Isolation, identification, and selection of strains as candidate probiotics and starters for fermentation of Swedish legumes

Background The non-dairy sector is growing, fermented alternatives to dairy are sparse. Adapted starter cultures to substituting raw materials needs to be developed. Objective Aims of this study were to isolate, identify, and phenotypically characterize lactic acid bacteria (LAB) that inhabit Swedish legumes, and assess properties necessary for selecting strains with the ability to ferment a bean beverage and with potential health beneficial properties. Design Isolates of presumed LAB were obtained from legumes collected at Öland, Sweden. Strain diversity was assessed by repetitive polymerase chain reaction (rep-PCR). The strains were identified using matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOF MS). Species belonging to Enterococcus were predominant along with Pediococcus and closely related Bacillus. Strains were tested for tolerance to low pH, phenol, and bile as well as their bile salt hydrolase (BSH) activity. In addition, Enterococcus strains were tested for antibiotic resistance, and Pediococcus strains for their ability to ferment a bean beverage. Results From the 25 strains characterized, five were found resistant to low pH, bile, and phenol, suggesting that they can survive a passage through the gastrointestinal tract (GIT) and hence potentially exert beneficial effects in the host. These are suggested for further investigation on specific host-beneficial properties. Two of these, belonging to Pediococcus pentosaceus, were able to ferment a bean beverage without any added nutrients, indicating that the Pediococcus strains are well adapted to the bean substrate. One of the P. pentosaceus strains were also able to markedly improve the reduction of phytate by the phytase-producing yeast strain Pichia kudriavzevii TY1322 during co-fermentation as well as increase the final cell count of the yeast strain. Conclusion Strain isolation and characterization performed in this study aids in selecting starter cultures for legume fermentation. Nutritional properties can be improved by co-fermentation with yeast indicating that novel nutritious fermented non-dairy products could be developed.


2
(page number not for citation purpose) Inger-Cecilia Mayer Labba et al. P ulses, such as beans and lentils, have a favorable nutritional composition with high contents of protein, dietary fibers, minerals, vitamins, and phytochemicals, and positive health effects have been demonstrated (1). Furthermore, pulses have been suggested as alternatives to resource-demanding animal products, and could contribute to climate change mitigation by lowering greenhouse gas emissions in arable systems (2).
Foods with a low environmental impact need to become more accessible, economically sustainable, and attractive for both farmer and consumer. In the Scandinavian countries, where cultivation of soy is not suited to the climate, other domestic protein crops, such as fava beans, have a large potential in partly replacing animal products. This raises the need to develop new palatable food products with low environmental impact. Fermentation is a well-known process with low energy use that can provide several positive nutritional and sensory properties to raw materials such as grain legumes (3,4). Depending on the organisms used, the end product will have different properties and nutritional values (4). Previous studies have shown that certain microorganisms can synthesize vitamins such as folic acid, riboflavin, niacin, thiamin (5,6), and B12 (7) using grain legumes as a substrate. Some bacteria are able to metabolize at least two of the main components (n-hexanal and pentanal) that give a bean-like flavor to soy milk (8). Furthermore, fermentation lowers the level of oligosaccharides, which can cause postprandial flatulence from legumes (9,10), and other anti-nutritional components such as tannins, phytate (11) , and trypsin inhibitors (12). Different strains within the same species can have completely distinct metabolic patterns, which, in turn, affect not only synthesis of vitamin and breakdown of undesirable components but also production of substances that affect taste and texture of the product, such as exopolysaccharides (13).
Lactic acid bacteria, LAB, are a vast group of bacteria within the phylum Firmicutes and the class Bacilli, showing considerable genotypic and phenotypic diversity. The closely related genus Bacillus belongs to the same class (Bacilli) and hence related to LAB. Both LAB and Bacillus include probiotic as well as pathogenic organisms. LAB are the most used bacteria within the food sector and many strains have been graded as Generally Recognized as Safe (GRAS) strains. Despite generally not being able to reduce the levels of phytate, LAB can provide favorable conditions for phytase activity and contribute indirectly to an improved phytate reduction important to improve bioavailability of essential dietary minerals, such as iron and zinc, in cereals and legumes (14). Co-fermentation between LAB and phytase-producing yeast strains takes place in many natural food fermentations (15,16). Cofermentation between LAB and yeast has been suggested to improve sensory properties and prolong shelf life as well as improve resistance to spoilage molds (17). This indicates interesting applications within the food industry for new strains of LAB.
Despite already extensively used in the food industry, there is still a need to find and develop starter cultures functionally adapted to different types of plant-based raw material as alternative to dairy products. Apart from the functional and nutritional improvements during fermentation, LAB can also have probiotic properties. According to the World Health Organization (WHO), probiotics are 'live microorganisms which, when administrated in adequate amounts, confer a health benefit to the host' (18). Health benefits include reduction in hypercholesterolemia (19), reduction in symptoms of irritable bowel syndrome (20,21), normalization of disturbed gut microbiota, regulation of intestinal transit, competitive exclusion of pathogens, and production of short chain fatty acids (22). The importance of gut microbiota in health and disease has become increasingly evident, and is reflected in the popularity of fermented and probiotic foods among consumers and their rapid market growth (23). The gut microbiota can be modified by the food consumed, for instance by dietary fiber present in legumes, as well as live microbes present in fermented food (24). New probiotics should be selected based on their strain characteristics. The main selection criteria for probiotics include many functional properties, such as tolerance to gastric acidity, bile, and phenol toxicity and ability to de-conjugate bile salts (25). In the present study, we searched for, and initiated studies of, strains that efficiently metabolize complex carbohydrates in pulses, with the long-term goal to develop new types of digestible, palatable, and nutritious fermented legume-based foods.
The aim was to identify and isolate strains with the ability to ferment a fava bean beverage and to characterize LAB isolated from Swedish grain legumes, perform co-fermentation to reduce phytate, a major inhibitor of iron and zinc absorption, and initial test of probiotic potential of the strains.

