Serine and threonine residues are central to post-translational modifications including phosphorylation and glycosylation, and they form the catalytic triads of serine hydrolases. Yet their hydroxyl groups are notoriously difficult to label selectively in complex peptides, because thiols and amines are more reactive, phenols compete for many electrophilic reagents, and most solvents that dissolve peptides do little to suppress off-target reactivity. Phosphorus(V) reagents reported by Baran achieve selectivity for serine over threonine in dimethylformamide, but routes to broader alcohol-side-chain editing remain limited. A general platform that labels both Ser and Thr rapidly, tolerates nearly all proteinogenic residues, and delivers multiple downstream transformations from a single tagged intermediate would substantially expand the toolkit for peptide chemical biology.
Researchers in the Chen Lab at Nankai University, published in the Journal of the American Chemical Society, describe how fluorosulfuryl isocyanate, FSI, reacts with Ser and Thr side chains in hexafluoroisopropanol, HFIP, solvent within one minute at room temperature to give fluorosulfuryl carbamate products in near-quantitative conversion. The team used model peptides to screen solvents systematically, performed density functional theory calculations to map transition-state energies for competing nucleophiles, and applied the optimized conditions to a library of natural and synthetic peptides as well as two model proteins.
The HFIP solvent does two distinct jobs. Its mild acidity protonates the ε-amino groups of lysine, histidine, and arginine, suppressing their reactivity toward FSI, while HFIP clusters act as proton shuttles that lower the activation barrier for alcohol addition to the C=N bond of FSI from 28.2 kcal/mol to 11.3 kcal/mol. Under these optimized conditions, all proteinogenic amino acids except cysteine are tolerated; tryptophan labels as a minor side product at below 10%, and methionine undergoes sulfoxide formation at below 3%. Cyclic peptide drugs including Atosiban, cyclo(-RGDfT), cyclosporin A, and the antimitotic payload monomethyl auristatin E all label in above 90% conversion by LC-MS. Labeling of the macrocyclic antibiotic daptomycin with a reduced FSI loading of 2.0 equivalents gives selective Ser modification in 79% yield. For proteins, lysozyme acquires up to nine carbamate tags, and tandem MS/MS confirms modification at 16 of 17 Ser and Thr residues with no off-target labeling; circular dichroism spectra show the labeled enzyme retains its native fold.
The fluorosulfuryl carbamate tag serves as a branch point for three further transformations. First, sulfur(VI) fluoride exchange, SuFEx, ligation with amines under Ca(OTf)2/DABCO conditions at room temperature converts the tag into sulfamide-linked conjugates; drug molecules such as Alogliptin, Amoxapine, and Duloxetine as well as a biotin handle and a fluorophore all ligate in good to excellent yields. Second, intramolecular SuFEx with a proximal lysine side chain generates Ser/Thr-Lys stapled peptides; a 10-mer α-synuclein fragment cyclizes in 91% yield, and Atosiban gives a bicyclic product selectively. Third, treatment with dilute base in water above pH 10.5 eliminates the carbamate to produce dehydroalanine or dehydrobutyrine residues, which undergo one-pot Michael addition with thiols, amines, or phosphines to deliver deoxygenative heteroatom-linked analogs; S-glycoconjugation with β-1-thioglucose proceeds in approximately 70% yield, and glutathione ligation reaches 91%.
The platform addresses a persistent gap in residue-selective chemistry by providing rapid, mild access to two of the most abundant yet underexploited side chains in peptide and protein sequences. Because the three downstream pathways, SuFEx ligation, stapling, and elimination-addition, diverge from a common tagged intermediate that is stable between pH 3.0 and 9.5 for at least 24 hours, practitioners can prepare and store labeled peptides before choosing a modification pathway. Extension to proteins in HFIP at low concentration suggests that the approach may complement existing bioconjugation strategies for antibody-drug conjugate construction and activity-based protein profiling where Ser- or Thr-directed chemistry is needed.
