6, 1025.
9
5504 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

4-Aryl-1,3,2-Oxathiazolylium-5-olates as NO-Donors
A R T I C L E S
Scheme 2. Syntheses of 3a-3da
a (i) Ph3P, DIAD, HSAc; (ii) NaOMe/MeOH; (iii) iBuONO/CH2Cl2; (iv) DCC/CH2Cl2.
Results and Discussion
In solution, the decomposition of 3b in the presence of stray
light was also monitored and only showed that 7% of the original
amount decomposed over a period of 3 h. The decomposition
of 3b is slower as compared to S-nitrosoglutathione (GSNO)
with 21% of the compound that decomposed over the same
period of time (Figure S3 of the Supporting Information). This
result indicates that aryloxathiazolyliumolates are relatively more
stable as compared to RSNO in aqueous solution at ambient
conditions. In acetonitrile, all of the aryloxathiazolyliumolates
were found to be stable with only minimal decomposition after
2 h at 37 °C (Figure 4).
Detection of Nitric Oxide. To verify if the decomposition
pathway for 3a-3d did involve NO release (Scheme 3), electron
paramagnetic resonance (EPR) spin trapping was used to directly
detect NO. The detection of NO was carried out using two
methods: (1) purging the acidified solution of 3a-3d with argon
and allowing the gas to pass through the [(MGD)2-FeII] spin
trap solution;47 and (2) directly mixing the acidified solution of
3a-3d with [(MGD)2-FeII].48 For both techniques, a triplet
signal with giso ) 2.041 and hyperfine splitting constant (hfsc)
of aN ) 12.70 G (lit. giso ) 2.04 and aN ) 12.70 G)49 as well
as line width of ∆Bpp ) 3.32 G were observed (Figure 5a).
This spectral feature is consistent with the formation of an
[(MGD)2-FeII-NO] adduct when SNAP was used as the
standard at the same pH of 5 (Figure 5b). The low EPR signal
intensity observed for the [(MGD)2-FeII-NO] adduct, despite
the already high concentration of [(MGD)2-FeII] used of 20
mM, could be due to the weaker basicity of NO in the presence
of an acid because competition between the two Lewis acids
(i.e., the proton and Fe2+) for NO can occur in solution. Heating
the acidified mixture containing the [(MGD)2-FeII] to 45 °C
for 10 min increased the intensity of the EPR signal 2-fold.
This increase in signal intensity at elevated temperature could
be due to the faster decomposition of compounds 3a-3d. There
was no evidence of [(MGD)2-FeII-NO] formation at neutral
pH, indicating that NO production from 3a-3d is initiated by
acid. The formation of [(MGD)2-FeII-NO] adduct was also
observed in the CH3CN-water (1:9) solvent system.
Synthesis. Four 4-aryl-1,3,2-oxathiazolylium-5-olate com-
pounds (3a-3d) were synthesized from mandelic acid deriva-
tives 1a-1d (Scheme 2). Compounds 2a-2d are key interme-
diates for the synthesis of compounds 3a-3d. Literature
methods for the syntheses of 2a-2d treated compounds 1a-
1d with a mixture of HBr and concentrated H2SO4, followed
by sodium thioacetate in anhydrous EtOH to afford compounds
2a-2d.34,44-46 However, this methodology resulted in poor
yields and made the product purification very difficult. More-
over, HBr and H2SO4 caused the cleavage of the methyl ether
in 1d. In this paper, we opted to exploit the Mitsunobu reaction
to prepare the key intermediates 2a-2d because it was a one-
step reaction and it provided clean products. The thiol-acetate
group was introduced to the R-carbons by treating 1a-1d with
triphenyl phosphate, diisopropyl azodicarboxylate (DIAD), and
thiolacetic acid to form the compounds 2a-2d with yields of
70-80%. The quantitative deacetylation of 2a-2d was done
using sodium methoxide in methanol under inert conditions.
The R-aryl-R-mercaptoacetic acid intermediates were nitrosated
and then cyclized to afford the heterocyclic compounds 3a-
3d in 55-65% yield. Compounds 3a-3d displayed vivid colors
depending on the type of substituent on the para-position; that
is, electron-withdrawing substituents were yellow solids, while
the electron-donating substituents were red. These compounds
were soluble in DMSO, acetonitrile, or methylene chloride, but
only slightly soluble in water.
Stability Study. In the solid state, compounds 3a-3d are
very stable and can be stored at room temperature without any
observable decomposition for months. However, a change in
color was observed upon treating the solutions of 3a-3d with
acid. The resulting decompositions of 3a-3d were then studied
spectrophotometrically in a phosphate buffer at pH 5.0. The
decay of absorbance for each compound was monitored. Using
the same initial concentrations, each of the compounds exhibited
varying decay rates with half-lives of approximately 1, 6, 15,
and 130 min for 3a, 3b, 3c, and 3d, respectively (Figure 2).
Compound 3d with an electron-donating group (-OCH3) was
more stable as compared to the unsubstituted 3b, or 3a with an
electron-withdrawing substituent (-CF3). The dependence of
the decay rate within the pH range of 5.0-7.0 was also
investigated for a representative compound, 3b (Figure 3). As
shown in Figure S2 of the Supporting Information, the rate of
decay for 3b increases with decreasing pH with calculated first-
order rate constants of k ) 0.005, 0.03, 0.06, 0.13, and 0.24
min-1 at pH 7.0, 6.5, 6.0, 5.5, and 5.0, respectively. This result
further indicates that the decomposition of aryloxathiazolyliu-
molates is acid-catalyzed.
The formation and decay of [(MGD)2-FeII-NO] was moni-
tored by EPR (Figure 6). The low-field peak intensity was
monitored over a period of 13 min using the direct mixing
technique at pH 5.0 and 7.0. The formation of [(MGD)2-FeII-
NO] was instantaneous after acidification of the mixture
containing 3a-3d and [(MGD)2-FeII], and then the signal
intensity gradually decays. Using the same concentration (i.e,
20 mM) for all compounds, different maxima for 3a-3d were
(47) Li, H.; Cui, H.; Liu, X.; Zweier, J. L. J. Biol. Chem. 2005, 280, 16594.
(48) Mordvintcev, P.; Muelsch, A.; Busse, R.; Vanin, A. Anal. Biochem. 1991,
199, 142.
(49) Xia, Y.; Cardounel, A. J.; Vanin, A. F.; Zweier, J. L. Free Radical Biol.
Med. 2000, 29, 793.
(45) Bruce, J. M.; Sutcliffe, F. K. J. Chem. Soc. Abstr. 1957, 4789.
(46) Gotthardt, H. Chem. Ber. 1972, 105, 188.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5505

A R T I C L E S
Lu et al.
