Ripe M. indica var. Edward was provided by Agromar S.A.C (Sullana, Perú), M. dubia fruits were purchased in a local market from the Cooperativa Agroindustrial Yarinacocha (Coopay), Pucallpa, Ucayali, Perú. The fresh whole milk was purchased from the livestock establishment in the agricultural technical college, located in Sullana. The powdered milk, sugar, carboxymethylcellulose (CMC), were also purchased in the local market of Sullana (Perú).

DPPH (2,2-diphenyl-1-picrylhydrazyl), gallic acid, Folin-Ciocalteu, sodium carbonate, oxalic acid, ascorbic acid, 2,6-dichlorophenolindophenol, ethanol (99.9% purity) and methanol (99.5% purity) were obtained from Sigma-Aldrich (ACS, Lima, Perú).

The selected fruits, M. dubia (intermediate ripening) and M. indica (mature), were washed, disinfected by immersion in water with sodium hypochlorite at 50 ppm and manually, with the help of a knife, the peel and seed were removed. Then independently chopped pulps were homogenized with a blender for 2 min, and subjected to heat treatment boiling temperature for 10 min, then cooled until 5°C, using ice water. The pulps were stored at 5°C until they were used for ice cream preparation.

In milk, the pH was analyzed using a potentiometer (Hanna Instruments, HI991001, serial N° B40848) and the density using a lactodensimeter calibrated at 15 °C, according to the methodology described by Rodrigues et al. (2020). Likewise, the methodology described by da Silva et al. (2020) to analyze pH using a potentiometer (Hanna Instruments, HI991001, #B40848) and total soluble solids through a refractometer (Yhequiomen, RHB0-90) at 20°C in Myrciaria dubia and Mangifera indica pulps.

The four ice cream treatments were prepared using the ingredients indicated in Table 2 and the percentages considered for each pulp (15% and 20% of M. indica, as well as 5% and 10% of M. dubia) were based on the study carried out by Barrios Renteria et al. (2022). The ice cream preparation was carried out according to Espinoza et al. (2020). First, the milk powder (2%) was dissolved in fresh milk, it was heated up to 75°C and the premixed dry ingredients (sugar and CMC) were added to it with constant stirring to avoid agglutination until the mixture reached 80°C. After constant stirring, the mixture was pasteurized at 80°C for 45 seconds, cooled until 5°C using ice water and kept refrigerated (at 5°C) for 24 h. Later, the corresponding M. indica and M. dubia pulps were added for each treatment. The churning process was carried out in an ice cream machine at −8°C for 20 min. The soft ice cream was packed in polystyrene containers (100 g capacity) and stored at −22°C for hardening (Figure 1) and subsequent determination of the ascorbic acid content, total phenolic content, antioxidant capacity and sensorial analysis was completed.

The determination of the ascorbic acid content was carried out by reduction of 2,6- dichlorophenolindophenol according to the method described by Hung and Yen (2002). Approximately 2 g of ice cream was mixed with 4 mL of 0.4% oxalic acid, the mixture was filtered through Whatman No. 2 paper and a solution was obtained. Then 200 μL of the solution were reacted with 1,800 μL 2,6 dichlorophenolindophenol and the absorbance values were recorded in a UV-Visible spectrophotometer (Genesys, S-150, 6287015), at 520 nm wavelength. For AA quantification, Equation (1) was used

The control absorbance was obtained by the reaction of 200 μL of 0.4% oxalic acid, with 1,800 μL of 2,6-dichlorophenolindophenol.

The total phenolic content was determined according to the method reported by Cornelio-Santiago et al. (2019). To obtain the extracts, approximately 2 g of ice cream were weighed, 2 mL of methanol were added in a 10 mL capacity tube, and then they were homogenized in a Vortex mixer at 2,500 rpm for 20 min. Additionally, it was centrifuged at 4,500 rpm for 20 min and the supernatant was removed. With the residue obtained, the aforementioned steps were repeated, obtaining a second supernatant. The two supernatants were mixed, and 1 mL of distilled water was added, finally centrifuged at 2,500 rpm for 2 min, the methanolic extract (ME) was refrigerated until use.

