Intercalation of Benzoxaborolate Anions in Layered Double

Hydroxides: Toward Hybrid Formulations for Benzoxaborole Drugs ... Extensive characterization was carried out on the materials, notably by powder X-ra...

0 downloads 165 Views 2MB Size
Article pubs.acs.org/cm

Intercalation of Benzoxaborolate Anions in Layered Double Hydroxides: Toward Hybrid Formulations for Benzoxaborole Drugs Saad Sene,† Sylvie Bégu,† Christel Gervais,‡ Guillaume Renaudin,§ Adel Mesbah,∥ Mark E. Smith,⊥,# P. Hubert Mutin,† Arie van der Lee,▽ Jean-Marie Nedelec,§ Christian Bonhomme,‡ and Danielle Laurencin*,† †

Institut Charles Gerhardt de Montpellier, UMR 5253, CNRS-UM2-UM1-ENSCM, Place E. Bataillon, CC1701, 34095 Montpellier cedex 05, France ‡ Sorbonne Universités, UPMC Université Paris 06, CNRS, Collège de France, UMR 7574, 11 place Marcelin Berthelot, Laboratoire de Chimie de la Matière Condensée de Paris, 75005, Paris, France § Clermont Université, ENSCCF, UMR 6296, Institut de Chimie de Clermont-Ferrand, BP 10448, 63000 Clermont-Ferrand, France ∥ Institut de Chimie Séparative de Marcoule, UMR 5257 CNRS/CEA/UM2/ENSCM, bat 426, BP 17171, 30207 Bagnols-sur-Ceze, France ⊥ Vice-Chancellor’s Office, University House Lancaster University, Lancaster, LA1 4YW, United Kingdom # Magnetic Resonance Centre, Department of Physics, University of Warwick, Coventry, CV4 7HS, United Kingdom ▽ Institut Européen des Membranes, CNRS-UMR 5635, Université de Montpellier 2, CC 047, Pl. E. Bataillon, 34095 Montpellier Cedex 5, France S Supporting Information *

ABSTRACT: Benzoxaboroles are a family of cyclic derivatives of boronic acids, whose reactivity makes them interesting candidates for the development of novel drugs. In this study, we describe the preparation of the first hybrid organic− inorganic materials involving benzoxaboroles, as a first step toward their use as new formulations for such drugs. The materials were prepared by intercalation of the simplest benzoxaborole (C7H6BO(OH), BBzx) or of its fluorinated analogue (C7H5FBO(OH), AN2690, a recently developed antifungal drug), both taken under their anionic form (benzoxaborolate), into a biodegradable inorganic matrix (a Mg−Al layered double hydroxide). Extensive characterization was carried out on the materials, notably by powder X-ray diffraction and multinuclear (11B, 27Al, 13C, 19F, 25Mg, and 1H) solid state NMR, in order to describe their structure, particularly in the vicinity of the organoboron species. Three crystalline phases involving benzoxaborolate anions in association with Ca2+ or Mg2+ cations were also prepared as part of this work (Mg(C7H6BO(OH)2)2·10H2O, Mg(C7H6BO(OH)2)2·7H2O and Ca3(C7H6BO(OH)2)5(C7H6BO2)·3H2O), in order to assist in the interpretation of the spectroscopic data. A DFT computational model of the interlayer space was proposed, which is consistent with the experimental observations. Several properties of the materials were then determined with a view of using them as part of novel formulations, namely the maximum loading capacity toward benzoxaborol(at)es, the optimal storage conditions, and the release kinetics in simulated physiological media. All in all, this study serves as a benchmark not only for the development of novel formulations for benzoxaborole drugs, but more generally for the preparation of a novel class of organic− inorganic materials involving benzoxaborol(at)es. functions exposed at the surface of cells4,6 and favoring the intracellular delivery of drugs or enzymes.6 Various nanoparticles exposing benzoxaborole functions at their periphery have also been prepared,5,7 allowing for example the rapid and selective enrichment of glycosylated proteins from complex

1. INTRODUCTION Benzoxaboroles are an emerging family of molecules which are finding an increasing number of applications, in molecular chemistry, biotechnology, medicine, and materials science.1,2 In particular, their ability to bind reversibly to sugars and diols at physiological pH,3 and with a higher affinity than the corresponding boronic acids, has been widely exploited. On the basis of this reactivity, benzoxaborole-based carbohydrate sensors have been designed,4,5 notably for targeting sugar © 2015 American Chemical Society

Received: November 13, 2014 Revised: January 18, 2015 Published: January 19, 2015 1242

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials biological media.7 Benzoxaborole-functionalized monolithic columns have also been elaborated for the separation of cisdiol containing biomolecules,8 showing an excellent selectivity and affinity toward nucleosides and glycoproteins. Furthermore, the association between polysaccharides and macromolecules bearing benzoxaborole groups has been looked into,9 leading for example to the preparation of temperature, pH, and glucose responsive gels.9a One of the most promising fields of applications of benzoxaboroles is in medicine, where they are mainly being studied as enzyme inhibitors.10,11 By varying the nature of the substituents on the heterocycle and aromatic rings, active benzoxaborole molecules have been developed for the treatment of onychomycosis,1,11,12 bacterial infections10a,11 or human African trypanosomiasis (sleeping sickness),10b,11 to name a few. The first benzoxaborole-based drug to receive FDA-approval is tavaborole (also referred to as AN2690, Figure 1B). It was developed by Anacor pharmaceuticals, approved in

applications such as catalyst supports,16 adsorbents for water purification17 or fillers in polymers.15b,18 Moreover, some LDH compositions are attractive in medicine,18−22 due to their biocompatibility, low cytotoxicity, high capacity of anionic loading, and also their possibility of surface functionalization. Mg/Al-LDH is for example already prescribed as an antiacid (Magaldrate, Riopan , 1998, Takeda). Furthermore, previous studies have also shown that LDHs can be used as inorganic carriers for drugs like ibuprofen,20 L-Dopa,21 pravastatin,22 or fluorouracil,23 taken under their anionic form. Intercalating benzoxaborolate anions in LDHs could thus be useful for the development of novel ways of formulating benzoxaborole-type drugs. In this article, we describe the first hybrid materials involving benzoxaborole drugs, by intercalation of benzoxaborolates into a Mg/Al-LDH. Two benzoxaborolates were considered: (i) anions deriving from the simplest benzoxaborole (i.e., the “privileged scaffold” C7H6BO(OH), referred to as “BBzx” below, Figure 1A), in order to have a general picture of the reactivity of this family of molecules with respect to LDH; and (ii) anions deriving from the fluorinated analog AN2690 (i.e., an active drug, Figure 1B),11,12 because of the fluorine atom which may help provide further insight into the structure of these materials using 19F NMR. Here, the synthesis of the hybrid materials will first be described, together with their characterization using a wide range of analytical techniques, including electron microscopy, X-ray diffraction and multinuclear (1H, 27Al, 11B, 13C, 19F, and 25Mg) solid-state NMR. Particular attention was paid to the benzoxaborolate local environment in the LDH, and it will be shown how it was elucidated using experimental data from three new crystalline phases (Mg(C7H6BO(OH)2)2·10H2O, Mg(C7H6BO(OH)2)2· 7H2O and Ca3(C7H6BO(OH)2)5(C7H6BO2)·3H2O), and density functional theory (DFT) modeling of the interlayer space. Finally, we will discuss the maximum loading capacity and best storage conditions of the intercalated LDH, as well as the release profiles which were evaluated in view of determining how to use these hybrid materials for the delivery of benzoxaborolate drugs.

Figure 1. Structures of benzoxaboroles: (A) the privileged scaffold BBzx and (B) the antifungal drug AN2690.

