Monitoring Commercial Starter Culture Development in Presence of Red Grape Pomace Powder to Produce Polyphenol-Enriched Fresh Ovine Cheeses at Industrial Scale Level

: Red grape Nero d’Avola cultivar grape pomace powder (GPP) was applied during fresh ovine cheese production in order to increase polyphenol content. Before cheeses were produced, the bacteria of a freeze-dried commercial starter culture were isolated and tested in vitro against GPP. Two dominant strains, both resistant to GPP, were identiﬁed. Thestarter culture was inoculated in pasteurized ewe’s milk and the curd was divided into two bulks, one added with 1% ( w / w ) GPP and another one GPP-free. GPP did not inﬂuence the starter culture development, since lactic acid bacteria (LAB) counts were 10 9 CFU/g in both cheeses at 30 d. To exclude the interference of indigenous LAB, the pasteurized milk was analyzed, and several colonies of presumptive LAB were isolated, puriﬁed and typed. Four strains were allotted into Enterococcus and Lacticaseibacillus genera. The direct comparison of the polymorphic proﬁles of cheese bacteria evidenced the dominance of the starter culture over milk LAB. The addition of GPP increased cheese total phenolic compounds by 0.42 g GAE/kg. Sensory evaluation indicated that GPP-enriched cheese was well appreciated by the judges, providing evidence that GPP is a suitable substrate to increase the availability of total phenolic content in fresh ovine cheese.


Introduction
Polyphenols are secondary plant-derived bioactive metabolites [1]. These compounds are able to prevent lipid oxidation, deterioration of polyunsaturated fatty acids and inhibit the growth of undesired pathogenic and spoilage microorganisms in food products [2,3]. Polyphenols also exert beneficial effects on the human health, including the prevention of different chronic diseases [4]. In the last decades, in light of the directives from the European Commission (1999/31/EC and 2008/98/EC) related to waste and environmental sustainability, fruit and vegetable processing industries have focused their efforts to diminish the environmental impact of wastes and by-products. One promising strategy to valorize these organic matrices is the extraction of active compounds [5,6]. In this context, particular attention has been paid to vinery by-products [7], in particular grape pomace, which represents approximately 20-25% of the grapes at harvest [8]. Grape pomace contain grape skin and seeds [9] and represents a waste biomass difficult to dispose of [10], since it cannot be directly applied as a soil conditioner or fertilizer due to high levels of phenolic compounds inhibiting the germination of plant seeds [11] and the growth of soil microor-

Isolation and Typing of Commercial Starter Culture
Freeze-dried starter preparation (5 g) was first homogenized with Ringer's solution (45 mL) by a stomacher Bag-Mixer 400 (Interscience, Saint Nom, France) for 2 min at the maximum speed and then serially diluted (1:10). Cell suspensions of freeze-dried starter culture were subjected to a plate count for the enumeration of mesophilic coccusshaped LAB on M17 agar (Biotec, Grosseto, Italy), incubated anaerobically for 48 h at 30 • C. After growth, the colonies of presumptive lactic acid bacteria (LAB) were randomly picked up, purified by successive subculturing and tested for the Gram reaction [26], and for catalase activity performed by addition of H 2 O 2 (3% (v/v)) to the colonies [27]. The purity and cell morphology of the isolates were microscopically investigated. All cultures were subjected to DNA extraction using a DNA-SORB-B kit (Sacace Biotechnologies Srl, Como, Italy) following the protocol provided by the manufacturer. The differentiation of the isolates at the strain level was performed by randomly amplified polymorphic DNA (RAPD)-PCR analysis using single primers M13 (5 -GAGGGTGGCGGTTCT-3 ), AB106 (5 -

Inhibitory Assay
In order to test the suitability of GPP to produce fresh ovine cheese, its potential inhibitory activity was tested against the strains isolated from the commercial starter culture. GPP was rehydrated in distilled water to a final concentration of 200 µg/mL as commonly performed for extracts of vegetable origin [29,30]. The water-soluble extract of GPP was tested against a cell density of 10 7 CFU/mL of each LAB strain in M17 soft agar (0.7% w/v) applying the paper disc diffusion assay as reported by Cruciata et al. [31]. Streptomycin (10% w/v) was used as positive control, with sterile water as negative control. The inhibitory activity was considered positive if a definite clear area was detected around the paper discs. This test was performed in duplicate.

