Archives
br Experimental and computation Standard samples of artemisi
Experimental and computation
Standard samples of artemisinin (1) and artemether (2) were purchased from China National Institutes for Food and Drug Control. Their ECD spectra, at a concentration of 1.0–5.0mg/mL in ethanol, were recorded in a quartz cuvette of 1mm optical path length using a Jasco J-815 CD spectrometer (Jasco Inc., Japan). Optical rotations of 1 and 2 were measured on a SGW-5 automatic polarimeter (INESA Inc., Shanghai) in ethanol at different wavelengths (633, 589, 578, 546, 436 and 405nm) and room temperature.
All quantum-chemical calculations were performed on structures with absolute configuration as depicted in Fig. 1. Initial conformational analysis was run using MMFF94 force field via the MOE software package [12]. The obtained MMFF94 conformers were further optimized with the software Gaussian 09 using the B3LYP functional at the 6−31G(d) and 6−31+G(d,p) basis set level [13]. These conformers were verified as minima on the potential 5-Methoxy-CTP mg surface by showing no imaginary frequency. Polarizable continuum model (PCM) was used to consider solvent effects using the dielectric constants 24.85 for ethanol. Oscillator strengths and rotational strengths in dipole velocity (Rvel) representation of the 20 lowest electronic transitions were calculated for each conformer. ECD spectra were then simulated by using a Gaussian function with band width σ=0.4eV. ORD curves were obtained by calculating the specific optical rotations at different wavelengths using the length representation together with gauge-invariant atomic orbitals.
Results and discussion
Because of the skeleton rigidity, both 1 and 2 have only one stable conformer in the equilibrium mixture. The optimized geometries at B3LYP/6−31+G(d,p) level are shown in Fig. 2. When compared with the single crystal X-ray data, the computed bond parameters coincided very well to the available experimental data, with mean standard derivation of bond length being ±0.012Å [14]. The peroxide bridge is regarded necessary for the biological activity, and its bond strength correlates with the antimalaria activity [15]. The computed OO bond lengths for 1 and 2 are 1.462 and 1.465Å, respectively, both located within the generally accepted range 1.458–1.467Å. Two adjacent CO bonds gave different bond lengths as 1.419 and 1.462Å, indicating the peroxide group was not strictly C2 symmetry. As to 1, the bond lengths of O1O2C3O13C12O11C10 chain are sequentially 1.462, 1.419, 1.449, 1.394, 1.459, and 1.356Å. The same carbon-oxygen chain in 2 gave a similar long-short-long bond length pattern. The nonplanar peroxide group of 1 arranged in a clockwise direction and gave a COOC dihedron angle as +47.2°, very close to the value +47.4° in single crystal. For 2, the COOC dihedron angle was also in a clockwise direction but the value decreased to +43.4°. The good match of theoretical and experimental structural data demonstrated that the B3LYP hybrid functional and the 6−31+G(d,p) basis set level could be a good choice to geometry analysis. The lower level of 6−31G(d) basis set could give similar results and need less computation time.
Both compounds 1 and 2 have been included in the Chinese Pharmacopeia (2010 version), and specific optical rotation at D-line ([α]D) is listed as the testing item of identification [16]. The permissible ranges are +75°−+78° and +168°−+173° for 1 and 2 in ethanol, respectively. In this study, their experimental [α]D values were +76.7° and +169.8°, both located in the qualified range. Both compounds showed positive optical rotations and monotonically decreasing ORD curves with the increase of wavelength over the range 400−600nm, but 1 gave smaller values than 2. Theoretical specific optical rotations of 1 and 2 were obtained using the B3LYP and CAM-B3LYP functionals, and the latter offered better results. The predicted [α]D values of 1 and 2 at CAM-B3LYP/6−31+G(d,p) level are +104.43° and +204.45°, both in positive sign and very close to the experiment. Their theoretical ORD curves reproduced remarkably well the experimental data (Fig. 3). Thus, internal environment is clear that correct absolute configurations of 1 and 2 could be assigned by comparison of the experimental and theoretical [α]D values and ORD curves.