Targeted RNA editing by adenosine deaminase-antisense oligonucleotide conjugates

Abstract

Disclosed is a method of targeted RNA editing comprising providing a protein- antisense oligonucleotide (ASO) conjugate comprising a protein covalently linked to an ASO, and delivering the protein-ASO conjugate to one or more cells to edit a target mRNA sequence. Providing the protein-ASO conjugate may comprise synthesizing a modified ASO comprising a reactive group, producing a recombinant fusion protein comprising the protein fused to a self- labeling protein tag, and forming the protein-ASO conjugate by incubating the modified ASO with the recombinant fusion protein. The self-labeling protein tag may include a modified haloalkane dehalogenase enzyme or a modified mammalian O6-alkylguanine-DNA- alkyltransferase (ACT) enzyme. The first protein may be a deaminase enzyme. The disclosure also provides bulge-forming antisense oligonucleotides (Bulge ASOs) comprising discontinuous binding regions configured to position a target mRNA base within a ssRNA loop between 4-9 nt in length, and a protein tag coupled to the nucleotide sequence at an intermediate location.

Description

[0001] Princeton - 104176

[0002] TARGETED RNA EDITING BY ADENOSINE DEAMINASE- ANTISENSE

[0003] OLIGONUCLEOTIDE CONJUGATES

[0004] CROSS-REFERENCE TO RELATED APPLICATIONS

[0005] This application claims priority to U.S. Provisional Application Nos. 63 / 705,196, titled "Targeted RNA editing by direct delivery of an adenosine deaminase-antisense oligo conjugate" and filed October 9, 2024, and 63 / 836,838, titled "Targeted RNA editing by direct delivery of an adenosine deaminase-antisense oligo conjugate", filed July 1, 2025, which are hereby incorporated by reference in their entirety.

[0006] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0007] This invention was made with government support under Grant No. GM 152748 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0008] FIELD OF INVENTION

[0009] The present disclosure relates to RNA editing technologies, and more particularly to targeted RNA editing by direct delivery of adenosine deaminase-antisense oligonucleotide conjugates for correcting disease-causing mutations at the RNA level.

[0010] BACKGROUND

[0011] RNA editing represents a post-transcriptional mechanism that can modify genetic information at the RNA level without altering the underlying DNA sequence. This process occurs naturally in cells through the action of various RNA-modifying enzymes, including adenosine deaminases acting on RNA (ADAR) and apolipoprotein B mRNA editing enzyme catalytic subunit (APOBEC) family members. These enzymes can convert adenosine to inosine or cytidine to uridine, respectively, thereby changing the coding sequence of mRNA transcripts and potentially altering protein function.

[0012] The therapeutic potential of programmable RNA editing has gained considerable attention as an alternative to genome editing approaches. Unlike DNA editing, which results in permanent genetic modifications, RNA editing offers transient effects that can be controlled through dosing and timing. This reversibility provides advantages in therapeutic applications Princeton - 104176 where temporary' correction of disease-causing mutations may be sufficient or where permanent genetic changes are undesirable due to safety’ concerns.

[0013] Current approaches to targeted RNA editing face several technical challenges. Many existing methods rely on the recruitment of endogenous cellular editing enzymes, such as AD ARI and ADAR2, using antisense oligonucleotides or guide RNAs. However, these natural enzymes have limited substrate specificity and often require complex guide RNA designs with specific secondary structures to achieve efficient editing. The substrate preferences of ADAR enzymes, for example, are restricted to certain sequence contexts, which limits their applicability7across different target sites.

[0014] Alternative strategies have employed CRISPR-based systems, such as catalytically inactive Cast 3 proteins fused to editing enzymes, to direct RNA modification activities to specific target sequences. While these approaches offer programmability through guide RNA sequences, they present delivery challenges due to the large size of the protein-RNA complexes and the need for co-delivery of multiple components.

[0015] The field has also explored the use of evolved base editing enzymes derived from bacterial systems, such as variants of the E. coli tRNA adenosine deaminase TadA. These engineered enzymes often exhibit enhanced activity7and broader substrate tolerance compared to their natural counterparts. However, directing these evolved enzymes to specific RNA targets in mammalian cells typically requires complex delivery systems or genetic engineering approaches.

[0016] Delivery of RNA editing components to target cells remains a significant hurdle in translating these technologies to therapeutic applications. Traditional methods often require genetic modification of target cells or complex formulations to deliver large protein-RNA complexes. The development of more direct and efficient delivery methods could expand the applicability of RNA editing technologies.

[0017] The ability7to achieve site-specific RNA editing with high efficiency while maintaining cellular viability represents an ongoing challenge in the field. Balancing editing specificity', minimizing off-target effects, and ensuring adequate delivery to target tissues are considerations that influence the development of practical RNA editing therapeutics.

[0018] SUMMARY

[0019] This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify Princeton - 104176 key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

[0020] According to an aspect of the present disclosure, a method of targeted RNA editing is provided. The method comprises providing a protein-antisense oligonucleotide (ASO) conjugate comprising a first protein covalently linked to an ASO. The method further comprises delivering the protein-ASO conjugate to one or more cells to edit a target mRNA sequence. This approach enables direct delivery of RNA editing components without requiring genetic modification of target cells or complex multi-component delivery systems. The covalent linkage between the protein and ASO ensures co-localization of the editing enzyme with its target sequence, enhancing editing specificity and efficiency.

[0021] According to other aspects of the present disclosure, the method may include one or more of the following features. Providing the protein-ASO conjugate may comprise synthesizing a modified ASO comprising a reactive group. The method may comprise producing a recombinant fusion protein comprising the first protein fused to a self-labeling protein tag. The method may comprise forming the protein-ASO conjugate by incubating the modified ASO with the recombinant fusion protein. The use of self-labeling protein tags facilitates efficient and specific conjugation between the protein and ASO components, enabling scalable production of homogeneous conjugates with defined stoichiometry'.

[0022] According to other aspects of the present disclosure, synthesizing the modified ASO may include solid phase oligonucleotide synthesis. Solid phase synthesis allows for precise control over ASO sequence and chemical modifications, enabling optimization of target binding affinity' and cellular stability7.

[0023] According to other aspects of the present disclosure, producing the recombinant fusion protein may include using heterologous expression in a host cell. Heterologous expression systems enable cost-effective production of recombinant proteins with consistent quality and activity7for therapeutic applications.

[0024] According to other aspects of the present disclosure, the host cell may be E. coli. E. coll expression systems provide rapid and economical protein production with well-established purification protocols. According to other aspects of the present disclosure, the host cell may be an organism other than E. coli. Alternative expression systems may provide enhanced protein folding, post-translational modifications, or improved solubility' for certain protein constructs.

[0025] According to other aspects of the present disclosure, the self-labeling protein tag may include a modified haloalkane dehalogenase enzyme or a modified mammalian O6- Princeton - 104176 alkylguanine-DNA-alkyltransferase (AGT) enzyme. These self-labeling systems enable rapid and irreversible conjugation under mild conditions, preserving protein activity and ASO integrity.

[0026] According to other aspects of the present disclosure, the method may further comprise allowing the recombinant fusion protein to recode at least some genetic information present in mRNA. RNA recoding enables correction of disease-causing mutations at the post- transcriptional level without permanent genomic alterations.

[0027] According to other aspects of the present disclosure, the method may further comprise allowing the recombinant fusion protein to regulate a disease-associated mRNA transcript. Targeted regulation of disease-associated transcripts provides therapeutic opportunities for conditions where protein function can be restored through RNA-level modifications.

[0028] According to other aspects of the present disclosure, the first protein may be a deaminase enzyme. Deaminase enzymes enable base conversion reactions that can correct point mutations or introduce stop codon readthrough.

[0029] According to other aspects of the present disclosure, the protein-ASO conjugate may be a protein-bulge-forming antisense oligonucleotide (ASO) (Bulge ASO) conjugate. Bulgeforming ASOs create specific structural contexts that enhance editing enzyme selectivity and reduce off-target modifications.

[0030] According to other aspects of the present disclosure, providing the protein-ASO conjugate may include protein trans-splicing utilizing split inteins. Split intein systems enable post-translational assembly of protein-ASO conjugates, providing flexibility in conjugate design and production.

[0031] According to other aspects of the present disclosure, providing the protein-ASO conjugate may include chemically conjugating the first protein to the ASO utilizing compatible reactive groups present on the first protein and the ASO. Chemical conjugation methods provide versatility in linking diverse protein and ASO components with controlled stoichiometry and orientation.

[0032] According to other aspects of the present disclosure, the ASO may have been modified to include a first compatible reactive group, and / or the first protein may have been modified through site-directed mutagenesis, protein semi-synthesis, or unnatural amino acid mutagenesis. These modification strategies enable introduction of specific reactive functionalities at defined positions, optimizing conjugation efficiency and maintaining biological activity. Princeton - 104176

[0033] According to other aspects of the present disclosure, the first compatible reactive group may be an azide, an alkyne, a tetrazine, an alkene, an amine, a thiol, a hydrazide, an aminooxy, an aldehyde, a ketone, an isothiocyanate, a carboxylate, a maleimide, an NHS ester, or a Gly- Gly-Gly peptide. These diverse reactive groups provide multiple conjugation strategies with varying reaction conditions and selectivities to accommodate different protein and ASO requirements.

[0034] According to other aspects of the present disclosure, the first compatible reactive group may be a functional group for bio-conjugation or bio-orthogonal chemistry. Bio-orthogonal chemistry enables selective conjugation reactions in complex biological environments without interfering with native cellular processes.

[0035] According to another aspect of the present disclosure, a method of delivery of an antisense oligo (ASO) conjugate and a gene encoding a recombinant fusion protein to a cell is provided. The method comprises transfecting the ASO conjugate into a cell via lipid nanoparticles (LNP). The method comprises introducing a recombinant fusion protein bytransfecting a plasmid vector tag into the cell via lipid nanoparticles (LNP). where the plasmid vector tag is a plasmid vector encoding a recombinant fusion protein comprising a deaminase enzyme fused to a self-labelling protein tag, delivering a gene encoding the recombinant fusion protein into the cell using modified mRNA, or delivering the gene encoding the recombinant fusion protein into the cell using a viral delivery vector. This dual-component delivery approach enables temporal control over conjugate formation within target cells and may enhance delivery efficiency for certain cell types or tissues.

[0036] According to other aspects of the present disclosure, the viral delivery vector may be an adeno-associated virus (AAV). AAV vectors provide efficient gene delivery with favorable safety profiles and tissue-specific targeting capabilities.

[0037] According to other aspects of the present disclosure, the self-labelling protein tag may include a modified haloalkane dehalogenase enzyme or a modified mammalian O6- alkylguanine-DNA-alkyltransferase (AGT) enzy me. These self-labeling systems enable intracellular conjugate formation with high specificity and efficiency.

[0038] According to other aspects of the present disclosure, the method may further compnse allowing an ASO of the ASO conjugate to regulate a disease-associated mRNA transcript. Intracellular conjugate formation enables targeting of disease-associated transcripts with precise spatial and temporal control. Princeton - 104176

[0039] According to other aspects of the present disclosure, an ASO of the ASO conjugate may be a bulge-forming antisense oligonucleotide (ASO) (Bulge ASO). Bulge-forming ASOs enhance editing selectivity by creating specific structural contexts for enzyme recognition.

[0040] According to another aspect of the present disclosure, a bulge-forming antisense oligonucleotide (ASO) (Bulge ASO) is provided. The Bulge ASO comprises a nucleotide sequence comprising discontinuous binding regions configured to position a target mRNA base within a ssRNA loop between 4 nt and 9 nt in length. The Bulge ASO comprises a protein tag coupled to the nucleotide sequence at an intermediate location within the nucleotide sequence. This structural design creates optimal substrate presentation for editing enzy mes while maintaining target specificity' through flanking binding regions.

[0041] According to other aspects of the present disclosure, each discontinuous binding region may be, independently, at least 5 nucleotides in length. Binding regions of sufficient length provide stable target recognition while allowing formation of the editing-competent bulge structure.

[0042] According to other aspects of the present disclosure, each discontinuous binding region may be, independently. 6-30 nucleotides in length. This length range balances target specificity with structural flexibility for optimal editing enzyme access.

[0043] According to other aspects of the present disclosure, the protein tag may be aHaloTag or SnapTag. These protein tags enable efficient and specific conjugation with compatible ligands under physiological conditions.

[0044] According to another aspect of the present disclosure, a method of correcting a genetic mutation is provided. The method comprises providing a bulge-forming antisense oligonucleotide (ASO) (Bulge ASO) coupled to a protein tag. The method comprises allowing discontinuous binding regions of the Bulge ASO to bind to a target sequence such that a target mRNA base is positioned within a ssRNA loop between 4 nt and 9 nt in length, the target mRNA base comprising a genetic mutation. The method comprises editing the target mRNA base. This approach enables precise correction of genetic mutations through controlled structural presentation of target bases to editing enzymes, providing therapeutic potential for genetic diseases.

[0045] The foregoing general description of the illustrative embodiments and the following detailed description thereof are merely exemplary aspects of the teachings of this disclosure and are not restrictive. Princeton - 104176

[0046] BRIEF DESCRIPTION OF FIGURES

[0047] Non-limiting and non-exhaustive examples are described with reference to the following figures.

[0048] FIG. 1 illustrates a flowchart of a method for providing a protein-antisense oligonucleotide conjugate, according to aspects of the present disclosure.

[0049] FIG. 2 illustrates a flowchart of a method for delivering an antisense oligonucleotide conjugate and a gene encoding a recombinant fusion protein to a cell, according to aspects of the present disclosure.

[0050] FIG. 3A illustrates chemical modifications for antisense oligonucleotides, according to aspects of the present disclosure.

