UC San Diego

Protein-protein interactions (PPIs) are ubiquitous in Nature and found in all domains of life. They mediate signal transduction, muscle contraction, electron transport and metabolic pathway processing, among many other processes, and the molecular underpinnings of PPIs remain a rich source of scientific exploration. One model system used to study PPIs is fatty acid biosynthesis (FAS) from Escherichia coli. In E. coli FAS each enzyme of the pathway is a discrete unit to which cargo is delivered by a central acyl carrier protein (EcACP). EcACP tethers the growing acyl chain, sequesters the chain within its alpha helical core to protect it from hydrolysis and cross reactivity, finds the correct partner enzyme, delivers the cargo to the correct partner enzyme and then finds the next partner enzyme in the cycle, all while navigating the crowded milieu of the cytoplasm. During E. coli FAS, EcACP engages in nine different PPIs, and has been implicated in at least 23 additional PPIs with enzymes and proteins from other metabolic pathways in E. coli alone. This wide range of PPIs somehow allow EcACP to deliver substrates to partner enzymes with exquisite selectivity and specificity.While numerous biophysical and structural tools such as isothermal titration calorimetry or X-ray crystallography exist to characterize PPIs, progress in the field has been hindered by a lack of universal, or rapid screen. This dissertation focuses on the development of four different fluorescent-based assays for the rapid visualization of PPIs between EcACP and its numerous partner enzymes, and for the discovery of new carrier protein interactions in previously uncultured marine bacteria. First, the utility of a 4-dimethylnaphthalene (4-DMN) pantetheine probe to monitor thirteen different EcACP PPIs was expanded. 4-DMN-pantetheine was appended to EcACP to generate 4-DMN-EcACP and it was shown that when a partner enzyme interacted with 4-DMN-EcACP, the 4-DMN cargo from EcACP was chain flipped into the active site of the partner enzyme, which led to an increase in fluorescence intensity. 4-DMN-EcACP was used to visualize PPIs between EcACP and enzymes from FAS, biotin and Lipid A biosynthesis, iron-sulfur cluster formation and chromosome trafficking to demonstrate the wide applicability of this fluorescent tool. It was also shown that this fluorescent tool could be used to detect the presence of active site inhibitors in partner enzymes. Next, in silico screening methods were used to identify a potential inhibitor of the PPI between EcACP and a dehydratase from E. coli FAS, FabA. Following identification of a lead compound, suramin, the inhibitory effect of suramin was tested in vitro using the previously developed solvatochromic assay, and the inhibition of suramin was confirmed using a secondary, crosslinking based assay. This work provided the basis for a major R01 grant to the NIH and demonstrated how detailed structural knowledge of these complexes can guide antibiotic discovery efforts and the development of a high-throughput screening program to be undertaken with collaborators at Sanford-Burnham-Prebys. The utility of 4-DMN-EcACP was also demonstrated for mapping residues on the surface of two partner enzymes from E. coli FAS, FabF and FabI. Previous structural methods had identified residues on the surface of FabF and FabI as being critical to the EcACP interaction with each enzyme, and each of these residues were systematically mutated to either alanine or a charge swap mutation. The effect of these mutations on the PPIs between EcACP and FabF and EcACP and FabI was studied using an adapted 4-DMN-EcACP assay. It was found that some point mutations could effectively destroy the PPI, while others left the PPI relatively unchanged. This work is significant in future PPI engineering efforts in being able to rapidly identify which residues on the surface of EcACP partner enzymes are most critical to the PPI. Finally, the same 4-DMN-pantetheine fluorescent tool was used to label individual bacterium in vivo from the microbiome of a tunicate. Sorting of these fluorescently labeled bacteria, followed by single-cell genomics and genome mining revealed that bacteria that up took the 4-DMN-pantetheine dye had a larger number of active secondary metabolite pathways compared to untreated samples. This study provides a workflow for the future discovery of secondary metabolite pathways from uncultured bacteria that might otherwise be overlooked by more traditional genome mining methods.