Fluorescence analysis, with its excellent sensitivity and accuracy, is widely used to study the interactions between DNA and drugs or poisonous molecules. Pyrene derivatives have fluorescence, shown in Figure 2, because of their large conjugated system and rigid planar structures. Pyrene groups in the excited state themselves could form charge transfer complexes, known as excited dimers. Therefore, pyrene derivatives present two fluorescence peaks, including the red shift emission of the dimer compared to the monomer.

The fluorescence spectra here are used to estimate the binding mode and the binding abilities of pyrene derivatives to DNA. After mixing with DNA at increasing concentrations, the pyrene derivatives solution shows a gradually weakening in fluorescence, as shown in Figure 2 for p53 DNA. When the molar ratio of DNA and pyrene derivatives is 20:1, the quenching is about 80%. This result shows that there may be energy and electron transfers occurring between pyrene derivatives and DNA, leading to the fluorescence quenching.

Although changes in the fluorescence of the pyrene derivatives solution in the presence of DNA act as evidence of interactions, it cannot be a direct proof of the binding mode of pyrene derivatives and DNA. It may because hydrophobic interactions between pyrene derivatives and DNA change the microenvironment of pyrene derivatives, and this hinders the electron transition [16]. The Stern–Volmer equation [Equation (1)] [17] and the Scatchard equation [Equation (2)] [18] can be used to describe the fluorescence quenching process of pyrene derivatives as follows: F₀/F = 1+ K_qτ₀ [DNA] = 1 + K_SV [DNA] (1) log[(F₀ − F)/F] = logK_b + n log[DNA] (2) where F₀ is the fluorescence intensity of pyrene derivatives. F is the fluorescence intensity of pyrene derivatives with the addition of the quenching agent, DNA. [DNA] is the concentration of DNA added. τ₀ is the average life expectancy of the fluorescent quencher molecule. For a biomacromolecule, τ₀ is 10 ns on average [19]. K_SV is the linear Stem-Volmer constant, which can be calculated from Equation (1). K_q is the rate constant of the quenching process, which is equal to K_SV divided by τ₀. K_b is the binding constant and n is the binding number of DNA to pyrene derivatives, which can be calculated from Equation (2).

The calculated results, K_q, K_SV, K_b and n, are available in Table 1. First, the quenching rates of DNA, K_q, are all above 10¹² L·mol⁻¹℘s⁻¹, indicating that their interactions with pyrene derivatives involve static quenching [20]. Second, the results suggest that the values of K_b of pyrene derivatives to p53 DNA are commonly larger than that for C-myc DNA, suggesting the binding specificity of pyrene derivatives to p53 DNA superior to that for C-myc DNA. Third, the shorter the side chain of pyrene derivatives, the larger the K_b values are obtained (4.04 × 10⁵ L·mol⁻¹ of 1-OHP and 1.39 × 10³ L·mol⁻¹ of 1-PBO for C-myc DNA, respectively). Therefore, for the same target (C-myc DNA or p53 DNA), the order of the binding ability is 1-OHP > 1-PBO.

The thermodynamic experiments were performed at room temperature and physiological temperature (298 K and 310 K, respectively). From the results of C-myc DNA in Figure 3, we can see that the slope of Stern-Volmer curves decrease as the temperature rises from 298 K to 310 K. That is, the Stern-Volmer constant, K_SV, is inversely proportional to the temperature, which confirms that static quenching of DNA with the pyrene derivatives happens [21].

The thermodynamic constants (Gibbs free energy ΔG, enthalpy ΔH, entropy ΔS) could be calculated according to Equations (3–5). The subscripts “1” or “2” in our work refers the condition of 298 K or 310 K, respectively: lnK_b1/K_b2 = (1/T₁ − 1/T₂) ×ΔH/R (3) ΔG = −RTlnK_b (4) ΔG = ΔH − TΔS (5)

The relevant thermodynamic constants are summarized in Table 2. The negative values of ΔH and ΔS indicate that pyrene derivatives bind to the DNA mainly via van der Waals forces and hydrogen bonds [22], which are widely considered to make an important contribution to the binding in the minor groove of DNA [16]. Thus, the minor groove binding of pyrene derivatives can be considered. Since the larger magnitude of ΔG indicates a greater interaction possibility, for the same target (C-myc DNA or p53 DNA), the order of the binding ability is 1-OHP>1-PBO, and the binding specificity of pyrene derivatives to p53 DNA is superior to that for C-myc DNA.

