ADENINE DNA BASE EDITOR VARIANTS WITH REDUCED OFF-TARGET RNA EDITING

Engineered adenine base editor (ABE) variants with reduced RNA editing activity, and methods of using the same.

1. An adenine base editor (ABE) variant comprising an adenosine deaminase and a programmable DNA binding domain, the adenosine deaminase comprising one or more E. coli TadA monomers, wherein at least one of the one or more E. coli TadA monomers comprise one or more mutations that decrease RNA editing activity while preserving DNA editing activity.

2. The ABE variant of claim 1, wherein the adenosine deaminase comprises a wild type or engineered E. coli TadA monomer, or a variation of homo- or heterodimers thereof.

3. The ABE variant of claim 1, wherein the adenosine deaminase comprises ABE 0.1, ABE 0.2, ABE 1.1, ABE 1.2, ABE 2.1, ABE 2.2, ABE 2.3, ABE 2.4, ABE 2.5, ABE 2.6, ABE 2.7, ABE 2.8, ABE 2.9, ABE 2.10, ABE 2.11, ABE 2.12, ABE 3.1, ABE 3.2, ABE 3.3, ABE 3.4, ABE 3.5, ABE 3.6, ABE 3.7, ABE 3.8, ABE 4.1, ABE 4.2, ABE 4.3, ABE 5.1, ABE 5.2, ABE 5.3, ABE 5.4, ABE 5.5, ABE 5.6, ABE 5.7, ABE 5.8, ABE 5.9, ABE 5.10, ABE 5.11, ABE 5.12, ABE 5.13, ABE 5.14, ABE 6.1, ABE 6.2, ABE 6.3, ABE 6.4, ABE 6.5, ABE 6.6, ABE 7.1, ABE 7.2, ABE 7.3, ABE 7.4, ABE 7.5, ABE 7.6, ABE 7.7, ABE 7.8, ABE 7.9, ABE 7.10, or ABEmax.

4. The ABE variant of claim 1, wherein the one or more mutations comprise one or more mutations at amino acid positions that correspond to residues of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: [[24]]34) as listed in Table A.

5. The ABE variant of claim 4, wherein the one or more mutations are at amino acid positions that correspond to residues Y10, W11, R13, T17, K20, R21, R23, E25, R26, A48, 149, A56, A58, Q71, N72, R74, D77, V82, V106, R107, N108, A109, K110,T111, H122, Y123, H128, R129, A138, D139, E140, A142, A143, F148, and/or R150, R153, V155 of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: [[24]]34).

6. The ABE variant of claim 5, wherein the one or more mutations comprise mutations that correspond to Y10A, W11A, R13A, T17A, K20A, R21A, R23A, R23W, E25A, R26A, A48G, I49A, A56G, A58G, Q71A, N72A, R74A, D77A, V82G, V106G, V106W, R107A, N108A, A109G, A109W, K110A, T111A, H122A, Y123A, H128A, R129A, A138W, A138G, D139A, E140A, A142W, A142G, A143G, F148A, R150A, R153A, V155G, and/or V155W of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: [[24]]34).

7. The ABE variant of claim 6, wherein the at least one of the one or more E. coli TadA monomers comprise mutations that correspond to:

R13A;

T17A;

K20A and R21A;

K20A, R21A, and R23A;

R23W;

E25A;

R26A;

A48G;

I49A;

A56G;

R74A;

D77A;

V82G;

W11A;

V106G;

N108A;

A109W;

K110A;

T111A;

A138G;

D139A and E140A;

A142G;

A143G;

R153A;

V155G;

V155W;

A58G;

N72A;

V106W;

K110A;

H128A and R129A;

A138W;

D139A and E140A;

A142W;

F148A; or

R150A

of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: [[24]]34).

8. The ABE variant of claim 1, further comprising one or more nuclear localization sequences (NLS).

9. The ABE variant of claim 1, comprising a linker between the adenosine deaminase monomers and/or between the adenosine deaminase monomer or between a single-chain dimer and the programmable DNA binding domain.

10. The ABE variant of 9claim 1, wherein the programmable DNA binding domain is a engineered C2H2 zinc-finger, a transcription activator effector-like effector (TALE), or a Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) Cas RNA-guided nucleases (CRISPR-Cas nuclease), or a variant thereof.

11. The ABE variant of claim 10, wherein the CRISPR-Cas nuclease is a single strand DNA (ssDNA) nickase or is catalytically inactive.

12. The ABE variant of claim 11, wherein the CRISPR-Cas nuclease is a Cas9 or Cas12a that has ssDNA nickase activity or is catalytically inactive.

13. A base editing system comprising:

(i) the ABE variant of claims 112claim 1, wherein the programmable DNA binding domain is a CRISPR Cas RGN or a variant thereof; and

(ii) at least one guide RNA compatible with the base editor that directs the base editor to a target sequence.

14. An isolated nucleic acid encoding the ABE variant of claims 112 claim

1.

15. A vector comprising the isolated nucleic acid of claim 14.

16. An isolated host cell, preferably a mammalian host cell, comprising the nucleic acid of claim 14.

17. The isolated host cell of claim 16, which expresses the ABE variant of claim 1.

18. A method of deaminating a selected adenine in a nucleic acid, the method comprising contacting the nucleic acid with the ABE variant of claim 1, and at least one guide RNA compatible with the ABE variant that directs the ABE variant to a target sequence comprising the selected adenine.

19. The method of claim 18, wherein the nucleic acid is in a cell.

20. The method of claim 19, wherein the cell is in a living subject.

21. The method of claim 20, wherein the living subject is a mammal.

22. A composition comprising a purified ABE variant of claim 1, and at least one guide RNA compatible with the ABE variant that directs the ABE variant to a target sequence.

23. The composition of claim 22, comprising one or more ribonucleoprotein (RNP) complexes.

This application claims the benefit of U.S. Provisional Application Ser. No. 62/800,974, filed on Feb. 4, 2019 and U.S. Provisional Application Ser. No. 62/844,717, filed on May 7, 2019. The entire contents of the foregoing are incorporated herein by reference.

This invention was made with Government support under Grant No. HG009490 awarded by the National Institutes of Health and HR0011-17-2-0042 awarded by the Defense Advanced Research Projects Agency (DARPA). The Government has certain rights in the invention.

The instant application contains a Sequence Listing which has been filed electronically in ASCII format and is hereby incorporated by reference in its entirety. Said ASCII copy, created on Mar. 27, 2020, is named 29539-0387001_SL.txt and is 201,290 bytes in size.

Described herein are variants of wild type and engineered E. coli TadA domains of the adenine DNA base editor (ABE) that have reduced unwanted off-target RNA editing activity.

Base editors represent a new genome editing platform that allows efficient installation of single base substitutions in DNA (Rees, H. A. & Liu, D. R. Base editing: precision chemistry on the genome and transcriptome of living cells. Nat. Rev. Genet. (2018); Komor, A. C., et al., Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage, Nature (2016); Gaudelli, N. M. et al., Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature 551, 464-471 (2017)). Adenine base editors (ABEs) are fusions of programmable DNA-binding domains (e.g, catalytically impaired RNA-guided CRISPR-Cas nucleases) linked to engineered adenosine deaminases that can induce programmable adenosine (A) to inosine (I) edits in single-stranded DNA that in turn result in A-to-G transitions after DNA repair or replication. In instances where the programmable DNA-binding domain is a CRISPR-Cas nuclease, targeted adenines lie within an “editing window” in the single-stranded (ss) DNA bubble (R-loop) induced by the CRISPR-Cas RNA-protein complex. The most commonly used ABEs comprise an adenosine deaminase heterodimer consisting of E. coli TadA (wild type) fused to an engineered E. coli TadA variant, as well as a nickase Cas9 and nuclear localization sequences (NLS) (Gaudelli, N. M. et al. Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature 551, 464-471 (2017)). ABEs have been used successfully for installation of A-to-G substitutions in multiple cell types and organisms and could potentially reverse a large number of mutations known to be associated with human disease (Rees, H. A. & Liu, D. R. Base editing: precision chemistry on the genome and transcriptome of living cells. Nat. Rev. Genet. (2018)).

Described herein are adenine base editors (ABEs) having reduced RNA editing activity. These ABEs comprise a programmable DNA-binding domain fused to an adenosine deaminase, e.g. TadA or previously described engineered TadA variants (e.g. ABEs 0.1, 0.2, 1.1, 1.2, 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 2.10, 2.11, 2.12, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 4.1, 4.2, 4.3, 5.1, 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8, 5.9, 5.10, 5.11, 5.12, 5.13, 5.14, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 7.1, 7.2, 7.3, 7.4, 7.5, 7.6, 7.7, 7.8, 7.9, 7.10, ABEmax as described in Gaudelli et al., Nature. 2017 Nov. 23; 551(7681): 464-471 and Koblan et al., Nat Biotechnol. 2018 October; 36(9):843-846), with one or more amino acid substitutions may decrease RNA editing activity while still preserving DNA editing activity (RRE or Reduced RNA Editing variants). The substitutions in E. coli Tad A (SEQ ID NO:1) or its engineered derivatives described herein are at the amino acid positions indicated in Table A. In some embodiments, the ABE comprises one or more mutations corresponding to a mutation listed in Table A.

In one aspect, the present disclosure relates to an adenine base editor (ABE) variant including an adenosine deaminase and a programmable DNA binding domain, the adenosine deaminase including one or more E. coli TadA monomers, where at least one of the one or more E. coli TadA monomers include one or more mutations that decrease RNA editing activity while preserving DNA editing activity.

In one embodiment, the adenosine deaminase includes a wild type or engineered E. coli TadA monomer, or a variation of homo- or heterodimers thereof.

In one embodiment, the adenosine deaminase includes ABE 0.1, ABE 0.2, ABE 1.1, ABE 1.2, ABE 2.1, ABE 2.2, ABE 2.3, ABE 2.4, ABE 2.5, ABE 2.6, ABE 2.7, ABE 2.8, ABE 2.9, ABE 2.10, ABE 2.11, ABE 2.12, ABE 3.1, ABE 3.2, ABE 3.3, ABE 3.4, ABE 3.5, ABE 3.6, ABE 3.7, ABE 3.8, ABE 4.1, ABE 4.2, ABE 4.3, ABE 5.1, ABE 5.2, ABE 5.3, ABE 5.4, ABE 5.5, ABE 5.6, ABE 5.7, ABE 5.8, ABE 5.9, ABE 5.10, ABE 5.11, ABE 5.12, ABE 5.13, ABE 5.14, ABE 6.1, ABE 6.2, ABE 6.3, ABE 6.4, ABE 6.5, ABE 6.6, ABE 7.1, ABE 7.2, ABE 7.3, ABE 7.4, ABE 7.5, ABE 7.6, ABE 7.7, ABE 7.8, ABE 7.9, ABE 7.10, or ABEmax.

In one embodiment, the one or more mutations include one or more mutations at amino acid positions that correspond to residues of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: 34) as listed in Table A. In one embodiment, the one or more mutations are at amino acid positions that correspond to residues Y10, W11, R13, T17, K20, R21, R23, E25, R26, A48, 149, A56, A58, Q71, N72, R74, D77, V82, V106, R107, N108, A109, K110,T111, H122, Y123, H128, R129, A138, D139, E140, A142, A143, F148, and/or R150, R153, V155 of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: 34).