Figure 2. Decomposition of 3a-3d (0.8 mM) in 0.1 M phosphate buffer, pH 5.0, 37 °C. UV was measured at 581 nm (3a), 587 nm (3b), 592 nm (3c), and
602 nm (3d).
Figure 4. Decompositions of 3a (O), 3b (b), 3c (2), and 3d (×) (0.4
Figure 3. Decomposition plot at 587 nm of 3b (0.4 mM) in 0.1 M
mM, in acetonitrile, 37 °C). Absorbance was measured at 581 nm (3a),
phosphate at pH 5.0-7.0.
587 nm (3b), 592 nm (3c), and 602 nm (3d).
observed using SNAP as standard, assuming that 99.9% of NO
was released from SNAP at pH 5. A concentration of 5.34 mM
of [(MGD)2-FeII-NO] was achieved for 3a, which corresponds
to ∼27% total NO, while the lowest NO concentration was
observed for 3d (∼9% yield). No evidence of [(MGD)2-FeII-
NO] formation was observed at pH 7.0. Because the unsubsti-
tuted aryloxathiazolyliumolates (3b) have an NO yield of 21%,
these results show that the presence of an electron-withdrawing
substituent at the para position of the aryl group (3a) enhances
the NO-releasing property of the aryloxathiazolyliumolates,
whereas the electron-donating substituent decreases it. This
substituent effect can be exploited as a means to “fine-tune”
the NO-releasing property of aryloxathiazolyliumolates.
Detection of Thiyl Radical Intermediates. We propose that
there is a formation of a thiyl radical when NO is released via
the homolytic S-N bond cleavage as has been found in the
decomposition pathway for many RSNO compounds.50 How-
ever, the heterolytic bond cleavage of S-N to form nitroxyl
9
5506 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

4-Aryl-1,3,2-Oxathiazolylium-5-olates as NO-Donors
A R T I C L E S
Figure 5. X-band EPR and simulated (overlapped) spectra of (a) NO-adduct arising from 3a (20 mM) in the presence of [(MGD)2-FeII] (20 mM), aN
)
12.70 G, ∆Bpp ) 3.32 G; (b) NO-adduct arising from SNAP (6 mM) in the presence of [(MGD)2-FeII] (20 mM), aN ) 12.64 G, ∆Bpp ) 3.35 G; (c)
S-centered spin adduct arising from DMPO (200 mM) and 3a (20 mM) all at pH 5, aN ) 14.74 G, aH ) 13.92 G, ∆Bpp ) 0.78 G; (d) hydroxyl radical adduct
from UV photolysis of 0.5% H2O2 in the presence of 25 mM DMPO, aN ) 14.97 G, aH ) 14.73 G, ∆Bpp ) 0.59 G.
Scheme 3. Radical-Radical Reactions for the Decomposition of 3a-3d upon Acidificationa
a Compounds 4-12 are part of the decomposition mechanism as shown in Tables 3 and 4.
(NO-) might also be possible, because the [(MGD)2-FeII-NO]
adduct can form from NO- and [(MGD)3-FeIII]. EPR spin
trapping was used to detect any other radical intermediates
formed other than NO to determine which type of S-N cleavage
was responsible for the release of NO. Figure 5c shows a typical
EPR spectrum obtained upon the acidification of compound 3a
in the presence of a nitrone spin trap, DMPO. The observed
overall hyperfine feature of 1:2:2:1 is similar to that found for
the DMPO-OH adduct (Figure 5d). However, the spectral
profile was significantly different for DMPO-3a adduct with
hfsc’s of aN ) 14.74 G and aH ) 13.92 G, as compared to the
DMPO-OH adduct with aN ) 14.97 G and aH ) 14.73 G (lit.
aN ) 14.9 G and aH ) 14.9 G).51 Moreover, the observed peak-
with hfsc’s that are different from 3a, for example, for 3b, aN
) 14.30 G and aH ) 15.31 G; 3c, aN ) 14.29 G and aH
)
15.32 G; and 3d, aN ) 14.29 G and aH ) 15.26 G. All of the
EPR spectra from the DMPO adducts of 3b-3d gave line widths
similar to that of 3a with ∆Bpp ) 0.70-0.73 G. The observed
EPR parameters were further supported by the EPR spectral
parameters reported for S-centered radicals generated from low
molecular weight thiols whose hfsc values were in the range of
aN ) 15.2-15.8 G and aH ) 15.2-18.0 G.52 The differences
in the EPR spectral parameters between DMPO-OH and the
spectra arising from the decomposition of 3a-3d in the presence
of DMPO indicate the formation of a radical other than HO•.
To further confirm the nature of the radical generated from
3, the spin trap R-phenyl-tert-N-butyl nitrone (PBN) was also
employed for the detection of thiyl intermediates by EPR. Using
the Fenton system, the PBN-OH adduct was generated giving
the following hfsc’s: aN ) 14.43 G and aH ) 2.42 G,53 aN )
14.76 G and aH ) 2.75 G. However, acidification of 3b solution
to-peak line width in Figure 5c is much broader with ∆Bpp
)
0.78 G as compared to DMPO-OH’s ∆Bpp ) 0.59 G (Figure
5d). Furthermore, the DMPO spin adducts arising from the
acidification of compounds 3b, 3c, and 3d gave EPR spectra
(50) Wang, P. G.; Xian, M.; Tang, X.; Wu, X.; Wen, Z.; Cai, T.; Janczuk, A.
J. Chem. ReV. 2002, 102, 1091.
(51) Villamena, F. A.; Hadad, C. M.; Zweier, J. L. J. Phys. Chem. A 2003, 107,
4407.
(52) Davies, M. J.; Forni, L. G.; Shuter, S. L. Chem. Biol. Interact. 1987, 61,
177.
(53) Hinton, R. D.; Janzen, E. G. J. Org. Chem. 1992, 57, 2646.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5507

A R T I C L E S
Lu et al.
Figure 6. Formation and decay of [(MGD)2-FeII-NO] adduct generated from 3a-3d (20 mM) in the presence of [(MGD)2-FeII] (20 mM) at pH 5.0 and
7.0. The arrow indicates the peak that is being monitored. All data were the average of three measurements with a standard deviation of less than 10%.
Table 1. ESI Mass of Various Products Generated from the
formed (0.84 mM). This preference of DMPO adduct formation
follows the same qualitative trend as observed for the NO adduct
formation. This further supports the electronic effect of the aryl
substituents on the radical generating ability of the aryloxathia-
zolyliumolate analogues.
