Sources of E. Coli isolates and detection of the class I integrase gene intI1

Of the 721 E. coli isolates examined in this study, 113 were from human patients from Sir Run Shaw Hospital and 608 were from food-producing animals raised on farms across Zhejiang Province (298 from pig anal swabs and 310 from poultry anal swabs) (Fig. 1).

Overall, 93 (12.90%) of the 721 E. coli isolates were positive for the class I integrase gene intI1. The prevalence of the intI1-positive isolates was 17.70% (20/113) in hospitalized patients, 17.45% (52/298) in pig samples, and only 6.77% (21/310) in poultry samples. Notably, intI1 was considerably more common in pig samples and hospitalized patients than poultry samples (p < 0.05).

Antimicrobial susceptibility of intI1-positive and -negative E. Coli isolates

The AMR profiles of 93 intI1-positive and 628 intI1-negative E. coli isolates are depicted in Fig. 2. Of the 93 intI1-positive E. coli isolates, 88 (94.62%), 82 (88.17%), 82 (88.17%), and 60 (64.52%) were resistant to SIZ, TMP, TET, and SM, respectively (Fig. 2A). In addition, 90 (96.77%) of the intI1-positive isolates were resistant to at least one antibiotic, with 82 (88.17%) categorized as MDR (Fig. 2B). Furthermore, 38 (40.86%) and 28 (30.11%) of the 93 intI1-positive isolates were resistant to three and four classes of antibiotics, respectively. A single intI1-positive E. coli isolate (1.08%) exhibited resistance across all seven antibiotic classes. Overall, there were 16 distinct AMR patterns, with resistance to KAN-CPL-TMP-SIZ emerging as the most prevalent at 33.33% (31/93) (Table S1).

Fig. 2

AMR profiles of intI1-positive and -negative E. coli isolates in this study. (A) AMR rates of 93 intI1-positive E. coli isolates. (B) The distribution of MDR strains among 93 intI1-positive E. coli isolates. (C) AMR rate of 628 intI1-negative E. coli isolates. (D) The distribution of MDR strains among 628 intI1- negative E. coli isolates. Kanamycin, KAN; streptomycin, SM; neomycin, NEO; chloramphenicol, CPL; florfenicol, FLO; ampicillin, AMP; meropenem, MRP; enrofloxacin, ENR; ofloxacin, OFX; polymyxin B, PB; sulfisoxazole, SIZ; tetracycline, TET; trimethoprim, TMP

Of the 628 intI1-negative E. coli isolates, 299 (47.61%), 276 (43.95%), 237 (37.74%), and 169 (26.91%) were resistant to CPL, AMP, SM, and TET, respectively (Fig. 2C). In addition, 471 (75.00%) of the 628 intI1-negative E. coli isolates exhibited resistance to at least one antibiotic class, with 241 (38.38%) categorized as MDR (Fig. 2D). Among these, 157 (25.00%), 127 (20.22%), and 127 (20.22%) were susceptible, one, and three classes of antibiotics, respectively. In addition, six intI1-negative E. coli isolates (0.96%) were resistant to all seven antibiotic classes. In total, 31 different AMR patterns were identified among the intI1-negative E. coli isolates, with CPL-AMP as the most common AMR pattern (9.39%, 59/628) (Table S2).

The MDR pattern was correlated with the presence of class I integrons (Table 1). The resistance rates to SM, CPL, FLO, TET, TMP and SIZ were significantly higher for isolates harboring class I integrons (p < 0.01).Also, the MDR rate was significantly higher for isolates containing class I integrons (p < 0.01). Meanwhile, there were no significant differences in the rates of resistance to KAN, NEO, AMP, MRP, ENR, OFX, and PB between isolates with and without class I integrons (p > 0.05).

