- Email: [email protected]
Carotenoid binding to proteins: Modeling pigment transport to lipid membranes
Archives of Biochemistry and Biophysics 584 (2015) 125e133
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi
Carotenoid binding to proteins: Modeling pigment transport to lipid membranes Emilia Reszczynska a, b, Renata Welc a, Wojciech Grudzinski a, Kazimierz Trebacz b, Wieslaw I. Gruszecki a, * a b
Department of Biophysics, Institute of Physics, Maria Curie-Skłodowska University, 20-031 Lublin, Poland Department of Biophysics, Institute of Biology and Biochemistry, Maria Curie-Skłodowska University, 20-031 Lublin, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 July 2015 Received in revised form 25 August 2015 Accepted 2 September 2015 Available online 8 September 2015
Carotenoid pigments play numerous important physiological functions in human organism. Very special is a role of lutein and zeaxanthin in the retina of an eye and in particular in its central part, the macula lutea. In the retina, carotenoids can be directly present in the lipid phase of the membranes or remain bound to the proteinepigment complexes. In this work we address a problem of binding of carotenoids to proteins and possible role of such structures in pigment transport to lipid membranes. Interaction of three carotenoids, beta-carotene, lutein and zeaxanthin with two proteins: bovine serum albumin and glutathione S-transferase (GST) was investigated with application of molecular spectroscopy techniques: UVeVis absorption, circular dichroism and Fourier transform infrared spectroscopy (FTIR). Interaction of pigmenteprotein complexes with model lipid bilayers formed with egg yolk phosphatidylcholine was investigated with application of FTIR, Raman imaging of liposomes and electrophysiological technique, in the planar lipid bilayer models. The results show that in all the cases of protein and pigment studied, carotenoids bind to protein and that the complexes formed can interact with membranes. This means that proteinecarotenoid complexes are capable of playing physiological role in pigment transport to biomembranes. © 2015 Published by Elsevier Inc.
Keywords: Retina Macula Carotenoids Xanthophylls Lutein Zeaxanthin Carotenoideprotein complexes
1. Introduction Carotenoid pigments (see Fig. 1 for chemical structures) are ubiquitous in the biosphere and play several important physiological functions, among which protection against oxidative damage of biological structures seems to be one of the most important [1,2]. Very special is a role of carotenoids in the retina of an eye and in particular in its central part called the macula lutea [2]. It is generally accepted that macular carotenoids act as antioxidants and play additionally a role of blue light filter, protecting photoreceptors against photo-degradation [2e4]. It is a matter of ongoing debate whether a natural environment of physiological functioning
Abbreviations: Zea, zeaxanthin; Lut, lutein; b-car, b-carotene; BSA, bovine serum albumin; HSA, human serum albumin; GST, glutathione s-transferase; FTIR, Fouriertransform infrared absorption spectroscopy; BLM, bimolecular lipid membrane; CD, circular dichroism; THF, tetrahydrofuran; EYPC, L-a-egg yolk phosphatidylcholine; PBS, phosphate-buffered saline; ATR, attenuated total reflection. * Corresponding author. E-mail address: [email protected] (W.I. Gruszecki). http://dx.doi.org/10.1016/j.abb.2015.09.004 0003-9861/© 2015 Published by Elsevier Inc.
of carotenoids in the retina is a lipid phase of the membranes or rather membrane-situated pigmenteprotein complexes [5,6]. Importantly, proteinecarotenoid complexes have been isolated from the human macula lutea [5]. The complexes were composed of the Pi isoform of the glutathione S-transferase (GST) and zeaxanthin (Zea), one of the principal macular xanthophyll. It is possible that GST acts solely as a pigment transporter to the membrane but also that proteinexanthophyll complexes remain intact in the membrane environment, in where they can act as antioxidants and blue-light filters. Polar carotenoids, such as Zea and lutein (Lut), present directly in the lipid phase of the membranes are reported to influence considerably the physical properties of lipid bilayers and to protect lipids against oxidative degradation [3,7]. In the present work we address the problem of binding of carotenoid pigments to proteins and interaction of such complexes with model lipid membranes. Several carotenoid-binding proteins have been isolated and described in vertebrate and invertebrate tissues [8]. Two essentially different proteins have been selected in the present work, to study molecular interactions with carotenoids. One, with reported ability to form specific pigmenteprotein complexes, the Pi
126
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
purification in our laboratory were presented previously [11,12]. Carotenoid concentration, in organic solvent solutions, were evaluated based on the molar extinction coefficients listed by Britton [13]. Proteins, bovine serum albumin (BSA) and glutathione stransferase from human placenta, containing the Pi isoform (GST) were purchased from Sigma Aldrich Chem. Co. (USA). Synthetic L-aegg yolk phosphatidylcholine (EYPC) and Phosphate-buffered saline (PBS, pH adjusted to 7.4) were also purchased from Sigma Aldrich Chem. Co. (USA). 2.2. Sample preparation
Fig. 1. Chemical structure of b-carotene, lutein and zeaxanthin.
isoform of the glutathione S-transferase [5], referred to GST, and BSA (bovine serum albumin) which is a prominent representative of the transport protein class. BSA is an analog of human serum albumin (HSA), which is the most abundant protein in human blood plasma and constitutes ca. half of the blood serum protein fraction. HSA has been shown not to form specific pigmenteprotein complexes with Zea but to bind relatively efficiently Lut, with a dissociation constant kd ¼ 0.54 mM [5]. It has been concluded that the carotenoid may occupy 3 binding sites in albumin molecules. A specific binding of Lut and Zea has an inverse affinity in GST as compared to HSA. The dissociation constant for Zea binding to GST has been determined at the level kd ¼ 0.14 mM while relatively weak binding of Lut has been observed (kd ¼ 1.30 mM) [9]. It has to be noticed that carotenoideprotein complexes which bind pigments specifically have been examined exclusively in the studies referred to above. The pigment fractions bound nonspecifically to protein surface have been removed by washing with hexane [5,9]. On the other hand, nonspecific binding of carotenoids to protein molecules may be also expected, in particular in water environment, owing to the fact that dietary carotenoids found in human blood plasma (such as b-carotene, lycopene, lutein and zeaxanthin [10]) are water insoluble. Hybrid pigmenteprotein molecular organization forms may play a physiological role in transporting carotenoids within an organism. Binding of carotenoid pigments to proteins and possible activity of GST and BSA in carotenoid transport to lipid membranes are addressed in the present study. 2. Materials and methods 2.1. Chemicals Zeaxanthin was isolated from fruits of Lycium barabarum, lutein was isolated from leaves of Spinacia oleracea while b-carotene was purchased from Sigma Aldrich Chem. Co. (USA). Zeaxanthin (Zea) and lutein (Lut) were purified chromatographically by HPLC technique (on the phase-reversed, C-18 column), flow velocity 0.8 mL/ min, mobile phase: acetonitrile:methanol:water (72:8:3, v:v:v). Synthetic b-carotene (b-car) was dissolved in the HPLC mobile phase, methanol:hexane (4:1, v:v) and purified as in the case of zeaxanthin and lutein. The pigments were purified directly before experiments. More details regarding pigment isolation and
Protein (BSA and GST) solutions were prepared in PBS (pH 7.4). The protein concentration in the solutions was (1 106 M). Carotenoid solutions were prepared in tetrahydrofuran (THF). In order to transfer carotenoids to THF, the solutions in other organic solvents were evaporated in darkness, under stream of gaseous Argon. Proteinecarotenoid complexation was achieved via injecting of the pigment solutions, prepared in THF, into the protein solution in PBS at 37 C. Final concentration of THF in the water phase was always 2% (by volume) and initial proteinepigment ratio was changed via variation in initial carotenoid concentration in a THF solution. THF has been selected as a solvent (following the protocol developed by Bhosale et al. [5]) owing to relatively high solubility threshold for carotenoids, very good miscibility with water and, importantly, because it does not cause any structural changes to proteins, at the concentrations applied. In the case of control samples carotenoid solutions in THF were injected into PBS (without protein). Following THF injection into the protein solutions in PBS, the samples were incubated for 1 h, at 37 C, with continuous shaking. After incubation, the samples were transferred to refrigerator (4 C) and incubated for 12 h. After that, possible aggregates and microcrystals of carotenoids, which were not bound to proteins and remained in the water phase, were removed from the samples by 20-min. centrifugation at 20 C at 15 000 g. Samples were monitored at each step of preparation by means of absorption measurements in the UVeVis spectral region. Liposomes were formed from EYPC according to a general procedure described previously [14]. EYPC solution in chloroform was dried under gaseous nitrogen and was further kept under vacuum for 30 min in order to remove traces of an organic solvent. Large multilamellar vesicles were prepared by rehydration with the PBS buffer and vortex mixing: 3 cycles for 15 s separated by 5 minincubation at 32 C. After that, small unilamellar vesicles were prepared by sonication with VCXe130 ultrasonic processor (Sonics Inc., USA) for fifteen cycles of 3 s with 100% amplitude with a titanium probe. Sonication was carried out in the water bath, in equilibrium with ice (~0 C). Final lipid concentration was 0.2 mg/mL. 2.3. Spectroscopic measurements UVeVis absorption spectra were recorded with Cary 50 spectrometer from Varian (Australia). Liquid samples were placed in quartz cells (1 cm optical path-length). The spectra were corrected by subtracting the Rayleigh-type light scattering background component proportional to l4 and subjected to smoothing procedure (SavitzkyeGolay, 2-nd order polynomial, 11 points). Infrared absorption spectra were recorded by Vector 33 Fourier Transform Infrared (FTIR) absorption spectrometer (Bruker, Germany). The spectrometer was equipped with attenuated total reflection (ATR) attachment. The internal reflection element was a ZnSe crystal (45 cut) yielding 10 internal reflections. Typically, 10 scans were collected, Fourier transformed and averaged for each measurement. Absorption spectra at a resolution of one data point every 4 cm1 were recorded in the region between 4000 and
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
