OUP user menu

Development of antibiotic resistance and up-regulation of the antimutator gene pfpI in mutator Pseudomonas aeruginosa due to inactivation of two DNA oxidative repair genes (mutY, mutM)

Lotte Frigaard Mandsberg, Maria D. Maciá, Kirsten R. Bergmann, Lasse E. Christiansen, Morten Alhede, Nikolai Kirkby, Niels Høiby, Antonio Oliver, Oana Ciofu
DOI: http://dx.doi.org/10.1111/j.1574-6968.2011.02383.x 28-37 First published online: 1 November 2011

Abstract

Prevention and correction of oxidative DNA lesions in Pseudomonas aeruginosa is ensured by the DNA oxidative repair system (GO). Single inactivation of mutT, mutY and mutM involved in GO led to elevated mutation rates (MRs) that correlated to increased development of resistance to antibiotics. In this study, we constructed a double mutant in mutY and mutM (PAOMY-Mgm) and characterized the phenotype and the gene expression profile using microarray and RT-PCR. PAOMY-Mgm presented 28-fold increases in MR compared with wild-type reference strain PAO1. In comparison, the PAOMYgm (mutY) single mutant showed only a fivefold increase, whereas the single mutant PAOMMgm (mutM) showed a nonsignificant increase in MR compared with PAO1 and the single mutants. Mutations in the regulator nfxB leading to hyperexpression of MexCD-OprJ efflux pump were found as the mechanism of resistance to ciprofloxacin in the double mutant. A better fitness of the mutator compared with PAO1 was found in growth competition experiments in the presence of ciprofloxacin at concentrations just below minimal inhibitory concentration. Up-regulation of the antimutator gene pfpI, that has been shown to provide protection to oxidative stress, was found in PAOMY-Mgm compared with PAO1. In conclusion, we showed that MutY and MutM are cooperating in the GO of P. aeruginosa, and that oxidative DNA lesions might represent an oxidative stress for the bacteria.

Keywords
  • Pseudomonas aeruginosa
  • mutator
  • oxidative repair

Introduction

The chronic lung infection with Pseudomonas aeruginosa in the lungs of patients with cystic fibrosis (CF) is characterized by the biofilm mode of growth and a chronic inflammation dominated by polymorphonuclear leucocytes (PMNs) (Koch & Hoiby, 1993; Bjarnsholt et al., 2009). Pseudomonas aeruginosa are exposed to many reactive oxygen species (ROS), both generated by its own metabolism and especially from a large number of PMNs which release ROS in response to the chronic CF-lung infection (Brown & Kelly, 1994; Suntres et al., 2002; Kolpen et al., 2010). In addition, exposure of the microorganisms to antipseudomonal antibiotics such as ciprofloxacin, which is an inhibitor of DNA-gyrase, can stimulate the bacterial production of ROS (Dwyer et al., 2007; Kohanski et al., 2007). To combat the mutagenic consequences of the ROS, the MutT, MutY and MutM of the DNA oxidative repair system (GO) play an important role (Mandsberg et al., 2009). The enzymes of the GO system repair and prevent the incorporation of an oxidative damaged form of guanosine, 7,8-dihydro-8-oxo-dGuanine (8-oxodG), in the DNA. The MutT is a nucleoside triphosphate pyrophosphohydrolase catalysing the conversion of 8-oxodGTP to 8-oxodGMP + PPi, preventing oxidized guanine from being incorporated under replication; MutM is a formamidopyrimidine DNA-glycosylase that removes 8-oxodG when base paired with cytosine, and MutY is an adenine glycosylase capable of removing adenine when paired with 8-oxodG minimizing GC : TA transversions(Livingston et al., 2008). Inactivation of the GO genes leads to increased spontaneous mutations characterized by GC : TA (mutM and mutY) or AT : CG (mutT) transversions (Michaels & Miller, 1992; Fowler et al., 2003; Eutsey et al., 2007). An additional gene, pfpI, has been shown to play an antimutator role due to the protective role of its product against the DNA damage caused by oxidative stress (Rodriguez-Rojas & Blazquez, 2009).

