Synthesis of a water-soluble cyclodextrin modified hypocrellin and ESR study of its photodynamic therapy properties

Article abstract of DOI:10.1039/b202439j

A water-soluble cyclodextrin modified hypocrellin B (HBCD) was designed and synthesized. HBCD retained the phototherapeutic properties and exhibited much stronger photoinduced damage to calf thymus DNA (CT DNA) than hypocrellin B and mercaptoacetic acid substituted hypocrellin B (MAHB). The mechanism of electron transfer from CT DNA to the triplet state of HBCD was confirmed by steady-state electron spin resonance (ESR) and a time-resolved ESR study.

Full text of DOI:10.1039/b202439j

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of a water-soluble cyclodextrin modified hypocrellin and
ESR study of its photodynamic therapy properties
Zhi-Ze Ou, Jing-Rong Chen, Xue-Song Wang, Bao-Wen Zhang* and Yi Cao
Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing 100101,
P. R. of China. E-mail: g203@ipc.ac.cn
Received (in Montpellier, France) 6th March 2002, Accepted 15th April 2002
First published as an Advance Article on the web 12th July 2002
A water-soluble cyclodextrin modified hypocrellin B (HBCD) was designed and synthesized. HBCD retained
the phototherapeutic properties and exhibited much stronger photoinduced damage to calf thymus DNA (CT
DNA) than hypocrellin B and mercaptoacetic acid substituted hypocrellin B (MAHB). The mechanism of
electron transfer from CT DNA to the triplet state of HBCD was confirmed by steady-state electron spin
resonance (ESR) and a time-resolved ESR study.
Hypocrellins, including hypocrellin A (HA) and hypocrellin B
HB), are new photodynamic agents isolated from the fungus
of Hypocrella bambuase. They have been successfully
analytical grade was dried over calcium hydride and was dis-
tilled in the presence of fresh calcium hydride just before use.
All the other solvents were redistilled before use.
(
employed in the photodynamic therapy (PDT) treatment of
certain skin diseases. But their poor solubility in water limits
Calf thymus DNA (CT DNA) and cytochrome c were pur-
chased from Sigma Chemical Company. 2,2,6,6-Tetraethyl-4-
piperidone (TEMP), dicyclohexylcarbodiimide (DCC),
1-hydroxybenzotriazole hydrate (HOBT), 5,5-dimethyl-1-pyr-
roline-N-oxide (DMPO), superoxide dismutase (SOD), ethi-
dium bromide (EB) and 9,10-diphenylanthracene (9,10-DPA)
were all purchased from Aldrich Chemical Company, USA.
All experiments involving CT DNA were performed in buf-
fer solution (10 mM acetic acid ammonium salt, 100 mM
sodium chloride, pH 7), unless otherwise noted. DNA solu-
tions were prepared by dispersing the desired amount of
DNA in buffer solution and stirring overnight at a temperature
1
their further use in photodynamic therapy. Many efforts have
been made to enhance the water solubility of hypocrellins: Zou
2
and coworkers used liposomes as a delivery system, He et al.
3
introduced mercaptoacetic acid into HB, and Ma et al. used
4
cysteine as hydrophilic groups. In recent years, cyclodextrins,
5
as traditional drug delivery agents, have been successfully
applied to improve the water solubility and biocompatibility
of some phototherapeutic agents, porphyrins, fullerenes,
6
7
etc., while maintaining their therapeutic activities. However,
there are still no reports utilizing cyclodextrin to improve the
water solubility of hypocrellins. Herein, we report on the
synthesis of a cyclodextrin modified hypocrellin derivative
and its PDT properties.
ꢀ
below 4 C. In the experiments where titration with DNA was
ꢀ
required, the DNA solution was sonicated at 0 C for 10 min
using a Brason probe ultrasonicator. This operation signifi-
cantly reduced the viscosity of the DNA solutions and
permitted more accurate and precise titrations. The
concentration of CT DNA is expressed as the concentration
of nucleotide and was calculated using an average molecular
weight of 338 for a nucleotide and an extinction coefficient
It has been reported that hypocrellins and their derivatives
exhibited photosensitized cleavage and damage to isolated
and cellular DNA, whether under aerobic or anaerobic condi-
8
–10
tions.
In the presence of oxygen, the photoinduced cleavage
8
,9
involves reactive oxygen species, while He et al. discovered
that the electron transfer between the triplet state of hypocrel-
lin and CT DNA, studied by UV-Visible spectroscopy, could
initiate the photodamage of CT DNA in anaerobic condi-
ꢁ
1
ꢁ1
of 6600 M cm at 260 nm.
1
0
Synthesis of cyclodextrin modified hypocrellin (Scheme 1)
tions. Clearly, ESR is a more powerful technique for investi-
gating photoreaction intermediates. In this paper, steady state
ESR (SS-ESR) and time-resolved ESR (TR-ESR) were
employed to detect the radical intermediates responsible for
the photodamage of CT DNA by our cyclodextrin modified
hypocrellin under anaerobic and aerobic conditions.
Compound 2 (MAHB). Compound 2 was synthesized by
modification of a literature procedure. Compound 1 (HB)
3
(
2.8 mM) and mercaptoacetic acid (0.2 M) were put into a
three-necked round-bottomed flask with 500 ml of metha-
nol–buffer (1:3 v/v, pH 11) solution. The reaction mixture
was bubbled with air (5 ml min ) and irradiated ( > 470
ꢁ
1
nm) with a medium pressure sodium lamp for 10 h. The mix-
ture was isolated by TLC, and 0.63 g of monosubstituted HB
Experimental
1
was obtained (yield: 67%). HNMR (400 MHz, CHCl
3
-d
1
, d):
Chemicals
2
15.97 (s, 2H, exchangeable with D O, phenolic OH), 6.56 (s,
1
H), 3.75–4.2 (m, 14H), 2.38 (s, 3H), 1.93 (s, 3H). FAB-MS
([M – H] ): 617.
