Bio‐Based Potential of Phenolic‐Rich Fabaceae Seeds (Gonocytisus angulatus, Podocytisus caramanicus): Chemical Standardization, Functional Screening, and Molecular Insights

ABSTRACT

This study comparatively evaluated the phytochemical composition and bioactivities of methanolic extracts and aqueous decoctions obtained from seeds of two Fabaceae taxa (Gonocytisus angulatus, Podocytisus caramanicus). Aqueous decoction yielded higher extraction efficiencies (up to 39.20%), whereas methanolic extraction recovered a broader range of phenolic compounds. Phytochemical analysis revealed that flavonoids and phenolic acids were the dominant constituents, with chlorogenic acid, rutin, quercetin, kaempferol, and gallic acid as the major compounds, particularly enriched in methanolic extracts. In vitro assays demonstrated moderate enzyme inhibitory effects. The highest α‐glucosidase inhibition was observed in the methanolic extract (20.22%), while α‐amylase inhibition reached 30.11%. Anticholinesterase activity was also weak, with maximum acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibition values of 15.27% and 19.93%, respectively. In contrast, antioxidant activity was more pronounced, with ABTS radical scavenging reaching 97.40%, comparable to the reference standard. Molecular docking supported these findings by indicating favorable binding interactions of major phenolic compounds with target enzymes.

This study provides the first comparative phytochemical and functional evaluation of Gonocytisus angulatus and Podocytisus caramanicus seeds. Methanolic extracts showed richer phenolic profiles, while aqueous decoctions provided higher extraction yields. P. caramanicus exhibited strong ABTS radical scavenging activity. Enzyme inhibition, multivariate analysis, and molecular docking further supported the functional potential of these phenolic‐rich seed extracts.

1. Introduction

Alzheimer's disease (AD) is a major global health challenge and ranks among the leading causes of death in the elderly. AD is a progressive neurodegenerative disorder and the most common cause of dementia worldwide. It is marked by amyloid‐β plaques, tau tangles, and the degeneration of cholinergic neurons, leading to reduced acetylcholine (ACh) levels and cognitive decline. Based on the cholinergic hypothesis, acetylcholinesterase (AChE) inhibitors such as donepezil, galantamine, and rivastigmine are used to manage symptoms, though they do not stop disease progression. Recent studies suggest some AChE inhibitors may also have neuroprotective effects beyond enzyme inhibition, highlighting the need for new treatments targeting both cholinergic and non‐cholinergic pathways [1, 2, 3].

Diabetes mellitus (DM) is a growing chronic metabolic disorder and a significant global health issue. Global DM prevalence was around 2.8% in 2000 and is projected to rise to approximately 4.4% by 2030. Modern dietary and lifestyle changes have increased the prevalence of type 2 diabetes mellitus (T2DM), now affecting a broader age range. Inhibiting key enzymes like α‐amylase and α‐glucosidase, which break down starch into glucose, is an effective way to manage postprandial hyperglycemia. Although synthetic inhibitors are used clinically, they often cause side effects and may lose effectiveness. Natural inhibitors from foods such as certain glycosides, peptides, and lipids offer safer alternatives and hold promise for use in functional foods as part of dietary management in T2DM [4, 5, 6].

T2D and AD share common risk factors, most notably aging, which is marked by oxidative stress, mitochondrial dysfunction, and chronic inflammation. Insulin resistance further links these conditions, with mitochondrial impairment and oxidative stress appearing early, even before clinical onset [7]. AD and DM share common genetic, metabolic, and pathological mechanisms, suggesting that antidiabetic therapies may also offer therapeutic potential in AD [8].

The genus name Gonocytisus is derived from the combination of “‐gono” and “cytisus.” The prefix “‐gono” (or “‐gone”) refers to reproduction, reproductive organs, or buds, while it can also mean “winged” or “prominently winged” when taken from “‐gonio.” Cytisus is of Greek origin and refers directly to the plant itself, meaning “broom,” a typical feature of the Fabaceae family. Together, the name Gonocytisus can be interpreted as “winged broom” [9].

Gonocytisus angulatus Spach (Fabaceae) is an erect, slender, upright shrub native to the Eastern Mediterranean [10]. G. angulatus, locally known as “Kara borcak” in Türkiye, is a leafy‐branched shrub commonly used as fodder [11]. Six flavonoids were isolated from the aerial parts of G. angulatus and identified as luteolin, luteolin 7‐O‐rutinoside, vitexin, isovitexin, rhamnosyl vitexin, and orientin using TLC, melting point, and spectroscopic (UV, IR, and ¹H NMR) analyses [10]. The in vitro antimicrobial activity of ethanolic extracts and crude alkaloid mixtures from the roots, twigs, leaves, flowers, and seeds of G. angulatus collected in Türkiye was evaluated using the microdilution method. The samples showed inhibitory effects against all tested bacteria (E. coli, P. aeruginosa, S. faecalis, and S. aureus) and yeast‐like fungi (C. albicans, C. parapsilosis, C. pseudotropicalis, and C. stellatoidea) [12].

The monotypic genus Podocytisus Boiss. and Heldr. (Fabaceae), represented by P. caramanicus, displays a striking disjunct distribution between southern Anatolia and the Balkans. It differs from other Genisteae members by its flattened, tardily dehiscent, dorsally winged legumes [13]. P. caramanicus Boiss. and Heldr. is a rare species occurring sporadically in inland regions of northern and central Greece, typically between 400–1300 m, with its highest recorded elevations (up to 1600 m) on Mt. Smolikas in Epirus [14]. In a screening of Greek Fabaceae, the methanolic extract of P. caramanicus exhibited proliferative effects on MCF‐7 human breast cancer cells. Initial validation of the cell model was performed using known estrogenic compounds (E2, genistein, and biochanin A). The extract's major flavone, 7‐O‐β‐D‐glucopyranosylchrysin (Glc‐chr), along with its aglycone chrysin, significantly enhanced ERα‐mediated transcriptional activity. This estrogenic response was fully inhibited by the antagonist ICI 182,780, indicating ERα‐dependent action [15].

To the best of our knowledge, this study represents the first comprehensive evaluation of the phytochemical profile and multitarget bioactivities of these two Fabaceae taxa, particularly focusing on their enzyme inhibitory potential and molecular interactions. These species were selected due to their limited phytochemical and biological investigation in the literature, despite belonging to a family known for its rich phenolic content and functional bioactive compounds. The aim of this study was to evaluate the inhibitory effects of aqueous and methanolic seed extracts of G. angulatus and P. caramanicus on AChe, butyrylcholinesterase (BChE), α‐amylase, and α‐glucosidase enzymes, and to characterize their chemical profiles through FTIR analysis, together with assessment of their antioxidant activities using DPPH and ABTS assays.

2. Results and Discussion

2.1. Extraction

Figure 1 summarizes the extraction yields and obtained amounts from the seeds of G. angulatus and P. caramanicus using methanolic extraction and aqueous decoction. For both species, aqueous decoction yielded higher extraction yields than methanolic extraction, indicating greater recovery of water‐soluble constituents. In G. angulatus, the decoction extract showed a markedly higher yield (39.20%) than the methanolic extract (27.00%), suggesting that polar compounds are predominant in the seed matrix. A similar trend was observed for P. caramanicus, with decoction yielding 33.60% and the methanolic extract 26.47%.

FIGURE 1.

Extraction yields and amounts of G. angulatus and P. caramanicus.

The extraction yields obtained from G. angulatus and P. caramanicus clearly demonstrate the influence of solvent polarity on extraction efficiency. In both species, aqueous decoction resulted in higher yields than methanolic extraction, indicating a predominance of polar and water‐soluble constituents in the seed matrix. This observation is consistent with previous studies reporting that water‐based extractions are particularly effective in recovering hydrophilic phytochemicals such as phenolic acids, sugars, glycosides, and certain flavonoids, especially when fruits or seeds are used as plant material [16]. Methanol, although widely used for analytical purposes due to its ability to extract a broad range of compounds, often yields lower total extractable material compared to water, particularly in matrices rich in highly polar compounds [17]. The higher yields observed in decoction extracts may also be attributed to thermal effects, as elevated temperatures can enhance cell wall disruption, improve solvent penetration, and facilitate the release of intracellular metabolites [18].

From an applied perspective, the higher yields obtained via decoction are noteworthy, as water‐based extraction methods are generally regarded as safer, more economical, and more compatible with food and traditional preparation practices. Recent reviews emphasize that extraction yield, while not a direct indicator of biological potency, is an important preliminary parameter when evaluating plant materials for food‐related or nutraceutical applications [19].

