Abstract
RNA-guided programmable cytosine and adenine base editors are a powerful class of genome editing tool for the introduction of localized base transitions without generating a double-stranded DNA break. Base editors (BE) have an optimal window of activity relative to the PAM recognized by the Cas9 enzyme and these constructs are strand selective. Here we demonstrate that fusion of a programmable DNA-binding domain (pDBD) or another Cas9 orthologue to spCas9-BE, we can produce an RNA-programmable Cas9-BE-pDBD chimera or Cas9-BE-Cas9 chimeras with dramatically improved activities and increased targeting range. Cas9-pDBD or Cas9-Cas9 fusion base editors display an expanded targeting repertoire and achieve highly specific genome editing, which can be tailored to achieve extremely precise genome editing at nearly any genomic locus.
Claims
1. A method, comprising; a) providing; i) a nucleic acid sequence encoding at least one mutated base pair; and ii) a fusion protein comprising a Cas9/sgRNA complex, a programmable DNA binding domain (pDBD) and an adenine base editor (ABE) protein or a cytosine base editor (CBE) protein; b) contacting said fusion protein with said mutated base pair; and c) reverting said base pair to a wild type base pair.
2. The method of claim 1, wherein said programmable DNA binding domain is a zinc finger protein (ZFP), a transcription activator-like effector (TALE) protein or an orthogonal dCas9/sgRNA complex.
3. The method of claim 1, wherein said fusion protein further comprises a cytidine deaminase protein or an adenine deaminase protein.
4. The method of claim 1, wherein said at least one mutated base pair is an MECP2 gene mutation.
5. The method of claim 1, further providing a biological sample comprising said at least one mutated base pair.
6. The method of claim 5, wherein said biological sample is a human biological sample.
7. The method of claim 1, further comprising administering said fusion protein to a patient exhibiting at least one symptom of a genetic disease.
8. The method of claim 7, further comprising reducing said at least one symptom of said genetic disease with said fusion protein.
9. The method of Clam 7, wherein said genetic disease is Rett syndrome.
10. The method of claim 1, wherein said adenine base editor or said cytosine base editor hybridizes proximate to a protospacer adjacent motif (PAM) containing a single G.
11. The method of claim 1, wherein said adenine base editor or said cytosine base editor hybridizes to a protospacer adjacent motif that is non-canonical for said Cas9/sgRNA complex.
12. The method of claim 1, wherein said fusion protein is selected from the group consisting of a CBE/ABE-nSpyCas9-ZFP fusion protein, a CBE/ABE-nSpyCas9-TALE fusion protein and a CBE/ABE-nSpyCas9-dSauCas9/dNme2Cas9 fusion protein.
14. The method of claim 1, wherein said reverting comprises a base conversion activity that has a two-fold greater efficiency than a standard base editor protein lacking a pDBD.
15. A composition comprising a Cas9/sgRNA framework attached to a programmable DNA binding domain and an adenine or a cytosine base editor protein that hybridizes proximate to a single G protospacer adjacent motif containing a single G.
16. The composition of claim 15, wherein said Cas9/sgRNA framework further comprises an adenine deaminase protein or a cytidine deaminase protein.
17. The composition of claim 15, wherein said programmable DNA binding domain is a zinc finger protein or an orthogonal dCas9/sgRNA complex.
18. The composition of claim 15, wherein said Cas9 has attenuated DNA binding affinity to a protospacer adjacent motif containing a dual G.
19. An attenuated Cas9 protein comprising a PAM recognition domain having at least two amino acid substitutions, wherein said PAM recognition domain has an attenuated affinity for its cognate PAM sequence.
20. The attenuated Cas9 protein of claim 19, wherein said at least two amino acid substitutions are R1333S and K1118S.
21. The attenuated Cas9 protein of claim 19, wherein said at least two amino acid substitutions are R1335K and E1219Q.
22. The attenuated Cas9 protein of claim 19, wherein said at least two amino acid substitutions are R1333S, E1219Q and K1118S.
23. The attenuated Cas9 protein of claim 19, wherein said attenuated Cas9 protein is attached to a pDBD protein.
24. The attenuated Cas9 protein of claim 19, wherein said pDBD protein is a zinc finger protein, a transcription activator-like effector (TALE) protein or a Cas9 protein.
Description
BRIEF DESCRIPTION OF THE FIGURES
[0069] The file of this patent contains at least one drawing executed in color. Copies of this patent with color drawings will be provided by the Patent and Trademark Office upon request and payment of the necessary fee.
[0070] FIG. 1 illustrates several types of base editors (BEs) described in this disclosure. (top) Adenine or cytosine base editors (ABE or CBE) composed of an SpyCas9 nickase (nSpyCas9) fused to a programmable DNA-binding domain (pDBD). The star indicates mutations in the PAM recognition residues that render SpyCas9 recognition dependent on the attached pDBD. The red circle indicates the strand that is directly modified by base editing. (bottom) Adenine or cytosine base editors (ABE or CBE) composed of an SpyCas9 nickase fused to a nuclease dead orthogonal Cas9 (dCas9), such as Nme2Cas9.
[0071] FIG. 2 illustrates several embodiments of CBE variants based on a BE4 framework.sup.18 with modifications to the nuclear localization signal (NLS) sequences to improve their nuclear localization. Slotted into the dotted position within the top construct are the various CBE platforms that were tested.
[0072] FIG. 3 presents exemplary data showing Relative positions of the ZFP (Zif268), TALE, dSauCas9 and dNme2Cas9 binding sites in the KANK3, PLXNB2 and TGM2 loci. All of the TALE binding sites (green) are on the top strand. The Zif268 binding sites (red) in the TGM2 and PLXNB2 binding sites are on the complementary strand to the indicated regions. PAM sequences for the SauCas9 (NNGRRT) and Nme2Cas9 (NNNNCC) are indicated (brown/magenta).
