Changes in Composition of Some Bioactive Molecules upon Inclusion of Lacticaseibacillus paracasei Probiotic Strains into a Standard Yogurt Starter Culture

Incorporation of probiotic Lacticaseibacillus paracasei into a standard yogurt starter culture can drastically improve its health promoting properties. However, besides being an advantage in itself, the incorporation of a new probiotic strain can significantly affect the overall composition of fermented milk. In this article, the effect of incorporation of the L. paracasei probiotic strains (KF1 and MA3) into several standard yogurt starter cultures (consisting of the following strains: Streptococcus thermophilus 16t and either Lactobacillus delbrueckii Lb100 or L. delbrueckii Lb200) was investigated. Such parameters as the degree of proteolysis, antioxidant activity, ACE-inhibitory activity, content of organic acids, profile of FAs and profile of volatile organic compounds were measured, and the influence of the starter culture composition on these parameters was described. It was demonstrated that, at least in the case of the studied strains, yogurt with L. paracasei had an advantage over the standard yogurt in terms of the content of acetoin, acetic acid, butyric acid and conjugated linoleic acid. Moreover, the incorporation of L. paracasei KF1 significantly improved the hypotensive properties of the resulting yogurt. Thus, the presented study provides insight into the bioactive molecules of probiotic yogurt and may be useful for both academia and industry in the development of new dairy-based functional products.


Introduction
Although fermented milk products have been known since the dawn of civilization, interest in this type of food is now higher than ever.Initially, fermentation was a way to preserve milk for long-term storage and fermented dairy products were considered mainly as a source of nutrition [1,2].Over time, the shelf-life and nutritional value of fermented milk products became less important, while the fresh aroma, pleasant texture and distinctive taste of these products began to dominate consumer preferences.Nowadays, fermented dairy product have become an archetype of functional food-food that provides health benefits beyond its nutritional value upon regular consumption [3,4].
Yogurt is currently one of the most popular dairy products in the world.Although yogurt has been known for centuries as a health-promoting food, the scientific basis for its health benefits was only discovered in the early 20th century [5,6].At this time, the Bulgarian Stamen Grigorov, who had just entered the medical faculty of the University of Geneva, discovered Bacillus bulgaricus (now Lactobacillus delbrueckii subsp.bulgaricus), to the consumption of which in 1909 Yulia Mechnikov attributed the longevity of the Bulgarian population [7,8].Now, countless studies have been conducted on the health benefits of yogurt.It was demonstrated that regular consumption of yogurt can improve digestion, stimulate the immune system, and reduce the risk of cardiovascular diseases and type 2 diabetes [9][10][11].
Classically, yogurt is prepared using only milk and a starter culture consisting of L. delbrueckii and Streptococcus thermophilus [5]; however, over the past decade, many promising varieties of yogurt have been invented, including fruit or flavored yogurt, frozen yogurt, soy yogurt, and others [12].Arguably, the most promising variety of yogurt in terms of its potential health benefits is probiotic yogurt, for which the starter culture contains additional types of lactic acid bacteria (LABs) with probiotic properties.
One of the most promising probiotic LABs that can be incorporated into the standard yogurt starter is Lacticaseibacillus paracasei (formerly Lactobacillus paracasei).L. paracasei is a part of the Lacticaseibacillus casei group that include such widely researched probiotic species as L. casei and Lacticaseibacillus rhamnosus [13,14].Many of the isolated L. paracasei strains have been used as single-strain probiotics or as part of a symbiotic consortium within formulations [15][16][17].It has been shown that regular consumption of certain strains of L. paracasei can strengthen the immune system, alleviate inflammation and reduce the risk of obesity [15,17,18].
Although the incorporation of L. paracasei into the yogurt starter is a benefit in and of itself, it should be mentioned that the fermentation of milk by a specific LAB consortium significantly changes almost all its components.The major carbohydrate present in milk, lactose, is hydrolyzed to glycose and galactose, both of which can be further fermented into organic acids [19].The increased content of certain organic acids during fermentation leads to the unfavorable environment for the growth of pathogenic microorganisms, increases the bioavailability of micronutrients (vitamins) and promotes the absorption of calcium by the intestinal epithelium [20].The major milk proteins (casein and whey proteins) can be break down to small molecular weight peptides (from 2 to 15 aa) with antioxidant, antihypertensive and hypocholesterolemic properties [21,22].
Usually, fermentation does not significantly affect the overall fatty acid (FA) composition of milk; however, enzymatic lipolysis during fermentation can release some flavor forming FAs [23], and more importantly, starter cultures can enrich fermented milk with specific very biologically active FAs, such as conjugated and branched chain FAs [24][25][26][27].Both these groups of FAs possess a suppressing effect on tumor development, artery plaque formation and neurological disturbances [24,25,28,29].Additionally, recent studies identified a very specific class of FA, branched-chain hydroxy-FA, as cell-autonomous metabolic regulators, which help to preserve whole-body glucose homeostasis and alleviate some consequences of obesity-linked type 2 diabetes [30].
All of the mentioned changes play a significant role in the transformation of fermented milk into a functional product, and the inclusion of new LABs in the starter culture can significantly affect them.Currently, there are very few publications that describe how the incorporation of L. paracasei into the standard yogurt starter culture influences the overall composition of the resulted product [31][32][33].
In this work, several previously described L. paracasei probiotic strains (KF1 and MA3) [16] were each incorporated into two different standard yogurt starter cultures, consisting of S. thermophilus strain 16t and either L. delbrueckii strain Lb100 or L. delbrueckii strain Lb200.In products obtained from various combinations of the mentioned strains (as well as in several single-strain fermented products) such parameters as the degree of proteolysis, antioxidant activity, Angiotensin-I-converting enzyme inhibitory (ACE-I) activity, content of organic acids, profile of FAs and profile of volatile organic compounds (VOCs) were measured, and the influence of strain composition on these parameters was described.