400 cm1 using a clean crystal as a background. ATR crystal was cleaned with organic solvents (e.g. ethanol). 2 h before experiments and in course of measurements, the chamber of the spectrometer was purged with gaseous Argon, in order to remove water vapors and protect samples against oxidation. Spectral analysis was performed with OPUS (Bruker, Germany) and Grams AI software from Thermo-Galactic (USA). Protein and proteinecarotenoid complex samples were deposited at the surface of ATR crystal by means of evaporation from water solution, under stream of gaseous Argon. 2.4. Monomolecular layers In order to monitor protein binding from water phase to lipid membranes, a lipid bilayer was constituted out of two monomolecular layers formed with EYPC: one deposited at the surface of ATR crystal and the second one, formed inside a spectrometer in the specially designed mini-trough [15]. Monomolecular lipid layers, before deposition, were formed in a Teflon trough. Ultrapure water used in all experiments was obtained from a Millipore water purification system (France). The water specific resistivity was higher than 18 MU cm. Surface pressure was monitored and experiments were controlled by a KSV system (Finland). Monolayer experiments were carried out at room temperature (22 ± 1 C). For the purpose of ATR-FTIR experiments. lipid monolayers were deposited at the ZnSe support playing a role of an ATR crystal: the one lipid monolayer was deposited to the crystal by means of the LangmuireBlodgett technique. Monomolecular lipid membranes were deposited to the crystal surface at the surface pressure of 25 mN/m. In order to form monomolecular layer, the lipid (EYPC) was deposited at the argonewater interface from solution in chloroform (50 mL). Before compression of a monolayer the system was left for 15 min required for chloroform evaporation and the system equilibration. In order to prepare stock solution of EYPC, the lipid molecular weight has been assumed arbitrary as 780 g/mol. 2.5. Planar lipid bilayers: FTIR experiments Interaction of BSA or GST with lipid bilayer membranes was studied in a model system, by means of attenuated total reflection FTIR spectroscopy (ATR-FTIR). A lipid bilayer was formed by assembly of EYPC monomolecular lipid films: one deposited at the surface of the ATR monocrystal and the second formed at the surface of water subphase in the mini-trough placed inside the FTIR spectrometer. Protein was injected into the water subphase, beneath the lipid bilayer, in water solution. A volume of 110 mL was injected into the water volume. Final molar concentration of protein in the subphase was 1.23 106 M. IR absorption spectra were recorded every 1 min, during the first 25 min after injection, and every 2 min afterwards. The measurement lasted about 2 h. The experimental system applied has been described in more detail in our previous work [15].
127
CurrenteVoltage characteristics were obtained by using a fourelectrode method. The Ag/AgCl electrodes were placed in pairs in two compartments separated by the membrane. One pair of them, connected to the amplifier (Elektrometer Electro 705, WPI) represented the voltage registration circuit and monitored by program Lab Scribe 2. Similar system has been used and described in more detail previously [18,19]. In order to study protein interaction with lipid bilayers, the protein solution was injected to one water compartment (50 mL into 10 mL) to the final concentration of 5 108 M. Electric measurements were done after 8 min required for system equilibration and stabilization of measurement indications. 2.7. Resonance Raman spectroscopy and Raman imaging microscopy Raman spectroscopy analyses were carried out with application of a inVia confocal Raman microscope (Renishaw, UK) with argon laser (Stellar-REN, Modu-Laser™, USA) operating at 457 nm (7 mW laser power at the sample), equipped with 50 long distance objective (NA 0,35, Olympus). The spectra were recorded at about 1 cm1 spectral resolution (2400 lines/cm grating) in the spectral region 330e1900 cm1 using EMCCD detection camera Newton 970 from Andor, UK. To improve signal-to-noise ratio spectral scans were accumulated and for each measurement 30 acquisition (1s) were collected. All spectra were pre-processed by cosmic ray removing, noise filtering, baseline correction, smoothing (SavitzkyeGolay, 2-nd order polynomial, 11 points) and finally normalized using WiRE 4.1 software from Renishaw. 3. Results and discussion 3.1. Carotenoid binding to proteins Figs. 2 and 3 present absorption spectra of three selected carotenoids, b-carotene (b-car), lutein (Lut) and zeaxanthin (Zea) bound to two proteins, BSA and GST respectively (additional absorption spectra reflecting carotenoideprotein interactions are presented in Supplementary material, Figs. S1eS6). As can be seen, binding efficiency is different, depending on the chemical structure
2.6. Planar lipid bilayers: electrophysiological experiments Planar lipid membranes were also studied, formed between two water phases, in the model system called BLM (Black Lipid Membrane) [16,17]. BLMs were formed of EYPC solution in n-decaneeisoprpanol (3:1, v:v). Bathing solution for BLM studies was PBS buffer (pH 7.4) on both sides. A small amount of forming solution was applied on the hole in the Teflon cup located in a Plexiglas chamber with two glass windows. During the experiments, solutions were stirred in both compartments with magnetic microstirrers. The process of membrane thinning to a bilayer stage was monitored by visual inspections through a microscope.
Fig. 2. Absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to BSA. Initial carotenoid:protein molar ratio was 3:1. Absorption spectra recorded for other carotenoid:BSA ratios, before and after centrifugation, are presented in Supplementary Figs. S1eS3.
128
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 3. Absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to GST. Initial carotenoid:protein molar ratio was 3:1. Absorption spectra recorded for other carotenoid:GST ratios, before and after centrifugation, are presented in Supplementary Figs. S4eS6.
of both a protein component and a carotenoid. This can be judged from the absorbance level. BSA appears to be particularly efficient in binding xanthophylls (Lut and Zea), in contrast to non-polar bcar. In the latter case, the absorbance level is lowest, despite identical protein concentration in all the samples and the same initial pigment concentration, exposed to interaction. Such an observation points to the presence of hydroxyl groups, located in the C3 and C30 positions of a pigment molecule, as an important factor governing binding of carotenoids to BSA. On the other hand, b-car appears to be relatively efficiently bound to GST (Fig. 3A), despite the fact that it binds also to BSA (see also ref. [20]). As can be seen from the analysis of absorbance level, particularly efficient binding to GST takes place in the case of Zea, in accordance with the discovery that this protein is a specific Zea-binding protein [5]. Analysis of the process of carotenoid binding to GST shows that pigments with two terminal b-ionone rings (b-car and Zea) are more efficient than Lut, possessing one b- and one ε-ring. The differences observed in a carotenoid affinity to the two selected proteins, are consistent with opinion on specificity and selectivity of proteinecarotenoid interactions. Interestingly, as can be concluded based on the shape of the absorption spectra of b-car, Lut and Zea, both in the environment of BSA and GST, pigments adopt different molecular organization forms. Absorption spectra representing monomeric form of b-car, typical for the pigment solution in organic solvents [13], can be observed in the case of both b-car-BSA (Fig. 2A) and b-car-GST (Fig. 3A) complexes. On the other hand, the xanthophylls, Lut and Zea, bound to proteins show the absorption spectra typical of molecular aggregates observed both in hydrated organic solvents and in the environment of lipid membranes [11]. Statistically, in the present experiment, three carotenoid molecules per protein have been introduced into the solution. On the other hand, the fact that the absorption spectra show appearance of aggregated forms of the xanthophylls suggests possibility that some protein molecules are not loaded with pigments and some contain carotenoids at higher concentrations. It is also possible that aggregated structures of the pigments are not totally eliminated from the samples during centrifugation and remain in the water phase as independent entities. In order to check this the control experiment with carotenoid injection to water phase was performed. The results presented in Fig. S7 show that centrifugation
eliminates to a large extent aggregated structures of Zea from the samples, which remain in the water phase. Based on the results it can be estimated that pigments associated with the protein account for more than 80% of the absorbance signal in the samples. Some pigments formed relatively densely packed aggregated forms (constituted solely of carotenoids) and have been removed from the protein solution during centrifugation. Therefore, it is highly probable that molecular aggregates formed of three or more pigment molecules are immobilized at the surface of proteins. Such spectral forms are particularly observed in the case of Lut. As can be seen, binding of Zea to both BSA and GST, is characterized by the relatively complex absorption spectra in which the components typical for aggregated forms, are combined with spectral components with distinctive vibrational sub-structure, typical for nonaggregated carotenoid molecules. Such an effect could be interpreted in terms of two molecular organization forms of Zea attached to GST: monomers, entering specifically protein docking niche and molecular aggregates of the pigment, formed at the proteinewater interface. Such an interpretation has a support from the circular dichroism (CD) spectral analysis. Figs. 4 and 5 present the CD spectra of the carotenoideprotein complexes. As can be seen, binding of Lut to BSA results in formation of chiral complexes of the pigment (Fig. 4B), reported also to be formed in the hydrated organic solvents and in complexes to proteins [21]. Formation of similar structures, by Lut, has not been observed in the case of GST (Fig. 5B). In this case, a CD spectrum looks distinctly different and has very low amplitude. Monomeric binding of b-car to BSA and GST is associated with lack of any CD activity of those pigmenteprotein complexes (Figs. 4A and 5A). A solution of monomeric bcarotene does not give rise to any CD activity due to lack of any chiral centers in the molecular structure. The fact that b-carotene bound to both BSA and GST does not form aggregated structures, as can be concluded from the UVeVis absorption spectra, is most probably responsible for the lack of CD activity of the carotenoideprotein complexes. A specific binding to the protein agrees with the concept that albumin is an effective b-car
Fig. 4. Comparison of circular dichroism (CD) and absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to BSA. Initial carotenoid:protein molar ratio was 3:1.
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 5. Comparison of circular dichroism (CD) and absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to GST. Initial carotenoid:protein molar ratio was 3:1.
129
Fig. 6. FTIR spectra in the amide I region, of BSA (dotted line) and BSA complexed to bcarotene (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
transporter [20]. Interestingly, a complex nature of the Zea CD spectra both in BSA (Fig. 4C) and GST (Fig. 5C) can be observed. One component, observed in the short-wavelength spectral region (350e400 nm), corresponds directly to the similar component of Lut. Interestingly, the second spectral component, visible in the longer wavelength region (450e550 nm), is very similar as reported for a dimeric structure of a polyene antibiotic amphotericin B [14]. It is possible that specific interaction of zeaxanthin to the proteins studied, involves binding of dimeric structures of the pigment. Interestingly, the number of 2 Zea molecules per 23-kD GST subunit, at a saturation level, has been reported for the specific carotenoideprotein complex formation [5]. This suggests that the samples formed in the present study represent, most probably, the specific Zea-GST complexes with additionally bound xanthophyll molecules, immobilized at the protein surface. Such forms may play, in principle, both the specific pigmenteprotein complex and carotenoid-transporter functions. 3.2. Proteinecarotenoid interaction Analyses of infrared absorption spectra of both the proteins show that upon interaction to carotenoids, proteins respond by slightly reorganizing their secondary structure. The most pronounced effects have been observed in the case of BSA-b-car, BSAZea and GST-Zea interactions. Figs. 6e8 present IR absorption spectra of the proteins before and after the association with carotenoids, recorded in the amide I spectral region. As can be seen, binding of carotenoids to BSA is associated with a relative increase in intensity of the spectral components between 1620 cm1 and 1600 cm1, which can be assigned to the b-sheet structure and aggregated strands (see Ref. [22] for assignments of spectral components to protein organization forms), in expense of an intensity of the spectral component peaking around 1650 cm1, which can be assigned to the a-helix structure (see Figs. 6 and 7). Qualitatively different effect can be observed in the case of GST (Fig. 8). In this case, interaction to Zea causes increase in intensities in the spectral
Fig. 7. FTIR spectra in the amide I region, of BSA (dotted line) and BSA complexed to zeaxanthin (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
region corresponding to turns and loops (1678 cm1) in expense of the b-sheet structure (1627 cm1). In all the cases the effects observed are relatively small but reproducible and confirm carotenoideprotein interactions. 3.3. Interaction of carotenoideprotein structures with lipid membranes Interaction of proteinecarotenoid hybrid structures with model lipid membranes was examined in the experimental system, schematically presented in Fig. 9, with application of the ATR-FTIR approach (attenuated total reflection-FTIR spectroscopy). Lipid bilayer was formed at the interface of water and ZnSe monocrystal (ATR element) and protein with bound carotenoids were injected
130
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 8. FTIR spectra in the amide I region, of GST (dotted line) and GST complexed to zeaxanthin (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
Fig. 10. FTIR spectra in the amide I region, of BSA-b-carotene complex bound to bilayer lipid membrane from the water phase after 3 min (dotted line) and after 67 min (solid line). The experimental system is presented in Fig. 9. The spectra were normalized by dividing by the area beneath each spectrum. Lower panel presents the difference spectrum obtained by subtracting the spectrum recorded after 3 min after the protein injection from the spectrum recorded after 67 min of the experiment.
Fig. 9. Schematic representation of the experimental set-up applied in examination of interaction of a protein, introduced into the water subphase, with planar lipid bilayer deposited between the ATR crystal and the water subphase. See the text for more explanations.
into the water subphase, beneath the lipid membrane. A reference spectrum was always recorded just before injection, which enabled monitoring of protein binding to the membrane by recording of IR absorption spectra in the amide I region [15]. The exemplary spectra recorded after 3 min and 67 min after injection are presented in Fig. 10. Figs. 11 and 12 present time dependencies of absorbance changes, monitored at 1630 cm1, reflecting kinetics of binding of BSA and GST and also carotenoideprotein structures, to a lipid bilayer membrane. Slight differences in the spectra recorded during the experiments show that in course of binding, protein molecules tend to aggregate in the lipid membrane system: increased intensity in the spectral region of 1612 cm1 in expense of the band at 1654 cm1 corresponding to a-helix (Fig. 10, bottom panel). The experimental points in Figs. 11 and 12 were fit with the equation:
AðtÞ ¼ Amax 1 evt
(1)
in which, Amax corresponds to the saturation level and v represents
Fig. 11. Time dependency of absorbance changes at 1630 cm1, monitoring binding of BSA and BSAecarotenoid complexes to lipid bilayer membrane deposited between the ATR crystal and the water subphase (see Fig. 9 for experimental set-up). Composition of a protein complex injected beneath lipid bilayer is shown in each panel. The experimental points are fit with Equation (1) (shown in dashed line). Parameters of the fits, Amax and v, are listed in Table 1.
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
131
Fig. 12. Time dependency of absorbance changes at 1630 cm1, monitoring binding of GST and GSTecarotenoid complexes to lipid bilayer membrane deposited between the ATR crystal and the water subphase (see Fig. 9 for experimental set-up). Composition of a protein complex injected beneath lipid bilayer is shown in each panel. The experimental points are fit with Equation (1) (shown in dashed line). Parameters of the fits, Amax and v, are listed in Table 1.
the first order kinetics rate parameter. The fit results (plotted with a dashed line) are superimposed on the experimental points in Figs. 11 and 12 and the parameters fitted for different experimental systems are listed in Table 1. As can be seen, carotenoid binding to both the proteins, BSA and GST, has an effect in decrease (in all the cases except BSA-Lut) in a saturation level. It is possible that modification of the protein surface by carotenoids is responsible for this effect. On the other hand, binding of Lut and Zea to GST
Table 1 Kinetic parameters characterizing protein binding to planar lipid bilayers formed with EYPC, described by Equation (1). Presented are mean values ± S.D. Components
Amax 102
BSA BSA-ZEA BSA-LUT BSA-b-car GST GST-ZEA GST-LUT GST-b-car
1.06 0.6 1.01 0.72 1.05 0.50 0.63 0.54
± ± ± ± ± ± ± ±
v 102 [min1] 0.14 0.02 0.17 0.09 0.06 0.01 0.52 0.13
6.46 6.83 4.35 5.03 7.04 10.68 14.13 7.98
± ± ± ± ± ± ± ±
2.56 0.45 0.13 2.91 0.49 0.28 0.54 1.10
Fig. 13. Electric currentevoltage (IeU) characteristics of planar lipid membrane formed in the BLM model system with EYPC and membranes exposed to interactions with BSAecarotenoid complexes (indicated). Specific electric resistivity of the membranes, calculated based on the IeU dependencies are listed in Table 2.
132
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
increases considerably a rate of a carotenoideprotein binding to the membranes. Binding of carotenoideprotein structures to bilayer lipid membranes was additionally tested with application of planar lipid membrane system (so called BLM, Bimolecular Lipid Membrane [16,17]), by means of electrophysiological technique. Figs. 13 and 14 present voltageecurrent (UeI) dependencies recorded for the membranes formed with pure lipids before and after injection of protein into the bathing solutions. Specific resistance of the membranes was evaluated based on the UeI dependencies. Values determined for different experimental systems are listed in Table 2. As can be seen, interaction of both GST and BSA results in a decrease in the membrane specific resistivity. Such a result is indicative of a protein binding to the membranes, most probably at the polar group surface zone, due to the largely hydrophilic nature of the proteins (in particular BSA). Interestingly, the effect is not as much pronounced in the case of the carotenoideprotein structures, in particular in the case of the carotenoid-binding GST. It is possible that carotenoid molecules immobilized at the protein surface, modify protein interactions with the polar headgroup zone of the lipid bilayer. Alternatively, the effect observed can be associated with carotenoid molecules incorporated into the lipid membranes [18], owing to interaction with proteins. In such a case, carotenoidbinding proteins would play a function of pigment transporters to lipid membranes. Such a possibility was tested in the experiment in which the Zea-GST sample was mixed with the suspension of liposomes. UVeVis absorption spectra of such a system showed evolution in time (Fig. S8). Moreover, from the shape of the spectra it can be concluded that aggregated structures of Zea dissolve gradually in the lipid membrane environment, most probably upon protein interaction with liposomes. Such a conclusion has a strong support from the result of the experiment performed with application of the resonance Raman spectroscopy and Raman imaging. Fig. 15 presents the Raman microscopy image of non-pigmented liposome suspension subjected to interaction with Zea-binding GST solution. An argon laser was tuned to 457 nm, in order to be selectively in resonance with Zea. Owing to this fact, exclusively those liposomes could be imaged which bind the xanthophyll. Very clear image of the liposomes demonstrates effective interaction of the carotenoideprotein with the lipid membrane surface. Such an interaction could result in the protein binding to the membrane surface or/and in the pigment transfer to the membranes. As can be seen, the principal band of the resonance Raman spectrum of Zea bound to GST, assigned to the C]C stretching (peaking at 1520 cm1), is shifted towards lower frequencies with respect to the band recorded after subjecting the Zea-GST sample to interaction with liposomes (1524 cm1, the spectrum recorded from a single liposome marked with white cross in the Raman microscopic image of the liposome suspension presented in the left-hand
Table 2 Specific resistance of planar lipid membranes formed with EYPC and exposed to interaction with proteins and proteinecarotenoid complexes. Presented are mean values ± S.E.
Fig. 14. Electric currentevoltage (IeU) characteristics of planar lipid membrane formed in the BLM model system with EYPC and membranes exposed to interactions with GSTecarotenoid complexes (indicated). Specific electric resistivity of the membranes, calculated based on the IeU dependencies are listed in Table 2.
Components
Specific resistance of membrane 106 [U cm2]
EYPC EYPC EYPC EYPC EYPC EYPC EYPC EYPC EYPC
3.03 0.88 1.1 1.29 1.10 1.11 1.06 1.06 1.10
þ þ þ þ þ þ þ þ
GST GST-LUT GST-b-car GST-ZEA BSA BSA-LUT BSA-b-car BSA-ZEA
± ± ± ± ± ± ± ± ±
0.03 0.02 0.03 0.04 0.03 0.11 0.07 0.05 0.05
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
133
Fig. 15. The results of the resonance Raman analysis of interaction of Zea-binding GST with liposome suspension. A: microscopic image of liposomes deposited at the surface of polylysine film, B: Raman spectrum recorded from the image at the position marked with the white cross (Zea-GST-Liposomes) compared to the spectrum recorded from the ZeaGST sample. The image presented is based on Raman intensity integration in the spectral window 1500e1540 cm1.
panel). Such a result is consistent with the mechanism of Zea-GST interacting with lipid membranes associated with the carotenoid disaggregation in the lipid phase [11,23]. 4. Conclusions
[6]
[7] [8]
The results of the experiments presented show that two proteins interesting from the physiological standpoint: BSA, the popular blood plasma transporter and GST, known to form specific complexes with xanthophylls, are able to bind carotenoids from the water phase. The carotenoid-binding protein structures formed, interact effectively with lipid membranes, changing their physical properties. The mechanism observed may be particularly important in the case of GST-Zea complex, which can play a role of specific transporter of this xanthophyll into the retina of a human eye.
[9]
[10]
[11]
[12]
Acknowledgments
[13]
This research has been performed within the framework of the project, Molecular Spectroscopy for BioMedical Studies” financed by the Foundation for Polish Science within the TEAM program (TEAM/2011-7/2). The research was carried out with the equipment purchased thanks to the financial support of the European Regional Development Fund in the framework of the Development of Eastern Poland Operational Programme.
[14]
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.abb.2015.09.004. References [1] G. Britton, Functions of intact carotenoids, in: G. Britton, S. Liaaen-Jensen, H. Pfander (Eds.), Carotenoids. Natural Functions, vol. 4, Birkhauser, Basel, Boston, Berlin, 2008. [2] J.T. Landrum, Carotenoids. Physical, Chemical and Biological Functions and Properties, CRC Press, Boca Raton, London, New York, 2010. [3] A. Sujak, J. Gabrielska, W. Grudzinski, R. Borc, P. Mazurek, W.I. Gruszecki, Lutein and zeaxanthin as protectors of lipid membranes against oxidative damage: the structural aspects, Arch. Biochem. Biophys. 371 (1999) 301e307. [4] J.T. Landrum, R.A. Bone, Lutein, zeaxanthin, and the macular pigment, Arch. Biochem. Biophys. 385 (2001) 28e40. [5] P. Bhosale, A.J. Larson, J.M. Frederick, K. Southwick, C.D. Thulin, P.S. Bernstein, Identification and characterization of a Pi isoform of glutathione S-transferase
[15]
[16]
[17] [18]
[19]
[20] [21]
[22] [23]
(GSTP1) as a zeaxanthin-binding protein in the macula of the human eye, J. Biol. Chem. 279 (2004) 49447e49454. P.S. Bernstein, F. Khachik, L.S. Carvalho, G.J. Muir, D.Y. Zhao, N.B. Katz, Identification and quantitation of carotenoids and their metabolites in the tissues of the human eye, Exp. Eye Res. 72 (2001) 215e223. W.I. Gruszecki, K. Strzalka, Carotenoids as modulators of lipid membrane physical properties, Biochim. Biophys. Acta 1740 (2005) 108e115. P. Bhosale, P.S. Bernstein, Vertebrate and invertebrate carotenoid-binding proteins, Arch. Biochem. Biophys. 458 (2007) 121e127. P.P. Vachali, B.X. Li, A. Bartschi, P.S. Bernstein, Surface plasmon resonance (SPR)-based biosensor technology for the quantitative characterization of protein-carotenoid interactions, Arch. Biochem. Biophys. 572 (2015) 66e72. F. Khachik, Distribution and metabolism of dietary carotenoids in humans as a criterion for development of nutritional supplements, Pure Appl. Chem. 78 (2006) 1551e1557. P. Adamkiewicz, A. Sujak, W.I. Gruszecki, Spectroscopic study on formation of aggregated structures by carotenoids: role of water, J. Mol. Struct. 1046 (2013) 44e51. J. Milanowska, W.I. Gruszecki, Heat-induced and light-induced isomerization of the xanthophyll pigment zeaxanthin, J. Photochem. Photobiol. B 80 (2005) 178e186. G. Britton, UV/visible spectroscopy, in: G. Britton, S. Liaaen-Jensen, H. Pfander (Eds.), Carotenoids Volume 1B: Spectroscopy, Birkhauser Verlag, Basel, 1995, pp. 13e62. J. Starzyk, M. Gruszecki, K. Tutaj, R. Luchowski, R. Szlazak, P. Wasko, W. Grudznski, J. Czub, W.I. Gruszecki, Self-association of amphotericin B: spontaneous formation of molecular structures responsible for the toxic side effects of the antibiotic, J. Phys. Chem. B 118 (2014) 13821e13832. M. Herec, M. Gagos, M. Kulma, K. Kwiatkowska, A. Sobota, W.I. Gruszecki, Secondary structure and orientation of the pore-forming toxin lysenin in a sphingomyelin-containing membrane, Biochim. Biophys. Acta 1778 (2008) 872e879. A. Wisniewska-Becker, W.I. Gruszecki, Biomembrane models, in: R. Pignatello (Ed.), Drug-biomembrane Interaction Studies. Application of Calorimetric Techniques, Woodhead Publishing Ltd., Philadelphia, 2013, pp. 47e95. H.T. Tien, Bilayer Lipid Membranes (BLM): Theory and Practice, Dekker, New York, 1974. K. Kupisz, A. Sujak, M. Patyra, K. Trebacz, W.I. Gruszecki, Can membranebound carotenoid pigment zeaxanthin carry out a transmembrane proton transfer? Biochim. Biophys. Acta 1778 (2008) 2334e2340. A. Wardak, R. Brodowski, Z. Krupa, W.I. Gruszecki, Effect of light-harvesting complex II on ion transport across model lipid membranes, J. Photochem. Photobiol. B 56 (2000) 12e18. X.R. Li, G.K. Wang, D.J. Chen, Y. Lu, beta-Carotene and astaxanthin with human and bovine serum albumins, Food Chem. 179 (2015) 213e221. F. Zsila, G. Nadolski, S.F. Lockwood, Association studies of aggregated aqueous lutein diphosphate with human serum albumin and alpha(1)-acid glycoprotein in vitro: evidence from circular dichroism and electronic absorption spectroscopy, Bioorg Med. Chem. Lett. 16 (2006) 3797e3801. L.K. Tamm, S.A. Tatulian, Infrared spectroscopy of proteins and peptides in lipid bilayers, Q. Rev. Biophys. 30 (1997) 365e429. R. Mendelsohn, R.W. Van Holten, Zeaxanthin ([3R,30 R]-beta, beta-carotene-330 diol) as a resonance Raman and visible absorption probe of membrane structure, Biophys. J. 27 (1979) 221e235.
