Proteins Switched On as Pseudomonas aeruginosa Adapts to Chronic Cystic Fibrosis Lungs

Pseudomonas aeruginosa, a Gram-negative bacterium that thrives everywhere from hospital water systems to contact lens solutions, is one of the most formidable opportunistic pathogens in modern medicine. The World Health Organization has placed it on its high-priority list of bacteria urgently requiring new antibiotics, and for people with cystic fibrosis it is a particularly stubborn adversary. Even in patients receiving more than a year of modern CFTR modulator therapy, P. aeruginosa persists in a substantial cohort, driving chronic airway inflammation and proving harder to monitor once modulator treatment begins. Now, a new proteomic study published in MicrobiologyOpen offers one of the clearest molecular portraits yet of how this pathogen evolves inside the cystic fibrosis lung, revealing a small set of proteins that appear to be positively selected as infection turns chronic.

The research team, led by Siobhán McClean of University College Dublin with collaborators including Joanna Drabinska, Lucia O’Connor and Caoilin McClean, took advantage of a uniquely valuable resource: an international reference panel of 41 fully sequenced P. aeruginosa strains assembled by De Soyza and colleagues. Within that panel sit three independent series of sequential isolates, each recovered from a different person with cystic fibrosis in a geographically distinct region, and each spanning roughly seven years of infection. The German series comprises AA2, AA43 and AA44, with the latter two isolated 7.5 years after the first and shortly before the patient’s death. The Seattle series from a pediatric patient includes AMT 0060-3, recovered when the child was 7.7 years old, and two later strains isolated 7.9 years afterward. A third series, AMT0023-30 and AMT0023-34, was isolated eight years apart from a patient who was only six months old at the first sampling; the later strain carried 68 unique mutations relative to the earlier one.

Previous phenotypic work on these strains had already established a consistent pattern: reduced virulence in the Galleria mellonella acute infection model was the only trait altered in every late isolate, while pyocyanin production, the blue-green pigment that contributes to tissue damage, fell in four of the five late strains. Motility traits such as swarming and swimming also declined in most late isolates, with some functions lost entirely. What remained unknown was whether these convergent phenotypes reflected convergent molecular changes. To find out, the researchers performed label-free quantitative proteomics on all eight strains, using a Bruker TimsTOF Pro mass spectrometer coupled to an Evosep One chromatography system with PASEF acquisition, and analyzed the data in MaxQuant against the PAO1 reference proteome with a 1 percent false discovery rate.

The scale of change within each series was striking. In the AMT0060 series, 138 proteins changed abundance by at least 1.5-fold in one late strain and 166 in the other, with 57 changes shared between the two. In the AA2 series, 78 proteins changed in AA43 and 267 in AA44, though only 30 overlapped between the two late strains. The AMT0023 pair showed 182 altered proteins. Given the enormous diversity of P. aeruginosa, the team specifically searched for proteins altered in two or more independent series, reasoning that such changes would represent conserved adaptation pathways rather than patient-specific quirks.

The result was remarkable. Only 16 proteins in total were altered across all three series, and 11 of them showed a consistent direction of change: increased abundance in the late strains of every series. In nearly every case, these proteins were undetectable in the early isolates, meaning the underlying genes had been switched on during years of colonization. The probability that the same three proteins, PA2572, PA3819 and PA5028, would show increased abundance in all five late strains from three independent early ancestors by chance alone was calculated at 5.06 × 10⁻⁵³, which the authors describe as very strong evidence of positive selection. Even for the eight proteins that appeared in only one late isolate per patient, the probability of such a pattern arising randomly in a genome of roughly 5,570 open reading frames was 1.6 × 10⁻¹⁹.

The identities of these 11 proteins tell a coherent evolutionary story. PA3819 is an outer membrane lipoprotein with a glycine zipper domain, encoded within the AlgU regulon that governs alginate production and membrane stress responses, and previously linked to Toll-like receptor signaling in mucoid strains. PA2572 is an HD-GYP domain two-component response regulator sitting beside a chemotaxis transducer gene; it binds the sensor PA2573, influences ExoS and pyocyanin production, dampens swarming motility, and does not hydrolyze c-di-GMP unlike its two paralogs. PA3702, better known as WspR, is the diguanylate cyclase response regulator of the Wsp surface-sensing system, which responds to cell envelope stress, suppresses flagellar motility and promotes biofilm formation, consistent with the loss of motility widely reported for chronic isolates. PA5028 and PA1462 are cytoplasmic membrane proteins of the ParAB family, partners of the DNA-binding partitioning protein ParB, whose systems regulate cell division and act as global regulators of multiple proteins. CifR, a TetR-family epoxide-responsive repressor controlling the CFTR inhibitory factor Cif, rose 10- to 11-fold in one late isolate of each series, matching earlier observations that CF isolates maintain CifR expression over time. The set is rounded out by PA2551, a probable LysR-family transcriptional regulator that may counteract stress-induced growth slowdown; PA2679, a methyltransferase whose expression rises under hypoxia; PA2883, a membrane protein co-expressed with the c-di-GMP-binding protein MapZ and strongly induced by airway epithelia; PA3084, a hypothetical protein identified as conditionally essential for cardiomyocyte infection; and PA3271, a two-component sensor kinase whose disruption alters virulence genes including pyochelin synthesis, elastase and flagellar proteins.

