HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Savli, H. Articles by Vahaboglu, H. Search for Related Content PubMed PubMed Citation Articles by Savli, H. Articles by Vahaboglu, H. Agricola Articles by Savli, H. Articles by Vahaboglu, H.

Expression stability of six housekeeping genes: a proposal for resistance gene quantification studies of Pseudomonas aeruginosa by real-time quantitative RT-PCR

^1–3Tibbi Biyoloji AD¹, Mikrobiyoloji ve Klinik Mikrobiyoloji AD² and Enfeksiyon Hastaliklari ve Klinik Mikrobiyoloji AD³, Kocaeli Universitesi, Tip Fakultesi, Sopali Ciftligi, 41900 Kocaeli, Turkey ⁴Genetik AD, DETAE, Istanbul Universitesi, Istanbul, Turkey

Correspondence Haluk Vahaboglu vahabo{at}hotmail.com

Received November 22, 2002
Accepted January 28, 2003

	Abstract

TOP Abstract INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
Constantly expressed genes are used as internal controls in relative quantification studies. Suitable internal controls for such studies have not yet been defined for Pseudomonas aeruginosa. In this study, the genes ampC, fabD, proC, pbp-2, rpoD and rpoS of P. aeruginosa were compared in terms of expression stability by real-time quantitative RT-PCR. A total of 23 strains with diverse resistance phenotypes were studied. Stability of expression among the housekeeping genes was assessed on the basis of correlation coefficients, with the best-correlated pair accepted as being the most stable one. Eventually, proC and rpoD formed the most stable pair (r = 0.958; P < 0.001). Next, in four ciprofloxacin-selected nfxC-like mutants, levels of oprD, oprM and oprN mRNA were compared with those of their wild-type counterparts. The comparison was made after correcting the raw values by the geometric mean of the internal control genes proC and rpoD. The level of oprN mRNA was significantly up-regulated, while the oprD gene was down-regulated (although this difference was statistically insignificant), in the mutants. This expression pattern was consistent with that of the expected expression profile of nfxC-type mutants; this experiment therefore lends further support to the use of proC and rpoD genes simultaneously as internal controls for such studies.

	INTRODUCTION

TOP Abstract INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
Pseudomonas aeruginosa is an important nosocomial pathogen, particularly in intensive care units (Yucesoy et al., 2000). Members of this species are inherently resistant to a range of antibiotics. In addition, they are capable of conferring resistance to others by shifting the regulation levels of various innate genes (Livermore, 1992).

Porins (particularly porin D) and efflux pumps of the resistance-nodulation-division (RND) family are the currently recognized genetic systems that are associated significantly with antibiotic resistance (Okamoto et al., 2001). The exact links between the expression patterns of these and other resistance phenotypes have not been fully elucidated, and inconsistent conclusions appear in the literature (Masuda et al., 2000; Morita et al., 2001; Sumita & Fukasawa, 1996; Trias et al., 1989). Therefore, studies focusing on this topic still attract considerable interest.

Levels of expression of these genes have traditionally been studied by Western blotting with the aid of mAbs. Unfortunately, this method is unable to help in studying different proteins simultaneously and so is weak in relative comparisons. Moreover, mAbs are not commercially available. The quantification of mRNAs by real-time RT-PCR has been used with great success in other fields (Johnson et al., 2000; Wang & Brown, 1999). We suggest that this highly sensitive method may also be useful in relative comparison of resistance gene expression.

The reliability of a relative comparison depends largely on the normalization of unwanted variations between samples. Constantly expressed genes, often selected from among housekeeping genes, are used as internal controls for normalization of the results. The proportion of mRNA of constantly expressed genes in the total cellular RNA is assumed to be equal between different samples. Normalization of raw values by means of internal controls therefore serves to eliminate sample-to-sample variation of the RNA isolation and reverse transcription steps and, even more importantly, serves to eliminate variations in total transcriptional activity between cells.

For this purpose, six housekeeping genes were compared in this study. These were pyrroline-5-carboxylate reductase (proC), malonyl CoA : acyl carrier protein (ACP) transacylase (fabD), sigma factors RpoD (rpoD) and RpoS (rpoS), penicillin-binding protein 2 (pbp-2) and chromosomal beta-lactamase (ampC). The expression stability of these housekeeping genes was first investigated as proposed by Vandesompele et al. (2002) and, later, the levels of oprM, oprN and oprD mRNA were compared in a set of P. aeruginosa strains.

	METHODS

TOP Abstract INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
Strains, resistance tests and beta-lactamase assays.

