Quorum sensing in Pseudomonas aeruginosa isolates from cystic fibrosis patients

Abstract

The ESKAPE pathogen P. aeruginosa causes a variety of severe infections in hospitalised and immunocompromised patients and chronic pulmonary infections which are the main cause of death in people affected by Cystic Fibrosis (CF), a genetic disease affecting about 1/2500 newborns in the Caucasian population (Welsh et al., 2001; Talbot et al., 2006; Driscoll et al., 2007; Döring et al., 2010; Stefani et al., 2017). P. aeruginosa infections are hard to eradicate, since this microorganism is intrinsically resistant to many antibacterials and it is particularly prone to acquire new resistances in the hospital environment by horizontal gene transfer (Latifi et al., 1995; Talbot et al., 2006; Driscoll et al., 2007). The production of virulence factors and biofilm formation in P. aeruginosa are strictly dependent on a global regulatory circuit known as quorum sensing (QS), a communication system that controls gene expression in response to cell density (Smith and Iglewski, 2003; Lee et al., 2006; Williams and Càmara 2009; Papenfort and Bassler, 2016). P. aeruginosa has four QS systems that are interconnected and hierarchically arranged: in rich medium, the las QS system is at the top of this hierarchy because it is required for full activation of the other QS systems, among which the pqs QS system (Williams and Càmara, 2009; Papenfort and Bassler, 2016). The las system is based on the QS signal receptor LasR (coded by the lasR gene), and on the QS signal synthase LasI (coded by the lasI gene), producing the QS signal molecule N- (3-oxododecanoyl)homoserine lactone (3OC12-HSL) (Schuster and Greenberg, 2006). The pqs QS system is based on signal molecules belongings to the chemical class of alkyl quinolones (AQs). The enzymes encoded by pqsABCDE operon are involved in the production of the signal molecule 2-heptyl-4-quinolone (HHQ), which is converted in 2-heptyl-3-hydroxy-4-quinolone (PQS) by the monooxygenase PqsH, encoded by the pqsH gene. PQS or HHQ binding activates the QS signal receptor PqsR. Another relevant AQ produced by P. aeruginosa at µM level is 4-hydroxy-2- heptylquinoline-N-oxide (HQNO), synthesized by the PqsL monooxigenase (coded by pqsL gene). HQNO does not activate the PqsR receptor and cannot be considered a QS signal molecule (Deziel et al., 2004; Rampioni et al., 2016). Interestingly, HQNO is a potent inhibitor of the cytochrome bc1 complex of Gram-positive bacteria, hence it contributes to competitiveness of P. aeruginosa in the environment (Heeb et al., 2011). Both LasR and PqsR activated by the respective signal molecules positively control the expression of hundred of genes, including virulence genes and genes involved in biofilm formation (Williams and Càmara, 2009; Papenfort and Bassler, 2016). Accordingly, P. aeruginosa QS-deficient strains are less virulent than the wild type counterparts in animal and plant infection models, and QS is considered a promising target for the development of new anti-virulence drugs (Rasko and Sperandio, 2010; Lasarre and Federle, 2013; Rampioni et al., 2014). The lung of CF patients is an heterogeneous and stressful environment for invading bacteria: the immune response causes oxidative and nitrosative stresses, while the viscous mucus typically produced by these patients hampers bacterial clearance and causes osmotic stress. Further stressful factors are the sub-lethal concentrations of antibiotics and the competition with other microorganisms. CF patients experience in pediatric age a sequence of P. aeruginosa intermittent lung infections, followed by the establishment of a chronic infection which cannot be eradicated with antibiotics. The P. aeruginosa chronic infection can last for decades in the CF patient, causing severe lung inflammation which is the ultimate cause of decline of lung function and death (Folkesson et al., 2012; Winstanley et al., 2016). During the long permanence in the CF lung, P. aeruginosa undergoes evolutionary changes and different sub-lineages coexist and interact within the lung. Overall, it is believed that clones particularly adapted to the CF lung emerge along years (Folkesson et al., 2012; Winstanley et al., 2016). Mutations conferring antibiotic resistance, lack of specific virulence factors production, overproduction of alginate and other factors associated to the biofilm matrix are frequently studied among P. aeruginosa CF isolates and can be correlated with the adaptation to the CF lung environment (Folkesson et al., 2012; Winstanley et al., 2016). The emergence of mutants expressing a lytic phenotype, postulated to be related to a mutation in the pqsL gene (D’Argenio et al., 2002), and the emergence of lasR-defective mutants (Smith et al., 2006; Hoffman et al., 2008; Feltner et al., 2016) in P. aeruginosa populations chronically infecting the CF lung has also being described. On these basis, the general aim of this thesis has been to increase our understanding about the las and pqs QS systems in P. aeruginosa, with special focus on the CF infection. The research described in chapter 3 has been aimed