Antimycobacterial and Nitric Oxide Production Inhibitory Activities of Goyazensolide and Centratherin Isolated From Eremanthus crotonoides by High‐Speed Countercurrent Chromatography

TL;DR

Two compounds from Eremanthus crotonoides show strong antimycobacterial and anti-inflammatory effects, especially against tuberculosis strains.

Contribution

First isolation of goyazensolide and centratherin from E. crotonoides using HSCCC and their evaluation against virulent M. tuberculosis strains.

Findings

Goyazensolide and centratherin showed potent antimycobacterial activity against M. tuberculosis H37Rv with MIC50 values of 1.5 and 2.5 µg/mL.

Both compounds exhibited strong nitric oxide inhibition with IC50 values of 0.45 and 0.34 µg/mL, respectively.

The compounds showed low cytotoxicity and reduced activity against the hypervirulent M299 strain.

Abstract

In this study, bioassay‐guided fractionation of the ethanolic extract from Eremanthus crotonoides leaves enabled the isolation of two bioactive sesquiterpene lactones, centratherin and goyazensolide, which were evaluated for their antimycobacterial and anti‐inflammatory properties. The extract exhibited MIC50 values of 42.0 ± 0.1 and 39.0 ± 0.1 µg/mL against Mycobacterium tuberculosis H37Rv and the hypervirulent M299 strain, respectively, along with notable NO inhibitory activity. Fractionation by high‐speed countercurrent chromatography (HSCCC), followed by purification also employing HSCCC, yielded centratherin and goyazensolide, representing the first report of their isolation from E. crotonoides using this technique. These compounds showed potent activity against H37Rv (MIC50 = 1.5 ± 0.1 and 2.5 ± 0.1 µg/mL, respectively) but markedly reduced activity against M299 (MIC50 = 92.7 ± 0.1 and 90.6 ± 0.1 µg/mL). Both compounds also exhibited IC50 values of 0.45 ± 0.1 and 0.34 ± 0.1 µg/mL for nitric oxide inhibition, with low cytotoxicity. Collectively, these findings provide the first evidence of the antimycobacterial potential of these compounds against virulent M. tuberculosis strains and underscore their promise as potential drug candidates.

Figures

Chromatogram obtained from the UHPLC‐DAD‐MS/MS analysis of the F1 fraction derived from the EtOH extract of E. crotonoides.

Chromatogram by HPLC‐DAD (λ = 254 nm) of compound 1 with its respective UV spectrum, isolated from the F1 fraction from E. crotonoides.

Chromatogram by HPLC‐DAD (λ = 254 nm) of compound 2 with its respective UV spectrum, isolated from the F1 fraction from E. crotonoides.

E. crotonoides annotated compounds proposed by mass spectrometry analysis, and further identified by NMR, from the bioactive fraction obtained from the ethanolic extract through HSCCC separations.

Effect of samples on Mtb growth in bacterial culture. Bacterial suspensions (1 × 106 CFU/well) of Mtb strain H37Rv (A) and clinical isolate M299 (B) were treated or untreated with samples (0.031, 0.16, 0.8, 4, 20, 100, and 500 µg/mL) for 5 days. The MTT test quantified bacterial growth in the resulting cultures. As a negative control (C‐), it was used to supplement 7H9 medium, and as a positive control (C+), the untreated mycobacterial suspension. The seven bars for each compound refer to concentrations tested in ascending order. Values are represented as mean ± standard deviation, and different groups were considered significant at p < 0.05 (*), p < 0.01 (), and p < 0.001 (*) (n = 3).

Tables

TABLE 1. MS data of major sesquiterpene lactones present in the F1 fraction.

Observed m/z (M+H+)	Retention time	Observed fragment ions [M+H+]	Loss	Compound
361	29 min	275;	[M+H—C3H5CO2H]+	Goyazensolide
		257;	[M+H—C3H5CO2H—H2O]+
		229	[M+H—C3H5CO2H—H2O‐CO]+
375	32 min	275;	M+H—C4H6CO2H]+	Centratherin
		257	[M+H—C4H6CO2H—H2O]+

TABLE 2. Results of the biological assays performed on the ethanolic extract of E. crotonoides leaves, its respective fractions, and the isolated compounds.

Sample	Mycobacterium tuberculosis			Mycobacterium tuberculosis					TNF‐α	Cytotoxicity
	H37Rv			M299			NO
	MIC50	MIC90	SI	MIC50	MIC90	SI	IC50	SI		CC50
	(µg/mL)	(µg/mL)	(CC50 / MIC50)	(µg/mL)	(µg/mL)	(CC50 / MIC50)	(µg/mL)	(CC50 / IC50)		(µg/mL)
E. crotonoides ethanolic extract a	42.0 ± 0.1	195.0 ± 0.1	7.1	39.0 ± 0.1	363.1 ± 0.1	7.6	0.61 ± 0.01	489.3	261.1 ± 0.1	298.5 ± 0.1
F1	24.0 ± 0.1	83.2 ± 0.1	0.3	55.0 ± 0.1	293.1 ± 0.1	0.1	2.0 ± 0.01	3.1	>500	6.3 ± 0.1
F2	>500	>500	XX	>500	>500	XX	21.6 ± 0.01	XX	>500	203.1 ± 0.1
F3	>500	>500	XX	>500	>500	XX	18.4 ± 0.01	XX	>500	220.1 ± 0.1
F4	>500	>500	XX	>500	>500	XX	58.3 ± 0.01	XX	368 ± 0.1	53.7 ± 0.1
F5	>500	>500	XX	>500	>500	XX	31.2 ± 0.01	XX	>500	105.9 ± 0.1
F6	>500	>500	XX	>500	>500	XX	43.9 ± 0.01	XX	>500	111.9 ± 0.1
F7	>500	>500	XX	>500	>500	XX	48.9 ± 0.01	XX	>500	336.6 ± 0.1
F8	>500	>500	XX	113 ± 0.1	>500	XX	83 ± 0.01	XX	>500	184.6 ± 0.1
F9	>500	>500	XX	188 ± 0.1	>500	XX	28.1 ± 0.01	XX	356 ± 0.1	150.4 ± 0.1
F10	>500	>500	XX	> 500	>500	XX	15.3 ± 0.01	XX	>500	138.4 ± 0.1
Centratherin a	1.5 ± 0.1	112.2 ± 0.1	209.8	92.7 ± 0.1	>500	3.4	0.45 ± 0.01	699.3	138.1 ± 0.1	314.7 ± 0.1
Goyazensolide a	2.5 ± 0.1	>500	127.7	90.6 ± 0.1	>500	3.5	0.34 ± 0.01	939.1	139.9 ± 0.1	319.3 ± 0.1
L‐NMMA b	XX	XX	XX	XX	XX	XX	19.37 ± 1.5	XX	XX	XX
Rifampicin c	0.009 ± 0.1	0.12 ± 0.1	XX	0.99 ± 0.1	2.75 ± 0.1	XX	XX	XX	XX	XX

