Research Collection

Doctoral Thesis

Identification of phage receptors and enzymes essential for phage infection in Erwinia amylovora

Author(s): Knecht, Leandra

Publication Date: 2019-05-10

Permanent Link: https://doi.org/10.3929/ethz-b-000360576

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library

Diss. ETH No. 26009

Identification of phage receptors and enzymes essential for phage infection in Erwinia amylovora

A thesis submitted to attain the degree of

DOCTOR OF SCIENCES of ETH ZURICH (Dr. sc. ETH Zurich)

presented by

LEANDRA EVA KNECHT

MSc ETH Zurich

born on 03.10.1988

citizen of Windisch AG, Switzerland

accepted on the recommendation of

Prof. Dr. Martin J. Loessner, examiner

Prof. Dr. Lars Fieseler, co-examiner

Prof. Dr. Julia Vorholt-Zambelli, co-examiner

2019

Table of Contents

Abbreviations ...... 1

Summary ...... 3

Zusammenfassung ...... 6

1. Introduction ...... 11 1.1. Fire blight ...... 11 1.1.1. Erwinia amylovora ...... 11 1.1.1. Virulence factors...... 11 1.1.2. The disease ...... 12 1.1.3. Treatment options ...... 14 1.1.4. Fire blight in Switzerland ...... 15 1.2. Bacteriophages ...... 16 1.2.1. History ...... 16 1.2.2. Phage classification ...... 17 1.2.3. Life cycle ...... 17 1.2.4. Phage resistance ...... 19 1.2.5. Phage biocontrol ...... 20 1.2.6. Phage biocontrol of E. amylovora ...... 21 1.2.7. Phages of E. amylovora ...... 22 1.3. Aim of the study ...... 23 1.4. References ...... 24 2. Manuscripts ...... 37 2.1. Manuscript I: Receptor identification for phage cocktail composition .... 39 2.2. Manuscript II: Phage infection of E. amylovora requires cellulose ...... 65 2.3. Manuscript III: The role of topB1, rfaE and pgm in phage resistance ...... 95 2.4. Manuscript IV: Y2 resistance affects phage infectivity ...... 123 3. Conclusions and Outlook ...... 149

4. Acknowledgements ...... 155

5. Curriculum Vitae ...... 157

I

Abbreviations

Abbreviations

CFBP Collection Française de Bactéries associées aux Plantes CFU Colony forming units CHF Swiss francs CPC Cetylpyridinium chloride Da Dalton ddH2O Double distilled water DNA Deoxyribonucleic acid ds Double stranded EOP Efficiency of plating EPS Extracellular polysaccharides bp Base pair LB Lysogeny broth LPS Lipopolysaccharide M Molarity min Minutes mM Milimolar MOI Multiplicity of infection NCBI National Centre for Biotechnology Information OD Optical density PAGE Polyacrylamide gel electrophoresis PBS Phosphate-buffered saline PCR Polymerase chain reaction PFU Plaque forming unit RBS Ribosomal binding site RT Room temperature SD Standard deviation SOC Super optimal broth with catabolite repression spp. Species pluralis T3SS Type three secretion system wt Wildtype

1

Summary

Summary

The Gram-negative bacterium Erwinia amylovora is the causative agent of fire blight. This plant disease was classified as one of the ten most devastating plant diseases and affects members of the Rosaceae family. Fire blight first appeared in 1780 in Northern America from where it developed into an almost global threat for apple and pear farms. Under favourable conditions, the highly contagious pathogen is able to infect an entire orchard within a single season. Disease treatments are time and cost intensive since susceptible plants must be monitored and sanitized regularly. Antibiotics such as streptomycin are efficient in controlling the plant infection. However, the rise of antibiotic resistant strains and increasing public health concerns prompted increasing numbers of countries to ban streptomycin for agricultural purposes. Alternatives to antibiotics are therefore urgently needed. Over recent years, bacteriophages have been emerging as possible alternatives to conventional antibiotic treatments. These bacterial viruses specifically target and destroy host cells by recognizing receptors on their surface and have several advantages over antibiotics. Phages pose the most abundant entity on earth and outnumber bacteria by ten fold. This abundance ensures almost unlimited possibilities in combining different phages. Phages are highly specific and can only target host bacteria, leaving potentially beneficial bacteria unharmed. Since phages require their host metabolism for reproduction, phage numbers will increase in the presence of their hosts but start to decay in their absence. Finally, phages are considered as non toxic and environmentally safe. Although phages can adapt to modifying hosts, the abundance of estimated 1025 phage infections occurring per second, forces bacteria to adapt to the phages and develop phage resistance. To ensure that phage biocontrol is an efficient and especially long-lasting treatment option against pathogenic bacteria, the risk of resistance development must be minimized. The targeted receptor on the host surface can be modified, thereby ensuring phage resistance. Combinations of phages targeting different receptors could be applied to circumvent resistance development, since the bacterium would have to mutate several receptors simultaneously. To identify host receptors targeted by six well-characterized, highly E. amylovora specific phages, a transposon mutagenesis library was screened for phage resistant mutants. Transposon insertions in Bue1 and Y2 resistant mutants could mostly be linked to LPS biosynthesis and LPS export. It is likely that both phages require LPS structure of E. amylovora for host identification and successful infection. The phages L1 and S2 were unable to lyse mutants with disrupted ams genes. The ams operon is responsible for encoding the amylovoran synthesis apparatus, which generates one of the major EPS components secreted by E. amylovora and functions as virulence factor in fire blight pathogenicity. These findings and the fact that both phages harbour depolymerases, which are able to degrade

3 Summary amylovoran suggests that amylovoran is targeted by L1 and S2 as receptor. M7 and S6 were unable to infect mutants with transposon insertions in the bcs operon. This operon encodes the bacterial cellulose synthase, which is required for bacterial cellulose production and secretion. This carbohydrate polymer is required for stable biofilm production. Deletion of the entire operon as well as deletion of the key genes bcsA encoding the catalytically active subunit, or bcsC, which encodes the outer membrane protein BcsC demonstrated M7 and S6 resistance. Experiments with the cellulose binding dye Congo Red also protected bacteria from phage infection, suggesting that both the cellulose synthase complex and cellulose are required for M7 or S6 infection. In addition, a collection of genes in the M7 and S6 genomes were identified that could encode enzymes with cellulase or endoglucanase activity. The incubation of entire phages with cellulose verified cellulolytic activity for S6 phages. These findings support the hypothesis that M7 and S6 specifically target bacterial cellulose secreted by the host bacterium. These enzymes could have potential to target biofilm-forming bacteria such as E. amylovora. The fact that phages encode cellulase and are able to target bacterial cellulose or the cellulase synthase complex as host receptors is, to our knowledge a novelty and should be further investigated. The results generated from the screen were used to test different phage combinations in vitro and on blossoms for their potential as biocontrol agent. M7 or S6 alone were the most potent single treatment in vitro. Combinations of phages belonging to different phage receptor groups were generally more potent than combinations from within a phage receptor group. Combining Bue1 with a phage from a different host receptor group demonstrate strong biocontrol ability. As L1 and S2, Bue1 was also observed to encode a depolymerase. Triple combinations of phages belonging to different host receptor groups, were observed to efficiently keep cell counts below the detection limit during the entire experiment. On blossoms the treatment with L1 alone, M7+S6 or Bue1+S2+S6 were identified as the most potent phage treatment. To estimate the danger of phage resistance, a special emphasis was laid on the occurrence of multi phage resistance, which could render a phage cocktail futile. The screen was used to identify mutants, which are able to resist attacks from phages belonging to different host receptor groups. These mutants revealed transposon insertions in the genes topB1 (NAD- dependent epimerase), rfaE (ADP-heptose synthase) or pgm (phosphoglucomutase). The encoded gene products are involved in different metabolic processes that can affect extracellular structures. Deletions of these genes was observed to impact LPS structure and reduces the amount of EPS and bacterial cellulose produced by the three mutants. These modifications can explain how these deletion mutants are able to resist infection of phages targeting these structures. In addition, the spontaneously Y2 resistant mutant 1430Y2R was observed to permanently expose Y2 resistance and demonstrate cross resistance. Whole genome sequencing revealed

4 Summary a nucleotide deletion in the gene with the locus tag EAMY_2231, a putative glycosyltransferase, as main source for the permanent resistance. Both LPS dependent phages Bue1 and Y2 were unable to infect this mutant suggesting that the gene affects LPS biosynthesis. The phages L1, S2 and S6 were observed to have a reduced infectivity towards 1430Y2R in vitro. Only M7 infectivity was unaffected by the mutation. Analysis of 1430Y2R revealed strong alteration in the LPS structure and a reduced EPS production. These modifications explain why Bue1 and Y2 are unable to infect and L1 and S2 are reduced in their efficacy of plating. The impact the EAMY_2231 mutation has on S6 is yet to be determined and should be further investigated. These four genes were observed to encode proteins involved in manifold process that influence several surface structures such as LPS, EPS and bacterial cellulose. Modification of these four genes can therefore result in multiple phage resistance. The risk such alterations in metabolic processes pose to phage cocktails, was evaluated by investigating potential fitness costs. In contrast to the wildtype mutants with topB1, rfaE or pgm deletions were incapable of colonizing and infecting the blossoms. Since EPS and especially amylovoran is essential to attaching the bacterium to the plant surface, evade the plant immune system and establish full virulence, we assume that the strong reduction of EPS production observed in these mutants is responsible for this phenotype. If a bacterium would apply the strategy of modulating one of these three genes or their gene products to evade phage attack, the alteration would render the pathogen avirulent. Hence, the plant would survive the pathogen attack. The spontaneous Y2 resistance modification of EAMY_2231 however, was observed to generate comparable disease symptoms as wildtype infected plants. It is yet unclear if the observed EAMY_2231 modification can also occur in planta after Y2 exposure. If such mutants can be isolated form infected blossoms treated with Y2, the phage should be excluded from phage cocktails against E. amylovora. By adapting the phage cocktail accordingly, the risk of phage resistance development after phage treatment is minimized.

5 Zusammenfassung

Zusammenfassung

Die Pflanzenkrankheit Feuerbrand befällt spezifisch Pflanzen der Rosengewächs Familie. Die Krankheit trat erstmals 1780 in Nordamerika auf und entwickelte sich zu einer beinahe globalen Bedrohung für die Produktion von Apfel- und Birnen. Unter günstigen Bedingungen kann der hoch ansteckende Feuerbranderreger Erwinia amylovora, ein Gram-negatives Bakterium, innerhalb einer Saison eine ganzen Obstanlage infizieren. Die Prävention und die Behandlung von Feuerbrand ist sehr zeit- und kostenintensiv, da anfällige Pflanzen konstant überwacht und im Falle eines Ausbruchs stark zurückgeschnitten oder vernichtet werden müssen. Feuerbrand wurde darum als eine der zehn verheerendsten Nutzpflanzen- krankheiten eingestuft. Das Antibiotikum Streptomycin wurde lange zur erfolgreichen Bekämpfung des Erregers eingesetzt. Hohe Antibiotikawerte in Honig, Angst vor antibiotikaresistenten Keimen und Gesundheitsbedenken bei der Bevölkerung sorgten dafür, dass die Schweiz den Einsatz von Streptomycin 2016 verbot. Da immer mehr Länder Streptomycin als Bekämpfungsstrategie verbieten, werden momentan alternative Behandlungsmethoden getestet. Der Einsatz von Bakteriophagen als möglich Alternative zu herkömmlichen Antibiotika könnte dabei behilflich sein. Bakteriophagen sind bakterielle Viren die Wirtszellen zur Vermehrung nutzen, bevor diese zerstört werden. Mit geschätzt zehn Mal mehr Phagen als Bakterien bilden die Phagen die häufigste biologische Einheit auf diesem Planeten. Diese Vielfalt gewährleistet, dass mutmasslich für jedes Bakterium ein passender Phage isoliert werden kann. Des Weiterns, biete die Diversität beinahe unbegrenzte Möglichkeiten verscheide Phagen zu kombinieren. Phagen haben mehrere Vorteile gegenüber herkömmlichen Antibiotika. Sie erkennen ihre Wirtszellen gezielt über Rezeptoren auf der Oberfläche der Bakterien und können daher nur Bakterien mit den passenden Rezeptoren infizieren. Die nützliche bakterielle Flora wird dabei von den Phagen verschont. Phagen sind zusätzlich hoch dynamisch. Da sie für ihre Vermehrung auf den Wirtsmetabolismus angewiesen sind, vermehren sie sich nur, wenn Wirtszellen vorhanden sind. Dabei erhöht sich der Phagen Titer. Ohne Wirtszellen werden Phagen zum Beispiel durch UV-Strahlung abgebaut. Da Phagen zudem als unbedenklich für Menschen und Tiere und umweltverträglich eingestuft werden, sind sie eine vielversprechende Alternative zu Antibiotika. Neuste Schätzungen gehen von 1025 Infektion durch Phagen pro Sekunden aus. Diese Häufigkeit zwingt Bakterien sich gegen die Phagen mittels Resistenzentwicklung zu schützen. Obwohl sich Phagen zu einem bestimmten Grad ihren Wirtszellen anpassen können, muss sichergestellt werden, dass sich keine langfristige Phagenresistenz einstellen kann, die die Phagenbehandlung nutzlos machen würde. Durch die Kombination von verschiedenen Phagen, die unterschiedliche Infektionsstrategien nutzen, kann das Risiko einer Resistenzentwicklung minimiert werden.

6 Zusammenfassung

Sechs virulente, E. amylovora spezifische Phagen wurden dazu genauer untersucht. Eine Transposon-Mutagenese Bibliothek wurden überprüft, um Phagenresistente Mutanten zu identifizieren. Obwohl die sechs untersuchten Phagen unterschiedlich sind und zu verschieden Gattungen gehören, konnten sie aufgrund ihrer Rezeptorenpräferenz in drei Gruppen eingeteilte werden. Mutanten die eine Resistenz gegenüber den Phagen Bue1 oder Y2 aufwiesen, hatten häufig eine Insertion in Genen, die zur LPS Biosynthese und Export gebraucht werden. Es ist daher wahrscheinlich, dass beide Phagen E. amylovora Wirtszellen an spezifischen LPS Strukturen erkennen. Die Phagen L1 und S2 waren nicht länger in der Lage, Mutanten mit Insertionen in Genen des ams Operons zu infizieren. Dieses Operon kodiert den Amylovoran-Biosynthese-Komplex, der verantwortlich für die Produktion von Amylovoran ist. Dieses extrazelluläre Polymer ist der Hauptbestandteil des produzierten EPS im Feuerbranderreger und ist einer der Hauptvirulenzfaktoren des Pathogen. Beide Phagen stellen zudem ähnliche Depolymerasen her, die in der Lage sind Amylovoran abzubauen. Da neben dem ams Operon keine extrazellulären Strukturen mit Hilfe der Transposon- Mutagenese identifiziert werden konnten, gehen wir davon aus, dass Amylovoran als hauptsächlicher Rezeptor zur Identifizierung eines Wirtsbakteriums verwendet wird. Mutanten mit Transposoninsertionen in Genen des bcs Operons konnten von den Phagen M7 und S6 nicht infiziert werden. Das bcs Operon kodiert die bakterielle Cellulosesynthase Untereinheiten, die für die Produktion und Sekretion von bakterieller Cellulose zuständig sind. Dieses Kohlenhydratpolymer ist für das Ausbilden eines stabilen Biofilmes erforderlich. Die Deletion des gesamten Operons sowie die Deletion der Schlüsselgene bcsA, dass die enzymatisch aktive Untereinheit kodiert oder bcsC, dass das Membranprotein BcsC kodiert, welches zuständig für den Export der Cellulose ist, sorgten dafür, dass die Mutanten resistent gegenüber M7 und S6 Infektionen waren. Der cellulosebindende Farbstoff Congo Rot schützen die Bakterien ebenfalls vor Phageninfektion. Dies legt nahe, dass sowohl der Cellulosesynthese Komplex als auch Cellulose selber für eine erfolgreiche Phageninfektion notwendig sind. Zusätzlich wurden in den Genomen der beiden Phagen mehrere Gene identifiziert, die Enzyme mit Cellulase- oder Endoglucase-Aktivität kodieren könnten. Die Inkubation ganzer Phagen mit Cellulose, bestätigte eine cellulytische Aktivität für den Phagen S6. Diese Ergebnisse stützen die Hypothese, dass M7 und S6 spezifisch bakterielle Cellulose erkennen und zur Infektion nutzen. Diese phagenkodierten Enzyme haben das Potenzial gegen biofilmbildende Bakterien eingesetzt zu werden. Die Tatsache, dass Phagen Cellulase kodieren und in der Lage sind, bakterielle Cellulose oder den Cellulase-Synthase-Komplex als Wirtsrezeptoren anzugreifen, ist unseres Wissens nach eine Neuheit und muss weiter untersucht werden. Die durch das Screening erhaltenen Ergebnisse wurden verwendet, um verschiedene Phagenkombinationen in vitro und an Blüten auf ihr Potenzial als Biokontrollmittel zu testen.

7 Zusammenfassung

Behandlungen mit M7 oder S6 alleine waren die wirksamste Einzelbehandlung in vitro. Kombinationen aus Phagen, die zu verschiedenen Rezeptorgruppen gehören, erwiesen sich im Allgemeinen als wirksamer als Kombinationen innerhalb einer Rezeptorgruppe. Die Kombination von Bue1 mit einem Phagen aus einer anderen Rezeptorgruppe zeigt hohe Effizienz in der Kontrolle der Bakterien. Triple-Kombinationen mit Phagen, die zu verschiedenen Rezeptorgruppen gehören, demonstrierten Potenzial die Bakterien Zellzahl während des gesamten Experiments unterhalb der Nachweisgrenze zu halten. Auf den Blüten wurde die Behandlung mit L1, M7 + S6 oder Bue1 + S2 + S6 als die vielversprechendsten Phagenbehandlung identifiziert. Diese Daten bilden eine Grundlage für weitere Blüten und Pflanzentests. Um die Gefahr der Phagenresistenz genauer abzuschätzen zu können, wurde ein besonderer Schwerpunkt auf das Auftreten von Multiphagenresistenz gelegt. Diese können einen Phagen Cocktail nutzlos machen. Die Transposon-Mutagenese wurde dazu genutzt Mutanten zu isolieren, die gegen mehrere Phagen resistent sind. Dabei wurden die Gene topB1 (NAD- abhängige Epimerase), rfaE (ADP-Heptosesynthase) pgm (Phosphoglucomutase) identifiziert. Mutanten die eine Transposon Insertion in einem dieser Gene aufwiesen, konnte Phagen aus unterschiedlichen Wirtsrezeptorengruppen abwehren. Verschiedene biologische Datenbanken wurden analysiert um die Funktion der kodierten Genprodukte zu erkennen. Diese scheinen an verschiedenen Stoffwechselprozessen beteiligt zu sein, die extrazelluläre Strukturen beeinflussen können. Mutanten mit Deletionen dieser Gene generierten abweichende LPS Strukturen und produzierten reduzierte Mengen an EPS und bakterieller Cellulose. Diese Änderungen sind wahrscheinlich verantwortlich, dass Phagen die diese Oberflächenstrukturen als Rezeptoren nutzen, nicht mehr in der Lage sind diese Mutanten zu infizieren. Zusätzlich zu diesen drei Genen, konnte anhand spontaner Resistenzexperimenten, ein permanenter Y2 resistenter Keim (1430Y2R) isoliert werden der zudem Kreuzresistenzen gegenüber anderen Phagen aufweist. Die Genomsequenzierung des isolierten Stamms zeigte, dass eine Nukleotiddeletion in einem Gen mit dem Locus-Tag EAMY_2231, einer potenziellen Glycosyltransferase, verantwortlich für die beobachtete Resistenz ist. Da die beiden LPS abhängigen Phagen Bue1 und Y2 unfähig sind diesen Stamm zu infizieren, kann daraus geschossen werde, dass EAMY_2231 die LPS Biosynthese beeinflusst. Bei L1, S2 und S6 wurde in vitro eine verringerte Infektiosität gegenüber der Y2-resistenten Mutante beobachtet. Nur M7 wurde von der Mutation nicht beeinträchtigt. Tatsächlich zeigten Analysen des 1430Y2R Keims, eine starke Veränderung der LPS Struktur, sowie eine Verringerung der sekretierten EPS Menge. Diese Modifikationen erklären, warum Bue1 und Y2 nicht infizieren können und L1 und S2 in ihrer Wirksamkeit eingeschränkt sind. Die Auswirkungen der EAMY_2231-Mutation auf S6 sind noch unklar und sollten weiter untersucht werden.

8 Zusammenfassung

Diese Daten zeigen, dass die vier untersuchten Gene topB1, rfaE, pgm und EAMY_2231 Einfluss auf die Herstellung oder den Export mehrere Oberflächenstrukturen wie LPS, EPS oder bakterieller Cellulose haben und somit multiple Phagenresistenz vermitteln. Um zu untersuchen, ob diese Gene und deren Genprodukte ein Risiko für die Behandlung von E. amylovora mittels Phagen Cocktails darstellen, wurden die Mutanten auf allfällige Fitness Einbußen untersucht. Im Gegensatz zum Wildtype waren Mutanten mit topB1, rfaE- oder pgm- Deletionen nicht in der Lage, die Blüten zu kolonisieren und zu infizieren. Da EPS und vor allem Amylovoran essentiell sind, um das Bakterium an die Pflanzenoberfläche anzuhaften, das pflanzliche Immunsystem zu umgehen und als Virulenzfaktor fungieren, kann davon ausgegangen werden, dass die bei diesen Mutanten beobachtete starke Verringerung der EPS-Produktion, für diesen Phänotyp verantwortlich ist. Für den Fall, dass ein Bakterium tatsächlich die Strategie eines dieser Gene zu modifizieren anwendet um einen Phagen Angriff vorzubeugen, würde diese Veränderung den Keim avirulent machen. Somit würde die Pflanze den Befall überleben. Bakterien mit EAMY_2231-Modifikation hingegen, generierten gleich starke Krankheitssymptome wie der Wildtype. Unsicher ist, ob die in vitro beobachtete EAMY_2231-Modifikation auch nach einer Y2-Exposition in planta auftreten kann. Für den Fall, dass solche Mutanten aus infizierten Blüten, die mit Y2 behandelt wurden, isoliert werden können, sollte der Phage nicht in einen Phagencocktail gegen E. amylovora integriert werden. Die entsprechenden Anpassungen des Phagencocktails sollten sicherstellen, dass das Risiko der Resistenzentwicklung nach einer Phagenbehandlung minimiert wird.

9

Introduction

1. Introduction

1.1. Fire blight

1.1.1. Erwinia amylovora

The Gram-negative, rod shaped bacterium Erwinia amylovora belongs to the order of Enterobacteriales and is the type strain for the newly established Erwiniaceae family (1). The family contains mostly plant pathogens, non-pathogenic epiphytes and even opportunistic human pathogens have been reported (2). E. amylovora is peritrichously flagellated and can reach a size of 0.5-1.0 x 1.0-3.0 µm. The bacterium is able to grow at temperatures ranging from 3 to 37°C with a growth optimum between 25 and 27°C (3). Growth is inhibited at pH = 5, and nearly abolished at lower pH (4).

1.1.1. Virulence factors

E. amylovora is the causative agent of fire blight, a plant disease affecting members of the Rosaeceae family (5–9). In addition to the metalloprotease (10), the sucrose metabolism (11) and the desferrioxamine siderophores (12, 13), E. amylovora relies on two main virulence factors. Its pathogenicity is dependent on a type three secretion system (T3SS) and the secreted extracellular polysaccharides (EPS). E. amylovora harbours a set of genes termed hypersensitive reaction and pathogenicity (hrp), which is located in an apparent pathogenicity island (PI). The PI of E. amylovora contains up to 60 genes that can be classified into four gene types. The first group is composed of several genes that are required for the transfer of the pathogenicity island. The second group comprises regulatory genes that regulate the expression of the hrp genes. Secretory genes, which encode the Hrp T3SS that spans the inner and outer membrane of the bacterium with its needle-like structure, are the third group. Finally, proteins that will be secreted by the T3SS compose the fourth group (14). The T3SS and the secreted proteins are required for pathogenicity and induce the hypersensitive reaction in non-host plants (14). These proteins include the harpins HrpN and HrpW, which are released into the apoplast of the plant tissue (15). There, HrpN, which is essential to generate full virulence, potentially forms pores in the plant plasma membrane and facilitates translocation of DspA/E into the plant cell (16). The injected disease specific effector protein DspA/E will interfere with plant signalling and induces cell death (17). The expression of these hrp genes is tightly regulated. Expression is only induced under certain conditions such as low nutrients or low pH (18). In response to these environmental factors, the genes hrpS, hrpX

11 Introduction and hrpY will be induced, which in turn regulate the expression of hrpL, the master switch of the hrp genes (19, 20). E. amylovora is able to produce a collection of EPS, the second major virulence factor. EPS surrounds the pathogen, protects it from harsh environmental conditions and helps to attach the pathogen to the plant surface (21, 22). By masking critical structures on the bacterial surface that could induce plant defence mechanisms, bacteria can evade the plant immune system (23, 24). E. amylovora secretes two different EPS required for full virulence. Levan is a homopolymer derived from sucrose. Mutants lacking the levansuccrase encoding gene lsc, that is responsible for levan formation, were affected in biofilm formation and cell to cell aggregation (22). The second EPS, amylovoran, is a complex polysaccharide with a galactose backbone. It can vary strongly in length due to environmental and cell-metabolism related factors (25). The amylovoran synthesis complex responsible for amylovoran production is encoded in the ams operon. Amylovoran production correlates with virulence in E. amylovora (26). Both amylovoran and levan are required for biofilm formation (27). In the plant tissue, the pathogen induces biofilm production to ensure nutrient uptake and protection from the plant immune system. In addition to amylovoran and levan, E. amylovora produces bacterial cellulose. This polysaccharide is responsible for the three-dimensional structure of the biofilm and thereby ensures its stability and integrity (28). Congestions of EPS and biofilm in the plant tissue are responsible for the clogging of xylem vessels that eventually results in the characteristic disease symptoms. Biofilm formation and the production of its components is tightly regulated by the intracellular second messenger cyclic-di-GMP (c-di-GMP) (29, 30). Recent studies suggest that c-di-GMP has a critical function in bacteria. It is suggested to mediate the transition between motile and sessile life style of bacteria by downregulating flagella and upregulating biofilm production (30–32). Furthermore, c-di-GMP is supposedly responsible to orchestrate the transition between chronic and acute infection by controlling T3SS expression (29, 33–35).

1.1.2. The disease

Members of the Rosaceae family can be infected by fire blight (5). The economically most valuable plants in this family are apples and pears. First observations of fire blight symptoms were made in 1780 in the Hudson valley of New York state (36). From there the disease spread slowly all over the entire country. With increasing human mobility and trade, fire blight started to spread to neighbouring countries at the beginning of the 20th century. Around 1950, first fire blight cases were observed in Europe, possibly due to infested wood or trees imported from Northern America (5). Until today, the disease has spread to multiple countries on all continents.

12 Introduction

Figure 1 Schematic disease cycle. Modified from (37)

E. amylovora can overwinter in infected plant parts. In spring, the pathogen starts to replicate in these cankers, which can start to develop sticky ooze droplets consisting of a hygroscopic polysaccharide matrix and bacteria. Through rain or insects, the pathogen reaches susceptible host plants (Figure 1). Although the pathogen is able to infect the plant through natural openings or wounds, its preferred route of entry is through the blossoms. There, E. amylovora replicates on the stigma and, under favourable conditions, can reach the critical number of 105-106 CFU/flower, whereupon infection through the nectaries is induced (38, 39). Once inside the plant, the pathogen migrates through the plant tissue, multiplies and starts to form a thick biofilm. This results in the collapse of the parenchyma and the clogging of the xylem vessels (19, 24, 28, 40, 41). Characteristic disease symptoms such as swelling of tissue and ooze droplets become visible (Figure 2). Eventually, infected plant parts start to wilt and appear blackened, hence the name fire blight (41). The formed ooze droplets are a source for secondary infection of shoots and fruits. Flower and shoot infection will progress to form cankers in late summer to ensure overwintering of the pathogen (9). The disease is highly infectious and difficult to manage. Under optimal conditions, the pathogen can spread rapidly and infect or destroy entire orchards within a single season. Time consuming and cost intensive sanitation of orchards is therefore required to manage the disease. Heavy fire blight outbreaks are registered regularly and can generate substantial financial losses in affected countries (5, 42).

13 Introduction

Figure 2 Fire blight symptoms on susceptible host plants. Top row: Infected tree with browned leaves and branches (left); infected twig (centre); infected flower cluster (right). Bottom row: Infected flowers with ooze droplets on twigs (left); Highly infected apple with ooze droplets (right). All pictures were taken from Agroscope with courtesy of Eduard Holliger.

1.1.3. Treatment options

With rising global temperatures and increasing use of susceptible cultivars, the disease is likely to generate more severe outbreaks. Therefore, both reliable and environmentally friendly treatment options and the research on fire blight robust apple and pear cultivars can help to manage the disease (43). The most promising stage to target E. amylovora is during the blossoming period. At this stage, treatments can lower E. amylovora counts below the critical number necessary for entry into the plant tissue and thereby prevent disease outbreak. Since the disease progression is tightly linked to environmental conditions such as temperature and rainfall, prognosis models can help to predict high-risk periods and enable coordinated treatment. Maryblyt, Cougar blight and RIMpro are three frequently used fire blight forecast models. Maryblyt indicates a possible infection day, when all of the following four criteria are fulfilled: (i) presence of intact and opening blossoms, (ii) accumulation of a minimum of 110 degree hours >18.3°C, (iii) average daily temp >15.6°C, (iv) rain fall or dew >0.25 mm (44, 45). The forecasting models Cougarblight and RIMpro use different temperature thresholds and also consider if a particular orchard had fire blight outbreaks before and if cankers were identified (39, 46). RIMpro further includes blossom age as indicator to calculate the infectious

14 Introduction potential (46). Although these forecasting models can misinterpret the compiled data, the use of them greatly facilitates and improves treatment options and application timing for farmers. A collection of chemicals, including copper compounds and antibiotics, were previously tested against the fire blight pathogen and are summarized by Psallidas et al. (47). Although copper compounds generated good results against other bacteria or fungal pathogens, they were observed to demonstrate phytotoxicity, especially against pear plants (48). Several antibiotics were tested for their potency against E. amylovora. Aside from oxytetracyclin and kasugamycin, streptomycin was shown to effectively treat the disease. Streptomycin was therefore extensively used to manage fire blight outbreaks during the blossoming period. However, the application of streptomycin resulted in elevated levels of antibiotics found in honey and the appearance of antibiotic resistant bacteria in several countries (4, 49, 50). The application of streptomycin has consequently been banned in an increasing number of countries (51). Treatments options with aluminium are currently under investigation in different countries (52, 53). The compounds potassium aluminium sulphate (LMA) and aluminium sulphate (Myco- sin) were observed to generate good symptom reduction especially when applied in short intervals but efficacy varied strongly (54, 55). However, aluminium compounds are not allowed in organic farming and aluminium accumulation could have an effect on soil and water environments (4). An environmentally friendly solution could be the use of antagonistic microorganisms, which can act against the pathogen either by competitive exclusion or the production of bactericides or antibiotics. A collection of organic treatment options including the yeast Aureobasidium pullulans (Blossom protect), Bacillus subtilis (Serenade), Pseudomonas fluorescens, Pantoea vagans (BlightBan), Pantoea agglomerans (Pomavita) and Metschnikovia pulcherrima are currently under investigation (4, 47, 52, 54, 56–60). Finally, the application of bacteriophages was shown to have the potential to eliminate E. amylovora efficiently and specifically (61–63). So far, most of the alternative treatments have drawbacks due to their environmental impacts, due to lower efficacy against fire blight compared to streptomycin or due to varying performance.

1.1.4. Fire blight in Switzerland

First appearance of fire blight symptoms were observed in 1989 on Cotoneaster sp. followed by the first disease outbreak in apple and pear orchards in 1991 (64, 65). Since then, the disease has reached every canton in Switzerland and caused more than 100 million US$ reparation and sanitation costs and economic losses (66). In 2007, high temperature and optimal disease conditions at the beginning of April caused the most severe outbreak in Switzerland. As a result of this outbreak, over 45,000 old standard trees were registered as infected and intensive sanitation of affected orchards had to be carried out. Approximatively

15 Introduction

10,000 standard trees and more than 100 hectares of half-standard orchards (approx. 250,000 trees) had to be eliminated (67). The damage generated by the fire blight outbreak in 2007 amounted to 30 million CHF (64). Consequentially, the Swiss Federal Office for Agriculture (FOAG) permitted the restricted use of streptomycin during the blossoming season in 2008 and the approval was renewed annually (68). Increased levels of antibiotics found in honey, the appearance of streptomycin resistant strains in other countries and public health concerns, however, motivated the FOAG to ban streptomycin as fire blight control agent in 2016 (4, 51, 69, 70). The FOAG stated that, aside from these concerns, the enforced collaboration between farmers and public authorities, the development of a fire blight tolerant apple cultivar (Ladina) and the approval of alternative treatment options could ensure sustainable fire blight management (70). In 2020, a new law will come into effect, which classifies E. amylovora as no longer required to be declared to the authorities or controlled, except for the protected area (canton Wallis) (71). Nevertheless, severe fire blight outbreaks can be observed since 2016, suggesting that the current treatment options struggle to manage the disease especially under optimal conditions for the pathogen.

1.2. Bacteriophages

1.2.1. History

Bacteriophages are viruses exclusively infecting bacteria. With an estimated number of 1031 (72), phages pose the most abundant biological entity on earth and outnumber bacteria by at least tenfold (73). Phages are ubiquitous and can readily be isolated from lakes, soils, sediments, biofilms and microfloras of humans, animals and plants. Since phages are dependent on their hosts’ metabolism for reproduction, they can often be found in close proximity to them (74). Phages were first identified in 1915 by Frederick Twort, who described the phenomenon of glassy transformation of bacteria (75). In 1917, Felix d’Hérelle independently discovered these “bacteria eating microorganisms”, which could lyse bacteria in liquid and produce plaques on agar plates and named them bacteriophage. He realised their potential and investigated them as remedy against a collection of bacterial pests (76, 77). The identification of penicillin in 1928 by Alexander Fleming (79), however, diverted the focus from phage therapy and shifted it to the newly discovered antibiotics (80). Eastern European countries with hotspots in Tbilisi (Georgia) and Wroclaw (Poland) maintained an interest in advancing phage therapy as treatment against pathogenic bacteria. In western countries, however, the potential of phage therapy sunk into oblivion. Instead of phages as antimicrobial agents, investigating them as model organisms became more popular. Several studies used phages to elucidate the basic principles of life including the nature and function of DNA and the genetic code, gene expression or studies on cell regulation (78). Whole genome

16 Introduction sequencing was first carried out for phages (79). Phage display or the recent exploit of the CRISPR/Cas system (Clustered regularly interspaced short palindromic repeats, CRISPR associate proteins) further highlight the benefits phages provide to research (80, 81). Phage research has therefore paved the way to many scientific discoveries in the last century and with the rise of antibiotic resistant pathogens, phage therapy is gaining global interest once more.

1.2.2. Phage classification

The International Committee on Taxonomy of Viruses (ICTV) provides guidelines to classify phages. According to the latest annual release of the ICTV, there are 9 orders, 131 families, 46 subfamilies, 803 genera and 4853 species but the numbers will surely already have increased further (82). Currently, bacteriophages are classified foremost according to their nucleic acid and morphology and can further be grouped depending on their genome configuration and size, host range and genetic similarities. However, this taxonomical system is not taking overall DNA and protein identity and phylogeny into account. With increasing numbers of sequenced phage genomes, more data is available for genomic comparisons. Hence, it is possible that, in the coming years, phage taxonomy will be reformed accordingly. Until now, phages with double stranded DNA and tail structures are classified as Caudovirales. The order Caudovirales currently consists of approx. 96% of all identified phages (83) and can be subdivided into five families: Myoviridae, Siphoviridae, Podoviridae, Ackermannviridae and Herelleviridae (84, 85).

1.2.3. Life cycle

Phage infection starts by identifying a potential host bacterium and adsorbing to a host receptor located on the bacterial surface. The host receptors can be integrated into the outer membrane (86). In Gram negative bacteria, sugar molecules of the LPS core polysaccharides and the O-polysaccharide chain are prominent receptors recognized by a variety of phages. Other phages target membrane proteins like OmpA, OmpC or LamB. In addition to membrane bound molecules, certain phages recognize other structures such as flagella (87, 88), pili (89), or capsules (90) as host receptors. An extensive overview of phage receptors can be obtained by Bertozzi et al (88). After attachment to the host cell, the phage injects its DNA into the bacteria. Phage can be distinguished according to their life styles as strictly lytic or temperate (Figure 3). Strictly lytic phages are exclusively performing the lytic cycle. They infect the host cell and immediately degrade the host DNA, hijack the bacterial metabolism, and force the host to produce phage progenies. After production and assembly of new phages, the host cell will be lysed by endolysins. These phage encoded enzymes are specialized in degrading the host’s peptidogylcan layer. This process losens the strucutral integrity of the bacterial cell wall

17 Introduction and eventually results in host lysis. Finally, the burst of the infected bacterium releases the newly formed infective phage particles (91, 92). Temperate phages, on the other hand, are equipped to insert their DNA into the host genome after infection. Alternatively, the phage DNA can remain in the host in a plasmid like state. In both cases, phage gene expression is downregulated to avoid detection and degradation by the bacterium. Unaware of the infection, the host cell can grow and replicate normally with the integrated phage DNA. With each replication of the host bacterium, the prophage will be transferred to each daugther cell. If the infected host cell experiences stress like UV- irradiation, starvation, antibiotics or experiences optimal conditions for phage replication, the prophage can reactivate itself and transition into the lytic cycle (91, 93).

Figure 3 Schematic phage life cycle. Adapted from (94)

Phage infection can have a strong impact on evolution through horizontal gene transfer. Imprecise excision of a prophage during the transition between lysogenic to lytic cycle, can result in the packaging of bacterial DNA up- or downstream of the inserted prophage. During the infection of a new host cell, phage DNA and the accidentally integrated bacterial DNA can be inserted into the genome of the new host. This transfer of bacterial DNA to another bacterial genome by temperate phages is termed specialized transduction. The transferred bacterial DNA can impact the phenotype or fitness of the new host cell (lysogenic conversion). The most prominent lysogenic conversion genes are the cholera toxin and the Shiga toxins. In both cases, the toxin encoding genes are transported by phages. Subsequent infection of the host bacteria Vibrio cholerae or Escherichia coli, respectively, converts the non toxic hosts into toxin producing pathogens (95, 96). Aside from the specialised transduction and the lysogenic conversion, temperate phage can impact a host bacterium by inserting its genome into a coding region. This will disrupt the gene in question. Once inserted into the host genome,

18 Introduction homologous recombination of prophages and host DNA can occur and rearrange the host genome (97, 98). All these modifications can have long lasting effects on the host bacterium. Strictly lytic phages can also accidentally package bacterial host DNA instead of phage DNA when the host genome is insufficiently degraded. These phages are still virulent and can infect new host cells. However, instead of phage DNA, these phages injects bacterial DNA into the new host bacterium. The injected bacterial DNA can then be integrated into the host genome through homologous recombination (99). This mechanism is termed generalized transduction and exclusively occurs in phages that use the headfull packaging strategy, which has a weak sequence specificity (100). If host DNA is degraded poorly during phage infection, substantial amounts of undegraded bacterial plasmid DNA can be released into the environment after cell lysis. Naturally competent bacteria can then take up the released DNA. This process can potentially contribute to the spread of antibiotic resistance in natural environments (101).

1.2.4. Phage resistance

It has been estimated that approximatively 1025 phage infections occur every second (102). Bacteria are therefore forced to adapt to this phage pressure. They have several mechanisms to interfere with phage infection at different stages (103, 104). By mutating or masking the targeted receptor, bacteria can protect themselves from phage infection (103, 105–109). Phage adsorption can also be blocked by competitive inhibitor molecules that outcompete the phage for receptor binding (110). The injection of phage DNA can be blocked by superinfection exclusion (sie) systems (111, 112). These systems are usually encoded by prophages to protect the lysogenized host from further phage infections (113). If phage DNA reaches the cytoplasm of the host, a collection of intracellular defence mechanisms can act to prevent phage replication. BREX (Bacteriophage exclusion) and DISARM systems (Defence Island System Associated with Restriction-Modification) are systems that can both block phage replication by marking the phage DNA as foreign (114, 115). Restriction-modification (R-M) systems recognize the incoming DNA as foreign and digest it unspecifically (116). CRISPR/Cas systems function similarly such that incoming phage DNA is identified and cleaved. In contrast to the R-M system, CRISPR/Cas systems are highly specific against certain phages (117). Both systems cleave phage DNA and ensure survival of the bacterium. The abortive infection system (abi), however, is fatal for the infected bacteria. These abi systems are diverse in sensing and reacting to phage infection. Once activated, the infected bacterium destroys itself thereby preventing phage proliferation. This sacrifice guarantees survival of the surrounding bacterial population (113). Phage resistance is observed to be either transient or permanent (118, 119). Permanent resistance against phages can involve spontaneous genetic mutations. Such mutations can possibly impact fitness or virulence of

19 Introduction the phage resistant strain (120–122). In the case of a transient resistance, the phage sensitive state is re-established as soon as phages are removed. Phase variation can be applied to generate transient resistance (123, 124). Alterations of prominent structures such as capsules and flagella can be regulated by phase variation as observed in Campylobacter jejuni (125) or Salmonella enterica (126, 127). Modifications in response to a particular phage can potentially entail resistance against other phages. The mutation of a surface receptor for example can mediate resistance against several phages that recognize this particular receptor. Mutation of a global regulator such as the alternative sigma factor rpoN can even generate resistance against a collection of phages with different infection mechanisms due to its involvement in multiple intracellular processes (128).

