Prevalence, antimicrobial susceptibility, and antibiotic resistance gene transfer of Bacillus strains isolated from pasteurized milk

Pasteurization is carried out in dairy industries to kill harmful bacteria present in raw milk. However, endospore-forming bacteria, such as Bacillus , cannot be completely eliminated by pasteurization. In this study, a total of 114 Bacillus strains were isolated from 133 pasteurized milk samples. Antibiotic susceptibility tests showed that the percentage of Bacillus with intrinsic resistance to ampicillin and penicillin were 80 and 86%, respectively. Meanwhile, some Bacillus isolates had acquired resistance, including trimethoprim-sulfamethoxazole resistance (10 isolates), clindamycin resistance (8 isolates), erythromycin resistance (2 isolates), and tetracycline resistance (1 isolate). To further locate these acquired resistance genes, the plasmids were investigated in these 16 Bacillus strains. The plasmid profile indicated that Bacillus cereus BA008, BA117, and BA119 harbored plasmids, respectively. Subsequently, the Illumina Novaseq PE150 was applied for the genomic and plasmid DNA sequencing. Notably, the gene tetL encoding tetracycline efflux protein was found to be located on plasmid pBC46-TL of B. cereus BA117. In vitro conjugative transfer indicated that pBC46-TL can be transferred into Bacillus invictae BA142, Bacillus safensis BA143, and Bacillus licheniformis BA130. The frequencies were of 1.5 × 10 −7 to 1.7 × 10 −5 transconjugants per donor cells. Therefore, Bacillus strains with acquired antibiotic resistance may represent a potential risk for the spread of antibiotic resistance between Bacillus and other clinical pathogens via horizontal gene transfer.


INTRODUCTION
Bovine milk is highly nutritious and has a near-neutral pH (6.6-6.8), which can be a suitable growth medium for various bacteria including Enterobacteriaceae, Streptococcaceae, and Bacillaceae (Bartoszewicz et al., 2008;Quigley et al., 2013). These bacteria can influence the quality and safety of fluid milk or fermented dairy products. During the dairy product manufacturing, most of these bacteria can be destroyed by pasteurization. However, endospore-forming bacteria, such as Bacillus cannot be completely eliminated (Gopal et al., 2015). It has been reported that Bacillus from pasteurized milk showed resistance to some antibiotics including ampicillin, lincomycin, erythromycin, and tetracycline Zhao et al., 2020). For example, Bacillus strains were isolated from pasteurized milk samples collected from dairies in France and showed resistance to penicillin or erythromycin (Perrin-Guyomard et al., 2005). Seventy Bacillus cereus strains were isolated form 258 pasteurized milk samples collected from 32 cities in China and most of the isolates were resistant to ampicillin (99%), penicillin (99%), and cefoxitin (95%; Gao et al., 2018). Thus, it is necessary to assess the prevalence and antimicrobial susceptibility of Bacillus strains in pasteurized milk.
In Bacillus, the resistance to antibiotics could be either intrinsic or acquired. Generally, genes encoding acquired antibiotic resistance are located on mobile genetic elements such as plasmids or transposons, which could lead to the spread of antibiotic resistance among Bacillus and other clinical pathogens via horizontal gene transfer (Navaneethan and Effarizah, 2021). The erythromycin resistance gene on plasmid pHT73 can be transferred from Bacillus thuringiensis ssp. kurstaki Prevalence, antimicrobial susceptibility, and antibiotic resistance gene transfer of Bacillus strains isolated from pasteurized milk Zhengyuan Zhai, 1,2 Chang Cui, 1 Xueli Li, 1 Juan Yan, 1 Erna Sun, 1 Chenyuan Wang, 1 Huiyuan Guo, 1,2 and Yanling Hao 1,2,3 * † KT0 to B. cereus and Bacillus mycoides by conjugation (Hu et al., 2004). The tetracycline resistance gene tet(M) located on a Tn916-like transposon can be transferred from B. cereus R89 to Enterococcus faecalis JH2-2 and Staphylococcus aureus 8794RF by filter mating (Agersø et al., 2002). Therefore, Bacillus strains in pasteurized milk could act as donors or reservoirs for antibiotic resistance genes (ARG), which can be transferred to pathogenic bacteria (von Wintersdorff et al., 2016). Moreover, European Food Safety Authority recommends that bacterial strains harboring transferable ARG should not be present in foods for human (European Food Safety Authority, 2007). Thus, it is necessary to determine the transferability of ARG from Bacillus strains in pasteurized milk, which is crucial for preventing the spread of ARG via food chain.
To assess the prevalence and antimicrobial susceptibility of Bacillus strains in pasteurized milk, 114 Bacillus strains were isolated and identified from the pasteurized milk collected from different regions in China between May 2019 and March 2020. Subsequently, the broth microdilution method was used to assess the antibiotic susceptibility profiles of these Bacillus strains. Some Bacillus isolates showed acquired resistance to trimethoprim-sulfamethoxazole, clindamycin, erythromycin, and tetracycline, respectively. Genomic and plasmid DNA sequencing indicated that a tetracycline resistance gene tetL was located on plasmid pBC46-TL in B. cereus BA117. Moreover, the plasmid pBC46-TL can be transferred into Bacillus invictae BA142, Bacillus safensis BA143, and Bacillus licheniformis BA130 by filter mating. Altogether, our results indicated that the prevalence and antimicrobial susceptibility of Bacillus strains isolated from pasteurized milk should be paid more attention to prevent the spread of ARGs.