Table 1 Association between antimicrobial resistance phenotypes of intI1 positive or negative strains in 721 E. Coli isolatesAMR gene patterns of intI1-positive E. Coli isolates

A diverse array of AMR genes was identified among the 93 intI1-positive E. coli isolates (Fig. 3A), which included genes conferring resistance to β-lactams (blaCTX−M, blaDHA−1, blaOXA, blaTEM), sulfonamides (sul1, sul2, sul3), tetracycline [tet(A)], aminoglycosides [aac(3)-VIa, aac(3’)-Ib, aac(6’)-Ib, aadA1, aadA2, aadA5, aadA8, aph(4)-Ia, aph(6)-Id], trimethoprim (dfrA1, dfrA5, dfrA7, dfrA12, dfrA17), fluoroquinolones (qnrB4, qnrS1, qnrS2), fosfomycins (fosA, fosA3), lipopeptides (mcr-1.1, mcr-9.1), macrolides [mph(A)], rifamycin (arr), and chloramphenicol (floR).

Fig. 3

ARG patterns of intI1-positive E. coli isolates. (A) Distribution of acquired ARGs. The red and yellow colors indicate the existence and absence of ARGs, respectively. (B) Different classes of ARGs acquired by intI1-positive E. coli isolates

Prevalent resistance genes carried by the isolates included those conferring resistance to sulfonamides (sul1 [41.94%], sul2 [44.09%], sul3 [13.98%]), trimethoprim (dfrA1 [11.83%], dfrA5 [2.15%], dfrA7 [9.68%], dfrA12 [19.35%], dfrA17 [29.03%]), aminoglycosides (aac(3)-VIa [10.75%], aac(3’)-Ib [6.45%], aac(6’)-Ib [30.11%], aadA1 [24.73%], aadA2 [17.20%], aadA5 [30.11%], aadA8 [2.15%], aph(4)-Ia [8.60%], aph(6)-Id [34.41%]), and β-lactams (blaCTX−M [21.51%], blaDHA−1 [4.30%], blaOXA [6.45%], blaTEM [53.76%]) (Fig. 3B).

MLST analysis of intI1-positive E. Coli isolates

All 93 intI1-positive isolates were sequenced and subjected to MLST analysis. The draft genome length of these isolates was 4.46–5.37 Mb. In total, 39 sequence types (STs) with three (3/39) unknown STs were observed from all 93 isolates. Three novel ST profiles of seven E. coli isolates were identified with the EnteroBase online resource for analysis and visualization of genomic variation within enteric bacteria (https://enterobase.warwick.ac.uk/species/ecoli/allele_st_search), which included ST237112 (traced back to four distinct E. coli isolates derived from pig samples in Hangzhou), ST237113 (associated with two E. coli isolates from pig samples in Lishui), and ST237114 (linked to a single E. coli isolate from a pig sample collected in Hangzhou).

Of the 93 intI1-positive isolates, ST10 (8.60%, 8/93) emerged as the most predominant ST. In addition, ST349, ST101 and ST1196, were each identified in 6.45% (6/93), 5.38% (5/93) and 5.38% (5/93) of the isolates. A detailed analysis of the 20 intI1-positive isolates derived from hospitalized patients revealed ST1196 and ST131 as the most common STs, each accounting for 20.00% (4/20). Meanwhile, ST349 was identified in six (11.32%) of the 53 intI1-positive E. coli isolates obtained from pig samples. Of the 20 intI1-positive isolates from poultry samples, ST10 was identified in five (25.00%). Notably, ST10 was present in isolates from hospitalized patients, as well as pig and poultry samples. ST101, ST156, ST165, ST457 and ST7508 were identified in both pig and poultry samples. Furthermore, ST1196 was common for pig samples and hospitalized patients, while ST744 was identified in both poultry samples and hospitalized patients. In summary, five distinct STs were observed across two or three different sources (Fig. 4).

