Physicochemical properties, antioxidant action and practical application in fresh cheese of the solid inclusion compound γ-cyclodextrin·quercetin, in comparison with β-cyclodextrin·quercetin

2.3.1 β-CD·Quercetin (1:1)

100 mg (0.076 mmol) of β-CD were dissolved in 10 mL of water at 37 °C. Separately, 23 mg (0.076 mmol) of quercetin were dissolved in 13 mL of ethanol at 37 °C. After full dissolution of both compounds, solutions were mixed, stirred for 3 h and subject to snap freezing (with liquid nitrogen). Solvent removal by freeze-drying afforded a voluminous pale-yellow solid in quasi-quantitative yield.

Anal. Calc. for (C₄₂H₇₀O₃₅)(C₁₅H₁₀O₇)·18(H₂O) (1760.6): C, 38.85; H, 6.640%; Found: C, 38.62; H, 6.813%. Thermogravimetric analysis up to 135 °C shows a mass loss of 13.7%, corresponding to the evaporation of 13 water molecules.

FT-IR (KBr, cm⁻¹): ν(tilde) = 3386vs, 2929m, 1660m, 1612m, 1560m, 1523m, 1450m, 1410m, 1383m, 1324m, 1305sh, 1263m; 1201m, 1157s, 1102sh, 1080s, 1029vs, 1002sh, 947m, 939m, 864m, 843m, 826m, 794w, 756m, 706m, 668w, 656w, 639w, 606m, 578m, 531m, 478w, 445w, 419w, 408w, 400w, 385w, 376w, 359w, 354w, 336w, 328w, 314w, 303w, 297vw, 290w.

Raman: ν(tilde) = 1660w, 1605vs, 1589m, 1547s, 1439s, 1399m, 1371m, 1327m, 1316m, 1294w, 1267w, 1219w, 1174w, 1112w, 1093vw, 1014vw, 994vw, 942m, 843m, 784m, 721w, 706w, 685 w, 660w, 639m, 603s, 574m, 520m, 478m, 471m, 253w, 231w, 209 vw.

¹³C{¹H} CP/MAS NMR: 174.1 (guest, C₄), 163.8 (guest, C₇), 157.2 (guest, C₅), 154.6 (guest, C₉), 147.9 (guest, C_4′), 146.2 (guest, C₂), 141.5 (guest, C_3′), 135.1 (guest, C₃), 125.9 (guest, C_1′), 122.0 (guest, C_6′), 115.5 (guest, C_5′), 112.1 (guest, C_2′), 104.2, 103.3, 102.2, 101.4 (all β-CD, C₁), 95.9 (guest, C_6,8), 84.2, 83.3, 82.3, 81.2, 80.4, 78.4 (all β-CD, C₄), 76.3, 75.2, 73.8, 73.0, 72.6, 71.6 (all β-CD, C_2,3,5), 63.7, 62.2, 60.3, 59.6, 58.7 (all β-CD, C₆) ppm.

NOTE: For comparison, the ¹³C{¹H} CP/MAS NMR spectrum of quercetin was also collected. The resonances of pure quercetin are as follows: 174.4 (C₄), 164.0 (C₇), 157.4 (C₅), 154.7 (C₉), 148.2 (C_4′), 146.1 (C₂), 141.3 (C_3′), 135.2 (C₃), 125.9 (C_6′), 121.8 (C_1′), 115.6 (C_5′), 112.2 (C_2′), 101.2 (C₁₀), 95.7 (C_6,8) ppm (see Fig. 1a for carbon labelling of quercetin).

2.3.2 γ-CD·Quercetin (1:1)

A solution of 200 mg of γ-CD (0.14 mmol) in 30 mL of water:ethanol (1:2), at ambient temperature, was added stepwise with solid quercetin (42 mg, 0.14 mmol). After full dissolution, the solution was stirred for 2 h, snap-frozen and freeze-dried to obtain a pale yellow solid in quasi-quantitative yield.

Anal. Calc. for (C₄₈H₈₀O₄₀)(C₁₅H₁₀O₇)·16(H₂O) (1886.6): C, 40.07; H, 6.517%; Found: C, 39.46; H, 6.178%. Thermogravimetric analysis up to 115 °C shows a mass loss of 16.2%, corresponding to the evaporation of 17 water molecules.

