Partial intercalative binding of the food colorant erythrosine to herring sperm DNA

Abstract

Erythrosine (Ery) is an artificial colorant extensively used in the food industry, but may have a potential safety risk. In this study, the characteristics of the interaction in vitro between Ery and herring sperm DNA (hsDNA) were determined by multi-spectroscopic techniques and docking simulations. The multivariate curve resolution-alternating least squares chemometrics approach was used to process the expanded UV-vis spectral data matrix, and both the equilibrium concentration profiles and the pure spectra for the components (Ery, hsDNA and Ery–hsDNA complex) extracted from the highly overlapping composite response were obtained simultaneously to monitor the Ery–hsDNA interaction process. Partial intercalation as the dominant mode of Ery binding to hsDNA was found based on the hypochromism and red shift effect of Ery, stronger double DNA effect and decrease in hsDNA viscosity. The Fourier transform infrared and nuclear magnetic resonance spectra showed that the benzene ring of Ery most likely inserted into the guanine and cytosine bases of hsDNA, and molecular docking predicted and visualized the specific binding. The circular dichroism and DNA cleavage assays suggested that Ery at high concentrations may perturb B-DNA conformation and cause slight DNA cleavage. This study has provided insights into the binding mechanism of Ery with hsDNA and potential hazards of the colorant.

---

1. Introduction

Erythrosine (Ery, Fig. 1), a synthetic food dye with high water solubility, is extensively used in food, such as candies, chocolates, garlic sausages, cake decorating, cocktails and ice creams.¹ This synthetic colorant can effectively improve the appearance and color of food products which may enhance consumer attractiveness. Compared to many natural dyes, Ery has more applications because of its high stability in light, color uniformity, low microbiological contamination and relatively lower production costs.^2,3 However, Ery may have a potential safety risk when it is excessively consumed.^4,5 It was reported that Ery induced DNA damage in the gastrointestinal organs like glandular stomach and colon at a low dose,⁶ and affected thyroid activity due to its high iodine content.⁷ Moreover, Ery showed a high cytotoxicity and cytostaticity with its concentrations at 1, 2, 4 and 8 mM.⁸ Excessive consumption of this dye was considered to have carcinogenic potency.⁹ In the light of these health hazards, the use of this artificial colorant must be strictly limited by laws, regulations and acceptable daily intake values.

Fig. 1 Molecular structure of erythrosine (Ery).

DNA is frequently the main molecular target for harmful chemicals, such as organic dyes, heavy metals, pesticides, etc., which enter into the body by inhalation, ingestion or skin.¹⁰ They can interact with DNA either covalently or non-covalently, which may interfere with transcription and/or DNA replication mechanisms, alter DNA structure and function, and even trigger processes like cell apoptosis and canceration.¹¹ Generally, the small molecules bind to DNA primarily through three modes: intercalation between stacked base pairs, non-covalent groove binding and electrostatic binding with the negatively charged nucleic acid sugar–phosphate structure.¹² In the past few years, more and more researchers investigated the interaction of small molecules with DNA such as drugs, metal complexes and organic dyes as it may provide information on the design and screening of novel and more efficient drugs targeting DNA.^13–15 In recent years, the interactions of synthetic food colorants with DNA have attracted much attention due to the potential hazards of some food colorants. Basu and Kumar¹⁶ reported that carmoisine bound to herring testes DNA via a groove binding mode and this colorant induced moderate conformational changes in the B-form structure of DNA in vitro. Shahabadi et al.¹⁷ found that quinoline yellow strongly competed with Hoechst 33258 for the groove binding site of calf thymus DNA (ctDNA), but it did not cause any significant ctDNA cleavage. Kashanian et al.^18,19 have investigated the interactions of the food dyes sunset yellow and tartrazine with ctDNA, and found that both of them bound to ctDNA via groove binding mode. Based on these results, the authors believed that more attention should be paid to prevent children from consuming large amounts of food containing the two colorants. Ery was reported to interact with ctDNA to form a new product, resulting in hypochromism of Ery in the band at 527 nm and hyperchromism of ctDNA in the band at 260 nm along with blue shift of 2–3 nm.⁸ However, more in-depth studies concerning the binding mode, binding affinity, interaction mechanism and structural change of DNA binding have not been reported, which attracts our interest.

Humans are exposed to different synthetic food additives through the food chain. These chemicals may exhibit their toxicity by interacting with the DNA and affecting DNA structure and function after entering human body. Exploring the binding mechanism of food additives with DNA in vitro is helpful for the assessment of their potential toxicological effects. Therefore, this work was aimed to further investigate the characteristics of binding between Ery and herring sperm DNA (hsDNA) in vitro and the conformational changes of hsDNA binding under simulated physiological conditions (Tris–HCl buffer, pH 7.4) by using multi-spectroscopic methods including fluorescence, UV-vis absorption, circular dichroism (CD), Fourier transform infrared (FT-IR), nuclear magnetic resonance (¹H NMR) spectroscopy, hsDNA melting and viscosity analyses coupled with molecular modeling studies. Moreover, the multivariate curve resolution alternating least squares (MCR-ALS) algorithm, a chemometrics method, was used to analyze the complicated UV-vis absorption spectral data matrix collected from the Ery–hsDNA interaction to understand the kinetic process of the binding reaction. Additionally, the interaction between Ery and supercoiled pUC18 plasmid DNA was studied by gel electrophoresis experiments to further evaluate the cleavage effect of Ery on the DNA structure. This study is expected to provide further understanding on the binding mechanism of Ery with hsDNA and offer useful information on the potential hazard of the food colorant.

