New Constituents from Zanthoxylum rhoifolium

Keywords

Zanthoxylum rhoifolium - Rutaceae - sesquiterpenes - α-ionone - DFT/NMR - antibacterial

Introduction

Within the Rutaceae family, the genus Zanthoxylum L. comprises 225 species of trees and shrubs that are native to subtropical and temperate regions worldwide [1]. Among these, the native South American tree Zanthoxylum rhoifolium Lam., commonly referred to as “Indian ash” and “mamica de cadela”, is part of this genus. This species is primarily distributed in Brazil in the Minas Gerais and Rio de Janeiro states along the eastern rainforest of the Atlantic coast; additionally, it is found in the states of Piauí and Ceará in the northeast region [2]. In traditional medicine practices, the extract resulting from Z. rhoifolium bark boiled in water is used to relieve symptoms associated with malaria [3]. This practice has been observed in the Patamona Indians of Huyana and Bolivia [4]. Several traditional applications for this plant have been described, focusing on the bark efficiency as a tonic and febrifuge, as well as its value as a mouth rinse for toothaches. It is also worth noting the external use of a decoction as an antibacterial agent used for eruptions on childrenʼs legs [5].

In addition to the traditional medicinal uses, numerous studies reported in the literature have assessed the biological activities of extracts obtained from this plant. The roots of Z. rhoifolium have demonstrated anti-plasmodic activity, supporting the traditional use of this plant as an antimalarial agent [6]. Extracts obtained from the stem bark have shown antinociceptive effects [7] and gastroprotective activity [8]. In a TLC bioassay, the crude plant bark extracts exhibited moderate antibacterial activity against gram-positive bacteria such as Staphylococcus aureus, Staphylococcus epidermidis, and Micrococcus luteus, as well as against three gram-negative bacteria, Klebsiella pneumoniae, Salmonella setubal, and Escherichia coli [9], while the CHCl₃-MeOH extract from the plant bark has demonstrated fungistatic activity against Botrytis cinerea, Sclerotinia sclerotiorum, Alternaria alternata, Colletotrichum gloeosporioides, and Clonostachys rosea in an earlier preliminary study [10].

Numerous phytochemical investigations detailed in the literature have examined extracts derived from the bark, leaves, and fruits of the plant, revealing the presence of different metabolites, including alkaloids, coumarins, lignans, terpenes, and flavonoids [6], [11], [12]; however, few studies have been reported to date on the separation and chemical identification of the nonpolar constituents [13], a part from the ones mainly focused on the essential oil composition [9].

In the framework of a project study devoted to the antimicrobial evaluation of plant and fungi extracts and/or pure secondary metabolites [14], [15], a phytochemical study of apolar extracts of the plant stem bark and leaves was carried out, leading to the isolation and characterization of 29 compounds, including three new sesquiterpenes and one new α-ionone. Moreover, the antimicrobial activity of the total extracts against a collection of gram-positive and gram-negative bacteria (E. coli, Staphylococcus aureus, Klebsiella sp., Streptococcus mutans, Citrobacter sp., Salmonella sp., Bacillus subtilis, Shigella sp., Enterococcus fecalis, Bacillus clausii, Pseudomonas aeruginosa, Acinetobacter baumannii, Streptococcus epidermidis, and Lysteria monocitogenes) was evaluated.

Results and Discussion

The Z. rhoifolium stem bark and leaves were extracted with solvents of increasing polarity. The petroleum ether, CHCl₃, and CHCl₃-MeOH extracts subjected to different column chromatographies yielded 29 compounds, including three new sesquiterpenes (1 – 3) and one new α-ionone (4) ([Fig. 1]).

Fig. 1 Structures of compounds 1 – 4.

Among the known derivatives, five α-ionones: corchoionoside C [16], 6R,9S-3-oxo-α-ionol-β-d-glucopyranoside [17], breyniaionoside A [18], blumenol C [19], and debiloside C [20]; four coumarins: pimpinellin [21], anisocoumarin H [22], auraptene [23], and acetoxyauraptene [23]; thirteen sesquiterpenes: (−)-1,12-oxaguai-10-(15)-ene [24], pancherione [25], 4α-hydroxy-11-hydroxy-guai-10(14)-ene [26], 1β,6α-dihydroxyeudesm-4(15)-ene [27], spathulenol [28], 7-epi-11-hydroxychabrolidione A [29], holostylactone [29], caryolane-1,9-β-diol [30], (4S,5S,7S,10S)-5,12-dihydroxyeudesma-1-one [31], (4R,5S,7S,10S)-5,12-dihydroxyeudesma-1-one [32], (−)-10-epi-5-β-11-dihydroxyeudesmane [33], (1α,4αH,5α,7αH,10αH)-1,5-epoxyguaian-11-ol [34], and (1β,4αH,5α,7αH,10αH)-1,5-epoxyguaian-11-ol [34]; and three lignans episesamin [35]: piperitol-3,3-dimethylallyl ether [36], and alangilignoside C [37] were characterized by comparison of their NMR and MS data with those reported in the literature.

