Utilizing Chemical Biology and Computational Tools to Understand and Engineer Metabolic Pathways

Within the last decade, efforts have revealed remarkable detail in the machines that regulate primary metabolism with salient examples including the ribosome, protein folding and most recently the spliceosome. Our laboratory has taken interest in the macromolecular modular machines that construct fatty acids, polyketides and non-ribosomal peptides. The machines that prepare these classes of secondary metabolites share a common choreographed method. Small molecules such as acetyl-coenzyme A (CoA) and malonyl-CoA are assembled sequentially to form complex natural products, which have uses ranging from commodity chemicals to therapeutics.Although the starter units of these synthases are free floating molecules, their intermediates are tethered to acyl carrier protein (ACP). The ACP carries this cargo along the assemblyline of modifying enzymes (i.e., ketoreductase (KR), dehydratase (DH), enoyl-acyl carrier protein reductase (ER)) until it is release by a thioesterase (TE) domain. While this modular machinery appears ideal for metabolic engineering, many of the leading efforts, such as domain swapping, have been met with limited success. This arises, in part, from our lack in understanding the protein • protein interactions that guide the processivity between the CP and its associated partner domains (KS, KR, ER, DH and TE). Unfortunately, structural studies on these systems continue to pose challenges due to the transient nature of these interactions. The first couple chapters describes how new small molecule inhibitors were discovered to inhibit the ER in the fatty acid biosynthesis pathway in Plasmodium falciparum, the causitive agent responsible for malaria. These chapters describe in detail the collaborative work between in silico and in vitro methods to discover novel small molecules that have been from repositioned and commercial small libraries. The next chapter discuses an extension of the our laboratory’s work dedicated to developing a suite of tools to study the interactivity between CP and associated partner domains. We first sought to leverage our previous work with covalent mechanism-based inactivators with the KS, TE and DH domains, and apply this to FabI. Because the reaction catalyzed by FabI is cofactor dependent and does not involve any active site residues in a covalent manner, designing a mechanistic based inactivator that covalently crosslinks AcpP and FabI poses a challenge. We hypothesized that if we attached a tight non-covalent FabI binder to AcpP via our chemoenzymatic methods, the binding affinity may be high enough to stabilize the complex for structural studies. The study extends our collection of chemoenzymatic AcpP tools with the first “inhibitor-based non-covalent probe” inspired by a well-characterized FabI tight-binding small molecule inhibitor. We synthesized a pantetheinamide derivative of TCL, appended it to AcpP and used this to study FabI-AcpP interactions. This approach shows that both protein-protein interactions and protein-substrate interactions are important for productive catalysis. We envision the use of this novel probe for structural characterization of the AcpP-FabI interaction, which has yet to be resolved in greater detail. The last chapter describes an expansion and discovery of new substrates for the the promiscuous 4’-phosphopantetheinyl transferase (PPTase) and acyl carrier protein hydrolase (AcpH) by artificial intelligence. With the development of conjugated-protein therapeutics over the last two decades, the need for robust protein labeling methods has intensified. Unlike conventional large fusion tag proteins, which can interfere with protein activity, a peptide tag only adds 8-20 amino acids to either terminus of a protein. Previous studies demonstrate that the post-translational enzyme, phosphopantetheinyl transferases (PPTase), can label YbbR, a short 11 amino acid sequence. More recently, our lab expanded on this system by demonstrating that the acyl carrier protein hydrolase (AcpH) can unlabel short peptide substrates. Our research highlights the discovery of peptide sequences that can be labeled selectively by either Sfp (PPTase 1) or AcpS (PPTase 2). and unlabeled by AcpH to obtain an orthogonal and reversible labeling system. Additionally, we have developed a statistical method, called Peptide Optimization with Optimal Learning (POOL), for efficiently discovering minimal peptide substrates from our high-throughput peptide screens. By leveraging post-translational modifying enzymes, our project will allow for site-selective functionalization