2. Materials and methods

2.1. Chemicals

Ery (analytical grade) was obtained from Aladdin Chemical Co. (Shanghai, China), and its stock solution (5.0 × 10⁻³ mol L⁻¹) was prepared in ultrapure water and kept in dark to prevent any light-induced photochemical changes. HsDNA was purchased from Sigma-Aldrich Co. (St. Louis, MO, USA), and its stock solution was obtained by dissolving an appropriate amount of hsDNA in a 0.1 mol L⁻¹ NaCl solution and stored at 4 °C. The purity of the final DNA stock solution was checked by determining the absorption ratio at 260 and 280 nm (A₂₆₀/A₂₈₀), and the ratio gave a value of 1.83, indicating that the hsDNA was free from protein contamination.²⁰ The concentration of hsDNA in stock solution was determined to be 1.28 × 10⁻³ mol L⁻¹ by UV-vis absorption at 260 nm using a molar absorption coefficient ε₂₆₀ = 6600 L mol⁻¹ cm⁻¹.²¹ Supercoiled pUC18 plasmid DNA was purchased from Beijing Solarbio Science and Technology Ltd (Beijing, China). All stock solutions were diluted to the required concentrations with 0.05 mol L⁻¹ Tris–HCl buffer, pH 7.4. All other chemicals were of analytical reagent grade, and ultrapure water was produced by Millipore Simplicity water purification system (Millipore, Molsheim, France) and used throughout the whole experiment.

2.2. Methods

2.2.1. Construction of augmented UV-vis spectral data matrix. Two separate experiments (Experiments 1 and 2) were carried out by monitoring the absorption spectra of Ery–hsDNA interaction with a Shimadzu UV-2450 spectrophotometer. The spectrophotometric changes of Ery or hsDNA solution at a constant concentration upon the addition of different amounts of hsDNA (0 to 5.97 × 10⁻⁵ mol L⁻¹ at an interval of 2.1 × 10⁻⁶ mol L⁻¹, total 29 samples) (Experiment 1) or Ery (0 to 2.75 × 10⁻⁵ mol L⁻¹ in an increment of 9.8 × 10⁻⁷ mol L⁻¹, total 29 samples) (Experiment 2) were monitored. Each sample was mixed thoroughly and allowed to equilibrate for 6 min, and then the UV-vis absorption spectra were recorded over a wavelength range of 200–600 nm at 1 nm interval (a total of 401 wavelengths). Two spectral data matrices, D^Ery (29 × 401) and D^hsDNA (29 × 401) from Experiment 1 and Experiment 2 respectively were obtained and combined to construct an augmented spectral data matrix [D^Ery, D^hsDNA].

2.2.2. Fluorescence and resonance light-scattering (RLS) spectra measurements. All fluorescence spectra were carried out on a spectrofluorimeter (Model F-7000, Hitachi) equipped with a 150 W xenon lamp and a thermostat bath. A 3.0 mL Ery solution (5.0 × 10⁻⁵ mol L⁻¹) was added to a 1.0 cm quartz cell. The hsDNA solution was then gradually titrated to the cell using a micropipette and allowed to stand for 6 min to equilibrate. The fluorescence emission spectra were then measured at three different temperatures (298, 304 and 310 K) in the range of 520–640 nm using 2.5/2.5 nm slit widths. The appropriate blanks corresponding to the buffer solution were subtracted to correct background of fluorescence. Due to no absorption for hsDNA under the conditions of the excitation and emission spectra of Ery, the possible re-absorption and inner filter effect arising from UV-vis absorption of hsDNA were neglected in this work. The fluorescence titration data were analyzed by using the Stern–Volmer plot:²²

F₀/F = 1 + K_SV[Q] (1)

where [Q] is the concentration of hsDNA. The K_SV is the Stern–Volmer quenching constant. F₀ and F represent the fluorescence intensities of Ery in the absence and presence of hsDNA, respectively.

The association constant (K_a) of Ery–hsDNA interaction was calculated by the modified Stern–Volmer equation:²³

where f_a is the fraction of accessible fluorescence, and K_a is the modified Stern–Volmer association constant for the accessible fluorophores, which is equal to the quotient of an ordinate 1/f_a and slope 1/f_aK_a. The dependence of F₀/(F₀ − F) on the reciprocal value of the quencher concentration 1/[Q] is linear.

The preparation of single-strand DNA (ssDNA) was prepared by the method reported previously.²⁴ The quenching effects of ss hsDNA and double-stranded hsDNA (ds hsDNA) were compared by gradually adding ss hsDNA or ds hsDNA solution to Ery solution at a constant concentration of 5.0 × 10⁻⁵ mol L⁻¹. The corresponding fluorescence intensity was recorded, and then the quenching constants were determined.

The RLS spectra were obtained by progressively titrating hsDNA to the fixed amount of Ery, and then synchronous scanning on the spectrofluorometer (Δλ = 0 nm) with the wavelength range of 250–700 nm at room temperature.

2.2.3. Viscosity measurements. Viscometric measurements were carried out by vertical immersing an Ubbelohde viscometer (Φ 0.7–0.8 mm, Shanghai Qianfeng Rubber and Glass Company, Shanghai, China) in an electronic thermostat water bath maintained at 25 ± 0.1 °C, following the common method reported previously.^13,25 A 15.0 mL of the hsDNA solution with a concentration of 5.12 × 10⁻⁵ mol L⁻¹ (hsDNA stock solution was diluted to this concentration by Tris–HCl buffer) was placed in the viscometer, and various concentrations of Ery solution were added directly into the viscometer to obtain increasing Ery/hsDNA molar ratio values, the solutions were mixed thoroughly, and kept a thermal balance by standing for 30 min. Flow times of hsDNA alone and hsDNA mixed with different molar ratios of Ery were measured with an accuracy of ±0.20 s by using a digital stopwatch. The mean values of five replicated measurements subtracted the flow time of buffer solution were used to evaluate the average relative viscosity of the hsDNA and Ery–hsDNA mixtures. Viscosity values were calculated from the observed flow time of hsDNA containing solutions (t) and corrected for buffer solution (t₀), η = (t − t₀)/t₀. The data was presented as (η/η₀)^1/3 versus the molar ratios ([Ery]/[hsDNA]), where η and η₀ are the viscosity contributions of hsDNA in the presence and absence of Ery, respectively.

2.2.4. CD measurements. CD spectra of hsDNA in the absence and presence of varying concentrations of Ery were recorded on a MOS 450 CD spectrometer (Bio-Logic, Claix, France). The CD measurements were performed in Tris–HCl buffer, pH 7.4 under constant nitrogen flush at room temperature. All observed CD spectra were corrected for the buffer signal.

