Molecular characteristics of Escherichia coli from bulk tank milk in Korea.

BACKGROUND: Escherichia coli, which causes subclinical or clinical mastitis in cattle, is responsible for transmitting antimicrobial resistance via human consumption of raw milk or raw milk products.
OBJECTIVES: The objective of this study was to investigate the molecular characteristics of 183 E. coli from bulk tank milk of five different dairy factories in Korea.
METHODS: The molecular characteristics of E. coli such as serogroup, virulence, antimicrobial resistance, and integron genes were detected using polymerase chain reaction and antimicrobial susceptibility were tested using the disk diffusion test.
RESULTS: In the distribution of phylogenetic groups, group D was the most prevalent (59.6%) and followed by group B1 (25.1%). The most predominant serogroup was O173 (15.3%), and a total of 46 different serotypes were detected. The virulence gene found most often was fimH (73.2%), and stx1, fimH, incC, fyuA, and iutA genes were significantly higher in isolates of phylogenetic group B1 compared to phylogenetic groups A, B2, and D (p < 0.05). Among 64 E. coli isolates that showed resistance to at least one antimicrobial, the highest resistance rate was observed for tetracyclines (37.5%). All 18 integron-positive E. coli carried the integron class I (int1) gene, and three different gene cassette arrangements, dfrA12+aadA2 (2 isolates), aac(6')-Ib3+aac(6')-Ib-cr+aadA4 (2 isolates), and dfrA17+aadA5 (1 isolate) were detected.
CONCLUSIONS: These data suggest that the E. coli from bulk tank milk can be an indicator for dissemination of antimicrobial resistance and virulence factors via cross-contamination.

INTRODUCTION

Escherichia coli can be found in animals’ digestive systems as commensal bacteria and are easily observed in the environment [1]. Therefore, E. coli can quickly contaminate the bovine mammary gland and cause clinical mastitis [2]. Although the efficacy of antimicrobial treatment for E. coli mastitis is reported as limited, broad-spectrum antimicrobials are commonly used in farms [3].

Recently, the spreading of antimicrobial-resistant E. coli are recognized more and concerns are increasing in humans and animals [14]. In particular, the continuous use of antimicrobials for the treatment and prevention of bovine mastitis has contributed to the emergence and sustenance of antimicrobial-resistant E. coli by genetic mutation or horizontal gene transfer [5]. Antimicrobial resistance genes are carried on plasmids, transposons, or integrons that can be vectors that transfer these genes to the same or another bacterial species, and gene transfer can potentially be a threat to public health [6].

The pathogenic trait of E. coli isolated from bulk milk can also play an important role in human health risk. E. coli strains are commonly classified into phylogenetic groups A, B1, B2, and D [7]. Groups B2 and D are associated with extraintestinal infections and regarded as more invasive strains [8]. Virulence factors also usually responsible for the ability to infect a host, and exchange of them can happen via horizontal gene transfer. Moreover, virulence factors contributing to iron uptake can provide antimicrobial resistance to host immunological defenses [9]. Although the characteristics of antimicrobial-resistant E. coli from bovine mastitic milk have been described in South Korea [510], those in bulk tanks which can transfer antimicrobial resistance asymptomatically have not been fully described. Therefore, this study aimed to determine the characteristics of E. coli isolated from bulk tank milk to provide important consideration for the management of dairy herds.

MATERIALS AND METHODS

Sample collection

A total of 1160 batches of bunk tank milk were collected from 290 dairy farms of five different dairy factories operated by three companies in Korea. Each 50 mL milk sample was aseptically collected twice, in summer (July 2019) and winter (December 2019), and sent to the laboratory under 4°C conditions.