HB was prepared by dehydration of HA and purified by
ꢁ
1
1
recrystallization twice from acetone. b-Cyclodextrin (b-CD)
of reagent grade was recrystallized twice from water and dried
ꢀ
over night in vacuo at 110 C. p-Tosyl chloride of reagent grade
ꢀ
Compound 4. A solution of p-tosyl chloride (6.1 g, 32.0
mmol) in dry pyridine (200 ml) was added into a solution of
b-CD (47.0 g, 41.4 mmol) and DCC (8.54 g, 41.4 mmol) in pyr-
was recrystallized from petroleum ether (60–90 C). N,N-
Dimethylformamide (DMF) was dried over K CO and dis-
tilled in the presence of calcium hydride in vacuo. Pyridine of
2
3
ꢀ
ꢀ
idine (600 ml) cooled below 0 C. After stirring for 8 h at 0 C
1
130
New J. Chem., 2002, 26, 1130–1136
DOI: 10.1039/b202439j
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
Scheme 1 Synthesis of cyclodextrin modified hypocrellin.
and 24 h at room temperature, the reaction was stopped by
adding 2 ml of water to the solution and the mixture was eva-
ulation. Samples were injected quantitatively into specially
made quartz capillaries for ESR analysis, and purged with
argon, air or oxygen for 30 min in the dark, depending on
the experimental requirements, and illuminated directly in
the cavity of the ESR spectrometer with a Nd:YAG laser
(532 nm, 5–6 ns pulse width, 10 Hz repetition frequency, 10
mJ pulse energy).
ꢀ
porated in vacuo (10 mmHg) at 40 C to dryness. The crude
product was precipitated two times from DMF–acetone (100
ml:1000 ml). The precipitate was collected and recrystallized
1
three times from water (42% yield). HNMR (DMSO-d , d):
6
7
(
.7 (d, 2H), 7.4 (d, 2H), 5.6 (s, 14H), 4.85 (s, 7H), 4.5–3.3
m, 47H), 2.42 (s, 3H).
ꢃ
ꢁ
The superoxide anion (O ), hydroxyl radical ( OH) and
2
ꢃ
1
2
singlet oxygen ( O ) were detected by ESR using DMPO or
Compound 5. Two grams of compound 4 were dissolved in
0 ml of ethylenediamine. After stirring for 20 h under reflux,
00 ml of acetone was added to the solution. The mixture was
TEMP as the spin traps, respectively.
3
3
TR-ESR measurements were also performed on a Bruker
ESP-300E spectrometer, while a boxcar integrator (Stanford
SR 250) and a digital oscilloscope (LeCrory 9350A) were
added to monitor transient signals. The data from a SR250
Computer Interface Module were transferred to a PC, in which
data was stored and processed using the SR272 Data Acquisi-
tion Program.
filtered and the precipitate was recrystallized from 95% etha-
1
nol. HNMR (DMSO-d
6
, d): 6.0–5.6 (br m, 14H), 4.90 (s,
H), 4.2–3.6 (br m, 47H), 2.70 (m, 4H), 2.5 (s, 1H), 2.3 (s, 2H).
7
Compound 6 (HBCD). Compound 2 (10 mmol) and com-
pound 5 (15 mmol) were dissolved in dry DMF (20 ml), and
then DCC (10 mmol) and HOBT (14 mmol) was added. After
stirring for 48 h at room temperature, 200 ml of acetone was
poured into the reaction mixture. The mixture was filtered.
The precipitate was dissolved in 10 ml distilled water and this
solution was directly loaded onto a Sephadex C-25 column (3
cm ꢂ 40 cm). The product was eluted from the column with
distilled water. Fractions, each of 10 ml, containing the pro-
duct were analyzed by TLC (eluant: acetic acid–ethyl acet-
1
Quantum yield of O
2
generation
The photobleaching of 9,10-diphenylanthracene (DPA) [eqn.
genera-
1
1)] was used to determine the quantum yields of O
(
tion by HBCD, MAHB and HB.
2
1
2
ate–water 8:1:2, R
product were combined and evaporated to dryness at 40 C
in vacuo (10 mm Hg). The final product (30% yield), 6-
f
¼ 0.50). The fractions containing the
ꢀ
0
ð1Þ
deoxy-6-N-[N -(5-monomercapto acid hypocrellin B)diami-
noethyl]-b-cyclodextrin (HBCD), was characterized by
MALDI-TOF and HNMR. HNMR: 8.04 (s, 1H), 6.42 (s,
1
1
1
2
H), 5.74 (s, 14H), 4.81 (s, 7H), 4.55 (s, 6H), 3.0–4.1 (62H),
.77 (s, 1H), 2.07 (s, 3H), 1.89 (s, 3H). MALDI-TOF: 1780
+
An oxygen-saturated solution of DPA and these three kinds
of hypocrellins in DMSO were irradiated with light of 436 nm
[
M – Na ].
1
separately.The quantum yields of singlet oxygen ( O ) sensi-
2
ESR measurements
The ESR spectra were recorded at 25 C on a Bruker ESP-
tized by HBCD and MAHB were determined by using HB
as a reference whose singlet oxygen quantum yield has been
ꢀ
1
2
3
00E spectrometer at 9.8 GHz, X-band with 100 Hz field mod-
measured (0.76).
New J. Chem., 2002, 26, 1130–1136
1131

View Article Online
Reduction of cytochrome c
PDT because both isomers have the same photochemical and
photophysical properties.
ꢃ
ꢁ
) generation is often
The efficiency of superoxide anion (O
determined by the reduction of cytochrome c.
2
1
3,14
Compound 4 was prepared by reaction of b-CD with tosyl
chloride in dry pyridine. A low yield (ca. 31%) was often
After reduc-
tion by O2 , the ferric cytochrome c (Fe -cyt c) is trans-
18
ꢃ
ꢁ
3+
2
+
obtained in this reaction because it is difficult to completely
dehydrate the b-cyclodextrin. When we added an equimolar
amount of dicyclohexylcarbodiimide (DCC) to serve as a dehy-
drating agent, the yield of compound 4 increased to 42%.
The UV-visible spectra of compound 6 (HBCD) and HB are
similar (Fig. 1). The maximum concentration of HBCD in the
buffer solution can reach 10 mM, while that of HB is less than
formed into ferrous cytochrome c (Fe -cyt c). UV-Vis
absorption of cytochrome c changes remarkably at 550 nm
where the molar extinction coefficient is 8900 M cm for
ꢁ
1
ꢁ1
3
+
ꢁ1
ꢁ1
2+
Fe -cyt c and 29500 M cm for Fe -cyt c. By monitoring
the optical density change of the photoreaction system at 550
nm, the percentage of reduced cytocrome c can be determined
according to eqn. (2):
0.05 mM, which means that the water solubility of HBCD is
enhanced significantly.
Að550Þ ꢁ Að550Þ
t
0
%
cytred ¼ 100 ꢂ
ð2Þ
eð550Þ
red
Að550Þ ꢂ
ꢁ Að550Þ0
0
eð550Þ
Electron transfer between HBCD and CT DNA under anaerobic
conditions
ox
0 t
where A(550) and A(550) are the absorbance at 550 nm initi-
ally and after irradiation for time t, respectively. The O
generation efficiency can thus be estimated.
ꢃ
ꢁ
After illuminating an argon-saturated DMSO–buffer solution
(1:1 v/v, pH 7) of HBCD (0.1 mM) for 15 min with the laser
at 532 nm, an ESR spectrum (Fig. 2, spectrum A) was
recorded. This ESR signal increased in intensity when increas-
ing the concentration of HBCD, suggesting that the self-elec-
tron transfer between the ground and the excited states of
HBCD:
2
Ethidium bromide (EB) assay for DNA cleavage
A simple assay for DNA cleavage was applied based on the ca.
2
EB upon intercalation into DNA.
0-fold enhancement of the fluorescence intensity exhibited by
When the concentration
1
5–17
HBCD þ HBCD ! HBCD þ HBCDꢃþ
ꢄ
ꢃꢁ
ð4Þ
of EB is more than twofold that of the DNA base pair, the
fluorescence intensity of EB is linearly proportional to the con-
centration of DNA base pair. Any process in which the poten-
tial EB binding site is destroyed results in a decrease in
fluorescence intensity.
After preparation of solutions of EB–DNA buffer solution
80 mM EB, 40 mM CT DNA), HB, MAHB or HBCD was
(
added into 8 ml of EB–DNA buffer solution at a final concen-
tration of these hypocrellins of 10 mM. These mixture solutions
were divided into 4 aliquots and irradiated in a ‘merry-go-
round’ apparatus with a medium pressure sodium lamp
1
0
(
hn > 470 nm). The aliquots were removed at various times
and analyzed with the fluorescence emission spectrum excited
at 510 nm and measured from 525–800 nm.
The percentage of binding site remaining (%BSR) at a given
time t was calculated from the following equation:
ꢀ
ꢁ
I0 ꢁ It
%
BSR ¼ 100 ꢂ 1 ꢁ
ð3Þ
I0 ꢁ Ibuffer
Fig. 1 Absorption spectra of HB, HBCD and MAHB in DMSO–
buffer solution (1:1 v/v, pH 7).
where I
0
, I
t
, and Ibuffer denote the integrated fluorescence
intensities before irradiation, after irradiation for time t, and
of DNA-free buffer, respectively.
Results and discussion
Synthesis
Compound 2 (MAHB) was once prepared with a yield of only
12% by photoreaction of mercaptoacetic acid with Hypocrellin
B (HB) in the presence of O . The mechanism for this photo-
3
2
reaction has been proven to be nucleophilic addition of mer-
captoacetic acid anion to triplet HB followed by oxidation.
The role of O
one hand, O is indispensable; on the other hand, too high a
2
in this reaction, however, is complex. On the
2
2
concentration of O will ether quench the HB triplet or speed
up the oxidation of mercaptoacetic acid anion, and conse-
quently reduce the yield of mono- or bimercaptoacetic acid
substituted HB. Therefore, it is necessary to optimize the con-
Fig. 2 Spectrum A, photoinduced ESR spectrum of a deaerated solu-
tion of HBCD (0.1 mM) in DMSO–buffer (1:1 v/v, pH 7), illuminated
with 532 nm laser light for 15 min. Spectrum B, similar to spectrum A
but in the presence of calf thymus DNA (2.85 mM). Spectrum C, simi-
lar to spectrum A, but in the presence of aniline (2.85 mM). Spectrum
D, similar to spectrum B but with the addition of methylene blue (0.1
mM). Spectral parameter settings: X-band, microwave power, 8.0 mW;
time constant, 20.48 ms; modulation amplitude, 2.0 G; receiver gain,
centration of O
factor was controlled carefully, and a collection yield as high
as 67% was obtained when O was bubbled into the reaction
2
to achieve high yields. In our experiments, this
2
ꢁ
1
system at the rate of 5 ml min and the molar ratio of mercap-
toacetic acid to HB was enhanced from 50 to 100 (Scheme 1).
