Synthesis, G-quadruplexes DNA binding, and photocytotoxicity of novel cationic expanded porphyrins

Article abstract of DOI:10.1016/j.bioorg.2015.05.001

Intensive reports allowed the conclusion that molecules with extended aromatic surfaces always do good jobs in the DNA interactions. Inspired by the previous successful researches, herein, we designed a series of cationic porphyrins with expanded planar substituents, and evaluated their binding behaviors to G-quadruplex DNA using the combination of surface-enhanced raman, circular dichroism, absorption spectroscopy and fluorescence resonance energy transfer melting assays. Asymmetrical tetracationic porphyrin with one phenyl-4-N-methyl-4-pyridyl group and three N-methyl-4-pyridyl groups exhibit the best G4-DNA binding affinities among all the designed compounds, suggesting that the bulk of the substituents should be matched to the width of the grooves they putatively lie in. Theoretical calculations applying the density functional theory have been carried out and explain the binding properties of these porphyrins reasonably. Meanwhile, these porphyrins were proved to be potential photochemotherapeutic agents since they have photocytotoxic activities against both myeloma cell (Ag8.653) and gliomas cell (U251) lines.

Full text of DOI:10.1016/j.bioorg.2015.05.001

Bioorganic Chemistry 60 (2015) 110–117
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Synthesis, G-quadruplexes DNA binding, and photocytotoxicity of novel
cationic expanded porphyrins
b,
c
d
a
a
Shu-fang Jin a, Ping Zhao a, , Lian-cai Xu , Min Zheng , Jia-zheng Lu , Peng-liang Zhao , Qiu-lan Su ,
⇑
⇑
Hui-xian Chen a, Ding-tong Tang a, Jiong Chen a, Jia-qi Lin a
a School of Chemistry and Chemical Engineering, Guangdong Pharmaceutical University, No. 280, Waihuandong Road, Education Mega Centre, Guangzhou 510006, PR China
b School of Material and Chemical Engineering, Zhengzhou University of Light Industry, Zhengzhou 450002, PR China
c School of Basic, Guangdong Pharmaceutical University, No. 280, Waihuandong Road, Education Mega Centre, Guangzhou 510006, PR China
d School of Pharmacy, Guangdong Pharmaceutical University, No. 280, Waihuandong Road, Education Mega Centre, Guangzhou 510006, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 February 2015
Available online 8 May 2015
Intensive reports allowed the conclusion that molecules with extended aromatic surfaces always do good
jobs in the DNA interactions. Inspired by the previous successful researches, herein, we designed a series
of cationic porphyrins with expanded planar substituents, and evaluated their binding behaviors to
G-quadruplex DNA using the combination of surface-enhanced raman, circular dichroism, absorption
spectroscopy and fluorescence resonance energy transfer melting assays. Asymmetrical tetracationic por-
phyrin with one phenyl-4-N-methyl-4-pyridyl group and three N-methyl-4-pyridyl groups exhibit the
best G4-DNA binding affinities among all the designed compounds, suggesting that the bulk of the sub-
stituents should be matched to the width of the grooves they putatively lie in. Theoretical calculations
applying the density functional theory have been carried out and explain the binding properties of these
porphyrins reasonably. Meanwhile, these porphyrins were proved to be potential photochemotherapeu-
tic agents since they have photocytotoxic activities against both myeloma cell (Ag8.653) and gliomas cell
(U251) lines.
Keywords:
Expanded cationic porphyrins
G-quadruplex DNA
Spectral method
Photocytotoxic
Ó 2015 Published by Elsevier Inc.
1. Introduction
reported that quadruplex-binding and quadruplex-promoting
small molecules can in principle effectively inhibit both the cat-
Telomerase is expressed in over 90% of tumor cell lines but in
only a few, specialized, normal healthy cells. Thus, telomerase inhi-
bition represents a potentially highly selective target for anticancer
drug design. Currently there is considerable interest in finding
molecules that can inhibit telomerase and potentially act as
anti-cancer drugs. For excellent recent reviews of the relationship
among telomeres, telomerase, telomerase inhibitors, and cancer,
see Saretzki and Eissenberg [1,2].
alytic and capping functions of telomerase [3–5]. There is now con-
siderable interest in designing new quadruplex binding ligands as
effective therapeutic agents. Many small aromatic ligands, includ-
ing porphyrins, quinacridones, anthraquinones, phenanthrolines,
substituted triazines and acridines, have been shown to bind and
stabilize the quadruplex structure of G4 DNA [4–6]. In
particular, the free base porphyrin (5, 10, 15, 20-tetra
(N-methyl-4-pyridyl)porphyrin) (TMPyP4, see molecular structure
in Fig. 2 as porphyrin 3) has been intensively researched as a good
G4-binding model because of its unique symmetric structure and
high positive charge [7–9]. Moreover, intensive researches on such
as anthraquinones, phenanthrolines and polypyridine, suggested
that extended aromatic surfaces always do favors to the
drug-DNA interactions [10–15]. However, although intensive
efforts were donated on the interactions between TMPyP4 and
G4-DNA, so far the reports on TMPyP4-like porphyrins with more
extended aromatic substituents are relatively rare.
The single-stranded telomere segment at the 30 ends of chromo-
somal DNA are able to form a variety of four-stranded structures
known as G-quadruplexes (G4) in vitro, based on a G-tetrad struc-
ture of four Hoogsteen-paired, coplanar guanines (Fig. 1). The
potential application of G4 to cancer treatment has been the driv-
ing force behind the investigation of ligands that stabilize G4
and/or induce their formation (G4-ligands). It is frequently
⇑
Corresponding authors. Fax: +86 760 88207939 (P. Zhao), +86 371 86639676
(L.C. Xu).
