Introduction

Some efforts have been attempted in recent years to incorporate selected fruit and vegetable into a blend to improve its nutritional and sensorial features, (Kahraman et al. 2017), while lowering the sugar and caloric content. In particular, carrots are a valuable source of natural antioxidants such as carotenoids, phenolics, vitamin C, and tocopherols (Ferrario et al. 2017). Moreover, orange juice possesses high levels of vitamin C, phenolic compounds, and flavonoids, which inhibit oxidative stress. Its intake has been related to antioxidant and anti-inflammatory activities (Mesquita and Monteiro 2018).

Yeasts are usually the main contaminants of naturally acidic foods like most fruits, being Candida, among other yeasts and molds, one of the most frequent genera isolated during bottling in the beverage industry (Tomičić et al. 2017). Thermal processing ensures an adequate shelf-life and stability of fruit and vegetable juices. However, this method may negatively affect sensorial, nutritional, and physicochemical quality (Petruzzi et al. 2017). In the last two decades, the food industry and academy have given increasing attention to non-thermal technologies as novel alternatives for ensuring the safety of food and beverages while maintaining or even improving their nutritional and sensorial quality as well (Knorr 2018). In particular, UV-C light is considered a promising alternative for the decontamination of freshly pressed juices. UV-C comprises the spectrum range from 200 to 280 nm, resulting in lethal effects to several microorganisms without generating significant toxic and non-toxic by-products (Tremarin et al. 2017), while also being energy efficient and low cost (Paniagua-Martínez et al. 2018). Exposure to UV-C light triggers the cross-linking of neighboring pyrimidine bases in the same deoxyribonucleic acid (DNA) strand, eventually leading to cell death (Nocker et al. 2018). The implementation of UV-C for food preservation purposes has been limited until now to clear and transparent liquids. The presence of suspended matter, colored compounds along with high turbidity of certain liquid foods such as juices or milk, negatively impacts on the UV-C treatment effectiveness (Müller et al. 2015). To compensate for these limitations, new UV-C light-based processes using the hurdle approach have been developed, which include the assistance of traditional or emerging factors to UV-C. Most published studies have been focused on the inactivation of native flora, foodborne pathogens, and/or some spoilage microorganisms, as well as the physicochemical changes after exposure of different juices to the combined UV-C plus mild heat treatment (Gouma et al. 2020; La Cava and Sgroppo 2019; Santhirasegaram et al. 2015). However, a few of them have analyzed the impact of UV-C combined with mild heat on the sensory quality of liquid foods (Ferrario et al. 2018; Vásquez-Mazo et al. 2019). Moreover, no studies have been reported up to now comparing the sensory quality of juices processed by UV-C light assisted by heat to those that have been pasteurized.

The mechanisms associated with cell inactivation, including the physiological changes that may occur due to UV-C exposure should be explored in deepen to better design this technology as an alternative for juice processing. When using milder food preservation techniques, it is important to detect the presence of sublethally damaged cells. These subpopulations may recover and grow under certain favorable conditions, thus representing a severe risk for food safety and/or quality (Schottroff et al. 2018). Accordingly, the use of flow cytometry (FC) has proved to be useful for the assessment of microbial physiological status on a single cell level analyzing large populations of cells in a very short time. This goal can be achieved by adding fluorescent dyes, which are targeted on specific cellular structures, physiological status/activity (Cao et al. 2015; García Carrillo et al. 2018). In particular, regarding FC in the study of the viability of Candida spp., no information has been reported on the physiological state of this spoilage yeast in juices processed by emerging technologies. Previous studies have focused on cell adhesion assays (Silva-Dias et al. 2012), and assessment of resistance to antifungals (Da Silva et al. 2016), but only a few of them have evaluated its viability after food and beverage processing (Petruzzi et al. 2017).

The present study was aimed to evaluate the inactivation and physiological state of Candida parapsilosis ATCC 22019 in a turbid carrot-orange juice blend processed by a continuous UV-C treatment with the assistance of mild heat (UV-C/H, 50 °C, 1720 mJ/cm2) by flow cytometry analysis. Additionally, the evolution of total polyphenol content, antioxidant activity, color, pH, °Brix, turbidity, and pectin methylesterase activity (PME) of the treated juice blend was assessed along refrigerated storage (4 °C). Also, a conjoint study was performed to investigate how certain claims informing about the juice characteristics, influenced consumers about product healthiness and their willingness to try it. The overall acceptability and certain attributes of the UV-C/H juice and a pasteurized juice used as a control (80 °C, 5 min) were assessed. Furthermore, a check-all-that apply (CATA) question was conducted to analyze which sensory attributes better described these juices and determine how they differed from an ideal juice.

Materials and Methods

Carrot-Orange Juice Blend Preparation

Carrots and oranges, purchased in a local market and immediately refrigerated until use, were disinfected and the juice blend was obtained following the same procedure described in García Carrillo et al. (2017). The juice blend was fractioned (745 mL) and stored in amber glass bottles and stored at − 80 °C until use.

Inoculum Preparation

Experiments were performed using Candida parapsilosis ATCC 22019. Initial yeast inoculum was prepared as stated in Fenoglio et al. (2020) using Sabouraud Dextrose Agar (SAB, Biokar Diagnostics, Beauvais, France), harvested by centrifugation (1475g, 10 min) (Labnet, USA), washed twice, and resuspended in peptone water to obtain a final cell density of 106–107 CFU mL−1.

UV-C Treatment

The UV-C device consisted of two serially connected UV-C lamps (TUV-30 W, 253.7 nm, Philips, The Netherlands), each one inside a glass tube (0.87 m-long), shaping an annular space (outer diameter, 0.031 m; inner diameter, 0.024 m and volume, 0.22 L) constituted as the irradiation chamber. For the treatments, the juice (745 mL) was recirculated (1.6 L/min, 15 min, 1720 mJ/cm2, D10 UV-dose = 3.1 mJ/cm2) into a double jacket vessel, which was connected to the UV-C chamber and to a water bath to attain 20 or 50 °C in the vessel, as depicted in García Carrillo et al. (2017). UV-C performed at 20 °C was considered a single treatment (UV-C), while UV-C assisted by mild heat (50 °C) corresponded to the combined treatment (UV-C/H). Additionally, a single thermal treatment (H) was performed at 50 °C in the same way but with lamps turned off. This value was selected based on previous studies performed in a carrot-orange juice, in which a synergistic inactivation of Escherichia coli ATCC 35218 and Pseudomonas fluorescens ATCC 49838 was observed only when applying this temperature; whereas, additive inactivation effects were recorded by applying 40 and 45 °C. Moreover, for the sensory studies, a juice blend thermally treated at 80 °C for 6 min (T-coil) by recirculation in a coiled tube (Peng et al. 2017) was used for comparison purposes. For each challenge assay, the juice was inoculated with the microbial suspension (5 mL) before the UV-C treatment, following the procedure described in García Carrillo et al. (2017), leading to an initial C. parasilopsis concentration of ~ 3 × 106 CFU/mL. Previous studies demonstrated that the average residence time (tm) was 19.8 s, being the calculated theoretical residence time (tt) of 20.6 s. Therefore, a ratio tm/tt of 0.96 was obtained (Fenoglio et al. 2020). These results indicated the effective volume of the reactor was practically the real fill volume at the used flow rate (Fenoglio et al. 2020). All assays were run in triplicate. The hydraulic Reynolds number (Reh), the delivered UV-C dose by actinometry (mJ/cm2), and the UV-dose distribution (D10 UV-dose, mJ/cm2) estimated with a microbial tracer, were previously determined and described in detail in Fenoglio et al. (2020). The Reh number was in the range from 822 to 1100, depending on the temperature used, which corresponded to laminar flow (Mansor et al. 2014). The use of UV-C devices operating under laminar flow conditions with recirculation mode had been previously validated (Kaya and Unluturk 2016). For instance, Antonio-Gutiérrez et al. (2019) reported 6 log reductions of S. cerevisiae in grape juice treated by an UV-C treatment (35.8 mJ/cm2) under recirculation mode.

Microbial Enumeration

Samples were taken at regular time intervals during treatments and serially ten-fold diluted in 0.1% w/v peptone water and surface plated onto SAB agar using a spiral plater (Autoplate 4000, Spiral Biotech, USA). Plates were incubated for 72 h at 27 ± 1 °C. Colonies were enumerated using a counting grid. When low counts were detected (less than 103 CFU/mL), 1 mL of sample was pour plated into each dish. Plots of log N/N0 (where N is the number of colony-forming units per juice milliliter (CFU/mL) at a certain time and N0 the initial number of CFU/mL) versus treatment time were obtained.

Flow Cytometry

For the flow cytometry study, FDA (Calbiochem, Darmstadt, Hesse, Germany) and PI (Sigma Aldrich, St Louis, Missouri, USA) were used as fluorochromes (García Carrillo et al. 2018). On the one hand, metabolically active cells with esterase activity can hydrolyze FDA into fluorescein (F), a polar fluorescent green compound, which is unable to diffuse out of the cells. Thus, only cells with non-injured membranes will maintain the green fluorescence caused by the presence of F. On the other hand, propidium iodide (PI) binds to DNA only in cells with compromised membranes. If membrane integrity is lost, PI diffuses into the cell and stains the DNA delivering red fluorescence (García Carrillo et al. 2018).

Positive and negative histograms gates were defined using non-treated stained cells. Heated cells (85 °C, 15 min) and stained only with PI were defined as the negative controls. Whereas, non-treated cells labeled only with FDA corresponded to positive controls. Besides, non-treated and non-stained cells were used to determine autofluorescence. Control and treated cells were incubated with FDA (0.3 μL, 5 mg/mL acetone, 37 °C, 30 min). Thereafter, cells were centrifuged twice (7880 g, 5 min) and resuspended in a saline solution (PBS buffer, 1 mL, pH 7.0). The resuspended cells were subsequently incubated with PI (0.5 μL, 1 mg/mL sterile water, 10 min 25 °C, darkness). Finally, the samples were placed on ice in darkness for a maximum of 1 h until subsequent analysis.