Plant samples
A total of 23 samples were collected from fields (n = 12) and from stored seeds (n = 11) derived from plants cultivated during 2015. The 12 field samples came from five different locations of Öland, Sweden. On each site, approximately 50 g were collected by picking whole pods directly from the plant with sterile gloves. Samples were placed in clean bags and stored at 4°C until they were used for analysis. The remaining 11 samples were seeds that were harvested at Öland, dried and stored at 4°C for 8 months by the producer. The samples were different varieties of Phaseolus vulgaris (black beans, brown beans, kidney beans, navy beans, borlotti beans, and yin yang), and two varieties of Pisum sativum (grey and yellow peas).

Characterization of LAB for fermentation of Swedish legumes
Isolation of LAB For the selection of LAB from the bean samples, about 1 g of bean pods were aseptically transferred to de Man, Rogosa and Sharpe (MRS) broth (Sigma-Aldrich, Missouri, USA) with pH adjusted to 5.7 using HCl (Sigma-Aldrich). The flasks were incubated at 37°C for 4 days in a shaking incubator (Barnstead Lab-Line, ON, Canada) at 190 rpm. Dilutions of the liquid were spread in parallels on MRS agar plates (Scharlab, Spain); these were then incubated at anaerobic conditions using Anaerocult ® A (Merck Millipore, Germany) at 37°C for 72 h. Colonies were picked out and re-streaked on MRS agar plates until isolated bacterial strains were obtained. Strain isolation was made for two subsequent times for each sample.

Identification of isolates
Colonies that were Gram-positive, oxidase-negative, and catalase-negative were selected for further analyses. These isolates were harvested for DNA extraction and stored at -80°C.

Repetitive polymerase chain reaction (rep-PCR) genomic fingerprinting
To minimize the risk of duplicate strains, we performed rep-PCR. Genomic DNA was extracted from 116 isolates according to a previously described protocol (26). One loop of bacteria incubated at 39°C for 74 h on horse blood agar plates (Scharlab) were inoculated in 50-µL lysis buffer (10 mmol/L Tris HCl, 1 mmol/L EDTA, and 10 mmol/L saline; Sigma-Aldrich) and incubated for 10 min at 95°C. Samples were then centrifuged for 5 min at 12,000 × g. The supernatant was collected, measured for DNA concentration using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA), and stored at −20°C until used.
Amplification products were resolved by electrophoresis on a 25-cm gel in a 40-cm cell (Bio-Rad, Sub Cell GT, California, USA) at a constant voltage of 1.5 V/cm for 16 h in 4-mm 1.5% (w/v) agarose (Thermo Fisher) gel in 1x TBE buffer (Sigma-Aldrich). A DNA molecular weight ladder (Thermo Fisher) was used as a reference. Gels were stained post-run with 3x Gelred in 0.1 M NaCl (Biotium, California, USA) and visualized by UV light using a Gel Doc (BioRad).
The PCR fingerprints were analyzed using BioNumerics (Applied Maths NV, Sint-Martens-Latem, Belgium). Similarity of digitalized band patterns was calculated using the Sørensen-Dice coefficient. Complete linkage algorithms were used to construct an average linkage dendrogram to show relationship of isolates.

