A New Generation of Adenine Base Editors Improves Editing in Primary Human Cells

By Susanna Bachle

Adenine base editors (ABE) mediate A•T-to-G•C base changes (Figure 1), but it can be challenging to make these base changes, especially in primary human cells. Now, scientists at Beam Therapeutics have found a way to improve editing in primary human cells (Gaudelli et al., 2020).

One of the widely used base editing systems, ABE7.10 (and the starting point for a new generation of ABEs), consists of 3 components: 

  • a deaminase (TadA, originally from E.coli, named TadA7.10 in ABE7.10) 
  • a catalytically impaired Cas protein (dCas or Cas nickase) 
  • a guide RNA that targets the complex of TadA and dCas to the genomic DNA of interest 
Figure ABE8 Beam blog post
Figure 1. When ABEs are in action they bind DNA so that the TadA domain can edit the short single-stranded DNA stretch that is accessible. Image from Gaudelli et al., 2020.

Directed evolution identifies adenine base editors with improved editing in cell lines

Nicole Gaudelli, the Associate Director and Head of Gene Editing Technology at Beam Therapeutics, used a previously developed bacterial selection strategy (Gaudelli et al., 2017) in combination with a synthetic library of TadA sequences to create the eighth generation of adenine base editors, called ABE8 (Gaudelli et al., 2020). The generated ABE8s needed to produce 3 concurrent base edits to withstand antibiotic selection. As a result, 8 mutations in TadA could create all three edits and were thus more efficient. Informed by these mutations and previously described TadA mutations (Gaudelli et al., 2017), they modified ABE7.10 accordingly which resulted in variants named “ABE8.x”: “8” stands for the eighth round of ABE evolution and the “x” is the placeholder for the number that identifies the mutation in the evolved TadA domain.

Because ABE7.10 contains a heterodimeric fusion of wild-type TadA and TadA7.10, they designed one set of ABE8’s similarly (ABE8.x-d). In order to assess the effect of removing the wild-type TadA from the complex the researchers developed a second set:

  • The ABE8.x-d version carries a fusion of wild-type TadA and the evolved TadA region.
  • The ABE8.x-m version contains only the evolved TadA and is ~500 base-pair smaller than ABE8.x-d. 

This resulted altogether in 40 new ABE8 variants. When comparing on-target DNA editing efficiency in a human cell line for these versions they found that the smaller version, ABE8.x-m, showed comparable editing behavior to ABE8.x-d. Overall, ABE8s showed a median increase of 1.94x editing activity compared to ABE7.10. Depending on the position in the protospacer, the editing capacity of the ABE8s ranged from 1.5x higher for positions A5-A7 to 3.2x higher for positions A3-A4 and A8-A10.

The researchers selected 8 of the tested ABE8 constructs (ABE8.8-m/d, ABE8.13-m/d, ABE8.17-m/d, ABE8.20-m/d) to evaluate further. 

The creation of accidental insertions and deletions (indel) by the catalytically impaired D10A nickase from ABE7.10 can be a problem. Therefore, the authors created ABE8 constructs using a catalytically “dead” S. pyogenes Cas9 (dC9-ABE8.x-m/d). The dC9-ABE8.x-m/d constructs demonstrated a 2.1x on-target DNA-editing efficiency in a human cell line while reducing indel frequency more than 90% compared to ABE7.10.

Choosing Cas9 proteins to increase the targeting scope of adenine base editors

The most commonly used S. pyogenes Cas9 requires the target site in the genomic DNA to contain a specific protospacer adjacent motif (PAM) made up of 5′-NGG-3′ (“N” can be any nucleotide base, followed by 2 guanines). This limits the use of ABEs as DNA sites that do not contain a suitable PAM cannot be edited efficiently. To expand the targeting scope of ABE8, the team replaced S. pyogenes Cas9 in the ABE complex with PAM-variant Cas9 proteins: the S. pyogenes NG-Cas9 (PAM: NG) to create NG-ABE8.xm/d and the S. aureus Cas9 (PAM: NNGRRT) to create Sa-ABE8.x-m/d. With these variants, ABE8 could target many additional locations in the genome. Compared to ABE7.10, NG-ABE8.xm/d showed a 1.6x median increase and Sa-ABE8.x-m/d a 2x median increase in editing frequency over ABE7.10. The ability to use non-standard Cas9 in the ABE8 complexes enables a broader targeting scope for base editing. 