[0073] FIG. 4 presents exemplary data of an aggregate heat map of CBE editing rates across 10 target sites containing an NGG PAM for the SpCas9 recognition element. The activity range scale is shown to the left of the heat map. A low level of indels (not C to T conversion events) was detected for all of the samples, the average of which is indicated in numbers on the far right side of the panel. The D1 and D2 nomenclatures indicate the two possible relative orientations of the Cas9-Cas9 binding sites, where D1 is recognition of opposite strands and D2 is recognition of the same strand. In all cases the SpCas9 recognition domain avoids overlap with the attached pDBD or orthogonal dCas9 binding site. The numbering scheme at the top indicates the position of the C relative to the PAM, where C1 is most distal from the PAM. Data are from biological triplicate experiments characterized by Illumina sequencing.
[0074] FIG. 5 presents exemplary data shown an aggregate heat map of CBE editing rates across 17 target sites containing an NGH PAM for the SpCas9 recognition element (H = A, C or T). The activity range scale is shown to the left of the heat map. A low level of indels (not C to T conversion events) was detected for all of the samples, the average of which is indicated in numbers on the far right side of the panel. The D1 and D2 nomenclatures indicate the two possible relative orientations of the Cas9-Cas9 binding sites, where D1 is recognition of opposite strands and D2 is recognition of the same strand. In all cases the SpCas9 recognition domain avoids overlap with the attached pDBD or orthogonal dCas9 binding site. The numbering scheme at the top indicates the position of the C relative to the PAM, where C1 is most distal from the PAM. Data are from biological triplicate experiments characterized by Illumina sequencing.
[0075] FIG. 6 presents exemplary data showing an aggregate heat map of CBE editing rates across 15 target sites containing an NHG PAM for the SpCas9 recognition element. The activity range scale is shown to the left of the heat map. A low level of indels (not C to T conversion events) was detected for all of the samples, the average of which is indicated in numbers on the far right side of the panel. The D1 and D2 nomenclatures indicate the two possible relative orientations of the Cas9-Cas9 binding sites, where D1 is recognition of opposite strands and D2 is recognition of the same strand. In all cases the SpCas9 recognition domain avoids overlap with the attached pDBD or orthogonal dCas9 binding site. The numbering scheme at the top indicates the position of the C relative to the PAM, where C1 is most distal from the PAM. Data are from biological triplicate experiments characterized by Illumina sequencing
[0076] FIGS. 7A and 7B present exemplary data showing activity profiles of SpyCas9 BE4 (gray bars) relative to SpCas9- dSauCas9 BE4 (A) or SpCas9-dNme2Cas9 BE4 (B) across the KANK3 locus. C to T conversion activity is indicated for 18 different target sites (TS#), where the bp number indicates the rough separation distance between the Cas9-Cas9 binding sites. Activities are shown for both the D1 and D2 orientation of Cas9-Cas9 binding sites (color indicated in panel legend). Note that the active target sites for SpCas9 BE4 are those with the NGG PAMs for SpCas9 (denoted by black dots below TS#). The black arrow indicates the presence of enhancement in base editing rates even 139 bp distant from the dSauCas9 binding site.
[0077] FIG. 8 presents exemplary data showing an activity profile of SpyCas9 BE4 (SpCas9, gray bars) relative to SpCas9-zif268 BE4 (SpCas9-Zif, red) across the KANK3 locus. C to T conversion activity is indicated for 18 different target sites (TS#), where the bp number indicates the rough separation between the Cas9-ZFP binding sites. Note that the active target sites for SpCas9 BE4 are those with the NGG PAMs for SpCas9 (denoted by black dots below TS#). The other sites contain NGH or NHG PAMs. The black arrow indicates the presence of enhancement in nSpyCas9 base editing rates 97 bp distant from the ZFP binding site.
[0078] FIG. 9 presents an illustrative schematic diagram of a SpyCas9 ABE. R-loop formation between the guide and the target sequence liberates one genomic DNA stand for base editing. A short segment of the single-strand DNA (Base conversion Window) is appropriately positioned and accessible to a fused adenine deaminase module. A to G conversion in the sequence (and T to C on the opposite strand) is driven via DNA repair by a nick introduced by Cas9 on the opposite DNA strand.
[0079] FIG. 10 presents an exemplary target site overview of five common MECP2 mutations. Local sequence surrounding common C>T pathogenic mutations in MECP2, where the position of the mutation (bold T) and the resulting amino acid change is noted. Coding strand is top strand. Bold A indicates the target base for deamination. The nearest base conversion window targetable by a standard SpyCas9 ABE (SpABE) based on the presence of an NGG PAM is indicated with a red bar. Only two of the 5 target adenines fall within this window. The position of the non-standard PAM utilized by our proposed SpCas9-DBD ABE fusion (*ABE) is underlined and indicated in brackets below the sequence [5’->3’]. All of these targets are accessible, and in all cases position the target A at the center of the window where base conversion rates are expected to be maximal.
[0080] FIG. 11 provides exemplary data showing an efficient base conversion of C to T at non-standard PAMs by SpyCas9-DBD cytidine deaminase. A SpyCas9 cytosine BE fused to a DBD was programmed with different guides (magenta boxes) to target neighboring regions of a gene to “walk” across the locus utilizing different non-standard PAMs (blue boxes). The SpyCas9-DBD BE was delivered by transient transfection into cells and after 3 days the population of cells was harvested and their genomic DNA amplified and sequenced to assess the rate of C to T conversion. All of the non-standard PAMs achieved functional C to T conversion (peaks indicated by *). Only the standard nGG PAM was functional with standard SpyCas9 BE (data not shown).
