SoxRS-Mediated Lipopolysaccharide Modification Enhances Resistance against Multiple Drugs in Escherichia coli

Strains, phages, and plasmids.All strains, phages, and plasmids used in this study are listed in Table 1. DH5α was used for the cloning of recombinant DNA, and GC4468 was used as a host strain to harbor chromosomal copies of various lacZ fusions and mutations. The promoter-probing plasmid pRS415, which contains the promoterless lacZYA genes, was used for the construction of promoter-lacZ fusions.

Culture conditions.LB medium (1% tryptone, 1% NaCl, and 0.5% yeast extract) was used for routine bacterial culture. Antibiotics were used at the following concentrations: ampicillin, 50 μg/ml; kanamycin, 25 μg/ml; tetracycline, 20 μg/ml; and chloramphenicol, 20 μg/ml. To determine the effects of various chemicals on gene expression, the lacZ fusion strains were grown in LB broth to an optical density at 600 nm (OD₆₀₀) of 0.2 with vigorous shaking, treated with the agents at the indicated concentrations for 1 h, and assayed for β-galactosidase activity as described by Miller (22).

DNA and RNA manipulation.Reactions for DNA manipulation were carried out according to standard protocols or as recommended by the manufacturers. We always confirmed the final sequences of the constructs after every recombination process with DNA. Cellular RNA was extracted with Ultraspec-II total RNA isolation kits (Biotecx Laboratories Inc.) as recommended by the manufacturer, except that the cells were first treated with lysozyme (4 mg/ml) in 50 mM glucose, 25 mM Tris-HCl (pH 8.0), and 10 mM EDTA for 5 min on ice.

Construction of single-copy lacZ fusions and P1_vir transduction. waaY promoters of various lengths were cloned into pRS415 (pJH97, pJH98, and pJH99) and transformed into GC4468. The resulting transformants were infected with phage λRZ5 to bring about homologous recombination between the plasmid and phage DNAs in vivo, as described previously (32). The recombinant phage were then lysogenized into GC4468 at the att site to make single-copy lysogens (JH101, JH103, and JH104), which were screened by the lowest basal level of β-galactosidase activity. Introduction of various mutations into these lacZ fusion strains was done through P1_vir transduction as previously described (31). The soxRS, oxyR, and waaYZ mutations were transduced from BW829 (sox-8::cat), BW831 (soxS3::Tn10), BW847 (soxR4::cat), BW900 (soxR9::cat), GSO18 (ΔoxyR::kan), JWK5249 (ΔmarA::kan), JH1001 (ΔwaaY::kan), and JH1003 (ΔwaaYZ::kan) into the recipient strains, such as GC4468 and soxS-lacZ (MS1343), rpoH P3-lacZ (CAG16037), and waaY-lacZ (JH101 and JH103) mutants.

Primer extension and Northern hybridization.Primer extension and Northern analysis were done as described by Sambrook et al. (28). For primer extension, an oligonucleotide (5′-AATAATTGATTTCGCATCTCGTGG-3′) (see Fig. 2B) complementary to the downstream region of the putative +1 site was labeled at the 5′ end with [γ-³²P]ATP (Amersham) by T4 polynucleotide kinase. One hundred micrograms of RNA and 5′-end-labeled primer (10⁴ to 10⁵ cpm) were hybridized, and cDNA was synthesized by avian myeloblastosis virus reverse transcriptase (Promega). The resulting cDNAs were analyzed by electrophoresis on a 6% polyacrylamide gel containing 7 M urea. For Northern analysis, denatured RNA samples were electrophoresed in a 1% agarose gel containing formaldehyde and transferred to a Hybond-N+ membrane (Amersham). The 809-bp SspI-SspI fragment covering the whole waaY gene and upstream 303-bp HindIII-ScaI fragment (Fig. 1A) were used as probes that were labeled by the random priming method.

• Open in new tab
• Download powerpoint

FIG. 1.

waa gene cluster of E. coli K-12, location of the paraquat-inducible promoter, and structure of the LPS core region and action site of WaaY. (A) waa gene cluster and waaYZU region. The small arrow in waaR indicates the location of the H73 promoter. (B) Structure of the LPS core region of E. coli K-12 and genes involved in biosynthesis, shown at their approximate sites of action. P, phosphate; Hep, heptose; Glc, glucose; Gal, galactose; KDO, 3-deoxy-d-manno-2-octulosonic acid; PEtN, 2-aminoethyl phosphate. The proposed site of waaY action is circled.

