Variation of photostability of DNA-sensitive styrylcyanine dyes caused by N-alkyl functionalization

© 2020 Y. V. Snihirova et al.; Published by the Institute of Molecular Biology and Genetics, NAS of Ukraine on behalf of Biopolymers and Cell. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited UDC 577.336


Introduction
Fluorescent detection and visualization of biological molecules, organelles, cells and tissues is a powerful method of research that is widely used for different biomedical and medical purposes [1]. One of the necessary components this method does require is efficient fluorescent probes for the detection of various biological objects; particularly important are nucleic acids (NA) probes [2]. The dyes used as NA probes have to enhance strongly the fluorescence intensity upon binding to target NA molecule, possess high affinity to nucleic acid, low detection limit, high molar extinction coefficient and fluorescent quantum yield when bound, and sufficient photostability when exposed to the high intensity light necessary for microscopy [1].
A common problem with various cyanine fluorescent dyes is low photostability during continuous irradiation in the presence of oxygen that causes irreversible chemical destruction of the fluorophore molecule which thus loses its ability of fluorescence [3,4]. On the other hand, biological object can be damaged by direct excitation of the luminescent probe in close contact with that object. These processes limit the applicability of fluorescent dyes for practical applications such as fluorescence microscopy and labeling of molecules.
Fluorescent cyanine dyes are molecules commonly used as probes or labels in different biological applications such as imaging, biomolecular labeling, and proteomics [5]. As a result of several advantages of styryl dyes as compared to classical cyanine dyes (better photostability [6], large Stokes shift, relative ease of synthesis), styrylcyanines are extensively studied as possible fluorescent probes for NA detection and visualization. Thus, styrylcyanines were shown to be low-toxic NAs-sensitive dyes, they demonstrated fluorescence intensity enhancement upon binding to NAs as well as high fluorescence quantum yield in the NAs presence [7][8][9][10]. Applicability of benzothiazole styryl dyes for two-photon excited fluorescent visualization of living cells was also demonstrated [11]. Earlier, it was shown that the binding affinity of styrylcyanine to NAs could be modified by variation of the structure of N-alkyl tail group of the dye [12][13][14].
Here, a series of the benzothiazole-based styrylcyanine dyes (Fig. 1) functionalized by different N-alkyl tail groups was synthesized and characterized for their photostability and fluorescent sensitivity to nucleic acids. The long alkyl linkage (n-butyl) was chosen to eliminate the effects of terminal side groups on electronic transitions of chromophore. We studied UV-VIS absorption and fluorescent spectra of these dyes both in the absence and in the presence of NA (both dsDNA and RNA). We evaluated photostability of dyes as a change in the dye absorption after 150 min of irradiation by VIS source, and analyzed the influence of interaction between dyes and DNA on dyes photostability. By confocal laser scanning microscopy, we evaluated ability of styrylcyanine to penetrate into eukaryotic living cells and visualize cell components.

Materials and Methods
General. dsDNA (salmon testes) and yeast total RNA were purchased from Sigma-Aldrich Co. Solvents were of analytical grade. 1 H NMR spectra were recorded on Bruker ARX 400 spectrometers; chemical shifts (δ) were given in ppm relative to SiMe4. 50mM Tris-HCl buffer (pH 7.9) was used in all assays described.