Full text

Introduction

1

Tuberculosis (TB) remains a major global health concern. In 2024, an estimated 10.7 million people developed TB, and approximately 1.23 million died from the disease worldwide, with Brazil ranking among the 30 countries with the highest burden. The growing incidence of drug‐resistant TB poses a significant challenge, with an estimated number of 390.000 in 2024 [1]. In severe cases, adjuvant therapies, such as anti‐inflammatory agents, are essential to mitigate excessive inflammation, a key factor in TB pathogenesis. Despite advances in therapy, prolonged treatment, adverse effects, and resistance continue to compromise adherence and cure rates [2]. The increasing number of strains resistant to first‐line treatment has led to a rise in drug‐resistant TB (DR‐TB) cases, making this scenario even more concerning, particularly for immunocompromised patients [3]. These challenges highlight the urgent need to explore new therapeutic alternatives, including bioactive natural products with dual antimycobacterial and anti‐inflammatory potential [4].

Currently, natural products represent a vast source of therapeutic alternatives for numerous health challenges. Historically, the discovery of streptomycin and the development of rifampicin, both derived from living organisms, were essential to the advancement of tuberculosis (TB) treatment [5]. Brazil harbors the world's largest and most diverse biodiversity, reflected in an unparalleled array of secondary metabolites, compounds that play essential roles in plants [6]. Among Brazil's biomes, the restingas are characterized by distinctive floristic composition and vegetation structure, covering approximately 80% of the Brazilian coastline [7].

The evaluation of plant species from restingas is fascinating due to the unique edaphoclimatic characteristics and the specific nutrient availability in these environments, which directly influence the local flora and, consequently, their chemical compositions [8]. Eremanthus crotonoides (DC) Sch. Bip. (Asteraceae) is a plant species native to the Brazilian restinga and cerrado and is used in folk medicine [9]. Regarding the chemical composition of Asteraceae species, this family is characterized by a predominance of terpenoids, including mainly sesquiterpene lactones [10]. Specifically, numerous sesquiterpene lactones have been consistently reported in the aerial parts of E. crotonoides, and interest in this class of compounds stems from their high and diverse biological potential [9, 11, 12].

In a previous screening study conducted by this research group [13], the ethanolic extract and nonpolar fractions of E. crotonoides demonstrated both antimycobacterial and anti‐inflammatory activities, highlighting this species as a promising source of antitubercular compounds with potential efficacy under hyperinflammatory conditions. Therefore, the present study aims to perform bioassay‐guided fractionation of the ethanolic extract from E. crotonoides leaves (ECE), isolate compounds from the most bioactive fraction, and evaluate their antimycobacterial activity against Mycobacterium tuberculosis (Mtb) H_37_Rv and the hypervirulent M299 strains, as well as their anti‐inflammatory activity. High‐speed countercurrent chromatography (HSCCC) was employed due to its advantages over traditional partitioning methods, including low solvent consumption, complete sample recovery, and reduced processing time [14]. This study reports, for the first time, the separation of E. crotonoides extract by HSCCC and the isolation of two sesquiterpene lactones, centratherin and goyazensolide. In addition, it presents the first evaluation of the antimycobacterial activity of these compounds against Mtb H_37_Rv and M299 strains.

Results and Discussion

2

Isolation and Characterization of Bioactive Compounds

2.1

The ethanolic extract of E. crotonoides (ECE) yielded 23.06 g (5%). Preparative separations of 350 and 700 mg of ECE (CCC1 and CCC2, respectively) using the n‐hexane:acetonitrile (1:1, v/v) system afforded 10 fractions in both cases. The intra‐apparatus volumetric scale‐up procedure applied to the CCC2 confirmed the method's reproducibility and increased fraction yields [15] (Figures S1 and S2; Table S1). In the UHPLC‐DAD‐MS/MS chromatogram of F1, two major peaks are observed at 29.183 and 31.997 min (Figure 1).

Chromatogram obtained from the UHPLC‐DAD‐MS/MS analysis of the F1 fraction derived from the EtOH extract of E. crotonoides.

Regarding the major peak (31.997 min), a precursor ion of m/z 375 (M + H)^+^ is detected, along with fragment ions at m/z 275 and 257 (Figure S3). This pattern was previously described in the annotation of centratherin, a sesquiterpene lactone present in the dichloromethane fraction of E. crotonoides leaves [11]. Therefore, the product ion at m/z 275 is consistent with the loss of the ester linked to carbon C‐8, eliminated as its corresponding carboxylic acid (C_4_H_6_CO_2_H), and the ion m/z 257 arises from the loss of a water molecule from the previous product ion. For the peak observed at 29.183 min, the precursor ion at m/z 361 is detected in the positive ion mode (Figure S3), fragment ions are formed at m/z 275, resulting from the loss of an ester group attached to C‐8, eliminated as a carboxylic acid (C_3_H_5_CO_2_H); m/z 257, arising from the loss of a water molecule, forming the ion [M+H—C_3_H_5_CO_2_H—H_2_O]^+^; and m/z 229, resulting from the loss of a CO molecule from the previous product ion. This fragmentation pattern is characteristic of goyazensolide, a sesquiterpene lactone [16]. These data are summarized in Table 1.