[0081] FIG. 12 presents an illustrative schematic overview of Cas9-DBD frameworks. SpyCas9 is fused to a DNA-binding domain (DBD) — either a Zinc finger protein (ZFP) or a nuclease-dead orthogonal Cas9 (dCas9) - that recognizes a neighboring sequence to the SpyCas9 target site. The DBD subunit delivers the Cas9 to the target region of the genome, which allows it to function at non-standard PAMs that have low affinity. Once R-loop formation is initiated the PAM element does not impact SpyCas9 catalytic activity. These SpyCas9-DBD systems increase the number of targetable sequences for Cas9, and can also increase the specificity of their activity within the genome.
[0082] FIG. 13 presents an illustrative schematic of a CRISPR adenine base editor reporter.
[0083] FIG. 14 presents exemplary data showing the quantification of a CRISPR adenine base editing rates at different PAM sequences.
[0084] FIG. 15 presents an illustrative schematic of a CRISPR cytosine base editor reporter
[0085] FIG. 16 presents exemplary data showing the quantification of a CRISPR cytosine base editing rates at different PAM sequences.
[0086] FIG. 17 presents illustrative schematics showing enhanced CBEs and ABEs. The dotted rectangles in the main constructs indicate the position where each Cas9 or Cas9-fusion variant was positioned in the construct. Examples of the constructs are displayed below. These examples are not exhaustive.
[0087] FIG. 18 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor.
[0088] FIG. 18A: An illustrative design of a CBE CopGFP reporter used to evaluate the activity of CBE constructs. Conversion of the H mutation to Y, red, restores green fluorescence; Spacer sequence, underlined; PAM, blue. Target sequence, red.
[0089] FIG. 18B: Quantification of CBE efficacy by calculating % GFP.sup.+ cells. Base editors used were nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4. The “n” prefix before the name of the SpCas9 version indicates the D10A nickase. Negative = no DNA control.
[0090] FIG. 19 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor having a TGT PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4 in HEK293T cells. Data is dispalyed as a heatmap of C-to-T editing frequencies induced by enhanced CBE systems at KANK3 TS1 (PAM = TGT). Intensity of square reflects the mean of three independent biological replicates.
[0091] FIG. 20 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor having a AGT PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4 in HEK293T cells. Data is displayed as a heatmap of C-to-T editing frequencies induced by enhanced CBE systems at KANK3 TS2 (PAM = ATG). Intensity of square reflects the mean of three independent biological replicates.
[0092] FIG. 21 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor having a TGA PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4 in HEK293T cells. Data is displayed as a heatmap of C-to-T editing frequencies induced by enhanced CBE systems at KANK3 TS3 (PAM = TGA). Intensity of square reflects the mean of three independent biological replicates.
[0093] FIG. 22 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor having a GTG PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4 in HEK293T cells. Data is displayed as a heatmap of C-to-T editing frequencies induced by enhanced CBE systems at KANK3 TS4 (PAM = GTG). Intensity of square reflects the mean of three independent biological replicates.
[0094] FIG. 23 presents exemplary data showing the improved activity of an enhanced Cas9/cytosine base editor having a GGG PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with BE4 in HEK293T cells. Data is displayed as a heatmap of C-to-T editing frequencies induced by enhanced CBE systems at KANK3 TS5 (PAM = GGG). Intensity of square reflects the mean of three independent biological replicates.
[0095] FIG. 24 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor.
[0096] FIG. 24A: An illustrative design of an ABE mCherry reporter used to evaluate the activity of each ABE construct. STOP codon, red, conversion to G1n codon restores mCherry signal; Spacer sequence, underlined; PAM, blue, target site, red, is on the complementary strand.
[0097] FIG. 24B: Quantification of ABE efficacy by calculating % mCherry+ cells. Base editors used were nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused to ABEmax. The “n” prefix before the name indicates the D10A nickase. Negative, no DNA control.
[0098] FIG. 25 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having a TGT PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Data is displayed as a heatmap of A-to-G editing frequencies induced by enhanced ABE systems at KANK3 TS1 (PAM = TGT). Intensity of square reflects the mean of three independent biological replicates.
[0099] FIG. 26 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an ATG PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Data is displayed as a heatmap of A-to-G editing frequencies induced by enhanced ABE systems at KANK3 TS2 (PAM = ATG). Intensity of square reflects the mean of three independent biological replicates.
[0100] FIG. 27 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an TGA PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Data is displayed as a heatmap of A-to-G editing frequencies induced by enhanced ABE systems at KANK3 TS3 (PAM = TGA). Intensity of square reflects the mean of three independent biological replicates.
[0101] FIG. 28 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an GTG PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Data is displayed as a heatmap of A-to-G editing frequencies induced by enhanced ABE systems at KANK3 TS4 (PAM = GTG). Intensity of square reflects the mean of three independent biological replicates.
[0102] FIG. 29 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an GGG PAM sequence. Base editing was performed with nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Data is displayed as a heatmap of A-to-G editing frequencies induced by enhanced ABE systems at KANK3 TS5 (PAM = GGG). Intensity of square reflects the mean of three independent biological replicates.
[0103] FIG. 30 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an NGG PAM sequence. Data is displayed as a heatmap depicting the summary of the base editing frequency at each adenine in the spacer region for ten guide RNAs targeting sites with NGG PAMs using nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Values and intensity of square reflect the mean of three independent biological replicates.
[0104] FIG. 31 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an NGH PAM sequence. Data is displayed as a heatmap depicting the summary of the base editing frequency at each adenine in the spacer region for 14 guide RNAs targeting sites with NGH PAMs using nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Values and intensity of square reflect the mean of three independent biological replicates.
[0105] FIG. 32 presents exemplary data showing the improved activity of an enhanced Cas9/adenine base editor having an NHG PAM sequence. Data is displayed as a heatmap depicting the summary of the base editing frequency at each adenine in the spacer region for 14 guide RNAs targeting sites with NHG PAMs using nSpCas9, nSpCas9-NG, nxCas9, nSpCas9-Zif268, nSpCas9-TALE, nSpCas9-dSaCas9, and nSpCas9-dNme2Cas9 fused with ABE7.10 in HEK293T cells. Values and intensity of square reflect the mean of three independent biological replicates.