In the recent years, much attention has been paid to the role of hypermutabillity in bacterial adaptation, and it is predicted that hypermutation is beneficial for niche specialization and survival in stressful and/or fluctuating environments such as CF-lung environment (Miller, 1996; Taddei et al., 1997; Blazquez, 2003; Woodford & Ellington, 2007). Pseudomonas aeruginosa mutators are often found in chronically infected CF patients (Oliver et al., 2000; Ciofu et al., 2005; Macia et al., 2005; Henrichfreise et al., 2007; Montanari et al., 2007), and it has also been reported that mutator strains more frequently are multidrug resistant compared with nonmutators (Miller et al., 2002; Blazquez, 2003; Ciofu et al., 2005; Macia et al., 2005).

The mechanisms involved in the occurrence of strong mutators imply defective mismatch repair systems caused by loss of function mutations in genes mutS, mutL, uvrD (Oliver et al., 2002; Hogardt et al., 2007; Montanari et al., 2007; Mena et al., 2008; Ciofu et al., 2009). Inactivation of the genes involved in the P. aeruginosa DNA oxidative repair system (GO) showed elevated mutant frequencies, which correlated to an increased development of resistance to antibiotics, indicating that oxidative stress might be involved in development of resistance to antibiotics(Morero & Argarana, 2009; Sanders et al., 2009). We have also reported the occurrence of mutations in the GO system in CF mutator P. aeruginosa isolates (Mandsberg et al., 2009).

As we found a large number of CF P. aeruginosa strains harbouring mutations in several of the DNA repair genes (Ciofu et al., 2009), and as it has been shown in Escherichia coli that mutY mutM double mutant has a 25- to 75-fold higher mutation rate (MR) than either mutator alone (Michaels et al., 1992; Tajiri et al., 1995), we were interested in studying the effect of inactivation of these two genes involved in GO repair system in P. aeruginosa. We investigated the development of antibiotic resistance and the survival of the double mutant in the presence of ciprofloxacin at concentrations just below minimal inhibitory concentration (MIC) in growth competition experiments with the wild-type strain. To get insight into the effect of a nonfunctional oxidative repair system on the global gene expression, we conducted gene expression analysis of the double mutant and of the wild-type strain.

Materials and methods

All strains and plasmids included in this study are described in Supporting Information, Table S1. As a reference strain, we used PAO1. The bacteria were grown in Luria-Bertani (LB) broth or LB agar containing the appropriate antibiotics. When grown in different media, this is mentioned.

In all experiments, the strains were cultured from stocks kept at −80 °C.

Inactivation of P. aeruginosamutM and mutY using the Lox system

Double knockout mutants in mutM and mutY were constructed using the Cre-lox system for gene deletion and antibiotic resistance marker recycling. Combined sacB-based negative selection and cre-lox antibiotic marker recycling for efficient gene deletion in P. aeruginosa were used (Quenee et al., 2005). Upstream and downstream PCR fragments (Primers listed in Table S1) of the wild-type mutM or mutY gene from P. aeruginosa strain PAO1 were digested with HindIII and either BamHI or EcoRI, and cloned by a three way ligation into pEX100Tlink deleted for the HindIII site and opened by EcoRI and BamHI. Eighty-four residues from position 268 were deleted, when the upstream and downstream mutM amplified fragments were joined in pEX100Tlink vector, and 76 residues from position 374 were deleted in mutY, respectively.

The resulting plasmids (pEXTMM and pEXTMY) were transformed into E. coli XL1Blue strain, and transformants were selected in 30 mg L−1 ampicillin LB agar plates. The lox flanked gentamicin resistance cassette (aac1) obtained by HindIII restriction of plasmid pUCGmlox was cloned into the HindIII sites in pEXTMM and pEXTMY formed by the ligation of the upstream and downstream PCR fragments. The resulting plasmids were transformed into E. coli XL1Blue strain, and transformants were selected on 30 mg L−1 ampicillin–5 mg L−1 gentamicin LB agar plates. The resulting plasmids (pEXTMMGm and pEXTMYgm) were then transformed into the E. coli S17-1 helper strain. Single knockout mutants were generated by conjugation, followed by selection of double recombinants using 5% sucrose-1 mg L−1 cefotaxime-30 mg L−1 gentamicin LB agar plates. Double recombinants were checked by screening for ticarcillin (100 mg L−1) susceptibility and afterwards by PCR amplification and sequencing. For the recycling of the gentamicin resistance cassettes, plasmid pCM157 was electroporated into different mutants. Transformants were selected in 250 mg L−1 tetracycline LB agar plates. One transformant for each mutant was grown overnight in 250 mg L−1 tetracycline LB broth to allow the expression of the cre recombinase. Plasmid pCM157 was then cured from the strains by successive passages on LB broth. Selected colonies were then screened for tetracycline (250 mg L−1) and gentamicin (30 mg L−1) susceptibility and checked by PCR amplification. The single knockout mutants obtained were named PAOMMgm and PAOMYgm. To obtain the double mutant, the conjugation experiments with pEXMMGm using PAOMY as recipients were performed as described above. MutY-mutM double mutant was named PAOMY-Mgm. The maximum growth rate was found to be the same for PAOMY-Mgm and PAO1 in LB (Philipsen et al., 2008).

Complementation assay

Shuttle vector, pUCP26, was used for the construction of recombinant plasmids containing the PA01 mutY and mutM gene, under the control of the plasmid born lac-promoter (Mandsberg et al., 2009). pLM100 with the mutY gene and pLM102 with the mutM gene were electroporated into PAOMY-Mgm separately as previously described (Mandsberg et al., 2009).

Measurement of mutant frequency

The method was modified after Oliver (Oliver et al., 2000). To determine the mutant frequency (MF) to rifampicin and streptomycin, an overnight culture of 20 mL LB media was centrifuged for 10 min at 6000 g, and resuspended in 1 mL of 0.9% NaCl.

Serial dilutions of the bacterial culture were made, and 100 μL of appropriate dilution was spread on LB, 300 mg L−1 rifampicin and 500 mg L−1 streptomycin. The plates were incubated at 37 °C for 36 h before the CFU was determined. The CFU on rifampicin and streptomycin plates were compared with the CFU from LB plates. At least five replicates were run for each experiment.

Measurement of mutation rate

The MRs are estimated using a fluctuation experiment, where a culture of each strain was diluted to 2 × 104 cells in 280 μL of LB and grown in 27 microtitre wells to stationary phase, then plated on 100 mg L−1 rifampicin LB agar plates to count the number of mutants. Three wells for each strain were used to estimate the CFU per well. The expected number of mutations per well was then estimated using the Ma—Sandri—Sarkar (MSS) maximum likelihood method described by Sarkar et al., (1992).

The MR was found by dividing the number of mutations by the final CFU per well.

Determination of antibiotic susceptibility

The MIC was determined on 105 CFU mL−1 using the E-test system (AB Biodisk, Solna, Sweden), according to instructions of the manufacturer. Experiments were run at least in triplicates.

Isolation of ciprofloxacin resistant mutants

To isolate ciprofloxacin resistant mutants, overnight cultures of PAO1 and PAOMY-Mgm were diluted and plated on 5% blood agar plates containing twofold dilutions of ciprofloxacin (Bagge et al., 2000; Mandsberg et al., 2009). Isolated resistant colonies were grown in LB without antibiotic, twice before E-test was performed with 100 μL of 10−4 dilution of an overnight culture.

Microarray data

The strains were cultured in LB to an OD600nm= 1.0. Four millilitres of each culture was harvested, and RNA isolation and purification were performed using RNA Protect Bacteria Reagent and RNeasy Mini Kit (Qiagen, Hilden, Germany). RQ1 RNAse free DNAse (Promega, Madison, WI) was added to remove contaminating DNA. The experiment was run in triplicates.

Processing of the P. aeruginosa GeneChip (Affymetrix) was performed at the Department of Clinical Biochemistry, Microarray Core Unit, Rigshospitalet, University of Copenhagen, Denmark.

The gene expression analysis was done using arraystar v.3 Software (DNASTAR), http://isim.ku.dk/units/ub/research/paoym_vs_pao1microarray.pdf/.

Real-time PCR

The level of expression of mexB, mexD, mexF and mexX was determined using real-time PCR in adapted isolates from the growth competition assays, and the level of expression of pfpI and PA5148 was determined to verify the microarray data. RNA from the logarithmic phase growth OD600 nm = 0.9 was purified with the RNeasy mini kit (Qiagen) followed by a DNase treatment with RQ1 RNase free DNase (Promega). The RNA was adjusted to a concentration of 140 ng μL−1. A total quantity of 280 ng RNA was then used for one-step reverse transcription using High Capacity RNA-to-cDNA Master Mix (Applied Biosystems). For quantitative real-time PCR, amplification was performed with Power SYBR Green Master Mix in a Step One Plus Thermal Cycler (Applied Biosystems). Forty cycles were run with denaturation at 95 °C for 15 s, annealing at 55 °C for 30 s and extension at 60 °C for 45 s. rpsL was used as reference gene to normalize the relative amount of mRNA. The mRNA levels of a specific gene were expressed by comparing with the expression of the reference gene on that strain and also in PAO1, and the expression levels were calculated on a standard curve (Oh et al., 2003). RT-PCR was carried out in triplicates. Primers used for RT-PCR investigations are described in Table S1.

Growth competition assays

PAO1 and PAOMY-Mgm had similar growth rates in LB or in LB supplemented with 0.1 mg L−1 ciprofloxacin. Competition experiments were carried out in a Bioscreen (Labsystem C, Bie og Berntsen) with and without antibiotic. We attempted to start with a ratio 1 : 1 of PAO1 and PAOMY-Mgm in each well. The inoculums in the start of the experiment were 3.5 × 108 CFU mL−1 for PAO1 and 2.4 × 108 CFU mL−1 for PAOMY-Mgm. A total quantity of 140 µL of each strain culture was transferred in 2 × 10 wells in microtitre plate, the growth was carried out at 37 °C, continuously shaking, and OD600 nm measurements were performed every 30 min for 24 h. The experiment was carried out for 5 days (start day 0, end day 4), and in each day 1 : 1000 dilutions of the cultures were transferred to a new microtitre plate for exponential growth throughout the experiment. Each day, the culture was serially diluted and plated on LB agar and on LB agar supplemented with 30 mg L−1 gentamicin, a concentration which is inhibitory for PAO1, but not for the PAOMY-Mgm mutant. The CFU mL−1 of PAOMY-Mgm was calculated on gentamicin plates and the PAO1 and PAOMY-Mgm mixture on LB plates. The CFU of PAO1 was calculated by subtracting the number of CFU mL−1 on gentamicin plates from the number of CFU mL−1 on LB plates. The ratio of PAO1 : PAOMY-Mgm, PAO1 : PAOMYgm and PAO1 : PAOMMgm was followed for 4 days.

Sequencing of genes involved in ciprofloxacin resistance

The efflux pumps transcriptional regulators nfxB, mexR, mexZ and mexT, and the ciprofloxacin target genes gyrA, gyrB, parC and parE, were sequenced in selected isolates from the antibiotic resistance development study and from the growth competition study. DNA was purified using Promega Wisart purification kit (Promega). Polymerase Dynazyme EXT (Finnzymes, Espoo, Finland) was used for PCR amplification. The sequencing was done on an automatic DNA sequencer ABI3700 (Macrogen Inc., Seoul, South Korea). The numbers of reads were between two and four for each gene of each strain. The sequence results were compared with the strain PAO1 sequence (www.pseudomonas.com) with dnasis® max version 2.0 (Hitachi Software Engineering) to determine the occurrence of sequence variants. Primers used for PCR amplification and sequencing are described in the Table S2.

Results

Mutant rates and frequencies of PAOMY-Mgm

The MRs on Rifampicin of the PAOMY-Mgm mutant were 28-fold higher compared with PAO1 (Table 1). As expected, due to accumulation of mutants during cell division, the MF was 1 log higher than the MR (Macia et al., 2006). Thus, the MF on rifampicin/streptomycin of the PAOMY-Mgm, double mutant was 2.76 E-6/3.08E-8 compared to 1.63E-8/1.11E-9 of PAO1, 1.36E-7/3.51E-9 of PAOMYgm (mutY) and 2.78E-8/1.69E-9 of PAOMMgm (mutM).

View this table:

Overview of the doubling times, mutation and the levels antibiotic susceptibility of PAO1, the double mutant PAOMY-Mgm, PAOMY-Mgm complemented with wild-type mutY (pLM100) and wild-type mutM (pLM102), and of the single mutants PAOMYgm (mutY) and PAOMMgm (mutM)

