In-situ Target Base Editing Combining with Biosensor-driven Strategy Reveals Critical Single Nucleotide Variants for Enhanced Recombinant Protein Secretion in Pichia pastoris

This study presents a novel BINDER strategy combining dual-base editor-mediated in-situ genome engineering with nanobody-regulated biosensor-driven droplet sorting to identify critical single nucleotide variants that significantly enhance recombinant human serum albumin secretion in *Pichia pastoris*, achieving a record-breaking titer of 23.43 g/L.

The Gist

Imagine you are trying to bake the perfect loaf of bread, but your kitchen (the yeast cell) keeps messing up the recipe. You want to produce a specific, high-value ingredient (a recombinant protein, like human serum albumin) that the world desperately needs. The problem is, your yeast is slow, messy, and often throws away the finished product before it can leave the kitchen.

For years, scientists have tried to fix this by either rewriting the whole recipe from scratch or adding more ingredients. But sometimes, the solution isn't a massive overhaul; it's a tiny, almost invisible tweak to the oven's temperature or the kneading speed.

This paper describes a brilliant new "smart kitchen" strategy called BINDER that found those tiny, perfect tweaks in record time. Here is how it works, broken down into simple steps:

1. The Problem: The "Needle in a Haystack"

Scientists knew that changing just one single letter in the yeast's DNA (a "Single Nucleotide Variant" or SNV) could make a huge difference. But finding the right letter to change among billions of possibilities is like trying to find a specific needle in a haystack while wearing thick gloves.

• Old way: You'd randomly break things and hope something gets better. It takes forever.
• The new way: You need a way to test millions of tiny changes instantly.

2. The Tool: The "Genetic Pencil" (Base Editors)

Instead of cutting out big chunks of DNA (which is like using a chainsaw), the researchers used a Dual-Base Editor. Think of this as a high-tech pencil that can erase a single letter and write a new one without damaging the rest of the page.

• They used this pencil to make a massive library of 113,632 different yeast strains, each with a slightly different "typo" in its DNA.
• Crucially, this pencil was designed to be temporary. Once it did its job, it could be removed, leaving behind a clean, edited yeast cell.

3. The Detector: The "Magic Flashlight" (Biosensor)

Now they had a haystack of 113,000+ needles, but how do you find the ones that actually bake better bread?

• The Challenge: You can't just look at the yeast; you have to wait for them to secrete the protein, which is slow and hard to measure.
• The Solution: They built a Nanobody Biosensor. Imagine a "Magic Flashlight" that only turns on when it sees the specific protein you want.
  • They attached a "tag" to the protein the yeast makes.
  • They added a "sensor" (a nanobody) that acts like a dimmer switch. When the protein is present, the sensor flips the switch, and the protein glows brightly.
  • The Result: The more protein the yeast makes, the brighter it glows.

4. The Sorting Machine: The "Droplet Lottery"

They couldn't check 113,000 jars one by one. So, they used Microfluidics (tiny channels for fluids).

• They trapped each yeast cell in its own tiny water droplet (like a microscopic bubble).
• They added the "Magic Flashlight" sensor to the bubbles.
• A machine shot these bubbles through a laser. If a bubble glowed super bright, the machine zapped it with an electric field and sorted it into a "Winner's Circle."
• They did this at lightning speed: 3,000 bubbles per second.

5. The Discovery: The "Secret Ingredient"

From the 113,000+ candidates, the machine picked the top performers. The researchers found two "winning" mutations, but one was a superstar: HAC1_S224L.

• What did it do? It changed a single letter in a gene called HAC1. This gene is like the "manager" of the yeast's protein factory.
• The Analogy: Imagine the factory was clogged because the manager was panicking and shutting down the assembly line too often. This tiny mutation calmed the manager down, allowing the factory to run smoothly and efficiently without breaking down.
• The Result: The yeast with this mutation secreted 1.78 times more protein than the original strain.

6. The Grand Finale: The "Super-Bakery"

To prove this wasn't just a lab trick, they put the winning yeast into a giant 5-liter bioreactor (a massive industrial fermentation tank).

• They fed it the right nutrients and let it work for days.
• The Outcome: They produced 23.43 grams of protein per liter. This is the highest amount ever recorded for this type of yeast.

Why Does This Matter?

This isn't just about making one protein. It's about the method.

• Speed: They went from "no idea" to "super-producer" in just two months.
• Versatility: Because the "Magic Flashlight" sensor works on any protein (as long as you tag it), this strategy can be used to make insulin, vaccines, or biofuels in the future.
• Efficiency: They didn't need to guess; they let the data and the machine find the perfect solution.

In a nutshell: The researchers built a super-fast, automated system that uses a "genetic pencil" to make millions of tiny changes and a "glowing sensor" to instantly spot the winners. They found a tiny switch that turned a regular yeast factory into a world-record-breaking production line.

Technical Summary

1. Problem Statement

The production of recombinant proteins (r-proteins) in Pichia pastoris (now Komagataella phaffii) is hindered by a significant gap between production and demand, limiting the commercial viability of bulk protein products. While strategies like codon optimization and signal peptide engineering have improved titers, the highest reported secretions (17–20 g/L) remain insufficient for industrial scale.
The core challenges identified are:

• Genotype-Phenotype Disconnect: There is a lack of systematic linkage between specific genetic mutations and secretion phenotypes due to the complexity of metabolic networks.
• Limitations of Existing Tools: Traditional CRISPR-Cas9 approaches often rely on gene knockouts or large insertions, missing the potential of Single Nucleotide Variants (SNVs). Existing high-throughput screening methods (e.g., fluorescent reporters) are often limited to specific enzymes or require complex sensor preparation and struggle with eukaryotic membrane penetration.
• Library Size vs. Screening Throughput: Generating large mutant libraries is difficult in P. pastoris, and current screening methods cannot keep pace with the diversity required to find rare hyper-secretors.

