Prenylated Isoflavanones with Antimicrobial Potential from the Root Bark of Dalbergia melanoxylon

Dalbergia melanoxylon Guill. & Perr (Fabaceae) is widely utilized in the traditional medicine of East Africa, showing effects against a variety of ailments including microbial infections. Phytochemical investigation of the root bark led to the isolation of six previously undescribed prenylated isoflavanones together with eight known secondary metabolites comprising isoflavanoids, neoflavones and an alkyl hydroxylcinnamate. Structures were elucidated based on HR-ESI-MS, 1- and 2-D NMR and ECD spectra. The crude extract and the isolated compounds of D. melanoxylon were tested for their antibacterial, antifungal, anthelmintic and cytotoxic properties, applying established model organisms non-pathogenic to humans. The crude extract exhibited significant antibacterial activity against Gram-positive Bacillus subtilis (97% inhibition at 50 μg/mL) and antifungal activity against the phytopathogens Phytophthora infestans, Botrytis cinerea and Septoria tritici (96, 89 and 73% at 125 μg/mL, respectively). Among the pure compounds tested, kenusanone H and (3R)-tomentosanol B exhibited, in a panel of partially human pathogenic bacteria and fungi, promising antibacterial activity against Gram-positive bacteria including methicillin-resistant Staphylococcus aureus (MRSA) and Mycobacterium showing MIC values between 0.8 and 6.2 μg/mL. The observed biological effects support the traditional use of D. melanoxylon and warrant detailed investigations of its prenylated isoflavanones as antibacterial lead compounds.


Introduction
The genus Dalbergia L.f. (Fabaceae) consists of approximately 274 species distributed in the tropics and subtropics. Among these, eight species naturally occur in Kenya [1,2]. Plants of this genus vary from shrubs and lianas to small trees referred to as rosewoods (e.g., D. odorifera T.C. Chen., D. latifolia Roxb. and D. melanoxylon Guill. & Perr.) due to their fine timber of high economic value [3]. The genus is extensively used in the

Extraction and Isolation
The ground root bark of D. melanoxylon (1.6 kg) was macerated in a 1:1 mixture of CH 2 Cl 2 and MeOH to yield a gummy extract (95.7 g). The crude extract was partitioned between CH 2 Cl 2 and H 2 O. After removal of the organic solvent 70.6 g CH 2 Cl 2 extract was obtained. A portion of this extract (51.3 g) was subjected to column chromatography on silica gel (600 g, 80 × 4 cm) eluting with n-hexane containing increasing amounts of EtOAc.
The results (mean value ± standard deviation, n = 6) were given as relative values (% inhibition) in comparison to the negative control (bacterial growth, 1% DMSO, without test compound). Negative values indicate an increase of bacterial growth. Calculations were performed applying the software Excel.

Antifungal Assays
The assays were performed according to the monitoring methods approved by the fungicide resistance action committee (FRAC) with minor modifications [32]. The phytopathogenic ascomycetes Botrytis cinerea Pers. and Septoria tritici Desm., and the oomycete Phytophthora infestans (Mont.) De Bary were used as test microorganisms. The crude extract and pure compounds were tested in 96-well microtiter plate assays at 125 and 42 µg/mL with DMSO used as negative control (max. concentration 2.5%), while epoxiconazole (100% inhibition at 42 µM) and terbinafine (67% inhibition at 42 µM) served as positive control. Five to seven days after inoculation, pathogen growth was evaluated by measurement of the optical density (OD) at λ 405 nm with a TecanGENios Pro microplate reader (5 measurements per well using multiple reads in a 3 × 3 square). Each experiment was carried out in triplicate.

Anthelmintic Assay
The anthelmintic bioassay was performed using the model organism Caenorhabditis elegans that previously was shown to correlate with anthelmintic activity against parasitic trematodes as described by Thomsen et al. [33]. The Bristol N2 wild-type strain of C. elegans was obtained from the Caenorhabditis Genetic Center (CGC), University of Minnesota, Minneapolis, USA. The nematodes were cultured on NGM (Nematode Growth Media) Petri plates using the uracil auxotroph E. coli strain OP50 as food source. In this assay, the solvent DMSO (2%) and the standard anthelmintic drug ivermectin (10 µg/mL, 100% dead worms after 30 min incubation) were used as negative and positive control, respectively.