Sodium chloride, NaCl, is used as the ionic strength modifier to determine whether there are electrostatic interactions between DNA and pyrene derivatives. The increased ionic strength of solution can inhibit the electrostatic interactions between DNA and binding molecules [23]. If the enhanced fluorescence of the DNA-pyrene derivatives solution is observed with added NaCl, indicating that the amplitude of the fluorescence quenching of pyrene derivatives binding with DNA is weakened, an electrostatic interaction can be concluded. There are little changes in the extent of quenching of DNA-pyrene derivative solutions in the absence and presence of NaCl (<0.1%), indicating that the pyrene derivatives don’t have electrostatic interaction with DNA.

Anionic quenchers, such as potassium iodide (KI) can quench the fluorescence of organic molecules. When the organic molecule binds into the groove region of DNA, I⁻ ions exposed in solution can increase the extent of quenching [24]. For intercalation or electrostatic effect of organic molecules, I⁻ ions will be protected by the phosphate backbone and base pairs of DNA, and the extent of quenching can be reduced. As shown in Figure 4, the quenching extents of pyrene derivatives with the addition of KI in the absence and presence of DNA are denoted by (F₀/F − 1). With the increasing concentration of KI, the (F₀/F − 1) value gradually increases. When the KI concentration continues to increase over 5 μmol/L, the increase of the (F₀/F − 1) value slows down since the reactions reach saturation.

This shows significantly that in the presence of KI, the quenching extent of pyrene derivatives binding to C-myc DNA is increased, indicating the existence of the groove binding mode of pyrene derivatives to DNA.

Ethidium bromide (EB) has weak fluorescence in solution. When EB intercalates base pairs of the DNA double helix, its fluorescence intensity can increase 50–100 fold. If the conformation of the duplex DNA changes, or there are organic molecules competing with EB for the DNA interaction, EB can be freed from DNA complexes, causing fluorescence quenching. Therefore, the fluorescence changes of the EB-DNA system in the absence and presence of organic molecules can be used to identify the DNA intercalation [25,26,27,28,29]. Here, EB is used as a fluorescent probe on interactions between organic molecules and DNA. As shown in Table 3, the fluorescence change of the EB-DNA system is represented by R (%) = (F_EB-P − F_EB)/F_EB × 100%, where F_EB and F_EB-P represent for the EB-DNA’s fluorescence intensity in the absence and presence of pyrene derivatives, respectively.

R (%) values from −5.46% to −13.17% indicate that both 1-OHP and 1-PBO can compete with EB as weak intercalating agents in DNA interactions. Comparing the R (%) values of 1-OHP, 1-OHP is intercalates easier into the base pairs of p53 DNA than those of C-myc DNA. As for 1-PBO, approximate values of R (%) (−8.74% and −8.85% for C-myc and p53 DNA, respectively) show that there is no selectivity of DNA intercalation.

Circular dichroism spectra can provide conformational information about biomacromolecules. The positive peak at 260–280 nm is the signal of accumulation of DNA base pairs, and the negative peak at 245 nm corresponds to the double helix structure of B-type DNA [30]. The parallel G-quadruplex has a positive peak at 265 nm and a negative peak at 245 nm. The anti-parallel G-quadruplex has apositive peak at 295nm and a negative peak at 265 nm. A shoulder peak at 280 nm indicates polyploidy formation [31,32]. Organic molecules without CD signals may have the induced signals (ICDs) observed when they bind to DNA. It is generally believed that intercalating agents often show a slight negative ICD signal, while those in the groove binding mode often appear as a positive ICD signal [33,34]. Figure 5 shows the CD and ICD spectra of DNA solutions with the increased concentration of 1-PBO. An obvious positive ICD signal appears in Figure 5C, indicating that 1-PBO binds into the groove of the duplex DNA. In Figure 5D, the positive ICD signal of p53 DNA, weaker than that of C-myc DNA, suggesting that 1-PBO plays a stronger role on the groove binding of p53 DNA than that for C-myc DNA. As shown in Figure 5A,B, when mixed with 1-PBO, the negative peak of DNA at 245 nm in CD spectra is reduced, suggesting that the O atom of 1-PBO may form hydrogen bonds with base pairs in the groove of DNA.

Besides, with increased concentration of 1-PBO, a shoulder peak appears at about 290 nm indicating that there may be the formation of the anti-parallel G-quadruplex of the G-rich DNA sequence. As for the 1-OHP, apart from the similar reduced negative peak at 245 nm, there is an apparent reduced positive peak at 265 nm in the CD spectra of DNA, due to the intercalation of 1-OHP. This result confirms that the binding of pyrene derivatives can affect the normal conformation of the duplex DNA. Combined with the results of iodide quenching and thermodynamic studies, it can be concluded that the DNA interactions of pyrene derivatives are dominated by groove binding. Besides, results of EB competition experiments confirm the existence of intercalation in the binding of pyrene derivatives into DNA base pairs.