The mixture of 5- or 8-monomercaptoacetic acid substituted
hypocrellin B (MAHB) is able to meet the requirements for
5
1 ꢂ 10 .
1
132
New J. Chem., 2002, 26, 1130–1136

View Article Online
might be responsible for the origin of the ESR signal. Due to
their strong oxidation ability, quinone radical cations can only
be detected in solutions with high ionization potential, such as
1
9
20
21
CFCl
3
,
ture. Hence this ESR signal is attributed to the HBCD radical
Freon-133 and trifluoacetic acid at low tempera-
anion. The g value of 2.002 is in good agreement with the lit-
2
erature. The quenching of this ESR signal by O
2
2
supports
such an assignment. Similar results were also obtained in the
case of HA or HB.
2
3
It is expected that the ESR signal intensity of HBCD radical
anion will increase if a good reductant is present in the irra-
diated system, such that the electron transfer from the reduc-
tant to the excited state of HBCD:
HBCD þ Reductant ! HBCDꢃꢁ þ Reductantꢃþ ð5Þ
ꢄ
becomes more favorable than the self-electron transfer process
of HBCD [eqn. (4)]. A remarkable increase in intensity of ESR
signal was indeed observed (Fig. 2, spectrum C) when aniline,
a strong electron donor, was added, and this observation in
turn further proves the assignment of this ESR signal to
HBCD radical anion. The intensity of the ESR signal also
increased when CT DNA was added (Fig. 2, spectrum B),
though the increase is smaller than in the case of aniline (see
Table 1) owing to the weaker reductive ability of CT DNA
relative to aniline. This suggests that CT DNA interacts with
the excited state of HBCD as an electron donor:
Fig. 3 Spectrum A, TR-ESR spectrum of argon-saturated HBCD (1
mM) DMSO–buffer solution (1:1 v/v, pH 7) measured 420 ns after the
laser flash. Spectrum B, similar to spectrum A but in the presence of
calf thymus DNA (2.85 mM). Spectrum C, similar to spectrum A
but in the presence of aniline (2.85 mM). Spectrum D, similar to spec-
trum B, but with added methylene blue (1 mM). Spectral parameter
settings: microwave power, 8.0 mW; time constant, 20.48 ms; modula-
5
tion amplitude, 2.0 G; receiver gain, 1 ꢂ 10 .
and the triplet HBCD. After pulsed laser illumination of
HBCD, the triplet state T was generated by efficient intersys-
1
HBCD þ CT DNA ! HBCD þ CT DNAꢃþ
ꢄ
ꢃꢁ
ð6Þ
1
tem crossing (ISC) from the excited singlet state S . The ISC
process is spin selective and creates spin polarized triplet states.
In the presence of CT DNA, a precursor complex between CT
DNA and triplet HBCD is formed in which electron transfer
takes place with CT DNA serving as the electron donor.
Accompanying the electron transfer reaction, the spin polari-
zation of the HBCD triplet state is transferred to the produced
radical anion. Thus, polarized HBCD anion radicals are
formed. The precursor, triplets of HBCD, may have popula-
and, accordingly, gives rise to a stronger ESR signal of HBCD
radical anion.
Photoreactions often occur in a non-thermal equilibrium
state, and as a result, the created radicals usually possess a
non-Boltzmann distribution of the spin state, which leads to
2
4
chemically induced dynamic electron polarization (CIDEP).
As a powerful technique for the study of CIDEP, transient
ESR often provides valuable information about the details of
the initial stages of reactions that cannot be obtained by other
methods. The short-lived intermediate radicals involved in the
reaction can be detected and identified and thus the reaction
mechanism can be clarified. Whether the reaction is taking
place from an excited singlet state or a triplet state can only
be easily deduced from the pattern of the spectrum in the form
of enhanced absorption and/or stimulated emission peaks in
tion in the T
mann equilibrium at the time of reaction. An excess
population in the T level produces a stimulated emission
+ ꢁ
or T spin states that deviate from the Boltz-
+
ESR signal. Similar to the SS-ESR, in the presence of aniline,
the TR-ESR signal (Fig. 3, spectrum C) increased in intensity,
while only a very weak emissive TR-ESR signal was obtained
at the absence of CT DNA or aniline (Fig. 3, spectrum A). The
trend in the changes of polarized ESR signal intensity for the
three solutions (see Table 1) agrees with the assignment that
the polarized signals originate from the HBCD radical anions.
In our TR-ESR experiment, only HBCD anion radical was
observed. There were some reports on electron transfer from
2
5
the ESR spectrum. Therefore, TR-ESR was applied in this
work to gain insight into the photodamage mechanism of
CT DNA by HBCD under anaerobic conditions.
