Fusion protein and application method thereof

Abstract

The invention discloses a fusion protein, the fusion protein comprises an antibody binding structural domain and an RNA editing structural domain, the antibody binding structural domain is a pAG structural domain and comprises a functional structural domain of Protein A and Protein G, and the RNA editing structural domain comprises a rat APOBEC1 enzyme deamination structural domain and / or a human ADAR2 enzyme deamination structural domain. The fusion protein can be used for simultaneously identifying an RBP target and providing transcriptome information in the same sample, can be used for representing an RNA-protein interaction spectrum, has relatively high reliability and sensitivity, is suitable for researching RBP functions in tissue slices and single cells, and has a wide application prospect in research of functional mechanisms of RBP in development and disease processes.

Description

Technical Field

[0001] The present invention relates to the technical field of RNA binding protein (RBP) analysis, in particular to a sequencing method and application for simultaneously identifying RBP targets and providing transcriptome information. Background Art

[0002] Regulation of gene expression at the RNA level is the basis of many biological processes. It controls the life cycle of RNA molecules in a complex and precise manner. This type of post-transcriptional regulation is mainly caused by the interaction between some intracellular proteins and RNA in a complex way, assembling to form dynamic ribonucleoprotein particles (RNPs) to coordinate a series of processes such as cell fate and adapting to environmental changes. RNA-binding proteins (RBPs) can fully participate in different biological stages after RNA transcription, and form molecular machines with rRNA, tRNA, etc., playing a pivotal role in the gene expression regulatory network. However, the functions and molecular mechanisms of many RBPs are still blank, resulting in slow progress in the study of the role of RBP in development and disease. Today, more and more studies have shown that many human diseases, including neurological diseases, metabolic diseases and cancer, are related to RBP dysfunction. At present, the number of disease-related RBPs identified has exceeded 1,000, but with a few exceptions, people lack an understanding of the physiological regulatory system and pathological mechanism of RBP. When the disease phenotype cannot be explained by the well-known functions of RBP, finding out which targets it specifically recognizes and using RBP as the core to detect abnormal interactions caused by pathogenic mutations is crucial to understanding the disease process.

[0003] Currently, the cutting-edge technologies for studying the binding of RBP to RNA can be divided into two categories: the first is the immunoprecipitation method that relies on cross-linking and antibodies; the second is the in situ detection method that relies on RNA processing enzymes.

[0004] RNA immunoprecipitation (RIP) and cross-linking immunoprecipitation (CLIP) are the most basic techniques for studying RNA-protein interactions in cells. They use specific antibodies to precipitate RNA-binding proteins (RBPs) and their associated RNA targets. However, these methods have significant limitations, including time-consuming and labor-intensive experimental processes, non-specific interactions identified in low-complexity libraries, and the need for large amounts of starting materials. These challenges make them difficult to apply in certain scenarios, especially when dealing with low-input or rare samples, or when a large number of samples need to be processed in parallel. More importantly, these methods still face challenges in resolving targets at the transcript isoform level and achieving single-cell resolution. In addition, none of the above methods support parallel transcriptome analysis in the same cellular environment, limiting their ability to directly establish the relationship between RBP binding and gene expression regulation.

[0005] Recently, RNA editing-based TRIBE and STAMP have been developed, which use ectopically expressed RBP-deaminase fusion proteins to tag RNA substrates. RNA deamination is mainly mediated by two types of enzymes: adenosine deaminases acting on RNA (ADARs) and the apolipoprotein B mRNA editing enzyme cytosine deaminase family (also known as the AID / APOBEC family), which are responsible for A-to-I and C-to-U editing, respectively. The ADAR structure is modular and consists of double-stranded RNA binding motifs (dsRBMs) and a catalytic deamination domain (ADARcd), both of which can work independently. ADARcd deaminates adenosine on RNA to inosine, which is recognized as guanosine by reverse transcriptase, so AG mutations can be determined by RNA sequencing to identify ADAR editing. Endogenous ADAR editing sites are usually in Alu repeat elements in non-coding regions because they are biased to form double-stranded RNA features. APOBEC1 has a weak RNA binding ability and cannot edit its substrates in the absence of the chaperone APOBEC1 coupling factor (ACF). Its editing may rely on C-terminus-mediated homodimerization. AOPBEC1 editing in cells has structural and sequence preferences, almost all of which occur in 3′UTRs, preferring to bind to AU-rich regions and preferring to edit 5′-AC-3′ in sequence. These editing enzyme methods avoid the target enrichment step and retain complete transcriptome information. Although these IP-free technologies allow substrates to be discovered from low-input materials (including single cells), the use of a single enzyme may result in false negative results, and their reliance on genetic manipulation limits their application in primary cells and clinical samples. In addition, ectopically expressed fusion proteins may introduce artificial factors or interfere with the function of endogenous proteins. Finally, the lack of temporal resolution severely limits their ability to study dynamic biological processes.

[0006] Therefore, in order to fully understand the regulatory role of RBP function in different environments, the field urgently needs a technology that is sensitive, universal, and convenient, capable of dual-omics co-analysis, has high temporal resolution, requires only a small amount of sample input, and can be applied to tissue samples. Summary of the invention

[0007] In view of the deficiencies of the prior art, the present invention provides a fusion protein and an application method thereof, which can simultaneously identify RBP targets and provide transcriptome information, thereby solving the problems existing in the prior art.

[0008] In one aspect of the present invention, a fusion protein is provided, which includes an antibody binding domain and an RNA editing domain, wherein the antibody binding domain is a pAG domain, including functional domains of Protein A and Protein G, and the RNA editing domain includes a rat APOBEC1 enzyme deamination domain and / or a human ADAR2 enzyme deamination domain.

[0009] In a specific embodiment of the present invention, the fusion protein includes the pAG domain, and the deamination domain of rat APOBEC1 enzyme and the deamination domain of human ADAR2 enzyme.

[0010] In another aspect of the present invention, a method for simultaneously identifying RBP targets and providing transcriptome information using the fusion protein disclosed in the present invention is provided, the method comprising the following steps:

[0011] 1. Expression and purification of fusion protein: using a baculovirus expression vector system to express and purify the fusion protein of the present invention;

[0012] 2. Fixation of cell / tissue sections: Fix the cell / tissue samples with formaldehyde solution or 3,3'-dithiodipropionic acid di(N-hydroxysuccinimide ester) (DSP);

[0013] 3. Cell binding to ConA magnetic beads: resuspend the fixed cells with DPBS, add ConA magnetic beads, incubate with rotation at room temperature, and transfer the cells bound to the magnetic beads to the PCR tube;

[0014] 4. Antibody incubation: Prepare antibody incubation buffer, add RBP antibody or IgG antibody buffer to the experimental group and control group respectively, resuspend the cells, gently mix by vortexing, wash the cells and resuspend;

[0015] 5. Binding of deaminase: prepare a buffer containing the fusion protein of the present invention in advance, wash the cells incubated with the secondary antibody obtained in step 4 and resuspend them in a buffer containing the fusion protein;

[0016] 6. Initiation and termination of editing reaction: prepare a deamination reaction buffer containing zinc ions for initiating the deamination reaction, wash the cells co-incubated with the fusion protein of the present invention and resuspend them in the reaction buffer to perform the editing reaction;

[0017] 7. RNA extraction and sequencing library construction: extract total cell RNA, build a library, detect the library concentration, and sequence the cDNA in the library;

[0018] 8. Bioinformatics analysis: The bioinformatics analysis includes data quality control, two rounds of unique alignment, fine-tuning alignment, identification of single nucleotide variants (SNVs) and differential editing analysis.

[0019] In one embodiment of the present invention, 3,3'-dithiodipropionic acid bis(N-hydroxysuccinimide ester) (DSP) is used to fix the cells in step 2 above.

[0020] In one embodiment of the present invention, there is provided use of the fusion protein of the present invention for identifying RBP targets and providing transcriptome information, characterized in that it is applied to tissue samples, frozen sections or single cell samples.

[0021] Beneficial Effects

[0022] The present invention provides a sequencing method and application for simultaneously identifying RBP targets and providing transcriptome information.

[0023] Compared with the prior art, it has the following beneficial effects:

[0024] 1. The fusion protein and application method disclosed in the present invention can simultaneously identify RBP targets and provide transcriptome information. Using the technical solution disclosed in the present invention, the inventors successfully identified the RNA targets of RBPs such as YTHDF2 and G3BP1 in HeLa cells, detected the transcripts and binding motifs bound to PTBP1 and G3BP1 in HEK293T cells, and compared the endogenous protein results obtained by the technical solution of the present invention with the gold standard CLIP-seq. The results fully verified the practicality and sensitivity of the technical solution of the present invention.

[0025] 2. Since the fusion protein and its application method disclosed in the present invention can provide transcriptome information while realizing RBP target detection with single-cell resolution, the transcriptome information can be used to distinguish cells in different stages of the cell cycle, revealing the G3BP1 targets in different cell phases. The fusion protein and its application method disclosed in the present invention is a multifunctional dual-omics platform, which can characterize the RNA-protein interaction spectrum. Compared with other methods, it has many advantages, effectively reduces the material input, does not require the construction of a heterologous expression model, is suitable for studying the RBP function in tissue sections and single cells, and has broad application prospects in the study of the functional mechanism of RNA in development and disease processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0027] Figure 2 Schematic diagram of the cumulative number of MAPIT-seq editing signals in a 400-nucleotide window near the CLIP enrichment peak region in different treatment groups of the present invention.

[0028] Figure 3 A schematic diagram of editing signals for the experimental group and the control group on the G3BP1 binding target of the present invention.

