Nuclease fusion protein, mitochondrial DNA base editing system and application

Abstract

The invention belongs to the technical field of gene editing, and discloses nuclease fusion protein, a mitochondrial DNA base editing system and application. The invention provides a nuclease fusion protein. The nuclease fusion protein comprises double-mutation APOBEC3A, a Cas12 protein and a uracil glycosylase inhibitor, the double mutation APOBEC3A is subjected to W104A mutation and Y132D mutation; the Cas12 protein is GoCas12m, MmCas12m or LbCas12a which is subjected to double mutation, and the Cas12 protein is GoCas12m, MmCas12m or D832A mutation and E925A mutation are carried out on the LbCas12a subjected to double mutation. The invention provides a mitochondrial DNA base editing system comprising the nuclease fusion protein. According to the invention, the obstacle that conventional sgRNA is difficult to efficiently enter mitochondria is overcome, the editing efficiency of a CRISPR system in the mitochondria can be remarkably improved, and the existing efficiency of 0.03-0.23% is improved to a clinical available level (gt; 30%); the problem of bacterial toxicity in the construction process of the mitochondrial gene editing system is solved, the yield and stability of plasmids are improved, and large-scale preparation and application are facilitated; and the positioning efficiency of the editing component in the mitochondrial matrix is maximized. The invention provides a unified mitochondrial gene editing platform with high efficiency, high specificity and broad-spectrum applicability, and a theoretical basis is provided for precise targeting of disease-related mutation.

Description

Technical Field

[0001] This invention relates to the field of gene editing technology, specifically to a nuclease fusion protein, a mitochondrial DNA base editing system, and their applications. Background Technology

[0002] Mitochondrial DNA (mtDNA) is a circular, double-stranded DNA genome within the cell, containing approximately 100 to 100,000 copies per cell. Human mtDNA consists of 16,569 base pairs and encodes 37 genes: 13 proteins required for oxidative phosphorylation, 22 transfer RNAs, and 2 ribosomal RNAs. Mutations in mtDNA can lead to mitochondrial dysfunction, triggering a spectrum of serious diseases affecting multiple organs and systems. According to analysis of the MITOMAP database, conversion mutations (especially C-to-T and G-to-A conversions) account for approximately 41% of known pathogenic mtDNA variants, highlighting the urgent need for precise mitochondrial gene repair tools.

[0003] In recent years, the main technical routes for mitochondrial genome editing include: TALE-based editing systems: DdCBEs technology (Mok et al., 2020, Nature), which uses isolated DddA cytosine deaminase fused with a TALE array to achieve a C-to-T conversion efficiency of 5-50%; TALEDs technology (Cho et al., 2022, Cell), which uses an evolved TadA8e adenine deaminase to achieve an A-to-G editing efficiency of approximately 27%; and MitoBEs technology (Yi et al., 2023, Nature Biotechnology), which integrates a TALE array with multiple deaminases and nickases to achieve chain-specific editing. An attempt at CRISPR / Cas-based mitochondrial editing: SpCas9-based systems (Bi et al., 2022, The Innovation; Bian et al., 2019, ACS Synthetic Biology; Feng et al., 2022, Computational and Structural Biotechnology Journal) achieved extremely low editing efficiencies of only 0.03-0.23% in a mitochondrial environment.

[0004] While each of the above methods has its own characteristics, they all have certain technical limitations. The TALE system requires complex engineering design, while the CRISPR system faces the problem of extremely low efficiency in the mitochondrial environment. Current mitochondrial DNA editing technologies have the following key drawbacks: 1) Complexity of engineering design: The TALE system typically requires the design of 12-20 repeating units of 34 amino acids each, with each unit recognizing a single nucleotide. This leads to a complex and costly construction process. The TALE array needs to be redesigned and reconstructed for each new target, severely limiting high-throughput applications and library screening; 2) Delivery difficulties: TALE system constructs are usually large in size, exceeding the packaging limits of commonly used viral vectors such as adeno-associated virus (AAV) (approximately 4.7 kb); the sgRNA of the SpCas9 system (approximately 100 bp) is difficult to efficiently cross the mitochondrial double membrane system due to its length, resulting in excessively low concentrations of the editing components within the mitochondria; 3) Efficiency limits: The editing efficiency of SpCas9-based systems in mitochondria is extremely low (only 0.03-0.23%), far below the threshold required for therapeutic applications; although the TALE system has higher efficiency (5-50%), it is still difficult to achieve high efficiency in mitochondria. 4) Limited multi-target editing: Existing systems struggle to simultaneously edit multiple sites in the mitochondrial genome, and multi-gene defects are common in mitochondrial diseases; 5) Insufficient strand-specific editing capability: The lack of efficient and differentiated editing of the H and L strands of mtDNA limits precise treatment strategies targeting mutations on specific strands; 6) Insufficient product specificity: Some existing systems produce a high proportion of unwanted byproducts, such as unexpected base transitions and insertions / deletions, increasing the risk of off-target effects; 7) Insufficient universality: Existing editing systems exhibit significant efficiency differences across different cell types, making it difficult to maintain consistent high-efficiency editing capabilities across various tissue types. These technical bottlenecks severely restrict basic research and clinical translational applications of mitochondrial gene editing technology, necessitating a systematic solution to comprehensively overcome these limitations. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a nuclease fusion protein, a mitochondrial DNA base editing system, and its applications. Based on the characteristic of the CRISPR system's crRNA library to recognize a large number of DNA targets, this invention constructs a library platform capable of high-throughput analysis and library screening of mtDNA targets and effective editing tools. This allows for the systematic characterization of mtDNA-specific editing features, providing a theoretical basis for precise targeting of disease-related mutations. A Cas12 technology capable of editing mtDNA in mouse embryos and live mouse tissues is also established. This invention overcomes the obstacle of conventional sgRNA's inefficient entry into mitochondria, significantly improving the editing efficiency of the CRISPR system in mitochondria, raising the existing efficiency from 0.03-0.23% to a clinically usable level (>30%). It also solves the bacterial toxicity problem in the construction of the mitochondrial gene editing system, improving plasmid yield and stability, facilitating large-scale preparation and application, and maximizing the localization efficiency of the editing components in the mitochondrial matrix. This invention provides a unified mitochondrial gene editing platform that combines high efficiency, high specificity, and broad applicability. It can meet the editing needs of different cell types and different mammalian sources (human and mouse). By utilizing the characteristics of the Cas12 system to process pre-crRNA, it develops mitochondrial multiplex gene editing technology based on "single crRNA array", expands the application scope of editing strategies, and provides a theoretical basis for precise targeting of disease-related mutations.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a nuclease fusion protein comprising a double-mutated APOBEC3A, a Cas12 protein, and a uracil glycosylation inhibitor; wherein the double-mutated APOBEC3A is subject to W104A and Y132D mutations; the Cas12 protein is Cas12m or a double-mutated LbCas12a; wherein the double-mutated LbCas12a is subject to D832A and E925A mutations.

[0007] This invention presents a highly efficient mitochondrial DNA base editing system that integrates catalytically inactivated LbCas12a (dLbCas12a) or Cas12m, optimized cytosine deaminase, a specific mitochondrial targeting sequence, a recombination dimerization domain, and regulatory elements. System characterization reveals unique H-chain bias and broad applicability, enabling in vivo editing in mice and providing a revolutionary tool for targeting disease-related mutations.

[0008] In a preferred embodiment of the nuclease fusion protein described in this invention, the Cas12m is GoCas12m or MmCas12m.

