Enhanced extraction and functional properties of 3-deoxyanthocyanidins from red sorghum leaf sheaths via ultrasound-assisted extraction with natural deep eutectic solvent Authors: Oscar Zannou, Mansuri M Tosif, Ayşe Erdoğan, Gulden Goksen, Reza Tahergorabi Published: 2026 ABSTRACT Graphical abstractUnlabelled ImageKeywords: Sorghum bicolor, 3-deoxyanthocyanidins, Bioaccessibility, Biological activity, Natural deep eutectic solvents, Thermal stability Introduction Sorghum bicolor (L.) Moench (sorghum) is commonly known as red sorghum and botanically belongs to the family Poaceae (Grass). During the vegetative growth stage, it exhibits high resistance to drought and elevated temperatures compared to other major cereal crops [1]. Grains of the red sorghum are a staple cereal crop consumed worldwide to prepare diverse dishes, especially in traditional communities. Different parts of the plant are used in folk medicine and as pigments for food, dyes, and livestock feed. In some traditional communities, the leaf sheath is processed into flour and then used in porridges, cakes, sauces, couscous, patties, tajine, bread, and cookies [2]. The leaf sheath of the red sorghum is also used to prepare hot or cold beverages such as tea and infusions. Although the grains of red sorghum are widely used, the other parts of the plant, like the leaf sheath, are not well valorized. The actual challenges related to health, economy, food security, and food safety open the door to the adoption of solutions that reduce food and plant waste production and valorize agricultural wastes. This supports the United Nations’ efforts, which have set the Sustainable Development Goals (SDGs), to address global challenges and advance sustainable development by 2030. The valorization of the leaf sheath of the red sorghum generates more income for the stakeholders, supporting SDGs 1, 2, and 8, which aim to end poverty in all its forms, mitigate waste, prevent hunger, and promote economic growth. The extraction of bioactive compounds from the leaf sheath of the red sorghum contributes to performing SDGs 3, 12, and 13, advocating the responsible consumption and production, preservation of consumers' health and protection of the environment. The leaf sheaths of red sorghum are a rich source of biocolorant 3-deoxyanthocyanidins. The 3-deoxyanthocyanidins are typical anthocyanidins that have a high deprotonation constant of their flavylium cation. In 3-deoxyanthocyanidins, the hydroxyl group is absent at the C-3 position of the C-ring, giving them the name 3-deoxy and making them more stable to pH changes than other anthocyanidins. The 3-deoxyanthocyanidins present great interest in the food industry thanks to their resistance to pH changes, bleaching additives, and heat treatment [2], [3], [4], [5], [6]. Likewise, they exhibit numerous health benefits, including antioxidant, anti-inflammatory, antimicrobial, antidiabetic, antihypertensive, and anticancer activities [7], [8], [9], [10], [11]. Apigeninidin (APG) and luteolinidin (LUT) have been identified as the major 3-deoxyanthocyanidins from the leaf sheaths of red sorghum [12], [13]. The extraction of 3-deoxyanthocyanidins from leaf sheaths of red sorghum has been extensively performed using traditional techniques, including heat-assisted extraction [3], [14] and stirring-assisted extraction [9], [15], [16], [17]. Santana et al. [18] successfully applied the pressurized liquid extraction to extract 3-deoxyanthocyanidins from the grains of red sorghum. Furthermore, only conventional solvents such as water, methanol, ethanol, acetone, and diethyl ether are used for the extraction [9], [17], [19]. Recently, Serrano et al. [19] investigated the extraction of 3-deoxyanthocyanidins from leaf sheaths of red sorghum using the heat-assisted extraction and ultrasound-assisted extraction (UAE) techniques and indicated that the UAE performed a higher extraction of 3-deoxyanthocyanidins than the heat-assisted extraction. However, the UAE conditions for the extraction of 3-deoxyanthocyanidins should be optimized to enhance cell disruption, mass transfer, and extraction yield, allowing better release of 3-deoxyanthocyanidins. Moreover,3-deoxyanthocyanidins are typically extracted from the red sorghum sheath using organic solvents [15], [17]. However, high quantities of the organic solvents have been revealed to be unsafe for the body and harmful to the environment [20]. Therefore, alternative green and sustainable solvents should be applied to efficiently extract 3-deoxyanthocyanidins from red sorghum. Natural deep eutectic solvents (NADESs) are considered green and sustainable alternative solvents, highly efficient in extracting plant bioactive compounds [21], [22]. NADESs are formed by a hydrogen bond donor (HBD) and a hydrogen bond acceptor (HBA). They have been extensively used to efficiently extract plant pigments such as anthocyanins [22], chlorophyll [23], carotenoids [24], and curcumin [25]. In the present study, the NADESs were used for the extraction of 3-deoxyanthocyanidins from the leaf sheath of red sorghum. This study investigated for the first time the 3-deoxyanthocyanidin extraction from sorghum leaf sheath using a dual-optimized NADES. Moreover, the extraction conditions implying UAE and NADESs were performed to produce a 3-deoxyanthocyanidin, namely APG and LUT-rich functional extract with high stability for food or pharmaceutical industry applications. In addition, the bioavailability, antioxidant, antidiabetic and antimicrobial activities of these 3-deoxyanthocyanidins were evaluated. Materials and methods Raw material and reagents Sun-dried leaf sheaths of red sorghum (S. bicolor) were ground, sieved through a 415-µm mesh and kept at −20 ˚C for further use. 2,2-diphenyl-1-picrylhydrazyl (DPPH) and 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) were purchased from Sigma Aldrich Chemical Co. (St Louis, MO, USA). Methanol and ethanol were bought from Honeywell (Indiana, USA). Acetic acid, lactic acid, and ethylene glycol were provided by Carlo Erba (Emmendingen. Germany). Glycerol (≥99.5%) was purchased from Kimyalab (Istanbul, Turkey). Choline chloride (≥ 98%) was provided by Aromel Kimya Medikal (Turkey). All salts for the simulated oral fluid (SOF), the simulated gastric fluid (SGF) and the simulated intestinal fluid (SIF) were prepared using analytical grade chemicals supplied by Sigma–Aldrich (Sigma Chemical Co. St. Louis, MO, USA). Bile extract (B8631, 031MO106V) and all other reagents were bought from Sigma–Aldrich (Sigma Chemical Co.). All enzyme activities were performed before starting the digestion experiments in accordance with the methodology outlined by Brodkorb et al. [26]. Methods Preparation of NADESs Binary NADESs were prepared with choline chloride (hydrogen bond acceptor) and lactic acid, acetic acid, glycerol, urea and ethylene glycol (hydrogen bond donors) at a 1:2 M ratio. Mixtures were heated at 80°C for 2 h. Choline chloride: acetic acid, choline chloride: glycerol, choline chloride: urea and choline chloride: ethylene glycol were inscribed as CHAC, CHGLY, CHUR and CHEGLY, respectively. Ultrasound-assisted extraction of 3-deoxyanthocyanidins The extraction was conducted following the protocol of Zannou et al. [27] with slight modifications. Briefly, 0.5 g of powder of sorghum leaf sheaths was mixed with 20 mL of NADESs as well as conventional solvents (distilled water, ethanol and methanol) and submitted to ultrasound-assisted extraction (Ultrasounic processor, Sonics and Material, Inc, Newtown, USA) at a frequency of 50 Hz, power of 750 W, temperature of 25 ˚C, amplitude of 40%, daylight, and extraction duration of 10 min. Determination of 3-deoxyanthocyanidins Luteolinidin and apigeninidin were estimated by the pH differential method using a UV–visible spectrophotometer (UV-5500PC, Shanghai Yuan Analysis Co., Ltd., China) [28], [29]. Briefly, a 50 µL aliquot of the extract was mixed with 2 mL of sodium acetate-acid buffer (pH 1), and another 50 µL aliquot of the extract was mixed with 2 mL of sodium acetate-acid buffer (pH 4.5). Afterward, the absorbances were recorded at 482 nm and 700 nm for luteolinidin, and at 468 and 700 nm for apigeninidin [28]. The total contents of APG and LUT were calculated in mg/g with the following equation: where A = (Absorbance468/482 nm − Absorbance700 nm) pH 1.0 − (Absorbance468/482 nm − Absorbance700 nm) pH 4.5; MW (molecular weight) = 255.24 and 271.24 g/mol for APG and LUT, respectively, DF = dilution factor established in D, l = pathlength in cm and ε= 30,400 and 31700 L /mol/cm for APG and LUT, respectively. Optimization of NADES components The NADES components, including hydrogen bond donor (acetic acid, mol) and water content, were optimized by applying the surface methodology based on a three-level central composite design (Design-Expert software 13.0). The design experiment included two factors such as X1 (CHAC, molar ratio) and X2 (water content, %) and the responses (Y) were APG and LUT. The actual and coded values of the factors are presented in Table 1. Table 1Actual and coded values of the independent variables for the central composite design.Coded valuesActual valuesX1X2−1.520.9810.09−11.5014.5002.7525.15+1435.80+1.524.5240.21X1: molar ratio; X2: water content (%). Actual and coded values of the independent variables for the central composite design. X1: molar ratio; X2: water content (%). The combination of the factors, such as CHAC (1:0.98, 1:1.5, 1:2.75 and 1:4.52 M ratios) and water content (10.09, 14.50, 25.15, 35.80 and 40.21%) was generated. The experimental points together with responses are shown in Table 2. Analysis of variance (ANOVA) was performed to justify the statistical significance of the regression model, and the p-value was employed to check the interaction effects of each variable. The second-order polynomial equation was determined for the response variables as follows: where, Y is the response variable, β0, βi, βii and βij represent the regression coefficients of intercept, linear, quadric and interaction, respectively. Xi and Xj are the independent variables. k is a variable number. Table 2The experimental points and responses for the central composite design.RunX1: Molar ratioX2: Water content (%)LUT (mg/g)APG mg/g)11.535.819.5815.3422.7525.1526.4127.5432.7510.0938.0041.2914414.5027.2731.3254.5225.1523.4023.3562.7525.1527.4426.147435.8028.4322.4781.514.5036.7238.4592.7525.1527.1126.43102.7540.2125.6419.24110.9825.1523.6924.26 The experimental points and responses for the central composite design. Optimization of extraction conditions The extraction parameters, such as time, amplitude and NADES content (mL), were optimized using response surface methodology (RSM) based on a three-level Box-Behnken design (Design Expert software 13.0). X1 (time, min), X2 (amplitude, %), and X3 (NADES content, mL) were independent variables, while APG and LUT were the responses (Y). The actual and coded values of the independent variables are shown in Table 3. Table 3Actual and coded values of independent variables for the Box-Behnken design.Coded valuesActual valuesX1X2X3−153010017.506025+1309040X1 (Extraction time); X2 (Amplitude, %) and X3 (NADES content, mL). Actual and coded values of independent variables for the Box-Behnken design. X1 (Extraction time); X2 (Amplitude, %) and X3 (NADES content, mL). The combinations of the time (5, 17.5 and 30 min), amplitude (30%, 60% and 90%) and NADES content (10, 25 and 40 mL) generated 15 experimental points as displayed in Table 4. The experimental data were set to the following quadratic polynomial model: Table 4The experimental points and responses for the Box-Behnken design.RunX1: Time (min)X2: Amplitude (%)X3: NADES content (mL)LUT (mg/g)APG (mg/g)15302540.6040.92217.5304020.3719.35317.5602528.1529.81417.5602528.0829.61517.5901020.702.5563030252.1531.23730902543.1652.38817.5904033.7730.179560408.4121.541030604027.2735.10113060109.297.20125601026.3810.971317.5602527.6027.911417.5301032.2710.4915590252.5528.94 The experimental points and responses for the Box-Behnken design. High-speed homogenate-assisted extraction (HSHAE) Powder of sorghum leaf sheaths (0.5 g) was mixed with 20 mL of NADES and placed in a high-speed homogenizer (HZ1, LABTron, Korea) for extraction for 29.50 min at room temperature under homogenization speed 15000 rpm. Stirring-assisted extraction (SAE) Powder of sorghum leaf sheaths (0.5 g) was mixed with 20 mL of solvents and stirred at 250 rpm for 29.50 min using a magnetic stirrer (Velp Scientifica, Italy). Determination of antioxidant activity The DPPH radical scavenging assay was implemented as detailed in Zannou et al. [29]. The DPPH radical scavenging activity was provided as mmol Trolox equivalent (TE)/g. The ABTS radical scavenging assay was implemented as detailed in Kong et al. [30]. The ABTS radical scavenging activity was provided as µmol Trolox equivalent (TE)/g. FRAP assay was carried out by the previous method described by Benzie & Strain [31]. The value of FRAP was obtained from a standard curve of FeSO4. The results were given as mmol FeSO4 equivalents per g (mmol ISE/g). Determination of in vitro antidiabetic activity The in vitro antidiabetic assay was performed using the alpha-amylase (α-amylase) inhibitory activity as described in Perumal et al. [32] with a slight modification. Briefly, a serial dilution of the crude extract ranging from 0.1 mg/mL to 40 mg/mL was prepared. 