Investigating the interactions of a novel anticancer delocalized lipophilic cation and its precursor compound with human serum albumin

Abstract

F16 is a novel identified delocalized lipophilic cation (DLC) which has been found to inhibiting a variety of tumor cell proliferation due to its selective accumulation in the mitochondria of carcinoma cells. To gain further insight into the thermodynamic properties of this small molecule, we chose human serum albumin (HSA) as the model protein, and investigated the interactions of F16 and its precursor compound PVI with HSA by comprehensive spectroscopy, electrochemistry and molecule modeling methods. The static fluorescence quenching of HSA suggests that both F16 and PVI can form complexes with HSA, though the binding mechanisms are different. The main driving forces for F16–HSA binding are typical hydrophobic interactions, while PVI–HSA binding takes place through electrostatic interactions. F16–HSA binding shows an adverse temperature dependence recognized as the effect of the high activation energy requirement in the binding process generated by the specific structural obstacle. Both F16 and PVI can bind with HSA and thus benefit their transportation and elimination in body, however, the positive charge of F16 may have negative effect on the binding interaction.

---

1. Introduction

Delocalized lipophilic cations (DLCs) are a class of compounds with a high hydrophobicity, a delocalized positive charge and rigid structures. A variety of DLCs display cytotoxicity for rapid accumulation in mitochondria and alter the mitochondria functions in response to the negative charge inside the transmembrane potentials,^1–4 and it has been found that they are selectively toxic to carcinoma cells compared with normal cells. The higher mitochondria transmembrance potential (ΔΨ_m) of carcinoma cells versus normal cells has been hypothesized to contribute to the increased uptake and prolonged retention of these DLCs, accounting for the selective tumor cell killing.^5,6 Many DLCs have been designed and synthesized as novel drugs and exhibit efficacy in killing carcinoma cells or inhibiting cell growth.^7–9

F16, (E)-4-(1H-indol-3-ylvinyl)-N-methylpyridinium iodide (Scheme 1), a novel DLC first identified by Fantin et al.,¹⁰ was found to accumulate in the mitochondria of mammary epithelial, oncogene neu-overexpressing cells, thus selectively inhibiting tumor cell proliferation. Further study showed that F16 could trigger necrosis in target carcinoma cells incapable of apoptosis by virtue of Bcl-2 overexpression.¹¹ The good antiproliferative property of F16 has been linked to its ability to dissipate the proton gradient across the inner mitochondrial membrane.¹⁰ Since the positive charge is the major factor that allows F16 to accumulate in the mitochondria of cancer cells, we suspect whether it will affect other thermodynamic performances of F16, such as the absorption, transportation, distribution and metabolism, considering that these properties could significantly alter drug efficacy.^12,13

Scheme 1 The structure of PVI (left) and F16 (right).

Human serum albumin (HSA) is the most abundant protein in blood plasma which facilitates the transportation and distribution of various endogenous and exogenous compounds, including nutriment and pharmaceuticals.^14,15 Due to its ability to reversibly bind various drug molecules and alter their pharmacokinetic properties, serum albumin has long been the center of attention of the pharmaceutical industry.¹⁶ Therefore, HSA has been chosen as the model protein to investigate the effect of the positive charge on the thermodynamic properties of F16. For clarifying the result, PVI, (E)-3-(2-(pyridine-4yl)vinyl)-1H-indole, the precursor compound of F16 with a similar structure but no positive charge (Scheme 1), is used as a control during the whole experiment process.

In this paper, we focus on the relationship between the structure factors and binding properties of F16 and PVI with HSA. For study convenience, F16 and PVI have been synthesized and characterized (Fig. SI 1 and 2†). The interactions of F16 and PVI with HSA have been investigated by comprehensive spectroscopy, electrochemistry and molecule modeling methods. A better understanding of the effect of the positive charge on the F16–protein binding should provide useful information for DLC discovery and drug efficacy evaluation.

2. Materials and methods

2.1 Reagents

Gramine and 4-pyridinecarboxaldehyde were used for the synthesis of PVI. They were obtained from Sigma Aldrich, were of analytical reagent grade and were used without further purification. HSA was purchased from Sigma Aldrich and was prepared in PBS (pH 7.4) at the concentration of 2 × 10⁻⁶ mol L⁻¹. All solutions were kept in dark place at 4 °C. All other reagents were of analytical reagent grade and doubly distilled water was used in all procedures.

2.2 Apparatus

All fluorescence spectra were recorded with a LS-55 spectrofluorimeter from Perkin-Elmer Corporate, which was equipped with quartz cells (1.0 cm) and a thermostat bath. The UV-vis absorption spectra were measured by an UNICO 4802 UV-vis Double Beam Spectrophotometer. A CHI 660C electrochemical workstation from Shanghai Chenhua Instrument Company equipped with a three-electrode system was used to measure the electrochemical experiments. The CD spectra were recorded on a Circular Dichroism Photomultiplier from Applied Photo Physics Limited with a quartz cell having a path length of 0.1 cm. For the synthesis and identification, the ¹H NMR spectra were recorded on a mercury 300 MHz NMR spectrometer using DMSO-d₆ as the solvent.