Strains and Cultivation Conditions
The strains used as classical yogurt starter culture-S.thermophilus 16t, L. delbrueckii Lb100 and L. delbrueckii Lb200-and probiotic strains-L.paracasei KF1 and L. paracasei MA3-were obtained from the Collection of the All-Russian Research Institute of the Dairy Industry (VNIMI, Moscow, Russia).Upon reception, all strains were stored at −80 • C in skim milk containing 20% (v/v) glycerol (i.e., glycerol-stock cultures).
To obtain the starting broth-stock cultures, the glycerol-stock cultures of S. thermophilus 16t and Lactobacillus spp.were inoculated into M17 broth medium (HiMedia Laboratories, Mumbai, India) and de Man, Rogosa and Sharpe (MRS) broth medium (HiMedia Laboratories, Mumbai, India), respectively.The incubation was carried out overnight at 37 • C for S. thermophiles and L. delbrueckii strains, and at 30 • C for L. paracasei strains.
The working culture of the LAB was obtained by inoculation of reconstituted (12%, w/v) commercial skim milk powder of the "Standard" brand (Complimilk, Slutsk cheesemaking plant, Slutsk, Belarus) with the broth-stock culture (3%, v/v).Before inoculation, the milk was sterilized at 110 • C for 10 min and cooled to approximately 30 • C. Incubation was carried out overnight at the optimal growth temperature for each LAB-37 • C for S. thermophiles 16t and both strains of L. delbrueckii, and 30 • C for both strains of L. paracasei.

Milk Fermentation
The reconstituted skim milk (RSM) was pasteurized at 85 • C for 30 min and then cooled to 40 ± 2 • C. The pasteurized RSM was aseptically inoculated (1.0%, v/v) with the different combinations of working cultures presented in Table 1.The fermentation was carried out at 37 • C until the coagulation of the milk, after which the samples were stored at 4 • C for 12 h for further analysis.All fermentations were performed in triplicate.

