Unnatural amino acids have become some of the most valuable building blocks in modern drug discovery, appearing in peptide therapeutics, antibody–drug conjugates, and enzyme inhibitors where a single noncanonical residue can dramatically improve stability, potency, or pharmacokinetics. Yet despite their pharmaceutical importance, most of these molecules are still produced through multi-step chemical synthesis that demands expensive catalysts, protecting groups, and laborious purification. A new study published in Biotechnology for Biofuels and Bioproducts describes a refreshingly different approach: instead of building elaborate amino acids from scratch, researchers have engineered bacteria to grow them one carbon at a time, using evolved transaminase enzymes paired with a chain-extension module borrowed from central metabolism.
The work, led by Rex C. Wen, Ryan S. Wang, Nancy N. Kang, Patrick Y. Lin, and colleagues under the corresponding authorship of Claire R. Shen at National Tsing Hua University in Hsinchu, Taiwan, presents what the authors call a modular strategy to elongate α-amino acids by single-carbon increments. The core insight is that any α-amino acid can, in principle, be traced back to its corresponding α-ketoacid precursor, and that ketoacids can be chemically lengthened by one methylene unit through well-characterized acetyl-CoA-dependent reactions. If a transaminase could then be evolved to convert the elongated ketoacid back into the corresponding amino acid, the cycle could, in theory, be repeated indefinitely, walking up the homologous series of amino acids one carbon at a time.
The challenge, of course, lies in the enzymes. Native Escherichia coli transaminases such as TyrB, an aromatic amino acid transaminase, and IlvE, a branched-chain amino acid transaminase, are exquisitely tuned to their natural substrates and typically reject ketoacids with unfamiliar side chains, particularly those bearing bulky tert-butyl groups, hydroxyl groups, or thioether linkages. To overcome this specificity barrier, the team turned to directed evolution, the Nobel-recognized technique of introducing mutations across an enzyme’s active site and screening vast libraries for improved function. But screening thousands of enzyme variants against dozens of exotic substrates is slow and expensive, so the researchers devised an elegant biological shortcut: a glutamate auxotroph selection system.
In this selection scheme, an E. coli strain unable to synthesize its own glutamate survives only if it can produce the amino acid through an alternative transaminase reaction. By wiring the engineered transaminase to the cell’s survival, the researchers converted an enzyme-engineering problem into a growth problem. Cells carrying transaminase variants capable of acting on a given unnatural ketoacid substrate would grow; those carrying inactive or poorly suited variants would not. This growth-based selection rapidly enriched beneficial mutations, and, crucially, it illuminated previously uncharacterized residues in both TyrB and IlvE that influence substrate recognition. Substitutions at these newly identified positions broadened the enzymes’ substrate scope toward hydroxyl-bearing, tert-butyl-bearing, and thioether-containing frameworks with extended side chains, expanding the catalytic repertoire far beyond what the wild-type enzymes could accomplish.
With an expanded transaminase toolbox in hand, the team coupled it to an acetyl-CoA-dependent ketoacid extension module that adds a single carbon to the α-ketoacid backbone before the amination step. Together, the two modules enabled the first biosynthetic access to a remarkable panel of chain-elongated unnatural α-amino acids produced directly from their minus-one precursors. Among the products were neopentylglycine, an amino acid with a bulky quaternary carbon in its side chain that is notoriously difficult to access biosynthetically; C5 and C6 hydroxy amino acids relevant to hydrophobic peptide design; homomethionine, the one-carbon-elongated cousin of the essential amino acid methionine; and homoglutamate, the elongated analogue of glutamate.
Perhaps the most striking demonstration of the platform’s generality is its iterative capability. The researchers showed that homoserine, the immediate biosynthetic precursor of threonine, could be pushed through two consecutive elongation cycles, generating both the plus-one and plus-two homologues in sequence. This means the system is not limited to a single elongation event but can genuinely walk up a homologous series, a property that dramatically multiplies the number of accessible noncanonical amino acids from a finite set of starting ketoacids. In effect, the platform functions as a molecular ratchet, converting cheap fermentable carbon sources into an expanding library of unnatural building blocks without requiring any new chemistry for each additional carbon.
To showcase pharmaceutical relevance, the team selected homoglutamate as a flagship target and pushed the system toward preparative scale. In ordinary shake flasks, without the benefit of controlled bioreactors or fed-batch optimization, the engineered strains accumulated homoglutamate at titers of up to 10 grams per liter. For an entirely novel biosynthetic route to a nonproteinogenic amino acid, this titer is notable, and it underscores the scale-up potential of the approach for industrial bioproduction. Fermentation at the multi-gram-per-liter level is the threshold at which microbial manufacturing becomes economically interesting, and reaching it in shake flasks suggests substantial headroom for further process development.
The significance of this work extends beyond the specific molecules produced. By establishing a generalizable enzymatic framework for stepwise expansion of α-amino acid scaffolds, the study offers synthetic biologists a composable strategy: any ketoacid that can be elongated by the acetyl-CoA module, and any transaminase that can be evolved to aminate the product, becomes a route to a new unnatural amino acid. Because the selection system couples enzyme activity to cell growth, the directed evolution step is cheap, scalable, and does not require specialized analytical screening for every variant. This combination of modularity, iterativity, and selection-driven evolvability distinguishes the platform from earlier biosynthetic efforts that typically addressed one target molecule at a time.
The implications for medicine and industry are considerable. Noncanonical amino acids are increasingly used to tune the protease resistance and half-life of peptide drugs, to install bioconjugation handles in protein therapeutics, and to probe biological mechanisms with chemical precision. A fermentative route to these molecules could reduce costs, eliminate hazardous reagents, and shrink the environmental footprint of amino acid manufacturing, aligning with the broader mission of biotechnology for sustainable fuels and products. At the same time, the identification of previously uncharacterized substrate-determining residues in TyrB and IlvE adds to the fundamental understanding of how PLP-dependent transaminases recognize their substrates, knowledge that will inform future enzyme engineering campaigns well beyond this particular system.
Funded by Taiwan’s National Science and Technology Council under grant NSTC111-2221-E007-007-MY3, the study was published as an open-access article with a permanent DOI, making the strains, methods, and data available to the wider community. As the authors note, the framework broadens access to noncanonical building blocks for peptide and bioproduction applications, and the iterative elongation concept invites exploration of even longer homologues, additional functional groups, and other transaminase scaffolds. What began as a question about whether enzymes could be taught to lengthen amino acids one carbon at a time has matured into a working platform that grows some of chemistry’s most sought-after molecules directly inside living cells, one methylene unit at a time.
Subject of Research: Directed evolution of E. coli transaminases for one-carbon biosynthetic elongation of unnatural α-amino acids
Article Title: Synthetic amino acid elongation via transaminase engineering
Article References: Wen, R. C., Wang, R. S., Kang, N. N., Lin, P. Y., Liao, V. Y., Mai, H.-T., Tsai, T. P., Lin, J. J., Wang, J. C., & Shen, C. R. (2026). Synthetic amino acid elongation via transaminase engineering. Biotechnology for Biofuels and Bioproducts. https://doi.org/10.1186/s13068-026-02821-y
Image Credits: AI Generated
DOI: 10.1186/s13068-026-02821-y
Keywords: unnatural amino acids, transaminase engineering, directed evolution, metabolic engineering, α-ketoacid extension, homoglutamate, neopentylglycine, homomethionine, Escherichia coli, biocatalysis, protein engineering, bioproduction
Cite Scienmag News
APA
MLA
Chicago
Drew Townsend. (September 21, 2026). Engineered Transaminases Grow Unnatural Amino Acids One Carbon at a Time. Scienmag. https://scienmag.com/engineered-transaminases-grow-unnatural-amino-acids-one-carbon-at-a-time/
Drew Townsend. “Engineered Transaminases Grow Unnatural Amino Acids One Carbon at a Time.” Scienmag, 21 September 2026, https://scienmag.com/engineered-transaminases-grow-unnatural-amino-acids-one-carbon-at-a-time/. Accessed 21 September 2026.
Drew Townsend. “Engineered Transaminases Grow Unnatural Amino Acids One Carbon at a Time.” Scienmag. September 21, 2026. https://scienmag.com/engineered-transaminases-grow-unnatural-amino-acids-one-carbon-at-a-time/
Copy citation
Download RIS
Tags: antibody-drug conjugate synthesisbiocatalysisbioproductionbiotechnological production of unnatural amino acidsdirected evolutionengineered bacteria for amino acid productionenzymatic chain elongation in amino acid biosynthesisenzyme evolution for amino acid modificationEscherichia colihomoglutamatehomomethioninemetabolic engineeringmetabolic engineering of amino acid pathwaysmicrobial synthesis of noncanonical amino acidsmodular biosynthetic pathway for amino acidsneopentylglycineone-carbon chain extension in amino acidspeptide therapeutic developmentProtein Engineeringtransaminase engineeringtransaminase enzyme engineeringunnatural amino acid biosynthesisunnatural amino acidsα-ketoacid extension