2.3 Spectra measurements

The fluorescence measurements were measured at different temperatures (308 K, 303 K, 298 K and 293 K); the slit widths for the excitation and emission were set to 15.0 and 12.0 nm throughout. For HSA, 295 nm was chosen as the excitation wavelength. The concentration of HSA for the fluorescence measurements was 2 × 10⁻⁶ mol L⁻¹. The UV-vis absorption spectra were measured with a 1 cm quartz cell at room temperature.

2.4 Electrochemical measurements

A gold disk electrode (diameter = 2 mm) was chosen as the working electrode in the three-electrode electrochemical testing system, while an Ag/AgCl electrode served as the reference electrode and a Pt wire served as the counter electrode.

For testing the binding between PVI and HSA, based on the dry adsorption method developed previously,¹⁷ a tiny drop of HSA was dropped onto the surface of the bare gold electrode and dried for several hours. 10 mL electrolyte containing 5 mmol L⁻¹ K₃Fe(CN)₆/K₄Fe(CN)₆ and 10 mmol L⁻¹ KCl was used, various volumes of PVI solution were added gradually to the electrolyte and stirred for 2 min, then rested for 2 min before testing. The scan range of the cyclic voltammetry was from 0.00 to 0.06 V and the scan rate was set as 0.05 V s⁻¹.

As F16 can be oxidized at the gold electrode, the bare gold electrode was used without further modification and PBS buffer (pH 7.4) was used as electrolyte.^18,19 The cyclic voltammograms of F16 in the absence and presence of HSA were recorded at the scan rate was 0.05 V s⁻¹. All solutions were purged with pure nitrogen for 10 min before testing.

2.5 Molecule docking investigation

The structures of F16 and PVI were generated by the Sybyl 8.1 package, the molecules were optimized using Tripos Force Field, while the setting of the energy termination gradient was 0.01 kcal mol⁻¹. Molecules of F16 and PVI were charged using the Gasteiger and Marsili method, and the formal charge of the F16 of atom N1 was changed to +1 before computing charging. The crystal structure of HSA was taken from the RCSB Protein Data Bank (PDB ID: 1h9z). The crystal structure was analyzed and fixed with Sybyl 8.1 software before docking. Docking studies were conducted using a Surflex Dock program in the Sybyl 8.1 package. The protomol for HSA were generated by the ligand mode and the Threshold was set at 0.50, while the Bloat was 0. The parameters in the docking work were set as follows: Additional Starting Conformation per Molecule: 20; Angstroms to Expand Search Grid: 6; Max Conformation per Fragment: 20; Max Number of Rotatable Bonds per molecule: 100.

3. Results and discussion

3.1 The effect on the HSA spectra

Fluorescence spectroscopy is a highly sensitive and important research tool in biochemistry and biophysics. In this experiment we chose 295 nm as the excitation wavelength of HSA, since it provides no excitation of the tyrosine residues, both emission and energy transfer to the lone indole side chain would be negligible, and the tryptophan residue was found to be the major contributor of protein fluorescence.²⁰ As Trp is sensitive to the microenvironment, it could be an indicator of protein configuration change and protein–drug interaction.²¹ The emission spectra of HSA in the absence and presence of the two drugs are shown in Fig. 1. The maximum emission wavelength of HSA was near 348 nm, and both F16 and PVI showed almost no emission under the measurement conditions. With the addition of F16 or PVI, we observed a regular and sharp decrease in the HSA fluorescence intensity. Moreover, the maximum emission wavelength shifted from 348 nm to 339 nm with the addition of F16, and from 348 nm to 329 nm for PVI. The blue shift indicated the increasing hydrophobicity around the tryptophan residue, and PVI was likely to have a greater impact on the local dielectric environment of HSA.

Fig. 1 The effect of (a) PVI and (b) F16 on the fluorescence spectra of HSA at 298 K. c(HSA) = 2.0 × 10⁻⁶ mol L⁻¹; c(PVI or F16)/(10⁻⁶ mol L⁻¹), A–L: 0; 0.5; 1.0; 2.0; 3.0; 4.0; 5.0; 6.0; 7.0; 8.0; 9.0; 10.0. The curve at the bottom shows the emission spectrum of the compound only under this condition, c = 2.0 × 10⁻⁶ mol L⁻¹. The insert plots correspond to the Stern–Volmer plots.

Fluorescence quenching can be classified as static and dynamic quenching according to their differing dependence on the temperature and viscosity.²² We further performed fluorescence tests at different temperatures (308 K, 303 K, 298 K and 293 K) and used the Stern–Volmer equation to analyze the data.

F₀ and F are the fluorescence intensities in the absence and presence of the drugs; [Q] is the concentration of F16 or PVI; K_SV is the Stern–Volmer quenching constant; τ₀ stands for the average lifetime of the fluorophore without the quencher and the fluorescence lifetime of biomolecules is 10⁻⁸ s; k_q is the apparent bimolecular quenching rate constant equal to K_SV/τ₀. The plots of the Stern–Volmer equation at different temperatures are shown in Fig. 2, and the values of K_SV and k_q are presented in Table 1. For PVI, the values of K_SV and k_q decreased with the rising temperature, indicating a static quenching. For F16, the values of K_SV and k_q increased as the temperature increased, suggesting the dynamic quenching; however, the value of k_q was much larger than 2.0 × 10¹⁰ L mol⁻¹ s⁻¹, exceeding the maximum quenching constant for dynamic quenching,^23–25 thus indicating a more complicated quenching mechanism.

