Affinity purification mass spectrometry (AP-MS) is a technique that allows for the detection and identification of protein-protein interactions (PPIs) present for a specific protein. PPIs provide insight into the pathways that a protein is involved in, which can be used to determine a proteins’ biological function. Previous work in the Backus Lab established a proteomic platform called neo-N-terminal SP3 pyridine carboxaldehyde enrichment (N-SPyCE), which enriches and labels protein neo-N-termini. This method revealed that during T cell stimulation, polypyrimidine tract binding protein 1 (PTBP1) is proteolytically cleaved into various proteoforms. PTBP1 contains various RNA recognition motif (RRM) domains that classify it as an RNA binding protein (RBP). The proteoforms of interest for my project are the full length (FL-PTBP1), one with a disrupted nuclear localization sequence (S35-PTBP1), and two proteoforms possessing only one half of PTBP1’s four RNA recognition motif (RRM) domains (RRM12-PTBP1, and RRM34-PTBP1). I hypothesize the formation of S35-PTBP1 induces a subcellular relocalization of the protein, thus affecting the interactome of the cleaved protein. Furthermore, I hypothesize that the formation of RRM12-PTBP1 and RRM34-PTBP1 alters the proteins ability to bind certain RNA and RBPs, similarly impacting its interactome. My project highlights a biological application of mass-spectrometry for determining the RRM domain dependent interactome of an RBP.

Peptide therapeutics, including clinically important compounds such as cyclosporin A and vancomycin, play significant roles in treating autoimmune diseases and bacterial infections. However, the chemical synthesis of complex cyclic peptides remains challenging, often limited by poor regioselectivity and stereoselectivity. Biocatalysis, therefore, emerges as a promising alternative, using enzymes to perform precise, selective transformations under mild conditions. In this study, we investigate a recently identified family of fungal enzymes, Domain of unknown function 3328 (DUF3328), which catalyze diverse oxidative transformations in ribosomally synthesized and post-translationally modified peptide biosynthetic pathways. To uncover new DUF3328 activities, we leverage genome mining to identify candidates associated with peptide substrates in biosynthetic gene clusters and characterize their activity in vitro. Our results reveal a previously unreported transformation in which a DUF3328 enzyme catalyzes oxidation of the serine side chain hydroxyl group, forming an aldehyde at the peptide C terminus. This modification generates an electrophilic handle not found in canonical amino acids, enabling selective chemical derivatization and bioconjugation. This work expands the functional repertoire of DUF3328 enzymes and shows their potential as biocatalytic tools for site-selective peptide modification, providing new strategies for peptide functionalization and therapeutic development.

Recent advances in protein structure prediction, particularly AlphaFold2, have achieved high accuracy but rely primarily on evolutionary sequence information. Deep mutational scanning (DMS) provides complementary experimental data by measuring how mutations affect protein function. Previous work with DMS-Fold showed that incorporating thermal stability data improves structure prediction, but it is unclear whether this approach generalizes to other types of experimental measurements. This project investigates whether DMS-Fold remains effective when applied to protein fitness scores. DMS-Fold was evaluated using experimental fitness data from the Human Domainome 1 library derived from abundance protein fragment complementation assays, as well as simulated fitness scores in ESgenerated with ESM1v. Model performance was assessed by comparing predicted structures to experimentally determined structures using RMSD and TM-score metrics. Ultimately, DMS-Fold improved structure prediction accuracy for 77% of proteins using experimental fitness data and 74% using simulated fitness data. These findings demonstrate that DMS-Fold can successfully leverage fitness-based measurements, not just stability data. This work expands the applicability of DMS-guided structure prediction and highlights the potential of integrating diverse experimental data types to improve protein modeling.

Organocatalysis, using catalytic activation to increase efficiency of a chemical reaction, has gained much traction in green chemistry methods and synthetic labs everywhere. Asymmetric catalysis particularly is integral in producing enantiopure compounds, especially for natural products. The Mitsunobu reaction is renowned for its stereoselectivity in converting a primary or secondary alcohol functional group into an ester. It is driven forward by a nucleophile—usually an acid—and two redox reagents: an azodicarboxylate for oxidation and a phosphine for reduction. This reaction is a more reliably useful stereospecific method in organic synthesis. However, there are a few issues regarding its efficiency and atom economy. This project is centered on creating a library of CarvoPhos catalysts to replace the most commonly chosen triphenylphosphine and on scaling up production of each derivative. The goal is to have a racemic alcohol mixture and an acid nucleophile undergo a catalytic, redox-driven reaction to create enantioenriched alcohol and nucleophilic products. This generation of phosphine catalysts, deemed CarvoPhos catalysts derived from the (R)-(−)-Carvone species, have been developed to improve upon our lab’s previously developed hydroxyproline-derived (HypPhos) catalysts. The building blocks and final products are more electron rich or sterically bulky compared to HypPhos catalysts and thus potentially more reactive for future experimentation—such as in a variant of the Mitsunobu reaction.

We report a novel alternating copolymerization of different epoxide monomers and N-carboxyanhydride to directly synthesize polyesters with hydrogen-bonding motifs on their main chain, making previously inaccesssible poly(amino ester)s. This directly incorporates secondary amine moieties into the polymer backbone, improving upon previous conventions that limit preparation of functionalized polyesters to side chain modifications.