[0009] As a preferred embodiment of the nuclease fusion protein described in this invention, it further includes at least one mitochondrial targeting sequence.

[0010] As a further preferred embodiment of the nuclease fusion protein described in this invention, the mitochondrial targeting sequence is cytochrome c oxidase subunit 8 (Cox) and / or neurosporidian ATPase subunit 9 (Su9).

[0011] As a further preferred embodiment of the nuclease fusion protein of the present invention, the mitochondrial targeting sequence is embedded at at least one site among the C-terminus, N-terminus, Cas12 protein, and uracil glycosylation inhibitor of the nuclease fusion protein.

[0012] As a preferred embodiment of the nuclease fusion protein of the present invention, it further includes at least one dimerizing domain GCN4.

[0013] As a further preferred embodiment of the nuclease fusion protein of the present invention, the dimerization domain GCN4 is embedded after at least one site of the N-terminus, C-terminus, or amino acids 1040, 1074, 1087, 1120, 1142, 1143, and 1158 of the double-mutated LbCas12a.

[0014] As a preferred embodiment of the nuclease fusion protein of the present invention, it further includes a mammal-specific intron; the mammal-specific intron is embedded in the double-mutated APOBEC3A; the mammal-specific intron contains a typical 5' splice donor site (GT) and a 3' splice acceptor site (AG).

[0015] As a preferred embodiment of the nuclease fusion protein of the present invention, its amino acid sequence is shown in any one of SEQ ID NO:2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 32, 46, 48.

[0016] In a second aspect, the present invention provides a nucleic acid encoding a nuclease fusion protein, the nucleotide sequence of which is shown in any one of SEQ ID NO: 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 31, 45, 47.

[0017] Thirdly, the present invention provides a mitochondrial DNA base editing system, comprising the nuclease fusion protein and crRNA described above.

[0018] Fourthly, the present invention applies the described nuclease fusion protein, the described nucleic acid, and the described mitochondrial DNA base editing system in any of the following fields: i. Gene editing; ii. To prepare anti-aging drugs; iii. Delivery of reagents into the mitochondria; v. To prepare medicines for the treatment and / or prevention of mitochondrial diseases; vi. To prepare drugs for the treatment and / or prevention of tumors; vii. To prepare medicines for the treatment and / or prevention of diabetes; viii. To prepare medicines for the treatment and / or prevention of obesity.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention, through systematic structural design and functional optimization, has achieved the following revolutionary technological breakthroughs in the field of mitochondrial gene editing: 1. A qualitative leap in editing efficiency. (1) Unprecedented mitochondrial editing efficiency: The highest C-to-T conversion rate reached 54.84% (ND2 site, after FACS enrichment), and multiple target sites could achieve high-efficiency editing of >30%, which is more than 200 times higher than the existing SpCas9-based system (0.03-0.23%). Even in U-2OS cells that are difficult to transfect, significant editing can be achieved; up to 31 target sites can be edited; high-efficiency editing can be achieved in a variety of cell lines, including human and mouse cell lines.

[0020] (2) Systematic improvement path for editing efficiency: Basic platform (mitoCas12-BE1): average editing efficiency of about 4-5%; introduction of GCN4 dimerization (mitoCas12-BE2): average editing efficiency increased to 20-40%; intron optimization and FACS enrichment (mitoCas12-BE3): editing efficiency further increased to 30-55%; achieving a leap from basic research level to potential clinical application level.

[0021] 2. Editing specificity and accuracy (1) Product specificity: C-to-T conversions accounted for >96% of editing events at most target sites, while undesired C-to-A / G conversions and insertions / deletions remained at <4%, and the product distribution was stable and consistent across different cell types; (2) Target site specificity: The systematized characterization of the editing window location features clarified the chain-specific editing preferences, providing a basis for precise targeting. The self-modification characteristics of PAM sequences provide an intrinsic mechanism to prevent over-editing.

[0022] 3. System's universal applicability (1) Cell type universality: It maintains high editing efficiency in different cell backgrounds such as HEK293T, HeLa and U-2OS, including human and mouse mitochondrial DNA; the efficiency difference between cell types is predictable, mainly reflecting the difference in transfection efficiency, providing a unified tool for basic research on multiple cell types in vitro; mitochondrial editing of mouse embryos and adult mouse liver tissues; (2) Cross-species applicability: It exhibits efficient editing capabilities on both human and mouse mtDNA, and the editing mode remains highly consistent across different species, laying the foundation for animal model research and cross-species applications.

[0023] 4. System Construction and Application Advantages (1) Overcoming key technical obstacles: The compact crRNA characteristics of the Cas12 system enable the editor to enter the mitochondria; the intron strategy effectively solves the bacterial toxicity problem; dual fluorescence enrichment improves the accuracy of editing analysis and significantly increases plasmid yield (>5 times), which is conducive to large-scale application; high-throughput library screening enables the characteristic analysis of the editor and the screening of efficient editors; high-throughput library screening enables the screening of mtDNA functional targets; the generation of pathogenic SNVs on mtDNA is realized; and in vivo editing in mice is realized.

[0024] (2) Potential for multifunctional applications: The first CRISPR-based editable mtDNA base editor, achieving C-to-T and editing up to 31 targets; enabling mtDNA editing in multiple cell lines and human / mouse sources; capable of autonomously editing C in PAM sequences to create new targets; utilizing the ability of LbCas12a to process tandem crRNA to provide a foundation for multi-gene editing; creating the first efficient crRNA library screening platform, greatly simplifying target optimization; and enabling the induction of pathogenic SNVs on mtDNA in human cell lines and live mice.

[0025] 5. Advantages compared to existing technologies (1) Compared with TALE-based systems: No complex protein engineering is required, target design is simplified, the construction process is simpler, the cost is significantly reduced, high-throughput library screening can be achieved, multiple gene editing can be achieved, and it is more suitable for research applications; (2) Compared with SpCas9-based systems: editing efficiency is increased by more than 200 times (from <0.23% to >50%), overcoming the key bottleneck of sgRNA mitochondrial delivery, and the construction size is smaller and more suitable for delivery systems. Attached Figure Description

[0026] Figure 1Construction and evaluation of mitochondrial-targeted dLbCas12a base editor; In the figure, A: crRNA easily enters mitochondria; B: Schematic diagram of three dLbCas12a base editor variants, showing the variants without mitochondrial targeting sequence, 3×Cox8 and 3×Su9; CF: Comparison of C-to-T conversion efficiency at four target sites.

[0027] Figure 2 This is the plasmid map of P11694.

[0028] Figure 3 To enhance editing efficiency based on the GCN4 dimerization strategy; in the figure, A: schematic diagram of the GCN4-mediated dimerization mechanism; B: design of the mitoCas12-BE1 variant embedding GCN4; C: embedding sites of GCN4 within lbCas12a, including E1040, I1074, E1087, D1120, T1142, G1143, and D1158; DG: editing efficiency of all variants at ND1 site 1 (D), ND1 site 2 (E), ND2 site 1 (F), and ND3 site 1 (G). The GCN4 (mitoCas12-BE2) located at 1087 showed the highest editing activity. A (GGGGS)×4 flexible linker embedding at 1087 was used for comparison.

