Thalhammer, K. O. | bioRxiv Science

bioRxiv2026

Redox-dependent lipophilicity of phenazine metabolites is modulated by intramolecular hydrogen bonds and controls their biological distribution

Phenazines are redox-active microbial metabolites produced and secreted in diverse ecological contexts from soils to chronic infections. In these disparate environments phenazines can function variously as antibiotics, extracellular electron shuttles, and nutrient scavengers. Key to understanding the impact of these functions is a robust expectation of phenazine retention or diffusion in a given context. But predicting phenazine fate and transport is difficult because of the chemical complexity of their local microenvironments. To address this challenge, we measured the octanol water distribution coefficient (LogD) as a proxy for lipophilicity of three naturally occurring phenazines produced by the opportunistic pathogen Pseudomonas aeruginosa: phenazine-1-carboxylic acid, phenazine-1-carboxamide, and pyocyanin. We investigated the behavior of both oxidized and reduced forms of these phenazines across broad ionic strength and pH conditions. While the ionic context exerts only small effects, the pH and redox state contribute strongly and independently to changes in phenazine lipophilicity. The pH trends are expected, but the observed redox dependence is generally missed by existing lipophilicity calculation methods. Additional LogD measurements with 1-hydroxyphenazine and unsubstituted phenazine, together with density functional theory modeling of phenazines in their reduced and oxidized forms, reveal that intramolecular hydrogen bonding contributes significantly to the increased lipophilicity of reduced phenazines that possess H-bond accepting substituents in the 1-position. These results explain phenazine behavior in a biological context: redox state alone significantly alters retention of pyocyanin in planktonic P. aeruginosa cells, with the reduced species being predominantly retained by membranes. We propose that the modulation of phenazine lipophilicity in response to the local redox environment has evolved to give a competitive advantage to bacteria by retaining or dispersing these bioactive molecules. Beyond improving our understanding of natural phenazine fate in diverse microbial contexts, our results emphasize an oft-overlooked theme relevant to rational drug and electrochemical shuttle design: redox state matters.

biochemistry↗

bioRxiv2026

Discovery of a phenazine thiol conjugase from sparse data using genome-informed machine learning

Machine learning has enabled powerful biological discoveries using models trained on large datasets. However, for many important biological questions, such as identifying enzymes that transform understudied substrates, sparsity of training data is often a major bottleneck. Here, using phenazine natural products as a case study, we show that integrating genome-informed data augmentation with contrastive learning in protein language space enables identification of phenazine-interacting proteins starting from only 14 known phenazine modifying sequences. Applying this framework led to the discovery of PTC (Phenazine-Thiol Conjugase), the first enzyme known to catalyze phenazine thioconjugation, a phenazine modification reaction long observed but previously presumed to occur only through non-enzymatic chemistry. In silico simulation and experimental measurements demonstrate that PTC binds to both phenazine and glutathione as substrates. Recombinant expression and biochemical characterization reveal that PTC promotes glutathione-dependent modification of phenazines, yielding distinct reaction outcomes that depend on substrate identity. Although thiol-conjugated phenazine products exhibit reduced toxicity to bacterial cells, deletion of the gene encoding PTC does not confer a strong fitness disadvantage, illustrating how direct learning of sequences can uncover relevant enzymes that might evade phenotype-based genetic screens. Together, these results demonstrate that coupling comparative genomics with protein machine learning can convert "small data" typically outside the scope of machine learning into actionable predictive power, thereby facilitating enzyme discovery. SignificanceMachine learning excels when large, well-labeled datasets are available, yet many biologically important problems lack sufficient experimental data to support such approaches to discovery. This limitation is particularly acute for identifying enzymes acting on rare or understudied substrates. Here, we show that genomic organization can be leveraged as an additional source of biological information to address data sparsity. Starting with only 14 enzymes experimentally shown to modify phenazines, we developed a model identifying phenazine-interacting enzymes by integrating genome-informed data augmentation with protein machine learning. Guided by the model, we discovered the first enzyme known to catalyze thioconjugation modifications of phenazines, demonstrating a simple yet powerful strategy for extracting predictive insight from sparse biological knowledge.

microbiology↗