[0106] FIG. 33 presents exemplary constructs of attenuated Cas9 proteins. The schematic shows enhanced CBEs and ABEs comprising an SpCas9 with multiple amino acid substitutions to further attenuate cognate cleavage activity in the absence of a fused DNA targeting unit such as a ZFP, as opposed to a wild type SpCas9 protein or a single amino acid substituted SpCas9 protein (e.g., R1333S or R1335S). The dotted rectangles in the main constructs indicate the position where each Cas9 or Cas9-fusion variant was positioned in the construct. Examples of the constructs are displayed below. These examples are not exhaustive.
[0107] FIG. 34 presents exemplary data showing the improvement in non-cognate base editing subsequent to attachment of a pDBD (e.g., ZFP) to a Cas9 protein. The data compares adenine base editing frequency between wild type nSpCas9, attenuated nSpCas9.sup.R1333S,K1118S and attenuated nSpCas9.sup.R1335K,E1219Q fused to the TadA8e domain with (+) or without (-) Zif268 targeting KANK3 TS1-TS5. The PAM for each target site is indicated above each set of bars.
[0108] FIG. 35 presents data demonstrating the dependence of the attenuated Cas9 base editor on the attached pDBD for target site editing. The activity of the nSpCas9.sup.R1335K,E1219Q,K1118S fused to the TadA8e domain was tested with and without the pDBD (zinc finger protein Zif268) at the KANK3 locus. In the absence of the pDBD, sanger sequencing of the genomic DNA of the population of treated cells indicates that there was minimal conversion of the adenines on the complementary strand (positions highlighted by red boxes), which would be read out as T to C conversion on the sequenced strand. In the presence of the Zif268 fusion, the SpCas9 ABE causes base conversion at two of the three highlighted positions. Numeral grad below indicates the estimate of each base at each DNA position based on the chromatogram. Control = untreated genomic DNA.
DETAILED DESCRIPTION OF THE INVENTION
[0109] The present invention is related to the field of gene editing. The use of the presently disclosed accessory pDBD and/or orthogonal Cas9 systems enhances gene editing rates and the position of editing within a target sequence. The improved CRISPR platform provides an efficient conversion of the target base, and for limiting the rate of “bystander” conversion of bases that would be undesirable, which could create unwanted mutations. These disclosed fusion systems should also allow higher specificity for the base editing process, such as reduced off-target editing.
I. Cytosine And Adenine Base Editor Proteins
[0110] Genome editing systems have been developed from these systems were recently described: cytosine.sup.1,2 and adenine.sup.3 base editors. These systems allow the conversion of cytosine to thymine or adenine to guanine within the DNA. These base editor systems can be used to revert point mutations.sup.4, introduce stop codons.sup.5, disrupt splicing sequences.sup.6, all of which can be used for therapeutic applications. One challenge with the current Cas9 base editing systems is the necessity to have a complementary PAM at the correct position and on the appropriate DNA strand to target the activity of the cytosine or adenosine base editors to precise genomic positions that are targeted for conversion, as base editors usually are strand-specific with regards to their activity. Consequently, there have been substantial efforts to broaden the targeting specificity of SpyCas9 through mutations that increase the number of PAMs that can be recognized. Two of the most prominent modified versions of Cas9 are xCas97 and Cas9-NG8 , both of which permit targeting some additional PAM elements.
[0111] A new class of genome editing systems developed from CRISPR/Cas9 systems were recently described: cytosine (Komor, et. al. 2016 (PMID 27096365) and Nishida, et al. 2016 (PMID 27492474)) and adenine (Gaudelli, et al. 2017 (PMID 29160308) base editors (CBE/ABE). These base editors typically contain two components: the adenine or cytidine deaminase and the Cas9/sgRNA complex (or Cas12a/crRNA complex), where the Cas9 component is mutated so that it cannot produce a double-strand break. Typically the Cas9 component will be a strand specific nickase (e.g. D10A mutant of SpyCas9). These systems allow the strand-specific conversion of cytosine to uracil or adenine to guanine within the DNA (Huang, et. al. 2021 (PMID 33462442)). These base editor systems can be used to revert point mutations, introduce stop codons, disrupt splicing sequences, all of which can be valuable for therapeutic applications.
[0112] In one embodiment, the present invention contemplates a Cas9-base editing platform that has a much broader targeting range for PAM recognition than the standard SpyCas9 systems. For example, the Cas9-base editing platform hybridizes proximate to a single G (NGN or NNG) rather than two Gs as in traditional NGG SpyCas9 PAM motifs. FIG. 1. This was achieved by appending programmable DNA-binding domains (pDBD).sup.9, such as zinc finger proteins (ZFP).sup.10 or TALE domains.sup.11, or an orthogonal dCas9.sup.12. Orthogonal Cas9 variants (e.g., Nme2Cas9) recognize C- rich PAM motifs and work with these same fusion strategies.sup.13. This platform allows nucleic acid sequence targeting almost anywhere in in the genome on either strand, dramatically expanding the number of disease-causing mutations that can potentially be corrected via base editors.
[0113] Since these base editing systems are not dependent on a specific stage of the cell cycle for function, and require no accessory elements beside the programmed guide RNA, they are able to function efficiently in post-mitotic cells.sup.14. Other favorable aspects of Cas9-pDBD or Cas9-Cas9 fusion systems includes, but is not limited to, the attenuation of the PAM recognition binding affinity of SpCas9 to render DNA recognition dependent on the associated pDBD or nuclease dead orthogonal Cas9. It has been reported that the SpyCas9 nuclease can dramatically improve the specificity of the Cas9 nuclease.sup.9,12. Since base editors can produce off-target DNA editing at near cognate target sequences within a genome.sup.7,15-17, the ability to limit the activity of the base editor to the DNA target sequence can provide many advantages that are compatible with the presently disclosed adenine-cytosine based editing Cas9 fusion systems.