[0029] Figure 4 To improve editing performance through intron insertion and dual-fluorescence enrichment strategies; in the figure, A: schematic diagram of the mitoCas12-BE3 structure, which contains mammalian-specific introns in the Beacon1 coding region that prevent bacterial expression; B: flowchart of the dual-fluorescence enrichment strategy: mitoCas12-BE3 (mCherry) and crRNA (EGFP) plasmids enrich dual-fluorescence positive cells by flow cytometry sorting (FACS); CF: comparison of the editing efficiency of mitoCas12-BE2, mitoCas12-BE3, and FACS-enriched mitoCas12-BE3 at ND1 site 1 (C), ND1 site 2 (D), ND2 site 1 (E), and ND3 site 1 (F). The highest editing efficiency of ND2 site 1 was 54.84% after enrichment.

[0030] Figure 5 The editing efficiency of mitoCas12-BE3 in HEK293T, HeLa, and U-2OS cells is shown in the figure. The editing efficiency of mitoCas12-BE3 on ND1 site 1 (A), ND site 2 (B), ND2 site 1 (C), and ND3 site 1 (D) in HEK293T, HeLa, and U-2OS cells is shown in the figure.

[0031] Figure 6 To validate the multi-gene site of mitoCas12-BE3 in human and murine mtDNA; in the figure, A: C-to-T editing efficiency of 10 human mtDNA targets in HEK293T cells, including examples of elimination or generation of stop codons at TRNH site 1 and ATP8 site 1; B: C-to-T editing efficiency of 15 murine mtDNA targets in AML12 cells, including nonsense mutations at CYTB site 1, ND4 site 1 and ND1 site 2; CD: Analysis of editing products of all targets, with C-to-T being the major editing product, and C-to-A / G and indel being byproducts with the smallest proportion.

[0032] Figure 7 To demonstrate multiplex gene editing of human and mouse mtDNA using a "single crRNA array"; Figure A: Schematic diagram of cleavage of the "single crRNA array" by Cas12a RNase activity to generate three mature and isolated crRNAs; B: Multiplex gene editing of three sites (ND1 site 1, ND2 site 1, and ND3 site 1) by mitoCas12-BE3 in HEK293T cells; C: Multiplex gene editing of three sites (CYTB site 2, D-loopsite 1, and TRNL1 site 1) by mitoCas12-BE3 in AML12 cells.

[0033] Figure 8 To characterize editing properties using a crRNA library system; in the figure, A: analysis flow of mtDNA crRNA library; B and C: editing efficiency of the mitoCas12-BE system in human mtDNA H chain (B) and L chain (C); D: comparison of average C-to-T editing efficiency of human mtDNA L chain and H chain; E and F: editing efficiency of the mitocas12-BE system in mouse mtDNA H chain (E) and L chain (F); G: comparison of average C-to-T editing efficiency of mouse mtDNA L chain and H chain; H and I: comparison of C-to-T editing efficiency of Beacon1 (mitoCas12-BE3) and 6 cytosine deaminases in HEK293T cells (H) and AML12 cells (I); J and K: changes in crRNA abundance after mitoCas12-BE3 editing HEK293T cells (J) and AML12 cells (K).

[0034] Figure 9To illustrate the induction of pathogenic mtDNA SNVs using mitoCas12-BE3 in human cell lines and in vivo mice; in the figure, A: Schematic diagram of premature generation of stop codons on human CYTB site 1 via C18-to-T conversion; B: Editing efficiency of CYTB site 1 in HEK293T cells; C: Increased ROS levels in HEK293T cells after truncating CYTB protein; D and E: Schematic diagram of premature stop codons and pathogenic SNVs in mouse CYTB site 1 (D) and ND1 site 1 (E); F: Schematic diagram of mitoCas12-BE3 systemic editing of mouse embryos; G and H: Editing efficiency of mitoCas12-BE3 on CYTB site 1 and ND1 site 1 in young mice; I: Schematic diagram of mitoCas12-BE3 systemic editing of mouse liver; J and K: Editing efficiency of mitoCas12-BE3 on CYTB site 1 and ND1 site 1 in mouse liver.

[0035] Figure 10 Mitochondrial DNA editing was achieved using two Cas12m editors; in the figure, A: schematic diagram of the Cas12m editor. GoCas12m and MmCas12m were used to replace LbCas12a, respectively; BE: editing efficiency of the two Cas12m editors at four target sites. Detailed Implementation

[0036] To better illustrate the objectives, technical solutions, and advantages of this invention, the invention will be further described below with reference to specific embodiments. Those skilled in the art should understand that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0037] Unless otherwise specified, the experimental methods used in the examples are conventional methods; the materials and reagents used are commercially available unless otherwise specified.

[0038] The plasmid for the Cas12 variant in this example was constructed as follows: pCMV-BEACON1 (Addgene#171697) was used as the basic backbone, PCR amplification was performed using KAPA HiFi HotStart Ready Mix (KAPA Biosystems#KK2602), seamless cloning was performed using 2X MultiF Seamless Assembly Mix (ABclonal#RK21020), and PuroR and BSD resistance genes were introduced as selection markers.

[0039] In the example, the crRNA expression plasmid was constructed as follows: based on the pU6-LbCas12a-crRNA backbone (modified from Addgene#84751), the target sequence was inserted by oligonucleotide annealing, the U6 promoter drove the crRNA transcriptional expression, and the EGFP expression cassette was integrated for dual fluorescence screening.

[0040] The cell culture conditions in this example were as follows: HEK293T cells: DMEM (Gibco#11965092), HeLa cells: DMEM (Gibco#11965092), U-2OS cells: McCoy's 5a Medium (Gibco#16600082); all culture media were supplemented with 1% penicillin-streptomycin (Gibco#15140122) and 10% fetal bovine serum (Gibco#A5670801), and cultured at 37°C under 5% CO2 conditions.

[0041] The transfection parameter optimization method in this example is as follows: 24-well plate, 5×10^4 cells seeded per well, cultured to 70% confluence, 750 ng mitoCas12-BE plasmid, 500 ng crRNA plasmid, using Hieff Liposome transfection reagent (Yeasen#40802ES08) was used. Cells were harvested and analyzed 5-7 days after transfection.

[0042] The editing efficiency analysis method in this embodiment is as follows: (1) Genomic DNA extraction and PCR amplification: using Genomic DNA Extraction Kit (Transgen#TEE101-01), two-step PCR construction of sequencing libraries: Step 1, 25 cycles, introduces gene-specific sequences; Step 2, 11 cycles, adds double indexing, purifies amplification products with magnetic beads, and mixes in equimolar amounts to construct sequencing libraries; (2) Deep sequencing and data analysis: Sequencing was performed on the MGI2000 platform, raw data quality control was performed using FASTP, paired-end reads were merged using FLASH software, and the type and efficiency of the edited products were analyzed using CRISPResso software. The editing efficiency was calculated as: C-to-T conversion reads at the target site / total reads × 100%.

[0043] The method for detecting reactive oxygen species (ROS) levels in this embodiment is as follows: Following the manufacturer's instructions, ROS levels were assessed using a ROS detection kit (Beyotime, S0033M). HEK293T cells were seeded in 6-well plates and transfected with 1000 ng mitoCas12-BEs plasmid and 3000 ng crRNA plasmid, respectively. 72 h post-transfection, cells were washed three times with PBS. Cells were incubated with 10 μM DCFH-DA probe at 37°C for 10 min. After incubation, cells were washed again to remove uninternalized probe. Intracellular ROS levels were detected by flow cytometry, with the intensity of DCFH-DA fluorescence used to assess ROS levels.

