Analysis of Efflux Pump Genes in β-lactam Resistant Clinical Isolates of Pseudomonas aeruginosa from a Tertiary Level Hospital in Ecuador

.- Pseudomonas aeruginosa is a nosocomial microorganism that causes a wide spectrum of infections and is known as one of the primary multi-resistant microorganisms against β-lactam antibiotics. One of the main resistance mechanisms found in P. aeruginosa is the efflux pumps. This study is aimed at characterizing this mecha - nism by analyzing the expression of four genes ( mexA, mexX, oprJ and oprM ) involved in antibiotic efflux pumps in Pseudomonas aeruginosa . Forty clinical isolates (20 resistant, 20 susceptible) were collected from the Bacte - riology Laboratory at the Carlos Andrade Marin Hospital, in Quito-Ecuador. Expression levels for the selected ge - nes were assessed by RT-qPCR assays using RpsL as a housekeeping gene for ΔΔCt adjusted relative quantitation analysis. The importance of efflux pumps as a resistance mechanism was corroborated through analysis of efflux pumps genes that showed overexpression in all phenotypically resistant isolates. Furthermore, phenotype/genotype analysis was performed comparing Antibiotic Susceptibility Testing (AST) results with expression profiles. Results for the mexA genotype showed correlation with the TPZ resistance phenotype and the mexX genotype with the IPM, MEM and FEP resistance phenotypes. In conclusion, the expression pattern of the efflux pump genes suggests re - sistance mechanisms that are due to horizontal transmission or pathogens spreading into the hospital environment.


INTRODUCTION
Pseudomonas aeruginosa is one of the main nosocomial microorganisms.It is responsible for 10% to 15% of nosocomial infections-such as pneumonia, urinary tract infections, wound infections and bloodstream infections-and it is mainly present in intensive care units and surgical wards (Lister et al. 2009).According to Guzmán-Blanco and Istúriz (2010), Pseudomonas aeruginosa infections have a prevalence of 16% in studies conducted in South American hospitals.In Ecuador, infectious episodes due to Pseudomonas aeruginosa rate at 23% of all β-lactam resistant episodes.
Pseudomonas aeruginosa shows a high level of resistance to the primary anti-pseudomonal antibiotics, such as β-lactams.Its genome is one of the largest of all microorganisms, allowing it to mutate and adapt to different stress conditions, such as exposure to antibiotics and through several complex resistance mechanisms (Meletis and Begkeri 2013).Resistance mechanisms are divided into three categories: acquired, adaptive and intrinsic.Adaptive resistance is due to environmental or nutritional stress, intrinsic resistance is caused by the low permeability of the outer membrane, while acquired mechanisms are represented by horizontal gene transfer and mutational events.The production of enzymes (β-lactamases), mutations in regulatory porines in the outer membrane and, most notably, overexpression of efflux pumps-encoding genes may result in a resistant phenotype (Breidenstein et al. 2011).Efflux pumps are part of the Nodular Resistance Division (RND), a tripartite system that includes a membrane fusion protein (MFP), an outer membrane factor (OMF) and a cytoplasmic membrane carrier.RND controls the inflow and expulsion of molecules from cytoplasm to periplasmic space.RND complex genes are arranged in operons in the P. aeruginosa chromosome.Twelve efflux pumps have been identified and are expressed in P. aeruginosa strains.MexAB-oprM, MexCD-oprJ, MexEF-OprN and MexXY have been widely studied, and all of them were identified in multiresistant isolates (Morita et al. 2012).Efflux pumps use antibiotics as substrate and expel them out of the cell by proton motive force using ions through the electrochemical gradient of the membrane, in a process known as chemiosmosis (Lister et al. 2009).
RT-qPCR assays allow RNA quantitation analysis.Their high sensitivity, reproducibility and efficiency have led qPCR to be the current main technique for gene expression analysis (Khan-Malek and Wang, 2011).Quantitative PCR (qPCR) uses specific primers and fluorescent probes.When the target sequence is detected, a fluorescent signal is emitted that has a higher intensity than the baseline-Threshold Cycle (Ct) (Nolan et al. 2006).Relative quantitation is the method used for measuring gene expression in quantitative PCR analysis.It relies on a housekeeping gene-whose expression is constant under different conditions-and utilizes the ratio of target gene expression over housekeeping gene expression for the quantitation of gene expression (Bolha et al. 2012).
The ΔΔCt adjusted method has been used for relative quantitation of gene expression in RT-qPCR.The method takes into consideration the assay´s Amplification Efficiency (AE) value.The Ct values of target and housekeeping genes are also considered in the equation and it allows for the calculation of gene expression ratios (Yuan et al. 2008).
The aim of this study was to implement RT-qP-CR systems for quantitation of the expression of gene encoding efflux pumps in clinical isolates of Pseudomonas aeruginosa and to compare expression ratios in susceptible and resistant isolates in order to evaluate the role of efflux pumps in antibiotic resistance mechanisms.