Gel mobility shift assay.The 97-bp AluI-AluI fragment of the waaY promoter (Fig. 2B) was size fractionated from an agarose gel and labeled with [α-³²P]dATP, using Klenow DNA polymerase. DNA binding reaction mixtures (20 μl) contained 10 mM Tris-HCl (pH 8.0), 75 mM KCl, 2 mM dithiothreitol, 10% (vol/vol) glycerol, 1 fmol of ³²P-labeled probe, 10 ng of poly(dI-dC):poly(dI-dC) (Sigma), and the indicated amounts of purified SoxS. The reaction mixture was electrophoresed in a 5% polyacrylamide gel (20 mM Tris-HCl, pH 8.0, 3 mM sodium acetate, 1 mM EDTA) and visualized by autoradiography on X-ray film or by phosphorimaging. HaeIII-digested fragments of pGEM-3zf(+) (Promega) and unlabeled probe were used as nonspecific and specific competitors, respectively. The SoxS protein used in this experiment was a native form which was overexpressed and purified as described by Li and Demple, using the SoxS expression plasmid pKOXS (20).

• Open in new tab
• Download powerpoint

FIG. 2.

Primer extension analysis and sequence of waaYp region. (A) Primer extension was carried out with total RNAs from paraquat-treated (0.8 mM) and untreated wild-type cells. +1 sites are indicated by arrows and asterisks. Growth conditions for RNA extraction were the same as those for the β-galactosidase assay. Induction was quantified as 19-fold by phosphorimager analysis (Bio-Rad). (B) Sequence of the waaY promoter region and details.

Gene disruption.Disruption of the waaY, waaZ, and waaYZ genes was done as described by Nagano et al. (25). The internal regions of the waaY, waaZ, and waaYZ genes (SspI-BssHII, ApaI-NsiI, and SspI-NsiI fragments, respectively) were displaced in vitro with a kanamycin/bleomycin resistance cassette from pUC4-KIXX (Pharmacia). The recombinant plasmids were transformed into a temperature-sensitive polA mutant (CP367) in which only cointegrates can form colonies at 42°C on antibiotic-containing plates. The cointegrates isolated at 42°C were further grown without antibiotics at 30°C for five consecutive generations to allow excision of the plasmid body by a second recombination event. The desired mutations were selected from Kan^r Amp^s colonies and transferred from the CP367 to GC4468 background by P1_vir transduction. The correct gene replacement in all mutants was confirmed by Southern hybridization. In-frame deletion of waaY was done as described by Baba et al. (6). The waaY::kan locus was transferred from JWK3600 to GC4468 by P1_vir transduction. The kanamycin resistance cassette was excised from the flanking FRT site by using an FLP helper plasmid (pCP20; Amp^r Chl^r), which was then removed by cultivation without antibiotics at 37°C. Colonies that lost all resistance were selected by replica plating. The correct in-frame deletion of the waaY gene was confirmed by PCR and sequencing.

LPS extraction and electrophoresis.LPS was extracted from whole cells by the hot phenol method as described by Chart (10). An overnight culture of cells was harvested, washed, and resuspended in TE buffer (10 mM Tris-Cl, pH 8.0, 1 mM EDTA). An equal volume of hot phenol (68°C) was added, and the culture was mixed carefully to form a uniform “milky” emulsion and incubated at 68°C for 15 min. After centrifugation at 3,000 × g for 45 min, the upper, aqueous phase was collected. Further extraction by cold phenol was done twice to remove remaining proteins and lipids. LPS in the aqueous pool was electrophoresed by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized by silver staining.

PM.Phenotypic microarray (PM) analysis was performed by Biolog Inc., essentially as described elsewhere (9). Briefly, this analysis uses a redox chemistry employing cell respiration as a reporter. If the treatment of a drug is strongly positive for cell growth, the cells respire actively, reducing a tetrazolium dye and forming a strong color. If it is weakly positive or negative, the respiration is slowed or stopped, and less color or no color is formed. The differences in colorimetric intensities of the wild-type and mutant cells were recorded, with the colorimetric scores for the wild-type cells set as a reference. Thus, positive differences mean that the drug treatment enhanced mutant cell growth, and negative scores mean that the mutant cells became sensitive to the drug. In our study, the waaYZ mutant (JH1003) was compared pairwise with the isogenic parental strain (GC4468). The consensus results were taken from two independent runs with each of the two strains.

