Hydrolysis and photolysis of 4-acetoxy-4-(benzothiazol-2-yl)-2,5- cyclohexadien-1-one, a model anti-tumor quinol ester

Article abstract of DOI:10.1021/jo9008436

(Chemical Equation Presented) 4-Acetoxy-4-(benzothiazol-2-yl)-2,5- cyclohexadien-1-one, 1, a quinol derivative that exhibits significant anti-tumor activity against human breast, colon, and renal cancer cell lines, undergoes hydrolysis in aqueous solution

Full text of DOI:10.1021/jo9008436

Hydrolysis and Photolysis of
4-Acetoxy-4-(benzothiazol-2-yl)-2,5-cyclohexadien-1-one,
a Model Anti-Tumor Quinol Ester
Yue-Ting Wang,† Kyoung Joo Jin,† Lauren R. Myers,† Stephen A. Glover,‡ and
Michael Novak*,†
Department of Chemistry and Biochemistry, Miami UniVersity, Oxford, Ohio 45056, and DiVision of
Chemistry, School of Science and Technology, UniVersity of New England, Armidale, 2351,
New South Wales, Australia
noVakm@muohio.edu
ReceiVed April 23, 2009
4-Acetoxy-4-(benzothiazol-2-yl)-2,5-cyclohexadien-1-one, 1, a quinol derivative that exhibits significant
anti-tumor activity against human breast, colon, and renal cancer cell lines, undergoes hydrolysis in aqueous
-
solution to generate an oxenium ion intermediate, 3, that is selectively trapped by N3 in an aqueous
environment. The 4-(benzothiazol-2-yl) substituent slows the rate of ionization of 1 compared to analogues
with 4-phenyl or 4-(p-tolyl) substituents, 4a or 4b. However, once generated, 3 is somewhat more selective
than the 4-phenyl-substituted cation 5a. Calculations performed at the B3LYP/6-31G(d) level agree that
the 4-(benzothiazol-2-yl) substituent does significantly stabilize 3. The structure of the major isolated
azide adduct, 4-(6-azidobenzothiazol-2-yl)phenol, 9, confirms that the positive charge is highly delocalized
in 3. The results of hydrolysis of 1 show that the 4-(benzothiazol-2-yl) substituent has a significant inductive
electron-withdrawing effect as well as a significant resonance effect that is electron-donating. Photolysis
of 1 in aqueous solution generates the quinol 2 as one of several photolysis products. The presence of
the quinol suggests that photolysis also leads, in part, to generation of 3, but photoionization of 1 is
significantly less efficient than is the case for the esters 4a and 4b. This study proves that 3 is generated
by ionization of 1 in an aqueous environment. A significant number of other 2-benzothiazole derivatives
that are not quinols, including ring-substituted derivatives of 2-(4-aminophenyl)benzothiazole 15, are
under development as anti-tumor agents as well. The possible generation of the reactive intermediate 17
by hydrolysis of the putative metabolite 16 is under investigation.
Introduction
human colon cancer cell lines HCT 116 and HT 29 as well as
against the breast cancer cell lines MCF-7 and MDA 468 with
IC50 values at the 40-800 nM level.1,2 Selective activity for
The 4-(benzothiazol-2-yl)-substituted quinol ester 1 and the
parent quinol 2 have been shown to have activity against the
† Miami University.
(1) Wells, G.; Bradshaw, T. D.; Diana, P.; Seaton, A.; Shi, D.-F.; Westwell,
A. D.; Stevens, M. F. G. Bioorg. Med. Chem. Lett. 2000, 10, 513–515.
‡ University of New England.
10.1021/jo9008436 CCC: $40.75
Published on Web 05/28/2009
2009 American Chemical Society
J. Org. Chem. 2009, 74, 4463–4471 4463

Wang et al.
both compounds against renal and colon cancer cell lines was
discovered on the NCI Developmental Therapeutics Screening
Program in vitro screen against 60 human cancer cell lines.2 In
vivo activity against the human renal tumor xenograft RXF
944LX in mice was also demonstrated.2 Structurally related
quinol derivatives, including some indol-2-yl and benzimidazol-
2-yl derivatives, have also shown activity against breast, colon,
and renal cancer cell lines.3-5 Thioredoxin has been identified
as one target of 2 and related quinols.6,7 Irreversible binding of
2 to the protein has been demonstrated, but the chemical nature
of the reaction has not been studied in detail.6 Other possible
protein targets of 2 have been identified, but nonprotein targets
have not be elucidated.8
constants of substituted benzoic acids.14,15 This substituent may
cause a significant deceleration of the ionization process that
has previously been shown to be sensitive to electron-withdraw-
ing substituents.11 It is not clear which effect of the benzothiazol-
2-yl group would be dominant in the chemistry of 1. If the
electron-withdrawing inductive effect destabilizes the transition
state leading to ionization sufficiently, it is possible that the
chemistry of 1 is dominated by reactions that do not generate
3. On the other hand, if the transition state for ionization is
sufficiently cation-like, the ability of the 2-benzothiazolyl group
to delocalize the positive charge via resonance may dominate.
In this paper, we provide evidence that both of the expected
electronic properties of the benzothiazol-2-yl group play a role
in the chemistry of 1. Ionization to generate 3 does occur,
although the rate of the reaction is significantly slower than that
of related quinol esters 4a and 4b with electron-donating 4-aryl
substituents. Once generated, 3 is a fairly selective ion that is
-
readily trapped by N3 at [N3-] < 0.05 M.
Previously, we have shown that 4a and 4b can be photolyzed
in aqueous solution to generate the oxenium ions 5a and 5b,
although this process occurs in competition with homolysis to
generate the corresponding radicals 6a and 6b.12,13 The ho-
molysis occurs exclusively in the less polar solvent CH3CN.13
Photolysis products of 1 in aqueous solution and in CH3CN
can also be explained, in part, by a combination of heterolysis
and homolysis, although some products appear to be generated
by other processes.
The irreversible inhibition of thioredoxin is thought to occur
by a double conjugate addition of two cysteine residues in the
active site of the protein directly on 2.6 However, our experience
with quinols closely related to 2 shows that they are not very
reactive with nucleophiles.9-13 Quinol esters similar to 1 have
recently been shown to generate oxenium ions in aqueous
solution.9-13 These cations are reactive with water but still react
very selectively with strong nucleophiles in an aqueous
environment.9-13 An alternative hypothesis to the direct reaction
of 2 with biological nucleophiles would be the reaction of the
oxenium ion 3, generated by ionization of 1, a likely metabolite
of 2, with these nucleophiles. This hypothesis relies on the ability
of the benzothiazol-2-yl group to stabilize 3 by delocaliza-
tion of the positive charge. Qualitatively, this seems likely, and
calculations performed at the B3LYP/6-31G(d) level presented
herein support that assumption. However, the benzothiazol-2-
yl group is a known inductive electron-withdrawing substituent
with a σp of 0.29 determined from its effect on the ionization
Results and Discussion
We have previously used ground state singlet energies of
aryloxenium ions to calculate ∆E of isodesmic reactions such
as eq 1.11,16 We have successfully correlated these ∆E with the
experimental lifetimes of 4′-substituted biphenylyloxenium
ions.11 Similar calculations have been used to understand the
properties of other aryloxenium and arylnitrenium ions.9,10,16,17
∆E of eq 1 measures the thermodynamic stability of 5a and 5b
relative to their respective hydration products 7a and 7b. At
the pBP/DN*//HF/6-31G(d) and BP/6-31G(d)//HF/6-31G(d)
levels, ∆E for eq 1 without ZPE corrections is 3.41 and 5.46
kcal/mol, respectively.11 The calculated stabilization of 5b
relative to 5a agrees well with the observed ca. 14-fold greater
lifetime of 5b in aqueous solution at 30 °C (170 ns for 5b vs
12 ns for 5a) in a correlation of log(lifetime) versus ∆E for
four 4′-substituted biphenylyloxenium ions.11 A similar cor-
relation of log(lifetime) versus ∆E was observed for a much
larger series of arylnitrenium ions and their hydration products
at the HF/6-31G(d)//3-21G level of theory.17
(2) Wells, G.; Berry, J. M.; Bradshaw, T. D.; Burger, A. M.; Wang, B.;
Westwell, A. D.; Stevens, M. F. G. J. Med. Chem. 2003, 46, 532–541.