FT-IR (KBr, cm⁻¹) ν(tilde) = 3388vs, 2928m, 1657m, 1630m, 1602m, 1566m, 1518m, 1446m, 1416m, 1374m, 1324m, 1250m; 1200m, 1159s, 1126sh, 1102s, 1080s, 1052s, 1026vs, 1002sh, 942m, 936sh, 860m, 843m, 815m, 760m, 706m, 668w, 638w, 605m, 583m, 529m, 479w, 446w, 413w, 362w, 290vw, 283w.

Raman: ν(tilde) = 1564w, 1606vs, 1590m, 1567m, 1548m, 1441s, 1401s, 1371m, 1318s, 1267w, 1217w, 1175w, 1112m, 1084w, 995w, 942w, 843w, 784w, 721w, 707w, 687w, 659w, 638m, 602s, 575m, 518m, 479s, 442w, 401w, 355w, 331w, 228w, 200w.

¹³C{¹H} CP/MAS NMR: 174.7 (guest, C₄), 164.3, 163.2 (guest, C₇), 160.2 (guest, C₅), 156.4 (guest, C₉), 146.9, 146.0 (guest, C_4′), 144.9, 144.1 (guest, C₂), 143.5 (guest, C_3′), 134.6 (guest, C₃), 121.6 (guest, C_1′), 120.0 (guest, C_6′), 115.0 (guest, C_5′+2′), 103.1 (γ-CD, C₁), 97.5 (guest, C₆), 95.0 (guest, C₈), 81.5 (γ-CD, C₄), 72.8 (γ-CD, C_2,3,5), 60.2 (γ-CD, C₆).

3.2.1 Vibrational spectroscopy

Quercetin possesses several functional groups which are likely to suffer shifts by encapsulation. The carbonyl stretching band, ν(C⚌O), is particularly well-observed in Fourier Transform infrared spectroscopy (FT-IR). It appears at relatively low frequency (1664 cm⁻¹) which, in early investigations, was attributed to an intramolecular C5—O—H⋯O⚌C hydrogen bond (Fig. 1b), referred to as a “conjugate-chelate” type (Briggs and Colebrook, 1962). The bond would, however, be quite strong, implying a ν(C—O) frequency even lower than the one observed. A more realistic model was recently presented. Structural studies of solid anhydrous quercetin using a combination of molecular modelling, PXRD and solid-state NMR data (Filip et al., 2013) have proposed that in solid quercetin there is a network of intra- and intermolecular hydrogen bonds, with a more homogeneous charge distribution and lower C—O polarisation (Fig. 1c).

After inclusion into the CDs the ν(C—O) band of quercetin suffers a shift, as well as other quercetin bands. These changes are listed in Table 1. Note how the spectrum γ-CD·quercetin presents several shifted guest bands. These bands are associated with stretching modes of guest aromatic rings, thus suggesting that the entire quercetin molecule is included into the wide cavity of γ-CD. In comparison, the host-guest interaction occurring in β-CD·quercetin appears to be weaker, as only two oscillators present shifts. The C—O stretch is redshifted by −4 cm⁻¹ and the C5—OH stretch is blueshifted by +5 cm⁻¹. These changes suggest the disruption of the network hydrogen bonds found in pure quercetin and to the formation of a new hydrogen bond with β-CD, most likely at the secondary hydroxyls located at its larger rim of this macrocycle, lowering the ν(C—O) frequency due to the increased degree of polarisation. The interaction of quercetin with β-CD is thus postulated to occur superficially, at the larger rim, instead of involving inclusion of quercetin into the cavity. This proposed interaction model is coherent with previous ¹H NMR studies in solution which show partial inclusion of quercetin into β-CD (Zheng et al., 2004), and also with the solid-state ¹³C{¹H} CP-MAS NMR data herein presented (Section 3.2.3).

The Raman spectrum of γ-CD·quercetin further confirms the inclusion. The carbonyl stretch occurs at 1654 cm⁻¹, redshifted by −6 cm⁻¹ in comparison with pure quercetin; other bands are blueshifted by c.a. +2 cm⁻¹, namely the C-C bands of rings B and A, found at 1548 and 1441 cm⁻¹ for γ-CD·quercetin and at 1546 and 1439 cm⁻¹, respectively, for pure quercetin.

3.2.2 Powder X-ray diffraction and morphology

The diffraction pattern for freeze-dried γ-CD·quercetin (Fig. 2) shows a quasi-amorphous material that crystallises as a new phase upon rehydration by storing the solid for a few hours in a vapour-saturated chamber. Note that this phase (Fig. 2d) is free from reflections of pure quercetin or γ-CD heptahydrate, evidencing that γ-CD·quercetin is a true and pure inclusion compound (Caira, 2001). Its stability against disruption by moisture demonstrates strong host-guest affinity. To further demonstrate the stability, the inclusion constant of γ-CD·quercetin was determined (see details in section S3 of the Electronic Supplementary information), having a value of K_app = 1805.4 M⁻¹.

Close

Figure 2 Experimental powder X-ray diffraction patterns of (a) quercetin, (b) γ-CD heptahydrate, (c) the freeze-dried γ-CD·quercetin product (obtained in an amorphous form) and (d) the same product after rehydration (in a vapour-saturated chamber).

Export to PPT

The diffraction pattern of the hydrated sample of γ-CD·quercetin provides information on the cyclodextrin packing mode. Note that the pattern shares good similarities with that of γ-CD·12-crown-4-ether (Kamitori et al., 1986), a model compound in which the units of inclusion complex are stacking to form channels (Fig. 3). According to isostructurality principle of CD inclusion compounds reported by Caira (Caira, 2001) such similarity implies that the γ-CD·quercetin complex units are also channel-packed.

Close

Figure 3 Comparison of the experimental powder diffraction pattern of the rehydrated sample of γ-CD·quercetin with that of a model inclusion complex, γ-CD·12-crown-4-ether. The trace of the model compound was calculated using the software Mercury (v3.5.1, ©CCDC 2001–2014) on the atomic coordinates of γ-CD·12-crown-4-ether (Kamitori et al., 1986). Structural details of this model complex are depicted above the diffractrogram, with the guest highlighted in purple colour.

Export to PPT

Freeze-dried β-CD·quercetin is a fully amorphous material, which, upon rehydration, evidences recrystallisation of a small portion of β-CD and quercetin, indicative of low host-guest affinity (Fig. S1 in the Electronic Supplementary Information). The inclusion constant (K_app) values for β-CD·quercetin, obtained from phase solubility studies, varies between 402 M⁻¹ (Pralhad and Rajendrakumar, 2004) and 602 M⁻¹ (Jullian et al., 2007), indicating a mild affinity. This means that water molecules in the saturated chamber are able to displace some of the quercetin molecules out of the β-CD cavities, leading these to re-crystallise as pure quercetin.

The powdered samples of pure quercetin, γ-CD heptahydrate (as-received from the manufacturer) and rehydrated γ-CD·quercetin were studied by scanning electron microscopy (SEM) to evaluate the following: i) crystal morphology, ii) crystal integrity, and iii) material homogeneity. Quercetin is entirely composed of very small needle-shaped crystals (Fig. 4b and e). The pure γ-CD host presents a good number of fragments of crystals and some particles with the rough shape of parallelepipeds albeit with rounded edges and multiple cracking on the surface (Fig. 4a and d). Given the high solubility of this cyclodextrin in water, the fragmentation and cracking are likely to reflect a fast drying method employed by the manufacturer to obtain it quickly and in a good yield. Regarding the inclusion compound, it is possible to observe the formation of microcrystals with plate-like shape throughout the bulk material (Fig. 4c and f). This morphology is coherent with the observations made by powder X-ray diffraction that evidence a new crystalline phase for this inclusion compound.

Close

Figure 4 SEM images of γ-CD (a and d), quercetin (b and e) and rehydrated γ-CD·quercetin (c and f) at two different magnifications: the images on the top line of the figure show a broader view at low magnification (scale bar represents 10 μm) and the images on the bottom line show the same samples with a higher magnification (scale bar represents 2 μm) for a more detailed view of the morphology.

