First Class of Phosphorus Dendritic Compounds Containing β-Cyclodextrin Units in the Periphery Prepared by CuAAC

A new class of phosphorus dendritic compounds (PDCs) having a cyclotriphosphazene (P3N3) core and decorated with six β-cyclodextrin (βCD) units, named P3N3-[O-C6H4-O-(CH2)n-βCD]6, where n = 3 or 4 was designed, and the synthesis was performed using copper (I) catalyzed alkyne-azide cycloaddition (CuAAC). To obtain the complete substitution of the P3N3, two linkers consisting of an aromatic ring and an aliphatic chain of two different lengths were assessed. We found that, with both linkers, the total modification of the periphery was achieved. The two new obtained dendritic compounds presented a considerably high water solubility (>1 g/mL). The compounds comprised in this new class of PDCs are potential drug carrier candidates, since the conjugation of the βCD units to the P3N3 core through the primary face will not only serve as surface cover but, also, provide them the faculty to encapsulate various drugs inside the βCDs cavities.


Introduction
The use of nanomedicine confers a substantial potentiality for drug delivery and targeted release, as it increases the safety by reducing toxic effects in nontargeted organs and tissues [1][2][3]. In recent years, many researchers have focused on the development of dendrimers, since they have distinctive properties, such as monodispersity, a large number of easily available functional groups on the surface and an extraordinary capability to encapsulate host molecules within their hydrophobic environment, making them ideal nanocarriers for the targeted delivery (with or without ligand) of therapeutic and diagnostic agents [4][5][6].
Dendrimers are well-defined hyper-branched three-dimensional macromolecules. Structurally, dendrimers have a core from which branches (or arms) are derived and end with multiple peripheral groups that determine their macroscopic properties [7]. Dendrimers can be constructed starting from the core towards the periphery (divergent synthesis) or through a top-down approach, from the the hydroquinone and the respective X-alkynes (X = Cl for n = 3 and I for n = 4), using K 2 CO 3 as a base and anhydrous N,N-dimethylformamide (DMF) as the solvent at 72 • C. Mono and difunctionalized products were obtained, and they could be separated by column chromatography. Subsequently, the fraction corresponding to the monofunctionalized product was recrystallized in a cold/hot hexane to give the compounds A and B [27]. Scheme 1. The intermediates A and B were prepared via a Williamson etherification reaction between the hydroquinone and the respective X-alkynes (X = Cl for n = 3 and I for n = 4), using K2CO3 as a base and anhydrous N,N-dimethylformamide (DMF) as the solvent at 72 °C. Mono and difunctionalized products were obtained, and they could be separated by column chromatography. Subsequently, the fraction corresponding to the monofunctionalized product was recrystallized in a cold/hot hexane to give the compounds A and B [27].
The synthetic route of intermediates P3N3-[O-C6H4-O-(CH2)n-alkyne]6 (n = 3 or 4) C and D is shown in Scheme 1. The preparation of the intermediates was achieved, as previously reported in the literature but using phenols with a different substitution pattern [28]. The reaction proceeded through a substitution reaction of the Cl atoms attached to the P3N3 core; this substitution was carried out in basic and anhydrous medium using intermediates A and B to obtain compounds C and D, respectively. The next step in the synthetic route was the formation of mOTs-βCD (E), followed by mN3-βCD (F), which is shown in Scheme 1. For the synthesis of mOTs-βCD, the most efficient technique reported in the literature was employed to modify a single position of the native βCD [29], with minimal differences in the purification of the final product. It is worth noticing that this synthetic step was crucial in the design of our PDCs. In particular, the βCD tosylation method provided high yields compared to other previously reported methods [30,31]. Then, the tosyl group of compound E was replaced by an azide group through a nucleophilic substitution reaction to give compound F [32].
Once the necessary intermediates, alkynes (P3N3-[O-C6H4-O-[CH2]n-alkyne]6) and azide (mN3-βCD) were obtained, the CuAAC reaction was carried out, as shown in Scheme 1. According to the design of these novel PDCs, we used two linkers with different lengths of the aliphatic chain to assess whether this factor had an influence on the complete functionalization of all positions of the P3N3 core. To carry out the CuAAC reaction, we used a Cu(I) catalyst synthesized in situ, with CuSO4 as The synthetic route of intermediates P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -alkyne] 6 (n = 3 or 4) C and D is shown in Scheme 1. The preparation of the intermediates was achieved, as previously reported in the literature but using phenols with a different substitution pattern [28]. The reaction proceeded through a substitution reaction of the Cl atoms attached to the P 3 N 3 core; this substitution was carried out in basic and anhydrous medium using intermediates A and B to obtain compounds C and D, respectively. The next step in the synthetic route was the formation of mOTs-βCD (E), followed by mN 3 -βCD (F), which is shown in Scheme 1. For the synthesis of mOTs-βCD, the most efficient technique reported in the literature was employed to modify a single position of the native βCD [29], with minimal differences in the purification of the final product. It is worth noticing that this synthetic step was crucial in the design of our PDCs. In particular, the βCD tosylation method provided high yields compared to other previously reported methods [30,31]. Then, the tosyl group of compound E was replaced by an azide group through a nucleophilic substitution reaction to give compound F [32].
Once the necessary intermediates, alkynes (P 3 N 3 -[O-C 6 H 4 -O-[CH 2 ] n -alkyne] 6 ) and azide (mN 3 -βCD) were obtained, the CuAAC reaction was carried out, as shown in Scheme 1. According to the design of these novel PDCs, we used two linkers with different lengths of the aliphatic chain to assess whether this factor had an influence on the complete functionalization of all positions of the P 3 N 3 core. To carry out the CuAAC reaction, we used a Cu(I) catalyst synthesized in situ, with CuSO 4 as the copper source and H 2 Asc as a reducing agent, using dimethylsulfoxide (DMSO) as the ideal solvent, since it solubilizes the azide and the respective alkynes. To ensure the complete functionalization of the six core groups of P 3 N 3 , an excess of mN 3 -βCD was added. The final PDCs G and H were purified by size exclusion chromatography, using water as the eluent, and were obtained in high purity and with yields of >50%. We found that the length of the linker did not affect the complete functionalization Molecules 2020, 25, 4034 4 of 12 of P 3 N 3 with the βCD units, since all six positions of the P 3 N 3 core were functionalized with both proposed linkers.

