Three new proaporphine alkaloids from Cissampelos capensis L.f. and their cytotoxic evaluation

Abstract Three new proaporphine alkaloids, cissamaline (4), cissamanine (5), and cissamdine (6), along with two known compounds (2 and 3) were isolated from total tertiary alakloidal extracts of the leaves and stems of Cissampelos capensis. The new alkaloids were elucidated through comprehensive spectroscopic analysis, including 1 D, 2 D NMR and High-resolution LCMS. The in vitro MTT cytotoxicity was evaluated on Caco-2 cell lines, where all the isolated compounds exhibited cytotoxic effects at concentrations above 250 µM. Graphical Abstract


Introduction
The Cissampelos genus, belonging to the Menispermaceae family, is widely used in traditional medicines worldwide as established in reports about its therapeutic and toxic capacity across cultures (Semwal et al. 2014;da Silva Mendes et al. 2020). The frequently studied species worldwide included C. mucronate, C. pareira, C. sympodialis, and C. capensis, where the isoquinolines were the main alkaloid type found. However, there remains a paucity in the pharmacological potential and phytochemical characterisation of the genus (da Silva Mendes et al. 2020). Recent research on the phytochemistry of the Cissampelos genus includes the cularine-type isoquinoline alkaloid isolated from C. pareira (Kumari et al. 2022). The research on the phytochemistry of Cissampelos capensis has mostly targeted the alkaloids, as previous reports suggested that the biological activity was due to its rich diversity of benzyltetrahydroisoquinoline alkaloids (De Wet et al. 2011). Key findings include the positive cytotoxicity against renal, melanoma, and breast cancer cells exhibited by an alkaloidal crude extract from the leaves and rhizomes of C. capensis (De Wet et al. 2009). To date, 21 chemical compounds have been reported from C. capensis, including three proaporphine alkaloids: crotsparine (1), glaziovine (2), and pronuciferine (3) (Figure 1). Where the proaprophines 1, 2, and 3 were successfully isolated from the leaf extract and identified through HPLC analysis in the stem extract of C. capensis (De Wet et al. 2011). Furthermore, some of the alkaloids found in the leaves of C. capensis are known to be toxic. Thus, to extend the phytochemistry of C. capensis further phytochemical analysis was conducted. Resulting in the isolation and elucidation of three new proaporphine alkaloids; compounds 4, 5, and 6. This paper explores their purification, structural elucidation, and cytotoxicity capacity.