MALDI-TOF MS identification of isolates
The 116 bacterial isolates were analyzed twice with matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) analysis using intact cell biomass obtained from overnight cultures of LAB at 37°C on horse blood agar plates. Samples were spotted in duplicates on disposable target plates ( bioMeriéux, Marcy l'Etoile, France) and overlaid with 1 µL VITEK ® MS-FA (Formic acid; bioMeriéux), then air-dried and again overlaid with 1 µL VITEK ® MS-CHCA matrix solution (α-cyano-4-hydroxycinnamic acid; bioMeriéux). MALDI-TOF MS analysis was performed using a VITEK ® MS instrument (bioMeriéux) in the mass range of 2,000-20,000 m/z, and the spectral data were analyzed using the VITEK ® MS IVD and VITEK ® MS SARAMIS/RUO databases (bioMeriéux).
Phenotypic tests related to probiotic potential A number of prerequisites need to be fulfilled for a strain to be considered a candidate for further investigation of its probiotic properties. We studied the ability of strains to survive the harsh conditions such as in the gastrointestinal tract (GIT), including tolerance to bile, low pH, and phenol.
Tolerance to low pH Survival of strains in acidic conditions was asessed by incubation in low pH growth media according to a previously described method (28) using pH 2.5. Unique strains, which we identified by MALDI-TOF, were propagated twice in MRS broth for 24 h at 37°C, collected by centrifugation (at 5,000 × g, for 5 min), and washed twice in phosphate buffer solution (PBS; Thermo Fisher). After propagation and dilution, cells were inoculated in triplicates into 8 mL of MRS broth adjusted to pH 2.5 to simulate gastric conditions. Cells were then incubated at 37°C for 2.5 h with continuous stirring using a tube rotator (New Brunswick Scientific, New Jersey, USA). Bacterial enumeration was performed using five replicates per dilution on MRS agar plates at 0, 0.5, 1, 1.5, 2, and 2.5 h.

Phenol tolerance
The isolates' ability to survive in the presence of phenol, which can be present in the gut, was examined according to a previously described method (29), with minor modifications. Unique strains were prepared in the same way as described for the low pH test. The OD 600 was adjusted to 0.4-0.6, then 200 μL of the diluted cells were inoculated in duplicates into 8 mL of MRS broth, pH 5.7, with 0.4% phenol (Sigma Aldrich). The cells were then incubated in a shaking incubator at 37°C for 24 h. Bacterial enumeration was performed using five replicates per dilution on MRS agar plates at 0 and 24 h.

Bile tolerance
Tolerance of the isolates to bile salts was assayed as described by Gilliland et al. with slight modifications (30). Prior to the bile tolerance test, strains were prepared in the same way as described for the low pH test. The OD 600 was then adjusted to 0.4-0.6 with PBS, and 200 μL of the diluted cells were inoculated in duplicates into 8 mL of MRS broth with 0.3% (w/v) bile salts (Sigma Aldrich). Cells were then incubated in a shaking incubator at 37°C for 48 h. The bacterial suspension was assessed as colony forming units/milliliter (CFU/mL) using five replicates on MRS agar plates at 0, 24, and 48 h.
Bile salt-MRS agar plates were plated in duplicates with each representative strain and incubated anaerobically using an Anaerocult ® A at 37°C for 72 h before observation. A precipitation zone surrounding colonies indicated BSH activity.
Screening of isolated Enterococcus strains for antibiotic resistance Isolates identified as Enterococcus species were tested for antibiotic resistance using the disk diffusion test according to the Clinical and Laboratory Standards Institute (CLSI) (32). The isolated Enterococcus was tested for resistance to ampicillin, trimethoprim, vancomycin, and nitrofurantoin.

Production of bean flour
Dried fava beans were soaked overnight in deionized water with 2% (w/v) NaHCO 3 (Sigma Aldrich), then peeled by hand and incubated at 95°C in deionized water with 0.5% (w/v) NaHCO 3 for 4 min in a water bath. The boiled beans were then blended in 90°C deionized water for 60 s and put on an ice bath to cool before freeze-drying. The freeze-dried flour was then stored at −18°C until use.