1.2.5. Phage biocontrol

Ever since Felix d’Hérelle isolated phages in 1917, their potential against pathogenic bacteria was investigated extensively (129–132). With the rise of antibiotic resistant bacteria, phage therapy moves once more into the spotlight (129). The idea behind phage therapy is captivatingly simple. Phages can be isolated readily against possibly any bacteria of interest. The infection of these bacteria will result in their destruction and simultaneously release more infectious phage particles rendering the treatment more potent. In the absence of host bacteria, phages will eventually decay e.g. through UV irradiation (133–135). In contrast to antibiotics, phages only recognize and target specific host bacteria, while other potentially beneficial bacteria are left unharmed (136). Phages do not cause side effects in humans and animals, and were considered as environmentally safe (137, 138). Nevertheless, phages for phage therapy should be well characterized and selected with great care. Phages suitable for biocontrol agents should be easy to propagate to high titres and have to be stable during application and storage. They should have a broad host range and be unable to infect the natural flora. Whole genome sequencing can reveal if phages encode toxins and pathogenicity associated factors and if they are prone to transfer bacterial DNA via generalized transduction. In general, phages that exclusively perform the lytic life cycle should be considered for phage therapy (138). Although temperate phages eventually lyse their host bacteria, the transition from lysogenic to lytic cycle is dependent on environmental factors and therefore not reliable. Furthermore, temperate phages can mediate specialized transduction or enhance the fitness of the host bacterium. The use of synthetically engineered temperate phages exhibits potential and is currently discussed (139), but the application of such phages should be evaluated thoroughly. Although phages have the ability to co-evolve with their host bacteria (103, 140), the development of phage resistance must be avoided. Hence, certain precautions should be taken to ensure treatment effectivity. Identifying phages that are known to have the ability to

20 Introduction bypass or outwit the applied resistance mechanism can help to minimize the risk of resistance development. By combining phages with different infection strategies, the risk of resistance through receptor modification can be reduced (130, 141, 142). Some phage mixtures even demonstrate synergistic effects, which render the treatment more potent (143–145). Certain phage were observed to characteristically modify their DNA to protect themselves from bacterial restriction enzymes (146). The addition of such phages to the cocktail could bypass resistance through RM systems. Phages that target virulence factors are also promising for phage cocktails. In response to these phages, the bacterium will modify the targeted receptors thereby reducing or abolishing its virulence (122). Finally, careful and thorough investigation of the host is most advantageous. Analysis of the host genome can detect CRISPR/Cas, sie systems and certain RM systems and the risk of phage resistance can thus be anticipated.

1.2.6. Phage biocontrol of E. amylovora

Phage are promising control agents against plant pathogens. Already in 1924, first experiments were conducted to use phage biocontrol agents as control measurement against Xanthomonas campestris pv. campestris (137). Since then, phage biocontrol was further developed and improved (131, 137, 147, 148). Several studies demonstrate that phage have the potential to control plant pathogens, such as Pectobacterium carotovorum ssp. carotovorum (149, 150), Dickeya solani causing soft rot (151), Ralstonia solanacearum causing bacterial wilt (152), Xanthomonas campestris pv. vesicatoria causing bacterial spot (153, 154) and Xylella fastidiosa causing Pierce’s disease (155). In 2005, the first phage based pesticide AgriPhage by Omnilytics was registered by the EPA. The first application of phages as biocontrol agent against fire blight dates back to 1973 (156) and their potential was investigated in different studies in the following years (61–63, 131, 142). The efficiency of phages against E. amylovora was tested on either immature pear fruit (156–158), detached blossoms (62, 63, 159–161) or entire trees (62, 142, 162). The most promising stage to target fire blight is during the blossoming period in spring. In general, apple and pear trees bloom only for a few days or weeks. Intensive spraying of phage biocontrol agents during this time could target E. amylovora in the blossoms and prevent disease outbreak. Hence, testing phage treatments on detached flowers or on trees would generate the most reliable predictions for spray application. Most single phage treatments or phage cocktails tested in blossoms and on trees were able to significantly reduce bacterial cell numbers or lower the severity of disease symptoms. However, it is difficult to compare these studies since settings and treatment procedures are not standardized (163). The amounts of pathogen used for infection ranged from 500 CFU/blossom up to 1012 CFU/ml. Similarly, the used phage titres ranged from 3x102 up to 1x1010 PFU/ml. Furthermore, different multiplicities of infections (0.01-100) were used to treat the infected blossoms. Most of these studies,

21 Introduction however, conclude that phage viability was crucial to generate sufficient pathogen control. Phage viability is therefore regarded as the limiting factor for phage biocontrol of E. amylovora. To enhance efficacy and stability of phage in the harsh environment of the phyllosphere, certain measures can be taken. Supplementing the phages with certain adjuvants could ensure stable phage titres. Different substances were previously tested by Born et al. for their protective ability against UV irradiation. Carrot, beet root and red pepper extracts as well as casein, soy peptone and Tween 80 amongst others mediated UV protection and had no compromising effect on phage infectivity (135). In addition, skim milk, sucrose or corn flour were previously approved to enhance phage stability and treatment potency (135, 153). In the absence of host bacteria, phages are at risk of decomposing. Providing phages with an alternative host bacterium could guarantee continuous infection and replication of these phages. An avirulent or a closely related strain could serve as carrier bacterium (63). Certain studies included the closely related bacterium Pantoea agglomerans into the tested biocontrol agent. P. agglomerans was already tested as antagonist of E. amylovora and can be purchased as Bloomtime. Studies have shown that the combination of phages with P. agglomerans as carrier bacterium can enhance the potency of the treatment (62, 63). Finally, the timing of the application of optimized phage biocontrol agents is crucial and can enhance treatment outcome as observed in other phage treatments (164).

1.2.7. Phages of E. amylovora

The first publication of E. amylovora specific phages was by Billing et al. in 1960 (165). Since then, more than 150 E. amylovora specific phages were identified and many have already been sequenced (7, 62, 159, 161, 166–180). In Switzerland, Born et al. isolated and characterized 24 E. amylovora specific phages (176). The phages L1 (39.3 kbp), M7 (84.7 kbp), S6 (74.7 kbp) and Y2 (56.6 kbp) were subsequently sequenced and further analysed. The podoviruses L1 and S6 are classified as T7-like or N4-like phages, respectively. M7 and Y2 were both recognized as myoviruses belonging to the genera FelixO1-like or GJ1-like phages respectively. All four phages are strictly lytic with a low risk of generalized transduction. The phages have a broad host range and their potential to control E. amylovora as single phage or in combination was tested in vitro. Especially L1+S6 and L1+Y2 were able to reduce the viable cell counts below the detection limit. Further investigations concerning the synergistic effect of L1 and Y2 revealed an enhanced adsorption to and killing of E. amylovora (176). L1 encodes a depolymerase specialized in degrading the capsules surrounding the bacterium (144). The degradation of this protective layer by the L1 depolymerase exposes the bacterial surface and increases the accessibility of host receptors. This can then be exploited by Y2 (144, 181). However, in all cases bacterial regrowth could be observed after 24 hours of incubation, suggesting resistance development against the phage combinations.

22 Introduction

1.3. Aim of the study

Even though phages are an effective alternative to antibiotics, the risk of resistance development must be minimized. Optimal combinations of potent phages can reduce the risk of phage resistance by circumventing bacterial resistance mechanisms. To ensure potency and simultaneously minimize the risk of resistance development, careful composition of phage biocontrol agents must be carried out. This study aimed to optimize phage cocktails by facilitating the process of selecting and integrating potent phages into a phage cocktail against E. amylovora. To generate a broader pool of potent phages for phage cocktails, the phage S2 and the newly isolated phage Bue1 were sequenced and investigated for their potential to be incorporated into a phage biocontrol cocktail. Subsequently, the study focused on host receptor identification of the well-characterized and highly potent phages Bue1, L1, M7, S2, S6 and Y2. A transposon mutagenesis of E. amylovora CFBP1430 revealed genes and structures required for successful phage infection. Phages with diverging host receptors were then incorporated into phage cocktails and efficacy against the pathogen and risk for phage resistance development was evaluated in vitro and on infected apple blossoms. Finally, the study aimed to anticipate the risk of resistance development against multiple phages that could render a phage cocktail futile. Transposon mutagenesis screen and spontaneous resistance experiments were preformed, to identify mutants with multi phage resistance. These experiments helped to reveal enzymes involved in manifold metabolic processes that can influence phage sensitivity. The knowledge gained from this part will help avoiding the development of multi resistant strains and estimating their impact on phage therapy against E. amylovora.

23 Introduction

1.4. References

1. Adeolu M, Alnajar S, Naushad S, Gupta RS. 2016. Genome-based phylogeny and taxonomy of the ‘Enterobacteriales.’ Int J Syst Evol Microbiol 66:5575–5599.

2. Polsinelli I, Borruso L, Caliandro R, Triboli L, Esposito A, Benini S. 2019. A genome- wide analysis of desferrioxamine mediated iron uptake in Erwinia spp. reveals genes exclusive of the Rosaceae infecting strains. Sci Rep 9:2818.

3. Billinge E, Baker LAE, Crosse JE, Garrett CME. 1961. Characteristics of english isolates of Erwinia amylovora (Burrill) Winslow et al. J Appl Bacteriol 24:195–211.

4. Gusberti M, Klemm U, Meier MS, Maurhofer M, Hunger-Glaser I. 2015. Fire blight control: The struggle goes on. A comparison of different fire blight control methods in Switzerland with respect to biosafety, efficacy and durability. Int J Environ Res Public Health 12:11422–47.

5. Bonn WG, van der Zwet T. 2000. Distribution and economic importance of fire blight, p. 37–53. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

6. Vanneste JL. 2000. What is fire blight? Who is Erwinia amylovora? How to control it?, p. 1–6. In Fire Blight. CAB International, Wallingford.

7. Gill JJ, Svircev AM, Smith R, Castle AJ. 2003. Bacteriophages of Erwinia amylovora. Appl Environ Microbiol 69:2133–2138.

8. Momol MT, Aldwinckle HS. 2000. Genetic diversity and host range of Erwinia amylovora, p. 55–72. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

9. Thomson S V. 2000. Epidemiology of fire blight, p. 9–36. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

10. Zhang Y, Bak DD, Heid H, Geider K. 1999. Molecular characterization of a protease secreted by Erwinia amylovora. J Mol Biol 289:1239–1251.

11. Bogs J, Geider K. 2000. Molecular analysis of sucrose metabolism of Erwinia amylovora and influence on bacterial virulence. J Bacteriol 182:5351–8.

12. Expert D, Dellagi A, Kachadourian R. 2000. Iron and fire blight: Role in pathogenicity of desferrioxamine E, the main siderophore of Erwinia amylovora, p. 179–195. In Fire Blight. CAB International, Wallingford.

13. Smits THM, Duffy B. 2011. Genomics of iron acquisition in the plant pathogen Erwinia amylovora: insights in the biosynthetic pathway of the siderophore desferrioxamine E. Arch Microbiol 193:693–699.

14. Oh CS, Beer S V. 2005. Molecular genetics of Erwinia amylovora involved in the development of fire blight. FEMS Microbiol Lett 253:185–192.

15. Boureau T, Siamer S, Perino C, Gaubert S, Patrit O, Degrave A, Fagard M, Chevreau E, Barny MA. 2011. The HrpN effector of Erwinia amylovora, which is involved in type III translocation, contributes directly or indirectly to callose elicitation on apple leaves.

24 Introduction

Mol Plant-Microbe Interact 24:577–584.

16. Bocsanczy AM, Nissinen RM, Oh C-SS, Beer S V. 2008. HrpN of Erwinia amylovora functions in the translocation of DspA/E into plant cells. Mol Plant Pathol 9:425–434.

17. Boureau T, El Maarouf-Bouteau H, Garnier A, Brisset MN, Perino C, Pucheu I, Barny MA. 2006. DspA/E, a type III effector essential for Erwinia amylovora pathogenicity and growth in planta, induces cell death in host apple and nonhost tobacco plants. Mol Plant-Microbe Interact 19:16–24.

18. Wei ZM, Sneath BJ, Beer S V. 1992. Expression of Erwinia amylovora hrp genes in response to environmental stimuli. J Bacteriol 174:1875–82.

19. Wei Z, Kim JF, Beer S V. 2000. Regulation of hrp genes and type III protein secretion in Erwinia amylovora by HrpX/HrpY, a novel two-component system, and HrpS. Mol Plant-Microbe Interact 13:1251–1262.

20. Wei ZM, Beer S V. 1995. hrpl activates Erwinia amylovora hrp gene transcription and is a member of the ECF subfamily of σ factors. J Bacteriol 177:6201–6210.

21. Ordax M, Marco-Noales E, López MM, Biosca EG. 2010. Exopolysaccharides favor the survival of Erwinia amylovora under copper stress through different strategies. Res Microbiol 161:549–555.

22. Koczan JM, McGrath MJ, Zhao Y, Sundin GW. 2009. Contribution of Erwinia amylovora exopolysaccharides amylovoran and levan to biofilm formation: implications in pathogenicity. Phytopathology 99:1237–1244.

23. Bellemann P, Geider K. 1992. Localization of transposon insertions in pathogenicity mutants of Erwinia amylovora and their biochemical characterization. J Gen Microbiol 138:931–940.

24. Koczan JM, Lenneman BR, McGrath MJ, Sundin GW. 2011. Cell surface attachment structures contribute to biofilm formation and xylem colonization by Erwinia amylovora. Appl Environ Microbiol 77:7031–7039.

25. Vrancken K, Holtappels M, Schoofs H, Deckers T, Valcke R. 2013. Pathogenicity and infection strategies of the fire blight pathogen Erwinia amylovora in Rosaceae: State of the art. Microbiology.

26. Maes M, Orye K, Bobev S, Devreese B, Van Beeumen J, De Bruyn A, Busson R, Herdewijn P, Morreel K, Messens E. 2001. Influence of amylovoran production on virulence of Erwinia amylovora and a different amylovoran structure in E. amylovora isolates from Rubus. Eur J Plant Pathol 107:839–844.

27. Koczan JM, McGrath MJ, Zhao Y, Sundin GW. 2009. Contribution of Erwinia amylovora exopolysaccharides amylovoran and levan to biofilm formation: implications in pathogenicity. Phytopathology 99:1237–1244.

28. Castiblanco LF, Sundin GW. 2016. Cellulose production, activated by cyclic di-GMP through BcsA and BcsZ, is a virulence factor and an essential determinant of the three- dimensional architectures of biofilms formed by Erwinia amylovora Ea1189. Mol Plant Pathol 19:90–103.

29. Edmunds AC, Castiblanco LF, Sundin GW, Waters CM. 2013. Cyclic di-GMP

25 Introduction

modulates the disease progression of Erwinia amylovora. J Bacteriol 195:2155–65.

30. Bobrov AG, Kirillina O, Perry RD. 2007. Regulation of biofilm formation in Yersinia pestis. Adv Exp Med Biol 603:201–10.

31. Omadjela O, Narahari A, Strumillo J, Mélida H, Mazur O, Bulone V, Zimmer J. 2013. BcsA and BcsB form the catalytically active core of bacterial cellulose synthase sufficient for in vitro cellulose synthesis. Proc Natl Acad Sci U S A 110:17856–61.

32. Kharadi RR, Castiblanco LF, Waters CM, Sundin GW. 2019. Phosphodiesterase genes regulate amylovoran production, biofilm formation, and virulence in Erwinia amylovora. Appl Environ Microbiol 85:e02233-18.

33. Ahmad I, Lamprokostopoulou A, Le Guyon S, Streck E, Barthel M, Peters V, Hardt W- D, Römling U. 2011. Complex c-di-GMP signaling networks mediate transition between virulence properties and biofilm formation in Salmonella enterica serovar Typhimurium. PLoS One 6:e28351.

34. Bobrov AG, Kirillina O, Ryjenkov DA, Waters CM, Price PA, Fetherston JD, Mack D, Goldman WE, Gomelsky M, Perry RD. 2011. Systematic analysis of cyclic di-GMP signalling enzymes and their role in biofilm formation and virulence in Yersinia pestis. Mol Microbiol 79:533–51.

35. Le Guyon S, Simm R, Rehn M, Römling U. 2015. Dissecting the cyclic di-guanylate monophosphate signalling network regulating motility in Salmonella enterica serovar Typhimurium. Environ Microbiol 17:1310–20.

36. Denning W. 1794. On the decay of apple trees. Trans Soc Promot Agric Arts Manuf 1:219–222.

37. Johnson KB. Fire blight of apple and pear. Plant Heal Instr.

38. Taylor R., Hale C., Gunson F., Marshall J. 2003. Survival of the fire blight pathogen, Erwinia amylovora, in calyxes of apple fruit discarded in an orchard. Crop Prot 22:603– 608.

39. Smith TJ, Pusey PL. 2011. Cougarblight 2010, a significant update of the cougarblight fire blight infection risk model. Acta Hortic 331–336.

40. Geider K. 2000. Exopolysaccharides of Erwinia amylovora: structure, biosynthesis, regulation, role in pathogenicity of amylovoran and levan, p. 117–140. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford, Wallingford.

41. Vanneste JL, Eden-Green S. 2000. Migration of Erwinia amylovora in host plant tissues, p. 73–83. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

42. Aćimović SG, Zeng Q, McGhee GC, Sundin GW, Wise JC. 2015. Control of fire blight (Erwinia amylovora) on apple trees with trunk-injected plant resistance inducers and antibiotics and assessment of induction of pathogenesis-related protein genes. Front Plant Sci 6:16.

43. Schlathölter I, Jänsch M, Flachowsky H, Broggini GAL, Hanke M-V, Patocchi A. 2018. Generation of advanced fire blight-resistant apple (Malus × domestica) selections of the

26 Introduction

fifth generation within 7 years of applying the early flowering approach. Planta 247:1475–1488.

44. Steiner PW. 1992. Predicting apple blossom infections by Erwinia amylovora using the MARYBLYT model. ISHS Acta Hortic 273:139–148.

45. Dewdney MM, Biggs AR, Turechek WW. 2007. A statistical comparison of the blossom blight forecasts of MARYBLYT and Cougarblight with receiver operating characteristic curve analysis. Phytopathology 97:1164–1176.

46. Philion V, Trapman M. 2011. Description and preliminary validation of RIMpro-Erwinia, a new model for fire blight forecast. Acta Hortic 307–317.

47. Psallidas PG, Tsiantos J, Johnson KB, Stockwell VO. 2000. Biological control of fire blight, p. 199–234. In Vanneste, JL (ed.), Fire blight: The disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

48. Martin, H; Woodcok D. 1973. The scientific principles of crop protection.7th edn. Edward Arnold, London.

49. Jones AL, Schnabel EL. 2000. The development of streptomycin-resistant strains of Erwinia amylovora, p. 235–251. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

50. Piqué N, Miñana-Galbis D, Merino S, Tomás JM. 2015. Virulence factors of Erwinia amylovora: A review. Int J Mol Sci 16:12836–54.

51. Russo NL, Burr TJ, Breth DI, Aldwinckle HS. 2008. Isolation of streptomycin-resistant isolates of Erwinia amylovora in New York. Plant Dis 92:714–718.

52. Reininger V., Schöneberg A. HE. 2017. Plant protection field trials against fire blight in Switzerland in 2015. J Plant Pathol 99:131–136.

53. Kunz S, Schmitt A, Haug P. 2008. Field testing of strategies for fire blight control in organic fruit growing.

54. Reininger V., Schöneberg A., Walch B. HE. 2017. Feuerbrand - pflanzenschutzmittelversuch 2017. Schweizer Zeitschrift für Obs und Weinbau 153:8– 11.

55. Bantleon G. 2012. Fire blight: Characterization and control measures. Ph.D. Thesis, University of Hohenheim, Stuttgart, Germany.

56. Johnson KB, Temple TN. 2013. Evaluation of strategies for fire blight control in organic pome fruit without antibiotics. Plant Dis 97:402–409.

57. Jin M, Liu L, Wright SA, Beer S V, Clardy J. 2003. Structural and functional analysis of pantocin A: an antibiotic from Pantoea agglomerans discovered by heterologous expression of cloned genes. Angew Chemie Int Ed 42:2898–2901.

58. Stockwell VO, Johnson KB, Sugar D, Loper JE. 2010. Control of fire blight by Pseudomonas fluorescens A506 and Pantoea vagans C9-1 applied as single strains and mixed inocula. Phytopathology 100:1330–1339.

59. Barbé S, Llop P, Blom J, Cabrefiga J, Goesmann A, Duffy B, Montesinos E, Smits THM,

27 Introduction

Lopez MM. 2013. Complete sequence of Erwinia piriflorinigrans plasmids pEPIR37 and pEPIR5 and role of pEPIR37 in pathogen virulence. Plant Pathol 62:786–798.

60. Johnson KB, Stockwell VO. 1998. Management of fire blight: a case study in microbial ecology. Annu Rev Phytopathol 36:227–248.

61. Svircev AM, Castle AJ, Lehman SM. 2010. Bacteriophages for control of phytopathogens in food production systems, p. 79–96. In Sabour, PM, Griffiths, MW (eds.), Bacteriophages in the control of food- and waterborne pathogens. ASM Press, Washington, DC.

62. Boulé J, Sholberg PL, Lehman SM, O’Gorman DT, Svircev AM. 2011. Isolation and characterization of eight bacteriophages infecting Erwinia amylovora and their potential as biological control agents in British Columbia, Canada. Can J Plant Pathol 33:308– 317.

63. Lehman SM. 2007. Ph.D. thesis: Development of a bacteriophage-based biopesticide for fire blight. Department of Biological Sciences, Brock University, St. Catharines.

64. Kohler S, Maranta A, Sautter C. 2008. Bekämpfung des Feuerbrands in der Schweiz Traditionelle Lösung oder Gentechnik?

65. Duffy B, Schärer HJ, Bünter M, Klay A, Holliger E. 2005. Regulatory measures against Erwinia amylovora in Switzerland. EPPO Bull 35:239–244.

66. Bravin E. 2013. Forschen für einen nachhaltigen Schweizer Obstbau trotz Feuerbrand. Agrar Schweiz 4:440–443.

67. Holliger E, Vogelsanger J, Schoch B, Duffy B, Lussi L, Bünter M. 2007. Das Feuerbrandjahr 2007. Schweizerische Zeitschrift für Obs und Weinbau 24:9–12.

68. FOAG. 2008. Bekämpfung von Feuerbrand: Streptomycin wird befristet zugelassen. Federal Office for Agriculture.

69. Manulis S, Zutra D, Kleitman F, Dror O, David I, Zilberstaine M, Shabi E. 1998. Distribution of streptomycin-resistant strains of Erwinia amylovora in Israel and occurrence of blossom blight in the autumn. Phytoparasitica 26:223–230.

70. Bundesamt für Landwirtschaft. 2016. Kein Einsatz von Streptomycin im Kampf gegen den Feuerbrand. news.admin.ch.

71. Bundesamt für Landwirtschaft. 2018. Wechsel bei der Regelung von Feuerbrand ab 2020.

72. Casjens SR. 2005. Comparative genomics and evolution of the tailed-bacteriophages. Curr Opin Microbiol 8:451–458.

73. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell 108:13–6.

74. Heller K, Loessner MJ, Fieseler L. 2011. Bakteriophagen : Grundlagen, Rolle in Lebensmitteln und Andwendungen zum Nachweis sowie der Kontrolle von Krankheitserregern. Behr, Hamburg.

75. Twort FWW. 1915. An investigation on the nature of ultra-microscopic viruses. Lancet 2:1241–1243.

28 Introduction

76. D’Herelle F. 2007. On an invisible microbe antagonistic toward dysenteric bacilli: brief note by Mr. F. D’Herelle, presented by Mr. Roux. Res Microbiol 158:553–554.

77. Summers WC. 2001. Bacteriophage therapy. Annu Rev Microbiol 55:437–451.

78. Keen EC. 2015. A century of phage research: bacteriophages and the shaping of modern biology. Bioessays 37:6–9.

79. Sanger F, Air GM, Barrell BG, Brown NL, Coulson AR, Fiddes CA, Hutchison CA, Slocombe PM, Smith M. 1977. Nucleotide sequence of bacteriophage phi X174 DNA. Nature 265:687–95.

80. Smith GP. 1985. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 228:1315–7.

81. Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. 2012. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science (80- ) 337:816–821.

82. King AMQ, Lefkowitz EJ, Mushegian AR, Adams MJ, Dutilh BE, Gorbalenya AE, Harrach B, Harrison RL, Junglen S, Knowles NJ, Kropinski AM, Krupovic M, Kuhn JH, Nibert ML, Rubino L, Sabanadzovic S, Sanfaçon H, Siddell SG, Simmonds P, Varsani A, Zerbini FM, Davison AJ. 2018. Changes to taxonomy and the International Code of Virus Classification and Nomenclature ratified by the International Committee on Taxonomy of Viruses (2018). Arch Virol 163:2601–2631.

83. Ackermann HW. 2007. 5500 Phages examined in the electron microscope. Arch Virol 152:227–243.

84. Adriaenssens EM, Wittmann J, Kuhn JH, Turner D, Sullivan MB, Dutilh BE, Jang H Bin, van Zyl LJ, Klumpp J, Lobocka M, Moreno Switt AI, Rumnieks J, Edwards RA, Uchiyama J, Alfenas-Zerbini P, Petty NK, Kropinski AM, Barylski J, Gillis A, Clokie MRC, Prangishvili D, Lavigne R, Aziz RK, Duffy S, Krupovic M, Poranen MM, Knezevic P, Enault F, Tong Y, Oksanen HM, Rodney Brister J. 2018. Taxonomy of prokaryotic viruses: 2017 update from the ICTV Bacterial and Archaeal Viruses Subcommittee. Arch Virol 163:1125–1129.

85. Barylski J, Enault F, Dutilh BE, Schuller MBP, Edwards RA, Gillis A, Klumpp J, Knezevic P, Krupovic M, Kuhn JH, Lavigne R, Oksanen HM, Sullivan MB, Jang H Bin, Simmonds P, Aiewsakun P, Wittmann J, Tolstoy I, Brister JR, Kropinski AM, Adriaenssens EM. 2018. Taxonomy proposal: To create one (1) new family, Herelleviridae, in the order Caudovirales. International Committee on Taxonomy of Viruses (ICTV).

86. Marti R, Zurfluh K, Hagens S, Pianezzi J, Klumpp J, Loessner MJ. 2013. Long tail fibres of the novel broad-host-range T-even bacteriophage S16 specifically recognize Salmonella OmpC. Mol Microbiol 87:818–834.

87. Shin H, Lee JH, Kim H, Choi Y, Heu S, Ryu S. 2012. Receptor diversity and host interaction of bacteriophages infecting Salmonella enterica serovar Typhimurium. PLoS One 7:e43392.

88. Bertozzi Silva J, Storms Z, Sauvageau D. 2016. Host receptors for bacteriophage adsorption. FEMS Microbiol Lett 363:fnw002.

89. Guerrero-Ferreira RC, Viollier PH, Ely B, Poindexter JS, Georgieva M, Jensen GJ,

29 Introduction

Wright ER. 2011. Alternative mechanism for bacteriophage adsorption to the motile bacterium Caulobacter crescentus. Proc Natl Acad Sci U S A 108:9963–8.

90. Fehmel F, Feige U, Niemann H, Stirm S. 1975. Escherichia coli capsule bacteriophages. VII. Bacteriophage 29-host capsular polysaccharide interactions. J Virol 16:591–601.

91. Ackermann HW. 1998. Tailed bacteriophages: the order Caudovirales. Adv Virus Res 51:135–201.

92. Wang IN, Smith DL, Young R. 2000. Holins: The protein clocks of bacteriophage infections. Annu Rev Microbiol 54:799–825.

93. Erez Z, Steinberger-Levy I, Shamir M, Doron S, Stokar-Avihail A, Peleg Y, Melamed S, Leavitt A, Savidor A, Albeck S, Amitai G, Sorek R. 2017. Communication between viruses guides lysis–lysogeny decisions. Nature 541:488–493.

94. Born Y. 2012. Ph. D. thesis: Characterization of bacteriophages for the control and detection of Erwinia amylovora. ETH Zürich.

95. Waldor MK, Mekalanos JJ. 1996. Lysogenic conversion by a filamentous phage encoding cholera toxin. Science (80- ) 272:1910–1914.

96. Plunkett G, Rose DJ, Durfee TJ, Blattner FR, Blattner FR. 1999. Sequence of shiga toxin 2 phage 933W from Escherichia coli O157:H7: Shiga toxin as a phage late-gene product. J Bacteriol 181:1767–78.

97. Canchaya C, Fournous G, Brussow H. 2004. The impact of prophages on bacterial chromosomes. Mol Microbiol 53:9–18.

98. Brussow H, Canchaya C, Hardt WD. 2004. Phages and the evolution of bacterial pathogens: from genomic rearrangements to lysogenic conversion. Microbiol Mol Biol Rev 68:560–602.

99. Touchon M, Moura de Sousa JA, Rocha EP. 2017. Embracing the enemy: the diversification of microbial gene repertoires by phage-mediated horizontal gene transfer. Curr Opin Microbiol 38:66–73.

100. Oliveira L, Tavares P, Alonso JC. 2013. Headful DNA packaging: Bacteriophage SPP1 as a model system. Virus Res 173:247–259.

101. Keen EC, Bliskovsky V V, Malagon F, Baker JD, Prince JS, Klaus JS, Adhya SL. 2017. Novel “superspreader”; bacteriophages promote horizontal gene transfer by transformation. MBio 8:e02115-16.

102. Pedulla ML, Ford ME, Houtz JM, Karthikeyan T, Wadsworth C, Lewis JA, Jacobs-Sera D, Falbo J, Gross J, Pannunzio NR, Brucker W, Kumar V, Kandasamy J, Keenan L, Bardarov S, Kriakov J, Lawrence JG, Jacobs WR, Hendrix RW, Hatfull GF. 2003. Origins of highly mosaic mycobacteriophage genomes. Cell 113:171–182.

103. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistance mechanisms. Nat Rev Microbiol 8:317–327.

104. Seed KD. 2015. Battling phages: How bacteria defend against viral attack. PLoS Pathog 11:e1004847.

30 Introduction

105. Riede I, Eschbach ML. 1986. Evidence that TraT interacts with OmpA of Escherichia coli. FEBS Lett 205:241–5.

106. Pedruzzi I, Rosenbusch JP, Locher KP. 1998. Inactivation in vitro of the Escherichia coli outer membrane protein FhuA by a phage T5-encoded lipoprotein. FEMS Microbiol Lett 168:119–125.

107. Harvey H, Bondy-Denomy J, Marquis H, Sztanko KM, Davidson AR, Burrows LL. 2018. Bacteria defend against phages through type IV pilus glycosylation 1. Nat Microbiol 3:47–52.

108. Forde A, Fitzgerald GF. 2003. Molecular organization of exopolysaccharide (EPS) encoding genes on the lactococcal bacteriophage adsorption blocking plasmid, pCI658. Plasmid 49:130–42.

109. Scholl D, Adhya S, Merril C. 2005. Escherichia coli K1’s capsule is a barrier to bacteriophage T7. Appl Environ Microbiol 71:4872–4874.

110. Destoumieux-Garzón D, Duquesne S, Peduzzi J, Goulard C, Desmadril M, Letellier L, Rebuffat S, Boulanger P. 2005. The iron–siderophore transporter FhuA is the receptor for the antimicrobial peptide microcin J25: role of the microcin Val 11 –Pro 16 β-hairpin region in the recognition mechanism. Biochem J 389:869–876.

111. Moak M, Molineux IJ. 2000. Role of the Gp16 lytic transglycosylase motif in bacteriophage T7 virions at the initiation of infection. Mol Microbiol 37:345–355.

112. Bebeacua C, Lorenzo Fajardo JC, Blangy S, Spinelli S, Bollmann S, Neve H, Cambillau C, Heller KJ. 2013. X-ray structure of a superinfection exclusion lipoprotein from phage TP-J34 and identification of the tape measure protein as its target. Mol Microbiol 89:152–165.

113. Dy RL, Richter C, Salmond GPC, Fineran PC. 2014. Remarkable mechanisms in microbes to resist phage infections. Annu Rev Virol 1:307–331.

114. Goldfarb T, Sberro H, Weinstock E, Cohen O, Doron S, Charpak-Amikam Y, Afik S, Ofir G, Sorek R. 2015. BREX is a novel phage resistance system widespread in microbial genomes. EMBO J 34:169–183.

115. Ofir G, Melamed S, Sberro H, Mukamel Z, Silverman S, Yaakov G, Doron S, Sorek R. 2018. DISARM is a widespread bacterial defence system with broad anti-phage activities. Nat Microbiol 3:90–98.

116. Tock MR, Dryden DT. 2005. The biology of restriction and anti-restriction. Curr Opin Microbiol 8:466–472.

117. Dupuis M-È, Villion M, Magadán AH, Moineau S. 2013. CRISPR-Cas and restriction– modification systems are compatible and increase phage resistance. Nat Commun 4:2087.

118. Moineau S. 1999. Applications of phage resistance in lactic acid bacteria, p. 377–382. In Lactic Acid Bacteria: Genetics, Metabolism and Applications. Springer Netherlands, Dordrecht.

119. Oechslin F. 2018. Resistance development to bacteriophages occurring during bacteriophage therapy. Viruses 10:351.

31 Introduction

120. Koskella B, Lin DM, Buckling A, Thompson JN. 2012. The costs of evolving resistance in heterogeneous parasite environments. Proc R Soc B Biol Sci 279:1896–1903.

121. Smits THM, Rezzonico F, Kamber T, Blom J, Goesmann A, Ishimaru CA, Frey JE, Stockwell VO, Duffy B. 2011. Metabolic versatility and antibacterial metabolite biosynthesis are distinguishing genomic features of the fire blight antagonist Pantoea vagans C9-1. PLoS One 6:e22247.

122. León M, Bastías R. 2015. Virulence reduction in bacteriophage resistant bacteria. Front Microbiol 6:343.

123. Seed KD, Faruque SM, Mekalanos JJ, Calderwood SB, Qadri F, Camilli A. 2012. Phase variable O antigen biosynthetic genes control expression of the major protective antigen and bacteriophage receptor in Vibrio cholerae O1. PLoS Pathog 8:e1002917.

124. van der Woude MW, Bäumler AJ. 2004. Phase and antigenic variation in bacteria. Clin Microbiol Rev 17:581–611.

125. Gencay YE, Sørensen MCH, Wenzel CQ, Szymanski CM, Brøndsted L. 2018. Phase variable expression of a single phage receptor in Campylobacter jejuni NCTC12662 influences sensitivity toward several diverse CPS-dependent phages. Front Microbiol 9:82.

126. Kim M, Ryu S. 2012. Spontaneous and transient defence against bacteriophage by phase-variable glucosylation of O-antigen in Salmonella enterica serovar Typhimurium. Mol Microbiol 86:411–425.

127. Cota I, Blanc-Potard AB, Casadesús J. 2012. STM2209-STM2208 (opvAB): A phase variation locus of Salmonella enterica involved in control of O-antigen chain length. PLoS One 7:e36863.

128. Wright RCT, Friman V-P, Smith MCM, Brockhurst MA. 2018. Cross-resistance is modular in bacteria–phage interactions. PLOS Biol 16:e2006057.

129. Gordillo Altamirano FL, Barr JJ. 2019. Phage therapy in the postantibiotic era. Clin Microbiol Rev 32:e00066-18.

130. Chan BK, Abedon ST, Loc-Carrillo C. 2013. Phage cocktails and the future of phage therapy. Future Microbiol 8:769–783.

131. Svircev A, Roach D, Castle A. 2018. Framing the future with bacteriophages in agriculture. Viruses 10:218.

132. Endersen L, O’Mahony J, Hill C, Ross RP, McAuliffe O, Coffey A. 2014. Phage therapy in the food industry. Annu Rev Food Sci Technol 5:327–349.

133. Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JA, Momol MT, Jones JB, Vallad GE. 2012. Soil-based systemic delivery and phyllosphere in vivo propagation of bacteriophages: Two possible strategies for improving bacteriophage persistence for plant disease control. Bacteriophage 2:215–224.

134. Jones JB, Vallad GE, Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JC, Momol MT. 2012. Considerations for using bacteriophages for plant disease control. Bacteriophage 2:208–214.

32 Introduction

135. Born Y, Bosshard L, Duffy B, Loessner MJ, Fieseler L. Protection of Erwinia amylovora bacteriophage Y2 from UV-induced damage by natural compounds. Bacteriophage 5.

136. Loc-Carrillo C, Abedon ST. 2011. Pros and cons of phage therapy. Bacteriophage 1:111–114.

137. Buttimer C, McAuliffe O, Ross RP, Hill C, O’Mahony J, Coffey A, O’Mahony J, Coffey A. 2017. Bacteriophages and bacterial plant diseases. Front Microbiol 8:34.

138. Hagens S, Loessner MJ. 2010. Bacteriophage for biocontrol of foodborne pathogens: calculations and considerations. Curr Pharm Biotechnol 11:58–68.

139. Monteiro R, Pires DP, Costa AR, Azeredo J. 2019. Phage therapy: Going temperate? Trends Microbiol 27:368–378.

140. Samson JE, Magadán AH, Sabri M, Moineau S. 2013. Revenge of the phages: defeating bacterial defences. Nat Rev Microbiol 11:675–687.

141. Chadha P, Katare OP, Chhibber S. 2016. In vivo efficacy of single phage versus phage cocktail in resolving burn wound infection in BALB/c mice. Microb Pathog 99:68–77.

142. Schnabel EL, Fernando WGD, Meyer MP, Jones AL, Jackson LE. 1999. Bacteriophage of Erwinia amylovora and their potential for biocontrol. ISHS Acta Hortic 489:649–653.

143. Naghizadeh M, Karimi Torshizi MA, Rahimi S, Dalgaard TS. 2018. Synergistic effect of phage therapy using a cocktail rather than a single phage in the control of severe colibacillosis in quails. Poult Sci 98:653–663.

144. Born Y, Fieseler L, Klumpp J, Eugster MR, Zurfluh K, Duffy B, Loessner MJ. 2014. The tail-associated depolymerase of Erwinia amylovora phage L1 mediates host cell adsorption and enzymatic capsule removal, which can enhance infection by other phage. Environ Microbiol 16:2168–2180.

145. Schmerer M, Molineux IJ, Bull JJ. 2014. Synergy as a rationale for phage therapy using phage cocktails. PeerJ 2:e590.

146. Adriaenssens EM, Ackermann H-W, Anany H, Blasdel B, Connerton IF, Goulding D, Griffiths MW, Hooton SP, Kutter EM, Kropinski AM, Lee J-H, Maes M, Pickard D, Ryu S, Sepehrizadeh Z, Shahrbabak SS, Toribio AL, Lavigne R. 2012. A suggested new bacteriophage genus: Viunalikevirus. Arch Virol 157:2035–46.

147. Balogh B, Jones JB, Iriarte FB, Momol MT. 2010. Phage therapy for plant disease control. Curr Pharm Biotechnol 11:48–57.

148. Nagy JK, Kiraly L, Ildiko S. 2012. Phage therapy for plant disease control with a focus on fire blight. Cent Eur J Biol 7:1–12.

149. Czajkowski R, Ozymko Z, de Jager V, Siwinska J, Smolarska A, Ossowicki A, Narajczyk M, Lojkowska E. 2015. Genomic, proteomic and morphological characterization of two novel broad host lytic bacteriophages ΦPD10.3 and ΦPD23.1 infecting pectinolytic Pectobacterium spp. and Dickeya spp. PLoS One 10:e0119812.

150. Lim J-A, Jee S, Lee DH, Roh E, Jung K, Oh C, Heu S. 2013. Biocontrol of Pectobacterium carotovorum subsp. carotovorum using bacteriophage PP1. J Microbiol Biotechnol 23:1147–1153.

33 Introduction

151. Adriaenssens EM, Van Vaerenbergh J, Vandenheuvel D, Dunon V, Ceyssens P-J, De Proft M, Kropinski AM, Noben J-P, Maes M, Lavigne R. 2012. T4-related bacteriophage LIMEstone isolates for the control of soft rot on potato caused by Dickeya solani. PLoS One 7:e33227.

152. Fujiwara A, Fujisawa M, Hamasaki R, Kawasaki T, Fujie M, Yamada T. 2011. Biocontrol of Ralstonia solanacearum by treatment with lytic bacteriophages. Appl Environ Microbiol 77:4155–4162.

153. Balogh B, Jones JB, Momol MT, Olson SM, Obradovic A, King P, Jackson LE. 2003. Improved efficacy of newly formulated bacteriophages for management of bacterial spot on tomato. Plant Dis 87:949–954.

154. Obradovic A, Jones JB. 2004. Management of tomato bacterial spot in the field by foliar applications of bacteriophages and SAR inducers. Plant Dis 88:736–740.

155. Das M, Bhowmick TS, Ahern SJ, Young R, Gonzalez CF. 2015. Control of pierce’s disease by phage. PLoS One 10:e0128902.

156. Erskine JM. 1973. Characteristics of Erwinia amylovora bacteriophage and its possible role in the epidemiology of fire blight. Can J Microbiol 19:837–845.

157. Müller I, Lurz R, Kube M, Quedenau C, Jelkmann W, Geider K. 2011. Molecular and physiological properties of bacteriophages from North America and Germany affecting the fire blight pathogen Erwinia amylovora. Microb Biotechnol 4:735–745.

158. Gill JJ. 2000. Ph.D thesis: Bacteriophages of Erwinia amylovora and their potential use in biological control. Brock University, St Catharines, Ontario.

159. Schwarczinger I, Kolozsváriné Nagy J, Künstler A, Szabó L, Geider K, Király L, Pogány M. 2017. Characterization of Myoviridae and Podoviridae family bacteriophages of Erwinia amylovora from Hungary - potential of application in biological control of fire blight. Eur J Plant Pathol 149:1–14.

160. Schwarczinger I, Kiss E, Süle S, Tóth M, Hevesi M. 2011. Control of fire blight by bacteriophages on apple flowers. ISHS Acta Hortic 896:457–462.

161. Muller I, Kube M, Reinhardt R, Jelkmann W, Geider K. 2011. Complete genome sequences of three Erwinia amylovora phages isolated in north america and a bacteriophage induced from an Erwinia tasmaniensis strain. J Bacteriol 193:795–796.

162. Schoofs H, Vandebroek K, Pierrard A, Thonart P, Lepoivre P, Beaudry T, Deckers T. 2002. Bacteriocin Serratine-P as a biological tool in the control of fire blight Erwinia amylovora. Meded Rijksuniv te Gent Fak van Landbouwkd en Toegepaste Biol Wet 67:361–8.

