In vitro assessment of safety and functional probiotic properties of Lactobacillus mucosae strains isolated from Iranian native ruminants intestine

Abstract The search for host isolated probiotic bacteria from animal intestine may discover new probiotic candidates with promising health properties. This study evaluated the safety and functional probiotic potential of the lactic acid bacteria (LAB) isolated from the Iranian native ruminants intestine under in vitro assays. The isolates were selected according to criteria including survivability in low pH, bile salts, pancreatic enzymes, different temperatures, NaCl concentrations, antibacterial activities, presence of adhesion genes and safety characteristics. The selected LAB were then identified to species level using 16S rRNA gene sequencing. The results showed that out of one hundred and eighty-seven LAB isolates, only six strains (NABRII50, NABRII51, NABRII52, NABRII53, NABRII54 and NABRII55) were tolerant to low pH, bile salt, pancreatin enzyme, 45 °C temperature and 2% sodium chloride. The six selected isolates belonged to Lactobacillus mucosae. Two of the adhesion genes (mub and map) were detected in all strains except NABRII53. The virulence factors were observed in NABRII50, NABRII53 and NABRII55. The tetracycline resistance gene (tet (S)) was detected in NABRII55. This study was the first effort to select Lb. mucosae strains with the probiotic potential from the Iranian ruminants intestine. These results revealed that the ruminant intestinal ecosystem could be considered as a valuable origin of probiotic candidates and all the selected LAB strains except NABRII50, NABRII53 and NABRII55 could be considered as promising probiotics. HIGHLIGHTS Lactobacillus mucosae strains isolated from Iranian native ruminants intestine including NABRII51, NABRII52 and NABRII54 showed probiotic potential under in vitro assays. The Lb. mucosae strains including NABRII51, NABRII52 and NABRII54 were able to survive in intestinal physiological conditions, and carried the adhesion genes, such as mub and map. In vitro biosafety assays confirmed that Lb. mucosae strains (NABRII51, NABRII52 and NABRII54) were safe to further studies.


Introduction
Currently, there is a great interest in probiotics application for health and improve livestock performance (Li et al. 2020). The application of probiotic products can improve immune function, digestion and feed efficiency. Lactic acid bacteria (LAB) have been used as probiotics to modulate the composition of intestinal microbiota to promote the health of the host intestine (de Moraes et al. 2017). LAB isolates have been isolated from a variety of sources, including fermented and dairy products, plants, soil and various organs of poultry, cattle or fish (Kuppusamy et al. 2020). In addition, large amounts of LAB with vital function have been found in the ruminant's digestive tract (Timmerman et al. 2006).
Based on the selection criteria for probiotic strains, a bacterial strain should be able to withstand low gastric pH and bile salts of the intestine, adhere to the intestinal mucosa and exceed safety criteria, such as the absence of haemolytic activity and sensitivity to antimicrobials of human and veterinary importance (Musikasang et al. 2009;Iñiguez-Palomares et al. 2011). Besides, resistance to osmolytes and temperature is considered as notable features that lead to the successful performance of probiotic bacteria (Salas-Jara et al. 2016;Aleksandrzak-Piekarczyk et al. 2019) and those candidates which meet the established criteria for probiotic can be used to produce probiotic supplements.
The genus Lactobacillus is one of the beneficial LAB. Most of the studies have shown that Lactobacillus sp. are able to survive in harsh conditions of the gastrointestinal tract (GIT). The organic acids produced by LAB form an acidic environment that can inhibit the viability of pathogenic bacteria (Dunne et al. 2001;Bernardeau et al. 2008). Among LAB species, Lactobacillus mucosae is one of the highly mucosaassociated subpopulations closely related to the animal and human intestine and other mucosal niches (Etzold et al. 2014;Drobn a et al. 2017). The results of various studies have shown that Lb. mucosae strains can promote host resistance against pathogens and improves mucosal immunity by increasing epithelial impermeability and barrier function, producing secondary metabolites and antimicrobial compounds (Pajarillo et al. 2017).
The indigenous livestock population of Iran is the unique reserves adapted to the various climatic conditions of this vast country. Adaptation to different climates has led to developing a specific microbial community in the GIT of ruminants (Ebrahimi et al. 2018;Naeemi et al. 2019). Given these points, this study aimed to investigate the functional probiotic properties and safety characteristics of autochthonous Lb. mucosae strains from the duodenal mucosal layer of native goat, sheep and cattle from west and north areas of Iran, using molecular and microbial approaches.

