Recognition and binding of voltage-dependent anion channel-1 with ATP and NADH by spectroscopic analysis and molecular docking

Abstract

The beta-barrel mitochondrial outer membrane protein voltage-dependent anion channel-1 (VDAC-1) plays an important role in the regulation of mitochondrial functions and the control of apoptosis. It appears to be a convergence point for a diversity of cell survival and cell death signals, which was mediated by its association with various ligands and proteins. In this study, the binding behavior of VDAC-1 with adenosine triphosphate (ATP) and the reduced form of nicotinamide adenine dinucleotide (NADH) were studied by spectroscopic analysis and molecular docking. The results showed that ATP and NADH bound to VDAC-1 in equimolar amounts. The association constants of VDAC-1 with ATP and NADH were 1.93 × 10⁴ and 1.15 × 10⁴ L mol⁻¹ at 20 °C, respectively. The data from the spectroscopic analysis and molecular docking studies revealed that ATP and NADH could bind to the nucleotide binding sites of VDAC-1 through electrostatic interactions. The structural difference and flexible feature in ligands could affect the binding affinity to VDAC-1, which provides useful help in the design and discovery of lead compounds for the treatment of VDAC-1 related diseases.

---

Introduction

Membrane proteins such as receptors and porins are essential to all life, and they take up thirty percent of the open reading frames in known genomes.^1–3 The targets of many pharmaceutical interventions are membrane proteins. Membrane proteins modulate vital functions such as recognition and transduction of signals, transport of ions and molecules, control of transmembrane potential, generation and transduction of energy, catalysis of chemical reactions and so on.^4–6 In the current validated medicines, about fifty to sixty percent of them are targeted to membrane proteins and they act as the principal target for new drug discovery.^7,8 Therefore, increasing recognition is being given to characterizing the membrane protein–ligand interaction.

The voltage-dependent anion channel (VDAC) protein encodes for a protein product length of 283. It belongs to the membrane proteins group and is also known as mitochondrial porin. VDAC serves as the primary avenue for anions, cations, and various metabolites including substrates and nucleotides between the mitochondrion and cytoplasm.⁹ Currently three isoforms of VDAC were found in human body, and they have been simply numbered in the order of their discovery as VDAC-1, VDAC-2 and VDAC-3.¹⁰ The three human VDAC isoforms show extensive similarities at gene, nucleotide and amino acid level.^11,12 The most information available on VDAC comes from the isoform VDAC-1 with the structure of a 19-strand β-barrel and a flexible N-terminal α-helix in the central pore adjacent to the wall (between residues 5–16).¹³ The channel protein has a 27 Å wide opening facing both cytosol and inter membrane space, but tapers to 14 Å wide in the center of the channel due to the presence of the helical segment.¹⁴ The N-terminus of VDAC-1 is 25 amino acids long and very flexible. Recently published 3D structure data of VDAC-1 showed that the N-terminus residues were within the pore.¹⁵ However, independent studies showed that it could also stretch out of the pore.¹⁶ The N-terminal α-helical segment is proposed to be involved in VDAC-1 gating function, where it could be acting as the voltage sensor and possibly regulating the conductance of ions and metabolites through the VDAC-1 pore.¹⁶

A survey of literatures pointing to the function of VDAC-1 in cell life and death highlights these functions in relation to cancer. VDAC-1 assumes to control the metabolic cross-talk and energy products between the mitochondria and the rest of the cells.^17,18 Accumulated evidence suggests that the functions of VDAC-1 are modulated by various physiological ligands, such as adenine nucleotides, Ca²⁺, the reduced form of nicotinamide adenine dinucleotide (NADH), glutamate, and non-physiological compounds, such as Koenig's polyanion, ruthenium red, dicyclohexylcarbodiimide, and 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid.^13,18–21 In this regulation, the state of the VDAC-1 channel (“open” or “closed”) serves as a key determinant: regulation of metabolic substances through the channel by bound them controlling transport activity would be of great importance. Unlike a large number of studies available for the channel functions of VDAC-1, the binding properties of some regulatory ligands to VDAC-1 have not received much attention. Therefore, studying the interaction between VDAC-1 and its regulatory ligands would be imperative and could give some mechanic information of the permeability of VDAC-1. ATP is the most commonly used “energy currency” of cells from most organisms, while NADH can act as a primary enzyme in the production of ATP. The mitochondria are the principal source of ATP generation where a high number of ATP molecules must exit the organelle-primarily through VDACs.¹⁴ NADH was found to regulate the gating of mammalian, fungal, and plant VDACs, and adenosine triphosphate (ATP) was shown to modify Neurospora crassa VDAC channel conductance in a biphasic manner, leading to the suggestion that VDAC contains at least one nucleotide-binding site (NBS).¹⁴ Due to the structural similarities of the two ligands (Scheme 1) and the important regulatory functions for them binding to VDAC-1, the interactions between VDAC-1 and the two ligands have been investigated using fluorescence spectroscopy and molecular docking in the present work.

Scheme 1 Chemical structures of ATP (A) and NADH (B).

Materials and methods

Chemicals and instruments

ATP and NADH (HPLC grade), ovolecithin, yeast powder, peptone and guanidine hydrochloride were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ni²⁺ Sepharose 6 Fast Flow pre-packed column (10 × 58 mm) was purchased from GE Healthcare Life Sciences (Uppsala, Sweden).

