All the chemicals used for analytical work were of ACS grade. Porcine pancreatic α-amylase (EC 3.2.1.1, A3176) and α-glucosidase from Saccharomyces cerevisiae (EC 3.2.1.20, G5003) were obtained from Sigma-Aldrich. Phosphate buffer and soluble starch were obtained from HiMedia Laboratories (India); 4-Nitrophenyl β-D-glucopyranoside (N7006) was obtained from Sigma-Aldrich. Standard polyphenolic compounds viz. gallic acid, chlorogenic acid, 4-hydroxybenzoic acid, vanillic acid, caffeic acid, rutin, sinapic acid, ferulic acid, naringin, and quercetin were of HPLC reference standard grade and obtained from Sigma-Aldrich.

Edible parts of Ipomoea aquatica (IA) consisting of leaves and tender shoots were obtained from wild natural habitats such as wetlands of the state of Assam, India in the month of July. Freshly collected leaves were washed, cleaned and freeze dried. The moisture content of the leaves was calculated according to the method outlined by AOAC method (930.15). The dry leaves were then ground to fine powder resulting ground dry matter (dm) and used for subsequent analysis.

The proximate composition and mineral profile in IA were determined on dry weight basis using respective AOAC methods (15) as mentioned in Supplementary Table S1. For estimation of total carbohydrate, the dry matter of the leaves were initially digested by 2.5 N HCl and estimated by anthrone reagent (16). The calorific value (kcal/100 g) was calculated from the given formula (17).

where, C, P and L are percentage of carbohydrate, protein and lipid content, respectively.

Extraction of amino acids was carried out according to method 994.12 of AOAC (15) with some modifications (18). Extracted amino acids were derivatized using FMOC and OPA followed by quantification in HPLC-DAD (Agilent, United States). The detailed method of amino acid extraction, derivatization and quantification is listed in Supplementary material.

The phytochemicals of IA were extracted using a hydroalcoholic solvent system consisting of 80% methanol. This solvent system has been previously demonstrated to effectively extract various phenolic components, including flavonoids and other secondary metabolites (19). 5 g of the ground dry matter of IA was extracted under constant agitation for 6 h at room temperature using about 100 mL of 80% methanol. The components were then sonicated for 30 min and extracted again for 90 min in water bath shaker at 40°C. The extracts were centrifuged, and concentrated using a rotary evaporator at 35°C and the remaining content was lyophilized. A portion of the lyophilized extract was reconstituted in HPLC grade methanol for subsequent phytochemical analysis.

Total phenolics content (TPC) was estimated by Folin–Ciocalteu method (20) and the results were expressed as gallic acid equivalent (mg GAE/g dm). Total flavonoid content (TFC) was estimated by aluminum chloride method (21) and expressed as rutin equivalent (mg RE/g dm).

In vitro antioxidant activity of the reconstituted crude IA extracts was determined using five different assay methods namely DPPH radical scavenging assay (22), Trolox equivalent antioxidant capacity (TEAC) assay (23), oxygen radical absorbance capacity (ORAC) assay (24), ferric reducing antioxidant power (FRAP) assay (25) and phosphomolybdate assay for total antioxidant capacity (TAC) (26).

The screening of different phytochemicals from the crude extract was carried out using an Agilent, 6,550 iFunnel quadrupole time of flight (QTOF) mass coupled with and Agilent LC/MS system. The LC method was optimized to separate and detect the maximum possible phytoconstituents from IA crude extract. 5 μL of redissolved lyophilized extract was injected using an autosampler into a reverse-phased column thermostatically maintained at 40°C. The mobile phase comprised of (A) 0.1% formic acid in water; (B) 90% acetonitrile with 10% water and 0.1% formic acid. The flow rate was 0.3 mL/min, and the gradient elution program was set as 95%A for 1 min 0%A for 20 min and 25 min 95%A for 26 min and 30 min.

Mass spectrometric acquisition was done in both positive and negative ESI mode using dual AJS ESI within the range of 100 to 1,100 m/z with MS and MS/MS scan rate 1 spectra/s. The scan source parameters were: VCap 3,500, nozzle voltage 1,000 V, fragmentor 175, skimmer 65 and octupole RFPeak 750. The temperature of gas and sheath gas were 250°C and 300°C, respectively, with flow of 13 L/min and 11 L/min. The data acquisition was performed by Mass Hunter software of Agilent and the detected compounds were identified by searching against a plant metabolite database available with the instrument.

