Isolation and Structural Elucidation of Compounds from Pleiocarpa bicarpellata and Their In Vitro Antiprotozoal Activity

Species of the genus Pleiocarpa are used in traditional medicine against fever and malaria. The present study focuses on the isolation and identification of bioactive compounds from P. bicarpellata extracts, and the evaluation of their antiprotozoal activity. Fractionation and isolation combined to LC-HRMS/MS-based dereplication provided 16 compounds: seven indole alkaloids, four indoline alkaloids, two secoiridoid glycosides, two iridoid glycosides, and one phenolic glucoside. One of the quaternary indole alkaloids (7) and one indoline alkaloid (15) have never been reported before. Their structures were elucidated by analysis of spectroscopic data, including 1D and 2D NMR experiments, UV, IR, and HRESIMS data. The absolute configurations were determined by comparison of the experimental and calculated ECD data. The extracts and isolated compounds were evaluated for their antiprotozoal activity towards Trypanosoma brucei rhodesiense, Trypanosoma cruzi, Leishmania donovani, and Plasmodium falciparum, as well as for their cytotoxicity against rat skeletal myoblast L6 cells. The dichloromethane/methanol (1:1) root extract showed strong activity against P. falciparum (IC50 value of 3.5 µg/mL). Among the compounds isolated, tubotaiwine (13) displayed the most significant antiplasmodial activity with an IC50 value of 8.5 µM and a selectivity index of 23.4. Therefore, P. bicarpallata extract can be considered as a source of indole alkaloids with antiplasmodial activity.


Introduction
Human protozoal diseases cause significant morbidity and mortality. Malaria is one of the most widespread and severe among such diseases. According to a recent report from the World Health Organization (WHO), almost half of the world's population is at risk of malaria, mostly in Africa where the disease is endemic. In 2021, 241 million malaria cases were estimated, and 627,000 people died from it, mainly children under the age of 5 and pregnant women [1].
Since 2006, WHO recommends the use of artemisinin-based combination therapies (ACTs) as the first-line treatment to reduce the risk of resistance. ACTs combine an artemisinin derivative with a longer half-life anti-malarial drug to improve efficacy and reduce the risk of emergence of new resistant strains. Unfortunately, resistance to ACT has recently been reported in several southeast Asian countries [2]. Hence, there is an urgent global need to search for new, safe, more effective and affordable antimalarial drugs.
Pleiocarpa is a genus belonging to the Apocynaceae family, native to tropical Africa and distributed from Senegal to Tanzania and Zimbabwe. This genus consists of six accepted species. Medicinal plants from this chemically under-investigated genus are well known for their use in the traditional treatment of fever, pain, and stomachache [3]. One of them, P. mutica, is used to treat fever and malaria in Ghana. A methanolic extract from the roots showed significant activity with an IC 50 value of 16.7 µg/mL against P. falciparum [4]. Antihypertensive and nematocidal activities have also been reported in previous studies [5]. Various secondary metabolites such as indole, bisindole alkaloids, and triterpenoids have been reported from Pleiocarpa species [3]. The occurrence of indole alkaloids appears to be important in the family and they were proposed as a novel chemical class of antiplasmodial agents [6]. To the best of our knowledge, the chemical and biological constituents of P. bicarpellata Staph. have not been reported so far.
As part of our ongoing research on the discovery of new antiparasitic compounds from African plants, the main purpose of this study was to investigate an extract of P. bicarpellata. The dichloromethane/methanol (1:1) extract from the roots of P. bicarpellata showed a strong antiplasmodial activity with an IC 50 value of 3.5 µg/mL against P. falciparum. This study led to the isolation of sixteen compounds, including seven indole alkaloids, four indoline alkaloids, two secoiridoid glycosides, two iridoid glycosides, and one phenolic glucoside. Among these, one quaternary indole alkaloid and one indoline alkaloid are described here for the first time. The antiprotozoal activities of all the isolated compounds were evaluated, and these results are detailed herein.

In Vitro Antiprotozoal Activity of the Extracts
The in vitro antiprotozoal activity (against T. b. rhodesiense, T. cruzi, L. donovani, and P. falciparum) and cytotoxicity against rat myoblast L6 cells of the P. bicarpellata extracts were evaluated. The extracts were considered inactive when their IC 50 values were >50 µg/mL against the parasites. The dichloromethane/methanol (1:1) extract from the roots demonstrated an activity against the NF54 strain of P. falciparum with an IC 50 value of 3.5 µg/mL without cytotoxicity towards mammalian L6 cells at a concentration of 100 µg/mL ( Table 1). Species of the genus Pleiocarpa are already known to be a rich source of important bioactive compounds such as indole, bisindole alkaloids, and triterpenoids. Indole alkaloids are reported to possess various biological and pharmacological activities such as antihistamine [7], antifungal [8], antimicrobial [9], antioxidant [10], plant growth regulator [11], anti-HIV [12], anticonvulsant [13], anti-inflammatory, cancer chemo-preventive [14], and analgesic [15] properties. Alkaloids isolated from the roots of P. mutica showed potent activity against P. falciparum and are responsible for the strong antiplasmodial activity of the extract. Despite the number of published studies on various Pleiocarpa species, P. bicarpellata has not been extensively studied from a chemical and pharmacological point of view.

LC-ESI-MS/MS Profiling
The extracts were analyzed by ultrahigh-performance liquid chromatography-highresolution mass spectrometry (UHPLC-HRMS/MS) and the data were used to generate a molecular network. Data generated in the positive mode demonstrated a better ionization efficiency, and therefore, were selected for the annotation of metabolites. MzMine [16] was used to process the data and Cytoscape [17] to visualize the molecular network. Various dereplication tools were used: Global Natural Product Social (GNPS) molecular network platform [18] and In-Silico Database of Natural Products (ISDB) [19]. The chemical profiling of the extracts suggested the presence of classes such as alkaloids and iridoid monoterpenes ( Figure S1, Supplementary Materials). The same molecules were putatively identified in the root and stem extracts of P. bicarpellata. Moreover, a fragment ion at m/z 144 [C 10 H 10 N] + indicative of monoterpene indole alkaloids [20] was observed for several compounds. According to the m/z value and MS/MS fragment ion, a node of the cluster at m/z 325.191 was attributed to tubotaiwine (13). It was assumed that the absence of annotation of the other isomers at m/z 325.1928 when dereplicated against all the databases indicated possible new analogues. Kopsinine (2) was observed and annotated for the node at m/z 339.2069 in the same cluster than tubotaiwine (13). Other indole alkaloids with an m/z at 517.2198, 313.1915 and 327.1069, which have never been described in the genus Pleiocarpa and/or for antiparasitic activities, were putatively identified. Therefore, these compounds were targeted for isolation.
The HMBC correlations from H-5, H-6, H-8, and H-15 to C-7 and from O-CH 3 and H-15 to C-2 revealed the indolinic moiety. The presence of an Aspidosperma skeleton [34] was supported by the HMBC correlations from H-15, H-3, and H-19 to C-21 and from H-18 and H-19 to C-20 ( Figure S17). Moreover, the HMBC correlations from O-CH 3 to C-17 and from H-15 to C-16 indicated the attachment of the methoxy group at C-16 ( Figure 2).
The ROEs correlations (Figures 3 and S18) were not sufficient to determine the configuration of the stereogenic carbons C-7, C-15, C-20, and C-21. The absolute configuration of these carbons was determined by comparison of the experimental and calculated ECD data [35]. The experimental spectrum (Figure 4)

Evaluation of the Antiprotozoal Activity
Compounds for which a sufficient amount was available (1-7, 10-13, and 15-16) were evaluated for their in vitro antiprotozoal activity against T. b. rhodesiense, T. cruzi, L. donovani, and P. falciparum, as well as for their cytotoxicity towards L6 cells. The results are summarized in Table 1. Compounds were considered inactive when their IC 50 values were >50 µM against the parasites, except for compound 12 that could not be tested above 10 µM due to the low amount of compound available. Secoxyloganin (16) exhibited antileishmanial activity with an IC 50 value of 25.3 µM against L. donovani and a selectivity index of 11.6. Tubotaiwine (13) was the most active compound with an IC 50 value of 8.5 µM against P. falciparum and a selectivity index of 23.4. To the best of our knowledge, this is the first report on the antileishmanial activity of compound 16 and the antiplasmodial activity of compound 13. In a previous study, secoxyloganin (16) was tested against T. cruzi and did not show any activity (IC 50 > 150 µM) [37], which confirms the results obtained here. Moreover, tubotaiwine (13) was previously reported in the literature for its antileishmanial activity against L. infantum with an IC 50 value of 17.3 µM, and no activity against T. cruzi [38]. Its isomer, (7S,15R,20R,21S)-tubotaiwine (15), showed no antiprotozoal activity (IC 50 > 50 µM), and none of the tested compounds showed toxicity towards L6 myoblast cells. A study revealed that the antiplasmodial activity of Pleiocarpa spp. was due to the presence of alkaloids [4]. Indeed, five alkaloids isolated from the methanol root extract of P. mutica were evaluated against P. falciparum, and pleiomutinine showed significant in vitro antiplasmodial activity with an IC 50 value of 5 µM. Conversely, kopsinine (2) and pleiocarpine (11) were inactive (IC 50 > 200 µM). These results are in accordance with our data. Nevertheless, in an in vivo mouse model, compound 11 was found to be moderately active against P. berghei, where daily doses of 25 mg/kg/day reduced parasitemia by 28.5% compared to untreated control mice [4].

General Experimental Procedures
Optical rotation measurements were performed using a JASCO P-1030 polarimeter (Easton, MD, USA; methanol, c in g/100 mL). The ECD spectra were acquired on a JASCO J-815 CD spectrometer (Easton, MD, USA; methanol). The UV spectra were recorded using a Perkin-Elmer Lambda-25 UV-vis spectrophotometer (Wellesley, MA, USA; methanol). IR spectra were obtained using a Perkin-Elmer Spectrum 100 spectrometer. NMR spectroscopic data were obtained on a Bruker Avance III HD 600 MHz NMR spectrometer equipped with a QCI 5 mm Cryoprobe and a SampleJet automated sample changer (Bruker BioSpin, Rheinstetten, Germany). Chemical shifts (δ) are given in parts per million (ppm) based on the methanol-d 4 signals (δ H 3.31; δ C 49.0) for 1 H-and 13 C-NMR experiments, respectively. Coupling constants (J) are reported in Hertz (Hz). HRMS data were measured on a Q Exactive Focus Hybrid quadrupole-orbitrap mass spectrometer (Thermo Scientific, Waltham, MA, USA) using electrospray ionization in positive-ion mode. UHPLC-PDA-MS measurements were performed using an Acquity UPLC I-class System (Waters, Milford, MA, USA) equipped with an Acquity PDA detector and a Quattro Micro triple quadrupole mass spectrometer (Waters) using an ESI source operating in positive-ion mode. The separation was performed on a Kinetex EVO C 18 UPLC column (100 × 2.1 mm i.d., 1.7 µm) (Waters). The flow rate was set to 0.5 mL/min using a gradient (acetonitrile and water both containing 0.1% formic acid) from 5 to 98% acetonitrile in 15 min. The column was then washed with 98% acetonitrile for 2 min and equilibrated with 5% acetonitrile for 2 min. The injection volume was 2 µL and the column temperature set to 40 • C. The UV absorbance was measured at 210 nm, and PDA absorption spectra were recorded between 190 and 500 nm (1.2 nm steps). Fractionation was done on an Armen Spot preparative chromatographic system (Interchim, Montluçon, France) equipped with a quaternary pump, a fraction collector, and a UV detector. Semi-preparative chromatography was carried out on an Armen Spot System (Saint-Avé, France) using a Kinetex Axia Core-Shell C 18 column (5 µm, 250 × 21.2 mm; Phenomenex, Torrance, CA, USA). Each fraction was analyzed on an Acquity UPLC System (Waters) with an Acquity BEH C 18 column (50 × 2.1 mm i.d., 1.7 µm) (Waters).

Plant Material
Pleiocarpa bicarpellata (Apocynaceae) was collected in Ndjoré at the Centre Region of Cameroon in February 2019 and identified by Victor Nana (Botanist at National Herbarium, Yaoundé, Cameroon). A voucher specimen (N • 30598/HNC) was deposited at the National Herbarium in Yaoundé, Cameroon.

Extraction and Isolation
The roots of P. bicarpellata (500 g) were air-dried, crushed into small pieces, and then powdered and extracted with dichloromethane/methanol (1:1) (3 × 5 L, room temperature, 48 h) to afford 4 g of crude extract. The stems of P. bicarpellata (800 g) were also dried, crushed into small pieces, powdered, and extracted with methanol (3 × 4 L, room temperature, 48 h) to yield 15 g of crude extract.
The methanol stem extract of P. bicarpellata (5 g) was mixed with 14 g of Celite 577 and introduced into a cartridge for a dry load injection. Fractionation was performed using a flash chromatography column (PF-C 18 HQ/300 g, 15 µm C 18 , Interchim) with a linear gradient of 20 to 25% methanol in 40 min and then up to 100% methanol in 10 min.
The flow rate was set to 50 mL/min, and UV detection was performed at 205 nm. The separations yielded 159 fractions that were combined into two fractions, fraction 1 and fraction 2, according to their chromatographic profiles. Fraction 2 (1.7 g) was mixed with 4 g of Celite 577 and introduced into a cartridge for a dry load injection. Fractionation was performed using a flash chromatography column (PF-C 18 HQ/120 g, 15 µm C 18 , Interchim) with a linear gradient of 18 to 25% methanol in 90 min followed by an increase to 100% methanol in 10 min. The flow rate was set to 30 mL/min, and UV detection was performed at 205 nm. The separation yielded 255 fractions that were combined into eight fractions according to their chromatographic profiles (fractions 21-28). Fraction 27 was selected for separation using two flash chromatography columns in series (PF-C 18 HQ/120 g, 15 µm C 18 , Interchim) with a linear gradient of 20 to 40% methanol in 100 min and then up to 100% methanol in 10 min. The flow rate was set to 30 mL/min, and UV detection was performed at 205 nm. The separation yielded 175 fractions that were combined into five fractions according to their chromatographic profiles (fractions 271-275). Fraction 272 was selected for further purification using a semi-preparative HPLC with an X-Select C 18 column (5 µm, 250 × 19.0 mm, Waters) using a linear gradient of 20 to 40% methanol in 70 min followed by an increase to 100% methanol in 30 min. The flow rate was set to 20 mL/min and UV absorbance was measured at 225 nm. This yielded (7S,15R,20R,21S)-tubotaiwine (15, 1.2 mg) and secoxyloganin (16, 2.0 mg).

MS Data Treatment, Molecular Network Generation, and Annotation
Thermo .RAW files were converted into .mzXML (mass spectrometry data format) using MSConvert software.36, part of the Proteowizard package (ProteoWizard, Palo Alto, CA, USA) [39]. The converted files were uploaded to MZmine software suite version 2.53 [16]. For mass detection at MS 1 level, the noise level was set to 1.0 × 10 6 . For MS 2 detection, the noise level was set to 0.00. The ADAP chromatogram builder parameters were set as follows: minimum group size of scans, 5; minimum group intensity threshold, 1.0 × 10 6 ; minimum highest intensity, 1.0 × 10 6 and m/z tolerance of 8.0 ppm. The ADAP algorithm (wavelets) was used for chromatogram deconvolution with the following parameters: S/N tolerance, 25; minimum feature height, 1.0 × 10 6 ; coefficient area threshold, 100; peak duration range, 0.02-1.0 min; RT wavelet range, 0.02-0.08 min. Isotopes were detected using the isotope peak grouper with a m/z tolerance of 8.0 ppm, a RT tolerance of 0.02 min (absolute), the maximum charge set at 1, and the representative isotope used was the most intense. Each file was filtered by retention time within a range from 0.70 to 8.00 min, and only the ions with an associated MS 2 spectrum were kept. Alignment was done with the join-aligner (m/z tolerance, 8.0 ppm; RT tolerance, 0.05 min) comparing the spectral similarity (spectral tolerance, 8.0 ppm; MS level, 2; Weighted dot-product cosine with default parameters). The resulting aligned peak list was exported for further analysis as a .mgf file.
The .mgf file was exported from MZmine to build the molecular network, using the online workflow (https://ccms-ucsd.github.io/GNPSDocumentation/) on the GNPS website (http://gnps.ucsd.edu). The data were clustered with the following parameters: precursor ion mass tolerance: 0.02 Da; MS/MS fragment ion tolerance: 0.02 Da; mini-mum cosine score: 0.7; and minimum matched peaks: 6. The spectra in the network were then searched against the spectral libraries of GNPS. The library spectra were filtered in the same manner as the input data. The required library matches were set to show a score above 0.6 and at least 3 matched peaks. The job can be found here: https: //gnps.ucsd.edu/ProteoSAFe/status.jsp?task=56ea2e86598848a5b6750a445be60c74. The output of the GNPS platform was compared against an in silico database to extend the rate of putative annotations [40]. This output was subjected to taxonomically informed metabolite annotation [19] to re-rank and clean up the output based on the taxonomy of the collection. The in silico database used for all this process includes the combined records of the Dictionary of Natural Products (https://dnp.chemnetbase.com/) and Lotus (https://lotus.naturalproducts.net/) [41].

ECD Computational Details
The absolute configuration of compounds 7 and 15 was assigned according to the comparison of the calculated and experimental ECD spectra. Based on the structure proposed by NMR experiments, conformers were generated using the MMFF94s force field with Spartan Student v7 (Wavefunction, Irvine, CA, USA). From the results obtained, the 10 isomers with the lowest energy were subjected to further successive PM3 and B3LYP/6-31G(d,p) optimizations with Gaussian 16 software (Gaussian Inc., Wallingford, CT, USA) using the CPCM model in methanol. All optimized conformer outputs were checked to avoid imaginary frequencies after each optimization. A cut-off of 4 kcal/mol was set as maximum difference between conformers. The remaining conformers were submitted to Gaussian 16 software for ECD calculations, using B3LYP/def2svp as basis set with the CPCM model in methanol. The computation in Gaussian was performed at the University of Geneva on the Baobab cluster (https://plone.unige.ch/distic/pub/hpc/baobab_en). The calculated ECD spectra were generated in SpecDis1.71 software (Berlin, Germany) based on a Boltzmann weighing average.