This work is supported by a U.S.-Israel Binational Science Foundation Grant (no. 2004118). We are grateful to Dr. Dmitry Mogilyanski and Shoshana Lach for the XRPD measurements on the powders obtained by traditional methods (Figure 3) and for technical assistance.
Designing a Cocrystal of γ-Amino Butyric Acid†
Article first published online: 9 NOV 2006
Copyright © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Angewandte Chemie International Edition
Volume 45, Issue 47, pages 7966–7969, December 4, 2006
How to Cite
Wenger, M. and Bernstein, J. (2006), Designing a Cocrystal of γ-Amino Butyric Acid. Angew. Chem. Int. Ed., 45: 7966–7969. doi: 10.1002/anie.200603241
- Issue published online: 27 NOV 2006
- Article first published online: 9 NOV 2006
- Manuscript Revised: 9 SEP 2006
- Manuscript Received: 9 AUG 2006
- U.S.-Israel Binational Science Foundation. Grant Number: 2004118
- crystal engineering;
- hydrogen bonds;
- pH control;
Organic salts are generally the preferred crystal form of active pharmaceutical ingredients because of their higher solubilities and/or increased degree of crystallinity. The potential number of suitable organic salts is limited to the counterions specified by the food and drug association (FDA) as GRAS (generally regarded as safe). A recently suggested alternative to the common inorganic salts is the preparation of cocrystals1 or organic salts, some of which have been shown to improve therapeutic utility while reducing side effects.2–4
Candidates for cocrystal formation with a particular structural motif for achieving desired physical properties5, 6 are chosen on the basis of their ability to utilize known intermolecular interactions in forming those crystals. Because hydrogen bonds are among the strongest and most preferentially directional intermolecular interactions, they have been the leading candidates for the interactions to be utilized in forming cocrystals,7–10 although other interactions have also been suggested.11a,b
Particular preference in the development of the strategy for cocrystal formation utilizing hydrogen bonds has traditionally been based on the tendency of carboxylic acids and amides12, 13 to form the homo- or heterointermolecular synthons14–16 described in Etter's graph-set notation17, 18 as an (8) ring (which interprets as a ring (R) motif containing eight atoms with two hydrogen-bond donors (subscript) and two hydrogen-bond acceptors (superscript); Scheme 1). 1However, other hydrogen-bonding motifs such as (4) (Scheme 1), 1have also been investigated.19, 20 Herein we demonstrate the potential for utilizing another, as yet little recognized and not yet utilized, hydrogen-bonded synthon (8) (Scheme 1).1
A survey of the Cambridge Structural Database (CSD version 1.8, 2006) has identified over 12 000 instances of this synthon, virtually all of which involve four individual and nonconnected (but not necessarily chemically different) moieties in the solid state. Many of these cases involve two chemically different moieties, often resulting in a pattern that is crystallographically centrosymmetric or pseudocentrosymmetric. Perhaps the most remarkable feature about this synthon in terms of cocrystal formation is that it potentially involves the intermolecular recognition and supramolecular synthesis of four different molecules.
In the CSD, there are 918 hits for structures utilizing the (8) synthon (Scheme 1) 1in which the donor is the amino group (NH2) and the acceptor is the carbonyl group (CO). These are broken down into specific functional groups like (CO)O−, N(CO), N(CO)N, (CO)OH, C(CO)C, C(CO)H, -NH3+, CNH2, NH4+, and NNH2. This survey clearly indicates that the highest incidence of participation in the (8) synthon is for carboxylate acceptors and ammonium cation donors.
In our cocrystallization attempts to achieve this synthon, we chose to use amino acids, starting with γ-amino butyric acid (GABA). As the CSD search indicated a preference for the ionized species -NH3+ and COO− in the formation of the desired (8), we attempted crystallization in the pH range for which the amino group of GABA will be in the form of -NH3+ and the carboxylic group of the oxalic acid and benzoic acid will be in the form of -CO2−. The molar fractions21 for the ionized species in different pH ranges can be calculated from the pKa values of the acid and base. Figure 1 1represents the calculated molar fraction for GABA and oxalic acid. For GABA, pKa=10.43, pKb=9.77, and the isoelectric point is pI=7.33. For oxalic acid, pKa1=1.23 and pKa2=4.19. It is clear from Figure 1 1that the crystallization for a 1:1 GABA/oxalic acid cocrystal should be carried out in the pH range 0–4.19 and crystallization for 2:1 cocrystal ratio should be carried out in the pH range 4.19–10.43. Our initial experiments were carried out at pH 5.
