AI is increasingly used to design novel proteins for drug discovery. However, the resulting proteins have never been expressed in cells and their large-scale production optimization may have unique challenges in achieving natural protein-like folding and functional expression.
In this blog, we examine how signal peptide engineering can address challenges in protein expression.
Missing the nature’s template
Natural proteins have undergone evolutionary selection for ensuring stability and expression in biological systems.
However, AI-designed proteins are based on computational predictions that may fail to fully account for cellular machinery constraints: de novo proteins may contain sequences that are difficult for cells to process, leading to misfolding, aggregation, or low yields.
Optimized selection of signal peptides can address several challenges related to production.
Signal peptides: Significant impact on folding, expression and post-translational modifications
Signal peptides direct synthesized proteins to the secretory pathway, which is essential for optimal protein folding, secretion and quality control. Their role extends beyond simply guiding proteins, as they also facilitate important processes within the ER lumen:
- Signal peptides can play a significant role in modulating glycosylation, disulfide bond formation, and other modifications, which are crucial for functional activity.
- Signal peptides can also regulate chaperone recruitment that is essential in correct protein folding. The interaction between signal peptides, chaperones, and translocation machinery ensures that proteins fold properly and avoid aggregation.
- Helps facilitate and rearrange disulfide bonds between cysteine residue which provides structural stability.
Look for the correct match among thousands, not dozens
Natural proteins produced for therapeutic purposes contain an established signal peptide that guides newly synthesized proteins to the endoplasmic reticulum and help regulate the protein’s folding, post-translational modification and expression rate. In industrial production, typically only a handful of signal peptides are tested during method optimization.
High throughput analysis of signal peptides allows the simultaneous testing of thousands of signal peptides to find the optimal match.
For AI-designed proteins, an experimental approach to choosing the optimal signal peptide is important: novel proteins do not have the nature’s template for optimal signal peptide selection nor dynamic production.
What are signal peptides?
Signal peptides are short sequences (typically 15-30 amino acids) found at the N-terminus of a newly synthesized protein. Signal peptides play a critical role in protein synthesis, folding, and cellular function beyond merely guiding proteins to their correct location.
Acting as a molecular “address tag,” signal peptides initiate the co-translational targeting of proteins to the endoplasmic reticulum (ER). This ensures that proteins destined for secretion, membrane integration, or organelle localization are properly processed.
There are over 100,000 signal peptides existing in nature and they express significant diversity. Advances in synthetic biology are enabling the design of engineered signal peptides to optimize protein production for industrial and therapeutic applications. However, signal peptides remain highly specific to their associated proteins: The reasons why a signal peptide effectively enhances the expression of one protein but does not do so for another are not yet fully understood.