Synthesis of the dyes
Referent dye with N-methyl substituent (Ref) was synthesized as described previously [15].
Monomeric benzothiazole styryl dyes (Sbt1, Sbt3) with positively charged tail groups were synthesized as described in [12]. The structures of the dyes were confirmed by 1 H NMR spectra, LC-MS and element analysis.
The dyes Sbt5, Sbt6 and Sbt7 were synthesized as follows:
Preparation of the solutions. Dye stock solutions were prepared by dissolving the dyes at 2 mM concentration in DMSO. Stock solutions of dsDNA and RNA were prepared by dissolving the NA in Tris-HCl buffer (50 mM, pH 7.9) at the concentration of 6.15 mM b.p. for dsDNA and 12.3 mM b. for RNA. Working solutions of free dyes were prepared by dilution of the dye stock solutions with Tris-HCl buffer (pH 7.9) to the concentration of 2 μM for absorption and fluorescence spectrophotometry and 5 μM for the photostability experiments. The working solutions of dye/NA mixtures for absorption and fluorescence spectrophotometry were prepared by mixing a dye aliquot (1 µL) and an aliquot of DNA or RNA stock solutions (10 µL) in Tris-HCl buffer (final concentration of dsDNA was 61.5 µM b.p. and RNA -123 µM b.). The working solutions of dye/NA mixtures for the photostability experiments were prepared by mixing a dye aliquot (7 µL) and an aliquot of DNA solution (30 µL) in 3 mL of Tris-HCl buffer (final concentration of dsDNA was 61.5 µM b.p.) Photostability. Photostability of the dyes in solution and in the presence of DNA was studied using the laboratory-designed equip-ment; its instrumental setup was as follows. The transparent glass vials with the samples (3 mL of each sample) were placed into the container with the reflecting inner surface, while on the top of the container the irradiating unit was installed. The irradiating unit consisted of 27 light-emitting diodes (blue light, 470 nm). To prevent the heating of the samples, cool water was placed inside the container that served as water bath for the glass vials with the samples. The distance between the samples and lamp was 2 cm. This way, photostability of the studied dyes was estimated both in a free state and in the presence of dsDNA (5 µM dye, 61.5 µM b.p. dsDNA) in 0.05 M Tris-HCl buffer (pH 7.9) and in 0.01 M phosphate buffer (pH 7.0); experiments were also performed for the dyes Sbt1 and Sbt3 in methanol. After irradiation of the mentioned samples during certain time (0.5, 1.5 and 2.5 hours), the irradiated solutions were placed into the quartz cell, absorption and fluorescence spectra of these solutions were measured, and the values of optical density (D) and fluorescence intensity (I) were obtained. Photostability of the dyes was then characterized by the ratios D/D 0 and I/I 0 , where D 0 and I 0 were optical density and fluorescence intensity of corresponding solutions prepared in the same way without irradiation. For each sample, the photostability measurements were repe ated two to four times, and standard deviations were calculated.
Determination of the binding constant (K b ) for the association of dsDNA with dyes Sbt1 and Sbt6. To estimate the stability of the association of selected dyes Sbt1 and Sbt6 with dsDNA, we conducted fluorescent titration of Sbt1 and Sbt6 with increasing dsDNA concentrations (0.2 -171 μM). Each experiment was performed three times. The titration curve is provided for the average values along with the standard deviations (SD). For the calculation of binding constant, we used the points corresponding to the excess of DNA. Thus we assumed that only negligible amount of the dye molecules will bind to dsDNA close to each other and affect each other's binding. Based on this assumption, the binding of dye to DNA could be described by the following equilibrium: dye + dsDNA ↔ dye-dsDNA (1) the constant of this equilibrium (binding constant, K b ) can be expressed with the equation (law of mass action [16]) below: (2) where C bd , C fd and C fDNA are concentrations of a bound dye, a free dye and free dsDNA binding sites, respectively. For the DNA concentrations starting from 20 μM, that is significantly higher than that of the cyanine dye (2 μM), the concentration of free dsDNA base pairs is roughly equal to the total dsDNA concentration C DNA ; C fDNA ≈ C DNA . Concentration of the free dye in equilibrium is C fd = C d -C bd (where C d is the total dye concentration). The measured fluorescence intensity (I) of the dye in the presence of dsDNA at the DNA concentration of C DNA can be expressed with the following equation where I 0 is the fluorescence intensity of the dye (2 µM) in the absence of dsDNA and I max is the fluorescence intensity of the dye (2 µM) in the presence of the indefinitely large dsDNA concentration. Equation (2) can be transformed into (3): where A= I max -I 0 .
K b and A values can be calculated as approximation parameters by fitting the experimentally obtained data I -I 0 versus C DNA by using equation (3). Fitting was performed and the values of K and A with their standard deviations were estimated by using Origin 8.0 program.
Eukaryotic cell staining. Mesenchymal stem cells were obtained like it was done before [17]. After transferring the slides to room temperature, the medium was removed. The cells were fixed on microscope slides at room temperature, and incubated with fluorescent dyes for 30 min. The washed cell samples were covered with a cover slip.
Confocal Laser Scanning Microscopy analysis (CLSM). CLSM analysis of the samples was done using Leica TCS SPE Confocal system with coded DMi8 inverted microscope (Leica, Germany). Images were acquired using excitation at 488 nm and emission collected at 510-605 nm for SYBR Green I, and excitation at 532 nm and emission collected at 600-740 nm for Sbt1.

