Raw materials

Sweet whey was provided by AURALAC S.A. (Antioquia, Colombia). Products such as sucrose, stabilizer, and commercial flavoring were supplied by Tecnas S.A. (Medellín, Colombia). Cape gooseberry fruits (Physalis peruviana L.) were purchased from a local supplier and selected in compliance with resolution 3929 of 2013 of the Ministerio de Salud y Protección Social de Colombia. The fruit selected had a state of maturity based on color of 4, 5, and 6 (NTC 4580).

The whey presented the following physicochemical characteristics: Acidity: 0.09±0.006% lactic acid (AOAC 947.05), pH=6.53±0.067 (NTC 4592), soluble solids 6.97±0.239 °Bx (AOAC 932.12), lactose 47.9±0.273 g L^-1 protein 0.85±0.066% (AOAC 990.03), ashes 0.54±0.038% w/w, calcium 366.64±46.817 mg kg^-1 (NTC 5151), magnesium 72.62±10.153 mg kg^-1 (NTC 5151). The lactose was quantified by high-performance liquid chromatography (HPLC) using a chromatograph AGILENT TECHNOLOGIES 1200 series with an AMINEX HPX-87H ion exchange column (300x7.8 mm), and as a mobile phase solution of H₂SO₄ 0,008 N at a constant flow of 0.6 mL min⁻¹ (^{Pérez-Escobar et al. 2020}). The cape gooseberry whole pulp presented the following physicochemical characteristics: total soluble solids of 15.17±0.10 °Bx (NTC 4580), acidity of 1.636±0.095% citric acid (NTC 4580), pH of 3.46±0.01 (NTC 3651), phenolic acids of 84.14±4.83 AGE 100 g^-1, total carotenoids of 33.59±1.278 mg kg^-1 antioxidant capacity with ABTS 53.6±5.11 mg trolox 100 g^-1 and DPPH 38.73±5.05 mg Trolox 100 g^-1. In addition, products such as sucrose, stabilizer (CMC FGHH, provided by Tecnas S.A.), and commercial flavoring (Passion Fruit MN-Y, provided by Tecnas S.A.) were provided by local companies in the city from Medellin.

Ultrafiltration permeate (PUF)

Sweet whey was conditioned at 43 °C, skimmed by centrifugation at 8,000 rpm (GEA Westfalia Separator AG, Type: MTA5-00-104), and pasteurized at 63 °C for 30 min. Subsequently, the product was subjected to an ultrafiltration process (Pilot filtration equipment, PERINOX, Series 0114/0115, Model E0FT) at 48 °C, with a concentration factor of 18, pressure and outlet pressure of 1 and 3 bar, respectively. The ultrafiltration pilot plant was equipped with a polyethersulfone spiral semipermeable membrane with a cut size of 10 kDa. The whey permeate by ultrafiltration had the following physicochemical characteristics: Acidity of 0.081±0.004% lactic acid (AOAC 947.05), pH of 66.47±0.15 (NTC 4592), soluble solids of 5.50±0.05 °Bx (AOAC 932.12), lactose of 50.052±0.134 g L^{- 1} (^{Pérez-Escobar et al. 2020}), ash of 0.481±0.032% w/w (NTC 5151), protein <2.5% (AOAC 990.03), calcium of 283.063±27.472 mg kg^-1 (NTC 5151), magnesium of 61.813±5.256 mg kg^-1 (NTC 5151).

Hydrolyzed PUF

Hydrolyzed permeate was obtained by cold hydrolysis using 0.075 mL L^-1 of lactase for 20 h, achieving a degree of hydrolysis of approximately 30%. The percentage of hydrolysis was measured using the cryoscopy point method described by ^{Llerena et al. (2019)}.

Preparation of cape gooseberry (skin and pulp)

The whole fruit was cleaned, disinfected (sodium hypochlorite at 100 mg L^-1), and homogenized using an industrial blender (CI TALSA LI30). The resulting product was passed through a USA Standard Mesh (18 mesh=1 mm sieve size) to remove the seeds. The processed fruit was stored frozen at -18 °C for later use.

Preparation of beverages

For the preparation of beverages, preliminary experiments were conducted varying the percentage of PUF and seeking to obtain an osmotic concentration following the regulations for isotonic beverage in normative 2229 of 1994 of the Ministerio de Salud de Colombia. Two formulations were developed: non-hydrolyzed isotonic beverage with cape gooseberry fruit (BIUN) and hydrolyzed isotonic beverage with cape gooseberry fruit (BIUH), formulated with non-hydrolyzed and hydrolyzed PUF, respectively (Table 1). Beverages were prepared by mixing the ingredients and homogenizing at 10.34 MPa (1,500 psi) (Homogenizer, St. Regis CP Division, Series: 3DD13-2941), then pasteurized (62 °C for 30 min), bottled in 250 mL PET flasks, and stored at 4 °C.

