1. Introduction
Sweet cherry,
Prunus avium L., is a native tree of Europe and western Asia and grows wild around the world [
1]. Its fruit is consumed fresh, but also after processing, as canned, dried, frozen, and in syrups and juices. The main producers of sweet cherry are Turkey, United States, Iran, Spain, Italy and Chile. In Italy the annual production is about 100,000 tonnes, and the Apulia region is the most relevant area of production [
2]. Sweet cherry is characterized by a high content of micronutrients and bioactive compounds, even if these attributes strongly depend on cultivar, ripening, growth condition, pre and post-harvest treatments [
1]. Sugars and organic acids in fruit have been reported in the 125–265 g/kg and 3.67–8.66 g/kg ranges on fresh weight basis, respectively [
3]. The balance between sweetness (sugars, mainly glucose and fructose) and sourness (acids, mainly malic acid) is paramount for the acceptance of the product by consumers [
4], as well as the fruit aroma, despite the fact that aromatic compounds represent only 0.001 to 0.01% of the total fruit weight [
5]. The aroma of cherry is related to a wide number of organic compounds, including aldehydes, alcohols, esters, acids and terpenes [
6]. Aldehydes and alcohols represent more than the 80% of total volatile compounds, followed by acids, esters and terpenes. In particular, the most represented compounds are hexanal, (E)-2-hexenal (green note and fresh green odours), 1-hexanol (floral and grape notes), (E)-2-hexen-1-ol (vegetable note), benzyl alcohol (floral note), and benzaldehyde as the most important contributor of the typical cherry note. Moreover, the aroma of sweet cherry fruits is also influenced by non-volatile glycosidically bound precursors that, in sweet cherry, are more concentrated than the free forms. Glycosylated forms of alcohols, terpenes, norisoprenoids and organic acids are well known, and it is now accepted that the release of these compounds could strongly modulate fruit flavor [
7,
8,
9].
Cherries are also rich in polyphenols, compounds derived from secondary plant metabolism and characterized by one or more hydroxylated aromatic rings: anthocyanins, phenolic acids, and flavonoids are the main phenolics observed in cherries.
Lactic acid bacteria (LAB) are the most widespread microorganisms involved in food fermentation, and their ability to convert phenolic compounds has been reported in literature [
10,
11,
12,
13,
14]. The field of fruit juice fermentation represents a new interesting ever-increasing line of research for product innovation [
15,
16], although the number of fermented commercial products is still limited. The starter strains generally used for fruit fermentation belong to
Lactobacillus plantarum species, recognized to be the most suitably adapted for these types of substrate, and autochthonous strains have often been used for the same reason. Lactic acid bacteria exhibit a great biodiversity, also derived from their ability to adapt to different environments, and this biodiversity can be exploited for different purposes, such as to enhance flavors [
17,
18]. Based on these assumptions, in the present work, the contribution of different LAB species, isolated from dairy and plant products, were evaluated in the framework of cherry juice fermentation. Although
L. plantarum was already used to ferment cherry juice [
19], the novelty of this study was to evaluate the contribution of different LAB species isolated from dairy and plant products. The ability to adapt to this specific matrix, the metabolism of sugars and organic acids, and the effect on the fruit volatile and phenolic profiles were investigated considering a large set of dairy strains as starters.
2. Materials and Methods
2.1. Chemicals
Analytical standards of 3-O-caffeoylquinic acid, 5-O-caffeoylquinic acid, 4-O-caffeoylquinic acid, protocatechuic acid, quercetin-3-rutinoside hydrate, quercetin dihydrate, kaempferol, phenyllactic acid, caffeic acid, naringenin, (+)-catechin, (−)-epicatechin, p-coumaric acid and toluene were from Sigma-Aldrich (St. Louis, MO, USA). Dihydrocaffeic acid, p-hydroxyphenyllactic acid were from Santa Cruz Biotechnology (Santa Cruz, CA, USA), while luteolin was from Extrasynthese (Genay, France). HPLC-grade acetonitrile was from Sigma-Aldrich (St. Louis, MO, USA), while HPLC-grade water and LC-MS grade formic acid were purchased from VWR International (Milan, Italy).