Cheese Productions and Sample Collection
Cheese production was performed under industrial conditions at a dairy factory scale (Biopek, Gibellina, Italy) applying "Primosale" type pressed cheese technology. The experimental plan included two different cheese productions: a control production (CP) inoculated with the same commercial starter culture characterized for GPP resistance, and an experimental production (EP) inoculated with the commercial starter culture and 1% (w/w of curd) of GPP. Each trial was performed in a stainless-steel vat previously sanitized with a PROMOX P900 solution (Leggiuno, Italy), with 500 L of pasteurized ewe's milk (72 • C for 15 s). Freeze-dried commercial starter culture (5 U) was reactivated in 5 L of pasteurized ewe's milk and kept for 10 min at 44 • C. After cooling at 38 • C, the bulk milk was inoculated with the starter culture preparation and kept under slow agitation for 30 min prior to adding liquid rennet (150 mL, Fromase ® 220 TL, DSM Bright Science Brighter Living, Heerlen, The Netherlands). After 40 min, the coagulum was mechanically cut to the dimension of rice grains. After draining whey, the control curd was immediately put into perforated plastic molds, while GPP was added to the experimental curd and then transferred into perforated plastic molds. Both curds were stewed at 40 • C for 2 h and turned every 30 min. At the end of the process, all cheeses were salted in saturated brine for 6 h, dried for 24 h at room temperature and ripened for 30 d at 13 • C and 80% relative humidity. The trials were carried out in triplicate over three consecutive weeks. Samples of GPP, raw milk, pasteurized milk, inoculated milk after addition of reactivated starter culture, curds, and cheeses after 30 d of ripening, were collected for analyses.

Microbiological Analyses
All milk samples (1 mL) were serially diluted directly in Ringer's solution (Sigma-Aldrich, Milan, Italy), while GPP, curd and cheese samples (10 g) were first homogenized in 90 mL of Ringer's solution in a stomacher and then serially diluted as reported above.
Cell suspensions of raw milk, pasteurized milk and GPP were subjected to plate counts for the enumeration of the following microbial groups: total mesophilic microorganisms (TMM) on plate count agar (PCA), incubated aerobically at 30 • C for 72 h; mesophilic rod shaped LAB on de Man-Rogosa-Sharpe (MRS) agar, acidified to pH 5.4 with lactic acid (5 mol/L) and incubated anaerobically for 48 h at 30 • C; mesophilic coccus-shaped LAB on Media 17 (M17) agar, incubated anaerobically for 48 h at 30 • C; members of the Enterobacteriaceae family on violet red bile glucose agar (VRBGA), incubated aerobically at 37 • C for 24 h; coagulase-positive staphylococci (CPS) on Baird-Parker (BP) agar added with rabbit plasma fibrinogen, incubated aerobically at 37 • C for 48 h; Listeria monocytogenes on Listeria selective agar base (LSAB) added with SR0140E supplement, incubated aerobically at 37 • C for 48 h; Escherichia coli and Salmonella spp., on Hektoen enteric agar (HEA), incubated aerobically at 37 • C for 24 h.
Cell suspensions of inoculated milk after addition of reactivated starter culture, curds and cheeses were analyzed for TMM and LAB exclusively on M17 agar as reported above. All media and supplements were purchased from Biotec, expect HEA provided by Microbiol Diagnostici (Uta, Italy). Plates counts were performed in duplicate.

Isolation, Genotypic Differentiation and Identification of Thermoduric Indigenous Milk LAB
The colonies of presumptive LAB developed from the highest dilutions of pasteurized milk sample suspensions were picked up, purified and tested for main LAB characteristics as reported above. All Gram positive and catalase-negative cultures were considered presumptive LAB and differentiated by RAPD-PCR analysis as reported above. Strain typing was performed using the software Gelcompare II version 6.5 (Applied-Maths, Sin Marten Latem, Belgium). Genotypic identification was performed by sequencing the 16S rRNA gene following the procedures applied by Gaglio et al. [32] using the primers rD1 (5 -AAGGAGGTGATCCAGCC-3 ) and fD1 (5 -AGAGTTTGATCCTGGCTCAG-3 ). The identities of the sequences were determined by comparison with those available in the GenBank/EMBL/DDBJ (http://www.ncbi.nlm.nih.gov) [33] and EzTaxon-e (http: //eztaxon-e.ezbiocloud.net/) [34] databases. The unequivocal identification of the Enterococcus species that showed a discrepancy between the two databases was confirmed using the multiplex PCR assay based on the sodA gene. PCR reactions were performed as described by Jackson et al. [35] using the primers FM1 (5 -GAAAAAACAATAGAAGAATTAT-3 ) and FM2 (5 -TGCTTTTTTGAATTCTTCTTTA-3 ) specific for Enterococcus faecium.