Compound 1 possessed a molecular formula of C₁₅H₂₆O₃ and three indices of hydrogen deficiencies, as determined by the HRESIMS spectrum (m/z 277.1766 [M + Na]⁺). Analysis of the 1D NMR data ([Table 1]) showed the presence of four methyl groups at δ _H 1.04 (3H, d, J = 6.8 Hz), 1.06 (3H, d, J = 6.5 Hz), and 1.18 (6H, s), one hydroxymethine (δ _H 4.13, br d, J = 6.0 Hz), four methylenes, three methines, and three oxygenated quaternary carbons, suggesting a tricyclic skeleton of 1. A comparison of ¹H and ¹³C NMR data, obtained with the analysis of HSQC and HMBC spectra ([Fig. 2]), with those of related guaiane sesquiterpenes suggested close similarities [34], [38]. A COSY experiment registered for 1, revealed the following connectivities: H-2–Me-15 (through H₂-3 and H-4) and H₂-6–Me-14, supporting the presence of a tertiary hydroxyl function at C-11. Signals at δ _C 73.0 and 74.7 indicated the presence of an epoxy ring. HMBC correlations between H₂-3 and C-4; H₂-6 and C-1, C-4, C-5, C-8, C-11; Me-12 and Me-13 and C-11; Me-14 and C-1 and C-10 suggested that the hydroxymethine function was placed at C-2, with the epoxy ring between C-1 and C-5. The determination of the relative configuration of C-1, C-2, C-4, C-5, C-7, and C-10 of 1 was performed by applying the quantum mechanical (QM) methods combined with NMR spectroscopy (QM/NMR) approach. This methodology, discovered and optimized by our group [39], is based on the computation of NMR properties (¹³C/¹H NMR chemical shift) at the density functional theory (DFT) level and the subsequent comparison of the experimental and predicted ¹³C and ¹H NMR chemical shifts using statistical parameters [39]. The latter step is fundamental for suggesting the proper prediction of the stereochemical assignment of organic molecules. First, using Monte Carlo molecular mechanics (MCMM), low-mode conformational sampling (LMCS), and molecular dynamics (MD) simulations, a thorough empirical conformational search related to all the possible investigated stereoisomers of 1 was conducted. Specifically, five stereocenters were undefined, since C-1 and C-5 were considered as a single stereocenter for the presence of the epoxide. Therefore, among the possible 32 stereoisomers, 16 diastereoisomers (1a-1p, vide infra) were considered for the QM/NMR predictions. The MM conformer ensembles were further subjected to geometry and energy optimization processes at the DFT at the MPW1PW91/6 – 31 g(d) level of theory. Finally, additional visual inspection was performed on the optimized conformers to eliminate any additional redundant ones. The Boltzmann distribution of the conformers for each stereoisomer derived at the same level of theory was then taken into consideration for predicting the ¹³C and ¹H NMR chemical shifts for 1a-1p at the MPW1PW91/6 – 31 g(d,p) level. The integrated equation formalism model (IEFPCM) for modelling methanol as a solvent was used for all DFT calculations (see Material and Methods). The computed and experimental values were then compared using the mean absolute error (MAE) values (see Computational Details, Material and Methods). The relative configuration of 1 was thus elucidated based on the MAE values for the ¹³C and ¹H NMR chemical shifts (¹³C MAE = 1.67 ppm, ¹H MAE = 0.10 ppm for 1b) (Tables 1S and 2S, Supporting Information). Specifically, the relative configuration was determined as follows: 1R*, 2R*, 4R*, 5S*, 7S*, 10R*. Additionally, we used the DP4+ method [40] to corroborate our findings further. Again, the stereoisomer 1b showed the highest DP4+ probability (100%), corroborating the relative configuration 1R*, 2R*, 4R*, 5S*, 7S*, 10R*. Thus, 1 was elucidated as (1R*,2R*,4R*,5S*,7S*,10R*)-epoxy-guaian-2,11-diol.

Table 1 ¹H and ¹³C NMR data of compounds 1 and 2 ^a.

Position	1		2
	δ H	δ C	δ H	δ C
aSpectra were recorded in CD3OD at 600 (1H) and 150 MHz (13C); J values are in parentheses and reported in Hz; chemical shifts are given in ppm; assignments were confirmed by COSY, 1D-TOCSY, HSQC, and HMBC experiments; boverlapped signal
1		75.6		158.9
2a	4.13 br d (6.0)	73.0	5.73 br s	125.1
2b
3a	1.62 dd (13.0, 8.0)	38.0	2.27 ddd (11.2, 9.0, 3.5)	38.8
3b	1.21 m		2.01b
4	2.28 m	36.2	2.38 m	41.5
5		73.4	2.29 m	48.8
6a	2.21 d (15.0)	26.0	2.01b	31.6
6b	1.41 m		0.92 m
7	1.73b	46.0	1.35 m	54.6
8a	1.74b	29.0	1.83b	25.2
8b	1.33b		1.25 m
9a	1.51 m	28.0	1.85b	43.0
9b	1.32b		1.66 br dd (14.0, 11.0)
10	2.44 m	29.0		75.0
11		74.7		73.8
12	1.18 s	27.0	1.18 s	27.4
13	1.18 s	27.0	1.15 s	26.3
14	1.06 d (6.5)	18.0	1.27 s	32.0
15	1.04 d (6.8)	13.5	1.07 d (6.5)	15.7

Fig. 2 Main HMBC correlations of compounds 1 – 4.

The HRESIMS of compound 2 showed a protonated molecular ion at m/z 239.2017 [M + H]⁺, consistent with the molecular formula C₁₅H₂₆O₂ and three indices of hydrogen deficiencies. Its 1D and 2D NMR ([Table 1]) features indicated the presence of a guaiane-type sesquiterpene and similarities with compound 1. The ¹³C NMR chemical shift assignments were derived from the analysis of HSQC and HMBC experiments ([Fig. 2]). COSY and HSQC experiments were able to establish the spin system in the molecule starting from the sp² proton a δ _H 5.73 at C-2. Two quaternary hydroxyl functions were also evident from the HMBC spectra at δ _C 73.8 and 75.0 that were located at positions 10 and 11 from the respective correlations between Me-14–C-1, Me-14–C-9, Me-14–C-10, H₂-9–C-10, Me-12–C-7, Me-12–C-11, and Me-13–C-7, Me-13–C-11. Keys HMBC correlations were also observed among H₂-3, H₂-6, H₂-9, and C-1, confirming the presence of a double bond between C-1/C-2. The relative configuration was obtained through the QM/NMR approach, as described for compound 1. After evaluating the obtained MAE, the isomer 2c featured the lowest one (¹³C MAE = 1.12 ppm, ¹H MAE = 0.10 ppm) (Tables 3S and 4S, Supporting Information), with the relative configuration of 4R*, 5R*, 7S*, and 10S*. Thus, compound 2 was identified as (4R*,5R*,7S*)-guaia-1,2-en-10,11-diol, also confirmed after applying the DP4+ approach, in which 2c showed the highest DP4+ probability (100%).