2.2. Bacterial Strains
Fourteen strains belonging to different species of lactic acid bacteria (
L. plantarum, Lactobacillus rhamnosus, Lactobacillus casei and
Lactobacillus paracasei) were singly used for the fermentation of a commercial cherry juice (
Table 1). All bacterial strains were stored at −80 °C in de Man Rogosa and Sharpe (MRS) medium (Oxoid, Milan, Italy) supplemented with 25% glycerol (
v/v). The cultures were propagated three times with about 3% (
v/v) of inoculum in MRS and incubated in anaerobiosis (AnaeroGen, Oxoid, Basingstoke, UK) overnight at 30 °C for
L. plantarum and 37 °C for
L. rhamnosus, L. casei and
L. paracasei.
2.3. Fermentation Process and Storage
A commercial pasteurized cherry juice (Bionaturae) was used for fermentation. The absence of microbial contamination in the juice was evaluated on Plate Count Agar at 30 °C and 37 °C. LAB strains were cultivated in MRS broth for 15 h at 30 °C for L. plantarum and at 37 °C for L. rhamnosus, L. casei and L. paracasei, to reach the late exponential growth phase, centrifuged at 10,000 g for 10 min at 4 °C, washed twice with Ringer’s solution (Oxoid, Milan, Italy) and re-suspended in sterile distilled water. These cultures were individually inoculated into cherry juice to reach the final concentration of ca. 7 Log CFU/mL. The juices were incubated at 30 °C for L. plantarum and at 37 °C for L. rhamnosus, L. casei and L. paracasei for 48 h and then stored for 12 days at 4 °C. Unfermented cherry juices (not added of starter cultures) were incubated at 30 °C and at 37 °C for 48 h, then stored for 12 days at 4 °C and used as controls. All the fermentations were carried out in triplicate.
2.4. Evolution of Bacterial Growth and Acidification of Cherry Juice
Cherry juice, inoculated with starters, was analyzed before and after fermentation (48 h) and after the storage period (12 days). Cultivable cells were determined using the standard plate count agar method as follows: decimal dilutions of samples were carried out in Ringer solution (Oxoid, Milan, Italy) and plated on MRS agar, then incubated at 30 °C (L. plantarum) and 37 °C (L. rhamnosus, L. casei and L. paracasei) for 48 h under anaerobic condition. The pH of samples was measured using a pH metre (Mettler Toledo, Greifensee, Switzerland). Plate count and pH measurement was carried out in triplicate.
2.5. Sugars and Organic Acid Analysis
Sugars and organic acids were analyzed after fermentation and storage, both for fermented juices and for controls. The method reported by Cirlini et al. [
20], with slight modifications, was applied. Briefly, 10 µL of samples was added to 1 mL of a solution containing two internal standards (turanose and glutaric acid 500 µg/mL each), dried under vacuum and dissolved with 500 µL of dimethylformamide. Subsequently, silylation was carried out adding 400 µL of hexamethyldisilazane and 200 µL of trimethylchlorosilane to the samples and heating for 30 min at 70 °C. Samples were analyzed by a Thermo Scientific Trace 1300 gas chromatograph coupled to a Thermo Scientific ISQ single quadrupole mass spectrometer equipped with an electronic impact (EI) source on a BP5MS capillary column (30 m × 0.25 mm, with 0.25 µm film thickness, SGE Analytical Science, Milan, Italy). Chromatographic conditions were the following: initial oven temperature, 60 °C, then increase of 20 °C/min up to 280 °C; carrier gas, (flow rate, 1 mL/min). Temperature of the transfer line was maintained at 280 °C, while the ion source was set at 230 °C. The acquisition mode was full scan (m/z: 40–550). Signals were identified on the basis of their mass spectra compared with those present in the instrument library (NIST 14). In addition, once the glucidic and organic acid fractions were recognized, proper analytical standards were used in order to confirm the identifications. The semi quantification of all detected gas-chromatographic signals was performed on the basis of the use of two internal standards: turanose for sugar quantification and glutaric acid for organic acid quantification. For each identified compound, the Response Factor (RF) was calculated and the values range between 0.8–1.2.
2.6. Characterization of the Volatile Profile
The volatile profile of fermented and unfermented samples was analyzed after 48 h of incubation and after 12 days of storage. Volatiles were characterized by HS-SPME/GC-MS (Head Space-Solid Phase Microextraction/Gas Chromatography-Mass Spectrometry) technique following the protocol reported by Ricci et al. [
18]. In brief, 2 mL of cherry juice was placed in a glass vial and added to an aqueous toluene standard solution (0.25 µg/mL). Head space micro-extraction was performed for 30 min at 40 °C after 15 min of equilibration time. A SPME fiber coated with 50/30 µm of Divinylbenzene–Carboxen–Polydimethylsiloxane (DVB/Carboxen/PDMS) was used (Supelco, Bellefonte, PA, USA). The desorption of volatiles was accomplished by exposing the fiber into the GC injector for 2 min at 250 °C. GC–MS analyses were performed on a Thermo Scientific Trace 1300 gas chromatograph coupled to a Thermo Scientific ISQ single quadrupole mass spectrometer equipped with an electronic impact (EI) source. All samples were injected in splitless mode. Helium was used as carrier gas, with a total flow of 1 mL/min. The separation was performed on a SUPELCOWAX 10 capillary column (Supelco, Bellefonte, PA, USA; 30 m × 0.25 mm × 0.25 µm) with the following program gradient: initial temperature, 50 °C for 3 min, linear increase by 5 °C per minute to 200 °C, then maintained for 12 min. The transfer line temperature was 250 °C. The signal acquisition mode was full scan (from 41 m/z to 500 m/z). The main volatile compounds of cherry juices were identified both on the basis of their mass spectra compared with the library NIST 14 mass spectra, as by calculation of linear retention indices (LRI). The semi-quantification of all detected gas-chromatographic signals was performed on the basis of the use of an internal standard (toluene).