Export to PPT

3.2.3 ¹³C{¹H} CP-MAS spectroscopy

The solid-state ¹³C{¹H} CP-MAS spectra of γ-CD·quercetin, pure quercetin and γ-CD are depicted in Fig. 5. Note how the resonances ascribed to the host present multiple signals in the case of pure γ-CD and appear as single signals in the spectrum of the inclusion compound, indicating that the ring of γ-CD has adopted a more symmetric conformation to present a larger cavity, able to accommodate quercetin better.

Close

Figure 5 Solid-state ¹³C{¹H} CP-MAS spectra of γ-CD·quercetin, γ-CD and pure quercetin. Spinning side bands are marked with asterisks.

Export to PPT

The resonances ascribed to the guest, quercetin, also feature relevant changes in the spectrum of γ-CD·quercetin, as listed in Table 2. The signals ascribed to the carbons 2′ and 5′ of the catechol (B ring) are superimposed while other carbons are presenting two or multiple resonances, namely C2, C4′, C7′ and the carbons 6 + 8, which give rise to a single resonance in pure quercetin. This indicates an overall chance in the chemical environment around these molecules and the occurrence of two or more different geometries of inclusion (Braga et al., 2002).

In the spectrum of β-CD·quercetin, the resonances of both the host and the guest carbons appear mostly unchanged in comparison with those observed in the spectra of pure quercetin and β-CD hydrate (see Table 2 and Fig. S2 in the Electronic Supplementary Information). This indicates a weak host-guest interaction and further confirms that quercetin is located outside the cavity of β-CD, rather interacting with its rim as postulated from FT-IR data.

3.2.4 Thermogravimetry

Thermal decomposition profiles for the pure CDs, quercetin, β-CD·quercetin, γ-CD·quercetin and their 1:1 physical mixtures (PMs) were collected and compared (Figs. 6 and 7).

Close

Figure 6 TG traces for β-CD, quercetin, β-CD·quercetin and the 1:1 physical mixture of β-CD and quercetin.

Export to PPT

Close

Figure 7 TG traces for γ-CD, quercetin, γ-CD·quercetin and the 1:1 physical mixture of γ-CD and quercetin.

Export to PPT

The TG traces of the pure hosts, β-CD (Fig. 6) and γ-CD (Fig. 7), are characterised by an initial dehydration step, which starts at ambient temperature and proceeds up to 75 °C for β-CD and to 95 °C for γ-CD. Both CDs have a mass loss from dehydration of 15%, that translates into 11 hydration waters for β-CD and 12 for γ-CD. Quercetin dehydrates between 65 and 135 °C, having no further mass loss until 290 °C, which marks the onset of its thermal oxidation (Da Costa et al., 2002; Masek and Chrzescijanska, 2015). From 360 °C onwards, there is a second and steeper degradation step and, at 515 °C, all quercetin is decomposed. The steps associated with quercetin’s oxidative and thermal degradation are absent in both the PMs and the adducts, evidencing the interaction of quercetin with the CDs. Decomposition of the host occurs around 250 °C for β-CD·quercetin and its PM, and at 270 °C for γ-CD·quercetin and its PM. The traces of the inclusion compounds differ from those of PMs only in the dehydration process which takes place in two separate steps in the PMs (CD and quercetin waters have different evaporation temperatures). In γ-CD·quercetin, mass loss from dehydration is 16%, corresponding to 17 water molecules, in good consonance with microanalysis and being higher than the water content of the PM and of pure γ-CD. The stacking of γ-CD molecules into channels (shown in PXRD), in turn arranged into squares, leaves wide inter-channel spaces suitable to accommodate a large number of water molecules (Fig. 3, inset).

3.3.1 Evaluation of the lipid peroxidation inhibition with the TBARS assay

Lipid peroxidation, by accumulation of reactive oxygen species (ROS) such as the superoxide anion, hydroxyl radicals, nitric oxide and the peroxy radical, is an important mechanism of food deterioration (Antolovich et al., 2002). The TBARS assay is an indirect measure of a compounds’ ability to block ROS and a good first approach to gather evidence on their capability to interfere with oxidation processes occurring in food matrices (Fig. 8a).

