Screening and Characterization of Lactic Acid Bacteria with Broad-Spectrum Antimicrobial Activity from Tibetan Qula, a Yak Milk Cheese

Lactic acid bacteria with natural, effective antibacterial activity, safe and reliable characteristic, gradually become one of the key technologies in food fermentation applications, food preservation and other elds. In this study, 112 presumptive lactic acid bacteria isolated from Tibetan Qula, a fermented yak cheese popular in the Tibetan plateau, were screened for potential probiotic microorganism with antimicrobial activity. glycol (PEG) 6000 was added to MRS medium. Additionally, Tween 80 at different concentrations (0, 1, 2, 3, 4, 5, and 6%) was used to determine the optimum concentration of Tween 80. In each experiment, antibacterial activity was conducted as previously described. The resistance of bacteriocin to heat, pH, and enzymes was examined. Cell-free culture supernatants were heated to 60, 80, or 100°C in a water bath and to 120°C in an autoclave for 15 and 30 min. To assess the sensitivity to different pH values, cell-free culture supernatants were adjusted to pH 2, 3, 4, 5, 6, 7, 8, 9, or 10 and incubated at 37°C for 2 h. Cell-free culture supernatants were treated with α-amylase (1 mg/ml), β-amylase (1 mg/ml), lipase (1 mg/ml), pepsin (3 mg/ml), trypsin (3 mg/ml), protease K (3 mg/ml), papain (3 mg/ml), or chymotrypsin (3 mg/ml) at 37°C for 2 h. Prior to the 2-h incubation, the pH of the cell-free supernatants was adjusted to the optimum pH of each enzyme. After all treatments, antibacterial activity was determined as previously described. analysis P