Fig. 2 The Stern–Volmer plots for the fluorescence quenching of HSA by (a) PVI and (b) F16 at different temperatures.

Table 1 The Stern–Volmer quenching constants of the PVI–HSA and F16–HSA systems at different temperatures (pH 7.4)

T (K)	PVI–HSA			F16–HSA
	KSV (10^4 M^−1)	kq (10^12 M^−1 s^−1)	R^a	KSV (10^4 M^−1)	kq (10^12 M^−1 s^−1)	R
a R is the correlation coefficient.
293	10.57	10.57	0.999	2.416	2.416	0.996
298	9.270	9.270	0.999	2.513	2.513	0.994
303	8.885	8.885	0.999	2.616	2.616	0.999
308	8.176	8.176	0.999	2.730	2.730	0.996

---

As a consequence of the ground-state complex formation, static quenching will result in the perturbation of the absorption spectrum of the fluorophore,²⁶ while dynamic quenching caused by collision will not. One additional method to distinguish static and dynamic quenching is by the careful examination of the absorption spectra of the fluorophore.²² We recorded the absorption spectra of HSA in the absence and presence of F16 or PVI. As shown in Fig. 3, the absorption spectra of HSA in the presence of F16 or PVI (line D) were different from the HSA-only absorption spectra (line A). The changes in the absorption spectra confirmed that the quenching process of F16 as well as PVI was static quenching.

Fig. 3 The effect of (a) PVI and (b) F16 on the UV-vis spectra of HSA. c(HSA) = c(PVI or F16) = 2.0 × 10⁻⁶ mol L⁻¹.

In order to explain the adverse dependence on temperature for the F16–HSA system, we referred to Arrhenius' theory. As previously introduced, the value of k_q for dynamic quenching is less than 2.0 × 10¹⁰ L mol⁻¹ s⁻¹ and is limited by excited molecules, diffusion and the probability of collision,^23–25 while for static quenching there are no such limitations. Higher temperatures typically result in the dissociation of weakly bound complexes, and hence a smaller amount of static quenching and a decreasing k_q. On the other hand, the rate constant k is a positive correlation function of temperature according to Arrhenius' theory, so the value of K_SV tends to increase as the temperature increases. The two factors impact on the quenching process simultaneously. According to Arrhenius' theory, the influence of the temperature on the rate constant can be inferred by the activation energy of the quenching process:

where k_q is the apparent quenching rate constant which is equal to K_SV/τ₀, E_a is the activation energy for the quenching process, the parameter A is the pre-exponential factor and R is the gas constant. The linear relationship of ln k_q vs. 1/T is shown in Fig. SI 3.† The value of E_a for the F16–HSA quenching process was 6.1 kJ mol⁻¹. The high activation energy might give an explanation of the adverse temperature dependence of the F16–HSA quenching process.

After confirming that the processes for PVI–HSA and F16–HSA were static quenching, the quenching data were analyzed according to the modified Stern–Volmer equation.

ΔF is the difference in the fluorescence intensity in the absence and presence of the quencher at concentration [Q], f_a is the fraction of the accessible fluorescence and K_a is the effective quenching constant. The plots of the modified Stern–Volmer equation at different temperatures are shown in Fig. 4, and the values of the binding constant K_a are presented in Table 2. It can be observed that the value of K_a for PVI was larger than that for F16 at the same temperature, which implied that the positive charge might have an adverse effect on the binding with HSA.

Fig. 4 The modified Stern–Volmer plots for the fluorescence quenching of HSA by (a) PVI and (b) F16 at different temperatures.

Table 2 The modified Stern–Volmer quenching constants and thermodynamic parameters of the PVI–HSA system at different temperatures (pH 7.4)

T (K)	Ka (10^4 M^−1)	ΔH (kJ mol^−1)	ΔG (kJ mol^−1)	ΔS (J mol^−1 K^−1)	R^a
a R is the correlation coefficient.
293	9.987	−18.18	−28.04	33.738	0.996
298	9.037		−28.27		0.999
303	7.760		−28.36		0.991
308	7.016		−28.57		0.999

---

The above discussion was based on the optical spectra. In order to avoid method error and give more information, we also carried out cyclic voltammetry (CV) measurements. Notably, the experimental methods utilized for testing F16 and PVI were distinct due to their differing electrochemical properties (see the Materials and methods section). For PVI with HSA (Fig. 5a), the equilibrium constant K_A can be calculated by the Langmuir equation:²⁷

where c is the concentration of PVI, ΔI_p is the current drop and ΔI_p-max represents the maximum current drop. K_A was calculated at the value of 1.125 × 10⁵ L mol⁻¹ at the temperature of 293 K, which is a little larger than the K_a obtained from the fluorescence measurement. This was acceptable as in electrochemical experiments K_A stands for all possible bindings lowering the value of I_p, while in fluorescence measurements only the bindings that occurred around the fluorophore are recorded.

Fig. 5 The cyclic voltammetric curves of PVI or F16 with HSA at 293 K. (a) The PVI–HSA system. c(PVI)/(10⁻⁶mol L⁻¹): A–K: 0; 1.0; 2.0; 3.0; 4.0; 5.0; 6.0; 7.0; 8.0; 9.0; 10.0. (b) F16 only and (c) F16 in the presence of HSA, c(HSA) = 1.0 × 10⁻⁵ mol L⁻¹. c(F16)/(10⁻⁶mol L⁻¹): A–G: 8; 10; 12; 14; 16; 18; 20.

For F16 with HSA (Fig. 5b and c), we referred to the method reported by Fotouhi et al.²⁸ and Qu et al.,²⁹ which assumed that HSA and the drug only produce a single complex HSA·F16_m according to the reaction as follows:

HSA + mF16 ↔ HSA·F16_m (5)

Based on the single complex reaction, the equilibrium constant of F16 can be calculated by the equation.

K_A stands for the equilibrium constant, m represents the total binding number of the ligands, ΔI is the peak current change of F16 in the absence and presence of HSA, and ΔI_max is the maximum current change. K_A was calculated at the value of 1.723 × 10⁴ L mol⁻¹ at the temperature of 293 K, which is slightly less than the value of K_a obtained from the fluorescence measurement, indicating that the positive charge of F16 might have a negative effect on binding with HSA under an electric field.

3.2 Binding force and properties

There are four main types of non-covalent driving forces in drug–HSA binding interactions; hydrogen bonds, van der Waals forces, electrostatic and hydrophobic interactions. They can be determined by the thermodynamic law concluded by Ross and Subramanian.³⁰ The thermodynamic parameters enthalpy change (ΔH) and entropy change (ΔS) can be calculated according to the van't Hoff equation;where K_a represents the effective quenching constant at the corresponding temperature and R is the gas constant. The plots of the van't Hoff equation are shown in Fig. 6. Considering the existence of systematic error and accidental error, a correlation coefficient of R = 1 was not expected. Moreover, K_a were obtained from the emission quenching at the corresponding temperature T, and the values of lnK_a and 1/T with approximate treatment were substituted into the van't Hoff equation, bringing in unavoidable deviation. Therefore, we considered that R > 0.99 was acceptable, which indicated that the enthalpy change (ΔH) remained constant in the discussed temperature range. Based on this, the Gibbs free energy change (ΔG) at different temperatures was obtained as follows.

ΔG = ΔH − TΔS = −RT lnK_a (8)

Fig. 6 The van't Hoff plots of (a) the PVI–HSA system and (b) the F16–HSA system at different temperatures.

The values of ΔH, ΔS and ΔG for PVI and F16 are presented in Table 2 and 3, respectively. The negative Gibbs free energy (ΔG) and positive entropy (ΔS) suggested that both quenching processes occurred spontaneously. However, the impacts of the enthalpy (ΔH) on these two binding processes were different. In the PVI–HSA quenching process, ΔH < 0, indicating that it is both entropy and enthalpy driven, thus electrostatic interactions played the major role and the hydrophobic interactions might also make a contribution. ΔH was positive in the F16–HSA quenching process, suggesting that it was mainly entropy driven, and the driving forces were typical hydrophobic interactions, while hydrogen bonds and van der Waals forces might have minor effects. Notably, the entropy change of the F16–HSA system was extraordinarily large. Taking the previous optical spectra and electrochemistry results into consideration, we should attribute the difference in the thermodynamic process to the delocalized positive charge of F16 affecting the structure and configuration of the ligand.

Table 3 The modified Stern–Volmer quenching constants and thermodynamic parameters of the F16–HSA system at different temperatures (pH 7.4)

T (K)	Ka (10^4 M^−1)	ΔH (kJ mol^−1)	ΔG (kJ mol^−1)	ΔS (J mol^−1 K^−1)	R^a
a R is the correlation coefficient.
293	2.490	11.41	−30.24	142.16	0.998
298	2.653		−30.95		0.999
303	2.916		−31.66		0.996
308	3.017		−32.38		0.996

---

3.3 Binding number and sites

For protein–drug complexes, the binding number (n) can be obtained by using the double-logarithmic equation.

F₀ and F are the fluorescence intensities in the absence and presence of the corresponding drug and K_b stands for the apparent binding constant. The data for n at different temperatures is shown in Table 4. It can be observed that the values of n were close to 1 for both PVI and F16, indicating that the two drugs preferred to form a 1 : 1 complex with HSA.

Table 4 The binding numbers of PVI or F16 with HSA at different temperatures (pH 7.4)

T (K)	PVI–HSA		F16–HSA
	n	R^a	n	R
a R is the correlation coefficient.
293	1.01	0.998	1.18	0.990
298	0.97	0.996	1.17	0.991
303	1.07	0.996	1.06	0.995
308	1.03	0.999	1.08	0.994

---

As is well accepted, HSA is made up of three homologous domains I–III, each composed of two subdomains (A and B). There are two typical binding sites on HSA called site I and site II in subdomains IIA and IIIA.^31,32 To identify the binding site for PVI and F16 on HSA, a site maker competitive experiment was carried out at 308 K. Warfarin and ibuprofen were used as site markers due to their specially binding to site I and site II, respectively.²⁴ In the experiment, HSA with warfarin or ibuprofen was first added at the concentration ratio of 1 : 1. Then, different concentrations of the drugs were added and the fluorescence intensity change was recorded (see in Fig. 7).