Table 1 Formulation of isotonic beverages. 

Physicochemical analysis

Osmolality was determined by cryoscopy procedure and following the NTC 3837 given by ^{ICONTEC (2009)}, as shown in equation 1:

where ∆T is the difference (°C) between the freezing temperature of water and the freezing point of the sample, ∆T was measured with a FUNKE GERBER CryoStar I single sample automatic cryoscope (DIN/ISO/IDF 5764). Determination of pH was measured by potentiometric method, after calibration with buffer solutions of pH=4 and 7 at 20 °C with an OHAUS STARTER 3100 potentiometer (AOAC Method 981.12/90). Acidity was measured by titration (AOAC Method 947.05/90). Total soluble solids (TSS) were expressed as Brix degrees (°Bx) and quantified using a digital refractometer (HI 96801) and following the NTC 4624 given by ^{ICONTEC (1999)}. Mineral content (calcium, magnesium, potassium, and sodium) was obtained by atomic absorption spectrometry (AOAC 985.35-1988), and total sugars by spectrophotometry UV-VIS (^{Nielsen 2010}). Antioxidant capacity was determined using a modification of the DPPH and ABTS radical cation trapping methods proposed by ^{Bravo et al. (2015}). Extraction: 10 g of each sample was taken in a falcon tube and mixed up to 20 mL with solvent (methanol/water, 70:30). The sample was vortexed at 30,000 rpm for 2 min and put in an ultrasonic bath for 10 min, then centrifuged at 8,000 rpm for 15 min. The supernatant was filtered into 25 mL volumetric flasks, completing the volume with solvent. DPPH was determined with 50 μL of the extract were taken in an Eppendorf tube, 950 μL of work solution (0.05 mM DPPH solution) were added, it was stirred at 2,000 rpm in a vortex for 30 s and left to react in darkness for 30 min. The absorbance of the extract was measured at a wavelength of 517 nm using a spectrophotometer (THERMO Scientific, Evolution 60). ABTS was determined with 50 μL of the extract were taken in an Eppendorf tube, 950 μL of work solution (ABTS solution) were added, it was stirred at 2,000 rpm in a vortex and left to react in darkness for 8 min. The absorbance of the extract was measured at a wavelength of 734 nm using a spectrophotometer (THERMO Scientific, Evolution 60). Total phenol content was determined using a modification of the Folin-Ciocalteu reagent method described by ^{Bravo et al. (2015)}. Finally, 100 μL of the extract were taken in an Eppendorf tube, 400 μL of 0.07 N sodium carbonate solution were added, it was stirred at 15,000 rpm in a vortex, and allowed to stand for 5 min. Then, 500 μL of Folin solution were added and stirred at 15,000 rpm in a vortex. The tube was capped and stored in darkness at room temperature for 2 h. The absorbance of the extract was measured at a wavelength of 760 nm using a spectrophotometer (THERMO Scientific, Evolution 60). Total Carotenoid content was determined using a modification of the method described by ^{Ferreira et al. (2009}). A 15 g sample was taken in a falcon tube and stirred with 15 mL of solvent (n-hexane/acetone, 6:4) for 10 min at room temperature and filtered through Whatman No.4 filter paper. The absorbance of the filtrate was measured at a wavelength of 450 nm using a spectrophotometer (THERMO Scientific, Evolution 60).

Microbiological analysis

Counts of total coliforms (AOAC 966.23C 2005), fecal coliforms (AOAC 966.24 2005), aerobic mesophiles (AOAC 988.18 2005), molds and yeasts (AOAC 995.21 2005), and Clostridium sulfite reducer spores (AOAC 972.41 2005) were conducted following the NTC 3837. Analyses were made for two months on beverage samples (BIUH and BIUN) for storage times of 4, 19, 39, and 54 days.