In the HRESIMS spectrum, compound 3 showed a sodiated molecular ion at m/z 277.1768 [M + Na]⁺, indicating a molecular formula of C₁₅H₂₆O₃, with three indices of hydrogen deficiencies. The ¹H NMR spectrum ([Table 2]) indicated three methyl groups as two singlet signals (δ _H 1.05 3H, 1.12 6H), one methyl group as a doublet signal (δ _H 1.05, J = 6.5 Hz) and four methylenes, three methines, and one hydroxymethine (δ _H 3.82, br dd, J = 2.5, 2.5 Hz). From the COSY experiment, the spin system H₂-2–H₂-9 was observed. The ¹³C NMR data suggested an eudesmane skeleton for 3 since one of the three hydrogen deficiencies was consumed by one keto group (δ _C 220.0) [32], [38]. An oxygenated tertiary carbon (δ _C 74.0) and a quaternary carbon (δ _C 53.4) were also evidenced. The HMBC experiment ([Fig. 2]) was helpful in establishing the substituent locations; thus, the keto group was placed at C-1 from the H-3–C-1, H-4–C-1, H₂-6–C-1, H₂-9–C-1, and Me-14–C-1 correlations. The hydroxy groups were located at C-3 and C-11 from the respective correlations Me-15–C-3 and H₂-6–C-11. Unfortunately, despite various attempts to derive the stereochemical configuration of compound 3, the obtained data lacked sufficient confidence to elucidate and suggest the stereogenic centers of this molecule properly. For this reason, only the novel 2D chemical scaffold was reported. Thus, we suggest compound 3 as 3,11-dihydroxyeudesma-1-one.

Table 2 ¹H and ¹³C NMR data of compounds 3 and 4 ^a.

Position	3		4
	δ H	δ C	δ H	δ C
aSpectra were recorded in CD3OD at 600 (1H) and 150 MHz (13C) for 3 and at 400 (1H) and 100 MHz (13C) for 4; J values are in parentheses and reported in Hz; chemical shifts are given in ppm; assignments were confirmed by COSY, 1D-TOCSY, HSQC, and HMBC experiments; boverlapped signal
1		220.0		39.3
2a	2.34 dd (15.0, 3.0)	33.6	2.43 d (13.0)	56.2
2b	1.60b		2.03 d (13.0)
3a	3.82 br dd (2.5, 2.5)	78.3		213.9
4a	2.43 m	32.3	2.32 ddd (13.0, 4.5, 2.3)	49.8
4b			2.16 br t (13.0)
5	2.47 m	56.3	1.94 m	34.4
6	1.79 br dd (9.0, 8.0)	22.0	2.01 dd (13.0, 2.0)	58.1
7	1.61b	44.0	5.65 br d (15.3, 9.0)	135.9
8a	1.62b	22.8	5.55 br d (15.3, 7.7)	131.3
8b	1.45 m
9a	2.20 ddd (16.0, 14.0, 3.5)	33.0	4.41 m	79.6
9b	1.31 m
10		53.4	3.64 m	66.5
11		74.0	0.82 s	21.0
12	1.12 s	24.9	0.99 s	31.8
13	1.12 s	27.3	1.01 d (6.0)	21.8
14	1.05 s	20.7
15	1.05 d (6.5)	17.6
Glc 1			4.38 d (7.8)	100.7
2			3.16 dd (9.0, 7.8)	74.9
3			3.31 dd (9.0, 9.0)	77.8
4			3.28 dd (9.0, 9.0)	71.4
5			3.20 m	77.8
6a			3.86 dd (12.0, 3.0)	62.4
6b			3.66 dd (12.0, 5.0)

Compound 4 (C₁₉H₃₂O₈) displayed a sodiated molecular ion at m/z 411.1985 [M + Na]⁺ and a fragment in the HRESIMS at m/z 249.13 [M + Na – 162]⁺, consistent with the presence of a hexose moiety in the molecule. Its NMR features ([Table 2]) suggested the presence of an α-ionone glycoside [18] for the signals attributable to two methyl groups as singlets (δ _H 0.82, 0.99), one methyl group as a doublet (δ _H 1.01, d, J = 6.0 Hz), one oxygenated methine (δ _H 4.41, m), one hydroxymethylene (δ _H 3.64, m), two methines, two methylenes, two olefinic protons (δ _H 5.55 br d, J = 15.3, 7.7 Hz, and 5.65 br d, J = 15.3, 9.0 Hz), one keto group (δ _C 213.9), and signals for a β-glucopyranose moiety. The COSY experiment suggested the spin sequence H₂-4–H-5–H-6–H-7–H-8–H-9–H₂-10. The HMBC spectrum ([Fig. 2]) led to the identification of the hydroxymethylene C-10, showing cross-peaks between H₂-10–C‐8 and H₂-10–C‐9 and the hydroxymethine C-9 thanks to the H-7–C-9 and H‐8–C-9 correlations. Finally, the relative configuration of compound 4 was obtained through the application of the QM/NMR combined approach, as described above. Specifically, we firstly performed the computation of the chemical shifts considering only the aglycone part of compound 4, thus removing the glucopyranose from the structure and replacing it with a simplified group (-OCH₃), considering four diastereoisomers (4a-4d). The obtained MAE highlighted two possible stereoisomers featuring the relative configuration of 5R*, 6R*, and 9R* (4a) and 5R*, 6R*, and 9S* (4b) (see Table 5S and 6S, Supporting Information), thus suggesting the relative configuration of the aglycone part as 5R* and 6R*. Therefore, in order to elucidate the relative configuration of C-9, the prediction of NMR parameters was carried out also considering the entire glycoside, accounting for 5R*, 6R*, and 9R* (4e) and 5R*, 6R*, and 9S* (4 f). In detail, the glucopyranose moiety included in this structure was considered for the computation of the relative energy and the final Boltzmann distribution and the prediction of the chemical shifts, while it was not considered for the computation of the statistical parameters for two reasons: (1) the sugar moiety was already identified as glucopyranose and (2) to avoid additional errors in the quantitative analysis that could compromise the identification of the correct stereoisomer. In particular, these steps were performed to evaluate the influence of the β-glucopyranose moiety on the conformational ensemble and, therefore, on the chemical shift predictions. Specifically, we manually selected the conformers considering (a) the chair arrangement of the aglycone part, (b) the arrangement of the alkene at the C-7 and C-8 position, and (c) the β-glucopyranose moiety orientation (Fig. 26S, Supporting Information), for a total of 12 conformers for each stereoisomer (4e and 4f). Finally, the stereoisomer 4e featured the lowest MAE (¹³C MAE = 1.27 ppm, ¹H MAE = 0.14 ppm) (Tables 7S and 8S, Supporting Information). Therefore, the relative configuration of compound 4 was suggested as 5R*, 6R*, and 9R*, and was confirmed by the obtained DP4+ probability of 99.71%. From all these data, 4 was characterized as (5R*,6R*,9R*)-10-dihydroxy-megastigman-7-en-3-one 9-O-β-d-glucopyranoside, a new specialized metabolite also for the aglycone portion.

Minimum inhibitory concentrations (MICs) were determined to assess the Z. rhoifolium extracts antimicrobial effect against a collection of gram-positive and gram-negative bacteria. The MIC of the crude extracts against the bacteria were determined by the reference protocol of the “Agar and broth dilution methods to determine the MIC of antimicrobial substances” [41]. The MIC was determined as the lowest extract concentration that inhibited visible bacterial growth. The results showed no particular antimicrobial activity, except for the chloroform-methanol bark extract that, at 500 µg/mL, gave a 20% vitality inhibition on S. mutans and Citrobacter sp.

Material and Methods

General experimental procedures

Optical rotations were measured on an Atago AP-300 digital polarimeter equipped with a sodium lamp (589 nm) and a 1-dm microcell. NMR data were acquired on a Bruker Ascend-600 NMR spectrometer (Bruker BioSpin GmBH) equipped with a Bruker 5-mm PATXI Probe and a Bruker DRX-400 spectrometer at 300 K. Data processing was carried out with Topspin 3.2 software. 2D NMR spectra were acquired in methanol-d ₄ and standard pulse sequences and phase cycling were used for COSY, HSQC, HMBC, 1D-TOCSY, and ROESY spectra. HRESIMS data were obtained in the positive ionization mode on a Q Exactive Plus mass spectrometer, Orbitrap-based FT-MS system, equipped by an electrospray ionization (ESI) source (Thermo Fischer Scientific Inc.). Column chromatographies were carried out over silica gel (70 – 220 mesh; Merck) and Sephadex LH-20 (40 – 70 µm; Pharmacia). RP-HPLC separations were conducted using a Shimadzu LC-8A series pumping system equipped with a Shimadzu RID-10A refractive index detector and Shimadzu injector on a C₁₈ µ-Bondapak column (Waters; 30 cm × 7.8 mm, 10 µm, flow rate 2.0 mL/min). TLC separations were carried out using silica gel 60 F₂₅₄ (0.20 mm thickness) plates (Merck) and cerium sulphate as a spray reagent.

Plant material

Z. rhoyfolium stem bark and leaves were collected in April 2010 near Merida, Venezuela, and identified by Eng. Juan Carmona, Universidad de Los Andes, Merida, Venezuela. A voucher specimen (n. 607) was deposited at Jardin de Plantas Medicinales de la Facultad de Farmacia y Bioanalisis, Merida, Venezuela.

Extraction and isolation

Dried stem bark (814 g) and leaves (800 g) of Z. rhoyfolium were extracted with solvents of increasing polarity, starting from petroleum ether, CHCl₃, CHCl₃-MeOH 9 : 1, and MeOH by exhaustive maceration (3 × 2.0 L) to give 20.0, 15.0, 35.0, and 50.0 g of the respective residues from the stem bark and 11.5, 22.0, 8.2, and 35.0 g of the respective residues from the leaves. Part of the petroleum ether stem bark extract (7.0 g) was submitted to silica gel column chromatography (5 × 30 cm), eluting with n-hexane, followed by increasing concentrations of CHCl₃ in n-hexane (between 1 and 100%), continuing with CHCl₃, followed by increasing concentrations of MeOH in CHCl₃ (between 1 and 50%). There were 300 fractions of 25 mL collected, analyzed by TLC (silica gel plates, in n-hexane, or mixtures of n-hexane-CHCl₃, 9 : 1, 3 : 7 and CHCl₃-MeOH, 95 : 5, 9 : 1), and grouped into ten major fractions, A₁-J₁ (A₁ = 1 – 52, B₁ = 53 – 118, C₁ = 119 – 148, D₁ = 149 – 177, E₁ = 178 – 203, F₁ = 204 – 221, G₁ = 222 – 243, H₁ = 244 – 264, I₁ = 265 – 290, J₁ = 291 – 300). An aliquot (100 mg) of fraction D₁ (727.2 mg) was dissolved in 1 mL of MeOH-H₂O (7 : 3) and submitted to RP-HPLC with the same solvent (10 injections) to give holostylactone (1.9 mg, t _R 9 min), (1α,4αH,5α,7αH,10αH)-1,5-epoxyguaian-11-ol (2.3 mg, t _R 17 min), (1β,4αH,5α,7αH,10αH)-1,5-epoxyguaian-11-ol (2.8 mg, t _R 18 min), and piperitol-3,3-dimethylallyl ether (2.2 mg, t _R 25 min). Fraction E₁ (85.1 mg) was dissolved in 850 µL of MeOH-H₂O (6.5 : 5.5) and purified by RP-HPLC with the same solvent (8 injections) to yield panchierone (3.2 mg, t _R 14 min). An aliquot (40 mg) of fraction G₁ (97.8 mg) was solubilized in 400 µL of MeOH-H₂O (3 : 2) and subjected to RP-HPLC with the same solvent (4 injections) to yield (4S,5S,7S,10S)-5,12-dihydroxyeudesma-1-one (2.2 mg, t _R 15 min). An aliquot (50 mg) of fraction H₁ (146.2 mg) was solubilized in 500 µL of MeOH-H₂O (3 : 2) and subjected to RP-HPLC with the same solvent (5 injections) to give 7-epi-11-hydroxychabrolidione A (2.0 mg, t _R 16 min), (−)-1,12-oxaguai-10-(15)-ene (3.2 mg, t _R 55 min), and 2 (1.5 mg, t _R 58 min). Finally, an aliquot (70 mg) of fraction I₁ (143.4 mg) was dissolved in 700 µL of MeOH-H₂O (5.5 : 4.5) and chromatographed by RP-HPLC with the same solvent (7 injections) to yield compounds 3 (2.3 mg, t _R 7 min), (H-5)-5,12-dihydroxyeudesma-1-one (3.0 mg, t _R 10 min), 1 (1.0 mg, t _R 14 min), (−)-10-epi-5-β-11-dihydroxyeudesmane (1.5 mg, t _R 20 min), 4α-hydroxy-11-hydroxy-guai-10(14)-ene (2.4 mg, t _R 22 min), and caryolane-1,9-β-diol (1.8 mg, t _R 30 min).

A side of the CHCl₃ leaf extract (7.0 g) was submitted to flash silica gel column chromatography (SNAP 340 g, 7 × 18 cm) by Biotage, eluting with n-hexane, followed by increasing concentrations of CHCl₃ in n-hexane (between 1 and 100%), continuing with CHCl₃, followed by increasing concentrations of MeOH in CHCl₃ (between 1 and 50%). There were 320 fractions of 25 mL collected, analyzed by TLC (silica gel plates, mixtures of n-hexane-CHCl₃, 7 : 3, 1 : 1, CHCl₃, and CHCl₃-MeOH, 95 : 5), and grouped into five major fractions, A₂-E₂ (A₂ = 1 – 93, B₂ = 94 – 188, C₂ = 189 – 230, D₂ = 231 – 292, E₂ = 293 – 320. An aliquot (120 mg) of fraction C₂ (537 mg) was solubilized in 1.2 mL of MeOH-H₂O (7.5 : 2.5) and purified through RP-HPLC with the same solvent (10 injections) to yield pimpinellin (2.7 mg, t _R 8 min), 1β,6α-dihydroxyeudesm-4(15)-ene (2.4 mg, t _R 9 min), anisocoumarin H (2.4 mg, t _R 12 min), episesamin (2.3 mg, t _R 14 min), acetoxyauraptene (3.8 mg, t _R 18 min), spathulenol (1.4 mg, t _R 25 min), and auraptene (1.8 mg, t _R 35 min).

A side of the CHCl₃-MeOH leaf extract (2.8 g) was chromatographed over Sephadex LH-20 column chromatography using MeOH as an eluent (3 × 100 cm; flow rate 0.8 mL/min, collection volume 10 mL), collecting 54 fractions that were grouped by TLC (silica gel plates, CHCl₃-MeOH-H₂O 80 : 18 : 2, n-BuOH-CH₃COOH-H₂O 60 : 15 : 25) into six major collections, A₃-F₃ (A₃ = 1 – 14, B₃ = 15 – 27, C₃ = 18 – 22, D₃ = 23 – 33, E₃ = 34 – 45, F₃ = 46 – 54). An aliquot (100 mg) of fraction B₃ (535.2 mg) was dissolved in 1.0 mL of MeOH-H₂O (3.5 : 6.5) and subjected to RP-HPLC with the same solvent (10 injections) to yield debiloside C (3.0 mg, t _R 9 min), breyniaionoside A (2.4 mg, t _R 12 min), corchoionoside C (1.8 mg, t _R 16 min), compound 4 (4.0 mg, t _R 74 min), 6R,9S-3-oxo-α-ionol-β-d-glucopyranoside (3.9 mg, t _R 66 min), and blumenol C (2.4 mg, t _R 21 min). An aliquot (80 mg) of fraction C₃ (295 mg) was solubilized in 700 µL of MeOH-H₂O (3.5 : 6.5) and purified through RP-HPLC with the same solvent (8 injections) to give alangilignoside C (1.4 mg, t _R 33 min).

Compound 1: amorphous powder; [α] _d ²⁵ + 168 (c 0.1, MeOH); ¹H and ¹³C NMR data, see [Table 1]; HRESIMS m/z 277.1766 [M + Na]⁺ (calcd. for C₁₅H₂₆O₃Na, 277.1780).

Compound 2: amorphous powder; [α] _d ²⁵ + 103(c 0.1, MeOH); ¹H and ¹³C NMR data, see [Table 1]; HRESIMS m/z 239.2017 [M + H]⁺ (calcd. for C₁₅H₂₇O₂, 239.2006).

Compound 3: amorphous powder; [α] _d ²⁵ + 145 (c 0.1, MeOH); ¹H and ¹³C NMR data, see [Table 2]; HRESIMS m/z 277.1768 [M + Na]⁺, 255.1958 [M + H]⁺ (calcd. for C₁₅H₂₆O₃Na, 277.1780).

Compound 4: amorphous powder; [α] _d ²⁵ − 70 (c 0.1, MeOH); ¹H and ¹³C NMR data, see [Table 2]; HRESIMS m/z 411.1985 [M + Na]⁺, 249.13 [M + Na – 162]⁺ (calcd. for C₁₉H₃₂O₈Na, 411.1995).

Computational details

Maestro 12.7 (Schrödinger Schrödinger Suite) was used for generating the starting 3D chemical structures of compounds 1, 2, and 4. Optimization of the 3D structures was performed with MacroModel (version 13.1, Schrödinger Suite, LLC) [42] using the OPLS force field and the Polak-Ribier conjugate gradient (PRCG) algorithm (maximum derivative less than 0.001 kcal/mol). In particular, for compound 1, showing five stereogenic centers (C-1 and C-5 were considered as unique center, since an epoxide was present), 16 isomers were considered:

• 1a (1R*, 2R*, 4R*, 5S*, 7R*, 10R*) (9 conformers), 1b (1R*, 2R*, 4R*, 5S*, 7S*, 10R*) (14 conformers), 1c (1R*, 2R*, 4R*, 5S*, 7R*, 10S*) (13 conformers), 1d (1R*, 2R*, 4R*, 5S*, 7S*, 10S*) (8 conformers), 1e (1R*, 2S*, 4R*, 5S*, 7R*, 10R*) (19 conformers), 1f (1R*, 2S*, 4R*, 5S*, 7S*, 10R*) (15 conformers), 1g (1R*, 2S*, 4R*, 5S*, 7R*, 10S*) (9 conformers), 1h (1R*, 2S*, 4R*, 5S*, 7S*, 10S*) (13 conformers), 1i (1R*, 2R*, 4S*, 5S*, 7R*, 10R*) (9 conformers), 1j (1R*, 2R*, 4S*, 5S*, 7S*, 10R*) (11 conformers), 1k (1R*, 2R*, 4S*, 5S*, 7R*, 10S*) (9 conformers), 1l (1R*, 2R*, 4S*, 5S*, 7S*, 10S*) (17 conformers), 1m (1R*, 2S*, 4S*, 5S*, 7R*, 10R*) (10 conformers), 1n (1R*, 2S*, 4S*, 5S*, 7S*, 10R*) (11 conformers), 1o (1R*, 2S*, 4S*, 5S*, 7R*, 10S*) (10 conformers), 1p (1R*, 2S*, 4S*, 5S*, 7S*, 10S*) (14 conformers).