July 2014, and is now being commercialized under the name of Keridyn for the treatment of mild to moderate onychomycosis. From a more general perspective, the benzoxaborole motif can now be considered as a new “privileged scaffold”,13 because its intrinsic reactivity (reversible binding to cis-diols, Lewis acidity of the boron) and its metabolic stability make it an interesting starting point for the design of novel drugs. Despite the numerous studies on benzoxaboroles, no attempt to associate these molecules to inorganic materials directly through the benzoxaborole function has been reported to date. Given the tremendous developments around hybrid organic− inorganic materials over the past 30 years,14 the elaboration of such hybrid phases could open up a whole new range of perspectives in terms of materials design around these molecules. For instance, the incorporation of benzoxaborole drugs into bioresorbable inorganic carrier matrices could be used as an alternative way of formulating these molecules, which would give access to different release kinetics compared to currently available formulations, and thereby open the way to novel possibilities in terms of treatments for patients. Here, as a very first step toward the elaboration of hybrid organic− inorganic materials involving benzoxaboroles, we decided to investigate their association with Layered Double Hydroxides (LDHs) in order to learn more about the behavior of formulated benzoxaboroles, and to evaluate the properties of these materials in view of the development of novel drugdelivery systems. LDHs are compounds of general formula {M2+1−xM3+x (OH)2}{(An−)x/n·mH2O}, where x ranges between 0.20 and 0.33.15 Their structure derives from the mineral brucite (Mg(OH)2) and is composed of positively charged layers of metal hydroxides (M2+ = Mg2+, Zn2+, Ca2+, etc. ; M3+ = Al3+, Fe3+, Mn3+, etc.), with interlayer spaces occupied by exchangeable An− anions (An− = NO3−, Cl−, etc.) and water. Due to their lamellar structure and tunable composition, LDHs show interesting properties and have been studied for many

2. EXPERIMENTAL SECTION 2.1. Materials. Magnesium nitrate (Mg(NO3)2·6H2O, SigmaAldrich, ≥ 99%), aluminum nitrate (Al(NO3)3·9H2O, Sigma-Aldrich, ≥ 98%), calcium chloride (CaCl2·2H2O, Sigma-Aldrich, ≥ 99%), sodium hydroxide (NaOH, Acros Organics) and 2-hydroxymethylphenylboronic acid cyclic monoester (C7H6BO(OH), benzoxaborole, abbreviated as BBzx, Sigma-Aldrich, 97%) were used as received. All reactions were carried out using ultrapure water and absolute ethanol. AN2690 (5-fluorobenzoxaborole) was synthesized according to the procedure reported previously.24 The purity of each batch of BBzx or AN2690 was checked by 11B solution NMR before use (to verify the absence of boric acid), and estimated to be better than 99%. The Mg/ Al-NO3 LDH starting material was synthesized based on a previously established procedure25 and characterized as described in the Supporting Information (SI, p S2). 2.2. Syntheses. 2.2.1. Preparation of LDH-BBzx (Intercalation of BBzx in LDH). BBzx (100.5 mg, 0.75 mmol) and NaOH microbeads (30.0 mg, 0.75 mmol) were dissolved in 20 mL of H2O under magnetic stirring. The solution was degassed by sonication for 5 min. An inert atmosphere was then set up by bubbling N2 in and over the solution under magnetic stirring. Mg/Al-NO3 LDH (200.0 mg, 2.29 mmol) was then added to the solution which was stirred for 5 min. The reaction container was then sealed and the suspension was magnetically stirred at 50 °C for 16 h. It was then centrifuged (10 min at 20 000 rpm). The supernatant was removed and the solid was 1243

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials washed with 10 mL of distilled water with manual shaking and centrifuged again (10 min at 20 000 rpm). After two steps of washing and centrifugation, the solid was finally dried under vacuum at room temperature for 24 h; this compound is referred to as LDH-BBzx. Yield: 265.8 mg. LDH-AN2690 was prepared using the same conditions, but replacing BBzx by AN2690. LDH-0 corresponds to a “control” sample prepared using similar conditions, but in absence of any organoboron molecule in solution. 2.2.2. Preparation of Crystalline Metal-Benzoxaborolate Phases (Also Referred to as “Model Materials” Herein). Mg(C7H6BO(OH)2)2·10H2O (MgBBzx·10H2O), Mg(C7H6BO(OH)2)2·7H2O (MgBBzx·7H 2 O) and Ca 3 (C 7 H 6 BO(OH) 2 ) 5 (C 7 H 6 BO 2 )·3H 2 O (CaBBzx·3H2O) were prepared by precipitating an alkaline solution of BBzx with Ca2+ or Mg2+. Full details on the reaction conditions are provided in the SI (p S3). 2.3. Characterization. Elemental analyses were carried out by the Service Central d’Analyse of the CNRS (Villeurbanne, France). Routine X-ray diffraction powder patterns were recorded using a PANalytical X’pert MPD-Pro diffractometer at the wavelength of Cu Kα1 (λ = 1.5405 Å) (45 kV and 20 mA) in Bragg−Brentano scanning mode. The program scanned angles (2θ) from 2° (or 4°) to 70° with a 0.01° step, and a step time of 40 s. For MgBBzx·7H2O, high resolution XRD powder patterns were collected at different temperatures at the powder station of the X04SA-MS beamline of the SLS synchrotron facility (Paul Scherrer Institute, Villigen, Switzerland), from a sample loaded in a 0.3 mm Hilgenberg glass capillary (Glas Nr 50), using Debye−Scherrer geometry, 12.4 keV radiation (λ = 0.999942 Å), and a Mythen detector (PSI Detector Group) covering 120° in 2θ with 0.0038° per step. For CaBBzx·3H2O and MgBBzx·10H2O, single crystals suitable for single-crystal X-ray diffraction analysis were obtained; their X-ray data were collected using an Agilent Technologies Xcalibur-I single-crystal diffractometer using graphite monochromated Mo-radiation (λ = 0.71073 Å). Further details on the XRD measurements and structure resolution of the three model materials can be found in the SI (p S5−S12). The .cif files of all crystal structures can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/ cif, with the following numbers: CCDC 1033689 for CaBBzx·3H2O, CCDC 1033690 to 1033692 for the structures of MgBBzx·7H2O at 125, 175, and 290 K, and CCDC 1033693 for MgBBzx·10H2O. SEM measurements were conducted on a Hitachi S4800 instrument under an excitation voltage between 0.5 and 8.0 kV depending on each powder’s surface charging. Powdered samples were simply deposited on double sided tape and then Pt-metalized by sputtering under vacuum. Samples for transmission electron microscopy (TEM) measurements were prepared by inclusion in Epon 812 resin. After 3 days of polymerization at 60 °C, layers of 70 nm thickness were cut using an ultramicrotome (Leica ultracut UCT). Slices were placed on a Cu grid and TEM measurements were carried out at 100 kV using a JEOL 1200 EXII microscope. Multinuclear (1H, 27Al, 11B, 13C, 19F, 25Mg, and 31P) solid-state NMR experiments were carried out on the samples at magnetic fields ranging from 9.4 to 20.0 T, using MAS (Magic Angle Spinning) or static conditions. Full details on the pulse sequences, acquisition parameters, and referencing can be found in the SI (p S13−S15). All solid state NMR experiments were performed under temperature regulation, in order to ensure that the temperature inside the rotor is ∼20 °C. 2.4. Computational Modeling and GIPAW DFT Calculations. All structural modeling calculations were performed using the ab initio plane-wave pseudopotential approach as implemented in the VASP code.26 The Perdew−Burke−Ernzerhof (PBE) functional27 was chosen to perform the periodic DFT calculations. The valence electrons were treated explicitly and their interactions with the ionic cores are described by the Projector Augmented-Wave method (PAW),28 which allows a low energy cutoff equal to 400 eV for the plane-wave basis to be used. The integration of the Brillouin zone was performed on the Γ-point, in all calculations.

For CaBBzx·3H2O and MgBBzx·10H2O model compounds, geometry optimization was carried out, starting from the experimental XRD structure and allowing the positions of protons to relax until the total energy differences between the loops is less than 10−4 eV. For the LDH-BBzx hybrid systems, models were prepared as follows. A super cell composed of 6 unit cells of Mg−Al LDH intercalated with carbonates (ICSD-86655) was used as a starting point. Carbonates were removed, as well as one of the metal-hydroxide sheets, to obtain a spacing between the layers adjusted at ∼16.2 Å (i.e., consistent with the spacing in the intercalated phase). In each layer, aluminum and magnesium atoms were positioned so that two AlO6 do not share an edge (i.e., are separated by MgO6)29 and respecting an Al to Mg ratio of 1:2. Different configurations were then built introducing benzoxaboroles, benzoxaborolates, nitrates and water molecules in the interlayer space. Atomic positions were first relaxed using ab initio molecular dynamics within the Born−Oppenheimer approximation at constant temperature (300 K), keeping only aluminum and magnesium atoms at fixed positions to maintain the spacing between the layers to the experimental value. The time step was set at 2.5 fs, and the geometries were sampled up to a few ps to have a reliable image of the equilibrium geometry at 300 K using a microcanonical ensemble. Tritium mass was set for all protons to avoid fluctuations due to the large time step chosen. Second, energetically favorable geometries were picked from the MD results and optimized at 0 K with a 300 eV energy cutoff. The first-principles NMR calculations were performed within Kohn−Sham DFT using the QUANTUM-ESPRESSO software.30 The PBE generalized gradient approximation27 was used and the valence electrons were described by norm conserving pseudopotentials31 in the Kleinman−Bylander form.32 The wave functions were expanded on a plane wave basis set with a kinetic energy cutoff of 1088 eV. The shielding tensor is computed using the GIPAW33 approach which permits the reproduction of the results of a fully converged allelectron calculation. The isotropic chemical shift δiso is defined as δiso = −[σ − σref] where σ is the isotropic shielding and σref is the isotropic shielding of the same nucleus in a reference system as previously described.24,34 The principal components Vxx, Vyy, and Vzz of the electric field gradient (EFG) tensor defined with |Vzz| ≥ |Vxx| ≥ |Vyy| are obtained by diagonalization of the tensor. The quadrupolar interaction can then be characterized by the quadrupolar coupling constant CQ and the asymmetry parameter ηQ, which are defined as CQ = eQVzz/h and ηQ = (Vyy − Vxx)/Vzz. 2.5. Release Assays. Release assays in dynamic conditions were performed using a standardized USP 4 flow cell designed for powdered samples35 and drug releases were monitored by UV spectroscopy with an Avaspec spectrometer (Avantes) and a micro flow Z-cell-10 (Avantes) for in situ measurements (see SI, p S16). Standard curves with calibrated solutions were first established by dissolving BBzx or AN2690 in a phosphate buffer saline (PBS, composition: NaCl ∼ 137 mM, Na2HPO4 ∼ 15 mM and KH2PO4 ∼ 1.4 mM) at pH 7.4; calibration curves were done in triplicate, with three sets of solutions. The solutions were thermostated at 37 °C, flowed through the USP 4 flow cell at a rate of 8.5 mL·min−1, and the intensity at the maximum absorption wavelength was then measured at 268 and 265 nm for BBzx and AN2690, respectively. For the release assays, 25 mg of LDH-BBzx or 12.5 mg of LDH-AN2690 were placed in the USP 4 flow cell and 25 mL of PBS were flowed through the flow cell at 37 °C. The Z-flow cell was inserted in the closed-circuit, connected through optical fibers to the spectrometer, allowing in situ measurements of the absorption over time. The weights of released molecules for each time point were calculated using the calibration curves and normalized to the drug content determined by elemental analysis. For each intercalated phase, the release assays were performed in triplicate. Mean values and standard deviations are presented in the graphs. Release assays in static conditions were monitored by UV spectroscopy using a UV/Visible PerkinElmer precisely Lambda 35. Standard curves with calibrated solutions (n = 3) were first established by dissolving BBzx or AN2690 in PBS, their maximum absorption wavelength being measured at 268 and 265 nm, respectively. Approximately 7 mg of the intercalated phase were placed in an 1244

