Ways to Boost Enzyme Expression: A Complete Guide in Five Layers

The expression level of a recombinant enzyme is the key economic factor that determines whether it can move from the lab to industrial applications. Whether it’s engineered enzymes in biopharmaceuticals, hydrolytic enzymes in industrial catalysis, or functional enzymes in synthetic biology, being able to produce enough of the enzyme is always the first hurdle for industrialization.

However, boosting enzyme expression isn’t something a single strategy can fix. Different enzyme sources, varying structural complexities, and host preferences—each variable can become a bottleneck for expression. This article systematically goes through a complete set of methods to increase enzyme expression from five aspects: host selection, gene and transcription optimization, translation and folding, fermentation processes, and AI-driven approaches.

1. Host Selection: Finding the Right “Production Workshop”

Multi-level strategies for enhancing enzyme expression

Choosing the host is the first-level decision in determining enzyme expression levels. Different hosts vary significantly in protein folding capability, post-translational modifications, secretion ability, and so on. There isn’t a 'best host,' only the 'most suitable host.'

E. coli is the most widely used prokaryotic expression system, with advantages like mature genetic manipulation, fast growth, and low cost. For enzymes with relatively simple structures, E. coli is usually the first choice. However, E. coli lacks complex post-translational modification capabilities, so for eukaryotic enzymes that need glycosylation or correct disulfide bond formation, inclusion bodies or insufficient activity can be a problem. For membrane proteins or toxic proteins, special strains like C41 (DE3) or C43 (DE3) can be tried. These strains use point mutations in the lacUV5 promoter to reduce the basal expression of T7 RNA polymerase, slow down transcription rates, mitigate the toxic effects of membrane/toxic proteins on the host, and lower the aggregation rate of nascent peptides, reducing inclusion body formation.

Komagataella phaffii (formerly Pichia pastoris) is a common eukaryotic expression host, combining the protein processing abilities of eukaryotes with the high-density fermentation traits of prokaryotes. For enzymes that need to be secreted or glycosylated, K. phaffii is often a better choice. Optimization strategies cover six dimensions: gene level, transcription level, translation level, protein folding and secretion, cell resistance, and fermentation process.

Bacillus subtilis is an important host for industrial enzyme production, especially suitable for secretion. Combining optimized promoters and signal peptides has achieved efficient secretion of various enzymes.

Saccharomyces cerevisiae performs well in expressing some bacterial proteases, yielding high-titer, highly soluble recombinant proteins without the risk of endotoxin contamination.

For certain enzymes, it’s often necessary to test multiple hosts in parallel to determine the optimal expression system.

2. Gene and Transcription Level Optimization: Boosting Expression from the Start

Once you've picked your host, the next big question is: how can you make the host cells "transcribe" the target gene more efficiently?

Codon optimization is the simplest and most effective method at the gene level. Different species have clear preferences for synonymous codons, and if your target gene has lots of rare host codons, ribosome translation can slow down. By swapping those codons for ones preferred by the host, you can significantly boost translation without changing the amino acid sequence. In fungal systems, codon optimization has steadily increased laccase production.

The bicistronic design (BCD) is a precise way to fine-tune protein expression. Researchers have developed a method using a BCD library for enzyme expression screening that can be optimized in just under 17 tests in two steps (expression profiling and focused screening). The BCD library covers a 992-fold range of expression, and protein abundance is measured via fluorescence. This strategy is currently mainly mature in prokaryotic systems.

Promoter engineering is key to transcription-level optimization. In Bacillus subtilis, screening strong native promoters revealed that PrapA, PmetE-1, and Phin-1 promoted alkaline pectinase expression 9.8, 4.8, and 3 times more than the commonly used P43 promoter, respectively. Plus, setting up a dual-promoter system can further ramp up expression.

3. Translation and Folding Level Optimization: Making Proteins "Born Right"

Representative molecular tools for improving soluble expression

The transcribed mRNA needs to be efficiently translated, and the resulting peptide chains need to fold correctly. These two steps often present bottlenecks that are trickier than transcription itself.

The transcribed mRNA needs to be efficiently translated, and the translated peptide chains need to fold correctly—the bottlenecks in these two stages are often more challenging than transcription.

Fusion tag technology is one of the most direct and effective approaches for improving soluble protein expression. Among the various fusion tags, SUMO stands out as one of the most prominent—under induction at 25℃, the soluble fraction of SUMO-tagged heparanase I increased significantly from (24.28 ± 1.63)% before fusion to (94.52 ± 3.99)% after fusion. Additionally, a few studies have reported that fusion with streptococcal protein G (SpG) can also enhance the solubility of certain specific enzymes.

Chaperone co-expression and folding pathway regulation constitute another core strategy for addressing folding bottlenecks. When the target protein is overexpressed, the host's own chaperone system often becomes overwhelmed, leading to misfolding of nascent peptide chains and formation of inclusion bodies. By co-expressing exogenous chaperones or regulating folding pathways, "folding assistants" can be provided for the target protein. Studies have shown that co-expression of the HAC1 transcription factor (activating the unfolded protein response pathway) significantly increased the yield of lipase LipA, reaching an enzyme activity of 3,233.2 U/mL, an increase of 47.6%. In Komagataella phaffii, the combined strategy of co-expressing the chaperone BiP and the UPR regulator HAC1 in a certain amylase expression case increased extracellular enzyme activity was increased by 602%, and this strategy is generally more effective for secretory proteins that are difficult to fold.

Protein rational design can also improve soluble expression at the sequence level. Through rational design strategies, replacing hydrophobic amino acids with basic amino acids (arginine and lysine) can significantly improve the soluble expression level of terminal deoxynucleotidyl transferase (TdT) in E. coli.

Secretion pathway engineering is particularly important for secretory enzymes. Through engineering synthetic signal peptides combined with strong promoters, secretory expression can be effectively enhanced. In Komagataella phaffii, knocking out the OCH1 gene can improve hyperglycosylation issues. Additionally, co-expressing vesicle transport factors can also improve secretion efficiency.

4. Optimizing Fermentation: Make Cells "Work More, Waste Less"

Tweaks at the gene and protein level solve the "can it be produced" question, while fermentation optimization tackles the "how much and how well" question.

High-cell-density fermentation (HCDC) is key for bumping up enzyme yields. By fine-tuning feed strategies to support steady cell growth, you can seriously increase enzyme output per volume.

Optimizing induction conditions directly affects how much of the target protein you get and its quality. Lowering induction temperature is a common way to cut down on inclusion body formation. Terminal deoxynucleotidyl transferase shows soluble expression when induced at a cool 15°C. For heparinase I, 25°C induction gives the best soluble expression with a SUMO fusion. Tweaking things like inducer concentration and induction time can also noticeably boost protein production.

Medium optimization matters too. In Bacillus subtilis, by adjusting the carbon source (25 g/L glycerol) and nitrogen source (20.75 g/L soy peptone), enzyme activity jumped from 19.08 U/mL to 80.36 U/mL — a 4.2-fold increase.

Controlling pH is a newer strategy. For alkaline-adapted enzymes, raising the expression medium pH to 9.0 and engineering E. coli to handle high pH improved the soluble activity of two alcohol dehydrogenases by 18.55-fold and 26.59-fold. This shows just how much pH affects protein folding and stability.

5. AI-Driven: From "Trial and Error" to "Global Optimization"

AI-driven global optimization paradigm

Traditional enzyme expression optimization often relies on a 'trial-and-error' approach—switching hosts, testing tags, adjusting conditions—each step requiring a lot of experimental validation. But AI technology is changing this model.

MatwingsVenus™ (XiaoWu™), a conversational protein R&D AI developed by Matwings Technology, provides a systematic solution for enzyme expression optimization. The platform leverages a dataset of tens of billions of protein sequences and over 200 protein design tools to directly predict the solubility, stability, and expression potential of a target protein from its sequence. Users just need to input their task goals in natural language, and the system will automatically break down the task, completing the full chain of R&D from sequence analysis and expression optimization design to candidate screening.

The platform's two core features, AI-directed evolution and AI enzyme mining, focus on different aspects of expression optimization: 'AI-directed evolution' can quickly predict beneficial mutation combinations based on a small amount of experimental data, greatly reducing the scale of experimental screening; 'AI enzyme mining' can directly discover naturally high-expressing enzymes from vast amounts of proteins with unknown functions.

On the industrial validation side, according to public reports, Matwings Technology has delivered more than 30 protein design projects. In collaboration with GenSci Pharmaceuticals, the platform improved protein alkalinity by 4 times, doubled its lifespan, and achieved successful 5,000-liter industrial production, saving the company over tens of millions of yuan annually.

Quick Reference for Common Issues in Increasing Enzyme Expression

Q1: With so many ways to boost expression, where should I start?

It's recommended to proceed by a "cost-benefit" gradient: first, optimize codons (lowest cost, most certain gain); second, test 2–3 fusion tags (SUMO, GST, MBP); third, if results are still unsatisfactory, consider changing hosts or co-expressing molecular chaperones. Solve the "can it be produced" problem first, then tackle "how much can be produced."

Q2: I did codon optimization, but expression still doesn’t improve. What’s wrong?

Codon optimization addresses whether the ribosome can translate smoothly. But if the protein tends to misfold, codon optimization won’t help. At this point, it’s better to first try fusion tags (SUMO tags have performed well in many studies) or reduce induction temperature (15–25°C). If that still doesn’t work, consider changing the host. Some enzymes simply don't fold well in E. coli, but switching to Pichia pastoris or Bacillus subtilis may yield surprises.

Q3: Which works best: SUMO, GST, or MBP tags?

There’s no absolute "best"—it depends on the enzyme’s properties. But SUMO tags have shown notable results: in heparinase I, solubility increased from about 24% to roughly 94%, and SUMO tags can be precisely removed by specific proteases without leaving extra residues. If conditions allow, it’s recommended to test 2–3 tags in parallel and choose the one that fits your enzyme best.

Q4: Can molecular chaperone co-expression and fusion tags be used together?

Yes, and they often have synergistic effects. Fusion tags improve the protein’s solubility, while chaperones enhance the cellular folding environment—they target different aspects and can work together. In Pichia pastoris, co-expressing BiP and HAC1 boosted extracellular enzyme activity of recombinant strains by 602%, proving the potential of combination strategies.

Q5: Should fermentation optimization come after genetic modification?

It’s usually suggested to first complete gene-level modifications to improve per-cell production, then optimize fermentation to increase biomass and production stability. But they’re not completely independent: gene changes like secretion signals and stress responses directly affect high-density fermentation adaptability, and process parameters like induction conditions and pH also significantly impact per-cell expression. In practice, iterative optimization between the two is common.

Q6: How can AI specifically help me increase enzyme expression?

AI can assist on three fronts: during design, AI models can directly generate sequences with high solubility and stability; during screening, AI-guided evolution can predict beneficial mutation combinations with minimal data, reducing experiments from tens of thousands of variants to just dozens; during discovery, AI can mine massive databases to find naturally high-expressing enzymes. Matwings Technology’s MatwingsVenus™ (Xiaowu™) platform integrates these capabilities into a closed loop, shortening traditional 2–5 year R&D cycles to 2–6 months.