Sesquiterpene lactones constitute an important class of terpenoids derived from the mevalonic acid pathway. Their structures consist of fifteen carbon atoms, comprising three isoprene units (C_5_), and they undergo regioselective oxidation (β‐carboxylation at C‐7). The addition of a hydroxyl group at C‐6 or C‐8 leads to the formation of the lactone ring [17]. This class is notable for its pharmacological properties, particularly its anti‐inflammatory, trypanocidal, antifungal, antitumor, anxiolytic, antibacterial, and analgesic activities. It has also shown promising results when combined with other drugs, thereby broadening its clinical applications [11, 12, 18, 19].

Fraction 1 (F1, tubes 1–6) exhibited the strongest antimycobacterial activity and was selected for further purification by HSCCC. Using the HEMWat‐DMSO system (4:6:4:6:0.1, v/v/v/v/v), F1 was resolved into 16 fractions, from which two sesquiterpene lactones were isolated: compound 1 (tubes 22–23, 1.8 mg) and compound 2 (tubes 39–42, 1.5 mg). HPLC‐DAD analysis of the fractions confirmed the isolation of two compounds with purity greater than 90%: compound 1 (λ = 268 nm, 96.1%) and compound 2 (λ = 266 nm, 91%) (Figures 2 and 3). The absorption maxima are consistent with those reported for this group of secondary metabolites [12].

Chromatogram by HPLC‐DAD (λ = 254 nm) of compound 1 with its respective UV spectrum, isolated from the F1 fraction from E. crotonoides.

Chromatogram by HPLC‐DAD (λ = 254 nm) of compound 2 with its respective UV spectrum, isolated from the F1 fraction from E. crotonoides.

The fractionation of the extract and the isolation of both compounds have already been reported in the literature; however, these studies employed conventional liquid–liquid partitioning and column chromatography to obtain the purified compounds. Such techniques are characterized by high solvent consumption and longer processing times, whereas the approach used in the present study features reduced solvent usage and a more time‐efficient workflow. Regarding yields, the recovery of centratherin was slightly lower than that reported in other studies. Nevertheless, it is well established that several environmental factors may contribute to variations in the levels of this metabolite, particularly considering that the plant material was collected during markedly different periods, since HSCCC overcomes the limitations associated with other separation techniques, especially regarding sample recovery, owing to the absence of a solid stationary phase [11, 12].

Compound 1 was obtained as an orange oil. The ^1^H NMR spectrum revealed a singlet at δ 5.93 (s, 1H) corresponding to H‐2 in the 3‐(2H)‐furanone core. The triplet at δ 6.07 (t, 1H) and the doublet at δ 5.66 (d, J = 2.7 Hz, 1H) represent, respectively, two olefinic hydrogens H‐3’a and H‐3’b forming a methacrylate, which is particular from this structure when compared with furanohealigolides already reported, such as lychnofolide. The doublet at δ 6.16 (d, J = 3.1 Hz, 1H) and the multiplet between δ 5.65–5.63 (m, 1H) are assigned to H‐13a and H‐13b, respectively, originating from a γ‐lactone conjugated with a methylene group. The doublet of quartets at δ 5.35 (dq, J = 5.0, 2.4 Hz, 1H) corresponds to H‐6. The doublet of triplets at δ 6.24 (dt, J = 3.2, 1.7 Hz, 1H) refers to the vinylic hydrogen at H‐5. The doublet of triplets at δ 4.53 (dt, J = 11.8, 2.4 Hz, 1H) corresponds to H‐8. The multiplet between δ 4.35 and 4.30 (m, 3H) is assigned to the methylene hydrogens H‐15a and H‐15b. The doublet of quartets at δ 3.86 (dq, J = 5.6, 2.8 Hz, 1H) represents H‐7. The double doublets at δ 2.74 (dd, J = 14.0, 11.8 Hz, 1H) and δ 2.26 (dd, J = 14.0, 2.0 Hz, 1H) correspond to H‐9a and H‐9b. Finally, the strong singlet at δ 1.53 (s, 3H) is attributed to the methyl hydrogens at carbon C‐14. The ^13^C spectra showed 20 carbon signals, including four methylene carbons δ 127.1 (C‐3’), 125.7 (C‐13), 63.2 (C‐15), and 44.5 (C‐9)), two methyl carbons δ 20.8 (C‐14) and 18.1 (C‐4’)) and four methyne carbons δ 137.0 (C‐5), 107.5 (C‐2), 83.4 (C‐6), 75.2 (C‐8), and 52.2 (C‐7)). The NMR data obtained were similar to those reported in the literature for goyazensolide (1) (Figure 4) and were consistent with data obtained by UHPLC‐DAD‐MS/MS [20].

E. crotonoides annotated compounds proposed by mass spectrometry analysis, and further identified by NMR, from the bioactive fraction obtained from the ethanolic extract through HSCCC separations.

Compound 2 was a white, amorphous powder. Its ^1^H NMR spectrum showed a quintuplet at δ 4.33 (q, J = 1.9 Hz, 1H), attributable to H‐8. The high‐field singlet at δ 1.51 (s, 3H) corresponds to the methyl hydrogens at C‐14. Another singlet appears at δ 5.94 (s, 1H), corresponding to H‐2, located in the 3‐(2H)‐furanone core, which characterizes the furanoheliangolide subclass. The doublet of triplets at δ 6.25 (dt, J = 3.2, 1.7 Hz, 1H) refers to the vinylic hydrogen at H‐5. At δ 2.25 (dd, J = 13.9, 1.9 Hz, 1H) and δ 2.74 (dd, J = 13.9, 12.0 Hz, 1H), the double doublets correspond to H‐9b and H‐9a, respectively. The pair of doublets at δ 5.63 (d, J = 2.7 Hz, 1H) and δ 6.17 (d, J = 3.1 Hz, 1H) are characteristic of a γ‐lactone conjugated with a methylene group and are assigned to the H‐13a and H‐13b hydrogens.

