Virus-free continuous directed evolution in human cells using somatic hypermutation

The authors developed CODE-HB, a virus-free continuous directed evolution platform that repurposes human B cell somatic hypermutation mechanisms to rapidly evolve diverse proteins, including neutralizing antibodies against influenza, directly within human cells without compromising genome-wide fitness.

The Gist

The Big Idea: Teaching Cells to "Evolve" on Demand

Imagine you are trying to invent a new key that fits a very specific, complicated lock (like a virus). In the past, scientists had to build thousands of keys by hand in a lab, test them one by one, throw away the bad ones, and repeat the process. It was slow, expensive, and required a lot of manual work.

This paper introduces a new method called CODE-HB. Think of it as giving the human cell a "magic wand" that allows it to invent new keys inside the cell itself, continuously, without stopping.

The Problem: Evolution is Usually Too Slow

In nature, evolution happens because of random mistakes (mutations) when cells copy their DNA. But these mistakes are rare. If you wanted a specific change in a gene, you might have to wait for a cell to divide a million times to get lucky. It's like waiting for a monkey typing randomly to accidentally type out a whole Shakespeare play.

The Solution: Borrowing from the Immune System

The authors looked at how our bodies naturally fight infections. Our immune system has special cells called B-cells. When a B-cell meets a virus, it doesn't just copy its DNA perfectly; it intentionally makes lots of mistakes in the part of the DNA that codes for antibodies. This is called Somatic Hypermutation (SHM). It's like the B-cell is saying, "I need a better key, so I'm going to try 1,000 different variations of this key right now!"

Usually, this happens only in the specific "antibody factory" section of the B-cell's DNA. The rest of the cell stays safe and doesn't get mutated.

The Breakthrough: CODE-HB

The researchers asked: Can we trick the B-cell into using this "mutation factory" for any protein we want, not just antibodies?

They built a system called CODE-HB (Continuous Directed Evolution in Human B-cells). Here is how it works, step-by-step:

1. The Safe House (The Locus): They used a tool called CRISPR (like molecular scissors) to cut a specific, safe spot in the B-cell's DNA (the "H11 locus"). This is like building a new workshop in a safe, stable part of the city that won't cause the whole city to collapse.
2. The "Mutation Magnet" (The ProXIV Sequence): They took a special DNA sequence from the natural antibody factory (the "ProXIV" sequence) and attached it to the front of their new gene. This sequence acts like a magnet. It attracts the cell's mutation machinery (the "error-makers") to this specific spot.
3. The Test Subject (eGFP): First, they tested this with a glowing protein called eGFP. They broke the protein so it wouldn't glow (eGFP*). Then, they let the cell run. Because of the "magnet," the cell started making random mistakes in the broken protein. Eventually, by pure chance, one of those mistakes fixed the protein, and the cell started glowing again.
   • Analogy: Imagine a broken flashlight. You shake it violently (mutation). Most of the time, it stays broken. But sometimes, a loose wire gets jiggled back into place, and suddenly—flash—it works again.

The Real-World Application: Fighting Bird Flu

Once they proved the system worked, they used it for something important: Antibodies against Bird Flu (H5N1).

1. The Surface Display: They engineered the B-cells to stick the "keys" (antibody fragments) on their outer skin, like flags on a ship.
2. The Selection: They introduced the "lock" (the Bird Flu virus protein). The cells that had mutated their "keys" to fit the lock better would stick to the virus more tightly.
3. The Sorting: They used a machine (FACS) to find the cells with the best "keys" and let only those cells multiply.
4. The Result: After a few rounds of this, they found new antibodies that were much better at grabbing the Bird Flu virus than the original ones.

Why This is a Big Deal

Most other methods for evolving proteins happen in bacteria or yeast, or they use viruses to force the evolution. This paper is special because:

• It's Virus-Free: They didn't need to infect the cells with a virus to make it work.
• It's in Human Cells: This is crucial because proteins often fold differently in human cells than in bacteria. If you evolve a drug in bacteria, it might not work in humans. Evolving it directly in human cells ensures it will work better in real life.
• It's "Broad": Unlike other methods that mostly just swap single letters (substitutions), this system creates deletions (cutting out parts) and insertions (adding new parts) naturally. It's like not just changing a letter in a word, but sometimes deleting a whole word or adding a new sentence. This allows for much more creative solutions.

The Bottom Line

The researchers have built a "living laboratory" inside human cells. Instead of manually designing new drugs or antibodies, they can now set up a system where the cells do the heavy lifting, constantly trying out millions of variations until they find the perfect one. It's like giving the cell a turbo-charged engine for innovation, allowing us to rapidly design better medicines to fight diseases like Bird Flu.

Technical Summary

1. Problem Statement

Directed evolution is a powerful method for engineering biomolecules with desired functions, but traditional approaches face significant bottlenecks:

• Inefficiency of Natural Mutation: Natural mutation rates in human cells are too low ($10^{-9}$ to $10^{-10}$ per base) to allow for rapid evolution within a reasonable timeframe.
• Labor-Intensive Iteration: Standard directed evolution relies on in vitro diversification (e.g., error-prone PCR, DNA shuffling) followed by screening, requiring repeated cycles of library generation and selection.
• Limitations of Existing Continuous Evolution Platforms:
  • Bacterial/Yeast Systems: While platforms like OrthoRep exist in model organisms, they are not directly applicable to human proteins that require specific mammalian post-translational modifications or folding environments.
  • Virus-Based Mammalian Systems: Many continuous evolution methods in mammalian cells rely on viral vectors (e.g., VEGAS, PACE). These often suffer from low integration efficiency, genomic instability, and the inability to sustain stable, long-term expression of the target gene.
  • Previous B-cell Attempts: Earlier attempts to use B-cell lines for evolution (e.g., integrating genes into immunoglobulin loci) were hindered by low integration frequencies, random genomic insertion, and instability of the immunoglobulin loci themselves.

The authors sought to develop a virus-free, continuous directed evolution platform in human B cells that leverages the cell's natural machinery to evolve proteins rapidly without compromising genomic fitness.

2. Methodology: CODE-HB Platform

The authors developed CODE-HB (Continuous Directed Evolution in Human B cell lines), a platform that repurposes the natural Somatic Hypermutation (SHM) machinery of B cells.

• Genome Editing Strategy:

  • Instead of random viral integration, the team used a virus-free, CRISPR/Cas9-mediated homology-directed repair (HDR) approach.
  • They targeted the H11 "safe harbor" locus on human chromosome 22, a non-immunoglobulin site known for stability and high expression.
  • A two-plasmid system (Cas9/sgRNA + donor template) was used to integrate a specific cassette containing a reporter gene (eGFP or Fab) and a puromycin resistance marker.
  • Result: Stable integration in >90% of the RA1 human B cell line population, ensuring high-level, uniform expression.

• Recruiting SHM Machinery:

  • To induce mutations specifically at this stable locus, the team identified and inserted specific DNA sequences upstream of the reporter gene. These sequences were derived from the immunoglobulin heavy chain (IgH) locus (specifically regions upstream of IgHV4-34 and IgHV4-55).
  • These sequences (termed proXIV) contain transcription factor binding sites (e.g., Oct motifs) that recruit the B cell's SHM machinery, including Activation-Induced Cytidine Deaminase (AID).
  • This creates an "orthogonal" mutagenesis zone where the target gene undergoes high-frequency mutation while the rest of the genome remains stable.

• Surface Display Platform:

  • To evolve binding proteins (antibodies), the authors engineered a human B cell surface display system.
  • They fused antibody Fragment antigen-binding (Fab) regions to a membrane localization sequence and an MHC I transmembrane helix.
  • A FLAG tag was included to distinguish surface expression from antigen binding during flow cytometry (FACS).

