Pseudomonas aeruginosa is a primary bacterial model to study cooperative behaviors because it yields exoproducts such as siderophores and exoproteases that act as public goods and can be exploited by selfish nonproducers behaving as social cheaters. Iron-limited growth medium, mainly casamino acids medium supplemented with transferrin, is typically used to isolate and study nonproducer mutants of the siderophore pyoverdine. However, using a protein as the iron chelator could inadvertently select mutants unable to produce exoproteases, since these enzymes can degrade the transferrin to facilitate iron release. Here we investigated the evolutionary dynamics of pyoverdine and exoprotease production in media in which iron was limited by using either transferrin or a cation chelating resin. We show that concomitant loss of pyoverdine and exoprotease production readily develops in media containing transferrin, whereas only pyoverdine loss emerges in medium treated with the resin. Characterization of exoprotease- and pyoverdine-less mutants revealed loss in motility, different mutations, and large genome deletions (13–33 kb) including Quorum Sensing (lasR, rsal, and lasI) and flagellar genes. Our work shows that using transferrin as an iron chelator imposes simultaneous selective pressure for the loss of pyoverdine and exoprotease production. The unintended effect of transferrin uncovered by our experiments can help to inform the design of similar studies.

Pseudomonas aeruginosa is the main bacterial model to study cooperative behaviors, since it yields exoproducts such as exoproteases and siderophores that act as public goods and can be exploited by selfish non-producers that behave as social cheaters. Non-producers of the siderophore pyoverdine are typically isolated in media with low free iron, mainly casamino acids medium supplemented with transferrin. Nevertheless, using a protein as the iron chelator could additionally select mutants unable to produce exoproteases that degrade the transferrin to facilitate iron release. Here, we investigated the dynamics of pyoverdine and exoprotease production in media in which iron was limited by using either transferrin or a cation chelating resin. Our experiments show that concomitant loss of pyoverdine and exoprotease production readily develops in media with transferrin whereas only lack of pyoverdine emerges in medium treated with the resin. Genomic characterization of the exoprotease- and pyoverdine-less mutants revealed large deletions (13 to 33 Kb) including Quorum Sensing (lasR, rsal and lasl) and flagellar genes. Complementation experiments, PCR and motility tests confirmed the deletions. Our work shows that using transferrin as an iron chelator imposes simultaneous selective pressure for the loss of pyoverdine and exoprotease production. The unintended effect of transferrin observed in our experiment settings can help revisiting or informing the design of similar studies.