An SPL 16 with UPF horizontal shaker (Labcon, Petaluma, CA, USA) was used for preparing the solvent extracts. A rotary evaporator (Buchi, Flawil, Switzerland) was used to evaporate the organic solvent in vacuo. Precision balance 220/C/2 (RADWAG Instruments, Radom, Poland) were used for weighing milligram and gram quantities. Finnpipettes™ (Thermo Labsystems, Helsinki, Finland) were used for spiking volumes. Glass pipettes were used for adding extraction solvents. Filter papers (Thermo Labsystems) were used for filtration. Thin-layer chromatography (TLC) plates were purchased from Sigma-Aldrich (Sigma-Aldrich, Johannesburg, South Africa).

Acetone, hexane, dichloromethane (DCM), methanol (MeOH), ethyl acetate (EtOAC) and sulphuric acid (H₂SO₄) were purchased from Merck (Merck, Darmstadt, Germany). Vanillin, 2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-phenyltetrazolium chloride was purchased from Sigma-Aldrich (Johannesburg, South Africa). Distilled water was prepared by a Millipore water system (Milli–QTM, Darmstadt, Germany). Ciprofloxacin and amoxicillin were purchased from Sigma-Aldrich.

The leaf (5 g) was extracted with acetone (50 mL). Acetone was chosen as the solvent in this comparative study because it extracts a range of polar and non-polar compounds (Levy & Marshall 2004). The mixture was vigorously agitated on a linear shaker and filtered. The filtrate was concentrated to dryness using a rotary evaporator and the residue reconstituted to 50 mg/mL in acetone.

The acetone leaf extract was subjected to various qualitative phytochemical tests using various methods (Boye et al. 2012; Omoruyi, Bradley & Afolayan 2012). The phytochemicals tested were alkaloids, saponins, phytosteroids, phenolics and tannin compounds.

TLC plates were loaded with the acetone leaf extracts (100 µg) of the leaf, developed using a mobile phase of hexane: EtOA_C (2:1), (1:2) and also in acetone only. The plates were each visualised under UV light at 254 nm and 365 nm before being sprayed with vanillin–sulphuric acid reagent and heated at 100^°C for 10 min until the colour changed. The replicates, which were not sprayed with reagent, were used for the bioautography assay. The retention factor (R_f) values of the separated spots or bands were determined as follows:

R_f = distance travelled by the compound / distance travelled by solvent front

A comparison was conducted of the antibacterial chemical components present in the leaves of C. anisata. Acetone extracts of the leaves were analysed on TLC plates using hexane EtOAC (2:1) as mobile phase and then sprayed with S. pyogenes, E. coli and S. aureus; we also developed TLC plates without bacteria. Clear lines show the inhibition of growth of bacteria.

TLC plates were loaded with 20 µL of extract, developed in mobile phase. The plates were dried in an oven at 40^°C for 1 h to make it free from solvents used to develop the chromatograms. The developed bands were sprayed with overnight cultures of E. coli, S. aureus and B. cereus on different TLC plates until completely wet and then were each incubated at 37^°C in a humidified chamber for 24 h. The plates were sprayed with r-iodonitrotetrazolium violet (INT) and incubated for an extra 24 h. The presence of clear bands on the plates against a purple background indicated growth inhibition. The bioautography plates were compared with the reference TLC plates prepared (Lekganyane et al. 2012; Masoko & Nxumalo 2013; Masoko, Picard & Eloff 2005).

The acetone leaf extract of C. anisata was evaporated to dryness and separated on a flash column filled with silica gel (216 g, 70–230 mesh; Merck), eluted with hexane–EtOA_C mixtures (in the ratios 1000 mL: 100 mL, 1600 mL: 800 mL), EtOA_C (1000 mL) and finally washed with MeOH (1000 mL). Fractions were combined according to their R_f values; a total of seven fractions (A–G) were obtained. Fraction A was singled out for structural elucidation because of the presence of a single band in the TLC. Fraction A was evaporated to dryness and re-dissolved in 100% acetone, loaded onto glass preparatory plates and developed with hexane: EtOA_C (2:1) in the glass tank. The final compound was characterised by a 600 MHz Bruker (Rheinstetten, Germany) ¹H NMR spectrometer equipped with a 5 mm dual probe, CDCL₃ was used as the solvent and chemical shifts were reported relative to the solvent peaks.

Fourier transform infrared spectroscopy (FT-IR) spectrum was recorded using a Shimadzu (Tokyo, Japan) FT-IR Model 8300 6AD spectrophotometer. The sample (Fractions C, D and G) was mixed with potassium bromide pellets (KBr) as a disc and read at a spectral range from 400 cm⁻¹ to 4000 cm⁻¹. The procedures of Shriner et al. (1979) and Al-Maliki (2011) were followed.