• Evolution Workflow:

  1. Integration: Integrate the target gene (with proXIV enhancer) into the H11 locus.
  2. Passaging: Cultivate cells to allow SHM to accumulate mutations over time.
  3. Selection: Use FACS to sort cells based on improved phenotypes (e.g., restored fluorescence for eGFP, or increased antigen binding for Fabs).
  4. Expansion: Expand the sorted population for the next round of evolution.

3. Key Contributions

• First Virus-Free Continuous Evolution in Human Cells: CODE-HB eliminates the need for viral vectors, offering a stable, non-integrating (in the viral sense) system for continuous evolution in human B cells.
• Orthogonal SHM Recruitment: The discovery that specific upstream IgH sequences (proXIV) can recruit SHM machinery to a non-immunoglobulin "safe harbor" locus, enabling targeted mutagenesis without genomic instability.
• Broad Mutational Spectrum: Unlike many continuous evolution systems that rely on specific error-prone polymerases (often yielding mostly point mutations), CODE-HB recapitulates the natural SHM profile, generating a diverse mix of substitutions, insertions, and deletions (indels).
• Modular Surface Display: The development of a robust surface display platform compatible with human B cells, allowing for the direct evolution of therapeutic antibodies.

4. Key Results

• Reporter Protein Evolution (eGFP):

  • The team started with a non-fluorescent eGFP variant (eGFP*, containing a T66I mutation).
  • Within 5 passages, cells containing the proXIV-1 enhancer sequence showed a significant recovery of fluorescence, whereas controls without the enhancer did not.
  • Sequencing Analysis: PacBio long-read sequencing revealed a broad mutational profile. The primary reversion was T197C (restoring fluorescence), but the population also contained diverse substitutions, deletions, and insertions. The mutation rate was estimated at $5 \times 10^{-5}$ to $9 \times 10^{-5}$ substitutions per base per generation, intermediate between OrthoRep and PACE.

• Antibody Evolution (CR9114):

  • The platform was used to evolve the broadly neutralizing antibody CR9114 to improve binding to H5 avian influenza hemagglutinin.
  • Selection: Iterative FACS sorting based on H5 binding vs. surface expression enriched high-affinity binders.
  • Mutational Insights: Sequencing identified variants with improved binding. Notably, the most potent variant, W154R, contained a mutation in the heavy chain constant region (not the variable CDR loops).
  • Functional Validation: The W154R variant showed a 4-fold improvement in viral neutralization potency (HAI assay) against H1N1 and maintained binding to H5. Structural modeling suggested this mutation optimized the antibody's structural network, enhancing functional avidity.

• Evolution of Escape Variants (047-09_1A02):

  • The platform evolved an antibody (047-09_1A02) to bind a contemporary H1N1 strain (Michigan/2015) it previously bound poorly.
  • Unique Indel Discovery: The evolution process identified a 7-amino-acid deletion in the sc60 linker between the light and heavy chains, combined with point mutations. This deletion was crucial for improving binding, demonstrating the platform's unique ability to utilize indels for optimization—a feature often missed in other continuous evolution systems.

5. Significance and Impact

• Therapeutic Relevance: CODE-HB enables the rapid engineering of human antibodies directly in human cells, potentially bypassing issues where proteins misfold or lack modifications in bacterial/yeast systems. This is critical for developing therapeutics against emerging pathogens (e.g., H5N1 avian flu).
• Mechanistic Insight: The platform provides a tool to study SHM mechanisms in a controlled setting, revealing how specific DNA sequences recruit mutagenic machinery and how constant regions of antibodies can influence binding affinity.
• Broad Applicability: The system is modular and can be applied to evolve cytosolic proteins, membrane proteins, enzymes, and RNA molecules.
• Superior Mutational Diversity: By capturing the full spectrum of SHM (including indels), CODE-HB can access evolutionary trajectories that are inaccessible to systems restricted to point mutations, potentially leading to more optimal biomolecules.

In conclusion, CODE-HB represents a paradigm shift in directed evolution, moving from in vitro or viral-dependent methods to a stable, virus-free, human-cell-based system that leverages the immune system's natural evolutionary power for biotechnology.