[0051] FIG. 3B illustrates a sequence diagram showing a process of targeted RNA editing, according to aspects of the present disclosure.

[0052] FIG. 4A illustrates a nucleic acid complex for targeted RNA editing, according to aspects of the present disclosure.

[0053] FIG. 4B illustrates a chemical structure of an internal dT-Halo ligand modification, according to aspects of the present disclosure.

[0054] FIG. 5 illustrates a diagram showing a bulge-forming antisense oligonucleotide structure, according to aspects of the present disclosure.

[0055] FIG. 6 illustrates a schematic showing synthesis ofNHS-Halo ligand.

[0056] DETAILED DESCRIPTION

[0057] The following description sets forth exemplary aspects of the present disclosure. It should be recognized, however, that such description is not intended as a limitation on the scope of the present disclosure. Rather, the description also encompasses combinations and modifications to those exemplary aspects described herein.

[0058] Referring to FIG. 1, a method 100 for targeted RNA editing provides a systematic approach for modifying specific mRNA sequences within cells through the use of protein- antisense oligonucleotide (ASO) conjugates. The method 100 represents a platform that enables precise targeting of cellular mRNA transcripts for therapeutic applications, including correction of disease-causing mutations and regulation of gene expression. In some cases, the method 100 may be applied to address various genetic disorders by modifying RNA sequences at the post-transcriptional level, offering advantages over permanent genome editing approaches due to the temporary nature of RNA modifications. Princeton - 104176

[0059] The method 100 begins with a providing step 102 that involves providing a protein- ASO conjugate comprising a first protein covalently linked to an ASO. The protein-ASO conjugate represents a bifunctional molecular entity that combines the sequence-specific targeting capability of antisense oligonucleotides with the catalytic activity of RNA-moditying enzy mes. In some cases, the first protein may be a deaminase enzyme capable of converting adenosine to inosine or cytidine to uridine within target RNA sequences. The covalent linkage between the first protein and the ASO ensures spatial proximity between the targeting element and the catalytic domain, facilitating efficient and selective RNA editing at desired sites.

[0060] In some implementations, the providing step 102 may include a synthesizing step 104 that involves synthesizing a modified ASO comprising a reactive group. The synthesizing step 104 may utilize standard solid phase oligonucleotide synthesis techniques to construct antisense oligonucleotides with specific chemical modifications that enable subsequent conjugation to protein components. In some cases, the antisense oligonucleotide sequence may be synthesized using 2'-methoxyethyl (MOE) phosphoramidites with optimized coupling parameters to ensure high-quality oligonucleotide products. The coupling time during synthesis may be reduced from the standard 12 minutes to 6 minutes as recommended for 2’-MOE phosphoramidites, allowing for efficient incorporation of modified nucleotides while maintaining synthesis fidelity7. The reactive group incorporated during the synthesizing step 104 may include a chloroalkane moiety, azide, alkyne, or other bioorthogonal functional groups that facilitate selective conjugation reactions with complementary reactive groups on protein partners.

[0061] Following the synthesizing step 104, the method 100 proceeds to a producing step 106 that involves producing a recombinant fusion protein comprising the first protein fused to a self-labeling protein tag. The producing step 106 may employ heterologous expression systems to generate fusion proteins that combine the catalytic activity of RNA-moditying enzymes with the conjugation capability of self-labeling protein tags. In some cases, the recombinant fusion protein may be expressed in E. coli host cells using standard molecular biology techniques, including transformation of expression plasmids and induction of protein expression with appropriate chemical inducers. The self-labeling protein tag may include modified haloalkane dehalogenase enzymes such as HaloTag or modified mammalian O6-alkylguanine-DNA- alkyltransferase (AGT) enzymes such as SnapTag, which enable covalent attachment to complementary reactive substrates. The producing step 106 may also involve purification of the recombinant fusion protein through multiple chromatographic steps to achieve the purity levels necessary for subsequent conjugation reactions and cellular applications. Princeton - 104176

[0062] With continued reference to FIG. 1 , the method 100 advances to a forming step 108 that involves forming the protein-ASO conjugate by incubating the modified ASO with the recombinant fusion protein. The forming step 108 may utilize the complementary reactive groups present on the modified ASO and the self-labeling protein tag to establish covalent linkages between these components under controlled reaction conditions. In some cases, the incubation may be performed in buffered aqueous solutions at physiological pH and temperature to promote efficient conjugation while maintaining the structural integrity and biological activity of both the ASO and protein components. The reaction conditions during the forming step 108 may be optimized to achieve high conjugation efficiency while minimizing non-specific interactions or degradation of the molecular components. The purification of recombinant fusion proteins may involve cobalt affinity chromatography followed by anion exchange chromatography to remove impurities and achieve the protein purity necessary for effective conjugate formation. The forming step 108 may also include quality control measures to verify the successful formation of protein-ASO conjugates through analytical techniques such as gel electrophoresis or mass spectrometry.

[0063] The method 100 proceeds to a delivering step 110 that involves delivering the protein- ASO conjugate to one or more cells to edit a target mRNA sequence. The delivering step 110 may utilize various cellular delivery mechanisms to transport the protein-ASO conjugate across cell membranes and into the cytoplasm where target mRNA molecules are located. In some cases, the delivering step 110 may employ electroporation, lipofection, or other transfection methods to achieve cellular uptake of the protein-ASO conjugate. The target mRNA sequence may include disease-associated transcripts, endogenous cellular mRNAs, or plasmid-encoded reporter sequences, depending on the specific application.

[0064] As further shown in FIG. 1, the method 100 enables delivery of biological effector proteins to disease-associated mRNA transcripts to regulate their behavior and control gene expression. The approach extends beyond simple base editing to encompass broader RNA modification strategies that may influence mRNA stability, translation efficiency, or subcellular localization. In some cases, the method 100 may incorporate proteins with diverse catalytic activities, including methyltransferases, demethylases, or other RNA-modifying enzymes, to achieve specific regulatory outcomes. The flexibility of the protein-ASO conjugate system allows for customization of both the targeting specificity through ASO sequence design and the functional outcome through selection of appropriate effector proteins.

[0065] The method 100 may be implemented with continuous dosing protocols to overcome the temporary nature of RNA editing and maintain therapeutic effects over extended periods. Princeton - 104176

[0066] Unlike genome editing approaches that produce permanent modifications, RNA editing through the method 100 results in transient changes that diminish as edited mRNA molecules undergo natural degradation and replacement. In some cases, repeated administration of protein-ASO conjugates may be employed to sustain desired editing levels and maintain therapeutic efficacy. The dosing regimen may be optimized based on factors such as mRNA half-life, cellular uptake efficiency, and the duration of therapeutic effect required for specific applications.

[0067] Referring to FIG. 2, a method 200 provides an alternative approach for delivering antisense oligonucleotide (ASO) conjugates and genes encoding recombinant fusion proteins to cells through dual-component delivery strategies. The method 200 offers flexibility in therapeutic applications by separating the delivery of the ASO component from the delivery of the protein-encoding genetic material, allowing for independent optimization of each delivery pathway. In some cases, the method 200 may be employed when direct delivery of preformed protein-ASO conjugates presents challenges related to conjugate stability, cellular uptake efficiency, or manufacturing considerations. The dual-component approach of the method 200 enables the use of established gene deliver}' technologies while maintaining the targeting specificity and functional capabilities of the protein-ASO system.

[0068] The method 200 begins with a transfecting step 202 that involves transfecting the ASO conjugate into a cell via lipid nanoparticles (LNP). The transfecting step 202 utilizes lipid- based delivery vehicles to facilitate cellular uptake of modified antisense oligonucleotides that retain reactive groups for subsequent conjugation with protein partners expressed within the target cells. In some cases, the lipid nanoparticles employed in the transfecting step 202 may be formulated with cationic lipids, helper lipids, and polyethylene glycol-lipid conjugates to optimize cellular delivery and endosomal escape of the ASO conjugate. The ASO conjugate delivered during the transfecting step 202 may include the same chemical modifications and reactive groups described in the synthesizing step 104, enabling intracellular conjugation with recombinant fusion proteins once both components are present within the target cells.

[0069] With continued reference to FIG. 2, the method 200 proceeds to an introducing step 204 that involves introducing a recombinant fusion protein through multiple alternative delivery pathways. The introducing step 204 provides three distinct options for delivering the genetic material encoding the recombinant fusion protein, allowing selection of the approach that provides optimal expression levels and cellular compatibility for specific applications. The flexibility of the introducing step 204 accommodates different cell types, therapeutic contexts, and regulatory considerations that may influence the choice of gene delivery method. In some Princeton - 104176 cases, the introducing step 204 may be performed simultaneously with the transfecting step 202, while in other applications, sequential delivery of the ASO conjugate and proteinencoding genetic material may be employed to optimize timing of conjugate formation within target cells.

[0070] The introducing step 204 includes a plasmid vector tag step 206 that involves transfecting a plasmid vector tag into the cell via lipid nanoparticles (LNP). where the plasmid vector tag represents a plasmid vector encoding a recombinant fusion protein comprising a deaminase enzyme fused to a self-labelling protein tag. The plasmid vector tag step 206 utilizes the same lipid nanoparticle delivery technology employed in the transfecting step 202, providing consistency in delivery methodology while enabling nuclear delivery of genetic material for transcription and subsequent protein expression. In some cases, the plasmid vector employed in the plasmid vector tag step 206 may include mammalian expression elements such as cytomegalovirus promoters, polyadenylation signals, and antibiotic resistance markers to facilitate stable transfection and sustained protein expression in target cells. The deaminase enzyme component of the recombinant fusion protein may include natural enzy mes such as ADAR1 / 2 and APOBEC family enzymes, or artificial deaminase enzymes such as Tad8.20. Tad7. 10, TadA8r, TadDE, and TadCBE, depending on the specific editing requirements of the application.

[0071] As further shown in FIG. 2, the introducing step 204 includes a modified mRNA step 208 that involves delivering a gene encoding the recombinant fusion protein into the cell using modified mRNA. The modified mRNA step 208 provides an alternative to plasmid-based delivery that bypasses nuclear entry requirements and enables direct cytoplasmic translation of the recombinant fusion protein. In some cases, the modified mRNA employed in the modified mRNA step 208 may include chemical modifications such as pseudouridine incorporation, 5'- cap structures, and optimized untranslated regions to enhance mRNA stability, translation efficiency, and immune evasion within target cells. The modified mRNA approach of the modified mRNA step 208 may offer advantages in applications where rapid protein expression is desired or where nuclear delivery of plasmid vectors presents technical challenges. The delivering step 1 10 may utilize different lipofection reagents including Lipofectamine 2000. Lipofectamine 3000, and Lipofectamine RNAiMAX to optimize cellular uptake of both ASO conjugates and modified mRNA components, with reagent selection based on cell type compatibility and transfection efficiency requirements.

[0072] The introducing step 204 also encompasses a viral delivery vector step 210 that involves delivering the gene encoding the recombinant fusion protein into the cell using a viral Princeton - 104176 deliver^' vector. The viral deliver}' vector step 210 may employ various viral systems engineered to carry therapeutic genetic payloads while maintaining safety’ profiles appropriate for clinical applications. In some cases, the viral delivery’ vector may be an adeno-associated virus (AAV) that provides efficient gene deliver}' yvith reduced immunogenicity compared to other viral systems. The viral deliver}' vector step 210 may offer advantages in applications requiring high transduction efficiency, tissue-specific targeting, or sustained protein expression over extended periods. The viral vectors employed in the viral delivery vector step 210 may be engineered to include tissue-specific promoters, safety syvitches, or other regulatory elements that enhance therapeutic specificity while minimizing off-target effects.

[0073] Referring to FIG. 3A, various chemical modifications may be incorporated into antisense oligonucleotides to enable enhanced stability, cellular uptake, and conjugation capabilities yvhile maintaining sequence-specific hybridization properties with target mRNA molecules. The antisense oligonucleotides may incorporate multiple types of chemical modifications that collectively improve their pharmacological properties and functional performance in cellular environments. In some cases, the chemical modifications may include sugar modifications, backbone modifications, and nucleobase modifications that work synergistically to optimize the antisense oligonucleotides for therapeutic applications. The structural features illustrated in FIG. 3 A demonstrate the integration of these modifications into a cohesive molecular architecture that supports both targeting specificity and protein conjugation functionality’.

[0074] The 2'-methoxyethyl (MOE) modification represents a sugar modification that may be incorporated at the ribose position of nucleotides within the antisense oligonucleotide sequence. The 2'-methoxyethyl modification involves the attachment of a methoxyethyl group to the 2'-hydroxyl position of the ribose sugar, creating a modified nucleotide that exhibits enhanced nuclease resistance and improved binding affinity to complementary RNA sequences. In some cases, the antisense oligonucleotide sequence may be fully 2'-methoxy ethyl (MOE)-modified, meaning that every nucleotide within the sequence contains the 2'- methoxyethyl modification. The comprehensive incorporation of 2'-M0E modifications throughout the antisense oligonucleotide sequence may provide uniform protection against enzymatic degradation while maintaining the ability to form stable duplexes with target mRNA molecules. The 2'-methoxyethyl modification may also influence the pharmacokinetic properties of the antisense oligonucleotides by affecting their distribution, metabolism, and clearance within biological systems. Princeton - 104176

[0075] As further shown in FIG. 3 A, the phosphate linkages connecting adjacent nucleotides within the antisense oligonucleotide may be modified to enhance stability and cellular compatibility. The antisense oligonucleotide sequence may incorporate a phosphodiester backbone that maintains the natural phosphate linkages found in native nucleic acids while providing compatibility with the 2'-methoxyethyl modifications. In some cases, the phosphodiester backbone may offer advantages in terms of reduced immunogenicity and improved cellular tolerance compared to other backbone modifications. Alternatively, the antisense oligonucleotide sequence may be fully 2'-methoxyethyl (MOE)-modified with a phosphorothioate backbone, where one of the non-bridging oxygen atoms in each phosphate linkage may be replaced with a sulfur atom. The phosphorothioate backbone modification may provide additional nuclease resistance and may influence the pharmacokinetic properties of the antisense oligonucleotides through interactions with cellular proteins and transport mechanisms.

