Since time immemorial, mankind has used plant extracts from different medicinal plants to cure many diseases and thus relieve physical agony, playing an essential role in the world’s healthcare (Yadav et al., 2006). A wide array of phytoconstituents, called secondary metabolites, which do not appear to contribute directly to plant growth and reproduction, are responsible for the pharmacological and therapeutic effects. Thus, purification, isolation, and bioactive assessment of such secondary metabolites from medicinal and vegetative plants are widely practiced (Duraipandiyan et al., 2006).

Araceae plants (family Arum ) are commonly known as “aroids” and are distributed all over the world, and are abundant in tropical and sub-tropical regions (Ara and Hassan, 2019). Colocasia is one of the 27 genera of the Araceae family available in Bangladesh. It is also native to southeastern Asia and the Indian subcontinent (Wagner et al., 1990; Ara and Hassan, 2019). In Bangladesh, this genus of flowering plants is known to contain the following nine species: C. affinis Schott, C. esculenta (L.) Schott, C. fallax Schott, C. gigantea (Blume) Hook. f., C. heterochroma H. Li et Z.X. and Wei, C. lihengiae C.L. Long et K.M. Liu, C. mannii Hook. f., C. oresbia A. Hay, and C. virosa Kunth (Ara and Hassan, 2012). Colocasia leaves have demonstrated antidiabetic, antihypertensive, immunoprotective, neuroprotective, and anticarcinogenic activities. Detailed assessment of the phytochemical compounds present in various extracts of the leaves has shown the presence of active chemical compounds like anthraquinones, apigenin, catechins, cinnamic acid derivatives, vitexin, and isovitexin, which are possibly responsible for the exhibited biological properties (Gupta et al., 2019). Phytochemical extraction and structural elucidation of Colocasia leaves also yield notable bioactive chemical compounds, such as isoorientin, orientin, isoschaftoside, Lut-6-C-Hex-8-C-Pent, vicenin, alpha-amyrin, beta-amyrin, monoglycerol stearic acid, penduletin anthraquinones, apigenin, catechins, cinnamic acid derivatives, vitexin, and isovitexin (Gupta et al., 2019).

Pathogenic bacteria are one of the leading causes of morbidity and mortality in humans, driving pharmaceutical companies to develop numerous new antibacterial agents to address infections, which have become a global concern. Clinical microbiologists are particularly interested in plant-derived antimicrobials for two primary reasons: the potential of phytochemicals to serve as effective antimicrobial agents prescribed by healthcare professionals, and the need to increase awareness of the risks associated with the misuse of conventional antibiotics. Additionally, diarrhea is a common condition marked by frequent bouts of watery bowel movements and abdominal pain. It is a major problem in developing countries, often leading to malnutrition and even death (Workneh et al., 2024). Plant extracts have shown promise in treating diarrhea by helping the body reabsorb water, preventing electrolyte loss, and slowing gut movements (Agbor et al., 2004; Alam et al., 2020). Furthermore, inflammation manifests through a multitude of pathways and different mediators, leading to a spectrum of adverse effects like necrosis, degeneration, and various forms of exudation (Medzhitov, 2008). To alleviate pain and inflammation, analgesic medications such as opioids (like morphine and fentanyl), NSAIDs, and emerging treatments such as gabapentin, carbamazepine, and ketamine are commonly employed. Glucocorticoids exert their effects by binding to receptors, resulting in enhanced transcription of anti-inflammatory proteins (such as IL-1 antagonists) and inhibition of activated transcription factors (e.g., NF-κB) (Barnes, 1998). NSAIDs also inhibit cyclooxygenase enzymes (COX-1 and COX-2), which are responsible for synthesizing various inflammatory mediators. Despite the array of NSAIDs available, they pose significant side effects, including gastrointestinal ulceration, liver toxicity, and kidney disease, particularly with prolonged usage (Sostres et al., 2010). More specifically, COX-2 selective inhibitors run the risk of harming the heart, while non-selective inhibitors tend to wreak havoc on the gut and kidneys (Emon et al., 2021a). Hence, exploring novel phytochemicals holds promise for the development of more effective alternatives (Liu, 2007).