Results and discussion
Absorption characteristics of the dyes and dyes' fluorescent properties in free state and in complexes with NA Earlier a series of styrylcyanine dyes were characterized as the fluorescent probes for DNA detection, and some of them showed good ability to enhance the fluorescent signal upon DNA binding [12]. Spectral properties of the dyes studied in this work (including Sbt1, Sbt3 and Ref dyes reported also in [12]) are provided in the Table 1. The absorption maxima of the styrylcyanines in 50 mM Tris-HCl buffer (pH 7.9) are located in the range of 510-524 nm, with moderate values of the molar extinction coefficients at maximum wavelengths (2.6-7.2)×10 4 M -1 cm -1 . Maxima of fluorescence excitation spectra of free dyes  are shifted for 11-27 nm to the long-wavelength region as compared to corresponding absorption maxima and are situated at 529-539 nm, whereas the maxima of fluorescence emission spectra are between 592 and 596 nm. The studied styrylcyanine dyes were found to be weakly fluorescent (14-31 a.u.) in the unbound state.
The presence of dsDNA leads to the longwavelength shift of the maxima of fluorescence excitation and emission spectra for 17-30 nm and 3-11 nm as compared to corresponding maxima of the free dyes; the excitation and emission maxima for the dyes in the presence of dsDNA are thus at 547-562 nm and 596-605 nm respectively. Upon binding with DNA, the dyes increased their fluorescence intensity up to 83 times (the most significant enhancement was observed for Sbt1 and Sbt3 dyes, the least significant -for Sbt7 and Sbt5). It could be concluded that the less significant values of fluorescence intensity enhancement were observed for zwitterionic dyes (Sbt5 and Sbt7), whereas the other dyes, which are positively charged, demonstrate higher enhancement values, and these values are higher for Sbt1 and Sbt3 containing the charge +2 compared to the dyes Sbt6 and Ref containing the charge +1. At the same time, the comparison of the dyes with the same charges points to a negative effect of -(СH 2 ) 4 COOCH 3 (comparison of Sbt6 and Ref), and to better effect of -(СH 2 ) 4 N + (CH 2 ) 2 CH 2 Ph group (Sbt1) as compared to -(СH 2 ) 4 dipyridyl one (Sbt3). Noteworthy, the presence of RNA (with the same concentration of nucleotide bases) results for the studied dyes (except for Sbt5) in about the same increase in fluorescence intensity as in the presence of dsDNA.

Photostability of the dyes
To investigate the relationship between molecular structures of the studied styrylcyanine dyes and their photostabilities, the effect of irradiation of the dyes solutions by the blue light (working wavelength about 470 nm) on the absorption and fluorescence spectra of these solutions was studied. The change of the characteristics (maximum wavelength and optical density) of the absorption spectra of the dyes (free and in the dsDNA presence) after 30 minutes and 150 minutes of irradiation could be seen in the Table 2.
It could be seen from the Table 2 that 30min irradiation leads to a small shift of the absorption maxima of some samples. At the same time, irradiation during 2.5 hours results in the short-wavelength shift for 1-17 nm (without any change in the peak shape) for all the studied samples. Noteworthy, for the dyes Sbt1, Sbt3 and Ref the shift for the free dye (10, 15 and 17 nm respectively) significantly exceeds that for the dye in dsDNA presence (3, 4 and 9 nm respectively). As for the dyes Sbt5, Sbt6 and Sbt7, the shift for the dyes in the dsDNA presence (3, 3 and 1 nm respectively) is about the same as for the free dyes (3 nm for all three dyes).
Besides the wavelength shift, irradiation of the dyes solutions also leads to the change in the optical density of the dyes. For all the dyes (except Sbt7), the optical density after 30-min irradiation became equal to 63-95 % of the corresponding value before irradiation for free dyes (the lowest values were observed for Sbt1 and Sbt3) and 75-93 % for dyes in the presence of DNA (Table 2; Fig. 2). Noteworthy, for the dyes Sbt1 and Sbt3, the addition of DNA caused noticeable increase in photosta-bility, whereas for other dyes the mentioned effect exceeding the experimental error was not observed. As for the dye Sbt7, irradiation during 30 minutes caused an increase in the optical density of this dye both in the presence and in the absence of DNA (Fig. 3). This result could be explained by bad solubility of Sbt7 in the buffer, which could be increased because of local heating as the result of light absorption by the low-fluorescent dye.
After 150 min of irradiation, for all the dyes (except Sbt7) further photodestruction was observed, and two tendencies described above became still more noticeable ( Table 2, Fig. 2).    Fig. 3). Thus, the obtained results regarding the irradiation effect on the dyes absorption spectra could be summarized as follows. First, it could be seen that the dyes with aromatic charged substituents Sbt1 (with N-alkyldimethylbenzyl substituent) and Sbt3 (with N-alkyldipyridyl substituent) bearing double positive charge are, on one hand, the least photostable in free state and, on the other hand, an addition of DNA to these dyes results in a strong increase of their photostability. On the other hand, the unsubstituted dye Ref as well as the dye Sbt6 (with uncharged aliphatic substituent) both bearing single positive charge are more photostable in free state than Sbt1 and Sbt3, whereas the dyes Sbt5 and Sbt7 (which are zwitterionic due to negatively charged N-substituents), are the most photostable. At the same time for these four dyes (Ref, Sbt6, Sbt5 and Sbt7), the changes in their photostability induced by the DNA presence do not exceed the experimental error.