The same calculations and molar fraction diagrams were prepared for the crystallization of GABA and benzoic acid. The pKa value for benzoic acid is 4.19, which is the same as GABA. The crystallization attempts were carried out at three different pH values: 4, 6, and 12. At pH 4, the two species of benzoic acid and the two species of first dissociation of GABA are in equilibrium (see the Supporting Information). Experiments on oxalic acid and GABA yielded a 1:2 cocrystal of GABA/oxalic acid. As expected from Figure 1, 1GABA is protonated and the oxalic acid is a dianion. Charge neutrality therefore requires a 1:2 ratio of GABA and oxalic acid.
The crystal structure shown in Figure 2 2clearly reveals three of the four hydrogen bonds required to complete the (8) synthon. The donor–acceptor (D⋅⋅⋅A) distances, a=1.913 Å, b=2.517 Å, and c=2.063 Å, are in the range of normal hydrogen bonds. The distance d (3.242 Å) is considerably long for an NH⋅⋅⋅O hydrogen bond, but the incipient pattern of the synthon (8) still appears to be utilized in this structure.
Additional crystallizations were carried out at pH 0, 3, and 12. The results of X-ray powder diffraction (XRPD) studies, when compared with those of the starting materials (Figure 3 A), 3indicate the presence of new crystalline phases (polymorph, hydrate, or differing stoichiometry) of the resulting powders. However, crystals suitable for structure analysis have not yet been obtained. These new phases can result from the different species coexisting at each pH value. From the crystal structure obtained at pH 5 (Figure 2), 2the acceptor that participates in the hydrogen bonds that are labeled a–d is the carboxylate group of GABA. This species can exist only above pH 4 (Figure 1). 1This fact can account for the different phases at the pH values of 0 and 3. One possibility for the appearance of a different phase obtained at pH 12 is that the donor for all the hydrogen bonds that participates in the hydrogen bonding is the amino group of GABA, which is protonated below pH 10 (see Figure 1); 1above this pH value (e.g. pH 12), the amino group will not be protonated and will likely participate in different hydrogen-bonding patterns.
A cocrystal of 1:1 benzoic acid/GABA was obtained from aqueous solution at pH 4. The small colorless crystals appeared a few minutes after adding water to a mixture of the solids. In the structure, the GABA is protonated and the benzoic acid is in the anionic form. Charge neutrality therefore requires a 1:1 ratio of GABA and benzoic acid. Figure 4 4clearly reveals that the crystal structure utilizes two instances of the (8) synthon with hydrogen bonds a and b common to the two synthons.
We have described a strategy to design cocrystals based on the (8) synthon by using pH as a controlling tool. Two cocrystals were obtained with the active pharmaceutical ingredient GABA, namely (GABA)2 oxalate and GABA benzoate. The identification of other robust and versatile synthons can be utilized as an aid in the selection of potential cocrystal formers. Note that in the course of this work, we have also isolated and obtained the crystal structure of two previously unreported hydrates of oxalic acid, a dimorphic sesquihydrate. These will be reported in due course.
Oxalic acid and GABA were weighed into a vial in a 1:1 stoichiometric ratio (likewise benzoic acid and GABA), water was added to dissolve the mixture, and then the pH was adjusted to pH 0, 3, 5, or 12 with NaOH or HCl solutions as needed.
X-ray crystal structural analysis for the cocrystal of oxalic acid/ GABA: Data were collected at 297 K by using a Bruker SMART 6000 CCD diffractometer with MoKα radiation (λ=0.71073 Å) Mr=98.76, monoclinic space group R21/c, a=7.4566(9), b=10.2685(13), c=9.7924(12) Å, β=108.478(3)°, V=711.13(15) Å3, Z=4, ρcalcd=1.384 g cm−3, μ(MoKα)=0.120 mm−1, F(000)=316, θmin=2.9°, θmax=28.3°, R=0.056, wR=0.1374 observed data (with I>2σI)=1646, total=7832, unique=1762, R(int)=0.0293, Goof=1.117 for 131 parameters.
X-ray crystal structural analysis for the cocrystal of benzoic acid/GABA: Data were collected at 297 K by using a Bruker SMART 6000 CCD diffractometer with MoKα radiation (λ=0.71073 Å) Mr=225.4, monoclinic space group P21/n, a=13.424(3) Å, b=6.212(1) Å, c=13.999(3) Å, β=101.748(4)°, V=1142.8(4) Å3, Z=4, ρcalcd=1.309 gcm−3, μ(MoKα)=0.100 mm−1, F(000)=480, θmin=1.9°, θmax=28.3°, R=0.046, wR=0.1128 observed data (with I>2σI)=1663, total=7108, unique=2710 R(int)=0.034, Goof=1.011 for 205 parameters.
- 14J. Chem. Soc. 1973, 2, 503–508., , ,