Introduction

In recent years, plant-derived polyphenols have been intensively studied because of their various biological and pharmacological activities. Chlorogenic acid (CGA) is a phenolic compound found in various parts of the plant, in fruits, vegetables, and herbs. It is present in large quantities in coffee beans, dry plums, cocoa fruits, hawthorn, nettle, bilberries, sunflower seeds, honeysuckle, potato tubers, and apples [1–6]. The biological properties of CGA are primarily attributed to its capacity to donate hydrogen atoms of the phenolic ring to free radicals, thus inhibiting oxidation process [7]. Research confirms that CGA, among the polyphenol compounds that occur in plants, is an important natural antioxidant [5, 6, 8–12]. The polyphenols are scavengers of free radicals which are responsible for oxidizing biological structures, with resultant damage and pathological states of an organism. Epidemiological studies have shown relationships between consumption of polyphenol-rich foods and prevention of diseases, such as coronary heart disease and osteoporosis [13]. Research showed that CGA has bacteriostatic, hypoglycemic, anticancer, anti-inflammatory, and antiviral activities [3, 14–17].

The mechanisms of the wide biological activity exhibited by CGA may involve its interactions with lipid bilayers. Understanding the effects of CGA on the model lipid bilayer may help in explaining the molecular mechanism of the action of this polyphenol as an antioxidant [18–23]. The ability of polyphenols to penetrate into lipid bilayers is undoubtedly crucial to the protection against oxidation [24]. The authors [25] suggest that polyphenols that partition in the nonpolar region of the bilayer can inhibit the propagation of lipid oxidation by intercepting intramembrane radicals and/or by increasing membrane fluidity which hinders their propagation. Thus, it is important to determine the precise location of CGA in membranes and to examine its effect on membrane fluidity and packing order. Despite great interest in this compound, its interaction with the lipid model membrane has not yet been investigated.

The aim of this paper was to present a study of the interactions of CGA with lipid membranes, and specifically to determine the effect of this compound on thermotropic phase behavior of lipids and their localization in the lipid bilayer. CGA effects were studied by means of differential scanning calorimetry (DSC) and steady-state fluorimetry. The above-mentioned methods have been employed to systematically monitor the influence of CGA on different membrane regions. In particular, the degree of packing order in the hydrophilic phase of the lipid bilayer was determined by the Laurdan and Prodan fluorescence probes, while the fluorescence anisotropy of the hydrophobic phase with the DPH and TMA-DPH probes.

Differential scanning calorimetry is useful to measure the thermal properties and the effects of degradation on the structure of biodegradable polymers, and is a useful method in studies on the nature of biological systems such as proteins or lipids [26–30]. Using DSC, we monitored thermal parameters which included the temperatures of the pretransition (T p) and the main transition peak (T m), the half-width (ΔT 1/2), and change of the enthalpy (ΔH) of the main transition of lipids. We used model systems of dimyristoylphosphatidylcholines, dipalmitoylphosphatidylcholines, and dipalmitoylphosphatidylcholines with cholesterol, such as multilamellar or unilamellar liposomes, as a very good platform to elucidate the effect of CGA on some physical properties of lipid membranes [18, 19, 31–35].

Experimental

Materials

The 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC) and 1,2-dimyristoyl-sn-glycero-3-phosphatidylcholine (DMPC) lipids, cholesterol (chol), and 3-caffeoyl-d-quinic acid (CGA) (Fig. 1) were purchased from Sigma Aldrich, Steinheim, Germany. The fluorescent probes—6-dodecanoyl-2-dimethylaminonaphthalene (Laurdan), 6-propionyl-2-dimethylaminonaphthalene (Prodan), 1,6-diphenyl-1,3,5-hexatriene (DPH) and 1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene p-toluenesulfonate (TMA-DPH)—were purchased from Molecular Probes, Eugene, Oregon, USA.

