Targeted Isolation of Anti-Trypanosomal Naphthofuran-Quinone Compounds from the Mangrove Plant Avicennia lanata

The discovery of new secondary metabolites from natural origins has become more challenging in natural products research. Different approaches have been applied to target the isolation of new bioactive metabolites from plant extracts. In this study, bioactive natural products were isolated from the crude organic extract of the mangrove plant Avicennia lanata collected from the east coast of Peninsular Malaysia in the Setiu Wetlands, Terengganu, using HRESI-LCMS-based metabolomics-guided isolation and fractionation. Isolation work on the crude extract A. lanata used high-throughput chromatographic techniques to give two new naphthofuranquinone derivatives, hydroxyavicenol C (1) and glycosemiquinone (2), along with the known compounds avicenol C (3), avicequinone C (4), glycoquinone (5), taraxerone (6), taraxerol (7), β-sitosterol (8) and stigmasterol (9). The elucidation and identification of the targeted bioactive compounds used 1D and 2D-NMR and mass spectrometry. Except for 6–9, all isolated naphthoquinone compounds (1–5) from the mangrove plant A. lanata showed significant anti-trypanosomal activity on Trypanosoma brucei brucei with MIC values of 3.12–12.5 μM. Preliminary cytotoxicity screening against normal prostate cells (PNT2A) was also performed. All compounds exhibited low cytotoxicity, with compounds 3 and 4 showing moderate cytotoxicity of 78.3% and 68.6% of the control values at 100 μg/mL, respectively.


Introduction
Mangrove plants as well as their endophytic fungi exhibit unique chemical diversity from various classes of compounds with promising biological activities [1][2][3]. Avicennia is the only mangrove genus belonging to the Avicenniaceae family; it is the most abundant genus in mangrove ecosystems and is widely distributed on tropical and subtropical coastlines. Eight to ten species have been recorded worldwide [4]. Avicennia lanata (synonym: A. rumphiana), locally known in Malaysia as 'Api-api bulu', is found mainly in sandy or firm silt substrate of middle to higher intertidal zones [5]. It is native and common throughout much of Peninsular Malaysia, Philippines and New Guinea. This tree is identified by its furry underside leaves and fruit. The pelt ('bulu' in Malay) on the leaves conserves water by trapping a layer of insulating air, thus reducing water loss through evaporation. The tree or shrub can grow up to 20 m tall; the bark is dark brown to black, warty or smooth, with pneumatophores 20-30 cm tall. The first phytochemical study on Avicennia sp. revealed

Dereplication Study on A. lanata Total Crude Extract
The total ion chromatogram of the crude extract of A. lanata (Figure 1) showed the distribution of known and unknown compounds present in the total extract ( Table 1). Some of the putatively identified compounds have been previously isolated from Avicennia sp. (Table 1). The values and predicted formulas of unknown compounds are also shown in Table 1, indicated with "NO HITS" found. The dereplication studies revealed that the plant extract possessed certain types of compounds, such as alkaloids, triterpenes and naphthoquinones, which have also previously been isolated from different Avicennia species, particularly A. marina and A. alba.
The total A. lanata crude extract was fractionated by using medium pressure liquid chromatography with gradient elution using hexane-ethyl acetate-methanol yielding nine major fractions. To proceed with efficient targeted isolation work of the active metabolites, nine fractions were preliminarily screened against T. b. brucei and subjected to HRESI-LCMS prior to multivariate analysis. A dereplication study was performed to obtain the metabolomic profile of each fraction.
Each fraction was screened at different concentrations of 20, 10 and 5 µg/mL. The assay was performed in duplicate for each sample. The results show the percentage growth of T. b. brucei (Table 2), with negative readings representing those with higher growth inhibition on the trypanosomal cells. The A. lanata crude extract showed marginal anti-trypanosomal activity, whereas after fractionation, the activity for fractions F5 to F8 increased, and decreased in fraction F9. Fractions F1 and F2 showed very weak anti-trypanosomal activity whereas fractions F3, F4, and F9 showed moderate activity. Meanwhile, fractions F5 until F8 showed significant activity in this screening test, thereby supporting further investigation of the biologically active compounds from this plant extract.
The relationship between the occurrence of the metabolites in the different A. lanata fractions and their bioactivity against T. b. brucei were evaluated through multivariate analysis. The unsupervised Principal Component Analysis (PCA) scores plot showed moderate separation of the A. lanata fractions ( Figure 2A). There was a clear separation between the bioactive fraction F5 and the other fractions that were also active against T. b. brucei. Fraction F4, which also possessed moderate activity, was likewise set apart from other fractions ( Figure 2A). Meanwhile, a supervised multivariate OPLS-DA scores plot analysis ( Figure 2B) exhibited two classes -fractions F1, F2, F3 and F9 were the inactive group and clustered very close together, while fractions F4, F5, F6, F7 and F8 were the active group, with fraction F5 being segregated from the rest of the active cluster. The segregation of F5 may indicate the presence of unique chemistry in F5. The OPLS-DA loadings plot ( Figure 2C) Table 1. Boxed in green are the isolated compounds.  Table 1. Boxed in green are the isolated compounds. ∆ As the number of "HITS" is mostly greater than 20, sources were filtered to the Genus Avicennia or biological sources only widely distributed in Southeast Asia and their marine-derived or associated endophytes that could include both bacteria and/or fungi as well as for the presence of a benzofuran unit for compounds with DBEs between 7 and 10. Albeit the number of "HITS" is less than 20 for a couple of ion peaks (e.g., P_6110), the biological source has a lower distribution to be found in the region of the collected material, then there is a higher probability that the detected metabolite could be a new natural product, which in this case cannot be properly validated. MS fragmentation interpretation was inconclusive for compounds with more than 20 HITS. For the benzofuran analogues, the neutral loss of several water units could be observed. The identification of the occurrence of respective compounds could only be validated with a reference Avicennia extract with known standard analogues in the database and by NMR after the initial fractionation. As for this instance, the dereplication result for P_3632 and P_3702 were then later described to be another compound after NMR analysis of the purified bioactive secondary metabolites. For this study, targeted isolation was only done on the predicted bioactive metabolites. *: Isolated bioactive compounds marked with * corresponding compound number used in the text. a % D, percentage viability of control (at 20 µg/mL, 10 µg/mL, 5 µg/mL) were determined by averaging of two independent assays (n = 2); suramin as positive control.
The OPLS-DA S-loadings plot ( Figure 2D) exhibited the discriminating metabolites at the end points of both the active and inactive groups, respectively. From the DNP database, it was putatively determined that fractions F1 and F2 contained mostly fatty acid oils, while fractions F3 and F4 comprised of terpenoid and sterol metabolites. Metabolites from fractions F5 to F9 were putatively identified to have mainly aromatic compounds, compounds perhaps contributing towards the activity of the fractions. Some of the ion peaks in the active fractions have been dereplicated as presented in Table 1 The aim of this study was to isolate compounds from the active fractions that were responsible for anti-trypanosomal activity. Using metabolomics and bioassay-guided isolation work to search for new bioactive secondary metabolites against the protozoan T. b. brucei, the organic crude extract of A. lanata afforded two new metabolites, (1-2) along with the known compounds (3-9) ( Figure 3) after a series of chromatographic techniques. The absolute elucidation and identification of these metabolites were achieved by using 1D and 2D-nuclear magnetic resonance (NMR) and high-resolution mass spectrometry (MS). The isolated pentacyclic triterpenes and sterols: taraxerone (6) taraxerol (7), β-sitosterol (8) and stigmasterol (9) were not initially detected from the crude extract by LCMS but were only evidenced from the NMR spectra of the extracts and the non-polar fractions, which were instead confirmed by GCMS. Meanwhile, the relative occurrence of the targeted metabolites in the bioactive fractions as well as their absence or lower abundance in the inactive fractions is shown in Figure 4.           Meanwhile, the known naphthofuranquinone deivatives avicenol C (3) and avicequinone C (4) were also isolated in this study. The structures of the molecules were confirmed by 1D and 2D-NMR as well as comparison with previous literature, namely as avicenol C and avicequinone C, respectively.
Compound 2 (6.0 mg) was isolated as a yellow oil and its molecular formula was determined by HRESI-MS ( Figure 6A Table 3) showed an ABCD spin system with proton signals at δ H 8.12 (d, J = 7.9 Hz, 1H), 7.72 (d, J = 7.8 Hz, 1H), 7.58 (t, J = 7.6 Hz, 1H) and 7.44 (t, J = 7.6 Hz, 1H) corresponding to protons H-8, H-5, H-6 and H-7, respectively. Two methyl signals located on aliphatic C-4 and C-5 remained the same to the known derivative glycoquinone (5). As in compound 5, the presence of a prenyl moiety was indicated by one methylene proton signal at δ H 2.98 (m, 2H, H-1 ) that correlated to a downfield proton at δ H 4.77 (dd, J = 8.6, 6.3 Hz, H-2 ) and two proton methyl singlets at δ H 1.49 (s, 3H, H-4 ) and 1.54 (s, 3H, H-5 ). Compound 2 differed to that of 5 by the incidence of a downfield methylene proton at δ H 3.00 (m, 2H, H-1 ) positioned on the dihydrofuran ring. The methylene proton was split to give a multiplet due to the unsaturation at carbons ∆ 3,4 and the anisotropy effect between C-3 and C-4. A methine proton signal linked to oxygen was indicated by a doublet of a doublet signal (dd, J = 9.0, 4.3 Hz, H-2 ) at δ H 3.93.
The 13 C NMR results (Table 4) showed 20 carbon signals, including two carbonyl carbon signals at δ C 181.8 (C-1) and 145.6 (C-4) which was different compared with the carbon signal in compound 1. This compound was proposed to have a carbonyl group attached at carbon position C-1. The remaining aromatic carbon signals were shown by three quaternary carbon signals at δ C 128.9, 142.0, and 112.6, which corresponded to C-4a, C-8a and C-3, respectively. The carbon signals for prenyl moiety were observed as a methylene carbon signal at δ C 29.6 (C-1 ), olefinic carbon signals at δ C 117.4 (C-2 ) and 138.6 (C-3 ) along with two methyl carbon signals at δ C 18.7 (C-4 ) and 26.3 (C-5 ). The remaining aliphatic carbon signals were found similar to 5.
The molecular structure was also established by 2D-NMR using COSY and HMBC experiments. Some of the important COSY correlations were between protons H-1 and H-2 and amongst the aromatic protons. The HMBC showed important C-H correlations between H-8 and the carbonyl carbon C-1 as well as methine carbons C-7 and C-8a. Further strong correlations were observed between H-1 and C-3 , H-7 and C-8/8a, and between the prenyl proton H-2 and C-4 /5 . The COSY and HMBC correlations were depicted in Figure 6B. The NMR spectra revealed that this compound was different only at C-4 in which the carbon carried a hydroxyl substituent hence affecting the splitting pattern of the aromatic protons when compared with compound 5 that has been described in Glycosmis pentaphylla [8]. Therefore, compound 2 was proposed to be a new semihydroquinone derivative of 5 and was assigned the trivial name (+)-glycosemiquinone. Glycosmis pentaphylla (Fam Rutaceae) is widely distributed in low altitude tropical forests of India, south China, Thailand, peninsular Malaysia, Indonesia and the Philippines Islands [38]. Interestingly, naphthofuranquinone compounds earlier reported for Avicennia species that included avicenol B and avicequinone C have also been isolated from the Glycosmis pentaphylla [8].
The molecular structure was also established by 2D-NMR using COSY and HMBC experiments. Some of the important COSY correlations were between protons H-1′ and H-2′ and amongst the aromatic protons. The HMBC showed important C-H correlations between H-8 and the carbonyl carbon C-1 as well as methine carbons C-7 and C-8a. Further strong correlations were observed between H-1′ and C-3′, H-7 and C-8/8a, and between the prenyl proton H-2″ and C-4″/5″. The COSY and HMBC correlations were depicted in Figure 6B. The NMR spectra revealed that this compound was different only at C-4 in which the carbon carried a hydroxyl substituent hence affecting the splitting pattern of the aromatic protons when compared with compound 5 that has been described in Glycosmis pentaphylla [8]. Therefore, compound 2 was proposed to be a new semihydroquinone derivative of 5 and was assigned the trivial name (+)glycosemiquinone. Glycosmis pentaphylla (Fam Rutaceae) is widely distributed in low altitude tropical forests of India, south China, Thailand, peninsular Malaysia, Indonesia and the Philippines Islands [38]. Interestingly, naphthofuranquinone compounds earlier reported for Avicennia species that included avicenol B and avicequinone C have also been isolated from the Glycosmis pentaphylla [8].
However, the ability of a hydroxyl moiety to decrease the bioactivity might slightly be different to the effect of a methoxy group. The presence of the hydroxyl groups, instead of the quinoid group in (−)-hydroxyavicenol C (1) had decreased the trypanocidal activity with the MIC value of 12.50 µM. There was no difference in the anti-trypanosomal activity of new derivative, (+)-glycosemiquinone (2) and glycoquinone (5) as the MIC for both compounds was 12.50 µM. The new derivative (2) with a ketone at C-1 and a hydroxyl at C-4 showed the same activity as the parent analogue 5. It showed that the presence of the prenyl group on both 2 and 5 decreased the activity against the protozoa. All naphthofuranquinone derivatives showed weak toxicities against normal prostate cells (PNT2A).
Meanwhile, the pentacyclic triterpenes-taraxerone (6) taraxerol (7), and β-sitosterol (8) were inactive against T. b. brucei with MICs of 154.20, 145.30 µM and 142.30 µM, respectively (Table 5). Stigmasterol (9) had a very low activity with MIC values of 126.40 against the protozoa. Two pentacyclic triterpenes possessing the same taraxerane-type skeleton, 6 and 7, have been isolated from the mangrove A. lanata, and differ only at C-3 where taraxerone has a carbonyl and taraxerol a β-OH. A previous study showed that β-amyrin had poor activity against bloodstream strain of T. b. brucei with an IC 50 of 126.9 µM [41]. The molecular structure of β-amyrin is similar to 6, the only difference being a shift in the double bond from ∆ 14,15 to ∆ 12,13 . Taraxerol has also been isolated from the Ecuadorian plant Cupania cinerea and showed IC 50 values of <10 µM against T. b. rhodesiense in vitro bloodstream trypomastigotes with low cytotoxicity [42]. A study on the activity of 6 and 7, which have been isolated from the bark of Cupania dentata (Sapindaceae), against flagellate protozoan Giardia lamblia trophozoites, found that both have potential giardicidal activity with IC 50 values of 26.7 and 37.8 µM, respectively [43]. The compounds also exhibited anti-plasmodial (Plasmodium falciparum), analgesic [44] and anti-inflammatory [45] activities. Other biological activities showed that 6 and 7 were also allelopathic [46] and displayed anti-fungal activity [47]. Meanwhile, the only difference between compound 8 and 9 is the presence of a double bond on the side chain at ∆ 22, 23 . Compound 8 which possessed a saturated side chain was inactive against T. b. brucei with an MIC value of 142.30 µM. However, compound 9 which has a double bond on the side chain showed slightly increased activity compared with 8, but still had a very low inhibitory effect on T. b. brucei (MIC of 126.40 µM).

Plant Materials
The twigs of Avicennia lanata were collected from Setiu Wetlands, Terengganu, Malaysia with the help of Mr Muhamad Razali Salam. The specimen was deposited in the Universiti Malaysia Terengganu herbarium with the voucher specimen code UMT-01.

General Experimental Procedures
The structural determination of the isolated compounds was based on MS and NMR spectroscopy data. One dimensional NMR (1D-NMR) data consisted of 1 H and 13 C NMR spectra captured using Jeol ( 1 H 400 MHz, 13 C 100.5 MHz, SIPBS, University of Strathclyde) and Bruker instruments ( 1 H 600 MHz, 13 C 150 MHz, Department of Pure and Applied Chemistry, University of Strathclyde) and was confirmed by two-dimensional NMR (2D-NMR) spectra such as HMQC or HSQC, HMBC, COSY and NOESY as well as comparisons with the literature. A pure sample was dissolved in 500 µL of a suitable deuterated solvent and transferred to 5 mm Norell NMR tube (NORELL Inc., Morganton, NC, USA). Samples that were low in quantity were analysed in Shigemi tubes (SHIGEMI Co., LTD., Hachioji City, Japan) with 180 µL of the appropriate deuterated solvent. Dimethyl sulfoxide-d 6 , chloroform-d, acetone-d 6, and methanol-d 3 bought from Sigma-Aldrich (USA) were the deuterated solvents used. The spectra were then processed with MestReNova-9.0 (MNova) 2.10 (Mestrelab Research, S.L, Santiago de Compostela, Spain). The optical rotation of the optically active compounds was measured with the digital polarimeter 341 (PerkinElmer Life and Analytical Sciences, Shelton, CT, USA) in which the pure compound was dissolved in in 2 mL of the suitable solvent (chloroform or acetone) to a concentration of 1 mg/mL.