Viability of Lactic Acid Bacteria and pH Measurement
The viability of LABs was expressed as the number of colony-forming units (CFUs) per mL of culture.For S. thermophilus the CFUs were selectively enumerated using M17 agar (HiMedia Laboratories, Mumbai, India), and for Lactobacillus spp. the CFUs were selectively enumerated using MRS agar (HiMedia Laboratories, Mumbai, India) adjusted to pH 5.2.The incubation was carried out anaerobically at 37 • C for 24 and 72 h for S. thermophilus and Lactobacillus spp., respectively.Anaerobic conditions were created using Anaero Bag System 24 (HiMedia Laboratories, Mumbai, India) [34].
The pH value of fermented samples was measured with a pH-meter (Mettler Toledo, Griefensee, Switzerland) at room temperature (20 ± 2 • C).

Proteolytic, Antioxidant and Angiotensin-I-Converting Enzyme Inhibitory Activities
The degree of proteolysis, antioxidant activity and ACE-I activity of the samples were measured after the removal of the milk clot.The clot was separated by centrifugation at 3000× g for 15 min at 4 • C (Eppendorf centrifuge 5430 R, Hamburg, Germany), and the resulting supernatant was stored at −20 • C until further use.Before measurements, the supernatant was thawed at 4 • C, centrifuged at 10,000× g for 3 min at room temperature, and filtered through a syringe filter with a 0.45 µm hydrophilic membrane (Merk Millipore, Darmstadt, Germany).
The proteolytic activity in the resulting supernatant was measured spectrophotometrically with the 2,4,6-trinitrobenzenesulfonic acid (TNBS; Sigma-Aldrich, St. Louis, MO, USA) as substrate according to Adler-Nissen [35] with some modifications, as described in Glazunova et al. [34].The degree of proteolysis was determined using a calibration curve constructed with L-leucine (L-Leu) in the concentration range from 0.1 to 2.0 mM, and the proteolytic activity was reported as L-Leu molar equivalents-mM (L-Leu).
The antioxidant activity in the resulting supernatant was determined by the trolox equivalent antioxidant capacity (TEAC) assay with the generation of the 2,2 -azinobis-(3-ethylbenzothiazoline-6-sulfonate radical cation (ABTS •+ ) according to Re et al. [36].The ABTS •+ was prepared by incubation of a solution containing 7 mM ABTS (Sigma-Aldrich, St. Louis, MO, USA) and 2.45 mM potassium peroxodisulfate (Sigma-Aldrich, St. Louis, MO, USA) in the dark at room temperature for 12-18 h.The reaction was recorded by the decrease in OD 734 nm for 40.5 min at 25 • C. The measurements were carried out on a Synergy 2 microplate photometer-fluorimeter (BioTek, Winooski, VT, USA).The antioxidant capacity of samples against ABTS •+ was reported as an amount of trolox molar equivalents-µM (TE).
The ACE activity was measured by the enzymatic cleavage of the o-Aminobenzoyl-Phe-Arg-Lys(dinitrophenyl)-Pro (Sigma-Aldrich, St. Louis, MO, USA), and the fluorescence from the cleaved substrate was recorded with a Synergy 2 microplate photometerfluorometer (BioTek, Winooski, VT, USA).The ACE-I activity of the resulting supernatant was expressed as its half maximal inhibitory concentration (IC 50 ), as described in Torkova et al. [37].

Organic Acid Profile Determination
Organic acid profiles of the fermented milk samples were determined by capillary electrophoresis as described in Rozhkova et al. [38].Analysis was performed using a capillary electrophoresis system (model Kapel-105M, Lumex Ltd., St. Petersburg, Russia) capable of determining the concentration of 10 organic acids, namely oxalic, formic, tartaric, malic, citric, succinic, lactic, acetic, ascorbi, and propionic acids.The data were processed using Elforun ® 205 software (St.Petersburg, Russia).