MATERIALS AND METHODS
Because no live human or animal subjects were used in this study, no approval from an Institutional Review Board or Institutional Animal Care and Use Committee was required.

Sample Collection and Preparation
To assess the prevalence of Bacillus in pasteurized milk products, 30 pasteurized milks (stored at 4°C and within the shelf life) were purchased from retail markets and supermarkets in Beijing, and 30 samples of pasteurized milk were provided by local dairy factories in Beijing between May 2019 and March 2020. Meanwhile, 73 raw milk samples were collected from local dairy farms in Beijing, Tianjin, Hebei, Henan, Jiangsu, Heilongjiang, and Inner Mongolia. All raw milk samples were collected from bulk tanks by staff with permission from the quality manager and filled into 50-mL sterile plastic tubes (Corning). These samples were kept on ice and transported to the laboratory within 24 h. To address the spore-forming Bacillus, the raw milks were pasteurized at 80°C for 10 min in the laboratory and kept below 4°C until analysis (Griffiths and Phillips, 1990).

Isolation and Identification of Bacillus Strains
Bacillus strains were isolated according to National Food Safety Standard (China, 2014) with minor modifications (Kwon et al., 2021). Briefly, 25-g samples were randomly collected from each pasteurized milk sample and put into a sterile blender jar with 225 mL of 0.85% sterile saline buffer, then homogenized for 2 min at high speed (10,000 to 12,000 rpm), followed by dilution until 10 −5 . An aliquot of 0.2 mL of each dilution was plated onto Mannitol-Egg-Yolk-Polymyxin Agar (AOBOX) and incubated at 30°C for 24 to 48 h. Pink colonies surrounded by a zone of precipitation were considered as presumptive B. cereus group strains, whereas yellow colonies without a zone of precipitation were considered as other Bacillus strains. Colonies on MYP plates were streaked onto Nutritional Agar plate and incubated at 30°C for 24 h before identification.
Genomic DNA of Bacillus strains was extracted using a TIANamp Bacteria DNA Kit (TIANGEN) according to the manufacturer's instructions. Concentration and purity of the DNA were measured by a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific). A 16S rDNA fragment was amplified by standard PCR using Q5 High-Fidelity DNA polymerase (NEB) and universal bacterial primers 27F and 1492R (Lane, 1991). The PCR products were sequenced by Sangon Biotech and analyzed by nucleotide blast (http: / / blast .ncbi .nlm .nih .gov/ Blast .cgi). The strains with high similarity to 16S rRNA sequence of the Bacillus reference strain (E-value = 0 and Max identity ≥98%) were regarded as Bacillus strains.

