Phytochemical Investigation and Biological Evaluation of Disynaphia filifolia: A Novel Source of Tremetone in Asteraceae

ABSTRACT

Chemical investigation of Disynaphia filifolia (Hassl.) R.M. King & H. Rob. (Asteraceae) led to the isolation of eight compounds, with tremetone being the major metabolite recovered. Tremetone has been suggested as the toxic compound in Ageratina altissima and Isocoma pluriflora, two species of the Asteraceae family. Although D. filifolia is not classified as a toxic species, tremetone was identified as the major compound in the crude extract of this species, constituting approximately 21% of the total content. Furthermore, 22 compounds were annotated using dereplication techniques, ultra‐high‐performance liquid chromatography coupled with high‐resolution tandem mass spectrometry, and a molecular networking approach. Antiproliferative assays indicated that tremetone was weakly active against the human tumor cell lines tested, proving to be even more toxic to non‐tumor cells compared to tumor cell lines. Moreover, the potential antioxidant and antibacterial activities of the crude extract and fractions of D. filifolia are reported herein.

1. Introduction

Continuing our investigation of Eupatorieae species in the Campos Gerais of Paraná, Brazil, an ecotone region between the Atlantic Forest and Cerrado biomes, we selected a species from the Disynaphia Hook. & Arn. ex DC. the genus for this study. This genus comprises 14 species [1], of which 12 occur in Brazil, including 8 endemic species [2].

To date, only two species of this genus, D. spathulata and D. multicrenulata, have been chemically investigated. From D. spathulata (formerly known as D. halimifolia), sesquiterpene lactones of the glaucolide, elemadienolide, and germacranolide types have been isolated, along with diterpenes and benzofuran derivatives [3].

A chemical study on D. multicrenulata from Paraguay reported the isolation of germacradienolides, heliangolides, ent‐kauranoic acids, and sesquiterpenes [4]. Flowers and leaves of this species from Argentina yielded germacradienolides, heliangolides, melampolides, a parthenolide derivative, cronquistiolides, eudesmanolides, an elemadienolide, a grazielolide, an isoguaiagrazielolide, diepoxygermacran‐8,12‐olides, kauranoic acids, pinoresinol, jaceosidin and the sesquiterpene oplopanone [5].

Aiming to expand studies on Disynaphia, the present work investigated the chemical profile and biological activity of D. filifolia. The chemical analysis led to the isolation of eight compounds, with tremetone being the major metabolite recovered. Tremetone has been suggested as the toxic principle in white snakeroot (Ageratina altissima) [6, 7] and rayless goldenrod (Isocoma plurifolia) [7, 8], two Asteraceae species known to cause trembles, milk sickness, or alkali disease in livestock and humans. Tremetone has also been found in other Asteraceae species, such as Parastrephia lepidophylla [9] and Wyethia arizonica [10].

Moreover, 22 compounds were additionally annotated through ultra‐high‐performance liquid chromatography coupled to high‐resolution tandem mass spectrometry (UHPLC‐HRMS/MS)‐based dereplication, including flavonoids and other benzofurans. Natural and synthetic benzofuran derivatives have demonstrated therapeutic potential, and the benzofuran nucleus is present in various clinical drugs. This class of compounds has been associated with a range of biological activities, including antibacterial and anticancer properties [11, 12]. Furthermore, flavonoids are well known for their antioxidant activity [13]. Taking this into account, and to further explore the biological potential of D. filifolia, the antioxidant, antibacterial, and antiproliferative activities of its crude extract and fractions were evaluated.

2. Results and Discussion

2.1. Chemical Composition of D. filifolia Aerial Parts

The chemical investigation led to the isolation of eight compounds from the aerial parts of D. filifolia (Figure 1). These compounds were identified by comparing their experimental nuclear magnetic resonance (NMR) spectroscopic data with previously reported data in the literature. The ¹H NMR spectra of the hexane (DFH) (Figure S1) and dichloromethane (DFD) (Figure S2) fractions revealed the presence of a major compound, which was subsequently isolated and identified as the benzofuran tremetone (1) (Figure S3) [14]. As previously mentioned, tremetone, and possibly other benzofuran ketones, are naturally occurring toxic compounds found in certain plants of the Asteraceae family [6, 8–10]. Although other benzofurans have been described in Disynaphia species [3], this is the first report of tremetone in this genus.

FIGURE 1.

Chemical constituents isolated from the aerial parts of Disynaphia filifolia.

The other known compounds isolated from D. filifolia were fomannoxin acid (2) [15], 2,4’‐dihydroxyacetophenone (3) [16], quercetin (4) [17], p‐hydroxybenzoic acid (5) [18], 3,4‐dihydroxybenzoic acid (6) [19], 6‐methoxy‐kaempferol (7) [20], and shikimic acid (8) [21]. The NMR data for all isolated compounds are provided in the supplementary material.

Fomannoxin acid (2) is a metabolite previously described in the pathogenic basidiomycete Heterobasidion annosum (syn. Fomes annosus) [22, 23]. This acid is uncommon in plants and has only been previously isolated from Gentiana algida (Gentianaceae) [15]. This study marks the first identification of this compound within the Asteraceae family.

There are few reports of flavonoids in the Disynaphia genus. Compounds of this class have been identified by HPLC analysis in D. multicrenulata and isolated from D. littoralis, which the authors cited under its basionym Eupatorium littorale [24]. Except for quercetin (4), all the other metabolites are being reported for the first time in the Disynaphia genus.

Because tremetone was identified as the major compound in both the DFH (Figure S1) and DFD (Figure S2) fractions, this metabolite was quantified in the D. filifolia crude extract. Using the previously isolated tremetone (1) as a standard, a reversed‐phase HPLC method was developed for its quantification (Figure 2).

FIGURE 2.

