Faculty of Science and Bio-engineering Sciences, Department of Bio-engineering Sciences, Research Group of Microbiology, Academic Year 2015-2016

Adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung and the role of shear stress in disrupting alginate matrix biofilms

Thesis submitted in fulfillment of the requirements for the degree of Doctor (Ph.D.) in Bio-engineering Sciences

Jozef Dingemans

Promotors: Em. Prof. Dr. Pierre Cornelis

Prof. Dr. Daniel Charlier

Co-promotors: Prof. Dr. Anne Malfroot (UZ Brussel)

Dr. Rob Van Houdt (SCK•CEN)

Members of the jury

Promotors:

Em. Prof. Dr. Pierre Cornelis (VUB)

Prof. Dr. Daniel Charlier (VUB)

Co-promotors:

Prof. Dr. Anne Malfroot (VUB, UZ Brussel)

Dr. Rob Van Houdt (SCK-CEN)

President:

Prof. Dr. Geert Angenon (VUB)

Secretary:

Prof. Dr. Wim Versees (VUB)

Internal jury members:

Prof. Dr. Guido Verniest (VUB)

Prof. Dr. Jo Van Ginderachter (VUB)

Prof. Dr. Kim Roelants (VUB)

External jury members:

Prof. Dr. Craig Winstanley (University of Liverpool)

Prof. Dr. Sylvie Chevalier (Université de Rouen)

Prof. Dr. René De Mot (KU Leuven)

The work described in this dissertation was performed at:

-Research Group of Microbiology, Department of Bio-engineering Sciences,

Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussel

-Unit of Microbiology, Institute for Health, Environment and Safety,

Studiecentrum voor kernenergie (SCK•CEN), Boeretang 200, 2400 Mol

This research was financially supported by:

- The government agency for Innovation by Science and Technology (IWT Vlaanderen).

- The Belgian Association against Cystic Fibrosis (Belgische Vereniging voor Strijd tegen Mucoviscidose).

- The Flanders Interuniversity Institute for Biotechnology (VIB).

Published by the Research Group Microbiology,

Department of Bio-engineering Sciences,

Vrije Universiteit Brussel,

Pleinlaan 2,

1050 Brussel,

Belgium

Copyright © 2015 by Jozef Dingemans

Apart from any fair dealing for the purposes of research or private study or criticism or review, this publication may not be reproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, scanning, or otherwise) without prior written permission of the publisher.

Acknowledgements

About five years ago I asked Pierre if it was possible to start a PhD in his lab. Although we have already met regularly in the scope of the microbiology courses that he taught, always in a warm atmosphere, I could never guess that my respect for him would reach these heights. Pierre, you were an excellent mentor for me in the scientific world, but also a guide in life. I could try to summarize the many conferences, dinners, movie nights, etc. that we have enjoyed together with the MICR group, but the printing costs of this thesis would accumulate to excessive amounts. Nevertheless, I want to thank you for all the freedom that you gave me in performing my research, the positive atmosphere that was omnipresent in the MICR group, and the many moments of friendship that helped both of us when skies were grey. I am convinced that the way you performed research, allowing serendipity to occur, is the right way for the VUB to continue in a world that often lacks the tolerance to allow freedom of tought and speech.

One of the things that touched me most during this PhD was listening to cystic fibrosis patients and their parents during a couple of meetings (one in Paris, one in Belgium). I was astonished by the optimism that these people exhibited when telling their story. Every single day I devoted to this PhD, I felt I could contribute to the change in life quality of these people, encouraging me to bring this project to fruition. Due to the medical nature of this project, I relied on the knowledge and expertise of a great team of physiotherapists (Hanneke Eyns, Nino Galasso and Sylvie Vanlaethem) and physicians (Prof. Anne Malfroot and Dr. Julie Willekens) at UZ Brussel. In particular I want to thank my co-promotor Prof. Anne Malfroot for giving me the opportunity to attend the ECFS meeting in Lisbon, but most importantly for the many fruitful multidisciplinary meetings and collaborations. I want to conclude by thanking all CF patients that have participated in our clinical study and who provided sputum samples during the optimalisation period of this project.

Next, I would like to thank all the collaborators that have contributed to this research. In all the labs I have worked during my PhD (MICR, SCK in Mol, and LabMCT in Neder-over- Heembeek), people were kind and helpful. Special thanks to Florence Bilocq for introducing me in the world of Diversilab, and to Dr. Daniel De Vos and Dr. Jean-Paul Pirnay for using the facilities at LabMCT and for the help in publishing my first 1st author paper. Also many thanks to the people of SCK. I really appreciated the help of Ilse, Wietse, Joachim, Bo, Hugo, Katinka (scanning electron microscopy), Pieter (RNAseq analysis), and especially my co- promotor Dr. Rob Van Houdt. Without your alertness, I would have missed several PhD days and I would have to ask every time I went to SCK for a new security badge! I also want to thank you for the excellent suggestions/corrections when proofreading my papers.

As I have already told, the atmosphere at MICR has always been exquisite, even though we had to share the office with more than 15 people (during Prof. Jeroen Raes’ era at the VUB). I will never forget the Friday afternoon drinks in Opinio with Falk, Raul, Sarah, Youssef, and Sam! But also the nice company of the people from the lab of Prof. De Vuyst (Luc, Henning, Frédéric, Andrea, David, Marko …) The interaction between these groups has resulted in a memorable visit to the Christmas market in Aachen! This reminds me of the great opportunities the VUB offers due to its small scale to meet other groups and to collaborate.

Pierre’s group has always been small, creating a cosy atmosphere in the lab that stimulated us to collaborate and perform sports after work (usually followed by a few beers in “het complex”). The strong cohesion in our group (Ken, Ye, Ameer, Wei, and myself) has always given me a warm feeling and motivated me to successfully finish this PhD. One of the key factors that created the group atmosphere was without any doubt the presence of our amazing secretary Linda! A day in the lab without Linda felt like a day without sunshine! Also after all these lovely people left the cosy VUB environment (I am often called “the last of the Mohicans”), MICR maintained a nice working environment. In particular due to the enthusiasm of Liesbeth, Charlotte, Indra, Han, and especially Karl. I think it will be difficult to find a running partner and Gin Connoisseur as enthusiastic as you, Karl! Furthermore, I want to thank Prof. Eveline Peeters for adopting me in her lab after the retirement of Pierre! Besides the MICR people, I want to express my appreciation to the people of GEVI (Prof. Jean-Pierre Hernalsteens, Sien, and Charlotte), Prof. Henri Degreve, Prof. Daniel Charlier (also for being my promotor after Pierre’s retirement), Kris, Sonya and many others. I want to conclude my gratitude to MICR by thanking my excellent thesis students Francesca, Michael, and Laila. I really enjoyed the scientific discussions that we had and the effort you put in our work!

Further, I want to acknowledge all collaborators that are co-authors, including Prof. René De Mot, Dr. Maarten Ghequire (KU Leuven), Prof. Dianne Newman, and Dr. Ryan Hunter (USA), and in particular Dr. Aurélie Crabbé (Universiteit Gent)! During your stay in Arizona it was always a pleasure to Skype and brainstorm with you, advising me in my research. Great

thanks to Prof. Laurence Van Melderen (ULB) and her lab too, it was always a pleasure meeting you during (and after) a day at a conference.

I am also grateful to the jury members for accepting the invitation to serve on the Examination Board and for the interesting discussions we have had during the private defense.

Finally, I want to thank my dear friends and family who always supported me during the more difficult moments of my PhD! Thanks to Hugo and Jens I could focus my thoughts on football when watching our beloved club RSC Anderlecht. When Wim, Loic, and Max joined, this regularly resulted in a night dancing in the Havana club. In addition, the nice excursions with Niels and Evelyne to Moroccan, Greek, and Portuguese restaurants and once in a while eating “fruits de mer” (Lobster time!) with Kurt helped me to reduce stress levels. Nevertheless, I want to express my gratitude to my parents who have always been there for me and supported me in every choice I have made so far in my life. Finally, I want to thank my lovely spouse Florien for being there for me every single moment of the day. I do realize that this PhD also put some pressure on her, but our love has passed this test with great distinction!

Abstract

1 Abstract

Pseudomonas aeruginosa is a Gram-negative gamma-proteobacterium that can be found in a wide range of environments including water, soil, animals, and humans. However, due to its metabolic versatility and large genome encoding multiple virulence factors, this opportunistic pathogen is able to infect patients with severe wounds, immunocompromised individuals, and most importantly cystic fibrosis (CF) patients. Chronic P. aeruginosa infections are the major cause of morbidity in CF patients due to persistent lung inflammation and the resulting irreversible lung damage. Often the initial infection is mediated by the acquisition of environmental P. aeruginosa strains, although inter-patient transmissions have also been reported. Subsequently, the initially colonizing P. aeruginosa strains adapt to the CF lung environment by undergoing extensive adaptation and switch from a planktonic to a biofilm lifestyle.

In the first part of this thesis, we have typed 54 P. aeruginosa CF isolates using a novel genotyping approach based on the combination of Rep-PCR and multiplex PCR targeting ferripyoverdine siderophore receptor and pyocin genes. Interestingly, we found that a number of strains have lost the fpvB receptor gene during adaptation of P. aeruginosa to the CF lungs. One of these strains, CF_PA39, was found to be present in several CF patients attending different CF reference centers in Belgium and hence was considered to be transmissible. Analysis of the whole genome sequence of this strain revealed the accumulation of several large deletions during colonization of the CF lung environment, including deletion of the entire type III secretion system, several virulence genes and TonB-dependent siderophore receptor genes. When analyzing a database of whole-genome sequenced P. aeruginosa strains belonging to the DK2 clone from patients attending the Copenhagen CF clinic, we found that the deletion of TonB-dependent receptor genes is part of the genome reduction process that occurs during adaptation of P. aeruginosa to the cystic fibrosis lungs. Finally, via qRT-PCR we have shown (in collaboration with the group of Dianne Newman) that multiple iron-uptake systems are being used by P. aeruginosa in the CF lung environment and that not only ferric, but also ferrous iron uptake is important in this condition.

In the second part of the thesis we have studied the effect of shear stress on the transmissible P. aeruginosa isolate CF_PA39 grown in artificial sputum medium (ASM) at the

i Abstract transcriptomic (via RNA sequencing) and phenotypic level. Under low fluid shear conditions, genes involved in stress response, alginate biosynthesis, denitrification, glycine betaine biosynthesis, glycerol metabolism and cell shape maintenance were up-regulated, while genes involved in phenazine biosynthesis, type VI secretion, and multidrug efflux, were down- regulated. In addition, a number of small RNAs appeared to be involved in the response to shear stress. Interestingly, the subtle effect of shear stress at the transcriptomic level was drastically pronounced at the phenotypic level since the formation of robust biofilms under low fluid shear conditions was not observed under high fluid shear conditions. Finally, quorum sensing signaling was found to be slightly affected in response to shear stress, resulting in higher production of auto-inducer molecules during growth under high fluid shear conditions. In summary, this part of our study has revealed a way to modulate the behaviour of a highly adapted P. aeruginosa CF strain by means of introducing shear stress, driving it from a biofilm into a more planktonic lifestyle.

In the third part of this thesis, we have determined the expression of marker genes associated with either the planktonic or biofilm lifestyle of P. aeruginosa, discovered in the course of this work, in sputum samples from CF patients receiving intrapulmonary percussive ventilation (IPV) at low or high frequency versus standard therapy. We found that IPV at high frequency was able to enhance pulmonary function and induce the expression of planktonic marker genes in a number of CF patients, indicating that this might be a promising treatment with regard to the disruption of biofilms in CF patients.

Finally, using the genome sequence of CF_PA39, we have identified an open reading frame (ORF) encoding a novel S-type pyocin. This novel pyocin sequence perfectly matched to the nucleotide sequences of both the receptor binding and translocation domains of pyocin S1. However, the nucleotide sequence corresponding to the killing domain of this novel pyocin sequence was completely different compared to that of pyocin S1. Using real-time PCR, we showed that the expression of the pyocin gene is induced under iron-limiting conditions. Finally, we have proven the functionality of this pyocin as it was able to inhibit the growth of a number of P. aeruginosa CF isolates.

ii Samenvatting

2 Samenvatting

Pseudomonas aeruginosa is een Gram-negatieve gamma-proteobacterie die kan worden teruggevonden in een brede waaier aan omgevingen, waaronder water en bodem, maar ook bij dieren en mensen. Door zijn metabolische veelzijdigheid en het grote genoom dat codeert voor meerdere virulentiefactoren, is deze opportunistische pathogeen in staat om patiënten met ernstige wonden, individuen met een beperkt immuunsysteem en voornamelijk mucoviscidosepatiënten, te infecteren. Chronische infecties met P. aeruginosa vormen de hoofddoodsoorzaak bij mucoviscidosepatiënten door een aanhoudende longontsteking en de daarmee gepaard gaande onomkeerbare longschade. Vaak ontstaat de initiële infectie door de opname van P. aeruginosa stammen uit de omgeving, maar ook besmettingen tussen patiënten werden reeds beschreven. In een volgende fase past de P. aeruginosa stam die de patiënt oorspronkelijk koloniseerde zich aan de mucoviscidoselongomgeving aan door een uitvoerige adaptatie te ondergaan en over te schakelen van een planktonische naar een biofilmlevensstijl.

In het eerste deel van deze thesis hebben we 54 P. aeruginosa mucoviscidose-isolaten getypeerd, gebruik makend van een nieuwe genotyperingsmethode, gebaseerd op de combinatie van Rep-PCR en multiplex PCR om ferripyoverdinesiderofoorreceptor- en pyocinegenen te amplificeren. We vonden dat een aantal stammen het fpvB receptorgen verloren tijdens de adaptatie van P. aeruginosa aan de mucoviscidoselong. Eén van deze stammen, CF_PA39, werd teruggevonden in meerdere mucoviscidosepatiënten die verschillende mucoviscidosereferentiecentra bezochten en werd aldus als overdraagbaar beschouwd. Analyse van de volledige genoomsequentie van deze stam onthulde de accumulatie van verscheidene grote deleties tijdens de kolonisatie van de mucoviscidoselongomgeving, met inbegrip van de deletie van het volledige type-III- secretiesysteem, verschillende virulentiegenen en TonB-afhankelijke siderofoorreceptorgenen. Tijdens de analyse van de databank van volledige genoomgesequeneerde P. aeruginosa stammen die behoren tot de DK2 kloon van patiënten uit het mucoviscidosereferentiecentrum in Kopenhagen, werd gevonden dat de deletie van TonB-afhankelijke receptorgenen deel uitmaakt van het genoomreductieproces dat optreedt tijdens adaptatie van P. aeruginosa aan de mucoviscidoselong. Uiteindelijk toonden we via real-time PCR aan (in samenwerking met de groep van Dianne Newman) dat meerdere

iii Samenvatting ijzeropnamesystemen door P. aeruginosa worden gebruikt in de mucoviscidoselongomgeving en dat niet enkel de opname van geoxideerd, maar ook van gereduceerd ijzer, belangrijk is in deze conditie.

In het tweede deel van de thesis hebben we het effect van schuifspanning op het volledige genoomgesequeneerde, overdraagbare P. aeruginosa isolaat, CF_PA39, opgegroeid in artificieel sputum medium (ASM), op het transcriptomisch (via RNA-sequenering) en fenotypisch niveau bestudeerd. Onder lage schuifspanning werden genen betrokken in stressrespons, alginaatbiosynthese, denitrificatie, glycine betaine-biosynthese, glycerolmetabolisme en onderhoud van de celvorm opgereguleerd, terwijl genen betrokken in phenazinebiosynthese, type-VI-secretie en multidrugefflux werden down-gereguleerd. Daarenboven bleek dat een aantal kleine RNAs betrokken waren in de respons op schuifspanning. Het subtiele effect van schuifspanning op het transcriptomisch niveau was sterk uitgesproken op het fenotypisch niveau, aangezien de vorming van robuuste biofilms onder lage schuifspanning niet werd teruggevonden onder hoge schuifspanning. Tenslotte bleek dat quorum sensingsignalisatie lichtjes beïnvloed werd in respons op schuifspanning, resulterend in een hogere productie van auto-inducer moleculen tijdens groei onder hoge schuifspanning. Samengevat, heeft dit deel van onze studie een weg aangetoond die kan gevolgd worden om het gedrag van een sterk geadapteerde P. aeruginosa stam te moduleren, waarbij deze stam wordt gedreven van een biofilmlevensstijl naar een meer planktonische levensstijl.

In het derde deel van deze thesis hebben we de expressie van merkergenen, geassocieerd met ofwel de planktonische, ofwel de biofilmlevensstijl van P. aeruginosa, die in de loop van dit werk werden geïdentificeerd, bepaald in sputumstalen van mucoviscidosepatiënten die intrapulmonale percussieventilatie (IPV) ontvingen op lage of hoge frequentie t.o.v. standaardtherapie. We stelden vast dat IPV bij hoge frequentie in staat was om de longfunctie te bevorderen en de expressie van planktonische merkergenen te induceren in een aantal mucoviscidosepatiënten, hetgeen aangeeft dat dit een veelbelovende behandeling kan zijn om de biofilmvorming door P. aeruginosa in de mucoviscidoselong te verstoren.

Tenslotte, gebruik makend van de genoomsequentie van CF_PA39, hebben we een open leesraam geïdentificeerd dat codeert voor een nieuw S-type pyocine. Deze nieuwe pyocinesequentie kwam perfect overeen met de nucleotidesequentie van zowel de receptorbindings- als translocatiedomeinen van pyocine S1. De nucleotidesequentie

iv Samenvatting corresponderend met het afdodingsdomein van deze nieuwe pyocinesequentie was echter volledig verschillend in vergelijking met die van pyocine S1. Gebruik makend van RT-PCR, hebben we aangetoond dat de expressie van het pyocinegen geïnduceerd wordt onder ijzer- limiterende condities. Tenslotte hebben we de functionaliteit van dit pyocine aangetoond, aangezien het in staat was om de groei van een aantal P. aeruginosa mucoviscidose-isolaten te inhiberen.

v

vi List of abbreviations

3 List of abbreviations

1-HP 1-hydroxyphenazine ABC ATP-binding cassette AD autogenic drainage ADP adenosine diphosphate ADPRT ADP ribosyltransferase AHLs N-acylhomoserine-lactones AIDS acquired immune deficiency syndrome AL acyl coenzyme A ligase AQs 2-alkyl-4-quinolones ASA aspartate β-semialdehyde ASL airway surface liquid ASM artificial sputum medium asRNA(s) antisense RNA(s) ATP adenosine triphosphate BLAST Basic Local Alignment Search Tool BSA bovine serum albumin BSRD Bacterial Small regulatory RNA Database c-di-GMP cyclic diguanosine monophosphate CAA casamino acids cAMP cyclic adenosine monophosphate cDNA complementary DNA CF cystic fibrosis CFTR cystic fibrosis transmembrane conductance regulator CFU colony-forming unit Cif CFTR inhibitory factor CM cytoplasmic membrane CO carbon monoxide COG clusters of orthologous groups CRISPR Clustered regularly interspaced short palindromic repeats Dab 2,4-diaminobutyrate

vii List of abbreviations

DGC(s) diguanylate cyclase(s) Dnr dissimilative nitrate respiration regulator DTPA diethylene triamine pentaacetic acid E epimerization domain ECF extracytoplasmic function elF2 elongation factor 2 ER endoplasmic reticulum FA fatty acid FDA Food and Drug Administration FDR false discovery rate FEV1 forced expiratory volume in one second fOHOrn N5-formyl-N5-hydroxyornithine FVC forced vital capacity GI(s) genomic island(s) GM-CSF granulocyte-macrophage colony-stimulating factor GSH glutathione Has heme acquisition system HCN hydrogen cyanide − HCO3 bicarbonate ion HHQ 2-heptyl-4(1H) quinolone HS high fluid shear HSI Hcp Secretion Island HTH helix-turn-helix IFN interferon IgA Immunoglobulin A IgG Immunoglobulin G IM inner membrane IPTG Isopropyl β-D-1-thiogalactopyranoside IPV intrapulmonary percussive ventilation IQS Integrated Quorum Sensing Signal IRF3 IFN regulatory factor 3 IS insertion sequence KD killing domain

viii List of abbreviations

LB lysogeny broth LES Liverpool epidemic strain LPS lipopolysaccharide LRP low-density receptor-related protein LS low fluid shear MAPK Mitogen-activated protein kinase MDR multidrug-resistant MDR-PA multidrug-resistant Pseudomonas aeruginosa MFP membrane fusion protein MIC minimum inhibitory concentration MPCAB 5-methylphenazine-1-carboxylic acid betaine mRNA messenger RNA MRSA methicillin-resistant Staphylococcus aureus MSSA methicillin-sensitive Staphylococcus aureus MVB multivesicular body MyD88 Myeloid Differentiation factor 88

N2O nitrous oxide NETs neutrophil extracellular traps NF-κB nuclear factor kappaB NO nitric oxide − NO2 nitrite − NO3 nitrate NRPS non-ribosomal peptide synthetases NTM nontuberculous mycobacteria OHOrn N5-hydroxyornithine OM outer membrane OMF outer membrane factor ORF(s) open reading frame(s) P box Promoter box PAMPs pathogen-associated molecular patterns PBS Phosphate-buffered saline PC phosphatidylcholine PCA phenazine-1-carboxylic acid

ix List of abbreviations

PCN phenazine-1-carboxamide PCR polymerase chain reaction PDE(s) phosphodiesterase(s) PG peptidoglycan PHAST PHAge Search Tool Phu Pseudomonas heme uptake PLCs phospholipases C PMNs polymorphonuclear neutrophils PP periplasm PQS Pseudomonas Quinolone Signal PVD pyoverdine QC quality control qRT-PCR quantitative real-time PCR QS quorum sensing RAST Rapid Annotations using Subsystems Technology RBD receptor binding domain RBS ribosome binding site Rep-PCR Repetitive Sequence-Based PCR RFU relative fluorescence units RGP(s) region(s) of genome plasticity RND Resistance-Nodulation-Division ROS reactive oxygen species rRNA ribosomal RNA RT-PCR reverse-transcriptase PCR RWV rotating wall vessel SCN− thiocyanate SD Shine-Dalgarno (sequence) SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SEM scanning electron microscopy SNP(s) single nucleotide polymorphism(s) sRNA(s) small RNA(s) SSU ribosomal small subunit T3SS type III secretion system

x List of abbreviations

Tat Twin arginine translocation TD translocation domain Te thioesterase domain TLR(s) Toll-like receptor(s) TNF-α tumor necrosis factor alpha TRIF TIR-domain-containing adapter-inducing interferon-β tRNA transfer RNA UTR untranslated region WT wild-type

xi

xii Table of contents

Table of contents

1 ABSTRACT ...... I 2 SAMENVATTING ...... III 3 LIST OF ABBREVIATIONS ...... VII 1 CHAPTER 1: LITERATURE OVERVIEW ...... 1 1.1 CYSTIC FIBROSIS ...... 3 1.1.1 Genetic basis ...... 3 1.1.2 Prevalence of cystic fibrosis ...... 7 1.1.3 Clinical manifestations of cystic fibrosis ...... 9 1.1.4 Pathogens associated with cystic fibrosis ...... 10 1.1.5 Host-pathogen interactions in the cystic fibrosis lung ...... 12 1.1.6 Available in vitro and animal models to study cystic fibrosis infections ...... 15 1.1.6.1 Tissue cultures ...... 15 1.1.6.2 The rotating wall vessel (RWV) ...... 15 1.1.6.3 Animal models ...... 16 1.2 PSEUDOMONAS AERUGINOSA ...... 17 1.2.1 Omnipresence of P. aeruginosa ...... 17 1.2.2 Genomic features of P. aeruginosa ...... 18 1.2.3 P. aeruginosa virulence factors ...... 20 1.2.3.1 Regulation of virulence via quorum sensing and two-component regulatory systems 21 1.2.3.2 Elastase and staphylolysin ...... 26 1.2.3.3 Rhamnolipids ...... 27 1.2.3.4 Pseudomonas quinolone signal ...... 28 1.2.3.5 Hydrogen cyanide ...... 28 1.2.3.6 Exotoxin A ...... 29 1.2.3.7 Phospholipase C ...... 29 1.2.3.8 Pyocyanin and other phenazines ...... 30 1.2.3.9 P. aeruginosa secretion systems ...... 32 1.2.3.10 Flagellum and type IV pili required for motility ...... 35 1.2.3.11 Pyoverdine and pyochelin involved in iron uptake ...... 36 1.2.3.12 Exopolysaccharides involved in biofilm formation ...... 42 1.2.3.13 Regulation of virulence gene expression via small RNAs ...... 44 1.2.4 Antibiotic resistance mechanisms in P. aeruginosa ...... 46 1.2.5 Pyocin production by P. aeruginosa ...... 47 1.2.6 The switch from planktonic to chronic lifestyle via c-di-GMP signaling ...... 50 1.3 P. AERUGINOSA IN THE CYSTIC FIBROSIS LUNG ...... 51 1.3.1 From the environment, over acute infection to a chronic lifestyle ...... 51 1.3.2 Genetic adaptation of P. aeruginosa to the cystic fibrosis lung ...... 53 1.3.3 Iron uptake by P. aeruginosa in the cystic fibrosis lung ...... 55 1.3.4 Biofilm formation by P. aeruginosa in the cystic fibrosis lung ...... 57 1.3.5 Growth of P. aeruginosa under microaerophilic/anaerobic conditions ...... 58 1.3.6 Effect of antibiotic treatment on P. aeruginosa in the cystic fibrosis lung ...... 60 1.4 TREATMENT OF CYSTIC FIBROSIS PATIENTS ...... 61

xiii Table of contents

1.4.1 Lung transplantation ...... 61 1.4.2 Gene therapy ...... 61 1.4.3 CFTR correctors and potentiators ...... 61 1.4.4 Intrapulmonary Percussive Ventilation ...... 63 2 AIM OF THE THESIS ...... 65 CHAPTER 2: THE DELETION OF TONB-DEPENDENT RECEPTOR GENES IS PART OF THE GENOME REDUCTION PROCESS THAT OCCURS DURING ADAPTATION OF PSEUDOMONAS AERUGINOSA TO THE CYSTIC FIBROSIS LUNG...... 67 2.1 INTRODUCTION ...... 69 2.2 MATERIALS AND METHODS...... 71 2.2.1 Strains ...... 71 2.2.2 Isolation of P. aeruginosa from CF sputum ...... 71 2.2.3 Typing of P. aeruginosa CF isolates ...... 71 2.2.3.1 Typing via Rep-PCR ...... 71 2.2.3.2 Typing via Multiplex PCR ...... 72 2.2.4 Amplification and sequencing of fpvB region ...... 73 2.2.5 Testing of antibiotic susceptibility of fpvAI, fpvB, and fpvAIfpvB mutants versus wild type P. aeruginosa PAO1 ...... 73 2.2.6 Whole genome sequencing ...... 73 2.2.7 Genome assembly ...... 74 2.2.8 Genome annotation and analysis ...... 74 2.2.9 Confirmation of genomic content via PCR ...... 75 2.2.10 Identification of deletions comprising TonB-dependent receptor genes amongst P. aeruginosa CF isolates of the DK2 lineage ...... 75 2.2.11 RNA extraction from CF sputum and bacterial cultures...... 75 2.2.12 cDNA synthesis ...... 77 2.2.13 Real-time PCR...... 77 2.2.14 Statistical analysis...... 77 2.3 RESULTS ...... 78 2.3.1 The P. aeruginosa CF population at the UZ Brussel is characterized by the absence of a dominant clone ...... 78 2.3.2 The fpvB gene is deleted in a small proportion of P. aeruginosa CF isolates ...... 80 2.3.3 Deletion of ferripyoverdine receptor genes does not confer resistance to antibiotics frequently used to treat CF infections ...... 81 2.3.4 Genome sequence analysis of a Belgian epidemic P. aeruginosa CF isolate reveals deletions in several genomic regions including two TonB-dependent receptor genes ...... 81 2.3.5 P. aeruginosa CF_PA39 belongs to a new sequence type ...... 85 2.3.6 Deletion of TonB-dependent receptor genes in P. aeruginosa strains of the DK2 lineage during adaptation to the CF lung ...... 85 2.3.7 Iron uptake by P. aeruginosa in the CF lung...... 86 2.4 DISCUSSION ...... 88 2.5 CONCLUSION ...... 93 2.6 SUPPLEMENTARY DATA ...... 94 3 CHAPTER 3: EFFECT OF SHEAR STRESS ON PSEUDOMONAS AERUGINOSA ISOLATED FROM THE CYSTIC FIBROSIS LUNG...... 99 3.1 INTRODUCTION ...... 101 3.2 MATERIALS AND METHODS...... 103

xiv Table of contents

3.2.1 Bacterial strains and culture conditions ...... 103 3.2.2 Preparation of artificial sputum medium ...... 103 3.2.3 Rotating wall vessel experiment ...... 103 3.2.4 Determination of bacterial counts...... 104 3.2.5 RNA isolation ...... 104 3.2.6 RNA sequencing and data analysis ...... 105 3.2.7 Reverse transcription ...... 106 3.2.8 Quantitative real-time PCR ...... 106 3.2.9 Scanning electron microscopy ...... 106

3.2.10 Quantification of 3-oxo-C12-HSL ...... 107 3.2.11 Determination of elastase production ...... 107 3.2.12 Qualitative determination of short-chain N-acylhomoserine-lactone production ...... 108 3.2.13 Statistical analyses ...... 108 3.3 RESULTS AND DISCUSSION ...... 108 3.3.1 Two different colony morphologies were identified after growth of P. aeruginosa CF_PA39 in artificial sputum medium ...... 108 3.3.2 High fluid shear levels preclude the formation of self-aggregating biofilms...... 110 3.3.3 Effect of shear stress on the transcriptome of P. aeruginosa CF_PA39 grown in artificial sputum medium ...... 111 3.3.3.1 Genes up-regulated under low fluid shear ...... 116 3.3.3.2 Genes down-regulated under low fluid shear ...... 121 3.3.4 Role of small RNAs in the shear stress response ...... 123 3.3.5 Quorum sensing molecules are slightly higher produced in response to shear stress .... 125 3.4 CONCLUSION ...... 126 3.5 SUPPLEMENTARY DATA ...... 128 4 CHAPTER 4: EFFECT OF INTRAPULMONARY PERCUSSIVE VENTILATION ON P. AERUGINOSA IN THE CF LUNG ...... 131 4.1 INTRODUCTION ...... 133 4.2 MATERIALS AND METHODS...... 134 4.2.1 Set-up of the clinical study and inclusion criteria ...... 134 4.2.2 Determination of lung functions ...... 134 4.2.3 Collection of sputum samples ...... 135 4.2.4 Determination of bacterial load and identification of microorganisms ...... 136 4.2.5 RNA isolation and cDNA synthesis ...... 136 4.2.6 Real-time PCR ...... 137 4.2.7 Statistical analysis ...... 138 4.2.8 Ethics statement ...... 138 4.3 RESULTS ...... 138 4.3.1 Effect of IPV on lung function ...... 138 4.3.2 Change in bacterial load of CF sputum during AD and IPV treatment ...... 139 4.3.3 Effect of IPV on P. aeruginosa gene expression in the CF lung ...... 142 4.4 DISCUSSION ...... 145 4.5 CONCLUSION ...... 146 4.6 SUPPLEMENTARY DATA ...... 147 5 CHAPTER 5: IDENTIFICATION AND FUNCTIONAL ANALYSIS OF A NOVEL S- TYPE PYOCIN...... 149 5.1 INTRODUCTION ...... 151

xv Table of contents

5.2 MATERIALS AND METHODS ...... 152 5.2.1 Strains and plasmids ...... 152 5.2.2 Secondary structure prediction of pyocin S6 ...... 152 5.2.3 RNA isolation and semi-quantitative reverse transcriptase (RT)-PCR ...... 153 5.2.4 Cloning of the pyocin S6 gene ...... 154 5.2.5 Overexpression and purification of the protein ...... 154 5.2.6 Pyocin sensitivity assay ...... 155 5.2.7 Screening of clinical strains for the presence of pyocin S6 ...... 155 5.3 RESULTS AND DISCUSSION ...... 155 5.3.1 Nucleotide and amino acid sequences of pyocin S6 ...... 155 5.3.2 Expression of pyocin S6 under iron-limited versus iron-repleted conditions ...... 159 5.3.3 Cloning, overexpression and purification of pyocin S6 ...... 161 5.3.4 Pyocin S6 activity ...... 162 5.3.5 Screening of clinical strains for the presence of pyocin S6 ...... 164 5.4 CONCLUSION ...... 164 5.5 SUPPLEMENTARY DATA ...... 166 6 CHAPTER 6: GENERAL DISCUSSION AND FUTURE PERSPECTIVES ...... 171 7 REFERENCES ...... 179 8 LIST OF PUBLICATIONS ...... 207

xvi Chapter 1

1 Chapter 1: Literature overview

1 Literature overview

2 Chapter 1

1.1 Cystic fibrosis

1.1.1 Genetic basis

Cystic fibrosis (CF) is a life-threatening hereditary disease that was first described in 1938 by doctor Dorothy Andersen who remarked shared pathologic conditions in very young patients (<1 week to 14.5 years old) who died from either gastrointestinal obstruction or respiratory failure (Andersen, 1938). Via pedigree analysis it was found that CF is an autosomal recessive disorder since ca. 25% of the offspring in 20 CF families was affected (Andersen & Hodges, 1946). In 1989, the genetic basis of cystic fibrosis was revealed as the disease appeared to be caused by a gene located on the long arm of chromosome 7 (Kerem et al., 1989; Riordan et al., 1989; Rommens et al., 1989). The gene product was found to share similarity with a large family of transporters and hence was named cystic fibrosis transmembrane conductance regulator (CFTR) (Riordan et al., 1989). The CFTR gene is approximately 189 kb long and contains 27 exons and 26 introns (Figure 1.1) (Tsui & Dorfman, 2013). The resulting gene product belongs to the superfamily of adenosine triphosphate-binding cassette (ABC) transporters, consists of 1480 amino acids and has two nucleotide-binding domains, two transmembrane domains and one regulatory domain (Figure 1.2).

Figure 1.1. Composition of the CFTR gene. A. Scheme of the exons and introns present in the CFTR gene. Originally, the gene was thought to be composed of 24 exons, later it appeared that exons 6, 14, and 17 contained small introns and were renamed, yielding a total of 27 exons. B. Rescaled exon scheme. C. Distribution of known mutations and polymorphisms. (Tsui & Dorfman, 2013).

3 Literature overview

The two transmembrane domains, each consisting of six α-helices, form the channel pore, while channel gating occurs via the two nucleotide-binding domains. Finally, the cytosolic regulatory domain regulates channel activity via protein kinase A-dependent phosphorylation (Kim & Skach, 2012).

Figure 1.2. CFTR protein structure. The two transmembrane domains (TMD) each consist of six α-helices (1-6 for TMD1 and 7-12 for TMD2). The nucleotide binding domains (NBD) and the intrinsically disordered regulatory domain (R) are located in the cytosol. ICL, intracellular loop. ECL, extracellular loop. ER, endoplasmic reticulum. (Kim & Skach, 2012)

It is assumed that channel gating occurs via the intracellular loops that are formed by the helical extensions of the two transmembrane domains. More specifically, these intracellular loops protrude into the cytosol and provide docking sites for the nucleotide-binding domains. ATP binding and hydrolysis at the interface of the two nucleotide-binding domains cause conformational changes that are transmitted to the transmembrane domains via the intracellular loops and control gating of the channel (Figure 1.2) (Kim & Skach, 2012). CFTR is a cyclic adenosine monophosphate (cAMP)-regulated ion channel that conducts − − chloride (Cl ) as well as bicarbonate (HCO3 ) ions across epithelial cell membranes. Several mutations of the CFTR gene have been documented that can interfere with virtually any step in the process of transcription, over pre-mRNA processing to protein folding and correct functioning of the protein at the apical cell membrane. Over 2000 mutations have been documented in the cystic fibrosis database (http://www.genet.sickkids.on.ca/StatisticsPage.html) and the great majority of these mutations have been categorized into six different classes (Figure 1.3). Class I mutations are nonsense or frameshift mutations that result in premature stop codons yielding truncated and/or non-functional CFTR protein. Missense mutations and in-frame deletions leading to misfolding of CFTR and/or incorrect trafficking of this protein form the most abundant class

4 Chapter 1

(class II) of mutations found in CF patients since more than 85% of CF patients carry a class II mutation on at least one CFTR allele (De Boeck et al., 2014). In particular, the ΔF508 mutation found at exon 11 (formerly named exon 10) is responsible for the overrepresentation of class II mutations among CF patients as its frequency (ca. 85%) is nearly identical to that of class II mutations in general (De Boeck et al., 2014; Riordan, 2008). Furthermore, about half of the European CF population was found to be homozygous for class II mutations (De Boeck et al., 2014). In contrast, class III and IV mutations are less common among CF patients. Mutations belonging to these classes do not interfere with protein folding and/or trafficking since CFTR is present at the cell surface, but instead lead to incorrect gating of the ion channel (class III mutations) or reduced conduction of ions through the channel pore (class IV mutations). Class V mutations cause incorrect splicing of pre-mRNA products leading to reduced amounts of CFTR protein expressed at the apical cell membrane, while class VI mutations affect stability of the CFTR protein at the membrane (Boyle & De Boeck, 2013).

Figure 1.3. Overview of CFTR mutation classes (Boyle & De Boeck, 2013).

As mentioned before, CFTR is a chloride channel and malfunctioning of this protein has dramatic physiological effects. Figure 1.4 shows the normal pathway of chloride secretion in secretory epithelial cells. The sodium (Na+)/potassium (K+) ATPase actively pumps Na+ out of the epithelial cell against its electrochemical gradient. The resulting negative membrane

5 Literature overview potential as well as the increased concentration of Na+ ions outside the cell cause an influx of Na+ via the K+/2Cl−/Na+ cotransporter. Since Cl− ions are cotransported into the epithelial cell against their electrochemical gradient, these ions will leave the cell down their electrochemical gradient via transport through the CFTR ion channel at the time the epithelial cell is stimulated to secrete. Na+ ions follow via a paracellular route and consequently water to restore the osmotic balance (Lyczak et al., 2002). In this way, the airway surface liquid (ASL, a biphasic layer consisting of an upper viscous mucus layer and a lower aqueous periciliary liquid layer, separating the mucus from the epithelial cells) that covers the respiratory epithelial cell lining remains sufficiently hydrated. However, since a malfunctioning or non-functioning CFTR protein is not able to secrete Cl− ions, Na+ is maintained inside the cell or even reabsorbed and the osmotic balance is disrupted, leading to a dehydrated ASL covering the epithelial cells.

Figure 1.4. Chloride secretion in healthy secretory epithelial cells expressing CFTR at the apical cell membrane. Step 1: Na+ ions are actively pumped out of the epithelial cell by a Na+/K+ ATPase increasing the extracellular Na+ pool. Step 2: Na+ follows its electrochemical gradient through a K+/2Cl−/Na+ cotransporter that concomitantly transports Cl− against its electrochemical gradient. Step 3: The epithelial cell is stimulated to secrete and Cl− ions are transported outside the cell through the CFTR ion channel. Step 4: Na+ ions follow the Cl− ions via a paracellular pathway leading to the formation of NaCl. In order to restore the osmotic balance, water diffuses out of the cell. (Lyczak et al., 2002).

It has been observed that mucin secretion is elevated in CF patients compared to healthy individuals (Henderson et al., 2014). In CF patients, dehydration of the ASL leads to a close interaction between the viscous mucus layer and the epithelial cell surface that is otherwise impaired by the aqueous periciliary liquid layer. The combination of a thickened mucus layer and an overall dehydrated ASL leads to blocking of the mucociliary movement that is maintained by the ciliated epithelium in healthy cells (Figure 1.5). Indeed, in CF patients various mucus plugs have been detected in the proximal (bronchi) and distal (bronchioles)

6 Chapter 1 airways (Henderson et al., 2014). In healthy individuals, the mucociliary movement ensures the unidirectional removal of microorganisms from the airways to the esophagus. In addition, the ASL of these individuals contains endogenous antimicrobials that reinforce bacterial clearance from the airways. On the contrary, in the case of CF patients this mucociliary clearance is impaired since the thickened mucus layer interferes with ciliary beating. − Furthermore, the lower levels of HCO3 , that cannot be transported through the defective CFTR ion channel and is important for pH buffering, lead to a decrease in the pH of the ASL. This decrease in pH affects the bacterial killing capacity of the ASL (Pezzulo et al., 2012). Consequently, many microorganisms are able to colonize the airways of CF patients (Figure 1.5) (Stoltz et al., 2015).

1.1.2 Prevalence of cystic fibrosis

Worldwide, about 70,000 people are affected by CF. However, the distribution of CF in the world is non-uniform as it is race-dependent. In particular, the European population is at risk since the birth prevalence there is ca. 1 in 3,000-5,000 live births (O'Sullivan & Freedman, 2009). Even in Europe, differences in prevalence of the disease exist as exemplified by the distribution of the frequency of the most common mutation ΔF508 that is most frequent in the northwest and decreases in frequency towards the southeast. Unsurprisingly, due to their European descent, a similar prevalence of 1 in 3,000 births has been found for white Americans and Australians, while 1 in 4,000-10,000 Latin Americans and 1 in 15,000-20,000 African Americans are affected by the disease (O'Sullivan & Freedman, 2009). In contrast, CF is uncommon in Asian and African populations. The life expectancy of CF patients has greatly increased due to the availability of pancreatic enzyme supplements, intense therapy and medical care in general. In Belgium, the median age of the CF population has increased from 14.8 years in 2000 to 18.9 years in 2010, while the proportion of adult CF patients augmented from 37.7% in 2000 to 52.9% in 2010 (http://www.cysticfibrosisdata.org/ReportsBelgium.html). In addition, it was estimated that British CF patients born between 2000 and 2003 and patients from the USA born in 2010 will have a life span of ca. 40 and 37 years, respectively (Dodge et al., 2007; MacKenzie et al., 2014). However, the extended life expectancy of CF patients implies that novel (respiratory) problems are encountered and this is reflected by a worsening quality of life experienced by CF patients (Sawicki et al., 2011).

7 Literature overview

Figure 1.5. Mucociliary clearance in the airways of healthy individuals versus CF patients. A. In healthy individuals the airway surface liquid (ASL) covering the respiratory epithelium contains sufficient levels of − HCO3 and is well hydrated. Consequently, bacteria that become trapped in the upper mucus layer of the ASL are transported out of the airways via ciliary beating and are simultaneously killed by endogenous antimicrobials. B. In CF patients, the ASL is dehydrated and mainly consists of a thick viscous mucus layer that − blocks the mucociliary clearance movement. Furthermore, HCO3 levels are reduced and endogenous antimicrobial killing mechanisms are less effective leading to the accumulation of bacteria in the CF airways (Stoltz et al., 2015).

8 Chapter 1

1.1.3 Clinical manifestations of cystic fibrosis

Clinical manifestations of cystic fibrosis can occur as early as at the fetal stage, when abnormally viscous pancreatic secretions (due to defective CFTR function) cause intestinal obstruction. After birth, the first feces of the newborn (called meconium) are blocked at the level of the small intestine (ileus), hence causing obstruction of the small intestine (also known as meconium ileus) (Lyczak et al., 2002; van der Doef et al., 2011). If not treated, this syndrome can be lethal to the newborn. Although meconium ileus is the earliest symptom of CF, in most cases the disease is detected via abnormally high concentrations of salt (NaCl) in the sweat of the patient. This is because in CF patients, Cl− is not properly absorbed out of the sweat by CFTR when the sweat travels from the sweat gland to the skin. This leads to higher Cl− concentrations in CF sweat (>60 mEq) compared to normal sweat (<40 mEq) (Lyczak et al., 2002). A sweat test is used to determine sweat Cl− concentrations. Briefly, a substance that stimulates sweat secretion (pilocarpine) is applied on the skin via a patch and two electrodes (one at the same spot of the pilocarpine substance, the other at a second spot) are applied to initiate the sweating process. Next, a filter paper is applied on the same place in order to collect the sweat during 30 minutes and the content is subsequently analyzed (Taylor et al., 2009). Basically, all organ systems that contain epithelia, such as the lungs, the pancreatic ducts, the sweat gland, the intestine, the biliary ducts of the liver, and the male reproductive system, are affected in CF patients. Obstruction of the pancreatic ducts is one of the major symptoms of CF patients. This obstruction leads to inflammation and subsequently destruction of the organ. The decrease or loss of pancreatic enzyme production causes malabsorption of a variety of nutrients, in particular fats. Consequently, patients used to die because of malnutrition and subsequent poor growth during the first decade of their live, when not treated. Nowadays, due to the replacement of pancreatic enzymes, the life span of patients suffering from exocrine pancreatic insufficiency, has been greatly extended (Cutting, 2015; Li & Somerset, 2014). CF patients also suffer from a specific type of diabetes (different from type 1 or 2 diabetes) at later ages (about half of the adult CF population has diabetes compared to 1.5% in children <10 years old) (Ode & Moran, 2013). It is still unclear whether the impaired functioning of the endocrine pancreas is due to direct effects of the malfunctioning CFTR channel or collateral damage suffered during destruction of the exocrine pancreas (Cutting, 2015; Ode & Moran, 2013). Chronic pulmonary disease is the major hallmark of CF. The chronic colonization of the CF airways by microorganisms leads to continuous inflammation, and eventually irreversible lung tissue damage. Therefore,

9 Literature overview chronic lung infections, in particular with P. aeruginosa, are considered to be the main cause of morbidity and mortality in a CF population with an increasing life expectancy. Although mutation of CFTR lies at the basis of all of these clinical manifestations, other factors have been shown to determine whether a CF patient experiences the corresponding symptoms, or not. These include variations of genes other than CFTR (genetic modifiers) as well as environmental and stochastic factors. In particular, the combination of certain types of CFTR mutations, genetic modifiers and environmental conditions determines the degree of disease severity (Cutting, 2015).

1.1.4 Pathogens associated with cystic fibrosis

Disruption of the mucociliary clearance mechanisms leads to the accumulation of a variety of microorganisms in the CF airways. However, the colonization of the CF lungs by certain pathogens appears to be age-dependent (Figure 1.6). In general, young CF patients are initially colonized by Staphylococcus aureus (S. aureus) and/or Haemophilus influenzae (H. influenza). At later ages, Pseudomonas aeruginosa (P. aeruginosa) becomes the dominant pathogen present in the CF lungs. It is believed that 60-80% of CF patients are colonized by P. aeruginosa at the age of 18 years (Harrison, 2007; Hauser et al., 2011). Chronic colonization of the CF lungs by this pathogen is correlated with poor clinical outcomes (Ahlgren et al., 2015). In addition, the majority of CF patients that have been colonized by P. aeruginosa, carry this bacterium for life, despite the temporary eradication, or more likely, reduction in abundance of the pathogen. Therefore, the apparent decline of P. aeruginosa in older CF patients (around the age of 30 years) (Figure 1.6) most probably reflects the fact that many people that were colonized by P. aeruginosa have already succumbed at this age. CF patients are often co-infected with both P. aeruginosa and S. aureus (Ahlgren et al., 2015; Goddard et al., 2012; Harrison, 2007), and it has been shown that S. aureus pre-colonization is even a risk factor for the later acquisition of P. aeruginosa (Maselli et al., 2003). The extensive use of antibiotics by CF patients leads to the emergence of multidrug-resistant (MDR) P. aeruginosa strains identified at later ages, peaking to 20% of the CF population at an age of about 35 years. Methicillin-resistant S. aureus (MRSA) have been found to reach a peak frequency at an even earlier age (> 20% of CF patients between 6-24 years). The methicillin resistance is obtained when a methicillin-sensitive S. aureus (MSSA) acquires the staphylococcal chromosomal cassette carrying the mecA gene (encoding a penicillin-binding protein) through horizontal gene transfer (Parkins & Floto, 2015). Similar to P. aeruginosa,

10 Chapter 1 the extensive exposure to antibiotics imposes a risk factor for acquiring MRSA. In addition, MRSA strains can also be obtained in the hospital environment or in the community. In the USA, the prevalence of MRSA among CF patients has steadily increased from ca. 10% in 2003 to 30% in 2012/2013 (Parkins & Floto, 2015). Stenotrophomonas maltophilia (S. maltophilia), Burkholderia cepacia complex (B. cepacia complex), both formerly being classified in the genus Pseudomonas (before the classification based on genotyping techniques), and Achromobacter xylosoxidans (A. xylosoxidans) occur at lower frequencies in the CF population (<20%) (Hauser et al., 2011).

Figure 1.6. Age-dependent prevalence of pathogens in the airways of CF patients. P. aeruginosa, Pseudomonas aeruginosa; H. influenzae, Haemophilus influenzae; S. aureus, Staphylococcus aureus; MDR-PA, multidrug- resistant Pseudomonas aeruginosa; S. maltophilia, Stenotrophomonas maltophilia; B. cepacia complex, Burkholderia cepacia complex; A. xylosoxidans, Achromobacter xylosoxidans; MRSA, methicillin-resistant Staphylococcus aureus (Folkesson et al., 2012).

B. cepacia complex encompasses a group of at least 18 species and some of these species (Burkholderia cenocepacia and Burkholderia multivorans) have been found to cause the cepacia syndrome. This syndrome occurs when the bacteria enter the bloodstream causing bacteremia and cause a multitude of acute symptoms including fever, leucocytosis, necrotizing pneumonia, and progressive deterioration leading to death of the CF patient (Hauser et al., 2011). Other microorganisms isolated from CF lung infections include nontuberculous mycobacteria (NTM) and the fungus Aspergillus fumigatus (A. fumigatus). NTM can cause chronic pulmonary infection and progressive inflammation of the lung tissue but can also transiently or permanently reside in the CF airways without causing symptoms.

11 Literature overview

A. fumigatus is frequently isolated from older CF patients and those that receive chronic inhaled antibiotics. Upon chronic colonization of the CF lung by A. fumigatus, some CF patients can develop an allergic reaction to it, called allergic bronchopulmonary aspergillosis (Hauser et al., 2011).

1.1.5 Host-pathogen interactions in the cystic fibrosis lung

The mucociliary clearance mechanism is the first line of innate immunity as this anatomical barrier normally clears the majority of microorganisms out of the lungs. However, since this barrier is not present (or at least functionally impaired) in the CF lung, major pathogens associated with CF such as P. aeruginosa and S. aureus accumulate in the viscous mucus layers that cover the respiratory epithelial cells. Although some studies have indicated that the transcription factor nuclear factor kappaB (NF-κB) that regulates the transcription of pro- inflammatory cytokines is already constitutively expressed before infection (Khan et al., 1995; Tirouvanziam et al., 2000; Verhaeghe et al., 2007), the presence of microorganisms enforces the pro-inflammatory response. More specifically, the pathogen-associated molecular patterns (PAMPs) derived from P. aeruginosa (i.e. flagellin and lipopolysaccharide (LPS)) and S. aureus (i.e. peptidoglycan) are recognized by the Toll-like receptors (TLRs) present at the apical surface of the epithelial cells (Figure 1.7A). TLR2 is activated upon recognition of peptidoglycan, while flagellin stimulates TLR5 (Kawai & Akira, 2007). In addition, TLR4, which recognizes LPS, is restricted to the endosome. Activation of the TLRs initiates a signaling cascade that leads to NF-κB activation and consequently expression of pro-inflammatory genes (Figure 1.7A). When inactive, NF-κB is sequestered to the cytoplasm by its inhibitory protein I-κB. However, stimulation of TLR2 and TLR5 leads to the activation of the adaptor protein Myeloid Differentiation factor 88 (MyD88). The MyD88-dependent signaling cascade will subsequently cause activation of the I-κB kinase complex, resulting in phosphorylation and finally proteasome-mediated degradation of I-κB, allowing NF-κB to enter the nucleus and initiate transcription of genes associated with inflammation (Cohen & Prince, 2012; Kawai & Akira, 2007). TLR4, whose expression is restricted to the endosome, has an additional pathway (next to the MyD88-dependent patway) that is activated at a later stage and leads to resolution of the acute inflammatory response. This pathway is called the TIR-domain-containing adapter-inducing interferon-β (TRIF)-dependent pathway and leads to activation of interferon (IFN) regulatory factor 3 (IRF3) via phosphorylation. Finally, phosphorylated IRF3 dimerizes and translocates into the nucleus to initiate transcription of

12 Chapter 1 type I IFN genes (i.e. IFN-β). The resulting type I IFN response is crucial for resolution of the acute inflammation as well as the dendritic cell-mediated activation of the adaptive immune system. However, TLR4 activation of TRIF from the endosome is impaired, leading to constitutive activation of NF-kB and inflammation of the CF airways without the efficient clearance of important CF pathogens such as P. aeruginosa and S. aureus. The pro- inflammatory cytokines (IL-8, GM-CSF, and TNF-α) recruit polymorphonuclear neutrophils (PMNs) that accumulate in the CF airways. These PMNs release reactive oxygen species (ROS) in the airways that cause oxidative damage to the bacterial cells as well as to the host cells (Figure 1.7B). The resulting oxidative stress activates Mitogen-activated protein kinase (MAPK)-signaling pathways that enhance the production of the pro-inflammatory cytokine IL-8, leading to new waves of neutrophil recruitment. Furthermore, due to the CFTR mutation, the antioxidants glutathione (GSH) and thiocyanate (SCN−) cannot be trafficked into the ASL and hence are unable to counter the oxidative damage (Figure 1.7B) (Cohen & Prince, 2012). One of the important defense mechanisms used by PMNs in the CF airways are neutrophil extracellular traps (NETs). These are meshworks of neutrophil DNA, histones, and proteolytic enzymes produced in order to trap CF pathogens (Yonker et al., 2015). Nevertheless, NETs have limited success in killing CF pathogens (in particular P. aeruginosa) and actually contribute to the viscosity of the CF sputum. Although the CF airways are characterized by an innate immune response leading to the production of pro-inflammatory cytokines, adaptive immunity plays a role in clearing CF pathogens since antibodies directed against alginate (Pedersen et al., 1990), exoproteins (Hollsing et al., 1987b), and the Type III secretion system of P. aeruginosa (Corech et al., 2005), as well as teichoic acid and alpha toxin of S. aureus (Hollsing et al., 1987a) have been detected in sputa from CF patients. Altogether, the CF lung is considered to be in a state of persistent inflammation, thereby inducing adaptation of the CF pathogens to survive in this hostile environment.

13 Literature overview

Figure 1.7. A. Pathogen-associated molecular patterns (PAMPs) originating from CF pathogens such as P. aeruginosa and S. aureus are recognized by Toll-like receptors (TLRs) 2 and 5 expressed at the apical membrane of the respiratory epithelial cells. This will initiate a signaling cascade via MyD88 that will lead to the phosphorylation and proteasome- dependent degradation of I-κB, releasing NF-κB. After entering the nucleus, NF-κB activates transcription of genes encoding pro-inflammatory cytokines such as IL-8 and GM-CSF. These cytokines recruit polymorphonuclear neutrophils (PMNs) to the respiratory epithelium. Intracellular activation of TRIF by TLR4 is impaired, preventing IRF3 from translocating into the nucleus and activating the production of type I interferons. This prevents resolution of the pro-inflammatory response, resulting in constitutive activation of NF-κB. B. The constitutive production of pro-inflammatory cytokines such as IL-8 causes an accumulation of PMNs in the CF airways. PMNs release reactive oxygen species (ROS) that cannot be counteracted by the antioxidants glutathione (GSH) and thiocyanate (SCN−), since they require a functional CFTR to be secreted. In addition, SCN− has an antimicrobial effect that cannot be exploited due to the CFTR mutation (Cohen & Prince, 2012).

14 Chapter 1

1.1.6 Available in vitro and animal models to study cystic fibrosis infections

Several in vitro and in vivo (animal) models exist to study CF, each of them having specific advantages and disadvantages.

1.1.6.1 Tissue cultures

Tissue cultures are the simplest model to use to study the interactions between epithelial cells and pathogens or to study the effect of novel drugs on these cells in order to determine if they have an effect in restoring CFTR function. With regard to airway infections, primary respiratory epithelial cells are the most relevant system. The disadvantage is that these cells are not easily accessible since they need to be obtained directly form the CF lung. An alternative is the reprogramming of human embryonic stem cells from cystic fibrosis patients to secretory epithelial cells (Wong et al., 2012) or the formation of intestinal organoids via the expansion of stem cells present in primary intestinal cultures (Dekkers et al., 2013). The advantage of all of these tissue culture models is that they enable the application of personalized medicine, so that each individual CF patient can be helped.

1.1.6.2 The rotating wall vessel (RWV)

Disruption of the mucociliary clearance mechanism in the CF lung changes the viscosity of the ASL that covers the respiratory epithelium and hence microorganisms experience completely different physical conditions in the CF lung environment. More specifically, it is believed that the CF lung is characterized by low levels of fluid shear (Blake, 1973; Knowles & Boucher, 2002). A model that simulates the low fluid shear levels experienced in the CF lung is the rotating wall vessel (RWV) (Figure 1.8A & B). When this cylindrical bioreactor is completely filled with culture medium and rotated on an axis parallel with the ground, this results in a solid body mass rotation of the medium creating a low fluid shear environment. When bacteria are inoculated in the culture medium, they will remain in suspension in a restricted fluid orbit (Figure 1.8C) (Crabbé et al., 2008; Nauman et al., 2007). In addition, this device can be used to design a 3-D monotypic alveolar lung epithelium model (Carterson et al., 2005) or even a multicellular immune-component model consisting of both lung epithelial cells and immune cells (Crabbé et al., 2011a). The advantage of the RWV is that a multitude

15 Literature overview of combinations of microorganisms and cell types can be combined allowing to study their interactions under similar physical conditions as those that are present in the CF lung.