Initial isolation, purification and phenotypic characterisation
The LAB colonies were isolated from duodenum mucosal layer suspensions of cows (n ¼ 3), sheeps (n ¼ 3) and goats (n ¼ 3) belonging to the North Providence, Guilan, and the west providences, Kermanshah and Hamedan of Iran. The samples were received from a livestock slaughterhouse located in the north of Iran (Deylam Sanat Shargh Livestock Industrial Slaughterhouse CO-OP CO (37 08 0 37.4 00 N 49 38 0 38.1 00 E)). The experimental protocols were carried out according to the principles of the Declaration of Helsinki (World Medical Association [WMA] 2008). The mucosal layer suspensions were spread on de Man, Rogosa and Sharpe (MRS) agar medium (Merck, Darmstadt, Germany), which were supplemented with 0.1% (w/v) L-cysteine (Merck, Darmstadt, Germany). After incubation at 37 C for 48 h, the selected colonies were identified by Gram stain and catalase activity tests. The Gram-positive and catalase-negative isolates were stored in the MRS broth containing 10% (w/v) skimmed milk, and 60% (1:1) glycerol at À80 C for further investigations.

Resistance to different acidic conditions, bile salts and pancreatin
Rapid preliminary screening for the acid tolerance of the LAB isolates was performed according to the method described by Ehrmann et al. (2002) with some modifications. Acid tolerance assay of each selected isolate from previous step was performed according to the method described by Yamazaki et al. (2012) and Grispoldi et al. (2020). Tolerance to the acidic conditions was determined by comparing number of bacterial colonies before (in neutralised PBS, as control) and after exposure to acidic conditions including pH ¼ 2.5 and pH ¼ 3 adjusted by 1 N HCl for 3 h.
The bacterial resistance to bile salts was performed according to the method described by Kumar and Kumar (2015) through measuring suspensions optical density at 630 nm (OD 630 ) before and after incubation in MRS broth containing 0.3% (w/v) Oxgall for 8 h. The suspensions without Oxgall were considered as control. At last, the LAB isolates which showed resistance to 0.3% (w/v) Oxgall more than 50% were considered as bile-resistant.
For evaluation of the bacterial cell resistance to pancreatin, 10 lL of each selected isolates suspension was inoculated into 170 lL MRS broth supplemented with 1.9 mg/mL of pancreatin (Sigma-Aldrich, St. Louis, MO), adjusted to pH ¼ 8.0. Also, MRS broth (pH ¼ 8.0) without pancreatin was considered as control. After incubating at 37 C for 3 h, OD 630 was recorded and the result was expressed based on growth rate (Hosseini et al. 2009;Sharifuzzaman et al. 2018).

Molecular identification
DNA extraction was performed with the Gram-positive bacteria DNA extraction kit (Cinaclone, Tehran, Iran) . The bacterial 16S ribosomal RNA-based polymerase chain reaction was carried out using universal primers 27f (5 0 AGAGTTTGATCMTGGCTCAG-3 0 ) and 1492r (5 0 TACGGYTACCTTGTTACGACTT-3 0 ) (Plessas et al. 2017). After purification, PCR products were sequenced by the automated DNA sequencing system (Macrogen, Seoul, Korea). The sequences were edited by Bioedit software version 7 (Hall 1999). A comparison of the sequence similarity was made by the Basic Local Alignment Search Tool (BLAST) of the National Centre of Biotechnology Information (NCBI). CLUSTALW program of the Bioedit software version 7 was used to sequences alignment. A phylogenetic tree was built using the neighbour-joining tree method based on the 16S rRNA gene sequence analysis by MEGA software version 6 (Tamura et al. 2013). In the phylogenetic tree construction, nucleotide sequences, including six isolated LAB sequences derived from this study and five sequences of the Lactobacillus species derived from Genbank, were involved. The Lactococcus lactis strain NCDO 604 was used as an out group.
Resistance assay of Lb. mucosae strains under temperature stress and sodium chloride concentrations The effects of different temperatures and sodium chloride concentrations on bacterial survivability were examined according to the procedure described by Mortezaei et al. (2020) and Aleksandrzak-Piekarczyk et al. (2019), respectively. During temperature treatments, each selected isolate was exposed to different temperatures (37, 45 and 50 C) for 48 h and formation of bacterial colonies was evaluated. For salinity resistance test, each selected isolate was inoculated on MRS-agar plates supplemented with or without (control) NaCl (2, 4, 6 or 8% (w/v)). After incubation at the 37 C for 48 h, presence and quality developed colonies were compared with control plates. The experiments were carried out in triplicate and the results were expressed by qualitative data.