On the other hand, in recent years, photodynamic therapy
(PDT), which requires a combination of photosensitizer, light, and
oxygen to promote the destruction of localized neoplastic lesions,
E-mail addresses: zhaoping666@163.com (P. Zhao), miss_xulc@126.com
(L.-c. Xu).
http://dx.doi.org/10.1016/j.bioorg.2015.05.001
0045-2068/Ó 2015 Published by Elsevier Inc.

S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
111
cytotoxicity, it is expected that these new porphyrins may bring
us new surprising on the discovery of ligands with higher bio-
chemical abilities.
2. Experimental
2.1. Synthesis
Classical Adler–Longo method [23] was used to synthesize the
porphyrins.
5, 10, 15, 20-tetra (phenyl-4-N-methyl-4-pyridyl) porphyrin
(1): A mixture of 120 mL propionic acid, 7 mL acetic anhydride
and 1.82 g 4-(4-pyridinyl)benzaldehyde (10 mmol) was heated at
130 °C with stirring for 0.5 h. 0.8 mL pyrrole (10 mmol) diluted to
10 mL with propionic acid was then added from the dropping fun-
nels to the refluxing mixture. The resulting mixture was refluxed
for 1.5 h. The solvent was subsequently evaporated under reduced
pressure. The residue was then purified by column chromatogra-
phy. The purple product was methylated by an excess amount of
methyl iodide in DMF under room temperature overnight.
Diethyl ether was then added and purple powder was collected
by centrifugation. Targeted compound 1 was afforded after dried
in vacuum. Yield: 780 mg (5.2%). 1H NMR (500 MHz, DMSO):
chemical shift d: 9.11(d, J = 9.2 Hz, 8H), 8.87 (d, J = 6.0 Hz, 6H),
8.27(d, J = 7.8 Hz, 6H), 7.91 (d, J = 6.4 Hz, 8H), 7.51(d, J = 7.2 Hz,
8H), 4.62(s, 12H), ꢀ2.79 (s, 2H). ES-MS [EtOH-CHCl3, m/z]: (245
Fig. 1. Structures of a G-quartet showing the anti-parallel G-quadruplex topology
in the sodium solution (a) and Hoogstaeen base pairing (b) of human telomeric
sequences AG3(T2AG3)3. Arrows denoted 50 ? 30 strand alignment for DNA.
has been emerging as an effective modality for the selective treat-
ment of tumors and neoplastic lesions [16–18]. Although is widely
employed in clinic, the mechanism of PDT still needs to be fully
understood. The attempts to design and synthesize new photosen-
sitizer with higher photocytotoxicity never stopped [19–22].
Porphyrin derivatives were frequently reported as photosensitizing
agents which could generate cytotoxic species upon light exposure.
The majority of these reports have been focused on TMPyP4 and its
metal complexes. Researches on the photocytotoxicity of
TMPyP4-like porphyrins with larger extended planes were rela-
tively rare.
Herein we synthesized a series of new expanded cationic por-
phyrins (see molecular structures in Fig. 2) and comparatively
studied their interactions with G4 DNA. This is the first time to
design TMPyP-like porphyirns with larger planar on the sub-
stituents. Since researches on small molecules such as polypyridine
and anthraquinone always allow the conclusion that bigger aro-
matic planes relate to higher DNA binding affinities and cell
([M-4I]4+). UV–Vis (10
422(4.83), 515(3.39), 563(3.48), 588(3.10), 638(2.92).
lM in Tris buffer), kmax(nm) (log e):
5, 10, 15-tris (N-methyl-4-pyridyl)- 20- (phenyl-4-N-methyl-
4-pyridyl) porphyrin (2): A mixture of 120 mL propionic acid,
7 mL acetic anhydride and 0.55 g 4-(4-pyridinyl)benzaldehyde
(3 mmol) was heated at 130 °C with stirring. 0.92 mL
4-pyridinecarboxaldehyde (9 mmol) and 1.1 mL pyrrole (12 mmol)
diluted to 10 mL with propionic acid were then added separately
from the dropping funnels to the refluxing mixture. The resulting
mixture was refluxed for 1.5 h. The solvent was subsequently
evaporated under reduced pressure. The purifying and methylation
R1
NH
N
N
1. CH3CH2COOH
2. CH3I
R4
+
OH
R2
R
N
H
HN
R3
Porphyrin
R1
R2
R3
R4
N CH3
N CH3
CH3
N CH3
CH3
N CH3
CH3
N CH3
CH3
1
2
3
4
5
N
N
N
N
N
N
N
CH3
CH3
CH3
N CH3
N CH3
N CH3
N CH3
N CH3
1. 130 oC, reflux; 2. DMF, room temperature, overnight.
Fig. 2. The molecular structures of porphyrins 1–5.

112
S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
processes were similar to porphyrin 1. Yield: 430 mg (3.4%). 1H
NMR (500 MHz, DMSO): chemical shift d: 9.37 (d, J = 5.94 Hz,
2H), 9.12 (d, J = 6.21 Hz, 6H), 8.87(d, J = 5.51 Hz, 6H), 8.28(d,
J = 8.13 Hz, 6H), 7.92(d, J = 6.8 Hz, 6H), 7.5(d, J = 8.06 Hz, 6H),
4.61(s, 12H), ꢀ2.85(s, 2H). ES-MS [EtOH-CHCl3, m/z]: (189
of 20 mW at the samples. UV–Vis spectra were recorded on a
Hitachi U-3900H spectrophotometer. FRET spectra and FRET melt-
ing assay experiments were recorded on a Sanco 970CRT spec-
trofluorophotometer. CD spectra were recorded on a JASCO-J810
spectrometer.
([M-4I]4+). UV–Vis (10
lM in Tris buffer), kmax(nm) (log e):
420(4.72), 518(3.39), 552(3.48), 580(3.06), 643(2.91).