In one embodiment, the one or more mutations include mutations that correspond to Y10A, W11A, R13A, T17A, K20A, R21A, R23A, R23W, E25A, R26A, A48G, 149A, A56G, A58G, Q71A, N72A, R74A, D77A, V82G, V106G, V106W, R107A, N108A, A109G, A109W, K110A, T111A, H122A, Y123A, H128A, R129A, A138W, A138G, D139A, E140A, A142W, A142G, A143G, F148A, R150A, R153A, V155G, and/or V155W of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: 34).

In one embodiment, the at least one of the one or more E. coli TadA monomers include mutations that correspond to: R13A; T17A; K20A and R21A; K20A, R21A, and R23A; R23W; E25A; R26A; A48G; 149A; A56G; R74A; D77A; V82G; W11A; V106G; N108A; A109W; K110A; T111A; A138G; D139A and E140A; A142G; A143G; R153A; V155G; V155W; A58G; N72A; V106W; K110A; H128A and R129A; A138W; D139A and E140A; A142W; F148A; or R150A of wild type E. coli TadA (SEQ ID NO:1) or E. coli TadA deaminase monomer with ABE 7.10 mutations (SEQ ID: 34).

In one embodiment, the ABE variant described herein further includes one or more nuclear localization sequences (NLS).

In one embodiment, the ABE variant described herein includes a linker between the adenosine deaminase monomers and/or between the adenosine deaminase monomer or between a single-chain dimer and the programmable DNA binding domain.

In one embodiment, the programmable DNA binding domain is a engineered C2H2 zinc-finger, a transcription activator effector-like effector (TALE), or a Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) Cas RNA-guided nucleases (CRISPR-Cas nuclease), or a variant thereof.

In one embodiment, the CRISPR-Cas nuclease is a single strand DNA (ssDNA) nickase or is catalytically inactive.

In one embodiment, the CRISPR-Cas nuclease is a Cas9 or Cas12a that has ssDNA nickase activity or is catalytically inactive.

In one aspect, the present disclosure relates to a base editing system including: (i) an ABE variant described herein, where the programmable DNA binding domain is a CRISPR Cas RGN or a variant thereof; and (ii) at least one guide RNA compatible with the base editor that directs the base editor to a target sequence.

In one aspect, the present disclosure relates to an isolated nucleic acid encoding an ABE variant disclosed herein.

In one aspect, the present disclosure relates to a vector including an isolated nucleic acid described herein.

In one aspect, the present disclosure relates to an isolated host cell, preferably a mammalian host cell, including a nucleic acid described herein.

In one embodiment, the isolated host cell described herein expresses any one of the ABE variant described herein.

In one aspect, the present disclosure relates to a method of deaminating a selected adenine in a nucleic acid, the method including contacting the nucleic acid with an ABE variant or a base editing system described herein.

In one embodiment, the nucleic acid is in a cell.

In one embodiment, the cell is in a living subject.

In one embodiment, the living subject is a mammal.

In one aspect, the present disclosure relates to a composition including a purified ABE variant or a base editing system described herein.

In one embodiment, the composition described herein includes one or more ribonucleoprotein (RNP) complexes.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Methods and materials are described herein for use in the present invention; other, suitable methods and materials known in the art can also be used. The materials, methods, and examples are illustrative only and not intended to be limiting. All publications, patent applications, patents, sequences, database entries, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control.

Other features and advantages of the invention will be apparent from the following detailed description and figures, and from the claims.

ABEs efficiently install A-to-G transitions in DNA (Gaudelli, N. M. et al. Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature 551, 464-471 (2017); Koblan, L. W. et al., Improving cytidine and adenine base editors by expression optimization and ancestral reconstruction, SUPPLEMENTS. Nat. Biotechnol. 36, 843-848 (2018)). However, the E. coli TadA protein (present in all commonly used ABEs) was originally discovered as a tRNA-specific adenosine deaminase (FIG. 1A) (Wolf, J. tadA , an essential tRNA-specific adenosine deaminase from Escherichia coli. EMBO J. 21, 3841-3851 (2002); Kim, J. et al. Structural and kinetic characterization of Escherichia coli TadA, the wobble-Specific tRNA deaminase. Biochemistry 45, 6407-6416 (2006)). Directed evolution and protein engineering were used to generate an enzyme that is capable of programmable A-to-I deamination on ssDNA, when fused to a catalytically impaired CRISPR/Cas mutant (Gaudelli, N. M. et al. Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature 551, 464-471 (2017)).

We sought to investigate if the RNA editing capability of the TadA enzyme might have been preserved or even expanded (e.g., to other RNA species) when present in an ABE context (FIG. 1B & C).

Thus, described herein are variants of wild type and engineered (ABE0.1-7.10 and ABEmax) TadA domains, each as monomers and/or combined as single-chain homodimers and/or single-chain heterodimers, bearing mutations that may exhibit reduced RNA editing (RRE) activities while preserving DNA deamination activities, optionally fused to an engineered DNA binding domain such as a CRISPR-Cas nuclease modified to either be a nickase or catalytically inactive, to enable DNA adenine base editing with reduced RNA mutation profiles. These SElective Curbing of Unwanted RNA Editing (SECURE)-ABE variants exhibit substantially reduced unwanted RNA editing activities while retaining robust and more precise on-target DNA editing.

Herein are described structure-guided engineering of SECURE-ABE variants that not only possess reduced off-target RNA editing with comparable on-target DNA activities but are also the smallest Streptococcus pyogenes Cas9 (SpCas9) base editors described to date. Finally, we discovered the important finding that ABEs that exhibit off-target editing activities can also self-edit their own transcripts. This hitherto unappreciated activity leads to substantial heterogeneity in base editor coding sequences and provides strong additional motivation for using variants with reduced RNA editing activities. In sum, our work describes broadly useful SECURE-ABE base editors, defines a new class of unintended alterations caused by base editor self-editing, and reinforces the importance of minimizing RNA editing activities of DNA base editors for research and therapeutic applications.

The work described here extends our understanding of the off-target RNA editing activities of DNA base editors, expands the options available to minimize these unwanted effects, and provides novel SECURE base editor architectures with other desirable properties. The successful engineering of SECURE-ABE variants shows that it is possible to minimize unwanted RNA editing while retaining efficient on-target DNA editing for an ABE. In the process of engineering these variants, we discovered a more extended consensus sequence motif for adenines edited with high efficiencies by ABEmax (CUACGAA) that appears to be recognized by the wild-type TadA part of this fusion. Deletion of this TadA domain abolished recognition of these high efficiency sites and also resulted in the generation of the smallest SpCas9 base editors (1605 amino acids in length) described to date. Our findings further expand the toolbox of base editors that can be used without inducing high-level RNA editing.

Our description of self-editing by DNA base editors provides yet another strong motivation to avoid the use of base editors that possess off-target RNA editing activities. Self-editing by ABEs potentially creates a heterogeneous population of base editor-encoding transcripts in human cells including missense mutations that might lead to the generation of novel epitopes or other gain/loss-of-function effects. The potential impacts of creating diverse mutated forms of base editor proteins in cells are particularly important to consider because these fusions will be highly overexpressed for most applications. One possibility is that these truncated forms might further exacerbate RNA editing activity levels because these proteins would still be expected to induce off-target RNA editing but not on-target DNA editing. Thus, the existence of self-editing further underscores the importance of using DNA base editors with reduced RNA editing activities for both research and therapeutic applications.

In some embodiments, the adenosine deaminase is TadA from E. coli, or an orthologue from a different prokaryote, e.g. S. aureus, or a homologue from the eukaryotic domain, such as yeast TAD1/2 or a mammalian species such as human (e.g. ADAT2). The tRNA-specific adenosine deaminase family members have high sequence homology. FIG. 5 shows the alignment of exemplary tRNA-specific adenosine deaminase orthologues and homologues across different domains, kingdoms and species listed in the uniprot database that are compatible with one or more of the amino acid substitutions in E. coli TadA expected to cause an RRE phenotype and described in this application.

Reduced RNA Editing (RRE) Base Editor Variants

Thus described herein are base editors comprising adenosine deaminases with one or more mutations to reduce undesirable RNA editing activity. In general, these base editors have one or more mutations as described herein. In some embodiments, they have mutations shown in Table A that correspond to residues in wild type (SEQ ID NO: 1) or engineered E. coli TadA (e.g. ABEs 0.1, 0.2, 1.1, 1.2, 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 2.10, 2.11, 2.12, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 4.1, 4.2, 4.3, 5.1, 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8, 5.9, 5.10, 5.11, 5.12, 5.13, 5.14, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 7.1, 7.2, 7.3, 7.4, 7.5, 7.6, 7.7, 7.8, 7.9, 7.10, ABEmax as described in Gaudelli et al., Nature. 2017 November 23; 551(7681): 464-471 (especially supplementary sequences 3 thereof) and Koblan et al., Nat Biotechnol. 2018 October;36(9):843-846). Alternatively, or in addition, they may have mutations in either one of the WT or engineered monomers solely and separately fused to a DNA binding protein such as nickase Cas9, or they might have the same mutations introduced to one or both TadA monomers which are then linked to each other, generating single-chain homo- or heterodimers that in turn are fused to a DNA binding domain.

The mutations can include substitution with any other amino acid other than the WT amino acid; in some embodiments the substitution is with alanine or glycine.

The wild type sequence of wild type E. coli TadA, available in uniprot at P68398, is as follows:

The engineered E. coli TadA sequence present in ABE7.10 and ABEmax is as follows:

In the most commonly used ABEs (ABE7.10 and ABEmax), these two proteins are fused using a 32 amino acid linker (bolded in sequence below), forming a heterodimer, the sequence of which is as follows:

Other exemplary sequences are shown in the list below as well as aligned to E. coli TadA in FIG. 5. These tRNA-specific adenosine deaminase orthologues and homologues also represent candidates for inclusion of the abovementioned mutations at analogous positions in these proteins.

In some embodiments, the base editors do not include catalytically dead adenine deaminase variants, e.g. E59A. (Gaudelli et al, 2017, PM ID: 29160308).

Programmable DNA Binding Domain

In some embodiments, the base editors include programmable DNA binding domains such as engineered C2H2 zinc-fingers, transcription activator effector-like effectors (TALEs), and Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) Cas RNA-guided nucleases (RGNs) and their variants, including ssDNA nickases (nCas9) or their analogs and catalytically inactive dead Cas9 (dCas9) and its analogs (e.g., as shown in Table C), and any engineered protospacer-adjacent motif (PAM) or high-fidelity variants (e.g., as shown inTable D). A programmable DNA binding domain is one that can be engineered to bind to a selected target sequence.

CRISPR-Cas Nucleases

Although herein we refer to Cas9, in general any Cas9-like nickase could be used (including the related Cpf1/Cas12a enzyme classes), unless specifically indicated.

These orthologs, and mutants and variants thereof as known in the art, can be used in any of the fusion proteins described herein. See, e.g., WO 2017/040348 (which describes variants of SaCas9 and SpCas 9 with increased specificity) and WO 2016/141224 (which describes variants of SaCas9 and SpCas 9 with altered PAM specificity).