Antimicrobial activity of Lb. mucosae strains
The inhibitory activity of the selected strains against four pathogenic bacteria (Salmonella typhimurium (ATCC 14028), Salmonella enteritidis (ATCC13076), Escherichia coli (o157) and Staphylococcus aureus (ATCC 25923)) was determined by the double agar layer method (Tour e et al. 2003;Gaudana et al. 2010). Briefly, 2 lL of each selected Lactobacillus strain grown overnight culture was spotted onto MRS agar plates. Then plates incubated at 37 C for 18 h in CO 2 incubator (5%). After colony development, the plates were overlaid with soft agar (containing 0.7% (w/v) agar and Trypticase Soy Agar (TSA) kept at 50 C), seeded with 1% (v/v) of an active overnight culture of each pathogen, and incubated aerobically at 37 C. After one night of incubation, the growth inhibition zones around LAB colonies were determined. The test was performed in triplicate.

Assessment of adhesion properties of Lb. mucosae strains
The adhesion properties of the selected Lb. mucosae strains were preliminarily determined by Congo red staining. Then, isolates were screened by PCR to investigate the presence of genes encoding adhesion proteins (msa, map, mub and ef-tu).

Congo red staining
The hydrophobicity of the Lb. mucosae strains was determined by Congo red staining (Leyva-Madrigal et al. 2011). The bacterial colonies were streaked on MRS agar plates containing 0.03% (w/v) Congo red (Merck, Darmstadt, Germany) and incubated at 37 C for 24 h anaerobically. Subsequently, red colonies were considered as hydrophobic strains, and white or transparent colonies were considered as non-hydrophobic.

PCR detection of adhesion encoding genes
The selected Lb. mucosae strains were screened for adhesion encoding genes according to the PCR protocol described by de Moraes et al. (2017) after bacterial DNA extraction was done by the Gram-positive bacteria DNA extraction kit (Cinaclone, Iran). The primers were employed for the amplification of msa, map, mub and ef-tu presented in Table 1. The amplified products were then separated by electrophoresis in 1.0% (w/v) agarose gels. Safety assessments of Lb. mucosae strains

Haemolytic activity
The haemolytic activity of selected isolates was performed on blood agar (Quelab, Montr eal, Canada), supplemented with 5% (v/v) of sheep blood. After 48 h incubation at 37 C in CO 2 incubator (0.5%), the plates were then examined for the halo of haemolysis. The bacterial isolates without displaying the signs of b-haemolysis around the colonies were classified as non-haemolytic (without b-haemolysis) (Maragkoudakis et al. 2009).

Antibiotic susceptibility
The assay for antibiotic susceptibility of the six Lb. mucosae strains were performed in 96-well plates using the broth-microdilution method for the eight antibiotics of human and veterinary importance (ampicillin, clindamycin, gentamicin, streptomycin, tetracycline, erythromycin, kanamycin and chloramphenicol) according to the European Food Safety Authority (EFSA 2012). Antibiotic susceptibility was expressed as the minimum inhibitory concentration (MIC, lg/mL) necessary for inhibition bacteria visible growth and compared with the lactobacilli MIC breakpoint values recommended by EFSA (2012).
PCR detection of genes for tetracycline resistance, virulence factors and biogenic amines production The presence of tetracycline resistance genes, virulence factors genes, such as gelE (gelatinase), hyl (hyaluronidase), asa1 (aggregation substance), esp (enterococcal surface protein), cylA (cytolysin), efaA (endocarditis antigen), ace (adhesion of collagen) and genes encoding biogenic amines production, such as histidine decarboxylase (hdc1), histidine decarboxylase (hdc2), tyrosine decarboxylase (tdc) and ornithine decarboxylase (odc) were evaluated according to the PCR protocols (Muñoz et al. 2014;Perin et al. 2014;de Moraes et al. 2017). The primers employed for amplifying the encoding genes are presented in Table 2. Briefly, extracted DNA using the Gram-positive bacteria DNA extraction kit was used for the PCR amplification according to the PCR protocol described by Muñoz et al. (2014). The amplified products were then separated by electrophoresis in 0.8 to 2.0% (w/v) agarose gels in 0.5Â TAE buffer.

