Application of Response Surface Methodology for the Extraction of Phytochemicals from Upcycled Kale (Brassica oleracea var. acephala)

Abstract

Kale (Brassica oleracea) is recognized as a ‘superfood’ among leafy vegetables due to its high carotenoid content and potential health benefits. This study aims to optimize ultrasound-assisted extraction (UAE) to enhance the recovery of carotenoids and other phytochemicals from upcycled kale using response surface methodology. The optimized extraction parameters for carotenoids, i.e., aqueous ethanol as solvent, temperature, and extraction time at a fixed solid-to-solvent ratio, were established using the central composite design. The optimized extraction method was compared with other reported extraction methods for total phenolic content (TPC) and total antioxidant capacity (ferric reducing antioxidant power (FRAP) and 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging). Ultra-high-performance liquid chromatography–electrospray ionization–mass spectrometry (UPLC-ESI-MS) analysis was also performed. Under the optimized UAE conditions of 100% ethanol at 57 °C for 30 min, total carotenoid content of 392 µg/g dry weight (DW) was observed, though the predicted carotenoid content was 550 µg/g DW. Under these conditions, TPC, FRAP, and DPPH were 10.5 mg gallic acid equivalents/g DW, 13.9 µmole Trolox equivalence/g DW, and DPPH radical scavenging activity as IC₅₀ of 2.04 ± 0.31 mg/mL, respectively. The UPLC-ESI-MS analysis showed the highest total phytochemicals recovered through microwave-assisted extraction, followed by UAE, compared to other tested extraction methods. In conclusion, the established optimized UAE process significantly enhanced the yield and quality of recovered phytochemicals from upcycled kale.

1. Introduction

Brassica oleracea var. acephala, commonly known as kale, is a green leafy vegetable in the family of Brassicaceae. It is often considered a “superfood” due to its nutritional value and health benefits [1]. Kale originated from the eastern Mediterranean and Asia Minor regions and is widely cultivated in North America and central and northern Europe [2]. The production of kale has experienced significant growth in North America, particularly in the United States. The cultivation of kale has increased from 2012 to 2022 by 2500 acres to 18,121 acres [3]. The cultivation in Canada is at 752 acres in 2022 [4]. The expansion of kale production has increased demand and potential uses of its food waste.

Globally, 2.5 billion tons of food were produced in 2021, with North America (USA and Canada) accounting for roughly 6.72% of annual food waste [5]. Therefore, there is an unmet need for effective strategies for food waste management. One of the innovative strategies is ‘upcycling food’, which provides additional value for generating ingredients from food waste. Upcycled foods are innovative products that repurpose ingredients that include surplus produce and by-products from production that are not used for human consumption. These foods are sourced and manufactured through verified supply chains, contributing to a sustainable environment [6]. A food product can be classified as upcycled if it meets the following criteria: (a) comprises ingredients that are waste, (b) is transformed into a consumable product with less energy consumption, and (c) has a process that improves the value of the product [7].

Among other crops in the Brassica family, kale contains bioactive compounds, namely dietary carotenoids, glucosinolates, flavonoids, and phenolic acids. These phytochemical compounds contribute to multiple pharmacological activities, which include anti-bacterial, anti-inflammatory, antioxidant, cancer, and cardiovascular disease prevention properties [8]. Kale is also among the major food waste due to supply chain inefficiencies, short shelf life, and imperfect appearance [9]. Efforts are being made to reduce food waste by upcycling and using them as a valuable source of natural bioactive compounds [10]. Efficient green extraction allows these compounds to recover upcycled waste and use them in many value-added applications [11].

The traditional extraction methods, such as percolation, Soxhlet, maceration, and hydro-distillation, degrade, ionize, and oxidize the targeted bioactive compounds due to the application of toxic solvents, high energy, and longer extraction time [12]. However, non-conventional green extraction methods such as ultrasound-assisted extraction (UAE), microwave-assisted extraction (MAE), pressurized liquid extraction, and supercritical fluid extraction are considered to reduce extraction time, be cost-effective, and have higher extraction efficiency [11]. Amongst extraction methods, UAE provides a high extraction yield within a shorter extraction time [13]. Bioactive recovery using UAE is influenced by the extraction solvent, extraction temperature, solid-to-solvent ratio, power, and extraction time [14].

To our knowledge, no reports exist describing the optimization of extraction of carotenoids from kale using response surface methodology (RSM) and central composite design (CCD). Therefore, this study aims to determine the optimum extraction parameters to recover bioactive compounds from upcycled kale using food-grade ethanol through the application of RSM and CCD. Moreover, a comparison of the optimized extraction method to reported extraction methods and solvents was performed.

2. Materials and Methods

2.1. Reagents and Chemicals

Acetone, acetate buffer, acetonitrile, β-carotene, chlorogenic acid, diethyl ether, 2,2-diphenyl-1-picrylhydrazyl (DPPH), anhydrous ethanol, ferrous tripyridyl triazine, ferulic acid, Folin–Ciocalteu reagent, formic acid, gallic acid, hexane, hydrogen chloride, iron (III) chloride hexahydrate, kaempferol, methanol, sodium carbonate, sodium chloride, Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid), deionized water, and quercetin were purchased from Sigma-Aldrich, Oakville, ON, Canada. The other chemicals and their suppliers were lutein (BOC Sciences, New York, NY, USA), DL-menthol (Thermo Fisher Scientific, Mississauga, ON, Canada), lactic acid (Honeywell Fluka Analytical, Quebec, QC, Canada), and canola oil (Sobeys, Truro, NS, Canada).

