Unveiling the tartrazine binding mode with ds–DNA by UV–visible spectroscopy, electrochemical, and QM/MM methods

Abstract

Here, we studied the interaction between the food colorant tartrazine (TZ) and double stranded DNA (dsDNA), using spectroscopic, electrochemical, and computational methods such as QM/MM combined with TD-DFT. Despite the UV–vis spectroscopy is widely used to study the interaction between molecules, for the case of TZ there are discrepancies in the analyses presented in the literature available, presenting both hyperchromic and hypochromic effects and consequently different rationalizations for their results. Herein we propose the combination of UV–vis experiments with the design of high-level computational models capable of reproducing the experimental behavior to finally define the proper binding mode at the molecular scale together with the rationalization of the experimental optical response due to the complex formation. To complement the UV–vis experiments, we propose the use of electrochemical measurements, to support the results obtained through UV–vis spectroscopy, as it has been successfully used for the determination of interaction modes between small molecules and biomolecules in any condition. Our UV–vis spectroscopy experiments showed only a hypochromic effect of the absorption spectra of TZ after interaction with DNA, indicative of TZ being deeply buried in the DNA structure. The effect of ionic strength in the experimental procedures led to the dissociation of TZ, thus indicating that the interaction mode was groove binding. On the other hand, the electrochemical studies showed an irreversible reduction peak of TZ, which after the interaction with DNA exhibited a positive shift in potential that can be attributed to groove binding. The binding constant for TZ-DNA was calculated as $4.45x{10}^4{M}^{-1}$ (UV–vis) and $5.75x{10}^4{M}^{-1}$ (electrochemistry), in line with other groove binder azo dyes. Finally, through the QM/MM calculations we found that the minor-groove binding mode interacting in zones rich in adenine and thymine was the model best suited to reproduce the experimental UV–vis response.

Introduction

Tartrazine (TZ) is a yellow synthetic azo dye used for coloring food, drugs, and cosmetic products. As an azo dye, its degradation/metabolism produces aromatic amines in the body, which are considered highly toxic. These molecules can damage urinary organs, the kidney, the stomach, and the liver [1], [2]. For example, the food azo dye Sudan I has both genotoxic and carcinogenic effects, and international panels such as the European Food Safety Authority (EFSA) assume that other azo dyes with similar structures are potentially genotoxic and possibly carcinogenic [3]. Therefore, a series of studies on azo dye interactions with biomolecules such as human hemoglobin [1], albumins [4], lysozyme [5], and deoxyribonucleic acid (DNA) [6], [7], [8], [9] have been reported.

Among these biomolecules, the interaction with DNA is one of the most important aspects in several areas of research because it helps to understand the mechanism of action or toxicity of different molecules. It is well known that DNA can be found unwrapped under several conditions of normal DNA reading and replication. The interaction of small molecules and DNA has been also demonstrated under physiological conditions, not only for dyes, but for a series of molecules, including drugs, which support their mechanism of action via interaction with DNA by different binding mechanisms [10], [11]. The interaction between molecules and DNA has been categorized into three types: (i) electrostatic (between positively charged molecules and the negative DNA phosphate groups), (ii) groove binding, and (iii) intercalation (associated with planar aromatic molecules) [11], [12], [13]. These types of interactions can change the DNA conformation, interrupt protein-DNA interactions and potentially lead to the breakage of DNA strands. All of these events can have substantial effects on gene expression [10], [12].

The commonly used techniques for studying the interaction between small molecules and DNA are UV–visible spectroscopy, fluorescence spectroscopy, circular dichroism spectroscopy, viscosity measurement studies, isothermal titration calorimetry, Fourier transform infrared spectroscopy, and voltammetry measurements [12]. In this respect, voltammetric techniques can work as an excellent complement to other techniques based on the optical response of the complex, especially when the results show weak absorption or fluorescence spectra and/or overlapping of electronic transitions with DNA. UV–visible spectroscopy is one of the most used techniques because of its simplicity. The interaction between small molecules and DNA leads to changes in the UV–visible spectra of both molecules. In this context, to be able to correlate the band intensity (hyper- or hypochromic effect) and/or any shift in the position of the peak in the spectra (batho- or hypsochromic shift) to a specific interaction mode is an important target of research, as it provides the insights to predict the interaction mode of chemically similar molecules depending on their optical response. According to Rehman et al. [14], when small molecules intercalate into DNA, the UV–visible spectrum of the molecule shows a bathochromic shift as well as a hypochromic effect. In the case of groove binding interactions, a smaller or no bathochromic shift is observed. On the other hand, in the case of electrostatic interactions, Horakova et al. [15] indicated that a hyperchromic effect is detected. If the band observed for the interaction comes from DNA absorption, the hyperchromism could be associated with partial uncoiling of the double helix, which exposes more bases of the DNA to the solution, explaining the increase in the DNA absorption band [16]. Therefore, the hyper- or hypochromic effect must be analyzed and assigned carefully, as there are several variables involved in the observed spectral changes after complexation with DNA.