Bacterial identification

The isolation and identification of E. coli were performed following the standard microbiological protocols published by the Ministry of Food and Drug Safety (2018) [11]. Briefly, 1 mL of each milk sample was inoculated in 9 mL mEC broth medium (Merck, Darmstadt, Germany). After incubation at 37°C for 24 h, a loopful of medium (~10 μL) was streaked onto MacConkey agar (BD Biosciences, NJ, USA) and incubated at 37°C for 24 h. The confirmation of E. coli was performed using polymerase chain reaction (PCR) to detect malB, as described previously [12]. If two isolates of the same origin showed the same antimicrobial resistance patterns, only one isolate was randomly chosen and included in this study.

Phylogenetic groups and serogrouping

All 183 E. coli isolates included in this study were categorized by the phylogenetic group and serogrouping with PCR-based typing, as described previously by Clermont et al. [7] and Iguchi et al. [13], respectively.

Antimicrobial susceptibility testing

Based on the Clinical and Laboratory Standards Institute guidelines [14], all E. coli isolates were investigated for antimicrobial resistance using the disk diffusion test with the following disks (BD Biosciences): ampicillin (AM; 10 μg), amoxicillin-clavulanate (AMC; 20 μg), chloramphenicol (C; 30 μg), ceftazidime (CAZ; 30 μg), cefadroxil (CDX; 30 μg), cephalothin (CF; 30 μg), ciprofloxacin (CIP; 5 μg), colistin (CL; 10 μg), cefotaxime (CTX; 30 μg), cefuroxime (CXM; 30 μg), cefazoline (CZ; 30 μg), cefepime (FEP; 30 μg), cefoxitin (FOX; 30 μg), gentamicin (GM; 10 μg), imipenem (IPM; 10 μg), nalidixic acid (NA; 30 μg), trimethoprim/sulfamethoxazole (SXT; 1.25 μg), and tetracycline (TE; 30 μg). E. coli ATCC 25922 was used as the quality control. Multidrug resistance (MDR) was defined as acquired resistance to at least one agent in three or more antimicrobial classes.

Detection of virulence, antimicrobial resistance, and integron genes

The detection of virulence, antimicrobial resistance, and integron genes was performed using PCR with the primers listed in Supplementary Table 1. The target virulence genes included in this study are eaeA, stx1, stx2, hly, fimH, iucC, fyuA, and iutA. The antimicrobial resistance determinants were genes conferring resistance to β-lactamase (bla, bla, bla, and bla), aminoglycosides [aac(6′)-Ib, aac(3)-II, and ant(2″)-I], tetracyclines (tetA, tetB, tetC, tetD, tetE, and tetG), sulfonamides (sul1 and sul2), and chloramphenicol (catA1 and cml1). In particular, β-lactamase and integron cassette gene amplicons were sequenced with an automatic sequencer (Cosmogenetech, Seoul, Korea) and compared to those in GenBank using the Basic Local Alignment Search Tool program available at the National Center for Biotechnology Information website (www.ncbi.nlm.nih.gov/BLAST) and Resfinder 4.0 (https://cge.cbs.dtu.dk/services/ResFinder/).

Statistical analysis

Statistical analyses were performed using SPSS version 25 (IBM Corp., USA). The prevalence of antimicrobial resistance and the distribution of virulence and antimicrobial genes of isolates by factories were compared using χ2 test. When the expected count for χ2 test is < 5, which can lead to inaccurate interpretation, Fisher's exact test was used as an alternative method, and p < 0.05 was considered statically significant [15].

RESULTS

Distribution of phylogenetic groups, serogroups, and virulence genes of E. coli isolates