2.7. Characterization of Polyphenolic Profile of Fermented and Unfermented Cherry Juices
All fermented and unfermented samples were analyzed by an Accela UHPLC 1250 equipped with a linear ion trap-mass spectrometer (MS) (LTQ XL, Thermo Fisher Scientific Inc, San Jose, CA, USA) fitted with a heated-electrospray ionization probe (H-ESI-II, Thermo Fisher Scientific Inc, San Jose, CA, USA). Separation was performed on an Acquity UPLC HSS T3 (2.1 × 100 mm) column coupled with a pre-column Acquity UPLC HSS T3 VanGuard (2.1 × 5 mm) (Waters, Milford, MA, USA). The volume injected was 5 μL, and oven temperature was set to 40 °C. Phenolic profiling was performed following the protocol reported by Ricci et al. [
12]. Briefly, the mobile phase was 0.1% (
v/v) acetonitrile (phase A) and 0.1% (
v/v) aqueous formic acid (phase B). Elution was performed at a flow rate of 0.3 mL/min. The gradient started with 95% B and 5% A for 0.5 min, then eluent B decreased at 49% and A increased at 51% in 9 min. After 0.5 min, the column was flushed, setting the eluent percentages at 20% B and 80% A for 11.00 min. Finally, the initial conditions were restored (total run time = 17 min). Data processing was performed using Xcalibur 2.2 software from Thermo Fisher Scientific Inc, (San Jose, CA, USA).
2.8. Statistical Analysis
To evaluate the normal distribution for each group of independent samples Shapiro-Wilk test was used. One-way ANOVA was applied to discriminate the significant differences among the samples, applying Bonferroni post hoc test and the results were considered different for values of
p < 0.05. All the detected compounds (volatiles, phenolics, organic acids and sugars) were used as variables for Principal Component Analyses (PCA), which was performed by applying a correlation matrix. All the mentioned analyses were performed on SPSS Statistics 21.0 software (SPSS Inc., Chicago, IL, USA), while hierarchical clustering and heat map were carried out using Heatmapper [
21].
4. Discussion
L. plantarum is the most employed species for lactic acid fermentation of fruit juices, as these strains isolated from plant environments are better adapted to these matrices [
10]. Indeed, chemical characteristics of fruits, such as pH and the concentration of sugars and organic acids, make them a hostile environment for microorganisms [
22,
23,
24]. However, a great biodiversity among LAB, species- and strain-dependent, was observed, which allowed them to adopt specific alternative metabolic pathways in adverse conditions using non-conventional carbon sources for the exploitation of alternative substrates or in a global stress response. This adaptation may result in formation of volatile compounds [
17,
18,
23], metabolism of phenolic compounds [
11,
25], and production of new molecules. Starting from these premises, a large number of dairy strains were used for cherry juice fermentation, and a global view on the metabolism of both dairy and plant-derived microorganisms was proposed.
Cherry juice is recognized to be a stressful substrate for microorganisms, due to its low pH, its high sugar and phenolic content, and the presence of malic acid [
23]. This unfavorable environment affects the growth and cell viability of most tested strains, with a particular negative effect on
L. casei and
L. rhamnosus, which were not able to grow satisfactorily in this medium. On the contrary, all
L. plantarum strains, both of dairy and plant origin, showed a good adaptability. Moreover,
L. rhamnosus 2360 and
L. paracasei 4186 were also able to ferment cherry juice components and to survive during refrigerated storage, also offering the possibility of conveying viable cells of non-plant origin in this fruit juice. PCA analysis showed a clear separation between unfermented and fermented juices, on account of the significant modulation of the composition of the juice matrix by lactic acid fermentation.