Fig. 1
figure 1

Chemical structure of chlorogenic acid (CGA)

Full size image

Methods

Differential scanning calorimetry (DSC)

Samples for DSC were prepared from multilamellar liposomes (MLV) of phosphatidylcholine (DPPC, DMPC) with the presence and absence of cholesterol. Pure lipid or lipid with cholesterol was dissolved in chloroform. Chloroform was very carefully evaporated to dryness under nitrogen, and thin film was formed on the flask wall. Samples were left in a vacuum pump for at least 2 h. After that, phosphate buffer of pH 7.4 and CGA were added. Next, the lipid film was washed away from the flask wall using a mechanical shaker. The solution penetrated in between the film bilayer causing its swelling and formation of multilamellar vesicles (MLV)—liposomes. The shaking was conducted in a temperature above the main phase transition of the lipid, until all the lipids were suspended as a homogenic, milky suspension. The obtained buffered suspension contained multilamellar liposomes. Final lipid concentration in the samples was 25 mg cm−3, while that of cholesterol was 2 and 10 mol%. The prepared MLV of pure lecithin or DPPC/chol (control sample) and lecithin with CGA were placed in Mettler Toledo standard aluminum crucibles with lid of 40 μl capacity. Next, the vessels were tightly closed using a special press, and incubated for 4 days at 4 °C. The measurements were made with Mettler Toledo Thermal Analysis System D.S.C. 821e, operated at the heating rate of 2 °C min−1 from 20 to 60 °C, of DPPC (from 10 to 50 °C of DMPC) and DPPC/chol. Thermal cycles were repeated three times, the experimental errors in temperature and thermal response were ±0.2 °C and ±5 %, respectively.

The data were analyzed using original software provided by Mettler Toledo in order to determine the temperature T p, T m, half-width (ΔT 1/2), and calorimetric enthalpy (ΔH) of the main phase transition.

Fluorescence spectroscopy

Samples for steady-state fluorimetry consisted of small unilamellar vesicles (SUV) with DPPC, DMPC, and DPPC with cholesterol—all modified with CGA (DPPC-CGA, DMPC-CGA, and DPPC/chol-CGA). First, the MLV were prepared similarly as for DSC measurements, and then SUV were formed by sonificating lecithin dispersion with phosphate buffer (pH 7.4) for 15 min at 20 kHz. Fluorescent probes (Laurdan, Prodan, and DPH or TMA-DPH) were added to SUV and after that samples were incubated for 15 min in darkness at room temperature. Then a solution of CGA was added, and samples were incubated for 30 min. Control samples contained lipid suspension and a suitable fluorescent probe at 1000:1 molar ratio, and appropriate compounds at 140 μM concentration were added to the remaining samples. The measurements were made at different temperatures—above and below the main phase transition of DPPC or DMPC. Thermal cycles (20–55 or 10–40 °C) were repeated three times.

The measurements were conducted with a CARY Eclipse of VARIAN fluorimeter equipped with a Peltier temperature controller DBS (temp. accuracy ± 0.1 °C). The excitation and emission wavelengths for DPH probe were λ ex = 360 nm, λ em = 425 nm, whereas for TMA-DPH probe: λ ex = 358 nm, λ em = 428 nm. The excitation wavelength for Laurdan and Prodan was 360 nm, and the emitted fluorescence was recorded at two wavelengths: 440 and 490 nm.

Fluorescence anisotropy (A) for DPH probe was calculated using the formula \( A = (I_{\text{II}} - GI_{ \bot } )/(I_{\text{II}} + GI_{ \bot } ) \), where I II and I ⊥—fluorescence intensities—were observed in directions parallel and perpendicular to the polarization direction of the exciting wave. G is an apparatus constant that depends on emission wavelength [30]. Changes in the polar group packing arrangement of the hydrophilic part of the membrane were investigated using the Laurdan and Prodan probes, on the basis of generalized polarization (GP), and were calculated using the formula \( GP = (I_{\text{g}} - I_{\text{l}} )/(I_{\text{g}} + I_{\text{l}} ) \), where I g and I l are the fluorescence intensities at the gel and fluid phase, respectively [35, 36].

Statistical analyses

Statistical analysis was carried out using Statistica 9.0 (StatSoft Inc.). All the experiments were performed at least in triplicate unless specified otherwise. Analysis of variance was carried out and significance between means was determined using Dunnett’s post hoc test. Results are presented as mean ± SD. Significant levels were defined at p < 0.05.

Results and discussion

Interactions of CGA with lipid bilayers were investigated by means of DSC and fluorescence spectroscopy. The molecular models used in our studies of the interaction of CGA with membranes are DPPC and DMPC bilayers, which are the representative phospholipids of the human erythrocyte membrane. To make the model membranes resemble the real ones as much as possible, they were modified with cholesterol. Cholesterol is essential for the normal growth and functioning of cells (e.g., by modulating the function of the membrane proteins through direct binding to sterol-specific sites on membrane proteins).