The distribution of phylogenetic groups and virulence genes of 183 E. coli isolates are shown in Table 1. Phylogenetic group D was the most prevalent (109 isolates; 59.6%), followed by group B1 (46 isolates; 25.1%), group A (23 isolates; 12.6%), and group B2 (5 isolates; 2.7%). One hundred fifty-seven (85.8%) isolates were classified into 46 different serotypes, with 26 isolates remaining ungrouped. The most predominant serogroup was O173 (28 isolates; 15.3%), which all belong to phylogenetic group D. The virulence gene found most often was fimH (134 isolates; 73.2%), followed by eaeA (37 isolates; 20.2%), iucC (26 isolates; 14.2%), fyuA (9 isolates; 4.9%), iutA (6 isolates; 3.3%), and stx1 (3 isolates; 1.6%). However, fimH was the most prevalent virulence gene in phylogenetic groups B1, B2, and D. In phylogenetic group A, only 1 (4.3%) of 23 isolates carried fimH. Interestingly, stx1, fimH, incC, fyuA, and iutA genes were significantly higher in isolates of phylogenetic group B1 than in phylogenetic groups A, B2, and D (p < 0.05). All isolates did not carry the stx2 and hly genes.

Table 1

Distribution of phylogenetic groups and virulence genes of 183 E. coli isolates from bulk tank milk of five dairy factories

Phylogenetic group	No. (%) of isolates	No. (%) of isolates with each virulence gene								O Serotypes (No. of isolates)
		eaeA	stx1	stx2	hly	fimH	iucC	fyuA	iutA
Group A	23 (12.6)	0 (0)a	0 (0)a,b	0 (0)	0 (0)	1 (4.3)a	0 (0)a	0 (0)a,b	0 (0)a,b	O3 (1), O6 (1), O10 (1), O29 (1), O60 (1), O86 (1), O88 (5), O98 (2), O109 (4), O117 (1), O127 (1), O156 (1), O158 (1), O165 (1), O179 (1)
Group B1	46 (25.1)	7 (15.2)a,b	3 (6.5)b	0 (0)	0 (0)	44 (95.7)b	19 (41.3)b	9 (19.6)b	6 (13.0)b	O3 (1), O10 (2), O11 (3), O15 (2), O26 (1), O39 (1), O102 (3), O109 (3), O112 (1), O117 (1), O142 (1), O147 (1), O158 (3), O170 (1), O174 (2), O177 (1), Ogp8 (1)*, ONT (18)†
Group B2	5 (2.7)	0 (0)a,b	0 (0)a,b	0 (0)	0 (0)	4 (80.0)b,c	1 (20.0)a,b	0 (0)a,b	0 (0)a,b	O44 (2), O99 (2), O126 (1)
Group D	109 (59.6)	30 (27.5)b	0 (0)a	0 (0)	0 (0)	85 (78.0)c	6 (5.5)a	0 (0)a	0 (0)a	O3 (4), O8 (1), O10 (3), O11 (2), O15 (12), O25 (1), O28 (1), O38 (1), O51 (3), O52 (1), O55 (1), O60 (1), O74 (1), O76 (2), O80 (1), O88 (13), O95 (1), O96 (6), O98 (4), O100 (1), O102 (1), O109 (4), O112 (1), O125 (1), O137 (1), O153 (3), O158 (2), O173 (28), ONT (8)
Total	183	37 (20.2)	3 (1.6)	0 (0)	0 (0)	134 (73.2)	26 (14.2)	9 (4.9)	6 (3.3)

Values within a column not having the same superscript letter differ significantly (p < 0.05).

*O107 or O117; †Not determined.

Prevalence of antimicrobial-resistant E. coli isolates

The prevalence of antimicrobial-resistant E. coli isolates by factories is shown in Table 2. Among 64 E. coli isolates that showed resistance to at least one antimicrobial, the highest resistance rate was observed in TE (37.5%), followed by CF (35.9%), AM (34.4%), GM (26.6%), CL (21.9%), C (14.1%), and CZ (10.9%). None of the isolates were resistant to NA, IPM, and CIP. In particular, the prevalence of resistance to AM, TE, SXT, GM, C, CL, and CF showed significant differences among the five dairy factories (p < 0.05, χ2 test). None of the 14 isolates from factory C-1 showed resistance to AM and GM. In contrast, all six antimicrobial-resistant isolates from factory C-2 showed resistance to AM and GM. Seven isolates (53.8%) from factory B-1 showed high resistance to C. In contrast, no isolates from factories A-1, C-2, and C-3 were resistant to C. Isolates from factories C-2 and C-3 did not show resistance to CF, but isolates from other factories showed relatively high resistance from 38.5% to 52.4%. Isolates from factories B-1, C-2, and C-3 also showed significantly higher resistance to TE than factories A-1 and C-1. The prevalence of antimicrobial-resistant E. coli isolates by the phylogenetic group was analyzed, but there were no significant differences among groups (data not shown).