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

Figure 2. Characterization of LDH-BBzx and LDH-0 by A/ SEM; B/ XRD; C/ 27Al MAS NMR, and D/ 11B MAS NMR.

Figure 3. A/ Coordination modes of the benzoxaborolates in the model materials (only the local environment of each inequivalent cation is represented here; H atoms bound to the carbon were omitted for clarity; for more complete representations of these structures, see SI p S17). B/ 11B MAS NMR spectra of CaBBzx·3H2O, MgBBzx·7H2O, and BBzx. Eppendorf and 1 mL of PBS was added before incubation at 37 °C. At different time intervals, the suspension was centrifuged, the supernatants were collected, and 1 mL of fresh PBS was added again before incubation. The quantity of BBzx or AN2690 released for each time point was determined by measuring the absorption of the supernatants by UV spectroscopy, a baseline substraction being applied between the

sample and a reference sample not containing any drug (LDH-0). The weights of released molecules for each time point were then calculated using the calibration curves, and were normalized based on the drug content obtained by elemental analysis. For each intercalated phase, the release assays were performed in triplicate. The mean values and standard deviations are presented in the graphs. 1245

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

3. RESULTS AND DISCUSSION 3.1. Intercalation of Benzoxaborolates in LDH. 3.1.1. Post-intercalation of BBzx. The LDH-BBzx hybrid material was prepared through an anion exchange process, by replacing the nitrates initially present between the layers of the Mg/Al LDH precursor with benzoxaborolates. The reaction was carried out at 50 °C overnight, and the sample was then washed twice to remove any adsorbed molecules and dried under vacuum at room temperature. A series of characterization experiments was then performed on this sample, and comparisons were made with a “control” LDH sample (LDH-0), for which a similar protocol was applied but without any benzoxaborolate in the solution. According to SEM, the overall aspect of the LDH sheets was maintained in the presence of benzoxaborolates (Figure 2A). However, the XRD powder patterns were distinct (Figure 2B). Indeed, compared to LDH-0, for which a basal spacing of 8.8 Å typical of a nitrateintercalated LDH was observed, new diffraction peaks were obtained for LDH-BBzx at low angles, with a basal spacing of 29.5 Å suggesting the successful intercalation of the BBzx anions in the interlayer space. Characterization by solid state NMR was performed to gain further insight into the structure of the intercalated phase, by looking at the inorganic sheets (using 27Al NMR), or at the intercalated BBzx (using 11B NMR). The 27Al MAS NMR spectrum of LDH-0 (Figure 2C) showed only one signal at ∼12 ppm,36 which corresponds to octahedral aluminum environments as expected for a pure LDH phase.37,38 LDHBBzx presented a very similar 27Al signal showing that AlO6 environments are still present. However, the 27Al line shape was slightly different, and additional high field fast 27Al → 1H CPMAS experiments demonstrated that modifications of the local environment of 27Al had occurred following the intercalation of BBzx between the layers (see SI, p S16). The 11 B MAS NMR spectrum of LDH-BBzx (Figure 2D) presented two different signals between 3 to 32 ppm. Given that no NMR studies on benzoxaborolate anions in the solid state had been reported yet, we prepared model compounds involving benzoxaborolate anions in order to help with the assignment of the 11B signals of LDH-BBzx. The model materials were obtained by precipitation of a solution of benzoxaborolates with an aqueous solution of metal cations. Magnesium and aluminum were first used as precipitating cations, for a direct comparison with the Mg/AlLDH materials. Despite numerous attempts, no crystalline aluminum benzoxaborolate phase could be obtained in our synthetic conditions, due to the competitive precipitation of Alhydroxides. However, two different crystalline phases were isolated with magnesium: MgBBzx·7H2O and MgBBzx·10H2O. In addition, a crystalline phase involving Ca2+ was also successfully synthesized, CaBBzx·3H2O, which was included as part of this study due to its interesting features in terms of benzoxaborolate geometry. As illustrated in Figure 3A, benzoxaborolates were found to either act as ligands in these materials through various coordination modes, or to simply play a charge compensating role (with no direct binding to the metal ions). The coordinated benzoxaborolates were also found to exist in two different configurations: as a tetrahedral anion (as seen in all 3 crystal structures), but also as a planar anion for one of the benzoxaborolates in CaBBzx·3H2O. It should be noted that only tetrahedral benzoxaborolate anions were initially expected, because all syntheses had been carried out

in water (where the tetrahedral form is predominant), and that the few examples of planar benzoxaborolate anions reported so far corresponded to species isolated in anhydrous conditions.39 Using these crystalline phases, the 11B NMR signatures of benzoxaborolate anions in the solid-state were determined (Figure 3B). For both magnesium phases, only one signal with an isotropic chemical shift at ∼8 to 9 ppm was observed on the MAS spectra, which was assigned unambiguously to the tetrahedral benzoxaborolate anion. CaBBzx·3H2O had two signals with isotropic chemical shifts at ∼9 and ∼31 ppm in a 5:1 ratio, which is the ratio observed in the crystal structure between the number of inequivalent tetrahedral and planar benzoxaborolate ligands. The signal at ∼9 ppm corresponded thus to the tetrahedral benzoxaborolate anion, while the signal at ∼31 ppm was assigned to the planar benzoxaborolate anion. On the basis of the 11B NMR signatures of the model materials (Figure 3B), the 11B signal of LDH-BBzx with an isotropic chemical shift36 at ∼9 ppm was assigned unambiguously to the tetrahedral benzoxaborolate form. This was the form expected given that the anion exchange process had been carried out in water at high pH. However, the broad signal with an isotropic chemical shift at ∼31 ppm could correspond either to planar benzoxaborolate anions (as in CaBBzx·3H2O) or to the benzoxaboroles in their acid form, which also have an isotropic chemical shift at ∼31 ppm (Figure 3B).24,40 To discriminate between both forms, LDH-BBzx, CaBBzx·3H2O and the acid form of BBzx were exposed to a basic NH3/H2O atmosphere for 24 h, and then characterized by 11B solid-state NMR (see the SI, p S18). For CaBBzx·3H2O, the boron signals remained unchanged after exposure to the basic atmosphere, while new signals appeared in the case of BBzx, with notably a strong signal at ∼9 ppm consistent with the presence of the tetrahedral anion. For LDH-BBzx, the comparison between the samples before and after exposure clearly showed the disappearance of the broad high frequency signal, demonstrating that it corresponded to the acid form and not to the planar benzoxaborolate anion. It is worth noting that the comparison of the 11B asymmetry parameters ηQ between LDH-BBzx (ηQ ≈ 0.36), CaBBzx·3H2O (ηQ ≈ 0.68) and the acid form of BBzx (ηQ ≈ 0.48) further supports this conclusion, as the asymmetry parameter of LDH-BBzx is closer to the one observed for the acid form BBzx. Overall, on the basis of all these 11B NMR characterizations, it appears that both the acid form of the benzoxaborole and the tetrahedral benzoxaborolate anion are present within the material. The presence of the acid form within the LDH phase was quite surprising, given that the pH measurements performed on the suspension during the preparation of LDH-BBzx and on the supernatants after the different washings all gave a pH value at 9.2 ± 0.2, which is well above the pKa of BBzx (pKa ≈ 7.3 at 25 °C). In order to probe the location of the acid form of BBzx within the LDH-BBzx material, a 2D 11B−11B NOESY NMR experiment was performed (see SI p S18). A cross-peak was observed between the signals of the tetrahedral benzoxaborolate anion and the acid form, thereby revealing that both species were present in the same interlayer space. The presence of the acid form in LDH-BBzx could be associated with the evolution of tetrahedral benzoxaborolate anions in the interlayer space, as further discussed below. Apart from BBzx and the tetrahedral benzoxaborolate, other anions were also found to be present in the interlayer space. Residual nitrates were identified by N elemental analyses, while the presence of carbonates was attested by 13C solid state NMR 1246

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

Figure 4. A/ XRD powder pattern and B/ 11B MAS NMR spectrum of LDH-AN2690, in comparison with LDH-BBzx (and LDH-0).

Figure 5. A/ Evolution of the XRD powder patterns of a wet paste of LDH-BBzx, exposed to successive scans under the X-ray beam and B/ Schematic representation of the evolution of the structure of LDH-BBzx toward a second staging structure (the BBzx species are represented by gray bars and the inorganic anions by purple circles).

AN2690. However, as shown in Figure 4A, the XRD powder patterns of the dry samples were different: LDH-AN2690 presented a basal spacing at ∼16.3 Å, which is much smaller than the ∼29.5 Å observed for LDH-BBzx. Furthermore, only the signature of the AN2690 anion with tetrahedral boron geometry was observed by 11B solid state NMR (Figure 4B), and not the acid form. The presence of nitrates and carbonates was also detected within this material, and the chemical analyses of LDH-AN2690 were consistent with the formula Mg0.67Al0.33(OH)2(C7H7BFO3)0.25(NO3)0.02(CO3)0.03·0.5H2O, which corresponds to an anionic exchange of 74% towards AN2690, close to what had been found for LDH-BBzx. 3.1.3. Structure of Benzoxaborolate-Intercalated LDH Phases. Despite the similarities between BBzx and AN2690 in terms of size and reactivity, the XRD data recorded for LDHBBzx and LDH-AN2690 clearly revealed different intercalation modes. In the case of LDH-AN2690, the basal distance ∼16.3 Å is comparable to what has been reported for the intercalation of small molecules of similar size like saccharin.43 Given that the estimated thickness of the metal-hydroxide layer is ∼5 Å, the interlayer spacing would be ∼11.3 Å, meaning that

studies of the intercalated phase (see SI p S19). It is worth noting that due to the high affinity of carbonates for LDH, it is very difficult to avoid carbonate contamination (even when working under an inert atmosphere), and traces of these anions were actually already present in the starting Mg/Al LDH precursor (see SI p S2). Having determined the nature of the different ionic and molecular species in the material, the exchange capacity of Mg/ Al LDH toward BBzx species was estimated as ∼66% based on H, C, B, and N elemental analyses.41 The following formula could be proposed for the hybrid material, Mg0.67Al0.33(OH)2.07(C7H8BO3)0.14(C7H7BO2)0.07(NO3)0.06(CO3)0.03·0.5H2O, in which (i) the acid form of BBzx was assumed to come from the dehydroxylation of the tetrahedral benzoxaborolate, and (ii) the ratio between the acid BBzx and tetrahedral benzoxaborolate form was determined by 11B solid state NMR. 3.1.2. Post-intercalation of AN2690. Mg/Al LDH was intercalated by AN2690 using the same conditions as for BBzx. Due to the structural similarity of BBzx and AN2690 and their close pKa values (∼7.3 and 6.7, respectively),1,42 similar behaviors were expected between LDH-BBzx and LDH1247

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

Figure 6. A/ Evolution of the 19F solid state NMR spectra of LDH-AN2690 as a function of time, when stored at room temperature ((*) correspond to spinning sidebands); B/ 19F19F 2D NOESY NMR spectrum of an LDH-AN2690 phase after 4 months of storage at room temperature exhibiting a second staging structure (the cross-peaks are here circled in red; the dashed red circle shows the absence of cross-peak); C/ 13C{19F} REDOR spectra of LDH-AN2690 recorded with (red) and without (black) 19F recoupling pulses.

some degree of interdigitation of the AN2690 anions must be present (the “length” of AN2690 is ∼7 Å). In the case of LDH-BBzx, the new basal spacing value is much larger (∼29.5 Å). Considering the size of the BBzx species (∼7 Å “length”), and the interlayer organization observed for the crystalline model materials MgBBzx·7H2O and MgBBzx·10H2O, which led to a spacing between the planes of cations ranging from ∼12.5 to ∼15.7 Å (see SI p S4), such a large basal spacing value can only be explained by a second staging structure (i.e., a structure in which different interlayer distances are observed, which alternate periodically throughout the material). Second staging phenomena are well-known for LDH materials,44−46 but the reasons for their appearance have not been systematically established. Various parameters such as the synthetic conditions, the LDH composition or the nature of the organic guests have been found to play a role, and some studies on second staging structures related their appearance to the drying step, and mainly to the dehydration of the interlayer space.45,46a In our case, we found that similar second staging structures are also obtained when intercalating BBzx in Zn/Al LDH or phenylboronic acid in Mg/Al LDH (see SI, p S19). However, interesting observations were made when looking at the structure of LDH-BBzx before the drying step. The wet paste of LDH-BBzx obtained after centrifugation of the reaction medium was characterized by XRD, and successive scans of the same fraction with the X-ray beam were performed (Figure 5A). The diffractogram during the first scan clearly showed a different diffraction pattern compared to the dry powder, with a basal spacing at ∼16.3 Å, similar to the one observed for AN2690. At the second scan, two different diffraction patterns coexisted corresponding to the former structure and the second staging structure. At this stage, it was also observed that the paste started to crack and dry under the X-ray beam. At the third scan, the second staging pattern was predominant. Thus, globally, the intercalation of BBzx in LDH seemed to occur in two steps, with a transient phase evolving

quickly toward a second staging structure during the drying. It can be suggested that the removal of water in the interlayer space induced the local segregation of some molecules, and/or their tilting in different orientations between the layers, causing the appearance of the second staging structure. Similar hypotheses have been proposed by Zhang et al.46a and Kooli et al.45 where the intercalated molecules could adopt horizontal or vertical orientations with respect to the layers. Among the two models describing the formation of staging structures, i.e., the Rüdorff47 and Daumas-Hérold48 models, the latter was believed to occur in our case implying a distortion of the LDH layers as shown schematically in Figure 5B. Indeed, by TEM, the layers seemed to be distorted for several LDH particles, and the interlayer spaces were uneven from one part of the crystallites to the other (see SI p S20).49 Moreover, such a model involves a minimal distortion of the metal-hydroxide sheets of the LDH, with the preservation of the interlayer distance at ∼16 Å. In addition, it should be noted that similar evolutions toward staging structures by distortion of LDH sheets upon sample dehydration have been proposed.46a Knowing that the second staging structure of LDH-BBzx appeared after drying, we tried to go back to the transient structure by rehydration. Exposure of a dry powder of LDHBBzx to a saturated aqueous atmosphere did not lead to the elimination of the second staging structure. However, when the dry powder was fully redispersed in water, the resulting paste exhibited peaks characteristic of both the regular and second staging structures (see SI p S20), showing the reversibility of the intercalation mode, depending on the hydration level of the interlayer space. Having observed the evolution of LDH-BBzx towards a second staging structure, the long-term stability of the LDHAN2690 anion between the LDH layers was also monitored. XRD powder patterns and 11B NMR spectra of LDH-AN2690 phases stored at room temperature for ∼4 months were recorded (see SI p S21). The XRD powder pattern of this 1248