Transient fluorescence spectroscopy has recently become a critical tool in the biochemical and biophysical field [35]. The transient fluorescence experiments on lifetimes are used to discuss the binding modes of pyrene derivatives with p53 DNA and C-myc DNA, as shown in Figure 6 and Table 4. The lifetime I of 1-OHP is observed as 13.8 ns. With the addition of DNA, its lifetime changes little. Consistent with results in steady state fluorescence quenching experiments, this indicates that static quenching dominates in the DNA-1-OHP binding process [36]. Due to the conformational complexity of 1-PBO with the long side-chain, there are two lifetimes, 90.7 ns and 4.90 ns. The distribution of the longer one is related to the DNA-bound structure of 1-PBO and the shorter one corresponds to the unbound structure. Besides, the magnitude of lifetime often reflects the variations in fluorophore-DNA interactions resulting from changes in the structure of the organic molecules or DNA molecules and environmental changes [37,38]. Since the unbound molecules have characteristic fluorescence lifetimes depending on their circumstances, the changes shown with the unbound lifetime component suggest they are sensitive to the changes in molecular environment [39]. According to the results, the lifetime of DNA-bound component changes little, which confirms the static quenching. The average lifetime of 1-PBO with DNA interaction is shorter than that of unbound 1-PBO, suggesting that 1-PBO does affect the DNA double helix structure, similar with some groove binding and intercalation agents in previous studies [38,40]. Moreover, combining the Stern-Volmer fluorescence data with lifetime measurement data discussion, we again confirmed the static quenching that existed in the interaction of pyrene derivatives with the two kinds of DNA.

Non-native polyacrylamide gel electrophoresis (PAGE) can be used for the separation and purification of biological macromolecules [41,42]. Especially, high concentration PAGE has a higher resolution in the separation of short DNA fragments. The effect of 1-OHP on the structure of p53 DNA is shown in Figure 7. Since G-quadruplex DNA of two hybrid G-rich sequences (referred as G-4) has more charge and a more tight structure than double-strand DNA (referred to as ds), G-4 will have a faster mobility rate than ds. As shown in Figure 7, the double-strand strips of p53 DNA slightly diffuse in the presence of 1-OHP, indicating that 1-OHP doesn’t affect the stability of the helix structure of double-strand DNA. However, as the molar ratio of 1-OHP and DNA increases, the G-4 strip becomes brighter, indicating the formation trend of G-quadruplex DNA, which is also confirmed by the CD results.

Fluorescence measurements were performed with a LS50B spectrofluorimeter (Perkin Elmer, Massachusetts, MA, USA). Transient fluorescence spectra were measured on a FM-4P-TCSPC Spectrofluorometer (Horiba Jobin Yvon, Paris, France). Circular dichroism spectra were recorded with a Chirascan circular dichroism spectrophotometer (Applied Photophysics, Leatherhead, UK). Non-native polyacrylamide gel electrophoresis was performed in a PROTEAN II xi Cell (Bio-Rad, Hercules, CA, USA), and the PAGE image was captured by GelDoc-It TS Imaging System (UVP, Citrus Heights, CA, USA). PH measurements were carried out with a Seven Multi pH digital pH-meter (Mettler Toledo, Shanghai, China) with a combined glass-calomel electrode. An electronic thermostat water-bath (Changzhou Guohua Medical Instrument Company, Changzhou, China) was used for controlling the temperature.

p53 gene P1 promoter fragment (ss1:5′-CCTCCTCCCCAACTCC-3′, ss2:3′-GGAGGAGGGGTT-GAGG-5′); C-myc gene promoter of NHE III1 fragment (ss1:5′-GGGAGGGTGGGGAAGG-3′, ss2:3′-CCCTCCCACCCCTTCC-5′). Single-stranded oligonucleotides were diluted with 100 mmol/L Tris-HCl buffer solution (pH = 7.4). For duplex DNA, two complementary single-stranded oligonucleotides were mixed in equimolar proportions, annealed at 85 °C and slowly cooled to room temperature to form of the duplex (ds) at 500 μmol/L: A stock solution (10 mmol·L⁻¹) of pyrene derivatives (AccuStandard, USA) was prepared by dissolving its crystals in methanol (Sigma-Aldrich, St Louis, MO, USA), and then diluted to different concentrations. Ethidium bromide (EB, AccuStandard, Palo Alto, CA, USA) solution, NaCl solution and KI solution (500 μmol·L⁻¹ and 100 μmol·L⁻¹) were all prepared by dissolving its crystals in water. SYBR Green I (Unique, Beijing, China); DNA Ladder, Ultra Low Range (Fermentas, Burlington, ON, Canada); 6 × Orange DNA Loading Dye (Fermentas). Pure 18-MΩ MilliQ water was used for preparation of all solutions.