Commercial Starter Culture Recognition
The persistence of the commercial starter culture, and its dominance over the indigenous milk LAB resistant to the pasteurization process, was assessed by microscopic inspection and RAPD profile comparison between the LAB collected during cheese making and those purified from the commercial starter culture.

Total Phenolic Content
Total phenolic content (TPC) of GPP and cheese samples was evaluated through an analyzer iCubio iMagic M9 (Shenzhen iCubio Biomedical Technology Co. Ltd. Shenzhen, China), which runs with full automation. This system automatically pipetted reagents and samples into the cuvette, allowed incubation at a controlled temperature, read absorbance at the specific wavelength and calculated the concentration of the analyses with a calibration curve. The parameters used in the automated photometric systems were temperature, 37 • C; wavelengths 340 nm and 415 nm (bichromatic) and optical path of 1 cm. The reagent used was Enzytec™ Polyphenols Cod. E2530. All reagents and standards were purchased from R-Biopharm AG (Darmstadt, Germany). All samples were homogenized before the extraction procedure. Conventional solid-liquid extraction of polyphenols and sample preparation were performed according to manufacturer's instructions: 20 g of each sample were treated with 80 mL of ethanol solution (70% v/v) and stirred for 60 min at 25 • C. The samples were then filtered and put into a 100 mL volumetric flask, brought to volume with distilled water. The final concentration of polyphenols was then reported as g of gallic acid equivalent (GAE)/kg of dry mass (DM) for GPP or g GAE/kg of product for cheeses. All measurements were performed in duplicate.

Sensory Evaluation
The control and experimental cheese productions, at 30 d of ripening, were evaluated for their sensory traits by a descriptive panel of 12 judges (six women and six men aged between 21 and 68 years old) selected for their cheese consumption habits.
The judges were trained for cheese attribute evaluation following the ISO 8589 [36] indications. For each cheese, the judges evaluated twelve descriptive attributes among those reported by Niro et al. [37] and evaluated by Costa et al. [38]. In particular, the evaluation considered the following aspects of the cheeses: intensity of odor and aroma, sweet, salt, acid, astringent, friability, fiber, adhesiveness, hardness, humidity and overall assessment. The quality of

Statistical Analyses
Microbiological data were subjected to One-Way Variance Analysis (ANOVA) using XLStat software version 7.5.2 for Excel (Addinsoft, New York, NY, USA). The Duncan test was applied to evaluate the level of significance between control and experimental samples. p < 0.05 was considered significant.

Isolation and Typing of Commercial Starter Culture
Dried starter culture was dominated by LAB with levels of mesophilic LAB cocci of 10.5 Log CFU/g. LAB (Gram positive and catalase negative) were isolated from M17 agar media incubated at 30 • C at the highest dilutions of cell suspensions. After microscopic inspection, all isolates showed coccus-shaped cells organized in short chains, typical of lactococci [40]. As declared by the producer company, RAPD analysis confirmed the presence of two strains ( Figure 1) included in the L. lactis freeze-dried starter preparation.
for their sensory traits by a descriptive panel of 12 judges (six women and six men aged between 21 and 68 years old) selected for their cheese consumption habits.
The judges were trained for cheese attribute evaluation following the ISO 8589 [36] indications. For each cheese, the judges evaluated twelve descriptive attributes among those reported by Niro et al. [37] and evaluated by Costa et al. [38]. In particular, the evaluation considered the following aspects of the cheeses: intensity of odor and aroma, sweet, salt, acid, astringent, friability, fiber, adhesiveness, hardness, humidity and overall assessment. The quality of each attribute was scored using a 70-mm visual analogue scale anchored on the left as reported by Faccia et al. [39].

Statistical Analyses
Microbiological data were subjected to One-Way Variance Analysis (ANOVA) using XLStat software version 7.5.2 for Excel (Addinsoft, New York, NY, USA). The Duncan test was applied to evaluate the level of significance between control and experimental samples. p < 0.05 was considered significant.