[0114] SpyCas9 base editors have been developed that facilitate the site-specific transition of cytosine to thymine (C to T) or adenine to guanine (A to G, which achieves T to C) within a specific genomic locus. 18, 39-40 The SpyCas9 base editing systems are believed to achieve base conversion by delivering a cytosine or adenine deaminase module to a specific genomic region where they can act on the single-stranded DNA region that is created upon Cas9 R-loop formation with its target sequence. FIG. 9. Fixation of the mutation within the genome is facilitated through the generation of a nick in the non- edited DNA strand..sup.39 These base editor systems are functional in vivo in post-mitotic cells,.sup.41 and do not require the production of a double strand break (DSB) to institute sequence modification, which mitigates the production of some forms of collateral DNA damage associated with nuclease-based DSB generation..sup.42 These conventional CBE and ABE gene editors have a primary disadvantage of not being validated as a base editor for each specific mutation of interest.
II. CRISPR CBE and ABE Gene Editing Platforms
[0115] CRISPR-Cas9-based genome editing systems have revolutionized genome editing approaches and are now being leveraged for a broad range of commercial and therapeutic applications. The present invention contemplates embodiments comprising a CRISPR platform integrated with CBE and/or ABE gene editing platforms comprising an enhanced activity and targeting range as compared to other previously reported CRISPR systems.
[0116] In one embodiment, the present invention contemplates compositions comprising a cytosine base editing (CBE) and/or an adenine base editing (ABE) platform including, but not limited to, CBE/ABE-nSpyCas9-ZFP fusions, CBE/ABE-nSpyCas9-TALE and CBE/ABE-nSpyCas9-dSauCas9/dNme2Cas9 frameworks. Although it is not necessary to understand the mechanism of an invention, it is believed that such CBE and ABE platforms can be used for efficient and specific base conversion in a variety of sequence contexts. The data included herein demonstrate the successful creation of robust CBE platforms in the CBE-nSpyCas9-ZFP fusions, CBE-nSpyCas9-TALE and CBE-nSpyCas9- dSauCas9/dNme2Cas9 frameworks that can target a far broader range of DNA sequences with higher efficiency than existing frameworks (e.g., SpyCas9, xCas9 or Cas9-NG). In one embodiment, the present invention contemplates a method for targeting disease alleles in patient-derived cell lines to examine the potential clinical efficacy of these systems with the presently disclosed CBE and ABE base editing CRISPR platforms. Although it is not necessary to understand the mechanism of an invention, it is believed that the presently disclosed CBE and ABE base editing CRISPR platforms may provide a therapeutic application for efficient base conversion in target tissue containing a pathogenic point mutation.
[0117] In one embodiment, the present invention contemplates a composition comprising a Cas9/sgRNA framework comprising a pDBD protein or a second Cas9 fusion protein integrated as an adenine base editor or a cytosine base editor (e.g., a BE4-based cytosine base editor.sup.18), wherein said base editor hybridizes proximate to a single G protospacer adjacent motif. See, FIG. 2. In one embodiment, a wild- type SpyCas9 nickase (nSpyCas9) is used to create a fusion protein. The activity of these constructs were tested across nucleic acid loci (e.g., KANK3, PLNXB2 & TGM2) spanning 42 different target sites, where these target sites contained a variety of NGG, NGH and NHG PAMs (H = A, C or T) for the SpyCas9 recognition module.
[0118] This BE4-based cytosine base editor framework was compared to the nickase versions of the “wild-type” SpyCas9, xCas9.sup.7 and Cas9-NG.sup.8. These latter two systems have been shown to facilitate the recognition of a broader set of PAMs beyond the standard NGG PAM for SpyCas9. In one embodiment, the present invention contemplates a CBE-Cas9 framework comprising a zinc finger within the Cas9-ZFP fusion system. In one embodiment, the zinc finger is employing Zif268, which contains three zinc fingers and has a well defined 10 bp recognition motif.sup.20 that is present in all three of the target loci.sup.19. For the Cas9-TALE constructs, an artificial TALE domain was generated for each tested nucleic acid loci using “golden gate” assembly methods.sup.21.
[0119] FIG. 3. In one embodiment, the present invention contemplates a composition comprising a Cas9-Cas9 fusion construct comprising an orthogonal nuclease-dead Cas9 (dCas9) with an sgRNA that is specific for each locus to anchor binding of the SpyCas9 nickase within the target locus. In one embodiment, the dCas9 comprises SauCas9.sup.22 or Nme2Cas9.sup.13. The presently disclosed data was performed in HEK293T cells by transient transfection of expression plasmids, with Illumina deep sequencing of PCR amplicons spanning the target site used for quantification of the editing rates.
[0120] At canonical NGG PAM target sites all of the CBEs are functional, although the Cas9-pDBD and Cas9-Cas9 fusion proteins outperform the single Cas9 constructs (SpCas9, SpCas9-NG and xCas9) in most instances even at canonical target sites. FIG. 4. The Cas9- ZFP fusions and the Cas9-TALE fusions perform particularly well with regards to achieving higher base conversion activity in this assay. At non-canonical NGH PAM target sites the “wild-type” SpCas9 BE4 construct displays little activity. FIG. 5. The SpCas9-NG and xCas9 display modest activity, with the Cas9-NG construct proving to be the most robust of these two. The Cas9-Cas9 fusion proteins outperform the single Cas9 constructs in most instances - in particular for the D1 orientation of the target sites. The Cas9-ZFP fusions and the Cas9- TALE fusions perform particularly well with regards to higher base conversion activity at the NGH PAM target sites.