Structural elucidation of the new proaporphine alkaloids
Compound 4, a deep purple powder, was isolated from the stem extract of C. capensis and appeared as a brown spot after staining with ninhydrin, implying an amine-containing compound. A nominal MS, m/z 314.1711 [M þ H] þ , was observed and agreed with a molecular formula of C 19 H 23 NO 3 . The 13 C ( Figure S1) and 1 H NMR ( Figure S2) spectra of compound 4 bared similarity to the proaporphine alkaloids: crotsparine (1), glaziovine (2), and pronuciferine (3), from the Southern African Menispermaceae family (De Wet et al. 2004, 2005, 2011. However, an additional strong carbonyl group (d C Figure 1. The proaporphine alkaloids molecular structures of 1-6 from Cissampelos capensis. 208.68) and methyl (d C 29.26) observed, in the 13 C NMR, did not fit the known proaporphine alkaloids. After comparison against various NMR databases compiled from the Menispermaceae family, proaporphine alkaloids and neighbouring families of alkaloids, the closest relation to the data was the aforementioned proaporphine alkaloids (Bernauer and Hofheinz 1968;Barbosa-Filho et al. 2000;De Lira et al. 2002;Do Nascimento et al. 2012;Semwal et al. 2014). Further investigation into the biosynthetic pathway of compounds 2 and 3 revealed the formation of an intermediate B (in the first step of the reaction) with two carbonyl substituents to its structure (Scheme S1) ( Barbosa-Filho et al. 2000). This intermediate structure and the NMR 2 D data obtained, assisted in the elucidation of compound 4. Of the previously isolated proaporphines alkaloids, compound 4 was compared to 2 as the biosynthesis pathway describes the formation of 2 after the C-ring formation in the intermediate B. Different from compound 2, however, 4 comprises three rings, A-B connected to the D-ring by a methylene group. The 13 C and 1 H NMR comparison of compounds 2 and 4 is shown in Table S1. The regioselective substitution of the functional groupscarbonyl group at C-1 (d C 208.68), methoxy group (d C 55.32) at C-2 (d C 153.54), and the N-CH 3 (d C 41.93) to the A and B-rings were observed through HMBC correlation as shown in Figure S3. The carbonyl C-1 (d C 208.68) position was confirmed by the HMBC correlation of the 7c-CH 3 proton (d H 2.05) with the carbonyl group. The C-7c (d C 59.90) position was confirmed by the H-7c (d H 3.49) correlation to C-7b (d C 144.21). The C-2 (d C 153.54) position was supported by the methoxy proton (d H 3.69) correlation. The A-ring arrangement was completed through the H-3 (d H 6.68) correlations to C-7b (d C 144.2) and C-3a (d C 132.05), while establishing the connection to the B-ring by C-4 (d C 26.60). Compared to compound 2, the B-ring chemical shifts remained unchanged, aside from the influence of the addition of the electronegative carbonyl group at C-1. The methylene bridging C-7 (d C 46.30) position was supported by H a -7 (d H 2.30) HMBC correlations to C-6a (d C 65.31), C-5 (d C 54.44) and C-3a (d C 132.05), while H b -7 (d H 2.14) correlated to the 7c-CH 3 (d C 29.26). The chemical shifts of the D-ring of compound 4 compared to 2 remain unchanged. Selected COSY correlations have also been highlighted ( Figure S4); H-7a (d H 3,41) to H-6a (d H 3.52), H-7c (d H 3.49) to H-6a (d H 3.52) and H a,b -7 (d H-a 2.30; d H-b 2.14), confirms the structure of compound 4. The structural depiction of selected HMBC and COSY correlations of compound 4 is shown in Figure  S5. The NMR data, MS data, and the supporting biosynthesis pathway for compound 4 have led to the characterisation as cissamaline.