We selected 17 P. aeruginosa strains from a set obtained from university hospitals in four different geographical regions of Turkey. The bacteria had been identified in those universities by various identification systems. The strains were, nevertheless, reidentified in our institute by classical methods and, if required, by the non-fermenter ID panel of the Becton Dickinson system (Diagnostic Instrument Systems).

Of the 17 strains selected, 10 were fully susceptible to antibiotics of the main classes, while the others were of different resistance phenotypes. We challenged the susceptible strains in Mueller–Hinton (MH) broth with ciprofloxacin (0.1 µg ml^-1) in order to obtain resistant variants. Consequently, six isogenic mutants with various resistance patterns were selected. These mutants, as well as their counterparts, were included in the study group.

Resistance patterns were determined by the disk diffusion method on MH agar plates. MICs were determined by either E-test strips on MH agar plates or by the agar dilution method as described by the NCCLS. Antibiotic disks and MH agar were obtained from Oxoid while E-test strips were sourced from AB Biodisc. Powder forms of the antibiotics were obtained as gifts from the respective companies.

Beta-lactamases were analysed as described elsewhere (Vahaboglu et al., 1998). Extracts obtained by freezing and thawing of dense bacterial suspensions were applied to polyacrylamide gels with ampholytes ranging from pH 3 to 10. Nitrocefin overlay and migration relative to TEM-1 and SHV-1 standards enabled us to evaluate the pI values of beta-lactamases. Precise identification of the beta-lactamases depended on sequence analysis of PCR products as described previously (Vahaboglu et al., 1998).

Random amplified polymorphic DNA (RAPD) typing.

Clonal variability was further ensured by RAPD typing of eight selected strains, two from each region. DNA was isolated from fresh overnight agar cultures. A loopful of bacteria was homogenized in 50 µl TE buffer (10 mM Tris/HCl, 1 mM EDTA, pH 8) and lysis was accomplished with 500 µl guanidium thiocyanate (6 M) plus 0.5 % sodium lauroylsarcosine for 10 min at room temperature. DNA was extracted first by phenol/chloroform and then by chloroform/isoamyl alcohol (24 : 1, v/v) and then precipitated with 0.1 vols sodium acetate (3 M, pH 5.4) plus an equal volume of 2-propanol at room temperature. Precipitates were collected by centrifugation (10 min at 12 000 g) and then the pellets were washed twice with 70 % ethanol, air-dried for 2 min and resuspended in 30 µl double-distilled water.

RAPD PCR was performed as described elsewhere (Kerr et al., 1995; Renders et al., 1996). The primers were ERIC-1R (5'-AAGCTCCTGG GGATTCA-3') and ERIC-2 (5'-AAGTAAGTGACTGGGGTGAGCG-3'). Master mixtures were prepared as described in the above references. However, the amplification program was modified as follows: one cycle of denaturation for 5 min at 95 °C followed by 25 cycles of 3 min at 39 °C (with an increase of 0.3 °C every cycle), 2 min at 72 °C and 1 min at 94 °C and 30 cycles of 2 min at 44 °C, 3 min at 72 °C and 1 min at 94 °C. A final extension for 1 h at 72 °C completed the procedure.

PCR products were separated on a 2 % agarose gel and visualized by ethidium bromide staining. For better resolution, they were also run on a 6 % acrylamide/bis-acrylamide gel (data not shown). The banding patterns of both gels were analysed by the freely distributed gel analysis software LabImage (version 2.62). Molecular sizes of the bands were calculated by this software relative to the marker DNA.

RNA isolation and reverse transcription.

Total RNA was isolated from 5 ml fresh overnight (approx. 18 h) broth culture (MH broth) by using the NucleoSpin RNA II kit (Macherey-Nagel), as described by the manufacturer. Genomic DNA was eliminated by RNase-free DNase I treatment during the isolation procedure. Finally, RNAs were eluted from the silica membranes in a volume of 40 µl diethyl pyrocarbonate-treated double-distilled water. The A₂₆₀ of the resulting RNA solution was between 1 and 10. Reverse transcription was performed at 42 °C for 90 min by using random hexamer primers so as to obtain cDNA copies of mRNAs (2 µl) with 100 IU MMuLV reverse transcriptase (MBI Fermentas) in 20 µl total volume. Concentrations of cDNAs were adjusted on a LightCycler (Roche Diagnostics). For every sample, 1 µl cDNA and 9 µl SYBR Green I (the same concentration as indicated by the manufacturer for the PCR assay) were mixed in capillary tubes. After incubation at 95 °C for 5 min, fluorescence emissions were read at 55 °C with the real-time fluorometry facility of the LightCycler. This enabled us to compare the total cDNA concentrations of the samples with the control transcript, which was approximately 1 mg ml^-1 at the highest dilution. Concentrations of cDNAs of the samples were adjusted to a level close to the second dilution (10^-1) of the control cDNA. This adjustment was critical for performing successful calculations. The aim was to keep the cDNA concentrations of the samples between the concentrations of the controls in order to avoid large variations during calculations by the LightCycler software.

