| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Matforsk, Norwegian Food Research Institute, Osloveien 1, N-1430 Ås, Norway
2 The Norwegian University of Life Science (UMB), PO Box 5003, N-1432 Ås, Norway
3 UR454 Microbiologie-Equipe QuaSA, INRA site de Theix, 63122 Saint-Genès Champanelle, France
4 Plate-forme Protéomique, INRA site de Theix, 63122 Saint-Genès Champanelle, France
Correspondence
Solveig Langsrud
solveig.langsrud{at}matforsk.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The array data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE6712.
A supplementary data file is available with the online version of this paper.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The QACs have been actively deployed since the 1930s and one has not seen a resistance development similar to that seen for antibiotics in the same period. Still, there are numerous reports of QAC resistance (Heir et al., 1999; Langsrud et al., 2003; Wright & Gilbert, 1987), most often associated with acquisition, or hyperexpression, of multi-drug efflux pumps (Li & Nikaido, 2004; Tikhonova & Zgurskaya, 2004), which has also been associated with changes in MIC of therapeutically important antibiotics serving as substrates to such pumps (Langsrud et al., 2004; Poole, 2004, 2005). Because of their relatively impermeable outer membrane, Gram-negative bacteria are intrinsically more resistant to QACs than Gram-positives (McDonnell & Russell, 1999), but both Gram-negative and Gram-positive BC-resistant bacteria have been isolated from the food industry (Aase et al., 2000; Heir et al., 1995; Langsrud et al., 2003; Soumet et al., 2005).
Several studies have linked intrinsic/natural resistance of Gram-negative bacteria to tenside-based disinfectants, such as QACs, to the low permeability of the outer membrane, or broad-spectrum efflux systems (Denyer & Maillard, 2002; Nikaido, 2001). Acquisition of resistance in Escherichia coli has been related mainly to changes in the composition of lipopolysaccharide (LPS) and fatty acids in the membrane (Ishikawa et al., 2002; Sakagami et al., 1989). In contrast to antibiotics, which have very specific targets within the bacterial cell, biocides affect multiple cellular components. We have earlier demonstrated that adaptation of E. coli to BC is followed by enhanced efflux activity and may result in cross-resistance to several antibiotics (Langsrud et al., 2004). In this study we describe changes in expression level at the transcriptomic and proteomic level for E. coli K-12 cells gradually adapted to a higher tolerance level to BC. The results reveal that BC treatment may result in superoxide stress in Gram-negative bacteria.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Cultures were stored in 16 % (v/v) glycerol at –80 °C. Unless otherwise stated the strain was cultivated at 37 °C with shaking (200 r.p.m.).
Adaptation to benzalkonium chloride.
E. coli K-12 was inoculated (two drops) into 5 ml TSB with BC in a range of concentrations in 2 µg ml–1 steps. The cultures were incubated overnight (20–24 h) and inspected visually for growth. The tube with highest BC concentration with visible growth was reinoculated (two drops) into tubes with fresh TSB, with a range of BC concentrations in steps of 2 µg ml–1. The cultures were incubated overnight and inspected visually for growth. This procedure was continued for a total of 30–40 days until five reinoculations did not result in growth at higher concentrations of BC. The adaptation process was performed three times, leading to the three biological repeats of adapted strains (A1, A2 and A3). Adapted strains and samples from several steps of the adaptation process were grown overnight in TSB without BC and kept frozen in 16 % glycerol until use. PFGE was run to confirm the identity of the final adapted strain, using a Chef Electrophoresis Cell and Pulswave 760 (Bio-Rad), and essentially following the protocol of Bohm & Karch (1992). All adapted strains were tested for stability of the tolerance level after freezing by four reinoculations and growth overnight in TSB without BC.
Harvesting of cells for RNA, DNA and protein analysis.
Frozen cultures (control, A1, A2 and A3) were inoculated into 5 ml TSB and incubated overnight. One millilitre of culture was inoculated in 99 ml TSB in a 500 ml baffled shake flask and incubated at 37 °C with shaking. Adapted cultures were in addition grown in TSB with 50 µg BC ml–1. Bacterial cells were harvested at mid-exponential phase (
108 c.f.u. ml–1). For protein samples the culture was centrifuged at 5000 g for 10 min and washed three times in TE buffer pH 7.5 (20 mM Tris, 5 mM EDTA and 5 mM MgCl2), resuspended in 1 ml TE buffer and frozen at –80 °C. For DNA arrays, bacterial cells were harvested directly from the shake flask into RNAprotect (Qiagen) to secure stable RNA before cell lysis. One volume of culture (1x108 c.f.u. ml–1) was added to two volumes of RNAprotect, vortexed and incubated at room temperature for 5 min. The sample was then centrifuged at 5000 g for 10 min and the supernatant removed.
Cultures for transcriptional and proteomic analysis were harvested three times on different days, resulting in three control samples and three samples for each adapted strain. Transcriptional analysis was performed for all samples. Protein analysis was performed for controls, A1 and A3 with 4–8 gels per sample.
Protein extraction – total soluble fraction.
The protocols were mainly based on methods described by Molloy et al. (1999). The sample was centrifuged at 10 °C and 7500 g for 10 min. The pellet was resuspended in 1 ml buffer 1 (40 mM Tris pH 9.5, 5 mM MgCl2 and 5 mM EDTA). Then 10 µl DNase I and 10 µl RNase A were added and the tube was vortexed and kept at room temperature for 15 min. The sample was then split into two fractions of 500 µl each and sonicated using a Vibra-cell sonifier (Bioblock Scientific) with a Microsonde l=138 mm and coupler l=90 mm. The sonication was run on ice as five cycles, alternating between 30 s at constant power level 5 and 2 min at power level 5 and 50 % of the active cycle, separated by 1 min on ice between each single step. Five hundred microlitres of buffer 2 (40 mM Tris pH 9.5, 8 M urea and 4 %, w/v, CHAPS) was added to each tube and the tubes vortexed. Just before use, 25 µl 40 % (v/v) ampholyte 3–10 and 20 µl 0.2 M tributylphosphine were added to each tube. The samples were then centrifuged at 10 °C, 13 000 g for 15 min. The supernatant was collected as the cellular protein fraction. The protein concentrations were determined for each fraction by the method of Bradford (1976) using the Bio-Rad protein assay kit with bovine serum albumin as the standard.