Statistical analysis
The data were expressed as the mean values (or log values) ± standard error (SE). All in vitro assays were performed in triplicate. p Values of less than .05 were considered statistically significant. Significant differences

Bacterial initial isolation and phenotypic characterisation
The bacterial isolation and phenotypic characterisation resulted in isolation of one hundred and eighty-seven Gram-positive, non-spore-forming, catalase-negative and non-motile bacteria from the duodenum mucosal layer of native Iranian cows, sheep and goats. The LAB colonies were rod-shaped or coccoid form and the cells were shown in pairs or short chains.

Survival assays under acidic conditions, bile salts and pancreatin resistance
Based on the previous studies, in vitro assessments, such as acid, bile and pancreatic enzymes tolerance, and survey some physiological characteristics (such as optimum growth temperature and salt sensitivity) have been considered as a good indicator to evaluate the probiotic properties of a bacterial isolates (Kim et al. 2019;Kuppusamy et al. 2020). After preliminary growth assays under acidic condition, seventy LAB isolates exhibited good survivability to pH ¼ 3 (data not shown).
In the small intestine, the presence of bile is the most crucial stress agent for ingested bacteria. According to the results, the six selected lactobacilli isolates showed an ability to grow over 50% at 0.3% Oxgall (Figure 1). Among them, NABRII52 and NABRII55 showed significantly higher (97.97%) and lower (58.88%) ability than other isolates in this regard, respectively (p < .05). 0.3% bile salt concentration is a critical concentration for evaluating the ability of LAB to tolerance bile salts, and those with resistance more than 50% at this range considered as bile resistant isolates (Sahadeva et al. 2011;Kumar and Kumar 2015).
In the ruminants, pH values can vary between 5.7 and 7.3 in the rumen and 2-3 in the abomasum (Gentile et al. 2004). The secretion of hydrochloric acid by the gastric cells is an important defence mechanism to protect the host body against the ingested pathogens (Smith 2003). The results of acid tolerance for all six Lactobacillus strains are shown in Table 3. All of the selected isolates showed a significant reduction in pH ¼ 2.5 than control (0 h) (p < .05), But their survival rate remained more than 90% after exposure to pH ¼ 2.5 for 3 h. In addition, all isolates' survivability in pH ¼ 3 was similar to the control (0 h  at pH ¼ 2.5 for 3 h. Generally, the selected LAB showed different levels of resistance to acid and bile salt which probably could be due to the straindependent mechanisms (Li et al. 2020).
Furthermore, all six selected Lactobacillus isolates exhibited strong tolerance to pancreatic enzymes ( Figure 2). Among them, NABRII52 showed significantly higher tolerance to 1.9 mg/mL of pancreatin (60.78%) after 3 h exposure in comparison to the other selected isolates (p < .05) except NABRII50. The results of tolerance to pancreatin are in agreement with the results obtained by Maragkoudakis et al. (2009) and Mahmoudi et al. (2016).

Molecular identification using 16S rRNA gene sequence
The results of comparative 16S rRNA gene analysis showed that the six selected LAB belonged to the genus Lactobacillus and were 99-100% similar to Lb. mucosae. The 16S rRNA gene sequences of the six Lb. mucosae strains were deposited in the GenBank database under the accession numbers MH595979.1 to MH595980.1 for isolates NABRII50 to NABRII55, respectively (Table 4). In this study, a phylogenetic tree (Figure 3) depicts the phylogenetic relationships between the six Lb. mucosae strains and five type Table 3. Viability of six selected lactobacilli (log cfu/mL) after 3, h exposure to pH 2.5 and 3 after 3, h compared with control (0, h  strains obtained from the Genbank based on 16S rRNA gene sequence analysis. Lc. lactis (NCDO 604) was used as the outgroup. The phylogenetic tree depicted that the six Lb. mucosae strains grouped into one leading group.
The Lb. mucosae was first isolated from the intestine of piglets as a new species (Iñiguez-Palomares et al. 2007). To our knowledge, the present research is the first report on the isolation of Lb. mucosae strains from the duodenum of Iranian sheep, goat and cow. Taxonomic studies have shown a close relation between Lb. mucosae and Lb. reuteri (Wang et al. 2016).