[0029] Figure 4 This is a schematic diagram of the comparison results between the G3BP1 target identified by MAPIT-seq of the present invention and PAR-CLIP.

[0030] Figure 5 Schematic diagram of the overlap ratio of the G3BP1 target identified by MAPIT-seq of the present invention compared with CLIP-seq and its variants, RIP-seq, and HyperTRIBE.

[0031] Figure 6 Schematic diagram of specific editing events on PTBP1 and YTHDF2 target mRNAs identified by MAPIT-seq of the present invention.

[0032] Figure 7 Schematic diagram of the overlap between the PTBP1 and YTHDF2 targets identified by MAPIT-seq of the present invention and the targets detected by CLIP.

[0033] Figure 8 Schematic diagram of the high-quality transcriptome profile corresponding to the cellular environment studied by the RBPs of the present invention.

[0034] Fig. 9 It is a schematic diagram of the PTBP1 binding motif detected by MAPIT-seq of the present invention.

[0035] Fig.10Schematic diagram of the YTHDF2 binding motif detected by MAPIT-seq of the present invention.

[0036] Fig.11 Schematic diagram of editing events enriched by MAPIT-seq on fixed frozen sections and fresh frozen sections of the present invention.

[0037] Fig.12 This is a schematic diagram of the G3BP1 editing signal and gene expression repeatability of the MAPIT-seq of the present invention applied to mouse embryonic tissue sections.

[0038] Fig.13 This is a schematic diagram of gene expression for the transition from neurogenesis to gliogenesis between E12.5 and E16.5 of the present invention.

[0039] Fig.14 This is a schematic diagram comparing the enrichment editing events of G3BP1-MAPIT and IgG-MAPIT in brain tissue sections of mouse embryos at two stages using MAPIT-seq of the present invention.

[0040] Fig.15 Schematic diagram of specific editing events generated by MAPIT-seq of the present invention on the G3BP1 targets Mapt and Cadm2.

[0041] Fig.16 Schematic diagram of the content of G3BP1 binding motifs identified by CLIP in targets identified by applying MAPIT-seq of the present invention on mouse embryonic brain tissue sections.

[0042] Fig.17 This is a schematic diagram showing the correlation between the gene expression profile of the G3BP1-MAPIT library fixed by DSP in the present invention and the untreated group.

[0043] Fig.18 Schematic diagram of the signal-to-noise ratio of bulk MAPIT-seq generated by DSP-MAPIT of the present invention.

[0044] Fig.19 It is a schematic diagram comparing the results of the G3BP1 binding substrate of the present invention, formaldehyde-MAPIT and ARTR-seq.

[0045] Fig. 20 This is a schematic diagram of the correlation between the gene expression profile of the scMAPIT-seq library fixed with DSP in the present invention and the untreated group.

[0046] Fig.21 Schematic diagram of the enrichment of dual editing events on G3BP1 targets in single-cell and single-cell aggregated G3BP1-MAPIT datasets of the present invention.

[0047] Fig. 22 This is a schematic diagram showing the comparison of the results of the scMAPIT-seq of the present invention, which identified the G3BP1 target after aggregation of all single cells, with those of bulk MAPIT-seq and ARTR-seq.

[0048] Fig.23 This is a schematic diagram of scMAPIT-seq of the present invention using gene expression to group cells into three cell cycle phases.

[0049] Fig.24 Schematic diagram showing the differences of the G3BP1 target in three stages (G1, S and G2 / M) of the present invention.

[0050] Fig.25 This is a schematic diagram of the correlation distribution between G3BP1 binding strength and RNA expression level during the cell cycle process detected by scMAPIT-seq of the present invention.

[0051] Fig.26 It is a schematic diagram comparing the sequencing read lengths of DSP-G3BP1-MAPIT and FA-G3BP1-MAPIT samples of the present invention.

[0052] Fig. 27 It is a schematic diagram comparing the editing signal intensities of DSP-G3BP1-MAPIT and FA-G3BP1-MAPIT samples of the present invention.

[0053] Fig.28 Schematic diagram comparing all targets of the long-read MAPIT-seq data of the present invention with those of the short-read MAPIT-seq data.

[0054] Fig.29 Schematic diagram of the comparison of editing signal intensities of the long-read MAPIT-seq data of the present invention on two transcript isoforms of YTHDF2, GDAP2 and ING3. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] The present invention provides a fusion protein, which includes an antibody binding domain and an RNA editing domain, wherein the antibody binding domain is a pAG domain, including functional domains of Protein A and Protein G, and the RNA editing domain includes a rat APOBEC1 enzyme deamination domain and / or a human ADAR2 enzyme deamination domain. On the basis of the above-mentioned fusion protein, the present invention provides a technical solution based on RNA editing enzyme and in situ capture of intracellular RBP targets. Its core principle is different from methods such as TRIBE, and does not require genetic manipulation, but relies on antibodies to guide recombinant editing enzymes (i.e., the fusion protein of the present invention) to target RBP binding positions for in situ deamination reactions. Therefore, this technology can not only realize the joint detection of RBP targets and transcriptomes, but also can draw dynamic RBP interactions.

[0057] The present invention designs key experimental processes based on the characteristics of RBP: 1) Use 0.5% formaldehyde to fix the sample. Since the binding of RBP to RNA is transient and dynamic, efficient and reversible cross-linking is a key step to capture RBP interactions and preserve high-quality transcriptomes in subsequent operations. 2) One of the cores of the present invention lies in the construction of RNA editing enzyme components, including antibody binding domains and RNA editing domains. For the antibody binding domain, the prior art generally uses Protein A, which binds well to the constant regions of rabbit, guinea pig and other antibodies, but binds poorly to the mouse-derived IgG1 subtype, so Protein G with higher affinity to mouse antibodies is required. In order to improve the universality of the fusion protein, the present invention adopts pAG-ERH constructed by Henikoff et al. (see Meers MP, Bryson TD, Henikoff JG, Henikoff S. Improved CUT&RUN chromatinprofiling tools. Elife. 2019Jun 24; 8: e46314. doi: 10.7554 / eLife.46314. PMID: 31232687; PMCID: PMC6598765. The full text of this document is introduced into this application), that is, a Protein G domain containing three point mutations is added to Protein A, which is sufficient to ensure affinity with common host antibodies. Secondly, as described in the background technology, TRIBE and STAMP technologies are limited by the efficiency of a single editing enzyme, and APOBEC and ADAR both have strong substrate and sequence preferences. The present invention fuses rat APOBEC1 (rAPOBEC1) and the deamination domain (hADAR2dd) of the human ADAR2 (E488Q) enzyme activity enhanced mutant with pAG through an XTEN linker, and the resulting double editing events can be combined to reduce false negative detection and improve the sensitivity of the technology. After testing vectors with different combinations, the inventors determined that rAPOBEC1-pAG-hADAR2 is the best performing combination. In order to ensure the biological activity and yield of the fusion protein, the inventors expressed rAPOBEC1-pAG-hADAR2 in eukaryotic insect cells sf21 and purified it through Strep magnetic beads and molecular sieves.

[0058] Based on the above ideas, the inventors developed MAPIT-seq technology (Modification Added to RBP Interacting Transcript-sequencing, adding modification and sequencing technology to RBP interacting transcripts). The specific principle is as follows Figure 1As shown, formaldehyde was used to fix the intracellular molecules in situ, and covalently cross-linked the RNA bound to RBP. Digitonin was added to permeabilize the cell membrane, and antibodies (RBP corresponding antibodies as experimental groups, IgG as control groups to remove non-specific editing) and fusion proteins rAPOBEC1-pAG-hADAR2dd were incubated in sequence at 4°C. After entering the cell, the fusion protein was guided by the pAG domain in the fusion protein to target the RBP antibody. The excess antibodies and enzymes were washed away in the corresponding steps, and then the fusion protein was deaminized in the RNA region bound to RBP at room temperature. Finally, the reaction was terminated and proteinase K was added to dissolve the cross-linking, RNA was extracted to build a library, and sequencing analysis was performed to identify the mutation site to identify the RBP substrate.

[0059] Example 1 MAPIT-seq experimental steps

[0060] 1 Expression and purification of fusion protein

[0061] The fusion protein rAPOBEC1-pAG-hADAR2dd used in the present invention is expressed and purified using a baculovirus expression vector system, with real protein modification to obtain maximum enzyme activity. General process: The recombinant pFastBac plasmid is transformed into DH10Bac cells (Biomed, BC112) to generate bacmids. The bacmid DNA is transfected into Sf21 cells using X-tremeGENE HP DNA transfection reagent (06366236001, Roche) and cultured for 4 days. Low-titer recombinant baculovirus is collected and amplified in Sf21 cells to produce high-titer virus. Next, the baculovirus infects Hi5 cells at a ratio of 1:200 and is cultured for 48-60 hours. The cells are lysed in a lysis buffer containing 50mM Tris-HCl (pH 8.0) and 200mM NaCl. The lysate was sonicated at 4°C using an ultrasonic processor (SCIENTZ), set to a cycle of 7 seconds on and 5 seconds off, at a power of 300W for 20 minutes. After sonication, the lysate was centrifuged at 16,000g, 4°C for 40 minutes. The deaminase fusion protein was then purified using Strep-Tactin beads (SA053, Smart Lifesciences). In brief, the supernatant was incubated with Strep-Tactin beads at 4°C for 1 hour. After washing the beads with 30 volumes of lysis buffer, the deaminase fusion protein was eluted with a lysis buffer containing 5mM desthiobiotin. Further purification was performed using a Superdex 200Increase column (GE Healthcare). The deaminase protein was eluted into a buffer containing 25mM HEPES (pH 7.5) and 150mM NaCl, and then supplemented with 5% glycerol. The deaminase protein was rapidly frozen in liquid nitrogen and stored at -80°C.

[0062] 2. Fixation of cells / tissue sections

[0063] For each experiment, 5 million to 500,000 cells were collected, digested with 0.1% Trypsin for 5 minutes, and collected into a centrifuge tube rinsed with 1% BSA in advance after the culture medium was terminated. The following description is based on a group of cells or tissue sections; centrifuged at room temperature at 250×g for 5 minutes, washed once with DPBS; prepared 0.2% (tissue) or 0.5% (cell) formaldehyde solution; added formaldehyde solution to each group of cells, incubated at room temperature for 5 minutes; added Glycine or Tris-HCl pH 7.4 to terminate crosslinking, incubated at room temperature for 5 minutes; centrifuged in a non-fixed centrifuge at 4°C, 450×g, 5 minutes. Washed cells 3 times with pre-cooled DPBS.

[0064] 3. Cell Binding to ConA Magnetic Beads

[0065] Pipette 5 μL ConA beads and add 10× binding buffer, mix well, place on the magnetic stand for 30s-2min, and remove the supernatant; repeat once and resuspend with the initial volume of binding buffer; resuspend the fixed cells with DPBS, add ConA magnetic beads, and incubate at room temperature for 10 to 20min; transfer the cells bound to the magnetic beads to the PCR eight-tube strip;

[0066] 4. Antibody Incubation

[0067] Prepare antibody incubation buffer (take two groups as an example, fresh preparation is required): (1mM PMSF, 1×proteaseinhibitor cocktail, 1U / μL RiboLock RNase Inhibitor (EO0382, ThermoFisher), 0.01% Digitonin, 1mM DTT, 2mM EDTA, and 0.1% BSA, supplemented to 100μL with DPBS); add RBP antibody or IgG antibody buffer (100:1) to the experimental group and the control group, resuspend the cells, and gently mix by vortexing; incubate at 4℃ for 3-4hr. Mix halfway; wash the cells once with DPBS, resuspend with secondary antibody incubation buffer (antibody buffer-EDTA), and incubate at 4℃ for 1hr.

[0068] 5. Binding of Deaminase

[0069] Prepare enzyme buffer in advance: 20mM HEPES-KOH pH 7.5 + 300mM NaCl to replace DPBS in the secondary antibody incubation buffer. Add 1μg of fusion protein rAPOBEC1-pAG-hADAR2dd to 50μL buffer. Wash the cells after secondary antibody incubation once with DPBS; resuspend the cells in enzyme buffer and incubate at 4℃ for 1hr.

[0070] 6 Editing reaction initiation and termination

[0071] Prepare deamination reaction buffer: (15mM HEPES pH 7.9, 60mM KCl, 15mM NaCl, 5% Glycerol, 0.5mM DTT, and 0.1μM ZnCl2), add Proteinase Inhibitor, RNAase Inhibitor and DTT freshly before use; wash the cells after incubation with deaminase three times with 200μL washing buffer (20mM HEPES-KOH pH7.5+300mM NaCl, 0.005% Digitonin); resuspend the cells in 40μL reaction buffer and react at 30℃ for 3hr;

[0072] 7RNA extraction and sequencing library construction

[0073] Add 60 μL proteinase K digestion buffer and 1 μL proteinase K, digest at 56°C for 1 hour. Use Magzol RNA extraction reagent to extract total RNA, and add glycogen to the trace cells to help precipitation. Take 500 ng RNA for library construction. Universal V6 RNA-seq Library Prep Kit for Kit, after Qubit detection of library concentration, 100 ng cDNA was sent to Illumina NovaSeq 6000 platform for sequencing.

[0074] 8 Bioinformatics analysis

[0075] After sequencing data quality control, TrimGalore was used to remove adapter sequences and low-quality bases. Before MAPIT-seq alignment, BWA was used to remove reads of high-abundance RNA in cells, including ribosomal RNA and transfer RNA. The data alignment reads used the "two-round unique alignment" strategy of HISAT2 and BWA, and then the GATK tool was used to correct the alignment file. HaplotypeCaller was used to identify SNV sites and remove known SNP sites. Finally, A-to-G and C-to-U editing sites were screened. Bedtools was used to annotate genes at the sites, such as transcript elements and repeat sequences. Differential editing analysis divides the transcripts into several small blocks according to a given window size (default 50bp), and calculates the editing index of each window, which is defined as the cumulative value of the editing rate of the editing sites in the specified region. Finally, the Wilcoxon signed rank test was used for the editing index of each transcript in the experimental group and the control group to screen out transcripts with significantly higher editing in the experimental group than in the control group (MAPITscore greater than 0.5, P-value less than 0.001 and editing multiples>2), which were RNA targets targeted by RBP, where MAPITscore was defined as the editing index within a transcript range in the experimental group minus the value of the control group. For Motif analysis, Poisson was used to screen out high-confidence editing sites, and editing sites within a neighbor distance of 150nt were clustered to obtain editing clusters. Editing clusters containing at least 3 sites and with a higher editing rate in the experimental group than in the control group were retained as high-confidence RNA editing segments, and RBP binding motifs were predicted using homer. For single-cell MAPIT-seq, the STARsolo module was used for alignment. After read segment deduplication and cell quality control, the alignment results were split by single cells, and RNA editing identification analysis was performed on each single cell. For long-read MAPIT-seq, Minimap2 was used for alignment. After quantification and quality control using isoquant, the alignment results were split according to transcript isoforms, and RNA editing identification analysis was performed on each transcript isoform. After sequencing data quality control, TrimGalore was used to remove adapter sequences and low-quality bases. For MAPIT-seq data, BWA was used to remove high-abundance RNAs in cells before alignment, including ribosomal RNA, mitochondrial RNA, and transfer RNA.