MIC (mg L−1) and size of resistant subpopulation
StrainDoubling time (min)Mutation rateMEPPIPCAZTOBCIPATM
PAO126.566.75E-90.12531.51.50.190.75
PAOMY-Mgm25.661.9E-70.125 (+++)4 (+++)1.5 (++)1.50.19 (+++)1 (++)
PAOMY-Mgm + pucP261.9E-70.094 (+++)3 (++)1.5 (+)1.50.19 (++)ND
PAOMY-Mgm + pLM100 (mutY)2.6E-90.125 (++)4 (+)1.51.50.190.75
PAOMY-Mgm + pLM102 (mutM)4.6E-80.125 (+++)4 (++)1.5 (+)1.50.191 (++)
PAOMYgm25.73.85E-80.094 (+++)2 (++)1.5 (++)1.5 (+)0.19 (+)1 (+)
PAOMMgm27.36.38E-90.094 (++)2 (++)1.5 (+)1.50.191.5
  • The numbers of resistant mutant colonies in the inhibition zone are as follows: +, < 10; ++, 10–100; +++, > 100 (Macia et al., 2004).

  • MEP, meropenem; PIP, piperacillin; CAZ, ceftazidime; TOB, tobramycin; CIP, ciprofloxacin; ATM, aztreonam.

Complementation of the PAOMY-Mgm double mutant with single wild-type mutY or mutM decreased the MR by 73-fold and by 4-fold (Table 1).

Increased development of resistance to antibiotics in GO double mutant

To evaluate the capacity of PAOMY-Mgm to develop resistance to antibiotics, we identified the presence of resistant mutant subpopulations within the inhibition zones of E-test strips and characterized their sizes by a ranking system described previously (Macia et al., 2004). The sizes of the resistant mutant subpopulation of PAOMY-Mgm were larger than those of the mutM single mutant (PAOMMgm) for all the tested antibiotics, and also larger than those of mutY single mutant (PAOMYgm) for ciprofloxacin, piperacillin and aztreonam (Table 1).

PAOMY-Mgm complemented with wild-type mutY showed no resistant subpopulations to ceftazidime, tobramycin, ciprofloxacin, aztreonam and showed a smaller resistant subpopulation to piperacillin and meropenem. The effect of complementation of PAOMY-Mgm with wild-type mutM was less pronounced, but eliminated the resistant subpopulation to ciprofloxacin (Table 1).

Mechanism of resistance to ciprofloxacin

To reveal the mechanism of resistance to ciprofloxacin, colonies of PAOMY-Mgm and PAO1 were collected from plates containing ciprofloxacin in concentration of fivefold MIC (1 mg L−1). The ciprofloxacin resistant colonies showed cross-resistance to several groups of antibiotics, and one of the PAOMY-Mgm colonies showed high-level resistance to ciprofloxacin (Table 2). The cross-resistance to several antibiotic groups indicated the involvement of an efflux pump as mechanism of resistance.

View this table:

Resistance phenotypes and type of mutations in nfxB in ciprofloxacin resistant isolates. Two colonies of PAO1 (mutant1 and mutant2) and two colonies of PAOMY-Mgm (mutant1 and mutant2) were picked from plates containing 1 mg L−1 ciprofloxacin

MIC (mg L−1)
Bacterial isolatesCIPTETCHLnfxB
PAO1 (control)0.19412NC
PAOMY-Mgm (control)0.19 (++)412 (+)NC
PAO1mutant10.532>256A115C
PAO1mutant20.7524>256Insertion G at position 540
PAOMY-Mgmmutant1148>256G358T: stop codon
PAOMY-Mgmmutant2>25632>256G331T: stop codon
  • +, < 10 mutants; ++, 10–100 mutant colonies in the inhibition zones.

  • This isolate presented also a C to A transversion at position 1397 in gyrB.

  • CIP, ciprofloxacin; CHL, chloramphenicol; TET, tetracycline; NC, no change in DNA sequence.

Sequencing of the transcriptional regulator nfxB allowed us to identify loss of function mutations in nfxB in four ciprofloxacin resistant isolates of PAO1 and PAOMY-Mgm, indicating that hyperexpression of MexCD-OprJ efflux pump was involved in the resistance to ciprofloxacin. However, the ciprofloxacin resistant isolates of PAOMY-Mgm showed G∙C→T∙A transversions characteristic for mutM and mutY mutants of the GO system, whereas the mutations identified in nfxB of PAO1 were base insertions and an A to C transversion (Table 2).

Interestingly, mutation G331T leading to a premature stop codon in nfxB of PAOMY-Mgm has been previously described in a ciprofloxacin resistant isolate, selected from the single mutY mutant of PAO1(Mandsberg et al., 2009).