H‐7 is represented by the doublet of quartets at δ 3.84 (dq, J = 5.5, 2.8 Hz, 1H). The structure shows a close similarity to compound 1; however, the presence of a multiplet at δ 1.82–1.79 (m, 3H) and δ 1.82 (p, J = 1.5 Hz, 3H) indicates the presence of two methyl groups, corresponding to the hydrogens attached to C‐5' and C‐4' (of the angelate group). The absence of signals at δ 6.02 and 5.55, along with the presence of a signal at δ 6.20‐6.12 (qq, H), indicates that only one hydrogen is attached to C‐3', giving an angelate rather than a methacrylate, as observed in compound 1. The ^13^C spectrum showed 20 carbon signals, including three methylene carbons δ 128.0 (C‐13), 63.2 (C‐15), and 44.7 (C‐9)), three methyl carbons δ 20.8 (C‐14), 20.2 (C‐5’) and 15.9 (C‐4’)) and five methyne carbons δ 141.4 (C‐5), 125.6 (C‐2), 83.6 (C‐6), 74.9 (C‐8), and 52.2 (C‐7)). These findings enabled the identification of centratherin (2) (Figure 4), another sesquiterpene lactone previously described in E. crotonoides [17].

The ^1^H and ^13^C NMR data acquired for both compounds are presented in Table S2 and S3 (supporting file S1). This is the first report of the isolation of these compounds by HSCCC. Moreover, the technique could separate two structurally similar compounds with high purity in a single chromatographic step.

Antimycobacterial and Anti‐Inflammatory Activities

2.2

For the assays, the laboratory reference strain Mtb H_37_Rv and the highly virulent clinical isolate M299, belonging to the modern Beijing sublineage, were used. Hypervirulent Mtb strains exhibit accelerated intracellular growth and can induce necrotic death of infected macrophages [21, 22]. Epidemiological studies have documented the increasing global dissemination of Beijing lineage strains, which are associated with increased pathogenicity, drug resistance, and disease outbreaks [21, 23]. The MIC_50_ values determined for ECE were 42.0 ± 0.1 µg/mL, with a selectivity index (SI) of 7.1 for the M. tuberculosis H_37_Rv strain, and 39.0 ± 0.1 µg/mL, with an SI of 7.6 for the clinical isolate M299 (Table 2). Although both values were below the threshold of 128 µg/mL, which is considered indicative of antimycobacterial activity, the M299 strain exhibited slightly higher sensitivity than H_37_Rv. Therefore, these data reinforce the potential of ECE as a source of antitubercular compounds and underscore the importance of accounting for strain‐specific differences in the development of therapeutic strategies.

Although the immune system plays a crucial role in controlling and/or eliminating Mtb infection, the balance between host‐ and bacillus‐related factors can trigger an excessive immunological response, leading to tissue damage. The modulation of NO and TNF‐α production by these samples was also evaluated, as Mtb‐infected macrophages upregulate NO and TNF‐α levels, and excessive production of both mediates pathogenicity and lethality in severe forms of TB [24, 25]. This dual approach, which investigates both antimycobacterial and anti‐inflammatory activities, is well established in the literature for more severe cases of TB and has been extensively explored by our research group [13, 26]. ECE showed IC_50_ values for NO and TNF‐α of 0.61 ± 0.01 and 261.1 ± 0.1 µg/mL, respectively, and a CC_50_ value of 298.5 ± 0.1 µg/mL in the cytotoxicity assay, indicating good selectivity and excellent inhibitory activity against nitric oxide synthesis (Table 2).

Among the fractions derived from ECE, F1 stood out with MIC_50_ values of 24.0 ± 0.1 and 55.0 ± 0.1 µg/mL against Mtb strains H_37_Rv and M299, respectively. Although all fractions exhibited inhibitory activity against NO production, F1 and 10 were notable, presenting IC_50_ values of 2.0 ± 0.01 and 15.3 ± 0.01 µg/mL, respectively. These values were lower than those obtained for the standard NG‐monomethyl‐L‐arginine (L‐NMMA), which is particularly noteworthy given that these samples were still fractions rather than isolated compounds. However, F1 showed the highest cytotoxicity (CC_50_ 6.3 ± 0.1 µg/mL) and lacked inhibitory activity on TNF‐α production (Table 2). Goyazensolide (1) and centratherin (2) displayed potent antimycobacterial activity against the Mtb H_37_Rv strain, with MIC_50_ values of 1.5 ± 0.1 and 2.5 ± 0.1 µg/mL, respectively (Figure 5A). These activities were associated with high selectivity, as reflected by SI values of 209.8 for goyazensolide and 127.7 for centratherin. In contrast, both compounds showed reduced activity against the hypervirulent Mtb M299 strain, with MIC_50_ values of 92.7 ± 0.1 and 90.6 ± 0.1 µg/mL, respectively, corresponding to SI values of 3.4 and 3.5, respectively (Figure 5B). The differential activity observed between the reference strain Mtb H_37_Rv and the hypervirulent clinical strain M299 reflects well‐characterized phenotypic and adaptive differences between these lineages. M299, exhibits a distinct virulence profile, including accelerated intracellular growth, induction of necrotic death in infected macrophages, and the ability to trigger an intense inflammatory response characterized by excessive neutrophil recruitment and extensive pulmonary necrotic lesions in murine models, features not observed during infection with H_37_Rv. These characteristics indicate relevant differences in bacterium: host interactions and intracellular adaptation. In addition, epidemiological studies have demonstrated the increasing global dissemination of Beijing lineage strains, which are associated with enhanced pathogenicity, drug resistance, and tuberculosis outbreaks [21, 22]. Collectively, these factors provide a consistent biological basis for the distinct behavioral and response profiles observed between the H_37_Rv and M299 strains.