The Cas9 nuclease from S. pyogenes (hereafter simply Cas9) can be guided via simple base pair complementarity between 17-20 nucleotides of an engineered guide RNA (gRNA), e.g., a single guide RNA or crRNA/tracrRNA pair, and the complementary strand of a target genomic DNA sequence of interest that lies next to a protospacer adjacent motif (PAM), e.g., a PAM matching the sequence NGG or NAG (Shen et al., Cell Res (2013); Dicarlo et al., Nucleic Acids Res (2013); Jiang et al., Nat Biotechnol 31, 233-239 (2013); Jinek et al., Elife 2, e00471 (2013); Hwang et al., Nat Biotechnol 31, 227-229 (2013); Cong et al., Science 339, 819-823 (2013); Mali et al., Science 339, 823-826 (2013c); Cho et al., Nat Biotechnol 31, 230-232 (2013); Jinek et al., Science 337, 816-821 (2012)). The engineered CRISPR from Prevotella and Francisella 1 (Cpf1, also known as Cas12a) nuclease can also be used, e.g., as described in Zetsche et al., Cell 163, 759-771 (2015); Schunder et al., Int J Med Microbiol 303, 51-60 (2013); Makarova et al., Nat Rev Microbiol 13, 722-736 (2015); Fagerlund et al., Genome Biol 16, 251 (2015). Unlike SpCas9, Cpf1/Cas12a requires only a single 42-nt crRNA, which has 23 nt at its 3′ end that are complementary to the protospacer of the target DNA sequence (Zetsche et al., 2015). Furthermore, whereas SpCas9 recognizes an NGG PAM sequence that is 3′ of the protospacer, AsCpf1 and LbCp1 recognize TTTN PAMs that are found 5′ of the protospacer (Id.).

In some embodiments, the present system utilizes a wild type or variant Cas9 protein from S. pyogenes or Staphylococcus aureus, or a wild type or variant Cpf1 protein from Acidaminococcus sp. BV3L6 or Lachnospiraceae bacterium ND2006 either as encoded in bacteria or codon-optimized for expression in mammalian cells and/or modified in its PAM recognition specificity and/or its genome-wide specificity. A number of variants have been described; see, e.g., WO 2016/141224, PCT/US2016/049147, Kleinstiver et al., Nat Biotechnol. 2016 August;34(8):869-74; Tsai and Joung, Nat Rev Genet. 2016 May;17(5):300-12; Kleinstiver et al., Nature. 2016 Jan. 28;529(7587):490-5; Shmakov et al., Mol Cell. 2015 Nov. 5;60(3):385-97; Kleinstiver et al., Nat Biotechnol. 2015 December;33(12):1293-1298; Dahlman et al., Nat Biotechnol. 2015 November;33(11):1159-61; Kleinstiver et al., Nature. 2015 Jul. 23;523(7561):481-5; Wyvekens et al., Hum Gene Ther. 2015 July;26(7):425-31; Hwang et al., Methods Mol Biol. 2015;1311:317-34; Osborn et al., Hum Gene Ther. 2015 February;26(2):114-26; Konermann et al., Nature. 2015 Jan. 29;517(7536):583-8; Fu et al., Methods Enzymol. 2014;546:21-45; and Tsai et al., Nat Biotechnol. 2014 June;32(6):569-76, inter alia. Concerning rAPOBEC1 itself, a number of variants have been described, e.g. Chen et al, RNA. 2010 May;16(5):1040-52; Chester et al, EMBO J. 2003 Aug. 1;22(15):3971-82.: Teng et al, J Lipid Res. 1999 April;40(4):623-35.; Navaratnam et al, Cell. 1995 Apr. 21;81(2):187-95.; MacGinnitie et al, J Biol Chem. 1995 Jun. 16;270(24):14768-75.;Yamanaka et al, J Biol Chem. 1994 Aug. 26;269(34):21725-34. The guide RNA is expressed or present in the cell together with the Cas9 or Cpf1. Either the guide RNA or the nuclease, or both, can be expressed transiently or stably in the cell or introduced as a purified protein or nucleic acid.

In some embodiments, the Cas9 also includes one of the following mutations, which reduce nuclease activity of the Cas9; e.g., for SpCas9, mutations at D10A or H840A (which creates a single-strand nickase).

In some embodiments, the SpCas9 variants also include mutations at one of each of the two sets of the following amino acid positions, which together destroy the nuclease activity of the Cas9: D10, E762, D839, H983, or D986 and H840 or N863, e.g., D10A/D10N and H840A/H840N/H840Y, to render the nuclease portion of the protein catalytically inactive; substitutions at these positions could be alanine (as they are in Nishimasu al., Cell 156, 935-949 (2014)), or other residues, e.g., glutamine, asparagine, tyrosine, serine, or aspartate, e.g., E762Q, H983N, H983Y, D986N, N863D, N863S, or N863H (see WO 2014/152432).

In some embodiments, the Cas9 is fused to one or more SV40 or bipartite (bp) nuclear localization sequences (NLSs) protein sequences; an exemplary (bp)NLS sequence is as follows: (KRTADGSEFES)PKKKRKV (SEQ ID NO: 23). Typically, the NLSs are at the N- and C-termini of an ABEmax fusion protein, but can also be positioned at the N- or C-terminus in other ABEs, or between the DNA binding domain and the deaminase domain. Linkers as known in the art can be used to separate domains.

TAL Effector Repeat Arrays

Transcription activator like effectors (TALEs) of plant pathogenic bacteria in the genus Xanthomonas play important roles in disease, or trigger defense, by binding host DNA and activating effector-specific host genes. Specificity depends on an effector-variable number of imperfect, typically ˜33-35 amino acid repeats. Polymorphisms are present primarily at repeat positions 12 and 13, which are referred to herein as the repeat variable-diresidue (RVD). The RVDs of TAL effectors correspond to the nucleotides in their target sites in a direct, linear fashion, one RVD to one nucleotide, with some degeneracy and no apparent context dependence. In some embodiments, the polymorphic region that grants nucleotide specificity may be expressed as a triresidue or triplet.

Each DNA binding repeat can include a RVD that determines recognition of a base pair in the target DNA sequence, wherein each DNA binding repeat is responsible for recognizing one base pair in the target DNA sequence. In some embodiments, the RVD can comprise one or more of: HA for recognizing C; ND for recognizing C; HI for recognizing C; HN for recognizing G; NA for recognizing G; SN for recognizing G or A; YG for recognizing T; and NK for recognizing G, and one or more of: HD for recognizing C; NG for recognizing T; NI for recognizing A; NN for recognizing G or A; NS for recognizing A or C or G or T; N* for recognizing C or T, wherein * represents a gap in the second position of the RVD; HG for recognizing T; H* for recognizing T, wherein * represents a gap in the second position of the RVD; and IG for recognizing T.

TALE proteins may be useful in research and biotechnology as targeted chimeric nucleases that can facilitate homologous recombination in genome engineering (e.g., to add or enhance traits useful for biofuels or biorenewables in plants). These proteins also may be useful as, for example, transcription factors, and especially for therapeutic applications requiring a very high level of specificity such as therapeutics against pathogens (e.g., viruses) as non-limiting examples.

Methods for generating engineered TALE arrays are known in the art, see, e.g., the fast ligation-based automatable solid-phase high-throughput (FLASH) system described in U.S. Ser. No. 61/610,212, and Reyon et al., Nature Biotechnology 30,460-465 (2012); as well as the methods described in Bogdanove & Voytas, Science 333, 1843-1846 (2011); Bogdanove et al., Curr Opin Plant Biol 13, 394-401 (2010); Scholze & Boch, J. Curr Opin Microbiol (2011); Boch et al., Science 326, 1509-1512 (2009); Moscou & Bogdanove, Science 326, 1501 (2009); Miller et al., Nat Biotechnol 29, 143-148 (2011); Morbitzer et al., T. Proc Natl Acad Sci U S A 107, 21617-21622 (2010); Morbitzer et al., Nucleic Acids Res 39, 5790-5799 (2011); Zhang et al., Nat Biotechnol 29, 149-153 (2011); Geissler et al., PLoS ONE 6, e19509 (2011); Weber et al., PLoS ONE 6, e19722 (2011); Christian et al., Genetics 186, 757-761 (2010); Li et al., Nucleic Acids Res 39, 359-372 (2011); Mahfouz et al., Proc Natl Acad Sci U S A 108, 2623-2628 (2011); Mussolino et al., Nucleic Acids Res (2011); Li et al., Nucleic Acids Res 39, 6315-6325 (2011); Cermak et al., Nucleic Acids Res 39, e82 (2011); Wood et al., Science 333, 307 (2011); Hockemeye et al. Nat Biotechnol 29, 731-734 (2011); Tesson et al., Nat Biotechnol 29, 695-696 (2011); Sander et al., Nat Biotechnol 29, 697-698 (2011); Huang et al., Nat Biotechnol 29, 699-700 (2011); and Zhang et al., Nat Biotechnol 29, 149-153 (2011); all of which are incorporated herein by reference in their entirety.

Zinc Fingers

Zinc finger (ZF) proteins are DNA-binding proteins that contain one or more zinc fingers, independently folded zinc-containing mini-domains, the structure of which is well known in the art and defined in, for example, Miller et al., 1985, EMBO J., 4:1609; Berg, 1988, Proc. Natl. Acad. Sci. USA, 85:99; Lee et al., 1989, Science. 245:635; and Klug, 1993, Gene, 135:83. Crystal structures of the zinc finger protein Zif268 and its variants bound to DNA show a semi-conserved pattern of interactions, in which typically three amino acids from the alpha-helix of the zinc finger contact three adjacent base pairs or a “subsite” in the DNA (Pavletich et al., 1991, Science, 252:809; Elrod-Erickson et al., 1998, Structure, 6:451). Thus, the crystal structure of Zif268 suggested that zinc finger DNA-binding domains might function in a modular manner with a one-to-one interaction between a zinc finger and a three-base-pair “subsite” in the DNA sequence. In naturally occurring zinc finger transcription factors, multiple zinc fingers are typically linked together in a tandem array to achieve sequence-specific recognition of a contiguous DNA sequence (Klug, 1993, Gene 135:83).

Multiple studies have shown that it is possible to artificially engineer the DNA binding characteristics of individual zinc fingers by randomizing the amino acids at the alpha-helical positions involved in DNA binding and using selection methodologies such as phage display to identify desired variants capable of binding to DNA target sites of interest (Rebar et al., 1994, Science, 263:671; Choo et al., 1994 Proc. Natl. Acad. Sci. USA, 91:11163; Jamieson et al., 1994, Biochemistry 33:5689; Wu et al., 1995 Proc. Natl. Acad. Sci. USA, 92: 344). Such recombinant zinc finger proteins can be fused to functional domains, such as transcriptional activators, transcriptional repressors, methylation domains, and nucleases to regulate gene expression, alter DNA methylation, and introduce targeted alterations into genomes of model organisms, plants, and human cells (Carroll, 2008,., 15:1463—Gene Ther68; Cathomen, 2008, Mol. Ther., 16:1200-07; Wu et al., 2007, Cell. Mol. Life Sci., 64:2933-44).