2.3.1. SERS experiment
5, 10, 15, 20-tetra (N-methyl-4-pyridyl) porphyrin (3, TMPyP4)
were similarly synthesized to porphyrin 1, with replacing
4-(4-pyridinyl)benzaldehyde by 4-pyridinecarboxaldehyde. Yield:
8.2%. 1H NMR (500 MHz, DMSO): chemical shift d: 9.45(d,
J = 6.0 Hz, 6H, 2, 6- pyridinium), 8.99–9.01(s, 8H, b-pyrrole),
8.27(d, J = 7.8 Hz, 6H, 3,5-pyridinium), 4.69(s, 12H, N+–Me),
ꢀ3.01(s, 2H, NH pyrrole). ES-MS [EtOH-CHCl3, m/z]: 169
Ag colloids were prepared by reducing AgNO3 with EDTA [26].
The SERS active systems were obtained by of mixing equal volume
of drugs (including AG22, porphyrins or a DNA/porphyrin ratio of
30:1) with Ag colloid, and the spectrum was immediately mea-
sured at room temperature.
2.3.2. Absorption, CD spectrum
([M-4I]4+). UV–Vis (10
lM
in Tris buffer), kmax(nm) (log
e):
Tris buffer was employed in the absorption and CD spectrum
experiments. For absorption spectra, the aliquot DNA prepared in
solution was added stepwise to the sample cell containing the por-
phyrins. After equilibration for 5 min, absorption or fluorescence
spectrum was recorded. The titration processes were repeated
until there was no change in the spectra for at least three titrations
indicating the binding saturation had been achieved.
As to the Scatchard equation, r/Cf = K(n – r), where r is the num-
ber of moles of porphyrin bound to 1 mol of G-quadruplex
(Cb/CDNA), n is the number of equivalent binding sites, and K is
the affinities of ligands for those sites [27]. The concentrations of
free porphyrin (Cf) and bound porphyrin (Cb) are calculated using
423(4.63), 517(3.39), 562(3.48), 585(3.00), 643(2.99).
5, 10, 15-tris(phenyl-4-N-methyl-4-pyridyl)-20-(4-biphenyl)p
orphyrin (4) was similarly synthesized to porphyrin 1, with replac-
ing
4-(4-pyridinyl)benzaldehyde
and
by
the
mixture
of
4-biphenylcarboxaldehyde
4-(4-pyridinyl)benzaldehyde.
Yield: 4.5%. 1H NMR (500 MHz, DMSO): chemical shift d: 9.11(d,
J = 9.2 Hz, 6H), 8.87 (d, J = 6.0 Hz, 6H), 8.27(d, J = 7.8 Hz, 6H), 7.91
(d, J = 6.4 Hz, 8H), 7.51(d, J = 7.2 Hz, 8H), 7.42(d, J = 6.4 Hz, 4H),
4.59(s, 9H), ꢀ2.79 (s, 2H). ES-MS [EtOH-CHCl3, m/z]: 323
([M-3I]3+). UV–Vis (10
lM in Tris buffer), kmax(nm) (log e):
425(5.09), 532(3.94), 560(3.40), 581(3.66), 648(3.18).
5, 15-di(4-biphenyl)- 10, 20-di(phenyl-4-N-methyl-4-pyridyl)
porphyrin (5) was similarly synthesized to porphyrin 4. Yield:
3.4%. 1H NMR (500 MHz, DMSO): chemical shift d: 9.09 (d,
J = 9.2 Hz, 4H), 8.84 (d, J = 6.0 Hz, 8H), 8.27 (d, J = 7.8 Hz, 4H),
7.91 (d, J = 6.4 Hz, 12H), 7.47(d, J = 7.2 Hz, 12H), 7.42(d, J = 6.4 Hz,
4H), 7.36(d, J = 7.2 Hz, 2H), 4.59 (s, 6H), ꢀ2.79 (s, 2H). ES–MS
Cf = C(1 – and Cb = C – Cf, respectively, where C is the total por-
a
)
phyrin concentration (5
lM). The fraction of bound porphyrin (
a)
was calculated using the equation,
a
= (Af – A)/(Af – Ab), where Af
and Ab are the absorbance of the free and fully bound porphyrin
at the Soret maximum of porphyrin, respectively, and A is the
absorbance at any given point during the titration. The percent
hypochromicity of the Soret band of porphyrin can be calculated
[EtOH-CHCl3, m/z]: 476 ([M-2I]2+). UV–Vis (10
lM in Tris buffer),
kmax(nm) (log
e): 428(4.63), 521(3.39), 569(3.18), 589(3.00),
using the equation
D
H = [(ef
–
e
b)/e
f] ꢁ 100%, where
eb = Ab/Cb.
644(3.00).
For CD and ICD spectra, porphyrins mixed with G4 DNA at a
ratio of [Drug]/[DNA] = 5 in the Tris buffer, and then the mixture
was incubated at 4 °C overnight. Each measurement was the aver-
age of three repeated scans recorded with a 0.1 mm quartz cell
with reaction volume of 1 mL at 25 °C. Final analysis of the data
was carried out using Origin 8.0 (Origin Lab Corp.).
The Tris buffer used in UV-Vis spectrum consists with 10 mM
Tris–HCl, 1 mM Na2EDTA, and 100 mM NaCl, pH 7.4. The 1H NMR
spectra of these porphyrins were given in Supplemental materials
as Fig. S1.
2.2. Materials
2.3.3. FRET experiment
The DNA oligonucleotides were purchased from the Shanghai
Sangon Biological Engineering Technology & Services Co., Ltd.
(China) in an HPLC-purified form. Nearest-neighbor approximation
method was employed to calculate the single-strand extinction
coefficients from mononucleotide data [24], using extinction coef-
Tris buffer was employed in the FRET experiment. Excitation
wavelength was scanned between 240 and 320 nm with selected
concentrations of G4 DNA, and fluorescence emission from the por-
phyrins (1
lM) was observed at 660 nm. Relative fluorescence
intensity (F) was calculated from the equation:
ficients at 260 nm of 228,500 Mꢀ1 cmꢀ1
.
