Biomolecules, Vol. 12, Pages 1762: Structural Studies of Piperine Inclusion Complexes in Native and Derivative β-Cyclodextrins

1. IntroductionBlack pepper (Piper nigrum) is one of the most important spices worldwide due to its many biological properties such as antioxidant, antimicrobial, anti-inflammatory, and anticancer activities [1]. Its characteristic pungent taste and flavor is attributed to the presence of the alkaloid piperine (1-(5-[1,3-benzodioxol-5-yl]-1-oxo-2,4-pentadienyl) piperidine), for short PN, Figure 1a). Although piperine is consumed as a dietary spice and has been considered as functional food [2], it has also many pharmaceutical benefits. More specifically, piperine is known as a phytochemical and antimicrobial [3] or antifungal agent acting as an inhibitor for certain enzymes [4], but it also affects the activity of catalase and glutathione peroxidase enzymes [5] preventing in this way the outbreak of negative biological procedures and diseases like Parkinson and Alzheimer [6,7]. However, its ability to reduce the risk of developing certain cancers [8] and to exhibit anticarcinogenic effects [9] have recently drawn great attention. Piperine can reverse multidrug resistance (MDR) in cancer cells and acts as a bioavailability enhancer for many chemotherapeutic agents [10], as many studies indicate that piperine shows synergistic effects when taken in combination with various classes of drugs [11].However, its potential application in functional foods and pharmaceutical formulations is incommoded as piperine is practically insoluble in water, slightly soluble in other permissible pure solvents [12] and unstable in ultraviolet light [13]. An already tested approach, in order to improve the physicochemical properties of piperine (i.e., increase its water solubility and protect it from degradation) and thus effectively enhance its bioavailability and activity, is by the formation of piperine inclusion complexes in suitable cyclodextrins (CDs) [14,15]. CDs are amphiphilic cyclic oligosaccharides consisting of at least six D-(+) glucopyranose units attached by α-(1, 4) glycosidic bonds. Their distinctive round conformation of a truncated cone facilitates the encapsulation of certain guest molecules inside their interior cavity. Beta-cyclodextrin (β-CD), which is comprised of seven glucose units is the most common representative in food and pharmaceutical industries due to its suitable cavity size for the accommodation of several food ingredients or drugs, and its low cost. The hydroxyl groups of its macrocycle rims participate in several intra- and intermolecular (host-host and host-guest) hydrogen bonds. Methylation of these hydroxyls in both rims results in its methylated derivatives that are characterized by significantly higher water solubility and flexibility compared to the parental β-CDs. (Heptakis(2,6-di-O-methyl)-β-cyclodextrin (DM-β-CD), heptakis(2,3,6-tri-O-methyl)-β-cyclodextrin (TM-β-CD) and randomly methylated β-cyclodextrin (RM-β-CD) (Figure 1b). Another β-CD derivative, obtained by substituting the aforementioned hydroxyls with 2-hydroxypropyl groups, is the 2-hydroxypropyl-β-Cyclodextrin (HP-β-CD), which is also well known for its pharmaceutical applications (Figure 1b). CDs have a distinct role in the pharmaceutical industry, either by their direct use as drugs [16] or as drug carriers participating in nanoparticle (NPs) formulations [17].The inclusion of piperine in native β-CD has been studied by various spectroscopic (phase solubility studies, Job’s Plot, FT-IR, near-infrared spectroscopy (NIR), Raman, 1H-NMR, UV-visible absorption and fluorescence intensity study, powder X-ray diffraction (PXRD)) and calorimetric (differential scanning calorimetry (DSC) and thermogravimetry (TG)) methods as well as by scanning electron microscopy (SEM) [18,19,20,21,22,23]. The above methods indicated the formation of PN/β-CD complexes of either 1:1 or 1:2 guest:host stoichiometry, according to the interactions found between the aromatic ring of PN and β-CD, with complexation efficiency varying from 7 to 70% for various stoichiometric ratios. Moreover, the dissolution rate behavior of the complex was examined via dissolution testing assays. Although in the above-mentioned works, the inclusion complex has been also studied in the solid state, no crystal structure of piperine complexes, either in native β-CD or in β-CD derivatives has been presented so far. In this work, the inclusion compounds of piperine in β-CD, DM-β-CD and TM-β-CD are investigated by single crystal X-ray diffraction (SC-XRD), revealing the stoichiometry, interactions and geometrical details of the complex in the crystalline state. Based on the crystallographically determined coordinates, molecular dynamics (MD) studies have also been performed in order to monitor the dynamic behavior and the stability of the complexes in aqueous environments and in the absence of crystal contacts. Finally, phase solubility studies in aqueous solution were carried out for PN/β-CD, PN/RM-β-CD and PN/HP-β-CD inclusion complexes, in order to examine the solubility profile and estimate the apparent stability constant (K1:1) and the compexation efficiency (CE) for these complexes. In the case of PN/HP-β-CD, where no crystal structure is available, the 1:1 guest:host stoichiometry, indicated by its solubility profile, was used for the preparation of the docked model which was further examined by MDs.The complementary structural analysis by experimental (X-ray crystallography) and theoretical (MD) studies presented in this work, sheds light on the structure-stability relationship of the examined “spicy” cyclodextrin inclusion complexes. Our understanding of the structural aspects of these complexes may be useful in the engineering of modified guest-host preparations with optimized pharmacological properties and shape future piperine applications [24]. 2. Materials and Methods 2.1. Materials

Piperine (97% pure) as a light yellow powder was purchased from Merck KGaA (Damstadt, Germany), while β-CD, DM-β-CD, TM-β-CD, RM-β-CD (degree of substitution (DS) ~12) and HP-β-CD (DS~4.5) of pharmaceutical grade quality as white powders were from Cyclolab Ldt. (Budapest, Hungary). Double distilled water was utilized for the preparation of all the examined solutions.

2.2. Phase Solubility StudyThe UV–visible (UV–Vis) spectrophotometer BK-S380 (BioBase Group, Jinan, Shandong, China) was utilized for all spectroscopic analyses concerning the solubility of pure PN and the phase solubility studies. The calibration curve was obtained at the visible absorption maximum of the PN (345 nm) as follows: Five standards of PN solution in a 1:1 methanol: water ratio corresponding to concentrations of 0.00125, 0.0025, 0.005, 0.01 and 0.02 mM were prepared and measured at 345 nm. The calibration curve was depicted by plotting the absorbance against the above PN concentrations and by applying the linear regression analysis according to a previously described procedure [25]. Subsequently, phase solubility studies were carried out according to the method reported by Higuchi and Connors [26]. More explicitly, an excess amount of PN (50 mg) was added to 10 mL of deionized water containing various concentrations between 1 to 15 mM for β-CD and 1 to 60 mM for both RM-β-CD and HP-β–CD. The solutions were further mixed using an orbital shaker (PHOENIX Instrument Laboratory Shaker RS-OS 5, Berlin, Germany) at 25 °C for 48 h to ensure equilibrium, and then the solutions were passed through a 0.45-μm filter to remove the undissolved solids. The filtered solutions were appropriately diluted with a 1:1 (v/v) methanol: water solution and measured at 345 nm. 2.3. Single-Crystal Preparation

In the crystal formation process of native CD supramolecular complexes, the slow cooling method, where the temperature of an aqueous saturated mixture solution of CD and the guest molecule is gradually decreased from 343 K to ambient temperature, was followed. More specifically, 20 mg of β-CD (0.0175 mmoles) were weighted into vials and 2 mL of distilled water was added. An equimolar quantity (5.0 mg, 0.0175 mmoles) of PN was added and the mixture was stirred for about four hours at 343 K until it was limpid. Clear light colourless prismatic-like crystals suitable for X-ray data measurements were obtained over a seven-day period.

