De impact van fagen op de fysiologie van Pseudomonas aeruginosa

Bacterial viruses or ‘bacteriophages’, the natural enemies of bacteria, have long since been considered an alternative to antibiotics to treat bacterial infections. This has led to the characterization of a high number of phages, which genomes contain a high number of genes with unknown functions. During millions of years of co-evolution a fraction of these genes have evolved to inhibit, activate and redirect the host towards efficient phage production, often through protein-protein interactions. As such, these newly characterized genomes provide a rich source of potentially interesting biotechnological and antibacterial proteins. To investigate this, a set of Pseudomonas phages were characterized and their early genes cloned in the bacterial genome by our research group. The goal of this dissertation was to identify early phage proteins that impact relevant bacterial processes for the development of phage-derived applications and to better understand the underlying biology. As an opportunistic human pathogen, known for its high and still increasing antibiotic resistance, there is a dire need for novel antibacterials against Pseudomonas aeruginosa. To this end, three previously identified toxic proteins were investigated in terms of understanding the mode of action for their toxicity and their potential in the development of novel small-molecule antibiotics. The most promising protein was LUZ24 gp9, which interacts with the gyrase B subunit and inhibits gyrase activity. Although we were not able to determine its 3D-structure, we could identify two amino acids, essential for the interaction. Our results enable the use of an interaction-based screen to identify lead molecules for the development of novel antibacterials. Next, we determined the 3D-structure of LUZ7 gp14, which shares homology to the OB-fold found in ssDNA binding proteins. Indeed, LUZ7 gp14 displayed an aspecific DNA binding capacity with a preference for ssDNA. We now propose that LUZ7 gp14 is a functional homologue of a previous identified aspecific ssDNA-binding protein in N4. The final protein, LUZ7 gp8, seems to have no bacterial interaction partner. Despite this it is an interesting protein for antibacterial development, as it is bacteriolytic and triggers cell wall permeabilisation. Thanks to the small size (8 kDa), a peptidomimetic might be the road to take with this protein to turn it into a phage-derived antibacterial. Biofilms are complex surface-attached multicellular structures, which protect the bacteria against predation and other environmental stresses. Hence, they have a detrimental impact in industrial systems (e.g. water systems) and also enable environmental bacteria (e.g. Pseudomonas) to successfully infiltrate our health care. Since P. aeruginosa is often used as model organism for Gram-negative biofilm formation, we set out to investigate if its phages encode proteins that might help us to combat these biofilms. Screening our library of 180 early phage genes for altered biofilm formation and excluding proteins previously found as antibacterial, identified a total of 35 proteins that significantly increased (15) or decreased (17) biofilm formation. To our knowledge, this is the first large-scale screen to identify early phage genes specifically targeting biofilm formation. Four phage proteins cause a >30% decrease in biofilm mass and of LUZ7 gp33 the 3D structure is known, making it the top priority for the development of a novel antibiofilm peptide or small molecule if an interaction is required. Intriguingly, the Pbunalikevirus LMA2 encodes the other three proteins, which indicates variability between phages in their impact on biofilm formation. We also identified four phage proteins that influence c-di-GMP signaling. YuA gp44, a predicted diguanylate cyclace (DGC) with the active site GGDEF, triggers wrinkled colony morphology. While 14-1 gp11, interacts with the membrane-bound DGC YfiN (PA1120) and has two close homologues (LMA2 gp11.1 & LBL3 gp11.1). Bacteria expressing these genes lose their motility, constistent with an increased activity of YfiN, and the deletion of YfiN abolishes 14-1 gp11 activity. Our hypothesis is that the interaction with 14-1 gp11 causes conformational changes in YfiN to activate its DGC domain. The resulting spike of c-di-GMP can trigger dispersion, to allow phage progeny to escape the biofilm matrix and maximize its chances to encounter new hosts. As an application, 14-1 gp11 could be a novel motility inhibitor to prevent establishment and spread of infections in fast-flow areas, e.g. bladder. In a third part, we wanted to investigate the uncharted territory of host metabolism take-over to identify novel tools for metabolic engineers to overcome bottlenecks in the host