500 µL of the crude extract was mixed with 600 μL of 0.1 M PBS (pH 5), followed by 200 μL of 1 U/ml of α-amylase enzyme (from Aspergillus oryzae) and the assembly was stored at 37°C for 15 min. After this, 500 μL of 0.25% starch was added to the mixture and stored 37°C for 15 min. The enzyme-substrate reaction was stopped by the addition of 200 μL of 1 M hydrochloric acid and stored again in a hot water bath at 75°C for 5 min. Finally, the mixture was completed with 500 μL of iodine reagent and then the absorbance was read at 620 nm. DMSO was used as a negative control. To prepare the extract blanks, 500 μL of crude extract was mixed with 2000 μL DMSO. The percentage of inhibition of α-amylase enzyme was expressed by the following equation: Where: Abs (N) stands for the absorbance of negative control, while ΔAbs (E) sands for the difference of the absorbances of crude extract and blank of crude extract. Antimicrobial activity analysis The antimicrobial activity of the samples was evaluated against microorganisms, including Listeria monocytogenes, Pseudomonas aeruginosa, Staphylococcus aureus and Escherichia coli. Prior to testing, bacterial strains were subcultured on Tryptic Soy Agar (TSA), followed by incubation at 37°C for 24 h. After incubation, single isolated colonies were selected from the agar plates for further testing. For suspension preparation, an overnight culture of each microorganism was inoculated into Tryptic Soy Broth (TSB) at a 1:100 dilution ratio. Cultures were incubated at 37°C under shaking conditions until they reached the exponential growth phase. The optical density at 600 nm (OD600) was then adjusted to 0.1–0.2 to standardize the microbial concentration prior to use. These standardized suspensions were subsequently used for both the disc diffusion assay and the percentage bacterial growth inhibition experiments as described by Erdoğan Eliuz [33]. Disc Diffusion Method: The standardized microbial suspensions were uniformly spread onto Mueller-Hinton Agar (MHA) plates using the spread plate technique. Sterile filter paper discs (6 mm in diameter) were impregnated with 30 µL of each test sample and then carefully placed onto the surface of the inoculated agar plates. Sterile distilled water was used as the negative control, while antibiotic discs (ampicillin and flaconazole) were prepared for 125 µg/mL and served as the positive control. All plates were incubated at 37°C for 24 h. After incubation, the antimicrobial activity was assessed by measuring the diameters of the inhibition zones around each disc in millimeters. All tests were conducted in triplicate, and results were expressed as mean ± standard deviation. Percentage Bacterial Growth Inhibition Experiment: The antimicrobial activity of the samples was further evaluated using a microdilution-based growth inhibition assay in 96-well ELISA microplates. Standardized microbial suspensions (OD600 = 0.1–0.2) were prepared as described previously. A volume of 50 µL of each bacterial suspension was added to individual wells of a sterile, flat-bottom 96-well ELISA plate. Test samples were added to the wells at defined concentrations (50 µL), and the final volume in each well was adjusted to 200 µL using Tryptic Soy Broth (TSB) for bacteria. Negative control is a microbial suspension with broth only (no test sample). Microplates were covered with a sterile lid and incubated at 37°C for 24 h. After incubation, bacterial growth was quantified by measuring optical density at 600 nm (OD600) using a microplate spectrophotometer. The percentage of bacterial growth inhibition was calculated using the following formula: Where: •ODSample = Absorbance of the test extract ODSample = Absorbance of the test extract •ODControl = Absorbance of the negative control (bacteria without extract) ODControl = Absorbance of the negative control (bacteria without extract) Results were expressed as mean ± standard deviation of three independent measurements. Determination of in vitro bioaccessibility of 3-deoxyanthocyanidins The 3-deoxyanthocyanidin-rich extracts were subjected to in vitro digestion using the INFOGEST protocol developed by Brodkorb et al. [26] and [34]. The digestive process is comprised of the oral, gastric, and intestinal phases. In the oral phase, the extract was taken in a 50 mL centrifuge tube, followed by the addition of the simulated salivary fluid and a calculated amount of CaCl2 solution (0.3 M) according to the protocol. The tubers were placed in a rotator, placed in an incubator and stirred at 150 rpm for 2 min at 37°C. In the gastric phase, the simulated gastric fluid was combined with the oral digesta (1:1 v/v) and adjusted to pH 3 using HCl (1 M) before adding porcine pepsin (2000 U/mL). The tubes were then placed in a rotator in an incubator and stirred for 120 min at 37°C. Afterwards, in the intestinal phase, gastric digesta was dissolved in simulated intestinal fluid (1:1, v/v) and added to CaCl2 solution and bile salt (at a final concentration of 10 μM) and the pH was adjusted to 7.0 with NaOH (0.1 M). Then, pancreatin (100 U/mL) was mixed with the intestinal phase. The tubes were kept in a rotator in an incubator and stirred for 120 min at 37°C. The digestion was stopped with a specific enzyme inhibitor (Pefabloc®, Sigma Aldrich), and the samples were rapidly frozen using liquid nitrogen. The in vitro bioaccessibility in percentage was obtained using the following equation. Thermal stability test 5 mL of the aqueous or NADES extract was put in the sealed glass tubes and kept at 60, 80 and 100°C for 60 and 120 min. After each temperature and time set, the extract was immediately cooled and the contents of APG and LUT were determined The first-order kinetic model of degradation was calculated with the following formula: In the formula; C0 is the initial content (t = 0); C is the content at time t; t = time (minutes); k represents the degradation rate constant (min−1). The half-life (t1/2) was calculated with the degradation rate constant using the following equation: In these formulas, Ea is the activation energy (kJ/mol); R is the universal gas constant (8.314 J/mol. K); T is the absolute temperature (K). Thermodynamic parameters of degradation such as Gibbs free energy (ΔG) (kJ/mol), enthalpy (ΔH) and activation entropy (ΔS) (kJ.mol.K−1) were calculated using the degradation rate constant and activation energy (Ea). Statistical analysis All results were expressed as the mean of three replicates ± standard deviation. The one-way analysis of variance ANOVA was done with Duncan's multiple range test and the differences at p < 0.05 were set as statistically significant. Results and discussion Selection of solvents The ability of NADESs to extract pigments 3-deoxyanthocyanidins (APG and LUT) from the leaf sheath of red sorghum was investigated and compared with the conventional solvents. The NADESs included as CHAC, CHGLY, CHUR and CHEGLY, while the conventional solvents investigated were distilled water, methanol and ethanol. The extraction of pigments was performed via ultrasound-assisted extraction (UAE), and the results are displayed in Fig. 1. As can be seen, the type of solvent induced significant changes in the extraction of APG and LUT (p < 0.05). The contents of APG and LUT were found in ranges of 2.25 ± 0.01–31.59 ± 0.09 mg/g and 1.36 ± 0.01–30.44 ± 0.11 mg/g, respectively. A previous study reported a similar content of APG of 31.30 mg/g from leaf sheaths of red sorghum, while a lower amount of LUT of 1.20 mg/g has been determined [9]. Petti et al. [35] reported low values of APG and LUT from the leaf sheaths of red sorghum, being 0.421 mg/g and 1.768 mg/g, respectively. These variations in the contents of 3-deoxyanthocyanidins extracted from leaf sheaths of red sorghum are due to the origin of the raw material and extraction conditions. Recovery of APG and LUT by NADESs and conventional solvents. In the present study, the highest content of 3-deoxyanthocyanidins was determined with CHAC (APG = 31.59 ± 0.09 mg/g and LUT = 30.44 ± 0.11 mg/g), followed by methanol, ethanol, CHEGLY, CHGLY and CHUR, respectively. Similarly, in previous