Plasmid DNA Extraction
Plasmid DNA of Bacillus strains that exhibited acquired antibiotic resistance was extracted using EZNA Plasmid DNA Mini Kit (Omega Bio-Tek Inc.) according to the manufacturer's instructions. Overnight cultures were inoculated (1% vol/vol) into 10 mL of Nutritional Broth and incubated at 30°C and 180 rpm. When cell density reached an OD 600 nm of 0.8, bacterial cells were collected by centrifugation at 6,000 × g for 10 min at 4°C for plasmid DNA extraction. Concentration and purity of the DNA were measured by a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific). Plasmid DNA samples were analyzed by electrophoresis on 1% wt/vol agarose gel (Pan et al., 2011). A supercoiled DNA ladder (catalog no. 3585A, Takara) was used to estimate the size of supercoiled plasmid. Agarose gel was stained with Gel Red (Biotium) in 1 × TAE buffer for 15 min. Then the plasmid DNA on gel was visualized with UV light (302 nm wavelength) and photographed by the gel documentation system (UVP).

Genome Sequencing and Detection of Antibiotic Resistance Genes
Genomic and plasmid DNA were sequenced using Illumina NovaSeq PE150 at the Beijing Novogene Bioinformatics Technology Co. Ltd., respectively. Briefly, 1 μg of DNA per sample was fragmented by sonication to a size of 350 bp, then DNA fragments were end-polished, A-tailed, and ligated with the full-length adaptor for Illumina sequencing with further PCR amplification. At last, PCR products were purified (AMPure XP system) and libraries were analyzed for size distribution by Agilent2100 Bioanalyzer and quantified using real-time PCR. The library was sequenced by Illumina NovaSeq PE150. The raw data obtained by sequencing was filtered to obtain Clean Data. The genome assembly with Clean Data was performed by SOAP denovo software. The functions of the predicted protein-coding genes were annotated with the Clusters of Orthologous Groups database using the WebMGA. Antibiotic Resistance Genes Database and Comprehensive Antibiotic Resistance Database were used for the prediction of antibiotic resistance genes (Liu and Pop, 2009;Alcock et al., 2020).

Conjugative Plasmid Transfer by Filter Mating
All bacterial strains used for conjugative plasmid transfer were listed in Table 1. Bacillus was grown in nutritional broth at 30°C. Lactobacillus was grown under anaerobic conditions at 37°C in a De Man-Rogosa-Sharpe medium. Escherichia coli was grown at 37°C in Luria Bertani broth. Transfer of tetracycline resistance gene tetL was determined by filter mating as described previously (Hinnekens et al., 2019) with slight modification. Both donor and recipient stains were incubated at 30 or 37°C to reach OD 600 nm of 0.8, respectively. Subsequently, 500 μL of donor and 500 μL of recipient were mixed well and passed through a 0.45-μm cellulose-nitrate membrane filter (Xingya). The filter was then collected and placed on a nonselective agar plate for 24 h. After mating, the filter was washed by 10 mL PBS. Cells were collected by centrifugation at 6,000 × g for 10 min at 4°C. Transconjugants were screened by plating serial dilutions of the mating cells on appropriate selection agar plates (Table 4). Antibiotics (Sigma-Aldrich) were added at the following final concentrations: 20 μg/mL of tetracycline for the donor strain BA117; 16 μg/mL of erythromycin for Bacillus recipient; 4 μg/mL of clindamycin for Bacillus recipient BA102, BA142, and BA143; 16 μg/mL of clindamycin for Bacillus recipient BA107, BA130, and BA131; 34 μg/mL of chloramphenicol for E. coli recipient; 8 μg/ mL of chloramphenicol for Lactobacillus recipient, respectively. Each mating experiment included a plating control of donor and recipient cells on selective medium to assess the presence of potential spontaneous mutants. The transfer frequencies were calculated as the ratio of transconjugants per donor cells (T/D) at the end of the conjugation time. The transfer of tetracycline resistance gene tetL was confirmed by plasmid profiling and PCR.