Figure 1.8. A. Set-up of the rotating wall vessel (RWV) system. B. Close-up of the RWV bioreactor showing the two sampling ports, the filling port and the gas-permeable membrane. C. During rotation, (bacterial) cells remain in suspension in a restricted fluid orbit (Crabbé et al., 2008).

1.1.6.3 Animal models

When studying genetic diseases, animal models are considered to be the gold standard. In total, five different CF animal models have been engineered (Cutting, 2015). However, due to the substantial physiological differences between the animals, the clinical symptoms developed by these models vary greatly (Table 1.1). Regarding the most crucial CF feature, pulmonary disease, rodents that have lost the CFTR gene, or carry a mutated copy, do not develop the characteristic symptoms, although displaying CFTR-related ion transport abnormalities. In contrast to mice, ferrets, and especially pigs do better resemble the human lung architecture. Although the CF pig develops pulmonary disease comparable to that observed in humans, it starts at an earlier stage of live. Next to pulmonary disease, several differences in clinical symptoms are observed, even between mammalian models. For instance, the mouse model displays minimal pancreatic exocrine dysfunction compared to humans, while this is severe in case of the porcine model. Despite the variation in several clinical CF symptoms, the pig model appears to closely resemble the pulmonary disease that characterizes CF patients and hence may be a valid animal model to study CF lung infections. Nevertheless, performing host-pathogen interaction studies using porcine models remains a great challenge.

16 Chapter 1

Table 1.1. Comparison of the characteristic CF features between different CF animal models (Cutting, 2015). Features* Human Pig Ferret Mouse Rat Zebrafish Aberrant chloride transport ↑ ↑ ↑ ↑ ↑ ↑ Intestinal obstruction ↑ ↑↑ ↑↑ ↑↑‡ ↑↑‡ – Growth disturbance ↑ ↑↑ ↑↑ ↑↑ ↑ – Maldevelopment of trachea ↑ ↑ ↑ ↑ ↑ – Pancreatic exocrine dysfunction ↑ ↑↑ ↑ – – – Obstructive lung disease ↑ ↑ ↑ – – – Liver dysfunction ↑ ↑↑ ↑ – – – Diabetes mellitus ↑ ↑ ↑ – – – Anomalous vas deferens ↑ ↑ ↑ – ↑ – ‡ Intestinal obstruction occurs after the neonatal period in murine models around the time of weaning.

1.2 Pseudomonas aeruginosa

1.2.1 Omnipresence of P. aeruginosa

P. aeruginosa is a Gram-negative, rod-shaped, aerobic bacterium that can be isolated from a broad range of environments. P. aeruginosa is ubiquitous in the environment as it has been commonly found in the rhizosphere, associated with plants, as well as in water samples and is considered to be harmless to healthy individuals (Hu et al., 2005; Pirnay et al., 2005; Pirnay et al., 2009; Selezska et al., 2012). However, people suffering from severe burn wounds, immunocompromised patients (cancer patients receiving chemotherapy and AIDS patients), as well as cystic fibrosis patients can be infected by this bacterium as it takes advantage of the opportunity to colonize these individuals (Hauser et al., 2011; Kerr & Snelling, 2009; Lyczak et al., 2002; Pirnay et al., 2009; Rabello et al., 2015; Rosenberg et al., 2001). Pirnay and colleagues found that the P. aeruginosa population structure is epidemic. An epidemic population structure is an intermediate between a panmictic structure, characterized by a random association of alleles and a clonal population structure characterized by non-random association of alleles, resulting in the frequent isolation of only a few of the many multilocus genotypes. More specifically, the P. aeruginosa population structure is predominantly panmictic, with some epidemic clonal complexes showing significant association between loci. This epidemic population structure of P. aeruginosa was also observed in the study of Cramer et al., were it was found that the five most common clones in the CF population can also be retrieved among the 20 most represented clones in the environment and non-CF

17 Literature overview disease habitats, thereby excluding the existence of one widespread transmissible P. aeruginosa CF clone (Cramer et al., 2012).

1.2.2 Genomic features of P. aeruginosa

The P. aeruginosa genome is constituted of a single circular chromosome and a variable number of plasmids. The size of this genome varies between different P. aeruginosa strains, ranging from 6.26 million bases (Mbp) to 7.50 Mbp, with an average genome size of 6.59 Mbp for the 26 completely sequenced P. aeruginosa genomes currently available (Table 1.2). Furthermore, the P. aeruginosa genome has an average GC content of 66.27% (range: 65.6- 66.60%) and contains on average 6087 genes (range: 5697-7070 genes) (Table 1.2). The large repertoire of transcriptional regulators, transporters, and two-component regulatory systems encoded by the P. aeruginosa genome is responsible for the metabolic versatility of this opportunistic pathogen, allowing it to utilize various nutrient sources and hence to colonize different environments. The P. aeruginosa genome can be subdivided in a highly conserved core genome that is shared among all of the P. aeruginosa isolates, showing sequence diversities of as low as 0.5-0.7%, and a highly variable accessory genome that is responsible for the diversity of P. aeruginosa isolates (Klockgether et al., 2011). In fact, the accessory genome consists of extrachromosomal elements (i. e. plasmids) and regions of genome plasticity (RGP) that contain genomic islands, integrated prophages, transposons or insertion sequence (IS) elements, all considered to be integrated in the chromosome subsequent to horizontal gene transfer (Klockgether et al., 2011). Recently, via sequence analysis of the genomes of isolates representing the 15 most frequent clonal complexes, it was estimated that the core genome of P. aeruginosa consist of at least 4,000 genes, while the accessory genome consists of about 10,000 genes in RGPs commonly found in P. aeruginosa clonal complexes. Furthermore, it was estimated that about 30,000 genes or more are present in only a few strains or clonal complexes, hence considered to be rare genes (Hilker et al., 2015). The resulting pangenome thus consists of a relatively small conserved core genome, a larger combinatorial accessory genome containing genes found in RGPs that are present in multiple clonal complexes and a very large portion of rare genes that are only found in a few isolates or clonal complexes (Hilker et al., 2015).

18 Chapter 1

Table 1.2. Features of completed P. aeruginosa genomes. Strain Source Genome size (Mbp) GC% Genes P. aeruginosa PAO1 Wound 6.26 66.60 5697 P. aeruginosa UCBPP-PA14 Wound 6.54 66.30 5980 P. aeruginosa PA7 Clinical, non-respiratory 6.59 66.40 6079 P. aeruginosa LESB58 CF 6.60 66.30 6132 P. aeruginosa PACS2 CF 6.49 66.30 - P. aeruginosa 19BR Clinical, non-CF 6.74 66.10 - P. aeruginosa 213BR Clinical, non-CF 6.72 66.10 - P. aeruginosa M18 Rhizosphere 6.33 66.50 5825 P. aeruginosa DK2 CF 6.40 66.30 5920 P. aeruginosa NCGM2.S1 Urinary tract infection 6.76 66.10 6278 P. aeruginosa B136-33 Infant with Shanghai fevera 6.42 66.40 5876 P. aeruginosa RP73 CF 6.34 66.50 5856 P. aeruginosa VRFPA04 Eye infection 6.82 66.50 6308 P. aeruginosa PA1 Oil waste 6.53 66.30 6055 P. aeruginosa PA1R Oil waste 6.31 66.30 5863 P. aeruginosa MTB-1 t-HCH contaminated soilb 6.58 66.20 6077 P. aeruginosa LES431 CF 6.55 66.30 6059 P. aeruginosa SCV20265 CF 6.72 66.30 6261 P. aeruginosa YL84 Compost 6.43 66.40 5908 P. aeruginosa F22031 Pubic bone from a cancer patient 6.60 66.20 6195 P. aeruginosa NCGM 1984 Urinary catheter 6.85 66.00 6399 P. aeruginosa NCGM 1900 Urinary catheter 6.81 66.00 6354 P. aeruginosa FRD1 CF 6.71 66.10 6179 P. aeruginosa Carb01 63 Rectal swab from hospitalized patient 7.50 65.60 7070 P. aeruginosa DSM 50071 Unknown 6.32 66.50 5803 P. aeruginosa F9676 Rice seed 6.37 66.50 5828 Average 6.59 66.27 6087 Based on information available on the NCBI website (http://www.ncbi.nlm.nih.gov/genome/genomes/187?). Mbp, million base pairs. -, information not available. aShanghai fever is a sepsis accompanied with enteric disease (diarrhea) caused by P. aeruginosa. bSoil contaminated with technical grade hexachlorocyclohexane (t-HCH).

19 Literature overview

1.2.3 P. aeruginosa virulence factors

Table 1.3. Overview of the major virulence factors produced by P. aeruginosa. Virulence factor Function Cell-associated/secreted Reference Elastase/LasB Elastolytic activity Secreted, TIISS (Bainbridge & Fick, 1989; protease Degradation of collagen Heck et al., 1990; Degradation of IgA and IgG Mariencheck et al., 2003; Degradation of lung surfactant Thibodeau & Butterworth, proteins A and D 2013) Disruption of epithelial cells Staphylolysin/LasA Elastolytic activity Secreted, TIISS (Kessler et al., 1993; protease Serine protease activity Kessler et al., 1998) Rhamnolipids Biosurfactant (motility) Secreted (Abdel-Mawgoud et al., Lysis of PMNs 2010; Alhede et al., 2009) Biofilm dispersion PQS QS signal molecule Extracellular (Diggle et al., 2007) Fe3+-chelator HCN Inhibition of aerobic Volatile (Blumer & Haas, 2000; respiration Lenney & Gilchrist, 2011) Exotoxin A Inhibition of protein synthesis Secreted, TIISS (Pollack, 1983) Phospholipases C Degradation of lung surfactant Secreted, TIISS (Holm et al., 1991; Vasil Haemolysin (PlcH) et al., 1991) Phenazines Oxidative stress (ROS) Extracelullar (Briard et al., 2015; Reduction of Fe3+ to Fe2+ Mavrodi et al., 2001; Fe2+-chelator (1-HP) Wang et al., 2011) ExoS, ExoT, ExoU, Toxins Secreted, TIIISS (Hauser, 2009) ExoY Modulation of immune system Disruption of epithelial and endothelial barriers Flagellum Swimming motility Cell-associated (Doyle et al., 2004) Swarming motility Type IV pili Twitching motility Cell-associated (Kearns, 2010) Adhesion Aggregation Pyoverdine High affinity siderophore Extracellular (Cornelis et al., 1989; Signal molecule Meyer, 2000; Visca et al., 2007) Pyochelin Low affinity siderophore Extracellular (Cox et al., 1981; Dumas et al., 2013) Alginate Biofilm formation Secreted (Franklin et al., 2011) Pel Biofilm formation Secreted (Franklin et al., 2011) exopolysaccharide Psl Biofilm formation Secreted (Franklin et al., 2011) exopolysaccharide TIISS, type II secretion system. PMN, polymorphonuclear neutrophil. HCN, hydrogen cyanide. 1-HP, 1- hydroxyphenazine. TIIISS, type III secretion system.

20 Chapter 1

1.2.3.1 Regulation of virulence via quorum sensing and two-component regulatory systems

In analogy to many other Gram-negative bacteria, P. aeruginosa relies on quorum sensing (QS) systems to regulate the expression of genes encoding virulence factors, allowing it to establish infection in the human host. Quorum sensing is a cell-to-cell communication mechanism that allows bacteria to coordinate gene expression in a cell density-dependent manner via the production of autoinducer signal molecules. In P. aeruginosa, three major QS systems have been described that control the expression of virulence genes in a hierarchical way. Two of these systems (las and rhl) rely on N-acyl-homoserine lactones (AHLs) as signal molecules, while one system (pqs) responds to 2-alkyl-4-quinolones (AQs) (Schuster & Greenberg, 2006; Williams & Camara, 2009). AHLs consist of fatty acids that are linked to a homoserine lactone moiety via a peptide bond (Figure 1.9). The fatty acid moiety can vary in length (usually 4-18 carbons in length) and can be modified via a 3-oxo or 3-hydroxy substituent (Hirakawa & Tomita, 2013). AHLs are synthesized by members of the LuxI family of proteins, while being sensed by members of the LuxR family of transcriptional regulators (Fuqua et al., 1996). The las QS system produces and responds to 3-oxo-C12-HSL, while the rhl QS system produces and is activated by C4-HSL. When a P. aeruginosa population reaches a certain density at which the threshold concentration of a certain AHL is reached, the AHL will form a complex with the cognate LuxR family member of transcriptional regulators, thereby activating the transcription of a number of virulence genes in combination with the activation of the gene encoding the LuxI family member, hence inducing its own production. In the case of the las QS system, the accumulation of 3-oxo-

C12-HSL above a certain threshold will result in the formation of a LasR-3-oxo-C12-HSL complex that activates the expression of genes encoding several virulence factors associated with acute infection (including lasA, lasB, and toxA) as well as the rhlR gene and lasI, inducing its own production (Figure 1.9). Next, the rhl QS system is activated after reaching a threshold concentration of C4-HSL, leading to the formation of a RhlR-C4-HSL complex that will bind to the promoter region of genes involved in the production of rhamnolipids, the LasA and LasB proteases (already induced by the las system), pyocyanin and hydrogen cyanide (HCN) among others (Figure 1.9). QS is tightly controlled by a number of transcriptional repressors such as RsaL. This repressor binds simultaneously with the LasR-3- oxo-C12-HSL complex to the lasI promoter, inhibiting its expression. In addition, it controls

21 Literature overview the expression of the genes involved in the biosynthesis of pyocyanin and HCN by directly binding to their promoters.

Figure 1.9. Las and Rhl quorum sensing sytems of P. aeruginosa. When a certain threshold concentration is reached, the autoinducer molecule 3-oxo-C12-HSL forms a complex with the transcriptional regulator LasR. The resulting autoinducer-LasR complex will subsequently activate the expression of several genes encoding virulence factors as well as the lasI gene, thereby inducing production of its own signaling molecule. In addition, the 3-oxo-C12-HSL–LasR complex will activate the rhlR gene in a hierarchical fashion. Upon reaching a threshold concentration, the C4-HSL signal molecule will form a complex with RhlR. The resulting complex will activate transcription of a multitude of genes encoding virulence factors associated with acute infection as well as the rhlR gene. Some genes are activated by LasR, others by RhlR or by both LuxR regulators (Jimenez et al., 2012).

The third QS system of P. aeruginosa, pqs, responds to the AQ signal molecules 2-heptyl-3- hydroxy quinolone, also known as the Pseudomonas Quinolone Signal (PQS) and its precursor 2-heptyl-4(1H) quinolone (HHQ) (Diggle et al., 2006). Biosynthesis of PQS starts with the condensation of anthranilate and a β-keto fatty acid (Pistorius et al., 2011). The anthranilate substrate can be obtained via two different pathways: either via the biodegradation of tryptophan via the kynurenin pathway or via the conversion of chorismate mediated by the PhnAB enzymes (Figure 1.10) (Jimenez et al., 2012). The enzymes PqsABCD catalyze the synthesis of the PQS precursor HHQ in a multistep process. Finally,

22 Chapter 1

HHQ is hydroxylated by PqsH, resulting in the formation of PQS, the pqsH gene being under the control of LasR (Diggle et al., 2006). Both PQS and HHQ can bind to the PqsR transcriptional regulator (also known as the multiple virulence factor regulator MvfR) resulting in the auto-induction of PQS biosynthesis as well as activating expression of the genes involved in the production of pyocyanin and pqsE by the PQS-PqsR and HHQ-PqsR complexes (Jimenez et al., 2012). Another gene belonging to the pqs operon, pqsE, is not required for the biosynthesis of PQS (Gallagher et al., 2002), but recently has been shown to contribute to HHQ synthesis via its role as a thioesterase (Drees & Fetzner, 2015). This role can be taken over by the broad-specificity thioesterase TesB, explaining why PqsE is not essential for HHQ and subsequently PQS biosynthesis (Drees & Fetzner, 2015). In addition to its role as a thioesterase, PqsE appears to be a major virulence effector acting downstream of the pqs QS system as it controls the expression of the virulence factors pyocyanin, lectins, hydrogen cyanide, and rhamnolipids. Finally, the pqs and AHL QS systems appear to be hierarchically connected since the pqsR gene is positively regulated by LasR, while being negatively regulated by RhlR. In addition, PqsE has been shown to activate the expression of PQS-regulated genes independently of PQS or PqsR. However, this regulatory role of PqsE is linked to the rhl QS system as it requires RhlR (Farrow et al., 2008). Recently, the existence of a fourth QS system has been suggested based on the discovery of a novel signal molecule, called Integrated Quorum Sensing Signal (IQS) (Lee et al., 2013). IQS, structurally known as 2-(2-hydroxyphenyl)-thiazole-4-carbaldehyde (Figure 1.11) and recently found to be identical to the previously described aeruginaldeyde, is in fact a byproduct of pyochelin and enantiopyochelin biosynthesis in respectively P. aeruginosa and Pseudomonas fluorescens (P. fluorescens) (Ye et al., 2014). Although the addition of exogenous IQS was able to restore C4- HSL and PQS production in a PAO1ΔambB mutant strain in a dose-dependent manner, the exact working mechanism of the iqs QS system remains largely unknown, since it is unlikely that the amb gene cluster is responsible for the production of IQS (Rojas Murcia et al., 2015; Ye et al., 2014).

23 Literature overview

Figure 1.10. The pqs quorum sensing system of P. aeruginosa. The PQS precursor HHQ can be synthesized from anthranilate in a multistep process involving the pqsABCD gene products. Anthranilate can either be derived from the degradation of tryptophan mediated by the kynurenin pathway or from the conversion of chorismic acid. Finally, HHQ is converted into PQS by PqsH via the addition of a hydroxyl group. Both PQS and HHQ can bind to PqsR and activate transcription of the pqs genes, thereby inducing their own production as well as the production of PqsE and pyocyanin. PqsE acts as a major virulence effector as it is regulates expression of genes involved in the biosynthesis of virulence factors such as pyocyanin, lectins, hydrogen cyanide and rhamnolipids (Jimenez et al., 2012).

Figure 1.11. IQS structure (Lee & Zhang, 2015).

24 Chapter 1

In addition to QS, P. aeruginosa utilizes two-component regulatory systems to control the production of several virulence factors. These two-component regulatory systems allow P. aeruginosa to sense environmental stimuli and accordingly switch between planktonic and biofilm modes of growth. In general, two-component systems consist of an inner-membrane spanning sensor kinase that detects an environmental stimulus and a cytoplasmic response regulator that classically binds DNA (Laub & Goulian, 2007). The genes encoding the sensor kinase and the cognate response regulators are often organized in operons. More specifically, the sensor kinase consists of an N-terminal periplasmic input domain and a C-terminal cytoplasmic transmitter domain that has ATP-binding and histidine kinase activity. The response regulator on the other hand consists of a conserved aspartate-containing receiver domain and a variable C-terminal output domain that often has DNA-binding activity. Once a dimeric membrane-bound sensor kinase senses a stimulus via its N-terminal periplasmic input domain, this will result in trans-autophosphorylation at conserved histidine residues in its C- terminal cytoplasmic transmitter domain (Laub & Goulian, 2007). Subsequently, the phosphoryl group is transferred from the transmitter domain of the sensor kinase to the aspartate residues at the receiver domain of the response regulator. Consequently, phosphorylation of the response regulator receiver domain results in conformational changes of the response regulator that alter the activity of the output domain. In total, more than 60 two-component regulatory systems have been described (Gooderham & Hancock, 2009; Jimenez et al., 2012).

One of the most important two component systems in regulating the switch from acute to chronic infections is the GacS/GacA system (Figure 1.12). When GacS senses a still unknown input signal, this will lead to the phosphorylation and subsequent activation of its cognate response regulator, GacA. Next, GacA up-regulates the expression of the small regulatory RNAs RsmZ and RsmY. These small RNAs form a complex with the RNA-binding protein RsmA, alleviating the translational repression of many virulence factors associated with chronic infection by this protein, while repressing virulence factors associated with acute infection (Heurlier et al., 2004). Several other kinases have been shown to interact with the GacS/GacA system. The sensor kinase LadS acts in parallel with GacS, and positively regulates the pel genes involved in biofilm formation, while negatively regulating the type III secretion genes (Ventre et al., 2006). In contrast to this, the sensor kinase RetS activates the expression of virulence factors associated with acute infection (Ventre et al., 2006). RetS is able to form heterodimers with GacS, preventing autophosphorylation of the transmitter

25 Literature overview domain of this sensor kinase and interfering with the phosphorylation of GacS. This results in lower levels of the RsmA-RsmZ complex and higher levels of free RsmA, resulting in repression of virulence factors associated with chronic infection and the increased expression of virulence factors associated with acute infection (Jimenez et al., 2012).

Figure 1.12. The GacS/GacA two-component regulatory system. The sensor kinase GacS senses an input signal at its N-terminal periplasmic domain resulting in autophosphorylation of the histidine residues at its C-terminal cytoplasmic transmitter domain. The transfer of a phosphoryl group from the transmitter domain of GacS to the receiver domain of GacA results in the up-regulation of the small RNAs RsmZ and RsmY by this protein. RsmZ and RsmY bind to RsmA that acts as a translational repressor. When RsmA is bound to these small RNAs, repression of genes involved in biofilm formation (pel and psl) and chronic infection in general is released, while the expression of virulence factors related to acute infection such as the type III secretion system is repressed. LadS works in parallel to GacS, while RetS blocks GacA phosphorylation via the formation of heterodimers with GacS, hence resulting in the activation of acute virulence factors and the repression of chronic virulence determinants (Jimenez et al., 2012).

1.2.3.2 Elastase and staphylolysin

P. aeruginosa produces two elastases, LasA (also called staphylolysin) and LasB, both capable of degrading elastin. These proteases are synthesized as preproenzymes that contain a signal sequence that is cleaved during transport across the inner membrane. Both proenzymes remain inactive inside the cell and require proteolytic cleavage of the N-terminal propeptide

26 Chapter 1 part in order to become enzymatically active (Kessler et al., 1998). In case of LasB, this is achieved by autocatalytic cleavage in the periplasm mediated by the enzyme itself. The propeptide remains covalently bound to the enzyme until dissociation of the complex takes place during translocation across the outer cell membrane. In contrast, LasA is transported as an unprocessed precursor enzyme and becomes active upon the proteolytic activity of the LasB elastase or a lysine-specific protease (Kessler et al., 1998). Both enzymes are secreted via the type II secretion system (Filloux, 2011). In addition to elastin, LasB is able to degrade collagen, immunoglobulins IgA and IgG, the opsonizing lung surfactant proteins A and D and is involved in the disruption of tight junctions (Bainbridge & Fick, 1989; Gellatly & Hancock, 2013; Heck et al., 1990; Mariencheck et al., 2003; Thibodeau & Butterworth, 2013). LasA has a lower elastolytic activity compared to lasB and rather has a synergistic effect on LasB- mediated elastin degradation. In fact LasA is a serine protease that is able to cleave the pentaglycine cross-link in the peptidoglycan of the S. aureus cell wall, hence being called Staphylolysin (Kessler et al., 1993). In the CF lung, another important source of elastase is present in the form of neutrophil elastase that is secreted during the inflammatory response. The combination of neutrophil elastase and the two elastases produced by P. aeruginosa leads to the degradation of elastin and collagen, both important for maintaining tissue plasticity, ultimately resulting in pulmonary fibrosis (Voynow et al., 2008).

1.2.3.3 Rhamnolipids

Rhamnolipids are glycolipidic biosurfactants that consist of one or two rhamnose moieties linked to each other via an α-1,2-glycosidic bond and one to three β-hydroxy fatty acid chains

(ranging in length from C8 to C16) linked to each other via an ester bond. The rhamnose moieties are linked to the fatty acid moieties via an O-glycosidic bond (Figure 1.13) (Abdel- Mawgoud et al., 2010).

Figure 1.13. Structure of α-L-rhamnopyranosyl-α-L-rhamnopyranosyl-β-hydroxyde-canoyl-β-hydroxydecanoate (Rha-Rha-C10-C10) (Abdel-Mawgoud et al., 2010).

27 Literature overview

Many variations exist in the rhamnolipid structure, depending on the number of rhamnose moieties, the number and chain length of the fatty acid moieties, the number of unsaturated bonds in the fatty acid chain, and the nature of the side chains on the rhamnose and fatty acid moieties. These possible combinations have led to the identification of more than 60 different rhamnolipid structures (Abdel-Mawgoud et al., 2010). Di-rhamno-di-lipid congeners appear to be the most abundantly produced in a batch culture of P. aeruginosa (67%), compared to mono-rhamno-di-lipid (22%), di-rhamno-mono-lipid (9%), and mono-rhamno-mono-lipid congeners (<3%). Rhamnolipids are involved in the swarming motility of P. aeruginosa since this type of motility involves the migration of bacterial cells across a surface. The rhamnolipids function as wetting agents, reducing the surface tension surrounding the bacterial colony, hence facilitating migration of P. aeruginosa cells across a particular surface. Indeed, it has been shown that mutation of the genes involved in rhamnolipid biosynthesis leads to aberrant swarming behavior of P. aeruginosa (Caiazza et al., 2005). Furthermore, rhamnolipids are involved in the dispersion of cells from P. aeruginosa biofilms (Boles et al., 2005) and in the maintenance of open water channels in mature biofilms (Davey et al., 2003). Finally, it was found that rhamnolipids are able to lyse PMNs and surround P. aeruginosa biofilms as a protective mechanism against approaching PMNs (Alhede et al., 2009).

1.2.3.4 Pseudomonas quinolone signal

Next to its role as a QS signal molecule, PQS is able to bind Fe(III). In fact, it has been shown that PQS functions as an iron trap, being closely associated with the bacterial cell, thus facilitating iron-uptake by the P. aeruginosa siderophores pyoverdine and pyochelin (Diggle et al., 2007). Recently, it has also been shown that this molecule is involved in cross-talk between P. aeruginosa and eukaryotic cells since it stimulates chemotaxis of PMNs (Hansch et al., 2014).

1.2.3.5 Hydrogen cyanide

Hydrogen Cyanide (HCN) is particularly produced by P. aeruginosa under microaerophilic growth conditions when the bacterium reaches high cell densities. In P. aeruginosa, three genes (hcnABC) are involved in the production of HCN, each encoding a subunit of a

28 Chapter 1 flavoenzyme (the HCN synthase) associated with the cytoplasmic membrane. More specifically, the HCN synthase generates HCN via oxidative decarboxylation of glycine (Blumer & Haas, 2000; Lenney & Gilchrist, 2011). HCN is extremely toxic since it interferes with aerobic respiration at the level of the electron transport chain by inhibiting cytochrome c oxidases. P. aeruginosa is resistant to the toxic activity of HCN by incorporating cyanide- insensitive terminal oxidases in its electron transport chain (Cunningham & Williams, 1995). HCN produced by P. aeruginosa has been shown to cause lethal toxicity in a Caenorhabditis elegans infection model (Gallagher & Manoil, 2001). Furthermore, high levels of HCN have been detected in sputum samples and the breath of CF patients. HCN is therefore considered to be a useful biomarker in detecting P. aeruginosa colonization of CF patients at an early stage as well as estimating P. aeruginosa population densities in the CF host during antibiotic treatment (Smith et al., 2013).

1.2.3.6 Exotoxin A

Exotoxin A is an adenosine diphosphate (ADP) ribosyltransferase (ADPRT) that inhibits protein synthesis in eukaryotes via ADP ribosylation of elongation factor 2 (elF2), consequently triggering cell death (Gellatly & Hancock, 2013; Pollack, 1983). In humans, it is known to bind to the low-density receptor-related protein (LRP), enabling its endocytosis (Kounnas et al., 1992). Similarly to the LasA and LasB proteases, Exotoxin A is secreted via the type II secretion system. The toxA gene is activated by the extracytoplasmic (ECF) sigma factor PvdS, which also controls pyoverdine production (Visca et al., 2007).

1.2.3.7 Phospholipase C

P. aeruginosa is able to produce at least three different phospholipases C (PLCs). One of these (PlcB) is a Zn-dependent PLC with a relatively unknown function. However, it has been shown that PlcB is involved in phospholipid chemotaxis (Barker et al., 2004). In addition, two larger, Zn-independent PLCs have been described, exhibiting hemolytic activity (PlcH) or not (PlcN) (Vasil et al., 1991). PlcH is very toxic to endothelial cells (at picomolar concentrations), and has been shown to inhibit angiogenesis (Vasil et al., 2009). It is proposed that P. aeruginosa PLCs are of major importance in the degradation of lung surfactant. Lung surfactant is composed of 90% lipids and 10% proteins and is crucial in reducing the surface

29 Literature overview tension in the alveoli. About 80% of the lipids present in lung surfactant consist of phosphatidylcholine (PC). This compound can be degraded by the action of PLCs (in combination with lipases) yielding phosphorylcholine, glycerol and fatty acids that can be further processed, serving as energy sources for the bacteria (Son et al., 2007). Degradation of the lung surfactant by PLCs can subsequently lead to an increase in surface tension in the alveoli, resulting in a collapse of the lung tissue (atelectasis) (Holm et al., 1991). All P. aeruginosa PLCs are secreted via the type II secretion system, but a difference exists in the transport across the inner membrane since PlcB is transported via the Sec translocon, while PlcH and PlcN are transported via the Tat translocon (Filloux, 2011).

1.2.3.8 Pyocyanin and other phenazines

Phenazines are pigmented, redox-active, nitrogen-containing heterocyclic molecules that are toxic to both eukaryotic and prokaryotic cells. Although the phenazine core structure is similar in different phenazines, the nature and position of the substituents on the heterocyclic ring mainly determines their color and biological properties (Price-Whelan et al., 2006). Two clusters of phenazine biosynthetic genes can be found in the P. aeruginosa genome, phz1 and phz2 (Mavrodi et al., 2001). These appear to be redundant for the major part since only small sequence variations are found in the phzA-G genes (Figure 1.14). However they differ in the fact that the phz1 cluster, but not phz2, contains the phzM and phzS genes. Another gene involved in phenazine biosynthesis (phzH) is located at a distant position (relative to both phz clusters) in the genome (Figure 1.14).

30 Chapter 1

Figure 1.14. Phenazine biosynthesis in P. aeruginosa. In general, the P. aeruginosa genome harbors two phenazine biosynthetic clusters that are located at different loci. The phzA-G genes (purple) are higly conserved between both gene clusters. The proteins encoded by these genes are required to convert chorismic acid to phenazine-1-carboxylic acid (PCA). In addition, the phz1 cluster contains the genes phzM (blue) and phzS (yellow) that are required for the conversion of PCA to 5-methylphenazine-1-carboxylic acid betaine (MPCAB) and 1-hydroxyphenazine (1-HP), respectively. MPCAB can be converted to pyocyanin via the action of PhzS. Finally, the phzH gene that is found at a different position in the genome than the phz gene clusters, encodes an enzyme required for the conversion of PCA to phenazine-1-carboxamide (PCN). Pyocyanin can act as a signaling molecule downstream of the PQS QS system and control the expression of various genes involved in drug efflux, redox processes and iron acquisition (Jimenez et al., 2012).

The phzA-G gene products are involved in the biosynthesis of phenazine-1-carboxylic acid (PCA) from chorismic acid. PCA can subsequently be converted to phenazine-1-carboxamide (PCN), 1-hydroxyphenazine (1-HP) and 5-methylphenazine-1-carboxylic acid betaine (MPCAB) by the action of the PhzH, PhzS and PhzM enzymes, respectively. In addition, MPCAB can be converted to pyocyanin via PhzS. Although it is believed that phenazines can contribute to virulence of P. aeruginosa in both acute and chronic infections, the phz1 and phz2 gene clusters have been found to differentially contribute to phenazine production during different modes of growth of P. aeruginosa (Recinos et al., 2012). More specifically, phz1 genes are more expressed during planktonic growth of P. aeruginosa, while genes in the phz2 cluster were almost exlusively contributing to phenazine biosynthesis in colony biofilms. Phenazines mediate their toxic activity to eukaryotes and other prokaryotes mainly by causing

31 Literature overview oxidative stress. Reduced phenazines are able to react with oxygen resulting in reactive oxygen species. Interestingly, it has been shown that PCA is able to reduce iron to its ferrous form, allowing the bacterium to form biofilms in the absence of ferric iron uptake systems (Wang et al., 2011). In addition, it was shown that at higher phenazine concentrations (>50 μM) the majority of iron in the CF lung is present in the ferrous form, indicating that P. aeruginosa can utilize phenazines to access alternative sources of iron in CF patients (Cornelis & Dingemans, 2013; Hunter et al., 2013). Recently, Briard and colleagues have revealed a role for 1-HP (but not PCA, PCN or pyocyanin) as a chelator, since this phenazine was able to bind to iron with a 2:1 stoichiometry (Briard et al., 2015). Finally, the blue- pigmented pyocyanin that is often overproduced by P. aeruginosa CF isolates (Fothergill et al., 2007) and exhibits toxic activity to several prokaryotic and eukaroytic organisms (Baron & Rowe, 1981; Lau et al., 2004), can also function as a signal molecule since it controls the expression of several genes involved in multridrug efflux, redox processes as well as iron acquisition (Dietrich et al., 2006).

1.2.3.9 P. aeruginosa secretion systems

In order to secrete proteins in the extracellular environment and to transfer toxins into eukaryotic as well as prokaryotic cells, P. aeruginosa uses five secretion systems (Filloux, 2011). The type IV secretion system is not used by this opportunistic pathogen. Protein secretion can occur via a two-step process (type II and type V secretion systems) or via a one- step process (type I, III, and VI secretion systems). In the case of a two-step mechanism, proteins first need to be transported to the periplasm via a general export pathway that is normally used to transport periplasmic and outer membrane proteins. The general export pathway can either be the Sec or Tat (Twin arginine translocation) pathway (Figure 1.15). Both pathways transport proteins that have a N-terminal signal peptide, but are selective based on the nature of the signal peptide. Proteins transported by the Tat pathway generally have signal peptides that are longer, contain a less hydrophobic middle region, and harbor tandem arginine residues in the N-terminal part (Filloux, 2011; Palmer & Berks, 2012) compared to the signal peptides of proteins transported by the Sec pathway. In addition, Sec-dependent substrates are translocated to the periplasm in an unfolded conformation, while Tat-dependent substrates are already folded when arriving in the periplasm. Upon cleavage of the signal peptide in the periplasm, the proteins are transferred outside the bacterial cell via a second transport complex.

32 Chapter 1

In case of the type II secretion system, many proteins are involved in the formation of the secretion complex. In P. aeruginosa, two type II secretion systems have been described, Xcp and Hxc, which are homologous (Figure 1.15). Xcp is constituted out of 12 different proteins, organized as a multiprotein complex within the cell envelope (Figure 1.15). More specifically, this complex consists of an inner membrane platform, a traffic ATPase, and an outer membrane pore (secretin) through which the exoproteins are secreted. Interestingly, five of the Xcp proteins are pseudopilins and constitute a pseudopilus. It is believed that extension and retraction of this pseudopilus expels the exoproteins through the secretin pore (Douzi et al., 2012).

Figure 1.15. Overview of the P. aeruginosa secretion systems (Filloux, 2011). In contrast to this complex system, type V secretion systems consist of a single polypeptide that contains multiple domains (Figure 1.15). The protein has a cleavable N-terminal signal peptide that confers its translocation to the periplasm in a Sec-dependent manner. Next, the C- terminal domain (translocator domain) is inserted in the outer membrane where it forms a β- barrel. Subsequently, the passenger (protease) domain travels through the channel formed by the β-barrel and is released from it via autocatalytic cleavage at the cell surface (Filloux, 2011). Originally, it was believed that type V secretion proteins were not assisted by other proteins, and hence were called autotransporters. However, recently it has been found that the β-barrel assembly complex (Bam), consisting of BamA-BamB-BamC-BamD-BamE, is

33 Literature overview required for the fast and efficient insertion of the translocator domain into the outer membrane (Costa et al., 2015; Leo et al., 2012; van Ulsen et al., 2014).

Type I secretion systems are tripartite systems consisting of an ABC protein, an inner membrane protein that largely protrudes into the periplasm and an outer membrane porin (Figure 1.15). An example of a type I secretion system in P. aeruginosa is the heme acquisition system (Has), in which the HasDEF system mediated secretion of the heme- binding protein HasAp. When bound to heme, HasAp binds to the cell surface receptor HasR. More complex secretion systems that require one step to transport proteins are the type III and type VI secretion systems.

The type III secretion system comprises a needle-like structure that injects toxins into the eukaryotic cell (Figure 1.15). In total, 36 genes encoded in five clustered operons and an additional six genes located at other loci on the chromosome are involved in type III secretion. Five of the 36 genes that comprise the large type III secretion system cluster are involved in regulation of this system, among which exsA is the master regulatory gene. Regulation of type III secretion via ExsA occurs via cAMP signaling and the Gac regulatory system (Hauser, 2009). In addition, type III secretion is triggered upon contact with host cells or during growth under low Ca2+ concentrations (Iglewski et al., 1978; Vallis et al., 1999). Five functional components constitute the type III secretion system. These are the needle complex, the translocation apparatus, regulatory proteins, the effector proteins (toxins), and the chaperones. The needle complex itself consists of a multi-ring base and a hollow needle- like filament through which the toxins are transported. In total, four different toxins (ExoS, ExoT, ExoU, and ExoY) have been shown to be secreted by the type III secretion system. These exotoxins cause direct damage to the host cells, are involved in modulation of the host immune response and can disrupt epithelial as well as endothelial barriers, allowing bacteria to enter the bloodstream (Hauser, 2009). Interestingly, P. aeruginosa strains either harbor exoS or exoU in their genomes, as it has been found that these genes are mutually exclusive (Feltman et al., 2001).

The type VI secretion system is another example of a complex system that results in the delivery of toxins into eukaryotic or prokaryotic cells. In fact, the type VI secretion system greatly resembles the contratile bacteriophage T4 tail that delivers DNA in the bacterial cell through puncturing of the outer membrane. Similarly, in type VI secretion, a tube formed of a

34 Chapter 1 polymer of the Hcp protein covered with a tip consisting of a trimer of the VgrG protein is enclosed by a sheath-like structure. Activation of the type VI secretion system results in contraction of the sheath-like structure and consequently puncturing of the (outer) cell membrane of the target cell by the VgrG tip (Figure 1.15). Subsequently, the effector toxins are injected in the target cell (Filloux, 2011; Kapitein & Mogk, 2013; Silverman et al., 2012). P. aeruginosa has three large gene clusters encoding type VI secretion systems: HSI (Hcp Secretion Island) I, HSI-II, and HSI-III (Chen et al., 2015; Filloux, 2011). In P. aeruginosa PA14, it has been shown that MfvR/PqsR and LasR negatively regulate HSI-I, while positively regulating HSI-II and HSI-III (Lesic et al., 2009).

1.2.3.10 Flagellum and type IV pili required for motility

P. aeruginosa can use several mechanisms to move across a surface or in liquids. The best known motility mechanisms are swimming, swarming, and twitching (Figure 1.16). P. aeruginosa has one polar flagellum that allows it to swim in liquid environments. Furthermore, P. aeruginosa is able to swarm using this flagellum. Swarming is the multicellular movement of P. aeruginosa across a surface.

Figure 1.16. Different types of motility exhibited by P. aeruginosa. Swarming and swimming rely on the use of the polar flagellum for P. aeruginosa to move across a surface or in a liquid environment, respectively. In contrast to swimming, swarming is a type of multicellular behavior. In order to swarm, P. aeruginosa utilizes rhamnolipids to lower the surface tension of the substrate. Twitching requires the production of type IV pili that extend, attach to the surface, and subsequently retract, resulting in movement of the bacteria (Kearns, 2010).

Movement across a surface imposes the problem of surface tension to the bacterium (Partridge & Harshey, 2013). However, during swarming, P. aeruginosa cells produce rhamnolipids that reduce the surface tension of the substrate and subsequently allow the

35 Literature overview bacteria to migrate via rotation of their helical flagellum. In addition, when moving across surfaces or swimming in viscous liquids, P. aeruginosa is able to use an alternative flagellar motor (more specifically the stator part of the motor) for movement under these conditions (Doyle et al., 2004; Toutain et al., 2005). A number of studies have even demonstrated that P. aeruginosa may produce two polar flagella during swarming across a surface (Kohler et al., 2000; Rashid & Kornberg, 2000). Another type of movement across surfaces exhibited by P. aeruginosa is twitching, mediated via type IV pili. In this type of motility, P. aeruginosa produces type IV pili (long fimbrial structures) that extrude through the outer membrane and attach to a surface, before retracting and bringing the cell closer to the site of attachment (Kearns, 2010). Next to their role in motility, type IV pili are also involved in adhesion, aggregation, and host cell invasion (Filloux, 2011).

1.2.3.11 Pyoverdine and pyochelin involved in iron uptake

Most of the environments from which P. aeruginosa can be retrieved are aerobic and consequently iron is particularly present in its oxidized form (Fe3+). However, oxidized iron is insoluble and therefore the bioavailablity of this metal to the bacteria is low. In order to circumvent this problem, P. aeruginosa has evolved elegant iron acquisition systems based on siderophores. Siderophores (“iron carriers”) are low-molecular weight molecules that chelate Fe3+ with a high affinity and are taken up by specific receptors. P. aeruginosa produces two siderophores, which differ in their affinity for iron. Pyoverdine has a high affinity for Fe3+ (1024 M-1 to 1027 M-1), while pyochelin has a rather low affinity for this metal (2.5 x 105 M-1) (Cox et al., 1981; Dumas et al., 2013; Meyer, 2000; Ravel & Cornelis, 2003; Visca et al., 2007). In fact, the low affinity siderophore pyochelin allows P. aeruginosa to acquire iron in conditions where the Fe3+ concentration is not extremely low (Cox et al., 1981; Dumas et al., 2013). Furthermore, it has been shown that pyochelin is produced prior to pyoverdine biosynthesis, which is triggered when the available iron concentration is too low (Dumas et al., 2013). P. aeruginosa strains produce one of three different types of pyoverdine (type I, II, or III) that, when bound to iron, are taken up by the bacterial cell via a specific ferri- siderophore receptor (FpvAI, II, or III) present at the outer membrane (Cornelis et al., 1989; de Chial et al., 2003; Meyer et al., 1997). In addition, P. aeruginosa strains usually have an alternative type I pyoverdine receptor, called FpvB that allows type II and III pyoverdine- producing strains to steal pyoverdine from type I producers (de Chial et al., 2003; Ghysels et al., 2004).

36 Chapter 1

Pyoverdine is composed of a dihydroxyquinoline chromophore that gives pyoverdine its characteristic green-yellow color and fluorescence under UV-light, a variable peptide chain that is linked by an amide group to the C1 carboxyl group of the chromophore, and an acyl side chain (either dicarboxylic acid or amide) bound to the aminogroup group of the chomophore (Figure 1.17A-C). The chromophore part, that contains a catechol group involved in Fe3+ chelation, is conserved among all known pyoverdine structures (Cornelis & Dingemans, 2013; Ravel & Cornelis, 2003; Visca et al., 2007). In addition to the catecholate group of the chromophore, the peptide chain contains two other hydroxamate iron-binding sites allowing Fe3+ to be bound with a high affinity. In the case of pyochelin, the phenolate and carboxylate groups both contribute to the chelation of Fe3+ (Figure 1.17D) (Mislin et al., 2006).

Figure 1.17. Structure of the pyoverdine and pyochelin siderophores. A. Structure of type I pyoverdine. B. Structure of type II pyoverdine. C. Structure of type III pyoverdine. D. Structure of pyochelin. aThr, allo- threonine; cDab, tetrahydropyrimidine ring generated by condensation of Dab with the preceding amino acid; Chr, chromophore; cOHOrn, cyclic-N5-hydroxyornithine; Dab, 2,4-diaminobutyrate; fOHOrn, N5-formyl-N5- hydroxyornithine.

37 Literature overview

Pyoverdine is synthesized by the action of large enzymes (the P. aeruginosa PAO1 PvdI protein comprises 5149 amino acids), called non-ribosomal peptide synthetases (NRPS) that are organized in modules. Each module catalyzes the incorporation of one amino acid into the growing peptide (Visca et al., 2007). All modules contain (1) a domain for the recognition and adenylation of the amino acid substrate, (2) a thiolation or peptidyl carrier domain to which a phosphopantetheine cofactor is covalently bound, and (3) a condensation domain that catalyzes the formation of the peptide bond via a nucleophilic attack by the next phosphopantetheine-bound amino acid (Figure 1.18A). In the module that incorporates the first amino acid, the condensation domain is not present (Figure 1.18A). PvdL is the first peptide synthetase in the process of pyoverdine synthesis. It is unique in that its starter module (PvdL-M1) contains an acyl coenzyme A ligase (AL) domain that couples coenzyme A with a fatty acid in an ATP-dependent reaction and subsequently transfers the acylated fatty acid to the L-Glu carried by the second module of PvdL (Mossialos et al., 2002). Next, a repetitive process takes place in which each module adenylates its cognate amino acid, followed by a covalent transfer to the phosphopantetheine factor, and finally peptide bond formation between the carboxyl group of the nascent peptide and the amino acid carried by the flanking module (Visca et al., 2007). In addition, some modules can contain accessory domains for epimerization or methylation of the amino acid substrates. Eventually, the thioester bond between the assembled peptide and the phosphopantetheine cofactor is hydrolyzed by the thioesterase domain of the last module (part of the PvdD NRPS) (Figure 1.18A). In total, four NRPS (PvdL, PvdI, PvdJ, and PvdD) are involved in pyoverdine biosynthesis and intermediates are being transferred from PvdL through PvdI and PvdJ to PvdD (Figure 1.18A). The resulting precursor peptide, acylated ferribactin, which is synthesized in the cytoplasm, does not contain a mature chromophore yet and needs to be transported to the periplasm via the PvdE ABC transporter to become functional (Hannauer et al., 2012; Visca et al., 2007). Once transported into the periplasm, ferribactin undergoes maturation, first by removal of the myristoyl group by PvdQ, and followed by maturation of the chromophore by PvdMNOP to yield mature pyoverdine that is exported out of the cell by the efflux pump PvdRT-OpmQ (Figure 1.18B, Figure 1.19). During the maturation process, the chromophore part that consists of the tripeptide L-Glu-D-Tyr-L-Dab, synthesized by PvdL, is modified (Visca et al., 2007). More precisely, L-Dab and D-Tyr are condensed to a tetrahydropyrimidine ring and finally modified to form the dihydroxyquinoline chromophore (Figure 1.18B).

38 Chapter 1

Figure 1.18. Pyoverdine biosynthetic pathway. A. Biosynthesis of ferribactin by non-ribosomal peptide synthetases (NRPS) containing modules consisting of adenylation domains (A), thiolation domains (T) and condensation domains (C). The acyl-CoA ligase domain (AL) from the initial module of PvdL is indicated in white. Auxiliary domains such as the epimerization domain (E) and the thioesterase domain (Te) are indicated by ovals. The predicted cytoplasmic membrane (CM) and periplasmic (PP) location of proteins (PvdE, PvdM, PvdN, PvdO, PvdP and PvdQ) which have roles in pyoverdine synthesis and/or export are shown. B. Proposed scheme for the maturation of the chromophore part. PVD, pyoverdine; ASA, aspartate β-semialdehyde; Dab, 2,4- diaminobutyrate; FA, fatty acid; fOHOrn, N5-formyl-N5-hydroxyornithine; OHOrn, N5-hydroxyornithine; OM, outer membrane; R1, acyl chain; R2, peptide chain (Visca et al., 2007).

39 Literature overview

Pyoverdine receptors belong to the family of TonB-dependent receptors since they require energy transmitted from the TonB-complex to open and transport the pyoverdine-iron complex (ferripyoverdine) into the periplasm (Cornelis, 2010). More specifically, the ferripyoverdine receptor has a β-barrel structure formed by 22 transmembrane β-strands, short periplasmic turns and large extracellular loops (Brillet et al., 2007; Wirth et al., 2007). In addition, the N-terminal part of the protein functions as a “cork” that plugs the barrel. The peptide chain of the pyoverdine molecules interacts with both the “cork” and extracellular loops upon binding, hence causing a conformational change in the plug. Furthermore, the N- terminal end of the receptor contains a domain called the TonB box that interacts with the inner membrane protein TonB, that together with ExbB and ExbD (constituting the TonB complex) transmits the energy of the proton motive force, allowing the porin to open and the ferripyoverdine to be translocated into the periplasm (Cornelis, 2010).

Fe3+

Pyoverdine

FpvA OpmQ

Chromophore maturation Dissociation? Recycling PvdONMP FpvC FpvF PvdR PvdQ acylase

TonB PvdT FpvE PvdE

FpvD

Acylated ferribactin 2+ Fe Acylated ferribactin

Figure 1.19. Mechanism of pyoverdine maturation and ferripyoverdine uptake. Acylated ferribactin precursor is transported to the periplasm by the PvdE ABC transporter and de-acylated by PvdQ. Further maturation of the chromophore is mediated by PvdONMP in the periplasm. FpvA is a β-barrel composed of 22 transmembrane β- strands that spans the outer membrane and has a N-terminal plug domain that blocks the gate in the absence of ferripyoverdine. Binding of ferripyoverdine causes a conformational change that triggers opening of the channel in a TonB-dependent way. Finally, the ferri-siderophore is transported into the periplasm, bound by two periplasmic binding proteins (FpvC and FpvF) which reduce iron to Fe2+, which is targeted for transport to the cytoplasm, mediated by an ABC transporter (FpvE and FpvD). In the periplasm the ferripyoverdine complex

40 Chapter 1

dissociates and pyoverdine is either recycled or degraded. Pyoverdine is exported via the PvdT-PvdR-OpmQ efflux system (Cornelis, 2014). From there, a periplasmic binding protein composed of FpvC and FpvF will bind the ferripyoverdine. Reduction to Fe2+ takes place in the periplasm and it is further transported inside the cytoplasm by the FpvE-FpvD ABC transporter. Finally, the iron is translocated into the cytoplasm using energy from ATP-hydrolysis and can be incorporated in the Fe-S clusters of various enzymes.

Iron uptake is strictly regulated by the ferric uptake regulator (Fur) that acts as a repressor (using Fe2+ as a co-repressor) assembling onto the promoter of iron uptake genes. Next to its role in iron uptake, pyoverdine also serves as a signal molecule. In the absence of ferripyoverdine, the TonB-dependent ferripyoverdine receptor FpvA is plugged. In addition, the anti-sigma factor FpvR sequesters the extracytoplasmic function (ECF) sigma factors FpvI and PvdS to the cytoplasmic membrane, resulting in low transcriptional activity of their target genes. When ferripyoverdine binds to the FpvA receptor, a conformational change will take place and a signal will be transmitted to the anti-sigma factor FpvR, resulting in activation of the ECF sigma factors FpvI and PvdS (Figure 1.20). It has been shown that FpvR is in fact degraded via the RseP/MucP protease (Draper et al., 2011). FvpI will guide the RNA polymerase to the fpvA promoter and initiate transcription. In contrast to this, activation of PvdS has a pleiotropic effect as it will initiate the transcription of genes involved in pyoverdine biosynthesis, exotoxin A production as well as synthesis of the PrpL endoprotease (Llamas et al., 2014).

The low affinity siderophore pyochelin also uses a TonB-dependent receptor (FptA) to transport the iron-siderophore complex into the bacterial cell. In contrast to pyoverdine, that is not supposed to enter the cytoplasm after uptake, pyochelin is transported into the cytoplasm by the inner-membrane permease FptX, after which the ferripyochelin complex dissociates and iron is released (Lamont et al., 2009). By analogy with the role of pyoverdine as a signal molecule, accumulation of (ferri)pyochelin in the cell causes the regulator PchR to activate expression of the two pyochelin biosynthetic operons pchDCBA and pchEFGHI as well as the fptA gene encoding the ferripyochelin receptor (Michel et al., 2007).

41 Literature overview

Figure 1.20. Role of pyoverdine as a signal molecule. In the absence of ferripyoverdine, the FpvR anti-sigma factor sequesters the extracytoplasmic (ECF) sigma factors FpvI and PvdS, resulting in low transcription levels of their target genes. In addition, the conformation of FpvA does not allow interaction with FpvR. When ferripyoverdine bind to FpvA, this will lead to a conformational change that triggers proteolytic degradation of FpvR by the RseP/MucP protease, releasing FpvI and PvdS. Whereas FpvI only activates expression of FpvA, PvdS will control a larger repertoire of virulence factors including exotoxin A and the endoprotease PrpL, next to the entire pyoverdine biosynthetic cluster (Llamas et al., 2014).

1.2.3.12 Exopolysaccharides involved in biofilm formation

In order to persist in unfavorable environments, bacteria can produce extracellular polysaccharides (exopolysaccharides), which protect them against dessication, oxidative stress and host immune responses. Expolysaccharides are major constituents of biofilms, together with other components such as extracellular DNA, proteins, and (in case of P. aeruginosa) rhamnolipids (Mann & Wozniak, 2012). P. aeruginosa is able to produce three different types of exopolysaccharides: alginate, Psl, and Pel. Usually, P. aeruginosa produces only one of these exopolysaccharides at a particular time point. Whereas Pel biofilms are formed in static liquid cultures (pellicle formation at the air-liquid interface) (Coulon et al.,

42 Chapter 1

2010), P. aeruginosa will rather secrete Psl to colonize surfaces (Ma et al., 2006) and has been shown to overproduce alginate during growth under certain physical conditions such as low fluid shear (Crabbé et al., 2008; Crabbé et al., 2010) or as the result of mutations acquired during growth in the CF lung environment (Feliziani et al., 2010; Lyczak et al., 2002; Pedersen, 1992). Although the structure of Pel remains to be elucidated, the structures of alginate and Psl are well known. Alginate is synthesized as a linear homopolymer of D- mannuronic acid in the cytoplasm, transported to the periplasm and subsequently modified there by epimerases and acetyltransferases to yield a random polymer of D-mannuronic acids, interspersed by L-guluronic acid residues and acetylated at C2’ and/or C3’ positions (Figure 1.21A). Due to these modifications, alginate is a high molecular weight, negatively charged acidic non-repetitive polymer that is characterized by its slimy appearance on agar plates covered with mucoid colonies. The O-acetylation of alginate has been shown to protect P. aeruginosa against antibody-independent opsonic phagocytosis, enhances microcolony formation and improves biofilm adhesion (Pier et al., 2001). Furthermore, it has been shown that the epimerization of D-mannuronic acid to L-guluronic acid increases the viscosity of alginate (Gacesa, 1988). The alginate biosynthetic operon, also known as the algD operon, comprises 12 genes (Figure 1.22). In total, three genes of this operon (algIJF) are involved in the O-acteylation of alginate, while algG provides the epimerase. Recently, it has been shown that in addition to the algD promoter, two internal promoters are present in the algD operon, just upstream of the algG and algI genes (Paletta & Ohman, 2012). This indicates that the properties of alginate and hence of the biofilm matrix can be adjusted in multiple ways.

Interestingly, the alginate and pel (but not psl) biosynthetic clusters both contain a gene encoding a c-di-GMP binding protein that mediates c-di-GMP regulation of exopolysaccharide biosynthesis (Lee et al., 2007; Merighi et al., 2007) (Figure 1.22). In contrast to alginate, Psl is a polymer consisting of repetitive pentamer subunits comprising three D-mannose residues, one L-rhamnose and one D-glucose residue (Figure 1.21B). Unlike alginate that is not directly associated with the cell surface, Psl forms a helical distribution surrounding the cell surface and is suspected to play a role in cell-cell and cell-surface interactions during biofilm formation (Ma et al., 2009).

43 Literature overview

Figure 1.21. Structure of alginate (A) and Psl (B). Alginate is a polymer of D-mannuronic acids that are interspersed with L-guluronic acids. O-acetylation of D-mannuronic acid residues can occur at the C2’ and/or C3’ positions. Due to the O-acetylation and epimerization, alginate has a random structure. B. Psl is composed of repeating pentamers of D-mannose (3x), L-rhamnose (1x), and D-glucose (1x) (Franklin et al., 2011).

Figure 1.22. Gene clusters involved in the biosynthesis of Pel, alginate, and Psl (Franklin et al., 2011).

1.2.3.13 Regulation of virulence gene expression via small RNAs

Bacterial non-coding small RNAs (sRNAs) comprise an important class of post- transcriptional regulators that vary in size from 50-400 nucleotides. These sRNAs can bind to

44 Chapter 1 target mRNAs and affect their stability and/or the degree of their translation (Figure 1.23). Often the interaction between the sRNA and the target mRNA is mediated through the RNA chaperone Hfq, as it has been shown that this protein binds a large group of sRNAs (Moller et al., 2002; Sonnleitner et al., 2008; Sonnleitner et al., 2012). sRNAs can also selectively promote degradation of single cistrons within a polycistronic transcript (Balasubramanian & Vanderpool, 2013). Recently, it has been shown that a significant number of genes are transcribed from the reverse complementary strand, generating antisense RNA molecules (asRNAs) that can vary in length from 10-1000 nucleotides (Gomez-Lozano et al., 2014; Thomason & Storz, 2010). Furthermore, due to extensive sequence complementarity between these asRNAs and their target mRNA, these asRNAs do not require Hfq binding. asRNAs can affect the expression of their target gene in several ways. Apart from the mechanisms mentioned in Figure 1.23 that are also valid for conventional sRNAs, transcription of the asRNAs could result (1) in convergent transcription in which the transcription of one promoter is suppressed by the promoter on the opposite strand (2) binding of the asRNA to the target mRNA could result in premature transcription termination (Thomason & Storz, 2010).