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi
Carotenoid binding to proteins: Modeling pigment transport to lipid membranes Emilia Reszczynska a, b, Renata Welc a, Wojciech Grudzinski a, Kazimierz Trebacz b, Wieslaw I. Gruszecki a, * a b
Department of Biophysics, Institute of Physics, Maria Curie-Skłodowska University, 20-031 Lublin, Poland Department of Biophysics, Institute of Biology and Biochemistry, Maria Curie-Skłodowska University, 20-031 Lublin, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 July 2015 Received in revised form 25 August 2015 Accepted 2 September 2015 Available online 8 September 2015
Carotenoid pigments play numerous important physiological functions in human organism. Very special is a role of lutein and zeaxanthin in the retina of an eye and in particular in its central part, the macula lutea. In the retina, carotenoids can be directly present in the lipid phase of the membranes or remain bound to the proteinepigment complexes. In this work we address a problem of binding of carotenoids to proteins and possible role of such structures in pigment transport to lipid membranes. Interaction of three carotenoids, beta-carotene, lutein and zeaxanthin with two proteins: bovine serum albumin and glutathione S-transferase (GST) was investigated with application of molecular spectroscopy techniques: UVeVis absorption, circular dichroism and Fourier transform infrared spectroscopy (FTIR). Interaction of pigmenteprotein complexes with model lipid bilayers formed with egg yolk phosphatidylcholine was investigated with application of FTIR, Raman imaging of liposomes and electrophysiological technique, in the planar lipid bilayer models. The results show that in all the cases of protein and pigment studied, carotenoids bind to protein and that the complexes formed can interact with membranes. This means that proteinecarotenoid complexes are capable of playing physiological role in pigment transport to biomembranes. © 2015 Published by Elsevier Inc.
Keywords: Retina Macula Carotenoids Xanthophylls Lutein Zeaxanthin Carotenoideprotein complexes
1. Introduction Carotenoid pigments (see Fig. 1 for chemical structures) are ubiquitous in the biosphere and play several important physiological functions, among which protection against oxidative damage of biological structures seems to be one of the most important [1,2]. Very special is a role of carotenoids in the retina of an eye and in particular in its central part called the macula lutea [2]. It is generally accepted that macular carotenoids act as antioxidants and play additionally a role of blue light filter, protecting photoreceptors against photo-degradation [2e4]. It is a matter of ongoing debate whether a natural environment of physiological functioning
Abbreviations: Zea, zeaxanthin; Lut, lutein; b-car, b-carotene; BSA, bovine serum albumin; HSA, human serum albumin; GST, glutathione s-transferase; FTIR, Fouriertransform infrared absorption spectroscopy; BLM, bimolecular lipid membrane; CD, circular dichroism; THF, tetrahydrofuran; EYPC, L-a-egg yolk phosphatidylcholine; PBS, phosphate-buffered saline; ATR, attenuated total reflection. * Corresponding author. E-mail address: [email protected] (W.I. Gruszecki). http://dx.doi.org/10.1016/j.abb.2015.09.004 0003-9861/© 2015 Published by Elsevier Inc.
of carotenoids in the retina is a lipid phase of the membranes or rather membrane-situated pigmenteprotein complexes [5,6]. Importantly, proteinecarotenoid complexes have been isolated from the human macula lutea [5]. The complexes were composed of the Pi isoform of the glutathione S-transferase (GST) and zeaxanthin (Zea), one of the principal macular xanthophyll. It is possible that GST acts solely as a pigment transporter to the membrane but also that proteinexanthophyll complexes remain intact in the membrane environment, in where they can act as antioxidants and blue-light filters. Polar carotenoids, such as Zea and lutein (Lut), present directly in the lipid phase of the membranes are reported to influence considerably the physical properties of lipid bilayers and to protect lipids against oxidative degradation [3,7]. In the present work we address the problem of binding of carotenoid pigments to proteins and interaction of such complexes with model lipid membranes. Several carotenoid-binding proteins have been isolated and described in vertebrate and invertebrate tissues [8]. Two essentially different proteins have been selected in the present work, to study molecular interactions with carotenoids. One, with reported ability to form specific pigmenteprotein complexes, the Pi
126
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
purification in our laboratory were presented previously [11,12]. Carotenoid concentration, in organic solvent solutions, were evaluated based on the molar extinction coefficients listed by Britton [13]. Proteins, bovine serum albumin (BSA) and glutathione stransferase from human placenta, containing the Pi isoform (GST) were purchased from Sigma Aldrich Chem. Co. (USA). Synthetic L-aegg yolk phosphatidylcholine (EYPC) and Phosphate-buffered saline (PBS, pH adjusted to 7.4) were also purchased from Sigma Aldrich Chem. Co. (USA). 2.2. Sample preparation
Fig. 1. Chemical structure of b-carotene, lutein and zeaxanthin.
isoform of the glutathione S-transferase [5], referred to GST, and BSA (bovine serum albumin) which is a prominent representative of the transport protein class. BSA is an analog of human serum albumin (HSA), which is the most abundant protein in human blood plasma and constitutes ca. half of the blood serum protein fraction. HSA has been shown not to form specific pigmenteprotein complexes with Zea but to bind relatively efficiently Lut, with a dissociation constant kd ¼ 0.54 mM [5]. It has been concluded that the carotenoid may occupy 3 binding sites in albumin molecules. A specific binding of Lut and Zea has an inverse affinity in GST as compared to HSA. The dissociation constant for Zea binding to GST has been determined at the level kd ¼ 0.14 mM while relatively weak binding of Lut has been observed (kd ¼ 1.30 mM) [9]. It has to be noticed that carotenoideprotein complexes which bind pigments specifically have been examined exclusively in the studies referred to above. The pigment fractions bound nonspecifically to protein surface have been removed by washing with hexane [5,9]. On the other hand, nonspecific binding of carotenoids to protein molecules may be also expected, in particular in water environment, owing to the fact that dietary carotenoids found in human blood plasma (such as b-carotene, lycopene, lutein and zeaxanthin [10]) are water insoluble. Hybrid pigmenteprotein molecular organization forms may play a physiological role in transporting carotenoids within an organism. Binding of carotenoid pigments to proteins and possible activity of GST and BSA in carotenoid transport to lipid membranes are addressed in the present study. 2. Materials and methods 2.1. Chemicals Zeaxanthin was isolated from fruits of Lycium barabarum, lutein was isolated from leaves of Spinacia oleracea while b-carotene was purchased from Sigma Aldrich Chem. Co. (USA). Zeaxanthin (Zea) and lutein (Lut) were purified chromatographically by HPLC technique (on the phase-reversed, C-18 column), flow velocity 0.8 mL/ min, mobile phase: acetonitrile:methanol:water (72:8:3, v:v:v). Synthetic b-carotene (b-car) was dissolved in the HPLC mobile phase, methanol:hexane (4:1, v:v) and purified as in the case of zeaxanthin and lutein. The pigments were purified directly before experiments. More details regarding pigment isolation and
Protein (BSA and GST) solutions were prepared in PBS (pH 7.4). The protein concentration in the solutions was (1 106 M). Carotenoid solutions were prepared in tetrahydrofuran (THF). In order to transfer carotenoids to THF, the solutions in other organic solvents were evaporated in darkness, under stream of gaseous Argon. Proteinecarotenoid complexation was achieved via injecting of the pigment solutions, prepared in THF, into the protein solution in PBS at 37 C. Final concentration of THF in the water phase was always 2% (by volume) and initial proteinepigment ratio was changed via variation in initial carotenoid concentration in a THF solution. THF has been selected as a solvent (following the protocol developed by Bhosale et al. [5]) owing to relatively high solubility threshold for carotenoids, very good miscibility with water and, importantly, because it does not cause any structural changes to proteins, at the concentrations applied. In the case of control samples carotenoid solutions in THF were injected into PBS (without protein). Following THF injection into the protein solutions in PBS, the samples were incubated for 1 h, at 37 C, with continuous shaking. After incubation, the samples were transferred to refrigerator (4 C) and incubated for 12 h. After that, possible aggregates and microcrystals of carotenoids, which were not bound to proteins and remained in the water phase, were removed from the samples by 20-min. centrifugation at 20 C at 15 000 g. Samples were monitored at each step of preparation by means of absorption measurements in the UVeVis spectral region. Liposomes were formed from EYPC according to a general procedure described previously [14]. EYPC solution in chloroform was dried under gaseous nitrogen and was further kept under vacuum for 30 min in order to remove traces of an organic solvent. Large multilamellar vesicles were prepared by rehydration with the PBS buffer and vortex mixing: 3 cycles for 15 s separated by 5 minincubation at 32 C. After that, small unilamellar vesicles were prepared by sonication with VCXe130 ultrasonic processor (Sonics Inc., USA) for fifteen cycles of 3 s with 100% amplitude with a titanium probe. Sonication was carried out in the water bath, in equilibrium with ice (~0 C). Final lipid concentration was 0.2 mg/mL. 2.3. Spectroscopic measurements UVeVis absorption spectra were recorded with Cary 50 spectrometer from Varian (Australia). Liquid samples were placed in quartz cells (1 cm optical path-length). The spectra were corrected by subtracting the Rayleigh-type light scattering background component proportional to l4 and subjected to smoothing procedure (SavitzkyeGolay, 2-nd order polynomial, 11 points). Infrared absorption spectra were recorded by Vector 33 Fourier Transform Infrared (FTIR) absorption spectrometer (Bruker, Germany). The spectrometer was equipped with attenuated total reflection (ATR) attachment. The internal reflection element was a ZnSe crystal (45 cut) yielding 10 internal reflections. Typically, 10 scans were collected, Fourier transformed and averaged for each measurement. Absorption spectra at a resolution of one data point every 4 cm1 were recorded in the region between 4000 and
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