Try ABE8 base editors for your research!

Base editing in primary human cells 

Cell therapy approaches can use base editors to correct mutations in stem cells and return the corrected cells back to the patients. For example, defects in hemoglobin (𝛽-globin), which cause sickle cell disease and 𝛽-thalassemia, can be mitigated with the expression of fetal hemoglobin. Previous studies have shown that one base change in the promoter region of gamma hemoglobin (HGB1/2) leads to continuous expression of fetal hemoglobin even in adult patients. In order to analyze the clinical use of ABE8 for primary human cell editing, the researchers tested ABE8 in CD34+ hematopoietic stem cells. ABE8 efficiently edited the fetal hemoglobin promoter and caused persistent expression of fetal hemoglobin to a greater level than ABE7.10.

In addition to stem cells the authors targeted 6 genes in primary human T cells using 8 selected ABE8s and ABE7.10. They quantified genomic editing by measuring the reduction in protein expression of these six genes. All ABE8s edited more efficiently than ABE7.10 with ABE8.20-m reducing protein expression the most in primary human T cells.

While editing single genes is very useful, sometimes, you need to modify multiple genes to generate a cellular phenotype. Again ABE8.20-m was the winner. It could simultaneously edit 3 genes (98.1%, 98.3%, or 98.6% efficiency, respectively) and outperformed ABE7.10 by at least 1.4 fold. The ability of ABE8.20-m to edit primary human T cells efficiently makes it a promising tool for T cell therapy development. 

Reducing off-target editing in ABE8s

While base editors are precise molecular tools, they do have undesired off-target editing effects. ABEs have been described to introduce base changes in cellular DNA and RNA dependent or independent of the gRNA used (Gaudelli et al., 2017). The authors tested ABE8.17-m and found that one base change mutation (ABE8.17-m+V106W) (Gaudelli et al., 2020) was able to reduce off-target RNA and gRNA-dependent DNA editing while maintaining on-target base editing functionality. 

The delivery method of base editors plays a role for off-target editing effects. ABEs can be introduced into cells as a plasmid or as mRNA constructs. In primary human cells mRNA delivery of ABEs resulted in more effective on-target editing and reduced off-target editing frequencies. For applications that require high DNA editing specificity, for example for potential therapeutic approaches, the authors recommend mRNA delivery, the V106W version of ABE8, and a thoughtful selection of guideRNAs. Of particular note, ABE8s reported in this work did not cause genome-wide, guide-independent off-target deamination, an important attribute for therapeutic application.

ABE8s demonstrate an overall improved base editing capacity, even at sites that used to be difficult to target. In particular, their ability for more robust editing in human cell lines and primary cells makes them a great addition to the base editing toolbox. 

Sharing science just got easier... Subscribe to the blog


Gaudelli NM, Komor AC, Rees HA, Packer MS, Badran AH, Bryson DI, Liu DR (2017) Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature 551:464–471 . https://doi.org/10.1038/nature24644

Gaudelli NM, Lam DK, Rees HA, Solá-Esteves NM, Barrera LA, Born DA, Edwards A, Gehrke JM, Lee S-J, Liquori AJ, Murray R, Packer MS, Rinaldi C, Slaymaker IM, Yen J, Young LE, Ciaramella G (2020) Directed evolution of adenine base editors with increased activity and therapeutic application. Nature Biotechnology. https://doi.org/10.1038/s41587-020-0491-6

Additional resources on the Addgene blog

Resources on Addgene.org

Topics: CRISPR, Base Editing

Leave a Comment

Sharing science just got easier... Subscribe to our blog