Table 2. Optical density (D) of dyes in a free state and in the presence of dsDNA
Though the detailed mechanisms of dyes photodestruction are not completely understood, this process is generally considered to be connected with the transition of the excited dye to the triplet state. The dye in the triplet state can then either induce reactive oxygen species which will further damage the dye, or directly react with redox-active molecules present in the solvent [4]. Thus, the obtained dependence of the free dye photostability on its electric charge and/or substituent could be explained by the effect of charge/substituent on chromophore interaction with the possible redox-active molecules or generated reactive oxygen species in the solution (due to the interaction of the substituent group either with the chromophore, or with reactive molecules). It should be added that photostability of the free dyes in phosphate buffer (data not pre- sented) does not essentially differ from corresponding values in Tris-HCl buffer. At the same time, photostability of Sbt1 and Sbt3 in methanol is significantly higher (optical density after 150 min of irradiation is 89±7 % and 103±2 % respectively of corresponding values before irradiation) than in both Tris and phosphate buffers. These facts could point to the role of the redox-active molecules present in water in photodestruction of the dyes.
As for the different influence of the DNA presence on the photostability of different studied dyes, two factors should be considered. First, the dyes Sbt1 and Sbt3 which increase their photostability in the DNA presence have also the highest increase in their fluorescence intensity upon DNA addition (Table 1). Taking into account that these dyes possess double positive charge, we could suppose that they bind to DNA with higher affinity as compared to other dyes. If so, the reason of the increase in the dye photostability should be its binding with DNA (despite the anticipated increase in the excited state lifetime). The possible explanation could be the restriction (at least partial) by DNA of the access to the dye molecule of the generated reactive oxygen species or redoxactive molecules. On the other hand, one cannot also exclude the fixation of the dye structure upon binding that does not permit the excited dye to acquire the reaction-suitable conformation. As for other studied dyes which either contain single positive charge or are zwitterionic, we could suppose that they bind with DNA with lower affinity. This means that lower percent of these dyes form complexes with DNA, and thus DNA influence on the dyes photostability is less significant. Another factor is that the dyes Sbt5, Sbt6 and Sbt7 are much more photostable in a free state as compared to Sbt1 and Sbt3, so that the influence of DNA on their photostability would not be as noticeable for the former dyes as for the latter ones.
For the dyes Sbt1, Sbt3 and Ref in the presence of dsDNA, we also compared the influence of irradiation on dye optical density and on its fluorescence intensity (Table 3). It is seen from the Table 3, that while for the dyes Sbt1 and Sbt3 fluorescence intensity is decreased to much lower extent as compared to optical density, for the dye Ref emission is suppressed almost as strongly as absorption (the difference between D/D 0 and I/I 0 for Ref is due to the nonlinear dependence of fluorescence intensity on optical density). Such difference between Ref and two other dyes corresponds to the difference between them in DNA effect on their photostability. In the case when the DNA-bound dye molecules are more photostable than the free ones, irradiation of dye-DNA solution leads to the destruction of mostly non-bound dye molecules. Since the emission intensity of DNA-bound dyes significantly exceeds that of the free dyes, such photodestruction makes stronger effect on dye absorption than on its emission. This feature