Effect of samples on Mtb growth in bacterial culture. Bacterial suspensions (1 × 106 CFU/well) of Mtb strain H37Rv (A) and clinical isolate M299 (B) were treated or untreated with samples (0.031, 0.16, 0.8, 4, 20, 100, and 500 µg/mL) for 5 days. The MTT test quantified bacterial growth in the resulting cultures. As a negative control (C‐), it was used to supplement 7H9 medium, and as a positive control (C+), the untreated mycobacterial suspension. The seven bars for each compound refer to concentrations tested in ascending order. Values are represented as mean ± standard deviation, and different groups were considered significant at p < 0.05 (), p < 0.01 (), and p < 0.001 () (n = 3).

This is the first report of their activity against Mtb strains with higher virulence, such as H_37_Rv and the hypervirulent M299 strain. Previously, the only reported antimycobacterial activity of these compounds was against the H_37_Ra strain, which exhibits reduced virulence [27]. It is essential to emphasize the efficiency of HSCCC in concentrating the most active compounds into a single fraction with a higher yield. Considering the decreasing IC_50_ values observed for the extract, fractions, and isolated compounds, it is evident that the antimycobacterial activity of the ECE is attributable to the presence of these sesquiterpene lactones.

Although they possess very similar chemical structures, centratherin bears a methacrylate ester substituent, whereas goyazensolide contains an angelate moiety. The main mechanism of action of this class of compounds is generally attributed to the presence of the α‐methylene‐γ‐butyrolactone moiety in their structures. Due to its high reactivity, this moiety can covalently bind to thiol groups (R–SH) present in cysteine residues, which are widely distributed in biomolecules. This binding occurs via Michael addition reactions, resulting in alkylation [29]. However, this variation in the substitution pattern can lead to differences in the three‐dimensional geometry and torsion angles, even though both compounds have the same absolute configuration at the core chiral centers (C‐6, C‐7, and C‐8), which can influence biological interactions [19, 29].

Therefore, mycothiol‐dependent redox enzymes and cysteine‐rich regulatory proteins are potential targets of these compounds. The markedly lower susceptibility observed for the hypervirulent M299 strain may reflect altered cell envelope permeability and/or increased detoxification mechanisms, all of which have been associated with hypervirulent phenotypes [28, 30].

The modulation of NO and TNF‐α production by these samples was also evaluated. In general, the major compounds in the fractions did not display significant inhibitory activity against TNF‐α production. Regarding NO inhibition, centratherin (1) exhibited an IC_50_ of 0.45 ± 0.1 µg/mL, while goyazensolide (2) showed an IC_50_ of 0.34 ± 0.1 µg/mL, both lower than that of L‐NMMA, highlighting the significant potential of these compounds as inhibitors of nitric oxide synthesis. In the TNF‐α inhibition assays, both compounds exhibited lower levels of activity, displaying IC_50_ values of 138.1 ± 0.1 and 139.0 ± 0.1 µg/mL, respectively (Table 2). Additionally, the potential of certain sesquiterpene lactones to inhibit nitrite accumulation, dependent on inducible nitric oxide synthase (iNOS), in LPS‐stimulated RAW 264.7 macrophages has already been reported, with himenin, alantolactone, and helenalin identified as the most active compounds. Moreover, the presence of the α‐methylene‐γ‐butyrolactone moiety has been identified as a key structural feature of NO‐inhibitory activity [31].

The in vivo inhibitory activity of TNF‐α synthesis by goyazensolide has also been previously described [32]. Importantly, neither goyazensolide nor centratherin showed significant cytotoxicity in our assays using RAW 264.7 macrophages, with CC_50_ values of 319.3 ± 0.1 and 314.7 ± 0.1 µg/mL, respectively, indicating selectivity across all evaluated activities (Table 2).

It is important to note that α‐methylene‐γ‐lactone‐containing compounds are associated with known pharmacological liabilities, including limited aqueous solubility and non‐selective Michael acceptor‐mediated binding, which have been documented for sesquiterpene lactones and can lead to increased toxicity, undesired effects, and shorter bioavailability [33, 34].

Despite the observed high potential, further studies are necessary to establish the antitubercular and anti‐inflammatory profiles of these compounds and to conduct in vitro evaluations against other clinically relevant strains, underscoring the need to isolate larger quantities of both compounds.

Conclusions

3

HSCCC proved to be an efficient, reproducible, and scalable approach for the bioassay‐guided fractionation of the ethanolic extract of E. crotonoides leaves by enabling the concentration of bioactive compounds into a single fraction (F1), which exhibited the strongest antimycobacterial activity among all fractions, with MIC_50_ values of 24.0 ± 0.1 µg/mL against Mycobacterium tuberculosis H_37_Rv and 55.0 ± 0.1 µg/mL against the hypervirulent M299 strain. The intra‐apparatus volumetric scale‐up confirmed the method's robustness and reproducibility while increasing fraction yields. Subsequent HSCCC purification of F1 afforded goyazensolide (1, 1.8 mg) and centratherin (2, 1.5 mg), two sesquiterpene lactones with purity above 90%. Structural identification was achieved by UHPLC‐DAD‐MS/MS and NMR analyses and was consistent with previously reported data. Notably, this represents the first isolation of these compounds by HSCCC. Goyazensolide and centratherin exhibited potent antimycobacterial activity against the virulent H_37_Rv strain, with MIC_50_ values of 2.5 ± 0.1 and 1.5 ± 0.1 µg/mL, respectively. In contrast, both compounds were markedly less active against the hypervirulent M299 isolate (MIC_50_ = 90.6 ± 0.1 and 92.7 ± 0.1 µg/mL, respectively). These results demonstrate a clear enrichment of antimycobacterial activity from the crude extract (MIC_50_ = 42.0 ± 0.1 µg/mL for H_37_Rv) to the isolated compounds.

