CHEMICAL CONSTITUENTS OF THE ROOT WOOD OF ERYTHRINA SACLEUXII AND DETERMINATION OF THE ABSOLUTE CONFIGURATION OF SUBERECTIN

Phytochemical investigation on the root wood of Erythrina sacleuxii (Leguminosae) led to the isolation of nine secondary metabolites (1-9). Compound 1 was isolated from the genus Erythrina for the first time. The pure compounds were identified on the basis of comprehensive spectroscopic and spectrometric analyses, while their absolute configurations were determined based on chiroptical measurements. Compounds 5 and 6 showed weak antifungal activity against Pyricularia oryzae with MIC values of 20 μg/mL.


INTRODUCTION
The genus Erythrina, a member of Leguminosae family, is made up of more than 110 species of red or orange flowered trees or shrubs widely distributed in the tropical and subtropical regions of the world [1][2][3]. Erythrina sacleuxii Hua is a 9-24 m tall tree, endemic to Kenya and Tanzania [4]. Traditionally, the tree is used for the treatment of malaria fever and microbial infections [5,6]. Previous phytochemical studies on the root and stem bark of the Kenyan specimens of E. sacleuxii had yielded flavones, isoflavones, isoflavanones, pterocarpans, and isoflav-3-enes [7][8][9][10], which have attracted interest on the account of their antiplasmodial activity [7]. We have recently reported some flavonoids from the stem bark and twigs of the Kenyan specimens of E. sacleuxii with antifungal and cytotoxic activities [11,12]. As part of our continuing investigation on E. sacleuxii, herein, we report the isolation and structural elucidation of nine known compounds from the root wood of E. Sacleuxii ( Figure 1) and their antifungal activity.

General experimental procedures
Melting points were determined on a Stuart melting point apparatus SMP1 (UK) and are uncorrected. Silica gel 60 Å (35-70 μm) from Acros Organics was used for column chromatography. Preparative TLC was performed on 20 cm × 20 cm glass plates precoated with silica gel 60 PF 254+366 having 0.75 mm layer thickness. The spots and bands were visualized under a UV lamp (254 and 366 nm). Optical rotation measurements were performed on a 241 model polarimeter (Perkin-Elmer, Waltham, MA, USA) at 22 o C. NMR spectra were recorded on an Avance-III spectrometer (Bruker, Karlsruhe, Germany) at 400 and 100 MHz ( 1 H and 13 C, 136 respectively) equipped with a 5 mm probe head. The spectra were referenced to the residual solvent signal [13,14]. ESI-HRMS was obtained with a Q-ToF ULTIMA-III quadrupole TOF mass spectrometer (Waters, Eschborn, Germany). A Tensor 27 spectrometer (Bruker Optics, Ettlingen, Germany) equipped with a diamond ATR unit was used to record IR spectra. UV and ECD measurements were performed using a J-815 spectropolarimeter (Jasco, Tokyo, Japan) in a quartz glass cuvette with a path length of 1 mm and a spectral range of 400-185 nm at a scan speed of 20 nm/min with 8 repetitions. The concentration of the acetonitrile solution of 1 amounted to 0.25 mmol/L. The baseline was corrected by subtraction of a solvent spectrum obtained with the same parameters.

Plant material
The root wood of Erythrina sacleuxii was collected from Mutsengo, Kilifi County, along the Kenyan coast in August 2015. Identification and authentification of the plant species was performed by Prof. S.T. Kariuki of Biological Sciences, Egerton University. A voucher specimen of E. sacleuxii was deposited at the Department of Biological Sciences, Egerton University (no. 0136EUH).
Compound 1 was isolated as a brown powder and assigned the molecular formula C 16 [24]. Two para-oriented aromatic proton singlets at δ H 7.24 and δ H 6.45 were assigned to H-5 and H-8, respectively, based on the HMBC cross-peaks of H-5 with C-4 (δ C 190.2), C-6 (δ C 143.3), C-7 (δ C 154.5), C-8 (δ C 103.4), C-9 (δ C 158.1), C-10 (δ C 112.7) and H-8 with C-4 (δ C 190.2), C-5 (δ C 106.9), C-7 (δ C 154.5), C-9 (δ C 158.1), and C-10 (δ C 112.7). The methoxyl group (δ H 3.81, δ C 55.5) was placed at C-6 based on its HMBC correlation to C-6 (δ C 143.23). The signals of aromatic protons of ring B comprised a doublet at δ H 7.05 (1H, 2.1 Hz, H-2') and a multiplet at δ H 6.86 (2H, m, H-5'/6'). The lack of splitting between H-5' and 6' could be due to dynamic process as a result of tautomerism between flavanone and chalcone. Furthermore, the small specific rotation value as a result of partial racemization due to inter-conversion between flavanone and chalcone suggests a free -OH at C-4′. The absolute configuration of (-)-suberectin which is reported for the first time was determined by comparing the experimental ECD spectrum with a TDDFT calculated ECD 138 spectrum of (-)-suberectin (Figure 2). The experimental ECD spectrum of (-)-suberectin exhibited a positive Cotton effect at around  = 340 nm as well as a negative Cotton effect at around 310 nm (Figure 2). Such observations in an ECD spectrum according to Gaffield, suggest that the compound should have an absolute configuration of (2S) [25]. To further confirm the absolute configuration of (-)-suberectin as (2S), its UV and ECD spectra were calculated using TDDFT at the TD-ωB97XD/6-311G++(d,p)/IEFPCM//B3LYP/6-311G(d,p)/IEFPCM level of theory. Comparison of the observed and theoretical spectra was performed using the procedure established by Bringmann and co-workers [26][27][28]. The calculated ECD spectrum of (S)-suberectin showed a similarity of 85.2% to the experimental one, whereas the calculated ECD spectrum of (R)-suberectin yielded only 1.23% similarity. The difference value (also called enantiomeric similarity index, ESI) therefore amounts to 83.97%. This observation further confirmed that (-)-suberectin has an absolute configuration of (2S). It is worth noting that (-)-suberectin can adopt four relevant conformations which only differ in the orientation of the hydroxyl groups of the phenyl substituent. The two most important conformers yield very similar ECD spectra ( Figure 3). For the lowest energy conformer, natural transition orbitals were calculated. The positive Cotton effect at approx.  = 340 nm seems to be caused by n→π* transition in the chromanone unit, whereas the negative Cotton effect at 310 nm is due to π→π* transition. At around 260 nm (scaled wavelength), two ECD bands appear to cancel each other. The positive Cotton effect at around 240 nm is predicted to originate from another chromanone π→π* transition.

Biological activity
The in vitro antifungal activity of the isolated compounds were tested against a panel of pathogens, including Botrytis cinerea, Candida albicans, Eremothecium coryli, Penicillium notatum, Pyricularia oryzae, and Rhizomucor miehei as agar diffusion assays as per the method described previously [29] with minor modifications. Compounds 5 and 6 showed weak activity against P. oryzae with MIC values of 20 µg/mL. The rest of the compounds did not show any appreciable activity. The positive control, ciclopirox, had an MIC of 5 µg/mL.