Crystal structure of a 2:1 co-crystal of meloxicam with acetylendicarboxylic acid

The crystal structure of a new 2:1 co-crystal of meloxicam and acetylendicarboxylic acid is reported


Chemical context
In recent years, crystal engineering has focused on finding new crystalline forms based on the multi-component crystallization of an active pharmaceutical ingredient (API) with a biologically inactive compound. These complexes are ultimately aimed at being employed in the pharmaceutical industry as tablets, suspensions, powders and any other solid forms for oral administration (Shakhtshneider et al., 2007a,b;Crowley & Zografi, 2002;Hancock & Parks, 2000;Shakhtshneider & Boldyrev, 1993;Willart & Descamps, 2008;Shakhtshneider et al., 2011;Stephenson et al., 2011). Coformers are typically chosen from among the dicarboxylic acids due to their favourable molecular shape and the presence of functional groups capable of forming multiple hydrogen bonds, combined with their affordability and availability. Meloxicam (MXM), 4-hydroxy-2-methyl-N-(5-methyl-2-thiazolyl)-2H-1,2-benzothiazine-3-carboxamide-1,1-dioxide, belongs to the oxicam family of APIs and is commonly used in the treatment of rheumatoid arthritis Weyna et al. 2012). MXM is known to cocrystallize with numerous aliphatic and aromatic dicarboxylic acids under various conditions (temperature, pressure, solvents). In particular, MXM is known to co-crystallize with dicarboxylic acids of C-C bond order 1 (succinic acid) and 2 (fumaric and maleic acids). The aim of this study was to obtain a co-crystal of MXM with a dicarboxylic acid of bond order 3: acetylenedicarboxylic acid (ACA).

Structural commentary
The crystal structure of MXM:ACA 2:1 is triclinic with an asymmetric unit that contains one MXM molecule and half of an ACA molecule. The formula unit is generated by an inversion centre which is located at the midpoint of the triple bond of the ACA molecule (Fig. 1). The two stereoisomers of MXM, which differ with respect to the nitrogen atom of the sulfonamide group, are related by an inversion centre in the crystal structure. The dihedral angles between the mean planes of the thiazole and benzene rings of MXM form an almost planar arrangement in terms of the following torsion angles: S2-C11-N2-H2 = À174.0 , S2-C11-N2-C10 = 6.0 (3) , H2-N2-C10-O4 = 176.5 , O4-C10-C8-C7 = 10.0 (3) , C8-C7-O3-H3 = À2.2 . The presence of an intramolecular O-HÁ Á ÁO hydrogen bond between the carbonyl and hydroxy groups belonging to MXM may account for the near planarity and the trans position of the N2-H2 group with respect to the carbonyl group C10-O4. The S1/ N1/C1/C6/C7/C8 ring is non-planar because of the presence of the sulfonamide group with nitrogen atom N1 in sp 3 hybridization, with angles S1-N1-C8 = 112.79 (12) , S1-N1-C9 = 117.11 (14) and C9-N1-C8 = 115.41 (17) (bond-angle sum = 345.3 ). The overall conformation of this ring is half-chair with atoms S1 and N1 being the out-of-plane atoms.

Supramolecular features
In the crystal, the components of the structure are linked by N-HÁ Á ÁO and O-HÁ Á ÁN hydrogen bonds between MXM and ACA, in addition to a long O-HÁ Á ÁO interaction, forming chains along [011] which incorporates both R 2 2 (8) and R 2 2 (12) rings. Similar structural motifs have been documented for other MXM co-crystals and in other crystal structures including pure MXM, MXM co-crystals and MXM salts. The structure-forming unit includes two molecules of MXM connected through a dicarboxylic acid molecule acting as a bridge, similar to what has been reported for other MXM cocrystals . Intra-and intermolecular hydrogen bonds are shown in Fig. 3 and their geometrical parameters are summarized in Table 1. The centroid-tocentroid distance between symmetry-related benzene and thiazole rings is 3.7383 (12) Å . These connect the chains into a three-dimensional network. Meloxicam (MXM) and acetylenedicarboxylic acid (ACA) molecules of the 2:1 co-crystal, showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 50% probability level. Only half of the ACA molecule belongs to the asymmetric unit, as the molecule lies across an inversion centre. Table 1 Geometrical parameters (Å , ) for the O-HÁ Á ÁO (1), O-HÁ Á ÁN (2) and N-HÁ Á ÁO (3) interactions in the MXM:ACA 2:1 co-crystal (see also Fig. 3).