Characterization
The full characterization of all the synthetic intermediates and final compounds was carried out. Regarding intermediates A and B, in the 1 H-NMR spectra (Figures S1 and S4 are available in Supplementary Materials (SM)) in DMSO-d 6 , it is possible to observe the signal corresponding to the proton of the alkyne group at 2.80 and 2.78 ppm, respectively. Moreover, the signal of the phenol proton appeared between 8.90 and 8.88 ppm. The structure of intermediates A and B was also confirmed by the 13  The presence of phosphorus in the core allowed simple monitoring by 31 P-NMR; furthermore, this allowed to verify the completion of the reactions in each synthesis step, as well as the integrity of the entire structure [33]. For the characterization of the intermediates C and D, a single signal corresponding to the complete functionalization was observed in the 31 P-NMR spectrum (Figures S9 and S13 are available in SM). Furthermore, the complete functionalization of the phosphorous core was confirmed with the rest of the characterization. In the 1 H-NMR spectrum (Figures S7 and S11 are available in SM), the disappearance of the signal corresponding to the phenol proton present in intermediates A and B, as mentioned above, is clearly observed due to its grafting on P 3 N 3 . The structure of C and D was also confirmed by 13  and Homonuclear Correlation Spectroscopy (COSY)) (Figures S15-S17 and S19-S22 are available in SM) was carried out in DMSO-d 6 . The classical signals of the linkers and native βCD units are in agreement with those reported in previous works [17,26]. Moreover, it was possible to observe the differentiation for some protons of the modified glucopyranose unit of the βCD (see Figure 1). The signals of protons H-1', H-5' and, particularly, H-6' appeared at a lower field than those of H-1, H-5 and H-6 of the nonfunctionalized subunits, due to the change in their chemical environment after the CuAAC reaction. In the same way, it was possible to identify the H-6" diastereotopic protons, corresponding to one -CH 2 -OH fragment contiguous to the substituted one at 3.12 and 2.94 ppm (see HMQC in Figure 2). Therefore, H-6" and the adjacent OH-6" on the primary face (around 4.33 ppm) appeared significantly upfield-shifted in comparison to their respective analogs H-6 and OH-6 because of the change in chemical environment due to the neighboring substitution. In this step, 31 P-NMR was used to assess the completion of the reactions and to assure the complete functionalization of the six positions of P 3 N 3 . This was confirmed by the appearance of a single signal (at 9.31 ppm) in the 31 P-NMR spectrum of P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -βCD] 6 (where n = 3, compound G). This signal exhibited an upfield shift compared to the signal corresponding to P 3 N 3 -[O-C 6 H4-O-(CH 2 ) n -alkyne] 6 (where n = 3, compound C) (at 9.93 ppm), because P 3 N 3 is protected by βCD molecules (Figure 3). The same behavior was observed in the 31  and H, respectively. Furthermore, the signals at 7034.596 m/z and 7120.029 m/z for compounds G and H, respectively, correspond to partial ionization, since complete ionization of the molecule is complex when using this technique. This difficulty has been previously reported for other phosphorus dendrimers [34].
Molecules 2020, 25, x FOR PEER REVIEW 5 of 12 same behavior was observed in the 31 P-NMR spectrum ( Figure S24 is available in SM) of P3N3-[O-C6H4-O-(CH2)n-βCD]6 (where n = 4, compound H). Finally, the structures of these new dendrimers were corroborated by MALDI-TOF mass spectrometry (Figures S18 and S23 are available in SM). The molecular ions appeared mainly at 8169.59 m/z and at 8254.30 m/z, corresponding to the molecular weights of the PDCs G and H, respectively. Furthermore, the signals at 7034.596 m/z and 7120.029 m/z for compounds G and H, respectively, correspond to partial ionization, since complete ionization of the molecule is complex when using this technique. This difficulty has been previously reported for other phosphorus dendrimers [34].

