Structure-oriented mining cytosine deaminase with high editing activity and no sequence preference and application of structure-oriented mining cytosine deaminase

Abstract

The invention provides structure-oriented excavated cytosine deaminase with high editing activity and no sequence preference and application of the structure-oriented excavated cytosine deaminase. The invention discloses a series of novel cytosine deaminase. The cytosine deaminase disclosed by the invention has enhanced targeted editing performance; the enhanced targeted editing performance includes improved sequence-free preference, improved editing efficiency, reduced off-target effect, and improved targeted specificity.

Description

Technical Field

[0001] The present invention belongs to the field of molecular biology, and more specifically, the present invention relates to structure-guided mined cytosine deaminase with high editing activity and no sequence preference and its application. Background Art

[0002] Mining new functional proteins is of great significance in both biomedical and agricultural applications. Traditional mining of new functional proteins mainly relies on the primary sequence information of protein functions, and new proteins are screened by sequence conservation comparison and classification. Protein primary structure sequence alignment is an important method for mining new functional proteins. However, this method ignores the three-dimensional structure of proteins, which is a decisive factor affecting protein function, thereby reducing the efficiency and accuracy of screening. Mining new proteins with the knowledge of the three-dimensional structure of proteins will greatly improve work efficiency. Previously, protein three-dimensional structure information was mainly obtained through methods such as cryo-electron microscopy, X-rays and nuclear magnetic resonance, which require a lot of manpower and time costs and are not suitable for large-scale protein mining. Artificial intelligence technologies represented by AlphaFold2 developed in recent years can quickly predict the 3D structure of a given protein sequence, which provides an important way to mine new functional proteins.

[0003] Although cytosine base editors (CBEs) have great application potential in biomedicine and agricultural breeding, their sequence preference for target sites and high off-target effects limit their further application. Cytosine deaminases, such as apolipoprotein B mRNA catalytic editing protease (APOBEC) and activation-induced cytosine deaminase (AID), are gene editing enzymes that catalyze the deamination of C bases to U bases in genomic DNA and induce single base substitutions. Cytosine deaminases such as rAPOBEC1, hA3A and hAID have been developed for various cytosine base editor systems. Although systematically optimized deaminases have shown more excellent properties through various sequence engineering evolutions, these optimized deaminases still have defects in some aspects of editing properties due to the limitations of the original basic sequences, such as the inability to take into account both high editing efficiency and low off-target effects, the existence of extensive indels effects, and low C-to-T editing efficiency and editing purity. More importantly, existing editors all have obvious sequence preferences, and no deaminase has been found that has no complete preference for any NC (AC / TC / CC / GC), which greatly limits the editing efficiency of CBE at certain important sites.

[0004] Therefore, discovering cytosine deaminases with new functional properties is of great significance for upgrading and expanding the editing capabilities of CBE. Summary of the invention

[0005] The object of the present invention is to provide structure-guided mining of cytosine deaminase with high editing activity and no sequence preference and its application.

[0006] In a first aspect of the present invention, a protein or a functional conservative variant thereof is provided for use as a cytosine deaminase or for preparing a cytosine base editor; wherein the protein is selected from the group consisting of: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD590, CD0058, and CD0149.

[0007] In another aspect of the present invention, a method for preparing a pyrimidine base editor is provided, comprising using a protein selected from the following group or a functional conservative variant thereof as a cytosine deaminase: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD0590, CD0058, and CD0149.

[0008] In one or more embodiments, the protein or its functional conservative variant has enhanced targeted editing performance; the enhanced targeted editing performance includes: improved sequence-free preference (i.e., improved editing versatility / extensiveness / broad spectrum), improved editing efficiency, reduced off-target effects, and improved targeting specificity.

[0009] In one or more embodiments, the cytosine deaminase or a cytosine base editor containing the same is used for C-to-T gene editing; preferably, the C-to-T gene editing includes: introducing a C-to-T mutation into a target site of a target gene in a cell, introducing a stop codon, mutating a target promoter, or introducing a variable splicing site.

[0010] In one or more embodiments, CD0208 is a protein having an amino acid sequence as shown in SEQ ID NO: 1; preferably, the functional conservative variant of CD0208 is based on the amino acid sequence as shown in SEQ ID NO: 1, having a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A.

[0011] In one or more embodiments, each mutant is as follows:

[0012] The CD0640 is a protein having an amino acid sequence shown in SEQ ID NO: 2;

[0013] The CD0181 is a protein having an amino acid sequence shown in SEQ ID NO: 3;

[0014] The CD0418 is a protein having an amino acid sequence shown in SEQ ID NO: 4;

[0015] The CD0458 is a protein having an amino acid sequence shown in SEQ ID NO: 5;

[0016] The CD0730 is a protein having an amino acid sequence shown in SEQ ID NO: 6;

[0017] The CD0902 is a protein having an amino acid sequence shown in SEQ ID NO: 7;

[0018] The CD0911 is a protein having an amino acid sequence shown in SEQ ID NO: 8;

[0019] The CD0755 is a protein having an amino acid sequence shown in SEQ ID NO: 9;

[0020] The CD0526 is a protein having an amino acid sequence shown in SEQ ID NO: 10;

[0021] The CD0085 is a protein having an amino acid sequence shown in SEQ ID NO: 11;

[0022] The CD0827 is a protein having an amino acid sequence shown in SEQ ID NO: 12;

[0023] The CD0236 is a protein having an amino acid sequence shown in SEQ ID NO: 13;

[0024] The CD0956 is a protein having an amino acid sequence shown in SEQ ID NO: 14;

[0025] The CD0931 is a protein having an amino acid sequence shown in SEQ ID NO: 15;

[0026] The CD0054 is a protein having an amino acid sequence shown in SEQ ID NO: 16;

[0027] The CD0701 is a protein having an amino acid sequence shown in SEQ ID NO: 17;

[0028] The CD0590 is a protein having an amino acid sequence shown in SEQ ID NO: 18;

[0029] The CD0058 is a protein having an amino acid sequence shown in SEQ ID NO: 19; or

[0030] The CD0149 is a protein having the amino acid sequence shown in SEQ ID NO: 20.

[0031] In another aspect of the present invention, a cytosine deaminase variant is provided, which is a functional conservative variant of CD0208, and based on the amino acid sequence shown in SEQ ID NO: 1, it has a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A; preferably, the functional conservative variant of CD0208 has enhanced targeted editing performance; the enhanced targeted editing performance includes: improved sequence-free preference, improved editing efficiency, reduced off-target effects, and improved targeting specificity.

[0032] In another aspect of the present invention, a cytosine base editor is provided, which includes a cytosine deaminase; the cytosine deaminase is selected from the following group proteins or their functional conservative variants: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD0590, CD0058, CD0149; preferably, the cytosine deaminase is CD0208, and its functional conservative variant is based on the amino acid sequence shown in SEQ ID NO: 1, having a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A.

[0033] In one or more embodiments, the cytosine base editor comprises the following elements connected in an operative sequence: cytosine deaminase-Cas enzyme; preferably, the N-terminus of the cytosine deaminase and / or the C-terminus of the Cas enzyme also includes: a nuclear localization signal (NLS).

[0034] In one or more embodiments, a linker sequence is also included between the cytosine deaminase and the Cas enzyme, between the cytosine base editor and the nuclear localization sequence (NLS), and between the Cas enzyme and the nuclear localization sequence.

[0035] In one or more embodiments, the C-terminus of the Cas enzyme (preferably between the Cas enzyme and the nuclear localization signal) also includes a uracil DNA glycosylase inhibitor (UGI).

[0036] In one or more embodiments, the cytosine base editor is encoded by the following construct: promoter (such as CMV)-NLS-cytosine deaminase-(Linker or none)-Cas enzyme-UGI-NLS-pA.

[0037] In one or more embodiments, the construct encoding the cytosine base editor is also operably linked to a construct encoding a reporter gene protein for tracing.

[0038] In one or more embodiments, the construct encoding the cytosine base editor (or also linked to a construct encoding a reporter gene protein) is contained in pCMV.

[0039] In one or more embodiments, the sgRNA that cooperates with the cytosine base editor is introduced into the cell by a viral transfection system.

[0040] In one or more embodiments, the Cas enzyme includes Cas9 enzyme and Cas12 enzyme.

[0041] In one or more embodiments, the Cas9 enzyme includes a mutant of Cas9; preferably, the mutant of Cas9 includes nCas9 or dCas9.

[0042] In one or more embodiments, the Cas9 enzyme includes: a mutated inactivated Cpf1 (dCpf1, also called dCas12a).

[0043] In another aspect of the present invention, an isolated polynucleotide or a recombinant expression vector containing the polynucleotide is provided; wherein the polynucleotide encodes the cytosine base editor.

[0044] In another aspect of the present invention, a genetically engineered host cell is provided, which contains the aforementioned recombinant expression vector, or has the aforementioned isolated polynucleotide integrated into its genome.

[0045] In another aspect of the present invention, the use of the cytosine base editor is provided for performing C-to-T gene editing (including gene editing with the main purpose of C-to-T); or for preparing a reagent or kit for performing C-to-T gene editing.

[0046] In one or more embodiments, the cytosine base editor has enhanced targeted editing performance; the enhanced targeted editing performance includes: improved sequence-free preference (i.e., improved editing versatility / extensivity / broad spectrum), improved editing efficiency, reduced off-target effects, and improved targeting specificity.

[0047] In one or more embodiments, the use of the cytosine base editor for gene editing is an in vitro method for non-living organisms.

[0048] In one or more embodiments, the subject of gene editing using the cytosine base editor will not develop into a living organism or is a human embryo within 14 days of fertilization that has not undergone in vivo development.

[0049] In one or more embodiments, the use of the cytosine base editor for gene editing is a cellular level method, including single cells, dual cells or multiple cells.

[0050] In one or more embodiments, the use of the cytosine base editor for gene editing is not for the direct purpose of disease diagnosis.