For compound 2, showing four stereogenic centers, 8 isomers were considered:

• 2a (4R*, 5R*, 7R*, 10R*) (17 conformers), 2b (4R*, 5R*, 7R*, 10S*) (15 conformers), 2c (4R*, 5R*, 7S*, 10R*) (14 conformers), 2d (4R*, 5R*, 7S*, 10S*) (14 conformers), 2e (4R*, 5S*, 7R*, 10R*) (16 conformers), 2 f (4R*, 5S*, 7R*, 10S*) (17 conformers), 2g (4R*, 5S*, 7S*, 10R*) (14 conformers), 2h (4R*, 5S*, 7S*, 10S*) (14 conformers).

For compound 4, showing three stereogenic centers considering the aglycone, four isomers of the aglycone (4a-4d) and two isomers (4e-4f) considering the β-glucopyranose were accounted for:

• 4a (5R*, 7R*, 9R*) (5 conformers), 4b (5R*, 7R*, 9S*) (8 conformers), 4c (5R*, 7S*, 9R*) (9 conformers), 4d (5R*, 7S*, 9S*) (5 conformers), 4e (5R*, 7R*, 9R*) (12 conformers), 4f (5R*, 7R*, 9S*) (12 conformers).

Using the produced 3D structures as input, exhaustive conformational searches were conducted at the empirical MM level. In particular, LMCS rounds (50 000 steps) and MCMM rounds (50 000 steps) were carried out. Moreover, MD simulations with a time step of 2.0 fs, an equilibration period of 0.1 ns, and a simulation duration of 10 ns were run at 450, 600, 700, and 750 K. A constant methanol dielectric term that mimicked the solvent presence was taken into account for each of these simulations. For each isomer, using the “Redundant Conformer Elimination” module of MacroModel (version 13.1, Schrödinger Suite, LLC) [42], the sampled conformers were then minimized (PRCG, maximum derivative less than 0.001 kcal/mol) for each isomer and compared. Specifically, a minimum cutoff of 0.5 Å RMSD (root-mean-square deviation) was set for saving structures. Afterwards, nonredundant conformers were considered for QM calculations using Gaussian 09 software [43]. Specifically, the sampled conformers were subjected to a further step of geometry optimization at the DFT level using the MPW1PW91 functional and the 6–31 G(d) basis set. MeOH solvent effects were reproduced by setting the related integral equation formalism version of the polarizable continuum model (IEFPCM) [44]. After that, the optimized geometries were additionally visually inspected in order to remove possible residual redundant conformers. The computation of the ¹³C and ¹H NMR chemical shifts was performed on each ensemble of the considered isomers of 1, 2, and 4, using the MPW1PW91 functional and the 6–31 G(d,p) basis set and MeOH (IEFPCM). For each case study compound, final ¹³C and ¹H NMR sets of data were produced for each investigated isomer, considering the weight of each conformer on the total Boltzmann distribution considering their relative energies. The multi-standard approach (MSTD) [45] was employed for the calibrations of predicted ¹³C/¹H chemical shifts and, followed this procedure:

1. chemical shifts for sp² carbons were computed accounting for benzene as a reference considering the carbonyl group (i.e., C-3 for compound 4);

2. chemical shifts for sp² carbons (i.e., C-7 and C-8 of compound 4) and their related protons were computed accounting for allyl alcohol as a reference for C-7, C-8, C-9 and O-β-d-Glc of compound 4 [45].

Tetramethylsilane (TMS) was used to compute chemical shift data of sp³ carbons and their related protons (Tables 1S–8S, Supporting Information). The comparison of experimental and calculated ¹³C and ¹H NMR chemical shifts was carried through the Δδ parameter (Tables 1S–8S, Supporting Information):

$\Delta \mathit{\delta }=\left|{\mathit{\delta }}_{\exp }–{\mathit{\delta }}_{\mathrm{calc}}\right|$

where δ _exp (ppm) and δ _calc (ppm) are the ¹³C/¹H experimental and predicted chemical shifts, respectively. In this way, the MAE values were computed for all the considered isomers of each case study compound using the following equation:

$\mathrm{MAE}=\frac{\sum \left(\mathrm{\Delta \delta }\right)}{n}$

namely, the sum (Σ) of the number of computed absolute error values (Δδ), normalized to the number of chemical shifts considered (n) (Tables 1S–6S, Supporting Information). DP4+ probabilities for each accounted isomer of 1, 2, and 4 were computed considering both ¹³C and ¹H NMR chemical shift values and the related experimental data. The chemical shift datasets obtained using only TMS as a reference compound were employed for the calculation of the DP4+ probabilities, manually selecting the sp² atoms in the available DP4+ Toolbox (Excel file) following the “multi-standard” approach.

Antibacterial activity

Gram-positive and gram-negative bacterial strains were Escherichia coli, Staphylococcus aureus, Klebsiella sp., Streptococcus mutans, Citrobacter sp., Salmonella sp., Bacillus subtilis, Shigella sp., Enterococcus fecalis, Bacillus clausii, Pseudomonas aeruginosa, Acinetobacter baumannii, Streptococcus epidermidis, and Lysteria monocitogenes and were purchased from ATCC. MICs were determined by the reference protocol of the “Agar and broth dilution methods to determine MIC of antimicrobial substances” [41]. All strains were grown aerobically in brain heart infusion (BHI) broth-rich medium at 37 °C. The samples were dissolved in 100% DMSO at different concentrations (extract: from 500 to 30 µg/mL), added to each well and bacterial suspensions (0.5 × 10⁵ CFU/mL), and then incubated at 37 °C for 24 h. Cell absorbance was measured at 600 nm using a Tecan Infinite 200 PRO spectrophotometer. A blank control (sterile culture medium, without compounds and suspensions of microorganisms), a vehicle control (sterile culture medium with DMSO), and an antimicrobial agent, chlorhexidine gluconate (CHX), were used. The MIC was determined as the lowest extract concentration that inhibited visible bacterial growth. Each experiment was performed with duplicate samples at each time point.