Oxidant-responsive promoter of waaY.Among paraquat-inducible promoters that were screened through cloning with the promoter-probing plasmid pRS415 (18, 19), a strongly induced one (H73) was located upstream of the waaY (formerly rfaY) gene, in the middle of the LPS core biosynthetic gene cluster (Fig. 1A). The waaQGPSBORYZU cluster has been suggested to constitute a single transcription unit, based on genetic studies (13, 29), and the WaaY protein has been suggested to be a kinase that phosphorylates HepII, a heptose of the LPS inner core (Fig. 1B) (40). The syntenic organization of the waaY gene is conserved in most E. coli strains, Shigella spp., and Salmonella spp. (13).

To precisely locate the inducible transcription start site of the waaY promoter (waaYp), we performed primer extension analysis of waaY transcripts in vivo. The results showed a dramatic induction of the transcripts by paraquat, and the primary start site was mapped 171 nucleotides upstream from the start codon for WaaY (Fig. 2A). Minor RNAs of shorter lengths might reflect either minor downstream start sites or non-full-length extension. From the major transcriptional start site (+1), the −35 and −10 elements of the paraquat-responsive waaYp were predicted. A consensus sequence for SoxS binding (soxbox) was located adjacent to the −35 element (Fig. 2B). The inducibility of waaYp by various oxidants was examined by monitoring β-galactosidase activity from a waaYp-lacZ fusion integrated into the chromosome as a single copy (JH101 strain). As demonstrated in Fig. 3, waaYp responded dramatically to superoxide generators, such as paraquat, lawsone, menadione, and plumbagin, and much less to H₂O₂ and ethanol. It did not respond to heat (42 to 50°C) or a reducing agent (dithiothreiol) (data not shown).

• Open in new tab
• Download powerpoint

FIG. 3.

Response of waaYp to various chemicals. Single-copy waaYp-lacZ fusion cells (JH101) in early exponential phase (OD₆₀₀ = 0.2) were treated aerobically with various concentrations of chemicals for 1 h at 37°C and then assayed for β-galactosidase activity (Miller units). The concentration of ethanol is indicated separately (%). These are the most representative results from multiple measurements.

SoxRS-dependent regulation of waaY transcription.Since waaYp responded almost exclusively to superoxide generators, we examined whether it is controlled by SoxRS. For this purpose, we introduced various soxRS and oxyR mutant loci into JH101 and estimated the expression of waaYp-lacZ (Fig. 4). As expected, the soxRS mutations (sox-8 and soxS3) abolished the paraquat induction of waaYp, whereas a soxR constitutive mutation (soxR4) elevated the expression in the absence of paraquat. The oxyR mutation did not affect induction. Therefore, it is clear that waaYp is under the control of SoxRS. To estimate the size of the waaY transcript, we performed Northern analysis (Fig. 4B). The size of the induced RNA, which was observed in the wild type but not in the soxRS mutant, was about 1.6 kb. This is much larger than the length of the waaY coding region and is predicted to encompass most of the waaZ gene. Since a probe upstream of waaYp (303-bp HindIII-ScaI fragment) (Fig. 1A) failed to detect paraquat-inducible transcripts (data not shown), waaYp most likely produces a waaY (or, at maximum, waaYZ) transcript under conditions of superoxide stress.

• Open in new tab
• Download powerpoint

FIG. 4.

soxRS-dependent induction of waaY. (A) waaYp activity was assayed in various soxRS and oxyR mutant backgrounds, using single-copy lacZ fusions. The isogenic wild type (WT) is JH101, and all mutants are derivatives of JH101: the sox-8::cat mutant (JH201) has a deletion mutation of soxRS, the soxS3::Tn10 mutant (JH301) has an insertion mutation of soxS, the soxR4::cat mutant (JH401) has a constitutive soxR mutation, and the oxyR::kan mutant (JH501) has an insertion mutation of oxyR. β-Galactosidase activity was assayed after paraquat (PQ) treatment for 1 h with vigorous aeration. (B) Northern analysis was done with RNAs from the wild type (GC4468) and the soxRS mutant (BW829). An SspI-SspI fragment including the entire waaY open reading frame was used as a probe (Fig. 1A). A 1.6-kb transcript was induced 11.3-fold (quantified by phosphorimaging). (C) A 98-bp AluI-AluI fragment, from positions −101 to −4 (Fig. 2B), was used as a probe for a gel shift assay with increasing amounts of purified SoxS. Lanes 1 to 3, 0, 65, and 130 ng of purified SoxS, respectively (0, 250, and 500 nM, respectively); lanes 4 to 8, 130 ng of SoxS (500 nM), with nonspecific competitor in 65-, 130-, and 650-fold molar excess (lanes 4 to 6, respectively) or with specific competitor in 5- and 10-fold molar excess (lanes 7 and 8, respectively) over the labeled probe.