[0051] In another aspect of the present invention, a method for C-to-T gene editing (including gene editing with the main purpose of C-to-T) is provided, comprising mediating gene editing with the cytosine base editor; preferably, the nucleic acid sequence encoding the cytosine base editor and the sgRNA are co-transfected or injected into a receptor to perform gene editing; preferably, the receptor includes: somatic cells or germ cells; preferably, the germ cells include embryonic cells or fertilized eggs.

[0052] In one or more embodiments, the gene editing method is an in vitro method for a non-living organism.

[0053] In one or more embodiments, the subject of the gene editing method will not develop into a living organism.

[0054] In one or more embodiments, the gene editing method is a cellular level method, including single cell, dual cell or multi-cell.

[0055] In one or more embodiments, the gene editing method is not used for the direct purpose of disease diagnosis.

[0056] In another aspect of the present invention, a reagent or kit for gene editing is provided, which contains any of the base editors described above; or, contains a polynucleotide encoding the base editor; or, contains a recombinant expression vector of the polynucleotide.

[0057] Other aspects of the present invention will be apparent to those skilled in the art in view of the disclosure herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 , Clustering of cytidine deaminases based on three-dimensional protein structure. a, Workflow using 3D protein structure prediction methods to mine potential cytosine deaminases with base editing application value. Candidate cytosine deaminases sequences were obtained based on homology sequences of the cytosine deaminase catalytic domain, and clustered according to the three-dimensional structural similarity of the predicted candidate cytosine deaminases. The target sites of the sgRNA-target library were then sequenced by Illumina NovaSeq, and the editing properties of each cluster deaminase were characterized in the endogenous sites of the cell. b, Hierarchical clustering of cytosine deaminases based on structural differences. The red and yellow colors in the heat map indicate the degree of structural similarity. The gradient from green to white indicates the cluster number. c, Phylogenetic analysis of cytosine deaminases based on predicted three-dimensional structures. Green represents 272 deaminases selected from 184 clusters, and red plus green represents a total of 1483 candidate deaminases.

[0059] Figure 2 , characterization of the editing properties of cytosine deaminases. a, Experimental strategy for high-throughput identification of the editing properties of candidate cytosine deaminases using sgRNA-target libraries. b, Average editing efficiency of 272 representative cytosine deaminases in 102 sgRNA-target libraries. The 102 sgRNA-target sites are arranged along the abscissa, and the 272 candidate deaminases are arranged along the ordinate. The red-yellow heat map gradient represents the editing efficiency. The #132, #145, and #147 clusters to which the cytosine deaminases belong are marked in red. c, The off-target effects of candidate cytosine deaminases were evaluated using the R-loop method in HEK293T cells. Each point represents the average editing efficiency of the 102 sgRNA-target sites (y-axis) and the average off-target effect of the four R-loop off-target sites (x-axis). The eight deaminases with the highest editing efficiency and the control deaminases (hA3A, rAPOBEC1, YE1, and eA3A) are marked in uppercase. d, Analysis of editing window mediated by CD. The red numbers indicate the editing efficiency of the site with the highest CD editing mediated by CD, and the red box is the editing window. e, Average editing efficiency of 10 CDs selected from cluster #147 in 102 sgRNA-target libraries. The 102 sgRNA-target sites are arranged along the horizontal axis, and the 10 candidate deaminases are shown on the vertical axis. The red-yellow heat map gradient indicates the editing efficiency. be, Data represent the average of four independent experiments, except Figure 2c The R-loop method was used to evaluate the off-target activity of candidate deaminases in HEK293T cells (n=3). The sgRNAs used are as follows:

[0060]

[0061]

[0062]

[0063] Figure 3 , high editing efficiency, sequence-free cytosine deaminases and their engineering modification to reduce off-target effects. a, Illumina NovaSeq sequencing analysis of 8 highly active cytosine deaminases ( Figure 2 ) and 4 well-characterized deaminases (hA3A, rAPOBEC1, YE1, and eA3A) in 11,868 sgRNA-target libraries. The middle line is the median, the bottom and top lines of the box represent the first and second quartiles, respectively, and the tails extend to the minimum and maximum values. b, Sequence preference analysis of 8 highly active cytosine deaminases ( Figure 2 c). c, Sequence preference of CD0208 CBE in 34 endogenous targets in HEK293T cells. d, Editing efficiency of CD0208 in the 2-7nt editing window of 8 representative endogenous targets in HEK293T cells. The data are the average of three independent experiments. e, Amino acid residues predicted to be involved in ssDNA interactions in the three-dimensional structure of CD0208. f, Detection of on-target and off-target editing efficiencies of CD0208 mutants. Each point represents the average editing efficiency of 11,868 sgRNA-target sites (y-axis) and the average off-target effect of 4 R-loop sites (x-axis). CD0208 P52A Variants and controls (hA3A, rAPOBEC1, YE1 and eA3A) are marked in large fonts. g, The on-target to off-target ratios of CD0208 variants calculated according to (f). rAPOBEC1, YE1 and eA3A were used as controls. CD0208 was selected P52A Variants (red arrows) were further evaluated. h, CD0208 detected in 11,868 sgRNA-target libraries P52A Sequence preference. rAPOBEC1, YE1, eA3A and CD0208 were used as controls. i. In the 11868 sgRNA-target library, CD0208 P52A, CD0208 and four well-characterized deaminases (hA3A, rAPOBEC1, YE1 and eA3A). The editing windows are shown from left to right in the horizontal axis. j, CD0208 in the 11868 sgRNA-target library P52A Distribution of editing types for a, CD0208, and four well-characterized deaminases (hA3A, rAPOBEC1, YE1, and eA3A). The numbers in each cell indicate the proportion of a certain editing type to all editing types. The error bars in b, h, and i represent the mean ± SE of the average editing efficiency of 11,868 sgRNA targets. The error bars in c represent the mean ± SE of three independent experiments. Examples of sgRNAs used are as follows:

[0064]

[0065]

[0066] Figure 4 、CD0208 P52A CBE introduces nonsense mutations in single-copy and multi-copy genes in mammalian cell lines. a, rAPOBEC1, YE1, hA3A, eA3A, CD0208, and CD0208 P52A C-to-T editing efficiency of CBEs at 11 target sites of Tyr genes in mouse N2A cells. b, Average C-to-T editing efficiency of each CBE in the editing window of the target site in N2A cells. c, Summary of editing efficiencies in a. d, Efficiency of each CBE in inducing stop codons at 11 target sites of Tyr genes in N2A cells. e, C-to-T editing efficiency of 8 target sites of 3 multi-copy genes (Rbmy1a1, Ssty1 and Ssty2) on the Y chromosome of mouse mESCs. f, Average C-to-T editing efficiency of each CBE within the editing window of 8 target sites. g, Summary of editing efficiencies in f. h, rAPOBEC1, YE1 and CD0208 P52A Efficiency of CBEs in introducing stop codons at eight target sites of Rbmy1a1, Ssty1, and Ssty2 genes in mESCs. i, rAPOBEC1, CD0208, and CD0208 P52A C-to-T editing efficiency of CBEs at 9 target sites of PERV pol gene in PK-15 cells. j, Average cytosine substitution efficiency of each CBE at the target site within the editing window of PK-15 cells. k, Summary of editing efficiency in i. l, rAPOBEC1, CD0208 and CD0208 P52AEfficiency of CBEs in introducing stop codons at 9 target sites of PERV pol gene in PK-15 cells. Error bars in a, b, e, f, i, and j represent the mean ± SE of three or more independent experiments. The center line in c, d, g, h, k, and l is the median, and the bottom and top lines of the boxes are the first and second quartiles of the editing efficiency obtained from three or more independent experiments, respectively. The tail extends to the minimum and maximum values. All P values ​​were calculated using a two-sided Wilcoxon rank sum test. The sgRNAs used are as follows:

[0067]

[0068]

[0069] Figure 5 、CD0208 P52A -dCpf1 CBE construction and base editing efficiency detection. a, Schematic diagram of the pCMV-dCpf1-CBE and pU6-sgRNA-EGFP plasmid structures. b, rAPOBEC1-dCpf1, YE1-dCpf1, hA3A-dCpf1, CD0208-dCpf1 and CD0208 P52A -d C-to-T editing efficiency of Cpf1 CBEs at 13 endogenous sites in HEK293T cells. Error bars represent the mean ± SE of three independent experiments. c, Summary of editing efficiencies from b. d, C-to-T editing efficiency of each CBE in the editing window of 13 endogenous sites in HEK293T cells. ef, Base substitution pattern (e) and indels analysis (f) of each CBE at 13 endogenous sites in HEK293T cells. The center line in c and f represents the median, and the bottom and top edges of the box represent the top and bottom quartiles of the editing efficiency. The tails extend to the minimum and maximum values. All P values ​​were calculated using a two-sided Wilcoxon rank sum test. The sgRNAs used are as follows:

[0070] sgRNA sequence SEQ ID NO: ADAM6 ATTCCATCACTCAGTGAGAGGAG 99 APOE GGATTACCTGCGCTGGGTGCAGA 100 CFTR CTGACTGCTCACTAGATGTTGAA 101 EMX1 TGGTTGCCCACCCTAGTCATTGG 102 GYPE CTGTTCCTGCTTAGATTCTAAGA 103 HBB GGGATCTGTCCACTCCTGATGCT 104 HMGA1 TGTCTCCCAGCATCCCAGCCATC 105 DNMT1 TTTCCCTTCAGCTAAAATAAAGG 106 DYRK1A GAAGCACATCAAGGACATTCTAA 107 FANCF GGATAAGCACAGTTTTAAATAGT 108 RUNX1 TTCTCCCCTCTGCTGGATACCTC 109 SITE3 GTTTAAACACACCGGGTTAATAA 110 VEGFA CTCTCAAGACCCACAATCCAGGC 111

[0071] Figure 6 , Cluster analysis of candidate cytosine deaminases. a, Clustering of 272 selected cytosine deaminases based on three-dimensional structural similarity. The gradient colors in the heat map indicate the degree of structural similarity. b, Phylogenetic analysis of cytosine deaminases based on primary amino acid sequences. The green line represents 272 deaminases selected from 184 clusters. The red plus green line represents the candidate 1483 cytosine deaminases.