In addition to in vitro and in vivo studies, in today’s scientific realms, computational biology plays a pivotal role, in facilitating the generation and validation of vast amounts of data used by today’s molecular and experimental biologists. Presently, this paradigm enables the meticulous exploration and verification of drug design endeavors for nascent molecules. The adoption of computer-aided drug discovery (CADD) techniques alongside molecular docking stands as an efficacious and expedient in silico approach (Parmar et al., 2023; Parmar et al., 2020). Prudent molecular docking methodologies must adeptly discern the positioning of the native ligand within the three-dimensional confines of the binding site of the protein structure while also taking into account their intricate physicochemical interactions (Parmar et al., 2022; Guedes et al., 2014).

The present work was carried out for the isolation and identification of bioactive secondary metabolites from the species C. gigantea and C. affinis, along with in vitro, in vivo, and in silico investigation of the bioactivities of the identified compounds. Isolation and identification were carried out by chromatographic separation, followed by ¹H-NMR analysis. Then, the identified compounds were employed for the assessment of their antibacterial antidiarrheal and analgesic potentials. Furthermore, the potential of the identified compounds for these activities was also assessed by molecular docking. An ADME/T study was done for their pharmacokinetic properties.

This study revealed the first-time report on compound isolation ascertained by structure elucidation employing NMR technology from two Colocasia species, namely, C. gigantea and C. affinis. Previous reports on these plants only focused on phytochemical profiling through HPLC-DAD and GC-MS (Zilani et al., 2021; Alam et al., 2024). Thus, phytochemical isolation and the elucidation of their structures are the novelty of this current research. These revealed three bioactive flavonoids along with two triterpenoids and one monoglyceride in total, which sheds light on the polyphenol-rich candidacy of these plants, which may result in promising lead compounds for future drug discovery and development. Dichloromethane fractions of both plants were used in phytochemical screening and isolation of compounds. A total of four compounds were elucidated from the dichloromethane soluble fraction of C. gigantea as a result of consecutive chromatographic separation and purification. The compounds isolated from C. gigantea are penduletin (1), a mixture of α-amyrin (2a) and β-amyrin (2b), and a monoglyceride of stearic acid (3). Another three compounds were isolated from the dichloromethane-soluble fraction of C. affinis following the same technique: penduletin (4) (which was also isolated from C. gigantea), 7,8 -(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol (5), and a mixture of 7,8 -(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol (5) and 4′,7,8-trihydroxy flavonol (6).

The ¹HNMR spectrum (500 MHz, CDCl3) of compounds 1 and 4 (Figure 4) in CDCl₃ indicated the presence of three methoxy groups at δ_H 4.04 (3H, s), δ_H 4.02 (3H, s), and δ_H 3.90 (3H, s). The two pairs of ortho-coupled (J = 9.0 Hz) doublets at δ_H 7.04 and δ_H 7.90 showed that ring B is monosubstituted at C-4’. The peak at δ_H 12.79 belongs to 5-OH. There is a one-proton singlet at δ_H 6.59. There are three possible positions for this singlet: C-3, C-6, and C-8. Another proton singlet at δ_H 6.37 could be assigned to the -OH group at C-4’. These ¹H NMR data are in close agreement with the published value of penduletin (Wang et al., 1989; Makhmoor and Choudhary, 2010), where the singlet at δ_H 6.59 is at the C-8 position, and the three methoxy groups are at the C-3, C-6, and C-7 positions. Thus, the structure of compounds 1 and 4 can be concluded as penduletin.

The ¹H NMR spectrum (400 MHz, CDCl₃) of compound 2 (Figure 5) showed two triplets at δ_H 5.32 ppm and δ_H 5.28 ppm characteristic of the olefinic proton (H-12) of α- and β-amyrin, respectively. In addition, eight singlets and two doublets of nCH₃ protons were identified in the range δ_H 1.28-0.79 ppm. The singlet at δ_H 0.89 (3H) and δ_H 0.88 ppm (3H) indicated the presence of CH₃-28. This means that the analyzed component was not an acid with the carboxyl group at C-17. The peaks of the methyl groups of α-amyrin were identified at δ_H 1.16 ppm (singlet; CH₃-27), δ_H 0.93 ppm (doublet-doublet; J = 3.2 Hz; CH₃-30), and δ_H 0.82 ppm (singlet; CH₃-29). The protons of the CH₃-27, CH₃-29, and CH₃-30 groups of β-amyrin had peaks at δ_H 1.28 ppm (singlet; 3H; CH₃-27) and δ_H 0.95 ppm (singlet; 6H; CH₃-29 and 30). The other signals were identical for both amyrins: δ_H 1.11 (6H, s; CH₃-26), δ_H 1.01 (6H, s; CH₃-28), δ_H 0.98 (6H, s; CH₃-25), δ_H 0.82 (6H, s; CH₃-23), and δ_H 0.80 (6H, s; CH₃-24). The above data identified compound 2 as a mixture of α-amyrin (2a) and β-amyrin (2b) with a ratio of 4:3 based on the ¹H NMR peak heights. This identity was further confirmed by a direct comparison of its ¹H NMR spectrum with that recorded for a mixture of α-amyrin and β-amyrin in CDCl₃ (400 MHz, CDCl₃) (Migas et al., 2005).

The structure of compound 3 can be proposed as a monoglyceride of saturated fatty acid. The ¹H NMR spectrum (400 MHz, CDCl₃) showed signals at δ_H 4.195 (2H, dd, J = 13.6, 5.6 Hz), δ_H 3.957 (1H, m), and δ_H 3.732 (2H, dd, J = 11.6, 4.0 Hz). This group of signals can be attributed to the protons of the (-CH₂-CH-CH₂-) backbone (glycerol structure). The upfield peak at δ_H 0.903 with three proton intensities corresponds to one –CH₃ proton. Two additional distinct groups of upfield signals at δ_H 2.374 (2H, t, J = 7.6 Hz), δ_H 1.64 (2H, m), δ_H 1.313 (8H, m), and δ_H 1.25 (20H, m) could correspond to 16 (sixteen) methylene groups. The presence of an additional oxymethine proton in the carbon chain can be justified by the signal at δ_H 3.617 (1H, dd, J = 11.6, 5.6 Hz). Comparing all the spectral data on the structure of compound 3 with the previously documented study (Kavadia et al., 2017) supports identifying this compound as a monoglyceride of stearic acid with the structure drawn in Figure 6.

The ¹H NMR spectrum (CDCl3, 500 MHz) of compound 5 (Figure 7) showed two aromatic systems, including one 1,4-disubstituted benzene ring with the doublet at δ_H 7.68 (2H, d, J = 8.5 Hz), one tetra-substituted benzene ring with the ¹H signals at δ_H 6.86 (1H, d, J = 8.5 Hz), δ_H 7.46 (1H, d, J = 8.5 Hz), and a 3″,3″-dimethyl-pyrano group. The presence of 3″, 3″-dimethyl-pyrano group can be proved by two tertiary methyl signals at δ_H 1.55 (6H, s) and two olefinic proton signals at δ_H 5.87 (1H, d, J = 12.5 Hz) and δ_H 6.96 (1H, d, J = 12.5 Hz). All these chemical shift values are in close agreement with the published values of citrusinol except for the presence of an extra aromatic proton signal that could be assigned to H-5. Another change in the chemical shift of the H-6 proton at δ_H 6.84 differs largely in values, which is at δ_H 6.21 for citrusinol (Qing-Hui et al., 2017). This high value of the chemical shift can be justified only if the position of the oxygen atom of the pyran ring is at C-8 rather than C-7. Thus, the structure of compound 5 can be proposed as 7, 8-(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol.