Fig. 4

Prevalence of 93 intI1-positive E. coli isolates in this study. Sankey diagram combining the cities, sampling sources, and STs based on 93 intI1-positive E. coli isolates. The diameter of the line is proportional to the number of strains, which is also labeled with a number

Characterization of class I integrons in E. Coli isolates

Of the 93 intI1-positive E. coli isolates, 59 (63.44%) harbored the classic class I integron, characterized by the intI1 gene in the 5’CS and the qacEΔ1 + sul1 genes in the 3’CS. The qacEΔ1 gene confers resistance to quaternary ammonium compounds, while the sul1 gene confers resistance to sulfonamides. Overall, 33 isolates from pig samples, 17 from hospitalized patients, and nine from poultry samples harbored the classic class I integron genetic structure. Sources, STs, and arrangement of AMR GCs among the 59 E. coli isolates carrying classic class I integrons are listed in Tables 2 and Fig. S1. In total, six distinct AMR GCs were identified, with dfrA17-aadA5 as the most prevalent at 33.40% (20/59), followed by dfrA12-aadA2 (27.11%, 16/59), dfrA1-aadA1 (22.03%, 13/59), dfrA7 (8.47%, 5/59), aac(6’)-Ib (5.08%, 3/59), and aadA1-aac(3)-VIa (3.39%, 2/59). Remarkably, all 59 isolates with AMR GCs exhibited MDR. As shown in Table S3, 34 non-classic class I integrons lacked either GCs or the 3’CS region. Of these, 11 possessed only the intI1 gene, five had GCs conferring resistance to trimethoprim, and 18 contained GCs associated with aminoglycoside resistance.

Table 2 Sources and sequence types of 59 classic class I integron-carrying E. Coli isolates and the arrangement of gene cassettesGenetic environment of classic class I integrons

Characterization of the genomic contexts of 59 classic class I integrons revealed the presence of six insertion sequences (IS1, IS6, IS21, IS91, IS110, and IS256) and one transposon (Tn3) adjacent to the integrons (Table S4). A comprehensive analysis to elucidate potential transmission pathways of these integrons both in the context of this study and globally found that all 20 integron sequences associated with the predominant GCs (dfrA17-aadA5) identified in this study were bordered by IS6 elements. A comparative sequence analysis using the NCBI database was conducted focusing on the “intI1-dfrA17-aadA5-qacEΔ1-sul1” class I integron sequence. Four E. coli isolates exhibiting the highest sequence homology were selected for in-depth comparisons. Notably, strains containing this specific integron sequence were documented in poultry (this study, 2022), an unspecified animal in Japan (2016), wastewater in Switzerland (2021), a Chinese hospital, and a human subject in the USA (both 2019). Each of these isolates harbored integrons flanked by IS6 elements (Fig. 5A). Furthermore, integron sequences with the second most common gene cassettes (dfrA1-aadA1) identified in this study was associated with IS1, IS6, IS21, or Tn3 elements. Similar sequences from the NCBI database were also affiliated with IS6100, Tn21, and TniB elements (Fig. 5B). These integron sequences were identified in various environments, including poultry, humans, and farms, and were geographically dispersed, with instances recorded in China, France, and the UK over different periods.

Fig. 5

Genetic environment of class I integrons with the most frequent GCs in the genomes of E. coli isolates. (A) Genetic environment of class I integrons with the GC array dfrA17-aadA5. (B) Genetic environment of class I integrons with the GC array dfrA1-aadA1. Arrows indicate the direction of transcription. Regions of > 90% homology are shaded in gray. Gene families are differentiated by different colors. * means missing the C-terminus

As shown in Fig. S2, further analysis was conducted on other frequently observed integron sequences, namely “IS6-intI1-dfrA12-aadA2-qacEΔ1-sul1-IS91” (Fig. S2A), “Tn3-intI1-dfrA7-qacEΔ1-sul1-IS21” (Fig. S2B), “IS6-intI1-aac(6’)-Ib-qacEΔ1-sul1-IS6” (Fig. S2C), and “Tn3-intI1-aadA1-aac(3)-VIa-IS91-IS256-qacEΔ1-sul1-IS110” (Fig. S2D). These sequences were compared with others retrieved from the NCBI database. Strains harboring these integron configurations were bordered by an assortment of insertion sequences and transposons. These E. coli isolates were identified in various host organisms and environmental samples, including animals and meat products (e.g., pigs, poultry, turkeys, veal, other avian species), blood and urine samples collected from hospitalized patients and the community, and environmental sources (wastewater and surface water). Geographic analysis revealed a wide distribution of these strains across continents, occurring in Asia (China and Singapore), Europe (Switzerland, Norway, Italy, France, and Spain), and North America (USA).