ORIGINAL PAPER Bioactive papaverine derivatives ...

Chemical Papers 66 (2) 79–84 (2012) DOI: 10.2478/s11696-011-0092-4

ORIGINAL PAPER

Bioactive papaverine derivatives bind G-quadruplexes selectively‡ a

Elzbieta Galezowska, a Joanna Kosman, a Agnieszka Stepien, b Blazej Rubis, b Maria Rybczynska, a Bernard Juskowiak* a Faculty

of Chemistry, Adam Mickiewicz University, Grunwaldzka 6, 60-780 Poznan, Poland

b Department

of Clinical Chemistry and Molecular Diagnostics, University of Medical Sciences, Przybyszewskiego 49, 60-355 Poznan, Poland

Received 24 May 2011; Revised 27 July 2011; Accepted 2 August 2011

G-quadruplexes are a family of DNA secondary structures resulting from the folding of a guaninerich sequence. Targeting quadruplexes by small molecules is an approach that is currently being studied with the aim of exploring their biological roles and developing new anti-cancer agents. There is evidence that the formation of G4 structures by telomeric DNA can be used to inhibit the enzyme activity of telomerase, and thereby to activate the pathway to senescence in tumour cells. It was previously shown that the papaverine oxidation products 6a,12a-diazadibenzo-[a,g]fluorenylium derivative (ligand I ) and 2,3,9,10-tetramethoxy-12-oxo-12H-indolo[2,1-a]isoquinolinium chloride (ligand II ) bind to G-quadruplex representing the human telomeric sequence. These ligands possess the ability to inhibit telomerase and polymerase action at the micromolar level. Here we report a DNA binding study on these two ligands and a new derivative 2-(2-carboxy-4,5-dimethoxyphenyl06,7-dimethoxyisoquiloliniuminner salt (ligand III ) in order to evaluate their binding selectivity to samples of nucleic acids (ssDNA, dsDNA, triplexes, and quadruplexes). Simultaneous investigations on several DNA–ligand complexes carried out using an equilibrium dialysis approach revealed pronounced binding selectivity of ligand I and ligand II to tetraplex DNA structures over the doublestranded DNA forms. c 2011 Institute of Chemistry, Slovak Academy of Sciences Keywords: DNA binding selectivity, equilibrium dialysis, fluorescence, G-quadruplex, papaverine derivatives, telomerase

Introduction In recent years a number of small molecules have been reported as stabilising G-quadruplex structures and inhibiting telomerase activity (Balasubramanian & Neidle, 2009; Mergny et al., 2002; Neidle, 2009; Neidle & Parkinson, 2002). DNA sequences with stretches of multiple guanines can form four-stranded tetraplex DNA structures known as guanine-quadruplexes or G4 DNA. In eukaryotic systems, guanine-rich sequences are concentrated at the ends of chromosomes and are known as telomeric DNA. In vitro characterisa-

tion of G-quadruplexes indicates four-stranded structures containing one or more nucleic acid strands, in parallel or anti-parallel orientations (Davis, 2004; Neidle & Balasubramanian, 2006; Simonsson, 2001). Four guanines on a plane interacting via Hoogsteenbonding form a G-quartet (Fig. 1A). Typically, three or four G-quartets are stacked and held together by π–π non-bonded attractive interactions, thus forming G-quadruplexes with different topological structures as shown in Figs. 1B–1D. The coordination of certain metal cations stabilises G-quadruplex, as do some small organic molecules

*Corresponding author, e-mail: [email protected] ‡ Presented at 38th International Conference of the Slovak Society of Chemical Engineering, Tatranské Matliare, 23–27 May 2011.

Unauthenticated Download Date | 9/24/15 11:48 PM

80

E. Galezowska et al./Chemical Papers 66 (2) 79–84 (2012)

Fig. 1. (A) Structure of G-quartet showing hydrogen bonds between four guanine units and interaction with a cation (filled circle). Schematic representation of G-quadruplex structures: intramolecular anti-parallel “chair-type” G-quadruplex with all lateral loops (B); intramolecular anti-parallel “basket-type” G-quadruplex with one diagonal and two lateral loops (C); an intramolecular parallel quadruplex with all loops positioned alongside the grooves (D).