Acidification of Aryloxathiazolyliumolate 3aa
product
16 or 18
17
19
20
21 or 22
Mobs
Mcalcd
M + Naobs
M + Nacalcd 493.4
471.7
471.4
493.5
not observed 383.7
191.6
193.2
427.1
427.4
449.5
449.4
237.2
259.2
259.2
383.4
407.4 not observed
405.4 215.2
The difference in the formation profile between Figures 5
and 6 for the spin adducts using [(MGD)2-FeII] and DMPO,
respectively, could be due to the fast rate of formation of the
[(MGD)2-FeII-NO] adduct with k ≈ 108 M-1 s-1 55 as
compared to the formation of the S-centered radical adduct with
DMPO, which was shown to be reversible with k ≈ 107-108
a Refer to Scheme 3 for the respective structures.
Table 2. Reduction Potentials E1/2 of 3a-3d Determined by Cyclic
Voltammetry Relative to Ag/AgCl in 1 mM TBABF4 Acetonitrile
compound
E1 2 (red1) (V)
/
E1 2 (red2) (V)
/
E1 2 (red3) (V)
/
M-1 s-1 56,57
.
3a
3b
3c
3d
-0.56
-0.65
-0.66
-0.67
-1.05
-1.19
-1.14
-1.19
-1.56
-2.40
-1.56
-2.48
Mass Spectral Analysis. Electrospray ionization (ESI) mass
spectrometric analysis of the decomposition products from the
acidified 3a was carried out to confirm the generation of the
radical intermediates. Product analysis of 3a before acidification
only yielded the parent ions of compound 3a; however, upon
treatment with an acid a variety of products were produced,
which is indicative of radical-radical reaction as shown in
Scheme 3 and Table 1. Based on Scheme 3, dimerization of
intermediates 13 or 14 yielded either compound 16 or 18, which
lends support to the observed mass at 471.7 m/z. The observed
mass of 427.1 m/z, which corresponds to the products 21 or
22, is evident of a mixed radical-radical reaction of intermedi-
ates 13 or 14 with 15. H-atom abstraction by the radicals 13,
14, and 15 gave compounds 17 and 20 and is consistent with
the observed masses at 259.2 m/z (M + Na) and 191.6 m/z,
respectively. Evidence of dimer 19 formation from the radical
intermediate 15 was observed with mass of 383.7 m/z. These
analyses reveal the formation of radical intermediates during
the decomposition path.
in the presence of PBN gave hfsc’s of aN ) 14.33 G and aH )
2.05 G, which are similar to those observed for the spin trapping
of low molecular weight thiyl radicals by PBN with hfsc’s of
aN ) 14.46 G and aH ) 1.53-2.09 G.54 Therefore, the radical
generated during acidification of aryloxathiazolyliumolates is
most probably an S-centered radical. Even though there is also
the possibility that an O-centered adduct is formed, an H-atom
migration is not likely to occur due to the endoergic nature of
this reaction (Table 4, reaction V).
No radical formation was observed at neutral pH using
DMPO as a spin-trap similar to when [(MGD)2-FeII] was used
(Figure 7). The DMPO spin adduct formation from 3a-3d was
quantified using a stable nitroxide, 3-CP, and is shown in Figure
6. The growth of the first low-field peak was monitored over a
period of 20 min. At the initial concentration of 20 mM, the
maximum concentration of DMPO-adduct formed from 3a is
the highest (6.5 mM) as compared to the amount of adduct
formed from 3b-3d with 3d giving the least amount of adduct
(55) Fujii, S.; Yoshimura, T. Coord. Chem. ReV. 2000, 198, 89.
(56) Davies, M. J.; Forni, L. G.; Shuter, S. L. Chem. Biol. Interact. 1987, 61,
177.
(57) Potapenko, D. I.; Bagryanskaya, E. G.; Tsentalovich, Y. P.; Reznikov, V.
A.; Clanton, T. L.; Khramtsov, V. V. J. Phys. Chem. B 2004, 108, 9315.
(54) Ionita, P.; Gilbert, B. C.; Whitwood, A. C. Lett. Org. Chem. 2004, 1, 70.
9
5508 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

4-Aryl-1,3,2-Oxathiazolylium-5-olates as NO-Donors
A R T I C L E S
Table 3. Aqueous Phase Reaction Enthalpies (∆H298K,aq) and Free Energies (∆G298K,aq) for Various Steps Leading to the Formation of
RSNO at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* Level in kcal/mol
∆H29 8K,aq
∆G29 8K,aq
I-A
I-B
I-C
I-D
I-E
II
I-A
I-B
I-C
I-D
I-E
II
3a (-CF3)
3c (-Cl)
8.9
8.1
6.4
7.2
20.5
19.9
18.9
14.9
16.2
15.4
14.6
14.9
-7.6
-11.1
-12.0
-16.7
-9.7
-10.7
-11.6
-12.1
-5.1
-4.7
-4.1
-3.3
9.0
8.0
6.3
7.6
17.5
16.8
16.0
37.9
14.8
13.9
13.1
13.7
-7.3
-10.7
-11.6
-16.2
-8.8
-10.1
-10.8
-11.3
-8.6
-8.3
-7.7
-6.9
3b (-H)
3d (-OCH3)
Table 4. Aqueous Phase Reaction Enthalpies (∆H298K,aq) and Free Energies (∆G298K,aq) for Various Steps after the Release of NO at the
PCM/B3LYP/6-31+G**//B3LYP/6-31G* Level in kcal/mol
∆H298K,aq
∆G29 8K,aq
III-A
III-B
III-C
IV
V
VI
VII
III-A
III-B
III-C
IV
V
VI
VII
3a (-CF3)
3c (-Cl)
42.2
43.5
44.2
44.7
32.4
32.1
31.5
30.9
69.6
59.8
59.7
46.8
3.2
3.9
1.3
4.2
9.5
8.3
11.7
6.8
-22.9
-20.2
-19.5
-15.5
-13.4
-11.8
-7.8
31.5
32.8
33.6
34.2
20.5
20.5
19.9
19.6
57.6
48.2
48.1
35.1
2.3
3.9
1.2
4.2
8.3
7.2
10.9
5.7
-33.2
-31.2
-30.1
-26.2
-24.9
-23.9
-19.3
-20.5
3b (-H)
3d (-OCH3)
-8.8
Reducibility of Aryloxathiazolyliumolates. The decomposi-
tion of 3a in the presence of two common bioreductants,58,59
L-ascorbic acid or reduced L-(-)-glutathione, was measured
spectrophotometrically. Because no change in the absorption
profile before and after the addition of the reductants was
observed, it can be concluded that 3a is stable in the presence
of these reductants. This observation was further confirmed
(58) Halliwell, B.; Gutteridge, J. M. C. Free Radicals in Biology and Medicine,
3rd ed.; Oxford University Press, Inc.: New York, 1999.
(59) Buettner, G. R. Arch. Biochem. Biophys. 1993, 300, 535.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5509

A R T I C L E S
Lu et al.