Among all the identified metabolites, 10 polyphenolic compounds viz. gallic acid, chlorogenic acid, 4-hydroxybenzoic acid, vanillic acid, caffeic acid, rutin, sinapic acid, ferulic acid, naringin, and quercetin were quantified from IA extract. Quantification was carried out in a UPLC-DAD system (Agilent, United States) with a polar RP C18 column (250 mm L, 4.6 mm ID, 5 μm particle size) using mobile phase 0.1% formic acid in HPLC grade water (A) and 100% acetonitrile (B). The column was thermostatically maintained at 40°C, flow rate was set at 1 mL/min and 15 μL was used as injection volume. The gradient program for separation was as follows: 90%B for 0 min, 70%B for 8 min, 50%B for 13 min, 40%B for 15 min, 90%B for 18 min and 20 min. DAD detector was set at 255 nm for 4-hydroxybenzoic acid, chlorogenic acid, and vanillic acid; 280 nm for cinnamic acid, gallic acid, naringin and p-coumaric acid; 320 nm for caffeic acid, sinapic acid and ferulic acid; 375 nm for rutin and quercetin.

The metabolite composition of IA was determined by GC 2010 plus with TP-8030 triple quadrupole MS (Shimadzu, Japan), fitted with EB-5 MS column (30 m); briefly – a portion of lyophilized IA extract was mixed with 200 μL methoximation mixture and incubated at 37°C for 90 min. This was followed by the addition of 150 μL BSTFA with 1% trimethylchlorosilane and incubation at 70°C for 60 min. 1 μL of the test portion was injected in splitless mode. Helium was used as carrier gas at a flow of 1 mL/min; columns were isothermally maintained at 80°C for 2 min followed by ramping at 10°C/min to 300°C followed by a hold for 6 min. Injector and ion source temperatures were kept at 280°C and 230°C along with 150°C transfer temperature to MS. The ionized mass fragments were recorded between 50–550 m/z.

Qualitative assessment of oxidative DNA damage prevention (genoprotective effect) of IA extract was carried out on PTZ57R (2,886 bp) plasmid (Fermentus, United States). Lyophilized IA extract was reconstituted on HPLC grade water. The reaction mixture was prepared by adding 16 μL TE buffer with 2 μL DNA and 5 μL extract at different concentrations (1–10 μg/mL) and 2 μL 30% H₂O₂. The negative control comprised of 5 μL water; a reaction as control was set without the addition of H₂O₂. The reaction mixture was exposed to UVC radiation for 15 min to produce hydroxyl radicals and induce DNA damage. After that, the reaction was terminated by the addition of loading dye (6X). The mixtures were transferred to a 1% agarose gel and electrophoresis was carried out followed by documentation in a gel documentation system. The densitometric processing of the gel image was carried out in ImageJ tool (27). The percentage of DNA damage inhibition (DI) was calculated by the following equation.

where, N_S and N₀ are percentage of nicked DNA in lanes with sample and without sample extracts, respectively.

For determination of inhibitory effects of IA on α-amylase, 0.5 U/mL solution porcine pancreatic α-amylase (PPA) was prepared in 0.2 M phosphate buffer, pH 6.5. Subsequently, 0.5 mL of PPA was incubated with different concentrations of 1 mL IA extracts (prepared in phosphate buffer) at 37°C for 30 min. This reaction was initiated by the addition of 0.5 mL soluble starch solution (5 mg/mL) and allowed to be hydrolyzed for 15 min at the same condition. A negative control and a reaction blank were processed in a similar manner without the addition of enzyme and extract, respectively. The reaction was then stopped by adding 0.3 mL dinitro salicylic acid (DNS) reagent and diluted using 0.5 mL phosphate buffer. The reaction mixture was then heated in a boiling water bath for 5 min and allowed to cool at room temperature. The absorbance of the colored complex was recorded at 540 nm using Varioskan™ LUX multimode microplate reader (Thermo Scientific, United States).

For α-glucosidase inhibitory assay, 0.15 U/mL α-glucosidase from Saccharomyces cerevisiae (YGU) was prepared in 0.2 M phosphate buffer, pH 6.5. The p-nitrophenyl α-D-glucopyranoside (pNPG) was used as a reaction substrate. The reaction volume comprised of 50 μL of YGU and 150 μL of inhibitor, i.e., IA extracts at different concentrations, was incubated for 10 min at 37°C. The reaction was initiated by addition of 50 μL of 0.2 mg/mL pNPG followed by incubation for 15 min at 37°C. The reaction was then terminated by addition of 50 μL 0.2 M Na₂CO₃ and the absorbance of the colored solution was recorded at 450 nm using Varioskan™ LUX multimode microplate reader (Thermo Scientific, United States). The results were quantified as maltose equivalent using a calibration curve.