Fig. 7 The fluorescence spectra of the PVI–HSA system and the F16–HSA system in the presence of site marker warfarin [a] and ibuprofen [b]. T = 308 K, λ_ex = 295 nm; c(warfarin) = c(ibuprofen) = c(HSA) = 2.0 × 10⁻⁶ mol L⁻¹; c(PVI or F16)/(10⁻⁶ mol L⁻¹), A–L: 0; 0.5; 1.0; 2.0; 3.0; 4.0; 5.0; 6.0; 7.0; 8.0; 9.0; 10.0.

The binding constants of the drugs with HSA in the absence and presence of the site markers were calculated by the modified Stern–Volmer equation and were compared. For the PVI–HSA system, the binding constant was 7.016 × 10⁴ L mol⁻¹, and this changed to 5.278 × 10⁴ L mol⁻¹ when warfarin was added. In the presence of ibuprofen the binding constant was 11.537 × 10⁴ L mol⁻¹. The result suggested that the major competitor of PVI was warfarin. In other words, PVI mainly bound to site I on HSA.

The result for F16 was similar. The binding constant of the F16–HSA system was 3.017 × 10⁴ L mol⁻¹, which changed to 2.541 × 10⁴ L mol⁻¹ and 3.901 × 10⁴ L mol⁻¹ in the presence of warfarin and ibuprofen, respectively. The decrease of the binding constant caused by the addition of warfarin indicated that F16 was mainly bound to site I. Therefore, we conclude that the main binding site for F16 and PVI on HSA is Sudlow's site I (subdomain IIA).

3.4 Conformation change

3.4.1 Synchronous fluorescence spectroscopy. PVI or F16 may induce a conformation change of HSA when binding to it, and these conformation changes should be significant in deciding the binding properties. We utilized synchronous fluorescence spectroscopy, circular dichroism spectra and three-dimensional fluorescence spectra to provide such information. As it is known, the position shift of the maximum emission wavelength corresponds to the change of polarity around the chromophore molecule.³³ In synchronous fluorescence spectroscopy, the characteristic information of the tryptophan residue can be obtained when the D-value (Δλ) between the excitation and emission wavelength is set as 60 nm.³⁴ The synchronous fluorescence spectra of HSA in the absence and presence of F16 or PVI are shown in Fig. SI 4.† It can be seen that the synchronous fluorescence intensity of HSA decreased regularly and acutely with the addition of F16 or PVI, while the maximum emission wavelength was slightly blue shifted, indicating the microenvironment perturbation of the tryptophan residue.

3.4.2 Circular dichroism (CD) spectra. The circular dichroism (CD) spectra is one of the most useful measurements in the investigation of the protein conformation.³⁵ In the CD spectra of HSA, two negative bands exhibit at 208 nm and 222 nm as the characteristic of the α-helix in the advanced structure of the protein.^36,37 Therefore, a change of the two bands represents the conformation conversion of the protein. As shown in Fig. 8, the ellipticity of the negative bands exhibited a decrease in the presence of PVI or F16, suggesting a decrease in the α-helix content. As calculated, the α-helix content reduced from 63.7% to 50.0% for PVI and from 63.7% to 60.5% for F16 when the concentration ratio of the drug and HSA was 1 : 1, indicating the changes in the protein secondary structure.³⁸

Fig. 8 The effect of (a) PVI and (b) F16 on the CD spectra of HSA. c(HSA) = 2.0 × 10⁻⁶ mol L⁻¹; c(PVI or F16)/(10⁻⁶ mol L⁻¹) (A–D): 0; 1; 2; 4.

As is well accepted, the biological activities of a protein are closely related to its secondary structure. The decrease of the α-helix content indicated that the normal structure of HSA had been disrupted. In other words, the two compounds might alter the biological activities of HSA by perturbing its conformation.

3.4.3 Three-dimensional fluorescence spectra. The three-dimensional fluorescence spectra of HSA contains three main fluorescence peaks. The peak on the left is the first-ordered Rayleigh scattering peak, whose emission wavelength equals the excitation wavelength, besides which there are two characteristic peaks named peak 1 and peak 2. Peak 1 (λ_ex = 230.0 nm and λ_em = 342.0 nm) represents the fluorescence characteristic of the polypeptide backbone structure,³⁹ while peak 2 (λ_ex = 280.0 nm and λ_em = 342.0 nm) mainly stands for the behavior of the tryptophan and tyrosine residues.⁴⁰ As seen in Fig. 9 and SI 5,† both peak 1 and 2 showed an intensity decrease in the presence of PVI or F16. For peak 1, the quenching ratio of the fluorescence intensity was 45.55% for the PVI–HSA system and 20.98% for the F16–HSA system. The large decrease indicated a disturbance of the polypeptide backbone structure of HSA caused by PVI or F16, which is in accordance with the CD spectra. In addition, the quenching ratio of peak 2 was 7.98% for the PVI–HSA system and 3.40% for the F16–HSA system, indicating that the two compounds also perturbed the microenvironment of the tryptophan residues. Combining the results of the synchronous fluorescence spectroscopy and the site maker competitive experiment, we hypothesized that the binding site in HSA was close to the residue of tryptophan. In addition, PVI absolutely had more influence on the structure of HSA compared to F16.

