Crystal Growth, Structure and Morphology of Rifapentine Methanol Solvate *

Chinese Journal of Chemical Engineering, 20(3) 602-607 (2012) Crystal Growth, Structure and Morphology of Rifapentine Methanol Solvate* ZHOU Kun (周堃)...
Author: Valentine McGee
0 downloads 1 Views 433KB Size
Chinese Journal of Chemical Engineering, 20(3) 602-607 (2012)

Crystal Growth, Structure and Morphology of Rifapentine Methanol Solvate* ZHOU Kun (周堃)1, LI Jun (李军)2,**, LUO Jianhong (罗建洪)2 and JIN Yang (金央)2 1 2

College of Materials and Chemical Engineering & Chemistry, Chengdu University of Technology, Chengdu 610059, China College of Chemical Engineering, Sichuan University, Chengdu 610065, China

Abstract Rifapentine, an important antibiotic, was crystallized from methanol solvent in the form of its methanol solvate. The crystal structure of rifapentine methanol solvate belongs to monoclinic, space group P21, with the unit cell parameters of a = 1.2278(3) nm, b = 1.9768(4) nm, c = 1.2473(3) nm, Z = 2, and β = 112.35(3)°. The parallelepiped morphology was also predicted by Materials Studio simulation program. The influence of intermolecular interaction was taken into account in the attachment energy model. The crystal shape fits the calculated morphology well, which was performed on the potential energy minimized model using a generic DREIDING 2.21 force field and developed minimization protocol with derived partial charges. Keywords rifapentine, crystal structure, morphology, crystallization

1

INTRODUCTION

Rifapentine is a brick red powdered crystal, which is a derivative of rifamycin with excellent activity against Mycobacterium tuberculosis in vitro and in animals [1, 2]. Rifapentine is also one of the most effective antibiotics, which can be administrated in low-dosage and long-term in pulmonary tuberculosis treatment. Rifapentine has an advantage of five times longer half-life than rifampicin and it is recommended for use in intermittent therapy [3]. The crystal habit is an important variable in pharmaceutical manufacturing which is affected during crystallization by some factors such as the presence of impurities in the solvent [4, 5]. Industrially, most crystals are grown from solution and very few ones in vapour phase. Sometimes crystal morphology is affected by the different interaction between solvent molecules and different crystal surfaces. Control of the crystal morphology is a field of vital interest to the industries ranging from hydrometallurgy to pharmaceuticals as many crystal physical properties are implicitly dependent on their shapes [6, 7]. Therefore, basic studies on crystal structure and properties are important for both research and industrial application. Herein, the single crystal of rifapentine solvate was obtained from methanol solvent and its basic data including crystal structure and morphology are reported. 2 CRYSTAL GROWTH OF RIFAPENTINE FROM METHANOL Brick red crystalline rifapentine powder (C47H64N4O12, molecular mass 877.04) used in the experiment was purchased from Leshan San Jiu-Long March Pharmaceuticals Co., Ltd., China. Its mass

fraction determined by HPLC is higher than 99.0%. A certain amount of rifapentine solid was dissolved in methanol at 30 °C above the saturation temperature. The saturation temperature was determined by the solubility curve for the known solution concentration [8]. After the clear solution was obtained, the supersaturation was generated by cooling the solution from the initial temperature to 25 °C at cooling rate of 1 °C·d−1 to yield single crystals. Parallelepiped shaped crystals of Rifapentine CH3OH solvate (RIF M) were obtained. 3

CRYSTAL STRUCTURE MEASUREMENT

The crystal structure of an air-dried crystal of RIF-M (0.33 mm×0.30 mm×0.26 mm) was measured using a Rigaku MM-007 diffractometer operating at 50 kV and 20 mA, using Mo-Ka radiation (λ = 0.07107 nm) at 113 K. Data were collected in frames using oscillation method with θ ranging between 1.77° and 27.87°. Image processing and data reduction were done by using the Rigaku/MSC software. 23270 diffraction points, among which 12944 points were independent, were collected through the ABSCOR automatic collection and refinement program. The structure was solved by direct methods using SHELXS-97 and SHELXL-97 least-squares on F2 [9]. 4