2.2. Qualitative Screening of Primary and Secondary Metabolites

Figure 2 presents the qualitative phytochemical profile of the methanolic and aqueous (decoction) seed extracts of G. angulatus and P. caramanicus. Overall, both species showed a highly consistent pattern, with flavonoids, tannins, saponosides, coumarins, starch, and carbohydrates detected in all extracts, suggesting that these taxa are rich in polar and semi‐polar constituents that are readily recoverable using either methanol or hot water.

FIGURE 2.

Data of qualitative analysis of secondary and primary metabolites of methanolic and aqueous extracts of G. angulatus and P. caramanicus.

Notably, alkaloids, anthocyanosides, anthracene heterosides, cardioactive heterosides, and cyanogenetic heterosides were not detected in any sample, indicating the absence (or negligible levels) of these metabolite classes in the tested seed matrices under the applied screening conditions. This profile supports a phytochemical composition dominated by phenolic‐type metabolites (e.g., flavonoid/tannin‐related constituents) rather than glycosides associated with cardiotonic or cyanogenic activity.

A clear solvent‐related difference was observed for lipids, which were present only in the methanolic extracts and absent in the decoction extracts for both species. This is consistent with the expected selectivity of aqueous decoction toward hydrophilic constituents, while methanol can solubilize a broader range of compounds, including relatively less polar fractions.

The qualitative phytochemical screening of the methanolic and aqueous seed extracts of G. angulatus and P. caramanicus revealed a consistent metabolite profile dominated by polar and semi‐polar constituents. The uniform detection of flavonoids, tannins, saponosides, coumarins, starch, and carbohydrates across all samples indicates that these metabolite classes represent the major chemical components of the seed matrices. This finding is in line with previous reports on Fabaceae taxa, where phenolic compounds and glycoside‐related metabolites are frequently reported as predominant constituents, particularly in fruit and seed tissues [20, 21].

The absence of alkaloids, anthracene heterosides, cardioactive heterosides, and cyanogenic heterosides in both extraction types suggests that these metabolite classes are either absent or present at levels below the detection limits of qualitative assays. This observation is consistent with earlier phytochemical surveys indicating that many Cytisus‐related genera are characterized by low alkaloid content and lack of cardiotonic or cyanogenic glycosides, which supports the chemical safety profile of these taxa from a food‐related perspective [21, 22]. A clear solvent‐dependent pattern was observed for lipid detection, with lipids present exclusively in the methanolic extracts and absent in decoction extracts. This result reflects the higher affinity of organic solvents for non‐polar or weakly polar compounds, while aqueous extraction favors hydrophilic constituents such as carbohydrates, starch, and phenolic glycosides [17]. Chlorogenic acid, one of the major phenolic compounds detected in the extracts, is widely recognized for its diverse biological properties, including antioxidant and metabolic effects [23]. Similarly, aqueous plant extracts may contain a broad spectrum of phenolic constituents, highlighting the importance of extraction conditions in determining phytochemical composition and associated biological activities [24].

From a methodological standpoint, qualitative phytochemical screening serves as a preliminary but informative approach, providing a rapid overview of dominant metabolite groups and guiding subsequent targeted analyses. Recent reviews emphasize that such screening data are particularly valuable in under‐studied taxa, where they help prioritize analytical strategies (e.g., LC–MS/MS‐based phenolic profiling) and biological assays without overstating biological or therapeutic relevance [16, 19].

2.3. Quantitative Analysis of Secondary Metabolites

The LC–MS/MS analysis revealed marked differences in the phenolic profiles of G. angulatus and P. caramanicus extracts, as well as pronounced effects of the extraction solvent. Overall, methanolic extracts exhibited substantially higher concentrations and a broader spectrum of phenolic compounds compared to the corresponding decoction extracts, indicating the superior efficiency of methanol for recovering diverse secondary metabolites (Table 1).

TABLE 1.

LC–MS/MS analysis results of the most effective extracts (mg/g).

		G. angulatus		P. caramanicus
No.	Compounds	MeOH	Decoction	MeOH	Decoction
1	Gallic acid	1.073	3.478	6.743	0.165
2	Caffeic acid	0.256	0.117	0.374	—
3	Ferulic acid	0.022	—	0.044	—
4	p‐Coumaric acid	0.022	—	0.069	—
5	Fumaric Acid	—	—	0.003	—
6	Rosmarinic Acid	0.568	—	0.888	—
7	4‐OH‐Benzoic Acid	—	—	0.008	—
8	Quercetin	1.972	0.592	1.902	—
9	Kaempferol	1.278	0.688	7.989	2.001
10	Rutin	2.294	1.284	4.298	1.984
11	Catechin	—	—	1.247	0.023
12	Epicatechin	—	—	0.007	0.003
13	Chlorogenic acid	3.502	1.732	6.707	0.017
14	Hesperidin	0.075	0.007	0.075	0.085

In G. angulatus, chlorogenic acid (3.502 mg/g), rutin (2.294 mg/g), quercetin (1.972 mg/g), and kaempferol (1.278 mg/g) were identified as the dominant constituents in the methanolic extract. Although these compounds were also detected in the decoction extract, their concentrations were markedly lower, suggesting either limited extraction efficiency or partial degradation under hot‐water conditions. An exception was observed for gallic acid, which was present at a higher level in the decoction extract (3.478 mg/g) than in the methanolic extract, consistent with its high polarity and thermal stability (Table 1).

In P. caramanicus, the methanolic extract contained higher levels of kaempferol (7.989 mg/g), chlorogenic acid (6.707 mg/g), gallic acid (6.743 mg/g), and rutin (4.298 mg/g), indicating a comparatively richer phenolic profile than G. angulatus. The decoction extracts also contained measurable amounts of gallic acid, kaempferol, rutin, and hesperidin, albeit at lower concentrations, reflecting the preferential extraction of highly polar constituents by water.

Several compounds, including ferulic acid, p‐coumaric acid, rosmarinic acid, and fumaric acid, were detected predominantly in methanolic extracts, highlighting the critical role of solvent polarity in shaping both qualitative and quantitative phytochemical profiles. The relatively high levels of bioactive flavonoids (quercetin, kaempferol, and rutin) and phenolic acids (chlorogenic, gallic, and caffeic acids) provide a chemical basis for the observed in vitro biological activities.

These findings are consistent with previous studies indicating that methanol is more effective than water for extracting medium‐ to low‐polarity phenolics. The results further suggest that the biological activities arise from the combined effects of multiple phenolic constituents rather than a single dominant compound. However, considering the moderate enzyme inhibition levels, these extracts should be regarded as sources of bioactive phytochemicals with supportive potential rather than direct therapeutic efficacy.

2.4. FTIR‐ATR Spectroscopy

FTIR analyses provide valuable information about the chemical profiles of plant extracts. FTIR‐ATR functional group analysis can provide information about the biochemical composition of plant metabolites such as polysaccharides, phenolics, terpenoids, and lipids, and this information reflects solvent polarity and organ‐specific metabolic variation. In this study, the spectra of aqueous and methanolic seed extracts of G. angulatus and P. caramanicus were examined. The obtained spectra are presented in Figure 3. The transmittance values in all spectra are generally in the range of 95%–100%, indicating that the samples are in good contact with the ATR crystal and that the spectra are of high quality [25].

FIGURE 3.

FTIR‐ATR spectra of the aqueous and methanolic seed extracts of G. angulatus and P. caramanicus. GA_SD: G. angulatus seed decoction; GA_SM: G. angulatus seed methanolic extract; PC_SD: P. caramanicus seed decoction; PC_SM: P. caramanicus seed methanolic extract.

Literature reviews indicate that the broad O–H bands observed in the FT‐IR spectroscopy between 3200–3400 cm⁻¹ represent hydrogen‐bonded hydroxyl (O–H) stretching vibrations and are used as a characteristic indicator of phenolic content in plant extracts [26, 27]. The FTIR‐ATR results obtained show that these bands have higher absorbance in methanol extracts, indicating that methanol extracts phenolic components better than water. The higher O─H band in methanol extracts indicates that these extracts are richer in phenolic hydroxyl groups.

The bands in the 2920–2850 cm⁻¹ region are associated with ─CH₂ and ─CH₃ groups, long aliphatic chains, fatty acids, and terpenoid structures, and these bands are more pronounced in methanol extracts. The more intense reporting of the 2920–2850 cm⁻¹ bands in methanol plant extracts is consistent with methanol being a polar solvent that dissolves polar components through strong hydrogen bonds, thereby increasing the solubility of plant phenolics [28, 29]. The relatively higher intensity observed in the 3000–3400 cm⁻¹ region in methanolic extracts may be associated with the semi‐polar nature of methanol, which can facilitate the extraction of compounds containing hydroxyl functional groups.