Figure 1.23. Positive and negative post-transcriptional regulation by sRNAs. Positive regulation of mRNA transcript levels can occur when (A) a sRNA binds to a 5’ untranslated region (5’UTR) that normally blocks the ribosome binding site (RBS) due to a secondary structure in the RNA. Upon binding, a conformational change in the secondary structure can liberate the RBS, allowing translation to proceed. sRNAs can also negatively affect mRNA transcript levels by either (B) blocking the RBS, while simultaneously promoting mRNA degradation in a RNase E-dependent way, (C) only block the RBS or only mediate degradation of the target mRNA via RNase E (Balasubramanian & Vanderpool, 2013).

45 Literature overview

1.2.4 Antibiotic resistance mechanisms in P. aeruginosa

P. aeruginosa has a high intrinsic resistance to antibiotics due to multiple reasons (Figure 1.24). The first reason is the very low membrane permeability of this bacterium. It has been shown that the P. aeruginosa outer membrane is only 8% as permeable as the membrane of E. coli (Hancock & Brinkman, 2002). A second reason can be found in the fact that P. aeruginosa has an extensive repertoire of multidrug efflux pumps. In total, P. aeruginosa harbors 12 operons encoding efflux pumps from the Resistance-Nodulation-Division (RND) family that extrude many substrates including antibiotics, biocides, dyes, QS molecules, etc. (Figure 1.25) (Lister et al., 2009). Next to these two intrinsic resistance mechanisms, P. aeruginosa can also acquire adaptive resistance during long-term exposure to antibiotics. Long term-exposure to polymyxin antibiotics such as colistin results in the modification of the lipid A component of LPS by the acquisition of mutations in the pmrB gene that comprises the sensor kinase component of the PmrAB two-component system (Moskowitz et al., 2012). Subsequently the response regulator becomes activated and turns on transcription of the genes required for the biosynthesis and transfer of 4-amino-L-arabinose to the lipid A moiety of LPS.

Figure 1.24. Overview of the intrinsic and adaptive resistance mechanisms of P. aeruginosa. LPS, lipopolysaccharide. PG, peptidoglycan. Intrinsic resistance to antibiotics comprises the presence of multiple efflux pumps and low membrane permeability, while modification of LPS and conversion to mucoidy are adaptive responses to antibiotic pressure encountered by P. aeruginosa (Park et al., 2014).

46 Chapter 1

In addition, P. aeruginosa may acquire mutations in the genes encoding the large outer membrane porins. Some of these porins are targets for certain antibiotics to enter the bacterial cell. This is most probably the reason why in many P. aeruginosa CF isolates the oprD gene is mutated since it is known that the positively charged antibiotic imipenem diffuses through this porin (Huang & Hancock, 1993; Marvig et al., 2015). Finally, mutations can occur in genes that are involved in the regulation of alginate biosynthesis, leading to the overexpression of this viscous exopolysaccharide that can subsequently “shield” the bacteria from certain antibiotics (Figure 1.24). Figure 1.25. Genetic organization, structure, and function of RND efflux pumps. An efflux pump belonging to the superfamily of Resistance- Nodulation-Division (RND) efflux pumps typically consists of a cytoplasmic membrane protein (RND), a periplasmic membrane fusion protein (MFP), and an outer membrane factor (OMF). This tripartite system spans the entire cell wall and mediates the proton-driven transport of both lipophilic and amphiphilic drugs out of the bacterial cell. Although the classic RND efflux system consists of three proteins, 6 out of the 12 operons in P. aeruginosa PAO1 do not contain the OMF gene. (Lister et al., 2009).

1.2.5 Pyocin production by P. aeruginosa

Due to its ubiquity, P. aeruginosa encounters many different species in the various environments it colonizes. As mentioned earlier P. aeruginosa can utilize several proteases, phenazines and many other secondary metabolites to compete with other microorganisms. However, this bacterium can also participate in intra-specific competition. In order to compete with other strains, a particular P. aeruginosa can produce toxins called bacteriocins. Bacteriocins are narrow-spectrum antibacterial peptides or proteins that are produced under

47 Literature overview stress conditions, such as overcrowding and nutrient depletion, allowing the producing strains to protect themselves against closely related competitors. The bacteriocins produced by P. aeruginosa are called pyocins. Three major types of pyocins can be distinguished; these are the R-, F-, and S-type pyocins (Ghequire & De Mot, 2014; Michel-Briand & Baysse, 2002). More recently, a fourth type of pyocin, the M-type pyocin, has been identified in the exoU- containing genomic islands of P. aeruginosa strains (Barreteau et al., 2009). M-type pyocins are lipid II-degrading bacteriocins that share homology with colicin M (Ghequire & De Mot, 2014). The R- and F-type pyocins are also referred to as tailocins due their structural similarities with bacteriophage tails (Figure 1.26). However, R-type pyocins are rigid and contractile (Figure 1.26), while F-type pyocins are flexible but non-contractile. In contrast to these two types of pyocin, S-type pyocins are soluble and highly modular, consisting of a receptor binding, translocation, and killing domain. The killing domain has either DNase (Duport et al., 1995; Sano & Kageyama, 1993; Sano et al., 1993; Seo & Galloway, 1990), tRNase (Elfarash et al., 2012), rRNase or pore-forming (Elfarash et al., 2014) activity. In order to protect itself from the toxic activity of the pyocin, the producing cell harbors an immunity gene that is typically organized in an operon together with the pyocin gene (Michel- Briand & Baysse, 2002). Production of the immunity protein leads to neutralization of the enzymatic activity of the pyocin due to interaction of the killing domain with the immunity protein (Figure 1.27). Figure 1.26. Structure of pyocin R2 as an example of R-type pyocins (obtained by cryo-electron microscopy). A. Surface view of a three-dimensional montage reconstruction of the entire precontraction pyocin. Pyocin R2 can be subdivided in three major parts: the baseplate, a trunk, and a collar. Six tail fibers extend from the outer side of the baseplate and serve as an anchor to attach to the target cell, while the inner side of the baseplate is connected to a central spike protein. The precontraction trunk is composed of an inner tube surrounded by a sheath and has a helical structure. Pyocin R2 has a sixfold axial symmetry and is composed of discs (or annuli) each composed of six tube subunits and six sheath subunits. Upon contact, the sheath will contract and the spike will pierce the target cell membrane. B-E. Surface views of the segmentation patterns shown at the left of each image highlighting the trunk structure of pyocin R2 (Ge et al., 2015).

48 Chapter 1

Figure 1.27. Interaction of the DNase domain of pyocin AP41 with its cognate immunity protein as an example of the ultra-high-affinity interaction between pyocin killing domains and immunity proteins. The immunity protein of pyocin AP41 (cartoon representation) binds to the DNase domain of AP41 (represented as a blue surface) at an “exosite” adjacent to the active site, thereby blocking DNA substrate binding, but leaving the catalytic residues completely exposed. This rather unusual mode of action (it would be expected that the immunity protein binds directly to the catalytic site) is common for DNase pyocins/colicins. Residues that are conserved between pyocins and colicins are colored in yellow and are located in the DNA binding groove, while residues that are not-conserved between pyocins and colicins are located in the periphery of the DNA binding region and are colored in pink (Joshi et al., 2015).

All S-type pyocins studied so far bind to TonB-dependent receptors involved in iron- siderophore uptake (Baysse et al., 1999; Denayer et al., 2007; Elfarash et al., 2012; Elfarash et al., 2014). Therefore, pyocin killing activity is greatly enhanced under iron-limited conditions. Regulation of pyocin expression is mediated by the PrtN, PrtR and RecA proteins. Regulatory sequences called P-boxes (often composed of several repeats of a particular motif), are located upstream of the R-, F-, and S-type pyocins and serve as binding sites for the PrtN protein, which is an activator of pyocin expression (Matsui et al., 1993). Under normal conditions, expression of the prtN gene is repressed due to binding of the transcriptional repressor PrtR to the prtN promoter region. However, under stress conditions that induce DNA-damage, RecA (involved in the SOS response) will become active and cleave the PrtR repressor, resulting in activation of prtN and subsequently pyocin expression (Figure 1.28).

49 Literature overview

Figure 1.28. Regulation of pyocin expression. The R-, F-, and S-type pyocin genes are preceded by non-coding regulatory boxes (P-boxes, black) that serve as binding sites for the PrtN activator of pyocin gene expression. Under normal conditions (A), prtN expression is repressed by the PrtR transcriptional repressor, preventing expression of the pyocin genes. Under stress conditions (exposure to DNA-damaging agents), the RecA protein mediates cleavage of the PrtR repressor, leading to activation of prtN expression and ultimately pyocin production. The recA SOS-box is indicated in purple (Ghequire & De Mot, 2014).

1.2.6 The switch from planktonic to chronic lifestyle via c-di-GMP signaling

The signal for P. aeruginosa to switch from its planktonic lifestyle, that is associated with the expression of many of the acute virulence factors discussed in the previous sections of this chapter, to the biofilm lifestyle, that is associated with chronic CF infections, is provided by the second messenger cyclic diguanosine monophosphate (c- di-GMP) (Jimenez et al., 2012). This second messenger is produced in response to certain extracellular signals and stimulates biofilm formation, while repressing motility and acute virulence factor production (Verstraeten et al., 2008). Two types of enzymes regulate c-di-GMP levels in the bacterial cell: diguanylate cyclases (DGCs), that contain GGDEF domains, convert two molecules of GTP into one molecule of c-di-GMP, while phosphodiesterases (PDEs) that contain domains with an EAL (or HD-GYP) motif are involved in the degradation of c-di-GMP (Figure 1.29). Furthermore, several proteins

50 Chapter 1 contain both GGDEF and EAL domains and it is believed that these enzymes are involved in balancing the c-di-GMP intracellular levels (Hengge, 2009). A number of response regulators contain GGDEF and/or EAL/HD-GYP domains, linking two- component regulatory systems to the switch from a planktonic to a biofilm lifestyle of P. aeruginosa, associated with acute and chronic infections, respectively (Kulasakara et al., 2006).

Figure 1.29. c-di-GMP signaling and the switch from planktonic to biofilm growth. The activity of diguanylate cyclases (DGCs) and phosphodiesterases (PDEs) determines the level of c-di-GMP in the bacterial cell by converting GTP into c-di-GMP or by degrading c-di-GMP, respectively. High levels of GTP are associated with a planktonic lifestyle, while high levels of c-di-GMP are associated with a biofilm lifestyle (Jimenez et al., 2012).

1.3 P. aeruginosa in the cystic fibrosis lung

1.3.1 From the environment, over acute infection to a chronic lifestyle

It has been widely accepted that the majority of initial P. aeruginosa infections are caused by environmental strains that have no history in the clinical environment (Burns et al., 2001; Manos et al., 2013; Rau et al., 2010). Nevertheless, the initially acquired P. aeruginosa isolates can be outcompeted by more adapted strains. Indeed, many inter-patient transmissions have been documented in which an environmental-like strain was replaced by a highly adapted clone (Anthony et al., 2002; Jelsbak et al., 2007; Marvig et al., 2015; Scott & Pitt,

51 Literature overview

2004). However, once established in the CF lung, a P. aeruginosa strain is often present in this habitat for the entire life span of the patient (Figure 1.30).

Figure 1.30. Overview of the different stages of CF lung colonization and adaptation by P. aeruginosa. At a certain age (typically during the late teenage years) CF patients are initially infected by (in most cases) an environmental P. aeruginosa strain. This strain will colonize the lungs as well as the paranasal sinuses and genetically adapt to the CF lung environment that is characterized by strong inflammation and a high antibiotic pressure. The penetration of antibiotics and the host immune response are more pronounced in the lungs than in the paranasal sinuses, leading to the occasional eradication of the P. aeruginosa strain in the lungs, but not the paranasal sinuses which function as a bacterial reservoir for future lung colonization. After several cycles of eradication due to extensive antibiotic usage, followed by re-colonization, the bacteria will become highly adapted, reaching a fitness peak in the CF environment and chronically colonize the CF lungs (Folkesson et al., 2012).

Due to the extensive application of antibiotics, P. aeruginosa titers might drop under the detection limit of conventional culturing methods and it may seem that the bacterium is temporarily eradicated. Unfortunately, other locations in the CF airways, such as the paranasal sinuses may serve as reservoirs for P. aeruginosa (Folkesson et al., 2012; Hansen et al., 2012). Once the antibiotic treatment has been finished, these strains can re-colonize the lower airways and increase in numbers. Since this process is repeated for several rounds, P. aeruginosa strains adapt to the harsh conditions in the CF lungs that are created by the mainly pro-inflammatory immune response, as well as the antibiotic pressure (Folkesson et al., 2012).

52 Chapter 1

1.3.2 Genetic adaptation of P. aeruginosa to the cystic fibrosis lung

In order to adopt the biofilm lifestyle associated with chronic infections that characterizes persistent P. aeruginosa strains, this bacterium undergoes a number of mutations. In a recent study, Marvig and colleagues have tried to identify the genes that become mutated during the adaptation of environmental P. aeruginosa strains to the CF lung environment (Marvig et al., 2015). In order to do this, they compared the genome sequences of P. aeruginosa isolates that were longitudinally collected and belonged to the same clone type. The identification of genes that had acquired non-synonymous mutations during adaptation of several P. aeruginosa clone types, independently from each other, allowed them to generate a list of the genes that are mutated as a result of adaptation to the disease habitat and hence were called pathoadaptive genes. In total, 52 pathoadaptive genes were identified, most of them involved in antibiotic resistance, alginate biosynthesis, synthesis of type IV pili, pyoverdine, Pel exopolysaccharide, and type VI secretion system (Marvig et al., 2015). By comparison of the whole genome sequences, they were also able to infer the order of mutations in a particular pathway. They found, in agreement with other studies, that mucA is always the first gene to become mutated, rendering the bacterium mucoid, followed by the later mutation of the algU gene, abolishing mucoidy. This was also the case for RetS and the GacS/GascA and LadS pathways, in which the mutation of retS always preceded mutation of any of the other three genes. These results indicate that environmental P. aeruginosa strains undergo convergent evolution in the CF airways. Finally, Rau and colleagues have indicated that during adaptation, the P. aeruginosa genome becomes reduced due to the accumulation of several large deletions (up to 150 kb), most probably via illegitimate recombination (Rau et al., 2012). On the other hand, the acquisition of novel genomic content was rare. Various studies have described the accumulation of mutations that inactivate the expression of virulence factors associated with acute infection, while genes involved in alginate biofilm synthesis become overexpressed during long-term colonization of the CF lung environment (Bjarnsholt et al., 2010; Chang et al., 2007; Feliziani et al., 2010; Hauser et al., 2011; Rau et al., 2012) (Figure 1.31). Genes that are typically inactivated comprise mucA (leading to alginate overproduction), lasR (defective las QS system), genes of the type III secretion system (loss of type III secretion system-mediated toxicity), and genes involved in motility (loss of flagella and pili) (Figure 1.31).

53 Literature overview

Figure 1.31. Genetic adaptation of P. aeruginosa during colonization of the CF lung environment. A “naïve” environmental P. aeruginosa strain occasionally colonizes the CF lung environment consisting of respiratory epithelial cells covered by layers of dehydrated, viscous mucus. Upon contact with the host cells, this strain will produce multiple virulence factors (type III secretion system, pili, flagella, LPS), associated with acute infections. LPS and flagellin stimulate the TLRs 4 and 5, hence activating a signaling cascade that results in the production of pro-inflammatory cytokines. In addition, TLR9 can sense CpG DNA, further contributing to the pro-inflammatory response. The CFTR inhibitory factor (Cif) produced by P. aeruginosa can mediate ubiquitination of CFTR, conferring it to the lysosomal degradation pathway, thereby facilitating colonization of the CF airways. P. aeruginosa also produces pyocyanin that contributes to oxidative stress via the generation of reactive oxygen species (ROS), in combination with the ROS produced by the PMNs. The long-term exposure to ROS will result in the accumulation of mutations in particular P. aeruginosa genes. One of the first genes to become mutated is mucA, resulting in overproduction of the major biofilm matrix constituent, alginate. This alginate protects P. aeruginosa from PMNs as well as antibiotics. In the meantime, the bacterium has shut down the production of nearly all of its immunogenic, energy-consuming acute virulence factors and focuses on persistence in the CF environment, avoiding immune clearance. Ab, antibiotic; GSH, glutathione; NAC, N- acetylcysteine (Cohen & Prince, 2012).

The inactivation of genes associated with acute virulence is most probably due to the strong immunogenicity that they exhibit. Indeed, the flagella (flagellin) and LPS stimulate TLRs that participate in the signaling cascade leading to NF-κB activation and the resulting pro- inflammatory response (Cohen & Prince, 2012; Kawai & Akira, 2007). In addition, several

54 Chapter 1 components of the type III secretion system have been shown to elicit antibody responses (Corech et al., 2005; Moss et al., 2001). The persistent inflammation that occurs in the CF lung environment exposes P. aeruginosa to ROS that can cause DNA damage, hence facilitating the genetic adaptation of this pathogen to the CF lung. At a certain time point the bacterium may reach a fitness peak in the CF environment due to the accumulation of many pathoadaptive mutations (Figure 1.31).

1.3.3 Iron uptake by P. aeruginosa in the cystic fibrosis lung

Many iron sources are available to P. aeruginosa in the CF lung. It has been shown that mutation of the CFTR gene results in an increased release of iron by human airway epithelial cells (Moreau-Marquis et al., 2008). Furthermore, the constitutive inflammation present in the CF lung can lead to the disruption of lung tissue and eventually to micro-bleeds, releasing hemoglobin into the lung. Both P. aeruginosa elastase (LasB) as well as neutrophil elastase are able to degrade human hemoglobin, although in vivo, neutrophil elastase seems to be the predominant protease (Cosgrove et al., 2011). P. aeruginosa can utilize two pathways to transport heme into the cell, these are the Heme acquisition system (Has) and Pseudomonas heme uptake (Phu) system (Figure 1.32). The Has system relies on the secretion of a heme- binding protein (HasAp), that serves as a hemophore, guiding heme to the HasR receptor, while the PhuR receptor directly binds hemoproteins. In the periplasm, heme is bound by a periplasmic binding protein, which delivers it to an ABC transporter. Next, the chaperone PhuS directs heme to the heme oxygenase HemO which mediates its degradation to yield Fe2+, CO, and biliverdin (Cornelis & Dingemans, 2013). Although iron concentrations in CF sputum are predicted to be high (Moreau-Marquis et al., 2008), this metal can be sequestered by host proteins such as lactoferrin and transferrin (Figure 1.32). However, the high-affinity siderophore pyoverdine is able to displace Fe3+ from these proteins, whereas the affinity of pyochelin is too low to enable this (Cornelis & Dingemans, 2013). In the CF lung environment, P. aeruginosa encounters different oxygen gradients, from aerobic, to microaerophilic, and even anaerobic regions (Kolpen et al., 2014; Quinn et al., 2014). It is likely that in the microaerophilic and/or anaerobic regions of the CF lung, iron is predominantly present in its reduced Fe2+ form.

55 Literature overview

Figure 1.32. Iron uptake systems that can be used by P. aeruginosa in the CF lung. The CF lung contains many sources of iron, both in its oxidized (Fe3+) and reduced (Fe2+) forms. Fe3+ is sequestered in the CF host by the host iron-binding proteins lactoferrin and transferrin. Although the low-affinity siderophore pyochelin is not efficient in displacing Fe3+ from these molecules, pyoverdine has a higher affinity for Fe3+ and will efficiently chelate these molecules, transporting Fe3+ into the bacterial cell via the FpvA receptor. Another iron source that is available to P. aeruginosa in the CF lung environment is heme, which is released from red blood cells that became lysed as a result of the inflammatory response. P. aeruginosa has two heme uptake systems: the Phu system that directly binds heme and the Has (Heme acquisition system) that relies on a secreted heme-binding protein that needs to bind to a specific receptor on the outer membrane. Finally, due to the microaerophilic and/or anaerobic conditions in the CF lung, Fe2+ might be more abundant in CF sputum than Fe3+. In addition, phenazines may contribute to the Fe2+ pool by reducing Fe3+ to Fe2+. P. aeruginosa can directly import Fe2+ via the FeoB transporter. Inside the bacterial cell, iron uptake systems are strictly regulated by the Ferric uptake regulator (Fur) that forms a complex with its co-repressor Fe2+ (Reid et al., 2009).

Furthermore, it has been shown that the phenazine PCA is able to reduce Fe3+ bound to host proteins to Fe2+ (Wang et al., 2011). In agreement with this finding, Fe2+ appears to be the predominant form of iron when CF sputum contains high phenazine levels (Hunter et al., 2013). Reduced iron uptake is energetically more efficient than pyoverdine uptake since it does not require siderophore production. Instead, Fe2+ diffuses through the outer membrane

56 Chapter 1 and is subsequently transported inside the cells via a system composed of three components: an inner membrane permease FeoB and two proteins, FeoA and FeoC (Figure 1.32) (Cornelis & Dingemans, 2013).

1.3.4 Biofilm formation by P. aeruginosa in the cystic fibrosis lung

One of the hallmarks of chronic CF infections caused by P. aeruginosa, is the formation of robust biofilms composed of the exopolysaccharide alginate. Alginate biosynthesis is normally regulated post-translationally through action of the MucA protease. MucA is an anti- sigma factor that sequesters the alternative sigma factor AlgU to the inner membrane, thereby preventing AlgR-mediated transcription of the algD operon (Figure 1.33).

Figure 1.33. Regulation of alginate production. A. Activation of the alginate biosynthetic operon is mediated via the alternative sigma factor AlgU that is normally sequestered to the inner membrane (IM) by the anti-sigma factor MucA. However, mutation of mucA will result in the liberation of AlgU that subsequently will activate the expression of the regulator of the algD operon, algR. In addition, the algC gene that is located at a distant position relative to the algD operon will also be activated. This will result in the constitutive overproduction of the exopolysaccharide alginate. B. Alginate biosynthesis is also regulated at the transcriptional level, as seven transcriptional regulators (AlgU, AlgR, IHF, AmrZ, Vfr, Hp-1/AlgP, CysB) bind to the algD promoter, causing DNA looping of the latter and enabling binding and transcription initiation by the RNA polymerase (Okkotsu et al., 2014).

57 Literature overview

However, since mutation of mucA is one of the first genetic adaptations of P. aeruginosa to the CF lung environment (Folkesson et al., 2012; Marvig et al., 2015), the algD operon becomes constitutively expressed, leading to alginate overproduction, causing the characteristic slimy appearance of mucoid colonies on an agar plate. Transcriptional regulation of alginate biosynthesis requires the binding of seven transcriptional regulators to the promoter region of algD, creating a DNA loop for the recruitment of RNA polymerase, thereby activating transcription.

1.3.5 Growth of P. aeruginosa under microaerophilic/anaerobic conditions

Although being classified as an aerobic microorganism, P. aeruginosa is able to grow under microaerobic and even anaerobic conditions (Hassett et al., 2002; Schobert & Jahn, 2010). In order to grow under anaerobic conditions P. aeruginosa can utilize nitrate or nitrite as terminal electron acceptors in the electron transport chain. If oxygen, nitrate and nitrite are not present, P. aeruginosa can grow or survive by fermenting arginine or pyruvate, respectively (Schobert & Jahn, 2010). Respiration of nitrate (denitrification) requires four reductases: the nitrate, nitrite, nitric oxide, and nitrous oxide reductases that catalyze the reduction of nitrate to dinitrogen (Figure 1.34). The process starts with the global oxygen-sensing regulator Anr that will form homodimers in the absence of oxygen. Subsequently, the active dimer binds to a recognition sequence called the Anr box that results in transcriptional activation of the target operon. Two other transcriptional regulators are needed in the denitrification process, the NarX-NarL two-component system and the Dnr (Dissimilative nitrogen respiration) regulator. NarX is a sensor kinase that detects nitrate and activates the response regulator NarL, resulting in transcription of the nar operon as well as dnr and the nirQOP operon. Dnr wil activate transcription of the nir, nor, and nos operons, in addition to nirQOP. NirQ is required for the post-translational maturation of NirS (Figure 1.34).

There is strong evidence that P. aeruginosa respires anaerobically in CF sputum via denitrification. In one study, the authors measured nitrous oxide (N2O) and nitric oxide (NO) production in freshly expectorated CF sputum samples from CF patients that were chronically infected with P. aeruginosa (Kolpen et al., 2014). They found that an initial increase in nitrous oxide (N2O) was followed by a decrease after six hours of monitoring in freshly − expectorated CF sputum. Furthermore, they noticed a decrease in the concentration of NO3 in sputum during 24 hours of incubation. In a second study, it was found that the addition of

58 Chapter 1 nitrate to LB medium, yielding physiological nitrate levels, resulted in increased growth rates of P. aeruginosa PAO1 as well as clinical P. aeruginosa CF isolates under anoxic conditions, comparable to those observed in CF lungs and sputum (Line et al., 2014)

Figure 1.34. Overview of the P. aeruginosa denitrification system. The complete conversion of nitrate into N2 requires the action of the nitrate reductase NarGHI, the nitrite reductase NirS, the nitric oxide reductase NorCB, and finally the nitrous oxide reductase NosZ. The genes encoding these reductases are part of four different operons (nar, nir, nor, and nos). An additional operon (nirQOP) encodes the NirQ protein that is involved in the post-translational maturation of NirS. Under oxygen-limited conditions, the global oxygen-sensing regulator Anr will activate the expression of both the narXL and dnr genes. Subsequently, NarX, which is the sensor kinase of - the NarX/NarL two-component system, will activate the nar operon in the presence of nitrate (NO3 ). The dissimilative nitrate respiration regulator Dnr, that senses nitric oxide (NO), will activate all four operons that encode the reductases (Schobert & Jahn, 2010).

59 Literature overview

1.3.6 Effect of antibiotic treatment on P. aeruginosa in the cystic fibrosis lung

CF patients receive many antibiotics, starting at a young age and continuing the treatment during their whole life span. Adopting a prophylactic antibiotic treatment strategy may delay the onset of chronic P. aeruginosa colonization. Table 1.4 provides an overview of the antibiotics that are generally used by CF patients to treat P. aeruginosa infections. The antibiotic treatment strategy largely depends on the policy of the CF reference center/hospital. Interestingly, a comparative study between the CF reference centers in Copenhagen and Hanover, demonstrated that the spread of dominant clones in the CF population is greatly dependent on the antibiotic treatment strategy and the applied infection control measures (Cramer et al., 2012). It was concluded that the frequent switch between antibiotics and the less conservative prescription thereof, in combination with only partial segregation of P. aeruginosa-positive and P. aeruginosa-negative patients, may have led to the replacement of initially acquired clones by P. aeruginosa clones with a superior fitness in the antibiotic-rich CF lung environment

Table 1.4. Overview of the antibiotics that are used to treat P. aeruginosa infections in CF patients.

Antibiotic type Antibiotic Primary target Species range Fluoroquinolones Aerobic Gram-positive and Gram- Topoisomerase II negative species, some anaerobic DNA synthesis inhibitor Ciprofloxacin (DNA gyrase), Gram-positive species (C. topoisomerase IV perfringens) and M. tuberculosis β-lactams Cell wall synthesis Carbapenems Penicillin-binding Aerobic and anaerobic Gram-positive inhibitors (imipenem, meropenem) proteins and Gram-negative species Lipopeptides Cell wall synthesis Colistin Cell membrane Gram-negative species inhibitors Aminoglycosides Protein synthesis Aerobic Gram-positive and Gram- Tobramycin 30S ribosome inhibitors negative species, and M. tuberculosis Macrolides Protein synthesis Aerobic and anaerobic Gram-positive Azithromycin 50S ribosome inhibitors and Gram-negative species C. perfringens, Clostridium perfringens. M. tuberculosis, Mycobacterium tuberculosis.

60 Chapter 1

1.4 Treatment of cystic fibrosis patients

1.4.1 Lung transplantation

The chronic inflammation as a consequence of the continuous interaction between the CF pathogens and the host immune system can lead to a deterioration of the clinical status of the patient. When the pulmonary function is drastically restricted (a forced expiratory volume in one second (FEV1) <30%) and no conservative treatment options are available anymore, lung transplantation may be considered by the affected CF patient. Apart from rejection, many complications can occur, including infections that are the main cause of death in the first year after transplantation (Hartert et al., 2014).

1.4.2 Gene therapy

Gene therapy is defined as the addition of a functional CFTR gene into the airway epithelial cells. In addition, gene therapy is often referred to as mutation-independent treatment of cystic fibrosis, since virtually every mutation can be treated in this way compared to the use of correctors or activators that only help specific groups of patients that have a particular mutation in their CFTR gene. Despite several attempts in testing the in vivo application of drugs based on gene therapy, no delivery strategy was successful so far. This is due to the complexity of the route that a vector has to follow. It has to cross several barriers including the viscous mucus, cilia beating, and cross the nuclear membrane. In the past, 26 clinical trials involving 450 patients have been carried out, but were not designed to study the clinical benefit. Instead, these trials were rather designed to assess safety and deliver a proof-of- concept. One clinical trial that was designed to improve lung function was based on adeno- associated viral vectors, but did not prove to be efficient in transducing the airway epithelial cells through the apical membrane (Lee & Southern, 2012).

1.4.3 CFTR correctors and potentiators

In contrast to gene therapy, CFTR correctors and potentiators have proven successful in restoring expression and activity of CFTR at the apical membrane of the airway epithelial cells. CFTR correctors rescue type II mutations such as ΔF508 by improving

61 Literature overview folding of the mutated CFTR in the endoplasmic reticulum (ER) (Figure 1.35). This will delay turnover at the plasma membrane. However, ultimately the CFTR is targeted for proteasomal degradation via ubiquitination and lysosomal transfer.

Figure 1.35. Role of aminoglycosides, effectors and potentiators in restoring CFTR function. A. Class I mutations (e.g. G542X) cause preliminary termination codons (PTC) resulting in truncated and non-functional CFTR. Aminoglycosides and the novel potent drug Ataluren can rescue these mutations by causing ribosomal readthrough of the PTC, hence restoring CFTR expression at the apical membrane of the epithelial cells. B. In the case of the ΔF508 mutation (and other class II mutations), the protein is incorrectly folded, recognized by the quality control (QC) system of the endoplasmic reticulum (ER) and conferred to ubiquitin-mediated proteasomal degradation. Correctors (e.g. Vx-809) can partially rescue the processing of the mutated CFTR by improving folding in the ER and delaying turn over at the plasma membrane (PM). However, the QC system at the plasma membrane will ultimately detect these incorrect CFTRs and they are targeted for proteasomal degradation via ubiquitination and lysosomal transfer. C. Class III mutations (e.g. G551D) do not impair CFTR expression at the cell membrane, but instead interfere with the proper gating activity of the channel. CFTR potentiators, such as the FDA-approved Ivacaftor can restore the CFTR activity. MVB, multivesicular body (Okiyoneda & Lukacs, 2012).

62 Chapter 1

Currently, the corrector Vx809 has been studied with great interest since it is able to treat ca. 85% of the CF patients. Furthermore, it increased the chloride transport of primary airway cells that carry the ΔF508 mutation up to 14% of the wild-type levels. In contrast to class II mutations, class III mutations lead to the correct folding and expression of CFTR at the apical membrane, but instead cause a gating defect. This type of mutation can be rescued by potentiators. One of these potentiators is Vx770 that restored chloride transport in primary airway cells carrying the G551D variant to ca. 50% of wild-type levels. In addition, Vx770 (potentiator) has a synergystic effect on the function of Vx809 (corrector). Finally, class I mutations that introduce premature termination codons can be rescued by an aminoglycoside or novel drugs that cause ribosomal readthrough, resulting in the expression of functional CFTR at the plasma membrane (Figure 1.35).

1.4.4 Intrapulmonary Percussive Ventilation

Intrapulmonary Percussive Ventilation (IPV) is a ventilation technique that introduces bursts of gas at high frequencies (100-400 bursts per minute) in the airways of patients via a mouthpiece (Lucangelo et al., 2003). Previously, a comparison was made between autogenic drainage (chest physiotherapy) preceded with IPV with saline and autogenic drainage preceded by wet inhalation of saline. No significant differences were found in oxygen saturation or the recovery of sputum wet weight, indicating that IPV did not enhance the removal of abnormal viscid bronchial secretions in cystic fibrosis patients (Van Ginderdeuren et al., 2008). Nevertheless, it is believed that the CF lung is characterized by low levels of fluid shear (Blake, 1973; Knowles & Boucher, 2002). Furthermore, it has been shown that under low fluid shear conditions, P. aeruginosa PAO1 produces biofilms in suspension, composed of the exopolysaccharide alginate (Crabbé et al., 2008; Crabbé et al., 2010), which is not observed under high fluid shear conditions. A possible way to increase the level of fluid shear in the CF airways could be the introduction of bursts of gas at high frequencies via the IPV system. This could result in the disruption of biofilm formation by P. aeruginosa, hence increasing the susceptibility of this bacterium to antibiotics and immune clearance.

63

64 Aim of the thesis

2 Aim of the thesis

Due to the availability of novel drugs as well as the progress in airway clearance techniques, the life expectancy of CF patients has been greatly increased. Unfortunately, the majority of CF patients that have achieved adolescence or young adulthood are colonized by P. aeruginosa strains, most often originating from the environment. Once the airways of a CF patient are chronically colonized by this bacterium, it is nearly impossible to eradicate it by means of antibiotic therapy due to the formation of protective alginate matrix biofilms. Therefore, it is a great challenge to modulate the behavior of this opportunistic pathogen in the CF lung environment, causing it to transition from a biofilm to a planktonic lifestyle. Since planktonic bacteria are more susceptible to antibiotics as well as the immune system, this could lead to the temporal eradication of P. aeruginosa or at least to a reduction in the bacterial load, hence improving the life quality of the CF patient. A possible way to modulate the behavior of P. aeruginosa is to introduce shear stress in the lungs of the CF patient.

The ultimate aim of this thesis was to study the effect of shear stress on biofilm formation by P. aeruginosa in the CF lung. In order to reach this, several preparatory steps have been followed:

In Chapter 2, we have typed the P. aeruginosa CF population at the UZ Brussel by using a combination of two genotyping methods and identified a highly adapted, transmissible P. aeruginosa CF strain (P. aeruginosa CF_PA39).

In Chapter 3, using this highly adapted, transmissible strain, the effect of shear stress was determined in an in vitro model that creates a low fluid shear environment, comparable to the CF lung, in artificial sputum medium. By comparing the transcriptome of P. aeruginosa CF_PA39 grown under low fluid shear (characterized by the formation of robust biofilms) to that under high fluid shear (characterized by planktonic cells), several marker genes for the biofilm and planktonic lifestyles were obtained.

The expression of these marker genes was determined in Chapter 4, in which the effect of intrapulmonary percussive ventilation on the lung function and the behavior of P. aeruginosa was studied.

65 Aim of the thesis

Finally, in Chapter 5, the identification of a novel S-type pyocin was revealed by screening of the genome sequence of P. aeruginosa CF_PA39, in an attempt to discover new antimicrobial molecules that can serve as an alternative to antibiotics.

66 Chapter 2

Chapter 2: The deletion of TonB-dependent receptor genes is part of the genome reduction process that occurs during adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung

This chapter is based on the following publication:

Dingemans, J., Ye, L., Hildebrand, F., Tontodonati, F., Craggs, M., Bilocq, F., De Vos, D., Crabbé, A., Van Houdt, R., Malfroot, A. & Cornelis, P. (2014). The deletion of TonB- dependent receptor genes is part of the genome reduction process that occurs during adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung. Pathog Dis 71, 26-38.

67 Deletion of TonB-dependent receptor genes

68 Chapter 2

2.1 Introduction

Due to the broad diversity of its metabolic and pathogenic pathways, Pseudomonas aeruginosa (P. aeruginosa) is able to colonize a variety of environments, ranging from water, soil and plants to burn-wound patients, immunocompromised patients and cystic fibrosis (CF) patients. Since most of the environments that P. aeruginosa colonizes are aerobic, with an abundance of Fe3+ over Fe2+, this bacterium needs to produce siderophores in order to acquire the extremely low soluble oxidized form of this metal. P. aeruginosa produces two different siderophores that can be distinguished based on their affinity for Fe3+, pyoverdine and pyochelin. The high affinity of pyoverdine for Fe3+ (1024 M-1 to 1027 M-1) allows P. aeruginosa to efficiently displace iron from host metal-binding extracellular proteins such as transferrin and lactoferrin, both having a lower affinity for Fe3+ (~1020 M-1) (Meyer, 2000; Ravel & Cornelis, 2003; Visca et al., 2007). On the other hand, the low affinity siderophore pyochelin (2.5 x 105 M-1) allows P. aeruginosa to acquire iron in conditions where the Fe3+ concentration is not extremely low (Cox et al., 1981; Dumas et al., 2013). Recently, it has been shown that pyochelin is in fact produced prior to pyoverdine (Dumas et al., 2013). In general, every P. aeruginosa isolate is able to produce one out of three different types of pyoverdine (Type I, II, and III) (Cornelis et al., 1989; Meyer et al., 1997). The difference between these three types of pyoverdine lies in the peptide chain, while the chromophore part of the molecule is conserved (Meyer, 2000; Mossialos et al., 2002; Ravel & Cornelis, 2003; Visca et al., 2007). In order to take up the Fe3+-bound pyoverdine, referred to as ferripyoverdine, P. aeruginosa uses a corresponding TonB-dependent ferripyoverdine receptor (FpvAI, FpvAII, and FpvAIII). Furthermore, P. aeruginosa can use the type I ferripyoverdine of other isolates via the alternative type I ferripyoverdine receptor FpvB, allowing type II and III pyoverdine-producing strains to use two types of ferripyoverdine (de Chial et al., 2003; Ghysels et al., 2004).

In addition, P. aeruginosa is able to colonize the microaerobic and/or anaerobic habitats that occur in the CF lung. The thick viscous mucus layer, altered metabolism of CF airway epithelial cells and an accumulation of multi-species biofilms in the respiratory tract of the CF patients are believed to be causative for the oxygen deprived CF lung mucus (Worlitzsch et al., 2002). In these environments, there is a higher ratio of Fe2+ versus Fe3+, correlating with an elevated production of phenazines by P. aeruginosa (Hunter et al., 2012; Hunter et al., 2013). Recently, it was shown that the phenazine pyocyanine and, to a higher extent, its

69 Deletion of TonB-dependent receptor genes precursor phenazine-1-carboxylic acid (PCA) are responsible for the reduction of Fe3+ to Fe2+, releasing the iron from host proteins (Cornelis & Dingemans, 2013; Hunter et al., 2013; Wang et al., 2011). In order to use the Fe2+ that is present in the CF sputum, P. aeruginosa can use the Feo system (Cornelis & Dingemans, 2013). Furthermore, it has been shown that in the CF lung, P. aeruginosa is not restricted to one form of iron, but both oxidation states of this metal play a role in its biofilm-forming capacity (Hunter et al., 2013). It has been observed that during the colonization of P. aeruginosa in the CF lung, pyoverdine-negative mutants can accumulate (De Vos et al., 2001; Lamont et al., 2009), either indicating that its production is too energetically costly or that there might be a selection pressure acting on this siderophore in these conditions. Like other bacteria, P. aeruginosa is able to produce bacteriocins, termed pyocins, which are able to kill close relatives of the same species (Michel-Briand & Baysse, 2002). Interestingly, several S-pyocins (soluble pyocins) target ferrisiderophore receptors to gain entry to the cells and kill them. This is the case for pyocins S2/S4 and S3, which use FpvAI and FpvAII as receptors, respectively (Baysse et al., 1999; Denayer et al., 2007; Elfarash et al., 2012) while the pore-forming pyocin S5 utilizes the FptA ferripyochelin receptor (Elfarash et al., 2014).

Recently, a number of P. aeruginosa CF genomes have been completely sequenced, yielding a total of 12 completed genome sequences. In the study of Rau et al. (Rau et al., 2012) it was shown that during the adaptation of P. aeruginosa to the CF lung, genome reduction, rather than gene acquisition, occurs in order to increase the fitness of this bacterium in its host environment.

In this study, we typed 54 P. aeruginosa CF isolates based on an innovative technique that combines the whole-genome typing method of Repetitive Sequence-Based PCR (Rep-PCR) with the more gene-specific typing method of multiplex PCR targeting ferripyoverdine receptors and S-pyocin genes. Using this typing method, we discovered that the fpvB gene can be deleted during the adaptation of P. aeruginosa to the CF lung. In order to detect deletions in other TonB-dependent receptor genes and/or iron uptake genes, we determined the genome sequence of one isolate of an epidemic P. aeruginosa CF strain. Next to the deletion of another TonB-dependent receptor gene, we found several other large deletions in the type III secretion system and the region corresponding to PA2171-PA2226 of the PAO1 genome. Finally, after performing a comparative analysis of the genome described in this study to the genomes of other P. aeruginosa CF isolates, we have found some general features that might

70 Chapter 2 contribute to the fitness of this bacterium in its CF host environment, such as the deletion of the fpvB gene.

2.2 Materials and methods

2.2.1 Strains

Non-clinical and clinical bacterial strains used in this study are listed in Table 2.1 and 2.2, respectively. Bacterial cultures were grown at 37°C in nutrient rich lysogeny broth (LB) medium (Life Technologies, Merelbeke, Belgium).

Table 2.1. Non-clinical strains used in this study

Strain Features Reference PAOI Wild type PAOI Stover et al., 2000 PAOIΔfpvA Chromosomal deletion mutant of fpvAI (PA2398) Ghysels, et al., 2004 PAOIΔfpvB Chromosomal deletion mutant of fpvB (PA4168) Ghysels, et al., 2004 PAOIΔfpvAfpvB Chromosomal deletion mutant of fpvAI and fpvB Ghysels, et al., 2004

2.2.2 Isolation of P. aeruginosa from CF sputum

Sputa were collected from a total of 22 CF patients (range: 10-49 years old) attending the CF Reference Center of the UZ Brussel (Belgium) in whom P. aeruginosa was previously detected over a time period of 20 months (Table 2.2). Sputum samples were diluted two times by adding an equal volume of PBS to reduce viscosity. Next, 100 µl of a serial dilution (10-1- 10-7) of this two-fold diluted sputum was plated on Pseudomonas agar P (Difco) to which Irgasan (1 mg/l) was added (Sigma-Aldrich). Finally, cultures were grown at 37°C for 48 hours.

2.2.3 Typing of P. aeruginosa CF isolates

2.2.3.1 Typing via Rep-PCR

DNA from a total of 54 P. aeruginosa isolates was extracted using the Ultra CleanTM Microbial DNA isolation Kit (MO-BIO Laboratories, CA, USA) starting from a single colony. Next, a Rep-PCR was carried out using the DiversiLab Pseudomonas Kit (BioMérieux, Marcy-l′Etoile, France). Amplification was performed in a VeritiTM 96 well Thermal Cycler detection system (Applied Biosystems) using the following thermocycling

71 Deletion of TonB-dependent receptor genes conditions: 94°C for 5 min, followed by 30 cycles of 94°C for 45 s, 55°C for 45 s, 72°C for 1 min and a final elongation step at 72°C for 10 min. For each reaction, a total volume of 25 μl was used including 18 μl of Rep-PCR MM1, 2.5 μl of GeneAmp 10x PCR buffer, 2 μl of primer mix, 0.5 μl of AmpliTaq DNA polymerase and 2 μl of template DNA. Finally, PCR products were separated via microfluidics electrophoresis in a Diversilab DNA LabChip (BioMérieux, Marcy-l′Etoile, France) using a 2100 Bioanalyzer (Agilent Technologies). Isolates sharing more than 95% identity were considered to be clonally related.

2.2.3.2 Typing via Multiplex PCR

Starting from an overnight culture of each of the P. aeruginosa isolates, DNA was extracted using the DNeasyTM Blood and Tissue kit (Qiagen) subsequent to an appropriate pretreatment for Gram-Negative Bacteria. A multiplex PCR assay was performed using primers targeting ferripyoverdine receptor genes. Primers were designed using the Primer3 software (Rozen & Skaletsky, 2000) and are listed in Table S2.1. Each 25 µl PCR reaction consisted of the following components: 14.75 µl of water, 5 µl of 5x Q solution (Qiagen), 2.5 µl of 10x Coral Load PCR buffer (Qiagen), 0.5 µl of 10 mM dNTP mixture (Thermo Scientific), 0.5 µl of forward and reverse primers (Sigma-Aldrich), 0.25 µl of Taq DNA polymerase (Qiagen), and finally 1 µl of genomic DNA. PCR amplification was performed using a TC-412 Thermocycler (Techne) with the following cycling parameters: 94°C for 5 min followed by 35 cycles of 94°C for 45 s, 55°C for 45 s, 72°C for 1 min, and a final extension at 72°C for 10 seconds. Four reference strains were used as a control: PAO1, W15 December 14, W15 December 8, and W15 August 2 from which the ferripyoverdine receptors have been previously described (Denayer et al., 2007; Pirnay et al., 2005). Multiplex PCR-amplified fragments were separated by gel electrophoresis on a 2% agarose gel and visualized under UV light subsequent to ethidium bromide staining. Similarly, a multiplex PCR typing assay using primers targeting S-pyocin genes was performed. Specific primers were designed for all previously characterized S-type pyocin genes (encoding pyocins S1, S2, S3, S4 and S5) using the Primer3 software (Rozen & Skaletsky, 2000) and are listed in Table S2.1. A separate PCR reaction using the thermocycling conditions mentioned above was performed in parallel for amplification of the pyocin AP41 gene (Table S2.1).

72 Chapter 2

2.2.4 Amplification and sequencing of fpvB region

Primers were designed targeting the genes upstream and downstream of the fpvB gene corresponding to the PAO1 genes PA4167 and PA4169, respectively (Table S2.2A). Amplification was performed in a TC-412 Thermocycler (Techne) using the following thermocycling conditions: 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 55°C for 1 min, 72°C for 7 min and a final elongation step at 72°C for 10 min. For each reaction, a total volume of 50 μl was used including 25 μl of nuclease-free water (Qiagen), 10 µl of 5x Q solution (Qiagen), 10 μl of 5x HiFi PCR buffer (Qiagen), 2 μl of primer mix, 2 μl of HiFi Taq DNA polymerase, and 1 μl of template DNA. Amplification of the fpvB region was verified by gel electrophoresis on a 2% agarose gel and visualized under UV light subsequent to ethidium bromide staining. Finally, PCR products were sequenced in double chain at the VIB genetic service facility (Wilrijk Belgium). The obtained sequences were screened and corrected for sequencing errors, assembled into contigs, and finally the resulting contigs were aligned to the corresponding region in the PAO1 genome in order to map deletions.

2.2.5 Testing of antibiotic susceptibility of fpvAI, fpvB, and fpvAIfpvB mutants versus wild type P. aeruginosa PAO1

In order to test the sensitivity of the wild type and mutant strains (Table 2.1) to several antibiotics frequently used to treat P. aeruginosa CF infections, a disc diffusion test was performed. More specifically, about 100 µl of an overnight culture was mixed with 5 ml of 6% (wt/vol) LB top agar and poured on an LB plate to form a top agar layer on which an antibiotic diffusion disc (Oxoid) was applied. The following antibiotics were tested: azithromycin (15 µg), ciprofloxacin (15 µg), colistin (10 µg), piperacillin (30 µg), tobramycin (10 µg), imipenem (10 µg), and meropenem (10 µg). After 24 hours of growth at 37°C, the diameter of the inhibition zone surrounding the diffusion disc was measured using a ruler.

2.2.6 Whole genome sequencing

Whole genome sequencing of P. aeruginosa CF_PA39 was done at the VIB nucleomics core using an Illumina Miseq sequencer. The library was constructed by using the Nextera kit, read length as 150 bp paired end. The genome was covered, on average, 62-fold. This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession JDVE00000000. The version described in this paper is version JDVE01000000.

73 Deletion of TonB-dependent receptor genes

2.2.7 Genome assembly

First the reads were quality filtered, retaining only reads that were > 50 bp, < 2 ambigous nucleotides, < 10 homonucleotides, > 20 average Quality using sdm (psbweb05.psb.ugent.be/lotus/). Furthermore, if the sequence was dropping below an average quality of 17 in a 15 bp window, the read was clipped 3’. Quality filtered sequences from the miSeq run were de novo assembled using Velvet version 1.2.08 (Zerbino & Birney, 2008). A range of parameter K was tested and based on the number of contigs, we chose a K of 97 with default options. The resulting contigs were scaffolded using SSPACE basic version 2.0 (Boetzer et al., 2011) using the options “-x 1 -p 1 -m 32 -t 5 -k 5 -a 0.70”. Further gaps within scaffolds we attempted to close using GapFiller (Boetzer & Pirovano, 2012) using the options “-m 75 -o 2 -r 0.7 -d 70 -t 10 -g 1”. The final genome assembly contained 100 contigs and scaffolds with a total length of 6,193,213 bp and an N50 of 113,134 bp. The whole genome sequence of P. aeruginosa CF_PA39 was deposited at GenBank as a total of 98 scaffolds and contigs (the two smallest scaffolds with a size <200 bp were excluded) under the accession JDVE00000000.

2.2.8 Genome annotation and analysis

Genome annotation was performed using RAST (Rapid Annotations using Subsystems Technology) (Aziz et al., 2008). In order to detect deleted regions, a comparative genome sequence analysis was performed via RAST using the genomes of P. aeruginosa PAO1, UCBPP-PA14, and LESB58. Genes that were conserved amongst these three reference genomes (> 95% shared protein sequence identity) but not in the draft genome of P. aeruginosa CFA39 (<70% protein sequence identity) were identified via the Sequence Based Comparison tool of RAST. Deleted regions (>1 kbp) were confirmed by manually analyzing the draft genome sequence via the bioinformatics software GeneiousTM (Biomatters Ltd.). Genomic islands were predicted using IslandViewer (Langille & Brinkman, 2009) and manually checked via sequence analysis in Geneious. In some cases, the start and stop positions of the genomic island were not correctly predicted and a manual correction was performed based on the characteristics of genomic islands as described by Juhas et al. (Juhas et al., 2009). In addition, regions in the CF_PA39 genome that were not present in the three reference genomes mentioned above were identified as genomic islands if significant hits were found against previously described genomic islands using NCBI BLASTn (Altschul et

74 Chapter 2 al., 1990) and/or the BLAST tool at the www.pseudomonas.com website (Winsor et al., 2011). Clustered regularly interspaced short palindromic repeats (CRISPR) regions were identified using CRISPRFinder (Grissa et al., 2007), while prophages were predicted via PHAST (PHAge Search Tool) (Zhou et al., 2011). A genome map of P. aeruginosa CF_PA39 was constructed using the CGView Server (Grant & Stothard, 2008).

2.2.9 Confirmation of genomic content via PCR

The presence of deleted regions, genomic islands, and CRISPR regions was confirmed via PCR amplification. Primers were designed using the Primer3 software (Rozen & Skaletsky, 2000) and are listed in Table S2.2A-B. Each 25 µl PCR reaction consisted of the following components: 14.75 µl of water, 5 µl of 5x Q solution (Qiagen), 2.5 µl of 10x Coral Load PCR buffer (Qiagen), 0.5 µl of 10 mM dNTP mixture (Thermo Scientific), 0.5 µl of forward and reverse primers (Sigma-Aldrich), 0.25 µl of Taq DNA polymerase (Qiagen), and finally 1 µl of genomic DNA. PCR amplification was performed using a TC-412 Thermocycler (Techne) with the following cycling parameters: 94°C for 5 min followed by 35 cycles of 94°C for 45 s, 55°C for 45 s, 72°C for 2 min, and a final extension at 72°C for 10 seconds.

2.2.10 Identification of deletions comprising TonB-dependent receptor genes amongst P. aeruginosa CF isolates of the DK2 lineage

Deleted regions in P. aeruginosa CF isolates of the DK2 lineage (Rau et al., 2012) were “in silico” screened for the presence of previously described TonB-dependent receptor genes using either the P. aeruginosa DK2 genome sequence (GenBank Accession No. CP003149) or the 71 contig files of the isolate CF510-2006 that are deposited in GenBank under the Accession No. AJHI00000000.

2.2.11 RNA extraction from CF sputum and bacterial cultures

Sputum samples were incubated with an equal volume of sputolysin reagent (Calbiochem, La Jolla, CA) and vortexed for 30s. Next, samples were incubated with DNase I (Roche) for 20 min with intermittent mixing to remove extracellular chromosomal DNA. Finally, 30 µl of proteinase K (Qiagen, Hilden Germany) was added per one ml of sputum sample and samples were incubated at RT with intermittent mixing to remove proteins. In order to lyse eukaryotic cells and remove soluble cellular debris, sputum samples were washed twice with ice-cold

75 Deletion of TonB-dependent receptor genes sterile double-distilled water. Total RNA was purified from the pellets using the RNeasy Mini Kit (Qiagen, Hilden, Germany). Prior to RNA purification, bacterial cells inside the sputum were first lysed by following the Enzymatic lysis of bacteria-protocol (Qiagen, Hilden Germany). The RNA was fixed by adding 1000 μl of RNAprotect Bacteria Reagent (Qiagen, Hilden Germany) to the bacterial culture in order to protect the RNA. Next, samples were vortexed for 5s and subsequently incubated at room temperature for 5 min. After a centrifugation step (5000 g, 10 min), the supernatant was decanted and 200 µl of TE-buffer (10 mM TrisCl, 1mM EDTA, pH 8.0) containing 3 mg/ml lysozyme was added to the pellets. Before adding 350 μl of RTL buffer, the mixture was incubated for at least 20 min with vortexing for 10 s every 2 min. Finally, 500 μl of ethanol was used to allow binding of the RNA on the column membrane. After this preparatory work, we proceeded to the purification step following the protocol for total RNA purification of bacterial lysates with the RNeasy mini kit (Qiagen, Hilden, Germany). The purified RNA was treated with DNase (Roche) in order to remove any contaminating DNA. To check whether the RNA extract contained no more residual DNA, a PCR was performed using primers (forward 5’- ATGAACAACGTTCTGAAATTCTCTGCT-‘3 and reverse 5’- CTTGCGGCTGGCTTTTTCCAG-’3) targeting the oprI gene in a Thermocycler (TC-412- Techne) with the following thermocycling conditions: 94°C for 5 min, followed by 38 cycles of 94°C for 45 seconds, 55°C for 45 seconds, 1 min at 72°C and a final elongation step at 72°C for 10 min. The RNA quantity was assessed via the Nanodrop ND-1000 spectrophotometer (Isogen LifeSciences, Wilmington, USA). The removal of genomic DNA was verified via 35 cycles of PCR amplification (5 min at 94°C, followed by 35 cycles of 45 s at 94°C, 45 seconds at 55°C, 1 min at 72°C, and a final extension step of 10 min at 72°C) of the oprI gene (Forward 5’-ATGAACAACGTTCTGAAATTCTCTGCT-‘3 and Reverse 5’- CTTGCGGCTGGCTTTTTCCAG-’3) of P. aeruginosa.

For RNA extraction from liquid bacterial cultures (P. aeruginosa PA6 grown in LB or CAA medium for 24 hours at 37°C in a shaking incubator at 200 rpm), exactly the same procedure was followed subsequent to the addition of RNAprotect Bacteria (Qiagen) to the liquid culture.

76 Chapter 2

2.2.12 cDNA synthesis

cDNA was synthesized starting from 200 ng of total RNA by means of the Protoscript M- MuLV First Strand cDNA Synthesis Kit (New England Biolabs). In order to verify efficient conversion of RNA to cDNA, a PCR was performed using primers (forward 5’- ATGAACAACGTTCTGAAATTCTCTGCT-3’ and reverse 5’- CTTGCGGCTGGCTTTTTCCAG-3’) that allow amplification of the oprI gene in a Thermocycler (TC-412-Techne) with the following thermocycling conditions: 94°C for 5 min, followed by 35 cycles at 94°C for 45 seconds, 55°C for 45 seconds, 1 min at 72°C and a final elongation step at 72°C for 10 min.

2.2.13 Real-time PCR

Real-time PCR was performed to determine the expression of genes involved in pyoverdine production (pvdA, pvdS) and uptake (fpvAI, fpvB), pyochelin uptake (fptA), ferrous iron uptake (feoA, feoB), and heme uptake (hasAp, phuT). oprI was used as the housekeeping gene. The real-time PCR was carried out in a 96 well plate within each well a 25 μl volume consisting of 9.5 μl nuclease-free water, 1µl (400 ng) of each forward and reverse primer (Table S2.1), 12.5 μl of the 2x iQTM SYBR® Green supermix (Biorad) and finally 1 μl of template cDNA (200ng/μl). The real-time PCR was performed in an iQ2 real-time PCR detection system (BioRad) with the following program: an initial cycle at 95°C for about 3 min for denaturation and enzyme activation, then 40 cycles of 95°C for 10 s and 55°C for 60 s to allow amplification. Finally, melt curves were determined to check for primer dimers. Primers were designed using the primer3 software (Rozen & Skaletsky, 2000) and are listed in Table S2.3 (supplementary information). Fold changes were calculated using the Livak method (Livak & Schmittgen, 2001). The experiment was at least performed in technical duplicates.

2.2.14 Statistical analysis

The disc diffusion test was performed in biological triplicate. In order to determine the statistical significance of differences between wild type and mutant strains, a two-tailed Student’s t-test was performed. P-values <0.05 were considered to be statistically significant.