[0076] The nucleobase modifications illustrated in FIG. 3A include m5C (5-methylcytosine) and m5U (5 -methyluracil) modifications that may be incorporated to enhance the stability and binding properties of the antisense oligonucleotides. The 5-methylcytosine modification involves the addition of a methyl group to the 5-position of cytosine bases, creating a modified nucleobase that may exhibit altered base-pairing properties and enhanced resistance to enzymatic modification. Similarly, the 5 -methyl uracil modification may be incorporated at uracil positions to provide comparable stability enhancements and functional benefits. In some cases, these nucleobase modifications may work in conjunction with the 2'-methoxy ethyl sugar modifications and backbone modifications to create antisense oligonucleotides with optimized performance characteristics for cellular applications. The combination of multiple modification types may enable fine-tuning of the antisense oligonucleotide properties to achieve desired levels of stability, binding affinity, and cellular compatibility.

[0077] The synthesis of chemically modified antisense oligonucleotides may be accomplished through solid phase oligonucleotide synthesis techniques that accommodate the incorporation of modified phosphoramidites and specialized coupling conditions. The antisense oligonucleotide sequence may be synthesized using standard solid phase oligonucleotide synthesis with 2'-M0E phosphoramidites and optimized reaction parameters to ensure efficient incorporation of the modified nucleotides. In some cases, the coupling time during synthesis may be reduced from the standard 12 minutes to 6 minutes as recommended for 2'-MOE phosphoramidites, allowing for efficient coupling reactions while minimizing potential side reactions or degradation of the modified phosphoramidites. The synthesis process may utilize Princeton - 104176 controlled pore glass (CPG) solid supports, automated synthesizers, and purification protocols specifically adapted for modified oligonucleotides to achieve high-quality products suitable for conjugation and cellular applications. The antisense oligonucleotide sequence may be 18 nucleotides in length, providing a balance between targeting specificity and practical synthesis considerations while maintaining compatibility with the chemical modification strategies.

[0078] The incorporation of reactive groups for protein conjugation may be achieved through the attachment of specialized linkers and functional groups during or after the oligonucleotide synthesis process. The antisense oligonucleotides may be modified to include a first compatible reactive group that enables selective conjugation with complementary reactive groups present on protein partners. In some cases, the first compatible reactive group may be an azide, an alkyne, a tetrazine, an alkene, an amine, a thiol, a hydrazide, an aminooxy, an aldehyde, a ketone, an isothiocyanate, a carboxylate, amaleimide, anNHS ester, or a Gly-Gly-Gly peptide, depending on the specific conjugation strategy7employed. The selection of the first compatible reactive group may be based on factors such as reaction efficiency, selectivity, stability under physiological conditions, and compatibility with the chemical modifications present in the antisense oligonucleotide. The first compatible reactive group may represent a functional group for bio-conjugation or bio-orthogonal chemistry, enabling selective conjugation reactions that proceed efficiently in biological environments without interfering with cellular processes or biomolecules.

[0079] The chemical conjugation of proteins to antisense oligonucleotides may utilize compatible reactive groups present on both the protein and ASO components to form stable covalent linkages. The conjugation process may involve the reaction between the first compatible reactive group incorporated into the antisense oligonucleotide and a complementary’ reactive group present on the protein component. In some cases, the protein may be modified through site-directed mutagenesis, protein semi-synthesis, or unnatural amino acid mutagenesis to introduce reactive groups that are compatible with the first compatible reactive group on the antisense oligonucleotide. These protein modification strategies may enable the incorporation of non-natural amino acids, chemical handles, or other reactive functionalities that facilitate efficient conjugation while maintaining the biological activity of the protein component. The resulting protein-ASO conjugates may exhibit enhanced stability, improved cellular uptake, and optimized functional properties compared to non-conjugated components, enabling effective targeted RNA editing applications.

[0080] Referring to FIG. 3B, the targeted RNA editing process involves a systematic sequence of molecular interactions that enable precise modification of specific mRNA sequences through Princeton - 104176 the coordinated action of antisense oligonucleotide targeting elements and catalytic protein components. An antisense oligonucleotide sequence component 300 provides the foundation for sequence-specific recognition and binding to target mRNA molecules within cellular environments. The antisense oligonucleotide sequence component 300 incorporates both targeting functionality and conjugation capability through the integration of complementary binding regions and reactive attachment sites. In some cases, the antisense oligonucleotide sequence component 300 may be designed to accommodate various target mRNA sequences while maintaining consistent conjugation properties and cellular compatibility. The modular architecture of the antisense oligonucleotide sequence component 300 enables customization of targeting specificity' through sequence design while preserving the structural elements necessary for protein conjugation and functional activity.

[0081] The antisense oligonucleotide sequence component 300 includes an antisense oligonucleotide sequence 302 that provides the sequence-specific binding capability for target mRNA recognition. The antisense oligonucleotide sequence 302 may be designed with complementary’ base-pairing regions that enable hybridization to specific sequences within target mRNA molecules through Watson-Crick base pairing interactions. In some cases, the antisense oligonucleotide sequence 302 may incorporate the chemical modifications described previously, including 2'-methoxy ethyl modifications, backbone modifications, and nucleobase modifications that enhance stability and binding affinity while maintaining sequence specificity. The length and composition of the antisense oligonucleotide sequence 302 may be optimized to achieve selective binding to target sequences while minimizing off-target interactions with non-complementary mRNA molecules. The antisense oligonucleotide sequence 302 may also include regions with partial complementarity or strategic mismatches that influence the binding kinetics and thermodynamic stability of the antisense oligonucleotide-target mRNA complex.

[0082] With continued reference to FIG. 3B, the antisense oligonucleotide sequence component 300 incorporates a protein tag handle 304 that enables covalent attachment of catalytic protein components to the targeting oligonucleotide. The protein tag handle 304 may include reactive functional groups or chemical linkers that facilitate conjugation reactions with complementary reactive groups present on protein partners. In some cases, the protein tag handle 304 may be positioned at the 5' terminus, 3' terminus, or internal positions within the antisense oligonucleotide sequence 302, depending on the specific conjugation strategy and functional requirements of the application. The protein tag handle 304 may incorporate bioorthogonal reactive groups that enable selective conjugation reactions without interfering Princeton - 104176 with the base-pairing properties of the antisense oligonucleotide sequence 302 or the catalytic activity of the attached protein components. The design of the protein tag handle 304 may also consider factors such as linker length, flexibility, and chemical stability to optimize the spatial positioning and functional accessibility of the conjugated protein components.

[0083] The targeted RNA editing process involves the interaction between the antisense oligonucleotide sequence 302 and a target strand 306 that represents the mRNA molecule containing the sequence to be modified (e.g., target site 308 to be edited). The target strand 306 may include various types of mRNA molecules, including endogenous cellular transcripts, disease-associated mRNA sequences, or plasmid-encoded reporter molecules, depending on the specific application and therapeutic objectives. In some cases, the target strand 306 may contain disease-causing mutations, regulatory sequences, or other functionally relevant regions that benefit from targeted modification through RNA editing approaches. The binding interaction between the antisense oligonucleotide sequence 302 and the target strand 306 may form a duplex structure that positions the catalytic protein components in proximity to the intended modification sites. The stability and specificity of the antisense oligonucleotide sequence 302 and target strand 306 interaction may influence the efficiency and selectivity of the subsequent RNA editing reactions.

[0084] As further shown in FIG. 3B, the target strand 306 contains a target site 308 that represents the specific nucleotide position or region within the mRNA molecule that undergoes modification during the RNA editing process. The target site 308 may include adenosine residues that undergo conversion to inosine through deaminase activity, or cytidine residues that undergo conversion to uridine, depending on the catalytic specificity of the protein components employed in the editing system. In some cases, the target site 308 may be positioned adjacent to or within the binding region of the antisense oligonucleotide sequence 302, enabling the conjugated protein components to access and modify the target nucleotides through proximity -mediated catalysis. The sequence context surrounding the target site 308 may influence the efficiency and selectivity of the editing reaction, with certain sequence motifs providing enhanced substrate recognition and catalytic activity. The protein component may preferentially edit TA dinucleotide motifs within target sequences, indicating that the target site 308 may be selected based on the presence of favorable sequence contexts that promote efficient editing reactions.

[0085] The catalytic functionality of the targeted RNA editing system may be provided by a recombinant fusion protein component 310 that combines the enzy matic activity of RNA- modifying enzymes with the conjugation capability’ of self-labeling protein tags. The Princeton - 104176 recombinant fusion protein component 310 may be designed to maintain the catalytic activity of the enzyme component while enabling covalent attachment to the antisense oligonucleotide sequence 302 through the protein tag handle 304. In some cases, the recombinant fusion protein component 310 may be expressed and purified using the producing step 106 described previously, involving heterologous expression in bacterial or eukary otic host cells followed by chromatographic purification to achieve the purity levels necessary for conjugation and cellular applications. The recombinant fusion protein component 310 may incorporate linker sequences or structural elements that optimize the spatial relationship between the catalytic domain and the conjugation domain while preserving the functional properties of both components.

[0086] The recombinant fusion protein component 310 includes a protein 312 that provides the catalytic activity for RNA modification reactions. The protein 312 may be a deaminase enzyme capable of converting specific nucleotides within RNA substrates through hydrolytic deamination reactions. In some cases, the protein 312 may be a natural deaminase enzyme including AD ARI 12 and APOBEC family enzymes that exhibit intrinsic RNA editing activity and substrate specificity patterns derived from their biological functions. Alternatively, the protein 312 may be an artificial deaminase enzyme including Tad8.20. Tad7.10. TadA8r. TadDE, and TadCBE that have been developed through directed evolution or rational design approaches to enhance catalytic activity7, substrate promiscuity7, or other functional properties. The selection of the protein 312 may be based on factors such as catalytic efficiency, substrate specificity, stability under cellular conditions, and compatibility with the conjugation and delivery strategies employed in the targeted RNA editing system.

[0087] With continued reference to FIG. 3B, the recombinant fusion protein component 310 incorporates a protein component 314 that facilitates the conjugation process and enables covalent attachment to the antisense oligonucleotide sequence 302. The protein component 314 may include self-labeling protein tags or other conjugation domains that react with complementary7functional groups present in the protein tag handle 304. In some cases, the protein component 314 may include modified haloalkane dehalogenase enzymes such as HaloTag or modified mammalian O6-alkylguanine-DNA-alkyltransferase enzymes such as SnapTag that enable covalent attachment to chloroalkane or benzylguanine substrates, respectively. The protein component 314 may also facilitate protein trans-splicing utilizing split inteins, providing an alternative conjugation mechanism that enables the formation of protein-ASO conjugates through protein-mediated ligation reactions. The choice of protein component 314 may influence the conjugation efficiency, stability of the resulting conjugates, and compatibility with cellular delivery and functional applications. Princeton - 104176

[0088] The recombinant fusion protein component 310 includes a protein tag 318 that represents the specific conjugation domain responsible for covalent attachment to the protein tag handle 304. The protein tag 318 may exhibit high specificity and efficiency for conjugation reactions with complementary substrates while maintaining stability’ and functionality under physiological conditions. In some cases, the protein tag 318 may be engineered to optimize conjugation kinetics, reduce non-specific interactions, or enhance the stability of the resulting protein-ASO conjugates. The protein tag 318 may also be selected based on compatibility with cellular environments, minimal immunogenicity, and ease of incorporation into fusion protein constructs without disrupting the catalytic activity of the protein 312. The spatial positioning and orientation of the protein tag 318 within the recombinant fusion protein component 310 may influence the accessibility of the conjugation site and the functional properties of the resulting conjugates.

[0089] The targeted RNA editing process results in the formation of an edited target site 316 that represents the modified nucleotide position within the target strand 306 following the catalytic action of the protein 312. The edited target site 316 may contain inosine residues resulting from adenosine deamination or uridine residues resulting from cytidine deamination, depending on the catalytic specificity of the protein 312 employed in the editing system. In some examples, the editing efficiency at the edited target site 316 has achieved at least a 66.9% A-to-I conversion at the major target site, demonstrating the effectiveness of the proximity- mediated catalysis enabled by the protein-ASO conjugate system. The editing reaction may reach maximum efficiency at 24 hours post-transfection, indicating that the temporal dynamics of the editing process may be influenced by factors such as conjugate stability, cellular uptake kinetics, and mRNA turnover rates. The formation of the edited target site 316 may enable the recombinant fusion protein to recode at least some genetic information present in mRNA, potentially correcting disease-causing mutations or introducing functional modifications that alter protein coding sequences, regulatory elements, or other functionally relevant regions within the target mRNA molecules.

[0090] The targeted RNA editing system may also enable the recombinant fusion protein to regulate a disease-associated mRNA transcript through modifications that influence mRNA stability, translation efficiency, or other post-transcriptional regulatory mechanisms. In some cases, the editing reactions may target regulatory sequences, splice sites, or other non-coding regions within disease-associated mRNA transcripts to achieve therapeutic outcomes through mechanisms beyond simple codon correction. The flexibility of the protein-ASO conjugate system allows for the incorporation of various catalytic activities and targeting specificities to Princeton - 104176 address diverse therapeutic applications and disease contexts. The edited target site 316 may serve as a foundation for downstream cellular processes that recognize and respond to the modified nucleotides, potentially leading to altered gene expression patterns, protein function, or cellular phenotypes that contribute to therapeutic efficacy.