We then examined the direct binding of purified SoxS protein to the DNA fragment containing the putative soxbox and found that it binds specifically to the waaYp fragment (Fig. 4C). An in vitro transcription assay also demonstrated that SoxS acts as a sole activator for RNA polymerase containing σ⁷⁰ to transcribe waaYp (data not shown).

waaYZ mutants become sensitive to menadione and plumbagin but not to paraquat.To find out the role of waaYZ genes in the oxidative stress response, we constructed ΔwaaY, ΔwaaZ, and ΔwaaYZ mutants and investigated their sensitivity to superoxide generators. While both ΔwaaY and ΔwaaYZ mutants became sensitive to menadione and plumbagin, the ΔwaaY mutant was a bit less sensitive than the ΔwaaYZ mutant, and the ΔwaaZ mutant was only slightly sensitive to these agents (data not shown). This indicated that waaY might play a major role in protection, but there might be a minor involvement of downstream genes, either waaZ or possibly waaU. To better address this possibility, we constructed an in-frame deletion mutant of waaY (ΔwaaY_inf ) to exclude polar effects and performed a complementation test with a plasmid encoding WaaY or WaaYZ. Similar results were obtained, showing that the ΔwaaY_inf mutant had significant sensitivity to menadione and plumbagin, but to a lesser extent than that of the ΔwaaYZ mutant (Fig. 5A). This phenotype was complemented by the plasmid carrying waaY, confirming the major role of WaaY, but the ΔwaaY_inf mutant sometimes showed partial complementation (Fig. 5A). We suggest that waaY plays an important role in resistance, but we do not rule out the contribution of downstream waaZ for full activity.

• Open in new tab
• Download powerpoint

FIG. 5.

Sensitivity of waaY mutant to superoxide stress. (A) Different amounts of wild-type (GC4468), ΔwaaYZ (JH1003), and ΔwaaY_inf cells were grown on LB plates containing superoxide generators for 16 h, and growth was compared. For complementation, a plasmid expressing WaaY (pWaaY) was transformed into the ΔwaaYZ and ΔwaaY_inf mutants, and the growth of the transformants was compared. Vec, empty plasmid control. (B) soxSp-lacZ fusions in the wild-type (MS1343) and ΔwaaYZ (JH2003) backgrounds were assayed for β-galactosidase activity to monitor SoxR activity after menadione, plumbagin, and paraquat treatment for 1 h. For complementation, a plasmid carrying waaYZ (pWaaYZ) was transformed into the wild-type and ΔwaaYZ strains, and cells were assayed for β-galactosidase activity. To see the SoxR dependence of the activation, a soxR null mutation (BW900) was introduced into both the wild-type and ΔwaaYZ strains and assayed for β-galactosidase activity.

Unexpectedly, waaYZ mutations did not increase the sensitivity to paraquat, and the ΔwaaY_inf mutant was even slightly resistant to paraquat (Fig. 5A). We do not understand this unexpected resistance of the ΔwaaY_inf mutant, but the selective sensitivity of the mutants could have arisen from a differential susceptibility toward the drugs, not from the loss of some defense mechanism toward superoxide in general. To test this hypothesis, we examined whether the waaY mutant exhibited differential sensitivity toward superoxide-generating drugs in inducing soxRS target genes. For this purpose, the ΔwaaYZ mutant allele was transduced into a soxSp-lacZ reporter strain, MS1343 (17), and the inducibility of LacZ activity was measured. Compared with the wild type, the ΔwaaYZ mutant became more sensitized to activate the SoxRS system in response to menadione and plumbagin, inducing soxSp-lacZ to higher levels at lower concentrations of oxidants (Fig. 5B). This induction was totally dependent on SoxR, since all of the response disappeared by introducing a soxR null mutation. However, upon paraquat treatment, there was no difference between the waaYZ mutant and the wild type (Fig. 5B). This coincides with no increase in susceptibility to paraquat. These results strongly support the hypothesis that the waaYZ genes determine susceptibility toward different drugs, not toward superoxide radical itself.