Resistance to temperatures and sodium chloride
Temperature also plays a vital role in LAB growth (Yang et al. 2018). The optimum growth temperature of lactobacilli lies between 30 and 40 C, but depending on their species; they can grow at different temperatures from 5 to 53 C (Ahmed et al. 2006). This study also revealed that a temperature of 50 C has a negative effect on the survivability of the isolated Lb. mucosae strains, and their respective optimum temperatures were 37 and 45 C (Table 5). Also, osmotic stress may be a significant inhibitor of bacterial growth and causes structural and functional damage to strains (Ge et al. 2011;Zhang et al. 2014). Furthermore, the Lb. mucosae strains in this study were salt sensitive and could not grow at more than 2% salt concentration (Table 5). Silva et al. (2019) found that Lb. reuteri strains had different growth rates towards 6% NaCl, while Lb. mucosae CRL 1508 was not resistant to the same NaCl concentration.

Antimicrobial activity of the selected Lb. mucosae strains
The Antimicrobial activity of the selected Lb. mucosae strains were evaluated using various Gram-positive (S. aureus) and Gram-negative (S. typhimurium, S. enteritidis, and E. coli) pathogenic bacteria ( Table 6). Results of this study showed that NABRII52 and NABRII53 significantly inhibited the growth of the S. typhimurium (ATCC 14028) and S. enteritidis, respectively (p < .05). Our results are consistent with the results of the study  by Bian et al. (2011), which reported that Lb. reuteri DPC16 cell-free supernatants significantly inhibited the growth of selected Gram-negative food-borne pathogens (S. Typhimurium and E. coli) compared to Gram-positive pathogens (Listeria monocytogenes and S. aureus). Some previous studies have suggested that the production of bacteriocin-like metabolites by Lb. mucosae strain may be the reason (Maldonado et al. 2018).

Adhesion properties of the six selected Lb. mucosae strains
Adhesive ability and effective colonisation in GIT is a desirable feature for probiotic bacteria because it can inhibit pathogens' growth in the lumen through competitive exclusion (London et al. 2014). Various mechanisms, such as the presence of some adhesins, fimbriae, pili or cell surface proteins may be related to these phenomena (Devi and Halami 2017). The hydrophobic nature of the outer surface may also play a role in the binding of bacteria to the host tissue. The evaluation hydrophobicity test using Congo red stain can confirm this nature (Leyva-Madrigal et al. 2011). A positive result indicates that the bacteria would not repel from the intestinal epithelium and have the ability to bind non-specifically to the intestinal epithelium by hydrophobic interactions (Leyva-Madrigal et al. 2011). In this study, all the Lb. mucosae strains had hydrophobic structures in the cell wall.
Previous studies have shown that mucus-targeting proteins or mucus-binding proteins (mub) and mucus adhesion-promoting protein (map), which were well characterised among Lactobacillus species, mediate the adherence of them to the intestinal mucosal layer (Buck et al. 2005;Devi and Halami 2017;Chatterjee et al. 2018). The presence of these genes strengthens the probiotic potential and mucus-binding ability of Lb. mucosae (Roos et al. 2000). Previously, it has been reported that this species usually carries the mub gene and can attach to the intestinal mucosal layer with this feature (Roos et al. 2000). The mub and map genes encode for extracellular mucus-binding proteins (mub) and a mucous adhesion-promoting protein (map), respectively (Buck et al. 2005). The presence of mub and map genes in the studied Lb. mucosae strains strengthens the probiotic potential of these strains. In this study, all the isolated Lb. mucosae strains except Lb. mucosae NABRII53, carried both genes (Table 7).

Haemolytic activity
The absence of haemolytic activity is one of the safety issues to select new potential probiotic strains because it indicates that the isolated bacteria do not have one of the virulence factors and potential possible negative effect on humans and animals (De Vuyst et al. 2003). The isolated strains displaying haemolytic activity (b or a-haemolysis) can produce toxins that induce þþ þþ b þ þ þ þ þ 50 c À À À À À À Salinity tolerance (w/v) 2, % þþ þþ þþ þþ þþ þþ 4, % À À À À À À 6, % À À À À À À 8, % À À À À À À a (þþ) Strong resistance. b (þ) Moderate resistance. c (À) Susceptible. erythrocyte lysis. The results of the haemolytic activity in this study showed no harmful effects under in vitro assays. Our results confirm the findings of Repally et al. (2018) and Adetoye et al. (2018), which reported that other Lb. mucosae strains isolated from sheep milk and cattle faeces had no haemolytic activity.