Doublets were detected and removed by analyzing the forward area (FSC-A) versus forward height (FSC-H) plot to ensure that only single cells were counted and analyzed as described in García Carrillo et al. (2018). Analysis of treated microbial cells was performed on a flow cytometer (BD FACSAria II, New Jersey, USA). Logarithmic signals were collected from the scatter and fluorescence of single cells traveling through the laser channel. The green fluorescence corresponding to FDA-stained cells was recorded in the FL1 channel (525 ± 15 nm). The red fluorescence corresponding to PI-labeled cells was registered in the FL2 channel (620 ± 15 nm). Flow rate and sample cell concentration were adjusted to keep constant the acquisition rate (200 cells/s), this registering up to 20,000 events per sample. All the treatments were performed in triplicate for each condition.

Juice Physicochemical Characterization Along Storage

Total phenolic content (TPC), total antioxidant activity (TAA), pH, °Brix, turbidity, and color of the control or UV-C, H, and UV-C/H processed juices were assessed after treatment and along refrigerated storage. Juice samples were aseptically distributed in 10-mL amber glass flasks and kept in the dark (4 ± 1 °C, up to 24 days). The control and single UV-C and H treatments were assessed for 13 days, while juice samples treated by the proposed UV-C/H treatment were evaluated up to 24 days. For each treatment, the storage time was selected based on previous native flora studies (Ferrario et al. 2018), which determined the maximum period of time at which the acceptable levels for aerobic mesophiles and total coliforms were below the limits established by The Health Protection Agency (HPA 2009). Three independent flasks corresponding to each treatment were taken during storage at preset time intervals (0, 1, 3, 5, 7, 9, 13) for subsequent analysis and the most representative were chosen to be plotted (0, 5, 9, 13).

The particle size of juice samples in the range from 0.1 to 1000 μm was measured by static light scattering using a Mastersizer 2000 (Malvern Instruments, Malvern, Worcestershire, UK) as stated in Ferrario et al. (2015). Juice turbidity and UV absorption coefficient were determined according to García Carrillo et al. (2017). The color of juices was measured with a handheld tristimulus reflectance spectrocolorimeter (Minolta Co. Model CM-508-d, Osaka, Japan) by using a 1.4 measuring aperture with white and black background. Three (3) mL of sample were measured using illuminant C and 2° standard observer angle. Before each measurement, the instrument was calibrated with a standard provided by the manufacturer. The CIE color coordinates (X, Y, Z) and the components of the CIELAB space, L* (lightness), a* (green-red), and b* (blue-yellow) were recorded to assess the color evolution during storage of control and processed samples as described in Ferrario et al. (2018).

The total polyphenol content (TPC) in the extracts was determined using the method detailed in Ferrario et al. (2018). The TPC values were derived from the calibration curve (y = 13.2465x + 0.0518, R2 = 0.99) and expressed as micrograms of Gallic acid equivalents per sample milliliter, μg GAE/mL. The total antioxidant activity (TAA) was evaluated by a colorimetric method based on the measure of the free radical scavenging capacity of the sample with the 2,2-diphenyl-1-picrylhydrazyl (DPPH) stable radical (Merck, Billerica, USA), as stated in Ferrario et al. (2018). TAA values were determined from the calibration curve (y = 0.00084x − 0.0024, R2 = 0.98), and expressed as milligrams of Trolox equivalents per milliliter (mg TroloxEq/mL).

Pectin Methylesterase Activity

Pectin methylesterase activity was measured according to the method described by Aghajanzadeh et al. (2016) for control and treated samples immediately after treatment and along storage. Citrus pectin was purchased from Sigma (St. Louis, MO). Juice blend and pectin-salt solution were previously incubated separately at 30 °C and adjusted to pH 7.0 with 2.0 N NaOH solution. A 5 mL aliquot of the juice blend was mixed with 20 mL of 1% pectin-salt solution (1 g pectin and 1.17 g NaCI diluted in 100 mL distilled water) and readjusted to pH 7.7 with 0.05 N NaOH at 30 °C under agitation conditions. Time to regain pH 7.7 was recorded after the addition of 100 μL of 0.05 N of NaOH. Pectin methylesterase units were expressed as PME (units/mL juice), which represents the milliequivalents of esters hydrolyzed per min per mL of juice, according to Eq. 1:

$$ \mathrm{PME}\ \left(\frac{\mathrm{units}}{\mathrm{mL}}\right)=\frac{\left(0.05\ \mathrm{N}\ \mathrm{N}\mathrm{aOH}\right)\mathrm{x}\ \left(0.1\ \mathrm{mL}\ \mathrm{of}\ \mathrm{NaOH}\right)}{\left(5\ \mathrm{mL}\ \mathrm{sample}\right)\mathrm{x}\ \left(\mathrm{time}\left(\min \right)\right)} $$
(1)

Sensory Studies

A hundred and thirty (130) unpaid volunteers (59 females and 71 males) recruited from personnel and students of Buenos Aires University, aged between 18 and 45, and selected as frequent consumers of fruit juices, participated in the sensory tests. Each panelist tested both, the UV-C/H and T-coil samples. For each sample (15 mL), they performed the following tests: conjoint study, consumer field test, and check-all-that-apply (CATA) question. The samples were presented to the consumers in a monadic sequential randomized order, with washout periods of enough elapsed time to allow the respondents to return to some baseline level of perception. The samples were served to the panelists at the ideal temperature for consumption (5–7 °C) in white plastic cups. The evaluations were run in individual booths under white light (ISO 8589:1988).