[0121] At non-canonical NHG PAM target sites, Cas9 base editor variants display little activity (e.g., SpCas9, SpCas9-NG and xCas9). FIG. 6. Cas9-Cas9 fusion proteins provided favorable activity, in particular, for the D1 orientation of the target sites. The Cas9-ZFP fusions and the Cas9-TALE fusions perform particularly well with regards to high base conversion activity at the NGH PAM target sites. The forty two target sites that were chosen across the three genomic loci also provide information on the proximity of the binding sites of the pDBD or dCas9 to the linked nSpyCas9 base editor with regards to the enhancement in activity. The data for the nSpyCas9-dSauCas9 or the nSpyCas9-dNme2Cas9 across target sites within the KANK3 locus show that there is appreciable enhancement in activity for the Cas9-Cas9 fusions relative to the SpCas9 BE4 for binding sites that have up to 139 bp distance in separation. FIG. 7. Thus, the separation between binding sites where enhancement can be achieved may be similar to the Cas9-Cas9 nuclease platform, which is on the order of 200 bp between the target sequences.sup.12.
[0122] A similar picture emerges for the analysis for the base editing activity of the nSpyCas9-ZFP BE4 construct relative to the SpCas9 BE4 across target sites within the KANK3 locus. The data for the nSpyCas9-ZFP BE4 shows that there is appreciable enhancement in activity for binding sites that have ~ 100 bp distance in separation. FIG. 8. Thus, the enhancement which can be achieved for the nSpyCas9-ZFP system appears to be more modest than the Cas9-Cas9 BE4 system, but the enhancement in base editing activity for the nSpyCas9-ZFP BE4 is more robust than for the Cas9-Cas9 BE4 system. Thus the Cas9-pDBD and Cas9-Cas9 cytosine base editors have a broader targeting range than any of the published Cas9 variant systems and also achieve higher base editing activity. In one embodiment, these frameworks further comprise adenine base editor systems.
[0123] A single copy cytosine base editor reporter (CBE reporter) transgene was generated in HEK293T cells to evaluate the efficiency of cytosine base editors (CBE) that target different PAM sequences. FIG. 13. This transgene contains a single C to T mutation that converts a Tyrosine (TAC) to a Histidine (CAC). The resulting reporter fluoresces blue (CFP). Conversion of the codon back to TAC shifts the emission wavelength to green (GFP). On the coding strand are denoted three different SpCas9 target sequences: one sequence with an optimal PAM [NGG], and two sequences shifted by a single base pair that harbor suboptimal PAMs [NGC or NCG]. Also denoted are neighboring binding sites for other Cas9 orthologs [SauCas9 or Nme2Cas9] that can be utilized as nuclease-dead modules in the context of SpyCas9-dSau/dNme2Cas9 cytosine base editors to localize them to the target site. Base conversion of cytosine to uracil (thymine analog) on the complementary strand will revert the CAC codon to TAC to change the color of the cells from blue to green, which permits a sensitive measure of the base editing rates. Utilizing the adenine base editor reporter (ABE reporter) HEK293T line the efficiency of three different adenine base editor (ABE) constructs were evaluated: 1) SpCas9 ABE, 2) SpCas9-dSaCas9 ABE, and 3) SpCas9-dNme2Cas9 ABE. FIG. 14. These were programmed with three different guide RNAs compatible with three different PAMs for SpCas9 gGG, gGT and tAG -the latter two of which are suboptimal. For the SpCas9-dSaCas9 or SpCas9-dNme2Cas9 ABEs additional guide RNAs were included to target the nuclease-dead orthogonal Cas9 to the indicated binding site in the ABE reporter sequence. FIG. 13. ABEs and their guides were delivered as expression plasmids by transient transfection (800 ng ABE vectors and 200 ng sgRNAs, 150k cells). Adenine conversion rate within the reporter cells was determined by FACS analysis based on the fraction of mCherry positive cells after 3 days. All three ABEs efficiently utilized the NGG PAM to correct the C to T mutation in the ABE reporter. However, only the SpCas9-dSa/dNme2Cas9 ABEs were able to efficiently utilize the NGT or NAG PAMs to achieve reporter correction.
[0124] A single copy cytosine base editor reporter (CBE reporter) transgene was generated in HEK293T cells to evaluate the efficiency of cytosine base editors (CBE) that target different PAM sequences. FIG. 15. This transgene contains a single C to T mutation that converts a Tyrosine (TAC) to a Histidine (CAC). The resulting reporter fluoresces blue (CFP). Conversion of the codon back to TAC shifts the emission wavelength to green (GFP). On the coding strand are denoted three different SpCas9 target sequences: one sequence with an optimal PAM [NGG], and two sequences shifted by a single base pair that harbor suboptimal PAMs [NGC or NCG]. Also denoted are neighboring binding sites for other Cas9 orthologs (e.g., SauCas9 or Nme2Cas9) that can be utilized as nuclease-dead modules in the context of SpyCas9-dSau/dNme2Cas9 cytosine base editors to localize them to the target site. Base conversion of cytosine to uracil (thymine analog) on the complementary strand will revert the CAC codon to TAC to change the color of the cells from blue to green, which permits a sensitive measure of the base editing rates. Utilizing the cytosine base editor reporter (CBE reporter) HEK293T line the efficiency of three different cytosine base editor (CBE) constructs were evaluated: 1) SpCas9 CBE, 2) SpCas9-dSaCas9 CBE, and 3) SpCas9-dNme2Cas9 CBE. FIG. 16. These were programmed with three different guide RNAs compatible with three different PAMs for SpCas9 cGG, gGC and tCG - the latter two of which are suboptimal. For the SpCas9-dSaCas9 or SpCas9-dNme2Cas9 CBEs additional guide RNAs were included to target the nuclease-dead orthogonal Cas9 to the indicated binding site in the CBE reporter sequence. FIG. 15. CBEs and their guides were delivered as expression plasmids by transient transfection (800 ng CBE vectors and 200 ng sgRNAs, 150k cells). Cytosine conversion rate within the reporter cells was determined by FACS analysis based on the fraction of GFP positive cells after 3 days. All three CBEs efficiently utilized the NGG PAM to correct the T to C mutation in the CBE reporter. However, only the SpCas9-dSa/dNme2Cas9 CBEs were able to efficiently utilize the NGC or NCG PAMs to achieve reporter correction.
