Multidrug resistance plasmids commonly reprogram the expression of metabolic genes in Escherichia coli

Abstract

Multidrug-resistant Escherichia coli is a leading cause of global mortality. Transfer of plasmids carrying genes encoding beta-lactamases, carbapenamases, and colistin resistance between lineages is driving the rising rates of hard-to-treat nosocomial and community infections. Multidrug resistance (MDR) plasmid acquisition commonly causes transcriptional disruption, and while a number of studies have shown strain-specific fitness and transcriptional effects of an MDR plasmid across diverse bacterial lineages, fewer studies have compared the impacts of different MDR plasmids in a common bacterial host. As such, our ability to predict which MDR plasmids are the most likely to be maintained and spread in bacterial populations is limited. Here, we introduced eight diverse MDR plasmids encoding resistances against a range of clinically important antibiotics into E. coli K-12 MG1655 and measured their fitness costs and transcriptional impacts. The scale of the transcriptional responses varied substantially between plasmids, ranging from >650 to <20 chromosomal genes being differentially expressed. However, the scale of regulatory disruption did not correlate significantly with the magnitude of the plasmid fitness cost, which also varied between plasmids. The identities of differentially expressed genes differed between transconjugants, although the expression of certain metabolic genes and functions were convergently affected by multiple plasmids, including the downregulation of genes involved in L-methionine transport and metabolism. Our data show the complexity of the interaction between host genetic background and plasmid genetic background in determining the impact of MDR plasmid acquisition on E. coli.

Importance: The increase in infections that are resistant to multiple classes of antibiotics, including those isolates that carry carbapenamases, beta-lactamases, and colistin resistance genes, is of global concern. Many of these resistances are spread by conjugative plasmids. Understanding more about how an isolate responds to an incoming plasmid that encodes antibiotic resistance will provide information that could be used to predict the emergence of MDR lineages. Here, the identification of metabolic networks as being particularly sensitive to incoming plasmids suggests the possible targets for reducing plasmid transfer.

Keywords: Escherichia coli; multidrug resistance; plasmids; transcriptomics.

Fig 1

Maps for the nine plasmids conjugated into E. coli MG1655 Turquoise = resistance genes, including ones not identified by ResFinder. Orange = conjugal transfer genes, including potential gene/gene fragments not identified by Prokka but found using manual BLAST queries. Maroon = transposase. Grey = hypothetical protein. Blue = annotated gene of other function.

Fig 2

Genes significantly upregulated (orange) and downregulated (blue) FDR threshold of P < 0.05 and an absolute log fold change (FC) of at least one] in the transconjugants in comparison to the plasmid-free control. Genes that had an absolute log FC of at least one but did not reach the FDR threshold are shown in gray and are considered to not be significantly differentially expressed. Plots are labeled by the donor with respect to Table 1. Select genes are labeled.

Fig 3

Correlation (Pearson) between a total number of differentially expressed genes and competition index (CI) of the transconjugants against a plasmid-free strain (mean of biological triplicate), whereby a value less than one indicates the plasmid-free is fitter in the competition condition. Individual CI values are plotted.

Fig 4

GO terms significantly enriched in the combined set of all transconjugants. A value greater than one indicates more genes observed than expected. Terms are grouped by their role in biological (orange box), molecular (pink), or cellular (blue) function. GO IDs and P values given in Table S3.

Fig 5

(A) Simplified L-methionine biosynthetic pathway, with key genes highlighted and (B) the logFC values for the genes downregulated on the acquisition of an MDR plasmid, colored by gene involved in L-methionine transporter and biosynthesis.

Fig 6

Genes upregulated (circle) and downregulated (square) across all transconjugants that relate to L-arginine biosynthetic pathways. Selected metabolites in each reaction are shown for simplicity, with key metabolites or reactions highlighted in bold.

Fig 7

Components of ABC transporters that are significantly upregulated (orange background) or downregulated (blue background) in 7E2 (pink), LL19 (purple), both (green), or neither (white). Simplified schematics of the periplasmic binding, membrane, and ATP-binding proteins are shown. Transporters with significant DE genes with unknown or putative specificity are not included but are detailed in the Supplementary Data Sheet.