All the samples were measured in triplicates. The percentage of inhibition (I) for both PPA and YGU were calculated using the following formula:

where, A_C, A_S, A_CB, A_SB are the absorbances for control, sample, control blank and sample blank, respectively.

The IC50 values for α-amylase and α-glucosidase inhibition were calculated using the following equation (28).

where, [I] is the concentration of IA extracts or acarbose, and Imax is the maximum percentage inhibition.

The kinetics of PPA inhibition were determined at different concentrations of starch in the range of 0.25–5 mg/mL. Similarly, for YGU, inhibition kinetics were determined at different concentrations of pNPG in the range of 0.1–2.0 mM. IA extract at various concentrations were used inhibitor ([I]); for PPA inhibitor concentration was in the range of 1–2.5 mg/mL and for YGU the range was within 0.05–0.2 mg/mL. Acarbose was used as a standard inhibitor and kinetic parameters were calculated in similar manner to that of IA extract. The enzymatic reactions and substrate utilization were measured spectrophotometrically using the same methods as described above. The reaction velocities were determined by recording the starch digestion at 0, 5, 10 and 15 min, respectively. The kinetic parameters were determined using Michaelis–Menten kinetic model where V is the rate of reaction, V_max is the maximum reaction rate of enzyme, [S] is the substrate concentration, and K_m is the Michaelis–Menten constant. The initial rate of enzyme reaction (V₀) was determined from the quantity of substrate hydrolyzed at different concentrations by plotting a calibration curve. The nature of inhibition was determined by plotting a Lineweaver-Burk double reciprocal plot of 1/V against 1/S for each substrate concentration and IA extract concentration used as inhibitor. The Dixon plot [Eq. (2)] and Cornish Bowden-Eisenthal plot [Eq. (3)] was used to calculate the binding constant K_i, the equation used was as follows:

In order to evaluate the interaction between IA polyphenols with human α-amylase and α-glucosidase, molecular docking was carried out in the active site of both the enzymes. Three dimensional structures of human α-amylase (1HNY), porcine pancreatic α-amylase (1OSE) and α-glucosidase (3L4Y) were obtained from protein data bank (PDB). For both α-amylase and α-glucosidase structures, acarbose interaction was modeled in all the proteins from acarbose co-crystallized templates available at PDB and active sites were mapped. The proteins were processed using the protein preparation wizard of Schrödinger maestro (Schrödinger suite 2021) and docking grid was generated. Structures of all the identified compounds from IA were obtained and processed using LigPrep tool allowing generation of tautomer. All the ligands were then docked in the acarbose binding site grid of all the prepared structures. The stability of the docked complexes was determined by molecular dynamic (MD) simulations. Briefly, the system was generated using orthorhombic boundary conditions containing TIP3P water model and 0.15 M NaCl within 10 Å buffer region. The system was equilibrated at 300 K and 1 bar using NVT and NPT ensembles. Nosé-Hoover chain thermostat and Martyna Tobias Klein barostat were used with isotropic coupling at 1 ps and 2 ps relaxation times. The system was relaxed using default options and MD simulations were performed for 100 ns in Desmond (academic version 2023) using the OPLS4 force field. The binding free energy of the complexes was calculated using the MM-GBSA method.

All the statistical analyzes and curve fitting were carried out using R programming and SPSS Statistics v28.0 (IBM Corp, United States). Data generated were analyzed in triplicates and expressed as mean ± standard error of mean (SEM).

The proximate composition of IA is listed in Table 1. The content of 8 individual mineral elements were shown in Table 2. IA exhibited relatively low protein content, analysis of amino acid profile revealed 20 amino acids with varying amounts (Table 2) and of the total amino acids 60.4% were essential amino acids. Among the essential amino acid lysine was highest with 2.141 g/100 g followed by phenylalanine with 1.891/100 g and isoleucine 1.674 g/100 g. Among the sulfur containing amino acids methionine occurred in relatively high amount which is 1.223 g/100 g and cystine in relatively low amount (0.291 g/100). Furthermore, among the aromatic amino acids, phenylalanine occurred in highest quantities (1.891 g/100 g) followed by tyrosine and tryptophan (Table 2).

