Carol A. Fierke

Aromatic π-π stacking interactions are fundamental to protein architecture, molecular recognition, and drug efficacy, yet directly quantifying them under near-physiological conditions has remained challenging. Here, we use a recently developed spectroscopic platform, thermostable Raman interaction profiling (TRIP), that enables direct, label-free detection and quantification of aromatic π-π interactions in complex protein environments. Using the SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) main protease (M ^pro ) as a biologically and clinically relevant model, we demonstrate that subtle changes in the phenylalanine benzene ring breathing (BRB) mode serve as a precise spectroscopic indicator of π-π stacking strength. This signal is highly responsive to both protein concentration-dependent dimerization and ligand-induced structural changes. M ^pro forms a catalytically active dimer stabilized by a conserved aromatic triad (phenylalanine-140, histidine-163, and histidine-172), providing an ideal system to interrogate π-stacking at an important protein interface. Potent inhibitors MPI8 and nirmatrelvir produced the strongest BRB spectral shifts, broadening, and intensity changes, consistent with enhanced aromatic stacking and dimer stabilization, whereas halicin and VB-B-145 showed weaker engagement. BRB spectral changes also showed quantitative correlation with dimerization efficiency, published IC ₅₀ (median inhibitory concentration) values, and antiviral efficacy in A549-ACE2 cells. Complementary density functional theory revealed electron density rearrangements and vibrational coupling patterns unique to stacked aromatic residues. This integrated spectroscopic-computational approach enables quantitative probing of π-π stacking in native-like protein environments and positioning TRIP as a generalizable tool for designing drugs targeting aromatic protein-protein interfaces.

Epigenetic reader proteins interpret histone epigenetic marks to regulate gene expression. Given their vital roles and the link between their dysfunction and various diseases, these proteins present compelling targets for therapeutic interventions. Nevertheless, designing selective inhibitors for these proteins poses significant challenges, primarily due to their unique properties such as shallow binding sites and similarities with homologous proteins. To overcome these challenges, we propose an innovative strategy that uses phage display with a genetically encoded noncanonical amino acid (ncAA) containing an epigenetic mark. This ncAA guides binding to the reader protein’s active site, allowing the identification of peptide inhibitors with enhanced affinity and selectivity. In this study, we demonstrate this novel approach’s effectiveness by identifying potent inhibitors for the ENL YEATS domain that plays a critical role in leukemogenesis. Our strategy involved genetically incorporating N^ε-butyryl-l-lysine (BuK), known for its binding to ENL YEATS, into a phage display library for enriching the pool of potent inhibitors. One resultant hit was further optimized by substituting BuK with other pharmacophores to exploit a unique π–π–π stacking interaction with ENL YEATS. This led to the creation of selective ENL YEATS inhibitors with a K_D value of 2.0 nM and a selectivity 28 times higher for ENL YEATS than its close homologue AF9 YEATS. One such inhibitor, tENL-S1f, demonstrated robust cellular target engagement and on-target effects to inhibit leukemia cell growth and suppress the expression of ENL target genes. As a pioneering study, this work opens up extensive avenues for the development of potent and selective peptidyl inhibitors for a broad spectrum of epigenetic reader proteins.

Numerous organic molecules are known to inhibit the main protease (M^Pro) of SARS-CoV-2, the pathogen of Coronavirus Disease 2019 (COVID-19). Guided by previous research on zinc-ligand inhibitors of M^Pro and zinc-dependent histone deacetylases (HDACs), we identified BRD4354 as a potent inhibitor of M^Pro. The in vitro protease activity assays show that BRD4354 displays time-dependent inhibition against M^Pro with an IC₅₀ (concentration that inhibits activity by 50%) of 0.72 ± 0.04 μM after 60 min of incubation. Inactivation follows a two-step process with an initial rapid binding step with a K_I of 1.9 ± 0.5 μM followed by a second slow inactivation step, k_inact,max of 0.040 ± 0.002 min^-1. Native mass spectrometry studies indicate that a covalent intermediate is formed where the ortho-quinone methide fragment of BRD4354 forms a covalent bond with the catalytic cysteine C145 of M^Pro. Based on these data, a Michael-addition reaction mechanism between M^Pro C145 and BRD4354 was proposed. These results suggest that both preclinical testing of BRD4354 and structure-activity relationship studies based on BRD4354 are warranted to develop more effective anti-COVID therapeutics.

