← Back to Home

Rewriting tRNA Ends to Expand the Chemical Space for Protein Design

A Harvard team has moved experiments with the genetic code into test tubes, using modified tRNAs and ribosomes to test new rules for protein assembly. The system could theoretically accommodate 34 amino acids, but validation so far remains limited to incorporating nonstandard amino acids and reassigning up to three codons.

By SURL BioNews

Giving proteins access to a wider variety of chemical building blocks could allow researchers to design functions rarely found in natural proteins. However, rewriting a cell’s manufacturing rules often affects its entire biological machinery. The AGENTEX platform, developed by a team at Harvard Medical School and the Wyss Institute, first tests new rules in a cell-free test-tube system, allowing exploration without having to begin each time by rewriting an organism’s genome.

The study, with Felix Radford as first author and George Church as senior author, was published in Nature on August 26, 2026. The key finding involved transfer RNA (tRNA): these molecules, which transport amino acids, typically have a CCA sequence at their ends. The team found that the enzymes in Escherichia coli responsible for attaching amino acids to tRNAs tolerate changes to these ends better than previously understood; many tRNAs with modified ends could still be loaded with amino acids.

With this flexibility, researchers could combine modified tRNAs with matching engineered ribosomes to produce peptides in cell extracts. This paired design is intended to reduce interference between the new translation rules and the existing system. The study also established an analytical method called tSCAN to screen for tRNA variations that still allow efficient amino acid attachment.

The headline-grabbing figure of “34 amino acids” needs to be understood within the scope of the experiments. Under the study’s design, reducing redundancy in the genetic code could theoretically make room for 14 building blocks beyond the 20 standard amino acids. However, the paper actually tested two alternative genetic codes and demonstrated the incorporation of nonstandard amino acids and the reassignment of up to three codons. It has not yet shown that a single system can simultaneously assemble 34 different amino acids into proteins.

Another focus of AGENTEX is linking repeated experiments into an automated workflow. According to the Wyss Institute, the team provides free software that can be used with Opentrons robots to prepare and test different combinations in batches. This could allow researchers to compare genetic code designs more quickly, then use the results to adjust the next round of experiments, gradually exploring assembly methods for new peptides and biomaterials.

There is still some distance to go from test tubes to practical use. The research currently focuses on cell-free translation systems, and how to achieve the desired performance in living cells or organisms remains to be clarified; expanding the range of available building blocks also cannot be directly equated with drug benefits. The study disclosed that Radford and Church are inventors on related patent applications, and that Church also has related company interests. At this stage, the most concrete value of the work is an experimental platform for repeatedly testing the genetic code, giving more protein chemistry designs a chance to be made and then evaluated.

References

  1. Phys.org
  2. Nature
  3. Wyss Institute at Harvard University
  4. Chemical & Engineering News