(3) Berry, J. M.; Bradshaw, T. D.; Fichtner, I.; Ren, R.; Schwalbe, C. H.;
Wells, G.; Chew, E.-H.; Stevens, M. F. G.; Westwell, A. D. J. Med. Chem.
2005, 48, 639–644.
(4) McCarroll, A. J.; Bradshaw, T. D.; Westwell, A. D.; Matthews, C. S.;
Stevens, M. F. G. J. Med. Chem. 2007, 50, 1707–1710.
(5) Lion, C. J.; Matthews, C. S.; Wells, G.; Bradshaw, T. D.; Stevens,
M. F. G.; Westwell, A. D. Bioorg. Med. Chem. Lett. 2006, 16, 5005–5008.
(6) Bradshaw, T. D.; Matthews, C. S.; Cookson, J.; Chew, E.-H.; Shah, M.;
Bailey, K.; Monks, A.; Harris, E.; Westwell, A. D.; Wells, G.; Laughton, C. A.;
Stevens, M. F. G. Cancer Res. 2005, 65, 3911–3919.
(7) Mukherjee, A.; Huber, K.; Evans, H.; Lakhani, N.; Martin, S. Br. J.
Pharmacol. 2007, 151, 1167–1175.
(8) Chew, E.-H.; Matthews, C. S.; Zhang, J.; McCarroll, A. J.; Hagen, T.;
Stevens, M. F. G.; Westwell, A. D.; Bradshaw, T. D. Biochem. Biophys. Res.
Commun. 2006, 346, 242–251.
(9) Novak, M.; Glover, S. A. J. Am. Chem. Soc. 2004, 126, 7748–7749.
(10) Novak, M.; Glover, S. A. J. Am. Chem. Soc. 2005, 127, 8090–8097.
(11) Novak, M.; Poturalski, M. J.; Johnson, W. L.; Jones, M. P.; Wang, Y.-
T.; Glover, S. A. J. Org. Chem. 2006, 71, 3778–3785.
(12) Wang, Y.-T.; Wang, J.; Platz, M. S.; Novak, M. J. Am. Chem. Soc.
2007, 129, 14566–14567.
(13) Wang, Y.-T.; Jin, K. J.; Leopold, S. H.; Wang, J.; Peng, H.-L.; Platz,
M. S.; Xue, J.; Phillips, D. L.; Glover, S. A.; Novak, M. J. Am. Chem. Soc.
2008, 130, 16021–16030.
(14) Bystrov, V. F.; Belaya, Zh. N.; Gruz, B. E.; Syrova, G. P.; Tolmachev,
A. I.; Shulezhko, L. M.; Yagupol’skii, L. M. Zh. Obshchei Khimii 1968, 38,
1001–1005.
At the B3LYP/6-31G(d) level of DFT, oxenium ions 3, 5a,
and 5b all have singlet ground states. Their corresponding triplet
(15) Durmis, J.; Karvas, M.; Manasek, Z. Collect. Czech. Chem. Commun.
1973, 38, 215–223.
(16) Glover, S. A.; Novak, M. Can. J. Chem. 2005, 83, 1372–1381.
(17) Novak, M.; Lin, J. J. Org. Chem. 1999, 64, 6032–6040.
4464 J. Org. Chem. Vol. 74, No. 12, 2009

Hydrolysis and Photolysis of a Model Anti-Tumor Quinol Ester
FIGURE 1. Bond lengths and charges computed at the B3LYP/6-
31G(d) level for (a) 3, (b) 5a, and (c) 2-methylbenzothiazole (charges
on H summed into the heavy atoms); (d) resonance forms for 3. Data
in red correspond to charge differences between atoms in 3 and those
of the corresponding atoms in 2-methylbenzothiazole.
states are computed to be 22.3, 22.3, and 23.2 kcal/mol,
respectively, higher in energy in keeping with T-S0 gaps
computed previously by DFT for phenyl- and 4-methylpheny-
loxenium ions.16 Using S0 energies, the energetics of the
isodesmic reactions of eq 1 and eq 2 were compared at
the B3LYP/6-31G(d) level with and without ZPE corrections.
The results are shown in the equations. ∆E calculated at this
level for eq 1 is comparable to the results previously obtained
at the pBP/DN*//HF/6-31G(d) and BP/6-31G(d)//HF/6-31G(d)
levels. The results for the isodesmic reaction of eq 2 suggest
that the 4-(benzothiazol-2-yl) substituent is similar to the 4-Ph
substituent in its ability to stabilize an aryloxenium ion. If 3
can be generated by ionization of 1, it should have a lifetime
C-6′ in 5a and 5b rather than an enhanced resonance interaction
with the benzothiazole system.
Like 5b,13 delocalization into the 4-phenyl ring of 5a is
evidenced by strong bond alternation (Figure 1b). Positive
charge is greatest at C-2, C-4, and C-4′, the known sites of
trapping by water and azide.9,10 Delocalization throughout 3 is
also evident from a comparison of the ground state structures
of 3 with 2-methylbenzothiazole computed at the same level.
While the carbon-carbon bond lengths in 2-methylbenzothia-
zole are similar (Figure 1c), in 3, they show a high degree of
bond length alternation (Figure 1a). Notably, the ring-junction
carbon-carbon bond is 0.032 Å longer in the oxenium ion. This,
together with significant lengthening of the N-(C-2′) bond,
corresponding contraction in the N-(C-3a′) and S-(C-7a′)
bonds, as well as a charge increase on S of +0.176, points to
a strong contribution of resonance form III in Figure 1d.
Substantial increases in charge at C-4′ and C-6′ relative to
2-methylbenzothiazole are also evident, pointing to resonance
representations IV and V.
long enough for it to be detected by N3 trapping.9-13
-
This conclusion is supported by the computed properties of
3, 5a, and 5b. B3LYP/6-31G(d) charges and bond lengths for
5b have recently been reported along with its computed and
experimental vibrational spectrum.13 Corresponding data for 3
and 5a are given in Figure 1a,b. All three ions exhibit
2,5-cyclohexadien-1-one-4-yl cationic character. The carbonyl
stretch frequency for 5b has been calculated at 1672 cm-1 in
the gas phase. This correlates with that observed by time-
resolved resonance Raman spectroscopy (1635 cm-1 in aqueous
solution).13 3 and 5a are computed to have similar scaled
carbonyl stretch vibrations at 1663 and 1672 cm-1.18 In all three
ions, the positive charge is essentially on the 4-position and on
the 4-aryl substituent rather than on the oxygen atom.
The dipole moments for 3, 5a, and 5b are 4.00, 4.29, and
4.45 D, respectively, and are aligned largely along the molecular
axes toward the substituent ring. The total Mulliken charges
on the substituent aryl groups in 3 and 5a are, respectively,
+0.49 and +0.47, while the corresponding charge on the p-tolyl
group in 5b is +0.52.13 Though slightly less efficient than 4-(p-
tolyl), both 4-phenyl and 4-(benzothiazol-2-yl) have a similar
capacity to delocalize the positive charge away from the
4-position of the aryloxenium ion. Interestingly, B3LYP/6-
31G(d) computes 5a and 5b to be slightly twisted, with inter-
ring dihedral angles of 19 and 14°, respectively, while 3 is
completely planar. The inter-ring bond in 3 (1.414 Å) is also
shorter than that in 5a and 5b (1.432 and 1.427 Å, respec-
tively).13 These differences probably reflect the reduced steric
demand of the heteroatoms in 3 as opposed to CH at C-2′ and
By analogy to 5a, in which water attacks at C-4 and azide
reacts at C-2 and C-4′,9,10 nucleophilic attack on 3 should occur
at C-2 and C-4 of the phenyl ring and at C-4′ and C-6′ on the
distal benzothiazole group. All these positions contribute
strongly to the LUMO of 3 shown in Figure 2.