The protein purification experiments were performed in an ÄKTA10 system with UNICORN 5.20 software (GE Healthcare). The Nathex-47 high-pressure filtration extruder was obtained from NTI Machinery Co., Ltd (Shanghai, China). Fluorescence measurements were performed with a Hitachi model F-4500 spectrofluorimeter (Tokyo, Japan) equipped with a 1.0 cm quartz cell, a 150 W xenon lamp source and a thermostat bath that keep temperature constant within 0.1 °C. The widths of both the excitation and emission slits were set at 10.0 nm. Absorption spectral measurements were carried out using a Hitachi U-3310 model UV-Visible spectrophotometer. Quartz cuvettes of path length 1 cm were used to record the absorption spectra. Time-resolved experiments were performed at room temperature on an Edinburgh Instruments FLS 920 time-correlated single photon counting fluorescence spectrometer with a picosecond pulsed diode source (EPLED-270). The wavelength, pulse width and bandwidth of the picosecond pulse diode source were 273.2 nm, 870.5 ps and 9.8 nm, respectively.

VDAC-1 expression and sample preparation from inclusion bodies

The cDNA for human VDAC-1 with the corresponding GenBank accession number NM_003374 was obtained from Sangon Biotech (Shanghai Sangon Co., Ltd., China). A C-terminally 6× His-tagged version of VDAC-1 was cloned by standard PCR methods into the Nco I/Xho I sites of pET28b. VDAC-1 was expressed in BL21(DE3) cells with kanamycin resistance and the cells were cultured until the OD values reached to 0.4–0.6, then induced with 1.0 mmol L⁻¹ IPTG at 30 °C for 6 h. The cells containing His-tagged VDAC-1 were collected by centrifugation at 6000 × g for 30 min at 4.0 °C with the supernatant discarded, and 30.0 g E. coli bacteria was obtained. The VDAC-1 was expressed in the form of inclusion bodies, so the extract of protein of interest should be prepared to be the soluble protein under denaturing conditions. Sodium phosphate buffer (PBS, 20.0 mmol L⁻¹, 300.0 mL) containing 500.0 mmol L⁻¹ NaCl (buffer A, pH 7.4) was added to suspend E. coli, and disintegrated using ice-water bath through an ultrasonic processor at 400 W with an endurance of 10 s for each time. The procedure was repeated 120 times with an interval span of 15 s between each treatment. The resulted suspension was centrifuged at 10 000 × g for 30 min to obtain E. coli extract for next use. The extract was sequentially washed twice by washing buffer A (10.0 mmol L⁻¹ Tris–HCl, 5.0 mmol L⁻¹ EDTA, 0.5% Triton X-100, pH 8.0) and washing buffer B (10.0 mmol L⁻¹ Tris–HCl, 5.0 mmol L⁻¹ EDTA, 0.5 mmol L⁻¹ guanidine hydrochloride, pH 8.0) to exclude a small amount of hybrid proteins. After centrifugation at 10 000 × g for 40 min, the precipitation was dissolved by 200 mL buffer C (10.0 mmol L⁻¹ Tris–HCl, 7.0 mol L⁻¹ guanidine hydrochloride, 10.0 mmol L⁻¹ β-mercaptoethanol, pH 8.0), then filtered through a 0.45 μm membrane and the filtrate was stored at 4 °C for further use.

Denaturing purification of VDAC-1

A Ni²⁺ Sepharose 6 Fast Flow pre-packed column (10 × 58 mm) was used to purify VDAC-1 under denaturing conditions. The absorbance of the eluate was continuously monitored at 280 nm. The column was first equilibrated with buffer D (10.0 mmol L⁻¹ Tris–HCl, 7.0 mol L⁻¹ guanidine hydrochloride, 500.0 mmol L⁻¹ NaCl, pH 8.0) with a flow rate of 2.0 mL min⁻¹, and the above prepared lysate of the VDAC-1 extract was loaded on the column at the speed of 1.0 mL min⁻¹. After re-equilibrated with buffer D, the column was sequentially washed by buffer E (10.0 mmol L⁻¹ Tris–HCl, 500.0 mmol L⁻¹ NaCl, pH 8.0) containing the gradually reducing concentrations of guanidine hydrochloride with a flow rate of 2.0 mL min⁻¹, i.e. 5.0 mol L⁻¹, 3.0 mol L⁻¹, 1.0 mol L⁻¹ and zero mole guanidine hydrochloride, each washing step was lasted for 30 min. Until the absorbance of 280 nm reached to baseline, the column was washed by 10% buffer F (buffer E in the presence of 500.0 mmol L⁻¹ imidazole, pH 8.0) and 90% buffer E at the speed of 2.0 mL min⁻¹. After flushing for 30 min, the column was eluted using 50% buffer F and 50% buffer E to achieve the VDAC-1 extract. The purified VDAC-1 with high concentration imidazole and NaCl buffer was desalted by dialysis against 4.0 L of buffer G (10.0 mmol L⁻¹ Tris–HCl, pH 7.4) in a 10 000 MWCO dialysis membrane. VDAC-1 was purified more than 95% according to the analysis made by SDS-PAGE in which the protein was present as a single band of ∼35 kDa. The concentration of the protein was determined by Lowry's method²² and the standard curve was plotted using bovine serum albumin.

VDAC-1 refolding

The liposome was prepared according to the protocols proposed by Ben-Hail et al.^23,24 In detail, 120 mg ovolecithin was dissolved in 60.0 mL mixture of methanol and dichloromethane (v : v = 1 : 2) and added to a 100 mL round-bottom flask with a long extension neck, the solvent was removed under reduced pressure by a rotary evaporator with the speed of 100 rpm at 40 °C, until a homogeneous monolayer phospholipid was formed on the surface of the flask, the evaporator was still worked for another 3 hours to remove the rudimental organic phase. The aqueous phase 40.0 mL of 5.0 mmol L⁻¹ phosphate buffer (pH 7.4) was added to the flask, the system was kept continuously under nitrogen and the resulting two-phase system was sonicated at least 30 min in a bath-type sonicator until the mixture became either a clear one-phase dispersion or a homogeneous opalescent dispersion. The resulting mixture was passed through the Nathex-47 high-pressure filtration extruder at least three times with a 100 nm pore size filter to obtain the liposome with particle size of about one hundred nanometers. A known concentration of VDAC-1 was added to the filtered liposome until the protein–liposome ratio reached to 1 : 60. The mixture was incubated in a water bath at 25 °C for overnight, sonicated for 30 min and freeze-thawing for at least three cycles to obtain the VDAC-1–liposome. Dynamic Light Scattering (DLS) was used to detect the size of liposome and VDAC-1 proteoliposome samples. Malvern Instruments Zetasizer Nano ZS90 (Zetasizer Ver. 6.11) was used to draw the phase analysis light scattering.