Two-dimensional electrophoresis (2-DE).
2-DE was performed essentially according to the method previously described by Folio et al. (2004). For isoelectric focusing, 18 cm pH 3–10 non-linear (NL) immobilized pH gradient (IPG) strips (Bio-Rad) were rehydrated overnight with 400 µl IEF buffer containing 60 µg solubilized proteins (Folio et al., 2004). To improve focusing of protein spots in the gels, 50 % trifluoroethanol was added to the rehydration buffer according to Deshusses et al. (2003).
Proteins were separated using a Multiphor II electrophoresis unit (Amersham Biosciences). After running the first dimension overnight, the IPG strips were equilibrated as previously described (Folio et al., 2004) and then placed on top of 12 % acrylamide gels. The second dimension (SDS-PAGE) was carried out in a Multicell Protean II XL system (Bio-Rad) at 15 mA per gel and 40 V constant for 1 h and then at 150 V constant for about 15 h.
Gels were silver stained according to the procedure of Rabilloud et al. (1992). For the identification of protein spots, semipreparative conditions were applied with 800 µg of proteins and spots were detected using colloidal Coomassie blue stain according to the protocol of Neuhoff et al. (1988).
Gel analysis.
Analysis of gel images was carried out using the softwares Z3 (Compugen) and Melanie 4.0 (GeneBio). 2-DE images of adapted and control E. coli were compared and qualitative and quantitative differences in spot patterns were analysed. Mean values using at least four gels from two independent protein extractions were taken into account for comparative analyses. The molecular masses and pI values were estimated via standard protein markers, and with theoretical IPG pH gradient and theoretical parameters of five identified spots.
Identification of proteins by MALDI-TOF-MS.
Proteins were identified by MALDI-TOF-MS analysis (Proteomic Platform, INRA site de Theix, Clermont-Ferrand, France), using the peptide mass fingerprinting approach. One microlitre each of sample and saturated 
-cyano-4-hydroxycinnamic acid matrix (10 mg ml–1) in 50 % acetonitrile and 0.1 % TFA were mixed onto the target and allowed to dry. Positive-ion MALDI mass spectra were recorded in the reflectron mode of a MALDI-TOF-MS (Voyager DE-Pro; PerSeptive BioSystems), using Voyager software for data collection and analysis. The MS was calibrated with a standard peptide solution (Proteomix; LaserBio Labs). Trypsin autolysis peptides were used for internal calibration of samples. Monoisotopic peptide masses were assigned and used for NCBI database searches with the Mascot software (http://www.matrixscience.com). The following parameters were considered for the searches: a maximum fragment ion mass tolerance of ±25 p.p.m., a maximum of one missed enzymic cleavage, modification of cysteines by carbamidomethylation and possible oxidation of methionine.
Arrays.
Total RNA was isolated using the RNeasy mini kit (Qiagen) according to the manufacturer's recommendations. The concentration of RNA was determined by measuring the absorbance at 260 nm and 280 nm in an Ultrospec 3000 spectrophotometer (Pharmacia Biotech). The ratio between the measured absorbance at the two wavelengths gives an estimate of RNA purity with respect to contamination (Qiagen Bench guide).
RNA degradation was checked with an Agilent 2100 Bioanalyser according to the userguide from Agilent Technologies. Before cDNA labelling, variations in RNA concentrations were adjusted to 1 µg µl–1 for all array hybridizations. Labelled cDNA was prepared according to the Panorama E. coli gene array protocol (Sigma Genosys), with some modifications. Superscript II Rnase H– reverse transcriptase (Invitrogen) and sheared salmon sperm DNA (Ambidon) replaced the equivalent components given in the Panorama protocol. The unincorporated nucleotides were removed with MicroSpinTM G-25 columns (Amersham Pharmacia Biotech), according to the protocol given. Purified labelled cDNAs were hybridized to the Panorama E. coli gene array, according to the protocol supplied. The membranes were wrapped in clear plastic film and exposed to a Storage Phosphor Screen (Amersham Biosciences) for 6–7 days. [33P]dCTP from Amersham Bioscience was used for all hybridizations but one, where the isotope was purchased from Montebello Diagnostics. The arrays were stripped between hybridizations according to the Panorama E. coli gene arrays technical protocol (Sigma Genosys).
The software Imagene (BioDiscovery) was used for quantification of data and the software Genespring (Silicon Genetics) for normalization.
Real-time quantitative PCR (RTQ-PCR).
The quality of the Taqman data was secured through the use of two references, housekeeping genes gapA and accD. The Taqman probes and primers (Table 1) were found using the software SDS 2.2 and purchased from Applied Biosystems. The protocol used was essentially after Rudi et al. (2003). PCR samples and controls were prepared in triplicate. The reaction tubes were 96-Well Optical Reaction Plate and tube caps were Optical Adhesive Covers. Applied Biosystems supplied all consumables. Reactions were performed in the 7900HT Sequence Detection System from Applied Biosystems.
|
, using known reference strains. The method is used in the Swedish (http://www.sva.se), SVARM, Norwegian (http://www.zoonose.no), NORM-VET and Finnish (http://www.eela.fi), FINRES, monitoring programmes (Petersen et al., 2003). Overnight cultures in TSB (37 °C, 200 r.p.m.) were diluted to 104–105 c.f.u. ml–1 and 50 µl was added to each of the 96 wells of the panel containing an antibiotic gradient for 13 different antibiotics. The panels were incubated for 2 days at 37 °C and inspected for visible growth indicated by a pellet at the bottom of the wells after 1 and 2 days. The screening was repeated twice.