2.2. Sample Preparation

The upcycled kale powder was supplied by Outcast Foods Ltd., Dartmouth, NS, Canada. Initially, verified supply chain production, kale (consisting of stalks, stems, rinds, and leaves) was collected and it was upcycled by the vacuum-drying method. Upcycled kale was ground into a fine powder using a grinder (Model 80393C Hamilton Beach Brands Inc., Glen Allen, VA, USA) and sieved to a size of 500 microns (Model 20210764 Retsch GmbH, Haan, Germany). The powder was stored in an airtight container at −20 °C until further use.

2.3. Experimental Design

RSM, a statistical technique, was used for modeling and to optimize response variables that are influenced by several factors. In RSM, a face-centered CCD model uses 20 runs at the low, center, and high axial levels of three independent factors. In the present study, the relation of independent variables includes ethanol concentration (%, X₁), temperature (°C, X₂), and time (min, X₃), which were evaluated for extraction to optimize the conditions using an ultrasonic bath (120 volts; amps 12 per 50/60 Hz; model 750D VWR International, West Chester, PA, USA). implemented in Minitab^® statistical software (version 21.1) (

Table 1. The independent variables and their coded levels with actual values for sonication.

Independent Variables	Levels
	−1.68	−1	0	1	+1.68
X1 Ethanol concentrations (%)	60	68	80	92	100
X2 Temperature (°C)	20	32	50	68	80
X3 Extraction duration (min)	30	85	165	245	300

). The solid-to-solvent ratio for the experiments is 0.5 g in 10 mL (Figure S1), which was constant for all the combinations. The objective was to determine the optimal setting of the independent variables to maximize the response variables (total carotenoid content, TCC) from kale extracts. The generalized second-order polynomial equation used in response surface analysis to fit the experimental data using regression analysis is shown in Equation (1).

$$\begin{gathered} \mathcal{Y}={\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+{\beta }_3{X}_3+{\beta }_1{X}_1^2+{\beta }_2{X}_2^2+{\beta }_3{X}_3^2+{\beta }_1{{\beta }_{2}X}_1{X}_2+ \\ {\beta }_1{{\beta }_{3}X}_1{X}_3+{\beta }_2{{\beta }_{3}X}_2{X}_3, \end{gathered}$$

(1)

where $\mathcal{Y}$ is the predicted response variable (TCC), ${\beta }_0$ is the intercept term of the model, $X_1$, $X_2$, and $X_3$ represent the independent variables of extraction parameters, ${\beta }_1$, ${\beta }_2$, and ${\beta }_3$ are considered linear effects, ${\beta }_1^2$, ${\beta }_1^2,$ and ${\beta }_1^2$ are quadratic effects, and ${\beta }_1{\beta }_2$, ${\beta }_1{\beta }_3,$ and ${\beta }_2{\beta }_3$ are the interaction effects of the regression coefficients of the corresponding independent variables, respectively.

2.4. Extraction, Purification, and Isolation of Carotenoids, Glucosinolates, and Flavonoids

2.4.1. Extraction Through Ultrasound-Assisted Technique

For this experiment, the UAE method was selected to extract targeted bioactive compounds from upcycled kale (Figure S2). UAE technique involves disruption of cell walls using ultrasound waves to enhance mass transfer and thereby increase the extraction yield. This method and selected extraction parameters are reported in the literature for kale (

Table 2. Central composite design and its response variable for carotenoid content extracted from upcycled kale.

Run Order	Ethanol Concentration (%)	Temperature (°C)	Extraction Time (min)	TCC (μg Carotenoid/g DW)
1	80 (0)	50 (0)	165 (0)	420.8
2	80 (0)	50 (0)	165 (0)	363.6
3	80 (0)	50 (0)	165 (0)	337.9
4	100 (+1.68)	50 (0)	165 (0)	444.7
5	80 (0)	50 (0)	165 (0)	376.3
6	80 (0)	50 (0)	165 (0)	405.7
7	68.1 (−1)	32.2 (−1)	245.3 (+1)	235.1
8	68.1 (−1)	67.8 (+1)	84.7 (−1)	178.3
9	60 (−1.68)	50 (0)	165 (0)	46.80
10	68.1 (−1)	67.8 (+1)	245.3 (+1)	153.0
11	91.9 (+1)	67.8 (+1)	245.3 (+1)	185.1
12	80 (0)	80 (+1.68)	165 (0)	259.0
13	80 (0)	50 (0)	30 (−1.68)	395.6
14	80 (0)	20 (−1.68)	165 (0)	254.8
15	68.1 (−1)	32.2 (−1)	84.7 (−1)	205.6
16	80 (0)	50 (0)	300 (+1.68)	191.3
17	91.9 (+1)	32.2 (−1)	245.3 (+1)	445.3
18	91.9 (+1)	32.2 (−1)	84.7 (−1)	403.0
19	91.9 (+1)	67.8 (+1)	84.7 (−1)	448.5
20	80 (0)	50 (0)	165 (0)	380.3

Note: Ethanol concentration (X₁), temperature (X₂), and extraction time (X₃). TCC, total carotenoid content.