Fluorescence spectroscopy measurement

The interaction of VDAC-1 with ATP and NADH was investigated by fluorometric titration. In detail, one milliliter of VDAC-1 proteoliposome (2.0 × 10⁻⁶ mol L⁻¹) was first added to the quartz cell (1.0 cm × 1.0 cm × 3.0 cm), and then titrated by successive additions of 1.0 × 10⁻² mol L⁻¹ ATP or 1.0 × 10⁻² mol L⁻¹ NADH with a 10.0 μL microsyringe to obtain a series of final concentrations of ATP/VDAC-1 and NADH/VDAC-1 proteoliposome mixtures. The titrations were operated manually with gentle stirring. The resulting molar ratio of ATP to VDAC-1 proteoliposome were 5, 10, 20, 30, 40, 50, 65, 80, 95, 110 and 125, respectively; while the molar ratio of NADH to VDAC-1 proteoliposome were 1, 4, 7, 10, 13, 16, 19, 22, 25 and 30, respectively. The liposome without VDAC-1 was used as a control with all the other procedures settled the same as the experimental group. Fluorescence emission spectra were measured at 10, 15, 20, 25 and 30 °C. The emission spectra were recorded from 300 to 400 nm with an excitation wavelength of 278 nm.

Molecular docking

The human VDAC-1 was used as macromolecular receptor in the docking simulation, whose crystal structure is available in Protein Data Bank, from which the three dimensional structure of VDAC-1 was imported (PDB code: 2JK4).¹⁵ The three-dimensional (3D) structures of ATP and NADH were generated by ChemDraw Ultra 8.0 and energies were minimized with the CHARMm program. Libdock program, implemented in the software platform of Discovery Studio 2.5 (DS 2.5, Accelrys Software Inc., San Diego, CA), was used for the docking study.²⁵ The binding sphere is selected as a radius of 8.825 Å, with the center coordinates of x: 27.101, y: 4.678, and z: −2.061 according to the previous proposed NBSs of rat VDAC-1 (ref. 14) and the binding sites information of human VDAC-1 predicted by DS 2.5. Ten automatically achieved docking modes were obtained with Libdock scoring results of the ligand–VDAC-1 combination.

Results and discussion

Characterization of VDAC-1 proteoliposome

VDAC-1 is a mitochondrial outer membrane protein. Membranes can not only provide native environment for membrane proteins, but also maintain the functional activity and stability of them.²⁶ Therefore, it is vital to reconstitute VDAC-1 into a phospholipid membrane structure. It has been proved that it is difficult to set up a stable and electrically quiet environment for the ion channel in planar lipid bilayer. Reconstitution of ion channels into proteoliposomes has proven to be a better alternative.²⁷ The VDAC-1 proteoliposome were produced by constituting VDAC-1 into liposomes. The sizes of liposome and VDAC-1 proteoliposome were detected by DLS with intensity weighted size distributions (Fig. 1). Both liposome and VDAC-1 proteoliposome showed rather narrow size distribution. The mean sizes of liposome and VDAC-1 proteoliposome were 123.4 nm and 382.0 nm, respectively. It is apparent that if VDAC-1 proteoliposome formed, the structure of the liposome should be changed. From the size change of liposome and proteoliposome, it can be concluded that VDAC-1 had been inserted into the liposomes. Moreover, the size of VDAC-1 proteoliposome showed high agreement with the Agrc proteoliposome (MW: 47 kDa; size: 301 nm) constructed by Wang et al.,²⁸ indicating a successful construction of VDAC-1 proteoliposome in this study.

Fig. 1 Characterization of liposome (A) and VDAC-1–liposome (B) using dynamic light scattering.

Fluorescence quenching of VDAC-1 induced by ATP and NADH

It is well studied that the intrinsic emission property of proteins comes mainly from the tryptophan (Trp), tyrosine (Tyr) and phenylalanine (Phe) residues.²⁹ The emission maximum wavelengths of the three residues were at 348, 303 and 282 nm, respectively.²⁹ The protein–ligand interactions often lead to changes in the intrinsic emission property of proteins. In this work, the effects of ATP and NADH on the intrinsic emission property of VDAC-1 were determined using emission spectroscopy. Fig. 2 represents the effect of increasing ATP and NADH concentration on the VDAC-1 proteoliposome fluorescence emission spectra in phosphate buffer (pH 7.0) by subtracting the control group at 20 °C. When exited at 278 nm, the emission spectra of VDAC-1 in the absence of ATP and NADH show the emission maximum at 333.3 nm. Considering the general emission property of Trp, Tyr and Phe, the excitation wavelength of 278 nm was chosen and the resulting emission spectrum is ascribed to the three kinds of amino acid residues. The fluorescent intensity decreases with the increasing of ATP and NADH concentration. However, the rate of quenching decreased with the increase in ATP and NADH concentration. When the ratio of VDAC-1 proteoliposome to ATP reached to 1 : 125, the fluorescence intensity variation of the system was almost remained unchanged. Therefore, the limit ratio of ATP to VDAC-1 proteoliposome which could cause the fluorescence quenching of VDAC-1 proteoliposome was 125 : 1. At this ratio, about 75.4% quenching of the fluorescence emission has happened. In addition, a 6 nm red shift, from 333.2 to 339.2 nm, was observed at the maximum emission wavelength of VDAC-1 proteoliposome. Similar phenomenon can be found in the system of VDAC-1 proteoliposome with NADH as well. In consequence, it can be concluded that VDAC-1 could interact with ATP and NADH.

Fig. 2 Fluorescence emission spectra of VDAC-1 proteoliposome in the presence of different concentrations of ATP (A) and NADH (B). (From the top to bottom, the molar ratio of ATP to VDAC-1 proteoliposome were 0, 5, 10, 20, 30, 40, 50, 65, 80, 95, 110 and 125, respectively; and the molar ratio of NADH to VDAC-1 proteoliposome were 0, 1, 4, 7, 10, 13, 16, 19, 22, 25 and 30, respectively. The concentration of VDAC-1 was 2.0 × 10⁻⁶ mol L⁻¹. All the experiments were performed in phosphate buffer (pH 7.0) and at 20 °C.)

Inner filter effect

The inner-filter effect refers to the absorbance or optical dispersion of light at the excitation or emission wavelength by compounds present in the solution.^29–31 One situation is occurred when absorption at the excitation wavelength is significant, less light reaches the center of the sample and the fluorescence of the fluorophore is reduced, while absorption at the emission wavelength reduces the emitted light that reached the detector.²⁹ On the other hand, an inner-filter effect can be caused also by non-absorbing ligands in case they induce significant light scattering, such as the protein–membrane interaction studies, in which titration is performed with a suspension of liposome. In this work, the proteoliposomes were used and thus, the contribution of inner filter effect to the observed fluorescence quenching can not be neglected.³¹

According to the literatures,^29–32 the correction of emission intensity can be achieved by measuring the absorbance values at the excitation and emission wavelength for each concentration of ATP and NADH, and then multiplying the observed emission intensity value. Consequently, the emission intensities can be corrected from:

F_corr = F_obs × e^{(A_exi+A_emi)/2} (1)

where F_corr is the corrected fluorescence intensity that would measured in the absence of inner filter effects, F_obs is the measured fluorescence, and A_exi and A_emi the measured absorbance at the excitation and emission wavelengths, respectively.

All the observed emission intensity F_obs were corrected according to eqn (1), and the corrected emission intensity F_corr were used in the following calculation.

Fluorescence quenching mechanism

The fluorescent quenching can occur in two different mechanisms, static quenching and dynamic quenching.³² During dynamic quenching, the fluorescence substance collides with quencher leading to a decrease in tryptophan emission intensity that may due to direct energy transfer from singlet exited state of tryptophan to quencher. However, in static quenching, a decrease in tryptophan emission may due to the formation of a ground-state complex by quencher and the fluorophores. The Stern–Volmer equation can be used to describe the above two fluorescence quenching mechanism.³²where F₀ and F are the fluorescence intensities in the absence and presence of quencher, respectively. K_sv is the Stern–Volmer quenching constant and it can be expressed as K_sv = k_Qτ₀, where k_Q is the biomolecular quenching rate constant and τ₀ is the average lifetime of the biomolecule without quencher (the value of τ₀ is 10⁻⁸ s), [Q] is the concentration of the quencher. The values of k_Q and K_sv at different temperatures can be used to distinguish static and dynamic quenching. The Stern–Volmer plots for the quenching of VDAC-1 fluorescence by ATP and NADH at different temperatures are displayed in Fig. 3. According to eqn (2), the values of K_sv and k_Q at different temperatures for ATP and NADH are listed in Table 1. Apparently, the values of K_sv decreased with the increase of temperature and the values of k_Q at all temperatures were far greater than the maximum dynamics quenching constant of 2.0 × 10¹⁰ L mol⁻¹ s⁻¹, indicating that the binding of VDAC-1 with ATP and NADH are both static quenching.

Fig. 3 Stern–Volmer plots for the quenching of VDAC-1 fluorescence by ATP (A) and NADH (B) at different temperatures. 10 °C (♦); 20 °C (■); 30 °C (●); 37 °C (▼); 45 °C (▲).

Table 1 Stern–Volmer parameters for ATP and NADH binding to VDAC-1

Ligands	T (K)		r^2	Ksv × 10^3 (L mol^−1)	kQ × 10^11 (L mol^−1 s^−1)
ATP	283		0.988	12.93	12.93
	288		0.973	11.24	11.24
	293		0.980	9.34	9.34
	298		0.961	6.56	6.56
	303		0.984	5.10	5.10
NADH	283		0.987	13.24	13.24
	288		0.991	11.34	11.34
	293		0.971	8.54	8.54
	298		0.979	7.69	7.69
	303		0.990	6.54	6.54

---

Fluorescence lifetime

Time-resolved measurements are widely used in fluorescence spectroscopy, particularly for studies of biological macromolecules.^32–34 It was reported that one can distinguish static and dynamic quenching using lifetime measurements.³² The formations of static ground-state complex do not decrease the decay time of the uncomplexed fluorophores because only the unquenched fluorophores are observed. However, dynamic quenching is a rate process acting on the entire exited-state population; it decreases the mean decay time of the entire excited-state population.³² In this case, transient decays experiments were carried out to further identify the binding mechanism of VDAC-1 with ATP and NADH. The decay time measurements results were shown in Fig. 4. Ludox was used as a reference to examine the shortest fluorescence lifetime of the time-resolved fluorescence spectrometer. Decay in the fluorescence intensity I(t) with time (t) was fitted by a single exponential function:

I(t) = I₀ × e^(−t/τ) (3)

where I₀ is the intensity at time 0 and the lifetime τ is the inverse of total decay rate. The fluorescence lifetime of VDAC-1 liposome was calculated to be (2.16 ± 0.02) ns. It was obviously from the decay curves in Fig. 4A that the fluorescence lifetime of VDAC-1 liposome was larger than Ludox. In addition, from Fig. 4A and B, the fluorescence lifetime of VDAC-1 liposome remained constant with the addition of ATP or NADH, indicating that the binding of VDAC-1 with ATP and NADH are both static quenching. The results of lifetime measurements also confirmed the reliability of Stern–Volmer equation data.

