Mimicry of protein secondary structure elements, such as α-helices and β-sheets, using conformationally constrained peptide macrocycles, can be utilized to disrupt native protein-protein and protein-nucleic acid interactions. Although α-helical stapled peptides have been extensively studied as pharmacological probes, the application of β-sheet and β-hairpin mimetics remains comparatively limited. Less is known about the structural and biophysical consequences of β-hairpin macrocyclization in the context of target binding. In this work, we use a poxvirus immune antagonist protein 018 as a template for the structure-based design of β-hairpin mimetic macrocyclic peptides targeting the STAT1 transcription factor. We demonstrate that successive orthogonal cyclizations have additive effects on the thermodynamic and kinetic properties of peptide binding, most notably slowing the dissociation from the target. We elucidate the structural and dynamic consequences of interstrand and head-to-tail cross-linking and propose a kinetic model explaining the gains in target residence. Finally, we highlight the pharmacological potential of these peptides by competitive inhibition of STAT1 binding to its cognate interferon receptor docking site. These data suggest that β-hairpin macrocyclization may represent a general strategy to extend target engagement, with implications for peptidic probe design.

Only a few sesquarterpenes (C35 terpenes) have been found in nature, highlighting a chemical space requiring focused exploration of compounds. Noncanonical class IB terpene synthases (IB-TPSs) are the only enzymes that can cyclize C35 prenyldiphosphate. Currently, ca. 6000 IB-TPS homologues from bacteria have been registered in the NCBI database. However, only two sesquarterpenes have been identified as IB-TPS products. In this study, we performed genome mining of 11 IB-TPSs from phylogenetically diverse bacterial species using an expression system in Bacillus subtilis. We revealed that 8 of the 11 homologues with ≥30% identity to Bacillus subtilis TPS (BsuTPS) synthesize the same product, tetraprenyl-β-curcumene, as BsuTPS. The absolute stereochemistry of the tetraprenyl-β-curcumene formed by BsuTPS and one of its homologues was determined to be (-)-7R using vibrational circular dichroism and specific optical rotation analyses. In contrast, 3 of the 11 IB-TPS homologues with <30% identity to BsuTPS produced 3 novel cyclic sesquarterpenes, 1 of which had an unprecedented 5/3-fused bicyclic sesquarterpene skeleton. These new sesquarterpenes could be synthesized through various carbocation quench mechanisms that differ from those of previously identified sesquarterpenes. In this study, we demonstrated that novel sesquarterpenes can be discovered with a high probability (3 out of 4) among IB-TPS homologues with <30% identity to BsuTPS, thereby expanding the structural diversity of sesquarterpenes.

The 8-OH of rifamycin is essential for its bioactivity, while its naphthalene ring formation with 8-OH has remained unclear. Biochemical and structural analysis has demonstrated that a pair of rare NAD+-dependent dehydrogenases, RifS and RifT, forms a S2T2 structure to dehydrogenate at C8 in a structure-dependent manner. RifS catalyzes 8-dehydrogenation through a canonical NAD+-dependent mechanism, while RifT acts as a noncatalytic partner. Finally, we proposed a plausible pathway for the transformation of benzene-type prorifamycin A (1) to naphthalene-type 34a-deoxyrifamycin W (1a) bearing the 8-OH group. These results provided direct evidence for the branch point of rifamycin and 8-deoxyrifamycin biosynthesis and paved an approach to engineering novel ansamycins.

Glycans on glycoproteins play roles that range from quality control in protein folding, to mediation of interactions with other proteins, to stabilization of the protein to which they are attached. Computation can suggest structures that underlie these roles, but confidence is limited by the accuracy of energetic calculations and their applicability to the aqueous environment in which proteins function. Experimental validation of suggested structures is therefore of primary importance. Here we use NMR data, including long-range pseudocontact shifts (PCSs) and residual dipolar couplings (RDCs), to screen structures produced by a version of accelerated molecular dynamics (Pep-GaMD). This version was designed to improve the search for peptide-protein interactions, but here it is successfully applied to glycans attached to a target protein. The target protein, the N-terminal domain of human CEACAM1, is expressed with homogeneous GlcNAc2Man5 glycans at its three N-glycosylation sites. One site (N104) is found to have preferred conformations that exploit hydrophobic interactions between its glycans and protein hydrophobic residues, potentially adding to protein stability and protection from adverse interactions.

Small molecules that bind specific sites in RNAs hold promise for altering RNA function, manipulating gene expression, and expanding the scope of druggable targets beyond proteins. Identifying binding sites in RNA that can engage ligands with good physicochemical properties remains a significant challenge. fpocketR is a software and framework for identifying, characterizing, and visualizing ligand-binding sites in RNA. fpocketR was optimized, through a comprehensive analysis of currently available RNA-ligand complexes, to identify pockets in RNAs able to bind small molecules possessing favorable properties, generally termed drug-like. Here, we demonstrate multiple, complex, uses of fpocketR to analyze RNA-ligand interactions and novel pockets in small and large RNAs, to assess ensembles of RNA structure models, to identify pockets in dynamic RNA systems, and to evaluate the shapes of RNA pockets. fpocketR performs best with RNA structures visualized at atomistic resolution but also provides useful information with lower resolution structures and computational models. fpocketR is a powerful, ligand-agnostic tool for discovery and analysis of targetable pockets in RNA molecules.

Aldehyde dehydrogenase 1A1 (ALDH1A1) is highly expressed in therapy-resistant and metastatic cancers and represents a clinically relevant biomarker for selective activation strategies. We report AAP, an OFF-ON photosensitizer activated through ALDH1A1-mediated oxidation that produces singlet oxygen upon light exposure. AAP uses a donor photoinduced electron transfer (d-PeT) mechanism to suppress intersystem crossing in its unreacted benzaldehyde form, which minimizes background activity. Oxidation by ALDH1A1 disrupts d-PeT and restores phototoxicity. AAP showed minimal off-target activation by other ALDH isoforms or oxidative stress. In vivo, AAP suppressed tumor growth in two non-small cell lung cancer (NSCLC) models. In the first, intratumoral delivery into established tumors confirmed efficacy and ALDH1A1 dependence. In the second, liposomal AAP enabled intravenous delivery to early stage lesions with limited vascularization where treatment remained effective. These findings establish d-PeT suppression of intersystem crossing as an effective chemical biology strategy for enzyme-activated photodynamic therapy.

Several luciferases have been developed for imaging and biosensing, and the collection continues to grow as new applications are pursued. The current workflow for luciferase optimization, while successful, remains laborious and inefficient. Mutant libraries are generated in vitro and screened, "winning" mutants are picked by hand, and the isolated sequences are subjected to additional rounds of mutagenesis and screening. Here, we present a streamlined platform for luciferase engineering that removes the need for manual library generation during each cycle. We purposed an orthogonal DNA replication (OrthoRep) system for continuous hypermutation of a well-known luciferase (GeNL). Short cycles of culturing and screening were sufficient to evolve the enzyme, with no repetitive manual library generation necessary. New GeNL variants were identified that exhibit improved light outputs with a noncognate and inexpensive luciferin. We further characterized the novel luciferases in cell models. Collectively this work establishes OrthoRep and continuous hypermutation as a viable method to engineer luciferases, and sets the stage for more rapid development of bioluminescent reporters.