Contributorsʼ Statement

Data collection: M. Di Stasi, V. Parisi, V. Hernandez, E. Gazzillo, G. Bifulco, N. De Tommasi, G. Donadio Conception and design of the work: M. Chini, G. Bifulco, A. Braca, N. De Tommasi Statistical analysis: M. Di Stasi, V. Parisi, V. Hernandez, E. Gazzillo, M. Chini, G. Donadio Analysis and interpretation of the data: M. Di Stasi, V. Parisi, V. Hernandez, E. Gazzillo, M. Chini, G. Bifulco, A. Braca, N. De Tommasi, G. Donadio Drafting the manuscript: A. Braca, M. Di Stasi, V. Parisi, V. Hernandez, E. Gazzillo Critical revision of the manuscript: M. Chini, G. Bifulco, A. Braca, N. De Tommasi, G. Donadio

• References

• 1 Appelhans MS, Reichelt N, Groppo M, Paetzold C, Wen J. Phylogeny and biogeography of the pantropical genus Zanthoxylum and its closest relatives in the proto-Rutaceae group (Rutaceae). Mol Phylogenet Evol 2014; 126: 31-44
• 2 Cruz GL. Dicionário de plantas úteis do Brasil, 5th Ed. Rio de Janeiro: Civilização Brasileira 1995.
• 3 Bertani S, Bourdy G, Landau I, Robinson JC, Esterre P, Deharo E. Evaluation of French Guiana traditional antimalarial remedies. J Ethnopharmacol 2005; 98: 45-54
• 4 De Lucca M, Zalles J. Flora Medicinal Boliviana. La Paz: Los Amigos del Libro Editions; 1992: 498
• 5 Grenand P, Moretti C, Jacquemin H, Prévost MF. Pharmacopées Traditionnelles en Guyane. Créoles, Palikur, Wayãpi. Paris: IRD Editions; 2004: 816
• 6 Jullian V, Bourdy G, Georges S, Maurel S, Sauvain M. Validation of use of a traditional antimalarial remedy from French Guiana, Zanthoxylum rhoifolium Lam. J Ethnopharmacol 2006; 106: 348-352
• 7 Pereira SS, Lopes LS, Marques RB, Figueiredo KA, Costa DA, Chaves MH, Almeida FR. Antinociceptive effect of Zanthoxylum rhoifolium Lam. (Rutaceae) in models of acute pain in rodents. J Ethnopharmacol 2010; 129: 227-231
• 8 Freitas FF, Fernandes HB, Piauilino CA, Pereira SS, Carvalho KI, Chaves MH, Soares PM, Miura LM, Leite JR, Oliveira RC, Oliveira FA. Gastroprotective activity of Zanthoxylum rhoifolium Lam. in animal models. J Ethnopharmacol 2011; 137: 700-708
• 9 de Abreu Gonzaga W, Weber AD, Giacomelli SR, Simionatto E, Dalcol II, Machado Dessoy EC, Morel AF. Composition and antibacterial activity of the essential oils from Zanthoxylum rhoifolium . Planta Med 2003; 69: 773-775
• 10 Carotenuto G, Carrieri R, Tarantino P, Alfieri M, Leone A, De Tommasi N, Lahoz E. Fungistatic activity of Zanthoxylum rhoifolium Lam. bark extracts against fungal plant pathogens and investigation on mechanism of action in Botrytis cinerea . Nat Prod Res 2015; 29: 2251-2255
• 11 de Moura NF, Ribeiro HB, Machado ECS, Ethur EM, Zanatta N, Morel A. Benzophenanthridine alkaloids from Zanthoxylum rhoifolium . Phytochemistry 1997; 46: 1443-1446
• 12 de Abreu Gonzaga W, Weber AD, Giacomelli SR, Dalcol II, Hoelzel SC, Morel AF. Antibacterial alkaloids from Zanthoxylum rhoifolium . Planta Med 2003; 69: 371-374
• 13 Arruda MSP, Fernandes JB, Vieira PC, Silva MFGF, Pirani JR. Chemistry of Zanthoxylum rhoifolium: A new secofuroquinoline alkaloid. Biochem Syst Ecol 1992; 20: 173178
• 14 Di Stasi M, Donadio G, Bader A, De Leo M, Braca A. Two new triterpenes from Commicarpus grandiflorus (A. Rich.) Standl. aerial parts exudate. Nat Prod Res 2023; 37: 3228-3236
• 15 Donadio G, Chini MG, Parisi V, Mensitieri F, Malafronte N, Bifulco G, Bisio A, De Tommasi N, Bader A. Diterpenoid constituents of Psiadia punctulata and evaluation of their antimicrobial activity. J Nat Prod 2022; 85: 1667-1680
• 16 Caliş I, Kuruüzüm-Uz A, Lorenzetto PA, Rüedi P. (6S)-Hydroxy-3-oxo-alpha-ionol glucosides from Capparis spinosa fruits. Phytochemistry 2002; 59: 451-457
• 17 Pabst A, Barron D, Sémon E, Schreier P. Two diastereomeric 3-oxo-α-ionol β-d-glucosides from raspberry fruit. Phytochemistry 1992; 31: 1649-1652
• 18 Morikawa H, Kasai R, Otsuka H, Hirata E, Shinzato T, Aramoto M, Takeda Y. Terpenic and phenolic glycosides from leaves of Breynia officinalis HEMSL. Chem Pharm Bull 2004; 52: 1086-1090
• 19 Miyase T, Ueno A, Takizawa N, Kobayashi H, Oguchi H. Studies on the glycosides of Epimedium grandiflorum Morr. var. thunbergianum (Miq.) Nakai. Chem Pharm Bull 1988; 36: 2475-2484
• 20 Xu XH, Tan CH, Jiang SH, Zhu DY. Debilosides A–C: Three new megastigmane glucosides from Equisetum debile . Helv Chim Acta 2006; 89: 1422-1426
• 21 Reed MW, Moore HW. Efficient synthesis of furochromone and furocoumarin natural products (khellin, pimpinellin, isophellopterin) by thermal rearrangement of 4-furyl-4-hydroxycyclobutenones. J Org Chem 1988; 53: 4166-4171
• 22 Zdero C, Bohlmann F, King RM, Robinson H. Diterpene glycosides and other constituents from Argentinian Baccharis species. Phytochemistry 1986; 25: 2841-2855
