The role of small open reading frames in Shewanella oneidensis phage λSo in host takeover and phage proliferation Dissertation Zur Erlangung des Doktorgrades der Naturwissenschaften (Dr. rer. Nat.) Dem Fachbereich 08 - Biologie und Chemie - der Justus-Liebig-Universität Gießen vorgelegt von Svenja Wiebke Thöneböhn Angefertigt im Institut für Mikro- und Molekularbiologie Gießen, August 2025 II Erstgutachter: Prof. Dr. Kai Thormann Institut für Mikrobiologie und Molekularbiologie Justus-Liebig-Universität Gießen Zweitgutachter: Prof. Dr. Julia Frunzke Institute of Bio- and Geoscience Forschungszentrum Jülich I Die während der Promotion erzielten Ergebnisse sind zum Teil in folgender Publikation veröffentlicht: Svenja Thöneböhn, Dorian Fischer, Vanessa Kreiling, Alina Kemmler, Isabella Ober- heim, Fabian Hager, Nicole E Schmid, Kai M Thormann: Identifying components of the Shewanella phage LambdaSo lysis system. Journal of Bacteriology, 21.04.2024 DOI: 10.1128/jb.00022-24 Eidesstattliche Erklärung II Eidesstattliche Erklärung Hiermit versichere ich, die vorgelegte Dissertation selbständig und ohne unerlaubte fremde Hilfe und nur mit den Hilfen angefertigt, die ich in der Dissertation angegeben habe. Alle Textstellen, die wörtlich oder sinngemäß aus veröffentlichten Schriften entnommen sind, und alle Angaben, die auf mündlichen Auskünften beruhen, sind als sol- che kenntlich gemacht. Bei den von mir durchgeführten und in der Dissertation erwähnten Untersuchungen habe ich die Grundsätze guter wissenschaftlicher Praxis, wie sie in der „Satzung der Justus-Liebig-Universität Gießen zur Sicherung guter wissenschaftlicher Praxis“ niedergelegt sind, eingehalten. Gemäß § 25 Abs. 6 der Allgemeinen Bestimmung für modularisierte Studiengänge dulde ich eine Überprüfung der Thesis mittels Anti-Plagiatssoftware. Datum: Unterschrift: Eidesstattliche Erklärung III Erklärung zur Verwendung von künstlicher Intelligenz bei der Erstellung von Aufsätzen und Abschlussarbeiten Name des Studierenden Familienname: Vorname: Titel der Dissertation The role of small open reading frames in Shewanella oneidensis phage λSo in host takeover and phage proliferation Bitte markieren: ☐ Ich habe bei der Erstellung dieses Textes kein KI-Tool verwendet. ☐ Ich habe ein KI-Tool in den folgenden Bereichen eingesetzt (Mehrfachnennungen möglich): ☐ Ideen finden, meine Kreativität anregen ☐ Verstehen von Konzepten, Recherche von Fakten und Definitionen ☐ Optimierung eines von mir verfassten Textes ☐ Erstellen ganzer Textpassagen nach meinen Vorgaben Wenn Sie ein KI-Tool verwendet haben, erklären Sie bitte, welche Teile Ihres Textes von dem Tool profitiert haben und wie. Fügen Sie bei Bedarf eine zusätzliche Seitehinzu oder geben Sie Hyperlinks zu den von Ihnen verwendeten Chat-Verläufen an. Verwendet wurden DeepL, DeepL-Write sowie ChatGPT zur Übersetzung, Suche von Synonymen für gewisse Worte sowie zur Verbesserung des Vokabulars und des Wortlautes des in Englisch verfassten Textes. Datum: Unterschrift: Thöneböhn Svenja Wiebke Abstract IV Abstract Bacteriophages are the most abundant biological entities on Earth. They wield an immense influence on microbial ecosystems in almost all habitats by regulating bacterial population dynamics. Most phages follow one of two well-characterised strategies for host exploitation: the lytic or the lysogenic cycle. In both pathways, host cell lysis represents the terminal event and is therefore central to phage fitness. The temperate phage λSo is one of four known prophages in the genome of Shewanella oneidensis MR-1 and has a genome size of about 51 kbp. During lysogeny, λSo remains integrated into the host chromosome, replicating in concert with the host cell. In this study, the lysis system of λSo was characterised as a pinholin-SAR endolysin-two- component spanin pathway. The λSo holin protein, SSo, contains two transmembrane do- mains and also produces an antagonistic isoform through an alternative translation start, named antiholin. This regulatory mechanism enables precise temporal control over the in- itiation of host lysis. In addition to the pinholin and the SAR endolysin, the lysis system requires a two-component spanin complex, made up of an inner membrane protein (i- Spanin, RzSo) and an outer membrane protein (o-Spanin, Rz1So). The corresponding genes are present in an overlapping reading frame structure, and the encoded proteins likely form a functional dimer of two dimers. This putative dimer enables the fusion of the inner and outer membrane. In addition, this work has shown that further, previously uncharacterised gene products are involved in cell lysis. Like many phages, λSo harbours genes encoding small proteins of unknown function. A gene cluster, so called cluster C, was identified, whose deletion significantly reduced the number of plaque-forming units. Cluster C consists of six genes (lcc1 - lcc6) encoding proteins between 41 and 137 amino acids in length that have no obvious homologies to known protein domains. Bioinformatic analysis suggests that Lcc4 and Lcc6 contain putative transmembrane domains. Functional characterisation revealed that Lcc6 plays a critical role in phage-induced host cell lysis. In lcc6 deletion strains, induction of the lytic cycle of λSo using mitomycin C resulted in the formation of phage particles, which, however, failed to lyse the host cells and are therefore not released. These findings Abstract V suggest that Lcc6 participates in an early phase of the lysis cascade, likely acting in concert with pinholin-mediated membrane disruption. The ectopic expression of the Lcc4 protein on the other hand resulted in a pronounced elongation of the host cells and delocalisation of the FtsZ rings - a phenotype that is compatible with a disruption of cell division. The modelling of plausible protein interactions confirmed that this phenotype results from a direct interaction of Lcc4 with key components of the bacterial divisome, particularly FtsZ and ZipA. Site-directed mutagenesis identified isoleucine residues at positions 16 and 19 as essential for the interaction with FtsZ, and tryptophan 80 and arginine 84 as critical for binding to ZipA. Taken together, these results suggest that Lcc4 specifically inhibits bacterial cytokinesis following prophage induction in order to maximise the availability of the metabolic resources of the host cell during phage replication. The Lcc proteins, encoded by genes of the cluster C, thus represents a previously undescribed phage-host effector system with profound influence on cellular organisation and the course of lysis. Zusammenfassung VI Zusammenfassung Bakteriophagen stellen die am weitesten verbreiteten biologischen Entitäten auf der Erde dar. Sie üben einen immensen Einfluss auf die mikrobiellen Ökosysteme nahezu aller Le- bensräume aus, insbesondere durch die Regulation bakterieller Populationen. Die meisten Phagen verfolgen dabei eine von zwei gut charakterisierten Strategien zur Wirtsausbeu- tung: den lytischen oder den lysogenen Zyklus. Die Lyse der Wirtszelle bildet dabei das finale und für die Phagenfitness zentrale Ereignis beider Strategien. Der temperente Phage λSo ist einer von vier bekannten Prophagen im Genom von Shewanella oneidensis MR-1 und weist eine Genomgröße von etwa 51 kbp auf. Im lysogenen Lebenszyklus ist λSo in das Wirtsgenom integriert und repliziert synchron mit der Wirtszelle. Im Rahmen dieser Arbeit konnte das Lysesystem von λSo als ein Pinholin-SAR-Endolysin-Zwei-Komponenten-Spanin-System charakterisiert werden. Das Holin-Protein SSo besitzt zwei Transmembrandomänen und generiert durch einen alternativen Translationsstart einen antagonistischen Isoform, welche als Antiholin fungiert und eine präzise zeitliche Regulation der Zelllyse ermöglicht. Ergänzend zum Pin- holin und dem SAR-Endolysin ist ein Zwei-Komponenten-Spanin-System erforderlich, bestehend aus einem inneren Membranprotein (i-Spanin, RzSo) und einem äu- ßeren Membranprotein (o-Spanin, Rz1So). Die entsprechenden Gene liegen in einer über- lappenden Leserasterstruktur vor und bilden vermutlich funktional einen Dimer aus zwei Dimeren, der die Fusion der inneren und äußeren Membran ermöglicht. Darüber hinaus konnte durch diese Arbeit gezeigt werden, dass weitere, bislang uncharakterisierte Genprodukte an der Zelllyse beteiligt sind. Viele Phagen enthalten zahlreiche Gene, die die für Proteine unterschiedlicher Größe kodieren, deren Funktion derzeit nicht klar ist. Das Genom von λSo enthält unter anderem ein Gencluster, Cluster C, dessen Deletion zu einer drastischen Abnahme der Pro- duktion von Plaque-bildenden Einheiten führt. Das Cluster besteht aus sechs Genen (lcc1 - lcc6), die für eher kleine Proteine im Bereich von 41 bis 137 Aminosäuren kodieren und keine offensichtliche Homologien zu bekannten Proteindomänen aufweisen. Zwei der Proteine, Lcc4 und Lcc6, besitzen vermutlich eine Transmembrandomäne. Das Protein Lcc6 konnte als zentraler Faktor in der Phagen- induzierten Zelllyse identifiziert werden. In lcc6-Deletionstämmen führte die Induktion des Zusammenfassung VII lytischen Zyklus mittels Mitomycin C zwar zur Bildung reifer Phagenpartikel, diese konn- ten jedoch die Wirtszellen nicht lysieren und die Zelle final verlassen. Dies legt nahe, dass Lcc6 in einen frühen, wahrscheinlich Pinholin-vermittelten Schritt der Zelllyse eingreift. Die ektopische Expression des Proteins Lcc4 hatte hingegen eine ausgeprägte Verlängerung der Wirtszellen sowie eine Delokalisierung der FtsZ-Ringe zur Folge: ein Phänotyp, der mit einer Störung der Zellteilung vereinbar ist. Durch Struktursimulationen konnte gezeigt werden, dass dieser Phänotyp auf eine Interaktion von Lcc4 mit Proteinen der Zellteilungsmaschinerie wie FtsZ und ZipA zurück zu führen ist. Funktionelle Mutagenese-Experimente identifizierten Isoleucin an Position 16 und 19 als essenziell für die Interaktion mit FtsZ sowie Tryptophan 80 und Arginin 84 für die Bindung an ZipA. Diese Ergebnisse deuten darauf hin, dass Lcc4 die Zellteilung gezielt hemmt, um die metabolischen Ressourcen der Wirtszelle während der Phagenreplikation vollständig verfügbar zu machen. Das hier identifizierte Lcc-System stellt damit ein neuartiges Phagen-Wirt-Effektorsystem dar, das tiefgreifenden Einfluss auf die zelluläre Organisation und den Verlauf der phagen-vermittelten Zelllyse besitzt. Table of contents VIII Table of contents Eidesstattliche Erklärung .................................................................................................. II Abstract ........................................................................................................................... IV Zusammenfassung........................................................................................................... VI Table of contents .......................................................................................................... VIII 1 Introduction ........................................................................................................... 11 1.1 General characteristics of bacteriophages ............................................... 11 1.1.2 Caudoviricetes ...................................................................................... 12 1.2 Life cycles ............................................................................................... 14 1.3 Phage mediated lysis ............................................................................... 17 1.3.1 Canonical holin lysis pathway .............................................................. 18 1.3.2 Pinholin SAR endolysin pathway ......................................................... 20 1.3.3 Spanin proteins...................................................................................... 22 1.4 Phage protein interactions with host cellular machinery ........................ 25 1.5 Bacterial cell division .............................................................................. 26 1.6 Model organism Shewanella oneidensis MR-1 ....................................... 