A sequence-independent way to edit protein backbones

Proteins carry out exceptionally diverse molecular functions—even though they are all built on an unvarying polyamide backbone. How much more chemical diversity might be unlocked by editing that backbone? Although researchers have worked hard to modify the protein-synthesizing ribosome to build new types of bond, the task has proven difficult. Now researchers led by Alanna Schepartz at the University of California, Berkeley, report that they’ve serendipitously found a way to install a new carbon-carbon bond into a polypeptide (JACS 2025, DOI: https://doi.org/10.1021/jacs.4c14103).

Postdoctoral fellow Carly Schissel stumbled across the reaction while trying to study intermolecular reactions between peptides containing unnatural amino acids. Extrapolating from a previous finding, that dehydroalanine in a small peptide could catalyze a reaction between peptides, she incorporated dehydrolactone (DHL) into a synthetic tripeptide to characterize its reactivity. But the reaction didn’t go as planned.

“I would never get the product I expected,” Schissel says. It was only when she ran a control experiment with the DHL-containing tripeptide on its own that she realized it was instead reacting with itself.

“It took about 24 h for Carly to go from ‘I’m not getting the product I want’ to ‘I can’t believe the product I am getting,’ ” Schepartz says. “It was incredibly exciting and just so much fun to watch.”

Schissel found that under basic conditions, the carbon and oxygen in DHL within the tripeptide switched places, installing a carbon atom in the peptide backbone—a notoriously difficult job.

The researchers worked out a mechanism for the intramolecular rearrangement. The reaction proceeds by way of a carbon nucleophile—a species that’s difficult to make because it is so reactive. But this nucleophile, Schepartz says, “never has a chance to do anything other than react with the proximal reactive carbonyl.”

The researchers replicated the reaction in peptides synthesized by a ribosome; to include DHL in the ribosomally synthesized peptide, they incorporated phenylselenocysteine, which can be converted to DHL by oxidation.

The blended amide-ketide backbone that the reaction yields could generate many interesting molecules, the researchers say. For example, a few natural products are made by complex enzyme cascades that blend ketones and amides in their backbones. With this reaction, researchers could generate libraries that resemble those molecules, produce peptide-like molecules resistant to cleavage by proteases, and embed rings in protein backbones.

Wilfred van der Donk, a chemist at the University of Illinois Urbana-Champaign, calls the new reaction unexpected and very interesting. He says that in contrast to methods for modifying peptide backbones that his lab team and others have reported, the method appears to be work independent of a protein’s amino acid sequence. “In principle, this is a more general way to put ketones into the backbone,” Van der Donk says.