III. MECP2 Gene Base-Editing Strategies To Treat Rett Syndrome
[0125] In one embodiment, the present invention contemplates a sequence-specific base editor (BE).sup.38. Although it is not necessary to understand the mechanism of an invention it is believed that the sequence-specific BE provides a direct reversion of common pathogenic mutations.
[0126] It has been reported that pathogenic mutations in the MECP2 gene account for about half of the disease alleles that are associated with this locus.sup.27. These lesions are most often reported to be C - T base transitions that produce either a missense or nonsense mutation. Table 1.
TABLE-US-00001 Representative MECP2 Mutations Mutation Result AA Change Mutation Freq (%) Bystander adenines c.473C>T Missense T158M 8.81 yes c.502C>T Nonsense R168X 7.63 no c.763C>T c.763C>T Nonsense R255X 6.68 no c.808C>T Nonsense R270X 5.80 no c.916C>T Missense R306C 5.17 no c.880C>T Nonsense R294X 5.00 yes c.397C>T Missense R133C 4.56 yes c.316C>T Missense R106W 2.77 no
Five of the eight most common Rett mutations would be suitable targets for adenine base editors in that they that do not have bystander adenines in danger of introducing new missense mutations at neighboring base pairs upon ABE treatment. Of these five suitable targets, the c.808C>T and c.316C>T mutations are targetable with standard SpyCas9 ABEs. The c.502C>T, c.763C>T and c.916C>T mutations are addressable with a Cas9-DBD ABEs.
[0127] In principle, an adenine base editor (ABE).sup.18 should be capable of reverting all eight of the common pathogenic MECP2 mutations, since it can drive T to C transitions in the context of a base pair. Implementation of the current generation of ABEs takes into account: 1) a complementary PAM at the correct position and on the desired DNA strand to allow base conversion, as ABEs have maximal activity on the ssDNA strand within a window roughly 13 to 16 nucleotides 5’ of the PAM element.sup.18, and 2) the absence of nearby adenines on the same strand (e.g., bystanders) that would also fall within the ABE active window, where their conversion to G would promote the generation of a missense mutation.
[0128] Four of the top five most frequent MECP2 mutations in Rett patients (R168X, R255X, R270X and R306C), which account for ~25% of all pathogenic mutations, do not have bystander concerns. Table 1. However, for SpyCas9 with its NGG PAM, only one out of these four mutant sequences (R270X) is targetable in the “sweet spot” of the ABE. FIG. 10. This fact highlights the importance of the density of available target sites for the Cas9 module within the ABE. The PAM recognition domain of SpyCas9 is a limitation that prevents maximal reversion efficiencies for many common MECP2 mutations. To address the issue of target density, an xCas9 base editor has been engineered to utilize an NGN PAM.sup.36, .sup.43, but independent studies using the xCas9 BE framework observed low base conversion at most NGN target sites..sup.43
[0129] Cas9-DNA-binding domain (Cas9-DBD) base editing platforms have been developed that have a much broader targeting range for PAM recognition than the standard SpyCas9 systems - effectively requiring only a single G within the PAM (NGN or NNG PAM) for function. FIG. 11. This more flexible BE platform is constructed based on an improved SpCas9 nuclease system with broader targeting range and specificity that employs a fusion to a programmable DNA-binding domain (either a Cys2-His2 zinc finger protein.sup.25 (ZFP) or an orthogonal nuclease-dead Cas9 (dCas9) to drive genome-locus-specific activity of the nuclease. FIG. 12.
IV. Development And Characterization Of Cas9-pDBD And Cas9-Cas9 Base Editors
[0130] In one embodiment, the present invention contemplates a fusion protein comprising an adenine or cytidine deaminase, a Cas9/sgRNA complex and a programmable DNA binding domain or a Cas9 base editor. In one embodiment, the pDBD base editor is an adenine base editor (ABE). In one embodiment, the pDBD base editor is a cytosine base editor (CBE).
A. Enhanced Adenine And Cytosine Base Editors (CBEs)
[0131] Conventionally, adenine and cytosine base editors are reported to comprise proteins such as, nickase SpCas9, nickase xCas9 or nSpCas9-NG. In one embodiment, the present invention contemplates fusion proteins comprising a Cas9/sgRNA complex and an enhanced adenine and cytosine base editors that include, but are not limited to, zinc finger proteins (ZFP), transcription activator-like effector (TALE) proteins, dead SaCas9 or dead Nm2Cas9. In one embodiment, the fusion protein is flanked by accessory proteins or domains including, but not limited to, adenine deaminase (hTadA-XTEN-hTadA*7.10, TadA8e) or cytidine deaminase (APOBEC1), nuclear localization signal (NLS) sequences (e.g., C-myc or SV40 NLS), intervening linkers (e.g., XTEN or other sequences) and/or uracil glycosylase inhibitor (UGI) proteins. See, FIGS. 17 and 33.
[0132] The improved activity of enhanced cytosine base editor embodiments were validated using a CopGFP reporter line. This reporter line shifts from a blue signal (BFP) to a green signal (GFP) subsequent to the modification of the trinucleotide target sequence from “cac” to “tat”. See, FIG. 18A. The data shows that the enhanced CBEs (blue/orange bars) contemplated herein have an approximate 2-fold increase in GFP fluorescence at target sites containing a PAM with a single G in comparison to previously reported CBEs (gray/turquoise/green bars). See, FIG. 18B. Similar data showing improved activity for enhanced CBEs as contemplated herein versus conventional CBE’s has been collected at KANK3 target sites having a variety of PAM sequences: i) TGT (FIG. 19); ii) ATG (FIG. 20); iii) TGA (FIG. 21); iv) GTG (FIG. 22); v) GGG (FIG. 23).