Fig. 9 The three-dimensional fluorescence spectra of (a) HSA only, (b) the PVI–HSA system and (c) the F16–HSA system. c(HSA) = c(PVI or F16) = 2 × 10⁻⁶ mol L⁻¹.

3.5 Molecule modeling study

As mentioned above, we have made a hypothesis that the binding site of PVI and F16 in HSA is close to the residue of tryptophan, therefore the molecule modeling method was employed to identify it and offer the accurate configuration of the two compounds in a visualized description.

The details of the molecule modeling results are shown in Fig. 10 and 11, and SI 6 and 7.† Fig. 10a and b exhibit the exact binding site of the two compounds on HSA. We can see clearly that both PVI and F16 are bound to HSA at Sudlow's site I, which is consistent with the result of the site marker competitive experiment. Trp214, the only tryptophan residue in HSA, had been involved in the binding site, thus the change of the microenvironment of the tryptophan residue altered by the compounds could be understood (see in Fig. SI 6 and 7†). The distances from Trp214 to the two compounds were reasonable so that the energy might be transported from the excited tryptophan to the drug by a dipole–dipole interaction, resulting in fluorescence quenching.

Fig. 10 The modeling results of the PVI–HSA system and the F16–HSA system. (a) and (b): the binding site of (a) PVI and (b) F16 in HSA. The two compounds are presented in a space-filling model. (c) and (d): the conformation of (c) PVI and (d) F16 in the binding site. The residues lying around the binding sites are presented in lines and are colored by a second structure. (e) The surface electrostatic potential map of HSA with PVI (from red to blue: negatively charged to positively charged).

Fig. 11 The surface model of (a) PVI and (b) F16 in the binding site of HSA. The surface of the HSA binding site is shown in purple and the surface of the compound is shown in green. The obstacles of the HSA–PVI system are shown in orange, while the outside obstacles of the HSA–F16 system are shown in orange and the inside one is shown in red.

From Fig. 10c we can see that the amino acid residues around the binding site of the PVI–HSA system included Lys199, Ala210, Phe211, Ala213, Trp214, Ala215, Arg218, Leu219, Arg222, Leu238, His242 and Ala291, forming a hydrophobic cavity (for detailed information, see Fig. SI 6†). From the surface electrostatic potential map (Fig. 10e) electrostatic forces would be the main driving factor for this binding interaction. The hydrophobic cavity suggested that hydrophobic interactions increased the binding stability.

For the F16–HSA binding system, the details changed as the molecule structure changed. The residues around the binding site were comprised of more than 10 amino acids, Phe211, Trp214, Ala215, Ala217, Leu219, Arg222, Phe223, Leu238, His242, Trp244, Leu260, Ala261, Leu264, Ser287, Ile290 and Ala291, forming two hydrophobic loops surrounding the indole group and the pyridine group (see Fig. 10d and SI 7†). Furthermore, as shown in Fig. SI 7e,† there was a hydrogen bond: NE2 in His242 formed a hydrogen bond with H10 in F16. According to the analyzed details, we suggested that the main forces for this binding process were hydrophobic interactions, and hydrogen bonds might have a minor contribution. These results were exactly corresponding to the analysis of the thermodynamic parameters previously.

From the optical spectra experiment, we concluded that the quenching mechanisms of F16 and PVI with HSA were different. The PVI–HSA system was a typical static quenching while the F16–HSA system was an unusual static quenching. The difference of the quenching mechanism could be explained by the surface modes. From Fig. 11 we can see that the binding sites for F16 and PVI were located deep in the body of HSA instead of being on the surface of the protein; moreover, the binding site of F16 was buried more deeply than that of PVI. The channel for F16 and PVI to settle in HSA could be seen in Fig. 11a and SI 6e,† respectively. In addition, from the figures we can find that some amino acid residues acted like stumbling blocks as they blocked the entrance. For PVI, the channel was probably comprised of Lys195, Lys199, His242 and Glu292, while Gln196 and His288 acted as stumbling blocks. Similarly, for F16 the entrance channel was made up of Phe211, Trp214, Arg218, His242 and Asp451, and the stumbling blocks were Lys195, Tyr452 and Ala291 (shown in Fig. 11b and SI 6f†). Comparing the two systems, the binding channel of PVI was just blocked by two residues while F16 was blocked by three, inducing a higher activation energy requirement for F16–HSA binding. The high activation energy led to the unusual static quenching and the unusual temperature dependence of the Stern–Volmer quenching constant.

In summary, the consequence obtained by molecule modeling were consistent with the experiment results, indicating that the 3D structure of HSA and the modeling method were receivable and credible for our work. Based on this, the deep insights into the structure–quenching mechanism relationship between HSA and F16 or PVI were explored, and the difference between the two binding process were presented successfully.

4. Conclusions

In this article, in order to explore the effect of the positive charge on the thermodynamic properties of F16, we synthesized F16 and its precursor compound PVI, and their interactions with HSA have been investigated. The differences in the quenching mechanism, binding constants and binding force indicate that they have different binding models with serum albumin. The quenching mechanism of PVI is static quenching driven by electrostatic interactions, while F16 is an untypical static quenching driven by hydrophobic interactions. We suggest that this untypical static quenching may be caused by the high activation energy requirement in the binding process, which is generated by the specific structural obstacle. By carefully analysing all the results and comparing the two compounds, we conclude that the difference in the electrical properties may be responsible for the different interactions with HSA. Despite playing an important role in the antiproliferative ability of F16, the positive charge may have a negative influence on the thermodynamic performance. Considering HSA dominating the disposition and transportation of various endogenous and exogenous compounds, the pharmaceutical effect of F16 in vivo may be influenced by the binding with serum albumin. This study revealed the different binding mechanism of F16 and its precursor compound with HSA, which can be a helpful guidance for the pharmacodynamic study of F16 and other DLCs.

Acknowledgements

We gratefully acknowledge financial support from the 973 program from the Chinese Ministry of Science and Technology (2012CB720600), National Natural Science Foundation of China (21225313), Program for Changjiang Scholars, Innovative Research Team in University (IRT1030) and National Science Foundation of Wuhan University (2042014Kf0287).