metabolism during the optimization of the production of a desired endproduct. In a way, phages do exactly the same to optimize viral replication and encode their own genetic toolbox for this. For this we first looked at the changing metabolic content of phage-infected P. aeruginosa cells, within a single infection cycle, for six phages and found phage-specific alterations of the host metabolism. A clear distinction can be made between ‘leeching’ phages, which take all existing resources for viral replication (e.g YuA), and phages that actively modulate the metabolism, by redirecting the cell’s metabolism to produce new resources required for phage production (e.g. phiKZ). We could show that metabolic pathways, targeted by phage-encoded auxiliary metabolic genes (AMGs), are statistically enriched in the set of metabolites changing upon infection. Hence, each specific set of AMGs will result in a unique metabolic phenotype. Thus not only the occurrence of phage infection but also the diversity of the phages present will impact the net metabolic effect of phage infections in complex bacterial populations. In the future, high-throughput metabolomics could identify novel ‘metabolic modulators’, which also enable linking the observed metabolic effects during infection to these encoded AMGs. Finally, we describe for the first time two phage-encoded acetyltransferases (LUZ19 gp13 & gp28) that use reversible lysine acetylation to alter host protein function during infection, which adds a, never before reported, layer of complexity to phage-host interactions. LUZ19 gp13 strongly acetylates two lysines in MetE and, to a lesser extent, two lysines in MetK. Upon expression of LUZ19 gp13, a rapid increase in the level of 5’-methylthioadenosine (MTA) and adenine was observed via metabolome analysis. While MetE and MetK are involved in the synthesis of S-adenosylmethionine (SAM), essential for polyamine biosynthesis, MTA and adenine are by-products of polyamine biosynthesis. We hypothesize that acetylation by gp13 activates MetE or/and MetK to increase SAM production and polyamine biosynthesis, which can be used as counter-ion for DNA packaging. From a biotechnological point of view, LUZ19 gp13 might be an interesting metabolic modulator to construct a superior strain for the production of SAM, a frequently used supplement in the US and Canada. LUZ19 gp28 triggers the cleavage of the α subunit from the RNAP complex, between Gln244-Glu245 in the flexible linker domain. This physical disconnection of the αCTD from the RNAP complex during late infection does not resemble any known mechanism of viral transcriptional shutdown. While acetylation was suspected to alter the substrate specificity of a host cytoplasmic protease, we could not identify large differences in the acetylation level of a single protease. In contrast, acetylation seemed to target two proteins of fatty acid biosynthesis (DesA and PA5174). We now propose that Gp28 shuts down de novo fatty acid biosynthesis, leading to oxidative stress that triggers an ‘emergency break’ of transcription by αRNAP cleavage. Once the bacterium has coped with the stress, normal transcription is restored. Preliminary data fits this hypothesis, however more research is required. Our broad approach to mine the enormous biotechnological potential embedded in the sequences of phage genomes, resulted in the find of some promising leads for phage-derived applications, e.g. the gyrase-binding LUZ24 gp9 as antibacterial, LUZ7 gp33 as novel biofilm inhibitor and LUZ19 gp13 as metabolic modulator. As these results originate from a library of a mere 150 early phage genes, the potential present in the virome might be an endless source for phage-based applications. Especially for antibacterial development, a bacteriophage-based platform for new target discovery and drug development could offer a solution to stay ahead of the inevitable development of microbial resistance. Aside from these potential applications, we managed to get a wealth of insights into phage biology. In the future, one of our leads will have to be developed to a finished application, proving the potential for the development of phage-derived applications towards industry. This could then hopefully trigger a shift in the focus of the bacteriophage research field from its current genome-orientation, towards the molecular investigation of these intracellular phage-host interactions during infection. From a more fundamental view, these increased efforts will also help to fill the gap between functional analyses of bacteriophage genes and the booming number of uncharacterized novel phage genome sequences. These will be exciting times as we will get a better understanding and appreciation for the complexity of viral replication at the system’s level.nrpages: 199status: publishe