|Home | About | Journals | Submit | Contact Us | Français|
The Arc (anoxic redox control) two-component signal transduction system, consisting of the ArcB sensor kinase and the ArcA response regulator, allows adaptive responses of Escherichia coli to changes of O2 availability. The arcA gene was previously known as the dye gene because null mutants were growth sensitive to the photosensitizer redox dyes toluidine blue and methylene blue, a phenotype whose molecular basis still remains elusive. In this study we report that the toluidine blue O (TBO) effect on the arc mutants is light independent and observed only during aerobic growth conditions. Moreover, 16 suppressor mutants with restored growth were generated and analyzed. Thirteen of those possessed insertion elements upstream of the cydAB operon, rendering its expression ArcA independent. Also, it was found that, in contrast to cythocrome d, cythocrome o was not able to confer toluidine blue resistance to arc mutants, thereby representing an intriguing difference between the two terminal oxidases. Finally, a mechanism for TBO sensitivity and resistance is discussed.
The Arc (anoxic redox control) two-component system is a key element in the complex transcriptional regulatory network that allows facultative anaerobic bacteria, such as Escherichia coli, to sense various respiratory growth conditions and adjust their gene expression accordingly (42). This system comprises the transmembrane sensor kinase ArcB (32) and the cytoplasmic response regulator ArcA (34). Under reducing conditions of growth, ArcB autophosphorylates at the expense of ATP, a process enhanced by various anaerobic metabolites, such as lactate and acetate (19, 49), and transphosphorylates the response regulator ArcA (22, 37). The phosphorylated form of ArcA, ArcA-P, in turn, regulates negatively the expression of many operons that code for enzymes involved in aerobic metabolism and activates the expression of genes encoding proteins involved in fermentative metabolism (40, 42). Under oxidizing conditions, the kinase activity of ArcB is inhibited by the quinone electron carriers through the oxidation of Cys 180 and Cys 241, which participate in intermolecular disulfide bond formation (20, 41), allowing dephosphorylation of ArcA (18, 45).
Before the identification and characterization of Arc as a two-component system, the arcA gene was known as the dye gene because it was observed that mutation in this gene conferred sensitivity to dyes such as toluidine blue O (TBO) and methylene blue (8). Later, it was observed that mutants carrying mutations in arcB and in the cytochrome d-encoding operon, cydAB, exhibit a similar TBO-sensitive phenotype (15, 32). However, the causes of the dye phenotype in these mutants remain so far unknown. It is of interest to mention that TBO and methylene blue are photosensitizers that in the presence of light are able to instigate redox reactions producing reactive oxygen species (ROS), which can damage nucleic acids and enzymes, leading to cell death (57). The utility of these photosensitizers against a range of bacterial strains has been reported extensively (59). In a recent study, it was proposed that TBO results in a significant increase of ROS in an arcA mutant and that this increase in ROS is the cause of the dye phenotype (50). Moreover, the heterologous expression of poly 3-hydroxybutyrate was shown to be able to suppress the dye sensitivity in arcA mutants by diminishing O2 consumption and the production of ROS (50).
In this work, we examined the effect of oxygen and light and also the effect of the antioxidants sodA, katG (hydroperoxidase I [HPI]), katE (HPII), Hmp (flavohemoglobin), and AhpCF (alkyl hydroperoxide reductase) and the carotenoid biosynthetic genes of Erwinia herbicola (46) on the TBO-dependent phenotype of arc mutants. Furthermore, we generated and characterized several suppressor mutant strains with abolished TBO-dependent growth defects. Our results demonstrate that cytochrome d, but not cytochrome o, serves as a key element in the protection of the cells against TBO by generating an anoxic intracellular environment. Finally, aerobic expression of cydAB in the isolated suppressor mutants was found to be achieved by the introduction of insertion sequences (IS) upstream of the cydAB operon, rendering its expression arc independent.