Antibiotics susceptibility
Acquired antibiotic resistance remains a serious concern due to the high risk of horizontal spread of resistance genes when a viable microorganism is used as an active agent in probiotic supplements (EFSA 2012;Frieri et al. 2017). According to the EFSA (2012), when a bacterial strain is inhibited by a concentration of a specific antimicrobial agent equal to or lower than the specified cut-off values, considered as a susceptible strain and no further antibiotic resistance studies are required (EFSA 2012). In this study, we investigated the antibiotic susceptibility profiles of the six Lb. mucosae strains from the Iranian native cows, sheep and goats. The results of the antibiotic susceptibility profiles showed that all Lb. mucosae strains were sensitive to clinically relevant antibiotics such as ampicillin, gentamicin, kanamycin, streptomycin, erythromycin, clindamycin and chloramphenicol. However, only Lb. mucosae NABRII53 and Lb. mucosae NABRII55 were phenotypically resistant to tetracycline (MIC > 8 lg/mL) (Table 8).

PCR detection of genes encoding tetracycline resistance
The tetracycline-resistant strains were screened by PCR for the presence of tetracycline resistance genes (tet (L), tet (M), tet (O), tet (S) and tet (w)) for identifying the resistance determinants responsible for the tetracycline resistance (Table 9). PCR analysis showed that Lb. mucosae NABRII55 carried the tet (S) and neither the genes encoding ribosomal protection proteins tet (M), tet (O), tet (S) or tet (W) nor gene encoding the tetracycline efflux pump tet (L) were detected in Lb. mucosae NABRII53. However, the absence of resistance determinants tet (M), tet (O), tet (S), tet (W), tet (L) and tet (K) may suggest a new mechanism of resistance which can be due either to acquired genes or to the mutation of indigenous genes (EFSA 2012). Energydependent efflux of tetracycline from the cell, ribosome protection and enzymatic inactivation are three mechanisms of tetracycline resistance in microorganisms (Schaechter and Lederberg 2004). Studies have shown that genes conferring resistance to tetracycline are commonly found in the human gut microbiota, both in healthy adults and in breast-fed infants (Gueimonde et al. 2006). Similarly, these genes have been found in several lactobacilli isolated from dairy foods and it is vital to avoid their spread to pathogens through the consumption of fermented foods (Muñoz et al. 2014). Similar to this study, the absence of the resistance genes tet (M), tet (O), tet (S), tet (W), tet (L) and tet (K) was reported in Lc. Pseudomesenteroides from fermented table olive (Muñoz et al. 2014).
Overall, lactobacillus resistance to antimicrobials is a Table 7. Presence of genes associated with adhesion properties in six selected lactobacilli.

Adhesion genes
Presence of genes in the selected isolates -a (-) Lack of genes encoding for tetracycline resistance. b (þ) Presence of genes encoding for tetracycline resistance.
relevant scientific topic and one of the crucial properties for identifying safety potential probiotics.