The quinol ester 1 was synthesized from the oxidation of the
corresponding phenol 8 by phenyliodonium diacetate (PIDA)
in AcOH in ca. 20% yield. This procedure has been applied to
synthesize other quinol esters in low to moderate yields.9,11
1
is stable enough in the dark to be purified by radial chroma-
tography on silica gel. Synthesis and characterization of all
1
compounds is presented in the Experimental Section, and H
and 13CNMR spectra are reported in the Supporting Information.
Kinetics of decomposition of 1 were measured at 80 °C in
HClO4 solutions (pH < 3.0) or in HCO2H/NaHCO2, AcOH/
AcONa, NaH2PO4/Na2HPO4 buffers (5 vol % CH3CN-H2O, µ
) 0.5 (NaClO4)) by monitoring changes in UV absorbance as
a function of time. All reactions were performed with an initial
ester concentration of ca. 5 × 10-5 M. Absorbance versus time
data were fit to a first-order or a consecutive first-order rate
equation. Rate constants taken at a minimum of three wave-
(18) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502–16513.
J. Org. Chem. Vol. 74, No. 12, 2009 4465

Wang et al.
TABLE 1. Derived Rate Parameters for the Decomposition of 1
and 2 and 4a and 4b
T
(°C)
ko (s-1
)
kH (M-1 s-1
)
kOH (M-1 s-1
)
1
2
80 (1.91 ( 0.08) × 10-4 (6.2 ( 0.8) × 10-3
4aa c 30 (2.50 ( 0.05) × 10-5 (1.45 ( 0.08) × 10-3
(5.5 ( 1.2) × 102
80
(2.5 ( 0.4) × 10-3
,
4bb c 30 (1.02 ( 0.01) × 10-3 (1.95 ( 0.08) × 10-2
,
4ad
80 (4.5 ( 0.4) × 10-3
4be
80 (1.0 ( 0.1) × 10-1
a Data from ref 9. b Data from ref 11. c Reaction conditions, except
temperature, identical to 1. d Extrapolated to 80 °C from data taken in
the 30 to 70 °C range. e Extrapolated to 80 °C from data taken in the 20
to 40 °C range.
pathways, respectively. The first two rate terms of eq 3 are
consistent with what we observed for the oxenium ion precursors
4a and 4b.9,11 For these two esters, the processes governed by
both ko and kH involve formation of the cations 5a and 5b.9-11
However, no rate term corresponding to kOH was observed for
4a and 4b in basic phosphate buffers.9,11 Quinol 2 decomposes
FIGURE 2. LUMO of 3 at the B3LYP/6-31G(d) level. Arrows denote
likely positions for nucleophilic attack.
2
slowly through an acid-catalyzed pathway governed by kH . This
process became exceedingly slow at pH > 3.0, so the decom-
position of 2 no longer interfered with the measurement of the
kinetics of decomposition of 1 except at pH > 7.0. HPLC
examination and repetitive wavelength scans showed that 2
slowly decomposes in basic phosphate buffers, but its decom-
position kinetics were not examined in detail under these
conditions. Negligible absorbance changes occurred for the
decomposition of 2 in phosphate buffers in the range from 250
to 270 nm, so it was possible to monitor the decomposition
kinetics of 1 in this wavelength range without interference from
the much slower decomposition of 2. Rate constants derived
from the fits to eq 3 and eq 4 for the decomposition of 1 and 2
are reported in Table 1 with some comparisons to 4a and 4b.
Comparison of ko1 to the corresponding rate constants for 4a
and 4b clearly shows that the rate of uncatalyzed decomposition
of 1 is significantly depressed compared to other quinol esters
that generate oxenium ions. At 80 °C, the uncatalyzed decom-
position of 1 occurs ca. 24-fold more slowly than the decom-
position of 4a and ca. 520-fold more slowly than 4b, while the
N3- trapping data shown below indicate that 1 does generate 3.
Therefore, the benzothiazol-2-yl group is clearly acting as an
electron-withdrawing group that destabilizes the transition state
for ionization and decreases the rate of ionization. The base-
catalyzed term kOH was not observed for 4a or 4b at pH < 8.0.9,11
If kOH is about the same magnitude for 4a and 4b at 80 °C as
kOH1, it would not be possible to detect the base-catalyzed
reaction for 4a below ca. pH 9.2 or for 4b below pH 10.6
because of the larger magnitudes of ko for these compounds
compared to 1. Since it is probable that the base-catalyzed
hydrolysis of 1 is accelerated by the electron-withdrawing
benzothiazol-2-yl group, it is likely that the pH estimates for
the onset of observable base-catalyzed hydrolysis of 4a and 4b
are lower limits.
FIGURE 3. The pH dependence of the decomposition of 1 and 2 at
80 °C. Red circles correspond to the hydrolysis rate constants for
decomposition of 1; blue triangles correspond to acid-catalyzed
decomposition of 2 determined by consecutive first-order fitting of the
rate data for 1; magenta triangles correspond to directly measured rate
constants for acid-catalyzed decomposition of authentic 2. Data were
fit to eqs 3 and 4 as described in the text.
lengths were averaged at each pH. Consecutive first-order
reactions were found only at pH < 3.0. Under these acidic
conditions, one rate constant was found to be experimentally
equivalent to the rate constant for acid-catalyzed decomposition
of 2, the major hydrolysis product of 1. The absorbance versus
time data for the decomposition of authentic 2 at pH < 3.0 fit
a standard first-order rate equation well. All kobs for 1 and 2 are
provided in the Supporting Information. The pH dependence
of the decomposition of 1 and 2 is summarized in Figure 3.
The kobs for decomposition of 1 at pH 1.0-7.5 was fit to the
rate law of eq 3. The kobs for decomposition of 2 at pH 1.0-2.5
was fit to the rate law of eq 4. Both 1 and 2 could be protonated
at N-3′ in sufficiently acidic solution, but neither the kinetic
data nor UV spectra of these compounds indicate that proton-
ation occurs at pH g 1.0.
1
1
1
kobs ) ko + kH [H+] + kOH1[OH-]
(3)
(4)
The HPLC examination of the decomposition of 1 showed
that, at pH > 3.0 in the absence of any nucleophile other than
H2O, only one significant product was detected. This product
was isolated from the hydrolysis of 1 in pH 4.61 acetate buffer
at 80 °C in 71% yield and was identified as the quinol 2. At pH
< 3.0, 2 decomposes into a new material that has nearly the
same HPLC retention time as 2. The new material can be
differentiated from 2 by its different absorbance characteristics
from 2 at the two wavelengths used to monitor the HPLC
2
2
kobs ) kH [H+]
1 exhibits pH-independent decomposition kinetics from pH
2 to 7 and pH-dependent decomposition at pH < 2 and pH > 7.