The strains and plasmids used in this study are listed in Table Table1.1. Strain IFC5001 was constructed by P1 transduction of the h-ns::Kanr allele derived from the Keio collection (3) into strain ECL5020 (arcA::Tetr) (21). Strains IFC5002, IFC5003, and IFC5004 were constructed by P1 transduction of the katG::Kanr, katE::Kanr and sodA::Kanr alleles, respectively, derived from the Keio collection into strain MC4100. Plasmid pMX513 was constructed by cloning a 2.3-kb SmaI fragment from pDT1.5 (54), containing the sodA gene under its own promoter, into the SmaI site of pACT3 (17). The genomic libraries of the suppressor strains were constructed by Sau3A1 restriction of the genomic DNA and cloning the obtained fragments into BamHI-digested pTZ19R (Fermentas Inc, Glen Burnie, MD).
For the PCR amplification of the cydAB promoter region, the reverse primer cydpromRev (5′-CCCGGATCCCATCATGACTCCTTGCTCATCGC-3′) was used with the forward primers cydpromFw700, 5′-GGAATTCTGGCAAGCGTAGCGCATCAGG-3; cydpromFw1600, 5′-CGGAATTCGTGTCCTGGTCCCTACACCT-3′; cydpromFw2100, 5′-CGGAATTCGCCATGCCCGTGAGATCGA-3′; and cydpromFw4000, 5′-CGGAATTCTGCCTGGGCGCAATGGC-3′, yielding products of 700, 1,600, 2,100, and 4,000 bp, respectively, when used with genomic DNA from a wild-type MC4100 strain (10) as a template.
The genes hmp and ahpCF were amplified by PCR using chromosomal DNA of strain MC4100 (10) as a template and the following primers: hmp-up (5′-CCCCATATGCTTGACGCTCAAAC-3′) and hmp-down (5′-CCCAAGCTTGCCGGATGTTTCCATCC-3′), yielding a 1,235-bp product; and ahpCF-up (5′-CCCCATATGTCCTTGATTAACACC-3′) and ahpCF-down (CCCAAGCTTCGGCGGCTAAGCAATTGC-3′), yielding a 2,426-bp product. Both PCR products were digested with NdeI-HindIII and cloned between the NdeI-HindIII sites of plasmid pMX020 (45) under the control of the arabinose promoter, generating the Hmp and AhpCF overexpressing plasmids pMX511 and pMX512, respectively.
Escherichia coli strains were routinely grown in Luria-Bertani (LB) medium or tryptone broth (10 g liter−1 tryptone, 8 g liter−1 NaCl) at 37°C. Unless otherwise specified, TBO sensitivity was tested on tryptone-TBO-agar (TTA) plates (10 g liter−1 tryptone, 8 g liter−1 NaCl, 0.2 mg ml−1 TBO , and 15 g liter−1 of Bacto agar). When necessary, ampicillin, kanamycin, tetracycline, and chloramphenicol were used at final concentrations of 100, 50, 12.5, and 34 μg ml−1, respectively. To induce expression of the ara promoter-controlled genes, arabinose was added to TTA plates to a final concentration of 0.13 mM.
To determine the catalase and superoxide dismutase (SOD) activities, the strain of interest was grown aerobically in LB medium to an optical density at 600 nm (OD600 of ~0.6). Cells were harvested and washed with 100 mM cold HEPES buffer (pH 7.2) and disrupted by sonication, and unbroken cells and debris were eliminated by centrifugation at 9,000 × g for 10 min at 4°C. Soluble proteins (6 to 30 mg) were separated on 8% native polyacrylamide gels at 4°C and at 150 V for 1 h or 2.5 h for the determination of SOD or catalase, respectively. Catalase activity was detected as described previously (12), whereas a modified photochemical method of Beauchamp and Fridovich (4) was used to locate SOD activities on gels. Briefly, the gel was first soaked in 25 ml of 1.23 mM nitroblue tetrazolium for 15 min and then in 25 ml of a solution containing 28 mM N,N,N′,N′′-tetramethylethylenediamine and 28 μM riboflavin for another 15 min. Both incubations where carried out in the dark. Subsequently, the gel was exposed to light, to initiate the photochemical reaction, until the bands were clearly distinguishable. Carotenoid content was determined by growing the cells in LB broth at 37°C with constant agitation. The cells were collected by centrifugation at 4,000 × g for 15 min. The pellet fraction was weighed, suspended in a 1:1 (vol/vol) mixture of chloroform-methanol, and vortexed for 2 min. The suspension was separated by centrifugation at 4,000 × g for 5 min, and the organic layer was used for measure of absorption spectra between 370 and 520 nm. The maximum wavelength for β-carotene in chloroform-methanol solvent was 460 nm, as previously reported (5).