One existing method for engineering zinc finger arrays, known as “modular assembly,” advocates the simple joining together of pre-selected zinc finger modules into arrays (Segal et al., 2003, Biochemistry, 42:2137-48; Beerli et al., 2002, Nat. Biotechnol., 20:135-141; Mandell et al., 2006, Nucleic Acids Res., 34:W516-523; Carroll et al., 2006, Nat. Protoc. 1:1329-41; Liu et al., 2002, J. Biol. Chem., 277:3850-56; Bae et al., 2003, Nat. Biotechnol., 21:275-280; Wright et al., 2006, Nat. Protoc., 1:1637-52). Although straightforward enough to be practiced by any researcher, recent reports have demonstrated a high failure rate for this method, particularly in the context of zinc finger nucleases (Ramirez et al., 2008, Nat. Methods, 5:374-375; Kim et al., 2009, Genome Res. 19:1279-88), a limitation that typically necessitates the construction and cell-based testing of very large numbers of zinc finger proteins for any given target gene (Kim et al., 2009, Genome Res. 19:1279-88).

Combinatorial selection-based methods that identify zinc finger arrays from randomized libraries have been shown to have higher success rates than modular assembly (Maeder et al., 2008, Mol. Cell, 31:294-301; Joung et al., 2010, Nat. Methods, 7:91-92; Isalan et al., 2001, Nat. Biotechnol., 19:656-660). In preferred embodiments, the zinc finger arrays are described in, or are generated as described in, WO 2011/017293 and WO 2004/099366. Additional suitable zinc finger DBDs are described in U.S. Pat. Nos. 6,511,808, 6,013,453, 6,007,988, and 6,503,717 and U.S. patent application 2002/0160940.

Variants

In some embodiments, the components of the fusion proteins are at least 80%, e.g., at least 85%, 90%, 95%, 97%, or 99% identical to the amino acid sequence of an exemplary sequence (e.g., a TadA or DBD as provided herein), e.g., have differences at up to 1%, 2%, 5%, 10%, 15%, or 20% of the residues of the exemplary sequence replaced, e.g., with conservative mutations, e.g., including or in addition to the mutations described herein. In preferred embodiments, the variant retains a desired activity of the parent, e.g., deaminase activity, and/or the ability to interact with a guide RNA and/or target DNA, optionally with improved specificity or altered substrate specificity.

To determine the percent identity of two nucleic acid sequences, the sequences are aligned for optimal comparison purposes (e.g., gaps can be introduced in one or both of a first and a second amino acid or nucleic acid sequence for optimal alignment and non-homologous sequences can be disregarded for comparison purposes). The length of a reference sequence aligned for comparison purposes is at least 80% of the length of the reference sequence, and in some embodiments is at least 90% or 100%. The nucleotides at corresponding amino acid positions or nucleotide positions are then compared. When a position in the first sequence is occupied by the same nucleotide as the corresponding position in the second sequence, then the molecules are identical at that position (as used herein nucleic acid “identity” is equivalent to nucleic acid “homology”). The percent identity between the two sequences is a function of the number of identical positions shared by the sequences, taking into account the number of gaps, and the length of each gap, which need to be introduced for optimal alignment of the two sequences. Percent identity between two polypeptides or nucleic acid sequences is determined in various ways that are within the skill in the art, for instance, using publicly available computer software such as Smith Waterman Alignment (Smith, T. F. and M. S. Waterman (1981) J Mol Biol 147:195-7); “BestFit” (Smith and Waterman, Advances in Applied Mathematics, 482-489 (1981)) as incorporated into GeneMatcher Plus™, Schwarz and Dayhof (1979) Atlas of Protein Sequence and Structure, Dayhof, M. O., Ed, pp 353-358; BLAST program (Basic Local Alignment Search Tool; (Altschul, S. F., W. Gish, et al. (1990) J Mol Biol 215: 403-10), BLAST-2, BLAST-P, BLAST-N, BLAST-X, WU-BLAST-2, ALIGN, ALIGN-2, CLUSTAL, or Megalign (DNASTAR) software. In addition, those skilled in the art can determine appropriate parameters for measuring alignment, including any algorithms needed to achieve maximal alignment over the length of the sequences being compared. In general, for proteins or nucleic acids, the length of comparison can be any length, up to and including full length (e.g., 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, or 100%). For purposes of the present compositions and methods, at least 80% of the full length of the sequence is aligned.

For purposes of the present disclosure, the comparison of sequences and determination of percent identity between two sequences can be accomplished using a Blossum 62 scoring matrix with a gap penalty of 12, a gap extend penalty of 4, and a frameshift gap penalty of 5.

Conservative substitutions typically include substitutions within the following groups: glycine, alanine; valine, isoleucine, leucine; aspartic acid, glutamic acid, asparagine, glutamine; serine, threonine; lysine, arginine; and phenylalanine, tyrosine.

Also provided herein are isolated nucleic acids encoding the base editor fusion proteins, vectors comprising the isolated nucleic acids, optionally operably linked to one or more regulatory domains for expressing the variant proteins, and host cells, e.g., mammalian host cells, comprising the nucleic acids, and optionally expressing the variant proteins. In some embodiments, the host cells are stem cells, e.g., hematopoietic stem cells. In some embodiments, the fusion proteins include a linker between the DNA binding domain (e.g., ZFN, TALE, or nCas9) and the BE domains. Linkers that can be used in these fusion proteins (or between fusion proteins in a concatenated structure) can include any sequence that does not interfere with the function of the fusion proteins. In preferred embodiments, the linkers are short, e.g., 2-20 amino acids, and are typically flexible (i.e., comprising amino acids with a high degree of freedom such as glycine, alanine, and serine). In some embodiments, the linker comprises one or more units consisting of GGGS (SEQ ID NO:24) or GGGGS (SEQ ID NO:25), e.g., two, three, four, or more repeats of the GGGS (SEQ ID NO:24) or GGGGS (SEQ ID NO:25) unit. Other linker sequences can also be used.

In some embodiments, the deaminase fusion protein includes a cell-penetrating peptide sequence that facilitates delivery to the intracellular space, e.g., HIV-derived TAT peptide, penetratins, transportans, or hCT derived cell-penetrating peptides, see, e.g., Caron et al., (2001) Mol Ther. 3(3):310-8; Lange!, Cell-Penetrating Peptides: Processes and Applications (CRC Press, Boca Raton Fla. 2002); El-Andaloussi et al., (2005) Curr Pharm Des. 11(28):3597-611; and Deshayes et al., (2005) Cell Mol Life Sci. 62(16):1839-49.

Cell penetrating peptides (CPPs) are short peptides that facilitate the movement of a wide range of biomolecules across the cell membrane into the cytoplasm or other organelles, e.g. the mitochondria and the nucleus. Examples of molecules that can be delivered by CPPs include therapeutic drugs, plasmid DNA, oligonucleotides, siRNA, peptide-nucleic acid (PNA), proteins, peptides, nanoparticles, and liposomes. CPPs are generally 30 amino acids or less, are derived from naturally or non-naturally occurring protein or chimeric sequences, and contain either a high relative abundance of positively charged amino acids, e.g. lysine or arginine, or an alternating pattern of polar and non-polar amino acids. CPPs that are commonly used in the art include Tat (Frankel et al., (1988) Cell. 55:1189-1193, Vives et al., (1997) J. Biol. Chem. 272:16010-16017), penetratin (Derossi et al., (1994) J. Biol. Chem. 269:10444-10450), polyarginine peptide sequences (Wender et al., (2000) Proc. Natl. Acad. Sci. USA 97:13003-13008, Futaki et al., (2001) J. Biol. Chem. 276:5836-5840), and transportan (Pooga et al., (1998) Nat. Biotechnol. 16:857-861).

CPPs can be linked with their cargo through covalent or non-covalent strategies. Methods for covalently joining a CPP and its cargo are known in the art, e.g. chemical cross-linking (Stetsenko et al., (2000) J. Org. Chem. 65:4900-4909, Gait et al. (2003) Cell. Mol. Life. Sci. 60:844-853) or cloning a fusion protein (Nagahara et al., (1998) Nat. Med. 4:1449-1453). Non-covalent coupling between the cargo and short amphipathic CPPs comprising polar and non-polar domains is established through electrostatic and hydrophobic interactions.

CPPs have been utilized in the art to deliver potentially therapeutic biomolecules into cells. Examples include cyclosporine linked to polyarginine for immunosuppression (Rothbard et al., (2000) Nature Medicine 6(11):1253-1257), siRNA against cyclin B1 linked to a CPP called MPG for inhibiting tumorigenesis (Crombez et al., (2007) Biochem Soc. Trans. 35:44-46), tumor suppressor p53 peptides linked to CPPs to reduce cancer cell growth (Takenobu et al., (2002) Mol. Cancer Ther. 1(12):1043-1049, Snyder et al., (2004) PLoS Biol. 2:E36), and dominant negative forms of Ras or phosphoinositol 3 kinase (P13K) fused to Tat to treat asthma (Myou et al., (2003) J. lmmunol. 171:4399-4405).

CPPs have been utilized in the art to transport contrast agents into cells for imaging and biosensing applications. For example, green fluorescent protein (GFP) attached to Tat has been used to label cancer cells (Shokolenko et al., (2005) DNA Repair 4(4):511-518). Tat conjugated to quantum dots have been used to successfully cross the blood-brain barrier for visualization of the rat brain (Santra et al., (2005) Chem. Commun. 3144-3146). CPPs have also been combined with magnetic resonance imaging techniques for cell imaging (Liu et al., (2006) Biochem. and Biophys. Res. Comm. 347(1):133-140). See also Ramsey and Flynn, Pharmacol Ther. 2015 Jul. 22. pii: S0163-7258(15)00141-2.

Alternatively or in addition, the deaminase fusion proteins can include a nuclear localization sequence, e.g., SV40 large T antigen NLS (PKKKRRV (SEQ ID NO:26)) and nucleoplasmin NLS (KRPAATKKAGQAKKKK (SEQ ID NO:27)). Other NLSs are known in the art; see, e.g., Cokol et al., EMBO Rep. 2000 Nov. 15; 1(5): 411-415; Freitas and Cunha, Curr Genomics. 2009 December; 10(8): 550-557.

In some embodiments, the deaminase fusion proteins include a moiety that has a high affinity for a ligand, for example GST, FLAG or hexahistidine (SEQ ID NO: 35) sequences. Such affinity tags can facilitate the purification of recombinant deaminase fusion proteins.

The deaminase fusion proteins described herein can be used for altering the genome of a cell. The methods generally include expressing or contacting the deaminase fusion proteins in the cells; in versions using one or two Cas9s, the methods include using a guide RNA having a region complementary to a selected portion of the genome of the cell. Methods for selectively altering the genome of a cell are known in the art, see, e.g., U.S. Pat. No. 8,993,233; US 20140186958; U.S. Pat. No. 9,023,649; WO/2014/099744; WO 2014/089290; WO2014/144592; WO144288; WO2014/204578; WO2014/152432; WO2115/099850; U.S. Pat. No. 8,697,359; US20160024529; US20160024524; US20160024523; US20160024510; US20160017366; US20160017301; US20150376652; US20150356239; US20150315576; US20150291965; US20150252358; US20150247150; US20150232883; US20150232882; US20150203872; US20150191744; US20150184139; US20150176064; US20150167000; US20150166969; US20150159175; US20150159174; US20150093473; US20150079681; US20150067922; US20150056629; US20150044772; US20150024500; US20150024499; US20150020223;; US20140356867; US20140295557; US20140273235; US20140273226; US20140273037; US20140189896; US20140113376; US20140093941; US20130330778; US20130288251; US20120088676; US20110300538; US20110236530; US20110217739; US20110002889; US20100076057; US20110189776; US20110223638; US20130130248; US20150050699; US20150071899; US20150050699; ; US20150045546; US20150031134; US20150024500; US20140377868; US20140357530; US20140349400; US20140335620; US20140335063; US20140315985; US20140310830; US20140310828; US20140309487; US20140304853; US20140298547; US20140295556; US20140294773; US20140287938; US20140273234; US20140273232; US20140273231; US20140273230; US20140271987; US20140256046; US20140248702; US20140242702; US20140242700; US20140242699; US20140242664; US20140234972; US20140227787; US20140212869; US20140201857; US20140199767; US20140189896; US20140186958; US20140186919; US20140186843; US20140179770; US20140179006; US20140170753; WO/2008/108989; WO/2010/054108; WO/2012/164565; WO/2013/098244; WO/2013/176772; US 20150071899; Makarova et al., “Evolution and classification of the CRISPR-Cas systems” 9(6) Nature Reviews Microbiology 467-477 (1-23) (June 2011); Wedenheft et al., “RNA-guided genetic silencing systems in bacteria and archaea” 482 Nature 331-338 (Feb. 16, 2012); Gasiunas et al., “Cas9-crRNA ribonucleoprotein complex mediates specific DNA cleavage for adaptive immunity in bacteria” 109(39) Proceedings of the National Academy of Sciences USA E2579-E2586 (Sep. 4, 2012); Jinek et al., “A Programmable Dual-RNA-Guided DNA Endonuclease in Adaptive Bacterial Immunity” 337 Science 816-821 (Aug. 17, 2012); Carroll, “A CRISPR Approach to Gene Targeting” 20(9) Molecular Therapy 1658-1660 (September 2012); U.S. Appl. No. 61/652,086, filed May 25, 2012; Al-Attar et al., Clustered Regularly Interspaced Short Palindromic Repeats (CRISPRs): The Hallmark of an Ingenious Antiviral Defense Mechanism in Prokaryotes, Biol Chem. (2011) vol. 392, Issue 4, pp. 277-289; Hale et al., Essential Features and Rational Design of CRISPR RNAs That Function With the Cas RAMP Module Complex to Cleave RNAs, Molecular Cell, (2012) vol. 45, Issue 3, 292-302. For methods in which the deaminase fusion proteins are delivered to cells, the proteins can be produced using any method known in the art, e.g., by in vitro translation, or expression in a suitable host cell from nucleic acid encoding the deaminase fusion protein; a number of methods are known in the art for producing proteins. For example, the proteins can be produced in and purified from yeast, E. coli, insect cell lines, plants, transgenic animals, or cultured mammalian cells; see, e.g., Palomares et al., “Production of Recombinant Proteins: Challenges and Solutions,” Methods Mol Biol. 2004;267:15-52. In addition, the deaminase fusion proteins can be linked to a moiety that facilitates transfer into a cell, e.g., a lipid nanoparticle, optionally with a linker that is cleaved once the protein is inside the cell. See, e.g., LaFountaine et al., Int J Pharm. 2015 Aug. 13;494(1):180-194.

Expression Systems

To use the deaminase fusion proteins described herein, it may be desirable to express them from a nucleic acid that encodes them. This can be performed in a variety of ways. For example, the nucleic acid encoding the deaminase fusion can be cloned into an intermediate vector for transformation into prokaryotic or eukaryotic cells for replication and/or expression. Intermediate vectors are typically prokaryote vectors, e.g., plasmids, or shuttle vectors, or insect vectors, for storage or manipulation of the nucleic acid encoding the deaminase fusion for production of the deaminase fusion protein. The nucleic acid encoding the deaminase fusion protein can also be cloned into an expression vector, for administration to a plant cell, animal cell, preferably a mammalian cell or a human cell, fungal cell, bacterial cell, or protozoan cell.

To obtain expression, a sequence encoding a deaminase fusion protein is typically subcloned into an expression vector that contains a promoter to direct transcription. Suitable bacterial and eukaryotic promoters are well known in the art and described, e.g., in Sambrook et al., Molecular Cloning, A Laboratory Manual (3d ed. 2001); Kriegler, Gene Transfer and Expression: A Laboratory Manual (1990); and Current Protocols in Molecular Biology (Ausubel et al., eds., 2010). Bacterial expression systems for expressing the engineered protein are available in, e.g., E. coli, Bacillus sp., and Salmonella (Palva et al., 1983, Gene 22:229-235). Kits for such expression systems are commercially available. Eukaryotic expression systems for mammalian cells, yeast, and insect cells are well known in the art and are also commercially available.

The promoter used to direct expression of a nucleic acid depends on the particular application. For example, a strong constitutive promoter is typically used for expression and purification of fusion proteins. In contrast, when the deaminase fusion protein is to be administered in vivo for gene regulation, either a constitutive or an inducible promoter can be used, depending on the particular use of the deaminase fusion protein. In addition, a preferred promoter for administration of the deaminase fusion protein can be a weak promoter, such as HSV TK or a promoter having similar activity. The promoter can also include elements that are responsive to transactivation, e.g., hypoxia response elements, Gal4 response elements, lac repressor response element, and small molecule control systems such as tetracycline-regulated systems and the RU-486 system (see, e.g., Gossen & Bujard, 1992, Proc. Natl. Acad. Sci. USA, 89:5547; Oligino et al., 1998, Gene Ther., 5:491-496; Wang et al., 1997, Gene Ther., 4:432-441; Neering et al., 1996, Blood, 88:1147-55; and Rendahl et al., 1998, Nat. Biotechnol., 16:757-761).

In addition to the promoter, the expression vector typically contains a transcription unit or expression cassette that contains all the additional elements required for the expression of the nucleic acid in host cells, either prokaryotic or eukaryotic. A typical expression cassette thus contains a promoter operably linked, e.g., to the nucleic acid sequence encoding the deaminase fusion protein, and any signals required, e.g., for efficient polyadenylation of the transcript, transcriptional termination, ribosome binding sites, or translation termination. Additional elements of the cassette may include, e.g., enhancers, and heterologous spliced intronic signals.

The particular expression vector used to transport the genetic information into the cell is selected with regard to the intended use of the deaminase fusion protein, e.g., expression in plants, animals, bacteria, fungus, protozoa, etc. Standard bacterial expression vectors include plasmids such as pBR322 based plasmids, pSKF, pET23D, and commercially available tag-fusion expression systems such as GST and LacZ.

Expression vectors containing regulatory elements from eukaryotic viruses are often used in eukaryotic expression vectors, e.g., SV40 vectors, papilloma virus vectors, and vectors derived from Epstein-Barr virus. Other exemplary eukaryotic vectors include pMSG, pAV009/A+, pMT010/A+, pMAMneo-5, baculovirus pDSVE, and any other vector allowing expression of proteins under the direction of the SV40 early promoter, SV40 late promoter, metallothionein promoter, murine mammary tumor virus promoter, Rous sarcoma virus promoter, polyhedrin promoter, or other promoters shown effective for expression in eukaryotic cells.

The vectors for expressing the deaminase fusion protein can include RNA Pol III promoters to drive expression of the guide RNAs, e.g., the H1, U6 or 7SK promoters. These human promoters allow for expression of deaminase fusion protein in mammalian cells following plasmid transfection.

Some expression systems have markers for selection of stably transfected cell lines such as thymidine kinase, hygromycin B phosphotransferase, and dihydrofolate reductase. High yield expression systems are also suitable, such as using a baculovirus vector in insect cells, with the gRNA encoding sequence under the direction of the polyhedrin promoter or other strong baculovirus promoters.

The elements that are typically included in expression vectors also include a replicon that functions in E. coli, a gene encoding antibiotic resistance to permit selection of bacteria that harbor recombinant plasmids, and unique restriction sites in nonessential regions of the plasmid to allow insertion of recombinant sequences.

Standard transfection methods are used to produce bacterial, mammalian, yeast or insect cell lines that express large quantities of protein, which are then purified using standard techniques (see, e.g., Colley et al., 1989, J. Biol. Chem., 264:17619-22; Guide to Protein Purification, in Methods in Enzymology, vol. 182 (Deutscher, ed., 1990)). Transformation of eukaryotic and prokaryotic cells are performed according to standard techniques (see, e.g., Morrison, 1977, J. Bacteriol. 132:349-351; Clark-Curtiss & Curtiss, Methods in Enzymology 101:347-362 (Wu et al., eds, 1983).

Any of the known procedures for introducing foreign nucleotide sequences into host cells may be used. These include the use of calcium phosphate transfection, polybrene, protoplast fusion, electroporation, nucleofection, liposomes, microinjection, naked DNA, plasmid vectors, viral vectors, both episomal and integrative, and any of the other well-known methods for introducing cloned genomic DNA, cDNA, synthetic DNA or other foreign genetic material into a host cell (see, e.g., Sambrook et al., supra). It is only necessary that the particular genetic engineering procedure used be capable of successfully introducing at least one gene into the host cell capable of expressing the deaminase fusion protein.

In methods wherein the fusion proteins include a Cas9 domain, the methods also include delivering at least one gRNA that interacts with the Cas9, or a nucleic acid that encodes a gRNA.

Alternatively, the methods can include delivering the deaminase fusion protein and guide RNA together, e.g., as a complex. For example, the deaminase fusion protein and gRNA can be can be overexpressed in a host cell and purified, then complexed with the guide RNA (e.g., in a test tube) to form a ribonucleoprotein (RNP), and delivered to cells. In some embodiments, the deaminase fusion protein can be expressed in and purified from bacteria through the use of bacterial expression plasmids. For example, His-tagged deaminase fusion protein can be expressed in bacterial cells and then purified using nickel affinity chromatography. The use of RNPs circumvents the necessity of delivering plasmid DNAs encoding the nuclease or the guide, or encoding the nuclease as an mRNA. RNP delivery may also improve specificity, presumably because the half-life of the RNP is shorter and there's no persistent expression of the nuclease and guide (as you'd get from a plasmid). The RNPs can be delivered to the cells in vivo or in vitro, e.g., using lipid-mediated transfection or electroporation. See, e.g., Liang et al. “Rapid and highly efficient mammalian cell engineering via Cas9 protein transfection.” Journal of biotechnology 208 (2015): 44-53; Zuris, John A., et al. “Cationic lipid-mediated delivery of proteins enables efficient protein-based genome editing in vitro and in vivo.” Nature biotechnology 33.1 (2015): 73-80; Kim et al. “Highly efficient RNA-guided genome editing in human cells via delivery of purified Cas9 ribonucleoproteins.” Genome research 24.6 (2014): 1012-1019.

The present invention also includes the vectors and cells comprising the vectors, as well as kits comprising the proteins and nucleic acids described herein, e.g., for use in a method described herein.

Methods of Use

The base editors described herein can be used to deaminate a selected adenine in a nucleic acid sequence, e.g., in a cell, e.g., a cell in an animal (e.g., a mammal such as a human or veterinary subject), or a synthetic nucleic acid substrate. The methods include contacting the nucleic acid with a base editor as described herein. Where the base editor includes a CRISPR Cas9 or Cas12a protein, the methods further include the use of one or more guide RNAs that direct binding of the base editor to a sequence to be deaminated.

For example, the base editors described herein can be used for in vitro, in vivo or in situ directed evolution, e.g., to engineer polypeptides or proteins based on a synthetic selection framework, e.g. antibiotic resistance in E. coli or resistance to anti-cancer therapeutics being assayed in mammalian cells (e.g. CRISPR-X Hess et al, PMID: 27798611 or BE-plus systems Jiang et al, PMID: 29875396).

The invention is further described in the following examples, which do not limit the scope of the invention described in the claims.

Methods

The following materials and methods were used in the Examples set forth herein.

Molecular Cloning

Expression plasmids are constructed by selectively amplifying desired DNA sequences using the PCR method such that they have significant overlapping ends and using isothermal assembly (or “Gibson Assembly”, NEB) to assemble them in the desired order in a CAG or CMV expression vector. PCR is conducted using Phusion HF polymerase (NEB). Cas9 gRNAs is cloned into the pUC19-based entry vector BPK1520 (via BsmBI) under control of a U6 promoter.

Guide RNAs

All gRNAs are of the form 5′-NNNNNNNNNNNNNNNNNNNNCGTTTTAGAGCTAGAAATAGCAAGTTAAAATAAGGCTAGTC CGTTATCAACTTGAAAAAGTGGCACCGAGTCGGTGCTTTTTTT-3′. (SEQ ID NO:28) Shown below are the protospacer regions (NNNNNNNNNNNNNNNNNNNN in SEQ ID NO:28) for these gRNAs (all written 5′ to 3′).

Cas9 guide RNA protospacer, RNF2 site 1: GTCATCTTAGTCATTACCTG (SEQ ID NO:30)

Cas9 guide RNA protospacer, RNF2 site 1: GTCATCTTAGTCATTACCTG (SEQ ID NO:30)

Cell Culture and Transfections

HEK293T cells (CRL-3216, ATCC) are grown in culture using Dulbeccos Modified Medium (Gibco) supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin solution (Gibco). Cells are passaged at ˜80% confluency every 2-3 days to maintain an actively growing population. HepG2 cells (HB80-65, ATCC) are grown in Eagle's Minimum Essential Medium (ATCC) supplemented with 10% FBS and 0.5% penicillin-streptomycin solution (Gibco). Cells are passaged at ˜80% confluency every 4 days. Both cell lines are used for experiments until passage 20 for HEK293T and passage 12 for HepG2. Cells are tested for mycoplasma bi-weekly.

For sorting experiments, transfections with 50 ug of transfection quality DNA (Qiagen Maxiprep) encoding desired ABEmax-P2A-EGFP fusion proteins or controls (same construct, lacking TadA-TadA* heterodimer, * marks the engineered variants, e.g. 7.10) and gRNAs (75:25%) were conducted by seeding 6×10⁶HEK293T or 15×10⁶HepG2 into TC-treated 150 mm plates 18-24 h prior to transfection to yield -80% confluency on the day of transfection. Cells are transfected at 60-80% confluency using TransIT-293 (HEK293T, Mirus) or tranfeX (HepG2, ATCC) reagents according to the manufacturers' protocols. To ensure maximal correlation of negative controls to ABE overexpression, cells of the same passage are transfected with bpNLS-32AAlinker-nCas9-bpNLS (negative control) and adenine base editors in parallel. RNA and gDNA is harvested after cell sorting. For experiments validating DNA on-target activity of ABE or ABEmax-RRE variants, 1.5×10⁴HEK293T cells are seeded into the wells of a 96-well plate and transfected 18-24 h after seeding with 220 ng DNA (ABE/nCas9-NLS control:gRNA ration of 75:25%). For these experiments, gDNA is harvested 72 h post-transfection.

FACS & RNA/DNA Harvest

Sorting of negative control and BE expressing cells as well as RNA/DNA harvest is carried out on the same day. Cells are sorted on a BD FACSARIAII 36-40 h after transfection. We gate on the cell population on forward/sideward scatter after exclusion of doublets. We then sort all GFP-positive cells and/or top 5% of cells with the highest FITC signal into pre-chilled 100% FBS and 5% of mean fluorescence intensity (MFI)-matched cells for nCas9-NLS negative controls, matching the MFI/GeoMean of top 5% of ABE or ABEmax-transfected cells. We use MFI-matching for these controls, as the bpNLS-32AAlinker-nCas9-bpNLS-P2A-EGFP (control) plasmid is smaller than ABEmax-P2A-EGFP—due to the lack of the TadA-TadA* heterodimer—and thus yields higher transfection efficiency and overall higher FITC signal. After sorting, cells are spun down, lysed using DNA lysis buffer (Laird et al, 1991) with DTT and Proteinase K or RNA lysis buffer (Macherey-Nagel). gDNA is extracted using magnetic beads (made from FisherSci Sera-Mag SpeedBeads Carboxyl Magnetic Beads, hydrophobic according to Rohland & Reich, 2012), after overnight lysis. RNA then is extracted with Macherey-Nagel's NucleoSpin RNA Plus kit.

High-throughput Amplicon Sequencing, RT-PCR & Base Editing Data Analysis

Genomic DNA is amplified using gene-specific DNA primers flanking desired target sequence. These primers include illumina-compatible adapter-flaps. The amplicons are molecularly indexed with NEBNext Dual Index Primers (NEB) or index primers with the same or similar sequence ordered from IDT. Samples are combined into libraries and sequenced on the Illumina MiSeq machine using the MiSeq Reagent Kit v2 or Micro Kit v2 (Illumina). Sequencing results are analyzed using a batch version of the software CRISPResso 2.0 (crispresso.rocks). Reverse transcription is performed using the High Capacity RNA-to-cDNA kit (Thermo Fisher) following the manufacturer's instructions. Amplicon PCR and library preparation for Next-Generation Sequencing (NGS) off of cDNA is done as described above for gDNA. If possible, we use exon-exon junction spanning primers to exclude amplification of gDNA.

RNA-seq and Single Nucleotide Variant Calling

RNA library preparation is performed using Illumina's TruSeq Stranded Total RNA Gold Kit with initial input of -500 ng of extracted RNA per sample, using SuperScript III for first-strand synthesis (Thermo Fisher). rRNA depletion is confirmed during library preparation by fluorometric quantitation using the Qubit HS RNA kit before and after depletion (Thermo Fisher). For indexing, we use IDT-Illumina Unique Dual Indeces (Illumina). Libraries are pooled based on qPCR quantification (NEBNext Library Quant Kit for IIlumina) and loaded onto a NextSeq (at MGH Cancer Center, PE 2x150, 500/550 MidOutput Cartridge) or HiSeq2500 in High Output mode (Broad Institute, PE 2x76). Illumina fastq sequencing reads are aligned to the human hg38 reference genome with STAR (Dobin et al., 2013, PMID: 23104886) and processed with GATK best practices (McKenna et al., 2010, PMID: 20644199: DePristo et al., 2011, PMID: 21478889). RNA variants are called using HaplotypeCaller, and empirical editing efficiencies are established on PCR-de-duplicated alignment data.

Variant loci in ABE/ABEmax overexpression experiments are further required to have comparable read coverage in the corresponding control experiment (read coverage for SNV in control >90th percentile of read coverage across all SNVs in overexpression). Additionally, the above loci are required to have a consensus of at least 99% of reads calling the reference allele in control.

Protein Structure Analysis and DNA/RNA Binding Prediction

We access the crystal structures of E. coli (1Z3A, DOI: 10.2210/pdb1Z3A/pdb) and S. aureus TadA (2B3J, DOI: 10.2210/pdb2B3J/pdb; NDB: PR0180), only the latter of which is in complex with RNA, from the protein databank PDB (rcsb.org). Using the software PyMOL (Schrödinger), the two three-dimensional crystal structures are aligned and residues in these two crystal structures are analyzed regarding proximity to RNA and the modeled enzymatic pocket adenosine deamination is localized at. DNA and RNA binding is predicted using the DRNApred interface (Yan&Kurgan, NAR 2017).

Alignment of tRNA Adenosine Deaminase Homologues and Orthologues

The amino acid sequence of E. coli TadA is aligned to other tRNA adenosine deaminase homologues or orthologues using Geneious 7 software. Amino acid sequences are obtained from the uniprot platform (uniprot.org).

To test whether ABEs might be capable of editing adenines in RNA, we assessed whether this base editor fusion could edit adenines transcriptome-wide using RNA-seq. To do this, we transfected human HEK293T cells with a plasmid that expressed an ABEmax-P2A-EGFP fusion protein (the P2A sequence mediates a post-translational cleavage that releases EGFP from the ABEmax part of the fusion) (Methods). At 36 hours after transfection, we then used flow cytometry to sort out the cells with the highest (top 5%) GFP/FITC signal and isolated total RNA from these cells. As a negative control, we transfected HEK293T cells in parallel with a plasmid that expressed a bpNLS-32AAlinker-nickase Cas9 (nCas9)-bpNLS-P2A-EGFP (called nCas9-NLS below) fusion protein (i.e., a plasmid identical to the ABEmax-P2A-EGFP expression plasmid but lacking the TadA-TadA* heterodimer within the ABEmax part of the fusion protein) and also sorted these for the top 5% GFP signal and isolated total RNA. We used a gRNA targeting a genomic site in the RNF2 gene and on-target DNA base editing was high (˜70% A-to-G, data not shown). Using RNA-seq, we found that ABEmax edited tens of thousands of adenosines in RNA with high efficiency (FIG. 2A-C). These experiments were done in triplicate and each ABEmax sample was RNA-sequenced with a matching negative control. The fraction of A-to-I changes with respect to all possible substitutions was >99.5% in all replicates (Table C).

Total transcriptome-wide numbers of edited adenosines in different biological replicates Cells were transfected 18-24 h after seeding and sorted 36-40 h after transfection for top 5% FITC signal (Methods).

These edited As were distributed throughout the human genome and had considerable editing efficiencies (FIGS. 2B-C). Sequence logos derived from edited As in each of these experiments showed the high prevalence of a T preceding the edited A (FIG. 2D), which is similar to the sequence context of adenosine 34 which gets physiologically edited on tRNA by TadA. Subsequent experiments are focused on determining the impact of ABE expression levels, the use of different gRNAs as well as potential differences in off-target A-to-I RNA editing induced by ABEs in different cell lines.

Given the transcriptome-wide RNA editing induced by ABEmax, it is desirable to create variants of the adenine base editor that would diminish this unwanted activity while retaining the desired capability to perform targeted DNA base editing (RRE or Reduced RNA Editing variants). We reasoned that the introduction of mutations into the TadA-TadA* part (* marking the engineered variant of E. coli TadA of the adenine base editor) of an ABE, either in one or both monomers, might accomplish this. We also reasoned that mutations might also be introduced into the TadA* part of an ABE harboring only a single adenine deaminase domain. To identify potential positions to alter, we first aligned the structures of S. aureus TadA bound to RNA and E. coli TadA on its own (no structure is available for this TadA bound to RNA) using PyMol software. All residues in proximity of the enzymatic pocket in which A34 gets deaminated, or residues which could establish contacts with the RNA near the TadA structure were identified using this structure-guided strategy (FIG. 3) and we reasoned that these are candidate positions at which to introduce amino acid substitutions that might have an RRE phenotype. To generate additional positions that might be mutated to achieve an RRE phenotype, we used DRNApred to predict differential DNA and RNA binding of positions in the TadA-TadA* single-chain heterodimer. Using this tool, we were able to identify residues that may predominantly bind to one or the other or both types of nucleic acids (i.e., DNA or RNA) (FIG. 4). Positions predicted to predominantly bind RNA are candidate residues at which amino acid substitutions may generate adenine deaminases with an RRE phenotype (FIG. 4). Table A lists all amino acid positions identified with these two orthogonal approaches. To test these variants with amino acid substitutions at these positions for an RRE phenotype, we generate variants of ABE with mutations at the one or more of the positions listed in Table A and test them in three steps: (1) testing for on-target DNA editing capabilities in cells using gRNAs targeted to 4 genomic sites, (2) all variants that preserve high and precise on-target DNA editing capabilities are then assayed for off-target RNA editing on 4 commonly RNA-edited transcriptomic sites determined using RNA-seq data from Example 1, and (3). the variants showing most promising on/off-target profiles are assayed using deep RNA-seq to validate reduction of transcriptome-wide RNA editing activity.

Methods:

The following materials and methods were used in Example 3.

PyMOL Analysis of TadA structures. Escherichia coli tRNA-specific adenosine deaminase (TadA, PDB 1Z3A) and Staphylococcus aureus TadA with tRNA (PDB 2B3J) structures were downloaded from the Protein Data Bank and visualized with PyMOL version 2.2.2. Subunit A (monomer) of S. aureus TadA with tRNA was superimposed with subunit A of E. coli TadA using the “super” command. All figures were generated with PyMOL (Schrodinger).

Plasmid cloning. All ABE constructs (reported in Supplementary Table 1) were cloned using the backbone and the P2A-EGFP-NLS fragment of ABEmax-P2A-EGFP-N LS (Agel/NotI digest; Addgene ID 112101). ABEmax and variants were expressed under the control of a pCMV promoter. For the P2A-EGFP fragments in these constructs, we used BPK4335 (pCMV-BE3-P2A-EGFP) as a template. Guide RNA (gRNA) plasmids were cloned using the SpCas9 gRNA entry vector BPK1520 (pUC19 backbone; Bsmbl cassette, Addgene ID 65777). All remaining constructs were generated using isothermal amplification (Gibson assembly, NEB). All gRNA and ABE plasmids were midi or maxi prepped using the Qiagen Midi/Maxi Plus kits.

Cell culture. HEK293T cells (CRL-3216) and HepG2 cells (HB-8065) were purchased from and STR-authenticated by ATCC. Cells were cultured in Dulbecco's Modified Eagle Medium (DMEM, Gibco) supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco) and 1% (v/v) penicillin-streptomycin (Gibco) or Eagle's Minimum Essential Medium with 10% (v/v) FBS and 0.5% (v/v) penicillin. Cells were passaged every 2-3 days when reaching around 80-90% confluency. Both cell lines were used only until passage 20 for all experiments, and the media was tested every two weeks for mycoplasma.

Transfections. For ABE DNA on-target screening experiments, 2×10⁴HEK293T cells were seeded into 96-well Flat Bottom Cell Culture plates (Corning), transfected 24 h post seeding with 165 ng base editor or negative control (bpNLS-32AA linker-nCas9(D10A)-bpNLS), 55 ng guide RNA expression plasmid, and 0.66 μL TransIT-293 (Mirus), and harvested 72 h after transfection for DNA. For ABE RNA off-target screening experiments, 2×10⁵HEK293T cells were seeded into 12-well Cell Culture plates (Corning), transfected 24 h post seeding with 1.65 μg base editor or negative control, 0.55 μg guide RNA, and 6.6 μL TransIT-293, and harvested 36 h after transfection for RNA. For experiments with FACS-sorted cells, 6.5-7×10⁶HEK293T cells were seeded into 150 mm Cell Culture dishes (Corning), transfected 24 h post seeding with 37.5 μg base editor or an appropriate negative control fused to P2A-EGFP, 12.5 μg guide RNA, and 150 μL TransIT-293. Sorting took place 36-40 h post transfection.

Fluorescence-activated cell sorting (FACS). Cells were prepared for sorting by diluting to 1×10⁷cells per ml with 1× Phosphate Buffer Saline (PBS, Corning) supplemented with 10% FBS and filtering through 35 μm cell strainer caps (Corning). Cells were sorted on a FACSAria II (BD Biosciences) using FACSDiva version 6.1.3 (BD Biosciences) after gating for single live cells. Cells treated with base editor were sorted for either all GFP signal (standard expression) or top 5% of cells with the highest GFP (FITC) signal (overexpression) into FBS; cells treated with nCas9 negative controls were sorted for either all GFP positive cells or the 5% of cells with a mean fluorescence intensity (MFI) matching that of the top 5% of cells treated with base editor.

DNA extraction. For ABE DNA on-target experiments, cells were lysed for DNA 72 h post-transfection with freshly prepared 43.5 μL DNA lysis buffer (50 mM Tris HCl pH 8.0, 100 mM NaCl, 5 mM EDTA, 0.05% SDS, adapted from ref. 15), 5.25 μL Proteinase K (NEB), and 1.25 μL 1M DTT (Sigma). For experiments with sorted cells, cells were centrifuged (200 g, 8 min) and lysed with 174 μL DNA lysis buffer, 21 μL Proteinase K, and 5 μL 1M DTT. Lysates were incubated at 55° C. on a plate shaker overnight, then gDNA were extracted with 2× paramagnetic beads (as described in ref. 16), washed 3 times with 70% EtOH, and eluted in 30 μL 0.1× EB buffer (Qiagen).

RNA extraction & reverse transcription. Cells were lysed for RNA 36 h-40 h post-transfection with 350 μL RNA lysis buffer LBP (Macherey-Nagel), and RNA were extracted with the NucleoSpin RNA Plus kit (Macherey-Nagel) following the manufacturer's instructions. RNA was then reverse transcribed into cDNA with the High Capacity RNA-to-cDNA kit (Thermo Fisher) following the manufacturer's instructions.

Library preparation for DNA or cDNA targeted amplicon sequencing. Next-generation sequencing (NGS) of DNA or cDNA was performed as previously described⁵. In summary, the first PCR was performed to amplify genomic or transcriptomic sites of interested with primers containing Illumina forward and reverse adapter sequences (see FIG. 6G, Supplementary Table 2 for primers and amplicons used in this study), following NEB Phusion High-Fidelity DNA Polymerase instructions. The first PCR products were cleaned with 0.7× paramagnetic beads, then the second PCR was performed to add barcodes with primers containing unique sets of p5/p7 Illumina barcodes (analogous to TruSeq CD indexes). The second PCR products were again cleaned with 0.7× paramagnetic beads. The libraries were then pooled based on concentrations measured with the QuantiFluor dsDNA System (Promega) and Synergy HT microplate reader (BioTek) at 485/528 nm. The final pool was quantified by qPCR with the NEBNext Library Quant Kit for Illumina (NEB) and sequenced paired-end (PE) 2×150 on the Illumina MiSeq machine using 300-cycle MiSeq Reagent Kit v2 or Micro Kit v2 (Illumina). FASTQs (post-demultiplexing) were downloaded from Illumina BaseSpace and analyzed using a batch version of CRISPResso 2.

RNA library preparation & sequencing. RNA-seq experiments were performed as previously described⁵. Briefly, RNA libraries were prepared with the TruSeq Stranded Total RNA Library Prep Gold kit (Illumina) following the manufacturer's instructions. SuperScript III (Invitrogen) was used for first-strand synthesis, and IDT for Illumina TruSeq RNA unique dual indexes (96 indexes) were used to avoid index hopping. The libraries were pooled based on qPCR measurements with the NEBNext Library Quant Kit for Illumina. The final pool was sequenced PE 2×76 on the Illumina HiSeq2500 machine (for the ABE experiment shown in FIG. 6b) or PE 2×100 on the NovaSeq6000 machine (for all remaining ABE experiments) at the Broad Institute of Harvard and MIT (Cambridge, Mass.). To account for variable sequencing depths, all RNA-seq libraries sequenced on the NovaSeq were uniformly downsampled to 100 million reads per library using seqtk version 1.0-r82-dirty (github.com/lh3/seqtk).

Amplicon sequencing analysis. Amplicon sequencing data was analyzed with CRISPResso2 v.2.0.27¹⁷. The heatmaps for the SECURE-ABE screening in FIG. 6f display the highest edited adenine at the target site. Editing efficiency values were log₂transformed with a pseudocount of 1, averaged over quadruplicates, and normalized to ABEmax. The remaining heatmaps showing ABE on-target DNA editing (FIG. 7 and FIG. 9) show an editing window that includes the edited Cs and a grey background for editing efficiencies smaller than 2%.

RNA Variant Calling Pipeline

All bioinformatic analysis was performed in concordance with GATK Best Practices^18,19for RNA-seq mutation calling as we have previously described⁵. Briefly, raw sequencing reads were two-pass aligned to the reference hg38 reference genome with STAR²° with parameters to discard multi-mapping reads. After PCR duplicate removal and base recalibration, mutations in RNA-seq libraries were called using GATK HaplotypeCaller. RNA edits in ABE overexpression experiments were identified using a downstream modification of the GATK pipeline output as we have previously described⁵. Specifically, mutation positions called by HaplotypeCaller were further filtered to include only those satisfying the following criteria with reference to the corresponding control experiments: (1) Read coverage for a given edit in control experiment should be greater than the 90th percentile of read coverage across all edits in the overexpression experiment. (2) 99% of reads covering each edit in the control experiment were required to contain the reference allele. Edits were further filtered to exclude those with fewer than 10 reads or 0% alternate allele frequencies. A-G edits include A-G edits identified on the positive strand as well as T-C edits identified on the negative strand.

Six A-to-I edits identified from the above pipeline were chosen to test SECURE ABE variants based on the following criteria. These were sites that had (1) read coverage of at least 50 in all replicates of control and overexpression experiments, (2) 99% reads in all control experiments containing reference allele and (3) at least 60% alternate allele frequencies in all replicates. From this list, primers were tested for the top 15 edited sites that were also within 150 bases of an exon-exon junction and the 6 highest edited sites with robust amplification from cDNA were chosen.

To identify self-edits occurring on the base-editing construct, we generated a modified hg38 reference genome with additional contigs for the gRNA and base editor constructs. These additional contigs were appended to the reference genome, and each library was re-processed using GATK best practices, including variant calling with HaplotypeCaller. Variants were then further filtered using a similar process as described above for the transcriptome (i.e. filtering for no more than 1% editing in the negative control) with the exception that positions poorly covered in the control due to differences in the construct design (i.e. the deaminase domain) were not filtered out. We note that since both control and BE constructs were expressed from plasmids, the overall expression of these transcripts is much higher than most detected genes which supersedes the control of coverage between control and BE expression in this analysis (see part 1 of transcriptome variant calling above). Editing efficiencies per position were computed based on the abundance of Gs (ABE) over total coverage from bam-readcount estimated on the PCR deduplicated .bam files. Edits were further filtered to exclude those with fewer than 50 reads or 0% alternate allele frequencies.

Results

To engineer SECURE-ABE variants, we first used a protein truncation strategy to reduce the RNA recognition capability of the widely used ABEmax fusion. ABEmax harbors a single-chain heterodimer of the wild type (WT) E. coli TadA adenosine deaminase monomer (which deaminates adenines on tRNA) fused to an engineered E. coli TadA monomer that was modified by directed evolution to deaminate DNA adenines^3,6,7(FIG. 6a). We hypothesized that the WT TadA monomer should still be capable of recognizing its tRNA substrate and therefore might recruit ABEmax to deaminate RNA adenines that lie in the same or a similar sequence motif to that present in the tRNA. Consistent with this idea, a re-analysis of our previously published RNA-seq data⁵revealed that adenines that are edited at the highest efficiencies (80-100%) are embedded in a more extended CUACGAA motif, which contrasts to the shorter UA sequence observed across all edits (FIG. 6b). Importantly, the CUACGAA motif is a perfect match to the sequence surrounding the adenine deaminated in the tRNA substrate of the WT E. coli TadA enzyme⁶. We reasoned that removing the WT TadA domain from ABEmax might reduce its RNA editing activity and we therefore generated a smaller ABEmax variant lacking this domain that we refer to as miniABEmax (FIG. 6a).

We used RNA-seq to compare the transcriptome-wide off-target RNA editing activities of miniABEmax to ABEmax in HEK293T cells. Both editors and a nickase Cas9 (nCas9) control were each assayed in biological triplicate with three different gRNAs: two targeted to endogenous human gene sites (HEK site 2 and ABE site 16)³and one to a site that does not occur in the human genome (NT)⁵. We performed these studies by sorting for GFP-positive cells (ABEmax was expressed as a P2A fusion with the base editor or nCas9 (Methods)). As an internal control, we first confirmed that ABEmax and miniABEmax both induced comparable on-target DNA editing efficiencies with HEK site 2 and ABE site 16 gRNAs (FIG. 9). Edited RNA adenines were identified as previously described⁵by filtering out background editing observed with read-count-matched negative controls (Methods). Surprisingly, the total number of edited adenines induced with miniABEmax expression was not consistently lower than what we observed with ABEmax—the two editors induced on average 80-fold and 54-fold more edited adenines relative to background (determined with a GFP-only negative control) (FIG. 6c and Extended Data Table 1). However, the overall distribution of individual RNA adenine edit efficiencies induced by miniABEmax were generally shifted to lower values (FIG. 6d and FIG. 10). In addition, the sequence logos of the adenines edited by miniABEmax now appear to be shorter GUA or UA motifs, in contrast to the more extended CUACGAA motif characteristic of ABEmax (FIGS. 11a and 11b).

We reasoned we might further reduce the off-target RNA editing activity of miniABEmax by altering amino acid residues within the remaining engineered E. coli TadA domain that could potentially mediate RNA recognition. However, although a crystal structure of isolated E. coli TadA has previously been solved⁸(PDB 1Z3A; FIG. 6e), no structural information was available to delineate how this protein might recognize its RNA substrate. To overcome this, we exploited the availability of a S. aureus TadA-tRNA co-crystal structure⁹(PDB 2B3J) (FIG. 6e and Methods). Although E. coli and S. aureus TadA share only partial amino acid sequence homology (39.5% identity; FIG. 12a), we found that these two proteins share a high degree of structural homology (FIG. 6e). This similarity enabled us to overlay the two structures and thereby to infer 31 amino acid residues in E. coli TadA that likely lie near the enzymatic pocket around the substrate tRNA (FIG. 6e). In addition, we mutated three positively charged residues (R13, K20, and R21) in TadA* that we imagined might make contacts to the phosphate backbone of an RNA substrate.

We generated a total of 34 miniABEmax variants bearing various substitutions at the amino acid positions described above and screened each editor for on-target DNA editing and off-target RNA editing activities in HEK293T cells. To assess on-target DNA editing, we examined the efficiencies of A-to-G edits induced by each of the 34 variants with four gRNAs targeted to different endogenous gene sequences. To screen for off-target RNA editing activities, we quantified editing by each of the 34 variants at six RNA adenines using standard plasmid expression conditions (i.e., without sorting for GFP expression; see Methods); these six adenines were previously identified as being highly edited with ABEmax overexpression in HEK293T cells⁸. These experiments revealed that 23 of the 34 variants induced robust on-target DNA editing at least comparable to that observed with miniABEmax and ABEmax (FIG. 6f and FIG. 12b). In addition, 14 of the 34 variants showed reduced editing activities on at least three of the six RNA adenines we examined relative to that observed with miniABEmax (FIG. 6f, FIG. 12b). Importantly, three of the 34 variants (miniABEmax-K20A/R21A, -K20A/R21A/R23A, and -V82G) showed both robust on-target DNA editing activity and substantially reduced RNA editing activities (FIG. 6f and FIG. 12b and Extended Data Table 2). Based on their DNA/RNA editing ratios, we chose to carry forward two miniABEmax variants (K20A/R21A and V82G) for more comprehensive characterization.

We characterized the transcriptome-wide off-target RNA editing profiles of the miniABEmax K20A/R21A and V82G variants using RNA-seq. The two variants were assessed in biological triplicate with the HEK site 2, ABE site 16, and NT gRNAs. In contrast to what we observed with miniABEmax, the K20A/R21A and V82G variants both induced substantially reduced numbers of edited adenines relative to ABEmax but still approximately four-fold and three-fold higher numbers, respectively, than background (determined with the GFP-only negative control) (FIG. 6c). In addition, the distribution of individual RNA adenine editing efficiencies for the two variants was shifted lower with both variants relative to ABEmax and miniABEmax (FIG. 6d and FIG. 10). Overall, these results demonstrate the feasibility of separating unwanted off-target RNA editing from desired on-target DNA editing activities with an ABE.

Finally, given their abilities to edit RNA transcripts, we wondered whether ABEs might also self-edit their own transcripts, thereby potentially generating a set of heterogeneous base editor proteins. To assess this, we applied our analysis pipeline to quantify self-edit events in our previously published RNA-seq data⁵obtained with BE3 expressed at standard or overexpression levels in HEK293T cells. These data showed ABEmax and miniABEmax both inducing dozens (29 to 67) of A-to-I changes throughout their own transcripts with editing efficiencies ranging from 7.3% to 58.7% among replicates performed with three different gRNAs (FIG. 8a; FIG. 13; Table 1). Nearly all of the edits induced by the ABEs may induce missense mutations (Table 1). On average, 53% of adenines self-edited by ABEmax appeared to be edited across all three replicates (FIG. 8c). Comparing the unions of self-edits from different gRNAs shows 68% of overlap between edits across the three gRNAs, suggesting that self-editing is independent of the gRNA with which the ABE was co-expressed (FIG. 8). Notably, the two miniABEmax variants showed substantially reduced self-editing activities: K20A/R21A induced only small numbers (range 1 to 3) of self-edits and V82G did not induce any detectable self-edits (FIG. 3d; FIG. 13; Table 1).

To screen for additional SECURE-ABE variants with minimized unwanted RNA editing activities that maintain efficient DNA on-target editing, we engineered 30 more miniABEmax variants and assessed their DNA and RNA editing efficiencies. In this second screen, we included two SECURE-ABE variants (miniABEmax-K20A/R21A and -V82G) with reduce RNA off-target editing. DNA on-target editing was examined with four gRNAs targeted to different endogenous gene sequences (HEK site 2, ABE site 2, site 3 and site 4), and 25 out of 30 variants induced DNA editing comparable to that observed with miniABEmax and ABEmax. RNA off-target editing was examined on six RNA sites that were previously identified to be highly edited with ABEmax and were used for first round of screening, and 24 out of 30 variants showed reduced RNA editing compared to miniABEmax on all 6 sites tested. Based on both DNA and RNA editing profiles, miniABEmax-W11A, -K110A, and -D139A/E140A showed the most promising characteristics to become SECURE-ABE variants, while A58G, N72A, V106W, K110A, H128A/R129A, A138W, D139A/E140A, A142W, F148A, and R150A all showed promising reductions of RNA off-targets as well, with reasonably maintained DNA on-target editing capabilities.

1 Rees, H. A. & Liu, D. R. Base editing: precision chemistry on the genome and transcriptome of living cells. Nat Rev Genet 19, 770-788, doi:10.1038/s41576-018-0059-1 (2018).

2 Komor, A. C., Kim, Y. B., Packer, M. S., Zuris, J. A. & Liu, D. R. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature 533, 420-424, doi:10.1038/nature17946 (2016).

3 Gaudelli, N. M. et al. Programmable base editing of A*T to G*C in genomic DNA without DNA cleavage. Nature 551, 464-471, doi:10.1038/nature24644 (2017).

4 Nishida, K. et al. Targeted nucleotide editing using hybrid prokaryotic and vertebrate adaptive immune systems. Science 353, doi:10.1126/science.aaf8729 (2016).

5 Grunewald, J. et al. Transcriptome-wide off-target RNA editing induced by CRISPR-guided DNA base editors. Nature, doi:10.1038/s41586-019-1161-z (2019).

6 Wolf, J., Gerber, A. P. & Keller, W. tadA, an essential tRNA-specific adenosine deaminase from Escherichia coli. EMBO J 21, 3841-3851, doi:10.1093/emboj/cdf362 (2002).

7 Koblan, L. W. et al. Improving cytidine and adenine base editors by expression optimization and ancestral reconstruction. Nat Biotechnol 36, 843-846, doi:10.1038/nbt.4172 (2018).

8 Kim, J. et al. Structural and kinetic characterization of Escherichia coli TadA, the wobble-specific tRNA deaminase. Biochemistry 45, 6407-6416, doi:10.1021/bi0522394 (2006).

9 Losey, H. C., Ruthenburg, A. J. & Verdine, G. L. Crystal structure of Staphylococcus aureus tRNA adenosine deaminase TadA in complex with RNA. Nat Struct Mol Biol 13, 153-159, doi:10.1038/n5mb1047 (2006).

10 Wang, X. et al. Efficient base editing in methylated regions with a human APOBEC3A-Cas9 fusion. Nat Biotechnol, doi:10.1038/nbt.4198 (2018).

11 Gehrke, J. M. et al. An APOBEC3A-Cas9 base editor with minimized bystander and off-target activities. Nat Biotechnol 36, 977-982, doi:10.1038/nbt.4199 (2018).

12 Komor, A. C. et al. Improved base excision repair inhibition and bacteriophage Mu Gam protein yields C:G-to-T:A base editors with higher efficiency and product purity. Sci Adv 3, eaao4774, doi:10.1126/sciadv.aao4774 (2017).

13 Sharma, S., Patnaik, S. K., Kemer, Z. & Baysal, B. E. Transient overexpression of exogenous APOBEC3A causes C-to-U RNA editing of thousands of genes. RNA Biol 14, 603-610, doi:10.1080/15476286.2016.1184387 (2017).

14 Fritz, E. L. et al. A comprehensive analysis of the effects of the deaminase AID on the transcriptome and methylome of activated B cells. Nat Immunol 14, 749-755, doi:10.1038/ni.2616 (2013).

15 Laird, P. W. et al. Simplified mammalian DNA isolation procedure. Nucleic Acids Res 19, 4293 (1991).

16 Rohland, N. & Reich, D. Cost-effective, high-throughput DNA sequencing libraries for multiplexed target capture. Genome Res 22, 939-946, doi:10.1101/gr.128124.111 (2012).

17 Clement, K. et al. CRISPResso2 provides accurate and rapid genome editing sequence analysis. Nat Biotechnol 37, 224-226, doi:10.1038/s41587-019-0032-3 (2019).

18 McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 20, 1297-1303, doi:10.1101/gr.107524.110 (2010).

19 DePristo, M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet 43, 491-498, doi:10.1038/ng.806 (2011).

20 Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21, doi:10.1093/bioinformatics/bts635 (2013).

It is to be understood that while the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.