To identify genes that are differentially expressed in response to TPEN, E. coli BL21(DE3) cells were grown in minimal medium (76) until the cultures reached mid-log phase. TPEN was introduced, and the cells were cultured for 5 h; this time was chosen so that the results would correspond to previous transcriptional response studies with excess Zn(II) (51). The cultures containing 5 μM TPEN were analyzed with DNA microarrays (E. coli K12 V2 array slides from MWG-BIOTECH), and the results were compared to those from E. coli cells grown in minimal medium containing no TPEN. A minimum of six slides was used for each experiment (three slides with one combination of Cy3 and Cy5 dyes and three slides with swapped dyes). Fifty-five transcripts were significantly up-regulated (twofold) (Table 1). Four genes (ykgM, znuA, znuC, and yodA) that are regulated by Zur, the Zn(II) uptake regulator (30), exhibited significant increases in expression. The remaining Zur-regulated transcript, znuB, was also up-regulated; however, this transcript did not meet our filtering criteria: (i) P values of ≤0.5, (ii) ≥2-fold changes in expression, and (iii) consistent data in all six slides. The expression of zntA, which is regulated by ZntR and encodes the high-affinity Zn(II) exporter in E. coli (7, 14, 78), was unchanged in E. coli cultures stressed with TPEN (see the complete set of DNA array data).

Twenty-nine of the up-regulated genes from TPEN-supplemented cultures are transcriptionally regulated by Fur, the iron uptake regulator (Table 1), and several are involved in Fe transport. The genes (entB, entA, entD, entE, and fes) which encode proteins involved with enterobactin synthesis and uptake (26, 29, 54) and the fec genes (fecR and fecI), which encode proteins involved in Fe import (23), were significantly up-regulated. Several other genes (fhuA, exbB, exbD, fhuD, fhuC, and fhuF) associated with ferrichrome transport in E. coli were also up-regulated (19, 56, 62).

Forty-six transcripts were significantly down-regulated in E. coli cells stressed with TPEN (Table 2). Two operons in E. coli, flgBCDEFGHIJKL and flgAMN, possess genes that encode proteins involved in flagellar biosynthesis (8). Some of these motility-related genes, namely, flgB, fliM, and motB, were previously reported as being up-regulated in E. coli cells stressed with excess Zn(II) (51). In addition, the expression of flagellar biosynthetic proteins in E. coli is affected by the concentration of copper, possibly exerting its effect via the OmpR or H-NS transcriptional regulators (42). All of the genes in the cus and cue systems, which confer copper tolerance to E. coli (68), were also down-regulated, although cusA and copA were filtered out of the data shown in Table 2. Previous studies have demonstrated that the expression levels of the cus and cue genes are dependent on aerobic/anaerobic conditions as well as on the levels of copper in the periplasm/cytoplasm (68). The expression levels of cytoplasmic ferritin (3) were also significantly down-regulated in TPEN-treated cells.

To validate the microarray data, two representative genes were selected for real-time PCR and assayed for the level of mRNA by a two-step, real-time PCR technique. The genes yodA and pdxH, which were up-regulated (21-fold) and exhibited no change (1.1-fold), respectively, were analyzed with real-time PCR. Real-time PCR results were as follows: yodA up-regulated, 17 ± 1; pdxH up-regulated, 1.1 ± 0.1. In order to probe whether the changes observed in cells grown in the presence of TPEN were due to the chelator, RT-PCR was used to probe for the expression levels of yodA in E. coli cells grown in minimal medium containing 5 μM TPEN and 30 μM Zn(II). There was no change in transcript levels of yodA when E. coli was cultured in this medium.

Despite the fact that TPEN is often referred to as a Zn(II)-specific chelator (17, 33, 38, 48, 59, 64, 71, 79, 80, 83, 88), the analyses of our microarray data suggest that the levels of other metal ions may have been affected by the presence of TPEN. TPEN has been reported to bind Cd(II) (K_d = 4.7 × 10⁻¹⁷), Co(II) (K_d = 2.6 × 10⁻¹⁷), Ni(II) (K_d = 2.8 × 10⁻²²), and Cu(II) (K_d = 2.9 × 10⁻²¹) more tightly than it binds Zn(II) (2). In addition, TPEN forms stable complexes with Fe(II) (K_d = 2.5 × 10⁻¹⁵) (2). Since TPEN forms much tighter complexes with Cu(II), it is likely that the down-regulation of the copper homeostasis/export transcripts, cueO, copA, cusA, cusB, cusC, and cusF, is due to low levels of intracellular copper. Twenty-nine Fur-regulated transcripts were up-regulated. Despite the fact that Fe(II) binds 1 order of magnitude less tightly to TPEN than Zn(II) (2), it is possible that TPEN lowered the intracellular concentrations of Fe(II), resulting in the transcription of Fur-regulated iron uptake proteins. The reduction in intracellular concentrations of Fe(II) could be due to direct chelation by TPEN or by oxidation of intracellular Fe(II) to Fe(III). The fact that several other SoxRS-regulated transcripts were up-regulated in cultures stressed with TPEN (Table 1) suggests oxidative stress in these cells (97). It is also possible that the reduction of intracellular Zn(II) caused by the presence of TPEN resulted in an improperly folded Fur, which requires one Zn(II) for proper structure/function (1).