Determination of Water Solubility for P3N3-(O-C6H4-O-(CH2)n-βCD)6 PDCs
Among all the special properties of dendrimers, their high solubility makes an important difference compared to linear polymers. Differences in solubility between hyper-branched and linear polymers can be of several orders of magnitude, up to 10 6 times, in some cases [35]. Among all the solvents used to dissolve dendrimers, water is especially important when considering biological applications. In recent years, the importance of water-soluble phosphorous-containing dendrimers has increased. In most cases, the water solubility of these compounds was dependent mainly on the reactivity of terminal groups in the periphery [36]. Therefore, the determination of the water solubility of our new PDCs was carried out, according to a method previously reported in the literature [37]. It was found that, for both PDCs, the solubility in water was >1g/mL. These results represent a considerably higher solubility compared to the solubility of native βCD (18.5 mg/mL) and of other commercial βCD derivatives, such as sulfobutylether-βCD (>500 mg/mL), O-methyl-βCD (>500 mg/mL) and 2-hydroxypropyl-βCD (> 600 mg/mL) [38]. Water-soluble phosphorouscontaining dendrimers with cationic or anionic end groups have been reported; nevertheless, their water solubility was dependent on the pH [17,35]. On the contrary, the advantage of our new P3N3-[O-C6H4-O-(CH2)n-βCD]6 PDCs (n = 3 and 4) is that their solubility in water is conferred by the βCD units in standard conditions.

