And finally, an incredibly-loud thank you to all my coauthors, especially @contaminatedsci.bsky.social , Greg Priebe, and Matt Schaefers, who mentored me throughout this endeavor. I began this project as a college freshman, and now, years later, I’m ecstatic to see it out in the world!

In our paper, we further comment on the coexistence of different O-antigen phenotypes over a decade, speculate on the clonality of B. dolosa transmission, and more! Please let us know any questions, criticisms, and comments, or just want to speculate about reversions.

Also see related theory by labmate Paul Torrilo,  who also speculates reversion may be masking evolution across human hosts {10.1101/2023.09.14.557751}

Our work shows that repeated, de novo mutation in ordinary genes is a viable strategy to traverse changing environments, and suggests pathogen evolution may be oversimplified without frequent sequencing to resolve reversions.

Through in vitro and murine in vivo models, we find a tradeoff between tissue localization -- O-antigen-absent isolates survive better in immune cells at the expense of lung cells. We speculate spatial localization may be driving our observed phenotypic switching.

This creates a fascinating story, where B. dolosa repeatedly disrupts O-antigen expression in early infection and restores it in chronic, decades-long infection across patients via de novo mutation alone.

To our surprise and initial confusion, the new patients were initially infected with an O-antigen-expressing strain. However, new de novo mutations disrupting the O-antigen-synthesis pathway quickly emerged within each patient, mimicking the outbreak ancestor!

Two new recently infected patients allowed us to address this question. We sequenced 931 B. dolosa isolates from the first 38 months of their infections and the lung autopsy of the chronically-infected (~10 years) suspected source.

This stop codon independently reverted in 9/14 studied patients across years of infection, leading to a new hypothesis: does O-antigen absence facilitate transmission at the expense of chronic infection?

Back in 2011, @contaminatedsci.bsky.social and Michel et al. followed a Boston-area outbreak and discovered that the founding B. dolosa strain contained a premature stop codon that abrogated its O-antigen, a lipopolysaccharide sugar chain that mediates immune recognition. {10.1038/ng.997}

Here, we trace how an ordinary, non-hypermutable pathway is toggled on-and-off throughout an outbreak of Burkholderia dolosa, a rare pathogen that infects people with cystic fibrosis.