References

1. A. Szewczyk and L. Wojtczak, Mitochondria as a pharmacological target, Pharmacol. Rev., 2002, 54, 101–127.
2. D. Pathania, M. Millard and N. Neamati, Opportunities in discovery and delivery of anticancer drugs targeting mitochondria and cancer cell metabolism, Adv. Drug Delivery Rev., 2009, 61, 1250–1275.
3. A. G. Kruglova, V. V. Teplovaa and N. L. Saris, The effect of the lipophilic cation lucigenin on mitochondria depends on the site of its reduction, Biochem. Pharmacol., 2007, 74, 545–556.
4. J. S. Modica-Napolitano and J. R. Aprille, Delocalized lipophilic cations selectively target the mitochondria of carcinoma cells, Adv. Drug Delivery Rev., 2001, 49, 63–70.
5. S. Davis, M. J. Weiss, J. R. Wong, T. J. Lampidis and L. B. Chen, Mitochondrial and plasma membrane potentials cause unusual accumulation and retention of rhodamine 123 by human breast adenocarcinoma-derived MCF-7 cells, J. Biol. Chem., 1985, 260, 13844–13850.
6. J. S. Modica-Napolitano and J. R. Aprille, Basis for the selective cytotoxicity of rhodamine 123, Cancer Res., 1987, 47, 4361–4365.
7. J. J. Min, S. Biswal, C. Deroose and S. S. Gambhir, Tetraphenylphosphonium as a novel molecular probe for imaging tumors, J. Nucl. Med., 2004, 45, 636–643.
8. E. Galeano, E. Nieto, A. I. García-Pérez, M. D. Delgado, M. Pinilla and P. Sancho, Effects of the antitumoural dequalinium on NB4 and K562 human leukemia cell lines: mitochondrial implication in cell death, Leuk. Res., 2005, 29, 1201–1211.
9. Y. S. Kim, C. T. Yang, J. Wang, L. Wang, Z. B. Li, X. Chen and S. Liu, Effects of targeting moiety, linker, bifunctional chelator, and molecular charge on biological properties of ⁶⁴Cu-labeled triphenylphosphonium cations, J. Med. Chem., 2008, 51, 2971–2984.
10. V. R. Fantin, M. J. Berardi, L. Scorrano, S. J. Korsmeyer and P. Leder, A novel mitochondriotoxic small molecule that selectively inhibits tumor cell growth, Cancer Cell, 2002, 2, 29–42.
11. V. R. Fantin and P. Leder, F16, a mitochondriotoxic compound, triggers apoptosis or necrosis depending on the genetic background of the target carcinoma cell, Cancer Res., 2004, 64, 329–336.
12. D. C. Carter and J. X. Ho, Structure of serum albumin, Adv. Protein Chem., 1994, 45, 153–196.
13. Y. J. Hu, Y. Liu and X. H. Xiao, Investigation of the interaction between berberine and human serum albumin, Biomacromolecules, 2009, 10, 517–521.
14. D. Shcharbin, E. Pedziwiatr, L. Chonco, J. F. Bermejo-Martin, P. Ortega, F. J. Mata, R. Eritja, R. Gómez, B. Klajnert, M. Bryszewska and M. A. Muñoz-Fernandez, Analysis of interaction between dendriplexes and bovine serum albumin, Biomacromolecules, 2007, 8, 2059–2062.
15. E. Lázaro, P. J. Lowe, X. Briand and B. Faller, New approach to measure protein binding based on a parallel artificial membrane assay and human serum albumin, J. Med. Chem., 2008, 51, 2009–2017.
16. N. A. Kratochwil, W. Huber, F. Müller, M. Kansy and P. R. Gerber, Predicting plasma protein binding of drugs: a new approach, Biochem. Pharmacol., 2002, 64, 1355–1374.
17. R. Li, F. Jiang, Q. Xiao, J. Li, X. Liu, Q. Yu, Y. Liu and C. Zeng, Microcalorimetric, spectroscopic and microscopic investigation on the toxic effects of CdTe quantum dots on Halobacterium halobium R1, Nanotechnology, 2010, 21, 475102.
18. C. Q. Xiao, F. L. Jiang, B. Zhou, R. Li and Y. Liu, Interaction between a cationic porphyrin and bovine serum albumin studied by surface plasmon resonance, fluorescence spectroscopy and cyclic voltammetry, Photochem. Photobiol. Sci., 2011, 10, 1110–1117.
19. C. Zhou, Y. Dong, Z. Li, X. Xu and Z. Liu, Electrochemistry of magnolol and interaction with DNA, J. Electroanal. Chem., 2010, 642, 115–119.
20. M. Dockal, D. C. Carter and F. Rüker, Conformational transitions of the three recombinant domains of human serum albumin depending on pH, J. Biol. Chem., 2000, 275, 3042–3050.
21. X. M. He and D. C. Carter, Atomic structure and chemistry of human serum albumin, Nature, 1992, 358, 209–215.
22. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New York, 2006, 3rd edn, pp. 277–330.
23. B. K. Sahoo, K. S. Ghosh and S. Dasgupta, Molecular interactions of isoxazolcurcumin with human serum albumin: spectroscopic and molecular modeling studies, Biopolymers, 2009, 91, 108–119.
24. Y. J. Hu, Y. O. Yang, C. M. Dai, Y. Liu and X. H. Xiao, Site-selective binding of human serum albumin by Palmatine: spectroscopic approach, Biomacromolecules, 2010, 11, 106–112.
25. F. F. Tian, J. H. Li, F. L. Jiang, X. L. Han, C. Xiang, Y. S. Ge, L. L. Li and Y. Liu, The adsorption of an anticancer hydrazone by protein: an unusual static quenching mechanism, RSC Adv., 2012, 2, 501–513.
26. E. Froehlich, J. S. Mandeville, C. J. Jennings, R. Sedaghat-Herati and H. A. Tajmir-Riahi, Dendrimers bind human serum albumin, J. Phys. Chem. B, 2009, 113, 6986–6993.
27. S. Q. Liu, M. L. Cao and S. L. Dong, Electrochemical and Ultraviolet-visible spectroscopic studies on the interaction of deoxyribonucleic acid with vitamin B₆, Bioelectrochemistry, 2008, 74, 164–169.
28. L. Fotouhi, S. Banafsheh and M. M. Heravi, Electrochemistry of the interaction of furazolidone and bovine serum albumin, Bioelectrochemistry, 2009, 77, 26–30.
29. F. Qu, N. Q. Li and Y. Y. Jiang, Electrochemical studies of NiTMpyP and interaction with DNA, Talanta, 1998, 45, 787–793.
30. P. D. Ross and S. Subramanian, Thermodynamics of protein association reactions: forces contributing to stability, Biochemistry, 1981, 20, 3096–3102.
31. G. Sudlow, D. J. Birkett and D. N. Wade, Further characterization of specific drug binding sites on human serum albumin, Mol. Pharmacol., 1976, 12, 1052–1061.
32. Y. J. Hu, Y. Liu, J. B. Wang, X. H. Xiao and S. S. Qu, Study of the interaction between monoammonium glycyrrhizinate and bovine serum albumin, J. Pharm. Biomed. Anal., 2004, 36, 915–919.
33. U. Kragh-Hansen, F. Hellec, B. de Foresta, M. le Maire and J. V. Møller, Detergents as probes of hydrophobic binding cavities in serum albumin and other water-soluble proteins, Biophys. J., 2001, 80, 2898–2911.
34. J. B. F. Lloyd, The nature and evidential value of the luminescence of automobile engine oils and related materials: part I. synchronous excitation of fluorescence emission, J. Forensic Sci. Soc., 1971, 11, 83–94.
35. J. N. Miller, Recent advances in molecular luminescence analysis, Proc. Anal. Div. Chem. Soc., 1979, 16, 203–208.
36. D. M. Charbonneau and H. A. Tajmir-Riahi, Study on the interaction of cationic lipids with bovine serum albumin, J. Phys. Chem. B, 2010, 114, 1148–1155.
37. P. F. Qin, R. T. Liu, X. R. Pan, X. Y. Fang and Y. Mou, Impact of carbon chain length on binding of perfluoroalkyl acids to bovine serum albumin determined by spectroscopic methods, J. Agric. Food Chem., 2010, 58, 5561–5567.
38. V. Oliveri and G. Vecchio, A novel artificial superoxide dismutase: non-covalent conjugation of albumin with a Mn^III salophen type complex, Eur. J. Med. Chem., 2011, 46, 961–965.
39. Y. Z. Zhang, B. Zhou, X. P. Zhang, P. Huang, C. H. Li and Y. Liu, Interaction of malachite green with bovine serum albumin: determination of the binding mechanism and binding site by spectroscopic methods, J. Hazard. Mater., 2009, 163, 1345–1352.
40. Y. Wang, B. Tang, H. Zhang, Q. Zhou and G. Zhang, Studies on the interaction between imidacloprid and human serum albumin: spectroscopic approach, J. Photochem. Photobiol., B, 2009, 94, 183–190.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra46997b

‡ Both authors have contributed equally to the paper.

---