Cytotoxicity Assay
Briefly, for the cytotoxicity assay, the human prostate cancer cell line PC-3 and the colon adenocarcinoma cancer cell line HT-29 (both from ATCC, Manassas, VA, USA) were used. The cell handling and assay techniques were in accordance with the method described by Khan et al. [34]. The extract was tested at the concentrations of 0.05 and 50 µg/mL. Anti-proliferative and cytotoxic effects of the extract were investigated by performing colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) and CV (crystal violet)-based cell viability assays (Sigma-Aldrich, Taufkirchen, Germany) after 48 h treatment time, respectively. The absorbance was measured with an automated microplate reader at 540 nm with a reference wavelength of 670 nm. Digitonin (125 µM) was used as positive control, which was set for data normalization to 0% cell viability. The results are presented as a percentage of control values obtained from untreated cultures.

Agar Diffusion Assay
The experiment was performed as previously published [35]. Briefly, test compounds were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 1 mg/mL. Ciprofloxacin and amphotericin B (both positive control) were provided at 5 µg/mL and 10 µg/mL, respectively. The following test strains were used: B. subtilis

Cytotoxicity Testing (Compound 7)
HeLa cells (DSM ACC 57) were grown in RPMI 1640 medium supplemented with 10 mL/L ultraglutamine 1 (CAMBREX 17-605E/U1), 550 µL/L gentamicin sulfate (50 mg/mL, CAMBREX 17-518Z) and 10% heat inactivated fetal bovine serum (GIBCO Life Technologies 10270-106) at 37 • C in a 5% CO 2 atmosphere in high density polyethylene flasks (NUNC 156340). Cells were pre-incubated for 48 h in the absence of test substances. Subsequently, HeLa cells were incubated with serial dilutions of test compounds in 96 well microplates for 72 h at 37 • C in a humidified atmosphere and 5% CO 2 . After incubation, the cytolytic effect of compounds was analyzed relative to the negative control (DMSO) using a colorimetric assay (methylene blue). The adherent HeLa cells were fixed by glutaraldehyde (MERCK 1.04239.0250) and stained with a 0.05% solution of methylene blue (SERVA 29198) for 15 min. After gentle rinsing, the stain was eluted through addition of 0.2 mL hydrochloric acid (0.33 M) to each well. The absorptions were measured at 660 nm in a SUNRISE microplate reader (TECAN). Four replicates were assayed for each substance. The half-cytotoxic concentration (CC 50 ) was defined as the test compound concentration required for 50% reduction of the viable cell count in the monolayer relative to the respective untreated control. All calculations of CC 50 values were performed with the software Magellan (TECAN).

Isolation and Structure Elucidation
Chromatographic separation of the extract from the root bark of D. melanoxylon afforded six hitherto-undescribed isoflavanones (1-6) alongside eight known secondary metabolites comprising isoflavonoids (7-10), neoflavones (11)(12)(13) and alkyl hydroxylcinnamates (14) (Figure 1). Based on HRESIMS, NMR and ECD spectra and comparison to published data, the known compounds were identified as kenusanone H (7; (11) from Dalbergia sissoo [37], dalbergin (12) from D. odorifera [38], melannein (13) from D. melanoxylon [28] and a mixture of cinnamic acid esters with the main compound being 3 ,4 -dihydroxyl-trans-cinnamic acid octacosyl ester (14) known from Gliricidia sepium [39]. With the exception of compounds 12 and 13, all known compounds were isolated for the first time from D. melanoxylon. For tomentosanol B (9) so far only the planar structure based on 1 H NMR data was described [18]. Herein we report its 13 C ( Table 2) and 2D NMR data ( Figures S9_3 and S9_4, Table S9). Based on ECD measurements ( Figure S9_6) the configuration at C3 was determined as R and compound 9 thus elucidated as (3R)-6-prenyl-3,4 ,5,7-trihydroxyl-2 -methoxy-3 -prenyl-isoflavanone (trivial name (3R)-tomentosanol B).  [16,40]. In addition, the NMR also exhibited signals of a methoxy (δ H 3.58, δ C 62.0) and a 3-hydroxylmethyl-3-methylbut-2-enyl [δ H 3. These observations were further supported with 2D spectra which showed cross-peaks in the 1 H-1 H COSY spectrum between H-6 and H-8 in the A-ring, and between H-5 and H-6 in the B-ring. Analysis of the 13 C-NMR spectrum indicated, in accordance with the molecular formula, the presence of 21 carbons with resonances ranging from δ C 195.7 (sp 2 hybridized ketone) to 12.6 (sp 3 hybridized methyl unit). The 13 C-NMR chemical shift of the deshielded methoxy group signal (δ C 62.0) indicated that it is di-ortho-substituted with two bulky groups, which is consistent with its placement at C-2 [16,41]. This finding was further confirmed with NOESY correlations observed between 2 -OMe (δ H 3.58) and H-2A (δ H 4.70). HMBC correlations from H-1 A (δ H 3.40) to C-2 (δ C 157.7), C-3 (δ C 122.1), C-4 (δ C 158.7) and C-2 (δ C 125.8) indicated the placement of the isoprenyl unit at C-3 . Analysis of the NMR spectroscopic data showed its structural similarity to kenusanone F 7-methyl ether (C 22 H 24 O 7 ) isolated previously from stem bark of D. melanoxylon [16] and to kenusanone F (C 21 H 22 O 7 , 8) obtained from stem bark of Erythrina brucei [42] and also isolated in this study. The difference is that compound 1 is missing one methyl group compared to  The absolute configuration of 1 was assigned by ECD spectroscopy. Usually, the octant rule modified for cyclic arylketones is applied to determine the stereochemistry of isoflavanones [43]. This predicts a positive Cotton effect (CE) for the n → π* carbonyl transition between 320-352 nm for (3R)-isoflavanones with the B-ring in the favored equatorial position [16,43]. However, it should be kept in mind that the priority order according to the Cahn-Ingold-Prelog rules changes when hydrogen at C-3 in isoflavanones is replaced with a hydroxyl group in 3-hydroxylisoflavanones. Thus, (3R)-isoflavanones show the same spatial arrangement as (3S)-hydroxylisoflavanones. However, at least for 3-hydroxylisoflavones, the octant rule is not fully reliable and seems to be prone to misinterpretation. The ECD spectrum of 1 shows intense positive Cotton effects at 237, 292 and 348 nm, and a weak negative one around 330 nm (Figure 2). The weak CEs in the long wavelength region, around 330 (negative CE) or around 348 nm (positive CE), may not be reliable for the assignment of the absolute configuration of compound 1. However, the ECD spectrum of 1 appears similar to the one calculated for (3S)-kenusanone F 7-methyl ether with a negative CE at 330 nm [42] and shows a mirror image to (3R)-kenusanone F (8, Figure 2; [42], hence it is consistent with (3S)-1 configuration. This previously undescribed compound (1) was therefore characterized as (3S)-3,4 ,5,7-tetrahydroxyl-2 -methoxy-3 -(4hydroxylprenyl)isoflavanone.
The absolute configuration of 1 was assigned by ECD spectroscopy. Usually, the octant rule modified for cyclic arylketones is applied to determine the stereochemistry of isoflavanones [43]. This predicts a positive Cotton effect (CE) for the n→π* carbonyl transition between 320-352 nm for (3R)-isoflavanones with the B-ring in the favored equatorial position [16,43]. However, it should be kept in mind that the priority order according to the Cahn-Ingold-Prelog rules changes when hydrogen at C-3 in isoflavanones is replaced with a hydroxyl group in 3-hydroxylisoflavanones. Thus, (3R)-isoflavanones show the same spatial arrangement as (3S)-hydroxylisoflavanones. However, at least for 3-hydroxylisoflavones, the octant rule is not fully reliable and seems to be prone to misinterpretation. The ECD spectrum of 1 shows intense positive Cotton effects at 237, 292 and 348 nm, and a weak negative one around 330 nm (Figure 2). The weak CEs in the long wavelength region, around 330 (negative CE) or around 348 nm (positive CE), may not be reliable for the assignment of the absolute configuration of compound 1. However, the ECD spectrum of 1 appears similar to the one calculated for (3S)-kenusanone F 7-methyl ether with a negative CE at 330 nm [42] and shows a mirror image to (3R)-kenusanone F (8, Figure 2; [42], hence it is consistent with (3S)-1 configuration. This previously undescribed compound (1) was therefore characterized as (3S)-3,4′,5,7-tetrahydroxyl-2′-methoxy-3′-(4-hydroxylprenyl)isoflavanone.     16.2 (C-9 ) and 17.7 (C-10 )] at C-6 in 2. The tail-to-head linkage of the two isoprenyl moieties to form the geranyl group was further corroborated using 1 H-1 H COSY correlations between H-5 /H-6 and H-5 /H-4 . HMBC crosspeaks from H-1 (δ H 3.22) to C-5 (δ C 162.8), C-6 (δ C 109.8), C-7 (δ C 166.0), C-2 (δ C 124.0) and C-3 (δ C 135.2) clearly establish the location of the geranyl substituent at C-6. A Cotton effect for n → π* transition was not observed in compound 2, probably due to low concentration, and hence could not be used to determine absolute configuration. On the other hand, as in compound 1, compound 2 showed a strong positive Cotton effect for π → π* transition at 309 nm, allowing the assignment of the same absolute configuration at C-3, but the designation is R (due to change in priority because of the absence of OH at C-3 in compound 2). Thus, compound 2 was elucidated as (3R)-6-geranyl-4 ,5,7-trihydroxyl-2methoxy-3 -prenylisoflavanone. . HMBC cross-peaks from H-1 (δ H 3.02) to C-2 (δ C 158.8) and C-4 (δ C 155.2) indicated that the epoxyprenyl moiety was located at C-3 . Hence, the planar structure of 5 was elucidated as 6-geranyl-4 ,5,7-trihydroxyl-2 -methoxy-3 -(2,3-epoxy-3-methyl-butyl)-isoflavanone. Reliable optical rotation, UV and ECD spectra could not be generated for compounds 3-5.

Biological Activity
Since Dalbergia species are known to exhibit a variety of biological activities, the partitioned crude extracts and the isolated compounds of D. melanoxylon were tested for their antibacterial, antifungal, anthelmintic and cytotoxic properties applying an established model organism non-pathogenic to humans ( Table 3). The crude CH 2 Cl 2 extract of the root bark induced nearly complete inhibition (97% ± 0%) of the Gram-positive bacterium Bacillus subtilis at the concentration of 50 µg/mL and complete inhibition (100% ± 0%) of the Gram-negative bacterium Aliivibrio fischeri at 500 µg/mL showing its potential especially against Gram-positive bacteria. The antifungal and anti-oomycetes activity was evaluated against the phytopathogens Septoria tritici, Botrytis cinerea and Phytophthora infestans, respectively. The extract showed promising activity against all phytopathogens at a concentration of 125 µg/mL. No anthelminthic activity against Caenorhabditis elegans could be detected at 500 µg/mL. Likewise, at low concentration (0.05 µg/mL) no antiproliferative or cytotoxic effects were observed against the human cancer cell lines PC3 and HT29 whereas a higher concentration (50 µg/mL) induced significant inhibition of cell growth and viability (Table S15). These results imply that the crude extract possesses moderate cytotoxic properties but might also show selective biological effects with focus on antibacterial and antifungal activities. Table 3. Antibacterial (Bacillus subtilis, Aliivibrio fischeri) and antifungal (Phytophthora infestans, Botrytis cinerea, Septoria tritici) activities of the CH 2 Cl 2 extract and isolated compounds from D. melanoxylon shown as growth inhibition [%] a . Data are presented as mean values ± standard deviation (n = 6 for antibacterial assays, n = 3 for antifungal assays. Based on the results of the crude extracts, the isolated major compounds (1, 2, 7, 9, 10) were subjected to a preliminary biological screening in antibacterial and antifungal assays (Table 3). For the antibacterial assays, the compounds were tested at concentrations of 1 and 100 µM, and for the antifungal assays of 42 and 125 µg/mL. In both B. subtilis and A. fischeri assays, (3R)-tomentosanol B (9) and sophoraisoflavanone A (10) inhibited nearly 100% of bacterial growth at a concentration of 100 µM after 16 h incubation time. Both compounds had also a good antifungal activity against S. tritici at 125 µg/mL (corresponding to 0.28 and 0.34 mM, respectively) (Table 3). Furthermore, kenusanone H (7) at 42 µg/mL (0.1 mM) showed a promising growth inhibition of B. cinerea and S. tritici. Thus, these compounds were also tested against a panel of human pathogenic bacteria (Table 4) and fungi (Table S15). Kenusanone H (7), (3R)-tomentosanol B (9) and sophoraisoflavanone A (10) exhibited promising antibacterial activity against Gram-positive bacteria including MRSA as shown by the induction of significant inhibition zones in agar diffusion assays. Even more importantly, these compounds also inhibited the growth of Mycobacteria vaccae, a nonpathogenic member of the tuberculosis inducing the Mycobacteriaceae family. Indeed, previous docking studies indicated the potential binding of 3-hydroxylisoflavanones from D. melanoxylon to different mycobacterial target enzymes [16]. In the present study kensuanone H (7) displayed MIC values of 1.56, 1.56 and 0.78 µg/mL (3.7, 3.7 and 1.8 µM) against S. aureus (MRSA), Enterococcus faecalis and Mycobacterium vaccae, respectively, while tomentosanol B (9) inhibited the growth of these bacteria with MIC values of 3.12, 6.25 and 1.56 µg/mL corresponding to 6.9, 13.8 and 3.4 µM (Table 4), respectively. In addition, compound 7 also exhibited moderate antifungal effects against Candida albicans, Penicillium notatum and Aspergillus fumigatus, compound 9 against P. notatum and 10 against S. salmicolor, C. albicans and P. notatum (Table S15). Except for compounds 8-10, the antimicrobial potential of the tested compounds is reported here for the first time. Nevertheless, in prior studies, kenusanone F (8), purified from E. brucei displayed moderate activity (MIC values ranging from 125 to 250 µg/mL) against four pathogenic test organisms, namely S. aureus, B. cereus, B. megaterium and E. coli [42], while tomentosanol B (9) showed antiplasmodial activity (IC 50 = 25.3 µM) and virtually no in vitro cytotoxicity against the Chinese hamster ovarian (CHO) cell line (selectivity index = 5) [47]. Sophoraisoflavanone A (10) isolated from Echinosophora koreensis was already previously described as compound with strong antifungal (C. albicans, S. cerevisiae) and antibacterial activity (E. coli, S. typhimurium, S. epidermis, S. aureus) showing MIC values around 60 and 20 µg/mL, respectively [19]. In addition, this compound has proven toxic (IC 50 = 22.1 µg/mL) to a human liver (HepG2) cell line [19]. Although we