Article pubs.acs.org/JPCB
Interaction of Cationic Protoberberine Alkaloids with Human Serum Albumin. No Spectroscopic Evidence on Binding to Sudlow’s Site 1 Milena Marszalek, Anna Konarska, Ewa Szajdzinska-Pietek,* and Marian Wolszczak* Institute of Applied Radiation Chemistry, Lodz University of Technology, Wroblewskiego 15, 93-590 Lodz, Poland S Supporting Information *
ABSTRACT: Physicochemical studies on drug interactions with human serum albumin (HSA) are relevant for elucidation, at the molecular level, of the processes occurring in vivo. In this work using optical spectroscopic methods (fluorescence, absorption, circular dichroism), we have investigated aqueous HSA solutions containing pharmaceutically important isoquinoline alkaloids, berberine and palmatine. The primary objective was to verify whether the two compounds are located in the subdomain IIA of the secondary HSA structure as reported in literature. We prove that the excited state of Trp214 residue is not quenched by the alkaloids; all observed changes in fluorescence spectra are due to inner filter effects. Furthermore, differential absorption spectra indicate that the ligands remain in a waterlike microenvironment. We infer that bound alkaloid molecules are located at the protein/water interface. Yet, such binding mode can induce some unfolding of the HSA molecule detectable in the far-UV circular dichroism (CD) spectra. We have also performed, for the first time, pulse radiolysis studies of hydrated electron scavenging in the HSA/alkaloid systems and have measured steady-state absorption spectra of irradiated samples. The results reveal that neither berberine nor palmatine is effectively protected by the protein against one-electron reduction, which is consistent with the aforementioned conclusion.
■
binding constants are about 104 M−1.17−22 The steady-state absorption and emission measurements, with berberine (BER) and palmatine (PAL) as quenchers of the intrinsic HSA fluorescence, were interpreted in terms of the alkaloid binding to Sudlow’s site 1 in the subdomain II A where the tryptophan residue, Trp214, is located.18,19,21 These findings, however, are not in line with other reports which have suggested that mainly anionic and neutral ligands are effectively bound in hydrophobic domains of the albumins.7,23 Therefore, we have undertaken a more detailed study of BER and PAL (cf. Chart 1 for the formulas) in aqueous HSA systems by optical spectroscopic techniques (UV−vis absorption, steadystate and time-resolved fluorescence, circular dichroism) with the
INTRODUCTION Protoberberine alkaloids display a variety of pharmacological and biological activities, and their use as medicines dates to several thousand years ago; they are the main active components of some Chinese herbal preparations such as Rhizoma Coptidis, Cortex Phellodendri, or Hydrastis Canadensis (goldenseal). Their antibacterial and antiviral properties have been well evidenced.1 More recent studies indicate that these alkaloids are also promising drugs for treatment of cancer2 and Alzheimer’s disease3 and can be used as cholesterol-lowering agents4 as well as sensitizers in photodynamic therapy5 and radiotherapy.6 Human serum albumin (HSA) is a transport protein that constitutes up to 60% of all proteins in the blood plasma. It consists of a single polypeptide chain made up of 585 amino acid residues (molecular weight 66.4 kDa). At biological pH, the HSA chain forms three homologous domains (I−III), which are arranged in a heart-shaped structure of the side lengths 83, 70, and 82 Å and of the thickness 30 Å. Each domain is divided into two subdomains A and B comprising specific binding sites for various endo- and exogenous substances, such as drugs, hormones, and other solutes present in the bloodstream.7−9 The distribution and metabolic lifetime of these compounds in the body as well as biological activity largely depend on their affinity to HSA. It is thus important to elucidate molecular mechanisms of the ligand−HSA binding, and there is extensive literature on this subject, for example, refs 10−22. Interaction of protoberberine alkaloids with serum albumins was examined by various methods,15−22 and the reported © 2013 American Chemical Society