Determination of Water Solubility for P 3 N 3 -(O-C 6 H 4 -O-(CH 2 ) n -βCD) 6 PDCs
Among all the special properties of dendrimers, their high solubility makes an important difference compared to linear polymers. Differences in solubility between hyper-branched and linear polymers can be of several orders of magnitude, up to 10 6 times, in some cases [35]. Among all the solvents used to dissolve dendrimers, water is especially important when considering biological applications. In recent years, the importance of water-soluble phosphorous-containing dendrimers has increased. In most cases, the water solubility of these compounds was dependent mainly on the reactivity of terminal groups in the periphery [36]. Therefore, the determination of the water solubility of our new PDCs was carried out, according to a method previously reported in the literature [37]. It was found that, for both PDCs, the solubility in water was >1g/mL. These results represent a considerably higher solubility compared to the solubility of native βCD (18.5 mg/mL) and of other commercial βCD derivatives, such as sulfobutylether-βCD (>500 mg/mL), O-methyl-βCD (>500 mg/mL) and 2-hydroxypropyl-βCD (>600 mg/mL) [38]. Water-soluble phosphorous-containing dendrimers with cationic or anionic end groups have been reported; nevertheless, their water solubility was dependent on the pH [17,35]. On the contrary, the advantage of our new P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -βCD] 6 PDCs (n = 3 and 4) is that their solubility in water is conferred by the βCD units in standard conditions.

General Notes
All starting materials were commercially available reagent grade and were used without any further purification. Hexachlorocyclotriphosphazene (P 3 N 3 Cl 6 ), 5-chloro-1-pentyne, 6-iodo-1-hexyne, hydroquinone, β-cyclodextrin (βCD), p-toluenesulfonyl chloride (Cl-Ts), p-toluenesulfonic acid (OH-Ts), N,N-dimethylformamide anhydrous (DMF), potassium carbonate (K 2 CO 3 ), cesium carbonate (Cs 2 CO 3 ), dimethylsulfoxide (DMSO), ascorbic acid (H 2 Asc), Bio-Gel P-10 medium from Bio-Rad (Hercules,  = 3 and 4) A and B The synthesis of the intermediates A and B was carried out according to a previously reported procedure, with some modifications [27,28]. A dry hydroquinone solution (24.03 mmol) in DMF (250 mL) was refluxed for 30 min at 72 • C, K 2 CO 3 (30.04 mmol) was added and the mixture was refluxed for 1 h. To this mixture, alkyne (12.02 mmol) was added dropwise over 2 h. The resulting mixture was refluxed for 36 h, then cooled to 25 • C and filtered. The filtrate was evaporated under reduced pressure. The resulting brown oil was dissolved in CH 2 Cl 2 (150 mL), and the solution was extracted with water (3 × 50 mL), the organic phase was dried with anhydrous Na 2 SO 4 and the solvent was evaporated. The crude product was composed of a mixture of unreacted hydroquinone, the monofunctionalized and the difunctionalized products, which were separated by column chromatography on silica gel using hexanes:EtOAc (8:2). Once the fraction corresponding to the monofunctionalized product was obtained, it was recrystallized from hot/cold hexane, and the final product was recovered by filtration; the solid was left to dry overnight under vacuum. HO-C 6 H 4 -O-(CH 2 ) 3 -alkyne was obtained as a beige solid (7.17 mmol, 29%). HO-C 6 H 4 -O-(CH 2 ) 4 -alkyne was obtained as a yellow solid (7.88 mmol, 66%).

Synthesis of P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -alkyne] 6 (n = 3 and 4) C and D
The synthesis of intermediates C and D was carried out according to a previously reported procedure, with some modifications [39]. A solution of intermediate A or B (2.84 mmol) in dry THF (25 mL) was stirred for 20 min. Cs 2 CO 3 (5.68 mmol) was added to this solution and stirred for 1 h. Afterwards, P 3 N 3 Cl 6 (0.32 mmol) was added to the reaction mixture, and it was left stirring at room temperature for 7 days, until the reaction was complete, monitored by 31 P-NMR. The reaction mixture was then centrifuged at 10,000 rpm for 20 min to remove inorganic salts, the supernatant was recovered and the solvent was evaporated under reduced pressure. The crude product contained fully functionalized P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -alkyne] 6 and the unreacted monoalkynes, which were separated by column chromatography on silica gel using CH 2 Cl 2 as the eluent. Once the fraction corresponding to the functionalized P 3 N 3 was obtained, it was recrystallized from cold isopropyl ether. P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -alkyne] 6 (n = 3 or 4) were obtained as white solids (0.19 mmol, 59% and 0.12 mmol, 52% for C and D, respectively).