Fatty Acid Profile Determination
The fats were extracted from the samples of fermented milk by the method proposed by Folch et al. [39].After the extraction, the fats were hydrolyzed with simultaneous derivatization of FAs using a commercial solution of 3 M HCl in methanol (Supelco, Bellefonte, PA, USA).The separation of the derivatized FAs was performed in the regime of temperature gradient on the GC 2010 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with an MDN-5 column (30 m × 0.25 mm; Bellefonte, PA, USA).Separated FAs were analyzed using a QP 2010 quadrupole mass spectrometer (Shimadzu, Kyoto, Japan).The complex qualitative mixture of FAs, PUFA-2 (Supelco, Bellefonte, PA, USA), was used as an analytical standard.The identification of the FAs was carried out as described in Moiseenko et al. [40].The relative intensities (further relative abundances) of the FA were obtained by normalization on the total intensity of the assigned peaks.All experiments were performed in triplicate.
The FA-related nutritional indices were calculated according to Chen et al. [41]: where MUFA stands for monounsaturated fatty acids; PUFA-polyunsaturated fatty acids; IA-index of atherogenicity; HPI-health-promoting index (which is the reciprocal of IA and is mainly used in research on dairy products); IT-index of thrombogenicity; HHhypocholesterolemic/hypercholesterolemic ratio; UI-unsaturation index.

Volatile Organic Compound Profile Determination
The VOC profiles of the fermented milk samples were determined as described in Moiseenko et al. [40].In brief, the VOCs were extracted from the samples of fermented milk by solid phase microextraction using SPME Fiber Assembly Polydimethylsiloxane/Divinylbenzene (PDMS/DVB) cartridges (Supelco, Bellefonte, PA, USA).Extracted VOCs were separated using a GS 2010 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with an Optima-1 column (25 m × 0.25 mm, Supelco, Bellefonte, PA, USA) and analyzed using QP 2010 quadrupole mass spectrometer (Shimadzu, Kyoto, Japan).To determine Kovats retention indices, the column was calibrated using retention index standards (Sigma, USA), containing a mixture of C8-C32 hydrocarbons.Mass detection was carried out in the range of 45-450 m/z, and the relative intensities (hereinafter referred to as relative contents) of individual VOCs were obtained by normalization to the total intensity of all peaks assigned to the VOCs.VOCs were identified using both Kovats retention indices and mass spectra.The mass spectra were compared with the reference spectra from the NIST-EPA-NIH Mass Spectral Database (NIST 11).Only peaks with an assignment reliability greater than or equal to 90% were considered valid for further analysis.

Statistical Analysis
All statistical comparisons between groups of samples were performed in two stages.Firstly, the one-way ANOVA omnibus F-Test was carried out to determine whether all compared groups were the same or if some groups were statistically significantly (p < 0.05) different from the others.If the presence of significantly different groups was detected, in the second stage, multiple pairwise comparisons using Tukey's honestly significant difference (Tukey's HSD) tests were carried out.The differences were considered significant if the p-values were less than 0.05 (i.e., p < 0.05).

Bacterial Growth and Medium Acidification
Data on the bacterial growth and medium acidification for the samples fermented by the different combinations of LABs are presented in Table 2. Generally, acidification capacity is an important technological factor in milk fermentation, affecting both organoleptic characteristics and storage stability of the resulting product [42].For all studied samples, no statistically significant differences (p > 0.05) between the pH values at the end of fermentation were detected, and the pH value comprised 4.3 ± 0.1.In all studied LAB consortia, S. thermophilus 16t was the predominant LAB.The final concentration of its cells was approximately an order of magnitude higher than that for Lactobacillus spp.During milk fermentation, changes in the viable cell counts of S. thermophilus 16t growing in monoculture was about 2.25 lg (CFU•mL −1 ), and its growth was not significantly affected by the presence of the other LABs (Table 2).Generally, CFUs of S. thermophilus 16t reached a value of 8.7 ± 0.1 lg (CFU•mL −1 ) at the end of fermentation; the exceptions were the Str16t + Lb200 and Str16t + Lb200 + KF1 samples, in which the viable cell counts at the end of fermentation were 8.9-9.0 lg (CFU•mL −1 ).
For the L. delbrueckii strains, their co-cultivation with S. thermophilus 16t (i.e., Str16t + Lb100 and Str16t + Lb200) resulted in a slightly higher number of viable cells (7.9 ± 0.1 lg (CFU•mL −1 )) compared to the mono-strain fermentations (7.2 ± 0.4 lg (CFU•mL −1 )).The incorporation of L. paracasei probiotic strains in yoghurt starters did not significantly affect the cell viability of Lactobacillus spp.; the only exception was the Str16t + Lb200 + KF1 sample in which the CFU of Lactobacillus spp. was significantly (p < 0.05) lower than in the other multi-strain fermentations, but at the same level as the fermentations with single strains of L. delbrueckii.

Proteolytic, Antioxidant and ACE-Inhibitory Activities
Proteolytic, antioxidant and ACE-I activities measured at the end of fermentation are presented in Table 3.The highest degree of proteolysis and antioxidant activity (p < 0.05) were detected during milk fermentation with monocultures of L. delbrueckii, and strain Lb100 showed higher proteolytic activity compared to strain Lb200.For the other samples, both the degree of proteolysis and antioxidant activity were the same (p > 0.05).The observed dependency between the degree of proteolysis and antioxidant activity of the samples substantiate previous observations about their positive correlation [43].
The values of ACE-I activity significantly varied between fermentations.It was demonstrated that the incorporation of L. paracasei MA3 to the starter containing S. thermophilus 16t and L. delbrueckii LB100 significantly (p < 0.05) increased ACE-I activity (i.e., decreased the IC 50 value), while its incorporation into the starter containing S. thermophilus 16t and L. delbrueckii LB200 did not produce any effects.The addition of the L. paracasei KF1 strain significantly (p < 0.05) increased ACE-I activity when it was supplemented to both Str16t + Lb100 and Str16t + Lb200 starters.Interestingly, previous in silico analysis of the L. paracasei KF1 and L. paracasei MA3 proteolytic systems [16] showed that the genome of L. paracasei KF1 contained two genes encoding cell envelope proteinases (CEPs)-prtP (WP_003601280) and prtB (WP_003603181)-while the genome of L. paracasei MA3 contained three CEP genes-two prtP (WP_018041452 and WP_216693535) and one prtB (WP_216693433).Hence, the higher number of prtP genes can result in an increased amount of prtP enzymes, which in turn results in a decreased number of ACE-I peptides.

Profile of Organic Acids
Among 10 organic acids, of which the concentration was measured in the studied samples, only citric, formic, acetic and lactic acids were present in detectable concentrations (Table 4).The concentration of citric acid was the same (p > 0.05) in all studied samples.The concentration of formic acid in all fermented samples was the same (p > 0.05) and its value was three-times lower than that in the unfermented milk.The concentration of acetic acid significantly (p < 0.05) decreased in the samples fermented with monoculture of S. thermophilus 16t compared to the unfermented milk; at the same time, its concentration significantly (p < 0.05) increased in the samples fermented with three-strain cultures.Expectedly, lactic acid was not detected in the unfermented milk.The highest concentration of lactic acid was detected in the milk fermented with individual L. delbrueckii strains (Lb100 and Lb200), while the lowest concentration was in the milk fermented with a monoculture of S. thermophilus 16t.The obtained data are generally consistent with those previously reported on organic acid composition of yogurts.Citric acid is always present in yogurt, since it is a natural compound of raw milk [44,45].Typically, the concentration of citric acid in milk amounts to ~1.5 mg•mL −1 and does not significantly change during yogurt fermentation [46,47].However, it should be noted that the small changes (less than 5%) in the concentration of citric acid are generally hard to detect by the current methods, since all these methods rely on an extraction procedure, the efficiency of which can vary greatly between replicates [46][47][48][49].The presence of formic and acetic acids in unfermented milk can be explained by autoclaving.It was previously shown that both of these acids can be readily formed in thermally treated milk [50,51].The significant increase in acetic acid concentration in the samples fermented with three-strain cultures can be explained by the presence of L. paracasei.L. paracasei is a facultative heterofermentative LAB that, in addition to lactic acid, can synthesize acetic acid, acetoin, acetaldehyde, and/or ethanol while grown on hexose sugars [13,52].Moreover, L. paracasei can produce lactic and acetic acids from pentose sugars [53].Additionally, it should be mentioned that L. paracasei and especially S. thermophilus are acetoin-producing LABs that can utilize citric acid through the diacetyl/acetoin pathway with concomitant production of acetic acid [54].