30 1.6.1 Phages of S. oneidensis MR-1 .................................................................... 30 1.7 Aim of this study ............................................................................................. 31 2 Results ................................................................................................................... 32 2.1 Analysis of the Shewanella prophage λSo .............................................. 32 2.1.1 Characterization of the λSo lysis cluster ............................................... 32 2.1.1.1 Further characterization of the holin protein ...................................... 39 2.1.1.2 Further characterization of the spanin system .................................... 41 2.1.1.3 Identification of a novel lysis component .......................................... 45 2.1.1.4 Additional proteins needed for λSo Pinholin-SAR-Endolysin lysis .. 49 2.1.2 Characterization of novel components in host takeover by λSo ........... 50 2.1.2.1 Identification of possible interaction partner of Lcc4 ........................ 55 2.1.2.2 Characterization of the interaction between Lcc4 and FtsZ ............... 63 3 Discussion.............................................................................................................. 70 3.1 Analysis of the Shewanella phage λSo ................................................... 70 3.1.1 λSo uses a pinholin-SAR-endolysin lysis pathway .................................... 71 3.1.2 The λSo lysis system requires a two component spanin system ................ 73 3.1.3 The λSo holin encodes its own inhibitor .................................................... 76 3.1.4 λSo lysis requires at least two more proteins for sufficient lysis ............... 79 3.1.5 Conclusion ................................................................................................. 80 3.2 Identification of novel host effector proteins in λSo ............................... 82 Table of contents IX 3.2.1 λSo protein Lcc4 inhibits the cell division machinery ............................... 83 3.2.2 Conclusion ................................................................................................. 88 3.3 Outlook .................................................................................................... 91 4 Materials & Methods ............................................................................................. 93 4.1 Materials .......................................................................................................... 93 4.2 Microbiological methods ............................................................................... 111 4.2.1 Cultivation of bacterial strains ................................................................. 111 4.2.2 Conjugation of S. oneidensis MR-1 cells ................................................. 111 4.2.3 Electroporation ......................................................................................... 112 4.2.4 Induction and cultivation of phages ......................................................... 112 4.2.5 Determination of phage lysis profiles ...................................................... 112 4.2.6 Microscopy ............................................................................................... 113 4.2.7 Determination of cell length..................................................................... 113 4.2.8 One-step growth experiment .................................................................... 114 4.2.9 Bacterial Two Hybrid Assay .................................................................... 114 4.2.10 Measurement of planktonic growth ....................................................... 115 4.2.11 Membrane depolarization assay ............................................................. 115 4.3 Molecular biological methods ....................................................................... 115 4.3.1 Polymerase chain reaction........................................................................ 115 4.3.2 Agarose gel electrophoresis ..................................................................... 116 4.3.3 In vitro digestion of DNA ........................................................................ 117 4.3.4 Plasmid and strain constructions .............................................................. 117 4.3.5 Gibson Assembly ..................................................................................... 117 4.4 Biochemical methods .................................................................................... 118 4.4.1 SDS- Page and Western blotting .............................................................. 118 4.4.2 Protein isolation ....................................................................................... 118 4.4.3 Sedimentation assay ................................................................................. 120 4.4.4 Bioinformatic approaches ........................................................................ 121 Appendix ....................................................................................................................... 122 A Figures ................................................................................................... 122 B Tables .................................................................................................... 133 References ..................................................................................................................... 136 Abbreviations ................................................................................................................ 164 Acknowledgments......................................................................................................... 168 X 1 Introduction 1.1 General characteristics of bacteriophages Bacteriophages, often called phages, are thought to be the most abundant biological entities on earth and are found in every explored biome 1–3. They outnumber bacteria by up to tenfold, as viral concentrations can reach up to 2.5 × 10⁸ viruses per millilitre in natural aquatic environments 4. Phages are obligate intracellular parasites of bacteria with diverse life cycles that play a crucial role in shaping bacterial evolution and community structure and thereby influencing ecosystem dynamics. Their evolutionary significance is emphasised by the enormous extent of phage-mediated horizontal gene transfer, which is estimated to be about 2 × 10¹⁶ events every second worldwide 5,6. As antibiotic-resistant bacteria continue to present a serious threat to global health, interest in bacteriophage-based strategies has grown significantly in recent years. Phages are estimated to account for up to 40% of bacterial mortality each day, underscoring their ecological importance and therapeutic potential. Over the past two decades, this has spurred a resurgence in research aimed at harnessing phages for use in medicine, agriculture, and food safety 7,8. Phage therapy, employing either naturally occurring or genetically engineered lytic phages, has yielded encouraging results in clinical contexts 7. In contrast to the slower regulatory uptake in healthcare, the food and agricultural sectors have adopted phage-based technologies more swiftly, with several products already available on the market 8. Moreover, due to their high host specificity, phages are increasingly being investigated as tools for the sensitive and accurate detection of bacterial pathogens. Bacteriophages display extraordinary diversity and are currently classified based on a combination of genomic features and morphological characteristics 9–11. With the rapid increase in publicly available viral genome sequences, genome-based taxonomic approaches have become widespread and are now a central method in phage classification 10. The genetic material of phages can consist of either double-stranded (ds) or single-stranded (ss) DNA or RNA and can include modified nucleotides as protection against restriction enzymes (Fig. 1) 12–14. Genome sizes are highly variable, ranging from 1 Introduction 12 approximately 3.5 kilobases in the ssRNA phage MS2 to around 500 kilobases in the dsDNA phage G 15. The genetic material is enclosed within a capsid, which can take various forms: polyhedral (Microviridae, Corticoviridae, Tectiviridae, Leviviridae, and Cystoviri- dae), filamentous (Inoviridae), pleomorphic (Plasmaviridae), attached to a tail (Caudoviri- cetes) or possessing lipid or lipoprotein envelopes 1,16,17. Figure 1: Viruses of bacteria. Bacterial virus sub-families are represented and grouped based on their Bal- timore classification. Relative sizes and symmetries are approximate. Modified after: Hay & Lithgow 17 1.1.2 Caudoviricetes As a model belonging to the Caudoviricetes class was utilised in this study, the following section will provide a more detailed description of the characteristics of this class. The class of the Caudoviricetes encompasses tailed bacteriophages with double-stranded DNA (dsDNA) genomes and represents the majority of phages characterised to date. This class can be divided into three morphological groups: myoviruses (phage T4), siphoviruses (phage λ) and podoviruses (phage T7) 1,9,15. These groups differ notably in their tail morphology: myoviruses possess long, rigid, contractile tails; siphoviruses are characterised by long, flexible, non-contractile tails; and podoviruses feature short, non- contractile tails (Fig. 1) 18,19. Tail structures play a crucial role in recognizing specific receptors, penetrating cell membranes, and delivering the viral genome into the bacterial cytoplasm 16. The tails of siphoviruses and myoviruses are composed of three key elements: 1 Introduction 13 the tail tip complex, which mediates host recognition and initiates the infection process; the tail tube, which serves as a conduit for the transfer of genomic DNA into the host cell; and the terminator proteins, which complete the tail assembly and form the interface for attachment to the phage head 20,21. Capsid size varies significantly among members of the Caudoviricetes with diameters ranging from 45 to 185 nm, typically correlating with genome size 22. The majority of phages in this family, around 75 %, possess icosahedral (isometric) capsids, while approximately 15% (e.g., T4) feature prolate heads. Prolate heads are icosahedral structures elongated along the five-fold axis, which aligns with the phage tail 23. The tail is connected to the head via a so-called connector protein, which is described as a dodecameric portal protein that binds to a special pentameric vertex of the phage capsid (Fig. 2). This protein forms a channel essential for genome packaging during virion assembly and for genome release during infection. Phage head assembly typically begins at this vertex, where the portal protein facilitates the organ- ization of scaffolding and major capsid proteins 16,24. Figure 2: Phage structure. Schematic representation of a siphovirus phage tail, neck and head 24. Capsid assembly is facilitated by internal scaffolding proteins and completed through the closure of the portal gate by head completion proteins (Fig. 2) 25. These proteins typically exist as monomers in solution and do not appear to interact with other phage 1 Introduction 14 proteins. However, they serve as a hub for tail assembly in podoviruses and as a platform for the attachment of the preassembled tail in siphoviruses and myoviruses 24. Phages belonging to those families generally possess two head completion proteins, each forming a ring beneath the portal. In the mature virion, these two rings are located between the portal protein and the tail. The protein shells of the mature capsids are remarkably stable and can withstand the high internal pressure exerted by the tightly packed, encapsulated DNA 26. 