) [11]. Herein, 0.5 g of the sample was weighed and added to 10 mL of anhydrous ethanol in an amber glass tube.

The extraction was using different ethanol concentrations, temperatures, and time points with intermittent shaking every 20 min as mentioned in Table 1. The upcycled kale extracts were centrifuged (Sorvall ST 18 centrifuge, Thermo Scientific Inc., Ottawa, ON, Canada) at 1957× g for 15 min to separate supernatant and pellet formation. The extracted supernatant layer of the samples was stored at −80 °C until the assays. Secondly, the optimized conditions from RSM were used for estimating the recovery percentage of TCC.

Further comparison of optimum conditions with different solvents and extraction methods was performed. The methods include UAE with canola oil (1 g in 10 mL for 45 °C, and 30 min) and deep eutectic solvent (DES) mixture of 8 mol of DL-menthol and 1 mol lactic acid (0.083 g in 10 mL of DES for 70 °C in 30 min), microwave-assisted extraction (MAE) (The Genius microwave 1100 W Model NNSG676W, Panasonic, Boston, MA, USA) with canola oil and 80% ethanol (0.25 g in 10 mL with 450 W for 2 min), water shaking bath (Model WS27 Sheldon Manufacturing Inc., Cornelius, IL, USA) 100% ethanol (0.5 g in 10 mL for 30 min at 57 °C) and Soxhlet extraction (Sigma-Aldrich, Oakville, ON, Canada) with 2:1:1 v/v of hexane–acetone–ethanol (2 g in 150 mL at 57 °C in 6 h) to determine TCC, total phenolic content (TPC), and antioxidant activity (ferric reducing antioxidant power (FRAP) and 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging. The extracts were also used for the determination of selected major carotenoids and phenolic compounds using ultra-high-performance liquid chromatography–electrospray ionization–mass spectrometry (UPLC-ESI-MS).

2.4.2. Determination of Total Carotenoid Content

The TCC was analyzed using a previously described method with some modification [15]. In a 15 mL Eppendorf tube, 1 mL of kale extract was added containing 1 mL of hexane–diethyl ether with a ratio of 9:1 v/v, followed by 2 mL of saturated NaCl solution, and vortex mixer (Cat no. 02215360 Fisher Scientific Inc., Ottawa, ON, Canada) for 1 min. The kale extract and hexane–diethyl ether were in equal volume in the solution. The upcycled kale sample was centrifuged at 700× g for 5 min to facilitate the separation of aqueous and organic layers. Half a milliliter of the organic layer was transferred and dried using a nitrogen evaporating system (Model N-Model 5085 EVAP^TM111, Organomation Associates Inc., Berlin, MA, USA). The dried extract was re-dissolved in 2 mL of methanol, and 200 µL of the sample was used to read the absorbance at 470 nm using a plate reader (Infinite^® 200 PRO, TECAN, Mannedorf, Switzerland). The TCC was estimated using the following formula:

$$\mathrm{T}\mathrm{C}\mathrm{C} \left(\mathsf{\mu }\mathrm{g}/\mathrm{g}\right)={\displaystyle \frac{AV\times {10}^4\times 4}{{\mathrm{A}}_{1\mathrm{c}\mathrm{m}}^{1\%}\times W}}$$

(2)

where A = absorbance at 450 nm; V = total volume of dissolved extract (mL); ${\mathrm{A}}_{1\mathrm{c}\mathrm{m}}^{1\%}$ = extinction coefficient for lutein set to 2550; W = weight of the sample (g); 10⁴ = conversion factor; and 4 = dilution factor.

2.4.3. Determination of Total Phenolic Concentration

The TPC of upcycled kale extracts was determined using the Folin–Ciocalteu method with modifications [16]. Briefly, 20 μL of kale extract was added to 100 µL of 0.2 N Folin–Ciocalteu reagent in a clear flat-bottom 96-well microplate (COSTAR 9017, Fisher Scientific, Ottawa, ON, Canada). Then, after 5 min of incubation, 7.5% sodium carbonate of 80 μL was added to the solution in the microplate. The 96-well microplate was kept in the dark using aluminum foil for 2 h at room temperature. A TECAN plate reader was used to record the sample absorbance at 760 nm (Infinite^® 200 PRO, TECAN, Mannedorf, Switzerland). The TPC was expressed using gallic acid equivalents (GAE) after establishing calibration curves using standards made from 59 to 1470 μM gallic acid.