1
1
Three rate constants, ko , kH , and kOH1, were used to fit the rate
1
data: ko corresponds to the pH-independent decomposition
pathway that is dominant over a wide pH range, while kH1 and
kOH1 describe acid-catalyzed and base-catalyzed decomposition
4466 J. Org. Chem. Vol. 74, No. 12, 2009

Hydrolysis and Photolysis of a Model Anti-Tumor Quinol Ester
SCHEME 1. Azide Trapping during Hydrolyis of 1
FIGURE 4. Azide trapping experiments for 1 performed at pH 6.5 in
0.02 M phosphate buffer. Data were fit by least-squares procedures to
the standard “azide clock” equations as described in the text. Inset:
Rate constants for decomposition of 1 in the presence of N3 in pH
6.5 0.02 M phosphate buffer.
attack of H2O and N3- is a common feature of both aryloxenium
and arylnitrenium ions.9-11,17,20 It is likely that N3- also attacks
these cations, at least in part, at C-4, but these products are
unstable and undergo rearrangement to more thermodynamically
-
chromatography (212 and 235 nm). The same material was also
detected by HPLC as the final product of the hydrolysis 1 at
pH < 3.0. This material was not isolated, but quinols similar to
2 have previously been shown to undergo dienone-phenol
rearrangements under these conditions.10,19 The yield of 2
determined by HPLC quantification in buffers from pH 3.5 to
6.5 after 8 half-lives (ca. 8 h) of the hydrolysis of 1 is somewhat
higher than the isolated yield at pH 4.6 at ca. 85-90%. In more
basic phosphate buffers, authentic 2 slowly decomposes, so its
yield determined after the same reaction time is lower. For
example, the yield of 2 from the hydrolysis reaction of 1 in pH
7.1 phosphate buffer determined after 8 h of reaction is
somewhat lower at 72%. For this reason, the wavelengths used
to monitor the decomposition kinetics of 1 in phosphate buffer
at pH 7.1 and 7.5 were chosen from the wavelength range in
which the decomposition of 2 did not show an absorbance
change. Quinol 2 is the expected product of the reaction of the
cation 3 with the aqueous solvent. Both the product and the
kinetics of the hydrolysis reaction of 1 are consistent with what
we have observed for other quinol esters that yield oxenium
ion intermediates,9-13 but these data do not require the
intermediacy of an oxenium ion since 2 would also be the
product of an ordinary ester hydrolysis.
stable N3 adducts.20 The product data obtained at different
-
[N3-] fit to the standard “azide clock” equations to generate
kaz/ks,21,22 the ratio of the second-order rate constant for the
-
reaction between 3 and N3 to the first-order rate constant for
the trapping of 3 by the aqueous solvent. This ratio describes
the selectivity of the intermediate toward trapping by N3- versus
the aqueous solvent. For 3, kaz/ks is 310 ( 25 M-1
.
The two oxenium ions 5a and 5b derived from 4a and 4b
have kaz/ks of 77 and 1.0 × 103 M-1, respectively, at 30 °C.9,11
Provided that these ratios are not strongly affected by temper-
ature, the ability of the benzothiazol-2-yl substituent to stabilize
the cation appears to be between that of phenyl and p-tolyl
substituents when they are placed at the 4-position in the aryl
ring of the cation. If the reaction between the intermediate 3
-
and N3 is diffusion-limited, as is the case for other reactive
oxenium and nitrenium ions,12,23,24 kaz will be ca. 6.5 × 109
M-1 s-1 at 20 °C.12 Extrapolation of kaz to 80 °C, based on the
temperature dependence of other diffusion-controlled reactions,
gives an estimate for the rate constant of ca. 1.6 × 1010 M-1
25,26
s-1
.
Then the estimated lifetime (1/ks) of 3 in aqueous
solution at 80 °C is ca. 20 ns. That lifetime would be expected
to be longer at 30 °C, so 3 would appear to be somewhat longer
lived than 5a (ca. 12 ns) under the same conditions.9 The
differences in behavior between 3 and 5a are small enough that
the lifetimes of these two cations in aqueous solution under
identical conditions are likely to be within a factor of 2 to 3 of
each other.
The combination of kinetic data and azide trapping results
shows that the 4-(benzothiazol-2-yl) substituent has two sig-
nificantly different electronic effects. Inductive electron with-
drawl dominates in destabilizing the transition state for ioniza-
tion as C-4 of 1 transforms from a saturated tetrahedral carbon
Results of azide trapping experiments performed on 1 in
phosphate buffer at pH 6.5 are summarized in Figure 4. In the
presence of N3-, the hydrolysis product 2 was replaced by
the azide trapping adduct 9 without an apparent increase in the
overall rate constant for decomposition of 1 (Figure 4, inset).
-
The N3- results require a two-step mechanism in which N3
and water compete for the same electrophilic intermediate
generated in a prior rate-limiting step. The intermediate most
consistent with the generation of 2 and 9 is the oxenium ion 3
(Scheme 1). Control experiments showed that the reaction
-
between N3 and quinol 2 is negligible under our reaction
(20) Novak, M.; Kahley, M. J.; Lin, J.; Kennedy, S. A.; James, T. G. J.
Org. Chem. 1995, 60, 8294–8304.
(21) Jencks, W. P. Acc. Chem. Res. 1980, 13, 161–169.
(22) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 4689-4691;
1982, 104, 4691-4692; 1984, 106, 1383-1396.
(23) Davidse, P. A.; Kahley, M. J.; McClellend, R. A.; Novak, M. J. Am.
Chem. Soc. 1994, 116, 4513–4514.
(24) McClelland, R. A.; Davidse, P. A.; Hadzialic, G. J. Am. Chem. Soc.
1995, 117, 4173–4174.
(25) Buxton, G. V.; Elliot, A. J. J. Chem. Soc., Faraday Trans. 1993, 89,
485–488.
conditions. All the loss of 2 can be attributed to the trapping of
the cation 3. There are two apparent azide trapping products
detected by HPLC, but only the major one 9 was successfully
isolated by multiple application of radial chromatography. The
structure of 9 indicates that the positive charge in 3 is
significantly delocalized into the thiazole substituent, particularly
at C-6′ of the benzothiazole ring. Differing regioselectivity for
(19) Vitullo, V. P.; Logue, E. A. J. Org. Chem. 1973, 38, 2265–2267.
(26) Sehested, K.; Christensen, H. Radiat. Phys. Chem. 1990, 36, 499–500.
J. Org. Chem. Vol. 74, No. 12, 2009 4467

Wang et al.
SCHEME 2. Photolysis Resultsa
Combined yields of these products account for ca. 50% of 1
that was consumed under both conditions. Some of the products
(2, 8, and 10) are analogous to those isolated from the photolysis
of 4a and 4b.13 The products 8, 10, and 11 appear to be derived
from the aryloxy/acetoxy radical pair 14, while 2 is the expected
hydration product of 3.13 A pathway for the generation of 10
from the corresponding aryloxy radical in oxygenated solutions
has been presented elsewhere.13 Combination of the aryloxy
radical with methyl radical generated from decarboxylation of
the acetoxy radical provides a path to 11. Surprisingly, some 2
is generated in CH3CN. The analogous products 7a and 7b were
not detected during photolysis of 4a and 4b in CH3CN, and 5b
could not be detected after laser flash photolysis of 4b in CH3CN
even though 5b is known to have a significant lifetime in
aqueous solution.13 This suggests that there may be a photolytic
route from 1 to 2 in CH3CN that does not involve 3 since this
cation is less stable than 5b. The yield of 2 significantly
increases in aqueous solution. This indicates that 3 is responsible
for the generation of a significant portion of 2 in aqueous
solution.