In parallel, both sesquiterpene lactones exhibited potent and selective inhibition of nitric oxide production in LPS‐stimulated macrophages, with IC_50_ values of 0.34 ± 0.1 µg/mL for goyazensolide and 0.45 ± 0.1 µg/mL for centratherin. In addition, neither compound showed significant cytotoxicity toward RAW 264.7 macrophages, with CC_50_ values of 319.3 ± 0.1 and 314.7 ± 0.1 µg/mL, respectively, indicating favorable selectivity indices.

Therefore, further in vivo and complementary in vitro studies, pharmacokinetic evaluations, and mechanistic investigations are warranted to characterize the antimycobacterial potential of these compounds fully and to explore formulation strategies and activity against additional clinically relevant Mtb strains.

Experimental Section

4

Chemicals and Reagents

4.1

HPLC‐grade organic solvents used in HSCCC (High‐Speed Countercurrent Chromatography), UHPLC‐DAD‐MS/MS (Ultra‐Performance Liquid Chromatography coupled to Diode Array and Mass Spectrometry detectors and HPLC‐DAD (High‐Performance Liquid Chromatography Coupled To Diode Array Detector) analyses were purchased from Tedia (Rio de Janeiro, Brazil), and the ultrapure water was supplied by an ElgaPurelab water purification system (Wycombe, UK). For TLC analysis, analytical‐grade organic solvents were purchased from Tedia (Rio de Janeiro, Brazil) and SiliaPlate TLC plates (Silica, F_254_) from Silicycle (Québec, Canada).

The OADC and ADC supplements were from BD Biosciences (BD, Sparks, MD, USA). The Mycobacteria‐specific Middlebrook 7H9 and 7H10 media were purchased from Difco (Detroit, MI, USA). DMEM‐F12 culture media from Gibco/Invitrogen (Grand Island, NY, USA). NG‐MonoMethyl‐l‐Arginine acetate salt (L‐NMMA)—inhibitor of NOS (cod. M7033), Rifampicin (cod. R7382), Lipopolysaccharide (LPS) from serotype 0111:B4 Escherichia coli, and 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) were obtained from Sigma‐Aldrich Co. (St. Louis, MO, USA). The samples were dissolved in dimethyl sulfoxide (DMSO, Sigma‐Aldrich, St. Louis, MO, USA) for use in cell culture.

Plant Material

4.2

E. crotonoides (DC) Sch. Bip., Asteraceae: leaves were collected in February 2020 at the Restinga de Jurubatiba National Park, Quissamã, Rio de Janeiro, Brazil (22.19828 °S; 41.46338 °W; 10 m altitude). The identification of the specimen was performed by the botanist Prof. Dra. Tatiana Ungaretti Paleo Konno, and the voucher (RFA‐38749) was deposited in the “Instituto de Biodiversidade e Sustentabilidade (NUPEM)” herbarium at the Federal University of Rio de Janeiro, Macaé, Brazil.

Extraction

4.3

The leaves of E. crotonoides (461.180 g) were air‐dried and extracted to exhaustion using analytical grade ethanol in a percolator. The resulting ethanolic extract was filtered and concentrated under reduced pressure using a rotary evaporator (Heidolph) at 200 mBar and 45°C–50°C to ensure complete solvent removal.

TLC Analyses

4.4

The choice of the solvent system and the chemical evaluation of the fractions obtained were performed using silica gel F_254_ TLC plates (Silicycle, Québec, Canada), developed with chloroform:methanol (9.5:0.5, v/v). The plates were observed under UV light (254–365 nm) and later treated with vanillin sulphuric acid reagent (vanillin 1% and sulphuric acid 10% in methanol).

HSCCC Equipment and Separation Procedure

4.5

The HSCCC separation of the ethanolic extract and the subsequent purification of the constituents of the bioactive fraction (F1) were performed on a QuikPrep Mk8 (AECS‐QuikPrep Ltd, United Kingdom) instrument equipped with two bobbins containing two polytetrafluoroethylene (PTFE) multi‐layer coils each (2.1 mm i.d., 4x112 mL). The system was operated at 860 rpm in head‐to‐tail mode. Both separations were carried out at a flow rate of 2.0 mL/min and a temperature of 30°C. Fractions of 4 mL/tube were collected. The established methodology was then applied to two columns to increase the yield of the fractions obtained. To this end, the two columns were connected, resulting in a total volume of 224 mL, a loop volume of 10 mL, a flow rate of 4 mL/min, 8 mL/tube, and a sample of 700 mg.

The n‐hexane:acetonitrile (1:1, v/v) system was employed for the preparative separation of 350 and 700 mg of E. crotonoides ethanolic extracts (CCC1 and CCC2, respectively). The detailed conditions are presented in Table S1. CCC1 resulted in 10 fractions after TLC analysis: tubes 1–6 (52 mg), tubes 7, 8 (10.1 mg), tubes 9–11 (22.8 mg), tubes 12–18 (23.9 mg), tubes 19–25 (20.9 mg), tubes 26–37 (22.7 mg), tubes 38–53 (8 mg), 54–63 (9 mg), 64–80 (9.2 mg), and tube 81 (50 mg).

In the second separation (CCC2), a scale‐up of CCC1 was performed by applying the TF≈2,2 scaling factor to the column volume, the mass of the ethanolic extract, the flow rate, and the sample volume. It also resulted in 10 fractions: tubes 1–6 (103.3 mg), tubes 7, 8 (23.5 mg), tubes 9–15 (46 mg), tubes 16–21 (48.1 mg), tubes 22–25 (42.2 mg), tubes 26–41 (44 mg), tubes 42–53 (14 mg), tubes 54–63 (18 mg), tubes 64–80 (18.1 mg), and tube 81 (100.1 mg).