Profile of Total Fatty Acids
The qualitative composition and relative abundances (percentage from total) of FA in the original milk and milk fermented by different combinations of LABs are shown in Table 5.In total, 43 different FAs were detected in all studied samples.All detected FAs were subdivided into eight general groups: medium and long-chain saturated fatty acids (MCSFAs and LCSFAs), monounsaturated fatty acids (MUFAs), polyunsaturated fatty acids (PUFAs), conjugated fatty acids (CFAs), branched-chain fatty acids (BCFAs), 2-hydroxy branched-chain fatty acids (2OH-BCFAs), and oxo-saturated fatty acids (oxo-SFA).The group of LCSFAs was the most abundant one, and all samples contained a similar (p > 0.05) total amount of LCSFAs (50 ± 2%).The second, third, fourth and fifth largest groups were the groups of MUFAs, MCFAs, PUFAs and BCFAs, respectively.Compared to unfermented milk, the total amount of MUFAs significantly (p < 0.05) decreased in the samples fermented with individual L. delbrueckii strains (Lb100 and Lb200) and two the LAB consortiums-Str16t + Lb200 and Str16t + Lb200 + MA3.The total amount of MCFAs significantly (p < 0.05) decreased in the samples fermented with S. thermophilus 16t either individually or as part of the consortium with other LABs and significantly (p < 0.05) increased in the sample fermented with L. delbrueckii Lb100.In contrast to MCFAs, the total amount of PUFAs significantly (p < 0.05) increased in the samples fermented with S. thermophilus 16t either individually or as part of the consortium, with the only exception being the samples fermented with Str16t + Lb100 + KF1 and Str16t + Lb200.The total amount of BCFAs was significantly (p < 0.05) decreased only in the sample fermented with L. delbrueckii Lb100 and significantly (p < 0.05) increased only in the sample fermented with Str16t + Lb200.
The CFAs and 2OH-BCFAs were not detected in the unfermented milk, while their abundance after fermentation varied significantly (p < 0.05) depending on the fermented strains.The highest CFA content (p < 0.05) was observed in the Str16t + Lb200 + KF1, while the highest content of 2OH-BCFAs (p < 0.05) was observed in the samples fermented with L. delbrueckii Lb100.Interestingly, it was previously shown that among dairy products fermented by eight different LAB strains, the milk fermented with L. delbrueckii Lb100 demonstrated significantly better hypotensive and hypocholesterolemic properties in the Spontaneously Hypertensive Rat (SHR) animal model [34].In all the samples only one oxo-SFA, 4-oxo-pentanoic acid (4O-C5:0), was detected; its maximal abundance was observed in the sample fermented with S. thermophilus 16t, and its minimal abundance was observed in the samples fermented with Str16t + Lb100 + MA3, Str16t + Lb200 and Str16t + Lb200 + KF1.
Although the tabular representation of the profile of total fatty acids is quite common, it usually contains an overwhelming amount of information.For a more compact and convenient presentation of the similarities and differences in the FA composition of the studied samples, the well-known dimensionality reduction technique, PCA, was performed.As can be seen from Figure 1, all the samples depicted in the plain formed by the first two principal components (which commonly explained 47% of all observed variations in the data) agglomerated into four clusters: the first cluster was formed by the biological replicas of the unfermented milk; the second and third clusters were formed by the biological replicas of the samples fermented with the individual L. delbrueckii strains Lb100 and