Antimicrobial susceptibility and virulence genes of clinical and environmental isolates of Pseudomonas aeruginosa

Background Pseudomonas aeruginosa is ubiquitous, has intrinsic antibiotic resistance mechanisms, and is associated with serious hospital-associated infections. It has evolved from being a burn wound infection into a major nosocomial threat. In this study, we compared and correlated the antimicrobial resistance, virulence traits and clonal relatedness between clinical and fresh water environmental isolates of P. aeruginosa. Methods 219 P. aeruginosa isolates were studied: (a) 105 clinical isolates from 1977 to 1985 (n = 52) and 2015 (n = 53), and (b) 114 environmental isolates from different fresh water sources. All isolates were subjected to ERIC-PCR typing, antimicrobial susceptibility testing and virulence factor genes screening. Results Clinical and environmental isolates of P. aeruginosa were genetically heterogenous, with only four clinical isolates showing 100% identical ERIC-PCR patterns to seven environmental isolates. Most of the clinical and environmental isolates were sensitive to almost all of the antipseudomonal drugs, except for ticarcillin/clavulanic acid. Increased resistant isolates was seen in 2015 compared to that of the archived isolates; four MDR strains were detected and all were retrieved in 2015. All clinical isolates retrieved from 1977 to 1985 were susceptible to ceftazidime and ciprofloxacin; but in comparison, the clinical isolates recovered in 2015 exhibited 9.4% resistance to ceftazidime and 5.7% to ciprofloxacin; a rise in resistance to imipenem (3.8% to 7.5%), piperacillin (9.6% to 11.3%) and amikacin (1.9% to 5.7%) and a slight drop in resistance rates to piperacillin/tazobactam (7.7% to 7.5%), ticarcillin/clavulanic acid (19.2% to 18.9%), meropenem (15.4% to 7.5%), doripenem (11.5% to 7.5%), gentamicin (7.7% to 7.5%) and netilmicin (7.7% to 7.5%). Environmental isolates were resistant to piperacillin/tazobactam (1.8%), ciprofloxacin (1.8%), piperacillin (4.4%) and carbapenems (doripenem 11.4%, meropenem 8.8% and imipenem 2.6%). Both clinical and environmental isolates showed high prevalence of virulence factor genes, but none were detected in 10 (9.5%) clinical and 18 (15.8%) environmental isolates. The exoT gene was not detected in any of the clinical isolates. Resistance to carbapenems (meropenem, doripenem and imipenem), β-lactamase inhibitors (ticarcillin/clavulanic acid and piperacillin/tazobactam), piperacillin, ceftazidime and ciprofloxacin was observed in some of the isolates without virulence factor genes. Five virulence-negative isolates were susceptible to all of the antimicrobials. Only one MDR strain harbored none of the virulence factor genes. Conclusion Over a period of 30 years, a rise in antipseudomonal drug resistance particularly to ceftazidime and ciprofloxacin was observed in two hospitals in Malaysia. The occurrence of resistant environmental isolates from densely populated areas is relevant and gives rise to collective anxiety to the community at large.