For the quantification of cytochrome d each strain was aerobically grown in 1.5 liter LB medium to middle-exponential phase (an OD600 of approximately 0.6). Cells were harvested and washed twice with 50 mM cold phosphate buffer (pH 7) containing 5 mM CaCl2 and 5 mM MgCl2. The cells were disrupted by sonication. Unbroken cells and debris were eliminated by centrifugation at 9,000 × g for 10 min. Membranes were prepared by centrifugation at 125,000 × g for 40 min and thereafter washed twice with phosphate buffer. Membranes were suspended in phosphate buffer, and protein concentrations were measured by a modification of the Lowry method (16).
Conventional optical absorption difference spectra were recorded at room temperature in an SLM-Aminco DW 2000 spectrophotometer (SLM Instruments Inc.) using 1-cm light path cuvettes. Samples were reduced with solid sodium dithionite, whereas references were oxidized with air (1 min of vortex agitation). The concentration of cytochrome d in membranes was calculated from the spectrum difference (dithionite-reduced samples minus air-oxidized references) at room temperature, by using the wavelength pairs 630 and 650 nm and an extinction coefficient (E) of 19 mM−1 cm−1 (26).
Deletion of the arcA, arcB, or cydAB genes has previously been reported to result in an increased growth-sensitive phenotype in the presence of the redox dyes TBO or methylene blue (8, 15, 32). In addition, it is well known that in the presence of light, these photosensitizer dyes generate ROS that damage nucleic acids and proteins, thereby leading to cell death (57). Therefore, to test whether ROS cause the TBO-dependent growth-sensitive phenotype, the arcA, arcB, and cydAB mutants (11, 33, 38) were plated out on TTA plates and incubated in the presence of light or in darkness. It was found that none of the three mutants was able to grow in either condition (Fig. 1A and B), indicating that either ROS do not cause the above-described phenotype or that TBO-dependent ROS production occurs even in darkness. To differentiate between the two possibilities, cultures of a wild type and its isogenic arcA mutant strain were challenged with 0.05 mg/ml TBO for 30 min in the presence of light or in darkness, and the activity of katG-encoded HPI, an OxyR-induced protein, was monitored. Curiously, a significant TBO-dependent increase of HPI activity, indicative of ROS production, was observed with both strains when incubated in the presence of light but also in darkness (Fig. (Fig.1C).1C). Therefore, we argued that if ROS were the principal cause for the TBO-dependent growth defect of the above-described mutants, overproduction of the antioxidant enzymes SodA (54), HPI (55), and HPII (58) or the carotenoid biosynthetic proteins of Erwinia herbicola (46) may protect the mutant cells and suppress their growth defect. To test this, plasmids carrying the above-described genes were transformed into a ΔarcA::kan mutant strain (11) and plated out on TTA plates. Although all the above-mentioned plasmid-borne genes were able to overexpress the corresponding activity (Fig. 2A to C), not one was able to suppress the dye phenotype of the arcA mutant strain (data not shown). Also, the simultaneous overexpression of katG or katE and sodA that act sequentially to inactivate ROS, did not have any effect on the growth of the mutants in the presence of TBO. Moreover, deletion of katG, katE, or sodA did not result in a TBO-dependent growth-sensitive phenotype of the respective mutant strains (data not shown). Finally, the possible suppressing effect of Hmp (flavohemoglobin) and AhpCF (alkyl hydroperoxide reductase), which are suggested to be involved in the repair of oxidative damaged lipid membranes and in the scavenge of endogenous hydrogen peroxide, respectively (7, 43), on the TBO-dependent growth defect of the arcA mutant was tested. The ΔarcA::kan mutant strain carrying plasmid pMX511 or pMX512, overproducing HMP or AhpCF, respectively, were not able to grow on arabinose containing TTA plates (data not shown). Thus, ROS per se might not be the principal cause for the TBO-dependent growth defect of the arc and cydAB mutants.