2.4.4. Determination of Total Antioxidant Capacity

FRAP Assay

The antioxidant capacity of the samples was estimated spectrophotometrically using a FRAP assay [17]. The stock solution consisted of 300 mM acetate buffer mixed with 270 mg of iron (III) chloride hexahydrate (FeCl₃·6H₂O) dissolved in 50 mL of deionized water. Thereafter, 150 mg of ferrous tripyridyl triazine complex (Fe (III)-TPTZ) solution was added to 150 mL of hydrochloric acid (HCl) dissolved in 50 mL of distilled water. The FRAP reagent comprised 5 mL of FeCl₃·6H₂O, 5 mL (Fe (III)-TPTZ), and 50 mL of acetate buffer solution. A 96-well plate was used to add the 10 µL sample with 300 µL of FRAP solution, and then it was kept for 5 min incubation at 35 °C in the dark. The calibration curve ranges from 5 to 450 µM to determine the antioxidant capacity by measuring the final-colored sample at 593 nm absorbance using a TECAN plate reader based on Trolox equivalents (TE) (Infinite^® 200 PRO, TECAN, Mannedorf, Switzerland).

DPPH Radical Scavenging Assay

A modified method was used with 96-well microplates for the DPPH radical scavenging assay [18]. For this experiment, the extracts of kale were dissolved into a concentration ranging from 50 to 1600 µg/mL. The DPPH reagent of 0.2 mM was prepared to pipette 150 µL into each well containing 150 µL of sample. The antioxidant activity inhibition percentage was calculated using the formula of % inhibition = ${\displaystyle \frac{\left(Ab blank-Ab sample\right)}{\left(Ab blank\times 100\right)}}$; where Ab represents the value of sample absorbance with kale extract and Ab blank represents the value of absorbance of the sample without the kale extract. The antioxidant capacity was expressed as IC₅₀ for the kale extracts, which denotes the concentration needed to achieve a 50% reduction in the initial DPPH concentration.

2.5. UPLC-ESI-MS Analysis of the Extracts

Quantification of major bioactive compounds present in upcycled kale was performed using the UPLC-ESI-MS, which consists of a Micromass Quattro Micro-API MS/MS system (Micromass, Cary, NC, USA) and was managed with a MassLynx V4.2 data analysis software system [19]. The samples used for UPLC analysis were filtered through 0.2 μm nylon filters and transferred into amber vials. Samples were injected into a UPLC system, equipped with a reverse-phased Waters BEH C₁₈ column (2.1 × 100 mm, 1.7 μm) (Waters, Milford, MA, USA). The flow rate was 0.2 mL/min, and the injection volume was 5 μL. Chromatographic separation of the phytochemicals was performed using a linear gradient profile. This profile consisted of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B), with the proportion of solvent B changing over time (min); (time, solvent B%): (0, 6%), (2, 16.5%), (2.6, 17%), (3.1, 17.5%), (3.63, 17.5%), (4, 18.5%), (4.7, 20%), (6.5, 80%), (8.25, 80%), and (12, 6%) for lutein, quercetin, and kaempferol, and (0, 80%), (15, 100%), (16, 100%), (17, 80%), and (20, 80%) for β-carotene, respectively. The electrospray ionization source (ESI) was set for nebulizing gas (N₂) at 375 °C and capillary voltage for 3000 V. The mass spectrum conditions for the analysis of lutein, β-carotene, quercetin, and kaempferol were ESI+ m/z 568.8, ESI+ m/z 536.5, ESI− m/z 300.7, and ESI− m/z 285.2, respectively. The UPLC analysis was determined with a calibration curve created using analytical standards (purity above 98%).

2.6. Statistical Methods

Statistical analysis and CCD generation were estimated using Minitab^® statistical software (version 21.1). The experimental design and data were analyzed using statistical methods as previously described [19]. Data were verified for the validity of normal distribution, constant variance, and independence of error terms. The model adequacy check was assured with a non-significant lack of fit (p > 0.05), indicating an adequate fit. The independent variables and their interactions were analyzed using analysis of variance (ANOVA) at a 95% confidence level; the p-value is less than 0.05. Response surface plots were generated to represent the relationship between the extraction parameters and carotenoid content. Optimal extraction conditions to achieve maximum yield of upcycled kale were determined based on the response optimizer. Further, the optimized extraction was compared among the different extraction methods using the one-way ANOVA and multiple mean comparisons (Tukey’s HSD method) for estimating the statistical significance.

3. Results and Discussion

Based on the green chemistry principle, low-energy-based extraction methods are used to extract bioactive compounds. UAE is extensively used to extract phytochemicals due to its rapid extraction time, low temperature, non-toxic solvents, cost-effectiveness, and improved yield [20]. The green extraction principles applied in this study include renewable plant resources and waste minimization (upcycled kale), agro and no-toxic solvent (ethanol), energy efficiency (UAE method), safe and robust processes (UAE), and biodegradable extract without contaminants [21].

Traditionally, carotenoids are extracted using nonpolar solvents such as acetone or hexane because of their nonpolar long carbon chain structure [22]. Nonetheless, these nonpolar solvent extracts have restricted use in food due to health and environmental risks [23]. Consequently, we aimed to identify an appropriate solvent that is both safe and environmentally friendly for application in food systems. This work employed food-grade ethanol as an extraction solvent for carotenoids, as it is classified as a GRAS (Generally Recognized as Safe) solvent by the U.S. Food and Drug Administration [24].

The present work optimized the factors affecting the extraction conditions by RSM-based CCD for the determination of total carotenoids in upcycled kale by the UAE method. Further, the bioactive phytochemicals in kale were quantified using UPLC-ESI-MS analysis, indicating the potential of the chosen upcycled kale extract to be used in food, nutraceutical, and other industry sectors.

3.1. Influence of UAE Parameters on TCC

RSM employs three-dimensional (3D) surface plots to represent the interaction of two independent variables on a response variable, while other factors remain constant. The present study describes the successful extraction conditions (independent variables) from upcycled kale using UAE for the estimation of TCC (response variable). The range for TCC of RSM obtained under 20 experimental runs from UAE conditions is shown in Table 1. The extraction yield value varied significantly from 46.8 µg/g DW to 449 μg/g of DW (Table 2). The highest TCC yield was obtained with 91.9% ethanol, 68 °C, and 85 min. The results demonstrated that extraction parameters have a significant impact on extraction yield. The 3D response surface plots illustrate the influence of ethanol concentration, extraction temperature, and extraction duration on the yield of TCC and were used to study the relationship between TCC and extraction parameters (Figure 1).

Based on the 3D response surface plot, a synergistic effect exists between ethanol concentration, extraction temperature, and time in maximizing carotenoid content. The carotenoid content increased with an increase in extraction temperature and time at a fixed ethanol concentration to a certain point, and afterwards, TCC decreased. The ultrasound waves create acoustic cavitation bubbles that disrupt the cell wall, facilitating the release of carotenoids to increase the efficiency of the extraction yield [25]. The interactive effect of ethanol concentration and temperature was demonstrated in Figure 1. As the ethanol concentration increased, the TCC increased, with a fixed extraction temperature to confirm the quadratic effect of ethanol concentration and time. The effect of ethanol concentration on yield reflects the solubility properties of the carotenoids, indicating that a higher proportion of ethanol enhances the solid-to-solvent interactions by increasing extraction efficiency [26]. These findings are consistent for extracting high carotenoid content from a green leafy vegetable, Gymnema lactiferum, with increased aqueous ethanol concentration [27]. The ability of ethanol to interact with polar and nonpolar molecules such as carotenoids makes it effective to extract [28].

The results of extraction time on extraction yield indicate that a longer extraction time negatively impacts the extraction yield, whereas a shorter extraction time (30 min) leads to a higher recovery of carotenoids (Figure 2). The carotenoids might potentially degrade due to extended extraction time and have reduced the extraction yield [29]. A similar study was found by [30] the recovery of carotenoids was demonstrated to be time-dependent, with an increase in yield as extraction time extended from 10 to 40 min. The ideal conditions from the study are an extraction duration of 39 min, a temperature of 47 °C, and a solid–to–liquid ratio of 30 g/100 mL. The duration of extraction significantly influences the total carotenoid yield, as prolonged solid-to-solvent interaction may improve molecule diffusion [31].

It is noted that a rise in temperature from 20 °C to 50 °C obtained maximum recovery, and afterwards a significant reduction in carotenoid content was observed (Figure 2). The significant increase in yield at higher temperatures accelerated mass transfer and enhanced solubility [32]. However, the temperature beyond the central point has a major drawback: the degradation of carotenoids or increased energy consumption.

The total carotenoids in annatto seeds showed a similar trend in the context of an interaction effect between temperature and time [33]. The positive correlation between extraction time and yield could be attributed to the cumulative effect of ultrasonic waves on cell disruption and solute diffusion. The observed result suggests that extended extraction time may not yield higher yields due to possible degradation of the compounds. One of the limitations of this study was the use of an ultrasonic bath for the UAE method, which did not have controls for extraction parameters such as sonication frequency and amplitude that might have influenced the extraction efficiency of carotenoids from upcycled kale.

3.2. UAE Model Fitting

The ANOVA for estimating TCC was conducted using CCD. The research data were fitted accordingly to a second-order polynomial 3D response surface plot and represented as a quadratic model for three factors, namely ethanol concentration, temperature, and extraction duration (

Table 3. ANOVA table for the response surface methodology of the CCD model.

Source of Variation	p-Value
Constant	0.000
% EtOH (X1)	0.000
Temp (X2)	0.126
Time (X3)	0.015
% EtOH * % EtOH	0.010
Temp * Temp	0.017
Time * Time	0.084
% EtOH * Temp (X1X2)	0.485
% EtOH * Time (X1X3)	0.152
Temp * Time (X2X3)	0.033
Lack of Fit	0.050
R-Square	90.18%
Adjusted R2	81.33%

).

The generated equation computed for the polynomial model is given below in coded variables:

Y = −3399 + 64.9 X₁ + 21.08 X₂ + 4.75 X₃ − 0.3026 X₁*X₁ − 0.1222 X₂*X₂ − 0.00402 X₃*X₃ − 0.0621 X₁*X₂ − 0.0295 X₁*X₃ − 0.0315 X₂*X₃.

(3)

Based on the regression model for this design, the experimental data were found to be statistically significant for X₂X₃ (where p < 0.05) at a 5% level of significance. The effect of the remaining variable interactions was found to be non-significant, indicating X₁X₂ and X₁X₃ had less effect (where p > 0.05) on the UAE in terms of TCC. The statistical analysis indicated the coefficient of determination (R²) value for the mentioned model was 0.9018, suggesting that 90% of the variability in the response variable (TCC) was explained by the model. By this, we can suggest that the developed model could appropriately represent the real relationship among the parameters chosen. In this RSM design, the non-significant lack of fit p-value is greater than 0.05 (p > 0.05), which implies that the proposed model fits the experimental data and suggests that independent variables had a considerable effect on the TCC.

3.3. Optimization of the Extraction and Model Validation

The effective extraction of bioactive compounds from upcycled kale was studied to optimize through CCD in RSM. Among the 20 experimental runs, the results for the optimal conditions of UAE to maximize the yield were found to be 100% (ethanol concentration), 57 °C (temperature), and 30 min (time) when the solid-to-solvent ratio (1:20 v/v) was kept constant (Figure 2). The TCC of 392 μg/g DW was obtained from the optimum conditions of UAE mentioned. Based on the RSM-CCD, the predicted results of 550 µg/g DW were higher than the actual results obtained using optimized extraction parameters (Figure 2). CCD is a powerful statistical tool for the extraction optimization processes to reduce experimental effort and provide cost-effective extraction for various food industries [34]. In this context of the study, limited data are available on the extraction optimization of the upcycled kale.

However, only a few studies have explored the optimization of carotenoid extraction parameters using the CCD. It was found that a temperature of 80 °C, a solvent mixture of 10 mL n-hexane and acetone, and an extraction time of 100 min were the best conditions for extracting 0.97 mg/g DW of carotenoids from pumpkin peel using UAE [35]. The optimization of carrot extraction was studied utilizing ultrasound and ethanol as a solvent. The ideal circumstances for achieving the maximum TCC content of 31.8 ± 0.55 μg/g DW of carrot were 51% ethanol, 32 °C temperature, and 17 min of duration [36]. There have been reports of RSM-CCD for extracts of Passiflora edulis, Cassia auriculata, and Sesbania grandiflora with higher carotenoid extraction efficiency [37,38,39]. Moreover, it is essential to consider green extraction methods due to the increasing demand for sustainable and environmentally friendly processes that produce products with higher quality, better functionality, reduced energy consumption, and minimized waste generation [20]. Furthermore, the optimal extraction conditions for upcycled kale were compared with those previously reported for cantaloupe waste [40]. Our results show that the carotenoids in upcycled kale are three times higher than those in cantaloupe waste. Therefore, this could signify a practical set of parameters for achieving higher carotenoid content compared to prior research.

3.4. The Recovery Percentage of Optimized Conditions for TCC

First, the powdered kale was extracted using optimum conditions for the extraction yield of the sample. The supernatant was collected after centrifuging to calculate the TCC. Next, the sample was used for the second extraction process of carotenoid content by following the same optimum conditions; this procedure was repeated three times for the total recovery of carotenoid content. The results were calculated for the recovery percentage of TCC with four times extraction. The multiple extraction process addresses the challenges, including lesser recovery of targeted compounds due to the complexity of plant matrices and uneven distribution of bioactive compounds within the plant extract [41].

The recovery percentages of TCC by three subsequent extractions of optimized UAE were investigated (

Table 4. Recovery of TCC after repeated extractions of UAE.

Extraction Run	Recovery
	TCC (μg/g)	%
1	392	71.4
2	99.4	18.1
3	35.4	6.46
4	21.8	3.98

). The data revealed that the first extraction provides about 71% of the potential total recovery of TCC, followed by 89.5% by the second extraction.

3.5. Comparison of Optimized Extraction Conditions of UAE with Different Extraction Methods

The extraction of bioactive compounds from kale was conducted using various techniques such as conventional and green extraction methods. The most used methods in laboratories were considered for the comparison of extraction methods. The extracts prepared from seven extraction methods (Section 2.4.1) were compared using TPC, TCC, FRAP, and DPPH assays (

Table 5. Comparing the effectiveness of the optimized UAE for bioactive extraction from upcycled kale with different extraction methods and solvents.

Extraction Methods	Solvent		Assay
	Type	Concentration	TCC (μg Carotenoid/g DW)	TPC (mg GAE/g DW)	FRAP (μmole TE/g DW)	DPPH (IC50 mg/mL)
Optimized UAE	Ethanol	100%	392 ±1.8 ab	10.5 ± 1.4 c	13.9 ± 1.5 b	2.04 ± 0.3 a
WSB	Ethanol	100%	345 ±16.4 bc	9.0 ± 0.9 c	11.8 ± 0.7 b	2.29 ± 0.4 a
MAE	Ethanol	80%	354 ± 7.6 abc	34.1 ± 3.3 a	34.2 ± 10.3 b	0.68 ± 0.05 c
UAE	Canola oil	100%	128 ± 1.7 d	1.75 ± 0.09 d	8.25 ± 4.8 b	1.15 ± 0.06 bc
MAE	Canola oil	100%	129 ± 1.5 d	6.03 ± 0.07 cd	14.8 ± 2.8 b	0.98 ± 0.05 bc
UAE	NADES (DL-menthol: Lactic Acid)	8:1 mol HBA/mol HBD	425 ± 48.2 a	19.4 ± 2.9 b	102 ± 37.0 a	NA *
Soxhlet Extraction	Hexane: Acetone: Ethanol	2:1:1 v/v/v	292 ± 53.08 c	3.46 ± 1.8 d	2.9 ± 1.7 b	1.4 ± 0.05 b

Note: Different superscript letters (a–d) within a column indicate statistically significant differences (p < 0.05) based on Tukey’s HSD test. Abbreviations: HBA, hydrogen bond acceptor; HBD, hydrogen bond donor; v/v, volume per volume; NADES, natural deep eutectic solvent; DW, dry weight basis. NA *, not available.