However, the yield and antimicrobial spectrum of bacteriocins is low; certain bacteriocins cannot inhibit Gramnegative bacteria, yeast, or mold. Moreover, the yield and antibacterial spectrum of different bacteriocins are dependent on the bacterial strain and habitat [6]. Therefore, it is important to identify LAB that synthesize high levels of broad-spectrum bacteriocins.
In this study, we screened and identi ed LAB strains producing a novel broad-spectrum bacteriocin. For maximum production and improved applications of bacteriocins, we optimized the bacteriocin production conditions, and the physicochemical characteristics of the resulting bacteriocins were evaluated. Further, this study was the rst to date to identify the bacteriocin-producing L. plantarum strains found in Qula cheese.
Moreover, it could not completely utilize D-arabitol and gluconate. For the 16S rRNA gene identi cation of QZ50, a phylogenetic tree was constructed using a neighbor-joining method ( Fig. 1). In the Lactobacillus cluster, strain QZ50 was grouped with L. pentosus, L. plantarum, and L. paraplantarum, but it could not be identi ed to a species level based on 16S rRNA gene sequence analysis. The ampli cation products gained from the recA gene was displayed in Fig. 2. Strain QZ50 and type strain L.
plantarum subsp. plantarum JCM 1149 T all produced 318 bp products. Thus, it was identi ed as L. plantarum subsp. plantarum. All the above strains with good antibacterial activity were identi ed, QZ11, QZ21, QZ71 and QZ81 were identi ed as Lactobacillus plantarum, QZ58 was identi ed as Lactococcus lactis, and QZ14, QZ71 were identi ed as Lactobacillus johnsonii.
Determination and Comparison of the performance of selected lactic acid bacteria.
The antibiotic susceptibility of the 12 selected strains with antibacterial activity was tested and the results were shown in Table 3. All the selected LAB strains were resistant to CN, CIP, CT, VA and P, while most strains exhibited sensitivity to AMP, TE and C. For erythromycin (E), the QZ11, QZ14, QZ21, QZ42,QZ57, QZ58 and QZ81 had sensitivity, while QZ29, QZ41, QZ50 and QZ67 were moderately resistant, and QZ71 was resistant to it. As to rifampicin (RD), QZ29, QZ41 QZ57, QZ58, QZ81 were resistant to it. Strain QZ57 was found to be resistant to most antibiotics except erythromycin and chloramphenicol.  The surface hydrophobicity and agglutination of the Lactobacillus strains were measured. As shown in Fig. 3, the surface hydrophobicity of the representative strains of LAB ranged from 19-48%, with strains QZ21, QZ58, QZ34 and QZ50 showing high surface hydrophobicity, while the remaining the remaining 2 strains showed lower surface hydrophobicity. As shown in Fig. 4, the self-agglutination of LAB strains varied from 26.3-58.3%, with strains QZ71 and QZ50 showing signi cantly higher self-agglutination than the remaining 4 strains. The self-agglutination of 2 strains, QZ71 and QZ81, was signi cantly lower than that of the remaining strains.
The acid production capacity of the six selected strains of LAB is shown in the Fig. 5, there was no signi cant difference in acid production capacity, and strain QZ50 has the strongest acid production capacity, at 36 hours, the pH was reduced to 3.98. Combined with the above experimental results and analysis of acid production capacity and surface hydrophobicity and self-aggregation and bacterial inhibition, three strains of bacteria, QZ14 QZ50 and QZ71 were selected for simulated gastrointestinal survival experiments. As shown in the Fig. 6, all three strains of LAB survived after 7h incubation in the simulated gastrointestinal tract. After 3h of arti cial gastric juice, the viable