Close

Figure 8 Antioxidant action of γ-CD·quercetin, compared with pure quercetin and β–CD·quercetin. (a) Results of the lipid peroxidation assay using buffered egg yolk; (b) results of the DPPH• scavenging assay, measured at 20 min of incubation. Data points represent the mean ± SD of three independent assays.

Export to PPT

β-CD·quercetin displayed the highest percentage of inhibition, followed by quercetin and γ-CD·quercetin. The activity is quasi dose-dependent; that is, it increases with the concentration, up to 0.8 mM. At higher concentration, the activity of β-CD·quercetin continues to increase but not in a dose-dependent mode, while the activities of quercetin and γ-CD·quercetin start to decrease, possibly due to occurrence of a pro-oxidant mechanism at a high concentration. The different behaviour of β-CD·quercetin may result from the fact that quercetin is more readily available in this adduct since it is not located in the host cavity. In addition, quercetin can dissociate easily (as demonstrated by powder diffraction after contact with moisture), being able to act freely while having higher solubility in the reaction medium due to the presence of β-CD.

3.3.2 Evaluation of radical scavenging properties with the DPPH assay

In this assay the test compound scavenges the free radical DPPH^•, causing the test solution to change colour and thus allowing the reaction progress to be easily monitored by UV–Vis spectroscopy (refer to Section S2 of the ESI for details). The percentage of unreacted DPPH^• at different concentrations of quercetin or inclusion compounds is given in Fig. 8b. The EC₅₀ is determined by simple interpolation.

The t_EC50 and the antiradical efficiency (AE) are presented in Table 3. These parameters can be determined by measuring the kinetics of inhibition using the concentration which corresponds to the EC₅₀ of the test compound (calculation details given in the Experimental Section).

Quercetin is a very efficient radical scavenger even at low concentrations. At concentrations above 40 μM, the activity reaches a plateau (not shown) and, overtime, quercetin may eventually begin to act as a pro-oxidant. The inclusion compounds present very similar DPPH^• scavenging curve profiles, both requiring high concentrations to exhibit radical scavenging action. The kinetics of reaction is, however, faster, as 50% of the DPPH^• is bleached in roughly one third of the time required with quercetin. The antiradical efficacy values (AE), weighing these two previous factors, are thus slightly higher for the inclusion compounds, particularly for β-CD·quercetin.

3.4.1 Sensory evaluation

Fortification of fresh cheese with CD·quercetin, even in a low percentage (0.03–0.04% of mass), produced minor changes in its organoleptic characteristics, particularly in the colour and texture. Fortified fresh cheese was more compact and less “watery”, likely owing to the ability of cyclodextrins to retain water molecules and to improve the texture and flavour of low fat products (Reineccius et al., 2004). Fortified fresh cheese presented a yellow colour and a “dotted” aspect, particularly with γ-CD·quercetin.

In the product preference test (refer to Table S1 for details), fortified and untreated fresh cheeses scored very similar, with p-values of 0.3 for β-CD·quercetin, and 0.4 for γ-CD·quercetin. These values, calculated for one degree of freedom (fortified vs non-treated), are far above the limit of statistically significant difference (p < 0.05) thus meaning that the incorporation of CD·quercetin does not cause significant differences to the acceptance of fresh cheese by tasters.

3.4.2 Shelf-life assessment of fresh cheese fortified with CD·quercetin adducts

To evaluate the stability over the time of storage, the pH of untreated and fortified fresh cheese samples was monitored once a week over three weeks (refer to Table S2 of the electronic supplementary Information for the list of measured pH values). After one week of storage, non-treated samples presented pH values around 5.5 and the fortified ones had pH around 5.7. After the second week, however, all the cheese samples had pH > 5, evidencing a degradation process most likely caused by fermentation of residual lactose to lactic acid by either pasteurisation-resistant indigenous bacteria in milk or post-production contamination with environmental bacteria. It can thus be concluded that the shelf-life of fresh cheese is not altered by the fortification with CD·quercetin adduct.