In partial fulfillment of the requirements for the degree of

Doctor of Philosophy in Biology

In the

School of Biological Sciences

Isaac Estrada

Will defend his dissertation

Defining and Engineering R-pyocins as Targeted Antimicrobials against Multidrug-Resistant Pseudomonas aeruginosa

23rd, November, 2026

3pm

Microsoft Teams meeting

Join: https://teams.microsoft.com/meet/242094461672285?p=tkn3wkcQrcQ9LBhWbE

Meeting ID: 242 094 461 672 285

Passcode: Jv9pr6TC

Thesis Advisor:

Steve Diggle, Ph.D.

School of Biological Sciences

Georgia Institute of Technology

Committee Members:

Sam Brown, Ph.D.

School of Biological Sciences

Georgia Institute of Technology

Marvin Whiteley, Ph.D.

School of Biological Sciences

Georgia Institute of Technology

Joanna Goldberg, Ph.D.

Department of Pediatrics

Division of Pulmonary, Asthma, Cystic Fibrosis, and Sleep

Emory University School of Medicine

Kendra Rumbaugh, Ph.D.

Department of Surgery

Texas Tech University Health Sciences Center

ABSTRACT:

Defining and Engineering R-pyocins as Targeted Antimicrobials

against Multidrug-Resistant Pseudomonas aeruginosa

Isaac Estrada

298 Pages

Directed by Dr. Stephen P. Diggle

Pseudomonas aeruginosa is one of the most difficult organisms to treat, and the high-risk clonal lineages disseminated worldwide leave few treatment options. R-type pyocins are narrow-spectrum, phage tail-like bacteriocins produced by P. aeruginosa to kill other strains of the same species and are currently classified into three main subtypes (R1, R2, and R5). They have been proposed as therapeutic agents for many years, and their mechanism of action is well understood. Their full killing specificity, their activity in infected tissue, and the purity of current preparations, however, remain poorly characterized.

In this work, we evaluated R-pyocin distribution and killing specificity across a clinical biobank of high-risk strains. Screening 1,086 genomes and the 100-strain high-risk clinical panel, we find that O-specific antigen serotype predicts R-pyocin subtype independent of clonal background, and that killing spectrum tracks tail fiber allelic variation rather than the subtype classification currently in use.

To determine how resistance to R-pyocins arises and how far it extends, we subjected laboratory strains and a pan-drug-resistant wound isolate to sustained R-pyocin selection. Resistance evolved rapidly, and sequenced resistant variants were cross-resistant to all subtypes. We found that resistance is gained via a large chromosomal deletion that removed galU, which is essential for the formation of the R-pyocin receptor. The same deleted region carries the mexXY efflux operon together with virulence and biofilm loci, so surviving mutants are resensitized to aminoglycosides, unable to form biofilms, and avirulent. Resistance to R-pyocins is therefore itself a therapeutic advantage.

To evaluate how much activity is retained from in vitro assays to in vivo infection, we tested fixed-dose preparations across increasingly complex systems: broth and wound media, a collagen biofilm matrix, explanted infected wound tissue, and a murine pulmonary infection. We found that activity is limited by access to the bacteria and contact time rather than by intrinsic potency.

Lastly, to determine whether R-pyocin preparations are cytotoxic to mammalian cells, and whether any cytotoxicity is a property of the particle itself, we applied nine independently produced preparations to three mammalian cell lines. Toxicity tracked the preparation rather than the subtype and could not be attributed to any property of the particle, indicating that current preparation methods cannot reliably produce non-cytotoxic R-pyocins. We therefore engineered a hexahistidine tag into the collar protein encoded by PA0615 and captured assembled particles from crude lysate in a single affinity chromatography step.

Overall, this work demonstrates that the limits on R-pyocin therapy are set by specificity, delivery, and purity of preparation rather than by the killing machinery itself. These findings, along with the affinity-tagged particle, provide the basis for TAILSTRIKE, a proposed plasmid-encoded chassis that separates the killing machinery from an exchangeable targeting module, allowing narrow-spectrum antimicrobials to be customized to a chosen target.