[0044] The RNA preparation and mouse embryo injection methods in this embodiment are as follows: mRNA was prepared using the HiScribe T7 ARCA mRNA Kit (tailed) (NEB, E2060S) according to the manufacturer's operating procedures. crRNA was chemically modified by thiophosphorylation and methoxy groups, and synthesized using GenScript (Nanjing, China). The RNA mixture contained 200 ng / μL mitoCas12-BE mRNA and 300 ng / μL crRNA. Adult female C57BL / 6 mice were injected with maternal serum gonadotropin and human chorionic gonadotropin (hCG) 48 h later. Subsequently, the female mice were mated with male mice. The female mice were euthanized with 40% CO2 12–24 h later. Fertilized eggs were collected from the oviducts. 8 μL of the RNA mixture was microinjected into the cytoplasm of the fertilized eggs. The injected oocytes were cultured at 37°C and 5% CO2 and transplanted into replacement mice the following day. Offspring were born 19–20 days post-transplantation. The mouse animal experiments were approved by the Women and Children's Hospital Affiliated to Zhongnan Hospital of Wuhan University.

[0045] The method of orthotopic injection into the mouse liver in the example is as follows: The mitoCas12-BE3 system and luciferase were packaged separately as chronic viruses. The two lentiviruses were mixed at a viral titer of 2:1. Mice were anesthetized and fixed ventrally upwards, and an incision of approximately 1–2 cm was made along the midline of the abdomen. The liver was gently extracted from the abdominal cavity using a cotton swab. One x 10^7 TU of the lentivirus mixture was slowly injected into the liver through a fine needle. The abdominal incision was closed with absorbable sutures. Two weeks later, mice were anesthetized and injected with 200 μL of 15 mg / ml d-luciferin potassium (Beyotime, ST196). Fifteen minutes later, the mice were euthanized with 40% CO2, and the liver was dissected and removed. The liver was irradiated for 20 seconds for fluorescence imaging. DNA was extracted and amplicon sequencing was performed from the liver region with the strongest fluorescence.

[0046] The amplicon sequencing method in this embodiment is as follows: use Genomic DNA was extracted from transfected cells using a genomic DNA kit (Transgen, TEE101-01). Genomic DNA was used for two-step PCR with KAPA HiFi HotStart Ready Mix. The first-step PCR primer pair, with 25 cycles, contained a gene-specific region and an external Illumina-compatible adapter sequence (see Table 1). The second-step PCR primer pair had 11 cycles and included the external adapter sequence. The amplicon library was then sequenced on the MGI2000 platform. After quality control of the raw sequencing data using default FASTP parameters, pair-end reads were merged into a single read using FLASH software. Base-editing products were determined by CRISPResso alignment with the mtDNA genomic sequence. The percentage of edited product types, i.e., editing efficiency, was calculated as the desired edited reads divided by the total sequencing reads for a given crRNA region.

[0047] Example 1: Construction and evaluation of a mitochondrial-targeted dLbCas12a base editor 1. Design and Construction of Mitochondrial Base Editor Component To design the basic architecture of a mitochondrial-targeted CRISPR-Cas12a base editing system, a fusion protein containing the following functional components was constructed: (1) Cytosine deaminase module: The high-fidelity APOBEC3A variant hA3A (W104A / Y132D) optimized by mutation was selected. This cytosine deaminase variant is also known as Beacon1. Beacon1 will be used as the name of this deaminase below. This variant has a low off-target rate while maintaining high catalytic activity. (2) Nuclease vector: The catalytically inactivated LbCas12a (dLbCas12a) is used, which contains double mutations of D832A and E925A, losing DNA cutting ability but retaining precise DNA targeting ability; (3) Repair inhibition component: Integrate uracil glycosylation inhibitor (UGI) to prevent base excision repair at the edit site and improve editing stability; (4) Mitochondrial targeting strategy: Three variants were systematically evaluated: a. Non-targeted control group (without mitochondrial targeting sequence), named Beacon1-dLbCas12a-UGI, with the following sequence structure: 16aa linker+Beacon1+16aa linker+dLbCas12a+4aa linker+UGI+4aa linker; its nucleotide sequence is shown in SEQ ID NO:1, and its amino acid sequence is shown in SEQ ID NO:2.

[0048] b. 3×Cox8-MTS construct: containing three tandem cytochrome c oxidase subunit 8 mitochondrial targeting sequences (i.e., Cox8), named Beacon1-dLbCas12a-UGI-3x Cox8, with the following sequence structure: Cox8+16aa linker+Beacon1+16aa linker+dLbCas12a+2aa linker+Cox8+4aa linker+UGI+4aa linker+Cox8; its nucleotide sequence is shown in SEQ ID NO:3, and its amino acid sequence is shown in SEQ ID NO:4.

[0049] c. 3×Su9-MTS construct: Contains three tandemly linked mitochondrial targeting sequences of the Neurospora ATPase subunit 9 (i.e., Su9), named Beacon1-dLbCas12a-UGI-3x Su9 (mitoCas12-BE1), with the following sequence structure: Su9+16aalinker+Beacon1+16aa linker+dLbCas12a+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:5, and its amino acid sequence is shown in SEQ ID NO:6.

[0050] All three constructs were constructed using the Gibson assembly method, placed in mammalian expression vectors, and expressed via the CMV promoter. Detailed structures of the constructs are shown below. Figure 1 As shown in B.

[0051] 2. crRNA Design and Target Site Selection To evaluate the functionality of the constructed editor, crRNAs targeting four sites in human mtDNA were designed: two sites in the ND1 gene (ND1 crRNA1 and ND1 crRNA2); one site in the ND2 gene (ND2 crRNA); and one site in the ND3 gene (ND3 crRNA).

[0052] These target sites all meet the TTTV PAM requirements recognized by LbCas12a (where V represents A, C, or G) and contain a target C site that can be recognized by cytosine deaminase. crRNA expression was driven by the U6 promoter (plasmid: P11694, addgene; plasmid map see [link]). Figure 2Downstream of the U6 promoter is an sgRNA sequence. Replacing the sgRNA with a crRNA sequence yields a crRNA plasmid (approximately 43 nt in length). The crRNA sequence structure includes: AATTTCTACTAAGTGTAGAT (19 nt DR region) GGCTACTGCTCGCAGTGCGCCGA (23 nt spacer sequence, i.e., the target sequence; the current target sequence identifies ND1 site 1. This sequence can be replaced with other target sequences listed in Table 1 to edit other gene sites). Replace the sgRNA sequence inside plasmid P11694 (addgene) with the 19 nt DR sequence and the 23 nt target sequence.

[0053] The target site sequences and primer sequences of the relevant mtDNA are shown in Table 1: Table 1. Sequences of target sites and primers 3. mtDNA editing efficiency assessment Three base editor constructs were co-transfected into HEK293T cells with four different crRNA plasmids. Cells were harvested 120 hours post-transfection, and total DNA was extracted and analyzed using amplicon sequencing (amplicon primers are listed in Table 1). Editing efficiency was defined as the percentage of C-to-T conversion reads at the target C site relative to the total reads. The experimental results showed: In the absence of mitochondrial targeting sequences (i.e., the variant Beacon1-dLbCas12a-UGI), C-to-T conversion was negligible (<1%). Beacon1-dLbCas12a-UGI-3xCox8 produced editing efficiencies of 4.26% and 2.30% at ND1 site 1 and ND2 site 1, respectively, while showing no significant improvement at the other two sites. Figure 1 (CF). In comparison, Beacon1-dLbCas12a-UGI-3xSu9 significantly improved editing efficiency, reaching 9.08% on ND1 site 1 (2.13 times higher than Beacon1-dLbCas12a-UGI-3xCox8), 0.32% on ND1 site 2 (3.59 times higher), 5.15% on ND2 site 1 (2.24 times higher), and 4.91% on ND3 site 1 (8.32 times higher). Figure 1 CF). Surprisingly, this Beacon1-dLbCas12a-UGI-3xSu9 catalyzed the conversion of PAM at ND3 site 1 from TTTC to TTTT ( ). Figure 1E). This PAM-containing editing activity expands the editable cytosine range beyond the conventional window (note: other base editors cannot edit PAM) and may help limit repeated editing and reduce the generation of nonspecific products.

[0054] These results indicate that Su9 is a good mitochondrial targeting sequence for introducing dLbCas12a into mitochondria. In this study, Beacon1-dLbCas12a-UGI-3xSu9 was designated as mitoCas12-BE1 and used as the base platform for subsequent optimization.

[0055] Example 2: Enhancing Mitochondrial Base Editing Efficiency Based on GCN4 Dimerization Strategy 1. Protein dimerization principles and strategy design Although the mitoCas12-BE1 system developed in Example 1 achieved base editing of mtDNA, its editing efficiency was low, which is insufficient for therapeutic applications. A protein dimerization strategy was designed to enhance the local concentration of the editor at the target site.

[0056] The GCN4 leucine zipper domain was chosen as the dimerization mediator. GCN4 is a 19-amino acid dimerization domain (sequence: EELLSKNYHLENEVARLKK), derived from a yeast transcription factor, capable of forming stable α-helical dimers through hydrophobic interactions (see...). Figure 3 A).

[0057] 2. System optimization of GCN4 domain integration location To systematically evaluate the GCN4-mediated dimerization effect, the crystal structure of LbCas12a (PDB:5XUT) was analyzed, and seven sites within the Nuc domain were selected for systematic evaluation, one site each at the N-terminus and C-terminus of LbCas12a. Figure 3 D). Finally, a strategy variant of mitoCas12-BE1 was designed ( Figure 3 These variants (DG) are constructed using a multi-fragment Gibson assembly method, with nine embedding GCN4 and one embedding a (GGGGS)×4 flexible linker. Details are as follows: a. N-terminal convergence (N-GCN4) A GCN4 domain was fused to the N-terminus of mitoCas12-BE1; named mitoCas12-BE1-1GCN4(N-terminal), with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1st aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:7, and its amino acid sequence is shown in SEQ ID NO:8.

[0058] b. Consumer-side integration (C-GCN4) A GCN4 domain was fused to the C-terminus (position 1228) of mitoCas12-BE1; named mitoCas12-BE1-1228GCN4(C-terminal), with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aalinker+dLbCas12a embedded GCN4 at 1228th aa+2aa linker+Su9+4aa linker+UGI+4aalinker+Su9; its nucleotide sequence is shown in SEQ ID NO:9, and its amino acid sequence is shown in SEQ ID NO:10.

[0059] c. Internal embedding (1040-GCN4) A GCN4 domain is embedded at position 1040 (i.e., after the 1040th amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1040GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1040th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:11, and its amino acid sequence is shown in SEQ ID NO:12.

[0060] d. Internal embedding (1074-GCN4) A GCN4 domain is embedded at position 1074 (i.e., after the 1074th amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1074GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1074th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:13, and its amino acid sequence is shown in SEQ ID NO:14.

[0061] e. Internal embedding (1087-GCN4) A GCN4 domain is embedded at position 1087 (i.e., after the 1087th amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1087GCN4 (mitoCas12-BE2), with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1087thaa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:15, and its amino acid sequence is shown in SEQ ID NO:16.

[0062] f. Internal embedding (1120-GCN4) A GCN4 domain is embedded at position 1120 (i.e., after the 1120th amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1120GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1120th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:17, and its amino acid sequence is shown in SEQ ID NO:18.

[0063] g. Internal embedding (1142-GCN4) A GCN4 domain is embedded at position 1142 (i.e., after the 1142nd amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1142GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1142th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:19, and its amino acid sequence is shown in SEQ ID NO:20.

[0064] h. Internal embedding (1143-GCN4) A GCN4 domain is embedded at position 1143 (i.e., after the 1143rd amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1143GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1143th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:21, and its amino acid sequence is shown in SEQ ID NO:22.

[0065] i. Internal embedding (1158-GCN4) A GCN4 domain is embedded at position 1158 (i.e., after the 1158th amino acid of dLbCas12a) within the Nuc domain of dLbCas12a; named mitoCas12-BE1-1158GCN4, with the following sequence structure: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GCN4 at 1158th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:23, and its amino acid sequence is shown in SEQ ID NO:24.

[0066] j. Flexible connector control (1087-GS) An equal-length (GGGGS)×4 flexible linker was inserted at position 1087 inside the Nuc domain of dLbCas12a as a non-dimerization control; named mitoCas12-BE1-1087Gslinker, its sequence structure is: Su9+16aa linker+hA3A(W104A / Y132D)+16aa linker+dLbCas12a embedded GSlinker at1087thaa+2aalinker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:25, and its amino acid sequence is shown in SEQ ID NO:26.

[0067] These constructs were built using the multi-fragment Gibson assembly method, validated by sequencing, and then functionally evaluated (see [link to Gibson assembly]). Figure 3 B).

[0068] 3. Comparative Analysis of Editing Efficiency The GCN4 variant was co-transfected with crRNA plasmids targeting four different sites into HEK293T cells, and the C-to-T conversion efficiency was analyzed by amplicon sequencing. The results showed that: Embedding GCN4 within the Nuc domain significantly increased the C-to-T conversion at all four targets, while the editing efficiency improvement of variants fused with N-terminus or C-terminus was minimal or negligible. Figure 3 Interestingly, the degree of improvement in editing efficiency depends on the embedding site, which may be due to differences in dimerization conformation.

[0069] Of all variants, the variant embedding GCN4 at position 1087 (i.e., mitoCas12-BE1-1087GCN4 listed below) produced the highest editing activity: 24.02% at ND1 site 1 (3.10 times that of mitoCas12-BE1), 4.01% at ND1 site 2 (7.51 times), 30.44% at ND2 site 1 (5.66 times), and 27.72% at ND3 site 1 (6.37 times). Figure 3 (DG). In contrast, embedding a (GGGGS)×4 flexible linker at position 1087 failed to improve editing performance, suggesting that its benefit comes from GCN4-mediated dimerization. We named the editor with GCN4 embedded at position 1087 our second-generation base editor mitoCas12-BE2.

[0070] Example 3: Further enhancement of mitoCas12-BE2 through intron optimization and dual fluorescence enrichment strategies 1. Intron optimization strategies to overcome bacterial toxicity In protein engineering, significant toxicity was observed in the expression of Beacon1-containing constructs in bacteria, leading to low plasmid yield and frequent escape mutations. To address this issue, an innovative intron insertion strategy was designed: Intron design: A 235 bp mammal-specific intron was inserted into the Beacon1 coding sequence, containing a typical 5' splice donor site (GT) and a 3' splice acceptor site (AG); Mammalian-specific expression: Since bacteria lack intron splicing mechanisms, this design can block the functional expression of Beacon1 in bacteria, while producing the complete protein through normal splicing in mammalian cells; Construction and Validation: Intron sequences were inserted into the mitoCas12-BE2 construct using the Gibson assembly method, named mitoCas12-BE2-intron (mitoCas12-BE3). Functional evaluation was performed after sequencing validation (see [link to documentation]). Figure 4 A).

[0071] Experimental results confirmed that the intron insertion strategy effectively eliminated bacterial toxicity, increased plasmid yield (>5-fold), and did not affect the editing function in mammalian cells.

[0072] 2. Optimization of Dual Fluorescence Enrichment Strategy To further improve system efficiency, an innovative dual-fluorescence expression and enrichment strategy was developed: Design of a dual-plasmid system: First plasmid: The plasmid expressing mitoCas12-BE2 also expresses mCherry, driven by the CMV promoter; Second plasmid: The plasmid expressing crRNA also integrates EGFP expression, driven by the CMV promoter. FACS enrichment strategy: 72 hours post-transfection, double-positive (mCherry) cells were separated using flow cytometry (FACS). + / EGFP + Cell population, ensuring that the cells analyzed contain both the editor and crRNA components (see cell population). Figure 4 B); The nucleotide sequence of mCherry is shown in SEQ ID NO:27, and the amino acid sequence is shown in SEQ ID NO:28. The nucleotide sequence of EGFP is shown in SEQ ID NO:29, and the amino acid sequence is shown in SEQ ID NO:30.

[0073] 3. Editing efficiency assessment Among the four target sites, the variant with only intron addition (mitoCas12-BE3 group) showed enhanced C-to-T conversion efficiency compared to mitoCas12-BE2, while cells obtained by flow cytometry sorting based on a dual fluorescence enrichment strategy (mitoCas12-BE3 with FACS group) had higher editing efficiency. Figure 4 Compared with the unsorted mitoCas12-BE2 population, the editing efficiencies of enriched cells at ND1 site 1, ND1 site 2, and ND3 site 1 were 33.77% (1.44-fold), 5.02% (1.79-fold), and 35.29% (1.41-fold), respectively, with the highest editing efficiency at ND2 site 1 at 54.84% (average 49.91%), which was 1.77-fold.

[0074] This indicates that the dual-fluorescence enrichment strategy can effectively improve editing efficiency, and the 54.84% editing efficiency of ND2 site 1 sets a new historical high for CRISPR systems in mitochondrial editing. Figure 4 E).

[0075] The sequence structure of mitoCas12-BE2-intron (mitoCas12-BE3) is as follows: Su9+16aa linker+hA3A(W104A / Y132D)with intron+16aa linker+dLbCas12a embedded GCN4 at 1087th aa+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:31, and its amino acid sequence is shown in SEQ ID NO:32.

[0076] Example 4: Cross-cell line validation To evaluate the generalizability of the mitoCas12-BE3 and dual-fluorescence enrichment strategy in Example 3, mtDNA editing was performed in three human cell lines: Cell lines selected: HEK293T (human embryonic kidney cell line), HeLa (human cervical cancer cell line), and U-2OS (human osteosarcoma cell line).

[0077] Editing efficiency comparison: The editing efficiency of each cell line was evaluated at four target sites.

[0078] Experimental results showed that mitoCas12-BE3 maintained robust editing capabilities in all tested cell lines, although efficiency varied with cell type: After dual-fluorescence flow cytometry sorting, mitoCas12-BE3 achieved approximately 1 / 2 and 1 / 3 of the editing efficiency in HeLa cells and U-2OS cells, respectively, compared to HEK293T cells. Notably, the editing efficiency at ND2 site 1 remained high in HeLa cells, reaching 39.02% (compared to 46.36% in HEK293T cells). Figure 5 AD).

[0079] This result confirms the broad applicability of the mitoCas12-BE3 system in different cellular contexts, and the efficiency differences may reflect differences in cell transfection efficiency, mitochondrial number, and mtDNA copy number.

[0080] Example 5: Multi-target validation in human and mouse mtDNA All mtDNA target site sequences and amplicon primer sequences in this embodiment are listed in Table 1.

[0081] To assess the breadth and specificity of mitoCas12-BE3, the editing test was extended to other endogenous sites of human and mouse mtDNA.

[0082] In human HEK293T cells, mitoCas12-BE3 efficiently edited 10 target sites ( Figure 6 A). Seven targets (TRNH site 1, ATP8 site 1, TRNN site 1, RNR site 1, ND5 site 1, ATP6 site 1, and RNR2 site 1) showed 20%–35% efficiency, while D-loop site 1, ND5 site 2, and COX2 site 1 showed approximately 10% editing efficiency.

[0083] Next, 15 targets were evaluated in AML12 cells from the source mice. Figure 6 B). Among the 15 targets, the editing efficiency of 1 target is greater than 30%, the editing efficiency of 5 targets is 20-30%, the editing efficiency of 4 targets is 10-20%, and the editing efficiency of 4 targets is 7-10%; while the editing efficiency of RNR2 site 2 is 2.5%.

[0084] Notably, mitoCas12-BE3 can induce the generation or elimination of stop codons, indicating its significant potential in correcting pathogenic point mutations in mtDNA. In human TRNH site 1, an editing efficiency of 30.49% was observed at C-1 and 26.71% at C26; C26-to-T converted the stop codon AGA to AAA, extending the open reading frame of the upstream ND4 protein by 32 amino acids. In human ATP8 site 1, C6-to-T (24.4% editing efficiency) prematurely terminated ATP8, shortening it from 68 amino acids to 50 amino acids. In mouse CYTB site 1, mitoCas12-BE3 achieved editing rates of 32.41% and 2.99% at C19 and C25, respectively; both C-to-T positions generated stop codons, truncating CYTB from 392 amino acids to 164 amino acids and 162 amino acids, respectively. Early termination can also be achieved at ND4 site 1 and ND1 site 2 in mice.

[0085] The editing results of mitoCas12-BE3 were highly specific. Of the 14 human and 15 mouse targets tested, 27 targets had C-to-T product purity ≥86%, with 13 targets having purity ≥96%. Figure 5 (CD). This combination of breadth, efficiency, and specificity underscores the accuracy, durability, and security of the mitoCas12-BE3.

[0086] In summary, these data demonstrate that mitoCas12-BE3 possesses robust editing capabilities at multiple mtDNA targets in both humans and mice.

[0087] Example 6: Multiplex gene editing of human and mouse mtDNA using a "single crRNA array" Cas12a can process pre-crRNA into a single mature crRNA. To perform multiplex gene editing in mtDNA, a single crRNA array with three spacer sequences was designed and coupled with mitoCas12-BE3 for editing (…). Figure 7 A).

[0088] In HEK293T cells, a single crRNA array can simultaneously achieve C-to-T editing of ND1 site 1, ND2 site 1, and ND3 site 1, with average editing efficiencies of 23.21%, 19.25%, and 13.85%, respectively. Figure 7B). In AML12 cells, single crRNA arrays targeting CYTB site 2, D-loop site 1, and TRNL1 site 1 also produced C-to-T editing with average efficiencies of 14.82%, 17.64%, and 14.82%, respectively. Figure 7 C).

[0089] Therefore, mitoCas12-BE3 supports multiplex gene editing of human and mouse mtDNA, which greatly reduces the number of transcripts and simplifies the difficulty of delivering crRNA that simultaneously targets multiple genes.