Chromatograms obtained by injection of tremetone (1) standard (A) and the crude extract of Disynaphia filifolia (B). Analysis conditions: elution isocratic: water/acetonitrile (8:2, v/v); flow rate 0.8 mL/min at λ = 280 nm.

The calibration curve demonstrated excellent linearity (R² = 0.9996) within the concentration range of 5.0–50 µg/mL. The limit of quantitation and the limit of detection were determined to be 2.5 and 0.757 µg/mL, respectively. The results obtained from the peak areas were used to generate the calibration curve (y = 24885x − 67794). In the crude extract of D. filifolia at a concentration of 100 µg/mL, the tremetone content was found to be 21.36 ± 0.27 µg/mL, corresponding to approximately 21% of the crude extract.

Although we found a high concentration of tremetone in the D. filifolia crude extract, factors such as different growing season periods and collection regions may influence the levels of this compound. These variations should be further explored in future studies. This is the first report providing a measured concentration of tremetone in the D. filifolia crude extract, highlighting this species as a potential alternative source of the compound.

To obtain the annotation of non‐isolated minor compounds present in D. filifolia and contribute to the understanding of its phytochemical profile, the DFD and ethyl acetate (DFEA) fractions were analyzed by UHPLC‐MS/MS quadrupole time‐of‐flight (Q‐ToF) in both positive and negative ionization modes. The obtained fragmentation data were organized into molecular networks using the Global Natural Products Social Molecular Networking (GNPS) platform [25]. The molecular network for the DFD and DFEA fractions was generated from 523 precursor ions in positive ionization mode and 534 in negative ion mode, visualized as nodes (Figures 3 and 4). Based on cosine similarity, 28 molecular families were formed in positive ionization mode, while 38 were formed in negative ion mode. Spectra that did not group into molecular families were represented as single nodes at the bottom of the network. Through dereplication, a total of 22 known compounds were putatively identified, including 12 flavonoids (DF‐1MS to DF‐12MS), five benzofurans (DF‐13MS to DF‐17MS), and five other compounds (DF‐18MS to DF‐22MS). Seventeen of these compounds were annotated using libraries available on GNPS [25]. Additionally, manual dereplication was performed to annotate compounds without library matches based on fragmentation patterns and literature data (Table 1).

FIGURE 3.

Molecular networking of the dichloromethane and ethyl acetate fractions of D. filifolia analyzed by ultra‐high‐performance liquid chromatography coupled to high‐resolution tandem mass spectrometry (UHPLC‐HRMS/MS) in the negative ionization mode. Nodes represent detected compounds and are colored according to the respective fraction.

FIGURE 4.

Molecular networking of the dichloromethane and ethyl acetate fractions of D. filifolia analyzed by UHPLC‐HRMS in positive mode. Nodes represent detected compounds and are colored according to the respective fraction.

TABLE 1.

Data of the annotated compounds in the dichloromethane and ethyl acetate fractions of Disynaphia filifolia by ultra‐high‐performance liquid chromatography coupled to high‐resolution tandem mass spectrometry (UHPLC‐HRMS/MS) and molecular networking.

Compound	Molecular formula	Theoretical mass (m/z)	Precursor ion (m/z)	Mass error /ppm	tR/min	Main fragment ions	Putative identification
Flavonoids
(4) #	C15H10O7 [M—H]‐	301.0354	301.0339	5.0	5.43	273; 271; 178; 151; 121	Quercetin [33]§
DF‐1MS *	C20H18O11 [M—H]‐	433.0776	433.0743	7.6	4.37	300; 223; 195; 165; 151	Quercetin pentoside derivative [34]§
DF‐2MS *	C21H20O11 [M—H]‐	447.0933	447.0904	6.5	4.81	315; 300; 271; 243	Methoxy‐quercetin pentoside derivative [35]§
DF‐3MS *	C26H28O15 [M—H]‐	579.1355	579.1304	8.8	4.59	315; 300	Methoxy‐quercetin pentosyl‐pentoside derivative [36]§
DF‐4MS *	C21H20O12 [M—H]‐	463.0882	463.0850	6.9	4.20	301; 300; 271; 178; 151	Quercetin hexoside derivative [34]§
DF‐5MS *	C22H22O12 [M—H]‐	477.1038	477.0999	8.2	4.53	315; 300; 271; 179	Methoxy‐quercetin hexoside derivative [37]§
DF‐6MS *	C27H30O16 [M—H]‐	609.1461	609.1410	8.4	4.05	301; 300; 271; 178; 225; 151	Quercetin hexosyl‐deoxyhexoside derivative [33]§
(7) #	C16H12O7 [M—H]‐	315.0510	315.0497	4.1	6.06	300; 272; 243; 181; 165;	6‐methoxy kaempferol [38]
DF‐7MS *	C15H10O6 [M—H]‐	285.0405	285.0390	5.3	6.00	257; 229; 213; 185; 169; 151; 136; 93	Kaempferol [39]§
DF‐8MS *	C20H18O10 [M—H]‐	417.0827	417.0798	6.7	4.66	285; 284; 256; 151; 119	Kaempferol pentoside derivative [40]§
DF‐9MS *	C15H10O5 [M—H]‐	269.0455	269.0441	5.2	5.88	233; 225; 177; 151; 119; 93	Apigenin [39]§
DF‐10MS *	C15H12O5 [M—H]‐	271.0612	271.0601	4.0	5.94	177; 151; 119; 93	Flavanone [41]§
DF‐11MS *	C16H14O6 [M—H]‐	301.0718	301.0700	6.0	6.04	286; 284; 177; 151	Flavanone [41]§
DF‐12MS *	C16H12O6 [M—H]‐	299.0561	299.0547	4.7	5.98	284; 163; 151	Methoxy‐luteolin derivative [42]§
Benzofurans
(1) #	C13H14O2 [M + H]+	203.1067	203.1062	2.5	5.30	161; 145; 133; 119; 107; 91	Tremetone [43]
(2) #	C12H12O3 [M + H]+	205.0855	205.0859	1.9	4.81	187; 159; 146; 133; 119; 105; 91	Fomannoxin acid [15]
DF‐13MS *	C13H12O2 [M + H]+	201.0910	201.0907	1.4	4.79	186; 173; 145; 131; 119; 105; 91	Dehydrotremetone [43]
DF‐14MS *	C13H16O2 [M + H]+	205.1223	205.1219	1.9	4.98	149; 135; 121; 107; 93	Methyl‐2‐(1‐methylethenyl)‐5‐benzofuranmethanol **
DF‐15MS *	C13H14O3 [M + H]+	219.1016	219.1012	1.8	4.73	201; 186; 173; 155; 145; 131; 105; 91	Hydroxytremetone or toxol [44]
DF‐16MS *	C13H12O3 [M + H]+	217.0859	217.0856	1.4	5.55	201; 186; 173; 155; 145; 131; 105; 91	Euparin [45]
DF‐17MS *	C15H16O4 [M + H]+	261.1121	261.1116	1.9	5.17	219; 201; 173; 155; 131; 107; 91	Toxyl acetate derivative **
Other compounds
DF‐18MS *	C11H16O3 [M + H]+	197.1172	197.1167	2.5	4.09	179; 161; 133; 107; 91	Loliolide§
DF‐19MS *	C11H16O3 [M + H]+	249.1485	249.1480	2.0	4.56	231; 203; 177; 157; 145; 135; 119; 107; 91	Eupatolide or Grandulin **
DF‐20MS *	C15H24O2 [M + H]+	237.1849	237.1844	2.1	5.38	219; 203; 177; 159; 145; 133; 119; 105; 93	Curcumol§
DF‐21MS *	C15H22O2 [M + H]+	235.1693	235.1687	2.5	4.70	217; 201; 175; 159; 145; 133; 119; 105; 93	Curcumenol or isocurcumenol§
DF‐22MS *	C18H28O2 [M + H]+	293.2111	293.2104	2.4	4.36	275; 247; 201; 187; 145; 133; 119; 107; 91	12‐oxophytodienoic acid§