RESULTS AND DISCUSSION

Cell parameters and other crystallographic information are given in Table 1 along with additional details concerning data collection. In the measurement, the R(int) value was 0.0353 and the goodness-of-fit on F2 was 1.048. The crystal structure was found to be

Received 2011-05-11, accepted 2011-08-20. * Supported by Open Fund of Mineral Resources Chemistry Key Laboratory of Scihuan Higher Education Institutions. ** To whom correspondence should be addressed. E-mail: [email protected]

Chin. J. Chem. Eng., Vol. 20, No. 3, June 2012

603

Figure 1 Molecular structure of rifapentine

Table 1

Crystal data and structure refinement for rifapentine+3 methanol

Identification code

Rifapentine+3 methanol

empirical formula

C47H64N4O12

formula mass

877.04

temperature

113 K

wavelength

0.07107 nm

crystal system

monoclinic

space group

P21

unit cell dimensions

a = 1.2278(3) mm b = 1.9768(4) mm

α = 90° β = 112.35°

c = 1.2473(3) mm

γ = 90°

2.8001(1) nm3

volume Z

2

density (calculated)

1.264 mg·m−3

absorption coefficient

0.093 mm−1

F(0 0 0)

1152

crystal size

0.33 mm×0.30 mm×0.26 mm

θ range for data collection

1.77°-27.87°

reflections collected

23270

independent reflections

12944

R(int) value

0.0353

goodness-of-fit on F

1.048

final R indices [I > 2sigma(I)]

R1 = 0.0631, wR2 = 0.1698

R indices (all data)

R1 = 0.0714, wR2 = 0.1769

2

monoclinic, space group P21 with unit cell parameters a = 1.2278(3) nm, b = 1.9768(4) nm, c = 1.2473(3) nm, Z = 2; F(0 0 0) = 1152; V = 2.8001(1) nm3, Dc = 1.264 g·cm−3. A single weighting scheme was applied, and the refinement continued until the final deviation factors, R1 and wR2, were 0.0631 and 0.1698, respectively. The molecular configuration of rifapentine and the atom numbering scheme are shown in Fig. 1. In the molecules of the compound, the five-membered ring of furan and naphthalin are planar. At the same time, the six-membered ring of piperazine and the five-membered of cyclopentyl have a chair conformation. The lengths of C H bond for C(13), C(18), C(21), C(24), C(27), C(29), C(31), C(35) and C(37) are all equal to 0.098 nm while that those for C(39), C(40), C(41), C(42), C(44), C(45), and C(47) are 0.099 nm. The lengths of C H bond for C(14), C(15), C(16), C(32), C(33) and C(38) are shortened to 0.095 nm due to the intensive attraction by double bonds. For the same reason, the lengths of C O double bond are shorter than C–O bond. The lengths of O H bond are all 0.084 nm no matter where the hydroxy group is located. There is an sp3 hybrid orbit between the two methyl groups and methanol, yielding an ideal tetrahedral structure with equal double bond angles of 109.5° [10]. The H C H bond angle in the six-membered ring of piperazine is 108.0°, while that in the five-membered ring of cyclopentyl is increased to some extent due to the relatively smaller steric hindrance. The H C H bond angle varies with its location in the five-membered ring of cyclopentyl. For example, the H(46A) C(46) H(46B)

604

Chin. J. Chem. Eng., Vol. 20, No. 3, June 2012

Table 2 D

Specified hydrogen bonds for rifapentine-methanol

H…A

D

H

H…A

D…A

D

H…A

O2

H2…O14

RIF- methanol 2

0.84

1.83

2.665(3)

170.3

O3

H3…O2

RIF- RIF

0.84

1.91

2.650(3)

146.5

H11…O10

RIF- RIF

0.84

1.71

2.462(3)

148.5

O11 O12 N1

H12…O8

RIF -RIF

0.84

1.78

2.608(3)