Changes in outer membrane structure, including LPS.Since WaaY and WaaZ are components of the LPS core synthetic system, disruption of their genes is likely to cause some alteration in the outer membrane structure, which can be monitored through activation of envelope stress-responsive genes. To do this, the ΔwaaYZ allele was introduced into a σ^E-dependent reporter strain containing an rpoH P3-lacZ fusion (11). The rpoH P3 promoter is recognized by σ^E, an ECF sigma factor that is activated in response to various forms of extracytoplasmic stresses, including abnormality in LPS (1, 24, 34). Compared with the isogenic parental strain (CAG16037), an about fivefold increase in rpoH P3 expression was observed in the ΔwaaYZ mutant (Fig. 6A), suggesting the presence of envelope stress in the mutant. To determine whether the mutant was indeed altered in LPS structure, we extracted LPSs from the ΔwaaY and ΔwaaYZ mutants and analyzed them by SDS-PAGE. LPSs from the mutants showed faster migration, revealing an alteration in LPSs from the mutants (Fig. 6B). These results indicated that WaaYZ-induced modification of LPS structure could serve as a barrier against certain oxidants and drugs. Since WaaY was suggested to be a kinase that phosphorylates heptose II in the LPS inner core (13, 40), a change in phosphorylation status might be one reason for this altered electrophoretic mobility.

• Open in new tab
• Download powerpoint

FIG. 6.

Changes in outer membrane structure of waaYZ mutant. (A) β-Galactosidase activity of the envelope stress reporter strain (rpoH P3-lacZ single-copy fusion) was assayed in the wild-type background (CAG16037) or ΔwaaYZ mutant background (JH3003). (B) LPSs were extracted from the wild-type (wt) (GC4468), ΔwaaY (JH1001), and ΔwaaYZ (JH1003) strains, separated by 15% SDS-PAGE, and visualized by silver staining.

WaaYZ confers a protective function against a large number of toxic drugs.If LPS modification serves as a barrier to certain drugs, the waaYZ mutant should be sensitive to a wider range of drugs in addition to superoxide generators. Consistently, the waaYZ mutant was sensitive to lawsone and some antibiotics, including chloramphenicol and 8-hydroxyquinoline (data not shown). For further examination of this possibility, we performed PM analysis of the waaYZ mutant to find out its phenotype under broader growth conditions (9, 42). Among about 2,000 test conditions, including a wide variety of C, N, P, and S sources, a wide pH range, and chemical agents that disrupt various biological pathways, we found that the ΔwaaYZ mutant showed dramatic sensitivity to a large number of chemicals compared with the reference strain (GC4468). Table 2 summarizes the list of chemicals that exhibited a pronounced growth inhibitory effect on the mutant relative to the wild type (with differential growth values of <−150). It was evident that the ΔwaaYZ mutant was sensitive to a number of antibiotics and toxic compounds, in addition to redox cycling chemicals. Therefore, WaaYZ-mediated LPS modification confers protection on E. coli cells against a wide variety of chemicals not restricted to superoxide generators. We observed that the waaYZ mutant became resistant to four antibiotics (bleomycin, phleomycin, dihydrostreptomycin, and neomycin) (not listed in Table 2). This most likely resulted from the presence of the kanamycin/bleomycin resistance cassette in the disrupted waaYZ gene (30) and can be regarded as a good positive control for the validity of the PM assay.

The waaY promoter is also induced by antibiotics through the MarA system.Since we found that WaaY confers resistance against a wide range of antibiotics, we were intrigued to find out whether the gene is inducible through the MarA system, which responds to various drugs and confers resistance against multiple antibiotics. We examined the effect of sodium salicylate, a known inducer of the marA regulon, in the waaYp-lacZ reporter strain with a different mutant background. The results in Fig. 7 demonstrate that sodium salicylate also induces the waaY gene as efficiently as paraquat does, and the induction is dependent on both the SoxRS and MarA systems. This contrasts with induction by paraquat, which was solely dependent on the SoxS system. We think that this partial SoxRS dependence of salicylate induction could be due to intracellular ROS generation by salicylate treatment, as reported recently (16). When the predicted soxbox sequence was deleted, induction by both paraquat and sodium salicylate was completely gone, indicating that MarA and SoxS activation was mediated through the same cis-acting site. Therefore, the waaY gene that is induced by superoxide generators (redox cyclers) is also induced by nonredox cyclers through the MarA system and confers resistance toward a wide range of chemicals.

• Open in new tab
• Download powerpoint

FIG. 7.

waaY promoter is also induced by antibiotics through the mar system. Wild-type single-copy waaYp-lacZ fusion cells (JH103) and derivatives of the ΔsoxRS (JH203) and ΔmarA (JH603) mutants were treated with paraquat (0.1 mM) and sodium salicylate (10 mM) at early exponential phase (OD₆₀₀ = 0.2), and after 1 h of incubation with vigorous aeration, β-galactosidase activity was assayed. For soxbox deletion in the wild-type background, JH104 was used. We confirmed that the soxbox was correctly deleted and the −35 box remained intact by sequencing.