Conjoint Study

A conjoint study was conducted in order to investigate the influence of the juice blend characteristics and type of processing on consumers’ perceived healthiness and willingness to try those natural juice blends as described in Ares et al. (2009). The 22-factorial design for the displayed claims consisted of two categorical factors: 1) juice profile and 2) juice processing, each one having two levels. For the first factor (juice profile), the following levels were used: 1-natural, with no preservatives nor sugar addition, rich in antioxidants and, 0- no information was provided about the juice profile. The two levels corresponding to the second factor (juice processing) were: 1- juice processed without using high temperature and, 0- without providing information about the juice processing. Therefore, each participant randomly evaluated the picture of the juice blend having one of the four claim combinations: carrot-orange juice (control, claim 1), (0,0); carrot-orange juice processed without elevated temperature (claim 2), (0,1); 100% natural carrot-orange juice, with no preservatives nor sugar added and rich in antioxidants (claim 3), (1,0) and 100% natural carrot-orange juice, with no preservatives nor sugar added and rich in antioxidants, and processed without elevated temperature (claim 4), (1,1). All the information was presented using paper-based ballots containing an image of the juice blend along with the different claims, while the codes were only used internally. The images were printed with high resolution in glossy paper and were presented to the participants numbered with three-digit random codes. They were asked to score the perceived healthiness of the different images, using a 7-box scale anchored on the left with “not at all healthy” and on the right with “very healthy” terms, and to score their willingness to try them using a 7-box scale anchored on the left with “I would definitely not try it,” on the middle with “Maybe yes, maybe not” and on the right with “I would definitely try it” (Ares et al. 2009).

Consumer Field Test

A consumer field test form was designed according to the general recommendations cited by Lawless and Heymann (2010) for this type of test. This study was performed to collect some diagnostic information on the reasons behind consumer likes and dislikes about the samples they tested. The overall acceptability was assessed in a 9-point hedonic scale (1—dislike extremely; 9—like extremely). Secondly, the panelists were asked to answer open-ended questions for liking or disliking with an appropriate “skip pattern.” The skip pattern dropped to reasons for liking if the respondent was positive, and then probed any dislikes, and vice versa. In addition, more specific attributes were investigated through the use of intensity and just right 5-point scales as described in Vásquez-Mazo et al. (2019). The juice attributes were evaluated according to their acceptability, intensity, and/or just-about-right perception in the juice sample. In particular, the carrot taste was evaluated in a short-intensity scale with labeled anchors (1: without carrot taste; 5: very intense carrot taste). Similarly, the bitterness of the samples was evaluated in a short-intensity scale with labeled anchors (1: not bitter to 5: extremely bitter). The sour-tasting, aroma, and body (a group of mouthfeel attributes corresponding to the sensations of viscosity and density perceived in the mouth such as watery, heavy, light, etc.) of the juice samples were evaluated using just-about-right (JAR) scales, which were labeled at the left end as “not sour enough,” “too week aroma,” or “without enough juice body”; at the right end as “extremely sour,” “very intense aroma,” or “with too much juice body,” respectively; and “just right” in the middle of all scales. The lexicon and the scales were explained until the panelists fully understood. Data was analyzed by converting assigned positions into numbers. Results were reported as an average of the individual values. Responses from open-ended questions were collected and qualitatively analyzed by grouping into categories the common attributes described by the panelists.

Check-All-That-Apply Question

With regard to the CATA question, all panelists answered a questionnaire, which consisted of sensory descriptors, generated in a preliminary round table session. The selected descriptors were: pleasant color, strong color, natural taste, intense taste, artificial taste, orange taste, carrot taste, fruity, not fruity, cooked taste, too particulate (presence of too many suspended particles in the juice), taste persistence (with aftertaste), strange taste, strange aroma, pleasant aroma, pleasant juice body. Consumers were explained that they had to choose all the terms mentioned in the checklist that they considered appropriate to describe each drink (Vidal et al. 2018). The CATA counts were totaled for each product and the resulting contingency table was used in subsequent analyses.

Statistical Analysis

Multivariate analysis of variance (MANOVA) was performed to establish significant differences in TPC, TAA, pH, °Brix, turbidity, and the color functions according to the “time” and “treatment” variables, and the corresponding interaction “treatment*time.” The significance level was set at p < 0.05. Multivariate outliers were detected by Mahalanobis distance and removed from the data set. In the case of finding significant differences, post hoc multiple comparisons among multivariate means of factors were performed by the Hotelling test based on the Bonferroni correction.

A two-way multivariate analysis of covariance (MANCOVA) was performed to investigate differences for the carrot-orange juice blend in healthiness and willingness to try, using gender, age, and diary intake of fruits per day as co-variables.

An agglomerative hierarchical cluster analysis was conducted on the hedonic scores assigned by panelists to determine groups of consumers´ preferences for the carrot-orange juice using the weighted average linkage and the Euclidean distance, as stated in Ferrario et al. (2018). The overall goodness of fit was measured by the cophenetic correlation coefficient (CCC) (Lawless 2013).

For the CATA question, a correspondence analysis (CA) was applied from the frequency table, which contained all the responses regarding sensory descriptors included in the CATA question for UV-C/H, T-coil samples, and the ideal beverage.

Statistical analyses were conducted using InfoStat 2009 (InfoStat Group, FCA-UNC, Córdoba, Argentina).