PCR detection of genes encoding virulence factors
Investigating the presence of virulence genes is another criterion for assessing the probiotic candidates' safety with potential applications in food products. The virulence factors are usually associated with competitive advantages of pathogenic strains, and their presence is more common in Enterococcus spp. and other clinical isolates (Semedo et al. 2003). The virulence factors usually locate in transferable plasmids. Therefore, due to the concern of transferring the mentioned genetic elements to the intestinal tract's pathogens, the detection of these genetic elements is inevitable (Eaton and Gasson 2001).
LAB species, such as Lactococcus spp. can carry different virulence genes, but their presence in the genome is not a definitive indicator of these species' pathogenesis because of the low capability of expressing these genes, which has been observed in different studies (Casalta and Montel 2008;Perin et al. 2014).
In this study, the Lb. mucosae NABRII53 and Lb. mucosae NABRII55 strains generated positive PCR results for the ace gene (adhesion of collagen protein). This protein facilitates the binding to collagen and may play a negative role during human infections (Girish and Kemparaju 2007); however, on the positive side, it can contribute to better adhesion and colonisation in the GIT . dos Santos et al. (2015) observed positive results for the ace gene in the studied Lb. rhamnosus and Lb. plantarum strains. Furthermore, two of the investigated lactobacilli strains in this study, including Lb. mucosae NABRII53 and Lb. mucosae NABRII50, similar to the study conducted by de Moraes et al. (2017), carried the cytolysin gene. The cytolysin (cylA) is related to haemolytic activity among enterococci (Jim enez et al. 2013). Despite the presence of cytolysin gene, these strains (Lb. mucosae NABRII53 and Lb. mucosae NABRII50) were not able to exhibit haemolytic activity under in vitro assays. The cytolysin is a virulent substance due to its haemolytic potential, but it is also considered as an antibacterial bacteriocin, according to Cotter et al. (2005) classification. The presence of cylA is not enough to activate haemolytic activity by bacteria because the cytolysin expression requires the presence and functionality of eight genes (Perin et al. 2014). Additionally, Lb. mucosae NABRII53 showed positive results for the hyl gene. The hyl gene is involved in the production of hyaluronidase enzymes, which break down hyaluronic acid. The ability of bacteria to degrade hyaluronic acid may be a virulence factor and allow hyaluronidase-producing pathogens to penetrate hyaluronic acid-rich tissues (Aubin et al. 2017). However, the consequences of hyaluronidase activity among lactobacilli are not clear, as this has not been reported within the context of virulence and pathogenicity yet (Franz et al. 2005). The gelE gene, which encodes for the production of Gelatinase, was detected in Lb. mucosae NABRII53. The gelE is commonly found in E. faecalis (Munoz-Atienza et al. 2013). However, the presence of gelE gene is not enough for gelatinase activity since the complete fsr operon seems to be essential for its expression (Lopes et al. 2006). Furthermore, the esp and asa1 genes were detected in Lb. mucosae NABRII53 strain. These genes are relevant virulence factor, which contributes to intestinal adhesion through encodes the production of extracellular surface protein (Valenzuela et al. 2009;de Moraes et al. 2017). According to the results of this study, due to the presence of virulence genes in Lb. mucosae NABRII53, Lb. mucosae NABRII50 and Lb. mucosae NABRII55 these isolates are not recommended for food applications.

PCR detection of genes encoding biogenic amines
Examining the presence of genes encoding biogenic amines is another aspect of assessing the probiotic candidates' safety because their products could cause  health problems. The hdc1, hdc2, tdc and odc genes that are involved in biogenic amine production were searched in the genome of the Lb. mucosae strains ( Table 11). None of the genes associated with the production of histidine decarboxylase, tyrosine decarboxylase and ornithine decarboxylase was detected in the selected strains. In this study, only hdc1gene was detected in NABRII55.
Biogenic amines production is an intrinsic property (Franz et al. 2005;Lorenzo et al. 2010). Lactobacillus strains are usually considered safe organisms in this respect (Arena et al. 2002). de Moraes et al. (2017) found none of the genes associated with the production of histamine and cadaverine in the genome of the Lb. mucosae strains. Mart ın et al. (2005) reported that Lb. gasseri and Lb. fermentum cannot produce biogenic amines. However, the results of the biogenic amines production gene showed that Lb. mucosae NABRII55 isolated from cow carries the hdc1 gene. The formation of biogenic amine from histidine is responsible for allergic reactions. Histamine is formed by the histidine decarboxylation activity of Gram-negative enteric bacteria. Furthermore, it is mainly produced by Gram-positive LAB in some fermented products. dos Santos et al. (2015) detected the hdc1 gene in the genome of the All Lb. rhamnosus strains isolated from Artisanal Coalho cheeses except for Lb. rhamnosus EM1107. Therefore, all examined isolates in this study except for Lb. mucosae NABRII55 are safe in this regard and could be used as starter cultures or other food supplements.

Conclusions
In conclusion, the results of this study indicate that all the Lb. mucosae strains identified from native ruminant intestine of Iran except the Lb. mucosae NABRII50, Lb. mucosae NABRII53 and Lb. mucosae NABRII55 strains were considered as safe for antibiotic resistance, carried virulence factor, and biogenic amine genes. Moreover, they have probiotic properties including ability to survive under simulated gastric conditions, inhibiting bacterial pathogens, harbouring genes related to intestinal adhesion properties under in vitro assessment. Thus, the three Lactobacillus strains could be considered as candidate probiotic strains and should be further studied for their health benefits.

Ethical approval
The protocol followed in this study was approved by the the Ethics Committee recommendations. Table 11. Presence of genes associated with biogenic amines production in Lb. mucosae strains.

Biogenic amine producing genes
Presence of genes in the selected isolates -----a (-) Lack of genes encoding for biogenic amines. b (þ) Presence of genes encoding for biogenic amines.