168.6

H1N…O15

RIF- methanol 3

0.82(4)

2.10(4)

2.891(4)

162(3)

O13

H13…O11

methanol 1-RIF

0.84

1.90

2.716(5)

164.0

O15

H15A…O2

methanol 3-RIF

0.84

2.02

2.837(3)

165.3

Figure 2 Molecular packing arrangement viewed along the a-axis of the united cell

bond angle is 108.6°, while the H(44A) C(44) H(44B) bond angle is 109.2° due to the hindrance effect of the six-membered ring of piperazine. The packing of the molecules is shown in Fig. 2, which indicates that the molecules are interlinked by hydrogen bonds when viewed down the a-axis. The carbon atoms of the methanol are C(48), C(49) and C(50). The hydrogen bonds exist between O(2)-O(3), O(8)-O(12), O(10)-O(11), O(2)-O(14), O(11)-O(13), O(2)-O(15) and N(1)-O(15). The atom O(13), O(14) and O(15) belong to methanol. The oxygen atom O(1), O(4), O(5), O(6), O(7) and O(9) of HC are not involved in hydrogen bonding. The most common way of linking between molecules is by head-to-tail hydrogen bonding of molecules in the crystal structure. Since methanol influences the crystal shape of rifapentine, it might be expected that changing the solvent would yield a different morphology [11]. 5 PREDICTION OF THE CRYSTAL MORPHOLOGY The molecular and crystalline modeling software based on molecular mechanics are now available to build new methods for prediction of crystal morphology. There are many methods to predict crystal structures, such as the Bravais Friedel Donnay Harker

(BFDH) method, the Periodic Bond Chains (PBC) model, the Attachment Energy (AE) method, and so on. Herein, the AE method is employed. From Bravais-Friedel Donnay-Harker (BFDH) theory a series of possible growth faces {h k l} can be determined by the crystal lattice and symmetry. In order to predict more accurately, the shape of a crystal, the energetics of the system must be taken into account. The attachment energy model of Hartman and Perdok makes use of the actual magnitude of molecular interactions to estimate relative growth rates of crystal faces [12]. The growth rate of the crystal face is assumed to be proportional to its attachment energy [13], that is, faces with the lowest attachment energies are the slowest growing and, therefore, have the most morphological importance. The attachment energy, Eatt, is defined as the energy released on attachment of a growth slice to a growing crystal surface. Eatt is usually computed as [14] Eatt = Elatt − Eslice (1) where Elatt is the lattice energy of the crystal and Eslice is the energy of a growth slice of thickness dhkl. In this work the crystal morphology of RIF-M was predicted using the Materials Studio simulation program. Appropriate force field and charge assigning methods are essential for the reliability of calculation. To validate the force fileds and charge assigning method

605

Chin. J. Chem. Eng., Vol. 20, No. 3, June 2012

(a) Figure 3

Predicted crystal morphology of rifapentine methanol solvate in vacuum (a) by BFDH model and (b) by AE model

for rifapentine, molecular dynamics in NPT ensembles for rifapentine crystal supercells with various combinations of force fields and charge assigning methods were performed. After equilibrating the system for 1 ns, the equilibrium crystal structure was calculated as the average of a 1 ns production run with a sampling interval of 50 s. The changes in the lattice parameters of the equilibrium crystal structure under different combination of force fields and charge assigning method were compared with the experimental X-ray structures. Among them, the Dreiding/Gasteriger set was found to give the lowest deviation in the lattice parameters. The ability of the determined Dreiding/Gasteriger set to reproduce the experimental density of rifapentine crystal was tested carefully when they were derived initially. Therefore, other validations were not performed. It was then concluded that this force field and charge assigning method were adequate for further simulations. Both the BFDH and AE models are used to predict the morphology of RIF-M in vacuum (Fig. 3). Both results are different from the experimentally observed morphology (Fig. 4).