Statistics.
The data were log-transformed and evaluated statistically according to the approach of Langsrud (2002, 2005) for analysing designed experiments with multiple responses. The analysis was based on an unbalanced analysis of variance model with the following model terms (degrees of freedom in parentheses): Strain (3), BC conc. (1), Biological repeats (2), StrainxBC conc. (2), Biological repeatsxStrain (6). The presence of the Biological repeatsxStrain interaction complicated the analysis. To ensure that our conclusions were general, we treated Biological repeats as a random factor. This means that the data were analysed according to a mixed-model approach where Strain was tested against Biological repeatsxStrain as its error term. Since the current software for Langsrud's approach (available at http://www.matforsk.no/ola/program.htm) does not handle such models, a specialized routine was programmed in Matlab. To separate the adapted effect from the general Strain difference, the main effect of Strain was decomposed into two parts: Adapted (one degree of freedom) and the remaining part with two degrees of freedom.
With several multiple responses, ordinary significance testing leads to a lot of false significant results. Therefore, the P-values for each gene were adjusted by rotation testing according to a false discovery rate criterion (Moen et al., 2005).
In the analysis of a total of 4033 genes, observed difference needed to be relatively large to be judged as significant. Therefore, in addition to this analysis of (nearly) all genes, we ran an extra analysis with a smaller group of 25 pre-selected genes (acrABEF, cysDK, emrABDKY, fliY, marABR, mltAB, ompFR, soxS, tolC, yciQ, yhfZ, yhiUV). These 25 genes had been chosen from the literature for their relevance to resistance and stress prior to the analyses. In this case it was easier to detect significant results, since these genes did not need to compete with all the other 4008 (randomly varying) genes.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Transcriptional and proteomic analysis
Initial experiments showed that it was difficult to extract sufficient protein from adapted cells pregrown in BC, and only results from adapted cells grown in TSB without BC were analysed further. More than 50 protein spots showed differences in expression between control and adapted strains, but for many of these spots there was high variation in expression between adapted strains, or it was not possible to identify the protein by MALDI-TOF-MS. Only identified proteins showing similar change of expression for all adapted strains and biological repeats analysed are discussed in the following.
Based on these criteria, the analyses of the protein data revealed 13 identified proteins (Table 2). For some proteins, such as RpsF and NfnB (Fig. 1; Table 2), several spots were identified and analysed.
|
|
). The PCR confirmed the array data and confirmed that marA was upregulated in all adapted strains, although at only a small level in one of them. The effect of BC in the medium (BC conc.) varied between adapted strains and no genes showed significance (P<0.05) for this. The statistical model term StrainxBC conc. showed significance (P<0.05) for 105 genes (results not shown). These 105 genes are listed in supplementary material available with the online version of this paper (and also at http://www.matforsk.no/lab/ebo/1). Significance for this interaction indicates that there is some difference in genetic expression between 50 µg ml–1 and no BC in the medium, but this effect varied between adapted strains (A1, A2, A3) and no significant main effect for BC conc. was shown. These genes are therefore not further discussed in the following.
|
). A large upregulation of the global regulator marA was also seen in two of the three adapted lines (A1 and A2) but only a small increase in the third adapted line (A3) placed marA outside the 5 % significance level (P=0.118). A significant effect of adapted line variation was also calculated for marA (P=0.018). RTQ-PCR confirmed that marA was upregulated in all adapted lines, but that the upregulation in one adapted line was small. In this strain (A3) the expression level of marA increased early in the adaptation process, but then came close to the expression level of the control strain towards the end of the adaptation process (Fig. 2). Our results showed a significant increase in expression of acrAB and tolC. Testing for adapted line variation showed significance for acrB (P=0.051), but RTQ-PCR confirmed the upregulation of acrB in all adapted lines (Table 3). RTQ-PCR also showed (Fig. 2) that the upregulation of acrB took place in the early part of the adaptation, without any further increase in the level of expression throughout the adaptation process.
|
, Fig. 2), through the gradual increase in BC tolerance level during the adaptation process. Both macroarrays and RTQ-PCR showed a striking upregulation for soxS. The curve throughout the adaptation process (Fig. 2) clearly shows high upregulated levels of soxS already in the early adaptation steps, but also that the level continued to increase throughout the adaptation process.
Although at a much lower level of overexpression than soxS, the sox regulon genes zwf and nfnB were both significantly upregulated in the adapted strains. The gene zwf encoding glucose-6-phosphate dehydrogenase, which is the first enzyme leading into the pentose phosphate pathway (PPP), is known to be trancriptionally activated by SoxS when exposed to oxidative stress (Greenberg et al., 1990) and by antibiotics through marA (Cohen et al., 1993). Another glucose dehydrogenase gene, gcd, was also shown to be upregulated in adapted strains at a similar ratio to zwf (Table 2), but the relation of this gene to the soxRS regulon has never been described.
NfnB is a minor E. coli oxygen-insensitive nitroreductase that uses both NADPH and NADH as cofactors. Transcription of nfnB is known to be highly increased by induced or constitutive chromosomal expression of marA, but is unaltered by constitutive expression of chromosomal soxS (Barbosa & Levy, 2002). In the protein gels, NfnB was identified in three spots with the same molecular mass, but different pI values (Fig. 1).