The procedures of Schreiber et al. (2010) were followed for the LC-MS analysis, performed in triplicate and in a randomised order on Fractions C, D and G. Analyses were performed with an Agilent LC-MS system (AB SCIEX, Framingham, MA, USA) Triple TOF 5600) with the analyser and electron spray ionisation source in positive mode. Source parameters were optimised to provide high sensitivity. The source parameters were as follows: gas temperature 600°C, drying gas flow rate 0.5 mL/min, nebuliser pressure 50 psi, capillary voltage 5500 V, separation was carried out by ultra-high pressure liquid chromatography using a Shimadzu UFLCXR with a Phenomenex (California, USA) Kinetex 2.6 µm (100 × 2.1) column. The mobile phase used was A: water (5 mM ammonium formate + 0.5% formic acid) and B (acetonitrile); the gradient program was as follows: 5% B for starting condition, increase up to 95% B in 20 min, hold 25 min, increase % B to 95% from 2 to 20 min, hold 25.20 min and decrease % B to 5% at the final step. The total run time was 30 min. The injection volume was 50 µl. The standardised collision energy was 35 V with collision energy of ±15 V; statistical data processing was carried out using MarkerView software. The detection was accomplished using MS/MS scan mode; compound identification and quantification were performed using PeakView software.

Isolation and identification of compounds in the selected fractions of C. anisata leaf extracts, using the FT-IR, LC-MS and NMR techniques, was achieved by comparison of retention times obtained at identical chromatographic conditions of the analysed fractions; the molecular mass was compared using MS data. Compounds were identified by analysis of the chromatograms using a PeakView library, R_f values, functional group frequency, literature and computer search with the aid of ChemSpider software.

Phytosteroids, saponins, alkaloids, tannins, phenolics and flavonoids were all present in the acetone leaf extracts of C. anisata (Table 1). Plant phenolics including flavonoids, phenols and tannins have been reported for their many medicinal attributes and capability of defending humans against foreign bodies and illness (Dai & Mumper 2010). With reference to the health benefit significance of plant phenolics to humans, it bears out the relevance of phytochemical content investigation in plant-based research, which is also in line with the view of Masoko and Nxumalo (2013).

TLC has been reported to be one of the best and most appropriate methods for chemical content profiling and a preliminary means for compound evaluation in plants (McGaw, Jager & Van Staden 2002). TLC of the acetone leaf extract was used as a guide to compare the different fractions obtained during purification using flash column chromatography; the various bands were observed (indicating different constituents and compounds) during daylight with an observation of light green, which indicates the presence of phenols, and light orange, which shows the presence of alkaloids (Figures 2 and 3). This was further confirmed, under ultraviolet ray light of 254 nm and 365 nm, which revealed the presence of phenols with light violet colour under 365 nm and light green under 254 nm in Fractions C and G, with R_f values of 0.46 and 0.40, respectively, while the fluorescent colour blue was observed in Fraction D, which indicated the presence of alkaloidic compounds, with R_f values of 0.30 and 0.15 (Table 2). This study is in total agreement with the report on isolation and identification of phenols and alkaloids in Elettaria cardamomum under several observations (Al-Maliki 2011). Different spraying reagent was also used to further confirm these compounds, Dragendorff’s reagent was used to confirm the presence of alkaloids while the orange colour confirmed the presence of alkaloid compounds in Fractions C and D. Vanillin, and vanillin in H₂SO₄, was used to confirm the presence of flavonoids, which was in line with the reports of some researchers, including Al-Maliki (2011), Sasidharan et al. (2011) and Shai & Eloff (2009)

Medicinal herbs constitute the major factors in herbalism and traditional healing practices all over the world. Plants serve as a reservoir for the storage of several classes of compounds; these compounds act as a defence mechanism against foreign bodies and therefore help in the management and cure of several ailments and health conditions in humans and animals. In the isolation process, the choice of solvent for extraction is vital. Some compounds have previously been isolated from the Rutaceae family and C. anisata is one of the species in this family with a similar compound (Figure 1). Some researchers, including Eloff (1999) and Masika and Afolayan (2003), have reported that acetone extracted both polar and non-polar compounds in plants. This is in support of the choice of solvent in this study, while the acetone leaf extract of C. anisata (Fractions C, D and G) was established to possess compounds that showed significant antibacterial activity against the test organisms S. aureus, E. coli and S. pyogenes compared to the other fractions (A, B, E and F), which showed no activity as revealed in the bioautography assay. Two components in C showed activity against the test organism with R_f values of 0.24, 0.52 and 0.70. One component in D inhibited the test organism with R_f values of 0.20 and 0.52, while two components in G also inhibited the growth of the test organism with R_f values of 0.20 and 0.54. The inhibition diameters of the fractions were also recorded. The fractions showed a relatively higher inhibition zone as compared to standard drugs (amoxicillin and ciprofloxacin) (Table 2). The effectiveness of the fractions obtained from acetone extracts corroborates with a similar antimicrobial study on Curtisia dentata growing in South Africa, while the compounds isolated from the acetone leaf extract showed significant inhibition against some microorganisms (Shai & Eloff 2009). Overall, the highest activity on the test organism was observed with Fractions D and G with inhibition diameters of 8 ± 2.0 and 6 ± 1.5 against S. pyogenes as compared to the standard drugs. Although antibacterial activities have been reported in other countries for C. anisata, for instance India, Ghana and in KwaZulu-Natal Province of South Africa (Agyepong et al. 2014; Parekh, Jadeja & Chanda 2005; York et al. 2012), this could justify its folkloric usage as one of the plants in the Eastern Cape for treating respiratory infections, including tuberculosis, as previously reported (Buwa & Afolayan 2009; Lawal et al. 2014).