1.2 Life cycles Before the required proteins can be expressed and virions formed, the phage must first infect a host cell. This process begins with the phage identifying and attaching to specific receptors on the surface of the host cell, such as surface proteins, lipopolysaccharides, or other molecules on the bacterial cell envelope like teichoic acids, fimbriae, and flagella 21,27,28. As soon as the phage successfully recognizes its target receptor, it permanently adsorbs to the host cell and injects its genetic material 16. The subsequent replication strategy, including the formation, release and transmission of virions, depends on whether the phage is virulent or temperate 29. There are four common phage life cycles: lytic, lysogenic, pseudolysogenic and chronic infection life cycle 30,31. In the lytic life cycle, which is utilised by virulent phages such as T4, the phage employs various strategies to hijack the host's metabolism and initiate the production of viral progeny 32,33. During this process, virion particles are assembled, packed with the respective viral nucleic acid, and ultimately released into the extracellular envi- ronment following the lysis of the infected host cell (Fig. 3) 16,34. This process requires the disruption of the bacterial membrane and cell wall, which relies on specific lysis proteins, including endolysins, holins, and spanins 35,36. In contrast, the lysogenic life cycle allows the phage not only to either lyse its host but also to establish a stable association, known as lysogeny (Fig. 3) 31. During lysogeny, the viral genome, known as a prophage, replicates in synchrony with the host DNA, either as a free, plasmid-like entity (e.g., phage P1) or integrated into the bacterial chromosome (e.g. phage λ) 34. A temperate phage has the capacity to remain dormant as a prophage, replicating alongside the host genome. In response to specific environmental cues, which often involve host cell stress, prophages have been shown to be triggered to exit lysogeny and transition 1 Introduction 15 to the lytic cycle. This results in the production of new virions that are subsequently released from the bacterium 37. The decision between lysogenic and lytic cycles in temperate phages is influenced by various factors, including carbon and nitrogen availabilty, salinity, UV radiation, temperature or pollutants 38–40. Several studies suggest that host density-dependent quorum sensing (QS) plays a crucial role in molecular communication, both between phages and among phages and hosts, potentially guiding the lysis-lysogeny decision 41,42. A long-term association between host and phage can be maintained over thousands of generations, potentially altering the phenotype of the host bacterium through the expression of genes that are not normally activated during infection: a phenomenon known as lysogenic conversion. A well-known example is the prophage CTXφ, which infects certain strains of Vibrio cholerae and encodes toxins responsible for cholera symptoms 43. Bacteria harbouring prophages may also develop immunity to subsequent infections by homologous phages and potentially exhibit enhanced fitness 44. Prophages that remain in the host genome but can no longer be induced are referred to as cryptic prophages 45–47. These can be divided into four different types based on the extent of genomic degradation resulting from the integration of viral and bacterial genetic material: defective prophages, satellite phages, bacteriocins, and other prophage-related entities, such as gene transfer agents. A prophage incapable of undergoing the lytic cycle due to loss of function through mutations or the deletion of genes essential for DNA packaging into the phage capsid is called a defective prophage 48. A satellite phage depends on a specific helper phage for its replication due to the absence of genes encoding structural proteins 49. As the phage genome is progressively reduced, leaving only a few genes encoding for specific proteins, it can give rise to bacteriocins and eventually gene transfer agents. These phage-like particles enable the transfer of random bacterial genomic fragments to other bacteria through a process resembling generalized transduction. However, unlike true phages, gene transfer agents lack the genetic elements required for encoding phage structural components 50. 1 Introduction 16 Figure 3: Possible phage life cycles. To infect a bacterial host, a phage must first recognize and bind specific surface receptors, then inject its genome into the cell. Virulent phages follow the lytic cycle: producing new virions and ultimately lysing the host. Temperate phages can either enter the lytic cycle or establish lysogeny, a dormant state maintained by a repressor protein that ensures replication of the prophage with the host genome. Stress conditions, such as those triggering the host SOS response (blue boxes), can induce the prophage to switch to the lytic cycle. In phage λ, the CI repressor maintains lysogeny by repressing early promoters (pL, pR) through octamer formation and DNA looping. Upon SOS induction, host RecA cleaves CI, reducing its levels and halting its synthesis, thereby initiating the irreversible transition to the lytic cycle 51. In addition to the well-characterised lytic and lysogenic life cycles, phages may also adopt alternative infection strategies. One such example is pseudolysogeny, a non-canonical state in which the phage neither integrates into the host genome nor initiates lysis, thereby failing to establish a stable, long-term relationship with the host 52,53. This condition typically arises under stressful environmental circumstances, such as nutrient starvation, and may revert to either the lysogenic or lytic cycle once favourable conditions are 1 Introduction 17 restored 54. A further alternative is the chronic infection cycle, observed in certain archaeal viruses, filamentous phages (e.g., rod-shaped single-stranded DNA phages), and plasmaviruses infecting Mycoplasma species. In this mode of infection, phage particles are continuously or intermittently released from the host cell without inducing immediate cell lysis, allowing prolonged coexistence between virus and host 23,29,34. 1.3 Phage mediated lysis Phage-mediated cell lysis is an autonomous, meticulously regulated, and temporally scheduled pathway that involves multiple proteins and manifests as a sudden, explosive burst with minimal time during which the host morphology is altered. The capacity to time this process with precision is a crucial fitness factor for the phage 55,56. The pathway is comprised of three distinct steps, each with two fundamentally different mechanisms: initiation through strictly coordinated triggering by holin or pinholin, cell wall degradation by endolysin or SAR endolysin, and outer membrane disruption by the i-spanin/o-spanin complex or the u-spanin 54. The holin function not only triggers the lysis pathway but also plays a crucial role in regulating its duration 57. Holin proteins always have a small cytoplasmic domain, a characteristic that is manifest at the C-terminus, in conjunction with a minimum of one transmembrane α-helical segment 58–60. It has been demonstrated that these proteins possess the capacity to be triggered, resulting in the formation of pores within the membrane. This, in turn, leads to a collapse of the membrane potential, disruption of active transport and an increase in permeability of the inner membrane 61,62. The classification of bacteriophage holins is determined by the topology of their transmembrane α-helical segments: Class I holins possess three transmembrane α-helical segments arranged in an N-out and C-in configuration (e.g. λ), Class II holins consist of two transmembrane α-helical segments arranged in an N-in and C-in configuration (e.g. ϕ21), and Class III holins have one transmembrane α-helical segment and a large periplasmic domain arranged in an N-in and C-out configuration (e.g. T4) 58,63,64. 1 Introduction 18 1.3.1 Canonical holin lysis pathway As the expression of late phage genes begins, holin proteins start to accumulate as freely mobile entities within the cytoplasmic membrane, while phage particle assembly continues (Fig. 4A). Holin proteins initially accumulate harmlessly as homodimers, with their hydrophilic faces sequestered against each other in the membrane 65–67. Simultaneously, the endolysin proteins, which possess transglycosylase activity, also accumulate in the cytoplasm as a monomeric, properly folded and fully active enzyme 68–70. Once the holin concentration reaches a critical, allele-specific threshold, those proteins suddenly cluster into large, two-dimensional aggregates known as death rafts (Fig. 4B, 5B) 71. These large, sparse structures are lipid-depleted due to the intimate helical packing of the holins, and are thus poor insulators. This leads to local reduction or collapse of the proton motive force (PMF), and consequently to a sudden halt in culture growth and respiration of the infected bacterial cell 72. The disturbance of the PMF gives rise to alterations in the orientation of certain TMDs of the holins, which in turn results in further leakage of protons 62,73. As the PMF continues to decline, the death rafts undergo a massive reorganization, terminating in the sudden formation of a small number of micron-scale holes lined by a single layer of holin molecules, averaging >340 nm in diameter 74. These large holes permit the endolysin to escape into the periplasmic space, thus facilitating the degradation of the peptidoglycan (Fig. 4B). The infected cells elongate until the actual lysis occurs, a process referred to as a 'blow out'. This is characterised by the sudden expulsion of cytoplasmic content including the progeny virions (Fig. 4C). This phenomenon is hypothesized to be caused by the degradation of the peptidoglycan in relation to the large lesions in the cytoplasmic membrane by escaped endolysins 75. 1 Introduction 19 Figure 4: Phage lysis, the canonical holin-endolysin lysis pathway. Cartoon model of one of the two path- ways of phage lysis of Gram-negative hosts exemplified by phage λ for the canonical holin-endolysin lysis pathway. During the expression of the late phage genes, the respective holins accumulate together with their antiholins in the inner membrane (A). In a similar manner, the endolysins have been observed to accumulate in the cytoplasm (A). Once a critical holin concentration is attained, holin triggering results in micron-scale holes (B), which release the endolysin into the periplasm (B). The subsequent peptidoglycan degradation results in the activation of the spanins, which overcome the barrier of the outer membrane by fusing the inner and outer membrane (see Chapter 2.1.1.2 and 3.1.2) and thus releasing the phage progeny (C) 54. In order to ensure that this process occurs in the correct sequence and at the appropriate time, the holin is controlled by the so-called antiholin prior to the triggering event 76. In phage λ, both protein products are generated from the same locus, which contains two Shine-Dalgarno (SD) sequences and two start codons 63,77. These elements are involved in ensuring an adequate translation, as well as a specific ratio of holin to antholin molecules. The holin-antiholin ratio for the bacteriophage λ is 2.5:1 under normal conditions, as a stem-loop structure in the mRNA determines the relative initiation frequency 78. The λ antiholin carries an additional positively charged residue at its N-terminus, thereby hindering the integration of this TMD1 into the membrane (Fig. 5A) 63. The antiholin then heterodimerizes preferentially with the holin, forming inactive holin-antiholin heterodimers that control the timing of lysis 79. However, depolarization induced after the triggering event ensures that the TMD1 of the antiholin can enter the membrane and convert the inactive antiholin to an active holin. As a result, given that only about one-third of the total holin gene products accumulate as homodimeric holins, while approximately two-thirds exist as holin-antiholin 1 Introduction 20 heterodimers, the instantaneous triggering event leads to a threefold increase in the number of active holin dimers 80. Figure 5: Holin structure. Schematic representation of the membrane topologies of two distinct classes of holin proteins: the canonical holin from bacteriophage λ (Panel A) and the pinholin from phage ϕ21 (Panel B). (A) The canonical holin S105 from phage λ features three transmembrane α-helices arranged in an N-out, C-in orientation. Its activity is negatively regulated by the antiholin S107. Upon a triggering signal, accompanied by the collapse of the proton motive force (PMF), the N-terminus of the holin undergoes a conformational change and flips into the periplasm (indicated by the blue arrow). This event facilitates the formation of large, micron-scale membrane holes. (B) The pinholin S2168 from phage ϕ21 consists of two transmembrane α-helices with both N- and C-termini located in the cytoplasm (N-in, C-in orientation). It is negatively regulated by the antiholin S2171. At the moment of triggering, the first transmembrane domain (TMD1) of S2171 exits the membrane and relocates to the periplasm (blue arrow), resulting in the assembly of small, heptameric pinholes in the membrane 60,66,81. 