studies, acidified or aqueous methanolic extracts have been shown to contain a high amount of 3-deoxyanthocyanidins [9], [35]. However, in the present study, the highest amount of 3-deoxyanthocyanidins with CHAC proved that it has a strong affinity for APG and LUT. It has been reported that CHAC has a great ability in extracting pigments such as anthocyanins and betalains [36], [37]. The ability of CHAC to highly extract pigments is due to the orchestration of a strong intramolecular structure. Furthermore, the presence of multiple hydrogen bonding possibilities in CHAC generates a strong interaction with 3-deoxyanthocyanidins, allowing their extraction. Similarly, the phenomenon has been previously reported between NADESs and phenolic compounds as well as pigments [37], [38], [39]. Moreover, the low pH value of CHAC (pH ≈ 2) promoted the extraction of APG and LUT from the leaf sheaths of red sorghum. Therefore, CHAC was chosen for further investigation. Optimization of CHAC components Model fitting The molar ratio and water content in NADES are key factors that can influence the efficiency of NADES in extracting bioactive compounds [40], [41]. In the present study, the molar ratio and water content of CHAC were optimized in order to set a suitable composition of CHAC that facilitates a maximum extraction of APG and LUT from leaf sheaths of red sorghum. The data of the central composite design showed that the contents of APG and LUT varied as 15.34–41.29 mg/g and 19.58–38.00 mg/g, respectively. The lowest contents of APG and LUT were determined at run 1 (1:1.5 M ratio and 35.80% water content), while the highest values were determined at run 3 (1:2.75 M ratio and 10.09% water content). Since CHAC has a hydrophilic character and low viscosity [37], [41], it performed an efficient extraction of APG and LUT, requiring less addition of water. The ANOVA results for models provided a p < 0.0001 and a non-significant lack of fit (p > 0.05), with R2 and adjusted R2 around 0.99 (Table 5), suggesting that the experimental and predicted values correlated with a high degree and the model is significantly fitted. Moreover, the models displayed low coefficient variation percentages, ranging from 1.66% to 1.97%, indicating good reproducibility and a reliable model. Table 5ANOVA results of the optimization of NADES components.SourceLUTAPGSum of SquaresF-valuep-valueSum of SquaresF-valuep-valueModel294.04281.21< 0.0001591.52422.26< 0.0001X1-Molar ratio0.12870.61560,46820.20730.74000.4290X2-Water content140.09669.86< 0.0001498.511779.31< 0.0001X1X283.77400.55< 0.000150.86181.53< 0.0001X1215.2272.790,000412.4744.520.0011X2235.17168.17< 0.000117.2261.460,0005Residual1.051.40Lack of Fit0.48530.57740.68380.32080.19810.8904Pure Error0.56031.08Cor Total295.09592.92R20.99650.9976Adjusted R20.99290.9953Predicted R20.98400.9921Adequate precision55.3465.7288C.V. %1.661.97 ANOVA results of the optimization of NADES components. Effects of molar ratio and water content on APG and LUT The linear, interactive and quadratic terms had a significant effect on APG (p < 0.0001–0.0011), except the linear term of molar ratio (X1; p = 4.290) (Table 5). The regression coefficients developed for the APG model revealed that APG is more impacted by the interactive term X1X2 and the quadratic term X2X2 (Eq. (13). The linear, interactive and quadratic terms had a significant effect on LUT (p < 0.0001–0.0004), except the linear term of molar ratio (X1; p = 4.4682) (Table 5). The regression coefficients developed for the LUT model revealed that LUT is more impacted by the interactive term X1X2 and the quadratic term X2X2 (Eq. 14). As shown in Fig. 2, the combination of water content and molar ratio had significant effects on APG and LUT, which were promoted till 14.50% water content and when the molar ratio was increased continuously. The increase of the molar ratio provoked the availability of H-bond donors, which favored the formation of a stronger hydrogen network between CHAC and pigments. Similarly phenomenon has been reported in various published studies where the increase of the molar ratio has facilitated the dissolution of bioactive compounds in NADESs, promoting the extraction yield [42], [43], [44]. 3D Graphics of LUT and APG obtained from the optimization of NADES components. Validation of optimization of CHAC components The optimal conditions for CHAC components were determined by considering the desirability function, where the independent variables were held in range and the responses were maximized. The predicted optimum conditions for CHAC were found as 1:1.5 M ratio and 14.50% water content, giving 38.58 mg/g and 36.73 mg/g for APG and LUT, respectively. Further experiments were conducted under these conditions and the results were found as 39.15 ± 0.58 mg/g and 37.89 ± 0.86 mg/g for APG and LUT, respectively. These predicted and the experimental values of APG and LUT were found close, confirming the reliability of the optimization process. Our results are supported by a previous study, which has reported a highest of APG to be 45.8 mg/g [9]. Optimization of extraction conditions Analysis of the developed model The extraction conditions were optimized using the Box-Behnken design, considering factors such as extraction time, amplitude, and NADES content, to adequately recover maximum APG and LUT from the leaf sheaths of red sorghum. The results showed that the contents of APG and LUT changed as 2.55–52.38 mg/g and 2.15–43.16 mg/g, respectively. The lowest value of APG was detected at run 5 (extraction time of 17.5 min, amplitude of 90% and NADES content of 10 mL), whereas the greatest value was detected at run 7 (extraction time of 30 min, amplitude of 90% and NADES content of 25 mL). For the LUT, the lowest value was detected at run 6 (extraction time of 30 min, amplitude of 30% and NADES content of 25 mL) and the highest value was obtained at run 7 (extraction time of 30 min, amplitude of 90% and NADES content of 25 mL). These findings showed that the extraction conditions affected differently each of APG and LUT. Similarly, previous studies have reported that the factors influence the individual compounds of anthocyanins and betalains [36], [45]. The ANOVA results for models provided a p < 0.0001 and a non-significant lack of fit of 1.29, implying that the model is significant. Moreover, the R2 and adjusted R2 were above 0.98 (Table 6), suggesting that the experimental and predicted values correlated with a high degree and the model is significantly fitted. Table 6ANOVA results of the optimization of extraction conditions.SourceSum of SquaresF-valuep-valueSum of SquaresF-valuep-valueModel2283.982402.73< 0.00012501.05145.06< 0.0001X1-Time1.9318.290.007969.2536.150.0018X2-Amplitude2.8727.140.003418.159.470.0275X3-NADES content0.17571.660.2536702.20366.55< 0.0001X1X21562.3014791.74< 0.0001274.38143.23< 0.0001X1X3323.103059.08< 0.000175.0539.180.0015X2X3155.771474.85< 0.000187.9045.890.0011X12201.241905.32< 0.0001140.2373.200.0004X228.9284.430.000335.3918.470.0077X3227.38259.28< 0.00011013.64529.12< 0.0001Residual0.52819.58Lack of Fit0.34791.290.46537.382.240.3236Pure Error0.18022.20Cor Total2284.502510.62R20.99980.9962Adjusted