[0076] The MAPIT-seq method for tissue sections is a modified version of the standard MAPIT-seq. A brief description is as follows:

[0077] Thaw sections stored at -80 °C at room temperature for 10 min.

[0078] Use a PAP pen to draw a hydrophobic circle around the tissue.

[0079] The sections were fixed with 0.2% formaldehyde.

[0080] The cell permeabilization and deamination procedures were basically the same as standard MAPIT-seq.

[0081] After the deamination reaction, the samples were digested with 0.4 mg / mL proteinase K at 55°C for 2 h.

[0082] Finally, the mixture was transferred from the slide to a 1.5 mL tube using a pipette for subsequent RNA extraction, library construction, and sequencing.

[0083] MAPIT-seq for high-throughput single-cells was performed using DSP immobilization combined with the 10x Genomics platform:

[0084] The cells were fixed with 1 mg / mL DSP solution at room temperature for 30 min.

[0085] Throughout the experiments, DSP-fixed cells were collected by centrifugation at 650 g for 5 min.

[0086] After the deamination reaction, 50 mM DTT was added for reverse cross-linking.

[0087] Cells were washed once and then resuspended in DPBS containing 0.05% BSA and 1 U / μL RiboLock RNase inhibitor.

[0088] Subsequently, samples were sorted on a MoFlo Astrios EQ (Beckman) to eliminate cell debris and aggregates.

[0089] Finally, the samples were loaded onto the 10x Genomics Chromium platform.

[0090] Example 2 MAPIT-seq achieves specific editing of RBP targets

[0091] In order to preliminarily verify the feasibility of the system, the inventors first selected the RNA binding protein YTHDF2 with clear functions and CLIP data sets. 6A plays a key role in regulation. And taking into account the efficiency of antibodies, the present invention constructed a cell line overexpressing RBP-Flag, and collected 250,000 cells for MAPIT-seq experiments. HeLa cells overexpressing YTHDF2-Flag were used as the experimental group for targeted editing using Flag antibodies. IgG / no antibody treatment was used as a simple control group, while Flag antibody treatment of cells expressing only Flag was used as a more reliable non-specific editing control. First, the inventors counted the cumulative number of MAPIT-seq editing signals in a 400 nucleotide window near the CLIP enrichment peak area. Within 400bp around the YTHDF2 binding site, the overall number of edits and efficiency of the YTHDF2-Flag experimental group were significantly increased compared with the three control groups ( Figure 2 ). Notably, the increase in editing events required both YTHDF2 binding to RNA and the corresponding antibody to recruit the recombinase for editing, demonstrating the specificity of MAPIT technology for RBP.

[0092] Example 3 MAPIT-seq combined detection of endogenous RBP binding motifs and gene expression

[0093] After verifying the feasibility of the system and optimizing the experimental conditions, the inventors applied MAPIT-seq to the endogenous protein G3BP1 in HEK293T cells. The inventors first checked the G3BP1 target gene previously reported by PAR-CLIP. The IGV (Integrative Genomic Viewer) results showed that the experimental group produced abundant dual editing signals on the target, which had a similar distribution to the G3BP1 binding sites detected by PAR-CLIP, while the control group had almost no editing in the binding area ( Figure 3 ). These results indicate that MAPIT-seq can enrich for target-specific editing, indicating RBP binding substrates. To evaluate the effectiveness of MAPIT-seq in identifying G3BP1 interacting RNAs across the entire transcriptome, the inventors applied a set of stringent criteria to define G3BP1 targets. In short, if a transcript had an editing fold enrichment >2, a MAPIT score >0.5, and a P value <0.001, the inventors considered it a G3BP1 target. To minimize false negatives, the inventors combined A to I and C to U editing to calculate the MAPIT score for each gene. The results showed that the G3BP1 targets identified by MAPIT-seq overlapped significantly with the PAR-CLIP results ( Figure 4 ), and most of the G3BP1 targets identified by MAPIT-seq were cross-validated by multiple independent methods, including CLIP-seq and its variants, RIP-seq, HyperTRIBE, and TRIBE-ID ( Figure 5), the above results can reflect the accuracy and reliability of MAPIT-seq in detecting RBP target genes.

[0094] To further explore the versatility of MAPIT-seq, the inventors also detected PTBP1 protein RNA substrates in HEK293T cells and YTHDF2 protein RNA substrates in HeLa cells. Similar to the findings for G3BP1, MAPIT-seq also revealed specific editing events on these RBP target mRNAs ( Figure 6 ). In addition, the targets identified by MAPIT-seq all overlap significantly with those detected by CLIP-based methods ( Figure 7 ), highlighting its reliability and broad applicability across different RBPs. Furthermore, MAPIT-seq provides high-quality transcriptome profiles corresponding to the cellular contexts in which these RBPs were studied ( Figure 8 ).

[0095] Finally, MAPIT-seq can also be used to detect the binding motifs of RBPs. The inventors focused on PTBP1, an RBP known to prefer CU-enriched sequences. Using the PTBP1-MAPIT edited cluster, the inventors identified a significant enrichment of the "CUUCUUUC" motif using the HOMER suite ( Fig. 9 ). Similarly, the inventors found that YTHDF2 was enriched with the conserved "GGAC" motif ( Fig.10 Collectively, these data demonstrate the effectiveness of MAPIT-seq in detecting RBP-binding targets and revealing their associated sequence motifs.

[0096] Example 4 MAPIT-seq detection of RBP targets in tissue sections

[0097] On the basis of fully verifying the advancedness and reliability of MAPIT-seq technology, the inventors further explored whether MAPIT-seq can be applied in tissues to provide RBP binding groups and gene expression information in tissue environments. Therefore, the inventors tested the MAPIT-seq method in mouse tissues. The inventors first tested the MAPIT-seq experimental conditions of G3BP1 on fixed and fresh frozen mouse embryo sections at embryonic day 10.5 (E10.5). The results showed that fresh frozen sections, but not fixed frozen sections, showed enriched editing events ( Fig.11 ), and the editing signals and gene expression of MAPIT-seq are highly reproducible in brain tissue ( Fig.12). Based on this, the inventors decided to use fresh frozen sections in subsequent experiments. To further investigate the function of G3BP1 in brain development, the inventors performed MAPIT-seq on fresh frozen sections of E12.5 and E16.5 mouse embryonic brains. Gene expression confirmed the transition from neurogenesis to gliogenesis between E12.5 and E16.5 ( Fig.13 The inventors found that G3BP1-MAPIT enriched more editing events than IgG-MAPIT in mouse brain slices at all stages ( Fig.14 G3BP1-MAPIT discovered specific editing events on the G3BP1 targets Mapt and Cadm2 previously identified by CLIP ( Fig.15 Overall, the G3BP1 targets identified by the inventors in E12.5 and E16.5 mouse brain slices were significantly enriched for the G3BP1 binding motifs identified by CLIP ( Fig.16 ), confirming the feasibility and specificity of MAPIT-seq in detecting RBP binding in tissue sections. GO analysis of subsequent targets can be performed to explore the regulatory function of G3BP1.

[0098] Example 5 MAPIT-seq detects RBP targets in single cells at a high throughput level

[0099] The inventors first explored the possibility of MAPIT-seq to reliably detect in situ RBP-RNA interactomes and transcriptomes at the single-cell level. The inventors first considered that detecting mutations in formaldehyde-fixed samples requires a decrosslinking step, which is incompatible with most high-throughput single-cell RNA-seq (scRNA-seq) platforms. To address this problem, the inventors adopted an alternative fixative that does not form covalent bonds and is compatible with scRNA-seq: 3,3'-dithiodipropionic acid di(N-hydroxysuccinimide ester) (DSP), a new crosslinker that can be easily reversed by dithiothreitol (DTT). The inventors first evaluated whether DSP can simultaneously capture protein-RNA interactions and gene expression profiles in bulk MAPIT-seq. The results showed that the gene expression profile of the G3BP1-MAPIT library fixed with DSP was highly correlated with that of untreated cells and was very similar to the expression profile of formaldehyde-fixed samples ( Fig.17 ). Notably, DSP-MAPIT yielded a significantly higher signal-to-noise ratio compared to formaldehyde-MAPIT, with a clear enhancement around the PAR-CLIP peak ( Fig.18 The G3BP1 binding substrates obtained by DSP-MAPIT were strongly correlated with the results of formaldehyde-MAPIT and ARTR ( Fig.19 ).

[0100] We next performed scMAPIT-seq using DSP fixation and confirmed that it successfully maintained transcriptome integrity in single cells without compromising RNA quality ( Fig. 20 ). These findings indicate that DSP is an effective fixative for MAPIT-seq and scMAPIT-seq, making it suitable for scaling up scMAPIT-seq to high-throughput applications. The inventors sequenced MAPIT-seq cells using the 10xGenomics single-cell workflow, capturing 3,400 G3BP1-MAPIT cells and 3,945 IgG-MAPIT cells. The results showed that dual editing events were significantly enriched on the G3BP1 target in both single-cell and single-cell aggregated G3BP1-MAPIT datasets compared to IgG-MAPIT ( Fig.21 ).

[0101] In addition, the inventors aggregated the signals of all individual cells and found that scMAPIT-seq maintained comparable sensitivity and accuracy to bulkMAPIT-seq and showed strong consistency with ARTR-seq ( Fig. 22 ). These data support that the inventors' high-throughput scMAPIT-seq can effectively reveal RBP targets and gene expression simultaneously at single-cell resolution. scMAPIT-seq maintains comparable sensitivity and accuracy to bulk MAPIT-seq.

[0102] Using the dual-omics capability of scMAPIT-seq, the present inventors further identified a library of cell cycle stage-specific G3BP1 targets ( Fig.23 The present inventors observed significant differences in G3BP1 targets among the three phases (G1, S and G2 / M) ( Fig.24 ). In addition, during the cell cycle, there is a positive and negative correlation between G3BP1 binding intensity and RNA expression level, with the positive correlation being more common ( Fig.25 Together, these data demonstrate the bi-omics power of scMAPIT-seq to reveal stage- and subpopulation-specific RBP interactions and highlight a specific role for G3BP1 during the cell cycle.

[0103] Example 6 Long-read MAPIT-seq reveals differential binding of RBP to RNA isoforms

[0104] The inventors further improved the resolution of MAPIT-seq at the molecular level to identify RBP binding events specific to individual transcript isoforms. The inventors input G3BP1-MAPIT samples of formaldehyde-fixed and DSP-fixed HEK293T into the PacBio HiFi sequencing platform. Compared with the formaldehyde-fixed samples, the DSP-G3BP1-MAPIT samples produced longer reads and higher editing signals ( Fig.26 and 27 ), indicating that DSP fixation enhances the target detection efficiency in long read sequencing. Based on these results, the inventors used DSP-MAPIT data for further analysis.

[0105] When comparing long-read MAPIT-seq data with short-read MAPIT-seq data, the inventors observed strong consistency in target gene identification between the two datasets ( Fig.28 ). This demonstrates the accuracy of long-read MAPIT-seq. Further analysis of editing events revealed that G3BP1 binding is isoform-specific for some genes, including YTHDF2, GDAP2, and ING3 ( Fig.29 ).

[0106] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0107] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fusion protein, comprising an antibody binding domain and an RNA editing domain, wherein the antibody binding domain is a pAG domain, comprising functional domains of Protein A and Protein G, and the RNA editing domain comprises a rat APOBEC1 enzyme deamination domain and / or a human ADAR2 enzyme deamination domain. 2 . The fusion protein according to claim 1 , comprising the pAG domain, and a rat APOBEC1 enzyme deamination domain and a human ADAR2 enzyme deamination domain.

3. A method for simultaneously identifying RBP targets and providing transcriptome information using the fusion protein of claim 1, the method comprising the following steps: 1). Expression and purification of fusion protein: using a baculovirus expression vector system to express and purify the fusion protein of claim 1; 2) Fixation of cell / tissue sections: Fix the cell / tissue samples with formaldehyde solution or 3,3'-dithiodipropionic acid di(N-hydroxysuccinimide ester); 3) Cell binding to ConA magnetic beads: resuspend the fixed cells with DPBS, add ConA magnetic beads, incubate with rotation at room temperature, and transfer the cells bound to the magnetic beads to the PCR tube; 4) Antibody incubation: Prepare antibody incubation buffer, add RBP antibody or IgG antibody buffer to the experimental group and control group respectively, resuspend the cells, and gently mix by vortexing. Wash the cells and resuspend them; 5) Binding of deaminase: prepare a buffer containing the fusion protein of the present invention in advance, wash the cells incubated with the secondary antibody obtained in step 4 and resuspend them in a buffer containing the fusion protein; 6) Initiation and termination of editing reaction: prepare a deamination reaction buffer containing zinc ions for initiating the deamination reaction, wash the cells co-incubated with the fusion protein of the present invention and resuspend them in the reaction buffer to perform the editing reaction; 7). RNA extraction and sequencing library construction: extract total cell RNA, build a library, detect the library concentration, and sequence the cDNA in the library; 8). Bioinformatics analysis: The bioinformatics analysis includes data quality control, two rounds of unique alignment, fine-tuning alignment, identification of single nucleotide variations (SNVs) and differential editing analysis.

4. The method according to claim 3, characterized in that In step 2), 3,3'-dithiodipropionic acid di(N-hydroxysuccinimide ester) was used to fix the cells.

5. Use of the fusion protein according to claim 1 for identifying RBP targets and providing transcriptome information, characterized in that: It is applied to tissue samples, frozen sections or single cell samples.

Images