Whole transcriptome analyses of Pseudomonas aeruginosa after infection by representatives of different phage clades

Detailed knowledge of the progression of bacteriophage infection has been limited to few model bacteriophages, mostly infecting Escherichia coli, as well as the results of low throughput and increasingly antiquated methods. However, state-of-the-art approaches now make it possible to explore decades of interest in the transcriptional strategies of phage as well as how they affect host transcription and metabolism, now re-sparked by modern therapeutic and biotechnological potential, as well as the ability to ask deeper more comprehensive questions. By performing RNA-Seq on the non-ribosomal RNA of synchronously infected cells we are able to compare host transcript data for early, middle and late transcription with six fundamentally different lytic phages. These include PhiKZ, PEV2, 14-1, LUZ19, YuA, and the novel F7 infecting P. aeruginosa PAO1. We also collaborated with colleagues at the Pasteur Institute to collect similar data for PAK_P3 and PAK_P4 infections of P. aeruginosa PAK. With this data we now have experimental evidence to support or revise in silico predictions of open reading frames, operons, promoters, 5’ untranslated regions, regulatory elements, and terminators as well as posit the presence of non-coding and small RNA species for the whole phage at once. With experimental evidence identifying and defining these features at each of the various stages of infection, we are able to construct a transcription level narrative for each phage that powerfully describes them on a fundamental level. Indeed, we have found a number of significant surprises including a large number of antisense RNA molecules that dominate middle transcription of structural and lysis coding sequences in PEV2 presumably silencing all of late translation until late protein expression is triggered. At the same time, unlike any other tailed virus that we are aware of, phage 14-1 appears to not differentially express its genome on a transcriptional level, and instead uses a translation level system to express gene features associated with converting the cell to a phage metabolism during early infection while expressing structural coding sequences in late infection. Similarly, this technology enables looking at the impact of phage infection on the relative abundance of host transcripts during infection. Having data for so many fundamentally different phage types, all infecting the same host, also provides the unique ability to distinguish phage-mediated manipulations of host transcription (by looking for effects that are specific to each phage), from host mediated responses to phage infection(establishing effects that are common to each studied infection). Chapters 3-9 detail and discuss the distinct ways in which the individual phages manipulate between 7% and 11% of the gene features of their hosts to their benefit. demonstrating for the first time that not only does this phenomenon exist but that it is diverse and a key feature of phage infection. At the same time, we have also found that 5.5% of gene features in P. aeruginosa PAO1 are differentially expressed by the host in response to each of the phage infections studied here. Interestingly, this includes the pqsABCDE-phnAB operons of P. aeruginosa PAO1 as well as their PhrS sRNA regulator. As the PQS quorum sensing system has been previously demonstrated to control a stress response encouraging metabolic dormancy, and as we have found phiKMV-like phage LUZ2 is only able to infect PA14 when PQS is knocked out, this may indicate the discovery of a novel micro-colony wide phage defense system. This research not only contributes significantly to our fundamental understanding of Pseudomonas phages (doubling the number of known transcription regulation mechanisms of the host), it also impacts bacteriophage biology and biotechnology in general. It also has direct implications for the safety of phage therapy as it is currently being assessed in medically vulnerable patients. Indeed, our finding that phage LUZ19 upregulates virulence factors responsible for dehydrogenating hydrogen cyanide from glycine not only suggests that the production of HCN might be a direct concern, but that host virulence factors in general have the potential to interfere.status: publishe