2. Methodology: The BINDER Strategy

The authors developed a novel strategy named BINDER (Dual-Base editor-mediated Base editing, In-situ genome engineering, Nanobody-regulated Droplet sorting, Enhance R-protein secretion). The workflow consists of three main pillars:

A. Dual-Base Editor (DBE) Development

• Goal: To create an in-situ mutagenesis tool capable of generating diverse SNVs without the need for donor DNA libraries.
• Construction: The team engineered a Dual-Base Editor by fusing an Adenine Base Editor (ABE8e) and a Cytosine Base Editor (CDA-UGI) to a catalytically impaired Cas9 (nCas9).
• Delivery: The DBE system was delivered via an episomal plasmid to avoid genomic integration toxicity and allow for easy curing (elimination) of the editing machinery post-mutagenesis.
• Efficiency: The system achieved high editing efficiency (>0.5%) with an editing window of up to 15 nucleotides upstream of the PAM, enabling C>T and A>G conversions.

B. Nanobody-Regulated Fluorescent Protein (NbFP) Biosensor

• Goal: To develop a universal, high-sensitivity biosensor for detecting secreted r-proteins in droplet microfluidics.
• Mechanism:
  • The target protein (e.g., Human Serum Albumin, HSA) is fused with a nanobody (GBP1) that binds GFP.
  • A mixture of GFP and an inhibitory nanobody (GBP4) is added to the system. GBP4 suppresses GFP fluorescence.
  • When the secreted GBP1-tagged protein enters the extracellular space, it competitively binds GFP, displacing GBP4 and restoring fluorescence.
• Advantages: This system is independent of the target protein's enzymatic activity, has a wide dynamic range (ΔF/F0 ~21.9), and avoids the membrane penetration issues of previous chemical sensors (like FlAsH).

C. High-Throughput Droplet Sorting (FADS)

• Process: The mutant library was encapsulated in microdroplets (3,000 droplets/sec) containing the NbFP biosensor reagents.
• Sorting: Fluorescence-Activated Droplet Sorting (FADS) was used to isolate the top 0.1% of hyper-secretors (200 droplets/sec).
• Scale: The system screened a library of 113,632 mutants generated from 261 sgRNAs targeting three transcriptional regulators: Mxr1, Yap1, and Hac1.

3. Key Results

Identification of Critical SNVs

• From the screened library, two critical SNVs were identified that significantly enhanced secretion:
  1. HAC1_S224L (Mutation M3): A serine-to-leucine substitution in the transcription factor HAC1.
  2. YAP1_A44T (Mutation M9): An alanine-to-threonine substitution in YAP1.
• Performance: The HAC1_S224L mutant (M3) showed a 1.78-fold increase in rHSA secretion titer compared to the control, outperforming strains with wild-type HAC1 overexpression.

Mechanism Elucidation (Transcriptomics)

• RNA-seq analysis of the M3 strain revealed that the mutation modulates the Hsp70 co-chaperone cycle.
• While most HAC1-dependent genes (e.g., SSA1, KAR2, ERO1) were downregulated (suggesting reduced ER stress), ERJ5 (an ER-resident J-protein) was upregulated.
• Validation: Overexpression of ERJ5 using its native promoter improved secretion by 11%, whereas overexpression using the strong pAOX1 promoter reduced titers. This indicates that the HAC1_S224L mutation fine-tunes the Hsp70 cycle balance, preventing the toxicity associated with excessive chaperone expression.

Industrial Scale Production

• Fed-batch Fermentation: The engineered strain (M3-HSA combined with HAC1_S224L overexpression) was tested in a 5-L bioreactor.
• Outcome: The strain achieved a record-breaking rHSA titer of 23.43 g/L, a 2.75-fold improvement over the baseline strain.
• Yield: The yield reached 0.198 g HSA/g DCW (Dry Cell Weight).

Versatility

• The HAC1_S224L mutation was successfully applied to other proteins, including a SARS-CoV-2 nanobody fusion and phytase, resulting in 1.38-fold and 1.12-fold titer increases, respectively.

4. Significance and Impact

• Methodological Breakthrough: BINDER bridges the gap between high-throughput genome perturbation and phenotypic screening in P. pastoris, enabling the discovery of rare SNVs in a large genetic space (100k+ mutants) within two months.
• Novel Biosensor: The NbFP biosensor provides a universal, tag-based solution for screening secreted proteins, overcoming the limitations of enzyme-specific or membrane-impermeable sensors.
• Scientific Insight: The study reveals that fine-tuning the Hsp70 co-chaperone cycle (specifically via ERJ5 modulation) is more effective for secretion than simple overexpression of chaperones, offering a new paradigm for strain engineering.
• Industrial Application: Achieving >23 g/L of rHSA sets a new benchmark for P. pastoris productivity, demonstrating the potential for cost-effective, large-scale production of therapeutic proteins and bulk food/feed ingredients.
• Scalability: The strategy is transferable to other microbial species and r-proteins due to the modularity of the base editors and the protein-independent nature of the biosensor.