UGI relocation inside Cas9 reduces Cas9 dependent off target effects in cytosine base editors

The first generation of cytosine base editors (CBEs) was constructed through the fusion of cytidine deaminase enzymes with CRISPR/Cas9 to enable efficient and precise C-to-T conversions without inducing double-stranded DNA breaks^1,2,3,4. CBEs rely on cytidine deaminase for cytosine (C) deamination to uracil (U)^5,6, where U in DNA adopts thymine (T)-like base-pairing properties. However, the persistence of U within DNA recognized by uracil DNA glycosylase (UDG) as DNA damage, and further triggers base excision repair that reverts U:G intermediates to original C:G pairs, thereby compromising the C-to-T editing efficiency⁷.

To address this limitation, the second-generation CBEs incorporated a bacteriophage-derived uracil glycosylase inhibitor (UGI) fused to the carboxyl-terminus of Cas9 nickase (nCas9)¹. This UGI addition effectively suppresses endogenous UDG activity and enhanced C-to-T conversion efficiencies. Nevertheless, UGI addition at nCas9 terminus enhanced both on-target and off-target C-to-T editing simultaneously, ultimately exacerbating off-target effects¹.

We propose that integrating UGI within nCas9 could lead to UGI spatial reorganization, which might reduce UNG inhibition at off-target regions while preserve localized inhibition of UNG at targeted genomic regions, thereby reducing off-target DNA effects. Since integration of deaminase domains within nCas9 architecture preserved efficient on-target base editing activity while enhancing editing precision and specificity^8,9,10, investigating the impact of UGI relocation within nCas9 on off-target effects is highly valuable for CBE improvement. Here we extend this principle by systematically evaluating UGI relocation within distinct nCas9 internal regions compared to canonical C-terminal fusion. This exploration aims to identify optimal protein architectures that simultaneously retain high editing efficiency while mitigating Cas9-dependent off-target effects, potentially informing next-generation CBE engineering strategies.

To systematically identify optimal UGI integration sites within nCas9 that improve both on-target efficiency and editing specificity, we selected classical CBE variant YE1, known for its minimal Cas9-independent off-target activity, as the framework for UGI relocation analysis^11,12,13. By systematic insertion of UGI at 23 distinct positions within nCas9 described previously^9,10, we generated a series of YE1-UGI-X CBE variants (where X denotes the insertion site) (Fig. 1a). These constructs were subsequently evaluated in HEK293T cells using the endogenous HEK4 locus for on-target assessment and the well-characterized off-target (OT) locus OT2 for off-target profiling¹⁴.

After co-transfection of candidate CBE variants and target sgRNA for 72 h in HEK293T cells, double positive cells were collected by flow cytometry for on- and off-target editing analysis. It was revealed that 20 out of 23 YE1-UGI-X variants maintained robust on-target editing at HEK4 (> 50% C-to-T conversion frequencies) (Fig. 1b). Notably, 20/23 engineered variants exhibited significantly reduced Cas9-dependent off-target activity at OT2 locus compared to the classical YE1 CBE (Fig. 1b, Sup. Table 1).

It has been reported that split Cas9 led to reduced off-target cleavage¹⁵, so we added P2A linker upstream of UGI combining the features of split Cas9 and UGI relocation, and further examined whether P2A-UGI insertion inside nCas9 could similarly preserve on-target efficacy while reducing Cas9-dependent off-target effects (Fig. 1c). Although P2A-UGI would lead to the generation of two separate protein fragments, sgRNA would serve as scaffold to mediate structural complementation and functional assembly. Among the 23 engineered YE1-2A-UGI-X CBE variants, 16 constructs retained robust HEK4 on-target editing with > 50% C-to-T conversion frequencies (Fig. 1d). Crucially, 21/23 YE1-2A-UGI-X CBE variants exhibited significantly reduced Cas9-dependent off-target activity at HEK4 OT2 locus compared to the YE1-2A-UGI-C control (Fig. 1d, Sup. Table 2). Notably, the effective positions differed between the YE1-UGI-X and YE1-2A-UGI-X CBE variants (Fig. 1b, d), which may be attributed to the P2A-mediated generation of two separate protein fragments in YE1-2A-UGI-X, necessitating additional structural and functional assembly to achieve efficient editing at target sites.

