Received 23 July 2013
Tuning solubility and stability of hydrochlorothiazide co-crystals
Hydrochlorothiazide (HCT), C7H8ClN3O4S2, is a diuretic BCS (Biopharmaceutics Classification System) class IV drug which has primary and secondary sulfonamide groups. To modify the aqueous solubility of the drug, co-crystals with biologically safe co-formers were screened. Multi-component molecular crystals of HCT were prepared with nicotinic acid, nicotinamide, succinamide, p-aminobenzoic acid, resorcinol and pyrogallol using liquid-assisted grinding. The co-crystals were characterized by FT-IR spectroscopy, powder X-ray diffraction (PXRD) and differential scanning calorimetry. Single crystal structures were obtained for four of them. The N-HO sulfonamide catemer synthons found in the stable polymorph of pure HCT are replaced in the co-crystals by drug-co-former heterosynthons. Isostructural co-crystals with nicotinic acid and nicotinamide are devoid of the common sulfonamide dimer/catemer synthons. Solubility and stability experiments were carried out for the co-crystals in water (neutral pH) under ambient conditions. Among the six binary systems, the co-crystal with p-aminobenzoic acid showed a sixfold increase in solubility compared with pure HCT, and stability up to 24 h in an aqueous medium. The co-crystals with nicotinamide, resorcinol and pyrogallol showed only a 1.5-2-fold increase in solubility and transformed to HCT within 1 h of the dissolution experiment. An inverse correlation is observed between the melting points of the co-crystals and their solubilities.
Design and synthesis of new multi-component molecular crystals by co-crystallization of an active pharmaceutical ingredient (API) is one of the important aspects of crystal engineering (Desiraju, 1995). Co-crystals are of importance for APIs because they offer opportunities to tune physicochemical properties such as solubility, bioavailability and stability (McNamara et al., 2006; Remenar et al., 2003; Good & Rodríguez-Hornedo, 2009). Co-crystals, as substances of new chemical composition, not only engineer new properties, but also can achieve new patent protection and commercial value because of their novelty, non-obviousness and utility (Hanna et al., 2009). A co-crystal strategy is particularly useful for APIs having non-ionizable functional groups, such as sulfonamide, carboxamide and phenol (Caira et al., 1995; Babu et al., 2012; Smith et al., 2011). Finding an optimum solid form with desired physicochemical properties is an important step during the drug development process.
Hydrochlorothiazide (HCT; see below), a diuretic drug, acts by inhibiting the kidney's ability to retain water (Dupont & Dideberg, 1972). It is a BCS (Biopharmaceutics Classification System) class IV drug (Amidon et al., 1995) of low solubility (0.7 g L-1 at 298 K) and low permeability (log P = -0.07). Due to its poor aqueous solubility, the API exhibits bioavailability only up to 65% (Patel et al., 1984). There is a report on solubility and bioavailability improvement for a drug-drug solid dispersion of hydrochlorothiazide and the diabetic drug captopril (Padmapriya et al., 2011). The recent literature also reports solubility enhancement of the API by forming solid dispersions with losartan potassium and urea (Trivedi et al., 2011). Hydrochlorothiazide exhibits four polymorphs, two of which [forms (I) and (II)] have published three-dimensional coordinates (Kim & Kim, 1984; Leech et al., 2008). Form (I) is the most stable, and its crystal structure contains catemeric hydrogen-bond motifs of the sulfonamide. To form co-crystals of hydrochlorothiazide, this catemer chain must be interrupted in the presence of co-formers by supramolecular heterosynthon formation (Shattock et al., 2008). With this in mind, biologically safe aromatic carboxylic acids, carboxamide and phenols are screened here in an effort to form co-crystals. Hydrochlorothiazide co-crystals with nicotinic acid and piperazine have been reported previously by PXRD patterns in a patent (Almarsson et al., 2007), but there is no report on their crystal structures. The crystal structure of an HCT-isonicotinamide co-crystal hydrate has also been reported recently (Clarke et al., 2010). Further, the API has been shown to form several solvates (Florence et al., 2005; Johnston et al., 2006a,b). There is no report in the literature on solubility improvement by co-crystallization of the API. Hence, we set out here to improve the solubility of HCT by preparing anhydrous co-crystals with GRAS (Generally regarded as safe; chemicals approved by the US-FDA) co-formers.
Hydrochlorothiazide (m.p. 542-454 K) was obtained from Sigma Aldrich Chemicals, Bangalore, India, and used directly for experiments. All other reagents were purchased from commercial sources and used without further purification. Melting points were measured on a Büchi melting point apparatus. Water used in the experiments was filtered through a double-distilled water purification system (Siemens, Ultra Clear, Germany).
Fourier transform IR (FT-IR) spectra were recorded as KBr pellets with a Perkin-Elmer (UK) spectrophotometer (4000-400 cm-1). Powder X-ray diffraction (PXRD) data were recorded using a Philips X'pert Pro X-ray powder diffractometer equipped with Cu K radiation and an X'celerator detector at room temperature over the range 2 = 5-40°, with step size 0.017°. HighScore Plus (Panalytical, 2012) and PowderCell (Kraus & Nolze, 1996) were used to compare experimental PXRD patterns with patterns calculated from the crystal structures. Differential scanning calorimetry (DSC) was performed on a Mettler Toledo DSC 822e instrument with heating rate 10 K min-1 over the temperature range 303-573 K under an N2 atmosphere. Liquid-assisted grinding, solution crystallization and slurry methods in polar solvents such as MeOH, CH3CN were used to obtain the co-crystals. All of the co-formers that were used for co-crystallization attempts with HCT are listed in the supporting information.1
HCT (100 mg , 0.33 mmol) and NIC (41.0 mg, 0.33 mmol) were ground with a mortar and pestle for 15 min in the presence of a few drops of MeOH and then the ground mixture was crystallized from MeOH to obtain single crystals. Colourless needle-like crystals were harvested after 3-4 days; m.p. 531-535 K.
HCT (100 mg, 0.33 mmol) and NCT (41.4 mg, 0.33 mmol) were ground with a mortar and pestle for 15 min in the presence of a few drops of MeOH and then the ground mixture was crystallized from MeOH to obtain single crystals. Colourless plate-like crystals were grown after 3-4 days; m.p. 446-448 K.
HCT (100 mg, 0.33 mmol) and SAM (20.2 mg, 0.33 mmol) were ground with a mortar and pestle for 15 min in the presence of a few drops of MeOH and then the ground mixture was crystallized from MeOH to obtain single crystals. Colourless block-like crystals were obtained after 3-4 days; m.p. 508-511 K.
HCT (100 mg, 0.33 mmol) and PABA (92.1 mg, 0.67 mmol) were ground with a mortar and pestle for 15 min in the presence of a few drops of MeOH and then the ground mixture was crystallized from MeOH to obtain single crystals. Colourless rod-like crystals were obtained after 3-4 days; m.p. 449-451 K.
HCT (100 mg, 0.33 mmol) and resorcinol (37.0 mg, 0.33 mmol) or pyrogallol (84.8 mg, 0.66 mmol) were ground with a mortar and pestle in separate experiments for 15 min in the presence of a few drops of MeOH. We were not able to produce single crystals in spite of several crystallization attempts; m.p. 446-450 K (HCT-RES) and 453-459 K (HCT-PYR).
Single-crystal X-ray data for all co-crystals were collected on a Rigaku Mercury 375/M CCD (XtaLAB mini) diffractometer using graphite-monochromated Mo K radiation. H atoms were treated by a mixture of constrained and free refinement. H atoms bound to C atoms were placed geometrically and treated as riding, while H atoms bound to N and of the carboxylic acid groups were located in difference Fourier maps and refined with isotropic displacement parameters. For HCT-NIC and HCT-NCT, the absolute structure was determined satisfactorily by refinement of the Flack parameter (Flack, 1983).
The absorption coefficients of HCT and the co-crystals were measured from the slope of the absorbance versus concentration curves for concentrated solutions in distilled water, measured with a Perkin-Elmer UV-vis spectrometer at 317 nm to avoid co-former interference (in the 250-280 nm region). The solubility of each solid was measured at 1, 4 and 24 h using the shake-flask method (Glomme et al., 2005).
Hydrochlorothiazide (HCT) exists as two crystalline polymorphs (Leech et al., 2008). The stable form (I) (space group P21) consists of N-HO hydrogen-bonded catemer chains of the primary sulfonamides, whereas the secondary sulfonamide supports the catemer chain by interaction with the NH of the primary sulfonamide (Fig. 1a). In the metastable form (II) (space group P21/c), the secondary sulfonamide forms N-HO hydrogen-bonded dimers, whereas the primary sulfonamide is involved in hydrogen bonding with the secondary amine NH (Fig. 1b). Various synthon motifs observed in HCT polymorphs and solvates are illustrated below: (a) primary sulfonamide catemer [observed in form (I)]; (b) primary sulfonamide dimer; (c) secondary sulfonamide dimer [observed in form (II)]; (d) dimer involving both primary and secondary sulfonamide.
| || Figure 1 |
Crystal structures of the reported HCT polymorphs: (a) the stable form (I) contains primary sulfonamide catemer chains; (b) the metastable form (II) contains secondary sulfonamide dimers.
We report here the structures of HCT co-crystals with nicotinic acid (NIC), nicotinamide (NCT), succinamide (SAM) and p-aminobenzoic acid (PABA). The crystallographic parameters and normalized hydrogen bonds for these co-crystals are summarized in Tables 1 and 2. The formation of HCT-RES and HCT-PYR co-crystals by liquid-assisted grinding was confirmed by powder X-ray diffraction, which showed clear differences from the API and co-former (see the supporting information ). Characteristic peaks are observed at 2 values of 11.7, 14.4, 15.2 and 26.8° for the HCT-RES co-crystal. Similarly, new diffraction peaks are observed at 2 values of 13.3, 17.3 and 23.9° for the HCT-PYR co-crystal. However, solvent re-crystallization of these products always resulted in separation of HCT and the co-former, and we have not been able to establish the structures of these co-crystals to date.
The crystal structure of HCT-NIC (space group P212121, Z = 4) comprises HCT molecules and nicotinic acid (NIC) in a zwitterionic form. The zwitterionic NIC molecules assemble by intermolecular charge-assisted N+-HO- hydrogen bonds (N4-H4NO6; Table 2), forming a one-dimensional chain parallel to the a axis. The sulfonamide catemer motif of the primary sulfonamide groups found in the stable polymorph of HCT is interrupted by the co-former in the co-crystals. N-HO- intermolecular hydrogen bonds (N3-H3AO5 and N3-H3BO5; Table 2) are formed between the primary sulfonamide NH group and the zwitterionic NIC carboxylate group (Fig. 2a). The HCT molecules assemble via C-ClO=S=O halogen bonds and N-HO interactions involving the secondary sulfonamide. The molecular planes lie approximately in layers parallel to the (010) planes (Fig. 2b), and each layer consists of chains of HCT and NIC molecules which interact with the next layer through N-HO (HCT) and N-HO- (NIC) interactions. The three-dimensional packing arrangement shows each chain of NIC molecules (running parallel to the a axis) surrounded by four chains of HCT molecules (Fig. 2b). The zwitterionic form of nicotinic acid is rare in the CSD (Cambridge Structural Database; Allen, 2002), and is present only in four co-crystals with L-ascorbic acid (CSD: RUWFAR; Kavuru et al., 2010), gallic acid (RUWHAT; Kavuru et al., 2010), rac-hesperetin (RUWHEX; Kavuru et al., 2010) and (+)-hesperetin (RUWHIB; Kavuru et al., 2010), and in one salt with picric acid (UBEHUG; Stilinovic & Kaitner, 2011).
| || Figure 2 |
Structure of the HCT-NIC co-crystal: (a) chains of HCT and NIC running along the a axis; (b) packing diagram viewed along the a axis.
The HCT-NIC and HCT-NCT co-crystals are isomorphs (Smolka et al., 1999; Sarma et al., 2008). The HCT-NCT co-crystal (space group P212121, Z = 4) contains nicotinamide molecules forming one-dimensional chains parallel to the a axis through amidepyridine synthons involving N-HN hydrogen bonds (N4-H4BN5; Table 2; Fig. 3a). The interactions between the HCT molecules are essentially identical to those described for HCT-NIC (N1-H1NO1), and N-HO intermolecular hydrogen bonds (N3-H3AO5) are formed between the primary sulfonamide NH group and the carbonyl group of the amide (Fig. 3b).
| || Figure 3 |
Structure of the HCT-NCT co-crystal: (a) chains of HCT and NCT running along the a axis; (b) packing diagram viewed along the a axis.
A competition experiment involving HCT, NIC and NCT in solution produced only the HCT-NIC co-crystal, confirming stronger interactions of HCT with NIC than with NCT. Similar co-formers such as benzoic acid, benzamide and salicylic acid failed to co-crystallize, indicating a need for the heterocyclic N atom within the co-former. However, we note that we were not able to co-crystallize HCT with isonicotinic acid.
The asymmetric unit of the HCT-SAM co-crystal (space group C2/c, Z = 4) consists of one HCT molecule and one succinamide molecule on a crystallographic twofold axis (thereby giving the 2:1 stoichiometry). Unlike the other co-crystals, the primary sulfonamide participates in a centrosymmetric N-HO hydrogen-bonded dimer (N3-H3AO2; Table 2), where the secondary sulfonamide NH supports the dimer motif through N-HO interactions (N2-H2O2; Fig. 4a). There is a further dimer formed by the HCT molecules, with an R22(16) ring motif, formed between the primary sulfonamide sulfonyl and secondary sulfonamide NH groups. Like the other co-crystals, the primary sulfonamide forms an N-HO hydrogen bond (N3-H3BO5) with the carbonyl acceptor of the co-former (here SAM), while the secondary sulfonyl group of HCT interacts with the amide of two succinamide molecules through N-HO interactions. Overall, each succinamide molecule is surrounded by six HCT molecules using its two carbonyls (as bifurcated hydrogen-bond acceptors) to four HCT molecules and two of its amine NH bonds as hydrogen-bond donors to two more molecules of HCT (Fig. 4b). Each HCT molecule is surrounded by three SAM molecules.
| || Figure 4 |
Structure of the HCT-SAM co-crystal: (a) interactions between HCT and SAM through N-HO hydrogen bonds, including the centrosymmetric primary sulfonamide dimer (centre); (b) one SAM molecule surrounded by six HCT molecules.
The HCT-PABA co-crystal (space group P21/c, Z = 4) contains carboxylic acid dimers (O5-H5O8; O7-H7OO6) formed between two symmetry-independent PABA molecules. The HCT molecules form ClCl type I halogen bonds (d = 3.33 Å, = 138.5°; Pedireddi et al., 1994), defining planar pairs (Fig. 5a). These pairs of molecules form N-HO hydrogen bonds to PABA molecules (N3-H3BO5) between the primary sulfonamide -NH2 and the -OH group of the carboxylic acid (Fig. 5a). HCT molecules related by the 21 screw axis along b interact through N-HO hydrogen bonds (N3-H3AO3) between the secondary sulfonamide -SO2 and primary sulfonamide NH, in a manner similar to that seen in HCT form (I) (Fig. 1a), and further interact with PABA via N-HN interactions (Fig. 5b). The three-dimensional packing resembles a layered structure, where hydrophobic and hydrophilic regions of HCT and PABA sit alternately (Fig. 5c).
| || Figure 5 |
Structure of the HCT-PABA co-crystal: (a) hydrogen bonding in HCT-PABA co-crystal; (b) 21 screw-axis related HCT molecules interact with PABA through N-HN hydrogen bonds; (c) packing diagram viewed down the a axis, showing hydrophobic and hydrophilic regions.