counts of the three strains of LAB decreased to different degrees, among which the viable counts of QZ14 and QZ71 were signi cantly reduced, indicating that these two strains were less tolerant to the arti cial gastric juice. After transferring the lactic acid bacteria from the simulated gastric juice to the arti cial intestinal uid, the viable counts of QZ14, QZ71 and QZ50 were signi cantly reduced, and the viable counts of all three strains of LAB were signi cantly reduced compared with the initial ones after 4h incubation in the arti cial intestinal uid. Collectively, the strains of LAB were less affected by the arti cial gastric uid, especially ZX50 had no signi cant decrease in viable bacteria count after incubation in gastric uid. In contrast, all strains showed a signi cant decrease of at least 0.89 log10 CFU/mL after incubation in intestinal uid.
Optimization of media and culture conditions Some more in-depth studies were performed on Strain QZ50, a broad-spectrum bacteriocin-producing strain with the highest antibacterial activity. The effects of different media, temperature, initial pH values, and inoculum amount on bacteriocin production are shown in Fig. 7. The highest inhibition zone diameter was obtained in MRS broth, at a temperature of 30°C, a pH value of 6.5, and an inoculum amount of 3%. The lowest inhibition zone diameter was obtained with M17 broth, at 25°C, a pH of 4.5, and a 1% inoculum amount.
Compared with the conditions of MRS broth, 30°C, pH 6.5, the antimicrobial activity of strain QZ50 under other conditions was signi cantly different (P < 0.001). Compared with the inoculum amount of 3%, other inoculum amount except 2% was signi cantly different (P < 0.001).

Optimization of medium components
The effects of different C and N sources on bacteriocin production are respectively shown in Fig. 8 and Fig. 9.
The antimicrobial activity of strain QZ50 due to different C sources was signi cantly different (P ≤ 0.001). As shown in Fig. 8, Glucose contributed to the highest inhibition zone diameter, while soluble starch contributed to no bacteriocin production. Lower inhibition zone diameters were obtained in a descending order as: fructose, sucrose, maltose, lactose, and cellobiose.
As shown in Fig. 9, yeast extract contributed to the highest diameter of inhibition, while inorganic N contributed to no bacteriocin production. Except ammonium citrate and ammonium chloride, the antimicrobial activity of strain QZ50 due to different N sources was signi cantly different (P 0.05). Lower inhibition zone diameters were obtained in a descending order as: tryptone, peptone, casein peptone and beef extract.
The effect of different stimulating factors on bacteriocin production is presented in Fig. 10(A). The diameter of inhibition of QZ50 was the highest with Tween 80 as surfactant. The antimicrobial activity of strain QZ50 due to different surfactants was signi cantly different (P ≤ 0.01). As shown in Fig. 10(B), the levels of stimulating factor in the culture medium had signi cant effect on bacteriocin production(P < 0.01). The diameter of inhibition of QZ50 was the highest with 2% v/v Tween 80 and the lowest with no Tween 80.
Resistance of bacteriocin against heat, pH, and enzymes As shown in Table 4, the antibacterial activity of plantaricin QZ50 completely disappeared following treatment with pepsin, protease K, trypsin, papain, and chymotrypsin, while amylase and lipase had no effect. Plantaricin QZ50 retained its activity following incubation at 60°C for 15 and 30 min, but activity slightly decreased at 80, 100, and 121°C. It also retained its activity at pH 2.0-8.0, but complete inactivation was founded at pH 9.0 or 10.0.