Fig. 4 Decay time measurements of VDAC-1 liposome. Time scan results of VDAC-1 liposome (A); fluorescence and lifetime changes of VDAC-1 liposome quenched by ATP (B) and NADH (C).

Binding parameters

In the case of static quenching, the number of substantive binding sites (n) and apparent binding constants (K_a) of ATP–VDAC-1 complex and NADH–VDAC-1 complex can be determined by the double log Stern–Volmer equation:

However, this equation suffers from some pitfalls: the formation of a non-fluorescent complex is assumed, and the added [Q] is generally equal to the free [Q] in the determined solution.³⁰ In our work, the two conditions were both achieved. Firstly, the static quenching was proved to be the quenching mechanism in the binding of VDAC-1 with ATP and NADH, and the non-fluorescent complex could form. Secondly, during the fluorescent emission experiments, the titrated amount of ATP and NADH were all larger than VDAC-1, (i.e. the molar ratio was (5–125) : 1 and (1–30) : 1 for ATP and NADH binding with VDAC-1) and accordingly, the added [Q] would be generally equal to the free [Q] if a simple 1 : 1 binding was generated. In this case, eqn (2) was used to determine the binding parameters of ATP and NADH with VDAC-1. Fig. 5 shows the plot of vs. lg[Q] based on eqn (4). The calculated values of n and K_a at different temperatures are listed in Table 2. The results indicated that ATP and NADH bound to VDAC-1 in equimolar amounts, and the association constants of the two ligands binding to VDAC-1 decreased with the temperature increasing.

Fig. 5 Plots of vs. lg[ATP] (A) and lg[NADH] (B) at different temperatures. 10 °C (●); 20 °C (▲); 30 °C (▼); 37 °C (×); 45 °C (✳).

Table 2 Binding parameters for ATP and NADH binding to VDAC-1

Ligands	T (K)		r^2	Ka × 10^4 (L mol^−1)	n
ATP	283		0.985	2.77	1.104
	288		0.980	2.40	1.111
	293		0.982	2.11	1.114
	298		0.994	1.86	1.137
	303		0.984	1.67	1.151
NADH	283		0.969	1.12	0.990
	288		0.983	0.97	0.991
	293		0.975	0.86	0.992
	298		0.974	0.76	0.998
	303		0.990	0.61	0.997

---

Although the determined association constants of the two ligands with VDAC-1 are in the same order of magnitude, the K_a values of ATP binding to VDAC-1 are larger than that of NADH with VDAC-1 in each temperature. Yehezkel et al.³⁵ revealed high and low affinity binding site of [α-32P]BzATP, a photoreactive ATP analog, in the VDAC molecule with apparent dissociation constant values of 15 and 53 μM, respectively. The corresponding K_a values of the above high and low affinity binding site are 6.67 × 10⁴ and 1.89 × 10⁴ M⁻¹, respectively. Lee et al.³⁶ showed the dissociation constant of VDAC-1 with NADH was 86 μM, which corresponds to the association constant of 1.16 × 10⁴ M⁻¹. These results indicated that the binding affinity of ATP with VDAC-1 was greater than that of NADH with VDAC-1 and also confirmed the reliability of our data.

Thermodynamic studies

In addition, the thermodynamic data (Table 3) from fluorescent spectra over the temperature range of 283 to 303 K were also estimated by Van't Hoff equation (eqn (5)) and Gibbs Heim Hertz equation (eqn (6)).³⁷

ΔG = ΔH − TΔS (6)

Table 3 Thermodynamic data for ATP and NADH binding to VDAC-1

Ligands	T (K)	ΔH (kJ mol^−1)	ΔG (kJ mol^−1)	ΔS (J mol^−1 K^−1)
ATP	283	−18.04	−24.05	21.25
	288		−24.16
	293		−24.27
	298		−24.37
	303		−24.48
NADH	283	−21.07	−21.98	3.21
	288		−21.99
	293		−22.01
	298		−22.03
	303		−22.04

---

The negative signs of ΔG for the binding of ATP and NADH to VDAC-1 indicated spontaneous bindings during the two processes. According to the semi-empirical Ross rule,³⁸ if ΔH < 0 and ΔS > 0 during the binding of a ligand to a protein, indicating that the main binding force is electrostatic interactions; if ΔH < 0 and ΔS < 0, hydrogen bond interaction and van der Waals force could play a decisive role in the binding; if ΔH > 0 and ΔS > 0, hydrophobic interaction may become the main binding force. Based on the data from Table 3, the negative signs of ΔH and positive signs of ΔS indicated that the binding of VDAC-1 with ATP and NADH should be driven by electrostatic forces.

