Interaction between mutations and regulation of gene expression during development of de novo antibiotic resistance

Abstract

Bacteria can become resistant not only by horizontal gene transfer or other forms of exchange of genetic information but also by de novo by adaptation at the gene expression level and through DNA mutations. The interrelationship between changes in gene expression and DNA mutations during acquisition of resistance is not well documented. In addition, it is not known whether the DNA mutations leading to resistance always occur in the same order and whether the final result is always identical. The expression of >4,000 genes in Escherichia coli was compared upon adaptation to amoxicillin, tetracycline, and enrofloxacin. During adaptation, known resistance genes were sequenced for mutations that cause resistance. The order of mutations varied within two sets of strains adapted in parallel to amoxicillin and enrofloxacin, respectively, whereas the buildup of resistance was very similar. No specific mutations were related to the rather modest increase in tetracycline resistance. Ribosome-sensed induction and efflux pump activation initially protected the cell through induction of expression and allowed it to survive low levels of antibiotics. Subsequently, mutations were promoted by the stress-induced SOS response that stimulated modulation of genetic instability, and these mutations resulted in resistance to even higher antibiotic concentrations. The initial adaptation at the expression level enabled a subsequent trial and error search for the optimal mutations. The quantitative adjustment of cellular processes at different levels accelerated the acquisition of antibiotic resistance.

FIG 1

Expression profiles of cells adapted to antibiotics and grown at 0.25× MIC, compared to the wild-type cells in the absence of antibiotics. (a) Number of differentially up- and downregulated genes in amoxicillin (Amx)-adapted (MIC, 512 μg/ml), tetracycline (Tetra)-adapted (MIC, 64 μg/ml), or enrofloxacin (Enro)-adapted (MIC, 512 μg/ml) E. coli cells compared to the wild-type cells. Genes are listed when expression is significantly (95% confidence level) changed by a factor exceeding 2. (b) Number of genes that are up- or downregulated, grouped according to the factor of the differential expression in E. coli cells resistant to enrofloxacin (Enro), tetracycline (Tetra), and amoxicillin (Amx).

FIG 2

Genetic modifications in the ampC promoter region of E. coli MG1655 during the acquisition of amoxicillin resistance. (a) The measured MIC as a function of the amoxicillin concentration in the culture and the mutations found in the ampC promoter sequence over the course of stepwise increasing amoxicillin concentrations for 7 replicate cultures of E. coli MG1655. For every concentration, PCR products of 2 clones were sequenced. The asterisks indicate mutations that were found only in one colony. Green color indicates mutations in the ampC attenuator region, yellow, mutations in the −10 box, blue, mutations in the interbox distance, and red, mutations in the −35 box. (b) Genomic location of the ampC promoter mutations identified during the acquisition of amoxicillin resistance in the genome of E. coli MG1655.

FIG 3

Genetic modifications in resistance-conferring genes of E. coli MG1655 during the adaptation to enrofloxacin. Top, plot of the measured MIC as a function of the enrofloxacin concentration in the medium. Bottom, mutations in gyrA, parC, and gyrB as a function of the MIC during growth at stepwise increasing enrofloxacin concentrations in 7 replicate cultures of E. coli MG1655. For every concentration, 2 clones were sequenced. The asterisks indicate mutations found only in one colony; blue, mutations in gyrA; red, mutations in parC; grey, mutations in gyrB.

FIG 4

Mutations found in resistance-conferring regions of gyrA, parC, and gyrB in 2 enrofloxacin-resistant E. coli replicates cultured for 30 days in the presence or absence of the antibiotic. Blue indicates mutations in gyrA; red, mutations in parC; grey, mutations in gyrB. The asterisks indicate mutations found in only one of the two colonies that were sequenced for each data point.

FIG 5

Overlap of differentially up- and downregulated genes in E. coli MG1655 wild-type and antibiotic-resistant cells. (a) Overlap of up- and downregulated genes in E. coli cells resistant to enrofloxacin (Enro), tetracycline (Tetra), and amoxicillin (Amx) in the absence of antibiotics compared to wild-type expression levels. The genes gadABC and hdeA were downregulated in all three antibiotic adapted strains. (b) Overlap of up- and downregulated genes in enrofloxacin-exposed (0.25× MIC, 0.125 μg/ml) wild-type and enrofloxacin-resistant E. coli cells (Enro, enrofloxacin adapted; Enro30, enrofloxacin adapted and cultured 30 days without the antibiotic) compared to wild-type expression levels. Only the NADH-quinone reductase azoR was downregulated in all 3 conditions.

FIG 6

Change in transcriptomic profile of wild-type and antibiotic-resistant E. coli cells in response to short-term (<10 generations) drug exposure and long-term adaptation (>100 generations). (a) Number of up- and downregulated genes after antibiotic exposure (0.25× MIC, 1 μg/ml amoxicillin, 0.25 μg/ml tetracycline, and 0.125 μg/ml enrofloxacin) in wild-type (WT) and antibiotic-resistant (AR) cells compared to wild-type cells. (b) Number of up- or downregulated genes in AR cells compared to wild-type cells for the wild-type cells exposed to 0.25× MIC enrofloxacin [WT(exposed)], enrofloxacin-adapted cells (Enro), enrofloxacin-adapted cells exposed to 0.25× MIC [Enro(exposed)], and enrofloxacin-adapted cells cultured for 30 days without the antibiotic (Enro 30).