Glycosylation-associated dysregulation of pyocyanin production in Pseudomonas aeruginosa: Implications for quorum sensing regulation

Pseudomonas aeruginosa (P. aeruginosa) is an important opportunistic pathogen associated with high mortality in pneumonia, sepsis, and cystic fibrosis. Lending to its ability to cause severe disease and death is its arsenal of virulence factors and host evasion tactics. In addition to various other regulatory systems, many of P. aeruginosa’s virulence factors are regulated by a population density dependent regulatory network known as quorum sensing (QS). Many regulatory systems are impacted by post-translational modifications of proteins. An underexplored physiological aspect of P. aeruginosa is its ability to glycosylate proteins and the subsequent impact of glycosylation on P. aeruginosa physiology and behavior. The goal of this study was to determine whether P. aeruginosa QS is regulated by glycosylation. Here we demonstrate that disruption of glycosylation dysregulates QS phenotypes, notably pyocyanin production, in P. aeruginosa PAO1. In this study, it was initially observed that deletion of the P. aeruginosa neuraminidase, PaNA, caused an increased production of pyocyanin in LB-Lennox broth compared to wildtype bacteria at identical population densities. To confirm that the increased pyocyanin production was due to QS, we performed induction experiments using 10% cell-free media harvested from overnight cultures. To determine whether the QS phenotype observed is specific to pseudaminic acid, the target of PaNA, or if it is a reflection of global changes in glycosylation, we measured QS in a library of mutant bacteria generated in an MPAO1 background containing transposon insertions in various glycosyl-associated enzymes. The pattern of dysregulated QS held true in these mutant strains as well. Overall these data indicate that in P. aeruginosa, glycosylation is an important determinant of QS.


Introduction
Signal transduction occurs across the bacterial membrane, as does 74 glycosylation [13,14]. Glycosylation is the covalent attachment of a glycan, or 75 carbohydrate chain, to a substrate such as a protein or a lipid [15,16]. While protein 76 glycosylation was previously viewed as a eukaryotic process, it is now recognized as a possesses an enzyme initially identified as a neuraminidase, but following X-ray 103 crystallography and in silico docking experiments, the enzyme was determined to be a 104 pseudaminidase [45,46]. This enzyme, PaNA, is encoded at locus PA2794 and was 105 initially of interest as a virulence factor. However, its role in the biology of the bacteria 106 remains uncharacterized [36,44]. 107 A strain of P. aeruginosa PAO1 from which the neuraminidase gene PA2794 has 108 been deleted [44], PAO1Δ2794, exhibited a pronounced over-expression of pyocyanin, 109 which is associated with the PQS arm of QS, compared to the wildtype strain. We 110 therefore hypothesized that the deletion of the neuraminidase resulted in an alteration of 111 the glycosylation of one or more proteins which led to this anomalous phenotype. 112 Indeed, lectin blots revealed a differential pattern of glycosylation between the wildtype 113 strain and the mutant Δ2794. We confirmed our initial observations by using a different true in these mutant strains as well. We show that disruption of PaNA, as well as other 120 glycosyl-associated enzymes, results in a QS phenotype-namely the over-production 121 of pyocyanin-which is decoupled from population density, overall suggesting that 122 bacterial glycosylation is a critical determinant of QS.

190
Disruption of PaNA does not alter growth dynamics, but promotes pyocyanin production 191 We initially observed that PAO1Δ2794, a strain with an allelic deletion of PaNA, 192 exhibits increased pyocyanin production compared to its wildtype counterpart, Pyocyanin production can be induced in the wildtype strains using conditioned media 251 We next asked whether this decoupling of pyocyanin production in the mutant where the mutant is compared to its corresponding wildtype at that timepoint.

271
Importantly, while the behavior is inducible earlier in wildtype strains grown in 272 10% supernatant, these cultures also produced pyocyanin at a higher magnitude than

282
Pyocyanin production relies on MvfR 283 We next asked whether pyocyanin production may be occurring through a 284 pathway other than the canonical PQS system of QS. To address this question, we 285 used an available mutant strain PW2812 which carries a disruption in the MvfR gene.

286
As expected, culture supernatant from the PaNA mutants, wildtypes, or PW2812 was 287 not able to induce pyocyanin production in the mvfR-mutant strain ( Figure 3A).  Figure 4A shows that the lectin-334 binding pattern of fucose-specific Lotus lectin is dynamic over growth in wildtype 335 MPAO1 (n=3). Figure 4B shows that there are differences in the glcosylation pattern of 336 MPAO1 when it is induced to QS by growth in 10% culture supernatant from PW2812.

337
PW2812 is included as a control, but also shows differences from wildtype. Namely a  assessed Las activation with the skim-milk clearance assay ( Figure 5D). Some strains 370 exhibited increased clearance compared with the wildtype strain, while others did not.

371
The assessment of rhl activation by use of the swarming assay showed no difference 372 among any of the strains in the ability to swarm ( Figure 5E). This ability to swarm is represented data are an average of an n=3 where the mutant is compared to its 385 corresponding wildtype at that timepoint. Significance is indicated when *p≤0.05, 386 **p≤0.01, ***p≤0.001, and ****p≤0.0001.

387
Finally, we assessed whether the signal was transferable from some of these 388 glycosyl-associated mutant strains to wildtype strains. We included two strains with 389 disrupted QS-regulator enzymes as a control [33,34]. Culture supernatants were able 390 to induce pyocyanin production above the background wildtypes in naïve culture media ( Figure 6). Even at 24 hours, the pyocyanin produced by induction was detected at a 392 higher magnitude than that of the uninduced wildtypes.

453
The present work shows that disruption of various glycosyl-associated enzymes 454 leads to an over-production of pyocyanin. Pseudomonas aeruginosa leads to an over-production of pyocyanin, suggesting that