Discussion
Since nisin was approved as a food preservative, LAB and other bacteriocin-producing bacteria have become an important research focus. In this study, 112 acid-producing bacteria strains were isolated from traditional Tibetan Qula cheese and screened for broad-spectrum antibacterial activity. In addition to bacteriocin, LAB produce other antibacterial compounds such as organic acids and hydrogen peroxide. Therefore, the isolates were screened for antibacterial activity after eliminating for organic acid and hydrogen peroxide effects. Among the isolates, only strain QZ50 retained its broad-spectrum antibacterial activity. The antimicrobial compound produced by strain QZ50 showed sensitivity towards proteases, indicative of a protein nature. In addition, QZ50 showed stronger inhibition against Micrococcus luteus ATCC 28001, Staphylococs aureus ATCC 26003, and Bacillus subtilis ATCC 63501 than against Escherichia coli ATCC 30105, Pseudomonas aeruginosa ATCC 10104, and Salmonella enterica ATCC 50094. These results, consistent with previous observations [21,22], con rmed that bacteriocins produced by Gram-positive bacteria generally exhibit stronger antimicrobial activity against Gram-positive bacteria than against Gram-negative bacteria.
Cheese represents a good source of LAB. Certain bacteriocin-producing LAB such as Enterococcus faecium MMT21 [23], Enterococcus faecalis WHE 96 [24], and Enterococcus faecium SD1 [25] have been successively isolated from cheese. However, to the best of our knowledge, this is the rst report on a bacteriocin-producing L.
plantarum strain isolated from Qula cheese.
Culture conditions and medium composition are very important for bacteriocin production. The production of bacteriocins usually require complex media, and several media have been tested for bacteriocin production [26].
In the present study, MRS was a more suitable medium for bacteriocin production than other media, consistent with the ndings of Karthikeyan and Santosh (2009) [27] and Hoda et al. (2013) [26]. Additionally, culture temperature and pH can affect bacteriocin production [26]. In this study, the maximum bactericidal activity was  [31], who reported that the optimum temperature for bacteriocin production from Leuconostoc carnosum LA54A and Enterococcus faecium RZS C5 was 25 and 35°C, respectively. Cell growth and metabolite accumulation are dependent on pH [32]. In this study, the production of plantaricin QZ50 was signi cantly affected by the initial pH value of the culture medium. This is consistent with observations obtained from Pediococcus acidilactici BA28 and Bacillus cereus XH25 [32,33]. However, in the study of Embaby et al. (2014) [34], the production of bacteriocin from Bacillus sp. YAS 1 was not affected by the initial pH. The optimum pH for bacteriocin production is not consistent across LAB strains. The optimum pH for bacteriocin production is 6 from Lactobacillus spp. [26], and < 5 from L. sakei CCUG 42687, L. acidophilus JCM 1132, and L. plantarum LL41 [29,35,36]. In this study, the optimum pH for bacteriocin production from L. plantarum QZ50 was pH 6.5, which might contribute to the transcription and synthesis of plantaricin QZ50. The composition of the medium has important effects on bacteriocin production. To enhance the production and reduce the cost of the medium, several studies have been performed to optimize the composition [26,37,38]. The C source, N source, and percentage of Tween 80 in MRS medium are the most common studied factors [38]. In this study, different C sources contributed to different levels of plantaricin QZ50. This nding was in agreement with earlier reports, which showed that the C source had a signi cant effect on bacteriocin production [38,39]. Additionally, glucose contributed to the highest bacteriocin production. Pal [34]. In this study, lactose was utilized by L. plantarum QZ50 to a lower level than fructose or sucrose. However, lactose favored the production of bacteriocin from Enterococcus faecium RZS C5 and Lactobacillus sp. MSU3IR [31,42]. Organic N had a signi cant effect on bacteriocin production, in agreement with earlier reports [38,39]. Our ndings revealed that yeast extract was the optimum N source, consistent with the reports of Kim et al. (2000) [43], but different from Pal et al. (2010) [38]. In the study of Pal et al. (2010), tryptone was optimum for production of bacteriocin by Weissella paramesenteroides DFR-8 [38]. Inorganic N resulted in no bacteriocin production. However, Iyapparaj et al. (2013) reported that inorganic N contributed to increased bacteriocin yield [42]. In this study, plantaricin QZ50 production was the highest with Tween 80. Similar observations were reported with Lactobacillus sp. MSU3IR [42] and L. plantarum 423 [44]. As a surfactant, Tween 80 reduces the surface tension between bacteriocins and bacteria, thereby improving permeability of cell membranes, which improves the release of bacteriocin from the cell surface [45].  [42] reported that the optimum Tween 80 concentrations ranged from 1 to 1.5%. Based on the results obtained, it is important to conduct optimization experiments for novel bacteriocin-producing isolates.
Next, the partial biochemical characteristics of plantaricin QZ50 were evaluated. In general, bacteriocins are sensitive to proteolytic enzymes. The antibacterial activity of plantaricin QZ50 was lost following treatment with pepsin, proteinase K, trypsin, papain, and chymotrypsin; however, it was not affected by amylase or lipase, which con rmed that plantaricin QZ50 was protein-based. Plantaricin QZ50 retained its antibacterial activity at 60-121°C. Bacteriocin from strain Bacillus sp., BAC YAS 1 from Bacillus sp. is stable at 45-80°C [34], Sh10 is stable at 30-121°C [46], and bacteriocin from Lactobacillus sake C2 remained activity at 80-121°C [47]. The resistance of plantaricin QZ50 against heat makes it applicable in food processing and preservation. In this study, the antibacterial activity of plantaricin QZ50 uctuated with increasing temperatures. This result did not agree with past studies, which reported that the activity of bacteriocin declined with increasing temperatures [34,42,47,48]. Plantaricin QZ50 was stable at pH 2-8, similar to bacteriocins from Bacillus subtilis 14B and Lactobacillus brevis FPTLB3 [48,49]. In studies by Embaby et al. (2014) [34], bacteriocins from Bacillus sp. YAS 1 exhibited a higher pH stability range (3)(4)(5)(6)(7)(8)(9)(10)(11)(12) than that of plantaricin QZ50. In the studies by Magnusson and Li et al. (2015) [11] and Schnürer (2001) [50], the activity of bacteriocins from strains ZZU 203 and 204 and Lactobacillus coryniformis subsp. coryniformis Si3 disappeared at pH > 6.0. The stability of plantaricin QZ50 at low pH values makes it useful for applications in acidic environments. Differences in physicochemical characteristics among various bacteriocins is mainly caused by differences in amino acid composition. The analysis of the amino acid composition of plantaricin QZ50 is currently underway. In addition, we can control and select the conditions of various bacteriocins based on their physicochemical properties, which will greatly improve their applications.