1.3.2 Pinholin SAR endolysin pathway This particular type of phage-mediated cell lysis is also initiated by the accumulation of the holins as dimers within the membrane (Fig. 6A). However, the structural characteristics of these proteins differ from the canonical holins. The so-called pinholins that are a part of this pathway are classified as class II holins and consequently possess two TMDs, exhibiting an N-in, C-in topology (Fig. 5B) 82,83. These proteins 1 Introduction 21 accumulate harmlessly with two TMDs in the bilayer, as the native topology is not capable of triggering. TMD1 is not essential for lesion formation and acts as a negative regulator of TMD2, which is the essential domain for hole formation 84. The TMD1 domains of both holin molecules of the accumulated dimers must exit the bilayer to activate the complex and to proceed down a pathway to triggering 85,66,86. These holins also reorganize into large, two-dimensional aggregates after reaching an allele-specific critical concentration - however, these are smaller and occur more frequently than in canonical holin lysis pathway (Fig. 5B, 6B) 82. The aggregates self-organize into so-called pinholes, which represent a homoheptamer of pinholins with the hydrophilic face of TMD2 facing the lumen 81. These pinholes have a diameter of under 2 nm, which is too small to allow most proteins to pass through 82,86. Therefore, this type of phage-mediated cell lysis requires so-called SAR endolysins to be able to attack the peptidoglycan 87. SAR endolysins (Signal Anchor Release endolysins) carry an N-terminal TMD, a signal anchor, which serves to engage the sec pathway of the host cell and then anchor the exported endolysin to the bilayer in a membrane potential-dependent fashion (Fig. 6A) 88,89. It should be noted that this form is enzymatically inactive. It has been observed that some of the SAR domains can leave the membrane spontaneously; however, the enzymes are released instantaneously when the PMF breaks down (Fig. 6B) 36. Subsequently, the enzymes fold into an enzymatically active form. In this process, the SAR domain plays a pivotal role in enzyme refolding. It facilitates structural stabilization through covalent disulfide bonds or non- covalent interactions with the enzyme's core 88–91. Enzyme activation promotes the degradation of peptidoglycan. After the pinholins are triggered, the pinholes are formed and the SAR endolysins are released, the infected cells begin to shorten and become round, ultimately resulting in their rupture (Fig. 6C). This further difference to the canonical holin lysis pathway is probably indicative of the even distribution of endolysins and thus peptidoglycan degradation. As previously outlined, TMD1 functions as an intrinsic intramolecular inhibitor of lysis when within the membrane. However, the presence of an antiholin within this lysis pathway has also been observed 83. The antiholin is distinguished by the presence of an additional positively charged residue at its N-terminus, a feature that hinders the release of TMD1 (Fig. 5B). The holins also dimerize with the antiholins in this lysis pathway, thereby slowing down the lysis clock. The decrease in PMF also leads to a topological change of 1 Introduction 22 TMD1 and thus converts the inactive holin-antiholin heterodimer into fully active molecules 81. Figure 6: Phage lysis, the pinholin-SAR endolysin lysis pathway. Cartoon model of one of the two path- ways of phage lysis of Gram-negative hosts exemplified by phage ɸ21 for the pinholin-SAR endolysin lysis pathway. During the expression of the late phage genes, the respective holins accumulate together with their antiholins in the inner membrane (A). In a similar manner, the endolysins have been observed to accumulate tethered to the membrane as SAR-endolysins during the pinholin lysis pathway (A). Once a critical holin concentration is attained, holin triggering results in small heptameric pinholes (B), which release the SAR endolysin from the inner membrane into the periplasm (B). The subsequent peptidoglycan degradation results in the activation of the spanins, which overcome the barrier of the outer membrane by fusing the inner and outer membrane and thus releasing the phage progeny (C) 54. 1.3.3 Spanin proteins For a considerable period, it was hypothesized that the degradation of the peptidoglycan and the resultant instability of the cell would be sufficient to overcome the barrier of the outer membrane and cause the infected cell to lyse. However, scientific research in this field over the past two decades has shown that this hypothesis is incorrect. To disrupt the outer membrane to allow phage egress, most phages studied to date fuse the inner and outer membrane of the host cell, a process that is facilitated by spanin proteins. These proteins had previously been overlooked due to the fragility of the cells in the context of a shaker flask and their resulting destruction by shear forces 75,92,93. To date, two different types of spanins have been identified: the so-called two-component spanins and the unimolecular 1 Introduction 23 spanins (Fig. 7). Current in silico studies have identified over 500 two-component spanins and just over 50 unimolecular spanins in the NCBI reference sequence database 36,75,94–96. Unimolecular spanins, also known as u-spanins, possess an outer membrane (OM) lipoprotein determinant and a C-terminal transmembrane domain (TMD) embedded in the inner membrane (IM) (Fig. 7A) 96. Notably, these spanins are encoded as a single gene within the bacteriophage lysis cassette 95. The two component spanins, such as the Rz and Rz1 proteins from phage λ, consist of a complex of an integral IM protein, the i-spanin, and an OM lipoprotein, the o-spanin, which spans the entire periplasm (Fig. 7B) 97,98. There are three different ways known thus far of encoding o-spanin and i-spanin in the phage genome: the two genes encoding these proteins can be nested (λ), overlapped (P2) or separated (T4) 95. Both spanin proteins are required for complete cell lysis. If either i-spanin or o-spanin is defective, cell lysis results in spherical cells lacking a cell wall and held together by the outer membrane 75. The spanin complex assembles in the cell envelope during morphogenesis via C-terminal interactions. Recent studies have shown that Rz and Rz1 accumulate as covalent homodimers, stabilized by three intermolecular disulfide bonds: two within the Rz dimer and one within the Rz1 dimer. Consequently, the λ spanin complex is structurally a dimer composed of Rz₂:Rz1₂ 99. During the early stages of lysis, these spanin complexes become trapped in the gaps of cross-linked peptidoglycan, rendering the PG layer a negative regulator of spanin activity 97,100. Following holin triggering, the collapse of the proton motive force, caused by pore formation in the inner membrane, facilitates the release of endolysins. This, in turn, leads to the degradation of the peptidoglycan layer. Degradation of the peptidoglycan layer permits lateral diffusion of spanin complexes, allowing them to assemble into functional oligomers 101. Conformational changes within these complexes generate the free energy required for outer membrane disruption 102. The resulting oligomerized structures have been demonstrated to mediate membrane fusion by bringing the inner and outer membranes into close apposition 103–105. 1 Introduction 24 Figure 7: Spanin structure. Schematic illustration of the membrane topology of the two-component spanin prototype from phage λ and the unimolecular spanin (u-spanin) from phage T1. (A) The uni- molecular spanin in T1 is composed of an N-terminus containing three fatty acyl chains (light blue) which attaches this part of the protein to the inner leaflet of the outer membrane. It is attached to the inner membrane via its C-terminal TMD (dark blue). The periplasmic domain of this u-spanin is predicted to be mainly extended beta sheets (turquoise diamonds) which connect the protein parts in the IM and OM through the PG meshwork. (B) In phage λ, the i-spanin is anchored in the inner membrane via an N-terminal transmembrane domain (green) and extends into the periplasm with a domain consisting of two α-helices connected by a linker region, likely forming a coiled-coil structure. The o-spanin is tethered to the inner leaflet of the outer membrane through three lipid modifications (orange square) and features a periplasmic segment (orange stick) that is predicted to be intrinsically disordered. The i- and o-spanins associate via their C-terminal regions to assemble into a spanin complex, bridging the IM and OM across the peptidoglycan layer 101,104,106. Bacteriophage-encoded lysis proteins are currently being developed for various applications in medicine, the food industry, biotechnology, and pharmaceuticals 7,8,107–112. For example, several research groups have demonstrated the potential for the use of spanins in the delivery of drugs and biochemicals into cells by means of membrane fusion 104. 1 Introduction 25 1.4 Phage protein interactions with host cellular machinery Phage-induced lysis represents a crucial and tightly regulated phase in the viral replication cycle. This event is preceded by an extensive remodelling of the host bacteriums metabolic and regulatory pathways. Such changes begin either immediately after infection or when the virus transitions from a lysogenic to a lytic life cycle. The genes involved in these early steps, commonly referred to as “early phage genes”, are among the first to be expressed. They typically encode for proteins that suppress the host's defence mechanisms, disrupt its gene regulation, and, quite often, arrest the bacterial cell cycle 113–115. These combined actions establish a cellular environment favourable to efficient phage propagation. In phage λ for example, expression of the Kil protein induces filamentation in host cells and inhibits cell division 116. Interestingly, Kil also slows down the timing of lysis by about 30% 117. This delay is thought to be due to Kil’s interaction with the bacterial protein FtsZ. By preventing FtsZ from polymerizing properly, Kil interferes with the formation of the Z-ring, which is essential for bacterial cytokinesis. Instead of forming a single, well-positioned division site, cells infected with λ exhibit a diffuse distribution of FtsZ or show misplaced rings. When Kil is overproduced, it appears to bind directly to FtsZ monomers, reducing their GTPase activity 117,118. Another example of a phage encoded protein that interferes with the division machinery is the Gp04 protein from phage T4, which also targets FtsZ. Gp04 binds to FtsZ and disrupts Z-ring assembly, resulting in elongated, filamentous bacterial cells with multiple mispositioned Z-rings. This represents evidence of a breakdown in the spatial control of division 119,120. Blocking bacterial division serves multiple strategic purposes for the phage. First, it helps to prevent the host from dividing too early, which could otherwise produce uninfected daughter cells 121. Additionally, by halting division, the phage ensures that all of the host’s metabolic assets, including energy, nutrients, and biosynthetic capacity, remain devoted to viral replication 122,123. Finally, inducing filamentous growth may provide more physical space for assembling and organizing large numbers of new phage particles 121,124. 