The rearrangement product 12 and the dimer 13 do not appear
to be derived from a cationic or radical intermediate but are
likely to be derived directly from excited states of 1. Products
similar to these were not detected during photolysis of 4a and
4b.12,13 The rearrangement product 12 is the expected product
of a dienone-phenol photorearrangement of 1.27 Photodimer-
ization of cyclohexenone derivatives usually yields cis-anti-cis
head-to-head (H-H) or head-to-tail (H-T) dimers as the major
products.28,29 We have tentatively assigned the H-H structure
1
to 13 based on an analysis of its H NMR spectrum. The peak
at 2.92 ppm assigned to Ha (Ha′) is a simple doublet with a
coupling constant of 5.0 Hz. It is coupled to the proton at 3.88
ppm assigned to Hb (Hb′). In the H-T structure Ha or Ha′ would
be expected to couple to both Hb and Hb′ to generate a peak for
Ha (Ha′) that would not be a simple doublet based on observed
1H NMR spectra of similar compounds.29
a Conditions: CH3CN, O2 (upper yields), or 20 vol % CH3CN/H2O, pH
5.1 acetate buffer, O2 (lower yields), 235-280 nm.
to the trigonal carbon of 3. Once the cation is fully formed,
charge delocalization through the π-system of the 4-(benzothia-
zol-2-yl) substituent stabilizes the cation to a somewhat greater
extent than does a 4-phenyl substituent.
Photolysis of 4a and 4b has previously been shown to
generate both the cations 5a and 5b and the aryloxy radicals
6a and 6b.12,13 Homolysis to form the radical is the exclusive
reaction in CH3CN, while both heterolyis and homolysis occur
in aqueous solution.13 The intermediates 5b, 6b, and 6a were
directly detectable by fast UV spectroscopy following nano-
second laser flash photolysis.12,13 The cation 5a was not directly
detectable due to its short lifetime, estimated to be 12 ns from
azide trapping data,9 but 7a, the hydration product of 5a, was
readily detected after photolysis in aqueous solution.13
The major differences between the photolysis of 1 and those
previously reported for 4a and 4b are the significantly lower
yield of the quinol from photolysis in H2O under similar reaction
conditions (7.8% of 2, 18% of 7a, and 33% of 7b) and the
greater yields of products from photolysis of 1 that are not
derived from cationic intermediates.13 It appears that photoion-
ization of 1 is suppressed compared to 4a and 4b. This would
lead directly to a lower yield of the quinol and to greater yields
of products derived from competing pathways. Suppression of
photoionization is probably due to the inductive electron-
withdrawing effect of the benzothiazol-2-yl group. This effect
may be particularly important in the photoionization reaction
Steady state photolysis of 1 was carried out at initial
concentrations of 1 of ca. 6 × 10-4 M in oxygenated CH3CN
and 20 vol % CH3CN/H2O buffered with 0.02 M 1/1 HOAc/
NaOAc. Some initial photoproducts (particularly 2 and 10) were
also photoreactive, so the reactions were halted when ca. 90%
of 1 had been consumed. Under these conditions, photolysis of
initial photoproducts was minimized, but not eliminated, so
yields of 2 and 10, in particular, are lower limits. The analogous
products from photolysis of 4a and 4b were also photoreactive,
so comparisons of relative yields can still be made.13 HPLC
analysis of reaction mixtures as photolysis proceeded showed
that at least 11 initial products were generated under both
photolysis conditions. Scheme 2 provides a summary of the
yields of six major photoproducts that were isolated and
(27) Schultz, A. G.; Hardinger, S. A. J. Org. Chem. 1991, 56, 1105–1111.
Schultz, A. G.; Green, N. J. J. Am. Chem. Soc. 1992, 114, 1824–1829.
(28) Lem, G.; Kaprinidis, N. A.; Schuster, D. I.; Ghatlia, N. D.; Turro, N. J.
J. Am. Chem. Soc. 1993, 115, 7009–7010.
(29) Schuster, D. I.; Greenberg, M. M.; Nunez, I. M.; Tucker, P. C. J. Org.
Chem. 1983, 48, 2615–2619. Yang, J.; Dewal, M. B.; Profeta, S.; Smith, M. D.;
Li, Y.; Shimizu, L. S. J. Am. Chem. Soc. 2008, 130, 612–621.
1
identified. Yields are based on a combination of H NMR
integration and HPLC quantification.
4468 J. Org. Chem. Vol. 74, No. 12, 2009

Hydrolysis and Photolysis of a Model Anti-Tumor Quinol Ester
TABLE 2. Wavelengths Used for Kinetic Analysis of the
Decomposition of 1
since the expected vertical ionization process would initially
lead to a cation with a trigonal pyramidal geometry at C-4 that
would not be effectively stabilized by resonance interaction with
the benzothiazol-2-yl group until it had undergone structural
reorganization.
reaction solution
pH
UV wavelengths monitored (nm)
HClO4 solutions
formate buffer
acetate buffer
phosphate buffer
phosphate buffer
1.0-2.5
3.0-4.0
4.5
5.5-6.5
7.1-7.5
212, 221, 235, 340
217, 257, 340
217, 257, 340
222, 257, 340
250, 257, 265
Conclusion
Azide trapping and kinetic results of the decomposition of
quinol ester 1 in aqueous solution and comparison to the
behavior of other quinol esters 4a and 4b that yield oxenium
ions 5a and 5b prove that oxenium ion 3 is generated from
uncatalyzed decomposition of 1 and that 3 is responsible for
the formation of the hydration product 2 under these conditions.
On the basis of the precedent of 4a, it is also likely that the
acid-catalyzed decomposition of 1 generates 3.10 Unlike the 4-Ph
and 4-(p-tolyl) substituents of cations 5a and 5b, the 4-(ben-
zothiazol-2-yl) substituent of 3 shows two significantly different
electronic effects at different stages of the reaction. At the
transition state for ionization of 1, the inductive electron-
withdrawing effect of the 4-(benzothiazol-2-yl) substituent
dominates and suppresses the transformation of C-4 from a
saturated tetrahedral carbon to a positively charged carbon with
trigonal geometry. That makes the generation of 3 significantly
slower than the generation of 5a or 5b at the same temperature.
Once the cation is formed, the same substituent stabilizes 3 by
electron delocalization through the π-system of the benzothia-
zole ring, confirmed by the magnitude of kaz/ks and by the
structure of the major azide trapping product 9. This stabilization
lies between that of the 4-phenyl and 4-(p-tolyl) substituents of
5a and 5b.
Compared to previously reported photochemistry of 4a and
4b, photolysis of 1 is more complicated. Besides heterolysis
and homolysis routes that generate cationic transient 3 and the
radical pair 14, respectively, additional routes, including
dienone-phenol photorearrangement and photodimerization,
appear to be involved in the formation of rearrangement product
12 and dimer 13. No products similar to 12 and 13 were found
previously after the photolysis reactions of other oxenium ion
precursors.13 Quinol 2, the expected product of 3 generated by
photolytic heterolysis of 1, is formed in aqueous solution but
in significantly lower yield than that of 7a and 7b observed
during photolysis of 4a and 4b under equivalent conditions. This
suggests that the photoionization of 1 is suppressed, probably
due to the inductive electron-withdrawing effect of the ben-
zothiazol-2-yl group. A very small amount of 2 is found after
photolysis of 1 in CH3CN as well, indicating that alternative
mechanisms for generation of 2 should be considered in this
case. We will continue this study with emphasis on direct
detection and structural characterization of 3 and the other
related transients by laser flash photolysis methods.