For food colorants, a few studies have confirmed the interaction of DNA with TZ [6], [9], [17], sunset yellow (SY) [7], [16], carmoisine (CR) [18], [19] and Allura Red (AR) [20] by UV–visible absorption spectroscopy. These four colorants have two aromatic moieties connected by an azo group; therefore, a similar interaction mode between them and DNA was expected. For SY, Asaadi et al. [7] added DNA to different concentrations of SY observing a hypochromic effect on the main band of the colorant, from which calculated an intrinsic binding constant (K) of $2.1\times {10}^3$ M⁻¹. Also, based on other experiments, they stated that the interaction of SY with DNA is a combination of intercalation and electrostatic modes due to the planar structure and positive charges of the colorant. Nevertheless, the problem with that binding mode proposal is that SY is in fact not planar; therefore, the intercalation mode should not be favored. In addition, the ${\mathit{pK}}_a$ of the colorant is 10.5 [21], which means that below pH 8.0, the molecule is actually in an anionic state, having a repulsive effect toward electrostatic interactions. On the other hand, Kashanian et al. [16] observed a hyperchromic effect for the same experiment and attributed the changes in the spectra to an external contact of the colorant with DNA, with a K of $7.0\times {10}^4$ M⁻¹. The authors concluded that the interaction mode of SY is groove binding by comparison of the K value with those of other groove binder agents reported. For CR, two reports showed opposite effects in the UV–vis spectra when different amounts of DNA were added to a fixed concentration of CR. Arvin et al. [19] observed a hyperchromism tendency with a K of 6.2$\times {10}^4$ M⁻¹. The interaction of CR with DNA was explained by the formation of hydrogen bonds between the amino and hydroxyl groups of the colorant with the DNA bases. On the other hand, Basu et al. [18] observed a hypochromic effect upon the addition of increasing amounts of DNA, indicating that the spectral changes can be attributed to a strong interaction between the ligand and the DNA base pair. The K value obtained was 2.9$\times {10}^4$ M⁻¹, and the groove binding mode was proposed using other techniques instead of spectral analysis. In the case of the red colorant AR, a hypochromism effect was observed after the addition of increasing amounts of DNA with a K of 7.5$\times {10}^5$ M⁻¹. The groove binding mode was deduced by comparison of the K value with those of other well-established groove binders [20]. Finally, for the case of TZ, the UV–vis spectra found in the literature also presented inconclusive and opposite results. Kashanian et al. [17] studied the interaction of TZ with DNA at a neutral pH of 7.4, and their results showed a hyperchromic effect in the UV absorption band of TZ. According to the authors, the binding of TZ damages the DNA structure and interacts with a K value of $3.75\times {10}^4$ M⁻¹. Using other techniques, the authors indicated that the interaction mode is groove binding. On the other side, Basu et al. [6], after conducting the same experiment, observed a hypochromic change after the complexation of TZ with DNA. In this case, the K value obtained for TZ was $3.53\times {10}^4$ M⁻¹, and the interaction mode through groove binding was also suggested by the results obtained with other techniques. For azo colorants, different authors proposed the same binding mode with similar values of the binding constant, but they observed opposite changes in the UV spectra. Considering that the main goal from the use of spectroscopic techniques in this context is to be able to correlate changes in the spectrum with a specific binding mode, the discrepancy between the binding mode and the optical response becomes relevant to be solved. This would provide a predictive character to the method allowing to search similar modifications in the patterns of the UV–vis spectrum after the binding of other chemically similar small molecules to DNA.

Because of this, the combination of electrochemical methods, UV–vis spectroscopy and computational models will provide sound evidence to solve the discrepancy in the correlation between the binding mode between TZ and ds-DNA with the UV–vis response. In recent decades, the use of electrochemical techniques have also demonstrated to provide detailed insights regarding DNA-drug interactions shown by previous studies [22], [23], which could be an excellent complement to the UV–visible methods due to their accessibility and low cost. The rationale behind the use of this technique comes from the fact that the peak potential and peak current of the compound change in the presence of DNA if the compound interacts with it and vice versa. Furthermore, the changes in the peak current of the molecule by the addition of increasing DNA concentration can be used for the determination of binding constants, while the shift in potential can be used to determine the mode of interaction [12]. For this purpose, some authors have reported that if the peak potential (E_p) of the molecule shifts to more negative values when it interacts with DNA, the interaction mode is dominated by electrostatic interactions. In contrast, when E_p shifts to more positive values, the interaction mode is intercalation or groove binding [22], [24]. To corroborate this, authors such as Catalán et al. [22] studied the effect of the ionic strength (IS) of the medium. At low IS values, the electrostatic interaction mode between a molecule and DNA is predominant; meanwhile, at high IS values, the charges of the DNA and the molecule are shielded, which reduces noncovalent interactions mediated by electrostatic interactions [14], [25].

Herein, to solve the presented discrepancy, we show the combination of UV–vis spectroscopy, electrochemical methods, and computational models including TDDFT calculations as a robust strategy to provide the necessary evidence to solve the discrepancy in the correlation between the binding mode between TZ and ds-DNA with the UV–vis response. The use of molecular models of the TZ-DNA complex and the calculation of the UV–vis spectra of several possible binding modes by using TDDFT allowed us to determine which interaction complex reproduces the experimental UV–vis, results supported by the electrochemical response and the binding mode predicted by this method. Additionally, to the best of our knowledge, there are no reports on electrochemical studies regarding the interaction between TZ and DNA or those complemented by computational models. Finally, the experimental results will be compared and used to determine the binding constant.