P. aeruginosa is considered as one of the harmless bacterial skin flora (Cogen, Nizet & Gallo, 2008).However, once inside the host, depending on the route of entry, it may express a series of pathogenic mechanisms.Flagella, pili and lipopolysaccharide are responsible for bacterial motility and adhesion; type I, II and III secretion systems, phenazine system and lectins are responsible for invasion and dissemination; latency and antimicrobial resistance are due to quorum-sensing and biofilm formation (Kipnis, Sawa & Wiener-Kronish, 2006).
Antibiotic resistance constitutes one of the most serious threats to the global public health and impacts all aspects of therapeutics, animal husbandry and agriculture; it is natural, ancient, and hard wired in the microbial pan-genome (Bhullar et al., 2012).Since the discovery of penicillin (Fleming, 2001), various natural and synthetic antimicrobials have been developed (Bassetti et al., 2013), but the rapid emergence of resistant bacteria in contrast to the slow development of drugs has resulted in a nearly ''empty'' antibiotic pipeline (Spellberg et al., 2008).Approximately 8% of all healthcare-associated infections in the USA are caused by P. aeruginosa, and 13% of them were found to be multidrug resistant (MDR) (CDC, 2013).P. aeruginosa is intrinsically resistant to many antimicrobials (Wroblewska, 2006) due to its low outer membrane permeability which is 100-fold less than Escherichia coli (Angus et al., 1982).Selective pressure due to antipseudomonal therapy and especially the use of imipenem has resulted in significantly higher risk of emergence of resistance than the use of ciprofloxacin or piperacillin, but ceftazidime had the lowest risk (Carmeli et al., 1999).Mechanisms responsible for the natural resistance of P. aeruginosa are: (i) efflux pumps, (ii) AmpC β-lactamase, (iii) loss of OprD porin and iv) mutations in the topoisomerase II and IV genes (Livermore, 2002), as well as acquired resistance due to aminoglycoside-modifying enzymes (Poole, 2005) and β-lactamases (class A, B and D) (Potron, Poirel & Nordmann, 2015).
The objectives of this study were to (a) determine the clonal relatedness of clinical and environmental isolates of P. aeruginosa, (b) test the antimicrobial susceptibility of clinical isolates from different isolation periods: 1977 to 1985 and 2015, (c) evaluate the antimicrobial susceptibility of environmental isolates recovered from fresh water sources in Malaysia in 2015, and (d) investigate the prevalence of virulence factor genes in both clinical and environmental isolates.

Bacterial isolates
One hundred and five non-duplicate clinical isolates of P. aeruginosa were obtained from 2 hospitals at different isolation periods: 52 isolates, 1977 to 1985 from the University Malaya Medical Centre (UMMC), Kuala Lumpur; and 53 from Hospital Sultanah Aminah (HSA), Johor Bahru (southern Malaysia) in 2015.One hundred and fourteen environmental isolates recovered in 2015 from different fresh water sources (ponds, waterfall, drains, well water, pools, paddy field and lakes) (Supplemental Information 1).

ERIC-PCR typing
Genomic DNA of the bacteria was extracted based on the approach as described previously (Teh, Chua & Thong, 2010).Strain diversity was determined by ERIC-PCR using primers as described (Versalovic, Koeuth & Lupski, 1991).PCR was performed in a final volume of 25 µl reaction mixture containing sterile MilliQ water, 4 µl 10× DreamTaq buffer, 0.25 mM dNTP mix, 50 pmol of each primers, 3.75 mM MgCl 2 , 2 U of Taq polymerase (Thermo Scientific, US) and 100 ng of template DNA.The reaction mixture was denatured at 95 • C for 7 min, and then subjected to 30 cycles at 90 • C for 30 s, annealing at 52 • C for 1 min, extension at 65 • C for 8 min and a final extension at 65 • C for 16 min (Szczuka & Kaznowski, 2004).P. aeruginosa PAO1 and P. aeruginosa ATCC R 27853 were used as controls.
Fingerprint analysis was carried out by the GelJ software (Heras et al., 2015) and similarity was calculated by the Dice coefficient and cluster analysis using the unweighted pair group method with average linkages (UPGMA).

Statistical analysis
Statistical analysis was carried out using the Minitab 18 software (http://www.minitab.com)and the Social Science Statistics website (http://www.socscistatistics.com/tests/ztest/Default2.aspx).A chi-square test was performed to compare the prevalence of antimicrobial resistance and virulence factor genes in all isolates.The calculations were carried out at 95% confidence interval and a p < 0.05 considered statistically significant.
Thirty (57.7%) from 1977 to 1985 and 36 (67.9%) from 2015 clinical isolates had fingerprints that were not seen in the rest of the isolates (Fig. 1).Twenty-two clinical isolates had common ERIC-PCR patterns.
Extensive heterogeneity, in comparison with clinical isolates, was observed in P. aeruginosa isolated from fresh water sources; 97 (85.1%) had distinct ERIC-PCR patterns.However, 10 isolates from different geographical areas exhibited 100% identical patterns.