Determination of binding constant
Based on the results of the dye fluorescence intensity increase in the dsDNA presence, we supposed above that increased photostability of the dyes Sbt1 and Sbt3 in the DNA presence is due to higher affinity of these dyes binding to DNA as compared to other studied dyes. To support this assumption, we have compared the affinity of dye binding to dsDNA for the dyes Sbt1 and Sbt6. Both dyes increase their fluorescence intensity upon the DNA binding, but while the presence of DNA strongly enhances the photostability of Sbt1, that of Sbt6 does not depend on the DNA presence. The equilibrium constant (K b ) of binding to dsDNA was determined for Sbt1 in [12]; for Sbt6, K b was found here using the approximation of the dye fluorescent titrations by dsDNA with the equation (3) (Fig. 4), in the same way as it was done for Sbt1 in [12]. Thus, the values of K b for Stb1 and Sbt6 were found to be 5.0×10 4 M -1 and 1.0×10 4 M -1 respectively. Such values of K b are typical for intercalating dyes, the usual range being between 10 4 and 10 6 M -1 , while groove binders have higher binding constants (10 5 -10 9 M −1 ) [18].
Since the K b value for Sbt6 is 5 times lower than the one for Sbt1, higher percent of dye molecules are bound to DNA in the case of the latter as compared to the former. For the obtained K b values and used concentrations (61.5 µM DNA and 5µM of dye), the rough estimation based on the mass action law results in 75 % and 38 % of the bound dye in the dye-DNA solution for Sbt1 and Sbt6 respectively. That may explain (at least partially) the fact that the DNA effect on the dye photostability was significant for Sbt1 and was not observed for Sbt6. Other alternative (or additional) explanations could be e.g. more rigid fixation in DNA or longer time of being bound (since dynamic equilibrium between free and DNAbound dyes takes place in dye-DNA solution) for Sbt1 as compared to Sbt6, or different mechanisms of fixation making different groups of the chromophore accessible to reactive oxygen species or redox-active molecules.

Staining by the dyes
Since the dye Sbt1 bearing N-alkylbenzylamine group has demonstrated rather high constant of binding to dsDNA, as well as strong increase in fluorescence intensity and photostability upon this binding, we have studied the possibility to use this dye as fluorescent stain for fluorescent microscopy. Thus, we have performed the staining of mesenchymal stem cells with Sbt1 in co-staining with DNAsensitive stain SYBR Green I (Fig. 5); the obtained images showed that Sbt1 brightly stains cytoplasm, so it could be suggested to visualize cytoplasmic RNA. The dye seemed to penetrate less efficiently into the nucleus, but nucleus components (probably nucleoli) stained by this dye were highly discernable. In contrast to styrylcyanine dye Sbt1, SYBR Green I stained mainly nucleus, and discrimination of nucleolus components with this dye was less pronounced (Fig. 5).

Conclusions
A series of the benzothiazole-based styrylcyanines functionalized with N-alkyl tail groups of different nature was synthesized and characterized for their photostability and fluorescence response upon binding to nucleic acids. These styrylcyanine dyes are weakly fluorescent in the unbound state, whereas the presence of NA caused an increase of the dyes fluorescence intensity up to 83 times. The intensity of dye fluorescence response depends on N-alkyl tail substituent of the dye. The most significant enhancement was observed for the dyes with positively charged aromatic substituents Sbt1 (with N-alkyldimethylbenzyl substituent) and Sbt3 (with N-alkyldipyridyl substituent), the least significant -for the dyes with negatively charged groups Sbt7 (sulfosubstituent) and Sbt5 (carboxy-substituent). The dyes Sbt1 and Sbt3 are, on one hand, the least photostable in a free state and, on the other hand, addition of DNA to these dyes results in strong increase of their photostabi lity. The dyes, which either bear single positive charge on chromophore (uncharged N-alkyl tail group) or are zwitterionic (due to negatively charged N-alkyl tail group), are more photostable when free, and the presence of DNA weakly affects their photostability. Thus, the charge and/or the nature of N-alkyl tail group makes strong effect on the photostabi lity of benzothiazole styrylcyanine dye in water solution. The dye bearing positively charged N-alkyl tail groups binds to DNA with higher affinity (binding constant for Sbt1 is 5×10 4 M -1 ) comparing to the dye Sbt6 with uncharged tail group (binding constant 10 4 M -1 ).
Staining of mesenchymal stem cells with the dye Sbt1 in co-staining with sensitive to DNA dye SYBR Green I has shown that Sbt1 brightly stains cytoplasm and nucleus components (probably nucleoli), whereas SYBR Green I stains mainly the nucleus. Presumably, Sbt1 is suggested to visualize cytoplasmic RNA and clusters of RNA in the nucleus.
Thus, variation of chemical nature of N-alkyl tail group is a way to design styrylcyanine dyes of different photostability. Functionalized styrylcyanines are suggested as handy and photostable fluorescent stains for microscopy techniques and nucleic acid detection in solution.