1 Introduction 26 1.5 Bacterial cell division In order to better understand the importance of bacterial cell division for the life cycle of a phage, it is relevant to summarise the key factors of bacterial cell division at this point. Bacterial cell division is orchestrated by the divisome, a highly sensitive and dynamic macromolecular complex made out of approximately a dozen proteins (Fig. 8) 125. The included components are tightly regulated in both space and time throughout the cell cycle 126. Their coordinated actions are essential for preserving cell wall integrity against the internal turgor pressure as well as ensuring that cytokinesis occurs only after DNA has been accurately replicated and segregated. At the heart of the divisome is FtsZ, a tubulin- like protein with GTPase activity. FtsZ is key to the early stages of divisome formation; it polymerizes to form the Z-ring, a structure that marks the future site of cell division 127. This structure serves as a scaffold, guiding the assembly of additional division proteins and thereby establishing the cell’s division plane 128–130. To form the Z-ring, FtsZ assembles into polymers along the division plane using its polymerizing GTPase domain, which is one of its three conserved domains 131. This domain, somewhat resembling tubulin, mediates filament formation through nucleotide binding. Upon assembly of FtsZ monomers into filaments, GTP hydrolysis leads to a conformational shift from a closed to an open state, resulting in a depolymerization and thereby contributing to the curved architecture of the Z-ring 132. Recent studies have demonstrated that the FtsZ filament undergoes circumferential movement around the division plane, a process known as treadmilling 133. While the individual monomers remain stationary within the filament, the movement occurs through a net addition of FtsZ subunits at one end of the filament 134. FtsZ itself is not a membrane protein and cannot directly bind to the membrane. This interaction is mediated by membrane anchor proteins, such as FtsA and ZipA, which bind to the conserved C-terminal domain of FtsZ 135. The absence of either anchor protein leads to defective cell division, ultimately resulting in lethality 136. FtsA, a bacterial actin homologue, is a widely conserved membrane-binding protein that binds ATP with low affinity 137. It localizes to the centre of the cell in an FtsZ-dependent manner and tethers FtsZ to the membrane through interactions with the C-terminus 138. ZipA, present only in Gram-negative gammaproteobacteria, is an essential bitopic integral membrane protein, consisting of a large cytoplasmic domain connected to a single N-terminal transmembrane 1 Introduction 27 domain by an extended linker 139. ZipA shares a partially overlapping function with FtsA, since both proteins facilitate the binding of FtsZ to the membrane. However, ZipA interacts with the C-terminus of FtsZ through conserved residues distinct from those involved in binding FtsA 140. The interaction of ZipA with FtsZ is also thought to contribute to the stabilisation of the Z-ring structure. Together, FtsZ, FtsA and ZipA form the proto-ring, which serves as the basis for subsequent protein recruitment 141. The FtsE-FtsX complex attaches to this proto-ring and plays a crucial role in the recruitment of additional divisome components 142. After a distinct delay, the second stage of divisome assembly, known as maturation, occurs just before constriction 143. During this phase, a number of proteins, including FtsK, FtsQ, FtsL, FtsB, FtsW, PBP3, (FtsI)-PBP1B, and FtsN, are sequentially incorporated into the divisome 144. FtsW and FtsI are important enzymes for peptidoglycan synthesis and are highly conserved in all bacterial species 145. Their recruitment to the midcell depends on FtsK and the FtsQ-FtsL-FtsB complex, both of which require FtsA and ZipA for proper functioning 146. Disruption of either of these proteins results in a lethal block to cell division. However, the division process can be restored in cells lacking FtsE or FtsX if they are cultured in a media with high osmotic pressure 145. In addition to the essential divisome components, there are also non- essential proteins, such as the Zap proteins. Although their absence does not directly prevent cell division, these proteins, when functioning together, play an important role in ensuring the proper progression of normal cell division 147. Once this stage of cell division is successfully completed, the constriction of the cell wall follows, which serves as the primary driving force behind cell division. The process of cell wall constriction is likely triggered by the arrival of the final divisome proteins that activate cytokinetic cell wall synthesis 148. Given that the cell wall is crucial for maintaining cellular integrity, it is essential that this process proceeds without disruption. The cell wall is primarily composed of peptidoglycan (PG), which is a network of glycan strands interconnected by short peptide bridges 149,150. In Gram-negative bacteria such as E. coli, the PG layer is typically a single, 3-6 nm thick layer. The enzymes responsible for synthesizing the glycan strands are glycosyltransferases (GTases), while transpeptidases (TPases) catalyse the crosslinking of peptide side chains. PG synthases are classified into three categories: bifunctional GTase/TPase enzymes (class A penicillin-binding proteins, PBPs), monofunctional TPases (class B PBPs) and 1 Introduction 28 monofunctional GTases 151–153. FtsW and FtsI serve as the primary GTase and TPase enzymes during bacterial cell division in E.coli, respectively 154,155. The enzymes needed for PG synthesis require activation signals from previously recruited divisome proteins to initiate the constriction process 156,157. It is also hypothesized that FtsZ contributes to the regulation of PG synthase activity through its treadmilling action 127. However, PG synthesis is not solely regulated by the activation of FtsW and FtsI through intracellular divisome proteins. Constriction also necessitates activation signals from the OM lipoproteins LpoA and LpoB that act from outside the sacculus 158,159. Figure 8: Bacterial divisiome during cell division. The schematic shows the essential members (with the exception of the Zap proteins ) of the divisome multiprotein complexes for peptidoglycan synthesis during cell division in E. coli according to known localization patterns and interactions in the cell. The precise molecular architecture of the divisome and how they insert new material into the existing peptidoglycan layer are not yet known. The co-localization of MurJ with the divisome requires an active FtsW. MurJ requires the PMF in order to drive its conformational changes needed for its intended lipid transport mechanism - and by this the delivery of lipid II to the complex via FtsW. Lipid II is produced at the inner membrane by MraY and MurG. For simplicity, peptidoglycan hydrolases known to associate with each complex and to be required for their proper function, are not shown 146,160. The process by which new material is incorporated into the existing PG has yet to be clarified. It is established that the PG precursor lipid II is synthesized at the inner leaflet of the cytoplasmic membrane and eventually translocated into the periplasmic space via the lipid II flippase FtsW 161,162. FtsW has been shown to 1 Introduction 29 interact with the two major PG synthases, PBP3 and PBP1B, and is essential for the recruitment of PBP3 to the divisome, likely acting in coordination with PBP1B 158,163. Current evidence suggests that PBP1B and PBP3 constitute the principal enzymatic activities that drive septal peptidoglycan synthesis during cytokinesis. Moreover, it has been proposed that PBP3 plays a central role in regulating the spatial and temporal initiation of new PG synthesis, thereby ensuring precise coordination with the division machinery 158,164,165. Several regulatory systems such as the Min system and the nucleoid occlusion system are essential for ensuring the correct spatial and temporal positioning of the divisome 166,167. The Min system in E. coli consists of three proteins, MinC, MinD, and MinE, that work together to prevent the polymerization of FtsZ near the cell poles and thereby promoting accurate Z-ring assembly at midcell 168. MinD is a deviant Walker-type ATPase that associates with the cytoplasmic membrane when bound to ATP 169. Its membrane- associated partner, MinE, stimulates the hydrolysis of ATP, causing MinD to dissociate from the membrane 170. Once MinD exchanges ADP for ATP, it can rebind to the membrane, initiating a new cycle. This dynamic interplay leads to the formation of oscillatory concentration gradients of Min proteins between the cell poles, eventually creating a bipolar gradient that restricts FtsZ assembly to the cell centre 171. MinC, which interacts directly with two domains of FtsZ, acts as a strong inhibitor of FtsZ polymerization 172,173. Through its association with MinD, MinC is also subject to an oscillatory movement and thereby contributing to the suppression of Z-ring formation at the poles, where MinD is predominantly localized 174. In the absence of a functional Min system, cell division may still occur at midcell; however, division often takes place near the poles, leading to the formation of DNA-less minicells due to mispositioned septation 175. The nucleoid occlusion system in E. coli involves the DNA-binding protein SlmA, which specifically associates with certain chromosomal regions that are located away from the replication terminus in the centre of the cell and closer to the replication origin (OriC) near the cell poles 176. These binding interactions create a bipolar gradient of SlmA within the cell. In addition to its DNA-binding activity, SlmA interacts with FtsZ and thereby inhibiting its polymerization 177. Through this dual functionality, SlmA prevents the assembly of the FtsZ ring in the vicinity of the nucleoid, especially in regions close to OriC, and thus contributes to the spatial regulation of cell division 178. 1 Introduction 30 1.6 Model organism Shewanella oneidensis MR-1 Current research indicates that the genus Shewanella consists of over 70 species, with the majority inhabiting aquatic environments 179. Several species have been identified as opportunistic pathogens in humans and aquatic animals 180. Shewanella spp. are also known to adhere to diverse surfaces and form biofilms 181. S. oneidensis MR-1 is a Gram-negative, facultatively aerobic gammaproteobacterium that serves as a key model organism in microbial research, because it is able to utilize a broad spectrum of terminal electron acceptors like manganese oxide under anaerobic conditions 182–186. The genome of S. oneidensis MR-1 is made up of a 4.9 Mbp circular chromosome predicted to encode 4,318 proteins. Additionally, this bacterium carries a 161 kbp megaplasmid with 149 protein-coding genes 185,186. S. oneidensis MR-1 exhibits a rod-shaped morphology, measuring 2-3 µm in length and 0.4-0.7 µm in diameter. It is motile, utilizing a single polar flagellum for swimming 187,188. 1.6.1 Phages of S. oneidensis MR-1 The genome of S. oneidensis MR-1 contains four prophages: λSo, MuSo1, MuSo2 185,189,190 and CP4So 191. Sequence analyses have shown that λSo and MuSo2 share homology with the E. coli phages λ (morphological group of siphoviruses) and Mu (morphological group of myoviruses), respectively. Both λSo and MuSo2 are capable of forming intact, infectious phage particles. In contrast, MuSo1 does not produce active particles, as shown by the absence of plaque formation, even when λSo and MuSo2 are deleted from the bacterial genome 189. The fourth prophage, CP4So, is a P4-like cryptic element that can only be induced under specific conditions 191. Prophages are known to carry genes that can provide beneficial traits to their bacterial hosts, such as enhanced resistance to antibiotics and environmental stressors 192–194. This phenomenon is also observed in S. oneidensis MR-1, where the prophages λSo and MuSo2 play critical roles in biofilm development. These prophages significantly influence proper biofilm formation by mediating processes such as cell lysis, which may be necessary for the release of extracellular DNA and other factors involved in cell-cell and cell-surface interactions. While Mu-like phages primarily affect the early stages of biofilm development, λSo is the principal contributor to the formation of complex, three- 1 Introduction 31 dimensional biofilm structures 189,195,196. λSo is strongly induced in cells attaching to a surface, which is regulated by intracellular iron levels 189,195,197. Elevated intracellular iron triggers the SOS response via RecA, which subsequently induces the lytic cycle of λSo. The prophage λSo in S. oneidensis MR-1 has a genome size of 51 kbp and encodes 78 annotated genes 185,186. It is integrated into a genomic region flanked by two genes of unknown function, which are conserved among closely related Shewanella species. Although the genomes of S. oneidensis and its prophage λSo have been sequenced and annotated, and despite the growing volume of available phage genomic data, 23 genes within the λSo genome still encode proteins of unknown function 11. 1.7 Aim of this study The aim of this study was to characterise selected gene products of the prophage λSo in greater detail, with a particular focus on their effects on the host cell following infection or prophage induction. Special emphasis was placed on λSo-mediated cell lysis, in order to gain deeper insights at the protein level into this largely unexplored process. 2 Results 32 2 Results 2.1 Analysis of the Shewanella prophage λSo Research over the past few decades has revealed that phage genomes frequently contain gene clusters encoding rather small proteins, many of which lack identifiable structural homologues in existing databases 114. These proteins are typically expressed immediately after infection or the initiation of the lytic cycle and are critical for the reprogramming of the host cell One of the primary objectives of this study was to characterise the genome of phage λSo, with a particular focus on genes encoding proteins of unknown function. Prior research had identified two distinct gene clusters whose products are essential for the fitness of λSo198. The proteins encoded by these clusters are examined in greater detail in the following sections. 2.1.1 Characterization of the λSo lysis cluster Previous studies by Binnenkade et al. identified a gene cluster spanning SO_2966 to SO_2974 as playing a key role in phage-mediated cell lysis in λSo. This cluster is made up of genes ranging from 162 to 2215 bp, encoding proteins between 53 and 737 amino acids in length (Tbl. 1, Fig. 9A). It has been demonstrated that a deletion of this gene cluster does not prevent the formation of phage particles; however, these particles are no longer capable of inducing host cell lysis 198. Table 1: Lysis gene cluster gene number basepairs amino acids annotation SO_2966 162 53 protein of unknown function SO_2968 2215 737 Lambda phage terminase A SO_2969 330 109 putative HNH nuclease YajD SO_2970 360 119 protein of unknown function SO_2971 270 89 putative holin SSo SO_2972 628 208 protein of unknown function SO_2973 513 170 putative endolysin RSo SO_2974 264 87 pyridoxal phosphate dependent enzyme 2 Results 33 Primary sequence based homology analyses using BLAST suggest that two of these genes, SO_2971 and SO_2973, may encode a putative holin and endolysin, respectively. However, these functional assignments had not been proven experimentally thus far. Figure 9: Genetic organization of Shewanella prophage λSo lysis genes. The upper panel displays the fully annotated genome of λSo. The predicted open-reading frames within the phage genome are represented as arrows indicating their transcriptional direction. The boxes labelled A and B show the position of the gene clusters involved in phage-induced cell lysis, which are displayed in the middle and lower panels in more detail. The deduced gene products are colour-coded according to their predicted function. (A) Genetic organization of the predicted lysis cluster (SO_2971 to SO_2973). Genes encoding proteins directly involved in cell lysis are indicated as coloured arrows. The predicted translation start of the spanin component Rz1So within the +1 frame of the gene encoding spanin component RzSo is highlighted in yellow; the upstream ribosome binding site is also indicated. The three alternative start sites of the pinholin- encoding gene sSo are accordingly highlighted in yellow. Bioinformatic predictions are based on NCBI BLASTP (National Library of Medicine) and PHAST analyses. (B) Gene organization of so-called λSo cluster C [lcc1 (SO_4794) to lcc6 (SO_4975)]. Predicted genes are indicated as arrows. lcc6, which is part of the lysis machinery, is coloured in blue, the other lcc genes are coloured in light-grey, and the neighbouring genes are shown in dark-grey. dcm encodes a DNA (cytosine-5-)-methyltransferase, and xis encodes the phage DNA excisionase Xis. Bioinformatic predictions are based on NCBI BLASTP (National Library of Medicine). The homology analyses suggest that the gene product of SO_2971 functions analogously to a pinholin, as the predicted protein possesses two transmembrane domains and a putative N-in/C-in topology (Fig. S1). Notably, the open reading frame contains two alternative translational start sites located six and eight amino acids downstream of the annotated start codon (Fig. 9A). Similar to the mechanism described for the E. coli phage 2 Results 34 ɸ21 84, these alternative start sites could give rise to distinct holin isoforms, potentially constituting a holin/antiholin regulatory system. Taken together, these features support the hypothesis that the SO_2971 gene product serves as the key effector in initiating host cell lysis via depolarization of the cytoplasmic membrane. Accordingly, it will hereafter be referred to as λSo holin S (SSo). The protein encoded by SO_2973 represents a strong candidate for the putative endolysin, as structural predictions indicate the presence of an N-terminal SEC signal peptide, although a clearly defined proteolytic cleavage site following the signal sequence is missing (Fig. S1). This suggests that the protein likely belongs to the class of SAR endolysins, which are translocated into the periplasm but remain anchored to the cytoplasmic membrane via their uncleaved signal peptide. Release and activation of these endolysins is typically triggered by membrane depolarization, facilitated by pinholin- mediated pore formation. Accordingly, this protein will be referred to as λSo endolysin R (RSo). Structure-based homology searches using the HHPred tool in the HAMMER database, which is focusing on predicted 3D structures rather than primary amino acid sequences, suggested that SO_2972 may encode an additional key lysis-related protein: a putative spanin. Spanins are known to facilitate the final step of host cell lysis by disrupting the outer membrane, complementing the roles of holins and endolysins. The predicted structure of the SO_2972 gene product is consistent with the cytoplasmic and periplasmic component of a canonical two-component spanin system 100,101. Specifically, the protein is predicted to contain an N-terminal transmembrane domain (amino acids 9-31), followed by a long α-helical region (amino acids 32-110) (Fig. 9B). Further genetic analysis revealed additional features characteristic of known two-component spanin systems: an alternative open reading frame, designated SO_2972b, initiates within SO_2972 in the +1 reading frame starting at nucleotide position 340 (Fig. 9B). A putative Shine-Dalgarno sequence (GAAAGG) is located 7 bp upstream of the predicted start codon. The SO_2972b ORF extends into the intergenic region between SO_2972 and SO_2971 and encodes a 95- amino-acid protein that contains a lipoprotein signal peptide. Using the SignalP-Tool, this peptide is predicted to be cleaved between residues 18 and 19, exposing an N-terminal cysteine required for lipid modification and anchoring to the outer membrane (Fig. S1). Consequently, these proteins are designated as λSo Spanin Rz and Rz1 (RzSo, Rz1So). 2 Results 35 To gain deeper insight into the roles of these proteins during host cell lysis, bacterial strains were engineered with individual deletions of the genes encoding each respective protein. These experiments were conducted using a S. oneidensis background strain in which the two endogenous prophages, MuSo1 and MuSo2, had been removed. This was done to eliminate potential disruptive effects and ensure that any observed phenotypes could be attributed specifically to λSo. For clarity and simplicity, this strain will hereafter be referred to as the wild type (WT). The engineered strains were subsequently characterised in greater detail using fluorescence microscopy, enabling a more precise assessment of phage infectivity, lytic activity, and potential morphological changes associated with the deletion of individual genes. For this experimental setup, exponentially growing cultures of S. oneidensis MR-1 harbouring the respective genetically modified lysogenic λSo prophages were treated with mitomycin C (MMC) to induce the lytic cycle. Mitomycin C is used to induce the lytic cycle in temperate phages (such as phage λ) because it specifically triggers DNA damage in the host bacterium. In E. coli, this process triggers the SOS response, which in turn activates RecA. RecA is responsible for cleaving the repressor protein CI of the λ phage. In the absence of the CI repressor, the phage DNA is activated for replication and the lytic cycle is triggered 199,200. Following induction, cells were monitored using time-lapse phase-contrast microscopy to observe the dynamics of lysis. To verify phage protein expression and confirm phage particle assembly, the experiment was repeated using a S. oneidensis MR-1 background strain in which the gene product of SO_2960, which encodes a head-tail joining protein, was fused to the fluorescent protein Venus (Fig. S2B, S2C). Importantly, this fusion does not impair phage infectivity (Fig. S2A), allowing direct visualization of phage assembly processes in real time without functional disruption. Following induction of λSo in wild-type cells, a pronounced elongation phase was observed, culminating in complete cell lysis approximately 180 minutes post-induction. In contrast, cells carrying the λSo variant with a deletion of the putative holin gene (ΔsSo) exhibited no lysis (Fig. 10B). In these cells, the expression of proteins that are part of the phage particles was confirmed via fluorescence microscopy using the strain in which a structural protein was fused to a fluorophore (Fig. 10A). Despite this, no plaque-forming units were detected in the culture supernatant, although a certain part of the λSo genome 2 Results 36 remained detectable within the cells via PCR (Fig. 11A). This suggests that λSo could be present in the cells without being lytically active. Complementation through ectopic expression of the deleted gene from a plasmid under the control of the putative late promotor of the λSo lysis cluster restored normal lysis (Fig. S3, Fig. 7). Figure 10: λSo-induced lysis of S. oneidensis MR-1 requires a pinholin, SAR endolysin and a spanin complex. (A) Microscopic images of S. oneidensis SO_2960-GGS-Venus cells (a strain, in which a head-tail- joining protein of λSo was fused to the fluorophore Venus) that bears deletions in the genes encoding for the pinholin (SSo), SAR endolysin (RSo) or i-spanin (RzSo) as indicated. The time points subsequent to λSo induc- tion by addition of mitomycin C (10 μg/mL) is 150 min. The images were taken in phase contrast and in the fluorescence channel of the fluorochrome YFP. The scale bar represents 5 μm. The liquid cultures were supplemented with 10 mM MgCl2. (B) Micrographs display a time-lapse series of S. oneidensis strains in which the genes encoding for the pinholin (SSo), SAR endolysin (RSo), i-spanin (RzSo) or o-spanin (Rz1So) were deleted from the λSo genome as indicated. The time points subsequent to λSo induction by addition of mitomycin C (10 μg/mL) are indicated above. The scale bar represents 5 μm. The corresponding complementation strains are shown in Fig. S3. 