I. INTRODUCTION
Many pharmaceutical compounds in daily use have a well defined molecular formula but were never completely characterized crystallographically (or results stay unpublished), so the question of a possible polymorphism has no answer, opening some place for industrial competition. A typical example is bethanechol chloride, also called carbamyl-β-methylcholine chloride (Major and Bonnett, Reference Major and Bonnett1943), a quaternary amine which is a cholinergic agent. This compound is a synthetic ester structurally and pharmacologically related to acetylcholine. It acts principally as stimulant of the parasympathetic nervous system. It has been used for a long time in urology and gastroenterology, and more recently it has found new applications in the treatment of Alzheimer’s disease (Cummings and Kaufer, Reference Cummings and Kaufer1996) and cerebral palsy (Carter, Reference Carter2008). The PDF contains two entries (00-043-1748 and 00-046-1964), both are unindexed and show some discrepancies. The aim of the present work is to provide more in depth results, in spite of the absence of single crystal, by solving the structure from the powder data by using methodologies which continuously have demonstrated their efficiency during repeated blind tests (Le Bail et al., Reference Le Bail, Cranswick, Adil, Altomare, Avdeev, Cerny, Cuocci, Giacovazzo, Halasz, Lapidus, Louwen, Moliterni, Palatinus, Rizzi, Schilder, Stephens, Stone and van Mechelen2009).
II. STRUCTURE DETERMINATION AND RIETVELD REFINEMENTS
Various samples, as received from the SERATEC company (www.serateclab.com), showed powder diffraction patterns with large discrepancies at first sight (Figure 1), but they all evolved with time (weeks) into one unique stable state. As the bethanechol chloride is said to be hygroscopic, this suggested that the final formula could be, in fact, C7H17ClN2O2.nH2O. However, no mass loss was detected on the thermogravimetric analysis before decomposition, leading to a hypothesis of polymorphism, one of the poly-morphs being possibly unstable at ambient conditions, transforming slowly into the final stable form. The powder pattern of this latter form could be indexed in a monoclinic cell by the MCMAILLE software (Le Bail, Reference Le Bail2004), the figures of merit (FoMs) for the pattern selected for this study being M 20=84.4 (de Wolff, Reference de Wolff1968), F 20=192.0 (0.0035, 30) (Smith and Snyder, Reference Smith and Snyder1979), and McM 20=198.0 (Le Bail, Reference Le Bail, Clearfield, Reibenspies and Bhuvanesh2008). Indeed, a series of ten data sets were proposed to be used as indexing benchmarks for the comparison of computer programs during an indexing round robin (Bergmann et al., Reference Bergmann, Le Bail, Shirley and Zlokazov2004), including a synchrotron pattern which was the easiest to index, leading to FoMs: M 20=168.0, F 20=888.2 (0.0008, 30), and McM 20=540.0. This synchrotron pattern, limited to 2θ max=28° at λ=0.6995 Å, is not used at the final Rietveld stage in this work because the corresponding sample either was not completely transformed into the stable polymorph or more probably suffered from its airplane low-pressure and low-temperature trip [such effects were observed for a vanadium hypophosphite (Le Bail et al., Reference Le Bail, Marcos and Amorós1994)], and the line profiles show anisotropic broadening, contrarily to the sample selected here, prepared in conditions so as to obtain directly the stable polymorph.
Figure 1. (Color online) Three representative bethanechol chloride samples as received from the SERATEC (www.serateclab.com). Samples b and c finally evolved slowly, without mass gain nor loss, into a powder pattern identical to that of sample a which could be indexed. The transformation has already started for sample c which contains the largest proportion of probably another still undefined and unstable polymorph.
TABLE I. Experimental and Rietveld refinement details for bethanechol chloride.
The cell volume suggested Z=4. Using the Le Bail fitting (Le Bail, Reference Le Bail2005) for evaluating the reliability of the indexing and for intensity extraction, the P21/n space group was determined unambiguously. The first structure solution attempts were done in direct space from the Cu Kα conventional laboratory data using the ESPOIR software (Le Bail, Reference Le Bail2001) moving in the cell by a Monte Carlo process, a NC4 tetrahedron, a CO2N triangle, and the chlorine and two carbon atoms at random (for a total of 21 degrees of freedom), expecting to complete the bethanechol molecule. This led to R=16% on the first 200 extracted hkl. Refinements were undertaken by the Reference RietveldRietveld (1969)FULLPROF software (Rodríguez-Carvajal, Reference Rodríguez-Carvajal1993). Some ambiguity for the NH2 position in the CO2NH2 group was removed by consideration on the Cl-N interatomic distances, the Cl atom being expected to point towards the direction of the hydrogen atoms of the NH2 group. At this stage (RP=9.05(, RWP=8.63, RB=5.03, and RF=7.09%)), the 17 independent hydrogen atoms were suggested by the OPEN BABEL software (Guha et al., Reference Guha, Howard, Hutchison, Murray-Rust, Rzepa, Steinbeck, Wegner and Willighagen2006) and their inclusion with 31 soft restraints on the C-H, N-H, and H-H distances in the refinement allowed to lower all the R factors by approximately 1.5% to the final values listed in Table I (see the Rietveld fit in Figure 2). Atomic coordinates, selected geometric parameters, and X-ray diffraction data are in Tables II-IV, respectively. The structure could also be solved later, during the indexing round robin, by J. van de Streek using the DASH software (David et al. 2006) applied to the synchrotron data.