2.2.5. FT-IR spectra. A Thermo Nicolet-5700 spectrometer (Thermo Nicolet Co., USA) equipped with a germanium attenuated total reflection (ATR) accessory, a DTGS KBr detector and a KBr beam splitter were used to measure FT-IR spectra of the solution at different Ery/hsDNA molar ratios following the methods reported previously.^25,26 After background spectra were collected, the infrared spectra were recorded in the wavelength range of 4000–650 cm⁻¹ with a resolution of 4 cm⁻¹ and the sample data were accumulated from 64 scans after a 2 h incubation of Ery with hsDNA. All measurements were carried out at room temperature and relative humidity of 45%.

2.2.6. ¹H NMR spectra. The ¹H NMR spectra were recorded with a Bruker AM-500 nuclear magnetic resonance spectrometer (Bruker, Germany). Deuterium oxide (D₂O, Beijing, 99.8% atom D) was used as the solvent. The solution of Ery was prepared by dissolving the required amount of Ery in D₂O, and its proton chemical shifts were measured in the absence and presence of hsDNA.

2.2.7. Molecular modeling. The molecular docking was performed by the MGL tools 1.5.6 matched with the software package AutoDock version 4.2.²⁷ The docking procedure was simply described in the following way: (i) construction of Ery molecule. The structure of Ery was generated by Sybyl-x 2.0 and subsequently optimized to minimal energy with the help of the Tripos force field using Gasteiger–Huckel charges; (ii) preparation of DNA molecule. The crystal structure of the B-DNA with identifier 425D was downloaded from the Protein Data Bank (http://www.rcsb.org/pdb); (iii) all of the water molecule were removed, and polar hydrogen atoms and Gasteiger charges were added to the macromolecule file; (iv) the grid spacing was set at 0.375 Å with a dimension of 120 Å × 74 Å × 120 Å (X–Y–Z) to encompass the whole DNA molecule. In the docking process, Lamarckian Genetic Algorithm was used as docking parameters algorithm and other miscellaneous parameters were assigned the default values given by AutoDock4.2.

2.2.8. Gel electrophoresis. The cleavage of supercoiled pUC18 plasmid DNA (2 μL, 400 μg mL⁻¹) in the absence and presence of Ery at different concentrations was monitored by using agarose gel electrophoresis (100 V, 1.0 h, 1× Tris–acetic acid–EDTA). The samples of DNA and the Ery–DNA complex were incubated for 12 h at 37 °C. Subsequently, a loading buffer (2 μL, 0.03% bromophenol blue, 0.03% xylene cyanol and 30% glycerol) was added and mixed adequately, and the mixtures were carefully loaded onto the sample wells. After electrophoresis, the agarose gel (1.0%) was stained by goldview, and then visualized and photographed using an UV illuminator.

2.2.9. MCR-ALS algorithm. Interactions of chemicals with DNA often cause the changes of absorption spectra, which may conveniently help us to analyze the binding properties and process. However, when the response signals of components in reaction system overlap with each other, it is usually difficult to distinguish the existing species and judge binding process. As a popular chemometrics method, MCR-ALS algorithm may overcome this limitation.²⁸ Without any previous knowledge or model about the studied process, MCR-ALS can successfully distinguish multiple component responses in unresolved mixtures, and output provides the concentration profiles and pure spectra of all the components.²⁹

The multivariate calibration algorithm MCR-ALS has been described in detail previously.^30–32 Therefore, a short specification was given here. The bilinear decomposition of the augmented matrix D is performed according to the following expression:

D = C × S^T + E (3)

where the rows of D contain the spectra (s) measured and corresponding the wavelengths (w), the matrix C contains the relative concentrations of the spectroscopic active species (f), the matrix of S^T relates the pure spectra with the f species in the samples, and E is the residual matrix with the data variance and not fitted by the model.³³

The related bilinear model for MCR-ALS analysis could be expressed by the following equation:

where [D^Ery, D^hsDNA] is the new augmented data matrix obtained from above absorption titrations experiments (Experiments 1 and 2, as described in Section 2.2.1), respectively. Then, this data matrix, D was imported into Matlab 6.5 programs, and subjected to MCR-ALS analysis to obtain the concentration profiles and pure spectra of all molecular components. Thus, a single concentration profile matrix, C, and a row-wise augmented matrix of pure spectra, S^T, was resolved by the MCR-ALS algorithm.

3. Results and discussion

3.1. Spectral decomposition: application of MCR-ALS

As shown in Fig. 2A, the UV-vis absorption spectra of Ery exhibited three absorption bands at 526, 308 and 261 nm respectively. When hsDNA was added to the solution, the absorption intensity of Ery at 526 nm decreased gradually accompanied by a remarkable increase at 261 nm (corresponding to curves from 1 to 29), and an isobestic point at 308 nm occurred. On the other hand, the maximum absorption peak of hsDNA at 261 nm (curve 1) increased gradually and two new absorption bands at 310 and 526 nm (Ery characteristic absorbance peak) were observed with the continuous addition of Ery (curves 1 to 29, Fig. 2B). Although the isobestic point at 308 nm and hypochromism effect at 526 nm were the evidence for the formation of Ery–hsDNA complex (Fig. 2A), the acquired UV-vis spectral profiles of Ery and hsDNA highly overlapped, and the accurate information including the concentration profiles of the components (Ery, hsDNA and Ery–hsDNA complex) as well as their spectral profiles could hardly be obtained by this conventional UV-vis spectrophotometry.