The effect of CGA on the phase transition of DPPC and DPPC/chol, determined in a DSC experiment, is shown in Fig. 2. MLV formed from DPPC showed the characteristic pretransition and the main transition. In the absence of cholesterol and CGA, DPPC liposomes display a low enthalpic pretransition at 35.4 °C and a sharp main transition at 41.4 °C (Fig. 2a). The phase transition temperatures of pure DPPC and the half-width of the main phase transition (ΔT 1/2 = 0.6 °C) are in good agreement with the literature data [37, 38]. Below the pretransition temperature (T p), the lamellar gel phase is present, in which the lipid chains are in all-trans conformation. Above the main transition temperature (T m), the fluid lamellar phase appears. In the temperature range T p −T m, an intermediate phase is observed, where bilayers are modulated by a ripple phase [38]. Cholesterol tends to fluidize the gel phase and to order the liquid-crystalline phase, thereby broadening or even eliminating the transition between the two phases (at concentration above 50 mol%). In our experiments, two different concentrations of cholesterol were used: 2 mol%, for which the pretransition could be observed, and 10 mol%, for which sharp main transition occurs [38–40].

Fig. 2
figure 2

DSC transition curves of MLV with DPPC, and DPPC modified with CGA, in the absence (a) and the presence of cholesterol (b, c)

Full size image

In the studied range of concentrations, CGA practically did not affect the main phase transition temperature of MLV from DPPC (Figs. 2a, 3a, b). The peak of the main transition of DPPC is a little broadened, but remains approximately symmetric (Figs. 2a, 3b). The same effect was observed for MLV from DMPC (data not shown). These results suggest that CGA does not intercalate with the hydrophobic bilayer core. The half-width of the main phase transition (ΔT 1/2) is an index of the cooperativity of this conversion: the narrower the peak, the higher the cooperativity [29]. Compounds that interact or intercalate with the acyl chains can significantly affect the cooperativity of the transition.

Fig. 3
figure 3

Main phase transition temperatures (T m) as a function of CGA concentration (a), peak half-width (ΔT 1/2) as a function of CGA concentration (b), and enthalpies (ΔH) of the main phase transition as a function of CGA concentration (c)

Full size image

In the case of MLV and the presence of cholesterol, we observed slightly greater change of T m and ΔT 1/2 than for pure lipids. It is interesting that CGA does not abolish the pretransition of DPPC but abolishes the pretransition of DPPC with 2 mol% cholesterol (Fig. 2b). These results confirm the suggestion that CGA most probably interacts with the surface of the bilayer by electrostatic interaction. Pretransition is very sensitive to any kind of perturbation, not only by compounds that penetrate the lipid bilayer but also by cosmotropic substances. Compounds that incorporate into the polar–unpolar region of the lipid bilayer abolish the DPPC pretransition even when used at low concentrations [33]. Additionally, the greater overall enthalpy suggests a lower ability of CGA to disrupt the bilayer structure (Fig. 3c). Summarizing, CGA had less influence on membranes formed from pure DPPC than on membranes modified with cholesterol. Analyzing the influence of cholesterol and investigated compound on the thermotropic parameters of membranes, it can be concluded that cholesterol (10 mol%) modifies them in greater extent than CGA, which may lead to the conclusion that CGA locates itself in the vicinity of the polar heads of phospholipids and does not disturb fluidity of the bilayer.

The phase behavior and membrane fluidity changes induced by CGA were analyzed using four probes: Laurdan, Prodan, DPH, and TMA-DPH. These fluorescent probes were used because each of them incorporates in a different region of the lipid bilayer. The TMA-DPH probe incorporates at the fourth carbon atom in the transient region between the hydrophobic and hydrophilic parts of the bilayer. The DPH probe locates in the hydrophobic, while Laurdan and Prodan—in the hydrophilic region. Such differentiated incorporation of the probes gives insight into the structural changes caused by incorporation of CGA [41, 42].

The effect of CGA on the fluidity and main phase transition of SUV formed of DPPC was studied on the basis of anisotropy measured with the DPH probe. Steady-state fluorescence anisotropy is primarily related to restriction of rotational motion of a dye in the hydrocarbon chain region. Therefore, a decrease in the anisotropy parameter can be explained by structural perturbation of the bilayer hydrophobic region due to incorporation of a compound [42]. The dependence of the DPH probe on temperature is presented in Fig. 4a. The presence of CGA practically does not change fluorescence anisotropy—we observed only slight decrease in the anisotropy of the gel-crystalline phase without changing T m. Additionally, we investigated the effect of CGA on fluidity of SUV formed of DMPC with TMA-DPH probe. While DPH is supposed to distribute in the hydrophobic bilayer, TMA-DPH is anchored at the hydrophobic–hydrophilic interface of the bilayer and fixed in the outer region of the acyl chains [43]. The fluorescence anisotropy of TMA-DPH in SUV as a function of temperature is presented in Fig. 4b. CGA practically does not change the temperature of the phase transition of DMPC, and does not change anisotropy. These results suggest that the investigated compound practically has no influence on fluidity in the hydrophobic region of the lipid bilayer, i.e., in the area of hydrocarbon chains.

