A research group at Japan’s Nagoya University has found that adding a four-amino-acid tag to the start of a protein can stop ribosomes from stalling and significantly improve how much protein E. coli produces. It is a small change with large consequences for how biologics are manufactured, and AI did much of the work of finding it.
Why E. coli matters
E. coli, a species of bacteria found in the gut, is essential for producing proteins that are later used as enzymes, antibodies and vaccine components. It is the workhorse of the field because it grows fast, is cheap to culture, is easy to genetically modify, produces large quantities of protein, and is one of the most researched organisms there is.
The problem comes during foreign protein production. The amount of protein produced is often reduced when ribosomes stall during translation, a common phenomenon when microbes are made to produce proteins that are not their own. When ribosomes pause or get stuck on the mRNA, protein synthesis stops, which means lower yield, higher cost and weaker efficiency. Trailing ribosomes stack up behind the stalled one, and the cell responds by cleaving the mRNA and tagging the ribosomal protein RPS3 for ubiquitination by HEL2.
The Nagoya group found that adding specific short peptide sequences, tetrapeptides that are four amino acids long, to the N-terminus of a protein (the start of the protein; the C-terminus is the end) can prevent this stalling and significantly improve expression. They initially found one sequence that works, Ser-Lys-Ile-Lys, and are now investigating other combinations that could boost translation further.
How they found it
The method started with a library of 160,000 different tetrapeptides. There are 20 natural amino acids and four slots, so 20^4 gives 160,000 unique sequences, every possible combination. The group then used AI to analyze 250 experimental datasets, identify which peptides prevent stalling, and predict the performance of all 160,000 sequences.
The relevance is that a tiny four-amino-acid tag that increases protein yield makes research and production considerably more efficient and faster. That includes cheaper manufacturing of biologics such as complex proteins and enzymes, more efficient enzyme production for biofuels and bioplastics, and more consistent biomanufacturing pipelines.
What comes next
Companies will begin screening these peptide chains with AI to predict which proteins need which tetrapeptide so ribosomal stalling does not occur. In theory that should make protein manufacturing cheaper and more efficient overall, especially for difficult-to-express proteins such as membrane proteins and complex enzymes.
This can push AI further into biopharma: designing peptides in an instant, predicting expression levels, and using CRISPR to insert these tetrapeptide sequences straight into DNA. Expect biotech startups and institutions to secure patents for peptide and tetrapeptide combination libraries and to build their own AI-enhanced sequence design tools and platforms.
Keep an eye out for further developments and new tetrapeptide discoveries, and for any M&A among the companies becoming leading candidates in this approach to protein production.