^# isolated compounds; ^*annotated compounds; t_R: retention time; ^§GNPS database;^** fragment data not found.

The molecular network for flavonoids obtained in negative ionization mode exhibited networks of both glycosylated (O‐glycosyl and O‐diglycosyl) and non‐glycosylated flavonoids (Figure 3). Fragmentation data of glycosylated flavonoids (Table 1) showed neutral losses of 132, 162, 264, and 308 Da, corresponding to the losses of sugar moieties such as pentoses, hexoses, a diglycoside containing two pentoses, and a diglycoside containing both hexose and deoxyhexose. The identification of the aglycones of O‐glycosyl flavonoids was based on the fragmentation of Y₀ ions, as described by Tsimogiannis et al. [26]. These aglycones were annotated as quercetin, kaempferol, luteolin, and apigenin. Molecular networking successfully grouped the detected aglycones based on their specific fragment ions. Analysis of MS/MS fragmentation spectra also indicated the presence of several methoxylated flavonoids, as evidenced by the losses of 15 Da.

It is important to note that methoxylated flavonoids have been frequently identified in our studies on Eupatorium s.l. species from the Campos Gerais region of Paraná State [27, 28, 29, 30, 31, 32]. Furthermore, as expected, glycosylated flavonoids were primarily identified in the DFEA fraction, while aglycone flavonoids were putatively identified in the DFD subfraction.

Some compounds belonging to the benzofuran class were tentatively identified based on the obtained MS² spectra from DFD. These spectra were grouped and analyzed as a molecular family in positive ionization mode (Figure 4). The node at m/z 219.1012 [M+H]⁺ was annotated as toxol or hydroxytremetone. The node at m/z 261.1116 differed from DF‐15MS by 42 Da, suggesting it is an acetylated derivative. The node at m/z 217.0856 differed by 2 Da, indicating a derivative with an additional unsaturation on the heterocyclic moiety. Other benzofurans were annotated but appeared as individual nodes, including the isolated benzofuran tremetone (1) (m/z 203.1062 [M+H]⁺) and fomannoxin acid (2) (m/z 205.0859 [M+H]⁺). Although benzofuran derivatives were isolated from D. spathulata in a previous study [3], this is the first identification of tremetone and other structurally related compounds.

Another notable molecular family grouped two metabolites annotated as curcumol (m/z 237.1844 [M+H]⁺) and curcumenol or isocurcumenol (m/z 235.1687 [M+H]⁺), indicating the presence of terpenoids in this species. These compounds are being reported for the first time in the Disynaphia genus.

2.2. Biological Screening

2.2.1. Antiproliferative Activity

Expressed as the sample concentration required to induce 50% cell growth inhibition (GI₅₀, µg/mL), the antiproliferative activity of the extract, fractions, and compounds 1 and 2 from D. filifolia aerial parts were evaluated following the NCI‐60 protocol. Using cell lines from different tissues, this experiment aims to assess a broad cytostatic profile, avoiding tissue‐directed selection [46]. According to the NCI's criteria for active extracts (log GI₅₀ ≤ 1.5), the crude extract (DFCE) exhibited a weak cytostatic effect against nearly all tumor cell lines evaluated. The DFH and DFD fractions showed a cytostatic effect similar to DFCE, while the DFEA, butanolic (DFBU), and hydromethanolic (DFHM) fractions were inactive (Table 2). By comparing the effective concentrations (GI₅₀) of each sample against non‐tumor (HaCaT) and tumor cell lines, the selectivity index (SI) provides an indication of whether the sample may affect normal proliferative tissues [47]. Among the active fractions, DFH demonstrated better selectivity against glioblastoma (U251, SI = 1.9), lung carcinoma (NCI‐H460, SI = 1.9), and prostate adenocarcinoma (PC‐3, SI = 1.9) cells. Additionally, the DFD fraction showed better selectivity against multidrug‐resistant ovarian adenocarcinoma cells (NCI‐ADR/Res, SI = 1.8) (Table 2).

