3D Genomics Reveals How Phage PAK_P3 Reorganizes Pseudomonas aeruginosa During Infection
A bacteriophage infection is usually described as a sequence of molecular events: attachment, genome injection, takeover of bacterial metabolism, viral replication, assembly and finally cell lysis. But inside the infected bacterium, another process is unfolding at the same time. Viral and bacterial DNA are physically reorganizing in three dimensions.
A study published in Science Advances on 26 August 2026 now provides an unusually detailed view of this hidden choreography. Researchers from Institut Pasteur, Université Paris Cité, Sorbonne Université, CNRS and the University of Oxford followed the virulent bacteriophage PAK_P3 during infection of Pseudomonas aeruginosa and reconstructed how both the viral genome and the bacterial chromosome change their spatial organization over the course of only a few minutes.
The work was led by Amaury Bignaud and colleagues, with Martial Marbouty as corresponding author and contributions from Quentin Lamy-Besnier, Devon E. Conti, Agnès Thierry, Fabien Girard, Pauline Misson, Romain Koszul and Laurent Debarbieux. The team combined time-resolved chromosome conformation capture, or Hi-C, with transcriptomic information to observe the infection as a dynamic structural process rather than simply measuring which viral genes were expressed.
PAK_P3 provides an especially useful model for such an experiment. It is a virulent double-stranded DNA phage with a linear genome of approximately 88 kilobases that infects P. aeruginosa. It belongs to the Nankokuvirus group and depends on the bacterial RNA polymerase for transcription. Its eclipse period lasts roughly 12 minutes and its complete infection cycle approximately 20 minutes, making it one of the faster known contractile-tailed dsDNA phages.
The researchers sampled infected bacteria at 0, 3, 5, 7, 10, 13 and 16 minutes after phage addition. This allowed them to follow the transition from genome entry through host takeover, phage DNA replication and eventual packaging of newly synthesized genomes into viral particles.
Before infection, the P. aeruginosa chromosome displayed a highly organized architecture. The researchers identified 37 chromosome interaction domains, or CIDs, whose boundaries were associated with highly transcribed regions. At smaller scales, strongly transcribed genes generated characteristic transcription-induced domains, while the chromosome also maintained long-range contacts between its two replichores.
PAK_P3 progressively dismantled much of this local organization.
As infection advanced, the bacterial chromosome became increasingly decondensed. Transcription-induced domains disappeared, local chromosome boundaries weakened and structural features such as loops and stripes were progressively lost. By around 13 minutes after infection, much of the local organization visible in uninfected bacteria had collapsed. This structural deterioration occurred alongside the previously observed shutdown of host transcription as PAK_P3 redirected the bacterial transcriptional machinery toward expression of its own genome.
Yet the host chromosome was not simply destroyed or completely randomized. Long-range replichore contacts remained detectable, suggesting that the global architecture maintained by the bacterial SMC-ScpAB machinery persisted even as local transcription-dependent folding disappeared. The infection therefore appears to selectively erase one level of chromosome organization while preserving another.
The viral genome underwent an equally dramatic transformation.
Inside the phage particle, PAK_P3 DNA is extremely compact. Hi-C measurements performed directly on virions confirmed this highly constrained state and supported a linear genome architecture with direct terminal repeats located around 63 kb. Once the DNA entered the bacterial cytoplasm, however, its spatial organization changed rapidly.
The genome decondensed during the first minutes of infection, reaching its most open state approximately five to ten minutes after entry. It then began to compact again. By 16 minutes, its contact profile was approaching the initial state observed inside the virion, consistent with newly synthesized genomes being packaged into nascent phage particles.
This restructuring was not random. Different regions of the PAK_P3 genome formed spatial domains that corresponded strikingly with the phage's transcriptional programme. Regions containing early, middle and late genes displayed changing patterns of medium-range contacts as their respective transcriptional programmes became active.
One particularly notable structure appeared only three minutes after infection. A roughly 14-kb region containing part of the early gene cluster separated spatially from the rest of the viral genome and formed what the researchers describe as a loose looping structure.
The observation suggests that phage chromosomes may possess considerably more sophisticated spatial organization than previously assumed. Instead of viral DNA simply floating inside the bacterial cytoplasm while genes are sequentially transcribed, genome folding itself could contribute to coordinating the infection programme.
This raises another question: what physically organizes the phage genome?
The authors discuss several possibilities, including phage-encoded nucleoid-associated protein-like factors, non-coding RNAs, DNA supercoiling, plectoneme formation or mechanisms analogous to loop extrusion. Among proteins previously identified in PAK_P3 viral particles, they highlight Gp160, a small basic predicted DNA-binding protein with characteristics reminiscent of bacterial nucleoid-associated proteins. Its involvement remains hypothetical, however, and the authors emphasize that mutant phages or chromatin immunoprecipitation experiments would be required to establish its role.
PAK_P3 DNA also appeared to circularize during infection.