Binding distances

Fluorescence resonance energy transfer (FRET) is applied to measure the distance between the tryptophan residues of VDAC-1 and the two ligands. According to the Förster's theory,³⁹ FRET can happen when the emission spectrum of fluorophore donor overlaps with the absorption spectrum of the acceptor molecule. The efficiency of energy transfer (E) is described by eqn (7):where r is the distance from the binding ligand on protein to the tryptophan residue, R₀ is the Förster critical distance at which 50% of the excitation energy is transferred to the acceptor. If the wavelength is expressed in cm and J(λ) is in units of M⁻¹ cm³, R₀ is given by the following equation:³²

R₀⁶ = 8.8 × 10⁻²⁵K²N⁻⁴φJ (in cm⁶) (8)

where φ is the quantum yield of donor in the absence of an acceptor, N is the refractive index of medium, and K² is the orientation factor. This factor describes the relative orientation in the space of transition dipoles of the donor and acceptor, and is usually assumed to be equal to 2/3, whereby J describes the extent of overlap between the normalized fluorescence emission spectrum of the donor and the acceptor absorption spectrum. J can be calculated by eqn (9):in the equation, f(λ) is the fluorescence intensity of the fluorescent donor at wavelength λ and is dimensionless, ε(λ) is the molar absorption coefficient of the acceptor at wavelength λ.

In this work, two solutions containing equimolar concentration of ATP and VDAC-1, NADH and VDAC-1 were prepared and their UV-Vis absorption and fluorescence spectra were recorded. As shown in Fig. 6, a good overlap is found between the absorption spectrum of ATP and the emission spectrum of VDAC-1, and so are NADH and VDAC-1. The efficiency of energy transfer and overlapping integration can be obtained using eqn (7) and (9). R₀ can be calculated by eqn (8) using K² = 2/3, N = 1.336 (the average refractive index of water and organics), φ = 0.118 (the quantum yield of tryptophan residue). The calculated distance r is the mean distance from the bound ligand to the tryptophan residues in VDAC-1. The values of the efficiency of energy transfer (E), the spectral overlapping integral (J), the distance between donor and acceptor (r) and the Förster distance (R₀) are listed in Table 4. It is obvious that the distances from the bound ATP and NADH to tryptophan residues of VDAC-1 are less than 7 nm, which indicates non-radiative energy transfer mechanism for these quenching. In addition, the short distances of tryptophan residues of VDAC-1 to ATP and NADH suggested the vital interaction of them.

Fig. 6 Fluorescence emission (solid lines) and UV-Vis absorbance (short dash lines) of VDAC-1 proteoliposome with ATP (A) and NADH (B). (The excitation wavelength was set at 278 nm and the concentrations of ATP, NADH and VDAC1 were 2.0 × 10⁻⁶ mol L⁻¹. All the experiments were performed in phosphate buffer (pH 7.0) and at 20 °C.)

Table 4 Parameters of Förster non-radiation energy transfer

Ligands	J (cm^3 L mol^−1)	E	R0 (nm)	r (nm)
ATP	1.65 × 10^−15	0.120	1.24	1.73
NADH	5.03 × 10^−15	0.067	1.49	2.31

---

In fact, using this analysis, two conditions must be considered: (1) FRET must be the only quenching mechanism; (2) the quenching efficiency should be determined under conditions of complete saturation. Otherwise, unquenched molecules will contribute to the overall emission intensity, making the FRET analysis meaningless.³⁰ In this work, FRET analysis was only used to determine the distance between tryptophan residues of VDAC-1 to ATP and NADH, and validated the essential interactions between VDAC-1 and the two ligands.

Molecular docking

Molecular docking can provide a general binding mode of ligands and proteins, the binding energy and the residues of protein involved in the interaction with ligands. In this work, molecular docking studies were performed by Libdock program. In this program, binding sphere is an important parameter for protein–ligand interaction and affects the reliability of docking results. It was reported that ATP and NADH could bind to VDAC-1 through NBSs and the locations of the NBSs in rat VDAC-1 had already proposed by Song, Colombini and Yehezkel.^35,40,41 It was proved that rat VDAC-1 possessed two or more NBSs, at least two located in each of the N- and C-terminal halves of VDAC-1.^14,35 The amino acids of human VDAC-1 that differ from rat VDAC-1 were Thr58, Met132, Ala163 and Ile230.³⁵ These four amino acids were not involved in the proposed NBSs. Taking into account the possible NBSs in VDAC-1, we selected the radius of 8.825 Å with the center coordinates of x: 27.101, y: 4.678, and z: −2.061, which makes the binding sphere just involved the NBSs of human VDAC-1.

In addition, Libdock score and the value of relative energy were important parameters to evaluate the best binding mode for ligands and receptors. The ligand–receptor complex with the highest Libdock score and the lowest relative energy was often considered to be the best binding pose. Taking into consideration the two values, we selected the best binding modes of ATP–VDAC-1 complex (Libdock score: 144.718; relative energy: 2.904; pose 4) and NADH–VADC-1 complex (Libdock score: 153.134; relative energy: 11.234; pose 6). The 2D overviews of ATP–VDAC-1 complex and NADH–VADC-1 complex are exposed in Fig. 7. The amino acid residues participate in hydrogen bonds, electrostatic and polarity interactions are shown in purple, while those involved in the van der Waals force are displayed in green. It is clear that the amino acid residues in purple are much more than the amino acid residues in green. Therefore, it can be inferred that hydrogen bonds, electrostatic and polarity interactions tend to be responsible for the binding of ATP and NADH to VDAC-1.

Fig. 7 2D overviews of ATP–VDAC-1 complex (A) and NADH–VADC-1 complex (B). (Amino acid residues participate in hydrogen bonds, electrostatic and polarity interactions are shown in purple. Amino acid residues participate in van der Waals force are displayed in green. Blue lines represent the hydrogen bonds and the arrows point to the backbone acceptor. Green lines stand for the hydrogen bonds and the arrows point to the side chain acceptor.)