It is well established that the ArcA/B system is active under reducing growth conditions (42). Also, under such conditions expression of the cydAB operon is activated (33, 56). Therefore, we tested whether the above-described mutants exhibit the same TBO-dependent growth defect under anaerobic growth. Surprisingly, the mutant cells grew equally as well as their wild-type counterparts (Fig. (Fig.3A).3A). Moreover, it was observed that the dye surrounding the bacterial colonies was reduced, as judged by the bleaching of the dye, although it was rapidly reoxidized by exposure of the plates to air oxygen (data not shown). It therefore seems that in the presence of oxygen the mutant cells do not have the reducing power to cope with the stress generated by TBO, which could perturb the electron transport chain by diverting electrons to redox cycling, thereby subverting energy production. To test this, the arcA, arcB, and cydAB mutant strains were plated on glucose-supplemented TTA plates, because in the presence of glucose the cells do not require efficient respiration to generate sufficient ATP production to support cell growth. Indeed, all mutant strains were able to grow on the glucose-supplemented TTA plates (Fig. (Fig.3B).3B). Hence, TBO appears to inhibit cell growth by impeding energy production through diverting electrons to futile flow.
In an attempt to test whether TBO exerts a bactericidal or bacteriostatic effect on the arc mutants, the ΔarcA::kan mutant strain was grown in LB, and ~103 cells were plated out either on tryptone-agar plates or on several TTA plates. As expected, the arcA mutant strain grew on the plates without TBO but not on the ones with TBO. Subsequently, every 24 h the cells of a TTA plate were transferred to a plate without TBO by replica plating in order to estimate the number of living cells. A twofold decrease in the number of CFU, indicative of cell death, was observed after day 4 (Fig. (Fig.2C).2C). Thus, the presence of TBO appears to lead to the death of the arcA mutant cells, although with a rather slow death rate.
Curiously, various colonies start appearing on the TTA plates after 4 days of incubation. These colonies (IFCS1 to -16) were picked and restreaked on TTA plates. All of them were resistant to kanamycin, indicating that they were not contaminants, and were able to produce colonies the size of those of a wild-type strain after overnight incubation (data not shown), indicative of a suppressor mutation event(s).
To characterize the suppression event(s), pTZ19R-based genomic libraries of the suppressor mutants were constructed and screened by transformation of the plasmid libraries into the arcA::kan mother cells, which in turn were plated out on TTA plates. Two plasmids, pMX515 and pMX516, derived from the libraries of strains IFCS1 and IFCS13 were found to suppress the TBO-sensitive phenotype of the arcA mutant (data not shown). Subsequent sequencing analysis of the two plasmids revealed the same 4,654-bp fragment containing the cydAB structural genes with 170 bp upstream of the cydA start codon and 1,758 bp downstream of the cydB stop codon. To find out whether cydAB or the downstream genes (ybgTEC and tolQ) were responsible for the suppression effect, the EcoRI-EcoRI fragment of plasmid pMX515 containing the cydAB operon was subcloned into the same vector, resulting in plasmid pMX515Eco, and transformed into the arcA::kan mutant cells. The transformed cells grew equally as well as the wild-type cells on TTA plates (data not shown). Thus, cydAB appears to be able to suppress the TBO-dependent growth defect of the arcA mutants.
To confirm that the increased expression of cydAB suppresses the TBO-dependent growth defect of arc mutants, the effect of ectopic expression of wild-type cydAB, using plasmid pNG2 (25), on the TBO growth-sensitive phenotype of the mutants was tested. In agreement with the above-described result, overexpression of cydAB allowed the arc mutants to grow on TTA plates (Fig. 4A and B), indicating that an increase in the expression of the cydAB operon in the suppressor strain is responsible for the suppression effect.