The adaptation to the acid environment of cherry juice is based on the ability of LAB to metabolize malic acid [
24], resulting in its almost complete degradation. Actually, all tested dairy and plant strains, independently of species and with the exception of
L. plantarum C1, completely converted malic acid into lactic acid. Indeed, most LAB can convert malic acid into lactic acid thanks to the malolactic enzyme, decarboxylating malate to lactate by a NAD
+ and Mn
2+-dependent malolactic enzyme [
23,
26,
27]. At low pH, the choice of malic acid as the preferred energy source over glucose had already been reported in literature and linked to the increase of intracellular pH and the increase of reducing power [
28], with an associated modification of cellular permeability [
23]. In the present work, LAB did not consume glucose and fructose during fermentation, whereas sucrose showed a marked decrease. The metabolic activity of the tested strains, especially in
L. rhamnosus and
L. paracasei, is mainly based on malolactic fermentation, aimed at the maintenance of vitality, rather than on sugars fermentation, which is mainly correlated with growth [
23]. The exploitation of an organic carbon source by a microorganism strongly depends on the type of substrate. Two different works on cherry juice fermentation reported the absence of sugars metabolism in one case [
23], and a slight consumption after fermentation in the other [
2], the latter being in very good agreement with our results. Observed differences may be related to cultivar, to ripening time, and to the applied processing conditions, which could affect the concentration of nutrients and/or acids [
1,
29]. In the present work, tartaric acid was also consumed by LAB, especially by
L. plantarum, in agreement with literature, where
Lactobacillus spp. were reported to be able to convert it into oxaloacetic acid (and then into lactic acid, acetic acid, and CO
2), thanks to the tartrate dehydratase enzyme [
30,
31,
32]. Nevertheless, the degradation of tartaric acid is not widespread in LAB, and it has been studied especially in wine. Moreover, not all LAB were able to metabolize tartaric acid in the same way; for instance, differently from
L. plantarum, in
Lactobacillus brevis, succinic acid can be produced instead of lactic acid [
33]. The acetic acid found in several fermented fruit juices [
18,
23,
34] is characterized by sharp, pungent and vinegar notes. However, thanks to specific bacterial metabolic traits, acetic acid can be converted to the corresponding esters [
35,
36], such as propyl acetate; this was especially observed after fermentation with
L. rhamnosus 2360 and at the end of storage using
L. paracasei 4186. The increase of β-linalool, observed in different fermented samples, could be ascribed to glycosylases produced by LAB, involved in the release of aglycones from glycosylated terpenes [
18]. Acetoin was also affected by fermentation and its bio-synthesis can be derived from citrate metabolism [
18,
27].
Phenolic compounds have been reported to exert health benefits in humans [
37], to exhibit antimicrobial activity, and to impact the flavor, taste and color [
10] of plant products [
38]. Their beneficial activities have been partially correlated with microbial metabolism that can occur during fermentation processes [
39]. Moreover, the metabolism of phenolics may contribute to bacterial stress response when microorganisms are in hostile conditions [
19,
40].
In the present study, hydroxycinnamic acids were transformed by LAB, ideally through phenolic acid decarboxylases or phenolic acid reductases [
19]. Caffeic acid was effectively converted into dihydrocaffeic acid, as also already reported in the literature for
L. plantarum POM1, [
11,
12,
19], with a potential impact on health, due to its more effective capacity to inhibit platelet activation than its phenolic precursor [
41] and to its antioxidant effect on endothelial cells [
42]. In addition to putative effects on the health-promoting activity of fermentation metabolites, the microbial transformation exerted by LAB may be relevant for other purposes. For example,
p-coumaric acid was decarboxylated into
p-vinylphenol and subsequently reduced, possibly by a phenolic acid reductase, to the phenolic volatile 4-ethylphenol, which contributes to aroma in fermented food [
19]. The reduction of
p-vinylphenol was favoured under anaerobic conditions or in absence of electron acceptors, i.e., when fructose is found at high concentration, to increase NAD
+ quantity [
19,
43]. Protocatechuic acid was also metabolized [
11,
12], in particular by
L. plantarum 285 and POM1, reaching complete depletion, but catechol was not detected. Finally, worthy of note is that phenyllactic acids were produced ex-novo by all the tested strains, probably deriving from amino acid metabolism. Phenylalanine can be converted into phenylpyruvic acid by a transamination reaction, and finally metabolized into phenyllactic acid by hydroxyacid dehydrogenase, while
p-hydroxyphenyllactic acid may originate from tyrosine metabolism [
44]. Their antimicrobial activity against pathogenic strains and moulds has been well documented [
44,
45], even if the concentrations found in the present study were lower compared to those explaining antimicrobial activity.