Flavodoxin mutants of Escherichia coli K-12

The flavodoxins are flavin mononucleotide-containing electron transferases. Flavodoxin I has been presumed to be the only flavodoxin of Escherichia coli, and its gene, fldA, is known to belong to the soxRS (superoxide response) oxidative stress regulon. An insertion mutation of fldA was constructed and was lethal under both aerobic and anaerobic conditions; only cells that also had an intact (fldA(+)) allele could carry it. A second flavodoxin, flavodoxin II, was postulated, based on the sequence of its gene, fldB. Unlike the fldA mutant, an fldB insertion mutant is a viable prototroph in the presence or absence of oxygen. A high-copy-number fldB(+) plasmid did not complement the fldA mutation. Therefore, there must be a vital function for which FldB cannot substitute for flavodoxin I. An fldB-lacZ fusion was not induced by H(2)O(2) and is therefore not a member of the oxyR regulon. However, it displayed a soxS-dependent induction by paraquat (methyl viologen), and the fldB gene is preceded by two overlapping regions that resemble known soxS binding sites. The fldB insertion mutant did not have an increased sensitivity to the effects of paraquat on either cellular viability or the expression of a soxS-lacZ fusion. Therefore, fldB is a new member of the soxRS (superoxide response) regulon, a group of genes that is induced primarily by univalent oxidants and redox cycling compounds. However, the reactions in which flavodoxin II participates and its role during oxidative stress are unknown.     

Hypochlorous acid activates the heat shock and soxRS systems of Escherichia coli.

A series of plasmids, containing fusions of different stress promoters to lux reporter genes, was used in an attempt to monitor the defense circuits activated upon exposure of Escherichia coli to sublethal doses of free chlorine. A significant level of activation was exhibited by promoters of three heat shock genes (grpE, dnaK, and lon), in an rpoH-dependent manner. The promoter of micF, a gene under the control of the soxRS regulon, was also strongly induced, but not in a soxR mutant. This induction was not affected by sodA and sodB mutations, implying that it did not involve oxygen radical activity. Free-chlorine activation of both heat shock and soxRS regulons required an exposure of less then I s in duration. The oxyR or the SOS regulons were apparently not induced by free chlorine (as judged by lack of activation of katG and recA, respectively), and neither was the universal stress (uspA) protein.     

Effects of overproduction of superoxide dismutase on the toxicity of paraquat toward Escherichia coli

Gross overproduction of the manganese-containing superoxide dismutase in Escherichia coli, by virtue of a multicopy plasmid bearing the sodA gene, decreases enumeration on paraquat-containing agar plates. This reflects growth inhibition, not lethality, since cells on these plates can be rescued by exclusion of dioxygen. Growth in liquid medium revealed that the control strain adapted to growth in the presence of paraquat more rapidly than did the overproducer. Glucose-6-phosphate dehydrogenase, taken as a representative of the superoxide-inducible soxR regulon, was induced during exposure to paraquat to a much greater extent in the control than in the superoxide dismutase-over-producing strain. These results support the view that overproduction of superoxide dismutase interferes with induction of the soxR regulon and thus prevents a balanced adaptation to the multiple aspects of the toxicity of aerobic paraquat.     

Activation of multiple antibiotic resistance and binding of stress-inducible promoters by Escherichia coli Rob protein.