C═O vibrations in the 1700–1730 cm⁻¹ region indicate phenolic acids, esterified compounds, and flavonoid carbonyl groups. They are particularly evident in the P. caramanicus seed–methanol extract. The clearer observation of carbonyl bands in seed extracts is consistent with the prevalence of ester and oxygen‐containing functional groups in seed tissues, and strong bands in the 1700 region of alcoholic seed extracts have been reported in the literature before [30, 31, 32].

The 1600–1510 cm⁻¹ region corresponds to aromatic C═C stretching vibrations and is characteristic of phenolic structures, which were observed in all extracts, indicating the presence of aromatic compounds in both species. In seed tissues, the presence of these bands is generally associated with the accumulation of phenolic acids and flavonoids that contribute to protective functions such as antioxidant defense and resistance to environmental stress. Unlike leaves, where phenolic compounds are closely linked to photosynthetic activity, seeds primarily accumulate these metabolites as part of their storage and defense systems. As reported by Schulz and Baranska (2007) [33], IR spectroscopy is a useful tool for detecting such aromatic structures and provides supportive evidence for the presence of phenolic constituents in plant matrices. The observed aromatic C═C bands in the present study therefore reflect the phenolic‐rich composition of the seed extracts, rather than photosynthesis‐related metabolic activity.

The C─O/C─O─C vibration peaks observed in the 1000–1100 cm⁻¹ region indicate the presence of glycosidic and polysaccharide compounds [34, 35]. These bands are more pronounced in aqueous extracts. This is because water efficiently extracts polar polysaccharides and oligosaccharides. These findings reveal that solvent polarity is decisive for the structural properties of the extracted compounds. Methanol preferentially extracts more lipophilic and aromatic components, while water preferentially extracts hydrophilic and glycosidic structures.

The extracts of G. angulatus, particularly the aqueous extract, exhibited bands indicative of a more pronounced polysaccharide and polyphenolic profile, whereas the methanolic extract of P. caramanicus appeared relatively richer in carbonyl and aliphatic compounds, consistent with the presence of lipophilic metabolites in seed tissues. These differences reflect solvent‐dependent extraction behavior, where aqueous extraction favors hydrophilic constituents such as polysaccharides and phenolic glycosides, while methanol enables the extraction of less polar compounds, including aliphatic and ester‐related structures.

Such variations are consistent with the known chemical composition of seeds, which typically contain both hydrophilic storage compounds and lipophilic constituents, including fatty acids and esterified molecules. However, these interpretations are based on general functional group assignments and should be considered with caution. In this context, FTIR‐ATR serves as a rapid and non‐destructive technique for preliminary characterization, providing supportive insights into overall extract composition rather than definitive or compound‐specific identification.

FTIR‐ATR spectroscopy was employed as a rapid and complementary technique to provide a general overview of functional group distributions in the extracts. The spectra indicated solvent‐dependent differences, with aqueous extracts showing relatively stronger bands associated with polysaccharide‐related structures, while methanolic extracts exhibited more pronounced signals corresponding to aromatic and aliphatic functional groups. These observations are consistent with the polarity‐driven extraction behavior and support the LC–MS/MS findings at a general level. However, it should be noted that FTIR analysis provides only broad functional group information and does not allow for compound‐specific identification; therefore, its role in this study is limited to supportive interpretation rather than definitive characterization.

2.5. In Vitro Antidiabetic and Anticholinesterase Activity Assays

Among the tested samples, P. caramanicus seed methanolic extract exhibited notable α‐glucosidase inhibition (20.22%), while α‐amylase inhibition was observed only in the methanolic seed extract of G. angulatus (30.11%) and the seed decoction of P. caramanicus (17.24%). Acarbose, the reference drug, showed the highest α‐glucosidase and α‐amylase inhibitory activities (35.05% and 70.38%, respectively) (Table 2).

TABLE 2.

Antidiabetic and anticholinesterase activities of samples.

		Antidiabetic activity assays		Anticholinesterase activity assays
Species	Used part	α‐Glucosidase inhibitory activity (%) (at 2500 µg/mL ± StD a )	α‐Amylase inhibitory activity (%) (at 5000 µg/mL ± StD a )	Acetylcholinesterase inhibitory activity (%) (at 100 µg/mL ± StD a ) c	Butyrylcholinesterase inhibitory activity (%) (at 1000 µg/mL ± StD a )
G. angulatus	SM	ND b	30.11 ± 0.31	15.27 ± 0.4 b	19.93 ± 0.94
	SD	ND	ND	4.68 ± 0.61 d	ND
P. caramanicus	SM	20.22 ± 0.05 (IC50 = 7408.93 µg/mL)	ND	8.65 ± 0.62 c	5.66 ± 1.97
	SD	ND	17.24 ± 0.19	5.48 ± 0.50 d	5.13 ± 0.68
Acarbose		35.05 ± 1.82 (IC50 = 4625.71 µg/mL)	70.38 ± 1.79 (IC50 = 305.02 µg/mL)	—	—
Donepezil		—	—	99.71 ± 0.43 a (IC50 = 0.31 µg/mL)	99.99 ± 0.19 (IC50 = 15.41 µg/mL)

The concentration ranges used for IC₅₀ determination were as follows: 1–5000 µg/mL for α‐glucosidase and α‐amylase assays, 1–100 µg/mL for acetylcholinesterase (AChE), and 1–1000 µg/mL for butyrylcholinesterase (BChE).

^a StD: Standard deviation.

^b ND: Not determined.

^c Statistical comparisons were performed using one‐way analysis of variance (ANOVA) followed by Tukey's honestly significant difference (HSD) post hoc test (p < 0.05). For AChE inhibitory activity, groups with different letters (a–d) differ significantly. For α‐glucosidase, α‐amylase, and BChE inhibitory activities, statistical comparisons were not applicable because of the limited number of measurable groups.

For anticholinesterase activity, the methanolic seed extract of G. angulatus demonstrated the weakest AChE (15.27%) and BChE (19.93%) inhibition among the plant samples. The seed methanolic extract and decoction of P. caramanicus exhibited moderate AChE inhibition (8.65% and 5.48%, respectively) and lower BChE inhibition. As expected, donepezil exhibited almost complete inhibition of both AChE (99.71%) and BChE (99.99%) (Table 2).

To the best of our knowledge, this study represents the first report investigating the α‐glucosidase, α‐amylase, and anticholinesterase inhibitory activities of G. angulatus and P. caramanicus, and more broadly, one of the limited studies addressing such activities within these genera. Previous studies on Fabaceae species have generally reported moderate enzyme inhibitory effects. For example, an evaluation of 19 traditionally used Fabaceae species demonstrated that, although certain taxa such as Caesalpinia pulcherrima (IC₅₀ = 436.97 µg/mL) and Bauhinia malabarica (IC₅₀ = 745.08 µg/mL) exhibited relatively higher activity, all extracts remained less active than the reference drug acarbose (IC₅₀ = 331 µg/mL), which is consistent with the trends observed in the present study [36]. Similarly, a systematic investigation of 29 Vicia faba varieties reported inhibition rates ranging from 15.17% to 43.33% for α‐amylase and from 4.26% to 22.05% for α‐glucosidase, indicating generally moderate activity levels across plant‐based extracts [37]. These findings support the notion that plant extracts typically exhibit lower inhibitory potency than standard drugs, while still providing a source of bioactive compounds. In this context, the enzyme inhibitory activities observed in the present study should be considered as moderate and indicative of potential functional relevance rather than direct pharmacological efficacy.

A study assessed the anti‐AChE activity of methanolic leaf extracts of Sophora mollis, Mucuna pruriens, and Indigofera atropurpurea from Fabaceae at different concentrations. Among the three species, S. mollis exhibited the most pronounced inhibitory effect, achieving 114.60% inhibition at 1000 µg/mL and showing the lowest IC₅₀ value (75.9 µg/mL), indicating strong enzyme inhibition. The other two species showed comparatively lower levels of AChE inhibition [38]. Another study provided a comprehensive metabolomic profiling and biological evaluation of five commonly consumed legume seeds: fava bean (Vicia faba), black‐eyed pea (Vigna unguiculata), kidney bean (Phaseolus vulgaris), red lentil (Lens culinaris), and chickpea (Cicer arietinum). Methanolic extracts were analyzed using UHPLC‐Q‐TOF‐MS, identifying 139 metabolites including flavonoids, isoflavonoids, saponins, alkaloids, phenolic acids, and lignans. In vitro assays revealed strong antioxidant capacities and significant cholinesterase inhibition, with chickpea showing the highest AChE inhibition (IC₅₀ = 2.66 µg/mL), followed by kidney bean. Pearson's correlation linked 18 metabolites such as sissotrin, genistin, catechin‐3‐glucoside, soyasapogenol E, β‐chaconine, and oleuropein to anti‐cholinesterase activity. Molecular docking confirmed promising binding affinities of these metabolites to AChE and BChE active sites, with β‐chaconine, dehydro‐soyasaponin I, kaempferol‐3‐O‐(2″‐O‐glucosyl)‐rutinoside, kaikasaponin III, N,N′‐diferuloyl putrescine, and oleuropein emerging as the most potent candidates [39].

Taken together, this study provided the first evidence of α‐glucosidase, α‐amylase, and anticholinesterase inhibitory activities of G. angulatus and P. caramanicus. Although their activities were lower than those of standard drugs, the methanolic extracts, particularly the methanolic seed extract of G. angulatus, showed relatively stronger inhibition than the corresponding decoction. These findings align with previous reports on Fabaceae species, where plant extracts generally exhibited lower potency compared to acarbose or donepezil, but still demonstrated promising bioactivity. Collectively, our results highlighted the potential of these species as sources of bioactive compounds and justify further phytochemical and in vivo studies to explore their therapeutic relevance.

2.6. Antioxidant Activity

The antioxidant activities of the samples were evaluated through their ABTS^{• +} and DPPH^• radical scavenging capacities. Both methods are commonly used in vitro tests for assessing antioxidant capacity [40]. The ABTS^{• +} radical scavenging activity results obtained in the study are presented in Figure 4. The DPPH^• radical scavenging activity results obtained in the study are presented in Figure 5.

FIGURE 4.

ABTS radical scavenging activity of methanolic extracts and aqueous decoctions of G. angulatus and P. caramanicus seeds. GA_SM: G. angulatus seed methanolic extract; GA_SD: G. angulatus seed decoction; PC_SM: P. caramanicus seed methanolic extract; PC_SD: P. caramanicus seed decoction. ABTS: 2,2′‐azino‐bis(3‐ethylbenzothiazoline‐6‐sulfonic acid).

FIGURE 5.

DPPH radical scavenging activity of methanolic extracts and aqueous decoctions of G. angulatus and P. caramanicus seeds. GA_SM: G. angulatus seed methanolic extract; GA_SD: G. angulatus seed decoction; PC_SM: P. caramanicus seed methanolic extract; PC_SD: P. caramanicus seed decoction. DPPH: 2,2‐diphenyl‐1‐picrylhydrazyl.

When the ABTS^{• +} scavenging activity results were evaluated, it was observed that the seed methanolic extract of P. caramanicus showed 97.40% ± 0.04% inhibition at a concentration of 80 µg/mL, and this value was comparable to the reference standard Trolox (96.85% ± 0.04%). The seed decoction of the same species also reached a high percentage inhibition value (90.01% ± 1.05%), indicating a strong antioxidant effect. In contrast, the seed methanol extract of G. angulatus showed a moderate ABTS^{• +} scavenging effect (42.08% ± 0.12%), while the activity of the seed decoction was found to be lower (7.82% ± 1.84%). Although the DPPH^• radical scavenging activity results were similar to those of the ABTS tests, it was determined that the extracts of both species exhibited relatively low activity against this radical. In particular, the methanol extract of P. caramanicus seed showed the highest DPPH• scavenging effect with 10.57% ± 0.51%, while this activity was not detected for the G. angulatus seed decoction (ND). In contrast, the reference antioxidants α‐tocopherol and Trolox showed high inhibition percentages in the DPPH^• test (89.41% ± 0.25% and 92.00% ± 0.03%) (Figure 6). Statistical analysis revealed significant differences among the extracts at comparable concentrations (p < 0.05), with methanolic extracts of P. caramanicus showing significantly higher ABTS radical scavenging activity compared to other samples. Similar trends were observed in DPPH assays, although overall activity levels were lower.

FIGURE 6.

ABTS (80 µg/mL) and DPPH (70 µg/mL) Radical Scavenging Activities (%); differences between groups were analyzed using one‐way ANOVA. Samples labeled with different letters are statistically significantly different (p < 0.001). F(5,12) indicates the degrees of freedom for the ANOVA F‐statistic (between‐group df = 5, within‐group df = 12).

One‐way ANOVA revealed statistically significant differences among groups for both ABTS (F(5,12) = 4736.1, p < 0.001) and DPPH (F(5,12) = 2427.9, p < 0.001) radical scavenging activities. Post hoc comparisons using Tukey's HSD test confirmed that PC_SM exhibited significantly higher ABTS scavenging activity than all other plant extracts (p < 0.001), while GA_SD showed the lowest activity, forming a distinct group. These results statistically validate the observed differences in antioxidant capacity among the tested extracts.

The antioxidant activity of plant extracts depends on many factors. It is particularly complex due to the structural diversity of the phytochemicals they contain and the interaction of these compounds through different reaction mechanisms. In addition, the techniques used during the preparation of the plant material also affect the activity. Therefore, evaluating the antioxidant capacity with a single method is not considered a sufficient and comprehensive approach [41]. For this reason, ABTS and DPPH radical scavenging activity tests were used to determine antioxidant capacity in our study.

A literature review revealed no direct studies on the antioxidant capacity of methanolic and aqueous extracts of G. angulatus and P. caramanicus species. However, the Fabaceae family is frequently studied in scientific research due to its numerous species possessing rich phytochemical content. A previous study evaluated 29 different varieties of Chinese faba beans and determined that they are rich in protein, dietary fiber, vitamins, and minerals, as well as containing significant levels of phytochemicals, particularly phenolic compounds. This phytochemical richness was reported to lead to significant differences in antioxidant capacity and enzyme inhibitory effects among faba bean varieties. The main findings indicate that flavonoids are the most important phytochemical groups contributing to antioxidant activity. Lincan 13 (Gansu) variety demonstrated the strongest antioxidant activity, associated with having the highest flavonoid content (40.51 mg rutin/g) (DPPH: 16.32 µmol Trolox/g; ABTS: 5.85 µmol Trolox/g; FRAP: 21.38 mmol Fe^{2 +}/g) [37]. In line with the literature, our study also showed that the methanol extract of P. caramanicus, which has the highest rutin content, exhibited the strongest free radical scavenging capacity.

When the ABTS•⁺ and DPPH^• radical scavenging activity results of the extracts are evaluated together with the phenolic compound profiles determined by LC–MS/MS analyses, it is seen that the antioxidant capacity is largely related to the phenolic content [42].

In a study conducted to determine the antioxidant and antibacterial potential of seeds and seed rinds of six plant species belonging to the Fabaceae family, Cajanus indica Spr., Pisum sativum L., Vicia faba L., Vigna mungo L., Vigna radiata (L.) Wilczek, and Vigna unguiculata (L.) Walp. were extracted separately using cold percolation with n‐hexane, ethyl acetate, methanol, and water. The antioxidant activities of the extracts were evaluated by DPPH free radical scavenging activity, superoxide anion radical scavenging activity, and reducing power analyses. The IC₅₀ values of methanolic extracts from seed were determined to be above 1000 µg/mL. While P. sativum and V. faba seed rind extracts showed high IC₅₀ values (>1000 µg/mL), the IC₅₀ values of the other four seed rind extracts were found to range from 39 to 560 µg/mL. The lowest IC₅₀ value (39 µg/mL) was detected in the methanolic extract of C. indica seed rind. The results showed that the methanol extract of C. indica seed rind exhibited the highest DPPH free radical and superoxide anion radical scavenging activity, as well as the highest reducing capacity and the highest total phenolic compound content. A significant and positive correlation was found between total phenolic compound content and antioxidant activity [43].

Phenolic acids and flavonoids are known as the main phytochemical groups that stabilize free radicals due to their ability to donate hydrogen or electrons [44].

2.7. Multivariate Analysis of Bioactivity and Phytochemical Profiles

Principal component analysis (PCA) was performed using integrated data derived from antioxidant activities (ABTS and DPPH), enzyme inhibitory activities (α‐glucosidase, α‐amylase, AChE, and BChE), and the major phytochemical compounds identified by LC–MS/MS analysis (Figure 7). The first two principal components accounted for 83.1% of the total variance, with PC1 and PC2 explaining 54.2% and 28.9% of the variance, respectively. PC1 was primarily associated with antioxidant activity parameters (ABTS and DPPH) and phenolic compounds such as chlorogenic acid, rutin, kaempferol, and quercetin, whereas PC2 was predominantly driven by α‐amylase, AChE, and BChE inhibitory activities. The methanolic extract of P. caramanicus (PC_SM) was positioned in the positive region of PC1, consistent with its high phenolic content and strong antioxidant activity, while the decoction extract of G. angulatus (GA_SD) was clearly separated on the negative side of both components, reflecting its comparatively low biological activity profile. Overall, the PCA biplot revealed a clear discrimination among samples based on species and extraction method and demonstrated strong relationships between phytochemical composition and biological activities.