IA exhibited total phenolics and flavonoid content of 142.43 mg GAE/g dm and 43.06 mg RE/g, respectively (Table 3). Assessment of different radical scavenging capabilities revealed that IA antioxidant can effectively inhibit ABTS radicals (IC50 of 0.087 mg/mL and 0.924 mg dm eq) as compared to DPPH radicals (IC50 of 0.122 mg/mL and 1.125 mg dm eq). Different antioxidant assay activities exhibited by IA were compared to that of trolox, which is a vitamin E analog, and hence results were expressed as trolox equivalent. In both the cases radical scavenging assays the IC50 values were similar to of the standard compound trolox (Table 3). ORAC assay demonstrated significant capability of IA extract to inhibit free radical damage by measuring the fluorescence intensity. Furthermore, upon assessment of radical scavenging capability by TEAC assay, the antioxidant capacity was found to be 699.35 μmol TE/g dm. Moreover, the ferric reducing power as determined by FRAP assay was observed to be 5762.88 μmol TE/g dm. Determination of total antioxidant capacity by phosphomolybdate assay revealed antioxidant activity of 17.93 mM TE/g dm.

In both positive and negative ESI mode, LC-QTOF-MS analysis has identified more than 65 compounds, and among them 36 compounds were important secondary metabolites (Tables 4, 5). The compounds were identified by database search of HRMS data and similarity percentage of the retrieved compounds were reported. Moreover, characteristic mass fragments of the compound hits were also listed. Total ion chromatograms in the ionization mode were shown in Figure 1. Among the compounds identified, the number of polyphenols and their derivatives were high, followed by some important alkaloids and terpenoids. Polyphenolic compounds such as phenolic acids and flavonoids were detected mostly in negative ESI mode whereas alkaloids and terpenoids were detected mostly in positive ESI mode. Major polyphenolic compounds in IA were comprised of quinic acid, caffeoylquinic acid and its derivatives including caffeic acid and chlorogenic acid. Furthermore, flavone glycosides constituted a dominant portion of IA phytochemical constituents. In addition, non-polyphenolic compounds such as norharman, asclepin, jervine, edulinine, cerbertine, euphornine, jatrophone etc. were also found to be a part of IA phytochemical constituent, which had not been reported in earlier works. Apart from the 36 secondary metabolites identified, the other compounds such as fatty acids, amino acid derivatives etc. which were obtained in database search are listed separately in Supplementary Tables S3, S4.

After identification of different polyphenols in LC-ESI-QTOF-MS analysis 10 major polyphenolic compounds were further quantified using a customized UPLC-DAD method (Figure 2). The quantitative composition of the individual polyphenolic compound is listed in Table 6 and the parameters of the quantification method are listed in Supplementary Table S6. Among the quantified polyphenols prominent were caffeic acid (1553.8 μg/g) and ferulic acid (222.1 μg/g). Furthermore, contents of quercetin and chlorogenic acids were also found marginally high, which was 53.22 μg/g and 27.15 μg/g, respectively. Gallic acid, 4-hydroxybenzoic acid and naringin content were found to be low and varied in the range of 10–15 μg/g.

GC–MS evaluation of metabolites revealed 65 metabolites which were mainly organic acids, fatty acids, amino acids, sugars, and other metabolites such as catecholamines, sterols etc. The compounds were listed based on their category, retention time and percent area of the peak (Table 7). The majority of compounds detected were organic acids, sugar and their derivatives, followed by fatty acids. The analysis revealed the presence of essential fatty acids, mostly poly unsaturated fatty acids along with omega 3 fatty acids such as α-linolenic acid and omega 6 fatty acid such as linoleic which are very rare in conventional food plants. Moreover, the presence of α-linolenic acid was also observed during UPLC-qTOF-MS analysis (as listed in Supplementary Tables S3, S4). Four essential amino acids viz. valine, isoleucine, threonine and tryptophan were detected in metabolic analysis as well as during HPLC quantification of amino acids. Furthermore, the present analysis also identified other plant secondary metabolites such as linalool, squalene etc.

IA crude extracts have exhibited a promising DNA damage prevention activity. The hydroxyl radicals produced during the UVC irradiation have conferred to a total 93.5% of DNA damage. However, the presence of IA extract has inhibited 48.9% of DNA damage at concentration of 10 μg/mL and 18.8% of DNA damage at lowest concentration levels up to 1 μg/mL (Figure 3). The DNA damage inhibition by different concentrations of IA crude extracts were represented by a calibration curve shown in (Figure 3C).