[0072] Figure 7 , CBE and sgRNA vector design. a, base C in 102 sgRNA (such as Figure 2b, GC content analysis of 102 sgRNAs. c, CBE vector backbone for 102 sgRNA-target library; the linker sequence is LSGSETPGTSESATPES (SEQ ID NO: 210). d, lentiviral vector backbone of sgRNA.

[0073] Figure 8 , editing efficiency of CBEs composed of candidate cytosine deaminases. Deep sequencing was used to statistically analyze the C-to-T editing efficiency of CBEs composed of candidate cytosine deaminases and four deaminases with clear editing characteristics (hA3A, rAPOBEC1, YE1, and eA3A) in 102 sgRNA-target libraries. The center line represents the median, and the bottom and top edges of the box represent the top and bottom quartiles of the editing efficiency. The tail extends to the minimum and maximum values. The sgRNA sequences are shown in Figure 2 Listed in.

[0074] Fig. 9 , Editing windows of CBEs composed of candidate deaminases. Deep sequencing was used to analyze the C-to-T editing windows of CBEs composed of candidate cytosine deaminases and four deaminases with clear editing characteristics (hA3A, rAPOBEC1, YE1, and eA3A) in a 102 sgRNA-target library. Data are the average of four independent experiments. sgRNA sequences are shown in Figure 2 Listed in.

[0075] Fig.10 , Editing efficiency of CBE composed of candidate deaminases under different base conversion modes. Average editing efficiency distribution of 272 deaminases at different base conversion types at 102 sgRNA-target sites. The x-axis represents the average editing efficiency of 272 deaminases at 102 sgRNA-target sites, and the y-axis represents the density distribution of editing efficiency. sgRNA sequence as Figure 2 Listed in.

[0076] Fig.11, the off-target effect detection of CBE composed of candidate deaminases was performed using R-loop experiments. a, Off-target effects of candidate deaminases at four dSaCas9-sgRNA recognition sites (Sa site 3, Sa site 4, Sa site 5, and Sa site 6). Error bars represent the mean ± SE of three independent experiments. b, Correlation analysis between on-target and off-target. Each point represents the average editing efficiency of different deaminase editors at 102 sgRNA-target sites (y-axis) and the average off-target effect at 4 R-loop off-target sites (x-axis). c, The ratio of on-target efficiency and off-target effect mediated by candidate deaminases and control deaminases (rAPOBEC1, YE1, and eA3A). The sgRNA sequences are as follows. Figure 2 List and group:

[0077] sgRNA sequence SEQ ID NO: Sa site 3 GTGTCAGGTAATGTGCTAAACA 112 Sa site 4 GGTGGAGGAGGGTGCATGGGGT 113 Sa site 5 TCTGCTTCTCCAGCCCTGGC 114 Sa site 6 GATGTTCCAATCAGTACGCA 115

[0078] Fig.12 , Editing properties of CBEs composed of deaminases derived from cluster #147. a, Average editing efficiency of CBEs composed of cytosine deaminases derived from cluster #147 in 102sgRNA-target library. Four well-characterized deaminases hA3A, rAPOBEC1, YE1 and eA3A were used as controls. The middle line represents the median, and the bottom and top lines of the box represent the first and second quartiles, respectively. The tail extends to the minimum and maximum values. b, Editing windows of 10 cytosine deaminases CBEs derived from cluster #147 in 102sgRNA-target library, with well-characterized deaminases hA3A, rAPOBEC1, YE1 and eA3A as controls. Data are the average of four independent experiments. sgRNA sequences are shown in Figure 2 Listed in.

[0079] Fig.13 , the editing window of CBE composed of 8 highly active deaminases. Deep sequencing analysis of the editing windows of 8 highly active deaminases (CD0458, CD0730, CD0208, CD0902, CD0640, CD0418, cd0181 and CD0911) and 4 deaminases with clear editing characteristics (hA3A, rAPOBEC1, YE1 and eA3A) in 11868 sgRNA-target libraries. The red numbers indicate the editing efficiency of different deaminases at the sites with the highest editing activity, and the red boxes represent the editing windows. The data represent the average editing efficiency of 11,868 sgRNA-target sites. The sgRNA sequences are shown in Figure 2 Listed in.

[0080] Fig.14 , CD0208 CBE editing efficiency. C-to-T editing efficiency of CD0208 CBE at positions 2-7 of 26 endogenous targets in HEK293T cells. Well-characterized deaminases eA3A, YE1, and rAPOBEC1 were used as controls. Data are the average of three independent experiments. sgRNA sequences are shown in Figure 3 Listed in.

[0081] Fig.15 , Prediction of amino acid residues involved in ssDNA interaction in CD0208. The amino acid residues involved in ssDNA interaction in CD0208 were predicted using DRNApred online software. The blue-white gradient indicates the predicted binding ability.

[0082] Fig.16 , CD0208 CBE off-target effects. R-loop assay was used to detect the off-target effects of CD0208 variant CBE at four dSaCas9-sgRNA recognition sites (Sa site 3, Sa site 4, Sa site 5, and Sa site 6). Error bars represent the mean ± SE of three independent experiments. sgRNA sequences are shown in Fig.11 Listed in.

[0083] Fig.17 , editing properties of CD0208 CBE. a, Editing efficiency of CD0208 mutants and three well-characterized deaminases (rAPOBEC1, YE1, and eA3A) in 11868 sgRNA-target libraries was analyzed using deep sequencing. b, Editing windows of CD0208 mutants and three well-characterized deaminases (rAPOBEC1, YE1, and eA3A). The data represent the average of the editing efficiencies of 11,868 sgRNA-target sites. c, Sequence preferences of CD0208 mutants and three well-characterized deaminases (rAPOBEC1, YE1, and eA3A). Error bars represent the average ± SE of the editing efficiencies of 11,868 sgRNA-target sites. sgRNA sequences are shown in Figure 3 Listed in.

[0084] Fig.18 、CD0208 P52A CBE introduced nonsense mutations into the Tyr gene in mouse N2A cells. a, CD0208, CD0208 P52Aa, C-to-T base editing efficiency of rAPOBEC1, YE1, hA3A and eA3A CBE in 11 target sites of Tyr gene in mouse N2A cells. The C base in red font indicates the site to be edited. The data are the average of three independent experiments. b, The efficiency of introducing stop codons in 11 target sites of Tyr gene in N2A cells. The error bars represent the mean ± SE of three independent experiments. sgRNA sequences are as follows Figure 4 Listed in.

[0085] Fig.19 、CD0208 P52A CBE induces nonsense mutations in multiple copies of genes in mESCs. a,CD0208 P52A , rAPOBEC1 and YE1 CBE C-to-T base editing efficiency at eight target sites of three multi-copy genes (Rbmy1a1, Ssty1 and Ssty2) on the Y chromosome of mESCs. The red font is the C base to be edited. b, CD0208 P52A The efficiency of introducing stop codons at eight target sites of three multi-copy genes in mESCs by rAPOBEC1, rAPOBEC1, and YE1 CBE. Data are the mean of three independent experiments. Error bars represent the mean ± SE of three independent experiments. sgRNA sequences are shown in Figure 4 Listed in.

[0086] Fig. 20 、CD0208 P52A CBE introduced nonsense mutations in multiple copies of genes in porcine PK-15 cells. a, CD0208, CD0208 P52A and rAPOBEC1 CBEs induce C-to-T base editing efficiency at nine target sites of PERV pol gene in PK-15 cells. The red font C indicates the desired editing site. b, CD0208, CD0208 P52A The efficiency of rAPOBEC1 and rAPOBEC1 CBEs in inducing stop codons at nine target sites of PERV pol gene in PK-15 cells. Data are the mean of three independent experiments. Error bars represent the mean ± SE of three independent experiments. sgRNA sequences are shown in Figure 4 Listed in.

[0087] Fig.21 , CD0208-dCpf1 CBE editing efficiency at endogenous loci in HEK293T cells. CD0208-dCpf1 and CD0208 P52AC-to-T editing efficiency of -dCpf1 CBEs at 13 endogenous loci in HEK293T cells, with rAPOBEC1-dCpf1, YE1-dCpf1, and hA3A-dCpf1 CBEs as controls. Data are the average of three independent experiments. sgRNA sequences are as follows Figure 5 Listed in. DETAILED DESCRIPTION

[0088] The inventors have revealed a series of novel cytosine deaminases through extensive screening, experimental analysis and optimization. Compared with existing cytosine deaminases, the cytosine deaminases of the present invention have enhanced targeted editing performance; the enhanced targeted editing performance includes: improved sequence-free preference, improved editing efficiency, reduced off-target effects, and improved targeting specificity.

[0089] In the present invention, AlphaFold2 is applied to generate the three-dimensional protein structures of 1483 cytidine deaminases predicted based on artificial intelligence, these deaminases are clustered according to structural similarity, and representative deaminases are selected from each cluster for base editing activity detection. On this basis, combined with research analysis and experiments, the inventors screened a series of deaminases with distinct characteristics, with higher editing efficiency, higher editing efficiency and off-target effect ratio, and diverse editing windows. Surprisingly, the inventors screened and obtained several cytosine deaminases that have no preference for editing site sequences, which show no selective preference for editing different AC / GC / TC / CC motifs (ie, they have the same C-to-T editing efficiency for AC / TC / CC / GC sites).

[0090] Through sequence-based amino acid residues, DNA interaction models and experimental analysis, the inventors predicted multiple amino acid sites that affect single-stranded DNA binding on candidate deaminases, and mutations in these sites further reduced the off-target effects of deaminases. In addition, these cytosine deaminase-based CBEs can introduce stop codons or mutant promoters and alternative splicing sites in single-copy and multi-copy genes in cells without causing double-strand breaks in DNA. The optimized cytosine deaminase is compatible with the dCpf1 CBE system and performs efficient and accurate C-to-T editing in human cells. The present invention highlights the great potential of artificial intelligence-based protein structure prediction tools in mining and screening new functional proteins, and provides new candidate cytosine deaminases for constructing CBE tools to promote the application of CBE in biomedicine.