[0091] Referring to FIG. 4A, the nucleic acid complex formation involves sophisticated molecular interactions that enable precise positioning of catalytic components relative to target mRNA sequences through the integration of targeting oligonucleotides and conjugated protein partners. The complex architecture illustrated in FIG. 4A demonstrates how the antisense oligonucleotide sequence 302 may be configured to hybridize with the target strand 306 while simultaneously providing attachment sites for recombinant fusion protein components that facilitate targeted RNA editing reactions. In some cases, the nucleic acid complex may incorporate structural features that optimize the spatial relationship between the targeting elements and the catalytic domains to enhance editing efficiency and selectivity. The formation of stable nucleic acid complexes may depend on factors such as sequence complementarity, thermodynamic stability, and the presence of chemical modifications that influence hybridization kinetics and duplex stability under physiological conditions.

[0092] The antisense oligonucleotide sequence 302 within the nucleic acid complex may be designed to form specific base-pairing interactions with complementary7regions of the target strand 306. creating a duplex structure that positions the target site 308 in proximity to the conjugated catalytic components. The hybridization between the antisense oligonucleotide sequence 302 and the target strand 306 may involve Watson-Crick base pairing interactions that provide sequence specificity while accommodating the presence of chemical modifications and conjugated protein components. In some cases, the duplex formation may create local structural changes in the target mRNA that enhance accessibility of the target site 308 to the catalytic activity of the attached protein components. The stability of the antisense oligonucleotide sequence 302 and target strand 306 interaction may influence the duration of the editing complex and the efficiency of the subsequent catalytic reactions, with stronger binding interactions potentially leading to more sustained editing activity.

[0093] With continued reference to FIG. 4A, the nucleic acid complex incorporates an internally installed handle 402 that provides a conjugation site positioned within the antisense oligonucleotide sequence 302 rather than at the terminal positions. The internally installed handle 402 may offer advantages in terms of spatial positioning and functional accessibility compared to terminal conjugation approaches by enabling the attachment of protein components at locations that optimize their proximity to target sites within the hybridized Princeton - 104176 complex. In some cases, the internally installed handle 402 may be positioned at specific nucleotide positions within the antisense oligonucleotide sequence 302 based on structural modeling or empirical optimization to achieve desired editing efficiency and selectivity profiles. The internally installed handle 402 may incorporate reactive functional groups or chemical linkers that enable covalent attachment of the recombinant fusion protein component 310 while maintaining the base-pairing capability of the surrounding nucleotides. The design of the internally installed handle 402 may also consider factors such as linker flexibility, chemical stability, and potential interference with duplex formation to ensure optimal performance of the resulting nucleic acid complex.

[0094] The recombinant fusion protein component 310 may be attached to the internally installed handle 402 through covalent conjugation reactions that form stable linkages between the protein and oligonucleotide components. The conjugation process may utilize the selflabeling protein tag functionality within the protein component 314 to react with complementary reactive groups present in the internally installed handle 402. In some cases, the self-labeling protein tag may include a modified haloalkane dehalogenase enzyme such as HaloTag that reacts with chloroalkane substrates incorporated into the internally installed handle 402. Alternatively, the self-labeling protein tag may include a modified mammalian O6- alkylguanine-DNA-alkyltransferase (AGT) enzyme such as SnapTag that reacts with benzylguanine or benzylcy tosine substrates present in the internally installed handle 402. The selection of the conjugation chemistry may influence the efficiency of complex formation, the stability of the resulting conjugates, and the functional properties of the assembled nucleic acid complex.

[0095] As further shown in FIG. 4A, the protein 312 within the recombinant fusion protein component 310 provides the catalytic activity for modifying nucleotides at the target site 308 within the hybridized complex. The spatial positioning of the protein 312 relative to the target site 308 may be influenced by the location of the internally installed handle 402 and the structural properties of the linker regions connecting the protein to the oligonucleotide. In some cases, the protein 312 may be a deaminase enzyme that converts adenosine residues to inosine or cytidine residues to uridine through hydrolytic deamination reactions that occur when the enzyme gains access to susceptible nucleotides within the target strand 306. The catalytic activity7of the protein 312 may be enhanced by the proximity-mediated targeting enabled by the nucleic acid complex, allowing for efficient modification of specific nucleotides while minimizing off-target effects on non-compl ementary mRNA molecules. The protein 312 may exhibit sequence-dependent catalytic preferences that influence the efficiency of editing Princeton - 104176 reactions at different target sites, with certain sequence contexts providing more favorable substrates for the deaminase activity.

[0096] The producing step 106 for generating the recombinant fusion protein component 310 may utilize heterologous expression in a host cell to achieve high-level production of the fusion protein with appropriate folding and functional properties. The heterologous expression system may involve the transformation of expression vectors containing the genetic sequences encoding the recombinant fusion protein component 310 into suitable host cells that provide the cellular machinery for protein synthesis, folding, and post-translational modifications. In some cases, the host cell may be E. coli, which offers advantages in terms of rapid grow th, high expression levels, and well-established protocols for protein production and purification. The use of E. coli as the host cell may enable cost-effective production of the recombinant fusion protein component 310 while providing sufficient protein yields for conjugation and cellular applications. Alternatively, the host cell may be an organism other than E. coli, such as yeast, insect cells, or mammalian cells, which may offer advantages in terms of protein folding, post- translational modifications, or functional activity for specific protein components that require eukaryotic expression systems.

[0097] Referring to FIG. 4B, the chemical structure of chloroalkane-modified antisense oligonucleotide conjugates demonstrates the integration of reactive functional groups that enable covalent attachment to HaloTag protein partners through nucleophilic substitution reactions. The chloroalkane modification may be incorporated into the antisense oligonucleotide through the attachment of a chloroalkane-terminated linker that provides the reactive substrate for HaloTag-mediated conjugation reactions. In some cases, the chloroalkane group may be connected to the oligonucleotide through a series of chemical linkages that include ether bonds, amide bonds, or other stable covalent connections that maintain the integrity of the reactive group while providing appropriate spacing and flexibility for conjugation reactions. The linker region connecting the chloroalkane group to the oligonucleotide may be designed to optimize the accessibility' of the reactive group to the HaloTag enzyme while minimizing interference with the base-pairing properties of the antisense oligonucleotide sequence 302. The chemical stability of the chloroalkane modification under physiological conditions may influence the shelf-life and functional performance of the modified oligonucleotides in cellular applications.

[0098] The chloroalkane-modified antisense oligonucleotide conjugates may incorporate multiple types of chemical modifications that work synergistically to enhance their functional properties and compatibility with protein conjugation reactions. The oligonucleotide portion of Princeton - 104176 the conjugate may include the 2'-methoxyethyl modifications, backbone modifications, and nucleobase modifications described previously, providing enhanced nuclease resistance and improved binding affinity to target mRNA sequences. In some cases, the combination of these modifications with the chloroalkane conjugation group may create antisense oligonucleotides with optimized properties for both targeting specificity and protein attachment capability. The chemical modifications may also influence the pharmacokinetic properties of the conjugates, affecting their distribution, cellular uptake, and clearance within biological systems. The integration of multiple modification types may enable fine-tuning of the conjugate properties to achieve desired levels of stability', binding affinity, and conjugation efficiency for specific applications.

[0099] With continued reference to FIG. 4B, the delivering step 110 may be performed using electroporation to introduce the protein-ASO conjugate into cells through the application of electrical pulses that create temporary pores in cellular membranes. The electroporation process may involve the application of controlled electrical fields to cell suspensions containing the protein-ASO conjugates, creating transient membrane permeabilization that allows for the uptake of large molecular complexes that would not normally cross intact cell membranes. In some cases, the electroporation parameters may be optimized based on factors such as cell type, conjugate size, and desired uptake efficiency to achieve effective cellular delivery while maintaining cell viability. The electroporation approach may offer advantages in terms of delivery efficiency and reproducibility compared to other transfection methods, particularly for applications involving primary cells or cell types that are resistant to lipid-based transfection reagents. The electrical parameters used during electroporation, including voltage, pulse duration, and pulse number, may be adjusted to optimize the balance between cellular uptake and cell survival for specific applications.

[0100] The method 100 may be performed with ASO concentrations as low as 6.25 nM while maintaining editing efficiency, demonstrating the high potency of the protein-ASO conjugate system for targeted RNA editing applications. The low concentration requirements may offer advantages in terms of reduced material costs, minimized potential for off-target effects, and improved therapeutic index for clinical applications. In some cases, the high potency of the system may result from the proximity-mediated catalysis enabled by the covalent conjugation between the targeting oligonucleotide and the catalytic protein components, allowing for efficient editing reactions even at low conjugate concentrations. The concentration-response relationship may exhibit a relatively flat profile across a range of concentrations, indicating that the system may be robust to variations in dosing or cellular uptake efficiency. In some Princeton - 104176 implementations, the method 100 may achieve maximal editing at between 12.5 nM and 25 nM ASO concentrations, providing a practical working range for experimental and therapeutic applications that balances editing efficiency with material usage and potential side effects.

[0101] Referring to FIG. 5, the bulge-forming antisense oligonucleotide structure represents an advanced molecular architecture that enables precise positioning of target nucleotides within accessible single-stranded RNA loops for enhanced editing efficiency. The antisense oligonucleotide sequence 302 may be configured with discontinuous binding regions that create specific hybridization patterns with the target strand 306, resulting in the formation of structured complexes that optimize the accessibility of target sites to conjugated catalytic proteins. In some cases, the discontinuous binding regions may be designed to form stable duplex regions flanking a central unpaired region, creating a bulge structure that positions specific nucleotides within a single-stranded loop where they become accessible to deaminase enzymes. Each discontinuous binding region may be at least 5 nucleotides in length, such as being independently 6-30 nucleotides in length.

[0102] The bulge-forming architecture may provide advantages over fully complementary antisense oligonucleotides by creating localized regions of single-stranded RNA that serve as preferred substrates for RNA editing enzy mes while maintaining overall sequence specificity through the flanking duplex regions.

[0103] The antisense oligonucleotide sequence 302 within the bulge-forming structure may incorporate discontinuous binding regions that are configured to hybridize with complementary sequences on either side of the target modification site. Each discontinuous binding region may be, independently, at least 5 nucleotides in length to provide sufficient binding stability and sequence specificity for target recognition. In some cases, each discontinuous binding region may be, independently, 6-30 nucleotides in length, allowing for optimization of binding affinity and specificity while accommodating various target sequence contexts and structural requirements. The length and composition of the discontinuous binding regions may7be adjusted based on factors such as target mRNA secondary7structure, thermodynamic stability requirements, and the specific positioning needs for optimal editing efficiency. The spacing between the discontinuous binding regions may be designed to create a bulge 502 of appropriate size and structure to accommodate the target nucleotides and provide accessibility for the conjugated protein components.

[0104] With continued reference to FIG. 5, the bulge 502 represents a single-stranded RNA loop structure that forms when the antisense oligonucleotide sequence 302 hybridizes with the target strand 306 through the discontinuous binding regions while leaving an intermediate Princeton - 104176 region unpaired. The bulge 502 may be configured to position a target mRNA base within a single-stranded RNA loop between 4 nucleotides and 9 nucleotides in length, providing an optimal substrate structure for deaminase enzy me activity. In some cases, the size of the bulge 502 may be controlled through the design of the antisense oligonucleotide sequence 302, with the spacing between the discontinuous binding regions determining the number of unpaired nucleotides that form the single-stranded loop. The bulge 502 may create a localized region of increased nucleotide accessibility’ that enhances the catalytic efficiency of RNA editing enzymes compared to fully duplex structures where target nucleotides may be less accessible to enzy matic modification. The structural properties of the bulge 502 may also influence the binding kinetics and residence time of the editing complex, potentially affecting the overall efficiency of the RNA editing process.

[0105] The target site 504 within the bulge 502 represents the specific nucleotide position that undergoes modification during the RNA editing process. The target site 504 may be positioned within the single-stranded RNA loop created by the bulge 502, where the unpaired nucleotides become accessible to the catalytic activity’ of the conjugated deaminase enzymes. The single stranded RNA (ssRNA) loop may be between 4 nt and 9 nt in length. In some cases, the target site 504 may contain adenosine residues that undergo conversion to inosine through the action of adenosine deaminases, or cytidine residues that undergo conversion to uridine through cytidine deaminase activity. The positioning of the target site 504 within the bulge 502 may be optimized through the design of the antisense oligonucleotide sequence 302 to ensure that the target nucleotides are located within the single-stranded loop region where they can be efficiently accessed by the conjugated protein components. The sequence context surrounding the target site 504 may influence the editing efficiency, with certain nucleotide environments providing more favorable substrates for deaminase activity than others.

[0106] As further shown in FIG. 5, the recombinant fusion protein component 310 may be attached to the antisense oligonucleotide sequence 302 through the internally installed handle 402, which may be positioned at an intermediate location within the nucleotide sequence to optimize the spatial relationship between the catalytic protein and the target site 504. The internally installed handle 402 may enable the attachment of the protein 312 at a position that provides optimal access to the bulge 502 and the target site 504 contained within the singlestranded loop. In some cases, the internally installed handle 402 may be coupled to a protein tag that facilitates the conjugation process, with the protein tag being a HaloTag or SnapTag that enables covalent attachment to complementary’ reactive substrates incorporated into the antisense oligonucleotide sequence 302. The positioning of the internally installed handle 402 Princeton - 104176 may be selected based on structural modeling or empirical optimization to achieve the desired spatial relationship between the catalytic domain and the target nucleotides within the bulge 502.

