Comparative metabolomics shows the metabolic profiles fluctuate in multi-drug resistant Escherichia coli strains.

Abstract

In this study, two susceptible strains and two multi-drug resistant clinical Escherichia coli strains were obtained by Kirby-Bauer method, and then a GC-MS-based metabolomics method was used to compare the differential expression of metabolites between two drug sensitive (CK1 and CK2) and two multidrug-resistant (MDR1 and MDR2) clinical strains of E. coli. We characterized a total of 273 metabolites, including 77 commonly altered metabolites, between MDR vs. antibiotic sensitive strains. Interestingly, the PCA score plot clearly discriminated drug sensitive and MDR strains. The following bioinformatics analysis showed that biosynthesis of amino acids, biosynthesis of phenylpropanoids and purine metabolism were commonly enriched in MDR strains. Moreover, microbial metabolism in diverse environments, carbon metabolism，and pyrimidine metabolism pathways were more likely to be enriched MDR1 strain, while ABC transporters, and cysteine and methionine metabolism pathways were enriched in MDR2 strains. The enzyme activities in several involved metabolic pathways were further measured and metabolite candidates were validated by GC-MS-SIM method. These results indicated that antibiotic resistance affects the metabolite profiles of bacteria. In general, our study provides evidence on the study and prediction of MDR characteristics and mechanisms in bacteria at the metabolite level. BIOLOGICAL SIGNIFICANCE: Overuse and abuse of antibiotics has led to the emergence of antibiotic-resistant strains of bacteria; however, relatively little is known about their resistance mechanisms. In this study, metabolomics method was used to compare the differential expression of metabolites between sensitive and multidrug-resistant clinical strains of E. coli. Results show that the PCA score plot clearly discriminated sensitive and MDR strains, indicating that they had different metabolic profiles. Further, bioinformatics analysis showed that biosynthesis of amino acids, biosynthesis of phenylpropanoids and purine metabolism may be related to resistance. Finally, the enzyme activities in several involved metabolic pathways were further measured and metabolite candidates were validated by GC-MS-SIM method. In general, our study provides evidence on the study and prediction of MDR characteristics and mechanisms in bacteria at the metabolite level.