163. Nagy JK, Kiraly L, Ildiko S, Király L, Schwarczinger I. 2012. Phage therapy for plant disease control with a focus on fire blight. Cent Eur J Biol 7:1–12.

164. Flaherty JE, Jones JB, Harbaugh BK, Somodi GC, Jackson LE. 2000. Control of bacterial spot on tomato in the greenhouse and field with h-mutant bacteriophages. HortScience 35:882–884.

165. Billing E, Crosse JE, Garrett CME. 1960. Laboratory diagnosis of fire blight and bacterial blossom blight of pear. Plant Pathol 9:19–25.

34 Introduction

166. Evans TJ, Trauner A, Komitopoulou E, Salmond GP. 2010. Exploitation of a new flagellatropic phage of Erwinia for positive selection of bacterial mutants attenuated in plant virulence: towards phage therapy. J Appl Microbiol 108:676–685.

167. Lehman SM, Kropinski AM, Castle AJ, Svircev AM. 2009. Complete genome of the broad-host-range Erwinia amylovora phage ΦEa21-4 and its relationship to Salmonella phage Felix O1. Appl Environ Microbiol 75:2139–2147.

168. Sharma R, Berg JA, Beatty NJ, Choi MC, Cowger AE, Cozzens BJR, Duncan SG, Fajardo CP, Ferguson HP, Galbraith T, Herring JA, Hoj TR, Durrant JL, Hyde JR, Jensen GL, Ke SY, Killpack S, Kruger JL, Lawrence EEK, Nwosu IO, Tam TC, Thompson DW, Tueller JA, Ward MEH, Webb CJ, Wood ME, Yeates EL, Baltrus DA, Breakwell DP, Hope S, Grose JH. 2018. Genome sequences of nine Erwinia amylovora bacteriophages. Microbiol Resour Announc 7:e00944-18.

169. Ritchie DF, Klos EJ. 1977. Isolation of Erwinia amylovora bacteriophage from aerial parts of apple trees. Phytopathology 67:101–104.

170. Ritchie DF, Klos EJ. 1979. Some properties of Erwinia amylovora bacteriophages. Phytopathology 69:1078–1083.

171. Lagonenko AL, Sadovskaya O, Valentovich LN, Evtushenkov AN. 2015. Characterization of a new ViI-like Erwinia amylovora bacteriophage phiEa2809. FEMS Microbiol Lett 362.

172. Meczker K, Dömötör D, Vass J, Rákhely G, Schneider G, Kovács T. 2014. The genome of the Erwinia amylovora phage phiEaH1 reveals greater diversity and broadens the applicability of phages for the treatment of fire blight. FEMS Microbiol Lett 350:25–27.

173. Vandenbergh PA, Wright AM, Vidaver AK. 1985. Partial purification and characterization of a polysaccharide depolymerase associated with phage infected Erwinia amylovora. Appl Environ Microbiol 49:994–996.

174. Dömötör D, Becságh P, Rákhely G, Schneider G, Kovács T. 2012. Complete genomic sequence of Erwinia amylovora phage phiEaH2. J Virol 86:10899.

175. Schnabel EL, Jones AL. 2001. Isolation and characterization of five Erwinia amylovora bacteriophages and assessment of phage resistance in strains of Erwinia amylovora. Appl Environ Microbiol 67:59–64.

176. Born Y, Fieseler L, Marazzi J, Lurz R, Duffy B, Loessner MJ. 2011. Novel virulent and broad-host-range Erwinia amylovora bacteriophages reveal a high degree of mosaicism and a relationship to Enterobacteriaceae phages. Appl Environ Microbiol 77:5945– 5954.

177. Knecht LE, Born Y, Pothier JF, Loessner MJ, Fieseler L. 2018. Complete genome sequences of Erwinia amylovora phages vB_EamP-S2 and vB_EamM-Bue1. Microbiol Resour Announc 7:e00891-18.

178. Buttimer C, Born Y, Lucid A, Loessner MJ, Fieseler L, Coffey A. 2018. Erwinia amylovora phage vB_EamM_Y3 represents another lineage of hairy Myoviridae. Res Microbiol 169:505–514.

179. Esplin IND, Berg JA, Sharma R, Allen RC, Arens DK, Ashcroft CR, Bairett SR, Beatty NJ, Bickmore M, Bloomfield TJ, Brady TS, Bybee RN, Carter JL, Choi MC, Duncan S,

35 Introduction

Fajardo CP, Foy BB, Fuhriman DA, Gibby PD, Grossarth SE, Harbaugh K, Harris N, Hilton JA, Hurst E, Hyde JR, Ingersoll K, Jacobson CM, James BD, Jarvis TM, Jaen- Anieves D, Jensen GL, Knabe BK, Kruger JL, Merrill BD, Pape JA, Payne Anderson AM, Payne DE, Peck MD, Pollock S V, Putnam MJ, Ransom EK, Ririe DB, Robinson DM, Rogers SL, Russell KA, Schoenhals JE, Shurtleff CA, Simister AR, Smith HG, Stephenson MB, Staley LA, Stettler JM, Stratton ML, Tateoka OB, Tatlow PJ, Taylor AS, Thompson SE, Townsend MH, Thurgood TL, Usher BK, Whitley K V, Ward AT, Ward MEH, Webb CJ, Wienclaw TM, Williamson TL, Wells MJ, Wright CK, Breakwell DP, Hope S, Grose JH. 2017. Genome sequences of 19 novel Erwinia amylovora bacteriophages. Genome Announc 16:e00931-17.

180. Park J, Lee GM, Kim D, Park DH, Oh C-S. 2018. Characterization of the lytic bacteriophage phiEaP-8 effective against both Erwinia amylovora and Erwinia pyrifoliae causing severe diseases in apple and pear. plant Pathol J 34:445–450.

181. Born Y, Fieseler L, Thöny V, Leimer N, Duffy B, Loessner MJ. 2017. Engineering of bacteriophages Y2::dpoL1-C and Y2::luxAB for efficient control and rapid detection of the fire blight pathogen, Erwinia amylovora. Appl Environ Microbiol 83:e00341-17.

36 Manuscripts

2. Manuscripts

Manuscript I: Receptor identification for phage cocktail composition

Manuscript II: Phage infection of E. amylovora requires cellulose

Manuscript III: The role of topB1, rfaE and pgm in phage resistance

Manuscript IV: Y2 resistance affects phage infectivity

37

Receptor identification for phage cocktail composition

2.1. Manuscript I: Receptor identification for phage cocktail composition

Screening of host receptors for effective and sustainable phage cocktail composition against the fire blight pathogen Erwinia amylovora

Leandra E. Knecht1,2, Jules Peter2, Marina Mahler2, Katja Felder2, Yannick Born1, David Vinzent1, Jonas Hofmänner1, Cosima Pelludat3, Martin J. Loessner2, Lars Fieseler1

1 Institute of Food and Beverage Innovation; Zurich University of Applied Sciences; Wädenswil, Switzerland. 2 Institute of Food; Nutrition and Health; ETH Zurich; Zürich, Switzerland. 3 Agroscope, Plant Pathology and Zoology in fruit and vegetable production, Wädenswil, Switzerland

*Correspondence: Lars Fieseler, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Einsiedlerstrasse 31, Wädenswil, Switzerland. Tel: +41 58 934 54 07; e-mail: [email protected]

39 Manuscript I

Abstract Bacteriophages are a promising alternative to antibiotics in the war against bacteria. These natural predators of bacteria recognize their hosts through specific surface receptors. To circumvent resistance development, combinations of phages targeting different receptors could be applied. Such a mixture could hinder the rise of phage insensitive bacteria and prolong the therapeutic effect. Hence, identifying and incorporating potent phages into an effective cockatil is crucial. A high throughput screening assay to identify phage receptors was therefore established. A Tn5 transposon mutant library of the plant pathogen Erwinia amylovora, the causative agent of fire blight, was constructed. Library screening was performed for the phages Bue1, L1, M7, S2, S6 and Y2 and transposon insertion sites were determined for phage resistant mutants. The screen revealed LPS (Bue1, Y2), amylovoran (L1, S2) and bacterial cellulose or the cellulose synthase complex (M7, S6) as most probable receptors. According to these results, different phage combinations were tested for their potential as biocontrol agents against E. amylovora in vitro and in blossoms. The most potent treatments in vitro were observed to be M7 or S6 alone, most combinations with Bue1 and certain combinations with S2. The most promising treatments of fire blight in blossoms were L1, M7+S6 and Bue1+S2+S6. The prediction of the treatment efficacy in blossoms was observed to be more complex and should be further investigated. The most effective phage treatments should be validated on blossoms and entire plants. In addition, modifications of the buffer composition and the supplementation of UV-protective substances could enhance the phage biocontrol formulation.

40 Receptor identification for phage cocktail composition

Introduction Bacteriophages (phages) have the potential to be a potent alternative to antibiotics (1). Phages are viruses exclusively infecting bacteria. They recognize potential host cells by specific receptors on the bacterial surface. These receptors can be proteins, sugar molecules or cell surface structures located either in the outer membrane (2), in slime layers or capsules (3), or in appendages such as pili (4) and flagella (5, 6). After adsorption to the host, phages inject their DNA and hijack the bacterial metabolism. Phage DNA is replicated and new phage particles are assembled before lysis of the host occurs. Phages have several advantages over antibiotics (7). Most phages have a narrow host range, where only certain bacteria will be recognized, while other potentially beneficial bacteria are left unharmed. Phages can be isolated from environmental samples and high phage titres can be achieved in the lab. Since phages use their host cells for replication, phage counts increase in the presence of host bacteria. In the absence of host cells, phage numbers remain stable or start to drop due to e.g. UV irradiation (8, 9). Although bacteria can develop phage resistance, phages are, in contrast to antibiotics, adaptable and have the potential to co-evolve with their host bacteria. In addition, phages were found to be environmentally safe and do not pose side effects (1, 10). As most common entities on the planet, phages outnumber bacteria by approximately a tenfold (11, 12). This vast number ensures specific phages possibly against any bacteria. The control of Erwinia amylovora is currently in urgent need for effective treatment. The Gram- negative bacterium is the causative agent of fire blight (13, 14). This plant disease affects Rosaceae family members and was classified as one of the ten most devastating plant pathogens in crop production (15, 16). Amylovoran production and biofilm formation as well as a type three secretion system with effector proteins are major virulence factors that contribute to the pathogenicity of E. amylovora (17). Infection of susceptible plants generally occurs in the blossoms where the pathogen replicates on the stigma. When the bacteria have reached the critical number of 105-106 CFU/flower, infection through the nectaries is induced (18, 19). In the plant tissue, E. amylovora starts to grow as biofilm and eventually blocks water flow in the xylem (20, 21). Disease symptoms such as swelling of tissue and ooze droplets become visible. Eventually infected plant parts start to wilt and appear blackened, hence the name fire blight (22). When the pathogen has reached the plant tissue, effective elimination of the disease can only be established by generously pruning infected tissue or eliminating the entire plant. The disease is highly infectious and can spread rapidly in orchards by insects, rain and wind as carrier (14, 23). Time consuming and cost intensive sanitation of orchards is essential to prevent disease outbreak or spread. The preventive treatment of susceptible plants during the blossoming season with antibiotics such as streptomycin is the most effective method to ensure plant health. High antibiotic levels in honey, appearance of antibiotic resistant bacteria and public health concerns are problematic (17, 24). The use of streptomycin

41 Manuscript I is consequentially banned in increasing numbers of countries (25). Phages could offer an effective remedy against the fatal disease and their potential is already under investigation (26–28). Although phages have the ability to adapt and co-evolve with their host bacteria, the development of phage resistance must be avoided to ensure treatment effectivity. Combinations of phages with complementing host ranges and different infection strategies can enhance the treatment (29). Certain phage mixtures were observed to exhibit synergistic effects rendering the combination more potent (30, 31). A previous study showed that the E. amylovora specific phages L1 and Y2 have synergistic effects when combined against the pathogen (32). L1 encodes a depolymerase, an enzyme able to break down exopolysaccharide (EPS) molecules surrounding the bacteria. The degradation of this protective layer exposes the bacterial surface and increases accessibility to surface receptors, which can be exploited by Y2. Bacteria can prevent phage attack by modifying the specific receptor recognized by the phage or altering its accessibility by covering it. By combining phages that target different host receptors, the risk of simultaneous receptor alterations is lowered. The identification of these receptors is therefore crucial to optimally combine different phages into a potent cocktail and minimize the risk of resistance development. In this study, an easy to handle high throughput screen using a Tn5 transposon mutagenesis library was established to identify phage receptors. Six E. amylovora specific phages were isolated previously from Swiss apple and pear orchards (33, 34). The phages can be classified as Ackermannviridae (Bue1), Myoviridae (M7 and Y2) or Podoviridae (L1, S2, S6) and belong to different genera. Bue1 was identified as Vi1-like, L1 as T7-like, M7 as Felix-O1, S2 as SP6- like, S6 as N4-like and Y2 as GJ1-like phage. The six phages have a broad and specific infection range and combinations can infect all E. amylovora strains tested. The risk of generalized transduction was regarded as minimal for all six phages (33, 34). The six phages have the potential to be used as phage biocontrol agent against the fire blight pathogen. The results generated from the screen facilitate the combination of phages that rely on different receptors for successful host infection. Selected phage combinations were generated and their potential as biocontrol agent against fire blight was tested in vitro and in blossoms.

42 Receptor identification for phage cocktail composition

Materials and Methods

Bacterial strains and culture conditions E. amylovora strains were cultivated on LB agar at 28°C. Kanamycin (50 µg/ml) was added if required.

Soft agar overlay and propagation of phages Phages were propagated using the soft agar overlay method (35). To generate semi confluent lysed plates, molten LB+ soft agar (LB Bouillon, 4 g/l agar, 2 mM MgSO4, 10 mM CaCl2) was supplemented with 90 µl bacterial overnight suspension and 10 µl diluted bacteriophage and spread evenly onto LB plates. After overnight incubation, 5 ml/plate SM buffer (100 mM NaCl,

8 mM MgSO4, 50 mM Tris-Cl, pH 7.4) were added and plates were incubated for 5 h at room temperature (RT) with shaking. The supernatant was supplemented with 0.5 M NaCl. The bacteriophage suspension was treated with polyethylenglycol (10% w/v PEG 8,000, ice bath overnight), purified (CsCl density gradient) (36) and dialyzed against SM buffer. The phage Y2 generated higher phage titres when the phage-PEG was resuspended in SM buffer and incubated at room temperature for 1 h. PEG was removed by centrifugation for 10 min, 5,000 xg at RT. The supernatant containing the phages was sterile-filtered (0.22 µm filter).

Transposon Mutagenesis An overnight culture of E. amylovora CFBP1430 was diluted 1:1,000 in fresh medium and incubated with shaking until an OD600 of 0.5 was reached. Cells were incubated for 30 min on ice. The cells were washed three times with ice cold 5% (v/v) glycerol before resuspending in 10% (v/v) glycerol. Electrocompetent cells were kept at -80°C until further use. To generate the transposon mutants, 40 µl of electrocompetent cells were supplemented with 1 µl of a transposome (EZ-Tn5 Tnp; Epicenter, Madison, USA) and electroporated (25 V, 25 µF, 200 Ω). Cells were immediately rescued with 4 ml SOC (2% w/v tryptone, 0.5% w/v yeast extract, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM MgSO4 and 20 mM glucose) (37) and incubated at 28°C with shaking for 3-4 h before a dilution series was plated on LB kanamycin plates. A total of 6768 single colonies were picked and stored in 96-well plates filled with LB supplemented with 10% glycerol at -86°C.

43 Manuscript I

Screening of library Phage resistant mutants were identified using a high throughput screen schematically shown in Figure 1. Overnight cultures of the transposon library were prepared as starter culture using a 96-well replica stamp. The next day, molten LB+ or LC soft agar was supplemented with kanamycin (25 µg/ml) and 5% glycerol if required. Two replicates of each plate were generated: one supplemented with the phages Bue1, L1, M7, S2, S6 or Y2 the other plate without phages in the soft agar. The soft agar was distributed into a 96-well flat bottom plate, 200 µl per well. Immediately after distribution, the bacterial starter culture was transferred into the prepared soft agar plate using a 96-well replica stamp. The plates were incubated overnight. The Phage Concentration Agar Glycerin Temp. Incubation 8 optical density of each well was Bue1 10 PFU/ml LB+ - 28°C 16 h L1 107 PFU/ml LC 5% 28°C 19 h measured using a Synergy H1 M7 106 PFU/ml LB+ 5% 28°C 17-20 h Multi-Mode Microplate Reader S2 107 PFU/ml LC 5% 28°C 19 h 5 (BioTek Instruments Inc). To S6 10 PFU/ml LB+ 5% 25°C 14-17 h Y2 105 PFU/ml LB+ - 28°C 14 h identify possibly resistant mutants, the difference in optical Figure 1 Schematic overview of the screening process (top). densities of the grown bacterial Adjusted screening settings for the phages Bue1, L1, M7, S2, S6 and Y2 (bottom). cultures inoculated with and without phages was calculated. Phage insensitive mutants grow similarly in absence and presence of phages resulting in a low difference. Mutants remaining sensitive to phages grow weaker in the presence of the phages; therefore, the calculated difference in optical density is larger. Transposon mutants, which were able to grow in the presence of phages to comparable

44 Receptor identification for phage cocktail composition optical densities as in the absence of phages, were validated in soft agar overlays for plaque formation. This method also allowed the detection of resistant mutants with low growth capacities.

Identification of Tn5 insertion sites in the E. amylovora genomes Arbitrary-primed PCR as described by Das et al. (2005) was performed (38). Arbitrary primers Arb-P1 (5’-GGC CAC GCG TGC ACT AGT CAN NNN NNN NNN GCT CG-3’), Arb-P2 (5’- GGC CAC GCG TGC ACT AGT CAN NNN NNN NNN GAC TC-3’) and Arb-P3 (5’-GGC CAC GCG TGC ACT AGT CAN NNN NNN NNN GAT AC-3’) were paired with the first nested primer (5’-ACC TAC AAC AAA GCT CTC ATC AAC C-3’). Using the purified PCR product as template, the anchor primer (5’-GGC CAC GCG TGC ACT AGT CA-3’) and the second nested primer (5’-TTC AGG GTT GAG ATG TGT ATA AGA GAC AG-3’) amplified a shorter segment, which was purified and sequenced (Microsynth, Switzerland). The obtained sequences were compared to the E. amylovora CFBP1430 genome (accession number FN434113) to locate the transposon insertion site.

In vitro infection assays Overnight cultures of E. amylovora CFBP1430 were washed twice in sterile SM buffer and

7 OD600 was adjusted to reach 10 CFU/ml. Subsequently, 20 µl of the washed cells were transferred to 1960 µl LB+ broth and supplemented with either 20 µl sterile SM buffer or with 20 µl bacteriophage with a concentration of 1010 PFU/ml. For combinations of bacteriophages, 10 µl of each bacteriophage were added to the bacteria. The mixtures were then added into

96-well flat bottom plates and incubated at 25°C with double orbital shaking for 30 h. OD600 measurements were carried out every half hour.

Detached flower assay A detached flower assay using fresh Golden Delicious blossoms from two-year-old apple trees was carried out to monitor efficacy of different phage treatments against E. amylovora (39). Racks were cleaned and autoclaved before the experiment and 24 wells per rack were filled with 2 ml H2O. The wells were sealed with scotch tape, which was perforated using a syringe. Blossom stems were freshly cut to ensure water uptake before being transferred through the scotch tape. Bacteria grown overnight on plates were scratched off and carefully resuspended

7 in PBS. OD600 was adjusted to 1.0 and a 1:50 dilution performed (approx.10 CFU/ml). The blossoms were inoculated by pipetting 20 µl bacterial suspension or PBS directly onto the receptacle and allowed to soak up the liquid. Afterwards, different treatments of phages or PBS as control were applied to the blossoms. In case of the phage treatments, 5 µl of phage suspension at different concentrations between 106 and 108 PFU/ml in total were added directly into the receptacle. Each storage box (5 l) was laid out with paper towels and filled

45 Manuscript I

with three racks. To ensure humidity, 100 ml H2O were added per box. The blossoms were stored at 26°C for 4-5 days. The read out was performed according to an adjusted rating system (40). Healthy blossoms without disease symptoms were classified as Grade 1. Visible symptoms on the blossom (browning of the calix) were referred to as Grade 2. Blossoms with infection spreading from the calix to the stipe of the blossoms corresponded to Grade 3.

46 Receptor identification for phage cocktail composition

Results Identification of phage resistant mutants was carried out using a Tn5 transposon mutant library of E. amylovora CFBP1430. Screening for phage resistant mutants was performed for 6 different phages (Bue1, L1, M7, S2, S6, Y2). Phage resistant mutants were validated and insertion sites were determined.

Bacterial cellulose synthase is important for M7 and S6 infection Table 1 summarizes M7 and S6 resistance mediating genes. One operon in particular was observed to be involved in M7 and S6 resistance. The genes bcsA, bcsB, bcsC, celA3, EAMY3608 and wssA are all located in the bcs operon. With the exception of bcsD, all genes in this operon were identified by the screen. This operon encodes the bacterial cellulose synthase. This complex is required to regulate, produce and secrete bacterial cellulose a carbohydrate polymer involved in biofilm formation (41). In contrast to M7, S6 is also relying on the gene celA3, another gene of the bcs operon, for infection. Additionally, the gene lon was identified to play a role in S6 resistance. Two genes unrelated to the bacterial cellulose synthase complex but associated with LPS production and secretion were identified to mediate M7 and S6 resistance. This suggests LPS involvement for M7 and S6 adsorption and infection. However, only a fraction of all waaL (M7: 4, S6: 4 out of 19) and wabM mutants (S6: 2 out of 18) were observed to be M7 and S6 resistant.

Table 1 Genes with Tn5 insertions mediating M7 or S6 resistance, their predicted gene product and function. Total hits indicate how many mutants with independent transposon insertions in this particular gene were detected overall the 6 independent screenings.

Total Gene Locus tag M7 S6 Product Function hits Glycosyltransferases, probably bcsA EAMY_3606 6 8 8 cellulose synthase catalytic subunit involved in cell wall biogenesis bcsB EAMY_3605 4 5 5 cellulose synthase regulator protein bcsC EAMY_3604 3 3 3 cellulose synthase operon protein C FOG: TPR repeat celA3 EAMY_3602 - 2 2 endoglucanase precursor EAMY hypothetical protein predicted by EAMY_3608 7 8 8 _3608 Glimmer/Critica hmsT EAMY_3035 1 1 1 Uncharacterized protein yhcK FOG: GGDEF domain DNA-binding ATP-dependent ATP-dependent Lon protease, lon EAMY_0985 - 1 1 protease La bacterial type pgm EAMY_1153 1 1 1 Phosphoglucomutase ADP-heptose synthase, bifunctional rfaE EAMY_0426 1 1 1 ADP-heptose synthase sugar kinase/adenylyltransferase topB1 EAMY_1960 7 8 8 DNA topoisomerase III* Topoisomerase IA Lipid A core - O-antigen ligase and waaL EAMY_0091 4 4 19 O-antigen ligase related enzymes Predicted glycosyl transferase, Glycosyltransferases involved in cell wabM EAMY_0090 - 2 18 family 2 wall biogenesis ATPases involved in chromosome wssA EAMY_3607 4 4 4 Uncharacterized protein yhjQ partitioning

47 Manuscript I

L1 and S2 rely on amylovoran for host recognition Disruptions in genes listed in Table 2 render the mutants L1 or S2 resistant. All ams genes belong to the amylovoran operon in E. amylovora. The operon is responsible for amylovoran production, a major virulence factor in the pathogen (42). Only three genes in this operon were not targeted by the transposon, or could not be identified to mediate phage resistance by the screen.

Table 2 Genes with Tn5 insertions mediating L1 or S2 resistance, their predicted gene product and function. Total hits indicate how many mutants with independent transposon insertions in this particular gene were detected overall the 6 independent screenings.

Total Gene Locus tag L1 S2 Product Function hits Uncharacterized protein involved in amsA EAMY_2250 1 1 1 putative tyrosine-protein kinase exopolysaccharide biosynthesis Glycosyltransferases involved in cell amsB EAMY_2249 12 12 12 glycosyltransferase amsB wall biogenesis Exopolysaccharide biosynthesis amsC EAMY_2248 11 11 11 protein amsD EAMY_2247 11 11 11 glycosyltransferase amsD Glycosyltransferase Glycosyltransferases involved in cell amsE EAMY_2246 6 6 6 putative glycosyltransferase wall biogenesis Amylovoran export outer Periplasmic protein involved in amsH EAMY_2252 2 2 2 membrane protein amsH polysaccharide export precursor Exopolysaccharide biosynthesis amsJ EAMY_2244 5 5 5 Uncharacterized conserved protein protein amsJ Exopolysaccharide biosynthesis amsK EAMY_2243 1 1 1 Glycosyltransferase glycosyl transferase Membrane protein involved in the amsL EAMY_2242 3 3 3 Flippase export of O-antigen and teichoic acid aroE EAMY_3345 1 1 1 Shikimate 5-dehydrogenase

galE EAMY_2240 2 2 2 UDP-glucose 4-epimerase delta(2)- tRNA delta(2)- miaA EAMY_3158 1 1 1 isopentenylpyrophosphate tRNA- isopentenylpyrophosphate adenosine transferase transferase pgm EAMY_1153 1 1 1 Phosphoglucomutase

prlC EAMY_3536 1 1 1 Zn-dependent oligopeptidase Zn-dependent oligopeptidases Response regulator containing a activator of exoploysaccharide rcsB3 EAMY_2342 1 1 1 CheY-like receiver domain and an synthesis HTH DNA-binding domain ADP-heptose synthase, bifunctional rfaE EAMY_0426 1 1 1 ADP-heptose synthase sugar kinase/adenylyltransferase Ribosomal protein L25 (general stress rplY EAMY_2320 1 1 1 50S ribosomal protein L25 protein Ctc) thdF EAMY_3683 1 1 1 tRNA modification GTPase trmE Predicted GTPase

topB1 EAMY_1960 1 - 8 DNA topoisomerase III* Topoisomerase IA Uncharacterized protein yceD Predicted metal-binding, possibly yceD EAMY_1469 1 1 1 (G30K) nucleic acid-binding protein

48 Receptor identification for phage cocktail composition

The same genes were identified to also mediate S2 resistance. However, S2 infectivity was not affected by the disruption of the gene topB1. Even though L1 infectivity was affected by disruption of topB1, only one out of the eight identified topB1 mutants was identified to affect L1. None of the other disrupted genes are annotated to encode a possible outer membrane protein or extracellular structure, indicating amylovoran as main identification structure.

Bue1 and Y2 potentially require LPS structures for host recognition The phage Bue1 was unable to infect mutants with insertions in genes listed in Table 3. The majority of these genes are clearly associated with LPS-synthesis and -export. The other targeted genes are almost exclusively annotated as glycosyltransferases or can be linked to glycosylation steps. The screen revealed that all Bue1 resistant strains also mediate Y2 resistance. Additionally to the genes involved in mediating Bue1 and Y2 resistance, the genes EAMY_0795, EAMY_1638, gtrB1, hslU, rfaH, rfbB3, yhjL and yjdB could be identified to be exclusively involved in Y2 resistance.

Table 3 Genes with Tn5 insertions mediating Bue1 or Y2 resistance, their predicted gene product and function. Total hits indicate how many mutants with independent transposon insertions in this particular gene were detected overall the 6 independent screenings.

Total Gene Locus tag Bue1 Y2 Product Function hits hypothetical protein predicted EAMY_0795 EAMY_0795 - 1 1 by Glimmer/Critica" hypothetical protein predicted EAMY_1638 EAMY_1638 - 1 1 by Glimmer/Critica Uncharacterized EAMY_2231 EAMY_2231 3 3 3 Glycosyltransferase glycosyltransferase MJ1607

EAMY_2232 EAMY_2232 5 7 7 putative glycosyl transferase Predicted glycosyltransferases

Lipopolysaccharide biosynthesis epsF EAMY_2234 5 6 6 putative glycosyl transferase proteins, LPS:glycosyltransferases NAD-dependent WcaG, Nucleoside-diphosphate- galE2 EAMY_2237 3 4 4 epimerase/dehydratase sugar epimerases family protein UTP-glucose-1-phosphate galF EAMY_2241 1 2 2 UDP-glucose pyrophosphorylase uridylyltransferase UDP-galactopyranose glf EAMY_2233 9 11 12 mutase Glutathione synthase/Ribosomal gshB EAMY_0598 1 1 1 glutathione synthetase protein S6 modification enzyme (glutaminyl transferase) bactoprenol glucosyl Glycosyltransferases involved in cell gtrB1 EAMY_0951 - 1 1 transferase wall biogenesis Hemolysin expression- hha EAMY_1001 1 1 1 modulating protein ATP-dependent hsl protease ATP-dependent protease HslVU hslU EAMY_0131 - 1 1 ATP-binding subunit hslU (ClpYQ), ATPase subunit

49 Manuscript I

Total Gene Locus tag Bue1 Y2 Product Function hits

pgm EAMY_1153 - 1 1 Phosphoglucomutase

ADP-heptose synthase, bifunctional rfaE EAMY_0426 - 1 1 ADP-heptose synthase sugar kinase/adenylyltransferase

rfaH EAMY_0219 - 1 1 Transcriptional activator RfaH Transcription antiterminator

ABC-type polysaccharide/polyol O-antigen export system rfbA1 EAMY_2236 1 3 3 phosphate export systems, permease protein rfbA permease component putative O-antigen export ABC-type sugar transport systems, rfbB1 EAMY_2235 1 1 1 system ATP-binding protein ATPase components dTDP-glucose 4,6- rfbB3 EAMY_2239 - 3 3 dTDP-D-glucose 4,6-dehydratase dehydratase

EAMY_ topB1 3 4 8 DNA topoisomerase III* Topoisomerase IA 1960 lipopolysaccharide ADP-heptose:LPS waaC EAMY_0093 1 1 1 heptosyltransferase-1 heptosyltransferase

waaG EAMY_0085 1 1 1 probable glycosyltransferase Glycosyltransferase

Lipid A core - O-antigen ligase and waaL EAMY_0091 16 18 19 O-antigen ligase related enzymes Lipopolysaccharide core wabK EAMY_0092 6 8 8 biosynthesis Glycosyltransferase glycosyltransferase Predicted glycosyl Glycosyltransferases involved in cell wabM EAMY_0090 14 18 18 transferase, family 2 wall biogenesis glycosyl transferase, group 1 walW1 EAMY_0084 2 2 2 Glycosyltransferase family protein Lipopolysaccharide walW3 EAMY_0088 1 1 1 biosynthesis protein UDP- UDP-N-acetylmuramyl pentapeptide GlcNAc:undecaprenylphosph phosphotransferase/UDP-N- wecA EAMY_0171 2 2 2 ate GlcNAc-1-phosphate acetylglucosamine-1-phosphate transferase transferase O-antigen translocase in LPS Uncharacterized membrane protein, wzx EAMY_0179 2 2 2 biosynthesis putative virulence factor"

wzy EAMY_0181 1 1 1 putative ECA polymerase

ybaJ EAMY_1002 2 2 2 Uncharacterized protein ybaJ

UDP-4-amino-4-deoxy-L- Predicted pyridoxal phosphate- arabinose- dependent enzyme apparently yhjL EAMY_0627 - 1 1 oxoglutarateaminotransferas involved in regulation of cell wall e biogenesis UPF0141 membrane protein Predicted membrane-associated, yjdB EAMY_1696 - 1 1 yjdB metal-dependent hydrolase

The results from the individual screens are summarized in Figure 2. The Venn diagram depicts the identified genes mediating resistance against each phage. The 6 tested phages can be classified into three host receptor groups. Bue1 and Y2 belong to the first group. Both phages rely on different LPS genes for host recognition. The phages L1 and S2 are in the second group and require the amylovoran operon for successful infection. The third group are M7 and S6, which both require the bacterial cellulose synthase operon bcs for host infection. Aside

50 Receptor identification for phage cocktail composition

from waaL and wabM, three genes were identified to mediate resistance against phages belonging to different receptor groups simultaneously. The gene topB1 is predicted to encode a DNA-topoisomerase. Amino acid sequence comparison, however, rather suggests similarities with an NAD-dependent epimerase. The gene rfaE is predicted to encode an ADP- heptose synthase. The gene pgm encodes a phosphoglucomutase.

S6 celA3 lon

M7 bcsA bcsB bcsC EAMY 3608 wssA hmsT

amsA amsJ prlC S2 amsB amsK rcsB3 amsC amsL rplY, pgm amsD aroE thdF rfaE topB1 amsE galE yceD Bue1 L1 amsH miaA

waaC EAMY 2231 waaG EAMY 0795 EAMY 2232 waaL EAMY 1638 epsF Figure 2 Venn diagram of transposon mutagenesis wabK gtrB1 galE2 screen, identifying genes involved in phage wabM hslU galF walW1 rfaH glf resistance mediation. Each phage is depicted in a walW3 rfbB3 gshB different colour. Genes encircled by a specific colour wecA yhjL hha wzx yjdB rfbA1 were detected in resistance against the phage with wzy rfbB1 the corresponding colour. ybaJ

Y2

Phage combinations have potential as biocontrol agents in vitro To verify whether combinations of phages that recognize different host receptors perform better than phages relying on the same receptor for infection, phage combinations were tested in vitro. Figure 3 summarizes the results obtained from in vitro infection assays. The application of single phages revealed that L1 and S2 have the lowest biocontrol potential since bacteria were able to regrow quickly. Bue1 and Y2 were both able to prevent bacterial regrowth for up to 20 h of incubation. The most potent single phage applications were the incubations with M7 or S6, where no growth was observed for up to 40 h. The most potent combinations of two phages were observed to be combinations with Bue1.

51 Manuscript I

Figure 3 In vitro infection assays. E. amylovora

CFBP1430 cells were adjusted to a final

concentration of 105 CFU/ml in LB+ medium and supplemented with a total of 108 PFU/ml single phage or different combinations of phages. Cells were incubated at 26°C for 40 h with double orbital shaking. OD600 was measured every 30

min. Error bars represent standard deviation

52 Receptor identification for phage cocktail composition

Only the combination of Bue1+Y2 was observed to have no synergistic effect since the combination was as potent as each phage alone. In addition, the combination of S2+S6 and S2+Y2 were also able to control bacterial growth during the entire 40 h of incubation. Generally, the combinations of phages were observed to be more potent than single phage treatments with the exception of M7 or S6 alone. Furthermore, the combinations of phages recognizing the same host receptors were observed to perform weaker compared to mixtures with phages from different receptor groups. According to the results obtained from the screens, triple phage combinations were generated whereby phages from different receptor groups were combined and tested for their therapeutic effect in vitro. All tested triple combinations generated prolonged biocontrol of bacterial cells in vitro. The mixture of Y2+S2+S6 was observed to be the weakest phage cocktail since bacterial growth was observed after 25 h of incubation.

Phage treatment against fire blight in blossoms The therapeutic effect of different phage treatments against E. amylovora CFBP1430 was tested in fresh apple blossoms. The concentrations 106, 107 and 108 PFU/ml were tested. Both 107 and 108 PFU/ml reduced disease symptoms. Since the phage Y2 is unreliable in generating high phage titres (>109 PFU/ml) after propagation, all phage treatments were tested at a concentration of 107 PFU/ml against E. amylovora in blossoms. The performances of the phage treated blossoms were evaluated by comparing disease stages of the Figure 4 Treatment of infected blossoms with L1. Blossoms were treated with the untreated blossoms. infected with 107 CFU/ml E. amylovora CFBP1430 and 107 As an example, Figure 4 shows the PFU/ml L1 and incubated for 4 days. Read out was performed by evaluating disease symptoms. Grade 1 indicates healthy results of L1 treatment. Of the infected blossoms without disease symptoms. Grade 2 refers to blossoms blossoms, 12.5% remained disease with first disease symptoms such as browning of the calix. free and healthy (Grade 1), 12.5% Blossoms showing advanced disease symptoms on both calyx showed first disease symptoms at the and stipe correspond to Grade 3. Each treatment was tested at least on 60 individual blossoms. calyx (Grade 2) and 75% were strongly

53 Manuscript I infected at both calyx and stipe (Grade 3). The treatment of infected blossoms with L1 generated 43.75% healthy, 18.75% Grade 2 and 37.5% fully infected blossoms (Grade 3). The L1 treatment was therefore able to reduce the amount of Grade 3 blossoms by 37.5 percentage points (p.p) and increased the amount of Grade 2 blossoms by 6.3 p.p. and Grade 1 blossoms by 31.3 p.p. The results of all the tested treatments are listed in Table 4. The most potent single phage treatment was observed to be L1 (-37.5 p.p., Grade 3). The combination of M7+S6 (-31.3 p.p.) generated the strongest reduction of Grade 3 blossoms generated by two phages in combination. Treatment with the cocktail Bue1+S2+S6 reduced the Grade 3 blossoms by 33.3 p.p.

Table 4 Effect of phage treatments against E. amylovora CFBP1430 on apple blossoms. Numbers depict difference in percent points between phage treated and untreated blossoms. Grade 1 are healthy blossoms, Grade 2 blossoms with first disease symptoms, Grade 3 fully infected blossoms. Green numbers indicate a net gain, red a net reduction of the corresponding disease grade. Treatments where E. amylovora CFBP1430 control generated less than 60% Grade 2 and 3 blossoms are listed in italics.

Difference in disease grade between

untreated and treated blossoms (p.p.) Treatment Grade 1 Grade 2 Grade 3 L1 +31.3 +6.3 -37.5 S6 0 +10.4 -10.4 Y2 +5.6 -6.9 +1.4 L1+Y2 +16.7 0 -16.7 M7+S6 +31.3 0 -31.3 S6+Y2 +10.4 +2.1 -12.5 Bue1+M7+S2 +14.6 +12.5 -27.1 Bue1+S2+S6 +20.8 +12.5 -33.3 S2 +11.2 +4.7 -15.9 Bue1+L1 +16.7 +8.3 -25.0 Bue1+M7 -6.3 +6.3 0 Bue1+S2 +8.3 -4.2 -4.2 Bue1+S6 -2.4 +10.0 -7.5 Bue1+Y2 +1.8 -4.1 +2.3 L1+M7 +16.7 0 -16.7 L1+S2 -23.2 +12.7 +10.5 L1+S6 -10.4 +20.8 -10.4 M7+S2 +14.6 0 -14.6 M7+Y2 -10.4 +18.8 -8.3 S2+S6 +8.3 +14.6 -22.9 S2+Y2 +8.3 +1.1 -9.4 Bue1+M7+S6 -16.7 +12.5 +4.2

54 Receptor identification for phage cocktail composition

Further phage combinations were tested the following year and are listed in Table 5. The treatment with S2 generated a strong reduction of Grade 3 blossoms by -48.5 p.p. Despite this strong reduction, the treatment only reduced the disease symptoms by one Grade (+38.1 p.p. Grade 2; +10.4 p.p. Grade 1). The treatment with L1 generated a strong reduction of Grade 3 blossoms (-33.8 p.p.) but, in contrast to the previous year, the increase of Grade 1 blossoms was much weaker (+9.4 p.p.) The combination of 5 phages simultaneously reduced the Grade 3 blossoms by 16.7 p.p. Overall, the phage treatments appeared less potent than the previous year. Strong Grade 3 reductions can be observed for several treatments, however, the reduction of disease symptoms was only by one Grade. This suggests that the treatments were slowing the infection down instead of preventing it.

Table 5 Effect of phage treatments against E. amylovora CFBP1430 on apple blossoms. Numbers depictdifference in percent points between phage treated and untreated blossoms. Grade 1 are healthy blossoms, Grade 2 blossoms with first disease symptoms, Grade 3 fully infected blossoms. Green numbers indicate a net gain, red a net reduction of the corresponding disease grade.

Difference in disease grade between

untreated and treated blossoms (p.p.) Treatment Grade 1 Grade 2 Grade 3 Bue1 +8.3 +10.4 -18.8 L1 +9.4 +24.4 -33.8 S2 +10.4 +38.1 -48.5 M7 +6.3 +2.1 -8.3 Bue1+S2 0 +22.9 -22.9 Bue1+S6 0 -4.2 +4.2 L1+S6 0 +11.3 -11.3 L1+M7 +12.5 +20.8 -33.3 M7+S2 0 +12.5 -12.5 L1+S2+Bue1+M7+S6 0 +16.7 -16.7

55 Manuscript I

Discussion With increasing numbers of antibiotic resistant bacteria, the threat of bacterial pathogens is rising. Effective alternatives are in urgent need. Bacteriophages could function as potent and highly specific weapons against bacteria. These bacterial viruses are specified in targeting particular bacterial host strains, infect and destroy them. Although phage resistance can be observed, combinations of different phages can prolong the therapeutic effect of the phage biocontrol (1, 32). In this study, a high throughput screen was established to identify host receptors recognized by different phages. The identification of these receptors could facilitate the combination of phages, to generate a highly effective phage cocktail, with minimized risk of resistance development. A transposon mutagenesis library of the fire blight pathogen E. amylovora CFBP1430 was established. The used Tn5 transposon was observed to insert randomly into the E. amylovora CFBP1430 genome since no insertion hot spots could be identified. A total of 56 different genes with transposon insertions were revealed to mediate resistance against at least one phage. More than half of these genes were identified more than once with independent insertions, though it is unclear if the screen was able to identify all these mutants. The screen was adapted for 6 different phages and offers a fast and potent tool to identify host receptors, reveal genes involved in resistance development or uncover mutants where phage infectivity is lowered. The majority of the genes that were revealed to induce M7 and S6 resistance are located in the bcs operon. This operon encodes the bacterial cellulose synthase complex, which is responsible to form and release bacterial cellulose (43). Bacterial cellulose is required to form stable biofilms in E. amylovora (20). The gene bcsC encodes the outer membrane porin BcsC, which is responsible for the release of the produced cellulose strand. This strongly suggests that the outer membrane protein BcsC or cellulose are required for successful M7 and S6 infection. The phages L1 and S2 were unable to lyse mutants with transposon insertions into the ams operon. The operon is required to synthesize amylovoran, a major virulence factor in E. amylovora (17). Furthermore, the gene rcsB3 is annotated as activator of exopolysaccharide synthesis. The majority of the other genes identified to mediate L1 and S2 resistance can be associated with glycosyltransferases or sugar metabolism. None of the other genes are annotated to encode a possible outer membrane protein or extracellular structure. Genome analysis of S2 revealed that the phage encodes a depolymerase (34). Since the depolymerases of L1 and S2 share a high amino acid homology, we can assume that both depolymerases are specific in degrading amylovoran. Furthermore, neither L1 nor S2 can infect the low amylovoran producing strain E. amylovora 4/82 (33).These findings strongly suggest that amylovoran is required for successful L1 and S2 infection.