400 cm1 using a clean crystal as a background. ATR crystal was cleaned with organic solvents (e.g. ethanol). 2 h before experiments and in course of measurements, the chamber of the spectrometer was purged with gaseous Argon, in order to remove water vapors and protect samples against oxidation. Spectral analysis was performed with OPUS (Bruker, Germany) and Grams AI software from Thermo-Galactic (USA). Protein and proteinecarotenoid complex samples were deposited at the surface of ATR crystal by means of evaporation from water solution, under stream of gaseous Argon. 2.4. Monomolecular layers In order to monitor protein binding from water phase to lipid membranes, a lipid bilayer was constituted out of two monomolecular layers formed with EYPC: one deposited at the surface of ATR crystal and the second one, formed inside a spectrometer in the specially designed mini-trough [15]. Monomolecular lipid layers, before deposition, were formed in a Teflon trough. Ultrapure water used in all experiments was obtained from a Millipore water purification system (France). The water specific resistivity was higher than 18 MU cm. Surface pressure was monitored and experiments were controlled by a KSV system (Finland). Monolayer experiments were carried out at room temperature (22 ± 1 C). For the purpose of ATR-FTIR experiments. lipid monolayers were deposited at the ZnSe support playing a role of an ATR crystal: the one lipid monolayer was deposited to the crystal by means of the LangmuireBlodgett technique. Monomolecular lipid membranes were deposited to the crystal surface at the surface pressure of 25 mN/m. In order to form monomolecular layer, the lipid (EYPC) was deposited at the argonewater interface from solution in chloroform (50 mL). Before compression of a monolayer the system was left for 15 min required for chloroform evaporation and the system equilibration. In order to prepare stock solution of EYPC, the lipid molecular weight has been assumed arbitrary as 780 g/mol. 2.5. Planar lipid bilayers: FTIR experiments Interaction of BSA or GST with lipid bilayer membranes was studied in a model system, by means of attenuated total reflection FTIR spectroscopy (ATR-FTIR). A lipid bilayer was formed by assembly of EYPC monomolecular lipid films: one deposited at the surface of the ATR monocrystal and the second formed at the surface of water subphase in the mini-trough placed inside the FTIR spectrometer. Protein was injected into the water subphase, beneath the lipid bilayer, in water solution. A volume of 110 mL was injected into the water volume. Final molar concentration of protein in the subphase was 1.23 106 M. IR absorption spectra were recorded every 1 min, during the first 25 min after injection, and every 2 min afterwards. The measurement lasted about 2 h. The experimental system applied has been described in more detail in our previous work [15].
127
CurrenteVoltage characteristics were obtained by using a fourelectrode method. The Ag/AgCl electrodes were placed in pairs in two compartments separated by the membrane. One pair of them, connected to the amplifier (Elektrometer Electro 705, WPI) represented the voltage registration circuit and monitored by program Lab Scribe 2. Similar system has been used and described in more detail previously [18,19]. In order to study protein interaction with lipid bilayers, the protein solution was injected to one water compartment (50 mL into 10 mL) to the final concentration of 5 108 M. Electric measurements were done after 8 min required for system equilibration and stabilization of measurement indications. 2.7. Resonance Raman spectroscopy and Raman imaging microscopy Raman spectroscopy analyses were carried out with application of a inVia confocal Raman microscope (Renishaw, UK) with argon laser (Stellar-REN, Modu-Laser™, USA) operating at 457 nm (7 mW laser power at the sample), equipped with 50 long distance objective (NA 0,35, Olympus). The spectra were recorded at about 1 cm1 spectral resolution (2400 lines/cm grating) in the spectral region 330e1900 cm1 using EMCCD detection camera Newton 970 from Andor, UK. To improve signal-to-noise ratio spectral scans were accumulated and for each measurement 30 acquisition (1s) were collected. All spectra were pre-processed by cosmic ray removing, noise filtering, baseline correction, smoothing (SavitzkyeGolay, 2-nd order polynomial, 11 points) and finally normalized using WiRE 4.1 software from Renishaw. 3. Results and discussion 3.1. Carotenoid binding to proteins Figs. 2 and 3 present absorption spectra of three selected carotenoids, b-carotene (b-car), lutein (Lut) and zeaxanthin (Zea) bound to two proteins, BSA and GST respectively (additional absorption spectra reflecting carotenoideprotein interactions are presented in Supplementary material, Figs. S1eS6). As can be seen, binding efficiency is different, depending on the chemical structure
2.6. Planar lipid bilayers: electrophysiological experiments Planar lipid membranes were also studied, formed between two water phases, in the model system called BLM (Black Lipid Membrane) [16,17]. BLMs were formed of EYPC solution in n-decaneeisoprpanol (3:1, v:v). Bathing solution for BLM studies was PBS buffer (pH 7.4) on both sides. A small amount of forming solution was applied on the hole in the Teflon cup located in a Plexiglas chamber with two glass windows. During the experiments, solutions were stirred in both compartments with magnetic microstirrers. The process of membrane thinning to a bilayer stage was monitored by visual inspections through a microscope.
Fig. 2. Absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to BSA. Initial carotenoid:protein molar ratio was 3:1. Absorption spectra recorded for other carotenoid:BSA ratios, before and after centrifugation, are presented in Supplementary Figs. S1eS3.
128
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 3. Absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to GST. Initial carotenoid:protein molar ratio was 3:1. Absorption spectra recorded for other carotenoid:GST ratios, before and after centrifugation, are presented in Supplementary Figs. S4eS6.
of both a protein component and a carotenoid. This can be judged from the absorbance level. BSA appears to be particularly efficient in binding xanthophylls (Lut and Zea), in contrast to non-polar bcar. In the latter case, the absorbance level is lowest, despite identical protein concentration in all the samples and the same initial pigment concentration, exposed to interaction. Such an observation points to the presence of hydroxyl groups, located in the C3 and C30 positions of a pigment molecule, as an important factor governing binding of carotenoids to BSA. On the other hand, b-car appears to be relatively efficiently bound to GST (Fig. 3A), despite the fact that it binds also to BSA (see also ref. [20]). As can be seen from the analysis of absorbance level, particularly efficient binding to GST takes place in the case of Zea, in accordance with the discovery that this protein is a specific Zea-binding protein [5]. Analysis of the process of carotenoid binding to GST shows that pigments with two terminal b-ionone rings (b-car and Zea) are more efficient than Lut, possessing one b- and one ε-ring. The differences observed in a carotenoid affinity to the two selected proteins, are consistent with opinion on specificity and selectivity of proteinecarotenoid interactions. Interestingly, as can be concluded based on the shape of the absorption spectra of b-car, Lut and Zea, both in the environment of BSA and GST, pigments adopt different molecular organization forms. Absorption spectra representing monomeric form of b-car, typical for the pigment solution in organic solvents [13], can be observed in the case of both b-car-BSA (Fig. 2A) and b-car-GST (Fig. 3A) complexes. On the other hand, the xanthophylls, Lut and Zea, bound to proteins show the absorption spectra typical of molecular aggregates observed both in hydrated organic solvents and in the environment of lipid membranes [11]. Statistically, in the present experiment, three carotenoid molecules per protein have been introduced into the solution. On the other hand, the fact that the absorption spectra show appearance of aggregated forms of the xanthophylls suggests possibility that some protein molecules are not loaded with pigments and some contain carotenoids at higher concentrations. It is also possible that aggregated structures of the pigments are not totally eliminated from the samples during centrifugation and remain in the water phase as independent entities. In order to check this the control experiment with carotenoid injection to water phase was performed. The results presented in Fig. S7 show that centrifugation
eliminates to a large extent aggregated structures of Zea from the samples, which remain in the water phase. Based on the results it can be estimated that pigments associated with the protein account for more than 80% of the absorbance signal in the samples. Some pigments formed relatively densely packed aggregated forms (constituted solely of carotenoids) and have been removed from the protein solution during centrifugation. Therefore, it is highly probable that molecular aggregates formed of three or more pigment molecules are immobilized at the surface of proteins. Such spectral forms are particularly observed in the case of Lut. As can be seen, binding of Zea to both BSA and GST, is characterized by the relatively complex absorption spectra in which the components typical for aggregated forms, are combined with spectral components with distinctive vibrational sub-structure, typical for nonaggregated carotenoid molecules. Such an effect could be interpreted in terms of two molecular organization forms of Zea attached to GST: monomers, entering specifically protein docking niche and molecular aggregates of the pigment, formed at the proteinewater interface. Such an interpretation has a support from the circular dichroism (CD) spectral analysis. Figs. 4 and 5 present the CD spectra of the carotenoideprotein complexes. As can be seen, binding of Lut to BSA results in formation of chiral complexes of the pigment (Fig. 4B), reported also to be formed in the hydrated organic solvents and in complexes to proteins [21]. Formation of similar structures, by Lut, has not been observed in the case of GST (Fig. 5B). In this case, a CD spectrum looks distinctly different and has very low amplitude. Monomeric binding of b-car to BSA and GST is associated with lack of any CD activity of those pigmenteprotein complexes (Figs. 4A and 5A). A solution of monomeric bcarotene does not give rise to any CD activity due to lack of any chiral centers in the molecular structure. The fact that b-carotene bound to both BSA and GST does not form aggregated structures, as can be concluded from the UVeVis absorption spectra, is most probably responsible for the lack of CD activity of the carotenoideprotein complexes. A specific binding to the protein agrees with the concept that albumin is an effective b-car
Fig. 4. Comparison of circular dichroism (CD) and absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to BSA. Initial carotenoid:protein molar ratio was 3:1.