The ¹H NMR spectrum (CDCl₃, 500 MHz) of compound 6 shows signals that exactly match those of compound 5 (7,8-(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol). But the presence of additional peaks suggest the presence of another compound as a mixture. The additional signals include two para-coupled doublets with J = 8.5 Hz at δ_H 8.00 and δ_H 6.88, each integrating for two protons, which were assigned to the coupled H-2′ and H-6′, and H-3′ and H-5′, respectively. A typical ortho-coupled signal for protons is found at C-5 and C-6 at δ_H 7.46 (1H,d, J = 9.0 Hz) and δ_H 7.30 (1H,d, J = 9.0 Hz). All these additional proton signal values are in close agreement with published values (Ponce et al., 2009) of the structure (Figure 8) and suggested as 4′,7,8-trihydroxy flavonol. Thus, the structure of compound 6 appeared as a mixture of 4′,7,8-trihydroxy flavonol with 7,8-(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol (compound 5) in a ratio of 1:2, according to the peak heights in the ¹H NMR spectrum.

Diarrhea, characterized by disrupted intestinal movements and fluid accumulation coupled with increased peristalsis, occurs due to disruptions in the electrolyte permeability of the intestinal membrane (Bristy et al., 2020). Infectious diarrhea primarily arises from the invasion and spread of pathogens, particularly Salmonella and Shigella (Panda et al., 2012), against which traditional remedies have proven effective (Koné et al., 2004). The study of phytochemicals extracted from medicinal plants showcased promising antidiarrheal effects attributed to their various phytoconstituents, notably flavonoids, which are known for their antidiarrheal properties (Otshudi et al., 2000). Flavonoids are thought to achieve this by lessening motility in both the small and large intestine and suppressing bowel contractions (Capasso et al., 1991; Meli et al., 1990). All of the identified phytochemicals exhibited more pronounced antidiarrheal activity at higher doses (20 mg/kg b.w.) than at lower doses (10 mg/kg b.w.) by reducing the average weight of total and wet feces, as well as decreasing the total number of feces at a dose of 20 mg/kg, akin to the efficacy of loperamide, the standard drug used to treat diarrhea in this investigation, despite not displaying quite a dose-dependent response. However, unlike flavonoids and amyrin, the ester did not exert noteworthy antidiarrheal activity in mouse models.

The human body orchestrates a myriad of chemical reactions through various catabolic and anabolic processes, yielding a plethora of substances. These reactions, which can induce pain, inflammation, and oxidative stress, are often triggered by inflammatory mediators and reactive oxygen and nitrogen species (RONS) (Wilhelm et al., 2016). Inflammatory cascades, spurred by triggers such as microbial invasion or tissue distress, prompt the mobilization of defense cells like leukocytes to the affected site, a phenomenon orchestrated by receptors such as toll-like receptors (TLRs) and NOD-like receptors (NLRs) (Abdel Motaal and Abdel Maguid, 2005). Furthermore, an array of inflammatory mediators such as eicosanoids, cytokines, chemokines, and vasoactive amines, displays the inflammatory milieu (Sun et al., 2014). Flavonoids have demonstrated the capacity to curtail the abundance of inflammatory cells and the synthesis of MMP-9 (matrix metalloproteinase) and other inflammatory mediators (Li et al., 2012). Moreover, several flavonoids have demonstrated anti-inflammatory prowess by inhibiting the expression of the COX-2 gene (Chen et al., 2000). In the present study, the isolated flavonoids exhibited significant inhibition of writhing and licking compared to the standard. Hence, it can be conjectured that the secondary metabolites, notably flavonoids, may exert an analgesic effect by quelling the inflammatory cascade. Interestingly, other compounds, especially a mixture of α- and β- amyrin, also showed noteworthy activity. Previous studies also support the candidacy of the analgesic activity of amyrin through a prospective association of prostaglandins and TNF-alpha inhibition (Aragao et al., 2008). A monoester extracted from the plant showed relatively mild analgesic activity compared to other compounds, which invalidates its strong candidacy as a lead compound for further drug development exploration.