[0091] These base editors can also introduce stop codons in single-copy and multi-copy genes in mammalian cells without double-strand breaks. This invention provides an efficient method for data mining and greatly expands the application of base editing.

[0092] the term

[0093] As used herein, the terms "cytosine base editor", "C-to-T base editor", "CBE", and "CBEs" are used interchangeably and refer to base editors that convert cytosine C to thymine T.

[0094] As used herein, the term "functional conservative variant" refers to a protein (polypeptide) that substantially retains the same biological function or activity of the protein (polypeptide). The "functional conservative variant" may be (i) a polypeptide in which one or more (e.g., 1-50, 1-30, 1-25, 1-20, 1-15, 1-10, 1-5, 1-3 or 1-2) conservative or non-conservative amino acid residues (preferably conservative amino acid residues) are substituted, and such substituted amino acid residues may or may not be encoded by the genetic code, or (ii) a polypeptide having at least 70%, at least 75%, at least 80%, preferably at least 85%, preferably at least 90%, preferably at least 95%, preferably at least 97% sequence identity, or (iii) a polypeptide having a substitution group in one or more amino acid residues. According to the teachings herein, these fragments, derivatives and analogs belong to the scope known to those skilled in the art.

[0095] As used herein, the term "variant" or "mutant" refers to a peptide or polypeptide that changes the amino acid sequence by insertion, deletion or substitution of one or more amino acids compared to a reference sequence but retains at least one biological activity. The mutant described in any embodiment herein includes an amino acid sequence that has at least 80%, preferably at least 85%, preferably at least 90%, preferably at least 95%, preferably at least 97% sequence identity with a reference sequence (such as any cytosine deaminase obtained by mining herein, specifically such as CD0208) and retains the biological activity of the reference sequence (such as as a glycosyltransferase). The sequence identity between the two aligned sequences can be calculated using, for example, NCBI's BLASTp. Mutants also include amino acid sequences that have one or more mutations (insertions, deletions or substitutions) in the amino acid sequence of the reference sequence while still retaining the biological activity of the reference sequence. The multiple mutations generally refer to within 1-20, such as 1-15, 1-10, 1-8, 1-5 or 1-3. The substitution is preferably a conservative substitution. For example, in the art, when conservative substitutions are made with amino acids with similar or similar properties, the function of a protein or polypeptide is usually not changed. "Amino acids with similar or similar properties" include, for example, families of amino acid residues with similar side chains, including amino acids with basic side chains (e.g., lysine, arginine, histidine), amino acids with acidic side chains (e.g., aspartic acid, glutamic acid), amino acids with uncharged polar side chains (e.g., glycine, asparagine, glutamine, serine, threonine, tyrosine, cysteine), amino acids with non-polar side chains (e.g., alanine, valine, leucine, isoleucine proline, phenylalanine, methionine, tryptophan), amino acids with β-branched side chains (e.g., threonine, valine, isoleucine), and amino acids with aromatic side chains (e.g., tyrosine, phenylalanine, tryptophan, histidine). Therefore, replacing one or several sites in a polypeptide with another amino acid residue from the same side chain class will not substantially affect its activity.

[0096] As used herein, the term "operably linked" or "operably connected to" refers to a situation where certain parts of a linear DNA sequence are able to regulate or control the activity of other parts of the same linear DNA sequence. For example, if a promoter controls the transcription of a sequence, it is operably linked to a coding sequence. "Operationally linked (connected)" means that the connection of each element is in a certain order, such as from amino acid to carboxyl terminus.

[0097] As used herein, "linker" refers to a chemical group or molecule that connects two molecules or parts, such as two domains of a fusion protein, such as a Cas9 enzyme and a deaminase. In some connection modes, the linker is located between or on the flanks of two groups, molecules or other parts, and the two are connected by a covalent bond. In some embodiments, the linker is an amino acid or multiple amino acids (such as a peptide or protein). In some embodiments, the linker is an organic molecule, a group, a polymer or a chemical part. The length and type of the linker can be designed as needed.

[0098] As used herein, "mutation" refers to the substitution of a residue within a sequence (e.g., a nucleic acid or amino acid sequence) with another residue or the change of one or more residues within a sequence to another residue, or the occurrence of a deletion or insertion. In the editor of the present invention, it is desired to enhance the targeted editing performance and generate as many "C-to-T" mutations as possible; the enhanced targeted editing performance includes: improved sequence-free preference (i.e., improved editing versatility / extensiveness / broad spectrum), improved editing efficiency, reduced off-target effects, improved targeting specificity, and improved editing activity, while reducing or eliminating mutations in other bases.

[0099] As used herein, "insertion / deletion" refers to the insertion or deletion (Indels) of a nucleotide base in a nucleic acid. Such insertions or deletions can lead to frameshift mutations in gene coding regions. In the present invention, it is desirable to induce fewer indels and non-C-to-T conversions. In some embodiments, Indels are used as an indicator for investigating base editors, and it is desirable to produce base editors that effectively modify (e.g., mutate or deaminize) specific nucleotides in nucleic acids without producing a large number of insertions or deletions (i.e., insertions / deletions) in nucleic acids.

[0100] Cytosine base editor

[0101] The cytosine base editor of the present invention includes the following elements connected in an operative sequence: cytosine deaminase-Cas enzyme. Preferably, the N-terminus of the cytosine deaminase and / or the C-terminus of the Cas enzyme further includes a nuclear localization signal (NLS).

[0102] The cytosine base editor of the present invention includes a cytosine deaminase or a functional conservative variant thereof obtained by screening in the present invention. In the present invention, "deaminase" refers to a protein or enzyme that catalyzes a deamination reaction. Preferably, the deaminase is a cytosine (cytidine) deaminase, which catalyzes the hydrolysis and deamination of cytidine or deoxycytidine to uridine or deoxyuridine.

[0103] There have been some studies on cytosine deaminase in the art, and a variety of cytosine deaminase have been proposed, including APOBEC1, APOBEC3G, APOBECA3A, YE1, etc. The novel cytosine deaminase in the present invention has better performance than the traditional deaminase.

[0104] In a preferred embodiment of the present invention, between the cytosine deaminase and the Cas enzyme, between the cytosine base editor and the nuclear localization sequence (NLS), and between the Cas enzyme and the nuclear localization sequence, a linker sequence is also included, and the linker sequence can be any linker sequence that does not affect the functions of the two, for example, it can be a tag sequence or some flexible linker sequences known in the art (such as P2A). Suitable tags can be used in the present invention. For example, the tag can be FLAG, HA, HA1, c-Myc, Poly-His, Poly-Arg, Strep-TagII, AU1, EE, T7, 4A6, gE or Ty1.

[0105] In the present invention, the Cas enzyme is an RNA-guided nuclease. Methods for site-specific cleavage (e.g., modifying a genome) using Cas enzymes such as (but not limited to) Cas9 or Cas12 are known in the art. In some embodiments, the Cas9 domain is a Cas9 domain from Streptococcus pyogenes (SpCas9). In some exemplary embodiments, the Cas9 enzyme is a SpCas9 nickase (nSpCas9); or a mutation-inactivated Cpf1.

[0106] In the modified base editor of the present invention, the elements used to construct the base editor, such as cytosine deamination (i.e., the enzyme selected, optimized or modified in the present invention), Cas enzyme, etc., can be recombinant proteins, natural proteins, synthetic proteins, preferably recombinant proteins. The protein of the present invention can be a naturally purified product, or a chemically synthesized product, or produced using recombinant technology from a prokaryotic or eukaryotic host (e.g., bacteria, yeast, higher plants, insects, and mammalian cells).

[0107] The present invention also includes function-conservative variants (including fragments, derivatives and analogs, etc.) of cytosine deaminase (i.e., the enzyme selected, optimized or modified in the present invention) or Cas enzyme.

[0108] The present invention also provides a polynucleotide sequence encoding the selected, optimized or modified enzyme of the present invention or its conservative variant protein.

[0109] The polynucleotide of the present invention can be in the form of DNA or RNA. The DNA form includes cDNA, genomic DNA or artificially synthesized DNA. The DNA can be single-stranded or double-stranded. The DNA can be a coding strand or a non-coding strand.

[0110] The polynucleotide encoding the mature protein of the mutant includes: a coding sequence encoding only the mature protein; a coding sequence of the mature protein and various additional coding sequences; a coding sequence of the mature protein (and optional additional coding sequences) and non-coding sequences.

[0111] A "polynucleotide encoding a protein" may be a polynucleotide that encodes the protein, or may be a polynucleotide that also includes additional coding and / or non-coding sequences.

[0112] The full-length nucleotide sequence of the enzyme selected for optimization or modification of the present invention or its fragment can usually be obtained by PCR amplification, recombination or artificial synthesis. For PCR amplification, primers can be designed based on the relevant nucleotide sequence disclosed in the present invention, especially the open reading frame sequence, and a commercially available cDNA library or a cDNA library prepared by conventional methods known to those skilled in the art is used as a template to amplify the relevant sequence. When the sequence is long, it is often necessary to perform two or more PCR amplifications, and then splice the fragments amplified in each time together in the correct order.

[0113] Once the relevant sequence is obtained, it can be obtained in large quantities by recombinant methods. This is usually done by cloning it into a vector, then transferring it into cells, and then isolating the relevant sequence from the propagated host cells by conventional methods.

[0114] In addition, artificial synthesis methods can also be used to synthesize relevant sequences, especially when the fragment length is shorter. Usually, by synthesizing multiple small fragments first and then connecting them, very long fragments of the sequence can be obtained. At present, the DNA sequence encoding the protein of the present invention (or its fragment, or its derivative) can be obtained completely by chemical synthesis. The DNA sequence can then be introduced into various existing DNA molecules (or such as vectors) and cells known in the art. In addition, mutations can also be introduced into the protein sequence of the present invention by chemical synthesis.