No mutations in other transcriptional regulators of the efflux pumps (mexR, mexZ and mexT) or in the genes gyrA, parC or parE encoding the targets of ciprofloxacin were identified in these isolates, with the exception of a gyrB mutation in a PAOMY-Mgm isolate (PAOMY-Mgmmutant2) with high-level resistance to ciprofloxacin. The mutation C1397A in gyrB was a G·C→T·A transversion characteristic for mutY and mutM mutants of the GO system leading to an amino acid substitution. Alteration of gyrB at position 1397 has previously been reported in a fluoroquinolone-resistant clinical strain of P. aeruginosa (Oh et al., 2003). Mutations in both gyrB and nfxB clarify the high-level resistance to ciprofloxacin (> 256 mg L−1) in this isolate.

PAOMY-Mgm survives better than the wild-type in sub-MIC concentrations ciprofloxacin

As ciprofloxacin can stimulate the bacterial production of ROS (Morero & Argarana, 2009; Kohanski et al., 2010), and as PAOMY-Mgm mutator is defective in the repair of DNA oxidative lesions, we decided to investigate the relative fitness of the PAOMY-Mgm mutator compared with PAO1 in the presence of 0.1 mg L−1 ciprofloxacin (MIC ciprofloxacin = 0.19 mg L−1 for PAO1 and 0.19 mg L−1 and resistant subpopulation (+++) for PAOMY-Mgm). Prior to the experiment, we ensured that the PAO1 and PAOMY-Mgm mutant have statistically the same growth rate in LB (doubling time ± SD: 26.5 ± 0.6 and 25.7 ± 0.7 min, respectively) and that the concentration of 0.1 mg L−1 ciprofloxacin, which is just below the strains MIC had statistically similar inhibitory effect on the growth rates of the two strains (doubling time ± SD: 66.6 ± 3.2 and 64.3 ± 3 min, respectively) (Philipsen et al., 2008). In the absence of selection pressure in the environment, the two bacterial populations co-existed and appeared equally fitted during the 5-day period of the experiment (Fig. 1a), whereas in the presence of 0.1 mg L−1 ciprofloxacin, the PAOMY-Mgm overtook the PAO1 population at day 3 (Fig. 1b). This was not seen for the single mutants inactivated in mutY or mutM (Fig. S1 C–F). This suggests occurrence of a tolerant bacterial population more fitted to grow in the presence of ciprofloxacin in the PAOMY-Mgm population. To investigate the cause of the better fitness of the PAOMY-Mgm population compared with PAO1, we searched for ciprofloxacin resistant mutants in the mutator population. The MIC levels of ciprofloxacin were increased only by twofold and of chloramphenicol by eightfold in the adapted isolates compared with control isolates (not exposed to ciprofloxacin) (Table 3). This phenotype was associated with moderate increases in the expression levels of some of the genes encoding efflux pumps. The expression levels of mexD were increased 7- to 15-fold and of mexB twofold to fourfold compared with control, untreated isolates (Table 3). No differences in the expression levels of mexE and mexF were found (data not shown). Sequence analysis of the transcriptional regulators of the efflux pumps including the promoter regions (nfxB and mexR) and of the genes responsible for ciprofloxacin resistance (gyrA, gyrB, parC and parE) did not show any alterations leading to amino acid changes in the adapted isolates (data not shown). This finding was not surprising, as higher expression levels of mexB and mexD are to be expected in mexR or nfxB mutants (Dumas et al., 2006). The nature of the specific beneficial adaptive mechanisms in the tolerant bacterial population is under investigation in our lab.

Growth competition assay of PAO1 with PAOMY-Mgm. The relative percentages of PAO1 and PAOMY-Mgm populations during the 4 days of the experiment in LB (a) and in LB containing 0.1 mg L−1 ciprofloxacin (b).

View this table:

Resistance phenotypes and efflux pumps expression levels in five PAOMY-Mgm isolates surviving the 0.1 mg L−1 ciprofloxacin treatment in the growth competition experiment. The time-point (day 2, 3 or 4) during the competition experiment when the isolates were selected is presented

cMIC (mg L−1)Fold changes (RT-PCR)
Bacterial isolatesCIPTETCHLmexFmexDmexBmexY
PAO1 (control)0.19432
PAOMY-Mgm (control)0.19 (++)432
PAOMY-Mgm isolate day2/W74/C20.38 (+)24>2562.038.812.261.28
PAOMY-Mgm isolate day3/W74/C30.38 (+)24>2560.787.112.610.99
PAOMY-Mgm isolate day3/W74/C50.3816>2561.020.744.551.44
PAOMY-Mgm isolate day4/W74/C10.3816960.8714.934.191.1
PAOMY-Mgm isolate day4/W74/C50.381296 (+)0.727.984.190.71
  • +, < 10 mutants; ++, 10–100 mutant colonies in the inhibition zones.