Lb200, respectively; and the fourth cluster was formed by the biological replicas of the samples fermented with S. thermophilus 16t either individually or as part of the consortium with other LABs.Moreover, all the samples in the fourth cluster were almost evenly spread around the samples fermented with a monoculture of S. thermophilus 16t.Hence, overall FA composition in all fermentations with a mixture of LABs was primarily guided by the presence of S. thermophilus 16t.It was previously hypothesized that in standard yogurt consortium (consisting of S. thermophilus and L. delbrueckii) S. thermophilus performs the main lipolytic function, while L. delbrueckii performs the main proteolytic function [55].Moreover, it was shown that the set of genes involved in long-chain fatty acid production is significantly upregulated in S. thermophilus during yogurt fermentation [56].In line with that, although the total abundance of LCSFAs was unchanged in our investigation, significant (p < 0.05) increases in the abundances of individual LCSFAs were detected (Table 4), and all these increases were observed in the samples fermented with S. thermophilus, either individually or as part of the consortium with other LABs.Additionally, in our study almost all significant increases in the abundances of individual PUFAs were associated with the presence of S. thermophilus.
To assess the possible impact of the studied samples on the cardiovascular health in individuals who regularly consume yogurt, several well-known nutritional indices related to the FA composition of the samples were calculated (Table 6).Compared to the unfermented milk, all studied samples demonstrated significantly (p < 0.05) decreased IA and increased HPI, both of which convey the same information regarding the probability of developing atherosclerosis with constant use of the sample.The HPI is the inverse of the IA.While the IA is wildly used in the calculation of various diets, the HPI is most often used to assess the FA quality of dairy products [41].The value of the IT, which reflects the probability of thrombosis development [57], was the same (p > 0.05) in unfermented milk and all fermented samples.The value of the HH, which reflects the probability of hypercholesterolemia development [58], was significantly increased (compared to unfermented milk) only in the sample fermented with Str16t + Lb200 + KF1.The value of the UI, which can be used to estimate the content of the high-quality PU-FAs necessary to maintain the fluidity of biological membranes [59,60], was significantly (p < 0.05) increased in the sample fermented with Str16t + Lb200 + KF1 and significantly (p < 0.05) decreased in the samples fermented with individual L. delbrueckii strains (Lb100 and Lb200).Regarding the general use of the FA nutritional indices for dairy products, their limitations should be mentioned.Although these indexes can be useful to some extend while comparing different dairy product with each other [34], they can be misleading while comparing dairy with other products.It is well known that dairy products contain high levels of saturated fatty acids (SFAs), the consumption of which can generally have detrimental effects on cardiovascular health [61,62].However, recent studies suggest that milk has a neutral effect on cardiovascular outcomes, and fermented milk products have a positive or neutral effect [63,64].In contrast, consumption of meat, which along with dairy products is one of the largest sources of SFAs in the human diet, was associated with a neutral or increased risk of cardiovascular disease [65].