[0115] The present invention also relates to a vector comprising the polynucleotide of the present invention, a host cell produced by genetic engineering using the vector of the present invention or a selected, optimized or modified enzyme coding sequence, and a method for producing the protein of the present invention by recombinant technology.

[0116] The polynucleotide sequence of the present invention can be used to express or produce recombinant selected, optimized or modified enzymes by conventional recombinant DNA technology. Generally speaking, the following steps are involved: (1) using the polynucleotide encoding the selected, optimized or modified enzyme of the present invention, or using a recombinant expression vector containing the polynucleotide to transform or transduce a suitable host cell; (2) culturing the host cell in a suitable culture medium; (3) isolating and purifying the protein from the culture medium or cells.

[0117] In the present invention, the selected optimized or modified enzyme polynucleotide sequence or the cytosine base editor polynucleotide sequence can be inserted into a recombinant expression vector. The term "recombinant expression vector" refers to bacterial plasmids, bacteriophages, yeast plasmids, plant cell viruses, mammalian cell viruses or other vectors well known in the art. In short, as long as they can replicate and stabilize in the host, a variety of plasmids and vectors can be used. An important feature of an expression vector is that it usually contains a replication origin, a promoter, a marker gene, and a translation control element. As a preferred embodiment of the present invention, the expression vector is an expression vector suitable for the introduction of a base editor and for intracellular expression, such as but not limited to a pCMV vector.

[0118] Methods well known to those skilled in the art can be used to construct expression vectors containing the selected optimized or modified enzyme polynucleotide sequence or the cytosine base editor polynucleotide sequence and suitable transcription / translation control signals. These methods include in vitro recombinant DNA technology, DNA synthesis technology, in vivo recombination technology, etc. The DNA sequence can be effectively connected to an appropriate promoter in the expression vector to guide mRNA synthesis. The expression vector also includes a ribosome binding site and a transcription terminator for translation initiation. The expression vector preferably contains one or more selective marker genes to provide phenotypic traits for selecting transformed host cells.

[0119] The vector containing the above-mentioned appropriate DNA sequence and an appropriate promoter or control sequence can be used to transform appropriate host cells or recipient cells.

[0120] Gene editing methods and applications

[0121] The present invention also provides a method for gene editing, including gene editing mediated by the cytosine base editor of the present invention. In addition to using the cytosine base editor of the present invention for gene editing, other aspects of gene editing reagents can be used as known in the art, for example, gRNA (or sgRNA) can be designed in a manner known in the art.

[0122] In the present invention, there is no particular limitation on the objects of gene editing, and they can be somatic cells or germ cells, animal cells or human cells.

[0123] Although cytosine base editors (CBEs) have broad prospects as powerful gene editors, their application in science and medicine is limited by off-target effects of DNA and RNA, low purity of editing products, too wide editing window, and low editing precision. In a specific embodiment of the present invention, the inventors screened enzymes from multiple sources in detail, combined with appropriate mutations, and measured them by measuring indicators such as sequence-free preference, editing efficiency, and off-target effects, and provided appropriate solutions.

[0124] Due to the great potential of CBE in clinical therapeutic applications, a lot of research work continues to focus on reducing the off-target effects of CBE. CBE produces off-target editing, including sgRNA-dependent off-target associated with sgRNA-Cas9, and sgRNA-independent off-target caused by random binding of deaminases to ssDNA or RNA. Optimizing sgRNA or improving Cas9 fidelity can generally reduce sgRNA-dependent off-target. In contrast, cytosine deaminases with low sgRNA-independent off-targets also generally exhibit low on-target editing efficiency. The inventors' previous studies have shown that mutations in the ssDNA / RNA binding domain of deaminases may reduce sgRNA-independent off-target effects. In the present invention, the inventors further focused on the candidate cytosine deaminase CD0208 with no sequence preference. The inventors used DRNApred online software to predict the amino acid residues on CD0208 that may interact with ssDNA / RNA. Then, a series of CD0208 variants were constructed by alanine scanning, and their editing efficiency and off-target effects were determined. Among these variants, CD0208 P52A The mediated off-target effect was significantly lower than that of CD0208, while its on-target efficiency was 2.2 times that of YEI.

[0125] In summary, the present invention has a very high editing efficiency and a cytosine deaminase with no sequence preference. High-efficiency deaminases are an important component of developing advanced CBEs. Since the most effective CBEs currently preferentially edit TC motifs, deaminases with no sequence preference can significantly expand the range of CBE editable targets and improve their application in research and clinical practice.

[0126] The present invention is further described below in conjunction with specific examples. It should be understood that these examples are only used to illustrate the present invention and are not intended to limit the scope of the present invention. The experimental methods in the following examples where specific conditions are not specified are usually carried out according to conventional conditions such as those described in J. Sambrook et al., Molecular Cloning Experiment Guide, 3rd Edition, Science Press, or according to the conditions recommended by the manufacturer.

[0127] Materials and methods

[0128] Three-dimensional structure prediction and clustering of cytosine deaminase

[0129] To obtain candidate cytosine deaminases, the APOBEC-like N-terminal domain (Pfam identifier PF08210) and APOBEC-like C-terminal domain (Pfam identifier PF05240) were downloaded from the Pfam database of InterPro as torrent files. The UniProt database was searched using HMMER (v3.3.1). The e-values ​​of the candidate deaminases were less than 0.01, and incomplete deaminases were removed. In order to quickly and efficiently discover potential deaminases with different characteristics, the deaminases were classified according to their three-dimensional structures, and their functions were further verified based on clustering. First, the three-dimensional structure prediction of the monomer model was performed using AlphaFold2; then the three-dimensional structural similarity of the proteins (TM score ≥ 0.7) was calculated using USalign (v20220924) software, and clustering was performed in combination with protein length and ranking. 10% of the proteins were randomly selected from each category for subsequent analysis (rounded off, if the number was less than 10, one was selected).

[0130] Plasmid construction

[0131] The codon-optimized CD4+ DNA sequence was synthesized at BGI and cloned into the pCMV-nSpCas9-CBE vector backbone ( Figure 7 C) or pCMV-dCpf1-CBE (Addgene Plasmid #107685). Oligonucleotides were synthesized in GENEWIZ and cloned into SpCas9-U6-sgRNA, pCMV-dSaCas9-U6-sgRNA and Cpf1-U6-sgRNA vectors after annealing to construct sgRNA expression vectors matching SpCas9, SaCas9 and Cpf1. For the sgRNA-target library, the synthesized sgRNA-target oligo library was cloned into the Lenti-sgRNA-EGFP vector backbone.

[0132] Cell culture

[0133] HEK293T, N2A and PK-15 cells were cultured in DMEM supplemented with 10% FBS and 1% PS. mESCs were cultured in DMEM supplemented with 15% FBS, 1% NEAA, 1% GlutaMAX, 1% Nucleosides, 1% mM β-mercaptoethanol, 1μM PD0325901, 3μM CHIR99021 and 1000units / mlmouse LIF. All cells were cultured in an incubator at 37°C and 5% CO2.

[0134] Preparation of lentivirus and stable cell lines

[0135] Lentiviral transfection method: 1.5×10 7 HEK293T cells were seeded in 15 cm culture dishes and transfected with 125 μl PEI to 22.5 μg of lentiviral packaging plasmid, 15 μg of lentiviral envelope plasmid, and 30 μg of plasmid expressing sgRNA library. The culture medium was replaced with fresh medium 6 h after transfection, and the culture medium was collected and the viral particles were concentrated 48 h after transfection. The virus titer was detected by qPCR. HEK293T cells were infected with lentivirus (MOI = 0.3) and fresh medium was replaced 6 h after infection. To establish a HEK293T cell library stably expressing 102 sgRNA-targets, GFP-positive cells were screened 7 days after lentiviral infection. To generate HEK293T cells expressing 11868 sgRNA, 2 μg / ml puromycin was added to the cell culture medium 72 h after infection to enrich transfection-positive cells.

[0136] Cell transfection

[0137] For the 102sgRNA-target library experiment, 1×10 6 HEK293T cells stably expressing sgRNA-target library were seeded in 6-well plates, and when they grew to 80% confluence, 6 μg CBE expression plasmid was transfected with 12 μl PEI. 24 hours after plasmid transfection, 3 μg / ml puromycin was added to the cell culture medium to enrich transfection-positive cells, and cells were collected 120 hours after plasmid transfection to detect the editing efficiency of the target site.

[0138] For the 11868 sgRNA-target library experiment, 6 × 10 6 HEK293T cells stably expressing the library were seeded in a 10 cm culture dish and transfected with 80 μl of PEI when they grew to 80% confluence. 72 h after plasmid transfection, 1×10 7Positively transfected cells were identified and genomic DNA was extracted.

[0139] For R-loop off-target detection experiments, 2×10 5 HEK293T cells were seeded in 24-well plates and co-transfected with 0.75 μg pCMV-dSaCas9-U6-sgRNA vector and 0.75 μg pCMV-nSpCas9-CBE vector using 3 μl PEI. 24 h after plasmid transfection, 3 μg / ml puromycin was added to the cell culture medium to enrich transfected positive cells, and cells were collected 120 h after transfection to detect off-target effects.

[0140] To induce a stop codon in the Tyr gene, 2 × 10 5 N2A cells were seeded in 24-well plates, and when they grew to 80% confluence, 0.75 μg of plasmid encoding sgRNA and 0.75 μg of plasmid expressing CBE were transfected with 3 μl PEI. 72 h after plasmid transfection, transfection-positive cells were enriched by flow cytometry and genomic DNA was extracted.