Compound 5, an orange powder, was isolated from the leaves of C. capensis and appeared as a brown spot after staining with ninhydrin, implying an amine-containing compound. A nominal MS, m/z 302.1345 [M þ H] þ , was observed and agreed with a molecular formula of C 17 H 19 NO 4 . Compound 5 produced similar 13 C ( Figure S6) and 1 H ( Figure S7) NMR spectral data as 4 as shown in Table S2; however, upon thorough inspection, deviations were observed. The major variations were the detection of an oxymethine carbon (d C 73.41), the lack of the N-CH 3 (d C 41.93) in the 13 C NMR, and no HMBC correlation at (d H 2.02) to the C-1 carbonyl group was observed. The structural elucidation of compound 5 was further characterised through HMBC ( Figure S8) and COSY ( Figure S9) experimentation. The decision of the N-H substituent to the B-ring was informed by the structural comparisons of compounds 1, 2, and 3, where the N-H substituent in 1 was determined based on the absence of the N-CH 3 (± d C 41.93) in the 13 C NMR (De Wet et al. 2004Wet et al. , 2005. The position of the hydroxy substituted C-7c (d C 73.41) was confirmed through the HMBC correlation with various neighbouring protons including H b -5 (d H 3.69), H-6a (d H 3.26) and H a -7 (d H 2.29). The 7c-OH proton (d H 5.15) displayed HMBC correlation to C-9 (d C 127.61) and COSY correlation to H a,b -7 (d H 2.95, 2.40) and H b -4 (d H 3.02), confirming the position of C-7c. The assignment of methine carbons C-7a (d C 59.86) and C-6a (d C 56.09) was determined by comparing their proton HMBC correlations. The proton at H-6a (d H 3.26) correlated to its neighbouring carbons C-7b (d C 73.41) and C-5 (d C 65.57) by a J 2 HMBC correlation. If the chemical shift was assigned to C-7a the C-5 correlation would be more (J 4 ) and thus less likely. Further, HMBC correlation of H-3 (d H 6.91) confirmed the position of C-4 (d C 23.64) and C-7c ( Figure S10. The NMR data and MS data led to the structural characterisation of cissamanine. Compound 6, a dark brown powder, was isolated from the leaves of C. capensis and appeared as a brown spot after staining with ninhydrin, implying an amine-containing compound. A nominal MS, m/z 286.1460 [M þ H] þ , was observed and in agreement with a molecular formula of C 17 H 19 NO 3 . The 13 C ( Figure S11) and 1 H ( Figure S12) NMR data of the compound 6 was consistent with that reported for compounds 4 and 5 (Table S2); however, the absence of the C-1 carbonyl group and C-3a (± d C 132.7) was observed in the 13 C NMR. The ring arrangement and regioselective substitution of compound 6 were further determined through HMBC ( Figure S13) and COSY ( Figure S14) experiments, analysing the C-H and H-H correlations ( Figure S4). The HMBC correlation observed for H-12 (d H 6.73) with C-8 (d C 151.22) and C-10 (d C 185.50), along with the COSY correlation of H-8 (d H 6.89) to H-9 (d H 6.35), and H-11 (d H 6.26) to H-12 (d H 6.73) confirmed the formation of the D-ring as previously observed in 4 and 5. The proton at H-9 (d H 6.35) and H-11 (d H 6.26) correlated respectively to methine C-7a (d C 50.85), completing the D-ring. The position of C-2 (d C 148.92) was confirmed by the HMBC correlation of the OCH 3 proton (d H 3.80) to C-2 (d C 148.92) and the H-3 (d H 6.58) correlation confirmed C-7b (d C 141.97). COSY correlation of OCH 3 (C-2) (d H 3.80) to -OH (C1) (d H 3.47) suggested a hydroxy substituent on C-1 with the chemical shift of (d C 70.56). The position of C-7b (d C 42.95) is confirmed through the correlation H a,b -7b (d H 2.43, 2.97) to H-6 (d H 4.33). The structural depiction of selected HMBC and COSY correlations of compound 6 is shown in Figure S15. For the proposed structure, C-3a as quaternary carbon was also suggested as signal d C 146.79; from the 13 C-NMR spectra this signal appears to integrate for two carbons. The structural elucidation of compound 6 was achieved through the comparison to 4 and 5, NMR COSY and HMBC data, and nominal MS analysis determined as cissamidine.
The Circular Dichroism (CD) spectrum of compound 2-6 ( Figure S16) exhibited a positive pp Ã Cotton effect at k ̴ 240 nm and a negative pp Ã Cotton effect at k ̴ 270 nm corresponding with a 6aS, 7aR configuration at chiral position C-6a and C-7a, respectively (Snatzke and Wollenberg 1966).