FIGURE 7.

PCA biplot showing the differentiation of samples and the influence of phytochemical variables on total variance.

The Pearson correlation matrix clearly demonstrates the presence of strong and statistically meaningful associations between biological activities and phytochemical constituents (Figure 8). In particular, phenolic compounds such as chlorogenic acid, rutin, gallic acid, and caffeic acid exhibited strong positive correlations with ABTS and DPPH radical scavenging activities, indicating that antioxidant capacity is largely driven by these molecules. The very strong correlation observed between α‐glucosidase inhibition and both rutin and chlorogenic acid suggests that these phenolics play a central role in the antidiabetic potential of the extracts. Likewise, the high correlation coefficients between AChE and BChE inhibitory activities and flavonoids such as quercetin and hesperidin indicate that anticholinesterase effects are closely associated with specific flavonoid compounds.

FIGURE 8.

Pearson correlation matrix showing the interrelationships between phytochemical compounds and bioactivities. Asterisks indicate statistical significance levels: ^* p < 0.05.

These findings carry biologically meaningful implications beyond simple statistical associations. The strong positive correlations observed between chlorogenic acid, rutin, and radical scavenging capacity (ABTS and DPPH) can be mechanistically explained by the structural characteristics of these compounds, particularly their multiple phenolic hydroxyl groups, which facilitate hydrogen atom donation and electron transfer to unstable free radicals, thereby mitigating oxidative damage. In terms of antidiabetic potential, the prominent contribution of rutin and chlorogenic acid to α‐glucosidase inhibition is in agreement with previous biochemical and molecular docking studies demonstrating that these phenolics can form hydrogen bonds and hydrophobic interactions within the enzyme's active site, thereby interfering with substrate binding and reducing postprandial glucose release [45, 46]. Regarding anticholinesterase activity, the correlations observed between quercetin and hesperidin and both AChE and BChE inhibition are consistent with the reported ability of flavonoids to interact with both the catalytic active site (CAS) and the peripheral anionic site (PAS) of cholinesterases. This dual binding behavior may enhance cholinergic neurotransmission and is considered relevant in the context of Alzheimer's disease‐related neurodegeneration [47, 48]. Overall, these results indicate that the biological activities of the extracts are not driven by a single dominant metabolite, but rather arise from the synergistic and coordinated effects of multiple phenolic constituents acting through complementary mechanisms.

2.8. Molecular Docking

To elucidate the molecular basis of the observed enzyme inhibitory potential and provide a theoretical mechanistic framework for the experimental findings, the major phenolic compounds identified in the extracts by LC–MS/MS were subjected to molecular docking against acetylcholinesterase (AChE, PDB ID: 4EY7), butyrylcholinesterase (BChE, PDB ID: 4TPK), α‐amylase (PDB ID: 1DHK), and α‐glucosidase (PDB ID: 3A4A) (Tables 3 and 4 and Figures 9, 10, 11, 12, 13, 14, 15, 16).

TABLE 3.

Binding energies of compounds.

Compounds	AChE (4EY7)	BChE (4TPK)	α‐Amylase (1DHK)	α‐Glucosidase (3A4A)
Gallic acid	−9.163	−6.006	−6.210	−5.100
Caffeic acid	−9.013	−5.991	−5.520	−4.080
Ferulic acid	−7.086	−5.191	−5.019	−3.741
p‐Coumaric acid	−6.174	−5.191	−4.740	−3.594
Fumaric Acid	−2.098	−1.969	−2.098	−2.625
Rosmarinic Acid	−6.443	−3.513	−7.853	−4.123
4‐OH‐Benzoic Acid	−6.142	−4.734	−5.591	−6.024
Quercetin	−11.939	−8.682	−7.235	−4.943
Kaempferol	−10.714	−7.389	−6.860	−4.070
Rutin	−11.232	−12.932	−8.043	−7.008
Catechin	−11.262	−8.649	−7.926	−5.030
Epicatechin	−11.704	−9.288	−7.634	−6.092
Chlorogenic acid	−11.519	−7.774	−7.378	−6.532
Hesperidin	−14.571	−9.483	−9.603	−11.058
Donepezil	−20.060	−11.115
Acarbose			−10.111	−8.29

TABLE 4.

Summary of molecular docking results showing binding energies (kcal/mol) and key interactions between major phenolic compounds and target enzymes.

Compound	Enzyme	Binding energy (kcal/mol)	Key residues	Bond type	Distances
Rutin	α‐Amylase	−8.043	Ser180 Glu191 Glu101 Asn115 Val195	H bond H bond H bond H bond Hydrophobic	2.78 Å 2.57 Å 2.7 Å 3.3 Å
Rutin	α‐Glucosidase	−7.008	Ser157 Tyr158 Leu313 Ser240 Pro312 Asp242 Thr310 His280	H bond H bond H bond H bond H bond H bond H bond π–π stacking	1.82 Å 3.38 Å 2.10 Å 2.30 Å 2.51 Å 1.59 Å 1.95 Å
Rutin	AChE	−11.23	Tyr341 Tyr341 Tyr341 Ser293 Leu76	π–π stacking π–π stacking H bond H bond Hydrophobic	1.87 Å
Rutin	BChE	−12.93	Tyr440 His438 Trp231 Thr284 Asp70 Ser79	H bond H bond π–π stacking H bond H bond H bond	3.3 Å 2.67 Å 2.78 Å
Chlorogenic acid	α‐Amylase	−7.378	Ser180 Val100 Val98	H bond H bond H bond	2.87 Å 2.79 Å 2.66 Å
Chlorogenic acid	α‐Glucosidase	−6.532	Arg442 Asp69 His112 His280 Arg213 Asp352 Glh277 Arg446 Tyr158 Val109	H bond H bond H bond H bond H bond H bond H bond H bond π–π stacking Hydrophobic	3.19 Å 2.80 Å 2.69 Å 2.99 Å 3.61 Å 2.59 Å 2.39 Å
Chlorogenic acid	AChE	−11.519	Ser293 Arg296	H bond H bond
Chlorogenic acid	BChE	−7.77	Ser198 His438 Gly117 Leu286 Phe329 Val288 Ala328	H bond H bond H bond H bond π–π stacking Hydrophobic Hydrophobic	2.97 Å 3.16 Å
Quercetin	α‐Amylase	−7.235	Asp103 Gly178 Gly178	H bond H bond H bond	2.76 Å 2.74 Å 3.12 Å
Quercetin	α‐Glucosidase	−4.943	Glu411	H bond	1.89 Å
Quercetin	AChE	−11.94	Trp286 Tyr72 Ser293 Phe295 Arg296	π–π stacking H bond H bond H bond H bond	2.38 Å 2.50 Å
Quercetin	BChE	−8.68	Ser198 His438 Gly117 Tyr440 Phe329 Phe329	H bond H bond H bond H bond π–π stacking π–π stacking	2.79 Å 2.91 Å
Kaempferol	α‐Amylase	−6.860	Val100 Val100	H bond H bond	2.96 Å 4.58 Å
Kaempferol	α‐Glucosidase	−4.070	Glh277 Arg315	H bond π‐cation	1.69 Å
Kaempferol	AChE	−10.71	Trp286 Trp286 Tyr341 Tyr341	π–π stacking π–π stacking π–π stacking π–π stacking
Kaempferol	BChE	−7.39	Tyr440 Phe329 Phe329 Ser198 Gly117	H bond π–π stacking π–π stacking H bond H bond	2.86 Å
Gallic acid	α‐Amylase	−6.21	Ser176 Ser176 Gly178 Gly178 Ser180 Arg42 Phe44	H bond H bond H bond H bond H bond H bond π–π stacking	2.76 Å 3.01 Å 2.74 Å 3.51 Å 3.00 Å 2.61 Å
Gallic acid	α‐Glucosidase	−5.1	Glu411 Arg315 Tyr158	H bond H bond π–π stacking	2.53 Å 3.16 Å
Gallic acid	AChE	−9.16	Ser293 Arg296 Trp286	H bond H bond π–π stacking
Gallic acid	BChE	−6.01	Trp82 His438	π–π stacking H bond	2.61 Å

FIGURE 9.

Molecular docking interactions of Chlorogenic acid, Rutin, Quercetin, Kaempferol, and Gallic acid with α‐Amylase (PDB ID: 1DHK): (a) Chlorogenic acid ‐ 1DHK complex, (b) Rutin ‐ 1DHK complex, (c) Quercetin ‐ 1DHK complex, (d) Kaempferol ‐1DHK complex, and (e) Gallic acid ‐ 1DHK complex.

FIGURE 10.