Collection and Storage of Pseudomonas aeruginosa
Isolates.-FortyP. aeruginosa isolates were collected from the Microbiology Laboratory at the Carlos Andrade Marin Hospital in Quito-Ecuador.Demographics and collection site information for patients and isolates for this study are shown in Table S1 (Supplementary Information).Antimicrobial Susceptibility Testing (AST) was performed in the hospital´s Bacteriology Laboratory by qualified and experienced microbiology technicians in accordance with to current standards (CLSI).This study was conducted with 20 resistant and 20 susceptible isolates (AST profiles for all isolates are shown in Table S2 (Supplementary Information).Suspensions in 15% v/v glycerol/water were cryopreserved in our laboratory at the Universidad de las Fuerzas Armadas-ESPE.
RNA Extraction, Purification and Quantitation.-Priorto RNA extraction, 100µl of the cellular suspension in 15% glycerol was cultured in 4ml of Brain Heart Infusion and incubated at 37ºC for 18 hours.Afterwards, the recommended protocol for the Purelink RNA Mini Kit (Ambion), DNa-se treatment was performed using TURBO DNase (Ambion), and a phenol/chloroform RNA extraction protocol was then performed (Jacobs and White, 2013).Purified RNA was quantified using a Qubit 1.0 fluorometer (Invitrogen) with RNA BR kit (Invitrogen).Finally, an aliquot of the extracted RNA was run in a 2% agarose electrophoresis gel in order to visually confirm RNA purification.

Design of Primers and
Probes.-Five pairs of primers and four probes (Table S3 Supplementary Information) were designed with aid of the Primer Express v3.0 software (Applied Biosystems).The additional bioinformatics tools BLAST, Clustal Omega and Oligo Analyzer were used for fine tuning of design parameters (melting temperature (Tm), primer size, GC content, homo dimers and hetero dimers formation).
RT-qPCR.-All 40 strains of P. aeruginosa were analyzed for each of the five genes included in the study.TaqMan RNA-to-Ct 1-Step Kit (Applied Biosystems) was used for mexA, oprM, oprJ and rpsL genes (housekeeping gene used as normalizer), and the Power SYBR Green RNAto-Ct 1-Step kit (Applied Biosystems) was used for mexX.The assays were carried out in a 7300 Real Time PCR System (Applied Biosystems).
For the mexA, oprM, oprJ and rpsL amplification assays, a 10µl final reaction volume was used.The reaction mix contained 5µl of 2X Taqman RT-PCR Mix, 0.9µl of 10µM forward and reverse primers, 1µl of 2000nM MGB probe, 0.25µl of TaqMan RT enzyme, 1µl of 10ng/µl RNA, and 0.95µl of nuclease-free water.The thermal cycler program included a retro transcription step at 48ºC for 20 minutes, an initial denaturation step at 95ºC for 10 minutes, followed by 35 cycles consisting of a denaturation step at 95ºC for 15 seconds, an annealing step at 51ºC (for mexA and oprJ), 59ºC (for oprM) and 53ºC (for rpsL) for 30 seconds, and a final extension step at 60ºC for 30 seconds.
For the mexX gene, the 10µl final reaction mixture contained 5µl of 2X Power SYBR green mix, 0.2µl of 10µM forward and reverse primers, 0.08µl of RT enzyme mix, 1µl of 10ng/µl RNA and 3.52µl of nuclease-free water.The PCR thermal cycler program consisted of a reverse transcription step at 48ºC for 20 minutes, an initial denaturation step at 95ºC for 10 minutes, followed by 35 cycles of a denaturation step at 95ºC for 15 seconds, and an annealing step at 55ºC for 30 seconds.A final extension step at 60ºC for 30 seconds was inclu-ded.For Tm analysis, the thermal cycler was programmed as follows: denaturation at 95ºC for 15 seconds, renaturation at 60ºC for 1 minute, denaturation at 95ºC for 15 seconds and final renaturation at 60ºC for 15 seconds.A dissociation curve was used to determine if the predicted Tm (87ºC) for the mexX amplicon corresponded to the actual Tm of the cDNA PCR fragment obtained in the assay.
Amplification Efficiency and Statistical Analyses.-Forcalculation of the assay´s Amplification Efficiency (AE), two clinical strains were used as assay controls, corresponding to a resistant and a susceptible isolate (Wong and Medrano 2005).Four RNA dilutions were prepared from the original concentrated RNA (10ng/µl), by adding DEPC treated water to RNA final concentrations of 5ng/µl, 1ng/ µl, 0.5ng/µl and 0.1ng/µl.A RT-qPCR reaction was run for all gene systems with the above RNA concentrations and Ct values were obtained for each PCR run.Ct values of the housekeeping gene (rpsL) were subtracted from the Ct values of the four target genes, and a ΔCt value for each system was obtained.A two-sample T-test was applied for comparison of Ct's for phenotypically resistant and susceptible strains in each targeted gene.A p-value of 0.05 was considered for statistical significance.
Relative Quantitation of Gene Expression.-Ct values were obtained and tabulated from the RT-qP-CR for the five genes (n= 40 isolates).The mathematical model proposed by Yuan et al. (2008) was used for calculation of expression ratios.
ΔΔCT adjusted = µ1 x AE1 -µ2 x AE2 -µ3 x AE3 + µ4 x AE4 (1) Being: µ1: Target gene Ct for sample X µ2: Control gene Ct for sample X µ3: Target gene Ct of control sample µ4: Control gene Ct of control sample AE1: Amplification Efficiency of target gene for sample.AE2: Amplification Efficiency of control gene for sample.AE3: Amplification Efficiency of target gene for control sample.AE4: Amplification Efficiency of control gene for control sample.
Phenotype/Genotype Correlation.-Acomparison between dominant (most common) phenotypes and genotypes was performed to esta-blish the correlation and associations among the genotypes responsible for antibiotic resistance.