Sulfonamide dimer/catemer synthons are so robust that it is sometimes difficult to break them even in the presence of co-formers (Ueto et al., 2012; Goud et al., 2011). In the co-crystals reported here, both NIC and NCT are able to disrupt the sulfonamide catemer synthon present in the stable polymorph of HCT to yield co-crystals. This indicates that the presence of the pyridine unit in the co-former is able to break the robust sulfonamide synthon motif. HCT-NIC and HCT-NCT co-crystals are isomorphous due to the similar size and shape of the co-formers. However, HCT-SAM and HCT-PABA co-crystals maintain sulfonamide dimer and catemer synthons, respectively. Among the reported HCT solvate crystal structures (Florence et al., 2005; Johnston et al., 2006a,b), N,N-dimethylacetamide, dioxane and DMF disrupt the sulfonamide synthon, whereas other solvates such as aniline, methyl acetate, DMSO and N-methyl-2-pyrrolidone maintain either the sulfonamide catemer or dimer synthons in their crystal structures.
Differential scanning calorimetry (DSC) shows a sharp melting onset at 541 K for HCT, confirming its purity. The co-crystals exhibit single melting onsets at 530.8 (HCT-NIC), 446.6 (HCT-NCT), 508.1 (HCT-SAM) and 448.9 K (HCT-PABA), supporting the homogeneity of the phases (see the supporting information ). Further, the HCT-RES (1:1) and HCT-PYR (2:1) ground mixtures exhibit melting onsets at 445.9 and 453 K, respectively, followed by broad endotherms around 503-553 K that correspond to the decomposition of HCT (Fig. 6). The single endotherm (prior to decomposition) of the ground mixture is again indicative of homogeneous co-crystal phases in each case. A 1:1 ground mixture of HCT and PYR also provided co-crystals, but with an excess of pyrogallol, which gives rise to an additional endotherm at 399-402 K, prior to melting of the co-crystal (see the supporting information ). Thus, the single endotherm observed for the DSC of the 2:1 mixture confirms the stoichiometry of the 2:1 HCT-PYR co-crystal.
| || Figure 6 |
DSC endotherms of HCT, HCT-RES and HCT-PYR co-crystals. The melting points of RES and PYR are 383-385 and 404-407 K, respectively.
The co-crystals were further characterized by FT-IR spectroscopy (Mukherjee et al., 2013). The replacement of the API sulfonamide catemer hydrogen bond by heterosynthons in the co-crystals is reflected in the IR spectra especially in the SO2 asymmetric/symmetric (1370, 1320/1150 cm-1) and NH stretching frequency (3360-3450 cm-1) regions (Fig. S3 ). The SO2 asymmetric region (1319-1380 cm-1) is quite broad in the FT-IR spectra of HCT, whereas sharp distinct bands appear in the co-crystals. Similar SO2 bands appear in the isostructural co-crystals with nicotinic acid and nicotinamide. Nicotinic acid is present as a zwitterionic form in the HCT-NIC co-crystal confirmed by the hypsochromic shift from 1707 cm-1 (neutral) to 1645 cm-1 (zwitterionic) in the IR spectra. Resorcinol and pyrogallol co-crystals were also confirmed by comparison of the vibrational frequencies. The HCT-RES co-crystal exhibits asymmetric SO2 stretching frequencies at 1320 (primary), along with 1364 cm-1 (secondary), which indicates the possible presence of a catemer chain motif of the sulfonamide, similar to that in the HCT-PABA co-crystal. The HCT-PYR co-crystal shows identical SO2 stretching frequencies as HCT-NIC, which indicates that sulfonamide dimer or catemer synthons may not be present. The FT-IR vibrational frequency comparisons are summarized in Table 3. Significant NH vibrational frequency changes are also observed due to the synthon variations and the presence of amide/amine co-formers in the co-crystals.
Hydrochlorothiazide is a poorly aqueous soluble drug (0.7 g L-1, 298 K) and because of its poor solubility (defined as the concentration in a solution that is in equilibrium with an excess amount of the solid), the bioavailability is also less (only 65%). To improve the aqueous solubility of HCT, solid dispersion methods have been reported, but only a limited improvement (2-3 times) was observed. We have attempted here to tune the solubility of HCT by the co-crystal approach using biologically safe co-formers (Aakeröy et al., 2009). However, solution-mediated phase transformation of co-crystals greatly reduces the apparent solubility and dissolution rate (Childs et al., 2004; Remenar et al., 2007). Dissolution rate (defined as the rate at which a substance dissolves and reaches equilibrium) is important in the case of phase transformations from one form to another. Apparent solubility, which is the time-dependent solubility of a molecule, is also prone to phase transformation. The aqueous solubility of HCT is 841 mg L-1 in distilled water at 310 K. The apparent solubilities of the co-crystals were measured at 1, 4 and 24 h to examine whether any phase transformation occurs during the solubility experiments. The measured solubility is significant if there is no phase transformation before the equilibrium between solvent and solute is reached. The HCT-NIC co-crystal exhibited the lowest apparent solubility, being 4 times less than HCT itself at 4 h. However, the equilibrium solubility of the co-crystal increased by 1.7 fold at 24 h (Table 4), and the co-crystal was stable even after 24 h slurry experiments in water, according to PXRD comparison (supporting information ). Comparatively, the HCT-NCT co-crystal exhibited a 1.5-fold increase in solubility compared with pure HCT at 4 h, but only 33% of the co-crystal remained intact because of the solvent-mediated phase transformation to pure HCT. The HCT-NCT co-crystals started to dissociate within 1 h of starting the dissolution experiment. The HCT-RES and HCT-PYR co-crystals were 2-3 times more soluble than HCT, but also dissociated to pure HCT within 1 h because of the significant solubility difference between HCT and the phenols. The HCT-PABA co-crystal showed the highest apparent solubility (a sixfold increase compared with pure HCT) within 4 h and was stable up to 24 h in the aqueous medium (Fig. 7). There was not much solubility change for the HCT-PABA co-crystal from 1 to 24 h, indicating that it had reached its saturation solubility at 1 h; this is quite significant for drug formulation.
| || Figure 7 |
PXRD comparisons of HCT-PABA co-crystals after 4 and 24 h slurry experiments in water indicating its superior stability.
There is an inverse correlation of calculated density and apparent solubility (4 h) for the co-crystals as follows: density (g cm-3): HCT-NIC (1.78) > HCT-NCT (1.71) > HCT-PABA (1.58) and apparent solubility (mg L-1): HCT-NIC (190) < HCT-NCT (1269) < HCT-PAB (5169). This indicates that HCT-NIC is the most stable, densely packed co-crystal, and hence exhibits relatively low solubility. Further, the solubility order of the co-former is NCT (500 g L-1) > NIC (18 g L-1) > PABA (6 g L-1). It is expected that NCT should dissolve quickly because of the high solubility difference between HCT and NCT, which ultimately results in the faster dissociation of the HCT-NCT co-crystal. By contrast, NIC and PABA form a congruent system (Rodríguez-Hornedo et al., 2006; Friscic et al., 2009) with HCT and are stable up to 24 h in the aqueous medium. Except for the HCT-NCT co-crystal, there is an inverse correlation between the melting points of the co-crystals and their aqueous solubilities (Sanphui et al., 2011; Table 4). HCT-NIC exhibits the highest m.p. and hence lowest solubility (Fig. 8). Further, HCT-PABA exhibits the lowest m.p. and also shows the highest solubility. In the HCT-PABA co-crystal, there is a clear separation of hydrophobic and hydrophilic regions (see Fig. 5c), which may help HCT to interact with the solvent and thereby improve the solubility of the co-crystal. Comparatively, the HCT-RES and HCT-PYR co-crystals exhibit intermediate m.p. and also moderate solubility. Among all of the co-crystals discussed, the HCT-PABA co-crystal is the best solid form in terms of high solubility and stability.
| || Figure 8 |
Inverse correlation of apparent solubility (4 h) and m.p. for the HCT co-crystals. The HCT-NCT co-crystal does not follow the general trend.
The poor solubility of the diuretic drug hydrochlorothiazide (HCT) was addressed by preparing co-crystals. The structural aspects of the HCT co-crystals with nicotinic acid (NIC), nicotinamide (NCT), succinamide (SAM) and p-aminobenzoic acid (PABA) have been discussed. The catemer synthon formed by the primary sulfonamide groups in the stable polymorph of HCT is replaced by HCTco-former interactions in the isomorphous co-crystals with NIC and NCT. The co-crystals with SAM and PABA showed primary sulfonamide dimer and catemer synthons for HCT. Co-crystals with resorcinol and pyrogallol were confirmed by PXRD, IR and DSC, but their structures have not been established to date. The aqueous solubility of the co-crystals suggests that the co-formers with relatively low solubility, PABA, SAM and NIC, form stable co-crystals in the aqueous medium. Among the stable binary systems, HCT-PABA co-crystals showed a sixfold solubility increase compared with pure HCT. The HCT-NCT co-crystal dissociated within 1 h of dissolution because HCT and the co-former form an incongruent system. Comparatively, the HCT-RES and HCT-PYR co-crystals improved the solubility by 2-3 times, but transformed to pure HCT within 1 h. In general, there is an inverse correlation observed between the melting points of the co-crystals and their aqueous solubilities.
PS thanks Ranbaxy Pvt. Ltd for a Fellowship. LR thanks the DST for a Young Scientist Fellowship. Both PS and LR thank Professor Gautam R. Desiraju for providing the laboratory facilities and valuable discussions and suggestions. We thank the Rigaku Corporation, Tokyo, for their support through a generous loan of a RigakuMercury375R/M CCD (XtaLAB mini) diffractometer.
Aakeröy, C. B., Forbes, S. & Desper, J. (2009). J. Am. Chem. Soc. 131, 17048-17049.
Allen, F. H. (2002). Acta Cryst. B58, 380-388.
Allen, F. H. & Bruno, I. J. (2010). Acta Cryst. B66, 380-386.
Almarsson, Ö., Hickey, M. B., Peterson, M., Zaworotko, M. J., Moulton, M. & Rodriguez-Hornedo, N. (2007). US Patent 2007/0059356 A.
Amidon, G. L., Lennernäs, H., Shah, V. P. & Crison, J. R. (1995). Pharmacol. Res. 12, 413-420.
Babu, N. J., Sanphui, P. & Nangia, A. (2012). Chem. Asian J. 7, 2274-2285.
Barbour, L. J. (2001). J. Supramol. Chem. 1, 189-191.
Caira, M. R., Nassimbeni, L. R. & Wildervanck, A. F. (1995). J. Chem. Soc. Perkin Trans. 2, pp. 2213-2216.
Childs, S. L., Chyall, L. J., Dunlap, J. T., Smolenskaya, V. N., Stahly, B. C. & Stahly, G. P. (2004). J. Am. Chem. Soc. 126, 13335-13342.
Clarke, H. D., Arora, K. K., Bass, H., Kavuru, P., Ong, T. T., Pujari, T., Wojtas, L. & Zaworotko, M. J. (2010). Cryst. Growth Des. 10, 2152-2167.
Desiraju, G. R. (1995). Angew. Chem. Int. Ed. 34, 2311-2327.
Dupont, L. & Dideberg, O. (1972). Acta Cryst. B28, 2340-2347.
Flack, H. D. (1983). Acta Cryst. A39, 876-881.
Florence, A. J., Johnston, A. & Shankland, K. (2005). Acta Cryst. E61, o2974-o2977.
Friscic, T., Childs, S. L., Rizvi, S. A. A. & Jones, W. (2009). CrystEngComm, 11, 418-426.
Glomme, A., März, J. & Dressman, J. B. (2005). J. Pharm. Sci. 94, 1-16.
Good, D. J. & Rodríguez-Hornedo, N. (2009). Cryst. Growth Des. 9, 2252-2264.
Goud, N. R., Babu, N. J. & Nangia, A. (2011). Cryst. Growth Des. 11, 1930-1939.
Hanna, M., Shan, N. & Cheney, M. L. (2009). US Patent 2009/0176787 A1.
Johnston, A., Florence, A. J. & Kennedy, A. R. (2006a). Acta Cryst. E62, o2926-o2928.