Molecular docking interactions of Chlorogenic acid, Rutin, Quercetin, Kaempferol, and Gallic acid with α‐Glucosidase (PDB ID: 3A4A): (a) Chlorogenic acid ‐ 3A4A complex, (b) Rutin–3A4A complex, (c) Quercetin–3A4A complex, (d) Kaempferol–3A4A complex, and (e) Gallic acid–3A4A complex.

FIGURE 11.

Molecular docking interactions of Chlorogenic acid, Rutin, Quercetin, Kaempferol, and Gallic acid with AChE (PDB ID: 4EY7): (a) Chlorogenic acid–4EY7 complex, (b) Rutin–4EY7 complex, (c) Quercetin–4EY7 complex, (d) Kaempferol–4EY7 complex, and (e) Gallic acid–4EY7 complex.

FIGURE 12.

Molecular docking interactions of Chlorogenic acid, Rutin, Quercetin, Kaempferol, and Gallic acid with BChE (PDB ID: 4TPK): (a) Chlorogenic acid–4TPK complex, (b) Rutin–4TPK complex, (c) Quercetin–4TPK complex, (d) Kaempferol–4TPK complex, and (e) Gallic acid–4TPK complex.

FIGURE 13.

Docking poses of compounds identified in G.angulatus and P. caramanicus seed extracts with AChE (AChE; PDB ID: 4EY7).

FIGURE 14.

Docking poses of compounds identified in G. angulatus and P. caramanicus seed extracts with butyrylcholinesterase (BChE; PDB ID: 4TPK).

FIGURE 15.

Docking poses of compounds identified in G. angulatus and P. caramanicus seed extracts with α‐amylase (PDB ID: 1DHK).

FIGURE 16.

Docking poses of compounds identified in G. angulatus and P. caramanicus seed extracts with α‐glucosidase (PDB ID: 3A4A).

The reliability of the docking protocol was validated by re‐docking the co‐crystallized ligand of each target into its respective active site. Comparison with the experimental binding pose yielded RMSD values of 0.17 Å for AChE, 1.85 Å for BChE, 1.87 Å for α‐amylase, and 0.16 Å for α‐glucosidase. Since all RMSD values were below the generally accepted threshold of 2.0 Å, the predicted binding poses were considered consistent with the experimental geometries, providing a reliable basis for subsequent analyses. Docking simulations for α‐amylase and α‐glucosidase were performed using PDB structures derived from the same species as those employed in the in vitro enzyme assays (porcine pancreatic α‐amylase, 1DHK; Saccharomyces cerevisiae, 3A4A), whereas human enzyme structures were selected for AChE and BChE. This species difference is expected to have minimal impact because the catalytic triad as well as the CAS and PAS, which govern substrate binding and access, are highly conserved between human and the corresponding animal enzymes. Nevertheless, sequence variations outside the catalytic core may partially influence binding kinetics, and this limitation was taken into consideration when interpreting the findings.

Molecular docking analyses performed against AChE, BChE, α‐amylase, and α‐glucosidase demonstrated that the identified phenolic compounds exhibited varying binding affinities toward all target enzymes. For AChE, hesperidin showed the highest binding affinity (−14.571 kcal/mol), followed by quercetin (−11.939 kcal/mol) and epicatechin (−11.704 kcal/mol). Against BChE, rutin (−12.932 kcal/mol) exhibited the strongest binding affinity, followed by hesperidin (−9.483 kcal/mol) and epicatechin (−9.288 kcal/mol). For α‐amylase, hesperidin (−9.603 kcal/mol), rutin (−8.043 kcal/mol), and catechin (−7.926 kcal/mol) exhibited the most favorable binding energies. Likewise, hesperidin (−11.058 kcal/mol), rutin (−7.008 kcal/mol), and chlorogenic acid (−6.532 kcal/mol) displayed the lowest binding energies against α‐glucosidase (Tables 3 and 4). Molecular docking analyses identified several phenolic constituents with favorable binding affinities toward the investigated enzymes, particularly hesperidin, rutin, chlorogenic acid, and quercetin. These findings provide molecular‐level evidence that the identified metabolites are capable of interacting with the catalytic sites of the target enzymes. However, docking simulations represent theoretical ligand–protein interactions and should be interpreted as complementary mechanistic information rather than direct evidence of biological potency. The relationship between the computational predictions and the experimental enzyme inhibition results is discussed in detail below.

This study provides a comprehensive phytochemical and biological evaluation of seed extracts from G. angulatus and P. caramanicus, highlighting the impact of extraction method on composition and bioactivity. Methanolic extracts were richer in phenolic compounds, while aqueous decoctions yielded higher total extract, reflecting the predominance of polar constituents. FTIR‐ATR results supported these solvent‐dependent differences at a general level. Among the species, P. caramanicus showed a comparatively higher phenolic content and stronger antioxidant activity.

Although enzyme inhibitory effects (α‐glucosidase, α‐amylase, AChE, and BChE) were moderate compared to standard compounds, they indicate a potential multitarget effect associated with the combined action of phenolics. Molecular docking results were consistent with the experimental findings, suggesting plausible interactions between major compounds and target enzymes. Although the investigated extracts demonstrated measurable enzyme inhibitory activities, the observed effects were generally moderate when compared to standard reference compounds. Therefore, these results should be interpreted with caution, and the extracts should not be considered as potent inhibitors in their current form. It is well established that crude plant extracts often exhibit lower activity than pure compounds or synthetic drugs; however, they may still serve as valuable sources of bioactive constituents due to the presence of multiple compounds acting through additive or synergistic mechanisms. In this context, the biological activities observed in this study should be regarded as indicative of functional potential rather than direct therapeutic efficacy. Further studies focusing on compound isolation, dose–response evaluation, and in vivo validation are necessary to better define their pharmacological relevance. This study has some limitations. Biological activities were assessed only in vitro without in vivo validation, and enzyme inhibition was moderate compared to standard drugs; therefore, the extracts should be considered as sources of supportive bioactive compounds rather than potent therapeutics.

It should be noted that the enzyme sources used for the in vitro assays and molecular docking were not completely identical. While AChE and BChE inhibition assays were performed using enzymes from electric eel (Electrophorus electricus) and equine serum, respectively, the molecular docking studies employed the corresponding human crystal structures (PDB IDs: 4EY7 and 4TPK). This approach is widely accepted in structure‐based drug discovery because high‐resolution human crystal structures provide detailed information on ligand binding and catalytic interactions. Moreover, the catalytic triad and the residues forming the CAS and PAS are highly conserved among vertebrate cholinesterases, resulting in very similar substrate recognition and inhibitor binding modes across species [49, 50]. Nevertheless, amino acid substitutions outside the conserved catalytic pocket may produce subtle differences in enzyme kinetics and ligand affinity. Therefore, the docking results should be interpreted primarily as mechanistic evidence supporting potential ligand–enzyme interactions rather than as direct quantitative predictors of the inhibition values obtained in vitro. This limitation has been taken into account when discussing the relationship between the computational and experimental findings [51].

The relatively weak enzyme inhibitory activities observed in the present study should be interpreted with caution. Unlike purified compounds, crude plant extracts represent highly complex matrices containing numerous primary and secondary metabolites that may interact through synergistic, additive, or antagonistic mechanisms, thereby modulating the overall biological response [52, 53, 54]. Consequently, the inhibitory activity of an extract cannot be directly attributed to the docking behavior of individual constituents. Although several phenolic compounds exhibited favorable binding energies in molecular docking analyses, these calculations were performed using isolated ligands under idealized conditions, whereas the biological assays evaluated the combined effects of all extract constituents. Moreover, LC–MS/MS quantification demonstrated that the detected phenolic compounds were present at relatively low concentrations in the analyzed extract solutions (ng/mL level), suggesting that the effective concentrations of individual metabolites during the enzyme inhibition assays were considerably lower than those modeled computationally. Therefore, even compounds displaying favorable theoretical binding affinities may contribute only marginally to the overall inhibitory activity of the crude extract. In this context, the docking analyses should be regarded as mechanistic evidence supporting potential molecular interactions rather than direct predictors of biological potency. Accordingly, the modest inhibition observed against α‐glucosidase, α‐amylase, AChE, and BChE is more appropriately described as marginal enzyme inhibitory activity, whereas the major biological strength of these extracts lies in their pronounced antioxidant capacity, particularly the ABTS radical scavenging activity, together with their rich phenolic composition revealed by LC–MS/MS analysis.

