Compounds with antibacterial and antioxidant activities from Cadia purpurea

Abstract Five compounds 1-5 were reported from leaves and roots of Cadia purpurea herein. 2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-17-(1-hydroxypent-4-en-2-yl)-10,13-dimethyl-1H-cyclopenta[a]phenanthren-3-ol (1), apigenin-7-O-rhaminoglucoside (2) and 13-O-pyrrolecarboxyl lupanine (3) were isolated from chloroform and methanol leaves extracts. And 5-((trihydro-2-oxy-4-(hydroxymethyl)-5-(3-hydroxyphenyl)furan-3yl)methyl)benzene-1,2,3,4-tetraol (4) and bis(2-hydroxybutyl) phthalate (5) were isolated from the chloroform: methanol (1:1) roots extract. The structures of the compounds were characterized using NMR, IR and GC/MS spectroscopic data. In vitro antibacterial and antioxidant activities of leaves extracts and compounds 1, 2, 4, and 5 were also evaluated. The present findings disclosed that tested compound isolates and extracts showed dose-dependent antibacterial and antioxidant activities. The molecular docking result indicates that isolated compounds show no violation of the Lipinski’s rule of five. Compounds 1, 2, 4, and 5 were reported herein for the first time from the leaves and roots of C. purpurea. Further biochemical investigation on whole parts of the plant may uncover additional chemical entities with better biological activities. Graphical Abstract


Introduction
Cadia purpurea (Fabaceae) is a poisonous green shrub distributed in Ethiopia, Eritrea, Yemen, Oman, North Somalia and North Kenya (Hedberg et al. 1989). In Ethiopia, people use roots of C. purpurea to cure severe wounds (Edwards et al. 1997) and the powder leaves along with coffee to heel fire burned body (Moravec et al. 2014). Flavonoids like apigenin, apigenin 7-O-glucoside and chrysoeriol were reported from leaves of the plant (Zweekhorst-Van Laer and Nelen 1976). Three novel quinolizidine alkaloids, viz., alkaloid I, II and III, were also reported from the leaves of C. purpurea of Ethiopian origin (van Eijk et al. 1976). Despite the extensive use of this plant as a remedy against various ailments in Ethiopia, reports on phytochemical constituents and biological activities of the leaves and roots parts were minimal. We report herein five compounds 1-5 along with their antibacterial and antioxidant activities, and drug-likeness characteristics from leaves and roots of C. purpurea.

Results and discussion
Herein, five compounds (Figure 1), three (1-3) from leaves extracts and two (4 and 5) from roots extracts of Cadia purpurea, were isolated and characterized (S1. Structural characterization). We also incorporated the report of the antibacterial and antioxidant activities of the leaves extracts and compounds 1, 2, 4, and 5. We reported the antibacterial and antioxidant activity of the roots extracts and compound 3 in our previous works (Kiros et al. 2022). Compound 5 was found as phthalate derivative which might be originated from possible plasticizers of industrial origin. This is because plants can possibly be contaminated by plasticizers which lead to the identification of phthalate esters in plant materials and/or extracts (Bianco et al. 2014;Venditti 2020).

Antibacterial and antioxidant activity evaluation
2.1.1. Antibacterial activity The zone of inhibition diameter (in mm) recorded by the extracts and compounds was presented in Tables S6 and S7, respectively and those showing an inhibitory value of less than 7 mm were considered as not active (Nascimento et al. 2000). All the tested concentrations of the chloroform, chloroform: methanol (1:1) and alcoholic extracts showed an activity against E. coli with the greater inhibitory value (15.7 ± 0.1 mm) provided by ethanol followed by methanol (15.3 ± 0.4 mm) and chloroform (15.0 ± 0.1 mm) extracts at 100,000 lg/mL (Table S6). Whereas the n-hexane extract did not indicate any activity against E. coli at all concentrations. The methicillin-resistant S. aureus bacterium showed a total resistance against all extracts at all tested concentrations. On the contrary, this n-hexane extract showed a potent activity against P. aeruginosa at all concentrations (10.2 ± 0.3 mm to 13.1 ± 0.1 mm). The inhibitory activity produced by all extracts against E. coli at all concentrations was found weak compared to chloramphenicol (23.1 ± 0.4 mm); while they exhibited a better activity than chloramphenicol (8.2 ± 0.6 mm) against P. aeruginosa at all concentrations.
Compound 2 recorded higher inhibitory values against E. coli at 1000 lg/mL (12.1 ± 0.1 mm) and 500 lg/mL (11.9 ± 0.0 mm) while it was completely resisted by S. aureus (Table S7). Compound 1 unveiled a weak activity against E. coli at all tested doses (6.8 ± 0.0-8.9 ± 0.1 mm) and also bared a mild activity against S. aureus merely at the higher amount of 1000 lg/mL (8.7 ± 0.0 mm). Compounds 4 and 5 also possessed a mild activity against E .coli even at the higher concentration of 1000 lg/mL (8.2 ± 0.2 mm and 7.5 ± 0.2 mm) whilst a better activity against S. aureus at the same dose (8.1 ± 0.1 mm and 7.7 ± 0.1 mm), respectively.