Fermentation of fava bean flour medium by
Pediococcus pentosaceus A bean flour medium was prepared using deionized water with 6% (w/v) fava bean flour brought to a boil, with pH adjusted to 6.
2, filtered through a cheese cloth and autoclaved. Selected strains belonging to Pediococcus pentosaceus were pre-cultured overnight in liquid MRS at 37°C in a shaking incubator. The cells were washed twice in PBS and diluted in PBS to an OD 600 of 0.5-0.6. In triplicates, 8 mL of the fava bean medium were inoculated with 200 μL of diluted cells from five selected strains. Bacterial enumeration was performed using five replicates per dilution on MRS agar plates at 0, 4, 6, 8, 12, 24, and 48 h.
Co-fermentation of fava bean flour medium Bean flour medium was prepared as described for fermentation by P. pentosaceus. Pichia kudriavzevii TY1322 (33) and P. pentosaceus strain 77 were pre-cultured overnight in liquid MRS and yeast extract peptone dextrose (YPD) medium respectively at 37°C and 220 rev/min in a rotary shaker. Cells were washed twice in PBS and diluted to an OD 600 of 0.55-0.60. In triplicates, 80 mL of the fava bean medium were inoculated in three combinations: (1) 600 μL of diluted yeast cells + 600 μL of diluted bacterial cells, (2) 600 μL of diluted bacterial cells, and (3) 600 μL of diluted yeast cells. Microbial enumeration was performed using five replicates per dilution on agar plates at 0, 4, 6, 8, 12, 24, 36, 48 h. YPD agar plates with 100-mg/L chloramphenicol (Sigma-Aldrich) were used for combinations 1 and 3. MRS agar plates with 0.5-mg/mL cycloheximide (Sigma-Aldrich) were used for combination 1 and 2. Each time, set microbial enumeration was performed and samples for phytic acid analysis were withdrawn from the fermentation media.

Phytate analysis
Fava bean flour medium (500 μL) was collected during fermentation and immediately mixed with 500 μL of 1 M HCl to stop enzymatic and microbial activity. Samples were stored at −20°C until analysis. Samples were thawed and mixed using a laboratory shaker (Heidolph Reax 2; Heidolph Instruments GmbH, Schwabach, Germany) for 3 h, centrifuged, and the supernatant was transferred to an HPLC vial. Phytate was analyzed as inositol hexaphosphate (InsP 6 ) by high-performance ion chromatography (HPIC) according to Carlsson et al. Characterization of LAB for fermentation of Swedish legumes column reaction with ferrous nitrate, and detected with at 290 nm in a UV-visible HPLC detector (UV-4075; Jasco). Each sample had a run time of 7 min, and the InsP 6 concentration was calculated using external standards covering the concentration range of 0.1-0.6 μM/mL.

Isolation and identification by Rep-PCR and MALDI-TOF MS
A total of 116 isolates of putative LAB were originally obtained from 23 legume samples. Strains fulfilling all criteria used to select LAB and closely related lactateproducing bacteria were isolated from each legume, except from the samples of black beans, from which we could not isolate any LAB. Using rep-PCR fingerprints, the 116 isolates were clustered into five main groups. Of these, fingerprints showing >90% similarity were assigned to the same genotype and were considered a cluster, as suggested by Maluping et al. (35).
From each cluster, one strain was randomly picked as a representative of the group for further analysis, resulting in a total of 25 unique strains (Fig. 1). MALDI-TOF MS identified the strains as Enterococcus hirae, Enterococcus faecium, Enterococcus mundtii, Bacillus coagulans, and P. pentosaceus.
The collected borlotti beans were found to have the highest LAB diversity, with strains from almost all identified genera present. Conversely, no LAB was isolated from black beans despite repeated enrichments. In total, eight unique strains of P. pentosaceus and six unique strains of B. coagulans were found from the plant samples (Table 1).

Phenotypic tests related to probiotic potential
Tolerance to low pH Each of the unique B. coagulans strains showed a high degree of tolerance to pH 2.5 and had an exponential growth during the entire incubation ( Fig. 2A). Strains belonging to E. faecium and E. mundtii showed a generally low tolerance to pH 2.5 and were completely depleted after incubation (Figs 2B and 2C). P. pentosaceusstrains (Fig. 2D) showed two distinctive growth patterns; strain number 77, 100, 101, and 176 showed tolerance to low pH with small variations of CFU/mL. On the contrary, P. pentosaceus strains 96, 97, and 177 did not survive incubation (Fig. 2D).

Phenol tolerance
Overall, strains from all genera showed tolerance to 0.4% phenol with a few exceptions. These were B. coagulans strains 222 and 224 (Fig. 3A) and E. mundtii strain 94 Fig. 1. Dendrogram generated after cluster analysis of the digitalized (GTG) 5 -PCR fingerprints of the isolated strains. Cluster analysis was made using the Sørensen-Dice coefficient and a complete linkage algorithm.    Characterization of LAB for fermentation of Swedish legumes (Fig. 3D). In this experiment, strains 222 and 224 showed a decrease of 30-40% in CFU/mL. Among other isolates, E. mundtii strains 87, 111, and 146 showed an average increase of 29% in CFU/mL (Fig. 3D). Strains belonging to P. pentosaceus (Fig. 3E) exhibited a high tolerance with a consistent level of CFU/mL before and after incubation for strains 77 and 101, while strains 96-100, 125, 176, and 177 showed an average increase of 39% in CFU/mL.

Fermentation of fava bean flour medium by P. pentosaceus
Each of the isolated and unique P. pentosaceus strains was able to grow in 6% fava bean medium without addition of carbon or any other nutrients. The strains showed a growth pattern with an initial lag phase followed by exponential growth. During the exponential phase, the strains differed in generation time. Strains 97 and 100 had a generation time of approximately 1 h, calculated by regression of data points between 2 and 8 h. Strains 77 and 96 had a generation time of approximately 2 h during the same time frame. Strain 176 had an intermediate generation time of 1.4 h under the given conditions. All strains reached a stationary phase after 8 h of incubation (Fig. 5).
Co-fermentation of fava bean flour medium by P. pentosaceus 77 and P. kudriavzevii TY1322 Yeast strain P. kudriavzeviiI TY1322 showed improved growth during co-fermentation with P. pentosaceus strain 77 compared to fermentation in monoculture. In both monoculture and co-fermentation, the yeast strain reached a stationary phase after 36 h, but in co-fermentation, the cell count of P. kudriavzeviiI TY1322 was 124% higher compared to fermentation in monoculture (Fig. 6).
The reduction of phytate in the fava bean medium by P. kudriavzevii TY1322 was significantly improved using a co-starter culture compared to fermentation monoculture (Fig. 7). After 48 h of fermentation, the phytate content in the medium fermented by P. kudriavzevii TY1322 in monoculture had decreased by 25%. The phytate content in the medium fermented by a mixture of P. kudriavzevii TY1322 and P. pentosaceus 77 was reduced by 77%. P. pentosaceus strain 77 was not able to reduce the phytate content by monoculture.

Discussion
There are strong arguments to change the food patterns in industrialized countries into more healthy and sustainable ones. Pulses, such as beans and lentils, are often described as a nutritious and sustainable protein source that could be used as alternatives to meat and dairy products.

Characterization of LAB for fermentation of Swedish legumes
Fermentation of legumes improve digestibility, natural preservation, and raise their nutritional value. Certain microbial strains may also be probiotic. Today, despite the growing demand for dairy-free products, only very few fermented non-dairy products exist in the market worldwide, and few strains available commercially are aimed for the fermentation of plantbased foods (36).
In the present study, 116 isolates were obtained. Subsequently, they were sorted into 25 unique strains from three different genera as confirmed by rep-PCR and MALDI-TOF MS. The strains were identified as LAB, and closely related B. coagulans. Since B. coagulans is a spore forming bacteria, it can survive extreme conditions and has previously been suggested as an health-promoting ingredient (37). Owing to the spore forming ability, B. coagulans can survive human ingestion and reach the intestine where it germinates. In the intestine, it can contribute to positive effects such as production of short-chain fatty acids, promoting populations of beneficial bacteria (38), and aid in the digestion of protein and carbohydrates (39). The ability to increase protein bioavailability is of relevance in the ongoing protein shift toward more plant-based protein. Probiotic food products containing B. coagulans are currently sold to consumers commercially (37), and B. coagulans has received the Generally    Recognized as Safe (GRAS) status and is reported as safe by the US Food and Drug Administration (FDA) (40). The 25 unique strains were characterized in vitro in order to select candidate strains suitable for further investigation related to probiotic properties. Strains belonging to E. faecium, E. mundtii, and P. pentosaceus showed a high level of tolerance to 0.3% bile. Tolerance to bile and low pH is important to evaluate, since it indicates survival during the passage through the upper GIT, a necessary step for probiotic organisms delivered via food (41). Among the strains isolated in this study, strains 77, 100, 101, and 176 belonging to B. coagulans and P. pentosaceus have shown tolerance to pH 2.5.
Tolerance to phenol was evaluated, since phenols may be formed in the gut by bacterial deamination of aromatic amino acids derived from dietary or endogenously produced proteins. Phenols and other products of protein degradation in the colon, such as ammonia, indoles, and amines, exert toxic effects in animal models and in vitro. These compounds are present in fecal samples and can exert bacteriostatic effects against some LAB (42).
We examined antibiotic resistance for Enterococcus strains because of the high occurrence of resistance within the Enterococcus genus, and because of the risk of pathogenicity. In the future studies, the presence of virulence genes should be addressed. Four of the Enterococcus strains (E. faecium 63 and E. mundtii 94, 146, and 189) were susceptible to all four tested antibiotics. Owing to their tolerance to salts and acids, Enterococcus strains are highly adapted to surviving in the GIT. Some Enterococcus strains have interesting properties relevant from food technology and medical perspective, such as the ability to produce multiple bacteriocins (43). Even though a few Enterococcus strains are currently commercialized as probiotics and used for treatment and prevention of certain diseases in humans and animals, these strains might originate from bovine feces used as manure in the fields, and may not be suitable to metabolize legumes (43,44).
The ability to produce BSH enzyme, and thus detoxify bile salts, has often been included as a criterion for selection of probiotic strain (45). BSH catalyzes the deconjugation of bile salts to liberate free bile acids (46). Strains neither belonging to P. pentosaceus nor to B. coagulans were positive for BSH activity. On the contrary, strains belonging to E. mundtii and E. faecium were able to hydrolyze the bile salt TDCA along with E. hirae strain 78. The BSH activity indicates that the isolates not only survive toxicity of these salts but also de-conjugate these salts and may help in the intestinal colonization of cells and in reducing blood cholesterol level in the host (47).
We tested five selected strains belonging to P. pentosaceus for their ability to grow in a fava bean medium with no additional carbon, thus indicating their potential to ferment a bean beverage. Since strains belonging to B. coagulans are mainly of value as spores and not as fermenting organisms, and because of the uncertainties regarding Enterococcus during fermentation, these strains were excluded from fermentation. The five P. pentosaceus strains that were used for fermentation in this study had a similar growth pattern: an initial lag phase followed by exponential growth until 8 h of incubation, when the cells reached a stationary phase.
Finally, we showed the ability of one of the isolated P. pentosaceus strains to improve the reduction of phytate by the yeast strain P. kudriavzevii TY1322 in a fava bean medium. The origin of the parent strain TY13 and its properties regarding phytase production are described in the previous work by Hellström et al (48,49,50). The symbiotic effect of P. pentosaceus strain 77 resulted in a 69% larger reduction compared to the monoculture and a 124% higher cell count of P. kudriavzevii TY1322 after 48 h of fermentation. This significantly improves the bioavailability of minerals such as iron and zinc, since phytate is a major inhibitor of iron and zinc absorption. This is especially important for individuals following a plantbased diet that does not contain dietary sources of heme iron.
P. pentosaceus strains have been previously isolated from legumes such as fava beans (51) and other unspecified beans (52). These findings are in line with our results. Strains of Pediococcus are currently used in the food industry as starter cultures for sausages, sauerkraut, cucumber, green bean, and soy sauce fermentation (53). Similarly, specific strains of Pediococcus are also used as probiotics for monogastric animals (54). The Pediococcus strains isolated in this study were able to use nutrients from fava beans, supporting their potential role as starters for further investigation and their possible use in the nondairy sector of food industry.

Conclusion
Some of the strains isolated can be of interest from the perspective of food technology as potential starters for fermentation of non-dairy legume-based products, and others may have potential to be probiotic candidate strains. Further investigation on probiotic properties are needed, such as adherence to gastric mucosa. Five strains showed tolerance to all three conditions tested (survival through GIT, tolerance to bile and phenol, and low pH). Three of these strains belong to B. coagulans (strains 51, 221, and 225), and the other two belong to P. pentosaceus (strains 77 and 176), which are of interest regarding fermentation, presenting an interesting possibility to produce a combined fermented and probiotic dairy-free food product. The P. pentosaceus strains that were isolated in the present study were suitable for the fermentation of a legume beverage. One of the P. pentosaceus strains was also able to improve the reduction of phytate by the phytase producing yeast strain P. kudriavzevii TY1322 during co-fermentation as well as increase the final cell count of yeast strain. This is a positive indication that new and improved fermented non-dairy products could be produced for commercial use and that strains with probiotic potential can be selected.