Although the investigated extracts exhibited measurable inhibitory effects against α‐glucosidase, α‐amylase, AChE, and BChE, these activities were generally weak to marginal and should not be interpreted as evidence of strong antidiabetic or anti‐Alzheimer potential. Instead, the principal contribution of this study lies in the comprehensive LC–MS/MS‐based phytochemical characterization of two previously underexplored Fabaceae species together with the demonstration of their remarkable antioxidant activity, particularly the ABTS radical scavenging capacity of the methanolic extract of P. caramanicus. Molecular docking analyses provide complementary mechanistic insights into possible ligand‐enzyme interactions but should be interpreted within the context of the complex chemical composition of crude plant extracts.

3. Experimental Section

3.1. Plant Material

Podocytisus caramanicus and G. angulatus of both plant samples were collected from the Serper vicinity (Ermenek/Karaman) in July 2020 and identified by Dr. Ömer ÇEÇEN. Voucher specimens have been deposited at the Herbarium of Biodiversity Application and Research Center, Karamanoglu Mehmetbey University (Podocytisus caramanicus KMUB 6559, Gonocytisus angulatus KMUB 6568).

3.2. Extraction

For the preparation of decoction and methanolic extracts, 15 g of dried seeds from G. angulatus and P. caramanicus were separately weighed and finely powdered. For decoction, the powdered samples were extracted with distilled water, filtered individually, frozen at −80°C, and subsequently lyophilized.

For methanolic extraction, each powdered sample was extracted separately with 70 mL of methanol under continuous mixing for 8 h. This extraction procedure was repeated three times to ensure maximum recovery of extractable compounds. The combined methanolic filtrates were then concentrated under reduced pressure using a rotary evaporator.

All obtained extracts were weighed individually, and extraction yields were calculated based on the dry weight of the starting plant material [55].

3.3. Qualitative Screening of Primary and Secondary Metabolites

3.3.1. Detection of Alkaloids

For alkaloid screening, 0.5 g of each methanolic extract was treated with 10 mL of 70% ethanol containing 6% H₂SO₄, gently heated for a short period, and allowed to cool. The mixture was then divided into test tubes, and Mayer's and Dragendorff's reagents were added separately to observe precipitate formation. Samples showing positive reactions were rendered alkaline with 25% Na₂CO₃, extracted with chloroform, and subsequently re‐extracted with 10% acetic acid. The acidic phase was re‐tested with Mayer's and Dragendorff's reagents to confirm the presence of alkaloids [55].

3.3.2. Detection of Flavonoids

Methanolic extracts (0.5 g) were diluted with distilled water and subjected to liquid–liquid extraction using ethyl acetate. The ethyl acetate residue was dissolved in a mixture of HCl, methanol, and water (1:1:1), followed by the addition of magnesium powder. The development of characteristic color changes was considered indicative of flavonoids [55].

3.3.3. Detection of Cardioactive Heterosides

For cardiac glycoside screening, 0.5 g of methanolic extract was heated with 10 mL of 70% ethanol in a water bath for 2 min and filtered. The filtrate was diluted with water (1:2) and treated with concentrated lead subacetate to remove interfering substances, followed by filtration. The clarified solution was extracted with chloroform, and the chloroform phase was divided into three portions for further testing.

1. Keller–Killiani Test: The chloroform extract was evaporated to dryness, and the residue was dissolved in 3 mL of 3.5% FeCl₃ prepared in glacial acetic acid. After 1 min, the solution was carefully layered over 2 mL of concentrated H₂SO₄. The appearance of a colored ring at the interface indicated a positive result.

2. Baljet Test: The dried chloroform residue was dissolved in 1 mL of ethanol and treated with Baljet reagent. The formation of a color change was interpreted as evidence of lactone‐containing cardiac glycosides [55].

3.3.4. Detection of Saponins

Methanolic extracts (0.5 g) were vigorously shaken with hot distilled water, and the persistence of stable foam for at least 10 min was recorded. Confirmation was achieved by acidifying the solution with 2 N HCl. Additionally, fluorescence was observed after dissolving the extract in chloroform and adding concentrated H₂SO₄ [55].

3.3.5. Detection of Anthocyanosides

Ethanolic extracts were treated separately with diluted H₂SO₄, NaOH, lead acetate, and amyl alcohol. The occurrence of characteristic color changes was used to indicate the presence of anthocyanosides [55].

3.3.6. Detection of Tannins

Aqueous extracts were analyzed using gelatin‐saline solution, ferric chloride (FeCl₃), and brominated water. The formation of precipitates or specific color reactions confirmed the presence of tannins [55].

3.3.7. Detection of Anthracene Heterosides

Methanolic extracts were hydrolyzed by boiling with diluted H₂SO₄, followed by extraction with ether. The ether phase was treated with ammonia solution, and the appearance of a color change in the ammoniacal layer indicated anthracene heterosides [45].

3.3.8. Detection of Coumarins

Methanolic extracts (0.5 g) were concentrated and treated with 1 N NaOH. The presence of coumarins was confirmed by observing fluorescence under UV light [55].

3.3.9. Detection of Starch

Aqueous extracts were treated with 0.1 N iodine solution. The development of a purple or blue coloration indicated the presence of starch [55].

3.3.10. Detection of Lipids

Petroleum ether extracts were concentrated and applied onto filter paper. The formation of a translucent oily spot was considered indicative of lipids [55].

3.3.11. Detection of Carbohydrates

Aqueous extracts were subjected to Fehling's test, Molisch reaction (1‐naphthol and H₂SO₄), and resorcinol‐HCl test. Positive color reactions confirmed the presence of carbohydrates [55].

3.3.12. Detection of Cyanogenic Glycosides

Methanolic extracts were tested using filter paper impregnated with picric acid and sodium carbonate. The development of a red coloration on the paper was interpreted as evidence of cyanogenic glycosides [55].

3.4. Quantitative Analysis of Secondary Metabolites

The quantitative determination of secondary metabolites in the most active extracts was performed using an Agilent 6460 Triple Quadrupole LC–MS/MS system (Agilent Technologies, USA) at the Atatürk University East Anatolia High Technology Application and Research Center (DAYTAM). Chromatographic separation was achieved on an Agilent Poroshell 120 EC‐C18 column (4.6 × 100 mm, 3.5 µm) coupled to an Agilent 1260 HPLC system. The analyses were carried out in positive electrospray ionization (ESI+) mode using multiple reaction monitoring (MRM), which enables sensitive and selective quantification of target compounds. Quantification was performed based on external calibration curves constructed using authentic reference standards prepared at different concentrations. The protonated molecular ions ([M+H]⁺) of each analyte were first identified, and specific precursor–product ion transitions were optimized for MRM analysis. Calibration curves were generated by plotting peak area versus concentration, and linearity was confirmed with high correlation coefficients (R ² > 0.99). The concentrations of the analytes in the extracts were calculated using these calibration curves and expressed accordingly. The method ensures reliable quantification through selective ion monitoring and comparison with standard compounds, in line with previously validated LC–MS/MS approaches reported in the literature by Akbaba et al., 2026. The mobile phase consisted of solvent A (water with 0.5% formic acid) and solvent B (acetonitrile with 0.5% formic acid), and the injection volume was set at 5.0 µL [56].

3.5. FTIR‐ATR Measurements

FTIR‐ATR spectroscopy measurements were performed using a Shimadzu IRSpirit series spectrophotometer. Spectra were recorded in the 400–4000 cm^{− 1} range with a resolution of 2 cm^{− 1}, averaging 20 scans per sample. Plant extract samples were compressed within the instrument, using an ATR crystal compatible with the solid sample. Spectra were subjected to ATR correction to compensate for the wavelength‐dependent penetration depth of the IR beam. After each sample, a background spectrum was recorded. The following spectrum processing procedures were applied: First, an air spectrum was collected under the same conditions as those used to study the samples. Next, the air absorption spectrum was automatically subtracted from the samples' spectra, leaving only the FTIR‐ATR spectra. The experiments were conducted at ambient temperature. SpectraGryph (v1.2.16.1) software was used to visualize the results. The use of a solid‐sample ATR crystal allowed for high‐quality surface measurements, while ATR correction minimized wavelength‐dependent variations in penetration depth [25]. These adjustments ensured reliable spectra for comparative analysis between extracts.