Synthesis and crystallization
MXM was purchased from Sigma Aldrich Co Ltd and acetone from Reaktiv. ACA was synthesized through a two-step process from fumaric acid. Fumaric acid was brominated in boiling water (Rhinesmith, 1938) and the resulting 2,3-dibromosuccinic acid was refluxed in potassium hydroxide methanolic solution. ACA was precipitated by adding a concentrated sulfuric acid solution and dried in vacuo (Rhinesmith, 1938). The purity of ACA and the absence of its monohydrate were checked by comparing its experimental powder X-ray diffraction powder (XRPD) pattern with the calculated XRPD patterns of ACA and ACA monohydrate (see S1 in Supporting information). Two polycrystalline samples were obtained by dry and slurry (with acetone) grinding of 1:2 molar mixture of reactants (0.035g, 0.1mmol MXM; 0.023g, 0.2mmol ACA). The 2:1 ratio would correspond to the target stoichiometry and is usually used for obtaining other MXM co-crystals with aliphatic dicarboxylic acids Weyna et al. 2012). However, to obtain MXM-ACA 2:1 co-crystals we used a 1:2 MXM:ACA ratio because ACA is highly hygroscopic and converts to its monohydrate form on grinding, not participating then in the co-crystallization. Acetone was used for slurry grinding because it completely dissolves the two starting components Weyna et al. 2012). All powder samples were characterized by XRPD using a Stoe Stadi-MP diffractometer with Cu K 1 radiation ( = 1.54060 Å ) at operating potential of 40 kV and electric current of 40 mA, and a Mythen 1K detector. All data were processed using WinXPOW (Stoe & Cie, 1999). Powder diffraction patterns for the samples obtained by grinding and slurry grinding were similar, confirming the possibility to obtain the same product both in the presence and in the absence of a specially added solvent (see S2 in Supporting information); the XRPD patterns of the co-crystal sample were compared with the patterns of the starting reactants, MXM and ACA (see S3 in Supporting information) to prove that a new phase (or a mixture of new phases) had been formed. The ground powder samples were subsequently dissolved in acetone and single crystals were obtained by slow evaporation. Selected crystals were investigated using singlecrystal X-ray diffraction.
1858 Tantardini  The molecular packing in the crystal structures of (a) pure MXM and its co-crystals (

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 2. All H atoms were initially located in a difference Fourier map. The positions of all H atoms were subsequently optimized geometrically and refined using a riding model, with the following assumptions and restraints: N-H = 0.86 Å and U iso (H)=1.2U eq (N) for -N(H)-group, C-H = 0.93 Å and U iso (H) = 1.2Ueq(C) for all C-H groups, O-H = 0.82 Å and U iso (H) = 1.5U eq (O) for all OH groups, C-H = 0.96 Å and U iso (H) = 1.5U eq (C) for CH 3 groups.
For single crystals of MXM:ACA (2:1), two data sets were collected. The first dataset was obtained from a crystal containing four domains, and the second from a single crystal. Unfortunately, the single crystal was very small and at d hkl ! 0.80 Å , R int was 10.2% and F 2 /(F 2 ) was 3.6. This was significantly worse than the data from the crystal that contained four domains [for the largest domain at d hkl ! 0.80 Å , R int was 2.50% and F 2 /(F 2 ) was 28.3]. Data obtained from the crystal that contained four domains were processed in three different ways: (1) taking into account the reflections from the largest domain only (one orientation matrix and 74.3% of all reflections); (2) processing the diffraction data as from multiple crystals (four different orientation matrices) using the hklf5file; (3) processing the diffraction data as from multiple crystals (4 different orientation matrixes) using the. hklf4-file from the largest domain (74.3% of all reflections). The first and the third processing methods gave approximately the same results, while the first methodology yielded the best results: R int = 0.025. This method was therefore chosen for the final structure solution and refinement.
The powder diffraction patterns calculated based on the X-ray single crystal diffraction data were compared with the experimental powder diffraction pattern measured for the sample obtained on grinding, to show that the latter contained a mixture of the MXM:ACA 2:1 co-crystal with some other phases, different from ACA, MXM, or ACA hydrate (see S4 in Supporting information).   (Agilent, 2013); program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: OLEX2 (Dolomanov et al., 2009) and Mercury (Macrae et al., 2008; software used to prepare material for publication: OLEX2 (Dolomanov et al., 2009) and publCIF (Westrip, 2010).

Special details
Experimental. Suitable-quality crystals were selected using polarised light under the microscope and mounted by means of MiTiGenMicroGrippers using MiTiGen LV Cryo Oil (LVCO-1) onto an Agilent Xcalibur (Ruby, Gemini Ultra) diffractometer. Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.