Nucleotide Sequence Accession Numbers
The genome sequence of B. cereus BA008, BA117, and BA119 was deposited at GenBank under the accession number PRJNA774561.

Prevalence Analysis of Bacillus in Pasteurized Milk
In this study, a total of 114 Bacillus strains were isolated from 133 pasteurized milk samples. These strains as shown in Table 2. The isolation rate of Bacillus was over 93.33% in pasteurized milk samples from market and dairy factory in Beijing. This agrees with the findings that the prevalence of Bacillus was high (~100%) in the pasteurized milk collected in China and Thailand (Chitov et al., 2008;Zhou et al., 2008). However, the prevalence of Bacillus was much lower in the milk samples pasteurized in the laboratory, which was 32.87% (24/73 samples). These findings were reasonable because the heat treatment of these raw milk samples was 80°C for 10 min, which was more sufficient than commercial pasteurization. In addition, postpasteurization contamination along the milk-processing lines was possibly a source of Bacillus in pasteurized milk. For instance, B. cereus showed outstanding ability to adhere to stainless steel surfaces of dairy plant and form biofilm (Kumari and Sarkar, 2016;Silva et al., 2018).
In this study, the isolate rate of Bacillus cereus was 41.23% (46/114), indicating that B. cereus is a common contaminant of pasteurized dairy products. It has been reported that the environments for milk production, handling, and processing could introduce B. cereus into dairy products . The heat-stable B. cereus spores in raw milk and the post-pasteurization contamination along the milk-processing lines were major sources of B. cereus contamination in pasteurized milk (Saleh-Lakha et al., 2017;Gao et al., 2018). The growth of B. cereus limits the shelf life of pasteurized milk. In addition to the risk of causing spoilage, Bacillus strains from pasteurized milk showed resistance to some antibiotics, such as ampicillin, erythromycin, and tetracycline Zhao et al., 2020). Therefore, Bacillus strains not only influence the shelf life of pasteurized milk, but also might be reservoir of ARG.

Antimicrobial Susceptibility Tests of Bacillus Isolates
The 114 Bacillus isolates were tested for susceptibilities to 13 antibiotics using the broth microdilution method. All tested isolates were susceptible to the remaining 5 antibiotics, including meropenem, vancomycin, gentamicin, ciprofloxacin, and rifampin. A high percentage of the isolates were resistant to ampicillin (80%, 91/114) and penicillin (86%, 98/114). This is consistent with previous studies that most isolates of Bacillus cereus group were intrinsically resistant to β-lactam antibiotics (da Silva Fernandes et al., 2014;Yibar et al., 2017) owing to β-lactamase production (King et al., 2017). However, the isolates identified as B. pumilus, B. safensis, and B. invictae were sensitive to both ampicillin and penicillin. These findings agree with 4 B. pumilus, 27 B. safensis, and 9 B. invictae isolated from different terrestrial sources and geographic locations were sensitive to both ampicillin and penicillin (Branquinho et al., 2015). It is important to note that a few Bacillus isolates displayed acquired resistance to some antibiotics. Ten Bacillus isolates (9%) were resistant to trimethoprim-sulfamethoxazole TET r = tetracycline resistance; CLI r = clindamycin resistance; ERY r = erythromycin resistance; CHL r = chloramphenicol resistance. Strain cannot be identified at species level by 16s rDNA sequencing.
( Figure 1). Eight strains were observed to be resistant to clindamycin. In addition, 2 isolates were resistant to erythromycin and 1 isolate named B. cereus BA117 was resistant to tetracycline (Table 3). Clindamycin, tetracycline, and erythromycin are frequently used for treatment of bovine mastitis in China (Gao et al., 2012). These antibiotics have been frequently found in cattle manure and wastewater (Zhou et al., 2013), which could create selective pressure for antibiotic-resistant Bacillus in dairy farms. The spores of these antibiotic-resistant Bacillus could be sources of contamination of raw milk. Notably, 14% (16/114) of the isolates exhibited multiple antibiotic resistance, which agrees with previous studies Gao et al., 2018). The antibiotic resistance of Bacillus isolates, especially acquired resistance, should be paid more attention to, as the acquired antibiotic resistance might  be further transferred to other pathogens via horizontal gene transfer .

Plasmid Profile of Bacillus Strains with Multiple Antibiotic Resistance
Antimicrobial resistance in bacteria is frequently encoded by genes carried on mobile genetic elements, in particular by plasmids (Partridge et al., 2018). The 16 Bacillus strains with acquired antibiotic resistance were tested for the presence of plasmids. The plasmid profile showed that 3 strains contained plasmids. B. cereus BA119, which was resistant to trimethoprimsulfamethoxazole, had a 3.5 kb plasmid. Another trimethoprim-sulfamethoxazole resistant strain, B. cereus BA008, had a large plasmid with an apparent size of 10 kb. Notably, B. cereus BA117, which was resistant to tetracycline, had at least 5 plasmids differing in size from 3.5 to 10 kb. To further identify and locate these ARG, both genomic and plasmid sequencing were carried out using Illumina NovaSeq PE150.

Detection of Antibiotic Resistance Genes by Genome and Plasmid Sequencing
The draft genome of B. cereus BA008 was ∼5.2 Mb with a G + C content of 34.98%. A total of 5,436 genes were predicted with functional assignments. Moreover, 91 RNA genes were detected, consisting of 73 tRNAs, 13 rRNAs, and 5 sRNAs. The ARG were predicted by the ARG databases Antibiotic Resistance Genes Database and Comprehensive Antibiotic Resistance Database. Genes blaA, blaC, or bla1 encoding class-A β-lactamase and pbp2B encoding penicillin-binding protein 2B could confer BA008 resistance to β-lactam. Gene dfrA encoding drug-insensitive dihydrofolate reductase could confer BA008 resistance to trimethoprim. These were consistent with the results of the antimicrobial susceptibility test. The draft genome of B. cereus BA119 was ∼5.4 Mb with a G + C content of 35.29%. A total of 5,630 protein-coding genes, 12 rRNA genes, 95 tRNAs, and 17 genomic islands were predicted in the genome of BA119. The presence of various ARG were observed, including for β-lactam resistance (bla1, mecA, mecC, and pbp2B), trimethoprim resistance (dfrA), and sulfamethoxazole resistance (sul3). The draft genome of B. cereus BA117 was ∼6.5 Mb with a G + C content of 35.44%. We observed 7,078 predicted genes, including 106 tRNAs, 36 rRNAs, and 9 sRNA as well as 28 genomic islands. The ARG prediction indicated β-lactam resistance genes blaA and mecA, tetracycline resistance gene tetM, and tetL were present in B. cereus BA117.
If these ARG are located on mobile genetic elements such as plasmids or integrative conjugative elements For some organism or antimicrobial combinations, the absence or rare occurrence of resistant strains precludes defining any results categories other than "susceptible" (Hindler and Richter, 2016 (ICEs), these genes will be easily transferred into other bacteria via horizontal gene transfer (Huddleston, 2014). Bioinformatics analysis indicated that the β-lactam resistance genes and trimethoprim or sulfamethoxazole resistance genes in BA008, BA117, and BA119 were located on the chromosome DNA. In addition, the tetracycline resistance gene tetM was also on the chromosome DNA of BA117. Moreover, ICE prediction with ICEberg 2.0 database indicated that these genes were not on mobile genetic elements (Liu et al., 2019). It is important to note that gene tetL was located on the 4.6 kb plasmid, which was designated as pBC46-TL. To further confirm the location of gene tetL, plasmid pBC46-TL was separated from other plasmids in B. cereus BA117 and purified by gel extraction (Figure 2A). A partial of tetL was obtained by PCR with purified pBC46-TL as a template, indicating that gene tetL was inside plasmid pBC46-TL ( Figure 2B). The complete sequence of pBC46-TL was obtained by PCR amplification with the primer pair (TL-F: 5′-GCTGGTGCTGGAATGAGTTTG −3′ and TL-R: 5′-TTGGTTGTGTCGTAAATTCG −3′) and DNA sequencing. The sequence analysis showed that the gene encoding the Mob protein and a putative transfer origin sequence oriT were present in pBC46-TL ( Figure 2C), indicating that this plasmid might be transferable via conjugation.

Conjugative Transfer of the tetL Gene by Filter Mating
The conjugative transfer capacities of the plasmid pBC46-TL were determined by filter mating using the tetracycline-resistant strain BA117 as donor. Eight Bacillus strains, 1 Escherichia coli strain, and 1 Lactobacillus paracasei strain were selected as recipients in this study (Table 1). Filter mating assays with BA117 (donor) and BA103 (recipient) revealed an average frequency of conjugation of 1.7 × 10 −5 (Table 4). When comparing the plasmid profile of the transconjugant with the donor strain, we have observed that the transconjugant harbored a small plasmid (∼4.6 kb) of similar size to that of the donor strain ( Figure 3). The tetL in the transconjugant was further confirmed by PCR amplification combined with DNA sequencing (data not shown).
In addition, transfer of pBC46-TL to B. invictae BA142, B. safensis BA143, and B. licheniformis BA130 occurred at a frequency of 1.5 × 10 −7 to 2.0 × 10 −6 transconjugants per donor cells (T/D). However, no transconjugant was detected when the conjugative transfer of pBC46-TL was performed with B. licheniformis BA107, Bacillus spp. BA131, B. pumilus BA102, B. licheniformis BA123, Escherichia coli Rosetta (DE3), and Lactobacillus paracasei sCAUH35-2. These results revealed highly significant associations between transfer frequency and donor or recipient species, which  The transfer frequencies were calculated as the ratio of transconjugants per donor cells (T/D) at the end of the conjugation time. was in accordance with previous studies (Leroy et al., 2019). Altogether, the filter mating experiments confirmed the transferability of tetL on plasmid pBC46-TL across some Bacillus strains, which might lead to the production of a new tetracycline-resistant strain in a natural environment.

CONCLUSIONS
A total of 114 Bacillus strains were isolated from 133 pasteurized milk samples. The antimicrobial susceptibility of these strains indicated that a high percentage of the isolates were intrinsically resistant to ampicillin and penicillin. Meanwhile, some Bacillus isolates displayed acquired resistance, including trimethoprimsulfamethoxazole, clindamycin, erythromycin, and tetracycline. Subsequently, genomic and plasmid DNA sequencing were carried out to determine the ARG in plasmid-containing strains B. cereus BA008, BA117, and BA119. It is important to note that gene tetL encoding tetracycline efflux protein was located on the plasmid pBC46-TL of B. cereus BA117. This plasmid can be transferred into other Bacillus strains by filter mating. Therefore, Bacillus strains in pasteurized milk is not only of concern with regard to the spoilage of the products, but also might be a reservoir of ARG. Although the detection of Bacillus is not currently required for the dairy microorganism test of Chinese food security standards, prudent use of antibiotics, improved hygienic conditions during milking, and removing spores with bactofugation should be taken into consideration to reduce the antibiotic resistance of Bacillus in pasteurized milk.