Results and Discussion

Effect of UV-C, H, and UV-C/H Treatments on C. parapsilopsis Inactivation

Inactivation curves corresponding to C. parapsilosis ATCC 22019 in the carrot-orange juice blend processed by the single UV-C and H treatments along with the UV-C assisted by mild heat (UV-C/H) treatment are shown in Fig. 1. Single UV-C and H showed almost linear and slight sigmoidal inactivation responses, inducing moderate reductions of 2.9 and 3.9 log cycles after 15 min of treatment, respectively (Fig. 1). Whereas, the inactivation curve corresponding to the combined treatment (UV-C/H) exhibited a sigmoidal shape, achieving a significantly higher inactivation effectiveness (p < 0.05) compared with the single treatments as 5.5 log reductions were determined after 10 min of processing. It is important to highlight that at least more than additive effects between UV-C and H were observed, as no colonies were detected in the combined treated juice samples after 10-min exposure.

Fig. 1
figure 1

Survival curves of Candida parapsilosis ATCC 22019 in carrot-orange juice blend treated by single UV-C (■), single mild heat at 50 °C (▲), and UV-C assisted by mild heat at 50 °C (●) (maximum dose 1720 mJ/cm2), determined by the plate count method. The results are means on data from three independent experiments with error bars indicating standard deviation (I)

Full size image

Gabriel et al. (2018) evaluated the inactivation of C. parapsilosis BC1 and SC1 in calamansi juice (pH 2.93, 11.10 °Brix) exposed to benchtop UV-C (55–143 mJ/cm2, 7-mm sample thickness). In agreement with the present work, they observed up to 2.0 to 2.5 log reductions of the yeast after single UV-C exposure. On the other hand, García Carrillo et al. (2018) evaluated the inactivation response of S. cerevisiae in a carrot-orange juice blend (pH 3.8, 10.6 °Brix) after UV-C (1060 kJ/cm2) assisted by mild heat (50 °C). In agreement, they reported an increased inactivation response (4.7 log reductions) when applying the combined treatment compared with single UV-C (3.3 log reductions) and mild heat (50 °C) (2.6 log reductions). Contrasting both studies, the inactivation curves of S. cerevisiae and C. parapsilosis after UV-C processing in the carrot-orange juice displayed different shapes, as upward concavity with a tail or sigmoidal shape were observed, respectively.

Flow Cytometry Study

Figure 2 shows the density plots corresponding to C. parapsilosis cells in a carrot-orange juice showing the evolution of enzyme activity and membrane damage after exposure to the single (UV-C, H) and combined (UV-C/H) treatments. The percentages of cells located in each gate can be visualized in Fig. 2 in the four edges of each plot.

Fig. 2
figure 2

Fluorescence density plots of Candida parapsilosis ATCC 22019 response to staining with FDA and PI in the carrot-orange juice blend treated (0–15 min) with single mild heat (H, 50 °C), single UV-C (0–1720 kJ/cm2) and UV-C assisted by mild heat (UV-C/H). The percentages of microbial populations that fall in each gate are displayed in the four edges of each plot

Full size image

Not all untreated C. parapsilosis cells (Fig. 2a, 0 min) yielded green fluorescence, possibly caused by extrusion of F outside the cell, mediated by an ATP-driven transport system. This possible efflux may also be also interpreted as an indicator of viability (Ferrario et al. 2017).

An increase in treatment exposure time of yeast cells resulted in a gradual shift of cells from gate #1 (cells with esterase activity and intact membrane) to gate #4 (cells with compromised membrane) (Fig. 2). Therefore, all the treatments disrupted the integrity of the cell membrane, allowing PI to penetrate the cell. The fastest shift to the death gate (#4) was more pronounced in the case of applying the combined UV-C/H treatment, reaching up to 98% of PI-stained cells. In contrast, only 12.0% and 46.2% were achieved after 15 min of exposure to the single H and UV-C treatments, respectively. The higher fraction of PI-stained cells observed for the combined treatment compared with single UV-C is in accordance with the higher inactivation observed by the plate count method illustrated in Fig. 1 and described above.

A small proportion of PI-stained cells (12.0%) was observed after 15 min of the single H treatment, compared with single UV-C (46.2%). Nevertheless, greater inactivation was achieved after single H (3.9 log reductions) compared with single UV-C (2.9 log reductions). This result suggests that other inactivation mechanisms, besides membrane damage, contributed to yeast inactivation by mild heat. It is important to highlight that no double-stained cells were recorded at any condition assayed, thus indicating that the proposed treatment ensured the absence of sublethally injured cells. The presence of viable but non-culturable cells (VBNC) could seriously affect the shelf-life of juices (Ferro et al. 2018) as these cells are metabolically active but cannot be detected by conventional culture methods (García Carrillo et al. 2018).

Consistently with the current work, in a previous study, García Carrillo et al. (2018) evaluated the induced damage of S. cerevisiae KE 162 cells in peptone water and in a carrot-orange juice blend after exposure to single UV-C (1060 kJ/cm2), single mild heat (50 °C) and UV-C assisted by mild heat by flow cytometry. They also observed a more pronounced shift to the PI+ gate after UV-C assisted by mild heat (99.0%) compared with the single treatments (single mild heat:17.5%, single UV-C:41.6%). When compared with the present study, a similar proportion of PI-stained cells was recorded after 15 min exposure to single UV-C (46.2%), single mild heat (12.0%), or combined UV-C/H (98.2%) for C. parapsilosis. Nevertheless, in that study was reported the presence of S. cerevisiae double-stained cells after 15 min and 1 min of exposure to single UV-C (8.5%) in the juice blend and peptone water (5%), respectively.

Physicochemical Characterization of Carrot-Orange Juice Blend Along Storage

Polyphenol Content and Total Antioxidant Activity

Figure 3 illustrates the total TPC and TAA values corresponding to untreated carrot-orange juice or treated by single UV-C, H or combined UV-C/H during refrigerated storage (4 ± 1 °C).

Fig. 3
figure 3

Total polyphenol content (TPC, μg GAE/mL) (a), and total antioxidant activity (TAA, mg Trolox Eq/mL) (b) of untreated carrot-orange juice (control, □) or processed by the single (■), single H (■), or combined UV-C/H (■) treatments during refrigerated storage (4 °C). Standard deviation (I). Different letters represent significant differences according to MANOVA test

Full size image

Control, UV-C and UV-C/H samples displayed similar TAA values of 0.9 ± 0.5, 0.9 ± 0.3 and 1.6 ± 1.3 mg Trolox Eq/mL, respectively, at the beginning of storage. Likewise, similar TPC values were observed for the control (205.0 ± 55.9 μg GAE/mL), UV-C (230 ± 31.0 μg GAE/mL), and UV-C/H (260 ± 90.1 μg GAE/mL) treated juices, immediately after exposure. Moreover, no differences were determined among the control, single UV-C, and combined UV-C/H treatments throughout storage (p > 0.05). A significant increase in TAA was observed after the single H (TAA = 2.4 ± 0.8 mg Trolox Eq/mL) treatment up to 5 days of storage (Fig. 3); but no differences were recorded between single H and control from the seventh day to the end of storage (Fig. 3). Therefore, the proposed treatment UV-C/H did not modify the TPC and TAA of natural fresh carrot-orange juice blend during storage compared with the untreated juice. The observed increase in TAA values after the H treatment may be attributed to a higher extractability of these compounds during mild processing, making them more accessible to react. On the same trend, Zaccari et al. (2015) assessed the carotenoid content of carrot cylinders (10-mm thickness) cooked in a steamer (20 min at 45 or 50 °C). These authors observed an increase in this biocompound in the cooked samples compared with the raw ones.

Literature reports variable results on the residual antioxidant activity and polyphenol content determined after UV-C treatments, depending on the juice type, UV-C treatment conditions, and the method employed to determine these bioactive compounds. Fundo et al. (2019) studied the antioxidant activity by ABTS and the total polyphenol content of melon juice (pH 6.3, 11.4 °Brix) treated by UV-C (13.44 W/m2, 4 lamps). Consistently with the present work, the UV-C treated samples showed similar total phenolic content and antioxidant activity to the untreated juice samples. Likewise, Caminiti et al. (2012) observed that the total phenol content and antioxidant activity values of a carrot-orange juice blend (1:1, pH 3.8, 9 °Brix) treated by UV-C (10,600 mJ/cm2) remained unchanged. In addition, they reported similar total phenolic content (498 and 505 μg GAE/mL) than those corresponding to the present study (205 and 306 μg GAE/mL) for untreated and UV-C treated juice samples, respectively. La Cava and Sgroppo (2019) examined the total polyphenol content and antioxidant activity of a grapefruit juice (pH 3.12, 9.6 °Brix, 2500 NTU) treated by UV-C (39.6 J/L) assisted by mild heat (65 °C) in a coiled tube reactor at laboratory scale, during 28 days of refrigerated storage. In contrast to the results shown in the present work, these authors observed a significant reduction of polyphenols (14%) and antioxidant activity determined by DPPH (19%) and ABTS (16%) when compared with the untreated juice. They also reported that the antioxidant activity gradually decreased with storage. This decrease may be attributed to the higher temperature applied compared with the temperature used in this study. In this regard, Ferrario et al. (2017), assessed the changes in the total antioxidant activity by DPPH in a carrot juice (pH 6.4, 11.3°Brix) processed by a hurdle technology involving different combinations of traditional factors such as mild heating (56, 58 and 60 °C), pH (4.5, 5.0, and 5.5) and time (2, 4, and 6 min). They reported a decrease in the antioxidant activity with temperature increase, probably due to the fact that higher temperature values were used compared with the present study.

Color, pH, °Brix, and Turbidity

No significant differences (p > 0.05) were recorded for L*, a*, and b* values immediately after juice processing among control, single H, and UV-C/H samples (Table 1). Moreover, for all the evaluated systems, the color values remained unaltered throughout storage (p > 0.05) (Table 1). These results indicated that all the treatments applied did not significantly alter the color of the freshly squeezed carrot-orange juice.

Table 1 Evolution of CIELab parameters (L*, a*, b*), pH, °Brix, and turbidity of the carrot-orange juice blend without treatment (control) or treated with single UV-C light (UV-C, 1720 mJ/cm2), mild heat (H, 50 °C), or UV-C light assisted by mild heat (UV-C/H) during refrigerated storage (4 °C)
Full size table

No changes in pH were observed between the control (4.21 ± 0.01) and all the treated samples (4.20–4.24), immediately after processing and throughout the refrigerated storage (Table 1). In addition, no differences were recorded in the °Brix and turbidity values among single UV-C, single H, and the combined UV-C/H (7.5–8.2 °Brix, 2540–3394) treatments from the onset and throughout the whole storage (Table 1).

As expected, the H treatment (50 °C) did not produce an impact on juice color, °Brix, pH and turbidity. Moreover, it is well known that UV-C does not generally modify these parameters when applied to juices and drinks for food preservation purposes. In particular, Fundo et al. (2019) evaluated changes in pH, °Brix, and color of melon juice (pH 6.3, 11.4 °Brix) treated by single UV-C (13.44 W/m2) and they did not observe changes in pH, °Brix, L*, a*, and b* parameters after UV-C processing compared with untreated juice samples. Similarly, Riganakos et al. (2017) reported no change in color parameters of a carrot juice (pH 6.3, > 11,000 NTU) after treated by UV-C (1.2 kJ/L).

Pectin Methylesterase Activity

The effect of the single or combined treatments on the pectin methylesterase activity (PME) was assessed immediately after processing (Fig. 4a), and also during refrigerated storage (Fig. 4b). PME was not reduced immediately after UV-C exposure, compared with the control samples (Fig. 4a). In contrast, the single H and UV-C/H treatments significantly reduced the enzyme activity (42.5 ± 1.5% and 44.6 ± 6.3%) immediately after processing and remained constant throughout storage (Fig. 4b), thus indicating that enzyme activity was mainly sensitive to the mild heat exposure. As an additional control, the juice blend was also thermally treated (T-coil) achieving a PME value of 6.8 ± 2.2% (Fig. 4a).

Fig. 4
figure 4

a Pectin methylesterase units (PME) corresponding to the carrot-orange juice blend untreated (control) or treated by single mild heat (H, 50 °C), UV-C (1720 mJ/cm2), UV-C assisted by mild heat (UV-C/H), and heat (T-coil, 80 °C) immediately after processing. b Pectin methylesterase units (PME) corresponding to the carrot-orange juice blend processed by UV-C assisted by mild heat (UV-C/H), during refrigerated storage (4 ± 1 °C): 0 (□), 6 (■), 13 (■) and only for the UV-C/H treatment, 24 days (■). Different letters represent significant differences according to MANOVA tests; *represents non-significant differences among storage days for the UV-C/H treatment

Full size image

In agreement with the current work when single UV-C was applied, Sew et al. (2014) reported that pectin methylesterase enzyme could not be inactivated in pineapple juice exposed to continuous UV-C (11.23 mJ/cm2, 2.47 s, T < 35 °C). Moreover, Biancaniello et al. (2018) reported that an UV-C (0.88–2.93 kJ/L, turbulent flow, coiled UV-C unit, T < 17 °C) treatment was unable to reduce the pectin methylesterase enzyme in a green juice blend. In contrast, Zhang et al. (2011), studied the inactivation of this enzyme in watermelon juice (24 mL) a coiled-tubing UV-C device (9.7 J/mL, 23 °C) and observed a reduction of 35% on its activity.

Sensory Analysis

Conjoint Study

When all the consumers used for this study were considered, willingness to try averaged 5.9 ± 1.3 points, which was close to the “I would definitely try it” category. Whereas, the average healthiness perception was of 6.4 ± 0.9 points, corresponding to the “very healthy” category. No differences in the willingness to try and the perceived healthiness were determined between consumers considering the information provided by the four claims (p > 0.05). In particular, consumers did not associate the use of high temperature to loss of product quality, probably because most of the panelists were very young (18–29 years old). Several studies reported that older people and consumers having knowledge on nutrition are more positive toward functional food and more receptive to health claims (Klopčič et al. 2020). Moreover, no differences in the hedonic scores were detected when the panelists were provided with information about juice health benefits (claims 3 and 4) of the product compared with those who did not receive any information (claim 1) (p > 0.05). However, women exhibited significantly higher willingness to try (6.6 ± 0.9) than men (5.3 ± 0.8) (p < 0.05), being those differences more pronounced when they received the claim 3 (100% natural, with no preservatives or sugar added, rich in antioxidants). No differences between genders were observed for the healthiness perception (p > 0.05) of the juice.

In agreement with the present work, Endrizzi et al. (2015) performed a conjoint study to evaluate the overall liking in a 9-point hedonic scale of apple discs while providing information about fiber and antioxidant content. They observed that nutrition information had no significant effect on the overall average preference. These authors proposed that healthiness perception might depend on the type of product, as it is difficult for consumers to perceive a well-known healthy food product as even more healthy. On the contrary, Zhu et al. (2018) evaluated the purchase intent and willingness to pay of a high-quality flavored tomato juice compared with one widely used by the consumers. They reported that there were no differences between the two juices regarding the purchase intent and willingness to pay before tasting. Notwithstanding, consumers’ willingness to pay dramatically fell after tasting the commercial juice compared with the high-quality tomato juice. Similarly, Gadioli et al. (2013) performed a conjoint study of commercial orange juices to assess the effect of the information provided in the labels. They reported that the nutritional information regarding vitamin C content was mostly considered a relevant factor, suggesting the concern of the participants about their preference to consume products that provide health benefits. Additionally, a claim alleging “homemade—with juicy bits” resulted in a positive impact as it reminded consumers of a homemade, not industrialized product.

Consumer Field Test

A consumer field test was performed to determine the overall impression and consumer’s perception of some relevant sensory attributes of the UV-C/H and T-coil juice samples. The overall acceptability of the UV-C/H and T-coil samples averaged 6.0 and 6.1 in the 9-point hedonic scale, respectively, which corresponded to the category “like slightly” in the used scale. Notwithstanding, two groups of consumers’ preferences emerged when a cluster analysis was applied: cluster 1 (C1), with 78 consumers, including the categories 6 to 8 in the 9-point hedonic scale, and cluster 2 (C2), with 52 consumers, encompassing the categories 2 to 5 (data not shown). Good adequacy of fit was achieved by the cluster analysis as the CCC value obtained was 0.81. The cluster C1 showed a marked interest in the product, exhibiting an average overall liking of 7.0 in the 9-point hedonic scale (corresponding to the category “like it moderately”). Figure 5a shows the scores assigned to the clusters 1 and 2 by the panelists for the carrot-orange juice blend samples treated by UV-C/H and T-coil. These samples were associated with having adequate carrot taste, bitterness, juice body, and intense fruity aroma. The lower scores assigned to C2 were attributed to significantly higher carrot taste and bitterness, compared with C1 (p < 0.01) (Fig. 5a). Hence, the UV-C/H juice would be targeted to a specific market segment in line with the segmentation approach, which has been embraced by manufacturers to a large extent. This approach is based on addressing specific products to different groups of consumers rather than developing a unique product aimed at all the consumers (Lawless 2013).

Fig. 5
figure 5

a Average scores assigned by the panelists corresponding to the clusters 1 (▬) and 2 (▬), in the consumer field test for the carrot-orange juice treated by UV-C assisted by mild heat (UV-C/H, 1720 mJ/cm2/ 50 °C). Average scores assigned by the panelists to the attributes corresponding to the thermally treated juice (T-coil, 80 °C - - -). b Correspondence analysis bi-plot representing term citation frequency according to the CATA question for the sensory evaluation of the UV-C/H and pasteurized juices, and contrasted to an ideal juice defined by consumers from a checklist included in the CATA questionnaire. Different letters indicate significant differences according to MANOVA test

Full size image

Check-All-That-Apply Question

The consulted consumers employed between 1 and 10 terms to describe the juice samples in the check-all-that-apply (CATA) question. The most frequently used terms for UV-C/H and T-coil samples were carrot taste, natural taste, and intense color. On the other hand, the least used terms were cooked taste and vegetable aroma.

Figure 5b shows a correspondence analysis (CA) performed for the studied sensory descriptors. The first and second dimensions of the CA calculated from CATA counts represented 95.8% and 3.2% of the experimental data variance, respectively. The UV-C/H juice exhibited positive and negative values according to the axes 1 and 2, respectively, and was described by the consumers as having the following attributes: carrot taste, not fruity, strange taste, strange aroma. Whereas, the T-coil sample showed positive values for both axes, being described as follows: cooked taste, artificial taste, too particulate, taste persistence, intense color, intense fruit-vegetable taste (Fig. 5b). Therefore, processing the juice blend by traditional pasteurization (80 °C) may have caused particle aggregation and undesirable cooked taste. The use of mild heat (50 °C) to assist the UV-C treatment prevented the juice from undergoing these changes. On the other hand, the ideal juice was defined as having the following attributes: natural taste, orange taste, pleasant juice body, pleasant color. Consequently, the comparison between the real juice samples and an ideal beverage defined by consumers should target specific further improvements that might be addressed to the studied samples. In particular, consumers would ideally have expected a stronger orange taste and weak carrot taste than those they actually perceived in the UV-C/H sample.

Conclusions

The present work demonstrated that the proposed UV-C (1720 mJ/cm2) assisted by mild heat (50 °C) treatment significantly reduced Candida parapsilosis ATCC 22019 in a turbid carrot-orange juice blend, while ensuring the absence of sublethally damaged cells. It also contributed to better understand the mechanism by which C. parapsilosis was inactivated in the carrot-orange juice blend treated by UV-C assisted by mild heat. The novel processing proposed resulted in a feasible alternative to thermal pasteurization by preserving the nutritional and sensorial juice quality, as it did not significantly modify the juice color, pH, turbidity, polyphenol content, and antioxidant activity during the whole refrigerated storage. Moreover, it induced an important decrease of the pectin methylesterase activity, which was mainly attributed to the effect of the mild heat assisting the UV-C treatment. Additionally, a group of consumers showed a strong interest in the carrot taste of the product and perceived it as very healthy. Moreover, the UV-C assisted by mild heat juice was described as having better sensory quality than a pasteurized juice, thus making the proposed juice processing a feasible alternative in the near future to the traditional thermal treatments. In addition, the CATA question indicated that changes in certain attributes should be addressed to the UV-C/H juice in order to better meet consumers’ expectations.

Further studies involving the incorporation of other hurdles and/or modifications in the UV-C device should be conducted in order to achieve an improved reduction of the pectin methylesterase activity.