DHFR, an enzyme present in all living organisms, plays a vital role in cellular metabolic pathways. It facilitates the conversion of dihydrofolate to tetrahydrofolate using reduced nicotinamide adenine dinucleotide phosphate (NADPH) as a cofactor. Tetrahydrofolate is further utilized in the synthesis of thymidylate, purines, and amino acids through biosynthetic processes (Canh Pham and Truong, 2022). Thus, binding to DHFR and inhibition of this enzyme leads to the disruption of nucleic acid synthesis, ultimately causing cell death (Kalogris et al., 2014). Although ciprofloxacin primarily targets GyrA and ParC, docking ciprofloxacin with DHFR would be an exploratory computational exercise to see if there are any meaningful interactions between the drug and the enzyme (Dhillon et al., 2013). In the in silico study, ciprofloxacin showed a docking score of −8.1 kcal/mol with the DHFR enzyme, including carbon-hydrogen bond, pi-alkyl, and unfavorable donor-donor interactions with the respective amino acids (Table 6; Figure 12). Compounds C5, C2a, and C2b surpassed this value and gave docking scores of −9.1 kcal/mol, −8.8 kcal/mol, and −8.8 kcal/mol, respectively (Table 5). Among all the compounds, C5 showed the highest binding affinity with interactions of pi-sigma, alkyl, pi-alkyl, and conventional hydrogen bonds (Table 6; Figure 9 ). The binding interactions for the other compounds are briefly depicted in Table 6. These binding interactions in molecular docking are important as they determine the strength and stability of the ligand binding to the target protein, which directly influences the efficacy of a potential drug or inhibitor (Morris and Lim-Wilby, 2008).

µ, κ, and δ are the opioid receptors that modulate the activity of the enteric nervous system and the release of neurotransmitters, thereby influencing both stimulatory and inhibitory motor pathways. This sequence of events influences gastrointestinal motility and stool consistency by slowing colonic transit, reducing the sensitivity of enteric nerves, and altering the secretion and transportation of fluids (Pannemans and Corsetti, 2018). Loperamide is an opioid agonist and serves as a potential antidiarrheal agent (Kopsky et al., 2019). In computational docking studies, for antidiarrheal activity, the standard loperamide drug molecule showed a binding affinity of −9.1 kcal/mol with the KOR receptor, including pi-sigma, pi-pi stacked, and alkyl bond interactions (Tables 5, 6; Figure 12). Compounds C2b, C5, and C2a had higher docking scores of −10.5 kcal/mol, −9.7 kcal/mol, and −9.4 kcal/mol, respectively, compared to the standard drug (Table 5). The compound β-amyrin (C2b) exhibited the highest affinity among all the compounds, including alkyl bond interactions with the amino acids Tyr 140, Ile 180, Trp 183, Leu 184, and Ile 191 (Table 5; Figure 10). Considering these binding interactions is for in predicting the binding affinity and stability of a ligand–protein complex, which ultimately affects the design of more effective drugs. Effective docking must consider these interactions to identify strong binders with good pharmacological potential (Morris and Lim-Wilby, 2008).