Fig. 2 UV-vis absorption spectra and the results analyzed by MCR-ALS. (A) Experiment 1: the UV-vis absorption spectra of Ery in the presence of varying concentrations of hsDNA. c(Ery) = 2.36 × 10⁻⁵ mol L⁻¹, and c(hsDNA) = 0, 0.21, 0.42, 0.64, …, 5.67, 5.88 × 10⁻⁵ mol L⁻¹ for curves 1 → 29, respectively. (B) Experiment 2: the UV-vis absorption spectra of hsDNA in the presence of Ery at different concentrations. c(hsDNA) = 2.31 × 10⁻⁵ mol L⁻¹, and c(Ery) = 0, 0.98, 1.96, 2.94, …, 26.46, 27.44 × 10⁻⁶ mol L⁻¹ for curves 1 → 29, respectively. Equilibrium concentrations of hsDNA, Ery and Ery–hsDNA complex resolved by MCR-ALS model for Experiment 1 (C) and Experiment 2 (D), respectively. (E) Extracted UV-vis absorption spectra for hsDNA, Ery and Ery–hsDNA complex by MCR-ALS method. Solid line: resolved spectra by MCR-ALS; dashed line: measured spectra.

Considering the complicated spectral characteristics of the interaction between Ery and hsDNA, a chemometrics method, MCR-ALS was used to interpret the augmented spectroscopic data matrix [D^Ery, D^ctDNA]. As previously described in Section 2.2.1, [D^Ery, D^ctDNA] was submitted for the purpose of extracting information including the number of significant factors, f and the relative concentrations of the various reactant and product species. The extracted first four eigenvalues by singular value decomposition (SVD) method were 121.12, 60.34, 10.36 and 2.48, implying that there were three significant factors, i.e., Ery, hsDNA and Ery–hsDNA complex in the reaction system.

The concentration profiles and spectral profiles of the components (Ery, hsDNA and Ery–hsDNA complex) were extracted by MCR-ALS approach (Fig. 2C–E). A gradual decrease in Ery concentration and increase in the concentration of Ery–hsDNA complex were observed upon the addition of hsDNA (Fig. 2C). Conversely, with the addition of Ery, the concentration of hsDNA decreased, accompanied by an increase in the concentration of Ery–hsDNA complex (Fig. 2D). The extracted spectra (solid line) of Ery and hsDNA by MCR-ALS were consistent with the measured spectra (dashed line, Fig. 2E), indicating that the concentration profiles of the three reacting components were resolved correctly.³² These results provided qualitative and quantitative evidence for the interaction between Ery and hsDNA associated with the formation of the Ery–hsDNA complex.

3.2. RLS spectra characteristics of Ery–hsDNA system

As shown in Fig. 3A, Ery (curve 1) and hsDNA (dashed curve) showed weak RLS signals under the scanning range from 250 to 700 nm. However, with an increase in the concentration of hsDNA, the RLS intensity and the peak shape of Ery were remarkably changed, accompanied by the appearance of a new peak at around 485 nm. The RLS signals are related to the formed aggregate of the particle dimension in the solution.^34,35 Generally, when the diameter of a particles increases, the light scattering signal should be dramatically enhanced. Thus, it can be inferred from the results that a new Ery–hsDNA complex was formed after mixing hsDNA with Ery in solution due to an increased light scattering signal. This result further supported the result of MCR-ALS analysis.

Fig. 3 (A) RLS spectra of the Ery–hsDNA system at pH 7.4 and room temperature. c(Ery) = 1.0 × 10⁻⁶ mol L⁻¹, and c(hsDNA) = 0, 0.42, 0.85, 1.28, 1.72 and 2.13 × 10⁻⁵ mol L⁻¹ for curves 1 → 6, respectively. Dashed line shows the RLS spectrum of hsDNA only, c(hsDNA) = 1.0 × 10⁻⁶ mol L⁻¹. (B) Absorption spectral titration of Ery (2.36 × 10⁻⁵ mol L⁻¹, curve 1) with increasing concentrations of hsDNA (curves 2 → 16), c(hsDNA) = 0.43, 0.85, 1.28, 1.71, …, 5.98 and 6.40 × 10⁻⁵ mol L⁻¹. (C) Fluorescence quenching plots of Ery by ds hsDNA and ss hsDNA at pH 7.4. c(Ery) = 5.0 × 10⁻⁵ mol L⁻¹, and c(hsDNA) = 0, 0.84, 1.66, 2.46, 3.24, 4.00, 4.74, 5.46, 6.17, 6.86 and 7.53 × 10⁻⁵ mol L⁻¹.

Previous studies have shown that tamoxifen³⁶ and oridonin³⁷ bound to ctDNA via an intercalation mode, resulting in a dramatical increase of RLS intensity. Similarly, it was reasonable to propose that the increase of the RLS signal of Ery may be due to the intercalation into hsDNA, inducing the formation of a new complex with large particle size.

3.3. UV-vis absorption spectroscopic characteristics

Due to a strong stacking interaction between the aromatic chromophore of the ligand and DNA base pairs, a large hypochromism and bathochromism or red-shift in the absorption spectra of ligand to DNA can be observed in the intercalation.^38,39 However, groove and electrostatic interaction do not show these changes in absorption spectroscopic characteristics.⁴⁰ Hypochromism and a significant red shift of the spectra of Ery at 526 nm were observed upon the addition of increasing concentrations of hsDNA to the Ery solution (Fig. 3B). The spectral characteristics indicated that a strong interaction in the molecular stack between the aromatic chromophore of Ery and the base pairs of hsDNA may occur due to the intercalative binding of Ery to hsDNA double-helix.^38,40

3.4. Effects of Ery on ssDNA and dsDNA

If the binding mode is intercalation, the quenching effect of ssDNA should be weaker than that of dsDNA for the release of the double strands of DNA. By contrast, the groove mode may lead to a greater quenching effect of ssDNA than dsDNA.⁴¹ As shown in Fig. 3C, ds hsDNA exhibited a higher slope than ss hsDNA, the corresponding K_SV values for ds hsDNA and ss hsDNA were determined to be 1.66 × 10⁴ and 1.10 × 10⁴ L mol⁻¹, respectively. The quenching effect of ds hsDNA was greater than that of ss hsDNA, which supported the intercalative binding between Ery and hsDNA.