2 Results 37 To investigate the subcellular localisation of the holin protein, a recombinant construct for ectopic expression, in which the holin was fused to the green fluorescent protein (GFP), was constructed using a plasmid under the control of an inducible arabinose promoter. This plasmid was introduced into a background strain in which the gene for the holin was deleted. Successful expression of the fusion protein was confirmed by immunoblot analysis (Fig. S4B). Fluorescence microscopy of the corresponding cells revealed distinct foci, predominantly arranged in large clusters at the periphery of the cells (Fig. S4A). This spatial distribution strongly suggests that the holin localises to the cytoplasmic membrane. However, multiple bands were observed in the corresponding western blot, which may indicate the presence of degradation products of the fusion protein. These findings suggest that the stability of the GFP-holin fusion is limited under the conditions tested. Moreover, subsequent to the fusion of Holin and GFP, the phage particles proved to be incapable of accurate host cell lysis. Consequently, the outcomes of this experiment should be interpreted with caution and may be in need of optimisation. To further investigate membrane integrity during phage induction, a membrane depolarization assay was conducted using DiBAC4(3) staining. Cells were stained with 1 mg/mL (wt/vol) DiBAC4(3) at a defined time point following λSo induction, and fluorescence intensity was measured using a TECAN plate reader. Four hours post- induction, a small but statistically significant reduction in depolarization was observed in ΔsSo cells compared to the wild type (Fig. 11B, Fig. S6). These results provide further evidence supporting SO_2971 as the functional pinholin in the λSo lysis system. Cells carrying a deletion of the SO_2973 gene, encoding the putative endolysin RSo, exhibited a comparable phenotype under microscopic observation: following phage induction, they underwent elongation but did not lyse, despite clear evidence of phage protein production (Fig. 11). Consistent with this, no plaque-forming units were detected in the culture supernatant (Fig. 11A). The lytic defect could be rescued by ectopic expression of the gene product of SO_2973 from a plasmid, confirming the functional role of the encoded protein (Fig. S3). However, these cells showed a significant increase in membrane depolarisation. This observation suggests that in the absence of the putative endolysin, depolarization, which is presumably initiated by holin activity, can proceed, but lysis is blocked, allowing the accumulation of depolarized cells. Thus, the signal persists longer in the population than in wild-type cells, where it would normally be rapidly lost upon cell lysis. 2 Results 38 Figure 11: λSo-induced lysis of S. oneidensis MR-1 requires a pinholin, SAR endolysin and a spanin complex. (A) Lysis profiles of λSo mutants by spot test analysis. Upper panel: phage lysates of λSo mutants were spotted in different dilutions onto a bacterial lawn containing cells of an exponentially grown S. oneidensis host culture. Lower panel: as a control, PCR was performed with the phage lysate to determine the presence of λSo in the respective lysates. (B) Membrane depolarization assay on λSo mutants using Di- BAC4(3). The cells were grown to exponential phase prior to λSo induction by mitomycin C phase and stained with DiBAC4(3) to visualize depolarization after 4 hours. Statistical significance was determined us- ing a two-way ANOVA (analysis of variance) and is indicated by the P value. NS, not significant, *P value = P ≤ 0.05, **P value ≤ 0.01, ***P value ≤ 0.001, and ****P-value ≤ 0.0001. Fluorescence microscopy images revealed relatively evenly distributed foci along the periphery of the cells (Fig. S4A), suggesting a uniform localisation of the SO_2973- encoded proteins within the cytoplasmic membrane. However, the correspondingly modified phage particles were unable to lyse the cells in the expected manner. This must be considered when evaluating the results. To assess the function of the putative spanin components, a deletion was performed targeting the gene SO_2972a, which encodes the predicted i-spanin RzSo (Fig. 9A). This deletion also disrupted the overlapping gene SO_2972b. To specifically inactivate SO_2972b without affecting the RzSo protein, a GTG- to-GTC substitution was introduced at position V132 in SO_2972a, resulting in a Cys19Ser change in SO_2972b that eliminates the cysteine required for lipid anchoring (Fig. 9, S2). Both resulting mutant strains displayed a striking phenotype in time-lapse microscopy: cells elongated normally after phage induction but, instead of undergoing lysis, a subset of both mutants rounded up into spherical forms (Fig. 10). When grown in planktonic shaking cultures, plaque-forming units were detected in the supernatant, but at levels three to four orders of magnitude lower than those of the wild type. Membrane depolarization 2 Results 39 measurements using DiBAC4(3) and TECAN analysis revealed a signal comparable to that observed in the SO_2973 deletion strain, indicating that depolarization occurs, but lysis is impaired. Fluorescence microscopy of the SO_2972a-GFP fusion protein revealed the presence of discrete, clustered foci localized along the cell periphery. In contrast, in a strain expressing a recombinant variant bearing the Cys19Ser mutation and thereby abolishing the function of the SO_2972b protein, a markedly altered distribution pattern was observed, characterised by uniform, non-clustered foci dispersed along the periphery of the cell. Since the lysis proteins of phages infecting Gram-negative bacteria are often functionally interchangeable despite differences in their structures 107,201, the identified lysis proteins of λSo were ectopically expressed from a plasmid in E. coli MG1655. As demonstrated in Figure 13, the λSo lysis proteins were also capable of inducing cell lysis in E. coli. However, it should be noted that the duration of this process was significantly longer in this particular setup. Figure 12: Proteins related to λSo cell lysis also cause cell lysis in E. coli. Time-lapse series of E. coli MG1655 in which the λSo pinholin, SAR-Endolysin, i-spanin and o-spanin were ectopically produced using the vector pBAD33. Cell cultures were supplemented with 0.2% arabinose to induce protein expression. The beginning of the time-lapse series (0 minutes) was defined as the time point at which arabinose was added to the respective cell culture. EVC resembles the empty vector control. The scale bar represents 5 μm. 2.1.1.1 Further characterization of the holin protein To gain deeper insights into the mechanism of phage-mediated lysis in λSo, the gene SO_2971 was subjected to detailed in silico analysis. This revealed the presence of two alternative translational start sites located six and eight amino acids downstream of the annotated start codon. These alternative start sites may give rise to shorter protein variants, potentially functioning as anti-holin. 2 Results 40 Figure 13: Alternative translational start sites encode a potential anti-pinholin. (A) Primary structure of the putative pinholin with corresponding predicted domains indicated. (B) 3D structures of the proteins of interest (λSo Holin) were predicted using Alphafold2 trough DeepMind´s Colab. The structures were visualised with Pymol. Certain protein domains are indicated using brackets. The structures are colour-coded by B-factor values. The respective amino acids that were modified are colour-coded as shown in the primary structure. (C) Micrographs display a time-lapse series of S. oneidensis strains in which the gene encoding for the pinholin (sSo) was deleted from the λSo genomes or modified as indicated. Cells were captured over a period of 120 minutes after being treated with 10 µg/ml MMC. The scale bar represents 2 μm. To test this hypothesis, a series of targeted genetic modifications were introduced into SO_2971, including deletions, truncations, and point mutations. Specifically, the first eight codons (SSo82) and the first ten codons (SSo80) were removed to evaluate whether these N- terminal regions influence the timing of cell lysis, as assessed by fluorescence microscopy (Fig 13A). Additionally, a base substitution (K4N) was introduced, replacing the positively charged lysine at position 4 with asparagine, a neutral but polar amino acid (Fig. 13A). This mutation was designed to probe the role of the N-terminal charge in modulating holin function. 2 Results 41 Following induction of the phage with MMC, wild-type cells exhibited pronounced filamentation, ending in cell lysis after approximately 120 minutes. As previously described, the ΔsSo mutant showed comparable filamentation but failed to undergo lysis. The SSo80 strain, in which the gene was truncated at the second alternative start codon, displayed a similar non-lytic phenotype. In contrast, the SSo82 mutant, which carried a truncation at the first alternative start codon of the gene, underwent significantly earlier lysis with cell rupture beginning around 90 minutes post-induction. Similarly, the point mutant SSo89 (K4N), in which lysine at position 4 was substituted with asparagine, also exhibited premature lysis, resembling the phenotype of the SSo82 strain (Fig. 13B). 2.1.1.2 Further characterization of the spanin system As previously noted, infectious plaque-forming phage particles were detected in the ΔrzSo deletion strain, although at levels three to four orders of magnitude lower than those observed for the wild-type phage. Prior studies have demonstrated that spanins play a particularly critical role in phage egress under conditions that stabilize the outer membrane, such as in the presence of Mg²⁺ ions 202. To investigate whether the spanin candidates identified in this study exhibit a similar dependency, both microscopic analyses and plaque assays were repeated for the ΔrzSo strain under addition of 10 mM MgCl₂. 