[0107] The bulge-forming antisense oligonucleotide structure may enable correction of various types of genetic mutations through targeted RNA editing at the post-transcriptional level. The method may be used to correct disease-causing missense mutations at the RNA level by converting specific nucleotides within coding sequences to restore proper amino acid sequences in the resulting proteins. In some cases, the bulge-forming structure may facilitate the correction of disease-causing nonsense mutations at the RNA level by converting premature stop codons to sense codons that allow for continued translation and production of functional proteins. The method may also be used to correct mutations that affect splicing at the RNA level by modifying splice site sequences or regulator}' elements that influence the splicing process. Additionally, the method may be used to correct mutations that affect translation initiation at the RNA level by modifying sequences within the 5' untranslated region or other regulatory elements that influence ribosome binding and translation efficiency.

[0108] The bulge-forming antisense oligonucleotide approach may be particularly applicable to the correction of mutations in disease-associated genes such as KRAS and PTEN. The method may target disease-causing mutations in KRAS and PTEN genes by designing antisense oligonucleotide sequences 302 with discontinuous binding regions that create bulge structures positioning the mutant nucleotides within accessible single-stranded loops. In some cases, the correction of KRAS mutations may involve the conversion of mutant codons back to wild- type sequences through targeted deamination reactions, potentially restoring normal protein function and cellular signaling pathways. Similarly, the correction of PTEN mutations may involve the restoration of tumor suppressor function through targeted RNA editing that corrects disease-causing sequence alterations. The bulge-forming structure may provide enhanced editing efficiency for these applications by creating optimal substrate structures for the conjugated deaminase enzymes while maintaining sequence specificity through the flanking duplex regions formed by the discontinuous binding regions.

[0109] The delivering step 110 may utilize electroporation as a delivery mechanism to introduce protein-ASO conjugates into target cells through the application of controlled electrical pulses that create transient membrane permeabilization. The electroporation process may involve the suspension of cells in conductive media containing the protein-ASO conjugates, followed by the application of electrical fields with specific voltage and pulse duration parameters that optimize cellular uptake while maintaining cell viability. In some Princeton - 104176 cases, the electroporation parameters may be adjusted based on cell type characteristics, with different cell lines requiring distinct voltage settings, pulse durations, and recovery conditions to achieve effective delivery. The electroporation approach may offer advantages for delivering large molecular complexes such as protein-ASO conjugates that may not efficiently cross cellular membranes through passive diffusion or receptor-mediated uptake mechanisms. The electrical parameters employed during electroporation may range from 100 to 160 volts with pulse durations of 10 to 20 milliseconds, depending on the specific cell type and delivery requirements for the application.

[0110] The delivering step 110 may alternatively employ lipofection techniques that utilize lipid-based transfection reagents to facilitate cellular uptake of protein-ASO conjugates through membrane fusion and endocytosis mechanisms. The lipofection process may involve the formation of lipid-nucleic acid complexes or lipoplexes that interact with cellular membranes to enable intracellular delivery' of the conjugated components. In some cases, various lipofection reagents may be employed to optimize delivery' efficiency for different cell types and experimental conditions, with reagent selection based on factors such as transfection efficiency, cell viability, and compatibility with the protein-ASO conjugate system. The lipofection approach may provide gentler delivery conditions compared to electroporation, potentially preserving the structural integrity' and functional activity' of sensitive protein components within the conjugates. The formation of lipoplexes may also protect the protein- ASO conjugates from degradation during the delivery process while facilitating their release within the cellular environment.

[0111] The viral delivery' vector step 210 may employ adeno-associated virus (AAV) systems as delivery vehicles for genetic material encoding recombinant fusion proteins in targeted RNA editing applications. The AAV delivery’ approach may offer advantages in terms of tissuespecific targeting, sustained gene expression, and reduced immunogenicity compared to other viral delivery' systems. In some cases, the AAV vectors may be engineered with tissue-specific promoters that direct expression of the recombinant fusion protein components to particular cell types or anatomical locations, enabling targeted therapeutic applications while minimizing off-target effects. The AAV delivery’ system may accommodate various serotypes that exhibit different tissue tropisms and transduction efficiencies, allowing for selection of appropriate viral variants based on the intended therapeutic application and target tissue characteristics. The genetic payload delivered through AAV vectors may include expression cassettes encoding the recombinant fusion protein component 310 with appropriate regulatory elements to ensure sustained protein expression levels that support effective RNA editing activity. Princeton - 104176

[0112] The modified mRNA step 208 may provide an alternative delivery approach that bypasses the need for nuclear entry and DNA integration by delivering modified mRNA molecules that undergo direct cytoplasmic translation to produce the recombinant fusion protein components. The modified mRNA approach may incorporate chemical modifications such as pseudouridine substitutions, optimized codon usage, and enhanced 5' cap structures that improve mRNA stability and translation efficiency while reducing innate immune activation. In some cases, the modified mRNA molecules may be formulated with lipid nanoparticles or other delivery vehicles that facilitate cellular uptake and protect the mRNA from degradation during the delivery' process. The modified mRNA approach may enable rapid protein expression kinetics compared to DNA-based delivery' methods, potentially allowing for more precise temporal control over the RNA editing process. The mRNA delivery system may also offer safety' advantages by avoiding permanent genetic modifications and enabling transient expression of the editing components for applications where temporary RNA modification may be desired.

[0113] The method 100 may target endogenous mRNA transcripts including ACTB mRNA to demonstrate the capability of the protein-ASO conjugate system for modifying naturally occurring cellular transcripts. The targeting of ACTB mRNA may serve as a model system for evaluating the efficiency' and specificity of RNA editing approaches on endogenous transcripts that are expressed at physiological levels within cells. In some cases, the antisense oligonucleotide sequence 302 may be designed with complementary regions that hybridize to specific sequences within the ACTB mRNA transcript, positioning the conjugated deaminase enzymes in proximity' to target adenosine or cytidine residues for editing reactions. The editing of endogenous ACTB mRNA may provide insights into the cellular uptake, stability, and functional activity of the protein-ASO conjugates under physiological conditions where the target transcripts are subject to normal cellular processing and regulation mechanisms. The ACTB targeting approach may also enable evaluation of editing efficiency across different regions of the mRNA molecule, including coding sequences, untranslated regions, and regulatory elements that may exhibit varying accessibility to the conjugated editing enzymes.

[0114] The method 100 may also target plasmid-encoded mRNA transcripts to enable controlled evaluation of RNA editing efficiency and specificity under defined experimental conditions. The plasmid-encoded targets may provide advantages for mechanistic studies and optimization experiments by allowing precise control over target sequence composition, expression levels, and cellular localization of the substrate mRNA molecules. In some cases, the plasmid-encoded mRNA transcripts may incorporate reporter sequences or selectable Princeton - 104176 markers that facilitate quantitative assessment of editing outcomes through functional assays or analytical techniques. The plasmid delivery approach may enable co-transfection of both the target mRNA and the protein-ASO conjugates, allowing for synchronized expression and editing reactions within the same cellular environment. The plasmid-encoded targets may also facilitate structure-activity relationship studies by enabling systematic variation of target sequence contexts, secondary structures, or regulatory elements that may influence editing efficiency and selecti \ i ty.

[0115] The method 100 may achieve editing in both 3'-untranslated region (UTR) and coding sequence regions of target mRNAs with different efficiencies that reflect the distinct structural and functional characteristics of these mRNA regions. The 3 '-UTR regions may provide different accessibility profiles for the conjugated editing enzymes compared to coding sequences due to differences in secondary structure, protein binding, or ribosome occupancy that may influence the ability of the protein-ASO conjugates to access target nucleotides. In some cases, the editing efficiency in 3'-UTR regions may be influenced by the presence of regulatory elements such as microRNA binding sites, AU-rich elements, or other sequence motifs that recruit cellular proteins and may affect the accessibility of target sites to the conjugated deaminase enzymes. The coding sequence regions may present different challenges for RNA editing due to the presence of ribosome binding sites, translation initiation complexes, or other protein-RNA interactions that occur during the translation process. The differential editing efficiencies observed in different mRNA regions may provide insights into the mechanistic factors that influence the activity of protein-ASO conjugates and may inform the design of optimized targeting strategies for specific therapeutic applications.

[0116] The transfecting step 202 may utilize lipid nanoparticles (LNP) as delivery' vehicles for the antisense oligonucleotide conjugates, providing a versatile platform that may accommodate various oligonucleotide modifications and conjugation chemistries. The lipid nanoparticle formulations may incorporate ionizable lipids, helper lipids, cholesterol, and polyethylene gly col-lipid conjugates that work together to facilitate cellular uptake and endosomal escape of the oligonucleotide cargo. In some cases, the lipid nanoparticle composition may be optimized based on the specific properties of the antisense oligonucleotide conjugates, including their size, charge, and chemical modifications, to achieve effective encapsulation and delivery'. The lipid nanoparticle approach may provide protection for the oligonucleotide conjugates during circulation and storage while enabling controlled release within target cells through pH- dependent membrane fusion mechanisms. The LNP delivery system may also offer scalability Princeton - 104176 advantages for therapeutic applications by providing a standardized formulation platform that may be adapted for different oligonucleotide sequences and conjugation strategies.

[0117] A recoding step 112 may enable the recombinant fusion protein to recode at least some genetic information present in mRNA through targeted deamination reactions that convert specific nucleotides within coding sequences. The recoding step 112 may involve the conversion of adenosine residues to inosine, which may be recognized as guanosine during translation, effectively changing the codon identity and the corresponding amino acid incorporated into the resulting protein. In some cases, the recoding step 112 may target diseasecausing mutations within coding sequences to restore wild-type amino acid sequences and protein function through post-transcriptional correction mechanisms. The recoding step 112 may also involve cytidine to uridine conversions that change codon assignments and may correct nonsense mutations by converting premature stop codons to sense codons that allow continued translation. The efficiency and specificity of the recoding step 112 may depend on factors such as the sequence context surrounding the target nucleotides, the accessibility' of the editing sites within the mRNA secondary structure, and the catalytic properties of the conjugated deaminase enzymes.

[0118] A regulating step 114 may allow the recombinant fusion protein to regulate a disease- associated mRNA transcript through modifications that influence mRNA stability', translation efficiency, or other post-transcriptional regulatory mechanisms. The regulating step 114 may involve targeting regulatory sequences within disease-associated transcripts, including splice sites, translation initiation sites, or microRNA binding sites that control gene expression at the post-transcriptional level. In some cases, the regulating step 114 may modily sequences within the 5' or 3' untranslated regions of disease-associated mRNAs to alter their interaction with regulatory proteins or non-coding RNAs that control mRNA fate and function. The regulating step 114 may also target sequences that influence mRNA localization, processing, or degradation pathways to achieve therapeutic outcomes through mechanisms that extend beyond simple codon correction. The regulatory modifications introduced during the regulating step 114 may provide therapeutic benefits by modulating the expression levels or functional properties of disease-associated proteins without requiring direct correction of coding sequence mutations.

[0119] The self-labelling protein tag within the recombinant fusion protein component 310 may include a modified haloalkane dehalogenase enzyme that enables covalent attachment to chloroalkane substrates incorporated into the antisense oligonucleotide sequence 302. The modified haloalkane dehalogenase enzy me may be engineered to exhibit enhanced stability, Princeton - 104176 reduced size, or improved conjugation kinetics compared to the wild-type enzyme while maintaining the ability to form covalent bonds with chloroalkane substrates through nucleophilic substitution reactions. In some cases, the haloalkane dehalogenase enzyme may be derived from bacterial sources and may be modified through directed evolution or rational design approaches to optimize its performance in mammalian cellular environments. The chloroalkane substrate recognition and binding may occur through specific active site interactions that position the reactive chloroalkane group for nucleophilic attack by a catalytic aspartate residue, resulting in the formation of a stable covalent bond between the protein and oligonucleotide components.

[0120] The self-labelling protein tag may alternatively include a modified mammalian O6- alkylguanine-DNA-alkyltransferase (AGT) enzyme that enables covalent attachment to benzylguanine or benzylcytosine substrates incorporated into the antisense oligonucleotide sequence 302. The modified AGT enzyme may be engineered to exhibit improved conjugation efficiency, reduced immunogenicity, or enhanced stability7compared to the wild-ty pe protein while maintaining the ability to transfer alkyl groups from benzylguanine substrates to a catalytic cysteine residue within the active site. In some cases, the AGT-based conjugation system may offer advantages in terms of reaction kinetics, substrate availability, or compatibility7with specific oligonucleotide modification strategies. The benzy lguanine substrate recognition may involve specific binding interactions within the AGT active site that position the substrate for nucleophilic attack by the catalytic cysteine residue, resulting in the formation of a stable thioether linkage between the protein and oligonucleotide components. The AGT conjugation approach may provide an alternative to haloalkane-based systems for applications where different conjugation chemistries or substrate properties may be advantageous for specific therapeutic or research applications.

[0121] Examples

[0122] Plasmid Construction for all examples. Gene blocks for human codon-optimized deaminases, HaloTag, and PTEN Q245X were obtained from Twist Biosciences, restriction enzyme-digested, and ligated into pcDNA5 vector with T4 DNA ligase (New England Biolabs). Wild-type PTEN control was synthesized by site-directed mutagenesis of the PTEN Q245X and overlap extension PCR. For REPAIRv2 and DECOR gRNAs, vectors for PspCasl3b (REPAIR) (Addgene #103854) and RfxCasl3d (DECOR) (Addgene #109053) were digested with Bbsl-HF (New England Biolabs) and purified following agarose gel electrophoresis. Targeting gRNA oligos were obtained from Sigma, annealed and Princeton - 104176 phosphorylated with T4 PNK (New England Biolabs), diluted 1:200, and 1 pL was ligated into 50 ng vector. For REPAIRv2 gRNA design, the A:C mismatch was positioned 20 nt from the 3'-end of the gRNA. For DECOR, the gRNA was designed to anneal 15 nt from the target edit site based on precedent. For all plasmid cloning, crude ligated product was transformed into DH5-alphaE. coli, liquid cultures were grown from single colonies, and miniprepped (Qiagen). Purified plasmids were validated by Sanger sequencing (Genewiz - Azenta Life Sciences).

