Synthesis and supramolecular properties of regioisomers of mononaphthylallyl derivatives of γ-cyclodextrin

Monosubstituted derivatives of γ-cyclodextrin (γ-CD) are suitable building blocks for supramolecular polymers, and can also serve as precursors for the synthesis of other regioselectively monosubstituted γ-CD derivatives. We prepared a set of monosubstituted 2I-O-, 3I-O-, and 6I-O-(3-(naphthalen-2-yl)prop-2-en-1-yl) derivatives of γ-CD using two different methods. A key step of the first synthetic procedure is a cross-metathesis between previously described regioisomers of mono-O-allyl derivatives of γ-CD and 2-vinylnaphthalene which gives yields of about 16–25% (2–5% starting from γ-CD). To increase the overall yields, we have developed another method, based on a direct alkylation of γ-CD with 3-(naphthalen-2-yl)allyl chloride as the alkylating reagent. Highly regioselective reaction conditions, which differ for each regioisomer in a used base, gave the monosubstituted isomers in yields between 12–19%. Supramolecular properties of these derivatives were studied by DLS, ITC, NMR, and Cryo-TEM.


Experimental part
General information -CD was purchased from Wako Chemicals (Germany), special chemicals and solvents from Sigma-Aldrich and common solvents from Penta or Lach-ner (Czech Republic). -CD was dried at 70 °C under reduced pressure (3 mbar). Solvents were distilled before use. DMSO was distilled and dried with 3 Å molecular sieves. DMF was predried with P 2 O 5 , then distilled and dried with 3 Å molecular sieves. Redistilled water was used.
EtONa was prepared by dissolving of Na (1.8 g) in absolute EtOH (30 ml). Residual solvent was then evaporated to yield solid EtONa. NaH was used as a 60% suspension in mineral oil.
LDA was used as a 1.5 M solution in cyclohexane, in complex with THF in ratio 1:1.
Cation exchanger DOWEX 50 was activated and converted to its H + form using standard methods and finally washed with 50% aqueous MeOH. Silica gel 60 (0.040-0.063 mm, Merck, Germany) was used for chromatography. TLC was performed on silica gel 60 F 254coated aluminum sheets (Merck, Germany) or reverse silica gel 60 RP-18 F 254 S (Merck, Germany). Spots were detected by a UV lamp (= 254 nm) or by spraying with 50% aqueous H 2 SO 4 followed by carbonization with a heat-gun. NMR spectra were recorded on a Bruker Avance III (600 MHz) ( 1 H at 600. 17 MHz, 13 C NMR at 150.04 MHz) and on a Varian VNMRS 300 ( 1 H at 299.94 MHz) in deuterated solvents and are referenced to the residual solvent peak. Chemical shifts are given in δ-scale, coupling constants J are given in Hz. The numbering of atoms for NMR spectra transcription was done analogous to Figure S1: Numbering of atomsFigure S1. The glucose unit bearing an alkyl substituent is numbered with " I " sign where the assignment is unambiguous. All other glucose signals are numbered indiscriminately. The alkyl substituent is numbered with "´" sign. Label 4' is used for all aromatic hydrogens of the naphthyl substituent. The assignment of the 1 H and 13
The reaction mixture was quenched by adding Me 2 S (2 ml) and the solvents were evaporated.
Purification of the crude product using column chromatography on silica gel (CHCl 3 /MeOH, 70/1) yielded 590 mg (73%) of 3 as a white solid. , S11 where  is an instrumental factor. For the diffusion of nanoparticles in liquid, the hydrodynamic radius R H or hydrodynamic diameter D H can be determined using the Stokes-Einstein equation 1 2 ) ( where k is the Boltzmann constant, n the refractive index and  the viscosity of the solvent. Most of the sample was removed by blotting (Whatman no. 1 filter paper) for approximately 1 s, and the grid was immediately plunged into liquid ethane held at -183 °C. The grid was then transferred without rewarming into the microscope. Images were recorded at an accelerating S12 voltage of 120 kV and with magnifications ranging from 11500× to 50000× using a Gatan UltraScan 1000 slow scan CCD camera in the low-dose imaging mode, with the electron dose not exceeding 1500 electrons per nm 2 . The magnifications resulted in final pixel size ranging from 1 to 0.2 nm, the typical value of applied underfocus ranged between 0.5 to 2.5 μm. The applied blotting conditions resulted in the specimen thickness varying between 100 and ca. 300 nm. Specimens were sonicated for 10 minutes with a Diogenode Bioruptor sonicator (Tosho Denki Co. Ltd., Japan).

Image processing
Acquired cryo-TEM images with pixel sizes of 0.3 nm and smaller were corrected for the contrast reversals caused by oscillations of the TEM's contrast transfer fuction by the GCTF programme [5] and band-pass filtered with ImageJ [6] in order to suppress density variations caused by changes of ice thickness and noise beyond the cutoff of 1 nm. Images with coarser pixel sizes were only band-pass filtered with the same settings. All distances were measured with ImageJ.

Measurements
Isothermal titration calorimetry was carried out as dilution experiments at 10, 25 and 45 C, using the ITC-200 instrument (MicroCal, Inc., Northampton, MA). In each titration, a series 0.5μL + 19  2 μL of 10 mM sample was injected into the measuring cell, initially filled with pure water, and the released heat was recorded for each injection. However, the dimerdissociation procedure of MicroCal Origin for ITC supplied with the instrument could not be used for the extraction of thermodynamic parameters from the obtained data because NA-γ-CDs have to be treated as monomers of AB type, capable of polymerization. Moreover, there S13 has been some controversy about implementation of the MicroCalc procedure [7]. Therefore the data were treated as described in the following section.

Data treatment
The measuring cell of the isothermal calorimeter has a fixed active volume V 0 fully filled with solution in which component X is at concentration c X,i-1 . Injection of solution from the syringe (concentration of X c X,S ) drives an equal volume ΔV i of initial solution from the active volume to the cell stem. After that, a new concentration c X,i is established due to mixing [8].
c X Is a total concentration of X, that is, the sum of concentrations of free and bound X.
Isodesmic supramolecular polymerization [9] of A-B monomer [10] can be used as a

γ-CD solution behavior
To obtain reference results, we also investigated the behavior of different concentration solutions of -CD in water (100, 50, 10, 1 mM). These samples prepared by serial dilution formed 1000 nm, 700 nm, 350 nm and 350 + 100 nm aggregates, respectively. Interestingly, the 10 mM sample of -CD prepared by direct dissolving of a solid compound contained just 1-2 nm particles according to its distribution function. The experiments were performed repeatedly, and the results are reproducible. Another series of -CD samples was prepared by S26 the dilution row again, but after that, the samples were stirred in a closed vial for 10 min at 120 °C, which led to formation of aggregates with sizes of 800 nm, 350 nm, 1.6, and 1.5 nm at 100, 50, 10, and 1 mM concentrations, respectively. Therefore, -CD aggregates survive in a large range of dilutions of the solution but are destroyed by boiling. However, Coleman [11] and Puskás [12] reported a different behavior of -CD solution. Puskás investigated aggregation properties of native CDs in water. While -CD and -CD showed similar results, forming primarily about 2 nm particles, -CD (1.0 wt% ~ 7.8 mM) formed aggregates with a broad size distribution from 50 to 600 nm. Coleman observed -CD (25 g/L ~19 mM) to form 200 nm aggregates. Occasionally, we also observed aggregates in 10 mM -CD prepared by the direct dissolving of a solid sample. However, in most cases, no aggregates of -CD were observed using this procedure. On the other hand, Messner et al. [13] summarized in their review that in ca. 10 mM aqueous solutions of native CDs the mass fraction of aggregated CD molecules is less than 1% (12 mM α-CD -0.8% [14], 10 mM β-CD -0.0011% [15][16], 12 mM γ-CD -0.02% [17]). This means that the behavior of a solution of native -CD depends significantly on the method of its preparation. We can conclude that many factors are influencing the formation of aggregates (temperature, filtration, dilution, etc.) which limit the reproducibility of the measurements.