2 Results 42 Figure 14: The spherical formation of the cells is due to the absence of SO_2972 and is stabilized by the addition of MgCl2. (A) Lysis profiles of λSo mutants after addition of divalent cations by spot test analysis. Upper panel: phage lysates of λSo mutants were spotted in different dilutions onto a bacterial lawn containing cells of an exponentially grown S. oneidensis host culture. Lower panel: as a control, PCR was performed with the phage lysate to determine the presence of λSo in the respective lysates. (B) Displayed are micro- graphs of a time-lapse series with S. oneidensis wild type and spanin-lacking (ΔrzSo; Rz1So C19S) cells treated with and without 10 mM MgCl2 subsequent to phage induction using mito- mycin C (10 μg/mL). The time points are indicated above; time point zero is defined as the start of induction. The scale bar represents 2 μm. (C) Morphology changes of cells lacking a spanin system after phage induc- tion. Micrographs were taken after the onset of sphere formation. Full conversion into a sphere takes about 60 seconds. The scale bar represents 2 μm. 2 Results 43 The membrane-stabilizing conditions induced by the presence of divalent cations prevented the detection of infectious phage particles in the supernatant of planktonic shaking cultures, as no plaques were formed, despite the presence of λSo DNA confirmed by PCR (Fig. 14A). Thus, the impaired cell lysis of the mutant phage is only observable under conditions lacking additional Mg²⁺. Microscopic analyses further revealed that the spheres formed by RzSo/Rz1So mutants were significantly more stable in the presence of Mg²⁺ and remained intact even 220 minutes post-induction with MMC (Fig. 14B). Moreover, the membrane-stabilizing conditions revealed that these morphological changes typically initiated at one pole of the elongated cell and progressed along its length within approximately one minute (Fig. 14C). To further characterise the identified spanins, an in silico structural analysis was performed using AlphaFold2 and PyMOL. This analysis suggested that RzSo and Rz1So most likely form dimers, which may assemble into a dimer of dimers (Fig. 15A, Fig. S1, Tbl. S3). The proteins appear to interact via their C-termini, while their N-termini are anchored in the inner membrane (RzSo) or outer membrane (Rz1So), respectively. The interaction between the dimers of the two spanins is predicted to induce a conformational change, potentially resulting in the "folding" of the RzSo dimers, a process that could facilitate membrane fusion. Using PyMOL, a putative interaction interface between the proteins was identified. To experimentally validate the in silico predictions, recombinant protein versions were designed in which selected amino acid residues within the proposed interaction interface were substituted in order to disrupt the interaction and to assess the impact on the lysis phenotype (Tbl. 2, Fig. S8). The mutant phage’s carrying the genes for these proteins were subsequently employed in plaque assays. Table 2: Amino acid substitutions in RzSo and Rz1So RzSo Rz1So T122S C47Y N129D Y52F C139R S54N N141D K55R C150R K56R N57G C74S 2 Results 44 Figure 15: Structures of the λSo spanin complex: a dimer of dimers. (A) 3D structures of the protein complex of interest (λSo i-Spanin; λSo o-Spanin) were predicted using Alphafold2 trough DeepMind´s Colab. The structures were visualised with Pymol. N- and C-termini are indicated using arrows. The structures are colour-coded by respective protein (i-spanin green, o-spanin yellow). The hypothesised interaction surface of the i- and o-spanin proteins is shown in magenta. (B) Lysis profiles of λSo mutants by spot test analysis. Phage lysates of λSo mutants were spotted in different dilutions onto a bacterial lawn containing cells of an exponentially grown S. oneidensis host culture. Quantification of Lcc mutant plaque formation on S. oneidensis host cells. To determine the phage titer, the plaques were counted, and PFU/mL was calculated. Error bars represent the standard error of three independent experiments. As shown in Figure 15B, the mutant phage’s exhibited plaque forming units (PFU) levels comparable to those of the ΔrzSo phage under conditions lacking divalent cation supplementation. Notably, phages carrying the substitutions N141D and N129D in RzSo, 2 Results 45 and K56R in Rz1So significantly reduced plaque formation compared to the control strain. Under membrane-stabilizing conditions, no plaques were detected for phage’s harbouring proteins with substitutions at S54 and C74 in Rz1So, a phenotype consistent with that of the ΔrzSo phage. In contrast, all other introduced substitutions resulted in improved lysis relative to the control strain. The most pronounced increase in PFU under membrane- stabilizing conditions was observed for phage’s carrying the C47Y and Y52F substitutions in Rz1So (Fig. 15B). 2.1.1.3 Identification of a novel lysis component To elucidate the effects of λSo on the host cell during infection or upon induction of the lytic cycle, RNA-seq analysis was performed on Shewanella oneidensis wild type strains lysogenized with λSo, as well as on a mutant strain in which the λSo gene cluster had been deleted (Δλ)198. The samples used for this approach were collected under standard growth conditions following induction of the SOS response via MMC treatment. The analysis revealed that, upon MMC addition, a gene cluster located near the end of the prophage genome was markedly upregulated. These genes ranked among the most highly expressed following induction. Notably, this cluster is situated upstream of the int and xis genes, which encode the integrase and excisionase enzymes, respectively. This gene cluster, designated Lambda cluster C (Lcc) based on its genomic position down- stream of two clusters named Cluster A and B, is made up of six genes (SO_4794, SO_3007 - SO_3010, and SO_4795) (Fig. 9B). These genes range in size from 126 to 414 bp and encode proteins between 41 and 137 amino acids in length (Tbl. 3). The Lcc proteins lack any clear homologies to known domains, except Lcc4 and Lcc6, which are predicted to harbour a transmembrane domain (Fig. S1). Previous studies demonstrated that the deletion of the Lcc gene cluster leads to a pronounced reduction in the number of plaque-forming units (PFU), suggesting that, while the cluster is not strictly essential for phage viability, it is crucial for efficient and productive phage propagation. 2 Results 46 Table 3: λ cluster C gene number basepairs amino acids annotation SO_4794, lcc1 249 82 protein of unknown function SO_3007, lcc2 126 41 protein of unknown function SO_3008, lcc3 414 137 protein of unknown function SO_3009, lcc4 276 91 protein of unknown function SO_3010, lcc5 201 66 protein of unknown function SO_4795, lcc6 198 65 protein of unknown function To shed light on the role of the Lcc proteins in host cell takeover following infection or induction of the lytic cycle, individual gene deletion strains were generated and subjected to further functional analyses. A single deletion of lcc1 could so far not be obtained. To assess the impact of each deletion on λSo infectivity, plaque assays were performed. For this purpose, phage lysates derived from the genetically modified λSo variants were prepared and applied at various dilutions to a S. oneidensis background strain in which the λSo integration site had been deleted. As a result, the phage was restricted to the lytic cycle, thereby allowing direct evaluation of lytic efficiency and infectivity. Analysis of the single-gene deletion mutants of cluster C revealed that the observed reduction in PFU is specifically attributable to the deletion of lcc6 (Fig. 16A). In contrast, deletion of lcc2, lcc3, lcc4, or lcc5 had no significant impact on λSo proliferation under the tested conditions (Fig. 16A). Microscopic examination further demonstrated that Lcc6 plays an essential role in cell lysis, as cells carrying the Δlcc6 mutation elongated, but failed to lyse following induction of λSo with MMC in a way that is highly reminiscent to the phenotype of holin or endolysin mutants (Fig. 16B, 17B). This defect was successfully complemented by ectopic expression of Lcc6 from a plasmid. Notably, ectopic expression of Lcc6 alone in a S. oneidensis MR-1 background strain did not result in any observable changes in cell morphology (Fig. 16C). 2 Results 47 Figure 16: λSo-induced lysis of S. oneidensis MR-1 also requires a small transmembrane protein Lcc6. (A) Quantification of Lcc mutant plaque formation on S. oneidensis host cells. To determine the phage titer, the plaques were counted, and PFU/mL was calculated. Error bars represent the standard error of three independent experiments. (B) Displayed are micrographs of a time-lapse series of S. oneidensis Δlcc6 mutant cells in which Lcc6 was produced after induction of the λSo lysis cassette. The cell culture was first treated with 10 μg/mL MMC for 180 minutes for induction of the lysis cassette and was then divided, with one supplemented with 0.2% arabinose to induce Lcc6 expression, while the other received no additional supplementation. After an incubation of 60 minutes, images were generated every 5 minutes for a total of 2 hours. The scale bar represents 2 μm. (C) Time-lapse series of S. oneidensis MR-1 ΔλSo ΔMuSo1 ΔMuSo2 in which Lcc6 was ectopically produced using the vector pBAD33. Cell cultures were supplemented with 0.2% arabinose to induce protein expression. The beginning of the time-lapse series (0 minutes) was defined as the time point at which arabinose was added to the respective cell culture. EVC resembles the empty vector control. The scale bar represents 5 μm. 2 Results 48 Figure 17: Lcc6 is required for λSo-induced cell lysis. (A) Membrane depolarization assay on λSo mutants using DiBAC4(3). The cells were grown to exponential phase prior to λSo induction by MMC and stained with DiBAC4(3) to visualize depolarization after the indicated time points. Statistical significance was determined using a two-way ANOVA (analysis of variance) and is indicated by the P value. NS, not significant, *P value = P ≤ 0.05, **P value ≤ 0.01, ***P value ≤ 0.001, and ****P-value ≤ 0.0001. (B) Lysis profiles of λSo mutants by bulk culture OD measurements. The respective λSo mutant cultures following inductions by MMC were incubated with constant shaking. Cultures to which 10 mM MgCl2 was added are displayed by open symbols; those with no additional MgCl2 are displayed by closed symbols. The shown data in this figure are representative of biological duplicates. (C) Visualization of the protein expression over time of the holin SSo fused to a sfGFP and Lcc6 fused to a single FLAG-Tag by western blot analysis. Time points are shown as minutes after induction of the phage with 10 μg/μL MMC. Polyclonal antibodies against GFP and 3xFLAG were used to detect the respective proteins. Sample normalization was achieved by adjusting cell suspensions to the same optical density at 600 nm (OD600) and analysis of stained SDS-PAGE gels. The original data file of the cropped and reassembled western blot is displayed in Fig. S5. (D) Bacterial two hybrid analysis of Lcc6 with holin SSo, endolysin RSo and the i-spanin RZSo. E. coli BTH101 was co-transformed with two-hybrid vector plasmids (put18/put18C, pKT25/pKNT25) expressing fusions of the protein of interest with the T18 and T25 domains, respectively. Transformants were spotted onto LB agar plates with IPTG and X-Gal and incubated at 30 °C for approximately 24 h. The plasmids put18C-zip and pKT25-zip were cotransformed as positive control (+). The negative control (-) contains the empty vectors put18c and pKT25. A blue colouring of the colonies (see positive control) shows a direct interaction of the hybrid proteins. 2 Results 49 The positioning of the lcc genes directly upstream of xis suggests that Lcc expression occurs early during phage activation, whereas the expression of lysis genes such as the holin is expected to take place during later stages of the phage life cycle. To better understand the role of Lcc6 in phage-mediated cell lysis and to determine at which stage of the lysis process the protein might act, mutant strains were generated in which sSo and lcc6 were chromosomally replaced by hybrid genes encoding N-terminally sfGFP-tagged holin or FLAG-tagged Lcc6. Although the resulting fusion proteins were non-functional, they allowed detection of the respective proteins by Western blotting. Following MMC induction, protein production was monitored, revealing that Lcc6 could be detected as early as 30 minutes post-induction, whereas holin expression was first observed after 90 minutes (Fig. 17C). As anticipated, Lcc6 is expressed early following λSo induction. This observation is further supported by the minimal membrane depolarization observed, whic