Table 2

Distribution of 64 antimicrobial-resistant E. coli isolates from bulk tank milk of five dairy factories

Variables	No. (%) antimicrobial-resistant isolates					Total
	Factory A-1	Factory B-1	Factory C-1	Factory C-2	Factory C-3
No. of isolates	40	41	36	29	37	183
No. (%) of antimicrobial-resistant isolates	21 (52.5)	13 (31.7)	15 (41.7)	6 (20.7)	9 (24.3)	64 (35.0)
AM	4 (19.0)a,b,c	6 (46.2)c,d	0 (0)a	6 (100)d	6 (66.7)b,c,d	22 (34.4)
AMC	2 (9.5)	1 (7.7)	1 (6.7)	1 (16.7)	0 (0)	5 (7.8)
TE	4 (19.0)a	8 (61.5)b	2 (13.3)a	4 (66.7)b	6 (66.7)b	24 (37.5)
SXT	1 (4.8)a,b	4 (30.8)a	0 (0)b	0 (0)a,b	0 (0)a,b	5 (7.8)
Na	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
IPM	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
CIP	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
GM	1 (4.8)a	4 (30.8)b	0 (0)a	6 (100)c	6 (66.7)b,c	17 (26.6)
C	0 (0)a	7 (53.8)b	2 (13.3)a	0 (0)a	0 (0)a	9 (14.1)
CL	6 (28.6)a,b	0 (0)a	5 (33.3)b	0 (0)a,b	3 (33.3)a,b	14 (21.9)
CZ	2 (9.5)	4 (30.8)	1 (6.7)	0 (0)	0 (0)	7 (10.9)
CF	11 (52.4)a	5 (38.5)a,b	7 (46.7)a,b	0 (0)b,c	0 (0)c	23 (35.9)
CDX	2 (9.5)	0 (0)	1 (6.7)	0 (0)	0 (0)	3 (4.7)
FOX	1 (4.8)	0 (0)	1 (6.7)	0 (0)	0 (0)	2 (3.1)
CXM	1 (4.8)	0 (0)	0 (0)	0 (0)	0 (0)	1 (1.6)
CTX	1 (4.8)	0 (0)	0 (0)	0 (0)	0 (0)	1 (1.6)
CAZ	0 (0)	0 (0)	3 (20.0)	0 (0)	0 (0)	3 (4.7)
FEP	1 (4.8)	0 (0)	0 (0)	0 (0)	0 (0)	1 (1.6)

Values within a column not having the same superscript letter differ significantly (p < 0.05).

AM, ampicillin; AMC, Amoxicillin-clavulanate; TE, Tetracycline; SXT, Trimethoprim sulfamethoxazole; Na, Nalidixic acid; IPM, Imipenem; CIP, Ciprofloxacin; GM, Gentamicin; C, Chloramphenicol; CL, Colistin; CZ, Cefazolin; CF, Cephalothin; CDX, Cefadroxil; FOX, Cefoxitin; CXM, Cefuroxime; CTX, Cefotaxime; CAZ, Ceftazidime; FEP, Cefepime.

Distribution of MDR E. coli isolates

The distribution of MDR E. coli isolates by factories is shown in Table 3. Fifteen isolates (23.4%) of the 64 antimicrobial-resistant E. coli were categorized as MDR. In particular, the number of MDR isolates from factory C-3 (6 isolates; 66.7%) was significantly higher than that in factories A-1 (2 isolates; 9.5%) and C-1 (0%; p < 0.05). However, the MDR against five and six antimicrobial classes was higher in isolates (15.4% each) from factory B-1, although there was no significant difference among factories.