Dereplication by Using HRESI-LCMS
The dereplication study on the total crude extract and fractions of the samples were performed using HRESI-LCMS and processed with the MZmine software [48][49][50], an in-house macro coupled with the Dictionary of Natural Products (DNP) 2015 and AntiMarin 2012, a combination of Antibase and MarinLit, and SIMCA 15 (Umetrics AB, Umeå, Sweden). The procedure and program for HRESI-LCMS was set up as described below. The total crude extract of 1 mg/mL in methanol was analysed on an Accela HPLC (Thermo Fisher Scientific, Waltham, MA, USA) coupled with a UV detector at 280 and 360 nm and an Exactive-Orbitrap high-resolution mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). A methanol blank was also analysed. The column attached to the HPLC was a HiChrom, ACE (Hichrom Ltd, Lutterworth, UK) C18, 75 mm × 3.0 mm, 5 µm column. The mobile phase consisted of micropore water (A) and acetonitrile (B) with 0.1 % formic acid for each solvent. The gradient program started with 10% B linearly increased to 100 % B within 30 min at a flow rate of 300 µL/min and remained isocratic for 5 min before linearly decreasing back to 10% B in 1 min. The column was then re-equilibrated with 10% B for 9 min before the next injection. The total analysis time for each sample was 45 min. The injection volume was 10 µL and the tray temperature was maintained at 12 • C. High-resolution mass spectrometry was carried out in both positive and negative ESI ionization switch modes with a spray voltage of 4.5 kV and capillary temperature at 320 • C. The mass range was set from m/z 150-1500 for ESI-MS range.
The mass spectral data was processed using the procedure by MacIntyre et al., [50] which was established in the Natural Products Metabolomics Group Laboratory at SIPBS as described here [35,50,51]. The LC-MS chromatograms and spectra were viewed using Thermo Xcalibur 2.1 or MZmine 2.20.

Extraction, Fractionation and Isolation of Metabolites from A. Lanata
The dried powdered twigs (4 kg) of A. lanata were macerated in methanol overnight (8L/extraction, 3×,) and the methanol extract was concentrated under vacuum using a rotary evaporator (Büchi Labortechnik AG, Flawil, Switzerland) to give 44.8160 g. The methanol extract was partitioned by liquid-liquid extraction three times with equal volumes of ethyl acetate to the aqueous phase (90% water + 10% methanol), to give an organic phase which was then concentrated under vacuum by a rotary evaporator (Büchi Labortechnik AG, Flawil, Switzerland) to afford the crude ethyl acetate extract weighing 14.4115 g. Thin layer chromatography (TLC) analysis was carried on A. lanata crude extract. The total crude extract was dissolved in any suitable solvent, mixed with Celite S (Merck KGaA, Darmstadt, Germany) then fractionated by medium pressure liquid chromatography (Büchi Labortechnik AG, Flawil, Switzerland) through gradient elution commencing with 100% hexane to 100% ethyl acetate for 20 min, followed by 100% ethyl acetate to 30% ethyl acetate and 70% methanol for 30 min at a flow rate of 50 mL/min. A VersaFlash silica column (Supelco Inc, Bellefonte, PA, USA) with dimensions of 4 × 150 mm and a particle size of 20-45 µm was used. The fraction collection volume was set at 100 mL/tube. TLC was carried out to monitor the separation profiles of the 25 fractions and similar fractions were pooled together. The pooled fractions were concentrated under vacuum by a rotary evaporator to give nine major fractions and were analysed using HRESI-MS for dereplication study and tested for anti-trypanosomal activity. The active major fractions were subjected to further isolation and purification using conventional gravity column or by Reveleris (W. R. Grace & Co.-Conn, Columbia, MD, USA) and Isolera One (Biotage AB, Uppsala, Sweden) high-throughput flash chromatography which can be either normal or reverse phase fitted with the respective commercially available pre-packed column either from Reveleris USA or SNAP Sweden, respectively. The two flash chromatography instruments were used to isolate and purify the active fractions or small quantity of the crude extracts. The non-UV active metabolites were purified by using Grace Reveleris, since dual detectors were coupled to the instrument. Meanwhile, the UV active metabolites were purified by using Biotage, since this instrument able to detect the UV active compounds in the 200-400 nm range. Open column chromatography was used with various column sizes and silica gel 60 (Kieselgel 60), 0.035-0.070 mm (220-440 mesh ASTM) (Alfa Aesar, Haverhill, MA, USA).