3.5. Viscosity

Viscosity was measured to further confirm the nature of the binding between Ery and hsDNA. A classical intercalation causes the viscosity increase of DNA as the space of adjacent base pairs needs to be large enough to accommodate the bound ligand,⁴² while a partial or non-classical intercalation ligand could bend (or kink) the DNA helix, typically reducing its effective length and concomitantly its viscosity.⁴³ As shown in Fig. 4, the relative viscosity of hsDNA obviously decreased with the successive addition of Ery, indicating that Ery molecules bound to hsDNA through a partial intercalation.

Fig. 4 Effect of increasing amounts of Ery on the relative viscosity of hsDNA at 298 K. c(hsDNA) = 5.12 × 10⁻⁵ mol L⁻¹.

3.6. FT-IR studies

The spectral features of free hsDNA and Ery–hsDNA complex were shown in Fig. 5A. The bands at position 1713, 1664, 1609 and 1489 cm⁻¹ primarily corresponded to bases stretching vibrations of guanine (G), thymine (T), adenine (A) and cytosine (C), respectively. The bands located at 1222 and 1089 cm⁻¹ are due to the asymmetric and symmetric phosphate stretching vibrations, respectively. The plots of the relative intensity (R) of several peaks of DNA in-plane vibrations related to A–T, G–C base pairs and the PO₂ stretching vibrations at different Ery concentrations were obtained after peak normalization using R_i = I_i/I₉₆₈, where I_i is the intensity of the absorption peak for pure DNA in the complex with i as ligand concentration, and I₉₆₈ is the intensity of the 968 cm⁻¹ peak (DNA internal reference).⁴⁴ The FT-IR spectra and relative intensity of hsDNA were changed by different [Ery]/[hsDNA] ratios, which offered valuable information concerning the binding sites of Ery with hsDNA.

Fig. 5 (A) FT-IR spectra and difference spectra [(hsDNA + Ery) solution − Ery solution] in the region of 1800–800 cm⁻¹ for free hsDNA and Ery–hsDNA complex in an aqueous solution. (B) Intensity ratio variations of several ctDNA vibrations as a function of different Ery/hsDNA molar ratios. c(hsDNA) = 1.28 × 10⁻³ mol L⁻¹.

The peak of guanine at position 1713 cm⁻¹ shifted downward to 1702 cm⁻¹ (by 11 cm⁻¹) with increasing concentrations of Ery. Simultaneously, cytosine band at 1489 cm⁻¹ was shifted upward to 1496 cm⁻¹ by Ery–hsDNA interaction (Fig. 5A). The position changes of these bands were also accompanied by the decreases in their vibrational intensity (Fig. 5B). Compared to guanine and cytosine bands, both thymine band at 1664 cm⁻¹ and adenine band at 1609 cm⁻¹ as well as the asymmetric and symmetric phosphate stretching vibrations located at 1222 and 1089 cm⁻¹ did not show significant spectral shifting. These spectral changes suggested that Ery molecules might primarily bind to the guanine and cytosine bases of hsDNA.²⁶ Small increases in vibrational intensity of the phosphate stretching peaks (both asymmetric and symmetric) in the presence of Ery at high concentrations were observed, suggesting non-interaction between Ery and phosphate group of hsDNA backbone, but the existence of Ery may result in some perturbation to the phosphate skeleton.²⁶

3.7. ¹H NMR characteristics

Additional evidence for the interaction of Ery with hsDNA was obtained by ¹H NMR study, an effective means of obtaining the binding mode of the interaction system. As shown in Fig. 6A, the ¹H NMR spectrum of Ery exhibited five peaks (6.870–6.882 ppm for H^b site, 7.482–7.484 ppm for H^c site, 7.570–7.571 ppm for H^d site, 7.626 ppm for H^a sites, and 7.715–7.727 for H^e, Table 1). Chemical shifts of the hydrogen atoms were observed after hsDNA was added into Ery solution. The hydrogen atoms (b) and (e) moved to upfield with the maximum chemical shift values of 0.021 and 0.012, while the hydrogen atoms (c) and (d) shifted to downfield with a new splitting peak, respectively. However, the hydrogen atoms (a) did not exhibit obvious chemical shift. Evidently, the chemical shift of proton resonances of benzene ring with carboxyl sodium exhibited a significant change, indicating that the benzene ring did insert into the base pairs of hsDNA, thus there was the magnetic shielding effect in Ery–hsDNA complex due to the changes of electron density in the surroundings of the ring.⁴⁵

Fig. 6 (A) ¹H NMR spectra of Ery (red) and Ery–hsDNA system (blue). (B) Molecular modeling results of the energy-minimized structure of Ery–DNA system. The green dashed lines stand for hydrogen bond.

Table 1 Chemical shifts (ppm) of protons of Ery in the absence and presence of hsDNA^a

System	H^a	H^b	H^c	H^d	H^e
a H^a, H^b, H^c, H^d and H^e represent the different hydrogen atoms of Ery corresponding to Fig. 1.
Ery	7.626	6.87	7.482	7.57	7.715
		6.882	7.484	7.571	7.727
			7.465	7.55
Ery–hsDNA	7.62	6.891	7.477	7.56	7.723
		6.904	7.49	7.563	7.739

---

3.8. Computational modeling of Ery–hsDNA complex

As it can provide the visual representation, and even predict the probable binding characteristics of ligands to DNA, molecular docking has been widely used to investigate the binding interaction of small molecules with DNA. The molecular docking of Ery with DNA was carried out to further clarify the binding mode of Ery with DNA and the steady conformation of Ery–hsDNA complex. After the 100 docking runs were performed, the energetically most favorable conformation of the docked pose was chosen (Fig. 6B). The benzene ring part of Ery nearly vertically intercalated into hydrophobic environment of hsDNA from the small groove, and surrounded by four bases (G23, C24, G4 and G5), suggesting the existence of hydrophobic interaction and π–π interaction between Ery and hsDNA.⁴⁶ Moreover, two hydrogen bonds were formed between the oxygen atoms O-1 and O-2 (Fig. 1) of Ery and the H-22 hydrogen atoms associated with N-2 of G4 and G5 of A chain, respectively. These results indicated that hydrogen bonds may play a major role in the stability of Ery–DNA structure.

3.9. CD spectroscopy

CD spectroscopy is deemed as an accurate and sensitive approach to monitor the conformational change of DNA.⁴⁷ As shown in Fig. 7A, the CD spectra of hsDNA exhibited a negative band at 245 nm and a positive band at 275 nm in the wavelength range of 220–320 nm, which was a signal characteristic of a typical B-DNA structure.⁴⁸ The two bands are extremely sensitive in detecting DNA interaction with small molecules, and the changes of these bands often attribute to the corresponding changes in the DNA structure. The decreases in CD signals of hsDNA at both 245 and 275 nm in the presence of Ery were observed (shifting to zero level), suggesting that hsDNA structure was changed by intercalation of Ery into the base pairs of hsDNA and subsequent reduction of its base stacking and loose of the right-handed helicity.⁴⁹ These results were consistent with the previous reports for other intercalators such as prodigiosin⁵⁰ and perfluoroalkyl acids.⁵¹

Fig. 7 (A) Circular dichroism spectra of hsDNA in the presence of increasing amounts of Ery at pH 7.4. c(hsDNA) = 5.0 × 10⁻⁴ mol L⁻¹. The molar ratios of Ery to hsDNA were 0 : 1, 1 : 1 and 2 : 1 for curves 1 → 3, respectively. c(Ery) = 5.0 × 10⁻⁴ mol L⁻¹ (dashed line). (B) Gel electrophoresis of plasmid pUC18 supercoiled DNA in the presence of increasing amounts of Ery after 24 h of incubation at 37 °C. Lanes 1, supercoiled plasmid DNA (control); 2 → 8, Ery + plasmid DNA, and c(Ery) = 0.25, 0.50, 0.75, 1.00, 1.25, 1.50, and 1.75 × 10⁻⁴ mol L⁻¹, respectively.

3.10. DNA cleavage

Supercoiled pUC18 plasmid DNA was used to determine the DNA cleavage activity of Ery by agarose gel electrophoresis. When circular plasmid DNA is run on horizontal gel using electrophoresis, supercoiled form (Form I) will be observed in the fastest migration. While one strand or even both strands are cleaved, the supercoils will relax to become a slower-moving open circular form (Form II), and a linear form (Form III) that migrates in between Form I and Form II, respectively.⁵² As shown in Fig. 7B, at low Ery concentrations (0 to 1.00 × 10⁻⁴ mol L⁻¹), Form I was only found in Lanes 1 to 5. When Ery concentration was increased (1.25 to 1.75 × 10⁻⁴ mol L⁻¹), a mild form II was found in Lanes 6 to 8, suggesting that Ery showed a slight cleavage activity of plasmid DNA at relatively higher concentrations.

3.11. Fluorescence quenching mechanism and binding constants

As shown in Fig. 8A, with increasing amounts of hsDNA, the fluorescence emission peak of Ery at 555 nm decreased gradually without apparent shift, indicating that hsDNA could interact with Ery. The fluorophore quenching results may attribute to static and dynamic quenching mechanisms. The static quenching is caused by the formation of a ground state complex, while the dynamic quenching is due to the collision between the fluorophore and the quencher.⁵³ Static and dynamic quenching can be distinguished by their different dependence on temperature. If the quenching mechanism is static, the quenching constants would decrease with increasing temperature, while the reverse effect is observed for dynamic quenching.⁵⁴

Fig. 8 (A) Effect of hsDNA on fluorescence spectra of Ery (T = 298 K, pH 7.4, λ_ex = 510 nm, λ_em = 555 nm); c(Ery) = 5.0 × 10⁻⁵ mol L⁻¹, and c(hsDNA) = 0, 0.84, 1.66, 2.46, 3.24, 4.00, 4.74, 5.46, 6.17, 6.86 and 7.53 × 10⁻⁵ mol L⁻¹, for curves 1 → 11, respectively. The Stern–Volmer plots (B) and the modified Stern–Volmer plots (C) for the fluorescence quenching of Ery by hsDNA at different temperatures.

The fluorescence data at different temperatures (298, 304 and 310 K) were analyzed by the Stern–Volmer plots to determine possible quenching mechanism. The plots showed a good linearity (Fig. 8B), suggesting that the predominance of only one type of quenching process occurred, either static or dynamic quenching. The K_SV values for the complexation between Ery and hsDNA decreased with rising temperature (Table 2), indicating that the quenching process was primarily due to complex formation viz., static quenching.⁵⁵

Table 2 The quenching constants (K_SV), association constants (K_a) and thermodynamic parameters for the interaction of Ery with hsDNA at three temperatures^a

T (K)	KSV (10^4 L mol^−1)	R^a	Ka (10^4 L mol^−1)	R^b	ΔH° (kJ mol^−1)	ΔS° (J mol^−1 K^−1)	ΔG° (kJ mol^−1)	R^c
a R^a is the correlation coefficient for the KSV values. R^b is the correlation coefficient for the Ka values. R^c is the correlation coefficient for the ΔH° values.
298	1.65	0.9969	2.14	0.9988	−27.65	−9.93	−24.69
304	1.27	0.9993	1.67	0.9985			−24.63	0.9874
310	1.01	0.9984	1.39	0.9983			−24.57

---

The spectral data obtained from fluorescence titrations were used to construct modified Stern–Volmer plots to determine the K_a value of Ery–hsDNA (Fig. 8C). As shown in Table 2, the values of K_a at different temperatures were in the order of 10⁴ L mol⁻¹ and tended to decrease with increasing temperature. These results were due to reduction of the Ery–hsDNA complex stability. In addition, the values of K_a were similar to those of other intercalative additives, such as indigo carmine,¹⁰ sunset yellow,¹⁸ butylated hydroxyanisole²⁵ and 2-tert-butyl-4-methylphenol.⁵⁶

3.12. Thermodynamic parameters and binding forces

The acting forces between small molecules and macromolecules mainly include hydrogen bonds, van der Waals force, electrostatic force and hydrophobic interaction.⁵⁷ The signs and magnitude of change in enthalpy (ΔH°), entropy (ΔS°) and free energy (ΔG°) for ligand–DNA interaction can be used to evaluate the main forces contributing to the formation of a ligand–DNA complex. According to the well known van't Hoff equation, the values of ΔH° and ΔS° were obtained from the slope and intercept of a linear van't Hoff plot between log K_a versus 1/T:where R is the gas constant (8.314 J mol⁻¹ K⁻¹). Knowing ΔH° and ΔS° values, the free energy change (ΔG°) can be calculated from the Gibbs–Helmholtz equation:

ΔG° = ΔH° − TΔS° (6)

It is reported that when ΔH° < 0 and ΔS° > 0, the main force is electrostatic force; when ΔH° < 0 and ΔS° < 0, the main force is attributed to van der Waals and hydrogen bond, and when ΔH° > 0 and ΔS° > 0, the main force is regarded as hydrophobic interaction.⁵⁸ As shown in Table 2, the values of ΔH° and ΔS° of Ery–hsDNA interaction were negative, suggesting that van der Waals and hydrogen bond played a major role in the binding of Ery to hsDNA and contributed to the stability of the complex,⁵⁹ and the binding reaction was spontaneous due to the negative free energy change (ΔG° < 0).

4. Conclusions

The current work investigated the binding affinity, mode of interaction, structural aspects of the Ery–hsDNA interaction through multifaceted spectroscopy, molecular modeling and agarose gel electrophoresis. A partial intercalative binding model for Ery–hsDNA complexation was supported by the emergence of hypochromism and significant bathochromism for Ery absorption, remarkable enhancement in RLS signals of Ery upon complexation with DNA, less quenching effect by denatured hsDNA and decrease in the relative viscosity of hsDNA. The benzene ring of Ery most likely insert into the G–C bases of hsDNA and the theoretical modeling predicted and visualized the preferential docking position of Ery in hsDNA. Ery induced moderate conformational perturbations in the secondary structure of hsDNA, and this dye could lead to slight cleavage of the plasmid DNA. Moreover, the MCR-ALS algorithm was used to resolve the expanded UV-vis spectral data matrix, and the spectra of Ery, hsDNA and especially the Ery–hsDNA complex along with their concentration profiles were extracted successfully to determine the progress of Ery interaction with hsDNA. These results suggested that more attention should be paid to re-evaluate Ery as a food additive.

Acknowledgements

This study is financially supported by the National Natural Science Foundation of China (No. 31460422, 21167013 and 31060210), the Joint Specialized Research Fund for the Doctoral Program of Higher Education (20123601110005), the Program of Jiangxi Provincial Department of Science and Technology (20141BBG70092), the Natural Science Foundation of Jiangxi Province (20142BAB204001 and 20143ACB20006), and the Research Program of State Key Laboratory of Food Science and Technology of Nanchang University (SKLF-ZZA-201302), which are gratefully acknowledged.

References

1. A. Mittal, J. Mittal, L. Kurup and A. K. Singh, J. Hazard. Mater., 2006, 138, 95–105.
2. M. Gómez, V. Arancibia, C. Rojas and E. Nagles, Int. J. Electrochem. Sci., 2012, 7, 7493–7502.
3. S. P. Alves, D. M. Brum, É. C. B. de Andrade and A. D. P. Netto, Food Chem., 2008, 107, 489–496.
4. B. Zhang, D. L. Du, Y. M. Yin, L. Zheng, J. H. Zhao, S. A. Eremin, V. B. Rybakov, M. Meng and R. M. Xi, Food Anal. Methods, 2014, 7, 1798–1803.
5. F. M. D. Chequer, V. de Paula Venâncio, M. D. L. P. Bianchi and L. M. G. Antunes, Food Chem. Toxicol., 2012, 50, 3447–3451.
6. L. Pereira, B. Ali, K. Mohite, P. Arora and C. V. Rao, J. Environ. Biol., 2000, 21, 309–315.
7. V. K. Gupta, A. Mittal, L. Kurup and J. Mittal, J. Colloid Interface Sci., 2006, 304, 52–57.
8. P. Mpountoukas, A. Pantazaki, E. Kostareli, P. Christodoulou, D. Kareli, S. Poliliou, C. Mourelatos, V. Lambropoulou and T. Lialiaris, Food Chem. Toxicol., 2010, 48, 2934–2944.
9. Y. F. Sasaki, S. Kawaguchi, A. Kamaya, M. Ohshita, K. Kabasawa, K. Iwama, K. Taniguchi and S. Tsuda, Mutat. Res., Genet. Toxicol. Environ. Mutagen., 2002, 519, 103–119.
10. Y. D. Ma, G. W. Zhang and J. H. Pan, J. Agric. Food Chem., 2012, 60, 10867–10875.
11. R. C. Gupta and G. Spencer–Beach, Regul. Toxicol. Pharmacol., 1996, 23, 14–21.
12. D. Sarkar, P. Das, S. Basak and N. Chattopadhyay, J. Phys. Chem. B, 2008, 112, 9243–9249.
13. S. Y. Bi, L. L. Yan, Y. Wang, B. Pang and T. J. Wang, J. Lumin., 2012, 132, 2355–2360.
14. I. Saha and G. S. Kumar, Dyes Pigm., 2013, 96, 81–91.
15. A. Tarushi, J. Kljun, I. Turel, A. A. Pantazaki, G. Psomas and D. P. Kessissoglou, New J. Chem., 2013, 37, 342–355.
16. A. Basu and G. S. Kumar, J. Agric. Food Chem., 2013, 62, 317–326.
17. N. Shahabadi and M. Maghsudi, Dyes Pigm., 2013, 96, 377–382.
18. S. Kashanian, S. H. Zeidali, K. Omidfar and N. Shahabadi, Mol. Biol. Rep., 2012, 39, 10045–10051.
19. S. Kashanian and S. H. Zeidali, DNA Cell Biol., 2011, 30, 499–505.
20. S. Ghosh, P. Kundu, B. K. Paul and N. Chattopadhyay, RSC Adv., 2014, 4, 63549–63558.
21. H. Joshi, A. Sengupta, K. Gavvala and P. Hazra, RSC Adv., 2014, 4, 1015–1024.
22. M. A. Husain, S. U. Rehman, H. M. Ishqi, T. Sarwar and M. Tabish, RSC Adv., 2015, 5, 64335–64345.
23. M. Yang, J. Jin, G. Xu, F. Cui and H. Luo, Chem. Phys., 2014, 428, 1–5.
24. S. Kashanian, S. Javanmardi, A. Chitsazan, M. Paknejad and K. Omidfar, J. Photochem. Photobiol., B, 2012, 113, 1–6.
25. G. W. Zhang, L. H. Wang, X. Y. Zhou, Y. Li and D. Gong, J. Agric. Food Chem., 2014, 62, 991–1000.
26. Y. Zhang, G. W. Zhang, Y. Li and Y. T. Hu, J. Agric. Food Chem., 2013, 61, 2638–2647.
27. A. Nematollahi, W. B. Church, N. A. Nadvi, M. D. Gorrell and G. Sun, Cent. Nerv. Syst. Agents Med. Chem., 2014, 14, 2–9.
28. M. Garrido, F. X. Rius and M. S. Larrechi, Anal. Bioanal. Chem., 2008, 390, 2059–2066.
29. J. Jaumot, R. Gargallo, A. de Juan and R. Tauler, Chemom. Intell. Lab. Syst., 2005, 76, 101–110.
30. J. Ghasemi, S. Ahmadi, A. I. Ahmad and S. Ghobadi, Appl. Biochem. Biotechnol., 2008, 149, 9–22.
31. M. Bosco, M. P. Callao and M. S. Larrechi, Talanta, 2007, 72, 800–807.
32. Y. N. Ni, M. Wei and S. Kokot, Int. J. Biol. Macromol., 2011, 49, 622–628.
33. M. Vives, R. Gargallo and R. Tauler, Anal. Chim. Acta, 2000, 424, 105–114.
34. R. F. Pasternack, C. Bustamante, P. J. Collings, A. Giannetto and E. J. Gibbs, J. Am. Chem. Soc., 1993, 115, 5393–5399.
35. T. T. Zhao, S. Y. Bi, Y. Wang, T. J. Wang, B. Pang and T. T. Gu, Spectrochim. Acta, Part A, 2014, 132, 198–204.
36. C. Q. Cai, X. M. Chen and F. Ge, Spectrochim. Acta, Part A, 2010, 76, 202–206.
37. Z. G. Chen, Z. Wang, J. H. Chen, X. Chen, J. H. Wu, Y. Y. Wu and J. Y. Liang, Eur. J. Med. Chem., 2013, 66, 380–387.
38. P. Paul and G. S. Kumar, Spectrochim. Acta, Part A, 2013, 107, 303–310.
39. H. Chao, W. J. Mei, Q. W. Huang and L. N. Ji, J. Inorg. Biochem., 2002, 92, 165–170.
40. M. Sirajuddin, S. Ali and A. Badshah, J. Photochem. Photobiol., B, 2013, 124, 1–19.
41. S. Kashanian, M. M. Khodaei, H. Roshanfekr, N. Shahabadi and G. Mansouri, DNA binding, Spectrochim. Acta, Part A, 2012, 86, 351–359.
42. N. Shahabadi, N. Fatahi, M. Mahdavi, Z. K. Nejad and M. Poirfoulad, Spectrochim. Acta, Part A, 2011, 83, 420–424.
43. J. H. Shi, J. Chen, J. Wang and Y. Y. Zhu, Spectrochim. Acta, Part A, 2015, 136, 443–450.
44. D. K. Jangir, G. Tyagi, R. Mehrotra and S. Kundu, J. Mol. Struct., 2010, 969, 126–129.
45. F. Wang, J. H. Yang, X. Wu, F. Wang and H. H. Ding, Spectrochim. Acta, Part A, 2007, 67, 385–390.
46. P. Deepa, P. Kolandaivel and K. Senthilkumar, Mater. Sci. Eng., C, 2012, 32, 423–431.
47. N. Shahabadi, M. Falsafi and N. H. Moghadam, J. Photochem. Photobiol., B, 2013, 122, 45–51.
48. A. Subastri, C. H. Ramamurthy, A. Suyavaran, R. Mareeswaran, P. L. Rao, M. Harikrishna, M. S. Kumar, V. Sujatha and C. Thirunavukkarasu, Int. J. Biol. Macromol., 2015, 78, 122–129.
49. X. Y. Xu, D. D. Wang, X. J. Sun, S. Y. Zeng, L. W. Li and D. Z. Sun, Thermochim. Acta, 2009, 493, 30–36.
50. L. Han, Y. H. Zhou, X. Q. Huang, M. S. Xiao, L. Zhou, J. H. Zhou, A. H. Wang and J. Shen, Spectrochim. Acta, Part A, 2014, 123, 497–502.
51. J. Cao, Y. Wei and Y. Cheng, Environ. Sci. Pollut. Res. Int., 2013, 20, 8355–8363.
52. G. W. Zhang, L. H. Wang, X. Y. Zhou, Y. Li and D. Gong, J. Agric. Food Chem., 2014, 62, 991–1000.
53. S. R. Feroz, S. B. Mohamad, N. Bujang, S. N. A. Malek and S. Tayyab, J. Agric. Food Chem., 2012, 60, 5899–5908.
54. J. R. Lakowicz, Principles of fluorescence spectroscopy, Springer Publications, New York, 3rd edn, 2006.
55. Y. Sameena and I. V. Enoch, J. Lumin., 2013, 138, 105–116.
56. S. Kashanian and J. E. N. Dolatabadi, Eur. Food Res. Technol., 2010, 230, 821–825.
57. T. Sarwar, M. A. Husain, S. U. Rehman, H. M. Ishqi and M. Tabish, Mol. BioSyst., 2015, 11, 522–531.
58. J. L. Yuan, H. Liu, X. Kang, Z. Lv and G. L. Zou, J. Mol. Struct., 2008, 891, 333–339.
59. P. D. Ross and S. Subramanian, Biochemistry, 1981, 20, 3096–3102.

---

Footnote

† These authors contributed equally to this work.

---