Evaluation of ferric reducing antioxidant power.
The ferric ion reducing power of each evaluated analytes was found positively related with corresponding concentrations (Table S9). The methanolic and ethanolic leaves extracts showed a better reducing power at the maximum concentration of 500 mg/mL with absorbance values of 0.72 ± 0.00 and 0.68 ± 0.00. At the same concentration, the chloroform and chloroform: methanol (1:1) extracts displayed weak reducing ability (0.39 ± 0.00 and 0.55 ± 0.00). Compound 4 indicated a notable ferric reducing strength with 0.62 ± 0.00 of absorbance at the concentration of 500 mg/mL. Compound 2 indicated a mild absorbance of reducing power (0.51 ± 0.00) at the same dose. The least absorbance values were shown in compounds 1 (0.401 ± 0.00) and 5 (0.47 ± 0.00) for similar concentration.

Molecular docking
The compounds 1, 2, 4, and 5 show some hydrogen bonding and hydrophobic including Van-der Walls interactions with some active site pockets of E. coli gyraseB (6F86) ( Table S10 and Figures S21-S25). All the isolated compounds had no violation of the Lipinski's rule of five (Table S11). Compounds 1 and 5 exhibit high gastro-intestinal absorption with skin permeation values (logP in cm/s) of À7.09 and À6.26 and show an inhibitory effect on CYP1A2 and CYP2D6, and CYP2C9 and CYP2D6, respectively; whereas compounds 2 and 4 exhibit À9.94 and À7.94 values of logP (in cm/s) with low gastro-intestinal absorption and no BBB permeability and are found as non-inhibitors of all the cytochrome-P (CYP) enzymes (Table S12). Compounds 1 and 4 were fall within the toxicity class of four (LD 50 values of 890 and 2000 mg/Kg), whereas compounds 2 and 5 were classified under the five class of toxicity with LD 50 values of 5000 and 2830 mg/Kg (Table S13).

General
Analytical grade organic solvents were used for the extraction and fractionation process. Extracts and collected fractions were analyzed on aluminum TLC sheet silica gel 60 F 254 (Merck) and visualized either by Uv-lamp (254 and 365 nm, UVP Chromato-Vue C-70G, Analytik Jena, USA), iodine, vanillin/methanol/H 2 SO 4 or Wagner's reagent. The melting point was measured by digital melting apparatus (SMP10, Bibby Scientific, UK) and the UV-Vis spectrophotometer analysis was conducted on Cecil CE4001 UV/VIS (England). The ASCII files of FT-IR were measured on Spectrum 65 FT-IR (PerkinElmer) in the wavenumbers range of 4000-400 cm À1 with 4 cm À1 resolution and 4 numbers of scans using the KBr pellets. The 1 H (400 MHz), 13 C and DEPT-135 (100 MHz) NMR experiments were conducted on a BRUKER ACQ 400 AVANCE spectrometer and all chemical shifts (d ppm ) of generated spectra were recorded using deuterated solvent peak as reference. The GC-MS experiment was performed on a 7890B GC coupled with 5977 A single quadrupole MS (Network, Agilent Technologies).

Extraction and isolation
Shed dried powder (100 g) of each part of C. purpurea was defatted with n-hexane (1 L x3) and successively extracted with chloroform, chloroform: methanol (1:1), methanol and ethanol solvents via shaking on orbital shaker (Hy-5A, Movel Scientific Instrument) for 24 h. Each extract was filtered by suction filtration and concentrated with rotary evaporator (Rotary vacuum, Jainsons) leading to 2.5, 1.8, 7.9, 8.0 and 0.3 g, respectively, of crude leaves extracts; and 0.6, 0.57, 4.7, 4.8 and 0.2 g, respectively, of roots extracts. Besides, an "acid-base shakeup" extraction technique was employed to target alkaloidal constituent from the leaves part. Dried powder leaves (100 g) were defatted with n-hexane, extracted with methanol, filtered and concentrated to afford 12 g. Then, defatted methanol extract (12 g) was suspended in 0.1 N tartaric acid and titrated to pH 5 with 10% Na 2 CO 3 . It was then partitioned in ethyl acetate followed by separation of the ethyl acetate portion and dried over anhydrous Na 2 SO 4 . The remaining aqueous-acidic phase was basified with 10% Na 2 CO 3 , to pH 11, followed by partitioning with ethyl acetate (50 mL, each) for several rounds. Thereafter, the ethyl acetate phase of each round was separated, combined, dried over anhydrous Na 2 SO 4 , filtered and concentrated affording 0.8 g.