Other transporter genes
Besides the sox- and mar-regulated acrAB multidrug exporter, the gene yhiV also shows a significant increase in its level of expression. This gene is a component of the YhiUV multidrug transporter tolC, which is also upregulated in the macroarrays, and is required for proper YhiUV transporter function (Zgurskaya & Nikaido, 2000). The gene yeeF encoding an amino acid APC transporter showed a reduced level of expression (Table 2).
Outer-membrane proteins (Omps)
The 2-DE shows a change in expression for several Omps, including TolC already mentioned (Table 2). OmpA, OmpF and OmpT were found to be downregulated while OmpC was upregulated in adapted strains. OmpF, OmpC and OmpT are general bacterial porins and it is well known that when osmotic pressure increases, OmpF decreases and OmpC increases and vice versa. Ishikawa et al. (2002) also found a higher relative content of OmpC than OmpF in accordance with higher BC tolerance in E. coli. Our Taqman analysis for ompF showed an initial response opposed to that of soxS, with a steep decrease in expression for initial adaptation steps. Then, the level of expression increased again and normalized half way through the adaptation process (Fig. 2) and stayed close to normal level for the last adaptation steps (50–100 µg ml–1). The Taqman results for the end-point in the adaptation of the different adapted lines (Table 3) indicate that this normalization of ompF expression level occurred in only one of the three adapted lines. The two other adapted strains remained downregulated to a significant level at the end of the adaptation process (Table 3).
Modifications of ribosomal subunits
A cluster of spots (Fig. 1) in the lower range of both pI and molecular mass was identified by MALDI-TOF as RpsF protein (30S ribosomal subunit S6). Two of these spots disappeared in patterns of control cells while RpsF spots with slightly higher molecular mass and lower pI appeared in patterns of adapted cells. These modifications indicate that RpsF seems to be post-translationally modified in the adapted strain. RpsF has two glutamic acid residues (Glu) in its C-terminus and it is known that additional Glu residues can be added post-translationally to the C-terminus by the enzyme RimK (Kang et al., 1989; Wilcox et al., 2001). Mutations in ribosomal subunits have been observed to confirm resistance to antibiotics (Allen & Noller, 1989; Bilgin et al., 1990; Chittum & Champney, 1994; Funatsu et al., 1972), and thus it can be assumed that the new RpsF spots observed in adapted cells correspond to post-translational modifications of the protein involved in cell tolerance to BC.
Other stress-related genes
The MdaB protein (NADPH quinone reductase) was significantly upregulated in the 2-DE gels (Table 2). Little has been published on this protein, but Wang & Maier (2004) concluded that it plays an important role in oxidative stress resistance in Helicobacter pylori. Moreover, Barbosa & Levy (2000) showed that the gene was upregulated along with constitutive expression of MarA. The mdaB gene seems to increase less during the first adaptation step than the sox regulon genes (Fig. 2), but it still showed a significant increase in the level of expression when compared to the control and it continued to increase throughout the adaptation process.
Ssb, the single-stranded DNA-binding protein, binds to single-stranded DNA at the lagging strand of the replication fork in E. coli cells (Eisenman & Craig, 2004; Witte et al., 2003). The 2-DE patterns show Ssb to be underexpressed in adapted strains. Others have recently shown Ssb to be upregulated in pH-tolerant oral streptococci (Len et al., 2004) and Ssb has also been shown to be induced upon UV treatment, but downregulated by mitomycin C-induced damage (Courcelle et al., 2001; Khil & Camerini-Otero, 2002).
The major global regulator gene dps showed a significant change in expression in the protein gels (Fig. 1). Dps (DNA-binding protein) is primarily produced in stationary phase and has been shown to be regulated by OxyR, 
38 and 70 (Altuvia et al., 1994; Azam et al., 1999; Lomovskaya et al., 1994). Nishino & Yamaguchi (2004) showed that another global regulator, H-NS, contributes to multidrug resistance, including to BC, by regulating the expression of multidrug exporter genes. They showed that a double deletion mutant hns acrAB had increased resistance to BC while no change was observed for an hns mutant and concluded that H-NS represses the expression of acrEF and yhiUV (mdtEF). However, our protein data for H-NS varied significantly between adapted lines, making it difficult to draw any final conclusions on the importance of this regulator.
Antibiotic tolerance of adapted strains
Table 4 shows antibiotic tolerance concentrations for the control and adapted strains. All adapted strains showed an increase in antibiotic tolerance for chloramphenicol, florfenicol, ciprofloxacin, nalidixic acid, ampicillin and cefotaxime. Increased susceptibility was shown for gentamicin, streptomycin and kanamycin.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Sakagami et al., 1989). We have earlier demonstrated that adaptation of E. coli to BC is followed by enhanced efflux activity and may result in cross-resistance to several antibiotics (Langsrud et al., 2004). The results of the current work showed that the BC-adapted E. coli strains have acquired defence mechanisms normally related to oxidative protection, and increased tolerance to antibiotics was confirmed.
soxS was by far the most upregulated gene in our arrays. It is known that soxS and marA contribute to intrinsic resistance to several antibiotics and chemicals in E. coli (for a review, see Randall & Woodward, 2002), and that overproduction of SoxS leads to a higher superoxide resistance than overproduction of MarA (Martin et al., 2000). Our results showed a striking upregulation of several genes known to belong to the soxRS regulon, as would be expected for cells exposed to an oxidative stress component. Basically, E. coli cells possess two genetic defence systems for response to oxidative stress: one specific defence against peroxides, mediated by the transcriptional activator OxyR (Zheng et al., 2001), and one against superoxide (Pomposiello et al., 2001) that is controlled by the two-stage soxRS system (for a review, see Cabiscol et al., 2000). Induction of defence mechanisms against oxidative stress has been linked to other stress responses such as acid stress responses. Maurer et al. (2005) showed that low pH accelerates acid consumption and proton export in E. coli, while co-inducing oxidative stress and heat shock regulons. Others have shown that overexpression of soxS in E. coli produces resistance to the antimicrobial agent triclosan (McMurry et al., 1998). When it comes to QACs such as BC, a recent study on BC effects on Chang conjunctival cells showed that QACs induced superoxide synthesis at low concentrations (Debbasch et al., 2001). To our knowledge there has not been published any clear link between acquired resistance to QACs, such as BC, and oxidative stress response in bacteria.