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 5. Comparison of circular dichroism (CD) and absorption spectra of b-carotene (A), lutein (B) and zeaxanthin (C) bound to GST. Initial carotenoid:protein molar ratio was 3:1.
129
Fig. 6. FTIR spectra in the amide I region, of BSA (dotted line) and BSA complexed to bcarotene (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
transporter [20]. Interestingly, a complex nature of the Zea CD spectra both in BSA (Fig. 4C) and GST (Fig. 5C) can be observed. One component, observed in the short-wavelength spectral region (350e400 nm), corresponds directly to the similar component of Lut. Interestingly, the second spectral component, visible in the longer wavelength region (450e550 nm), is very similar as reported for a dimeric structure of a polyene antibiotic amphotericin B [14]. It is possible that specific interaction of zeaxanthin to the proteins studied, involves binding of dimeric structures of the pigment. Interestingly, the number of 2 Zea molecules per 23-kD GST subunit, at a saturation level, has been reported for the specific carotenoideprotein complex formation [5]. This suggests that the samples formed in the present study represent, most probably, the specific Zea-GST complexes with additionally bound xanthophyll molecules, immobilized at the protein surface. Such forms may play, in principle, both the specific pigmenteprotein complex and carotenoid-transporter functions. 3.2. Proteinecarotenoid interaction Analyses of infrared absorption spectra of both the proteins show that upon interaction to carotenoids, proteins respond by slightly reorganizing their secondary structure. The most pronounced effects have been observed in the case of BSA-b-car, BSAZea and GST-Zea interactions. Figs. 6e8 present IR absorption spectra of the proteins before and after the association with carotenoids, recorded in the amide I spectral region. As can be seen, binding of carotenoids to BSA is associated with a relative increase in intensity of the spectral components between 1620 cm1 and 1600 cm1, which can be assigned to the b-sheet structure and aggregated strands (see Ref. [22] for assignments of spectral components to protein organization forms), in expense of an intensity of the spectral component peaking around 1650 cm1, which can be assigned to the a-helix structure (see Figs. 6 and 7). Qualitatively different effect can be observed in the case of GST (Fig. 8). In this case, interaction to Zea causes increase in intensities in the spectral
Fig. 7. FTIR spectra in the amide I region, of BSA (dotted line) and BSA complexed to zeaxanthin (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
region corresponding to turns and loops (1678 cm1) in expense of the b-sheet structure (1627 cm1). In all the cases the effects observed are relatively small but reproducible and confirm carotenoideprotein interactions. 3.3. Interaction of carotenoideprotein structures with lipid membranes Interaction of proteinecarotenoid hybrid structures with model lipid membranes was examined in the experimental system, schematically presented in Fig. 9, with application of the ATR-FTIR approach (attenuated total reflection-FTIR spectroscopy). Lipid bilayer was formed at the interface of water and ZnSe monocrystal (ATR element) and protein with bound carotenoids were injected
130
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
Fig. 8. FTIR spectra in the amide I region, of GST (dotted line) and GST complexed to zeaxanthin (solid line), panel A. Initial carotenoid:protein molar ratio was 3:1. The spectra were normalized by dividing by the area beneath each spectrum. Panel B presents the difference spectrum obtained by subtracting the spectrum of pure protein from the spectrum of the proteinecarotenoid complex.
Fig. 10. FTIR spectra in the amide I region, of BSA-b-carotene complex bound to bilayer lipid membrane from the water phase after 3 min (dotted line) and after 67 min (solid line). The experimental system is presented in Fig. 9. The spectra were normalized by dividing by the area beneath each spectrum. Lower panel presents the difference spectrum obtained by subtracting the spectrum recorded after 3 min after the protein injection from the spectrum recorded after 67 min of the experiment.
Fig. 9. Schematic representation of the experimental set-up applied in examination of interaction of a protein, introduced into the water subphase, with planar lipid bilayer deposited between the ATR crystal and the water subphase. See the text for more explanations.
into the water subphase, beneath the lipid membrane. A reference spectrum was always recorded just before injection, which enabled monitoring of protein binding to the membrane by recording of IR absorption spectra in the amide I region [15]. The exemplary spectra recorded after 3 min and 67 min after injection are presented in Fig. 10. Figs. 11 and 12 present time dependencies of absorbance changes, monitored at 1630 cm1, reflecting kinetics of binding of BSA and GST and also carotenoideprotein structures, to a lipid bilayer membrane. Slight differences in the spectra recorded during the experiments show that in course of binding, protein molecules tend to aggregate in the lipid membrane system: increased intensity in the spectral region of 1612 cm1 in expense of the band at 1654 cm1 corresponding to a-helix (Fig. 10, bottom panel). The experimental points in Figs. 11 and 12 were fit with the equation:
AðtÞ ¼ Amax 1 evt
(1)
in which, Amax corresponds to the saturation level and v represents
Fig. 11. Time dependency of absorbance changes at 1630 cm1, monitoring binding of BSA and BSAecarotenoid complexes to lipid bilayer membrane deposited between the ATR crystal and the water subphase (see Fig. 9 for experimental set-up). Composition of a protein complex injected beneath lipid bilayer is shown in each panel. The experimental points are fit with Equation (1) (shown in dashed line). Parameters of the fits, Amax and v, are listed in Table 1.
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
131
Fig. 12. Time dependency of absorbance changes at 1630 cm1, monitoring binding of GST and GSTecarotenoid complexes to lipid bilayer membrane deposited between the ATR crystal and the water subphase (see Fig. 9 for experimental set-up). Composition of a protein complex injected beneath lipid bilayer is shown in each panel. The experimental points are fit with Equation (1) (shown in dashed line). Parameters of the fits, Amax and v, are listed in Table 1.
the first order kinetics rate parameter. The fit results (plotted with a dashed line) are superimposed on the experimental points in Figs. 11 and 12 and the parameters fitted for different experimental systems are listed in Table 1. As can be seen, carotenoid binding to both the proteins, BSA and GST, has an effect in decrease (in all the cases except BSA-Lut) in a saturation level. It is possible that modification of the protein surface by carotenoids is responsible for this effect. On the other hand, binding of Lut and Zea to GST
Table 1 Kinetic parameters characterizing protein binding to planar lipid bilayers formed with EYPC, described by Equation (1). Presented are mean values ± S.D. Components
Amax 102
BSA BSA-ZEA BSA-LUT BSA-b-car GST GST-ZEA GST-LUT GST-b-car
1.06 0.6 1.01 0.72 1.05 0.50 0.63 0.54
± ± ± ± ± ± ± ±
v 102 [min1] 0.14 0.02 0.17 0.09 0.06 0.01 0.52 0.13
6.46 6.83 4.35 5.03 7.04 10.68 14.13 7.98
± ± ± ± ± ± ± ±
2.56 0.45 0.13 2.91 0.49 0.28 0.54 1.10
Fig. 13. Electric currentevoltage (IeU) characteristics of planar lipid membrane formed in the BLM model system with EYPC and membranes exposed to interactions with BSAecarotenoid complexes (indicated). Specific electric resistivity of the membranes, calculated based on the IeU dependencies are listed in Table 2.