Due to its bioactivity, sample F1, corresponding to the first fraction (tubes 1–6), underwent a third HSCCC (CCC3) fractionation to isolate its sesquiterpene lactones. In this case, the solvent system used was HEMWat with DMSO 4:6:4:6:0.1 (v/v/v/v/v). Sixteen final fractions were obtained: tubes 1–5 (0 mg), tubes 6–18 (7 mg), tubes 19–21 (1.9 mg), tubes 22, 23 (1.8 mg), tubes 24, 25 (1.6 mg), tubes 26–38 (6.1 mg), tubes 39–42 (1.5 mg), tubes 43–47 (3.7 mg), tubes 48–50 (2.1 mg), tubes 51–56 (4.9 mg), tubes 57–64 (5.3 mg), tubes 65–67 (4.2 mg), tubes 68, 69 (3.8 mg), tubes 70–75 (5.1 mg), tubes 76–78 (10.6 mg), and tubes 79–90 (15.3 mg). This procedure allowed the isolation of centratherin (1.5 mg) and goyazensolide (1.8 mg).

HPLC‐DAD

4.6

HPLC‐DAD analyses were performed on a Shimadzu HPLC system (Kyoto, Japan) equipped with an LC‐20AT quaternary solvent pump unit, SPD‐M20A diode array detector, SIL‐20A automatic injector, CBM‐20A controller, CTO‐20A column oven, DGU‐20A5 degasser, LC solution software, and DAD set between 200 and 450 nm. The chromatographic column used was a Phenomenex C18 Luna (250 mm x 4.6 mm; 5 µm). The samples were prepared at 1 mg/mL in acetonitrile. The mobile phase consisted of water and acetonitrile (95:5, A) and acetonitrile and water (95:5, B), both acidified with 0.05% trifluoroacetic acid (TFA). The elution system employed was: 0 min (10% B); 7 min (20% B); 22 min (40% B); 25 min (50% B); 43 min (100% B); 55 min (10% B). The flow rate was 1 mL/min, and the temperature was 40°C [9].

UHPLC‐DAD‐MS/MS

4.7

UHPLC‐DAD‐MS/MS analyses were performed on a Thermo Scientific LCQ FLEET UHPLC system (Thermo Fisher Scientific, Waltham, USA) equipped with a DAD detector and coupled to a mass spectrometer with an atmospheric pressure chemical ionization (APCI) source (LCQ, Thermo Fisher Scientific, Waltham, USA). A Kinetex Polar C18 column (100 x 4.6 mm; 2.6 µm) was used. The substances were separated using the following gradient: 0 min (5% B); 5 min (5% B); 50 min (95% B); 55 min (95% B); 55.1 min (5% B); 60 min (5% B), with a flow rate set to 0.5 mL/min and temperature maintained at 40°C. The mobile phase consisted of water (A) and acetonitrile (B).

MS measurements were performed with helium (99.999% purity) as the collision gas in the ion trap and nitrogen as the sweep, auxiliary, and sheath gas in the source. MS parameters were tuned in the APCI positive‐ionization mode: capillary temperature, 275°C; source voltage, 5.50 kV; mass range, 500–1200 Da. The CID (Collision‐induced dissociation) was 35 eV.

NMR Identification of the Compounds

4.8

^1^H NMR and ^13^C data for the purified compounds were obtained on a Varian VNMRSYS500 (California, USA) at 25°C, operating at 500 MHz. The samples were prepared in CD_3_OD.

Determination of Mycobacterial Growth Inhibition

4.9

The 10 fractions of the ethanolic extract of E. crotonoides, as well as the samples isolated from the F1 fraction, underwent biological assays to determine the inhibition potential of mycobacterial growth.

The laboratory reference strain Mtb H_37_Rv was obtained from ATCC (27294). The Beijing Mtb strain, Mtb M299, is a clinical isolate from Mozambican TB patients, provided by Dr. Philip Suffys (Oswaldo Cruz Institute/RJ), and was previously identified as hypervirulent [21]. The strains were cultivated in Middlebrook 7H9 medium (DIFCO, Detroit, MI), supplemented with 0.05% Tween 80 and 10% ADC (albumin, dextrose, catalase—Difco Laboratories) for 7 days until reaching the mid‐log phase (maximum growth) in a bacteriological incubator at 37°C and 5% CO_2_. Mycobacterial growth was monitored by measuring optical density (OD at 600 nm) using a spectrophotometer (Libra S6, Bichrom).

Upon reaching the logarithmic growth phase, Mtb cultures in suspension were vortexed (Biomatic) and sonicated in an ultrasonic bath (Ultrasonic Maxi Cleaner 800, Unique) for 1 min to dissolve clumps. The adjustment of mycobacterial concentration was done by reading 1 mL of the culture suspension (7H9) in a spectrophotometer (U1100 model, Hitachi) at 600 nm (subtracting the OD of the 7H9 culture medium only supplemented). Subsequently, the cultures were diluted according to the OD‐bacillus count criterion (OD 0.100–2 × 10^7^ bacilli/mL) for both strains. Fifty microliters of the suspension of each strain were plated on 96‐well plates (1 x 10^7^ CFU/mL). Then, 50 µL of E. crotonoides ethanolic extract and fractions (4, 20, 100, and 500 µg/mL), and the isolated compounds (0.032, 0.16, 0.8, 4, 20, 100, and 500 µg/mL) were added in duplicate for both strains. Subsequently, the plates were incubated at 37°C for 5 days. At the end of this period, 10 µL of MTT solution was added to each well. After 3 h, 100 µL of lysis buffer composed of 20% w/v Sodium Dodecyl Sulfate (SDS) and 50% v/v Dimethylformamide (DMF) (Sigma Aldrich) was added to each well. The reading was performed on a plate spectrophotometer at 570 nm. The supplemented 7H9 culture medium with bacteria and the untreated culture medium were used as positive and negative controls, respectively. It is worth noting that all experiments involving Mtb adhered to biosafety level III standards.