at shedding new light on the role played by PqsL and HQNO synthesis in P. aeruginosa physiology. It has been shown that the abolition of HQNO synthesis in P. aeruginosa PAO1 pqsL causes accumulation of HHQ. This induces the activation of the lytic cycle in the Pf4 prophage, with consequent formation of lytic plaques in colony biofilms. Moreover, 101 strains isolated from CF patients have been screened for the lytic phenotype. Overall, 26 out of 34 (76%) CF strains expressing the lytic phenotype carry a deleted or mutated pqsL gene, and the wild type (non-lytic) phenotype is restored in 6 out of the 8 CF strains in which a functional pqsL could be expressed in trans. To the best of our knowledge, this is the first evidence of a clear genetic correlation between the expression of the lytic phenotype and the emergence of the pqsL mutation in clinical P. aeruginosa strains. Overall, our results suggest that the production of HQNO confers an adaptive advantage to P. aeruginosa wild type not only by inhibiting the growth of competitors in polymicrobial communities (Heeb et al., 2011) but also by limiting HHQ accumulation and consequent cellular lysis due to prophage(s) activation. It should be highlighted that the first lung infections experienced by CF patients are usually caused by Staphylococcus aureus, which is subsequently overpowered by P. aeruginosa (Filkins et al., 2015; Nguyen et al., 2016). Hence it could be reasoned that when P. aeruginosa prevails upon earlier S. aureus in the CF lung infection and establishes a single-species infection, the selective pressure for the maintenance of HQNO synthesis could decrease and the emergence of a subpopulation of pqsL mutants could became advantageous in this challenging environmental niche. In this view, the increased antibiotic resistance of the P. aeruginosa PAO1 pqsL biofilm with respect to the wild type shown in this study, together with other studies suggesting the existence of a link between prophage activation, bacterial autolysis and biofilm resistance to antibiotics (Webb et al., 2004; Allesen-Holm et al., 2006; Rice et al., 2009; Chiang et al., 2013), could at least in part explain the emergence of pqsL mutants during the infection in antibiotic-treated CF patients. The research described in chapter 4 has been specifically aimed at investigating the suitability of anti-virulence drugs targeting the las QS system in CF therapy. Anti-virulence compounds are promising as alternative or adjuvants of conventional antibiotics, and the las QS system is considered a good target for the development of anti-virulence drugs against P. aeruginosa (Rasko and Sperandio, 2010; Rampioni et al., 2014). However, the frequent isolation of P. aeruginosa lasR mutants in CF patients has opened a debate about the suitability of drugs targeting the las QS system in CF therapy (Ciofu et al., 2015; Winstanley et al., 2016). Here, a collection of 100 P. aeruginosa isolates from CF patients infected for different years has been tested. In particular these strains have been characterized for: i) antibiotic susceptibility; ii) QS signal molecules production; iii) susceptibility to niclosamide (NCL; Imperi et al., 2013), a strong QS inhibitor targeting the las QS system in the model strain PA14. Our analysis showed that the majority of P. aeruginosa strains isolated from CF patients for the first time or with chronic infection established up to 7 years are able to synthetize 3OC12-HSL, hence they should be considered susceptible to drugs targeting the las QS system. However, our results show that the production of 3OC12-HSL and of 3OC12-HSL dependent virulence factors is overall scarcely affected in these CF isolates treated with the model QS inhibitor NCL. This result is in agreement with the results obtained with the QS inhibitor furanone C-30 (Garcìa-Contreras et al., 2015) and indicates that future anti-virulence drugs targeting the las QS system could be successfully developed and used in CF therapy only if shown to be active against a large proportion of CF isolates. Understanding the mechanisms underlying the lack of susceptibility of CF isolates to anti-QS compounds could drive further medicinal chemistry studies aimed at potentiating these drugs and at expanding their range of activity. It should also be highlighted that even if a QS inhibitor would reach the clinical use, the high variability of CF strains will require preliminary tests to ascertain the susceptibility of the specific P. aeruginosa strain(s) colonizing a patient toward the QS inhibitor, a practice routinely used before administering antibiotics. To this aim, methods for rapid and convenient detection of QS signal molecules have already been developed (Rampioni et al., 2018, Supplementary) and could be further improved and translated to diagnostic laboratories. Overall the research carried out in this thesis has clarified some aspects of the pqs QS system, which could be relevant to understand and combat the lethal P. aeruginosa infections in CF patients, and has also highlighted the limitations that should be overcome for successfully translating anti-QS drugs to the CF therapies.