Synthesis of 6-O-monotosyl-β-cyclodextrin (mOTs-βCD) E
Following the procedure previously reported [29], in a round-bottom flask, p-toluenesulfonic acid (0.0101 mol) and p-toluenesulfonyl chloride (0.039 mol) were dissolved in 50 mL of CH 2 Cl 2 and stirred at room temperature for 12 h. Then, the reaction mixture was filtered, and the filtrate was concentrated under reduced pressure. The residue was recrystallized (3×) with cold hexane, and the product (Ts 2 O) was allowed to dry under vacuum. Afterwards, in a round-bottom flask, βCD (0.0051 mol) and Ts 2 O (0.0076 mol) were dissolved in 150 mL of H 2 O. The reaction mixture was stirred for 2 h at room temperature; after this time, a 2.5-M aqueous NaOH solution was added. The reaction mixture was stirred for 10 min. Subsequently, the mixture was filtered, and the pH of the filtrate was adjusted to 8 with a saturated solution of ammonium chloride to give a precipitate. The mixture was filtered, and the precipitate was recrystallized (3×) in acetone. The product E was obtained as a white solid (0.0021 mol, 42%).

Synthesis of 6-O-monoazido-β-cyclodextrin (mN 3 -βCD) F
Following a previously reported procedure [32], in a round-bottom flask, mOTs-βCD (0.0016 mol), NaN 3 (0.005 mol) and KI (0.0008 mol) were dissolved in 8 mL of anhydrous DMF. The reaction mixture was stirred at 80 • C for 48 h. After this time, DMF was evaporated under reduced pressure, and the residue was recrystallized in a mixture, H 2 O:Acetone (1:1), and allowed to dry under vacuum. The product F was obtained as a white solid (0.0014 mol, 88%). The obtained solid was purified by size exclusion chromatography using Bio-Gel ® P-10 medium and water as the eluent. Once the corresponding fractions were obtained, they were lyophilized to evaporate the water. PDCs G and H were obtained as white solids (0.041 mmol, 61% and 0.048 mmol, 88% for G and H, respectively).

Determination of Water Solubility for P 3 N 3 -[O-C 6 H 4 -O-(CH 2 )n-βCD] 6 PDCs
The determination of the solubility of the P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -βCD] 6 PDCs was carried out according to a method previously reported by Jozwiakowski and Connors [36], with modifications. One gram of the compound was placed in three amber vials of 5 mL with a screw cap, and 1 mL of water was added. The vials were sealed with parafilm to avoid water evaporation. They were stirred in an oil bath at a constant temperature of 25 ± 0.01 • C for 48 h. The supernatant was separated from the solid phase by filtration through a Milli-Q membrane (pore size of 0.45 µm) by injection of the mixture into disposable plastic syringes (Franklin Lakes, NJ, USA) of 3 mL at 25 • C. The supernatant of each sample was placed in three different vials. The samples were lyophilized for 48 h, and the obtained solids were weighed on a scale with an uncertainty of ±0.0001 g.

Conclusions
In this work, two novel phosphorus dendritic compounds (PDCs) containing a P 3 N 3 as the core and βCD units as terminal groups were designed. The synthesis was carried out using the CuAAC reaction, giving high yields and products that were purified by a simple method. The complete functionalization of the P 3 N 3 core of the new molecules was carried out using two aliphatic chains of different lengths as the linker. We found that there was no significant impact of the aliphatic chain length, since, with both linkers, the functionalization was complete. This is the first report for PDCs of the type P 3 N 3 -[O-C 6 H 4 -O-(CH 2 ) n -βCD] 6 arising from a novel design of water-soluble PDCs. The potential use of these new dendrimers as drug carriers for biomedical applications is currently under study.