• 23 Chen IS, Lin YC, Tsai IL, Teng CM, Ko FN, Ishikawa T, Ishii H. Coumarins and anti-platelet aggregation constituents from Zanthoxylum schinifolium . Phytochemistry 1995; 39: 1091-1097
• 24 Sarker SD, Armstrong JA, Waterman PG. (−)-1, 12-oxagual-10 (15)-ene: A sesquiterpene from Eriostemon fitzgeraldii . Phytochemistry 1995; 40: 1159-1162
• 25 Raharivelomanana P, Bianchini JP, Faure R, Cambon A, Azzaro M. Two guaiane and eudesmane-type sesquiterpenoids from Neocallitropsis pancheri . Phytochemistry 1996; 41: 243-246
• 26 Zdero C, Bohlmann F, Niemeyer HM. Isocedrene and guaiane derivatives from Pleocarphus revolutus . J Nat Prod 1988; 51: 509-512
• 27 Ohmoto T, Ikeda K, Nomura S, Shimizu M, Saito S. Studies on the sesquiterpenes from Ambrosia elatior Linne. Chem Pharm Bull 1987; 35: 2272-2279
• 28 Iwabuchi H, Yoshikura M, Ikawa Y, Kamisako W. Studies on the sesquiterpenoids of Panax ginseng C.A. Meyer. Isolation and structure determination of sesquiterpene alcohols, panasinsanols A and B. Chem Pharm Bull 1987; 35: 1975-1981
• 29 Pereira MDP, da Silva T, Lopes LMX, Krettli AU, Madureira LS, Zukerman-Schpector J. 4, 5-seco-guaiane and a nine-membered sesquiterpene lactone from Holostylis reniformis . Molecules 2012; 17: 14046-14057
• 30 Heymann H, Tezuka Y, Kikuchi T, Supriyatna S. Constituents of Sindora sumatrana Miq. I. Isolation and NMR spectral analysis of sesquiterpenes from the dried pods. Chem Pharm Bull 1994; 42: 138-146
• 31 Minnaard AJ, Wijnberg JB, de Groot A. The synthesis of (+)-hedycaryol, starting from natural (−)-guaiol. Tetrahedron 1994; 50: 4755-4764
• 32 Xiong J, Wang LJ, Qian J, Wang PP, Wang XJ, Ma GL, Zeng H, Li J, Hu JF. Structurally diverse sesquiterpenoids from the endangered ornamental plant Michelia shiluensis . J Nat Prod 2018; 81: 2195-2204
• 33 Mathela CS, Melkani AB, Pant A, Dev V, Nelson TE, Hope H, Bottini AT. A eudesmanediol from Cymbopogon distans . Phytochemistry 1989; 28: 936-938
• 34 Tissandié L, Gaysinski M, Brévard H, Meierhenrich UJ, Filippi JJ. Revisiting the chemistry of guaiacwood oil: Identification and formation pathways of 5, 11- and 10, 11-epoxyguaianes. J Nat Prod 2017; 80: 526-537
• 35 Pelter A, Ward RS, Rao EV, Sastry KV. Revised structures for pluviatilol, methyl pluviatilol and xanthoxylol: General methods for the assignment of stereochemistry to 2, 6-diaryl-3, 7-dioxabicyclo [3.3. 0] octane lignans. Tetrahedron 1976; 32: 2783-2788
• 36 Yang CH, Cheng MJ, Lee SJ, Yang CW, Chang HS, Chen IS. Secondary metabolites and cytotoxic activities from the stem bark of Zanthoxylum nitidum . Chem Biodivers 2009; 6: 846-857
• 37 Achenbach H, Benirschke M, Torrenegra R. Alkaloids and other compounds from seeds of Tabernaemontana cymosa . Phytochemistry 1997; 45: 325-335
• 38 Boudermine S, Parisi V, Lemoui R, Boudiar T, Chini MG, Franceschelli S, Pecoraro M, Pascale M, Bifulco G, Braca A, De Tommasi N, De Leo M. Cytotoxic sesquiterpenoids from Ammoides atlantica aerial parts. J Nat Prod 2022; 85: 647-656
• 39 Lauro G, Bifulco G. Elucidating the relative and absolute configuration of organic compounds by quantum mechanical approaches. European J Org Chem 2020; 2020: 3929-3941
• 40 Grimblat N, Zanardi MM, Sarotti AM. Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nat Protoc 2008; 3: 163-175
• 41 Wiegand I, Hilpert K, Hancock REW. Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nat Protoc 2008; 3: 163-175
• 42 Watts KS, Dalal P, Tebben AJ, Cheney DL, Shelley JC. Macrocycle conformational sampling with MacroModel. J Chem Inf Model 2014; 54: 2680-2696
• 43 Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson GA, Nakatsuji H, Caricato M, Li X, Hratchian HP, Izmaylov AF, Bloino J, Zheng G, Sonnenberg JL, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Vreven T, Montgomery jr. JA, Peralta JE, Ogliaro F, Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN, Kobayashi R, Normand J, Raghavachari K, Rendell A, Burant JC, Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene M, Knox JE, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Martin RL, Morokuma K, Zakrzewski VG, Voth GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD, Farkas O, Foresman JB, Ortiz JV, Cioslowski J, Fox DJ. Gaussian 09, Revision A.02 (2009). Wallingford CT: Gaussian, Inc.; 2009
• 44 Tomasi J, Mennucci B, Cammi R. Quantum mechanical continuum solvation models. Chem Rev 2005; 105: 2999-3094
• 45 Sarotti AM, Pellegrinet SC. Application of the multi-standard methodology for calculating ¹H NMR chemical shifts. J Org Chem 2012; 77: 6059-6065