Fig. 4
figure 4

Fluorescence anisotropy of DPH in SUVs prepared from DPPC (a), and fluorescence anisotropy of TMA-DPH in SUVs prepared from DMPC (b), as a function of temperature

Full size image

Using the Laurdan and Prodan probes, we investigated the degree of order in the hydrophilic part of liposomes formed of DMPC and DPPC in the absence and presence of cholesterol. These probes are particularly sensitive to polarity of the environment and monitor relevant differences in polarities of the different phase states of lipid bilayers. Laurdan is located in the hydrophilic–hydrophobic interface of the bilayer, and its fluorophore locates near the phospholipid glycerol groups and is sensitive to polarity changes and dynamic properties at the membrane lipid–water interface [44]. The calculated values of generalized polarization (GP) of the Laurdan probe for SUV formed of DPPC and DPPC/chol lipids are presented in Fig. 5a, b. The values of GP are high and positive in the gel phase, and are much lower in the liquid-crystalline phase. CGA slightly lowers the GP coefficient both in the gel- and liquid-crystalline phases but does not change the temperature of the phase transition. Similar changes were observed for membranes containing 10 mol% cholesterol and for membranes formed from DPPC and DMPC (Fig. 6a).

Fig. 5
figure 5

Values of general polarization of Laurdan (a) and Prodan (b) in DPPC SUV in the absence (open symbol) and the presence (filled symbol) of 10 mol% cholesterol, unmodified and modified with CGA in concentrations of 140 μM, as a function of temperature

Full size image
Fig. 6
figure 6

Values of general polarization of Laurdan (a) and Prodan (b) in DMPC SUV unmodified and modified with CGA in concentrations of 140 μM, as a function of temperature

Full size image

Fluorescence intensity of the Prodan probe, like that of Laurdan, depends on polarity of the medium and exhibits a red shift—a fact that signifies a change of phase in the membrane and enables GP determination. Compared to Laurdan, Prodan has a different location in the lipid membrane, which is closer to the aqueous phase. Such a location allows Prodan to be also sensitive to the pretransition that occurs in the lipid polar head-groups region [44]. For that reason, we observed the occurrence of two phase transitions: a pretransition (lower-temperature transition) is connected with a change of packing and hydration of the lipid groups and the main phospholipid phase transition (higher-temperature transition). Figs. 5b and 6b allow to compare the values of GP coefficients calculated on the basis of fluorescence intensity measured at different temperatures with the Prodan probe. For liposomes formed of DPPC, large changes in GP values are observed in the gel phase. At the concentration of 140 μM, CGA does not remove the pretransition totally but changes its temperature (Fig. 5b) without changing T m. These results suggest that CGA strongly influences the gel state of the DPPC bilayer. The presence of CGA also causes a slight decrease in GP values obtained for liposomes formed of DPPC with 10 mol% cholesterol added, both in the gel and crystalline phases. Similar results were observed for SUV made of DMPC (Fig. 6b).

Conclusions

In our study, we analyzed the effect of CGA on thermotropic phase behavior of lipids and their localization in the lipid bilayer by means of DSC and fluorescence spectroscopy. We observed slight decrease of the lipid gel–liquid-crystalline phase transition temperature T m and small change of the transition enthalpy, together with a very slight broadening of the transition peaks. This suggests that CGA does not penetrate deep into the lipid bilayer and probably is located in the polar head-group region of the membrane. This conclusion is confirmed by the fluorimetric studies with DPH, TMA-DPH, Laurdan, and Prodan probes. CGA has almost no influence on fluorescence anisotropy both in the hydrophobic area and hydrophobic–hydrophilic interface of the lipid bilayer, and thus it does not change the fluidity of the bilayer in the hydrophobic area. At the same time, it slightly decreases GP, which indicates an increased disorder in the hydrophilic parts of the lipid bilayer. These effects are most pronounced in the presence of cholesterol.

In conclusion, one can infer that the protective action of CGA with respect to biological membranes depends on the extent of its incorporation in the hydrophilic part of the membrane, which is in agreement with the results we obtained for polyphenolic compounds [34].