Multiple antibiotic resistance in Escherichia coli can be mediated by induction of the SoxS or MarA protein, triggered by oxygen radicals (in the soxRS regulon) or certain antibiotics (in the marRAB regulon), respectively. These small proteins (SoxS, 107 residues; MarA, 127 residues) are homologous to the C terminus of the XylS-AraC family of proteins and are more closely related to a approximately 100-residue segment in the N terminus of Rob protein, which binds the right arm of the replication origin, oriC. We investigated whether the SoxS-MarA homology in Rob might extend to the regulation of some of the same inducible genes. Overexpression of Rob indeed conferred multiple antibiotic resistance similar to that known for SoxS and MarA (against chloramphenicol, tetracycline, nalidixic acid, and puromycin), as well as resistance to the superoxide-generating compound phenazine methosulfate. The Rob-induced antibiotic resistance depended only partially on the micF antisense RNA that down-regulates the OmpF outer membrane porin to limit antibiotic uptake. Similar antibiotic resistance was conferred by expression of a Rob fragment containing only the N-terminal 123 residues that constitute the SoxS-MarA homology. Both intact Rob and the N-terminal fragment activated expression of stress genes (inaA, fumC, sodA) but with a pattern distinct from that found for SoxS and MarA. Purified Rob protein bound a DNA fragment containing the micF promoter (50% bound at approximately 10(-9) M Rob) as strongly as it did oriC, and it bound more weakly to DNA containing the sodA, nfo, or zwf promoter (50% bound at 10(-8) to 10(-7) M). Rob formed multiple DNA-protein complexes with these fragments, as seen previously for SoxS. These data point to a DNA-binding gene activator module used in different protein contexts.     

Isolation and characterization of Escherichia coli strains containing new gene fusions (soi::lacZ) inducible by superoxide radicals.

Gene fusions in Escherichia coli that showed increased beta-galactosidase expression in response to treatment with a superoxide radical (O2-) generator, methyl viologen (MV), were obtained. These fusions were constructed by using a Mud(Ap lac) phage to insert the lactose structural genes randomly into the E. coli chromosome. Ampicillin-resistant colonies were screened for increased expression of beta-galactosidase on X-Gal (5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside) plates containing MV at 1.25 micrograms/ml. Other O2- generators, menadione and plumbagin, also induced beta-galactosidase activity in these fusion strains. The induction by these drugs occurred only under aerobic conditions. Hyperoxygenation also elicited an induction of the fusions. On the other hand, no significant induction was observed with hydrogen peroxide and cumene hydroperoxide. The induction of these fusions by MV was not dependent on the peroxide stress control mediated by the oxyR gene or on the recA-dependent SOS system. These fusions were named soi (superoxide inducible)::lacZ. The induction of beta-galactosidase was significantly reduced by introducing a soxS::Tn10 locus into the fusion strains, indicating that the soi genes are members of the soxRS regulon. Five of the fusions were located in 6 to 26 min of the E. coli genetic map, while three fusions were located in 26 to 36 min, indicating that these fusions are not related to genes already known to be inducible by O2- under the control of soxRS. At least five mutants containing the soi::lacZ fusion were more sensitive to MV and menadione than the wild-type strain, suggesting that the products of these soi genes play an important role in protection against oxidative stress.     

An oxidative stress-specific bacterial cell array chip for toxicity analysis

An oxidative stress-specific bacterial cell array chip was fabricated and implemented in the analysis of various different chemicals. The chip consisted of twelve toxicity responsive strains that respond specifically to different oxidative toxicities such as the generation of the superoxide radical, except for strain EBMalK, which was included as a negative control. Each bioluminescent strain carried a fusion of a stress gene promoter (sodA, pqi-5, soxR, fumC, soxS, inaA, hmp, malK, katG, zwf, fpr or pgi) to the bacterial lux reporter genes. A total of nine chemicals were selected to exhibit the capabilities of this array when analyzing different oxidative toxicities. Each of the chemicals were categorized according to their structure and their ability to form radicals in vivo: (I) paraquat, an active radical producer, (II) structural analogs of paraquat that produce radicals, (III) chemicals that are distinct from paraquat but still produce radicals and (IV) chemicals having similar structures as paraquat but do not produce radicals. The results found that each strain was responsive to one or more of the compounds tested but, as a definitive factor, the responses from the chip were dependent upon the production of radicals, i.e., the strains were unresponsive to compounds that were similar in structure to paraquat but lacked the ability to generate radicals. The specificity of the strains used in the chip was also demonstrated by their ability to discriminate between the superoxide radical and hydrogen peroxide. Therefore, this cell array chip could be implemented in characterizing and understanding the toxic impacts of newly synthesized chemicals and drugs in terms of toxicity classification and the nature of oxidative damage experienced by cells.