Isolation and Typing of Commercial Starter Culture
Dried starter culture was dominated by LAB with levels of mesophilic LAB cocci of 10.5 Log CFU/g. LAB (Gram positive and catalase negative) were isolated from M17 agar media incubated at 30 °C at the highest dilutions of cell suspensions. After microscopic inspection, all isolates showed coccus-shaped cells organized in short chains, typical of lactococci [40]. As declared by the producer company, RAPD analysis confirmed the presence of two strains ( Figure 1) included in the L. lactis freeze-dried starter preparation.

Evaluation of Suitability of GPP for Cheese Making
The suitability of GPP for the production of cheeses depends on its harmlessness towards LAB, which are the main fermentation agents necessary to transform milk into cheese [41]. The results of the paper disc diffusion assay showed that GPP did not inhibit

Evaluation of Suitability of GPP for Cheese Making
The suitability of GPP for the production of cheeses depends on its harmlessness towards LAB, which are the main fermentation agents necessary to transform milk into cheese [41]. The results of the paper disc diffusion assay showed that GPP did not inhibit the growth of the L. lactis strains isolated from the dried starter culture (results not shown), confirming its suitability in fresh ovine cheese production.

Microbiological Evolution during Cheese Productions
Microbiological investigation was carried out throughout cheese production from GPP to final 30-d ripened cheeses. No colonies of bacteria and yeasts were detected in GPP. The absence of these microorganisms is a consequence of the thermal treatment applied during the oven-drying process [42]. Statistically significant differences (p < 0.001) were observed for the levels of all microbial groups investigated between the raw milk and pasteurized  (Figure 2). The results of L. monocygenes and Salmonella spp. are not reported in the figure because their levels were below the detection limit for both matrices.
confirming its suitability in fresh ovine cheese production.

Microbiological Evolution during Cheese Productions
Microbiological investigation was carried out throughout cheese production from GPP to final 30-d ripened cheeses. No colonies of bacteria and yeasts were detected in GPP. The absence of these microorganisms is a consequence of the thermal treatment applied during the oven-drying process [42]. Statistically significant differences (p < 0.001) were observed for the levels of all microbial groups investigated between the raw milk and pasteurized milk (Figure 2). The results of L. monocygenes and Salmonella spp. are not reported in the figure because their levels were below the detection limit for both matrices. The levels of TMM, mesophilic coccus and rod LAB of raw milk were 6.7, 6.5 and 5.9 Log CFU/mL, respectively, while they were recorded at about 3 Log cycles lower in pasteurized milk. These results are not surprising since indigenous milk LAB are able to survive the heat treatment [43,44]. Regarding the undesired bacterial groups, especially members of Enterobacteriaceae family, CPS and E. coli, they were in the range 10 2 -10 3 CFU/mL in raw milk and completely disappeared after pasteurization.
The growth of the fermenting agent, added as starter culture, during cheese productions is reported in Table 1.  The levels of TMM, mesophilic coccus and rod LAB of raw milk were 6.7, 6.5 and 5.9 Log CFU/mL, respectively, while they were recorded at about 3 Log cycles lower in pasteurized milk. These results are not surprising since indigenous milk LAB are able to survive the heat treatment [43,44]. Regarding the undesired bacterial groups, especially members of Enterobacteriaceae family, CPS and E. coli, they were in the range 10 2 -10 3 CFU/mL in raw milk and completely disappeared after pasteurization.
The growth of the fermenting agent, added as starter culture, during cheese productions is reported in Table 1. The levels of TMM and L. lactis were superimposable for the samples analyzed. The milk inoculated with the reactivated starter culture showed levels of these microorganisms at about 10 7 CFU/mL. After curdling, TMM and L. lactis were counted both in control and  [45]. In the final control and experimental cheeses, the levels of TMM and LAB were above 9 Log CFU/g. These data highlighted the ability of the commercial starter culture to drive the fermentation process in presence of 1% (w/w) of GPP.

Identification of Thermoduric Indigenous Milk LAB
Twenty-one presumptive LAB (Gram positive and catalase negative) isolated from pasteurized milk were purified and analysed by RAPD-PCR in order to perform their strain typing. The last analysis, commonly used to discriminate LAB strains isolated from food matrices [46], showed the presence of four different strains (Figure 3). within a column followed by the same letter are not significantly different according to Duncan's test. Abbreviations: TMM, total mesophilic microorganisms; L., Lactococcus; CP, control production inoculated with a commercial starter culture; EP, experimental production inoculated with a commercial starter culture +1% of GPP.
The levels of TMM and L. lactis were superimposable for the samples analyzed. The milk inoculated with the reactivated starter culture showed levels of these microorganisms at about 10 7 CFU/mL. After curdling, TMM and L. lactis were counted both in control and experimental curd at 10 8 CFU/g showing an increase of about 1 Log cycle determined as a direct concentration of bacteria due to whey draining [45]. In the final control and experimental cheeses, the levels of TMM and LAB were above 9 Log CFU/g. These data highlighted the ability of the commercial starter culture to drive the fermentation process in presence of 1% (w/w) of GPP.