56 Receptor identification for phage cocktail composition

The majority of the genes involved in Bue1 or Y2 resistance can be associated with LPS synthesis. The other genes are almost exclusively annotated as glycosyltransferases or can be linked to glycosylation steps. In addition, 8 genes were revealed to induce Y2 resistance. These genes can also be associated with sugar transfer processes, with cell wall synthesis or transcription activator processes. The gene yjdB, although annotated as potential membrane protein, is suggested to encode a phosphoethanolamine transferase (BLAST search). BLAST search for the gene EAMY_0795 suggested that the gene encodes a protein belonging to the type III secretion system. This system is used to inject effector genes into other cells (44). It is located in the membrane of bacteria and could potentially be targeted by Y2 as receptor. Nevertheless, the overwhelming amount of genes involved in LPS biosynthesis that mediate Bue1 and Y2 resistance suggests that certain LPS structures are most likely identified and targeted by these two phages. Aside from these surface structures identified by the screen, three genes were revealed to mediate multiple phage resistance. The genes rfaE, pgm and topB1 are all predicted to encode intracellular enzymes. The function and importance of these three genes should be further investigated to assess the risk of simultaneous resistance development against phages from different receptor groups that could render the cocktail ineffective (chapter 3). The results obtained from the screen suggest that although the six phages belong to different genera they can be classified into 3 groups according to the targeted receptor. Hence, combination of phages from different receptor groups should generate the most potent treatment with the lowest risk of resistance development. To confirm this hypothesis, in vitro infection assays were carried out for single phage and phage cocktails. Especially M7 and S6 alone were potent in controlling the pathogen for prolonged time (40 h). Combinations of two phages were generally observed to be more potent than the single phages. The mixtures S2+S6 and S2+Y2 and combinations with Bue1 were able to keep pathogen cell counts below the detection limit for the entire duration of the experiment. Genome analysis of Bue1 revealed that the phage encodes a depolymerase, which could be involved in degrading amylovoran (34). In contrast to the L1 and S2 depolymerases, the Bue1 depolymerase shares higher amino acid identity with the bacterial amsF. This gene is located in the ams operon and contributes to amylovoran synthesis. It could be possible that the Bue1 depolymerase degrades amylovoran in a different manner than the other two depolymerases. Further analysis revealed that Bue1 resists digestion by most restriction enzymes. This was also observed for other phages belonging to the Vi1-like genus. The replacement of thymidine with a modified form of uracil observed in these phages, could be responsible for the resistance. This alteration could ensure that the phage DNA evades host restriction modification systems (45). Both factors could render Bue1 such a potent partner in phage cocktails. Even though the combination L1+S2 (amylovoran receptor group) was observed to be more potent than L1

57 Manuscript I or S2 alone, the mixtures of Bue1+Y2 (LPS) and M7+S6 (Cellulose) were found to be less potent than each phage alone. These findings support the theory of combining phages recognizing different receptors to minimize the risk of phage resistance development. By adding a third phage, which recognizes a different receptor, the appearance of phage resistant bacteria can further be prevented. With the exception of Y2+S2+S6 where bacterial regrowth was observed after 25 h, all triple phage cocktails were shown to efficiently control pathogen growth in vitro during the entire experiment. These results ensure that triple phage cocktails are potent in pathogen control and demonstrate low risk of resistance development in vitro. Tests in blossoms revealed that, aside from Bue1+S2+S6, treatments with M7+S6 were the most potent combinations. Even though M7 and S6 both belong to the same receptor group, the combination generated the second best disease reduction. Either the two phages recognize host bacteria by different structures, which cannot be modified simultaneously, or the receptor is essential for the bacterial infection of the plant. Cellulose was previously suggested to be crucial for biofilm formation and is a key component of a stable biofilm (20). Biofilm formation in the apoplast is a hallmark of the disease and contributes to xylem clogging that result in the characteristic disease symptoms. It is possible that by attacking the cellulose synthase complex, E. amylovora is impaired in establishing a stable biofilm. However, further experiments are required to elucidate the interaction of these two phages with the pathogen and their impact on cellulose production. Interestingly, L1 alone was observed to have a minimal effect on E. amylovora growth in vitro. The treatment of blossoms with L1, however, generated the strongest reduction of disease symptoms in the first experiment, compared to the other phages and phage cocktails tested. A previous study showed that L1 is potent in reducing cell counts by two logs in the first hour in vitro (33). Although cell counts increased rapidly after the first drop, this initial reduction could be crucial to prevent disease outbreak. In addition, the screen revealed that L1 requires amylovoran for successful host recognition and infection. By targeting amylovoran, L1 could force the pathogen to restrict its amylovoran production in order to avoid phage infection. The lack of amylovoran however, renders the pathogen avirulent. S2 treatment reduced the amount of Grade 3 blossoms by 48.5 p.p. and slowed the spread of infection. It should be verified if similar mechanisms as in L1 treatment are applied by S2. These findings suggests that even though triple phage cocktails containing phages from different receptor groups were shown to effectively control the pathogen in vitro, the potency of different phage therapies in blossoms is more complex and difficult to predict and should be analysed carefully. It is possible that certain phages in a cocktail will quickly predominate, depending on latency period and burst size. These underlying mechanisms and dynamics should be investigated further. They could help to facilitate efficient phage cocktail composition and understand phage resistance development.

58 Receptor identification for phage cocktail composition

The triple phage cocktails as well as the most potent phage treatments should be validated in blossoms. However, certain steps of the experiment should be evaluated and adjusted. Plants were kept at 4°C for storage before the blossoming period was induced. The later the blossoming period was induced the more deformed blossoms were observed. This could account for the reduced infectivity of the plant pathogen. The repetition of the phage treatments should therefore be carried out as close to the natural blossoming period as possible. In addition to the detached blossoms, treatments should be tested on entire apple plants against E. amylovora. Adjustments to the pathogen and phage concentrations should be considered. The currently applied pathogen concentration is artificially higher than observed in nature to generate fast readouts. To ensure more physiological conditions, lower pathogen concentrations should be applied. Also the applied phage concentration might need adjustment. In the food industry, experts recommend the use of >108 PFU/ml to tackle pathogenic bacteria. Previous studies utilized phage biocontrol agents with high phage titres to successfully control fire blight (27, 46, 47). Comparison of these results is however difficult due to differences in phage and pathogen concentrations and multiplicity of infection as well as diverging infection settings (timing, method of application and incubation). Most of the previous studies on phage biocontrol of fire blight conclude that maintaining high phage titres on blossoms is crucial for effective eradication of the pathogen. Therefore, different formulations and carrier substances should be evaluated to enhance phage stability and treatment potency. A previous study tested different substances to protect phages from UV-irradiation and identified carrot and red pepper extracts as suitable UV protectant (9). Balogh et al. tested different substances for phage stability against Xanthomonas campestris pv. vesicatoria. Mixtures of phages with pregelatinized corn flour, sucrose, casecrete or skim milk significantly increased phage longevity on the plant surface (48). Furthermore, the addition of non pathogenic alternative host bacteria should be considered. This would ensure continuous phage replication and maintain high phage titres. Good results could be achieved in previous study by Boulé et al. (27), where P. agglomerans Eh21-5 was added as alternative host. In addition, P. agglomerans strains were identified to produce small peptide antibiotics that affect E.amylovora growth. To enhance the phage biocontrol agent further, phage treatments could be fortified by adding certain phage encoded enzymes. Depolymerases (L1, S2, Bue1) can reduce the amount of amylovoran that surrounds the bacterium. Thereby the enzymes expose the surface of E. amylovora, which facilitates receptor accessibility for other phages. The amylovoran capsule also protects the bacterium form the plant immune system by covering pathogen-associated molecular patterns such as LPS. By removing this protective layer, the plant can identify the pathogen and induce immune response accordingly. All these measurement should help to improve efficacy and stability of a potent phage biocontrol agent against fire blight.

59 Manuscript I

Acknowledgments This work was funded by the Swiss National Science Foundation (SNF) grant 310030_156947.

60 Receptor identification for phage cocktail composition

References

1. Buttimer C, McAuliffe O, Ross RP, Hill C, O’Mahony J, Coffey A, O’Mahony J, Coffey A. 2017. Bacteriophages and bacterial plant diseases. Front Microbiol 8:34.

2. Marti R, Zurfluh K, Hagens S, Pianezzi J, Klumpp J, Loessner MJ. 2013. Long tail fibres of the novel broad-host-range T-even bacteriophage S16 specifically recognize Salmonella OmpC. Mol Microbiol 87:818–834.

3. Fehmel F, Feige U, Niemann H, Stirm S. 1975. Escherichia coli capsule bacteriophages. VII. Bacteriophage 29-host capsular polysaccharide interactions. J Virol 16:591–601.

4. Guerrero-Ferreira RC, Viollier PH, Ely B, Poindexter JS, Georgieva M, Jensen GJ, Wright ER. 2011. Alternative mechanism for bacteriophage adsorption to the motile bacterium Caulobacter crescentus. Proc Natl Acad Sci U S A 108:9963–8.

5. Shin H, Lee JH, Kim H, Choi Y, Heu S, Ryu S. 2012. Receptor diversity and host interaction of bacteriophages infecting Salmonella enterica serovar Typhimurium. PLoS One 7:e43392.

6. Bertozzi Silva J, Storms Z, Sauvageau D. 2016. Host receptors for bacteriophage adsorption. FEMS Microbiol Lett 363:fnw002.

7. Loc-Carrillo C, Abedon ST. 2011. Pros and cons of phage therapy. Bacteriophage 1:111–114.

8. Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JA, Momol MT, Jones JB, Vallad GE. 2012. Soil-based systemic delivery and phyllosphere in vivo propagation of bacteriophages: Two possible strategies for improving bacteriophage persistence for plant disease control. Bacteriophage 2:215–224.

9. Born Y, Bosshard L, Duffy B, Loessner MJ, Fieseler L. Protection of Erwinia amylovora bacteriophage Y2 from UV-induced damage by natural compounds. Bacteriophage 5:e1074330.

10. Hagens S, Loessner MJ. 2010. Bacteriophage for biocontrol of foodborne pathogens: calculations and considerations. Curr Pharm Biotechnol 11:58–68.

11. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell 108:13–6.

12. Cobián Güemes AG, Youle M, Cantú VA, Felts B, Nulton J, Rohwer F. 2016. Viruses as winners in the game of life. Annu Rev Virol 3:197–214.

13. Bonn WG, van der Zwet T. 2000. Distribution and economic importance of fire blight, p. 37–53. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

14. Thomson S V. 2000. Epidemiology of fire blight, p. 9–36. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

15. Momol MT, Aldwinckle HS. 2000. Genetic diversity and host range of Erwinia amylovora, p. 55–72. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

61 Manuscript I

16. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald P, Dow M, Verdier V, Beer S V., Machado MA, Toth I, Salmond G, Foster GD. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol 13:614–629.

17. Piqué N, Miñana-Galbis D, Merino S, Tomás JM. 2015. Virulence factors of Erwinia amylovora: A review. Int J Mol Sci 16:12836–54.

18. Smith TJ, Pusey PL. 2011. Cougarblight 2010, a significant update of the cougarblight fire blight infection risk model. Acta Hortic 331–336.

19. Philion V, Trapman M. 2011. Description and preliminary validation of RIMpro-Erwinia, a new model for fire blight forecast. Acta Hortic 307–317.

20. Castiblanco LF, Sundin GW. 2016. Cellulose production, activated by cyclic di-GMP through BcsA and BcsZ, is a virulence factor and an essential determinant of the three- dimensional architectures of biofilms formed by Erwinia amylovora Ea1189. Mol Plant Pathol 19:90–103.

21. Geider K. 2000. Exopolysaccharides of Erwinia amylovora: structure, biosynthesis, regulation, role in pathogenicity of amylovoran and levan, p. 117–140. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

22. Vanneste JL, Eden-Green S. 2000. Migration of Erwinia amylovora in host plant tissues, p. 73–83. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

23. Ordax M, Piquer-Salcedo JE, Santander RD, Sabater-Muñoz B, Biosca EG, López MM, Marco-Noales E. 2015. Medfly Ceratitis capitata as potential vector for fire blight pathogen Erwinia amylovora: Survival and transmission. PLoS One 10:e0127560.

24. Jones AL, Schnabel EL. 2000. The development of streptomycin-resistant strains of Erwinia amylovora, p. 235–251. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

25. Russo NL, Burr TJ, Breth DI, Aldwinckle HS. 2008. Isolation of streptomycin-resistant isolates of Erwinia amylovora in New York. Plant Dis 92:714–718.

26. Svircev AM, Castle AJ, Lehman SM. 2010. Bacteriophages for control of phytopathogens in food production systems, p. 79–96. In Sabour, PM, Griffiths, MW (eds.), Bacteriophages in the control of food- and waterborne pathogens. ASM Press, Washington, DC.

27. Boulé J, Sholberg PL, Lehman SM, O’Gorman DT, Svircev AM. 2011. Isolation and characterization of eight bacteriophages infecting Erwinia amylovora and their potential as biological control agents in British Columbia, Canada. Can J Plant Pathol 33:308– 317.

28. Lehman SM. 2007. Ph.D. thesis: Development of a bacteriophage-based biopesticide for fire blight. Department of Biological Sciences, Brock University, St. Catharines.

29. Chadha P, Katare OP, Chhibber S. 2016. In vivo efficacy of single phage versus phage cocktail in resolving burn wound infection in BALB/c mice. Microb Pathog 99:68–77.

30. Naghizadeh M, Karimi Torshizi MA, Rahimi S, Dalgaard TS. 2018. Synergistic effect of

62 Receptor identification for phage cocktail composition

phage therapy using a cocktail rather than a single phage in the control of severe colibacillosis in quails. Poult Sci 98:653–663.

31. Schmerer M, Molineux IJ, Bull JJ. 2014. Synergy as a rationale for phage therapy using phage cocktails. PeerJ 2:e590.

32. Born Y, Fieseler L, Klumpp J, Eugster MR, Zurfluh K, Duffy B, Loessner MJ. 2014. The tail-associated depolymerase of Erwinia amylovora phage L1 mediates host cell adsorption and enzymatic capsule removal, which can enhance infection by other phage. Environ Microbiol 16:2168–2180.

33. Born Y, Fieseler L, Marazzi J, Lurz R, Duffy B, Loessner MJ. 2011. Novel virulent and broad-host-range Erwinia amylovora bacteriophages reveal a high degree of mosaicism and a relationship to Enterobacteriaceae phages. Appl Environ Microbiol 77:5945– 5954.

34. Knecht LE, Born Y, Pothier JF, Loessner MJ, Fieseler L. 2018. Complete genome sequences of Erwinia amylovora phages vB_EamP-S2 and vB_EamM-Bue1. Microbiol Resour Announc 7:e00891-18.

35. Adams MH. 1959. Bacteriophages. Interscience Publishers Inc., New York.

36. Sambrook J, Russell DW. 2001. Molecular cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

37. Hanahan D. 1983. Studies on transformation of Escherichia coli with plasmids. J Mol Biol 166:557–580.

38. Das S, Noe JC, Paik S, Kitten T. 2005. An improved arbitrary primed PCR method for rapid characterization of transposon insertion sites. J Microbiol Methods 63:89–94.

39. Pusey PL. 1997. Crab apple blossoms as a model for research on biological control of fire blight. Phytopathology 87:1096–102.

40. Llop P, Cabrefiga J, Smits THM, Dreo T, Barbé S, Pulawska J, Bultreys A, Blom J, Duffy B, Montesinos E, López MM. 2011. Erwinia amylovora novel plasmid pEI70: complete sequence, biogeography, and role in aggressiveness in the fire blight phytopathogen. PLoS One 6:e28651.

41. Castiblanco LF, Sundin GW. 2016. New insights on molecular regulation of biofilm formation in plant-associated bacteria. J Integr Plant Biol 58:362–372.

42. Bugert P, Geider K. 1995. Molecular analysis of the ams operon required for exopolysaccharide synthesis of Erwinia amylovora. Mol Microbiol 15:917–933.

43. Römling U, Galperin MY. 2015. Bacterial cellulose biosynthesis: diversity of operons, subunits, products, and functions. Trends Microbiol 23:545–57.

44. Deng W, Marshall NC, Rowland JL, McCoy JM, Worrall LJ, Santos AS, Strynadka NCJ, Finlay BB. 2017. Assembly, structure, function and regulation of type III secretion systems. Nat Rev Microbiol 15:323–337.

45. Adriaenssens EM, Ackermann H-W, Anany H, Blasdel B, Connerton IF, Goulding D, Griffiths MW, Hooton SP, Kutter EM, Kropinski AM, Lee J-H, Maes M, Pickard D, Ryu S, Sepehrizadeh Z, Shahrbabak SS, Toribio AL, Lavigne R. 2012. A suggested new

63 Manuscript I

bacteriophage genus: Viunalikevirus. Arch Virol 157:2035–46.

46. Erskine JM. 1973. Characteristics of Erwinia amylovora bacteriophage and its possible role in the epidemiology of fire blight. Can J Microbiol 19:837–845.

47. Müller I, Lurz R, Kube M, Quedenau C, Jelkmann W, Geider K. 2011. Molecular and physiological properties of bacteriophages from North America and Germany affecting the fire blight pathogen Erwinia amylovora. Microb Biotechnol 4:735–745.

48. Balogh B, Jones JB, Momol MT, Olson SM, Obradovic A, King P, Jackson LE. 2003. Improved efficacy of newly formulated bacteriophages for management of bacterial spot on tomato. Plant Dis 87:949–954.

64 Phage infection of E. amylovora requires cellulose

2.2. Manuscript II: Phage infection of E. amylovora requires cellulose

Bacterial cellulose is required for phage M7 and S6 infection in Erwinia amylovora

Leandra E. Knecht1,2, Nadine Heinrich2, Katja Felder2, Yannick Born1, Cosima Pelludat3, Martin J. Loessner2, Lars Fieseler1*

1 Food Microbiology Research Group, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Wädenswil, Switzerland 2 Institute of Food, Nutrition and Health, ETH Zurich, Switzerland 3 Agroscope, Plant Pathology and Zoology in fruit and vegetable production, Wädenswil, Switzerland

*Correspondence: Lars Fieseler, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Einsiedlerstrasse 31, Wädenswil, Switzerland. Tel: +41 58 934 54 07; e-mail: [email protected]

65 Manuscript II

Abstract The plant pathogen Erwinia amylovora, the causative agent of fire blight, affects members of the Rosaceae family. One of its major virulence factors is the ability to form biofilms leading to clogged plant vessels, which causes disease symptoms. Although streptomycin may be used to control the plant infection, it has been banned in an increasing number of countries. Hence, alternative approaches are urgently required. Over recent years, bacteriophages have been emerging as possible alternatives to conventional treatments. These bacterial viruses specifically target and destroy host cells by recognizing receptors on their surface. The identification of these molecules is crucial for effective phage cocktail formulation. In a previous study, a high throughput screen of an E. amylovora CFBP1430 Tn5 transposon mutagenesis library revealed genes directly influencing phage infection. Phages M7 and S6 were unable to infect mutants with defective genes in the bacterial cellulose synthase operon (bcs). The constitutively expressed Bcs complex is responsible for producing and secreting bacterial cellulose, a carbohydrate polymer associated with biofilm. Deletion of the operon or single genes (bcsA, bcsC and bcsZ) verified their role for M7 and S6 infection. Genes encoding putative cellulases and endoglucanases in the genomes of M7 and S6 also suggested an interaction with bacterial cellulose. This hypothesis was further supported using the cellulose binding dye Congo Red, which effectively blocked phage attachment and infection. These findings suggest that M7 and S6 either recognize components of the Bcs complex such as the outer membrane protein BcsC, or require cellulose itself as binding ligand.

66 Phage infection of E. amylovora requires cellulose

Introduction The Gram-negative, facultative anaerobe Erwinia amylovora is the causative agent of fire blight (1, 2). This serious plant disease affects members of the Rosaceae family and was classified as one of the ten most devastating plant pathogens in crop production (3–6). E. amylovora is able to produce three different exopolysaccharides (EPS), which act as main virulence factors (7). Amylovoran is a complex polysaccharide with a galactose backbone encoded by the ams operon. Levan is a homopolymer derived from sucrose. Bacterial cellulose is produced and secreted by the bacterial cellulose synthase encoded by the bcs operon. Cellulose is required for modulating the structure of the biofilm (8–10). Both amylovoran production and biofilm formation are tightly regulated by the intracellular second messenger cyclic-di-GMP (c-di-GMP) (10–12). Mutants deficient in EPS production are unable to trigger disease (8, 13). As a main route of infection, E. amylovora grows in the blossoms before entering the plant tissue through the nectaries. Once in the plant tissue, the pathogen forms a thick biofilm eventually clogging the xylem vessels (10, 14–17). This leads to desiccation of affected tissues resulting in the characteristic disease symptoms (17). At this stage, the progressing infection can only be stopped by generous pruning or elimination of the entire plant. In many countries, early targeting of fire blight in the blossoming season is achieved by streptomycin. By lowering the cell counts of E. amylovora below the critical number necessary for entry into the plant tissue, the antibiotic is able to prevent epidemic disease outbreaks. However, antibiotic-resistant strains emerge quickly and have already been reported (18, 19). The application of streptomycin has therefore been banned in an increasing number of countries (20). As a consequence, environmentally friendly alternative treatments are required and are currently being investigated. Biological control of fire blight by bacteriophages could be such an alternative. Phages are viruses exclusively infecting bacteria. They pose the most abundant biological entity on earth reaching an estimated number of 1031 (21–23), and outnumber bacteria by at least tenfold (24). This large variety ensures a collection of highly efficient phages against possibly any bacterial species. Phages have several advantages over antibiotics (25). They can be isolated from natural environmental samples and can be propagated to high titres in a laboratory. In contrast to antibiotics, only the specific host bacteria will be affected by any phage, leaving other and potentially beneficial bacteria unharmed (25). Additionally, phages do not cause side effects in humans and animals, and are considered as environmentally safe (26, 27). Phages are furthermore highly dynamic and adaptable. In the presence of their specific host bacteria, phage numbers increase by utilizing the host cell for replication. In their absence, phages start to decay due to environmental factors such as UV irradiation (28–30). Although

67 Manuscript II phage resistance can be frequently observed, phages are able to adapt to the changing host, bypassing its resistance (31, 32). The existing evidence clearly shows that phage treatments have the potential to control different plant pathogens (26, 33), such as Pectobacterium carotovorum ssp. carotovorum (34, 35), Dickeya solani (36), Xanthomonas campestris pv. vesicatoria (37, 38), Xylella fastidiosa (39) and also E. amylovora (40–42). The abundance of phages in nature generates almost unlimited possibilities of combining different phages into suitable cocktails. In contrast to single phage application, mixtures may be more efficient in treating pathogens (43). Phages combined in a cocktail often demonstrate synergistic effects, rendering the treatment more potent (44–46). An example of such effects are enhanced adsorption to and killing of E. amylovora when exposed to the phages L1 and Y2 (47). L1 encodes a depolymerase, which is able to break down amylovoran (45). Such accessory enzymes have the potential to be powerful weapons against biofilm forming pathogens. By removing the protective capsule, the L1 depolymerase facilitates the access of Y2 to the surface (45, 48), which effectively kills the bacteria. Phages recognise their host through binding to a specific receptor. These receptors can be proteins, sugar molecules or cell surface structures located either in the outer membrane (49), in slime layers or capsules (50), or located in appendages such as pili (51) and flagella (52, 53). Bacteria challenged with a specific phage can develop phage resistance by modifying or obstructing the targeted receptor (54). A bacterium confronted with several phages has to adapt to different recognition, infection and replication strategies simultaneously. The risk of phage resistance development is therefore reduced (55, 56). The identification of phage receptors on the host surface is crucial to anticipate the potency of phages and the possibility of phage resistance. The E. amylovora specific phages M7 and S6 were isolated from Swiss apple and pear orchards in 2011 (47). The FO1-like myovirus M7 is highly potent against all tested E. amylovora strains, but could also infect certain closely related strains. The podovirus S6 encodes a putative virion-encapsidated RNA polymerase, a hallmark of N4-phages (57). Genome analysis revealed that M7 and S6 can be classified as strictly lytic, since neither of the two phages encode an integrase or a lysogeny control region. Determination of the genome packaging mechanisms of the two phages suggests a low risk for generalized transduction of these two viruses (47). Both phages have been tested for possible use in biocontrol of E. amylovora. Although single application as well as combined application of M7 and S6 reduced the bacterial cell counts only minimally, bacterial growth was affected after 24 hours where 2.5 (M7), 2.0 (S6) or 1.6 (M7+S6) log reductions compared to non infected bacteria were observed. Furthermore, combination of S6 with other phages such as L1 reduced cell counts below the detection limit (47). These data suggest that M7 and S6 are not only safe, but also offer the potential for control of E. amylovora.

68 Phage infection of E. amylovora requires cellulose

The aim of this study was to identify the phage receptors recognized by M7 and S6 on the surface of E. amylovora. It could be revealed that bacterial cellulose and the bacterial cellulose synthase complex (Bcs) are key components for successful infection of E. amylovora CFBP1430 by the phages M7 and S6. The identification of the binding sites enabled a better understanding of the first steps of the phage infection process.

69 Manuscript II

Materials and Methods

Bacterial strains and culture conditions E. amylovora strains were cultivated on LB agar at 28°C, Escherichia coli strains at 37°C (Table 1) The antibiotics kanamycin (50 µg/ml), streptomycin (100 µg/ml) and ampicillin (100 µg/ml) were added if required.

Table 1 Bacterial strains, plasmids, phages and antibodies used in this study

Bacterial strains Characterization Reference

E. amylovora

CFBP1430 Wild type strain, isolated in France from Crataegus sp; propagation strain for all phages Zhang and Geider, 1997 CFBP1430SmR Streptomycin resistant CFBP1430 used for knock-out generation CFBP1430 [pBAD18] Empty vector control This work

CFBP1430Δbcs Deletion of entire bacterial cellulose synthase operon (bcs) This work CFBP1430Δbcs+bcsC Deleted bcs operon complemented with bcsC This work CFBP1430ΔbcsA Deletion of the enzymatically active subunit BcsA This work CFBP1430ΔbcsA+bcsA Complementation of bcsA This work CFBP1430+bcsA Overexpression of bcsA This work

CFBP1430ΔbcsC Deletion of the outer membrane protein BcsC This work CFBP1430ΔbcsC+bcsC Complementation of bcsC This work CFBP1430+bcsC Overexpression of bcsC This work CFBP1430ΔbcsZ Deletion of periplasmic protein BcsZ This work CFBP1430ΔbcsZ+bcsZ Complementation of bcsZ This work CFBP1430+hmsT Overexpression of hmsT This work 4/82 Isolated in Egypt from Pyrus communis; low EPS-producer Zhang and Geider, 1997

E. coli Miller & Mekalanos, S17λ-pir 1988; Simon et al. 1988 Dh5α supE44 ∆lacU169 (φ80lacZ∆M15) hsdR17 recA1 endA1 gyrA96, thi-1 relA1 Hanahan, 1983

XL1-Blue NEB (Ipswich, USA)

Plasmids

pSB315 Containing kanamycin cassette without transcriptional terminator, AmpR, KanR Galan et al. (1992) Skorupski & Taylor pKAS32 Suicide vector with rpsL gene, AmpR (1996) pKAS32::deletion bcs Deletion vector to exchange bcs for a kanamycin cassette through homologous recombination This work pKAS32::deletion bcsA Deletion vector to exchange bcsA for a kanamycin cassette through homologous recombination This work pKAS32::deletion bcsC Deletion vector to exchange bcsC for a kanamycin cassette through homologous recombination This work pKAS32::deletion bcsZ Deletion vector to exchange bcsZ for a kanamycin cassette through homologous recombination This work pBAD18 Complementation vector, arabinose induced pBAD promotor Guzman et al 1995 pBAD18:: bcsA Complementation vector for bcsA, artificial RBS, arabinose induced pBAD promoter This work pBAD18:: bcsC Complementation vector for bcsC, artificial RBS, arabinose induced pBAD promoter This work pBAD18:: bcsZ Complementation vector for bcsZ, artificial RBS, arabinose induced pBAD promoter This work pBAD18:: hmsT Complementation vector for hmsT, artificial RBS, arabinose induced pBAD promoter This work

Phages

M7 E. amylovora specific myovirus, FO1-like Born et al, 2011 S6 E. amylovora specific podovirus, N4-like Born et al, 2011 L1 E. amylovora specific podovirus, T7-like Born et al, 2011 Y2 E. amylovora specific myovirus, GJ1-like Born et al, 2011

Antibodies

rabbit Anti BcsC -1 Anti- GYTGENVASNSC, Genscript This work

rabbit Anti BcsC -2 Anti- MDFSKNLSGFSLGQC, Genscript This work rabbit Anti BcsC -3 Anti- CYAKEAYFSGSSKSG, Genscript This work Abcam (Cambridge, Goat anti-rabbit IgG (HRP) Anti-Rabbit, Abcam Great Britain)

70 Phage infection of E. amylovora requires cellulose

Soft agar overlay and propagation of M7 and S6 Phages were propagated using the soft agar overlay method (58). Four millilitres molten LB+ soft agar (LB Bouillon, 4 g/l agar, 2 mM MgSO4, 10 mM CaCl2) were supplemented with 90 µl bacterial overnight culture and 10 µl diluted bacteriophage and spread on LB plates to generate semi confluent lysed plates. After overnight incubation, 5 ml SM buffer (100 mM

NaCl, 8 mM MgSO4, 50 mM Tris-Cl, pH 7.4) were added per plate and incubated for 5 h at room temperature (RT) with shaking. The decanted supernatant was supplemented with 0.5 M NaCl. The bacteriophage suspension was treated with polyethylenglycol (10% w/v PEG 8,000, ice bath overnight), purified (CsCl density gradient) (59) and dialyzed against SM buffer.

Generation of knockout mutants A previously performed Tn5 library screen revealed that gene disruptions in the bacterial cellulose synthase operon bcs renders E. amylovora CFBP1430 phage resistant (chapter 1). To elucidate the impact of the bcs operon on M7 and S6, the most promising genes bcsA, bcsC and bcsZ in the operon were replaced by a kanamycin resistance gene. Knockout mutants were generated by allelic exchange using the suicide plasmid pKAS32 carrying a R6K origin of replication (60). Flanking regions of the gene of interest and a kanamycin cassette (aphT) amplified from pSB315 (61) were integrated into the multiple cloning site of pKAS32. The two flanking regions included approximately 1,000 bp up- and downstream of the gene to be knocked out. The three fragments were amplified using Gibson assembly primers designed by the NEBuilder tool (NEBuilder Assembly Tool v1.12.18) using the KAPA HIFITM PCR kit (KAPA Biosystems, Wilmington, USA). The fragments were joined by overlap extension PCR. The vector pKAS32 was extracted from E. coli S17-1 λpir [pKAS32] according to the NucleoSpin® Plasmid Kit (Macherey-Nagel; Düren Germany). The vector pKAS32 was linearized by EcoRV. The product was anayzed by electrophoresis and purified from the 1% agarose gel by the DNA Clean & Concentrator™-5 Kit by Zymo Research. To prevent re- circularization of the linearized vector, a dephosphorylation step was carried out (CIP, 1 U/

250 ng vector). The plasmid was purified and eluted in 15 µl ddH2O. The insert was phosphorylated to facilitate formation of the vector. The suicide plasmid was generated either through Gibson assembly (62), or through classic ligation using a T4 polynucleotide ligase (Thermo Fisher Scientific) according to the manufacturer’s protocol. The plasmids were introduced into electrocompetent, streptomycin resistant E. amylovora CFBP1430 by electroporation. After electroporation, the cells were resuspended in SOC and incubated at 30°C for 1 h with vigourous shaking. Subsequently, the cells were plated on LB plates complemented with kanamycin and streptomycin. Correctness of deletion was checked by PCR.

71 Manuscript II

Table 2 Primers used in this study

Primer Sequence 5'-3' Characterization Reference

pSB315 f GAA AGC CAC GTT GTG TCT C Amplification of the Kanamycin cassette in pSB315 Galan et al. 1992

pSB315 r CCT TCA TTA CAG AAA CGG C Amplification of the Kanamycin cassette in pSB315

pBAD f CTG TTT CTC CAT ACC CGT T Verification of plasmid insertion

pBAD r CTC ATC CGC CAA AAC AG Verification of plasmid insertion

bcs, bcsZ pKAS32-up f CAG ATC TGC GCG CGA TCG ATT GAT TTG CAT Gibson assembly Δbcs, ΔbcsZ in E.amylovora This work GCG TGT TAC CFBP 1430 bcs, bscZ up-kan f CGT CCC TTA GAT TAG AAA AAC TCA TCG AGC Gibson assembly Δbcs, ΔbcsZ in E.amylovora This work ATC CFBP 1430 bcs, bscZ up-kan r GAG TTT TTC TAA TCT AAG GGA CGA AAG GTG G Gibson assembly Δbcs, ΔbcsZ in E.amylovora This work CFBP 1430 bcs kan-down f TAT GGC TCA TCA GTA TGA TCC CCA AGG TAG Gibson assembly Δbcs in E.amylovora CFBP 1430 This work bcs kan-down r GAT CAT ACT GAT GAG CCA TAT TCA ACG G Gibson assembly Δbcs in E.amylovora CFBP 1430 This work bcs down-pKAS32 r CGC AAA TTT AAA GCG CTG ATG ATG GAT CTG Gibson assembly Δbcs in E.amylovora CFBP 1430 This work CAA CAG GAC bscZ kan-down f TAT GGC TCA TAA AGC CAG CCC TGA GTA TGG Gibson assembly ΔbcsZ in E. amylovora CFBP This work CC 1430 bscZ kan-down r GGC TGG CTT TAT GAG CCA TAT TCA ACG G Gibson assembly ΔbcsZ in E. amylovora CFBP This work 1430 bcsZ down pKAS32 r CGC AAA TTT AAA GCG CTG ATA TGG CGA ACG Gibson assembly ΔbcsZ in E. amylovora CFBP This work CCG GGG CT 1430 pBAD18::bcsZ f CCC GTT TTT TTG GGC TAG CGG GAG GTC GTA Gibson assembly, complementation bcsZ mutant, This work ATG CTG TTG GCG CTC ATA ATG pBAD18 pBAD18::bcsZ r CTC ATC CGC CAA AAC AGC CAT TAG CGG TCT Gibson assembly, complementation bcsZ mutant, This work TTA AAC GCC pBAD18 bcsC pKAS32-up f CAG ATC TGC GCG CGA TCG ATG GAG GGG TTC Gibson assembly ΔbcsC in E. amylovora CFBP This work AGC ACC AC 1430 bcsC up-kan f TGA CTC ATG CAT TAG AAA AAC TCA TCG AGC Gibson assembly ΔbcsC in E. amylovora CFBP This work 1430 bcsC up-kan r GAG TTT TTC TAA TGC ATG AGT CAG TTA CTC Gibson assembly ΔbcsC in E. amylovora CFBP This work CAG G 1430 bcsC kan-down f TAT GGC TCA TTT ATT TTT CAT TGT CTT TTT TCT Gibson assembly ΔbcsC in E. amylovora CFBP This work CGT CAG AAG 1430 bcsC kan-down r ATG AAA AAT AAA TGA GCC ATA TTC AAC GG Gibson assembly ΔbcsC in E. amylovora CFBP This work 1430 bcsC down-pKAS32 r CGC AAA TTT AAA GCG CTG ATG CAA TAT ACG Gibson assembly ΔbcsC in E. amylovora CFBP This work CTG CGG CTC 1430 pBAD18::bcsC f CCC GTT TTT TTG GGC TAG CGG GAG GTC GTA Gibson assembly, complementation bcsC mutant, This work ATG CAC GGC GCC GTC TGC pBAD18 pBAD18::bcsC r CTC ATC CGC CAA AAC AGC CAT TAT TTG CCC Gibson assembly, complementation bcsC mutant, This work CCC AGC ATG TAG CG pBAD18 bcsA pKAS32-up f CAG ATC TGC GCG CGA TCG ATC ATT TTG TTC Gibson assembly ΔbcsA in E. amylovora CFBP This work GGG CAG CTG GC 1430 bcsA up-kan f GCG CGG CGG TCT TTC CAC ATT AGA AAA ACT Gibson assembly ΔbcsA in E. amylovora CFBP This work CAT CGA GC 1430 bcsA up-kan r GCT CGA TGA GTT TTT CTA ATG TGG AAA GAC Gibson assembly ΔbcsA in E. amylovora CFBP This work CGC CGC GC 1430 bcsA kan-down f TAT GGC TCA TGA GTC ATC CTG GAA AGC ATA Gibson assembly ΔbcsA in E. amylovora CFBP This work TCG TTA TCA TAC GTG AC 1430 bcsA kan-down r AGG ATG ACT CAT GAG CCA TAT TCA ACG G Gibson assembly ΔbcsA in E. amylovora CFBP This work 1430 bcsA down-pKAS32 r CGC AAA TTT AAA GCG CTG ATC CAC GGC GCG Gibson assembly ΔbcsA in E. amylovora CFBP This work GAT CAC CC 1430 pBAD18::bcsA f CCC GTT TTT TTG GGC TAG CGG GAG GTC GTA Gibson assembly, complementation bcsA mutant, This work ATG AAT AAA GTC CTG TTT TAT CTG CTG TTG C pBAD18 pBAD18::bcsA r CTC ATC CGC CAA AAC AGC CAT CAT GCC GCG Gibson assembly, complementation bcsA mutant, This work CCA TCC TC pBAD18 pBAD18::hmsT f CCC GTT TTT TTG GGC TAG CGA ATT CGG AGG Gibson assembly, complementation hmsT mutant, This work TCG TAA TGA ATT TGC AAA GCT ACG pBAD18 pBAD18::hmsT r CTC ATC CGC CAA AAC AGC CAA GCT TTT ATA Gibson assembly, complementation hmsT mutant, This work CGA TAT CTG CAG GTT TAC pBAD18

Complementation of knockout mutants and generation of gene overexpressions The arabinose inducible plasmid pBAD18 (63) holding an ampicillin resistance was linearized with EcoRI and HindIII and purified. The inserts and their ribosomal binding sites were amplified from E. amylovora CFBP1430 by PCR with Gibson primers (NEB). PCR products with the correct length were recovered from a 1% agarose gel. The linearized vector pBAD18 and the insert were joined by Gibson assembly (62). The newly formed plasmids were introduced into electrocompetent E. coli XL1-Blue cells for amplification. Cells were recovered in SOC and incubated for 1 h at 37°C with vigourous shaking before plating onto LB ampicillin plates. Correct plasmid insertion was verified by PCR with the primer pair pBAD fw and pBAD rev (Table 2). Correct plasmids were extracted and introduced into electrocompetent E.

72 Phage infection of E. amylovora requires cellulose amylovora CFBP1430 and knockout mutants. Plaque assays involving different dilutions of M7 and S6 were carried out and the plaque forming ability was monitored. Soft agar and LB plates were supplemented with ampicillin and 0.2% arabinose for promoter induction or 0.2% glucose for promoter repression.

Growth curves The impact of the generated gene modifications on the growth of these mutants was monitored. Bacteria were washed twice in SM buffer and OD600 was adjusted to 0.1. Bacteria were diluted in LB medium supplemented with MgSO4 (2 mM) and CaCl2 (10 mM) to reach a starting concentration of 105 CFU/ml. If required, the medium was supplemented with kanamycin and/or ampicillin. Cultures were shaken at 150 rpm (double orbital) at 28°C for 24 h in a plate reader.

SDS-PAGE and western blot Western blots using polyclonal Anti-BcsC antibodies were carried out to establish whether BcsC was produced in the different generated mutants. Mutants were grown overnight and supplemented with antibiotics if required. The cells were transferred into 50 ml of the corresponding prewarmed medium and incubated at 28°C with vigorous shaking until an OD600 of 0.5 was reached. Cells carrying a pBAD18 based construct were induced with 0.2% arabinose and incubated over night at 17°C with vigorous shaking before cells were harvested (4,000 xg, 20 min.). The cells were subsequently lysed using Bacterial Protein Extraction Reagent B-PER (Thermo Fisher) (4 ml/g pellet), lysozyme (8 µl/g pellet) and DNase I (8 µl/g pellet) and incubated for 15 min at RT in an overhead rotator. Two millilitres of each lysate were centrifuged at 10,000 xg for 20 min. at 4°C. The supernatant was collected and stored at -20°C and the pellet was resuspended in 100 µl 1x SDS-PAGE sample buffer (Tris-base 90 mM, SDS 2%, Bromophenol Blue 0.02%, glycerol 20% in ddH2O and pH adjusted to 6.8). Five microliters of each pellet sample were supplemented with 4 µl 1x SDS-PAGE sample buffer and 1 µl DTT (1 M) and boiled for 5 min at 95°C. Three microliters per sample were loaded on a 1 mm 7.5% resolving, 4% stacking gel and let run for 25 min at 200 V. The transfer was carried out using a TransBlot-Turbo System by Biorad and a PVDF Transfer Pack and let run for 10 min (1.3 A, 25 V). The bands were visualized by adding 5 ml Ponceau S solution to the membrane. After removing the Ponceau S solution, the membrane was blocked for 1 h at RT using blocking buffer (5% w/v milk powder, Triton X-100 0.1% in 1x PBS). The membrane was washed three times with PBS before the rabbit anti-BcsC Antibodies (GenScript) were added (1/1,000 in blocking buffer) and incubated at RT for 1 h under constant shaking. The membrane was washed three times with PBS before the secondary goat anti-rabbit IgG (HRP) antibody (abcam) was added (1/2,500 x in blocking buffer) and incubated for 1 h at RT with

73 Manuscript II constant shaking. Signals were detected through chemiluminescence by applying 10 ml ECL buffer (Tris 100 mM in ddH2O, pH 8.6), 22 µl p-coumaric acid, 50 µl luminol and 3 µl H2O directly onto the membrane for 90 seconds (64). The chemiluminescence solution was carefully removed and signals were detected by an Azure Biosystem C300.

