Journal of Inorganic Biochemistry 137 (2014) 12–21

Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry

j ourna l homepage: www.e lsev ie r .com/ locate / j inorgb io

A platinum(II) complex of liriodenine from traditional Chinese medicine(TCM): Cell cycle arrest, cell apoptosis induction and telomeraseinhibition activity via G-quadruplex DNA stabilization

Yu-Lan Li 1, Qi-Pin Qin 1, Yan-Cheng Liu ⁎, Zhen-Feng Chen, Hong LiangState Key Laboratory Cultivation Base for the Chemistry andMolecular Engineering ofMedicinal Resources, School of Chemistry & Pharmacy of Guangxi Normal University, Guilin 541004, PR China

⁎ Corresponding author. Tel./fax: +86 773 2120958.E-mail address: ycliugxnu@aliyun.com (Y.-C. Liu).

1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.jinorgbio.2014.04.0010162-0134/© 2014 Elsevier Inc. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 November 2013Received in revised form 1 April 2014Accepted 2 April 2014Available online 13 April 2014

Keywords:LiriodeninePlatinum complexCell apoptosisG-quadruplex DNATelomerase

Liriodenine (L), an antitumor active ingredient from the traditional Chinese medicine (TCM), Zanthoxylumnitidum, afforded a platinum(II) complex (1) of L, cis-[PtCl2(L)(DMSO)], which previously reported for itsin vitro antitumor activity and intercalative binding with DNA. In this study, complex 1 was further discussedfor its antitumor mechanism and structure–activity relationship, comparing with L and cisplatin. Towards themost sensitive BEL-7404 human hepatoma cells, complex 1 significantly induced cell cycle arrest at both G2/Mphase and S phase. It suggests that double helix DNA is not the simplex intracellular target for 1. On the otherhand, the BEL-7404 cells incubatedwith 1 and stained byHoechst 33258 andAO/EB showed typical cell apoptosisin dose-dependent manner. The BEL-7404 cells incubated with 1 and stained by JC-1 were also characteristic forcell apoptosis on the loss of mitochondrial membrane potential. Furthermore, the G-quadruplex DNA bindingproperty of complex 1 was also investigated by spectroscopic analyses, fluorescent indicator displacement(FID) assay and fluorescence resonance energy transfer (FRET) assay. The results indicated that 1 stabilized thehuman telomeric G4-HTG21 DNA better than L. The telomerase inhibition ratio of 1 ((62.50 ± 0.03)%), whichwas examined by telomerase polymerase chain reaction-enzyme-linked immunosorbent assay (PCR-ELISA),was much higher than L ((21.77 ± 0.01)%). It can be ascribed to the better G4-HTG21 DNA stabilization of1 than L. The results suggested that the nuclei, mitochondria and telomerase via G-quadruplex DNA stabilizationall should be key targets for the antitumor mechanism of 1, in which the central platinum(II) played a key role.

© 2014 Elsevier Inc. All rights reserved.

1. Introduction

DNA is regarded as the primary molecular target in tumor cells formost of the antitumor drugs or agents. The design and developmentof new antitumor agents targeting specific DNA secondary structuresor sequences remain to be a challenge [1]. Guanine-rich (G-rich)stretches of DNA have a high propensity to self-associate into planarguanine quartets (G-quartets), which will assemble to give unusualstructures called G-quadruplex. G-quadruplex exists inmany importantregulatory regions in human genome, including telomeric ends, im-munoglobulin switch regions, mutational hot spots, and regulatoryelements in oncogene promoters [2]. Since the specific structure ofG-quadruplex DNA can be potentially targeted by small molecules,increasing number of researches have focused on G-quadruplexDNA.

Telomere is an important section of chromosome, which guaranteesproper replication and protection of chromosome ends [3,4]. One ofthe most abundant sources of DNA sequences capable of forming

G-quadruplex structures can be found in telomeres, called telomericDNA. Telomeric DNA is composed of a repeated double-stranded[TTAGGG/CCCTAA]n sequence except in the 3′-terminal region,which consists of a single-stranded [TTAGGG] repeated sequenceover several hundred bases [5,6]. The rich [TTAGGG] sequencesform planar G-tetrads by hydrogen-bonding interactions betweenthe Watson–Crick edge and the Hoogsteen edge, which can be furtherstabilized by K+/Na+ and consequently form G-quadruplex DNA [7,8].

Moreover, telomerase has been found to be over expressed in 85–90% of the human tumor cells, while its activity is undetectable inmost of the normal somatic cells [9,10]. So the telomerase as an attrac-tive and specific target for developing targeting antitumor drugs oragents has raised many researchers' interests [11–13]. Furthermore,many studies have also shown that stabilization of G-quadruplexes orligand-induced quadruplex formation can inhibit the activity of telome-rase [14–16]. Therefore, G-quadruplex DNA has also become an impor-tant molecular target for developing new anticancer agents withtargeting effects [17].

On the other hand, new strategies for developing the metal-basedanticancer agents have focused on the selection or design of functionalligands in recent years [18–20], in which the selection of traditionalChinese medicine active ingredients (TCMAI) as bioactive ligands is

13Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

appreciated [21,22]. Liriodenine is a representative oxoaporphinealkaloid mainly occurring in the families of Rutaceae, Annonaceae,Magnoliaceae, etc. [23–25]. Liriodenine has shown awide range of phar-macological activities, including the antitumor, antibacterial, antifungaland antileishmanial activities, in which its antitumor activity wasmostly considered [26–28]. It is generally accepted that its aromaticplanar structure as oxoaporphine mainly contributes to this significantantitumor activity, as shown in Scheme 1. The chemical structure ofliriodenine bears the 7-N and 8-O chelating sites. However, since theliriodenine contents in most of the plants worldwide are very low,its coordination chemistry has not yet been developed except in ourprevious work [29,30].

In our previouswork, double helix DNA and topoisomerase had beenproven to be key intracellular targets for two synthesized platinum(II)complexes of liriodenine. As our continuous study, one platinum(II)complex (1) of liriodenine with well characterized structure (shownin Scheme 1) was selected for further study. And considering thatthe BEL-7404 human hepatoma cell line ranks among the most sen-sitive tumor cell lines towards complex 1 as previously reported(IC50 = 6.908 ± 1.333 μM), it was selected in this study to investigatethe detailed antitumor mechanism of complex 1. Furthermore, consid-ering the well retained planar structure of complex 1 in solution, weassume that the G-quadruplex DNA is another key molecular targetrather than the double helix DNA for complex 1, and thus the potentialtelomerase inhibition ability of 1 may be also expected consequently.All these studies were carried out by means of FID assay, fluorescencetitration analyses, circular dichroism (CD) spectral analyses, fluores-cence resonance energy transfer (FRET) assay and telomerase polymer-ase chain reaction-enzyme-linked immunosorbent assay (PCR-ELISA).Liriodenine as the bioactive ligand and cisplatin as the classic platinum(II)complex were also studied for comparison.

2. Materials and instruments

2.1. Materials

All chemical reagents were commercially available and receivedwithout further purification, unless noted specifically. The isolationand characterization of liriodenine (L) from TCM, as well as the synthe-sis of the title platinum(II) complex, cis-[PtCl2(L)(DMSO)] (complex 1),were according to our previous report [29].

Thiazole orange (TO), Tris, acridine orange (AO), and ethidium bro-mide (EB) were purchased from Sigma-Aldrich (USA). c-kit-1 (5′-CG3-

CG3CACGAG3AG3T-3′), c-kit-2 (5′-CG3CG3CGCTAG3AG3T-3′), HTG21(5′-G3TTAG3TTAG3TTAG3-3′), Pu22 (5′-TGAG3TG3TAG3TG3TAA-3′),RNAse A and propidium iodide (PI) were purchased from Sangon(Shanghai, China). Bisbenzimide (Hoechst 33258) and JC-1 mitochon-drial membrane potential detection assay kit was purchased fromBeyotime Institute of Biotechnology (China). Telomerase PCR-ELISA kitwas purchased from Roche (USA).

The formation of all intramolecular G-quadruplexes was analyzed asfollows: the oligonucleotide samples were dissolved in Tris–KCl–HClbuffer and heated to 90 °C for 5 min. Then they were gently cooled to

Scheme 1. Chemical structure of liriodenine (L, left) and its platinum(II) complex(1, right).

room temperature and incubated at 4 °C overnight. The human telome-rase was temporarily prepared in Tris buffer solution. All the spectro-scopic experiments were performed at 25 °C.

2.2. Instrumentation and methods

Fluorescence spectrawere obtained on ShimadzuRF-5301PC spectro-fluorophotometer (Ex (excitation)= 420 nm). CD spectrumwas record-ed on Jasco J-810-150L spectropolarimeter with scanning band between200 and 400 nm under the scanning rate of 100 nm/min. Fluorescentintercalator displacement (FID) assay was performed on M1000microplate reader (Tecan Trading Co. Ltd, Shanghai) using TO asfluorescence probe (Ex/Em (emission) = 501 nm/530 nm). Cellcycle analysis was recorded on FACS (Fluorescence Activated CellSorter) Aria II Flow Cytometer (BD Biosciences, San Jose, USA) (Ex/Em = 536 nm/617 nm). FRET assay was performed on 7500 fastReal-Time PCR (ABI Co. Ltd, USA) (for FAM staining: Ex/Em =494 nm/522 nm, for TAMRA staining: Ex/Em =560 nm/582 nm).Cell morphology was examined by a fluorescence microscope(Nikon TE2000, Japan) (for Hoechst 33258 staining: Ex/Em =346 nm/460 nm, for AO/EB staining: Ex/Em = 488 nm/510–550 nm,for JC-1 staining: Ex = 488 nm, Em = 535 nm (green), Em =595 nm (red)). Note: Ex, excitation wavelength; Em, emissionwavelength.

2.3. Cell culture and treatment

BEL-7404 human hepatoma cell line was obtained from the Instituteof Biochemistry and Cell Biology, Chinese Academy of Sciences(Shanghai, China). The cells were maintained in DMEM (Dulbecco'smodified eagle medium) supplemented with 10% fetal calf serum,100 units/mL ampicillin and 100 mg/mL streptomycin sulfate at37 °C in a humidified atmosphere under 5% CO2. Each compoundwas dissolved in DMSO at a concentration of 2 mM as a stock solution.The stock solution was diluted by DMEM to the required concentrationimmediately before use.

2.4. Flow cytometric analysis

Thedistribution of cell cycle phaseswas analyzed byflow cytometry.In the cell cycle assay, the BEL-7404 human hepatoma cells treatedwith0, 5, 10, 20 μM of 1 and L were harvested by trypsinization and rinsedwith PBS (phosphate buffered saline). After centrifugation, the pellet(105–106 cells) was suspended in 1 mL PBS. The cell suspension wasthen fixed by dropwise addition of 9mL of precooled (4 °C) 70% ethanolunder violent shaking. Fixed samples were kept at –20 °C before use.For staining, cells were centrifuged, resuspended in PBS, digestedwith 500 μL RNAse A (250 μg/mL), then were incubated for 30 minat 37 °C, and treated with 25 μL PI (100 μg/mL) for use. PI-positive cellswere counted with a FACS Aria II Flow Cytometer.

2.5. Fluorescence morphological examination

Cellular apoptotic morphological changes, which can be detected byHoechst 33258 staining and AO/EB staining respectively, were studiedby fluorescence microscopy. To examine whether L and 1 induce apo-ptosis on BEL-7404 human hepatoma cell line, the cells were plated insix-well plates, and treated with 0, 5, 10, and 20 μM of L and 1 for 24 h,respectively.

For Hoechst 33258 staining, the cells after treatment were washedwith PBS and fixed for 10 min at room temperature. The cells wererinsed twice in PBS and stained with Hoechst 33258 fluorescent dye,in dark for 10 min at room temperature. The cells were then washedtwice with PBS, examined and immediately photographed under thefluorescence microscope with excitation wavelength of 330–380 nm.

14 Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

Apoptotic cells were defined based on the nuclear morphology changessuch as chromatin condensation and fragmentation.

The AO/EBmolecular probes were also used to detect apoptotic cells[31]. Briefly, the cells were trypsinized and harvested, suspended in PBS,stained with 100 μg/mL AO and 100 μg/mL EB for 10 min at room tem-perature and visualized immediately by the fluorescence microscopywith excitation wavelength of 545 nm. More than 200 cells in randomwere assayed.

2.6. Measurement of mitochondrial membrane potential

The loss of mitochondrial membrane potential (Δψ) was assessedusing a lipophilic cationic fluorescent probe, JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine). Cells treated with 0, 5,10, and 20 μM of 1 and L for 24 h were incubated with 5 μg/mL JC-1 for20 min at 37 °C and examined under the fluorescence microscopy. Theemission fluorescence for JC-1 was monitored at 535 and 595 nm,under the excitation wavelength at 488 nm. The orange-red emissionof the dye is attributed to a potential-dependent aggregation in themitochondria, which reflects the Δψ. Green fluorescence reflects themonomeric form of JC-1, appearing in the cytosol after mitochondrialmembrane depolarization.

2.7. FID assay

Room temperature was kept constant inside the microplate reader.Each compound was tested on a line of the microplate in triplicate.The 96 microplate was filled with (a) 10 μL of a solution of G4-c-kit-1,G4-c-kit-2, G4-HTG21, G4-Pu22 (5 μM) and TO (10 μM) as fluorescentprobe; (b) a freshly prepared DMSO solution of each compound underconcentration gradient (0 to 2.5 μM along the line of the 96 microplate,from line 1–12: 0, 0.125, 0.25, 0.375, 0.5, 0.625, 0.75, 1.0, 1.25, 1.5, 2.0and 2.5 μM) and (c) a 200 μL of sodium cacodylate buffer solution.After shaking at 500 rpm for 5 min, the fluorescence was measuredby microplate reader. The percentage of displacement, (100 − [(F /F0) × 100])%, was calculated from the fluorescence intensity (F),where F and F0 refer to the fluorescence intensities of TO boundedwith DNA in the presence and absence of each compound, respec-tively. The percentage of displacement is then plotted as a functionof the concentration of the added compound. The DNA affinity wasevaluated by the concentration of compound required to decrease the50% of fluorescence of the probe (DC50), and was determined afternon-linear fitting for the displacement curve.

2.8. Fluorescence titration analysis

The fluorescence emission spectra were recorded under the ex-citation wavelength (Ex) as 420 nm with slit width as 3 nm/3 nm.The concentration of each compound for the working solution waskept at 20 μM. The G-quadruplex DNA was added with increasing[G4-HTG21]/[compound] ratios ranging from 0.05 to 0.5 at every0.05 interval.

The quenching constant, KSV, was calculated according to the classicStern–Volmer equation [32]:

I0=I ¼ 1þ KSV � Q½ �

where I0 and I arefluorescence intensities before or after the complex asthe quencher is added, respectively, KSV indicates apparent intrinsicbinding constant of interaction of complex with G4-HTG21, and [Q] rep-resents the concentration of as the quencher. In the plot of I0/I versus

Fig. 1. Cell cycle percentage of BEL-7404 human hepatoma cells detected by flow cytometry indgradient concentrations of 5, 10, and 20 μM for 48 h. The percentages of cells in the different p

[G4-HTG21], the apparent intrinsic binding constant, KSV, can be obtain-ed by linear fitting method.

2.9. Circular dichroism (CD) spectra

Two different buffers were prepared: (a) 10 mM Tris–HCl and100 mM KCl buffer pH 7.4 and (b) 10 mM Tris–HCl buffer pH 7.4 inthe absence of KCl. The corresponding oligonucleotide was dissolvedin buffer to prepare a stock solution of 100 μM. HTG21, G4-HTG21 aswell as each compound were diluted from the stock solutions to thefinal concentration. The concentrations of HTG21 and G4-HTG21 weremaintained at 5 μM. The concentration of each compound was kept at10 μM as working solution, with [HTG21 or G4-HTG21]/[compound] =1:2. The CD spectra were recorded in the wavelength range of220–600 nm in the 1 cmquartz cuvette. The instrumentwas flushed con-tinuously with pure evaporated nitrogen throughout the assay. All CDspectra were buffer-baseline-corrected.

2.10. FRET melting assay

Fluorescence resonance energy transfer (FRET) melting assay wascarried out on a real-time PCR apparatus. The labeled oligonucleotideF21T DNA (5′-FAM-GGGTTAGGGTTAGGGTTAGGG-TAMRA-3′), de-rived from the human telomeric sequence HTG21, was set as theFRET probe. FAM (6-carboxyfluorescein) and TAMRA (6-carboxy-tetramethylrhodamine) are two fluorophores that act as donorand acceptor, respectively. F21T was diluted by stock solution tothe required concentration of 10 μM in Tris buffer (10 mM Tris–HCland 100 mM KCl, pH 7.4). The ratios of [compound]/[F21T] were 0,0.2, 0.5, 0.8, 1.0, 1.2, and 1.5, respectively. The resulted 20 μL solu-tions were added into 96-well RT-PCR plates, and the measurementswere made in triplicate. Fluorescence intensities were recorded atevery 0.90 °C interval in the range of 37–95 °C, with the temperatureconstantly maintained for 30 s prior to each reading to ensurestability.

2.11. Telomerase activity assay

The BEL-7404 human hepatoma cells (5 × 106) in exponential phasewere pelleted and lysed on ice for 30 min. The lysate was centrifuged at13,000 rpm for 30 min at 4 °C and the supernatant was collected andstored at –80 °C before use. Telomerase inhibition ability of the com-pound was based on the TRAP–ELISA assay referring to themanufacturer's protocol of the telomerase PCR-ELISA kit (Roche,USA). Absorbance read by the microplate reader was reported asA450 nm against the blank (for reference wavelength as A690 nm). Forthe RNase treatment, the maximum absorbance for the negative con-trol (2.5 μL of the corresponding RNase treated or heated cell extract)should be lower than 0.25 × (A450 nm − A690 nm) units. The absor-bance for the positive control examined in 30 min reaction (cell ex-tract was prepared from the immortalized telomerase-expressinghuman kidney cells, and lyophilizate was supplied with the kit)should be higher than 1.5 × (A450 nm − A690 nm) units.

3. Results and discussion

3.1. Cell cycle arrest assay

As previously reported, complex 1 showed significant growthinhibition on the BEL-7404 human hepatoma cells, with IC50 value of6.91 ± 1.33 μM, which was much lower than that of free L (IC50 =

icating the G2/M and S phase cell cycle arrest respectively treatedwith L and complex 1 athases of cell cycle were reported inside the relative histogram.

}

15Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

Fig. 2. Cell morphological observation for cell apoptosis induction on the BEL-7404 cells treated by 1, L and cisplatin of 5, 10.0, and 20.0 μM for 24 h, respectively. Cells were stained byHoechst 33258 and selected visual fields illustrating the condensed chromatin (white arrow heading) as occurrence of cell apoptosis were shown.

16 Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

33.75± 3.10 μM) and cisplatin (IC50 = 98.04± 17.45 μM) [29]. To fur-ther investigate whether the cell cycle arrest and/or cell apoptosis canbe induced by complex 1, the cell cycle percentage of BEL-7404 cellstreated with complex 1was detected primarily by flow cytometry anal-ysis under the PI staining,with L and cisplatin for comparison. As shownin Fig. 1e–h, after treated with 1 at gradient concentrations of 5, 10, and20 μM for 48 h, respectively, the percentage of BEL-7404 cells at boththe G2/M phase and S phase increased. The cell percentage at G2/Mphase obviously increased from 7.33% to 13.16%, 26.18% and 31.03%under the concentration gradient of 1, while for S phase, the cell per-centage also slightly increased from 26.27% to 31.93%, 29.61% and34.15%. It indicated that complex 1 significantly induced cell cycle arrestmainly at G2/M phase along with a minor S phase arrest. It was foundthat free L also induced both S phase and G2/M phase cell cycle arrestunder the same condition. However, L induced S phase arrest moresignificantly than G2/M phase arrest, viewed from the detected cellcycle distribution in Fig. 1a–d (for S phase: from 26.27% up to 30.84%,38.14% and 39.68%; for G2/M phase: from 7.33% up to 10.50%, 14.16%and 18.57%). Comparatively, the cisplatin-induced cell cycle arrest onBEL-7404 cells was indicated to be a simplex S phase arrest, as shownin Fig. S1 (for S phase: from 25.32% up to 32.75%, 37.49% and 51.04%under the same condition). The characteristic S phase arrest inductionof cisplatin well agreed with the general acceptance that cisplatinacted as a DNA covalent-binding anticancer agent to block the DNAreplication. So it suggests the different action mechanism of complex1 to that of the mono-functional platinum(II) complex, cisplatin. Itmay also be different to that of free L, in which the central platinum(II)cation was expected to play key role.

3.2. Morphological observation of cell apoptosis by Hoechst 33258 andAO/EB staining

Since the cell cycle arrest induction was mostly correlated with thecell apoptosis induction [33], the potential cell apoptosis induction on

BEL-7404 cells by complex 1 was further examined by analyzing thecell morphology. After the cells were respectively treated with 1, Land cisplatin at gradient concentrations of 5, 10, and 20 μM for 24 hand followed by the membrane-permeable Hoechst 33258 staining,the cell morphologies were observed and analyzed comparatively, asshown in Fig. 2. Viewed from the control group, the nuclei of BEL-7404 cells retained the regular round contours, and cells with smallernuclei and condensed chromatin were rarely seen. After exposure toeach compound under the concentration gradient, the changes on thecell morphology can be clearly observed. It was found that even treatedwith complex 1 at lower concentrations of 5 and 10 μM, the contours ofsome of the BEL-7404 cells became irregular, the nuclei condensed (asbrightly blue fluorescence indicated) and the apoptotic bodies appeared(as arrows heading in Fig. 2f, g). When treated with 1 at higher con-centration of 20 μM, the nuclei of much more cells highly condensedand the apoptotic bodies were pervasive in the visual field (seeFig. 2h). It suggested the significant cell apoptosis induction of 1 onBEL-7404 cells. In contrast, less cell apoptosis induction on BEL-7404 cells was observed when treated with L or cisplatin under thesame condition. No obvious cell apoptosis can be observed underthe treatment of L or cisplatin at 5 and 10 μM. Till treated with L orcisplatin at the higher concentration of 20 μM, the characteristicsfor cell apoptosis, such as condensed nuclei and apoptotic bodies,can be observed.

Cell apoptosis induction via cell morphology observation was alsoexamined by AO/EB staining. As indicated in Fig. 3, the cells emittinggreen fluorescence in each group refer to the live BEL-7404 cells excitedby AO staining. After treated with 1 at increasing concentrations from 5to 20 μM,more andmore apoptotic cells with condensed nuclei and ap-optotic bodies (as white arrows heading) can be observed, along with asmall amount of necrotic cells (as red arrows heading), as shown inFig. 3f–h. Comparatively, seldom apoptotic cells were observed whentreated with 5 or 10 μM of L, until at the higher concentration of20 μM, a small amount of apoptotic and necrotic tumor cells were

Fig. 3. Cell morphological observation for cell apoptosis induction on the BEL-7404 cells treated by 1, L and cisplatin of 5, 10.0, and 20.0 μM for 24 h, respectively. Cells were stained by AO/EB and selected visual fields illustrating the corresponding live cells (green arrow heading), apoptotic cells (white arrow heading) and death cells (orange arrow heading) were shown.

17Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

observed. On the other hand, cisplatin induced more significant cell ap-optosis along with less necrosis than did L at the higher concentrationsof 10 and 20 μM. However, the cell apoptosis induction of cisplatin wasstill less than 1 under the same concentration.

Combining the cell morphological observation by Hoechst 33258and AO/EB staining, it strongly suggests the significantly higher cellapoptosis induction of complex 1 than the free L and cisplatin, whichshould be correlated with the different antitumor mechanism of 1.

3.3. Mitochondrial membrane potential assay for cell apoptosis

Since themitochondrionwas accepted as a vital cell organ to controlthe cell apoptosis by regulating the mitochondrial membrane potential(Δψ), it has been taken into account as a new antitumor target [34]. Theloss of Δψ is regarded to be characteristic for the cell apoptosis. Aimingto the mitochondria as new antitumor target, the mitochondrial mem-brane potential assay was then carried out to discuss if complex 1 mayinduce apoptosis on BEL-7404 cells via mitochondrial pathway. Thechanges on Δψ in BEL-7404 cells indicated by JC-1 staining wereshown in Fig. 4. In the live cells, JC-1molecules existed in themitochon-dria matrix of cytosol and accumulated to form J-aggregates, whichemitted orange fluorescence. While in the apoptotic and necrotic cells,due to the loss of Δψ, JC-1 molecules existed in monomeric form andstained the cytosol to emit green fluorescence. As indicated in Fig. 4a,e and i, all the cells stained by JC-1 in the control group emitted orangefluorescence, suggesting the coupled mitochondria with a normal Δψ.After treated with 1 at 5 μM, a few of the cells emitted green fluores-cence (as white arrows heading), which referred to the apoptotic cells(see Fig. 4f). However, when treated with 1 at higher concentrationsof 10 and 20 μM, as shown in Fig. 4g and h, the cells emitting green fluo-rescence were pervasive in the field of vision, suggesting the significantloss of Δψ for most of the tumor cells. Comparatively, the presence of Lor cisplatin under the same concentration gradient did not cause signif-icant loss of Δψ in the tumor cells, as shown in Fig. 4, except for the

20 μMof cisplatin-treatment, which induced considerable cell apoptosis(see Fig. 4l). This result is consistent with that of the cell morphologyassay above. It can be concluded that the significant cell apoptosis in-duction on BEL-7404 cells by complex 1 is neither simply by a DNA in-tercalation action like L nor simply by a DNA covalent binding actionlike cisplatin. Complex 1 was expected to exert a synergistic effect onthe antitumor mechanism when combining the L and the platinum(II)cation together.

3.4. FID assay for G-quadruplex DNA binding property

G-quadruplex DNA (G4-DNA) has become an important target forthe antitumor action of some planar and aromatic molecules, becausethe ligand-induced quadruplex formation and stabilization of G-quadruplexes by the telomeric G-rich strand were proven to be able toinhibit the telomerase activity, whichhas been regarded as a specific an-titumor target [35,36]. G4-FID assay is a simple and fast semi-quantitative method that is allowed to evaluate the binding affinity ofa small molecule for the G-quadruplex DNA, by using TO (thiazole or-ange) as a competitive binding probe. When TO interacts with G4-DNA, it can be protected by the hydrophobic environment betweenthe neighboring layers of G4-DNA, so that it may not be quenched bysolvent molecules and emits characteristic fluorescence. If TO wasforced out by other competitive molecules, its fluorescent emissionwill be quenched. Thus the displacement of TO by other moleculeswill provide an effective evidence for the competitive binding withG4-DNA. To quantify the binding affinity of complex 1 with G4-DNA,the concentration of the tested compound at which the TO fluorescenceintensity was quenched by 50% (DC50) was calculated, with L and cis-platin for comparison [37].

In the FID assay, four G-quadruplexes respectively derived from thetelomeric sequenceHTG21and oncogene sequences c-kit-1, c-kit-2, andPu22were selected for comparison. As shown in Fig. 5, the fluorescenceintensity of TO respectively boundwith G4-HTG21, G4-c-kit-1, G4-c-kit-

Fig. 4. Collapse ofmitochondrial membrane potential in the BEL-7404 cells treated by 1, L and cisplatin of 5, 10.0, and 20.0 μM for 24 h, respectively. Cells were stained by JC-1 and selectedvisual fields illustrated the corresponding live cells (orange-red) and apoptotic cells (green, shown as white arrow heading).

18 Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

2 and G4-Pu22 can be quenched in various degree under the treatmentwith each compound at increasing concentrations from 0.125 to 2.5 μM.Specifically, the DC50 values for 1 to replace the TO-bound G4-HTG21,G4-c-kit-1, G4-c-kit-2 and G4-Pu22 were determined to be 1.3 μM,2.2 μM, 2.2 μM and 2.3 μM, respectively (see Fig. 5b). While those for Lunder the same condition were determined to be 1.6 μM, 2.3 μM,2.5 μM and N2.5 μM, respectively (see Fig. 5a). For cisplatin, the DC50values were 2.0 μM towards G4-HTG21, and all N2.5 μM towards G4-c-kit-1, G4-c-kit-2 and G4-Pu22 (see Fig. 5c). From the results above, itclearly demonstrated that: (1) complex 1 displaced the TO and boundwith different types of G4-DNA more effectively, most probably byπ–π stacking, than did L and cisplatin; (2) even though the addition ofcisplatin also quenched the fluorescence of TO, there must existdifferent quenching actions between cisplatin and complex 1, due totheir totally different structures; (3) complex1, L and cisplatin all showedhigher binding affinity with the telomeric sequence G4-HTG21 than withthe oncogene sequences G4-c-kit-1, G4-c-kit-2 and G4-Pu22. So theG4-HTG21with the highest binding affinitywas selected for further com-parison for the G-quadruplex DNA binding affinity of complex 1, L andcisplatin using the following fluorescence and circular dichroism spectralanalyses.

Fig. 5. FID assay for the binding properties of 1, L and cisplatin towards the human te

3.5. Fluorescence spectral titration analysis

The fluorescence spectral analyses for complex 1 and L bound withG4-HTG21 was primarily performed. The fluorescence emission ofcisplatin was too weak to be examined. As shown in Fig. 6, in theabsence of DNA, both complex 1 and L in the aqueous solutions emittedthe characteristic fluorescence at around 500 nm under the excitationwavelength of 420 nm. With the addition of G4-HTG21 at gradient con-centrations from 0.5 to 4.0 μM, the peak fluorescence emission of 1 andL was both quenched gradually, but with different degrees. The totalquenching ratios of 1 and L were 20.3% and 8.4%, respectively, underthe addition of G4-HTG21 till 4.0 μM. And the apparent quenchingconstant, KSV, was consequently estimated to be 5.86 × 104 and1.65 × 104 M−1 for 1 and L, respectively (see the inset plots inFig. 6), which strongly suggested the much higher binding affinitywith G4-HTG21 of complex 1 than L.

3.6. Circular dichroism spectral analysis

CD is a useful and sensitive technique to assess whether the macro-molecules undergo conformational alterations as a result of complex

lomeric G4-HTG21 DNA and the oncogene G4-c-kit-1, G4-c-kit-2, G4-Pu22 DNA.

Fig. 6. Fluorescence spectral analysis of L (a) and complex 1 (b) under the titration of G4-HTG21with increasing [G4-HTG21]/[compound] ratios in the range from0.05 to 0.50 in every 0.05interval.

19Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

formation or changes under the environmental conditions, which issuitable for monitoring the conformational changes on G-quadruplexDNA [38,39]. The intramolecular and intermolecular conformationalconversions between various human telomeric G-quadruplexes, de-pending on the strand orientation,were initially observed during the in-vestigation on G-quadruplex stabilization [40]. The CD spectral analysiswas thus carried out to discuss whether complex 1 could better induceand stabilize the formation of G4-HTG21 DNA than L and cisplatin, so asto confirm the G-quadruplex binding priority of 1.

Firstly, CD spectral analysis was performed in the presence of K+,which facilitated the HTG21 sequence to form G-quadruplex DNA.As shown in Fig. 7, the CD spectrum of G4-HTG21 alone showed anegative absorption at 230 nm and two positive absorptions at253 nmand 293 nm, which suggested amixture of antiparallel and par-allel G-quadruplex conformations [41]. The addition of 2 equivalents of

Fig. 7. (upper) CD spectra of G4-HTG21 (10 μM) in Tris–KCl buffer solution in the absence (dashHTG21 sequence (10 μM) in K+-free Tris buffer solution in the absence (dashed line) and pres

L or 1 to G4-HTG21 solution induced regular CD spectral changes. In-duced by the addition of 1, the positive absorption around 293 nmshowed a slight decrement of 5.9%, and the positive absorption around253 nm showed an increment of 29.8%. Simultaneously, the negativeabsorption around 230 nm disappeared, as shown in Fig. 7b. It demon-strated that the antiparallel structure of G4-HTG21 tended to transforminto the more stable parallel structure under the interaction with 1.Comparatively, the addition of L induced a slight 5.3% decrement onthe positive absorption around 293 nm and a slight 7.0% increment onthe positive absorption around 253 nm (see Fig. 7a), which is not as in-tensive as 1 did while a slight 3.1% decrement around 293 nm and an11.7% increment around 253 nm were achieved under the addition ofcisplatin (see Fig. S2). From the results, complex 1 induced the mostsignificant spectral changes on the CD spectral of G4-HTG21, suggestingits higher binding affinity with G4-HTG21 than L and cisplatin. It can be

ed line) and presence (solid line) of 2.0 equivalents of L (panel a) or 1 (panel b); (lower)ence (solid line) of 2.0 equivalents of L (panel c) or 1 (panel d).

20 Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

explained as complex 1 having both the aromatic planar structure andthe central platinum(II) cation, which both facilitate the π–π stackingbetween 1 and G4-HTG21.

Secondly, whether complex 1 could induce the formation ofG-quadruplex DNA in the absence of K+ was also studied. This wasinvestigated by monitoring the CD spectra of HTG21 sequence in aK+-free buffer solution. As shown in Fig. 7c and d, the CD spectrumof HTG21 sequence alone showed a negative absorption around233 nm, a major positive absorption around 255 nm and a minorpositive absorption around 295 nm, which might be ascribed to thesignal of the randomly coiled HTG21 (characterized by the positiveabsorption around 257 nm) [42,43]. Under the addition of 2 equiva-lents of 1, as shown in Fig. 7d, the positive absorption around 295 nmincreased by 51.0% (corresponding to the antiparallel quadruplexDNA) and the positive absorption around 255 nm decreased by47.0% (corresponding to the unannealed HTG21 sequence), alongwith the dramatic decrease on the negative absorption around233 nm. The addition of L induced similar CD spectral alterations ofHTG21 sequence as 1 did, with a 20.2% decrease at 255 nm and a9.8% increase at 295 nm. The negative absorption around 233 nmalso decreased dramatically as induced by L, as shown in Fig. 7c. Asfor cisplatin, it also induced a 17.0% decrease at 253 nm and a 43.7% in-crease at 293 nm, respectively (see Fig. S2). These results are consistentwith the discussions above. It also indicated that complex 1 could in-duce the HTG21 sequence to transform into antiparallel G-quadruplexwith higher efficiency than L and cisplatin, even in the absence of K+.

3.7. FRET melting assay

Fluorescence resonance energy transfer (FRET) is a convenientmethod to monitor the distance from 3′- to 5′-end of oligonucleotides[44,45]. To the unfolded form of oligonucleotides, the average distancebetween two chromophores is generally larger than the critical Försterdistance (calculated to be ca. 5.0 nm), so the energy transfer is notexpected. However, intramolecular folding of the oligonucleotides isable to bring the two labeled chromophores at the 3′- and 5′-end beingclose enough for energy transfer, and thus the fluorescence quenchingof the chromophores will be observed. So the FRET melting assay wasalso applied to further investigate the G-quadruplex DNA stabilizationof complex 1 towards the labeled human telomeric sequence F21T(5′-FAM-GGGTTAGGGTTAGGGTTAGGG-TAMRA-3′), which could mimicthe human telomeric repeat [45].

As shown in Fig. 8, with increasing concentrations of 1 or L added till15 μM (as the [compound]/[G4-F21T] ratio increased up to 1.5), themelting point (Tm) of G4-F21T enhanced gradiently, which suggestedthat the stabilization of both complex 1 and L towards G4-F21Tappeared to be concentration-dependent. It strongly suggested that

Fig. 8. FRETmelting curves of oligonucleotideG4-F21T (5′-FAM-GGGTTAGGGTTAGGGTTAGGG-ing concentrations; r = [compound]/[G4-F21T].

both 1 and L could stabilize the G4-F21T in an intercalative bindingmode via π–π stacking, due to their same aromatic planar structure.However, under the same condition, complex 1 enhanced the meltingpoint of G4-F21T with a total ΔTm of 14.7 °C (see Fig. 8b). It is muchhigher than that of L, which led to a total ΔTm of only 5.9 °C, as shownin Fig. 8a. The higher binding affinity of complex 1 towards G4-F21Tshould be attributed to the extra electrostatic interaction between G4-F21T and the positively charged platinum(II). Comparatively, asshown in Fig. S3, the addition of cisplatin did not induce significant en-hancement on the melting point of G4-F21T, resulting from the non-aromatic planar structure of cisplatin.

3.8. Telomerase activity inhibition studies

It is generally accepted that the stabilization of G-quadruplex DNAby small molecules or the ligand-induced quadruplex formation bythe telomeric G-rich strand might be closely related to the inhibitionon telomerase activity [8,14–16]. The results above had shown good co-herence in terms of the higher binding affinity of complex 1 than LwithG4-HTG21 DNA. So as the final confirmation, the inhibition abilities ontelomerase activity of complex 1 and L were compared by using thetelomerase PCR-ELISA assay, from which we may conclude the highertelomerase inhibition of complex 1. To avoid the false positive resultsdue to the drug interference, both the negative control and positivecontrol were utilized in this assay. The calculated data (as OD valuesand inhibition ratios under the triplicate experiments) were summa-rized in Table 1, from which it was found that both complex 1 and Lexhibited the inhibition abilities on telomerase activity. However, the62.50% of inhibition ratio for 1 was about three times higher than the21.77% of inhibition ratio for L. Obviously, the intuitive result of the tel-omerase activity inhibition was well consistent with the above resultsfrom the spectral analyses and FRET assay, which together confirmedthe significantly higher inhibition ability on telomerase activity of com-plex 1 than free L.

4. Conclusions

In this study, one platinum(II) complex (1) of liriodenine (as TCMactive ingredient) with significant growth inhibition against BEL-7404humanhepatoma cell line as previously reportedwasdiscussed indetailon its potential antitumor mechanism on both the cellular and molecu-lar levels. The primary flow cytometry assay revealed that complex 1caused both the G2/M phase and S phase cell cycle arrest, suggestingits different antitumor mechanism of 1 to liriodenine and cisplatin.From the results of the H33258 and AO/EB staining assay by fluores-cence microscope observation, complex 1 was confirmed to effectivelyinduce apoptosis on BEL-7404 cells, which suggested that 1 is not a

TAMRA-3′) (1 μM, in Tris–KCl buffer, pH 7.4) bound by L (a) or complex 1 (b) with increas-

Table 1Inhibition abilities of1 and L on the telomerase activity at concentration of 2.0 μMbyusingthe telomerase PCR-ELISA assay (each experiment was performed in triplicate).

Average (OD) Inhibition ratios (%)

Complex 1 1.35 ± 0.01 62.50 ± 0.03L 2.81 ± 0.00 21.77 ± 0.01Negative control 0.05 ± 0.01 –

Positive control 1.54 ± 0.01 –

Control 3.59 ± 0.02 –

21Y.-L. Li et al. / Journal of Inorganic Biochemistry 137 (2014) 12–21

simplex cytotoxic agent. By JC-1 staining assay, complex 1 was alsofound to collapse the mitochondria membrane potential, which sug-gested the cell apoptosis induction by 1 via the mitochondrial pathway.The following spectroscopic analyses indicated that complex 1 signifi-cantly induced the formation and stabilization of the telomeric G-quadruplex DNA, most probably by intercalative binding mode viaπ–π stacking and by exterior electrostatic attraction between the posi-tively charged platinum(II) and G-quadruplex DNA. The final confirma-tion on the telomerase inhibition ability of 1was proven by telomerasePCR-ELISA assay, which provided another possible apoptotic pathwayfor 1 via inhibiting the telomerase activity based on the G-quadruplexDNA stabilization. These results demonstrated that complex 1 exertedmultiple antitumormechanisms, including cell cycle arrest, cell apopto-sis induction via mitochondrial pathway and telomerase inhibition,which showed promising potentials for this kind of complexes to devel-op new platinum-based antitumor medicines.

Abbreviations

TCM traditional Chinese medicineTCMAI traditional Chinese medicine active ingredientsHoechst 33258 bisbenzimide H 33258AO/EB acridine orange/ethidium bromideJC-1 5,5′,6,6′-tetrachloro-1,1′,3,3′-

tetraethylbenzimidazolylcarbocyanineFID fluorescent indicator displacementCD circular dichroismFRET fluorescence resonance energy transferHTG21 5′-G3TTAG3TTAG3TTAG3-3′PCR polymerase chain reactionIC50 half maximal inhibitory concentrationTO thiazole orangec-kit-1 5′-CG3CG3CACGAG3AG3T-3′c-kit-2 5′-CG3CG3CGCTAG3AG3T-3′Pu22 5′-TGAG3TG3TAG3TG3TAA-3′PI propidium iodideFAM 6-carboxyfluoresceinTAMRA 6-carboxytetramethylrhodaminePBS phosphate buffered salineF fluorescence intensityDC50 decrease the 50% of fluorescence of the probeF21T 5′-FAM-GGGTTAGGGTTAGGGTTAGGG-TAMRA-3′RT-PCR reverse transcription-polymerase chain reactionTRAP telomere repeat amplification protocolTm melting point

Acknowledgments

This work was financially supported by the National Natural ScienceFoundation of China (No. 21271051), the Natural Science Foundation ofGuangxi Province (Nos. 2012GXNSFDA053005, 2012GXNSFDA385001,2013GXNSFAA019044) and BAGUI scholar program of GuangxiProvince(No. 2011A003).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jinorgbio.2014.04.001. These data include MOL filesand InChiKeys of the most important compounds described in thisarticle.

References

[1] P. Wang, C.H. Leung, D.L. Ma, S.C. Yan, C.M. Che, Chem. Eur. J. 16 (2010) 6900–6911.[2] T.M. Ou, J. Lin, Y.J. Lu, J.Q. Hou, J.H. Tan, S.H. Chen, Z. Li, Y.P. Li, D. Li, L.Q. Gu, Z.S.

Huang, J. Med. Chem. 54 (2011) 5671–5679.[3] T. de Lange, Genes Dev. 19 (2005) 2100–2110.[4] V.A. Zakian, Science 270 (1995) 1601–1607.[5] R. McElligott, R.J. Wellinger, EMBO J. 16 (1997) 3705–3714.[6] N.W. Kim, M.A. Piatyszek, K.R. Prowse, C.B. Harley, M.D. West, P.L.C. Ho, G.M.

Coviello, W.E. Wright, S.L. Weinrich, J.W. Shay, Science 266 (1994) 2011–2014.[7] G.W. Collie, R. Promontorio, S.M. Hampel, M. Micco, S. Neidle, G.N. Parkinson, J. Am.

Chem. Soc. 134 (2012) 2723–2731.[8] A.M. Burger, F. Dai, C.M. Schultes, A.P. Reszka, M.J. Moore, J.A. Double, S. Neidle,

Cancer Res. 65 (2005) 1489–1496.[9] T. Tauchi, K. Shin-Ya, G. Sashida, M. Sumi, S. Obake, J.H. Ohyashiki, K. Ohyashiki,

Oncogene 25 (2006) 5719–5725.[10] J.L. Zhou, Y.J. Lu, T.M. Ou, J.M. Zhou, Z.S. Huang, X.F. Zhu, C.J. Du, X.Z. Bu, L. Ma, L.Q.

Gu, Y.M. Li, A.S.C. Chan, J. Med. Chem. 48 (2005) 7315–7321.[11] J.L. Mergny, J.F. Riou, P. Mailliet, M.P. Teulade-Fichou, E. Gilson, Nucleic Acids Res. 30

(2002) 839–865.[12] A.D. Cian, L. Lacroix, C. Douarre, N.T. Smaali, C. Trentesaux, J. Riou, J.L. Mergny,

Biochimie 90 (2008) 131–155.[13] E.M. Rezler, D.J. Bearss, L.H. Hurley, Annu. Rev. Pharmacol. Toxicol. 43 (2003) 359–379.[14] S. Redon, P. Reichenbach, J. Lingner, Nucleic Acids Res. 38 (2010) 5797–5806.[15] R. Kieltyka, J. Fakhoury, N. Moitessier, H.F. Sleiman, Chem. Eur. J. 14 (2008)

1145–1154.[16] T.M. Ou, Y.J. Lu, J.H. Tan, Z.S. Huang, K.Y. Wong, L.Q. Gu, ChemMedChem 3 (2008)

690–713.[17] G.R. Suzanne, S.P. Daniel, L. Angela, L. Leroy, J.L. Edmond, E.R. Joseph, J. Med. Chem.

53 (2010) 3632–3644.[18] H. Kostrhunova, J. Malina, A.J. Pickard, J. Stepankova, M. Vojtiskova, J. Kasparkova, T.

Muchova, M.L. Rohlfing, U. Bierbach, V. Brabec, Mol. Pharm. 8 (2011) 1941–1954.[19] J. Zhu, Y. Zhao, Y. Zhu, Z. Wu,M. Lin,W. He, Y. Wang, G. Chen, L. Dong, J. Zhang, Y. Lu,

Z. Guo, Chem. Eur. J. 15 (2009) 5245–5253.[20] J. Reedijk, Eur. J. Inorg. Chem. (2009) 1303–1312.[21] L. Wang, S. Gou, Y. Chen, Y. Liu, Bioorg. Med. Chem. Lett. 15 (2005) 3417–3422.[22] K.K.W. To, S.C.F. Au-Yeung, Y.-P. Hou, Anti Cancer Drugs 17 (2006) 673–683.[23] F. Tillequin, Phytochem. Rev. 6 (2007) 65–79.[24] K. Deepralard, T. Pengsuparp, M. Moriyasu, K. Kawanishi, R. Suttisri, Biochem. Syst.

Ecol. 35 (2007) 696–699.[25] C.Y. Chen, L.Y. Huang, L.J. Chen, W.L. Lo, S.Y. Kuo, Y.D. Wang, S.H. Kuo, T.J. Hsieh,

Chem. Nat. Compd. 44 (2008) 137–139.[26] H.-C. Chang, F.-R. Chang, Y.-C.Wu, Y.-H. Lai, Kaohsiung, J. Med. Sci. 20 (2004) 365–371.[27] M.M. Rahman, S.S. Lopa, G. Sadik, Harun-or-Rashid, R. Islam, P. Khondkar, A.H.M.K.

Alam, M.A. Rashid, Fitoterapia 76 (2005) 758–761.[28] E.V. Costa, M.L.B. Pinheiro, C.M. Xavier, J.R.A. Silva, A.C.F. Amaral, A.D.L. Souza, A.

Barison, F.R. Campos, A.G. Ferreira, G.M.C. Machado, L.L.P. Leon, J. Nat. Prod. 69(2006) 292–294.

[29] Z.-F. Chen, Y.-C. Liu, L.-M. Liu, H.-S. Wang, S.-H. Qin, B.-L. Wang, H.-D. Bian, B. Yang,H.-K. Fun, H.-G. Liu, H. Liang, C. Orvig, Dalton Trans. (2009) 262–272.

[30] Z.-F. Chen, Y.-C. Liu, Y. Peng, X. Hong, H.-H. Wang, M.-M. Zhang, H. Liang, J. Biol.Inorg. Chem. 17 (2012) 247–261.

[31] W.M. Zawada, D.L. Kirschman, J.J. Cohen, K.A. Heidenreich, C.R. Freed, Exp. Neurol.140 (1996) 60–67.

[32] A. Ambrus, D. Chen, J. Dai, T. Bialis, R.A. Jones, D.Z. Yang, Nucleic Acids Res. 34 (2006)2723–2735.

[33] C.-C. Chou, J.-S. Yang, H.-S. Lu, S.-W. Ip, C. Lo, C.-C. Wu, J.-P. Lin, N.-Y. Tang, J.-G.Chung, M.-J. Chou, Y.-H. Teng, D.-R. Chen, Arch. Pharm. Res. 33 (2010) 1181–1191.

[34] D.R. Green, Cell 94 (1998) 695–698.[35] A.J. Zaug, E.R. Podell, T.R. Cech, Proc. Natl. Acad. Sci. U. S. A. 102 (2005)

10864–10869.[36] K. Suntharalingam, A.J.P. White, R. Vilar, Inorg. Chem. 48 (2009) 9427–9435.[37] D.D. Sun, Y.N. Liu, D. Liu, R. Zhang, X.C. Yang, J. Liu, Chem. Eur. J. 18 (2012)

4285–4295.[38] S. Schäfe, I. Ott, R. Gust, W.S. Sheldrick, Eur. J. Inorg. Chem. (2007) 3034–3046.[39] V. Rajendiran, M.Murali, E. Suresh, M. Palaniandavar, V.S. Periasamy, M.A. Akbarsha,

Dalton Trans. (2008) 2157–2170.[40] R. Giraldo, M. Suzuki, L. Chapman, D. Rhodes, Proc. Natl. Acad. Sci. U. S. A. 91 (1994)

7658–7662.[41] E.M. Rezler, J. Seenisamy, S. Bashyam, M.Y. Kim, E. White, W.D.Wilson, L.H. Hurley, J.

Am. Chem. Soc. 127 (2005) 9439–9447.[42] D.P.N. GonÅalves, R. Rodriguez, S. Balasubramanian, J.K.M. Sanders, Chem. Commun.

(2006) 4685–4687.[43] T.J. Meade, J.F. Kayyem, Angew. Chem. Int. Ed. 34 (1995) 352–354.[44] R.M. Clegg, A.I.H. Murchie, A. Zechel, C. Carlberg, S. Diekmann, D.M.J. Lilley, Biochem-

istry 31 (1992) 4846–4856.[45] J.L. Mergny, J.C. Maurizot, ChemBioChem 2 (2001) 124–132.