IA crude extracts have exhibited significant levels of inhibition for both the hydrolase enzymes as represented by IC50 values (Table 8). However, YGU inhibition potential was substantially stronger as compared to PPA. The IC50 values of IA inhibition were 0.106 mg/mL for YGU and 1.316 mg/mL for PPA. Likewise, for acarbose IC50 values were 0.210 mg/mL for YGU and 0.115 mg/mL for PPA. The inhibition potential of IA extract was compared to acarbose (Figure 4) and observed that IA extracts have a significantly higher YGU inhibition potential as compared to acarbose. However, in the case of PPA inhibition acarbose inhibition was marginally high.

In order to explore the inhibition kinetics of IA phytochemicals against PPA and YGU, the enzyme kinetic reactions of both the enzymes were investigated with different concentrations of their substrate and IA extract as inhibitor. The kinetics of inhibition was compared to that of acarbose. In order to outline the inhibitory mechanism, the Lineweaver-Burk plots (1/[S] against 1/V) were generated as shown in Figure 5. IA extracts at different concentrations were taken as inhibitors and represented by a series of lines with different slopes and intercept at Y axis. With increasing concentration of IA extracts the value of slope (Km/Vmax) as well as the intercepts (1/Vmax) has increased significantly. This indicated the decrease of Vmax with increase of Km which is a nature of mixed inhibition type. The same nature of inhibition toward PPA was also observed in case of acarbose, however, in case of YGU the nature of inhibition exhibited by acarbose was different. In case of acarbose YGU inhibition, Vmax remained constant with increase of Km representing a competitive inhibition, unlike IA extract which was mixed. All the kinetic parameters of PPA and YGU inhibition were listed in Table 8. Dixon plot was generated by plotting concentration of inhibitor ([I]) against Km/Vmax values (Figure 5). By comparison of Ki values, it was observed that IA exhibited Ki value of 0.278 mg/mL for PPA and 0.017 mg/mL for YGU. Moreover, smaller Ki value YGU inhibition as compared to PPA inhibition indicates that the extracts can strongly inhibit YGU as compared to PPA. Similarly, upon comparison with Ki of acarbose – YGU inhibition (43.12 μg/mL) it has been revealed that IA has potentially high inhibitory effects toward YGU.

The compounds identified in LC-ESI-QTOF-MS and GC/MS analysis were evaluated for their interaction with different α-amylases (AML) and α-glucosidase (AGL). A set of 35 different compounds were selected resulting in a ligand library of 355 molecules after generation of tautomeric forms. The docking scores of all these molecules were listed in Supplementary Table S5 and the protein-ligand interaction of some compounds were shown in Figure 6. The stability of different interactions was represented by RMSD analysis of the MD simulation trajectory for 100 ns (Figure 7). Polyphenols showed a predominant interaction with AML and AGL enzymes, while a limited number of non-polyphenolic compounds such as jervine, asclepin and norharman also displayed significant levels of interaction with these enzymes. Oolonghomobisflavan A, 6″-caffeoylhyperin and herbacetin 3,8-diglucoside represented one of the strongest interactions for AMLs (Figure 6). Flavone glycosides such as rutin, herbacetin 3,8-diglucoside, myricetin 7-rhamnoside etc. represented strong docking score and binding energies for both AML and AGL. Moreover, caffeoylquinic acid derivatives and compounds like jervine and asclepin represented stronger docking score against AML and compounds like norharman for AGL. Although protein structures of human AML and PPA exhibited similar docking scores, the stability of similar compounds docked in the active sites varied nominally (Figure 7). This suggests that some compounds in IA extracts may be more effective at inhibiting human α-amylases as compared to PPA. For instance, kaempferol 3-(2″,3″-diacetyl-4″-p-coumaroylrhamnoside) may interact strongly and stably with human α-amylase as compared to PPA. The docking scores and binding energies of all the molecules were subjected to principal component analysis (Figure 8). PC1 comprised of total 90% variance and represented the compounds with or without significant interaction with α-amylases and α-glucosidase. PC2 comprised of 8% variance and distinguished the compounds that interacted with α-amylases and α-glucosidase. It was observed that flavone glucosides and rhamnosides interact strongly with AGL as compared to AML. However, 6″-caffeoylhyperin and caffeoylquinic acid derivatives were equally significant for interaction with both AML and AGL.