A pure emissive TR-ESR signal (Fig. 3, spectrum B) was
observed 420 ns after illuminating a DMSO–buffer solution
2
6–28
pyrimidine to triplet states of anthraqunone derivatives,
(
1:1 v/v, pH 7) of CT DNA and HBCD with a 532 nm pulsed
where the pyridine cation radicals were detected by Fourier
Transform ESR (FT-ESR), which has a high resolution on
the nanosecond time scale. An alternative approach to observe
ESR signals of cation radicals of DNA bases is to carry out
laser. This signal is located at the same position as that of the
HBCD anion radical generated in SS-ESR (Fig. 2, spectrum B)
and therefore was attributed to the polarized ESR signal of
HBCD radical anion. This assignment is supported by the fact
that the signal can be quenched efficiently by O . This comple-
2
tely pure emissive TR-ESR signal can be attributed to the tri-
plet mechanism (TM) of the photoreaction between CT DNA
2
9
measurements at very low temperature (77 K). However,
the CT DNA cation radical signal was not observed in our
,
experiments, probably because of the limitation of our ESR
instrument.
When methylene blue was added into the solution mixture
of CT DNA and HBCD (spectra D in Figs. 2 and 3), both
the SS-ESR and the TR-ESR signal dramatically weakened.
This is because methylene blue quenchs the triplet of HBCD
and inhibits the electron transfer from CT DNA to triplet
HBCD. From all of these results, electron transfer via CT
DNA and the triplet HBCD can be definitely concluded.
Table 1 Relative intensity of ESR signal obtained with HBCD,
HBCD + CT DNA, HBCD + aniline, HBCD + CT DNA + MB
Relative Intensity
Sample
SS-ESR
TR-ESR
ꢃ
ꢁ
)
HBCD
1.00
1.54
1.77
0.58
1.00
4.13
7.43
1.09
Generation of superoxide anion radical (O
2
HBCD + DNA
HBCD + aniline
HBCD + DNA + MB
ꢃ
ꢁ
The photoinduced SS-ESR spectrum of HBCD disappeared
when oxygen was bubbled through the argon-saturated
DMSO solution containing HBCD (0.1 mM) after irradiation.
New J. Chem., 2002, 26, 1130–1136
1133

View Article Online
Fig. 4 Spectrum A, ESR spectrum of DMPO–superoxide radical
adduct produced from irradiation of an oxygen-saturated DMSO solu-
tion of HBCD (0.2 mM) and DMPO (50 mM) for 2 min. Spectrum B,
in the absence of HBCD, oxygen or light. Spectrum C, similar to spec-
ꢁ1
trum A but in the presence of SOD (40 mg ml ). Spectral parameter
settings: microwave power, 1.0 mW; time constant, 20.48 ms; modula-
tion amplitude, 1.0 G; receiver gain, 1 ꢂ 10 .
5
Fig. 5 Reduction of cytochrome c (58 mM) in oxygen-saturated buf-
fer solution (pH 7) by hypocrellins (10 mM) as a function of irradiation
time.
One possible explanation for this observation is that electron
ꢃ
ꢁ
ꢃ
ꢁ
transfer from HBCD to oxygen takes place and O2 is pro-
duced:
HBCD þ O2 ! HBCD þ O2ꢃꢁ
ꢃꢁ
ð7Þ
The following experiments were carried out to examine this
possibility. When an oxygen-saturated DMSO solution con-
taining HBCD (0.1 mM) and DMPO (50 mM) was irradiated,
ꢃ
ꢁ
a typical spectrum of the DMPO–O
2
adduct was immedi-
ately seen(Fig. 4, curve A). The hyperfine coupling constants
and the position is in good agreement with the literature
N
H
H
30
(
a
¼ 0.13 mT, a
b
¼ 0.99 mT, a
g
¼ 0.015 mT). Control
ꢁ
experiments confirmed that oxygen, HBCD, and light were
all necessary to produce DMPO–O
ESR spectrum could be suppressed by the addition of SOD
ꢃ
2
(Fig. 4, curve B). The
Fig. 6 Spectrum A, ESR spectrum of DMPO–hydroxyl radical
adduct resulting from irradiation of an oxygen-saturated buffer solu-
tion of HBCD (0.4 mM, pH 7) and DMPO (50 mM). Spectrum B,
in the absence of HBCD, oxygen or light. Spectrum C, similar to spec-
trum A, but in the presence of sodium benzoate (10 mM). Spectrum D,
similar to spectrum A, but with HB instead of HBCD. Spectrum E,
similar to spectrum A, but MAHB was used instead of HBCD. Spec-
ꢁ
1
(
40 mg ml ), an efficient scavenger of superoxide anion (Fig.
4
ESR spectrum to DMPO–O
Although the spin trapping method is more sensitive than
the reduction of cytochrome c for the measurement of O
the cytochrome c method is a more traditional method to
, curve C), which further confirms the assignment of the
adduct.
ꢃ
ꢁ
2
ꢃ
ꢁ
,
2
tral parameter settings: microwave power, 1.0 mW; time constant,
2
5
0.48 ms; modulation amplitude, 2.0 G; receiver gain, 1 ꢂ 10 .