Johnston, A., Florence, A. J. & Kennedy, A. R. (2006b). Acta Cryst. E62, o1730-o1732.
Kavuru, P., Aboarayes, D., Arora, K. K., Clarke, H. D., Kennedy, A., Marshall, L., Ong, T. T., Perman, J., Pujari, T., Wojtas, L. & Zaworotko, M. J. (2010). Cryst. Growth Des. 10, 3568-3584.
Kim, B. H. & Kim, J. K. (1984). Arch. Pharm. Res. 7, 47-52.
Kraus, W. & Nolze, G. (1996). J. Appl. Cryst. 29, 301-303.
Leech, C. K., Fabbiani, F. P. A., Shankland, K., David, W. I. F. & Ibberson, R. M. (2008). Acta Cryst. B64, 101-107.
McNamara, D. P., Childs, S. L., Giordano, J., Iarriccio, A., Cassidy, J., Shet, M. S., Mannion, R., O'Donnell, E. & Park, A. (2006). Pharm. Res. 23, 1888-1897.
Mukherjee, A., Tothadi, S., Chakraborty, S., Ganguly, S. & Desiraju, G. R. (2013). CrystEngComm, 15, 4640-4654.
Padmapriya, S., Rajendran, N. N. & Ravichandiran, V. (2011). Int. J. PharmTech Res. 3, 1509-1514.
Panalytical (2012). HighScore Plus. Panalytical, BV Almelo, The Netherlands.
Patel, R. B., Patel, U. R., Rogge, M. C., Shah, V. P., Prasad, V. K., Selen, A. & Welling, P. G. (1984). J. Pharm. Sci. 73, 359-361.
Pedireddi, V. R., Reddy, D. S., Goud, B. S., Craig, D. C., Rae, A. D. & Desiraju, G. R. (1994). J. Chem. Soc. Perkin Trans. 2, pp. 2353-2360.
Remenar, J. F., Morissette, S. L., Peterson, M. L., Moulton, B., MacPhee, J. M., Guzmán, H. R. & Almarsson, O. (2003). J. Am. Chem. Soc. 125, 8456-8457.
Remenar, J. F., Peterson, M. L., Stephens, P. W., Zhang, Z., Zimenkov, Y. & Hickey, M. B. (2007). Mol. Pharm. 4, 386-400.
Rigaku (2009). CrystalClear. Rigaku Corporation, Tokyo, Japan.
Rodríguez-Hornedo, N., Nehm, S. J., Seefeldt, K. F., Pagan-Torres, Y. & Falkiewicz, C. J. (2006). Mol. Pharm. 3, 362-367.
Sanphui, P., Goud, N. R., Khandavilli, U. B. R. & Nangia, A. (2011). Cryst. Growth Des. 11, 4135-4145.
Sarma, B., Reddy, L. S. & Nangia, A. (2008). Cryst. Growth Des. 8, 4546-4552.
Shattock, T. R., Arora, K. K., Vishweshwar, P. & Zaworotko, M. J. (2008). Cryst. Growth Des. 8, 4533-4545.
Sheldrick, G. M. (2008a). Acta Cryst. A64, 112-122.
Sheldrick, G. M. (2008b). SADABS. University of Göttingen, Germany.
Smith, A. J., Kavuru, P., Wojtas, L., Zaworotko, M. J. & Shytle, R. D. (2011). Mol. Pharm. 8, 1867-1876.
Smolka, T., Boese, R. & Sustmann, R. (1999). Struct. Chem. 10, 429-431.
Spek, A. L. (2009). Acta Cryst. D65, 148-155.
Stilinovic, V. & Kaitner, B. (2011). Cryst. Growth Des. 11, 4110-4119.
Trivedi, R. V., Admane, P. S., Taksande, J. B., Mahore, J. G. & Umekar, M. J. (2011). Der Pharmacia Lettre, 3, 8-17.
Ueto, T., Takata, N., Muroyama, N., Nedu, A., Sasaki, A., Tanida, S. & Terada, K. (2012). Cryst. Growth Des. 12, 485-494.