Conclusions
In this study, some strains of LAB isolated from the Qula in the Tibetan plateau had good antibacterial activity which could be considered as potential probiotic. Especially, L. plantarum QZ50 had obvious broad-spectrum antibacterial activity and the inhibition ability was in uenced by the change of conditions. The optimum medium, temperature, initial pH, and inoculum amount for plantaricin QZ50 production were Man, Rogosa, and Sharpe (MRS), 30°C, 6.5, and 3%, respectively. Overall, the strain of L. plantarum QZ50, with a broad-spectrum, stable, safe, and natural antibiotic, has potential applications as a food biopreservative.

Samples and LAB isolation
A total of 31 traditional hand-made Qula cheese samples were collected from herdsmen's families in different areas of Qinghai province, China. The samples were collected in sterilized bags, stored in ice boxes, and immediately sent to the laboratory. Qula samples (10 g) were mixed with 90 ml sterilized water and serially diluted (10 ×) with sterilized water. The 10 − 1 to 10 − 5 dilutions (10 µl) were inoculated into Man, Rogosa, and Sharpe (MRS) agar [1] and incubated at 30°C for 48 h under anaerobic conditions. Following the 48-h incubation, we selected single colonies based on size, shape, and color and puri ed them on MRS agar. Gram staining and catalase test were performed on the puri ed colonies for the rst identi cation of isolates. Only Gram-positive and catalase-negative isolates were selected for the tests described below.

Preliminary Screening of of Lactic Acid Bacteria with Antibacterial Activity
The Oxford cup double-layer plate method described by Muhammad et al in 2019 [7] was used for the preliminary screening of all the isolated presumptive lactic acid bacteria from the hand-made Qula cheese samples, screening out the lactic acid bacteria with antimicrobial activity, the indicator strains were E. coli K88 and Salm. enterica ATCC 13076, which were kept in our laboratory. The materials and methods were as follows: measure 15 mL of NA solid medium poured into a Petri dish and cooled as the lower layer of medium, optical density (OD) at 600 nm adjusted to 1 and inoculated into the NA solid medium cooled to about 50°C at 3% inoculum, mixed well, and poured 5 mL into the lower medium with a pipette. 200µL of overnight culture of Lactobacillus cell-free supernatant in the Oxford cup, incubate at 37℃ for 24h, and observe whether there are inhibition circles, and measure the diameter of inhibition circles using vernier calipers.

Physiological and Biochemical Characteristics
All of the strains were initially denti ed by physiological and biochemical characteristics, including experiments of Gram reactions, catalase activity and gas production in the presence of glucose as described by Kozaki

Screening of LAB strains producing broad-spectrum bacteriocins
The antimicrobial activity of the isolates was examined by the agar diffusion assay method [10] with Micrococcus luteus ATCC 28001, Staphylococcus aureus ATCC 26003, Bacillus subtilis ATCC 63501, Escherichia coli ATCC 30105, Pseudomonas aeruginosa ATCC 10104, and Salmonella enterica ATCC 50094 as indicator strains. Overnight cultures in MRS broth were centrifuged at 10,000 g for 20 min. Each indicator strain was suspended in sterile water and standardized to an absorbance of 1 at 600 nm following activation. An aliquot of cell-free supernatant (300 µl) was placed in wells (7.80 mm in diameter) of MRS agar plates and inoculated with 10% (v/v) of indicator strains. Following incubation at 37°C for 24 h, the diameters of the zone of inhibition were measured.
To eliminate the antimicrobial effect of organic acids, the pH value of the cell-free supernatants and MRS broth (control) was adjusted to 6.0 with NaOH (2 mol/L) and lactic acid (2 mol/L), respectively. To eliminate the antimicrobial effect of hydrogen peroxide, we incubated supernatants (pH 6.0) at 37°C for 2 h with catalase (5 mg/ml) and without catalase (control) [11].
Cell-free supernatants still remaining antimicrobial activity were screened for testing the possible protein nature of antimicrobial compounds. The cell-free supernatants (pH 6.0) were incubated at 37°C for 2 h in the presence of trypsin (3 mg/ml), protease K (3 mg/ml), and no enzyme (control).
Identi cation of selected strains by 16S rRNA gene sequence analysis