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

shared in one part of the interlayer space, while in the other only tetrahedral benzoxaborolate anions are present. Finally, complementary solid state NMR experiments were carried out to try to position more precisely the carbonate anions within the material: a 13C{19F} REDOR (rotational echo double resonance) NMR experiment was performed on the LDH-AN2690 phase to probe the 13C···19F through-space proximities. As shown on the spectra Figure 6C, it was possible to completely “turn off” the signals belonging to AN2690 in the presence of 19F recoupling pulses, while the carbonate signal did not dephase. This suggests that the carbon atom of the carbonates is separated by more than 4.9 Å from the fluorine atom of AN2690 (i.e., the longest C···F distance in AN2690), and thus that these anions are not in close vicinity to the AN2690-benzoxaborolate species. Given that some carbonates had been detected in the starting material, and that these anions are known to be difficult to exchange, we interpret this observation by the presence of some fully carbonated domains within the LDH-AN2690 and LDH-BBzx samples, which were not replaced by benzoxaborolates. Overall, this implies that the inorganic anions present in the same interlayer space as the benzoxaborol(at)e species are mainly nitrates (represented by purple circles in Figure 5B). 3.1.4. Modeling of the Interlayer Space. On the basis of the numerous NMR experiments, each of which is informative on a specific structural feature of the LDH-BBzx and LDH-AN2690 phases, we felt that it should be possible to propose a model of the interlayer space (dashed orange box in Figure 5B). Such a model would provide a better comprehension of these systems, and help to rationalize in particular the possibility of having stabilized the acid form of benzoxaboroles between the LDH layers. Computational models were thus developed using molecular dynamics, followed by DFT relaxations, and the validity of the calculations was verified using DFT GIPAW calculations (Gauge Including Projector Augmented Wave), by looking at the agreement between calculated and experimental NMR parameters. It is worth noting that prior to looking at complex LDH-benzoxaborol(at)e models, a validation of this kind of NMR calculation for benzoxaborolate phases was first performed using the model materials MgBBzx·10H2O and CaBBzx·3H2O, whose structures had been solved by singlecrystal X-ray diffraction (see SI p S23). A simulation of the interlayer space of LDH-BBzx is shown in Figure 7A. For this model, an interlayer distance of ∼16.2 Å was used, which is representative of one of the interlayer spaces of the second staging structure (Figure 5B). Benzoxaborole species were introduced as two tetrahedral anions and one neutral molecule between the layers, which corresponds to the ratio observed between the corresponding 11B resonances in the 11B NMR spectrum of LDH-BBzx. Nitrates were also used to compensate the charge, while carbonates were avoided, as they were found to be segregated in a different part of the material. The chemical formula of this first model was thus Mg 0.67 Al 0.33 (OH) 2 (C 7 H 8 BO 3 ) 0.11 (C 7 H 7 BO 2 ) 0.06 (NO 3 ) 0.22 · 0.56H2O. Attempts to enhance the quantity of BBzx species in the interlayer space (and thereby reduce the nitrate content) were found to be very challenging due to steric hindrances between the molecules.52 Moreover, it was observed that hydroxyl anions could not be used as charged compensating anions, as they were unstable “computationally” during the structure optimization, which is why nitrates were used instead in this first model. This suggests that the exact formula of the LDH-BBzx phase exhibiting a second staging structure may be

phase clearly showed the presence of a second staging structure, suggesting a similar evolution as for the LDH-BBzx system, due to the dehydration of the interlayer space. Moreover, new signals appeared in the 11B NMR spectrum, with notably the 11 B NMR signature of the acid form of AN2690 (isotropic chemical shift at ∼31 ppm), showing that partial dehydroxylation of the tetrahedral benzoxaborolate form of AN2690 has occurred between the layers. It is worth noting that all samples presenting the particular second staging structure also presented the acid form of benzoxaboroles between the layers, which made us think that both were linked, meaning, in other words, that having a second staging structure may have implied the presence of the acidic form, and vice versa. To test this hypothesis, the previously mentioned phase, obtained by exposing the LDHBBzx sample to a NH3,H2O atmosphere was thus characterized by XRD, as we had seen that it had allowed conversion of the acid form into the tetrahedral anion. The XRD powder pattern of this sample exhibited the same second staging structure as before exposure to NH3,H2O, meaning that the presence of the acid form was not necessary for observing the second staging phenomenon; however, this experiment did not allow a conclusion as to whether the initial presence of the acid form between the layers had triggered the evolution toward a second staging structure. To investigate in more detail the structure and composition of the interlayer space, 25Mg and 19F solid state NMR experiments were carried out. The 25Mg NMR spectra of LDH-0 and LDH-BBzx are shown in the SI (p S22). For LDH-0, one main resonance was observed, whose NMR parameters were similar to those previously reported for a pure nitrate LDH phase, 50 corresponding predominantly to Mg(OAl)3(OMg)3 environments. The 25Mg NMR spectrum of LDH-BBzx showed similar features as LDH-0, with very close NMR parameters, the local environment of 25Mg being only slightly more distributed after intercalation of BBzx. More importantly, no broad signal attesting to the formation of Mg−O(H)−B bonds (such as those present in the MgBBzx· 7H2O phase) was observed, suggesting that all the benzoxaborolates are simply in electrostatic interaction between the layers, and not covalently attached to the metal hydroxide sheets.51 In the case of LDH-AN2690, 19F NMR experiments were performed. The spectrum of a fresh phase exhibiting a regular structure displayed a unique fluorine signal at ∼−120 ppm (Figure 6A). However, after its evolution to a staging structure, new signals were found at ∼−119 and ∼−115 ppm which could correspond to the AN2690 anions, and at ∼−107 ppm, which is close to the fluorine signature of AN2690 under its acidic form (∼−108.6 ppm).24 This 19F NMR spectrum is thus consistent with the formation of acid AN2690 molecules between the layers, when switching from the homogeneous to the staging structure, and also shows that two types of tetrahedral benzoxaborolate species are actually present in the material (which had not been resolved using 11B solid state NMR). By performing a 2D 19 F19 F NOESY NMR experiment on a LDH-AN2690 sample exhibiting a staging structure (Figure 6B), a correlation was observed between the 19 F signals at ∼−115 ppm and ∼−107 ppm, showing the close proximity between the tetrahedral anion and the acid form, while the 19F signal at ∼−119 ppm did not correlate. Such results complement the second staging model proposed for LDH-BBzx (Figure 5B), by suggesting that the clustered benzoxaborole species (tetrahedral anions and acid form) are 1249

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

diffraction, and computational modeling provided a detailed description of the structure of the material, notably at the organic−inorganic interface, and of the local environment of the benzoxaborol(at)e species. Given that these materials can be useful to the development of novel formulations of benzoxaborole-type drugs, several of their properties were looked into in view of using them for this application: their maximum loading capacity toward benzoxaborol(at)es, their optimal storage conditions, and the release kinetics. 3.2. Evaluation of the Properties of the Materials for the Development of Novel Formulations. 3.2.1. Maximum Loading Capacity. The experimental procedures presented above for BBzx and AN2690 had led to exchange capacities ∼70% in benzoxaborol(at)es. For each kind of intercalated organo-anion, LDH materials present a specific exchange capacity, which can be modulated in various ways, for example by changing the nature of the starting LDH phase or the synthetic conditions used, such as an anion exchange process or a one-step synthesis.53 Different parameters were studied to try to enhance the BBzx content between the layers, as summarized in the SI (p S26). Using an anionic exchange process, parameters like temperature (50 °C or room temperature), concentration (1 equiv or 3 equiv of BBzx compared to nitrates), time (16 or 168 h of reaction) and protocol (3 successive ion exchange processes or 1) were varied. None of these parameters were found to enhance the BBzx loading.54 Although Cl− anions are considered as slightly less easy to exchange than NO3−,55 we also tried to intercalate BBzx in a chloride Mg/Al LDH. However, no significant intercalation was observed. Attempts to intercalate BBzx in Mg/Al LDH were also made using a one-step synthesis, which consisted of synthesizing the materials by the same coprecipitation method as for Mg/Al LDH at pH 10, but in the presence of benzoxaborolate anions in solution. The resulting materials exhibited a similar structure compared to the pure Mg/Al LDH phase, and no significant amount of BBzx was intercalated. Overall, it was found that preparing LDH-BBzx at 50 °C with 1 equivalent of BBzx led to the best results, yet with an exchange capacity around 66 ± 4% (standard deviation on 3 samples) corresponding to a loading of ∼30 mg of BBzx for 100 mg of LDH-BBzx. A slightly higher value was found for LDH-AN2690 (74 ± 8%). Such values can thus be considered as a benchmark for further potential developments of intercalated-LDH phases for benzoxaborole drug formulations. It is worth noting that similar exchange capacities (∼69%) have been reported for the intercalation of small organic species like theobromine or cyclamate in Mg/Al LDH,43 while full intercalation has been reported in some cases like for acesulfame43 or ibuprofen.20 Here, the limited loading capacity can be related to various factors such as the steric hindrance between benzoxaborol(at)es in the interlayer space (as suggested by computational modeling), the poor affinity of the benzoxaborolate function with the sheets of LDH, and the presence of residual carbonates which are known to be very difficult to exchange. 3.2.2. Optimal Storage Conditions. LDH-BBzx samples were found to be sensitive to temperature. Indeed, when dried at 50 °C (instead of under vacuum at room temperature), additional signals were observed in the 11B NMR spectra, which were resolved using a 3Q-MAS experiment, and assigned to boric acid and borate degradation products (see SI p S27). It was also found that even when BBzx species were intercalated under an intact form, the hybrid materials tended to evolve over

Figure 7. A/ Model of the interlayer space of LDH-BBzx (chemical formula = Mg0.67Al0.33(OH)2(C7H8BO3)0.11(C7H7BO2)0.06(NO3)0.22· 0.56H2O) and B/ 11B NMR parameters of LDH-BBzx compared to those calculated by GIPAW (calculated values are averages of 3 similar configurations).

more complex than the one proposed based on elemental analyses. Three different starting configurations corresponding to the composition given above were optimized. In all cases, the acid form of BBzx was found to be stable between the layers (in agreement with experimental observations), while various attempts to stabilize “computationally” the planar benzoxaborolate anion between the LDH layers were unsuccessful, as it systematically reconverted to the acidic form during the geometry optimization by deprotonating the metal hydroxide layers. GIPAW calculations were carried out on each of the three models. When comparing the average 11B NMR parameters calculated for these models with the experimental ones (Figure 7B), a fairly good agreement was found for both the isotropic chemical shift and the quadrupolar parameter CQ. It is worth noting that similar models involving the intercalation of the AN2690 tetrahedral benzoxaborolate anion have also been looked into, and preliminary results show that 19F chemical shifts are expected around −118/−120 ppm for these anions, as observed experimentally. Although additional calculations on LDH-BBzx and LDH-AN2690 phases would merit to be carried out (using for example larger supercells, and varying the water content between the layers), these were computationally prohibitive at this stage. Nevertheless, the initial models developed here help rationalize some of the experimental observations, and they can be seen as a very first step toward the modeling of hybrid interfaces between benzoxaborol(at)es and inorganic materials. All in all, it was found that it is possible to prepare hybrid organic−inorganic materials by intercalation of BBzx and AN2690 in Mg/Al LDH. As demonstrated above, the combination of multinuclear solid state NMR, powder X-ray 1250

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

Figure 8. Release kinetics of LDH-BBzx and LDH-AN2690 in dynamic and static conditions.

intercalated with AN2690 and BBzx in both conditions are presented in Figure 8. Globally, the hybrid materials exhibited a fast release, all molecules being released in ∼2 h in dynamic conditions and ∼3 h in static conditions. They also presented a high burst release with 50 to 65% of the molecules being released between 10 and 60 min, depending on the conditions. Moreover, no clear link was found between the release kinetics and the structures of LDH-AN2690 (regular structure) and LDH-BBzx (second staging structure). Depending on the molecules intercalated, different release rates have been reported for LDH systems,19b like a fast release with ibuprofen (100% of release in ∼2 h),20 or a slow release with heparin (∼40% of release in ∼120 h).59 Such differences in behavior can be attributed to differences in the affinity of the molecules to the LDH sheets and also to changes in the release mechanisms, which can be due to an anion exchange between the layers or to the dissolution of the LDH sheets. In order to determine our release mechanism, the resulting powder was recovered after complete release of the drugs, dried, and characterized. The diffraction peaks at low angles attributed to the intercalation of BBzx were lost after release (see SI p S28): for the recovered phases, the lowest basal spacing was at ∼8.7 Å, which is similar to the basal spacing of the starting LDH. No change was observed at high angles suggesting that the inorganic sheets of the LDH were preserved during the release, and therefore that an anionic exchange had occurred between the layers. Both chloride and hydrogen phosphate anions, which are in high concentration in PBS, can a priori participate in this release. Indeed, chloride anions have a higher affinity for Mg/Al LDH than benzoxaborolates (see the above section on the “Maximum Loading Capacity”). Moreover, hydrogen phosphates have been shown to have an even higher affinity than chlorides for LDH (especially around pH 7),60 and they were actually detected in the material recovered after release by 31P solid-state NMR (see SI p S28). All this information will be useful for future developments of inorganic nanocarriers for benzoxaborolates drugs and can serve as a starting point for future studies on these systems. More complex release media will of course need to be considered to study the final formulations, depending on the therapeutic target.

time when they were stored at room temperature. Indeed, a series of 11B MAS NMR spectra of a LDH-BBzx sample after different durations of storage at room temperature were recorded, and the degradation of BBzx is clearly observed, with the formation of boric acid and borate, whose concentrations increase over time (see SI p S28). Similar observations were made for phases intercalated by AN2690. This suggests that the confinement of benzoxaborol(at)es in a basic Mg/Al LDH matrix favors their degradation. To check whether the degradation was indeed due to the basicity of this host-matrix,56 a study was performed in order to compare the stability over time of BBzx in 2 different LDHs: the Mg/Al LDH and a Zn/Al LDH known to be less basic.57,58 Both intercalated LDH displayed similar signatures by XRD and 11B MAS NMR. They were stored at −22 °C, and their degradation was followed over time by 11B solid-state NMR (see SI p S28). Freshly synthesized phases exhibited only BBzx signals. After 6 months, the signals due to degradation of BBzx started to appear in the Mg/Al-LDH sample, with the presence of a weak signal of borate corresponding to ∼3% of the total boron. After one year, no further degradation of the BBzx intercalated in the Mg/Al phase was observed, while for the Zn/Al LDH all BBzx species were still intact. These results clearly demonstrate the fragility of the BBzx inside LDH hosts: the confinement in this basic matrix can participate in the cleavage of the BC bond. However, when handling carefully the LDH-BBzx phases (i.e., drying at room temperature under vacuum), and when storing them at low temperatures (−22 °C), the degradation of the intercalated species can be avoided over months. From a more general perspective, all these observations show that the potential fragility of benzoxaborol(at)es in contact with specific inorganic matrices should be borne in mind when developing hybrid materials involving these molecules. 3.2.3. Release Kinetics. As intercalating the benzoxaboroles in LDH could be valuable for the development of new formulations of such molecules, we were interested in the release profiles from the resulting hybrid materials. A simple phosphate buffer (PBS, pH 7.4) was used as release medium in these initial tests, in line with previous in vitro studies on intercalated LDH.19b Benzoxaborole release assays were performed in both dynamic and static conditions, so as to evaluate how the material would behave to different solution fluxes, such as those encountered if the material was to be administered orally (dynamic conditions), or under the form of an implant (static conditions). Release kinetics of the LDH

4. CONCLUSIONS In this study, we described the preparation of the first hybrid organic−inorganic materials involving benzoxaboroles (BBzx and AN2690), by associating them with an inorganic matrix, 1251

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials Notes

Mg/Al LDH. The hybrid materials were fully characterized, and to help with the establishment of the structural and spectroscopic signatures of the benzoxaborolate anions in the solid state, three new crystalline phases were also prepared (CaBBzx·3H2O, MgBBzx·7H2O, and MgBBzx·10H2O). DFT computational modeling of the interlayer space was also performed, helping to rationalize the possibility of having both the acidic form of the benzoxaborole and the tetrahedral anion in the same interlayer space. Among our various investigations, we determined some of the properties of these intercalated phases, which could be useful for drug-delivery applications of these materials. In particular, the maximum loading capacity was found to be ∼30 wt %, and the fragility of the benzoxaboroles confined inside the basic LDH matrix was highlighted. The release profiles of LDH-benzoxaborole phases were determined in vitro, showing a fast release governed mainly by an anionic exchange of the benzoxaborolates between the layers. For such materials to be used for specific drug delivery applications, investigations on their shaping will next need to be considered, for instance to prepare LDH-based nanocarriers of controlled size and shape, such as those recently described in the literature.61 More generally, all the information gathered throughout this study can be seen as a milestone for the development of other hybrid formulations for benzoxaborole-based drugs, and we are currently focusing on means to modulate the release kinetics in order to have formulations which are adaptable to specific therapeutic needs. We expect that such hybrid formulations will open new opportunities for the numerous benzoxaborole-based drugs currently being developed. Beyond the potential drug-delivery applications, several important features could be highlighted through this first investigation of hybrid materials involving benzoxaborol(at)es. Indeed, it was shown that although highly attractive, the direct association of benzoxaboroles to inorganic hosts is not necessarily straightforward, because benzoxaboroles can be sensitive to temperature when confined in basic matrices, leading to the formation of byproducts. Such considerations are thus essential when trying to prepare hybrid organic−inorganic materials involving benzoxaboroles, and more generally boronic acids. On a different point, it should be noted that through the crystalline model phases developed in this work, it was shown for the first time that benzoxaborolates exhibit a very rich coordination chemistry. Indeed, they can act as multidentate bridging ligands, making them attractive for the preparation of materials like metal organic frameworks (MOFs). As such, they could lead to a novel family of porous materials, and we are currently looking in this direction.



The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the ANR (ANR JCJC “BOROMAT”), the seventh European framework program (Marie Curie ERG 239206), and a CNRS PICS project (QMAT-NMR). Synchrotron X-ray experiments were conducted at the Paul Scherrer Institute at Villigen in Switzerland (project 20120147), and A. Cervellino is acknowledged for his assistance in setting up the beamline. The UK 850 MHz solidstate NMR Facility used in this research was funded by EPSRC and BBSRC, as well as the University of Warwick including via part funding through Birmingham Science City Advanced Materials Projects 1 and 2 supported by Advantage West Midlands (AWM) and the European Regional Development Fund (ERDF). Financial support from the TGIR RMN THC FR3050 is acknowledged, and S. Cadars is thanked for his assistance in the NMR measurements performed in this context. DFT calculations were performed on the IDRIS supercomputer centre of the CNRS (Project 091461). M.-A. Pizzoccaro, F. Godiard, V. Richard, and Y. Guari are thanked for their assistance in sample preparation and electron microscopy measurements.



(1) (a) Baker, S. J.; Tomsho, J. W.; Benkovic, S. J. Chem. Soc. Rev. 2011, 40, 4279. (b) Liu, C. T.; Tomsho, J. W.; Benkovic, S. J. Bioorg. Med. Chem. 2014, 22, 4462. (2) Adamczyk-Wozniak, A.; Cyranski, M. K.; Zubrowska, A.; Sporzynski, A. J. Organomet. Chem. 2009, 694, 3533. (3) (a) Dowlut, M.; Hall, D. G. J. Am. Chem. Soc. 2006, 128, 4226. (b) Bérubé, M.; Dowlut, M.; Hall, D. G. J. Org. Chem. 2008, 73, 6471. (c) Peters, J. A. Coord. Chem. Rev. 2014, 268, 1. (4) Pal, A.; Bérubé, M.; Hall, D. G. Angew. Chem. 2010, 49, 1492. (5) Schumacher, S.; Katterle, M.; Hettrich, C.; Paulke, B.-R.; Hall, D. G.; Scheller, F. W.; Gajovic-Eichelmann, N. J. Mol. Recognit. 2011, 24, 953. (6) (a) Ellis, G. A.; Palte, M. J.; Raines, R. T. J. Am. Chem. Soc. 2012, 134, 3631. (b) Biswas, S.; Kinbara, K.; Niwa, T.; Taguchi, H.; Ishii, N.; Watanabe, S.; Miyata, K.; Kataoka, K.; Aida, T. Nat. Chem. 2013, 5, 613. (7) Zhang, Y.; Ma, W.; Li, D.; Yu, M.; Guo, J.; Wang, C. Small 2014, 10, 1379. (8) Li, H.; Wang, H.; Liu, Y.; Liu, Z. Chem. Commun. 2012, 48, 4115. (9) (a) Kotsuchibashi, Y.; Agustin, R. V. C.; Lu, J.-Y.; Hall, D. G.; Narain, R. ACS MacroLett. 2013, 2, 260. (b) Lin, M.; Sun, P.; Chen, G.; Jiang, M. Chem. Commun. 2014, 50, 9779. (c) Kotsuchibashi, Y.; Ebara, M.; Sato, T.; Wang, Y.; Rajender, R.; Hall, D. G.; Narain, R.; Aoyagi, T. J. Phys. Chem. B 2014, doi: 10.1021/jp506478p. (10) (a) Mendes, R. E.; Alley, M. R. K.; Sader, H. S.; Biedenbach, D. J.; Jones, R. N. Antimicrob. Agents Chemother. 2013, 57, 3334. (b) Jacobs, R. T.; Nare, B.; Wring, S. A.; Orr, M. D.; Chen, D.; Sligar, J. M.; Jenks, M. X.; Noe, R. A.; Bowling, T. S.; Mercer, L. T.; Rewerts, C.; Gaukel, E.; Owens, J.; Parham, R.; Randolph, R.; Beaudet, B.; Bacchi, C. J.; Yarlett, N.; Plattner, J. J.; Freund, Y.; Ding, C.; Akama, T.; Zhang, Y.-K.; Brun, R.; Kaiser, M.; Scandale, I.; Don, R. PLoS Negl. Trop. Dis. 2011, 5, e1151. (c) Akama, T.; Dong, C.; Virtucio, C.; Freund, Y. R.; Chen, D.; Orr, M. D.; Jacobs, R. T.; Zhang, Y.-K.; Hernandez, V.; Liu, Y.; Wu, A.; Bu, W.; Liu, L.; Jarnagin, K.; Plattner, J. J. Bioorg. Med. Chem. Lett. 2013, 23, 5870. (d) Akama, T.; Baker, S. J.; Zhang, Y.-K.; Hernandez, V.; Zhou, H.; Sanders, V.; Freund, Y.; Kimura, R.; Maples, K. R.; Plattner, J. J. Bioorg. Med. Chem. Lett. 2009, 19, 2129. (11) http://www.anacor.com.

ASSOCIATED CONTENT

S Supporting Information *

Additional data on the model materials (including synthesis conditions and.cif files); additional experimental details; further X-ray diffraction and solid state NMR data on LDH-BBzx and LDH-AN2690-type phases; additional computational models. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] 1252

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials

Chem. 2011, 50, 7802. (b) Sene, S.; Reinholdt, M.; Renaudin, G.; Berthomieu, D.; Zicovich-Wilson, C. M.; Gervais, C.; Gaveau, P.; Bonhomme, C.; Filinchuk, Y.; Smith, M. E.; Nedelec, J.-M.; Begu, S.; Mutin, P. H.; Laurencin, D. Chem.Eur. J. 2013, 19, 880. (35) U.S. Pharmacopeia & National Formulary, 1999. (36) For quadrupolar nuclei like 27Al and 11B, the isotropic chemical shift is positioned on the left extremity of MAS NMR signal, at the foot of the peak. (37) McKenzie, K.J. D.; Smith, M. E. Multinuclear Solid-State NMR of Inorganic Materials; Pergamon: New York, 2002. (38) Vyalikh, A.; Massiot, D.; Scheler, U. Solid State Nucl. Magn. Reson. 2009, 36, 19. (39) (a) Jaskowska, E.; Justyniak, I.; Cyranski, M. K.; AdamczykWozniak, A.; Sporzynski, A.; Zygado-Monikowska, E.; Ziemkowska, W. J. Organomet. Chem. 2013, 732, 8. (b) Ma, X.; Yang, Z.; Wang, X.; Roesky, H. W.; Wu, F.; Zhu, H. Inorg. Chem. 2011, 50, 2010. (40) Weiss, J. W. E.; Bryce, D. L. J. Phys. Chem. A 2010, 114, 5119. (41) It is worth noting that the Mg/Al ratio was checked by energydispersive X-ray spectroscopy (EDXS), and that no differences in this ratio were observed between LDH, LDH-0, or LDH-BBzx. Further details on the elemental analyses of intercalated LDH can be found in section 3.2 of the manuscript, and in the SI (p S26). (42) Tomsho, J. W.; Pal, A.; Hall, D. G.; Benkovic, S. J. ACS Med. Chem. Lett. 2012, 3, 48. (43) Markland, C.; Williams, G. R.; O’Hare, D. J. Mater. Chem. 2011, 21, 17896. (44) (a) Lerf, A. Dalton Trans. 2014, 43, 10276. (b) Pisson, J.; Taviot-Guého, C.; Israëli, Y.; Leroux, F.; Munsch, P.; Itié, J.-P.; Briois, V.; Morel-Desrosiers, N.; Besse, J.-P. J. Phys. Chem. B 2003, 107, 9243. (c) Feng, Y. J.; Williams, G. R.; Leroux, F.; Taviot-Gueho, C.; O’Hare, D. Chem. Mater. 2006, 18, 4312. (d) Ma, S.; Fan, C.; Du, L.; Huang, G.; Yang, X.; Tang, W.; Makita, Y.; Ooi, K. Chem. Mater. 2009, 21, 3602. (45) Kooli, F.; Chisem, I. C.; Vucelic, M.; Jones, W. Chem. Mater. 1996, 8, 1969. (46) (a) Zhang, Y.; Tan, H.; Zhao, J.-X.; Li, X.; Ma, H.; Chen, X.; Yang, X. Phys. Chem. Chem. Phys. 2012, 14, 9067. (b) Thyveetil, M.-A.; Coveney, P. V.; Greenwell, H. C.; Suter, J. L. J. Am. Chem. Soc. 2008, 130, 12485. (47) Rüdorff, W. Z. Z. Phys. Chem. B 1940, 45, 42. (48) Daumas, N.; Hérold, A. C. R. Séances Acad. Sci., Ser. C 1969, 268, 373. (49) We were unable to measure the interlayer spaces accurately using TEM, due to the sensitivity of the benzoxaborole species which were degraded almost as soon as the electron beam was focused on the crystallites. (50) Sideris, P. J.; Nielsen, U. G.; Gan, Z.; Grey, C. P. Science 2008, 321, 113. (51) The possibility of having formed Al-O-B bridges between the benzoxaborolate and the LDH was ruled out on the basis of computational modeling and GIPAW calculations (see the SI, p S29). (52) In some extreme cases, when trying to increase the benzoxaborolate content between the LDH layers, we even fell upon chemically invalid structures, in which the benzoxaborolate cycle had opened. (53) (a) Chibwe, K.; Jones, W. J. Chem. Soc., Chem. Commun. 1989, 926. (b) Leroux, F.; Adachi-Pagano, M.; Intissar, M.; Chauvière, S.; Forano, C.; Besse, J.-P. J. Mater. Chem. 2001, 11, 105. (c) Choy, J.-H.; Kwak, S.-Y.; Jeong, Y.-J.; Park, J.-S. Angew. Chem., Int. Ed. 2000, 39, 4041. (54) Only temperature was found to have a small effect on the diffraction peaks of the resulting materials, which were sharper when the reaction was performed at 50 °C. (55) Miyata, S. Clays Clay Miner. 1983, 31, 305. (56) The stability of the BBzx in basic media was studied. A solution of benzoxaborolates at pH > 13 was prepared and characterized by 11B solution NMR. It showed only one signal at ∼9 ppm. The same solution was preserved at room temperature and characterized again 3 weeks after. The spectrum exhibits a new signal at ∼2 ppm assigned to

(12) (a) Baker, S. J.; Zhang, Y.-K.; Akama, T.; Lau, A.; Zhou, H.; Hernandez, V.; Mao, W.; Alley, M. R. K.; Sanders, V.; Plattner, J. J. J. Med. Chem. 2006, 49, 4447. (b) Rock, F. L.; Mao, W.; Yaremchuk, A.; Tukalo, M.; Crépin, T.; Zhou, H.; Zhang, Y.-K.; Hernandez, V.; Akama, T.; Baker, S. J.; Plattner, J. J.; Shapiro, L.; Martinis, S. A.; Benkovic, S. J.; Cusack, S.; Alley, M. R. K. Science 2007, 316, 1759. (c) Hui, X.; Baker, S. J.; Wester, R. C.; Barbadillo, S.; Cashmore, A. K.; Sanders, V.; Hold, K. M.; Akama, T.; Zhang, Y.-K.; Plattner, J. J.; Maibach, H. I. J. Pharm. Sci. 2007, 96, 2622. (13) Welsch, M. E.; Snyder, S. A.; Stockwell, B. R. Curr. Opin. Chem. Biol. 2010, 14, 347. (14) (a) Nicole, L.; Laberty-Robert, C.; Rozes, L.; Sanchez, C. Nanoscale 2014, 6, 6267. (b) Sanchez, C.; Boissiere, C.; Cassaignon, S.; Chaneac, C.; Durupthy, O.; Faustini, M.; Grosso, D.; LabertyRobert, C.; Nicole, L.; Portehault, D.; Ribot, F.; Rozes, L.; Sassoye, C. Chem. Mater. 2014, 26, 221. (c) Sanchez, C.; Shea, K. J.; Kitagawa, S. Chem. Soc. Rev. 2011, 40, 471 (themed issue on hybrid materials). (15) (a) Newman, S. P.; Jones, W. New J. Chem. 1998, 105. (b) Layered Double Hydroxides: Present and Future Rives, V., Ed.; Nova Science Publishers: New York, 2001. (c) Khan, A. I.; O’Hare, D. J. Mater. Chem. 2002, 12, 3191. (d) Leroux, F.; Taviot-Guého, C. J. Mater. Chem. 2005, 15, 3628. (e) Ma, R.; Liang, J.; Liu, X.; Sasaki, T. J. Am. Chem. Soc. 2012, 134, 19915. (16) (a) Tichit, D.; Coq, B. Cattech 2003, 7, 206. (b) He, S.; An, Z.; Wei, M.; Evans, D. G.; Duan, X. Chem. Commun. 2013, 49, 5912. (17) (a) Theiss, F. L.; Ayoko, G. A.; Frost, R. L. J. Colloid Interface Sci. 2013, 402, 114. (b) Yu, Y.; Chen, J. P. J. Mater. Chem. A 2014, 2, 8086. (18) Evans, D. G.; Duan, X. Chem. Commun. 2006, 5, 485. (19) (a) Del Hoyo, C. Appl. Clay Sci. 2007, 36, 103. (b) Bi, X.; Zhang, H.; Dou, L. Pharmaceutics 2014, 6, 298. (20) Ambrogi, V.; Fardella, G.; Grandolini, G.; Perioli, L. Int. J. Pharm. 2001, 220, 23. (21) Wei, M.; Pu, M.; Guo, J.; Han, J.; Li, F.; He, J.; Evans, D. G.; Duan, X. Chem. Mater. 2008, 20, 5169. (22) (a) Panda, H. S.; Srivastava, R.; Bahadur, D. J. Phys. Chem. B 2009, 113, 15090. (b) Cunha, V. R. R.; Petersen, P. A. D.; Gonçalves, M. B.; Petrilli, H. M.; Taviot-Gueho, C.; Leroux, F.; Temperini, M. L. A.; Constantino, V. R. L. Chem. Mater. 2012, 24, 1415. (23) Chen, C.; Yee, L. K.; Gong, H.; Zhang, Y.; Xu, R. Nanoscale 2013, 5, 4314. (24) Sene, S.; Berthomieu, D.; Donnadieu, B.; Richeter, S.; Vezzani, J.; Granier, D.; Begu, S.; Mutin, P. H.; Gervais, C.; Laurencin, D. CrystEngComm 2014, 16, 4999. (25) Tichit, D.; Lutic, D.; Coq, B.; Durand, R.; Teissier, R. J. Catal. 2003, 219, 167. (26) (a) Kresse, G.; Hafner, J. Phys. Rev. B 1993, 47, 558. (b) Kresse, G.; Hafner, J. Phys. Rev. B 1994, 49, 14251. (27) (a) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. (b) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1997, 78, 1396. (28) (a) Blochl, P. E.; Jepsen, O.; Andersen, O. K. Phys. Rev. B 1994, 49, 16223. (b) Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, 1758. (29) Cadars, S.; Layrac, G.; Gérardin, C.; Deschamps, M.; Yates, J. R.; Tichit, D.; Massiot, D. Chem. Mater. 2011, 23, 2821. (30) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; Dal Corso, A.; Fabris, S.; Fratesi, G.; de Gironcoli, S.; Gebauer, R.; Gerstmann, U.; Gougoussis, C.; Kokalj, A.; Lazzeri, M.; Martin-Samos, L.; Marzari, N.; Mauri, F.; Mazzarello, R.; Paolini, S.; Pasquarello, A.; Paulatto, L.; Sbraccia, C.; Scandolo, S.; Sclauzero, G.; Seitsonen, A. P.; Smogunov, A.; Umari, P.; Wentzcovitch, R. M. J. Phys.−Condens. Matter 2009, 21, 395502. (31) Troullier, N.; Martins, J. L. Phys. Rev. B 1991, 43, 1993. (32) Kleinman, L.; Bylander, D. M. Phys. Rev. Lett. 1982, 48, 1425. (33) Pickard, C. J.; Mauri, F. Phys. Rev. B 2001, 63, 245101. (34) (a) Reinholdt, M.; Croissant, J.; Di Carlo, L.; Granier, D.; Gaveau, P.; Bégu, S.; Devoisselle, J.-M.; Mutin, P. H.; Smith, M. E.; Bonhomme, C.; Gervais, C.; van der Lee, A.; Laurencin, D. Inorg. 1253

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254

Article

Chemistry of Materials borates, which means that over time and in highly basic media, degradation of BBzx occurs. (57) Kagunya, W.; Hassan, Z.; Jones, W. Inorg. Chem. 1996, 35, 5970. (58) The Zn/Al LDH was synthesized using the same method by coprecipitation at constant pH (pH 8.0 instead of pH 10.0 for the Mg/ Al LDH). It was then intercalated with BBzx using the same intercalation process, and then drying the sample under vacuum at room temperature. (59) Gu, Z.; Thomas, A. C.; Xu, Z. P.; Campbell, J. H.; Lu, G. Q. Chem. Mater. 2008, 20, 3715. (60) Ookubo, A.; Ooi, K.; Hayashi, H. Langmuir 1993, 9, 1418. (61) (a) Gunawan, P.; Xu, R. Chem. Mater. 2009, 21, 781. (b) Bao, H.; Yang, J.; Huang, Y.; Xu, Z. P.; Hao, N.; Wu, Z.; Lud, G. Q.; Zhao, D. Nanoscale 2011, 3, 4069. (c) Bi, X.; Fan, T.; Zhang, H. ACS Appl. Surf. Interface 2014, 6, 20498. (d) Layrac, G.; Destarac, M.; Gérardin, C.; Tichit, D. Langmuir 2014, 30, 9663. (e) Gu, Z.; Atherton, J. J.; Xu, Z. P. Chem. Commun., doi: 10.1039/c4cc07715.

1254

DOI: 10.1021/cm504181w Chem. Mater. 2015, 27, 1242−1254