Covalent Protein Inhibitors via Tyrosine Conjugation with Cyclic Imine Mannich Electrophiles

Abstract

Targeted covalent inhibitors (TCIs) have increased in popularity among drug candidates and chemical probes. Among current TCIs, the chemistry employed is largely limited to labeling cysteine and lysine side chains. Tyrosine is an attractive residue for TCIs due to its enrichment at protein-protein interfaces. Here, we investigate the utility of cyclic imine Mannich electrophiles as covalent warheads to specifically target a pro-tein tyrosine adjacent to an inhibitor binding pocket. We characterized the intrinsic reaction rates of several cyclic imines to tyrosine and identified the iminolactone to be suitable for a covalent inhibitor (second order rate constant of 0.0029 M-1 s-1). We appended the cyclic imine warheads to a CBX8 chromodomain inhibitor to label a non-conserved tyrosine, which markedly improves both the potency and selectivity of the inhibitor for CBX8 in vitro and in cells. These results indicate that Mannich electrophiles are promising and robust chemical warheads for tyrosine bioconjugation and covalent inhibitors.

Emerging and Re-emerging Warheads for Targeted Covalent Inhibitors: An Update

“…Based on previous work using a three-component Mannich reaction to selectively target tryptophans and tyrosines for protein bioconjugation 462 and further development of this strategy utilizing a two-component version of the aforementioned reaction, 463 the groups of Dykhuizen and Krusemark, in a preprint publication, 464 recently demonstrated the utility of cyclic imine Mannich electrophiles for targeting the electronrich aromatic ring system of tyrosine in a PPI. As tyrosines are especially enriched in histone binding pockets of epigenetic regulator proteins containing BRDs and chromodomains, 465 the authors first set out to determine the reactivity of several cyclic imine derivatives to then append the best electrophiles onto a chromobox (CBX) inhibitor, in order to generate a selective CBX8 inhibitor that labels a non-conserved tyrosine.…”

“…While imino­lactams ( 273 , Figure A) showed acceptable rates of modification toward N -acetyl- l -tyrosine methyl­amide ( t 1/2 = 8 h, pH 7, k 2nd = 2.4 × 10 –4 M –1 s –1 ), imino­lactones ( 274 , Figure A) were even more reactive and demonstrated similar rates as acrylamides do toward cysteines ( t 1/2 = 40 min, k 2nd = 2.9 × 10 –3 M –1 s –1 ), also revealing that the reactivity can be tuned with EWGs adjacent to the imine. Both warheads showed high stability in buffer but reacted with other amino acids harboring an electron-rich aryl system like tryptophan and pyrrolated lysine, a lesser-known PTM (see also section ), as well as with cysteine. These characteristics are complementary to the more common sulfonyl fluorides, which tend to show limited hydrolytic stability and have different off-targets.…”

Graph Neural Networks for Identifying Protein-Reactive Compounds

“…In some cases, an atypical Michael acceptor warhead is successfully identified by the GNN, such as the tyrosineconjugating cyclic imine Mannich elctrophiles reported by Krusemark and coworkers. 98 In contrast, the novel ring-strain bicyclobutane carboxylic amides warheads were not indicated by the GradCAM map, so many of these compounds were incorrectly classified as non-reactive. In the cases where these compounds were positively classified, it appears to be a case of being "right for the wrong" reason because the GradCAM heatmap highlighted marginally electrophilic groups such as alkynes and alkenes rather than the warhead substructure.…”