132
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
increases considerably a rate of a carotenoideprotein binding to the membranes. Binding of carotenoideprotein structures to bilayer lipid membranes was additionally tested with application of planar lipid membrane system (so called BLM, Bimolecular Lipid Membrane [16,17]), by means of electrophysiological technique. Figs. 13 and 14 present voltageecurrent (UeI) dependencies recorded for the membranes formed with pure lipids before and after injection of protein into the bathing solutions. Specific resistance of the membranes was evaluated based on the UeI dependencies. Values determined for different experimental systems are listed in Table 2. As can be seen, interaction of both GST and BSA results in a decrease in the membrane specific resistivity. Such a result is indicative of a protein binding to the membranes, most probably at the polar group surface zone, due to the largely hydrophilic nature of the proteins (in particular BSA). Interestingly, the effect is not as much pronounced in the case of the carotenoideprotein structures, in particular in the case of the carotenoid-binding GST. It is possible that carotenoid molecules immobilized at the protein surface, modify protein interactions with the polar headgroup zone of the lipid bilayer. Alternatively, the effect observed can be associated with carotenoid molecules incorporated into the lipid membranes [18], owing to interaction with proteins. In such a case, carotenoidbinding proteins would play a function of pigment transporters to lipid membranes. Such a possibility was tested in the experiment in which the Zea-GST sample was mixed with the suspension of liposomes. UVeVis absorption spectra of such a system showed evolution in time (Fig. S8). Moreover, from the shape of the spectra it can be concluded that aggregated structures of Zea dissolve gradually in the lipid membrane environment, most probably upon protein interaction with liposomes. Such a conclusion has a strong support from the result of the experiment performed with application of the resonance Raman spectroscopy and Raman imaging. Fig. 15 presents the Raman microscopy image of non-pigmented liposome suspension subjected to interaction with Zea-binding GST solution. An argon laser was tuned to 457 nm, in order to be selectively in resonance with Zea. Owing to this fact, exclusively those liposomes could be imaged which bind the xanthophyll. Very clear image of the liposomes demonstrates effective interaction of the carotenoideprotein with the lipid membrane surface. Such an interaction could result in the protein binding to the membrane surface or/and in the pigment transfer to the membranes. As can be seen, the principal band of the resonance Raman spectrum of Zea bound to GST, assigned to the C]C stretching (peaking at 1520 cm1), is shifted towards lower frequencies with respect to the band recorded after subjecting the Zea-GST sample to interaction with liposomes (1524 cm1, the spectrum recorded from a single liposome marked with white cross in the Raman microscopic image of the liposome suspension presented in the left-hand
Table 2 Specific resistance of planar lipid membranes formed with EYPC and exposed to interaction with proteins and proteinecarotenoid complexes. Presented are mean values ± S.E.
Fig. 14. Electric currentevoltage (IeU) characteristics of planar lipid membrane formed in the BLM model system with EYPC and membranes exposed to interactions with GSTecarotenoid complexes (indicated). Specific electric resistivity of the membranes, calculated based on the IeU dependencies are listed in Table 2.
Components
Specific resistance of membrane 106 [U cm2]
EYPC EYPC EYPC EYPC EYPC EYPC EYPC EYPC EYPC
3.03 0.88 1.1 1.29 1.10 1.11 1.06 1.06 1.10
þ þ þ þ þ þ þ þ
GST GST-LUT GST-b-car GST-ZEA BSA BSA-LUT BSA-b-car BSA-ZEA
± ± ± ± ± ± ± ± ±
0.03 0.02 0.03 0.04 0.03 0.11 0.07 0.05 0.05
E. Reszczynska et al. / Archives of Biochemistry and Biophysics 584 (2015) 125e133
133
Fig. 15. The results of the resonance Raman analysis of interaction of Zea-binding GST with liposome suspension. A: microscopic image of liposomes deposited at the surface of polylysine film, B: Raman spectrum recorded from the image at the position marked with the white cross (Zea-GST-Liposomes) compared to the spectrum recorded from the ZeaGST sample. The image presented is based on Raman intensity integration in the spectral window 1500e1540 cm1.
panel). Such a result is consistent with the mechanism of Zea-GST interacting with lipid membranes associated with the carotenoid disaggregation in the lipid phase [11,23]. 4. Conclusions
[6]
[7] [8]
The results of the experiments presented show that two proteins interesting from the physiological standpoint: BSA, the popular blood plasma transporter and GST, known to form specific complexes with xanthophylls, are able to bind carotenoids from the water phase. The carotenoid-binding protein structures formed, interact effectively with lipid membranes, changing their physical properties. The mechanism observed may be particularly important in the case of GST-Zea complex, which can play a role of specific transporter of this xanthophyll into the retina of a human eye.
[9]
[10]
[11]
[12]
Acknowledgments
[13]
This research has been performed within the framework of the project, Molecular Spectroscopy for BioMedical Studies” financed by the Foundation for Polish Science within the TEAM program (TEAM/2011-7/2). The research was carried out with the equipment purchased thanks to the financial support of the European Regional Development Fund in the framework of the Development of Eastern Poland Operational Programme.
[14]
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.abb.2015.09.004. References [1] G. Britton, Functions of intact carotenoids, in: G. Britton, S. Liaaen-Jensen, H. Pfander (Eds.), Carotenoids. Natural Functions, vol. 4, Birkhauser, Basel, Boston, Berlin, 2008. [2] J.T. Landrum, Carotenoids. Physical, Chemical and Biological Functions and Properties, CRC Press, Boca Raton, London, New York, 2010. [3] A. Sujak, J. Gabrielska, W. Grudzinski, R. Borc, P. Mazurek, W.I. Gruszecki, Lutein and zeaxanthin as protectors of lipid membranes against oxidative damage: the structural aspects, Arch. Biochem. Biophys. 371 (1999) 301e307. [4] J.T. Landrum, R.A. Bone, Lutein, zeaxanthin, and the macular pigment, Arch. Biochem. Biophys. 385 (2001) 28e40. [5] P. Bhosale, A.J. Larson, J.M. Frederick, K. Southwick, C.D. Thulin, P.S. Bernstein, Identification and characterization of a Pi isoform of glutathione S-transferase
[15]
[16]
[17] [18]
[19]
[20] [21]
[22] [23]
(GSTP1) as a zeaxanthin-binding protein in the macula of the human eye, J. Biol. Chem. 279 (2004) 49447e49454. P.S. Bernstein, F. Khachik, L.S. Carvalho, G.J. Muir, D.Y. Zhao, N.B. Katz, Identification and quantitation of carotenoids and their metabolites in the tissues of the human eye, Exp. Eye Res. 72 (2001) 215e223. W.I. Gruszecki, K. Strzalka, Carotenoids as modulators of lipid membrane physical properties, Biochim. Biophys. Acta 1740 (2005) 108e115. P. Bhosale, P.S. Bernstein, Vertebrate and invertebrate carotenoid-binding proteins, Arch. Biochem. Biophys. 458 (2007) 121e127. P.P. Vachali, B.X. Li, A. Bartschi, P.S. Bernstein, Surface plasmon resonance (SPR)-based biosensor technology for the quantitative characterization of protein-carotenoid interactions, Arch. Biochem. Biophys. 572 (2015) 66e72. F. Khachik, Distribution and metabolism of dietary carotenoids in humans as a criterion for development of nutritional supplements, Pure Appl. Chem. 78 (2006) 1551e1557. P. Adamkiewicz, A. Sujak, W.I. Gruszecki, Spectroscopic study on formation of aggregated structures by carotenoids: role of water, J. Mol. Struct. 1046 (2013) 44e51. J. Milanowska, W.I. Gruszecki, Heat-induced and light-induced isomerization of the xanthophyll pigment zeaxanthin, J. Photochem. Photobiol. B 80 (2005) 178e186. G. Britton, UV/visible spectroscopy, in: G. Britton, S. Liaaen-Jensen, H. Pfander (Eds.), Carotenoids Volume 1B: Spectroscopy, Birkhauser Verlag, Basel, 1995, pp. 13e62. J. Starzyk, M. Gruszecki, K. Tutaj, R. Luchowski, R. Szlazak, P. Wasko, W. Grudznski, J. Czub, W.I. Gruszecki, Self-association of amphotericin B: spontaneous formation of molecular structures responsible for the toxic side effects of the antibiotic, J. Phys. Chem. B 118 (2014) 13821e13832. M. Herec, M. Gagos, M. Kulma, K. Kwiatkowska, A. Sobota, W.I. Gruszecki, Secondary structure and orientation of the pore-forming toxin lysenin in a sphingomyelin-containing membrane, Biochim. Biophys. Acta 1778 (2008) 872e879. A. Wisniewska-Becker, W.I. Gruszecki, Biomembrane models, in: R. Pignatello (Ed.), Drug-biomembrane Interaction Studies. Application of Calorimetric Techniques, Woodhead Publishing Ltd., Philadelphia, 2013, pp. 47e95. H.T. Tien, Bilayer Lipid Membranes (BLM): Theory and Practice, Dekker, New York, 1974. K. Kupisz, A. Sujak, M. Patyra, K. Trebacz, W.I. Gruszecki, Can membranebound carotenoid pigment zeaxanthin carry out a transmembrane proton transfer? Biochim. Biophys. Acta 1778 (2008) 2334e2340. A. Wardak, R. Brodowski, Z. Krupa, W.I. Gruszecki, Effect of light-harvesting complex II on ion transport across model lipid membranes, J. Photochem. Photobiol. B 56 (2000) 12e18. X.R. Li, G.K. Wang, D.J. Chen, Y. Lu, beta-Carotene and astaxanthin with human and bovine serum albumins, Food Chem. 179 (2015) 213e221. F. Zsila, G. Nadolski, S.F. Lockwood, Association studies of aggregated aqueous lutein diphosphate with human serum albumin and alpha(1)-acid glycoprotein in vitro: evidence from circular dichroism and electronic absorption spectroscopy, Bioorg Med. Chem. Lett. 16 (2006) 3797e3801. L.K. Tamm, S.A. Tatulian, Infrared spectroscopy of proteins and peptides in lipid bilayers, Q. Rev. Biophys. 30 (1997) 365e429. R. Mendelsohn, R.W. Van Holten, Zeaxanthin ([3R,30 R]-beta, beta-carotene-330 diol) as a resonance Raman and visible absorption probe of membrane structure, Biophys. J. 27 (1979) 221e235.