Fig. 2. Chemical structures of papaverine derivatives (ligands I–III ) used in this study.

that promote the formation and/or stabilisation of Gquadruplexes (Davis, 2004; Simonsson, 2001). Some of these molecules have been shown to induce telomere shortening and instability, triggering apoptosis in various tumour cell lines (Balasubramanian & Neidle, 2009; Neidle, 2009). Consequently, there is considerable interest in the design of ligands that target G-quadruplex DNA. In addition, an evaluation of quadruplex–ligand affinity and quadruplex–duplex DNA selectivity is required to select the putative telomerase inhibitors. To be effective, screening for selectivity has to be performed at the early stage of development and this should, ideally, be rapid and easy to implement. Several methods are currently employed including competitive or equilibrium dialysis (Ragazzon et al., 2007; Czerwinska & Juskowiak, 2011). Competition dialysis has proved to be a powerful and versatile tool to study the molecular recognition of new targets. The method is based on thermodynamic equilibration, is simple to implement, and allows the rapid identification of structure-selective binding interactions. We recently reported a new G-quadruplex interacting ligand, 6a,12a-diazadibenzo-[a,g]fluorenylium derivative (ligand I, Fig. 2), which binds to Gquadruplex with a much higher affinity than to single- and double-stranded DNAs (Juskowiak et al., 2004; Madry et al., 2006). This fluorenylium derivative was obtained as a by-product of papaverine oxidation with Hg(II) acetate (Hermann et al., 2002). On the other hand, photo-oxidation of papaverine gives another product with a rigid

planar structure, 2,3,9,10-tetramethoxy-12-oxo-12Hindolo[2,1-a]isoquinolinium cation (ligand II, Fig. 2) (Girreser et al., 2003). Both ligands exhibit remarkable cytotoxicity and ability to inhibit telomerase activity (Galezowska et al., 2007; Rubis et al., 2009). The family of papaverine derivatives was recently increased by the addition of a new compound, 2-(2-carboxy-4,5dimethoxyphenyl)-6,7-dimethoxyisoquinolinium inner salt (ligand III, Fig. 2), reported by Girreser et al. (2009). Since selective recognition of a particular DNA structure (e.g. duplex vs. quadruplex) by an interacting ligand is a crucial parameter of the putative drug, we report here on the study of selective binding to various forms of nucleic acids (ssDNA, dsDNA, triplexes, and quadruplexes) that were carried out using the equilibrium dialysis assay.

Experimental The 6a,12a-diazadibenzo-[a, g]fluorenylium derivative, 2,3,9,10-tetramethoxy-12-oxo-12H-indolo[2,1-a] isoquinolinium chloride, and 2-(2-carboxy-4,5-dimethoxyphenyl)-6,7-dimethoxyisoquinolinium inner salt (ligand I, II, and III, respectively, Fig. 2) were prepared in accordance with the procedures described elsewhere (Hermann et al., 2002; Girreser et al., 2003, 2009). The stock solutions (ca 2.5 mM) were prepared in pure EtOH (ligand I ) or in aqueous solution of EtOH–H2 O (ϕr = 50 %) (ligand II and ligand III ); working solutions were prepared by dilution with water or other solvents. The nucleic acid samples


81


Table 1. Nucleic acid samples used in equilibrium dialysis assay Structure

Entry

Nucleic acid

Monomeric unit

Single-stranded

1 2 3

Poly [dA] Poly [rA] Poly [rU]

Nucleotide

Double-stranded

4 5 6 7 8 9

Poly [dA]*Poly [dT] Poly [dG]*Poly [dC] Poly [dA-dT]*Poly [dA-dT] Poly [dG-dC]*Poly [dG-dC] 5’-CAATCGGATCGAATTCGATCCGATTG-3’ 5’-GAGAGAGAGAGAGAGAGAGAGAGA-3’

Base pair

Triple-stranded

10

Four-stranded

11

TT-CTTTCTCTCCTCC-3’ 5’-GAAAGAGAGGAGG-3’ CC-CTTTCTCTCCTCC-5’ Poly [dT]*Poly [dA]*Poly [dT]

12 13 14

5’-AGGGTGGGGAGGGTGGGG-3’ 5’-AGGGTTAGGGTTAGGGTTAGGG-3’ 5’-TTGGGGGGGGGGGGGGGGGGGGTT-3’

used in the equilibrium dialysis experiments are listed in Table 1. Polynucleic acids were purchased from Sigma Chemical Co. (St. Louis, USA) and were used as received. Oligonucleotides were synthesised and HPLC-purified by Metabion Int. AG (Martinsried, Germany). Other reagents were of analytical grade purity and were used as received. High-purity water (Polwater, Poland) was used throughout this study. Absorption spectra were recorded with a Specord M40 spectrophotometer (Jena, Germany). Steady-state fluorescence measurements were carried out on a RF5000 spectrofluorimeter (Shimadzu, Japan) with excitation and emission bandwidths of 5 nm. Cell compartments were thermostated at 25 ◦C. All measurements were carried out using a 10 mm quartz cell, and the fluorescence spectra were not corrected. Absorption and fluorescence measurements in 96-well microplates were performed with an Infinity M200 microplate reader (Tecan, Austria) controlled by Magellan software. Equilibrium dialysis assay: Nucleic acid samples were prepared at the same concentration of 75 µM in terms of the monomeric units (base, base-pair, triplet, quartet). The triplex and quadruplex samples were initially heated at 90 ◦C for 4 min in a water bath followed by slow cooling. All samples were incubated at 4 ◦C for 24 h prior to the dialysis experiment. Fifteen dialysis units (Slide-A-Lyzer MINI dialyser unit (Pierce, IL, USA), MWCO (Molecular weight cut-off of 3500)), each containing 100 µL volume of different nucleic acid and one blank unit (buffer only), were placed into a beaker containing 250 mL of dialysate solution: 1 µM ligand in sodium cacodylate buffer (15 mM, pH 6.85) containing 10 mM of MgCl2 and 185 mM of NaCl. The beaker was covered with aluminium foil and equilibrated under continuous stirring at room temperature for 24 h. Finally, a 90 µL aliquot

Triplet

Quartet

of each sample was transferred into a microplate and SDS (sodium dodecylsulphate) solution was added to dissociate the ligand from the ligand–DNA complex (a final SDS concentration of 1 % was used). The absorption and emission spectra of all samples were recorded with a microplate reader in the following wavelength ranges: ligand I, of 230–500 nm (absorption) and 400–700 nm (fluorescence, λexc 330 nm); ligand II, 230–550 nm (absorption) and 380–780 nm (fluorescence, λexc 310 nm); ligand III, 230–450 nm (absorption) and of 420–780 nm (fluorescence, λexc 330 nm). The concentration of bound ligand (Cb ) was calculated from Cb = Ct – Cf , where the total concentrations of ligand (Ct ) and the free ligand concentration (Cf ) were determined using absorption or fluorescence calibration graphs. The free ligand concentration (reagent blank) did not vary appreciably from its initial value of 1 µM.

Results and discussion To evaluate the DNA-binding selectivity of papaverine-derived ligands I–III, competition dialysis experiments were carried out using the 14 nucleic acid structures listed in Table 1 that were known to form a variety of structures: single-stranded (entries 1–3), double-stranded (entries 4–9) with entry 9 that forms a parallel duplex (Rippe et al., 1992), triplex (entries 10 and 11), and quadruplex (entries 12–14). It should be stated that oligonucleotide entry 12 possesses cmyc oncogene sequence and oligonucleotide entry 13 represents the human telomere sequence. Both can form intramolecular G-quadruplex structures by the folding of a single oligonucleotide molecule. On the other hand, sample entry 14 forms an intermolecular parallel G-quadruplex due to the association of four DNA molecules. In competition dialysis experiments,


82

Ligand concentration/µM


buffer 1

2

3

4

5

6

7

8

9 10 11 12 13 14

Ligand concentration/µM

Fig. 3. Results obtained by equilibrium dialysis method for papaverine derivatives (ligands I–III ). Measurements were performed in sodium cacodylate buffer (15 mM, pH 6.85) containing 10 mM of MgCl2 and 185 mM of NaCl. Entries 1–14 stand for nucleic acid samples and are explained in Table 1.

buffer 1

2

3

4

5

6

7

8

9 10 11 12 13 14

Fig. 4. Comparison of equilibrium dialysis results for ligand II obtained from absorbance ( ) and fluorescence ( ) measurements.

the amount of ligand accumulated in the dialysis tube containing a particular nucleic acid structural form can be directly related to the binding affinity of this structure. Fig. 3 shows the results obtained in the form of a bar graph representing the ligand concentration accumulated by each DNA sample. Ligand I exhibits pronounced binding selectivity to triplex and quadruplex forms of DNA (entries 10–14). Single-stranded nucleic acids bind the ligand very weakly but duplexes present a moderate affinity. Only duplex sample entry 4, representing poly [dA]*poly [dT], appears to be the preferred structure that is consistent with the high affinity of ligand I towards the triplex poly [dA]*(poly [dT])2 (entry 10). The preference for quadruplex over duplex binding selectivity of ligand I observed in the

equilibrium dialysis experiments is consistent with previous reports based on spectrophotometric titration measurements (Juskowiak et al., 2004; Madry et al., 2006). The highest DNA-binding affinity is observed for ligand II–c-myc G-quadruplex system (entry 12). Ligand II also interacts preferentially with other quadruplex and triplex structures, whereas it shows a weak affinity towards single-stranded samples (entries 1–3) and duplex structures (entries 4–9). The exception is poly [dG]*poly [dC] (entry 5), to which ligand II binds with an affinity comparable with that of the triplex forms. Ligand III exhibited the poorest DNA-binding properties, as the amounts of ligand III accumulated in the particular dialysis tubes were comparable with the control value (buffer). Variations in the DNA-binding affinities of these three ligands can be explained in terms of their structural differences. As shown in Fig. 2, ligands I and II are featured with extended planar aromatic systems and possess positive charges that enable efficient stacking and electrostatic interactions with nucleobases and phosphate groups, respectively. By contrast, ligand III has only the small planar isoquinolinium ring substituted with the twisted carboxydimethoxyphenyl group. Interaction with DNA is further suppressed by the lack of a permanent positive charge in this ligand (inner salt structure). It should be noted, however, that the equilibrium dialysis results for ligand I and III were consistent and reproducible using both absorbance and fluorescence measurements, whereas the results for ligand II obtained with these two techniques were ambiguous. Fig. 4 compares the equilibrium dialysis results for ligand II calculated using absorbance and fluorescence measurements, respectively. The amounts of ligand accumulated in the dialysis tubes calculated from fluorescence measurements are several times higher than those obtained spectrophotometrically. There are two reasons that support the validity of the absorbancebased results. Firstly, the fluorescence-based results suggest that ligand II binds G4 DNA with a much higher affinity than ligand I. This was not true, since the results of the Scatchard analysis of titration data reported earlier revealed that ligands I and II interacted with G-quadruplex DNA showing a comparable binding affinity (Madry et al., 2006; Galezowska et al., 2007). Secondly, the fluorescence spectra of ligand II in the dialysed samples show significant changes when compared with those for the reference solutions prepared for the calibration graph as shown in Fig. 5A. The dramatic difference observed in the fluorescence characteristics cannot be explained by the presence of DNA in the dialysed solution, since both spectra were recorded in the same environment after the addition of SDS (see experimental procedure). An additional emission band at ca 590 nm is clearly seen in the spectrum of the dialysed c-myc quadruplex– ligand II system (dashed line in Fig. 5A) that sug-


83


B

Fluorescence intensity/a.u.

A

entry entry entry entry entry

Fig. 5. (A) Examples of normalised fluorescence spectra of ligand II: solid line represents the reference spectrum of ligand II in buffer and dashed line shows the spectrum of the quadruplex-containing solution (entry 13) after the equilibrium dialysis experiment. (B) Fluorescence spectra of other DNA samples dialysed with ligand II.

gests the appearance of a new emitting species. The fluorescence spectra of the other dialysed samples also exhibited this long-wavelength band with an intensity that depended on the binding affinity of nucleic acid as shown in Fig. 5B. Moreover, the dual-band emission for dialysed samples was also observed prior to SDS addition. Hypotheses that a long incubation time and contact with DNA during dialysis may affect the properties of ligand II or that the SDS-driven dissociation of ligand–DNA complex is very slow were negatively verified by recording the fluorescence spectra of entry 13-ligand II solution incubated in the dark for 24 h. No spectral changes in the fluorescence band were observed with the time of incubation, which may suggest an important role of the dialysis membrane or specific permeation processes. HPLC analysis of the dialysed solution and G4 DNA (entry 13)–ligand II control mixture before and after incubation did not reveal new peaks or significant differences in retention times of oligonucleotide and ligand. Further studies are required for an explanation of these interesting fluorescence results for ligand II, involving other analytical methods, for example HPLC–MS technique. Acknowledgements. This work was supported in part by Research Grant No. N N401 223534 from the Ministry of Science and Higher Education, Poland.

References Balasubramanian, S., & Neidle, S. (2009). G-quadruplex nucleic acids as therapeutic targets. Current Opinion in Chemical Biology, 13, 345–353. DOI: 10.1016/j.cbpa.2009.04.637. Czerwinska, I., & Juskowiak, B. (2011). Differences in affinity of arylstilbazolium derivatives to tetraplex structures. Chemical Papers, 65, 242–245. DOI: 10.2478/s11696-010-0078-7. Davis, J. T. (2004). G-quartets 40 years later: From 5’-GMP to molecular biology and supramolecular chemistry. Angewandte Chemie International Edition, 43, 668–698. DOI: 10.1002/anie.200300589. Galezowska, E., Masternak, A., Rubis, B., Czyrski, A., Rybczy´ nska, M., Hermann, T. W., & Juskowiak, B. (2007). Spectroscopic study and G-quadruplex DNA binding affin-

ity of two bioactive papaverine-derived ligands. International Journal of Biological Macromolecules, 41, 558–563. DOI: 10.1016/j.ijbiomac.2007.07.008. Girreser, U., Czyrski, A., & Hermann, T. W. (2009). Synthesis and structure elucidation of a new isoquinolinium inner salt. Tetrahedron Letters, 50, 4610–4612. DOI: 10.1016/j.tetlet.2009.05.099. Girreser, U., Hermann, T. W., & Piotrowska, K. (2003). Oxidation and degradation products of papaverine, part II[1]: Investigations on the photochemical degradation of papaverine solutions. Archiv der Pharmazie, 336, 401–405. DOI: 10.1002/ardp.200300751. Hermann, T. W., Girreser, U., Michalski, P., & Piotrowska, K. (2002). Oxidation and degradation products of papaverine. Part I: Gadamer and Schulemann’s papaverinol synthesis revisited. Archiv der Pharmazie, 335, 167–169. DOI: 10.1002/1521-4184(200204)335:43.0. CO;2-O. Juskowiak, B., Galezowska, E., Koczorowska, N., & Hermann, T. W. (2004). Aggregation and G-quadruplex DNA-binding study of 6a,12a-diazadibenzo-[a,g]fluorenylium derivative. Bioorganic & Medicinal Chemistry Letters, 14, 3627–3630. DOI: 10.1016/j.bmcl.2004.05.029. Madry, L., Galezowska, E., Gluszynska, A., Hermann, T. W., Zabel, M., & Juskowiak, B. (2006). Interactions of the oxidation products of papaverine with guanine quadruplexes. Polish Journal of Chemistry, 80, 921–929. Mergny, J.-L., Riou, J.-F., Mailliet, P., Teulade-Fichou, M.-P., & Gilson, E. (2002). Natural and pharmacological regulation of telomerase. Nucleic Acids Research, 30, 839–865. DOI: 10.1093/nar/30.4.839. Neidle, S. (2009). The structures of quadruplex nucleic acids and their drug complexes. Current Opinion in Structural Biology, 19, 239–250. DOI: 10.1016/j.sbi.2009.04.001. Neidle, S., & Balasubramanian, S. (2006). Quadruplex nucleic acids. Cambridge, UK: RSC Publishing. Neidle, S., & Parkinson, G. (2002). Telomere maintenance as a target for anticancer drug discovery. Nature Reviews Drug Discovery, 1, 383–393. DOI: 10.1038/nrd793. Ragazzon, P. A., Garbett, N. C., & Chaires, J. B. (2007). Competition dialysis: A method for the study of structural selective nucleic acid binding. Methods, 42, 173–182. DOI: 10.1016/j.ymeth.2006.09.010. Rippe, K., Fritsch, V., Westhof, E., & Jovin, T. M. (1992). Alternating d(G-A) sequences form a parallel-stranded DNA homoduplex. The EMBO Journal, 11, 3777–3786.


84


Rubis, B., Kaczmarek, M., Szymanowska, N., Galezowska, E., Czyrski, A., Juskowiak B., Hermann, T., & Rybczynska, M. (2009). The biological activity of G-quadruplex DNA binding papaverine-derived ligand in breast cancer cells. Investigational New Drugs, 27, 289–296. DOI: 10.1007/s10637-0089173-9.

Simonsson, T. (2001). G-Quadruplex DNA structures variations on a theme. Biological Chemistry, 382, 621–628. DOI: 10.1515/BC.2001.073.