Cationic surfactants, including QACs, are known to disrupt cell membranes and interrupt protein functions, releasing K+ and other cell constituents and inducing cell autolysis. Majtan et al. (1995) reported that the QAC (1-methyldodecyl)trimethylammonium bromide inhibits respiratory activity in Pseudomonas aeruginosa. Under aerobic growth, disruption of the cell membrane by BC activity could disrupt the electron-transport chain and thereby increase the level of highly reactive oxygen radicals in the cell. This indirect oxidative stress would cause a depletion of NADPH followed by activation of the soxRS regulon as observed in our results. 2-DE showed a significant increase in the expression level of manganese superoxide dismutase (MnSOD) (Table 2, Fig. 1). MnSOD is the primary antioxidant enzyme that protects cells from oxidative stress by catalysing dismutation of superoxide (
) to hydrogen peroxide and oxygen. It is uniformly distributed throughout the cytoplasm in prokaryotic cells, binds to DNA and protects it from oxidative damage (Steinman & Weinstein, 1993; Steinman et al., 1994). The MnSOD gene sodA is part of the soxRS regulon and is transcriptionally regulated by at least six control systems (Compan & Touati, 1993; Nunoshiba et al., 1992; Wu & Weiss, 1992). The increase in MnSOD levels in the adapted cells helps to minimize depletion of NADPH, as does also induction of the PPP through Zwf, by increasing production of NADPH. Besides zwf, we also observed an increase in the expression of gcd, encoding another dehydrogenase. Gcd is known to be involved in the formation of membrane potentials, feeding electrons obtained from the oxidation of glucose to ubiquinone and ubiquinol oxidase (van Schie et al., 1985; Yamada et al., 1993b). Gcd is negatively regulated by cAMP and positively by oxygen (Yamada et al., 1993a).
FldA (flavodoxin I), one of the two known flavodoxins in E. coli and also known to belong to the soxRS oxidative stress regulon, was significantly upregulated in the adapted strains (Table 2). The same was the case for fumC (fumarase C) (Paterson et al., 2002). Upregulation of fumC, as observed in our data (Table 2), is also a part of the cellular defence against oxidative stress through the replacement of redox-sensitive isozymes by redox-resistant isozymes (fumarase C). FumC activity has also been shown to be upregulated as a response to high oxygen levels (Liochev & Fridovich, 1992; Tseng et al., 2001).
Our results showed significant upregulation of the nfnB gene (synonym nfsB), encoding an oxygen-insensitive nitroreductase, while we saw no evidence for increased expression of nfsA. On the other hand, we found the ybjC gene (conserved hypothetical protein) to be upregulated in adapted strains. This gene is located directly upstream of nfsA in the sox regulon, cotranscribed with nfsA, which suggests that these two genes form an operon (Paterson et al., 2002). NfsA is the major nitroreductase and uses NADPH as the electron donor. In a macroarray study, Barbosa & Levy (2000) showed nfnB, as well as nfsA, to be MarA responsive and they found a significant change for nfnB and ybjC, but not for nfsA under constitutive expression of MarA. Their data suggested that MarA is preferred over SoxS in upregulating nfnB and further work by the same authors supported this conclusion (Barbosa & Levy, 2002). Our results corroborate those of Barbosa & Levy (2000) for many of the observed changes in our adapted strains. The expression level of nfnB is lower for the adapted line (Table 3, A3) that showed the lowest level of marA expression. Due to the increase in expression we observed for marA and several genes known to be affected by constitutive expression of MarA, it is likely that this protein plays a regulatory role in the increased resistance to BC in the adapted cells. However, the variation between adapted lines for this gene, as opposed to the highly statistically significant upregulation of soxS, points to soxS as the most important regulator in the general adapted resistance to BC. Both marA and soxS encode positive regulators of acrAB and upregulate acrAB, which encodes a stress-induced multidrug efflux pump formed by the AcrAB–TolC complex across the membranes (Ma et al., 1995). This AcrAB–TolC efflux pump has been linked to intrinsic resistance to BC (Moken et al., 1997). Our results showed a significant increase in expression of acrAB and tolC.
There was no support in our results for any variation of gene or protein expression in the OxyR system, except a reduced expression of Dps in adapted cells as observed in 2-DE gels. dps is also part of the RpoS growth phase regulon and most work published on this protein refers to upregulation of its expression as a defence mechanism against oxidative damage by peroxides during the exponential growth phase (Almiron et al., 1992; Choi et al., 2000; Martinez & Kolter, 1997; Martinez-Martinez et al., 2000). Our data also showed a downregulated level of another member of the RpoS regulon, the WrbA protein. It therefore stands to reason that the RpoS regulon could be repressed in the adapted strains, although further work would have to be done to confirm this. Results of bactericidal tests (not shown) showed a much higher survival ratio for E. coli control cells in the exponential phase than in the stationary phase. In the stationary phase, Dps is known to be active in DNA protection by presence in high molecule numbers and against multiple stress factors such as UV, copper, thermal, acid and base shock (Almiron et al., 1992; Nair & Finkel, 2004; Wolf et al., 1999). Kirkpatrick et al. (2001) showed repressed levels of Dps and other RpoS regulon proteins for E. coli exposed to formate stress, as opposed to acetate stress, which led to increased expression of the same proteins. They hypothesized that the level of acetyl-CoA could be involved in regulation of this response. The activation of the soxRS regulon leads to induction of the PPP through glucose-6-phosphate dehydrogenase (Zwf) with corresponding reversion of the glycolysis (glyconeogenesis) to meet the urgent demand for reducing power (NADPH). The need to metabolize pyruvate through other pathways then arises, resulting in a reduction of actyl-CoA levels, which could explain the reduced level of the Dps protein in our adapted strains.