77 Deletion of TonB-dependent receptor genes

2.3 Results

2.3.1 The P. aeruginosa CF population at the UZ Brussel is characterized by the absence of a dominant clone

In total, 54 P. aeruginosa isolates were collected from sputum samples of 22 different CF patients based on distinct colony morphology. In order to type these isolates and gather a better view of the P. aeruginosa population structure amongst CF patients attending the UZ Brussel, two different typing methods were adopted. In a first approach, the clonal relatedness of the isolates was determined using the DiversiLab method (Ratkai et al., 2010). Isolates sharing more than 95% similarity were considered to be clonally related (Figure S2.1). However, in order to increase the discriminatory power of this previous method, the clonally relatedness of these isolates was linked to the presence of ferripyoverdine and/or pyocin genes, tested via a multiplex PCR approach (Bodilis et al., 2009; Ghysels et al., 2004). In total, 18 different clones could be distinguished by combining both methods (Table 2.2). Although most clones were unique in the patient population, some clones were isolated from several patients (UZBC2, UZBC9, and UZBC15), probably indicating the transmission of these P. aeruginosa clones. Interestingly, the UZBC15 clone was retrieved from four different patients and persisted in the lungs of patient CF13 for a time period of at least nine years (Table 2.2). Furthermore, only one patient (CF7) was colonized by more than one clone. Together, these results indicate that although inter-patient transmission is likely to occur at the UZ Brussel CF reference center, no dominant P. aeruginosa CF clone was found to be widespread.

Table 2.2. Overview of the features of all P. aeruginosa isolates and clones that were identified in the patient population attending the CF reference center at UZ Brussel.

Clonea Patient Patient age Isolateb Year of isolation fpv receptor genes Pyocin genes UZBC1 CF19 27 PA29 2012 fpvA IIb S1, AP41 UZBC1 CF19 27 PA32 2012 fpvA IIb + fpvB S1 UZBC2 CF5 21 PA6 2011 fpvAI + fpvB S2, AP41 UZBC2 CF5 22 PA46 2012 fpvAI + fpvB S2, AP41 UZBC2 CF5 22 PA47 2012 fpvAI + fpvB S2, AP41 UZBC2 CF5 22 PA48 2012 fpvAI + fpvB S2, AP41 UZBC2 CF5 22 PA49 2012 fpvAI + fpvB S2, AP41 UZBC2 CF8 18 PA10 2011 fpvAI + fpvB S2, AP41 UZBC2 CF8 18 PA11 2011 fpvAI + fpvB S2, AP41 UZBC3 CF12 16 PA15 2011 fpvA IIb + fpvB - UZBC4 CF14 32 PA19 2011 fpvA III S2, S5 UZBC5 CF6 23 PA7 2011 fpvAI + fpvB S2

78 Chapter 2

Clonea Patient Patient age Isolateb Year of isolation fpv receptor genes Pyocin genes UZBC6 CF15 29 PA20 2011 fpvA III + fpvB S2 UZBC7 CF4 32 PA4 2011 fpvAI + fpvB S2 UZBC8 CF7 30 PA8 2011 fpvAI + fpvB S3, S4, S5 UZBC8 CF7 30 PA9 2011 fpvAI + fpvB S3, S4, S5 UZBC8 CF7 31 PA35 2012 fpvAI + fpvB S3, S4, S5 UZBC8 CF7 31 PA36 2012 fpvAI + fpvB S3, S4, S5 UZBC9 CF10 30 PA13 2011 fpvA IIb + fpvB S2 UZBC9 CF16 36 PA22 2011 fpvA IIb + fpvB S2 UZBC9 CF16 36 PA26 2011 fpvA IIb + fpvB S2 UZBC10 CF18 32 PA27 2011 fpvA IIb + fpvB - UZBC10 CF18 32 PA28 2011 fpvA IIb + fpvB - UZBC11 CF7 31 PA33 2012 fpvAI + fpvB S1, S4, S5 UZBC11 CF7 31 PA34 2012 fpvAI + fpvB S1, S4, S5 UZBC12 CF11 27 PA14 2011 fpvA IIb + fpvB S1, AP41 UZBC12 CF11 28 PA37 2012 fpvA IIb + fpvB S1 UZBC12 CF11 28 PA38 2012 fpvA IIb + fpvB S1 UZBC13 CF20 19 PA30 2012 fpvA III + fpvB S1, S5 UZBC13 CF20 19 PA31 2012 fpvA III + fpvB S1, S5 UZBC13 CF20 20 PA50 2012 fpvA III + fpvB S1, S5 UZBC13 CF20 20 PA51 2012 fpvA III + fpvB S1, S5 UZBC13 CF20 20 PA52 2012 fpvA III + fpvB S1, S5 UZBC13 CF20 20 PA53 2012 fpvA III + fpvB S1, S5 UZBC14 CF1 23 PA1 2011 fpvAI + fpvB S2, S4, S5 UZBC15 CF22 21 PA111 2003 fpvA IIb S1 UZBC15 CF9 38 PA12 2011 fpvA IIb S1 UZBC15 CF13 21 PA87 2003 fpvA IIb S1 UZBC15 CF13 22 PA84 2004 fpvA IIb S1 UZBC15 CF13 29 PA16 2011 fpvA IIb S1 UZBC15 CF13 29 PA17 2011 fpvA IIb S1 UZBC15 CF21 37 PA39* 2012 fpvA IIb S1 UZBC15 CF21 37 PA40 2012 fpvA IIb S1 UZBC15 CF21 37 PA41 2012 fpvA IIb S1 UZBC15 CF21 37 PA42 2012 fpvA IIb S1 UZBC15 CF21 37 PA43 2012 fpvA IIb S1 UZBC15 CF21 37 PA44 2012 fpvA IIb S1 UZBC15 CF21 37 PA45 2012 fpvA IIb S1 UZBC15 CF21 37 PA54 2013 fpvA IIb S1 UZBC16 CF2 49 PA2 2011 fpvAI + fpvB - UZBC16 CF2 49 PA24 2011** fpvAI - UZBC17 CF3 26 PA3 2011 fpvAI + fpvB S2, S4 UZBC17 CF3 26 PA5 2011** fpvAI + fpvB S2, S4 UZBC18 CF17 10 PA23 2011 fpvAI + fpvB S3 aP. aeruginosa clones based on the combination of Diversilab and multiplex typing results. bP. aeruginosa isolates that were collected during the present study are given the name “PA” and a number. PA7 and PA14 have no relationship with the previously whole-genome sequenced isolates PA7 and PA14 being a multidrug-resistant non-respiratory (Roy et al., 2010) and virulent burn wound isolate (Lee et al., 2006), respectively. *PA39 is referred to as CF_PA39 in this article because of practical reasons. **Isolated at a later time point. Isolates belonging to different clones are separated by a solid line, while isolates that belong to the same clone but were isolated from different patients are separated by a dashed line.

79 Deletion of TonB-dependent receptor genes

2.3.2 The fpvB gene is deleted in a small proportion of P. aeruginosa CF isolates

An interesting feature that was shared by all the 14 isolates belonging to the UZBC15 clone was the lack of amplification of the fpvB gene, encoding the alternative type I ferripyoverdine receptor (Ghysels et al., 2004). In total, amplification of this gene was unsuccessful in 17 out of 54 P. aeruginosa isolates (4 out of 18 clones, corresponding to 22%). Noteworthy, in two patients (CF2 and CF19) the loss of this receptor gene could be detected in the course of the period of sputum collection. In order to verify if the lack of amplification was due to the absence of the fpvB gene and not because of SNP’s or small deletions affecting the primer annealing, primers were designed to anneal to the genes flanking fpvB (PA4167 and PA4169 in the PAO1 genome) (Table S2.1). A PCR fragment of ca. 1400 bp was detected for all the UZBC15 isolates, while the PAO1 reference strain and other isolates that were fpvB+ via multiplex PCR typing, yielded a fragment of the expected size (3571 bp) (Figure 2.1A).

Figure 2.1. A. Gel electrophoresis of amplified fragments of the fpvB region using primers fpvBreg_F and fpvBreg_R. In case of PA20 and PAOI an expected fragment around 3500 bp was observed, while in case of PA12, PA16, and PA17 a fragment below 1500 bp was observed indicating that a large part of the fpvB gene was deleted. In some cases (PA19 and PA24 are indicated) no amplification was observed, probably being the result of a larger deletion including the primer binding sites. The SmartLadder MW-1700-10B (Eurogentec, Charleroi, Belgium) was used as a molecular weight marker. B. Schematic representation of the fpvB region in PA12, PA16, and PA17. After sequencing of the fpvB region using primers fpvBreg_F and fpvBreg_R, it was observed that most of the fpvB gene was deleted (black), leaving only portions of the original sequence (green). Multiple deletion events probably occurred as one large deletion started in the intergenic region between PA4167 and fpvB and ended inside the gene, and two other deletions are located inside the gene.

80 Chapter 2

Furthermore, for all the isolates that were fpvB- via multiplex PCR typing, no PCR product was obtained indicating that both fpvB and its adjacent genes are absent in the genome of these isolates. Sequence analysis of the amplified fragment revealed that about 86% of the fpvB gene was deleted, without affecting the neighboring genes (Figure 2.1B).

2.3.3 Deletion of ferripyoverdine receptor genes does not confer resistance to antibiotics frequently used to treat CF infections

In order to test if the loss of the fpvB gene could confer resistance to one of the antibiotics that are commonly used to treat P. aeruginosa CF infections, mutants lacking fpvAI, fpvB or both were constructed in a PAOI background. No significant difference in resistance to any of the seven tested antibiotics was observed between the mutants and parent strain PAOI, indicating that none of both ferripyoverdine receptors facilitate the uptake of these antimicrobial compounds (Figure S2.2).

2.3.4 Genome sequence analysis of a Belgian epidemic P. aeruginosa CF isolate reveals deletions in several genomic regions including two TonB-dependent receptor genes

Following the finding that the fpvB gene can be lost by a number of P. aeruginosa strains during the colonization of the CF lung environment, next to the fact that one of the transmissible P. aeruginosa clones was characterized by the absence of this gene, we decided to sequence the complete genome of CF_PA39, an isolate belonging to the UZBC15 clone in order to verify if the deletion of TonB-dependent receptor genes is common during the patho- adaptation of P. aeruginosa to the CF lung. Using the RAST server, deleted regions were identified by comparing the genomic content of CF_PA39 with that of the P. aeruginosa strains PAO1, PA14 (both burn wound isolates), and PALESB58 (a CF isolate). A list of the deleted regions is shown in Table 2.3. Remarkably, next to the absence of two TonB- dependent receptor genes (fpvB and PA2070 in the PAO1 genome), the genome of this P. aeruginosa CF isolate lacked two large regions including the entire type III secretion system (T3SS) and the region corresponding to PA2171-PA2228 of the PAO1 genome that include exoY, the hcnABC genes involved in the production of hydrogen cyanide (Pessi & Haas, 2000; Pessi & Haas, 2001), PA2206 encoding a LysR regulator involved in oxidative stress defense (Reen et al., 2013), and the global AraC regulator gene vqsM (PA2227) (Dong et al., 2005).

81 Deletion of TonB-dependent receptor genes

In addition, next to the large region encoding the T3SS machinery and its regulators, the genes encoding the T3SS toxins ExoS, ExoT, and ExoY (located at different positions in the PAO1 genome) as well as exoU were absent. Other interesting regions that are absent in the CF_PA39 genome are involved in phenazine biosynthesis (phzH), exopolysaccharide production (pslAB), and rhamnolipid biosynthesis (rhlC).

Conversely, by identifying regions that were present in the genome of CF_PA39, but absent in the three reference genomes, it was possible to gain insight into the accessory genome of this transmissible clone.

Table 2.3. Deleted regions in the genome of CF_PA39.

Deleted regiona Size (bp) Function PA2171-PA2228 57,822 Carbohydrate transport, amino acid transport, virulence (catalase, ExoY toxin, hydrogen cyanide), antibiotic resistance PA1691-PA1725 24,322 Type III secretion system/virulence PA1015-PA1025 12,771 Fatty acid and phospholipid metabolism PA2594-PA2603 9,318 Sulfur metabolism PA1914-PA1918 6,364 Amino acid transport PA2307-PA2314 5,711 Inorganic ion transport and metabolism, carbohydrate transport and metabolism PA2123-PA2125 4,139 Energy production and conversion PA4895-PA4897 4,103 Iron sensing and uptake PA2229-PA2232 3,376 Exopolysaccharide/biofilm formation (pslAB) PA2036-PA2037 2,739 Coenzyme metabolism PA2931-PA2933 2,458 Energy production and conversion, carbohydrate transport and metabolism PA0050-PA0051 1,136 Phenazine biosynthesis/virulence (phzH) PA0051 1,124 Phenazine biosynthesis/virulence (phzH) PA1243 2,150 Two-component regulatory system PA2070 2,108 TonB-dependent receptor PA4822 2,043 Inorganic ion transport and metabolism PA3124-PA3125 1,986 Amino acid transport and metabolism PA0043-PA0044 1,696 Type III secretion system/virulence (ExoT toxin) PA4168 1,563 TonB-dependent receptor/iron uptake (fpvB) PA2920 1,528 Cell motility and secretion /signal transduction PA3841-PA3843 1,508 Type III secretion system/virulence (ExoS toxin + chaperone) PA2416 1,479 Carbohydrate transport and metabolism PA4869 1,466 Unknown PA2933-PA2934 1,436 Virulence (CFTR inhibitory factor, Cif) PA5032-PA5033 1,339 Unknown PA1130-PA1131 1,295 Rhamnolipid biosynthesis/virulence (rhlC) PA4108 1,245 Cyclic di-GMP phosphodiesterase/signal transduction PA5031-PA5032 1,182 Secondary metabolites biosynthesis, transport, and catabolism PA4915 1,180 Cell motility and secretion /signal transduction PA1930 1,126 Cell motility and secretion /signal transduction PA5218-PA5219 1,014 Iron uptake, carbohydrate transport and metabolism aRelative to PAO1 genome.

82 Chapter 2

In a second approach, the draft genome sequence of CF_PA39 was screened for novel and previously described genomic islands. The accessory genome is listed in Tables 2.4-2.7 and shown in Figure 2.2 and includes genomic islands (Table 2.4), prophage sequences (Table 2.5), and CRISPRs (Table 2.6). Next to the identification of four unique genomic islands, a number of genomic islands were found to be shared with other P. aeruginosa CF genomes including the gene cluster encoding biosynthesis of the antifungal compound pyoluteorin that was found in the genomes of PALESB58 (Winstanley et al., 2009), PACS171b and PACS88 (Hayden et al., 2008). Since we have collected P. aeruginosa isolates belonging to the UZBC15 clone over a period of nine years and from different patients, we tested whether the majority of the identified deleted and acquired genomic content was present in different P. aeruginosa UZBC15 isolates including PA12 (CF9), PA54 (CF21), PA16, PA17, PA87, PA84 (CF13), and PA111 (CF22) using a PCR-based approach (Primers are listed in Table S2.2A-B). Only one deletion (region PA2307-PA2314 in the PAO1 genome) was absent in the genome of the earliest isolate (PA111) compared to the other isolates, probably indicating that the majority of genomic rearrangements took place at a time point before the collection of these P. aeruginosa isolates while genomic reorganization rarely occurred afterwards.

Table 2.4. Genomic islands of CF_PA39.

Genomic Position in draft Size (bp) Predicted function Presence of region in other island genome (bp) P. aeruginosa genomes PA39-GI1 6,706-43,106 36,401 Type I restriction- Noa,b modification system PA39-GI2 666,991-672,838 5,848 Pathogenecity Noa,b,c PA39-GI3 1,085,223-1,109,178 23,956 Motility/Virulence/Resist Noa,b,c ance PA39-GI4 1,543,067-1,555,779 12,713 Energy production and P. aeruginosa PA7 (RGP76) conversion PA39-GI5 2,351,774-2,362,251 10,478 Unknown P. aeruginosa LESB58 (LESGI-4-like) PA39-GI6 2,836,652-2,867,814 31,163 Pyoluteorin biosynthesis P. aeruginosa LESB58 (LESGI-2) PA39-GI7 3,019,071-3,063,598 44,528 Unknown Noa,b,c PA39-GI8 3,489,638-3,505,243 15,606 O-antigen biosynthesis P. aeruginosa DK2 (PADK2 GI2), P. aeruginosa C3719 PA39-GI9 4,095,818-4,111,195 15,378 Type II secretion pathway P. aeruginosa PA7 (RGP69) cluster PA39-GI10 5,167,649-5,182,443 14,795 Type I restriction- Noa,b,c modification system aAltered GC content. bPresence of genes associated with mobile genetic elements. cAdjacent to tRNA gene.

83 Deletion of TonB-dependent receptor genes

Table 2.5. Potential prophages in the genome of CF_PA39.

Integrated Phage Position in draft genome (bp) Size (bp) Number of genes Type of prophage Prophage 1 715,728-733,596 17,869 23 Questionnable prophage* Prophage 2 5,329,633-5,339,106 9,474 9 Incomplete prophage * In fact this genomic region corresponds to the gene cluster of an F-type pyocin that has not yet been functionally characterized.

Table 2.6. Predicted CRISPR loci in the genome of CF_PA39.

CRISPR Position in draft Size cas genes Presence of CRISPR locus (apart from locus genome (bp) (bp) present spacers) in other P. aeruginosa genomes CRISPR 1 2,593,034-2,603,775 10,742 6 P. aeruginosa DK2 (PADK2 GI5) CRISPR 2 3,942,051-3,943,402 1,352 No No

Figure 2.2. Genome map of P. aeruginosa CF_PA39. Blast results are shown for the comparison between the genome of P. aeruginosa CF_PA39 and that of P. aeruginosa PAO1, PALESB58 and PA14. Genomic islands (GIs) are indicated and are represented by an orange arrow in accordance with their size. Prophage sequences are indicated by black arrows, while CRISPR loci are represented by purple arrows.

84 Chapter 2

2.3.5 P. aeruginosa CF_PA39 belongs to a new sequence type

By means of the genome sequence of P. aeruginosa CF_PA39, it was possible to determine the multilocus sequence type (MLST) profile of this strain as it was submitted to the P. aeruginosa PubMLST database (pubmlst.org/paeruginosa; Jolley & Maiden, 2010). P. aeruginosa CF_PA39 belongs to a new sequence type (ST-1744) to which no other isolates have been assigned so far.

2.3.6 Deletion of TonB-dependent receptor genes in P. aeruginosa strains of the DK2 lineage during adaptation to the CF lung

All of our previous findings indicate that genomic rearrangements can occur during adaptation of P. aeruginosa to the CF lung environment, including the deletion of TonB- dependent receptor genes. In order to confirm this hypothesis, we decided to use a comparative genomics approach using the genomes of the DK2 isolates. This collection of 45 P. aeruginosa isolates from 16 individuals can be considered as a gold mine to study the evolution of a P. aeruginosa clone in the CF lung environment since they have been collected over a period of 35 years (Yang et al., 2011). In total about seven different TonB-dependent receptor genes were deleted during the colonization of P. aeruginosa isolates in the lungs of CF patients attending the Copenhagen Cystic Fibrosis Clinic (Table 2.7). In the case of PA4156 and fpvB this deletion occurred at respectively three and two independent events. Finally the fpvA receptor gene that was lost during the adaptation of P. aeruginosa in the lungs of one patient in the Danish patient cohort also appeared to be absent in the genome of P. aeruginosa PACS2 (Hayden et al., 2008).

Table 2.7. List of TonB-dependent receptor genes hat have been deleted in P. aeruginosa isolates of the DK2 lineage during colonization of the CF lung.

Deleted gene (relative to PAOI Number of independent deletion Core/Accessory genome genome) eventsa PA1322 1 Core PA2289 1 Core PA2335 1 Accessory PA2398 (fpvA) 1 Accessory PA4156 3 Accessory PA4168 (fpvB) 2 Core PA4221 (fptA) 1 Core aA deletion event was considered independent if the deleted region was identified in the genome of a distinct P. aeruginosa isolate at a distinct time point in a distinct CF patient. The presence of the gene in the core or accessory genome is according to the study of Cornelis & Bodilis (2009).

85 Deletion of TonB-dependent receptor genes

2.3.7 Iron uptake by P. aeruginosa in the CF lung.

The observation that deletions in the ferripyoverdine uptake system are regularly observed in P. aeruginosa CF isolates, lead us to the hypothesis that during growth in the CF lung, P. aeruginosa is able to utilize alternative iron sources such as heme and Fe2+. In order to test this hypothesis, we have determined the expression of genes involved in pyoverdine production (pvdA), pyochelin uptake (fptA), ferrous iron uptake (feoA, feoB), and heme uptake (hasAp) in the sputum of seven Belgian and nine American CF patients, in collaboration with the lab of Dianne Newman (Caltech, CA, USA). It was observed that multiple iron uptake systems were simultaneously expressed and that relative expression levels of each gene varied over five orders of magnitude between patients (Figure 2.3).

Figure 2.3. Expression of iron uptake genes by P. aeruginosa in the sputum of seven Belgian and nine American CF patients. The expression levels of feoA, feoB (Fe2+uptake,), pvdA (pyoverdine production; Fe3+ uptake), fptA (ferripyochelin uptake; Fe3+ uptake), and hasAp (heme uptake) in the sputa of different CF patients are indicated by dark gray dots. The black, white and light gray dots represent the expression levels of these iron uptake genes in planktonic cultures to which respectively 50 μM Fe(II), 50 μM Fe(III) or nothing was added, but are not part of this thesis (Hunter et al., 2013).

In order to test if these genes are responsive to varying iron concentrations, we have compared the expression of pvdA, pvdS, fpvAI, fpvB, fptA, feoA, feoB, hasAp, and phuT in the sputum of patient CF5 (Table 2.2) to the expression levels of these genes in planktonic cultures of the isolate PA6 (collected from a sputum sample of patient CF5) in LB (iron- replete) and CAA (iron-poor) medium.

86 Chapter 2

Figure 2.4. Expression of genes involved in pyoverdine production (pvdA, pvdS) and uptake (fpvAI, fpvB), ferripyochelin uptake (fptA), ferrous iron uptake (feoA, feoB), and heme uptake (hasAp, phuT) in CF sputum relative to LB (black) and CAA (white).

Interestingly, the expression of genes involved in pyoverdine biosynthesis and uptake was highly enhanced in CF sputum relative to LB medium, indicating that pyoverdine is used to acquire iron in the CF lung (Figure 2.4). The feoA and feoB genes that are involved in the uptake of ferrous iron were also induced in CF sputum compared to laboratory media, indicating that P. aeruginosa also acquires iron in its reduced form in the CF lung. In addition, the hasAp and phuT genes were up-regulated in CF sputum relative to the laboratory media (with the exception of hasAp that was down-regulated relative to CAA), thereby suggesting that heme is used as an alternative iron source by P. aeruginosa in the CF lung environment. Finally, the fptA gene was not induced in CF sputum relative to LB, while it was highly down-regulated in CF sputum compared to CAA. These results confirm the previous observations that multiple iron uptake systems are being used by P. aeruginosa in the CF lung.

87 Deletion of TonB-dependent receptor genes

2.4 Discussion

In order to get an insight into the P. aeruginosa population that is present amongst CF patients attending the CF reference center at UZ Brussel, a combination of two previously described typing methods was used. The first approach comprised typing P. aeruginosa CF isolates, which is based on the amplification of repetitive elements in the genome via the DiversiLab method which has been used extensively to rapidly detect hospital outbreaks of P. aeruginosa (Deplano et al., 2011; Doleans-Jordheim et al., 2009; Fluit et al., 2010; Masoud- Landgraf et al., 2012; Ratkai et al., 2010). However, a number of studies have shown that typing solely based on DiversiLab can lead to an underestimation of the number of P. aeruginosa clone types present in the population (Fluit et al., 2010; Maatallah et al., 2013). Therefore, we have followed a second typing approach using a multiplex PCR assay that allows to type P. aeruginosa isolates based on the presence of different types of ferripyoverdine genes (Bodilis et al., 2009) and pyocin genes. Although a number of studies have reported the presence of highly transmissible P. aeruginosa clones dominating CF populations (Anthony et al., 2002; Cheng et al., 1996; Fothergill et al., 2012; Jelsbak et al., 2007; Scott & Pitt, 2004), we were not able to detect such a clone in our population as the majority of patients were colonized by a distinct P. aeruginosa clone. In a recent comparative study between the CF reference centers in Copenhagen and Hanover, it was found that the spread of dominant clones in the CF population is greatly dependent on the antibiotic treatment strategy and the followed infection control measures (Cramer et al., 2012). It has been suggested that the frequent switch between antibiotics and the less conservative prescription thereof, in combination with only partial segregation of P. aeruginosa-positive and P. aeruginosa-negative patients, may have led to the replacement of initially acquired clones by P. aeruginosa clones with a superior fitness in the antibiotic-rich CF lung environment. At the CF reference center of the UZ Brussel, an intermediate antibiotic treatment strategy was adopted (in general, switching of antibiotics only occurred if resistance was detected among bacterial isolates of a CF patient), while segregation in combination with strict hygiene measurements have been applied at all times, the UZ Brussel center being rather small with less than 200 patients. This could account for the absence of a dominant P. aeruginosa clone among CF patients. Nevertheless, a few cases of potential inter-patient transmission were observed, although contact between patients of the seven different Belgian CF centers as well as patient transfer to another Belgian center, is very limited. Interestingly, one particular clone (UZBC15) was found in four different CF patients and has been reported

88 Chapter 2 to be distributed in at least three other Belgian CF reference centers, belonging to a national cluster of patients that harbor the same P. aeruginosa genotype.

A common feature of the UZBC15 isolates was the lack of amplification of the fpvB gene during the multiplex PCR assay. In a previous study this gene was reported to be amplified in 100% of the CF isolates tested in patients from different geographical origin (Pirnay et al., 2009). Using a sequencing-approach, it was shown that about 86% of this gene was deleted in all isolates of the UZBC15 clone. Furthermore, three isolates, each belonging to a unique P. aeruginosa CF clone in the population, that did not show amplification of the fpvB gene, appear to have undergone larger deletions including fpvB and its adjacent genes. Remarkably, in the case of one patient (CF2), a P. aeruginosa isolate (PA2) containing both the fpvAI and fpvB genes was collected, while six months later a clonally related isolate (PA24) only containing the fpvAI gene was retrieved from the sputum of this patient. In addition, two clonally related P. aeruginosa isolates, from which one had lost the fpvB gene, were collected simultaneously from the sputum of patient CF19. Finally, we have detected two independent deletion events (in two different patients) comprising the fpvB gene in P. aeruginosa DK2 strains during CF lung colonization. Altogether, these observations indicate that deletion of the fpvB gene can occur during colonization of the CF lung environment and might be part of the adaptation process that P. aeruginosa undergoes. Next to fpvB, six other TonB-dependent receptor genes have been found to be missing compared to PAO1 during lung colonization of the CF lung. These receptor genes can be part of either the core or accessory genome of P. aeruginosa (Cornelis & Bodilis, 2009). Strikingly, one receptor (PA4156) was deleted at three independent events in the P. aeruginosa DK2 population.

Several hypotheses can explain the deletion of TonB-dependent receptor genes during colonization of the CF lung by P. aeruginosa:

A first hypothesis is based on the fact that some antibiotics might contain hydroxamate or catecholate groups, allowing them to bind oxidized iron. Subsequently, this antibiotic-iron complex could bind to a specific TonB-dependent receptor, facilitating its uptake into the bacterial cell. In the case of FpvAI and FpvB, we have confirmed that this hypothesis does not apply since no resistance against any of the frequently used antibiotics to treat P. aeruginosa CF infections was obtained after deletion of fpvAI and/or fpvB in a PAO1 background.

A more likely hypothesis is that pyocins exert a negative selection pressure on TonB- dependent receptors during colonization of the CF lung by P. aeruginosa since they have been

89 Deletion of TonB-dependent receptor genes proposed to shape the evolutionary adaptation of ferripyoverdine receptors (Smith et al., 2005; Tummler & Cornelis, 2005). It has been shown that different S-type pyocins utilize TonB-dependent receptors involved in iron-uptake to enter P. aeruginosa cells (Baysse et al., 1999; Denayer et al., 2007; Elfarash et al., 2012; Elfarash et al., 2014). Pyocins could indeed play a particularly important role during co-colonization of the CF lung by P. aeruginosa isolates that are genetically unrelated and possess different pyocin/immunity genes. A P. aeruginosa strain harboring the immunity gene for the pyocin produced by the co-colonizing strain and producing itself a pyocin to which the latter is sensitive could lead to the out- competition of that particular P. aeruginosa clone. Hence, pyocins can account for the negative selection pressure that leads to the deletion of certain TonB-dependent genes. In fact, negative selection pressure exerted on fpvA and fptA, both deleted in P. aeruginosa DK2 isolates, could have been caused by the pyocins S2/S4 and S5, respectively (Denayer et al., 2007; Elfarash et al., 2012; Elfarash et al., 2014).

Immune pressure against TonB-dependent receptors could be another reason for the observed deleted regions. Due to their large size and location on the outer membrane, TonB-dependent receptors could be potentially immunogenic molecules. During long-term colonization of the CF lung environment, the negative selection pressure exerted by the immune system could result in adaptive deletions. Recently, it has been described that the iron-acquisition-related TonB-dependent receptor TdrA of a pathogenic Pseudomonas fluorescens was able to elicit strong immune responses in Japanese flounder (Hu et al., 2012). This TonB-dependent outer membrane receptor possesses several conserved structural domains of other TonB-dependent outer membrane receptors indicating that certain epitopes on these proteins may act as immunogens of P. aeruginosa in the CF lung environment.

Finally, relaxation of selection could be one of the most important triggers of deletion events during adaptation of P. aeruginosa to the CF lung environment. This could particularly apply to the deletion of the fpvB gene in P. aeruginosa CF isolates. The FpvB receptor allows P. aeruginosa isolates producing type II or III pyoverdines to utilize type I pyoverdine produced by other P. aeruginosa strains hence offering a competitive advantage. However, after long- term colonization of the CF lung, this niche is often dominated by one P. aeruginosa strain. If this is a type II or III pyoverdine-producing strain, the fpvB gene would be of no further use and could be lost since there is no selection pressure acting on the gene. Furthermore, it has been shown that P. aeruginosa can utilize multiple iron-uptake systems in the CF lung to acquire iron in both its ferric and ferrous form (Hunter et al., 2013). It is known that in some

90 Chapter 2 region of the CF lung acidic and microaerobic conditions can occur, increasing the ratio of ferrous versus ferric iron. In these regions, the pyoverdine-mediated iron uptake system may be less used and hence this siderophore as well as its receptor(s) may be relieved from selection.

Indeed, we have shown that multiple iron uptake systems are simulteansously used by P. aeruginosa in the CF lung by determining their gene expression levels in the sputum of 16 CF patients, representing the Belgian as well as the American CF populations. In addition, we have determined the expression of several iron uptake genes in the sputum of a particular CF patient relative to planktonic cultures of a P. aeruginosa isolate collected from the sputum of this patient under iron-limited (CAA medium) and iron-replete (LB medium) conditions. We found that genes involved in the uptake of pyoverdine, ferrous iron, as well as heme were induced in CF sputum compared to laboratory conditions, indicating that iron is present in the CF lung in both its oxidized and reduced form.

Recently, Andersen and colleagues have investigated what selection pressures could lead to the loss of the ferripyoverdine receptor genes by analyzing pyoverdine production as well as the whole genome sequences of longitudinally sampled isolates from a large group of CF patients (Andersen et al., 2015). They found that social interactions drive selection of the ferripyoverdine receptor genes and that the loss of function of the ferripyoverdine receptor proceeds faster when cheaters (strains that are able to uptake pyoverdine but do not produce it) are co-infected with pyoverdine-proficient rather than pyoverdine-deficient strains. Remarkably, they found that the majority of mutations that occurred overtime in the fpvA gene mapped to the extracellular loops to which ligands bind, indicating that pyocins could account for the selection pressure acting on the ferripyoverdine receptor genes. Finally, they conclude that pyoverdine as well as its receptors are targets for selection, ultimately leading to the loss of both, thereby causing a shift from pyoverdine-mediated iron uptake to the acquisition of iron via alternative systems.

Because of the fact that UZBC15 showed deletions in the fpvB region, we decided to select the CF_PA39 isolate of this clone to perform a genome sequence analysis. In addition to fpvB, another TonB-dependent receptor gene (PA2070) was absent in the genome of this isolate. Remarkably, when comparing the genome of this isolate to the genomes of the P. aeruginosa reference strains PAO1, PALESB58, and PA14, we found large potential deletions including the entire T3SS and the region corresponding to PA2171-PA2226 of the PAO1 genome.

91 Deletion of TonB-dependent receptor genes

Interestingly, a number of genes that were absent in the genome of CF_PA39 were also found to be deleted during colonization of P. aeruginosa DK2 isolates in the CF lung environment (Rau et al., 2012). These include the exopolysaccharide genes pslAB, the T3SS effector toxin gene exoY, the gene cluster for biosynthesis of hydrogen cyanide, and a catalase gene. These shared characteristics could indicate parallel evolution of two distinct CF clones that have been in the clinic for a long period. The loss of the T3SS genes can be explained by the presence of a negative selection pressure as it is known that antibody responses can be elicited against epitopes on T3SS structural proteins and effector toxins in CF patients (Corech et al., 2005; Moss et al., 2001). Another reason can be found in the fact that P. aeruginosa simply does not utilize its T3SS when present in biofilms, and that this system is lost due to relaxation of selection. From a recent study it appeared that under oxygen-limited conditions, the T3SS could be regulated by isocitrate lyase (ICL, encoded by aceA) in an ExsA- independent way (Chung et al., 2013). Furthermore, ICL also appears to be involved in the positive regulation of the production of a precursor of hydrogen cyanide (HCN) (Hagins et al., 2009). Since the aceA gene is crucial in the glyoxylate shunt pathway during anaerobic growth and the subsequent up-regulation of T3SS and HCN biosynthesis genes is very energy-consuming, it is possible that the only way to inactivate these genes was by deleting them from the genome of CF_PA39. In contrast to the loss of the T3SS, which has been associated with acute infections, the Psl-biofilm system is considered to be important during the chronic infection stage of P. aeruginosa. However, Banin et al. have shown that loss of pyoverdine production or its uptake interfered with the Psl-type of biofilm formation (Banin et al., 2005). Recently, it has been shown that mutation of multiple iron-uptake systems (i.e. pyoverdine including the receptor gene fpvA and pyochelin) lead to iron-constitutive alginate expression in P. aeruginosa strains that contain mutations in mucA, mucB, mucC, or mucD, while single mutants in only one of the muc genes produced higher levels of alginate only under iron-limited conditions (Wiens et al., 2014). This iron-constitutive expression of alginate genes has also been observed in chronic P. aeruginosa CF isolates.

Next to the list of deletions, several genomic islands have been detected in the genome sequence of CF_PA39. Although four novel genomic islands have been identified, half of them were shared with other P. aeruginosa strains. One of the most interesting genomic islands is PA39-GI6 (LESGI-2) which contains the entire gene cluster for pyoluteorin biosynthesis. This antifungal compound could offer a competitive advantage to P. aeruginosa in the CF lung. Finally, two prophages and CRISPR regions have been found from which one

92 Chapter 2

CRISPR locus was identical (with exception of spacers) to the sequence in the DK2 genome and was constituted out of six cas genes. Remarkably, except for one deletion, no differences were observed between the earliest and latest isolates of UZBC15, indicating that this clone has adapted extensively to the CF lung environment and has reached a fitness peak in this particular environment.

2.5 Conclusion

In conclusion, we gained insight in the P. aeruginosa population structure via the use of a combination of genotyping techniques. No dominant P. aeruginosa CF clone could be retrieved from the patient population studied at UZ Brussel. However, a number of potential inter-patient transmissions were detected. Deletions, comprising TonB-dependent receptor genes regularly occurred during adaptation of P. aeruginosa to the CF environment. Whole- genome sequencing of one selected isolate of a potentially transmissible P. aeruginosa CF clone revealed several large deletions that were also found in previously documented P. aeruginosa CF isolates, indicating the parallel evolution of distinct P. aeruginosa clones. Finally, we have shown that P. aeruginosa is able to utilize multiple iron-uptake systems in the CF lung environment.

93 Deletion of TonB-dependent receptor genes

2.6 Supplementary data

Table S2.1. Primers used in this study for amplification of ferripyoverdine receptor and pyocin gene fragments.

Primer Sequence (5’→3’) Expected size (bp) Reference fpvAI_Fa CGAACCCGACGAAGGCCAGA 324 Bodilis et al., 2009 fpvAI_Ra GTAGCTGGTGTAGAGGCTCAA 324 Bodilis et al., 2009 fpvAIIa_Fa TACCTCGACGGCCTGCACAT 908 Bodilis et al., 2009 fpvAIIa_Ra GAAGGTGAATGGCTTGCCGT 908 Bodilis et al., 2009 fpvAIIb_Fa GAACAGGGCACCTACCTGTA 863 Bodilis et al., 2009 fpvAIIb_Ra GATGCCGTTGCTGAACTCGTA 863 Bodilis et al., 2009 fpvAIII_Fa ACTGGGACAAGATCCAAGAGA 505 Bodilis et al., 2009 fpvAIII_Ra CTGGTAGGACGAAATGCGA 505 Bodilis et al., 2009 fpvB_Fa GCATGAAGCTCGACCAGGA 562 Bodilis et al., 2009 fpvB_Ra TTGCCCTCGTTGGCCTTG 562 Bodilis et al., 2009 S1_Fb CGACCCATTGCTGACCTTAT 646 This study S1_Rb CCGCCTCAATACTTGCTTTG 646 This study S2_Fb ATGCTTTGCCTCAACTGACC 117 This study S2_Fb TCAAGGCATTGTTTGCAGTC 117 This study S3_Fb CCGCTGGAAGTGGACATTTA 312 This study S3_Rb CCCCTCCTCTAGCAATCCTT 312 This study S4_Fb GAGGGTGGACTGAGGTTGAA 200 This study S4_Rb ATCGTTCCGATCTGCAATTT 200 This study S5_Fb ACGCAGGACAGAAGCAGAAC 459 This study S5_Rb GCAATACCCACAAGCCAACT 459 This study AP41_Fc GGCGTTGCTACCGGTAATG 304 This study AP41_Rc CGCTTTCCAGAGGGACTACA 304 This study aPrimers used in the multiplex PCR targeting ferripyoverdine receptor genes. bPrimers used in the multiplex PCR assay targeting S-type pyocin genes. cPrimers used in a normal PCR assay targeting S-type pyocin AP41.

94 Chapter 2

Figure S2.1. Dendrogram based on Diversilab typing of the P. aeruginosa CF population at UZ Brussels. The genetic distance between P. aeruginosa isolates and clades is indicated on the scale at the bottom. In some cases, the results obtained by Diversilab and multiplex PCR targeting ferripyoverdine and S- type pyocin genes did not correspond, leading to the classification of closely related isolates as distinct clones. Clones that were distributed among different patients, indicating putative transmission events, are shown in red, while clones that were present in only one patient are indicated in blue. P. aeruginosa PAOI was also included.

95 Deletion of TonB-dependent receptor genes

Figure S2.2. Antibiotic susceptibility profiles of wild-type (WT) P. aeruginosa PAOI and its mutants obtained via the disc diffusion method. *Not significant (P>0.05).

Table S2.2A. Primers used in PCR assay to confirm the presence of putative deletions in the genome of isolates belonging to the UZBC15 clone.

Primer Sequence (5’→3’) Deleted regiona Size (bp)b, c TIIISS_F GCCTCAGCTACACCGAGTTC PA1690-PA1725 1389b/26416c TIIISS_R ACGTGCGCGTCTACGATTAC PA1690-PA1725 1389b/26416c fpvBreg_F AGGCTACGTGAGTTCGGAGA PA4167-PA4169 1324b/3571c fpvBreg_R TCGATCTGGTTGGTGGAT PA4167-PA4169 1324b/3571c G2_F ACTGCGGCACCTCCTATG PA1243 442b/2811c G2_R GTCGGGAAACTGTTGCTCAG PA1243 442b/2811c G3_F GCAGGGCTATCCCAATCTC PA1015-PA1025 1003b/13786c G3_R GTCGAATACGGCAACAGACC PA1015-PA1025 1003b/13786c G4_F CTGGGGTCCCAGACATACTC PA1914-PA1918 684b/7059c G4_R TTCGATAGCGAACGCCTCT PA1914-PA1918 684b/7059c G5_F CGTTCGATACGGTCGATTTC PA2036-PA2037 976b/3718c G5_R ATGGTCTGGGCAGTGATCTT PA2036-PA2037 976b/3718c G6_F ACCCGGTCCAGGTAGAGGT PA2070 1207b/3382c G6_R TGTTGCGATAAAGCTCGATG PA2070 1207b/3382c G7_F GGGCTACCTCAGCTACGAGA PA2123-PA2125 1217b/5346c G7_R CGACTACAACCCCAATACCG PA2123-PA2125 1217b/5346c G8_F ATGTCCCTGAGGGTGCAG PA2229-PA2232 1048b/4424c G8_R CGACCGTAGATGTCGTTGAA PA2229-PA2232 1048b/4424c G9_F GAAGGCGAACAGGGAGAAC PA2307-PA2314 1081b/6775c G9_R TGCTGATCTTCTTCGTCTGC PA2307-PA2314 1081b/6775c G10_F CACCGGTTTGGGTAGCAGT PA2594-PA2603 521b/9945c G10_R GACTGGGTACCGATGATGCT PA2594-PA2603 521b/9945c

96 Chapter 2

Primer Sequence (5’→3’) Deleted regiona Size (bp)b, c G11_F AAGCCTGGAACACCTGGAC PA2931-PA2933 719b/4613c G11_R ATTGAACTCGATCGCTTCGT PA2931-PA2933 719b/4613c G12_F CAACCCTGGCTCATATCGAC PA3124-PA3125 707b/2747c G12_R GGAGTACCTACCCGCCCTAC PA3124-PA3125 707b/2747c G13_F CCACCTGTTCATCGACCTG PA4822 1355b/3399c G13_R GGGCGCCYCTTTTCCAAC PA4822 1355b/3399c G14_F GATCCAGGGAAAGCGTCAT PA0043-PA0044 701b/2397c G14_R ATGTAGCTGGAGCGGTTGTC PA0043-PA0044 701b/2397c G15_F CGAACTTCAGCAAGGAGGAA PA2416 1253b/2732c G15_R GACCTGGTGATGAAGGTGCT PA2416 1253b/2732c G16_F TGCTCTTCGGCAAACTCTG PA2920 1203b/2733c G16_R GCCTGACCTTCACCGAATAC PA2920 1203b/2733c G17_F CCTGCCGCTACTGAACTTTG PA3841-PA3843 1149b/3131c G17_R GCCTACGACTGGCTGGAG PA3841-PA3843 1149b/3131c G18_F ATGGTCGCTATCGATGCTGT PA4108 1389b/2729c G18_R AGCATCGTCTTCGACACCTC PA4108 1389b/2729c G19_F TCGCCAAGTACACCATCAAC PA4869 759b/2223c G19R CAGGAACAGATCGACAGCAC PA4869 759b/2223c G20_F CTGTACGCGCTCGGTTTC PA4895-PA4897 1398b/5850c G20_R AACTGAAGATCGGCGAGATG PA4895-PA4897 1398b/5850c G21_F CCACGCTTCGACACCTTCTA PA4915 1192b/2429c G21_R TGAGCGTGTTCTACCTGGTG PA4915 1192b/2429c G22_F GTGCTCTCGGTGGTTTACCT PA5031-PA5032 1099b/3620c G22_R CGTTGATCACGAACAGCTTG PA5031-PA5032 1099b/3620c aRelative to PAOI genome. bSize expected during PCR amplification of this region in the genome of CF_PA39. cSize expected during PCR amplification of this region in the genome of PAOI.

Table S2.2B. Primers used in PCR assay to confirm presence of genomic islands, putative prophages, and CRISPR regions.

Primer Sequence (5’→3’) Size (bp) GI1_F a ACTACCGCTCGGTGTCTCTG 1600 GI1_Ra ACTGCTGGATGCACTGATTG 1600 GI2_F CTGTTCGTCTAGCGGCTCTT 1397 GI2_R ATGTTGACGCATCGGCTACT 1397 GI3_F GATGCCCGTGACATCTTCAT 1394 GI3_R CACACGCGTAAGACCAAGAA 1394 GI4_F GAAATCGACCTGCTCCTGA 1066 GI4_R GAACATCTGGAAGGCGATGT 1066 GI5_F GTGAATTCGAAAGCCCTGTG 1197 GI5_R GTCGAAGCCTTTGCTCACC 1197 GI6_F GACATCCGCATCAAGGAGTT 822 GI6_R GTTGCACCTGTTCGTAGCAA 822 GI7_F GAGAAGTGGCGTTGAAGGAG 1385 GI7_R CCACGAGTTGATTTCCGAGT 1385 GI8_F CTGCAGATTTCGGATGAACA 1402 GI8_R CGCAATGAAAGAACTGCGTA 1402

97 Deletion of TonB-dependent receptor genes

Primer Sequence (5’→3’) Size (bp) GI9_F ATTCATCTCGAGCTCCATCG 1209 GI9_R GAAGAAATACGGCGGGTTC 1209 G10_F AAAAGCTCATCCCCCACTTC 1200 G10_R AGGAACTGAGCGCAAAGACT 1200 Prophage1_F AGTTCCATGTCGCTGTGATG 1205 Prophage1_R CTTGACCTCGGCCTTGAG 1205 Prophage2_F CCTCCAGAACTGGAACAACG 1200 Prophage2_R TTCGTCCTGTTCACCCTCTT 1200 CRISPR1_F CCCGAGGTAACAGAATCGTC 1287 CRISPR1_R CGGCAGTGTTGCAGGTAGTA 1287 CRISPR2_F CGATTCGTAGGGCGAATAAC 1612 CRISPR2_R TTTCGGACGATTTCTTACGC 1612 aThe primers were also used to confirm the deletion of the PA2171-PA2228 region since the genomic island GI1 has substituted all genes between PA2170 and PA2229.

Table S2.3. Primers used for real-time PCR . Target gene Primer Sequence ( 5’-3’) Primer efficiency (%) pvdA fpvdA F CACAGCCAGTACCTGGAACA 99 fpvdA R GGGTAGCTGTCGTTGAGGTC pvdS pvdS F ACCGTACGATCCTGGTGAAG ND pvdS R TGAACGACGAAGTGATCTGC fpvA1 fpvAI F GAGCCTCAGGACCAGTTGAG ND fpvAI R CGGGGATTGTTGTAGACCAT fpvB fpvB F TGATGGCCGGCTATACCTAC ND fpvBR GTATAGCTGGTGGCGAGCTT fptA fptA F GGACCGCGACTACTTCTACG 97 fptA R TCGAGTCGATGTGCTGGTAT feoA feoA F AACCGTCCCGTTCCTACC 107 feoA R CAGAAGCCCCATGGAGAA feoB feoB F GAGCGGCTGATTACCATCAT 90 feoB R CGAGCAGGTACAGGGAGAAG hasAp hasAp F AAGGTGGTCTACGGCCTGAT 87 hasAp R ACTGGTCGAAGGTGGAGTTG phuT phuT F GACGTCTGCTGGTGCTTGAC ND phuT R CTTCCGTACTCAGCGGTTTC oprI oprI F AGCAGCCACTCCAAAGAAAC 99 oprI R CAGAGCTTCGTCAGCCTTG F, forward primer. R, reverse primer. ND, not determined.

98 Chapter 3

3 Chapter 3: Effect of shear stress on Pseudomonas aeruginosa isolated from the cystic fibrosis lung

The content of this chapter has been submitted as a manuscript to mBio, entitled:

“Effect of shear stress on Pseudomonas aeruginosa isolated from the cystic fibrosis lung" by Jozef Dingemans, Pieter Monsieurs, Sung-Huan Yu, Aurélie Crabbé, Konrad Förstner, Anne Malfroot, Pierre Cornelis, and Rob Van Houdt.

99 Effect of shear stress on P. aeruginosa in vitro

100 Chapter 3

3.1 Introduction

Pseudomonas aeruginosa is a Gram-negative γ-proteobacterium that can dwell in a wide range of environments, including water, soil, animals and the human host (Pirnay et al., 2009). Although this bacterium is harmless to the healthy human host, it poses great danger for individuals that suffer from burn wounds, immunodeficiency and in particular cystic fibrosis (CF), and is therefore considered an opportunistic pathogen (Kerr & Snelling, 2009; Lyczak et al., 2002). This opportunistic lifestyle is facilitated by the multitude of virulence factor-encoding genes present in its large genome (6-7 Mbp) (Mathee et al., 2008; Silby et al., 2011) as well as its high metabolic versatility (Frimmersdorf et al., 2010). The lungs of CF patients can initially be infected by P. aeruginosa via two different routes, either by environmental strains that have no clinical case history or by the transmission of adapted strains colonizing other CF patients (Anthony et al., 2002; Jelsbak et al., 2007; Marvig et al., 2015; Scott & Pitt, 2004; van Mansfeld et al., 2009). Since the development of whole-genome sequencing approaches (Dark, 2013), more insight into the adaptation of environmental P. aeruginosa strains to the CF lung has been gained. Multiple studies have reported that P. aeruginosa strains acquire deletions in genes that appear to be less vital for persistence in the CF lung environment, while the horizontal acquisition of novel genomic content has been less frequently observed (Dingemans et al., 2014; Lucchetti-Miganeh et al., 2014; Rau et al., 2012; Smith et al., 2006; Stewart et al., 2014). Additionally, P. aeruginosa rewires its global regulatory networks in order to survive in the hostile CF lung environment that is characterized by the presence of immune cells, competing pathogens and excessive antibiotic treatment (Damkiaer et al., 2013; Marvig et al., 2013; Marvig et al., 2015). Furthermore, it is believed that the shear stress in the CF lung is low due to the presence of viscous sputum that impairs the shear-causing mucociliary movement (Blake, 1973; Knowles & Boucher, 2002).

Previously, we have used the rotating wall vessel (RWV) bioreactor technology to study the response of P. aeruginosa to low (LS) and high (HS) fluid shear regimes (Crabbé et al., 2008; Crabbé et al., 2010). The RWV is a cylindrical bioreactor that, when completely filled and rotated on an axis parallel with the ground, results in solid body mass rotation of the culture medium, hence creating a low fluid shear environment (Nauman et al., 2007). Addition of different types of beads or horizontal positioning of the RWV has been reported to enhance fluid shear levels in the RWV (Crabbé et al., 2008; Crabbé et al., 2010; Nauman et al., 2007). Low fluid shear conditions were previously shown to affect gene expression and

101 Effect of shear stress on P. aeruginosa in vitro phenotypic traits of the pathogens Salmonella enterica serovar Typhimurium, Escherichia coli, and Staphylococcus aureus as compared to higher fluid shear controls (Castro et al., 2011; Lynch et al., 2004; Nauman et al., 2007; Pacello et al., 2012).

Using the rotating wall vessel (RWV), we previously demonstrated that culturing of the P. aeruginosa PAO1 reference strain in LB medium in the LS environment of the RWV bioreactor leads to the formation of biofilms in suspension (Crabbé et al., 2008). However, when the fluid shear in the RWV module was increased by means of a ceramic bead, a more surface-attached biofilm phenotype was observed. Furthermore, the alternative sigma factor AlgU mainly appeared to orchestrate this response to LS conditions, resulting in elevated levels of the exopolysaccharide alginate (Crabbé et al., 2010). These data suggested that the RWV bioreactor creates environmental conditions that trigger phenotypic traits in P. aeruginosa relevant to those in CF lung mucus, since alginate-containing biofilms are one of the hallmarks of chronic P. aeruginosa infections in this environment (Hogardt & Heesemann, 2010; May et al., 1991; Pedersen, 1992). However, the PAO1 reference strain has no CF background, since it was originally isolated from a wound (Holloway, 1955; Stover et al., 2000), and LB medium does not mimic the content and viscosity of sputum present in the lung environment of CF patients. Recently, we have reported the presence of a transmissible P. aeruginosa CF strain in Belgian CF patients, distributed among different CF reference centers (Dingemans et al., 2014). This strain has been found in CF reference centers for more than ten years and whole-genome sequencing revealed that its adaptation to the CF lung conditions involved the accumulation of numerous deletions.

In this study, we have scrutinized the effect of shear stress on the behavior of this well-characterized, highly adapted and transmissible CF strain at the transcriptomic, biofilm, and quorum sensing (QS) level in artificial sputum medium using the RWV bioreactor. Biofilm formation in response to shear stress was assessed via scanning electron microscopy, while an RNA sequencing approach was adopted to determine the effect of shear stress on the transcriptome of P. aeruginosa CF_PA39. In addition, small RNAs (sRNAs) in the genome of this strain were de novo predicted and the expression of these sRNAs along with previously confirmed sRNAs in other P. aeruginosa genomes was quantified. Finally, the production of both short-chain and long-chain (3-oxo-C12-HSL) QS molecules was determined to assess the role of QS in the response of P. aeruginosa CF_PA39 to fluid shear.

102 Chapter 3

3.2 Materials and methods

3.2.1 Bacterial strains and culture conditions

All bacterial strains used in this study are listed in Table S3.3. Bacterial overnight cultures were grown at 37°C in nutrient rich lysogeny broth (LB) medium (Life Technologies) with continuous shaking at 200 rpm unless mentioned otherwise, prior to inoculation in ASM.

3.2.2 Preparation of artificial sputum medium

Artificial sputum medium (ASM) was prepared as described by Fung et al. (Fung et al., 2010) without the addition of antibiotics. In order to obtain sterile ASM, stock solutions of bovine serum albumin (BSA), casamino acids (CAA), salts, diethylene triamine pentaacetic acid (DTPA) and salmon sperm DNA were filter-sterilized prior to use. Furthermore, the porcine stomach mucin solution was autoclaved at 121°C for 15 min since the viscosity of this solution did not allow filter-sterilization. Next, a 500 ml solution of ASM was prepared by combining all of the-above mentioned solutions with a sterile egg yolk solution in a laminar flow cabinet. In contrast to the study of Fung et al. (Fung et al., 2010), the pH of the ASM was adjusted to 6.8, in agreement with the study of Palmer et al. (Palmer et al., 2005) in which this value was found to be representative of the slightly acidic conditions found in the CF lung. Sterility of the ASM was verified by plating 100 µl of this solution on LB, followed by incubation at 37°C for 72h.

3.2.3 Rotating wall vessel experiment

An isolated mucoid colony of P. aeruginosa CF_PA39 was inoculated in 5 ml of LB medium and grown overnight with continuous shaking at 200 rpm. Next, the OD600nm was adjusted to a value of 1.0 using LB medium and 1 ml of the adjusted bacterial culture was centrifuged at 8000 x g to collect the cells. The resulting pellet was washed once with 1x PBS and re- suspended in 1 ml of this buffer. Subsequently, this bacterial culture was diluted 1000x in ASM, briefly mixed by vortexing for 5 s, and ~50 mL of this mixture was transferred to the RWV bioreactor (Synthecon Inc.) via the filling port with (high fluid shear, HS) or without (low fluid shear, LS) two glass beads with a diameter of 4 mm (Merck Millipore). Air bubbles were removed via the sampling ports using a 5 ml syringe. Finally, the RWV bioreactors were

103 Effect of shear stress on P. aeruginosa in vitro incubated at 37°C for 24 h at 25 rotations per minute (rpm) while humidity of the incubator was maintained. The experiment was performed in biological triplicate.

3.2.4 Determination of bacterial counts

In order to recover the entire bacterial population after 24h of growth in the RWV bioreactor, about half of the bacterial culture volume was transferred to a 50 ml Falcon tube, while the other half was vortexed for 30 s inside the RWV bioreactor. In this way, bacterial cells that would be attached to the gas-permeable silicone membrane were included. Finally, the two volumes were pooled and briefly vortexed for 5 s. Serial dilutions of the biological triplicates of each condition prepared in PBS were plated on LB medium and after 24h of growth at 37°C, colonies were quantified according to their phenotype.

3.2.5 RNA isolation

Multiple 2 ml aliquots of the bacterial cell culture, recovered from the RWV bioreactor as mentioned above, were centrifuged for 5 min at 8000 g. The resulting pellet was flash-frozen in liquid nitrogen and stored at -80°C until further processing. RNA was extracted using the SV Total RNA Isolation System (Promega). In summary, the thawed cell pellet was re- suspended in 200 µl of a freshly prepared lysozyme solution (3 mg/ml in Tris-EDTA buffer) followed by 5 min of incubation at room temperature. Next, this mixture was divided into two smaller volumes. Upon cell lysis and addition of RNA dilution buffer, samples were centrifuged for 15 min at maximum speed and the clear supernatant was transferred to fresh 1.5 ml microcentrifuge tubes. This step was necessary to avoid blocking of the column since the lysates appeared to be viscous, most likely due to the presence of alginate and/or mucin. Upon the addition of ethanol and a washing step, an on-column DNase treatment was performed, followed by two washing steps. Finally, RNA was eluted in nuclease-free water. Prior to RNA sequencing and qRT-PCR, an additional Turbo DNase (Ambion) treatment was performed via two 30 min incubation steps in the presence of 1 µl of (2U/µl) Turbo DNase. In a next step, the Turbo DNase-treated RNA was purified and concentrated using the RNA Clean and Concentrator Kit (Zymo Research) allowing the recovery of total RNA (>17 nucleotides). RNA quantity was determined using the Nanodrop ND-1000 spectrophotometer (NanoDrop technologies), while RNA quality was assessed via the Agilent 2100 Bioanalyzer using the Agilent RNA 6000 Nano Kit (Agilent Technologies). The removal of genomic DNA was verified via 35 cycles of PCR amplification (5 min at 95°C, followed by 35 cycles of 45 s

104 Chapter 3 at 95°C, 1 min at 55°C, 1 min at 72°C, and a final extension step of 7 min at 72°C) of the uvrD gene (Forward pimer: 5’-GTAGCGAGACCTACAACAAGGTTTC-3’; Reverse primer: 5’-TGGACAGGCGCACTTCCT-3’) of P. aeruginosa.