Between approximately three and ten minutes, Hi-C detected increasing interactions between the two ends of the viral genome. The authors interpret this as evidence that the linear genome circularizes, probably through recombination involving its direct terminal repeats. The signal subsequently decreased around 13 minutes as replication progressed and genomes began entering newly assembled capsids.
The replication pattern itself provided another clue. Between seven and ten minutes, sequencing coverage changed in a way compatible with bidirectional replication from a single region of the circular genome, supporting a theta-like replication mechanism. The authors nevertheless caution that theta replication could potentially coexist with rolling-circle replication, as has been observed in other phage systems.
Perhaps the most unexpected discovery came when the researchers examined where the phage genome was positioned relative to the bacterial chromosome.
The two genomes did not interact uniformly.
Contacts between PAK_P3 DNA and the P. aeruginosa chromosome increased during the early phase of infection, peaking around seven minutes before declining as viral genomes were progressively packaged. More importantly, specific regions of the bacterial chromosome consistently showed enriched or depleted interactions with phage DNA.
To determine whether this simply reflected the general accessibility of bacterial DNA, the researchers compared PAK_P3 with the multicopy plasmid pJN105. The two mobile genetic elements behaved very differently.
At five minutes after infection, PAK_P3 contacts were negatively correlated with both local bacterial Hi-C interactions and transcriptional activity. The correlations were −0.4257 with local chromosome contacts and −0.2573 with transcription. The plasmid showed the opposite pattern, with positive correlations of +0.6723 and +0.4123 respectively.
In other words, PAK_P3 preferentially associated with relatively weakly transcribed regions of the bacterial chromosome rather than the transcriptionally active areas favored by the plasmid.
The authors propose a provocative spatial interpretation. Because transcriptionally active bacterial chromatin has been reported to localize preferentially toward the nucleoid periphery, the observed interaction pattern could mean that PAK_P3 positions its genome more centrally within the bacterial nucleoid. Such positioning might help the phage access metabolites and cellular machinery or potentially avoid some host defensive mechanisms. At present, however, this remains a hypothesis: direct imaging of PAK_P3 DNA inside infected bacteria will be needed to determine its actual intracellular location.
The model presented in Figure 5 of the paper brings these observations together. A tightly packed linear viral genome enters the bacterium and rapidly decompacts. Early regions acquire a particular spatial organization, the genome circularizes, and replication begins while the bacterial chromosome simultaneously loses much of its transcription-associated local structure. The viral genome may position itself within particular regions of the nucleoid before newly replicated copies are eventually condensed again and packaged into virions.
For phage biology, these observations introduce an additional dimension to host takeover. A successful infection is not simply a matter of which genes are present and when they are expressed. The physical arrangement of those genes inside the cell may itself help regulate transcription, replication and access to bacterial resources.
This is particularly relevant for P. aeruginosa, an opportunistic pathogen of major interest for phage therapy. Therapeutic phage selection usually emphasizes characteristics such as bacterial host range, adsorption, replication efficiency, burst size, resistance development and the absence of undesirable genes. Studies such as this suggest that another layer may eventually need to be considered: how efficiently a phage reorganizes its genome and establishes an intracellular environment favorable to replication after entry into the bacterial cell.
The therapeutic implications should nevertheless be kept in perspective. The study did not test whether the observed genome structures improve treatment outcomes, nor did it compare successful and unsuccessful therapeutic phages. It also relied on infection at a high multiplicity of infection of 25 under controlled laboratory conditions. The proposed central positioning of PAK_P3 within the bacterial nucleoid and the involvement of candidate structural proteins such as Gp160 remain hypotheses requiring direct experimental validation.
What the study demonstrates much more firmly is that a phage genome is not a passive strand of DNA waiting to be transcribed. During infection, PAK_P3 undergoes rapid and coordinated three-dimensional transformations closely associated with its gene-expression programme, while simultaneously reshaping the chromosome of its host and establishing non-random spatial relationships with bacterial DNA.
That finding adds a new layer to the molecular arms race between bacteriophages and bacteria. Phages hijack enzymes, metabolism and transcription, but they may also exploit physical space inside the cell.
For phage therapy, understanding this intracellular choreography could ultimately help explain why some phages are exceptionally efficient at taking control of a pathogen while others that successfully attach and inject their DNA nevertheless fail to complete infection.
Sources :
Science Advances — Amaury Bignaud et al., “Bacteriophage PAK_P3 genome structuration and dynamics during infection of Pseudomonas aeruginosa reveal specific interaction patterns,” Science Advances 12, eaef2512, published 26 August 2026. DOI:
https://doi.org/10.1126/sciadv.aef2512
Full article:
https://www.science.org/doi/10.1126/sciadv.aef2512
The study was supported by the French ANR PhaStGut project, the European Research Council, the ANR COMET chair and an ENS fellowship from the French Ministry of Higher Education, Research and Innovation. The authors report no competing interests.
The article is distributed under a Creative Commons Attribution 4.0 (CC BY) licence, which is useful for The Phage Therapy if you want to reuse one of the paper's figures with proper attribution.
.png){kind=logo}
Comments
Post a Comment