Identification of Thermoduric Indigenous Milk LAB
Twenty-one presumptive LAB (Gram positive and catalase negative) isolated from pasteurized milk were purified and analysed by RAPD-PCR in order to perform their strain typing. The last analysis, commonly used to discriminate LAB strains isolated from food matrices [46], showed the presence of four different strains (Figure 3). These strains were identified at the species level by sequencing of the 16S rRNA gene, indicating that the thermoduric indigenous milk LAB included species within the genera Enterococcus and Lactobacillus. The last genus included two strains belonging to the species Lacticaseibacillus casei (PBM19) and Lacticaseibacillus paracasei (PBM7) (named Lactobacillus casei and Lactobacillus paracasei, respectively, before reclassification of Zheng et al. [47]). Regarding the Enterococcus genus, one strain was identified as Enterococcus faecalis (PBM1), while the strain PBM16 could not be identified at the species level due to the different results of the BLAST and EzTaxon searches. This strain was further analyzed by a speciesspecific multiplex PCR assay and classified as Enterococcus faecium. All the species identified are components of the nonstarter LAB (NSLAB) community implicated in cheese maturation [39], commonly associated with raw milk [48] and cheeses [44,49]. However, their presence in pasteurized milk is mainly due to postpasteurization contamination [50] or to their ability to survive the pasteurization process [51,52] These strains were identified at the species level by sequencing of the 16S rRNA gene, indicating that the thermoduric indigenous milk LAB included species within the genera Enterococcus and Lactobacillus. The last genus included two strains belonging to the species Lacticaseibacillus casei (PBM19) and Lacticaseibacillus paracasei (PBM7) (named Lactobacillus casei and Lactobacillus paracasei, respectively, before reclassification of Zheng et al. [47]). Regarding the Enterococcus genus, one strain was identified as Enterococcus faecalis (PBM1), while the strain PBM16 could not be identified at the species level due to the different results of the BLAST and EzTaxon searches. This strain was further analyzed by a species-specific multiplex PCR assay and classified as Enterococcus faecium. All the species identified are components of the nonstarter LAB (NSLAB) community implicated in cheese maturation [39], commonly associated with raw milk [48] and cheeses [44,49]. However, their presence in pasteurized milk is mainly due to postpasteurization contamination [50] or to their ability to survive the pasteurization process [51,52].

Commercial Starter Culture Recognition
In order to evaluate the ability of the commercial starter culture to dominate over the thermoduric indigenous milk LAB and to drive the fermentation process in presence of GPP, a total 101 isolates, collected from inoculated milk to final cheeses, were identified using a polyphasic approach based on microscopic inspection and RAPD-PCR analysis. The results of microscopic inspection confirmed that all isolates were cocci with cells occurring in long chains. The direct comparison of the polymorphic profiles of LAB isolated from pasteurized milk with those of L. lactis strains originating from commercial starter LAB, allowed the assessment of the evolution of the added LAB strains (results not shown). This approach confirmed the dominance of the added L. lactis strains, originating from commercial LAB starter culture, over thermoduric indigenous milk LAB. Moreover, these data confirmed those obtained by plate counts that showed the ability of commercial starter LAB to drive the fermentation process in presence of GPP.

Total Phenolic Content
The results of TPC are reported in Figure 4. GPP was characterized by a concentration of TPC of 110.7 g GAE/kg DM, which is higher than the value reported by Harsha et al. [53] for red wine grape pomace of a different Italian cultivar.
curring in long chains. The direct comparison of the polymorphic profiles of LAB isolated from pasteurized milk with those of L. lactis strains originating from commercial starter LAB, allowed the assessment of the evolution of the added LAB strains (results not shown). This approach confirmed the dominance of the added L. lactis strains, originating from commercial LAB starter culture, over thermoduric indigenous milk LAB. Moreover, these data confirmed those obtained by plate counts that showed the ability of commercial starter LAB to drive the fermentation process in presence of GPP.