Antibiotic Susceptibility
Antibiotic susceptibility testing was performed by the drug-sensitive tablet method described by Wang X et al [14] to test the susceptibility to 10 common antibiotics. Measure 15 mL of MRS solid medium into a Petri dish for cooling as the lower medium, add the overnight culture of Lactobacillus strains. to MRS agar medium at 5% inoculum and mix well, weigh 5 mL and pour it onto the cooled lower plate, and after cooling and solidifying, place the antibiotic susceptibility tablets on the agar plate with sterile forceps and incubate anaerobically at 30°C for 24 h. After that, use Vernier calipers to measure the diameter of the inhibition circle, 3 parallel experiments were performed to take the average value.

Cell Surface Hydrophobicity and Aggregation Assay
The cell surface hydrophobicity and auto-aggregation assay of selected strains were performed according to methods described by Sirichokchatchawan et al in 2018 [15]. Firstly, the preparation of Lactobacillus suspension was carried out, and Lactobacillus was cultured overnight in MRS liquid, centrifuged to take the bacteria, washed twice with PBS at pH 6.5, and then the bacteria were resuspended in PBS and the absorbance was adjusted to 0.6 at 600 nm (A0). The hydrophobicity of Lactobacillus cell surface was measured by the method of microbial adherence to hydrocarbons. The absorbance (At) of the aqueous phase at 600 nm was measured by taking 3 mL of Lactobacillus cell suspension and mixing it well with 1 mL of xylene and leaving it at 37°C for 40 min. The cell surface hydrophobicity (%) was calculated by the following formula: (1-At/A0)×100.
Self-agglutination was measured according to the method of De Souza et al in 2019 [16]. Firstly, 1 mL of lactic acid bacteria suspension was taken and left for 2 h at 37°C, and the absorbance (A1) of its upper liquid layer at 600 nm was measured, then its self-agglutination (%) = Optimization of plantaricin QZ50 production Effect of different media. Complete medium (CM) [18], MRS, medium 17 (M17) [19], and all-purpose Tween (APT) [20] were used to assess the optimum medium for bacteriocin production at 30°C for 24 h. Broth medium (pH 7.0) was used as the fermentation medium.
Optimization of culture conditions. The effect of temperature, inoculum amount, and initial pH on bacteriocin production were evaluated. To evaluate the effect of stimulating factor on bacteriocin production, 2% Tween 80, Tween 20, or polyethylene glycol (PEG) 6000 was added to MRS medium. Additionally, Tween 80 at different concentrations (0, 1, 2, 3, 4, 5, and 6%) was used to determine the optimum concentration of Tween 80. In each experiment, antibacterial activity was conducted as previously described.
Resistance of bacteriocin against heat, pH, and enzymes The resistance of bacteriocin to heat, pH, and enzymes was examined. Cell-free culture supernatants were heated to 60, 80, or 100°C in a water bath and to 120°C in an autoclave for 15 and 30 min. To assess the sensitivity to different pH values, cell-free culture supernatants were adjusted to pH 2, 3, 4, 5, 6, 7, 8, 9, or 10 and incubated at 37°C for 2 h. Cell-free culture supernatants were treated with α-amylase (1 mg/ml), β-amylase (1 mg/ml), lipase (1 mg/ml), pepsin (3 mg/ml), trypsin (3 mg/ml), protease K (3 mg/ml), papain (3 mg/ml), or chymotrypsin (3 mg/ml) at 37°C for 2 h. Prior to the 2-h incubation, the pH of the cell-free supernatants was adjusted to the optimum pH of each enzyme. After all treatments, antibacterial activity was determined as previously described.

Statistical analysis
Each assay was repeated on three independent occasions with triplicate determinations. Statistical analysis was performed using SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA) with statistical signi cance determined at P < 0.01 or 0.05. Results are expressed as the mean and standard error of the mean of three independent experiments. One-way ANOVA followed by Least signi cant difference test was used to determine signi cant differences of the antimicrobial activity due to the different media, culture conditions, medium components. Availability of data and materials All data generated or analysed during this study are included in this published article and are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.