).

3.5.1. Total Carotenoid Content

The effect of extraction conditions on the TCC from upcycled kale was conducted. The one-way ANOVA analysis was performed for optimized extraction conditions, and Tukey’s test was used for the comparison of different extraction methods. The results showed that UAE with solvent as natural deep eutectic solvent (NADES, DL-menthol: lactic acid) extracted significantly higher carotenoid content (425 ± 48.2 μg/g DW), followed by optimized UAE with ethanol as solvent (392 ± 1.8 μg/g DW) compared to methods mentioned in Section 2.4.1. The lower amounts of carotenoids were observed using canola oil as a solvent for both UAE and MAE methods. Advanced extraction methods such as UAE can extract bioactive compounds in less time, at low temperatures, with fewer solvents, and with lower energy requirements with higher extraction efficiency. Numerous experimental studies have been investigated to evaluate the potential of NADES as a solvent. According to the study by [42], lycopene extracted from tomatoes using the UAE method with NADES–DL-menthol–lactic acid as solvent at 70 °C for 10 min provided a higher extraction yield compared to other NADES solvents. Similar extraction methods and parameters were considered for this study to compare. However, it should be noted that drawbacks of NADES, such as high viscosity, problematic GC analysis, and high cost compared to ethanol, would affect the application to the food and nutraceutical industry [43]. Owing to the limitations of NADES, food-grade ethanol with UAE showed great ability to extract carotenoids from upcycled kale, providing better applications to the food industry.

3.5.2. Total Phenolic Content

The content of TPC in kale extracts ranges from 1.75 to 34.1 GAE mg/g DW for the mentioned extraction methods (Table 5). The MAE with solvent as ethanol 80%, power as 450 W, and time as 20 s generated the highest TPC of 34.2 mg GAE/g DW. The highest recovery of TPC was reported in mango waste using MAE with minimal degradation of the extract, and similar experimental conditions were considered [44]. As shown in Table 3, one-way ANOVA was performed, and Tukey’s HSD comparison was performed, which concluded that MAE provided significantly greater TPC compared to other extraction methods. However, for the recovery of phenolics, it has been comprehensively reported that various factors, including temperature, solvent concentration, extraction time, and solid-to-solvent ratio, can influence the concentration and distribution of secondary metabolites [45].

3.5.3. Total Antioxidant Capacity

The FRAP and DPPH radical scavenging assays were performed to determine the ability of the extract of kale to inhibit oxidation. The strongest antioxidant capacity was observed by UAE at 70 °C for 10 min using NADES (DL-menthol: lactic acid) as solvent. The statistical analysis indicated that UAE with NADES was significantly different among the extraction methods mentioned for comparison. The data indicates that it could significantly influence the antioxidant activity of kale extract. A similar trend was also noticed in the recovery of antioxidant activity by NADES solvent compared with ethanol, canola oil, hexane, and acetone [43,46]. On the other hand, the two extraction methods, UAE and WSB, using ethanol as a solvent, were recorded for the highest scavenging activity of kale extract compared to other extraction methods. The optimal antioxidant activity was found in kale extracts with 100% ethanol as the solvent concentration. Based on the DPPH radical scavenging, antioxidant capacity was the highest in the extract prepared using ethanol/MAE (IC₅₀ of 0.68 ± 0.05 mg/mL) (Table 5). Previously, [47] observed significantly higher DPPH inhibition % of pumpkin extracts using corn oil by similar extraction conditions of the MAE method mentioned in this experiment. Overall, the antioxidant capacity of upcycled kale was comparable to that of fresh kale (Figure S3).

3.6. Phytochemical Characterization of Kale Extracts Using UPLC-ESI-MS Analysis

The major phytochemicals of upcycled kale extracts were determined by UPLC-ESI-MS analysis obtained at optimal extraction conditions compared with different extraction methods (

Table 6. The major bioactive compounds of various extracts of upcycled kale determined using UPLC-ESI-MS.

Bioactive Compounds (μg/g DW)	Different Extraction Methods
	Ultrasound-Assisted Extraction	Microwave-Assisted Extraction	Water-Shaking Bath	Soxhlet Extraction
	Ethanol	Ethanol	Ethanol	Hexane–Acetone–Ethanol
Lutein	879 ± 125 b	1229 ± 369 b	843 ± 130 b	2323 ± 220 a
β-carotene	947 ± 95.7 a	492 ± 63 b	1009 ± 186 a	964 ± 45.5 a
Total carotenoid content	1829 ± 221	1721 ± 431	1853 ± 317	3287 ± 265
Kaempferol	2.83 ± 1.06 a	2.79 ± 0.59 a	1.70 ± 0.33 a	2.70 ± 0.3 a
Total quercetin	73.2 ± 13.8 b	267 ± 21.3 a	68.3 ± 1.95 b	87.4 ± 10.2 b
Total flavonoid content	76.1 ± 14.9	270 ± 21.9	70 ± 2.3	90.1 ± 10.5
Chlorogenic acid	84.4 ± 20 b	536 ± 18.1a	74.6 ± 2.2 b	101 ± 7.8 b
Ferulic acid	92.2 ± 2.9 c	146 ± 11.8 b	93.4 ± 0.3 c	234 ± 12.1 a
Total phenolic content	177 ± 22.1	682 ± 29.9	168 ± 2.5	336 ± 19.9
Total phytochemicals	2082	2673	2091	3713

Note: total carotenoid content (lutein and β-carotene) and total phenolic content (kaempferol, total quercetin, chlorogenic acid, and ferulic acid). Different superscript letters (a–c) within a raw indicate statistically significant differences (p < 0.05) based on Tukey’s HSD test.

). The results revealed various phytochemicals that are present in upcycled kale extracts, including carotenoids, flavonoids, and phenolic compounds. This study required modifications to the mobile phase composition and gradient elution system to effectively separate the compounds of interest from the analyte. The results showed a significant difference (p < 0.05) for lutein using Soxhlet extraction compared to other extraction methods. The traditional extraction (Soxhlet) process requires large volumes of organic solvents such as acetone and hexane, prolonged time, and makes it less desirable compared to green extraction methods [48]. In this study, a higher concentration of lutein was observed in upcycled kale with MAE (1229 ± 369 μg/g DW) followed by UAE (879 ± 125 μg/g DW) extraction methods. However, WSB and UAE methods showed higher extraction efficiency for β-carotene from upcycled kale extracts (p < 0.05). Based on our findings, a similar study was performed using UPLC-ESI-MS analysis to identify and quantify carotenoids in green kale and reported that lutein (524 ± 28.4 μg/g DW) and β-carotene (307 ± 2.92 μg/g DW) had higher content compared to other carotenoids [49]. Additionally, researchers conducted a study on the carotenoid profile of the Brassicaceae family using HPLC-DAD analysis. The results of the study highlight that kale recorded the highest amount of lutein and β-carotene among the 27 tested Brassica varieties [50]. Dietary supplementation with lutein in an in vivo mice model was found to be effective for skin photoprotection against reactive oxygen species (ROS) generation due to exposure to ultraviolet radiation. These findings indicate that lutein inhibits UV-induced DNA damage and immunosuppression in the skin [51,52]. Zheng et al. (2022) investigated the anti-aging properties of β-carotene in mesenchymal stem skin cells (MSCs) both in vitro and in vivo. According to the experimental results, β-carotene showed a significant difference in slowing down anti-aging markers such as p16 and p21 by controlling the KAT7-P15 signaling pathway [53]. The differences in carotenoid levels for kale are due to variety, genetic variation, and environmental factors affecting growth [54].

Furthermore, the major and selected phenolic compounds present in upcycled kale were quantified (Table 6). A significant difference was found between the extraction methods for the most abundant phenolic compounds, including total quercetin, chlorogenic acid, and ferulic acid (p < 0.05). The MAE provided the greatest polyphenol concentration compared to the other methods. A previous study has shown that flavonoids from kale seeds were determined using LC-MS/MS for kaempferol (0.42 ± 0.10 μg/g DW) and quercetin (0.20 ± 0.02 μg/g DW) [55]. Phenolic acids in kale seeds have also been reported as chlorogenic acid (29.8 ± 2.56 μg/g DW) and ferulic acid (10 ± 0.56 μg/g DW) [56]. The polyphenolic content in kale seeds is significantly lower compared to upcycled kale extracts in this study. The identified phenolic compounds of kale have health benefits, for instance, antioxidant, anti-inflammatory, antimicrobial, and antitumor activity [57,58,59].

Amongst the bioactive compounds present in upcycled kale extract, the MAE with ethanol as solvent provided the highest total carotenoid content, with enhanced extraction of more polar compounds compared to conventional extraction, i.e., hexane. However, the UAE and WSB extraction showed improved extraction for less polar bioactive compounds, including total carotenoids and phenolic compounds. The quantified phytochemicals mentioned above demonstrated significant biological properties, including antioxidant, anti-inflammatory, and free radical scavenging activity [60]. Overall, phytochemicals such as carotenoids, flavonoids, and other polyphenols of these extracts have the potential to be used in functional foods, nutraceuticals, and cosmeceutical formulations.

4. Conclusions

The statistical method, RSM-CCD, was successfully used for the optimization of the extraction parameters based on the TCC of the kale extracts. The results demonstrate that the optimization of extraction conditions led to the production of carotenoid-rich extracts from upcycled kale, with ethanol concentration and extraction time proving to be the most influential parameters in the process. UAE appears to be an optimal processing method for the extraction of carotenoids from upcycled kale without the need for conventional organic solvents such as hexane. To overcome the limitation of ultrasound baths, parameters such as frequency and amplitude can be addressed by using advanced ultrasonic extraction systems. Based on the UPLC-ESI-MS analysis of the optimized kale extract, the most abundant bioactive phytochemicals were lutein, quercetin, and chlorogenic acid. The potential of upcycled kale extract can be further investigated for potential use in food, nutraceutical, and cosmeceutical development.