ꢃ
ꢁ 31
quantitatively analyze O
2
. When oxygen and cytochrome
3
+
c (cyt Fe ) were present in the buffer solution containing
HBCD, irradiation resulted in an increase in absorption inten-
sity at 550 nm, which is characteristic of the reduction of cyto-
spins of the nitrogen atom and the hydrogen atom in the b-
3
+
2+
N
H
chrome c (cyt Fe ! cyt Fe ). If SOD is also included in the
position: a ¼ a ¼ 14.9 G. The two identical coupling con-
stants from b-N and b-H give a four-line ESR spectrum with
an intensity ratio of 1:2:2:1. The values of the coupling con-
reaction solution, the reduction of cytochrome c is inhibited,
suggesting that the cytochrome is reduced by O
ꢃ
ꢁ
2
rather than
3
2
by direct electron transfer from hypocrellin anion radicals to
the cytochrome c. Attempts were also made to compare the
O2 generation efficiency of HBCD with that of HB and
stants are in good agreement with the literature.
Control experiments ensured that no signal was obtained
without light, oxygen, or HBCD (Fig. 6, spectrum B). The pre-
sence of sodium benzoate, a scavenger of hydroxyl radical,
decreases the ESR signal significantly (Fig. 6, spectrum C).
These observations confirm the assignment of the spectrum
ꢃ
ꢁ
ꢃ
ꢁ
with a
MAHB. Fig. 5 suggests that HBCD generates O
lower efficiency than that of MAHB, but higher than that of
2
ꢃ
ꢁ
HB. The higher O
2
generation efficiency of MAHB can be
ꢃ
attributed to its stronger electron donating properties due to
the presence of the mercapto group. When cyclodextrin is
introduced, the bulkiness of cyclodextrin weakens the electron
transfer between the exited state of HBCD and ground state
shown in Fig. 6 (spectrum A) to the DMPO– OH radical
adduct.
ꢃ
In addition, we compared the OH generation efficiency of
HBCD with HB and MAHB at the same concentration (Fig.
6, spectra D and E). It can be seen that HBCD generates
ꢃ
ꢁ
HBCD and reduces the generation of HBCD , from which
ꢁ
ꢃ
ꢃ
ꢁ
ꢃ
O
2
can be generated via eqn. (7). As a result, the O
eration efficiency of HBCD is lower than that of MAHB.
2
gen-
OH with lower efficiency than MAHB, but higher than that
ꢃ
ꢁ
of HB. Interestingly, HBCD, MAHB and HB generate O
2
with the same trend (c.f. Fig. 4).
Generally speaking, there are two pathways to generate
hydroxyl radical in the hypocrellin photosensitization sys-
tem. One is via the superoxide-driven Fenton reaction
[eqns. (8) and (9)]:
ꢃ
Generation of hydroxyl radical ( OH)
3
3,34
After addition of DMPO (50 mM) to an oxygenated buffer
solution of HBCD (0.6 mM, pH 7), irradiation with the laser
at 532 nm gave rise to an ESR spectrum of the spin adduct
DMPO– OH (Fig. 6, spectrum A), which is characterized by
two coupling constants connected with the non-zero nuclear
Fe þ O2ꢃꢁ ! Fe þ O2
3
þ
2þ
ð8Þ
ð9Þ
ꢃ
2þ
3þ
ꢁ
ꢃ
Fe þ H2O2 ! Fe þ OH þ HO
1
134
New J. Chem., 2002, 26, 1130–1136

View Article Online
in which H
radical:
2
O
2
originates from dismutation of the superoxide
O2 þ O2ꢃꢁ ! H2O2 þ O2
ꢃꢁ
ð10Þ
3
+
2+
while Fe , Fe or other transition metal ions with multiple
oxidation states are often present in trace amounts in aqueous
solution.
The other pathway is via the reaction of semiquinone radical
anion with H O :
2
2
ꢃꢁ
ꢁ
ꢃ
HBCD þ H2O2 ! HBCD þ OH þ HO
ð11Þ
Both pathways involve the generation of superoxide anion
radical; this is why the hypocrellin derivative with higher
ꢃ
ꢁ
ꢃ
O
2
generation efficiency can produce OH more efficiently.
Generation of singlet oxygen
1
It has been observed that O
2
was involved in many photoox-
idation reactions sensitized by hypocrellins. O is generated
Fig. 8 DPA (0.6 mM) bleaching in oxygen-saturated DMSO photo-
sensitized using hypocrellins ([HB] ¼ 11.2 mM, [HBCD] ¼ 10 mM,
[MAHB] ¼ 13.5 mM) by measuring the absorbance decrease (DA) at
374 nm for different irradiation times.
3
5 1
2
via energy transfer from the triplet of hypocrellins to the
ground state oxygen. Moan and Wold have reported that
3
6
1
2
TEMP can easily react with O to yield a nitroxide, TEMPO,
which is long-lived and hence can be detected readily by SS-
ESR spectroscopy. Irradiation of oxygen-saturated DMSO
ꢃ
ꢁ
ꢃ
1
, OH and O by HBCD photosen-
The generation of O
2
2
ꢁ1
sitization indicates that it maintains the photodynamic activity
in terms of type I and type II mechanisms.
solutions containing HBCD (0.1 mM), SOD (40 mg ml ),
HCOONa (0.1 M) and TEMP (20 mM) afforded a typical
three-line ESR spectrum (Fig. 7, spectrum A). The coupling
N
constants a ¼ 16.0 G and g ¼ 2.0056 are in good agree-
3
7
Photoinduced damage to CT DNA
ment with those of TEMPO in the literature. Control
experiments confirmed that oxygen, HBCD, and light are all
necessary to produce this ESR spectrum (Fig. 7, spectrum
It has been reported that the combining of cyclodextrin with
the potent DNA intercalating agent anthrylamine provided
for a superior water solubility as well as an increased affinity
3
B). The addition of NaN (1 mM), a traditional singlet oxygen
scavenger, diminishes the ESR signals significantly (Fig. 7,
spectrum C).