E. coli is able to use two different terminal oxidases: cytochrome o, which is encoded by the cyoABCD operon and is used during oxygen-rich conditions, and cytochrome d, encoded by cydAB, which is preferably expressed in microaerobic growth conditions. Since expression of cytochrome d appears to suppress the TBO-dependent phenotype of the arc mutants, we tested whether the ectopic expression of cyoABCD, using plasmid pRG110 (2), also suppresses the growth defect of the arc mutants. No suppression was observed (Fig. 4A and B), indicating that in contrast to cytochrome d, cytochrome o is not able to protect the arc mutants from the deleterious effect of TBO.
Considering that the two oxidases differ, in that cytochrome d has high affinity to O2, whereas cytochrome o has low affinity to O2 (13, 14, 48), we argued that this higher affinity to O2 might provide the basis for the TBO growth defect suppression of the arc mutant. To test this, we examined whether the heterologous expression of recombinant hemoglobin 1 (rHb1) and leghemoglobin a (Lba), two plant-derived hemoglobins with high affinity to O2, in arcA and cydAB mutants is sufficient to suppress their TBO sensitivity. To this end, the arcA and cydAB mutants were transformed with plasmid prHb1 (1) or pLba1 (29), carrying rHb1 of Oryza sativa and Lba of Glycine max, respectively, and the transformants were plated out on TTA plates. Interestingly, the ectopic expression of both hemoglobins suppressed the TBO growth defect of both the arcA and cydAB mutants (Fig. (Fig.4C),4C), suggesting that the high affinity of cytochrome d to O2 provides the key element for the TBO-dependent growth phenotype of the arcA mutant. This, in combination with the fact that oxygen convection on agar plates is too slow to compensate for respiration, raises the possibility that cytochrome d provides TBO resistance by promoting an anoxic intracellular environment. If true, cytochrome d should fail to protect the cells when grown in air-saturated liquid cultures in the presence of TBO, because local oxygen concentrations cannot be depleted. To test this, the wild type, the arcA mutant, and the arcA mutant strain complemented with cydAB-carrying plasmid pNG2 were inoculated in tryptone broth in the presence or absence of TBO, and their growth was monitored by counting the number of CFU of each culture. It was found that neither strain was able to grow in air-saturated liquid cultures in the presence of TBO (Fig. (Fig.4D).4D). Therefore, it can be concluded that cytochrome d protects the cells against TBO on TTA plates by creating an anaerobic intracellular environment.
Because the ectopic expression of cytochrome d plays a central role in conferring TBO resistance to arc mutants, we tested whether aerobic cydAB expression is elevated in the suppressor mutants compared to the arcA::kan mother cells. To this end, the wild-type strain, the arcA mutant, and the mutant suppressor strains were grown aerobically to mid-exponential growth phase, and the amount of cytochrome d was measured by spectroscopic analysis of the membrane fractions (Fig. (Fig.5).5). It was found that membranes of the wild-type strain contained circa 60 pmol of cytochrome d per mg of protein, whereas no cytochrome d in the membranes of the arcA mutants was detected. As expected, membranes of the arcA mutant harboring either pNG2 or pMX515Eco contained amounts of cytochrome d ~2.5-fold larger than the ones of the wild-type strain. Finally, a considerable amount of cytochrome d, ranging from 5 to 120 pmol/mg of protein, was detected in all suppressor strains (Fig. (Fig.5).5). It thus appears that the suppression event(s) leads to an ArcA-independent expression of cydAB.
It has been demonstrated previously that the expression of cydAB is negatively regulated by the histone-like protein H-NS (24) and the transcriptional regulator Fnr (56). Therefore, the ArcA-independent expression of cydAB in the suppressor mutants might be a result of its derepression, due to a possible inactivation of H-NS. The transcriptional regulator Fnr was discarded as a candidate because its activity is practically null in the presence of molecular oxygen (36, 53). To test whether inactivation of H-NS suppresses the ΔarcA growth phenotype by derepressing expression of the cydAB operon, we created a ΔarcA Δh-ns double mutant strain and tested its ability to grow on TTA plates. No growth was observed (data not shown), thereby eliminating the possibility of H-NS inactivation as an explanation for the ArcA-independent expression of cydAB.
We then examined whether the ArcA-independent expression of cydAB in the suppressor mutants is a result of mutations in the complex promoter region of the cydAB operon. To this end, the cydAB promoter region of each suppressor strain was PCR amplified using the forward primers cydpromFw700, cydpromFw1600, cydpromFw2100, or cydpromFw4000 and the reverse primer cydpromRev. Curiously, only the reactions with DNA from strains IFCS7, -8, and -13 yielded a PCR product of the expected size, whereas the reactions with DNA from all other suppressor strains yielded specific but variably sized PCR products. Sequencing analyses of the obtained PCR products revealed that suppressor strains IFCS7, -8, and -13 possessed a wild-type promoter sequence, whereas IS elements (IS1, IS2, and IS5) were integrated into the regulatory promoter region of cydAB of all other suppressor strains (Table (Table22 and Fig. Fig.6A).6A). Noteworthy, strains IFCS7, -8, and -13, which possess a wild-type promoter sequence together with IFCS16, which has an IS5 integrated at approximately position −500 relative to the start codon of cydAB, were the ones with the smallest amount of spectroscopically detectable cytochrome d (5 to 20 pmol/mg of protein). Moreover, it was observed that the promoter of cydAB in strains IFCS1 and IFCS9 to -12, in addition to the insertion of IS1, suffered deletions of 800 to 2,850 bp upstream of the positions of insertion (Table (Table22 and Fig. Fig.6A6A).
Subsequent analysis of the new generated sequences, using the Neural Network Promoter Prediction software (47), revealed that potential chimeric promoters were generated for suppressor strain IFCS1, -2 to -6, -9 to -12, and -14 to -16 (Fig. (Fig.6B).6B). These hybrid promoters combine a −35 region provided by the IS with an endogenous −10 region, thereby enabling the ArcA-independent expression of cytochrome d. However, no explanation can be provided for strains IFCS7, -8, and -13, which possess a wild-type promoter. One possibility might be that a transcriptional factor has been altered in such a way that it now is able to activate expression of cytochrome d. Alternatively, a cytochrome d-independent mechanism might be responsible for the suppression of the dye sensitivity in these mutant suppressor strains.
Almost 30 years ago, arcA was discovered as a gene with different phenotypic properties, and therefore, it was designated dye (8), fexA (39), msp (9), seg (30), or sfrA (6). The designation dye described the growth sensitivity of these mutants to redox dyes, such as toluidine blue and methylene blue. Later on, a similar phenotype was described for mutants of arcB (32) and cydAB (15). Redox dyes, such as TBO and methylene blue, in the presence of light are able to produce ROS that were suggested to be the cause for the dye sensitivity of the arcA mutants (50). However, the results presented in this study indicate that the dye phenotype of arc mutants but also the TBO-dependent production of ROS is light independent. Also, they demonstrate that the overproduction of the antioxidant enzymes HPI, HPII, SodA, HMP, and AhpCF or the carotenoid biosynthetic enzymes from Erwinia herbicola, which are known to provide protection to oxidative stress given by H2O2 and near-UV in recombinant E. coli strains (5, 51), are not able to suppress the TBO-sensitive growth of arc and cydAB mutants. Although, these results do not support the suggestion that ROS formation is the main cause for the dye phenotype, the possibility of the participation of ROS in the dye phenotype cannot be discarded. Nevertheless, the cellular responses against ROS do not appear sufficient to cope with this stress.