3.2.6 RNA sequencing and data analysis

Ten μg of extracted total RNA was treated with the Ribo Zero kit (Epicentre) to enrich for mRNA, by removing the 16S and 23S rRNA. Paired-end libraries were prepared according to the TruSeq™ RNA Sample Preparation Guide (Illumina). The library preparation and Illumina RNA sequencing was performed by BaseClear (Leiden, the Netherlands). Obtained reads were aligned using the BWA software using default parameters (Li & Durbin, 2010). Raw counts per gene were calculated based on the genome annotation of Pseudomonas aeruginosa CF_PA39 (accession number NZ_JDVE00000000). Reads were allowed to map 50 bp upstream of the start codon or 50 bp downstream of the stop codon. Reads mapping to ribosomal or transfer RNA were removed from the raw counts data to prevent bias in detecting differential expression. Differential expression was calculated using the edgeR package (version 3.2.4) (Robinson et al., 2010) in BioConductor (release 3.0, R version 3.1.2), resulting for each gene in a fold change and a corresponding p-value corrected for multiple testing. Genes found to be differentially expressed (≥1.50-fold; p<0.05, FDR<0.05) were blasted using the Pseudomonas Genome Database (Winsor et al., 2011) in order to obtain information about their presence in other P. aeruginosa genomes, the functional classes they belong to, their genetic organization, and related literature.

For the detection of new sRNAs located in intergenic regions the RNA-Seq reads were mapped via the READemption pipeline 0.3.4 (using segemehl .0.1.7 (Forstner et al., 2014)) followed by coverage calculations. The sRNA prediction was conducted using ANNOgesic (Yu et al., unpublished). For this, positions with a coverage higher than 5 reads were combined to transcripts (gaps of maximum 5 nt with lower coverage were accepted). Transcript shorter than 20 nt or longer than 500 nt and transcripts that were overlapping in sense or anti-sense with known genes were discarded. For the remaining candidates the secondary structure was predicted by RNAfold (part of the Vienna package (Lorenz et al., 2011)) and only sRNA candidates that were able to form a secondary structure were kept. The sRNAs were aligned against the sRNA entries in BSRD (Bacterial Small regulatory RNA Database) (Li et al., 2013) with BLAST 2.2.28+ (Camacho et al., 2009) but non showed significant homology (i.e. an e-value below 0.0001).

105 Effect of shear stress on P. aeruginosa in vitro

3.2.7 Reverse transcription

cDNA was prepared using the iScript cDNA synthesis kit (Biorad), starting from 1 µg of DNA-free total RNA. The resulting cDNA was diluted 5x prior to use in qRT-PCR. In order to verify efficient conversion of RNA to cDNA, a PCR was performed using primers (forward 5’-ATGAACAACGTTCTGAAATTCTCTGCT-3’ and reverse 5’- CTTGCGGCTGGCTTTTTCCAG-3’) that allow amplification of the oprI gene in a Thermocycler (TC-412-Techne) with the following thermocycling conditions: 94°C for 5 min, followed by 35 cycles at 94°C for 45 seconds, 55°C for 45 seconds, 1 min at 72°C and a final elongation step at 72°C for 10 min.

3.2.8 Quantitative real-time PCR

Since the aim of this study was to identify genes that are differentially regulated in response to shear stress and that could later serve as marker genes in vivo, we decided to design qRT- PCR primers that are able to anneal to target genes in various P. aeruginosa strains without any mismatches. All primers used in qRT-PCR amplification were designed via Primer3 (Rozen & Skaletsky, 2000) and are listed in Table S3.4. Amplification was performed in a 96- well plate, in which each well contained 25 μl of volume consisting of 9.5 μl nuclease-free water, 1 µl of a each primer (10 μM), 12.5 μl of the 2x iQTM SYBR® Green supermix (Biorad) and finally 1 μl of template cDNA (5x diluted). The PCR amplification was performed via the iQ2 real-time PCR detection system (BioRad) using the following program: an initial cycle at 95°C for 3 min for denaturation and enzyme activation, then 40 cycles of 95°C for 10 s and 55°C for 60 s. Finally, melt curves were determined to identify primer dimer formation. qRT-PCR results were normalized against the housekeeping gene oprI encoding the major outer membrane lipoprotein I. Fold changes were calculated using the Livak method (Livak & Schmittgen, 2001). The experiment was performed in biological and technical triplicates.

3.2.9 Scanning electron microscopy

Bacterial cells were recovered from the RWV bioreactor as mentioned before (section 3.2.4) and subsequently 1000x diluted in PBS via serial dilutions. Next, 500 μl of each dilution was concentrated on a Nuclepore TrackEtch Polycarbonate membrane filter with 0.2 μm pore size, followed by two fixation steps of 20 min with a 3% gluteraldehyde (w/v) in 0.15 M

106 Chapter 3 cacodylate solution. In a following step, the membrane was washed three times with the 0.15 M cacodylate wash solution and stored overnight at 4°C. Dehydration was obtained by rinsing the filter surface with an ascending series of ethanol (30, 50, 70, 90, 95, 100% (v/v) in Milli-Q water). The final 100% ethanol solution was replaced three times. Next, the ethanol solution was replaced with hexamethyldisilazane and this was repeated three times. Finally, the membrane filters were air-dried at room temperature in a desiccator overnight, taped on a messing stub using carbon tape, and ultimately sputter coated with gold particles. SEM analysis was performed on a JEOL JSM-840 (Jeol Ltd.) equipped with a secondary electron and backscatter electron detector (point electronic GmbH) at a working distance of 37 mm and a 15-kV acceleration.

3.2.10 Quantification of 3-oxo-C12-HSL

Supernatant of RWV cultures was obtained via centrifugation of 25 ml of culture medium at 10,000 g for 10 min and stored as 2 ml aliquots at -20°C prior to use. An overnight culture of the E. coli indicator strain MH155 (Table S3.3) that produces GFP when sensing 3-oxo-C12-

HSL (Hentzer et al., 2003) was diluted to OD600nm = 1.0. Next, 100 µl of this diluted overnight culture and 200 µl of RWV culture supernatant were added to 5 ml of LB medium and incubated at 37°C for 24h with continuous shaking at 200 rpm. Finally, 200 µl of the bacterial culture was transferred to a 96-well plate and the OD595nm as well as the relative fluorescence units (RFU) obtained at an excitation and emission wavelength of 485 nm and 527 nm, respectively, were measured using a Fluoroskan Ascent fluorometer (Thermo

Scientific). Background-corrected RFU values were normalized to the OD595nm of the reporter strain and to the log10 CFU/ml for each replicate. The experiment was performed in biological and technical triplicates.

3.2.11 Determination of elastase production

The amount of extracellular elastase was assessed by means of a Congo Red colorimetric assay (Aendekerk et al., 2002). One hundred microlitres of defrosted supernatant was added to glass test tubes containing 10 mg of elastin Congo Red (Sigma-Aldrich) in 900 µl of 0.1 M Tris-HCl (pH = 7.2). After 6 h of incubation at 37°C, the tubes were centrifuged (10 min, 10,000g) and 250 µl of supernatant was pipetted in a 96-well microtiter plate. Finally, the optical density at 495 nm was measured using a Multiskan Spectrum spectrophotometer

107 Effect of shear stress on P. aeruginosa in vitro

(Thermo Scientific). Finally, background-corrected OD495nm values were normalized to the log10 CFU/ml for each replicate. The experiment was performed in biological and technical triplicates.

3.2.12 Qualitative determination of short-chain N-acylhomoserine-lactone production

LB plates (25 ml of LB agar) were covered with a 5 ml 0.6% LB top agar layer and 100 µl of the indicator strain Chromobacterium violaceum CV026 (Table S3.3) that produces violacein, characterized by a deep purple color, when sensing short-chain N-acylhomoserine-lactones (AHLs) (McClean et al., 1997). After allowing this top agar layer to dry for 10 minutes in a laminar flow cabinet, a diffusion disk (Oxoid) containing 70 µl of culture supernatant was applied on top of this soft agar by gently pressing on top of the disk by means of a forceps. Finally, the plates were incubated at 30°C for 24 hours. The experiment was performed in biological and technical triplicates.

3.2.13 Statistical analyses

All experiments were performed in triplicate. A paired, two-tailed Student’s t test was applied on data obtained from the bacterial count, 3-oxo-C12-HSL, and elastase experiments in order to detect differences between the LS and HS conditions. p-values <0.05 were considered to be statistically significant. Clusters of gene ontology (COG) classes enriched in either LS or HS were identified by means of hypergeometric distribution.

3.3 Results and discussion

3.3.1 Two different colony morphologies were identified after growth of P. aeruginosa CF_PA39 in artificial sputum medium

Growth of the CF lung-adapted transmissible P. aeruginosa CF_PA39 strain in ASM in the RWV bioreactor resulted in the formation of a non-mucoid and a mucoid colony morphotype (Figure 3.1A). More specifically, non-mucoid colonies were more abundant than mucoid colonies under both low (LS) and high (HS) fluid shear conditions, while no statistically significant difference in the number of non-mucoid or mucoid colonies was observed when

108 Chapter 3 comparing both culture conditions (Figure 3.1B). The ratios of non-mucoid over mucoid colonies recovered from LS and HS conditions were 1.78 and 1.92, respectively. Although the overall viable count was slightly higher under the HS condition (3.50 x 109 ± 1.41 x 109 CFU/ml) compared to the LS condition (2.44 x 109 ± 6.36 x 108 CFU/ml), the difference between both conditions was not significant.

Figure 3.1. A. Phenotype of colonies that were recovered from the RWV bioreactor after 24h of growth in ASM and subsequently plated on LB medium. Mucoid colonies are indicated by a black arrow, while non-mucoid colonies are indicated by a white arrow. B. Quantification of bacteria that were recovered from the RWV bioreactor after 24h of growth in ASM and subsequently plated on LB medium. LS, low fluid shear. HS, high fluid shear. NM, non-mucoid colonies. M, mucoid colonies. NS, not statistically significant (p>0.05).

A similar observation has been done by Woo and colleagues (Woo et al., 2012), where a chronic P. aeruginosa CF isolate was inoculated in a flow-cell and four extra colony morphotypes were obtained from the biofilm effluent next to the mucoid wild type after nine days of growth. When the genetic basis of this short-term diversification was revealed, it was found that many of the phenotypic variants from the dispersal population had acquired mutations in genes involved in alginate biosynthesis and c-di-GMP metabolism (McElroy et al., 2014). These results indicate that the occurrence of the non-mucoid phenotype under both LS and HS might be the result of a genetically diversifying population during growth in ASM, resembling the CF lung conditions. The non-mucoid phenotype might represent a more motile dispersal variant that is able to colonize a newly developed niche as a consequence of biofilm development and/or maturation.

109 Effect of shear stress on P. aeruginosa in vitro

3.3.2 High fluid shear levels preclude the formation of self-aggregating biofilms.

Scanning electron microscopy showed that for each of the biological replicates, the LS condition was characterized by the presence of numerous clusters of closely associated cells (Figure 3.2A-D) next to planktonic cells. The size of these clusters varied from small (containing dozens of cells, Figure 3.2A-B) to extremely large (thousands of cells, Figure 3.2C). Inside these clusters, frequent cell-to-cell contacts were observed (Figure 3.2B, D). In contrast to these findings, no clusters of P. aeruginosa cells could be observed for samples under the HS condition (Figure 3.2E-H). Consequently, all cells adopted a unicellular planktonic lifestyle under the HS condition (Figure 3.2F, H). Remarkably, elongated P. aeruginosa cells were regularly observed in both LS (Figure S3.1A) and HS (Figure S3.1B) conditions. However, because of the many cell clusters found under the LS condition, it was not possible to quantify this phenotype. In summary, these observations indicate that biofilm formation by P. aeruginosa CF_PA39 was dependent on the prevailing shear stress.

Figure 3.2. Scanning electron micrographs of P. aeruginosa CF_PA39 grown under low fluid shear (A-D) or high fluid shear conditions (E-H). B and D represent magnifications of the areas indicated by the white boxes in A and C, respectively. F and H represent magnifications of the areas indicated by the white boxes in E and G, respectively. The magnification and scale bars are shown below each picture. Images are representative for different biological repeats.

110 Chapter 3

3.3.3 Effect of shear stress on the transcriptome of P. aeruginosa CF_PA39 grown in artificial sputum medium

In order to detect differentially expressed genes in response to shear stress during growth of P. aeruginosa CF_PA39 in ASM in the RWV bioreactor, an RNA sequencing approach was followed. A complete list of the transcriptomic data is provided in Table S3.1. Differential shear stress induced subtle differences in P. aeruginosa CF_PA39 gene expression as the most up-regulated gene (trpB) under the LS condition was only 2.89-fold higher expressed than under the HS condition, while the most down-regulated gene (PA1923) was only 2.57- fold less expressed (Table S3.1). Because of these subtle differences, all genes that were ≥1.50-fold differentially regulated and statistically significant (p<0.05, FDR<0.05) were included. In total, 133 and 107 genes were up- and down-regulated under the LS condition compared to the HS condition, respectively. In order to look at the global response to shear stress, differentially expressed genes were first grouped according to their clusters of orthologous groups (COG) class and the most affected functional classes were identified. Genes involved in nucleotide transport and metabolism were overrepresented among both up- and down-regulated genes under LS conditions (Table S3.2). Besides this functional class, mainly genes involved in transcription and especially translation were up-regulated under LS, while those involved in carbohydrate transport and metabolism were down-regulated (Table S3.2). Although this approach allows the identification of majorly affected functional classes, it has a limited resolution. Therefore, all differentially expressed genes were blasted using the Pseudomonas Genome Database and classified into specific functional classes (PseudoCAP, COG, KEGG) based on the available gene information (Table S3.1). Based on these data, a comparison of the more specific functional classes that were up-regulated or down-regulated under LS could be made (Table 3.1). The majority of the up-regulated genes under LS (without considering the hypothetical function class) belonged to translation (13.53%) and transcriptional regulation (9.77%) as predicted in the previous analysis (Table 3.1). Next to these two large functional classes, genes involved in stress response (4.51%), denitrification (3.76%), glycerol metabolism (3.76%), alginate biosynthesis (3.01%), glycine betaine biosynthesis (1.50%), cell division (1.50%), tryptophan biosynthesis (1.50%) and type II secretion (1.50%) were identified among the ≥1.50-fold up-regulated genes under LS, and were not present in the list of ≥1.50-fold down-regulated genes. In contrast, genes involved in phenazine biosynthesis (2.80%), type VI secretion system (2.80%), cell motility (2.80%),

111 Effect of shear stress on P. aeruginosa in vitro lipid transport and metabolism (2.80%), Psl exopolysaccharide biosynthesis (1.87%), and secreted factors (1.87%) were exclusively found among down-regulated genes (Table 3.1).

Table 3.1. Comparison of the proportion of functional classes that were represented amongst ≥1.50-fold up- regulated or down-regulated genes under low fluid shear versus high fluid shear conditions.

Up-regulated Down-regulated Functional class Number of % Number of % genesa genesb Alginate biosynthesis 4 3.01 0 0.00 Amino acid transport and metabolism 4 3.01 14 13.08 Antibiotic resistance and susceptibility 3 2.26 2 1.87 Aromatic compound catabolism 1 0.75 0 0.00 Carbohydrate transport and metabolism 6 4.51 5 4.67 Carbon compound catabolism 0 0.00 1 0.93 Cell cycle control, cell division, chromosome 2 0 1.50 0.00 partitioning Cell motility 0 0.00 3 2.80 Cell wall/membrane/envelope biogenesis 5 3.76 2 1.87 Coenzyme transport and metabolism 5 3.76 2 1.87 Denitrification (anaerobic respiration) 5 3.76 0 0.00 Energy production and conversion 12 9.02 12 11.21 Glycerol metabolism 5 3.76 0 0.00 Glycine betaine biosynthetic process from choline 2 1.50 0 0.00 Glycine betaine catabolism 0 0.00 1 0.93 Glyoxylate and dicarboxylate metabolism 1 0.75 0 0.00 Inorganic ion transport and metabolism 4 3.01 1 0.93 Intracellular trafficking, secretion, and vesicular 3 2 2.26 1.87 transport Iron metabolism 2 1.50 0 0.00 Iron uptake 1 0.75 0 0.00 Lipid A biosynthetic process 1 0.75 0 0.00 Lipid transport and metabolism 0 0.00 3 2.80 Nucleotide transport and metabolism 6 4.51 3 2.80 Phenazine biosynthesis 0 0.00 3 2.80 Phosphonate metabolism 0 0.00 1 0.93 Posttranslational modification, protein turnover, 5 2 3.76 1.87 chaperones Psl biosynthesis 0 0.00 2 1.87 Replication, recombination and repair 2 1.50 2 1.87 Rhamnolipids biosynthesis 0 0.00 1 0.93 Secreted Factors (toxins, enzymes, ...) 0 0.00 2 1.87 Signal transduction mechanisms 1 0.75 2 1.87 Stress response 6 4.51 0 0.00 TonB-dependent receptors 0 0.00 1 0.93

112 Chapter 3

Up-regulated Down-regulated Functional class Number of % Number of % genesa genesb Transcriptional regulation 13 9.77 5 4.67 Translation, ribosomal structure and biogenesis 18 13.53 4 3.74 Transport of small molecules 1 0.75 2 1.87 Tryptophan biosynthesis 2 1.50 0 0.00 Type II secretion system 2 1.50 0 0.00 Type IV pilus biogenesis 0 0.00 1 0.93 Type VI secretion system 0 0.00 3 2.80 Unknown 36 27.07 39 36.45 aOn a total of 133 genes. bOn a total of 107 genes. LS, low fluid shear. HS, high fluid shear.

The expression of a selection of genes (most differentially regulated and/or part of majorly affected functional classes) was confirmed via qRT-PCR (Table 3.2-3.3). Overall, gene expression results obtained with RNA sequencing and qRT-PCR overlapped for the selected genes (Table 3.2-3.3). Besides the genes belonging to the aforementioned functional classes, the differential expression of the lasI gene, encoding the 3-oxo-C12-HSL synthase, and lasB, encoding elastase, was confirmed as both genes were slightly (<1.5-fold, p<0.05) down- regulated under the LS condition (Table 3.3). Interestingly, numerous genes that were up- regulated under the LS condition were previously identified as being up-regulated in CF sputum and/or chronic infection compared to planktonic cultures of the same P. aeruginosa strains (Table S3.1) (Bielecki et al., 2013; Son et al., 2007). Furthermore, a large number of genes that were differentially regulated, belong to the same operon (Table 3.2-3.3, Table S3.1). When taking a lower cutoff (≥1.20-fold differential expression; p<0.05, FDR<0.05), a number of operons containing genes of majorly affected functional classes can be distinguished (Figure 3.3).

113 Effect of shear stress on P. aeruginosa in vitro Table 3.2. Selection of genes that were up-regulated under the low fluid shear condition compared to the high fluid shear condition according to RNAseq analysis and whose expression was confirmed via qRT-PCR analysis.

Gene Product Function Operon FC RNAseq FC qRT-PCR PA0036 (trpB) Tryptophan synthase beta chain (EC Tryptophan biosynthesis; Amino acid transport PA0036 (trpB)-PA0035 (trpA) 2.89 4.50 ± 1.58 4.2.1.20) and metabolism PA5374 (betI) HTH-type transcriptional regulator BetI Transcriptional regulation; Glycine betaine PA5374 (betI)-PA5372 (betA) 2.85 3.52 ± 0.60 biosynthetic process from choline; Stress response PA3584 (glpD) Glycerol-3-phosphate dehydrogenase Glycerol metabolism; Energy production and 2.29 2.92 ± 0.31 conversion PA0523 (norC) Nitric oxide reductase subunit C (EC Denitrification (anaerobic respiration) PA0523 (norC)-PA0525 2.21 1.85 ± 0.13 1.7.99.7) PA0579 (rpsU) SSU ribosomal protein S21p Translation, ribosomal structure and biogenesis PA0579 (rpsU)-PA0578 2.17 2.95 ± 0.73 PA3391 (nosR) Nitrous oxide reductase maturation protein Denitrification (anaerobic respiration) PA3391 (nosR)-PA3396 (nosL) 1.94 2.84 ± 0.82 NosR PA3550 (algF) Alginate O-acetyltransferase AlgF Alginate biosynthesis PA3540 (algD)-PA3551 (algA) 1.83 1.94 ± 0.29 PA4481 (mreB) Rod shape-determining protein MreB Cell cycle control, cell division, chromosome PA4481 (mreB)-PA4479 (mreD) 1.80 2.26 ± 0.30 partitioning PA3551 (algA) Mannose-1-phosphate guanylyltransferase Alginate biosynthesis PA3540 (algD)-PA3551 (algA) 1.65 2.01 ± 0.13 (GDP) (EC 2.7.7.22) / Mannose-6- phosphate isomerase (EC 5.3.1.8) FC, fold change. EC, enzyme class. HTH, helix-turn-helix.

114 Chapter 3

Table 3.3. Selection of genes that were down-regulated under the low fluid shear condition compared to the high fluid shear condition according to RNAseq analysis and whose expression was confirmed via qRT-PCR analysis.

Gene Product Function Operon FC RNAseq FC qRT-PCR PA0121 Transcriptional regulator, GntR family Transcriptional regulation -2.39 -2.19 ± 0.50 PA5481 Inhibitor of vertebrate lysozyme precursor Periplasmic protein PA5482-PA5481 -2.20 -2.26 ± 1.02 PA3372 Metal-dependent hydrolases of the - Phosphonate metabolism PA3384 (phnC)-PA3372 -2.10 -5.54 ± 2.95 lactamase superfamily I; PhnP protein PA1922 Colicin I receptor precursor TonB-dependent receptors; Inorganic ion PA1922-PA1925 -2.07 -2.37 ± 0.46 transport and metabolism PA4599 (mexC) Multidrug efflux RND membrane fusion Antibiotic resistance and susceptibility PA4599 (mexC)-PA4597 (oprJ) -2.04 -2.39 ± 0.94 protein MexC PA4171 ThiJ/PfpI family protein Unknown PA4171-PA4172 -1.99 -3.40 ± 0.52 PA4600 (nfxB) Transcriptional regulator NfxB Transcriptional regulation; Antibiotic resistance -1.65 -1.65 ± 0.64 and susceptibility PA5040 (pilQ) Type IV pilus biogenesis protein PilQ Type IV pilus biogenesis; Cell motility PA5044 (pilM)-PA5040 (pilQ) -1.59 -1.28 ± 0.32 PA4190 (pqsL) 2-octaprenyl-3-methyl-6-methoxy-1,4- Energy production and conversion -1.50 -1.14 ± 0.48 benzoquinol hydroxylase (EC 1.14.13.-) PA4209 (phzM) Phenazine-specific methyltransferase Phenazine biosynthesis -1.45 -2.14 ± 0.77 PhzM PA3724 (lasB) Vibriolysin, extracellular zinc protease Secreted Factors (toxins, enzymes, alginate) -1.31 -1.21 ± 0.16 (EC 3.4.24.25); Pseudolysin, extracellular zinc protease (EC 3.4.24.26) PA1432 N-acyl-L-homoserine lactone synthetase Quorum sensing -1.28 -1.07 ± 0.27 LasI FC, fold change. RND, resistance-nodulation-cell division superfamily. EC, enzyme class.

115 Effect of shear stress on P. aeruginosa in vitro

3.3.3.1 Genes up-regulated under low fluid shear

Under LS, the algD (alginate biosynthesis), norCBD, nosRZDFYL (denitrification), betIBA (glycine betaine biosynthesis from choline), mreBCD (cell cycle control, cell division, and chromosome partitioning) and trpBA (tryptophan biosynthesis) operons contain at least two ≥1.50-fold up-regulated genes (Figure 3.3A). Accordingly, the P. aeruginosa CF_PA39 strain formed robust biofilms under the LS condition, in agreement with our previous study (P. aeruginosa PAO1 grown in LB in LS) (Crabbé et al., 2008). In contrast, under a high fluid shear condition, no clusters of cells could be observed as all cells appeared to be unicellular and planktonic (Figure 3.2). These phenotypic results matched with the transcriptomic data since the exopolysaccharide alginate (encoded by the algD operon) is involved in P. aeruginosa biofilm formation (Hentzer et al., 2001; Nivens et al., 2001; Wei & Ma, 2013). Next to alginate biosynthesis, a prominent role for genes involved in denitrification was observed under the LS condition. More specifically, norB, norC (part of the norCBD operon), nosR and nosY (part of the nosRZDFYL operon) were identified. This indicates that the accumulation of nitric oxide (NO) is avoided during growth under LS, since the nor (NO reductase) and nos (nitrous oxide reductase) genes are involved in the reduction of NO to N2. Furthermore, the gene encoding the transcriptional regulator Dnr (Dissimilative Nitrate Respiration regulator) was found to be up-regulated under the LS condition. Dnr is activated in the presence of NO and the expression of nirSMC (part of the nirSMCFDLGHJEN operon), nirQ, norCB and nosZ was found to be dependent on this regulator during growth in anaerobic conditions (Trunk et al., 2010). In addition, the anr (Fumarate and nitrate reduction regulatory protein) gene, that orchestrates the P. aeruginosa response to anaerobic conditions, was slightly (1.24-fold change), but significantly (p<0.05, FDR<0.05), up-regulated during growth under LS. These data indicate that P. aeruginosa CF_PA39 experiences microaerobic conditions during growth under LS in ASM. In fact, our previous study showed that growth of P. aeruginosa PAO1 in LB led to a decreased oxygen transfer rate under LS compared to a higher shear control (Crabbé et al., 2010). Such microaerobic or anaerobic conditions could occur because of the low mixing capacities of the viscous ASM under LS conditions and/or locally inside the alginate-enclosed biofilms. A recent study strongly indicated that P. aeruginosa respires anaerobically in CF sputum via denitrification since an initial increase in nitrous oxide (N2O) was followed by a decrease after six hours of monitoring in freshly expectorated CF sputum (Kolpen et al., 2014). Moreover, the addition of nitrate to LB medium, yielding physiological nitrate levels, resulted in increased growth rates of P.

116 Chapter 3 aeruginosa PAO1 as well as clinical P. aeruginosa CF isolates under anoxic conditions, comparable to those observed in CF lungs and sputum (Line et al., 2014). Genes involved in denitrification were not found to be differentially regulated during growth of P. aeruginosa PAO1 in LB medium under LS (Crabbé et al., 2010), which was most probably due to the low nitrate levels in this medium. However, a recent study has shown that nitrate levels in ASM (identical to the ASM used here, except for the addition of 3 µg/ml ferritin) are comparable to those found in CF sputum samples (in the low millimolar range) (Quinn et al., 2014).

Figure 3.3. Overview of the genetic regions that contain up-regulated (A) or down-regulated (B) genes under low fluid shear versus high fluid shear conditions for the key affected functional classes. X and Y represent the PA14 genes PA14_33980 and PA14_33970 that are not present in the PAO1 genome, respectively. LS, flow fluid shear. HS, high fluid shear. The // symbol indicates that this gene is located at a distant position in the genome. All adjacent genes that are transcribed in the same direction are considered to constitute an operon here.

Another important pathway that was induced under the LS condition was the biosynthesis of glycine betaine. This compound can be synthesized from choline, via the intermediate betaine aldehyde, by means of the enzymes BetA (choline dehydrogenase) and BetB (betaine

117 Effect of shear stress on P. aeruginosa in vitro aldehyde dehydrogenase). Under LS conditions, the betI (encoding a transcriptional repressor whose repression activity is abolished in the presence of choline) and betB genes were among the most highly up-regulated genes. In contrast to this finding, six out of eight genes involved in catabolism of glycine betaine were ≥1.20-fold down-regulated (p<0.05, FDR<0.05) under LS. Interestingly, the betI and betB genes were found among the highest up-regulated genes in one study in which the gene expression of P. aeruginosa in CF sputum in vivo was compared to that of a planktonic grown pool of isogenic isolates (Son et al., 2007), and in another study in which P. aeruginosa gene expression shared by three chronic conditions (tumour, burn wound and CF) was compared to the planktonic stage of growth (Bielecki et al., 2013). From our gene expression results, it can be deduced that the choline degradation pathway is shifted towards glycine betaine accumulation and not towards energy production via further catabolism of this compound, which is another option for P. aeruginosa as discussed by Wargo (Wargo, 2013). Glycine betaine is an important osmoprotectant for many bacteria and is produced in response to stress conditions. We have developed a hypothetical model for the role of this important molecule under the LS and CF condition (Figure 3.4). In robust biofilms, in which P. aeruginosa cells are enclosed by dense layers of alginate, a hyperosmotic condition could occur because of the local accumulation of cellular debris, extracellular DNA and secreted factors such as secondary metabolites and enzymes. In order to respond to this situation, P. aeruginosa would import choline from the environment and catabolizes it to obtain glycine betaine. Indeed, the betT1 and betT3 genes, which encode high affinity choline transporters, were significantly (≥1.30-fold; p<0.05, FDR<0.05) up-regulated under LS, in addition to the glycine betaine biosynthetic genes and in contrast to the glycine betaine catabolic genes (Table S3.1). A probable explanation why genes involved in glycine betaine biosynthesis were not up-regulated under LS in the previous RWV study (Crabbé et al., 2010) is most probably the absence of a choline source in the growth medium. The ASM used in the current study contained egg yolk, which is a major source of (phosphatidyl) choline. In their study Son et al. (Son et al., 2007) stated that (lung) phospatidylcholine induced the expression of genes involved in fatty acid degradation, choline degradation and glycerol metabolism. Furthermore, it was shown that several lipases and phospholipases were induced in the presence of phosphatidylcholine. These phospolipases could cleave phosphatidylcholine, generating fatty acids, choline and glycerol. Although increased expression of the phospholipase genes was not observed under the LS condition, genes involved in glycerol uptake (glpF) and metabolism (glpK, glpD) were ≥1.50-fold up-regulated under LS. Recently, it has been shown that CF-adapted P. aeruginosa isolates utilize glycerol

118 Chapter 3 more efficiently as a carbon source than non-adapted isolates (Daniels et al., 2014). In addition, mutation of the glpD gene (encoding a glycerol-3-phosphate dehydrogenase) resulted in lower levels of alginate production, indicating that the glycerol catabolic pathway is indispensable for full virulence of P. aeruginosa in chronic CF infection (Daniels et al., 2014).

Figure 3.4. Hypothetical model of the adaptation of P. aeruginosa CF_PA39 to the low fluid shear condition at the level of osmoprotection. Under the low fluid shear condition, P. aeruginosa forms biofilms of closely associated cells that are surrounded by alginate layers. The production of several secreted molecules as well as extracellular DNA and cell debris that result from dying cells, create a local hyperosmotic environment. In order to protect itself against this hyperosmotic condition, P. aeruginosa imports choline via the BetT1 and BetT3 transporters, thereby releasing repression of the betIBA operon (and of the choline transporter genes betT1 and betT3) by the BetI repressor and switches on the genes that are required for glycine betaine biosynthesis using choline as a substrate. At the same time, the majority of genes involved in the catabolism of glycine betaine to glycine are down-regulated, leading to an accumulation of the osmoprotectant glycine betaine. Genes highlighted in green were found to be up-regulated, whereas genes highlighted in red were found to be down-regulated under low fluid shear versus high fluid shear conditions. The genes involved in the biosynthetic process from choline to glycine betaine are shown below. The glycine betaine biosynthetic genes betA and betB are shown in green, the betI repressor gene is shown in orange, and the choline transporter genes are shown in blue. The BetI protein is represented by an orange open cylindrical shape. The // symbol indicates that this gene is located at a distant position in the genome. IM, inner membrane.

119 Effect of shear stress on P. aeruginosa in vitro

Besides the role of glycine betaine biosynthesis and glycerol metabolism, cell division appears to play an important role under LS since the mreBCD genes were up-regulated under this condition. These genes are involved in maintenance of the characteristic rod-like cell shape of bacteria through cell division (Errington, 2015). Recently, it has also been observed that mreB (the most highly up-regulated gene in the operon) is involved in osmotolerance in Escherichia coli (Winkler et al., 2014). Furthermore, mutation of this gene appears to result in impairment of the correct localization of type IV pili and hence greatly affects motility of P. aeruginosa (Cowles & Gitai, 2010). In the present study we have regularly observed elongated cells, most probably as a defect in cell division since the characteristic shape of the septum is still visible on the surface of these cells (Figure S3.1). Such elongated cells have been observed in another study when P. aeruginosa PAO1 was grown under anaerobic conditions (Yoon et al., 2011). Furthermore, in the same study it was shown that upon mutation of nirS, required for the reduction of nitrite to NO, or addition of a NO-antagonist to the culture medium, this phenotype was no longer observed and less robust biofilms were formed. These data indicate that P. aeruginosa responds to anaerobic conditions by changing its cell shape. In addition, two studies have described the elongation/filamentation of Pseudomonas putida cells when this bacterium was grown at low, but not at high shaking speed (Crabbé et al., 2012; Jensen & Woolfolk, 1985). Proteomic analysis of P. putida cultures grown at low shaking speed indicated that the elongated cell shape was most probably adopted as a survival strategy in the oxygen-limited conditions that are inherent to a lower shaking speed (Crabbé et al., 2012). Although we observed elongated cells under both LS and HS conditions, it was impossible to quantify this phenotype, since many bacterial cells under LS were grouped together in tight clusters.

Finally, the two genes of the trpBA operon, necessary for tryptophan biosynthesis, were among the four most up-regulated genes under LS (≥2.0-fold). In agreement with this observation, two other RWV studies have identified a role for tryptophan metabolism under LS. More specifically, the trpD gene was found among the 68 up-regulated genes when S. enterica serovar Typhimurium was grown under LS compared to normal gravity (Wilson et al., 2002). In a second study, a tryptophan permease-encoding gene was indispensable for the increased adherence of adherent-invasive E. coli to cell cultures under LS conditions (Allen et al., 2008). A more recent study showed that up-regulation of the trpD gene under LS compared to normal gravity is not ubiquitous among all genera of the Enterobacteriaceae family (Soni et al., 2014). The in vivo importance of the trpBA operon is less obvious since

120 Chapter 3 these genes were up-regulated in CF sputum in one study (Son et al., 2007), while being down-regulated in two others in either CF sputum (Palmer et al., 2005) or CF-sputum- containing medium (Fung et al., 2010), compared to laboratory media.

3.3.3.2 Genes down-regulated under low fluid shear

Among the genes down-regulated under the LS condition, many genes of unknown function were found, in addition to genes involved in cell motility, Psl biosynthesis, phenazine biosynthesis, type VI secretion system and multidrug resistance. All genes of the PA1922- PA1925, PA3370-PA3371, and PA5482-PA5481 operons were found to be ≥1.50-fold down- regulated, although their function is mentioned as being unknown (Figure 3.3B). In fact, PA5481 (Ivyp2) appears to control the autolytic activity of lytic transglycosylases in bacteria that do not O-acetylate their peptidoglycan, rather than inhibiting lysozymes (Clarke et al., 2010). In addition, the mexC-oprJ and PA2365-PA2374 operons encoding the MexCD-OprJ multidrug efflux pump and the HSI-III type VI secretion system, respectively, contain several ≥1.20-fold down-regulated genes (Figure 3.3B).

PA1922, a TonB-dependent receptor gene that is part of the P. aeruginosa core genome (Cornelis & Bodilis, 2009), shares homology with the colicin I receptor of Escherichia coli (Nau & Konisky, 1989). Previously, we have shown that deletion of the TonB-dependent receptor genes occurs frequently during adaptation of P. aeruginosa to the CF lung environment (Dingemans et al., 2014). Two divergent hypotheses can explain this observation. First of all, the biofilm lifestyle of P. aeruginosa in the CF lung could reduce selection pressure, leading to the loss of these genes. Secondly, pyocins, known to enter P. aeruginosa cells via TonB-dependent receptors (Baysse et al., 1999; de Chial et al., 2003; Denayer et al., 2007; Elfarash et al., 2012; Elfarash et al., 2014), could select for their deletion. Therefore, TonB-dependent genes may rather be down-regulated in biofilm-like than planktonic conditions due to rewiring of the regulatory networks that control expression of these genes. In agreement with this, no TonB-dependent receptor-encoding genes were detected among the up-regulated genes under the LS condition.

Although the function of the genes in the PA5482-PA5481 operon remains to be elucidated, they have been associated with acute infection, since their expression was elevated in a non- CF pneumonia isolate of the Liverpool Epidemic Strain (LES) compared to a chronic CF LES isolate during growth in LB medium (Salunkhe et al., 2005).

121 Effect of shear stress on P. aeruginosa in vitro

With regard to genes of known function, it appeared that those involved in motility were down-regulated under the LS condition. These genes are associated with the planktonic lifestyle of P. aeruginosa and have been shown to be prone to deletion during adaptation to the CF lung environment (Hauser et al., 2011). More specifically, the pilQ gene, necessary for the formation of type IV pili, was significantly down-regulated under LS. Mutations in this gene have been frequently observed during P. aeruginosa colonization of the CF lung (Chang et al., 2007; Marvig et al., 2015). Several genes involved in phenazine biosynthesis were down-regulated under LS conditions. This finding is in contrast with a previous study (Hunter et al., 2013) where it was shown that the concentration of phenazines is positively correlated with the presence of ferrous iron in CF sputum. Ferrous iron concentrations were found to be higher in sputum from patients with deteriorating lung functions, most probably because of the microaerobic or anaerobic conditions encountered by P. aeruginosa in this environment. The results from the latter study suggest that phenazine biosynthesis is associated with the biofilm rather than the planktonic lifestyle of P. aeruginosa. However, another study (Recinos et al., 2012) showed that the phz1 gene cluster is more expressed during planktonic growth, while phz2 almost exclusively contributed to phenazine biosynthesis in colony biofilms. Since only a draft genome of P. aeruginosa CF_PA39 is available, it was not possible to determine if both phenazine gene clusters were down-regulated under the LS condition. Nevertheless, the phzM and phzS genes, that are exclusive to the phz1 gene cluster, were down-regulated 1.45-fold and 1.30-fold under LS, respectively, indicating that phz1 is less expressed under this condition.

Interestingly, two genes of the psl gene cluster, that was found to be up-regulated during growth of P. aeruginosa PAO1 under LS (Crabbé et al., 2008), were ≥1.50-fold down- regulated under LS in the study conducted here. It is noteworthy to mention that the relevance of this up-regulation relative to Psl biosynthesis is low since P. aeruginosa CF_PA39 has a 3376 bp-deletion in this gene cluster, comprising pslAB (Dingemans et al., 2014). However, it is interesting to observe that the genes involved in Psl biosynthesis and alginate production are differentially regulated. Recently, it was shown that the transcription factor AmrZ differentially regulates both operons as it represses the psl operon via binding to the pslA promoter region, while activating the alginate biosynthetic operon (Jones et al., 2013). In P. aeruginosa CF_PA39, the amrZ gene appeared to be slightly, but significantly up-regulated in response to growth under low fluid shear conditions (Table S3.1). However, the regulatory role of AmrZ needs to be investigated in this strain since the psl promoter region has a

122 Chapter 3 completely different genetic architecture compared to that of P. aeruginosa PAO1. Interestingly, several genes of the HSI-III type VI secretion system were significantly down- regulated under the LS condition (Figure 3.3). P. aeruginosa CF_PA39 has a PA14-like HSI- III type VI secretion system including the PA14_33980 and PA14_33970 genes. In PA14, it has been shown that MfvR and LasR negatively regulate HSI-I, while positively regulating HSI-II and HSI-III (Lesic et al., 2009). In accordance to this, mvfR is slightly down-regulated under LS (Table S3.1). Finally, the genes encoding the multidrug efflux pump MexCD-OprJ were down-regulated, as well as its transcriptional repressor gene, nfxB. Although it has been shown that this efflux system can be up-regulated under envelope stress conditions in an algU-dependent way (Fraud et al., 2008), no differential expression of the algU gene was observed in this study (Table S3.1). In a recent paper it was shown that nfxB is essential for optimal fitness of P. aeruginosa PAO1 and PA14 during growth in MOPS-sputum medium (Turner et al., 2015). In addition, this study mentioned that several efflux genes were required for fitness of P. aeruginosa grown in sputum, depending on the strain studied, hence indicating that the role of these genes in the fitness of a specific P. aeruginosa strain depends on the genetic framework they are part of.

3.3.4 Role of small RNAs in the shear stress response

Small RNAs (sRNAs) have been shown to be important regulators of gene expression in many bacteria since they are involved in post-transcriptional modification of mRNA transcripts (Gottesman, 2005; Storz et al., 2011). They can regulate mRNA levels either positively, by enhancing ribosome binding, or negatively, by blocking the ribosome-binding site and/or enhancing RNase E-mediated degradation of the target transcript (Balasubramanian & Vanderpool, 2013). The interaction between the sRNA and a target mRNA can be mediated through the RNA chaperone Hfq and this has been described in E. coli, S. enterica, and P. aeruginosa among others (De Lay et al., 2013; Sonnleitner et al., 2008). Interestingly, hfq as well as other genes belonging to the Hfq regulon have been observed to be majorly involved in the response to LS in different studies involving S. enterica serovar Typhimurium, P. aeruginosa, and S. aureus (Crabbé et al., 2008; Crabbé et al., 2010; Crabbé et al., 2011b; Wilson et al., 2007; Wilson et al., 2008). In accordance with the growth of P. aeruginosa PAO1 under LS in LB medium in our previous study (Crabbé et al., 2010), hfq was slightly up-regulated (1.3-fold; p<0.05, FDR<0.05) under LS in the current study.

123 Effect of shear stress on P. aeruginosa in vitro

The expression of experimentally validated sRNA genes (Tsai et al., 2015) as well as newly predicted small RNA genes based on secondary structure prediction and RNA sequencing analysis for P. aeruginosa CF_PA39 was determined under LS and HS conditions. In total, three sRNA genes were found to be statistically significantly (p<0.05, FDR<0.05) up- regulated under LS (SPA0071, SPA0102, P34) (Table 3.4) while four sRNA genes were down-regulated (srna10, SPA0117, P8, and SPA0003) (Table 3.5). All, but one, of these differentially regulated sRNA genes are present in the PAO1 genome and their position is indicated in Tables 3.4 and 3.5. Both the sRNA gene SPA0102 and its adjacent gene PA3162 are up-regulated under the LS condition. In contrast, the sRNA gene P8 and its adjacent gene PA1030 are both down-regulated. Interestingly, the de novo predicted sRNA gene srna10 was highly down-regulated under the LS condition (2.35-fold), while the genes PA3966 and PA3967 that are in close proximity to this sRNA gene were ≥1.50-fold up-regulated. An interesting future perspective is to test whether these sRNAs regulate their adjacent genes. Interestingly, the sRNA P34, which was more expressed under the LS condition, was also found to be up-regulated in stationary phase planktonic cultures and static biofilms of P. aeruginosa PA14 grown in LB (Dotsch et al., 2012).

Table 3.4. List of small RNA genes that were significantly up-regulated (≥1.50-fold; p<0.05, FDR<0.05) under the low fluid shear condition compared to the high fluid shear condition.

Small RNA gene Length (bp) Position in PAO1 genome Experimentally Fold change RNAseq validateda SPA0071 201 IR PA0805-PA0806 YES 1.94 SPA0102 301 IR PA3162 (rpsA)-PA3163 YES 1.78 (cmk); overlapping cmk P34 399 IR PA5181-PA5182 YES 1.50 IR, intergenic region. aListed by Tsai et al (Tsai et al., 2015).

Table 3.5. List of small RNA genes that were significantly down-regulated (≥1.50-fold; p<0.05, FDR<0.05) under the low fluid shear condition compared to the high fluid shear condition.

Small RNA gene Length (bp) Position in PAO1 genome Experimentally Fold change RNAseq validateda srna10b 202 IR PA3964-PA3965 NO -2.35 SPA0117 201 IR PA3049 (rmf)-PA3050 YES -1.94 (pyrD) ; overlapping both genes P8 78 IR PA1030-PA1031 YES -1.85 SPA0003 137 IR PA2729-PA2730 YES -1.58 IR, intergenic region. aListed by Tsai et al (Tsai et al., 2015). bThis small RNA gene was de novo predicted in this study.

124 Chapter 3

3.3.5 Quorum sensing molecules are slightly higher produced in response to shear stress

In agreement with the RNA sequencing and qRT-PCR data, a subtle, but statistically significant difference was found between 3-oxo-C12-HSL production under LS compared to HS (Figure 3.5A). In addition, elastase production was significantly higher under the HS condition compared to the LS condition (Figure 3.5B). For the qualitative determination of short-chain AHLs (e.g. C4-HSL), the C. violaceum indicator strain CV026, which produces the deep purple compound violacein in the presence of short-chain AHLs, was used. A deep purple color was observed under the HS condition, whereas only a light purple color was present on plates containing supernatant from LS replicates, indicating that the production of short-chain AHLs is higher under HS compared to LS (Figure 3.5C). The elevated production of QS molecules and elastase under HS is in contrast with our previous RWV study (Crabbé et al., 2008) in which P. aeruginosa PAO1 was grown in LB medium. Nevertheless, the different genetic background, the complexity of the medium utilized in this study, and most importantly the highly restructured regulatory network as a consequence of chronic adaptation to the CF lung condition, are suggested to cause this discrepancy. Interestingly, the dispersal variants identified after nine days of growth in the study of Woo et al. (Woo et al., 2012), produced significantly higher amounts of the 3-oxo-C12-HSL and short-chain AHLs as well as elastase. Similarly, the subtle increase in the production of QS molecules and the QS- dependent product elastase observed under the HS condition in this study might be the result of a slightly enriched biofilm dispersal population that has adopted a more planktonic lifestyle. Taken together, these data indicate that a phenotypic variant might have emerged as a result of genotypic diversification, similar to the situation in the CF lung.

125 Effect of shear stress on P. aeruginosa in vitro

Figure 3.5. Production of QS molecules and elastase during growth under different shear stress conditions. A. 3- oxo-C12-HSL production. B. Elastase production. C. Production of short-chain (C4-C8) AHL molecules by P. aeruginosa CF_PA39 grown under low fluid shear (plate shown on the left) or high fluid shear (plate shown on the right) conditions. The picture shown here is representative for all three technical replicates of each biological replicate. LS, low fluid shear. HS, high fluid shear. RFU, relative fluorescence units. *p<0.05. 3.4 Conclusion

In this study, the response of a transmissible CF-adapted P. aeruginosa isolate to differential shear stress was studied at the transcriptomic as well as the phenotypic level in a medium resembling CF sputum. Following an RNA sequencing approach, genes involved in alginate biosynthesis, denitrification, cell shape determination, glycine betaine biosynthesis, glycerol metabolism, and tryptophan biosynthesis were found to be up-regulated under the LS condition, which presumably contributed to the observed biofilm formation. In contrast, genes involved in motility, phenazine biosynthesis, type VI secretion, and multidrug efflux were down-regulated, as well as many hypothetical genes. Overall, these transcriptomic results are in agreement with the SEM observations that revealed the formation of robust biofilms only under the LS condition. Furthermore, a number of sRNA genes might play a role in this switch from the biofilm to the planktonic lifecycle. When comparing the shear stress response

126 Chapter 3 observed in the study conducted here to that of a previous study that used P. aeruginosa PAO1 in LB medium, both similarities and differences were observed. Commonalities between both studies include the formation of self-aggregating biofilms in low shear, as well as the induction of genes involved in alginate synthesis, stress response and low oxygen conditions. Interestingly, the use of a highly adapted P. aeruginosa CF isolate and growth medium directly relevant to the CF lung environment resulted in the induction of additional pathways that have previously been shown to play a role in the metabolism and virulence of this pathogen in the CF patient, that were not found in our previous study using a non-CF P. aeruginosa strain and LB medium. We hypothesize that the combination of physicochemical factors (such as fluid shear, viscosity, and nutritional content) and relevant bacterial genetic background in the present study induced phenotypic and molecular genetic traits in P. aeruginosa that have been observed previously in vivo. Finally, since high fluid shear conditions precluded the formation of CF-like biofilms by P. aeruginosa, the results presented in this study are promising with regard to future in vivo applications that introduce shear stress, including certain types of physical therapy of CF patients, with the aim of disrupting P. aeruginosa biofilms.

127 Effect of shear stress on P. aeruginosa in vitro

3.5 Supplementary data

Table S3.1. Overview of the RNA sequencing results for all CDS and small RNA genes annotated in P. aeruginosa CF_PA39 (xlsx file).

Table S3.2. Analysis of the relative proportion of COG-functional classes that were represented among differentially expressed genes (xlsx file).

Figure S3.1. Elongated P. aeruginosa cells observed under low fluid shear (A) and high fluid shear (B) conditions via scanning electron microscopy. Images are representative for different biological repeats. Elongated cells are indicated by white arrows. Red arrows indicate the putative septum that was formed during cell division.

Table S3.3. Bacterial strains used in this study.

Strain Genotype Reference CF_PA39 P. aeruginosa CF isolate; MLST profile Dingemans et al., 2014 acs:118/aro:106/gua:85/mut:86/nuo:72/pps:73/trp:71 MH155 E. coli strain carrying plasmid pMHLAS (pUCP22NotI- PlasB::gfp(ASV)Plac::lasR) CV026 C. violaceum mutant strain harboring a mini-Tn5 insertion in the C6- HSL-encoding gene cviI (Smr mini-Tn5 Hgr cviI::Tn5xylE Kmr)

Table S3.4. Primers used for qRT-PCR amplification.

Primer Sequence (5’→3’) Amplicon Applicability to other genomesa size (bp) trpB_F CCTACTTCCAGCGCGACTAC 118 PAO1/LESB58/DK2/PA14/RP73/B136- trpB_R CCGGTATGGTTCAGCTCCT 33/PACS2 PA0121_F CCACCAGCCATTTCGTCTAC 88 PAO1/LESB58/DK2/PA14/RP73/B136- PA0121_R AGGGTTTCCTTGCTCAGCTT 33/PACS2/PA7 norC_F GCTTCAACACCTTCCTCCAG 139 PAO1/LESB58/DK2/PA14/RP73/B136-

128 Chapter 3

norC_R TCGATCTTCGAGCTCCACTT 33/PACS2/PA7 rpsU_F GCCAGCCGTCAAAGTAAAAG 122 PAO1/LESB58/DK2/PA14/RP73/B136- rpsU_R TGGGCTTCTCGTAGAACTCG 33/PACS2/PA7 lasI_F CGTGCTCAAGTGTTCAAGGA 131 PAO1/LESB58/DK2/PA14/RP73/B136- lasI_R AAAACCTGGGCTTCAGGAGT 33/PACS2 PA1922_F CTGGAGCTGAGCCAGAAACT 101 PAO1/LESB58/DK2/PA14/RP73*/B136- PA1922_R TAGGTGAAGTGCGTGTCGTC 33/PACS2 oprI_F AGCAGCCACTCCAAAGAAAC 108 PAO1/LESB58/DK2/PA14/RP73/B136- oprI_R CAGAGCTTCGTCAGCCTTG 33/PACS2/PA7 PA3372_F TCCTCCAGACCCATTACCAC 150 PAO1/LESB58/DK2/PA14/RP73/B136- PA3372_R GCTGGCTGAAATCGAGGAT 33/PACS2 nosR_F CTTCCGCGACCTCGACTAC 136 PAO1/LESB58/DK2/PA14/RP73/B136- nosR_R GAGCAACTCCAGGGTCCAG 33/PACS2/PA7 algF_F TCGAAGCTGACCCTGAAGAC 143 PAO1/LESB58/DK2/PA14/RP73/B136- algF_R TTCAGGTCGCTGACCTTCTT 33/PACS2 algA_F CAAGCAGTACCCCAAGCAGT 121 PAO1/LESB58/DK2/PA14/RP73/B136- algA_R GTGCTCCTTGTTGCACACC 33/PACS2/PA7 glpD_F GTGTTCCTTTGCGAACAGC 105 PAO1/LESB58/DK2/PA14/RP73/B136- glpD_R CAGGCGGAATTCGTAGTGTT 33/PACS2/PA7 lasB_F GTCATCGACGCCAAGACC 97 PAO1/LESB58/DK2/PA14/RP73/B136- lasB_R ACTTGCCGATCTTCTGGTTG 33/PACS2/PA7 PA4171_F CCTGGTCAGGGAATTCAGC 124 PAO1/LESB58/RP73/B136-33/PACS2 PA4171_R GATACGCACGCTGGAATAGG pqsL_F CGGCTATTTCATCCTCATGC 100 PAO1/LESB58/DK2/PA14/RP73/B136- pqsL_R GATGCGGGTCTCGAACAG 33/PACS2 phzM_F CGGCGAAGACTTCTACAGCTA 142 PAO1/LESB58/DK2/PA14/RP73/B136- phzM_R ACCGACGTCGACGAAGCTA 33/PACS2 mreB_F GGCTCGATGGTCGTAGACA 128 PAO1/LESB58/DK2/PA14/RP73/B136- mreB_R ACGTAGGTGACGATGGCTTC 33/PACS2 mexC_F CGATCTATGCGGATTTCACC 125 PAO1/LESB58/DK2/PA14/RP73/B136- mexC_R GTAGGGCGTCCCTTCGAC 33/PACS2 nfxB_F GGCAGTCCTACCTGGAAGC 118 PAO1/LESB58/DK2/PA14/RP73/B136- nfxB_R CCGTAGACCAGGGTGATGAA 33/PACS2 pilQ_F ACCTGGAGAAACTCGACGTG 107 PAO1/LESB58/DK2/PA14/RP73/B136- pilQ_R CGGCTGCTCGATGGTATAG 33/PACS2/PA7 PA5481_F GAACCTGGTGGAAGACGAGA 103 PAO1/LESB58/DK2/PA14/RP73/B136- PA5481_R CAGATACTTGTCGCCCTGGT 33/PACS2/PA7 betI_F ACGGCATCATCAGCCACTAC 125 PAO1/LESB58/DK2/PA14/RP73/B136- betI_R CGGACTGTCGTCGTAGAGC 33/PACS2/PA7 aNo mismatches were found when the corresponding primers were aligned to the indicated reference genomes. *The PA1922 gene of P. aeruginosa RP73 contains a 262 bp deletion, causing a frameshift and preliminary stop codon, thereby truncating the predicted amino acid sequence from 653 to 194 amino acids.

129

130 Chapter 4

4 Chapter 4: Effect of Intrapulmonary Percussive Ventilation on P. aeruginosa in the CF lung

131 Effect of IPV on P. aeruginosa in the CF lung

132 Chapter 4

4.1 Introduction

The impaired mucociliary clearance in the cystic fibrosis lung leads to the accumulation of various microorganisms, including P. aeruginosa that colonizes the airways of the CF patient. Genetic adaptation of P. aeruginosa to the CF environment results in the switch from a planktonic to a biofilm lifestyle that is characterized by an overproduction of the exopolysaccharide alginate and a down-regulation of genes involved in acute virulence. Biofilms protect P. aeruginosa from antibiotics as well as from the immune system thereby allowing this bacterium to persist in the hostile CF environment for the rest of the patient’s life (Anwar et al., 1992; Jensen et al., 1990; Taylor et al., 2014). The continuous interactions between P. aeruginosa as well as other CF pathogens and the immune system cause persistent inflammation and ultimately lead to lung dysfunction and death of the patient. In a previous study it was demonstrated that high levels of fluid shear interfere with the formation of alginate biofilms by the P. aeruginosa reference strain PAO1 (Crabbé et al., 2008; Crabbé et al., 2010). Additionally, in the previous chapter we have shown that even a highly adapted, transmissible, P. aeruginosa CF strain is unable to form robust alginate biofilms under high fluid shear conditions. A possible way to increase the fluid shear levels in the CF lung is the use of intrapulmonary percussive ventilation (IPV). This is an airway clearance technique in which bursts of gas are introduced into the lungs at frequencies of 100-400 bursts per minute (bpm) via a mouth piece and provides an alternative to the conventional airway clearance techniques such as autogenic drainage (Lucangelo et al., 2003; Varekojis et al., 2003). Autogenic drainage (AD) was developed by Jean Chevaillier in 1967 in Belgium and relies on breathing control via adjustment of the rate, depth and location of respiration within the thoracic cavity in order to clear viscous pulmonary secretions (Agostini & Knowles, 2007; Chevaillier, 1984; Schoni, 1989). Although this technique did not enhance the clearance of viscous sputum from the lower airways in CF patients when preceding autogenic drainage compared to wet inhalation of saline (Van Ginderdeuren et al., 2008), it could be efficient in introducing shear stress in the CF lung, hence disrupting biofilm formation by P. aeruginosa. In this chapter, we have evaluated the effect of IPV on lung function, bacterial load in CF sputum and P. aeruginosa gene expression by subjecting eight different CF patients to AD and IPV at medium (200 bpm) and high (400 bpm) frequency. Our results show that although there are no significant differences between the three treatments at the level of lung function for the whole population, IPV at high frequency is able to induce the expression of bacterial

133 Effect of IPV on P. aeruginosa in the CF lung planktonic marker genes, decrease the bacterial load (including P. aeruginosa) and restore lung function in certain patients.

4.2 Materials and methods

4.2.1 Set-up of the clinical study and inclusion criteria

An overview of the set-up of the clinical study is provided in Figure 4.1, while the patient features are listed in Table 4.1. Eight different CF patients (Table 4.1) were subjected to three different treatments (AD, IPV at a medium frequency of 200 bpm, and IPV at a high frequency of 400 bpm), with a period of three months between two therapies during which the patients received AD as the standard airway clearance technique. Similarly, before the first treatment, AD was used as standard airway clearance technique. Two sputum samples were taken at three different time points: one that was used for determination of the bacterial load, while the other was immediately fixed with an equal volume of RNAlater (Qiagen) and was stored at -20°C for subsequent gene expression analysis. In addition, at each time point, the lung function was measured. Inclusion criteria for this study were: A minimum age of six years old, the ability to produce sputum, clinical stability at the start of the study, and the ability to tolerate two IPV or standard sessions for airway clearance. Patients who received lung transplantation, exhibited massive hemoptysis, pneumothorax, received invasive ventilation, or were pregnant, were excluded. CF patients who were colonized by P. aeruginosa as well as patients who are colonized by other pathogens were included in the study. However, among the P. aeruginosa-positive patients, only chronically colonized patients were included. This study was performed blind and was not unblinded before all data were obtained.

4.2.2 Determination of lung functions

Lung function was determined using spirometry testing, via which the forced expiratory volume in one second (FEV1) and the forced vital capacity (FVC) parameters were measured using a Jaeger Masterscreen (Acertys Healthcare, Belgium).

134 Chapter 4

4.2.3 Collection of sputum samples

Sputum samples were obtained in the hospital within 30 min after receiving one of the three airway clearance treatments via assistance of a physiotherapist, subsequently stored and preserved at 4°C during transport to the laboratory.

Figure 4.1. Set-up of the clinical study. Patients underwent three different treatments, each separated by three months. During each treatment, lung functions were measured and sputum samples were collected at three different time points: a first time point at which the patient was hospitalized, a second time point in the middle of the hospitalization period (day 3-5), and a final time point at the day the patient left the hospital (day 5-10). All patients were minimally hospitalized for a period of five days. The study was peformed blind until all data were obtained. AD, autogenic drainage.

Table 4.1. Patient characteristics. Patient CFTR mutation Age Sex FEV1 FVC Identified microorganismsc (%)a (%)b 01 F508del/F508del 25 M 64.7 82.8 B. multivorans 02 F508del/F508del 28 F 60 75 P. aeruginosa, S. aureus, S. mitis, Staphylococcus spp. (non-S. aureus) 03 F508del/F508del 12 F 44.7 63 P. mirabilis, Staphylococcus spp. (non-S. aureus) 04 F508del/F508del 20 F 32.7 49 P. aeruginosa 05 F508del/F508del 12 F 35.5 50.9 Species unknown 06 L927P/2183AA-G 21 M 41 51 Species unknown 07 F508del/1717-1G-A 34 M 24 58 P. aeruginosa 08 F508del/3849+10kbC-T 25 F 31 59 P. aeruginosa aFEV1 measured before the first treatment. bFVC measured before the first treatment. FEV1 and FVC are expressed as a percentage of their predicted values. cThese bacteria were identified in this study by either using selective media (S. aureus), genotyping (P. aeruginosa) or 16S rRNA sequencing (B. multivorans, S. mitis, P. mirabilis). B. multivorans, Burkholderia multivorans. P. aeruginosa, Pseudomonas aeruginosa. S. aureus, Staphylococcus aureus. S. mitis, Streptococcus mitis. P. mirabilis, Proteus mirabilis. 2183AA-G and 1717-1G-A are class I CFTR mutations. F508del is a class II CFTR mutation. L927P is a class IV mutation. 3849+10kbC-T is a class V mutation.

135 Effect of IPV on P. aeruginosa in the CF lung

4.2.4 Determination of bacterial load and identification of microorganisms

Sputum samples were first two-fold diluted using PBS and vortexed to reduce viscosity. Next ten-fold serial dilutions were plated on Pseudomonas agar P (Difco) to which Irgasan (1 mg/l) was added (Sigma-Aldrich), mannitol salt agar (Difco), and LB to detect P. aeruginosa, S. aureus, and other CF microorganism, respectively. Finally, cultures were grown at 37°C for 48 hours. P. aeruginosa isolates were further genotyped using the multiplex amplification of ferripyoverdine receptor genes described in Chapter 2. S. aureus was identified by the formation of yellow colonies on mannitol salt agar. Other microorganisms (not showing a colony morphology characteristic of P. aeruginosa or S. aureus) that were identified on LB medium (or occasionally on Pseudomonas agar), were identified via 16S rRNA sequencing. In order to do this, DNA was extracted from an overnight culture of the uknown microorganism grown in LB medium, using the DNeasy Blood & Tissue kit (Qiagen). The resulting genomic DNA was used as a template in a PCR reaction containing the universal 16S rRNA primers (27f 5’-AGAGTTTGATCMTGGCTC-3’ and 1492r 5’- GGYTACCTTGTTACGACTT-3’) (Frank et al., 2008). Amplification was performed in a Thermocycler (TC-412-Techne) using the following thermocycling conditions: 94°C for 5 min, followed by 30 cycles of 94°C for 1 min, 48°C for 1 min, 2 min at 72°C and a final elongation step at 72°C for 10 min.. Finally, PCR products were sequenced in double chain at the VIB genetic service facility (Wilrijk Belgium). Obtained sequences were analyzed using the nucleotide blast tool of the NCBI website at the species level.

4.2.5 RNA isolation and cDNA synthesis

Prior to RNA extraction, the sputum samples were incubated with an equal volume of sputolysin reagent (Calbiochem, La Jolla, CA) and vortexed for 30s. Next, samples were incubated with DNase I (Roche) for 20 min with intermittent mixing to remove extracellular chromosomal DNA. Finally, 30 µl of proteinase K (Qiagen, Hilden Germany) was added per one ml of sputum sample and samples were incubated at RT with intermittent mixing to remove proteins. In order to lyse eukaryotic cells and remove soluble cellular debris, sputum samples were washed twice with ice-cold sterile double-distilled water. Total RNA was purified from the pellets using the RNeasy Mini Kit (Qiagen, Hilden, Germany). Prior to

136 Chapter 4

RNA purification, bacterial cells inside the sputum were first lysed by following the Enzymatic lysis of bacteria-protocol (Qiagen, Hilden Germany). The RNA was fixed by adding 1 ml of RNAprotect Bacteria Reagent (Qiagen, Hilden Germany) to the bacterial culture in order to protect the RNA. Next, samples were vortexed for 5s and subsequently incubated at room temperature for 5 min. After a centrifugation step (5000 g, 10 min), the supernatant was decanted and 200 µl of TE-buffer (10 mM TrisCl, 1mM EDTA, pH 8.0) containing 3 mg/ml lysozyme was added to the pellets. Before adding 350 μl of RTL buffer, the mixture was incubated for at least 20 min with vortexing for 10 s every 2 min. Finally, 500 μl of ethanol was used to allow binding of the RNA on the column membrane. After this preparatory work, we proceeded to the purification step following the protocol for total RNA purification of bacterial lysates with the RNeasy mini kit (Qiagen, Hilden, Germany). Prior to qRT-PCR, an additional Turbo DNase (Ambion) treatment was performed via two 30 min incubation steps in the presence of 1 µl of (2U/µl) Turbo DNase. In a next step, the Turbo DNase-treated RNA was purified and concentrated using the RNA Clean and Concentrator Kit (Zymo Research). RNA quantity was determined using the Nanodrop ND-1000 spectrophotometer (NanoDrop technologies), while RNA quality was assessed via the Agilent 2100 Bioanalyzer using the Agilent RNA 6000 Nano Kit (Agilent Technologies). In general, the RNA quality was significantly lower than that of RNA extracted from bacterial cultures grown in laboratory media. However, samples that exhibited extensive degradation were omitted for further processing. The removal of genomic DNA was verified via 35 cycles of PCR amplification (5 min at 94°C, followed by 35 cycles of 45 s at 94°C, 45 min at 55°C, 1 min at 72°C, and a final extension step of 10 min at 72°C) of the oprI gene (Forward 5’- ATGAACAACGTTCTGAAATTCTCTGCT-‘3 and Reverse 5’- CTTGCGGCTGGCTTTTTCCAG-’3) of P. aeruginosa. cDNA was prepared using the iScript Select cDNA synthesis kit (Biorad), starting from 1 µg of DNA-free total RNA. The resulting cDNA was diluted 5x prior to use in qRT-PCR. In order to verify efficient conversion of RNA to cDNA, a PCR was performed using the oprI- specific primers and the PCR conditions described above.

4.2.6 Real-time PCR

Since the P. aeruginosa strains harbored by the CF patient are expected to be substantially different at the genotypic level due to their (often) environmental origin, we decided to design qRT-PCR primers that are able to anneal to target genes in various P. aeruginosa strains

137 Effect of IPV on P. aeruginosa in the CF lung without any mismatches. All primers used in qRT-PCR amplification were designed via Primer3 (Rozen & Skaletsky, 2000) and are listed in Table S3.4 of Chapter 3. Amplification was performed in a 96-well plate, in which each well contained 25 μl of volume consisting of 9.5 μl nuclease-free water, 1 µl of a each primer (10 μM), 12.5 μl of the 2x iQTM SYBR® Green supermix (Biorad) and finally 1 μl of template cDNA (5x diluted). The PCR amplification was performed via the iQ2 real-time PCR detection system (BioRad) using the following program: an initial cycle at 95°C for 3 min for denaturation and enzyme activation, then 40 cycles of 95°C for 10 s and 55°C for 60 s. Finally, melting curves were determined to identify primer dimer formation. qRT-PCR results were normalized against the housekeeping gene oprI encoding the major outer membrane lipoprotein I. Fold changes were calculated using the Livak method (Livak & Schmittgen, 2001). The experiment was at least performed in technical duplicates.

4.2.7 Statistical analysis

In order to determine if there was a difference in the effect on lung functionality (reflected by a change in FEV1 and/or FVC) between the three treatments, a one-way ANOVA was performed using the Prism5 software (Graph Pad, La Jolla, CA, USA). In order to compare two treatments, a two-tailed student’s t test was performed using the Prism5 software (Graph Pad, La Jolla, CA, USA).

4.2.8 Ethics statement

This clinical study was approved by the Ethics Committee of the UZ Brussel. An informed consent was read and signed by the eight patients that participated in this clinical trial.

4.3 Results

4.3.1 Effect of IPV on lung function

Both AD (7.3 ± 6.3% for FEV1 and 8.7± 6.6% for FVC) and IPV at 400 bpm (7.4 ± 2.6% for FEV1 and 10.6 ± 2.6%) appeared to cause an increase in lung function compared to IPV at 200 bpm (Table 4.2). Although there was no significant difference between the treatment groups, the p-values for both FEV1 (p = 0.09) and FVC (p = 0.06), were only slightly below the level of significance. More specifically, when comparing IPV at 200 bpm to IPV at 400 bpm, a significant increase in FEV1 (p = 0.03) was observed for IPV at high frequency. Since

138 Chapter 4 no data were available for a number of subjects, it was not possible to perform an ANOVA using correlated samples for the whole population (Table 4.2). However, it appears that for the two subjects from whom data were collected during all three therapies (patients 02 and 04), IPV at 400 bpm was slightly more effective in increasing FEV1 (11% versus 8% for patient 02 and 5% versus 1% for patient 04) and FVC (13% versus 7% for patient 02 and 12% versus 5% for patient 04) than AD.

Table 4.2. Change in lung function during AD, IPV200 and IPV 400 treatmentsa. FEV1 FVC Patient AD IPV 200 IPV 400 AD IPV 200 IPV 400 01 6% N/A N/A 5% N/A N/A 02 8% 4% 11% 7% 5% 13% 03 N/A 5.8% N/A N/A 8.6% N/A 04 1% 0% 5% 5% -1% 12% 05 N/A -4.7% 6.5% N/A -21.8% 10.3% 06 19% N/A N/A 22% N/A N/A 07 3% -1% N/A 6% -2% N/A 08 7% N/A 7% 7% N/A 7% Average 7.3% 0.8% 7.4% 8.7% -2.2% 10.6% SD 6.3% 4.2% 2.6% 6.6% 11.8% 2.6% aThe change in lung function was determined by calculating the difference between the lung function (FEV1 or

FVC) measured at the end of the treatment and that measured at the start of the treatment (ΔFEV1= FEV1end-

FEV1start; ΔFVC1= FVC1end-FVC1start). N/A, data not available. SD, standard deviation.

4.3.2 Change in bacterial load of CF sputum during AD and IPV treatment

Four of the patients in our study were colonized with P. aeruginosa (Table 4.1). Although it was a prerequisite to be chronically colonized with P. aeruginosa in order to participate in the clinical study, we verified if the same genotype remained present in the sputum of the CF host during the different treatments by performing a multiplex PCR, amplifying fragments of the ferripyoverdine receptor genes.

139 Effect of IPV on P. aeruginosa in the CF lung

Figure 4.2. Change in bacterial count during IPV treatment of patients 02 (A-C) and 03 (D-F). The log10

CFU/ml is presented in function of the time point in treatment. The limit of detection was reached at a log10 CFU/ml value of 4.6. Day 1 indicates the first day of hospitalization. At this time point the patient has not received treatment yet. In case of patient 03, the bacterial count was not determined during the first day of the third treatment.

No change in ferripyoverdine receptor gene profile was observed during any of the three treatments for the CF patients. Patient 02 was colonized by a P. aeruginosa harboring fpvAIIb and fpvB, sputum from patient 04 contained a P. aeruginosa harboring fpvAIII and fpvB, while patients 07 and 08 remained colonized by P. aeruginosa strains containing the fpvAI and fpvB genes. Sequencing of the 16S rRNA genes revealed the presence of a number of other bacteria in the sputum of CF patients participating in this study. Patient 01 was colonized by Burkholderia multivorans (B. multivorans), while patient 03 was infected by Proteus mirabilis (Table 4.1). Sputum from patient 02 contained Streptococcus mitis, that has not been considered as a typical CF pathogen (Table 4.1).

140 Chapter 4

Figure 4.3. Change in bacterial count during IPV treatment of patients 04 (A-C) and 05 (D-E). The log10

CFU/ml is presented in function of the time point in treatment. The limit of detection was reached at a log10 CFU/ml value of 4.6. D1 indicates the first day of hospitalization. At this time point the patient has not received treatment yet. In case of patient 04, the bacterial count was not determined during the first day of the second treatment. Samples collected during the third treatment of patient 05 did not contain detectable amounts of bacteria using the plating assays applied in this study.

No general conclusions can be made when looking at the change in bacterial load of the sputum from the different subjects. In the case of patient 02, we observed a strong decrease (108 to 106 CFU/ml) in P. aeruginosa as well as total bacterial counts during the IPV at 400 bpm, but not during the other treatments (Figure 4.2), while for patient 04 this was the case for AD compared to the other treatments (Figure 4.3). For the other patients that received multiple treatments (patients 03, 05, and 07), comparable levels of bacterial counts were observed during the different treatments (Figures 4.2-4.4). This variability in observations might be a consequence of the antibiotic treatment received by the patients before and during the hospitalization period. Although the antibiotic treatment was mainly conserved during the three treatments, the addition or removal of certain antibiotics when comparing one treatment to another may account for the observed drop in bacterial numbers (Table S4.1).

141 Effect of IPV on P. aeruginosa in the CF lung

Figure 4.4. Change in bacterial count during IPV treatment of patients 01, 06, and 07. The log10 CFU/ml is presented in function of the time point in treatment. The limit of detection was reached at a log10 CFU/ml value of 4.6. D1 indicates the first day of hospitalization. At this time point the patient has not received treatment yet. Patients 01 (C) and 06 (D) underwent only one treatment, while the bacterial count in samples collected during both AD (A) and IPV at 200 bpm (B) was determined for patient 07.

4.3.3 Effect of IPV on P. aeruginosa gene expression in the CF lung

In the previous chapter we have identified a number of genes that could serve as marker genes for the planktonic or biofilm mode of growth. Selected genes associated with planktonic growth are PA1922 and PA0121, while the biofilm-associated genes are norC (denitrification), algA (alginate production), glpD (glycerol metabolism), and mreB (cell shape maintenance). The pvdS gene was included to study the effect of these treatments on the pyoverdine biosynthetic pathway. Since pvdS is an extracytoplasmic sigma factor that regulates the expression of toxA as well as that of PrPL endoprotease and the pyoverdine biosynthetic cluster it may also rather be considered as a gene that tends to be associated with the planktonic mode of growth. In addition, treatment of the CF lung by the introduction of bursts of gas might also lead to a better aeration and increased levels of oxygen that render iron in its oxidized form, hence making it available for pyoverdine-mediated uptake.

When looking at the change in gene expression profiles of P. aeruginosa, it appears that there is a lot of variation between patients, most likely reflecting the complexity of the regulatory networks in this pathogen (Figure 4.5). In some patients, the expression levels of genes are

142 Chapter 4 highly similar when comparing the middle and late sampling points (six of the seven genes are regulated in the same direction in the case of the IPV200 sample of patient 02, and the IPV400 sample of patient 08) (Figure 4.5). In contrast to this, the marker genes in the AD sample of patient 07 are inversely regulated when comparing different time points. Remarkably, the PA0121 gene, the second most down-regulated gene under low fluid shear conditions, and algA are often (in eight of the 12 samples) inversely regulated in the same sample. In addition, during IPV at 400 bpm of patient 02, both planktonic marker genes as well as pvdS are highly up-regulated, while algA is down-regulated, in contrast to the IPV200 sample of this patient. However, in patient 07, the planktonic genes are down-regulated during IPV at 200 bpm, while algA being up-regulated, which is opposite to the expression of these genes in the AD sample. One of the subjects, patient 04, received twice the IPV at 400 bpm and therefore was suitable for validation of the reliability of our gene expression assay. A similar expression pattern was observed for P. aeruginosa in both IPV400 samples at nearly the same time point since five out of seven genes were regulated in the same direction.

143 Effect of IPV on P. aeruginosa in the CF lung

Figure 4.5. Change in P. aeruginosa gene expression over time in function of the applied treatment. All values are expressed as log2 fold changes. D1 indicates the first day of hospitalization. At this time point the patient has not received treatment yet. PA1922 and PA0121 are marker genes for the planktonic lifestyle of P. aeruginosa, while norC, algA, glpD, and mreB are marker genes for the biofilm lifestyle of this pathogen according to results obtained in a previous in vitro study.

144 Chapter 4

4.4 Discussion Although several other studies have investigated P. aeruginosa gene expression in the CF lung in a cross-sectional way (Barthe et al., 2015; Bielecki et al., 2013; Konings et al., 2013; Son et al., 2007), we believe this is the first study to report the longitudinal monitoring of the P. aeruginosa transcriptome in response to different treatments that are not focused on antibiotic-mediated effects. In this clinical study, we have evaluated the potential of IPV treatment on manipulating P. aeruginosa gene expression driving it from a biofilm into a planktonic mode of growth by potentially introducing shear stress in the CF lungs. Interestingly, there was a nearly significant difference between the three treatment groups. More specifically, IPV at a low frequency did not improve pulmonary function, in contrast to AD or IPV at a high frequency. Furthermore, the lung function (reflected by FEV1 and FVC) of two patients (patients 02 and 04) that have been chronically colonized by (mucoid) P. aeruginosa was enhanced to a greater extent by the IPV treatment at high frequency compared to AD. Furthermore, one of these patients (patient 02) showed a strong decrease in bacterial sputum load, in contrast to the other treatments. However, the second chronically colonized CF patient only showed a strong decrease in bacterial sputum load when receiving AD. These conflicting results might have arisen due to the addition and/or removal of one of the antibiotics in the antibiotic repertoire received by these patients before and during the clinical study. The addition of azithromycin for instance might have resulted in a strong decrease of the bacterial sputum load, since this antibiotic was combined with the IPV400, but not the other treatments of patient 02, while patient 04 only received this antibiotic during AD (Figure S4.1). In agreement with the increase in lung function and the decrease in bacterial load, the planktonic marker genes of P. aeruginosa that are associated with high fluid shear conditions, were up-regulated in response to IPV at high, but not medium frequency. In addition, the PA0121 planktonic marker gene and algA, one of the key genes in alginate biosynthesis, were inversely regulated in the great majority of sputum samples. Concerning the other marker genes, it was difficult to make general conclusions since their expression in response to the different treatments varied strongly between patients. Similarly, in their recent study, Barthe and colleagues could not associate the expression of well-known marker genes (algD, algR, antB, lasA, and pqsA) with the clinical status of CF patients (Barthe et al., 2015). Although it is generally accepted that P. aeruginosa undergoes convergent evolution during adaptation to the CF lung environment (Marvig et al., 2015), the different P. aeruginosa strains that colonize CF patients have a different origin, and hence genetic make-

145 Effect of IPV on P. aeruginosa in the CF lung up. The difference in only one or a few regulators in the extremely complex regulatory network of P. aeruginosa (Balasubramanian et al., 2013) can result in major differences in the response of this opportunistic pathogen to the multiple stimuli it receives in the CF lung. The effect of a treatment on the response of P. aeruginosa may therefore greatly depend on the time that the bacterium has had to adapt to the CF environment and the different mutational routes that have lead to this adaptation.

4.5 Conclusion

In this clinical study, we have evaluated the potential use of IPV in manipulating the behavior of P. aeruginosa, driving it from a biofilm lifestyle into a planktonic mode of growth. We have found differences between AD, IPV at medium frequency, and IPV at high frequency in enhancing lung function, although slightly below the level of significance. Combining all data sets (lung function, bacterial counts, and gene expression by P. aeruginosa), we can conclude that IPV at a medium frequency does not contribute to an increase in pulmonary function, neither to enhanced bacterial clearance or down-regulation of bioflm-associated genes. In contrast, AD and IPV at a high frequency do improve lung function, although this appears to be more pronounced for IPV at high frequency in patients that are chronically infected by P. aeruginosa. Although these results are promising, a larger group of patients should be included in a study that compares the effects of IPV at 400 bpm on lung function and P. aeruginosa biofilm formation to those of AD while avoiding changing the antibiotic treatment during the study.

146 Chapter 4

4.6 Supplementary data

Table S4.1. Overview of the antibiotic treatment received by the clinical study participants before and during IPV treatment. Patient AB pre T1 AB during T1 01 Coli/TOBI inh Glazidim 1g/Negaban 2g IV/Floxapen 500mg PO 02 Coli 2MU inh,Ciproxine,Floxapen Coli 3x1MU/Glazidim 9g/24u IV/Floxapen 3x1g PO 03 Floxapen 2x2g, Amukin 500mg 1x/d inh Tazocin 4x3,5g, Amukin 1g IV, Floxapen 3x 1g PO Meronem 3x2g, Coli 3MU, Azythromicine 500 3d/w 04 TOBI/AZLI inh, Azythromicine 500 3d/w PO PO Tazocin 3x3g, Obracin 360mg, Coli 3x1MU IV, 05 Coli/TOBI inh, Azithromycine 250mg 3d/w Amoxicilline 3x500 mg, Coli/TOBI inh 06

Coli 2MU inh, Azithromycine 500 mg 3d/w Tazocin 4x4g, Coli 3x1MU IV, Azithromycine 07 PO 500mg 3d/w PO 08 Ciprofloxacine 750mg/d, AZLI Meronem, Coli Patient AB pre T2 AB during T2 01

Coli 2MU Temocilline2x2g/Coli 3x1MU IV,Floxapen3x1g PO, 02 inh,Ciproxine,Floxapen,Azythromicine 500 Coli 2MU inh 3d/w PO 03 Floxapen 2x2g, Amukin 500mg 1x/d inh Augmentin 4x1g, Amukin 1g IV Meronem 3x2g, Coli 3MU, Azythromicine 500 3d/w 04 TOBI/AZLI inh, Azythromicine 500 3d/w PO PO Tazocin 3x3g, Obracin 360mg, Coli 3x1MU IV, 05 Coli/TOBI inh, Azithromycine 250mg 3d/w Augmentin 3x500 mg, Coli/TOBI inh 06

Coli 2MU inh, Azithromycine 500 mg 3d/w Tazocin 4x4g, Coli 3x1MU IV, Azithromycine 07 PO 500mg 3d/w PO 08

Patient AB pre T3 AB during T3 01

Coli 2MU Glazidim 9g/24u/Amikacine 750mg/d IV, Floxapen 02 inh,Ciproxine,Floxapen,Azythromicine 500 3x1g, Azythromicine 500 3d/w PO, Coli 2MU 2x/d 3d/w PO 03 Floxapen 2x2g, Amukin 500mg 1x/d inh Tazocin 4x3,5g, Amukin 1g IV, Floxapen 3x 1g PO 04 Coli 2MU inh, Azythromicine 500 3d/w PO Coli 3x1MU, Negaban 2x2g IV, Coli 2MU inh Glazidim 3x2g, Coli 3x 1MU IV, 3x500mg 05 Coli/TOBI inh, Azithromycine 250mg 3d/w Amoxycilline 06

07

08 Ciprofloxacine 750mg/d, AZLI Glazidim 6g, Coli 3x1MU IV, AZLI inh Coli, colistin. TOBI, tobramycin. Ciproxin, ciprofloxacin. Floxapen, floxacillin. AZLI, aztreonam. Negabam, temocillin. Meronem, meropenem. Tazocin, piperacillin + tazobactam. Amukin, amikacin. Obracin, tobramycin sulfate. Augmentin, amoxicillin + clavulanic acid. Azythromycine, azithromycin. Amoxycilline, amoxicillin. AB, antibiotics.

147

148 Chapter 5

5 Chapter 5: Identification and functional analysis of a novel S-type pyocin

The content of this chapter has been submitted as a manuscript to MicrobiologyOpen, entitled:

“Identification and functional analysis of S6 pyocin from a cystic fibrosis Pseudomonas aeruginosa clinical isolate" by Jozef Dingemans, Michael Craggs, and Pierre Cornelis.

149 Identification of pyocin S6

150 Chapter 5

5.1 Introduction

P. aeruginosa is a ubiquitous Gram-negative γ-proteobacterium that can be isolated from various sources, ranging from environmental samples to the human host (Pirnay et al., 2009). In the human host, it can cause disease in patients that are at risk including immunocompromised as well as cystic fibrosis (CF) patients (Kerr & Snelling, 2009; Lyczak et al., 2002). Although the environments from which P. aeruginosa strains can be retrieved are diverse, most of them, if not all, share the feature of harboring entire microbial communities (Kent & Triplett, 2002; McGuigan & Callaghan, 2015; Renwick et al., 2014). These communities are shaped by many symbiotic as well as antagonistic interactions. In order to compete with phylogenetically distant species, P. aeruginosa can produce several antagonistic molecules such as phenazines, pyoluteorin, and staphylolysin (LasA protease) among others (Gross & Loper, 2009; Kessler et al., 1993). On the other hand, a particular P. aeruginosa strain could encounter other pseudomonads within the same niche. To protect itself against these closely related organisms, but nevertheless competitors, P. aeruginosa is able to produce bacteriocins as weapons.

Bacteriocins are narrow-spectrum antibacterial proteins which differ from traditional antibiotics in that they generally only affect members of related genera or even species. They are also much bigger than antibiotics; their molecular weight ranges from 2 to 85 kDa. These toxins are generally produced under conditions of stress, such as nutrient depletion or overcrowding (Riley & Gordon, 1999; Riley & Wertz, 2002a; Riley & Wertz, 2002b). Bacteriocins produced by P. aeruginosa are called pyocins and several types of these bactericidal molecules have been described (Ghequire & De Mot, 2014; Michel-Briand & Baysse, 2002). R- and F-type pyocins resemble the bacteriophage tails (hence being called tailocins) and are insoluble (Michel-Briand & Baysse, 2002). They differ in the fact that R- type pyocins are contractile but non-flexible, while the F-type pyocins are flexible but non- contractile. Another large group of pyocins are soluble (S-type) pyocins. These pyocins are highly modular and consist of three domains: a receptor binding domain (generally N- terminal), a translocation domain, and a killing domain (C-terminal) (Michel-Briand & Baysse, 2002). Although recently several S-type pyocins encoding genes have been identified by screening draft and whole-genome sequences of P. aeruginosa strains (Ghequire & De Mot, 2014), only six different S-type pyocins have been so far functionally characterized, either having DNase (S1, S2, S3, AP41) (Duport et al., 1995; Sano & Kageyama, 1993; Sano

151 Identification of pyocin S6 et al., 1993; Seo & Galloway, 1990), tRNAse (S4) (Elfarash et al., 2012) or pore forming activity (S5) (Ling et al., 2010). At this moment, the majority of S-type pyocins have been found to enter the bacterial cell via binding to a TonB-dependent receptor involved in ferrisiderophore uptake. More specifically, pyocins S2 and S4 enter via the ferripyoverdine receptor FpvAI (Denayer et al., 2007; Elfarash et al., 2012), pyocin S3 via FpvAII (Baysse et al., 1999), while pyocin S5 recognizes the FptA ferripyochelin receptor (Elfarash et al., 2014).

In this study, we have identified a novel S-type pyocin by screening the genome sequence of the epidemic P. aeruginosa strain CF_PA39, isolated from a Belgian cystic fibrosis (CF) patient (Dingemans et al., 2014), for pyocin-related sequences.

5.2 Materials and Methods

5.2.1 Strains and plasmids

Bacterial strains and plasmids used in this study are described in Table 5.1. Bacteria were grown at 37°C in rich lysogeny broth (LB) medium (Life Technologies) or in iron-poor Casamino Acids (CAA) medium (Difco Laboratories) and cultures were shaken in a New Brunswick Innova 4000 shaker at 200 rpm.

Table 5.1. Strains and vectors used in this study. Strains or plasmids Features References Pseudomonas aeruginosa PAO1 Wild type P. aeruginosa Stover et al., 2000 CF_PA1 to CF_PA125‡ CF clinical P. aeruginosa strains This study Escherichia coli DH5α F– Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 Hanahan, 1985 hsdR17 (rK–, mK+) phoA supE44 λ– thi-1 gyrA96 relA1 – – – BL21(DE3)pLysS E. coli B F dcm ompT hsdS(rB mB ) gal λ(DE3) [pLysS Weiner et al., 1994 Camr] Plasmids pET15b Expression vector, N-terminal his-tag, Apr Studier et al., 1990 ‡With the exception of CF_PA8, CF_PA18, CF_PA21, CF_PA25, CF_PA29 to CF_PA33, CF_PA52 and CF_PA95 to CF_PA99. Camr and Apr indicate resistance to chloramphenicol and ampicillin, respectively.

5.2.2 Secondary structure prediction of pyocin S6

The PHYRE2 server (Kelley et al., 2015) was used to acquire a predictive secondary structure model of pyocin S6 and its killing domain, as it reportedly provides a high accuracy of protein structure and trans-membrane protein prediction.

152 Chapter 5

5.2.3 RNA isolation and semi-quantitative reverse transcriptase (RT)-PCR

Using the RNeasy Mini kit (Qiagen), total RNA was isolated from the clinical P. aeruginosa strains CF_PA17 and CF_PA39 after 24h of growth in CAA medium (iron-limited growth condition) or LB medium (iron-saturated growth condition). Prior to reverse transcription, the RNA was treated with Turbo DNase (Ambion) via two 30 min incubation steps in the presence of 1 µl of (2U/µl) Turbo DNase. The purity and concentration of the RNA was determined by gel electrophoresis using oprI primers (Table 5.2) (De Vos et al., 1997) and spectrophotometry (NanoDrop, Thermo Scientific). Next, cDNA was synthesized starting from one microgram of Turbo DNase-treated (Ambion) total RNA using the iScript cDNA synthesis kit (Biorad). A negative control (no reverse transcriptase added) was included for each sample. RT-PCR was performed on the cDNA templates using primers S6_Fw & S6_Rv and S6I_Fw & S6I_Rv (Table 5.2). The PCR conditions included an initial denaturation of 5 min at 95°C, followed by the first of 35 cycles: 45 s denaturation at 95°C, 45 s annealing at 55°C and 2 min extension at 72°C, followed by a final extension at 72 °C for 10 min. The PCR products were loaded onto an ethidium bromide-stained, 2 % agarose gel using 1xTBE running buffer and visualized under U.V. light. A Smart Ladder MW-1700-10B molecular weight marker (Eurogentec) was run additionally to confirm the expected molecular weight of the amplification product. Table 5.2. List of primers used in this study Name Primer sequence (5` 3`) RE Product size Pyocin S6 primers S6_Fw GTCTCCAGATCCGCATGAAT - 884 bp S6_Rv CGGAGCAGGATGGTAACTGT Immunity primers S6I_Fw CCTAGCATCGGGAAATGATG - 127 bp S6I_Rv TACTCCAATCCAACCGGAAG Housekeeping gene primers oprI_Fw ATGAACAACGTTCTGAAATTCTCTGCT - 248 bp oprI_Rv CTTGCGGCTGGCTTTTTCCAG Pyocin S6 cloning primers S6C_Fw GGAATTCCATATGGCACGACCCATTGCTGACCTTA NdeI 1946 bp S6C_Rv CGGGATCCCTAGGCGTAAACCCCAATAAAATAC BamHI Pyocin S6 sequencing primers 715_S6_Fw TAATACGACTCACTATAGGG - 1133 bp 715_S6_Rv AGGGCGTAACGAACGCTAT 716_S6_Fw GGCTTTGCCAGTCTGACCTA - 1106 bp 716_S6_Rv GCTAGTTATTGCTCAGCGG The underlined sequences, added to the primer fragments, indicate the recognition sites of the restriction enzyme (RE) used for the cloning experiments.

153 Identification of pyocin S6

5.2.4 Cloning of the pyocin S6 gene

The pyocin S6 gene (1946 bp) was PCR-amplified from CF_PA39 (Table 5.1) genomic DNA using primers S6C_Fw and S6C_Rv (Table 5.2) and a KAPA HiFi PCR Kit. Amplified fragments were introduced into pET15b (+) (Merck, Germany) (Figure S5.1) by NdeI / BamHI double digestion, ligation, and introduced in DH5α competent cells. A colony PCR using the same primers was performed as selection step for colonies positive for the gene of interest. Plasmids from these positive colonies were purified using the PureYield Plasmid Miniprep system (Promega) and sent for sequencing at the VIB Genetic Service Facility (Wilrijk) using two sets of primers (external 715_S6_Fw & 716_S6_Rv and internal 715_S6_Rv & 716_S6_Fw, Table 5.2), ensuring complete coverage of the sequence.

5.2.5 Overexpression and purification of the protein

Firstly, the plasmid, of which the sequence was confirmed to contain the gene encoding pyocin S6, was introduced into BL21(DE3)pLysS competent cells using a heat shock protocol. Subsequently, a colony PCR was performed to verify whether the transformants contained the gene of interest. For overexpression of the cloned gene, the transformants with the recombinant plasmids were induced by growing them overnight at 28°C in the presence of

1.0 mM IPTG after the OD600 reached 0.7. The harvested cells were re-suspended in TGE buffer (50 mM Tris-HCl (pH 7.5), 10% glycerol, 1 mM EDTA, and 10 mM imidazole) and disrupted by French pressure cell press. The lysate was centrifuged at 10,000 rpm for 15 min, and the clear supernatant was loaded onto a Hi-Trap FF column (Amersham Biosciences, GE Healthcare) integrated by AKTA TM FPLC system (Amersham Biosciences, GE Healthcare). The His-tagged proteins were eluted using 500 mM imidazole in TGE buffer (pH 7.5). The purity of the His-tagged proteins was confirmed after 12% SDS polyacrylamide gel electrophoresis (SDS-PAGE, Invitrogen). The gels were stained with Coomassie Blue and a Thermo Scientific 10-170 kDa protein ladder was used to confirm band size. The purified proteins were dialyzed against 2 litres TGE buffer (pH 7.5). The protein concentration was determined using a NanoDrop 1000 spectrophotometer (Thermo Scientific). The pooled pure proteins were divided into small aliquots and stored at -20°C, which were frozen and thawed individually before each manipulation.

154 Chapter 5

5.2.6 Pyocin sensitivity assay

In order to verify the sensitivity of different clinical isolates (Table 5.1) to the purified pyocin S6, a pyocin spotting assay was performed. This involved spotting 10 µl of the purified pyocin protein (10 mg ml-1) onto CAA and LB plates with bacterial cell layer containing 5×106 cells ml-1 and incubated at 37°C for 24 h. Strains were classified as sensitive if a growth inhibitory zone appeared on and/or around the spotting site, and resistant if no such inhibition zone was apparent.

A minimum inhibitory concentration (MIC) test was performed in triplicate by spotting a 2- fold serial dilution (starting concentration, 17 mg ml-1) of pyocin S6 onto a lawn of the sensitive P. aeruginosa strain CF_PA109 (5×106 cells ml-1), which was grown overnight on a CAA plate at 37°C for 24 h.

5.2.7 Screening of clinical strains for the presence of pyocin S6

Chromosomal DNA from our collection of 110 P. aeruginosa CF strains was prepared using the DNeasy Blood & Tissue Kit (Qiagen) and screened for the gene encoding the S6 pyocin using primers S6_Fw and S6_Rv (Table 5.2). The PCR conditions included an initial denaturation of 5 minutes at 95°C, followed by the first of thirty cycles: 45 s denaturation at 95°C, 45 s annealing at 60°C and 1 min extension at 72°C, followed by a final extension at 72°C for 10 min.

5.3 Results and Discussion

5.3.1 Nucleotide and amino acid sequences of pyocin S6

The genome sequence of P. aeruginosa CF_PA39, an isolate belonging to an epidemic P. aeruginosa CF clone (Dingemans et al., 2014), was screened for the presence of existing pyocins. Strikingly, although nucleotide sequences showing high identities to the genes encoding the receptor binding (RBD) and translocation domains (TD) of pyocin S1 (Table 5.3) were detected in the genome of P. aeruginosa CF_PA39, the sequence of the gene encoding the killing domain (KD) of this novel pyocin, termed S6, was completely different from S1 as it shared 69% similarity with the gene encoding the killing domain of a putative colicin E3 homologue of Pseudomonas fluorescens F113 (Table 5.3).

155 Identification of pyocin S6

Table 5.3. Sequence comparisons between pyocin S6, S1 and colicin E3 homologue using a standard nucleotide BLAST Pyocin S6 RBD (S1) TD (S1) KD (E3) Nucleotide overlap 723 726 150 Mismatches 3 13 46 Identity 99.6% 98.2% 69.3% (Pyocin S1 and colicin E3 nucleotide sequences acquired from www.pseudomonas.com (Winsor et al., 2011)). RBD: receptor binding domain, TD: translocation domain, KD: killing domain.

Furthermore, a second ORF (start codon ATG), overlapping with the first ORF encoding the novel pyocin gene (stop codon TGA), was identified. Considering its genetic organization and size, this ORF likely encodes the immunity domain of pyocin S6.

The position of the ORFs encoding the putative novel S-pyocin and its immunity gene corresponds to the pyocin S2 gene locus in P. aeruginosa PAO1 (Figure 5.1A). Although the toxA and PA1153 genes are well-conserved between P. aeruginosa CF_PA39 and PAO1, the genetic region in between these genes is not. Remarkably, P. aeruginosa PAO1 harbors the pyocin S6 immunity gene although the pyocin S6 gene is absent. Such an “orphan” gene could confer a competitive advantage to the bacterium as the latter is likely able to neutralize a larger spectrum of invading pyocins.

The nucleotide sequence of the two tandem ORFs in the genome of P. aeruginosa CF_PA39 spans a region of 1,946 basepairs (bp) (Figure 5.1B; Figure S5.2). The first ORF encodes a protein of 571 amino acids and the second ORF specifies a small protein of 77 amino acids. The predicted molecular weights of these proteins are 60,901 and 8,586 respectively (Figure 5.1B). As found by Sano et al. (Sano et al., 1993) for pyocin S1, a regulatory P-box was discovered 74 bp upstream of the 5’ start codon (Figure 5.1B; Figure S5.2). The P-box is always located 60 to 100 bp upstream of the Shine-Dalgarno sequence of the toxin gene and serves as a binding site for the PrtN protein, which positively regulates pyocin expression (Ghequire & De Mot, 2014; Matsui et al., 1993). The sequence is usually composed of four repeats of 10 to 11 nucleotides and each repeat has a consensus sequence of ATTGnn(n)GTnn(n) (Sano et al., 1993). Similar to other characterized S-type pyocins, a putative ribosome binding site of the immunity gene was found in the region encoding the C- terminal portion of the toxin protein, located eight bases upstream of the methionine start codon. An inverted repeat of eight nucleotides, separated by eight nucleotides, was found downstream of the second ORF, which could serve as a transcription terminator by forming a stem and loop structure (Figure 5.1B; Figure S5.2).

156 Chapter 5

Figure 5.1. Genetic organization of the pyocin S6 locus in the genome of P. aeruginosa CF_PA39. A. Comparison between the pys6 locus in P. aeruginosa CF_PA39 and the pyocin S2 pys2 locus in P. aeruginosa PAO1. Genes that are present in both the P. aeruginosa CF_PA39 and PAO1 genomes are shown in grey, while genes exclusive to one of both genomes are shown in black. B. Detailed overview of the pys6/imm6 locus. The stop codon of the pys6 gene is underlined once, while the start codon of imm6 is underlined twice. P box, promoter box. SD, Shine-Dalgarno sequence. ORF, Open Reading Frame. RBD, Receptor Binding Domain. TD, Translocation domain. KD, Killing domain. pys6, pyocin S6 gene. imm6, pyocin S6 immunity gene (Not drawn to scale).

The structural domain organization of pyocin S6 (Figure 5.2A) was found to be the same as reported for pyocin S1, with an N-terminal polypeptide of 241 amino acids, corresponding to the RBD, a central TD (245 amino acids) and a C-terminal KD (85 amino acids). The S6 immunity protein, which is assumed to confer immunity to the pyocin producing cell, is predicted to be a polypeptide of 77 amino acids.

To gain a global view on the secondary structure of pyocin S6 (Figure 5.2B; Figure S5.3), its predicted amino acid sequence was loaded into PHYRE2, a protein homology/analogy recognition server that generates secondary structure models based on alignments to known

157 Identification of pyocin S6 protein structures using PSY-BLAST. A comparison with the RBD and TD of pyocin S1 could not be made, as the crystal structure of the latter has yet to be solved.

The predicted secondary structure of the RBD of pyocin S6 is mainly dominated by two α- helical structures, the first spanning from residue 91 to 168 and the second extending from residue 172 to 241. These two helices could form a hairpin, as was found for the RBD of colicin E3 (Soelaiman et al., 2001), that docks into the binding pocket of its cognate receptor through hydrophobic interactions between exposed hydrophobic residues in and around the hairpin region, and the “floor” of the binding pocket. Several hydrophobic residues were found in this area of the RBD of pyocin S6, notably A162, V163, I168, P171, A173, Y174, M175, F176 and L177. Mutagenesis experiments should be carried out on these residues to test their roles in pyocin S6 function.

The TD of pyocin S6 shares 24% sequence identity with the TD of colicin E7, a bacteriocin of E. coli with DNase activity (Cheng et al., 2006). They both share similar structures, with a disordered N-terminal end and central layers of β-sheets that are flanked by α-helices.

The expected amino acid sequence of the KD of pyocin S6 was loaded separately into PHYRE2 (Figure 5.2C). The alignment showing the highest sequence identity (53%), was the ribonuclease domain of colicin E3 (Figure 5.2D). The crystal structure of colicin E3 has been solved and its KD forms a six-stranded, highly twisted, antiparallel β-sheet flanked by a two- turn α-helix at its N-terminal end (Soelaiman et al., 2001). In comparison, the predicted secondary structure of pyocin S6 was similar as it showed a 5-stranded antiparallel β-sheet with a centrally located α-helix.

Site-specific mutagenesis experiments in the KD of colicin E3 identified D510, H513, E517 and R545 as active site residues (Soelaiman et al., 2001). In the alignment with the KD of pyocin S6, these residues (D46, H49, E53 and R80) were found to be conserved.

The similar secondary structure, together with 53% sequence identity of the KDs and conservation of catalytic residues, suggest that in target cells pyocin S6 uses a killing action comparable to colicin E3 by inactivation of the host’s protein biosynthetic machinery through cleavage of a single ribonucleotide bond at the ribosomal A-site of 16S rRNA.

The fact that pyocin S6 shares similarity with colicins is not so surprising, since the same has been found for pyocins S1, S2 and AP41, where the DNase domains and the immunity proteins of these pyocins were found to be homologous to the corresponding regions of E2

158 Chapter 5 group colicins; E2, E7, E8, and E9 (Sano et al., 1993). This demonstrates that, over the course of evolution, pyocins have been subjected to forms of recombination and domain shuffling events with colicins and possibly bacteriocins of other bacteria. Support for this hypothesis can be found in the striking structural organization that both bacteriocins have in common, even though the putative domains of pyocin S6 (RBD, TD and KD) are arranged differently from those in colicin E3, in which the order is TD-RBD-KD from N to C-terminal end.

Figure 5.2. The pyocin S6 domains, and secondary structure analysis of pyocin S6 and colicin E3. A. Schematic diagram of the pyocin S6 domains. B. The predicted three-dimensional structure of pyocin S6. (C) The predicted secondary structure of the killing domain of pyocin S6 shows 53% sequence identity (D) with the ribonuclease domain of colicin E3. The images in B and C are colored according to the rainbow color scheme (N-terminal = blue, C-terminal = red).

5.3.2 Expression of pyocin S6 under iron-limited versus iron-repleted conditions

In order to test whether the genes encoding the pyocin S6 and its immunity protein are expressed under iron rich (LB) or poor (CAA) conditions, a semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) was performed (Figure 5.3). To determine the relative abundance of expressed genes, primers specific for the housekeeping gene oprI were used, using genomic DNA from P. aeruginosa CF_PA39 as a positive control. The test was performed on two strains, CF_PA17 and CF_PA39, belonging to the same clone but isolated from different CF patients, grown overnight in both LB and CAA medium.

The obtained results show that signal intensities with regard to the housekeeping gene oprI were similar, indicating comparable mRNA levels (Figure 5.3). Further, it was found that the

159 Identification of pyocin S6 expression of the pyocin S6 gene is elevated under iron-limited conditions (CAA). This suggests that, similar to other soluble pyocins, pyocin S6 possibly targets sensitive cells by binding to an iron-regulated outer membrane receptor. Although expression of the immunity gene was enhanced under iron-limited conditions, it was strongly detected under both iron- limited and iron-saturated conditions. Interestingly, the expression levels of the immunity gene are considerably higher than those of the pyocin S6 gene under both conditions. This is in contradiction with the genomic organization of the two ORFs that constitute an operon. However, it has been shown that the expression of genes in an operon can be discoordinately regulated by means of small noncoding RNAs (sRNAs) (Balasubramanian & Vanderpool, 2013). More specifically, the sRNA GlmZ has been shown to regulate the discoordinate expression of the glmU and glmS genes that constitute the dicistronic glmUS operon in E. coli (Urban & Vogel, 2008). Upon cleavage of the dicistronic mRNA by RNase E at the glmU stop codon, GlmZ stabilizes the glmS transcript and activates its translation through disrupting the secondary hairpin structure that blocks the glmS ribosome binding site. In contrast, the glmU mRNA is not stabilized and becomes quickly degraded. In case of pyocin S6, such a discoordinate regulatory mechanism might prevent the bacterial cell from spending energy in the production of the pyocin molecule, while under iron-limited conditions, when the target receptors are abundantly expressed, a different regulatory mechanism may stabilize the pyocin transcript. On the other hand, high basal expression of the immunity gene protects the pyocin producing cell against invading pyocins and in addition ensures that it can rapidly respond to changing conditions. Recently, Gomez-Lozano et al. (Gomez-Lozano et al., 2014) identified three antisense sRNAs inside the pyocin S3 gene and one antisense sRNA in the pyocin S5 gene. These antisense sRNAs could interfere with the trancription process itself or affect stability of the pyocin mRNA. Similarly, the pyocin S6 mRNA levels could be regulated by such antisense sRNAs, thereby affecting its expression depending on the iron concentrations experienced by the bacterium.

160 Chapter 5

Figure 5.3. Effect of iron limitation on the expression of pys6 and imm6. Using semi-quantitative RT-PCR, the expression of pys6 (884 bp) and oprI (248 bp) (A) and imm6 (127 bp) and oprI (248 bp) (B) was determined for P. aeruginosa CF_PA17 and P. aeruginosa CF_PA39 grown in LB or CAA. +RT and -RT indicate that reverse transcriptase was added or not, respectively. Fifty nanograms of genomic DNA isolated from P. aeruginosa CF_PA39 was included as a positive control (lanes marked with a “+”). The molecular-weight size marker is shown in the lanes marked with an “L”. The lower bands observed in the -RT reactions represent primer dimers and are not observed in the +RT reactions. pys6, pyocin S6 gene. imm6, pyocin S6 immunity gene.

5.3.3 Cloning, overexpression and purification of pyocin S6

The DNA sequence encoding both pyocin S6 and its corresponding immunity domain (1,946 bp) was successfully cloned into a pET15b vector. Plasmid pET15b provides an inducible promoter (Plac) for cloned genes and allowed expression of the P. aeruginosa DNA in E. coli

BL21 (DE3) pLysS. At an OD600nm of 0.7 and 1.5 hours later, samples were taken of the E. coli culture before and after IPTG induction respectively, to test for basal expression of the cloned fragment. Subsequently, the samples were analysed by 12% SDS PAGE (Figure 5.4). A clear band below 70 kDa (expected size of the His-tagged pyocin S6 protein = 63 kDa) was only observed for the induced culture, confirming that the cloned fragment was under control of Plac and not under its native promoter. Pyocin S6 was purified from E. coli BL21 (DE3) pLysS as described under “Materials and Methods”. The purified protein appeared as a single band after 12% SDS PAGE.

161 Identification of pyocin S6

Figure 5.4. SDS-PAGE of recombinant pyocin S6 protein. S, size markers; Lane 1: before IPTG induction; Lane 2: after 1.5 hours IPTG induction; Lane 3; purified pyocin S6 protein.

5.3.4 Pyocin S6 activity

To verify the activity of the purified protein, a sensitivity assay was performed where pyocin S6 was spotted on bacterial lawns of 110 P. aeruginosa CF isolates with known ferri- pyoverdine receptor types, grown on both LB and CAA. Pyocin S6 was found to be active on a total of 22 strains (20%) (Table 5.4), confirmed by either clear punched-out zones of no growth or thinning of growth around the spotting area. However, susceptibility was more pronounced under iron limitation, as growth-inhibitory effects were only clearly visible for strains grown on CAA plates (Figures 5.5A, B). This can be correlated to the results obtained in the semi-quantitative RT-PCR, where it was found that the gene encoding pyocin S6 is expressed under iron-limiting conditions, suggesting once again that in order for the protein to gain access to a sensitive cell it has to bind to a ferri-siderophore receptor.

As the ferri-pyoverdine receptor types were previously determined for the 110 strains by multiplex PCR, we tried to establish a relationship between the killing activity of pyocin S6 and the presence of a specific ferri-pyoverdine receptor. However, the results presented in Table 5.4 do not show any clear correlation, suggesting pyocin S6 does not target a specific ferri-pyoverdine receptor type.

162 Chapter 5

Table 5.4. Phenotypes of sensitivity or resistance to pyocin S6 of 110 P. aeruginosa CF isolates previously checked for the type of FpvA receptor gene by multiplex PCR.

FpvAI FpvAI + FpvB FpvAIIb FpvAIIb + FpvB FpvAIII FpvAIII + FpvB S6 Resistant 0 23 20 16 0 29 S6 Sensitive 2 11 0 7 1 1

The insensitivity to pyocin S6 might stem from (i) resistant strains not having a specific receptor on the cell surface, (ii) tolerant strains not having the translocation system for pyocin S6, and/or (iii) the immunity protein expressed in the insensitive strains might inactivate pyocin S6 action by binding to the C-terminal end of pyocin S6.

To test the purified pyocin for the presence of bacteriophage and to establish a MIC, a 2-fold serial dilution (starting concentration, 17 mg ml-1) of pyocin S6 was assayed against the sensitive indicator strain CF_PA109 (Figure 5.5C). If any inhibitory activity is due to the presence of bacteriophage, the more diluted fractions will produce individual plaques of phage lysis. If inhibition zones appear without plaque formation it is assumed that only pyocin activity is present (Osman, 1965). No bacteriophage activity was found and the MIC was determined to be 260 µg ml-1.

Figure 5.5. Determination of pyocin S6 activity and MIC. Of the 110 tested strains, 22 were found to be sensitive to the purified protein. Clear zones of inhibition only appeared when grown on CAA rather than on LB medium, as demonstrated here for the strain PA109 in figures B and A, respectively. (C) The MIC was determined for CF_PA109 by a two-fold serial dilution; number 1 represents the undiluted purified fraction (17 mg ml-1). As dilutions increase, the killing zones become smaller and then become gradually more opaque due to incomplete inhibition.

163 Identification of pyocin S6

5.3.5 Screening of clinical strains for the presence of pyocin S6

A PCR was performed to determine the occurrence of the gene encoding pyocin S6 in the clinical isolates of this study. 56 of the 110 (51%) strains proved to be positive for the gene, immediately explaining the relative low percentage of inhibitory action of the pyocin (20%), as in principle the 51% S6-positive cells should be immune due to the presence of the immunity gene which is constitutively expressed (Figure 5.3) and its encoded protein inhibiting the killing protein. It should be mentioned that in our collection of P. aeruginosa CF strains, there is a bias towards strains belonging to the P. aeruginosa CF_PA39 clone (20 of the 110 tested isolates), explaining the high frequency of isolates harboring the pyocin S6 gene and consequently the lower number of sensitive isolates. When screening the genomes submitted to the Pseudomonas Genome Database (Winsor et al., 2011), the pyocin S6 gene was only retrieved in a small number of other P. aeruginosa strains (6 out of 1001 strains; <1%), including a strain associated with bloodstream infections (P. aeruginosa PA45; Segata et al., 2013), a strain associated with urinary tract infections, another strain associated with intra-abdominal infections, and two respiratory P. aeruginosa strains. This indicates that although the pyocin S6 gene is not restricted to P. aeruginosa strains colonizing the CF lung environment, the pyocin S6 gene appears to be overrepresented in the P. aeruginosa population present at the CF reference center of the UZ Brussel compared to the global P. aeruginosa population.

5.4 Conclusion

P. aeruginosa is an opportunistic pathogen and is frequently isolated from the lungs of cystic fibrosis patients. Under conditions of iron-limitation, different strains of P. aeruginosa become more susceptible to pyocins because iron-siderophore uptake receptors are involved in the pyocin uptake (Baysse et al., 1999; de Chial et al., 2003; Denayer et al., 2007; Elfarash et al., 2012; Elfarash et al., 2014).

This study has provided the first experimental evidence of a nucleic acid sequence encoding a novel pyocin, termed pyocin S6, and its cognate immunity domain in the genome of a cystic fibrosis isolate. Sharing the receptor binding and translocation domain of pyocin S1, its killing domain has a 53% structural identity with the killing domain of colicin E3; a bacteriocin from

164 Chapter 5

E. coli that kills through site specific cleavage of a single phosphodiester bond of 16S ribosomal RNA in the 30S subunit (Soelaiman et al., 2001). This is reminiscent of domain shuffling in molecular evolution and it is therefore not unreasonable to hypothesize that pyocin S6 and colicin E3 share a similar mode of killing sensitive cells.

Via semi-quantitative RT-PCR analysis it was revealed that the genes encoding pyocin S6 and its immunity are transcribed and up-regulated in at least two clinical strains grown under iron- limiting conditions. Subsequently, both components were cloned, expressed and purified. The activity of this bacteriocin was tested against a collection of 110 CF strains, of which 22 proved to be sensitive. Against CF_PA109, a sensitive clinical strain, the MIC of pyocin S6 was determined to be 260 µg ml-1.

Finally, it was demonstrated that the pyocin S6 gene occurs at an intermediate frequency in the genomes of 110 clinical strains.

165 Identification of pyocin S6

5.5 Supplementary data

Figure S5.1. pET15b vector map.