Figure 2. (Color online) Rietveld plot for C7H17ClN2O2.
TABLE II. Fractional atomic coordinates and isotropic displacement parameters.
TABLE III. Selected geometric parameters (Å, deg).
a Symmetry code: x+1/2, −y+1/2, z+1/2.
b Symmetry code: −x+3/2, y+1/2, −z+1/2.
c Symmetry code: x+1/2, −y+1/2, z−1/2.
d Symmetry code: −x+1, −y+1, −z.
III. DISCUSSION
Bethanechol (Figure 3) is a quaternary ammonium compound, a derivative of acetylcholine. The positively charged quaternary ammonium group imparts a very high polarity to the molecule based on its permanent positive charge. The Cl−ions are in a cage of hydrogen atoms belonging to four neighbouring bethanechol molecules (Figure 4). Seven of the Cl-H distances are smaller than 3 Å, the shortest being with H17 from the NH2 groups (Table III). Indeed, the shortest distances with a non-H atom are with the N2 atom of that group from two different molecules [Cl-N2=3.277(11) and 3.337(8) Å]. But the majority of the hydrogen atoms around the chlorine atoms belong to the ammonium group (H2, H4, H6, and H7), in accordance with the location of the positive charge of the molecule.
One should not expect much accuracy from that Rietveld refinement, no more than from a single-crystal study in the 1960s to 1970s (and the refinement from the synchrotron data leads to similar atomic coordinates and estimated standard deviations). Alerts on deviating methyl C1, C2, and C6 and low precision on C-C bonds were delivered when using the PLATON software (Spek, Reference Spek2003) for validation. Moreover, the refined thermal parameters are unusually high. The NC4, CO2N, and C3O groups are tetrahedral, planar, and triangular pyramidal, respectively, as expected. However, stronger restraints or DFT calculations would be necessary for obtaining more acceptable bonds and angle values. In any case, the results shown here are the best obtainable from the powder data only when some parameters are allowed to be refined (at least all the non-H atoms) freely during the Rietveld refinement. The CIF was deposited at the Crystallography Open Database (www.crystallography.net) (Gražulis et al., Reference Gražulis, Chateigner, Downs, Yokochi, Quirós, Lutterotti, Manakova, Butkus, Moeck and Le Bail2009).
When attempting to index the powder pattern of sample c (Figure 1) after removing the peaks clearly belonging to the stable polymorph, a probable orthorhombic cell with parameters a=8.857(10) Å, b=16.466(13) Å, c=7.053(6) Å, V=1028.6 Å3, M 20=15.4, F 20=24.2 (0.0150, 55), and McM 20=39.5 was obtained (Table V). This suggests that both structures could be quite similar, which would explain the absence of a clear event on the differential scanning calorimetry before the decomposition at 227 °C for all samples. In the absence of details about the way of synthesis from the SERATEC Co., it may be suggested that the higher symmetry second polymorph would be, in fact, a slightly higher temperature form which would be the first appearing when the preparation conditions are at a temperature higher than some value, above ambient conditions. More work is needed in order to confirm that hypothesis.
IV. CONCLUSION
In the absence of a suitable single crystal, modern powder diffraction methodologies have the power sometimes to reveal crystal structures, including those of pharmaceutical compounds in use for a long time but still uncharacterized crystallographically. Even if all problems are not solved here, such a structure determination of a defined polymorph gives the chance to a pharmaceutical company to provide fine details about the exact quality of their product, a quality of purity required by the international market where national agencies such as the FDA have to be convinced before delivering import permission. Moreover, every polymorph represents a drug candidate which in turn may increase profits. Powder diffraction may therefore be of great help in such important subjects.
TABLE IV. X-ray diffraction data of bethanechol chloride (Cu Kα1).
AKNOWLEDGMENTS
Thanks are due to P. W. Stephens for the recording of a synchrotron powder pattern at the National Synchrotron Light Source at Brookhaven National Laboratory, which is supported by the Division of Materials Sciences and Division of Chemical Sciences, U.S. Department of Energy. The SUNY X3 beamline at NSLS is supported by the Division of Basic Energy Sciences of the U.S. Department of Energy under Grant No. DE-FG02-86ER45231.
Figure 3. (Color online) Bethanechol molecule with atom numbering.
Figure 4. (Color online) Projection of the structure in a direction close to the a axis, showing the bethanechol molecules stacking. The seven shorter Cl-H bonds interlinking four molecules are marked for one Cl atom only for the sake of clarity.
TABLE V. The 20 peaks used for indexing the sample c routine powder pattern shown on Figure 1, excluding peaks obviously corresponding to the sample a (Cu Kα1).