Yet, our results demonstrate that cytochrome d plays a vital role in the TBO resistance of E. coli. Cytochrome d, used mainly under oxygen-limiting growth, and cytochrome o, predominating under highly aerated growth, are the two terminal quinol oxidases in the respiratory chain of E. coli (31). Despite that both these enzymes catalyze the same reaction, that is, the reduction of oxygen to water, they differ in that cytochrome d has high affinity to O2 and a low Vmax, whereas cytochrome o has low affinity to O2 and a high Vmax (13, 14, 48). Now, our results demonstrate that expression of cytochrome d but not cytochrome o suppresses the TBO growth-sensitive phenotype of the arc mutants. Moreover, they clearly demonstrate that the high affinity to O2 is the key characteristic of cytochrome d that explains the dye phenotype of the arc and cydAB mutants. Therefore, it is tempting to speculate that cytochrome d serves as a potent O2 scavenger able to generate an anoxic intracellular environment, thereby restricting TBO-generated stress by not permitting TBO reoxidation. In such a scenario, the reducing power of the cell would be sufficient to reduce TBO and abolish its detrimental effects on cell growth. Moreover, the fact that the dye phenotype is observed only in the presence of oxygen raises the possibility that TBO perturbs the electron transport chain by diverting electrons to redox cycling, thereby restricting energy production. This is supported by the previous observation demonstrating that cyanide-treated E. coli membranes that are unable to respire recover oxygen consumption in the presence of TBO (50) and also by the fact that the presence of glucose reverts the TBO-dependent growth defect of the arc mutants. Interestingly, a protective role for cytochrome d in generating an intracellular anaerobic environment has previously been observed with Azotobacter vinelandii, which uses the terminal oxidase to protect nitrogenase, an oxygen-sensitive enzyme, enabling nitrogen fixation even during aerobic growth (35).
The expression of the cydAB operon is negatively regulated by the histone-like protein H-NS and the transcriptional regulator Fnr and positively regulated by the ArcA/ArcB two-component system (23, 24, 40, 56). In agreement, we observed significant amounts of spectroscopically detectable cytochrome d with a wild-type strain but not with the arcA mutant. Thus, the TBO-dependent growth defect of the arc mutants might be attributed to the lack of cytochrome d. This is supported by the fact that ectopic expression of cydAB suppresses the growth defect of the arc mutants and further supported by the fact that considerable amounts of cytochrome d were expressed in 13 out of 16 arcA suppressor mutants. ArcA-independent cydAB expression in the suppressor strains appears to be achieved by the integration of IS in different locations of the cydAB promoter region. The activation of expression of bacterial silent genes (cryptic genes) by insertion of IS elements is reported widely. In general, transposable elements contribute positively to the fitness of the cells by activating catabolic operons that are otherwise silent, and their transposition rates increase under such conditions, such that the activation of those operons is beneficial (27). Examples include the E. coli operons bgl, cel, and asc, which code for enzymes involved in the utilization of β-glucoside sugars (28). Noteworthy, similarly to the case reported here, the insertion position of the IS in the cryptic operons bgl and cel is not at a single site but confined to a region of 223 (52) and 108 bp (44), respectively. However, whether the presence of TBO itself enhances the occurrence of IS transposition remains unknown.
It would be interesting to identify the suppressor gene(s) in strains IFCS7, IFCS8, and IFCS13, as they may provide further insights into the mechanism(s) by which bacteria copes with stress generated by redox dyes, such as TBO and methylene blue.
We thank Raul Arredondo-Peter for plasmids prHb1 and pLba1, NBRP (NIG, Japan) for h-ns, katG, katE, and sodA E. coli mutants, Claudia Rodriguez for technical assistance, Diego Gonzalez Halphen and Bertha Michel for helpful discussions and for critically reading the manuscript, and the Unidad de Biología Molecular from the Instituto de Fisiología Celular, Universidad Nacional Autónoma de México for oligonucleotide synthesis and sequencing.
This work was supported by grants 37342-N from the Consejo Nacional de Ciencia y Tecnología (CONACyT) and IN221106/17 from DGAPA-PAPIIT, UNAM.
Published ahead of print on 6 November 2009.