Too much of a good thing: Adaption to iron (II) intoxication in Escherichia coli

Abstract

Background: There has been an increased usage of metallic antimicrobial materials to control pathogenic and multi-drug resistant bacteria. Yet, there is a corresponding need to know if this usage leads to genetic adaptations that could produce more harmful strains.

Methodology: Experimental evolution was used to adapt Escherichia coli K-12 MG1655 to excess iron (II) with subsequent genomic analysis. Phenotypic assays and gene expression studies were conducted to demonstrate pleiotropic effects associated with this adaptation and to elucidate potential cellular responses.

Results: After 200 days of adaptation, populations cultured in excess iron (II), showed a significant increase in 24-h optical densities compared to controls. Furthermore, these populations showed increased resistance toward other metals [iron (III) and gallium (III)] and to traditional antibiotics (bacitracin, rifampin, chloramphenicol and sulfanilamide). Genomic analysis identified selective sweeps in three genes; fecA, ptsP and ilvG unique to the iron (II) resistant populations, and gene expression studies demonstrated that their cellular response may be to downregulate genes involved in iron transport (cirA and fecA) while increasing the oxidative stress response (oxyR, soxS and soxR) prior to FeSO₄ exposure.

Conclusions and implications: Together, this indicates that the selected populations can quickly adapt to stressful levels of iron (II). This study is unique in that it demonstrates that E. coli can adapt to environments that contain excess levels of an essential micronutrient while also demonstrating the genomic foundations of the response and the pleiotropic consequences. The fact that adaptation to excess iron also causes increases in general antibiotic resistance is a serious concern. Lay summary: The evolution of iron resistance in E. coli leads to multi-drug and general metal resistance through the acquisition of mutations in three genes (fecA, ptsP and ilvG) while also initiating cellular defenses as part of their normal growth process.

Keywords: E. coli; experimental evolution; gene expression; iron resistance; pleiotropy.

Figure 1.

Twenty-four hours growth assay to assess the progression and overall timeframe of iron adaptation. All five individual control populations (C1–C5), all five individual iron adapted populations (Fe²⁺_1-Fe²⁺_5) and four replicates of the ancestral population were grown for 24-h in increasing concentrations of FeSO₄ in triplicate. Optical densities were then measured at 602 nm. The mean of the controls (red), ancestral (green) and adapted (blue) populations were then plotted, error bars represent standard error among the populations. (A) After 200-days of adaptation, the control populations showed almost complete inhibition at 2500 mg/l. The iron (II)-selected populations display increasing optical densities at low concentration and steady growth at higher concentrations, all of which are significantly higher than what was exhibited by the controls and the ancestral population. This data giving strong evidence for the successful adaptation to optimized growth in iron (II) after 200-day of selection. (B) 24-h growth in Fe₂(SO₄)₃, (C) 24-h growth in Ga(NO₃)₃, (D) 24-h growth in AgNO₃ and (E) 24-h growth in Na₂SO₄. All iron (II)-adapted populations showed increased resistance to all four of the metals tested. Albeit, the overall growth rate increase was most apparent in iron (III) and the iron-analog gallium (III). Optical densities were severely hindered in both silver and sulfate with the iron (II)-adapted populations slightly outperforming the controls. This data demonstrates the pleiotropic effects associated with iron (II) adaptation which leads to general metal resistance. For statistical analysis, ANOVA analysis using a generalized linear model was performed to evaluate the treatment effect associated with selection in FeSO₄. Both the F-statistic and P-values are reported in Table 1

Figure 2.

Twenty-four hours growth assays to assess general metal and antibiotic resistance associated with FeSO₄ adaptation. Twenty-four hours optical assays were constructed similarly to those in Fig. 1 in seven antibiotics with varying mechanisms of action. (A) Ampicillin, (B) bacitracin, (C) polymyxin-B, (D) rifampin, (E) tetracycline (F) chloramphenicol and (G) sulfanilamide. All iron (II)-adapted populations showed increased resistance to four of the seven antibiotics tested. This data demonstrates the pleiotropic effects associated with iron (II) adaptation which leads to general antibiotic resistance. Statistical analysis was performed identically to Fig. 1 with both the F-statistic and P-values are displayed Table 1

Figure 3.

Differential gene expression in response to FeSO₄ acclimation. Gene expression studies were conducted at day 200 using NanoString technologies of 48 genes selected for their acquisition of mutations during the selection regime or for their role in general iron metabolism. Here, fold changes in RNA counts between each averaged set of populations for 38 of the 50 selected genes are reported (the remainder can be seen in Supplementary Table S2). For analysis, all genes that had a calculated P-value > 0.05 for their fold change were deemed zero. Each averaged population [controls or iron (II) adapted] were analyzed in pairwise combinations for fold-changes and represented in the heat map. Most notably, the adapted populations’ downregulate genes involved in iron transport and increase expression of oxidative stress response genes even in absence of exposure to iron. Normalized NanoString gene expression data was analyzed in terms of ratios and then converted into log₂ ratios, (fold-changes). For statistical analysis, a t-test was performed on log₂-transformed count data for each pairwise comparison and reported in Supplementary Table S2

Figure 4.

Mapping of the FecA mutations. Sequencing of the 200-day iron (II)-adapted populations identified three mutations in the fecA genes (D87Y, G201C and A526T). (A) The apo-FecA structure (PDB1PNZ) was used to map these three residues. D87Y (black) which is unique to the 200-day populations resides adjacent to the TonB-binding box (pink), G210C (yellow) in the internal plug domain important for ligand transport (blue) and A526T (yellow) on the periplasmic lid (green). Both G210C and A526T were also detected in our 45-day iron (II) adapted populations [27]. In red are two mutations (N169K and G400S) which were detected in gallium (iron analog) resistant strains [38] and as shown, map to similar regions as the iron resistant mutations. (B) The iron-citrate complexed structure (PDB1PO3) was also used to map two of the three iron resistant residues (D87Y lies in an unstructured region and therefore could not be mapped) and the gallium mutations. This shows the major rearrangement of the periplasmic lid (green) and the significant movement of A526 upon ligand binding. (C) The iron-citrate binding site of the bound structure (PDB1PO3) to demonstrate the proximity of two of the mapped residues (G210C and A526T) and the gallium mutations (N169K and G400S) to the ligand binding site. The importance of all three regions indicate that these mutations may prevent iron-citrate transport and therefore limit entry of excess iron into the cell by occlusion thereby increasing chances of survival and evidence for genetic adaptation