R20.99940.9893Predicted R20.99740.9510Adequate precision153.466642.4853C.V. %1.395.49 ANOVA results of the optimization of extraction conditions. Effects of extractions on APG and LUT The linear, interactive and quadratic terms of the developed model showed a significant effect on APG (p < 0.0001–0.0077) (Table 6). The regression coefficient showed that APG is positively impacted by the terms X1, X2, X3, X1X2, X1X3, X2X3, X1X1 and X2X2 while the term X3X3 had a negative effect (Eq. 15). The linear, interactive and quadratic terms had a significant effect on LUT (p < 0.0001–0.0079), except the linear term of NADES content (X3; p = 0.2536) (Table 6). The regression coefficients of the model of LUT showed that the terms X1, X2, X3, X1X2, X1X3, X2X3 and X2X2 had a positive impact, while the terms X1X1 and X3X3 had a negative effect (Eq. (16). The interactive effects of the independent variables are displayed in Fig…As can be seen, the combination of extraction time and amplitude had significant effects on APG and LUT, with the amplitude showing a continuous effect while the extraction time reached maximum values around 10–15 min and then declined. The interactive reaction of extraction time and NADES content was positive and continuous on APG and LUT, even though the effect of NADES content was sharper (Fig. 3). Likewise, the combination of amplitude and NADES content continuously promoted the extraction of APG and LUT (Fig. 3). Our results are in agreement with the previous studies, which stated that the extraction time, ultrasound amplitude and NADES content significantly affect the extraction of bioactive compounds from various raw materials [46], [47], [48]. Interactive effects of the extraction parameters on LUT and APG. Validation of optimization of extraction conditions Based on the desirability function and by keeping the independent variables in range and maximizing the responses, the predicted optimum conditions were 29.93 min, 89.86% and 27.95 mL for extraction time, amplitude and NADES content, respectively. Under these conditions, the predicted responses were 53.87 mg/g and 45.74 mg/g for APG and LUT, respectively. For validation, further analyses were implemented and the results were found as 54.97 ± 0.96 mg/g and 48.62 ± 0.44 mg/g for APG and LUT, respectively. The predicted and the experimental values of APG and LUT were found to be similar, confirming the reliability of the optimization process. Comparison of extraction methods The optimum extraction conditions obtained with CHAC and UAE were compared with high-speed homogenate-assisted extraction (HSHAE) and stirring-assisted extraction (SAE) techniques in order to determine the most efficient extraction systems for the extraction of APG and LUT. The results of the different extraction techniques are shown in Fig. 4. As can be observed, the application of different extraction methods had a significant effect on the extraction of APG and LUT (p < 0.05) (Fig. 4). The contents of APG and LUT were found in 54.97 ± 0.96 mg/g and 48.62 ± 0.44 mg/g for UAE, 50.53 ± 0.41 mg/g and 43.80 ± 0.35 mg/g for HSHAE, and 47.52 ± 0.48 mg/g and 41.65 ± 0.59 mg/g, respectively. For both APG and LUT, the UAE performed the highest extraction of 3-deoxyanthocyanidins, followed by HSHAE and SAE, respectively. Similarly, a previous study found that the HSHAE and UAE combined with NADESs are prominent for the extraction of bioactive compounds from olive pomace [49]. Furthermore, the UAE and homogenate-assisted extraction have shown high extraction efficiency of phenolic compounds from kinkeliba [50] and anthocyanins from sumac [51]. The increase of the extraction performance by the application of UAE is associated with the acoustic cavitation phenomenon and the bubbles generated, which cause the breakdown of the cell walls [49], [51]. Moreover, the occurrence of cavitation creates a swelling and enlargement of pores in the sample, leading to an easy solvent penetration and a high extraction efficiency [52]. The HSHAE is sustained by a rotating inner cutter driven through a high-speed motor, which generates a powerful shear force between the inner and external cutters, causing the disruption of the cellular walls and mixing the solid material with the solvent [49]. This induces a mass transfer between the solute and the solvent, increasing the extraction efficiency. Impacts of extraction techniques on the recovery of APG and LUT. Antioxidant activity of 3-deoxyanthocyanidin-rich extract The antioxidant activity of the 3-deoxyanthocyanidin-rich extracts obtained by UAE, HSHAE and SAE was determined using the ABTS, DPPH and FRAP assays (Table 7A). As can be observed, the extraction techniques induced a significant difference in the antioxidant activity of the 3-deoxyanthocyanidin-rich extracts of leaf sheaths of red sorghum (p < 0.05). The UAE was determined as the most efficient, providing 1117.07 ± 21.92 µmol TE/g for ABTS assay, 837.36 ± 16.60 mmol TE/g for DPPH assay and 5151.03 ± 39.44 mmol ISE/g for FRAP assay. The UAE was followed by HSHAE, giving 892.36 ± 16.44 µmol TE/g for ABTS assay, 829.06 ± 8.30 mmol TE/g for DPPH assay and 4667.87 ± 29.58 mmol TE/g for FRAP assay. The extract obtained from the SAE showed the lowest antioxidant activity, exhibiting 744.38 ± 10.96 µmol TE/g for ABTS assay, 746.07 ± 8.30 mmol TE/g for DPPH assay and 4658.01 ± 19.72 mmol ISE/g for FRAP assay. In contrast to our results, previous studies have reported low antioxidant activity for the alcoholic or aqueous extracts of leaf sheaths of red sorghum [9], [53]. Kayodé et al. [9] determined that the antioxidant activity of leaf sheaths of red sorghum was found to be 3.76–5.58 mmol/g with ABTS assay and 0.35–1.06 mmol/g with FRAP assay. Tugli et al. [53] found an antioxidant activity of leaf sheaths of red sorghum ranging from 0.957 mmol/g to 2.452 mmol/g using the FRAP assay. The highest values found in the present can be linked to the use of NADES, which enabled a high extraction of not only APG and LUT but also other antioxidant compounds. Moreover, the application of the emerging extraction techniques like UAE, HSHAE and SAE favored the extraction of pigments from the leaf sheaths of red sorghum and then increased the antioxidant activity. These findings are supported by previously published studies, which concluded that the NADESs combined with emerging extraction techniques increased the extraction of antioxidant compounds [37], [54], [55], [56]. Table 7Antioxidant activities of 3-deoxyanthocyanidin-rich extract (A), Inhibition zone diameter (mm) of the extract samples for different bacterial strains (B) and In vitro bioaccessibility of LUT and APG extracted from leaf sheaths of red sorghum (B).Table 7AExtraction methodsAntioxidant activitiesABTS (µmol TE/g)DPPH (mmol TE/g)FRAP (mmol ISE/g)UAE1117.07 ± 21.92a837.36 ± 16.60a5151,03 ± 39.44aHSHAE892.36 ± 16.44b829.06 ± 8.30a4667.87 ± 29.58bSAE744.38 ± 10.96c746.07 ± 8.30c4658.01 ± 19.72bSame letters in the column mean no statistical difference p < 0.05.Table 7BBacteriaInhibition zone diameter (mm)CHAC extractWater extractL. monocytogenes45.55±1.48a28.23±2.96cS. aureus46.01±1.10a22.94±1.98dP. aeruginosa43.82±2.06ab18.64±1.61eE. coli42.00±1.53b15.66±1.25eSame letters mean no statistical difference p < 0.05.Table 7CSolventsPhasesLUTAPGContent (mg/g)Bioaccessibility (%)Content (mg/g)Bioaccessibility (%)Distilled waterOral0.46±0.02d5.79±0.22d0.61±0.01d7.61±0.11dGastric0.30±0.00d3.78±0.00e0.37±0.01de4.61±0.11eIntestine0.16±0.00e2.00±0.00f0.15±0.01f1.82±0.11fCHACOral37.35±0.08ba76.83±0.16a38.56±0.04a70.15±0.08aGastric17.89±0.05b36.80±0.11b17.80±0.15b32.38±0.28bIntestinal17.34±0.22c35.66±0.45c17.26±0.03c31.41±0.06cSame letters in the column mean no statistical difference p < 0.05. Antioxidant activities of 3-deoxyanthocyanidin-rich extract (A), Inhibition zone diameter (mm) of the extract samples for different bacterial strains (B) and In vitro bioaccessibility of LUT and APG extracted from leaf sheaths of red sorghum (B). Same letters in the column mean no statistical difference p < 0.05. Antimicrobial activity of 3-deoxyanthocyanidin-rich extract The antimicrobial activity of 3-deoxyanthocyanin-rich extract obtained from the red sorghum leaf sheaths using CHAC and UAE was evaluated and compared with the aqueous extract obtained in the same conditions. The antibacterial effect of 3-deoxyanthocyanin-rich extracts was performed on four bacteria of interest in the food industry, including L. monocytogenes (Gram + ), P. aeruginosa (Gram −), S. aureus (Gram + ) and E. coli (Gram −) (Table 7B and Fig. 5). The most potent antimicrobial activity was observed for CHAC extract against L. monocytogenes, S. aureus, P. aeruginosa and E. coli, exhibiting the inhibition zone diameter of 45.55 ± 1.48, 46.01 ± 1.10, 43.82 ± 2.06 and 42.00 ± 1.53 mm, respectively. Whereas the water extract showed the lowest antimicrobial activity, exhibiting an inhibition zone diameter of 28.23 ± 2.96, 22.94 ± 1.88, 18.64 ± 1.61 and 15.66 ± 1.25 mm for L. monocytogenes, S. aureus, P. aeruginosa and E. coli, respectively. These findings revealed that the 3-deoxyanthocyanin-rich extracts showed antibacterial activity against the tested pathogenic microorganisms, with the NADES extract exhibiting the strongest antibacterial activity. Recently, has been reported that NADESs could not only fine-tune solvent physicochemical characteristics but also improve the antimicrobial activity of dissolved active compounds [57], [58], [59]. In the present study, the percentage of inhibition of 3-deoxyanthocyanin-rich extracts was summarized in Fig. 5. As can be seen, CHAC extract displayed the highest percentage of inhibition on L. monocytogenes, S. aureus, P. aeruginosa and E. coli, representing 94.98 ± 1.29%, 93.88 ± 2.63%, 85.01 ± 3.29% and 81.80 ± 2.73%, respectively. In contrast, the water extract showed a lower percentage of inhibition for all the tested microorganisms compared with the NADES extract. These results suggested that 3-deoxyanthocyanin-rich extracts obtained from the leaf sheaths of red sorghum, especially NADES extract, exerted a strong antimicrobial effect against L. monocytogenes, S. aureus, P. aeruginosa and E. coli. Effect of APG and LUT-rich extract on different bacterial strains growth inhibition. Bioaccessibility of APG and LUT extracted from leaf sheaths of red sorghum The bioaccessibility of the APG and LUT contained in the CHAC extract obtained at the optimum conditions was assessed and the results were compared with the aqueous extract obtained in the same conditions. The bioaccessibility of APG and LUT was conducted through the digestive tract, considering the oral, gastric, and intestinal phases, and the results are summarized in Table 7C. In the water, the concentration of LUT decreased from 7.88 ± 0.13 mg/g to 0.46 ± 0.02 mg/g in the oral phase, to 0.30 ± 0.00 in the gastric phase, and to 0.16 ± 0.00 mg/g in the intestinal phase, while the APG content dropped from 8.03 ± 0.13 mg/g to 0.61 ± 0.01 mg/g in the oral phase, to 0.37 ± 0.01 mg/g and to 0.15 ± 0.01 mg/g in the intestinal phase. For the CHAC extract, the concentration of LUT reduced from 48.62 ± 0.44 mg/g to 37.35 ± 0.08 mg/g in the oral phase, to 17.89 ± 0.05 mg/g in the gastric phase and to 17.34 ± 0.22 mg/g in the intestinal phase, whereas the concentration of APG decreased from 54.97 ± 0.96 mg/g to 38.56 ± 0.04 mg/g in the oral phase, to 17.80 ± 0.15 mg/g in the gastric phase and to 17.26 ± 0.03 mg/g in the intestinal phase. As can be evidenced, both the type of solvent and the digestive environment affected the retention of LUT and APG. The concentrations of LUT and APG dropped during the in vitro bioaccessibility test from one digestive phase to another for both water and CHAC. Nonetheless, it was noticed that the highest amounts of LUT and APG were obtained with CHAC throughout the digestive tract. This is due to the strong molecular network formed between CHAC, resulting in a resistant and protective effect throughout the digestive tract. Our results are in accordance with the previous studies, which reported that the bioactive compounds, including pigments (betalains, carotenoids and anthocyanins) and phenolic compounds obtained with NADESs, have been mostly preserved throughout the digestive tract [2], [36], [60], [61]. At the oral phase, the bioaccessibility of LUT and APG was 5.79 ± 0.22% and 7.61 ± 0.11% in the aqueous extract, while 76.83 ± 0.16% and 70.15 ± 0.08% in the CHAC extract (Table 7C). As evidenced, at the oral step, the bioaccessibility of LUT and APG was significantly affected by the type of solvent, with the CHAC providing the highest values. Zannou, et al. [11] evaluated the bioaccessibility of the anthocyanins obtained from the NADES extract and reported 13.17–21.04% for cyanidin-3-glucoside, 22.02–28.81% for cyanidin-3-rutinoside, 16.37–18.51% for pelargonidin-3-glucoside and 25.84–32.27% for cyanidin chloride in the mouth. From these results, it can be affirmed that at the oral phase, LUT and APG were more bioaccessible than the common anthocyanins (cyanidin-3-glucoside, cyanidin-3-rutinoside, pelargonidin-3-glucoside and cyanidin chloride). At the gastric phase, the bioaccessibility of LUT and APG in the aqueous extract was 3.78 ± 0.00% and 4.61 ± 0.11%, whereas 36.80 ± 0.11% and 32.38 ± 0.28% in the CHAC extract. Similar to the oral phase, the bioaccessibility of LUT and APG at the gastric environment was significantly affected by the type of solvent, with the CHAC yielding the highest bioaccessibility. Similar to the finding of the present study, in a previous study, the bioavailability of anthocyanins obtained from NADES was determined as 24.95––33.54% for cyanidin-3-glucoside, 29.17–33.54% for cyanidin-3-rutinoside, 10.71–26.89% for pelargonidin-3-glucoside and 34.16–42.05% for cyanidin chloride [11]. Finally, the bioaccessibility of LUT and APG at the intestinal environment was determined as 2.00 ± 0.00% and 1.82 ± 0.11% in the aqueous extract and 35.66 ± 0.45% and 31.41 ± 0.06% in the CHAC. CHAC provided the highest bioaccessibility of LUT and APG compared with the aqueous extract. Similar to our findings, a bioaccessibility of 26.06–33.56% for cyanidin-3-rutinoside extracted with NADES from blackberry fruit at the intestinal environment, while lower values of 26.06–33.56% and 20.63–25.96% for cyanidin-3-rutinoside and cyanidin chloride, respectively [13]. Moreover, the bioaccessibility of the anthocyanins extracted from Jussara pulp using NADES has been reported to be 30.80% in the intestinal phase [61]. The bioaccessibility of LUT and APG extracted from leaf sheaths of red sorghum was assessed for the first time, and when comparing our results with those of the previously published data on anthocyanins, it can be assumed that the bioaccessibility of LUT and APG obtained using NADES was high and acceptable. In the present study, the bioaccessibility values of 3-deoxyanthocyanidins were found to be higher in CHAC than in distilled water, indicating that the 3-deoxyanthocyanidins of sorghum leaf sheaths are more stable and bioaccessible in NADESs. Supportively, previous studies reported that the NADES media offer a strong stabilization and high bioaccessibility pigments [13], [27]. Furthermore, the highest solubility of 3-deoxyanthocyanidins in NADES might contribute to the increased bioaccessibility of 3-deoxyanthocyanidins. Likewise, NADESs form strong hydrogen bonds with biomolecules, resulting in increasing the bioaccessibility of these compounds by reducing their degradation [27], [62]. Antidiabetic activity of 3-deoxyanthocyanidin-rich extract The 3-deoxyanthocyanin-rich extract obtained from the leaf sheaths of red sorghum using CHAC and UAE was evaluated for its activity against alpha-amylase, an enzyme affecting diabetes. The result of the NADES extract was compared with the aqueous extract. The results of enzymatic inhibitory activity against enzymes affecting diabetes (alpha-amylase) showed that the enzyme inhibitory activity of the aqueous extract was significantly (p < 0.05) lower than that of the NADES extract. The IC50 values of NADES and aqueous extracts were determined as 1.15 ± 0.05 mg/mL and 27.90 ± 0.10 mg/mL, respectively. Similar to our findings, Olfat et al. [46] reported that the alpha-amylase inhibitory activity of the NADES extract was significantly lower than the extracts obtained by conventional solvents, providing the IC50 values of 1.21, 1.46, 1.35 and 1.62 mg/mL for NADES, 80% methanol, 80% ethanol and distilled water, respectively. Likewise, Zengin et al. [63] determined that NADES extracts of C. hypocistis have shown a higher alpha-amylase inhibitory activity compared with extracts obtained from hexane, ethyl acetate, dichloromethane, ethanol, watery ethanol and water. The high efficiency of the NADES-based 3-deoxyanthocynidin-rich extract in inhibiting the alpha-amylase activity is due to the NADES' ability to highly extract LUT and APG. Moreover, the Pearson correlation results showed an extremely strong and statistically significant relationship (r = − 1.000**; p < 0.001 (2-tailed)) between the variables total 3-deoxyanthocyanidins and α-amylase inhibitory activity. Stability of 3-deoxyanthocyanidins during heating Thermal degradation kinetics of 3-deoxyanthocyanidins The thermal stability test of LUT and APG extracted from the leaf sheaths of red sorghum was carried out at 60, 80 and 100°C for 1 h and the results of the degradation kinetics are shown in Fig. 6. The thermal degradation rate of LUT and APG increased with the increment of the temperature during heating. As can be evidenced, the degradation of LUT and APG was faster and more extensive in water, while it was slow in the NADES under the same conditions. The degradation of LUT and APG in the watery extract was at 60°C, but it was found to be acute at 80°C and 100°C. Nonetheless, the degradation of LUT and APG in CHAC extract was very slow during the heating process. The LUT obtained from the aqueous extract showed a degradation percentage of 29% at 60°C, 78% at 80°C and 85% at 100°C, while its degradation percentage in CHAC was found to be 1% at 60°C, 10% at 80°C and 14% at 100°C. APG showed the same tendencies, giving a degradation percentage of 23% at 60°C, 49% at 80°C and 84% at 100°C with aqueous extract, while it was found to be 2% at 60°C, 18% at 80°C and 19% at 100°C in the CHAC extract. These findings revealed that the CHAC showed a high protective effect on the LUT and APG. Similar results have been reported for the common anthocyanins such as cyanidin-3-glucoside, cyanidin-3-rutinoside, pelargonidin-3-glucoside and cyanidin chloride, where NADESs have shown more protective effects under heat and light exposure [13], [41]. It has been reported that the degradation of the common anthocyanins is accentuated at 70-100°C, even in NADESs [13], [41], [64], [65]; however, the results of the present study showed that the increase of temperature had a very slight impact on LUT and APG obtained with NADES. This suggested that the 3-deoxyanthocyanidins are more stable than the common anthocyanins, especially in NADES. The NADES form extensive hydrogen-bonding interactions with LUT and APG tha stabilizes their molecular structure and reduces susceptibility to thermal cleavage. Furthermore, NADES provides a less polar or structured environment than water, protecting flavylium cation forms (important for anthocyanidin stability) and preventing hydration and ring opening. Degradation kinetics of LUT and APG during the heating at 60, 80 and 100°C for 1 h. Half-life (t1/2) and first-order reaction constant (k) of 3-deoxyanthocyanidins The t1/2 values for LUT were determined as 274.75 ± 1.48–2891.66 ± 39.72 min in CHAC and 21.90 ± 1.13–122.87 ± 2.32 min in water (Table 8). For APG, the t1/2 values were found to be 204.07 ± 2.76–2002.08 ± 6.16 min in CHAC and 22.91 ± 1.84–122.87 ± 2.32 min in water. The t1/2 values of LUT and APG decreased as the temperature increased. However, their degradation was slow, especially in NADES. The k value indicates the degree of degradation of 3-deoxyanthocyanins. A lower value of k suggests a high degree of stability of LUT and APG. In the present study, results of k values were determined as 0.24·10-3 ± 3.36·10-5-2.53·10-3 ± 1.37·10-4 for LUT and 0.35·10-3 ± 1.1·10-5-3.46·10-3 ± 0.47·10-3 for APG in NADES, while 5.64·10-3 ± 10.67·10-5-31.73·10-3 ± 1.63·10-3 for LUT and 4.46·10-3 ± 5.70·10-5-30.46·10-3 ± 2.45·10-3 in water. These data showed that k values increased with increasing temperature, indicating that 3-deoxyanthocyanins were affected when heated at 100°C for 1 h, even though they were less affected in NADES. In the previous studies, different NADESs have shown more protective effects on pigments during the heating process [13], [41], [66]. Table 8Thermodynamic properties of LUT and APG extracted with CHAC and water.PigmentsTemperature, (°C)t1/2, minkΔG0, kJ.mol−1ΔH0, kJΔS0, kJ.mol.K−1Ea, kJ.mol−1CHACWaterCHACWaterCHACWaterCHACWaterCHACWaterCHACWaterLUT602891.66 ± 39.72a122.87 ± 2.32 fg0.24·10-3 ± 3.36·10-5e5.64·10-3 ± 10.67·10-5d104.96 ± 0.38c96.24 ± 0.05 h57.84±4.97a41.82±1.82b−0.14±0.01a−0.16±0.01a60.61±4.97a44.59±1.82b80417.37 ± 5.66c27.66 ± 3.79 h1.69·10-3 ± 22.96·10-5de25.53·10-3 ± 3.49·10-3b105.75 ± 0.40c97.79 ± 0.40 g57.67±4.97a41.66 ± 1.82b−0.14±0.02a−0.16 ± 0.01a100274.75 ± 1.48de21.90 ± 1.13 h2.53·10-3 ± 1.37·10-4de31.73·10-3 ± 1.63·10-3a110.64 ± 0.17a102.79 ± 0.16e57.51±4.97a41.49 ± 1.82b−0.14±0.01a−0.16±0.01aAPG602002.08 ± 6.16b155.61 ± 2.01f0.35·10-3 ± 1.1·10-5e4.46·10-3 ± 5.70·10-5de103.97 ± 0.09d96.90 ± 