Sensory analysis

The sensory profiles of the beverages were evaluated by the multidimensional approach method, following the Colombian Technical Standards NTC 3501 (^{ICONTEC 2012}), NTC 3932 (^{ICONTEC 1996}), and the Colombian Technical Guides GTC 165 (^{ICONTEC 2014}) and GTC 226 (^{ICONTEC 2012}). A set of relevant descriptors were identified and selected; the intensities were assessed by five trained judges, with an age range between 25 and 60 years. The samples were evaluated under the temperature of 23 °C and relative humidity of 61% (room conditions), and the rating scale from 0 to 5 for all the descriptors was established. Overall quality was rated on a scale from 1 to 3, where 3 is high and 1 is low. Student´s t-test for independent samples was used for statistical analysis.

Statistical Analysis

The study was conducted using a completely randomized design with three replicas. The physicochemical properties of each treatment were followed for two months (days 0, 7, 14, 21, 28, 35, 42, 49, and 56). Bioactive compounds were measured at weeks 0, 4, and 8. Results were analyzed using one-way analysis of variance (ANOVA) with P<0.05. Means were compared by the least significant difference method (LSD, P<0.05), using Minitab® 19.1.1 (Minitab, LLC., 2021).

Mineral and sugar content

Table 2 shows the results of the proximate analyses. There were statistically significant differences (P<0.05) in the content of calcium and magnesium between the two beverages -BIUN and BIUH. This difference may be due to the interaction of calcium and magnesium ions with the enzyme used for the hydrolysis of the BIUH beverage (^{Zolnere et al. 2017}), as well as the effect of additional heat treatment applied to the beverage to inactivate the enzyme (^{De La Fuente et al. 1999}; ^{Rojas Silva 2016}).

Table 2 Content of sugars and minerals in BIUN and BIUH (100 g of beverage). 

^{Ferreira et al. (2020}) prepared a beverage based on whey permeate and jaboticaba peel, reporting similar values of total sugars (4.8 g), calcium (34.1 mg), and magnesium (6.2 mg). ^{Hattem et al. (2010}) prepared a sports drink from whey permeate added to mango and obtained similar contents of total sugars (6.48 g), calcium (25.1 mg), and magnesium (5.2 mg). In this way, it can be observed that minerals such as magnesium, calcium, potassium, and sodium provided by the sweet whey permeate satisfy the requirements for an isotonic beverage according to Colombian regulations NTC 3837.

Physicochemical stability

According to Figure 1, the osmolality of the hydrolyzed beverage (BIUH) was higher (P<0.05), this is due to the lactose hydrolysis process increasing the concentration of solutes in the beverage. An increase in the osmolality of both beverages was observed over time (P<0.05): from 304.83 to 324.13 mOsm kg^-1 for the BIUN beverage and from 330.1 to 350.53 mOsm kg^-1 for the BIUH beverage. This increase is due to the usual hydrolysis of sucrose and other more complex carbohydrates present in the samples during storage (^{Sollanek et al. 2019}). ^{Ferreira et al. (2021}) obtained similar values to BIUN in osmolarity for an isotonic beverage made from whey permeate and pequi powder.

Figure 1 Osmolality stability during storage at 4 °C (confidence intervals). 

There were no statistically significant differences (P >0.05) in the beverages during the storage time to total soluble solids (TTS). Figure 2 shows that during the first week, there was a slight increase in TSS in the drinks (between 3 and 4%). This can be attributed to the partial enzymatic hydrolysis of complex carbohydrates present in the samples (^{Naik et al. 2009}). In the second week, the values remained stable between 6.9 and 7.1 °Bx for both drinks.

Figure 2 Total soluble solids (TTS) stability during storage at 4 °C (confidence intervals). 

According to Figure 3, there were no statistical significance (P>0.05) in the lactic acid content in both beverages nor in storage time. It can be seen that the acidity remained around 0.11% v/v lactic acid during the storage time. ^{Naik et al. (2009}) reported comparable results in a whey drink added with watermelon pulp, the slight increase in acidity after 30 days of storage was attributed to the conversion of lactose into lactic acid.

Figure 3 Acidity stability during storage at 4 °C (confidence intervals). 

There were statistically significant differences in the pH values during the storage time (P<0.05) but not between the two treatments (Figure 4). The variation in pH can be attributed to the formation of lactic acid during storage; it was observed that the pH of the beverages decreased over time (1.1%). These results differ slightly from those reported by ^{Hattem et al. (2010}), when studying sports drinks based on whey permeate and mango and strawberry pulp, pH decreased by around 5% in 15 days of storage at 4 °C. ^{Elsabie and Aziz (2011}) found a pH decrease of 10% in a beverage based on permeate and sweet potato after three weeks of storage at 7 °C. These differences may be due to better hygiene conditions during the preparation of the beverages and to lower storage temperature.

Figure 4 pH stability during storage at 4 °C (confidence intervals). 

Bioactive compounds

There were no significant differences (P >0.05) in the content of total phenols between the two treatments (Figure 5). The BIUN beverage presented an average content of 9.64 mg-GAE 100 g^-1, while the average content in the BIUH beverage was 9.72 mg-GAE 100 g^-1.

Figure 5 Total phenols stability during storage at 4 °C (confidence intervals). 

^{Porfírio et al. (2020}) developed a soft drink with jaboticaba (Plinia cauliflora) extract obtaining a 5-fold greater amount of total phenols (48.56 mg-GAE 100 g^-1), while ^{Atallah (2015}) reported a beverage with mango pulp with 50% less total phenols (4.32 mg-GAE 100 g^-1) than the developed beverages.

Statistically significant differences (P<0.05) were found in the total carotenoid content of the BIUH beverage throughout the storage time (Figure 6). However, there was no effect due to the hydrolysis treatment of the beverages (P>0.05). A decrease in the content of total carotenoids is observed during the study, this may be due to the fact that these chemical compounds are susceptible to oxidation and isomerization reactions during storage, due to their highly unsaturated structure (^{Ferreira et al. 2021}). The total carotenoid content decrease in the non-hydrolyzed beverage was about 20%, while the hydrolyzed beverage showed a decrease close to 34%.

Figure 6 Total carotenoid stability during storage at 4 °C (confidence intervals). 

There were no significant differences (P>0.05) in the antioxidant capacity, for both DPPH (Figure 7A) and ABTS radicals (Figure 7B) between the two treatments. During storage under refrigeration, there was a decrease of approximately 30 and 12% for DPPH and ABTS radicals, respectively, after the first 4 weeks; between weeks 4 and 8 there were no differences. The decrease in the antioxidant capacity may be due to the relatively low pasteurization temperature, since it may have not been enough to sufficiently inactivate the enzymes of the food matrix, leading to an alteration of the nutritional properties (^{Santander et al. 2017}).

Figure 7 Antioxidant capacity during storage at 4 °C: A. Antioxidant activity by DPPH radical; B. Antioxidant activity by ABTS radical (confidence intervals). 

The results found are similar to those reported by ^{Santander et al. (2017}), who found that the antioxidant capacity of a whey-based beverage added with tamarillo (Solanum betaceum) decreased gradually during storage at 4 °C as storage time increased, and after one month the loss of antioxidant capacity was close to 30%. Other studies such as that of ^{Ferreira et al. (2021}) have found that the addition of fruits enriched these types of drinks with nutraceutical components such as antioxidants.

Microbiological stability

According to Table 3, the beverages met the microbiological requirements established in the Colombian technical standard NTC 3837/2009 and normative 2229 (1994) of Ministerio de Salud de Colombia. This shows that the use of LTLT (low temperature/long time) pasteurization as heat treatment was sufficient to ensure the safety of both beverages.

Table 3 Microbiological results for isotonic beverages (BIUH and BIUN). 

Sensory analysis

A set of relevant descriptors that give the maximum information on the sensory attributes of beverages were identified and selected. Aiming to establish a sensory profile, the intensities were assessed by trained judges on a rating scale from 0 to 5 for all the descriptors. Overall quality was rated on a scale from 1 to 3, where 3 is high and 1 is low. Figure 8 shows the sensory profile of the samples.

Figure 8 Sensory profile by multidimensional approximation for BIUN vs. BIUH samples. 

The BIUN beverage sample presented a fruity flavor of cape gooseberry, yellow fruits, sweet granadilla, and passion fruit, and a balanced flavor. In the first olfactory phase, fruity notes of passion fruit, and dairy, a fruity smell similar to soursop were perceived. The fruity smell was more intense than the taste. The beverage was of high overall quality and had a good balance of notes. The BIUH beverage sample presented fruity notes of cape gooseberry, granadilla, passion fruit, and a slight barnyard flavor that persists over time. The olfactory perception of the fruity attribute was higher than the taste; metallic taste, residual milky notes, and fewer astringent sensations were perceived. Statistically significant differences (P<0.05) were found in the attributes of sweet and sour aromas, fruity flavor, and mouth coating, thus obtaining a higher score for the attributes described above in the non-hydrolyzed isotonic beverage sample. ^{Beucler et al. (2006}) and ^{Nemati et al. (2020}) compared sensory attributes of hydrolyzed and non-hydrolyzed whey permeate and found significant differences in acid, sweet, and astringent taste.