Figure 7. Formation and decay of S-centered DMPO adduct generated from 3a-3d (20 mM) in the presence of DMPO (200 mM) at pH 5.0 and 7.0. Arrow
indicates the peak that is being monitored. All data were the average of three measurements. The reproducibility of peak intensities is reasonable with a
standard deviation of less than 10%.
using EPR spin trapping, which did not show any radical adduct
formation when L-ascorbic acid or reduced L-(-)-glutathione
was introduced into the solution containing 3a and DMPO. Both
results from spectrophotometric and EPR experiments are
consistent with the reduction potential observed for 3a using
cyclic voltammetry as shown in Table 2. The observed first
reduction (E1/2) for 3a of -0.56 V is higher as compared to
those found for ascorbate and glutathione of -0.2858 and -0.26
V,60 respectively (the literature E1/2 values for other RSNO’s
are -0.80 to -1.14 V).61 The observed reduction potentials for
3b-3d range from -0.65 to -0.67 V, and, therefore, this series
of aryloxathiazolyliumolates (3a-3d) exhibit relative stability
in the presence of bioreductants. Stability in the presence of
metal ion was also explored, and results show that 3b (0.4 mM)
in the presence of 10 µΜ Fe2+ decomposed to about 16% over
a period 20 min, and this rate of decomposition is similar to
that of using GSNO (0.8 mM) in the presence of 20 µΜ Fe2+
(Figure S4 of the Supporting Information).
Based on our experimental evidence, aryloxathiazolyliumolate
compounds are relatively stable at neutral pH and NO is
produced only upon treatment with acid. Furthermore, our results
also suggest that the decomposition of aryloxathiazolyliumolate
is acid-catalyzed as shown in Figure 3 and that their NO-
releasing property is influenced by the substituents on the aryl
moiety. This computational study has two goals: (1) to
determine a plausible mechanism for the decomposition of
aryloxathiazolyliumolates; and (2) to determine whether aryl
substitution by electron-donating or electron-withdrawing groups
has an effect on the favorability of decomposition of arylox-
athiazolyliumolates. The PCM/B3LYP/6-31+G**//B3LYP/6-
31G* level of theory that takes into account the solvation effect
of water was employed. It has been reported that the B3LYP
density functional theory method overestimates S-N bond
lengths by 0.5-0.1 Å, and, therefore, the activation energies
obtained are lower than they should be.64 Nevertheless, we chose
this level of theory to achieve the best compromise between
computational cost and accuracy due to the size of our molecules
and the consideration of solvation effects in the calculation.
In this study, only the probable stable intermediates have been
calculated, and thus the values in Table 3 only represent the
favorability in reaction free energies (∆G). Based on the
assumption that the decomposition of 3 occurs after introduction
of H+, then protonation of 3 should be the first step to the
decomposition pathway and should also be energetically favor-
able (see Table 3).
Computational Results. Over the last 15 years, many
different decomposition pathways of RSNOs have been reported
such as unimolecular homolytic cleavage, metal-catalyzed, and
photolytic processes.62 However, it has been accepted that the
decomposition pathway is highly dependent on the reaction
conditions, while the factors that determine their stability in
solution are not yet fully understood.63 Nevertheless, ongoing
research has added to our understanding of the chemistry of
nitrosothiols; most recently, for example, a pathway has been
proposed where decomposition is catalyzed by nitrosonium.62
Optimization of aryloxathiazolyliumolates (3′) gave a struc-
ture with C-O bond length of 1.21 Å similar to those found in
esters with a CdO bond of 1.23 Å. It is, therefore, reasonable
to assume that the starting structure resembles that of lactone
compounds (see structure 3, Table 3). The favorability of
protonation on five possible sites of the pentacylic ring was
(60) Millis, K. K.; Weaver, K. H.; Rabenstein, D. L. J. Org. Chem. 1993, 58,
4144-6.
(61) Hou, Y.; Wang, J.; Arias, F.; Echegoyen, L.; Wang, P. G. Bioorg. Med.
Chem. Lett. 1998, 8, 3065.
(62) Zhao, Y.-L.; McCarren, P. R.; Houk, K. N.; Choi, B. Y.; Toone, E. J. J.
Am. Chem. Soc. 2005, 127, 10917.
(63) de Oliveira, M. G.; Shishido, S. M.; Seabra, A. B.; Morgon, N. H. J. Phys.
Chem. A 2002, 106, 8963.
(64) Baciu, C.; Gauld, J. W. J. Phys. Chem. A 2003, 107, 9946.
9
5510 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

4-Aryl-1,3,2-Oxathiazolylium-5-olates as NO-Donors
A R T I C L E S
(3d) and may not reflect the relative rates of reaction as observed
experimentally. Calculation of activation energy barriers (∆Gq)
for reactions IE and II would have provided a more definitive
trend on the relative rates of hydrolysis. Nevertheless, the overall
exoergic ∆G298K,aq values suggest that reactions IE and II are
the most preferred pathways for the decomposition of 3 under
acidic condition.
Shown in Table 4 are the enthalphies and free energies of
NO release from the hydrated aryloxathiazolyliumolates (reac-
tion III-B) with ∆H298K,aq values of 30.9-32.4 kcal/mol close
to the reported63,64,67 S-N bond dissociation energies of ∼30
kcal/mol for RSNOs with a benzyl group or a non-conjugated
alkyl system. The free energies (∆G298K,aq) of 19.6-20.5 kcal/
mol for NO release from compounds 9a-d were found to be
similar to those reported for related RSNO compounds with
values of 20-22 kcal/mol.64,68-70 The small differences in free
energies of reaction for step III-B (∼1 kcal/mol) indicate that
the substituents may have little effect on the favorability of NO
release from the hydrated compounds 9a-d.
The favorability of heterolytic S-N bond cleavage to form
the nitrosonium cation (NO+) (reaction III-A) and the ground
triplet nitroxyl (3NO-) (reaction III-C) from 9a-d was also
computationally explored to determine whether a more energeti-
cally favorable pathway exists for the decomposition of 9a-d.
Results indicate, however, that the formation of NO+ and 3NO-
from 9a-d is more endoergic as compared to the formation of
NO via homolytic cleavage (reaction III-B). The decomposition
pathways following release of NO (reactions V-VII) show that
the release of CO2 via reaction VII is the most preferred path
for the final product formation. The formation of the radical 15
is further supported by product analysis using MS (Scheme 3,
Table 1) as mentioned previously.