although the putative metabolite 16 had not previously been
isolated.31 It is our hypothesis that the biological activity of 15
and related derivatives is due to ionization of 16 and related
compounds to generate selective nitrenium ions of general
structure 17.32 We have recently synthesized 16 and have
initiated a study of its aqueous solution chemistry that includes
kinetic analysis, trapping of reactive intermediates, and possible
photolytic generation of 17. Results of these studies will be
reported at a later date.33
Experimental Section
Kinetics and Product Analyses: Hydrolysis reactions were
performed in 5 vol % CH3CN-H2O, µ ) 0.5 (NaClO4), at 80 °C
for 1 and 2. The pH was maintained with HClO4 solutions (pH <
3.0) or with HCO2H/NaHCO2, AcOH/AcONa, and NaH2PO4/
Na2HPO4 buffers. All pH values were measured at ambient (20-23
°C) temperature and are uncorrected. Reactions were initiated by
injecting 15 µL of a ca. 0.01 M CH3CN solution of 1 or 2 into 3
mL of the reaction solution incubated at 80 °C for 20 min prior to
the injection. The initial concentration of 1 or 2 in the reaction
was ca. 5 × 10-5 M. Kinetics were monitored by changes in UV
absorbance at a minimum of three wavelengths. The rate constant
for each reaction was obtained as an average of all rate constants
obtained at each wavelength. The wavelengths chosen to monitor
the decomposition of 1 in solutions at pH 1.0-6.5 varied slightly
in different reaction solutions due to slight shifts of the maximum
absorbance changes observed in these solutions. In phosphate
buffers at pH 7.1 and 7.5, the decomposition kinetics of 1 were
monitored at wavelengths chosen from the UV range in which no
observable absorbance change was found during the slow decom-
position of 2 at the same pH (Table 2). The kinetic studies of the
decomposition of 2 were performed at pH < 3. The wavelengths
used the decomposition of 2 were 212, 233, and 260 nm.
For 1, absorbance versus time data were fit to a double
exponential rate equation at pH < 3 and to a standard first-order
rate equation at pH > 3. For 2, a standard first-order rate equation
was used for data fitting. Two rate constants were observed for the
hydrolysis of 1 in solutions with pH < 3. Product yields were
monitored by HPLC on the same solutions used for kinetics after
8 half-lives of the hydrolysis reactions. HPLC conditions: 20 µL
injections on a 4.7 mm × 250 mm C-8 reverse phase column, 65/
Derivatives of 2-(4-aminophenyl)benzothiazole 15, including
one compound in phase 1 clinical trials, have been under
development as anti-tumor agents.30 Experimental data suggest
that 15 is activated by N-hydroxylation and esterification,
(30) Shi, D.-F.; Bradshaw, T. D.; Wrigley, S.; McCall, C. J.; Lelieveld, P.;
Fichtner, I.; Stevens, M. F. G. J. Med. Chem. 1996, 39, 3375–3384. Bradshaw,
T. D.; Wrigley, S.; Shi, D.-F.; Schultz, R. J.; Paull, K. D.; Stevens, M. F. Br. J.
Cancer 1998, 77, 745–752. Bradshaw, T. D.; Shi, D.-F.; Schultz, R. J.; Paull,
K. D.; Stevens, M. F. G. Br. J. Cancer 1998, 78, 421–429. Hutchinson, I.; Chua,
M.-S.; Browne, H. L.; Trapani, V.; Bradshaw, T. D.; Westwell, A. D.; Stevens,
M. F. G. J. Med. Chem. 2001, 44, 1446–1455. Hutchinson, I.; Jennings, S. A.;
Vishnuvajjala, B. R.; Westwell, A. D.; Stevens, M. F. G. J. Med. Chem. 2002,
45, 744–747.
(31) Bradshaw, T. D.; Westwell, A. D. Curr. Med. Chem 2004, 11, 1009–
1021. O’Brien, S. E.; Browne, H. L.; Bradshaw, T. D.; Westwell, A. D.; Stevens,
M. F. G.; Laughton, C. A. Org. Biomol. Chem. 2003, 1, 493–497.
(32) Novak, M.; Rajagopal, S. AdV. Phys. Org. Chem. 2001, 36, 167–253.
Novak, M.; Rajagopal, S.; Xu, L.; Kazerani, S.; Toth, K.; Brooks, M.; Nguyen,
T.-M. J. Phys. Org. Chem. 2004, 17, 615–624.
(33) Novak, M.; Chakraborty, M.; Brewer, S. to be submitted.
J. Org. Chem. Vol. 74, No. 12, 2009 4469

Wang et al.
35 MeOH/H2O elution solvent, flow rate 1.0 mL/min, monitoring
wavelengths ) 212, 235 nm.
The isolation of 2, the hydrolysis product of 1, was performed
in 0.02 M pH 4.61 acetate buffer with concentration of 1 of ca. 7
× 10-4 M. The rate of the decomposition of 2 is negligible at this
pH. The procedure for isolation and purification of 2 is described
in the synthesis section.
2. Isolation, Purification, and Characterization of Photoly-
sis Products: Photoproducts were isolated from large-scale pho-
tolysis of quinol ester 1 in aqueous buffers (primarily 2, 8, 11, 12,
13) or in CH3CN (primarily 8, 10, 11, 13). The general procedures
for product isolation are as follows. For the reactions run in
CH3CN-H2O mixture, the solution was extracted with CH2Cl2 (3
× 50 mL). The combined extract was dried over anhydrous Na2SO4,
filtered, and evaporated to dryness. The residue was subjected to
further purification as described below. For the reactions run in
CH3CN, photoproducts were subjected to separation and purification
directly after removing the solvent by rotary evaporation. An initial
separation of photoproducts was performed by radial chromato-
graphy on silica gel with 1:1 hexanes/EtOAc. The fractions
containing photoproducts were evaporated to dryness under vacuum,
and the residues were purified individually by radial chromatog-
raphy with 9:1 CH2Cl2/CH3CN. Multiple applications of TLC on
silica gel with 3:2 hexanes/EtOAc were used to purify 12. Final
purification of 10 and 13 was accomplished by radial chromato-
graphy on silica gel with 13:7 hexanes/EtOAc. The peroxy
compound 10 decomposes slowly during characterization. Purified
products were subjected to HPLC analysis for purity before final
characterization. Quinol 2 and the phenols 8 and 11 were identified
by direct comparisons to synthesized samples.
Azide Trapping Studies: Studies of the hydrolysis of 1 in the
presence of N3- at 80 °C were performed in 0.02 M 1/1 NaH2PO4/
Na2HPO4 buffer (pH 6.6, 5 vol % CH3CN-H2O, µ ) 0.5(NaClO4))
at various N3- concentrations (0.1, 0.01, 0.008, 0.006, 0.004, 0.002
-
M). The direct reaction of 2 with N3 was negligible under these
conditions. Kinetics of the decomposition of 1 in the presence of
N3- were determined by following the change in the UV absorbance
at 310 and 330 nm. Absorbance versus time data were fit to a
standard first-order rate equation. The reported rate constant is the
average taken at both wavelengths. The yields of products were
determined after 2 h reaction time by HPLC using the same
conditions as those described above. The yields were determined
after 2 h instead of after the completion of the reaction because
some of the products decompose slowly in the reaction media. The
residual concentration of 1 was monitored by HPLC to ensure that
the individual reactions had proceeded to the same extent. Two
new HPLC peaks with longer retention time were observed. Only
the major adduct was isolated as follows. The ester 1 (50.0 mg,
0.18 mmol), dissolved in 1 mL of CH3CN, was added in 0.1 mL
aliquots every 1 h to 250 mL of a 0.02 M phosphate buffer (pH
6.7, 5 vol % CH3CN-H2O, µ ) 0.5 (NaClO4)) containing 0.008
M NaN3 that was incubated in the dark at 80 °C. After the last
addition, the mixture was incubated in the dark at 80 °C for an
additional 3.5 h. The mixture was then refrigerated overnight and
was allowed to reach room temperature before extracting with
CH2Cl2 (5 × 50 mL). The combined extract was dried over
anhydrous Na2SO4, filtered, evaporated to dryness, and subjected
to multiple radial chromatography with two solvent systems (5:1
CH2Cl2/EtOAc, then 20:1 CH2Cl2/CH3CN). The major azide adduct
was separated as a low melting point solid.