166 Chapter 5

1 GTCATTGGATACGGGATATTGCAGTTTGATTGCTGTGTATTTCCAGTGTATTGGCGTTCA

61 TTTGGCAAGGGGCGTTTAAAAAGTCTAATTCGACGTAAATAATAAGCCATGCACTCATGG

121 AAGTAGGAGATACATATGGCACGACCCATTGCTGACCTTATCCACTTCAACTCTACAACT M A R P I A D L I H F N S T T

181 GTCACGGCAAGCGGAGACGTTTATTACGGCCCTGGGGGAGGTACCGGCATTGGCCCCATT V T A S G D V Y Y G P G G G T G I G P I

241 GCCAGACCTATAGAGCACGGCTTGGATTCGTCCACTGAAAATGGCTGGCAAGAGTTTGAA A R P I E H G L D S S T E N G W Q E F E

301 AGTTATGCTGATCTGGGCGTTGACCCCAGACGCTATGTTCCTCTTCAGGTTAAAGAAAAA S Y A D L G V D P R R Y V P L Q V K E K

361 CGCAGGGAGATCGAGCTTCAGTTCCGAGATGCCGAGAAAAAACTTGAGGCGTCGGTACAA R R E I E L Q F R D A E K K L E A S V Q

421 GCCGAGCTGGATAAGGCTGATGCCGCTCTTGGTCCGGCAAAGAATCTTGCACCATTGGAC A E L D K A D A A L G P A K N L A P L D

481 GTCATCAACCGCAGTCTGACCATCGTTGGAAACGCCCTCCAGCAAAAGAATCAAAAACTA V I N R S L T I V G N A L Q Q K N Q K L

541 CTGCTGAATCAGAAGAAGATTACCAGCCTGGGTGCAAAGAATTTCCTTACCCGTACGGCG L L N Q K K I T S L G A K N F L T R T A

601 GAAGAGATCGGTGAACAAGCGGTGCGAGAAGGCAATATTAACGGACCTGAAGCCTATATG E E I G E Q A V R E G N I N G P E A Y M

661 CGCTTCCTCGACAGGGAAATGGAAGGTCTCGCGGCAGCTTATAACGTAAAACTCTTCACC R F L D R E M E G L A A A Y N V K L F T

721 GAAGCGATCAGTAGTCTCCAGATCCGCATGAATACGTTGACCGCCGCCAAAGCAAGTATT E A I S S L Q I R M N T L T A A K A S I

781 GAGGCGGCCGCAGCAAACAAGGCGCGTGAACAAGCAGCGGCTGAGGCCAAACGCAAAGCC E A A A A N K A R E Q A A A E A K R K A

841 GAAGAGCAGGCCCGCCAGCAAGCGGCGATAAGAGCTGCCAATACCTATGCCATGCCGGCC E E Q A R Q Q A A I R A A N T Y A M P A

901 AATGGCAGCGTTGTCGCCACCGCCGCAGGCCGGGGTCTGATCCAGGTCGCACAAGGCGCC N G S V V A T A A G R G L I Q V A Q G A

961 GCATCCCTTGCTCAAGCGATCTCCGATGCGATTGCCGTCCTGGGCCGGGTCCTGGCTTCA A S L A Q A I S D A I A V L G R V L A S

1021 GCACCCTCGGTGATGGCCGTGGGCTTTGCCAGTCTGACCTACTCCTCCCGGACTGCCGAG A P S V M A V G F A S L T Y S S R T A E

1081 CAATGGCAGGACCAAACGCCCGATAGCGTTCGTTACGCCCTGGGCATGGATGCCGCTAAA Q W Q D Q T P D S V R Y A L G M D A A K

1141 TTGGGGCTTCCCCCAAGCGTAAACCTGAACGCGGTTGCAAAAGCCAGCGGTACCGTCGAT L G L P P S V N L N A V A K A S G T V D

1201 CTGCCGATGCGCCTGACCAACGAGGCACGAGGCAACACGACGACCCTTTCGGTGGTCAGC L P M R L T N E A R G N T T T L S V V S

167 Identification of pyocin S6

1261 ACCGATGGTGTGAGCGTTCCGAAAGCCGTTCCGATCCGGATGGCGGCCTACAATGCCACG T D G V S V P K A V P I R M A A Y N A T

1321 ACAGGCCTGTACGAGGTTACGGTTCCCTCTACGACCGCAGAAGCACCGCCACTGATCCTG T G L Y E V T V P S T T A E A P P L I L

1381 ACTTGGACGCCGGCGAGTCCTCCAGGAAACCAGAACCCTTCGAGTACCACTCCGGTCGTA T W T P A S P P G N Q N P S S T T P V V

1441 CCGAAGCCGGTGCCGGTATATGAGGGAGCGACCCTTACACCGGTGAAGGCCAAACCAGAA P K P V P V Y E G A T L T P V K A K P E

1501 ACCTATCCAGGAGTGATGACGCTACCGGATGATCTGATCATCGGCTTCCCGGCCGACTCG T Y P G V M T L P D D L I I G F P A D S

1561 GGGATCAAGCCGATCTATGTGATGATAAGCCGAGAGCACAGTTACCATCCTGCTCCGGCA G I K P I Y V M I S R E H S Y H P A P A

1621 ACACTACCTGCCTTCCCCGACGCCTTGAGAGCTAAGCCAAAAACCTCTATTCAAGGTGGT T L P A F P D A L R A K P K T S I Q G G

1681 GGTGGGTTGCGTAAGCGTTGGAAGGATAAAAAAGGGAATATCTATGAATGGGACTCTCAG G G L R K R W K D K K G N I Y E W D S Q

1741 CATGGCAGTGTTGAAATGTATGACAAGAGAGGGCGGCACTTGGGAGAGTTTAATTCTAAT H G S V E M Y D K R G R H L G E F N S N

1801 ACAGGTGCTCGCACGAAACCCGCTGATCCAAAAAGGAGTGTAGAACCATGACCGTATTGA T G A R T K P A D P K R S V E P * M T V L

1861 TTTCTGGATACCTAGCATCGGGAAATGATGACTCCTTGAAGTATGAAAAGGCCGTTCCTT

I S G Y L A S G N D D S L K Y E K A V P

1921 CCGGATGCATCTCGCAAGTTATGGAGGTGATGCGCTGGAAGAAAGGGGAGAATATTGAAG

S G C I S Q V M E V M R W K K G E N I E

1981 GTGAATATCCCGTCAAGGATAATGATGTCGTAAGGATTGAGGAGATTCTAGGCGAAAAAC

G E Y P V K D N D V V R I E E I L G E K

2041 TTCCGGTTGGATTGGAGTATTTTATTGGGGTTTACGCCTAGCTACAAGTTCTAGAGGAAG

L P V G L E Y F I G V Y A *

2101 GCCCTACTGCATGGGCCTTCTTTGTTAAGTTGGACAGCTATTGGTGGAGTGGTTTTCTTT

2161 GGTCTGATAGTTGTAAATAAAGAGAGCACTACTGGAAAGGCCTAGATTTGTA

Figure S5.2. Nucleotide sequences of the genes encoding the killing protein and immunity protein for pyocin S6 and predicted amino acid sequences. Singly and doubly underlined DNA sequences indicate the P-box and Shine-Dalgarno sequences, respectively. The characteristic ATTGnn(n)GTnn(n) motif repeats in the P-box are shown in bold. Start codons are highlighted in bold and doubly underlined. A putative transcription terminator is indicated by bold lines. Stop codons are indicated by an asterisk.

168 Chapter 5

Figure S5.3. Predicted secondary structure model of the entire pyocin S6 (PHYRE2).

169

170 Chapter 6

6 Chapter 6: General discussion and future perspectives

171 General discussion and future perspectives

172 Chapter 6

The life span of cystic fibrosis patients has been greatly extended due to the availability of pancreatic enzyme supplements, advanced airway clearance techniques and antibiotics. Whereas in 1980 the life expectancy of a CF patient was 14 years, it is now 35 years. However, there appears to be a typical pattern in the colonization of CF patients by particular pathogens. At very young ages, CF patients are often colonized with H. influenzae and/or S. aureus, while at the onset of adulthood, P. aeruginosa becomes the predominant pathogen. This implies that the majority of CF patients will be colonized by this bacterium at a particular moment during their lives. Once a patient is chronically colonized by P. aeruginosa, it is expected that the bacterium will persist during the rest of the patient’s life. Typically, an initial P. aeruginosa infection occurs through the acquisition of an environmental P. aeruginosa isolate that has no history in the CF lung environment. This strain is not well adapted to the hostile conditions present in the CF lungs and will cause a persistent inflammation. Ultimately, via the acquisition of several pathoadaptive mutations, this strain will adapt to the CF environment and concomitantly switch its lifestyle, from a planktonic growth to a biofilm mode of growth. Once this step has occurred, it is difficult to eradicate P. aeruginosa from the CF host. The ongoing interaction between P. aeruginosa and the host immune system, that is typically biased towards a pro-inflammatory response, results in irreversible lung damage and ultimately death of the CF patient. It can therefore be concluded that biofilm formation by P. aeruginosa, which is characterized by production of the viscous exopolysaccharide alginate, is the crucial step in determining the clinical outcome. With regard to the treatment of P. aeruginosa infections in CF patients, it is important to reduce the bacterial load, and hence dampen the immune reponse. Therefore, P. aeruginosa biofilms need to be disrupted in order to ensure optimal killing activity of antibiotics and the immune system. In this work, we have explored possible ways to reduce biofilm formation based on the introduction of shear stress into an in vitro model that mimicks the CF environment as well as directly into the CF lung. However, prior to study the effect of shear stress on biofilm formation, it was necessary to use a relevant P. aeruginosa strain that is well adapted to the CF lung conditions.

173 General discussion and future perspectives

Adaptation of P. aeruginosa to the CF lung occurs through deletion of TonB-dependent receptor genes and implies a shift towards the uptake of ferrous iron.

In the first chapter of this thesis we have first characterized the P. aeruginosa population present among CF patients attending the CF reference center of the UZ Brussel using a novel genotyping method. This method consisted of a combination of Rep-PCR, which is used to determine the clonally relatedness of P. aeruginosa isolates at the whole-genome level, and a multiplex PCR in which specific genes involved in ferripyoverdine uptake and pyocin synthesis, were amplified. Although a number of potential inter-patient transmissions were detected, the P. aeruginosa population present at the CF reference center of the UZ Brussel was diverse and was not dominated by a highly-transmissible clone. This indicates that the segregation of CF patients, the application of hygiene measures, as well as the antibiotic usage (since antibiotics can select for particular clones, reducing diversity (Cramer et al., 2012)) proved succesful in this CF reference center. Interestingly, when determining the ferripyoverdine receptor types of this P. aeruginosa CF population, we have observed that the fpvB gene, which is conserved among environmental P. aeruginosa isolates, was deleted in several P. aeruginosa strains. In addition, by screening the genome sequence of longitudinally sampled clonally related iolates from the Copenhagen CF reference center, we have shown that the deletion of ferripyoverdine receptor genes occurs frequently and that large deletions comprising these genes accumulate during adaptation of P. aeruginosa to the CF environment. One of the P. aeruginosa isolates in our collection, P. aeruginosa CF_PA39, that contained a fpvB deletion and was found to be transmissible, was subsequently selected for whole genome sequencing to reveal other deletions that may have occured during the adaptation process. Indeed, several large deletions (up to 58 kb) were identified comprising the entire gene cluster encoding the type III secretion system and genes involved in the production of virulence factors (Cif, HCN) as well as an additional TonB-dependent receptor. These results confirm the findings of other groups, that adaptation of P. aeruginosa to the CF lung involves genome reduction (Rau et al., 2012; Smith et al., 2006). Furthermore, we have shown that the deletion of ferripyoverdine receptor genes was not related to a higher resistance to any of the conventional antibiotics used to treat CF infections. It may therefore be another factor, such as a specific S-type pyocin, immune component or molecule produced by another CF pathogen, that selects for the loss of these receptors during colonization of the CF habitat. The deletion of ferripyoverdine receptor genes has consequences for the acquisition of iron by P. aeruginosa, since they are necessary to bind the ferri-siderophore.

174 Chapter 6

Therefore, it was suggested that alternative iron sources are available to P. aeruginosa in the CF lung. Indeed, it was confirmed in collaboration with the group of Dianne Newman, that multiple iron uptake systems are simultaneously used by P. aeruginosa, including ferrous iron uptake. Additionally, by comparing the expression of several iron uptake genes in CF sputum versus planktonic cultures of one of the P. aeruginosa isolates from the same sputum sample, we have shown that ferrous iron uptake significantly contributes to iron acquisition in the CF lung environment. It would be interesting to determine the expression of genes involved in iron uptake in sputum samples that were longitudinally collected from the same CF patients. Furthermore, recently colonized patients should be included in this study and lung functions should be measured. In this way, the evolution of iron uptake systems can be monitored and eventually linked to the disease state of the patient.

Introducing shear stress causes P. aeruginosa to switch from a biofilm to a planktonic lifestyle.

Previously it has been shown that growth of the P. aeruginosa PAO1 reference strain under low fluid shear conditions, obtained by means of the RWV bioreactor, in LB medium resulted in the formation of biofilms in suspension (Crabbé et al., 2008). However, since P. aeruginosa PAO1 is a wound isolate and LB medium is not representative for the viscous CF sputum, we have repeated this experiment using the highly adapted, transmissible P. aeruginosa CF_39 isolate and ASM. Interestingly, SEM revealed that robust biofilms were formed by this strain under LS conditions, while under HS conditions nearly all cells appeared to be planktonic. These phenotypic results were also reflected in the transcriptomic response of P. aeruginosa to shear stress, since genes involved in alginate biosynthesis, denitrification and stress response were up-regulated under LS. These results indicate that introducing shear stress in a CF-like environment can modulate the behavior of even a highly adapted P. aeruginosa CF strain, driving it from a biofilm mode of growth into a planktonic mode of growth. An experiment that would validate our RWV model as a standard model to study the behavior of P. aeruginosa in the CF lung, would be to inoculate a strain without a CF background (or eventually a strain that has only recently colonized a CF patient) in the RWV and allow the experiment to continue for several days or months while sampling aliquots from the bacterial culture at regular time intervals. Finally, whole-genome sequencing analysis of longitudinally collected isolates would reveal if mutations in genes that are known to be mutated in the CF lung are accumulated during growth under LS in ASM.

175 General discussion and future perspectives

The application of IPV at high frequency might disrupt P. aeruginosa biofilms in the CF lung and partially restore pulmonary function.

A possible way to increase fluid shear levels in the CF lung is the use of IPV. Although this airway clearance technique did not prove to enhance sputum clearance (Van Ginderdeuren et al., 2008), it might disrupt biofilm formation by P. aeruginosa by introducing shear stress in the CF lung. In order to test this, we have performed a clinical study comprising eight different subjects, from whom four have been chronically colonized with P. aeruginosa, in order to determine the effect of three different treatments on pulmonary function, bacterial clearance, and P. aeruginosa gene expression. Although no significant difference was found between the three treatments (however, approaching the level of significance), lung function clearly improved using AD and IPV at high frequency, but not IPV at low frequency. Interestingly, IPV at high frequency was more potent than AD in enhancing the lung function of two of the CF patients that were chronically colonized with P. aeruginosa. Furthermore, for one of these patients, IPV at high frequency was able to cause a decrease in the bacterial sputum load and resulted in the up-regulation of planktonic marker genes. Therefore, it can be concluded that IPV at high frequency might enhance the diruption of P. aeruginosa biofilms in the CF lung, most probably by increasing the level of fluid shear in this environment. Nevertheless, this study should be repeated with a larger number of individuals from whom lung functions should be measured continuously in order to allow the comparison of paired data when comparing different treatments. Further, it may be promising to combine the IPV treatment at high frequency with the addition of (FDA-approved) chelators that are specific for Fe2+ as well as Fe3+ during a clinical study as these treatments could act synergystically in order to disrupt biofilm formation by P. aeruginosa.

Identification of pyocin S6

Screening the genome sequence of P. aeruginosa CF_PA39, we were able to identify a novel S-type pyocin that shared the receptor binding and translocation domains with pyocin S1, but instead had a Colicin E3-like killing domain. Since the catalytic residues at its active site have been conserved, it is likely that this pyocin has an rRNase killing activity. However, in order to test this, these catalytic residues (D510, H513, E517 and R545) could be substituted with a different amino acid (i.e. alanine) via site-directed mutagenesis to confirm their role in the enzymatic activity of pyocin S6, providing evidence that it has rRNase activity. Subsequently, mutated versions of the pyocin as well as the wild-type could be incubated with purified

176 Chapter 6 bacterial total RNA and degradation of the 16S rRNA could be measured by subjecting the incubated total RNA sample to denaturing agarose gel electrophoresis. In addition, cell-free supernatant (or purified pyocin) of the mutants as well as the wild-type should be spotted on a sensitive P. aeruginosa strain to confirm the importance of each potential catalytic residue in the killing activity of the pyocin.

It was shown that pyocin S6 expression was enhanced under iron-poor conditions, while expression of the downstream immunity gene was present under both iron-poor and iron- abundant conditions (but higher under iron-poor conditions), although these genes are organized as an operon. This suggests an iron-dependent regulatory mechanism, most probably involving small RNAs. Recently, its has been shown that three antisense RNAs (asRNAs) are present in the pyocin S3 gene and one in the pyocin S5 gene, indicating that these asRNAs might regulate mRNA stability of the pyocin trancripts (Gomez-Lozano et al., 2014). In order to confirm the presence of an asRNA, RNAseq of RNA extracted from P. aeruginosa CF_PA39 grown under iron-limited versus iron-abundant conditions could be performed.

Finally we have demonstrated the killing activity of pyocin S6 towards a number of P. aeruginosa CF isolates. Interestingly, a number of P. aeruginosa strains (including PAO1 and DK2) harbor the pyocin S6 immunity gene, strongly indicating that the pyocin S6 operon might once have been present in these strains, while the immunity gene is still present as an “orphan” gene. Furthermore, this gene might offer a competitive advantage to the host strain in protecting itself against the killing action of pyocins produced by other P. aeruginosa strains. It can be concluded that pyocins are interesting molecules that could be used as narrow-spectrum antibiotics and exhibit an increased killing activity under iron-poor conditions, that prevail in several human infections. However, the application of S-type pyocins to cure P. aeruginosa infections is restricted to topical usage since these molecules are protease-sensitive. Therefore, the application of a combination (“cocktail”) of several S- type pyocins targeting different (ferri-siderophore) receptors on burn wounds could be a promising therapy, without the risk of evoking antibiotic resistance.

177

178 References

7 References

Abdel-Mawgoud, A. M., Lepine, F. & Deziel, E. (2010). Rhamnolipids: diversity of structures, microbial origins and roles. Appl Microbiol Biotechnol 86, 1323-1336. Aendekerk, S., Ghysels, B., Cornelis, P. & Baysse, C. (2002). Characterization of a new efflux pump, MexGHI-OpmD, from Pseudomonas aeruginosa that confers resistance to vanadium. Microbiology 148, 2371-2381. Agostini, P. & Knowles, N. (2007). Autogenic drainage: the technique, physiological basis and evidence. Physiotherapy 93, 157-163. Ahlgren, H. G., Benedetti, A., Landry, J. S., Bernier, J., Matouk, E., Radzioch, D., Lands, L. C., Rousseau, S. & Nguyen, D. (2015). Clinical outcomes associated with Staphylococcus aureus and Pseudomonas aeruginosa airway infections in adult cystic fibrosis patients. BMC Pulm Med 15, 67. Alhede, M., Bjarnsholt, T., Jensen, P. O., Phipps, R. K., Moser, C., Christophersen, L., Christensen, L. D., van Gennip, M., Parsek, M., Hoiby, N., Rasmussen, T. B. & Givskov, M. (2009). Pseudomonas aeruginosa recognizes and responds aggressively to the presence of polymorphonuclear leukocytes. Microbiology 155, 3500-3508. Allen, C. A., Niesel, D. W. & Torres, A. G. (2008). The effects of low-shear stress on Adherent-invasive Escherichia coli. Environ Microbiol 10, 1512-1525. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. J Mol Biol 215, 403-410. Andersen, D. H. (1938). Cystic fibrosis of the pancreas and its relation to celiac disease: a clinical and pathologic study. Am J Dis Child 56, 344-399. Andersen, D. H. & Hodges, R. G. (1946). Celiac syndrome; genetics of cystic fibrosis of the pancreas, with a consideration of etiology. Am J Dis Child 72, 62-80. Andersen, S. B., Marvig, R. L., Molin, S., Krogh Johansen, H. & Griffin, A. S. (2015). Long-term social dynamics drive loss of function in pathogenic bacteria. Proc Natl Acad Sci U S A 112, 10756-10761.

Anthony, M., Rose, B., Pegler, M. B., Elkins, M., Service, H., Thamotharampillai, K., Watson, J., Robinson, M., Bye, P., Merlino, J. & Harbour, C. (2002). Genetic analysis of Pseudomonas aeruginosa isolates from the sputa of Australian adult cystic fibrosis patients. J Clin Microbiol 40, 2772-2778. Anwar, H., Strap, J. L. & Costerton, J. W. (1992). Susceptibility of biofilm cells of Pseudomonas aeruginosa to bactericidal actions of whole blood and serum. FEMS Microbiol Lett 71, 235-241. Aziz, R. K., Bartels, D., Best, A. A., DeJongh, M., Disz, T., Edwards, R. A., Formsma, K., Gerdes, S., Glass, E. M., Kubal, M., Meyer, F., Olsen, G. J., Olson, R., Osterman, A. L., Overbeek, R. A., McNeil, L. K., Paarmann, D., Paczian, T., Parrello, B., Pusch, G. D., Reich, C., Stevens, R., Vassieva, O., Vonstein, V., Wilke, A. & Zagnitko, O. (2008). The RAST Server: rapid annotations using subsystems technology. BMC Genomics 9, 75.

179 References

Bainbridge, T. & Fick, R. B., Jr. (1989). Functional importance of cystic fibrosis immunoglobulin G fragments generated by Pseudomonas aeruginosa elastase. J Lab Clin Med 114, 728-733. Balasubramanian, D., Schneper, L., Kumari, H. & Mathee, K. (2013). A dynamic and intricate regulatory network determines Pseudomonas aeruginosa virulence. Nucleic Acids Res 41, 1-20. Balasubramanian, D. & Vanderpool, C. K. (2013). New developments in post- transcriptional regulation of operons by small RNAs. RNA Biol 10, 337-341. Banin, E., Vasil, M. L. & Greenberg, E. P. (2005). Iron and Pseudomonas aeruginosa biofilm formation. Proc Natl Acad Sci U S A 102, 11076-11081. Barker, A. P., Vasil, A. I., Filloux, A., Ball, G., Wilderman, P. J. & Vasil, M. L. (2004). A novel extracellular phospholipase C of Pseudomonas aeruginosa is required for phospholipid chemotaxis. Mol Microbiol 53, 1089-1098. Baron, S. S. & Rowe, J. J. (1981). Antibiotic action of pyocyanin. Antimicrob Agents Chemother 20, 814-820. Barreteau, H., Bouhss, A., Fourgeaud, M., Mainardi, J. L., Touze, T., Gerard, F., Blanot, D., Arthur, M. & Mengin-Lecreulx, D. (2009). Human- and plant-pathogenic Pseudomonas species produce bacteriocins exhibiting colicin M-like hydrolase activity towards peptidoglycan precursors. J Bacteriol 191, 3657-3664.

Barthe, C., Nandakumar, S., Derlich, L., Macey, J., Bui, S., Fayon, M., Crouzet, M., Garbay, B., Vilain, S. & Costaglioli, P. (2015). Exploring the expression of Pseudomonas aeruginosa genes directly from sputa of cystic fibrosis patients. Lett Appl Microbiol. Baysse, C., Meyer, J. M., Plesiat, P., Geoffroy, V., Michel-Briand, Y. & Cornelis, P. (1999). Uptake of pyocin S3 occurs through the outer membrane ferripyoverdine type II receptor of Pseudomonas aeruginosa. J Bacteriol 181, 3849-3851. Bielecki, P., Komor, U., Bielecka, A., Musken, M., Puchalka, J., Pletz, M. W., Ballmann, M., Martins dos Santos, V. A., Weiss, S. & Haussler, S. (2013). Ex vivo transcriptional profiling reveals a common set of genes important for the adaptation of Pseudomonas aeruginosa to chronically infected host sites. Environ Microbiol 15, 570-587. Bjarnsholt, T., Jensen, P. O., Jakobsen, T. H., Phipps, R., Nielsen, A. K., Rybtke, M. T., Tolker-Nielsen, T., Givskov, M., Hoiby, N., Ciofu, O. & Scandinavian Cystic Fibrosis Study, C. (2010). Quorum sensing and virulence of Pseudomonas aeruginosa during lung infection of cystic fibrosis patients. PLoS One 5, e10115. Blake, J. (1973). A note on mucus shear rates. Respir Physiol 17, 394-399. Blumer, C. & Haas, D. (2000). Mechanism, regulation, and ecological role of bacterial cyanide biosynthesis. Arch Microbiol 173, 170-177. Bodilis, J., Ghysels, B., Osayande, J., Matthijs, S., Pirnay, J. P., Denayer, S., De Vos, D. & Cornelis, P. (2009). Distribution and evolution of ferripyoverdine receptors in Pseudomonas aeruginosa. Environ Microbiol 11, 2123-2135.

180 References

Boetzer, M., Henkel, C. V., Jansen, H. J., Butler, D. & Pirovano, W. (2011). Scaffolding pre-assembled contigs using SSPACE. Bioinformatics 27, 578-579. Boetzer, M. & Pirovano, W. (2012). Toward almost closed genomes with GapFiller. Genome Biol 13, R56. Boles, B. R., Thoendel, M. & Singh, P. K. (2005). Rhamnolipids mediate detachment of Pseudomonas aeruginosa from biofilms. Mol Microbiol 57, 1210-1223. Boyle, M. P. & De Boeck, K. (2013). A new era in the treatment of cystic fibrosis: correction of the underlying CFTR defect. Lancet Respir Med 1, 158-163. Briard, B., Bomme, P., Lechner, B. E., Mislin, G. L., Lair, V., Prevost, M. C., Latge, J. P., Haas, H. & Beauvais, A. (2015). Pseudomonas aeruginosa manipulates redox and iron homeostasis of its microbiota partner Aspergillus fumigatus via phenazines. Sci Rep 5, 8220. Brillet, K., Journet, L., Celia, H., Paulus, L., Stahl, A., Pattus, F. & Cobessi, D. (2007). A beta strand lock exchange for signal transduction in TonB-dependent transducers on the basis of a common structural motif. Structure 15, 1383-1391. Burns, J. L., Gibson, R. L., McNamara, S., Yim, D., Emerson, J., Rosenfeld, M., Hiatt, P., McCoy, K., Castile, R., Smith, A. L. & Ramsey, B. W. (2001). Longitudinal assessment of Pseudomonas aeruginosa in young children with cystic fibrosis. J Infect Dis 183, 444-452. Caiazza, N. C., Shanks, R. M. & O'Toole, G. A. (2005). Rhamnolipids modulate swarming motility patterns of Pseudomonas aeruginosa. J Bacteriol 187, 7351-7361. Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer, K. & Madden, T. L. (2009). BLAST+: architecture and applications. BMC Bioinformatics 10, 421. Carterson, A. J., Honer zu Bentrup, K., Ott, C. M., Clarke, M. S., Pierson, D. L., Vanderburg, C. R., Buchanan, K. L., Nickerson, C. A. & Schurr, M. J. (2005). A549 lung epithelial cells grown as three-dimensional aggregates: alternative tissue culture model for Pseudomonas aeruginosa pathogenesis. Infect Immun 73, 1129-1140. Castro, S. L., Nelman-Gonzalez, M., Nickerson, C. A. & Ott, C. M. (2011). Induction of attachment-independent biofilm formation and repression of Hfq expression by low-fluid- shear culture of Staphylococcus aureus. Appl Environ Microbiol 77, 6368-6378. Chang, Y. S., Klockgether, J. & Tummler, B. (2007). An intragenic deletion in pilQ leads to nonpiliation of a Pseudomonas aeruginosa strain isolated from cystic fibrosis lung. FEMS Microbiol Lett 270, 201-206. Chen, L., Zou, Y., She, P. & Wu, Y. (2015). Composition, function, and regulation of T6SS in Pseudomonas aeruginosa. Microbiol Res 172, 19-25. Cheng, K., Smyth, R. L., Govan, J. R., Doherty, C., Winstanley, C., Denning, N., Heaf, D. P., van Saene, H. & Hart, C. A. (1996). Spread of beta-lactam-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. Lancet 348, 639-642. Cheng, Y. S., Shi, Z., Doudeva, L. G., Yang, W. Z., Chak, K. F. & Yuan, H. S. (2006). High-resolution crystal structure of a truncated ColE7 translocation domain: implications for colicin transport across membranes. J Mol Biol 356, 22-31.

181 References

Chevaillier, J. (1984). Autogenic drainage. In Cystic fibrosis horizons, pp. 235. Edited by D. Lawson. Chichester: John Wiley. Chung, J. C., Rzhepishevska, O., Ramstedt, M. & Welch, M. (2013). Type III secretion system expression in oxygen-limited Pseudomonas aeruginosa cultures is stimulated by isocitrate lyase activity. Open Biol 3, 120131. Clarke, C. A., Scheurwater, E. M. & Clarke, A. J. (2010). The vertebrate lysozyme inhibitor Ivy functions to inhibit the activity of lytic transglycosylase. J Biol Chem 285, 14843-14847. Cohen, T. S. & Prince, A. (2012). Cystic fibrosis: a mucosal immunodeficiency syndrome. Nat Med 18, 509-519.

Corech, R., Rao, A., Laxova, A., Moss, J., Rock, M. J., Li, Z., Kosorok, M. R., Splaingard, M. L., Farrell, P. M. & Barbieri, J. T. (2005). Early immune response to the components of the type III system of Pseudomonas aeruginosa in children with cystic fibrosis. J Clin Microbiol 43, 3956-3962. Cornelis, P., Hohnadel, D. & Meyer, J. M. (1989). Evidence for different pyoverdine- mediated iron uptake systems among Pseudomonas aeruginosa strains. Infect Immun 57, 3491-3497. Cornelis, P. & Bodilis, J. (2009). A survey of TonB-dependent receptors in fluorescent pseudomonads. Environ Microbiol Rep 1, 256-262. Cornelis, P. (2010). Iron uptake and metabolism in pseudomonads. Appl Microbiol Biotechnol 86, 1637-1645. Cornelis, P. & Dingemans, J. (2013). Pseudomonas aeruginosa adapts its iron uptake strategies in function of the type of infections. Front Cell Infect Microbiol 3, 75. Cornelis, P. (2014). Iron uptake and homeostasis in prokaryotic Microorganisms. In Binding, uptake and storage of metal ions in biological cells, pp. 303-332. Edited by W. Maret & A. Wedd. Cambridge, UK: Royal Society of Chemistry. Cosgrove, S., Chotirmall, S. H., Greene, C. M. & McElvaney, N. G. (2011). Pulmonary proteases in the cystic fibrosis lung induce interleukin 8 expression from bronchial epithelial cells via a heme/meprin/epidermal growth factor receptor/Toll-like receptor pathway. J Biol Chem 286, 7692-7704.

Costa, T. R., Felisberto-Rodrigues, C., Meir, A., Prevost, M. S., Redzej, A., Trokter, M. & Waksman, G. (2015). Secretion systems in Gram-negative bacteria: structural and mechanistic insights. Nat Rev Microbiol 13, 343-359. Coulon, C., Vinogradov, E., Filloux, A. & Sadovskaya, I. (2010). Chemical analysis of cellular and extracellular carbohydrates of a biofilm-forming strain Pseudomonas aeruginosa PA14. PLoS One 5, e14220. Cowles, K. N. & Gitai, Z. (2010). Surface association and the MreB cytoskeleton regulate pilus production, localization and function in Pseudomonas aeruginosa. Mol Microbiol 76, 1411-1426.

182 References

Cox, C. D., Rinehart, K. L., Jr., Moore, M. L. & Cook, J. C., Jr. (1981). Pyochelin: novel structure of an iron-chelating growth promoter for Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 78, 4256-4260. Crabbé, A., De Boever, P., Van Houdt, R., Moors, H., Mergeay, M. & Cornelis, P. (2008). Use of the rotating wall vessel technology to study the effect of shear stress on growth behaviour of Pseudomonas aeruginosa PA01. Environ Microbiol 10, 2098-2110. Crabbé, A., Pycke, B., Van Houdt, R., Monsieurs, P., Nickerson, C., Leys, N. & Cornelis, P. (2010). Response of Pseudomonas aeruginosa PAO1 to low shear modelled microgravity involves AlgU regulation. Environ Microbiol 12, 1545-1564. Crabbé, A., Sarker, S. F., Van Houdt, R., Ott, C. M., Leys, N., Cornelis, P. & Nickerson, C. A. (2011a). Alveolar epithelium protects macrophages from quorum sensing-induced cytotoxicity in a three-dimensional co-culture model. Cell Microbiol 13, 469-481. Crabbé, A., Schurr, M. J., Monsieurs, P., Morici, L., Schurr, J., Wilson, J. W., Ott, C. M., Tsaprailis, G., Pierson, D. L., Stefanyshyn-Piper, H. & Nickerson, C. A. (2011b). Transcriptional and proteomic responses of Pseudomonas aeruginosa PAO1 to spaceflight conditions involve Hfq regulation and reveal a role for oxygen. Appl Environ Microbiol 77, 1221-1230. Crabbé, A., Leroy, B., Wattiez, R., Aertsen, A., Leys, N., Cornelis, P. & Van Houdt, R. (2012). Differential proteomics and physiology of Pseudomonas putida KT2440 under filament-inducing conditions. BMC Microbiol 12, 282. Cramer, N., Wiehlmann, L., Ciofu, O., Tamm, S., Hoiby, N. & Tummler, B. (2012). Molecular epidemiology of chronic Pseudomonas aeruginosa airway infections in cystic fibrosis. PLoS One 7, e50731. Cunningham, L. & Williams, H. D. (1995). Isolation and characterization of mutants defective in the cyanide-insensitive respiratory pathway of Pseudomonas aeruginosa. J Bacteriol 177, 432-438. Cutting, G. R. (2015). Cystic fibrosis genetics: from molecular understanding to clinical application. Nat Rev Genet 16, 45-56. Damkiaer, S., Yang, L., Molin, S. & Jelsbak, L. (2013). Evolutionary remodeling of global regulatory networks during long-term bacterial adaptation to human hosts. Proc Natl Acad Sci U S A 110, 7766-7771. Daniels, J. B., Scoffield, J., Woolnough, J. L. & Silo-Suh, L. (2014). Impact of glycerol-3- phosphate dehydrogenase on virulence factor production by Pseudomonas aeruginosa. Can J Microbiol 60, 857-863. Dark, M. J. (2013). Whole-genome sequencing in bacteriology: state of the art. Infect Drug Resist 6, 115-123. Davey, M. E., Caiazza, N. C. & O'Toole, G. A. (2003). Rhamnolipid surfactant production affects biofilm architecture in Pseudomonas aeruginosa PAO1. J Bacteriol 185, 1027-1036. De Boeck, K., Zolin, A., Cuppens, H., Olesen, H. V. & Viviani, L. (2014). The relative frequency of CFTR mutation classes in European patients with cystic fibrosis. J Cyst Fibros 13, 403-409.

183 References de Chial, M., Ghysels, B., Beatson, S. A., Geoffroy, V., Meyer, J. M., Pattery, T., Baysse, C., Chablain, P., Parsons, Y. N., Winstanley, C., Cordwell, S. J. & Cornelis, P. (2003). Identification of type II and type III pyoverdine receptors from Pseudomonas aeruginosa. Microbiology 149, 821-831. De Lay, N., Schu, D. J. & Gottesman, S. (2013). Bacterial small RNA-based negative regulation: Hfq and its accomplices. J Biol Chem 288, 7996-8003. De Vos, D., Lim, A., Jr., Pirnay, J. P., Struelens, M., Vandenvelde, C., Duinslaeger, L., Vanderkelen, A. & Cornelis, P. (1997). Direct detection and identification of Pseudomonas aeruginosa in clinical samples such as skin biopsy specimens and expectorations by multiplex PCR based on two outer membrane lipoprotein genes, oprI and oprL. J Clin Microbiol 35, 1295-1299. De Vos, D., De Chial, M., Cochez, C., Jansen, S., Tummler, B., Meyer, J. M. & Cornelis, P. (2001). Study of pyoverdine type and production by Pseudomonas aeruginosa isolated from cystic fibrosis patients: prevalence of type II pyoverdine isolates and accumulation of pyoverdine-negative mutations. Arch Microbiol 175, 384-388. Dekkers, J. F., Wiegerinck, C. L., de Jonge, H. R., Bronsveld, I., Janssens, H. M., de Winter-de Groot, K. M., Brandsma, A. M., de Jong, N. W., Bijvelds, M. J., Scholte, B. J., Nieuwenhuis, E. E., van den Brink, S., Clevers, H., van der Ent, C. K., Middendorp, S. & Beekman, J. M. (2013). A functional CFTR assay using primary cystic fibrosis intestinal organoids. Nat Med 19, 939-945. Denayer, S., Matthijs, S. & Cornelis, P. (2007). Pyocin S2 (Sa) kills Pseudomonas aeruginosa strains via the FpvA type I ferripyoverdine receptor. J Bacteriol 189, 7663-7668. Deplano, A., Denis, O., Rodriguez-Villalobos, H., De Ryck, R., Struelens, M. J. & Hallin, M. (2011). Controlled performance evaluation of the DiversiLab repetitive-sequence-based genotyping system for typing multidrug-resistant health care-associated bacterial pathogens. J Clin Microbiol 49, 3616-3620. Dietrich, L. E., Price-Whelan, A., Petersen, A., Whiteley, M. & Newman, D. K. (2006). The phenazine pyocyanin is a terminal signalling factor in the quorum sensing network of Pseudomonas aeruginosa. Mol Microbiol 61, 1308-1321. Diggle, S. P., Cornelis, P., Williams, P. & Camara, M. (2006). 4-quinolone signalling in Pseudomonas aeruginosa: old molecules, new perspectives. Int J Med Microbiol 296, 83-91. Diggle, S. P., Matthijs, S., Wright, V. J., Fletcher, M. P., Chhabra, S. R., Lamont, I. L., Kong, X., Hider, R. C., Cornelis, P., Camara, M. & Williams, P. (2007). The Pseudomonas aeruginosa 4-quinolone signal molecules HHQ and PQS play multifunctional roles in quorum sensing and iron entrapment. Chem Biol 14, 87-96. Dingemans, J., Ye, L., Hildebrand, F., Tontodonati, F., Craggs, M., Bilocq, F., De Vos, D., Crabbé, A., Van Houdt, R., Malfroot, A. & Cornelis, P. (2014). The deletion of TonB- dependent receptor genes is part of the genome reduction process that occurs during adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung. Pathog Dis 71, 26-38. Dodge, J. A., Lewis, P. A., Stanton, M. & Wilsher, J. (2007). Cystic fibrosis mortality and survival in the UK: 1947-2003. Eur Respir J 29, 522-526.

184 References

Doleans-Jordheim, A., Cournoyer, B., Bergeron, E., Croize, J., Salord, H., Andre, J., Mazoyer, M. A., Renaud, F. N. & Freney, J. (2009). Reliability of Pseudomonas aeruginosa semi-automated rep-PCR genotyping in various epidemiological situations. Eur J Clin Microbiol Infect Dis 28, 1105-1111. Dong, Y. H., Zhang, X. F., Xu, J. L., Tan, A. T. & Zhang, L. H. (2005). VqsM, a novel AraC-type global regulator of quorum-sensing signalling and virulence in Pseudomonas aeruginosa. Mol Microbiol 58, 552-564. Dotsch, A., Eckweiler, D., Schniederjans, M., Zimmermann, A., Jensen, V., Scharfe, M., Geffers, R. & Haussler, S. (2012). The Pseudomonas aeruginosa transcriptome in planktonic cultures and static biofilms using RNA sequencing. PLoS One 7, e31092. Douzi, B., Filloux, A. & Voulhoux, R. (2012). On the path to uncover the bacterial type II secretion system. Philos Trans R Soc Lond B Biol Sci 367, 1059-1072. Doyle, T. B., Hawkins, A. C. & McCarter, L. L. (2004). The complex flagellar torque generator of Pseudomonas aeruginosa. J Bacteriol 186, 6341-6350. Draper, R. C., Martin, L. W., Beare, P. A. & Lamont, I. L. (2011). Differential proteolysis of sigma regulators controls cell-surface signalling in Pseudomonas aeruginosa. Mol Microbiol 82, 1444-1453. Drees, S. L. & Fetzner, S. (2015). PqsE of Pseudomonas aeruginosa Acts as Pathway- Specific Thioesterase in the Biosynthesis of Alkylquinolone Signaling Molecules. Chem Biol 22, 611-618. Dumas, Z., Ross-Gillespie, A. & Kummerli, R. (2013). Switching between apparently redundant iron-uptake mechanisms benefits bacteria in changeable environments. Proc Biol Sci 280, 20131055. Duport, C., Baysse, C. & Michel-Briand, Y. (1995). Molecular characterization of pyocin S3, a novel S-type pyocin from Pseudomonas aeruginosa. J Biol Chem 270, 8920-8927. Elfarash, A., Wei, Q. & Cornelis, P. (2012). The soluble pyocins S2 and S4 from Pseudomonas aeruginosa bind to the same FpvAI receptor. Microbiologyopen 1, 268-275. Elfarash, A., Dingemans, J., Ye, L., Hassan, A. A., Craggs, M., Reimmann, C., Thomas, M. S. & Cornelis, P. (2014). Pore-forming pyocin S5 utilizes the FptA ferripyochelin receptor to kill Pseudomonas aeruginosa. Microbiology 160, 261-269. Errington, J. (2015). Bacterial morphogenesis and the enigmatic MreB helix. Nat Rev Microbiol 13, 241-248. Farrow, J. M., 3rd, Sund, Z. M., Ellison, M. L., Wade, D. S., Coleman, J. P. & Pesci, E. C. (2008). PqsE functions independently of PqsR-Pseudomonas quinolone signal and enhances the rhl quorum-sensing system. J Bacteriol 190, 7043-7051. Feliziani, S., Lujan, A. M., Moyano, A. J., Sola, C., Bocco, J. L., Montanaro, P., Canigia, L. F., Argarana, C. E. & Smania, A. M. (2010). Mucoidy, quorum sensing, mismatch repair and antibiotic resistance in Pseudomonas aeruginosa from cystic fibrosis chronic airways infections. PLoS One 5.

185 References

Feltman, H., Schulert, G., Khan, S., Jain, M., Peterson, L. & Hauser, A. R. (2001). Prevalence of type III secretion genes in clinical and environmental isolates of Pseudomonas aeruginosa. Microbiology 147, 2659-2669. Filloux, A. (2011). Protein Secretion Systems in Pseudomonas aeruginosa: An Essay on Diversity, Evolution, and Function. Front Microbiol 2, 155. Fluit, A. C., Terlingen, A. M., Andriessen, L., Ikawaty, R., van Mansfeld, R., Top, J., Cohen Stuart, J. W., Leverstein-van Hall, M. A. & Boel, C. H. (2010). Evaluation of the DiversiLab system for detection of hospital outbreaks of infections by different bacterial species. J Clin Microbiol 48, 3979-3989. Folkesson, A., Jelsbak, L., Yang, L., Johansen, H. K., Ciofu, O., Hoiby, N. & Molin, S. (2012). Adaptation of Pseudomonas aeruginosa to the cystic fibrosis airway: an evolutionary perspective. Nat Rev Microbiol 10, 841-851. Forstner, K. U., Vogel, J. & Sharma, C. M. (2014). READemption-a tool for the computational analysis of deep-sequencing-based transcriptome data. Bioinformatics 30, 3421-3423. Fothergill, J. L., Panagea, S., Hart, C. A., Walshaw, M. J., Pitt, T. L. & Winstanley, C. (2007). Widespread pyocyanin over-production among isolates of a cystic fibrosis epidemic strain. BMC Microbiol 7, 45. Fothergill, J. L., Walshaw, M. J. & Winstanley, C. (2012). Transmissible strains of Pseudomonas aeruginosa in cystic fibrosis lung infections. Eur Respir J 40, 227-238.

Frank, J. A., Reich, C. I., Sharma, S., Weisbaum, J. S., Wilson, B. A. & Olsen, G. J. (2008). Critical evaluation of two primers commonly used for amplification of bacterial 16S rRNA genes. Appl Environ Microbiol 74, 2461-2470. Franklin, M. J., Nivens, D. E., Weadge, J. T. & Howell, P. L. (2011). Biosynthesis of the Pseudomonas aeruginosa Extracellular Polysaccharides, Alginate, Pel, and Psl. Front Microbiol 2, 167. Fraud, S., Campigotto, A. J., Chen, Z. & Poole, K. (2008). MexCD-OprJ multidrug efflux system of Pseudomonas aeruginosa: involvement in chlorhexidine resistance and induction by membrane-damaging agents dependent upon the AlgU stress response sigma factor. Antimicrob Agents Chemother 52, 4478-4482. Frimmersdorf, E., Horatzek, S., Pelnikevich, A., Wiehlmann, L. & Schomburg, D. (2010). How Pseudomonas aeruginosa adapts to various environments: a metabolomic approach. Environ Microbiol 12, 1734-1747. Fung, C., Naughton, S., Turnbull, L., Tingpej, P., Rose, B., Arthur, J., Hu, H., Harmer, C., Harbour, C., Hassett, D. J., Whitchurch, C. B. & Manos, J. (2010). Gene expression of Pseudomonas aeruginosa in a mucin-containing synthetic growth medium mimicking cystic fibrosis lung sputum. J Med Microbiol 59, 1089-1100. Fuqua, C., Winans, S. C. & Greenberg, E. P. (1996). Census and consensus in bacterial ecosystems: the LuxR-LuxI family of quorum-sensing transcriptional regulators. Annu Rev Microbiol 50, 727-751. Gacesa, P. (1988). Alginates. Carbohydr Polym 8, 161-182.

186 References

Gallagher, L. A. & Manoil, C. (2001). Pseudomonas aeruginosa PAO1 kills Caenorhabditis elegans by cyanide poisoning. J Bacteriol 183, 6207-6214. Gallagher, L. A., McKnight, S. L., Kuznetsova, M. S., Pesci, E. C. & Manoil, C. (2002). Functions required for extracellular quinolone signaling by Pseudomonas aeruginosa. J Bacteriol 184, 6472-6480. Ge, P., Scholl, D., Leiman, P. G., Yu, X., Miller, J. F. & Zhou, Z. H. (2015). Atomic structures of a bactericidal contractile nanotube in its pre- and postcontraction states. Nat Struct Mol Biol 22, 377-382. Gellatly, S. L. & Hancock, R. E. (2013). Pseudomonas aeruginosa: new insights into pathogenesis and host defenses. Pathog Dis 67, 159-173. Ghequire, M. G. & De Mot, R. (2014). Ribosomally encoded antibacterial proteins and peptides from Pseudomonas. FEMS Microbiol Rev 38, 523-568. Ghysels, B., Dieu, B. T., Beatson, S. A., Pirnay, J. P., Ochsner, U. A., Vasil, M. L. & Cornelis, P. (2004). FpvB, an alternative type I ferripyoverdine receptor of Pseudomonas aeruginosa. Microbiology 150, 1671-1680. Goddard, A. F., Staudinger, B. J., Dowd, S. E., Joshi-Datar, A., Wolcott, R. D., Aitken, M. L., Fligner, C. L. & Singh, P. K. (2012). Direct sampling of cystic fibrosis lungs indicates that DNA-based analyses of upper-airway specimens can misrepresent lung microbiota. Proc Natl Acad Sci U S A 109, 13769-13774. Gomez-Lozano, M., Marvig, R. L., Tulstrup, M. V. & Molin, S. (2014). Expression of antisense small RNAs in response to stress in Pseudomonas aeruginosa. BMC Genomics 15, 783. Gooderham, W. J. & Hancock, R. E. (2009). Regulation of virulence and antibiotic resistance by two-component regulatory systems in Pseudomonas aeruginosa. FEMS Microbiol Rev 33, 279-294. Gottesman, S. (2005). Micros for microbes: non-coding regulatory RNAs in bacteria. Trends Genet 21, 399-404. Grant, J. R. & Stothard, P. (2008). The CGView Server: a comparative genomics tool for circular genomes. Nucleic Acids Res 36, W181-184. Grissa, I., Vergnaud, G. & Pourcel, C. (2007). CRISPRFinder: a web tool to identify clustered regularly interspaced short palindromic repeats. Nucleic Acids Res 35, W52-57. Gross, H. & Loper, J. E. (2009). Genomics of secondary metabolite production by Pseudomonas spp. Nat Prod Rep 26, 1408-1446. Hagins, J. M., Locy, R. & Silo-Suh, L. (2009). Isocitrate lyase supplies precursors for hydrogen cyanide production in a cystic fibrosis isolate of Pseudomonas aeruginosa. J Bacteriol 191, 6335-6339. Hanahan, D. (1985). Techniques for transformation of E. coli. In DNA Cloning, a Practical Approach, pp. 109-135. Edited by D. M. Glover. London, UK: IRL Press, Ltd. .

187 References

Hancock, R. E. & Brinkman, F. S. (2002). Function of pseudomonas porins in uptake and efflux. Annu Rev Microbiol 56, 17-38. Hannauer, M., Schafer, M., Hoegy, F., Gizzi, P., Wehrung, P., Mislin, G. L., Budzikiewicz, H. & Schalk, I. J. (2012). Biosynthesis of the pyoverdine siderophore of Pseudomonas aeruginosa involves precursors with a myristic or a myristoleic acid chain. FEBS Lett 586, 96-101. Hansch, G. M., Prior, B., Brenner-Weiss, G., Obst, U. & Overhage, J. (2014). The Pseudomonas quinolone signal (PQS) stimulates chemotaxis of polymorphonuclear neutrophils. J Appl Biomater Funct Mater 12, 21-26. Hansen, S. K., Rau, M. H., Johansen, H. K., Ciofu, O., Jelsbak, L., Yang, L., Folkesson, A., Jarmer, H. O., Aanaes, K., von Buchwald, C., Hoiby, N. & Molin, S. (2012). Evolution and diversification of Pseudomonas aeruginosa in the paranasal sinuses of cystic fibrosis children have implications for chronic lung infection. ISME J 6, 31-45. Harrison, F. (2007). Microbial ecology of the cystic fibrosis lung. Microbiology 153, 917- 923. Hartert, M., Senbaklavacin, O., Gohrbandt, B., Fischer, B. M., Buhl, R. & Vahld, C. F. (2014). Lung transplantation: a treatment option in end-stage lung disease. Dtsch Arztebl Int 111, 107-116. Hassett, D. J., Cuppoletti, J., Trapnell, B., Lymar, S. V., Rowe, J. J., Yoon, S. S., Hilliard, G. M., Parvatiyar, K., Kamani, M. C., Wozniak, D. J., Hwang, S. H., McDermott, T. R. & Ochsner, U. A. (2002). Anaerobic metabolism and quorum sensing by Pseudomonas aeruginosa biofilms in chronically infected cystic fibrosis airways: rethinking antibiotic treatment strategies and drug targets. Adv Drug Deliv Rev 54, 1425-1443. Hauser, A. R. (2009). The type III secretion system of Pseudomonas aeruginosa: infection by injection. Nat Rev Microbiol 7, 654-665. Hauser, A. R., Jain, M., Bar-Meir, M. & McColley, S. A. (2011). Clinical significance of microbial infection and adaptation in cystic fibrosis. Clin Microbiol Rev 24, 29-70. Hayden, H. S., Gillett, W., Saenphimmachak, C., Lim, R., Zhou, Y., Jacobs, M. A., Chang, J., Rohmer, L., D'Argenio, D. A., Palmieri, A., Levy, R., Haugen, E., Wong, G. K., Brittnacher, M. J., Burns, J. L., Miller, S. I., Olson, M. V. & Kaul, R. (2008). Large- insert genome analysis technology detects structural variation in Pseudomonas aeruginosa clinical strains from cystic fibrosis patients. Genomics 91, 530-537. Heck, L. W., Alarcon, P. G., Kulhavy, R. M., Morihara, K., Russell, M. W. & Mestecky, J. F. (1990). Degradation of IgA proteins by Pseudomonas aeruginosa elastase. J Immunol 144, 2253-2257. Henderson, A. G., Ehre, C., Button, B., Abdullah, L. H., Cai, L. H., Leigh, M. W., DeMaria, G. C., Matsui, H., Donaldson, S. H., Davis, C. W., Sheehan, J. K., Boucher, R. C. & Kesimer, M. (2014). Cystic fibrosis airway secretions exhibit mucin hyperconcentration and increased osmotic pressure. J Clin Invest 124, 3047-3060. Hengge, R. (2009). Principles of c-di-GMP signalling in bacteria. Nat Rev Microbiol 7, 263- 273.

188 References

Hentzer, M., Teitzel, G. M., Balzer, G. J., Heydorn, A., Molin, S., Givskov, M. & Parsek, M. R. (2001). Alginate overproduction affects Pseudomonas aeruginosa biofilm structure and function. J Bacteriol 183, 5395-5401. Hentzer, M., Wu, H., Andersen, J. B., Riedel, K., Rasmussen, T. B., Bagge, N., Kumar, N., Schembri, M. A., Song, Z., Kristoffersen, P., Manefield, M., Costerton, J. W., Molin, S., Eberl, L., Steinberg, P., Kjelleberg, S., Hoiby, N. & Givskov, M. (2003). Attenuation of Pseudomonas aeruginosa virulence by quorum sensing inhibitors. EMBO J 22, 3803-3815. Heurlier, K., Williams, F., Heeb, S., Dormond, C., Pessi, G., Singer, D., Camara, M., Williams, P. & Haas, D. (2004). Positive control of swarming, rhamnolipid synthesis, and lipase production by the posttranscriptional RsmA/RsmZ system in Pseudomonas aeruginosa PAO1. J Bacteriol 186, 2936-2945. Hilker, R., Munder, A., Klockgether, J., Losada, P. M., Chouvarine, P., Cramer, N., Davenport, C. F., Dethlefsen, S., Fischer, S., Peng, H., Schonfelder, T., Turk, O., Wiehlmann, L., Wolbeling, F., Gulbins, E., Goesmann, A. & Tummler, B. (2015). Interclonal gradient of virulence in the Pseudomonas aeruginosa pangenome from disease and environment. Environ Microbiol 17, 29-46. Hirakawa, H. & Tomita, H. (2013). Interference of bacterial cell-to-cell communication: a new concept of antimicrobial chemotherapy breaks antibiotic resistance. Front Microbiol 4, 114. Hogardt, M. & Heesemann, J. (2010). Adaptation of Pseudomonas aeruginosa during persistence in the cystic fibrosis lung. Int J Med Microbiol 300, 557-562. Holloway, B. W. (1955). Genetic recombination in Pseudomonas aeruginosa. J Gen Microbiol 13, 572-581. Hollsing, A. E., Granstrom, M. & Strandvik, B. (1987a). Prospective study of serum staphylococcal antibodies in cystic fibrosis. Arch Dis Child 62, 905-911. Hollsing, A. E., Granstrom, M., Vasil, M. L., Wretlind, B. & Strandvik, B. (1987b). Prospective study of serum antibodies to Pseudomonas aeruginosa exoproteins in cystic fibrosis. J Clin Microbiol 25, 1868-1874. Holm, B. A., Keicher, L., Liu, M. Y., Sokolowski, J. & Enhorning, G. (1991). Inhibition of pulmonary surfactant function by phospholipases. J Appl Physiol (1985) 71, 317-321. Hu, H. B., Xu, Y. Q., Chen, F., Zhang, X. H. & Hur, B. K. (2005). Isolation and Characterization of a New Fluorescent Pseudomonas Strain that Produces Both Phenazine 1- Carboxylic Acid and Pyoluteorin. J Microbiol Biotechnol 15, 86-90. Hu, Y. H., Dang, W. & Sun, L. (2012). A TonB-dependent outer membrane receptor of Pseudomonas fluorescens: virulence and vaccine potential. Arch Microbiol 194, 795-802. Huang, H. & Hancock, R. E. (1993). Genetic definition of the substrate selectivity of outer membrane porin protein OprD of Pseudomonas aeruginosa. J Bacteriol 175, 7793-7800. Hunter, R. C., Klepac-Ceraj, V., Lorenzi, M. M., Grotzinger, H., Martin, T. R. & Newman, D. K. (2012). Phenazine content in the cystic fibrosis respiratory tract negatively correlates with lung function and microbial complexity. Am J Respir Cell Mol Biol 47, 738- 745.

189 References

Hunter, R. C., Asfour, F., Dingemans, J., Osuna, B. L., Samad, T., Malfroot, A., Cornelis, P. & Newman, D. K. (2013). Ferrous iron is a significant component of bioavailable iron in cystic fibrosis airways. MBio 4. Iglewski, B. H., Sadoff, J., Bjorn, M. J. & Maxwell, E. S. (1978). Pseudomonas aeruginosa exoenzyme S: an adenosine diphosphate ribosyltransferase distinct from toxin A. Proc Natl Acad Sci U S A 75, 3211-3215. Jelsbak, L., Johansen, H. K., Frost, A. L., Thogersen, R., Thomsen, L. E., Ciofu, O., Yang, L., Haagensen, J. A., Hoiby, N. & Molin, S. (2007). Molecular epidemiology and dynamics of Pseudomonas aeruginosa populations in lungs of cystic fibrosis patients. Infect Immun 75, 2214-2224. Jensen, E. T., Kharazmi, A., Lam, K., Costerton, J. W. & Hoiby, N. (1990). Human polymorphonuclear leukocyte response to Pseudomonas aeruginosa grown in biofilms. Infect Immun 58, 2383-2385. Jensen, R. H. & Woolfolk, C. A. (1985). Formation of Filaments by Pseudomonas putida. Appl Environ Microbiol 50, 364-372. Jimenez, P. N., Koch, G., Thompson, J. A., Xavier, K. B., Cool, R. H. & Quax, W. J. (2012). The multiple signaling systems regulating virulence in Pseudomonas aeruginosa. Microbiol Mol Biol Rev 76, 46-65. Jolley, K. A. & Maiden, M. C. (2010). BIGSdb: Scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics 11, 595. Jones, C. J., Ryder, C. R., Mann, E. E. & Wozniak, D. J. (2013). AmrZ modulates Pseudomonas aeruginosa biofilm architecture by directly repressing transcription of the psl operon. J Bacteriol 195, 1637-1644. Joshi, A., Grinter, R., Josts, I., Chen, S., Wojdyla, J. A., Lowe, E. D., Kaminska, R., Sharp, C., McCaughey, L., Roszak, A. W., Cogdell, R. J., Byron, O., Walker, D. & Kleanthous, C. (2015). Structures of the Ultra-High-Affinity Protein-Protein Complexes of Pyocins S2 and AP41 and Their Cognate Immunity Proteins from Pseudomonas aeruginosa. J Mol Biol 427, 2852-2866. Juhas, M., van der Meer, J. R., Gaillard, M., Harding, R. M., Hood, D. W. & Crook, D. W. (2009). Genomic islands: tools of bacterial horizontal gene transfer and evolution. FEMS Microbiol Rev 33, 376-393. Kapitein, N. & Mogk, A. (2013). Deadly syringes: type VI secretion system activities in pathogenicity and interbacterial competition. Curr Opin Microbiol 16, 52-58. Kawai, T. & Akira, S. (2007). Signaling to NF-kappaB by Toll-like receptors. Trends Mol Med 13, 460-469. Kearns, D. B. (2010). A field guide to bacterial swarming motility. Nat Rev Microbiol 8, 634- 644. Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N. & Sternberg, M. J. (2015). The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc 10, 845-858.