[0090] Example 7: Characterization of the editing properties of mitoCas12-BE3 and analysis of mitochondrial DNA editing properties using a crRNA library system 1. crRNA library construction and screening strategies Compared to the TALE system, the CRISPR system only requires simple guide RNA programming to target different sequences, greatly simplifying the target site screening and optimization process. Leveraging this advantage, a systematic crRNA library screening platform was developed, and two crRNA libraries were constructed, targeting 645 human mtDNA targets and 844 mouse mtDNA targets, respectively, for large-scale screening in a high-throughput manner. Figure 8 A). The library construction method was as follows: Ten cycles of PCR amplification were performed using KAPA HiFi HotStart Ready Mix to amplify an oligonucleotide pool containing crRNA, followed by purification using magnetic beads (Cytiva, 65152105050250). The Lenti-gRNA-Puro plasmid (Addgene, 84752) was linearized using PCR. The pooled PCR product was introduced into the linearized Lenti-gRNA-Puro plasmid using a 2X MultiF seamless assembly mix. The recombinant plasmid was transformed into *E. coli*, and colonies were scraped for plasmid purification. The lentiviral plasmid library was transfected into HEK293T cells to construct the lentiviral library.

[0091] A crRNA library with an MOI of 20 was transduced into HEK293T or AML12 cells to construct a crRNA cell library. Each crRNA should cover at least 500 cells. After 24 hours of transduction, the crRNA cell library was cultured for 10 days in fresh medium containing 2.5 μg / mL puromycin. Subsequently, the mitoCas12-BEs plasmid was transfected into the crRNA cell library. Genomic DNA was extracted from the cells for long-fragment amplicon sequencing.

[0092] The results show: Editing efficiency analysis revealed a significant asymmetric pattern between strands: library analysis showed that the mitoCas12-BE3 system exhibited extremely strong strand bias. In human mtDNA, the H strand showed editing efficiency greater than 1% at C1, C2, C7, C18, C19, C22, and TTTC PAM, while the L strand showed 0.2-0.4% activity at C14, C19, and TTTC PAM. Figure 8 Similarly, in mouse mtDNA, the H chain showed editing efficiency exceeding 1% at C1, C5-C9, C19, and TTTC PAM, while the L chain maintained 0.1-0.3% at C4, C7, C22, and TTTC PAM. Figure 8 EF). On average, the editing efficiency of the H chain was 8.05 times (humans) and 9.02 times (mice) higher than that of the L chain. Figure 8 D and G). A similar preference was also observed in mitoCas12-BE1-1040GCN4 ( Figure 8 (BG). This significant chain preference provides important guidance for mitochondrial editing, particularly for disease-related mutations located on different chains, allowing for the optimization of editing strategies.

[0093] Efficient screening of mtDNA-editable deaminases: Beacon1 was replaced with six cytosine deaminases to evaluate their editing performance in mtDNA. These six cytosine deaminases were evoAPOBEC1, evoCDA1, evoFERNY, pmCDA1, CBE6b, and CBE6d. In HEK293T cells, CBE6d exhibited the highest editing activity at C1-C4, followed by evoAPOBEC1. Figure 8 H). In AML12 cells, the activities of evoAPOBEC1 and Beacon1 were comparable, while the activities of the other deaminases were relatively poor. Figure 8 These results indicate that the activities of different deaminases are highly cell type dependent and suggest that CBE6d and evoAPOBEC1 are promising alternatives for mtDNA editing.

[0094] Efficient screening of mtDNA functional sites: In HEK293T cells, after editing with mitoCas12-BE3, the abundance of 219 crRNAs increased, and the abundance of 51 crRNAs decreased. Figure 8 J). In AML12 cells, the abundance of 35 crRNAs was upregulated and the abundance of 4 crRNAs was downregulated (J). Figure 8 Decreased abundance of crRNA may indicate that the target impairs the cell's ability to proliferate or survive.

[0095] Furthermore, under all test conditions, the C in the PAM sequence TTTC could be effectively edited to T, providing a mechanistic basis for multi-round editing. This high-throughput platform allows for the systematic characterization of the editing properties and preferences of mitochondrial editing systems. The similar editing characteristics observed in human and mouse mtDNA demonstrate the conservation of the developed mitochondrial editing system mitoCas12-BE and its potential for cross-species application.

[0096] In summary, crRNA libraries can be used to more effectively identify high-efficiency targets in mtDNA, analyze the editing characteristics of different deaminases, and search for pathogenic or special-function targets in mtDNA.

[0097] The sequence information for the above six cytosine deaminases is as follows. Replacing mitoCas12-BE3 with the corresponding deaminase sequence will edit the mtDNA. The nucleotide sequence of evoAPOBEC1 is shown in SEQ ID NO:33, and the amino acid sequence is shown in SEQ ID NO:34. The nucleotide sequence of evoCDA1 is shown in SEQ ID NO:35, and the amino acid sequence is shown in SEQ ID NO:36. The nucleotide sequence of evoFERNY is shown in SEQ ID NO:37, and the amino acid sequence is shown in SEQ ID NO:38. The nucleotide sequence of pmCDA1 is shown in SEQ ID NO:39, and the amino acid sequence is shown in SEQ ID NO:40. The nucleotide sequence of CBE6b is shown in SEQ ID NO:41, and the amino acid sequence is shown in SEQ ID NO:42. The nucleotide sequence of CBE6d is shown in SEQ ID NO:42, and the amino acid sequence is shown in SEQ ID NO:43.

[0098] Example 8: Implantation of pathogenic SNVs in human cells and in vivo mice using mitoCas12-BE3 To construct a cell model of impaired mitochondrial function, a TAA stop codon was generated at C18 of CYTBsite 1 in HEK293T cells using mitoCas12-BE3, truncating the CYTB protein from 386 aa to 140 aa. Figure 9 A). In HEK293T cells, editing CYTB site 1 resulted in an average C18-to-T editing efficiency of 18.19% ( Figure 9 B). Functionally, the levels of reactive oxygen species (ROS) in these edited HEK293T cells were 1.65 times that of the control group, indicating that CYTB protein truncation leads to impaired mitochondrial respiration. Figure 9 C).

[0099] Next, the feasibility of in vivo editing of mitoCas12-BE3 on mouse CYTB site 1 (which generates two premature stop codons) and ND1 site 1 (corresponding to pathogenic SNVs in human Leber hereditary optic neuropathy, i.e., LHON syndrome) was evaluated. Figure 9 DE). Microinjection of transcribed mitoCas12-BE3 mRNA and crRNA into mouse embryos resulted in detectable editing in young mice. Figure 9 F). On CYTB site 1, the average editing efficiency of C19-to-T was 5.02%, with a maximum of 6.46%; while the average editing efficiency of C25-to-T was 0.41%. Figure 9 G). On ND1 site 1, the average editing efficiency of C5-to-T is 2.58% ( Figure 9 H). To test in vivo tissue editing, mitoCas12-BE3 was packaged into a lentivirus and injected into the liver of mice (H). Figure 9 I). Two weeks post-injection, at CYTB site 1, the mean edit-incision rate for C19-to-T was 2.58%, and for C25-to-T, it was 0.26%. At ND1 site 1, the mean edit-incision rate for C5-to-T was 1.80%. Figure 9 JK).

[0100] In summary, these data demonstrate that mitoCas12-BE3 can induce pathogenic SNVs in human cells, mouse embryos, and tissues, providing important experimental data support for the study of the mechanisms of mtDNA diseases and the clinical evaluation of mitochondrial therapies.

[0101] Example 9: mtDNA editing using the Cas12m system To further expand the Cas12 mitochondrial editing system, LbCas12a in Beacon1-dLbCas12a-UGI-3x Su9(mitoCas12-BE1) was replaced with GoCas12m and MmCas12m.

[0102] The Beacon1-GoCas12m-UGI-3xSu9 sequence structure is: Su9+16aa linker+Beacon1+16aalinker+GoCas12m+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:45, and its amino acid sequence is shown in SEQ ID NO:46.

[0103] The Beacon1-MmCas12m-UGI-3xSu9 sequence structure is: Su9+16aa linker+Beacon1+16aalinker+MmCas12m+2aa linker+Su9+4aa linker+UGI+4aa linker+Su9; its nucleotide sequence is shown in SEQ ID NO:47, and its amino acid sequence is shown in SEQ ID NO:48.

[0104] Editing efficiency was tested on four target sites: ND1 site 1, ND1 site 2, ND2 site 1, and ND3 site 1. Figure 10 As shown, at ND1 site 1, ND2 site 1, and ND3 site 1, the editing efficiency of Beacon1-GoCas12m-UGI-3xSu9 reached approximately 25%, while Beacon1-MmCas12m-UGI-3xSu9 achieved approximately 25%, 10%, and 15%, respectively. However, at ND1 site 1, the editing efficiency of both variants was only about 3%. These results indicate that multiple Cas12 systems can achieve highly efficient editing of mitochondrial mtDNA.

[0105] The application technologies of this invention include: Chain-specific editing mechanisms: the differential editing efficiency and patterns of the system on the H and L chains of mtDNA, the mechanism basis and application of efficient H chain editing (7-9 times that of L chain), chain-specific target selection and optimization strategies, and targeting methods for pathogenic mutations on specific chains; Multi-gene editing methods: using LbCas12a to process tandem crRNA arrays to achieve multi-target editing; the mechanism of PAM sequence modification to generate new targets and its tandem application; combination strategies for simultaneous editing of multiple sites and their synergistic effects; and the application of multi-gene editing in the treatment of mitochondrial diseases. Edit window optimization strategy: Different edit window features and applications on H chain and L chain, target design strategy based on position-dependent editing efficiency, customized editing schemes for different chain sites, and expanding the technical approaches and application scenarios of edit windows.

[0106] The application system technology of this invention includes: crRNA library screening platform: a genome-wide mitochondrial crRNA library design method, a high-throughput crRNA library construction and transduction system based on lentiviruses, a long-distance amplicon sequencing method for evaluating mitochondrial editing efficiency, and a bioinformatics workflow for library data analysis and editing characteristic characterization; extended screening strategies for mtDNA deaminase types; and the ability to screen mitochondrial mtDNA functional target sites.

[0107] Mitochondrial editing optimization system: Parameter optimization methods for different cell types, system regulation strategies based on cell state, mitochondrial number, and transfection efficiency, optimization methods for balancing editing efficiency and cytotoxicity, and process flow for large-scale mitochondrial editing applications.

[0108] Specific disease applications of this invention include: Mitochondrial disease treatment strategies: design of editing protocols targeting specific pathogenic mtDNA mutations, threshold regulation strategies for heterogeneous mutation treatment, disease model construction methods based on the mitoCas12-BE3 system, and methods for evaluating the relationship between editing efficiency and clinical phenotypic improvement; Delivery system integration schemes: system optimization strategies compatible with viral vectors (AAV, lentivirus, etc.), adaptation methods for non-viral vector (liposomes, nanoparticles, etc.) delivery, synergistic optimization of tissue-specific delivery and editing systems, and research methods on the relationship between in vivo delivery efficiency and editing effect.

[0109] Potential applications of this invention include: Mitochondrial disease model construction: Using the mitoCas12-BE3 system to introduce specific mtDNA mutations, mitochondrial disease cell and animal models are constructed for mechanism research and drug screening; Treatment of hereditary mitochondrial diseases: Develop specific editing protocols for common mitochondrial disease-related mutations such as MELAS, MERRF, and LHON (e.g., LHON-related pathogenic SNVs induced in mice in Example 8), providing new strategies for gene therapy; Aging research: mtDNA mutations are closely related to cellular senescence. This system is used to study the mechanism of action of mtDNA mutations in the aging process and to develop intervention strategies. Cancer mechanism research: Mitochondrial dysfunction is associated with the occurrence and development of various cancers. This system is used to explore the relationship between mtDNA mutations and tumor metabolic reprogramming. Metabolic disease research: Mitochondria are the center of cellular energy metabolism. This system is used to study the association mechanism between mtDNA variation and metabolic diseases such as diabetes and obesity.

[0110] Further optimization directions of this invention include: Editing efficiency is further improved: explore novel dimerization or multimerization domains, optimize deaminase activity and substrate specificity, and develop mitochondrial-specific expression systems; Expanding editing types: Developing other base conversion systems such as A-to-G and G-to-A, integrating multiple types of editors to achieve complex editing, and developing mitochondrial-specific DNA cutting and repair systems; Delivery system optimization: Develop AAV vectors suitable for mitochondrial targeting, design non-viral vector systems for mitochondrial targeting, and develop tissue-specific delivery strategies; Clinical translational research: Conduct in vivo editing efficiency and safety assessments, establish a threshold analysis system for heterogeneous mutation editing, and develop personalized editing protocols for specific diseases.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A nuclease fusion protein, characterized in that, This includes a double-mutated APOBEC3A, Cas12 protein, and a uracil glycosylation enzyme inhibitor; the double-mutated APOBEC3A is subject to W104A and Y132D mutations; the Cas12 protein is Cas12m or a double-mutated LbCas12a; the double-mutated LbCas12a is subject to D832A and E925A mutations.

2. The nuclease fusion protein according to claim 1, characterized in that, The Cas12m is either GoCas12m or MmCas12m.

3. The nuclease fusion protein according to claim 1, characterized in that, It also includes at least one mitochondrial targeting sequence; the mitochondrial targeting sequence is cytochrome c oxidase subunit 8 and / or neurosporidian ATPase subunit 9.

4. The nuclease fusion protein according to claim 3, characterized in that, The mitochondrial targeting sequence is embedded at at least one site between the C-terminus and N-terminus of the nuclease fusion protein, the Cas12 protein, and the uracil glycosylation inhibitor.

5. The nuclease fusion protein according to claim 1, characterized in that, It also includes at least one dimerizing domain GCN4; the dimerizing domain GCN4 is embedded after at least one site of the N-terminus, C-terminus, or amino acids 1040, 1074, 1087, 1120, 1142, 1143, or 1158 of the double-mutated LbCas12a.

6. The nuclease fusion protein according to claim 1, characterized in that, It also includes mammal-specific introns; said mammal-specific introns are embedded in said double-mutant APOBEC3A; said mammal-specific introns contain typical 5' splice donor sites (GT) and 3' splice acceptor sites (AG).

7. The nuclease fusion protein according to claim 1, characterized in that, Its amino acid sequence is shown in any one of SEQ ID NO:2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 32, 46, 48.

8. A nucleic acid encoding a nuclease fusion protein, characterized in that, Its nucleotide sequence is shown in any one of SEQ ID NO:1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 31, 45, 47.

9. A mitochondrial DNA base editing system, characterized in that, Includes the nuclease fusion protein and crRNA as described in any one of claims 1-7.

10. The use of the nuclease fusion protein according to any one of claims 1-7, the nucleic acid according to claim 8, or the mitochondrial DNA base editing system according to claim 9 in any of the following fields: i. Gene editing; ii. To prepare anti-aging drugs; iii. Delivery of reagents into the mitochondria; v. To prepare medicines for the treatment and / or prevention of mitochondrial diseases; vi. To prepare drugs for the treatment and / or prevention of tumors; vii. To prepare medicines for the treatment and / or prevention of diabetes; viii. To prepare medicines for the treatment and / or prevention of obesity.

Images