38
for the negatively charged groove of double-strand DNA.
So it could be expected that this cyclodextrin modified hypo-
crellin derivative (HBCD) also possesses a higher affinity for
the negatively charged groove of CT DNA than HB and
MAHB, and therefore induces greater photodamage to the
DNA.
When HB was added into the anaerobic buffer solution of
EB–DNA (80 mM EB, 40 mM CT DNA), 20% of binding sites
were destroyed during a 30 min irradiation as calculated fro-
meqn. (3) (Table 2). Under the same conditions, MAHB and
HBCD exhibited similar destruction ability, a little higher than
HB. Under anaerobic conditions, the photodamage of CT
DNA results from the electron transfer between CT DNA
and triplet HB, MAHB or HBCD. But DNA is not a strong
reductant and therefore the poor photoinduced electron
The photooxidation of DPA to its endo-peroxide derivative
by singlet oxygen is usually used to detect singlet oxygen
1
2
formed from hypocrellin. In order to compare the quantum
yield of HBCD with that of HB and MAHB, the DPA bleach-
ing method was adopted by using HB as reference.
Control experiments indicated that the presence of O2 ,
hypocrellins, and light are all essential for DPA bleaching.
Moreover, the addition of NaN (1 mM) almost completely
3
inhibited DPA bleaching. These results confirmed that the
bleaching of DPA resulted from the reaction of DPA with
1
1
2 2
O formed by hypocrellin photosensitization. The O quan-
tum yields for both HBCD and MAHB in DMSO are esti-
mated to be 0.22 by using HB (0.76) as reference (Fig. 8).
Table 2 Photocleavage of CT DNA by the hypocrellins (10 mM)
detected by remaining binding site (BSR%) of ethidium bromide to
the damaged CT DNA under different conditions [CT DNA] ¼ 40
mM, [EB] ¼ 80 mM.
Irradiation time/min
Sample
10
20
30
a
Control experiment
97.6
90
95.6
82.4
85.3
84.2
48
94.1
77.1
80.8
77.6
40.6
68.5
55.9
59.5
55.9
53.7
HBCD + N
HB + N
MAHB + N
2
2
89.6
89.1
64.6
79.5
75.8
80.3
73.6
76.2
2
HBCD + O
HB + O
MAHB + O
2
2
74.1
61.9
68.8
61.4
65.6
2
Fig. 7 Spectrum A, ESR spectrum produced by irradiation of an
oxygenated DMSO–buffer solution (1:1 v/v, pH 7) containing HBCD
b
HBCD + O
HBCD + O
HBCD + O
2
+ SOD
c
ꢁ1
2
2
+ NaN
5
+ C H COONa
3
(
0.1 mM), SOD (40 mg ml ), HCOONa (0.1 M) and TEMP (10 mM)
for 2 min. Spectrum B, in the absence of HBCD, oxygen or light. Spec-
trum C, similar to spectrum A but in the presence of NaN (1 mM).
d
6
a
In the absence of hypocrellins. [SOD] ¼ 40 mg mlꢁ1
b
c
3
.
[NaN
3
] ¼ 1
Spectral parameter settings: microwave power, 3.17 mW; time con-
d
mM. [C
4
6
H
5
CO
2
Na] ¼ 10 mM
stant, 10.24 ms; modulation amplitude, 1.0 G; receiver gain, 1 ꢂ 10 .
New J. Chem., 2002, 26, 1130–1136
1135

View Article Online
transfer efficiency perhaps accounts for the none observation
of higher photodamage with HBCD, contrary to expectation.
In the presence of oxygen, the photosensitized cleavage effi-
ciency of HB, MAHB, and HBCD to CT DNA all increased
References
1
N. W. Fu, Y. X. Chu and J. Y. An, Acta Pharm. Sin., 1989, 10,
71.
3
2
3
W. Zou, J. Y. An and L. J. Jiang, Sci.China, Ser. B, 1996, 26, 206.
Y. Y. He, J. Y. An, W. Zou and L. J. Jiang, J. Photochem. Photo-
biol., B, 1998, 44, 45.
(
Table 2). However, HBCD inflicts stronger photoinduced
damage to CT DNA than HB and MAHB. Besides the elec-
tron transfer damage pathway, the reactive oxygen species
4
5
6
L. Ma, M. H. Zhang and L. J. Jiang, Kexue Tongbao, 1991, 36,
1551.
D. Duchen, Cyclodextrins and their Industrial Use, Edition de
Sante, Paris, 1987.
J. G. Moser, A. Ruebner, A. Vervoots and B. Wangner, in Pro-
ceedings of the Eighth International Symposium on Cyclodextrins,
ed. J. Szetli, Kluwer, Boston, 1996, p. 71.
ꢃ
ꢁ 1
, O
ꢃ
(
photosensitized cleavage of CT DNA, which is confirmed by
O
2
2
, OH) are included in the process of hypocrellin-
the addition of SOD, NaN
tion (Table 2).
Although HBCD generates O
3
or sodium benzoate into the solu-
ꢃ
ꢁ
ꢃ
2
, OH less efficiently than
1
MAHB and generates O
2
less efficiently than HB, the water
7
8
S. Samal and K. E. Geckler, Chem.Commun., 2000, 1101.
E. H. En, S. M. Xin and L. S. Cheng, Int. J. Radiat. Biol., 1992,
solubility and affinity for DNA are improved significantly rela-
tive to HB and MAHB, and as a resultant, HBCD exhibits
more efficient photodamage to CT DNA.