What is important is that most of the amino acid residues of VDAC-1 interacted with ATP and NADH are electrically charged ones such as threonine, lysine and arginine. ATP and NADH are both negative charged in the neutral condition. The amino acid residues on the binding surface of VDAC-1 and the two ligands posses opposite charge, indicating that ATP–VDAC-1 complex and NADH–VDAC-1 complex are mainly stabilized through electrostatic interactions. It can be seen that nine hydrogen bonds are formed between ATP and VDAC-1, the participated amino acid residues of VDAC-1 are Lys15, Arg18, Thr22, Gly24, Thr207 and Lys227. For the binding of NADH and VDAC-1, Lys15, Arg18, Asp19, Gln199, Thr207 and Lys227 of VDAC-1 are involved in the interaction and seven hydrogen bonds are formed. These results also indicated that hydrogen bonds also contributed to the binding of VDAC-1 to the two ligands. In addition, the relative energy of ATP–VDAC-1 complex is lower than that of NADH–VDAC-1 complex, while the number of the hydrogen bonds of ATP–VDAC-1 complex is more than that of NADH–VDAC-1 complex. This phenomenon reveals the possible stronger binding force between VDAC-1 and ATP compared with the binding of VDAC-1 to NADH. These results are in agreement with the data of thermodynamic studies. Moreover, lysine, glycine and threonine residues are also typical of the consensus sequences A and B that are common to many other nucleotide-binding proteins proposed by Walker,⁴² indicating a possible binding of VDAC-1 with ATP and NADH through NBSs.

Binding of VDAC-1 with ATP and NADH

As shown in Scheme 1, it is obvious that the two molecules share the same basic parent structure, indicating that they may have the similar functions during their binding with VDAC-1. As a matter of fact, ATP can modify the conductance of VDAC-1 and NADH can regulate the gating function of VDAC-1, indicating that the two ligands bind with VDAC-1 to change the permeability of the channel protein.¹⁴ It is easily obtained that ATP and NADH posses four and two negative charges in neutral conditions, respectively. In the view of the quantity of electrostatic charge, the affinity of the four charged ATP binding to VDAC-1 should be larger than that of the two charged NADH. Molecular docking results showed that two more hydrogen bonds were formed during the binding of ATP to VDAC-1 compared with the interaction between NADH and VDAC-1, indicating a stronger force of ATP and VDAC-1. It is notable that three phosphate groups are in the molecular of ATP while only two in NADH. The structure of tetraphydro-2-(4-(1-iminoethyl)pyridin-1(4H)-yl)-5-methylfuran-3,4-diol exists in NADH as well. Although this structure could provide more hydrogen donor during the binding process with VDAC-1 due to the existence of oxygen atom, nitrogen atom and hydroxyl groups, it may also increase steric hindrance of NADH and subsequently decrease the numbers of hydrogen bond of the complex NADH–VDAC-1. These results showed that the binding affinity of ATP and NADH to VDAC-1 was closely related to the structural difference and the flexible feature between them. It has been confirmed that ATP can modify the conductance of VDAC and NADH can regulate the gating of the channel protein.¹⁴ This conductance and gating functions of VDAC are often in relation to the cell life and death, making VDAC become a key player in mitochondria-mediated apoptosis.¹² As a metabolite transporter, VDAC-1 contributes to the metabolic phenotype of cancer cells. If we can control the gating functions of VDAC-1, cancer will be cured. In this work, the structural difference and the flexible features of ATP and NADH were considered to be the main influencing factors of their functions to VDAC-1, which provides useful help in the design and discovery of lead compounds for the treatment of VDAC-1 related diseases, such as cancer. In addition, many researchers have studied the binding and transport of negatively charged nucleotides by VDAC-1, and they suggested that the interactions between these negatively charged nucleotides and VDAC-1 were mainly through electrostatic interaction.^35,42 These results indicated that electrostatic interaction rather than hydrogen bond is mainly responsible for the binding of VDAC-1 with ATP and NADH. This hypothesis can be also confirmed by Yehezkel,³⁵ Lee³⁶ and our work.

Conclusions

This study provides valuable information about interaction of VDAC-1 with ATP and NADH. The obtained values for the numbers of binding sites and the binding constants (∼10⁴ M⁻¹) revealed that ATP and NADH have mild affinity to VDAC-1 and both the two ligands form 1 : 1 complexes with the channel protein. Fluorescence quenching showed that ATP and NADH can quench the fluorescence emission of VDAC-1 through static mechanism. The thermodynamic data obtained from fluorescent spectroscopy indicated the negative values of enthalpy changes and positive signs of entropy changes for both ATP and NADH binding to VDAC-1. Therefore, electrostatic interactions were the main driving force during the binding processes. Molecular docking results indicated that ATP and NADH could bind to VDAC-1 through NBS. This study provides an important insight into the binding mechanism of VDAC-1 to its ligands, which may have great value for the development of innovative drugs to cure VDAC-1 related diseases.

Acknowledgements

We are thanks to the financial support from the National Natural Sciences Foundation of China (No. 21075097, 21475103, 81303185), ‘The Drug Discovery Initiative’ in the Major Natural Science and Technology Project of China (No. 2009X09103-696), the 211 innovative talent project for graduate of The Twelfth Five Year Plan from Northwest University (No. YZZ14073), the Research Fund for the Doctorial Program of Higher Education of China (No. 20136101120030), the National Science Foundation of Shaanxi Province (No. 2015JM2072), the Program for Changjiang Scholars and Innovative Research Team in University (No. IRT_15R55), the Project for Innovative Research Team of Research and Technology of Shaanxi Province (No. 2013KCT-24), Scientific research plan projects of Shaanxi Education Department (No. 14JK1571), and the National Key Scientific Instrument and Equipment Development Project of China (No. 2013YQ170525; subproject: 2013YQ17052509). We also appreciate the technical support of Professor Ru Jiang in Fourth Military Medical University for the performance of molecular docking.

References

1. G. S. Patching, Biochim. Biophys. Acta, Biomembr., 2014, 1838, 43.
2. K. R. Vinothkumar and R. Henderson, Q. Rev. Biophys., 2010, 43, 65.
3. D. J. Müller, N. Wu and K. Palczewski, Pharmacol. Rev., 2008, 60, 43.
4. Y. Arinaminpathy, E. Khurana, D. M. Engelman and M. B. Gerstein, Drug Discovery Today, 2009, 14, 1130.
5. P. V. Escribá, J. M. González-Ros, F. M. Goñi, P. K. Kinnunen, L. Vigh, L. Sánchez-Magraner, A. M. Fernández, X. Busquets, I. Horváth and G. Barceló-Coblijn, J. Cell. Mol. Med., 2008, 12, 829.
6. J. P. Whitelegge, Anal. Chem., 2013, 85, 2558.
7. S. Lindert, I. Maslennikov, E. Chiu, L. C. Pierce, J. A. McCammon and S. Choe, Biochem. Biophys. Res. Commun., 2014, 445, 724.
8. V. Cherezov, E. Abola and R. C. Stevens, Methods Mol. Biol., 2010, 654, 141.
9. V. Shoshan-Barmatz and D. Ben-Hail, Mitochondrion, 2012, 12, 24.
10. A. Messina, S. Reina, F. Guarino and V. D. Pinto, Biochim. Biophys. Acta, Biomembr., 2012, 1818, 1466.
11. M. T. Eddy, L. Andreas, O. Teijido, Y. Su, L. Clark, S. Y. Noskov, G. Wagner, T. K. Rostovtseva and R. G. Griffin, Biochemistry, 2015, 54, 994.
12. V. De Pinto, F. Guarino, A. Guarnera, A. Messina, S. Reina, F. M. Tomasello, V. Palermo and C. Mazzoni, Biochim. Biophys. Acta, Bioenerg., 2010, 1797, 1268.
13. V. Shoshan-Barmatz and D. Mizrachi, Frontiers in Oncology, 2012, 2, 164.
14. O. P. Choudhary, A. Paz, J. L. Adelman, J. P. Colletier, J. Abramson and M. Grabe, Nat. Struct. Mol. Biol., 2014, 21, 626.
15. M. Bayrhuber, T. Meins, M. Habeck, S. Becker, K. Giller, S. Villinger, C. Vonrhein, C. Griesinger, M. Zweckstetter and K. Zeth, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 15370.
16. C. Guardiani, M. A. Scorciapino, G. F. Amodeo, J. Grdadolnik, G. Pappalardo, V. De Pinto, M. Ceccarelli and M. Casu, Biochemistry, 2015, 54, 5646.
17. W. Z. Tan, Biochim. Biophys. Acta, Biomembr., 2012, 1818, 1555.
18. A. Israelson, S. Abu-Hamad, H. Zaid, E. Nahon and V. Shoshan-Barmatz, Cell Calcium, 2007, 41, 235.
19. D. Incel, H. Zaid and V. Shoshan-Barmatz, Biochem. J., 2001, 358, 147.
20. N. Keinan, H. Pahima, D. Ben-Hail and V. Shoshan-Barmatz, Biochim. Biophys. Acta, Mol. Cell Res., 2013, 1833, 1745.
21. A. Lawena, J. D. Ly, D. J. R. Lane, K. Zarschler, A. Messina and V. D. Pinto, Int. J. Biochem. Cell Biol., 2005, 37, 277.
22. O. H. Lowry, N. J. Rosebrough, A. L. Farr and R. J. Randal, J. Biol. Chem., 1951, 193, 265.
23. D. Ben-Hail and V. Shoshan-Barmatz, Cold Spring Harbor Protocols, 2014, 2, 195–198.
24. U. Pick, Arch. Biochem. Biophys., 1981, 212, 186.
25. S. N. Rao, M. S. Head, A. Kulkarni and J. M. LaLonde, J. Chem. Inf. Model., 2007, 47, 2159.
26. A. D. Goddard, P. M. Dijkman, R. J. Adamson, R. I. dos Reis and A. Watts, Methods Enzymol., 2015, 556, 405.
27. H. H. Shen, T. Lithgow and L. L. Martin, Int. J. Mol. Sci., 2013, 14, 1589.
28. L. N. Wang, C. S. Quan, Y. B. Xu, X. H. Li, X. J. Qu and S. D. Fan, Acta Chim. Sin., 2014, 72, 233.
29. A. S. Sharma, S. Anandakumar and M. Ilanchelian, RSC Adv., 2014, 4, 36267.
30. M. van de Weert and L. Stella, J. Mol. Struct., 2011, 998, 144.
31. A. S. Ladokhin, S. Jayasinghe and S. H. White, Anal. Chem., 2000, 285, 235.
32. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Plenum Press, New York, 1999.
33. W. R. Ware, Pure Appl. Chem., 1975, 41, 635.
34. R. G. Bennett, Rev. Sci. Instrum., 1960, 31, 1275.
35. G. Yehezkel, N. Hadad, H. Zaid, S. Sivan and V. Shoshan-Barmatz, J. Biol. Chem., 2006, 281, 5938.
36. A. C. Lee, X. F. Xu and M. Colombini, J. Biol. Chem., 1996, 271, 26724.
37. I. Bijari, S. Ghobadi, H. Mahdiuni, R. Khodarahmi and S. A. Ghadamia, Int. J. Biol. Macromol., 2015, 80, 189.
38. P. D. Ross and S. Subramanian, Biochemistry, 1981, 20, 3096–3102.
39. T. Förster, Modern Quantum Chemistry, Academic Press, New York, 1996.
40. M. Colombini, Mol. Cell. Biochem., 2004, 256, 107.
41. J. Song and M. Colombini, J. Bioenerg. Biomembr., 1996, 28, 153.
42. J. E. Walker, M. Saraste, M. J. Runswick and N. J. Gay, EMBO J., 1982, 1, 945.

---