Table 3

Distribution of multidrug resistance of 64 antimicrobial-resistant E. coli isolates

No. of antimicrobial-resistant classes	No. of isolates (%)
	Factory A-1 (n = 21)	Factory B-1 (n = 13)	Factory C-1 (n = 15)	Factory C-2 (n = 6)	Factory C-3 (n = 9)	Total (n = 64)
1	13 (61.9)	4 (30.8)	8 (53.3)	1 (16.7)	3 (33.3)	29 (45.3)
2	8 (38.1)a	4 (30.8)a,b	7 (46.7)a	3 (50.0)a	0 (0)b	22 (34.4)
3	1 (4.8)a	1 (7.7)a,b	0 (0)a	2 (33.3)b,c	6 (66.7)c	10 (15.6)
4	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
5	1 (4.8)	2 (15.4)	0 (0)	0 (0)	0 (0)	3 (4.7)
6	0 (0)	2 (15.4)	0 (0)	0 (0)	0 (0)	2 (3.1)
No. (%) of MDR*	2 (9.5)a	5 (38.5)b	0 (0)a	2 (33.3)b	6 (66.7)b	15 (23.4)

Values within a column not having the same superscript letter differ significantly (p < 0.05).

*Multidrug resistance was defined as the acquired resistance to at least one agent in three or more antimicrobial classes.

Distribution of antimicrobial resistance genes of E. coli isolates

The distribution of antimicrobial resistance genes of 64 antimicrobial-resistant E. coli is shown in Table 4. In the distribution of β-lactamase genes, the bla gene alone and both bla and bla genes were detected in 4 (6.3%) and 19 (29.7%) isolates, respectively. In particular, the isolates from factories B-1, C-2, and C-3 showed a significantly high prevalence of bla and bla genes compared to that from factories A-1 and C-1 (p < 0.05). The other β-lactamase genes, such as bla and bla, were not detected in all resistant isolates. In the antimicrobial resistance genes encoding for aminoglycoside-modifying enzymes, although both aac(6′)-Ib and aac(3)-II genes were detected in isolates from factories A-1 and B-1, the distribution of single aac(3)-II gene showed significantly high isolates from factory C-3 compared to other factories (p < 0.05). Among TE resistance genes, the tetB gene (18 isolates; 28.1%) was predominant and showed significantly high isolates from factories C-2 and C-3 compared to that from factories A-1 and C-1 (p < 0.05).

Table 4

Distribution of antimicrobial resistance genes of 64 antimicrobial-resistant E. coli isolates from bulk tank milk

Antimicrobial class	Antimicrobial resistance gene	No. (%) of isolates that carried the target gene
		Factory A-1 (n = 21)	Factory B-1 (n = 13)	Factory C-1 (n = 15)	Factory C-2 (n = 6)	Factory C-3 (n = 9)	Total (n = 64)
β-lactamases	blaOXA	3 (14.3)	0 (0)	1 (6.7)	0 (0)	0 (0)	4 (6.3)
	blaCTX-M	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	blaSHV	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	blaTEM + blaOXA	1 (4.8)a	6 (46.2)b	0 (0)a	6 (100)b	6 (66.7)b	19 (29.7)
Aminoglycoside-modifying enzymes	aac(3)-II	0 (0)a	1 (7.7)a	0 (0)a	0 (0)a,b	6 (66.7)b	7 (10.9)
	ant(2′′)-I	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	aac(6′)-Ib + aac(3)-II	1 (4.8)	3 (23.1)	0 (0)	0 (0)	0 (0)	4 (6.3)
Tetracycline	tetA	1 (4.8)	3 (23.1)	0 (0)	0 (0)	0 (0)	4 (6.3)
	tetB	2 (9.5)a	4 (30.8)a,b	1 (6.7)a	4 (66.7)b	6 (66.7)b	17 (26.6)
	tetC	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	tetD	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	tetE	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	tetB + tetG	0 (0)	0 (0)	1 (6.7)	0 (0)	0 (0)	1 (1.6)
Sulfonamide	sul1	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
	sul2	0 (0)	2 (15.4)	0 (0)	0 (0)	0 (0)	2 (3.1)
	sul1 + sul2	1 (4.8)	2 (15.4)	0 (0)	0 (0)	0 (0)	3 (4.7)
Chloramphenicol	catA1	0 (0)	2 (15.4)	0 (0)	0 (0)	0 (0)	2 (3.1)
	cml1	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)