Vasodilation Studies. The ability of the aryloxathiazolyliu-
molate compounds to induce NO-dependent vasodilation was
investigated using rat aortic rings. Rings were stretched to
generate a resting tension of 1 g and allowed to equilibrate for
20 min. Following the equilibration period, the vessels were
constricted with phenylephrine (0.5 µM), and the relaxation
response to the aryloxathiazolyliumolate compounds was mea-
sured. The % relaxation was then compared among the drug-
treated groups. These results demonstrated that exposure to
compounds 3a-3d resulted in a pH-dependent relaxation with
maximal effects observed at pH 6.0 exhibiting 59%, 35%, 63%,
and 71% relaxation, respectively (Figure 9). The relaxation
response was significantly less pronounced at pH 7.4, in which
16%, 8%, 37%, and 44% relaxation was observed. These results
demonstrate that the aryloxathiazolyliumolates have biological
activity, and upon NO release induce vascular relaxation. The
pH dependence of the relaxation response is consistent with
our previous results demonstrating enhanced NO release in
acidic environments.
Figure 8. Lowest energy molecular orbitals of 3b, 6b, 7b, and 9b (red
lines) calculated at 6-31+G** in Gaussian 03 to show the π electron
delocalization between the aryl and pentacyclic ring systems and its
subsequent disruption upon hydration.
investigated (i.e., C-1, S-2, N-3, O-4, and O-6). Table 3 shows
that the protonations at the benzylic C-1 (reaction I-A), ester
O-4 (reaction I-C), and S-2 (reaction I-B) were all endoergic
by ∼7, ∼14, and ∼17 kcal/mol, respectively. Protonation on
the carbonyl O-6 (reaction I-E) or N-3 (reaction I-D) was found
to be exoergic by 7-16 kcal/mol. These results correlate with
the natural bond orbital (NBO) analysis65 of the charges on the
relevant atoms of the pentacyclic SNO of 3a-d showing the
highest charge density on the carbonyl O-6 with -0.65 e
followed by the benzylic C-1 (-0.36 e to -0.38 e), N-3 (-0.32
e to -0.36 e), and O-4 (-0.33 e to -0.34 e) with S-2 being
the most positive with a charge density of 0.88 e-0.92 e. It is
also interesting to note that the endoergic protonation on C-1,
O-4, and S-2 resulted in the opening of the pentacyclic ring,
while this was not observed upon protonation at O-6 and N-3.
Because the energetics of protonation at the O-6 and N-3 are
almost the same and no pentacyclic ring opening occurred, it
can be assumed that both of the protonated species, 6 and 7,
could exist in solution. Molecular orbital analyses of the starting
molecules 3a-d showed π electron conjugation between the
pentacyclic SNO ring and the aryl system (see Figure 8 for 3b).
Because of this extensive π electron delocalization between the
two ring systems, the aryl substituents can exert an electronic
effect on the pentacyclic ring. Both the O-6 and the N-3
protonation follow a trend that is consistent with the electron-
donating or -withdrawing capabilities of the aryl substituents:
ease of protonation follows -OCH3 > -H > -Cl > -CF3.
Moreover, the π electron conjugation is conserved upon
protonation of the carbonyl O to yield compounds 6a-d.
We hypothesize that the mechanism for decomposition of
6a-d would resemble that of ester hydration, where the rate-
determining step would be the hydration step.66 Because of the
high positive charge density on the C-5 atom (0.73 e) (see Table
3, reaction II), the most favorable site for nucleophilic addition
of H2O to compounds 6a-d is at the C-5 position to give 9a-
d. Reaction II shows pentacyclic ring opening with loss of π
electron conjugation on 9a-d with exoergic free energies of
reaction (∆G298K,aq) that range from -6.9 to -8.6 kcal/mol.
Substituent effect has also been observed for reaction II with
the -CF3 compound being the most susceptible to water
addition, and -OCH3 being the least susceptible.
Experimental Section
General Procedures. All reagents were purchased from commercial
sources and used without further purification. 1H NMR and 13C NMR
(67) Lue, J.-M.; Wittbrodt, J. M.; Wang, K.; Wen, Z.; Schlegel, H. B.; Wang,
P. G.; Cheng, J.-P. J. Am. Chem. Soc. 2001, 123, 2903.
(68) Arulsamy, N.; Bohle, D. S.; Butt, J. A.; Irvine, G. J.; Jordan, P. A.; Sagan,
E. J. Am. Chem. Soc. 1999, 121, 7115.
The overall ∆G298K,aq values (reaction IE + reaction II) are
(in kcal/mol): -17.4 (3a), -18.5 (3b), -18.4 (3c), and -18.2
(69) Grossi, L.; Montevecchi, P. C. Chem.-Eur. J. 2002, 8, 380.
(70) Toubin, C.; Yeung, D. Y. H.; English, A. M.; Peslherbe, G. H. J. Am.
Chem. Soc. 2002, 124, 14816.
(65) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899.
(66) Bender, M. L. Chem. ReV. 1960, 60, 53.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5511

A R T I C L E S
Lu et al.
General Synthesis of 4-(Aryl)-1,3,2-oxathiazolylium-5-olate (3a-
3d). Compound 2a, 2b, 2c, or 2d (7-8 mmol) was dissolved in
anhydrous MeOH (120 mL). The solution was flushed with Ar for 5
min, and then sodium methoxide was gradually added until the pH
reached 9-10. After the mixture was stirred for 5 h under N2 at room
temperature, the acid resin (Amberlyst 15 ion-exchange resin) was
added to ensure a pH of 2-3. The resin was then removed by filtration,
and the solvent was evaporated. The residue was dissolved in dry CH2-
Cl2 (50 mL) and was allowed to cool in an ice bath. Isobutylnitrite
(1.9 mL, 16 mmol) was added and stirred for 2 h while the mixture
was protected from light. The reaction mixture was diluted with dry
CH2Cl2 (200 mL), and DCC (5.00 g, 24 mmol) was added. The mixture
was stirred for an additional 2 h at 0 °C, and H2O (0.3 mL) was added
to quench the reaction. The reaction mixture was passed through a Celite
pad, and the filtrate was concentrated and purified by silica gel column
chromatography using EtOAc/hexanes 1:10 as the eluent.
Figure 9. Vasodilation effects of 20 µM of 3a-3d at pH 6.0 and 7.4.
spectra were recorded on a 400 MHz spectrometer. Chemical shifts
(δ) are given in ppm relative to tetramethylsilane (TMS) as internal
standard. Silica gel plates (Merck F254) and silica gel 60 (Merck, 70-
230 mesh) were used for thin-layer chromatography (TLC) and column
chromatography, respectively. Molecular weight was verified using an
ESI mass spectrometer.