Total Phenolic Content
The results of TPC are reported in Figure 4. GPP was characterized by a concentration of TPC of 110.7 g GAE/kg DM, which is higher than the value reported by Harsha et al. [53] for red wine grape pomace of a different Italian cultivar. To evaluate the effects of GPP addition to cheese, TPC was determined in control and experimental cheese samples. Statistically significant differences (p < 0.0001) were found among cheeses. As expected, the amount of TPC detected in control cheese was lower than the corresponding GPP enriched cheeses. In particular, the addition of 1% (w/w) of GPP resulted in an increase of 0.42 g GAE/kg of TPC in experimental cheeses. These results confirmed that the addition of GPP to dairy products results in an increase of TPC [21,23,54]. To evaluate the effects of GPP addition to cheese, TPC was determined in control and experimental cheese samples. Statistically significant differences (p < 0.0001) were found among cheeses. As expected, the amount of TPC detected in control cheese was lower than the corresponding GPP enriched cheeses. In particular, the addition of 1% (w/w) of GPP resulted in an increase of 0.42 g GAE/kg of TPC in experimental cheeses. These results confirmed that the addition of GPP to dairy products results in an increase of TPC [21,23,54].

Sensory Test
The spider plot reported in Figure 5 shows the sensory analysis of control and experimental cheeses after 30 d of ripening. Except salt, all other sensory attributes evaluated were different between control and experimental cheeses. The highest scores were registered for odor and aroma intensity in experimental cheeses. The scores for sweet, acid perception, fiber sensation, friability, adhesiveness and humidity of GPP-enriched cheeses were higher than those registered in control cheeses. On the contrary, the cheeses enriched with GPP showed lower scores of sweetness and hardness.
The spider plot reported in Figure 5 shows the sensory analysis of control and experimental cheeses after 30 d of ripening. Except salt, all other sensory attributes evaluated were different between control and experimental cheeses. The highest scores were registered for odor and aroma intensity in experimental cheeses. The scores for sweet, acid perception, fiber sensation, friability, adhesiveness and humidity of GPP-enriched cheeses were higher than those registered in control cheeses. On the contrary, the cheeses enriched with GPP showed lower scores of sweetness and hardness.

Figure 5.
Spider diagrams corresponding to the descriptive sensory analysis of cheeses. Abbreviations: CP, control production inoculated with a commercial starter culture; EP, experimental production inoculated with a commercial starter culture +1% of GPP.
The addition of vegetable by-products to dairy products confirmed the favorable influence on their sensory attributes [20,38,55,56]. As reported by Torri et al. [57], GPP-enriched cheese was described in terms of appearance as marbled with sensations perceived at low intensity of vegetables. However, although the addition of GPP determined large differences in terms of sensory characteristics, the overall assessment of experimental cheese, intended as the degree of overall satisfaction, was higher than the control cheese, indicating a certain appreciation of this GPP-enriched fresh ovine cheese.

Conclusions
The obtained results revealed that addition of GPP into ovine cheese at a concentration of 1% (w/w) did not alter the microbiological parameters during fermentation carried out with the commercial starter culture and resulted in a general appreciation by the judges. Furthermore, GPP addition enriched ewe's milk "Primosale" cheese with TPC. This result is of relevance to improve the antioxidant properties of cheeses, even though further studies, such as in vitro gastrointestinal digestion and radical scavenging activities, are necessary to better investigate the beneficial effect of this novel cheese on human health. The present work provided evidence on the industrialization of the process of GPP-enriched cheese making using commercial starter culture. The addition of vegetable by-products to dairy products confirmed the favorable influence on their sensory attributes [20,38,55,56]. As reported by Torri et al. [57], GPPenriched cheese was described in terms of appearance as marbled with sensations perceived at low intensity of vegetables. However, although the addition of GPP determined large differences in terms of sensory characteristics, the overall assessment of experimental cheese, intended as the degree of overall satisfaction, was higher than the control cheese, indicating a certain appreciation of this GPP-enriched fresh ovine cheese.

Conclusions
The obtained results revealed that addition of GPP into ovine cheese at a concentration of 1% (w/w) did not alter the microbiological parameters during fermentation carried out with the commercial starter culture and resulted in a general appreciation by the judges. Furthermore, GPP addition enriched ewe's milk "Primosale" cheese with TPC. This result is of relevance to improve the antioxidant properties of cheeses, even though further studies, such as in vitro gastrointestinal digestion and radical scavenging activities, are necessary to better investigate the beneficial effect of this novel cheese on human health. The present work provided evidence on the industrialization of the process of GPP-enriched cheese making using commercial starter culture.