Values within a column not having the same superscript letter differ significantly (p < 0.05).

Characterization of integron-positive E. coli isolates

The characterization of 18 integron-positive isolates (28.1%) among 64 antimicrobial-resistant E. coli are shown in Table 5. All integron-positive E. coli carried the integron class I gene (int1). Among them, five isolates showed the three different gene cassette arrangements, dfrA12 + aadA2 (2 isolates), aac(6′)-Ib3 + aac(6′)-Ib-cr + aadA4 (2 isolates), and dfrA17 + aadA5 (1 isolate). Eleven (61.1%) of 18 isolates were also categorized as MDR. No isolates were positive for the integron class II gene (int2).

Table 5

Characterization of the 18 integron-positive E. coli isolates from bulk tank milk

Factory	Isolates	Integron gene	Cassette array for class I integrons	Other antimicrobial resistance genes	Antimicrobial resistance profiles
A-1	19-A1-025-1	int1	-	-	CF, CL
A-1	19-A1-025-2	int1	-	-	CF
A-1	19-A1-026-1	int1	-	-	CF
A-1	19-A1-027-2	int1	-	-	CF, CL
A-1	19-A1-041-1	int1	dfrA12, aadA2	blaOXA , sul1, sul2, aac(6′)-Ib, aac(3)-II	CF, AM, AMC, SXT, GM
B-1	19-B1-002-2	int1	-	blaTEM , blaOXA	AM, C
B-1	19-B1-068-1	int1	dfrA17, aadA5	blaTEM , blaOXA , sul1, sul2, catA1, aac(6′)-Ib, aac(3)-II, tetB	AM, TE, SXT, GM, C
B-1	19-B1-107-1	int1	aac(6′)-Ib3, aac(6′)-Ib-cr, aadA4	blaTEM , blaOXA , sul2, aac(3)-II, tetA	CZ, CF, AM, TE, SXT, GM, C
B-1	19-B1-107-2	int1	aac(6′)-Ib3, aac(6′)-Ib-cr, aadA4	blaTEM , blaOXA , sul2, aac(3)-II, tetA	CZ, AM, TE, SXT, GM, C
B-1	19-B1-110-1	int1	-	tetB	CZ, CF, TE
B-1	19-B1-115-1	int1	dfrA12, aadA2	blaTEM , blaOXA , sul1, sul2, aac(3)-II	AM, SXT, GM
C-1	19-C1-008-2	int1	-	-	CF, C
C-1	19-C1-016-2	int1	-	blaOXA	CZ, CF, FOX, AMC, CDX
C-1	19-C1-043-1	int1	-	-	CF
C-2	19-C2-061-1	int1	-	blaTEM , blaOXA , aac(3)-II, tetB	AM, TE, GM
C-2	19-C2-071-1	int1	-	blaTEM , blaOXA , aac(3)-II, tetB	AM, TE, GM
C-3	19-C3-098-1	int1	-	blaTEM , blaOXA , aac(3)-II, tetB	AM, TE, GM
C-3	19-C3-119-1	int1	-	blaTEM , blaOXA , aac(3)-II, tetB	AM, TE, GM

AM, ampicillin; AMC, amoxicillin-clavulanate; TE, tetracycline; SXT, trimethoprim sulfamethoxazole; GM, gentamicin; C, chloramphenicol; CL, colistin; CZ, cefazolin; CF, cephalothin; CDX, cefadroxil; FOX, cefoxitin.

