Continuous hypermutation and evolution of noncanonical amino acid synthases

This study establishes an end-to-end in vivo framework for the continuous evolution of amino acid synthases using OrthoRep and aaRS-based biosensors, enabling yeast to efficiently biosynthesize diverse noncanonical tyrosine analogs from simple precursors to support scalable genetic code expansion.

The Gist

Imagine you are a master chef trying to create a brand-new, never-before-seen dish. You have a recipe book (your DNA) that tells your kitchen staff (the cell) how to build proteins. But there's a catch: your recipe book only has instructions for 20 standard ingredients (the 20 natural amino acids).

For years, scientists have been trying to sneak in "exotic" ingredients (noncanonical amino acids) to make proteins with superpowers, like glowing in the dark or sticking to cancer cells. The problem? These exotic ingredients don't grow in the cell's garden. You have to buy them from a store, ship them in, and dump them into the kitchen. This is expensive, messy, and hard to scale up.

This paper describes a brilliant solution: teaching the cell to grow its own exotic ingredients.

Here is how the scientists did it, using a few fun analogies:

1. The "Magic Factory" (The Cell)

Think of the yeast cell as a tiny, self-contained factory. Inside, there's a specific machine called Tyrosine Synthase (TmTyrS). Normally, this machine is supposed to build a standard ingredient called Tyrosine. But the scientists tweaked it so it could build exotic versions if given the right raw materials.

The problem was that the factory's machine was lazy. It was slow and inefficient. It couldn't make enough of the exotic ingredients to keep the factory running.

2. The "Evolutionary Video Game" (OrthoRep)

To fix the lazy machine, the scientists didn't just tweak it once. They put it inside a special "Evolutionary Video Game" mode called OrthoRep.

• The Glitch: Usually, cells copy their DNA perfectly. But OrthoRep is a "glitchy" copy machine. It makes mistakes on purpose, creating thousands of slightly different versions of the Tyrosine Synthase machine every time the cell divides.
• The Loop: Imagine a video game where you play a level, die, and immediately respawn with a random new power-up. The scientists let the yeast play this game continuously. Most versions of the machine were worse or useless, but a few were slightly better at making the exotic ingredients.

3. The "Bouncer and the VIP Pass" (The Selection System)

How did the scientists know which yeast cells had the "super-machines"? They set up a clever security system.

• The Bouncer (aaRS): Inside the cell, there is a bouncer (an enzyme called an aminoacyl-tRNA synthetase). This bouncer only lets a specific "VIP Pass" (a tRNA) into the main club (the protein factory) if it is holding a specific exotic ingredient.
• The Club (The Reporter): The club has a sign that says "Open" (Green Light) only if the VIP Pass gets in. If the club is closed, the sign stays red.
• The Test: The scientists added a cheap, simple raw material (like a phenol analog) to the factory.
  • If the lazy machine was there, it couldn't turn the raw material into the exotic ingredient. The bouncer stayed at the door. The club stayed closed (Red Light).
  • If a mutated, super-machine was there, it quickly turned the raw material into the exotic ingredient. The bouncer let the VIP Pass in. The club opened (Green Light).

The scientists used a giant laser sorter (FACS) to look at millions of yeast cells. They only kept the ones glowing green (the ones that successfully made the ingredient) and threw away the red ones. They did this over and over, letting the "winners" reproduce and mutate again.

4. The Result: A Self-Sufficient Factory

After many rounds of this "survival of the fittest" game, they found a champion machine: TmTyrS9.

This new machine was a superstar. It could take very cheap, simple chemicals (like 2-iodophenol, which costs pennies) and turn them into high-value exotic ingredients (like 3-iodo-tyrosine, which costs thousands of dollars per gram).

Why is this a big deal?

• Cost: It turns expensive ingredients into cheap ones.
• Scalability: You don't need to ship ingredients anymore; the factory grows them itself.
• Versatility: The machine they evolved could make four different types of exotic ingredients, not just one.

The Big Picture

Think of this paper as the invention of a self-sustaining garden. Before, if you wanted a rare flower, you had to buy it from a florist. Now, the scientists have taught the garden to grow that rare flower from a common weed, using a "glitchy" evolution machine to find the perfect gardener.

This creates a complete loop: The cell evolves the tool to make the ingredient, and the tool makes the ingredient to help the cell evolve better tools. It's a step toward creating "super-organisms" that can manufacture their own complex medicines and materials from scratch.

Technical Summary

1. Problem Statement

Genetic Code Expansion (GCE) allows for the site-specific incorporation of noncanonical amino acids (ncAAs) into proteins, enabling novel functions in biology and therapeutics. However, a major bottleneck limits the scalability and cost-effectiveness of GCE: the reliance on exogenously supplied, chemically synthesized ncAAs.

• Limitations: Most ncAAs are absent from endogenous metabolism, requiring expensive chemical synthesis, high concentrations for cellular uptake, and complex purification.
• The Gap: While engineered aminoacyl-tRNA synthetases (aaRSs) can recognize a vast array of ncAAs, there is a lack of robust, scalable pathways for the intracellular biosynthesis of these ncAAs from simple, cheap precursors. Existing biosynthetic pathways are often inefficient or non-existent for many ncAA classes.

2. Methodology

The authors developed an end-to-end experimental evolution framework using OrthoRep (an orthogonal DNA replication system in Saccharomyces cerevisiae) to evolve ncAA synthases.

• Core Platform:

  • OrthoRep: An error-prone DNA polymerase (epDNAP) was integrated into the yeast genome to continuously replicate a specific plasmid (p1) carrying the target gene (TmTyrS) at a high mutation rate ($1.6 \times 10^{-5}$ to $3.9 \times 10^{-5}$ substitutions/base), while leaving the nuclear genome untouched.
  • Target Enzyme: The study focused on evolving Tyrosine Synthase (TmTyrS), an engineered variant of Thermotoga maritima tryptophan synthase $\beta$-subunit (TrpB), which converts phenol analogs and L-serine into noncanonical tyrosines (ncTyrs).
  • Biosensor System: To link ncAA production to cellular fitness, the authors repurposed orthogonal aaRSs (specifically NitroY-F5 and its evolved variant NitroY-F5/3FY-D) as biosensors.
    • Reporter: A ratiometric "RXG" reporter containing an amber stop codon between Red Fluorescent Protein (RFP) and Green Fluorescent Protein (GFP).
    • Mechanism: If TmTyrS produces ncTyrs, the aaRS charges the orthogonal tRNA with the ncAA. This allows readthrough of the amber codon, producing GFP.
    • Selection Metric: Cells are sorted via FACS based on the Relative Readthrough Efficiency (RRE), defined as the ratio of GFP/RFP in the RXG reporter (amber) normalized to an RYG reporter (sense codon). High RRE indicates successful ncAA biosynthesis.

• Evolutionary Strategy:

  1. Positive Selection: Cells grown with phenol analogs (e.g., 2-iodophenol) were sorted for the top 0.05% highest RRE.
  2. Negative Selection: Cells grown without phenol analogs were sorted for the bottom 5% (lowest RRE) to eliminate "cheater" mutations (e.g., amber-to-sense codon reversion).
  3. Iteration: This cycle was repeated over multiple generations to accumulate beneficial mutations.

3. Key Contributions

• Unified Evolution Framework: The paper completes a proposed framework where both the aaRS (for ncAA recognition) and the ncAA synthase (for ncAA production) are evolved using the same OrthoRep-driven continuous hypermutation strategy.
• In Vivo Biosynthesis of ncTyrs: Successfully evolved TmTyrS variants capable of producing high levels of ncTyrs from simple, inexpensive phenol precursors directly inside yeast cells.
• Novel Enzyme Variants: Generated TmTyrS9, a variant with 19 cumulative amino acid substitutions, exhibiting unprecedented catalytic efficiency for multiple ncTyrs.
• Cost Reduction: Demonstrated the biosynthesis of 3-methyl-L-tyrosine, a compound that is chemically difficult and expensive to synthesize (~$1,600/g), from cheap 2-methylphenol (<$0.1/g).

4. Key Results

• Evolutionary Trajectory: Starting from a moderately active precursor (TmTyrSc), three sequential evolution campaigns yielded variants TmTyrS7, TmTyrS8, and finally TmTyrS9.
• Performance Metrics:
  • TmTyrS9 achieved an RRE of ~1.0 (equivalent to native sense codon translation) when supplied with 50 $\mu$M 2-iodophenol.
  • It showed high activity for 3-bromo-L-tyrosine (RRE ~1.0 at 158 $\mu$M 2-bromophenol).
  • It enabled the production of 3-chloro- and 3-methyl-L-tyrosine, substrates for which the parental enzyme (TmTyrS6) showed negligible activity.
• Structural Insights:
  • D300N Mutation: A critical substitution in the catalytic pocket (Aspartate to Asparagine) was identified. This residue is highly conserved as Aspartate or Arginine in natural TrpB enzymes but is rare (0.33% frequency) in evolutionary datasets. This mutation likely alters substrate binding and allosteric regulation.
  • F200S Mutation: Another rare substitution (Phenylalanine to Serine) affecting the helix near the catalytic lysine.
  • Reversions: Several mutations reverted to residues found in the wild-type TrpB or earlier engineering campaigns, suggesting a re-optimization of the enzyme's core catalytic machinery.
• Fidelity Confirmation: Mass spectrometry of sfGFP produced in the presence of 2-iodophenol confirmed the incorporation of 3-iodo-L-tyrosine at the amber codon position, proving that the biosynthesized ncAA was correctly recognized by the aaRS and incorporated into the protein.

5. Significance

• Scalability and Economics: This work removes the dependency on expensive, externally supplied ncAAs. By enabling cells to biosynthesize ncAAs from cheap precursors (phenols and serine), it drastically lowers the cost barrier for GCE applications.
• Platform Versatility: The framework is not limited to tyrosine synthases. Because aaRSs can be evolved for diverse specificities, this platform can be adapted to evolve biosynthetic pathways for a wide range of ncAAs, expanding the "chemical space" accessible to synthetic biology.
• End-to-End GCE: It establishes a complete pipeline for creating organisms that not only recognize but also produce their own expanded genetic code components, paving the way for more autonomous and scalable bio-manufacturing of novel proteins and therapeutics.