[0141] To induce stop codons in multicopy genes, mESCs were seeded in 24-well plates and, when grown to 80% confluence, lipofectamine TM 0.25 μg of plasmid encoding sgRNA and 0.25 μg of CBE were transfected with 3000 transfection reagent. 72 h after plasmid transfection, transfection-positive cells were enriched by flow cytometry and genomic DNA was extracted. 6 PK-15 cells were seeded in 6-well plates and grown to 80% confluence using a Lonza nucleoector TM Plasmid encoding 3 μg sgRNA and 3 μg CBE expression plasmid were transfected using 2b nucleofection system and T-020 program. 72 h after mass transfection, transfection-positive cells were sorted by flow cytometry and genomic DNA was extracted.

[0142] R-loop off-target detection analysis

[0143] R-loop experiments were used to evaluate the sgRNA-independent off-target effects of different base editors. Specifically, four dSaCas9-sgRNA recognition sites (Sa site3, Sa site4, Sa site5, and Sa site6) were selected as potential sgRNA-independent DNA off-target sites. 5HEK293T cells were seeded in 24-well plates, and 0.75μg dSaCas9-sgRNA expression plasmid and 0.75μg plasmid encoding CBEs were co-transfected into HEK293T cells by 3ul PEI. Then the cells were treated with 3ng / ml puromycin, and the cells were collected 120h after transfection to extract genomic DNA. Finally, the editing efficiency of the genomic site was detected by Illumina NovaSeq sequencing platform.

[0144] Targeted deep sequencing

[0145] The extracted cell genome was used as a template for PCR amplification using Takara Ex Taq Polymerase. In the second round of PCR, a barcode was added to distinguish different samples. The PCR product was purified using a DNA purification kit. The purified DNA was sequenced with 150 bp end pairs on the Illumina NovaSeq sequencing platform.

[0146] Bioinformatics analysis

[0147] Sequencing reads were demultiplexed using FASTQ-multx (v1.4.2), and FASTQ files were analyzed using CRISPResso2 (v2.2.7). Based on the results of the three-dimensional structural similarity analysis, the UPGMA (unweighted Pair Group Method with Arithmetic Mean) phylogenetic tree was constructed using the Phangorn package. The three-dimensional structural similarity matrix was constructed using PheatMap (v1.0.12). The phylogenetic tree was drawn using the ggtree (v3.8.0) package in R (v4.1). Another article image was drawn using R packages such as ggplot2 (v3.4.2) and pheatmap.

[0148] Sequence information

[0149] The sequences of some cytosine deaminases in the present invention are shown in Table 1.

[0150] Table 1

[0151]

[0152]

[0153]

[0154] Example 1: Mining and screening of cytosine deaminases based on three-dimensional structural classification

[0155] In order to improve the efficiency of screening for cytosine deaminases with editing activity or showing special functions, the inventors applied artificial intelligence-mediated protein structure analysis tools. For this method, a database search based on cytosine deaminase protein sequence homology was performed, 3D structure prediction was performed on the candidate cytosine deaminases screened, they were clustered by 3D structural similarity, and then a subset of cytosine deaminases was selected from each cluster, and their C-to-T editing activity in CBE was compared by Illumina NovaSeq sequencing ( Figure 1 ). For sequence homology search, two cytosine deaminase catalytic activity domains (APOBEC-like N-terminaldomain (PF08210) and APOBEC-like C-terminal domain (PF05240)) were selected from the Pfam database of Interpro. Protein sequences carrying these two domains were searched in the Uniprot database using amino acid primary sequence alignment. The top 1483 homologous protein sequences (e-values ​​were all less than 0.01, and incomplete proteins were removed) were selected for further experimental analysis.

[0156] Given the importance of three-dimensional structure to protein function, the inventors believe that clustering based on the complete protein structure may reflect functional specificity. Through the Alphafold2 three-dimensional structure prediction of these proteins, 1483 candidate deaminase proteins were divided into 184 clusters according to their structural similarities ( Figure 1 b). Next, 272 cytidine deaminases were randomly selected from the 184 clusters for subsequent experiments ( Figure 1 c and Figure 6 a). To validate the initial clustering analysis, the inventors re-clustered the 272 selected cytidine deaminases based on their amino acid sequences, reproducing the structure-based clustering observed in the full set of candidate enzymes ( Figure 6 b).

[0157] To facilitate the expression and function of these cytosine deaminases in mammalian cells, each deaminase sequence was optimized for human codons before synthesis and cloned into the CBE expression vector ( Figure 7 c).

[0158] Finally, 272 candidate CBE editors were transfected into a human HEK293T cell line stably expressing the sgRNA-target library, and their activities were compared using Illumina NovaSeq sequencing.

[0159] Example 2: Cytosine deaminase with high editing efficiency, low off-target effect, and diverse editing windows

[0160] The sgRNA-target library detection strategy was used to detect the editing activity of CBEs containing candidate deaminases ( Figure 2 a). To this end, the inventors synthesized a rationally designed oligo sequence library containing 102 sgRNAs and corresponding target sequences. Specifically, these sgRNAs included at least one C located within 2-8 nt of the PAM end, and 84% of the sgRNAs had a GC content between 40-65% ( Figure 7 a, b). The oligonucleotide library was then cloned into a lentiviral expression vector ( Figure 7 d), and stably integrated into the genome of HEK293T cells by lentivirus, and HEK293T cells were enriched by flow cytometry. After each CBE was transfected into cells stably expressing the sgRNA-target library, the C-to-T editing activity mediated by these CBEs was detected by the Illumina NovaSeq sequencing platform, with an average coverage of ≥1000× for each sgRNA and an average coverage of 2000× for the target site PCR product.

[0161] Here, the inventors used rAPOBEC1, YE1, hAPOBEC3A (hA3A) and engineered hA3A (eA3A) as controls to evaluate the editing activities mediated by 272 deaminases. Sequencing results of the target site PCR products showed that the average C-to-T editing efficiency of 71 editors and 102 sgRNA target sites was >10% ( Figure 2 b and Figure 8 Notably, 47 candidate CBEs had higher editing efficiencies than YE1 (21.4%), 22 candidate CBEs had higher editing efficiencies than rAPOBEC1 (34.9%), and 6 candidate CBEs had higher efficiencies than hA3A (53.6%), one of the most efficient deaminases reported to date ( Figure 2 c and Figure 8 ). Sequencing analysis of the editing window widths of 272 candidate deaminases revealed the diversity of the editing windows of these candidate deaminases ( Figure 2 d and Fig. 9), in the present invention, a lower threshold of maximum editing efficiency ≥ 50% is used to define the editing window. Relative to the editing window of positions 4-7 of rAPOBEC1 (PAM was identified as positions 21-23), the editing windows of some deaminases detected were shifted forward (e.g., C2-C7 for CD0208 and CD0640; C2-C6 for CD0256; C1-C7 for CD0293), or some were shifted backward (e.g., C3-C12 for CD0596; C2-C10 for CD0452; C4-C12 (CD0602), while some deaminases had a wider editing window (e.g., C1-C9 for CD0458, CD0730 and CD0181; C1-C10 for CD0336), and some deaminases had a very narrow editing window (e.g., C4-C6 for CD0354, CD0237, CD0371 and CD0230) ( Figure 2 d and Fig. 9 ). These high editing efficiencies and diverse editing windows observed in the candidate deaminases suggest that there is room for further improvement in both the editing efficiency and targeting range of CBEs compared to current editors. In addition, further analysis of base transitions other than C-to-T revealed that C-to-G transitions occurred more frequently than other types, suggesting that these cytosine deaminases have the feasibility of being designed as C-to-T base editors ( Fig.10 ).

[0162] Cytosine deaminases can bind to single-stranded DNA or RNA in cells and induce deamination, resulting in widespread sgRNA-independent off-target effects at the genome level. Evaluating these sgRNA-independent off-target effects induced by cytosine deaminases is critical for the development of high-fidelity cytosine deaminases. Here, the sgRNA-independent off-target effects induced by these deaminase editors were detected in HEK293T cells using the SaCas9 R-loop experiment, and four SaCas9 target sites were selected for off-target detection analysis. The detection structure showed that the off-target efficiency of all deaminases at the four potential off-target sites showed high consistency ( Fig.11 a), which shows the reliability of the detection system. The inventors also noticed that the on-target efficiency and off-target effect of most deaminases are highly positively correlated, that is, deaminases with high on-target efficiency often also have high off-target effects ( Figure 2 c and Fig.11b). However, a group of deaminases were also found to have a high on target and off target ratio (for example, CD0755, CD0526, CD0085, CD0827, and CD0236 had on target / off target values ​​that were 4.0, 3.8, 3.5, 2.8, and 2.6 times that of rAPOBEC1, respectively) ( Figure 2 c and Fig.11 c), indicating that they have high targeting specificity. In addition, CD0208 and CD0640 have higher on-target efficiency than hA3A (1.1-fold and 1.1-fold, respectively), but lower off-target effect than hA3A (0.7-fold and 0.7-fold, respectively) ( Figure 2 c and Fig.11 c). Some deaminases (e.g., CD0755, CD0526, and CD0085) have higher ontarget / offtarget values ​​than the widely used high-specificity deaminase YE1 (1.6-fold, 1.5-fold, and 1.4-fold, respectively) ( Figure 2 c and Fig.11 c).

[0163] In addition, the inventors re-studied the above-mentioned cluster analysis based on 3D structure and verified whether the similarity of the editing activity of the 10 deaminases with the highest editing efficiency (CD0181, CD0208, CD0288, CD0336, CD0418, CD0458, CD0640, CD0730, CD0902 and CD0911) was reflected in the structural features. The analysis showed that these 10 proteins were indeed clustered into three clusters (CD0208 and CD0640 belonged to cluster #132, CD0181, CD0336 and CD0418 belonged to cluster #145, and CD0288, CD0458, CD0730, CD0902 and CD0911 belonged to cluster #147) ( Figure 2b), which suggests that the editing efficiency of these deaminases may be closely related to their 3D protein structures. To further verify the hypothesis that deaminases with similar functions (e.g., high editing activity) have similar 3D structures and are clustered in the same cluster, the inventors speculated that there are more cytosine deaminases with similarly high editing activities that are clustered into the same cluster. The inventors selected 10 uncharacterized cytidine deaminases (CD0590, CD0058, CD0149, CD0956, CD0054, CD0931, CD0746, CD0027, CD0289 and CD0701) from cluster 147 and examined their editing activities. The experiment showed that 4 out of 10 candidate deaminases (40%) had editing efficiencies comparable to that of hA3A (65.6% for CD0956, 65.0% for CD0931, 60.3% for CD0054, and 52.6% for CD0701) (53.6%), which was higher than the proportion of 10 highly efficient deaminases screened from 224 candidate proteins in all clusters (4.5%) ( Figure 2 e and Fig.12 a, b). These results demonstrate the efficiency and reliability of deaminase mining via 3D structure classification strategy.

[0164] Example 3: Highly active cytosine deaminase with no selective preference for target site sequence

[0165] To precisely characterize the editing features and perform unbiased screening of candidate deaminases, we next tested the editing activity of candidate deaminases in a larger library of 11,868 sgRNA-target sequences, which was constructed in the same way as the 102 sgRNA-target library ( Figure 2 a). The inventors detected 8 deaminases with editing activities higher or close to hA3A in the 102sgRNA-target library (CD0458 was 66.5%, CD0730 was 64.6%, CD0208 was 61.6%, CD0902 was 60.9%, CD0640 was 60.3%, CD0418 was 55.5%, CD0181 was 52.9%, and CD0911 was 51.8%) ( Figure 8Illumina NovaSeq sequencing analysis showed that these eight base editors also showed very high C-to-T editing efficiencies in the 11,868 sgRNA-target libraries (52.2% for CD0458, 47.1% for CD0730, 48.7% for CD0208, 54.5% for CD0902, 45.8% for CD0640, 45.0% for CD0418, 55.2% for CD0181, and 48.4% for CD0911) compared with hA3A (51.5%), rAPOBEC1 (39.0%), YE1 (31.5%), and eA3A (19.9%). Figure 3 a and Figure 8 In terms of editing window, all eight deaminases showed a wide editing window (C1-C11 for CD0458 and CD0730, C1-C8 for CD0208, C1-C12 for CD0902, C2-C8 for CD0640, C1-C9 for CD0418, and C1-C12 for CD0181), and all had the highest editing activity in C5 ( Fig.13 ). These results are consistent with those obtained in the 102sgRNA-target library.

[0166] Since the application of many commonly used deaminases is limited by some sequence priority editing principles, the inventors then studied the editing sequence preferences of the above 8 deaminases through Illumina NovaSeq sequencing data. The analysis results showed that the editing sequence preferences of these 8 deaminases were divided into 4 categories: CD0458 and CD0730 had obvious preferential targeting of TC motifs; CD0181 and CD0418 showed high editing activity at GC and TC sites, among which the activity at GC sites was the highest, which made up for the low editing efficiency of eA3A, YE1 and rAPOBEC1 at GC; CD0902 and CD0911 had the highest efficiency for AC motifs, and preferred RC (AC / GC>TC / CC), and CD0911 preferred AC / TC rather than GC / CC; most notably, CD0208 and CD0640 showed no preferential editing of all editing sites, which means that the editing of C at all AC / TC / GC / CC sites was non-selective ( Figure 3 b). In particular, CD0208 exhibited high editing activity comparable to that of hA3A, while the off-target activity was only 0.72-fold that of hA3A ( Figure 3 a and Fig.11 c), which makes it have the potential to develop into a highly versatile editing tool.

[0167] To further characterize the editing properties of CD0208, the inventors tested its editing efficiency and sequence preference of 34 endogenous targets in HEK293T cells. In this experiment, HEK293T cells were co-transfected with vectors expressing CD0208 CBE and sgRNA, and Illumina NovaSeq sequencing analysis showed that CD0208 CBE showed no selective editing preference for all NC sites within the 3-7nt editing window ( Figure 3 c, d and Fig.14 ). Comparison with other CBEs showed that the overall editing efficiency of CD0208 and rAPOBEC1 was significantly higher than that of eA3A and YE1, which preferentially edited TC sites. In addition, although rAPOBEC1 had comparable efficiency to CD0208 at AC / CC / TC motifs, its editing activity at GC sites was significantly lower than that of CD0208 ( Figure 3 c, d and Fig.14 ).

[0168] These results indicate that CD0208-based CBEs can efficiently edit any NC site within the editing frame of the target site with little sequence preference.

[0169] Example 4: Rational engineered mutations reduce off-target effects of non-sequence-biased deaminases

[0170] The inventors' previous studies have shown that the sgRNA-independent off-target effects caused by deaminases can be reduced by mutating single or multiple amino groups on deaminases that are related to single-stranded DNA binding. Therefore, the same method is used here to reduce the sgRNA-independent off-target effects of CD0208 deaminase-induced. In order to select amino acid residues that may be involved in the interaction between CD0208 and ssDNA, the online DNA and RNA binding prediction tool DRNApred (http: / / biomine.cs.vcu.edu / servers / DRNApred / ) was used. Analysis found that 27 amino acid residues were most likely to be involved in ssDNA binding (score>0.4; Figure 3 e and Fig.15 Given that alanine scanning is an effective strategy to study the functions of amino acid residues, these 27 amino acid residues were replaced by alanine (A) to construct CD0208 mutant editors, and their editing efficiencies were tested in HEK293T cells expressing 11,868 sgRNA-target libraries.

[0171] In addition, we also tested whether mutations at four R-loop potential off-target sites affect off-target effects. The results showed that 11 mutants had lower sgRNA-independent off-target effects than CD0208, includingP52A (40.3% of CD0208), while 15 mutants had higher editing efficiencies than CD0208, including R15A (1.2 times of CD0208) Figure 3 f and Fig.16 , 12a). These results suggest that rationally designed amino acid residue substitutions can reduce off-target effects and also increase the editing activity of CBEs in mammalian cells.

[0172] To further evaluate the editing specificity of each mutant, the inventors eliminated four cytidine deaminase mutants with low editing activity, including CD0208 W169A 、CD0208 S168A 、CD0208 Y199A and CD0208 R197A Comparison of the on target / off target values ​​of CD0208 CBE variants showed that six variants had higher editing specificity than CD0208 (CD0208 P52A 4.4,CD0208 N45A 2.8,CD0208 H188A 2.3,CD0208 R53A 2.3,CD0208 R51A 2.2,CD0208 T170A is 2.0, CD0208 is 1.9). P52A The editing specificity of α-amylase was the best, 2.3 times that of CD0208 and 1.7 times that of YE1 ( Figure 3 g). These results support the hypothesis that residue P52 of CD0208 is involved in ssDNA binding and that CD0208 P52A The mutation can significantly reduce off-target effects while maintaining high cytidine deamination activity.

[0173] At the same time, the inventors also characterized the sequence preferences and editing windows of these mutation editors and found that almost all mutants, including CD0208 P52A , all showed a non-sequential preference characteristic comparable to CD0208 ( Figure 3 h and Fig.17 c). However, the editing window of some variants was changed compared with CD0208 ( Figure 3 i and Fig.17 b) For example, CD0208 W2A and CD0208 R15A Shows a wider editing window than CD0208 ( Figure 3 i and Fig.17 a,b). Analysis of editing types showed that CD0208 P52ACompared with other cytidine deaminases, high C-to-T editing purity is maintained ( Figure 3 j).

[0174] In summary, structure-guided mutagenesis can reduce off-target effects and improve the editing specificity of CD0208.

[0175] Example 5, CD0208 P52A CBE can efficiently induce nonsense mutations in single-copy and multi-copy genes in mammalian cell lines

[0176] To study CD0208 P52A To investigate whether CBE can introduce nonsense mutations in single-copy genes without causing DNA double-strand breaks (DSBs), the inventors tested the editing efficiency of introducing stop codons into endogenous loci in mouse N2A cells. Eleven sgRNAs targeting Tyr genes that can induce stop codons were designed ( Figure 4 a). And these sgRNAs were combined with CD0208 P52A CBE were co-transfected into mouse N2A cells, and rAPOBEC1, YE1, hA3A, eA3A and CD0208 CBE were used as controls. Illumina NovaSeq sequencing analysis showed that CD0208 P52A The C-to-T editing efficiency of CBE (34.3%) was significantly higher than that of all control editors (rAPOBEC1 CBE: 17.1%, YE1 CBE: 9.5%, hA3A CBE: 25.7%, eA3A CBE: 15.2%), and also higher than that of CD0208 CBE (29.0%). The editing window was narrowed compared with that of CD0208 CBE (CD0208 P52A C4-C8, CD0208 is C2-C8)( Figure 4 a, b, c and Fig.18 a). This is consistent with the results of the inventors in the sgRNA-target library ( Figure 3 i and Fig.17 b) Quantification of the efficiency of introducing stop codons showed that CD0208 P52A CBE (25.0%) has the same efficiency as CD0208 and hA3A CBE (25.7% and 22.2%, respectively), but is significantly higher than rAPOBEC1, YE1, and eA3A CBE (16.4% for rAPOBEC1 CBE, 6.9% for YE1 CBE, and 9.2% for eA3A CBE) ( Figure 4 d and Fig.18 a, b). These results show that CD0208 P52ACBE can effectively induce nonsense mutations in the genome of mouse N2A cells.