[0133] The improved activity of enhanced adenine base editor embodiments were validated using an mCherry reporter line. This reporter line shifts from no signal to a red signal subsequent to the modification of the codon target sequence from “tag” to “cag”. See, FIG. 24A. The data shows that the enhanced ABEs (blue/orange bars) contemplated herein have an approximate 2-fold increase in GFP fluorescence at target sites containing a PAM with a single G in comparison to previously reported ABEs (gray/white/green bars). See, FIG. 24B. Similar data showing improved activity for enhanced ABEs as contemplated herein versus conventional ABE’s has been collected at KANK3 target sites having a variety of PAM sequences: i) TGT (FIG. 25): ii) ATG (FIG. 26); iii) TGA (FIG. 27); iv) GTG (FIG. 28); v) GGG (FIG. 29); vi) NGG (FIG. 30); viii) NGH (FIG. 31); ix) NHG (FIG. 32).
V. Attenuated Cas9 Proteins
[0134] Although it is not necessary to understand the mechanism of an invention, it is believed that an attenuated nSpyCas9 system provides an avenue to dramatically reduce the off-target editing rates for any base editing system. In one embodiment, these base editing constructs target pathogenic mutations. In particular, the PAM recognition domain has a reduced affinity for the cognate PAM of a specific Cas9 protein. It is believed that this attenuation facilitates pDBD-mediated discrimination of binding between target and non-cognate target sites as described herein. Previous reporting has identified that, in the SpyCas9 an R1333S or R1335S substitution may result in attenuated Cas9 binding to the cognate PAM.
[0135] In one embodiment, an attenuated Cas9 protein comprises an amino acid substitution. In one embodiment, the amino acid substitution is in the PAM recognition domain. In one embodiment, the amino acid substitution comprises R1333S and K1118S. In one embodiment, the amino acid substitution comprises R1335K and E1219Q. In one embodiment, the amino acid substitution comprises R1333S, E1219Q and K1118S. In one embodiment, the attenuated Cas9 protein further comprises a pDBD protein. In one embodiment, the pDBD protein is a zinc finger protein. See, FIG. 33.
REFERENCES
[0136] 1. 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 (2016). [0137] 2. Nishida, K. et al. Targeted nucleotide editing using hybrid prokaryotic and vertebrate adaptive immune systems. Science (2016). doi:10.1126/science.aaf8729 [0138] 3. Gaudelli, N. M. et al. Programmable base editing of A.Math.T to G.Math.C in genomic DNA without DNA cleavage. Nature 551, 464-471 (2017). [0139] 4. Rees, H. A. & Liu, D. R. Base editing: precision chemistry on the genome and transcriptome of living cells. Nat Rev Genet 70, 3240 (2018). [0140] 5. Kuscu, C. et al. CRISPR-STOP: gene silencing through base-editing-induced nonsense mutations. Nature Methods (2017). doi:10.1038/nmeth.4327 [0141] 6. Gapinske, M. et al. CRISPR-SKIP: programmable gene splicing with single base editors. Genome Biol 19, 107 (2018). [0142] 7. Hu, J. H. et al. Evolved Cas9 variants with broad PAM compatibility and high DNA specificity. Nature (2018). doi:10.1038/nature26155 [0143] 8. Nishimasu, H. et al. Engineered CRISPR-Cas9 nuclease with expanded targeting space. Science (2018). doi:10.1126/science.aas9129 [0144] 9. Bolukbasi, M. F. et al. DNA-binding-domain fusions enhance the targeting range and precision of Cas9. Nature Methods 12, 1150-1156 (2015). [0145] 10. Persikov, A. V. et al. A systematic survey of the Cys2His2 zinc finger DNA-binding landscape. Nucleic Acids Research 43, 1965-1984 (2015). [0146] 11. Miller, J. C. et al. Improved specificity of TALE-based genome editing using an expanded RVD repertoire. Nature Methods 12, 465-471 (2015). [0147] 12. Bolukbasi, M. F. et al. Orthogonal Cas9-Cas9 chimeras provide a versatile platform for genome editing. Nature Communications 9, 4856 (2018). [0148] 13. Edraki, A. et al. A Compact, High-Accuracy Cas9 with a Dinucleotide PAM for In Vivo Genome Editing. Molecular Cell (2018). doi:10.1016/j.mo1ce1.2018.12.003 [0149] 14. Yeh, W.-H., Chiang, H., Rees, H. A., Edge, A. S. B. & Liu, D. R. In vivo base editing of postmitotic sensory cells. Nature Communications 9, 2184 (2018). [0150] 15. Liang, P. et al. Genome-wide profiling of adenine base editor specificity by EndoV-seq. Nature Communications 10, 67 (2019). [0151] 16. Kim, D. et al. Genome-wide target specificities of CRISPR RNA-guided programmable deaminases. Nature biotechnology (2017). doi:10.1038/nbt.3852 [0152] 17. Gehrke, J. M. et al. An APOBEC3A-Cas9 base editor with minimized bystander and off-target activities. Nature Biotechnology (2018). doi:10.1038/nbt.4199 [0153] 18. 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 (2017). [0154] 19. Pavletich, N. P. & Pabo, C. O. Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at 2.1 A. Science 252, 809-817 (1991). [0155] 20. Wolfe, S. A., Greisman, H. A., Ramm, E. I. & Pabo, C. O. Analysis of zinc fingers optimized via phage display: evaluating the utility of a recognition code. J Mol Biol 285, 1917-1934 (1999). [0156] 21. Cermak, T. et al. Efficient design and assembly of custom TALEN and other TAL effector-based constructs for DNA targeting. Nucleic Acids Research 39, e82-e82 (2011). [0157] 22. Ran, F. A. et al. In vivo genome editing using Staphylococcus aureus Cas9. Nature 520, 186-191 (2015).