3.6. In Vitro Antidiabetic and Anticholinesterase Activity Assays

3.6.1. α‐Glucosidase Inhibitory Activity

The inhibition of α‐glucosidase was assessed following the general protocol of Bachhawat et al. (2011) with slight modifications [47, 57]. In a 96‐well plate, 50 µL of phosphate buffer (50 mM, pH 6.9) was combined with 10 µL of α‐glucosidase solution (1 U/mL, derived from S. cerevisiae) and 20 µL of the test sample. The mixture was preincubated at 37°C for 5 min. The enzymatic reaction was initiated by adding 20 µL of 3 mM p‐nitrophenyl‐α‐D‐glucopyranoside (pNPG). After 10 min of incubation at 37°C, the reaction was stopped with 50 µL of 0.1 M sodium carbonate. The formation of p‐nitrophenol was measured at 405 nm using a microplate reader. Acarbose served as the reference inhibitor. Each measurement was performed in triplicate, and inhibition was expressed as a percentage using the formula:

$$Inhibition\left(\%\right)=\left[\left(A_{control}-A_{sample}\right)/A_{control}\right]\times 100.$$

3.6.2. α‐Amylase Inhibitory Activity

The α‐amylase inhibition test was conducted based on Nampoothiri et al. (2011) with slight adjustments [58]. Briefly, 100 µL of each sample was mixed with 100 µL of a 1% soluble starch solution (prepared in 20 mM sodium phosphate buffer, pH 6.9, containing 6 mM NaCl). The mixture was incubated at 25°C for 10 min. Afterwards, 100 µL of α‐amylase from porcine pancreas (0.5 mg/mL) was added, and the mixture was further incubated for another 10 min at 25°C. The enzymatic reaction was terminated with 200 µL of dinitrosalicylic acid (DNS) reagent, followed by heating at 100°C for 5 min. After cooling, 50 µL of the solution was transferred to a new plate, diluted with 200 µL distilled water, and absorbance was read at 540 nm. Acarbose was used as a positive control. Each experiment was repeated three times, and the inhibition percentage was calculated using the formula:

$$Inhibition\left(\%\right)=\left[\left(A_{control}-A_{sample}\right)/A_{control}\right]\times 100.$$

3.6.3. AChE and BChE Inhibition

The inhibitory effects on AChE and BChE were determined using a modified version of the Ellman method described by Ingkaninan et al. (2000) [59]. In each well, 125 µL of DTNB solution (3 mM) was mixed with 25 µL of substrate (acetylthiocholine iodide for AChE or butyrylthiocholine iodide for BChE, 15 mM), 50 µL of Tris‐HCl buffer (50 mM, pH 8.0), and 25 µL of the test sample. The reaction was initiated by adding 25 µL of the corresponding enzyme (AChE from electric eel or BChE from equine serum). After incubating for 10 min (AChE) or 15 min (BChE) at room temperature, absorbance was measured at 405 nm. Donepezil was used as the reference standard. Each measurement was performed in triplicate, and inhibition percentages were calculated according to the equation:

$$Inhibition\left(\%\right)=\left[\left(A_{control}-A_{sample}\right)/A_{control}\right]\times 100.$$

The enzyme inhibition assays were performed using α‐glucosidase from S. cerevisiae, α‐amylase from porcine pancreas, AChE from electric eel, and BChE from equine serum. All enzymes were purchased from Sigma–Aldrich (St. Louis, MO, USA) and used according to the manufacturer's recommendations.

3.7. Antioxidant Activity

3.7.1. ABTS^{• +} Radical Scavenging Activity

The ABTS cation radical scavenging activities of the samples were determined based on the method developed by Re et al. (1999) [60]. In this method, α‐tocopherol and Trolox were used as reference antioxidants, and antioxidant activities were tested with a 2 mM ABTS^{• +} solution. The percentage inhibition values of the standards against the ABTS^{• +} radical at different concentrations (1‐100 µg/mL) were determined, and calibration curves were created based on these values (R ²>0.99). The concentration range to be studied for the samples was determined in accordance with the obtained calibration curves. Serial dilutions of the extracts were prepared in the concentration range of 10–100 µg/mL, and their antioxidant activities against the 2 mM ABTS^{• +} solution were measured. All absorbance measurements were performed at a wavelength of 734 nm against a blank sample containing phosphate buffer [61]. Each analysis was repeated three times. The ABTS^{• +} radical scavenging capacities of the samples were calculated using the following equation:

$$Inhibition\left(\%\right)=\left[\left(A_{control}-A_{sample}\right)/A_{control}\right]\times 100.$$

3.7.2. DPPH^• Radical Scavenging Activity

The DPPH radical scavenging activities of the samples were determined based on the method developed by Blois, 1958 [62]. In this method, α‐tocopherol and Trolox were used as reference antioxidants, and antioxidant activities were tested with a 1 mM DPPH^• solution. The percentage inhibition values of the standards against the DPPH^• radical at different concentrations (1‐100 µg/mL) were determined, and calibration curves were created based on these values (R ²>0.99). The concentration range to be studied for the samples was determined in accordance with the obtained calibration curves. Serial dilutions of the extracts were prepared in the concentration range of 10–100 µg/mL, and their antioxidant activities against the 1 mM DPPH^• solution were measured. All absorbance measurements were performed at a wavelength of 517 nm against a blank sample containing absolute ethanol [61]. Each analysis was repeated three times. The DPPH^• radical scavenging capacities of the samples were calculated using the following equation:

$$Inhibition\left(\%\right)=\left[\left(A_{control}-A_{sample}\right)/A_{control}\right]\times 100.$$

3.8. Molecular Docking

Molecular docking was performed using the Schrödinger Suite. The crystal structures of AChE (4EY7) [63], BChE (4TPK) [64], α‐amylase (1DHK) [65], and α‐glucosidase (3A4A) [66] were retrieved from the PDB. Protein structures were prepared using the protein preparation wizard by assigning bond orders, adding hydrogen atoms, optimizing hydrogen‐bonding networks, and minimizing the structures using the OPLS4 force field. Receptor grids were generated by centering the grid box on the co‐crystallized ligand within the active site of each enzyme. Ligands were obtained from PubChem with the following CIDs: gallic acid (370), caffeic acid (689043), ferulic acid (445858), p‐Coumaric acid (637542), fumaric acid (444972), rosmarinic acid (5281792), 4‐OH‐benzoic acid (135), quercetin (5280343), kaempferol (5280863), rutin (5280805), catechin (9064), epicatechin (72276), chlorogenic acid (1794427), and hesperidin (10621). Ligand ionization states were generated using Epik within the LigPrep module. Molecular docking calculations were performed using Glide in extra precision (XP) mode. The best‐ranked pose for each ligand was selected according to the GlideScore value and subsequently analyzed for hydrogen bonding, π–π stacking, hydrophobic interactions, and other non‐covalent interactions with active‐site residues. To validate the docking protocol, the co‐crystallized ligand of each target enzyme was re‐docked into the corresponding active site. The reproduced binding pose was compared with the experimental crystallographic conformation, and RMSD values below 2.0 Å were considered indicative of an acceptable docking protocol.

3.9. Data Analysis

Statistical analyses were carried out using R software (Version 2025.09.2 Build 418). Differences among antioxidant activities (ABTS and DPPH) and enzyme inhibitory activities (α‐glucosidase, α‐amylase, AChE, and BChE) were evaluated by ANOVA. Post hoc comparisons between groups were performed using Tukey's HSD multiple comparison test, and statistical significance was accepted at p < 0.05. To provide an integrated assessment of similarities and differences among samples, PCA was applied, and PCA score plots and biplot visualizations were generated using the FactoMineR and factoextra packages. The direction and strength of associations between biological activity parameters and phytochemical constituents were examined using Pearson correlation analysis, with the statistical significance of correlation coefficients tested at the p < 0.05 level. The correlation matrix was visualized using the ggcorrplot package. To minimize the influence of differences in measurement units and variable scales, all data were subjected to autoscaling (mean‐centering and standardization) prior to analysis.

Author Contributions

Burak Bayrak: data curation, formal analysis, software, writing – original draft. Ferhat Güneş: software, data curation, resources, visualization, writing – review and editing. Beyzanur Akyüz: data curation, software, writing – original draft, writing – review and editing. Elif Beyza Özer: data curation, formal analysis, software, writing – original draft. Amenah Imad Abdulqader: data curation, software, writing – original draft, writing – review and editing. Ayşe Çivaş: data curation, software, writing – original draft, writing – review and editing. İsranur Küçüksu: data curation, software, writing – original draft, writing – review and editing. Hatice Rümeysa Sönmez: data curation, software, writing – original draft, formal analysis. Ömer Çeçen: investigation, resources, writing – original draft. Bilge Aydın: data curation, investigation, formal analysis, software, writing – original draft. Furkan Çoban: formal analysis, methodology, visualization, validation, writing – original draft, writing – review and editing. Hafize Yuca: methodology, software, conceptualization, validation, writing – original draft, writing – review and editing. Songül Karakaya: methodology, formal analysis, conceptualization, supervision, software, writing – original draft, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Declaration of Generative AI and AI‐Assisted Technologies in the Writing Process

The authors acknowledge the use of ChatGPT Plus (OpenAI) for English‐language editing. All suggestions were reviewed and verified by the authors. The graphical abstract was prepared independently by the authors without AI‐generated visual content.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request and with the permission of the authors.