Pull-down assay The adsorption of the bacteriophages to phage resistant mutants was tested by pull-down assays (45). E. amylovora CFBP1430 was used as positive control, E. coli Dh5α and medium or buffer only as negative controls. Cells from an overnight culture were washed with SM

Buffer, PBS or LB medium twice and diluted to an OD600 of 1.0. Ten microliters of S6 or M7 (109 PFU/ml) were added to 990 µl bacteria. After incubating at RT for 10, 20, 30, 40, 50, 60, 70 min with shaking, samples were centrifuged for 5 min at 10,000 xg or 5,000 xg at 4°C. Unbound phages in the supernatant and cell bound phages in the pellet were quantified by the soft agar overlay plating on E. amylovora CFBP1430.

Cellulose visualization To investigate alterations in the cellulose production provoked by the gene modifications, qualitative cellulose visualization was performed. LB agar was supplemented with 40 μg/ml Congo Red and 20 μg/ml Coomassie Brilliant Blue before pouring. Ampicillin and/or 0.2% arabinose were added if required. Bacteria were spotted (5 µl per sample) on plates and incubated for two days at RT. Bacteria with the ability to produce cellulose appeared as pink colonies while non-producing bacteria remained white (15).

EPS Measurement Impact of bcs modifications on EPS production was monitored using the EPS- Cetylpyridinium chloride (CPC) precipitation assay (65). Bacteria were grown in MM2 (65) at 28°C under shaking for 24 h. Samples were adjusted to an OD600 of 0.7 and 1 ml per sample was centrifuged at 10,000 xg for 5 min. Nine hundred and fifty microliters of supernatant were mixed with 50 μl CPC (50 mg/ml) and incubated for 10 min at RT. Finally, OD600 was measured. The low EPS producing strain E. amylovora 4/82 was used as control.

Detached flower assay A detached flower assay using fresh Golden Delicious blossoms from two-year-old apple trees was carried out to monitor virulence of E. amylovora mutants (66). Blossoms were treated with either E. amylovora CFBP1430, E. amylovora CFBP1430 [pBAD18] or PBS as mock infection. Racks were cleaned and autoclaved before the experiment and 24 wells per rack were filled with 2 ml H2O. The wells were sealed with scotch tape, which was perforated using a syringe. Blossom stems were freshly cut to ensure water uptake before being transferred through the

74 Phage infection of E. amylovora requires cellulose scotch tape. Bacteria grown overnight on plates were scratched off and carefully resuspended

7 in PBS. OD600 was adjusted to 1.0 and a 1:50 dilution performed (approx.10 CFU/ml). The blossoms were inoculated by pipetting 20 µl bacterial suspension or PBS directly onto the receptacle. Each storage box (5 l) was laid out with paper towels and filled with three racks.

To ensure humidity, 100 ml H2O were added per box. The blossoms were stored at 26°C for 4- 5 days. The read out was performed according to an adjusted rating system (67). Healthy blossoms without disease symptoms were classified as grade 1. Visible symptoms on the blossom (browning of the calix) were referred to as grade 2. Blossoms with infection spreading from the calix to the stipe of the blossoms corresponded to grade 3.

Soft agar overlays with Congo Red Four millilitres molten LB+ soft agar supplemented with different concentrations of Congo Red (10, 20, 40 µg/ml) and Coomassie Brilliant Blue (5, 10, 20 µg/ml) were mixed with 90 µl bacterial overnight culture before pouring on to LB plates with the corresponding amount of Congo Red and Coomassie Brilliant Blue. Phage dilution rows was spotted (5 µl each) on to the dried soft agar and incubated overnight. L1 and Y2 were used as negative control.

In vitro infection assays with Congo Red Overnight cultures of E. amylovora CFBP1430 were washed twice in sterile SM buffer and

7 OD600 was adjusted to reach approximately 10 CFU/ml. Subsequently, 20 µl of the washed cells were transferred to 1960 µl LB+ broth and supplemented with either 20 µl sterile SM buffer or with 20 µl bacteriophage (1010 PFU/ml). The mixtures were then added to 96 well flat bottom plates and incubated at 25°C with double orbital shaking for 30 h in a plate reader.

OD600 measurements were carried out every half hour. For experiments involving the cellulose binding dye Congo Red, different concentrations of Congo Red (10, 20, 40 µg/ml) and Coomassie Brilliant Blue (5, 10, 20 µg/ml) were added.

Cellulase functionality To control the functionality of the putative cellulases of the phages M7 and S6, a DNS (Dinitrosalicylic acid) test was performed. The experiment measured the amount of released glucose after incubation of cellulose with different cellulases. A modified protocol of Ghose

1987 was used (68). The substrate Carboxymethylcellulose sodium salt (CMC) or H2O as control as well as the four phages M7, S6, L1 and Y2 (all 5x109 PFU/ml) as samples and the cellulase from Trichoderma reesei (aqueous solution, >700 units/g Sigma Aldrich, Switzerland) as positive control, were prewarmed to 40°C. A total of 250 µl substrate were combined with 250 µl cellulase sample in a 2 ml test tube and mixed well. The tubes were subsequently incubated at 40°C for exactly 10 min. Afterwards, 1 ml of DNS solution (DNS

32.85 mM, NaOH 0.3 M, potassium sodium tartrate 1.05 M, lactose 87.5 mM in H2O) was

75 Manuscript II added to each tube and the reaction was stopped by boiling the samples at 100°C for 15 min. After cooling to room temperature, the samples were centrifuged at 3,000 xg for 4 min. Finally, the optical density was measured at 540 nm. A standard curve with known quantities of glucose (5, 10, 15 µmol/l) was performed. The glucose samples were mixed with 250 µl H2O and treated according to the protocol described above. The obtained values were compared to the standard curve to identify the total amount of released glucose.

76 Phage infection of E. amylovora requires cellulose

Results

The bacterial cellulose synthase operon is essential for infection by M7 and S6 A Tn5 transposon mutant library of E. amylovora CFBP1430 was screened previously for mutants with the ability to resist M7 and S6 infection (chapter 1). One operon was particularly targeted in the screens of both M7 and S6. M7 was unable to lyse bacteria with transposon insertions in the genes EAMY3608 and wssA renamed into bcsO and bcsQ, respectively (69), bcsA, bcsB and bcsC. S6 resistant transposon mutants revealed insertions in the genes bcsO, bcsQ, bcsA, bcsB, bcsC and celA3 renamed into bcsZ. All these genes belong to the bacterial cellulose synthesis (bcs) operon (Figure 1). The only gene not targeted in this operon was bcsD. In addition to the bcs operon, M7 and S6 were both unable to infect a mutant with a disrupted hmsT gene. hmsT is annotated to encode a diguanylate cyclase and could be involved in the production of cyclic di-GMP a second messenger that induces cellulose production. All genes but hmsT were disrupted multiple times at independent sites reducing the risk of bias.

Figure 1 Schematic overview of the bacterial cellulose synthase operon (bcs). A. Overview of the bcs operon in E. amylovora CFBP1430. Red arrows indicate independent Tn5 insertion events identified during the screening process. B. Predicted localization of the encoded Bcs components.

Due to the insensitivity of these mutants against the phages M7 and S6, knockout mutants were generated. The most promising gene for phage adsorption is bcsC since the encoded protein is localized in the outer membrane where it could be targeted by the phages as receptor (ΔbcsC). In order to establish if the phages require cellulose for infection, a deletion of the catalytic subunit BcsA was produced (ΔbcsA). A bcsZ knockout was generated (ΔbcsZ)

77 Manuscript II since M7, but not S6, was able to infect the transposon mutants (chapter 1). Additionally, a deletion of the complete bcs operon was generated (Δbcs). To verify if BcsC is present in the generated mutants, western blots were carried out using polyclonal Anti-BcsC antibodies (data not shown). The deletions of bcsA, bcsC or the entire operon rendered the mutants phage- insensitive (Figure 2). Deletion of bcsZ however, only affected S6 infection. The mutants were complemented in trans. Both phages were able to successfully lyse the complemented and overexpressed mutants.

Figure 2 Plaque assay of M7 and S6 on all bcs mutants. Dilution rows of M7 and S6 (approx. 1012 PFU/ml) were spotted on mutant lawns on LB (left), LB supplemented with glucose (middle) or arabinose (right).

78 Phage infection of E. amylovora requires cellulose

Complementing the operon deletion mutant with bcsC (Δbcs+bcsC), however, could not restore the phage sensitive status. The hmsT gene could not be deleted. The generated mutants were also tested for their growth capacities. Growth comparisons were generated in LB medium and minimal medium MM2. No significant difference in growth capacity was observed (data not shown).

Lack of cellulose synthesis does not attenuate E. amylovora virulence on detached flowers The Bcs complex is responsible for cellulose production. Hence, the mutants’ ability to produce cellulose was evaluated with the dye Congo Red, which binds to 1,4-β-glucose molecules. Cellulose producing strains grow as pink colonies, whereas cells lacking cellulose appear as white colonies (Figure 3). The mutants lacking the enzymatically active subunit BcsA were not able to produce cellulose and appeared as white colonies. The bcsC knockout mutant was observed to grow as light rose-coloured colony. Since the cellulose producing subunit BcsA is present, cellulose is produced but not secreted in the absence of BcsC. ΔbcsZ colonies were observed to be light pink. Complementations of the knocked-out genes were evaluated. When arabinose was added, induction of the pBAD promoter resulted in restoration of the pink colony morphology.

Figure 3 Cellulose production monitored through Congo Red. A total of 5 µl overnight culture of each bcs mutant were spotted on LB plates supplemented with Congo Red and Coomassie Brilliant Blue (left), 0.2% glucose (middle) or 0.2% arabinose (right). Cellulose producing strains grow with a pink phenotype whereas cellulose lacking strains grow as white colonies.

Possible effects on the EPS production of the generated mutation in the E. amylovora CFBP1430 background were evaluated by EPS-CPC precipitations (65). The low EPS producing strain 4/82 was used as control (45), along with CFBP1430 and CFBP1430 [pBAD18] (Figure 4). The generated mutants Δbcs, ΔbcsA, ΔbcsC and ΔbcsZ as well as the complementations obtained comparable results to the control. Additionally, the overexpression of hmsT, bcsA and bcsC generated similar levels of EPS as the controls. These findings suggest that modifications in the bcs operon do not affect production of amylovoran and levan.

79 Manuscript II

Figure 4 EPS measurement in bcs mutants. Optical density of EPS CPC precipitation was measured at 600 nm. The low EPS producing strain E. amylovora 4/82 was used as negative control. Each experiment was carried out three times for each mutant. Error bars indicate standard deviation.

Virulence of the generated knockout mutants was tested for variation (Figure 5). The wildtype (wt) strain and the strain carrying an empty vector were used as controls. Both strains were potent in infecting the blossoms and generated overall 91% and 92% grade 3 disease symptoms (infection symptoms visible in the blossoms and stipe), respectively. Comparable results were obtained when blossoms were infected with Δbcs, Δbcs+bcsC, ΔbcsA, ΔbcsA+bcsA, ΔbcsC and ΔbcsC+bcsC. The infection with the two mutants ΔbcsZ and ΔbcsZ+bcsZ generated less grade 3 disease symptoms on infected blossoms namely 66% and 72%, respectively. The overexpression of hmsT, which positively regulates the bacterial cellulose synthase through cyclic-di-GMP production, generated 88% of grade 3 blossoms when applied. These findings indicate that virulence on detached flowers is not affected by the modifications in the bcs operon.

Figure 5 Virulence of bcs mutants tested in detached apple blossoms. Blossoms were infected with approx. 107 CFU/ml bacteria and incubated for 4 days. Read out was performed by evaluating disease symptoms. Grade 1 indicates healthy blossoms. Grade 2 refers to blossoms with first disease symptoms such as browning of the calix. Blossoms showing advanced disease symptoms on calyx and stipe correspond to Grade 3. Each mutant was tested at least on 16 individual blossoms.

80 Phage infection of E. amylovora requires cellulose

Congo Red inhibits phage attachment and infection Phage dilutions were spotted on a CFBP1430 bacterial lawn on LB plates supplemented with Congo Red. In contrast to CFBP1430 grown on LB plates, M7 and S6 were not able to infect the wildtype on LB plates with the cellulose binding dye Congo Red supplemented in the soft agar (Figure 6). When Congo Red was added at lower concentrations to the soft agar, the protective effect of Congo Red against M7 and S6 was no longer as pronounced. To verify if Congo Red generally blocks phage infection by unspecifically binding to surface structures that would obstruct host receptor binding, the phages L1 and Y2 were used as control. Neither L1 (T7-virus; podovirus) nor Y2 (GJ1-virus; myovirus) supposedly recognizing amylovoran or LPS, respectively (45, 48), were affected by the dye. These results indicate that by specifically binding to cellulose, Congo Red blocks phage infection of M7 and S6.

Figure 6 Dose dependent Congo Red protection against S6 and M7 infection. Left panel: Dilution rows of M7 and S6 (1012 PFU/ml), Right panel: L1 (2x1010 PFU/ml) and Y2 (3x109 PFU/ml). E. amylovora CFBP1430 bacterial lawns were grown on A. LB only, B. LB supplemented with 40 µg/ml Congo Red and 20 µg/ml Coomassie Brilliant Blue, C. LB supplemented with 20 µg/ml Congo Red and 10 µg/ml Coomassie Brilliant Blue. The E. amylovora specific podovirus L1 and the myovirus Y2 were used as controls

The ability of M7 and S6 to adsorb to the surface of E. amylovora was tested in pull down experiments. The assay was carried out in different settings and with various buffers and media. All the tested settings resulted in ambiguous results with lower phage levels in the media/buffer control (data not shown). These results suggests pull down experiments are not feasible for the two phages M7 and S6.

81 Manuscript II

In vitro infections of wildtype bacteria and Δbcs using both phages were carried out in LB medium supplemented with Congo Red. The addition of S6 to the wildtype in LB medium revealed a delaying growth effect of 5 hours before growth was detectable in the plate reader (Figure 7). This delaying effect was no longer visible when Congo Red was added, suggesting a loss of S6 infectivity. Neither M7 nor S6 were able to affect the growth of the Δbcs mutant with or without the supplemented Congo Red. In contrast to S6, M7 was able to delay measurable growth of the wildtype for up to 40 hours. The addition of Congo Red revoked this effect and wildtype bacteria grew at a comparable rate to the sample without phage addition. Growth of the Δbcs mutant was unaffected by both the phage and Congo Red supplementation. To verify if Congo Red affects the phages directly, phages were recovered after the experiment, washed and a dilution row of the samples was spotted on an E. amylovora CFBP1430 lawn on LB. Both phages were able to infect the wildtype indicating no effect of Congo Red on the phages.

Figure 7 In vitro infection through M7, S6 and the influence of Congo Red. Growth curve of E. amylovora CFBP1430 (top) and Δbcs (bottom) infected with M7 (left) and S6 (right). Bacterial concentrations of 105 CFU/ml were infected with 108 PFU/ml phage and incubated with Congo Red 20 µg/ml and Coomassie Brilliant Blue 10

µg/ml over 30 hours. OD600 was measured regularly at 30 minute intervals. Error bars indicate standard deviation.

82 Phage infection of E. amylovora requires cellulose

Genome analysis revealed that both phages encode proteins with potential cellulolytic abilities. In order to estimate the functionality of these enzymes, the phages ability to degrade cellulose was tested by a DNS-assay. Whole phages were incubated directly with cellulose to observe enzyme activity. The phages L1 and Y2 do not encode cellulose-degrading enzymes and were therefore used as negative controls. The cellulase from Trichoderma reesei was used as positive control. In contrast to M7, phage S6 was able to degrade cellulose to glucose molecules (Figure 8). A total of 1.4 µmol/l glucose within 10 min were generated by S6.

Figure 8 Cellulase activity of whole phages. Potential cellulase activity of the phages M7 and S6 was tested. The amount of released glucose after digestion of CMC by the phage cellulases was measured using the colorimetric shift of DNS upon encounter of reducing sugars. Two dilutions of the cellulase of Trichoderma reesei (0.2 units; 0.02 units) were used as positive control, the two phages L1 and Y2 as negative controls since neither of the two phages encodes cellulases. The phage stocks were diluted to reach approx. 1010 PFU/ml before incubation with CMC. Error bars indicated the standard deviation of three independent experiments.

83 Manuscript II

Discussion Identification of a specific phage receptor is a crucial step in elucidating the infection mechanism of a particular phage, evaluating the potency of this phage against the targeted pathogen and anticipating phage resistance development. In a previous study, a mutant library of E. amylovora was screened to reveal genes responsible for rendering mutants phage-insensitive. The localization of the transposon insertions revealed that both phages M7 and S6 were unable to infect mutants with disruptions in the bacterial cellulose synthase (bcs) operon. This constitutively active operon encodes the bacterial cellulose synthase (Bcs). It is composed of several subunits and is required for the production and secretion of bacterial cellulose (70). The function of the first gene in this operon, bcsO (EAMY_2231) is yet unknown (10, 71). BcsQ (wssA) is suggested to be crucial for localizing the Bcs complex at the cell poles (69). Located in the inner membrane, the catalytically active subunit BcsA contains a large glycosyltransferase domain, generating the characteristic β-1,4-glycosidic linkage in cellobiose (70). BcsB is co-localized with BcsA and ensures BcsA stabilization (72). Cellulose strands are crystallized by the periplasmic protein BcsD and exported through the outer membrane porin BcsC (71, 73). The gene bcsZ encodes a putative endoglucanase and is involved in cellulose regulation (10). Essential for cellulose production are the subunits BcsQ, BcsA, BcsB, BcsC and BcsZ. Cellulose production and biofilm formation are tightly regulated. The intracellular second messenger cyclic di-GMP (c- di-GMP) is involved in regulating planktonic and sessile states of bacteria. High levels of intracellular c-di-GMP can activate biofilm production by inducing the cellulase synthesis complex (11). The enzymatically active subunit BcsA binds c-di-GMP upon which the glycosyltransferase domain starts linking UDP-glucose building blocks (71, 74). Production of c-di-GMP is performed by diguanylate cyclase after induction by environmental signals (75). In addition to the bcs disrupted mutants, a mutant with an insertion into the hmsT gene was revealed to have a phage resistant phenotype. The gene hmsT harbours a particular GGDEF domain indicating a diguanylate cyclase activity. An overexpression of hmsT generated deep pink colonies, suggesting augmented cellulose output in this mutant, thereby linking HmsT function to cellulose production. Bacterial cellulose was recently found to be produced by E. amylovora during the infection of susceptible plant tissues in the apoplast (10). Cellulose is suggested to be essential for the maintenance of biofilm integrity and stability in E. coli (76) and could have the same function in fire blight infection. Castiblanco et al. proposed that cellulose synthesis is critical to form a mature biofilm, in order to withstand disturbances generated by the xylem-sap flow (10). Nevertheless, a defective cellulose synthesis operon did not impair virulence in the constructed mutants. It is therefore suggested that although cellulose is key for biofilm stabilization in the plant tissue, bacterial virulence is mostly amylovoran dependent.

84 Phage infection of E. amylovora requires cellulose

Even though cellulose is not responsible for virulence, it seems crucial for phage infection by M7 and S6. Neither of the two phages are able to infect mutants lacking the cellulose producing subunit BcsA, the outer membrane protein BcsC or mutants without the entire bcs operon. The effect of cellulose on the two phages was monitored with Congo Red. The dye appears to have a blocking effect on the phages, since phage infection of the host was completely abolished when incubated on Congo Red agar. Lower Congo Red concentrations could partially restore the infectivity of the phages but with lower efficacy of plating suggesting a concentration dependent competition of Congo Red and the phages for binding. By binding cellulose, Congo Red blocks phage infection of M7 and S6. This result therefore clearly links M7 and S6 infection to cellulose recognition. Data generated for both of the phages in this and in a previous study (47) suggest that although both phages strongly rely on the cellulose synthase and cellulose, they do not recognize the exact same structure on the host surface. Discrepancy in the infectivity of M7 and S6 on the bcsZ knockout, differences in host range and varying potency when incorporated in a phage cocktail suggest different modes of recognition and infection. Bertozzi et al. analysed the preference of Caudovirales for specific receptor types and compositions (53). Myoviruses were shown to bind to proteinaceous receptors (mainly outer membrane proteins, n=6), combinations of protein and sugar moieties (n=6) as well as sugar moieties of polysaccharides (e.g. cell wall composing sugars, n=14) with a tendency to the latter. Receptors targeted by podoviruses were all polysaccharides (n=17). Although the number of analysed phage-receptor pairs was rather small, the exclusivity of podoviruses to bind to polysaccharide structures is striking. Another study from Roach et al. investigated the impact of EPS levels in E. amylovora on phage infection. The tested podoviruses were shown to be highly infectious towards high EPS producing hosts whereas the tested myoviruses appeared to be better adapted to infecting low EPS producing strains (77). Considering these results, this could indicate that S6 indeed could recognize cellulose whereas M7, in addition to cellulose, could identify the host cell through BcsC or another receptor. A majority of phages currently under investigation require two independent receptors for successful infection. A primary receptor, to which the phage particles bind with their long tail fibres, reversibly triggers a conformational change revealing the short tail fibres. These short fibres subsequently recognize and irreversibly bind the secondary receptor, which induces the infection process (78). Cellulose could act as such a first receptor for M7 inducing a conformational change whereupon the actual receptor-binding domain could be uncovered hence ensuring host specificity. Nevertheless, the initial question of the phage specific receptors remains. Although the transposon mutagenesis revealed a key function of the bcs operon in the infection of M7 and S6, bacteria lacking the enzymatically active subunit BcsA, where cellulose production is

85 Manuscript II abolished but BcsC is still incorporated, were not affected by the phages. This emphasizes a crucial role of cellulose for successful phage infection. Pull down experiments would provide an insight into the binding affinity of the phages towards the generated mutants. However, even though various settings were tested, pull down experiments with M7 and S6 showed poor reproducibility (data not shown). The negative control of phages incubated without bacteria, revealed a loss of phages in particular, which is as of yet unexplained. In vitro infection assays offer an alternative. Although they are not able to detect actual adsorption to the host cell, the experiment can show successful recognition and infection of bacteria. Genome analysis of M7 and S6 revealed a collection of different genes encoding putative endoglucanases, cellulases or cellobiosidases. Indeed, it was possible to show that entire S6 phages are able to degrade cellulose. The fact that phages could harbour functional cellulose- degrading enzymes is a novelty to our knowledge. Phages encoding degrading enzymes, such as depolymerases, are widely distributed and analysed for their enzymatic activity. Such enzymes can give phages an advantage, especially when their hosts are prone to biofilm production or capsule formation. This capsule can obstruct the access of phages to the host surface. Certain phages not only encode EPS degrading enzymes, but even have the ability to use these layer as receptors (45, 53, 79, 80). M7 and S6 could either use cellulose directly as receptor or by decomposing the cellulose layer could reach the bacterial surface and encounter a secondary receptor that triggers host infection. Infection assays carried out by Born et al. revealed that both phages only weakly affect bacterial growth over a period of 8 hours when viable cell counts were measured in vitro (47). This study involved a longer monitoring period. Both phages were able to prevent E. amylovora growth, in the case of M7 for up to 24 hours. These findings suggest that M7 and S6 rather control and stabilize bacterial growth instead of rapidly eradicating their host bacteria, which could lead to fast resistance development. Fujiwara and colleagues obtained similar findings for the Ralstonia solanacearum phage RSL1 (81). Instead of quickly killing all host bacteria, RSL1 kept the cell density at a low level. Tomato plants pre-treated with the phage could survive up to 4 months after R. solanacearum infection. They therefore proposed a second method for biocontrol, wherein a phage maintaining host-phage coexistence protects a susceptible plant. Our findings suggest that M7 and S6 could establish such an equilibrium between bacterial growth and host lysis. Considering the importance of cellulose for M7 and S6 infection, we hypothesize that cellulose could be key for the observed co-existence of host and phage. Cellulose production is tightly regulated and induced in the presence of high cell counts (82, 83). High bacterial density will initiate cellulose secretion upon which the phages are able to attack the bacteria. The resulted lysis of cellulose forming bacteria will reduce the total number of host cells thereby halting cellulose production. In the absence of cellulose, the

86 Phage infection of E. amylovora requires cellulose phages are unable to recognize the host bacteria permitting a recovery of the bacterial population. Regrowth will occur until cell density favours biofilm assembly anew, whereupon phages are again able to infect host cells. This scenario would thereby ensure co-existence of phage and host cells. The results obtained from this study could show that the phages M7 and S6 heavily rely on the bacterial cellulose synthase operon and cellulose for infection. Both phages are not able to lyse mutants deficient in cellulose production. The identification of encoded putative cellulases in the phage genomes further substantiates the crucial role of cellulose in the phage infection process. We further propose balanced co-existence of phages and their host in vitro. We therefore suggest that both M7 and S6 have a high potential in controlling fire blight.

Acknowledgments

This work was funded by the Swiss National Science Foundation (SNF) grant 310030_156947.

87 Manuscript II

References

1. Bonn WG, van der Zwet T. 2000. Distribution and economic importance of fire blight, p. 37–53. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

2. Vanneste JL. 2000. What is fire blight? Who is Erwinia amylovora? How to control it?, p. 1–6. In Fire Blight. CAB International, Wallingford.

3. Gill JJ, Svircev AM, Smith R, Castle AJ. 2003. Bacteriophages of Erwinia amylovora. Appl Environ Microbiol 69:2133–2138.

4. Momol MT, Aldwinckle HS. 2000. Genetic diversity and host range of Erwinia amylovora, p. 55–72. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

5. Thomson S V. 2000. Epidemiology of fire blight, p. 9–36. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

6. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald P, Dow M, Verdier V, Beer S V., Machado MA, Toth I, Salmond G, Foster GD. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol 13:614–629.

7. Ayers AR, Ayers SB, Goodman RN. 1979. Extracellular polysaccharide of Erwinia amylovora: a correlation with virulence. Appl Environ Microbiol 38:659–666.

8. Koczan JM, McGrath MJ, Zhao Y, Sundin GW. 2009. Contribution of Erwinia amylovora exopolysaccharides amylovoran and levan to biofilm formation: implications in pathogenicity. Phytopathology 99:1237–1244.

9. Gross M, Geier G, Rudolph K, Geider K. 1992. Levan and levansucrase synthesized by the fireblight pathogen Erwinia amylovora. Physiol Mol Plant Pathol 40:371–381.

10. Castiblanco LF, Sundin GW. 2016. Cellulose production, activated by cyclic di-GMP through BcsA and BcsZ, is a virulence factor and an essential determinant of the three- dimensional architectures of biofilms formed by Erwinia amylovora Ea1189. Mol Plant Pathol 19:90–103.

11. Edmunds AC, Castiblanco LF, Sundin GW, Waters CM. 2013. Cyclic di-GMP modulates the disease progression of Erwinia amylovora. J Bacteriol 195:2155–65.

12. Kharadi RR, Castiblanco LF, Waters CM, Sundin GW. 2019. Phosphodiesterase genes regulate amylovoran production, biofilm formation, and virulence in Erwinia amylovora. Appl Environ Microbiol 85:e02233-18.

13. Maes M, Orye K, Bobev S, Devreese B, Van Beeumen J, De Bruyn A, Busson R, Herdewijn P, Morreel K, Messens E. 2001. Influence of amylovoran production on virulence of Erwinia amylovora and a different amylovoran structure in E. amylovora isolates from Rubus. Eur J Plant Pathol 107:839–844.

14. Wei Z, Kim JF, Beer S V. 2000. Regulation of hrp genes and type III protein secretion in Erwinia amylovora by HrpX/HrpY, a novel two-component system, and HrpS. Mol Plant-Microbe Interact 13:1251–1262.

88 Phage infection of E. amylovora requires cellulose

15. Geider K. 2000. Exopolysaccharides of Erwinia amylovora: structure, biosynthesis, regulation, role in pathogenicity of amylovoran and levan, p. 117–140. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

16. Koczan JM, Lenneman BR, McGrath MJ, Sundin GW. 2011. Cell surface attachment structures contribute to biofilm formation and xylem colonization by Erwinia amylovora. Appl Environ Microbiol 77:7031–7039.

17. Vanneste JL, Eden-Green S. 2000. Migration of Erwinia amylovora in host plant tissues, p. 73–83. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

18. Jones AL, Schnabel EL. 2000. The development of streptomycin-resistant strains of Erwinia amylovora, p. 235–251. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

19. Piqué N, Miñana-Galbis D, Merino S, Tomás JM. 2015. Virulence factors of Erwinia amylovora: A review. Int J Mol Sci 16:12836–54.

20. Russo NL, Burr TJ, Breth DI, Aldwinckle HS. 2008. Isolation of streptomycin-resistant isolates of Erwinia amylovora in New York. Plant Dis 92:714–718.

21. Hatfull GF. 2008. Bacteriophage genomics. Curr Opin Microbiol 11:447–453.

22. Cobián Güemes AG, Youle M, Cantú VA, Felts B, Nulton J, Rohwer F. 2016. Viruses as winners in the game of life. Annu Rev Virol 3:197–214.

23. Casjens SR. 2005. Comparative genomics and evolution of the tailed-bacteriophages. Curr Opin Microbiol 8:451–458.

24. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell 108:13–6.

25. Loc-Carrillo C, Abedon ST. 2011. Pros and cons of phage therapy. Bacteriophage 1:111–114.

26. Buttimer C, McAuliffe O, Ross RP, Hill C, O’Mahony J, Coffey A, O’Mahony J, Coffey A. 2017. Bacteriophages and bacterial plant diseases. Front Microbiol 8:34.

27. Hagens S, Loessner MJ. 2010. Bacteriophage for biocontrol of foodborne pathogens: calculations and considerations. Curr Pharm Biotechnol 11:58–68.

28. Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JA, Momol MT, Jones JB, Vallad GE. 2012. Soil-based systemic delivery and phyllosphere in vivo propagation of bacteriophages: Two possible strategies for improving bacteriophage persistence for plant disease control. Bacteriophage 2:215–224.

29. Jones JB, Vallad GE, Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JC, Momol MT. 2012. Considerations for using bacteriophages for plant disease control. Bacteriophage 2:208–214.

30. Born Y, Bosshard L, Duffy B, Loessner MJ, Fieseler L. Protection of Erwinia amylovora bacteriophage Y2 from UV-induced damage by natural compounds. Bacteriophage 5:e1074330.

89 Manuscript II

31. Marston MF, Pierciey FJ, Shepard A, Gearin G, Qi J, Yandava C, Schuster SC, Henn MR, Martiny JBH. 2012. Rapid diversification of coevolving marine Synechococcus and a virus. Proc Natl Acad Sci 109:4544–4549.

32. Buckling A, Rainey PB. 2002. Antagonistic coevolution between a bacterium and a bacteriophage. Proc R Soc B Biol Sci 269:931–936.

33. Svircev A, Roach D, Castle A. 2018. Framing the future with bacteriophages in agriculture. Viruses 10:218.

34. Czajkowski R, Ozymko Z, de Jager V, Siwinska J, Smolarska A, Ossowicki A, Narajczyk M, Lojkowska E. 2015. Genomic, proteomic and morphological characterization of two novel broad host lytic bacteriophages ΦPD10.3 and ΦPD23.1 infecting pectinolytic Pectobacterium spp. and Dickeya spp. PLoS One 10:e0119812.

35. Lim J-A, Jee S, Lee DH, Roh E, Jung K, Oh C, Heu S. 2013. Biocontrol of Pectobacterium carotovorum subsp. carotovorum using bacteriophage PP1. J Microbiol Biotechnol 23:1147–1153.

36. Adriaenssens EM, Van Vaerenbergh J, Vandenheuvel D, Dunon V, Ceyssens P-J, De Proft M, Kropinski AM, Noben J-P, Maes M, Lavigne R. 2012. T4-related bacteriophage LIMEstone isolates for the control of soft rot on potato caused by Dickeya solani. PLoS One 7:e33227.

37. Balogh B, Jones JB, Momol MT, Olson SM, Obradovic A, King P, Jackson LE. 2003. Improved efficacy of newly formulated bacteriophages for management of bacterial spot on tomato. Plant Dis 87:949–954.

38. Obradovic A, Jones JB. 2004. Management of tomato bacterial spot in the field by foliar applications of bacteriophages and SAR inducers. Plant Dis 88:736–740.

39. Das M, Bhowmick TS, Ahern SJ, Young R, Gonzalez CF. 2015. Control of pierce’s disease by phage. PLoS One 10:e0128902.

40. Svircev AM, Castle AJ, Lehman SM. 2010. Bacteriophages for control of phytopathogens in food production systems, p. 79–96. In Sabour, PM, Griffiths, MW (eds.), Bacteriophages in the control of food- and waterborne pathogens. ASM Press, Washington, DC.

41. Boulé J, Sholberg PL, Lehman SM, O’Gorman DT, Svircev AM. 2011. Isolation and characterization of eight bacteriophages infecting Erwinia amylovora and their potential as biological control agents in British Columbia, Canada. Can J Plant Pathol 33:308– 317.

42. Lehman SM. 2007. Ph.D. thesis: Development of a bacteriophage-based biopesticide for fire blight. Department of Biological Sciences, Brock University, St. Catharines.

43. Chadha P, Katare OP, Chhibber S. 2016. In vivo efficacy of single phage versus phage cocktail in resolving burn wound infection in BALB/c mice. Microb Pathog 99:68–77.

44. Naghizadeh M, Karimi Torshizi MA, Rahimi S, Dalgaard TS. 2018. Synergistic effect of phage therapy using a cocktail rather than a single phage in the control of severe colibacillosis in quails. Poult Sci 98:653–663.

45. Born Y, Fieseler L, Klumpp J, Eugster MR, Zurfluh K, Duffy B, Loessner MJ. 2014. The

90 Phage infection of E. amylovora requires cellulose

tail-associated depolymerase of Erwinia amylovora phage L1 mediates host cell adsorption and enzymatic capsule removal, which can enhance infection by other phage. Environ Microbiol 16:2168–2180.

46. Schmerer M, Molineux IJ, Bull JJ. 2014. Synergy as a rationale for phage therapy using phage cocktails. PeerJ 2:e590.

47. Born Y, Fieseler L, Marazzi J, Lurz R, Duffy B, Loessner MJ. 2011. Novel virulent and broad-host-range Erwinia amylovora bacteriophages reveal a high degree of mosaicism and a relationship to Enterobacteriaceae phages. Appl Environ Microbiol 77:5945– 5954.

48. Born Y, Fieseler L, Thöny V, Leimer N, Duffy B, Loessner MJ. 2017. Engineering of bacteriophages Y2::dpoL1-C and Y2::luxAB for efficient control and rapid detection of the fire blight pathogen, Erwinia amylovora. Appl Environ Microbiol 83:e00341-17.

49. Marti R, Zurfluh K, Hagens S, Pianezzi J, Klumpp J, Loessner MJ. 2013. Long tail fibres of the novel broad-host-range T-even bacteriophage S16 specifically recognize Salmonella OmpC. Mol Microbiol 87:818–834.

50. Fehmel F, Feige U, Niemann H, Stirm S. 1975. Escherichia coli capsule bacteriophages. VII. Bacteriophage 29-host capsular polysaccharide interactions. J Virol 16:591–601.

51. Guerrero-Ferreira RC, Viollier PH, Ely B, Poindexter JS, Georgieva M, Jensen GJ, Wright ER. 2011. Alternative mechanism for bacteriophage adsorption to the motile bacterium Caulobacter crescentus. Proc Natl Acad Sci U S A 108:9963–8.

52. Shin H, Lee JH, Kim H, Choi Y, Heu S, Ryu S. 2012. Receptor diversity and host interaction of bacteriophages infecting Salmonella enterica serovar Typhimurium. PLoS One 7:e43392.

53. Bertozzi Silva J, Storms Z, Sauvageau D. 2016. Host receptors for bacteriophage adsorption. FEMS Microbiol Lett 363:fnw002.

54. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistance mechanisms. Nat Rev Microbiol 8:317–327.

55. Schnabel EL, Fernando WGD, Meyer MP, Jones AL, Jackson LE. 1999. Bacteriophage of Erwinia amylovora and their potential for biocontrol. ISHS Acta Hortic 489:649–653.

56. Chan BK, Abedon ST, Loc-Carrillo C. 2013. Phage cocktails and the future of phage therapy. Future Microbiol 8:769–783.

57. Kazmierczak KM, Rothman-Denes LB. 2006. Bacteriophage N4, p. 302–314. In Calendar, R (ed.), The bacteriophages. Oxford University Press, New York, NY.

58. Adams MH. 1959. Bacteriophages. Interscience Publishers Inc., New York.

59. Sambrook J, Russell DW. 2001. Molecular cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

60. Skorupski K, Taylor RK. 1996. Positive selection vectors for allelic exchange. Gene 169:47–52.

91 Manuscript II

61. Dahan S, Wiles S, La Ragione RM, Best A, Woodward MJ, Stevens MP, Shaw RK, Chong Y, Knutton S, Phillips A, Frankel G. 2005. EspJ is a prophage-carried type III effector protein of attaching and effacing pathogens that modulates infection dynamics. Infect Immun 73:679–86.

62. Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison Iii CA, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343.

63. Guzman LLM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J Bacteriol 177:4121–4130.

64. Mruk DD, Cheng CY. 2011. Enhanced chemiluminescence (ECL) for routine immunoblotting. Spermatogenesis 1:121–122.

65. Bellemann P, Bereswill S, Berger S, Geider K. 1994. Visualization of capsule formation by Erwinia amylovora and assays to determine amylovoran synthesis. Int J Biol Macromol 16:290–296.

66. Pusey PL. 1997. Crab apple blossoms as a model for research on biological control of fire blight. Phytopathology 87:1096–102.

67. Llop P, Cabrefiga J, Smits THM, Dreo T, Barbé S, Pulawska J, Bultreys A, Blom J, Duffy B, Montesinos E, López MM. 2011. Erwinia amylovora novel plasmid pEI70: complete sequence, biogeography, and role in aggressiveness in the fire blight phytopathogen. PLoS One 6:e28651.

68. Ghose TK. 1987. Measurement of cellulase activities.Pure &AppI. Chem.

69. Le Quéré B, Ghigo J-M. 2009. BcsQ is an essential component of the Escherichia coli cellulose biosynthesis apparatus that localizes at the bacterial cell pole. Mol Microbiol 72:724–40.

70. Omadjela O, Narahari A, Strumillo J, Mélida H, Mazur O, Bulone V, Zimmer J. 2013. BcsA and BcsB form the catalytically active core of bacterial cellulose synthase sufficient for in vitro cellulose synthesis. Proc Natl Acad Sci U S A 110:17856–61.

71. Römling U, Galperin MY. 2015. Bacterial cellulose biosynthesis: diversity of operons, subunits, products, and functions. Trends Microbiol 23:545–57.

72. Morgan JLW, Strumillo J, Zimmer J. 2013. Crystallographic snapshot of cellulose synthesis and membrane translocation. Nature 493:181–6.

73. Whitney JC, Hay ID, Li C, Eckford PDW, Robinson H, Amaya MF, Wood LF, Ohman DE, Bear CE, Rehm BH, Howell PL. 2011. Structural basis for alginate secretion across the bacterial outer membrane. Proc Natl Acad Sci U S A 108:13083–8.

74. Morgan JLW, McNamara JT, Zimmer J. 2014. Mechanism of activation of bacterial cellulose synthase by cyclic di-GMP. Nat Struct Mol Biol 21:489–96.

75. Ren G-X, Fan S, Guo X-P, Chen S, Sun Y-C. 2016. Differential regulation of c-di-GMP metabolic enzymes by environmental signals modulates biofilm formation in Yersinia pestis. Front Microbiol 7:821.

92 Phage infection of E. amylovora requires cellulose

76. Hung C, Zhou Y, Pinkner JS, Dodson KW, Crowley JR, Heuser J, Chapman MR, Hadjifrangiskou M, Henderson JP, Hultgren SJ. 2013. Escherichia coli biofilms have an organized and complex extracellular matrix structure. MBio 4:e00645-13.

77. Roach DR, Sjaarda DR, Castle AJ, Svircev AM. 2013. Host exopolysaccharide quantity and composition impact Erwinia amylovora bacteriophage pathogenesis. Appl Environ Microbiol 79:3249–56.

78. Crowther RA, Lenk E V., Kikuchi Y, King J. 1977. Molecular reorganization in the hexagon to star transition of the baseplate of bacteriophage T4. J Mol Biol 116:489– 523.

79. Lindberg AA. 1973. Bacteriophage receptors. Annu Rev Microbiol 27:205–241.

80. Rakhuba D V., Kolomiets EI, Dey ES, Novik GI, Szwajcer Dey E, Novik GI. 2010. Bacteriophage receptors, mechanisms of phage adsorption and penetration into host cell. Polish J Microbiol 59:145–155.

81. Fujiwara A, Fujisawa M, Hamasaki R, Kawasaki T, Fujie M, Yamada T. 2011. Biocontrol of Ralstonia solanacearum by treatment with lytic bacteriophages. Appl Environ Microbiol 77:4155–4162.

82. Lamas A, Miranda JM, Vázquez B, Cepeda A, Franco CM. 2016. Biofilm formation, phenotypic production of cellulose and gene expression in Salmonella enterica decrease under anaerobic conditions. Int J Food Microbiol 238:63–67.