Cell Culture and Treatment With E. crotonoides Fractions

4.10

Murine RAW 264.7 macrophage cell lines from ATCC (VA, USA) were cultured in DMEM‐F12 culture medium supplemented with 10% fetal bovine serum (FBS) (Gibco/Invitrogen, NY, USA) at 37°C and 5% CO_2_. For experiments, RAW 264.7 macrophages were seeded in 96‐well plates (5x10^5^ cells/mL). After 24 h, the culture medium was replaced by DMEM‐F12 medium supplemented with 2% FBS, in the presence or absence of E. crotonoides ethanolic extract and fractions (4, 20, 100, and 500 µg/mL), the isolated compounds (0.032, 0.16, 0.8 e 4 µg/mL) and/or LPS (1 µg/mL, E. coli 055:B5; Sigma‐Aldrich, USA). The plate was maintained at 37°C in 5% CO_2_. After 24 h, the culture supernatant was collected for the evaluation of inflammatory parameters. Murine L929 fibroblasts (2.5x10^5^ cells/mL) were plated with DMEM‐F12 supplemented with 10% FBS 24 h before collecting supernatant from the stimulated macrophage culture. For the assay, the supernatant from the L929 cell culture was removed, and 50 µL of DMEM‐F12 supplemented with 10% FBS and actinomycin D (2 µg/mL) was added to each well. Then, 50 µL of the supernatant collected from the LPS‐stimulated RAW 264.7 plate was added to each well.

Quantification of Nitric Oxide (NO) by the Griess Method

4.11

By determining nitrite concentration, NO synthesis by macrophages was estimated. For this, 50 µL of each supernatant (as described in the previous section) was added to 50 µL of Griess reagent (consisting of 1% p‐aminobenzenesulfonamide + 0.1% naphthylethylenediamine dihydrochloride in 5% phosphoric acid; Sigma‐Aldrich) and incubated for 10 min. Absorbance was then measured at 570 nm using a plate spectrophotometer (Dinatech MR5000). The nitrite concentration in the supernatant was determined in µM using the sodium nitrite curve and then subtracted from the value obtained with cell‐free reagents. The NO concentrations (µM) in the supernatant of each sample were compared with the negative control (macrophage culture without stimulation and/or treatment) and the positive control (macrophages stimulated with LPS, without treatment). The NO inhibitor L‐NMMA (20 µg/mL) was used as the standard.

Evaluation of Cytotoxicity in Macrophages

4.12

Cellular cytotoxicity of E. crotonoides fractions was determined by assessing the metabolization of 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium (MTT) to a formazan crystal in living cells. The cell culture was treated with the samples according to the previously described method, and after 24 h of incubation, 10 µL of MTT solution (5 mg/mL; Sigma‐Aldrich) was added to each well. Incubation was carried out for 2 h at 37°C and 5% CO_2_. Finally, the cell culture supernatant was removed, and the formazan crystals formed were diluted in acidified isopropanol with 4.0 mM hydrochloric acid (HCl). Reading was also performed on a plate spectrophotometer at 570 nm (Dinatech MR5000). Viability (%) determined by the MTT method was converted to cytotoxicity percentage using values obtained from positive and negative controls, following the formula: 100 ‐ ((OD.sample—OD.C^+^) * 100)/(OD.C^–^ OD.C^+^). The culture of stimulated and untreated macrophages was used as a positive control, supplemented with 10 µL of 0.1% Triton X‐100. The negative control, by contrast, consisted only of cultures of stimulated and untreated macrophages. Additionally, it is worth noting that the solvent control, DMSO, used for substance dilution was applied in this experiment.

Indirect Bioassay for TNF‐α Production Quantification

4.13

Using an indirect bioassay, TNF‐α can be assessed in the culture supernatant of RAW 264.7 macrophages. Murine L929 fibroblasts are known to be sensitive to the presence of TNF‐α; therefore, the amount of TNF‐α in the culture supernatant correlates with the cell death index.

L929 fibroblast culture was plated at a concentration of 2.5 x 10^5^ cells/mL in a 96‐well plate and incubated at 37°C, 5%CO_2_ for 24 h. Subsequently, L929 cells were incubated for 24 h in the presence or absence of supernatant from RAW 264.7 macrophages, as previously described. Using the MTT method, cellular viability was determined. The concentration of TNF‐α in the supernatant was quantified in pg/mL using a standard curve of recombinant murine TNF‐α. Macrophages without stimulation and treatment were used as the negative control for TNF‐α production, while stimulated and untreated macrophages were used as the positive control.

Statistical Analysis

4.14

The CC_50_, MIC_50_, and IC_50_ values were determined through nonlinear regression analysis of concentration–response curves, based on data from three independent experiments, each performed in triplicate. The curves were constructed from the logarithms of sample concentrations, and the analysis was performed using GraphPad Prism version 4.0 (GraphPad Software Inc., San Diego, CA, USA).

Statistical analysis was performed using one‐way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test, using GraphPad Prism 4 to assess significant differences between groups. A p‐value < 0.05 was considered statistically significant.

Author Contributions

Natalie Giovanna da Rocha Ximenes: conceptualization (lead), data acquisition, analysis and interpretation (equal), writing – original draft (lead), writing – review and editing (equal). Sanderson Dias Calixto: conceptualization (lead), data acquisition, analysis, and interpretation (equal); writing – review and editing (equal). Thatiana Lopes Biá Ventura: writing – review and editing (equal). Elena Lassuskaia: conceptualization (supporting), writing – review and editing (equal). Patrícia Homobono de Brito Moura: writing – review and editing (equal). Ivana Correa Ramos Leal: writing – review and editing (equal). Michelle Frazão Muzitano: conceptualization (lead), writing – review and editing (equal). Shaft Correa Pinto: conceptualization (lead), writing – review and editing (equal). All authors approved the final version of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: cbdv71068‐sup‐0001‐SuppMat.pdf.