Fractionation of leaves extracts
The chloroform extract (1 g) was adsorbed on silica gel and applied on top of column chromatography packed with silica gel (150 g, 230-400 mesh size). Elution was performed with CHCl 3 /Acetone with increasing polarity and one hundred ten (10-50 mL) fractions (Frs) were collected. Fr90-94 (eluted with 100% acetone) resulted in compound 1 (26 mg).

Fractionation of ethyl acetate leaves extract
The ethyl acetate extract (0.8 g) of the defatted methanol leaves extract was subjected to silica gel column chromatography with the eluents DCM, EtOAc and MeOH with gradient of polarity resulted in sixty (10-25 mL) fractions. Fr41-52, eluted with EtOAc/ MeOH (3:1), was found active against Wagner's reagent with slight impurity (500 mg). Portion of this fraction (300 mg) was further purified on silica gel column chromatography with the same solvent systems mentioned above leading to seventy sub-fractions (5 mL, each). Among them, sub-fr61-70 yielded compound 3 (67 mg).

Antibacterial activity
Three concentrations (50,000, 25,000 and 12,500 lg/mL) of n-hexane, chloroform, chloroform: methanol, methanol and ethanol leaves extracts, and four concentrations (500, 300, 100 and 50 lg/mL) of isolated compounds 1, 2, 4, and 5 were prepared from corresponding stock solutions (100,000 mg/mL for extracts and 1000 mg/mL for compounds in DMSO) and studied for their antibacterial activity, using agar disc-diffusion technique (Balouiri et al. 2016), against Staphylococcus aureus (S. aureus, ATCC 25923), Escherichia coli (E. coli, ATCC 25922) and Pseudomonas aeruginosa (P. aeruginosa, ATCC 27853) standard human pathogens collected from Ethiopian Public Health Institute. Standard antibiotic disc of chloramphenicol (30 mg/disc) and DMSO solvent were served as positive and negative controls. Experiment was done in duplicate and result was expressed as mean ± standard deviation.

Antioxidant potential evaluation
3.5.1. DPPH free radical assay The DPPH radical scavenging activity of extracts and isolated compounds was investigated in reference to ascorbic acid (Khorasani et al. 2015). Six various concentrations (500, 250, 150, 100, 50 and 25 mg/mL) were prepared from corresponding stock solutions (1 mg/mL in MeOH) and freshly prepared DPPH solution (2 mL, 0.004% w/v in MeOH) was added to each concentration and incubated for 30 min at room temperature. Thereafter, absorbance of each concentration was measured at 517 nm using UV-Vis spectrophotometer. The antioxidant activity against DPPH free radical of tested analytes was expressed in terms of scavenging percentage calculated using the following formula (1) and IC 50 value.

DPPH scavenging activity %
In which, A and A 0 are absorbance of analytes and DPPH solution, respectively. Each experiment was done in duplicate and obtained values were described as mean-± standard deviation.

Ferric reducing antioxidant power (FRAP) assay
The FRAP of extracts and isolated compounds was examined with six concentrations, 500, 250, 150, 100, 50 and 25 mg/mL in H 2 O. Each concentration was mixed with 0.2 M potassium phosphate buffer (2 mL, pH 6.6) and potassium ferricyanide (2.5 mL, 10% w/ v) solutions followed by incubation at 40 C for 30 min. Trichloroacetic acid (2.5 mL, 10% w/v) was added to the incubated solutions; centrifuged for 10 min at 3000 rpm and a supernatant (5 mL) of each solution was mixed with distilled water (2 mL) and ferric chloride (0.5 mL, 0.1% w/v) (Do et al. 2014). Finally, absorbance of each sample was read at 700 nm of UV-Visible spectrophotometer and expressed as mean ± standard deviation.

Molecular docking study
The molecular docking analysis of compounds 1, 2, 4, and 5 was studied by docking into E. coli DNA gyraseB protein model (PDB ID: 6F86). AutoDock software version 4.2 was applied for the protein optimization and entire molecular docking analysis. The energy grid maps of generated protein and ligand structures, saved in the PDBQT format, were calculated and analyzed using AutoDock vina. The grid box size was set at 46 Â 46 Â 46 Ð points by considering the maintaining space of the grid box at 0.375 Ð. The drug-likeness, ADME and toxicity predictions of the isolated compounds were computed by SwissADME, PreADMET and OSIRIS property explorer software.

Conclusion
Phytochemical investigation on leaves and roots of Cadia purpurea led to the isolation of five compounds 1-5, in which compound 5 was found as plasticizer derived phthalate ester. Isolated compounds and extracts showed a dose-depended antibacterial and antioxidant activities. All the compounds were found with zero violation of the Lipinski's rule of five against E. coli gyraseB. Compounds 1, 2, 4, and 5 were reported herein for the first time from leaves and roots of Ethiopian Cadia purpurea. Further chemical investigation on the plant may lead to isolation of additional chemical entities with encouraging biological activities.