  • CIP, ciprofloxacin; CHL, chloramphenicol; TET, tetracycline.

Influence of inactivation of mutY and mutM on global gene expression

To gain insight into the effect of inactivation of genes involved in the GO system on the global gene expression, we studied the transcriptional changes produced by inactivation of the two genes. The inactivation of both mutY and mutM caused significant changes (more than twofold and P-value < 0.05 compared with PAO1) in six genes (Table 4). Remarkable was the up-regulation of pfpI whose product has been shown to have a protective role against DNA damage caused by the oxidative stress (Rodriguez-Rojas & Blazquez, 2009). This can be considered a compensatory mechanism for the protection of the DNA in PAOMY-Mgm with impaired repair of the DNA oxidative damage. Interestingly, the most significantly down regulated gene was PA5148 involved in iron trafficking.

View this table:

Genes with modified transcription (equal or more than twofold) in the PAOMY-Mgm compared with PAO1 grown in LB

GeneDescription of the gene productIncrease/decrease transcription (fold)P-value
PA0355pfpI2.1200.005
PA0358Hypothetical protein2.0940.001
PA5148Hypothetical protein (Fe(II) trafficking)−4.26610−5
PA5149Hypothetical protein (oxido-reductase domain-containing protein)−2.3170.001
PA5507Hypothetical protein2.5260.005
PA5508Hypothetical protein2.0530.02
  • The P-value between the wild-type and the mutant (unpaired Student's t-test) was set to P = 0.05 for significance.

The modified expression of pfpI and PA5148 expression in PAOMY-Mgm compared with PAO1 was confirmed using RT-PCR, which showed up-regulation of pfpI (9 ± 2.3-fold) and down-regulation of PA5148 (4.04 ± 2.7-fold). Complementation of PAOMY-Mgm, which showed high expression levels of pfpI and low expression levels of PA5148 compared with PAO1 with wild-type mutM or mutY, reduced the level of pfpI up-regulation to 6 ± 2.4-fold and 4.6 ± 2.4-fold, respectively and the level of PA5148 down-regulation to 1.6 ± 0.09-fold and 2.1 ± 0.25-fold, respectively.

Discussion

Pseudomonas aeruginosa, which colonizes and persists within the highly ROS-rich CF airways has to protect itself against the mutagenic effect of ROS and it uses the GO system, consisting of MutT, MutY and MutM to prevent or eliminate the oxidized form of guanine, which is a mutagenic lesion. Homologue proteins are present in other microorganisms as well as in eukaryotic cells. Inactivation of each of the three genes encoding for the respective proteins led to various degree of increase in the spontaneous MF with mutants in mutY and mutM exhibiting a moderate and weak mutator phenotype (increase in MF < 20 times the MF of PAO1) (Mandsberg et al., 2009; Morero & Argarana, 2009; Sanders et al., 2009). In the present study, we show for the first time that the mutY and mutM double mutant (PAOMY-Mgm) showed a strong mutator phenotype providing evidence for the cooperation of MutM and MutY to prevent mutagenesis in P. aeruginosa, in a similar manner as in E. coli (Michaels et al., 1992; Tajiri et al., 1995).

It has been shown that hypermutability plays an important role in the adaptive evolution of P. aeruginosa in the CF lung (Mena et al., 2008), and it has been demonstrated that mutator populations are amplified by hitchhiking with adaptive mutations. The selective pressure exerted by antibiotics plays an important role in the adaptive process of P. aeruginosa from CF patients that are treated continuously with antipseudomonal drugs during their lifetime.

Previous studies (Oliver et al., 2000; Ciofu et al., 2005; Ferroni et al., 2009) showed that hypermutability is associated especially with multi-drug resistance development. Accordingly, we found that the increase in the frequency of mutation of PAOMY-Mgm correlated with the development of resistant subpopulations to several antipseudomonal drugs. The size of ciprofloxacin resistant subpopulation of the double GO mutant was larger compared with the single GO mutants demonstrating a faster accumulation of mutations responsible for antibiotic resistance.