As the COVID-19 pathogen, SARS-CoV-2 relies on its main protease (M^Pro) for pathogenesis and replication. During crystallographic analyses of M^Pro crystals that were exposed to the air, a uniquely Y-shaped, S-O-N-O-S-bridged post-translational cross-link that connects three residues C22, C44, and K61 at their side chains was frequently observed. As a novel covalent modification, this cross-link serves potentially as a redox switch to regulate the catalytic activity of M^Pro, a demonstrated drug target of COVID-19. The formation of this linkage leads to a much more open active site that can potentially be targeted for the development of novel SARS-CoV-2 antivirals. The structural rearrangement of M^Pro by this cross-link indicates that small molecules that lock M^Pro in the cross-linked form can potentially be used with other active-site-targeting molecules such as paxlovid for synergistic effects in inhibiting SARS-CoV-2 viral replication.

Eleven-nineteen leukemia protein (ENL) plays pivotal roles in the leukemogenesis. As a YEATS domain protein, ENL reads histone acylation marks and recruits key transcription factors to leukemic drivers such as HOXA9, MEIS1 , and MYB and therefore promotes leukemia development. The histone-reading function of ENL has been proven essential in the onset and progression of several acute leukemias, suggesting a putative therapeutic window for ENL inhibition. In this study, we developed a phage-assisted, active site-directed ligand evolution (PADLE) approach for the identification of potent and selective ENL inhibitors, where N ^ε -butyryl- l -lysine (BuK) that possesses known target-protein interactions with the ENL YEATS domain was genetically incorporated into a phage display library to serve as a warhead to direct displayed peptides to the active site of ENL YEATS for enrichment. Using this novel strategy in combination with structure-activity relationship that replaced BuK with other ncAAs for de novo π-π-π stacking interactions with two aromatic residues in ENL YEATS, selective and potent ENL inhibitors with a K _d value as low as 2.0 nM were identified. One pentapeptide inhibitor tENL-S1f displayed selective inhibition of ENL over other YEATS domains as well as strong cellular target engagement and on-target effects in inhibiting leukemia cell growth and suppressing the expression of ENL target genes. As the first of its kind study, the current work opens a large avenue of research of using PADLE to develop selective and potent peptidyl inhibitors for a large variety of epigenetic reader proteins.

Determination of metal ions such as zinc in solution remains an important task in analytical and biological chemistry. We describe a novel zinc ion biosensing approach using a carbonic anhydrase–Oplophorus luciferase fusion protein that employs bioluminescence resonance energy transfer (BRET) to transduce the level of free zinc as a ratio of emission intensities in the blue and orange portions of the spectrum. In addition to high sensitivity (below nanomolar levels) and selectivity, this approach allows both quantitative determination of “free” zinc ion (also termed “mobile” or “labile”) using bioluminescence ratios and determination of the presence of the ion above a threshold simply by the change in color of bioluminescence, without an instrument. The carbonic anhydrase metal ion sensing platform offers well-established flexibility in sensitivity, selectivity, and response kinetics. Finally, bioluminescence labeling has proven an effective approach for molecular imaging in vivo since no exciting light is required; the expressible nature of this sensor offers the prospect of imaging zinc fluxes in vivo.

Phage‐assisted, active site‐directed ligand evolution (PADLE) is a recently developed technique that uses an amber codon‐encoded noncanonical amino acid (ncAA) as an anchor to direct phage‐displayed peptides to a target for an enhanced ligand identification process. 2‐Amino‐8‐oxodecanoic acid (Aoda) is a ketone‐containing ncAA residue in the macrocyclic peptide natural product apicidin that is a pan‐inhibitor of Zn²⁺‐dependent histone deacetylases (HDACs). Its ketone serves as an anchoring point to coordinate the catalytic zinc ion in HDACs. Using a previously evolved N^????‐acetyl‐lysyl‐tRNA synthetase in combination with tRNA^Pyl, we showed that Aoda was efficiently incorporated into proteins in Escherichia coli by amber suppression. By propagating an amber codon‐obligate phagemid library in E. coli encoding Aoda, we generated an Aoda‐containing phage‐displayed peptide library. Using this library to conduct PADLE against HDAC8 revealed a 7‐mer peptide GH8P01F1 with Aoda‐flanking amino acid residues that matched existing peptide sequences in identified HDAC8 substrates. Switching Aoda in GH8P01F1 to a more Zn²⁺‐chelating ncAA S‐2‐amino‐8‐hydroxyamino‐8‐oxooctanoic acid (Asuha) led to an extremely potent compound GH8HA01, which has an HDAC8‐inhibition K_i value of 0.67 nM. GH8HA01 and its 5‐mer truncation analogue Ac‐GH8HA01Δ1Δ7 that has an HDAC8‐inhibition K_i value of 0.31 nM are two of the most potent HDAC8 inhibitors that have been developed. Furthermore, both are highly selective against HDAC8 compared with other HDACs tested, demonstrating the great potential of using PADLE to identify highly potent and selective ligands for targets with conserved active sites among homologues.