[0123] General cell culture for all examples. All cells were grown at 37 °C in a humidified atmosphere with 5% CO2 in DMEM (Life Technologies) supplemented with 10% fetal bovine serum (Atlanta), lx penicillin-streptomycin (Gibco Life Technologies), and 2 mM L-glutamine (Life Technologies).

[0124] Stable cell line generation for all examples. To generate stable cell lines expressing editor-HaloTag fusion proteins, Flp-In TRex 293 cells were seeded at 6 x 10A5 cells per well in six-well plates, and co-transfected with the appropriate pCDNA5 / FRT / TO-editor-HaloTag plasmid (0.2 pg) and pOG44 plasmid (L8pg, Thermo Fisher). After selection with 100 pg / mL hygromycin B and 15 pg / mL blasticidin, colonies were expanded.

[0125] Example 1 - ASO-directed RNA editing with TadA8.20

[0126] Bioconjugation and ASO-based targeting of evolved TadA-derived base editors using the HaloTag system as explored. A chloroalkane-modified version of the 18 nt fully 2’- methoxy ethyl (MOE)-modified therapeutic ASO Spinraza50 (“Halo-Spinraza 1”) with a phosphodi ester backbone (rather than the phosphorothioate backbone in the drug) was syntheized. See FIGS. 3A-3B. The chloroalkane Halo ligand was appended at the 5’ terminus of the ASO by incorporation of a 5 ’-amino modifier during solid phase synthesis, followed by reaction with NHS-Halo ligand (see FIG. 6 and Table 1)

[0127] Next, Flp-In TRex 293 cells were generated that express a tetracycline-inducible RNA editor-HaloTag fusion. The engineered adenosine deaminase enzyme TadA8.2044 w as chosen for this example, which has been shown to possess efficient RNA editing activity. Halo- Spinraza 1 was co-transfected together with a plasmid expressing an mRNA reporter containing the complementary 18 nt Spinraza binding sequence. After 24 h. 41.6% in ce / Zw / ocrosslinking was observed between TadA8.20-HaloTag fusion protein and Halo- Spinraza 1 by Western blot.

[0128] Next. A-to-I editing was quantified in a 250 bp region centered around the Spinraza binding site by RT-PCR followed by Sanger sequencing. 21 A-to-I editing events demonstrating statistically significant mutation rates were identified as compared to control Princeton - 104176 cells expressing only the TadA8.20-HaloTag fusion. The major site of editing, with 68.6% A- to-I conversion, occurred at the first A residue immediately 3 ’ downstream of the ASO binding site (adjacent to the 5’ end of the ASO where TadA8.20-HaloTag protein is tethered). Statistically significant editing sites were also detected both 5’ upstream of the ASO binding site as well as further downstream in the 3 ’ direction, but editing efficiency was <25% for these sites, and editing events were considerably reduced in frequency and conversion efficiency at a distance >50 nt away from the ASO binding site.

[0129] Table 1. ASO and gRNA Sequences. Items with an * are fully 2'-M0E, all C are m5C, all U are m5U; Halo = chloroalkane Hal oTag -ligand connected through 5'-Amino-5 modifier or amino modifier C2-dT (Glen Research). sgRNAs for REPAIRv2 and DECOR are canonical RNA nucleotides. Bolded sequences indicate the direct repeat sequence unique to PspCasl3b (REPAIRv2) or RfxCas 13d (DECOR). Princeton - 104176

[0130] The effect of ASO concentration on RNA editing was also investigated. A dose titration was performed with Halo-Spinraza 1 and observed minimal concentration dependence. Editing levels at the major target site immediately 3' to the ASO binding site were nearly constant from 6.25 nM (63.6% editing) to 200 nM ASO (61.8% editing). Similarly, production of crosslinked

[0131] ASO-editor complex was largely invariant over this ASO concentration range. Editing levels did decrease marginally at concentrations higher than 25 nM, suggesting a mild Hook effect. Princeton - 104176

[0132] Maximal editing was observed at 12.5 nM (68.6%) and 25 nM was used in all further experiments.

[0133] The effect of time on RNA editing was also investigated. Editing was assessed after 8,16,24,48, and 72 h post-transfection. After 8 h , editing reached 38.2% at the major edit site, and editing at other positions was low (<10%). Editing increased to 51.9% after 16 h , with editing at secondary edit sites increasing but < 15%. Maximal editing was found at 24 h (68.9%) and then slowly decreasing editing levels thereafter (61.9% at 48 h; 59.8% at 72 h). Interestingly, in contrast to the major site, the editing of other A sites continued to increase over time. Levels increased from 8.3% at 8 h to 35.0% at 72 h for the second-most highly edited site, which is the first A' upstream of the ASO binding region. Taken together, the data shows that ASO-directed proximity-mediated RNA editing can enable selective and high-efficiency RNA editing events with evolved TadAderived deaminase enzy mes.

[0134] Example 2 - ASO-directed RNA editing with a panel of RNA / DNA editing enzymes

[0135] The editing behavior of different deaminase enzymes with Halo-Spinraza 1 was also evaluated. A panel of adenine and cytosine editors was selected, including the RNA deaminases ADAR2 and APOBEC1, and the evolved DNA base editors TadA8r, TadA7.10, TadDE, and TadCBE. Inducible Flp-In TREx 293 cells containing HaloTag-editor fusions were generated and Halo-Spinraza 1 was co-transfected together with a plasmid mRNA reporter as described above for TadA8.20. Conjugation was validated between Halo-Spinraza 1 and each HaloTag- deaminase fusion protein by Western blot. Editing activity was quantified by Sanger sequencing following RT-PCR.

[0136] For RNA isolation and Sanger sequencing analysis, RNA isolation was performed with TRIzol (Thermo Fisher), TRI Reagent (Molecular Research Center), or RNAzol RT (Molecular Research Center) following the manufacturer's protocol but with 250-500 pL reagent per sample and proportionally scaled volumes of the other reagents utilized in the procedure. After isopropanol precipitation and washing with 75% ethanol, the RNA pellets were dissolved in DEPC-treated water, DNase buffer, and treated with DNase I (NEB) or TURBO DNase (Thermo Fisher) for 30 min at 37 °Cbefore ethanol precipitation at -80 °C. Samples were then pelleted by7centrifugation, washed with 70% ethanol, and resuspended in DEPC water. The concentration was then determined by Nanodrop (Thermo Fisher). 1-2 pg of total RNA was used in the reverse transcription reactions using SuperScript II (Invitrogen) or in-house-purified MMLV (Addgene #153312) following the Invitrogen SuperScript II protocol. Generally. 4 pL of crude reverse transcription reaction was used as the template in a 25 pL PCR reaction Princeton - 104176 employing Taq DNA polymerase (NEB). Products were analyzed by agarose gel electrophoresis and gel purified before Sanger sequencing (Genewiz - Azenta Life Sciences). EditR (moriaritylab. shinyapps. io / editr_vl0 / ) was used to quantify editing stoichiometry from Sanger sequencing data, and a student's t-test was used to assess statistical significance.

[0137] Among evolved TadA-derived adenine base editors, similar patterns of A-to-I editing were observed, with 32.5 - 51.6% A-to-I editing at the major editing site (3' to the ASO binding site) and secondary editing < 11.9% at all other sites. Whereas editing levels at the major site were lower for these enzymes than for TadA8.20, the prevalence of secondary edits was also reduced, indicating a trade-off between selectivity and activity7. To further demonstrate this, the effect of time on RNA editing with TadA7.10, a less active precursor to TadA8.20, was evaluated. Unlike TadA8.20, editing at the primary site with TadA7.10 continued to increase until 72 h , achieving 48.8% A-to-I conversion, with editing at other sites < 10%. This data also suggests that editor- ASO fusions remain active in cells up to 72 h and shows how editing efficiency and selectivity with TadA-derived adenine editors can be readily tuned by choosing among available enzyme variants.

[0138] TadA-denved base editors reported to catalyze C-to-U editing on DNA were also explored. For TadDE, a dual function editor that can perform A-to-I and C-to-U editing on DNA, C -to-U editing was unable to be observed, whereas A-to-I editing activity7was similar to TadA7.10. Similarly, C-to-U editing for TadCBE which was specifically evolved for C-to- U editing on DNA, suggesting that DNA C-to-U activity among TadA-derived editors is not readily transferable to RNA. Additionally, the catalytic domain from hADAR2 containing the hyperactive E488Q mutation (hADAR E488Qcd) and rAPOBECl (APOBEC) were evaluated as these enzymes have been used for targeted RNA editing and proximity-based RNA-protein interaction analysis. As expected, no significant editing was observed with hADAR E488Qcd, likely due to its requirement for dsRNA substrates. For APOBEC, 30.9% C-to-U editing was observed immediately 3' of the ASO binding site, indicating the promise of proximity -directed APOBEC editing, consistent with prior studies. These data demonstrate that whereas localized ASO-directed RNA editing can be achieved with both native APOBEC and evolved TadA- derived enzymes, editing efficiency is considerably higher with TadA-derived deaminases, with minimal compromise in selectivity.

[0139] Example 3 - Preferred editing substrate motif for TadA8.20

[0140] It was hypothesized that editing selectivity in the disclosed system was likely a product of both proximity-mediated ASO recruitment and substrate specificity of the editing enzyme. Princeton - 104176

[0141] WT E. coli TadA selectively modifies the sequence UACG within the anticodon loop of IRNAA,S2. and evolved TadA-based editors have been reported to preferentially deaminate TA / UA sequence motifs. Within the example mRNA reporter plasmid, the major editing site modified by TadA-based editors occurs within the 5 mer sequence GTACA, and therefore, it was investigated how varying this sequence affects A-to-I conversion.

[0142] Editing reporter variants were constructed containing single, double, or triple mutations. Referring to FIG. 3B. 5 mer sequences centering around the target site 308 were created, with "A" as the nucleotide at the target site. Then, editing was evaluated with Halo- Spinraza 1 and TadA8.20HaloTag. The data showed a preference for editing at TA dinucleotides, as all three 5 mer sequences with this motif (i.e., ATACA, CTACA, and GTATG) were edited between 59.7-69.4%, similar to the parent GTACA sequence . Mutation to GA, AA, or CA (i.e., GGAGG, GAACA, GCACA) reduced editing levels to between 22.8% - 34.6%. The data suggests that mutation of bases 5' or 3' to the TA dinucleotide has minimal impact on editing within this substrate.

[0143] Example 4 - ASO-directed RNA editing of endogenous mRNA

[0144] Having demonstrated editing on a plasmid-based reporter mRNA, editing on endogenous mRNA was then evaluated . ACTB was chosen as a model system and chloroalkane-modified ASOs (ACTB ASO 1-6) were synthesized (See Table 1) targeting 6 sites across the coding sequence (CDS) and 3'-UTR.

[0145] ASOs were designed with similar thermodynamic characteristics to HaloSpinraza 1 (see Table 2) and targeted a "TA" motif 7-9 nt downstream of the ASO binding site.

[0146] Table 2. ASO thermodynamics. Data was generated from the Oligoscreen web server (rna.urmc.rochester.edu / RNAstructureWeb / Servers / oligoscreen / oligoscreen.html) using the RNA setting and 310.15 K temperature. The overall duplex energy was determined from the IntaRNA web server (ma.informatik.uni-freiburg.de / IntaRNA / Input.jsp). Princeton - 104176

[0147] ACTB ASOs were then transfected into the TadA8.20-HaloTag-expressing cell line and measured editing using ACTB-specific RT-PCR primers (see Table 3) and Sanger sequencing. Table 3. Primer Sequences Princeton - 104176 Princeton - 104176 Princeton - 104176

[0148] ASO-dependent editing was observed for all six ASO constructs tested, with the major edit site occurring 3' of the ASO binding site, consistent with tethering of TadA8.20-HaloTag at the 5' end of the ASO. The two ASOs that bind to the 3'-UTR (ACTB ASO 1 and ACTB ASO 2) resulted in 36-56% editing at adjacent A residues, comparable to editing efficiency observed in our plasmid-based reporter mRNA. Of these, ACTB ASO 1, which binds a site used in a previous ADAR study, edited multiple A residues 3' to the ASO binding site at high efficiency (7-56%). Editing was also observed at these sites in the absence of ASO, suggesting they are edited by TadA8.20-HaloTag independent of ASO-directed recruitment. Mitigation of promiscuous editing was enabled by applying ACTB ASO 1 in the TadA7.10-HaloTag expressing cell line, however, target site editing efficiency was also modestly decreased to 23.0%. In contrast, for ACTB ASO 2. which also binds in the 3'-UTR. a single major editing site was observed as being modified at 36.7% and minimal editing in the absence of ASO. It Princeton - 104176 was speculated that the absence of off-target editing with ACTB ASO2 likely stems from a lack of additional substrate sites in this region, but may also depend upon ASO binding ability. Minor editing (<12%) does occur 5' upstream of the ACTB ASO 2 binding site at other TA dinucleotide motifs.