General Synthesis of S-Acetyl-r-aryl-r-mercaptoacetic Acids
(2a-2d). A mixture of compound 1a, 1b, 1c, or 1d (10 mmol) and
thiolacetic acid (20 mmol) in THF (25 mL) was added dropwise to a
stirred THF solution (50 mL) of the preformed adduct of triphenyl
phosphate (20 mmol) and diisopropyl azodicarboxylate (20 mmol) at
0 °C. After the addition was completed, the mixture was stirred at
0 °C until the solution turned clear. The ice bath was removed, and
stirring was continued for 1 h at room temperature. The solution was
evaporated, and the residue was dissolved in CH2Cl2 (50 mL) and
extracted twice with 1 M NaHCO3 solution (30 mL). The basic aqueous
layer was washed with CH2Cl2 (20 mL × 3). The aqueous layer was
acidified with 12 M HCl to pH ≈ 3, and then extracted with CH2Cl2
(20 mL × 3). The combined organic layers were dried over MgSO4,
and the solvent was evaporated.
4-(p-Trifluoromethylphenyl)-1,3,2-oxathiazolylium-5-olate (3a).
For compound 2a, 2.27 g, 8.17 mmol was used. Compound 3a (1.16
1
g, 4.70 mmol, yield 58%) was obtained as yellow needle crystals. H
NMR (500 MHz, CDCl3): δ 8.11 (d, J ) 8.3 Hz, 2H), 7.88 (d, J )
8.3 Hz, 2H). 13C NMR (100 MHz, CDCl3): δ 174.7, 132.0 (q, J )
33.0 Hz), 129.7, 126.5 (q, J ) 4.0 Hz), 126.1, 123.5 (q, J ) 270.0
Hz), 119.1. HRMS (M + Na+) (ESI) calcd for C9H4F3NO2SNa+
269.9813, found 269.9813.
4-Phenyl-1,3,2-oxathiazolylium-5-olate (3b). For compound 2b,
1.68 g, 8.0 mmol was used. Compound 3b (916 mg, 5.02 mmol, yield
1
64%) was obtained as yellow needle crystals. H NMR (400 MHz,
CDCl3): δ 7.89 (d, J ) 7.8 Hz, 2H), 7.51-7.48 (m, 3H). 13C NMR
(100 MHz, CDCl3): δ 174.9, 130.0, 129.2, 126.6, 126.0, 120.4. HRMS
(M + Na+) (ESI) calcd for C8H5NO2SNa+ 201.9939, found 201.9941.
4-(p-Chlorophenyl)-1,3,2-oxathiazolylium-5-olate (3c). For com-
pound 2c, 1.76 g, 7.2 mmol was used. Compound 3c (1.16 g, 4.70
1
mmol, yield 58%) was obtained as yellow needle crystals. H NMR
(400 MHz, CDCl3): δ 7.92 (d, J ) 8.3 Hz, 2H), 7.59 (d, J ) 8.3 Hz,
2H). 13C NMR (100 MHz, CDCl3) δ 175.4, 134.9, 129.7, 128.1, 126.1,
120.0, 83.7. HRMS (M + Na+) (ESI) calcd for C8H4ClNO2SNa+
235.9545, found 235.9540.
S-Acetyl-r-(p-trifluoromethylphenyl)-r-mercaptoacetic Acid (2a).
The crude product 2a (2.27 g, 8.17 mmol, yield 82%) was obtained
1
4-(p-Methoxyphenyl)-1,3,2-oxathiazolylium-5-olate (3d). For com-
pound 2d, 1.75 g, 7.3 mmol was used. Compound 3d (823 mg, 3.94
and used for the deprotection without further purification. H NMR
(500 MHz, CDCl3): δ 10.71 (brs, 1H), 7.59 (d, J ) 8.3 Hz, 2H), 7.51
(d, J ) 8.3 Hz, 2H), 5.35 (s, 1H), 2.36 (s, 3H). 13C NMR (125 MHz,
CDCl3): δ 193.1, 174.7, 138.3, 130.9 (q, J ) 32.8 Hz), 129.0, 125.9
(q, J ) 3.6 Hz), 123.7 (q, J ) 272.3 Hz), 50.2, 30.0. HRMS (M +
Na+) (ESI) calcd for C11H9F3O3SNa+ 301.0122, found 301.0120.
S-Acetyl-r-phenyl-r-mercaptoacetic Acid (2b). The crude product
2b (1.68 g, 8.0 mmol, yield 80%) was directly used for the deprotection
without further purification. 1H NMR (400 MHz, CDCl3): δ 10.28 (brs,
1H), 7.41-7.32 (m, 5H), 5.31 (s, 1H), 2.36 (s, 3H). 13C NMR (100
MHz, CDCl3): δ 194.0, 175.7, 134.2, 129.1, 128.9, 128.7, 51.0, 30.1.
HRMS (M + Na+) (ESI) calcd for C10H10O3SNa+ 233.0248, found
233.0247.
1
mmol, yield 54%) was obtained as yellow needle crystals. H NMR
(250 MHz, CDCl3): δ 8.87 (d, J ) 8.9 Hz, 2H), 8.09 (d, J ) 8.9 Hz,
2H). 13C NMR (63 MHz, CDCl3): δ 175.0, 160.7, 127.7, 121.1, 119.1,
114.8, 55.5. HRMS (M + Na+) (ESI) calcd for C9H7NO3SNa+
232.0044, found 232.0046.
Decomposition Studies of 3a-3d. In each study, UV-vis absor-
bance reading of the solvent was used to correct the actual readings
from the samples. All solutions were protected from light by covering
the exposed parts of the cuvette with aluminum foil, and the absorbance
reading was monitored at the wavelengths of 580, 587, 592, and 602
nm for 3a-3d, respectively, at 37 °C. Acidic condition: 0.8 mM
solutions of compounds 3a-3d were prepared in a mixture of 0.1 M
phosphate buffer (882 µL, pH 5.0) and acetonitrile (98 µL). The
absorbances of 3a, 3b, 3c, or 3d were immediately monitored over a
period of 5, 40, 80, or 500 min, respectively. Neutral condition: 0.4
mM solution of 3b was prepared in 0.1 M Tris-HCl buffer (882 µL,
pH 7.0) and acetonitrile (98 µL) solvent system. The absorbance was
monitored for 20 h. Acetonitrile: Decomposition studies of 3a-3d (0.4
mM) in acetonitrile were also carried out, and absorbance readings were
monitored for 2 h. Presence of room light: 1 mL of a solution of 0.4
mM 3b in 0.1 M phosphate buffer (882 µL, pH 7.0) and acetonitrile
(98 µL) was exposed to room light, and the decay of absorbance was
monitored over a period of 3 h. Similar experiment was done using a
0.8 mM solution of S-nitrosoglutathione (GSNO) using the same solvent
system. Presence of Fe2+: Fe(NH4)2(SO4)2 (10 µM) solution was added
to a solution of 3b (0.4 mM) in PBS buffer. Likewise, Fe(NH4)2(SO4)2
(20 µM) solution was added to a solution of GSNO (0.8 mM) in PBS
S-Acetyl-r-(p-chlorophenyl)-r-mercaptoacetic Acid (2c). Com-
pound 2c (1.76 g, 7.2 mmol, yield 72%) was obtained as a white solid
after purification using a silica gel column (EtOAc/hexanes 1:5 (0.5%
1
HOAc) as eluent). H NMR (500 MHz, CDCl3): δ 11.43 (brs, 1H),
7.32 (d, J ) 8.8 Hz, 2H), 7.29 (d, J ) 8.8 Hz, 2H), 5.27 (s, 1H), 2.34
(s, 3H). 13C NMR (100 MHz, CDCl3): δ 193.5, 175.4, 134.7, 132.7,
129.5, 128.6, 50.1, 30.0. HRMS (M + Na+) (ESI) calcd for C10H9-
ClO3SNa+ 266.9849, found 266.9846.
S-Acetyl-r-(p-methoxyphenyl)-r-mercaptoacetic Acid (2d). Com-
pound 2d (1.75 g, 7.3 mmol, yield 73%) was obtained as a white solid
after purification by silica gel column chromatography (EtOAc/hexanes
1
1:2 (0.5% HOAc) as eluent). H NMR (400 MHz, CDCl3): δ 9.60
(brs, 1H), 7.32 (d, J ) 8.7 Hz, 2H), 6.86 (d, J ) 8.7 Hz, 2H), 5.26 (s,
1H), 3.79 (s, 3H), 2.35 (s, 3H). 13C NMR (100 MHz, CDCl3): δ 194.1,
175.5, 159.9, 129.7, 125.9, 114.4, 55.3, 50.3, 29.9. HRMS (M + Na+)
(ESI) calcd for C11H12O4SNa+ 263.0354, found 263.0350.
9
5512 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

4-Aryl-1,3,2-Oxathiazolylium-5-olates as NO-Donors
A R T I C L E S
buffer. The decay of absorbances at wavelengths 587 and 335 nm was
monitored for 3b and GSNO, respectively, over a period of 20 min.
Acid Catalysis of 3b. Various solutions at pH 5.0, 5.5, 6.0, 6.5,
and 7.0 of 0.4 mM 3b in 0.1 M phosphate buffer (882 µL) and
acetonitrile (98 µL) were used. The absorbance of 3b was immediately
monitored over a period of 10 min for each solution. The rate constant
for the decomposition was obtained by plotting the ln of concentration
versus time for the first 5 min, and the slope was calculated using the
equation ln[3b] ) -k·t ) -k′[H+]·t.
EPR Spin Traps. Materials. The spin trap, 5,5-dimethyl-1-pyrroline
N-oxide (DMPO) and N-benzylidene-tert-butylamine N-oxide (PBN)
>99%, were obtained commercially and showed no paramagnetic
impurities. Sodium N-methyl-D-glucamine dithiocarbamate (MGD) was
synthesized using the procedure developed by Shinobu et al.71
Phosphate-buffer saline (pH 7.0) was used with 100 µM diethylenetri-
aminepentaacetic acid (DTPA) as a metal chelating agent for spin
trapping studies using DMPO and PBN.
quantified using 1-8 mM (at increment of 2 mM) solutions of 2,2,5,5-
tetramethyl-3-pyrrolin-1-oxyl-3-carboxylic acid (3-CP) in PBS. Spin
trapping of thiyl radical intermediate generated from 3b was also carried
out using R-phenyl-tert-N-butyl nitrone (PBN) as spin trap. The pH of
PBN (200 mM) and 3b (20 mM) PBS solution was adjusted to pH ∼5
by addition of 1 M H2SO4, and the formation of spin adduct was
monitored by EPR. The EPR spectrum was also obtained using PBS
solution of PBN (200 mM), H2O2 (20 mM), and ∼100 µM FeCl2 for
comparison.
Stability Studies. Stability of compounds 3a-3d in the presence
of ascorbic acid and glutathione was investigated. EPR measurements
were carried out using PBS solution of 200 mM DMPO, 20 mM 3a-
3d, 1 mM L-ascorbic acid, or 1 mM reduced L-(-)-glutathione.
Spectrophotometric analysis of the reaction mixture was carried out
using a solution of 0.5 mM 3a and 0.1 mM l-ascorbic acid. Peak decay
at 581 nm was monitored. This procedure was repeated using 0.1 mM
reduced glutathione.
Mass Spectral Analysis. Electrospray ionization (ESI) mass spec-
trometric analyses were performed immediately on the acidified and
nonacidified solutions of 3a (0.1 mM) in DMSO. Mass analyses were
performed with a 3-tesla Fourier transform mass spectrometer in positive
ion detection.
Preparation of [(MGD)2-FeII] Complex. Freshly prepared
[(MGD)2-FeII] was used in all of the NO-trapping studies by dissolving
MGD (790 mg, 3 mmol) in 10 mL of distilled water. The solution was
then purged with Ar gas for ∼5-10 min, and Fe(NH4)2(SO4)2·6H2O
(390 mg, 1 mmol) was added. [(MGD)2-FeII] gave a clear yellowish
solution, which oxidizes to a dark brown solution over time.49
EPR Measurements. EPR measurements were carried out at room
temperature on an EMX X-band spectrometer equipped with HS
resonator. General instrument settings are as follows. NO-trapping:
microwave power, 20 mW; modulation amplitude, 4.00 G; receiver
gain, 2.00 × 105; scan time, 42 s; time constant, 82 ms; sweep width
80 G. S-centered radical trapping: microwave power, 20 mW;
modulation amplitude, 1.00 G; receiver gain, 2.00 × 105; scan time,
42 s; time constant, 164 ms; sweep width 100 G. Measurements were
performed using a 50 µL glass capillary tube.
Spin Trapping of Nitric Oxide from 3a-3d. Method I. Purging
with argon: Nitric oxide was generated in a 5 mL conical flask by
adding 20 mM DMSO solution of 3a-3d to a PBS buffer containing
100 µM DTPA. It should be noted that the pH of PBS with 50% DMSO
alone is ∼10. The pH was carefully adjusted to ∼5.0 by adding 1 M
H2SO4. The flask was covered with a rubber septa and purged with Ar
using a needle syringe. The purged gas was allowed to flow through a
tube connected to a separate reaction vessel containing 1 mL of 10
mM [(MGD)2-FeII]. EPR spectrum was obtained for each 50 µL of
[(MGD)2-FeII] solution taken at various time intervals. This procedure
was repeated without H2SO4 as the acid control. Method II. Direct
mixing: In a typical experiment, the NO adduct was generated from
50 µL of PBS solution containing [(MGD)2-FeII] (20 mM) and 3a-
3d (20 mM, note: stock solutions of 3a-3d are in DMSO). The pH
of the solution was carefully adjusted to ∼5 by adding 1 M H2SO4.
The mixture was transferred to a 50 µL capillary tube and was used
for EPR measurements. The formation of the low-field peak intensity
was monitored over a period of 13 min. This procedure was repeated
without acidification. Quantitation of the intensities was performed using
the signal intensity of the first low-field peak arising from the
[(MGD)2-FeII-NO] adduct, which was generated from 0.6 to 6.0 mM
(with an increment of 0.6 mM) of S-nitroso-N-acetylpenicillamine
(SNAP) in the presence of [(MGD)2-FeII].
Cyclic Voltammetry Measurements. Cyclic voltammetry was
performed using a potentiostat, and Ag/AgCl (0.01 M) as reference
electrode, platinum electrode as the working electrode, and an auxiliary
electrode. The concentration of 3a-3d used was 0.1 mM in 1 mM
tetrabutylammonium tetrafluoroborate (TBABF4) in acetonitrile.
General Computational Methods. The initial conformational search
was carried out using the OPLS2001 force field of the Macromodel/
Maestro software package. Density functional theory72,73 was applied
in this study to determine the optimized geometry, vibrational frequen-
cies, and single-point energy of all stationary points.74-77 To include
solvation effects into our calculations, polarizable continuum model
(PCM)78-82 calculation was performed on the single-point energies. All
calculations were done using Gaussian 0383 at the Ohio Supercomputer
Center. Single-point energies were obtained at the B3LYP/6-31+G**
level based on the optimized B3LYP/6-31G* geometries. Natural
population analyses (NPA)84 were also performed at the B3LYP/6-
31+G** level of theory. These basis set calculations used the standard
six Cartesian d functions. Stationary points for all compounds have
zero imaginary vibrational frequencies as derived from a vibrational
frequency analysis (B3LYP/6-31G*). A scaling factor of 0.9806 was
used for the zero-point vibrational energy (ZPE) corrections with the
B3LYP/6-31G* level of theory.85 Spin contamination for all of the
stationary point of the radical structures was negligible, that is, 0.75 <
S2 < 0.78.
Vascular Reactivity Study. Contraction and relaxation of isolated
aortic rings were measured in an organ bath containing modified
Krebs-Henseleit solution (118 mM NaCl, 24 mM Na2HCO3, 4.6 mM
KCl, 1.2 mM NaH2PO4, 1.2 mM CaCl2, 4.6 mM HEPES, and 18 mM
glucose) aerated with 95% CO2-5% O2, 37 °C, and pH 6.5. Aortic
(72) Labanowski, J. W.; Andzelm, J. Density Functional Methods in Chemistry;
Springer: New York, 1991.
(73) Parr, R. G.; Yang, W. Density Functional Theory in Atoms and Molecules;
Oxford University Press: New York, 1989.
(74) Becke, A. D. Phys. ReV. 1988, 38, 3098.
(75) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(76) Becke, A. D. J. Chem. Phys. 1993, 98, 1372.
(77) Hehre, W. J.; Radom, L.; Schleyer, P. V.; Pople, J. A. Ab Initio Molecular
Orbital Theory; John Wiley & Sons: New York, 1986.
(78) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027.
(79) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996,
255, 327.
(80) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210.
(81) Barone, V.; Bencini, A.; Cossi, M.; Di Matteo, A.; Mattesini, M.; Totti, F.
J. Am. Chem. Soc. 1998, 120, 7069.
(82) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999.
(83) Frisch, M. J.; et al. Gaussian 03, revision B.04; Gaussian, Inc.: Pittsburgh,
PA, 2003.
(84) Reed, A. E.; Curtiss, L. A.; Weinhold, F. A. Chem. ReV. 1988, 88, 899.
(85) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
Spin Trapping of Sulfur-Centered Radical Intermediate from
Compounds 3a-3d. Spin trapping of S-centered radical intermediate
was carried out using DMPO. In a typical experiment, the pH of 50
µL of PBS buffer solution containing DMPO (200 mM) and 3a, 3b,
3c, or 3d (20 mM) was adjusted to pH ∼5 by addition of 1 M H2SO4.
The formation of the low-field EPR peak intensity was monitored over
a period of 20 min using the time-sweep setting. This procedure was
repeated without acidification of the solution. Peak intensities were
(71) Shinobu, L. A.; Jones, S. G.; Jones, M. M. Acta Pharm. Toxicol. 1984, 54,
189.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5513

A R T I C L E S
Lu et al.
rings were cut into 3 mm segments and mounted on a wire myograph
(Danish Myo, GA). Contraction was measured via a force transducer
interfaced with Chart software for data analysis. Following a 30 min
equilibration period, the rings were stretched to generate a tension of
1.0 g. The optimum resting force of the aortic rings was determined
by comparing the force developed by 40 mM KCl under varying resting
force. The effects of aryloxathiazolyliumolates on vascular relaxation
were then determined following phenylephrine (0.5 µM)-induced
contraction. The effects of 20 µM of compounds 3a-3d on vascular
relaxation were then measured. Results represent the mean ( SD from
n ) 4 groups.
These aryloxathiazolyliumolates were also found to be stable
in the presence of biological reductants. Vasodilation studies
on rat aortic rings revealed that vascular relaxation in the
presence of aryloxathiazolyliumolates is pH dependent where
a higher vasodilating effect was observed at pH 6 than at pH
7.4. This comprehensive study demonstrated the potential
application of aryloxathiazolyliumolates as future pro-drugs of
S-nitrosothiols for the controlled NO release in biological
systems.
Conclusion
Acknowledgment. This work is supported by a grant from
the NIH (GM54074). F.A.V. acknowledges NIH grant HL081248.
J.L.Z. acknowledges support from NIH grants HL38324,
HL63744, and HL65608. The Ohio Supercomputer Center
(OSC) is acknowledged for their generous computational support
of this research. We acknowledge the Center for EPR Spec-
troscopy and Imaging at OSU. We also wish to thank Mr. Jie
Shen for helpful discussions.
In summary, a novel series of aryloxathiazolyliumolates were
developed whose NO-releasing property was controlled by the
pH of their environment. Para-substitution of the aryl moiety
with electron-withdrawing groups showed higher NO-releasing
capabilities as compared to electron-donating substituents. The
decomposition pathway was supported by detection of NO and
S-centered radical intermediates using EPR spin-trapping. Also,
product analysis using mass spectrometry further confirmed the
formation of the various types of radical intermediates.
Computational studies suggest that the most plausible pathway
for aryloxathiazolyliumolates decomposition is via acid-
catalyzed concerted nucleophilic addition of water and ring
opening of the RSNO pentacylic ring, similar to that observed
in ester hydrolysis, and is followed by homolytic cleavage of
S-N bond to form NO.
Supporting Information Available: Complete ref 83. NMR,
EPR, MS, and cyclic voltammetry spectra. Geometries, energies,
enthalpies, and free energies for all compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA0682226
9
5514 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007