The first step in transfer RNA (tRNA) maturation is the cleavage of the 5’ end of precursor transfer RNA (pre-tRNA) catalyzed by ribonuclease P (RNase P). RNase P is either a ribonucleoprotein (RNP) complex with a catalytic RNA subunit or a pro tein-only R Nase P (PRORP). In most land plants, algae, and Euglenozoa, PRORP is a single-subunit enzyme. There are currently no inhibitors of protein-only RNase P that can be used as tools for studying the biological function of this enzyme. Therefore, we screened for compounds that inhibit the activity of a model PRORP from A. thaliana organelles (PRORP1) using a high throughput fluorescence polarization (FP) cleavage assay. Two compounds, gambogic acid and juglone (5-hydroxy-1,4-naphthalenedione) that inhibit PRORP1 in the 1 μM range were identified and analyzed. These compounds similarly inhibit human mtRNase P, a multi-subunit protein enzyme, and are 50-fold less potent against bacterial RNA-dependent RNase P. Biochemical measurements indicate that gambogic acid is a rapid-binding, uncompetitive inhibitor that targets the PRORP1-substrate complex while juglone acts as time-dependent inhibitor of PRORP1. X-ray crystal structures of PRORP1 in complex with juglone demonstrate the formation of a covalent complex with cysteine side chains on the surface of the protein. A model consistent with the kinetic data is that juglone binds to PRORP1 rapidly to form an inactive enzyme-inhibitor (EI) complex, and then undergoes a slow step to form an inactive covalent adduct with PRORP1. These inhibitors have the potential to be developed into tools to probe PRORP structure and function relationships.

The phage-assisted, active site-directed ligand evolution (PADLE) is a recently developed technique that uses an amber codon-encoded noncanonical amino acid (ncAA) as an anchor to direct phage-displayed peptides to a target for an enhanced ligand identification process. 2-Amino-8-oxodecanoic acid (Aoda) is a ketone-containing ncAA residue in the macrocyclic peptide natural product apicidin that is a pan-inhibitor of Zn ²⁺ -dependent histone deacetylases (HDACs). Its ketone serves as an anchoring point to chelate the catalytic zinc ion in HDACs. Using a previously evolved N ^ε - acetyl-lysyl-tRNA synthetase in combination with tRNA ^Pyl , we showed that Aoda was efficiently incorporated into proteins in Escherichia coli by amber suppression. By propagating an amber codon-obligate phagemid library in E. coli encoding Aoda, we generated an Aoda-containing phage-displayed peptide library. Using this library to conduct PADLE against HDAC8 revealed a 7-mer peptide GH8P01F1 with Aoda-flanking amino acid residues that match existing peptide sequences in identified HDAC8 substrates. Switching Aoda in GH8P01F1 to a more Zn ²⁺ -chelating ncAA S -2-amino-8-hydroxyamino-8-oxooctanoic acid (Asuha) led to an extremely potent GH8HA01 that has an HDAC8-inhibition K _i value as 0.67 nM. GH8HA01 and its 5-mer truncation analogue Ac-GH8HA01 Δ 1 Δ 7 that has an HDAC8-inhibition K _i value as 0.31 nM represent two most potent HDAC8 inhibitors that have been developed so far and both are highly selective against HDAC8 over the other three tested HDACs, demonstrating the great potential of using PADLE to identify highly potent and selective ligands for targets.