top of page

Development of a New Platform for High-throughput Base Editor Engineering and Profiling

16 Jun 2023

A base editor is a next generation CRISPR-Cas9 genome editing tool that has great editing precision. However, choosing or designing the optimal base editor for a desired therapeutic or research application can be challenging. A research team from the Centre for Oncology and Immunology (COI) and the University of Hong Kong (HKU) have developed the first platform for engineering and comparing the performance of base editors en masse. Their research was published in Cell Systems and a patent application has been filed based on this work. 


The advent of CRISPR-Cas9 technologies has revolutionized scientific research, enabling scientists to selectively modify genes by creating double strand breaks in specific sections of the DNA and allowing endogenous repair mechanisms to re-join the broken ends. Despite its enormous potential, several hurdles have limited its therapeutic application, such as low editing efficiency and the introduction of unintended off-target effects, including chromosomal rearrangements, or large DNA insertions or deletions. To address these shortcomings, base editors were developed. Similar to traditional CRISPR-Cas9 systems, base editors consist of a single guide RNA (sgRNA) that serves to locate the target DNA and a Cas9 protein that cuts the DNA. However, the Cas9 protein in base editors is a mutant form referred to as Cas9 nickase (nCas9), which only cuts one of the DNA strands instead of both, reducing off target effects. In addition, nCas9 is fused with a deaminase protein that catalyses the generation of targeted point mutations in the DNA. Thus, the development of base editors represents an important step forward in precise gene editing technology, but more work remains before it can be used to treat human disease.

Base editor performance can vary depending on the target sequence, with each application requiring a distinct base editor variant to make a particular type of edit, at an exact location in the DNA, and in a specific genomic context. Therefore, selecting or engineering the best possible base editor variant for a particular purpose is important for maximizing editing efficiency and minimizing undesired effects. Currently, tens of base editor variants, each with differing editing efficiency and substrate preference, have been reported and tediously characterized one by one. Unfortunately, this small number of base editor variants is not enough to meet the demand for precise editing of a wide spectrum of target sequences. Hence, there is an unmet need to engineer base editor variants and evaluate their performance at scale. The COI/HKU research team has developed the first platform for creating and analysing a library of engineered base editor variants, allowing scientists to quantitatively measure each variant’s editing efficiency, purity, motif preference, bias in generating single and multiple base conversions, as well as the generation rate for insertions/deletions and non-canonical edits. Therefore, this constitutes a powerful tool by which a large number of base editor variants can be engineered and compared head-to-head in an unbiased way to benchmark their performance in a context specific manner.

Research Methods

Previously, the research team had developed a platform called CombiSEAL, a method for high-throughput assembly and barcoding of protein variants. In their new base editor engineering and profiling platform, they have incorporated CombiSEAL into their pipeline to mass assemble base editor variants and label each one with a reporter and a unique barcode. Using this system, each variant is linked to its editing performance at single-base resolution in single cells within the cell population, and can be tracked in a pooled analysis by high-throughput short-read sequencing.


As proof-of-principle, the team used the platform to generate a base editor variant library consisting of every possible combination of 22 different deaminase proteins, 15 of which were newly engineered by the research team, and 5 different Cas9 nickases, resulting in a total of 110 base editor variants for analysis. They then used lentiviruses to deliver the library into human cells in culture, such that the cells were infected with only a single copy of each variant. Finally, after 7 days in culture, the genomic DNA from the pooled cell population was extracted and deep sequencing was performed to reveal the editing outcome of each base editor variant. As a result, the team was able to make head-to-head comparisons between all 110 base editor variants in terms of quantitative measurements, such as editing efficiency, editing purity, motif preference, bias for generating single and multiple base conversions, and the rate at which insertions, deletions and non-canonical edits were generated. In addition, the team generated a library of 312 sgRNA scaffold variants with linked reporter sequence and barcode to identify the best performing scaffold for increasing base editor activity. Similarly, they delivered this library into human cells and performed deep sequencing to evaluate the head-to-head editing performance of each variant. Remarkably, they were able to identify an sgRNA scaffold variant that could more than double base editor activity at the CXCR4 gene, compared to a wild-type scaffold. These results demonstrate the utility of this platform for facilitating the selection of the most suitable base editors to achieve the desired genetic edits for a particular application.

Impact of the Study

Prior to the development of this platform, the existing methods for engineering and validating base editor performance were labour intensive and time consuming. Base editor variants were generated one or a few at a time and characterized individually. With this new platform developed by the COI/HKU research team, it is now possible to engineer and simultaneously evaluate hundreds of base editor variants in an unbiased, high-throughput manner. Dr Alan Siu Lun Wong explains, “It is like an accelerated check-out process in stores. Since all product items (i.e., base editor variants) are tagged with a barcode, when it comes to the check-out counter barcode scanner, we need to only put all items in bulk into the basket at the check-out counter. The scanner can automatically identify all items and complete the payment (i.e., base editing performance analysis in our case). There is no need to individually test each base editor one-by-one.”

Thus, this platform raises the current capability to engineer and characterise base editors in a pooled format at single-base resolution, drastically reducing the time, cost and efforts required during the process of optimizing, profiling, and selecting the most efficient variant for a particular target sequence. Moreover, this platform has the potential to be adapted for high-throughput engineering and profiling of prime editors, another next generation CRISPR-Cas9 genome editing tool that has been gaining attention recently. Overall, this platform offers a powerful one-pot solution for obtaining the most precise genome editor with the least undesired effects for a given application.


About the research team

The research was led by Dr Alan Siu Lun Wong, a principal investigator at the Centre for Oncology and Immunology, Ltd. and an associate professor at the School of Biomedical Sciences, The University of Hong Kong. The first author was Mr John H.C. Fong, a PHD student, and other key researchers included Dr Hoi Yee Chu and Dr Peng Zhou, post-doctoral fellows in Dr Alan Wong’s lab.

bottom of page