A Biophysical Investigation of DNA-Binding Interactions of Push-Pull Dibenzodioxins and Implications for in Vitro anti-Cancer Activity

Abstract Dibenzodioxin-DNA interactions were studied using biophysical techniques. UV-visible spectral titrations were performed against calf thymus (CT)-DNA to determine the binding constants. Spectrofluorimetric dye-displacement assay and DNA viscometric assay were used to understand the nature of binding interactions. Molecules with intercalative binding preference increased DNA viscosity, similar to a standard intercalator. Dibenzodioxin cytotoxicity was evaluated against HeLa cells. The most active molecules also had the highest binding constants with CT-DNA. Select dibenzodioxins were docked against DNA in silico. The binding behavior was compared using a known DNA binding small molecule. Dibenzodioxins with good cytotoxicity also had high DNA-binding affinity, implying that DNA could be a likely target for these dibenzodioxins. One fluorescent dibenzodioxin was used to study intra-cellular localization in HeLa cells using fluorescence microscopy. This derivative with weaker cytotoxicity was located mainly in the cytoplasm than in the nuclei, again hinting toward dibenzodioxin-DNA-binding interaction as the main mechanism of cytotoxicity.


Introduction
In course of our research into development of planar heteroaromatic push-pull molecules with good emission characteristics coupled with promising in vitro anti-cancer activity, we have developed a number of novel dibenzodioxin motifs over the last five years. [1][2][3][4][5] Figure 1 lists out all the different types of dibenzodioxin molecules prepared thus far including in this present work (compounds 16 and 17; prepared using our established protocols [1][2][3][4][5]. They can be broken down into two broad categoriesthe tricyclic 2,3 and tetracyclic dibenzodioxins. 4,5 In these molecules, the terephthalonitrile ring acts as a powerful electron acceptor because of presence of two electron withdrawing cyano groups, while the central dioxin ring acts as a weak electron donor. Due to stabilization of the lowest unoccupied molecular orbital (LUMO) energy level through such electron push-pull arrangement, red-shifts are produced in the optical spectra and greater fluorescent character also is imparted onto the dibenzodioxins, compared to the parent dibenzodioxin structure. 2,3 These properties are further enhanced in the derivatives with at least one additional electron donor group and more so in the tetracyclic dibenzodioxins, all of which have enhanced charge transfer character in them. Interestingly, the saturated amine containing derivatives (2,4) have more bathochromically shifted optical properties among all the tricyclic derivatives. Another noteworthy point is that cyclic diamino moieties (11, 12, 14 and 15) are the most red-shifted in terms of their photophysical properties than all the others. 4,5 In addition, since the dibezodioxins are known DNA intercalators 6 due to their planar anthracenelike structure and have proven anti-cancer activity against a range of different cell lines, 7,8 we also evaluated the cytotoxicity of some of our molecules (depicted in red in Figure 1) against the wellknown human cervical cancer cell line HeLa and found promising results with IC 50 values in the range of 1.4-10 mM. 3,4 HeLa cell lines provide a reliable platform for evaluation of cytotoxicity of new chemical entities which prompted us to continue our investigations using them, like before. The most cytotoxic compounds were those bearing an ortho-or meta-carboxy ester (compounds 5, 6) reminiscent of the strong bioactivity of dibenzodioxin-1-carboxylic acid and carboxamides. 7,8 Also the ones with an additional electron donor group placed proximal to the electron-withdrawing nitrile groups (compounds 2, 3 and 4) had similar cytotoxicity. The carboxy ester groups in 5 and 6 would interfere with the internal charge transfer (ICT) from the dioxin ring to the nitrile groups, or in case of compounds 2, 3 and 4 the additional single donor group would push extra electron density into one of the nitrile groups, resulting in non-uniform distribution of electronic charge in these molecules, in contrast to structure 1. Clearly there is a link between molecular shape and electronic charge delocalization and the bioactivity of these compounds. We also investigated the DNA-binding properties of two of our dibenzodioxins (1 and 2) and found that they possess a multitude of non-covalent interactions with CT-DNA such as intercalation and groove binding, 9 both of which lent further credence to their observed anti-cancer activity against HeLa cell lines. It must be noted that these types of noncovalent interactions between small molecules and DNA is crucial for the inhibition of the Figure 1. Examples of push-pull tricyclic and tetracyclic dibenzodioxins prepared in our laboratory. Top blue frame depicts the tricyclic ones, while the bottom blue frame shows the tetracyclic ones. Furthermore, the molecules with electron-donating groups in each category are highlighted by the black frame. The molecules in red are the ones for which cytotoxicity studies has been carried out by us earlier, while those in black are the ones whose cytotoxicity has been explored here.
topoisomerase class of enzymes that are needed to uncoil the DNA for functioning of replication pathway, in cancer cells. 10,11 Therefore, in order to fully understand the role of substituents in imparting bioactivity to our dibenzodioxin compounds, we decided to investigate the in vitro cytotoxic action of all the remaining derivatives which had not been previously tested (depicted in black in Figure 1) against HeLa cell lines, and also study the nature of their DNA-binding interactions, which would have a direct imprint on their anti-cancer activity. Figure 2 depicts the results of cytotoxicity study of all the previously untested tricyclic and tetracyclic dibenzodioxins (compounds 7, 8, 9, 12, 13, 16, 17, 18 from Figure 1), performed using a standard MTT based cell proliferation assay against cultured HeLa cell lines. 12 The cells were incubated with 2, 5, 10, 25, and 50 mM doses of these compounds for 24 h duration. Cell viability was measured against control (DMSO solvent). All the eight compounds screened exhibited cytotoxicity against HeLa cells to varying extents. The IC 50 values were estimated by the TREND function in MS Excel. The tricyclic and tetracyclic compounds bearing the -CN, -Br and -NO 2 groups (7, 8, 9 and 16, 17, 18) exhibited poor activity in even at the highest dose of 50 mM. Only compound 8 showed a reduction of viability of HeLa cells down to 44%, with a calculated IC 50 value of 40 mM. However, the results were not statistically significant for any of the doses for compound 8. In case of compounds 7 and 9, the reduction of cell viability observed at 50 mM dose was 51.5% and 53% respectively, based on which it can be inferred that their IC 50 values were marginally higher than 50 mM. As for compounds 16-18, the viability of the cultured and treated HeLa cells was found to be 54%, 71%, 85% respectively at the 50 mM dose. This implied that for all three compounds the predicted IC 50 values were well above the highest does of 50 mM.

Determination of the DNA binding constants and fluorimetric dye-displacement assay
We had previously demonstrated that compounds 1 and 2 exhibited strong binding affinity to CT-DNA through spectrophotometric determination of their intrinsic binding constants (K b ). 9 They also possessed intercalative as well as minor groove binding interactions to variable extents with CT-DNA. 9 A small molecule may exhibit multiple modes of non-covalent interactions with DNA leading to disruption of its secondary structure, which would inhibit DNA replication and eventually promote cell death. The intercalative and minor groove binding interactions are crucial since they can inhibit the activity of the topoisomerase class of enzymes, necessary for DNA replication. Presence of intercalators or minor groove binders on the DNA double helix can interfere with the functioning of these enzymes leading to inhibition of DNA replication. 10,11 Therefore understanding the strength and nature of molecule-DNA interactions would help in unraveling the mechanism of their anti-cancer activity. Table 1 compiles the spectrophotometrically determined intrinsic CT-DNA binding constants (K b ) and binding mode preferences (determined by a fluorimetric dye displacement assay and a viscometric assay) for all the dibenzodioxins depicted in Figure 1. 13 UV visible spectrophotometric titration of dibenzodioxin-DNA adducts were performed, by monitoring the change in absorbance of the compounds at their respective longest wavelengths with addition of incremental amounts of CT DNA to a solution of dibenzodioxins. We observed that in all cases there was a hypochromic effect with increase in DNA concentration, until saturation behavior was reached. Notably, binding of a compound to a DNA via intercalating mode is often reflected through such spectroscopic behavior, while hyperchromicity if observed, would indicate minor groove binding. The K b values were estimated from the Benesi Hildebrand equation (1): Here, [DNA] ¼ concentration of DNA, e f and e a represent the extinction coefficients of free compound without DNA and extinction coefficient of compound bound to DNA respectively. 14 The intrinsic binding constant was calculated from a ratio of the slope [1/(e a -e f )] to intercept [1/K b (e ae f )] of the linear plot and log(K b ). The UV titration plots and Benesi-Hildebrand plots for all the other compounds depicted in Figure 1 are given in the supporting information file ( Figures  S9a-f). A look at Table 1 reveals that the best molecules with very high DNA-binding constants Compound 9.62 [3] 5.38 [9] 140 360 2.57 : 1 0.93 2 1.47 [3] 4.97 [9] 82.5 55 1 : 1.50 0.32 [ [above log(K b ) ¼ 5] were 1, 3, 4, 5, 6 from the tricyclic group and molecules 10, 11, 12, 13, 14, 15 from the tetracyclic group. The fluorimetric dye-displacement method described by Williams et. al. 15 and also effectively utilized previously by us in determining the nature of dibenzodioxin-DNA interactions for compounds 1 and 2, 9 was once again used to understand the binding mode reference of all of our dibenzodioxins. Figure 3 depicts the behavior of dibenzodioxin compound 6 in displacing a known intercalator (EB) 16 and a known minor groove binder (DAPI), 17 from their respective binding sites in CT-DNA. While these two standard dyes are essentially non-fluorescent in nature, upon binding to DNA they undergo a significant fluorescence enhancement at their respective emission maxima of 598 nm and 451 nm. When incremental amounts of a second test molecule (6) are added, it starts to displace the standards (EB and DAPI) from their DNA-binding sites, leading to a loss of emission intensity, as can be seen in Figure 3. Figure 4A compares the concentrations (C 50 ) of compound 6 required to induce 50% reduction of fluorescence of EB and DAPI in their respective CT-DNA adduct solutions. This was obtained by plotting the normalized emissions at 598 nm (EB) and 451 nm (DAPI) from Figure 3 against the added concentrations of dibenzodioxin 6 for each titration. A close look at Figure 4A shows that 6 has some preference toward displacing EB bound to CT-DNA in comparison to displacing DAPI bound to CT-DNA. The C 50 values of compound 6 as revealed from Figure 4A for EB and DAPI are 37.6 mM and 111 mM respectively. The C 50 (G): C 50 (I) ratio of 1: 2.95 for 6, implies a mixed binding behavior with preference for intercalative binding, as is known for dibenzodioxins. Table 1 shows that most molecules have a mixed preference of interaction through both intercalation and groove binding. Among the others molecules, in the tricyclic derivatives molecules 1 and 5 have a slight preference for intercalative binding mode, while molecules 12 and 18 exhibit modest intercalative preference from among the tetracyclic dibenzodioxins.

Viscosity measurement studies
Viscosity (g) of a DNA solution containing an intercalator generally is higher compared to that of a DNA solution containing a groove binder. This is because intercalation of small molecules causes elongation of the DNA and increases its viscosity. Therefore by studying the progressive viscosity changes in a DNA solution in presence of increasing amounts of a ligand molecule, the intercalative or groove binding type behavior of the ligand could be determined. 18,19 In order to generate two standard sets of data for benchmarking the behavior of our compounds, we measured the changes in CT-DNA viscosity with incremental amounts of a known intercalator ethidium bromide (EB) (Dg 1/3 ¼ 1.24) and also with a known minor groove binder (DAPI) (Dg 1/3 ¼ 0.21). Therefore, by comparison of Dg 1/3 values of CT-DNA in presence of our dibenzodioxin molecules with that of EB and DAPI, we were able to understand their preferred DNA binding modes. We also observed a good correlation between the data generated from the viscosity studies and the fluorimetric dye displacement studies to unearth the intercalative versus groove binding nature of our dibenzodioxins. Figure 4B shows a plot of variation of CT-DNA viscosity with increasing amounts of added compounds (EB, DAPI and dibenzodioxin 6). We can see that the trace for 6 is closer to that of EB than DAPI. While the standard intercalator EB produced a Dg 1/ 3 value of 1.24, the standard groove binder DAPI expectedly showed a much smaller value of only 0.21. Gradual addition of varying quantities of 6 to the CT-DNA solution resulted in a Dg 1/3 value of 1.09. The spectrofluorimetric dye displacement titration plots, the corresponding C 50 estimation plots and the CT-DNA viscometric titration plots for all the other compounds from Figure 1 are given in the supporting information file (Figures S10a-f, S11a-c, and S12a-c, respectively). Table 1 lists the final data for each compound.

Comparative in silico docking studies of dibenodioxins 6 and 12 against nuclear stain DAPI
In order to understand better the observed higher cytotoxicity of selected tricyclic dibenzodioxins over the tetracyclic variants and supplement our experimental findings, we decided to carry out in silico rigid body docking experiments of select derivatives against DNA. This is because dibenzodioxins are known to have strong affinity for DNA, as seen from our present biophysical studies and also as understood from the observed nuclear fragmentation induced by dibenzodioxins in HeLa cells, in one of our earlier publications. 3 We aimed to see how well our molecules could fit as a ligand against the DNA macromolecule as a host. Figure 5 depicts the results of the docking studies with molecules 6 (from the tricyclic group) and 12 (from the tetracyclic group) against the host DNA. We chose the well-known nuclear stain DAPI as a standard molecule for sake of comparison with our dibenzodioxins. DAPI is known to fit well inside the minor groove of DNA with high degree of conformational stabilization. We discovered that both molecules 6 and 12 bind to DNA in a pocket similar to DAPI in their respective lowest energy conformations or best poses. Due to presence of the carboxy methyl ester group in 6, it is capable of exhibiting greater interactions with the nucleobases, in contrast to 12. Interestingly, the binding energy of 6 with respect to DNA was À9.62 kcalsmol À1 and for 12 it was À7.82 kcalsmol À1 . This result compares well with the observed higher cytotoxicity of 6 compared to 12 as reflected in their IC 50 values (Table 1). Thus, we could hypothesize that DNA binding affinity of our dibenzodioxins does correspond to the strength of its in vitro bioactivity, when DNA is the main drug target and other mitigating factors are absent.

Microscopic imaging of HeLa cells using green emissive dibenzodioxin 12
Tetracyclic dbenzodioxin compound 12 was earlier found to be emissive in the green region of the visible spectrum with an emission maximum at 548 nm. 5 Figure 6 depicts the fluorescence microscopic image of immobilized HeLa cells incubated with 50 mM dose of 12 and also counterstained with the nuclear stain Hoechest 20 as a control. The images were recorded after incubation of the cells for 24 h with 12 using a green filter for emission from 12 and also a matching filter for Hoechest. A merged image depicted lesser extent of overlap between the green and blue regions, compared to what we had seen earlier for the strongly emissive compound 11. 4 Therefore, the cytoplasmic localization of 12 could be inferred, as evident from the strong green glow emanating from areas outside of the blue glowing Hoechest stained regions. This is in contrast to our earlier observations for compound 11 which showed good overlap between the green and blue regions for compound 11 and the nuclear stain DAPI respectively. 4 One more important difference between the behavior of 11 and 12 is the fact that in case of the former, the cellular emission was successfully observed within 2 hours incubation time, while for 12, 4 there was no observable cellular emission under same conditions. Only upon prolonged incubation time of 24 h with compound 12, the HeLa cells exhibited enhanced autofluorescence as seen in Figure 6. Table 1 lists the IC 50 values of all dibenzodioxins shown in Figure 1, evaluated in in vitro cell proliferation assays against cultured HeLa cell lines; for comparison with their DNA-binding properties. We can see that the compounds with the best IC 50 values indeed have very high DNA-binding constants. The most cytotoxic molecules (3, 2, 4, 6, 14, 5, 10, 15, 13, 11 and 1); also possessed the strongest CT-DNA binding interactions, with log(K b ) values ranging from 4.97 to 6.04. We also noticed that compounds 7-9 and 16-18 with the -CN, -Br and -NO 2 substituents were not necessarily very strongly cytotoxic against HeLa cell lines, with IC 50 values greater than 40 mM. These compounds incidentally also showed the lowest values of log(K b ) 21 (3.10 to 4.69); Table 1. These findings about the relatively poor bioactivity of compounds 7, 8, 9 and 16, 17, 18 reveal that presence of EWGs like -CN, -Br and -NO 2 has a detrimental effect on the cytotoxic action of the dibenzodioxins, irrespective of whether they are tricyclic or tetracyclic. This observation also contrasts with the good cytotoxicity of the compounds 5, 6 and 14, 15, which emphasizes about the importance of the carboxy ester group for its bioactivity. Further studies with tetracyclic derivatives of the parent compound 1 (compounds 12 and 13) also proved this fact as there was no significant improvement in the IC 50 value for these compounds (8.18 mM for 13 and 23.8 mM for 12, compared to the earlier published 9.62 mM for 1). In fact, we did not come across any tetracyclic dibenzodioxin that had a lower IC 50 value than any tricyclic versions. The most cytotoxic tetracyclic dibenzodioxins were compounds 10, 14 and 15 with IC 50 values between 1.8-4.5 mM. Therefore, the we could infer that a shorter total molecular length was preferred for better cytotoxicity, possibly due to a better fit inside the DNA binding pocket. This is also what we observe from our in silico studies when comparing the docking of 6 versus 12 with DNA. Also, no dibenzodioxin compound with poor log(K b ) value, yet exhibiting excellent IC 50 value was found in our list of compounds, implying that interaction with DNA and that through intercalative as well as groove binding mode is main mechanism of action of all the cytotoxic dibenzodioxins depicted in Figure 1. Additionally, majority of the most bioactive dibenzodioxins exhibited a mixed DNA binding behavior having close by C 50 values for both the DAPI and EB displacement ( Table 1). The molecules 6, 5, 15 and 1 showed good intercalative DNA binding, also with smaller extent of groove binding. The viscometric studies also showed a steady increase in CT-DNA viscosities as we compared molecules with similar C 50 values to compounds 6, 5, 15 and 1. For compounds with similar I versus G behavior from dye displacement analysis, their traces in the viscometric titration plots (Figures 12a-c in SI) had a slope that was intermediate between the two traces of EB and DAPI. Hence, we can conclude from the C 50 (G): C 50 (I) ratios and (Dg 1/3 ) values listed in Table 1 for each compound that both the spectrofluorimetric dye displacement studies and viscometric studies, show good correlation between them. Compound 12 is interesting it its bioactivity profile, since the IC 50 value of 23.8 mM against HeLa cells was intermediate between that of compound 6 on one hand and compound 18 on the other, implying only modest activity. A plausible explanation could be found based on the cell imaging results and also the earlier observed weaker DNA-binding affinity of À7.82 kcalsmol À1 compared to 6. When we compare the calculated LogP values 21 for related tetracyclic dibenzodioxin compounds 11, 12, 13, we find that as we increase the alkyl type substitution on the cyclic amines, the lipophilicity also increases [cLogP ¼ 2.32 (11), 2.81 (12) and 2.59 (13)]. Therefore, we could surmise that the cellular uptake of 12 would be slower compared to 11 and 13. Hence, this is reflected in a lower cytotoxicity for compound 12 and also a time delayed cellular autofluorescence, only after a 24 h incubation period. Additionally, the imaging experiment shows lower nuclear penetration by compound 12 in the incubated HeLa cells. Since DNA is the main target of the dibenzodioxin class of anti-tumor drugs, the poorer activity of 12 against cultured HeLa cell lines therefore could be explained on the basis of its greater non-polar character, which prevents it from reaching its target in sufficient doses.

Discussion
Based on our observations regarding the cytotoxicity and DNA-binding patterns of our dibenzodioxins, we have proposed a model structure in Figure 7, reflecting the desirable location of the suitable functional groups. A structure-activity relationship (SAR) can be established from Figure  7 to highlight the key subunits essential for good anti-proliferative activity (IC 50 < 3 mM) in all the different types of dibenzodioxins. This is compiled based on the present data and also the earlier published data for all compounds shown in Figure 1. The best IC 50 values recorded were for the tricyclic compounds 2-6 (in the range of 1.4-3 mM, Table 1), which compare well with other well-known anti-proliferative agents against HeLa cell lines such as 5-fluorouracil, tamoxifen and nedaplatin (2.9 mM, 8.85 mM, 1.76 mM respectively). 22 Here, EDG (electron donating group) could be an alkoxy group or a 2 or 3 amino group. When the EDGs were a cyclic diamine or dioxin ring, this feature did not significantly improve the IC 50 values over that of tricyclic compounds 2-6 ( Table 1, compounds 10, 14 and 15). This fact in combination with the similar IC 50 values as observed for compounds 12 and 13 (when compared to parent dibenzodioxin 1), implies that adding an additional heterocyclic ring in dibenzodioxins is not essential for imparting good bioactivity. Figure 7 also opens up a future possibility of a number of new derivatives for further exploring the bioactivity, i.e., the ones including both the methoxy ester and one EDG on the two peripheral ends of the dibenzodioxin molecule.

Conclusion
In conclusion, we here have investigated the cytotoxic behavior of all different types of dibenzodioxins prepared in our laboratory, using biophysical and cell biology techniques. Based upon their DNA-binding properties measured against CT-DNA, we can infer that all cytotoxic molecules have a tendency for strong interaction with DNA in solution. There was good correlation between data generated by fluorimetric dye-displacement assay and CT-DNA viscometric assay in order to determine the intercalative and or minor groove binding preference for all the compounds. We also found that the tricyclic dibenzodioxins bearing carboxy ester groups or an unsymmetrical push-pull functional group arrangement were the most cytotoxic candidates. Lastly, we were able to implicate the nuclear material of the incubated HeLa cells as the likely target of the cytotoxic dibenzodioxins. This was possible through fluorescence imaging of the cells incubated with one of the more emissive derivatives and a known nuclear stain as a control. This experiment showed that the candidate molecule was mostly localized in the cytoplasm of the HeLa cells which resulted in a weaker bioactivity of this compound, compared to some of the other derivatives. Finally, the utility of the fluorescent nature of our push-pull dibenzodioxins in monitoring intracellular events leading to greater understanding of the mechanism of their anticancer activity was also highlighted.

Materials and methods
All chemicals were reagent grade, purchased from commercial vendors. They were used as purchased without further purifications. Calf thymus DNA (CT-DNA) was obtained from Sigma Aldrich, UV-Vis JASCO V-770 spectrophotometer was employed to check DNA purity (A260: A280 > 1.80) and concentration (e ¼ 6600 M À1 cm À1 at 260 nm). 13 Interactions of the compounds with CT-DNA were studied using solutions of the compound in DMSO and CT-DNA in Tris-HCl buffer (pH 7.2) containing 5 mM Tris-HCl. The buffer solution was prepared with double-distilled water.

Synthesis of dibenzodioxins 16 and 17
Compounds 16 and 17 were prepared using the standard protocols. [1][2][3][4][5] 5.3. (4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay 12 This was carried out to evaluate the anti-proliferative activity against HeLa cells of prepared phenazine molecules (Figure 1 molecule 3 and 4). HeLa cells were plated on 96 well plates at density of 4 Â 10 4 per well and incubated with our compounds in concentrations of 2, 10 and 50 mM. After certain intervals, media was aspirated and fresh media with MTT (5 mg/ml) added. After 4 h, the MTT solution was removed and 100 lL of DMSO was added. Absorbance of the colored solution was measured at 570 nm with a reference at wavelength of 620 nm. The absorbance obtained from compound treated cells was always a fraction of absorbance obtained from untreated cells.

Microscopic cell imaging
HeLa cells were cultured on 6-well plate at a density of 4 Â 10 4 cells per well. On attaining 70-80% density, cells were treated with DMEM-F12 (1:1) (as described in MTT assay) followed by 50 mM compound 12 and incubated for another 24 h. The cells were then fixed with 4% of paraformaldehyde for 10 minutes at RT and nucleus were stained with Hoechest (1 mg/ml). Cells were observed under the fluorescent microscope (Leica DM2000 LED) using appropriate emission filters.

UV-vis measurements
UV absorption spectra recorded on JASCO V-770 UV-visible spectrophotometer at 298 K in the wavelength range of 250-700 nm. Quartz cuvette with 1 cm path length was used.

Fluorescence based competitive dye displacement assay
The fluorescence spectra were measured on JASCO FP-8500 spectrofluorimeter. The excitation wavelengths were 526 nm and 358 nm for EB and DAPI respectively and emission recorded between 400-650 nm. The excitation and emission slit widths were maintained at 2.5 nm each. The fluorescence spectrum of CT-DNA was recorded at 298 K in presence of DAPI and EB in 5 Â 10 À5 M and 3.03 Â 10 À4 M concentrations respectively. Competitive dye displacement assay was carried out by adding compounds 1-4 in different concentrations. The DAPI-CT-DNA complex and EB-CT-DNA complex were titrated with increasing concentration of compounds from 0 to 200 mM range.

Viscosity measurements
To determine the binding mode of compounds, viscosity measurements were performed by keeping the CT-DNA concentration constant (0.5 mM) and varying the concentration of compounds. Viscosity experiments were carried out using a Brookfield DV1 viscometer at 25 C. The data were presented as (g/g 0 ) 1/3 versus [compound]/[DNA] ratio, where g and g 0 are the viscosity of DNA in the presence and absence of compounds respectively. 5.6. In silico docking of select dibenzodioxins with ds-DNA Docking of molecules 6, 12, and known DNA binding agent DAPI with DNA were done using AutoDock 4.2. Structural coordinates for DNA are obtained from 1dnh.pdb. 23 The coordinates for 6, 12, and DAPI were generated using Avogadro 1.2. 24 The binding affinities and the docking poses of the ligands were analyzed with AutoDock Tools. 25 To cover the entire DNA (blind docking), the center of the grid box was chosen to be 14.993 (X), 21.223 (Y), and 8.453 (Z) with 60 points in each dimension with a spacing of 0.375 Å. The search algorithm used for the docking is the Lamarckian Genetic Algorithm with 50 runs using population size of 300 per each run. Maximum number of energy evaluations used was 2500000 and the maximum number of generations was kept at 27000. The remaining parameters used were the default parameters. All the three docking results were analyzed with Autodock tools and the best binding pose, as determined by the cluster having the best binding affinity (kcal/mol), was chosen for further analysis. The interaction profile between the molecule and the DNA was derived manually using Pymol.

Supporting information summary
The details about the synthetic procedures for compounds 16 and 17 with complete spectroscopic characterization of the prepared molecules, UV visible spectrophotometric titration plots with CT-DNA, Benesi-Hildebrand plots, fluorimetric dye-displacement plots, CT-DNA viscometric plots, experimental details about MTT cell proliferation assay and in vitro cell imaging can be found in the Supporting Information (SI) file.