[0149] ASOs binding the CDS generally showed lower levels of editing than 3'-UTR-binding ASOs, consistent with previous reports. ACTB ASO 3 produced 25.8% editing at an adjacent GTATG sequence with minor editing elsewhere. Similarly. ACTB ASO 4 directed 26.8% editing at an adjacent CTACG sequence with one other major edit. ACTB ASO 5 produced 17.8% editing at the adjacent CTATG site with several low-level but statistically significant editing events downstream. ASO 6 produced 20.4% editing at the adjacent GTACG sequence with one major edit upstream of the ASO binding site (16.7%) and one major edit further downstream (20.1%) along with other low- level edits. In the cases of ACTB ASO 4 and ACTB ASO 6 targets, several A-to-G mutations were observed that were independent of ASO and found in WT cells lacking TadA8.20- HaloTag. These are likely due to the amplification of related actin transcript isoforms. Taken together, the results demonstrate efficient proximity-based ASO-directed editing at multiple sites on an endogenous mRNA and indicate that single-site editing can be achieved through judicious selection of ASO binding site and surrounding sequence.

[0150] Example 5 - Comparison against CRISPR Cas-directed RNA editing methods

[0151] To benchmark the system, it was compared against CRISPR / Cas-based editing strategies, as these methods are compatible with multiple different RNA editing / modifying enzymes, and have been used with TadA-derived base editors for RNA editing. Editing was investigated at the ACTB ASO 2 target site using REPAIRv2, an ADAR method, and DECOR, which uses TadA8e (a TadA-derived adenine base editor with activity intermediate between TadA7.10 and TadA8.20). CRISPR sgRNAs were chosen for each method following literature precedent.

[0152] HEK293T cells were seeded on a poly -L-ly sine-coated 12 -well plate at 1.3 x 105cells / well 24 h before transfection. For REPAIRv2, 750 ng of editor expression plasmid (Addgene #103871) and 375 ng of gRNA-expressing plasmid was diluted to 125 pL in OptiMEM and mixed with 2.5 pL Lipofectamine 2000 diluted to 125 pL separately in OptiMEM. Complexes w ere incubated for 20 min and added to fresh 750 pL media. In the case of PTEN Q245X-directed editing (see, e.g., Example 9), 200 ng of PTEN Q245X plasmid was also included in the transfection. The conditions were the same for DECOR, except that 500 Princeton - 104176 ng DECOR expression plasmid (Addgene #219545) and 750 ng gRNAexpressing plasmid were used. Cells were harvested after 24 h for RNA isolation and sequencing analysis as described above.

[0153] For REPAIRv2, the sgRNA contained an A:C mismatch over the target site positioned 20 nt from the 3'-end of the sgRNA; using this construct only 7.4% A-to-I editing was observed, 5-fold lower than ASO-directed editing with TadA8.20.

[0154] Using the DECOR method, a CRISPR sgRNA was designed positioning the target editing site 15 nt from the 5'-end of the sgRNA and also observed poor editing (9.0%) . Whereas REPAIRv2 and DECOR have been validated on other mRNA substrates, the results here show that these methods are context-dependent (or require empirical optimization of sgRNA sequence), and highlights an mRNA transcript context where the disclosed ASO-guided approach performs more efficiently. The DECOR system also features dual SV40 nuclear localization sequences (NLS), reported to result in higher mRNA editing. An analogous TadA8.20-NLS-HaloTag-NLS construct was tested via transient transfection with ACTB ASO 2 and measured 27.4% editing, slightly lower than editing without the dual NLS sequences.

[0155] Example 6 - RNA editing via cellular delivery of a recombinant deaminase-ASO

[0156] To extend the application of the disclosed method to cells that do not require genetic engineering or transgene delivery and reduce off-target editing, it was hypothesized that cellular delivery of a recombinant TadA8.20-ASO oligo-protein complex formed in the test tube could mediate site specific RNA editing. TadA8.20-ASO delivery by electroporation or liposome transfection was explored.

[0157] Recombinant Tad8.20-HaloTag protein was first generated through heterologous expression in E. coli and demonstrated efficient in vitro labeling with a choroalkane-modified ASO. TadA8.20-HaloTag-ASO conjugates were then formed with ACTB ASO 1, 2, or 3, as these ASOs performed well in the Tad8.20-HaloTag-expressing cell line and delivered the oligo-protein conjugates into WT HEK293T cells.

[0158] Synthesis of chloroalkane-modified antisense oligonucleotides (Halo-ASOs). NH2- modified ASOs were synthesized on an ABI 394 oligonucleotide synthesizer (Applied Biosystems) using standard RNA synthesis conditions except for the phosphorami dite coupling time, which was reduced from 12 min to 6 min as recommended for 2'-M0E phosphoramidites. All oligo synthesis reagents were purchased from Glen Research. ASOs were synthesized on Glen Unysupport 500 (Glen Research 20-5140-41) CPG resin unless otherwise indicated Princeton - 104176

[0159] (Table 1), and a primary amine handle was installed on the 5' end of the ASOs with the 5'- amino- 5 modifier (Glen Research 10-1905-90) or internally with the amino modifier C2-dT (Glen Research 10-1037-90). Prior to coupling, the final product was detritylated manually by flowing 3% dichloroacetic acid through the column for 2 min. Oligos were then rinsed with ACN and dried under N2. Oligos were cleaved from the CPG resin and deprotected with a 1 :1 mixture of NHrOH and aq. 40% MeNH2 (AMA) for 1 h at 65°C, following the manufacturer's recommendations. The supernatant was collected and dried by speedvac. The material was redissolved in 0. 1 M triethylammonium acetate (TEAA) and purified via reverse-phase HPLC (0-30% acetonitrile / 0.1 M TEAA gradient). The product was lyophilized and redissolved in DEPC- H2O. 5'- NEb-modified ASOs were coupled to NHS-Halo ligand as follows: to approximately 200 nmol of ASO was added buffer containing 0.5 M MOPS pH 8.0, 0.5 M NaCl and 0. 1 M NHS-Halo ligand in DMSO to give a final concentration of 0.2 M MOPS pH 8.0, 0.2 M NaCl, and 10 mM NHS-halo ligand. Reactions were carried out for 2 hr at room temperature, diluted with 0. 1 M TEAA, purified by RP-HPLC (0-30% acetonitrile in 0. 1 M TEAA), and lyophilized. Purified product was dissolved in DEPC-H2O, analyzed by ESI-TOF MS (Agilent 6545XT), and stored at -20 °C.

[0160] TadA8.20-HaloTag-HA expression and purification. pET28a-6xHis-TadA8.20- HaloTag-HA was transformed into E. coli BL21 and expressed at 16 °C with 0.1 mM isopropyl-B-D-thiogalactopyranoside (IPTG) for 20 hr . Lysis by sonication was performed in buffer containing 50 mM Tris pH 7.5, 300 mM NaCl, 10 mM imidazole, 10% glycerol, 1 mM DTT, 1 mM PMSF (added fresh), and protease inhibitor tablet (Roche, added fresh). Lysate was clarified at 14,000 g for 30 min. Clarified lysate was incubated with 2 ml of fresh HisPur cobalt resin (Thermo Fisher) for 1 hr at 4 °C, washed with lysis buffer, and eluted in the same buffer but with 250 mM imidazole. Fractions were measured by Bradford assay and analyzed by SDS-PAGE, pooled, and dialyzed overnight against 2 L of 50 mM Tris pH 7.5, 100 mM NaCl, 10% glycerol, and 1 mM DTT. Dialyzed protein was then purified via FPLC (GE Healthcare) using a HiTrap QHP 1 mL anion exchange column (100 mM to 1 M NaCl in 50 mM Tris pH 7.5. 10% glycerol, and 1 mM DTT) over 30 CV at 0.5 mL / min. Fractions were analyzed by SDS-PAGE, pooled, and concentrated in 50 mM Tris pH 7.5, 100 mM NaCl, 10% glycerol, 1 mM DTT to > 1 mg / mL, aliquoted at 58.5 pM as determined by Bradford assay, flash frozen, and stored at -80 °C.

[0161] RNA editing with Halo-ASOs in deaminase-HaloTag-expressing Flp-In 293 cells. Culture plates were first coated with poly-L-lysine by diluting 10X poly-L-lysine (Sigma) into Princeton - 104176

[0162] IX DPBS (Gibco Life Technologies) and incubating culture wells for 5 min at RT, aspirating, and washing the wells with 1 X DPBS. Wells were allowed to dry for 30 min before plating cells. Editor-HaloTag-expressing Fip-In 293 cells were seeded in a 24 -well plate at 7 x 104cells in 500 pL media. After 24 h, 1 pg / mL tetracycline was added to induce protein expression. After 24 hours, the media was changed, and fresh tetracycline was supplied. Cells were transfected with HaloASO (25 nM final concentration in media) together with pcDNA5 plasmid encoding the mRNAreporter (200 ng) unless otherwise indicated. Transfection mixtures were prepared by diluting reagents in OptiMEM with 1.25 pL Lipofectamine 2000 diluted in a separate tube. The solutions were mixed and incubated for 20 min before adding to the wells. Cells were harvested 24 h after transfection unless otherwise indicated.

[0163] RNA editing by lipofection of the TadA8.20-HaloTag-ASO complex. HEK293T or HeLa cells were first plated in a poly-L-ly sine-coated 24 -well plate at 7.5 x 104cells / well in 500 pL 24 h before transfection. Shortly before transfection, purified TadA8.20-HaloTag-HA protein was incubated with an equimolar (10 pM) concentration of Halo-ASO in IX PBS at 37 °C for 30 min. The crude reaction was diluted directly in OptiMEM and 3 pL of Lipofectamine 2000, Lipofectamine 3000, or RNAiMAX was diluted in OptiMEM separately. The solutions were mixed (50 pL total volume) and incubated for 20 min before adding to the wells without changing the media to achieve a final enzyme- ASO complex concentration of 25 nM in 550 pL media. Cells were harvested after 24 h for RNA analysis.

[0164] RNA editing by electroporation of the TadA8.20-HaloTag-ASO complex. HEK293T cells were grown to 80% confluency before being trypsinized, pelleted, and resuspended in OptiMEM to a concentration of 5.556 x 106cells / mL. TadA8.20-HaloTag-ASO complex was prepared freshly as described above, and 10 pL of the reaction was mixed with 90 pL of cell suspension to achieve a 1 pM enzyme- ASO complex concentration in 100 pL with 5 x 105cells. The mixture was transferred to a 2 mm electroporation cuvette (Universal Medical) and electroporated (100 - 160 V, 10 ms or 20 ms time constant) using a BTX ECM 830 device. Cells were allowed to recover for 10 min at room temperature before mixing with 1.9 mL culture media and plating in a 6 -well plate. Cells were harvested after 24 h for RNA analysis.

[0165] Different concentrations (25 - 100 nM) of TadA8.20-HaloTag-ACTB ASO 1 complex ("RNP complex") were tested with different commercial transfection reagents and electroporation and found similar editing patterns on ACTB mRNA as in previous experiments with TadA8.20-HaloTag expressing cells transfected with ACTB ASO 1. Editing at the target position measured 24.0%, about half of what was observed in TadA8.20-HaloTag-expressing cells. Major differences in editing were not observed across the transfection conditions assayed, Princeton - 104176 and further experiments were performed using Lipofectamine 2000 and 25 nM deaminase- ASO complex. For the ACTB ASO 2 conjugate, 49.7% editing was measured at the primary editing site with minimal secondary editing; the efficiency and specificity, in this case, were better than in the TadA8.20-HaloTag-expressing cell line. Targeting the CDS with the TadA8.20-HaloTag-ACTB ASO 3 complex yielded 8.3% editing at the primary site, lower than editing when the ASO was transfected into TadA8.20-HaloTag-expressing cells.

[0166] The efficacy of RNP delivery and editing in HeLa cells was also evaluated to establish generality. Targeted editing with both ACTB ASO 1 RNP complex and ACTB ASO 2 RNP complex was observed, with 17.8% and 24.8% A-to-I conversion, respectively, at the target position. For both ACTB RNP complexes, editing efficiency was lower in HeLa cells than in HEK293T. which may be attributable to lower transfection efficiency in HeLa cells. Collectively, these data demonstrate editing of endogenous mRNA by direct cellular delivery of a recombinant deaminase- ASO conjugate.

[0167] Example 7 - RNP delivery reduces transcriptome-wide off-target editing.

[0168] It was next investigated whether RNP delivery reduced transcriptome-wide off-target editing as compared to a genetically encoded TadA8.20-HaloTag construct. RNA-seq libraries for Illumina sequencing from WT HEK293T cells treated with the ACTB ASO 2 RNP complex were prepared and analyzed A-to-G mutations (comparing against untreated cells) following literature precedent.

[0169] RNA-seq library preparation. HEK293T cells were plated on poly-L-lysine-coated 6 - well plate at 2.75 x 105cells / well 24 h before transfection. RNA-protein complexes were formed in vitro as described above. The crude reaction was diluted directly in OptiMEM, and 12 pL of Lipofectamine 2000 was diluted in OptiMEM separately. The solutions were mixed (200 pL total volume) and incubated for 20 min before adding to the wells without changing the media to achieve a final enzyme- ASO complex concentration of 25 nM in 2.2 rnL media. RNA was isolated after 24 h using TRI Reagent (Molecular Research Center) and treated with DNase I (New England Biolabs) prior to mRNA isolation via polyA pulldown with oligo dT(25) (SEQ ID NO: 57) magnetic beads (New England Biolabs). 50 ng of polyA mRNA was subject to library preparation using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England Biolabs). Quality control w as determined by gel electrophoresis and Bioanalyzer. The libraries were sequenced on an Illumina NovaSeq 6000 instrument (paired-end; R180 nt, R2 250 nt). Princeton - 104176

[0170] A-to-I editing was also analyzed in cells transfected with a plasmid encoding TadA8.20-HaloTag and cells treated with a Halo-Spinraza 1 RNP complex (as a non-targeting control). For HaloSpinraza 1 RNP transfection, A-to-I edits were not detected on SMN2 mRNA; however, the transcriptome-wide analysis contained no reads mapped to the intronic region in SMN2 premRNA where Spinraza binds, likely due to the short lifetime of the pre- mRNA species and the polyA-enrichment step used to prepare mRNA for RNA-seq library preparation.

[0171] It was found that RNP delivery resulted in ~ 100-1,000-fold reduction in transcriptome-wide off-target editing events as compared to plasmid-based expression of TadA8.20-HaloTag. Transfection of cells with plasmid-encoded TadA8.20-HaloTag resulted in an average of 27,439 transcriptome-wide A-to-I edits between two independent biological replicates. Although these examples did not co-transfect a chloroalkane-containing targeting ASO in this condition, it is estimated that only 40% of the HaloTag-editor is covalently conjugated to Halo-ASO in cells. In contrast, lipofection of Halo-Spinraza 1 RNP and ACTB ASO 2 RNP resulted in 36 and 144 transcriptome- wide editing events, respectively.

[0172] Comparative analysis of transcriptome-wide off-target editing using published RNA- seq data from the RNA editing methods DECOR, REPAIRv2, and SNAP-ADAR was also performed. The transcriptome-wide off-target profiles for ACTB ASO 2 RNP and Halo- Spinraza 1 RNP are most similar to SNAP-ADAR2 (SA2) (78 edit sites) and superior to DECOR. SNAP-ADAR2(E488Q) (SA2Q). and REPAIRv2, which contained 17.036. 1,979, and 673 edit sites, respectively. Importantly, the major ACTB ASO 2 editing site in the 3'-UTR of ACTB observed by RT-PCR and Sanger sequencing was comparably edited at 49.7% in the transcriptome-wide analysis. The substrate motif for TadA8.20 was determined by analyzing 27,439 edit sites observed in the TadA8.20-HaloTag plasmid transfection condition, which revealed a preferred "TA" dinucleotide motif, corroborating previous findings for the substrate preference of TadA variants and the disclosed findings with different substrate motifs.

[0173] The editing stoichiometry of sites detected in the ACTB ASO 2 RNP condition were further compared with those detected in SA2, as SA2 gave a similar number of transcriptomewide editing events. Although the number of off-target sites is similar, the sites detected in S A2 are modified at higher stoichiometry compared to those in ACTB ASO 2 RNP. Editing stoichiometry for ACTB ASO 2 RNP was also low er across most sites compared to TadA8.20- HaloTag.

[0174] The transcript location distribution of edit sites detected in ACTB ASO 2 RNP and TadA8.20-HaloTag plasmid conditions was also evaluated. The majority of edit sites occurred Princeton - 104176 in UTRs and CDSs, with a minor fraction occurring in stop and start codons. Additionally, ACTB ASO 2 RNP gave more sites that occurred in UTRs and fewer in CDSs relative to TadA8.20-HaloTag. Altogether, these data highlight the potential of RNP delivery with covalent deaminase-ASO conjugates as a strategy to achieve efficient, site-specific RNA editing in unmodified cells without off-target transcriptome-wide editing.

[0175] Example 8 - Alternative ASO architectures increase RNA editing selectivity)

[0176] Whereas RNP delivery dramatically reduced transcriptome-wide off-target editing, transcript-specific secondary editing was largely unaffected. Alternative ASO architectures were, therefore, explored to further restrict TadA8.20 activity to the target site of interest. Since mRNA editing was biased towards the 5'-end of the ASO (where the editor is tethered), tethering the editing enzyme to a nucleotide located in the middle of the ASO was evaluated.

[0177] A 2'-MOE-modified phosphodiester ASO was synthesized with the same nucleotide sequence as Halo-Spinraza 1 but the chloroalkane Halo ligand was installed at the C5 position of an internal dT residue (see FIG. 4A) located 10 nt from the 5'-end of the ASO ("Halo- Spinraza 2"). Editing was evaluated on the same mRNA reporter used in previous examples, and 49.4% editing was observed at the major target position (as compared to 68.6% with the 5'-Halo-modified Spinraza ASO) but with considerably lower secondary editing at other positions on the transcript.

[0178] Building on the improvements conferred by this simple modification in ASO-enzyme conjugate design, a "bulge-forming" ASO architecture was tested in which ASO-mRNA binding is predicted to generate a ssRNA "bulge" opposite the editing enzyme tethered at the middle of the ASO. See FIG. 5. It was reasoned that editing would be biased towards the ssRNA bulge, whereas adjacent bases would be protected by base pairing with the ASO. ASOs were designed with discontinuous binding regions intended to position the target mRNA base within a ssRNA loop varying between 4 to 9 nt, and with the Halo ligand at the middle of the ASO sequence ("Bulge ASO 1-3"). Gratifyingly, editing using the bulge-forming ASOs was efficient and specific for the target base, with a similar profile of on-target to off-target activity observed for the 6 nt and 9 nt bulge architectures as observed for a continuous ASO design with the editor linked at the middle.

[0179] It w as further evaluated whether alternative ASO architectures w ere compatible with other RNA deaminases in addition to TadA8.20. For TadA7.10, 16.6% editing was measured with Halo-Spinraza 2 at the target position with only one low-level statistically significant secondary edit. Editing with Bulge ASO 2 resulted in 30.2% A-to-I conversion at the target Princeton - 104176 site, also with only one low-level statistically significant secondary edit, similar to results with the 5'-Halo-ligand-modified Halo-Spinraza 1. Thus. TadA7.10 editing is compatible with internal ASO conjugation and the bulgeforming ASO architecture; however, TadA7.10 demonstrated high target site selectivity with the original 5'-Halo ASO design, and neither ASO design strategy further improved editing efficiency or selectivity7in the reporter system.

[0180] For APOBEC1, a C-to-U deaminase, it was found that editing at the target base directed by HaloSpinraza 2 was low efficiency (7.4%) as compared to Halo-Spinraza 1 (30.9% conversion). In contrast, efficient APOBEC1 -mediated C-to-U editing was found when directed by bulge-forming ASOs with fewer secondary7editing sites. In particular, Bulge ASO 2 and Bulge ASO 3, which induce 6 nt and 9 nt bulges in the mRNA substrate, respectively, resulted in 21-30% editing at the target site. Editing with Bulge ASO 1, which generates a smaller 4 nt bulge, was less efficient with 10% conversion. Importantly, the two secondary edits observed with Halo-Spinraza 1 were eliminated with the bulge-forming ASOs while retaining on-target efficiency. Collectively, these data indicate that off-target editing activity of diverse A-to-I and C-to-U RNA deaminases that preferentially modify ssRNA can be controlled through rational ASO design and modification without compromising on-target editing at the primary site.

[0181] Example 9 - Correction of a disease-causing mutation with ASO-direcled RNA editing Finally, it was tested whether the disclosed techniques could correct nonsense mutations in a disease-associated mRNA and restore protein expression. Inactivating mutations in PTEN are common in tumors and other germline diseases such as Cowden's syndrome. The Q245X mutation in PTEN was selected, which yields a truncated and inactive version of the tumor suppressor, as a candidate mutation to test in the disclosed system. PTEN Q245X has been subjected to small-molecule drug readthrough experiments with inconsistent results ranging from < 1% to 11.7% restoration.

[0182] The mutant PTEN gene was cloned into a plasmid and co-transfected it into TadA8.20- HaloTag-expressing cells together with either bulge-forming ASOs or a 5'-Halo-modified ASO targeting the mutation site. Editing was measured 24 h after transfection by RT-PCR and Sanger sequencing. The 5'-Halo ASO design (PTEN ASO 1) gave only 9.5% editing. Using a 7 nt bulge-forming ASO (PTEN ASO 2), 33.4% editing was achieved at the target site, with up to 12.9% editing at other sites within the gene. PTEN ASO 2 with TadA7. 10-HaloTag gave only 13.4% editing but eliminated most other statistically significant edit sites. Increasing the length of the ASO binding regions in PTEN ASO 2 from 22 nt to 26 nt generated PTEN ASO Princeton - 104176

[0183] 3, which increased editing to 36.4% at the target position and up to 18.2% at secondary sites. An ASO construct that reduces the bulge size from 7 nt to 5 nt (PTEN ASO 4) modestly decreased editing efficiency to 23.4%. Importantly, for all PTEN ASOs, the two highest secondary edits at nucleotide positions 30 and 38 relative to the PTEN ASO 1 binding site are conservative mutations, yielding lie to Vai and Vai to Vai, respectively. These editing approaches rescued PTEN expression by as much as 23.5% of the wild-type control (after 24 h ) as evaluated by Western blot.

[0184] The abilities of REPAIRv2 and DECOR to edit PTEN Q245X were also evaluated. Both methods were less efficient than the disclosed approach at primary site editing, with varying degrees of off-target editing. REPAIRv2 gave 22.6% on-target editing with minimal off-target editing at other positions within the gene. DECOR gave 17.3% on-target editing with multiple secondary edit sites. Therefore, the disclosed approach performs favorably as compared to CRISPR / Cas methods, with superior on-target editing efficiency.

[0185] A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the disclosure. Accordingly, other implementations are within the scope of the following claims.

Claims

Princeton - 104176CLAIMS1. A method of targeted RNA editing, comprising: providing a protein-antisense oligonucleotide (ASO) conjugate comprising a first protein covalently linked to an ASO; and delivering the protein-ASO conjugate to one or more cells to edit a target mRNA sequence.

2. The method of claim 1, wherein providing the protein-ASO conjugate comprises: synthesizing a modified ASO comprising a reactive group; producing a recombinant fusion protein comprising the first protein fused to a selflabeling protein tag; and forming the protein-ASO conjugate by incubating the modified ASO with the recombinant fusion protein.

3. The method of claim 2, wherein synthesizing the modified ASO includes solid phase oligonucleotide synthesis.

4. The method of claim 2 or 3. wherein producing the recombinant fusion protein includes using heterologous expression in a host cell.

5. The method of claim 4, wherein the host cell is E. coli.

6. The method of claim 4, wherein the host cell is an organism other than E. coll.

7. The method of any one of claims 2-6. wherein the self-labeling protein tag includes a modified haloalkane dehalogenase enzyme or a modified mammalian O6-alkylguanine-DNA- alkyltransferase (AGT) enzyme.

8. The method of any one of claims 2-7, further comprising allowing the recombinant fusion protein to recode at least some genetic information present in mRNA.

9. The method of any one of claims 2-8, further comprising allowing the recombinant fusion protein to regulate a disease-associated mRNA transcript.

10. The method of any one of claims 1-9, wherein the first protein is a deaminase enzyme.

11. The method of any one of claims 1-10, wherein the protein-ASO conjugate is a proteinbulge-forming antisense oligonucleotide (ASO) (Bulge ASO) conjugate.

12. The method of any one of claims 1-11, wherein providing the protein-ASO conjugate includes protein trans-splicing utilizing split inteins.

13. The method of any one of claims 1-12, wherein providing the protein-ASO conjugate includes chemically conjugating the first protein to the ASO utilizing compatible reactive groups present on the first protein and the ASO.Princeton - 10417614. The method of claim 13, wherein the ASO has been modified to include a first compatible reactive group, and / or wherein the first protein has been modified through site-directed mutagenesis, protein semi-synthesis, or unnatural amino acid mutagenesis.

15. The method of claim 14, wherein the first compatible reactive group is an azide, an alkyne, a tetrazine, an alkene, an amine, a thiol, a hydrazide, an aminooxy, an aldehyde, a ketone, an isothiocyanate, a carboxylate, a maleimide, an NHS ester, or a Gly-Gly-Gly peptide.

16. The method of claim 14, wherein the first compatible reactive group is a functional group for bio-conjugation or bio-orthogonal chemistry'.

17. A method of delivery' of an antisense oligo (ASO) conjugate and a gene encoding a recombinant fusion protein to a cell, comprising: transfecting the ASO conjugate into a cell via lipid nanoparticles (LNP); and introducing a recombinant fusion protein by: transfecting a plasmid vector tag into the cell via lipid nanoparticles (LNP), where the plasmid vector tag is a plasmid vector encoding a recombinant fusion protein comprising a deaminase enzyme fused to a self-labelling protein tag; delivering a gene encoding the recombinant fusion protein into the cell using modified mRNA; or delivering the gene encoding the recombinant fusion protein into the cell using a viral delivery vector.

18. The method of claim 17, wherein the viral delivery vector is an adeno-associated virus (AAV).

19. The method of claim 17 or 18, wherein the self-labelling protein tag includes a modified haloalkane dehalogenase enzyme or a modified mammalian O6-alkylguanine-DNA- alkyltransferase (AGT) enzy me.

20. The method of any one of claims 17-19, further comprising allowing an ASO of the ASO conjugate to regulate a disease-associated mRNA transcript.

21. The method of any one of claims 17-20. wherein an ASO of the ASO conjugate is a bulge-forming antisense oligonucleotide (ASO) (Bulge ASO).

22. A bulge-forming antisense oligonucleotide (ASO) (Bulge ASO), comprising: a nucleotide sequence comprising discontinuous binding regions configured to position a target mRNA base within a ssRNA loop between 4 nt and 9 nt in length; andPrinceton - 104176 a protein tag coupled to the nucleotide sequence at an intermediate location within the nucleotide sequence.

23. The Bulge ASO of claim 22, wherein each discontinuous binding region is, independently, at least 5 nucleotides in length.

24. The Bulge ASO of claim 23, wherein each discontinuous binding region is, independently, 6-30 nucleotides in length.

25. The Bulge ASO of any one of claims 22-24. wherein the protein tag is a HaloTag or SnapTag.

26. A method of correcting a genetic mutation, comprising: providing a bulge-forming antisense oligonucleotide (ASO) (Bulge ASO) coupled to a protein tag: allowing discontinuous binding regions of the Bulge ASO to bind to a target sequence such that a target mRNA base is positioned within a ssRNA loop between 4 nt and 9 nt in length, the target mRNA base comprising a genetic mutation; and editing the target mRNA base.

Images