Bioassays
MIC assays were performed for samples having greater than 90% inhibition at a concentration of 20 µM in the initial screening campaign. The MIC values of the isolated compounds against T. b. brucei were determined by averaging the results of two independent assays. The concentrations were averaged and converted to µM.
Meanwhile, the initial screening for cytotoxicity activity of the isolated compounds was performed on human normal prostatic epithelial cells (PNT2A) at 100 µg/mL. The % D of control was determined by averaging the results of three independent assays. In the initial screening, if the cell viability is less than 60%, a concentration response test at 0.3 to 300 µg/mL was carried on.

Anti-Trypanosomal Assay
The activity of the pure compounds was tested in vitro against the blood stream form of Trypanosomal brucei brucei (T. b. brucei) S427. The activity of the samples was determined using the well-established Alamar blue™ 96-well microplate assay, in which the screening procedure was modified from the microplate Alamar blue assay [52], to determine the drug sensitivity of African trypanosomes. The samples were initially screened at one concentration (20 µg/mL crude extracts or fractions, 20 µM for pure compounds) to determine their in vitro activity. Stock solutions of tested samples in DMSO (Acros Organics BVBA, Janssen Pharmaceutical, Geel, Belgium) were prepared at concentrations of 10 mg/mL (extracts) or 10 mM for pure compounds. The concentration of DMSO should not exceed 0.5% of the final test solution.

Cytotoxicity Assay
The pure compounds from mangrove plant, A. lanata were tested for cytotoxicity in vitro against human normal prostatic epithelial cell line (PNT2A) derived from ECACC (Sigma-Aldrich, Dorset, UK).
The cytotoxicity activity of the pure compounds was determined using the well-established Alamar Blue™ redox-based 96-well microplate assay, in which the screening procedure was modified from [53]. PNT2A cells were seeded into 96-well microplate assay at density of 0.5 × 10 4 cells/well in 100 µL of Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Paisley, UK) and incubated at 37 • C, 5% CO 2 with a humidified atmosphere for 24 h. The tested compound was prepared at desired concentrations in DMEM solutions, while DMSO and Triton X as a negative and positive controls, respectively, were added into the microplate to give a total volume of 200 µL. The microplate plate was incubated at 37 • C, 5% CO 2 with a humidified atmosphere for 24 h, then 10 µL of Alamar blue was added. The microplate was further incubated for another 20 h, then the fluorescence was measured using a Wallac Victor 2 microplate reader (Perkin Elmer, Cambridge, UK) (excitation: 530 nm, emission 590 nm). The results were calculated as % of the DMSO control values.

Conclusions
The main aims in the present work were to isolate secondary metabolites for anti-trypanosomal derived drugs from the mangrove plant, Avicennia lanata by utilising metabolomics and bioassays-guided approaches to aid in the preliminary screening, fractionation and purification of the targeted bioactive compounds. To achieve the goal of the discovery for new potential active secondary metabolites, the use of metabolomics tools assisted in the decision making of which fractions with targeted bioactivity should be prioritized for further isolation work. By means of high-resolution liquid chromatography-mass spectrometry, the crude fractions obtained from the A. lanata crude extract were preliminarily screened for bioactive molecules against T. b. brucei and were analysed by multivariate analysis such as PCA and OPLS-DA. A dereplication study was used to screen the known metabolites and predict the novelty of metabolites from the crude fractions prior to purification work to avoid repetitive work with the same bioactivity. Metabolomics and bioassay-guided isolation of potential anti-trypanosomal secondary metabolites were identified from the crude extracts of the mangrove plant A. lanata, which included two new naphthofuranquinone derivatives, hydroxyavicenol C and glycosemiquinone along with seven known compounds that included naphthoquinones, triterpenes and sterols. The naphthofuranquinone derivatives were active against the protozoa, T. b. brucei while the triterpenes and sterols were found inactive.