ꢀ
ꢁ ꢀ
ꢁ
Ik
I320
I320
Ik
The formation of intra- and inter-molecular G-quadruplexes
was carried out as follows: the oligonucleotide sample
50-AG3T2AG3T2AG3T2AG3-30 (AG22) and fluorescent-labeled
oligonucleotide sample 50-FAM-G3T2AG3T2AG3T2AG3 -TAMRA-30
(F21T, FAM: 6-carboxyfluorescein, TAMRA: 6-carboxytetramethylr
hodamine), dissolved in the Tris buffer solution mentioned in the
UV-Vis spetrum for AG22 or a PBS buffer (consisting of 10 mM
K2HPO4/KH2PO4, 40 mM KCl, 100 mM NaCl, pH 7.4) for F21T, was
heated to 90 °C for 5 min, gently cooled to room temperature,
and then incubated at 4 °C overnight. The formation of G4-DNA
was affirmed by the appearance of positive peak near 290 nm
and negative trough near 265 nm in CD spectra, which is character-
istic of antiparallel-stranded G-quadruplexes [25].
F ¼
b
f
where If and Ib are the fluorescence intensity of free and bound por-
phyrins, respectively. Fluorescence intensity excited at 320 nm was
used for normalization because of the very low absorption of
nucleic acids at this wavelength. Inner-filter effect was neglected
since fluorescence intensity was proportional to porphyrin concen-
trations below 2 lM in our condition.
2.3.4. FRET melting assay experiment
FRET melting assay was performed according to previous proto-
cols [12]. FRET melting assay was conducted on Perkin–Elmer Ls55
spectrofluorophotometer with a temperature controller. The fluo-
rescence dual labeled G-quadruplex DNA sample F21T was pre-
pared according to the method described above and diluted to
2.3. Instruments and Methods
Surface enhanced raman spectra (SERS) was carried out on an
inVia Laser Micro-Raman Spectrometer of Renishaw, with a power
0.2
lM with PBS buffer, and it was further incubated for 1 h with
different concentration compounds added. Excitation wavelength

S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
113
was 483 nm, and both the excitation and emission slit widths were
5 nm. The emission spectrum was recorded from 500 to 700 nm.
The fluorescence intensity at 533 nm was recorded by 28–38 inter-
vals over the range of 20–95 °C, and at each temperature incuba-
tion was continued for 5 min before reading the values. Final
analysis of the data was carried out with Origin 8.0 software
(OriginLab Corp.).
2.3.5. Theoretical detail
The full geometry optimization and further frequency analysis
of all studied porphyrins 1–5 were carried out with the
DFT-B3LYP method and 6-31G basis set, and single point calcula-
tion was further performed based on the level of B3LYP/6-31G⁄.
In order to vividly depict the detail of the frontier molecular orbi-
tal, the stereographs of some related frontier molecular orbits of
these complexes were drawn with the Molden v 4.2 program based
on the calculated results. All calculations were performed with the
G03 (d01) program-package.
Fig. 3. SERS spectra (300–1800 cmꢀ1) of porphyrins in the absence and presence of
G4. (—), (
) for free porphyrin 1 and G4, respectively; (
), (
) (
), (
)
2.3.6. Photocytotoxicity assays in vitro
(
), for G4 complexes with 1–5, respectively. [AG22]/[Por] = 30:1.
The capacities of porphyrins to interfere with the growth of
myeloma cells (Ag8.653) and gliomas cells (U251) were deter-
mined by WST assay. Compounds were dissolved in DMSO and
diluted with Sera Free Cryopreservation Media (RPMI) 1640 to
the required concentrations prior to use. The control well was pre-
Table 1
Raman frequencies and assignments for porphyrinsa.
Raman shift (cmꢀ1
)
Assignmentsb(cmꢀ1
)
pared by addition of culture medium (100 lL). Wells containing
330
411
810
1004
1191
1214
d(por)
d(por) +
1242
1337
1366
1443
1554
1635
m (C – pyr)
ms (N – C)
ms (C – Cb)
ms (C – Cb)
culture medium without cells were used as blanks. Myeloma cells
(Ag8.653) and gliomas cells (U251) with a density 2 ꢁ 104 cells per
well were precultured into 96-well microtiter plates for 48 h at
37 °C with 5% CO2. Upon completion of the incubation, stock
WST-1 solution was added to each well. After 4 h incubation in
dark, visible or UV-A light, the cell viability was determined by
m
(Ag-N)
m
m
(N+–CH3), ds(por)
(benz)
a
a
d(pyr),
d(pyr),
m
m
(N+–CH3)
(C – N)
m
(Cb – Cb)
ms (C – Cb)
a
a
Band assignments are according to Refs. [26–35].
Abbreviations: m, stretching mode; d, bending mode; s, symmetric mode; pyr,
b
measuring the absorbance of each well at 450 nm using
a
N-methylpyridinium; por, porphyrin core.
Multiskan SSCENT microplate reader. IC50 values were determined
by plotting the percentage viability versus concentration on a log-
arithmic graph and reading off the concentration at which 50% of
cells remain viable relative to the control.
is interpreted as a loss of accessibility for ligands to the Ag colloids.
The slightly decreased SERS signals suggest that there are still large
amount of porphyrins exposed in the solvent which are still detect-
able by SERS. The porphyrins 4 and 5 could not be fully protected
by the G4 DNA molecular structure and are supposed to externally
bind on the surface of G4 DNA. In contrast, the porphyrins 1, 2 and
3 present intercalating or end-stacking mode in G-quadruplexes,
because the molecules are buried in the bases or loops of
G-tetrads and thus become undetectable by SERS [28,31].
Moreover, it should be pointed out that the chosen protocol for
the SERS sample preparation allows the G4-porphyrin adduct to be
formed before its addition to the colloid solution (see
Experimental). Consequently, the spectral changes confirm that
the porphyrins efficiently bind to G4 DNA even in Ag sols. In this
medium, the preference of cationic porphyrins for polyanionic
G4-DNA prevails over negatively charged Ag nanoparticles
[32,33]. This result confirms the tight binding affinities between
the cationic porphyrins and these DNA structures.
3. Results and discussion
3.1. SERS investigation
SERS is
a
powerful and sensitive tool to study the
drug-biomacromolecules interactions at very low concentrations.
No substantial SERS signal is given by AG22, since the Ag colloids
have negative charges on the surface, which repels the adsorption
of negatively-charged G4 molecules (Fig. 3). In the absence of
G4-DNA, porphyrin’s characteristic SERS signals were clearly
observed, which were assigned in Table 1 [28–30].
However, in the presence of G4-DNA, most bands of the
porphyrin/G4 complexes decreased, indicating that the interac-
tions of porphyrins with AG22 can largely reduce the amount of
porphyrins adsorbed on Ag colloids [29]. Especially, the quenching
of the band at 1554 cmꢀ1, which is mainly assigned to the stretch-
ing and is the marker band of porphyrin ring, doubtlessly proves
the large decrease of porphyrin molecules surrounding Ag colloids
after the interaction between porphyrins and G4 DNA. The weaken
bands centered at 1001, 1098, 1242, 1635 nm also indicate the por-
phyrins are protected by G4 DNA from adsorbing on Ag particles
[12,28–30].
3.2. CD and ICD spectra
The CD spectral method, which is a very useful method in dis-
tinguishing the DNA secondary structures, was employed to study
the interaction of these porphyrins with G4 DNA. From Fig. 4, in the
presence of Na+, the CD spectrum of AG22 shows a strong positive
band at around 290 nm and a negative band at around 265 nm,
which indicates the formation of an anti-parallel G-quadruplex
structure [25]. In the presence of the porphyrins 1–4, the intensi-
ties of both positive and negative bands for G4-DNA increased
It is notable that the SERS of porphyrins 4 and 5 in the presence
of G4 DNA told us a quite different story from the other three por-
phyrins. Their SERS signals, although the intensities were
decreased, still remained. In terms of the short-range character of
the Raman enhancement in a colloid system, the SERS quenching

114
S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
Fig. 4. a: CD spectra of AG22 in the absence (
) and in the presence of 1 (
), 2 (
), 3 (
), 4 (
), 5 (
) in Tris buffer. [Drug]/[DNA] = 5. b: ICD spectra of 10 lM
porphyrins 1 (
), 2 (
), 3 (
), 4 (
), 5 ( ) in the presence of AG22 in Tris buffer. [DNA]/[Drug] = 5.
Table 2
Physical data of absorption titration for porphyrins binding with AG22.
significantly (Fig. 4a), which suggests that an anti-parallel struc-
ture is enhanced and stabilized by compounds [34]. However,
excepted for slight red shifts in the positive peak and negative
bands, no substantial changes were observed for G4-DNA in the
presence of porphyrins 5, implying that the porphyrin 5 affected
the structure of G4-DNA negligibly. The distinct CD spectral change
indicates that the porphyrins 1–4 may employ a more efficient
binding with the porphyrins than the porphyrin 5. This result is
in high consistent with the experiments above.
Porphyrin
UV titation
k (nm)
Ka
DT/°C
D
H%
1
2
3
4
5
7
32.5
49.8
44.6
30.6
15.4
9.2 ꢁ 105
8.2 ꢁ 106
5.6 ꢁ 106
1.5 ꢁ 105
3.5 ꢁ 103
8.9
12.4
9.2
4.3
2.1
13
10
5
1
ICD spectral method has also been employed to probe the inter-
action modes between quadruplexes and bound ligands.
Intercalation into guanine quadruplexes is characterized by a
strong negative ICD at the Soret band of porphyrin, while
end-stacking and groove binding generate a bisignate and a weak
positive ICD peak, respectively [34–35]. As depicted in Fig. 4b, free
base porphyrin 1 shows negligible ICD signals in the absence of
AG22 and similar phenomenon were observed for all the por-
phyrins. In the presence of G4-DNA, negative peaks were observed
for porphyrins 1, 2 and 3, suggesting that the free base porphyrin
should intercalate between G-tetrads when fully bound. As to the
tricationic porphyrin 4, bisignate ICD signal appeared. This sug-
gests a different binding mode of porphyrins 1-3, which is consis-
tent with end-stacking binding mode. On the other hand,
dicationic porphyrin 5, showed relatively silent CD signal in the
presence of G4-DNA, suggesting that the interaction between
G4-DNA and porphyrin 5 is comparatively weak and the porphyrin
may only externally bind on the surface of the DNA through elec-
trostatic strength. The result of ICD experiment unambiguously
proves the binding modes we proposed in the experiments above.
spectra of 4 and 5 proved that positive charges are important fac-
tors in binding with G4 DNA.
The Scatchard analysis can determine the numbers of equiva-
lent binding sites and affinities of ligands to binding sites, and thus
is a useful way to treat the binding data quantitatively. Fig. S3 casts
the data from each absorbance titration into Scatchard plots, and it
was found that the Scatchard plots produced from the UV–vis
absorption titration are nonlinear. Deviations from linearity may
arise from the unequal association constants of porphyrins for
the different binding sites in G-quadruplexes. Data points at lower
r values between 0.83 and 1.8 for 1, 0.75 and 1.61 for 2, 0.68 and
1.52 for 3, 0.41 and 1.43 for 4, 0.31 and 1.3 for 5, respectively, were
fit to the Scatchard model. The fit gave approximate Ka and binding
numbers n values for these porphyrins in terms of quadruplexes,
which were shown in Table 2.
The determined Ka for porphyrins binding with AG22 follow an
order as 2 > 3 > 1 ꢂ 4 ꢂ 5. Tetracationic porphyrins 1, 2 and 3 have
largest Ka while dicationic porphyrin 5 has the smallest Ka. Positive
charge plays a much important role in the ligand-DNA binding
behaviors, which consists with the conclusion of SERS.
Meanwhile, asymmetrical porphyrin 2 exhibit higher binding
affinities than the other porphyrins with same positive charges,
suggesting that the specific molecular structures of porphyrins also
affect the binding mode between porphyrins and G4 DNA.
3.3. Absorption titrations
Absorption titration experiment was employed to examine the
association of the porphyrins with AG22 and the absorption spec-
tral changes of 4 upon AG22 addition were given in Fig. S2 in sup-
plemental materials. Red shift and hypochromicity was observed
in Fig. S2 and similar results were observed for other porphyrins.
Table 2 summarizes the detailed titration data of all these por-
3.4. FRET spectroscopy
The porphyrins’ fluorescence emission is negligible in the wave-
length region between 240 nm and 320 nm. It is widely accepted
that once porphyrins closely contact with nucleobases, fluores-
cence in this region will be enhanced as a consequence of energy
transfer from DNA bases to porphyrins. In general, FRET has been
assumed to be observed specifically for a ligand intercalated
between DNA basepairs [35,36]. Recently, it was also reported that
FRET occurs when porphyrins bound externally to DNA in the DNA
phyrins at their absorption bands. Large red shift
hypochromicity H were induced by the G4 DNA binding of por-
phyrins 1, 2 and 3 in the Soret band, while the absorption spectral
change for porphyrin 4 were much smaller. The smallest k and
H were observed for titrations of AG22 to porphyrin 5. Since por-
Dk and
D
D
D
phyrin 4 is tricationic and 5 is dicationic, which are less positive
than the porphyrins 1, 2 and 3, the smaller change in absorption

S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
115
Fig. 6. FRET Melting curves of F21T in the absence (j) and presence of porphyrins
1( ), 2( ), 3( ), 4( ), 5 ( ) in PBS buffer.
Fig. 5. FRET spectra of 10 lM porphyrins 1( ), 2(d), 3( ), 4( ), 5 ( ) in the
presence of AG22 in Tris buffer at a molar ratio of [AG22]/[Por] = 10:1.
As revealed by the increase of Tm, the quadruplex structure was
groove [37–39]. Accordingly, the observation of FRET is indicative
of the close contact of ligands with DNA bases. In addition, spectral
profile of FRET should change depending on the type of a base,
which closely interacts with ligands. Generally speaking, the inter-
calation will give much higher fluorescence energy transfer effi-
ciency than the end-stacking binding mode [39].
Fig. 5 depicts the FRET spectra of the G4 DNA/porphyrin com-
plexes. For porphyrins 1, 2 and 3, significant increases were
observed in the relative quantum yield region from 240 nm to
320 nm, demonstrating the presence of strong energy transfer
from DNA bases to porphyrin. Meanwhile, the maxima FRET spec-
tra for porphyrins 1, 2 and 3 concentrated in the region of
280-290 nm. Thus, it is suggested that these porphyrins should
commonly locate close to guanine and/or thymine bases, because
the FRET spectra of both guanine and thymine showed the bands
around 280 nm [37–39]. Moreover, the increase in quantum yield
in case of the porphyrin 2 are much higher than that for the por-
phyrins 1 and 3, indicating that binding of porphyrins 2 to G4
DNA results in substantially greater energy transfer than those of
the porphyrins 1 and 3.
stabilized by these porphyrins. The values of
DTm in the presence of
these porphyrins are relatively large, 2.1–12.4 °C, indicating that
the porphyrins contribute differently in stabilizing the quadruplex
structure. Noteworthy is the fact that the tetracationic porphyrins
1, 2 and 3 were much more effective in the stabilization of the
quadruplex structure than the tricationic and dicationic porphyrins
4 and 5, which suggests that higher positive charges play a key role
in porphyrin-DNA binding.
Moreover, porphyrin 2, which replaces only one of the
N-methylpyridine group of TMPyP4 with more extended sub-
stituent, resulted in greater Tm increase than that of TMPyP4 (por-
phyrin 3). However, at the extreme, when all four groups were
replaced by large substituents, the melting temperature increased
a lower extent than TMPyP4. The crystal structure of G4 DNA scaf-
fold showed that the grooves of the intramolecular DNA quadru-
plex are not all the same sizes: a very narrow groove lies
opposite a wide groove, with two intermediate-width grooves
between. The results above can be rationalized in terms of the lar-
ger N-methylbenzonpyridine-4-yl group better fitting into the
wide groove of the quadruplex and therefore allowing more effec-
tive binding of the porphyrin to its target.
No energy transfer from DNA bases to porphyrins 4 and 5 for
this complex is evident since the relative quantum yields are lower
than 1 (Fig. 5). It is widely accepted that molecules bound on the
surface of DNA due to electrostatic interaction are not situated cor-
rectly for efficient energy transfer through the bases [37–39,40].
Thus, the porphyrins 4 and 5 are evidenced to bind on the surface
of G4 due to electrostatic interaction.
3.6. Energy transfer explanation from DFT calculations
According to the frontier molecular orbital theory, the higher
occupied molecular orbital energy of one donor molecule and the
lower unoccupied molecular orbital energy of the acceptor are
advantageous to the interaction between the two molecules,
because electrons more easily transferred from the occupied MOs
of donor to the unoccupied MOs of acceptor [42–45]. The energies
of the HOMO and HOMOꢀx (x = 1–6) for the DNA structures calcu-
lated by the authors were ꢀ1.27, ꢀ1.33, ꢀ1.69, ꢀ1.79, ꢀ1.98,
ꢀ2.06 and ꢀ2.08 eV, respectively. Fig. 7 displays some frontier
molecular orbital energy levels of the title compounds calculated
at the level of B3LYP/6-31G⁄. It is easy to find that the energies
of the LUMO and LUMO + x (x = 1–4) of these porphyrins are from
ꢀ10.19 eV to ꢀ5.55 eV, which are much lower than those of the
HOMO and HOMOꢀx (x = 1–6) of the DNA base-pairs model.
Therefore, the five porphyrins can interact with DNA strongly
and should be good electron acceptors when binding to DNA.
Moreover, from Fig. 7, it is effortless for one to conclude that, lower
energies of LUMO and LUMO + x (x = 1–3) always relates to more
positive charges. Tetracationic porphryins 1, 2 and 3 have LUMO
3.5. FRET Melting assay studies
FRET Melting assay is a convenient spectroscopic method that
provides significant information on macromolecules in solution,
and it is particularly suited to evaluate the structural changes in
proteins and nucleic acids [37,41]. Fig. 6 shows FRET melting
curves of G-quadruplexes in the presence of different concentra-
tions of compounds, and the data are compiled in Table 2.
It has been reported that the telomerase inhibition of drugs was
strongly related to the stabilization of the quadruplex structure
[12,14,37]. Under the present experimental conditions, the melting
curve has
a transition, and the Tm value of free F21T is
(61.9 0.2) °C; when mixed with the porphyrins, the observed
melting temperatures of G4 DNA increase to different extents
and the increases of Tm
(DTm) are 8.9 °C, 12.4 °C, 9.2 °C, 4.3 °C
and 2.1 °C for porphyrins 1, 2, 3, 4 and 5, respectively.

116
S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
the most important roles in G4-DNA binding affinities and photo-
cytotoxicity, and (2) substitution of extended aromatic plane is
only partially tolerated on the meso positions of the porphyrin ring,
and the bulk of the substituents should be matched to the width of
the grooves in which they putatively lie. This SAR rule may facili-
tate the research on porphyrin drugs targeting at human telomeric
quadruplex DNA.
Acknowledgment
This work was financially supported by the National Nature
Science Foundation of China (21101033, U1204209) and the
Outstanding Young Teachers’ Fund in Colleges and Universities of
Guangdong Province.
Appendix A. Supplementary material
Fig. 7. Frontier unoccupied orbital of LUMO and LUMO + x (x = 1–3) energies (in eV)
of porphyrins 1–5 calculated at the level of B3LYP/6-31G⁄.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bioorg.2015.05.
001.
Table 3
The IC50 values for porphyrins against Ag8.653 and U251 in vitro.
References
Porphyrin
Ag8.653 (lM)
U251 (lM)
[1] G. Saretzki, Curr. Pharm. Des. 20 (2014) 6386–6403.
[2] J.C. Eissenberg, Missouri Med. 110 (2013) 11–16.
[3] M. Nakashima, J. Nandakumar, K.D. Sullivan, J.M. Espinosa, C.R. Thomas, J. Biol.
Chem. 288 (2013) 33171–33180.
[4] P.G. Smiraldo, J. Tang, J.W. Shay, W.E. Wright, Telomerases 11 (2012) 243–263.
[5] A. Apurva, D. Shweta, G. Reema, Recent Pat. Anti-Canc. 7 (2012) 102–117.
[6] L. Laureline, P. Enora, R. Aurore, N. Marie-Laure, B. Romaric, H. Angeline van
Der, P. Christian, G. Dennis, M. Jean-Louis, J. Helene, Inorg. Chem. 53 (2014)
12519–12531.
Dark
dark Visible
UV-A
Dark
Visible
UV-A
1
2
3
4
5
>100
>100
>100
>100
>100
7.52
6.33
0.57
12.33
15.45
9.91
6.74
1.92
17.41
25.45
>100
>100
>100
>100
>100
10.43
8.38
0.64
18.27
14.37
10.15
7.89
1.77
22.63
28.24
[7] A.B. Chinen, C. Guan, M.A. Chad, Angew. Chem. Int. Ed. 54 (2015) 527–531.
[8] M.J. Morris, K.L. Wingate, J. Silwal, T.C. Leeper, S. Basu, Nucl. Acids Res. 40
(2012) 4137–4145.
[9] S. Cogoi, E.X. Luigi, Chem. Commun. 46 (2010) 7364–7366.
[10] S. Claudia, P. Manlio, Curr. Pharm. Des. 20 (2014) 6489–6509.
[11] W. Derrick, R. Nihar, S. Kumar, C.J. Gong, D.P. Arya, Bioorg. Med. Chem. Lett. 23
(2013) 6695–6699.
energies than the tri- or dicationic porphyrions 4 and 5, and thus
have much higher binding affinities to DNA molecules.
3.7. Photocytotoxicity assays
[12] L.H. Wang, C.Y. Wei, Chem. Biodivers. 10 (2013) 1154–1164.
[13] G.J. Lin, G.B. Jiang, Y.Y. Xie, H.L. Huang, Z.H. Liang, Y.J. Liu, J. Biol. Inorg. Chem.
18 (2013) 873–882.
[14] P. Zhao, J.Z. Lu, J. He, W.H. Chen, P.P. Chen, D.W. Chen, Q.Y. Bin, Nucleos.
Nucleot. Nucl. 33 (2014) 597–614.
[15] P. Zhao, J. Li, L.J. Yang, J.Z. Lu, H.M. Guo, L.N. Ma, B.H. Ou, J. Coord. Chem. 66
(2013) 4220–4236.
[16] D. Kumar, B.A. Mishra, K.P.C. Shekar, A. Kumar, K. Akamatsu, R. Kuriharab, T.
Ito, Org. Biomol. Chem. 11 (2013) 6675–6679.
[17] D. Kumar, K.P.C. Shekar, B. Mishra, R. Kurihara, M. Ogura, T. Ito, Bioorg. Med.
Chem. Lett. 23 (2013) 3221–3224.
[18] P. Mroz, A. Yaroslavsky, G.B. Kharkwal, M.R. Hamblin, Cancers 3 (2011) 2516–
2539.
[19] S.M. Abdelghany, D. Schmid, J. Deacon, J. Jaworski, F. Fay, K.M. McLaughlin,
Biomacromolecules 14 (2013) 302–310.