Additionally, we included a no-UGI CBE control variant (YE1-no UGI) and observed significantly lower C-to-T conversion efficiency (average: 12.6%) compared to YE1-UGI-X and YE1-2A-UGI-X variants (averages: 16.3–91.7%; Fig. 1b, d, Fig. S1a–b, Sup. Table 3). Since UGI depletion is known to elevate C-to-A and C-to-G conversions^16,17, we quantified these non-C-to-T edits in our constructs. Both YE1-UGI-X and YE1-2A-UGI-X variants induced low frequencies of C-to-A (average: 0.02–2.1%) and C-to-G (average: 0.1–10.1%) conversions while YE1-no UGI induced significantly higher C-to-A and C-to-G conversion frequencies (average: 8.8% for C-to-A; average 36.3% for C-to-G) (Fig. S1a–b, Sup. Tables 3–4). Therefore, these results indicated that UGI is functionally intact in the CBE variants YE1-UGI-X and YE1-2A-UGI-X.

Above screening suggested that UGI relocation within nCas9 could enhance base editor specificity. According to the on- and off-target editing effects, five representative YE1-UGI-X variants (X = 113, 535, 1029, 1154, 1282) and five YE1-2A-UGI-X CBE variants (X = 312, 459, 715, 1154, 1282), demonstrating preserved editing activity with reduced Cas9-dependent off-target profiles were selected for further characterization.

To comprehensively evaluate the editing specificity of selected CBE variants, we performed systematic off-target profiling at nine validated off-target (OT) loci associated with EMX1, HBG1, and BCL11A targets^14,18. It was revealed that the classical YE1 CBE exhibited substantial off-target activity at three EMX1-associated OT sites (editing range: 6.44–19.62%) and one BCL11A OT site (4.39%). Strikingly, the engineered YE1-UGI-1282 variant showed dramatic reductions in off-target editing across all examined loci, achieving average efficiencies of 0.47% (EMX1-OT2), 4.65% (EMX1-OT3), 0.50% (EMX1-OT4), and 0.84% (BCL11A-OT1) at these sites (Fig. 2, Sup. Table 5).

Notably, the on-target/off-target selectivity ratios of YE1-UGI-1282 exhibited 37- to 104-fold improvements over the classical YE1 system. For EMX1-associated OT sites, the selectivity ratios reached 122.2 (vs. YE1: 3.3), 12.4 (vs. 6.1), and 115.3 (vs. 10.0). Similarly, at HBG1-related OT loci, ratios improved to 345.4 and 334.5 compared to YE1’s 128.4 and 132.1. Enhanced specificity was also observed at BCL11A OT sites with ratios of 84.6 (vs. 21.1) and 103.4 (vs. 76.5), demonstrating the broad-spectrum specificity enhancement achieved through UGI repositioning (Fig. 3).

To confirm the broader applicability of our engineering strategy, we investigated whether UGI integration at the 1282 locus of nCas9 could enhance specificity across diverse deaminase-derived CBEs beyond the YE1 system. We engineered two reprogrammed TadA8e variants: GC (V28G-N46C) and GGATY (V28G-A48G-I49A-V82T-N108Y), both optimized for exclusive C-to-T conversion with eliminated adenine base editing activity¹⁹. Comparative analysis revealed that canonical TadA8e(GC)- and TadA8e(GGATY)-CBEs exhibited substantial Cas9-dependent off-target effects at HEK4-OT2 (30.6% and 38.1%, respectively) and all three EMX1-associated OT loci (GC: 39.9%/5.43%/11.98%; GGATY: 30.2%/3.02%/6.57%). Remarkably, corresponding 1282-UGI-modified counterparts displayed reduced HEK4-OT2 off-target activity to 4.3% (TadA8e(GC)-UGI-1282) and 3.03% (TadA8e(GGATY)-UGI-1282), while EMX1-OT efficiencies decreased to 3.90%/0.08%/0.34% and 8.72%/1.72%/1.52%, respectively (Fig. 4, Sup. Table 6).

