Inclusion of trans-resveratrol in methylated cyclodextrins: synthesis and solid-state structures

Summary The phytoalexin trans-resveratrol, 5-[(1E)-2-(4-hydroxyphenyl)ethenyl]-1,3-benzenediol, is a well-known, potent antioxidant having a variety of possible biomedical applications. However, its adverse physicochemical properties (low stability, poor aqueous solubility) limit such applications and its inclusion in cyclodextrins (CDs) has potential for addressing these shortcomings. Here, various methods of the attempted synthesis of inclusion complexes between trans-resveratrol and three methylated cyclodextrins (permethylated α-CD, permethylated β-CD and 2,6-dimethylated β-CD) are described. Isolation of the corresponding crystalline 1:1 inclusion compounds enabled their full structure determination by X-ray analysis for the first time, revealing a variety of guest inclusion modes and unique supramolecular crystal packing motifs. The three crystalline inclusion complexes were also fully characterized by thermal analysis (hot stage microscopy, thermogravimetric analysis and differential scanning calorimetry). To complement the solid-state data, phase-solubility studies were conducted using a series of CDs (native and variously derivatised) to establish their effect on the aqueous solubility of trans-resveratrol and to estimate association constants for complex formation.


Thermal characterization of trans-resveratrol (RSV)
Commercial resveratrol (RSV) is a microcrystalline white powder with a thermal profile typical of an anhydrous drug. In the DSC curve of commercial RSV, only one endothermic effect appeared at T peak,m = 266.3(4) °C (T onset,m = 265.1(3) °C; H m = 279(2) J g -1 ) due to the drug melting. The anhydrous nature of commercial RSV was confirmed by TGA analysis that revealed mass loss at 275 °C attributable to sample decomposition.
To further characterize the solid state of RSV, it was analysed by FT-IR spectroscopy and the spectrum obtained presents a broad band at 3184 cm -1 due to the O-H stretching vibration, a set of typical bands at 1605, 1581 and 1383 cm -1 corresponding to C-C aromatic double bond stretching, C-C olefinic stretching and C-O stretching, respectively. The band at 965 cm -1 is attributable to the trans olefinic bond.
Commercial RSV was treated by kneading and microwave irradiation, the same techniques used to prepare binary systems with cyclodextrins, and the thermic and enthalpic parameters of the solid phases obtained are reported in Table S1.
Compared with the untreated material, the absence of differences in enthalpic and thermal parameters confirmed that the same solid phase of RSV was retained. By means of the decomposition temperatures extrapolated by TGA measurements, it was possible to confirm the physical stability of RSV to the treatments used.

S5
The thermal data are supported by FT-IR analysis ( Figure S3). The shifts of some bands to higher wave numbers in the spectrum of the microwave product confirm the solid state interaction.

DSC and FT-IR data for DMB-RSV interaction
In the DSC profile of the physical mixture ( Figure S4, curve c) an endothermic effect appears at T peak = 207.4 ± 0,5 °C due to a solid state interaction between the drug and DMB by simple physical mixing; this interaction in the KN (curve d) and MW (curve not shown) products is evident as a complete amorphisation of the system.
The appearance of an endo-exothermic effect in the DSC curve of the system obtained by crystallization (curve e) is attributed to the inclusion of the drug in the CD cavity.

S7
The solid-state interaction between the drug and cyclodextrin is confirmed by FT-IR spectra of the treated physical mixtures. As an example, in Figure S5 the spectrum (d) of the KN product is reported, from which it is evident that the characteristic bands of RSV at 1605 and 1581 cm -1 that are still present in the spectrum (c) of the physical mixture no longer appear as a resolved doublet.

1 H NMR integration for complex host-guest stoichiometries
Following selection of single crystals of the inclusion complexes, the specimens were thoroughly surface-dried on filter paper, dissolved in acetone-d 6 (TMA·RSV, TMB·RSV) and DMSO-d 6 (DMB·RSV) and their 1 H NMR spectra were recorded on a 400MHz Bruker AMX 400 spectrometer. The data were analysed using the Bruker software Topspin, Mestronova or MestRe-C. Proton integration data confirming the 1:1 host-guest stoichiometries are tabulated below. The atomic labelling schemes are appended to the table. Interpretation of the data appears in the manuscript. Interpretation of the data appears in the manuscript. 8. X-ray crystal structure refinement details.

Complex TMA·RSV·6.25H 2 O
The asymmetric unit was found to contain two TMA molecules (A, B) and two RSV The remaining significant electron density peaks not associated with either a host or guest molecule were assigned as water molecules. A total of 12.5 water molecules were assigned. This is in good agreement with the TGA results, which revealed a

Modelling the trans-resveratrol guest
Electron density peaks found within the host molecule A were assigned as the nonhydrogen atoms of guest molecule A. No disorder was evident and the U iso values were found to be stable. The non-hydrogen atoms were subsequently refined anisotropically. The guest molecule included in CD host molecule B was found to be disordered over two positions. After some effort, the disordered components Isolated electron density peaks were assigned as water molecules. Atoms O4W and O5W were found to be disordered over two positions. The U iso values for O6W and O7W were abnormally high indicating partial occupancy. An average U iso was applied to these and their s.o.f.s were allowed to refine freely. Once the s.o.f.s had settled, these were fixed and the U iso values were allowed to refine freely. Electron density peaks corresponding to the expected positions of the hydrogen atoms of the water molecules were identified and assigned for atoms O1W, O2W, O3W and S14 O4W. The positions of the hydrogen atoms were fixed using the DANG and DFIX commands. A total of 5.6 water molecules were assigned over 9 sites (Table S4 lists

Modelling the trans-resveratrol guest
The   The key for these parameters is given as a footnote** in Table S5.

Experimental and simulated PXRD patterns for the inclusion complexes
The experimental PXRD patterns were collected at 21 C on a BRUKER D8 Advance X-ray diffractometer using CuK α -radiation (λ = 1.5406 Å) with generator settings 30 kV, 40 mA. Samples were placed on a zero-background sample holder and scanned over the range 4°-40° 2 with a step-size of 0.02. Simulated PXRD patterns were computed using the program Lazy Pulverix 1 with the refined single crystal X-ray structural data determined at 173(2) K as input.
In comparing the experimental and simulated patterns, allowance should be made for the distinctly different temperatures involved. This has the general effect that the 2-angles of peaks in the calculated patterns (lower temperature) generally tend to occur at higher values than those of the corresponding peaks in the experimental traces. Discrepancies in peak intensities are attributed to preferred orientation in the experimental samples.