Figure 4

(b)

Crystal morphology of RIF-M observed by SEM

This difference is ascribed to the influence of solvent on the growth of the experimental crystal and hence further modification is made to the calculation. These later calculations were made using the AE model which usually predicts the shape of the crystal more accurately because it takes the energetics of the system into account. Figure 5 lists the cleaved main growth faces of the crystal using the surface builder control panel. This feature is also useful for visualizing growth faces [15]. As methanol takes part in the intermolecular hydrogen bonds linked with O(2) and O(11) of rifapentine, the growth rate of different faces should be influenced by the position and amount of oxygen atoms. The number of exposed oxygen atoms is different on different cleaved main faces. The number of oxygen atoms on (1 0 0), (−1 0 1) and (0 2 0) faces are more than that of on (0 0 1) and (−1 −1 1) faces, resulting in stronger interaction between methanol solvent and faces (1 0 0), (−1 0 1) and (0 2 0). According to the Hartman-Perdok theory, the stable growth planes are those which consist of two-dimensional connected nets of strong bonds in the surface plane. Generally, those planes will have a non-zero step energy hindering nucleation and a low attachment energy [16]. To calculate the attachment energy the solvent and crystal surface, a supercell of rifapentine crystal with periodic boundary was built and relaxed using molecular dynamics at the temperature of 298 K and pressure of 0.0001 GPa. The particle mesh Ewald summation method was used to correct for the electrostatic interactions. After the system was equilibrated for 1 ns, the equilibrium crystal structure was calculated as the average of a 1 ns run with the sampling interval of 50 ps. Then the attachment energies were calculated using the equilibrium crystal structure. The expected crystal surface was obtained by cleaving the equilibrium crystal structure.

606

Chin. J. Chem. Eng., Vol. 20, No. 3, June 2012

(−1 −1 1)

(0 2 0)

(1 0 0)

(0 0 1) Figure 5

(−1 0 1) Cleaved main crystal faces of rifapentine methanol solvate

A solvent molecule was located on the center of a given surface and only the docked molecule was allowed to be movable. Dynamics of each surface at 386K were performed with NVT ensemble, 1 fs time step, and 100 ps dynamics time. Configurations of the simulated systems were taken every 10 ps. Coulombic interaction of the present model system was calculated by the Ewald summation method with accuracy of 4.184 J·mol−1. Finally, interaction energy of various outputed configurations between surface and adsorbed single molecule were mass-averaged. The developed minimization protocol with the derived partial charges and crystal-solvent interaction energy (Es) were kept for crystal morphology prediction, and the Es values were introduced into the calculation of modified attachment energy (Em). Eq. (2) was used to calculate the Em. (2) Em = Eatt − Es The energy value for the main faces of rifapentine crystal grown from methanol is given in Table 3. The attachment energy method was used to account for the probable nonbonded energetic interactions during crystal growth. The more negative the attachment energy in a particular direction, the faster the crystal growth rate and the less morphologically important the crystal face bounding that growth direction [17]. According to

Table 3

Energy value for the main faces of rifapentine crystal grown from methanol

Crystal plane

Es/kJ·mol−1

Eatt/kJ·mol−1

Em/kJ·mol−1

(−1 0 1)

−15.3762

−48.54

−33.163841

(0 0 1)

−0.59177

−67.346

−66.754235

(1 0 0)

−6.06795

−45.228

−39.160049

(0 2 0)

0.355289

−37.145

−37.500289

(−1 −1 1)

−1.06349

−62.471

−61.407508

the Em value, stronger interaction was found with the (0 0 1) and (−1 −1 1) faces and their growth rates are also faster than that of the (0 2 0), (−1 0 1) and (1 0 0) faces. Much stronger solvent interaction was found with the faces (0 2 0), (−1 0 1) and (1 0 0), which are more morphologically important crystal faces. The shape of RIF-M crystal predicted by the modified attachment energy model is shown in Fig. 6. It can be seen that the predicted morphology fits the experimental observed morphology very well. 6