RESULTS
RT-qPCR.-The assay tested the purified RNA of all forty P. aeruginosa strains for each of the five target genes.Ct values, amplification and dissociation curves were obtained through the 7300-software system (Applied Biosystems) from all forty P. aeruginosa isolates.Amplification curves for mexA and mexX resistant isolates (Figures 1A and  1B) had a lower Ct than susceptible ones.A melting curve analysis for mexX (87 ºC) was used for confirmation of specific amplification of the PCR fragment (Figure 1C).Finally, the constant expression of the housekeeping gene (rpsL) is showed for both susceptible and resistant isolates (Figure 1D).Amplification Efficiency and Statistical Analyses.-AEvalues were obtained from the resistant and susceptible isolates used as controls for each of the five gene expression systems.The Ct va-lues obtained in all serial dilutions were analyzed by a two-sample T-test (Table 1 and Table 2).
Gene Expression.-Based on Yuan et al. (2008), Statistical methods for efficiency adjusted real-time PCR quantification, gene expression was quantified using Eq (1) and (2) (see Materials and Methods).Gene expression ratios for all strains and genes are showed in Table S4 (Supplementary information).A comparison of the expression ratios of resistant and susceptible phenotypes (average) was performed for each gene as shown in Figure 2. As expected, the expression ratios were higher in resistant isolates in comparison to susceptible ones.Over expression of efflux pump genes in clinical strains of P. aeruginosa was thus confirmed, stressing its importance as one of the main resistance mechanisms against β-lactam antibiotics.Furthermore, expression ratios for oprM and oprJ genes are lower than expression ratios for mexA and mexX genes, in spite of over-expression resistant isolates being greater than that in susceptible ones.Phenotype/Genotype Correlation.-Correlationbetween genotype and phenotype was performed using gene expression results as well as AST profiles.The results presented in Figure 3 showed mexA and mexX as dominant genotypes, data that was confirmed by the higher expression ratios shown in Table S4 Supplementary information.Out of the 20 resistant isolates tested in this study, in 95% of the isolates the mexA gene is overexpressed and in 100% the mexX gene is overexpressed, showing the presence of active genes encoding efflux pumps in clinical isolates of P. aeruginosa.
Furthermore, AST analysis showed that the main resistance phenotypes were to MEM (meropenem) and to IPM (imipenem), with a 100% resistance profile, and to TPZ (tazobactam) and to FEP (cefepime), with a 95% resistance profile (Figure 4).
AST profiles for susceptible isolates showed that more than 50% of the isolates express resistance to at least one β-lactam antibiotic.When a comparison performed of the antibiotic susceptibili-ty profiles of the resistant strain with expression ratios for both susceptible and resistant isolates (Figure 5), we concluded that the mexA overexpression genotype is closely related to TPZ phenotype, while the mexX over expression genotype correlates to IMP, FEP and MEM resistance.

DISCUSSION
In this study, we developed RT-qPCR assays for relative quantification of the expression of P. aeruginosa genes encoding efflux pumps (mexA, mexX, oprJ and oprM).We then analyzed changes in gene expression in both resistant and susceptible clinical isolates.This technique allowed us to confirm the important role of efflux pumps as one of the main resistant mechanisms expressed in P. aeruginosa clinical isolates in the collection  analyzed.This finding is supported by the expression analysis of four genes performed in this study.Despite different expression ratios, over-expression in resistant isolates is much higher than in susceptible ones (Bolha et al. 2012).
The RT-qPCR technique developed in this study is as a robust evidence for gene expression analysis in susceptible and resistant P. aeruginosa The model proposed by Yuan et al. (2008) considers amplification efficiency to be reliable method for relative quantification, since 100% amplification value is not possible to achieve because due to the different substrate conditions and the efficiency of the nucleic acid extraction kit used.Usually, amplification efficiencies for RT-qPCR assays are within 70-100%.We obtained an average efficiency of 90% in all PCR systems tested.Expression ratios were considerably higher in all resistant isolates.mexA and mexX genes are the dominant genotypes, being over-expressed in almost all resistant isolates.Overexpression of efflux pump genes might suggest the presence of a nosocomial strain of P. aeruginosa in the hospital environment.However, it is important to note that horizontal transmission might be another mechanism for the spreading of infection among patients coming from another health facility.
The evident difference in the expression ratios of mexA and mexX in comparison with oprM and oprJ is related with their function in the RND complex.
The mexA and mexX genes are important Membrane Fusion Proteins (MFP) the pump structure.They act as assemble adaptors that ensure proper channel conformation in order to connect the cytoplasm and the membrane.Therefore, the overexpression of MFP's in all strains suggests a well-conformed and functional efflux pump.Low expression of the oprJ gene in comparison to the other three sys-  tems analyzed is due to the fact that it is acquired from plasmids, rather than encoded in the P. aeruginosa genome (Lister et al. 2009).
The importance of analyzing the phenotype and genotype correlation in multiresistant P. aeruginosa was first stressed by Pommerenke and collaborators (2010).In the present study, we compared AST profiles for susceptible and resistant clinical isolates to the isolate's expression ratios.Our findings show that a number of susceptible isolates have measurable gene expression for the resistance-related genes mexA and mexX and show resistance to one or more of the phenotypically dominant antibiotics (tazobactam (TPZ), imipenem (IMP), cefepima (FEP) and meropenem (MEM)).
The most remarkable case was the resistant isolate R11, which showed mexX overexpression, providing resistance to ATM, MEM and IPM.Phenotypically susceptible isolates (5, 7, 8, 10, 13, 15, and 20), which presented resistance to TPZ, showed high mexA expression.Therefore, we were able to establish a correlation for the presence of the mexA genotype in the TPZ-resistant phenotype and a correlation of the mexX genotype the ATM, MEM and IPM-resistant phenotypes.

CONCLUSIONS
Gene expression (mexA, oprM, mexX and oprJ) pattern analysis in resistant isolates confirmed the predominate role of efflux pumps as one of the most important β-lactam resistance mechanisms.
Efflux pumps encoding genes are mainly acquired by mutations and horizontal gene transfer among strains present in the inner hospital environment.P. aeruginosa resistant isolates play an important role on this acquired resistance mechanism inside the hospital.This study found the same genetic pattern in more than 95% percent of all clinical isolates studied.However, larger studies are recommended in order to achieve a better comprehension of the mechanisms involved in β-lactam resistance in Pseudomonas aeruginosa clinical isolates.S4.Expression ratios obtained for all efflux pumps encoding genes in susceptible and resistant strains.
Table S4.Expression ratios obtained for all efflux pumps encoding genes in susceptible and resistant strains.

Figure 1
Figure 1 a) Duplicate amplification curves for mexA gene, difference between resistant (red/violet) and sensitive (dark purple/ green) isolates.b) Duplicate amplification curves for mexX gene, difference between resistant (blue/green) and sensitive (violet/fuchsia) strains are showed.c) Duplicate dissociation curve for mexX gene (red/green), showing a melting temperature of 87 ºC.d) 10 amplification curves for rpsL gene with constant expression for resistant (purple spectrum) and sensitive isolates (green spectrum).

Table 1 .
Amplification Efficiency for each gene system

Table 2 .
Two-sample t-test results for resistant and sensitive isolates (p-value is < 0.05 in all cases)

Table 2 .
Two-sample t-test results for resistant and sensitive isolates (p-value is < 0.05 in all cases)