On the other hand, the slow evaporation method, which is more suitable for crystallizing inclusion complexes of methylated CDs, was used in the cases of DM-β-CD and TM-β-CD inclusion complexes. Briefly, a suitable amount of PN was added to aqueous solutions of DM-β-CD or TM-β-CD (0.004 M) at 1:1 host: guest mole ratios. The two mixtures were stirred for 1 h at room temperature and subsequently maintained at 321 K for a period of one week. Clear light colourless rod-like and prism-like crystals, suitable for X-ray data collection, were obtained in the case of PN/DM-β-CD and PN/TM-β-CD, respectively.

2.4. X-ray Diffraction Experiments

Data collection was performed using CuKa radiation (λ = 1.54178 Å) in a Bruker D8-VENTURE diffractometer equipped with the CMOS-based detector PHOTON 100. The tested specimens were harvested from the mother liquor, cryo-protected by rapid immersion in paraffin oil and flash frozen using a continuous nitrogen-flow cooling device (Oxford Cryosystems Ltd., Long Handorough, UK) at 100 or 120 K.

The data were processed with the Bruker SAINT Software package [27] using a narrow-frame algorithm and were corrected for absorption effects using the multi-scan method (SADABS) [28].In the case of the PN/β-CD complex, a total of 2494 frames were collected in a total exposure time of 15.1 h. The crystal was twinned. Based on a determined triclinic unit cell, two main domains were detected related by a rotation angle of 179.91°. The integration of the images using both domains and the subsequent scaling of the data by using the TWINABS program [29] yielded 16,593 and 17,055 total reflections with I/σ(I) of 33.0 and 33.2 for domain 1 and 2, respectively, whereas 17,875 reflections were attributed as composites (overlapped reflections from domain 1 and 2). The twin fraction was close to 0.5. The resolution of the measured data was up to 0.84 Å. However, due to the low completeness of the measured data set (94.4%), a lower resolution limit of 0.86 Å with a completeness of 98%, was used for the refinement of the crystal structure. The final cell constants and refinement details are listed in Table 1.The X-ray diffraction from a crystal of PN/DM-β-CD complex, resulted in a total of 2860 frames which were collected in 21 h. A monoclinic unit cell of P21 space group symmetry was defined and the integration procedure yielded a total of 59,020 reflections to a maximum θ angle of 63.89° (0.86 Å resolution), of which 12,372 were independent (average redundancy 4.770, completeness 99.1%, Rint = 7.19%, Rsig = 5.26%) and 10,678 (86.31%) were greater than 2σ(F2). The final cell constants, quoted in Table 1, are based upon the refinement of the XYZ-centroids of 9685 reflections above 20 σ(I) with 7.934° Finally, in the case of the crystal of PN/TM-β-CD complex, a total of 1706 frames were collected in 14 h. The unit cell was also a monoclinic P21 and the integration of the collected data yielded a total of 73,723 reflections to a maximum θ angle of 50.66° (1.00 Å resolution). The final cell constants, based upon the refinement of the XYZ-centroids of 9801 reflections above 20 σ(I) with 6.544° Table 1 along with other refinement information. The structure solution of these CD complexes was based on the Patterson-seeded dual-space recycling utility of the SHELXD program [30]. The structures were refined by full-matrix least squares against F2 by using SHELXL-2014/7 [31] in the SHELXLE GUI [32]. Due to the limited resolution, structural complexity and disorder of the final models, soft restraints on bond lengths and angles, generated from the PRODRG2 webserver [33], were applied on the host and guest molecules of the inclusion complexes. Anisotropic displacement parameters were refined using soft restraints (SIMU) [34] implemented in the SHELXL program where necessary.

