ABSTRACT: Metals at high concentrations can exert toxic effects on microorganisms. It has been widely reported that lowering environmental pH reduces effects of cadmium toxicity in bacteria. Understanding the effects of pH-mediated cadmium toxicity on bacteria would be useful for minimizing cadmium toxicity in the environment and gaining insight into the interactions between organic and inorganic components of life. Growth curve analysis confirmed that cadmium was less toxic to Escherichia coli at pH 5 than at pH 7 in M9 minimal salts medium. To better understand the cellular mechanisms by which lowering pH decreases cadmium toxicity, we used DNA microarrays to characterize global gene expression patterns in E. coli in response to cadmium exposure at moderately acidic (5) and neutral (7) values of pH. Higher expression of several stress response genes including hdeA, otsA, and yjbJ at pH 5 after only 5 minutes was observed and may suggest that acidic pH more rapidly induces genes that confer cadmium resistance. Genes involved in transport were more highly expressed at pH 7 than at pH 5 in the presence of cadmium. Of the genes that showed an interaction between pH and cadmium effects, 46% encoded hypothetical proteins, which may have novel functions involved in mitigating cadmium toxicity. GROWTH CONDITIONS FOR MICROARRAY EXPERIMENTS: Two overnight cultures of E. coli K-12 were started in M9 medium. A 250 mL Erlenmeyer flask containing 100 mL of M9 medium was inoculated with 0.5 mL for each of the two overnight cultures, each of which was considered a biological replicate. The cultures were grown on a rotary shaker (200 rpm) at 37 °C until the contents of the flask reached an OD600 of 0.3 (mid-log phase of growth). Each culture was divided into four 25 mL aliquots, transferred to 50 mL conical tubes (Corning), and centrifuged at 2540 x g for 12 minutes. The supernatant was decanted, and the cells were resuspended in 25 mL of M9 medium at pH 7 or pH 5 in the presence or absence (two cultures of each) of 5.4 µM (1 µg/mL) total cadmium, added as CdCl2. The cultures were incubated at 25 °C for either 5 or 15 minutes with manual rotation of the flasks once per minute. After the appropriate amount of time, 15 mL of RNAProtect Bacteria Reagent (Qiagen) was added to each culture to immediately halt all metabolic processes. The solutions were vortexed, incubated at 25 °C for 5 minutes, and centrifuged for 12 minutes at 3750 x g. RNA was extracted from the cell pellets immediately following centrifugation. RNA EXTRACTION AND HYBRIDIZATION PROCEDURES: RNA was extracted and purified using a Masterpure RNA purification kit (Epicentre Technologies). The quantity and quality of the RNA samples were determined spectrophotometrically. Preparation of the cDNA, labeling with Cy3 and Cy5, and successive hybridizations were accomplished using a 3DNA Array 900MPX kit following the manufacturer’s protocols (Genisphere) with the following modifications. 3DNA reverse transcriptase enzyme (Genisphere # RT300320) rather than SuperScript II was added to 1 µg of RNA and 2 µL of a random primer (1 µg/µL). The final cDNA hybridization mix contained 29 µL 2X enhanced cDNA hybridization buffer rather than 2X SDS-based hybridization buffer or 2X formamide-based hybridization buffer. The cDNA mix was hybridized to a cDNA microarray printed by the Microarray and Proteomics Facility at the University of Alberta (Operon version 1.0 oligonucleotides). The arrays were scanned with a Versarray ChipReader (BioRad) with laser power at 75%, photomultiplier tube (PMT) sensitivity at 800 V, and detector gain at 1. DATA ANALYSIS: Each array directly compared transcription at pH 5 and pH 7 for a given cadmium treatment (0 or 5.4 µM cadmium) and exposure time. Dye swaps were performed for each biological replicate for each of the following treatments (8 total arrays): 5 minutes with cadmium exposure, 5 minutes without cadmium exposure, 15 minutes with cadmium exposure, and 15 minutes without cadmium exposure. The 0-minute exposure to cadmium treatment was obtained from the 5-minute microarray without cadmium exposure. Spot intensities and locations were determined using TIGR Spotfinder, Version 3.1.1. All subsequent analyses were performed using the ma-anova package in the open-source statistical software package, R (www.r-project.org), Version 2.4.1. The data were normalized using the regional lowess method. Following normalization the median expression values of genes represented in triplicate on each array were determined for each gene. A mixed model two-way ANOVA for the main fixed effects of pH, cadmium, and their interaction (array and spot were the random effects) were performed (using Type III F-tests) separately for each time point to identify genes for which pH and cadmium interacted to significantly affect expression (FDR-adjusted p ? 0.05).