As previously found in single GO mutants (Mandsberg et al., 2009; Morero & Argarana, 2009), the resistance to ciprofloxacin of the PAOMY-Mgm occurred through hyperexpression of the MexCD-OprJ due to mutation in the transcriptional regulator nfxB. The types of mutations in nfxB of PAOMY-Mgm resistant mutants were G∙C→T∙A transversions, which are specific for unrepaired oxidized guanines. High level of ciprofloxacin resistance has been linked to mutations in the DNA-gyrase and topoisomerase genes gyrA, parC, gyrB and parE (Oh et al., 2003; Lee et al., 2005). In accordance, an isolate with high-level resistant phenotype (> 256 mg L−1) showed mutations in both gyrB and nfxB.

The global transcription study of PAOMY-Mgm showed up-regulation of pfpI gene, which has been shown to provide protection to oxidative stress (Rodriguez-Rojas & Blazquez, 2009) and down-regulation of genes involved in iron trafficking and metabolism compared with PAO1. Repression of genes involved in iron metabolism have been reported in oxidative stress situation such as exposure to H2O2 (Chang et al., 2005) and can be explained as a protection mechanism used by the bacteria against Fenton-reaction, which requires iron and results in ROS production. Thus, the unrepaired DNA oxidative lesions that occur in PAOMY-Mgm during growth in LB seem to trigger an oxidative stress response. It has been reported in unicellular eukaryotes such as Saccharomyces cerevisiae that various types of DNA damage are capable of causing an increase in intracellular ROS, which will function as secondary signal for a generalized stress response (Rowe et al., 2008). Such a DNA damage-induced increase in intracellular ROS levels as a generalized stress response might function in prokaryotes as well, especially as ROS has been shown to act as a secondary signal for antibiotic stress in bacteria (Kohanski et al., 2010). Ciprofloxacin is one of the antibiotics that can stimulate the bacterial production of ROS (Morero & Argarana, 2009; Kohanski et al., 2010), and therefore we were interested in investigating the survival of PAOMY-Mgm mutator in competition with the wild-type PAO1 in the presence of this antibiotic. The mutators overtook the entire population after 3 days, and the tolerant population had just a marginal increase in the MIC of ciprofloxacin and showed moderate increases in the expression of efflux pumps. As no mutations in specific ciprofloxacin target genes or in efflux pumps were identified, mutations in genes responsible for low-level resistance to ciprofloxacin could be responsible for this phenotype. Few fold up-regulation of the efflux pumps characterizes the persister phenotype (Su et al., 2010), and an increased number of ‘persister mutants' were found in mutS mutant P. aeruginosa isolate (Mulcahy et al., 2010); therefore, occurrence of an increased percentage of persisters in the PAOMY-Mgm compared with PAO1 might be an alternative explanation of our findings. Further studies are needed to verify the oxidative stress response in P. aeruginosa GO mutators. It would be interesting in the future to study the effect of exogenous ROS sources on the expression levels of pfpI and of genes involved in iron metabolism in the double PAOMY-Mgm mutant.

In conclusion, by revealing the cooperation of MutM and MutY in P. aeruginosa, our findings provide new insights into the functionality of the GO system in P. aeruginosa and suggest that unrepaired DNA oxidative lesions are triggering an oxidative stress response in the bacteria.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Growth competition assay of PAO1 with PAOMYgm (mutY) and PAOMMgm (mutM).

Table S1. Strains and plasmids used in this study.

Table S2. Primers for construction of PAOMY-Mgm and sequencing of parC, parE, gyrA, gyrB and nfxB, and RT-PCR of pfpI and PA5148.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Acknowledgements

We thank Tina Wassermann for her efforts and excellent technical assistance. This study was supported by grant from The Danish Research Council for Technology and Production Sciences, through Grant 274-05-0117. ‘M.D.M. and A.O. are supported by the Ministerio de Ciencia e Innovación of Spain and Instituto de Salud Carlos III, through the Spanish Network for the Research in Infectious Diseases (REIPI C03/14 and RD06/0008)'. Transparency declarations: The authors have nothing to declare.

Footnotes

  • Editor: Rustam Aminov

References

View Abstract