Chart 1. Structural Formulas of Berberine (a) and Palmatine (b) Cations
Received: September 3, 2013 Revised: October 16, 2013 Published: November 11, 2013 15987
dx.doi.org/10.1021/jp408827b | J. Phys. Chem. B 2013, 117, 15987−15993
The Journal of Physical Chemistry B
Article
solution, two additional cells were inserted in the fluorimeter chamber: one on the path of exciting light (E) and the other one on the path of emitted light (F); they could be filled with the neat buffer or with the quencher solution as an external filter (cf. section Berberine Effect on Fluorescence Spectra of HSA). Circular dichroism spectra (195−260 nm) were recorded using a J-815 CD spectropolarimeter (Jasco) with a thermostated (25 °C) cell holder. The optical path length of the cell was 0.5 cm. Before and during the experiment, the instrument was thoroughly purged with N2 (purity 99.9%). Measurement parameters were as follows: bandwidth, 1.0 nm; slit width, auto; response, 1 s; scan speed, 50 nm/min; step resolution, 1 nm; number of scans, 2 or 3. The dynode voltage during the measurement never exceeded 600 V. The spectra of HSA/BER systems were corrected by subtracting a weak background signal recorded for the respective BER solution in neat buffer and were analyzed for the α-helix content using the CDNN program provided by the instrument producer. Time-Resolved Fluorescence Measurements. The FL 900 spectrofluorimeter (Edinburgh Instruments), equipped with a 295 nm pulsed light-emitting diode (NanoLED-295 from Horiba Jobin Yvon, typical pulse width < 1.2 ns), was used for time-correlated single-photon counting (TCSPC) measurements of HSA fluorescence lifetimes at the emission wavelength 345 nm. Data analysis was performed with the nonlinear leastsquares deconvolution software provided with the instrument. The fluorescence lifetimes were also measured by the stroboscopic technique with an EasyLife V OBB fluorimeter using a 295 nm LED as an excitation light source and a 360 nm interference filter for emission detection. Four scans were acquired by the arithmetic data collection method. The number of channels used for every scan was 400, and the time of integration over which the signal was averaged for every point was 1 s. The recorded signals were analyzed with the commercial PTI T-Master software. In both time-resolved techniques, the instrumental response function (IRF) was determined experimentally on the basis of the light signal scattered from Ludox (colloidal silica in water) and was used for subsequent deconvolution of the fluorescence signal. Biexponential functions were used to fit the fluorescence data, and the fit goodness was evaluated by χ2 values and visual inspection of residuals. The average lifetime was calculated according to the equation
focus on evaluation of the inner filter effects on the protein fluorescence response. To gain additional support for our conclusions, which rather contradict those reported previously,18,19 we have also examined the effect of HSA on the alkaloid reduction by hydrated electrons (eaq−) produced in the aqueous phase by electron-beam irradiation.
■
EXPERIMENTAL METHODS
Materials and Sample Preparation. Essentially fatty acid free albumin from human serum and berberine and palmatine chlorides (97%) were purchased from Sigma-Aldrich and were used as received. Water was purified with the Millipore Milli-Q Plus system. Sodium phosphates of 99% purity were used to prepare 10 mM buffer, pH = 7.2. The examined solutions were prepared by mixing appropriate volumes of the HSA solution in buffer with the alkaloid stock (1−2 mM in water). Final concentrations of the solutes were verified spectrophotometrically using the molar absorption coefficients determined in this work: ε280 = 35 500 M−1 cm−1 for HSA, ε344 = 22 500 M−1 cm−1 for BER, and ε343 = 25 000 M−1 cm−1 for PAL (the values are consistent with literature data24−26). Samples subjected to electronbeam irradiation contained also 0.1 M tert-butanol (t-BuOH) as a scavenger of OH radicals (the reactive species generated along with eaq−, with the same yield27) and were saturated with highpurity N2 or were deaerated under vacuum. HSA solutions were equilibrated at room temperature at least for an hour and were used within 24 h. With the exception of circular dichroism (CD) experiments, vide infra, all samples were measured in quartz cells of the optical path 1 cm at room temperature (25 ± 1 °C). Steady-State Optical Measurements. Absorption and fluorescence spectra were recorded with Carry 5E (Varian) or Perkin-Elmer 750 spectrophotometers and Aminco-Bowman Series 2 spectrofluorimeter, respectively. Spectral resolution in absorption measurements was 0.5 nm. In emission measurements, the excitation wavelength was 295 nm, and excitation and emission slits were set to 4 and 2 nm, respectively. To assess experimentally the inner filter effects on the fluorescence spectra of HSA/alkaloid systems, we have used the setup presented in Chart 2. In addition to the measuring cell (M) containing HSA Chart 2. Schematic Representation of the System Used in the Steady-State Fluorescence Experimentsa
τav =
α1τ12 + α2τ22 α1τ1 + α2τ2
(1)
where α1 and α2 are the pre-exponential factors and where τ1 and τ2 are the respective decay times. Pulse Radiolysis. Electron pulses of 17 ns duration were delivered from the linear accelerator (LINAC) ELU-6E operating in a single pulse mode. Radiation-generated transient species were monitored using the real time optical detection system (UV−vis spectrophotometry) consisting of a Xe lamp (150W Hamamatsu, L2273), Spectra Pro-275 monochromator (Acton Research Corp.), photodetector (five-stage photomultiplier tube Hamamatsu R-955), and digital storage oscilloscope Tektronix TDS540 or Tektronix DPO7254. For measurements of transient absorption spectra, a flow cell system was used and the solution was purged with nitrogen at least 20 min before and all the time during the experiment. Rate constant measurements were done for samples vacuum deaerated in a suprasil cell equipped with a Teflon vacuum
E, excitation light filter cell; F, emitted light filter cell; M, measuring cell (all of the optical path length 1 cm); PMT, photomultiplier. a
15988
dx.doi.org/10.1021/jp408827b | J. Phys. Chem. B 2013, 117, 15987−15993
The Journal of Physical Chemistry B
Article
PAL.18,19 In both cases, a lowering of HSA fluorescence intensity, accompanied by spectral shape changes, was observed with increasing the alkaloid concentration; the characteristic nonresolved band (λmax = 340 nm) gradually split into two peaks (at ∼327 nm and ∼360 nm). The authors argue that these findings indicate the BER/PAL binding to Sudlow’s site 1, but in data analysis, they have ignored possible inner filter effects. Because absorption and fluorescence spectra of HSA overlap with the alkaloid absorption spectrum, cf. Figure 1 (or Figure S1 of the Supporting Information for PAL) and Figure 2A (top
valve; such samples were also used for the steady-state absorption measurements after irradiation with increasing number of pulses. The dose absorbed by a sample in a single pulse was in the range 55−60 Gy (J·kg−1) as determined from the initial absorbance of hydrated electrons at 720 nm using the molar absorption coefficient 18 500 M−1 cm−1.28
■
RESULTS AND DISCUSSION Absorption Spectra. Figure 1 shows absorption spectra of BER solutions (in buffer) in the absence and in the presence of
Figure 1. Absorption spectra of BER/HSA (curve 1) and BER (curve 2) solutions, and their difference (curve 3 = curve 1 − curve 2); [HSA] = 37 μM, [BER] = 27.5 μM in both systems. Inset: absorption spectra of 3 μM and 37 μM HSA solutions (curves a and b, respectively).
HSA and the differential spectrum BER/HSA minus BER. The latter is the same as the spectrum of HSA blank presented by curve b in the inset. Similar results have been obtained for the palmatine system (Figure S1 of the Supporting Information). Thus, in contradiction to the previous work of Hu and coworkers,18,19 we do not observe any alkaloid effect on the protein spectrum which might be taken as evidence of the HSA−alkaloid complex formation. In our opinion, the differential spectrum shown in ref 12 (cf. Figure 3 therein) is an artifact because of incorrect absorption measurements in the low wavelength region (