Then further characterization across five additional endogenous loci confirmed that the 1282-UGI relocation retained highly active on-target editing functionality comparable to classical C-terminal UGI placement (Fig. 5). Additonally, it was revealed that YE1-NO UGI induced the lowest C-to-T conversions as compared to selected CBE variants described in this study (Fig. 5, Sup. Table 7). However, YE1-NO UGI induced the highest C-to-A, and C-to-G conversion frequencies (an average of 0.5% to 70.7%) to CBEs variants reported in this study (all < 1%) (Fig. S2). These findings demonstrate that UGI relocation represents a universal strategy for specificity enhancement in CBEs, while maintaining efficient on-target editing activity critical for precision genome editing applications.

Multifaceted engineering strategies have been explored for the development of high-performance CBEs, including systematic screening and rational engineering of deaminase variants^{11,13,19,20,21,22,23,24,25,26,27,28}, internal integration of deaminase domains inside the nCas9 architecture^10,29, UGI dosage optimization through copy number modulation³⁰ and fusion of auxiliary protein domains³¹ to enhance editing precision or efficiency.

While existing engineering strategies have accelerated the evolution of CBE systems, notably achieving substantial mitigation of RNA and Cas9-independent DNA off-target effects, persistent challenges remain in addressing Cas9-dependent DNA off-target activity. Studies mainly focused on Cas9 engineering to generate high-fidelity Cas9 variants^32,33,34,35, which have been leveraged to reduce the Cas9-dependent DNA off-target activity of base editors. However, these engineered Cas9 variants generally led to reduced on-target binding capacity, thereby restricting the applications of derived CBEs. Our findings demonstrate that spatial relocation of UGI within the nCas9 architecture constitutes a critical determinant of editing precision, operating through mechanisms distinct from conventional UGI dosage optimization. Although the UGI relocation strategy is specific to CBE engineering, as UGI is not required for other precision genome editing tools such as ABEs and prime editors, the underlying concept of functional domain relocation is broadly applicable and could be extended to other functional domains. Through systematic screening, we identified multiple optimal integration sites within nCas9, with the strategically positioned 1282 locus emerging as a representative insertion site to minimize Cas9-dependent DNA off-target effects. This engineering paradigm expands the CBE optimization toolkit, providing complementary enhancement to current base editing platforms through architectural fine-tuning of different components.

Plasmid construction

The UGI coding sequence was integrated into designated nCas9 loci through BamHI/NheI (New England Biolabs, Ipswich, MA) dual restriction enzyme cloning. To generate insertion-compatible nCas9 backbones, site-directed mutagenesis was performed on the PX461 vector (Addgene #48140) using the KOD-Plus Mutagenesis strategy (Toyobo, Osaka, Japan; Cat# KOD-201). Cytidine deaminase variants were subsequently cloned into PvuI/BglII-linearized (New England Biolabs) nCas9-UGI constructs. For sgRNA expression vector assembly, the U6-sgRNA-EF1α-T2A-mCherry backbone was linearized with BsaI (New England Biolabs), followed by insertion of annealed oligonucleotides encoding target-specific guide sequences.

Cell culture and transfection procedure

HEK293T cells (GNHu17, Cell Bank of Chinese Academy of Sciences, Shanghai) were maintained in DMEM (Gibco, Waltham, MA; Cat# C11995500) supplemented with 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO; Cat# F0193) at 37 °C under 5% CO₂. Cells were seeded into 48-well plates (Corning, Corning, NY) 24 h prior to transfection. Plasmid DNA mixtures (base editor: sgRNA vector = 3:1 mass ratio) were complexed with PEI 25 K™ transfection reagent (Polysciences, Warrington, PA; Cat# 23966) in serum-free medium for 30 min at room temperature before adding into the wells. Cells were cultured for another 72 h and GFP/mCherry double-positive cells were isolated using a BD FACSAria™ Fusion cell sorter (BD Biosciences, San Jose, CA) for downstream analysis.

Base editing analysis

Sorted cells were lysed with DirectPCR reagent (Viagene Biotech, Ningbo, China; Cat# 302-C), followed by target amplification using KOD One™ PCR Master Mix (Toyobo; Cat# KMM-201). Amplicons were subjected to high-throughput sequencing (Illumina NovaSeq 6000) with paired-end 150 bp reads. Editing profiles were quantified using CRISPResso2 (v2.2.11) with default parameters.

1. Komor, A. C., Kim, Y. B., Packer, M. S., Zuris, J. A. & Liu, D. R. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature 533, 420 (2016).

   Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

2. Ma, Y. et al. Targeted AIDID-mediated mutagenesis (TAM) enables efficient genomic diversification in mammalian cells. Nat. Methods 13, 1029 (2016).

   Article  CAS  PubMed  Google Scholar 

3. Hess, G. T. et al. Directed evolution using dCas9-targeted somatic hypermutation in mammalian cells. Nat. Methods 13, 1036 (2016).

   Article  CAS  PubMed  PubMed Central  Google Scholar 

4. Nishida, K. et al. Targeted nucleotide editing using hybrid prokaryotic and vertebrate adaptive immune systems. Science 353, aaf8729 (2016).

   Article  PubMed  Google Scholar 

5. Pecori, R., Di Giorgio, S., Paulo Lorenzo, J. & Nina Papavasiliou, F. Functions and consequences of AID/APOBEC-mediated DNA and RNA deamination. Nat. Rev. Genet. 23, 505–518 (2022).

   Article  CAS  PubMed  PubMed Central  Google Scholar 

6. Conticello, S. G. The AID/APOBEC family of nucleic acid mutators. Genome Biol. 9, 229 (2008).

   Article  PubMed  PubMed Central  Google Scholar 

7. Kunz, C., Saito, Y. & Schar, P. DNA repair in mammalian cells: Mismatched repair: Variations on a theme. Cell. Mol. Life Sci. 66, 1021–1038 (2009).

   Article  CAS  PubMed  PubMed Central  Google Scholar 

8. Nguyen Tran, M. T. et al. Engineering domain-inlaid SaCas9 adenine base editors with reduced RNA off-targets and increased on-target DNA editing. Nat. Commun. 11, 4871 (2020).

   Article  ADS  PubMed  PubMed Central  Google Scholar 

9. Li, S. et al. Docking sites inside Cas9 for adenine base editing diversification and RNA off-target elimination. Nat. Commun. 11, 5827 (2020).

   Article  ADS  PubMed  PubMed Central  Google Scholar 

10. Liu, Y. et al. A Cas-embedding strategy for minimizing off-target effects of DNA base editors. Nat. Commun. 11, 6073 (2020).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

11. Kim, Y. B. et al. Increasing the genome-targeting scope and precision of base editing with engineered Cas9-cytidine deaminase fusions. Nat. Biotechnol. 35, 371 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

12. Doman, J. L., Raguram, A., Newby, G. A. & Liu, D. R. Evaluation and minimization of Cas9-independent off-target DNA editing by cytosine base editors. Nat. Biotechnol. 38, 620–628 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

13. Zuo, E. et al. A rationally engineered cytosine base editor retains high on-target activity while reducing both DNA and RNA off-target effects. Nat. Methods 17, 600–604 (2020).

    Article  CAS  PubMed  Google Scholar 

14. Tsai, S. Q. et al. GUIDE-seq enables genome-wide profiling of off-target cleavage by CRISPR-Cas nucleases. Nat. Biotechnol. 33, 187–197 (2015).

    Article  CAS  PubMed  Google Scholar 

15. Zetsche, B., Volz, S. E. & Zhang, F. A split-Cas9 architecture for inducible genome editing and transcription modulation. Nat. Biotechnol. 33, 139–142 (2015).

    Article  CAS  PubMed  Google Scholar 

16. Sakata, R. C. et al. Base editors for simultaneous introduction of C-to-T and A-to-G mutations. Nat Biotechnol 38, 865–869 (2020).

    Article  CAS  PubMed  Google Scholar 

17. Kurt, I. C. et al. CRISPR C-to-G base editors for inducing targeted DNA transversions in human cells. Nat. Biotechnol. 39, 41–46 (2021).

    Article  CAS  PubMed  Google Scholar 

18. Han, W. et al. Base editing of the HBG promoter induces potent fetal hemoglobin expression with no detectable off-target mutations in human HSCs. Cell Stem Cell 30, 1624–1639 (2023).

    Article  CAS  PubMed  Google Scholar 

19. Zhang, S. et al. TadA reprogramming to generate potent miniature base editors with high precision. Nat. Commun. 14, 413 (2023).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

20. Cheng, T. L. et al. Expanding C-T base editing toolkit with diversified cytidine deaminases. Nat. Commun. 10, 3612 (2019).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

21. Yu, Y. et al. Cytosine base editors with minimized unguided DNA and RNA off-target events and high on-target activity. Nat. Commun. 11, 2052 (2020).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

22. Koblan, L. W. et al. Improving cytidine and adenine base editors by expression optimization and ancestral reconstruction. Nat. Biotechnol. 36, 843 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

23. Rees, H. A., Wilson, C., Doman, J. L. & Liu, D. R. Analysis and minimization of cellular RNA editing by DNA adenine base editors. Sci. Adv. 5, eaax5717 (2019).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

24. Chen, L. et al. Re-engineering the adenine deaminase TadA-8e for efficient and specific CRISPR-based cytosine base editing. Nat. Biotechnol. 41, 663–672 (2023).

    Article  CAS  PubMed  Google Scholar 

25. Neugebauer, M. E. et al. Evolution of an adenine base editor into a small, efficient cytosine base editor with low off-target activity. Nat. Biotechnol. 41, 673–685 (2023).

    Article  CAS  PubMed  Google Scholar 

26. Lam, D. K. et al. Improved cytosine base editors generated from TadA variants. Nat. Biotechnol. 41, 686–697 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

27. Grunewald, J. et al. Transcriptome-wide off-target RNA editing induced by CRISPR-guided DNA base editors. Nature 569, 433–437 (2019).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

28. Zhou, C. et al. Off-target RNA mutation induced by DNA base editing and its elimination by mutagenesis. Nature 571, 275–278 (2019).

    Article  CAS  PubMed  Google Scholar 

29. Wang, Y., Zhou, L., Liu, N. & Yao, S. BE-PIGS: A base-editing tool with deaminases inlaid into Cas9 PI domain significantly expanded the editing scope. Signal Transduct. Target. Ther. 4, 36 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

30. Wang, L. et al. Enhanced base editing by co-expression of free uracil DNA glycosylase inhibitor. Cell Res. 27, 1289–1292 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

31. Komor, A. C. et al. Improved base excision repair inhibition and bacteriophage Mu Gam protein yields C:G-to-T: A base editors with higher efficiency and product purity. Sci. Adv. 3, eaao4774 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

32. Bravo, J. P. K. et al. Structural basis for mismatch surveillance by CRISPR-Cas9. Nature 603, 343–347 (2022).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

33. Slaymaker, I. M. et al. Rationally engineered Cas9 nucleases with improved specificity. Science 351, 84–88 (2016).

    Article  ADS  CAS  PubMed  Google Scholar 

34. Kleinstiver, B. P. et al. High-fidelity CRISPR-Cas9 nucleases with no detectable genome-wide off-target effects. Nature 529, 490–495 (2016).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

35. Chen, J. S. et al. Enhanced proofreading governs CRISPR-Cas9 targeting accuracy. Nature 550, 407–410 (2017).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Authors and Affiliations

1. Institute of Pediatrics, Children’s Hospital, Institutes for Translational Brain Research, State Key Laboratory of Medical Neurobiology, MOE Frontiers Center for Brain Science, Fudan University, Shanghai, 200032, China

   Zehao Shi & Tian-Lin Cheng

Contributions

T.L.C. conceived and designed the study. Z.H.S. executed molecular cloning and cellular experiments. Z.H.S. performed bioinformatics analysis. T.L.C. supervised all research phases and finalized the manuscript with input from all authors.

Corresponding author

Correspondence to Tian-Lin Cheng.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions