Safety Evaluation of Bifidobacterium breve IDCC4401 Isolated from Infant Feces for Use as a Commercial Probiotic

Previously, our research group isolated Bifidobacterium breve IDCC4401 from infant feces as a potential probiotic. For this study, we evaluated the safety of B. breve IDCC4401 using genomic and phenotypic analyses. Whole genome sequencing was performed to identify genomic characteristics and investigate the potential presence of genes encoding virulence, antibiotic resistance, and mobile genetic elements. Phenotypic analyses including antibiotic susceptibility, enzyme activity, production of biogenic amines (BAs), and proportion of D-/L-lactate were evaluated using E-test, API ZYM test, high-performance liquid chromatography (HPLC), and D-/L-lactic acid assay respectively. The genome of B. breve IDCC4401 consists of 2,426,499 bp with a GC content of 58.70% and 2,016 coding regions. Confirmation of the genome as B. breve was provided by its 98.93% similarity with B. breve DSM20213. Furthermore, B. breve IDCC4401 genes encoding virulence and antibiotic resistance were not identified. Although B. breve IDCC4401 showed antibiotic resistance against vancomycin, we confirmed that this was an intrinsic feature since the antibiotic resistance gene was not present. B. breve IDCC4401 showed leucine arylamidase, cystine arylamidase, α-galactosidase, β-galactosidase, and α-glucosidase activities, whereas it did not show production of harmful enzymes such as β-glucosidase and β-glucuronidase. In addition, B. breve IDCC4401 did not produce any tyramine, histamine, putrescine, cadaverine, or 2-phenethylamine, which are frequently detected BAs during fermentation. B. breve IDCC4401 produced 95.08% of L-lactate and 4.92% of Dlactate. Therefore, our findings demonstrate the safety of B. breve IDCC 4401 as a potential probiotic for use in the food industry.


Antibiotic Susceptibility of B. breve IDCC4401
The antibiotic susceptibility of B. breve IDCC4401 was determined by the E-test method against 9 antibiotics according to the European Food Safety Authority (EFSA) guidelines [17]. An overnight culture of bacteria (10 8 CFU/ml) was swabbed onto 15-cm diameter MRS agar plates with a sterilized cotton swab prior to placing an E-test strip (Liofilchem Inc., Italy) on the surface of the plate. After incubation at 37°C for 18 h, the relevant inhibition ellipse intersected the strip and the minimum inhibitory concentration was determined at complete inhibition. Finally, antibiotic susceptibility of B. breve IDCC4401 was determined following the guidelines of the EFSA [17].

Enzyme Activities of B. breve IDCC4401
The enzyme activities of B. breve IDCC4401 were determined using the API ZYM Kit (Biomerieux Inc., France) according to the manufacturer's instructions. The overnight culture of B. breve IDCC4401 (10 9 CFU/ml) was added into a cupule containing different substrate solutions and incubated at 37°C for 4 h. One drop of ZYM A and ZYM B reagents was added sequentially prior to incubation for 5 min at RT. Color change of the mixture was graded from zero (no activity) to five by comparing color intensity with the color chart provided by the manufacturer. Positive enzyme activities were determined to be above three intensity levels of color change.

BA Production of B. breve IDCC4401
Production of BAs by B. breve IDCC4401 was investigated following the method described in a previous study [22] with minor modifications. The overnight cultured B. breve IDCC4401 was centrifuged at 2,300 ×g for 5 min at 4°C. An aliquot of 0.75 ml of supernatant was mixed with the same volume of 0.1 M HCl and filtered through a 0.45-μm membrane to extract BAs. For derivatization of the BAs, 1 ml of filtered BAs was incubated at 70°C for 10 min prior to addition of 200 μl of saturated NaHCO 3 , 20 μl of 2 M NaOH, and 0.5 ml of dansyl chloride solution (10 mg/ml of acetone). The derivatized BAs were mixed with 200 μl of proline (100 mg/ml of H 2 O) and incubated for 15 min at RT in the dark. The derivatized BAs were then separated and quantified using HPLC (LC-NETII/ ADC, JASCO Inc., Japan) with an Athena C18 column (4.6 mm × 250 mm, ANPEL Laboratory Technologies Inc., China). Aqueous acetonitrile solution (Sigma-Aldrich Co., USA) was used as a mobile phase and the flow rate was adjusted to 0.8 ml/min. Finally, a peak was detected at 254 nm using a UV detector (UV-2075 Plus, JASCO Inc., Japan). The detected BAs were quantified from calibration curves of BAs including tyramine, histamine, putrescine, 2-phenethylamine, and cadaverine (Sigma-Aldrich Co.).

Proportion of D-/L-Lactate of B. breve IDCC4401
An overnight culture of B. breve IDCC4401 was centrifuged at 2,300 ×g for 30 min at 4°C and the supernatant was collected. Following supernatant filtration using a 0.2-μm pore size membrane, the filtrate was mixed with the agents in a D-/L-Lactic Acid (D-/L-Lactate) (Rapid) Assay Kit (Megazyme, Ireland). Absorbances of the mixture were measured at 340 nm and the concentration of D-/L-lactate was calculated according to the manufacturer's protocol.

Whole Genome Sequencing of B. breve IDCC4401
Whole genome sequencing of B. breve IDCC4401 was performed for the identification and confirmation of genes encoding antibiotic resistance, virulence, and mobile genetic elements to ensure safety. The assembled genome consisted of 2,426,499 bp with a GC content of 58.70% (Fig. 1). The strain was confirmed as B. breve with a similarity of 98.93% with B. breve DSM20213 based on ANI analysis. Of a total of 2,016 CDSs of B. breve IDCC4401, 1,583 CDSs were annotated as functional genes involved in translation, ribosomal structure, biogenesis, RNA processing, modification, transcription, replication, recombination, repair, cell cycle control, cell division, chromosome partitioning, defense mechanisms, signal transduction mechanisms, and 433 unknown genes (Table 1). Based on CARD and VFDB, genes associated with virulence and antibiotic resistance were not found in B. breve IDCC4401, respectively. Although 33 transposases, as mobile genetic elements, were identified, these mobile elements were not involved in the acquisition and transfer of antibiotic resistance genes due to the absence of virulence and antibiotic resistance genes in B. breve IDCC4401. Therefore, this result confirmed the safety of B. breve IDCC4401 for use as a probiotic based on our thorough genome analysis.

Antibiotic Susceptibility (MICs) of B. breve IDCC4401
To ensure safety, the phenotypic antibiotic susceptibility of B. breve IDCC4401 was investigated against 9 antibiotics including ampicillin (one representative of a β-lactam antibiotic), gentamicin, streptomycin, erythromycin, clindamycin, tetracycline, chloramphenicol, vancomycin, and kanamycin, using the E-test method [18]. Determination of resistance and susceptibility against each antibiotic followed the EFSA guidelines. As shown in Table 2, B. breve IDCC4401 was susceptible to ampicillin, gentamicin, streptomycin, erythromycin, clindamycin, tetracycline, and chloramphenicol but resistant to vancomycin. Susceptibility to kanamycin could not be determined because EFSA did not provide a definite cut-off value. However, the MIC value (256 μg/ml) against kanamycin was 2 times lower than that of B. longum BORI (512 μg/ml), and 4 times lower than those of B. longum BB536, B. breve M-16, B. bifidum BGN4, and B. lactis BB-12 (1,024 μg/ml) that were considered as GRAS [16].  isolated from human gastrointestinal tract and found resistances against five β-lactam antibiotics (penicillin G, amoxycillin, cepharadine, ceftizoxime and cefotaxime) among 44 antibiotics considered as "last resort antibiotics" [23]. In the meantime, B. breve IDCC4401 exhibited susceptibility to ampicillin. In addition, two other B. breve strains (IF2-173 and IF2-174) also isolated from breast-fed infants were susceptible to ampicillin, streptomycin, and chloramphenicol whereas they were resistant to tetracycline, erythromycin, and vancomycin [10]. When compared with previous study [10], B. breve IDCC4401 showed narrow antibiotic resistance only to vancomycin. Although B. breve IDCC4401 showed antibiotic resistance against vancomycin, this result could be related to an intrinsic feature of B. breve IDCC4401 as vancomycin resistance is a general feature of most Bifidobacterium spp. as shown in a study by Charteris et al. [23]. Vancomycin resistance is thought to be a result of the presence of D-alanine residues in the cell wall preventing vancomycin binding [24,25].
Desjardins et al. 's study demonstrated that four B. breve strains (ATCC 15698, ATCC 15699, ATCC 15700, and ATCC 15701) obtained from the American Type of Culture Collection (ATCC) exhibited the activities of esterase lipase, leucine aminopeptidase, acid phosphatase, phosphoamidase, α-galactosidase, β-galactosidase, α-glucosidase, and β-glucosidase [27]. In addition, B. breve ATCC 15699 showed β-glucuronidase activity. The enzyme activities of α-galactosidase, β-galactosidase and α-glucosidase were found to be consistent with those of a previous study [27]. Chevalier et al. (1990) demonstrated that the presence of Bifidobacterium spp. in the feces of the healthy infants was accompanied by α-galactosidase and α-glucosidase, which are characteristics of Bifidobacterium spp. [28]. Further, evidence proved that α-galactosidase and β-galactosidase enzyme activities were highly produced in B. longum RD 47, suggesting that these enzymes are involved in hydrolyzing galactosides such as lactose [29]. These findings are consistent with our results obtained from this study. Although probiotics can improve digestion and probiotic enzymes can act as natural substances for digestion of food in the human body, some enzyme activities may produce compounds that are to the host [30]. Cole and Fuller (1986) reported that βglucosidase may produce aglycones which are linked to the development of colorectal cancer [31]. Moreover, βglucuronidase might be linked to carcinogenic compounds for colorectal cancer [32]. Since B. breve IDCC4401 did not show the activity of β-glucosidase and β-glucuronidase, it appears to be safe and suitable for use as a probiotic.

BA Production of B. breve IDCC4401
Lactic acid bacteria (LAB) including Bifidobacterium spp. and Lactobacillus spp. can produce BAs during the fermentation process [33]. Previous studies [4,34] reported that tyramine, histamine, putrescine, cadaverine, 2phenethylamine, and spermidine can be frequently detected in fermented foods. The problem is that these BAs, at high concentration, can cause toxicological effects to humans with certain symptoms such as respiratory distress, heart palpitation, hypertension or hypotension, headaches, and allergenic disorders [33]. Verifying the formation of BAs by B. breve IDCC4401 is required to ensure the safety of B. breve IDCC4401. As shown in Table 4, B. breve IDCC4401 did not produce any of the five representative BAs, indicating that it did not produce any harmful BAs.
According to Lorencova et al. 's study [35], five Bifidobacterium strains (Bifidobacterium spp., B. adolescentis, B. lactis, B. bifidum, and B. longum) produced tyramine, cadaverine, putrescine, and spermidine among the eight BAs including the five we tested and three others (tryptamine, speramine, and spermidine). However, Kim et al. 's study [16] exhibited putrescine formation by B. bifidum BGN4 and B. longum BORI, but not cadaverine, histamine, or tyramine formation. Contradictory to previous studies, Ku et al. [4] demonstrated that B. lactis AD011 did not form any BAs when tested for cadaverine, histamine, putrescine, and tyramine formation, in accordance with our results. Thus, the absence of BAs suggested the potential of B. breve IDCC4401 as a commercial probiotic.

Proportion of D-/L-Lactate of B. breve IDCC4401
Lactate produced in either D-form or L-form isomers during fermentation by LAB and exhibits distinct biological effects in human [36]. However, Lactobacillus and Bifidobacterium have been known as D-lactate producers [37,38]. Unlike L-lactate, excess D-lactate produced by LAB cause short bowel syndrome, chronic fatigue and metabolic disorders, especially when jejunoileal bypass surgery is performed [39]. Due to these  toxicological effects, the proportion of D-/L-lactate of B. breve IDCC4401 needs to be clarified for its safety evaluation. As shown in Table 5, B. breve IDCC4401 produced 95.08% (21.26 g/l) of L-lactate and 4.92% (0.93 g/l) of D-lactate. Since the guidelines provided by FAO/WHO did not provide any clear criterion for the ratio of L-lactate and Dlactate, these results were compared with other studies recognized as GRAS. According to Munoz et al. 's study [40], B. longum CECT 7210 produced 2.22% (0.06 g/l) of D-lactate, which was lower than the L-lactate (97.78%, 2.64 g/l). Although the amount and ratio of D-lactate produced by B. breve IDCC4401 was higher than that of B. longum CECT 7210, it was lower than that of L-lactate and this pattern was similar to a previous study. In the case of B. lactis BB-12 ® which was certified as GRAS by the FDA, the ratio of L-lactate was more than 95% [41]. In this aspect, it can be assumed that B. breve IDCC4401 poses no safety concern over this property.
In conclusion, since the safety of B. breve IDCC4401 isolated from infant feces has not yet been investigated for its commercial usage as a probiotic, we investigated the presence of genes encoding virulence, antibiotic resistance, mobile genetic elements, and deleterious metabolic activities in this strain to ascertain its safety as a probiotic. Although B. breve IDCC4401 was evaluated for safety in this study, further studies are required on functional characteristics such as resistance to gastric acid and bile salts, adherence ability (aggregation properties and cell hydrophobicity), and antimicrobial activity to ensure its safety as a commercial probiotic in the food industry.