The identified compounds were docked against the COX-2 enzyme to investigate the peripheral analgesic properties. The prostaglandins synthesized from arachidonic acids by the COX-2 enzyme are responsible for the sensation of pain and inflammation (Lee et al., 2005). Therefore, COX-2 inhibitors are considered to work in chronic pain relief and post-surgical analgesia by reducing the production of prostaglandins responsible for inflammation, pain, and fever, along with the minimization of gastrointestinal side effects (Camu et al., 2003). Diclofenac sodium is a well-established COX-2 inhibitor, and it interferes with the production of prostaglandin G2, which is the precursor to other prostaglandins (Kaur and Sanyal, 2011). In computational docking modeling, this standard drug showed a binding affinity of −7.8 kcal/mol against the COX-2 enzyme, including pi-sigma, amide-pi stacked, pi-alkyl, and conventional hydrogen bond interactions (Tables 5, 6; Figure 12). The compound β-amyrin (C2b) showed a higher docking score of −8.7 kcal/mol than the standard and demonstrated a strong conventional hydrogen bond interaction with the amino acid Asp 347 (Tables 5, 6; Figure 11). Compounds C5 and C6 also exhibited notable binding affinities of −7.2 kcal/mol and −7.3 kcal/mol, respectively. These interactions are crucial for specific binding, as they enhance the affinity and selectivity of the ligand toward the target. The interactions improve docking scores, lead to more stable complexes, and play a key role in fine-tuning the fit of the ligand within the binding pocket, ensuring proper spatial alignment (Morris and Lim-Wilby, 2008).

The ADME/T study of the identified compounds also exhibited promising results, which are listed in Table 7. According to Lipinski, if the following requirements, such as molecular weight <500 amu, hydrogen bond acceptor sites <10, hydrogen bond donor sites <5, and lipophilicity value LogP ≤5 are met, then that particular substance is likely to be orally active (Zhang and Wilkinson, 2007). Compounds penduletin (C1/C4) and 4′,7,8-trihydroxy flavonol (C6) did not violate the five Lipinski principles. Additionally, 7,8-(3″,3″-dimethyl-pyrano)-4′-hydroxy flavonol (C5) violated only one of Lipinski’s rules. However, the compounds α- and β-amyrin (C2a and C2b) each had two violations, and the monoglyceride of stearic acid (C3) compound had three violations. The results suggest that the compounds C1/C4 and C5 conform to Lipinski standards, signifying their oral safety and potential as promising candidates for medicinal purposes.

pkCSM is a novel approach to predicting pharmacokinetic and toxicological consequences that use graph-based signatures to represent the chemistry and topology of small compounds (Pires et al., 2015). Table 7 shows that all the test compounds possessed negative water solubility values (log mol/L), suggesting that they are lipophilic, which promotes effective absorption (Kuentz and Arnold, 2009). The bioavailability scores of all compounds were found to be 0.55%. None of the identified test compounds exhibited any hepatotoxicity. Moreover, none of the compounds was found to inhibit hERG I, and except for C2a, C2b, and C5, none of them inhibited hERG II, indicating that these compounds may not be cardiotoxic (Muster et al., 2008).

Finally, from all the hypotheses of the in silico study, the identified compounds that gave promising docking scores with the respective receptors, adhered to Lipinski’s rule, and conformed satisfactorily with pharmacokinetic and toxicological assumptions should be considered for their drug candidacy and further investigation. Compared with existing therapeutic agents, the identified compounds exhibited remarkable effects in bioassays as well as remarkable binding affinities with the corresponding biomolecules.

Vegetables are a source of nutrition and disease management and can be crucial as nutritional strategies for integrative healthcare. The discovery of promising pharmacological activities and bioactive phytochemicals in locally consumed vegetables can add a new dimension to therapeutics, pharmacological research, and development. The present study represents the first-time report on compound isolation from C. gigantea and C. affinis, employing NMR technique. The outcome reveals three bioactive flavonoids, two triterpenoids, and a monoglyceride, underscoring these species’ polyphenol-rich attributes. Upon biological investigations, these compounds showed prospective antimicrobial, analgesic, and antidiarrheal potentials, justifying the tribal use of these taro vegetables. Further extensive scientific research on the isolated compounds in dosage form on a clinical trial basis is highly recommended to ascertain their safety and efficacy profile and mode of action to ease the process of future drug discovery and development.