Notably absent from this list are antibiotic resistance mechanisms, even though several resistance-associated proteins, such as the MexA efflux component, rose dramatically in some late strains. The authors suggest that differing antibiotic regimens among the three patients may explain the lack of a conserved resistance signature. Convergence was also evident at the pathway level rather than the individual protein level: virulence-associated secretion systems and secreted factors declined across multiple strains, and phenazine biosynthesis enzymes such as PhzB, PhzF and PhzG dropped in the AA2 and AMT0023 series but not in AMT0060, indicating that reduced pyocyanin production can arise through multiple regulatory routes, including quorum sensing and quinolone signaling, where LasR and PQS-associated proteins were also reduced.

To test whether the upregulated proteins actually confer a functional advantage, the team compared gene-deletion mutants of PA2572 and PA3819 with the wild-type MPAO1 strain under stresses that mimic the cystic fibrosis lung. Under normal conditions the mutants grew identically to the wild type, with comparable growth kinetics and stationary-phase densities. But the ΔPA3819 mutant showed clearly impaired survival under oxidative stress from 1 mM hydrogen peroxide, and both mutants fared worse under osmotic stress from 0.5 M sodium chloride and elevated temperature of 45°C. Antibiotic susceptibility testing added a further twist: loss of PA2572 increased susceptibility to imipenem, aztreonam and norfloxacin, while loss of PA3819 increased susceptibility to aztreonam alone. Together, these results suggest that the elevated abundance of these proteins in late isolates directly supports bacterial fitness under chronic-infection conditions and may contribute to the enhanced antibiotic resistance that emerges over years of treatment.

Intriguingly, quantitative PCR showed that the increased protein abundance was not mirrored at the transcript level. Expression of PA3819 was actually repressed 3.5-fold in two late strains, and PA2572 was repressed 5-fold in AA44, pointing to translational or post-translational control. Searching the amino acid sequences of the three consistently upregulated proteins revealed multiple predicted phosphorylation, glycosylation and N-myristoylation sites associated with protein stability, and a shared conserved motif resembling tyrosine phosphorylation sites found in bacterial effectors such as Tir of enteropathogenic Escherichia coli and Tarp of Chlamydia trachomatis. Tyrosine phosphorylation is known to regulate bacterial virulence traits, and the authors propose that such modifications may stabilize these proteins during chronic infection. Because earlier comparative studies of sequential isolates relied on less sensitive methods such as two-dimensional electrophoresis, none of these 11 proteins had previously been flagged as consistently increased, underscoring the power of modern high-throughput proteomics. Each of the 11 proteins, the authors conclude, represents a potential target for adjuvant therapies designed to disable the adaptation process itself and prevent acute infection from hardening into lifelong chronic colonization.

Subject of Research: Proteomic analysis of sequential Pseudomonas aeruginosa strains from cystic fibrosis patients to identify proteins under positive selection during chronic infection

Article Title: Proteomic Analysis of Three Independent Series of Sequential Cystic Fibrosis Strains in an International Pseudomonas aeruginosa Reference Panel Indicates Positive Selection in Late Infection Strains

Article References: Drabinska, J., O'Connor, L., McClean, C., & McClean, S. (2026). Proteomic Analysis of Three Independent Series of Sequential Cystic Fibrosis Strains in an International Pseudomonas aeruginosa Reference Panel Indicates Positive Selection in Late Infection Strains. MicrobiologyOpen, 15(5), Article e70417. https://doi.org/10.1002/mbo3.70417

Image Credits: AI Generated

DOI: 10.1002/mbo3.70417

Keywords: Pseudomonas aeruginosa, cystic fibrosis, proteomics, chronic infection, positive selection, bacterial adaptation, virulence, two-component regulators, antibiotic resistance, WspR, hypoxia response, host adaptation