From the FT-IR chromatograms of the isolated alkaloid compounds, for Fraction C there was indication of a strong band at 1719 cm⁻¹ belonging to the (C=O) α, β unsaturated ester group of alkaloids, while strong bands at 726 cm⁻¹ belonging to an aromatic group (C–H) were observed, as was a broad band (3632 cm⁻¹) representing the stretching vibration of a phenolic or alcoholic (OH) group with the hydrogen bonding. The two medium bands at 2725 cm⁻¹ and 2940 cm⁻¹ represent aliphatic C–H; aldehyde/ketone group (H-C=O:C–H), (C–H rock and C–H bend) alkanes at 1370 cm⁻¹ and 1455 cm⁻¹ were also observed in the chromatograms. The medium band at 1157 cm⁻¹ confirms the alkaloidic stretch of aliphatic amine (C–N). Fraction D revealed a sharp band at 3631 cm⁻¹ representing (O–H) stretch free hydroxyl groups of alcohol or phenols, while a medium band of 3160 cm⁻¹ representing (O-H) stretch of carboxylic acids was observed; at 2726 cm⁻¹ and 2945 cm⁻¹ a medium stretch of aldehyde (H–C=O:C=O–) and alkanes group (C–H) were also observed. The alkane groups of C–H rock and C–H bend at 1370 cm⁻¹ and 1455 cm⁻¹ were also represented. A medium stretch of aliphatic amine (C–N) was indicated at 1071 cm⁻¹, reflecting the presence of an alkaloidic compound. Fraction G affirms the presence of phenols (flavonoid) at the strong stretch of free hydroxyl of phenol or alcohol with the frequency of 3622 cm⁻¹, which represented the stretching vibration of hydrogen bonding, indicating the presence of the flavonoid. The frequencies 2923 cm⁻¹ and 2725 cm⁻¹ represent groups of medium stretch alkanes (C–H) and aldehyde (H–C–C=O:C–H), respectively, while the frequency of 1653 cm⁻¹ represents a medium stretch of the alkene group (–C=C). This study is in line with a similar study on the identification of alkaloids and phenolic compounds and their functional groups frequencies predetermining the classes and group of compounds in medicinal plants (Al-Maliki 2011; Rashed & Butnariu 2014).

The acetone extract of C. anisata leaves was subjected to column chromatography and TLC using a mobile phase of hexane: EtOAC (2:1, 1:1 and 1:2) and washed with MeOH to yield Fractions A-G. ¹H NMR results revealed that Fraction A, which was a yellowish oil, is a mixture of 5,7-disubstituted coumarin derivatives, which were found to be enantiomers 4a-b; NMR peak values are shown in Table 3 and the spectrum in Figure 4. The scientific name of the new compound isolated is (S/R)-3-(1,1,3-trimethyl-2-butenyl)-5-(2,3-dihydroxy-3-methylbutane)-7-methyl-2H-1-benzopyran-2-one (Figure 5 d), depending on the configuration of the function group at position 2’.

Compounds Compounds A and B (Figure 5) obtained from Fraction A are swiete coumarins, which are formed from anisocoumarin A in which the aldehyde group (CHO) was oxidised into the group with asterisk (*). Compounds A and B differ at position 6’. Because of the chiral centre at 2’, this newly isolated compound might display be peaks that are doubled (Figure 5).

LC-MS is regarded as one of the best techniques to provide detailed information on targeted and non-targeted compound in natural products (Soam et al. 2013; Wolfender et al. 2003). The LC-MS results are presented in Figures 7 and 8. The following compounds were identified: Fraction C (clausine B), Fraction D (clausenocoumarin and α-terpineol) and Fraction G (quercetin 3, 4-dimethyl ether) (Figure 6). These compounds were identified at different retention times and molecular weights. They are alkaloids and flavonoids, compounds isolated proven to be very active against the test bacteria; this is also in support of the previously reported potency of some alkaloid compounds and flavonoids isolated from plants against several diseases and health conditions such as tuberculosis, oxidative stress such as arthritis, fungal infections and inflammation. Clausine B and quercetin 3, 4-dimethyl ether have been previously isolated from Murraya koenigii and Dillenia indica, respectively (Chakarbotty, Barman & Bose 1964; Gowsala & Uvais 1975). However, these compounds were identified in the extract of C. anisata growing in South Africa using LC-MS profiling and were screened for their antibacterial activities using bioautography. Therefore, this study further supports the belief of the traditional healers and herbalists in the Eastern Cape using C. anisata for the management of some infectious diseases.