[20] J.T.F. Lau, P.C. Lo, W.P. Fong, D.K.P. Ng, J. Med. Chem. 55 (2012) 5446–5454.
[21] J. Berlanda, T. Kiesslich, V. Engelhardt, B. Krammer, K. Plaetzer, J. Photochem.
Photobiol. B 100 (2010) 173–180.
[22] R. Ruiz-Gonzalez, P. Acedo, D. Sanchez-Garcia, S. Nonell, M. Cañ ete, J.C.
Stockert, Eur. J. Med. Chem. 63 (2013) 401–414.
[23] A.D. Adler, F.R. Longo, J.D. Finarelli, J. Assour, L. Korsakoff, J. Org. Chem. 32
(1967) 476–479.
[24] C.R. Cantor, M.W. Warshaw, H. Shapiro, Biopolymers 9 (1970) 1059–1077.
[25] P. Zhao, S.F. Jin, J.Z. Lu, J.L. Lv, G.Q. Wu, P.P. Chen, C.L. Tan, D.W. Chen,
Spectrochim. Acta Part A 137 (2015) 227–235.
[26] P. Simakova, M. Prochazka, E. Kocisova, Adv. Biomed. Spectrosc. 7 (2013) 171–
175.
The antitumor activities of these porphyrins against Ag8.653
and U251 cell lines were evaluated by WST assay under dark, vis-
ible or UV-A conditions. The IC50 values obtained for the por-
phyrins against two tumor cell lines are shown in Table 3. Under
dark, the synthetic expanded porphyrins show negligible inhibi-
tion to cancer cells. However, all these porphyrins exhibit broad
inhibition on the two tested human cancer cell lines in visible
and UV-A light with relatively low IC50 values. This enhanced pho-
tocytotoxicity relative to the dark control is an essential property
of a photochemotherapeutic agent [16–22].
Meanwhile, it is notable that tetracationic porphryins 1, 2 and 3
have much lower IC50 than the tri- or dicationic porphyrions 4 and
5, indicating that positive charges are very significant in
anti-tumor activities. However, it is unexpected to find that por-
phyrin 2, whose DNA binding affinity is higher than that of por-
phyrin 3 (TMPyP4), possess a relatively lower inhibitory effect
against the two cell lines compared with TMPyP4. Moreover, As
to the IC50 values of porphyrins 1 and 2, the results under visible
light were comparable with UV-A light, which is irregular compar-
ing with other porphyrins and the previous reports [16,17,19].
More efforts about the mechanism to understand these should be
paid.
[27] J.B. Chaires, Meth. Enzymol. 340 (2001) 3–22.
[28] C.Y. Wei, G.Q. Jia, J.L. Yuan, Z.C. Feng, C. Li, Biochemistry 45 (2006) 6681–6691.
[29] V. Heleg-Shabtai, A. Zifman, S. Kendler, Adv. Exp. Med. Biol. 733 (2012) 53–61.
[30] S. Petra, P. Marek, J. Mol. Struct. 993 (2011) 425–427.
[31] S. Petra, G. Juliette, P. Marek, H.A. Katel, I. Chourpa, J. Phys. Chem. C 118 (2014)
7690–7697.
4. Conclusion
[32] S. Murphy, L.B. Huang, P.V. Kamat, J. Phys. Chem. C 117 (2013) 4740–4747.
[33] P. Simakova, M. Prochazka, E. Kocisova, Spectroscopy 27 (2012) 449–453.
[34] Y.H. Kim, C.Y. Lee, S.K. Kim, S.C. Jeoung, Biophys. Chem. 190–191 (2014) 17–24.
[35] M. Jamlet, K. Evgeni, G. Maya, K. David, K. Maya, L. Eteri, M. Hakob, H. Nune,
Am. J. Anal. Chem. 3 (2012) 698–703.
Herein a serious of expanded cationic porphyrins was designed
and their G4-binding abilities were comparatively studied with
TMPyP4, from which rules for a structure–activity relationship
(SAR) have been discerned: (1) positively charged substituents play
[36] G. Srijana, F.E. Phillip, M.R. David, Inorg. Chem. 53 (2014) 11108–11118.

S.-f. Jin et al. / Bioorganic Chemistry 60 (2015) 110–117
117
[37] R.M. Jenifer, D. Meo, F. Lo, M. Delcourt, C. Clement, S. Norman, P. Richeter, S.
Linares, M. Surin, Org. Biomol. Chem. 13 (2015) 2453–2463.
[38] L. Aurelien, D. Nicolas, S. Loic, R. Philippe, G.P. Claude, M. David,
ChemMedChem 9 (2014) 2035–2039.
[39] Q.L. Yue, T.F. Shen, C.N. Wang, L. Wang, H.B. Li, S.L. Xu, H.S. Wang, J.F. Liu,
Biosens. Bioelectron. 40 (2013) 75–81.
[41] L.H. Wang, Y. Wen, J. Liu, J. Zhou, C. Li, C.Y. Wei, Bioorg. Med. Chem. 21 (2013)
3379–3387.
[42] A. Kumar, M.D. Sevilla, J. Phys. Chem. B 118 (2014) 5453–5458.
[43] G.V. Palamarchuk, O.V. Shishkin, L. Jerzy, J. Phys. Chem. B 117 (2013) 2841–
2849.
[44] P. Venkata, A. Kumar, M.D. Sevilla, J. Phys. Chem. B 115 (2011) 15090–15097.
[45] J. X Wu, Y. Fu, Z.H. He, Y. Han, L. Zheng, J.L. Zhang, W. Li, J. Phys. Chem. B 116
(2012) 1655–1665.
[40] P. Zhao, J. Z Lu, F.Y. Hong, B.H. Ou, F.D. Zhang, L. N Ma, H.M. Guo, Spectrochim.
Acta Part A 108 (2013) 1–7.