MTT cytotoxicity evaluation of compounds 2-6
The MTT cytotoxicity assay was performed at 48 h for the five isolated proaporphine alkaloids, compounds 2-6 from C. capensis using Caco-2 cells. After 48 h, compounds 3, 4, and 5 exhibited cell viability above 50% at 250 mM, while the remaining compounds exhibited lower cell viability as shown in Figure S17. Compounds 2 and 3 have both been reported to possess cytotoxic effects at high concentrations (Sonnet and Jacobson 1971;Ma et al. 2014), which was similar to the findings here. Literature further reported limited tumour inhibition activity in vitro for compound 2 against the 9-KB tumour test system (Sonnet and Jacobson 1971).

General experimental procedures
1D ( 1 H, 13 C and DEPT135) and 2 D (HSQC, HMBC and COSY) NMR spectra were recorded on the Bruker Avance III 400. All proton and carbon chemical shifts are quoted relative to the relevant solvent signals (CDCl 3 [d H 7.26, d C 77.0 ppm], CD 3 OD [d H 3.33, d C 49.0 ppm]) (CDCl 3 /CD 3 OD: 99.8 atom % D, Merck, Germany). Proton-proton coupling constants (J value ) are reported in Hertz (Hz). All experiments were conducted at 300 K unless specified otherwise. LCMS and HRMS were conducted at the Central Analytical Facilities (CAF) at Stellenbosch University, South Africa. High-resolution LCMS was carried out on the Water Synapt G2 qTOF mass spectrometer; samples were introduced by ESI probe and detected by an ESI positive source with a 15 V cone voltage. The Waters BEH C18 column, 2.1 Â 100 mm, 1.7 mm, with Acetonitrile (ACN) and Water as a solvent system (both solvents lines contained 0.1% formic acid). Infrared spectra were obtained on the Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) TENSOR 27 spectrophotometer, fitted with a Diamond/ZnSe internal reflectance element capable of single bounce mode. The Cirular Dichroism (CD) were recorded on a Chirascan Plus Spectrapolarimeter at Stellenbosch University, South Africa.

Plant material
The C. capensis plant material was collected by Mr J Block from the rocky mountainous slopes of Joubertina, Eastern Cape (33.814558 S, 23.844843 E). Authentication of the plant as Cissampelos capensis L.f. was performed by Prof E. Campbell at the Botany Department at the Nelson Mandela University. The specimen was awarded the identification code PEU 25069 and stored in the university herbarium.

MTT cytotoxicity assay
Caco-2 cell viability was assessed by measuring the intracellular enzymatic conversion of the 3-(4,5-Dimethylthiazo-1-2-yl)-2,5-Diphenyltetrazolium bromide (MTT) substrate to the reduced formazan precipitate. Caco-2 cells were seeded in 96-well microtiter plates at a density of 3000 cells/well using 100 ml aliquots in complete medium and incubated for 24 h at 37 C and 5% CO 2 to allow for attachment. Cells were treated by adding 100 ml of plant extracts (solubilised in DMSO and diluted with DMEM:10% FBS) at three concentrations (125, 250 and 500 mg/ml). The final concentration of DMSO never exceeded 0.25% at any concentration. Melphalan (500 mM) was used as a positive control. Treatments were incubated for 24 h and 48 h at 37 C and 5% CO 2 , and then aspirated and replaced with 0.5 mg/mL MTT reagent (prepared in complete medium). After 2 h, contents were aspirated and replaced with DMSO to dissolve the formazan crystals. Absorbance was read at 540 nm using the BioTek PowerWave XS Microplate Reader (BioTek-Instruments, USA). Experiments were performed in quadruplicate and repeated three independent times (n ¼ 3), unless stated otherwise. SD of three independent experiments was calculated and is represented using error bars in graphs. Statistical significance was determined by means of the two-tailed student's ttest, where p < 0.05 ( Ã ) and < 0.005 ( ÃÃ ) was deemed significant relative to the untreated control, unless stated otherwise.
The Caco-2 (immortalized cell line of human colorectal adenocarcinoma) cells were purchased from Cellonex, South Africa. The Caco-2 cells were grown in TPP V R tissue culture plates employing low glucose Dulbecco's Modified Eagle Medium (DMEM), comprising glucose, L-glutamine, and sodium pyruvate, supplemented with 10% Fetal Bovine Serum (FBS). The cells were grown in a controlled atmosphere (5% CO 2 and 95% relative humidity) at 37 C. The cells were subcultured at a ratio of 1:3 upon reaching 80-90% confluency. Cells were detached using trypsin (0.25%) and routinely maintained with medium changes every 2 to 3 days.

Isolated compounds
Physical data HRMS and IR data can be found in appendix A.