[0177] Complete knockout of multi-copy genes in mammalian cells has long been a challenge for many commonly used editing tools. The inventors next evaluated CD0208 P52A Whether CBE can also efficiently introduce nonsense mutations in multi-copy genes. In this analysis, the efficiency of inducing stop codons in multi-copy genes in mouse embryonic stem cells (mESCs) and porcine kidney 15 cells (PK-15) was examined. Specifically, Rbmy1a1, Ssty1, and Ssty2 are multiple copy genes present on chromosome Y in the mouse genome (Rbmy1a1>50 copies, Ssty1>35 copies, Ssty2>30 copies), and targeted cutting of these genes using Cas9 in cells and mouse embryos induces Y chromosome deletions. In order to introduce stop codons, 2, 3, and 3 sgRNAs were designed for Rbmy1a1, Ssty1, and Ssty2, respectively ( Figure 4 e). Then these sgRNA expression vectors and CD0208 P52A CBE vectors were co-transfected into mESCs, and CBE containing rAPOBEC1 or YE1 was used as a control. Illumina NovaSeq sequencing analysis showed that CD0208 P52A The C-to-T editing efficiency of the CBE group was significantly higher than that of the rAPOBEC1 CBE and YE1 CBE groups ( Figure 4 e, f, g, and Fig.19 a). CD0208 P52A The average C-to-T editing efficiency of CBE at 8 sgRNA sites reached 73.0%, which was 1.2 and 1.7 times that of rAPOBEC1 CBE (60.9%) and YE1 CBE (42.8%). Figure 4 e, f, g, and Fig.19 a). The inventors also noted that CD0208 P52A The efficiency of CBE inducing stop codons (62.0%) was significantly higher than that of cells treated with rAPOBEC1 CBE (55.1%) and YE1 CBE (33.3%). Figure 4 h and Fig.19 a,b).

[0178] There are multiple copies of the porcine endogenous retrovirus (PERVs) genome in the pig genome, and PERVs are at high risk of infecting humans through pig organ transplantation. Previous studies have shown that the removal of PERVs from pig cells by CRISPR-Cas9 often leads to activation of the P53 pathway, and the resulting DSBs further lead to cell apoptosis. In order to validate CD0208P52A Whether CBE can induce nonsense mutations in PERVs genes without generating DSBs in pig cells, 9 sgRNAs with premature termination codons were designed for the pol gene, which is essential for PERV replication and infection. Figure 4 i), and co-transfected sgRNAs and CBE into PK-15 cells, with rAPOBEC1 CBE and CD0208 CBE as controls. Sequencing analysis showed that the editing efficiency of CD0208 P52A The C-to-T editing efficiency mediated by CBE (58.4%) was comparable to that of CD0208 CBE (54.3%), and significantly higher than that of rAPOBEC1CBE (32.1%) ( Figure 4 i, j, k, and Fig. 20 a). At the same time, these results show that CD0208 P52A The efficiency of CBE (32.9%) in inducing stop codons through C-to-T conversion was significantly higher than that of rAPOBEC1CBE (18.9%) ( Figure 4 l and Fig. 20 a,b). In addition, CD0208 P52A CBE and CD0208 CBE showed high editing activity towards almost all NC sites in the editing window, while rAPOBEC1 CBE had low editing activity at GC motifs ( Fig. 20 a). These results indicate that CD0208 P52A CBE can efficiently introduce nonsense mutations into multi-copy genes in mammalian cells.

[0179] Example 6, CD0208 P52A -dCpf1 CBE can be efficiently edited in human cells

[0180] The most widely used CBEs are constructed using nCas9 fused to cytosine deaminase, but harmful byproducts and NGG PAM limit their application. Previous studies have shown that rAPOBEC1-dCpf1 CBE can recognize T-rich PAM sequences and induce fewer indels and non-C-to-T conversions than other editors. Therefore, the inventors combined dCpf1 and CD0208 P52A The fusion constructs CD0208 with high C-to-T editing activity and high editing purity without sequence preference. P52A -dCpf1 CBE( Figure 5 a) Place CD208 P52A-dCpf1 CBE targeting 13 endogenous target sites in HEK293T cells with TTTN PAM, followed by deep sequencing to assess editing efficiency and specificity. P52A The C-to-T editing efficiencies of -dCpf1 CBE (27.3%) and CD0208-dcpf1 CBE (27.1%) were significantly higher than those of the well-characterized editors rAPOBEC1-dCpf1, YE1-dCpf1, and hA3A-dCpf1 CBE (8.6% for rAPOBEC1-dCpf1 CBE, 4.1% for YE1-dCpf1 CBE, and 18.2% for hA3A-dCpf1 CBE) ( Figure 5 b, c, d and Fig.21 ).

[0181] Consistent with the above findings of the inventors, CD0208 P52A -dCpf1 CBE editing window (C7-C11) maintains high editing efficiency. CD0208 P52A The C-to-A and C-to-G substitution rates of -dCpf1 CBE were 2.0% and 1.0%, respectively ( Figure 5 e), while CD0208 P52A The indels induction rate of -dCpf1 CBE (0.1%) was comparable to that of rAPOBEC1-dCpf1 and YE1-dCpf1 groups (0.1% and 0.1%, respectively). Figure 5 f).

[0182] In addition, CD0208 P52A -dCpf1 CBE can efficiently edit almost any NC target site within the editing window ( Fig.21 ).

[0183] These results indicate that CD0208 P52A -dCpf1 CBE can mediate efficient, sequence-free C-to-T editing at multiple target sites, thus expanding the application range of CBE.

[0184] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims. At the same time, all the documents mentioned in the present invention are cited as references in this application, just as each document is cited as reference separately.

Claims

1. Use of a protein or a functional conservative variant thereof as a cytosine deaminase or for preparing a cytosine base editor; wherein: The protein is selected from: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD590, CD0058, and CD0149.

2. A method for preparing a pyrimidine base editor, comprising using a protein selected from the following group or a functional conservative variant thereof as a cytosine deaminase: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD0590, CD0058, and CD0149.

3. The use according to claim 1 or the method according to claim 2, characterized in that The protein or its functional conservative variant has enhanced targeted editing performance; the enhanced targeted editing performance includes: improved sequence-free preference, improved editing efficiency, reduced off-target effects, and improved targeting specificity; or The cytosine deaminase or the cytosine base editor containing it is used for C-to-T gene editing; preferably, the C-to-T gene editing includes: introducing a C-to-T mutation into a target site of a target gene in a cell, introducing a stop codon, mutating a target promoter or introducing a variable splicing site.

4. The use according to claim 1 or the method according to claim 2, characterized in that The CD0208 is a protein having an amino acid sequence as shown in SEQ ID NO: 1; preferably, the functional conservative variant of CD0208 is based on the amino acid sequence as shown in SEQ ID NO: 1, and has a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A; The CD0640 is a protein having an amino acid sequence shown in SEQ ID NO: 2; The CD0181 is a protein having an amino acid sequence shown in SEQ ID NO: 3; The CD0418 is a protein having an amino acid sequence shown in SEQ ID NO: 4; The CD0458 is a protein having an amino acid sequence shown in SEQ ID NO: 5; The CD0730 is a protein having an amino acid sequence shown in SEQ ID NO: 6; The CD0902 is a protein having an amino acid sequence shown in SEQ ID NO: 7; The CD0911 is a protein having an amino acid sequence shown in SEQ ID NO: 8; The CD0755 is a protein having an amino acid sequence shown in SEQ ID NO: 9; The CD0526 is a protein having an amino acid sequence shown in SEQ ID NO: 10; The CD0085 is a protein having an amino acid sequence shown in SEQ ID NO: 11; The CD0827 is a protein having an amino acid sequence shown in SEQ ID NO: 12; The CD0236 is a protein having an amino acid sequence shown in SEQ ID NO: 13; The CD0956 is a protein having an amino acid sequence shown in SEQ ID NO: 14; The CD0931 is a protein having an amino acid sequence shown in SEQ ID NO: 15; The CD0054 is a protein having an amino acid sequence shown in SEQ ID NO: 16; The CD0701 is a protein having an amino acid sequence shown in SEQ ID NO: 17; The CD0590 is a protein having an amino acid sequence shown in SEQ ID NO: 18; The CD0058 is a protein having an amino acid sequence shown in SEQ ID NO: 19; or The CD0149 is a protein having the amino acid sequence shown in SEQ ID NO:

20.

5. A cytosine deaminase variant, which is a functional conservative variant of CD0208, based on the amino acid sequence shown in SEQ ID NO: 1, having a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A; preferably, the functional conservative variant of CD0208 has enhanced targeted editing performance; the enhanced targeted editing performance includes: Improved non-sequence preference, improved editing efficiency, reduced off-target effects, and improved targeting specificity.

6. A cytosine base editor, comprising a cytosine deaminase; the cytosine deaminase is selected from the following group proteins or their functional conservative variants: CD0208, CD0640, CD0181, CD0418, CD0458, CD0730, CD0902, CD0911, CD0755, CD0526, CD0085, CD0827, CD0236, CD0956, CD0931, CD0054, CD0701, CD0590, CD0058, CD0149; preferably, the cytosine deaminase is CD0208, and its functional conservative variant is based on the amino acid sequence shown in SEQ ID NO: 1, having a mutation selected from the following group: P52A, N45A, H188A, R53A, R51A, T170A, W2A, R15A.

7. The cytosine base editor according to claim 6, wherein It includes the following elements connected in an operative sequence: cytosine deaminase-Cas enzyme; preferably, the N-terminus of the cytosine deaminase and / or the C-terminus of the Cas enzyme also includes: a nuclear localization signal.

8. An isolated polynucleotide or a recombinant expression vector containing the polynucleotide; wherein: The polynucleotide encodes the cytosine base editor of claim 6 or 7.

9. Use of the cytosine base editor according to claim 6 or 7 for C-to-T gene editing; or Reagents or kits for preparing C-to-T gene editing. Preferably, the cytosine base editor has enhanced targeted editing performance; the enhanced targeted editing performance includes: Improved non-sequence preference, improved editing efficiency, reduced off-target effects, and improved targeting specificity.

10. A method for C-to-T gene editing, comprising: mediating gene editing with the cytosine base editor of claim 6 or 7; preferably, the nucleic acid sequence encoding the cytosine base editor and the sgRNA are co-transfected or injected into a receptor to perform gene editing; preferably, the receptor comprises: Somatic or germ cells; Preferably, the reproductive cells include embryonic cells or fertilized eggs.

Images