DISCUSSION

Although contamination of E. coli in raw milk is at low level, its transmission via human consumption, such as raw milk or raw milk products, could be a source of both enteric disease and antimicrobial resistance [1]. In this study, the characteristics of 183 E. coli isolates from bulk tank milk of five different dairy factories were investigated. The phylogenetic characterization of E. coli is an important tool to improve the understanding of the relationship between strains and disease [16]. E. coli is commonly classified into four major phylogenetic groups A, B1, B2, and D [81718]. Group D was the most prevalent (59.6%), followed by B1 (25.1%). Previous studies have reported that the intestinal pathogenic E. coli mostly belong to groups A, B1, and D [17]. In contrast, the most virulent extraintestinal pathogenic strains belong to phylogenetic groups B2 and D [19]. In Korea, the study of the phylogenetic group of E. coli from milk samples was only reported by Tark et al. [5]. E. coli isolates from bovine mastitic milk were mainly assigned to phylogenetic group A (60.0%), which was different from the results of this study. However, the distribution of phylogenetic group D in E. coli isolates from dairy products was continuously reported in Korea (20.0%; Tark et al. [5]), China (21.7%; Ali et al. [20]), and Iran (28.6%; Jamali et al. [21]). Moreover, E. coli isolates from bovine mastitis in China were mainly categorized as phylogenetic group B1 (58.6%; Liu et al. [22]). E. coli isolates from a dairy farm in the United States were also predominantly phylogenetic group B1 (63.9%; Son et al. [23]). The correlation between the phylogenetic group and pathogenic trait of E. coli in bovine mastitis should be further investigated.

In the study, the mannose-specific adhesion type 1 fimbriae gene, fimH, among eight virulence profiles (eaeA, stx1, stx2, hly, fimH, iucC, fyuA, and iutA genes), was the most prevalent virulence gene (73.2%). Other studies also reported that fimH was the most widespread virulence gene of E. coli isolates from the subclinical mastitis milk samples (93%) in Egypt [24] and clinical mastitis milk samples (84.2%) in Taiwan [25]. The fimH gene was associated with the initial colonization of tissues, and uropathogenic E. coli strains were reported to express genes by more than 90% [24]. In contrast, only one isolate (4.3%) from phylogenetic group A, considered commensal, carried the fimH gene in this study. Intimin plays a role in the attachment of bacterial cells to host cells [26]. In this study, 37 (20.2%) isolates carried the intimin gene, eaeA, and isolates in phylogenetic group D were significantly higher than in group A (p < 0.05). However, stx1, fimH, incC, fyuA, and iutA genes were significantly higher in isolates of phylogenetic group B1 than that of phylogenetic groups A, B2, and D (p < 0.05). Shiga toxin, encoded by stx1 and stx2, is one of the most potent virulence factors in E. coli and is associated with hemolytic uremic syndrome and neurological disorder [1727]. Aerobactin synthesis genes, iucC and iutA, are conserved in virulent strains [28]. The yersiniabactin receptor gene, fyuA, is correlated with other virulence factors contributing to urinary tract infection [29]. Although Ombarak et al. [17] also reported that E. coli isolates belonged to phylogenetic group B1 in raw milk showing a high prevalence of virulence genes, E. coli isolates were mostly classified into phylogenetic group B1 (63.5%). Only 3.8% belonged to phylogenetic group D. Group D was the most prevalent, but E. coli belonging to group B1 carried more virulence genes. Therefore, further studies about the link between the virulence of a strain and its phylogenetic group are needed to better understand the disease pathogenesis of E. coli isolated from dairy products.