TABLE 2.

Antiproliferative activity of the crude extract, fractions, and compounds tremetone (1) and fomannoxin acid (2) obtained from Disynaphia filifolia aerial parts.

Cell line	Parameter	Doxorubicin	DFCE	DFH	DFD	DFEA	DFBU	DFHM	Tremetone (1)	Fomannoxin acid (2)
U251	GI50	< 0.025	25 *	25 *	25 *	92.2 ± 18.8	>250	>250	50.1 ± 26.2	60.0 ± 57.4
	logGI50	n.c.	1.4 (W)	1.4 (W)	1.4 (W)	2.0 (I)	n.c.	n.c.	1.7 (I)	1.8 (I)
	SI	n.c.	1.0	1.9	1.5	1.7	n.c.	n.c.	0.5	2.9
MCF‐7	GI50	< 0.025	25 *	44.0 ± 4.9	25 *	231.32 ± 0.03	>250	>250	32.1 ± 3.4	80.3 ± 52.8
	logGI50	n.c.	1.4 (W)	1.6 (I)	1.4 (W)	2.4 (I)	n.c.	n.c.	1.5 (W)	1.9 (I)
	SI	n.c.	1.0	1.1	1.5	0.7	n.c.	n.c.	0.8	2.2
NCI‐ADR/Res	GI50	0.033	25 *	63.5 ± 28.8	21.6 ± 7.5	62.5 ± 30.1	>250	>250	94.2 ± 11.3	171 *
	logGI50	−1.6 (P)	1.4 (W)	1.8 (I)	1.3 (W)	1.8 (I)	n.c.	n.c.	2.0 (I)	2.2 (I)
	SI	n.c.	1.0	0.8	1.8	2.5	n.c.	n.c.	0.3	1.0
NCI‐H460	GI50	< 0.025	49.5 ± 35.4	25 *	25 *	103.0 ± 84.4	>250	>250	32 *	>250
	logGI50	n.c.	1.7 (I)	1.4 (W)	1.4 (W)	2.0 (I)	n.c.	n.c.	1.5 (W)	n.c.
	SI	n.c.	0.5	1.9	1.5	1.5	n.c.	n.c.	0.8	n.c.
PC‐3	GI50	0.103	25 *	25 *	25 *	65.5 ± 8.6	>250	>250	45.3 ± 7.6	>250
	logGI50	−1.6 (P)	1.4 (W)	1.4 (W)	1.4 (W)	1.8 (I)	n.c.	n.c.	1.7 (I)	n.c.
	SI	n.c.	1.0	1.9	1.5	2.4	n.c.	n.c.	0.5	n.c.
HT‐29	GI50	0.097	17*	50.2 ± 17.7	25*	145.0 ± 123.1	>250	>250	28*	>250
	logGI50	−1.6 (P)	1.2 (W)	1.7 (I)	1.4 (W)	2.2 (I)	n.c.	n.c.	1.4 (W)	n.c.
	SI	n.c.	1.5	1.0	1.5	1.1	n.c.	n.c.	0.9	n.c.
HaCaT	GI50	< 0.025	25.4 ± 20.0	48.6 ± 23.4	38.1 ± 6.4	155.1 ± 33.6	>250	>250	24.3 ± 18.5	174.0 ± 67.2
	logGI50	n.c.	1.4 (W)	1.7 (I)	1.6 (I)	2.2 (I)	n.c.	n.c.	1.4 (W)	2.2 (I)

Parameters: a) GI₅₀: concentration (µg/mL) required for 50% of cell growth inhibition followed by standard error calculated by sigmoidal regression using Origin 8.0 software; b) log GI₅₀: results expressed as logarithm and classified according to NCI's criteria (W, weak activity: 1.1 ≤ log GI₅₀ < 1.5; M, moderate activity: 0 ≤ log GI₅₀ < 1.1; P, potent activity: log GI₅₀ <0); (c) SI: selectivity index calculated as GI₅₀ HaCaT/ GI₅₀ Tumor cell line.

^*approximated value (experimental data did not converge); n.c.: not calculated.

DFCE = crude extract; DFH = hexane fraction; DFD = dichloromethane fraction; DFEA = ethyl acetate fraction; DFBU = butanol fraction; DFHM = hydromethanolic fraction.

Human tumor cell lines: U251 (glioblastoma); MCF‐7 (breast, adenocarcinoma); NCI‐ADR/RES (ovary, multi‐drug resistant adenocarcinoma); NCI‐H460 (lung, large cell carcinoma PC‐3 (prostate, adenocarcinoma), HT‐29 (colon, adenocarcinoma). Human non‐tumor cell lines: HaCaT (immortalized keratinocyte).

Similar to fomannoxin [48, 49], fomannoxin acid (2) was inactive against both human tumor and non‐tumor cell lines. Isolated from both DFH and DFD, tremetone (1) exhibited a weak cytostatic effect (log GI₅₀ = 1.4 to 1.5) with a low SI (0.8–0.9) against MCF‐7 (breast adenocarcinoma), NCI‐H460 (lung non‐small cell carcinoma), and HT‐29 (colon adenocarcinoma) cell lines (Table 2). Tremetone has previously been reported to affect cell viability only at high concentrations, with an IC₅₀ of 490 µM in human neuroblastoma SH‐SY5Y cells [50] (Table 2 and Figure S4).

As mentioned earlier, tremetone has been suggested as the toxic principle in certain Asteraceae species. Regarding pharmacokinetic parameters, in silico prediction using SwissADME indicated good oral absorption and the ability to cross the blood‐brain barrier (Figure S5). These findings support reports of toxicity following the oral consumption of tremetone‐containing plants.