6
1, 213.
W. Zou, J. Y. An and L. J. Jiang, J. Photochem. Photobiol., B,
996, 33, 73.
9
1
10
11
12
Y. Y. He and L. J. Jiang, Biochem. Biophys. Acta, 2000, 1523, 29.
K. H. Zhao and L. J. Jiang, Youji Huaxue, 1989, 9, 252.
Z. J. Diwu and J. W. Lown, J. Photochem. Photobiol., A, 1992, 64,
Conclusions
273.
I. Fridovich, J. Biol. Chem., 1970, 245, 4053.
C. Hadjur, A. Jeunet and P. Jardonj, J. Photochem. Photobiol., B,
By strictly controlling the reaction conditions of HB photo-
reaction with the mercaptoacetic acid, the collection yield of
monosubstituted HB increased from 12% to 67%. The yield
from the tosylation of the b-cyclodextrin was likewise
increased from 31% to 42% when DCC was used as dehydrant.
Profitting from the above improvements, the cyclodextrin
modified HB (HBCD) was successfully synthesized. The water
solubility of HBCD was enhanced significantly because of the
introduction of the b-cyclodextrin group.
The electron transfer between triplet HBCD and CT DNA
was confirmed by SS-ESR and TR-ESR studies. A pure emis-
sive CIDEP signal of the hypocrellin, generated via the triplet
mechanism, is first reported in this paper. When CT DNA is
added to the HBCD solution, this CIDEP signal increases in
an obvious fashion. The increase of the CIDEP signal intensity
results from the electron transfer from CT DNA to the triplet
HBCD, which was proved by various control experiments. The
electron transfer between CT DNA and HBCD initiates the
photodamage of CT DNA under anaerobic conditions.
HBCD also preserves the photodynamic properties of HB
and can generate active oxygen species such as oxygen anion
radical, hydroxyl radical and singlet oxygen. All these active
oxygen species can inflict damage on CT DNA. The b-cyclo-
dextrin group reinforces the affinity of HBCD for CT DNA,
and allows the active oxygen species formed by HBCD to
damage CT DNA more efficiently. As a result, HBCD exhibits
better photodynamic therapy properties than HB and MAHB.
In a word, the introduction of the b-cyclodextrin group
enhances both water-solubility and the PDT suitability of
HBCD.
1
1
3
4
1994, 26, 67.
W. A. Pruetz, Radiat. Environ. Biophys., 1984, 23, 1.
1
16 W. A. Pruetz, Radiat. Environ. Biophys., 1984, 23, 7.
5
17 H. C. Birnboim and J. J. Jevcak, Cancer Res., 1981, 41, 889.
1
1
8
9
Y. Matsui and A. Okimoto, Bull. Chem. Soc. Jpn., 1978, 51, 3030.
H. Chandra, L. Portwood and M. C. R. Symons, J. Chem. Soc.,
Faraday Trans. 1, 1989, 85, 1799.
2
0
J. Mayer and R. Krasluklanls, J. Chem. Soc., Faraday Trans. 1,
1
991, 87, 2943.
21 M. C. Depew, Z. L. Liu and J. K. S. Wan, J. Am. Chem. Soc.,
1983, 105, 2480.
2
2
2
2
3
4
L. Y. Zang, B. R. Misra and H. P. Misra, Photochem. Photobiol.,
992, 56, 453.
1
Y. Z. Hu, J. Y. An, L. J. Jiang and D. W. Chen, J. Photochem.
Photobiol., A, 1995, 89, 45.
Y. N. Molin, Spin Polarization and Magnetic Effects in Radical
Reactions, Elsevier, Amsterdam, 1984.
25 K. A. McLauchlan, Modern Pulsed and Continious-Wave Electron
Spin Resonance, John Wiley and Sons Inc., New York, 1990.
2
2
6
7
J. Geimer and D. Beckert, J. Phys. Chem., 1999, 103, 3991.
J. Geimer, O. Brede and D. Beckert, Chem. Phys. Lett., 1997, 276,
411.
J. Geimer and D. Beckert, Chem. Phys. Lett., 1998, 288, 449.
2
29 M. D. Sevilla, J. Phys. Chem., 1976, 80, 1898.
8
30 J. R. Harbour and M. L Hair, J. Phys. Chem., 1978, 82, 1397.
3
3
3
1
2
3
S. P. Sanders, S. J. Harrison, P. Kuppusamy, J. T. Sylvester and
J. L. Zweier, Free Radical Biol. Med., 1994, 16, 753.
K. Lang, M. Wagnerova, P. Stopka and W. Dameran, J. Photo-
chem. Photobiol., A, 1992, 67, 187.
Y. Y. He, J. Y. An and L. J. Jiang, Free Radical Biol. Med., 1999,
27, 203.
34 Y. Z. Hu and L. J. Jiang, J. Photochem. Photobiol., B, 1996,
3, 51.
3
L. J. Jiang, Kexue Tongbao, 1990, 35, 1608.
3
3
3
5
6
7
Acknowledgements
J. Moan and E. Wold, Nature, 1979, 279, 450.
Y. Lion, M. Delmelle and A. V. Vorst, Nature, 1976, 263, 442.
This research was supported by NNSFC (No. 29971031,
2
38 T. Ikeda, K. Yoshida and H. J. Schneider, J. Am. Chem. Soc.,
1995, 117, 1453.
0073050) and CAS.
1
136
New J. Chem., 2002, 26, 1130–1136