Changes in the cell membrane with alteration of the LPS composition have been described for QACs and this might lead to a tighter membrane structure, making it harder to extract proteins. On the other hand, recent work by Braoudaki & Hilton (2005) concluded that a sublethal concentration of BC did not lead to any significant changes in outer membrane or LPS, but that preadapted strains proved less hydrophobic than resistant strains. We observed that the pellets of cells grown in the presence of BC were dark brownish in colour and stuck strongly to the inside of the test tubes. The lower OD600 we observed for adapted cells at mid-exponential phase (108 c.f.u. ml–1) could also indicate a change in cell size.
Observed changes in Omps and in particular the upregulation of the small porin OmpC and the downregulation of the larger porin OmpF is well described in the literature as a mechanism to reduce access through the cell membrane (Ishikawa et al., 2002). The Taqman results for ompF, which showed a normalization of the ompF level after an initial decrease in expression level for one of the adapted lines, could indicate that the ompF response is not critical for the BC-resistant strain. One explanation could be that BC cannot pass through the OmpF porin, but that the switch from the large porin OmpF to the smaller porin OmpC is a general stress response, not actually protecting the cell from BC. Another explanation could be that as other adapted resistance mechanisms become stronger, the OmpF downregulation in early adapted cells is no longer needed to cope with the BC threat. This presupposes that BC can enter the cell through the OmpF porin, which is yet to be verified. Loss of OmpF has been shown to be one of the resistance mechanisms elaborated by E. coli against 
-lactam antibiotics (Martinez-Martinez et al., 2000) and fluoroquinolones (Tavio et al., 1999). It is known that -lactams such as ampicillin, which are molecules close to BC in size, can bind to and pass through the OmpF porin (Nestorovich et al., 2002) and interestingly the antibiotic panel testing (Table 2) showed an increase in tolerance to ampicillin.
The protein data indicate post-translational modifications of the S6 ribosomal subunit encoded by the rpsF gene. Several of the antibiotics showing increased tolerance for adapted strains in our screening tests (gentamicin, tetracycline, chloramphenicol, streptomycin) have killing mechanisms linked to ribosomal activity. One might speculate whether ribosomal modification is a likely cross-resistance mechanism between BC and certain antibiotics. Further work is needed to clarify this issue.
The variation in curve profiles for the adaptation process indicates that different genetic approaches could lead to the same resistance level and/or that some adaptation steps could be more difficult to acquire. This is further supported by the significant differences between adapted lines in the expression of genes such as marA and ompF.
The gapA gene is a common reference for Taqman results, but our results showed a variation in the level of expression of the GapA protein. Therefore, we decided to use another commonly used reference and housekeeping gene, accD, as a second reference for these experiments. The results showed that one of the adapted lines differed from the others when gapA was used as a reference. This confirmed that the expression of gapA was more strongly upregulated in some adapted lines than in others. For two out of three adapted lines, however, the gapA reference gave the same pattern of change in level of expression as accD. The change in gapA level in adapted strains could be explained by increased flux through the PPP, producing NADPH and returning glyceraldehyde 3-phosphate to glycolytic metabolism and then running either glyconeogenesis, or the second part of glycolysis to pyruvate.
The average signal in the DNA arrays was relatively low. Therefore there are probably smaller (<2-fold) but significant changes in the expression of several genes that are not seen in these results. The data presented represent the largest changes in expression level between adapted strains and control strain and are clearly significant for repeated experiments. There is overlap between protein data and DNA array data, and the PCR data confirm the other data.
In summary, E. coli K-12 adapted to higher tolerance to BC acquired several general resistance mechanisms including responses normally related to the multiple antibiotic resistance (Mar) regulon and protection against oxidative stress. Adapted strains showed increased tolerance to several antibiotics and the results revealed that BC treatment might result in superoxide stress in E. coli.
|
ACKNOWLEDGEMENTS |
|---|
The authors acknowledge the assistance of Signe M. Drømtorp in performing RTQ-PCR, Dr Kristine Naterstad for sharing her experience in PFGE, Dr Lars Axelsson and Dr Akild Holck for fruitful discussions on the subject of stress responses and Dr Jean Labadie for initiating the co-operation on this particular project between Matforsk-Norway and INRA-France.
Edited by: I. R. Henderson
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Allen, P. N. & Noller, H. F. (1989). Mutations in ribosomal protein-S4 and protein-S12 influence the higher-order structure of 16-S ribosomal-RNA. J Mol Biol 208, 457–468.[CrossRef][Medline]
Almiron, M., Link, A. J., Furlong, D. & Kolter, R. (1992). A novel DNA-binding protein with regulatory and protective roles in starved Escherichia coli. Genes Dev 6, 2646–2654.
Altuvia, S., Almiron, M., Huisman, G., Kolter, R. & Storz, G. (1994). The Dps promoter is activated by Oxyr during growth and by Ihf and a Sigma(S) in stationary-phase. Mol Microbiol 13, 265–272.[Medline]
Azam, T. A., Iwata, A., Nishimura, A., Ueda, S. & Ishihama, A. (1999). Growth phase-dependent variation in protein composition of the Escherichia coli nucleoid. J Bacteriol 181, 6361–6370.
Barbosa, T. M. & Levy, S. B. (2000). Differential expression of over 60 chromosomal genes in Escherichia coli by constitutive expression of MarA. J Bacteriol 182, 3467–3474.
Barbosa, T. M. & Levy, S. B. (2002). Activation of the Escherichia coli nfnB gene by MarA through a highly divergent marbox in a class II promoter. Mol Microbiol 45, 191–202.[CrossRef][Medline]
Bilgin, N., Richter, A. A., Ehrenberg, M., Dahlberg, A. E. & Kurland, C. G. (1990). Ribosomal-RNA and protein mutants resistant to spectinomycin. EMBO J 9, 735–739.[Medline]
Blattner, F. R., Plunkett, G., Bloch, C. A., Perna, N. T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J. D., Rode, C. K. & other authors (1997). The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1474.
Bohm, H. & Karch, H. (1992). DNA fingerprinting of Escherichia coli O157-H7 strains by pulsed-field gel-electrophoresis. J Clin Microbiol 30, 2169–2172.
Bradford, M. M. (1976). Rapid and sensitive method for quantitation of microgram quantities of protein utilizing principle of protein-dye binding. Anal Biochem 72, 248–254.[CrossRef][Medline]
Braoudaki, M. & Hilton, A. C. (2005). Mechanisms of resistance in Salmonella enterica adapted to erythromycin, benzalkonium chloride and triclosan. Int J Antimicrob Agents 25, 31–37.[CrossRef][Medline]
Cabiscol, E., Tamarit, J. & Ros, J. (2000). Oxidative stress in bacteria and protein damage by reactive oxygen species. Int Microbiol 3, 3–8.[Medline]
Chittum, H. S. & Champney, W. S. (1994). Ribosomal-protein gene sequence changes in erythromycin-resistant mutants of Escherichia coli. J Bacteriol 176, 6192–6198.
Choi, S. H., Baumler, D. J. & Kaspar, C. W. (2000). Contribution of dps to acid stress tolerance and oxidative stress tolerance in Escherichia coli O157 : H7. Appl Environ Microbiol 66, 3911–3916.
Cohen, S. P., Hachler, H. & Levy, S. B. (1993). Genetic and functional analysis of the multiple antibiotic-resistance (Mar) locus in Escherichia coli. J Bacteriol 175, 1484–1492.
Compan, I. & Touati, D. (1993). Interaction of six global transcription regulators in expression of manganese superoxide dismutase in Escherichia coli K-12. J Bacteriol 175, 1687–1696.
Courcelle, J., Khodursky, A., Peter, B., Brown, P. O. & Hanawalt, P. C. (2001). Comparative gene expression profiles following UV exposure in wild-type and SOS-deficient Escherichia coli. Genetics 158, 41–64.
Debbasch, C., Brignole, F., Pisella, P. J., Warnet, J. M., Rat, P. & Baudouin, C. (2001). Quaternary ammoniums and other preservatives' contribution in oxidative stress and apoptosis on Chang conjunctival cells. Invest Ophthalmol Vis Sci 42, 642–652.
Denyer, S. P. & Maillard, J. Y. (2002). Cellular impermeability and uptake of biocides and antibiotics in Gram-negative bacteria. J Appl Microbiol 92 (Suppl.), 35S–45S.
Deshusses, J. M. P., Burgess, J. A., Scherl, A., Wenger, Y., Walter, N., Converset, V., Paesano, S., Corthals, G. L., Hochstrasser, D. F. & Sanchez, J. C. (2003). Exploitation of specific properties of trifluoroethanol for extraction and separation of membrane proteins. Proteomics 3, 1418–1424.[CrossRef][Medline]
Eisenman, H. C. & Craig, E. A. (2004). Activation of pleiotropic drug resistance by the J-protein and Hsp70-related proteins, Zuo1 and Ssz1. Mol Microbiol 53, 335–344.[CrossRef][Medline]
Folio, P., Chavant, P., Chafsey, I., Belkorchia, A., Chambon, C. & Hébraud, M. (2004). Two-dimensional electrophoresis database of Listeria monocytogenes EGDe proteome and proteomic analysis of mid-log and stationary growth phase cells. Proteomics 4, 3187–3201.[CrossRef][Medline]
Funatsu, G., Wittmann, H. G. & Schiltz, E. (1972). Ribosomal-proteins. 27. Localization of amino-acid exchanges in protein S5 from two Escherichia coli mutants resistant to spectinomycin. Mol Gen Genet 114, 106–111.[CrossRef][Medline]
Gilbert, P. & Moore, L. E. (2005). Cationic antiseptics: diversity of action under a common epithet. J Appl Microbiol 99, 703–715.[CrossRef][Medline]
Greenberg, J. T., Monach, P., Chou, J. H., Josephy, P. D. & Demple, B. (1990). Positive control of a global antioxidant defense regulon activated by superoxide-generating agents in Escherichia coli. Proc Natl Acad Sci U S A 87, 6181–6185.
Heir, E., Sundheim, G. & Holck, A. L. (1995). Resistance to quaternary ammonium compounds in Staphylococcus spp. isolated from the food industry and nucleotide sequence of the resistance plasmid pST827. J Appl Microbiol 79, 149–156.[Medline]
Heir, E., Sundheim, G. & Holck, A. L. (1999). The qacG gene on plasmid pST94 confers resistance to quaternary ammonium compounds in staphylococci isolated from the food industry. J Appl Microbiol 86, 378–388.[CrossRef][Medline]
Ishikawa, S., Matsumura, Y., Yoshizako, F. & Tsuchido, T. (2002). Characterization of a cationic surfactant-resistant mutant isolated spontaneously from Escherichia coli. J Appl Microbiol 92, 261–268.[CrossRef][Medline]
Kang, W. K., Icho, T., Isono, S., Kitakawa, M. & Isono, K. (1989). Characterization of the gene rimk responsible for the addition of glutamic-acid residues to the C-terminus of ribosomal protein-S6 in Escherichia coli K12. Mol Gen Genet 217, 281–288.[CrossRef][Medline]
Khil, P. P. & Camerini-Otero, R. D. (2002). Over 1000 genes are involved in the DNA damage response of Escherichia coli. Mol Microbiol 44, 89–105.[CrossRef][Medline]
Kirkpatrick, C., Maurer, L. M., Oyelakin, N. E., Yoncheva, Y. N., Maurer, R. & Slonczewski, J. L. (2001). Acetate and formate stress: opposite responses in the proteome of Escherichia coli. J Bacteriol 183, 6466–6477.
Langsrud, Ø. (2002). 50-50 multivariate analysis of variance for collinear responses. J R Stat Soc Ser D 51, 305–317.
Langsrud, Ø. (2005). Rotation tests. Stat Comput 15, 53–60.
Langsrud, S., Sundheim, G. & Borgmann-Strahsen, R. (2003). Intrinsic and acquired resistance to quaternary ammonium compounds in food-related Pseudomonas spp. J Appl Microbiol 95, 874–882.[CrossRef][Medline]
Langsrud, S., Sundheim, G. & Holck, A. L. (2004). Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed to stress-inducers. J Appl Microbiol 96, 201–208.[CrossRef][Medline]
Len, A. C. L., Harty, D. W. S. & Jacques, N. A. (2004). Stress-responsive proteins are upregulated in Streptococcus mutans during acid tolerance. Microbiology 150, 1339–1351.
Li, X. Z. & Nikaido, H. (2004). Efflux-mediated drug resistance in bacteria. Drugs 64, 159–204.[CrossRef][Medline]
Liochev, S. I. & Fridovich, I. (1992). Fumarase-C, the stable fumarase of Escherichia coli, is controlled by the SoxRS regulon. Proc Natl Acad Sci U S A 89, 5892–5896.
Lomovskaya, O. L., Kidwell, J. P. & Matin, A. (1994). Characterization of the sigma(38)-dependent expression of a core Escherichia coli starvation gene, pexb. J Bacteriol 176, 3928–3935.
Ma, D., Cook, D. N., Alberti, M., Pon, N. G., Nikaido, H. & Hearst, J. E. (1995). Genes acrA and acrB encode a stress-induced efflux system of Escherichia coli. Mol Microbiol 16, 45–55.[CrossRef][Medline]
Majtan, V., Majtanova, L., Hostacka, A., Hybenova, D. & Mlynarcik, D. (1995). Effect of quaternary ammonium salts and amine oxides on Pseudomonas aeruginosa. Microbios 84, 41–51.[Medline]
Martin, R. G., Gillette, W. K. & Rosner, J. L. (2000). Promoter discrimination by the related transcriptional activators MarA and SoxS: differential regulation by differential binding. Mol Microbiol 35, 623–634.[CrossRef][Medline]
Martinez, A. & Kolter, R. (1997). Protection of DNA during oxidative stress by the nonspecific DNA-binding protein Dps. J Bacteriol 179, 5188–5194.
Martinez-Martinez, L., Conejo, M. C., Pascual, A., Hernandez-Alles, S., Ballesta, S., De Arellano-Ramos, E. R., Benedi, V. J. & Perea, E. J. (2000). Activities of imipenem and cephalosporins against clonally related strains of Escherichia coli hyperproducing chromosomal beta-lactamase and showing altered porin profiles. Antimicrob Agents Chemother 44, 2534–2536.
Maurer, L. M., Yohannes, E., Bondurant, S. S., Radmacher, M. & Slonczewski, J. L. (2005). pH regulates genes for flagellar motility, catabolism, and oxidative stress in Escherichia coli K-12. J Bacteriol 187, 304–319.
McDonnell, G. & Russell, A. D. (1999). Antiseptics and disinfectants: activity, action, and resistance. Clin Microbiol Rev 12, 147–179.
McMurry, L. M., Oethinger, M. & Levy, S. B. (1998). Overexpression of marA, soxS, or acrAB produces resistance to triclosan in laboratory and clinical strains of Escherichia coli. FEMS Microbiol Lett 166, 305–309.[CrossRef][Medline]
Moen, B., Oust, A., Langsrud, O., Dorrell, N., Marsden, G. L., Hinds, J., Kohler, A., Wren, B. W. & Rudi, K. (2005). Explorative multifactor approach for investigating global survival mechanisms of Campylobacter jejuni under environmental conditions. Appl Environ Microbiol 71, 2086–2094.
Moken, M. C., McMurry, L. M. & Levy, S. B. (1997). Selection of multiple-antibiotic-resistant (Mar) mutants of Escherichia coli by using the disinfectant pine oil: roles of the mar and acrAB loci. Antimicrob Agents Chemother 41, 2770–2772.[Abstract]
Molloy, M. P., Herbert, B. R., Williams, K. L. & Gooley, A. A. (1999). Extraction of Escherichia coli proteins with organic solvents prior to two-dimensional electrophoresis. Electrophoresis 20, 701–704.[CrossRef][Medline]
Nair, S. & Finkel, S. E. (2004). Dps protects cells against multiple stresses during stationary phase. J Bacteriol 186, 4192–4198.
NCCLS (2002). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically. NCCLS approved standard M7-A5. Wayne, PA: National Committee for Clinical Laboratory Standards.
), marA (
), nfnB (
), acrB (
), zwf (
), mdaB (
) and ompF (
) through the stepwise adaptation of E. coli ATCC 47076 to BC from an initial MIC of 13 µg ml–1 to a final MIC of 92 µg ml–1 in the adapted strain A3. An expression level of 1 in the graph indicates an equal level of expression in control and adapted strain. An expression level >1 indicates increased expression while an expression level <1.00 indicates decreased expression level of a gene through adaptation. The data in the figure are normalized to the expression of the housekeeping gene accD.