CONCLUSIONS The crystal structure of rifapentine methanol

Chin. J. Chem. Eng., Vol. 20, No. 3, June 2012

3

4

5

6 7

Figure 6 Crystal morphology of RIF-M predicted by AE in methanol

solvate belongs to monoclinic, space group P21, and with the unit cell parameters of a = 1.2278(3) nm, b = 1.9768(4) nm, c = 1.2473(3) nm, and Z = 2. The included methanol takes part in the intermolecular hydrogen bonding of the rifapentine solvate and the crystal habit of rifapentine may varies with different solvent. The parallelepiped experimental morphology observed by SEM was successfully predicted by the Materials Studio simulation program using a potential energy minimized model with a generic DREIDING 2.21 force field and a developed minimization protocol with derived partial charges. REFERENCES

8

9

10 11

12 13

14

15 16

1

2

Dickinson, J.M., Mitchison, D.A., “In vitro properties of rifapentine (MDL473) relevant to its use in intermittent chemotherapy of tuberculosis”, Tubercle, 68, 113-118 (1987). Ji, B., Truffot-Pernot, C., LaCroix, M.C., “Effectiveness of rifampin, rifabutin, and rifapentine for preventive therapy of tuberculosis in

17

607

mice”, Rev. Respir. Dis., 148, 1541-1546 (1993). Miyazaki, E., Chaisson, R.E., Bishai, W.R., “Analysis of rifapentine for preventive therapy in the cornell mouse model of latent tuberculosis”, Antimicrob. Agents Chemother., 43, 2126-2130 (1999). Nokhodchi, A., Bolourtchian, N., Dinarvand, R., “Crystal modification of phenytoin using different solvents and crystallization conditions”, Int. J. Pharm, 250, 85-97 (2003). Nie, Q., Wang, J.K., Wang, Y.L., Bao, Y., “Effects of solvent and impurity on crystal habit modification of llα-hydroxy-16α, 17α-epoxyprogesterone”, Chin. J. Chem. Eng., 15 (5), 648-653 (2007). Rohl, A.L., “Computer prediction of crystal morphology”, Curr. Opin. Sol. State Mater., 7, 21-26 (2003). Long, C.X., Zhang, L.J., Qian, Y., “Preparation and crystal modification of ibuprofen-loaded solid lipid microparticles”, Chin. J. Chem. Eng., 14 (4), 518-525 (2006). Zhou, K., Li, J., Ren, Y.S., “Solubility of rifapentine in the binary system of acetic acid and n-octanol solvent mixtures”, J. Chem. Eng. Data, 53, 998-999 (2008). Sheldrick, G.M., SHELXS-97, “Program package for the solution and refinement of crystal structures, Release 97-2”, University of Göttingen, Germany (1997). Zhou, G. D., Structure Chemistry, Chemical Industry Press, Beijing (1995). Chen, J.X., Wang, J.K., Zhang, Y., Wu, H., Chen, W., Guo, Z.C., “Crystal growth, structure and morphology of hydrocortisone methanol solvate”, J. Cryst. Growth., 265, 266-273 (2004). Hartman, P., Morphology of Crystals, Part A, Reidel, Dordrecht (1988). Woensdregt, C.F., “Computation of surface energies in an electrostatic point charge model: I. Theory”, Phys. Chem. Min., 19, 52-58 (1992). Docherty, R., Clydesdale, G., Roberts, K.J., Bennema, P., “Application of Bravais-Friedel-Donnay-Harker, attachment energy and Ising models to predicting and understanding the morphology of molecular crystals”, J. Phys. D: Appl. Phys., 24, 89-99 (1991). Myerson, A.S., Handbook of Industrial Crystallization, 2nd edition., Butterworth-Heinemann, Boston (2002). Grimbergen, R.F.P., Bennema, P., Meekes, H., “On the prediction of crystal morphology. II. symmetry roughening of pairs of connected nets”, Acta Crystallogr., A54, 501-510 (1998). Givand, J.C., Rousseau, R.W., Ludovice, P.J., “Characterization of L-isoleucine crystal morphology from molecular modeling”, J. Cryst. Growth, 194, 228-238 (1998).