All hydrogen atoms were placed in geometric positions and treated as riding on their parent atoms with dC–H = 0.95–1.00 Å (depending on the hybridization of carbon atom) and dO–H = 0.84 Å. Uiso(H) values were assigned in the range 1.2–1.5 times Ueq of the parent atom. Hydrogen atoms of water molecules were not included in any of the final structural models. In an effort to maintain a relatively high (>6.0) data/parameters ratio, anisotropic thermal parameters were imposed to selected, non-H atoms of the host molecules. Extinction corrections were applied to the PN/DM-β-CD case, while 8, 33 and 12 reflections that exhibited poor agreement to the refined models were omitted in the PN/β-CD, PN/DM-β-CD and PN/DM-β-CD cases, respectively.

The programs Mercury [35], Pymol [36] and Olex2 [37] were used to explore the crystal packing, analyze the structure geometry and visualize the asymmetric unit, the intermolecular interactions and the crystal packing of the complexes. Crystallographic data are given in Table 1. Crystallographic information files with embedded structure factors have been checked and validated for the consistency and integrity of crystal structure determinations according to IUCr standards and have been deposited in the Cambridge Structural Database (CSD) under the deposition numbers CCDC: 2063126, 2022888 and 2063127. 2.5. Computational MethodsFive separate MD simulations involving piperine inclusion complexes with the native and three modified β-CDs, illustrating different guest:host ratios and different inclusion modes were carried out in an effort to monitor their time-resolved motions in aqueous media. For the PN/β-CD complex, the crystallographically determined atomic coordinates of four adjacent β-CD molecules (two of them forming a dimer) and two encapsulated guests (PN1 and PN2) inside their cavity comprise the staring model of a 2:4 guest:host stoichiometry. For the inclusion complexes of piperine in methylated CDs, PN/DM-β-CD and PN/TM-β-CD, the initial models were also based on the crystallographically determined atomic coordinates (both with a 1:2 guest:host stoichiometry). Finally, in order to examine the PN/HP-β-CD inclusion complex for which no crystal structure is available, two stable binding models from a docking analysis using AutoDoc Vina [38], both having an 1:1 guest:host stoichiometry but different inclusion modes, were used as the starting structures of MD simulations in case of PN/HP-β-CD. The 1:1 stoichiometry for this complex has been shown by Jadhav [39] using Job’s plot analysis and further supported by the AL-type profile in the present phase solubility studies. More explicitly, the HP-β-CD molecule with degree of substitution (DS) 5 was built using VEGAZZ [40] by removing the methyl groups and arbitrarily adding five 2-hydroxypropyl groups to both rims of DM-β-CD. In order to examine the effect of piperine ’s inclusion mode (insertion of the aromatic or piperidine ring in the cavity) in HP-β-CD, as the determination of a 3D structure is unattainable, two different models (with high docking scores) representing two different inclusion modes were selected as the starting systems for the PN/HP-β-CD case. Charges to 2-hydroxypropyl groups of modified CDs were applied with GAMESS [41]. Consequently, MD simulations were performed for the five complexes by using the Amber12 suite [42]. The q4 md-CD force field [43] was applied to all modified CD atoms, while atoms of native β-CD molecules were treated with GLYCAM [44]. The piperine geometry was optimised following the AM1BCC methodology with the program ANTECHAMBER. In all cases, the formed cyclodextrin inclusion complex was initially solvated with TIP3P waters [45] in a periodic, octahedral box forming a 10 Å thick water shell around the structure using xLEaP. Hydrogen atoms were also added with xLEaP in all systems.MD calculations and minimizations were carried out with SANDER. Periodic boundary conditions were imposed by means of the particle mesh Ewald method using a 10 Å limit for the direct space sum. The protocol included energy minimization for hydrogen atoms with positional restraints of 50 kcal mol−1 Å−2 on the non-hydrogen atoms, heating equilibration of the solvent in the canonical (NVT) ensemble using positional restraints and the Berendsen thermostat algorithm with coupling constants of 0.5 ps to control temperature and pressure, unrestrained energy minimization, gradual temperature increase from 5 to 300 K with 10 kcal mol−1 Å−2 restraints on the atoms of the inclusion complex, gradual release of the restraints in successive steps at 300 K and density equilibration in the isobaric-isothermal (NPT) ensemble for 250 ps. Subsequently, production runs using a Berendsen-type algorithm with coupling constants of 1.0 ps were carried out under physiological conditions for additional 12 ns in the NPT ensemble. Root mean square deviation (RMSD) calculations, as well as geometric (H-H bond distance monitoring) analysis of the examined systems were performed by the CPPTRAJ module [46] of Amber12 and VMD [47].The well-known post-processing single-average MM/GBSA methodology [48] implemented in AMBER [49] was used for investigating the binding free energy of PN to CD hosts in solution. In the MM/GBSA approach, the free energy ΔG binding = Gcomplex − Ghost − Gguest for the binding of the guest to the host to form the complex, can be expressed as [50]:

ΔGbinding = ΔH − T ΔS = ΔEMM + ΔGsolvation − T ΔS

(1)

where ΔEMM, ΔGsolvation and −T ΔS are the changes in the gas-phase molecular mechanics (MM) energy, solvation free energy, and conformational entropy upon PN encapsulation in the host cavity. The ΔEMM is achieved from the combination of the electrostatic (ΔEele) and van der Waals (ΔEvdW) energies, whereas the ΔGsolvation is calculated using Equation (2),

ΔGsolvation = ΔGGB + ΔGnonpolar

(2)

where ΔGGB is the electrostatic solvation energy (polar contribution) calculated using the GB model and ΔGnonpolar is the nonpolar contribution between the solute and the continuum solvent, estimated using the solvent-accessible surface area (SASA) [51,52]:

ΔGnonpolar = γ·SASA + b

(3)

The change in entropy –T ΔS is obtained using the NMODE module of AMBER [49]. However, the estimation of the entropy term is often problematic as the normal mode lacks information of the conformational entropy and alternative methods do not give converged results [48]. Thus, although calculated, this term is usually neglected in the comparison between binding affinities of similar inclusion complexes. 4. DiscussionX-ray crystallography studies of the PN inclusion complexes in β-CD and its methylated derivatives heptakis(2,6-di-O-methyl)-β-Cyclodextrin (DM-β-CD) and heptakis(2,3,6-tri-O-methyl)-β-Cyclodextrin (TM-β-CD), revealed the formation of inclusion complexes with 1:2 guest:host stoichiometry in the crystalline state. In all determined crystal structures, a piperine molecule is found encapsulated in an extended hydrophobic cavity formed by two hosts which are arranged in a tail-to-tail mode (narrow rim facing the narrow rim) in the case of PN/β-CD and in a head-to-tail mode (wide rim facing the narrow rim) in the cases of PN/DM-β-CD and PN/TM-β-CD. Interestingly, in all cases it is the guest PN molecule that mainly interconnects the two host CDs, threading their cavities. Native β-CD hosts usually tend to form head-to-head dimers via a network of intermolecular H-bonds between their secondary hydroxyls. The formation of these dimers is observed in the crystalline state of PN/β-CD complexes, however this dimeric cavity is filled by water molecules tethered to the protruding part of an encapsulated PN in the adjacent tail-to-tail β-CD couple. This inclusion mode, revealed by the determined crystal structure of the PN/β-CD complex, is consistent and further explains previous findings for the PN/β-CD complex by DSC and 1H-NMR studies. In particular, the existence of water molecules in the β-CD cavity have been indicated by previous DSC studies of the PN/β-CD complex [18], whereas the interactions between the PN aromatic ring and the H3 and H5 of the β-CD that have been previously reported by 1H-NMR studies [19] were also observed by monitoring the host-guest H-H proximity during the MD simulations performed in this work for the PN/β-CD complex in an aqueous environment. The H-bonds network, which forms the above mentioned head-to-head β-CD dimers, that interconnect two adjacent tail-to-tail complex units, is not maintained during the time-frame of the MD simulations, although the inclusion complexes remain stable. On the other hand, in the cases of the DM-β-CD and TM-β-CD hosts, this kind of head-to-head dimers cannot be formed due to the limitation or complete absence of hydroxyls in their rims. Thus, the couple of hosts in the corresponding complex units is arranged in a head-to-tail mode forming columns in the crystalline state.

The stability of the inclusion modes revealed by X-ray crystallography was examined in an explicit water environment and in the absence of crystal contacts by MD studies. All inclusion complexes remain stable during the 12 ns simulations and MM/GBSA calculations showed the high binding affinity of PN for the β-CD, DM-β-CD and TM-β-CD hosts in their 1:2 guest:host complexes. On the other hand, in MD simulations performed for the PN/HP-β-CD inclusion complex by using docked starting models of 1:1 guest:host stoichiometry, as indicated by the solubility profile of the complex, although no complex dissociation was observed, the estimated binding affinities were significantly lower. This result, although expected due to the 1:1 guest:host stoichiometry of the complex, is not consistent with the rank order of the PN affinity with β-CD, RM-β-CD and HP-β-CD hosts (RM-β-CD > HP-β-CD > β-CD), according to the estimated Kc and CE values by phase-solubility studies. The solubility profiles for these complexes indicate a BS-type for PN/β-CD and an AL-type for PN/RM-β-CD and PN/HP-β-CD, thus the CE and Kc calculations were made for 1:1 guest:host complexes.

Ezawa et al. [18] have indicated a 1:1 guest:host stoichiometry for PN/β-CD in solution by a Job’s plot. Moreover, in a following work of Ezawa et al. [19] a B-type solubility profile, similar to the one observed in this work, and a significant higher estimation binding constant of 3244 M−1 was reported for PN/β-CD. However, in a recent work by Alshehri et al. [60] a binding constant of just 287 M−1, estimated by an A-type solubility profile of PN/β-CD, was reported. This low value and the A-type profile is disputable for the following reason: It is well documented that CD inclusion complexes, especially those of the natural CDs, have tendency to self-assemble in aqueous solutions to form aggregates, giving rise to characteristic B-type phase-solubility diagrams [61]. The self-aggregation increases with increasing cyclodextrin concentration. The increase of β-CD concentration in the phase solubility study by Alshehri et al. is not high enough to display the B-curve plateau. By comparing the binding constant of 3244 M−1 estimated by Ezawa et al. [19] to the one presented in this work (1800 ± 300) M−1, we should note that Kc as well as CE values for the B-type complexes are estimated as 1:1 guest:host complexes by the limited linear portion of the profile. Moreover, the KC value determined by Equation (4) is strongly affected by the S0 value which is usually very inaccurate for compounds with S062] like piperine. All these may cause the difference observed between these estimations. In addition, the theoretical calculations of this work, based on the 1:2 guest:host complex revealed by crystallography in aqueous solution, resulted in a high binding affinity for this complex. These results can justify the high Kc value estimated by Ezawa et al. [19] considering that the stability constants obtained from phase-solubility diagrams are estimations for 1:1 complexes but are the most frequently observed constants that are composed of a number of true stability constants for multiple types of coexisting water-soluble drug complexes in the aqueous complexation media. Moreover, it is well documented that the self-aggregation that causes a B-type solubility profile, as the one observed for this complex, is usually observed for multi-component ternary and quaternary CD complexes [63]. Thus, according to our theoretical calculations, the overall Kc obtained by the phase solubility diagrams, is expected to be high (at least higher than that of PN/HP-β-CD) as estimated by Ezawa et al. [19].Moreover, in crystalline state, it is observed that the formation of β-CD channels hosting PN molecules bridged by water molecules which are entrapped in the head-to-head β-CD interfaces. Although this structure resembles the “molecular necklace” of a linear CD pseudopolyrotaxane obtained in the solid state [64], the inclusion compounds of the linear assembly threaded by the CD rings are not directly interconnected, resulting in an unstable “necklace” in solution. Theoretical studies by Anconi et al. [65] have shown that CD pair interaction plays a major role in the probability distribution of the entities formed in the self-assembly system of polymers threaded by CDs. Although in this case no polymer is threaded in the CDs’ nanotube, it is interesting to notice the following: A tail-to-tail inclusion mode, as revealed by the crystal structures of PN/β-CD presented in this work, and other similar structures that will be presented in forthcoming works (e.g., the inclusion complex of capsaicin in β-CD; unpublished data), results in a Bs-type solubility profile, meaning a low solubility of the complex caused by self-aggregation of the complex units. On the other hand, a 1:2 guest:(head-to-head) host inclusion mode, as observed in the crystal structure of cholesterol in β-CD [66] is correlated with an A-type solubility profile [67], indicating that the complex units do not aggregate so intensively. Thus, the inclusion mode in the complex units of the 1:2 guest:host β-CD dimers, in the sense of whether the wide or the narrow rims are the open ends interconnecting adjacent complex units, plays a crucial role in the self-aggregation and thus solubility of the inclusion complex. However, this topic has to be further investigated.For all the above reasons, it is safer to compare the KC and CE values estimated by phase solubility studies solely between the PN/RM-β-CD and PN/HP-β-CD inclusion complexes, that both present an AL-type solubility profile. From this comparison, a stronger binding affinity is indicated for PN in RM-β-CD than HP-β-CD. The low binding constant of PN/HP-β-CD in solution has also been reported by Imam et al. [68] estimating a value of 238 M−1 by phase solubility studies. Thus, in agreement with our MM/GBSA calculations, PN has lower affinity for HP-β-CD than methylated β-CDs. 5. Conclusions

The determination of the crystal structure of the PN inclusion complexes in β-CD, DM-β-CD and TM-β-CD presented in this work, gives unique and valuable information about the stoichiometry and geometry of the complexes that conclusively clarifies the inclusion mode of piperine in these host molecules. In all determined structures, inclusion complexes of 1:2 guest:host stoichiometry were found in the crystalline state. The guest PN molecule threads the hydrophobic cavities of the hosts which are arranged as couples in a tail-to-tail mode in the case of PN/β-CD and in a head-to-tail mode in the cases of PN/DM-β-CD and PN/TM-β-CD.

Complement MD studies were performed for the crystallographically determined structures in order to monitor the dynamic behavior and the stability of the complexes in an aqueous environment and in the absence of crystal contacts. All inclusion complexes remain stable during the 12 ns simulations and MM/GBSA calculations showed the high binding affinity of PN for the β-CD, DM-β-CD and TM-β-CD hosts. Moreover, MD simulations were performed for the PN/HP-β-CD inclusion complex by using docked starting models of 1:1 guest:host stoichiometry, as indicated by the solubility profile of the complex. By monitoring the trajectories during the time-frame of the simulations the complex also remains stable although the absolute values of the estimated binding affinities (neglecting the entropic term) were significantly lower than those of the other examined complexes.

Finally, phase solubility studies were carried out for the PN/β-CD, PN/RM-β-CD and PN/HP-β-CD inclusion complexes in order to examine the solubility profile and estimate the apparent stability constant (K1:1) and the complexation efficiency (CE) for these complexes in aqueous solution. The low solubility of PN/β-CD resulted in a BS solubility profile and the estimation of the K1:1 and CE values, under the assumption of a 1:1 guest:host complex, is quite disputable. Although the complex units in solution may differ from those in crystalline state, the overall 1:1 guest:host stoichiometry used in these estimations, resulted to low values relative to the other complexes. This is inconsistent with the high binding affinity values given by MM/GBSA end-state energy calculations that were based on the MD simulations of the 1:2 guest:host crystal structure in an aqueous environment. Thus, it is safer to compare the estimated K1:1 and CE values only between the PN/RM-β-CD and PN/HP-β-CD complexes, both of AL-type profile, which show a higher binding affinity for the former than the latter complex in accordance with the calculated values using the MM/GBSA method.

留言 (0)

沒有登入
gif