Base editors enable precise conversion of single nucleotides without causing DNA double-strand breaks. This new generation of gene editing technology can theoretically correct most known pathogenic single-base mutations in humans, showing tremendous potential for treating genetic diseases.
However, the deaminases in base editors not only modify target bases but may also modify other bases within the editing window, a phenomenon known as "bystander editing". This has raised significant concerns about the feasibility and safety of base editing technology in clinical applications. Although researchers have attempted various strategies to reduce bystander editing, including rational design of deaminases with narrower editing windows and improving guide RNA (gRNA), these methods often come at the cost of reduced editing efficiency and their effectiveness is highly dependent on sequence context.
Recently, George Church's team at Harvard University published a research paper titled "Engineered base editors with reduced bystander editing through directed evolution" in Nature Biotechnology. This study proposes a multi-pronged approach to minimize bystander editing by optimizing gRNA and deaminases, thereby improving base editing precision without sacrificing editing efficiency. The research establishes a scalable framework for precision engineering of base editors, addressing a major challenge in genome editing applications.
To fully realize the therapeutic potential of base editing, it is essential to develop robust and flexible methods for designing new base editing strategies that achieve both high efficiency and precision.
In this latest study, the research team proposed a multi-pronged approach to design novel base editing systems that maintain high efficiency in vitro while minimizing bystander editing. This strategy integrates three complementary technologies:
- Engineering gRNA to reduce bystander editing;
- Using phage-assisted non-continuous evolution (PANCE) to selectively evolve base editors with higher precision;
- Leveraging protein language models (PLM) for rational design of deaminases with optimized activity.
To engineer gRNA, the research team designed and tested a library containing approximately 60,000 different 3'-end extensions of sgRNA, known as anchor-guide RNA (agRNA), to improve the precision of adenine base editors (ABE).
The most promising agRNA candidates were selected from this library and then used as part of the phage-assisted non-continuous evolution (PANCE) system to evolve more precise TadA-8e enzymes. The team designed a clever screening circuit that links base editing activity to phage replication capability-only when base editing occurs precisely at the target site (rather than bystander editing) can the phage replicate effectively. This design creates a dual selection pressure that promotes the evolution of base editors toward greater precision, successfully identifying several variants with narrower editing windows.
Notably, the V28C variant obtained through PANCE evolution showed significantly improved editing efficiency at target sites while substantially reducing bystander editing. Analysis of editing patterns for approximately 12,000 pathogenic mutations showed that V28C's precision was about two to three times that of ABE8e at tested sites, with efficiency increased by approximately 20%.
Figure 1. Performance of V28C as an editing tool for correcting human pathogenic mutations. (Perrotta R M, et al., 2025)
To further optimize the precision and efficiency of base editors, the research team used protein language models (PLM) to predict mutations that could enhance the precision and efficiency of TadA-8e deaminase. Among the 21 predicted potentially beneficial mutations, the M151E mutation significantly narrowed the editing window while improving target site editing efficiency in experimental validation. Additionally, the research team conducted tests in two clinically relevant scenarios. Results showed that the V28C variant achieved efficient and highly precise editing of the cardiovascular disease-related target PCSK9 and the early-onset Parkinson's disease-related mutation SNCA E46K.
| Cat.No. | Product Name | Price |
|---|---|---|
| CLKO-0239 | PCSK9 KO Cell Lysate-HeLa | Inquiry |
| CLOE-1307 | Human PCSK9(His) HEK293 Cell Lysate | Inquiry |
| CLOE-1310 | Human PCSK9 HEK293 Cell Lysate | Inquiry |
| CLOE-2195 | Rat Pcsk9 (His) HEK293 Cell Lysate | Inquiry |
| CLOE-2807 | Mouse Pcsk9 (His) HEK293 Cell Lysate | Inquiry |
| CSC-DC011470 | Panoply™ Human PCSK9 Knockdown Stable Cell Line | Inquiry |
| CSC-RT0735 | Human PCSK9 Knockout Cell Line-HeLa | Inquiry |
| CSC-SC011470 | Panoply™ Human PCSK9 Over-expressing Stable Cell Line | Inquiry |
| AD11969Z | Human PCSK9 adenoviral particles | Inquiry |
| LV21261L | human PCSK9 (NM_174936) lentivirus particles | Inquiry |
In summary, by integrating gRNA engineering, directed evolution, and machine learning, this work provides a systematic strategy to improve base editing precision without compromising editing efficiency. These advances pave the way for safer and more effective therapeutic applications of base editing.
Reference
Perrotta R M, et al. Engineered base editors with reduced bystander editing through directed evolution. Nature Biotechnology, 2025: 1-10.