Profile of Volatile Organic Compounds (VOCs)
The qualitative composition and relative abundances (percentage from total) of VOCs in the original milk and milk fermented by different combinations of LABs are shown in Table 7.In total, 11 different VOCs were detected in all studied samples.All detected VOCs were subdivided into four general groups: short-chain fatty acids (SCFAs), medium-chain methyl ketones (MCMKs), furan-containing compounds (FCCs), and other VOCs.The measurement error did not exceed the one unit of the least significant digit.ND-not detected.Str16t-Streptococcus thermophilus 16t; Lb100-Lactobacillus delbrueckii Lb100; Lb200-L.delbrueckii Lb200; KF1-Lacticaseibacillus paracasei KF1; MA3-L.paracasei MA3.
While SCFAs were not detected in the samples of unfermented milk, the group of SCFAs was the predominant one in almost all fermented samples.These FAs can be released from the milk fat through the complex biochemical transformations that mainly occur inside the LABs cells.Generally, FAs enter the cell either directly, through special transport systems, or indirectly in the form of triacylglycerides (constituting 98% of total milk fat), which are further hydrolyzed by intracellular lipases and esterases.Inside the cell, saturated FAs can be shortened due to the reaction of the β-oxidation cycle.Each complete cycle removes two carbon atoms from the carboxylic end of the FA.In addition, an incomplete β-oxidation cycle can lead to the formation of β-keto acids, the decarboxylation of which results in the formation of methyl ketones, which are one carbon atom less than the original FA [53,66].Both FAs formed inside the cell and methyl ketones can subsequently be released outside the cell either by diffusion or with the participation of yet unknown transport systems [67].Not surprisingly, the second most abundant VOC determined in our study was methyl ketones with an odd number of carbon atoms.The presence of such ketones in the original milk can be explained either by the metabolic activity of rumen microbiota or by the thermal treatment of milk before fermentation [68,69].Similarly, the presence of 2-amylfuran and furfuryl alcohol is also a result of the milk thermal treatment during sterilization [69].The acetoin was detected only in the samples fermented with S. thermophilus 16t, either individually or as part of the consortium with other LABs, and benzoic acid, only in the sample fermented with L. delbrueckii Lb200.
To investigate how the composition of the starter culture influences the VOC profile, a clustered heat-map was constructed for the volatile organic compounds (VOCs) determined in the studied samples (Figure 2).As can be seen from the clustering pattern of the columns, the VOC profile of the studied samples was primarily guided by the number of LABs in the fermented starter-the three-LABs starter cultures formed a distinct cluster, while the two-LABs starter cultures cluster was nested in a bigger cluster with a single LAB starter culture.Interestingly, the closest to the VOC profile of the cluster containing the two-LABs starter cultures was the VOC profile of the sample fermented only by S. thermophilus 16t.Additionally, the heat map revealed several interesting patterns.VOC profiles of the single and two-LABs starter cultures were both characterized by the higher amount of SCFAs and MCMKs compared to the three-LABs starter cultures.At the same time, VOC profiles of the three-LABs starter cultures were significantly richer in acetoin and slightly richer in butyric acid (with the exception of Str16t + Lb200 + MA3).The acetoin production can be related to the same biochemical root as the production of acetic acidcitrate metabolism [54].In Section 3.3, the significant increase in acetic acid concentration in the samples fermented with the three-strain cultures was especially highlighted, and the lack of citrate depletion was explained by a higher extraction error than the citrate consumption, which is required to produce the observed changes in both acetic acid and acetoin contents.

Conclusions
Summarizing all obtained results, it can be concluded that incorporation of the L. paracasei probiotic strains (KF1 and MA3) into the several standard yogurt starter cultures (consisting of S. thermophilus 16t and either L. delbrueckii Lb100 or L. delbrueckii Lb200) generally did not significantly deteriorate the composition of the resulting yogurts in terms of the biologically active molecules.Moreover, some health-related parameters were improved.The presence of either tested L. paracasei strains increased the content of acetic acid, butyric

Figure 1 .
Figure 1.Results of principal component analysis (PCA) performed on the FA profiles of the studied samples.

Figure 2 .
Figure 2. Clustered heat-map for the volatile organic compounds (VOCs) determined in the studied samples.

Table 1 .
Combination of working LAB cultures used for milk fermentation.

Table 2 .
Bacterial growth and medium acidification in the studied samples.

Table 3 .
Proteolytic, antioxidant and ACE-Inhibitory activities in the studied samples.

Table 4 .
Profile of organic acids in the studied samples.

Table 5 .
Profile of total fatty acids (FAs) in the studied samples.

Table 6 .
Fatty acid (FA)-related nutritional indices of the studied samples.

Table 7 .
Profile of volatile organic compounds (VOCs) in the studied samples.