83. Bai AJ, Rai VR. 2011. Bacterial quorum sensing and food industry. Compr Rev Food Sci Food Saf 10:183–193.

93

The role of topB1, rfaE and pgm in phage resistance

2.3. Manuscript III: The role of topB1, rfaE and pgm in phage resistance

Modification of enzymes involved in the synthesis of different phage receptors renders Erwinia amylovora CFBP1430 multi phage resistant

Leandra E. Knecht1,2, Yannick Born1, Cosima Pelludat3, Martin J. Loessner2, Lars Fieseler1*

1 Food Microbiology Research Group, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Wädenswil, Switzerland 2 Institute of Food, Nutrition and Health, ETH Zurich, Switzerland 3 Agroscope, Plant Pathology and Zoology in fruit and vegetable production, Wädenswil, Switzerland

*Correspondence: Lars Fieseler, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Einsiedlerstrasse 31, Wädenswil, Switzerland. Tel: +41 58 934 54 07; e-mail: [email protected]

95 Manuscript III

Abstract With the rise of antibiotic resistant pathogens, bacterial infections are an increasing threat. Bacteriophage could have the potential to complement the current treatment options. These bacteria infecting viruses are natural predators of bacteria and are in a continuous evolutionary arms-race with them. To ensure that phage biocontrol is an efficient and long-lasting treatment option against pathogenic bacteria, the risk of resistance development must be minimized. Therefore, phage resistance mechanisms applied by the pathogen should be thoroughly investigated. Especially resistance towards multiple phages can render a phage cocktail futile. Understanding and estimating the risk of multi phage resistance is essential for the establishment of phage therapy as long lasting solution against pathogenic bacteria. In a previous study, a transposon mutagenesis library of the fire blight pathogen E. amylovora was screened to identify multi phage resistant mutants. These mutants had transposon insertions in the genes topB1 (putative gene product: NAD-dependent epimerase), rfaE (ADP- heptose synthase) or pgm (phosphoglucomutase) and were able to survive attacks of phages that require different receptors for host infection. In this study knock-out mutants of these three genes were generated and analysed for modification of outer membrane structures and for potential fitness loss. Indeed, multiple alterations that could explain phage resistance were observed experimentally. LPS structures were altered in all three deletion mutants. EPS production was diminished or almost absent and cellulose production is affected in the deletion mutants. Researching the function of these genes indicated that the encoded proteins are involved in different metabolic pathways that affect these surface structures. All these findings strongly indicate that these enzymes are essential for synthesis of receptors that are targeted by the six phages. However, modification of these genes as resistance strategy against phage attack does not pose a risk to the treatment, since deletion mutants demonstrated reduced virulence in blossoms.

96 The role of topB1, rfaE and pgm in phage resistance

Introduction Bacteriophages are viruses exclusively infecting bacteria. With an estimated number of 1031 (1–3), they are the most abundant biological entity on earth and outnumber bacteria by at least tenfold (4). Phages were first identified in 1915 by Frederick Twort (5) and independently by Felix d’Hérelle in 1917 and their potential as remedy against bacterial pests was investigated (6, 7). The identification of penicillin by Alexander Fleming in 1928 (8) suddenly withdrew the focus on phage therapy and shifted it to the newly discovered antibiotics (9). The potential of phage therapy sunk into oblivion in western countries, which started to rely on antibiotics. With the rise of antibiotic resistant bacteria, phage therapy moves once more into the spotlight (10). In contrast to antibiotics, phages only recognize and target specific host bacteria, while other and potentially beneficial bacteria are left unharmed (11). Phages do not cause side effects in humans and animals, and were considered as environmentally safe (12, 13). Since phages require the host metabolism for replication, phage numbers will increase in presence of the host bacteria making the treatment more potent. If host bacteria are absent, phages will eventually decay e.g. through UV irradiation (14–16). All these abilities make phages a promising agent against pathogenic bacteria. Indeed, efficient phage treatments are already under investigation in humans and food production and against plant pathogens (12). Nevertheless, phage resistance can arise. Phages are highly specialized predators, with estimated 1025 phage infections occurring every second (17). This forces bacteria to adapt and develop resistance mechanisms. It is essential to advance our knowledge of the resistance mechanisms applied by bacteria, to generate efficient phage therapies and to avoid resistance development that render the treatment futile. Hence, the lessons learned from antibiotic resistance development should if possible be integrated, to generate potent and long lasting phage treatments that minimize the risk of phage resistance development. Bacterial resistance mechanisms can target the phage infection at different stages (18, 19). The first point of interaction between phages and potential host cells is the receptor identification on the surface of the bacterium by the phage and subsequent binding. Bacteria can mutate these receptors or modify their accessibility so that the phage can no longer recognize them as potential host cell (20–22). By combining phages that target different host receptors, the risk of resistance development through receptor alteration could be minimized. Injection of phage DNA can be blocked by superinfection exclusion systems (sie) (23, 24). Such sie systems are generally encoded by integrated prophages to prevent premature host lysis through other phages (25). Two measures can be taken to prevent resistance by sie systems. Genome analysis of the host bacterium could reveal possible prophage incorporation. Secondly, phages integrated into a phage cocktail should be strictly lytic to avoid phage DNA integration into the host genome. If the phage DNA is successfully injected into the host cell, intracellular systems can prevent the phage replication. CRISPR/Cas systems

97 Manuscript III were observed to function as bacterial immune system, which protects the cell from certain viral infections (26). Pathogen of interest should therefore be tested for established and functional CRISPR/Cas systems. This could help to estimate the risk of resistance development through spacer integration and CRISPR/Cas immunization. The restriction modification system protects the bacterial cell by marking the injected phage DNA as foreign. Digestion or methylation of phage DNA can ensure the degradation of the intruding DNA and thereby protects the bacterium from phage infection (27). Certain phage DNAs were observed to be characteristically modified to resist a collection of restriction enzymes (28). Incorporation of such phages in a cocktail could escape the bacterial restriction modification system. Some bacteria were identified to harbour an abortive infection system (abi). In contrast to the previously mentioned intracellular systems, the abi system induces controlled apoptosis of the infected cell (25). This self-sacrifice prevents phage replication and release of infectious phage progenies that could attack the bacterial population. Finally, mutation of a global regulator can result in multi phage resistant bacteria. The alternative sigma factor RpoN was previously identified to provide cross resistance against phages targeting different host receptors when mutated (29). Global regulator are likely to control the expression of several phage receptors simultaneously and can therefore affect infectivity of different phages. The danger these global regulators pose to phage therapy are yet to be discovered. In general, careful and thorough investigation of the bacterial host is most advantageous. Identifying and investigating phages that are known to have the ability to bypass or outwit the applied resistance mechanism should further help to minimize the risk of resistance development. Phage therapy could be an optimal solution for the plant pathogen Erwinia amylovora, the causative agent of fire blight (30). This severe plant disease affects members of the Rosaceae family and was classified as one of the 10 most devastating plant diseases affecting crop production (31, 32). The main route of infection is through the nectaries in the blossoms (33). Since the blossoming season spans several days, it is therefore the most promising stage to manage and prevent the disease. Fire blight outbreaks were successfully treated with antibiotics, but the appearance of antibiotic resistant strains and increasing public health concerns led to the abandoning of this therapy (34, 35). Phages could be employed as biocontrol agents to manage the disease, and different phage treatments for fire blight are currently under investigation (36–38). Optimal combination of potent phages ensures full infection coverage of E. amylovora strains and minimizes the risk of phage resistance development. A previous study analysed different E. amylovora genomes and revealed that none had an integrated prophage. Form these findings and further experiments the group concludes that lysogeny is possibly absent in E. amylovora (39). The lack of temperate E. amylovora phages could suggest that resistance development through prophage encoded sie systems is

98 The role of topB1, rfaE and pgm in phage resistance negligible in fire blight biocontrol. In the recent years identification of CRISPR/Cas system was performed with several E. amylovora strains. Indeed the pathogen was observed to harbour such a system (40). Further studies concerning the functionality of the system suggested that the E. amylovora CRISPR/Cas system is specialized in target plasmids but was unaffected by phage invasion (41). These findings suggest that although an active CRISPR/Cas system is present its involvement in phage resistance is minimal. The risk of multi phage resistance in E. amylovora though modification of a global regulator is yet to be determined. This study aimed to identify mutants that resists infection of phages that use different infection strategies. Three genes were identified in E. amylovora CFBP1430, which generate, when disrupted or deleted, multiple phage resistant mutants. The effects of these genes on the bacterium and on the phages were further investigated.

99 Manuscript III

Materials and Methods

Bacterial strains and culture conditions E. amylovora strains were cultivated on LB agar at 28°C, Escherichia coli strains at 37°C. Kanamycin (50 µg/ml) or ampicillin (100 µg/ml) was added if required. All used and generated strains are summarized in Table 1.

Table 1 Overview of used and generated bacterial strains, plasmids and phages

Bacterial strains Characterization Reference

E. amylovora CFBP1430 Wild type strain, isolated in France from Crataegus sp; Zhang and Geider, 1997 CFBP1430SmR Streptomycin resistant CFBP1430 used for knock-out generation CFBP1430 [pBAD18] Empty vector control This work CFBP1430ΔtopB1 Deletion of topB1, replaced by a kanamycin cassette This work CFBP1430ΔtopB1+topB1 Complementation of topB1 This work CFBP1430ΔrfaE Deletion of rfaE, replaced by a kanamycin cassette This work CFBP1430ΔrfaE+rfaE Complementation of rfaE This work CFBP1430Δpgm Deletion of pgm, replaced by a kanamycin cassette This work CFBP1430Δpgm+pgm Complementation of pgm This work CFBP1430Δbcs Deletion of the entire bcs operon Chapter 2 4/82 Isolated in Egypt from Pyrus communis; low EPS-producer Zhang and Geider, 1997 E. coli Miller & Mekalanos, 1988; Simon S17λ-pir et al. 1988 XL1-Blue NEB (Ipswich, USA) Dh5α supE44 ∆lacU169 (φ80lacZ∆M15) hsdR17 recA1 endA1 gyrA96, thi-1 relA1 Hanahan, 1983 Plasmids

pSB315 Containing kanamycin cassette without transcriptional terminator, AmpR, KanR Galan et al. (1992) pKAS32 Suicide vector with rpsL gene, AmpR Skorupski & Taylor (1996) pKAS32::deletion topB1 Deletion vector exchanging topB1 for a kanamycin cassette This work pKAS32::deletion rfaE Deletion vector exchanging rfaE for a kanamycin cassette This work pKAS32::deletion pgm Deletion vector exchanging pgm for a kanamycin cassette This work pBAD18 Complementation vector, arabinose induced pBAD promoter Guzman et al 1995 pBAD18:: topB1 Complementation vector for topB1, artificial RBS This work pBAD18:: rfaE Complementation vector for rfaE, artificial RBS This work pBAD18:: pgm Complementation vector for pgm, artificial RBS This work Phages

Bue1 E. amylovora specific ackermannvirus, Vi1-like Knecht et al, 2018 L1 E. amylovora specific podovirus, T7-like Born et al, 2011 M7 E. amylovora specific myovirus, FO1-like Born et al, 2011 S2 E. amylovora specific podovirus, SP6-like Knecht et al, 2018 S6 E. amylovora specific podovirus, N4-like Born et al, 2011 Y2 E. amylovora specific myovirus, GJ1-like Born et al, 2011

100 The role of topB1, rfaE and pgm in phage resistance

Soft agar overlay and propagation Phages were propagated using the soft agar overlay method (42). Molten LB+ soft agar (LB broth, 4 g/l agar, 2 mM MgSO4, 10 mM CaCl2) was supplemented with 90 µl of a bacterial overnight suspension and 10 µl of a phage dilution row. Soft agar was then evenly spread onto LB plates to generate semi confluent lysed plates. The plates were incubated overnight before 5 ml SM buffer (100 mM NaCl, 8 mM MgSO4, 50 mM Tris-Cl, pH 7.4) per plate were added and plates were incubated for 5 h at room temperature (RT) under shaking. NaCl was added (0.5 M) to the supernatant and the mixtures were incubated for 30 min at RT before centrifugation (10 min, 10,000 xg). Phages in the supernatant were then polyethylene glycol treated (10% w/v PEG8000, ice bath overnight), spun down (15 min, 10,000 xg, 4°C), CsCl density gradient purified (43) and dialyzed three times against SM buffer (Bue1, L1, M7, S2 and S6). In the case of Y2, PEG was removed by incubating the PEG-phage mixture at RT for 1 h. The solution was then centrifuged at 5,000 xg for 10 min. Phages in the supernatant were sterile filtered (0.22 µm filter). The phages were stored at 4°C. Table 2 Overview of the used primers for the gene deletions and the complementations

Primer Sequence 5'-3' Characterization Reference

pSB315 f GAA AGC CAC GTT GTG TCT C Kanamycin cassette Galan et al. 1992 pSB315 r CCT TCA TTA CAG AAA CGG C Kanamycin cassette Check plasmid pBAD f CTG TTT CTC CAT ACC CGT T insertion Check plasmid pBAD r CTC ATC CGC CAA AAC AG insertion pKAS32-topB1_up-fw CAG ATC TGC GCG CGA TCG ATG CGC TTA TCT CTG CAA ACA G CFBP1430ΔtopB1 This work pKAS32-topB1_up-rev GAG TTT TTC TAA TGG CAG ACT GAT GGC CTG AAA AAA AAG CFBP1430ΔtopB1 This work

pKAS32-topB1_kan-fw TCA GTC TGC CAT TAG AAA AAC TCA TCG AGC CFBP1430ΔtopB1 This work pKAS32-topB1_kan-rev CGA GGG CAC TAT GAG CCA TAT TCA ACG G CFBP1430ΔtopB1 This work pKAS32-topB1_down- CFBP1430ΔtopB1 TAT GGC TCA TAG TGC CCT CGT TAA TTA TGC AAT G This work fw pKAS32-topB1_down- CFBP1430ΔtopB1 CGC AAA TTT AAA GCG CTG ATC CAC CGT TCC GGT CGG TA This work rev Complementation pBAD-topB1-fw TGGGCTAGCGAATTCGAGCTCAGGAGGTTCGTATGAAATATCTGGTCACC This work CFBP1430ΔtopB1 Complementation pBAD-topB1-rev TGCATGCCTGCAGGTCGACTCTAGATTACTTGTGATAAAACTCTTTATAC This work CFBP1430ΔtopB1 pKAS32-rfaE_up-fw TGCGCATGCTAGCTATAGTTCTAGATAATGAAGGTCAGCGATC CFBP1430ΔrfaE This work pKAS32-rfaE_up-rev TATGGCTCATTCCAGAGACTCCAGACAG CFBP1430ΔrfaE This work

pKAS32-rfaE_kan-fw AGTCTCTGGAATGAGCCATATTCAACGG CFBP1430ΔrfaE This work pKAS32-rfaE_kan-rev CCGTAATCCGTTAGAAAAACTCATCGAGC CFBP1430ΔrfaE This work

pKAS32-rfaE_down-fw GTTTTTCTAACGGATTACGGCTGCCCGC CFBP1430ΔrfaE This work pKAS32-rfaE_down-rev TGGAATTTCCCGGGAGAGCTCGTTGTTCTGCACCATCATGGGC CFBP1430ΔrfaE This work Complementation pBAD-rfaE-fw TGGGCTAGCGAATTCGAGCTCTGGAGTCTCTGGAATGAAAATTAC This work CFBP1430ΔrfaE Complementation pBAD-rfaE-rev TGCATGCCTGCAGGTCGACTCTAGATCAGTTTTGCCCGGTTCC This work CFBP1430ΔrfaE pKAS32-pgm_up-fw TGCGCATGCTAGCTATAGTTCTAGATATCCTGGGCCATTAGCGG CFBP1430Δpgm This work pKAS32-pgm_up-rev TATGGCTCATTGGCGCTTCTCCCTGACG CFBP1430Δpgm This work

pKAS32-pgm_kan-fw AGAAGCGCCAATGAGCCATATTCAACGG CFBP1430Δpgm This work pKAS32-pgm_kan-rev GCCGGTCAAATTAGAAAAACTCATCGAGC CFBP1430Δpgm This work

pKAS32-pgm_down-fw GTTTTTCTAATTTGACCGGCGGTGAAAAG CFBP1430Δpgm This work pKAS32-pgm_down- CFBP1430Δpgm TGGAATTTCCCGGGAGAGCTCATAATGTTGATAAAGACCCATACGTC This work rev Complementation pBAD-pgm-fw TGGGCTAGCGAATTCGAGCTCAGGGAGAAGCGCCAATGG This work CFBP1430Δpgm Complementation pBAD-pgm-rev TGCATGCCTGCAGGTCGACTCTAGATCAGGCGTTATTCAGAACCG This work CFBP1430Δpgm

101 Manuscript III

Generation of knockout mutants The Tn5 library screen revealed mutants with transposon insertions into the genes rfaE, pgm and topB1 to be resistant against multiple phages (chapter 1). To investigate the impact of these genes on the phage infection process, knock-out mutants were generated (Table 1). The gene deletions were generated by allelic exchange using the suicide plasmid pKAS32 carrying a R6K origin of replication (44). Flanking regions (approx. 1,000 bp up- and downstream) of the genes of interest and a kanamycin cassette (aphT) amplified from pSB315 (45) were integrated into the multiple cloning site of pKAS32. The three fragments were amplified using Gibson assembly primers designed by the NEBuilder tool (NEBuilder Assembly Tool v1.12.18) using the KAPA HIFITM PCR kit (KAPA Biosystems, Wilmington, USA). The fragments were joined by overlap extension PCR. The vector pKAS32 was extracted from E. coli S17-1 λpir [pKAS32] according to the NucleoSpin® Plasmid Kit (Macherey-Nagel; Düren Germany) and linearized by SacI and XbaI. The product was anayzed by electrophoresis and purified from a 1% agarose gel by the DNA Clean & Concentrator™-5 Kit by Zymo Research. The suicide plasmid was generated either through Gibson assembly (46), or through classic ligation using a T4 polynucleotide ligase (Thermo Fisher Scientific) according to the manufacturer’s instructions. The plasmids were introduced into electrocompetent, streptomycin resistant E. amylovora CFBP1430 by electroporation Cells were resuspended in SOC and incubated at 30°C with vigourous shaking for 1 h. The cells were then plated on LB plates complemented with kanamycin and streptomycin to select for deletion mutants. Alternatively, the deletion vector was introduced by mating. Overnight cultures of E. amylovoraSmR and E. coli S17λ-pir containing the deletion vector were grown in LB containing streptomycin (100 µg/ml) or ampicillin. Bacteria were washed twice in PBS buffer and mixed at different ratios (E. coli : E. amylovora 1:10, 1:100). Cells were then centrifuged (5,000 xg, 2 min) and the pellet was spotted onto LB plates and incubated at 37°C. After 3, 6 and 24 h, cells were scraped off and plated onto MM2 plates supplemented with kanamycin and streptomycin to select correct mutants. Correctness of deletion was checked by PCR.

Complementation of knockout mutants The arabinose inducible plasmid pBAD18 (47) holding an ampicillin resistance was linearized with SacI and XbaI, and purified. The inserts and their ribosomal binding sites were amplified from E. amylovora CFBP1430 by PCR with Gibson primers (NEB). PCR products were recovered from a 1% agarose gel. The linearized pBAD18 vector and the insert were joined by Gibson assembly (46). The plasmids were then introduced into electrocompetent E. coli XL1-Blue cells for amplification. Cells were recovered in SOC and incubated for 1 h at 37°C with vigourous shaking before plating onto LB plates containing ampicillin. Correct insertion

102 The role of topB1, rfaE and pgm in phage resistance was verified by PCR. Correct plasmids were extracted and introduced into electrocompetent E. amylovora CFBP1430 knockout mutants. Phage resistance of the generated mutants was monitored using the spot on the lawn technique with dilution rows of Bue1, L1, M7, S2, S6 and Y2.

In vitro infection assay The infectivity of the six phages towards the generated mutants was tested. In vitro infections were carried out to monitor growth of the mutants in the presence of phages. E. amylovora CFBP1430 and the generated mutants were grown overnight in LB medium. Cells where then

7 washed twice in LB+ medium and OD600 was adjusted to 10 CFU/ml. A total of 20 µl per sample were transferred to 1960 µl LB and supplemented with 20 µl phage (1010 PFU/ml). Mixtures where then distributed into a 96-well flat bottom plate. The samples were incubated at 25°C with double orbital shaking in a plate reader for 40 h. OD600 was measured every half hour.

Pull-down assay Receptor binding affinity of the generated mutants was tested using pull down experiments (48). Overnight cultures of the mutants were washed with LB medium twice and diluted to an

OD600 of 1.0. A total of 990 µl per sample was transferred to 2 ml test tubes and 10 µl Bue1, L1, S2 or Y2 (109 PFU/ml) were added and incubated for 10 min at RT under shaking. Samples were then centrifuged at 10,000 xg for 10 min at 4°C. Unbound phages in the supernatant were quantified using the soft agar overlay method plating on E. amylovora CFBP1430. As controls, phages were incubated with wildtype or medium only. The phages M7 and S6 were excluded from the pull down experiment since previous experiments have demonstrated poor reproducibility (Knecht et al. unpublished)

EPS measurement The impact of rfaE, pgm and topB1 deletions on EPS synthesis was analysed using the EPS- Cetylpyridiniumchlorid (CPC) precipitation assay (49). The generated mutants and the wildtype were grown in MM2 (49) at 28°C under shaking for 24 h. The samples were adjusted to an OD600 of 1.0. A total of 1 ml per sample was centrifuged for 5 min at 10,000 xg and 950 µl supernatant per sample were mixed with 50 µl CPC (50 mg/ml). After incubating the samples for 10 min at RT, OD600 of each sample was measured. The low EPS producing strain E. amylovora 4/82 was used as control.

LPS silver staining Potential alterations in the LPS structure were analysed by silver staining. Cells were grown in LB and washed twice in PBS buffer before the OD600 was adjusted to 1.0 and 1 ml per

103 Manuscript III sample was centrifuged at 8,000 xg for 5 min. The pellet was resuspended in 100 µl SDS sample buffer (90 mM Tris base, 2% SDS, 0.02% Bromophenol blue, 20% sucrose, pH adjusted to 6.8 in H2O) and boiled for 10 min at 100°C. After cooling the samples to room temperature, 2.5 µl proteinase K (20 µg/µl) were added and samples were incubated for 1 h at 60°C. The samples were subsequently loaded onto an SDS PAGE gel (12% resolving/4% stacking gel) and let run with 35 A for 2 h. The gels were washed in ddH2O before being soaked with fresh fixing solution (40% ethanol and 5% acetic acid in H2O) for 1 h. The fixing solution was subsequently replaced with fresh oxidizing solution (fixing solution supplemented with 30 mM periodic acid) for 5 min. Oxidizing solution was removed and the gels were washed

3 to 5 times with at least 500 ml of ddH2O for 15 min to completely remove the oxidizing solution. The gels were soaked in freshly prepared staining solution (1.5 ml ammonium hydroxide solution 33%, 14 mM NaOH solution, 0.5 % AgNO3 in 200 ml H2O) for 15 min. After washing the gels several times in ddH2O, gels were developed with freshly prepared developer solution (200 ml ddH2O supplemented with 50 mg citric acid and 100 µl formaldehyde solution) until bands appeared. The development was stopped with several charges of ddH2O.

Cellulose visualization To investigate alterations in the cellulose production provoked by the gene modifications, qualitative cellulose visualization was performed. LB agar without NaCl was supplemented with 40 μg/ml Congo Red and 20 μg/ml Coomassie Brilliant Blue before pouring. Ampicillin and/or 0.2% arabinose were added if required. Per sample 5 µl were spotted onto plates and incubated for two days at RT. Cellulose producing bacteria appear as pink colonies while non- producing bacteria remain white (15).

Growth curves Impact of the mutations on the fitness of the generated strains was investigated. Bacteria were washed twice in SM buffer and OD600 was adjusted to 0.1. Cells were diluted in LB medium supplemented with MgSO4 (2 mM) and CaCl2 (10 mM) or MM2 medium (49) to a concentration of 105 CFU/ml. Cultures were incubated at 28°C for 24 h with double orbital shaking (150 rpm) in a plate reader.

BIOLOG To investigate if the carbon metabolism in the generated deletion mutants alters from the wildtype, growth was tested on different carbon sources. Samples were grown overnight on fresh LB plates before bacteria were prepared according to the manufacturer’s protocol (BIOLOG, Hayward, CA, USA; PM1-2). Bacterial cells were transferred to the BIOLOG PM1- 2 and incubated in the OmniLog PM system for 72 h at 25°C. Each sample was tested twice on PM1 and PM2. As control, the wildtype CFBP1430 was used.

104 The role of topB1, rfaE and pgm in phage resistance

Detached flower assay Virulence of the generated E. amylovora mutants was tested by a detached flower assay using fresh blossoms from 2 year old Golden Delicious apple trees (50). Blossoms were treated with E. amylovora CFBP1430, the knock-out mutants or PBS buffer (3 mM KCl, 137 mM NaCl, 2 mM KH2PO4, 10 mM Na2HPO4) as mock infection. Racks were autoclaved before the experiment and 24 alternating wells per rack were filled with 2 ml H2O. The wells were sealed with perforated scotch tape. Stems were cut freshly to ensure water uptake before transferring the blossoms into the filled wells. Bacteria grown overnight on plates were carefully scratched off and resuspended in PBS and the optical density was adjusted to 1.0. Subsequently, a 1:50 dilution was performed to generate approx. 107 CFU/ml. A total of 20 µl per treatment were pipetted directly onto the receptacle. The racks were placed into storage boxes (5 l) laid out with paper towels and soaked with 100 ml H2O to ensure humidity. Boxes were incubated at 26°C for 4-5 days before the read out according to an adjusted rating system was carried out (51): Healthy blossoms without disease symptoms are classified as grade 1. Visible symptoms on the blossom (browning of the calix) are referred to as grade 2. Blossoms with disease symptoms in the calix and the stipe of the blossoms correspond to grade 3. Cell counts on 12 blossoms per treatment were determined after 4 days of infection as described by Zengerer et al. (52). Briefly, petals, pedestals, stamps and stigmas were stripped of the blossoms and the remaining plant parts were transferred into a test tube supplemented with 1 ml PBS buffer. Samples were incubated for 30 min under shaking (1,400 rpm) and vortexed for 30 s. The generated serial dilution was spotted (3 µl/dilution) in duplicates onto LB plates containing kanamycin and incubated until counting was possible.

105 Manuscript III

Results Phage therapy has the potential to be an effective weapon against pathogenic bacteria. The development of phage resistant mutants should, however, be eliminated. Aside from single phage resistance, multiple phage resistance can transform a potent phage cocktail into an ineffective one. To anticipate the threat they can pose to phage biocontrol agents, genes involved in mediating multiple phage resistance should be identified. In a previous study, a Tn5 transposon mutagenesis library was screened to identify multi phage resistant mutants (chapter 1). The screen was carried out for 6 phages classified as ackermannvirus (Bue1), podoviruses (L1, S2, S6) and myoviruses (M7, Y2), belonging to different genera. Mutants with transposon insertions into rfaE, pgm and topB1 were observed to be able to resist infection of multiple phages. One mutant with an insertion into rfaE was identified as L1, M7, S2, S6 and Y2 resistant. Only the phage Bue1 was observed to lyse the transposon disrupted rfaE mutant. The gene rfaE is predicted to encode an ADP-heptose synthase. In E. coli K12, the bifunctional protein HldE, a homologue protein to RfaE, catalyses the phosphorylation of D-glycero-D-manno-heptose-7-phosphate to form D-glycero-beta-D-manno-heptose-1,7- bisphosphate and the ADP transfer from ATP to D-glycero-beta-D-manno-heptose-1- phosphate, generating ADP-D-glycero-beta-D-manno-heptose (53). The gene pgm was identified once to be involved in phage resistance mediation. As the rfaE transposon mutant, the pgm disrupted mutant was observed to be insensitive against L1, M7, S2, S6 and Y2 but not Bue1. pgm is annotated to encode a phosphoglucomutase, an enzyme involved in breakdown and synthesis of glucose (54). A total of 8 mutants with independent transposon insertions into the gene topB1 were identified during the screen (chapter 1). Phage resistance for these mutants was observed against Bue1, Y2, M7 and S6. The phage S2 was still able to lyse all the tested topB1 transposon mutants. L1 was unable to infect one topB1 mutant but was observed to lyse the other 7 mutants. Although the gene topB1 is annotated to encode a topoisomerase B1, amino acid BLAST suggests that the gene encodes an NAD-dependent epimerase. These proteins use nucleotide-sugar substrates for different chemical processes with NAD as cofactor (55). In order to exclude possible downstream effects of the transposon insertions, knock-out mutants were generated. To verify if the gene deletions still had an impact on phage infectivity, resistance was tested in soft agar overlays (Table 3). The deletion of rfaE was observed to mediate resistance against Bue1, Y2, M7, S6 and S2. The phage L1 was shown to be unaffected by the rfaE deletion. Mutants with a deletion of the phosphoglucomutase encoding gene pgm were able to resist infection by all phages with the exception of M7. If the LB agar and soft agar were supplemented with 0.2% arabinose, Bue1, L1 and S2 were also able to infect Δpgm.

106 The role of topB1, rfaE and pgm in phage resistance

Removing the gene topB1 rendered the knock-out mutant resistant against Bue1, Y2, M7 and S6. Both L1 and S2 were still able to lyse the topB1 mutant. Complementation of the three genes restored phage infectivity when arabinose was added to the medium to induce the promoter. In certain cases, the complementation already restored the phage sensitive phenotype in absence of arabinose. In general, the addition of arabinose or glucose (data not shown) was observed to affect the phage resistance of the deletion mutants. Complementing with the natural promoter could eliminate effects of inducing sugars on the mutants.

Table 3 Infection of the generated mutants by the six E. amylovora specific phages. Spot on the lawn experiments were carried out on LB plates. Arabinose 0.2% was added to induce the promoter of the complementants. As control, the wildtype and the empty vector control (CFBP1430 [pBAD18]) were used. The panel A represents the infection of the deletion mutants whereas panel B summarizes the results of phage infection of the complementants. Successful infection is depicted by +, - represents lack of infection and plaques.

Reduced phage adsorption in the generated mutants To identify how the generated mutants affect the six different phages, a collection of experiments were carried out. Pull-down experiments for Bue1, L2, S2 and Y2 were performed to verify if adsorption to the generated deletion mutants was modified. As illustrated in Figure 1, Y2 was unable to adsorb to all of the three deletion mutants. Thus, lack of recognition and binding, accounts for the inability of Y2 to infect the mutants. L1 and S2 were both still able to bind to the generated mutants. However, binding affinity was strongly reduced compared to the wildtype. L1 and S2 adsorption to Δpgm was observed to be the weakest of all three tested mutants. In contrast to the other phages, Bue1 generally showed intermediate adsorption with strong variation. The phages M7 and S6 generated no reproducible data in the pull down experiments and were therefore excluded.

107 Manuscript III

Figure 1 Pull down experiments of all generated mutants against Bue1, L1, S2, and Y2. The ability of these four phages to bind to the different mutants was tested. Each sample was tested at least three times independently. Error bars indicate standard deviation. Statistical analysis was performed using one way ANOVA, p < 0.05.

Alterations in LPS of all three deletion mutants The binding affinity of Y2 toward the knock-out mutants was observed to be abolished. Since Y2 was suggested to recognize LPS structures on the host surface, the mutants’ LPS was analysed for possible modifications. Alterations were observed for all three knock-out mutants (Figure 2). The topB1 deletion mutant lacks a total of four bands between 15-25 kDa and possibly at 40 kDa. The deletion of rfaE generated a strong alteration of the Lipid A and the core region. In addition, three bands between 15-25 kDa and possibly bands at 40 and 70 kDa are absent in the LPS pattern of ΔrfaE compared to the wildtype. The mutant lacking the gene pgm was observed to have the closest LPS pattern compared to the wildtype. As for the other two mutants, a band at ca. 18 kDa is missing and bands are more prominent between 55-70 kDa. In general, the band at 35 kDa appears much stronger for all three mutants compared to the wildtype.

108 The role of topB1, rfaE and pgm in phage resistance

Figure 2 LPS silver staining of the generated mutants. The wildtype (wt), ΔtopB1, ΔrfaE, and Δpgm were silver stained to analyse alterations of the LPS pattern. Image was taken by an Azure biosystems C300. Stars indicate alterations in the band pattern compared to the wildtype.

EPS production is affected in mutants Previous studies indicated that L1 and possibly S2 require the exopolysaccharide amylovoran for host recognition and infection (chapter 1). Therefore, the secreted amount of EPS was monitored for the deletion mutants and summarized in Figure 3. The strain E. amylovora 4/82 was used as negative control since the strain is a low EPS producer. All deletions were observed to strongly affect EPS production. The deletion of topB1, rfaE and pgm significantly reduced the amount of EPS produced by these strains. Δpgm was observed to generate the lowest amount of EPS overall compared to the wildtype. The complementants of the three mutants were induced with 0.2% arabinose and were all observed to produce comparable levels of EPS as the vector control. The uninduced complementation of topB1 and rfaE were observed to already produce higher levels of EPS compared to the deleted mutant.

109 Manuscript III

Figure 3 EPS-CPC precipitation. Quantitative EPS measurements were carried out for the generated mutants. Produced EPS was precipitated with CPC and measured by optical density. The EPS lacking strain 4/82 was used as negative control alongside CFBP1430 and CFBP1430 carrying the empty vector as positive controls. All samples were prepared in MM2 medium (black) or in MM2 with 0.2% arabinose to induce the promoter (white). Each sample was tested at least three times independently. Statistical analysis was done by a two way ANOVA, p < 0.05 and samples were compared to the corresponding wildtype. Error bars indicate standard deviation.

Alterations in cellulose production The two phages M7 and S6 were identified to require cellulose for successful infection of host cells (chapter 1,2). M7 and S6 show difficulties in infecting the generated knockout-mutants. The mutants were therefore tested for alterations in cellulose production (Figure 4). The wildtype and the cellulose operon lacking strain Δbcs were used as positive and as negative control, respectively. All three mutants were observed to generate colonies appearing lighter pink than the wildtype. Especially the strain carrying the pgm deletion was observed as light pink colony suggesting lowered cellulose production.

Figure 4 Cellulose production monitored through Congo Red. Overnight culture of the wildtype, Δbcs as negative control and each mutant were spotted on LB plates supplemented with Congo Red and Coomassie Brilliant Blue. Cellulose producing strains grow with a pink phenotype whereas cellulose lacking strains grow as white colonies

Impact of topB1, rfaE and pgm deletion on fitness Phenotypical alterations and possible fitness reduction due to the gene deletions were tested. Growth of the mutants was tested in LB medium, where both the Δpgm and the ΔtopB1 mutant were observed to perform comparably to the wildtype (Figure 5). The deletion of rfaE affected the growth of the mutant in LB. Similar results were obtained for the complementants in LB

110 The role of topB1, rfaE and pgm in phage resistance medium supplemented with 0.2% arabinose, where the pgm and topB1 complemented mutants were able to grow as the wildtype but the rfaE complementant was still observed to perform weaker. In MM2 medium, mutants were observed to perform weaker than the wildtype. All complementants but the pgm complemented strain were able to grow as well as the wildtype in MM2. The growth of the pgm complemented strain was observed to be delayed.

Figure 5 Growth curves of different mutants in LB or MM2 medium. Cells were adjusted to 105 CFU/ml and incubated in LB medium (A, B) or in MM2 medium (C, D). Arabinose 0.2% was added to the complementants to induce the vector promoter (B, D). The strain E. amylovora CFBP1430 (wt) and the strain complemented with the empty complementation vector pBAD18 (vector control) were used as control. Optical density was measured after every half hour for 24 h. Error bars indicate standard deviation.

To test whether the generated knock-out mutants are able to metabolize different sugars, a BIOLOG assay was carried out testing a collection of C-sources. All mutants were able to grow comparably to the wildtype on all tested C-sources. No significant differences were observed (data not shown). Virulence of the three generated knock-out mutants was tested with a detached flower assay (Figure 6). In case of the pgm and rfaE deleted mutants, all infected blossoms remained healthy without disease symptoms. The ΔtopB1 mutant generated 28.1% grade 3 blossoms with fully developed infection. A total of 93.75% of the blossoms infected with the wildtype were classified as grade 3.

111 Manuscript III

Figure 6 Virulence of topB1, pgm and rfaE deletion mutants tested on detached apple blossoms. Two-year-old Golden Delicious apple blossoms were infected with approx. 107 CFU/ml bacteria and incubated for 4 days. The read out was performed by evaluating disease symptoms. Grade 1 indicates healthy blossoms without disease symptoms. Grade 2 refers to blossoms with first disease symptoms. Grade 3 represents fully infected blossoms with advanced disease symptoms in the stipe.

However, when bacteria were reisolated form the infected blossoms, both Δpgm and ΔrfaE were observed to be unable to colonize the blossoms. Hence, the absence of pathogen is responsible for the healthy blossom phenotype in the case of pgm and rfaE deletion. Deletion mutants of topB1 were shown to grow on blossoms but to much lower concentrations compared to the wildtype. This reduced growth could account for the results obtained by disease symptoms.

112 The role of topB1, rfaE and pgm in phage resistance

Discussion Bacteriophages are a promising alternative to antibiotics in the fight against pathogenic bacteria. As with antibiotics, phage resistant bacteria will emerge eventually. By combining phages that apply different infection mechanisms, the risk of resistance can be controlled, since the targeted bacterium has to adapt to multiple phages simultaneously. Modification of a global regulator, however, can establish multiple phage resistance. Identification and characterization of factors that can generate multi phage resistant pathogens is therefore required to prevent the evolvement of phage resistance. In a previous study, a transposon screen identified such factors in the plant pathogen E. amylovora (chapter 1). In this study the effect of these factors namely the genes topB1, rfaE and pgm and their involvement in multiple phage resistance were investigated further. The gene topB1 is annotated to encode a topoisomerase IA. These enzymes cut and reseal the double stranded DNA to relax the DNA supercoils, which is required for transcription. However, amino acid comparisons by BLAST suggest that the protein rather functions as NAD-dependent epimerase (56, 57). Orthologue analysis by the KEGG database suggests a function as UDP-glucoronate 4-epimerase (58). Mutants with a deleted topB1 gene were observed to be resistant towards four (Bue1, M7, S6, Y2) out of six phages tested. The phages Bue1 and Y2 require certain LPS structures for successful host recognition and infection (chapter 1). This suggests that LPS is altered in the ΔtopB1 mutant. Silver staining of the extracted LPS revealed multiple alterations in the LPS structure of the ΔtopB1 mutant compared to the wildtype. Several bands are absent between 15-25 kDa. The STRING database (string-db.org, Version 11.0) was applied to identify known or predicted protein- protein interactions (59). TopB1 was suggested to interact with a collection of predicted glycosyltransferases (GalU, GalF, OtsA3) and the proteins WalW1 and WaaG, both involved in LPS biosynthesis. This could account for the modification of the LPS structures in the ΔtopB1 mutant and therefore the resistance towards Bue1 and Y2. In addition, the STRING database predicts an inter-pathway connection between TopB1 (Amino sugar and nucleotide sugar metabolism, KEGG) and BcsA (starch and sucrose metabolism, KEGG), the catalytic active subunit of the bacterial cellulose synthase complex. M7 and S6 were previously shown to rely on cellulose or bacterial cellulose synthase for successful infection (chapter 1,2). Neither of the two phages can infect the ΔtopB1 mutant. Cellulose production of the ΔtopB1 mutant was therefore monitored on Conge Red plates. CFBP1430ΔtopB1 was observed to generate less cellulose than the wildtype. These findings suggest that the deletion of topB1, aside from the observed LPS modifications, has an impact on cellulose production and thus abolishes infection by M7 and S6. Finally, the two phages L1 and S2 are not affected by the deletion of ΔtopB1. Both phages were observed to require the EPS amylovoran for host

113 Manuscript III recognition and infection (chapter 1). EPS production was monitored to investigate possible alterations. Although the ΔtopB1 mutant was observed to produce significantly less EPS compared to the wildtype, both phages were still able to infect the knock-out mutant. We conclude that the secreted amount of EPS must still be sufficient for phage infection. The fact that L1 and S2 are still able to infect the mutant indicates that the amylovoran structure is unaffected by the deletion. Summarizing these findings, we could show that the deletion of the gene topB1 affects LPS structure, bacterial cellulose production and lowers the total amount of secreted EPS. These modifications result in multiple phage resistance. The second gene identified to mediate multi phage resistance when disrupted was rfaE. This gene is annotated to encode the bifunctional protein HldE, an ADP-heptose synthase. The STRING database indicates protein-protein interaction of HldE with a collection of proteins essential for LPS synthesis (WaaF, WaaQ3, WaaC, WaaF2). The KEGG database suggests that HldE is involved in the LPS biosynthesis pathway and generates ADP-D-glycero-β-D- manno-heptose, which is a building block of the inner core oligosaccharide. Studies in E. coli and Salmonella typhimurium also link rfaE with the LPS biosynthesis mechanism (60, 61). Indeed, the ΔrfaE mutant was observed to express a strongly altered LPS structure, especially the Lipid A and the inner core compared to the wildtype. This explains why neither Bue1 nor Y2 were able to infect ΔrfaE. The mutant was observed to produce similar amounts of cellulose as ΔtopB1, which could account for M7 and S6 resistance. However, the link between rfaE and the cellulose synthase complex is not as clear. Further investigation should be carried out to understand the relationship between rfaE, cellulose and M7 or S6 infection. EPS production was monitored for the ΔrfaE mutant to reveal lowered EPS production compared to the wildtype. In contrast to the ΔtopB1 deletion, the ΔrfaE strain was only able to resist one amylovoran specific phage. L1 infectivity was unaffected by the deletion of rfaE. This poses the question on how this discrepancy can arise. Possible alterations in the amylovoran structure or the total amount of secreted EPS could affect S2 infectivity. Finally, the deletion of the gene pgm was observed to mediate resistance against Bue1, L1, S2, S6 and Y2. Only M7 infectivity was unaffected by the removal of pgm. The gene pgm is annotated to encode a phosphoglucomutase, which harbours four phosphoglucomutase domains (Pfam domains (62)). STRING database analysis suggests protein interaction with glycosyltransferases but no link between pgm and LPS or cellulose synthesis could be established. Nevertheless, neither the LPS recognizing phages Bue1 and Y2 nor the cellulose requiring phage S6 were observed to be able to infect the deletion mutant. Indeed, cells grow as whitish colonies on Congo Red, suggesting reduced cellulose production and LPS structure was observed to be slightly altered (55-70 kDa, Lipid A region), which probably accounts for the S6, Bue1 and Y2 resistance. Further analysis of the synthesised LPS structure should be

114 The role of topB1, rfaE and pgm in phage resistance carried out to reveal detailed structural modifications. In contrast to the deletion of topB1 or rfaE, both L1 and S2 were observed to be unable to lyse the Δpgm mutant. The production of EPS in Δpgm was investigated and observed to be minimal. The absence of EPS could therefore explain why neither L1 nor S2 were able to infect the Δpgm mutant. Interestingly, all phages were able to reinfect the Δpgm mutant when arabinose was supplemented in the medium. This raises the question if arabinose somewhat can bypass deletion and re-establish the phage sensitive phenotype. The KEGG database suggests that Pgm participates in different pathways such as the glycolysis/gluconeogenesis, the biosynthesis of antibiotics, streptomycin biosynthesis etc. In addition, Pgm is supposedly involved in starch and sucrose metabolism (cellulose) and in the amino sugar and nucleotide sugar metabolism. These data strongly indicate that Pgm is a multifunctional protein. Pgm can affect different metabolic processes which are required for the biosynthesis of several outer membrane molecules. As a result the deletion or mutation of pgm can have an impact on outer membrane structures and therefore on the phage infectivity. We conclude that the genes topB1, rfaE and pgm are involved in different metabolic pathways that can modify various surface structures recognized by phages. Although the mutation or deletion of these genes results in the observed multi phage resistance we decline to classify them as global regulators. In contrast to the previously identified global regulator RpoN, an alternative sigma factor, the three identified factors cannot influence gene regulation or expression. All three enzymes are involved in metabolic processes and are required for biosynthesis of several molecules. Their function is restricted to specific steps in this metabolic process. It is therefore rather the generated building block or a modified protein that has a subsequent impact on these outer membrane structures than the enzymes themselves. By properly functioning these enzymes, enable the correct production of the receptors that are then recognized and targeted by different phages. Nevertheless, mutation of these enzymes results in multiple phage resistant mutants. To verify the likelihood that such gene mutation can occur in nature, fitness of the generated deletion mutants was investigated. On blossoms, the deletion mutants were unable to colonize and infect the plant tissue. It is unclear what prevented the mutants form establishing growth in the blossoms, especially since the mutants were able to perform similarly as the wildtype on different C-sources. An explanation for this phenomenon could be the reduced or abolished EPS production of the deletion mutants. EPS is known to attach bacteria to plant surfaces and protect them from water or nutrient loss under harsh conditions (49, 63, 64). In addition, several studies suggest that EPS is crucial in bypassing the plant defence system (49, 65). Without the protective EPS layers, the bacterium is exposed to both harsh environmental conditions and plant defence and is therefore encountering a much more hostile habitat. From our results, we cannot exclude that mutation

115 Manuscript III of these genes occurs in nature. If, however, such a strategy would be applied by E. amylovora to protect itself from a phage cocktail, the alteration of these genes would render the pathogen more vulnerable to environmental conditions or the plant defence system. Both could ensure successful elimination of the pathogen.