could not demonstrate anthelmintic activity for the crude extract of D. melanoxylon, mild anthelmintic effects of prenylated isoflavonoids have been reported [48]. Neoflavonoids (represented e.g., by methyl dalbergin (11) and dalbergin (12)) were not included in the biological testing in our study but were previously shown to possess osteogenic properties [37] whereas structurally related dalbergiones from D. melanoxylon exhibited anti-inflammatory effects [49]. Prenylated flavonoids and isoflavonoids play important roles in the defense strategy of plants by protecting them against diseases through a broad inhibition profile against bacteria and fungi [50]. At the same time, these compounds represent promising starting points for the development of new, natural therapeutics against MRSA and other Grampositive bacteria [51]. Increased hydrophobicity and bioavailability (mediated by one or two prenyl groups) and electrostatic interactions are the main determinants for the anti-MRSA activity of prenylated isoflavonoids [51]. The effects might be mediated by damaging the membrane or cell wall function [19] whereby interaction with bacterial membranes reduces the fluidity of outer and inner membrane layers [52]. Specifically, prenylation at C-8, as present in kenusanone H (7), seems to be connected to strong biological activity, and also hydroxylation at C-3 in 10 plays a role for several biological effects [52]. In contrast, introduction of a hydroxyl group in the prenyl chain as in compound 1 seems to be connected to a reduction of activity. However, prenyl substitution increases antibacterial but also cytotoxic properties [52]. For the most promising candidate, kenusanone H (7), the cytotoxicity against HeLa cells was determined with a CC 50 of 1.8 ± 1.4 µg/mL (4.2 µM). Prenylated flavonoids and isoflavonoids show moderate cytotoxic properties [19,50] which would have to be considered for potential applications or development.
Altogether, fourteen compounds including six new isoflavanones were isolated from the root bark of D. melanoxylon, a medicinal plant largely used for the treatment of infectious diseases. The crude CH 2 Cl 2 extract of the root bark of D. melanoxylon induced in a concentration dependent manner different degrees of inhibition against the tested microorganisms. Among the tested compounds, 7 and 9 showed strong activities against several pathogenic microbes, while compound 10 was selective towards M. vaccae 10670 M4. It is worth noting that compounds 7 and 9 showed superior activity against S. aureus (MRSA) 134/93 R9 compared to the reference drug ciprofloxacin. Despite these activities, neither the crude extract, nor the tested compounds showed considerable anthelminthic and cytotoxic activities. Hence, the observed biological effects support the traditional use of D. melanoxylon against several conditions, which appear to be connected to bacterial or fungal infections [6,25]. The prenylated isoflavanone constituents proved to be of relevant bioactivity and are likely responsible for the activity of the roots of this plant, suggesting future investigations in terms of structure-activity-relationship, mode of action and in vivo experiments.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/metabo13060678/s1, Figure S1: NMR, HRMS, UV and CD spectra of compound 1; Figure S2: NMR, HRMS, UV and CD spectra of compound 2; Figure S3: NMR and HRMS spectra of compound 3; Figure S4: NMR and HRMS spectra of compound 4; Figure S5: NMR and HRMS spectra of compound 5; Figure S6: NMR and HRMS spectra of compound 6; Figure S7: NMR, UV and CD spectra of compound 7; Figure S8: NMR, UV and CD spectra of compound 8; Figure S9: NMR, HRMS, UV and CD spectra of compound 9; Figure S10: NMR, HRMS, UV and CD spectra of compound 10; Figure S11: NMR and HRMS spectra of compound 11; Figure S12: NMR and HRMS spectra of compound 12; Figure S13: NMR and HRMS spectra of compound 13; Figure S14: NMR and HRMS spectra of compound 14; Table S1: NMR data of compound 1; Table S2: NMR data of compound 2; Table S3: NMR data of compound 3; Table S4: NMR data of compound 4; Table S5: NMR data of compound 5; Table S6: NMR data of compound 6; Table S7: NMR data of compound 7; Table S8: NMR data of compound 8; Table S9: NMR data of compound 9; Table S10: NMR data of compound 10; Table S11: NMR data of compound 11; Table S12: NMR data of compound 12; Table S13: NMR data of compound 13; Table S14: NMR data of compound 14; Table S15: Cytotoxic activities of crude extract of D. melanoxylon against human cancer cell lines; Table S16: Antifungal activity of compounds from D. melanoxylon against human pathogens.

Informed Consent Statement: Not applicable.
Data Availability Statement: Additional information related to this manuscript can be found in the supporting information. Further data are available on request. Data is not publicly available due to privacy.