0.04 h56.46±4.32a46.82±2.42ab−0.14±0.01a−0.15±0.01a59.23±4.32a49.59 ± 2.42ab80328.29 ± 1.93d63.03 ± 0.58gh2.12·10-3 ± 0.13·10-3de11.09·10-3 ± 1.03·10-3c105.07 ± 0.18c100.22 ± 0.27f56.29±4.32a46.65±2.42ab−0.14±0.01a−0.15 ± 0.01a100204.07 ± 2.76ef22.91 ± 1.84 h3.46·10-3 ± 0.47·10-3de30.46·10-3 ± 2.45·10-3a109.69 ± 0.42b102.93 ± 0.25e56.13±4.32a46.48±2.42ab−0.14±0.01a−0.15 ± 0.01aSame letters in the columns of t1/2, k, ΔG0, ΔH0, ΔS0 and Ea mean no statistical difference p < 0.05. Thermodynamic properties of LUT and APG extracted with CHAC and water. Same letters in the columns of t1/2, k, ΔG0, ΔH0, ΔS0 and Ea mean no statistical difference p < 0.05. Activation energy (Ea) and Gibbs free energy change (ΔG°) In the present study, the activation energy (Ea) is associated with the minimum amount of energy that 3-deoxyanthocyanidins must absorb to undergo breakdown due to heat. The higher Ea, the higher amount of energy is required to break down the compounds [41], [67], and then the reaction is more resistant to the heat [68]. In this study, the Ea for LUT and APG were determined to be 60.61 ± 4.97 and 59.23 ± 4.32 kJ.mol−1 in NADES and 44.59 ± 1.82 and 49.59 ± 2.42 kJ.mol−1 in water, respectively (Table 8). These results revealed that LUT and APG required a high amount of energy to initiate the thermal degradation. The higher Ea for the NADES extract than for the aqueous extract suggested that a larger amount of activation energy would be required for the NADES extract than for the aqueous extract. Similar to our findings, Zannou, Koca, et al. [41] reported that the aqueous extract of anthocyanins showed a lower Ea value while choline chloride-glycerol-based NADES extract provided a higher Ea value, indicating that NADES showed a stabilizing effect on anthocyanins. The Gibbs free energy (ΔG) of the thermal stability reaction indicates whether the reaction is spontaneous or not. The ΔG° with a negative value expresses that the reaction takes place spontaneously. In the present study, LUT presented the ΔG° values of 104.96 ± 0.38–110.64 ± 0.17 kJ.mol−1 in NADES and 96.24 ± 0.05–102.79 ± 0.16 kJ.mol−1 in water (Table 8). Similarly, APG presented the ΔG° values of 103.97 ± 0.09–109.69 ± 0.42 kJ.mol−1 in NADES and 96.90 ± 0.04–102.93 ± 0.25 kJ.mol−1 (Table 8). For both NADES and water, the ΔG° values obtained for LUT and APG were positive and high, with the highest values determined with NADES. Thus, it can be concluded that the thermal reactions for LUT and APG were not spontaneous. These results are in agreement the previous studies, which have reported that the thermal degradation of anthocyanins is a non-spontaneous reaction [41], [69]. Enthalpy change (ΔH°) and entropy change (ΔS°) The Enthalpy (ΔH°) represents the total heat content of a system and is expressed as an endothermic (positive value) or exothermic (negative value) reaction. The ΔH° values for LUT were 57.51 ± 4.97–57.84 ± 4.97 kJ in NADES and 41.49 ± 1.82–41.82 ± 1.82 kJ in water (Table 8). APG displayed similar ΔH° values, being 56.13 ± 4.32–56.46 ± 4.32 in NADES and 46.82 ± 2.42–46.48 ± 2.42 kJ in water (Table 8). As can be noticed, both NADES and water provided positive enthalpy values, suggesting that these thermal reactions were endothermic. These results are supported by the previous studies, which affirmed that the anthocyanin degradation is an endothermic reaction [41], [69], [70]. Moreover, NADES gave higher enthalpy values compared with water, indicating that the LUT and APG obtained from NADES required a greater energy absorption before being degraded during the heating process. The activation entropy (ΔS°) change measures the disorder change of the molecules in the system. In the present study, negative entropy values were found for LUT and APG in NADES or water (Table 8). This suggests that the transition state has lower structural freedom than the reactants, and also confirms that it is an irreversible process [69], [70]. Our results are in agreement with those reported previously for the individual anthocyanins extracted from NADES or water, where negative values of entropy were determined [41]. Conclusion 3-deoxyanthocyanidins, namely APG and LUT were extracted from the leaf sheaths of red sorghum using a greener strategy which combined a multiple optimization of UAE and NADESs, including choline chloride: glycerol (CHGLY), choline chloride: ethylene glycol (CHEGLY), choline chloride: lactic acid (CHLA) and choline chloride: acetic acid (CHAC). In addition, the extraction with the conventional solvents, including distilled water, methanol and ethanol, was investigated for comparison. The optimization of the prominent NADES (CHAC) component using the central composite design increased APG and LUT, being 39.15 ± 0.58 mg/g for APG and 37.89 ± 0.86 mg/g for LUT at the optimum conditions of 1:1.5 M ratio and 14.50% water content. The optimization of the extraction conditions using the Box-Behnken design promoted the extraction of APG and LUT, being 54.97 ± 0.96 mg/g for APG and 48.62 ± 0.44 mg/g for LUT at the optimum conditions of 29.93 min, 89.86% and 27.95 mL for extraction time, amplitude and NADES content, respectively. The antioxidant activity of 3-deoxyanthocyanidin-rich extract was found to be 1117.07 ± 21.92 µmol TE/g, 837.36 ± 16.60 mmol TE/g and 5151.03 ± 39.44 mmol ISE/g, for ABTS, DPPH and FRAP assays, respectively. 3-deoxyanthocyanidin-rich extract obtained from CHAC exhibited the highest in vitro bioaccessibility throughout the digestive tract compared with the aqueous extract. Moreover, the greatest antidiabetic and antimicrobial activities were determined with CHAC compared with the aqueous extract. In addition, CHAC showed good thermal properties, presenting the best values of t1/2, Ea, ΔG°, ΔH° and ΔS°. This study showed that NADESs, especially CHAC, are prominent green solvents for an efficient extraction of biomolecules, especially 3-deoxyanthocyanidin. Further study is needed to investigate the effective application of the NADES extract of 3-deoxyanthocyanidins to food products. Moreover, future work should focus on DES recovery, purification strategies, or formulation approaches. CRediT authorship contribution statement Oscar Zannou: Writing – original draft, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Mansuri M. Tosif: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Elif Ayşe Erdoğan Eliuz: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis. Gulden Goksen: Writing – review & editing, Visualization, Validation, Supervision, Software, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Reza Tahergorabi: Writing – review & editing, Visualization, Validation, Project administration, Methodology, Funding acquisition. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Dr. Oscar Zannou thanks to UNESCO-TWAS programme, “Seed Grant for African Principal Investigators” financed by the German Ministry of Education and Research (BMBF). Contributor Information Oscar Zannou, Email: zannouoscar@gmail.com. Gulden Goksen, Email: guldengoksen@tarsus.edu.tr. Data availability Data will be made available on request. References Associated Data Data Availability Statement Data will be made available on request.