Acknowledgments This work was funded by the Swiss National Science Foundation (SNF) grant 310030_156947.

116 The role of topB1, rfaE and pgm in phage resistance

References

1. Hatfull GF. 2008. Bacteriophage genomics. Curr Opin Microbiol 11:447–453.

2. Cobián Güemes AG, Youle M, Cantú VA, Felts B, Nulton J, Rohwer F. 2016. Viruses as winners in the game of life. Annu Rev Virol 3:197–214.

3. Casjens SR. 2005. Comparative genomics and evolution of the tailed-bacteriophages. Curr Opin Microbiol 8:451–458.

4. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell 108:13–6.

5. Twort FWW. 1915. An investigation on the nature of ultra-microscopic viruses. Lancet 2:1241–1243.

6. D’Herelle F. 2007. On an invisible microbe antagonistic toward dysenteric bacilli: brief note by Mr. F. D’Herelle, presented by Mr. Roux. Res Microbiol 158:553–554.

7. Summers WC. 2001. Bacteriophage therapy. Annu Rev Microbiol 55:437–451.

8. Fleming A. 1929. On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae. Br J Exp Pathol X:226.

9. Salmond GPC, Fineran PC. 2015. A century of the phage: past, present and future. Nat Rev Microbiol 13:777–786.

10. Gordillo Altamirano FL, Barr JJ. 2019. Phage therapy in the postantibiotic era. Clin Microbiol Rev 32:e00066-18.

11. Loc-Carrillo C, Abedon ST. 2011. Pros and cons of phage therapy. Bacteriophage 1:111–114.

12. Buttimer C, McAuliffe O, Ross RP, Hill C, O’Mahony J, Coffey A, O’Mahony J, Coffey A. 2017. Bacteriophages and bacterial plant diseases. Front Microbiol 8:34.

13. Hagens S, Loessner MJ. 2010. Bacteriophage for biocontrol of foodborne pathogens: calculations and considerations. Curr Pharm Biotechnol 11:58–68.

14. Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JA, Momol MT, Jones JB, Vallad GE. 2012. Soil-based systemic delivery and phyllosphere in vivo propagation of bacteriophages: Two possible strategies for improving bacteriophage persistence for plant disease control. Bacteriophage 2:215–224.

15. Jones JB, Vallad GE, Iriarte FB, Obradović A, Wernsing MH, Jackson LE, Balogh B, Hong JC, Momol MT. 2012. Considerations for using bacteriophages for plant disease control. Bacteriophage 2:208–214.

16. Born Y, Bosshard L, Duffy B, Loessner MJ, Fieseler L. Protection of Erwinia amylovora bacteriophage Y2 from UV-induced damage by natural compounds. Bacteriophage 5:e1074330.

17. Pedulla ML, Ford ME, Houtz JM, Karthikeyan T, Wadsworth C, Lewis JA, Jacobs-Sera D, Falbo J, Gross J, Pannunzio NR, Brucker W, Kumar V, Kandasamy J, Keenan L, Bardarov S, Kriakov J, Lawrence JG, Jacobs WR, Hendrix RW, Hatfull GF. 2003. Origins of highly mosaic mycobacteriophage genomes. Cell 113:171–182.

117 Manuscript III

18. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistance mechanisms. Nat Rev Microbiol 8:317–327.

19. Seed KD. 2015. Battling phages: How bacteria defend against viral attack. PLoS Pathog 11:e1004847.

20. Riede I, Eschbach ML. 1986. Evidence that TraT interacts with OmpA of Escherichia coli. FEBS Lett 205:241–5.

21. Pedruzzi I, Rosenbusch JP, Locher KP. 1998. Inactivation in vitro of the Escherichia coli outer membrane protein FhuA by a phage T5-encoded lipoprotein. FEMS Microbiol Lett 168:119–125.

22. Harvey H, Bondy-Denomy J, Marquis H, Sztanko KM, Davidson AR, Burrows LL. 2018. Bacteria defend against phages through type IV pilus glycosylation 1. Nat Microbiol 3:47–52.

23. Moak M, Molineux IJ. 2000. Role of the Gp16 lytic transglycosylase motif in bacteriophage T7 virions at the initiation of infection. Mol Microbiol 37:345–355.

24. Bebeacua C, Lorenzo Fajardo JC, Blangy S, Spinelli S, Bollmann S, Neve H, Cambillau C, Heller KJ. 2013. X-ray structure of a superinfection exclusion lipoprotein from phage TP-J34 and identification of the tape measure protein as its target. Mol Microbiol 89:152–165.

25. Dy RL, Richter C, Salmond GPC, Fineran PC. 2014. Remarkable mechanisms in microbes to resist phage infections. Annu Rev Virol 1:307–331.

26. Dupuis M-È, Villion M, Magadán AH, Moineau S. 2013. CRISPR-Cas and restriction– modification systems are compatible and increase phage resistance. Nat Commun 4:2087.

27. Tock MR, Dryden DT. 2005. The biology of restriction and anti-restriction. Curr Opin Microbiol 8:466–472.

28. Adriaenssens EM, Ackermann H-W, Anany H, Blasdel B, Connerton IF, Goulding D, Griffiths MW, Hooton SP, Kutter EM, Kropinski AM, Lee J-H, Maes M, Pickard D, Ryu S, Sepehrizadeh Z, Shahrbabak SS, Toribio AL, Lavigne R. 2012. A suggested new bacteriophage genus: Viunalikevirus. Arch Virol 157:2035–46.

29. Wright RCT, Friman V-P, Smith MCM, Brockhurst MA. 2018. Cross-resistance is modular in bacteria–phage interactions. PLOS Biol 16:e2006057.

30. Thomson S V. 2000. Epidemiology of fire blight, p. 9–36. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

31. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald P, Dow M, Verdier V, Beer S V., Machado MA, Toth I, Salmond G, Foster GD. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol 13:614–629.

32. Momol MT, Aldwinckle HS. 2000. Genetic diversity and host range of Erwinia amylovora, p. 55–72. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

118 The role of topB1, rfaE and pgm in phage resistance

33. Bubán T, Orosz-Kovács Z. 2003. The nectary as the primary site of infection by Erwinia amylovora (Burr.) Winslow et al.: a mini review. Plant Syst Evol 238:183–194.

34. Jones AL, Schnabel EL. 2000. The development of streptomycin-resistant strains of Erwinia amylovora, p. 235–251. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

35. Piqué N, Miñana-Galbis D, Merino S, Tomás JM. 2015. Virulence factors of Erwinia amylovora: A review. Int J Mol Sci 16:12836–54.

36. Svircev AM, Castle AJ, Lehman SM. 2010. Bacteriophages for control of phytopathogens in food production systems, p. 79–96. In Sabour, PM, Griffiths, MW (eds.), Bacteriophages in the control of food- and waterborne pathogens. ASM Press, Washington, DC.

37. Boulé J, Sholberg PL, Lehman SM, O’Gorman DT, Svircev AM. 2011. Isolation and characterization of eight bacteriophages infecting Erwinia amylovora and their potential as biological control agents in British Columbia, Canada. Can J Plant Pathol 33:308– 317.

38. Lehman SM. 2007. Ph.D. thesis: Development of a bacteriophage-based biopesticide for fire blight. Department of Biological Sciences, Brock University, St. Catharines.

39. Roach DR, Sjaarda DR, Sjaarda CP, Ayala CJ, Howcroft B, Castle AJ, Svircev AM. 2015. Absence of lysogeny in wild populations of Erwinia amylovoraand Pantoea agglomerans. Microb Biotechnol 8:510–518.

40. Smits THM, Rezzonico F, Kamber T, Blom J, Goesmann A, Frey JE, Duffy B. 2010. Complete genome sequence of the fire blight pathogen Erwinia amylovora CFBP 1430 and comparison to other Erwinia spp. Mol Plant-Microbe Interact 23:384–393.

41. Yagubi A. 2017. Ph.D. thesis: Phage-mediated biological control of Erwinia amylovora: The role of CRISPRs and exopolysaccharide. Brock University.

42. Adams MH. 1959. Bacteriophages. Interscience Publishers Inc., New York.

43. Sambrook J, Russell DW. 2001. Molecular cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

44. Skorupski K, Taylor RK. 1996. Positive selection vectors for allelic exchange. Gene 169:47–52.

45. Dahan S, Wiles S, La Ragione RM, Best A, Woodward MJ, Stevens MP, Shaw RK, Chong Y, Knutton S, Phillips A, Frankel G. 2005. EspJ is a prophage-carried type III effector protein of attaching and effacing pathogens that modulates infection dynamics. Infect Immun 73:679–86.

46. Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison Iii CA, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343.

47. Guzman LLM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J Bacteriol 177:4121–4130.

119 Manuscript III

48. Born Y, Fieseler L, Klumpp J, Eugster MR, Zurfluh K, Duffy B, Loessner MJ. 2014. The tail-associated depolymerase of Erwinia amylovora phage L1 mediates host cell adsorption and enzymatic capsule removal, which can enhance infection by other phage. Environ Microbiol 16:2168–2180.

49. Bellemann P, Bereswill S, Berger S, Geider K. 1994. Visualization of capsule formation by Erwinia amylovora and assays to determine amylovoran synthesis. Int J Biol Macromol 16:290–296.

50. Pusey PL. 1997. Crab apple blossoms as a model for research on biological control of fire blight. Phytopathology 87:1096–102.

51. Llop P, Cabrefiga J, Smits THM, Dreo T, Barbé S, Pulawska J, Bultreys A, Blom J, Duffy B, Montesinos E, López MM. 2011. Erwinia amylovora novel plasmid pEI70: complete sequence, biogeography, and role in aggressiveness in the fire blight phytopathogen. PLoS One 6:e28651.

52. Zengerer V, Schmid M, Bieri M, Müller DC, Remus-Emsermann MNP, Ahrens CH, Pelludat C. 2018. Pseudomonas orientalis F9: A potent antagonist against phytopathogens with phytotoxic effect in the apple flower. Front Microbiol 9:145.

53. Kneidinger B, Marolda C, Graninger M, Zamyatina A, McArthur F, Kosma P, Valvano MA, Messner P. 2002. Biosynthesis pathway of ADP-L-glycero-beta-D-manno-heptose in Escherichia coli. J Bacteriol 184:363–9.

54. Joshi JG, Handler P. 1964. Phosphoglucomutase.I. Purification and properties of phosphoglucomutase from Escherichia coli. J Biol Chem 239:2741–51.

55. Thoden JB, Hegeman AD, Wesenberg G, Chapeau MC, Frey PA, Holden HM. 1997. Structural analysis of UDP-sugar binding to UDP-galactose 4-epimerase from Escherichia coli. Biochemistry 36:6294–304.

56. Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389–402.

57. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J Mol Biol 215:403–410.

58. Kanehisa M, Goto S, Sato Y, Kawashima M, Furumichi M, Tanabe M. 2014. Data, information, knowledge and principle: back to metabolism in KEGG. Nucleic Acids Res 42:D199–D205.

59. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, Jensen LJ, Mering C von. 2019. STRING v11: protein- protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 47:D607–D613.

60. Nakao R, Ramstedt M, Wai SN, Uhlin BE. 2012. Enhanced biofilm formation by Escherichia coli LPS mutants defective in Hep biosynthesis. PLoS One 7:e51241.

61. Jin UH, Chung TW, Lee YC, Ha SD, Kim CH. 2001. Molecular cloning and functional expression of the rfaE gene required for lipopolysaccharide biosynthesis in Salmonella typhimurium. Glycoconj J 18:779–87.

120 The role of topB1, rfaE and pgm in phage resistance

62. El-Gebali S, Mistry J, Bateman A, Eddy SR, Luciani A, Potter SC, Qureshi M, Richardson LJ, Salazar GA, Smart A, Sonnhammer ELL, Hirsh L, Paladin L, Piovesan D, Tosatto SCE, Finn RD. 2019. The Pfam protein families database in 2019. Nucleic Acids Res 47:D427–D432.

63. Ordax M, Marco-Noales E, López MM, Biosca EG. 2010. Exopolysaccharides favor the survival of Erwinia amylovora under copper stress through different strategies. Res Microbiol 161:549–555.

64. Nwodo UU, Green E, Okoh AI. 2012. Bacterial exopolysaccharides: functionality and prospects. Int J Mol Sci 13:14002–15.

65. Koczan JM, Lenneman BR, McGrath MJ, Sundin GW. 2011. Cell surface attachment structures contribute to biofilm formation and xylem colonization by Erwinia amylovora. Appl Environ Microbiol 77:7031–7039.

121

Y2 resistance affects phage infectivity

2.4. Manuscript IV: Y2 resistance affects phage infectivity

Spontaneous persistent resistance of Erwinia amylovora CFBP1430 against Y2 affects infectivity of multiple phages

Leandra E. Knecht1,2, Yannick Born1, Cosima Pelludat3, Joël F. Pothier4, Theo H.M. Smits4 Martin J. Loessner2, Lars Fieseler1*

1 Food Microbiology Research Group, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Wädenswil, Switzerland 2 Institute of Food, Nutrition and Health, ETH Zurich, Switzerland 3 Agroscope, Plant Pathology and Zoology in fruit and vegetable production, Wädenswil, Switzerland 4 Environmental Genomics and Systems Biology Research Group, Institute of Natural Resource Sciences, Zurich University of Applied Sciences (ZHAW), Wädenswil, Switzerland

*Correspondence: Lars Fieseler, Institute of Food and Beverage Innovation, Zurich University of Applied Sciences (ZHAW), Einsiedlerstrasse 31, Wädenswil, Switzerland. Tel: +41 58 934 54 07; e-mail: [email protected]

123 Manuscript IV

Abstract Broad application of antibiotics gave rise to increasing numbers of antibiotic resistant bacteria. Therefore, effective alternatives are currently investigated. Bacteriophages, natural predators of bacteria, could work as such a substitute. Although phages are highly effective in specifically eliminating bacteria, phage resistances can be observed after application. The nature of this resistance however, can differ depending on the phage. Exposing Erwinia amylovora CFBP1430, the causative agent of fire blight, to the phages Bue1, L1, S2, S6 or M7 led to transient resistance. These cells reversed to a phage sensitive state after the phage was eliminated. When wildtype bacteria were incubated with Y2, permanently resistant colonies (1430Y2R) formed spontaneously. In addition, 1430Y2R revealed cross-resistance against other phages (Bue1) or lowered the efficiency of plating (L1, S2, S6). Pull down experiments revealed that Y2 is no longer able to bind to the mutant suggesting mutation or masking of the Y2 receptor. Other phages tested were still able to bind to 1430Y2R. Bue1 was observed to still adsorb to the mutant but no host lysis was found. These findings indicated that, in addition to the alterations of the Y2 receptor, the 1430Y2R mutant might block phage attack at different stage of infection. Whole genome sequencing of 1430Y2R revealed a deletion in the gene with the locus tag EAMY_2231. The gene, which encodes a putative galactosyltransferase, was truncated due to the resulting frameshift. The mutant 1430Y2R was monitored for potential defects or fitness loss. Weaker growth was observed in LB compared to the wildtype but not in minimal medium. 1430Y2R was still highly virulent in blossoms even though exopolysaccharide production was observed to be reduced. Additionally, LPS structures were analysed and were clearly shown to be altered in the mutant. The truncation of EAMY_2231 can therefore be associated with manifold modifications in 1430Y2R, which can affect different phages.

124 Y2 resistance affects phage infectivity

Introduction Bacteriophages are viruses exclusively infecting bacteria. They pose the most abundant biological entity on earth reaching an estimated number of 1031 (1–3) and outnumber bacteria by a tenfold (4). It has been estimated that approximatively 1025 phage infections occur every second (5). Bacteria are therefore forced to adapt to this pressure. The co-evolution of bacteria and phages generates a continuous arms race (6, 7). Bacteria have several mechanisms to interfere with phage infection, which can target the process at different stages (6, 8). Phages recognize their host bacteria by binding to specific receptors on the host surface. These receptors can be proteins, sugar molecules or cell surface structures located either in the outer membrane (9), in capsules (10), or located in appendages such as pili (11) and flagella (12, 13). By mutating or masking the targeted receptor, bacteria can protect themselves from phage infection at this early stage (14–16). Another option to block phage adsorption is the production of competitive inhibitor molecules, that outcompete the phage for receptor binding (17). Bacteria can also reduce the accessibility of the receptor by secreting exopolysaccharides (EPS) covering the receptor (18, 19), or, if challenged with EPS-specific phages, they can reduce the amount of EPS secretion (6). After the attachment to the host receptor, the phage injects its DNA into the host cell. Phage resistance can be established by blocking DNA entry into the cell through superinfection exclusion (sie) systems (20, 21). These systems are usually encoded by prophages to protect the lysogenized host from further phage infections (22). After successful attachment and DNA injection, the bacterial metabolism is hijacked by the phage and modified into producing and assembling new phage particles. If phage DNA reaches the cytoplasm of the host cell, a collection of intracellular defence mechanisms can act to prevent phage replication. Restriction-modification (R-M) systems recognize the incoming foreign DNA and digest it (23). CRISPR/Cas systems function similarly such that incoming phage DNA is identified and cleaved. In contrast to the R-M system, CRISPR/Cas systems are highly specific against certain phages (24). In contrast to the R-M and the CRISPR/Cas system, where the bacteria are able to survive the phage attack, the abortive infection system is fatal for the infected bacteria. These systems are diverse in sensing and reacting to phage infection. However, once activated, the infected bacterium destroys itself thereby preventing phage proliferation (22). This sacrifice ensures survival of the surrounding bacterial population. Phage resistance can occur spontaneously where it was observed to be either transient or permanent (25, 26). Permanent resistance against phages can for example involve genetic mutations. Mutations in genes that encode the receptor or are crucial for receptor biosynthesis prevent infection by phages that would have targeted the particular receptor (27). Many studies however, have associated phage resistance with fitness loss (28, 29). Such a fitness loss could favour the susceptible over the phage resistant state. In the case of transient

125 Manuscript IV resistance, the phage sensitive state will be re-established after elimination of the phage. Bacteria can for example, apply phase variation to generate transient resistance (30, 31). Alterations of prominent structures such as capsules and flagella can be regulated by phase variation as observed in Campylobacter jejuni (32) or Salmonella enterica (33, 34). Modifications in response to a particular phage can potentially entail resistance against other phages. Phages adsorbing to the same receptor will be unable to bind to a modified receptor. Erwinia amylovora is the causative agent of fire blight, a plant disease affecting members of the Rosaceae family (35–39). Fire blight was previously classified as one of the ten economically most damaging plant disease (40). The Gram-negative bacterium has a collection of factors contributing to its pathogenicity such as EPS production and a type three secretion system (41). The pathogen proliferates in the blossoms on the stigmas before entering the plant tissue. Inside the plant tissue, a thick biofilm is produced, which clogs the xylem vessels (42–46). The affected tissues start to desiccate (46). At this stage, the disease spread can only be stopped by pruning of infected tissues or eliminating the entire plant. Antibiotics such as streptomycin are the most efficient treatment of fire blight during the blossoming season. The emergence of antibiotic resistant bacteria (41, 47) resulted in banning the antibiotic for agricultural application in an increasing number of countries (48). Therefore, efficient and environmentally friendly alternatives are currently under investigation. Bacteriophages were shown to have the potential to eliminate E. amylovora (49–51) efficiently and specifically. E. amylovora specific phages Bue1, L1, M7, S2, S6 and Y2 were isolated previously from Swiss orchards (52, 53). The phages are classified as myoviruses (M7, Y2), podoviruses (L1, S2, S6) and ackermannvirus (Bue1) all belonging to different genera (52, 53). All six phages were shown to have a broad host range and are strictly virulent. The phages were tested for their phytotherapeutic effect against E. amylovora in vitro and in blossoms. Different combinations of these phages were shown to have the potential to control bacterial cell counts over a prolonged time. Even though phages are an effective alternative to antibiotics, resistance development should be avoided. In order to prevent quick phage resistance development in the plant pathogens, the underlying mechanisms generating phage resistance and cross-resistance against other phages must be further investigated. This study aimed to anticipate the risk of spontaneous phage resistance by E. amylovora CFBP1430 against a variety of phages. In contrast to Bue1, L1, M7, S2 and S6, the phage Y2 induced a permanent resistance in the bacteria. The obtained mutant further revealed cross-resistance against the other phages.

126 Y2 resistance affects phage infectivity

Materials and Methods

Culture conditions E. amylovora strains were cultivated on LB agar at 28°C, Escherichia coli strains at 37°C (Table 1). Ampicillin was added at a concentration of 100 µg/ml if required.

Soft agar overlay and propagation All phages used in this study are listed in Table 1. The phages were propagated using the soft agar overlay method (54). Four millilitres molten LB+ soft agar (LB broth, 4 g/l agar, 2 mM

MgSO4, 10 mM CaCl2) were supplemented with 90 µl bacterial overnight suspension and 10 µl diluted phage and spread on LB plates to generate semi confluent lysed plates. After overnight incubation, 5 ml SM buffer (100 mM NaCl, 8 mM MgSO4, 50 mM Tris-Cl, pH 7.4) were added per plate and incubated for 5 h at room temperature (RT) under shaking. The supernatant was supplemented with 0.5 M NaCl and incubated for 30 min at RT before centrifugation (10 min, 10,000 xg). Phages in the supernatant were then polyethylenglycol treated (10% w/v PEG8000, ice bath overnight, and pelleting phages 15 min, 10’000 xg, 4°C),

CsCl2 density gradient purified (55) and dialyzed against SM buffer (Bue1, L1, M7, S2 and S6). In the case of Y2, PEG was removed by incubating the PEG-phage mixture at RT for 1 h. The solution was then centrifuged at 5,000 xg for 10 min. Phages in the supernatant were sterile filtered (0.22 µm filter). All phages were stored at 4°C.

Spontaneous resistance To identify spontaneous resistance against phages, an overnight culture of E. amylovora CFBP1430 was diluted to 105 CFU/ml in LB+ broth and supplemented with 108 PFU/ml phages. The suspension was incubated for 5 h under shaking before a dilution row (10-1 to 10- 8) was plated onto LB plates. After overnight incubation, 10 colonies per phage treatment were selected and tested for phage resistance on LB+ soft agar overlays. The supernatants of the overnight cultures were tested for remaining phages. The Y2 resistant mutant termed 1430Y2R was further investigated.

Phage infectivity To investigate the ability of different phages to infect the 1430Y2R strain, 5 µl of phage dilution rows were spotted onto a LB soft agar lawn inoculated with 100 µl 1430Y2R overnight culture. The plates were kept at RT until the spots were dried, before incubation over night at 28°C.

In vitro infection assay The impact of 1430Y2R on phage infectivity in liquid medium was evaluated by in vitro infection assays. Overnight cultures of E. amylovora CFBP1430 and 1430Y2R were washed twice in

127 Manuscript IV

7 sterile SM buffer and OD600 was adjusted to reach approximately 10 CFU/ml. Subsequently, 20 µl of the washed cells were transferred to 1960 µl LB+ broth and supplemented with either 20 µl sterile SM buffer or with 20 µl phages with a concentration of 1010 PFU/ml. The mixtures were then added to 96 well flat bottom plates and incubated at 25°C with double orbital shaking for 24 h in a Biotek Synergy H1 Hybrid plate reader. OD600 was measured every half hour.

Growth curves Impact of the mutations on fitness of 1430Y2R was investigated. Bacteria were washed twice in SM buffer and OD600 was adjusted to 0.1. Cells were diluted in LB medium supplemented

5 with MgSO4 (2 mM) and CaCl2 (10 mM) or MM2 medium (56) to a concentration of 10 CFU/ml. Cultures were incubated at 28°C for 24 h with double orbital shaking (150 rpm) in a plate reader.

Detached flower assay To monitor virulence of the E. amylovora mutant, a detached flower assay using fresh blossoms from 2 year old Golden Delicious apple trees was carried out (57). Blossoms were treated with either E. amylovora CFBP1430, the mutant 1430Y2R or PBS buffer (3 mM KCl, 137 mM NaCl, 2 mM KH2PO4, 10 mM Na2HPO4) as mock infection. Racks were cleaned and autoclaved before the experiment. Alternating 24 wells per rack were filled with 2 ml H2O. The wells were sealed with scotch tape, which was perforated. Stems were cut freshly to ensure water uptake before transferring the blossoms into the filled wells. Bacteria grown overnight on plates were scratched off and resuspended in PBS. OD600 was adjusted to 1.0 and a 1:50 dilution was performed to generate approx. 107 CFU/ml. A total of 20 µl bacterial suspension or PBS were pipetted directly onto the receptacle. The racks were transferred into storage boxes (5 l) which were laid out with paper towels soaked with 100 ml H2O to ensure humidity. The blossoms were incubated at 26°C for 4-5 days before the read out according to an adjusted rating system was carried out (58): Healthy blossoms without disease symptoms are classified as grade 1. Visible symptoms on the blossom (browning of the calix) are referred to as grade 2. Blossoms with disease symptoms in the calix and the stipe of the blossoms correspond to grade 3 (59).

Whole genome sequencing Identification of genetic modifications in 1430Y2R was achieved by whole genome sequencing. DNA was extracted as described previously (52) and sheared into 550 bp fragments on a Covaris E220 (Covaris, Woburn, MA). Libraries were prepared on an Illumina NeoPrep System (Illumina, San Diego, CA) using a TruSeq Nano DNA kit (Illumina) with six PCR cycles according to manufacturer’s instructions. Paired-end sequencing of 300 bp was performed on a MiSeq instrument (Illumina) using a 600-cycle MiSeq Reagent Kit v3 (Illumina) following

128 Y2 resistance affects phage infectivity manufacturer’s instructions. Mapping of the reads was done against the earlier published genome of CFBP 1430 (GenBank accession numbers FN434113 (chromosome) and FN434114 (pEa29) (60) using SeqMan Ngen v12 (DNASTAR, Madison, WI, USA) and checked for single nucleotide polymorphisms (SNPs) using the DNASTAR LASERGENE Package subroutine SeqMan. A total of three SNPs that were differential between both sequenced strains are reported in this study.

Complementation of 1430Y2R mutants The plasmid pBAD18 (61) was linearized with EcoRI and HindIII and purified. The gene with the locus tag EAMY_2231 and its ribosomal binding site were amplified from E. amylovora CFBP1430 by PCR with Gibson primers (NEBuilder Assembly Tool v1.12.18) EAMY_2231 fw

(5’-TGGGCTAGCGAATTCGAGCTCAGGAGGTCGTAATGCATAAGATCTGCTATTTC-3’) and EAMY_2231 rev (5’-

TGCATGCCTGCAGGTCGACTCTAGACTATATTAATTCGTTATAGGCGG-3’) using the KAPA HIFITM PCR kit (KAPA Biosystems, Wilmington, USA). The PCR product with the correct length was recovered from a 1% agarose gel using the DNA Clean & Concentrator™-5 Kit by Zymo Research. The linearized vector pBAD18 and the insert were joined by Gibson assembly (62). The newly formed plasmid was introduced into electrocompetent E. coli XL1-Blue cells for amplification. Cells were recovered in SOC and incubated for 1 h at 37°C under vigorous shaking before plating onto LB plates containing ampicillin. Correct plasmid insertion was verified by PCR with the primer pair pBAD fw (5’-CTGTTTCTCCATACCCGTT-3’) and pBAD rev (5’-

CTCATCCGCCAAAACAG-3’). Correct complementation plasmid pBAD::EAMY_2231 was extracted with the NucleoSpin® Plasmid Kit (Macherey-Nagel; Düren Germany) and introduced into electrocompetent E. amylovora CFBP1430 and 1430Y2R. All generated constructs and strains are listed in Table 1. Spot on the lawn experiments were carried out. Dilution rows of the six different phages were spotted on the wildtype, the empty vector control, 1430Y2R, 1430Y2R [pBAD::EAMY_2231] and 1430 [pBAD::EAMY_2231] and their plaque forming ability was monitored. Soft agar and LB plates were supplemented with ampicillin and 0.2% arabinose for pBAD18 promoter induction.

Phage adsorption pull down assay To investigate the adsorption of the phages to phage resistant mutants, pull down assays were carried out (63). E. amylovora CFBP1430 was used as a positive control, medium only was used as a negative control. Cells from overnight cultures were washed twice with LB medium

9 and diluted to an OD600 of 1.0. Ten microliters of Bue1, L1, S2 and Y2 (10 PFU/ml) were added to 990 µl bacteria. After 10 min of incubation at RT under shaking, samples were centifuged for 5 min at 10,000 xg at 4° C. Free unbound phages in the supernatant were quantified using the soft agar overlay method plating on E. amylovora CFBP1430.

129 Manuscript IV

Table 1 Overview of the used strains, generated mutants and applied phages

Bacterial strains Characterization Reference E amylovora Wild type strain, isolated in France from Crataegus sp; CFBP1430 Zhang and Geider, 1997 propagation strain for all phages CFBP1430 [pBAD18] Empty vector control This work CFBP1430Y2R Spontaneous permanent Y2 resistant strain This work CFBP1430+ EAMY_2231 Overexpression of EAMY_2231 This work CFBP1430Y2R+EAMY_2231 Complementation of 1430Y2R with EAMY_2231 This work 4/82 Isolated in Egypt from Pyrus communis; low EPS-producer Zhang and Geider, 1997

E. coli supE44 ∆lacU169 (φ80lacZ∆M15) hsdR17 recA1 endA1 Dh5α Hanahan, 1983 gyrA96, thi-1 relA1 XL1-Blue MRF’ NEB (Ipswich, USA)

Plasmids Complementation vector, arabinose induced pBAD pBAD18 Guzman et al 1995 promoter Complementation vector for EAMY_2231,arabinose pBAD18::EAMY_2231 This work induced pBAD promoter Phages Bue1 E. amylovora specific Ackermann virus, Vi1-like Knecht et al, 2018 L1 E. amylovora specific podovirus, T7-like Born et al, 2011 M7 E. amylovora specific myovirus, FO1-like Born et al, 2011 S2 E. amylovora specific myovirus, SP6-like Knecht et al, 2018 S6 E. amylovora specific podovirus, N4-like Born et al, 2011 Y2 E. amylovora specific myovirus, GJ1-like Born et al, 2011

EPS quantification through EPS-CPC precipitation Impact of EAMY_2231 modifications on EPS production was monitored using the EPS- Cetylpyridiniumchlorid (CPC) precipitation assay (56). Bacteria were grown in MM2 minimal medium composed of K2HPO4 (1.6 g), NaCl (3.0 g), Sorbitol (10 g), Nicotic acid (0.2 mg/ml),

Thiamine hydrochloride (0.2 mg/ml), L-asparagine (4 mg/ml) and MgSO4 (0.205 mg/ml) per litre

H2O (56) at 28°C under shaking for 24 h. OD600 of all samples were adjusted to 1.0 and 1 ml per sample was centrifuged at 10,000 xg for 5 min. A total of 950 µl supernatant of each sample was mixed with 50 µl CPC (50 mg/ml) and incubated for 10 min at RT before OD600 was measured. As control, the low-EPS producing E. amylovora strain 4/82 was used.

LPS extraction and analysis Alterations in the LPS structure were analysed by silver staining. Cells were grown overnight in LB medium and washed twice in PBS buffer before the OD600 was adjusted to 1.0. A total of 1 ml per sample was centrifuged at 8,000 x g for 5 min. The pellet was resuspended in 100 µl SDS sample buffer (90 mM Tris base, 2% SDS, 0.02% Bromophenol blue, 20% sucrose, pH adjusted to 6.8 in H2O) and boiled for 10 min at 100°C. Samples were cooled down to room temperature before 2.5 µl proteinase K (20 µg/µl) were added. After incubation at 60°C for 1 h, 10 µl of sample were loaded onto an SDS-PAGE gel (12% resolving/4% stacking gel) and

130 Y2 resistance affects phage infectivity

let run with 35 A for 2 h. The gels were quickly washed in ddH2O before being soaked with fresh fixing solution (40% ethanol and 5% acetic acid in H2O) for 1 h. The fixing solution was subsequently replaced with fresh oxidizing solution (fixing solution supplemented with 30 mM periodic acid) for 5 min. After the incubation, the gels were washed 3 to 5 times with at least

500 ml of ddH2O for 15 min to completely remove the oxidizing solution. The gels were soaked in freshly prepared staining solution (1.5 ml ammonium hydroxide solution 33%, 14 mM NaOH solution, 0.5% AgNO3 in 200 ml H2O) for 15 min. After washing the gels 3 times in ddH2O, gels were developed with freshly prepared developer solution (200 ml ddH2O supplemented with 50 mg citric acid and 100 µl formaldehyde solution) until bands appeared. The development was stopped with several charges of ddH2O.

131 Manuscript IV

Results Spontaneous Y2 resistance in CFBP1430 generates cross resistance Spontaneous resistance development in E. amylovora CFBP1430 was tested for the phages Bue1, L1, M7, S2, S6 and Y2. Spontaneous resistant bacteria were generated for all six phages. The observed resistances against Bue1, L1, M7, S2 and S6 were all shown to be transient. Colonies were anew phage sensitive after the phages were eliminated. Bacteria exposed to Y2 spontaneously generated permanent phage resistant colonies. Even in the absence of Y2, the resistance was maintained. The strain termed 1430Y2R could be passaged several times without losing the phage resistant trait. Cross resistance of 1430Y2R against other phages was tested by spotting phage dilutions on 1430Y2R lawns. Indeed, the 1430Y2R strain could no longer be infected by Bue1. In addition, reduced efficiency of plating for the two phages L1 and S2 was observed. M7 and S6 infectivity did not seem to be affected by the 1430Y2R modification. Furthermore, in vitro infection assays were carried out to verify possible cross-resistance (Figure 1). Growth of the wildtype strain incubated with all six phages was compared to the growth of 1430Y2R in the presence of the phages. Y2 had a strong impact on CFBP1430 cells, preventing regrowth for 25 h. The mutant 1430Y2R however, was able to induce exponential growth in the presence of Y2. Nevertheless, weaker growth was observed for 1430Y2R incubated with Y2 compared to 1430Y2R without Y2. Incubation with Bue1 generated similar results as the Y2 incubation. Bue1 is able to prevent wildtype growth for up to 18 hours, whereas no effect was observed, when incubated with 1430Y2R. The results for Bue1 infection are in accordance with the spotting results. L1 and S2 infectivity was observed to be weaker in 1430Y2R. In contrast to the spotting assays, the S6 infectivity was strongly impacted in the 1430Y2R mutant. When S6 was incubated with the wildtype, the phage was able to maintain the cell counts below detection level for up to 22 h before regrowth was observed. When mixed with the1430Y2R mutant however, the strain was able to regrow after 13 hours.

Figure 1 In vitro infection assay. Growth of A) E. amylovora CFBP1430 or B) 1430Y2R infected with phages. Bacterial concentrations of 105 CFU/ml were infected with 108 PFU/ml phage and incubated over 30 hours. Optical density (OD600) was measured regularly at 30-minute intervals. Error bars indicate standard deviations

132 Y2 resistance affects phage infectivity

Both strains were successfully infected and controlled by M7. This result suggests that infectivity of Bue1 and Y2 is completely abolished and L1, S2 and S6 infectivity is reduced in the 1430Y2R strain. Only M7 infectivity was shown to be unaffected by the modification.

Spontaneous Y2 resistance has little fitness impact To test whether the mutation in the 1430Y2R strain has an influence on the fitness of the mutants growth and virulence was monitored. Bacteria were tested for growth defects in LB and MM2 (Figure 2). While no difference could be observed in MM2 medium between CFBP1430 and 1430Y2R, the mutant strain grew weaker in LB medium.

Figure 2 Growth curves of E. amylovora CFBP1430 and 1430Y2R in A) LB or B) MM2 medium. Cell counts were adjusted to 105 CFU/ml before incubation at 28°C under double orbital shaking for 24 h in a plate reader. Optical density (OD600) was measured at 600 nm every 30 min. Each mutant was measured twice independently with 8 replicates per run. Error bars indicate the standard deviation.

Virulence was tested on fresh apple blossoms. As control, the blossoms were mock infected with PBS only. The wildtype was able to trigger disease symptoms spreading to the stipe of the blossoms (stage 3) in 98.8% of the infected blossoms (Figure 3). The 1430Y2R strain generated 88.6% stage 3 disease symptoms in the infected blossoms.

Figure 3 Virulence of 1430Y2R in detached apple blossoms. Blossoms were infected with 107 CFU/ml bacteria. Read out was performed by evaluating disease symptoms. Grade 1 indicates healthy blossoms. Grade 2 refers to blossoms with first disease symptoms such as browning of the calix. Blossoms showing advanced disease symptoms on both calyx and stipe correspond to Grade 3. Each treatment was tested on 60 individual blossoms.

133 Manuscript IV

EAMY_2231, the source of phage resistance In order to identify genetic modifications in 1430Y2R that renders the mutant phage resistant, whole genome sequencing applied for both the 1430Y2R mutant and its wildtype as control. The whole genome sequencing revealed three single nucleotide polymorphisms (SNP) in 1430Y2R as compared to the wildtype strain. A silent mutation was found in the gene rpIV, a ribosomal protein. The second SNP is located in the gene ytfB of 1430Y2R. It contains an adenosine instead of a guanosine at the gene position 472. This mutation results in a glycine to serine transition. The I-TASSER protein structure & function prediction tool (64) was used to evaluate the impact of the SNP on the predicted structure of YtfB and suggested only minor or no alterations due to the mutation. The third SNP is a nucleotide deletion in the gene EAMY_2231. The generated frameshift truncates the protein to 127 amino acids whereby the frameshift affects 20 C-terminal amino acids. The gene EAMY_2231 is annotated to encode a glycosyltransferase family 1. The encoded protein harbours two separate glycosyltransferase domains. Proteins encoding such domains are responsible for transferring UDP-, ADP-, GDP- or CDP-linked sugars to a variety of substrates and are important in biosynthetic processes. These proteins can be involved in the synthesis of exopolysaccharides or lipopolysaccharide cores (65, 66). It is hypothesized that the encoded protein EAMY_2231 could, similarly to cap1E in Streptococcus pneumoniae, be required for the synthesis of type 1 capsular polysaccharides (67, 68). Concluding from these results, the SNP in EAMY_2231 is the most promising cause for the observed phage resistance. A deletion of the entire gene was attempted. Although different approaches were tested, no mutant could be recovered. A complementation of the complete EAMY_2231 gene was generated to recover the 1430Y2R modification.

Table 2 Spotting of phages on the generated mutants. – symbolizes no plaque formation. Visible plaques are indicated by +. ++ indicate plaque formation visible on the dilution 10-2 and 10-4. +++ (up to 10-6) and ++++ (up to 10-8). Plaque formation was tested on LB (top) or LB supplemented with 0.2% arabinose (bottom).

Y2R 1430 Y2R 1430 LB wt vector control 1430 +EAMY2231 +EAMY2231 M7 ++++ ++++ ++++ ++++ ++++ S6 ++++ +++ +++ +++ +++ Bue1 ++ ++ ++ - ++ L1 ++++ ++++ ++++ ++ ++++ S2 ++++ ++++ ++++ ++ ++++ Y2 ++ ++ ++ - ++

Y2R 1430 Y2R 1430 LB Ara 0.2% wt vector control 1430 +EAMY2231 +EAMY2231 M7 ++++ ++++ ++++ ++++ ++++ S6 +++ ++++ ++++ +++ +++ Bue1 +++ +++ ++ - ++ L1 ++++ ++++ ++++ +++ ++++ S2 ++++ ++++ ++++ +++ ++++ Y2 +++ ++ ++ - ++

134 Y2 resistance affects phage infectivity

Complementation of EAMY_2231 in 1430Y2R rendered the mutant Y2 and Bue1 sensitive again (Table 2). The wildtype phenotype could therefore be re-established. Furthermore, the complementation of EAMY_2231 generated comparable infectivity levels to the wildtype for L1, S2 and S6 in spotting assays. The impact of the EAMY_2231 modifications on the phages interactions were investigated (Figure 4).