Previous studies reported various serotypes in E. coli isolates from bulk tank milk or raw milk [303132]. In this study, 183 E. coli isolates were classified into 46 serotypes, and the most prevalent type was O173 (28 isolates; 15.3%). Although Linnerborg et al. [33] reported that O173 is grouped as enteroinvasive E. coli, the virulence of O173 was rarely reported. The O173 isolates in this study did not showed any typical virulence factors and various phylogenetic groups and diverse distribution of virulence genes were observed.

In this study, 64 E. coli showed resistance to at least one antimicrobial, and the highest resistance rate was observed for TE (37.5%), followed by CF (35.9%) and AM (34.4%). Tark et al. [5] also reported that the most prevalent antimicrobial resistance in E. coli isolates from bovine mastitic milk in Korea was TE (23.3%), followed by AM (16.6%). Moreover, the distribution of antimicrobial resistance of E. coli and MDR also showed significant differences among the factories in this study. Isolates from factories B-1, C-2, and C-3 also showed significantly higher resistance to TE than that of A-1 and C-1. Also, isolates from factory C-1 did not show resistance to AM and GM, but all isolates from factory C-2 were resistant to AM and GM. Based on the antimicrobial consumption report in Korea, AM and TE are widely used antimicrobial agents for cattle in Korea [34]. However, the prevalence of antimicrobial resistance among factories showed significant differences because dairy factories allow producers to combine different biosecurity and sanitation practices, housing technologies, and feeding regimens, including the use of antimicrobials. In addition, isolates from factory B-1 showed high resistance to C, although C has been withdrawn from veterinary use in Korea in 1992 [10]. It may be because florfenicol can give resistance to C, as they have partially overlapped drug-binding sites [35].

The most prevalent antimicrobial resistance genes against β-lactamase class were bla + bla (29.7%), with a significant difference among factories (p < 0.05). Although bla, one of the most commonly detected β-lactamase genes [520], was not revealed in this study, the TEM gene has also been identified in clinical E. coli isolates from foods, humans, and healthy animals [363738]. β-Lactamase-producing E. coli are resistant to most β-lactams and other antimicrobials, such as aminoglycosides, C, quinolones, sulfonamides, tetracyclines, and trimethoprim [6]. Therefore, monitoring of β-lactamase-producing E. coli by dairy factories is crucial to prevent the emergence and dissemination of MDR strains [538].

Among antimicrobial genes for tetracyclines, tetB (26.6%) was the most prevalent gene, followed by tetA (4 isolates, 6.3%) in this study. In particular, the tetB gene’s distribution was significantly different among factories (p < 0.05). It is analogous with the findings of a study in Canada, which showed tetB (31.0%) and tetA (28.6%) as the most prevalent antimicrobial resistance genes of E. coli isolates from milk samples [21]. Abdus Sobur et al. [12] also reported that E. coli isolates from dairy farms and its environment carried tetA (80.51%) as the most prevalent antimicrobial resistance gene in Bangladesh.

In this study, 18 integron-positive isolates (28.1%) among 64 antimicrobial-resistant E. coli carried the int1 gene, and 5 integron-positive isolates harbored aadA and/or dfrA gene cassettes. Moreover, 11 (61.1%) of 18 int1-positive isolates showed MDR. Previous studies reported that the int1 gene is important for the transfer of resistance genes, and the aadA and dfrA genes, which are resistant to GM and SXT, respectively, are the most frequent gene cassettes in humans and animals in Korea [39]. In particular, Seo et al. [39] reported that isolates harboring aadA or dfrA showed higher antimicrobial resistance. Integrons have been continuously considered responsible for the emergence of MDR by enhancing gene expression and acquiring new gene cassettes [640].

The occurrence of antimicrobial-resistant bacterial infection can be problematic not only for the deterioration of milk quality management but also for limiting antimicrobial therapeutic choices in humans [21]. Although the characteristics of E. coli from bovine mastitic milk have been reported in the world [25], the results suggest that E. coli from bulk tank milk, which is not mastitic milk, can also become a reservoir to disseminate antimicrobial resistance and virulence factors via cross-contamination.