Spinach (Spinacia oleracea) was provided by Unicoop Tirreno S.C. (Viterbo, Lazio, Italy) and kept at 4°C until use. Enzyme preparations were as follows: cellulase (0.8 U/mg) and polygalacturonase (≥0.3 U/mg) from Aspergillus niger and xylanase (≥2500 units/mg) from Aspergillus oryzae, all obtained from Merck (Milan, Italy). In addition, all reagents, including acetone for solvent extraction and chloroplast isolation kit were purchased from Merck (Milan, Italy).

Enzyme-Assisted Extraction (EAE). The EAE was performed following the method developed by Lombardelli et al. (2024). Briefly, 25 g of spinach was blanched at 100°C for 5 sec, blended and added to 440 mL of McIlvaine buffer (0.1 M, pH 5) containing the tailored enzyme mix (total dose 0.1 U/g consisting of: cellulase 40% w/w, xylanase 41% w/w, and pectinase 19% w/w). The incubation time (1–3 h) (Lombardelli et al., 2024; Mazzocchi et al., 2023) and the total dose of metal ions (165 ppm) (Cu²⁺: Zn²⁺, 1:10) were varied according to the experimental plan drawn up by the design of experiment (DOE, Table 1). The temperature was set at 25°C. At the end of the incubation, the extract was centrifuged (4500 rpm, 10 min at 4°C, Heraeus Megafuge 16R Centrifuge, Thermo Scientific, Milan, Italy) and the pellet was subjected to analysis using a chloroplast isolation kit.

Solvent Extraction (SE). The SE was performed following the method developed by Ašimović et al. (2016) using 80% (v/v) acetone as a solvent. The extraction was performed maintaining the same solid–liquid ratio and the same operating conditions applied for EAE (Table 1). After centrifugation, the pellet was used for chloroplast isolation kit.

In order to determine (i) the effect of extraction time and the total dose of metal ions on the Chl-containing chloroplast and (ii) the best conditions to maximize the recovery of intact chloroplast, a DOE based on a central composite design with the quadratic model was applied. Considering that the yield of isolated chloroplasts is expressed on a unit Chl basis (mg of Chl/mL), the measured dependent variable (y) was the concentration of Chl (mg/mL). The variance for each evaluated factor was divided into linear, quadratic, and interactive components, and represented using the second-order polynomial function as follows:

y = b₀+b₁x₁+b₂x₂+b₁₁x₁²+b₂₂x₂²+b₁₂x₁x₂(1)

The coefficients of the polynomial were represented by b₀ (constant term), b₁ and b₂ (linear coefficient), b₁₁ and b₂₂ (quadratic coefficient), and b₁₂ (interactive coefficient). The significance of all the terms in the polynomial function was statistically assessed using the F-value at a probability (p) of 0.001, 0.01, or 0.05. The regression coefficients were then used to generate contour plots. The experimental design, statistical analysis, and response surface methodology (RSM) were performed using Minitab 17.1 (Minitab Inc., Pennsylvania, USA). RSM was used to verify the presence of the combined effect between the total dose of metal ions and extraction time (independent variables). These variables were coded as X1 and X2, respectively, and each had three levels: −1, 0, and +1. A total of 18 combinations, which included three replicates of the center point, were carried out in random order according to the central composite design. The level of variables for (x1 and x2) and (X1 and X2) are shown in Table 1.

The isolation of intact chloroplast was performed using the specific kit supplied by Merck (Milan, Italy), following the procedure reported in the technical data sheet.

Briefly, the extract is gradually passed through a filter mesh into 50 mL tubes. The filtrate is then divided into four 50 mL tubes, ensuring that the volume in each tube does not exceed two-thirds of its total capacity. To remove unwanted whole cells and cell wall debris, each tube is centrifuged for 3 min at 200 rpm, resulting in the precipitation of a white pellet at the bottom of the tube. For the isolation of chloroplast, the supernatant is carefully transferred into a fresh, pre-chilled 50 mL tube and centrifuged for 7 min at 1,000 rpm. This step sediments the chloroplasts, which form a distinct green pellet. The supernatant is discarded, and the pellet is resuspended in 2 mL of 1× chloroplast isolation buffer solution containing bovine serum albumin by gently pipetting up and down, ensuring that foaming is avoided during the process. The resuspended pellet is pooled into a single tube for subsequent steps. To purify intact chloroplasts, the separation of intact from broken chloroplasts is performed by centrifugation, applying a 40% Percoll^® layer to achieve effective segregation.

At the end of the procedure, a suspension containing only intact chloroplasts was obtained. This suspension was used to determine the yield of chloroplast, expressed on a unit Chl basis (mg of Chl/mL). This requires the extraction of the Chl from the chloroplast suspension with an organic solvent: 10 µL of the chloroplast suspension was added to 1 mL of an 80% acetone solution and mixed. Then, the sample was centrifuged for 2 min at 3000 rpm, retaining the supernatant and reading the absorbance at 652 nm (A₆₅₂) (UV-visible, Shimadzu UV 2450, Milan, Italy), using 80% acetone solution as reference blank. To obtain the mg of Chl per milliliter of chloroplast suspension, the following equation was applied:

As reported in literature (Gedi et al., 2017; Henriques, 2001), the Chl concentration serves as an indirect indicator of the amount of intact chloroplasts.

The optimal freeze-dried extract, characterized by the highest Chl content and the most suitable colorimetric parameters, was used in the preparation of a real food. Meringue, a sweet baked good typically white, has been selected in order to better appreciate the coloring impact of the recovered green pigment.

Meringues were prepared by mixing three egg whites (100 mL) with 220 g of powdered sugar (added in two steps) using electric whisks (Phillip, Milan, Italy) at speed 3 (600 rpm) for 10 min to obtain soft/stiff peak batter structure. The prepared meringue batter was divided into six equivalent portions, and the appropriate dosage of freeze-dried colorant was added (0, 1, 2, 3, 5, and 7% w/w sugar basis) and squeezed into waxed paper trays before baking. The baking temperature was adjusted to 75°C for 3 h using an oven (Venticell, MMM Medcenter Einrichtungen GmbH, Planegg/München, Germany). The cooked meringues (n = 150) were subjected to colorimetric and sensory analyses.

Color measurement of the extract and meringue was carried out using a CR-5 colorimeter (Konica Minolta, Tokyo, Japan) by a D65 illuminant on the basis of L*, a*, and b* values, where the L* value indicates lightness, and its value ranges from 0 to 100; the a* value gives the degree of the red-green color, with a lower negative a* value suggesting more green (−a); and the b* value indicates the degree of the yellow-blue color, with a higher positive b value suggesting more yellow (+b). The colorimeter was calibrated using a standard white and black plate. The ΔE (color difference) between samples [estimated as the difference with trial 3 (reference sample with the lowest a*)] was calculated by equation reported in Lombardelli et al. (2021).

Meringue samples were analyzed for their organoleptic characteristics by a panel of 25 trained members (26–47 years of age, included 14 female and 11 male panelists) using a five-point hedonic scale. The panelists scored different attributes, described each one to facilitate scoring (Table 2). The meringues were evaluated for their brightness, color homogeneity, green color intensity, smell (herbaceous), crumb color, internal color, texture, hardness, taste, herbaceous flavor, and overall acceptability. The samples were placed on white plates and identified with random three-digit numbers. The panelists evaluated all samples in a testing area with good light condition and were instructed to rinse their mouth with water between samples to minimize any residual effects.

All trials were performed in triplicate, and results were expressed as the mean ± standard deviation. The mean comparisons were carried out using an analysis of variance (ANOVA) to find the effect of individual factors and their interaction on colorimetric parameters (p < 0.01). Tukey’s post-hoc test (HSD) was also performed (p < 0.05) using EXCEL® extension DSAAS-TAT for multiple comparisons between samples.

The recovery yield of intact chloroplasts by EAE and SE, varying the total metal dose (0–165 ppm) and the extraction time (1–3 h), was investigated. From a first visual analysis, the difference between the two extraction methods was already evident. In fact, in Figure 1, it is possible to distinguish broken (top) and intact (bottom) chloroplasts; the latter seems to be more numerous in Figure 1A, suggesting EAE allowed the recovery of a greater amount of intact chloroplasts compared to SE (Figure 1B).

The enzymatic approach for extracting pigments retained within their structural environment has already proven successful. Cuccolini et al., (2013) utilized a commercial enzymatic preparation, whereas Lombardelli et al., (2020) applied a tailored enzymatic mixture specifically designed according to the characteristics of the vegetable matrix. Akin to the approach adopted in this study for the recovery of chloroplast, carotenoids still contained within chromoplasts were obtained, thereby preventing their degradation (Cuccolini et al., 2013; Lombardelli et al., 2020).

The effect of process variables (time and total metal dose) on the amount of Chl in intact chloroplasts was evaluated by RSM. Table 1 shows the results of the full factorial design, and Table 3 displays the coefficients of the mathematical model and statistical parameters, proving that the RSM developed was adequate (Models p-value = 0.000). For the response variable considered (Chl concentration), the R² values were equal to 0.96 for both extraction methods (EAE and SE), indicating that the regression models were able to effectively explain how the variables and their interactions affected the response.

Furthermore, the significance of each equation coefficient was determined using the p-value (Table 3) (Mazzocchi et al., 2023). The analysis of variance was performed to determine the significance of the linear, quadratic, and interaction effects of the independent variables (extraction time and total metals dose) on the dependent variable (Chl concentration). Considering the EAE process, both the extraction time and metal dose had a statistically significant linear effect (p-value < 0.001, Table 3), contributing to the recovery of intact chloroplasts. The quadratic effect was significant (p-value <0.05) only for metal dose (Table 3). Otherwise, no interactions among the variables were revealed (p-value > 0.05). The results allowed us to develop a useful equation to predict the amount of intact chloroplasts recovered with varying metal dose (X1) and extraction time (X2). Considering SE, only the linear and the quadratic terms of time were significant (p-value < 0.001).

The interactions between the variables are depicted through the contour plots (Figure 2). It is interesting to note that for EAE, the greatest amount of recovered pigment and, therefore, the highest yield of intact chloroplasts was obtained as the extraction time and metal dose increased (Figure 2A). For the SE, however, the yield of intact chloroplasts was greater as the extraction time decreased, regardless of the metal dose (Figure 2B), proving the detrimental effect of solvents on the integrity of organelles (Harwood, 1998).

Considering data reported in Figure 3A, it was confirmed that the amount of Chl in intact chloroplasts depends both on the extraction time and on the presence of metals. Indeed, the greatest concentration was obtained in the samples recovered using the maximum dose of metals (165 ppm) for 2 h (trial 14) with no remarkable difference extending the extraction time up to 3 h (trials 3 and 9). Irrespective of the process time, a notable difference in the amount of intact chloroplasts was observed between the sample with the maximum dose of metals (trials 2, 6, 14, 3, and 9) and the sample without metals (trials 1, 7, 16, 8, and 10). Trials with halved metal dose were characterized by an intermediate Chl content (Figure 3A). For the SE, data reported in Figure 3B confirmed that the amount of intact chloroplasts depended on the extraction time, whereas the presence of metals was not relevant. Indeed, with the same extraction time and different metal dose, the concentration of chloroplasts was similar (trial 1 and 7 vs trials 2 and 6 vs trial 12). Unlike the enzymatically extracted samples, the greater content of intact chloroplasts was identified in the tests with a shorter extraction time (1 h), probably because a prolonged contact with organic solvent destroyed the organelles (Lombardelli et al., 2020).

In addition to the quantification of pigment, CIELa*b* parameters have been considered for characterizing the extracts (Tables 4 and 5). In particular, the greenness is expressed by the a* value; the more negative it is, the greener the sample is. For EAE trials, the total color difference (ΔE) was also evaluated in order to better highlight the color changes. In detail, a total color difference ranging from 0 to 2 is considered unrecognizable, a ΔE ≥ 5 is rated significant and recognizable, whereas a remarkable variation is found when 5 ≤ ΔE < 12, and a different color if ΔE ≥ 12.

The lowest a* value for EAE samples (Table 4) was observed by applying the highest metal dose at 2h (trial 14) and 3h (trials 3 and 9), while preserving the greenness. By decreasing the metal dose, the color shifted towards duller nuances until obtaining the worst green color (higher a* value) for the tests without metal ions (trials 1,7 and 16), as also highlighted by the ΔE value.

Considering the colorimetric data of the SE samples (Table 5), the presence of metal ions did not preserve the color which tended to turn on duller and less brilliant nuances, tending to grayish hue when the extraction time extends. These results fit with what has been reported in the literature. It is well known that the use of organic solvents may easily lead to a partial degradation of the extracted molecules and, consequently, to a qualitative color deterioration (Taghavi et al., 2023).

Considering the obtained results, it is possible to hypothesize that the synergistic use of a tailored enzymatic mix and a suitable combination of metal ions (Zn²⁺ and Cu²⁺) may be an efficient method for the recovery of Chl within intact chloroplasts.

Meringues are a baked mixture of egg white and sugar, and the quality is mainly linked to the processing methods, the ingredients of the product recipes (Yüceer & Caner, 2021), as well as to the typical pure white. The freeze-dried Chl-based sample, containing intact chloroplast (trial 14), was used in the formulation of meringues at different doses (0, 1, 2, 3, 5, and 7 % w/w sugar based). The pictures and the coordinates of the tristimulus values (CIELa*b*) in Table 6 suggested a significant variation of the colorimetric parameters L*, a*, and b* between the meringue without colorant (0% w/w) and the product at lowest concentration (1% w/w). Concerning a*, it significantly varied toward negative values (more intense green nuance) up to a Chl-based extract concentration equal to 3% w/w. Over this dose, the addition of increased amount of colorant did not affect the colorimetric parameters a* and b* (Table 6).

The effect of the incorporation of powder colorant on the organoleptic characteristics of meringues is presented in Figure 4. The sensory evaluation suggested that, irrespective of the dosage, the addition of colorant had a significant effect on the quality of meringues (p < 0.05) except for internal color. Excluding the attributes related to the intensity of the green color and the overall acceptability, it may be noted that, considering the dose of the colorant added, no significant differences were observed among the control (0%) and the 1 and 2% samples for taste, smell (herbaceous), texture, hardness, and herbaceous flavor. The sample containing 3% pigment slightly differed from the 0–2% samples, with the only exception of brightness (Figure 4).

The meringues with the highest dose of colorant (5 and 7%) were similar, but completely different from the other samples with the exception of crumb color and color homogeneity. Regarding the intensity of the green color, as expected, the 3%, 5%, and 7% samples obtained the highest score (Figure 4). The overall acceptability on a five-point hedonic scale was higher than 4.6 for meringues at 1–3% and did not differ from the control. However, for 5 and 7%, the overall acceptability was rated as poor (3.2 and 2.6, respectively) because of the strong taste and aroma of these samples. The sensory analysis data suggested that the most suitable dose of colorant that can be used in an original and innovative green meringue preparation is 3% (w/w), without significantly affecting the sensory characteristics.