Later, 1.364 mL of distilled water were mixed in a test tube with 0.3 mL of methanolic extract, which reacted with 136 μL of Folin-Ciocalteu reagent, the resulting solution was left to stand at room temperature for 8 min in a dark room. Then, 1.2 mL of 7.5% Na₂CO₃ were added, the reaction was completed after standing for 2 h in the dark at room temperature, observing a dark blue color. Absorbance was recorded on a UV-Visible spectrophotometer (Genesys, S-150, 6287015) at 760 nm. The blank was prepared in the same way, replacing the methanolic extract with 0.3 mL of methanol and water (80:20, v/v). The TPC was calculated using the gallic acid calibration curve at the concentrations of 10, 20, 30, 40, 50 and 60 mg/L.

The antioxidant capacity was determined by the method of 1,1-diphenyl-2-picrylhydrazyl picrilhidrazil (Tian et al., 2018). The array comprised the combination of 100 uL of ME obtained with 2 mL of DPPH solution (0.8 mmol/L, solution in absolute ethanol), the reaction was completed after 10 min of rest at room temperature, the absorbance was read at 517 nm using a UV-Visible spectrophotometer (Genesys 150, 6287015). The percentage of inhibition was calculated using Equation (2).

To express in IC₅₀, the same test is carried out working with different concentrations of the extract (120, 100, 50, 25, 12, 6 mg/ml). The standard reaction required the combination of 100 uL of 80% methanol with 2 mL of absolute ethanol.

Where:

Abs control: absorbance of the DPPH solution at 0.8 mmol/L.

Abs sample: sample absorbance after 10 min.

The sensorial evaluation was carried out by 40 semi-trained panelists (10 women and 30 men) whose ages ranged from 18 to 73 years. The evaluation took place between 10:00 a.m. and 12:30 p.m. inside the Functional Foods and Bioprocesses laboratory at the Universidad Nacional de Frontera. Ice cream samples between 80 and 100 g were delivered in disposable plastic containers at −5°C, previously coded with 3-digit random numbers. The four ice cream treatments and commercial mango ice cream were subjected to a 4-attribute preference or ranking test (flavor, texture, color and general preference) using a scale between 1 and 5 (1 = least preferred and 5 = most preferred) (Rogers, 2010; Espinoza et al., 2020).

The four ice cream treatments were prepared using a 2² completely randomized two-factorial design with two variables (M. indica and M. dubia) and with two levels in each case (15% and 20% of M. indica, as well as 5% and 10% of M. dubia) and as control, a commercial mango ice cream was used. The statistical model proposed according to Montgomery and Runger (2010) is the following:

Where, Yij denotes any of the observations, μ is the Population mean, αi is the Effect of the i-th α (different M. indica ice cream formulations), Bj is the Effect of the j-th b (different M. dubia ice cream formulations), (αβ)ij is the interaction effect of factor α and factor β (different formulations of M. indica and M. dubia ice cream), εij is the Error effect.

The values were collected in triplicate and the data obtained from each the experiment were subjected to an analysis of variance (ANOVA), and Tukey's multiple comparison test, at a confidence level of 95%. The effects of the addition of M. indica pulp and M. dubia on the response variables studied were also identified. Statistical analyses were performed with STATGRAPHICS software (StatPoint, Inc. v.16.1.03). The curves were drawn by Microsoft Excel MSO (version 2019; 16.0.10390.20024; Microsoft Office Standard 2019).

Table 3 shows the results of the routine analyses carried out on the fruit pulp and the milk used to make the ice cream. The density and pH values of the milk complied with the quality limits for fresh milk established by Peruvian regulations (Instituto Nacional de Calidad, 2016).

The pH and the content of sugars in the fruits are important factors to take into account when carrying out the treatments, since they directly influence the sweet and acid taste of the product. It is observed that the pH of M. indica was higher than that reported by previous studies, which were between 3.8 to 4.7 (Musyimi, 2013). The content was between 10 to 20% higher for ripe M. indica (Lawson et al., 2019). Regarding M. dubia, higher values of pH (3.24) and soluble solids (7.20 °Brix) have been reported (Rodrigues et al., 2020) to those obtained in this study. These variations are attributed to the variety of the fruit, the weather conditions and agricultural practices (Bekele et al., 2020).

Table 3 shows the results of pH and soluble solids of the formulated ice creams. The pH values ranged between 4.24 and 4.53, while the soluble solids were between 16 and 18 °Brix. The control ice cream had a very close pH value (4.11); however, the soluble solids content was much higher (25 °Brix). The pH values were close to those reported in similar studies on ice cream with the addition of grape by-product powder (pH = 4.52–4.54) (Nascimento et al., 2018). In ice creams with the addition of 20 to 61% of M. dubia pulp, the soluble solids were around 24 to 51.30%, while the pH values ranged from 3.71 to 4.42; the higher the % of M. dubia, the lower the pH (Grigio et al., 2019). In the same way, an ice cream made with 50% soursop pulp, presented a soluble solids content of 22.70 °Brix (Virgen-Ceceña et al., 2019). It is important to consider that the high acidity and consequently a low pH value, generates higher viscosity, makes it difficult to shake, affects the stability of the mixture and generates protein denaturation, which affects the texture of the ice cream (Ismail et al., 2020).

Table 4 shows the results of the ascorbic acid content of the four ice cream treatments and the control. Ascorbic acid (AA) content was found in the range of 42.33–83.46 mg/100 g of ice cream; values well above the control (1.61 mg/100 g of ice cream). In the main effects analysis for AA content in ice cream treatments (Figure 2), it was evidenced that M. dubia pulp (PMd, %), M. indica pulp (PMi, %) and the interaction PMi x PMd significantly influence the AA content. Basically, those samples whose content in M. dubia was higher (T2 and T4), presented an increase of 98% and 69% in the content of ascorbic acid, respectively. This behavior was attributed to the high AA content of M. dubia between 1.7 to 6.11g/100 g of fresh pulp (Santos et al., 2022), higher content than that shown by M. indica (4.10–92.8 mg/100 g) (Gentile et al., 2019; Qureshi et al., 2020; Gupta et al., 2022). In addition, the AA present in the pulp of M. indica could be degraded by the pasteurization treatment. It has been reported that high temperature conditions favor the oxidation of ascorbic acid to dehydroascorbic acid, followed by hydrolysis to 2,3-diketogulonic acid and, finally, polymerization to other nutritionally inactive components (Bekele et al., 2020), which would explain the little significance of the addition of M. indica on AA in the different ice cream formulations. In a study carried out on frozen popsicles, 27% of M. dubia pulp was added, finding 84 mg of AA per 100 g of sample. However, in the present study, 17% less pulp was used to achieve a similar concentration of AA (T2), attributable to the origin of the fruits and possibly AA degradation by various factors (temperature, pH, enzymes, light and oxygen) (Grigio et al., 2019). Similarly, the content of ascorbic acid in yogurt and frozen dessert using 15% and 50% soursop pulp was between 11.25 and 9.84 mg/100g. In each case, a content of AA below our findings, the consumption of 100 g of frozen dessert would only provide between 10 and 13% of the Dietary Reference Intake (DRI) of AA for men and women. However, the T2 treatment of the present study would provide between 93.3% and 100% of the DRI, respectively (Virgen-Ceceña et al., 2019).

The TPC is shown in Table 4. The TPC for all treatments ranged from 41.23 to 80.10 mg GAE/100 g of ice cream, values higher than those found in the control (3.64 mg/100 g of ice cream). All treatments were statistically different. T2 and T4 were the treatments with the highest phenolic content. The low TPC of the control may be due to the low mango incorporation and the few phenolic compounds present in the milk from the cows' diet. However, the heat treatment leading to the Maillard reaction could favor the increase in TPC (Pandey et al., 2021).

According to the analysis of the main effects in Figure 3, the PMd (%), PMi (%) and PMd x PMi, affect the TPC of the different treatments, the addition of PMd exerts the greatest influence on the TPC in the treatments of ice cream depending on the added doses. Studies reported an increase in TPC (0.7 to 4.4 mg GAE/g) as the percentage of substitution of wheat flour by powder of by-products of M. dubia increased (0 to 20%) in cookie production (das Chagas et al., 2021). Similar behavior was found in yogurt enriched with M. dubia seed extract, where the sensorially accepted sample (0.25 g/100 g of extract) with respect to the TPC, showed about 20 mgGAE/100 g, values exceeded by all treatments of the present study (Fidelis et al., 2020). In the same way, the incorporation of black carrot concentrate (7.5%), produced ice creams with TPC of 513.67 mgGAE/100 g. It is highlighted that the carrot concentrate presented 30 times more TPC than the fresh roots, which supports the high TPC values shown in the ice cream (Pandey et al., 2021). In the literature, the degradation of TPC is observed during storage at −22 °C, after 30 and 72 days, reducing between 10 and 80% of TPC in samples of ice cream enriched with herbal extracts (mate and lemon balm), which suggests a short life span for this type of product (Gremski et al., 2019).

The antioxidant capacity (AC), is an important parameter that determines the potential to sequester and extinguish free radicals. The AC values found for the different ice cream treatments, expressed in IC₅₀ (mg/mL), were higher than 60 mg/mL, as seen in Table 4. The control sample exhibited the lowest DPPH radical scavenging power, requiring 2.5 times more samples compared to ice cream formulations with M. indica and M. dubia pulps.

Consequently, all the treatments were statistically different. According to the Pareto diagram for the main effects in Figure 4, both PMd, PMi, PMd x PMi showed a significant effect at 95% confidence, in which the greatest influence was exerted by the addition of PMd, and to a lesser extent the PMd x PMi interaction, M. indica by itself, showed a minor influence. Similar results have been observed in ice cream made with the addition of black carrot concentrate (CZ), in which the higher the CZ, the CA increased (Pandey et al., 2021). In the same way, the addition of M. dubia seed extract in a dairy product (Fidelis et al., 2020). The values of the correlation analysis between TPC vs CA and AA vs CA are shown in Figure 5. As can be seen, the P-value is 0.000 < 0.05; demonstrating that there is a relationship between TPC and AA with the CA of the ice cream. The correlation for both variables (TPC vs. CA and AA vs. CA) was negative; because a higher content of these compounds led to a lower antioxidant capacity value (expressed as IC₅₀). These findings are consistent with the results of ice cream made with hazelnut milk, which showed a positive correlation between TPC and CA (Atalar et al., 2021). Study shows antioxidant effects of an ice cream using cocoa, hazelnut extract and green tea. After the consumption of the ice cream, a reduction in oxidative stress and a greater availability of nitric oxide were observed, which improves the vascular function, due to the influence on arterial dilation and increased physical performance of healthy subjects. From this approach it is proposed that the synergy of different ingredients improves the antioxidant capacity of the final product, unlike using a single source of antioxidants. Another aspect to consider is that the bioavailability of phenolic compounds and other bioactive compounds from natural sources is greater than that of supplements (Sanguigni et al., 2017).

Figure 6 shows the results of the sensory evaluation of the ice cream. The organoleptic characteristics for the different attributes such as taste, texture, color and overall preference show significant differences between the different treatments. Two of the ice creams (T3 and T4) were the most preferred by the panelists. On the other hand, T1 and T2 showed no differences with respect to the control ice cream. In terms of taste, T1 and T2 ice creams were comparable to the control; however, T2 and T4 ice creams were perceived as less sweet than the others. On the other hand, three of the ice creams (T1, T2, and T3) were significantly milder than the control and T4. And as for the predominant color in the four ice creams developed, it corresponded to the natural color provided by the fruits. Therefore, the control ice cream received a better assessment by showing greater intensity of the yellow-orange color.

It was observed that the appreciation of the sweet taste of ice cream tends to decrease when the amount of M. dubia pulp is increased from 5 to 10%. This effect may be due to the increase of the characteristic sour taste of the fruit. A similar effect was demonstrated by Grigio et al. (2019); however, higher sweetness scores were also linked to higher sugar addition.

Ice creams containing 20% (T1 and T2) to 25% (T3) of added fruit showed a softer texture. Despite the fact that no milk fat is added in the development of the ice creams, the creaminess of the ice creams (T1, T2, and T3) was superior to the control ice cream. The fruit pulp contributes to improving the texture of the ice cream. Moreover, it intensifies the flavor and expands the smell of the ice cream, even during storage (Kowalczyk et al., 2021). On the other hand, the addition of CMC played an important role in the stabilization of the ice cream texture, as it simulates the physical properties of fat provided by its water retention capacity, gel formation and thickener function (Regand and Goff, 2002). The use of hydrocolloids in ice cream has improved the inhibition of recrystallisation, reduced the melting speed, improved the stability of the ingredients in the mixture, increased the viscosity and the overrun rate (Kamińska-Dwórznicka et al., 2022).

Probably the presence of yellow-orange coloring in the control ice cream led to better evaluations by the panelists. However, there was a greater preference for the ice creams with a higher percentage of added M. indica. Consequently, the ice cream with 20% M. indica and 5% M. dubia (T3), was the best rated against the other ice creams. The interaction of sweet-acid taste in this ice cream was very pleasant, which even surpassed the control in some organoleptic characteristics.