Mobility of β-Lactam Resistance Under Bacterial Co-infection and Ampicillin Treatment in a Mouse Model

Abstract

Ingestion of food- or waterborne antibiotic-resistant bacteria may lead to the dissemination of antibiotic-resistance genes in the gut microbiota and the development of antibiotic-resistant bacterial infection, a significant threat to animal and public health. Food or water may be contaminated with multiple resistant bacteria, but animal models on gene transfer were mainly based on single-strain infections. In this study, we investigated the mobility of β-lactam resistance following infection with single- versus multi-strain of resistant bacteria under ampicillin treatment. We characterized three bacterial strains isolated from food-animal production systems, Escherichia coli O80:H26 and Salmonella enterica serovars Bredeney and Heidelberg. Each strain carries at least one conjugative plasmid that encodes a β-lactamase. We orally infected mice with each or all three bacterial strain(s) in the presence or absence of ampicillin treatment. We assessed plasmid transfer from the three donor bacteria to an introduced E. coli CV601gfp recipient in the mouse gut, and evaluated the impacts of the bacterial infection on gut microbiota and gut health. In the absence of ampicillin treatment, none of the donor or recipient bacteria established in the normal gut microbiota and plasmid transfer was not detected. In contrast, the ampicillin treatment disrupted the gut microbiota and enabled S. Bredeney and Heidelberg to colonize and transfer their plasmids to the E. coli CV601gfp recipient. E. coli O80:H26 on its own failed to colonize the mouse gut. However, during co-infection with the two Salmonella strains, E. coli O80:H26 colonized and transferred its plasmid to the E. coli CV601gfp recipient and a residential E. coli O2:H6 strain. The co-infection significantly increased plasmid transfer frequency, enhanced Proteobacteria expansion and resulted in inflammation in the mouse gut. Our findings suggest that single-strain infection models for evaluating in vivo gene transfer may underrepresent the consequences of multi-strain infections following the consumption of heavily contaminated food or water.

Keywords: antibiotic resistance; antibiotic treatment; gut microbiota; intestinal inflammation; plasmid transfer.

FIGURE 1

Maps of plasmids: (A) IncI2 plasmid carried by the Escherichia coli O80:H26 (EC-107) donor, (B) IncN plasmid carried by the Salmonella Bredeney (SA20114778WT) donor, (C) IncA/C2 plasmid carried by the Salmonella Heidelberg (SL-312) donor. AMR genes (red), origin of transfer (green) and other genes (blue).

FIGURE 2

Enumeration of donor, recipient and transconjugant bacteria (mean + SE) in mouse fecal samples. Mice received the inoculation of donor bacteria: (A) EC = Escherichia coli O80:H26 (EC-107), number of mice (n) = 6. (B) SB = Salmonella Bredeney (SA20114778WT), n = 6. (C) SH = Salmonella Heidelberg (SL-312), n = 5. (D) Mix = the mixture of EC, SB and SH, n = 4 with (Amp) or without ampicillin treatment. RE = the E. coli CV601gfp recipient, TR = Transconjugants, dpi = day post infection. Conjugation frequency on 1 dpi (E) was expressed as the ratio of transconjugants to donors (sum of all three donors used in the Mix-Amp group). The conjugation frequency in the Mix-Amp group is significantly greater than that in other groups (P < 0.001) based on the one-way ANOVA test.

FIGURE 3

Representative images of hematoxylin and eosin (H&E)-stained cecum and colon sections. (A1) Normal cecum section with thin mucosa and extremely scant mononuclear cell infiltrate in lamina propria. (A2) Inflamed cecum section with markedly thickened mucosa and hyperplastic crypts, extensive crypt loss and proprial inflammatory cells (white asterisks), submucosal edema (black asterisk) and transmural inflammation (long arrow); surface epithelium is cuboidal and immature (short arrow) or eroded and ulcerated (arrowhead). (B1) Normal colon section with columnar surface epithelium and numerous mucous cells lining colonic crypts. (B2) Inflamed colon section with focal loss of crypts and proprial inflammation (white arrowhead), marked loss of mucous cells and crypt hyperplasia (black asterisk); and low columnar and more basophilic surface epithelium. Histologic inflammatory scores of the cecum (C1) and colon (C2) in each group of mice that received inoculation of donor bacteria and ampicillin treatment (Amp), Ctl = no bacteria, n = 6; EC = Escherichia coli O80:H26 (EC-107), number of mice (n) = 6; SB = Salmonella Bredeney (SA20114778WT), n = 6; SH = Salmonella Heidelberg (SL-312), n = 5; Mix = the mixture of EC, SB and SH, n = 4; or in all groups of mice (X) that received corresponding bacterial inoculation but no ampicillin treatment, n = 27. The percentage of mice that developed inflammation in the cecum or in the colon in the Mix-Amp group is significantly greater than that in the EC-Amp treatment group (P < 0.05) based on the Fischer’s exact test.

FIGURE 4

Microbial community composition analyzed by sequencing of the 16S rRNA gene from mouse fecal samples. Mice received the inoculation of donor bacteria: (A) Ctl = no bacteria, number of mice (n) = 6. (B) EC = Escherichia coli O80:H26 (EC-107), n = 6. (C) SB = Salmonella Bredeney (SA20114778WT), n = 6. (D) SH = Salmonella Heidelberg (SL-312), n = 5. (E) Mix = the mixture of EC, SB and SH, n = 4 with (Amp) or without ampicillin treatment. Data on the relative abundance of Proteobacteria from panels (A–E) were used to determine the relative changes of Proteobacteria responding to various treatments and expressed as the ratio of the relative abundance on 1, 2 or 5 day post infection (dpi) to that on day 0 within each treatment group (panel F). The relative abundance of Firmicutes, Bacteroidetes and Proteobacteria on 1 dpi is significantly different from that on day 0 within each ampicillin treatment group (P < 0.05). The relative fold change of Proteobacteria in the SB-Amp, SH-Amp or Mix-Amp treatment group is significantly greater than that in the Ctl-Amp or EC-Amp group on 1, 2 or 5 dpi (P < 0.05) based on the one-way ANOVA test.

FIGURE 5

Relative abundance of the genera of Escherichia-Shigella and Salmonella in the mouse gut microbiome. Mice received the inoculation of donor bacteria: (A) Ctl = no bacteria, number of mice (n) = 6. (B) EC = Escherichia coli O80:H26 (EC-107), n = 6. (C) SB = Salmonella Bredeney (SA20114778WT), n = 6. (D) SH = Salmonella Heidelberg (SL-312), n = 5. (E) Mix = the mixture of EC, SB and SH, n = 4 with (Amp) or without ampicillin treatment. The relative abundance of Escherichia-Shigella on 1, 2 or 5 dpi is significantly greater than that on day 0 in the Mix-Amp treatment group (P < 0.05) based on the one-way ANOVA test.