Real-time PCR.

The sequences of the genes studied were obtained from GenBank and the primers were designed with the aid of the OLIGO software (version 5.0; Molecular Biology Insights). The sequences of the primers are shown in Table 1.

PCR was performed in the LightCycler in capillary glass tubes with the LightCycler FastStart DNA Master SYBR Green I kit (Roche). Work was always carried out on desktop coolers (pre-cooled to 4 °C). Master mixtures were prepared exactly as recommended by the manufacturer, except for the concentration of Mg²⁺. The final concentrations of Mg²⁺ and primers were respectively 2.5 mM and 50 pmol per reaction.

The control cDNA was from P. aeruginosa ATCC 27853 and the primers of the control reactions were specific for pbp-2. An arbitrary concentration value of 1.5 x 10⁴ copies of the pbp-2 gene was assigned to the control transcript. Tenfold dilutions of this down to 15 copies of the pbp-2 gene were always included in the reactions.

PCR was accomplished after a 5 min activation and denaturation step at 95 °C, followed by 45 cycles of 15 s at 95 °C, 10 s at 60 °C and 15 s at 72 °C.

Primer dimers and other artefacts were evaluated by melting curve analysis and eventually only dimer- and artefact-free reactions were considered valid. Results were read with the ‘second derivative maximum’ algorithm of the software provided. The LightCycler software generated a standard curve by plotting ‘crossing cycle number’ versus logarithms of the given concentrations for each control. Eventually, a regression line was drawn between these points. The software calculated the concentrations of the studied genes with the aid of this standard curve.

Statistical analysis.

The stability of mRNA expression was assessed by calculation of Spearman's correlation coefficients of the raw concentration data with the aid of the statistical package SPSS (version 9.0). The best-correlated pair was considered to be the most stable one.

Stability was further evaluated by a freely distributed MS Excel application (geNorm). Detailed information on this application can be obtained from Vandesompele et al. (2002). This approach assumes that minimally regulated, stably expressed genes stay in a constant ratio to each other. In other words, in a given set of genes, it must be the pair of most stable genes that will be able to keep the ratio to each other constant in different samples. Importantly, co-regulated genes are exceptions to this assumption and they are not included.

The applet geNorm helps to calculate the gene expression stability measure (M), which is the mean pair-wise variation for a gene from all other tested control genes (Vandesompele et al., 2002). A higher value of M means greater variation in expression. The stepwise exclusion of genes with the highest M values allows the ranking of the tested genes according to their expression stability. The proposed threshold for eliminating a gene as unstable was an expression stability measure of 0.5.

Raw quantities were corrected by dividing a value by the geometric mean of proC and rpoD genes of the same sample. Relative comparisons were done between corrected values with the ANOVA test for significance.

	RESULTS AND DISCUSSION

TOP Abstract INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

 
A total of 23 strains were included in the study. The group was composed of 17 wild-type strains and six ciprofloxacin-selected, isogenic mutants. Clonal heterogeneity was investigated by DNA typing of eight of these wild-type strains. One sample did not provide a readable pattern, probably because of inhibitors carried over from the DNA isolation step. However, the other seven were amplified successfully and the fingerprints were sufficiently polymorphic to confirm the clonal diversity (Fig. 1).

View larger version (62K): [in this window] [in a new window]   Fig. 1. (a) RAPD fingerprints (negative exposure). Lanes: M, molecular size marker (100 bp DNA ladder); 1–7, selected strains (strains 2 and 5–7 are the parents of nfxC mutants). (b) Fingerprints according to the lengths (bp) of fragments calculated by gel analysis software. Lanes are the same as in (a).

Of the 17 wild-type strains, seven were already resistant to various antibiotics. It is noteworthy that three of these did not produce a beta-lactamase other than the chromosomal one (Table 2). These strains were probably resistant because of activated porin and/or efflux systems. Of the six ciprofloxacin-selected mutants, two were resistant only to the fluoroquinolones, while the other four expressed nfxC-like multiple-resistance phenotypes (Table 3). In addition to fluoroquinolones, the MIC of imipenem increases among nfxC mutants because of the concurrent down-regulation of oprD (Maseda et al., 2000; Ochs et al., 1999). The resistance phenotypes of the mutants were in agreement with this.