Although the exact mechanism of action remains incompletely understood, tremetone is believed to interfere with mitochondrial function, particularly by disrupting oxidative phosphorylation, leading to a failure in energy metabolism. This mechanism underlies symptoms such as muscle weakness, tremors, and metabolic acidosis observed in tremetone poisoning. Because mitochondrial dysfunction affects multiple systems, many of its toxic effects are interconnected [51, 52, 53, 54]. Furthermore, using SwissTargetPrediction, tremetone was predicted to affect key kinases such as MAPKs and CDK/cyclins (Figure S6 and Table S2). Because alterations in mitochondrial metabolism play a significant role in the induction of certain regulated cell death mechanisms [55], further studies should investigate whether the observed cytostatic effect of tremetone is linked to its action on mitochondria and kinase interactions.

2.2.2. Antioxidant Assays

The antioxidant activity of the crude extract and fractions from D. filifolia was evaluated using the 1,1‐diphenyl‐2‐picrylhydrazyl (DPPH) and ferric‐reducing ability power (FRAP) assays (Table 3). Among the tested fractions, DFEA and DFBU exhibited the highest activity. The DFEA fraction showed the strongest activity in the DPPH radical‐scavenging assay, while the DFBU fraction was the most active in the FRAP assay.

TABLE 3.

Antioxidant activity by 1,1‐diphenyl‐2‐picrylhydrazyl (DPPH) and ferric‐reducing ability power (FRAP) assays for the crude extract and fractions of the aerial parts of D. filifolia.

Extract and fractions	DPPH (µmol TE g−1 ± SD)	FRAP (µmol Fe II g−1 ± SD)
DFCE	576.54 ± 2.82	245.84 ± 5.73
DFH	353.88 ± 8.38	87.37 ± 2.50
DFD	535.55 ± 0.96	178.11 ± 0.55
DFEA	991.66 ± 3.67	680.74 ± 6.41
DFBU	710.10 ± 3.15	1465.18 ± 12.9
DFHM	449.44 ± 1.92	70.15 ± 0.64

Data are expressed as the mean ± standard deviation (SD); DFCE (crude extract); DFH (Hexane fraction); DFD (dichloromethane fraction); DFEA (Ethyl acetate fraction); DFBU: (Butanolic fraction); DFHM (Hydromethanolic fraction); TE: (Trolox Equivalent); DPPH: (1,1‐diphenyl‐2‐picrylhydrazyl); FRAP: (Ferric‐Reducing Antioxidant Power).

The antioxidant activity exhibited by the DFEA fraction in the DPPH assay can be attributed to the presence of various phenolic compounds, such as 2,4’‐dihydroxyacetophenone (3), quercetin (4), p‐hydroxybenzoic acid (5), 3,4‐dihydroxybenzoic acid (6), and 6‐methoxy‐kaempferol (7). These compounds, which were isolated from this fraction, are frequently reported in the literature as potential antioxidants [56, 57, 58, 59]. Additionally, other flavonoids, including kaempferol, luteolin, and apigenin, were putatively identified in this fraction by UHPLC‐HRMS/MS in negative ionization mode, and they may also contribute to its antioxidant activity [13, 60]. The antioxidants present in the DFBU fraction demonstrated strong reducing power, as indicated by its higher FRAP value.

A previous study demonstrated that tremetone does not exhibit antioxidant activity [14]. As expected, the DFH and DFD fractions, which are rich in tremetone, showed lower antioxidant activity in both the DPPH and FRAP assays than in the DFEA and DFBU fractions.

2.3. Antibacterial Activity

The crude extract and fractions of the aerial parts of D. filifolia were screened against nine bacterial species: Staphylococcus aureus, Bacillus subtilis, and B. cereus (Gram‐positive), as well as Salmonella enterica, Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, Shigella flexnerii, and Shigella dysenteriae (Gram‐negative). The results were expressed as the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) (Table S1).

The crude extract inhibited P. aeruginosa and B. subtilis (MIC = 5.0 mg/mL) and was bactericidal against B. subtilis at the same concentration. The DFH fraction was active against S. dysenteriae, with MIC and MBC values of 2.5 and 5.0 mg/mL, respectively, and also inhibited P. aeruginosa (MIC = 5.0 mg/mL). The DFD and DFEA fractions showed inhibition against S. dysenteriae and S. aureus (MIC = 5.0 mg/mL), as well as P. aeruginosa for DFD and S. aureus for DFEA. In these last two fractions, flavonoid compounds annotated as quercetin, kaempferol, luteolin, and apigenin are known to possess antibacterial activity [61].

Flavonoids and tremetone in the crude extract and fractions may explain their antimicrobial activity. Although glycosylated flavonoids were detected and are known to have lower antimicrobial activity than their corresponding aglycones, [62] their presence in higher quantities or potential interactions may still contribute to the observed effects. While the DFBU fraction exhibited the highest FRAP value, this fraction and the DFHM fraction were not active against bacteria at the highest concentration tested. The positive control (ampicillin) showed inhibition against all microorganisms, with MIC and MBC values ranging from 0.016 to 0.250 and 0.063 to 0.5 mg/mL, respectively.

3. Conclusions

In this study, we explored for the first time the chemical and biological potential of the aerial parts of D. filifolia, a species commonly found in the grasslands of the Campos Gerais region, Paraná State, southern Brazil. Twenty‐nine known compounds were identified, including methoxylated flavonoids, which have been frequently detected in our studies on Eupatorium s.l. species from this region.

Tremetone was identified as the major compound in the extract of D. filifolia, accounting for approximately 21% of the crude extract. As mentioned earlier, tremetone has been suggested as a toxin in A. altissima and Isocoma pluriflora, and its isolation from the aerial parts of a species that naturally occurs in grasslands highlights the significance of this study.

In summary, our results revealed the chemical profile of D. filifolia and identified it as a potential source of antioxidant and antibacterial compounds. Because tremetone was detected in this species, further studies are necessary to determine whether D. filifolia should also be classified as a toxic plant.

4. Experimental

4.1. General Experimental Procedures

Thin‐layer chromatography was carried out on normal‐phase pre‐coated silica gel 60G or 60GF₂₅₄ (Merck) plates. Visualization of the compounds was achieved using ultraviolet (UV) irradiation at 254 and 366 nm and/or by spraying with an H₂SO₄/anisaldehyde/acetic acid/methanol (5:0.5:10:85 mL, v/v) solution, followed by heating at 100°C. Column chromatography (CC) separations were performed using silica gel 60 (70–230 mesh; Merck), silica gel flash (230–400 mesh; Merck), or Sephadex LH‐20 (Sigma‐Aldrich). HPLC analyses were conducted on a Shimadzu Prominence instrument equipped with two LC‐20AR pumps, a DGU‐20ASR degasser, an SPD‐M20A diode array detector, and an automatic SIL‐10AF injection system, using a Shim‐pack PREP‐ODS column (250 × 20 mm). The mobile phase consisted of water (Milli‐Q; Millipore) and methanol (Merck, Chemicals Co). NMR spectra were recorded on a Bruker spectrometer operating at 300 MHz and 75.5 MHz, as well as on a Bruker Avance III HD spectrometer operating at 500 and 125 MHz, using CDCl₃ and dimethyl sulfoxide (DMSO)‐d6 as solvents (Sigma‐Aldrich).

4.2. Plant Material

The aerial parts of D. filifolia were collected at Campos Gerais National Park, Ponta Grossa, Paraná State, Brazil, approximately (25°08′46″ S, 049°057′025″ W) in March 2016 and identified by Dr Marta Regina Barrotto do Carmo. A voucher specimen was deposited in the herbarium of Universidade Estadual de Ponta Grossa (HUPG 22451; Registration code at SISGEN A610E3A).

4.3. Extraction and Isolation of Compounds

The aerial parts of D. filifolia (2,000 g) were dried at 40°C for 72 h, powdered using a knife mill, and exhaustively extracted with ethanol (Synth, 99.5%) at room temperature. After extraction, the solvent was removed under vacuum, yielding 123 g of DFCE. A portion of the crude extract (103.4 g) was suspended in 400 mL of a MeOH:H₂O (1:1) solution and successively partitioned with n‐hexane (Synth, 100%), CH₂Cl₂ (Synth, 100%), EtOAc (Synth, 100%), and BuOH (Synth, 99.8%). The solvents were removed under reduced pressure, resulting in the DFH (50.0 g), DFD (13.3 g), DFEA (10.2 g), DFBU (9.8 g), and DFHM (7.7 g) fractions.

A portion of the DFH fraction (20.0 g) was subjected to vacuum liquid chromatography (Φ = 56 cm × h = 20 cm) over silica gel 60, using hexane/EtOAc mixtures in increasing proportions as eluents, yielding 11 subfractions. Subfraction DFH‐3 (1.13 g) was further purified by CC (Φ = 2.5 cm × h = 28 cm) on silica gel flash, using an n‐hexane/EtOAc gradient as the eluent, resulting in subfractions DFH‐3‐1 to DFH‐3‐11. From subfraction DFH‐3‐2, compound 1 (228 mg) was isolated. Compound 2 (10.9 mg) was obtained from subfraction DFH‐4 (0.63 g) after CC on silica gel flash (Φ = 2.3 cm × h = 26.0 cm), using hexane, EtOAc, and MeOH in increasing polarity as eluents.

The DFD fraction (8.7 g) was subjected to CC (Φ = 3.3 cm × h = 33 cm) on silica gel using an n‐hexane/EtOAc gradient as the eluent, yielding subfractions DFD‐1 to DFD‐21. Subfraction DFD‐7 afforded a mixture of compounds 1 and 2 (476 mg). Subfraction DFD‐16 (690 mg) was further purified by silica gel CC (Φ = 2.9 cm × h = 23.4 cm), using a gradient system of n‐hexane/EtOAc/MeOH, resulting in 11 subfractions (DFD‐16‐1 to DFD‐16‐11). Subfraction DFD‐16‐1 led to the reisolation of compound 2 (12.8 mg).

A portion of the DFEA fraction (3.0 g) was subjected to Sephadex LH‐20 CC (Φ = 2.3 cm × h = 24.5 cm) and eluted with MeOH/H₂O in ratios of decreasing polarity, yielding 12 subfractions (DFEA‐1 to DFEA‐12). Subfractions DFEA‐3 and DFEA‐11 afforded compounds 3 (13.6 mg) and 4 (5.2 mg), respectively. The DFEA‐7 fraction (70.5 mg) was further purified by Sephadex LH‐20 CC filtration (Φ = 1.0 cm × h = 25 cm), eluted with MeOH/H₂O in a decreasing polarity gradient, resulting in subfractions DFEA‐7‐1 to DFEA‐7‐6, from which DFEA‐7‐6 yielded a mixture of compounds 5 and 6 (12 mg). Subfraction DFEA‐9 was also subjected to Sephadex LH‐20 CC (Φ = 1.0 cm × h = 29 cm), eluted with MeOH/H₂O in a decreasing polarity gradient, generating 10 subfractions (DFEA‐9‐1 to DFEA‐9‐10), with subfraction DFEA‐9‐8 affording a mixture of compounds 4 and 7 (11.5 mg).

A portion of the DFBU fraction (2.13 g) was subjected to Sephadex LH‐20 CC filtration (Φ = 2.3 cm × h = 21 cm), eluted with MeOH/H₂O in ratios of decreasing polarity, yielding eight subfractions (DFBU‐1 to DFBU‐8). Subfraction DFBU‐2 (57.2 mg) was further purified by Sephadex LH‐20 CC (Φ = 1.0 cm × h = 25 cm), eluted with MeOH, resulting in four subfractions (DFBU‐1 to DFBU‐4), from which DFEA‐4 afforded compound 8 (19.6 mg).

4.4. HPLC‐Diode Array Detection Analysis

The crude extract, fractions, and tremetone from D. filifolia were analyzed using HPLC with a diode array detector (Shimadzu, Prominence). Chromatographic separation was performed on a Supelcosil LC‐18 column (25 cm × 4.6 mm; 5 µm) (Merck, Germany). Various elution methods using methanol or acetonitrile in water were tested to optimize the separation of the constituents of interest. The optimal condition was determined to be a mobile phase of acetonitrile/water (2:8) with a flow rate of 0.8 mL min⁻¹ in isocratic mode, with detection at a wavelength (λ) of 280 nm.

Tremetone (1) was quantified in the crude extract of D. filifolia using a seven‐point calibration curve, prepared with the previously isolated compound from the DFH fraction of D. filifolia. The calibration curve was generated in acetonitrile over a concentration range of 10.0–50.0 µg mL⁻¹ by serial dilution (n = 4). The limits of detection and quantification for compound 1 were determined by diluting the lowest standard (10.0 µg/mL) until the measured peak heights were 3 × (3:1) and 10 × (10:1) the height of the baseline noise, respectively.

4.5. Dereplication by UHPLC‐HRMS/MS Analysis and Molecular Networking

4.5.1. UHPLC‐HRMS/MS Analysis

The DFD and DFEA fractions of D. filifolia were analyzed using the UHPLC‐HRMS/MS method as described by Ramos et al. [11]. UHPLC analysis was performed on a Shimadzu Nexera X2 instrument equipped with a CBM‐20A system controller, two LC‐30AD pumps, a CTO‐30A column oven, and a SIL‐30AC autosampler. Mass spectra were recorded on a Bruker IMPACT II mass spectrometer, using an electrospray ionization source in both positive and negative ionization modes (separately), a Q‐ToF analyzer, and a multichannel plate detector.

The samples were resuspended in methanol/acetonitrile (1:1, 2.0 mg/mL, v/v), and chromatographic separations were performed using UHPLC on a Symmetry C18 column (75 × 2.0 mm i.d.; 1.6 µm, Shim‐pack XR‐ODS III). The mobile phase consisted of H₂O and CH₃CN with 0.1% formic acid (solvent B) for positive ionization mode and H₂O and CH₃CN for negative ionization mode. The gradient program was as follows: initial 0–1 min, elution A–B (95:5, v/v); 1–3 min (30:70, v/v); 3–12 min (5:95, v/v); maintained at 95% B from 12–16 min; and 5% B from 16 to 17 min. The flow rate was set at 0.2 mL/min, with an injection volume of 3 µL and a column oven temperature of 40°C. High‐resolution mass spectrometry analysis was performed using a Q‐ToF mass spectrometer via an electrospray ionization interface [63]. Data processing was carried out using Data Analysis 4.3 (Bruker), and mass error values were calculated, with molecular formulas ≤10 ppm of error considered in this study.

4.5.2. Classical Molecular Networking Analysis

A molecular network was created using the online workflow on the GNPS website (http://gnps.ucsd.edu). Initially, both fractions and blank analyses (acetonitrile/water, 1:1 v/v) were converted to mzXML format using DataAnalysis 4.3 software. These files were used to generate molecular networks and perform library searches on the GNPS12 platform, with positive and negative ionization mode data considered separately. The data were filtered by removing all MS/MS fragment ions within ±17 Da of the precursor m/z. MS/MS spectra were window‐filtered by selecting only the top six fragment ions in the ±50‐Da window throughout the spectrum. The precursor ion and MS/MS mass tolerance were set to 0.02 Da. A network was then constructed, where edges were filtered to retain only those with a cosine score of >0.7 and more than four matched peaks. Additionally, edges between two nodes were included in the network only if each node appeared in the other's top 10 most similar nodes. The maximum size of a molecular family was set to 100, and the lowest‐scoring edges were removed until the molecular family size met this threshold. The clustered spectra were also searched against the GNPS public spectral libraries, applying a cosine threshold of 0.7 and requiring at least four matched fragment ions. Network visualization was performed using Cytoscape software version 3.8.1 (Cytoscape Consortium, San Diego, CA, USA) [64]. Nodes corresponding to blank analyses were removed from the molecular network, and node colors were assigned based on the MS/MS data source files, with a different color for each fraction.

4.6. Biological Screening

4.6.1. Antiproliferative Activity

In vitro antiproliferative activity experiments were conducted according to Monks et al. [46]. The crude extract, fractions, and benzofurans 1 and 2 of D. filifolia were evaluated in vitro against 10 human tumor cell lines: U251 (glioma, CNS), UACC‐22 (melanoma), MCF‐7 (breast), NCI‐ADR/RES (ovarian, expressing the multiple drug resistance phenotype), 786‐0 (renal), NCI‐H460 (lung, non‐small cell), PC‐3 (prostate), OVCAR‐3 (ovarian), HT‐29 (colon), and K‐562 (leukemia), kindly provided by the National Cancer Institute (Frederick, MA, USA). All samples were also tested against the immortalized human keratinocytes (HaCat) cell line (provided by Prof. Dr. Ricardo Della Coletta, UNICAMP). Stock solutions (0.1 g/mL) of each sample were prepared in DMSO and successively diluted in RPMI 1640, supplemented with 5% fetal bovine serum and 1% penicillin:streptomycin mixture (1000 IU/mL:1000 µg/mL) to final concentrations of 0.25, 2.5, 25, and 250 µg/mL for the extract and fractions, and 0.15, 1.50, 15.0, and 150.0 µg/mL for the isolated compounds. The chemotherapeutic doxorubicin hydrochloride (0.025, 0.25, 2.5, and 25.0 µg/mL) was used as a positive reference standard to determine cell line sensitivity. Cells in 96‐well plates (100 µL cells/well) were exposed to sample concentrations, in triplicate, for 48 h at 37°C with 5% CO₂. The final DMSO concentration (≤0.25%) did not affect cell viability. Prior to (plate control) and after sample addition, cells were fixed with 50% trichloroacetic acid, and cell proliferation was assessed by spectrophotometric quantification of cellular protein content at 540 nm (VersaMax model; Molecular Devices) using sulforhodamine B. Two effective concentrations, GI₅₀ and total growth inhibition, were calculated by non‐linear regression analysis (sigmoidal fit) using Origin 7.5 (OriginLab Corporation). The SI was calculated as SI = GI₅₀ HaCat/GI₅₀ tumor cell line [65].

4.6.2. Antibacterial Activity

Antibacterial assays were performed using the microdilution method in sterile 96‐well microplates, according to the Clinical and Laboratory Standards Institute [66, 67]. The antibacterial potential was tested against Gram‐positive strains S. aureus (ATCC 25923), B. subtilis (ATCC 6633), B. cereus (INCQS 00003), and Gram‐negative bacteria S. enterica subsp. enterica serovar Enteritidis (ATCC 13076), E. coli (ATCC 25922), P. aeruginosa (ATCC 15442), K. pneumoniae (ATCC 700603), and S. flexnerii (ATCC 12022). Serial dilutions of each extract or fraction (10 mg/mL) were prepared on a microdilution plate containing 100 µL of Mueller–Hinton broth. The inoculum was then added to each well, resulting in a final density of approximately 5.0 × 10⁵ CFU/mL. The microplates were incubated at 35 ± 2°C for 24 h. Ampicillin (1.0 mg mL⁻¹) was used as a positive control. The MIC was defined as the lowest concentration that resulted in the inhibition of visual growth. The MBC was determined as the lowest concentration yielding negative subcultures in Mueller–Hinton agar from each well, showing no growth. These tests were performed in triplicate.

4.7. Antioxidants Assays

4.7.1. DPPH Method

The free radical scavenging effect of the crude extract and fractions of D. filifolia was investigated using the DPPH (Sigma‐Aldrich) assay [68]. Initially, a calibration curve with the Trolox 2000 µmol/L standard was prepared under the shelter of light, immediately before the analysis. A solution of DPPH (9.13 × 10⁻² mmol/L) was prepared in an amber flask wrapped in aluminum foil to prevent degradation by light. Concentration dilutions were made from the Trolox standard solution, ranging from 100 to 1500 µmol/L. The samples were prepared by dissolving 1.0 mg of the crude extract and fractions of D. filifolia in 1 mL of ethanol. Then, 50 µL of the samples and 3 mL of the DPPH solution were pipetted into a 15‐mL Falcon tube protected from light. The mixture was thoroughly vortex‐mixed and kept in the dark for 30 min. Absorbance was measured using a UV‐visible spectrophotometer at 517 nm. All samples were tested in triplicate. The results were expressed as µmol of Trolox equivalent (TE) per gram of extract (µmol TE g⁻¹).

4.7.2. FRAP Method

Analysis of reducing power using the FRAP assay was performed in triplicate according to the method described by Benzie and Strain [69], with modifications from Boroski et al. [68]. The FRAP reagent consisted of a mixture of 300 mmol/L sodium acetate buffer solution (pH 3.6), 10 mmol/L TPTZ solution, and 20 mmol/L FeCl₃·6H₂O solution in a volume ratio of 10:1:1. Aliquots (100 µL) of each extract and fraction (2.0 mg/mL) were homogenized with 3 mL of FRAP solution and 300 µL of distilled water. After incubation for 40 min at 36°C, the absorbance was measured at 593 nm, with water as a blank. The antioxidant potential of the extracts to reduce Fe(III) to Fe(II) was expressed as µmol Fe(II) g⁻¹ using the calibration curve obtained with FeSO₄·7H₂O (0–2000 µmol/L).

Author Contributions

Debora Cristina Baldoqui and Maria Helena Sarragiotto conceived the present idea, developed the method, and supervised the findings of this work. Yasmin Rodrigues Chierici and Drielli Rhiane Peres Colhado Areas performed the phytochemical studies, identified the isolated compounds, and analyzed the raw data obtained from UHPLC‐HRMS/MS with the molecular networking strategy. Anderson Valdiney Gomes Ramos and Helena Mannochio‐Russo performed the analysis of the dichloromethane fraction by UHPLC‐HRMS/MS and contributed to the molecular network. Anderson Valdiney Gomes Ramos supervised the semi‐preparative HPLC analysis and contributed to the isolation and identification of the compounds. Biological assays were conducted and analyzed by Alana Carla Battistella, Solange Maria Cottica, Tatiana Shioji Tiuman, Ana Lucia Tasca Gois Ruiz, and Mary Ann Foglio. Marta Regina Barrotto do Carmo do Carmo collected and identified the aerial parts of the species. The manuscript was prepared by Yasmin Rodrigues Chierici, Anderson Valdiney Gomes Ramos, and Debora Cristina Baldoqui All authors discussed the results and contributed to the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting information for this article is available on the WWW under https://doi.org/10.1002/MS‐number.

Chierici Y. R., Ramos A. V. G., Areas D. R. P. C., et al. “Phytochemical Investigation and Biological Evaluation of Disynaphia filifolia: A Novel Source of Tremetone in Asteraceae.” Chemistry & Biodiversity 22, no. 9 (2025): 22, e202402696. 10.1002/cbdv.202402696

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.