Figure 4 Phage impact on growth of different strains. The wildtype, the empty vector control, the overexpression of EAMY_2231 the spontaneous Y2 resistant strain 1430Y2R and the complementation of 1430Y2R+EAMY_2231 were incubated with the phages Bue1, L1, S2, S6, M7 and Y2 or without phages as control. Error bars indicated standard deviation.

135 Manuscript IV

Bue1 was able to successfully infect the wildtype, the empty vector control, 1430+EAMY_2231, and the non induced strain 1430Y2R+EAMY_2231. Weak infection could be observed when Bue1 is incubated with the induced 1430Y2R+EAMY_2231. Bue1 had no effect on 1430Y2R. L1 and S2 were observed to have no or very weak impact on all the tested mutants. M7 was able to completely control all tested mutants for up to 40 hours. In the last hour of incubation a weak regrowth of 1430Y2R but not the other strains was observed, suggesting M7 infectivity might be weakly affected by the modification found in 1430Y2R. S6 generated similar results as M7. The only exceptions are 1430Y2R where only weak infectivity was measured and the non induced complementation of 1430Y2R+EAMY_2231 where S6 was unable to control mutant growth after 20 hours. Y2 was able to prevent mutant growth for up to 20 hours of incubation. No infectivity was observed when Y2 was incubated with 1430Y2R. In addition to the optical density, cell counts were monitored in the first hours of infection (data not shown). Y2 is shown to be more infective when EAMY_2231 is complemented in 1430Y2R or EAMY_2231 is overexpressed in the wildtype. Generally, the growth of mutants with an induced EAMY_2231 complementation vector were observed to be weaker than the induced vector control. This suggests strong growth deficiencies when EAMY_2231 is over expressed. The non induced complementation of 1430Y2R+EAMY_2231 however generated comparable results to the empty vector control. This indicates both that the used promotor could be leaky and that EAMY_2231 is required in low concentrations to generated wildtype phenotype. To test whether the generated modifications have an effect on phage adsorption, pull down assays were carried out (Figure 5). L1 showed stable binding towards the empty vector control, the overexpressed EAMY_2231 mutant and the complemented 1430Y2R. A reduction of adsorption of L1 was observed to the 1430Y2R strain and the non-induced complemented mutant. Similar results were obtained for S2. Y2 was shown to be unable to bind to the 1430Y2R mutant. Complete adsorption was restored in both the non-induced and the induced complemented mutant. No alteration in adsorption was observed in the EAMY_2231 overexpressed strain. Pull down experiments carried out for Bue1 showed weaker adsorption to the wildtype compared to the other phages tested. Although adsorption was observed to be generally weaker for Bue1, the phage was able to bind comparably to 1430Y2R and the other tested mutants as towards the wildtype. Weaker binding affinity was observed for the empty vector control. Pull down experiments carried out with M7 and S6 did not result in reproducible data and were therefore excluded.

136 Y2 resistance affects phage infectivity

Figure 5 Pull down experiment of all generated mutants against L1, S2, Bue1 and Y2. The ability of these four phages to bind to the different mutants. Error bars indicate standard deviation. Statistical analysis was done by an one way ANOVA p<0.0001. Each sample was tested at least three times independently.

EAMY_2231 affects both EPS and LPS The EPS levels of the generated mutants were tested by EPS-CPC precipitation. As control, the low EPS producing strain E. amylovora 4/82 was used. Compared to the wildtype and the empty vector control, a clear reduction in EPS production could be observed in the EAMY_2231 overexpressed mutant and the 1430Y2R strain (Figure 6). The non-induced complementation of EAMY_2231 in 1430Y2R restored EPS production to wildtype level. However, the induced complementant generated similar EPS amounts as the 1430Y2R strain. These results suggest a connection between EAMY_2231 and EPS production.

137 Manuscript IV

Figure 6. EPS-CPC precipitation. Quantitative EPS measurements were carried out for the generated mutants. Produced EPS was precipitated with CPC and measured by optical density. The EPS lacking strain 4/82 was used as negative control alongside CFBP1430 and CFBP1430 carrying the empty vector as positive controls. Error bars indicate standard deviation. Statistical analysis was done by an ordinary one way ANOVA comparing samples to the wildtype p<0.0001. Each sample was tested at least three times independently.

LPS extraction and analysis was carried out to verify if the EAMY_2231 modification results in altered LPS structures. Modifications in the Lipid A and the LPS core of 1430Y2R could be observed (Figure 7). An altered band pattern is visible for the intermediate and the long O- antigen section of the 1430Y2R strain. The complemented 1430Y2R strains (induced and non- induced) both exhibit a comparable band pattern as the wildtype, the empty vector control and the EAMY_2231 overexpressed mutant.

Figure 7 LPS silver staining of the 1430Y2R related mutants. Silver staining of the wildtype (wt), empty vector control, +EAMY_2231, 1430Y2R and the complementation of 1430Y2R+EAMY_2231 revealed alterations of the LPS pattern. Image of the left gel was taken by an Azure biosystems C300, whereas the image on the right was taken by a standard photo camera.

138 Y2 resistance affects phage infectivity

Discussion Spontaneous resistance development against phages can occur quickly since bacteria adapt to the phage pressure. Bacteria can apply a collection of strategies to circumvent phage infection. Different stages of the phage infection process can be targeted to generate phage resistance. In this study, E. amylovora CFBP1430 was tested for its potential to induce spontaneous resistance against six different phages. In the case of Bue1, L1, M7, S2 and S6, the obtained resistance was observed to be transient. After removal of all phages, the spontaneous resistant bacteria quickly reverted to a phage sensitive state. This was not the case for cells incubated with Y2. Y2 resistant cells were observed to be permanent phage resistant and could be passaged multiple times and stored without losing the phage resistant phenotype. In addition to the permanent resistance against Y2, the 1430Y2R strain was cross- resistant against other phages. Bue1 was no longer able to infect this mutant. L1, S2 and S6, although still able to infect, were shown to have a reduced infectivity. In order to identify possible permanent alterations in 1430Y2R, whole genome sequencing was carried out and revealed three SNPs in the 1430Y2R genome compared to the wildtype. The most promising mutation was the nucleotide deletion resulting in a frame shift in the gene EAMY_2231. Sequence similarities of the gene EAMY_2231 suggest that the encoded protein can be classified as glycosyltransferase family 1 (69). These proteins can be involved in synthesis of exopolysaccharides or lipopolysaccharide core.

Figure 8 Genome section of E. amylovora CFBP1430. The section shows the gene with the locus tag EAMY_2231 (yellow) and the surrounding genes (blue). The red triangle indicates the location of the deleted nucleotide responsible for the frameshift. The section has a length of 33670 bp.

EAMY_2231 is located in close proximity to the ams operon required for amylovoran biosynthesis (Figure 8). Additionally, the STRING database (string-db.org, Version 11.0) indicates interactions between EAMY_2231 and the proteins encoded by amsA (chain length determinant protein) and amsH (polysaccharide export protein) both involved in the amylovoran biosynthesis (60, 70). This suggests an impact on amylovoran production in 1430Y2R. The generated amount of EPS in the mutants was measured and a reduction in EPS in 1430Y2R could be observed. This reduction could account for the reduced adsorption of L1 and S2 to the mutant strain and the reduced efficiency of plating, since both phages require amylovoran for host recognition (chapter 1). The complementation of the gene EAMY_2231 however could only establish similar amounts of EPS as the wildtype when the promoter was not induced. The induction of EAMY_2231 in the complemented 1430Y2R and the

139 Manuscript IV overexpression of EAMY_2231 in the wildtype generated reduced levels of secreted EPS similar to the 1430Y2R mutant. EAMY_2231 might be involved in different metabolic processes, which account for the reduction of EPS when EAMY_2231 is overexpressed. The current state of knowledge is that EPS is tightly linked to virulence on apple blossoms (71). Although the secreted EPS amount in the phage resistant 1430Y2R is reduced, virulence was shown to be unaffected, indicating that the amount and composition of the produced EPS is sufficient to trigger full disease outbreak in blossoms. M7 and S6 were shown to rely on cellulose and the cellulose synthase complex (chapter 2). Cellulose production was monitored and 1430Y2R was shown to produce normal levels of cellulose (data not shown). This explains why M7 can still successfully lyse the mutants. S6, on the other hand, was observed to have a reduced infectivity in vitro. These findings suggest that aside from cellulose or the cellulose synthase operon, correct EAMY_2231 function is required for optimal S6 infectivity. Y2 and Bue1 were unable to infect 1430Y2R. Both Y2 and Bue1 are suggested to recognize their host cells through particular LPS structures (chapter 1). This strongly indicates EAMY_2231 involvement in LPS biosynthesis. Three genes in close proximity of EAMY_2231, namely epsF, rfbB1 and rfbA1, are annotated to be involved in LPS biosynthesis (69). rfbB1 and rfbA1 are supposedly involved in the O-antigen export system as ABC transporter. The gene epsF encodes a glycosyltransferase family 8 and is suggested to be required for LPS biosynthesis. The string database revealed a protein-protein interaction of RfsA and RfaI, both homologous proteins of epsF and EAMY_2231, respectively in E.coli K12 (72). Therefore, the produced LPS was controlled for modifications in the generated bacterial strains. Indeed strong alterations in the band pattern of 1430Y2R could be observed. The LPS structure of 1430Y2R will be further analysed in detail to identify the modifications that render the strain phage resistant. Uncovering the exact LPS structure of 1430Y2R will reveal the particular receptors recognized by Y2 and Bue1. The screening or a Tn5 transposon mutagenesis library of E. amylovora CFBP1430 for phage resistant mutants also identified mutants with disrupted EAMY_2231 to be Y2 and Bue1 resistant (chapter 1). It is however, unclear why the permanent resistance only occurs after Y2 exposure. Especially since the modification of EAMY_2231 also renders the mutant Bue1 resistant. The fact that Bue1 exposure results in spontaneous resistance instead of permanent resistance as observed with Y2, suggests that different mechanisms are involved in mediating resistance towards these two phages. Resistance against Y2 is clearly achieved by modification of the phage receptor. Pull down experiments revealed that Y2 is unable to adsorb when EAMY_2231 is truncated. Complementing EAMY_2231 in 1430Y2R fully recovered the binding affinity of Y2 to wildtype levels. Experiments measuring bacterial cell counts upon phage infection revealed that Y2 had a much stronger effect on bacteria when

140 Y2 resistance affects phage infectivity

EAMY_2231 was overexpressed. We hypothesize that EAMY_2231 positively regulates Y2 infection by directly contributing to Y2 receptor synthesis. Bue1 on the other hand could adsorb to both the wildtype and 1430Y2R comparably. The truncation of EAMY_2231 is therefore not affecting the binding affinity of Bue1. Although Bue1 is able to bind to both the wildtype and 1430Y2R, the phage is unable to lyse the latter, indicating that the mutant must use an alternative strategy to resist Bue1 infection. Concluding these observations, we hypothesize that EAMY_2231 directly affects Y2 binding. Overexpression of EAMY_2231 increases the infectivity of Y2. Bue1 does not require EAMY_2231 for adsorption but rather for successful infection in a yet unknown fashion. Phage resistance can occur at different stages. Since 1430Y2R affects a single gene, it is unlikely that the R-M system or the Abi system are responsible for the resistance development against Bue1 in 1430Y2R. The lack of incorporated Y2 DNA in 1430Y2R excludes the possibility of sie or CRISPR/Cas involvement. The fact that the gene responsible for the resistance has strong similarities to a glycosyltransferase suggests that resistance against Bue1 in 1430Y2R is probably not intracellularly but at the DNA injection stage. We propose to perform further experiment to verify if Bue1 DNA can be injected into 1430Y2R or not. Understanding interactions between phages and bacteria and investigating the applied resistance mechanisms towards the phages are important to elucidate and anticipate the observed evolutionary arms-race between phages and bacteria. Our experiments could identify the gene EAMY_2231 as main source for the spontaneous and permanent Y2 resistance. We could further show that modification of EAMY_2331 affects four other phages. In contrast to other findings, the observed phage resistance in 1430Y2R is only minimally linked to fitness loss in the mutant. LPS silver staining, phage-sensitivity and gene location suggest that EAMY_2231 is involved in LPS modification. Additionally, we could show that 1430Y2R produces lower amounts of EPS compared to the wildtype. All these results suggest that EAMY_2231 might be involved in several intracellular processes. A possible explanation for these observations could be the fact that EAMY_2231 harbours two glycosyltransferase domains. Glycosyltransferases are specialized to transfer particular sugar residues to a specific acceptor substrate (73).The domains incorporated in EAMY_2231 could be involved in glycosylating peptidoglycan, lipopolysaccharides and EPS capsules. This might explain why the EAMY_2231 modification could have such a profound impact on different phages. We therefore propose that EAMY_2231 plays a crucial role in mediating resistance or reduced infectivity against multiple phages simultaneously.

Acknowledgments

This work was funded by the Swiss National Science Foundation (SNF) grant 310030_156947.

141 Manuscript IV

References

1. Hatfull GF. 2008. Bacteriophage genomics. Curr Opin Microbiol 11:447–453.

2. Cobián Güemes AG, Youle M, Cantú VA, Felts B, Nulton J, Rohwer F. 2016. Viruses as winners in the game of life. Annu Rev Virol 3:197–214.

3. Casjens SR. 2005. Comparative genomics and evolution of the tailed-bacteriophages. Curr Opin Microbiol 8:451–458.

4. Brüssow H, Hendrix RW. 2002. Phage genomics: small is beautiful. Cell 108:13–6.

5. Pedulla ML, Ford ME, Houtz JM, Karthikeyan T, Wadsworth C, Lewis JA, Jacobs-Sera D, Falbo J, Gross J, Pannunzio NR, Brucker W, Kumar V, Kandasamy J, Keenan L, Bardarov S, Kriakov J, Lawrence JG, Jacobs WR, Hendrix RW, Hatfull GF. 2003. Origins of highly mosaic mycobacteriophage genomes. Cell 113:171–182.

6. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistance mechanisms. Nat Rev Microbiol 8:317–327.

7. Samson JE, Magadán AH, Sabri M, Moineau S. 2013. Revenge of the phages: defeating bacterial defences. Nat Rev Microbiol 11:675–687.

8. Seed KD. 2015. Battling phages: How bacteria defend against viral attack. PLoS Pathog 11:e1004847.

9. Marti R, Zurfluh K, Hagens S, Pianezzi J, Klumpp J, Loessner MJ. 2013. Long tail fibres of the novel broad-host-range T-even bacteriophage S16 specifically recognize Salmonella OmpC. Mol Microbiol 87:818–834.

10. Fehmel F, Feige U, Niemann H, Stirm S. 1975. Escherichia coli capsule bacteriophages. VII. Bacteriophage 29-host capsular polysaccharide interactions. J Virol 16:591–601.

11. Guerrero-Ferreira RC, Viollier PH, Ely B, Poindexter JS, Georgieva M, Jensen GJ, Wright ER. 2011. Alternative mechanism for bacteriophage adsorption to the motile bacterium Caulobacter crescentus. Proc Natl Acad Sci U S A 108:9963–8.

12. Shin H, Lee JH, Kim H, Choi Y, Heu S, Ryu S. 2012. Receptor diversity and host interaction of bacteriophages infecting Salmonella enterica serovar Typhimurium. PLoS One 7:e43392.

13. Bertozzi Silva J, Storms Z, Sauvageau D. 2016. Host receptors for bacteriophage adsorption. FEMS Microbiol Lett 363:fnw002.

14. Riede I, Eschbach ML. 1986. Evidence that TraT interacts with OmpA of Escherichia coli. FEBS Lett 205:241–5.

15. Pedruzzi I, Rosenbusch JP, Locher KP. 1998. Inactivation in vitro of the Escherichia coli outer membrane protein FhuA by a phage T5-encoded lipoprotein. FEMS Microbiol Lett 168:119–125.

16. Harvey H, Bondy-Denomy J, Marquis H, Sztanko KM, Davidson AR, Burrows LL. 2018. Bacteria defend against phages through type IV pilus glycosylation 1. Nat Microbiol 3:47–52.

142 Y2 resistance affects phage infectivity

17. Destoumieux-Garzón D, Duquesne S, Peduzzi J, Goulard C, Desmadril M, Letellier L, Rebuffat S, Boulanger P. 2005. The iron–siderophore transporter FhuA is the receptor for the antimicrobial peptide microcin J25: role of the microcin Val 11 –Pro 16 β-hairpin region in the recognition mechanism. Biochem J 389:869–876.

18. Forde A, Fitzgerald GF. 2003. Molecular organization of exopolysaccharide (EPS) encoding genes on the lactococcal bacteriophage adsorption blocking plasmid, pCI658. Plasmid 49:130–42.

19. Scholl D, Adhya S, Merril C. 2005. Escherichia coli K1’s capsule is a barrier to bacteriophage T7. Appl Environ Microbiol 71:4872–4874.

20. Moak M, Molineux IJ. 2000. Role of the Gp16 lytic transglycosylase motif in bacteriophage T7 virions at the initiation of infection. Mol Microbiol 37:345–355.

21. Bebeacua C, Lorenzo Fajardo JC, Blangy S, Spinelli S, Bollmann S, Neve H, Cambillau C, Heller KJ. 2013. X-ray structure of a superinfection exclusion lipoprotein from phage TP-J34 and identification of the tape measure protein as its target. Mol Microbiol 89:152–165.

22. Dy RL, Richter C, Salmond GPC, Fineran PC. 2014. Remarkable mechanisms in microbes to resist phage infections. Annu Rev Virol 1:307–331.

23. Tock MR, Dryden DT. 2005. The biology of restriction and anti-restriction. Curr Opin Microbiol 8:466–472.

24. Dupuis M-È, Villion M, Magadán AH, Moineau S. 2013. CRISPR-Cas and restriction– modification systems are compatible and increase phage resistance. Nat Commun 4:2087.

25. Moineau S. 1999. Applications of phage resistance in lactic acid bacteria, p. 377–382. In Lactic Acid Bacteria: Genetics, Metabolism and Applications. Springer Netherlands, Dordrecht.

26. Oechslin F. 2018. Resistance development to bacteriophages occurring during bacteriophage therapy. Viruses 10:351.

27. Nesper J, Kapfhammer D, Klose KE, Merkert H, Reidl J. 2000. Characterization of Vibrio cholerae O1 antigen as the bacteriophage K139 receptor and identification of IS1004 insertions aborting O1 antigen biosynthesis. J Bacteriol 182:5097–5104.

28. Koskella B, Lin DM, Buckling A, Thompson JN. 2012. The costs of evolving resistance in heterogeneous parasite environments. Proc R Soc B Biol Sci 279:1896–1903.

29. Smits THM, Rezzonico F, Kamber T, Blom J, Goesmann A, Ishimaru CA, Frey JE, Stockwell VO, Duffy B. 2011. Metabolic versatility and antibacterial metabolite biosynthesis are distinguishing genomic features of the fire blight antagonist Pantoea vagans C9-1. PLoS One 6:e22247.

30. Seed KD, Faruque SM, Mekalanos JJ, Calderwood SB, Qadri F, Camilli A. 2012. Phase variable O antigen biosynthetic genes control expression of the major protective antigen and bacteriophage receptor in Vibrio cholerae O1. PLoS Pathog 8:e1002917.

31. van der Woude MW, Bäumler AJ. 2004. Phase and antigenic variation in bacteria. Clin Microbiol Rev 17:581–611.

143 Manuscript IV

32. Gencay YE, Sørensen MCH, Wenzel CQ, Szymanski CM, Brøndsted L. 2018. Phase variable expression of a single phage receptor in Campylobacter jejuni NCTC12662 influences sensitivity toward several diverse CPS-dependent phages. Front Microbiol 9:82.

33. Kim M, Ryu S. 2012. Spontaneous and transient defence against bacteriophage by phase-variable glucosylation of O-antigen in Salmonella enterica serovar Typhimurium. Mol Microbiol 86:411–425.

34. Cota I, Blanc-Potard AB, Casadesús J. 2012. STM2209-STM2208 (opvAB): A phase variation locus of Salmonella enterica involved in control of O-antigen chain length. PLoS One 7:e36863.

35. Bonn WG, van der Zwet T. 2000. Distribution and economic importance of fire blight, p. 37–53. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

36. Vanneste JL. 2000. What is fire blight? Who is Erwinia amylovora? How to control it?, p. 1–6. In Fire Blight. CAB International, Wallingford.

37. Gill JJ, Svircev AM, Smith R, Castle AJ. 2003. Bacteriophages of Erwinia amylovora. Appl Environ Microbiol 69:2133–2138.

38. Momol MT, Aldwinckle HS. 2000. Genetic diversity and host range of Erwinia amylovora, p. 55–72. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

39. Thomson S V. 2000. Epidemiology of fire blight, p. 9–36. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

40. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald P, Dow M, Verdier V, Beer S V., Machado MA, Toth I, Salmond G, Foster GD. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol 13:614–629.

41. Piqué N, Miñana-Galbis D, Merino S, Tomás JM. 2015. Virulence factors of Erwinia amylovora: A review. Int J Mol Sci 16:12836–54.

42. Wei Z, Kim JF, Beer S V. 2000. Regulation of hrp genes and type III protein secretion in Erwinia amylovora by HrpX/HrpY, a novel two-component system, and HrpS. Mol Plant-Microbe Interact 13:1251–1262.

43. Geider K. 2000. Exopolysaccharides of Erwinia amylovora: structure, biosynthesis, regulation, role in pathogenicity of amylovoran and levan, p. 117–140. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

44. Koczan JM, Lenneman BR, McGrath MJ, Sundin GW. 2011. Cell surface attachment structures contribute to biofilm formation and xylem colonization by Erwinia amylovora. Appl Environ Microbiol 77:7031–7039.

45. Castiblanco LF, Sundin GW. 2016. Cellulose production, activated by cyclic di-GMP through BcsA and BcsZ, is a virulence factor and an essential determinant of the three- dimensional architectures of biofilms formed by Erwinia amylovora Ea1189. Mol Plant Pathol 19:90–103.

144 Y2 resistance affects phage infectivity

46. Vanneste JL, Eden-Green S. 2000. Migration of Erwinia amylovora in host plant tissues, p. 73–83. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

47. Jones AL, Schnabel EL. 2000. The development of streptomycin-resistant strains of Erwinia amylovora, p. 235–251. In Vanneste, JL (ed.), Fire blight: the disease and its causative agent, Erwinia amylovora. CAB International, Wallingford.

48. Russo NL, Burr TJ, Breth DI, Aldwinckle HS. 2008. Isolation of streptomycin-resistant isolates of Erwinia amylovora in New York. Plant Dis 92:714–718.

49. Svircev AM, Castle AJ, Lehman SM. 2010. Bacteriophages for control of phytopathogens in food production systems, p. 79–96. In Sabour, PM, Griffiths, MW (eds.), Bacteriophages in the control of food- and waterborne pathogens. ASM Press, Washington, DC.

50. Boulé J, Sholberg PL, Lehman SM, O’Gorman DT, Svircev AM. 2011. Isolation and characterization of eight bacteriophages infecting Erwinia amylovora and their potential as biological control agents in British Columbia, Canada. Can J Plant Pathol 33:308– 317.

51. Lehman SM. 2007. Ph.D. thesis: Development of a bacteriophage-based biopesticide for fire blight. Department of Biological Sciences, Brock University, St. Catharines.

52. Born Y, Fieseler L, Marazzi J, Lurz R, Duffy B, Loessner MJ. 2011. Novel virulent and broad-host-range Erwinia amylovora bacteriophages reveal a high degree of mosaicism and a relationship to Enterobacteriaceae phages. Appl Environ Microbiol 77:5945– 5954.

53. Knecht LE, Born Y, Pothier JF, Loessner MJ, Fieseler L. 2018. Complete genome sequences of Erwinia amylovora phages vB_EamP-S2 and vB_EamM-Bue1. Microbiol Resour Announc 7:e00891-18.

54. Adams MH. 1959. Bacteriophages. Interscience Publishers Inc., New York.

55. Sambrook J, Russell DW. 2001. Molecular cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

56. Bellemann P, Bereswill S, Berger S, Geider K. 1994. Visualization of capsule formation by Erwinia amylovora and assays to determine amylovoran synthesis. Int J Biol Macromol 16:290–296.

57. Pusey PL. 1997. Crab apple blossoms as a model for research on biological control of fire blight. Phytopathology 87:1096–102.

58. Llop P, Cabrefiga J, Smits THM, Dreo T, Barbé S, Pulawska J, Bultreys A, Blom J, Duffy B, Montesinos E, López MM. 2011. Erwinia amylovora novel plasmid pEI70: complete sequence, biogeography, and role in aggressiveness in the fire blight phytopathogen. PLoS One 6:e28651.

59. Zengerer V, Schmid M, Bieri M, Müller DC, Remus-Emsermann MNP, Ahrens CH, Pelludat C. 2018. Pseudomonas orientalis F9: A potent antagonist against phytopathogens with phytotoxic effect in the apple flower. Front Microbiol 9:145.

60. Smits THM, Jaenicke S, Rezzonico F, Kamber T, Goesmann A, Frey JE, Duffy B. 2010.

145 Manuscript IV

Complete genome sequence of the fire blight pathogen Erwinia pyrifoliae DSM 12163T and comparative genomic insights into plant pathogenicity. BMC Genomics 11:2.

61. Guzman LLM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J Bacteriol 177:4121–4130.

62. Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison Iii CA, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343.

63. Born Y, Fieseler L, Klumpp J, Eugster MR, Zurfluh K, Duffy B, Loessner MJ. 2014. The tail-associated depolymerase of Erwinia amylovora phage L1 mediates host cell adsorption and enzymatic capsule removal, which can enhance infection by other phage. Environ Microbiol 16:2168–2180.

64. Zhang Y. 2008. I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 9.

65. Stingele F, Newell JW, Neeser JR. 1999. Unraveling the function of glycosyltransferases in Streptococcus thermophilus Sfi6. J Bacteriol 181:6354–60.

66. Cote JM, Taylor EA. 2017. The glycosyltransferases of LPS core: A review of four heptosyltransferase enzymes in context. Int J Mol Sci 18.

67. Muñoz R, Mollerach M, López R, García E. 1997. Molecular organization of the genes required for the synthesis of type 1 capsular polysaccharide of Streptococcus pneumoniae: formation of binary encapsulated pneumococci and identification of cryptic dTDP-rhamnose biosynthesis genes. Mol Microbiol 25:79–92.

68. Marchler-Bauer A, Bo Y, Han L, He J, Lanczycki CJ, Lu S, Chitsaz F, Derbyshire MK, Geer RC, Gonzales NR, Gwadz M, Hurwitz DI, Lu F, Marchler GH, Song JS, Thanki N, Wang Z, Yamashita RA, Zhang D, Zheng C, Geer LY, Bryant SH. 2017. CDD/SPARCLE: functional classification of proteins via subfamily domain architectures. Nucleic Acids Res 45:D200–D203.

69. Smits THM, Rezzonico F, Kamber T, Blom J, Goesmann A, Frey JE, Duffy B. 2010. Complete genome sequence of the fire blight pathogen Erwinia amylovora CFBP 1430 and comparison to other Erwinia spp. Mol Plant-Microbe Interact 23:384–393.

70. Bugert P, Geider K. 1995. Molecular analysis of the ams operon required for exopolysaccharide synthesis of Erwinia amylovora. Mol Microbiol 15:917–933.

71. Ayers AR, Ayers SB, Goodman RN. 1979. Extracellular polysaccharide of Erwinia amylovora: a correlation with virulence. Appl Environ Microbiol 38:659–666.

72. Babu M, Bundalovic-Torma C, Calmettes C, Phanse S, Zhang Q, Jiang Y, Minic Z, Kim S, Mehla J, Gagarinova A, Rodionova I, Kumar A, Guo H, Kagan O, Pogoutse O, Aoki H, Deineko V, Caufield JH, Holtzapple E, Zhang Z, Vastermark A, Pandya Y, Lai CC, El Bakkouri M, Hooda Y, Shah M, Burnside D, Hooshyar M, Vlasblom J, Rajagopala S V, Golshani A, Wuchty S, F Greenblatt J, Saier M, Uetz P, F Moraes T, Parkinson J, Emili A. 2017. Global landscape of cell envelope protein complexes in Escherichia coli. Nat Biotechnol 36:103–112.

73. Lairson LL, Henrissat B, Davies GJ, Withers SG. 2008. Glycosyltransferases:

146 Y2 resistance affects phage infectivity structures, functions and mechanisms. Annu Rev Biochem 77:521–555.

147

Conclusions and Outlook

3. Conclusions and Outlook

The bacterium E. amylovora is the causative agent of fire blight, a devastating plant disease affecting pome fruit production. Since the ban of streptomycin as potent treatment, alternative options to control the pathogen are scarce. Bacteriophages could have the potential to fill this demand. These natural enemies of bacteria are highly specific in recognizing and eliminating their particular host bacteria. Combinations of several phages that target different receptors on the host surface and apply various strategies to evade the host immune system can further render the treatment more potent while simultaneously reduce the risk of resistance development towards the applied phages. To identify which receptor is targeted by a particular phage, a transposon mutagenesis library was generated to identify mutants able to resist phage infection. The screen was fine-tuned for six well characterized Erwinia amylovora specific phages with the potential to be used as phage biocontrol agents. Although the six phages Bue1, L1, M7, S2, S6 and Y2 expose diverging host ranges and belong to different genera, the screen classified them into three host receptor groups. The majority of transposon disrupted genes identified in Bue1 and Y2 resistant mutants can be associated with LPS biosynthesis suggesting that Bue1 and Y2 require certain LPS structure for host identification and infection. The phages L1 and S2 are grouped into a second receptor group. Both phages require an intact ams operon for host infection. This operon encodes the amylovoran synthesis apparatus. The produced molecule amylovoran is an EPS component, which is secreted by E. amylovora and functions as one of its major virulence factors. The results generated by the screen are in accordance with the fact that neither of the two phages L1 and S2 is able to infect low EPS producing strains and both phages were observed to encode depolymerase that are able to degrade amylovoran. Since the screen was unable to reveal an alternative surface structure, which could be used as receptor, it cannot be excluded that amylovoran is the sole receptor required for L1 or S2 host recognition and infection. Finally, the phages M7 and S6 rely on an intact bcs operon for infection. The operon encodes the bacterial cellulose synthase complex, required for cellulose production and stable biofilm formation. Indeed neither M7 nor S6 were able to infect mutants without the operon or key genes in the operon. Especially mutants, which are unable to produce bacterial cellulose, were insensitive to M7 or S6 exposure. Phage infection could also be abolished, when bacteria were incubated with the cellulose binding dye Congo Red. Both results indicate that bacterial cellulose plays an essential role in M7 and S6 infection. It is however, unclear if the phages solemnly rely on bacterial cellulose for identification and infection of host bacteria. Since bacterial cellulose can be produced by various bacteria it is unsure how the phages would be able to differentiate between host and non-host cells. Modification of bacterial cellulose by certain bacteria has been observed previously but it has yet to be determined if this is also the case for

149 Conclusions and Outlook

E.amylovora. Alternatively, M7 and S6 could require a secondary receptor to complete host infection. Similarly, to L1 and S2 the screen could not identify another surface structure that could be recognized by M7 and S6 leaving the question of an alternative receptor unanswered. Finally, genome analysis of M7 and S6 revealed a collection of putative cellulases and endoglucanase encoded by the two phages. Experiments with cellulose and entire phages, demonstrated cellulolytic functionality in the case of S6. The fact that phages encode such enzymes and specifically target bacterial cellulose is a novelty to our knowledge. Enzyme purification and functionality experiment should be performed as a next step to further characterize these phage encoded cellulases. Since bacterial cellulose is a major component of biofilms and even function as virulence factors in certain bacteria, these enzymes could be specifically applied to target biofilm-forming bacteria. In addition to revealing potential phage receptors, the screen was also used to identify mutants that were insensitive to several phages. Mutants with transposon insertions into the genes topB1 (NAD-dependent epimerase), rfaE (ADP-heptose synthase) or pgm (phosphoglucomutase) were observed to be multiple phage resistant. They were not only resistant towards phages belonging to the same receptor group but were also unaffected by exposure towards phages belonging to different receptor groups. Analysis of clean knock-out mutants revealed modification in LPS structure and reduced EPS and bacterial cellulose production. These alterations suggest that the encoded proteins are involved in manifold metabolic pathways. The impact these gene deletions have on LPS, EPS and bacterial cellulose simultaneously could explain how these mutants resist phages targeting different surface structures. By examining spontaneously resistant mutants, a permanent Y2 resistant mutant termed 1430Y2R was identified. It is yet unsure why spontaneous resistance was detected to be permanent only towards Y2. The other five phages Bue1, L1, S2, S6 and M7 were also able to trigger spontaneous resistance in E. amylovora, which revolved however to a phage sensitive state after the phages were removed. Further investigation to identify the underlying mechanism involved in generating either permanent or transient resistance will be performed as a next step. The mutant 1430Y2R not only survived Y2 exposure but also demonstrated cross-resistance against other phages. Bue1 was unable to infect 1430Y2R and the infectivity of the phages L1, S2 and S6 towards 1430Y2R was reduced. Only M7 infectivity was unaffected by the mutation. LPS structure, EPS and bacterial cellulose levels were measured for 1430Y2R to identify reasons for altered phage infectivity. Indeed 1430Y2R revealed strong alterations in LPS structures compared to the wildtype explaining why the two LPS dependent phages Bue1 and Y2 are unable to infect. The amount of secreted EPS was diminished, explaining the reduced efficiency of plating of L1 and S2. The effect 1430Y2R has on S6 infection remains unclear.

150 Conclusions and Outlook

Whole genome sequencing identified a nucleotide deletion as main source for the permanent resistance. The affected gene with the locus tag EAMY_2231 encodes a putative glycosyltransferase with two distinct enzymatically active domains whereby the deletion is affecting the C-terminal domain. Deletion of this gene has so far been unsuccessful. However, by reintroducing the complete gene in 1430Y2R, full susceptibility towards all tested phages could be restored. To conclude these findings, manipulation of the genes topB1, rfaE, pgm or EAMY_2231 can mediate multiple phage resistance targeting various receptors on the host surface. The encoded proteins were observed to affect the metabolic process of several extracellular structures such as LPS, EPS and bacterial cellulose simultaneously. The risk these genes and their gene products pose to phage cocktails was evaluated by investigating potential fitness costs. The most striking fitness loss was observed for the deletions of topB1, rfaE and pgm where the deletion mutants were unable to grow in the blossoms and establish disease symptoms. We assume that the substantial reduction of amylovoran produced by these mutants accounts for the inability to successfully colonize the blossoms. If such a strategy would be applied by the pathogen to evade the exposure of a phage cocktail, the modification of one of these genes would render the bacterium avirulent, thereby the phage cocktail protects the plant from fire blight outbreak. Interfering with such crucial metabolic processes required for survival in blossoms or virulence could be an alternative strategy to control fire blight outbreaks. By applying molecules that specifically target these processes the pathogen would be rendered avirulent and the disease stopped. However, it is yet to be determined how such molecules could be identified and isolated. Unlike topB1, rfaE and pgm, the modification of EAMY_2231 have no impact on virulence in blossoms. Since this mutation appeared spontaneously after Y2 exposure in vitro, it is crucial to verify if this mutation can also occur in blossoms. If such permanent phage resistant mutants can be isolated from blossoms after treatment with Y2, the phage should be excluded from the phage cocktail. This measure ensures a potent phage cocktail and simultaneously maintains a low risk of resistance development towards multiple phages. In a further step, the phages were tested for their biocontrol abilities against E. amylovora. To compare the efficacy of single phages and phage cocktails, both settings were tested in vitro and on blossoms. Furthermore, different combinations of the six phages Bue1, L1, M7, S2, S6 and Y2 were monitored. Combinations of phages belonging to different phage receptor groups were generally more potent in reducing bacterial growth in vitro than combinations from within a specific phage receptor group. The combinations were demonstrated to be efficient in controlling bacterial cell counts during the entire experiment. In blossoms, the treatments with L1 alone, M7+S6 or Bue1+S2+S6 generated the strongest reduction of disease symptoms. Triple combinations of phages belonging to different receptor groups as

151 Conclusions and Outlook well as the most potent treatments e.g. L1 and M7+S6, should be validated in blossoms and further tested on apple trees to confirm these results. In addition, resistance development of the bacteria towards the applied phages should be monitored and phage cocktails should be adjusted accordingly. Previous studies, which evaluated the potential of E. amylovora specific phages, concluded that the success of a phage cocktail is highly dependent on maintaining a stable phage titre over a prolonged period and an even distribution of the treatment on the plant surface. Different measures should be taken accordingly to achieve this. Spraying the phage treatment in the evening or during the night could reduce damaging UV exposure. Supplementation of the phage cocktail with certain adjuvants could further enhance the efficacy and stability of the treatment. These additives should protect the phages from UV irradiation and reduce the surface tension to ensure even distribution and protection on the blossoms. Substances that lower the pH could further be considered as additives since growth of the pathogen is reduced at low pH. In addition to these auxiliary substances, the phage treatments could be fortified by supplementing certain proteins and enzymes. Phage encoded depolymerases (L1, S2, and Bue1) or cellulases (M7, S6) can reduce the amount of amylovoran or bacterial cellulose produced by E. amylovora as protective layers and thereby further weaken the pathogen. Bactericides could further impede fire blight outbreak by directly targeting bacteria. Finally, supplementing the phage cocktail with antagonistic bacteria such as Pantoea agglomerans, certain Pseudomonas or antagonistic yeasts as Aureobasidium pullulans the antagonistic organisms in Blossom ProtectTM, could be promising. Such antagonistic bacteria could act against the pathogen either through direct interaction and elimination of the pathogen or through occupation of niches required by the pathogen to colonize the blossoms before entering into the plant tissue. Especially antagonistic bacteria, which can be used by the phages as alternative hosts, should be investigated. Adding these bacteria could ensure continuous replication of phages and therefore ensures maintenance of high titres of highly infective phages on the blossoms while simultaneously protect the blossoms through niche occupation. All these measures could enhance the potency, distribution, stability and sustainability of the phage biocontrol agent and ensure a low risk of phage resistance development. The results, information and conclusions gathered from this work will help advance the establishment of phages as successful biocontrol agents against the devastating plant disease fire blight. Minimizing the risk of phage resistance development was identified as key issue to introduce phage biocontrol as potent and long-lasting treatment option. Therefore, certain measures should be taken. The risk of multiple phage resistance has to be evaluated and adaptions to the phage cocktail must be carried out accordingly. Phages that target different host receptors on the bacterial surface and are able to evade the bacterial immune system

152 Conclusions and Outlook should be incorporated into phage cocktails. Finally, to enhance the efficacy of the phage cocktail, stabilizing adjuvants or auxiliary enzymes should be tested. This work is a valuable contribution to the fight against fire blight with phages.

153

Acknowledgements

4. Acknowledgements

The work presented here was generated over a prolonged period of time and is the result of paper reading, lab work, field work, team work, data analysis, student tutoring, scientific discussions, journal clubs and finally writing. On this journey many people contributed with knowledge, pipetting skills, blossom picking abilities, entertainments, guacamole and so much more. This work could not have been realised without them. I would therefore like to thank them all warmly at this point. Foremost I would like to thank Prof. Martin J. Loessner for giving me the opportunity to carry out this exciting research project. As doctor father, he provide me with scientific guidance and constructive inputs and feedback throughout the entire project. To Prof. Lars Fieseler who had the brilliant idea to initiate this project in the first place and therefore made this work possible. I would like to thank him for sharing his phage passion, his creative ideas, his scientific knowledge and his support in both research problems and everyday tasks. For participating in my scientific committee, I would like to thank Prof. Julia Vorholt. She contributed her knowledge and experience to advance this project. The project would only have been half as much fruitful if Dr. Yannick Born would not have been working in the lab. He introduced me to phage research and supported me with scientific knowledge, long lasting experience and insight into this project and moral support during this time. I am very grateful towards the members of the food microbiology and food biotechnology groups for their company and all the sophisticated discussions and entertaining activities I experienced with them. I would therefore like to thank Marjan Veljkovic, Nadine Heinrich, Giovanna Spielmann-Prada, Felix Stöppelmann, Lukas Reinau, Gabi Fahrer-Gurt, Tania Torossi, Rawaa Al Toma Sho, Dr. Titu Staubli, Prof. Susanne Miescher Schwenninger, Andrea Tönz, Denise Müller, Sandra Mischler, Susette Freimüller Leischtfeld and Dr. Edwina Romanens. They were continuously able to generate a productive and constructive but also highly enjoyable atmosphere during this project. I am also very thankful towards Stefanie Balada and Dr. Sabina Gerber for their knowledge and the help they provided in the lab. Jonas Hofmänner, David Vinzent, Katja Felder, Marina Mahler, Roman Müller, Jules Peter and Nadine Heinrich should be thanked for strongly contributing to this work in the context of their semester, bachelor or master thesis. Melinda, Mary and Goliath should be mentioned for their loyal and mostly effortless work during this entire project. The blossom experiments were carried out with the help of Dr. Cosima Pelludat, her co workers and Edi Holliger at Agroscope in Wädenswil. I would like to thank Cosima for ordering

155 Acknowledgements and managing the apple trees during the blossoming period, for constructive inputs, vast scientific knowledge, contribution to the journal club and the enjoyable Thai lunches. The Microbiology and Immunology PhD (MIM) program with its many students should be thanked for providing a large scientific community of other PhD students who passionately work in science and experience a similar journey. Judith Zingg in particular should be mentioned for her help and support in any administrative or personal issues and the excellent work she does for the MIM program. I am most grateful towards my family and friend for their incredible support during this intense time. All dinners, discussions and after work drinks were highly appreciated. I would like to thank my younger brother Niklaus for lending his vast graphical talent, Matthias and Fionna for my nieces who show me how enjoyable life outside the lab can be. Daniel and especially to my mother Eva for supporting my scientific interest. Last but not least Oliver, whom I thank with all my heart for sticking it out with me.

156 Curriculum Vitae

5. Curriculum Vitae

157