4-(Benzothiazol-2-yl)-4-(methylperoxy)cyclohexa-2,5-di-
enone (10): IR (thin film) 3063, 2924, 1675, 1635, 1612, 1458,
1435 cm-1;1H NMR (500 MHz, DMSO-d6) δ 3.93 (3H, s), 6.48
(2H, d, J ) 10.2 Hz), 7.40 (2H, d, J ) 10.1 Hz), 7.50-7.58 (2H,
m), 8.04 (1H, dd, J ) 7.7 Hz, 0.8 Hz), 8.19 (1H, dd, J ) 8.0 Hz,
0.9 Hz); 13C NMR (125.8 MHz, DMSO-d6) δ 64.5, 72.4, 122.6,
123.3, 126.2, 126.8, 130.2, 134.7, 144.7, 151.8, 167.6, 192.7; high-
resolution MS (ES, positive) C14H12NO3S (M + H) calcd 274.0532,
found 274.0548.
4-Acetoxy-3-(benzothiazol-2-yl)phenol (12): IR 3233, 3063,
1
2922, 1757, 1606, 1512, 1432, 1342, 1177, 1108, 1013 cm-1; H
NMR (500 MHz, CD2Cl2) δ 2.44 (3H, s), 6.99 (1H, dd, J ) 8.7
Hz, 3.0 Hz), 7.12 (1H, d, J ) 8.7 Hz), 7.43 (1H, td, J ) 7.5 Hz,
1.2 Hz), 7.53 (1H, td, J ) 7.7 Hz, 1.2 Hz), 7.80 (1H, d, J ) 3.0
Hz), 7.97 (1H, dd, J ) 7.5 Hz, 0.9 Hz), 8.06 (1H, dd, J ) 7.5 Hz,
0.6 Hz); 13C NMR (125.8 MHz, CD2Cl2) δ 21.85 (2.44), 116.04
(7.80), 118.84 (6.99), 121.85 (7.97), 123.63 (8.06), 125.34 (7.12),
125.86 (7.43), 126.78 (7.53), 127.24, 135.92, 142.45, 153.18,
154.08, 162.32, 169.92; LC-MS (ESI, positive) m/e 308 (M + Na)+,
286 (M + H)+, 244 (M - CH2CO + H)+; (ESI, negative) m/e 284
(M - H)-; high-resolution MS (ES, positive) C15H12NO3S (M +
H) calcd 286.0532, found 286.0548.
4-(6-Azidobenzothiazol-2-yl)phenol (9): IR 3062, 2097, 1606,
1
1557, 1451, 1435, 1279, 1225, 1168 cm-1; H NMR (500 MHz,
DMSO-d6) δ 6.93 (2H, d, J ) 8.5 Hz), 7.22 (1H, dd, J ) 8.8 Hz,
2.3 Hz), 7.91 (2H, d, J ) 8.5 Hz), 7.93 (1H, d, J ) 2.0 Hz), 7.97
(1H, d, J ) 8.5 Hz), 10.22 (1H, s); 13C NMR (125.8 MHz, DMSO-
d6) δ 112.63 (7.93), 116.37 (6.93), 118.66 (7.22), 123.55 (7.97),
124.13, 129.24 (7.90), 136.05, 136.63, 151.59, 160.81, 167.61; LC-
MS (ESI, positive) m/e 241 (M - N2 + H)+ (ESI, negative) m/e
239 (M - N2 - H)-; high-resolution MS (ES, positive) C13H9N4OS
(M + H) calcd 269.0492, found 269.0512.
1,8-Di(benzothiazol-2-yl)-4,5-dioxo-1,4,4a,4b,5,8,8a,8b-octahy-
drobiphenylene-1,8-diyl diacetate (13): IR 3066, 1764, 1706,
1
1574, 1436, 1369, 1198, 1179, 1014 cm-1; H NMR (500 MHz,
Steady State Photolysis Experiments. 1. Photolysis: Steady
state photolysis of 1 in O2-saturated pH 5.1, 0.02 M 1/1 acetate
buffer (20 vol % CH3CN-H2O, µ ) 0.5 (NaClO4)) and in O2-
saturated CH3CN were performed in a Rayonet photochemical
reactor in a jacketed quartz vessel kept at 30 °C. Luzchem LZC-
UVC lamps that have emission in the range of 235-280 nm were
used as the UV source. A brief description of the apparatus for
steady state photolysis has been published.12 Before irradiation,
an initial concentration of ca. 6 × 10-4 M of 1 was obtained by
adding 5 mL of ca. 0.024 M solution of 1 (33.5 mg, 0.12 mmol) in
CH3CN to 200 mL of O2-saturated acetate buffer, or by directly
adding the ester 1 (33.5 mg, 0.12 mmol) into 200 mL of
O2-saturated CH3CN. Reaction solutions were mixed thoroughly
and then subjected to UV radiation with 1.5-2 min irradiation
intervals. The yield of products reached a maximum after ca. 8.5
min of irradiation based on the HPLC analysis that was performed
before irradiation and at the end of each irradiation interval (C-8
reverse phase column, 65/35 MeOH/H2O elution solvent, 1 mL/
min, monitored by UV absorbance at 220 and 260 nm). Quantifica-
tion of products was determined by a combination of HPLC and
1H NMR of reaction mixtures.
CD2Cl2) δ 2.15 (3H, s), 2.92 (1H, d, J ) 5.0 Hz), 3.88 (1H, dd, J
) 5.0 Hz, 3.0 Hz), 6.00 (1H, dd, J ) 5.5 Hz, 1.0 Hz), 7.41 (1H,
td, J ) 7.5 Hz, 5.0 Hz), 7.45 (1H, m), 7.50 (1H, td, J ) 8.2 Hz,
5.0 Hz), 7.88 (1H, d, J ) 8.0 Hz), 7.96 (1H, d, J ) 8.0 Hz); 13C
NMR (125.8 MHz, CD2Cl2) δ 20.76 (2.15), 37.14 (2.92), 37.81
(3.88), 77.96, 122.10 (7.88), 123.56 (7.96), 125.89 (7.41), 126.87
(7.50), 132.93(6.00), 135.10, 153.29, 153.90, 168.08, 170.66,
199.79; LC/MS (ESI, positive) m/e 593 (M + Na)+, 308 (M/2 +
Na)+, 286 (M/2 + H)+, 249 (M/2 - OAc + Na)+; high-resolution
MS (ES, positive) C30H22N2O6S2Na (M + Na) calcd 593.0811,
found 593.0839.
Synthesis. 4-(Benzothiazol-2-yl)phenol (8):34 4-Hydroxyben-
zaldehyde (3.66 g, 23 mmol) and 2-aminothiophenol (7.51 g, 60
mmol) were dissolved in 100 mL of Et2O. Silica gel (30 g) was
added to the mixture, and the solvent was slowly evaporated under
vacuum. The dry mixture was divided into four portions, placed
into four sealable Teflon vessels, and subjected to microwave
heating at a maximum power of 650 W for 3 min. The silica gel
(34) Kodomari, M.; Tamaru, Y.; Aoyama, T. Synth. Commun. 2004, 34,
3029–3036.
4470 J. Org. Chem. Vol. 74, No. 12, 2009

Hydrolysis and Photolysis of a Model Anti-Tumor Quinol Ester
was washed with 9:1 hexanes/EtOAc (4 × 50 mL), EtOAc (4 ×
50 mL), and EtOH (4 × 50 mL). Silica gel (15 g) was added to
the EtOAc wash again, and the solvent was removed. The dry
residue was washed again with 9:1 hexanes/EtOAc (2 × 50 mL),
EtOH (2 × 50 mL), followed by hot EtOH (1 × 50 mL). The
hexanes/EtOAc washes were discarded. All the EtOH filtrates were
combined and evaporated to dryness under vacuum to yield crude
8. A clean sample of 8 was obtained by recrystallization from EtOH:
mp 226.5-227.5 °C, lit34 mp 227 °C; 1H NMR (300 MHz, DMSO-
d6) δ 6.93 (2H, d, J ) 8.5 Hz), 7.39 (1H, td, J ) 7.5 Hz, 1.2 Hz),
7.49 (1H, td, J ) 7.7 Hz, 1.2 Hz), 7.92 (2H, d, J ) 8.7 Hz), 7.97
(1H, d, J ) 7.8 Hz), 8.07 (1H, d, J ) 7.8 Hz), 10.21 (1H, s); 13C
NMR (125.8 MHz, DMSO-d6) δ 116.04, 122.05, 122.25, 124.01,
124.84, 126.36, 128.99, 134.07, 153.69, 160.48, 167.40.
4-(Benzothiazol-2-yl)-2-methylphenol (11): This compound was
synthesized from 4-hydroxy-3-methylbenzaldehyde and 2-ami-
nothiophenol as described above for 8. The crude product was
purified by recrystallization from EtOH: mp 229-230 °C; IR 3061,
2923, 1601, 1401, 1277, 1120 cm-1; 1H NMR (500 MHz, DMSO-
d6) δ 2.21 (3H, s), 6.93 (1H, d, J ) 8.5 Hz), 7.38 (1H, td, J ) 7.6
Hz, 1.3 Hz), 7.49 (1H, td, J ) 7.6 Hz, 1.2 Hz), 7.75 (1H, dd, J )
8.3 Hz, 2.3 Hz), 7.83 (1H, d, J ) 1.5 Hz), 7.95 (1H, d, J ) 7.5
Hz), 8.06 (1H, d, J ) 7.5 Hz), 10.12 (1H, s); 13C NMR (125.8
MHz, DMSO-d6) δ 15.81 (2.21), 115.11 (6.93), 122.04 (8.06)
122.17 (7.95), 123.83, 124.77 (7.38), 124.99, 126.34 (7.49), 126.44
(7.75), 129.59 (7.83), 134.05, 153.70, 158.68, 167.58; LC/MS (ESI,
positive) m/e 242 (M + H)+, (ESI, negative) m/e 240 (M - H)-;
high-resolution MS (ES, positive) C14H12NOS (M + H) calcd
242.0634, found 242.0644.
4-Acetoxy-4-(benzothiazol-2-yl)-2,5-cyclohexadien-1-one (1):
8 (0.455 g, 2 mmol) was dissolved in 30 mL of AcOH, and the
resulting solution was stirred and kept in a water bath at 30-33
°C as a solution of PIDA (0.708 g, 2.2 mmol) in 15 mL of AcOH
was added dropwise over a period of 1.5 h. The reaction solution
was stirred for another 30 min at room temperature after completion
of the addition. AcOH was removed by rotary evaporation under
vacuum, and the residue was left on the vacuum pump overnight.
The dry residue was dissolved in ca. 25 mL of EtOAc followed by
filtration. The filtrate was collected and evaporated to dryness and
then subjected to trituration with hot 75/25 hexanes/EtOAc. The
solvent was removed, and oily yellow 1 was obtained. Further
purification was performed by multiple application of radial
chromatography on silica gel with 2:1 hexanes/EtOAc (yield
10-20%): mp 98-100 °C; IR 3065, 3050, 2930, 1751, 1668, 1629,
1432, 1370, 1214, 1163, 1040 cm-1; 1H NMR (300 MHz, DMSO-
d6) δ 2.19 (3H, s), 6.43 (2H, d, J ) 10.2 Hz), 7.36 (2H, d, J ) 9.9
Hz), 7.47-7.52 (2H, m), 8.02 (1H, dd, J ) 7.8 Hz, 1.5 Hz), 8.16
(1H, dd, J ) 7.5 Hz, 1.5 Hz); 13C NMR (75.5 MHz, DMSO-d6) δ
20.46, 75.66, 121.85, 122.76, 125.50, 126.18, 128.60, 134.22,
144.38, 151.89, 166.83, 167.90, 184.02; high-resolution MS (ES,
positive) C15H12NO3S (M + H) calcd 286.0532, found 286.0549,
C15H11NO3SNa (M + Na) calcd 308.0352, found 308.0370.
4-(Benzothiazol-2-yl)-4-hydroxy-2,5-cyclohexadien-1-one (2):
2 was obtained from hydrolysis of 1. The procedure is as follows:
250 mL of 0.02 M, 1/1 AcOH/AcONa buffer (pH 4.61 5 vol %
CH3CN, µ ) 0.5 (NaClO4)) was prepared and incubated in a water
bath at 80 °C prior to the addition of 1. A freshly made solution of
1 (50 mg, 0.175 mmol) in 1 mL of CH3CN was added in 0.1 mL
aliquots every 1 h to the incubated acetate buffer over a period of
10 h. After the last addition, the reaction mixture was kept in the
dark at 80 °C for another 5 h before refrigerating overnight. It was
brought back to room temperature the next day and extracted with
CH2Cl2 (5 × 50 mL). The combined CH2Cl2 extract was dried over
anhydrous Na2SO4, filtered, and evaporated to dryness under
vacuum. The crude product was purified by radial chromatography
with 3:2 hexanes/EtOAc (yield 71%): mp 108-109.5 °C, lit2 mp
63-66 °C; IR 3168, 1674, 1607, 1494, 1434, 1386, 1147, 1069,
1055 cm-1; 1H NMR (300 MHz, CD2Cl2) δ 4.35 (1H, s), 6.32 (2H,
d, J ) 9.9 Hz), 7.01 (2H, d, J ) 9.9 Hz), 7.42-7.55 (2H, m), 7.93
(1H, dd, J ) 8.7 Hz, 0.6 Hz), 8.01 (1H, dd, J ) 8.5 Hz, 0.8 Hz);
13C NMR (75.5 MHz, CD2Cl2) δ 71.76, 122.68, 124.06, 126.55,
127.24, 129.37, 136.82, 148.04, 153.51, 171.45, 185.00; high-
resolution MS (ES, positive) C13H10NO2S (M + H) calcd 244.0427,
found 244.0439.
Calculations: Density functional calculations were carried out
using the Gaussian 03 suite of programs.35 Geometry optimization
of the ground state oxenium ions 3 and 5a and the quinols 2, 7a,
and 7b were executed at the B3LYP/6-31G(d) level, and their
verification as a minimum and derivation of IR vibrational
frequencies and normal modes of vibration were executed using a
harmonic frequency analysis, which afforded all real frequencies.
Calculations at this level were previously reported for 5b.13
Vibrational frequencies for comparison with experimental data were
scaled by a factor of 0.9614 according to Scott and Radom.18
Acknowledgment. We thank the Donors of the American
Chemical Society Petroleum Research Fund (Grant # 43176-
AC4) for partial support of this work. Y.-T.W. thanks the
Department of Chemistry and Biochemistry at Miami University
for a Dissertation Fellowship, and K.J.J. thanks Miami Univer-
sity for an Undergraduate Summer Scholars Award and an
Undergraduate Research Award.
Supporting Information Available: Table S1, containing
all rate constants collected for 1 and 2; Tables S-2-S-10,
containing the Z-matrices for the optimized structures of singlet
(So) 2, 3, 5a, 7a, 7b, and 2-methylbenzothiazole and triplet (T1)
1
3, 5a, and 5b at the B3LYP/6-31G(d) level of theory; H and
13C NMR spectra for 1, 2, 9, 10, 11, 12, and 13; and complete
ref 35. This material is available free of charge via the Internet
at http://pubs.acs.org.
JO9008436
(35) Frisch, M. J.; et al. Gaussian 03, revision E.01; Gaussian, Inc.:
Wallingford, CT, 2005.
J. Org. Chem. Vol. 74, No. 12, 2009 4471