Notes.
Significant difference in the prevalence of virulence factor genes: a, p < 0.05; b, p < 0.01; c, p < 0.001.However, those from tissue and urine were resistant to all of the antimicrobials, but blood samples were sensitive to imipenem and amikacin.Most of the clinical and environmental isolates were sensitive to almost all of the antipseudomonal drugs (above red line), except for ticarcillin/clavulanic acid (Fig. 3).Increased resistant isolates (below red line) was seen in 2015 compared to that of the archived isolates.Four MDR strains (J3, J11, J20 and J25) were detected, all were from 2015 clinical sources (Supplemental Information 1).
Environmental isolates from fresh water exhibited consistent susceptibility to almost all of the antimicrobials.However, some were non-susceptible to ciprofloxacin, piperacillin, carbapenems (doripenem, meropenem and imipenem) and piperacillin/tazobactam; almost all (n = 110) were non-susceptible to ticarcillin/clavulanic acid.

Virulence factor genes
Both clinical and environmental isolates of P. aeruginosa showed high (>60%) prevalence of virulence factor genes, except for exoT, exoU, pvdA and pilB (Table 1).None of the clinical isolates from both isolation periods harbored the exoT gene (p < 0.001).No virulence factor genes were present in ten (9.5%) clinical and eighteen (15.8%) environmental isolates (Supplemental Information 1).
Resistance to carbapenems (meropenem, doripenem and imipenem), β-lactamase inhibitors (ticarcillin/clavulanic acid and piperacillin/tazobactam), piperacillin, ceftazidime and ciprofloxacin was observed in some of the virulence-negative isolates (Fig. 5).Five isolates without virulence factor genes were susceptible to all of the antimicrobials.Only one MDR strain (J3) was absent of any of the virulence factor genes (Supplemental Information 1).

DISCUSSION
We found high genetic heterogenicity in both categories of isolates as only four clinical strains (J43, x117, PA37 and PA40) displayed 100% clonality to seven environmental isolates, reflecting previous studies where P. aeruginosa clinical isolates harboured unique genotypes with low genetic similarity to environmental isolates (Martins et al., 2014;Tumeo et al., 2008).The difference in genetic makeup is probably due to various P. aeruginosa biotypes existing in nature, and only those with high adaptability can survive in wide ranging habitats.
Three of the four clonal clinical strains, i.e., PA37, PA40 and x117 were recovered more than 30 years ago indicating that some extraordinary genotypes can persist in the environment for many years and become transmissible.For example, PA14 of sequence type (ST)-253 isolated in the USA 15 years ago, became globally distributed and was found in Queensland, Australia (Kidd et al., 2012).
Aquatic habitats could be a source of pseudomonal infections in humans, as our findings are consistent with reports of environmental isolates exhibited similar genotypes to clinical isolates (Kidd et al., 2012;Pellett, Bigley & Grimes, 1983;Romling et al., 1994).
P. aeruginosa is a common cause of healthcare-associated infections such as bloodstream infections, urinary tract infections, surgical site infections and pneumonias especially in CF patients (CDC, 2013).Current antipseudomonal drugs were introduced in 1960s and since then, only few new drugs have been approved for clinical use (Bassetti et al., 2013;Monnet & Giesecke, 2014).Our clinical isolates recovered from 1977 to 1985 exhibited relatively low resistance to amikacin and imipenem but were totally sensitive to ceftazidime and ciprofloxacin probably due to absence or low usage in treatment.Later isolates from the same hospital in 2005 exhibited higher resistance rates than our archived isolates: piperacillin/tazobactam (9.4%), imipenem (9.9%), amikacin (6.73%), gentamicin (12.9%), netilmicin (10.1%), ciprofloxacin (11.3%) and ceftazidime (10.9%) (Raja & Singh, 2007).High resistance rates to these drugs was also documented from Malaysia (Pathmanathan, Samat & Mohamed, 2009).In general, over a period of 30 years (1977 to 2009) there has been a rise in resistance to the core antipseudomonal drugs in Malaysian isolates probably due to selection pressure.The increased resistance to ceftazidime (10.9%) and ciprofloxacin (11.3%) pose a public health challenge.The National Surveillance of Antibiotic Resistance (NSAR) by the Ministry of Health Malaysia (2015) reported that the resistance patterns of P. aeruginosa clinical isolates were considerably stable from 2013 to 2015, with a slight decrease in resistance to most of the antipseudomonal drugs but a slight increase in piperacillin/tazobactam resistance (from 4.6% to 5.6%) (Ministry of Health Malaysia, 2015); this was similar to our clinical isolates recovered in 2015, probably due to effective surveillance program.Some of our aquatic isolates were resistant to ciprofloxacin, piperacillin/tazobactam, piperacillin and carbapenems (imipenem; meropenem; doripenem) which is unusual.A recent study reported 100% antimicrobial susceptibility in P. aeruginosa isolated from water samples; however, resistance to meropenem (30.4%), piperacillin/tazobactam (10.6%) and ceftazidime (4.2%) was observed in other Pseudomonas spp.isolated from the same sampling points (Kittinger et al., 2016).Another recent study on aquatic isolates of P. aeruginosa showed resistance to imipenem (9.43%), ticarcillin/clavulanic acid (1.88%) and co-resistance to piperacillin and ticarcillin/clavulanic acid (1.88%) (Schiavano et al., 2017).
Antibiotic biosynthesis and resistance is believed to be ancient and occurred naturally even before the introduction of antibiotics (Barlow & Hall, 2002;D'Costa et al., 2011).Bacteria isolated from the ancient Lechuguilla Cave of four million years showed most to be multidrug resistant to natural antibiotics.Physiological changes such as the production of antimicrobials occur in these bacteria under nutrient-limited cave environment and bacteria develop resistance as a defence mechanism (Bhullar et al., 2012).The plasmidmediated quinolone resistance determinant (Qnr) occurs naturally in aquatic reservoirs, and probably enables gene transfer between different waterborne bacteria in habitats where quinolones are not present (Poirel et al., 2005).P. aeruginosa possesses inherent resistance to many classes of drugs attributed to the chromosomal-encoded AmpC β-lactamases and efflux pumps, and its lower membrane permeability (Masuda et al., 2000;Poole & Srikumar, 2001).We believe that our resistant environmental isolates had probably acquired resistance in order to survive and persist in diverse natural habitats.
Soil samples from the Netherlands, spanning pre-and post-antibiotic eras (1940 to 2008) had shown increased antibiotic resistance genes in the recent soil samples, with some being more than 15 times more abundant than those in the 1970s (Knapp et al., 2010).The utilization of non-degradable synthetic antibiotics (e.g., quinolones) in aquaculture, extensive use of antibiotics in livestock, broken sewage pipes, hospital effluents and runoff from farms fertilized with livestock faeces, may contribute to the selection of resistant bacteria in natural habitats such as surface waters, ground water, drinking water or sediments (Bartlett, Gilbert & Spellberg, 2013;Goni-Urriza et al., 2000;Kummerer, 2004).
Increased concentration of antibiotics in the environments due to extensive use in clinical and agricultural settings affects the evolution of bacterial resistance and virulence.The interplay between resistance and virulence is postulated to follow a Darwinian model, in which more resistant and virulent isolates will be selected in the population (Beceiro, Tomas & Bou, 2013).In most cases, increased resistance is associated with decreased virulence and fitness (Geisinger & Isberg, 2017); however, no obvious correlation between antimicrobial resistance and virulence was observed in our P. aeruginosa isolates.
Four effectors ExoS, ExoT, ExoU and ExoY are present in the T3SS system and the secretion of ExoS and ExoT in combination reduce anti-internalization by phagocytic cells (Shaver & Hauser, 2004).The absence of the exoT gene in our clinical isolates is similar to a report (Finnan et al., 2004) and it is possible that clinical isolates may delete a less virulent exoT gene to prevent the antagonizing effect of multiple effectors.
A small number (30 clinical and nine environmental) of our P. aeruginosa harboured the exoU gene, probably acquired via horizontal transfer (Berthelot et al., 2003) to become highly virulent and cytotoxic (Schulert et al., 2003;Wong-Beringer et al., 2008).This acquisition of exoU probably occurs only under selective pressure resulting in low prevalence in nature.The ExoU-positive environmental isolates were mostly from recreational parks situated in densely populated areas.
The co-existence of exoS and exoU is probably mutually exclusive in P. aeruginosa due to their distinct loci in the genome (Bradbury et al., 2010).Only eight of the total 219 isolates contained both genes probably providing a selective advantage for the survival of P. aeruginosa in a specific niche.Over time, a change in the genotype may take place by the deletion of one or the other gene to prevent antagonism.Therefore, the universal genotype of P. aeruginosa is either exoU or exoS.
Expression of T3SS by P. aeruginosa is associated with increasing virulence, but T3SSnegative isolates have been recovered from patients, which may have been contaminants or probably had remained dormant to evade host immune system for long-term survival (Jain et al., 2004).The expression of virulence genes involves multiple regulatory and metabolic networks (Winstanley, O'Brien & Brockhurst, 2016).A full set of T3SS effectors was only detected in our environmental isolates probably providing selective advantage to P. aeruginosa under harsh natural environments.
Phenazine-modifying enzymes phzM, phzS and phzH in P. aeruginosa (Recinos et al., 2012) are toxic and pH-dependent (Cezairliyan et al., 2013), and we observed that many harboured all the 3 phzH, phzM and phzS genes.It is likely that positive isolates produce more than one type of phenazine toxin that act over a wide pH range to ensure bacterial survival and colonization under different environmental conditions (Bradbury et al., 2010;Finnan et al., 2004).
Uptake of iron is crucial for colonization and P. aeruginosa is able to acquire Fe 3+ from the host by producing iron chelating siderophore pyoverdine; the responsible gene (pvdA) was present in 48% to 70% of our clinical isolates, but may lose this ability during long periods of colonization (De Vos et al., 2001).

CONCLUSION
P. aeruginosa is ubiquitous and an opportunistic pathogen causing infections especially in immunocompromised patients.It is equipped with natural drug resistance and virulence mechanisms for survival in harsh environments.However, it can become resistant under selective pressure leading to increase in pseudomonal infections and possibly therapeutic failures.
Our findings indicate a rise in resistance to antipseudomonal drugs in two hospitals in Malaysia over the past 30 years.Therefore, it is necessary to implement a programme of periodic surveillance and standardization of a protocol for antipseudomonal therapy by the relevant authorities.In addition, the observation of antimicrobial resistance in environmental isolates from densely populated areas highlights the importance of increased public health awareness.
The limitation of this study was the small number of isolates, but our findings provide basic knowledge of epidemiology, antimicrobial resistance and virulence traits of P. aeruginosa.Works involved other typing methods such as multi-locus sequencing typing (MLST) or multi-virulent sequencing typing (MLVA) (Teh, Chua & Thong, 2011) could also be carried out to gather more differential information between clinical and environmental isolates of Pseudomonas aeruginosa.A robust surveillance of antimicrobial susceptibility should be implemented to monitor and prevent dissemination of pathogenic multidrug resistant strains in Malaysia.

Figure 1
Figure 1 Venn diagram of ERIC-PCR patterns.Orange/Green/Blue sections and intersections represent the number of isolates with distinct and similar ERIC-PCR fingerprints, respectively.Subsets (grey) indicates the number of isolates with 100% clonality in the respective category.Full-size DOI: 10.7717/peerj.6217/fig-1

Figure 2 Figure 3 LiewFigure 4
Figure 2 An overview of antimicrobial resistance of P. aeruginosa in clinical specimens from two isolation periods: archive (1977 to 1985) and 2015.No antimicrobial resistance was found in the following specimen categories: sputum, cerebrospinal fluid (CSF), slough, peritoneal fluid and discharge/drainage (ear, eye).NA indicates source of specimen is not available.Full-size DOI: 10.7717/peerj.6217/fig-2

Table 1 Distribution of virulence factor genes in clinical (1977 to 1985 and 2015) and environmental isolates of P. aeruginosa.
IsolatesNo. of isolates with virulence genes (%)