190 References

Kent, A. D. & Triplett, E. W. (2002). Microbial communities and their interactions in soil and rhizosphere ecosystems. Annu Rev Microbiol 56, 211-236. Kerem, B., Rommens, J. M., Buchanan, J. A., Markiewicz, D., Cox, T. K., Chakravarti, A., Buchwald, M. & Tsui, L. C. (1989). Identification of the cystic fibrosis gene: genetic analysis. Science 245, 1073-1080. Kerr, K. G. & Snelling, A. M. (2009). Pseudomonas aeruginosa: a formidable and ever- present adversary. J Hosp Infect 73, 338-344. Kessler, E., Safrin, M., Olson, J. C. & Ohman, D. E. (1993). Secreted LasA of Pseudomonas aeruginosa is a staphylolytic protease. J Biol Chem 268, 7503-7508. Kessler, E., Safrin, M., Gustin, J. K. & Ohman, D. E. (1998). Elastase and the LasA protease of Pseudomonas aeruginosa are secreted with their propeptides. J Biol Chem 273, 30225-30231. Khan, T. Z., Wagener, J. S., Bost, T., Martinez, J., Accurso, F. J. & Riches, D. W. (1995). Early pulmonary inflammation in infants with cystic fibrosis. Am J Respir Crit Care Med 151, 1075-1082. Kim, S. J. & Skach, W. R. (2012). Mechanisms of CFTR Folding at the Endoplasmic Reticulum. Front Pharmacol 3, 201. Kleanthous, C. & Walker, D. (2001). Immunity proteins: enzyme inhibitors that avoid the active site. Trends Biochem Sci 26, 624-631. Klockgether, J., Cramer, N., Wiehlmann, L., Davenport, C. F. & Tummler, B. (2011). Pseudomonas aeruginosa Genomic Structure and Diversity. Front Microbiol 2, 150. Knowles, M. R. & Boucher, R. C. (2002). Mucus clearance as a primary innate defense mechanism for mammalian airways. J Clin Invest 109, 571-577. Kohler, T., Curty, L. K., Barja, F., van Delden, C. & Pechere, J. C. (2000). Swarming of Pseudomonas aeruginosa is dependent on cell-to-cell signaling and requires flagella and pili. J Bacteriol 182, 5990-5996. Kolpen, M., Kuhl, M., Bjarnsholt, T., Moser, C., Hansen, C. R., Liengaard, L., Kharazmi, A., Pressler, T., Hoiby, N. & Jensen, P. O. (2014). Nitrous oxide production in sputum from cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection. PLoS One 9, e84353. Konings, A. F., Martin, L. W., Sharples, K. J., Roddam, L. F., Latham, R., Reid, D. W. & Lamont, I. L. (2013). Pseudomonas aeruginosa uses multiple pathways to acquire iron during chronic infection in cystic fibrosis lungs. Infect Immun 81, 2697-2704.

Kounnas, M. Z., Morris, R. E., Thompson, M. R., FitzGerald, D. J., Strickland, D. K. & Saelinger, C. B. (1992). The alpha 2-macroglobulin receptor/low density lipoprotein receptor-related protein binds and internalizes Pseudomonas exotoxin A. J Biol Chem 267, 12420-12423.

191 References

Kulasakara, H., Lee, V., Brencic, A., Liberati, N., Urbach, J., Miyata, S., Lee, D. G., Neely, A. N., Hyodo, M., Hayakawa, Y., Ausubel, F. M. & Lory, S. (2006). Analysis of Pseudomonas aeruginosa diguanylate cyclases and phosphodiesterases reveals a role for bis- (3'-5')-cyclic-GMP in virulence. Proc Natl Acad Sci U S A 103, 2839-2844. Lamont, I. L., Konings, A. F. & Reid, D. W. (2009). Iron acquisition by Pseudomonas aeruginosa in the lungs of patients with cystic fibrosis. BioMetals 22, 53-60. Langille, M. G. & Brinkman, F. S. (2009). IslandViewer: an integrated interface for computational identification and visualization of genomic islands. Bioinformatics 25, 664- 665. Lau, G. W., Hassett, D. J., Ran, H. & Kong, F. (2004). The role of pyocyanin in Pseudomonas aeruginosa infection. Trends Mol Med 10, 599-606. Laub, M. T. & Goulian, M. (2007). Specificity in two-component signal transduction pathways. Annu Rev Genet 41, 121-145. Lee, D. G., Urbach, J. M., Wu, G., Liberati, N. T., Feinbaum, R. L., Miyata, S., Diggins, L. T., He, J., Saucier, M., Deziel, E., Friedman, L., Li, L., Grills, G., Montgomery, K., Kucherlapati, R., Rahme, L. G. & Ausubel, F. M. (2006). Genomic analysis reveals that Pseudomonas aeruginosa virulence is combinatorial. Genome Biol 7, R90. Lee, J., Wu, J., Deng, Y., Wang, J., Wang, C., Wang, J., Chang, C., Dong, Y., Williams, P. & Zhang, L. H. (2013). A cell-cell communication signal integrates quorum sensing and stress response. Nat Chem Biol 9, 339-343. Lee, J. & Zhang, L. (2015). The hierarchy quorum sensing network in Pseudomonas aeruginosa. Protein Cell 6, 26-41. Lee, T. W. & Southern, K. W. (2012). Topical cystic fibrosis transmembrane conductance regulator gene replacement for cystic fibrosis-related lung disease. Cochrane Database Syst Rev 10, CD005599. Lee, V. T., Matewish, J. M., Kessler, J. L., Hyodo, M., Hayakawa, Y. & Lory, S. (2007). A cyclic-di-GMP receptor required for bacterial exopolysaccharide production. Mol Microbiol 65, 1474-1484. Lenney, W. & Gilchrist, F. J. (2011). Pseudomonas aeruginosa and cyanide production. Eur Respir J 37, 482-483. Leo, J. C., Grin, I. & Linke, D. (2012). Type V secretion: mechanism(s) of autotransport through the bacterial outer membrane. Philos Trans R Soc Lond B Biol Sci 367, 1088-1101. Lesic, B., Starkey, M., He, J., Hazan, R. & Rahme, L. G. (2009). Quorum sensing differentially regulates Pseudomonas aeruginosa type VI secretion locus I and homologous loci II and III, which are required for pathogenesis. Microbiology 155, 2845-2855. Li, H. & Durbin, R. (2010). Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26, 589-595. Li, L., Huang, D., Cheung, M. K., Nong, W., Huang, Q. & Kwan, H. S. (2013). BSRD: a repository for bacterial small regulatory RNA. Nucleic Acids Res 41, D233-238.

192 References

Li, L. & Somerset, S. (2014). Digestive system dysfunction in cystic fibrosis: challenges for nutrition therapy. Dig Liver Dis 46, 865-874. Line, L., Alhede, M., Kolpen, M., Kuhl, M., Ciofu, O., Bjarnsholt, T., Moser, C., Toyofuku, M., Nomura, N., Hoiby, N. & Jensen, P. O. (2014). Physiological levels of nitrate support anoxic growth by denitrification of Pseudomonas aeruginosa at growth rates reported in cystic fibrosis lungs and sputum. Front Microbiol 5, 554. Ling, H., Saeidi, N., Rasouliha, B. H. & Chang, M. W. (2010). A predicted S-type pyocin shows a bactericidal activity against clinical Pseudomonas aeruginosa isolates through membrane damage. FEBS Lett 584, 3354-3358. Lister, P. D., Wolter, D. J. & Hanson, N. D. (2009). Antibacterial-resistant Pseudomonas aeruginosa: clinical impact and complex regulation of chromosomally encoded resistance mechanisms. Clin Microbiol Rev 22, 582-610. Livak, K. J. & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402-408. Llamas, M. A., Imperi, F., Visca, P. & Lamont, I. L. (2014). Cell-surface signaling in Pseudomonas: stress responses, iron transport, and pathogenicity. FEMS Microbiol Rev 38, 569-597. Lorenz, R., Bernhart, S. H., Honer Zu Siederdissen, C., Tafer, H., Flamm, C., Stadler, P. F. & Hofacker, I. L. (2011). ViennaRNA Package 2.0. Algorithms Mol Biol 6, 26.

Lucangelo, U., Fontanesi, L., Antonaglia, V., Pellis, T., Berlot, G., Liguori, G., Bird, F. M. & Gullo, A. (2003). High frequency percussive ventilation (HFPV). Principles and technique. Minerva Anestesiol 69, 841-848, 848-851. Lucchetti-Miganeh, C., Redelberger, D., Chambonnier, G., Rechenmann, F., Elsen, S., Bordi, C., Jeannot, K., Attree, I., Plesiat, P. & de Bentzmann, S. (2014). Pseudomonas aeruginosa Genome Evolution in Patients and under the Hospital Environment. Pathogens 3, 309-340. Lyczak, J. B., Cannon, C. L. & Pier, G. B. (2002). Lung infections associated with cystic fibrosis. Clin Microbiol Rev 15, 194-222. Lynch, S. V., Brodie, E. L. & Matin, A. (2004). Role and regulation of sigma S in general resistance conferred by low-shear simulated microgravity in Escherichia coli. J Bacteriol 186, 8207-8212. Ma, L., Jackson, K. D., Landry, R. M., Parsek, M. R. & Wozniak, D. J. (2006). Analysis of Pseudomonas aeruginosa conditional psl variants reveals roles for the psl polysaccharide in adhesion and maintaining biofilm structure postattachment. J Bacteriol 188, 8213-8221. Ma, L., Conover, M., Lu, H., Parsek, M. R., Bayles, K. & Wozniak, D. J. (2009). Assembly and development of the Pseudomonas aeruginosa biofilm matrix. PLoS Pathog 5, e1000354. Maatallah, M., Bakhrouf, A., Habeeb, M. A., Turlej-Rogacka, A., Iversen, A., Pourcel, C., Sioud, O. & Giske, C. G. (2013). Four genotyping schemes for phylogenetic analysis of Pseudomonas aeruginosa: comparison of their congruence with multi-locus sequence typing. PLoS One 8, e82069.

193 References

MacKenzie, T., Gifford, A. H., Sabadosa, K. A., Quinton, H. B., Knapp, E. A., Goss, C. H. & Marshall, B. C. (2014). Longevity of patients with cystic fibrosis in 2000 to 2010 and beyond: survival analysis of the Cystic Fibrosis Foundation patient registry. Ann Intern Med 161, 233-241. Mann, E. E. & Wozniak, D. J. (2012). Pseudomonas biofilm matrix composition and niche biology. FEMS Microbiol Rev 36, 893-916. Manos, J., Hu, H., Rose, B. R., Wainwright, C. E., Zablotska, I. B., Cheney, J., Turnbull, L., Whitchurch, C. B., Grimwood, K., Harmer, C., Anuj, S. N., Harbour, C. & group, A. s. (2013). Virulence factor expression patterns in Pseudomonas aeruginosa strains from infants with cystic fibrosis. Eur J Clin Microbiol Infect Dis 32, 1583-1592. Mariencheck, W. I., Alcorn, J. F., Palmer, S. M. & Wright, J. R. (2003). Pseudomonas aeruginosa elastase degrades surfactant proteins A and D. Am J Respir Cell Mol Biol 28, 528- 537. Marvig, R. L., Johansen, H. K., Molin, S. & Jelsbak, L. (2013). Genome analysis of a transmissible lineage of Pseudomonas aeruginosa reveals pathoadaptive mutations and distinct evolutionary paths of hypermutators. PLoS Genet 9, e1003741. Marvig, R. L., Sommer, L. M., Molin, S. & Johansen, H. K. (2015). Convergent evolution and adaptation of Pseudomonas aeruginosa within patients with cystic fibrosis. Nat Genet 47, 57-64. Maselli, J. H., Sontag, M. K., Norris, J. M., MacKenzie, T., Wagener, J. S. & Accurso, F. J. (2003). Risk factors for initial acquisition of Pseudomonas aeruginosa in children with cystic fibrosis identified by newborn screening. Pediatr Pulmonol 35, 257-262. Masoud-Landgraf, L., Badura, A., Eber, E., Feierl, G., Posch, J., Zarfel, G., Zach, M. & Marth, E. (2012). Molecular epidemiology of Pseudomonas aeruginosa in cystic fibrosis patients from Southeast Austria. Wien Klin Wochenschr 124, 262-265. Mathee, K., Narasimhan, G., Valdes, C., Qiu, X., Matewish, J. M., Koehrsen, M., Rokas, A., Yandava, C. N., Engels, R., Zeng, E., Olavarietta, R., Doud, M., Smith, R. S., Montgomery, P., White, J. R., Godfrey, P. A., Kodira, C., Birren, B., Galagan, J. E. & Lory, S. (2008). Dynamics of Pseudomonas aeruginosa genome evolution. Proc Natl Acad Sci U S A 105, 3100-3105. Matsui, H., Sano, Y., Ishihara, H. & Shinomiya, T. (1993). Regulation of pyocin genes in Pseudomonas aeruginosa by positive (prtN) and negative (prtR) regulatory genes. J Bacteriol 175, 1257-1263. Mavrodi, D. V., Bonsall, R. F., Delaney, S. M., Soule, M. J., Phillips, G. & Thomashow, L. S. (2001). Functional analysis of genes for biosynthesis of pyocyanin and phenazine-1- carboxamide from Pseudomonas aeruginosa PAO1. J Bacteriol 183, 6454-6465. May, T. B., Shinabarger, D., Maharaj, R., Kato, J., Chu, L., DeVault, J. D., Roychoudhury, S., Zielinski, N. A., Berry, A., Rothmel, R. K. & et al. (1991). Alginate synthesis by Pseudomonas aeruginosa: a key pathogenic factor in chronic pulmonary infections of cystic fibrosis patients. Clin Microbiol Rev 4, 191-206.

194 References

McClean, K. H., Winson, M. K., Fish, L., Taylor, A., Chhabra, S. R., Camara, M., Daykin, M., Lamb, J. H., Swift, S., Bycroft, B. W., Stewart, G. S. & Williams, P. (1997). Quorum sensing and Chromobacterium violaceum: exploitation of violacein production and inhibition for the detection of N-acylhomoserine lactones. Microbiology 143 ( Pt 12), 3703- 3711. McElroy, K. E., Hui, J. G., Woo, J. K., Luk, A. W., Webb, J. S., Kjelleberg, S., Rice, S. A. & Thomas, T. (2014). Strain-specific parallel evolution drives short-term diversification during Pseudomonas aeruginosa biofilm formation. Proc Natl Acad Sci U S A 111, E1419- 1427. McGuigan, L. & Callaghan, M. (2015). The evolving dynamics of the microbial community in the cystic fibrosis lung. Environ Microbiol 17, 16-28. Merighi, M., Lee, V. T., Hyodo, M., Hayakawa, Y. & Lory, S. (2007). The second messenger bis-(3'-5')-cyclic-GMP and its PilZ domain-containing receptor Alg44 are required for alginate biosynthesis in Pseudomonas aeruginosa. Mol Microbiol 65, 876-895. Meyer, J. M., Stintzi, A., De Vos, D., Cornelis, P., Tappe, R., Taraz, K. & Budzikiewicz, H. (1997). Use of siderophores to type pseudomonads: the three Pseudomonas aeruginosa pyoverdine systems. Microbiology 143 ( Pt 1), 35-43. Meyer, J. M. (2000). Pyoverdines: pigments, siderophores and potential taxonomic markers of fluorescent Pseudomonas species. Arch Microbiol 174, 135-142. Michel, L., Bachelard, A. & Reimmann, C. (2007). Ferripyochelin uptake genes are involved in pyochelin-mediated signalling in Pseudomonas aeruginosa. Microbiology 153, 1508-1518. Michel-Briand, Y. & Baysse, C. (2002). The pyocins of Pseudomonas aeruginosa. Biochimie 84, 499-510. Mislin, G. L., Hoegy, F., Cobessi, D., Poole, K., Rognan, D. & Schalk, I. J. (2006). Binding properties of pyochelin and structurally related molecules to FptA of Pseudomonas aeruginosa. J Mol Biol 357, 1437-1448. Moller, T., Franch, T., Hojrup, P., Keene, D. R., Bachinger, H. P., Brennan, R. G. & Valentin-Hansen, P. (2002). Hfq: a bacterial Sm-like protein that mediates RNA-RNA interaction. Mol Cell 9, 23-30. Moreau-Marquis, S., Bomberger, J. M., Anderson, G. G., Swiatecka-Urban, A., Ye, S., O'Toole, G. A. & Stanton, B. A. (2008). The DeltaF508-CFTR mutation results in increased biofilm formation by Pseudomonas aeruginosa by increasing iron availability. Am J Physiol Lung Cell Mol Physiol 295, L25-37. Moskowitz, S. M., Brannon, M. K., Dasgupta, N., Pier, M., Sgambati, N., Miller, A. K., Selgrade, S. E., Miller, S. I., Denton, M., Conway, S. P., Johansen, H. K. & Hoiby, N. (2012). PmrB mutations promote polymyxin resistance of Pseudomonas aeruginosa isolated from colistin-treated cystic fibrosis patients. Antimicrob Agents Chemother 56, 1019-1030. Moss, J., Ehrmantraut, M. E., Banwart, B. D., Frank, D. W. & Barbieri, J. T. (2001). Sera from adult patients with cystic fibrosis contain antibodies to Pseudomonas aeruginosa type III apparatus. Infect Immun 69, 1185-1188.

195 References

Mossialos, D., Ochsner, U., Baysse, C., Chablain, P., Pirnay, J. P., Koedam, N., Budzikiewicz, H., Fernandez, D. U., Schafer, M., Ravel, J. & Cornelis, P. (2002). Identification of new, conserved, non-ribosomal peptide synthetases from fluorescent pseudomonads involved in the biosynthesis of the siderophore pyoverdine. Mol Microbiol 45, 1673-1685. Nau, C. D. & Konisky, J. (1989). Evolutionary relationship between the TonB-dependent outer membrane transport proteins: nucleotide and amino acid sequences of the Escherichia coli colicin I receptor gene. J Bacteriol 171, 1041-1047. Nauman, E. A., Ott, C. M., Sander, E., Tucker, D. L., Pierson, D., Wilson, J. W. & Nickerson, C. A. (2007). Novel quantitative biosystem for modeling physiological fluid shear stress on cells. Appl Environ Microbiol 73, 699-705. Nivens, D. E., Ohman, D. E., Williams, J. & Franklin, M. J. (2001). Role of alginate and its O acetylation in formation of Pseudomonas aeruginosa microcolonies and biofilms. J Bacteriol 183, 1047-1057. O'Sullivan, B. P. & Freedman, S. D. (2009). Cystic fibrosis. Lancet 373, 1891-1904. Ode, K. L. & Moran, A. (2013). New insights into cystic fibrosis-related diabetes in children. Lancet Diabetes Endocrinol 1, 52-58. Okkotsu, Y., Little, A. S. & Schurr, M. J. (2014). The Pseudomonas aeruginosa AlgZR two-component system coordinates multiple phenotypes. Front Cell Infect Microbiol 4, 82. Okiyoneda, T. & Lukacs, G. L. (2012). Fixing cystic fibrosis by correcting CFTR domain assembly. J Cell Biol 199, 199-204. Osman, M. A. (1965). Pyocine Typing of Pseudomonas aeruginosa. J Clin Pathol 18, 200- 202. Pacello, F., Rotilio, G. & Battistoni, A. (2012). Low-Shear Modeled Microgravity Enhances Salmonella enterica Resistance to Hydrogen Peroxide Through a Mechanism Involving KatG and KatN. Open Microbiol J 6, 53-64. Paletta, J. L. & Ohman, D. E. (2012). Evidence for two promoters internal to the alginate biosynthesis operon in Pseudomonas aeruginosa. Curr Microbiol 65, 770-775. Palmer, K. L., Mashburn, L. M., Singh, P. K. & Whiteley, M. (2005). Cystic fibrosis sputum supports growth and cues key aspects of Pseudomonas aeruginosa physiology. J Bacteriol 187, 5267-5277. Palmer, T. & Berks, B. C. (2012). The twin-arginine translocation (Tat) protein export pathway. Nat Rev Microbiol 10, 483-496. Park, A. J., Surette, M. D. & Khursigara, C. M. (2014). Antimicrobial targets localize to the extracellular vesicle-associated proteome of Pseudomonas aeruginosa grown in a biofilm. Front Microbiol 5, 464. Parkins, M. D. & Floto, R. A. (2015). Emerging bacterial pathogens and changing concepts of bacterial pathogenesis in cystic fibrosis. J Cyst Fibros 14, 293-304.

196 References

Partridge, J. D. & Harshey, R. M. (2013). Swarming: flexible roaming plans. J Bacteriol 195, 909-918. Pedersen, S. S., Espersen, F., Hoiby, N. & Jensen, T. (1990). Immunoglobulin A and immunoglobulin G antibody responses to alginates from Pseudomonas aeruginosa in patients with cystic fibrosis. J Clin Microbiol 28, 747-755. Pedersen, S. S. (1992). Lung infection with alginate-producing, mucoid Pseudomonas aeruginosa in cystic fibrosis. APMIS Suppl 28, 1-79. Pessi, G. & Haas, D. (2000). Transcriptional control of the hydrogen cyanide biosynthetic genes hcnABC by the anaerobic regulator ANR and the quorum-sensing regulators LasR and RhlR in Pseudomonas aeruginosa. J Bacteriol 182, 6940-6949. Pessi, G. & Haas, D. (2001). Dual control of hydrogen cyanide biosynthesis by the global activator GacA in Pseudomonas aeruginosa PAO1. FEMS Microbiol Lett 200, 73-78. Pezzulo, A. A., Tang, X. X., Hoegger, M. J., Alaiwa, M. H., Ramachandran, S., Moninger, T. O., Karp, P. H., Wohlford-Lenane, C. L., Haagsman, H. P., van Eijk, M., Banfi, B., Horswill, A. R., Stoltz, D. A., McCray, P. B., Jr., Welsh, M. J. & Zabner, J. (2012). Reduced airway surface pH impairs bacterial killing in the porcine cystic fibrosis lung. Nature 487, 109-113. Pier, G. B., Coleman, F., Grout, M., Franklin, M. & Ohman, D. E. (2001). Role of alginate O acetylation in resistance of mucoid Pseudomonas aeruginosa to opsonic phagocytosis. Infect Immun 69, 1895-1901.

Pirnay, J. P., Matthijs, S., Colak, H., Chablain, P., Bilocq, F., Van Eldere, J., De Vos, D., Zizi, M., Triest, L. & Cornelis, P. (2005). Global Pseudomonas aeruginosa biodiversity as reflected in a Belgian river. Environ Microbiol 7, 969-980. Pirnay, J. P., Bilocq, F., Pot, B., Cornelis, P., Zizi, M., Van Eldere, J., Deschaght, P., Vaneechoutte, M., Jennes, S., Pitt, T. & De Vos, D. (2009). Pseudomonas aeruginosa population structure revisited. PLoS One 4, e7740. Pistorius, D., Ullrich, A., Lucas, S., Hartmann, R. W., Kazmaier, U. & Muller, R. (2011). Biosynthesis of 2-Alkyl-4(1H)-quinolones in Pseudomonas aeruginosa: potential for therapeutic interference with pathogenicity. ChemBioChem 12, 850-853. Pollack, M. (1983). The role of exotoxin A in Pseudomonas disease and immunity. Rev Infect Dis 5 Suppl 5, S979-984. Price-Whelan, A., Dietrich, L. E. & Newman, D. K. (2006). Rethinking 'secondary' metabolism: physiological roles for phenazine antibiotics. Nat Chem Biol 2, 71-78. Quinn, R. A., Lim, Y. W., Maughan, H., Conrad, D., Rohwer, F. & Whiteson, K. L. (2014). Biogeochemical forces shape the composition and physiology of polymicrobial communities in the cystic fibrosis lung. MBio 5, e00956-00913. Rabello, L. S., Silva, J. R., Azevedo, L. C., Souza, I., Torres, V. B., Rosolem, M. M., Lisboa, T., Soares, M. & Salluh, J. I. (2015). Clinical outcomes and microbiological characteristics of severe pneumonia in cancer patients: a prospective cohort study. PLoS One 10, e0120544.

197 References

Rashid, M. H. & Kornberg, A. (2000). Inorganic polyphosphate is needed for swimming, swarming, and twitching motilities of Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 97, 4885-4890. Ratkai, C., Peixe, L. V., Grosso, F., Freitas, A. R., Antunes, P., Fodor, E., Hajdu, E. & Nagy, E. (2010). Successful application of the DiversiLab repetitive-sequence-based PCR typing system for confirmation of the circulation of a multiresistant Pseudomonas aeruginosa clone in different hospital wards. Diagn Microbiol Infect Dis 67, 202-206. Rau, M. H., Hansen, S. K., Johansen, H. K., Thomsen, L. E., Workman, C. T., Nielsen, K. F., Jelsbak, L., Hoiby, N., Yang, L. & Molin, S. (2010). Early adaptive developments of Pseudomonas aeruginosa after the transition from life in the environment to persistent colonization in the airways of human cystic fibrosis hosts. Environ Microbiol 12, 1643-1658. Rau, M. H., Marvig, R. L., Ehrlich, G. D., Molin, S. & Jelsbak, L. (2012). Deletion and acquisition of genomic content during early stage adaptation of Pseudomonas aeruginosa to a human host environment. Environ Microbiol 14, 2200-2211. Ravel, J. & Cornelis, P. (2003). Genomics of pyoverdine-mediated iron uptake in pseudomonads. Trends Microbiol 11, 195-200. Recinos, D. A., Sekedat, M. D., Hernandez, A., Cohen, T. S., Sakhtah, H., Prince, A. S., Price-Whelan, A. & Dietrich, L. E. (2012). Redundant phenazine operons in Pseudomonas aeruginosa exhibit environment-dependent expression and differential roles in pathogenicity. Proc Natl Acad Sci U S A 109, 19420-19425. Reen, F. J., Haynes, J. M., Mooij, M. J. & O'Gara, F. (2013). A non-classical LysR-type transcriptional regulator PA2206 is required for an effective oxidative stress response in Pseudomonas aeruginosa. PLoS One 8, e54479. Reid, D. W., Anderson, G. J. & Lamont, I. L. (2009). Role of lung iron in determining the bacterial and host struggle in cystic fibrosis. Am J Physiol Lung Cell Mol Physiol 297, L795- 802. Renwick, J., McNally, P., John, B., DeSantis, T., Linnane, B., Murphy, P. & Shield, C. F. (2014). The microbial community of the cystic fibrosis airway is disrupted in early life. PLoS One 9, e109798. Riley, M. A. & Gordon, D. M. (1999). The ecological role of bacteriocins in bacterial competition. Trends Microbiol 7, 129-133. Riley, M. A. & Wertz, J. E. (2002a). Bacteriocins: evolution, ecology, and application. Annu Rev Microbiol 56, 117-137. Riley, M. A. & Wertz, J. E. (2002b). Bacteriocin diversity: ecological and evolutionary perspectives. Biochimie 84, 357-364. Riordan, J. R., Rommens, J. M., Kerem, B., Alon, N., Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J. L. & et al. (1989). Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245, 1066-1073. Riordan, J. R. (2008). CFTR function and prospects for therapy. Annu Rev Biochem 77, 701- 726.

198 References

Robinson, M. D., McCarthy, D. J. & Smyth, G. K. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139- 140. Rojas Murcia, N., Lee, X., Waridel, P., Maspoli, A., Imker, H. J., Chai, T., Walsh, C. T. & Reimmann, C. (2015). The Pseudomonas aeruginosa antimetabolite L -2-amino-4- methoxy-trans-3-butenoic acid (AMB) is made from glutamate and two alanine residues via a thiotemplate-linked tripeptide precursor. Front Microbiol 6, 170. Rommens, J. M., Iannuzzi, M. C., Kerem, B., Drumm, M. L., Melmer, G., Dean, M., Rozmahel, R., Cole, J. L., Kennedy, D., Hidaka, N. & et al. (1989). Identification of the cystic fibrosis gene: chromosome walking and jumping. Science 245, 1059-1065. Rosenberg, A. L., Seneff, M. G., Atiyeh, L., Wagner, R., Bojanowski, L. & Zimmerman, J. E. (2001). The importance of bacterial sepsis in intensive care unit patients with acquired immunodeficiency syndrome: implications for future care in the age of increasing antiretroviral resistance. Crit Care Med 29, 548-556. Roy, P. H., Tetu, S. G., Larouche, A., Elbourne, L., Tremblay, S., Ren, Q., Dodson, R., Harkins, D., Shay, R., Watkins, K., Mahamoud, Y. & Paulsen, I. T. (2010). Complete genome sequence of the multiresistant taxonomic outlier Pseudomonas aeruginosa PA7. PLoS One 5, e8842. Rozen, S. & Skaletsky, H. (2000). Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 132, 365-386. Salunkhe, P., Smart, C. H., Morgan, J. A., Panagea, S., Walshaw, M. J., Hart, C. A., Geffers, R., Tummler, B. & Winstanley, C. (2005). A cystic fibrosis epidemic strain of Pseudomonas aeruginosa displays enhanced virulence and antimicrobial resistance. J Bacteriol 187, 4908-4920. Sano, Y. & Kageyama, M. (1993). A novel transposon-like structure carries the genes for pyocin AP41, a Pseudomonas aeruginosa bacteriocin with a DNase domain homology to E2 group colicins. Mol Gen Genet 237, 161-170. Sano, Y., Matsui, H., Kobayashi, M. & Kageyama, M. (1993). Molecular structures and functions of pyocins S1 and S2 in Pseudomonas aeruginosa. J Bacteriol 175, 2907-2916. Sawicki, G. S., Rasouliyan, L., McMullen, A. H., Wagener, J. S., McColley, S. A., Pasta, D. J. & Quittner, A. L. (2011). Longitudinal assessment of health-related quality of life in an observational cohort of patients with cystic fibrosis. Pediatr Pulmonol 46, 36-44. Schobert, M. & Jahn, D. (2010). Anaerobic physiology of Pseudomonas aeruginosa in the cystic fibrosis lung. Int J Med Microbiol 300, 549-556. Schoni, M. H. (1989). Autogenic drainage: a modern approach to physiotherapy in cystic fibrosis. J R Soc Med 82 Suppl 16, 32-37. Schuster, M. & Greenberg, E. P. (2006). A network of networks: quorum-sensing gene regulation in Pseudomonas aeruginosa. Int J Med Microbiol 296, 73-81. Scott, F. W. & Pitt, T. L. (2004). Identification and characterization of transmissible Pseudomonas aeruginosa strains in cystic fibrosis patients in England and Wales. J Med Microbiol 53, 609-615.

199 References

Segata, N., Ballarini, A. & Jousson, O. (2013). Genome Sequence of Pseudomonas aeruginosa PA45, a Highly Virulent Strain Isolated from a Patient with Bloodstream Infection. Genome Announc 1. Selezska, K., Kazmierczak, M., Musken, M., Garbe, J., Schobert, M., Haussler, S., Wiehlmann, L., Rohde, C. & Sikorski, J. (2012). Pseudomonas aeruginosa population structure revisited under environmental focus: impact of water quality and phage pressure. Environ Microbiol 14, 1952-1967. Seo, Y. & Galloway, D. R. (1990). Purification of the pyocin S2 complex from Pseudomonas aeruginosa PAO1: analysis of DNase activity. Biochem Biophys Res Commun 172, 455-461. Silby, M. W., Winstanley, C., Godfrey, S. A., Levy, S. B. & Jackson, R. W. (2011). Pseudomonas genomes: diverse and adaptable. FEMS Microbiol Rev 35, 652-680. Silverman, J. M., Brunet, Y. R., Cascales, E. & Mougous, J. D. (2012). Structure and regulation of the type VI secretion system. Annu Rev Microbiol 66, 453-472. Smith, D., Spanel, P., Gilchrist, F. J. & Lenney, W. (2013). Hydrogen cyanide, a volatile biomarker of Pseudomonas aeruginosa infection. J Breath Res 7, 044001. Smith, E. E., Sims, E. H., Spencer, D. H., Kaul, R. & Olson, M. V. (2005). Evidence for diversifying selection at the pyoverdine locus of Pseudomonas aeruginosa. J Bacteriol 187, 2138-2147. Smith, E. E., Buckley, D. G., Wu, Z., Saenphimmachak, C., Hoffman, L. R., D'Argenio, D. A., Miller, S. I., Ramsey, B. W., Speert, D. P., Moskowitz, S. M., Burns, J. L., Kaul, R. & Olson, M. V. (2006). Genetic adaptation by Pseudomonas aeruginosa to the airways of cystic fibrosis patients. Proc Natl Acad Sci U S A 103, 8487-8492. Soelaiman, S., Jakes, K., Wu, N., Li, C. & Shoham, M. (2001). Crystal structure of colicin E3: implications for cell entry and ribosome inactivation. Mol Cell 8, 1053-1062. Son, M. S., Matthews, W. J., Jr., Kang, Y., Nguyen, D. T. & Hoang, T. T. (2007). In vivo evidence of Pseudomonas aeruginosa nutrient acquisition and pathogenesis in the lungs of cystic fibrosis patients. Infect Immun 75, 5313-5324. Soni, A., O'Sullivan, L., Quick, L. N., Ott, C. M., Nickerson, C. A. & Wilson, J. W. (2014). Conservation of the Low-shear Modeled Microgravity Response in Enterobacteriaceae and Analysis of the trp Genes in this Response. Open Microbiol J 8, 51- 58. Sonnleitner, E., Sorger-Domenigg, T., Madej, M. J., Findeiss, S., Hackermuller, J., Huttenhofer, A., Stadler, P. F., Blasi, U. & Moll, I. (2008). Detection of small RNAs in Pseudomonas aeruginosa by RNomics and structure-based bioinformatic tools. Microbiology 154, 3175-3187. Sonnleitner, E., Romeo, A. & Blasi, U. (2012). Small regulatory RNAs in Pseudomonas aeruginosa. RNA Biol 9, 364-371. Stewart, L., Ford, A., Sangal, V., Jeukens, J., Boyle, B., Kukavica-Ibrulj, I., Caim, S., Crossman, L., Hoskisson, P. A., Levesque, R. & Tucker, N. P. (2014). Draft genomes of 12 host-adapted and environmental isolates of Pseudomonas aeruginosa and their positions in the core genome phylogeny. Pathog Dis 71, 20-25.

200 References

Stoltz, D. A., Meyerholz, D. K. & Welsh, M. J. (2015). Origins of cystic fibrosis lung disease. N Engl J Med 372, 351-362. Storz, G., Vogel, J. & Wassarman, K. M. (2011). Regulation by small RNAs in bacteria: expanding frontiers. Mol Cell 43, 880-891. Stover, C. K., Pham, X. Q., Erwin, A. L., Mizoguchi, S. D., Warrener, P., Hickey, M. J., Brinkman, F. S., Hufnagle, W. O., Kowalik, D. J., Lagrou, M., Garber, R. L., Goltry, L., Tolentino, E., Westbrock-Wadman, S., Yuan, Y., Brody, L. L., Coulter, S. N., Folger, K. R., Kas, A., Larbig, K., Lim, R., Smith, K., Spencer, D., Wong, G. K., Wu, Z., Paulsen, I. T., Reizer, J., Saier, M. H., Hancock, R. E., Lory, S. & Olson, M. V. (2000). Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406, 959-964. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. (1990). Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol 185, 60-89. Taylor, C. J., Hardcastle, J. & Southern, K. W. (2009). Physiological measurements confirming the diagnosis of cystic fibrosis: the sweat test and measurements of transepithelial potential difference. Paediatr Respir Rev 10, 220-226. Taylor, P. K., Yeung, A. T. & Hancock, R. E. (2014). Antibiotic resistance in Pseudomonas aeruginosa biofilms: towards the development of novel anti-biofilm therapies. J Biotechnol 191, 121-130. Thibodeau, P. H. & Butterworth, M. B. (2013). Proteases, cystic fibrosis and the epithelial sodium channel (ENaC). Cell Tissue Res 351, 309-323. Thomason, M. K. & Storz, G. (2010). Bacterial antisense RNAs: how many are there, and what are they doing? Annu Rev Genet 44, 167-188. Tirouvanziam, R., de Bentzmann, S., Hubeau, C., Hinnrasky, J., Jacquot, J., Peault, B. & Puchelle, E. (2000). Inflammation and infection in naive human cystic fibrosis airway grafts. Am J Respir Cell Mol Biol 23, 121-127. Toutain, C. M., Zegans, M. E. & O'Toole, G. A. (2005). Evidence for two flagellar stators and their role in the motility of Pseudomonas aeruginosa. J Bacteriol 187, 771-777. Trunk, K., Benkert, B., Quack, N., Munch, R., Scheer, M., Garbe, J., Jansch, L., Trost, M., Wehland, J., Buer, J., Jahn, M., Schobert, M. & Jahn, D. (2010). Anaerobic adaptation in Pseudomonas aeruginosa: definition of the Anr and Dnr regulons. Environ Microbiol 12, 1719-1733. Tsai, C. H., Liao, R., Chou, B., Palumbo, M. & Contreras, L. M. (2015). Genome-wide analyses in bacteria show small-RNA enrichment for long and conserved intergenic regions. J Bacteriol 197, 40-50. Tsui, L. C. & Dorfman, R. (2013). The cystic fibrosis gene: a molecular genetic perspective. Cold Spring Harb Perspect Med 3, a009472. Tummler, B. & Cornelis, P. (2005). Pyoverdine receptor: a case of positive Darwinian selection in Pseudomonas aeruginosa. J Bacteriol 187, 3289-3292.

201 References

Turner, K. H., Wessel, A. K., Palmer, G. C., Murray, J. L. & Whiteley, M. (2015). Essential genome of Pseudomonas aeruginosa in cystic fibrosis sputum. Proc Natl Acad Sci U S A 112, 4110-4115. Urban, J. H. & Vogel, J. (2008). Two seemingly homologous noncoding RNAs act hierarchically to activate glmS mRNA translation. PLoS Biol 6, e64. Vallis, A. J., Yahr, T. L., Barbieri, J. T. & Frank, D. W. (1999). Regulation of ExoS production and secretion by Pseudomonas aeruginosa in response to tissue culture conditions. Infect Immun 67, 914-920. van der Doef, H. P., Kokke, F. T., van der Ent, C. K. & Houwen, R. H. (2011). Intestinal obstruction syndromes in cystic fibrosis: meconium ileus, distal intestinal obstruction syndrome, and constipation. Curr Gastroenterol Rep 13, 265-270. Van Ginderdeuren, F., Verbanck, S., Van Cauwelaert, K., Vanlaethem, S., Schuermans, D., Vincken, W. & Malfroot, A. (2008). Chest physiotherapy in cystic fibrosis: short-term effects of autogenic drainage preceded by wet inhalation of saline versus autogenic drainage preceded by intrapulmonary percussive ventilation with saline. Respiration 76, 175-180. van Mansfeld, R., Willems, R., Brimicombe, R., Heijerman, H., van Berkhout, F. T., Wolfs, T., van der Ent, C. & Bonten, M. (2009). Pseudomonas aeruginosa genotype prevalence in Dutch cystic fibrosis patients and age dependency of colonization by various P. aeruginosa sequence types. J Clin Microbiol 47, 4096-4101. van Ulsen, P., Rahman, S., Jong, W. S., Daleke-Schermerhorn, M. H. & Luirink, J. (2014). Type V secretion: from biogenesis to biotechnology. Biochim Biophys Acta 1843, 1592-1611. Varekojis, S. M., Douce, F. H., Flucke, R. L., Filbrun, D. A., Tice, J. S., McCoy, K. S. & Castile, R. G. (2003). A comparison of the therapeutic effectiveness of and preference for postural drainage and percussion, intrapulmonary percussive ventilation, and high-frequency chest wall compression in hospitalized cystic fibrosis patients. Respir Care 48, 24-28. Vasil, M. L., Graham, L. M., Ostroff, R. M., Shortridge, V. D. & Vasil, A. I. (1991). Phospholipase C: molecular biology and contribution to the pathogenesis of Pseudomonas aeruginosa. Antibiot Chemother (1971) 44, 34-47. Vasil, M. L., Stonehouse, M. J., Vasil, A. I., Wadsworth, S. J., Goldfine, H., Bolcome, R. E., 3rd & Chan, J. (2009). A complex extracellular sphingomyelinase of Pseudomonas aeruginosa inhibits angiogenesis by selective cytotoxicity to endothelial cells. PLoS Pathog 5, e1000420. Ventre, I., Goodman, A. L., Vallet-Gely, I., Vasseur, P., Soscia, C., Molin, S., Bleves, S., Lazdunski, A., Lory, S. & Filloux, A. (2006). Multiple sensors control reciprocal expression of Pseudomonas aeruginosa regulatory RNA and virulence genes. Proc Natl Acad Sci U S A 103, 171-176. Verhaeghe, C., Delbecque, K., de Leval, L., Oury, C. & Bours, V. (2007). Early inflammation in the airways of a cystic fibrosis foetus. J Cyst Fibros 6, 304-308. Verstraeten, N., Braeken, K., Debkumari, B., Fauvart, M., Fransaer, J., Vermant, J. & Michiels, J. (2008). Living on a surface: swarming and biofilm formation. Trends Microbiol 16, 496-506.

202 References

Visca, P., Imperi, F. & Lamont, I. L. (2007). Pyoverdine siderophores: from biogenesis to biosignificance. Trends Microbiol 15, 22-30. Voynow, J. A., Fischer, B. M. & Zheng, S. (2008). Proteases and cystic fibrosis. Int J Biochem Cell Biol 40, 1238-1245. Wang, Y., Wilks, J. C., Danhorn, T., Ramos, I., Croal, L. & Newman, D. K. (2011). Phenazine-1-carboxylic acid promotes bacterial biofilm development via ferrous iron acquisition. J Bacteriol 193, 3606-3617. Wargo, M. J. (2013). Homeostasis and catabolism of choline and glycine betaine: lessons from Pseudomonas aeruginosa. Appl Environ Microbiol 79, 2112-2120. Wei, Q. & Ma, L. Z. (2013). Biofilm matrix and its regulation in Pseudomonas aeruginosa. Int J Mol Sci 14, 20983-21005. Weiner, M. P., Anderson, C., Jerpseth, B., Wells, S., Johnson-Browne, B. & Vaillancourt, P. (1994). Studier pET system vectors and hosts. Strateg Mol Biol 7, 41-43. Wiens, J. R., Vasil, A. I., Schurr, M. J. & Vasil, M. L. (2014). Iron-regulated expression of alginate production, mucoid phenotype, and biofilm formation by Pseudomonas aeruginosa. MBio 5, e01010-01013. Williams, P. & Camara, M. (2009). Quorum sensing and environmental adaptation in Pseudomonas aeruginosa: a tale of regulatory networks and multifunctional signal molecules. Curr Opin Microbiol 12, 182-191. Wilson, J. W., Ramamurthy, R., Porwollik, S., McClelland, M., Hammond, T., Allen, P., Ott, C. M., Pierson, D. L. & Nickerson, C. A. (2002). Microarray analysis identifies Salmonella genes belonging to the low-shear modeled microgravity regulon. Proc Natl Acad Sci U S A 99, 13807-13812. Wilson, J. W., Ott, C. M., Honer zu Bentrup, K., Ramamurthy, R., Quick, L., Porwollik, S., Cheng, P., McClelland, M., Tsaprailis, G., Radabaugh, T., Hunt, A., Fernandez, D., Richter, E., Shah, M., Kilcoyne, M., Joshi, L., Nelman-Gonzalez, M., Hing, S., Parra, M., Dumars, P., Norwood, K., Bober, R., Devich, J., Ruggles, A., Goulart, C., Rupert, M., Stodieck, L., Stafford, P., Catella, L., Schurr, M. J., Buchanan, K., Morici, L., McCracken, J., Allen, P., Baker-Coleman, C., Hammond, T., Vogel, J., Nelson, R., Pierson, D. L., Stefanyshyn-Piper, H. M. & Nickerson, C. A. (2007). Space flight alters bacterial gene expression and virulence and reveals a role for global regulator Hfq. Proc Natl Acad Sci U S A 104, 16299-16304. Wilson, J. W., Ott, C. M., Quick, L., Davis, R., Honer zu Bentrup, K., Crabbé, A., Richter, E., Sarker, S., Barrila, J., Porwollik, S., Cheng, P., McClelland, M., Tsaprailis, G., Radabaugh, T., Hunt, A., Shah, M., Nelman-Gonzalez, M., Hing, S., Parra, M., Dumars, P., Norwood, K., Bober, R., Devich, J., Ruggles, A., CdeBaca, A., Narayan, S., Benjamin, J., Goulart, C., Rupert, M., Catella, L., Schurr, M. J., Buchanan, K., Morici, L., McCracken, J., Porter, M. D., Pierson, D. L., Smith, S. M., Mergeay, M., Leys, N., Stefanyshyn-Piper, H. M., Gorie, D. & Nickerson, C. A. (2008). Media ion composition controls regulatory and virulence response of Salmonella in spaceflight. PLoS One 3, e3923.

203 References

Winkler, J. D., Garcia, C., Olson, M., Callaway, E. & Kao, K. C. (2014). Evolved osmotolerant Escherichia coli mutants frequently exhibit defective N-acetylglucosamine catabolism and point mutations in cell shape-regulating protein MreB. Appl Environ Microbiol 80, 3729-3740. Winsor, G. L., Lam, D. K., Fleming, L., Lo, R., Whiteside, M. D., Yu, N. Y., Hancock, R. E. & Brinkman, F. S. (2011). Pseudomonas Genome Database: improved comparative analysis and population genomics capability for Pseudomonas genomes. Nucleic Acids Res 39, D596-600. Winstanley, C., Langille, M. G., Fothergill, J. L., Kukavica-Ibrulj, I., Paradis-Bleau, C., Sanschagrin, F., Thomson, N. R., Winsor, G. L., Quail, M. A., Lennard, N., Bignell, A., Clarke, L., Seeger, K., Saunders, D., Harris, D., Parkhill, J., Hancock, R. E., Brinkman, F. S. & Levesque, R. C. (2009). Newly introduced genomic prophage islands are critical determinants of in vivo competitiveness in the Liverpool Epidemic Strain of Pseudomonas aeruginosa. Genome Res 19, 12-23. Wirth, C., Meyer-Klaucke, W., Pattus, F. & Cobessi, D. (2007). From the periplasmic signaling domain to the extracellular face of an outer membrane signal transducer of Pseudomonas aeruginosa: crystal structure of the ferric pyoverdine outer membrane receptor. J Mol Biol 368, 398-406. Wong, A. P., Bear, C. E., Chin, S., Pasceri, P., Thompson, T. O., Huan, L. J., Ratjen, F., Ellis, J. & Rossant, J. (2012). Directed differentiation of human pluripotent stem cells into mature airway epithelia expressing functional CFTR protein. Nat Biotechnol 30, 876-882. Woo, J. K., Webb, J. S., Kirov, S. M., Kjelleberg, S. & Rice, S. A. (2012). Biofilm dispersal cells of a cystic fibrosis Pseudomonas aeruginosa isolate exhibit variability in functional traits likely to contribute to persistent infection. FEMS Immunol Med Microbiol 66, 251-264. Worlitzsch, D., Tarran, R., Ulrich, M., Schwab, U., Cekici, A., Meyer, K. C., Birrer, P., Bellon, G., Berger, J., Weiss, T., Botzenhart, K., Yankaskas, J. R., Randell, S., Boucher, R. C. & Doring, G. (2002). Effects of reduced mucus oxygen concentration in airway Pseudomonas infections of cystic fibrosis patients. J Clin Invest 109, 317-325. Yang, L., Jelsbak, L., Marvig, R. L., Damkiaer, S., Workman, C. T., Rau, M. H., Hansen, S. K., Folkesson, A., Johansen, H. K., Ciofu, O., Hoiby, N., Sommer, M. O. & Molin, S. (2011). Evolutionary dynamics of bacteria in a human host environment. Proc Natl Acad Sci U S A 108, 7481-7486. Ye, L., Cornelis, P., Guillemyn, K., Ballet, S. & Hammerich, O. (2014). Structure revision of N-mercapto-4-formylcarbostyril produced by Pseudomonas fluorescens G308 to 2-(2- hydroxyphenyl)thiazole-4-carbaldehyde [aeruginaldehyde]. Nat Prod Commun 9, 789-794. Yonker, L. M., Cigana, C., Hurley, B. P. & Bragonzi, A. (2015). Host-pathogen interplay in the respiratory environment of cystic fibrosis. J Cyst Fibros 14, 431-439. Yoon, M. Y., Lee, K. M., Park, Y. & Yoon, S. S. (2011). Contribution of cell elongation to the biofilm formation of Pseudomonas aeruginosa during anaerobic respiration. PLoS One 6, e16105.

204 References

Zerbino, D. R. & Birney, E. (2008). Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res 18, 821-829.Pseudomonas aeruginosa Zhou, Y., Liang, Y., Lynch, K. H., Dennis, J. J. & Wishart, D. S. (2011). PHAST: a fast phage search tool. Nucleic Acids Res 39, W347-352.

205

206 List of publications

8 List of publications

Journal articles:

Goyvaerts, C., De Groeve, K., Dingemans, J., Van Lint, S., Robays, L., Heirman, C., Reiser, J., Zhang, X. Y., Thielemans, K., De Baetselier, P., Raes, G. & Breckpot, K. (2012). Development of the Nanobody display technology to target lentiviral vectors to antigen- presenting cells. Gene Ther 19, 1133-1140. Goyvaerts, C., Dingemans, J., De Groeve, K., Heirman, C., Van Gulck, E., Vanham, G., De Baetselier, P., Thielemans, K., Raes, G. & Breckpot, K. (2013). Targeting of human antigen- presenting cell subsets. J Virol 87, 11304-11308. Hunter, R. C., Asfour, F., Dingemans, J., Osuna, B. L., Samad, T., Malfroot, A., Cornelis, P. & Newman, D. K. (2013). Ferrous iron is a significant component of bioavailable iron in cystic fibrosis airways. MBio 4. Cornelis, P. & Dingemans, J. (2013). Pseudomonas aeruginosa adapts its iron uptake strategies in function of the type of infections. Front Cell Infect Microbiol 3, 75. Elfarash, A., Dingemans, J., Ye, L., Hassan, A. A., Craggs, M., Reimmann, C., Thomas, M. S. & Cornelis, P. (2014). Pore-forming pyocin S5 utilizes the FptA ferripyochelin receptor to kill Pseudomonas aeruginosa. Microbiology 160, 261-269. Dingemans, J., Ye, L., Hildebrand, F., Tontodonati, F., Craggs, M., Bilocq, F., De Vos, D., Crabbe, A., Van Houdt, R., Malfroot, A. & Cornelis, P. (2014). The deletion of TonB- dependent receptor genes is part of the genome reduction process that occurs during adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung. Pathog Dis 71, 26-38. Ghequire, M. G., Dingemans, J., Pirnay, J. P., De Vos, D., Cornelis, P. & De Mot, R. (2014). O serotype-independent susceptibility of Pseudomonas aeruginosa to lectin-like pyocins. Microbiologyopen 3, 875-884. Ye, L., Hildebrand, F., Dingemans, J., Ballet, S., Laus, G., Matthijs, S., Berendsen, R. & Cornelis, P. (2014). Draft genome sequence analysis of a Pseudomonas putida W15Oct28 strain with antagonistic activity to Gram-positive and Pseudomonas sp. pathogens. PLoS One 9, e110038. Vercammen, K., Wei, Q., Charlier, D., Dotsch, A., Haussler, S., Schulz, S., Salvi, F., Gadda, G., Spain, J., Rybtke, M. L., Tolker-Nielsen, T., Dingemans, J., Ye, L. & Cornelis, P. (2015). Pseudomonas aeruginosa LysR PA4203 regulator NmoR acts as a repressor of the PA4202 nmoA gene, encoding a nitronate monooxygenase. J Bacteriol 197, 1026-1039. Dingemans, J., Monsieurs, P., Yu, S.-H., Crabbe, A., Förstner, K., Malfroot, A., Cornelis, P. & Van Houdt, R. Effect of shear stress on Pseudomonas aeruginosa isolated from the cystic fibrosis lung. MBio., under submission.

Dingemans, J., Craggs, M. & Cornelis, P. Identification and functional analysis of S6 pyocin from a cystic fibrosis Pseudomonas aeruginosa clinical isolate. Microbiologyopen., under submission.

207 List of publications

Oral presentations at international conferences:

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Pseudomonas aeruginosa adapts to the cystic fibrosis lung environment by acquiring iron via pyoverdine-independent mechanisms. Oral presentation at the 8th International Biometals Symposium, Brussels, Belgium, July 18, 2012.

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Pseudomonas aeruginosa adapts to the cystic fibrosis lung environment by acquiring iron via pyoverdine-independent mechanisms. Oral presentation at the 7th European CF Young Investigator Meeting, Paris, France, February 28, 2013.

Dingemans J, Craggs M, Crabbé A, Malfroot A, Cornelis P. Genome sequence analysis of the highly adapted epidemic Pseudomonas aeruginosa isolate CF_PA39, retrieved from a Belgian cystic fibrosis patient. Oral presentation at the 14th International Conference on Pseudomonas, Lausanne, Switzerland, September 10, 2013.

Dingemans J, Crabbé A, Van Houdt R, Malfroot A, Cornelis P. Effect of shear stress on Pseudomonas aeruginosa isolated from the cystic fibrosis lung. Oral poster presentation at the Society for General Microbiology Annual Conference, Maastricht, The Netherlands, June 10, 2015.

Poster presentations at international conferences:

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Typing Pseudomonas aeruginosa isolates from Belgian cystic fibrosis patients attending the UZ Brussels by means of ferripyoverdine receptor typing, analysis of virulence factor production and real-time PCR. Poster presentation at the 13th International Conference on Pseudomonas, Sydney, Australia, September 5, 2011.

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Pseudomonas aeruginosa adapts to the cystic fibrosis lung environment by acquiring iron via pyoverdine-independent mechanisms. Poster presentation at the 8th International Biometals Symposium, Brussels, Belgium, July 16-18, 2012.

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Pseudomonas aeruginosa adapts to the cystic fibrosis lung environment by acquiring iron via pyoverdine-independent mechanisms. Poster presentation at the 14th International Symposium on Microbial Ecology, Copenhagen, Denmark, August 23-24, 2012.

Dingemans J, Crabbé A, Van Houdt R, Cornelis P, Malfroot A. Pseudomonas aeruginosa adapts to the cystic fibrosis lung environment by acquiring iron via pyoverdine-independent mechanisms. Poster presentation at the 7th European CF Young Investigator Meeting, Paris, France, February 28, 2013.

208 List of publications

Dingemans J, Craggs M, Bilocq F, Willekens J, Eyns H, Crabbé A, Pirnay J-P, De Vos D, Malfroot A, Cornelis P. The use of rep-PCR (DiversilabTM, BioMérieux) in combination with multiplex PCR (targeting virulence genes) reveals the transmission of Pseudomonas aeruginosa Isolates among cystic fibrosis patients in a hospital background. Poster presentation at the 36th European Cystic Fibrosis Conference, Lisbon, Portugal, June 14, 2013.

Dingemans J, Hildebrand F, Ye L, Crabbé A, Malfroot A, Cornelis P. Genome sequence analysis of the highly adapted epidemic Pseudomonas aeruginosa isolate CF_PA39, retrieved from a belgian cystic fibrosis patient. Poster presentation at the 5th Congress of European Microbiologists (FEMS 2013), Leipzig, Germany, July 22, 2013.

Dingemans J, Hildebrand F, Ye L, Crabbé A, Malfroot A, Cornelis P. Genome sequence analysis of the highly adapted epidemic Pseudomonas aeruginosa isolate CF_PA39, retrieved from a Belgian cystic fibrosis patient. Poster presentation at the 14th International Conference on Pseudomonas, Lausanne, Switzerland, September 8-10, 2013.

Dingemans J, Craggs M, Crabbé A, Malfroot A, Cornelis P. Identification and functional characterization of a novel S-type pyocin, produced by an epidemic Pseudomonas aeruginosa cystic fibrosis clone. Poster presentation at the 14th International Conference on Pseudomonas, Lausanne, Switzerland, September 8-10, 2013.

Dingemans J, Hildebrand F, Ye L, Crabbé A, Malfroot A, Cornelis P. Genome sequence analysis of the highly adapted epidemic Pseudomonas aeruginosa isolate CF_PA39, retrieved from a Belgian cystic fibrosis patient. Poster presentation at the 16th International Congress on Infectious Diseases (ICID), Cape Town, South Africa, April 5, 2014.

Dingemans J, Ye L, Hildebrand F, Tontodonati F, Craggs M, Bilocq F, De Vos D, Crabbé A, Van Houdt R, Malfroot A, Cornelis P. Deletions comprising TonB-dependent receptor genes frequently occur during adaptation of Pseudomonas aeruginosa to the CF lung environment. Poster presentation at the Society for General Microbiology Annual Conference, Liverpool, United Kingdom, April 16, 2014.

Dingemans J, Ye L, Hildebrand F, Tontodonati F, Craggs M, Bilocq F, De Vos D, Crabbé A, Van Houdt R, Malfroot A, Cornelis P. Genome sequence analysis of the highly adapted epidemic Pseudomonas aeruginosa isolate CF_PA39, retrieved from a Belgian cystic fibrosis patient and identification of a novel S-type pyocin. Poster presentation at the 114th General Meeting of the American Society for Microbiology, Boston, Massachusetts, USA, May 18, 2014.

Dingemans J, Crabbé A, Van Houdt R, Malfroot A, Cornelis P. Effect of shear stress on Pseudomonas aeruginosa isolated from the cystic fibrosis lung. Poster presentation at the Society for General Microbiology Annual Conference, Maastricht, The Netherlands, June 10, 2015

209