Interestingly, the MICs of ceftazidime were variable; one was increased and one decreased while the other two were unchanged. One study reported an unexplained decrease in the ceftazidime MICs of nfxC mutants (Maseda et al., 2000). However, another study showed that the functional subunit of the mexEF–oprN operon is not related to beta-lactam hypersusceptibility (Maseda et al., 2000). These observations suggest the existence of other co-operating systems responsible for beta-lactam susceptibility or resistance in nfxC mutants. This issue remains unresolved.

The housekeeping metabolism of prokaryotes has been shown to be highly variable (Vandecasteele et al., 2001), so genes expressed stably under one condition might not be so under others. Stability in terms of mRNA expression in prokaryotic cells, therefore, should be tested under equivalent conditions with the investigated setting. In other words, internal controls intended for use in resistance gene quantification studies should be tested in strains with changing resistance phenotypes. The diversity in the resistance phenotypes of this study group fulfils this requirement.

Six genes were compared in this study group. In order to avoid co-regulated genes, we carefully selected genes that are distantly related in metabolic function and chromosomal order. The selected genes participate in critical functions. pbp-2 has a central role in peptidoglycan metabolism, while it has some relation to the rod-shape-determining protein. ampC is involved in cell-wall recycling (Jacobs et al., 1995). The sigma factor-encoding rpoD is a critical housekeeping gene (Schnider et al., 1995). proC is involved in amino acid biosynthesis, while fabD is involved in a different class of metabolic function (Kutchma et al., 1999). However, their metabolic importance was not the only reason for selecting these genes. Equally important is that they all were shown in an Escherichia coli DNA array study to be expressed at levels comparable to outer-membrane proteins (Wei et al., 2001). These data led us to assume that these genes might also be expressed in sufficient quantities in P. aeruginosa, a further advantage in optimization of the PCR test.

The raw quantities of mRNA of the six genes studied obtained by RT-PCR are shown in Table 4. Correlation coefficients indicated proC and rpoD as the most significant pair (r = 0.958; P < 0.001). Similarly, the stepwise exclusion of the genes with the highest M values by geNorm left proC (M = 0.36) and rpoD (M = 0.36) as the most stable genes.

Next, the levels of oprD, oprM and oprN mRNA of the four nfxC-like isogenic mutants were compared with those of their wild-type counterparts. In this experiment, the geometric mean of the levels of proC and rpoD in a sample was its normalization factor. Comparison of the normalized quantitative values of these genes is shown in Table 5. The mean level of oprN was 1.3 times higher in the mutants, while the mean level of oprD was 1.28 times lower. Values were comparable for all the mutants. Interestingly, the statistical comparison was significant only for oprN concentrations.

Data on expression of these resistance genes in the literature have to date been obtained by immunoblotting. Evaluations were dependent on crude differences and lacked precise numerical values. Therefore, comparison of the results of this study with the literature was not possible. Moreover, immunoblotting indicates differences in the level of mature protein. It is known that mRNA expression predicts mature protein levels poorly and there may be up to 30-fold differences (Gygi et al., 1999). Transcription and translation are regulated individually to some extent. The levels of mRNA of these genes may, therefore, not be exactly in accord with protein levels. However, the quantification results with RT-PCR in this study were in agreement with the expected expression profile of nfxC-type mutants.

Resistance due to the altered regulation of intrinsic genes in P. aeruginosa is not well understood. This type of resistance may depend on the regulation of a more composite co-regulated network of multiple operons as well as the ‘quorum-sensing systems’ of P. aeruginosa (Kohler et al., 2001; Poole, 2001). Unfortunately, our understanding of this is extremely limited due to the limited power of the methods used at present. Therefore, further studies using new methods deserve increased scientific interest. We believe that real-time quantification with the selection of suitable internal control genes will facilitate studies and provide new insights into the regulatory alterations of innate genes and the multiple antibiotic resistance problem of P. aeruginosa.

This study showed that proC and rpoD form the most stable pair in a set of clonally unrelated P. aeruginosa strains with diverse resistance phenotypes. Thus, this pair may be used as internal controls in relative comparison studies of resistance genes in P. aeruginosa.

	Acknowledgments

 
We are grateful to Dr Serhat Unal (Hacettepe Universitesi), Dr Mehmet Ali Inal (Ege Universitesi) and Dr Hakan Leblebicioglu (Ondokuz Mays Universitesi) for providing strains and to Jason John Nash and Victor L. Yu for English reading of this manuscript.

	REFERENCES

TOP Abstract INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES