Epigenetic modulation system

Abstract

Described herein is a nucleic acid comprising a sequence encoding an epigenetic effector fusion comprising at least two epigenetic effectors. Described herein is an epigenetic editing system comprising any nucleic acid described herein and a nucleic acid encoding a gRNA comprising an RNA aptamer sequence that binds to the RNA binding protein. Described herein is a method of regulating gene expression, the method comprising introducing into a cell: any of the epigenetic modulation system described herein.

Description

Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025EPIGENETIC MODULATION SYSTEM

[0001] This International Patent Application claims the benefit of and priority to U.S. Provisional Application No. 63 / 696,705, filed September 19, 2024, entitled “EPIGENETIC MODULATION SYSTEM,” the content of which are hereby incorporated by reference in their entirety.

[0002] All patents, patent applications and publications cited herein are hereby incorporated by reference in their entirety. The disclosures of these publications in their entireties are hereby incorporated by reference into this application.

[0003] This patent disclosure contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure as it appears in the U.S. Patent and Trademark Office patent file or records, but otherwise reserves any and all copyright rights.BACKGROUND OF THE INVENTION

[0004] Epigenetic regulation in mammalian cells is a highly coordinated process, essential for organizing genetic information and orchestrating pivotal cellular functions like gene regulation, differentiation, and DNA replication or repair. Central to this regulation are large, multi-protein complexes that control epigenetic states through intricate combinations of post-translational modifications of DNA-scaffolding histone proteins and epigenetic modifications of the genome itself. These complexes often contain numerous subunits with distinct functionalities, such as the Polycomb repressive complexes, that coordinate a complex code of different histone modifications to enable context-dependent control of gene expression and chromatin state.

[0005] To date, the development of epigenetic editors has largely relied on rational design or low- throughput, guess-and-test approaches. Recent advances in pooled gene synthesis have enabled the functional screening of short peptide libraries of ~80 amino acids tiling larger effector protein candidates, further allowing the annotation of novel bivalent interactions that are synergistic or antagonistic. However, the synthesis length restriction of current-generation oligonucleotide pools may not accommodate the size of functionally active elements, and the use of engineered reporter loci may not extrapolate to endogenous genetic contexts. One recent study used barcoded multi ci stronic adaptors to clone and screen pairs of transcription factors up to 5.8 kilobases in length using a T cell knock-in system, butACTIVEUS 211418665 1Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 the generalizability and scale of this approach is limited by the requirement of a cell-type specific functionally monoallelic target locus and a sharp length-dependent integration bias.SEQUENCE LISTING

[0006] The instant application contains a Sequence Listing which has been submitted electronically in XML format and is hereby incorporated by reference in its entirety. Said XML copy, created on September 18 2025, is named 2220476_00130W01_SL.xml and is 324,713bytes in size.SUMMARY OF THE INVENTION

[0007] It is understood that any of the embodiments described below can be combined in any desired way, and that any embodiment or combination of embodiments can be applied to each of the aspects described below, unless the context indicates otherwise.

[0008] In certain aspects, described herein is a nucleic acid comprising a sequence encoding an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1 or comprising at least two epigenetic effectors of Table 2. In some embodiments, the first epigenetic effector comprises a Kruppel associated box (KRAB) domain. In some embodiments, the second epigenetic effector comprises L3MBTL3 sterile alpha motif (SAM) domain. In some embodiments, the first epigenetic effector comprises p65. In some embodiments, the second epigenetic effector comprises HSF1. In some embodiments, the first epigenetic effector comprises TET1. In some embodiments, the second epigenetic effector comprises the RING finger domain of RNF20, the SET domain of SMYD1, or the PR / SET domain of PRDM1. In some embodiments, the first and the second epigenetic effectors are connected by a linker. In some embodiments, the epigenetic effector fusion further comprises at least one RNA binding protein. In some embodiments, the RNA binding protein comprises MS2 bacteriophage coat protein (MCP) or PP7 bacteriophage coat protein (PCP). In some embodiments, the epigenetic effector fusion further comprises a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain. In some embodiments, the Cas protein comprises a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g that lacks nuclease and / or nickase activity. In some embodiments, the Cas protein comprises spdCas9. In some embodiments, the nucleic acid described herein further comprises an inducible promoter.ACTIVEUS 211418665 2Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0009] In certain aspects, described herein is a vector comprising any of the nucleic acids described herein.

[0010] In certain aspects, described herein is a host cell comprising any of the nucleic acid described herein or any of the vector described herein, wherein the host cell expresses the epigenetic effector fusion. In some embodiments, the transcription of a target DNA locus is reduced by about 5% to about 95% relative to the expression of the target DNA locus before expression of the nucleic acid. In some embodiments, the transcription of a target DNA locus is increased by about 5% to about 95% relative to the expression of the target DNA locus before expression of the nucleic acid. In some embodiments, the first epigenetic effector comprises TET1 and transcription of a target DNA locus is decreased for at least 5 days. In some embodiments, the first epigenetic effector comprises TET1 and transcription of a target DNA locus is decreased for at least 60 days. In some embodiments, the epigenetic effector fusion is expressed in the host cell transiently.

[0011] In certain aspects, described herein is an epigenetic editing system comprising any of the nucleic acid described herein.

[0012] In certain aspects, described herein is an epigenetic editing system comprising any nucleic acid described herein and a nucleic acid encoding a gRNA comprising an RNA aptamer sequence that binds to the RNA binding protein. In some embodiments, the RNA binding protein is MCP and the RNA aptamer sequence comprises MS2. In some embodiments, the RNA binding protein is PCP and the RNA aptamer sequence comprises PP7. In some embodiments, gRNA comprises a spacer sequence that targets the gRNA to a gene of interest. In some embodiments, the epigenetic editing system further comprises a nucleic acid encoding a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g lacks nuclease and / or nickase activity that binds to the gRNA. In some embodiments, the epigenetic editing system further comprises a nucleic acid encoding a spdCas9.

[0013] In certain aspects, described herein is a host cell comprising any of the epigenetic editing system described herein, wherein the host cell expresses the epigenetic editing system. In certain aspects, described herein is a host cell comprising any of the epigenetic editing system described herein, wherein the host cell expresses spdCas9.

[0014] In certain aspects, described herein is a method of regulating gene expression, the method comprising introducing into a cell: any of the epigenetic modulation system describedACTIVEUS 211418665 3Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 herein. In some embodiments, the epigenetic modulation system reduces or inhibits transcription of a target DNA locus.

[0015] In certain aspects, described herein is a nucleic acid comprising a sequence encoding a first epigenetic effector that reduces expression of a first target DNA locus and a first RNA binding protein and a second epigenetic effector that increases expression of a second target DNA locus and second RNA binding protein. In some embodiments, the first epigenetic effector comprises KRAB domain. In some embodiments, the second epigenetic effector comprises HSF1. In some embodiments, the first epigenetic effector comprises an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1. In some embodiments, the first epigenetic effector comprises L3MBTL3 sterile alpha motif (SAM) domain. In some embodiments, the second epigenetic effector comprises an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1. In some embodiments, the second epigenetic effector comprises p65. In some embodiments, the epigenetic effectors of the first epigenetic fusion and / or second epigenetic fusion are connected by a linker. In some embodiments, the first RNA binding protein comprises MS2 bacteriophage coat protein (MCP) and the second PP7 bacteriophage coat protein (PCP), or vice-versa. In some embodiments, the nucleic acid further comprises an inducible promoter.

[0016] In certain aspects, described herein is a nucleic acid comprising a sequence encoding a first epigenetic effector fusion that reduces expression of a first target DNA locus fused to a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain, and a second epigenetic effector fusion that increases expression of a second target DNA locus fused to a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain. In some embodiments, the first epigenetic effector fusion comprises at least two epigenetic effectors of Table 1 and the second epigenetic effector fusion comprises at least two epigenetic effectors of Table 1. In some embodiments, the first epigenetic effector fusion comprises KRAB domain. In some embodiments, the second epigenetic effector comprises HSF1. In some embodiments, the first epigenetic effector fusion comprises L3MBTL3 SAM domain. In some embodiments, the second epigenetic effector fusion comprises p65. In some embodiments, the epigenetic effectors of the first epigenetic fusion and / or second epigenetic fusion are connected by a linker. In some embodiments, the first and second epigenetic fusions are fused to a Cas protein. In some embodiments, the Cas protein comprises a Cas9, Cpfl, Cas 12b, Cas 12c, Cas 12d, Casl2e, Casl2f, Casl2h, Casl2i, or Cas 12g that lacks nuclease and / or nickase activity or a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f,ACTIVEUS 211418665 4Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Casl2h, Casl2i, or Casl2g with nuclease and / or nickase activity. In some embodiments, the Cas protein comprises spdCas9. In some embodiments, the first epigenetic effector and second epigenetic effector are on separate nucleic acid molecules. In some embodiments, the first epigenetic effector fusion and second epigenetic effector fusion are on separate nucleic acid molecules.

[0017] In certain aspects, described herein is a vector comprising any of the nucleic acids described herein.

[0018] In certain aspects, described herein is a host cell comprising any of the nucleic acid described herein or the vector described herein, wherein the host cell expresses the epigenetic effectors or epigenetic effector fusions. In some embodiments, expression of the first target DNA locus is reduced by about 5% to about 95% relative to the expression of the first target DNA locus before expression of the nucleic acid and wherein expression of the second target DNA locus is increased by about 5% to about 95% relative to the expression of the second target DNA locus before expression of the nucleic acid.

[0019] In certain aspects, described herein is an epigenetic editing system comprising any of the nucleic acids described herein. In some embodiments, the nucleic acid encoding a first gRNA comprising an RNA aptamer sequence that binds to the first RNA binding protein and a nucleic acid encoding a second gRNA comprising an RNA aptamer sequence that binds to the second RNA binding protein. In some embodiments, the first RNA binding protein is MCP and the RNA aptamer sequence comprises MS2 and wherein the second RNA binding protein is PCP and the RNA aptamer sequence comprises PP7. In some embodiments, the second RNA binding protein is MCP and the RNA aptamer sequence comprises MS2 and wherein the first RNA binding protein is PCP and the RNA aptamer sequence comprises PP7. In some embodiments, the first gRNA comprises a spacer sequence that targets the gRNA to a first gene of interest and the second gRNA comprises a spacer sequence that targets the gRNA to a second gene of interest. In some embodiments, the epigenetic editing system further comprises a nucleic acid encoding a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g lacks nuclease and / or nickase activity that binds to the gRNA. In some embodiments, the epigenetic editing system further comprises a nucleic acid encoding a spdCas9.

[0020] In certain aspects, described herein is a host cell comprising any of the epigenetic editing system described herein, wherein the host cell expresses the epigenetic editingACTIVEUS 211418665 5Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 system. In certain aspects, described herein is a host cell comprising any of the epigenetic editing system described herein, wherein the host cell expresses spdCas9.

[0021] In certain aspects, described herein is a method of regulating gene expression of at least two genes, the method comprising introducing into a cell: any of the epigenetic modulation system described herein. In some embodiments, the epigenetic modulation system reduces or inhibits expression of a first target DNA locus and increase expression of a second target DNA locus.BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The patent or application file contains at least one drawing executed in color. To conform to the requirements for PCT patent applications, many of the figures presented herein are black and white representations of images originally created in color.

[0023] FIGS. 1A-I show that COMBINE discovers epigenetic effector pairs that regulate endogenous human transcription. Figure 1A shows a conceptual framework for interrogating combinatorial biological processes through HTS of an N x N combinatorial domain library. Schematics of representative multi-protein complexes, including the Polycomb repressive complex 2 (PRC2) and the activating MLL3 / COMPASS complex, are shown. N x N combinatorial domains are recruited to the endogenous target gene loci by a dCas9 / MS2 system. Subpopulations of cells are enriched based on the target gene expression level in a high-throughput manner. Following the NGS readout, the combinatorial landscape of transcriptional perturbation is analyzed. Figure IB shows a schematic of Library 1 composed of epigenetic readers, recruiters, structural factors, and writers / erasers curated from diverse multiprotein complexes with an established ability to mediate transcriptional repression. Two examples of possible domain pairs from Library 1 are shown. A full list of members is available in Figure 37. Figure 1C shows a schematic of the composition of Library 2 including direct writers and erasers of diverse epigenetic modifications, curated agnostic of their known effect on transcription. Two examples of bivalent domain candidates from Library 2 are shown. A full list of members is available in Figure 38. Figure ID shows the experimental procedure for the combinatorial HTS of epigenetic effector pairs. Combinatorial domain candidates are recruited to the endogenous CD81 locus for 5-6 days with the addition of doxycycline. After staining with anti-CD81 antibody, cells are sorted into two bins (HIGH and LOW) based on the CD81 expression level. NGS is performed to measure the enrichment of bivalent domain candidates within sorted populations. Multiple timepoints are takenACTIVEUS 211418665 6Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 following the doxycycline removal - Day 0, Day 6, and Day 12 for Library 1 and Day 0 and Day 12 for Library 2. For more detailed descriptions, refer to methods and Figure 6A-D and Figure 7A-E. Figure IE shows the theoretical size distribution of bivalent domain candidates from Library 1 and 2 assuming uniform coverage. Figure IF shows scater plots of enrichment scores from Library 1 at Day 0. Effectors containing the repressive ZNF10 KRAB domain are highlighted in red. Figure 1G shows a scatter plot of enrichment scores from Library 1 at Day 12. Effectors containing the repressive ZNF10 KRAB domain are highlighted. Figure 1H shows a scatter plot of enrichment scores from Library 2 at Day 0. Effectors containing the repressive RYBP domain are highlighted in dark gray. Effectors containing the HSF1 activator are highlighted in light gray. Figure II shows a scatter plot of enrichment scores from Library 2 at Day 12. Effectors containing the repressive RYBP domain are highlighted in dark gray. Effectors containing the HSF1 activator are highlighted in light gray.

[0024] FIGS. 2A-D show classifications and size distributions of individual domains from Library 1 and Library 2. Figure 2A shows classification of Library 1 members. Figure 2B shows size distributions of specific classes from Library 1. Figure 2C shows classification of Library 2 members. DNA demethylation machinery (DNDM), DNA methyltransferase (DNMT), E2 and E3 ubiquitin ligases (UBL), histone acetyltransferase (HAT), histone arginine methyltransferase (HRMT), histone deacetylase (HDAC), histone demethylase (HDM), histone deubiquitinase (DUB), histone kinase (HK), histone lysine methyltransferase (HKMT), and histone phosphatase (HP). Figure 2D shows size distributions of specific classes from Library 2.

[0025] FIGS. 3A-C show Library 1 combinatorial cloning strategy. Figure 3A shows a schematic of combinatorial effector candidate cassette in lentiviral vector for HTS of Library DNA sequence encoding for Glutamic acid (E) - Serine (S) serves as a 5' overhang for the golden gate cloning. DNA sequence encoding for Serine (S) - Glycine (G) serves as a 3' overhang for the golden gate cloning. Cyan amino acid symbols indicate amino acids already encoded in the XTEN linker CDS. Orange amino acid symbols indicate amino acids inserted for the golden gate cloning overhang. Figure 3B shows a schematic of gene fragments ordered for Library 1 (Figure 37). Although PCR amplification handles are included in the design, it was not necessary to PCR amplify the gene fragments for cloning. Figure 3B discloses SEQ ID NOS 85-86, respectively, in order of appearance. Figure 3C shows library cloning procedure for the generation of Library 1 bivalent domain candidates.ACTIVEUS 211418665 7Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025The procedure consists of three sequential pooled Golden Gate assemblies. The first Golden Gate reaction involves the library cloning of pooled gene fragments into the N-term / KanRl receiver(pCMOOl). Each gene fragment was pooled with a molar ratio linear to its size including the PCR amplification handles. 1.5-fold of the 450 bp gene fragment was added compared to the 300 bp gene fragment. The second Golden Gate reaction involves the library cloning of pooled gene fragments into the C-term / KanR2 receiver (pCM002). The final Golden Gate reaction involves the library cloning of the N-terminus library and C-terminus library into the LV AmpR b ackbone (pCM003).

[0026] FIGS. 4A-K show Library 2 combinatorial cloning and barcode mapping strategy. Figure 4A shows schematic of clonal gene fragments ordered for Library 2 with dual golden gate overhangs. Figure 4A discloses SEQ ID NOS 87-89, respectively, in order of appearance. Figure 4B shows how clonal fragments of individual effectors were pooled manually to form the Effector 2 pool. Figure 4C shows that the amount of plasmid DNA added per effector increased 3.5-fold per kb. Figure 4D shows that the first golden gate assembly reaction for cloning effectors from Effector 2 pool into Effector 1 landing pad plasmid. Figure 4E shows a schematic of cloned Effector 1 plasmid with golden gate overhangs. Figure 4E discloses SEQ ID NOS 90-91, respectively, in order of appearance. Figure 4F shows that the XTEN16 linker was ordered as ssDNA and annealed with proper golden gate overhangs. Figure 4F discloses SEQ ID NOS 92-93 and 66, respectively, in order of appearance. Figure 4G shows that barcodes were ordered as a ssDNA fragment containing a 20N sequence. Polymerase extension was used to fill in barcodes, forming a dsDNA fragment, before the fragment was predigested with Esp3I to form proper golden gate overhangs. Figure 4G discloses SEQ ID NOS 94-95, respectively, in order of appearance. Figure 4H shows the final 5-piece golden gate reaction for generating Library 2 bivalent effectors for HTS. Figure 41 shows a schematic of Library 2 bivalent effectors cloned into a lentiviral vector for HTS. Golden gate assembly overhang positions are indicated as A-E. Barcode mapping Cas9 guides are indicated in red. Figure 41 discloses SEQ ID NOS 96-98, respectively, in order of appearance. Figure 4 J shows a heatmap of barcodes mapped per bivalent effector, with effectors organized by increasing size. Figure 4K shows a scatterplot of bivalent effector length versus number of mapped barcodes.

[0027] FIGS. 5A-H show doxycycline-inducible dCas9 / MS2 system enables temporal control of transcriptional perturbation. Figure 5A shows an illustration of dCas9 / MS2 system for Library 1 HTS. Tet-On dCas9 K562 (without CD81 guide array) cell line was used for theACTIVEUS 211418665 8Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 experiment. Combinatorial effectors are fused to stdMCP and delivered by lentiviral infection. Both dCas9 and combinatorial effectors are inducible by doxycycline. 3X CD81 guide array is introduced by plasmid nucleofection. Figure 5B shows the timeline for testing dCas9 / MS2 system used for Library 1 with established epigenetic repressor KRAB from ZNF10 protein, activator VP64, and no effector control. Figure 5C shows representative histograms of CD81 expression at 5 days post-nucleofection of CD81 targeting 3X guide array plasmids. Figure 5D shows the time course of CD81 repressed cells after nucleofection of CD81 targeting 3X guide array plasmids. Error bars indicate the standard deviation of 3 biological replicates. Figure 5E shows an illustration of dCas9 / MS2 system for Library 2 HTS. Tet-On dCas9 K562 (with CD81 guide array) cell line was used for the experiment. Combinatorial effectors were fused to MCP and introduced by lentivirus. CD81 targeting sgRNAs were constitutively expressed. Both dCas9 and combinatorial effectors are inducible by doxycycline. Figure 5F shows timeline for testing dCas9 / MS2 system used for Library 2 with established epigenetic repressor RYBP, activator HSF1, and neutral control 400 aa fragment of DMD. Figure 5G shows violin plots of CD81 expression at experimental Day 0 with or without the addition of doxycycline. Figure 5H shows violin plots of CD81 expression 5 days after washout of doxycycline after initial 5 days of doxycycline treatment.

[0028] FIGS. 6A-D show detailed procedures for Library 1 HTS. Figure 6A shows the timeline and NGS strategy for Library 1 HTS. CD81 targeting guide array plasmid was nucleofected, and nucleofected cells were sorted based on the mCherry marker present in the plasmid (pMH222). Figure 6B shows representative FACS plots of mCherry level at 3 days post-nucleofection of 3X guide array plasmids which express mCherry as a marker. Cells that were successfully nucleofected with 3X guide array plasmids were sorted based on mCherry level. To reduce heterogeneity in the sgRNA level, moderate mCherry-expressing cells were sorted. Figure 6C shows FACS plots of multiple conditions from Library 1 HTS. Approximate sorting gates are drawn. The actual gating was continuously amended during sorting to match the aimed HIGH / LOW percentage of the sorted cells. This procedure was necessary as the staining (CD81, APC) histogram shifted as the sorting procedure prolonged. Figure 6D shows number of cells sorted per screening condition.

[0029] FIGS. 7A-E show Detailed procedures for Library 2 HTS and barcode mapping. Figure 7A shows timeline and NGS strategy for Library 2 HTS. CD81 targeting sgRNAs were constitutively expressed from Tet-On dCas9 K562 cell lines (with CD81 guide array). Figure 7B shows FACS plots of Library 2 HTS on Day 0 illustrating 25th and 75th percentileACTIVEUS 211418665 9Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 gates, which were continuously adjusted throughout the sort. Figure 7C shows FACS plots of Library 2 HTS on Day 12 illustrating 25th and 75th percentile gates, which were continuously adjusted throughout the sort. Figure 7D shows number of cells sorted per screening condition. Figure 7E shows number of NGS reads per screening condition with the fraction of reads mapped to bivalent effectors indicated.

[0030] FIGS. 8A-8B show the investigation of lentiviral barcode swapping by targeted nanopore sequencing. Figure 8A shows schematic of genome integrated bivalent effectors from Library 2. Cas9 guide positions for targeted nanopore sequencing are indicated. Figure 8A discloses SEQ ID NOS 96-98, respectively, in order of appearance. Figure 8B shows number and percentage of genome-integrated nanopore reads with barcodes matching or mismatching effector pairs mapped by nanopore plasmid sequencing.

[0031] FIGS. 9A-E show number of acquired NGS reads versus bivalent effector length. Figure 9A shows number of NGS reads versus length of bivalent effector from Library 1 HTS Day 0 data. Figure 9B shows number of NGS reads versus length of bivalent effector from Library 1 HTS Day 6 data. Figure 9C shows number of NGS reads versus length of bivalent effector from Library 1 HTS Day 12 data. Figure 9D shows number of NGS barcode reads versus length of the corresponding bivalent effector from Library 2 HTS Day 0 data. Figure 9E shows number of NGS barcode reads versus length of the corresponding bivalent effector from Library 2 HTS Day 12 data.

[0032] FIGS. 10A-F show epigenetic pairs generate a rich combinatorial landscape of transient gene expression outcomes. Figure 10A shows a heatmap of enrichment scores generated from Library 1 at Day 0. Log2 transformed geometric means of HIGH / LOW enrichment scores from both replicates are plotted. Color scale centered at distribution mode. For this repressor-focused library, a lower log2 score indicates stronger repression. Dropouts are indicated in gray. Figure 10B shows Heatmap of enrichment scores generated from Library 2 at Day 0. Log2 transformed geometric means of HIGH / LOW enrichment scores from both replicates are plotted. Color scale centered at distribution mode. Dropouts are indicated in gray. For this library composed of putative repressors and activators, a negative score indicates repression and a positive score indicates activation. Histone deubiquitinase (DUB), histone acetyltransferase (HAT), histone arginine methyltransferase (HRMT), histone demethylase (HDM), histone lysine methyltransferase (HKMT), histone phosphatase (HP), E2 ubiquitin ligases (E2 UBL), E3ubiquitin ligases (E3 UBL), histone kinase (HK), DNA demethylation machinery (DNDM), histone deacetylase (HD AC), and DNAACTIVEUS 211418665 10Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 methyltransferase (DNMT). Figure IOC shows violin plots of all possible bivalent combinations within the same classes from Library 1. A breakdown of the recruiter class into KRAB and non-KRAB pairs is also shown (not shaded). The orange bar represents the median. Box spans quartile 1 to quartile 3. Whiskers extend to minimum and maximum. Number of combinations per class is indicated. The dashed line indicates the distribution mode of all pairs. Figure 10D shows violin plots of all possible bivalent combinations within the same classes from Library 2. The orange bar represents the median. Box spans quartile 1 to quartile 3. Whiskers extend to minimum and maximum. Number of combinations per class is indicated. The dashed line indicates the distribution mode of all pairs. Figure 10E shows violin plots of all possible bivalent combinations with the specified HAT effectors on the C terminus. The orange bar represents the median. Box spans quartile 1 to quartile 3. Whiskers extend to minimum and maximum. Number of combinations per effector is indicated. The dashed line indicates the distribution mode of all pairs. Figure 10F shows violin plots of all possible bivalent combinations with the specified HD AC effectors on the C terminus. The orange bar represents the median. Box spans quartile 1 to quartile 3. Whiskers extend to minimum and maximum. Number of combinations per effector is indicated. The dashed line indicates the distribution mode of all pairs.

[0033] FIGS. 11A-C show heatmaps of enrichment scores from later HTS timepoints. Figure HA shows Enrichment score heatmap from Library 1 HTS on Day 6. Figure 11B shows enrichment score heatmap from Library 1 HTS on Day 12. KRAB-DNMT3A and DNMT3A-KRAB combinations are marked with green boxes. Figure 11C shows enrichment score heatmap from Library 2 HTS on Day 12.

[0034] FIGS. 12A-D show N- versus C-terminal position effects. Figure 12A shows scatter plot of enrichment scores from X-Y versus Y-X bivalent effector pairs from Library 1 HTS on Day 0. Effectors with strong positional preference for the N-terminus are indicated in the X position and the Y position. Figure 12B shows scatter plot of enrichment scores from X-Y versus Y-X bivalent effector pairs from Library 2HTS on Day 0. Figure 12C shows a volcano plot illustrating positional preference of individual effectors from Library 1. X-axis score calculated as the difference in mean log2(HIGH / LOW) between matched effector pairs with given effector at N- versus C-terminus (C -terminus minus N-terminus), multiplied by -1 for effectors with repressive marginal scores and +1 for effectors with activating marginal scores. Q-values were calculated as FDR-corrected p-values from a two-sample t-test. OnlyACTIVEUS 211418665 11Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 effectors with >30 matched pairs are shown. Figure 12D shows a volcano plot illustrating positional preference of individual effectors from Library 2. Calculated as Figure 12C.

[0035] FIGS. 13A-G show marginal effector analysis reveals how individual domains perturb transcription across partners. Figure 13A shows schematic of the dCas9 / MS2 epigenetic editing complex bound at CD81 locus. Complex expression is induced by the addition of doxycycline and is recruited to the CD81 loci for 5-6days prior to screen readout. Figure 13B shows the concept of marginal effector analysis. For a given effector, the marginal score is calculated as the difference in mean log2(HIGH / LOW) between all pairs containing the effector versus not containing the effector. Q-values were calculated as FDR- corrected p-values from Welsh’s t-test. Figure 13C shows a volcano plot of marginal scores from Library 1 at Day 0. KRAB family members are highlighted. Vertical dashed line indicates the average marginal score across the KRAB family. Horizontal dashed line indicates q-value of 0.05. Figure 13D shows violin plots of CD81 expression after expression of individual effectors from Library 1 in individual validation (3 days post nucleofection of effector encoding plasmids). Mock transfection of pUC19 colored in gray. 2 independent replicates illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed line indicates repression gate at the 1st percentile of the pUC19 condition. The average percentages of repressed cells in each condition are indicated. The average percent changes in CD81 mean fluorescence intensity (MFI) versus pUC19 for each condition are indicated. Figure 13E Volcano plot of marginal scores from Library 2 at Day 0. Effectors from HD AC, DNMT, HAT, and DUB classes are highlighted. Vertical dashed lines indicate average marginal score for each respective colored class. Horizontal dashed line indicates q-value of 0.05. Figure 13F shows violin plots of CD81 expression after expression of the HDAC7 deacetylase domain and its mutants in individual validation (5 days post nucleofection and doxycycline induction). WTHDAC7 is the middle column. Mutants are the left and right columns. MCP-DMD neutral condition is colored gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed line indicates repression gate at the 1stpercentile of the neutral condition. The average percentages of repressed cells in each condition are indicated. The average percent changes in CD81 MFI versus DMD for each condition are indicated. *proposed mutant inferred from HDAC4. Figure 13G shows violin plots of CD81 expression after expression of individual effectors from Library 2 in individual validation (5 days post nucleofection and doxycycline induction). WT effectors are coloredACTIVEUS 211418665 12Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 green. Catalytic mutants are colored orange. Repressing and activating controls are colored respectively. MCP-DMD neutral condition is colored gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed lines indicate repression and activation gate at the 1stpercentile of neutral condition. The average percentages of repressed and activated cells in each condition are indicated.

[0036] FIGS. 14A-H show that COMBINE uncovers diverse modes of epigenetic interactions. Figure 14A shows the concept of synergy score calculation to quantify synergy and antagonism. Marginal scores of each individual effector are used to calculate an additive range for each bivalent combination. The lower bound of the additive range is defined as the minimum of the N-terminal effector’s marginal score, the C-terminal effector’s marginal score, and the sum of the marginal scores. The upper bound is defined as either the sum of the marginal scores or the minimum of the N-terminal and C-terminal marginal scores (whichever is higher). The magnitude of the synergy score represents how far outside the additive range the mean-centered observed bivalent enrichment score falls. Synergy is indicated by positive synergy scores, while antagonism is indicated by negative synergy scores. Two example cases illustrate the calculation of synergy scores. Library 1 synergy scores are all calculated in the repression direction. Library 2 synergy scores are calculated both in the activation and repression direction depending on whether the sum of marginal scores is positive or negative, respectively. Figure 14B shows a volcano plot of average synergy scores with HD AC family from Library 2 on Day 0. Effectors are colored based on the average log2(HIGH / LOW) enrichment scores across all combinations of the given effector with significant HD AC repressors. Q-values were calculated as FDR corrected p- values from a one-sample Wilcoxon test. Horizontal dashed line indicates q-value of 0.05. Figure 14C shows violin plots of CD81 expression at 5 days after nucleofection and induction of the HDAC7 +UBE2E1 combination and its mutants. WT domains are colored green. Mutant combinations are colored orange. MCP-DMD neutral condition is colored in gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed line indicates repression gate at the 1st percentile of neutral condition. The average percentages of repressed cells in each condition are indicated. The average percent changes in CD81 MFI versus DMD for each condition are indicated. *proposed mutant inferred from homology to HDAC4. Figure 14D shows violin plots of CD81 expression at 5 days after nucleofection and induction of HDAC4 +VRK1 andACTIVEUS 211418665 13Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 enzymatic mutants. WT domains are located in the second column. Mutant combinations are in all other columns. MCP-DMD neutral condition colored gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed line indicates repression gate at the 1st percentile of neutral condition. The average percentages of repressed cells in each condition are indicated. The average percent changes in CD81 MFI versus DMD for each condition are indicated. *proposed mutant inferred from HDAC7. Figure 14E shows the percent change in MFI relative to the neutral DMD control following dual nucleofection of plasmids encoding key transient effectors from Library 2 and 3X guide arrays targeting CD55, CD58, CD151, and CD155. CD81 percent MFIs are from single plasmid nucleofection into Tet-On dCas9 cell line with preinstalled 3X CD81 guide array. The dCas9 / MS2 complex was induced by doxycycline for 5 days. Error bars represent the standard deviation between 3 independent replicates.*proposed mutant inferred from homology. Lack of C-terminal effector indicates that the effector was tested monovalently. Arrows point from WT effector to mutant form of the same effector. Gray boxes indicate the same effector. Additional detailed data are shown in Figures 20A-D. Figure 14F shows heatmap of Pearson correlations of percent change in MFI versus the neutral DMD control between each gene tested. Datasets consist of the 13 effectors shown in Figure 14E. Additional detailed data are shown in Figure 21. Figure 14G shows a volcano plot of average synergy scores with KRAB family from Library 1 on Day 6. Effectors are colored based on the average log2(HIGH / LOW) enrichment scores across all combinations of the given effector with significant KRAB repressors. Q-values were calculated as FDR corrected p-values from a one-sample Wilcoxon test. Horizontal dashed line indicates q-value of 0.05.

[0037] Figure 14H shows a time course of CD81 repression following nucleofection of plasmids encoding KRAB combinations (n=2 independent replicates).

[0038] FIGS. 15A-C show heatmaps illustrating synergy scores for bivalent combinations in Library 1 on Day 0 (A), Day 6(B), and Day 12 (C). Red indicates antagonism where measured scores were less repressive than expected, and blue indicates synergy where scores were more repressive than expected. The magnitude of the score represents how far measured scores were outside of the expected (additive) range.

[0039] FIGS. 16A-B shows heatmaps illustrating synergy scores for bivalent combinations in Library 2 on Day 0 (A) and Day 12 (B). Lower values indicates antagonism where measured scores were less repressive or activating than expected, and higher valuesACTIVEUS 211418665 14Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 indicates synergy where scores were more repressive or activating than expected. Scores were computed in the direction of the stronger effector. The magnitude of the score represents how far measured scores were outside of the expected (additive) range.

[0040] FIGS. 17A-E show synergistic interaction between the H2AK119ubl reader and a subset of PRC 1 recruiters. Figure 17A shows sub-heatmap of synergy scores from Library1 HTS Day 0 data with recruiters of PRCl&2on the N-terminus and readers of epigenetic modifications on the C-terminus. The box indicates combinations between PRC1 recruiters and the H2AK119ubl reader domain from the RYBP protein. The PRC1 recruiting domain from the RYBP protein is indicated by a horizontal arrow. Figure 17B shows a volcano plot of average synergy scores with PRC1 recruiters on the N-terminus from Library HTS Day 0 data. Combinations with the KRAB family on the C-terminus are colored. Combination with the H2AK119ubl mark reader, the zf-RanBP domain from RYBP protein, on the C-terminus is colored blue. Q-values were calculated as FDR-corrected p-values from a one-sample Wilcoxon test. Horizontal dashed line indicates q-value of 0.05. Figure 17C shows an illustration of PRC 1 recruiter + H2AK119ubl reader fusion proteins tested for validation. Figure 17D shows percentage of CD81 repressed cells at 4 days post-nucleofection of fusion protein plasmids. CBX2 and CBX7 are components of canonical PRC1 (cPRCl) whereas RYBP is a component of variant PRC1 (vPRCl). Error bars indicate the standard deviation of2 biological replicates. Figure 17E shows illustration of full-length RYBP and potent fusion protein identified from Library 1 HTS on Day 0.

[0041] FIGS. 18A-B show HD AC and KRAB members included in the average synergy score calculation. Figure 18A shows marginal scores of HD AC members from Library 2 stratified by HD AC subclass. The dashed line indicates the threshold for inclusion in the average synergy score calculation. Only members with marginal scores less than -0.5 were included. Figure 18B shows marginal scores of KRAB members from Library 1. The dashed line indicates the threshold for inclusion in the average synergy score calculation. Only members with marginal scores less than -0.5 were included.

[0042] FIG. 19 shows visualization of endogenous CD81 locus with mapped epigenetic modifications. Vertical sky blue bars indicate positions of sgRNA binding. Plot generated with UCSC Genome Browser athttp: / / genome.ucsc.edu.

[0043] FIGS. 20A-D show key effectors from Library 2 transiently perturb CD55, CD58, CD151, and CD155. Figure 20A shows violin plots of CD55 expression at 5 days after nucleofection and induction of key effectors identified from Library 2. WT domains andACTIVEUS 211418665 15Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 mutants or combinations containing mutants are denoted with different colors. Repressing and activating controls are colored red and blue respectively. MCP-DMD neutral condition is colored in gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed lines indicate repression and activation gate at the 1st and 99th percentile of neutral condition. The average percentages of repressed and activated cells in each condition are indicated. The average percent changes in MFI versus the neutral DMD domain for each condition are indicated. *proposed mutant inferred from homology. Lack of C-terminal effector indicates that the effector was tested monovalently. Arrows point from WT effector to mutant form of the same effector. Gray boxes indicate the same effector. Figure 20B shows violin plots of CD58 expression at 5 days after nucleofection and induction of key effectors identified from Library 2. Described as Figure 20A. Figure 20C shows violin plots of CD151 expression at 5 days after nucleofection and induction of key effectors identified from Library 2. Described as Figure 20A. Figure 20D shows violin plots of CD155 expression at 5 days after nucleofection and induction of key effectors identified from Library 2. Described as Figure 20A.

[0044] FIG. 21 shows scatter plots illustrating gene-by-gene comparisons of percent change in MFI versus the DMD neutral control for each of the 13 effectors tested in Figures 20A-D. Pearson correlation for each comparison is indicated. KRAB and p65-HSFl controls are not included in correlation calculation.

[0045] FIGS. 22A-D show synergistic partners of KRAB and antagonistic partners of KRAB and SID4X. Figure 22A shows probability histograms of average HTS (LOW / HIGH) values of the selected KRAB combinations from Library 1 HTS on Day 6. Figure 22B shows arrayed validation of synergistic KRAB combinations including L3MBTL3 and EPOP. KRAB domain is from the ZNF10 protein. DMD is an 80 aa fragment stuffer that has been validated to be well expressed and does not have transcriptional perturbation activity from a previous study. Figure 22C shows violin plots of CD81 expression 3 days after nucleofection of individual effectors or bivalent effectors from Library 1. Mock transfection of pUC19 colored in gray. 2 independent replicates illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed line indicates repression gate at the 1st percentile of the pUC19 condition. The average percentages of repressed cells in each condition are indicated. The average percent changes in CD81 mean fluorescence intensity (MFI) versus pUC19 for each condition are indicated. Figure 22D shows theACTIVEUS 211418665 16Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 distribution of Library 1 HTS Day 0 synergy scores from the bivalent combinations that include SID4X.

[0046] FIGS. 23A-E show a combination of KRAB and L3MBTL3 enables robust and bidirectional transcriptional perturbation. Figure 23A shows an illustration of the exemplary clustering of the dCas9 / MS2 complex due to the multimerization property of the SAM domain of L3MBTL3. Figure 23B shows the doxycycline response curve of CD81 repression by stdMCP-KRAB and stdMCP-KRAB-L3MBTL3 4 days after doxycycline induction (n = 4 biological replicates). KRAB domain is from ZNF10. Figure 23C shows target gene repression 5 days following lentiviral transduction with sgRNA and doxycycline induction of the effector components (n > 2 biological replicates). Distance between the sgRNA target site and the TSS is indicated in parentheses. Error bars indicate the standard deviation of biological replicates. * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001; two-tailed t-test. Figure 23D shows a 2D-contour plot and kernel density estimate (KDE) plots of populations transduced with dual MS2 / PP7 sgRNAs. stdMCP-KRAB was recruited for CD81 gene repression and stdPCP-p65-HSFl was recruited for the activation of CD274 gene expression. CD81 / CD274 expression profiles were measured 5 days following lentiviral transduction with sgRNAs and doxycycline induction of the effector components. Genes within parentheses indicate the target genes for MS2 and PP7 sgRNAs, respectively. For example, (CD81, NT1) denotes that the CD81 gene is targeted by the MS2 sgRNA, while NT1 (non-target 1) is a control sgRNA used with PP7. Figure 23E shows 2D-contour plot and kernel density estimate (KDE) plots of populations infected with dual MS2 / PP7 sgRNAs. stdMCP-KRAB-L3MBTL3 was recruited for CD81 gene repression and stdPCP-p65-HSFl was recruited for the activation of CD274 gene expression. CD81 / CD274 expression profiles were measured 5 days following lentiviral transduction with sgRNAs and doxycycline induction of the effector components. Genes within parentheses indicate the target genes for MS2 and PP7 sgRNAs, respectively. For example, (CD81, NT1) denotes that the CD81 gene is targeted by the MS2 sgRNA, while NT1 (non-target 1) is a control sgRNA used with PP7.

[0047] FIGS. 24A-D show quantification of bidirectional perturbations by MS2 / PP7 system. Figure 24A shows a percentage of CD81 repressed cells 5 days post-infection of dual MS2 / PP7 sgRNAs to the respective cell lines. Error bars indicate the standard deviation between 3 independent replicates. Figure 24B shows the percentage of CD274 activated cells 5 days post-infection of dual MS2 / PP7 sgRNAs to the respective cell lines. Error bars indicate the standard deviation between 3 independent replicates. Figure 24C shows meanACTIVEUS 211418665 17Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 fluorescence intensity (MFI) of CD81 repressed cells 5 days post-infection of dual MS2 / PP7 sgRNAs to the respective cell lines. Expression level was normalized to control cells with no effectors. Error bars indicate the standard deviation between 3 independent replicates.Figure 24D shows the mean fluorescence intensity (MFI) of CD274 activated cells 5 days post-infection of dual MS2 / PP7 sgRNAs to the respective cell lines. Expression level was normalized to control cells with no effectors. Error bars indicate the standard deviation between 3 independent replicates.

[0048] FIGS. 25A-F show DNA methylation writers and erasers are essential for the memory of silencing, repression, and activation of gene expression. Figure 25A depicts a schematic showing loss of the dCas9 / MS2 epigenetic editing complex 12 days post withdrawal of doxycycline. CD81 expression perturbations rely on sustained epigenetic modifications. Figure 25B shows a volcano plot of marginal scores from Library 2 on Day 12. Horizontal dashed line indicates q-value of 0.05. Figure 25C shows a time course of CD81 repression following nucleofection of plasmids encoding DNMT3A-3L combinations and induction of the dCas9 / MS2 complex by doxycycline treatment for 5 days. Day 0 indicates the day of doxycycline washout. Each bar represents 3 timepoints collected within the specified days. The bar heights represent the average perturbation across the timepoints, and the error bars represent the standard deviation between 3 independent replicates. *proposed mutant inferred from PRDM7. Lack of C-terminal effector indicates that the effector was tested monovalently. Arrows point from WT effector to mutant form of the same effector. Additional detailed data are shown in Figures 27A-E. Figure 25D shows a time course of CD81 MFI relative to the neutral DMD control following nucleofection of plasmids encoding UBE2E1 + DNMT3 A combinations and induction of the dCas9 / MS2 complex by doxycycline for 5 days. Day 0 indicates the day of doxycycline washout. Each bar represents timepoints collected within the specified days. The bar heights represent the average perturbation across the timepoints, and the error bars represent the standard deviation between independent replicates. Arrows point from WT effector to mutant form of the same effector. Additional detailed data are shown in Figures 28A-D. Figure 25E shows a time course of CD81 MFI relative to the neutral DMD control following nucleofection of plasmids encoding TET1 combinations and induction of the dCas9 / MS2 complex by doxycycline for 5 days. Day 0 indicates the day of doxycycline washout. Each bar represents 3 timepoints collected within the specified days. The bar heights represent the average perturbation across the timepoints, and the error bars represent the standard deviation between 3 independentACTIVEUS 211418665 18Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 replicates. Lack of C-terminal effector indicates that the effector was tested monovalently. Arrows point from WT effector to mutant form of the same effector. Additional detailed data are shown in Figures 29A-D and Figures 30A-C. Figure 25F shows representative CD81 expression profiles illustrating three types of gene expression memory. CD81 expression levels were measured 12 days after doxycycline washout following 5 days of doxycycline induction.

[0049] FIGS. 26A-C show marginal effector and synergy score analysis on Day 12. Figure 26A shows a volcano plot of marginal scores from Library 1 HTS on Day 12. DNMT3A with a marginal score of -0.53 is highlighted in orange. Horizontal dashed line indicates q-value of 0.05. Figure 26B shows a volcano plot of average synergy scores with KRAB family from Library 1 HTS on Day 12. Effectors are colored based on the average log2(HIGH / LOW) enrichment scores across all combinations of the given effector with significant KRAB repressors. Q-values were calculated as FDR-corrected p-values from a one-sample Wilcoxon test. Horizontal dashed line indicates q-value of 0.05. Figure 26C shows a volcano plot of average synergy scores with HD AC family from Library 2 HTS on Day 12. Effectors are colored based on the average log2(HIGH / LOW) enrichment scores across all combinations of the given effector with significant HD AC repressors. Q-values were calculated as FDR-corrected p-values from a one-sample Wilcoxon test. Horizontal dashed line indicates q-value of 0.05.

[0050] FIGS. 27A-E show DNMT3 A-3L combinations induce long-term partial silencing of CD81. Figure 27A shows a violin plot time course of CD81 expression following 5 days transient recruitment of DNMT3A-3L and catalytic mutant. MCP-DMD neutral condition colored in gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed lines indicate repression, activation, and off gate. Repression and activation gates defined as 1st and 99th percentile of neutral condition. Off gate defined as 99th percentile of unstained WT K562s. The average percentage of each population in each condition is indicated. Figure 27B shows a violin plot time course of CD81 expression following 5 days transient recruitment of UHRF1 + DNMT3A-3L and catalytic mutant of UHRFl. Described as Figure 27A. Figure 27C shows a violin plot time course of CD81 expression following 5 days transient recruitment of PRDM4 + DNMT3A-3L and catalytic mutant of PRDM4. Described as Figure 27A. Figure 27D shows the time course of CD81 repressed cells following 5 days recruitment of WT combinations and mutant combinations. Error bars denote standardACTIVEUS 211418665 19Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 deviation between 3 independent replicates. Figure 27E shows the average fold change in CD81 repressed cells for WT effectors versus mutant effectors in DNMT3A-3L combinations. Error bars denote standard deviation between ratios calculated from average percentages across the 50 day time course.

[0051] FIGS. 28A-D show UBE2E1 + DNMT3A induces long-term repression of CD81. Figure 28A shows a violin plot time course of CD81 expression following 5 days transient recruitment of UBE2E1 + DNMT3A and UBE2E1(C131A) + DNMT3A. MCP-DMD neutral condition colored in gray. 3independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed lines indicate repression, activation, and off gate. Repression and activation gates defined as 1st and 99th percentile of neutral condition. Off gate defined as 99th percentile of unstained WT K562s. The average percentage of each population in each condition is indicated. Figure 28B shows a violin plot time course of CD81 expression following 5 days transient recruitment of UBE2E1 + DNMT3A(C710A) and UBE2E1(C131A) + DNMT3A(C710A). Described as Figure 28 A. Figure 28C shows a time course of % change in CD81 MFI versus neutral DMD 1120 control following 5 days recruitment of UBE2E1 + DNMT3A and mutant combinations. Error bars denote standard deviation between 3 independent replicates. Figure 28D show the average fold change in % change in CD81 MFI for UBE2E1 + DNMT3A versus UBE2E1(C131 A) + DNMT3 A. Error bars denote standard deviation between ratios calculated from average percentages across the 50 day time course, excluding outlier at day 44 timepoint.

[0052] FIGS. 29A-D show violin plot time courses of TET1 combinations. Figure 29A shows a violin plot time course of CD81 expression following 5 days transient recruitment of TET1 and TET1(H1672Y+D1674A). MCP-DMD neutral condition colored in gray. 3 independent replicates are illustrated as translucent overlays. The geometric means of each replicate are shown as solid black lines. Dashed lines indicate repression, activation, and off gate. Repression and activation gates defined as 1st and 99th percentile of neutral condition. Off gate defined as 99th percentile of unstained WT K562s. The average percentage of each population in each condition is indicated. Figure 29B shows a violin plot time course of CD81 expression following 5 days transient recruitment of RNF20 +TET1 and RNF20(K959E) + TET1. Described as Figure 29A. Figure 29C shows violin plot time course of CD81 expression following 5 days transient recruitment of SMYD1+ TET1 and SMYD1(R19A) + TET1. Described as Figure 29A. Figure 29D shows a violin plot timeACTIVEUS 211418665 20Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 course of CD81 expression following 5 days transient recruitment of PRDM1+ TET1 and PRDMl(Y200A) + TET1. Described as Figure 29A.

[0053] FIGS. 30A-C show TET1 combinations induce long-term activation of CD81. Figure 30A shows a time course of percent change in CD81 MFI versus neutral DMD control following nucleofection of MCP-TET1 combination encoding plasmids and 5 days recruitment of dCas9 / MS2 complex. Day 0 indicates the day of doxycycline washout. Each bar represents 3 timepoints collected within the specified days. The bar heights represent the average perturbation across the timepoints, and the error bars represent the standard deviation between independent replicates. Figure 30B shows a time course of % change in CD81 MFI versus neutral DMD control following 5 days recruitment of TET1 combinations and mutants. Error bars denote standard deviation between 3 independent replicates. Figure 30C shows the average fold change in % change in CD81 MFI for WT effectors versus mutant effectors in TET1 combinations. Error bars denote standard deviation between ratios calculated from average percentages across the 50 day time course.

[0054] FIG. 31 shows the two distinct epigenetic libraries used in the screen. Library 1 consists of 155 members of known readers, recruiters, writers, and structural factors. Library 2 consists of 198 catalytic epigenetic editors.

[0055] FIG. 32 shows the total number of unique pairwise combinations in each library.

[0056] FIG. 33 shows a volcano plot of marginal scores computed on only the subset of combinatorial effectors that contain KRAB family members from Library 1 on Day 6. Q- values were calculated as FDR742 corrected p-values from a two-sample t-test. Horizontal dashed line indicates q-value of 0.05.

[0057] FIG. 34 shows an embodiment of the enhancement of KRAB-based transcriptional repressors using L3MBTL3 SAM domain. See Example 10.

[0058] FIG. 35 shows an embodiment of the bidirectional transcriptional perturbation system for simultaneous gene up- and down-regulation. See Example 11.

[0059] FIG. 36 shows an embodiment of epigenetic effectors for long-term gene upregulation via hit-and-run mechanism. See Example 12.

[0060] FIG. 37 shows Library 1 candidates.

[0061] FIG. 38 shows Library 2 candidates.

[0062] FIG. 39 shows sequences of exemplary epigenetic effectors and epigenetic effector fusions.ACTIVEUS 211418665 21Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0063] FIGS. 40A-C show a combination of KRAB and L3MBTL3 is potent across multiple gene targets. Figure 40A shows CD81 repression of 3x guide RNA expressing cells that have undergone plasmid nucleofection with the effector plasmid (stdMCP-L3MBTL3, stdMCP-KRAB, or stdMCP-KRAB-L3MBTL3). Figure 40B shows CD81 repression of cells that have undergone sequential lentiviral infection with effector (stdMCP-KRAB or stdMCP- KRAB-L3MBTL3) and lx guide RNA as a function of doxycycline concentration. Figure 40C shows NF 1 mRNA repression 2 days following nucleofection of with sgRNA plasmid to the cells that have undergone lentiviral infection with effector (stdMCP-KRAB or stdMCP- KRAB -L3 MB TL3). Doxycycline was added 1 day before the nucleofection of sgRNA plasmid to induce the effector components (n > 2 biological replicates). Distance between the sgRNA target site and the TSS is indicated in parentheses. Error bars indicate the standard deviation of biological replicates. * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001; two-tailed t-test.

[0064] FIG. 41 shows an embodiment of a growth screen for KRAB-L3MBTL3 efficacy. See Example 14.

[0065] FIG. 42 shows guide RNA depletion score of stdMCP-KRAB-L3MBTL3 expressing K562 cells versus guide RNA depletion score of dCas9 expressing K562 cells.

[0066] FIG. 43 shows guide RNA depletion score of stdMCP-KRAB-L3MBTL3 expressing K562 cells versus guide RNA depletion score of stdMCP-KRAB expressing K562 cells.

[0067] FIGS. 44A-C show KRAB + L3MBTL3 synergy test on different KRAB paralogs. L3MBTL3 is fused with the KRAB domain of the following KRAB paralogs: ZNF10, ZIM3, or ZNF705.

[0068] FIG. 45 shows exemplary sequences of sgRNAs used in FIGS. 44A-C.

[0069] FIGS. 46A-B show nucleic acid sequences of (A) ZIM3 (80 AA)-L3MBTL3, ZIM3 (99 AA)-L3MBTL3, ZNF705-L3MBTL3 dual effectors, and (B) ZIM3 (80 AA), ZIM3 (99 AA), and ZNF705 single effectors.

[0070] FIG. 47 shows amino acid sequences of (A) ZIM3 (80 AA)-L3MBTL3, ZIM3 (99 AA)-L3MBTL3, ZNF705-L3MBTL3 dual effectors and (B) ZIM3 (80 AA), ZIM3 (99 AA), and ZNF705 single effectors.

[0071] FIG. 48 shows a list of genes targeted for the growth screen. See Example 14.

[0072] FIG. 49 shows a comparison between dCas9-KRAB and L3MBTL3-dCas9-KRAB in Ngn2-induced neurons from human embryonic stem cells. The following constructsACTIVEUS 211418665 22Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 were used: L3MBTL3-XTEN80-dCas9-HA tag-2xSV40 NLS-TagBFP-GGGGG-KRAB (KOX1) and dCas9-HA tag-2xSV40 NLS-TagBFP-GGGGG-KRAB (KOX1). L3MBTL3- XTEN80 was added to the N-terminus of the dCas9-HA tag-2xSV40 NLS-TagBFP- GGGGG-KRAB (KOX1) construct, which is derived from pHR-SFFV-dCas9-BFP-KRAB (Addgene Plasmid #46911; Gilbert, Luke A., et al. CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell 154.2 (2013): 442-451). See Example 16.DETAILED DESCRIPTION

[0073] The present invention relates to a gene regulating system that incorporates one or more epigenetic effectors. In some embodiments, one or more epigenetic effectors are connected to an RNA-binding protein (e.g., MS2 bacteriophage coat protein (MCP)), in some embodiments via a linker. In some embodiments, one or more epigenetic effectors are connected to an RNA-guided nuclease (e.g., dCas9, dCasl2), in some embodiments via a linker. In some embodiments, the linker is XTEN80. By utilizing an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA-binding domain (e.g., Zinc Fingers or TALENs) that lacks nuclease and / or nickase activity, in conjunction with an RNA aptamer, the system is capable of modulating gene expression. By utilizing an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA-binding domain (e.g., Zinc Fingers or TALENs) with nuclease and / or nickase activity, in conjunction with an RNA aptamer, the system is capable of modulating gene expression. Specifically, it can enhance, reduce or inhibit the expression of one or more target genes.

[0074] The present invention relates to a gene regulating system that incorporates at least two epigenetic effectors. The two epigenetic effectors are connected as a fusion, in some embodiments via a linker. By utilizing an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA-binding domain (e.g., Zinc Fingers or TALENs) that lacks nuclease and / or nickase activity, in conjunction with an RNA aptamer, the system is capable of modulating gene expression. By utilizing an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA- binding domain (e.g., Zinc Fingers or TALENs) with nuclease and / or nickase activity, in conjunction with an RNA aptamer, the system is capable of modulating gene expression. Specifically, it can enhance, reduce or inhibit the expression of one or more target genes.

[0075] The practice of aspects of the present invention can employ, unless otherwise indicated, conventional techniques of cell biology, cell culture, molecular biology, transgenic biology, microbiology, recombinant DNA, and biochemistry, which are within the skill of theACTIVEUS 211418665 23Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 art. Such techniques are explained fully in the literature. See, e.g., Molecular Cloning A Laboratory Manual, 3rdEd., ed. By Sambrook (2001), Fritsch and Maniatis (Cold Spring Harbor Laboratory Press: 1989); DNA Cloning, Volumes I and II (D. N. Glover ed., 1985); Oligonucleotide Synthesis (M. J. Gait ed., 1984); Mullis et al. U.S. Pat. No: 4,683,195; Nucleic Acid Hybridization (B. D. Hames & S. J. Higgins eds. 1984); Transcription and Translation (B. D. Hames & S. J. Higgins eds. 1984); Culture Of Animal Cells (R. I. Freshney, Alan R. Liss, Inc., 1987); Immobilized Cells and Enzymes (IRL Press, 1986); B. Perbal, A Practical Guide To Molecular Cloning (1984); the series, Methods In Enzymology (Academic Press, Inc., N.Y.), specifically, Methods In Enzymology, Vols. 154 and 155 (Wu et al. eds.); Gene Transfer Vectors For Mammalian Cells (J. H. Miller and M. P. Calos eds., 1987, Cold Spring Harbor Laboratory); Immunochemical Methods In Cell And Molecular Biology (Caner and Walker, eds., Academic Press, London, 1987); Handbook Of Experimental Immunology, Volumes I-FV (D. M. Weir and C. C. Blackwell, eds., 1986); Manipulating the Mouse Embryo, (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1986) and subsequent versions thereof, the contents of each of which are hereby incorporated by reference in their entireties.

[0076] One skilled in the art can obtain a protein in several ways, which include, but are not limited to, isolating the protein via biochemical means or expressing a nucleotide sequence encoding the protein of interest by genetic engineering methods, including, but not limited to, cell-based methods and cell-free methods.

[0077] A protein is encoded by a nucleic acid (including, for example, genomic DNA, messenger RNA (mRNA), complementary DNA (cDNA), synthetic DNA, as well as any form of corresponding RNA). Nucleic acids encoding a protein can be produced via recombinant DNA technology and such recombinant nucleic acids can be prepared by conventional techniques, including chemical synthesis, genetic engineering, enzymatic techniques, or a combination thereof.Epigenetic Effector Fusions

[0078] The present invention relates to a fusion polypeptide of at least two epigenetic effectors, which are also referred to herein as epigenetic effector fusions. In some embodiments, a fusion polypeptide comprises heterologous polypeptides comprising a histone deubiquitinase, a histone acetyltransferase, a histone arginine methyltransferase, a histone demethylase, a histone lysine methyltransferase, a histone phosphatase, a E2 ubiquitin ligase, a E3 ubiquitin ligase, a histone kinase, a DNA demethylation machinery, a histoneACTIVEUS 211418665 24Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 deacetylase, a DNA methyltransferase, a Kriippel associated box (KRAB) domain, a sterile alpha motif (SAM) domain, polypeptides that can recruit Polycomb repressive complex 2 (PRC2) and any combination thereof. Additionally, a fusion polypeptide may be combined with other domains disclosed herein. Non-limiting examples of epigenetic effectors are disclosed in Tables 1 and 2 and Figures 39, 46 or 47.

[0079] The term ‘Kriippel associated box domain,’ ‘KRAB,’ or ‘KRAB domain’ refers to a class of transcriptional repression domains found in approximately 400 human zinc finger protein-based transcription factors. KRAB domains typically consist of about 45 to 75 amino acid residues. Detailed descriptions of KRAB domains, including their functions and uses, can be found in sources such as Ecco, G., Imbeault, M., Trono, D., KRAB zinc finger proteins, Development 144, 2017; Lambert et al., The human transcription factors, Cell 172, 2018; and Gilbert et al., Cell (2013, 2014), all of which are incorporated herein by reference in their entirety. In some embodiments, the KRAB domain comprises the KRAB domain of ZNF10. In some embodiments, the KRAB domain comprises the KRAB domain of ZNF705. In some embodiments, the KRAB domain comprises the KRAB domain of ZIM3. In some embodiments the KRAB domain comprises the KRAB domain of a KRAB paralog. In some embodiments, the KRAB domain consists of the KRAB domain of ZNF10. In some embodiments, the KRAB domain consists of the KRAB domain of ZNF705. In some embodiments, the KRAB domain consists of the KRAB domain of ZIM3. In some embodiments the KRAB domain consists of the KRAB domain of a KRAB paralog. See Table 1 and Figures 39, 46, and 47.

[0080] In some embodiments, a fusion peptide comprises other domains disclosed herein. Non-limiting examples of epigenetic effectors are disclosed in Tables 1 and 2 and Figures 39, 46, and 47.

[0081] In some embodiments, an epigenetic effector domain is not fused to another epigenetic effector domain, and is connected directly or indirectly (e.g. as a fusion with, or as part of a system) with an RNA-guided nuclease or DNA-binding protein which directs the epigenetic effector domain to a specific target site as described herein. In some embodiments, an epigenetic effector domain comprises the catalytic domain of KDM2B (SEQ ID NO: 76) which can function as an antagonistic partner of KRAB-mediated transcriptional repression. In some embodiments, an epigenetic effector domain comprises the catalytic domain of KDM2B (SEQ ID NO: 76) or the catalytic domain of KDM5B (SEQ ID NO: 77) not fused to another epigenetic effector and instead these domains are connectedACTIVEUS 211418665 25Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 directly or indirectly (e.g. as a fusion with, or as part of a system) with an RNA-guided nuclease or DNA-binding protein which directs the KDM2B or KDM5B to a specific target site. In some embodiments, the catalytic domain of KDM2B (SEQ ID NO: 76) is an antagonistic partner of ZNFlO-mediated transcriptional repression. In some embodiments, the catalytic domain of KDM2B (SEQ ID NO: 76) is an antagonistic partner of SID4X-mediated transcriptional repression. In some embodiments, an epigenetic effector domain comprises the catalytic domain of KDM5B (SEQ ID NO: 77) which can function as an antagonistic partner of KRAB-mediated transcriptional repression. In some embodiments, the catalytic domain of KDM5B (SEQ ID NO: 77) is an antagonistic partner of ZNFlO-mediated transcriptional repression. In some embodiments, the catalytic domain of KDM5B (SEQ ID NO: 77) is an antagonistic partner of SID4X-mediated transcriptional repression. In some embodiments, targeting KDM2B or KDM5B to target genes can be used to antagonize the repressive action of oncogenic KRAB-ZFPs. In some embodiments, this is useful to ameliorate cancer progression or treat or prevent cancer. In some embodiments, the target gene is a gene repressed by a KRAB-ZFP (see Figure 4 of Sobocinska J, Molenda S, Machnik M, Oleksiewicz U. KRAB-ZFP Transcriptional Regulators Acting as Oncogenes and Tumor Suppressors: An Overview. Int J Mol Sci. 2021 Feb 23;22(4):2212, the content of which is hereby incorporated by reference in its entirety). In some embodiments, the catalytic domain of KDM2B or KDM5B are connected to an RNA-binding protein (e.g., MS2 bacteriophage coat protein (MCP)), in some embodiments via a linker. In some embodiments, the catalytic domain of KDM2B or KDM5B are connected to an RNA-guided nuclease (e.g., dCas9, dCasl2), in some embodiments via a linker.

[0082] In some embodiments, the epigenetic effector fusion comprises the epigenetic fusions disclosed in Table 1. In some embodiments, the epigenetic effector fusion comprises a linker between the epigenetic effectors listed in Table 1. The use of “Epigenetic Effector 1- Epigenetic Effector 2” fusion is intended to encompass both embodiments unless specified otherwise (i.e., in “KRAB-L3MBTL3” indicates both a direct bond or a linker between the KRAB and L3MBLT3 portions of the KRAB-L3MBTL3 fusion protein). In some embodiments, the epigenetic effector fusions comprise Epigenetic Effector 1 and Epigenetic Effector 2 in either N- to C- terminal orientation. For example, the amino acid sequence of “KRAB-L3MBTL3” is provided as SEQ ID NO: 1 and is exemplified as a fusion of KRAB N-terminal to the SAM domain of L3MBTL3, however, it should be understood that an epigenetic effector fusion can comprise a fusion of the SAM domain of L3MBTL3 N-ACTIVEUS 211418665 26Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 terminal to KRAB. Thus, in some embodiments the epigenetic effector fusions comprise Epigenetic Effector 1 of Table 1 fused N-terminal to Epigenetic Effector 2 of Table 1. In some embodiments the epigenetic effector fusions comprise Epigenetic Effector 2 of Table 1 fused N-terminal to Epigenetic Effector 1 of Table 1. In some embodiments, the inventive fusions are connected, directly or indirectly, to a CRISPR / Cas system, which directs the fusions to a specific target site. In some embodiments, the inventive fusions are connected, directly or indirectly, to a DNA-binding domain (e.g., zinc finger; TALEs), which directs the fusions to a specific target site. Without being bound by theory, these fusions will enhance, reduce or inhibit the expression of one or more target genes.

[0083] In some embodiments, the epigenetic effector or epigenetic effector fusion is connected to a CRISPR / Cas system. In some embodiments, Epigenetic Effector 1 is connected to a CRISPR / Cas system at either the N-terminus or the C-terminus of the CRISPR / Cas system. In some embodiments, Epigenetic Effector 2 is connected to a CRISPR / Cas system at either the N-terminus or the C-terminus of the CRISPR / Cas system. In some embodiments, KRAB is connected to dCas9 at either the N-terminus or the C- terminus of dCas9. In some embodiments, L3MBTL3 (e.g., SAM domain) is connected to dCas9 at either the N-terminus or the C-terminus of dCas9. In some embodiments, dCas9 is positioned between KRAB and L3MBTL3 (e.g., SAM domain). In some embodiments, the epigenetic effector fusion is connected to a CRISPR / Cas system via a linker. In some embodiments, the epigenetic effector fusion is connected to a CRISPR / Cas system using more than one linker. In some embodiments, epigenetic effector fusion comprises KRAB, a CRISPR / Cas system, L3MBTL3 (e.g., SAM domain) and at least one linker. In some embodiments, the linker is XTEN80.

[0084] Table 1 shows epigenetic effector pairs that regulate transcription. See FIGs. 39, 46 and 47.Table 1ACTIVEUS 211418665 27Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025ACTIVEUS 211418665 28Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0085] In some embodiments, the epigenetic effector fusion comprises a KRAB domain (e.g., KRAB domain of ZNF10, ZNF705, or ZIM3) and a SAM domain (e.g., SAM domain of L3MBTL3). KRAB domain refers to Kriippel associated box (KRAB) domain, which is described as InterPro domain number IPR001909. In some embodiments, the KRAB domain comprises the KRAB domain of any of the proteins identified on InterPro entry IPR001909 as having a KRAB domain (see www.ebi.ac.uk / interpro / entry / InterPro / IPR001909 / protein / UniProt / #table, incorporated herein by reference in its entirety). In some embodiments, the KRAB domain comprises the KRAB domain of any of the reviewed proteins identified on InterPro entry IPR001909 as having a KRAB domain. In some embodiments, the KRAB domain comprises the KRAB domain of any of the reviewed human proteins identified on InterPro entry IPR001909 as having a KRAB domain. In some embodiments, the KRAB domain comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%,ACTIVEUS 211418665 29Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 or 99% identity to the amino acid sequence of the KRAB domain of ZNF10 (SEQ ID NO:27). In some embodiments, the KRAB domain comprises the KRAB domain of ZNF10 (SEQ ID NO:27). In some embodiments, the KRAB domain comprises the KRAB domain of ZNF705 (SEQ ID NO: 163). In some embodiments, the KRAB domain comprises the KRAB domain of ZIM3 (SEQ ID NOS: 161-162). In some embodiments, the KRAB domain comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the KRAB domain of ZNF705 (SEQ ID NO: 163). In some embodiments, the KRAB domain comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the KRAB domain of ZIM3 (SEQ ID NOS: 161-162).

[0086] In some embodiments, the SAM domain (e.g., SAM domain of L3MBTL3) is used to enhance KRAB-mediated transcription repression. SAM domain refers to sterile alpha motif domain, which is described as InterPro domain number IPR001660. In some embodiments, the SAM domain comprises the SAM domain of any of the proteins identified on InterPro entry IPR001660 as having a SAM domain (see www.ebi.ac.uk / interpro / entry / InterPro / IPR001660 / protein / UniProt / #table, incorporated herein by reference in its entirety). In some embodiments, the SAM domain comprises the SAM domain of any of the reviewed proteins identified on InterPro entry IPR001660 as having a SAM domain. In some embodiments, the SAM domain comprises the SAM domain of any of the reviewed human proteins identified on InterPro entry IPR001660 as having a SAM domain. In some embodiments, the SAM domain comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the SAM domain comprises the SAM domain of L3MBTL3 (SEQ ID NO:28).

[0087] KRAB-mediated transcriptional repressors are widely used in a variety of configurations with diverse applications. See e.g., Ecco, Gabriela, Michael Imbeault, and Didier Trono. "KRAB zinc finger proteins." Development 144.15 (2017): 2719-2729. In some embodiments, described herein is an epigenetic effector fusion comprising a KRAB repressor system further comprising a SAM domain (e.g., SAM domain of L3MBTL3). In some embodiments, described herein is an epigenetic effector fusion comprising a KRAB repressor system further comprising an amino acid sequence having 70%, 75%, 80%, 85%,ACTIVEUS 211418665 30Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 202590%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, described herein is an epigenetic effector fusion comprising a KRAB repressor system further comprising an amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the KRAB repressor system further comprises an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA-binding domain (e.g., Zinc Fingers or TALENs) in either N- to C-terminal orientation. For example, the KRAB repressor system comprises dCas9-KRAB, dCasl2-KRAB, nCasl2-KRAB (nickase), KRAB-zinc finger (ZNF) ((U.S. Pub. Pat. No. 20240018203A1) Novel zinc finger protein transcription factors for repressing alpha-synuclein expression; Monteferrario, Davide, et al. "Epigenetic control of multiple genes with a lentiviral vector encoding transcriptional repressors fused to compact zinc finger arrays." Molecular Therapy Methods & Clinical Development 32.2 (2024)), or KRAB-TALE (Cong, Le, et al. "Comprehensive interrogation of natural TALE DNA-binding modules and transcriptional repressor domains." Nature communications 3.1 (2012): 968). For example, the KRAB repressor system comprises KRAB-dCas9 (Gilbert, Luke A., et al. "CRISPR- mediated modular RNA-guided regulation of transcription in eukaryotes." Cell 154.2 (2013): 442-451; Gilbert, Luke A., et al. "Genome-scale CRISPR-mediated control of gene repression and activation." Cell 159.3 (2014): 647-661; Addgene Plasmid #46911), KRAB- dCasl2 (Campa, Carlo C., et al. "Multiplexed genome engineering by Casl2a and CRISPR arrays encoded on single transcripts." Nature Methods 16.9 (2019): 887-893; Guo, Lucie Y., et al. "Multiplexed genome regulation in vivo with hyper-efficient Cast 2a." Nature cell biology 24.4 (2022): 590-600), KRAB-nCasl2 (nickase) (Hsiung, CC-S., et al. "Engineered CRISPR-Casl2a for higher-order combinatorial chromatin perturbations." Nature Biotechnology (2024): 1-15.), zinc-finger (ZNF)-KRAB or TALE-KRAB. In some embodiments, the epigenetic effector fusion comprises a KRAB-dCas9 or a dCas9-KRAB further comprising a SAM (e.g., SAM domain of L3MBTL3). In some embodiments, the epigenetic effector fusion comprises a KRAB-dCas9 or a dCas9-KRAB further comprising an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB- dCas9 or a dCas9-KRAB further comprising an amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-dCasl2 or a dCasl2-KRAB further comprising a SAM domain (e.g., SAM domainACTIVEUS 211418665 31Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 of L3MBTL3). In some embodiments, the epigenetic effector fusion comprises a KRAB- dCasl2 or a dCasl2-KRAB further comprising an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-dCasl2 or a dCasl2-KRAB further comprising an amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-ZNF or a ZNF-KRAB further comprising a SAM (e.g., SAM domain of L3MBTL3). In some embodiments, the epigenetic effector fusion comprises a KRAB-ZNF or a ZNF-KRAB further comprising an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-ZNF further comprising an amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-TALE or a TALE-KRAB further comprising a SAM (e.g., SAM domain of L3MBTL3). In some embodiments, the epigenetic effector fusion comprises a KRAB-TALE or a TALE-KRAB further comprising an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SAM domain of L3MBTL3 (SEQ ID NO:28). In some embodiments, the epigenetic effector fusion comprises a KRAB-TALE or a TALE-KRAB further comprising an amino acid sequence of the SAM domain of L3MBTL3 (SEQ ID NO:28). The SAM domain can be incorporated into a KRAB-repressor sequence either at the N- or C-terminus of the KRAB domain, and optionally via a linker as described herein. In certain aspects, described herein is a nucleic acid encoding these epigenetic effector fusions comprising a KRAB domain and SAM domain.

[0088] In some embodiments, an epigenetic modulation system comprises a catalytic domain of KDM2B (SEQ ID NO: 76) or KDM5B (SEQ ID NO: 77), which in some embodiments acts as an antagonist to a KRAB repressor. In some embodiments, the system comprises an RNA-guided nuclease (e.g., CRISPR / Cas systems) or a DNA-binding domain (e.g., Zinc Fingers or TALENs).

[0089] In some embodiments, the epigenetic effector fusion comprises the epigenetic fusions disclosed in Table 2. In some embodiments, the epigenetic effector fusion comprises a linker between two epigenetic effectors listed in Table 2. The use of “Epigenetic Effector 1-ACTIVEUS 211418665 32Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Epigenetic Effector 2” is intended to encompass both embodiments unless specified otherwise (i.e., in “RNF20-TET1” indicates both a direct bond or a linker between the RNF20 and TET1 portions of the RNF20-TET1 fusion protein). In some embodiments, the epigenetic effector fusions comprise Epigenetic Effector 1 and Epigenetic Effector 2 in either N- to C- terminal orientation. For example, the amino acid sequence of “RNF20-TET1” is provided as SEQ ID NO: 29 and is exemplified as a fusion of RNF20 N-terminal to TET1, however, it should be understood that an epigenetic effector fusion can comprise a fusion of the TET1 N-terminal to RNF20. Thus, in some embodiments the epigenetic effector fusions comprise Epigenetic Effector 1 of Table 2 fused N-terminal to Epigenetic Effector 2 of Table 2. In some embodiments the epigenetic effector fusions comprise Epigenetic Effector 2 of Table 2 fused N-terminal to Epigenetic Effector 1 of Table 2. In some embodiments, the inventive fusions are connected, directly or indirectly, to a CRISPR / Cas system, which directs the fusions to a specific target site. In some embodiments, the inventive fusions are connected, directly or indirectly, to a DNA-binding domain (e.g., zinc finger; TALEs), which directs the fusions to a specific target site. Without being bound by theory, these fusions will induce long-term enhanced, reduced or inhibited expression of one or more target genes. In some embodiments, the long-term effect lasts at least 5 days, 10 days, 15 days, 20 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, or 60 days.

[0090] Table 2 shows epigenetic effector pairs that induce long-term silencing or longterm enhancement of the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). See FIG. 39.Table 2ACTIVEUS 211418665 33Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025ACTIVEUS 211418665 34Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0091] In some embodiments, the epigenetic effector fusion comprises a TET1 DNA demethylase and an epigenetic effector. In some embodiments, the epigenetic effector is an E3 ubiquitin ligase domain (e.g., the RING finger domain of RNF20). In some embodiments, the epigenetic effector is a histone methyltransferase domain (e.g., the SET domain of SMYD1). In some embodiments, the epigenetic effector is a HKMT-like domain (e.g., the PR / SET domain of PRDM1). In some embodiments, TET1 comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of SEQ ID NO:57. In some embodiments, TET1 comprises SEQ ID NO:57. In some embodiments, the RING finger domain of RNF20 comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%,ACTIVEUS 211418665 35Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 202594%, 95%, 96%, 97%, 98%, or 99% identity to SEQ ID NO: 58. In some embodiments, the RING finger domain of RNF20 comprises SEQ ID NO:58. In some embodiments, the SET domain of SMYD1 comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to SEQ ID NO:59. In some embodiments, the SET domain of SMYD1 comprises SEQ ID NO:59. In some embodiments, the PR / SET domain of PRDM1 comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to SEQ ID NO:60. In some embodiments, the PR / SET domain of PRDM1 comprises SEQ ID NO:60. The epigenetic effector domain can be incorporated into a TET1 sequence either at the N- or C- terminus and optionally via a linker as described herein. In certain aspects, described herein is a nucleic acid encoding these epigenetic effector fusions comprising TET1 and an epigenetic effector.

[0092] In some embodiments, any of the epigenetic effector fusions described herein can further comprise a localization (e.g. nuclear import or export) signal. For example, one or more nuclear localization signal(s) can be included at the N-terminus, C-terminus, or both of the epigenetic effector fusion.Epigenetic Effectors

[0093] An epigenetic effector is a molecule or protein that influences gene expression without altering the underlying DNA sequence. These effectors can modify the structure of chromatin (the complex of DNA and proteins in the nucleus) or interact with other molecules to regulate how genes are turned on or off. Epigenetic effectors are composed of diverse classes including readers, recruiters, structural factors, and catalytic writers or erasers.

[0094] The term “epigenetic effector,” as used herein, refers to a protein or a domain of a protein that modulates epigenetic regulation. The epigenetic effector may be an epigenetic reader or a domain of an epigenetic reader which recognizes and binds to specific chemical modifications on histones or DNA. The epigenetic effector may be an epigenetic recruiter or a domain of an epigenetic recruiter which helps guide other proteins, such as writers, erasers or readers to a specific location on the DNA or chromatin. The epigenetic effector may be an epigenetic writer or a domain of an epigenetic writer which adds specific chemical groups to DNA or histones. The epigenetic effector may be an epigenetic eraser or a domain of an epigenetic eraser which removes specific chemical modifications from DNA or histones, effectively reversing the changes made by epigenetic writers.ACTIVEUS 211418665 36Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0095] In some embodiments, two or more epigenetic reader, recruiter, writer and eraser domains are combined to form an epigenetic effector fusion. See Figure IB. In some embodiments, the epigenetic effector fusion comprises a fusion of two or more different classes of epigenetic effector. For example, the epigenetic effector fusion comprises an epigenetic writer and an epigenetic eraser. See Figure IB and 1C. In some embodiments, the epigenetic effector fusion comprises an epigenetic reader and an epigenetic recruiter. In some embodiments, the epigenetic effector fusion comprises an epigenetic reader and an epigenetic writer. In some embodiments, the epigenetic effector fusion comprises an epigenetic reader and an epigenetic eraser. In some embodiments, the epigenetic effector fusion comprises an epigenetic recruiter and an epigenetic writer. In some embodiments, the epigenetic effector fusion comprises an epigenetic recruiter and an epigenetic eraser.

[0096] Examples of epigenetic effectors provided herein or useful in the nucleic acids, polypeptides, compositions, systems and methods disclosed herein include, but not limited to sterile alpha motif (SAM) domain of L3MBTL, PALI1, EPOP, DNMT3A-3L, DNMT3A, TET1, KDM2A, or KDM5B. See Tables 1 and 2 Figures 39, 46, and 47.

[0097] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors described herein. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors described herein. In some embodiments, the effector fusion comprises (a) the amino acid sequence of any of the epigenetic effectors disclosed in Tables 1 and 2, or (b) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (a). In some embodiments, the effector fusion consists of (a) the amino acid sequence of any of the epigenetic effectors disclosed in Tables 1 and 2, or (b) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (a).

[0098] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors comprising the amino acid sequence of any of the epigenetic effectors disclosed in Tables 1 and 2. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors comprising the amino acid sequence of any of epigenetic effectors disclosed in Tables 1 and 2. In some embodiments, the amino acid sequence of one or both of the epigenetic effectors of the epigenetic effector fusion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%,ACTIVEUS 211418665Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the epigenetic effectors disclosed in Tables 1 and 2.

[0099] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors described herein. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors described herein. In some embodiments, the effector fusion comprises (a) the amino acid sequence of any of the epigenetic effectors disclosed in Figures 39 and 47, or (b) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (a). In some embodiments, the effector fusion consists of (a) the amino acid sequence of any of the epigenetic effectors disclosed in Figures 39 or 47, or (b) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (a).

[0100] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors comprising the amino acid sequence of any of the epigenetic effectors disclosed in Figures 39 or 47. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors comprising the amino acid sequence of any of epigenetic effectors disclosed in Figures 39 and 46. In some embodiments, the amino acid sequence of one or both of the epigenetic effectors of the epigenetic effector fusion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the epigenetic effectors disclosed in Figures 39 or 47.

[0101] In some embodiments, the epigenetic effector fusion comprises ZNF10 KRAB domain (SEQ ID No: 27) and sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28). In some embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163) and sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID Nos: 161-162) and sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28). In some embodiments, the epigenetic effector fusion comprises ZNF10 KRAB domain (SEQ ID No: 27) and PALI1 (SEQ ID No: 61). In some embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163) and PALI1 (SEQ ID No: 61). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID Nos: 161-162) and PALI1 (SEQ ID No: 61). In some embodiments, the epigenetic effector fusion comprises ZNF10 KRAB domain (SEQ ID No: 27) and EPOP (SEQ ID No: 62). In someACTIVEUS 211418665 38Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163) and EPOP (SEQ ID No: 62). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID Nos: 161-162) and EPOP (SEQ ID No: 62). In some embodiments, the amino acid sequence of one or both of the epigenetic effectors of the epigenetic effector fusion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to these amino acid sequences.

[0102] In some embodiments, the epigenetic effector fusion comprises TET1 and any of the epigenetic effectors disclosed in Tables 1 and 2 or Figures 39, 46, or 47. In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID No: 57). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and RNF20 (SEQ ID No: 58). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and RNF20 (K959E) (SEQ ID No: 63). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and SMYD1 (SEQ ID No: 64). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and SMYD1 (R19A) (SEQ ID No: 64). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and PRDM1 (SEQ ID No: 60). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57) and PRDM1 (Y200A) (SEQ ID No: 65). In some embodiments, the amino acid sequence of one or both of the epigenetic effectors of the epigenetic effector fusion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to these amino acid sequences.

[0103] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors, wherein the epigenetic effector comprises epigenetic effector means for mediating gene regulation of a target gene. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors, wherein the epigenetic effector comprises epigenetic effector means for mediating gene regulation of a target gene.Linkers

[0104] In some embodiments, the fusion between the epigenetic effectors may include a linker. In some embodiments, the epigenetic effectors may further be fused to an RNA- binding protein, an RNA-guided nuclease, or a DNA binding protein via a linker. The term “linker,” as used herein, refers to a chemical group or a molecule linking two molecules or moieties. Typically, the linker is positioned between, or flanked by, two groups, molecules, or other moieties and connected to each one via a covalent bond, thus connecting the two. InACTIVEUS 211418665 39Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 some embodiments, the linker is an amino acid or a plurality of amino acids (e.g., a peptide or protein). In some embodiments, the linker is an organic molecule, group, polymer, or chemical moiety. In some embodiments, the linker may comprise a peptide or a non-peptide moiety. In some embodiments, the linker is 2-100 amino acids in length, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 30-35, 35-40, 40-45, 45-50, 50-60, 60-70, 70-80, 80-90, 90-100, 100-150, or 150-200 amino acids in length. Longer or shorter linkers are also contemplated.

[0105] In some embodiments, XTEN linkers are used in the fusions described herein.For example, in some embodiments, XTEN16 (SGSETPGTSESATPES (SEQ ID NO: 66)) is used. In some embodiments, XTEN80 (SEQ ID NO: 67) is used. In some embodiments, additional XTEN16 or XTEN80 repeats can be used to provide suitable lengths, as required. In some embodiments, XTEN80 (SEQ ID NO: 67) is connected to L3MBTL3 (e.g., SAM domain) at either N-terminus or the C-terminus of L3MBTL3 (e.g., SAM domain).

[0106] In some embodiments, a linker for use in the fusions described herein can comprise a combination of one or more of a GlySer linker, an XTEN linker, and / or a 2A self- cleaving peptides described above.

[0107] In other embodiments, the linker is at least 3 amino acids, at least 4 amino acids, at least 5 amino acids, at least 6 amino acids, at least 7 amino acids, at least 8 amino acids, at least 9 amino acids, at least 10 amino acids, at least 11 amino acids, at least 12 amino acids, at least 13 amino acids, at least 14 amino acids, at least 15 amino acids, at least 16 amino acids, at least 17 amino acids, at least 18 amino acids, at least 19 amino acids, at least 20 amino acids, at least 30 amino acids, at least 40 amino acids, at least 50 amino acids, at least 60 amino acids, at least 70 amino acids, at least 80 amino acids, at least 90 amino acids, at least 100 amino acids, at least 200 amino acids, at least 300 amino acids, at least 400 amino acids or at least 500 amino acids in length.

[0108] In some embodiments, the epigenetic effectors are fused directly to another epigenetic effector by a covalent bond. In some embodiments, the epigenetic effector fusions are fused directly to an RNA-binding protein or to an RNA-guided nuclease or to a DNA binding protein by a covalent bond. In certain embodiments, the covalent bond is a peptide bond, carbon-carbon bond, disulfide bond, carbon-heteroatom bond, a carbon-nitrogen bond of an amide linkage, etc. In certain embodiments, the epigenetic effectors are fused by a linker that is a peptide or based on amino acids. In other embodiments, the linker is not peptide-like. In certain embodiments, the linker is a cyclic or acyclic, substituted orACTIVEUS 211418665 40Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 unsubstituted, branched or unbranched aliphatic or heteroaliphatic linker. In certain embodiments, the linker is polymeric (e.g., polyethylene, polyethylene glycol, polyamide, polyester, etc.). In certain embodiments, the linker comprises a monomer, dimer, or polymer of aminoalkanoic acid. In certain embodiments, the linker comprises an aminoalkanoic acid (e.g., glycine, ethanoic acid, alanine, beta-alanine, 3- aminopropanoic acid, 4-aminobutanoic acid, 5-pentanoic acid, etc.). In certain embodiments, the linker comprises a monomer, dimer, or polymer of aminohexanoic acid (Ahx). In certain embodiments, the linker is based on a carbocyclic moiety (e.g., cyclopentane, cyclohexane). In other embodiments, the linker comprises a polyethylene glycol moiety (PEG). In certain embodiments, the linker comprises an aryl or heteroaryl moiety. In certain embodiments, the linker is based on a phenyl ring. The linker may include functionalized moieties to facilitate attachment of a nucleophile (e.g., thiol, amino) from the peptide to the linker. Any electrophile may be used as part of the linker. Exemplary electrophiles include, but are not limited to, activated esters, activated amides, Michael acceptors, alkyl halides, aryl halides, acyl halides, and isothiocyanates. In other embodiments, the linker comprises amino acids. In certain embodiments, the linker comprises a peptide.

[0109] In certain aspects, described herein is an epigenetic effector fusion wherein the two or more epigenetic effectors are fused directly. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion wherein a first epigenetic effector is fused directly to a second epigenetic effector. In some embodiments, the epigenetic effectors may further be directly fused to an RNA-binding protein or to an RNA-guided nuclease or a DNA-binding protein. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion wherein an epigenetic effector fusion is fused directly to an RNA- binding protein or to an RNA-guided nuclease or to a DNA-binding protein. In certain aspects, described herein are epigenetic effectors fused together via a peptide linker. In some embodiments, the epigenetic effectors may further be fused to an RNA-binding protein or to an RNA-guided nuclease or to a DNA-binding protein via a peptide linker. In some embodiments, the peptide linker is 2 to 100 amino acids long. In some embodiments, the peptide linker is 2 to 50 amino acids long. In some embodiments, the peptide linker is 2 to 30 amino acids long. In some embodiments, the peptide linker comprises glycine and serine residues. In some embodiments, the peptide linker comprises only glycine and serine residues. In some embodiments, the peptide linker is 2 to 30 amino acids long and comprises only glycine and serine residues. In some embodiments, the peptide linker is 24 amino acidsACTIVEUS 211418665 41Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 long and comprises only glycine and serine residues. In some embodiments, the peptide linker is 30 amino acids long and comprises only glycine and serine residues. In some embodiments, the peptide linker comprises GGS repeats. In some embodiments, the peptide linker comprises an XTEN16 sequence. In some embodiments, the peptide linker consists of an XTEN16 sequence. In some embodiments, the peptide linker comprises an XTEN80 sequence. In some embodiments, the peptide linker comprises multiple XTEN16 or XTEN80 repeats. In some embodiments, the peptide linker consists of multiple XTEN16 or XTEN80 repeats. In some embodiments, the peptide linker consists of an XTEN80 sequence. In some embodiments, the peptide linker comprises an F2A, E2A, P2A or T2A sequence. In some embodiments, the peptide linker consists of an F2A, E2A, P2A or T2A sequence. In some embodiments, the peptide linker comprises an XTEN16 sequence and one or more glycine, serine, proline, or glutamate residues at the N- or C-terminus of the XTEN16 sequence. In some embodiments, the peptide linker comprises an XTEN80 sequence and one or more glycine, serine, proline, or glutamate residues at the N- or C-terminus of the XTEN80 sequence.

[0110] In certain aspects, described herein is an epigenetic effector fusion comprising a peptide linker means for fusing two or more epigenetic effectors. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising a peptide linker means for fusing two or more epigenetic effectors. In certain aspects, described herein is an epigenetic effector fusion comprising a peptide linker means for fusing an epigenetic effector fusion to an RNA-binding protein or an RNA-guided nuclease or a DNA-binding protein. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising a peptide linker means for fusing an epigenetic effector fusion to an RNA- binding protein or an RNA-guided nuclease or a DNA-binding protein. In some embodiments, the linker sequence comprises an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the linker sequence disclosed in SEQ ID NO: 1 or 7.[OHl] In some embodiments, the fusion protein further comprises or consists essentially of or consists of a localization (nuclear import or export) signal as, or as part of, the linker between the epigenetic effectors. HA or Flag tags are also within the gambit of the invention as linkers. The linkers allow the user to engineer appropriate amounts of “mechanical flexibility”.ACTIVEUS 211418665 42Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0112] Longer linkers are preferable as well, (e.g., “(GGSS)?”, “(GGGGS)e” and “(GGS)8”) are disclosed as SEQ ID Nos: 68-70. Linker flexibility is also a factor, as more flexible linkers (GGS and GGGGS (SEQ ID NO: 82)) are preferable than more rigid linkers (XTEN16) in the epigenetic effector fusions.

[0113] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors and at least one linker described herein. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors and at least one linker portion described herein. In some embodiments, the epigenetic effector fusion comprises two or more epigenetic effectors and a linker, each epigenetic effector comprising (a) the amino acid sequence of any of the epigenetic effectors disclosed in Tables 1 and 2 or Figures 39 or 47, (b) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (a), or (c) epigenetic effector means for mediating gene regulation of a target gene; and the linker comprising (d) a peptide linker, (e) a peptide linker comprising one or more glycine-serine repeats, (f) a peptide linker comprising one or more XTEN linkers, (g) a peptide linker comprising one or more glycineserine repeats and one or more XTEN linkers, (h) an amino acid comprising the linker sequences disclosed in SEQ ID NOs: 1 or 7, (i) an amino acid sequence having 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of (h), or (j) peptide linker means for fusing together two or more epigenetic effectors. In some embodiments, the epigenetic effector fusion further comprises an RNA- binding protein or an RNA-guided nuclease or a DNA-binding protein, and at least one linker described herein linking the epigenetic effector fusion and the RNA-binding protein or RNA- guided nuclease or DNA-binding protein.

[0114] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors comprising the amino acid sequence of the epigenetic effectors disclosed in Tables 1 and 2 or Figures 39 and 47, and a linker comprising the amino acid sequence of the linker sequences disclosed in SEQ ID Nos.: 1 or 7. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors comprising any of the amino acid sequences disclosed in Tables 1 and 2 or Figures 39 and 46, and a linker portion comprising the amino acid sequence of the linker sequences disclosed SEQ ID NOs: 1 or 7. In some embodiments, the amino acid sequence of the epigenetic effectors has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%,ACTIVEUS 211418665 43Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the epigenetic effectors disclosed in Tables 1 and 2 or Figures 39 or 47. In some embodiments, the amino acid sequence of the linker portion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the linker sequences disclosed in SEQ ID NOs: 1 or 7. In some embodiments, the epigenetic effector fusion further comprises an RNA-binding protein or an RNA-guided nuclease or a DNA-binding protein, and at least one linker described herein linking the epigenetic effector fusion and the RNA-binding protein or RNA-guided nuclease or DNA-binding protein.

[0115] In some embodiments, the epigenetic effector fusion comprises ZNF 10 KRAB domain (SEQ ID No: 27), sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 1). In some embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163), sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 1). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID Nos: 161-162), sterile alpha motif (SAM) domain of L3MBTL3 (SEQ ID No: 28) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g, SEQ ID NO: 1). In some embodiments, the epigenetic effector fusion comprises ZNF10 KRAB domain (SEQ ID No: 27), PALI1 (SEQ ID No: 31) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 3). In some embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163), PALI1 (SEQ ID No: 31) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 3). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID Nos: 161-162), PALI1 (SEQ ID No: 31) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 3). In some embodiments, the epigenetic effector fusion comprises ZNF 10 KRAB domain (SEQ ID No: 27), EPOP (SEQ ID No: 62) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 5). In some embodiments, the epigenetic effector fusion comprises ZNF705 KRAB domain (SEQ ID No: 163), EPOP (SEQ ID No: 62) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 3). In some embodiments, the epigenetic effector fusion comprises ZIM3 KRAB domain (SEQ ID No: 161-162), EPOP (SEQ ID No: 62) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 3). In some embodiments, the amino acid sequence of the epigenetic effectors has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to these amino acid sequences. In some embodiments, the amino acid sequence of the linker portion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%,ACTIVEUS 211418665 44Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202598%, or 99% identity to the amino acid sequence of the linker sequence disclosed in SEQ ID NOs: 1 or 7.

[0116] In some embodiments, the epigenetic effector fusion comprises TET1 and any of the epigenetic effectors disclosed in Tables 1 and 2 or Figures 39, 46, or 47 and a linker disclosed in SEQ ID NO: 1 or 7. In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID No: 57) and a linker disclosed in SEQ ID NO: 1 or 7. In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), RNF20 (SEQ ID No: 58) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 29). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), RNF20 (K959E) (SEQ ID No: 63) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 31). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), SMYD1 (SEQ ID No: 64) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 33). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), SMYD1 (R19A) (SEQ ID No: 64) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 35). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), PRDM1 (SEQ ID No: 60) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 37). In some embodiments, the epigenetic effector fusion comprises TET1 (SEQ ID NO: 57), PRDM1 (Y200A) (SEQ ID No: 65) and a linker disclosed in SEQ ID NO: 1 or 7 (e.g., SEQ ID NO: 39). In some embodiments, the amino acid sequence of the epigenetic effectors has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to these amino acid sequences. In some embodiments, the amino acid sequence of the linker portion has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence of the linker sequence disclosed in SEQ ID NOs: 1 or 7.

[0117] In certain aspects, described herein is an epigenetic effector fusion comprising two or more epigenetic effectors comprising epigenetic effector means for mediating gene regulation of a target gene and at least one peptide linker means for fusing two or more epigenetic effectors. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion comprising two or more epigenetic effectors comprising epigenetic effectors means for mediating gene regulation of a target gene and at least one peptide linker means for fusing together two or more epigenetic effectors.

[0118] In certain aspects, described herein is an epigenetic effector fusion comprising the amino acid sequences provided in Tables 1 and 2 or Figures 39 and 47. In someACTIVEUS 211418665 45Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 embodiments, the amino acid sequence has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the amino acid sequence provided in Tables 1 and 2 or Figures 39 and 47. In certain aspects, described herein is a nucleic acid encoding an epigenetic effector fusion consisting of the amino acid sequences provided Tables 1 and 2 or Figures 39 and 46. In some embodiments, the nucleic acid sequence has 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity to the nucleic acid sequences provided in Tables 1 and 2 or Figures 39 and 46. In some embodiments, the epigenetic effector fusion further comprises an RNA-binding protein or an RNA-guided nuclease or a DNA-binding protein, and at least one linker described herein linking the epigenetic effector fusion and the RNA-binding protein or RNA-guided nuclease or DNA- binding protein.

[0119] In any of the embodiments described herein, a nucleotide sequence encoding the epigenetic effector fusion polypeptide or the epigenetic effector and / or linker portions and / or RNA-binding protein or an RNA-guided nuclease portions or a DNA-binding protein thereof can be codon-optimized. This type of optimization is known in the art and entails the mutation of foreign-derived DNA to mimic the codon preferences of the intended host organism or cell while encoding the same protein. Thus, the codons are changed, but the encoded protein remains unchanged. For example, if the intended target cell was a human cell, a human codon-optimized KRAB protein (or variant) would be a suitable epigenetic effector. Any suitable epigenetic effector can be codon optimized. As another non-limiting example, if the intended host cell were a mouse cell, then a mouse codon-optimized KRAB protein (or variant) would be a suitable epigenetic effector. While codon optimization is not required, it is acceptable and may be preferable in certain cases.Epigenetic Editing System

[0120] In certain aspects, described herein is an epigenetic editing system comprising a first nucleic acid encoding an epigenetic effector fusion as described herein tagged with a RNA-binding protein (e,g., MS2 bacteriophage coat protein (MCP), PP7 bacteriophage coat protein (PCP), lambda bacteriophage antiterminator protein N (kN), Qbeta coat protein (QPCP)), a second nucleic acid encoding a gRNA tagged with a RNA aptamer (e.g. MS2, PP7, boxB, QP), and a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2). It should be understood that the systems, vectors, cell lines etc. described herein referring to epigenetic effector fusions also refer to embodiments described herein where theACTIVEUS 211418665 46Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 epigenetic effector is not fused to another epigenetic effector (e.g., embodiments using the catalytic domain of KDM2B or KDM5B).

[0121] In some embodiments, the epigenetic editing system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with a RNA-binding protein (e,g., MCP, PCP, AN, QPCP), wherein the epigenetic effector fusion represses, reduces, or inhibits the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.), a second nucleic acid encoding a gRNA tagged with a RNA aptamer (e.g. MS2, PP7), and a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2) . In some embodiments, the epigenetic editing system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with a RNA-binding protein (e,g., MCP, PCP, AN, QPCP), wherein the epigenetic effector fusion activates or increases the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.), a second nucleic acid encoding a gRNA tagged with a RNA aptamer (e.g. MS2, PP7), and a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2).

[0122] In certain aspects, described herein is an epigenetic editing system comprising a first nucleic acid encoding an epigenetic effector fusion as described herein tagged with an RNA-guided nuclease (e.g., dCas9, dCasl2), and a second nucleic acid encoding a gRNA.

[0123] In certain aspects, described herein is an epigenetic editing system comprising a nucleic acid encoding an epigenetic effector fusion as described herein tagged with an DNA binding protein (e.g., zinc-finger (ZF), or transcription activator-like effector (TALE)). In some embodiments, the epigenetic effector fusion is a KRAB-SAM domain of L3MBTL3 fusion.

[0124] In some embodiments, the epigenetic editing system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with an RNA-guided nuclease (e.g., dCas9, dCasl2), wherein the epigenetic effector fusion represses, reduces, or inhibits the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) and a second nucleic acid encoding a gRNA. In some embodiments, the epigenetic editing system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with an RNA-guided nuclease (e.g., dCas9, dCasl2), wherein the epigenetic effector fusion activates or increases the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) and a second nucleic acid encoding a gRNA.ACTIVEUS 211418665 47Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0125] In some embodiments, the gRNA encoded by the nucleic acid comprises a spacer sequence portion and a tracr RNA portion, wherein the nucleic acid sequence of the spacer sequence portion is the same as a target nucleic acid sequence, except that T in the target nucleic acid sequence is U in the spacer sequence portion. In some embodiments, the spacer sequence portion is 16 to 20 nucleotides long. In some embodiments, the gRNA encoded by the nucleic acid is an sgRNA. In some embodiments, immediately 3’ to the target nucleic acid sequence on the DNA of interest is a PAM sequence. In some embodiments, the gRNA is bound by the RNA-guided nuclease. In some embodiments, the sequence of the spacer portion of the gRNA is designed to target the epigenetic effector fusion to a gene or genetic location of interest for effectuating its epigenetic effector function. In some embodiments, the sequence of the spacer portion of the gRNA is designed to target the epigenetic effector fusion to the transcriptional start site (TSS) of a gene for effectuating its epigenetic effector function. In some embodiments, an enhancer or a repressor is designed to target the epigenetic effector fusion to an enhancer of a gene for effectuating its epigenetic effector function.

[0126] In some embodiments, the RNA-guided nuclease is a Cas protein. Nuclease-null Cas variants that have no substantial nuclease activity are useful to localize proteins and RNA to nearly any set of dsDNA sequences. In some embodiments, the RNA-guide nuclease is a Cas protein with nuclease and / or nickase activity. Hsiung, CC-S., et al. "Engineered CRISPR-Casl2a for higher-order combinatorial chromatin perturbations." Nature Biotechnology (2024): 1-15. In some embodiments, the RNA-guide nuclease is Cas variants with nuclease and / or nickase activity.

[0127] In some embodiments, the DNA binding protein is a zinc finger (ZF) or TALE. A zinc finger DNA binding protein (or binding domain) is a protein, or a domain within a larger protein, that binds DNA in a sequence-specific manner through one or more zinc fingers, which are regions of amino acid sequence within the binding domain whose structure is stabilized through coordination of a zinc ion. The term zinc finger DNA binding protein is often abbreviated as zinc finger protein or ZFP. A TALE DNA binding domain or TALE is a polypeptide comprising one or more TALE repeat domains / units. The repeat domains are involved in binding of the TALE to its cognate target DNA sequence. A single repeat unit (also referred to as a repeat) is typically 33-35 amino acids in length and exhibits at least some sequence homology with other TALE repeat sequences within a naturally occurring TALE protein. Each TALE repeat unit includes 1 or 2 DNA-binding residues making up theACTIVEUS 211418665 48Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Repeat Variable Diresidue (RVD), typically at positions 12 and / or 13 of the repeat. Zinc finger and TALE binding domains can be engineered to bind to a predetermined nucleotide sequence, for example via engineering (altering one or more amino acids) of the recognition helix region of a naturally occurring zinc finger or TALE protein. Therefore, engineered DNA binding proteins (zinc fingers or TALEs) are proteins that are non-naturally occurring.

[0128] None-limiting examples of the RNA aptamer sequences include portions of MS2, PP7, AN, or Qp. In some embodiments, the RNA aptamer sequences comprise more than one repeat of the aptamer sequences. In some embodiments, the RNA aptamer sequences repeat 2, 5, 10, 15, 20, 25, 30 or more times. For example, the RNA aptamer comprises repeats of SEQ ID NO: 71. In some embodiments, the RNA aptamer binds to its cognate RNA-binding protein.

[0129] Non-limiting examples of RNA-binding proteins include MS2 bacteriophage coat protein (MCP), PP7 bacteriophage coat protein (PCP), lambda bacteriophage antiterminator protein N (AN), and Qbeta coat protein (QPCP). The RNA-binding protein recognizes and binds to an RNA aptamer, such as the specific binding sites of MS2, PP7, boxB, and QP. In some embodiments, the RNA-binding protein comprises SEQ ID NO: 72, 73, 74 or 75.

[0130] In certain aspects, described herein is an epigenetic editing system comprising the system means for mediating gene regulation of a target gene. In certain aspects, described herein is an epigenetic editing system comprising the system means for reducing or inhibiting the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). In certain aspects, described herein is an epigenetic editing system comprising the system means for activating or increasing the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.).

[0131] The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be reduced by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system.

[0132] The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be increased by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system.ACTIVEUS 211418665 49Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0133] The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be decreased for at least 5 days, 10 days, 15 days, 20 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, or 60 days. In some embodiments, the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be decreased for at least 5 days, 10 days, 15 days, 20 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, or 60 days, wherein one or more of the nucleic acids of the system are no longer expressed. In some embodiments, the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be decreased for at least 5 days, 10 days, 15 days, 20 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, or 60 days, wherein the nucleic acids of the system are expressed transiently. In some embodiments, the epigenetic effector fusion of such a system comprises a TET1 fusion as described herein.

[0134] As described herein, an epigenetic effector fusion can be tagged with a RNA- binding protein, such as MCP, PCP, AN, or QPCP by fusing the amino acid sequence encoding the RNA-binding protein to the epigenetic effector fusion. The RNA-binding protein can be fused to the N-terminus or C-terminus of the epigenetic effector fusion. In some embodiments, the fusion between the epigenetic effector fusion and the RNA-binding protein may include a linker. In some embodiments, the epigenetic effector fusion is fused directly to the RNA-binding protein.

[0135] As described herein, an epigenetic effector fusion can be tagged with a RNA- binding protein, such as MCP, PCP, AN, or QPCP, by fusing the amino acid sequence encoding the RNA-binding protein to the epigenetic effector fusion. The RNA-binding protein can be fused to the N-terminus or C-terminus of the epigenetic effector fusion. In some embodiments, the fusion between the epigenetic effector fusion and the RNA-binding protein may include a linker. In some embodiments, the epigenetic effector fusion is fused directly to the RNA-binding protein.

[0136] As described herein, a gRNA can be tagged with an RNA aptamer by including the nucleotide sequence encoding the RNA aptamer within the gRNA molecule.

[0137] As described herein, an epigenetic effector fusion of any of the embodiments herein that represses, reduces, or inhibits the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.)comprises a KRAB-L3MBTL3 SAM domain fusion. As described herein, an epigenetic effector fusionACTIVEUS 211418665 50Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 of any of the embodiments herein that activates or increases the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) comprises a p65-HSFl fusion.Bidirectional Transcriptional Perturbation System

[0138] In certain aspects, described herein is a bidirectional transcriptional perturbation system comprising two or more epigenetic editing systems.

[0139] In some embodiments, the bidirectional transcription perturbation system comprises a first nucleic acid encoding a first epigenetic effector as described herein tagged with a first RNA-binding protein (e,g., MCP, PCP, AN, QPCP), a second nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the first RNA-binding protein (e.g. MS2, PP7, boxB, QP), a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2), a fourth nucleic acid encoding a second epigenetic effector as described herein tagged with a second RNA-binding protein that is different to the first RNA-binding protein (e,g., MCP, PCP, AN, QPCP), and a fifth nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the second RNA-binding protein (e.g. MS2, PP7, boxB, QP). In some embodiments, the first epigenetic effector is a KRAB domain as described herein. In some embodiments, the second epigenetic effector is a HSF1. In some embodiments, the first epigenetic effector is a KRAB domain as described herein and the second epigenetic effector is an epigenetic effector fusion that activates or increases the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). In some embodiments, the first epigenetic effector is a KRAB domain as described herein and the second epigenetic effector is a p65-HSFl fusion. In some embodiments, the first epigenetic effector is a HSF1 domain as described herein and the second epigenetic effector is an epigenetic effector fusion that represses, reduces, or inhibits the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). In some embodiments, the first epigenetic effector is a HSF1 domain as described herein and the second epigenetic effector is a KRAB-L3MBTL3 SAM domain fusion.

[0140] In some embodiments, the bidirectional transcription perturbation system comprises a first nucleic acid encoding a first epigenetic effector fusion as described herein tagged with a first RNA-binding protein (e,g., MCP, PCP, AN, QPCP), a second nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the first RNA-binding protein (e.g. MS2, PP7, boxB, QP), a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2), a fourth nucleic acid encoding a second epigenetic effector fusion asACTIVEUS 211418665 51Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 described herein tagged with a second RNA-binding protein that is different to the first RNA- binding protein (e,g., MCP, PCP, AN, QPCP), and a fifth nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the second RNA-binding protein (e.g. MS2, PP7, boxB, QP).

[0141] In some embodiments, the bidirectional transcription perturbation system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with a first RNA- binding protein (e,g., MCP, PCP, AN, QPCP), wherein the epigenetic effector fusion represses, reduces, or inhibits the expression of a first target gene, a second nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the first RNA-binding protein (e.g. MS2, PP7), and a third nucleic acid encoding an RNA-guided nuclease (e.g., dCas9, dCasl2), a fourth nucleic acid encoding an epigenetic effector fusion tagged with a second RNA-binding protein that is different to the first RNA-binding protein (e,g., MCP, PCP, AN, QBCP), wherein the epigenetic effector fusion activates or increases the expression of a second target gene, and a fifth nucleic acid encoding a gRNA tagged with a RNA aptamer that binds to the second RNA-binding protein (e.g. MS2, PP7).

[0142] In some embodiments, the bidirectional perturbation system comprises a first nucleic acid encoding an epigenetic effector fusion as described herein tagged with an RNA- guided nuclease (e.g., dCas9, dCasl2), a second nucleic acid encoding a gRNA, a third nucleic acid encoding an epigenetic effector fusion as described herein tagged with an RNA- guided nuclease (e.g., dCas9, dCasl2), and a fourth nucleic acid encoding a gRNA.

[0143] In some embodiments, the bidirectional perturbation system comprises a first nucleic acid encoding an epigenetic effector fusion tagged with an RNA-guided nuclease (e.g., dCas9, dCasl2), wherein the epigenetic effector fusion represses, reduces, or inhibits the expression of a first target gene, a second nucleic acid encoding a gRNA, a third nucleic acid encoding an epigenetic effector fusion tagged with an RNA-guided nuclease (e.g., dCas9, dCasl2), wherein the epigenetic effector fusion activates or increases the expression of a second target gene, and a fourth nucleic acid encoding a gRNA.

[0144] In some embodiments, the gRNA encoded by the nucleic acid comprises a spacer sequence portion and a tracr RNA portion, wherein the nucleic acid sequence of the spacer sequence portion is the same as a target nucleic acid sequence, except that T in the target nucleic acid sequence is U in the spacer sequence portion. In some embodiments, the spacer sequence portion is 16 to 20 nucleotides long. In some embodiments, the gRNA encoded by the nucleic acid is an sgRNA. In some embodiments, immediately 3’ to the target nucleicACTIVEUS 211418665 52Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 acid sequence on the DNA of interest is a PAM sequence. In some embodiments, the gRNA is bound by the RNA-guided nuclease. In some embodiments, the sequence of the spacer portion of the gRNA is designed to target the epigenetic effector fusion to a gene or genetic location of interest for effectuating its epigenetic effector function.

[0145] In some embodiments, the RNA-guide nuclease is a Cas protein. Nuclease-null Cas variants that have no substantial nuclease activity are useful to localize proteins and RNA to nearly any set of dsDNA sequences. In some embodiments, the RNA-guide nuclease is a Cas protein with nuclease and / or nickase activity. Hsiung, CC-S., et al. "Engineered CRISPR-Casl2a for higher-order combinatorial chromatin perturbations." Nature Biotechnology (2024): 1-15. In some embodiments, the RNA-guide nuclease is Cas variants with nuclease and / or nickase activity.

[0146] None-limiting examples of the RNA aptamer sequences include portions of MS2, PP7, AN, or Qp. In some embodiments, the RNA aptamer sequences comprise more than one repeat of the aptamer sequences. In some embodiments, the RNA aptamer sequences repeat 2, 5, 10, 15, 20, 25, 30 or more times. For example, the RNA aptamer comprises repeats of SEQ ID NO: 71. In some embodiments, the RNA aptamer binds to its cognate RNA-binding protein.

[0147] Non-limiting examples of RNA-binding proteins include MS2 bacteriophage coat protein (MCP), PP7 bacteriophage coat protein (PCP), lambda bacteriophage antiterminator protein N (AN), and Qbeta coat protein (QPCP). The RNA-binding protein recognizes and binds to an RNA aptamer, such as the specific binding sites of MS2, PP7, boxB, and QP. In some embodiments, the RNA-binding protein comprises SEQ ID NO: 72, 73, 74 or 75.

[0148] In certain aspects, described herein is a bidirectional perturbation system means for mediating gene regulation of two or more target genes. In certain aspects, described herein is a bidirectional perturbation system comprising the system means for reducing or inhibiting the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). In certain aspects, described herein is a bidirectional perturbation system comprising the system means for activating or increasing the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.). In certain aspects, described herein is a bidirectional perturbation system comprising the system means for repressing, reducing or inhibiting and activating or increasing the expression of multiple genes.ACTIVEUS 211418665 53Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0149] The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be reduced by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system.

[0150] The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be increased by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system.

[0151] As described herein, an epigenetic effector fusion can be tagged with a RNA- binding protein, such as MCP, PCP, AN, or QPCP by fusing the amino acid sequence encoding the RNA-binding protein to the epigenetic effector fusion. The RNA-binding protein can be fused to the N-terminus or C-terminus of the epigenetic effector fusion. In some embodiments, the fusion between the epigenetic effector fusion and the RNA-binding protein may include a linker. In some embodiments, the epigenetic effector fusion is fused directly to the RNA-binding protein.

[0152] As described herein, a gRNA can be tagged with a RNA aptamer by including the nucleotide sequence encoding the RNA aptamer within the gRNA molecule.

[0153] As described herein, an epigenetic effector fusion of any of the embodiments herein that represses, reduces, or inhibits the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) comprises a KRAB-L3MBTL3 fusion. As described herein, an epigenetic effector fusion of any of the embodiments herein that activates or increases the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) comprises a p65-HSFl fusion.Vectors and Cell Lines

[0154] Several aspects of the invention relate to vector systems comprising one or more vectors, or vectors as such comprising nucleic acid sequences encoding the epigenetic effector fusions described herein and / or encoding guide polynucleotides described herein. Vectors can be designed for expression of transcripts (e.g. nucleic acid transcripts, proteins, or enzymes) in prokaryotic or eukaryotic cells. For example, transcripts can be expressed in bacterial cells such as Escherichia coh, insect cells (using baculovirus expression vectors), yeast cells, or mammalian cells as is known in the art. Alternatively, the recombinantACTIVEUS 211418665 54Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 expression vector can be transcribed and translated in vitro, for example using T7 promoter regulatory sequences and T7 polymerase.

[0155] Vectors may be introduced and propagated in a prokaryote. In some embodiments, a prokaryote is used to amplify copies of a vector to be introduced into a eukaryotic cell or as an intermediate vector in the production of a vector to be introduced into a eukaryotic cell (e.g., amplifying a plasmid as part of a viral vector packaging system). In some embodiments, a prokaryote is used to amplify copies of a vector and express one or more nucleic acids, such as to provide a source of nucleic acid constructs or one or more proteins for delivery to a host cell or host organism.

[0156] As used herein, minicircle refers to small circular plasmids or DNA vectors that are episomal and are produced as a circular expression cassette devoid of any bacterial plasmid backbone. They can be generated from a parental bacterial plasmid that contains a heterologous nucleic acid and two recombinase target sites by intramolecular (cis-) recombination using a site-specific recombinase, such as PhiC31 integrase. Recombination between the two sites generates a minicircle and a leftover miniplasmid. The minicircle can be recovered via separation from the miniplasmid.

[0157] In some embodiments, a vector is a yeast expression vector. Examples of vectors for expression in yeast Saccharomyces cerivisae include pYepSecl (Baldari, et al., 1987. EMBO J. 6: 229-234), pMFa (Kuijan and Herskowitz, 1982. Cell 30: 933-943), pJRY88 (Schultz et al., 1987. Gene 54: 113-123), pYES2 (Invitrogen Corporation, San Diego, Calif.), and picZ (InVitrogen Corp, San Diego, Calif.).

[0158] In some embodiments, a vector drives protein expression in insect cells using baculovirus expression vectors. Baculovirus vectors available for expression of proteins in cultured insect cells (e.g., SF9 cells) include the pAc series (Smith, et al., 1983. Mol. Cell. Biol. 3: 2156-2165) and the pVL series (Lucklow and Summers, 1989. Virology 170: 31-39), the contents of each of which are hereby incorporated by reference in their entireties.

[0159] In some embodiments, a vector is capable of driving expression of one or more sequences in mammalian cells (e.g., but not limited to, human embryonic stem cells, HEK cells, hepatocellular carcinoma cells) using a mammalian expression vector. Examples of mammalian expression vectors include pCDM8 (Seed, 1987. Nature 329: 840) and pMT2PC (Kaufman, et al., 1987. EMBO J. 6: 187-195), the contents of each of which are hereby incorporated by reference in their entireties. When used in mammalian cells, the expression vector’s control functions are typically provided by one or more regulatory elements. ForACTIVEUS 211418665 55Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 example, commonly used promoters are derived from polyoma, adenovirus 2, cytomegalovirus, simian virus 40, and others disclosed herein and known in the art. In some embodiments, the promoter is an Efl a promoter. In some embodiments, the promoter is a U6 promoter. For other suitable expression systems for both prokaryotic and eukaryotic cells see, e.g., Chapters 16 and 17 of Sambrook, et al., Molecular Cloning: A Laboratory Manual. 2nded., Cold Spring Harbor Laboratory, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989, the contents of each of which are hereby incorporated by reference in their entireties.

[0160] In some embodiments, a vector is capable of driving expression of one or more sequences in plant cells using a plant cell expression vector.

[0161] In some embodiments, the recombinant mammalian expression vector is capable of directing expression of the nucleic acid preferentially in a particular cell type (e.g., tissuespecific regulatory elements are used to express the nucleic acid). Tissue-specific regulatory elements are known in the art.

[0162] Methods for introducing proteins and nucleic acids described herein into a cell are known in the art. In some embodiments, delivery is via a nucleic acid (e.g., plasmid(s)) transfected into a cell. The transfected nucleic acids (e.g., plasmid(s)) can comprise an expression vector for an epigenetic effector fusion, and an expression vector for guide polynucleotides (e.g., gRNA or sgRNA). In some embodiments, plasmid(s) can further comprise an expression vector for an RNA-guided nuclease. In some embodiments, the cell may already express an RNA-guided nuclease.

[0163] The nucleic acids may be delivered using adeno associated virus (AAV), lentivirus, adenovirus or other viral vector types, or combinations thereof. The nucleic acids can be packaged into virions using appropriate packaging cells lines as known in the art. In some embodiments, the epigenetic effector fusion protein and one or more exogenous nucleic acids are delivered to a cell using a lentivirus particle.

[0164] In some cases, expression of the epigenetic effector fusions described herein and / or the guide polynucleotides are under the control of an inducible promoter or repressor element. The inducible promoter or repressor element can be inserted into the promoter region of a nucleic acid sequence encoding the epigenetic effector fusions described herein and / or the guide polynucleotides to provide temporal and / or spatial control of the expression or activity.ACTIVEUS 211418665 56Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0165] Upon delivery of a nucleic acid encoding an epigenetic effector fusion to a cell, the nucleic acid can be transcribed and translated into an epigenetic effector fusion protein. In some embodiments, the nucleic acid encoding an epigenetic effector fusion can be delivered to the cell along with a nucleic acid encoding the gRNA and / or RNA-guided nuclease.Applications

[0166] Described herein are several applications of the epigenetic editing system, including but not limited to use as a research tool, for drug discovery or functional genomics, gene therapy and synthetic biology. For example, the epigenetic editing system can be utilized to study gene function and regulation, identify gene targets, test potential therapeutics, and enhance control over gene expression in engineered biological systems. Additionally, the epigenetic editing system can be used to silence or activate genes for therapeutic applications. Target genes include, but are not limited to, BCL11 A for treating sickle cell disease, CCR5 for treating HIV / AIDS, HTT for treating Huntington’s disease, SOD1 for treating amyotrophic lateral sclerosis, and PD-L1 for treating cancers.

[0167] Described herein are several applications of the epigenetic editing system, including but not limited to research tools, gene therapy, regenerative medicine, cancer research and treatment, cellular reprogramming, drug discovery, and agricultural biotechnology. For example, the epigenetic system can be utilized to study the long-term consequences of gene activation in development and disease and investigate epigenetic memory and inheritance. The epigenetic system can be utilized to develop effective treatments for genetic disorders caused by insufficient gene expression, create therapies with reduced need for repeated administration, induce long-term expression of factors important for cell differentiation or tissue regeneration. Target genes include, but are not limited to, SMN2 for treating spinal muscular atrophy, F9 for treating Hemophilia B, HBB for treating beta-thalassemia, GATA2 for treating HSC haploinsufficiency, SOX9 for cartilage regeneration, HGF for liver regeneration, VEGF for angiogenesis. The epigenetic editing system can be utilized to reactivate silenced tumor suppressor genes, upregulate tumor suppressor genes to prevent cancer, and study the long-term effects of gene reactivation on cancer progression. The epigenetic system can be utilized to improve efficiency of iPSC generation and direct cell fate conversion. The epigenetic system can be utilized to study long-term effects of gene upregulation or downregulation without continuing drug administration. The epigenetic system can be utilized to create crop varieties with enhancedACTIVEUS 211418665 57Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 traits through stable gene activation or repression and develop new strategies for pest resistance or stress tolerance in plants

[0168] Described herein are several applications of the bidirectional transcriptional perturbation system, including but not limited to research tools, drug discovery or functional genomics, gene therapy, cell therapy and synthetic biology. For example, the bidirectional transcriptional perturbation system may be utilized for studying complex gene interactions and regulatory networks, investigating cellular responses to activation and repression of different signaling pathways and creating more sophisticated cellular models for disease research. The bidirectional transcriptional perturbation system may be utilized for developing advanced screening platforms for identifying drug targets and creating cellular assays that mimic complex disease states. The bidirectional transcriptional perturbation system may be utilized for designing and implementing complex genetic circuits, or engineering cells with finely tuned, multiplexed gene expression profiles for various applications. The bidirectional transcriptional perturbation system may also be used for developing sophisticated gene therapy approaches that can regulate multiple targets. For example, p53 can be upregulated and MDM2 can be downregulated for treating cancer; ADAMI 0 can be upregulated and BACE1 can be downregulated for treating Alzheimer’s disease; GLUT4 can be upregulated and PTP1B can be downregulated to treat diabetes; utrophin can be upregulated and myostatin can be downregulated for treating muscular dystrophy; PINK1 can be upregulated and LRRK2 can be downregulated to treat Parkinson’s disease; and UCP1 can be upregulated and SCD1 can be downregulated for treating obesity. The bidirectional transcriptional perturbation system can also be utilized to develop sophisticated engineered cells, enhancing the efficiency of cell therapy. For example, BCL2 can be upregulated and FAS can be downregulated in CAR-T cells for cancer therapy; HIF1 A can be upregulated and PHD2 can be downregulated in islet cells for diabetes treatment; IL-10 can be upregulated and IL-6 can be downregulated in regulatory T cells for autoimmune disease therapy; PD-L1 can be upregulated and CD47 can be downregulated for in regulatory macrophages for graft-versus- host disease prevention; and LDHA can be upregulated and PDK1 can be downregulated in CAR-NK cells for solid tumor therapy. The bidirectional transcriptional perturbation system can also be used to optimize cellular metabolism for improved bioproduction processes and to create designer cells with precisely controlled gene expression for various industrial applications.ACTIVEUS 211418665 58Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0169] In some embodiments, the transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be reduced by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system. The transcription of a target DNA locus (e.g., transcribable regions such as a gene, transposable element, enhancer RNAs, etc.) may be increased by at least about 5% to about 95%, e.g., about 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, relative to the expression of the target gene before expression of the system.

[0170] In some embodiments, the one or more nucleic acids encoding an epigenetic effector fusion, a guide polynucleotide(s), and an RNA-guided nuclease as described herein are used to produce a non-human transgenic animal or transgenic plant or transgenic organoid. In some embodiments, the transgenic animal is a mammal, such as a mouse, rat, or rabbit. In certain embodiments, the organism or subject is a plant. In certain embodiments, the organism or subject or plant is algae or crops. In certain embodiments, the subject is an organoid. Methods for producing transgenic plants, organoids, and animals are known in the art, and generally begin with a method of cell transfection, such as described herein. Transgenic animals are also provided, as are transgenic plants, especially crops and algae. The transgenic animal or plant may be useful in applications outside of providing a disease model. These may include food or feed production through expression of, for instance, higher protein, carbohydrate, nutrient or vitamins levels than would normally be seen in the wildtype. In this regard, transgenic plants, especially pulses and tubers, and animals, especially mammals such as livestock (cows, sheep, goats and pigs), but also poultry and edible insects, are preferred.

[0171] Transgenic algae or other plants such as rape may be particularly useful in the production of vegetable oils or biofuels such as alcohols (especially methanol and ethanol), for instance. These may be engineered to express or overexpress high levels of oil or alcohols for use in the oil or biofuel industries

[0172] In plants, pathogens are often host-specific. For example, Fusarium oxysporum f. sp. Lycopersici causes tomato wilt but atacks only tomato, and F. oxysporum f. dianthii Puccinia graminis f. sp. Tritici attacks only wheat. Plants have existing and induced defenses to resist most pathogens. Mutations and recombination events across plant generations lead to genetic variability that gives rise to susceptibility, especially as pathogens reproduce with more frequency than plants. In plants there can be non-host resistance, e.g., the host andACTIVEUS 211418665 59Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 pathogen are incompatible. There can also be Horizontal Resistance, e.g., partial resistance against all races of a pathogen, typically controlled by many genes and Vertical Resistance, e.g., complete resistance to some races of a pathogen but not to other races, typically controlled by a few genes. In a gene-for-gene level, plants and pathogens evolve together, and the genetic changes in one balance changes in other. Accordingly, using natural variability, breeders combine most useful genes for yield, quality, uniformity, hardiness, resistance. The sources of resistance genes include native or foreign varieties, heirloom varieties, wild plant relatives, and induced mutations, e.g., treating plant material with mutagenic agents. Using the present invention, plant breeders are provided with a new tool to control the expression level of a target gene. Accordingly, one skilled in the art can analyze the genome of sources of resistance genes, and in varieties having desired characteristics or traits employ the present invention to induce the rise of resistance genes, with more precision than previous methods and hence accelerate and improve plant breeding programs.

[0173] The invention comprehends the use of the nucleic acids, polypeptides, compositions, systems, and methods disclosed herein to establish and utilize transgenic cells / animals / organoids. Disclosed herein is a non-naturally occurring or engineered composition, or one or more polynucleotides encoding components of said composition, or vector or delivery systems comprising one or more polynucleotides encoding components of said composition for use in a modifying a target cell in vivo, ex vivo or in vitro and, may be conducted in a manner alters the cell such that once modified the progeny or cell line of the modified cell retains the altered phenotype. The modified cells and progeny may be part of a multicellular organism such as a plant or animal with ex vivo or in vivo application of the epigenetic effector fusion system to desired cell types.

[0174] The invention may be a therapeutic method of treatment. The therapeutic method of treatment may comprise gene or genome editing, or gene therapy. A method of the invention may be used to create a plant, an animal or cell that may be used to model and / or study genetic or epigenetic conditions of interest, such as through a model of mutations of interest or as a disease model. As used herein, “disease” refers to a disease, disorder, or indication in a subject. For example, a method of the invention may be used to create an animal or cell that comprises a modification in one or more nucleic acid sequences associated with a disease, or a plant, animal or cell in which the expression of one or more nucleic acid sequences associated with a disease are altered. Such a nucleic acid sequence may encode a disease associated protein sequence or may be a disease associated control sequence.ACTIVEUS 211418665 60Attorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Accordingly, it is understood that in embodiments of the invention, a plant, subject, patient, organism, or cell can be a non-human subject, patient, organism or cell. Thus, the invention provides a plant, animal or cell, produced by the present methods, or a progeny thereof. The progeny may be a clone of the produced plant or animal, or may result from sexual reproduction by crossing with other individuals of the same species to introgress further desirable traits into their offspring. The cell may be in vivo or ex vivo in the cases of multicellular organisms, particularly animals or plants. In the instance where the cell is cultured, a cell line may be established if appropriate culturing conditions are met and preferably if the cell is suitably adapted for this purpose (for instance a stem cell). Bacterial cell lines produced by the invention are also envisaged. Hence, cell lines are also envisaged.

[0175] In a preferred embodiment, the gene therapy vehicle according to the invention can also be used in conjunction with another therapeutic reagent. An effective amount of a pharmaceutical composition according to the invention is administered, optionally in combination with another therapeutic treatment or agent, such as an immunosuppressing drug.

[0176] In a further embodiment, the present invention provides an ex vivo method for transfecting the epigenetic effector system described herein in relevant host cells (e.g. stem cells). In one embodiment, suitable cells are isolated from the mammal, eventually differentiated in vitro and incubated with an effective amount of a pharmaceutical composition of the present invention. Thereafter, the treated (transfected) cells are reintroduced into the organism.

[0177] In some methods, the disease model can be used to study the effects of mutations on the animal or cell and development and / or progression of the disease using measures commonly used in the study of the disease. Alternatively, such a disease model is useful for studying the effect of a pharmaceutically active compound on the disease.

[0178] In some methods, the disease model can be used to assess the efficacy of a potential gene therapy strategy. That is, a disease-associated gene or polynucleotide can be modified such that the disease development and / or progression is inhibited or reduced. In particular, the method comprises modifying a disease-associated gene or polynucleotide such that an altered protein is produced and, as a result, the animal or cell has an altered response. Accordingly, in some methods, a genetically modified animal may be compared with an animal predisposed to development of the disease such that the effect of the gene therapy event may be assessed.ACTIVEUS 211418665 61Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0179] In another embodiment, this invention provides a method of developing a biologically active agent that modulates a cell signaling event associated with a disease gene. The method comprises contacting a test compound with a cell comprising one or more vectors that drive expression of the epigenetic effector fusion system of the present invention; and detecting a change in a readout that is indicative of a reduction or an augmentation of a cell signaling event associated with, e.g., a mutation in a disease gene contained in the cell.

[0180] A cell model or animal model can be constructed in combination with the method of the invention for screening a cellular function change. Such a model may be used to study the effects of a genome sequence modified by the epigenetic effector fusion of the invention on a cellular function of interest. For example, a cellular function model may be used to study the effect of a modified genome sequence on intracellular signaling or extracellular signaling. Alternatively, a cellular function model may be used to study the effects of a modified genome sequence on sensory perception. In some such models, one or more genome sequences associated with a signaling biochemical pathway in the model are modified.

[0181] A transgenic cell in which one or more nucleic acids encoding one or more of the components of the present invention are provided or introduced can be operably connected in the cell with a regulatory element comprising a promoter of one or more gene of interest. As used herein, the term “epigenetic effector fusion transgenic cell” refers to a cell, such as a eukaryotic cell, in which an epigenetic effector fusion has been genomically integrated. The nature, type, or origin of the cell are not particularly limiting according to the present invention. Also the way in which the epigenetic effector fusion transgene is introduced in the cell may vary and can be any method as is known in the art. In certain embodiments, the epigenetic effector fusion transgenic cell is obtained by introducing the epigenetic effector fusion transgene in an isolated cell. In certain other embodiments, the epigenetic effector fusion transgenic cell is obtained by isolating cells from an epigenetic effector fusion transgenic organism. By means of example, and without limitation, the epigenetic effector fusion transgenic cell as referred to herein may be derived from an epigenetic effector fusion transgenic eukaryote, such as an epigenetic effector fusion knock-in eukaryote. By means of example, the epigenetic effector fusion protein transgene may be delivered in for instance eukaryotic cell by means of vector (e.g., AAV, adenovirus, lentivirus) and / or particle and / or nanoparticle delivery, as also described herein elsewhere.

[0182] In certain aspects, described herein is a cell comprising a nucleic acid encoding any of the epigenetic editing systems disclosed herein. Such a cell line can be used in aACTIVEUS 211418665 62Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 method wherein a nucleic acid comprising any of the epigenetic effector fusions disclosed herein is introduced into the cell to generate an engineered cell line. In certain aspects, described herein is a cell comprising a nucleic acid encoding any of the epigenetic effector fusions disclosed herein. Such a cell line can be used in a method wherein a nucleic acid comprising any of the gRNA(s) and / or RNA-guided nucleases disclosed herein is introduced into the cell to generate an engineered cell line.

[0183] In some embodiments, described herein is a kit comprising a cell, the cell comprising a nucleic acid encoding any of the epigenetic editing systems disclosed herein. In some embodiments, the kit further comprises a nucleic acid vector (e.g. plasmid) comprising the epigenetic effector fusions disclosed herein. In some embodiments, described herein is a kit comprising a cell, the cell comprising a nucleic acid encoding any of the epigenetic effector fusions disclosed herein. In some embodiments, the kit further comprises a nucleic acid vector (e.g. plasmid) comprising any of the gRNA(s) and / or RNA-guided nucleases disclosed herein. In some embodiments the cell is a human cell. In some embodiments, the cell is a human embryonic stem cell. In some embodiments, the cells is a Hl human embryonic stem cell. In some embodiments, the cell is a human cancer cell. In some embodiment, the cell is a human cancer cell line. In some embodiments, the cell is a human liver cancer cell line. In some embodiments, the cell is a hepatocellular carcinoma cell line. In some embodiments, the cell line is HepG2 hepatocellular carcinoma cell line. In some embodiments, the cell is a HEK cell.

[0184] In some embodiments, the condition may be sickle cell disease. In some embodiments, the condition may be HIV / AIDS. In some embodiments, the condition may be Huntington Disease. In some embodiments, the condition may be Amyotrophic Lateral Sclerosis. In some embodiments, the condition may be cancer. In some embodiments, the condition may be Alzheimer’s Disease. In some embodiments, the condition may be diabetes. In some embodiments, the condition may be muscular dystrophy. In some embodiments, the condition may be Parkinson’s Disease. In some embodiments, the condition may be cystic fibrosis. In some embodiments, the condition may be obesity. In some embodiments, the condition may be spinal muscular atrophy. In some embodiments, the condition may be Hemophilia. In some embodiments, the condition may be HSC haploinsufficiency.

[0185] An example of Sickle cell disease-related protein includes BCL11 A.

[0186] An example of HIV / AIDS-related protein includes CCR5.

[0187] An example of Huntington’s Disease-related protein includes HTT.ACTIVEUS 211418665 63Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0188] An example of amyotrophic lateral sclerosis-related protein includes SOD1.

[0189] Examples of cancer-related proteins includes, but is not limited to PD-L1, p53, MDM2,

[0190] Examples of Alzheimer's Disease-related proteins includes, but is not limited to ADAM 10, BACE1.

[0191] Examples of diabetes-related proteins includes, but is not limited to GLUT4, PTP1B

[0192] Examples of muscular dystrophy-related proteins includes, but is not limited to utrophin, myostatin

[0193] Examples of Parkinson’s Disease-related proteins includes, but is not limited to PINK1, LRRK2

[0194] Examples of cystic fibrosis-related proteins includes, but is not limited to CFTR, ENaC

[0195] Examples of diabetes-related proteins includes, but is not limited to UCP1, SCD1

[0196] An example of spinal muscular atrophy-related protein includes SMN2.

[0197] An example of hemophilia-related protein includes F9.

[0198] An example of beta-thalassemia-related protein includes HBB.

[0199] An example of HSC haploinsufficiency-related protein includes GATA2.

[0200] A composition of the present invention can be provided in unit dosage form wherein each dosage unit, e.g., an injection, contains a predetermined amount of the composition, alone or in appropriate combination with other active agents. The term unit dosage form as used herein refers to physically discrete units suitable as unitary dosages for human and animal subjects, each unit containing a predetermined quantity of the composition of the present invention, alone or in combination with other active agents, calculated in an amount sufficient to produce the desired effect, in association with a pharmaceutically acceptable diluent, carrier, or vehicle, where appropriate. The specifications for the novel unit dosage forms of the present invention depend on the particular pharmacodynamics associated with the pharmaceutical composition in the particular subject.EXAMPLES

[0201] Examples are provided below to facilitate a more complete understanding of the invention. The following examples serve to illustrate the exemplary modes of making andACTIVEUS 211418665 64Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 practicing the invention. However, the scope of the invention is not to be construed as limited to specific embodiments disclosed in these Examples, which are illustrative only. EXAMPLE 1: A combinatorial domain screening platform reveals epigenetic effector interactions for transcriptional perturbation

[0202] Epigenetic regulation involves the coordinated interplay of numerous multicomponent protein complexes that can read, recruit, write, or erase diverse combinations of epigenetic marks. Recent advances in gene synthesis and functional screening have enabled the testing of thousands of sequences in a scalable fashion, yet current approaches have primarily explored short fragments of individual protein domains. Herein disclosed is COMBINE (combinatorial interaction exploration), a high-throughput screening platform that tests over 50,000 pairs of epigenetic effector domains ranging from 176-2094 amino acids in length for their ability to modulate transcription of an endogenous human gene. COMBINE nominated diverse synergistic and antagonistic domain interactions, including a potent KRAB-L3MBTL3 fusion that enhanced gene silencing up to 34-fold and enabled robust bidirectional CRISPR perturbation. Inducible screening revealed that DNA methylation modifiers were essential for establishing epigenetic memory, with distinct combinations eliciting long-term silencing, repression, or activation. Notably, novel TET1- based combinations were identified that induce hit-and-run upregulation of a basally expressed gene over 50 days, demonstrating long-term transcriptional activation. This systematic analysis of pairwise domain interactions provides a rich resource for understanding epigenetic crosstalk and developing next-generation epigenome editing tools. More broadly, COMBINE offers a generalizable platform to functionally characterize combinatorial biological processes at scale.

[0203] Epigenetic regulation in mammalian cells is a highly coordinated process, essential for organizing genetic information and orchestrating pivotal cellular functions (Goldberg, A.D., et al, 2007) like gene regulation, differentiation, and DNA replication (Vogelaur, M., et al, 2002) or repair (Hauer, M.H, et al, 2017). Central to this regulation are large, multi-protein complexes that control epigenetic states through intricate combinations of post-translational modifications of DNA-scaffolding histone proteins and epigenetic modifications of the genome itself (Zhang, T., et al, 2015). These complexes often contain numerous subunits with distinct functionalities (Policarpi, C., et al, 2024; Lukauskas, S., et al, 2024), such as the Polycomb repressive complexes (Kuzimichev, A., et al, 2002), thatACTIVEUS 211418665 65Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 coordinate a complex code of different histone modifications to enable context-dependent control of gene expression and chromatin state (Kuzimichev, A., et al, 2002).

[0204] Inspired by these natural mechanisms, the epigenome editing toolbox leverages similar combinatorial concepts that bring together programmable DNA-binding domains with epigenetic effector proteins (Thakore, P.I., et al, 2016; Nakamura, M. et al, 2021). For maximal activity, transcriptional activators such as the synergistic activation mediator (SAM) complex fuse a viral activation domain to dCas9 in addition to aptamer-based recruitment of two other activators (Konermann, S., et al, 2015). Epigenetic silencers such as CRISPRoff or EvoETR have also benefited from combining DNA methyltransferase domains with Kriippelassociated box (KRAB) repressors, enabling long-term gene silencing with hit-and- run delivery of the editor system (Amabile, A., et al, 2016; Nunez, J.K, et al, 2021; Cappelluti, M.A., et al, 2024).

[0205] To date, the development of epigenetic editors has largely relied on rational design or low throughput, guess-and-test approaches (Konermann, S., et al, 2015; Amabile, A., et al, 2016; Nunez, J.K., et al, 2021; Cappelluti, M.A., et al, 2024; Konermann, S., et al, 2013; Chavez, A., et al, 2015; Bintu, L. et al, Stepper, P., et al, 2017; Yeo, N.C., et al, 2018; O’Geen, H., et al., 2019; Alerasool, N., et al., 2020; Van, M.V, et al, 2021; Qian, J., et al, 2023, Neumann, et al, 2024). Recent advances in pooled gene synthesis have enabled the functional screening of short peptide libraries of ~80 amino acids tiling larger effector protein candidates (Tycko, J., et al., 2020; Alerasool, N., et al., 2022; DelRosso, N., et al., 2023), further allowing the annotation of novel bivalent interactions that are synergistic or antagonistic (Mukund, A.X., et al, 2023). However, the synthesis length restriction of current-generation oligonucleotide pools may not accommodate the size of functionally active elements, and the use of engineered reporter loci may not extrapolate to endogenous genetic contexts. One recent study used barcoded multi ci stronic adaptors to clone and screen pairs of transcription factors up to 5.8 kilobases in length using a T cell knock-in system, but the generalizability and scale of this approach is limited by the requirement of a cell-type specific functionally monoallelic target locus and a sharp length-dependent integration bias (Blaeschke, F., et al 2023).

[0206] To address these challenges, a generalizable high-throughput screening platform was developed to functionally characterize combinations of long and heterogenous epigenetic effectors, with the goal of 1) identifying new effector combinations that increase the potencyACTIVEUS 211418665 66Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 or durability of transcriptional perturbations, and 2) uncovering previously unknown interactions between individual effector domains.

[0207] Herein COMBINE (combinatorial interaction exploration) is reported, a high- throughput and inducible screening platform that can accommodate heterogeneously-sized fusion elements, including large catalytic domains up to 6.3 kilobases in length. COMBINE is applied to measure the transcriptional perturbation landscape of an endogenous human gene with over 50,000 bivalent epigenetic effectors, composed of pairs of protein domains from diverse classes including readers, recruiters, structural factors, and catalytic writers or erasers. Systematic analysis and validation of combinatorial interactions enabled the discovery of previously unknown synergistic or antagonistic interdomain interactions that control epigenetic regulation of transcription in human cells. The screening data allowed us to nominate new epigenetic effector domain combinations for synthetic biology and genetic medicine, including a potent KRAB-L3MBTL3 fusion that enhanced gene silencing up to 34- fold and enabled a robust bidirectional CRISPR perturbation system to simultaneously activate one target gene and silence another. To examine durable long-term perturbations in our screen, novel combinations containing CpG DNA methylation-related domains that drive heritable silencing, repression, or activation of an endogenous target gene were identified. COMBINE should be generalizable to a wide variety of lentivirus-transducible cell types and protein domain classes, as well as DNA and RNA elements, providing a broad platform to interrogate biological interactions at scale.EXAMPLE 2: MethodsCell lines and cell culture

[0208] All experiments in this study were carried out in K562 cells (ATCC, CCL-243, female). Cells were cultured in a humidified incubator at 37°C and 5% CO2, in RPMI 1640 (Gibco,) media supplemented with 10% fetal bovine serum (FBS) (Gibco, 26140095) and 1% penicillin-streptomycin (Gibco, 15140122). For Library 1 lentivirus generation, 293FT (ThermoFisher Scientific) or Lenti-X 293T (Takara) were cultured in a humidified incubator at 37°C and% CO2, in DMEM (Gibco, 10566016) media supplemented with 10% fetal bovine serum (FBS)(Gibco, 26140095) and 1% penicillin-streptomycin (Gibco, 15140122). For Library 2 lentivirus generation, viral production cells (Gibco, A35347) were cultured according to the manufacturer’s instructions.Guide design and cloningACTIVEUS 211418665 67Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0209] CRISPRa guides were generated using CHOPCHOP for the CD81, CD55, CD58, CD151, and CD155 genes, and three were selected to tile each promoter region from -500-0 to the TSS (Guides) (Labun, K., et al, 2019). A tiered golden gate approach was developed to assemble a 3X guide array and to easily insert this array into a landing pad PiggyBac vector, which was cloned in part using the EMMA toolkit (Martella, A., et al, 2017), between divergent expression cassettes pTRE3G-dCas9-tagBFP and pEFla-BlastR-P2A-rTA3 (Plasmids, pMH224). Briefly, the three guides were cloned separately via BsmBI golden gate into three distinct guide expression plasmids containing different golden gate overhangs around the guide cassette (Plasmids, pMH213-pMH215). The guide scaffold contained MS2 aptamers in stem loop two and the tetraloop and was sourced from recent optimization (Heidersbach, A. J., et al, 2023). These individual guide plasmids were then assembled via Bsal golden gate into a RFP dropout landing pad containing pSV40-mCherry for mammalian expression (Plasmids, pMH222). Guides were arranged in the 3X array such that the closest guide to the TSS was placed first, followed by the middle guide and then the furthest guide. The CD81 3X guide array plasmid was used directly for guide nucleofection in screening of Library 1 (Plasmids, pMH228). This 3X array was further assembled into pMH224 via a final BsmBI golden gate for PiggyBac insertion and constitutive guide expression, which was used for screening Library 2 (Plasmids, pMH260). Additional sgRNAs targeting before the TSS of CD55, CD58, CD151, and CD274 were generated using CHOPCHOP (Guides) (Labun et al, 2019). CD55, CD58, and CD151 guides were inserted into a lentiviral vector containing the recently optimized MS2 guide scaffold (Heidersbach, A. J., et al, 2023) driven by human U6 promoter and SV40 promoter driven-mCherry marker. For bidirectional perturbation experiments, the CD274 guide (Guides) was cloned into the human U6-driven PP7 guide scaffold within a lentiviral vector, where both MS2 sgRNA and PP7 sgRNA are expressed from distinct mouse and human U6 promoters, respectively.PB Tet-On dCas9 cell line generation:

[0210] To generate Tet-On dCas9 K562 cell lines expressing pTRE3G-dCas9-tagBFP, pEFla-BlastR-P2A-rTA3 with or without the 3X CD81 guide array, K562 cells were nucleofected with 600 ng of pMH224 or pMH260 and 200 ng of a separate PiggyBac transposase plasmid using SF Cell Line 96-well Nucleofector kit and protocol (Lonza, V4SC- 2096) (Plasmids). Two days following nucleofection, a 10-day selection began 1201 with 10 pg / mL blasticidin (Gibco, Al 113903). After selection, cells were expanded and assessed for BFP induction in response to 20 pg / mL doxycycline. This Tet-On dCas9 K562 (with CD81ACTIVEUS 211418665 68Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 guide array) cell line was used for Library 2 HTS and Library 1 validation experiments, while this Tet-On dCas9 K562 (without CD81 guide array) was used for validation of Library 2 effectors on additional genes. An additional Tet-On dCas9 K562 (without CD81 guide array) cell line was generated that expresses pTREtight-dCas9 and pEFla-BlastR-P2A-rTA3. K562 cells were co-transfected with Piggybac plasmid encoding both pTREtight-dCas9 and pEFla- BlastR-P2A-rTA3 (without transposases) and PiggyBac transposase plasmid using Lipofectamine 2000 (Invitrogen). Two days following transfection, 20 pg / mL blasticidin (Gibco, Al 113903) selection was performed for at least 10 days. This Tet-On dCas9 K562 (without CD81 guide array) cell line was used for Library 1 HTS.Library 1 design and cloning:

[0211] A literature review was conducted to cover critical components of the key repressive epigenetic multiprotein complexes and structural factors for chromatin organization. The majority of library members were selected based on previous arrayed studies employing truncation-based co-immunoprecipitation experiments or luciferase assays, except for the few members from high-throughput studies (Tycko, J., et al, 2020). Peptides shorter than 80 amino acids were expanded equally on either side to reach 80 amino acids in total. Start codons were removed as effectors were to be fused to the C-terminus of aptamer binding protein. Amino acid sequences were back-translated with mammalian species codon optimization and forbidding 3' end 6-mer of both amplification primers motifs (GGTGTG and ATGGCC) by Geneious Prime. Second codon optimization was performed for human codon usage, removing several type IIS restriction enzyme sites (BbsI, Bsal, BsmBI, BspQI, BtgZI, and SapI) and constraining GC content to between 35% and 65% in every 50 nucleotides window by DNA chisel (Zulkower, V., et al, 2020). Forward amplification primer binding handle (TCCaGACCGTTCaGGTGTG (SEQ ID NO: 83) adapted from previous study (Tycko et al, 2020) ), Bsal binding site (GGTCTCT), and overhang (GTCA) facilitating ‘Glutamic-acid (E) - Serine (S)’peptide overhang for C-terminus of XTEN16 linker were appended to 5' of the gene fragment (Figure 3 A). Overhang (AGTG) for ‘Serine (S) - Glycine (G)’ peptide overhang at N-terminus of XTEN16 linker, Bsal binding site (CGAGACC), and reverse amplification primer binding handle (gGCCaTGCGGaATGGGTTA (SEQ ID NO: 84)) were appended to 3' of the gene fragment. Optimized golden gate assembly overhangs were selected based on a previous study (Potapov, V., et al 2018). A few BsmBI sites, Bsal sites, and 3' end 6-mer of both amplification primers motifs (GGTGTG and ATGGCC) that were unintentionally generatedACTIVEUS 211418665 69Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 by former procedures were manually modified by silent mutations. Finally, library members that did not pass the Gblock or Eblock complexity test were human codon-optimized by IDT webtool, and again BsmBI, Bsal sites, and 3' end 6-mer of both amplification primers motifs were manually modified by silent mutations. Double-stranded gene fragments were synthesized (IDT and Twist Biosciences) and then pooled manually without PCR amplifications to make a 20 nM solution with the molarity of each fragment increasing proportional to its length with a slope of 1. 1.5 fold of the 450 bp gene fragment was added compared to the 300 bp gene fragment. The pooled library was subcloned intoKanRl receiver (pCMOOl) and KanR2_reveiver (pCM002) plasmids by two separate golden gate reactions. For a total volume of 40 pL golden gate reaction, 185ng of KanRl (pCMOOl) or KanR2 (pCM002) receiver plasmid, 1245 10 pL of 20 nM pooled library for~2: 1 molar ratio between backbone:insert, 2 pL of BsaI-HFv2 (NEB, 20000 U / mL), 2 pL of T4DNA ligase (NEB, 400000 U / mL), and 4 pL of 10X T4 DNA ligase buffer (NEB) was used. The thermocycling protocol was 30 cycles of 5 minutes digestion at 37°C and 5 minutes ligation at 16°C, followed by a final 5 minute digestion at 37°C, 5 minute heat inactivation at 60°C and an extra 20 minute heat inactivation at 70°C. The reactions were then purified with DNA Clean & Concentrator-5 (Zymo) and were eluted by 8 pL of nuclease-free water (Ambion). 1 pL of purified DNA was transformed into 25 pL of Endura electrocompetent cells (Lucigen) following the manufacturer’s instructions with 2 mL of prewarmed recovery media. 2 mL of cells were plated onto a 245 mm x 245 mm plate, and serially diluted cells were plated onto 100 mm diameter plates for the estimation of colony numbers. After overnight incubation at 30°C, colonies were collected, and plasmid libraries were extracted with Maxiprep (Machery- Nagel). Colony coverage was at least 34000X for 155 library members. For the final bivalent library construction, the LV AmpR b ackbone (pCM003) was predigested with FastDigest Esp3I (Thermo Fisher Scientific) and gel extracted. For a total volume of 100 pL golden gate reaction, 800 ng of predigested and gel-extracted LV AmpR b ackbone (pCM003), 1000 ng of N-terminus effector (KanRl recevier) library, lOOOng of C-terminus effector (KanR2_receiver) library, 10 pL of FastDigest Esp31 (Thermo Fisher Scientific), 5 pL of T4 DNA ligase (NEB, 400000 U / mL), and 10 pL of 10X T4 DNA ligase buffer(NEB) were used. The thermocycling protocol was initial 5 minute digestion at 37°C, 30 cycles of 5 minute digestion at 37°C and 10 minute ligation at 16°C, final 20 minute ligation at 16°C, final 5 minute digestion at 37°C, and 20 minute heat-inactivation at 75°C. The golden gate reaction was then isopropanol precipitated into 10 pL of nuclease-free water (Ambion)ACTIVEUS 211418665 70Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025(Joung, J., et al, 2017). Again, 1 pL of purified DNA was transformed into 25 pL of Endura electrocompetent cells (Lucigen) following the manufacturer’s instructions with 2 mL of prewarmed recovery media. 2 mL of cells were plated onto a 245 mm x 245 mm plate and serially diluted cells were plated onto 100 mm diameter plates for the estimation of colony numbers. After overnight (15 hours) incubation at 30°C, colonies were collected, and plasmid pools were extracted with Maxiprep (Machery-Nagel). Colony coverage for 1 bioassay plate was at least 1100X for 1552 library members. Two separate maxi-prepped DNA solutions from two bioassay plates were mixed to achieve at least 2200X colony coverage of library members.High-throughput assay to measure transcriptional effects of Library 1:

[0212] Low passage 293FT cells (Invitrogen) or Lenti-X (Takara) 293T cells were grown inDMEM (Invitrogen, 10566-016) supplemented with 10% FBS (Gibco) and 1% Penicillin- Streptomycin (Gibco). For one T225 flask, 39.9 pL of 1 mg / mL PEI MAX (Polysciences) and ImL of DMEM were mixed by inversion and incubated at room temperature for 10 min. During incubation, 5.6 pg of pMD2.G, 11.3 pg of psPAX2, and 22.7 pg of target plasmid or plasmid library were added to new 1 mL of DMEM and were thoroughly mixed by vortexing. After incubation, plasmids + DMEM mixture was added to PEI MAX + DMEM mixture followed by 30sec vortexing. The final ~2 mL mixture of plasmids + PEI MAX + DMEM was incubated at room temperature for 30 min. During incubation, 293FT or Lenti-X 293 T cells were split by TrypLE(Gibco) and 33.4 million cells were added to a standing T225 flask in 5 mL of fresh growth media. After the 30 min incubation, plasmids + PEI MAX + DMEM mixture was added to the flask, and the flask was gently mixed by swirling. The flask was 1289 incubated for 5 min in the upright position. Following the 5 min incubation, fresh growth media was added up to 45 mL, and cells were incubated at 37°C, 5% CO2 overnight. The next day (Day 1), the media was changed to 45 mL of fresh growth media and incubated for an additional 2 days, after which the supernatant was collected and concentrated by Lenti- X concentrator (Takara) following the manufacturer’s protocol. Upon infection of 40 million PB Tet-On dCas9 cells (without CD81 guide array) with 200pL of lentivirus per replicate on Day -13, cells were cultured in growth media containing l-2pg / mL doxycycline (Figure 6). In parallel, 2 control cell lines (negative control: PB Tet-On dCas9 cells without any effector & positive control: PB Tet-On dCas9 cells with LV-pTRE3G-PuroR-T2A-stdMCP-eGFP- KRAB (ZNF10) preinstalled by lentiviral infection) started doxycycline incubation as well. Doxycycline was replaced or added daily, maintaining the concentration of 0.5-2 pg / mL toACTIVEUS 211418665 71Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 compensate for the short half of doxycycline in the media. 1 day post-infection (Day -12), puromycin selection started with a 1 pg / mL concentration. To simultaneously measure the lentiviral titer, subsets of cells were passaged to 96 well plates for the viability assay. 3 days after the start of puromycin selection (Day -9), lentiviral titer was measured by the CellTiter- Glo 2.0 Cell Viability Assay (Promega G9242) as previously described (Joung, J., et al, 2017). Measured viability was 0.25-0.29 suggesting successful low MOI infection. Puromycin selection was continued for an additional 3 days until Day -6. There was an increase in the percentage of live cells by trypan blue staining (Thermo Fisher Scientific) during the additional 3 days of selection, suggesting the outgrowth of infected cells over debris from dead cells. After a total of 6 days of puromycin selection and outgrowth (Day -6), 39.2 million surviving cells were mixed with 0.8 million positive control cells (LV-pTRE3G- PuroR-T2A1313stdMCP-eGFP-KRAB (ZNF10)) per replicate. On Day -6, mixed cells were nucleofected with the CD81 targeting 3X guide array plasmid (pMH228). For each cuvette, 4 million cells were nucleofected with 4.5 pg of pMH228 following the manufacturer’s K562- specific protocol (4D X1316unit, FF-120 program, Lonza). For each replicate, 40 million cells were nucleofected by 10 cuvettes. 4 million positive control cells were also nucleofected by 1 cuvette. 3 days post-nucleofection (Day -3), mCherry-positive cells were sorted. To reduce the variability generated by sgRNA expression level, a population expressing moderate levels of mCherry was sorted (Figure 6B). In total, ~9.7 million cells (-400X coverage) were sorted for each replicate. 3 days post-sorting (Day 0), at least 11.5 million cells per replicate were frozen for future FACS sorting. The remaining half of at least 11.5 million cells were centrifuged and passaged to media without doxycycline to stop the expression of dCas9 and the stdMCP-combinatorial effectors. Cells were continuously expanded until Day 6 in culture media without doxycycline. On Day 6, 50 million cells were frozen for each replicate. Cells were again continuously expanded until Day 12. On Day 12, 120 million cells were frozen for each replicate. On the day of FACS, frozen cells from various timepoints (Day 0, 6, and 12) were thawed briefly and washed twice with eBioscience Flow Cytometry Staining Buffer (Invitrogen). APC Mouse Anti -Human CD81 antibody (Cat# 561958, BD Bioscience) was diluted 1 : 10 in lOOpL stain buffer per 0.1 million cells. Staining was performed on a 3D rotary mixer in a cold room(4°C) for 1 hour. Cells were washed twice with stain buffer and sorted using a FACS Aria Fusion Special Order Research Product (BD Bioscience). mCherry-negative singlet populations were separated into two populations based on CD81 level (Figure 6C). Sorted cells were frozen at -ACTIVEUS 211418665 72Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 20254°C until genomic DNA extraction. Genomic DNA (gDNA) was extracted using columns from the Quick-DNA Miniprep Plus kit (Zymo Research) and reagents from the Quick-DNA Midiprep Plus kit (Zymo Research) with the following modifications in the protocol. Cells were split into multiple columns if the number of cells was larger than 3 million. Per 3 million cells, excess genomic lysis buffer (Beta-Mercaptoethanol added) with a volume of 1.6 mL from the Quick-DNA Midiprep Plus kit (ZymoResearch) was used instead of BioFluid & Cell Buffer from the Quick-DNA Miniprep Plus kit (Zymo Research). After the addition of 1.6 mL genomic lysis buffer to the cells, the mixture was vortexed thoroughly. Cells were lysed in genomic lysis buffer for 1 hour at room temperature with 3D rotation. Excess genomic lysis buffer ensured complete lysis of the cells, yielding a high amount of extracted gDNA. Except for the described cell lysis step, the manufacturer’s protocol was followed for the Quick-DNA Miniprep Plus kit (Zymo Research). Enriched combinatorial library members were PCR amplified from the gDNA using the following conditions. For one 50 pL of the PCR reaction, 25 pL of 2X NEBNext HiFi (NEB), 1 pg of gDNA, and a final concentration of 0.5 pM for each forward and reverse NGS primers were prepared and mixed on ice. One-step NGS primers were designed to have 0-7 random spacer nucleotides to increase the base diversity during the initial sequencing cycles (Naik, T., et al 2023). The thermocycling protocol was 1 cycle of 98°C for 3 min, 27 cycles of 98°C for 10 seconds, 64°C for 30 seconds, 72°C for 75 seconds, and final 1 cycle of 75°C for 2 minutes. After thermocycling, PCR reactions were pooled and run through multiple lanes in 1.2% TBE gel. Rectangular regions of the gel spanning from 800 bp ~ 2 kb were excised, and PCR products were extracted using Monarch DNA Gel Extraction Kit (NEB). The gel-extracted library was quantified by Nanodrop (Thermo Fisher Scientific), Qubit HS kit (Thermo Fisher Scientific), Bioanalyzer (Agilent Technologies), and KAPA Library Quantification Kit (Roche).Approximate agreement in quantification between Qubit and Bioanalyzer was observed while KAPA library quantification exhibited the most discrepancy, potentially due to differences in the size between the library and the standards provided by the kit. Based on the results of Qubit and Bioanalyzer, the library was pooled with -10% PhiX control (Illumina) and sequenced on a NextSeq2000 (Illumina) with 2 x 300 bp P2 kits (Illumina). In total, two 2 x 300 bp P2 kits were used. Libraries from Day 0 and 6 were pooled together and sequenced with one kit, and the Day 12 library was sequenced separately using the second kit.NGS analysis of bivalent domains from Library 1:ACTIVEUS 211418665 73Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0213] To address the issue of lower R2 sequencing quality, which is occasionally observed during paired-end readout of long fragments (Tan, G., et al, 2019), an approach of spreading ten 15-mer short identifiers (IDs) was implemented across the entire 301 bp sequencing region. If any of the 10 IDs exactly matched the NGS read, the read was assigned to the corresponding domain. Unique 15mer IDs for identifying N- and C-terminal domains were generated by the following steps. Step 1) From fasta files composed of expected 301 bp NGS reads for every domain (Rl_Forward_P5_301cycles. fasta & R2_Reverse_P7_301 cycles. fasta), all possible unique 15mers were generated using UniqueKMER with the following commands (Chen et al, 2021). $ uniquekmer -f Rl_Forward_P5_301cycles. fasta -k 15 -o "kmercollection_Rl"$ uniquekmer -f R2_Reverse_P7_301 cycles. fasta -k 15 -o "kmercollection_R2"Step 2). If a 15 mer ID for one domain had a Hamming distance of less than 2 with an ID from another domain, one of the IDs was deleted from the domain with a larger number of IDs by executingLibrary l_step2.py. Output files were numerically sorted using Seqkit with the following commands (Shen et al, 2024). $ seqkit sort -nN Rl_15mer_HD2_UMIs. fasta -o Rl_15mer_HD2_UMIs_sorted.fasta$ seqkit sort -nN R2_15mer_HD2_UMIs. fasta -o R2_15mer_HD2_UMIs_sorted.fastaStep 3). Using pairwise2 (Biopython) local alignment mode with the scoring scheme of (match: 1, mismatch: 0, opening gap: -1, extending gap: -1), IDs that had an alignment score higher than with any other domains’ expected 301 bp NGS reads were removed. Also, 10 IDs were chosen to be spread across the 301 bp region as evenly as possible by executingLibraryl_step3.py.Step 4) For domains that ended up with fewer than 10 IDs, previously removed IDs with alignment scores (pairwise2) of 13 were recalled and assigned to ensure that the total number of IDs reached 10 for every domain by executing Library l_step4.py. Finally, 10 IDs (Rl_15mer_UMIs_10per_effector.fasta and R2_15mer_UMIs_10per_effector.fasta) were generated for each domain. By matching R1 and R2 reads, a count table for bivalent candidates was generated. There was a noticeable trend of count numbers for homo-bivalent candidates being higher than hetero-bivalent candidates which might be caused by recombination artifacts. Due to the difficulty in discerning whether the high count number arises from transcriptional perturbation of CD81 or from recombination artifacts, homo-bivalent candidates were not analyzed. If the sum of NGS reads from HIGH and LOW bins was less than 40 in at least one replicate, the bivalent candidate was considered a dropout. The ratio of a bivalent candidate in a specific bin was calculated by the following formula.ACTIVEUS 211418665 74Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0214] The enrichment score of bivalent candidate “i” was calculated by dividing Ratio;, CD81 High bin by RatOi,CD81 Low bin.Secondary validation of individual and bivalent effectors from Library 1

[0215] To harness previously ordered effector gene fragments, a series of landing pad plasmids for nucleofecti on-based delivery of effectors were generated. For bivalent domain candidates, similar to the library cloning procedure, 3 sequential golden gate reactions were performed. The landing pad plasmid for the 1st golden gate reaction was the sameKanRl receiver (pCMOOl) as the library cloning, whereasKanR2_individual_validation_receiver (pCM023) was used as the landing pad plasmid for the 2ndgolden gate reaction. For the final golden gate reaction, the landing pad was pCAG AmpR backbone (pCM024), which enables the constitutive expression of the effector by the CAG promoter. For monovalent domain plasmid cloning, 2 sequential golden gate reactions were performed. For the 1st golden gate reaction, the landing pad plasmid was KanR monovalent receiver (pCM025). For the 2nd golden gate reaction, the landing pad plasmid was the same as the bivalent domain candidate cloning, pCAG_AmpR_backbone(pCM024). For cloning linear gene fragments into KanRl receiver (pCMOOl ),KanR2_individual_validation_receiver (pCM023), andKanR monovalent receiver (pCM025),pL of 60 ng / pL landing pad plasmids, 2.5 pL of 20 nM gene fragment, 0.5 pL of 10X T41igase buffer (NEB), 0.25 pL BsaI-HFv2 (NEB, 20000 U / mL), 0.25 pL T4 Ligase (NEB, 400000U / mL) and 1.25 pL nuclease-free water (Ambion) was mixed well resulting in a total of 5 pL reaction. The thermocycling condition was 30 cycles of 37°C for 1 minute and 16°C for 1 minute followed by 1 cycle of 37°C for 5 minutes and 75°C for 5 minutes. For the final golden gate reaction, 0.5pL of 100 ng / pL pCAG AmpR backbone (pCM024), 1 pL of 50 ng / pL N-terminus effector plasmid, 1 pL of 50 ng / pL C-terminus effector plasmid, 1 pL of FastDigest Esp3I (Thermo FisherScientific), 0.5 pL of T4 DNA ligase (NEB, 400000 U / mL), 1 pL of 10X T4 buffer (NEB), and 5 pL of nuclease-free water (Ambion) were mixed well resulting in a total of 10 pL reaction.KanR monovalent receiver (pCM025) was used instead of KanRl receiver (pCMOOl) or KanR2_individual_validation_receiver (pCM023) for cloning the monovalent effector plasmid. The thermocycling protocol was the same: 30 cycles of 37°C for 1 minute and 16°C for 1 minute followed by 1 cycle of 37°C for 5 minute and 75°C for 5 minute. NucleofectionACTIVEUS 211418665 75Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 of effector plasmids was performed with SF Cell Line 4D-Nucleofector XKit S following the manufacturer’s protocol (4D X-unit, FF-120 program, Lonza). PB Tet-On dCas9 K562 cell line (with CD81 guide array) was preinduced with 1-2 pg / mL doxycycline starting 1-3 days before the day of nucleofection..

[0216] On the day of measurement, -300 pL of cells were washed twice with eBioscience Flow Cytometry Staining Buffer, and 99 pL staining buffer with 1 pL of APC Mouse Anti-Human CD81 antibody (Cat# 561958, BD Bioscience) was used per well in 96 well plates. Following 1 hour of incubation at 4°C, cells were washed twice, and the CD81 level was measured by Atune NxTFlow Cytometer (Thermo Fisher Scientific). Empty vector pUC19 nucleofected cells were used as control cells for gating on BFP, GFP, mCherry, and CD81 (APC) levels.Library 2 design, cloning, and barcode mapping

[0217] A literature review was conducted to identify minimally active catalytic epigenetic effectors across 11 well-established classes: DNA demethylation machinery (DNDM), DNA methyltransferase (DNMT), E2 and E3 ubiquitin ligases (UBL), histone acetyltransferase (HAT), histone arginine methyltransferase (HRMT), histone deacetylase (HDAC), histone demethylase (HDM), histone deubiquitinase (DUB), histone kinase (HK), histone lysine methyltransferase(HKMT), and histone phosphatase (HP). Human proteins associated with each class were identified using Gene Ontology terms, except for HKMTs, which were identified as human proteins containing SET domains in addition to GO terms. Within each list, each protein was manually reviewed to assess for possible catalytic activity and to search for catalytically active truncations. In total, 195 catalytic domains were identified across all classes, with approximately 75% sourced directly from literature, while the remaining domain bounds were sourced from alignment with closely related proteins or simply estimated from Pfam annotations. Additionally, three non-catalytically active controls were included: RYBP for silencing (Zhao et al, 2020), a 400 aa fragment of DMD determined to be neutral from HT-recruit24, and HSF1 for activation (Konermann et al, 2015). All domains and references are included in Figure 38. Amino acid sequences for each effector were codon optimized for 1463 human expression using DNA Chisel with constraints to avoid restriction enzyme sites (Bsal, BsmBI, BbsI, BtgZI), avoid patterns (5 x 3-mer, 2 x 15-mer, 2 x 12-mer, 8 x homopolymer), and enforce GC content between 35-65% in each 50 bp window (Zulkoer et al, 2020). Each effector was ordered from Twist Biosciences as a clonal gene in pTwist Kan High Copy with flanking sequences containing nested BsmBI and Bsal golden gate sitesACTIVEUS 211418665 76Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 to facilitate cloning bivalent combinations (Figure 4A, and 3' Effector Flank). All golden gate overhangs were sourced from a 15 member high fidelity set, and all landing pads contained CcdB dropout genes to improve cloning efficiency87. Individual effector plasmids were transformed into MACH1 cells (Thermo Fisher Scientific) and miniprepped using QIAprep Spin Miniprep Kit (Qiagen, 27106). Because lentiviral transduction efficiency decreases roughly 3.5X per additional kilobase of DNA29, effector plasmids were pooled exponentially by effector size by manually adding 0.2*3.5A(x / 1000) pmol of plasmid DNA, where x is the length of the effector in base pairs (Figure S3B and S3C). The BsmBI sites in this pool (EF2 pool) were later used to clone into the N-terminal Effector 2 position, and the Bsal sites were used to subclone the pooled effectors into a new high copy Chloramphenicol resistant backbone via golden gate reaction (50 ng pMH249 (Plasmids), 2-fold molar excess EF2 pool, 1 pL BsaI-HFv2 (NEB, R3733L), 1 pL T4 ligase (NEB, R0202L), 2.5 pL T41igase buffer (NEB, B0202S), and H2O to 25 pL incubated at 37°C for 5 min, followed by 30 cycles of 37°C for 5 minutes and 16°C for 10 minutes, followed by final ligation at 16°C for 20 minutes, final digestion at 37°C for 30 minutes, and heat inactivation at 80°C for 20 minutes) (Figure 4D). The reaction was then cleaned using DNA Clean & Concentrator- 5 (Zymo, D4013) and eluted in 6pL H2O, before electroporating 2 pL into 25 pL of Endura electrocompetent cells following the manufacturer’s instructions (Biosearch Technologies, 60242-1). Recovered cells were plated onto a 245 mm x 245 mm plate and incubated overnight at 37°C. The next day, colonies were scraped from the plate and the plasmid pool was maxiprepped (Machery-Nagel, 740424.50). The resulting EFl pool contained BsmBI golden gate sites at the same 4 bases as the Bsal insertion of the effector to allow for subsequent cloning into the C-terminal Effector 1 position (Figure 4E). Before the final bivalent assembly of the EFl and EF2 pool into the lentiviral landing pad, the XTEN16 linker and variable barcodes were prepared. The XTEN16 linker was ordered as two ssDNA oligos with the correct overhangs for assembly between the EFl and EF2position and annealed by equimolar mixing at room temperature (Figure 4F, Misc.). Variable barcodes were ordered from Integrated DNA Technologies (IDT) as a 20xN ssDNA oligo flanked by NGS primer binding sites and BsmBI golden gate sites for assembly after EF2(Figure 4G, Misc.). The oligo was ordered preannealed with the reverse complement of the 3' flanking sequence to facilitate filling in of the bottom strand of the barcode and 5' flanking sequence by polymerase extension (2.5 pL of preannealed oligos at 10 uM, 12.5pL KAPA HiFi HotStart Ready Mix (Roche KK2601), and H2O to 25 pL followed by incubation a 98°C for 3 minutesACTIVEUS 211418665 77Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 and 72°C for 30 minutes). The extended barcode was cleaned with DNA Clean &Concentrator-5 (Zymo, D4013) and eluted in 20 pL H2O, before predigestion (1 pg extendedbarcode, 2 pL Esp3I (Thermo Fisher Scientific, FD0454), 2 pL lOX FastDigest buffer (ThermoFisher Scientific, B64), and H2O to 20 pL incubated at 37°C for 25 min). The digested barcode was cleaned with DNA Clean & Concentrator- 5 (Zymo, D4013) and diluted to 2 ng / pL. Final bivalent assembly was performed via 5-piece golden gate reaction into the lentiviral landing pad, which was cloned in part using the EMMA toolkit85, containing pTRE3G-MCP-XTEN80 andpSV40-PuroR cassettes (250 ng pMH276 (Plasmids), IX molar EFl pool, IX molar annealed XTEN16 linker, IX molar EF2 pool, IX molar digested barcode, 1 pL Esp3I (ThermoScientific, FD0454), 1 pL T4 ligase (NEB, R0202L), 2.5 pL T4 ligase buffer (NEB, B0202S), and H2O to 25 pL incubated at 37°C for 5 minutes, followed by 30 cycles of 37°C for 5 minutes and 16°C for 10 minutes, followed by final ligation at 16°C for 20 minutes, final digestion at 37°C for 30 minutes, and heat inactivation at 80°C for 20 minutes) (Figure 4H). The reaction was then cleaned using DNA Clean& Concentrator-5 (Zymo, D4013) and eluted in 6 pL H2O, before electroporating 2x 2 pL into 25pL of Endura electrocompetent cells each following the manufacturer’s instructions (Biosearch Technologies, 60242-1). Recovered cells were serially diluted and plated onto 100 mm diameter plates for the estimation of colony numbers, while the remaining cells were plated onto several mm x 245 mm plates. After overnight incubation at 30°C, colonies on dilution plates were counted, and enough large plates were scraped and maxiprepped (Machery- Nagel, 740424.50) such that the average coverage was between 100-150X colonies per member. Strict coverage constraints were necessary both to adequately represent shorter effector pairs and to ensure that there were few enough barcodes to map via long-read sequencing. Long-read nanopore sequencing was used to map barcodes to bivalent effector pairs. The plasmid pool was prepared for sequencing using two Cas9 digestion steps followed by nanopore adapter ligation in an attempt to bias reads towards the barcoded bivalent effector region (Figure 41). First, two Alt-R S.p. Cas9 (IDT, 1081060) RNP pools were made, one containing the outer guides Ml and Pl and the other containing the inner guides M2 and P2, by prepooling 1 pL of each 100 pM crRNA and following the Alt-R CRISPR-Cas9 System protocol(Guides). The plasmid pool was then digested with the outer guides Ml and Pl (5 pgplasmid pool, 10 pL Ml+Pl Cas9 RNP, 4 pL 10X rCutSmart buffer (NEB, B6004S), and H2O to pL incubated at 37°C for 1 hour). The reaction was stopped by the addition of Proteinase K(+5 pL 20 mg / mL Proteinase K (Zymo, D3001-2-20), incubated at 56°C for 10ACTIVEUS 211418665 78Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 minutes). A IXAMPure XP bead (Beckman Coulter, A63881) cleanup was performed before dephosphorylation of ends followed by second digestion with the inner guides M2 and P2 (24 pL eluate, 3 pL Quick CIP (NEB, M0525S), and 3 pL 10X rCutSmart buffer (NEB, B6004S)incubated at 37°C for 10 minutes and heat inactivated at 80°C for 2 minutes, followed by addition of lOpL M2+P2 Cas9 RNP and incubation at 37°C for 1 hour). The reaction was stopped by the addition of Proteinase K (+5 pL 20 mg / mL Proteinase K (Zymo, D3001-2-20), incubated at 56°Cfor 10 minute). A 0.5X AMPure XP bead (Beckman Coulter, A63881) cleanup was performed before dA tailing using Taq polymerase (34 pL eluate, 1 pL 10 mM dNTP mix (NEB, N0447S), IpL Taq polymerase (NEB, M0267S), and 4 pL 10X rCutSmart buffer (NEB, B6004S) incubated at 72°C for 5 min). Another 0.5X AMPure XP bead (Beckman Coulter, A63881) cleanup was performed before nanopore adapter ligation (60 pL eluate, 25 pL Ligation Buffer (LNB, OxfordNanopore, SQK-LSK114), 10 pL NEBNext Quick T4 DNA Ligase (NEB, E6056S), and 5 pL Ligation Adapter (LA, Oxford Nanopore, SQK-LSK114) incubated at room temperature for 20 minutes). A final AMPure XP bead (Beckman Coulter, A63881) cleanup was performed following theSQK-LSKl 14 protocol using the Short Fragment Buffer (SFB, Oxford Nanopore, SQK-LSK114). The resulting library was quantified using Qubit IX dsDNA High Sensitivity Assay (Invitrogen,Q33231) and diluted to 20 fmol in 12 pL elution buffer (EB, Oxford Nanopore, SQK-LSK114) for sequencing on a MinlON flow cell (Oxford Nanopore, FLO-MINI 14) followed by a PromethlONflow cell (Oxford Nanopore, FLO-PRO114M). Nanopore reads were base called and duplexed using Dorado’s super-accuracy model v4.1.0 before conversion to .fastq file format by executing supduplex2fastq.sh on an appropriately configured GPU virtual machine. Barcode to effector pairings were then mapped by executing bc extraction.py. Briefly, Effector 1, Effector 2, and barcode sequences were extracted from each read using cutadapt and appropriate flanking sequences before effectors were mapped to individual effectors using minimap2 (Chen, S., et al 2018; Marin, M., et al 2011). Effector mappings with <75% identity were discarded. The mapped barcodes from all sequencing runs were then combined into a single table of unique pairings and filtered to remove any duplicate barcodes(mappedBCs filt.csv). In total, 7.8 million barcodes were mapped and represented every possible bivalent combination with the number of barcodes mapped scaling exponentially by bivalent effector length (Figures 4J and 4K).ACTIVEUS 211418665 79Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025High-throughput assay to measure transcriptional effects of Library 2

[0218] A pooled lentiviral library was generated from Library 2 using the LV-MAX kit (Gibco, A35348) and protocol in a IL flask. Lentivirus was concentrated 10X using Lenti-X concentrator following the manufacturer’s instructions (Takara, 631232). Functional titer was measured by transducing 40,000 K562 cells per well of a 96 well plate with serial dilutions of lentivirus and 8pg / mL polybrene (Millipore, TR-1003-G) via 90 minute spinfection at 1000g and 32°C. After overnight incubation, media containing lentivirus was replaced with fresh media and cells were allowed to recover for another 24 hours before being split equally into media with and without puromycin (2 pg / mL final, Gibco, Al 113803). After five days of selection, cell survival in each well was quantified on a Tecan Spark plate reader using CellTiter-Glo 2.0 Cell Viability Assay (Promega, G9242). Percent survival was calculated as the ratio of luminescence in the presence versus absence of puromycin for each lentiviral dilution, and functional lentiviral titer was calculated and averaged for all dilutions with 5- 30% survival. Two independent full scale transductions were then performed by transducing the PB Tet-On dCas9 3X CD81 K562 cell line at 100X library coverage and an MOI <0.1 with 8 pg / mL polybrene (Millipore, TR-1003-G) via 90 minute spinfection at 1000g and 32°C. After overnight incubation, media containing lentivirus was replaced with fresh media and cells were allowed to recover for another 24 hours before starting selection with 2 pg / mL puromycin for 5 days (Gibco, Al 113803). To assess bivalent chromatin effector activity in high-throughput, engineered cells were induced with 1 pg / mL doxycycline in 200 mL media in a IL Erlenmeyer flask, maintaining >1000X library coverage and replacing doxycycline media each day. After 5 days of induction (Day 0 timepoint), roughly 150 million cells (-3750X coverage) were frozen in 10 mL media supplemented with 10% DMSO at -80°C in a 15 mL falcon tube. Roughly 50 million remaining cells (-1250X coverage) were passaged to assess memory by washing twice with DPBS (Gibco) and resuspending in 400 mL fresh media without doxycycline in a 2L Erlenmeyer flask. Every 3 days for 24 total days, two frozen stocks of -3000X cells were made as described above, and cells were diluted to a minimum density of le5 / mL in 400 mL media in a 2L Erlenmeyer flask. Two frozen timepoints, Day 0 and Day 12, were chosen to assess transient effector activity and epigenetic memory on CD81 (Figure 7A). For a given timepoint and replicate, the frozen aliquot was thawed in a 37°C water bath for 4 min before being diluted in mL stain buffer (BD, 554656) and pelleted at 300g 1594 for 10 min. The cell pellet was resuspended in 14 mL stain buffer containing 750 pL PE-conjugated CD81 monoclonal antibody (Invitrogen, MAI-10292), andACTIVEUS 211418665 80Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 the tube was rotated for 30 min at 4°C. After staining, cells were spun down at 300g for 8 min and washed twice with 15 mL stain buffer before resuspending in mL and straining through a 40 pM filter. Using a BD FACS Aria Fusion Special Order Research Product, between 7-13 million cells were sorted from the top and bottom 25% of the PE fluorescence distribution (Figures 9B-9D). To measure bivalent effector representation in each sorted population, genomic DNA was harvested using NucleoSpin Blood L kit and protocol (Machery-Nagel, 740954.20), adding pL Monarch RNAse A (NEB, T3018L) before lysis and using high-yield elution recommendations. Then NGS libraries of bivalent effector barcodes were prepared using two sequential PCR steps. In PCR1, barcodes were amplified from gDNA using primer pairs containing staggers to diversity NGS reads (Primers, MH160- MH175) (up to 100 pg gDNA, 24 pL PCR1 primer set at 50 pM each, 600 pL NEBNext Ultra II Q5 MM (NEB, M0544L),and H2O to 1200 pL, divided into 12x 100 pL reactions in a PCR plate and incubated at 98°C for 10 seconds, 20 cycles of 98°C for 10 seconds and 65°C for 5 minutes, final extension at 65°C for 5 minutes). Split reactions were pooled and mixed thoroughly before 25 pL were cleaned using a two sided AMPure XP bead (Beckman Coulter, A63881) cleanup from 0.7X to 1.8X. In PCR2, sample indices were added along with flow cell binding sequences p5 and p7 (50-100 ng purified PCR1, 5 pL PCR2 primer set at 10 pM each, 25 pL NEBNext Ultra II Q5 MM (NEB, M0544L), and H2Oto 50 pL, incubated at 98°C for 30 seconds, 7 cycles of 98°C for 10 seconds and 65°C for 75 seconds, final extension at 65°C for 5 minutes). The reactions were cleaned using a two sided AMPure XP bead (Beckman Coulter, A63881) cleanup from 0.7X to 1.5X. Each sample was quantified using QubitX dsDNA High Sensitivity Assay (Invitrogen, Q33231) and 60 ng of each were pooled into a single NGS library that was sequenced on a NextSeq 2000 using a P3 2x51 cycle paired-end run (Illumina, 20040559).Targeted nanopore sequencing to assess barcode swapping in Library 2

[0219] To address the issue of barcode swapping, which is commonly associated with the use of barcodes in lentiviral libraries (Sack, L.M., et al, 2016), Library 2 contained only a short 39 bp homologous sequence between the end of the effector fusion and the barcode. Still, to quantitatively measure lentiviral recombination between bivalent effector pairings and barcodes, barcode swapping was assessed post-integration via long-read nanopore sequencing using Cas9-guided adapter ligation (Figure 8A) (Gilpatrick, T., et al, 2020). Genomic DNA from one full-scale lentiviral infection replicate was extracted using Quick- DNA Midiprep Plus kit and protocol (Zymo, D4075). An Alt-R S.p. Cas9(IDT, 1081060)ACTIVEUS 211418665 81Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025RNP pool was made by prepooling 1 pL of crRNAs Ml, M2, Pl, and P2 at lOOpM and following the Alt-R CRISPR-Cas9 System protocol (Guides). Genomic DNA was dephosphorylated before Cas9 cleavage and dA-tailing (5 pg gDNA, 3 pL Quick CIP (NEB,M0525S), 3 pL 10X rCutSmart buffer (NEB, B6004S), H2O to 30 pL incubated at 37°C for lOmin and heat inactivated at 80°C for 2 min, followed by addition of 10 pL M1+M2+P1+P2 Cas9RNP, 1 pL dATP at 10 mM (NEB, N0440S), and 1 pL Taq polymerase (NEB, M0267S) and incubation at 37°C for 20 min and 72°C for 5 min). The reaction was cleaned using AMPure XPbeads (Beckman Coulter, A63881) at 0.5X and DNA eluted in 60 pL H2O. Nanopore adapters were ligated following the SQK-LSK114 protocol, and the resulting library was sequenced on a MinlON flow cell (Oxford Nanopore, FLO-MINI 14). Nanopore reads were 1638 base called and converted to .fastq as previously described. Sequences containing complete reads through both effector positions were extracted using cutadapt and appropriate flanking sequences94. Barcode to effector pairings were mapped using bc extraction.py, modified to retain all found barcodes, which were then referenced against mappedBCs filt.csv. Reads from any mismatches between barcode to effector pairings were manually examined to determine the mismatch position (Figure 8B).NGS analysis of barcodes and bivalent domains from Library 2

[0220] NGS reads were converted into barcode counts using the script count barcodes.py. In summary, reads were merged with fastp before barcodes were excised with cutadapt using forward and reverse NGS primer sequences (Chen, S., et al, 2018;Martin, M., et al, 2011). Then, barcodes were dereplicated and tabulated using SeqFu, and the results were combined into a single table of barcode occurrences by experimental condition (Telatin, A., et al, 2021). After counting, any barcodes that were not mapped to effector pairs via nanopore were discarded, leaving an average of 76.6% of NGS reads (Figure 7E), and any barcodes containing homopolymer repeats of 8 or more were discarded due to limitations of nanopore sequencing. Next, barcode counts mapping to the same effector pair were combined to produce a total read count for each bivalent effector in each condition. Effector pairs were considered dropouts if there were less than 100 cumulative reads in both bins in either replicate and / or less than 5 reads in any bin of either replicate to prevent artifactual inflation or deflation of enrichment scores. Then read counts were normalized by the sum of each bin, and the enrichment scores were calculated as the ratio of the normalized counts in the HIGH bin versus the LOW bin. The average enrichment score from both replicates was computed using a geometric mean of the HIGH / LOW enrichment ratio.ACTIVEUS 211418665 82Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025Secondary validation of individual and bivalent effectors from Library 2

[0221] Selected individual and bivalent effectors from Library 2 were individually assessed via transient transfection into the PB Tet-On dCas9 K562 cell line with the 3X CD81 guide array (for assessing activity on CD81) or without the 3X CD81 guide array (for assessing activity on additional genes). To clone the constructs for transient transfection, a CcdB dropout landing pad was first made to assemble the pTRE3G-MCP-XTEN80-effector fusion while also expressing pEFla-EGFP-P2A-PuroR in a divergent orientation (Plasmids, pMH308). The individual clonal effectors ordered from Twist Biosciences were directly used for insertion into the Effector 2 position via the BsmBI sites in the flanking sequences. To generate the proper overhangs for insertion into the Effector 1 position, clonal effectors were predigested with Bsal (600 ng clonal fragment, 1 pL BsaI-HFv2 (NEB, R3733L), 1 pL Quick CIP (NEB, M0525S), 1 pLIOX rCutSmart buffer (NEB, B6004S), and H2O to 10 pL incubated at 37°C for 2 hours and 80°C for 1 hour). The same XTEN16 linker was used for bivalent assembly, while an Effector 2stuffer was used in place of both Effector 2 and the XTEN16 linker for monovalent assembly. To prepare the XTEN16 linker and Effector 2 stuff er, individual ssDNA oligos were ordered, annealed, and phosphorylated (1 pL 100 pM F oligo, 1 pL 100 pM R oligo, 1 pLT4 ligase buffer (NEB, B0202S), 1 pL T4 PNK (NEB, M0201S), and H2O to 6 pL incubated at 37°C for 30 minutes, 65°C for 20 minutes, and 95°C for 5 minutes before cooling down to 4°C over 15 minutes). The final assembly of bivalent effectors was performed using Golden Gate assembly (50 ng pMH308, 0.5 pL EFl predigest reaction, 0.2 pL XTEN16 oligo anneal, 0.5 pL EF2 clonal fragment at ~60 ng / pL, 1 pL T4 ligase buffer (NEB, B0202S), 0.5 pL Esp3I (Thermo Fisher Scientific, FD0454), 0.5 pL T4 ligase (NEB, R0202L), and H2O to 10 pL incubated at 37°C for 5 minutes, followed by 30 cycles of 37°C for 5 minutes and 16°C for 10 minutes, followed by final ligation at 16°C for 20 minutes, final digestion at 37°C for 30 minutes, and heat inactivation at 80°C for 20 minutes). The final assembly of monovalent effectors was also performed using golden gate assembly (50ng pMH308, 0.5 pL EFl predigest reaction, 0.2 pL EF2 stuffer oligo anneal, 1 pL T4 ligase buffer (NEB, B0202S), 0.5 pL Esp3I (Thermo Scientific, FD0454), 0.5 pL T4 ligase (NEB,R0202L), and H2O to 10 pL incubated at 37°C for 5 minutes, followed by 30 cycles of 37°C for 5 minutes and 16°C for 10 minutes, followed by final ligation at 16°C for 20 minutes, final digestion at 37°C for 30 minutes, and heat inactivation at 80°C for 20 minutes). 2 pL of each reaction were transformed into MACH1 cells (Thermo Fisher Scientific) before individual colonies were picked and grown overnight in 3 mL TB forACTIVEUS 211418665 83Atorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 plasmid purification with NucleoSpin Plasmid Transfection-grade miniprep kit (Machery- Nagel, 740490). All plasmids were sequence confirmed via full plasmid nanopore sequencing. Catalytic mutations for specific effectors were sourced through literature review. To clone mutations into sequence confirmed individual validation constructs generated above, 2-3 PCR fragments were made to overlap over the mutation site or sites and the GFP gene for Gibson assembly of the mutant plasmids (Primers, MH302-MH345). PCR fragments were generated from corresponding unmutated plasmids (10 ng parent plasmid, 0.75pL 10 pM F primer, 0.75 pL 10 pM R primer, 12.5 pL KAPA HiFi HotStart Ready Mix (Roche,KK2601), H2O to 10 pL incubated at 95°C for 3 minutes, 18 cycles of 98°C for 20 seconds and 72°C for 3 minutes, and 72°C for 6 minutes, followed by addition of 1 pL Dpnl (NEB, R0176S) and incubation at 37°C for 1 hour and 80°C for 20 min). 2 pL of the resulting reactions were analyzed for product length and purity via gel electrophoresis before mutant plasmids were assembled via Gibson assembly (1 pL each corresponding PCR reaction, 10 pL 2X Gibson MM, and H2O to 10 pL incubated at 50°C for 1 hour). 2 pL of each reaction were transformed into MACH1 cells (Thermo Fisher Scientific) before individual colonies were picked and grown overnight in 3 mL TB for plasmid purification with NucleoSpin Plasmid Transfection-grade miniprep kit (Machery -Nagel, 740490). All mutant plasmids were sequence confirmed via full plasmid nanopore sequencing. Sequence confirmed validation plasmids were nucleofected into the PB Tet-On dCas9K562 cell lines using SF Cell Line 96-well Nucleofector kit (Lonza, V4SC-2096). The manufacturer's instructions were followed using 800 ng plasmid DNA per nucleofection. For testing effectors on CD81, only the effector plasmid was nucleofected into the Tet-On dCas9cell line with the 3X CD81 guide array. For testing effectors on additional genes, the effector plasmid and the corresponding 3X guide array plasmid were nucleofected at a 1 : 1 mass ratio into the Tet-On dCas9 cell line without the 3X CD81 guide array. After nucleofection, cells were plated directly into 1 pg / mL final concentration of doxycycline to induce dCas9 and MCP- effector expression. One day after nucleofection and every subsequent day until 5 days post nucleofection, media was replaced with fresh media containing 1 pg / mL doxycycline and 2pg / mL puromycin (Gibco, Al 113803). Five days post-nucleofection, cells were analyzed for transient effector activity on the gene target by transferring to 96 well U-bottom plates and spinning down at 300g for 3 min. Media was aspirated, and cells were resuspended in 100 pL stain buffer (BD, 554656) per well. Then 100 pL stain 1726 buffer containing 1.25 pL fluorescently conjugated antibody was added to each well and mixed by pipetting beforeACTIVEUS 211418665 84Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 incubation at 4°C for min (APC CD55 antibody (Biolegends, 311311), Alexa Fluor 647 CD58 antibody (BDPharmingen, 563567), PE CD81 antibody (Invitrogen, MAI-10292), APC CD151 antibody(Biolegends, 350405), and APC CD155 antibody (eBioscience, 2H7CD155). After incubation, cells were washed twice with 200 pL stain buffer before final resuspension in 200 pL stain buffer per well. Stained cells were analyzed by flow cytometry on an Attune NxT Flow Cytometer (Thermo Fisher Scientific). To assess for long-term effector activity, cells at five days post nucleofection were transferred to 96 well V-bottom plates and washed twice with 200 pL DPBS (Gibco, 14190144) per well before resuspending in fresh media without doxycycline or puromycin. Three days later and at subsequent timepoints, cells were split to maintain growth, and the remaining cells were stained and analyzed as above. To systematically analyze large batches of flow cytometry data, Cytoflow’s Jupyter notebook integration was used (Teague, B., 2022). Live and singlet gates were first defined on WT K562 cells measured using the same cytometer settings, and the gates were applied to each sample file. Then, for transient timepoint analysis, cells were further gated for BFP and GFP, indicating dCas9 expression and the presence of the MCP- effector plasmid. Transient timepoints on genes other than CD81 were also gated on mCherry, indicating the presence of the 3X guide array plasmid. Threshold gates were defined at the 99.9th percentile of the channel measurements in WT K562 cells. For memory timepoints, no gating beyond live and singlet gates was performed. After gating, mean fluorescence intensity (MFI) was calculated for each sample using the flow.geom mean function in Cytoflow on the channel corresponding to the fluorescent antibody. Percent change in MFI was calculated using the MFI from the neutral DMD control as a reference. Percent cells activated and percent cells repressed were calculated using threshold gating at the 1st and 99th percentile of the neutral DMD control effector, and the percent cells silenced was calculated using threshold gating at the 99th percentile of unstained WT K562 cells.Analysis of individual and bivalent effector strength

[0222] To calculate the total number of significant effector pairs for each library, the raw count matrices were analyzed with pyDESeq2 (Muzellec, B., et al, 2023) which corresponds to DESeq2 (vl.34.0) with single factor analysis, Wald tests, and LFC shrinkage mode. For screen 1, candidates with total NGS reads of at least 80 were analyzed. For screen 2, the total NGS reads threshold was set as 100. A marginal effect test was defined to assess the marginal, or average, effect of a given individual effector. To conduct the test for a given effector, the bivalent pairs and corresponding average fold enrichment scores were split intoACTIVEUS 211418665 85Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 two groups, those containing the effector and those not containing the effector. The marginal score was defined as the difference in mean fold enrichment between the two groups, and if > 30 pairs were present in both groups, a Welsh’s t test was conducted. It was determined that a difference of means t-test was necessary, rather than a one sample t-test, to correct for library composition bias. In other words, the mean fold enrichment of any given individual effector would naturally be positive or negative if the overall library had a bias in either direction simply because the given effector is paired with every other library member. A Welsh’s t-test was chosen as opposed to a Student’s t-test because it was expected that individual effectors may vary in their interactions with other library members, thus not only altering the mean fold enrichment but also the standard deviation of fold enrichment scores. Once all the effectors were analyzed, p-values were corrected for a false discovery rate of 5%.Synergy scoring to identify synergistic and antagonistic combinations

[0223] To identify bivalent combinations with unexpected activities, synergy scores were calculated for each bivalent combination based on the expected effects of each individual effector (Figure 14A). First, the individual effects of each effector were calculated using the marginal score metric, with the difference that each effector was scored for both the N- terminal and C-terminal position. Then, the sum of the corresponding marginal scores for each bivalent combination was computed. Because effectors could still be additive in nature even if the measured score did not exactly match the sum, a range for additive activity was defined that assumes that effectors may not contribute their full marginal strength to the combination but that the contribution of each effector should still be in the expected direction. For expected repressors, the lower bound of the additive range was defined as the minimum of the N-terminal marginal score, the C-terminal marginal score, and the sum of the marginal scores (min(N-term, C-term, Sum)), and the upper bound of the range was defined as the maximum of the sum of the marginal scores and the minimum of the two marginal scores (max(Sum, min(N-term, C-term))). For expected activators, the expressions are reversed and the additive range becomes (min(Sum, max(N-term, C-term))) to (max(N-term, C-term, Sum)). Log2 fold enrichment scores from HTS data were first mean centered, a logical step considering that marginal scores are relative contributions and are unaffected by central tendency, and subsequently compared to the additive range. In all cases, if the observed data fell within the additive range, the pair was considered additive and assigned a synergy score of 0. For expected repressors, bivalent combinations were considered “synergistic” if the measured score fell below the additive range, i.e. it was more repressiveACTIVEUS 211418665 86Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 than expected, and the combination was considered “antagonistic” if the measured score fell above the additive range, i.e. it was less repressive than expected. Scores for expected activators were determined in the same way but in the opposite direction, with synergistic combinations exhibiting unexpectedly high observed scores and antagonistic combinations exhibiting unexpectedly low observed scores. In both cases, the synergy score was defined as how far outside the additive range the observed score fell, being positive for synergy and negative for antagonism. Because Library 1 was entirely composed of expected repressors, every combination was scored as a repressor. For Library 2, the sum of the marginal scores determined whether the combination was scored as an activator or a repressor with positive sums being scored as expected activators and negative sums being scored as expected repressors. To test if individual effectors broadly synergize or antagonize with KRAB members from Library 1 or HD AC members from Library 2, the synergy scores from all combinations containing the given effector with a KRAB or HD AC member were tested against the null theory that the average synergy score was zero using a one-sample Wilcoxon test. Only KRAB or HD AC members with marginal scores less than -0.5 were considered, and p-values were subsequently FDR corrected for multiple testing. In addition to performing significance testing, the average synergy score and the average log2 fold enrichment score for the combinations were calculated.Comparison between stdMCP-KRAB and stdMCP-KRAB-L3MBTL3

[0224] pTRE3G-PuroR-T2A-stdMCP-XTEN80-KRAB(ZNF10)-L3MBTL3-WPRE and pTRE3G-PuroR1818T2A-stdMCP-XTEN80-KRAB(ZNF10)-WPRE were infected by lentivirus to Tet-On dCas9 K562(without CD81 guide array) in similar MOI (0.35 ± 0.04 for KRAB-L3MBTL3 and 0.31 ± 0.01 for KRAB). For comparing across different panels of genes and guides, an equal amount of lentivirus encoding both sgRNA and mCherry marker were infected to both cell lines. Doxycycline concentration was maintained at 1-2 pg / mL since the LV-sgRNA infection. 5 days after lentiviral infection of sgRNAs, cells were stained with corresponding antibodies (APC anti 1824human CD55 antibody (Cat# 311311, Biolegends), Alexa Fluor 647 Mouse Anti-Human CD58 antibody (Cat# 563567, BD Pharmingen), APC Mouse Anti -Human CD81 antibody (Cat#, BD Bioscience), and APC anti-human CD151 antibody (Cat# 350405, Biolegends)) and expression level was measured by Attune NxT Flow Cytometer (Thermo Fisher Scientific). For the bidirectional perturbations using the MS2 / PP7 system, pTREtight-stdPCP-p65-HSFl-T2A1829BFP were lentivirally infected to the previous two cell lines described above. pTREtight-ACTIVEUS 211418665 87Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 stdPCP1830p65-HSFl-T2A-BFP was cloned from the construct pHR-CMV-stdPCP-p65- HSF1-T2A-GFP (Addgene #199456) (Liang et al 2022). The same amounts of lentivirus encoding both mouse U6 promoter driven 2x MS2 sgRNA and human U6 promoter-driven 2x PP7 sgRNA were infected to both cell lines. Doxycycline concentration was maintained at 1-2 pg / mL since the day of LV-dualMS2 / PP7 sgRNAs infection. CD81 / CD274 expression profiles were measured 5 days post infection of dual MS2 / PP7 sgRNAs following staining with antibodies against both targets (APC Mouse Anti-Human CD81 antibody (Cat# 561958, BD Bioscience) and Alexa Fluor 488 Mouse Anti-Human CD274 antibody (Cat# 53-5983- 42, Invitrogen)) by Attune NxT Flow Cytometer(Thermo Fisher Scientific).Additional MaterialsNote 1

[0225] Recognition of epigenetic modifications by a reader domain is a crucial step in establishing epigenetic feedback loops, representing a fundamental aspect of epigenetic interactions. Specifically, PRC 1 and PRC2, responsible for installing H2AK119ubl and H3K27me3, respectively, are well-known for their intricate feedback mechanisms. These mechanisms involve the spreading of a single epigenetic modification or cross-talk between two distinct modifications (Zhang, T., et al, 2015), all of which require the recognition of epigenetic modifications by the reader domain (Lukauskas, S., et al, 2024). By focusing on the sub-heatmap defined by recruiters of PRC 1 and PRC2 on the N terminus and readers of epigenetic modifications on the C-terminus, a prominent signal was noticed that emerged at the intersection of a subset of PRC 1 recruiters and theH2AKl 19ubl reader (Figure 17A). Out of 483 combinations within the sub-heatmap, 5 combinations out of the top 8 most repressive combinations were from combinations betweenPRCl recruiters and the H2AK119 monoubiquitination (H2AK119ubl) reader domain from RYBP protein. Also, theH2AK1 19ubl reader domain was nominated as the strongest synergistic C-terminus partner of PRC 1 recruiters excluding the KRAB family, although the q value was higher than the threshold of 0.05 (Figure 17B).

[0226] To validate these combinations between PRC1 recruiters and the H2AK119ubl reader, several PRC1 recruiter domains encompassing domains from canonical PRC1(CBX2 / CBX7) and variant PRC1 (RYBP) were tested, in combination with the readers of distinct histone modifications (H2AK119ubl by the RYBP Ranbp2-ZnF domain, H3K27me by the CBX6 Chromo domain, H3K9me by the CBX5 Chromo domain, and H3K4me by the ING1 PHD domain) (Figure 17C). After 4 days of recruitment, the resultsACTIVEUS 211418665 88Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 confirmed that only the combinations containing the H2AK119ubl reader domain from RYBP were sufficient to induce substantial gene silencing (CBX2 + H2AK119ubl reader: 12.7%, CBX7 + H2AK119ubl reader: 59.7%, RYBP + H2AK119ubl reader: 60.5% CD81 repressed cells), irrespective of whether thePRCl component originated from the canonical or variant complex (Figure 17D). Interestingly, the validated fusion protein containing the PRC1 recruiter domain and H2AK119ubl reader domain from RYBP essentially reconstitutes full- length RYBP with some truncations (Figure 17E). Variant PRC1 (vPRCl) comprising RYBP subunits has the unique ability to both read and write the H2AK119ubl mark owing to the H2AK1 19ubl reader domains found within RYBP protein (Zhao, J., et al, 2020). This ability to read and write the same epigenetic modifications has been speculated as an efficient strategy for propagating silencing marks to adjacent regions (Grewal, S.I.S, 2023), suggesting that interactions detected from the HTS experiment may recapitulate a natural architecture inherent within the PRC1 complex (Zhao, J., et al, 2020)EXAMPLE 3: Combinatorial interaction exploration (COMBINE) enables high- throughput bivalent domain screening.

[0227] Considering the combinatorial nature of epigenetic regulation (Figure 1 A), the first step was to assess the functional effects of epigenetic domain pairs on gene expression to uncover new biology and potential domain combinations for epigenome editing tools. The limitations of currently available bivalent domain screening technologies necessitated the development of a new pairwise domain screening approach that could accommodate larger domain sizes (many over 1000 amino acids in length) to assess combinations of epigenetic reader, recruiter, writer, and eraser domains for their effects on gene expression in a high- throughput pooled screen.

[0228] To achieve this, two complementary libraries of epigenetic effectors were curated, each with distinct design principles. Library 1 was designed to identify combinatorial epigenetic repressors and is composed of readers, recruiters, structural factors, writers, and erasers from known repressive epigenetic multiprotein complexes (Figures IB; Figure 37). It contains a total of 155 domains ranging from 80 aa to 545 aa (Figures 2A and 2B). Library 2 was designed to investigate how diverse types of direct epigenetic modifications interact to modulate transcription in both directions (repression or activation). This library is exclusively composed of catalytic writers and erasers of histone and DNA modifications, irrespective of their known effects on transcription (Figure 1C; Figure 38). It contains 198 domains, with a wider size distribution from 100 aa to 1,036 aa (Figures 2C and 2D). To mitigate length biasACTIVEUS 211418665 89Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 in lentivirus generation, transduction (Sweeney, N.P., and Vink, C.A. (2021) and NGS readout (Gohl, D.M., et al, 2019), library members of each library were pooled in a weighted fashion, and sequential golden gate assembly with CcdB dropout selection was utilized to increase cloning efficiency (Methods, Figures 3 and 4)

[0229] To recruit these epigenetic effectors to an endogenous gene promoter, dCas9 targeting was used in combination with the MS2-MCP aptamer system derived from RNA bacteriophage MS2, analogous to the prior designs that utilize sgRNAs with embedded MS2 stem-loops (Konermann, S., et al, 2018; Heidersbach, A. J., et al., 2019). For temporal control over the editing complex, doxycycline-inducible promoters drove the expression of both dCas9 and the effector combination fused to monomeric MS2-coat protein (MCP) or synonymous mutated tandem MCP (stdMCP) (Figures 5A and 5E) (Wu, B., et al., 2015). To assess transcriptional outcomes on an endogenous target locus, the gene encoding membrane protein CD81, whose expression can be measured using flow cytometry, was chosen. A three sgRNA expression cassette targeted upstream of the CD81 transcription start site (TSS) was designed, and its ability to transiently activate and repress CD81 expression with this experimental setup was confirmed. Using well-characterized effectors including KRAB and the transcriptional activation domain of HSF1, CD81 expression levels showed a good dynamic range of both repression and activation during doxycycline treatment and returned to baseline at 9 days after the removal of doxycycline (Methods, Figures 5B-D and 5F-5H).

[0230] For this screen, lentiviral libraries encoding Library 1 or Library 2 were transduced into K562 cells containing Tet-On dCas9 at a low multiplicity of infection (MOI<0.3) with a minimum effector coverage of 100X (Methods, Figures ID, 6A, and 7A). The combinatorial effector was recruited to CD81 loci for 5-6 days by adding doxycycline, followed by cell sorting based on CD81 expression (Day 0 timepoint to evaluate immediate effects on gene expression) (Methods, Figures 6B-D and 7B-D). This was followed by a 12- day effector washout period without doxycycline and a final round of cell sorting (Day 12 timepoint to evaluate durable effects on gene expression in the absence of effectors).

[0231] The size differences between Library 1 and Library 2 effector domain combinations (Figure IE) necessitated different NGS readout strategies. The majority (-82%) of bivalent members in Library 1 fell below the 1500 nucleotide upper limit for short-read NGS platforms34, enabling us to directly sequence the N- and C-termini of effector pairs using short-read sequencing (Figure 6A). In contrast, Library 2 often exceeded this limit. To overcome this, a short 20N barcode was included in the cloning process, which was initiallyACTIVEUS 211418665 90Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 mapped to the effector pair by nanopore sequencing and read out by short-read sequencing at the end of the screen (Methods, Figures 4G-K, 7A, and 7E). Targeted nanopore sequencing of integrated Library 2 members indicated less than 10% barcode swapping (Figure 8). This barcoded approach with exponential effector pooling for Library 2 resulted in a reduction of the Day 0 dropout rate from% in Library 1 to 4.2% and successfully offset the exponential decay of effector coverage with increasing length that was observed for Library 1 and previous combinatorial screening approaches (Figures 9A and 9D) (Blaeschke, F., et al, 2023; Gordon, K.S., et al, 2022). Effector dropout on Day 12 remained consistent at 42.1% for Library 1 and increased to 19.3% for Library 2 (Figures 9B, 9C, and 9E).

[0232] The transcriptional effects of each domain combination were determined by measuring the relative abundance of each unique effector pair in the sorted cells from the high and low CD81 expression bins, calculated as log2 fold enrichment. Both libraries effectively captured the temporal dynamics of CD81 transcriptional perturbation induced by well-established control effectors. In Library 1, target gene repression by combinations containing the KRAB domain from ZNF10 was strong at Day 0 and disappeared by Day 12 (log2 fold score of -4.7 and -0.24, Figures IF and 1G). Similarly, repression by full-length RYBP, a known epigenetic silencer that recruits Polycomb repressive complex 1 (PRC1) (Zhao, J., et al, 2020), and activation by the activation domain of HSF1 were captured at Day 0 and dissipated by Day 12 after dox removal for Library 2 (average log2 fold scores of -3.6 and 0.083 for RYBP and 1.9 and 0.047 for HSF1, Figures 1H and II). Based on both CD81 expression levels measured by flow cytometry and HTS enrichment scores (Figures 6C), the majority of transcriptional effects disappeared by 12 days after the removal of doxycycline (Day 12), highlighting the transient nature of perturbations introduced by the majority of epigenetic and transcriptional modifiers.EXAMPLE 4: COMBINE generates rich combinatorial landscapes of pairwise epigenetic interactions

[0233] To visualize transient perturbation outcomes across the large space of combinations tested, enrichment score heatmaps were generated (Figures 10A, 10B, and 11). Clear diagonal symmetry was observed between the X-Y and Y-X orientations of the same two effectors, with a high degree of correlation across pairs, indicating that most effector pairs behaved similarly regardless of orientation (Library: 1 r = 0.75; Library 2: r = 0.69, Figure 12).ACTIVEUS 211418665 91Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025

[0234] To identify general trends for the transcriptional impact of effector classes, pairs were grouped based on the families of individual effectors (Figures 10C and 10D). While Library 1 revealed the strongest repressive phenotypes among KRAB family pairs, Library 2 revealed strong repression across pairs of histone deacetylases (HDACs) and DNA methyltransferases (DNMTs). A surprising trend was also observed toward mild target gene repression with pairs of DNA demethylation machinery (DNDM) domains. Pairs of histone acetyltransferases (HATs)tended to activate gene expression, as did histone deubiquitinases (DUBs) despite their ability to remove both activating and repressive histone ubiquitination marks. Finally, pairs of histone lysine methyltransferases (HKMTs) and histone kinases (HKs) ranged broadly across both activation and repression outcomes.

[0235] Next, directionality of individual effectors in the context of these class-level trends was examined. In the HAT class, pairs containing C-terminal EP300 (a broad HAT capable of acetylating all four core histones (Ogryzko, V.V., et al, 1996) tended to activate transcription, whereas combinations containing C-terminal CDY1 (which displays a strong acetylation preference for histone H4 (Lahn, B.T., et al, 2002)) tended to repress transcription (Figure 10E). In addition, some HAT effectors, such as KAT6A or KAT14 (a broad histone and weak H4 acetyltransferase, respectively (Champagne, N, et al, 2001; Gueman, S., et al, 2009)), exhibited more partner-dependent effects.

[0236] Compared to the HAT class, the HD AC domain class had more members that consistently repressed transcription irrespective of their N-terminal partners such as SIRT1 (a broad histone deacetylase and decrotonylase (Wei, W., et al, 2017)) and class Ila HDACs (HDAC5, HDAC4, and HDAC7, Figure 10F). However, the degree of repression ranged widely depending on the exact effector pairing. Without being bound by theory, the combination of EP300 and HDAC5 activated transcription with an average log2fold enrichment of 1.1, while the pairing of UBE2E1 and HDAC5 repressed transcription with a score of -4.5. These global observations motivated us to quantitatively analyze the contribution of individual domain effects and domain-domain interactions to gene expression outcomes.EXAMPLE 5: Marginal effector analysis reveals how individual domains 234 affect transcription irrespective of partners

[0237] To quantitatively assess how each of the individual effector domains in the two libraries contribute to epigenetic control of gene expression, a ‘marginal effector score’ was calculated for each domain (Figures 13 A and 13B). This score reflects the averageACTIVEUS 211418665 92Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025 enrichment value across all bivalent combinations containing the domain of interest relative to all combinations lacking it, yielding a measurement of each domain’s relative contribution to target gene expression (positive or negative) across pairings.

[0238] Marginal effector analysis of Library 1 clearly reflected the well-known repressive effect of KRAB domains on gene expression (Bintu, L., et al 2016; Tycko, J. et al, 2020). The average marginal score across the KRAB domain class was -3.8, indicating an average 14- fold decrease in enrichment scores when KRAB is present in an effector pair (23 8~ 14, Figure 13C). The ZNF10 (K0X1) KRAB domain - which is commonly used in CRISPR interference (CRISPRi) applications (Gilbert, L.A., et al, 2013) - exhibited the lowest marginal score at -5.8. The more recently identified ZIM3 KRAB domain also exhibited strong repression with a score of -4.520. Notably, only two out of 25 tested KRAB domains did not consistently repress gene expression (ZNF791 and ZNF823; marginal scores of -0.1 and 1.0), suggesting that this family contains a wide spectrum of additional potential tool domains that have yet to be explored.

[0239] Outside the KRAB domain family, two other individual domains were identified with significantly repressive marginal scores in Library 1 : SID4X (marginal score -4.3) and the sterile alpha motif (SAM) domain of L3MBTL3 (marginal score -1.1). SID4X is an engineered ternary version of the Sin3 interacting domain (SID) from the MAX dimerization protein 1 (MXD1) that efficiently recruits the Sin3-HDAC complex (Cong, L., et al 2012), and the SAM domain of L3MBTL3 is known to multimerize and recruit lysine-specific histone demethylase 1A (LSD1 / KDM1A) (Knight, M.J., et al, 2011; Xu, T., et al, 2017), providing clear mechanistic theories for the repressive effects of each of these domains.

[0240] To independently validate each domain’s individual contribution to transcriptional repression outside of the bivalent effector context, ZNF10 KRAB, SID4X, or L3MBTL3 SAM was fused to stdMCP and expressed in the Tet-On dCas9 K562 cell line alongside a CD81 -targeting guide array (Figure 13D). Repression of CD81 target gene expression, as measured by the change in total mean fluorescence intensity (MFI), correlated with each domain’s relative marginal effect score: ZNF10 KRAB decreased CD81 MFI by 99%, SID4X by 90%, and L3MBTL3 by 46% relative to the negative control. Notably, unlike the all-or-nothing silencing pattern of KRAB (Bintu, L., et al, 2016), neither SID4X nor L3MBTL3 fully silenced CD81, but rather resulted in more graded repression.

[0241] The marginal scores calculated from Library 2 were lower in magnitude, likely reflecting the more varied composition of this library in comparison to the repression-focusedACTIVEUS 211418665 93Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Library 1 (Figure 13E). At the class level, the DNMT class exhibited the highest average repression, with a moderate effect size but high consistency across members (average marginal score -0.87). The DUB class resulted in the highest average gene activation with moderate but consistent effects (average marginal score 0.47). At the individual 277 effector level, five of the seven top repressive domains in Library 2 belonged to the HD AC class (including HDAC7 and HDAC4 with scores of -2.2 and -2.1, respectively), while the HAT class contained the two strongest individual activators (EP300 and KAT5 with scores of 1.8 and 1.6, respectively).

[0242] Next, the effects of 5 individual domains were confirmed by fusing them to MCP and transiently expressing them in the same cell line used above. The strongest repressor, HDAC7 (marginal score -2.2), reduced CD81 MFI the most, by 64%, and was modulated by the introduction of point mutations (Figure 13F). The hyperactive H843Y mutant (Schuetz, A., et al, 2008) increased repression to 82%, and the hypoactive H670L mutant (Wang, A.H., et al, 1999) diminished repression to 48% (Methods). Three additional repressors from diverse classes also repressed CD81 in concordance with their marginal scores. RAG1 (marginal score -0.72), an E3 ubiquitin ligase (E3 UBL) associated with histone H3 monoubiquitination during V(D)J recombination (Grazini, U. et al 2010; Deng, Z., et al, 2015), repressed CD81 by 39%;UBE2E1 (marginal score -0.69), an E2 ubiquitin ligase (E2 UBL) associated with PRC1 (Wheaton, K, et al, 2017), repressed CD81 by 35%; and KMT5B (marginal score -0.12), a histone lysine methyltransferase(HKMT) that writes H3K20 methylation (Weirich, S., et al 2016) - a modification for which no targeted epigenetic editing tools currently exist, repressed CD81 by 32% (Figure 13G). Catalytic mutations of these three enzymes (Yurchenko, V., et al 2003; Sun, L. and Fang, J., 2016; Wu, H., et al, 2013) reduced CD81 repression to 24%, 9%, and 10% respectively.

[0243] For activation, KAT5 (marginal score 1.6) was the second-ranking activator of gene expression in the screen, and validation experiments confirmed that it resulted in a 15% increase in CD81MFI (Figure 13G) which was completely abolished upon introduction of a catalytic point mutation. Altogether, the marginal effector analysis and validations of domains from Library 2 indicated a wide range of repressive and activating effects. Importantly, it was found that the observed transcriptional effects of each of these enzymatic domains were mediated by their catalytic activity as the incorporation of hypoactive or hyperactive variants led to expected changes in gene expression outcomes.ACTIVEUS 211418665 94Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025EXAMPLE 6: COMBINE reveals synergy and antagonism in epigenetic domain interactions

[0244] Beyond the effects of individual domains on target gene expression, a key advantage of COMBINE is the ability to quantify synergistic or antagonistic interactions among domain pairs. To systematically assess domain interactions, a ‘synergy score’ was calculated for each effector pair that uses individual marginal scores to calculate the difference between expected and observed effector strength for each domain combination (Methods, Figure 14A). Positive scores indicate synergy and negative scores indicate antagonism in the direction of the stronger marginal effector. Across both libraries, many strong synergistic or antagonistic interactions were identified between domains that provide a rich resource for further exploration (Figures 15 and 16). Compellingly, this analysis highlighted synergistic interactions between several PRC1 recruiters and a reader of H2AK1 19 monoubiquitination, recapitulating the known reader- writer synergy of variant PRC1 (Figure 17; Note 1) (Zhao, J. et al, 2020).

[0245] When analyzing the synergy scores for all KRAB and HD AC members with clear individual repressive effects (marginal scores below -0.5, Figure 18), additive, synergistic, and antagonistic effector pairings with other library members were able to be classified. Without being bound by theory, it was observed that the strongly activating HAT members KAT5 and EP300 fully canceled out the repressive effect of HDACs in a way that was explained by a direct additive relationship (Figure 14B). In contrast, 46 effectors resulted in synergistic enhancement of repression by HD AC domains. Two pairs were selected with particularly high synergy scores for experimental validation: HDAC7 + UBE2E1 (synergy score 2.4) and HDAC4+ VRK1 (synergy score 2.0).

[0246] In validation experiments, it was found that the HDAC7 + UBE2E1 and HDAC4 + VRK1 combinations reduced CD81 MFI by 75% and 73%, respectively, relative to the neutral control. This enhanced repressive effect was at least partially dependent on the catalytic activities of both effectors, as hypoactive mutations in the HD AC partners (Wang, A.H., et al, 1999) reduced each combination’s repression level to 63% and 36%, respectively, while loss-of-function mutations in the UBE2Eland VRK1 partners (Sun, L and Fang, J., 2016; Kang, T.H., et al, 2007) attenuated repression to 57% and 48%, respectively (Figure 14C and 14D). Strikingly, installing hyperactive mutations into the HD AC partner (Schuetz,ACTIVEUS 211418665 95Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025A., et al 2008) boosted CD81 repression to 92% for both of these combinations, further emphasizing the role of HD AC catalytic activity in this outcome.

[0247] As an E2 UBL associated with PRC1, UBE2E1 participates in the ubiquitination of H2AK119 to promote gene repression (Wheaton K., et al, 2017). Without being bound by theory, gene repression by HDACs is enhanced by these epigenetic effectors since the removal of histone acetylation marks near the CD81 locus (Figure 19A) (Kent, W.J., et al, 2002) could increase the efficiency of writing repressive marks through established PRC crosstalk mechanisms (Zhang, T., et al, 2015). While future mechanistic studies are needed, these results demonstrate the ability of the combinatorial screening platform to generate novel mechanistic theories about fundamental biological processes.

[0248] To address the generalizability of the potent HD AC combinations as well as individual domains identified from Library 2, 13 key effectors on a panel of 4 additional basally expressed genes including CD55, CD58, CD151, and CD155 were tested. Five days post-nucleofection of both effector and 3X guide array plasmids, it was observed that the percent changes in MFI for each of the effectors were similar across all tested genes (Figure 14E and 20). The percent changes were highly correlated between genes, spanning a range of r = 0.87-0.98 and demonstrating that the findings are broadly applicable to other target genes (Figure 14F and 21).

[0249] Next, synergy scores among the KRAB family of repressive domains were examined to discern partner domains that could enhance or antagonize KRAB activity. The synergy score analyses identified the SAM domain from L3MBTL3 and two PRC2 recruiters (PALI1 and EPOP) as the most potent synergistic partners of KRAB domains, all exhibiting both high average synergy scores and low enrichment scores on Day 6 (average synergy scores 4.0, 2.0, and 1.7 and log2fold enrichments -5.7, -3.7, and -3.4, respectively) (Figures 14G and 15B). In individual validation of these domain combinations, all three KRAB combinations - L3MBTL3, PALI1, and EPOP - further enhanced repression of CD81 compared to the ZNF10 KRAB domain alone (L3MBTL3 + KRAB, KRAB + PALI1, GFP + KRAB were 97.7%, 86.6%, and% of CD81 -repressed cells, Figures 14H, 22A, and 22B).

[0250] PALI1 and EPOP are vertebrate-specific and mutually exclusive PRC2 subunits (Conway, E., et al, 2018) that can recruit the PRC2.1 complex, thereby promoting the deposition of H3K27 trimethylation (Zhang, Q, et al, 2021). In the screen, the PRC2 interacting domains (PID) of PALI1 [1058 - 1250 aa] and EPOP [300 - 379aa] exhibited synergy with KRAB. KRAB + PALI! PID and KRAB + EPOP PID combinations areACTIVEUS 211418665 96Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 expected to install both H3K9 methylation marks via KRAB and H3K27 trimethylation marks through recruitment of PRC2. Notably, in its endogenous context, the full-length PALI1 protein [1 - 1557 aa] is also capable of installing both H3K9 and H3K27 methylation marks, as it contains both the H3K9 methyltransferase G9A interacting domain [310 - 723 aa] and PRC2 interacting domain [1058 - 1250 aa] (Fong, K.-W., et al, 2022). This demonstrates the ability of COMBINE to highlight naturally occurring synergistic combinations of epigenetic domains.

[0251] Moving beyond these synergistic partners, two strong antagonistic partners of KRAB were also identified: the catalytic JmjC domains of histone lysine demethylases KDM2B and KDM5B (average synergy scores -1.4 and -1.5, respectively) (Figures 14G and 33). It was confirmed that fusing these JmjC domains to the ZNF10 KRAB domain abolished the KRAB domain's repressive activity, resulting in a complete recovery to baseline CD81 expression (Figure 22C). On the other hand, no significant change in CD81 expression was found upon the recruitment of these JmjC domains alone to the CD81 locus. The demethylation substrate preferences of these two domains have been only partially characterized, with some reports suggesting that KDM2B demethylates H3K4me3 (Janzer, A., et al, 2012), H3K36mel / 2 (He, J., et al, 2008), and H3K79me2 / 3 (Kang, J.-Y, et al, 2018), while KDM5B has been reported to demethylate H3K4me2 / 3 (Yamane, K., et al, 2007) and H2BK43me2 (Biggar, K.K., et al, 2020). Most of these previously described demethylation substrates are histone marks indicative of active transcriptional status.

[0252] Interestingly, it was also observed and validated negative synergy scores for these JmjC domains when paired with another transcriptional repressor, SID4X, suggesting that these domains also antagonize SID4X-mediated transcriptional repression (Figures 22C and 22D). SID4X is known to recruit the Sin3-HDAC complex, which can erase histone acetylation marks, resulting in the repression of target gene transcription (Cong, L. et al, 2010). Future work will dissect the precise mechanisms by which these JmjC domains reverse KRAB- and SID4X-mediated target gene repression.EXAMPLE 7: KRAB + L3MBTL3 combination enables potent transcriptional repression in challenging conditions

[0253] Individual KRAB domains, including those of ZNF10 and ZIM3, have been harnessed as epigenetic repressors in many contexts (Alerasool, N., et al, 2020; Gilbert, L.A., et al, 2013) and yielded strong individual marginal scores in the screen (-5.8 and -4.5, respectively). However, in certain contexts - for example, under dose limited conditions dueACTIVEUS 211418665 97Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 to viral delivery, or when recruited to a locus in a more indirect manner -further enhancement of repressive activity would be beneficial. To provide new tool options for these challenging conditions, the screen data was examined for synergistic combinations that would enhance the potency of these KRAB domains.

[0254] In the screen data as well as in individual validation, the synergistic pairing of ZNF10 KRAB with the SAM domain from L3MBTL3 was the most potent combination identified. Notably, the SAM domain is known for its ability to multimerize (Knight, M.J., et al, 2011), potentially recruiting multiple KRAB domains to the target gene (Figure 23A). This combination was tested under dose-limiting conditions using low MOI lentiviral transduction of K562 cells with either the stdMCP-KRAB or stdMCP-KRAB-L3MBTL3 effectors combined with a single sgRNA - instead of three - targeting CD81. Significantly enhanced repression of the combined KRAB + L3MBTL3 effector under these challenging conditions was observed, particularly at low levels of doxycycline induction where we achieved up to ~34-fold increased repression (KRAB: 1.4% vs KRAB + L3MBTL3: 48.1% at 5 ng / pL dox) (Figure 23B). Even at standard 1000 ng / pL doxycycline conditions, the KRAB + L3MBTL3 combination exhibited ~6-fold increased repression compared to KRAB alone.

[0255] To test the generalizability of this effect, three additional gene targets encoding cell surface proteins with basal expression in K562 cells were selected (CD55, CD58, and CD151). The KRAB +L3MBTL3 effector consistently demonstrated superior repression across three additional target genes compared to KRAB alone (Figure 23C), ranging from 1.2-fold improvement with the highly effective CD151 targeting sgRNA to 8.0-fold for a weaker CD81 targeting sgRNA. Together, these results support the ability of this domain combination to overcome the limitations of weaker sgRNAs or lower effector expression to achieve robust target gene repression across diverse gene targets.

[0256] Beyond the general enhancement of CRISPRi activity in challenging settings, it was reasoned that the increased efficacy afforded by the KRAB + L3MBTL3 combination could be particularly enabling for bidirectional CRISPR perturbations, in which two orthogonal epigenetic effectors can be targeted to two distinct loci to achieve upregulation of one gene and simultaneous repression of another (Knoermann, S., et al 2015, Zalatan, J.G., et al 2015; Martella, A., et al, 2019). To assess this possibility, the stdMCP fusion strategy was used to recruit KRAB alone or KRAB + L3MBTL3 to an MS2-containing CD81 sgRNA or non-targeting control guide, while using an analogous setup to recruit the p65-HSFlACTIVEUS 211418665 98Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025CRISPRa effector (Liang, Y. et al, 2022) using the orthogonal stdPCP-PP7 system (Wu, B, et al. 2015; Zalatan, J.G., et al., 2015). Cell lines were engineered to express dCas9, stdMCP- KRAB or stdMCP-KRAB-L3MBTL3, and stdPCP-p65-HSFl in a doxycycline-inducible manner and introduced four dual MS2 / PP7 sgRNA combinations - each containing either a CD81 -targeting or a non-targeting MS2 sgRNA paired with a CD274-targeting or a nontargeting PP7sgRNA - using lentivirus.

[0257] Importantly, it was found that the KRAB + L3MBTL3 combination fully silenced CD81 expression to produce four distinct populations, unlike the KRAB-only condition (Figures 23D and 23E). This improved separation resulted from augmented repression by the KRAB + L3MBTL3 combination (KRAB: 41.5% vs KRAB + L3MBTL3: 89.4% CD81 repressed cells) without interfering with CD274 activation by the PP7 system (KRAB: 66.5% vs KRAB + L3MBTL3: 74.5% CD274 activated cells) (Figure 24). The KRAB + L3MBTL3 combination therefore enabled us to overcome limitations in the efficacy of other MS2-based repressors and develop a robust MS2 / PP7-based bidirectional perturbation system, which may be applied to future bidirectional CRISPR screening (Boettcher, M. et al, 2018; Pacalin, N.M., et al, 2024) or therapeutic applications.EXAMPLE 8: Long-term epigenetic memory requires modifiers of DNA methylation

[0258] Thus far, COMBINE has been shown to identify new individual effectors and combinations of effectors that transiently perturb gene expression, accurately measuring effector strength and uncovering unexpected synergistic and antagonistic interactions. Next, durable effects on target gene expression produced by transient expression of the epigenetic editing complex were investigated (Figure 25A). Following the initial 5-6 day effector recruitment phase, doxycycline was removed to halt any further production of dCas9 and the (std)MCP-fused effectors. Using the pilot studies as a reference, in which short-term repression by KRAB dissipated after twelve days (Figures 5B-D), cells were cultured for twelve days following doxycycline removal to the final screening timepoint. By Day 12, the majority of transcriptional effects had dissipated (Figures 1G and II).

[0259] Implementing the marginal score metric developed earlier allowed for quantitative determination of the magnitude of long-term changes to target gene expression induced by each individual effector across different combinations. Across both libraries, only three individual effectors induced significant long-term transcriptional changes with absolute marginal scores greater than 0.5, and all of these effectors participate in DNA methylation (Figures 25B and 26A). Two of the identified effectors, DNMT3A-3L17 and DNMT3A,ACTIVEUS 211418665 99Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 stably repressed target gene expression, while one effector, TET1, stably activated target gene expression. In contrast to DNMT3 A-3L and DNMT3 A, which are capable of de novo methylation of CpG sites (Gowher, H., and Jeltsch, A., 2001; Jurkowska, R.Z., et al, 2008; Liu, X.S., et al, 2016), TET1 oxidizes methylated cytosines to promote DNA demethylation (Liu, X.S., et al, 2016; Tahiliani, M., et al, 2009). Notably, both DNMTs have been shown to induce long-term silencing (Nunez, J.K, et al, 2021; Stepper, P., et al, 2017; O’Geen, H., et al, 2019), while TET1 has been used to reactivate disease-relevant or synthetically silenced genes in long-term (Nunez, J.K., et al, 2021; Liu, X.S., et al, 2018) and upregulate target gene expression in a transient manner (Liu, X.S., et al, 2016; Xu, X., et al, 2016; Choudhury, S.R., et al, 2016). Implementing the synergy score metric on Day 12screen data recovered the established synergistic interaction between the KRAB domain and DNMT3 A - the basis for CRISPRoff technology (Amabile, A., et al, 2016; Nunez, J.K, et al, 2021) - as the strongest hit in Library 1 (Figures 26B and 26C).

[0260] Each of these three domains were further investigated - including different pairings suggested by the screen data - with individual validation experiments. DNMT3 A-3L exhibited moderate but stable long-term repression both individually as well as in combinations identified to induce repressive memory in the screens (DNMT3A-3L, UHRF1 + DNMT3A-3L, PRDM4 + DNMT3A-3L, 23.4%, 19.7%, and 18.5% of cells repressed 12 days after doxycycline removal, respectively, Figures 25C and 27A-D). This repression was sustained for up to 50 days. A hypoactive mutant of UHRF1 and an inferred hypoactive mutant of PRDM4 were further tested by mutating a conserved tyrosine that is catalytically active in closely related PRDM9 (Jenkins, Y., et al. 2005; Blazer, L.L., et al, 2016). Incorporating these mutations led to a 1.4-fold and 2.4-fold reduction in repression on average over the 50-day time course (Figure 27E). The RING finger domain of UHRF1 is an E3 UBL that catalyzes ubiquitination of H3K18 and H3K23, which in turn recruits and stimulates DNMT1 to faithfully maintain DNA methylation (Liu, X., et al, 2013; Nishiyama, A, et al, 2013; Qin, W., et al, 2015; Li, T., et al, 2018). It follows that DNMT1 may aid DNMT3 A-3L in establishing methylation patterns that are heritable, for example by converting de novo hemimethylated CpGs into fully methylated CpG.

[0261] For DMNT3 A, the combination with UBE2E1 was selected for individual validation. Interestingly, it was observed that this combination led to minimal complete silencing (2.1% cells fully silenced at days) but did result in a 40.5% reduction in mean CD81 expression across all cells (measured as MFI) at 12 days, which lasted up to 50 days (FiguresACTIVEUS 211418665 100Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202525D and 28A-28C). Notably, this effect was reduced by 2.3-fold when UBE2E1 was mutated at its catalytic cysteine and completely abolished when DNMT3 A was catalytically inactivated (Figure 28D) (Sun, L., et al, 2016; Dukatz, M., et al, 2019). These results indicate that DNMT3 A is the primary mediator of this repressive epigenetic memory, but that UBE2E1 augments DNMT3A’s activity, at least partially through its catalytic function as an E2 UBL. The ability to stably reduce gene expression without silencing will enable future functional genomics studies interrogating gene dosage as well as offer therapeutic downregulation of genes in situations where maintaining low levels of target gene expression is required.

[0262] In addition to long-term repression, long-term epigenetic activation catalyzed by TET1 combinations was observed. While TET1 has been previously used to reactivate the expression of target genes silenced by DNA methylation in synthetic or disease contexts (Nunez, J.K., et al, 2021; Liu, X.S., et al, 2018), marginal effector score analysis suggests that TET1 can also increase the expression of a basally expressed target gene not only transiently (Liu, X.S., et al, 2016; Xu, X., et al, 2016; Choudhury, S.R., et al, 2016)but also in a longterm manner. For individual validation, the three TET1 partners with the highest activation from the screen were selected - RNF20’s RING finger domain, SMYDl’s SET domain, and PRDMl’s PR / SET domain. All three TET1 combinations led to an increase in CD81 expression that lasted up to 50 days and reached 40.3%, 28.7%, and 35.6% respectively at 6 days post-doxy cy cline washout (Figures 25E and 29). Activation by TET1 alone, on the other hand, only lasted up to 21 days and reached 19.5% at day 6. Mutation of the known catalytic residues of TET1 not only abrogated the activation phenotype (Tahiliani,et al, 2009) but surprisingly led to stable gene repression, while catalytic mutations in the TET1 partners had a neutral or positive effect on long-term activation compared to their unmutated counterparts (Figure 30). These results indicate that other as-yet unknown factors, such as for example the recruitment of endogenous machinery, can augment long-term activation induced by TET1.

[0263] In summary, three distinct modes of epigenetic memory mediated by modulators of DNA methylation were delineated: silencing, reduction in gene expression, and activation (Figure 25F). The ability to tune the degree of long-term repression on endogenous genes and establish long-term activation of a basally expressed gene on a single cell level were further demonstrated. Future work will delineate the mechanistic differences between the precise interactions of effector pairs and their resulting modes of epigenetic memory.ACTIVEUS 211418665 101Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1)Date of Electronic Filing: September 18, 2025EXAMPLE 9: Discussion

[0264] In this work, a new platform technology called COMBINE was developed for systematic, high throughput functional screening of domain pairs. Importantly, COMBINE overcomes the length limitations and length-based biases of previous screening methodologies (Mukund, A.X, et al, 2023; Blaeschke, F., et al, 2023; Gordon, K.S, et al, 2022) to enable the characterization of large protein domains and beyond, which was demonstrated by testing combinatorial domain candidates up to 2094 amino acids in length.

[0265] Applying COMBINE to epigenetic regulation of gene expression, over 50,000 pairs of epigenetic reader, recruiter, writer, and eraser domains across two libraries were tested to understand emergent properties of combinatorial domain interactions that illuminate new biology and provide a rich ground for future epigenome editing tool development. In particular, this systematic analysis captured broad, class-level trends of epigenetic effector activity, uncovered new and unexpected synergistic or antagonistic domain interactions, and allowed for the quantitative understanding of the contribution of individual domains and domain pairs to the modulation of target gene expression. Importantly, the screen results established herein are quantitative and generalizable across 5 endogenous genes through extensive secondary validation. New domain combinations that may prove useful for nextgeneration epigenome editing tools are highlighted, including the pairing of the ZNF10 KRAB domain with the SAM domain from L3MBTL3 which improved target gene silencing by up to 34-fold in challenging, dose limiting conditions. This robust activity enabled the substantial improvement of an aptamer based bidirectional CRISPR perturbation setup (Zalatan, J.G., et al, 2015; Martella, A., e al, 2019), in which one target gene is silenced while another is activated using an orthogonal guide RNA-mediated effector recruitment strategy.

[0266] The inducible nature of the COMBINE platform allowed for interrogating the durable effects of domain pairs on target gene transcription following inducer washout. From this long-term analysis, DNA methylation was nominated as the key modification driving heritable epigenetic regulation. Expanding beyond previously observed epigenetic silencing memory (Amabile, A. et al, 2016; Nunez, J.K., et al, 2021), domain combinations that elicited long-term graded target gene repression of approximately 40%, or long-term gene activation of 30-40% were observed- both of which were sustained for at least 50 days. Importantly, the latter example establishes the ability of TET1 pairs to initiate heritable gene activation from baseline expression levels and emphasizes the utility of the combinatorial screening approach to identify unexpected pairings that greatly enhance the potency andACTIVEUS 211418665 102Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 durability of this effect. It is anticipated that these new epigenetic memory effectors will expand the toolbox of long-term expression modulation technologies, complementary to existing approaches such as CRISPRoff (Nunez, J.K, et al, 2021).

[0267] This rich dataset quantifying the transcriptional output induced by over 50,000 epigenetic effector pairs provides a broad resource for the epigenetics community to explore new mechanistic theories and pursue new biotechnological tools. Further, the COMBINE platform is generalizable to many cell types due to its lentiviral delivery and is applicable beyond epigenetics to study many outstanding questions in combinatorial biology. The ability to systematically evaluate the functional interplay between entities-whether that includes DNA elements, RNA species, protein domains, or entire proteins-will unlock a sophisticated understanding of interactions in biology, where nothing acts in isolation.

[0268] References1. Goldberg, A.D., Allis, C.D., and Bernstein, E. (2007). Epigenetics: a landscape takes shape. Cell 128, 635-638.2. Vogelauer, M., Rubbi, L., Lucas, E, Brewer, B.J., and Grunstein, M. (2002). Histone acetylation regulates the time of replication origin firing. Mol. Cell 10, 1223-1233.3. Hauer, M.H., and Gasser, S.M. (2017). Chromatin and nucleosome dynamics in DNA damage and repair. Genes Dev. 31, 2204-2221.4. Zhang, T., Cooper, S., and Brockdorff, N. (2015). The interplay of histone modifications - writers that read. EMBO Rep. 16, 1467-1481 - 1481.5. Policarpi, C., Munafo, M., Tsagkris, S., Carlini, V., and Hackett, J. A. (2024). Systematic epigenome editing captures the context-dependent instructive function of chromatin modifications. Nat. Genet. 56, 1168-1180.6. Lukauskas, S., Tvardovskiy, A., Nguyen, N.V., Stadler, M., Faull, P., Ravnsborg, T., Ozdemir Aygenli, B., Dornauer, S., Flynn, H., Lindeboom, R.G.H., et al. (2024). Decoding chromatin states by proteomic profiling of nucleosome readers. Nature 627, 671-679.7. Kuzmichev, A., Nishioka, K., Erdjument-Bromage, H., Tempst, P., and Reinberg, D. (2002). Histone methyltransferase activity associated with a human multiprotein complex containing the Enhancer of Zeste protein. Genes Dev. 16, 2893-2905.8. Thakore, P.I., Black, J.B., Hilton, I.B., and Gersbach, C.A. (2016). Editing the epigenome: technologies for programmable transcription and epigenetic modulation. Nat. Methods 13, 127-137.ACTIVEUS 211418665 103Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 20259. Nakamura, M., Gao, Y., Dominguez, A. A., and Qi, L.S. (2021). CRISPR technologies for precise epigenome editing. Nat. Cell Biol. 23, 11-22.10. Konermann, S., Brigham, M.D., Trevino, A.E., Joung, J., Abudayyeh, O.O., Barcena, C., Hsu, P.D., Habib, N., Gootenberg, J.S., Nishimasu, H., et al. (2015). Genome-scale transcriptional activation by an engineered CRISPR-Cas9 complex. Nature 517, 583-588. 83.11. Amabile, A., Migliara, A., Capasso, P., Biffi, M., Cittaro, D., Naldini, L., 1970 and Lombardo, A. (2016). Inheritable Silencing of Endogenous Genes by Hit-and-Run Targeted Epigenetic Editing. Cell 167, 219-232. el4.12. Nunez, J.K., Chen, J., Pommier, G.C., Cogan, J.Z., Replogle, J.M., Adriaens, C., Ramadoss, G.N., Shi, Q., Hung, K.L., Samelson, A.J., et al. (2021). Genome-wide programmable transcriptional memory by CRISPR-based epigenome editing. Cell 184, 2503-2519. el7.13. Cappelluti, M.A., Mollica Poeta, V., Valsoni, S., Quarato, P., Merlin, S., Merelli, I., and Lombardo, A. (2024). Durable and efficient gene silencing in vivo by hit-and-run epigenome editing. Nature 627, 416-423.14. Konermann, S., Brigham, M.D., Trevino, A., Hsu, P.D., Heidenreich, M., Cong, L., Platt,R.J., Scott, D.A., Church, G.M., and Zhang, F. (2013). Optical control of mammalian endogenous transcription and epigenetic states. Nature 500, 472-476.15. Chavez, A., Scheiman, J., Vora, S., Pruitt, B.W., Tuttle, M., P R Iyer, E., Lin, S., Kiani,S., Guzman, C.D., Wiegand, D.J., et al. (2015). Highly efficient Cas9-mediated transcriptional programming. Nat. Methods 12, 326-328.16. Bintu, L., Yong, J., Antebi, Y.E., McCue, K., Kazuki, Y., Uno, N., Oshimura, M., and Elowitz, M.B. (2016). Dynamics of epigenetic regulation at the single-cell level. Science 351, 720-724.17. Stepper, P., Kungulovski, G., Jurkowska, R.Z., Chandra, T., Krueger, F., Reinhardt, R., Reik, W., Jeltsch, A., and Jurkowski, T.P. (2017). Efficient targeted DNA methylation with chimeric dCas9-Dnmt3a-Dnmt3L methyltransferase. Nucleic Acids Res. 45, 1703- 1713.18. Yeo, N.C., Chavez, A., Lance-Byrne, A., Chan, Y., Menn, D., Milanova, D., Kuo, C.-C., Guo, X., Sharma, S., Tung, A., et al. (2018). An enhanced CRISPR repressor for targeted mammalian gene regulation. Nat. Methods 15, 611-616.19. O’Geen, H., Bates, S.L., Carter, S.S., Nisson, K.A., Halmai, J., Fink, K.D., Rhie, S.K.,ACTIVEUS 211418665 104Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Farnham, P.J., and Segal, D.J. (2019). Ezh2-dCas9 and KRAB-dCas9 enable engineering of epigenetic memory in a context-dependent manner. Epigenetics Chromatin 12, 26.20. Alerasool, N., Segal, D., Lee, H., and Taipale, M. (2020). An efficient KRAB domain for CRISPRi applications in human cells. Nat. Methods 17, 1093-1096.21. Van, M.V., Fujimori, T., and Bintu, L. (2021). Nanobody-mediated control of gene expression and epigenetic memory. Nat. Commun. 12, 537.22. Qian, J., Guan, X., Xie, B., Xu, C., Niu, J., Tang, X., Li, C.H., Colecraft, H.M., Jaenisch, R., and Liu, X.S. (2023). Multiplex epigenome editing of MECP2 to rescue Rett syndrome neurons. Sci. Transl. Med. 15, eadd4666.23. Neumann, E.N., Bertozzi, T.M., Wu, E., Serack, F., Harvey, J.W., Brauer, P.P., Pirtle, C.P., Coffey, A., Howard, M., Kamath, N., et al. (2024). Brainwide silencing of prion protein by AAV-mediated delivery of an engineered compact epigenetic editor. Science 384, ado7082.24. Tycko, J., DelRosso, N., Hess, G.T., Aradhana, Baneijee, A., Mukund, A., Van, M.V., Ego, Yao, D., Spees, K., et al. (2020). High-Throughput Discovery and 2009 Characterization of Human Transcriptional Effectors. Cell 183, 2020-2035. el 6.25. Alerasool, N., Leng, H., Lin, Z.-Y., Gingras, A.-C., and Taipale, M. (2022). Identification and functional characterization of transcriptional activators in human cells. Mol. Cell 82, 677- 695. e7.26. DelRosso, N., Tycko, J., Suzuki, P., Andrews, C., Aradhana, Mukund, A., Liongson, E, Ludwig, C., Spees, K., Fordyce, P., et al. (2023). Large-scale mapping and mutagenesis of human transcriptional effector domains. Nature 616, 365-372.27. Mukund, A.X., Tycko, J., Allen, S.J., Robinson, S.A., Andrews, C., Sinha, J., Ludwig, C.H., Spees, K., Bassik, M.C., and Bintu, L. (2023). High-throughput functional characterization of combinations of transcriptional activators and repressors. Cell Sy st 14, 746-763. e5.28. Blaeschke, F., Chen, Y.Y., Apathy, R., Daniel, B., Chen, A.Y., Chen, P.A., Sandor, K., Zhang, W., Li, Z., Mowery, C.T., et al. (2023). Modular pooled discovery of synthetic knockin sequences to program durable cell therapies. Cell 186, 4216-4234. e33.29. Sweeney, N.P., and Vink, C.A. (2021). The impact of lentiviral vector genome size and producer cell genomic to gag-pol mRNA ratios on packaging efficiency and titre. Mol Ther Methods Clin Dev 21, 574-584.30. Gohl, D.M., Magli, A., Garbe, J., Becker, A., Johnson, D.M., Anderson, S., Auch, B.,ACTIVEUS 211418665 105Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025Billstein, B., Froehling, E., McDevitt, S.L., et al. (2019). Measuring sequencer size bias using REcount: a novel method for highly accurate Illumina sequencing-based quantification. Genome Biol. 20, 85.31. Peabody, D.S. (1993). The RNA binding site of bacteriophage MS2 coat protein. EMBO J. 12, 595-600.32. Heidersbach, A.J., Dorighi, K.M., Gomez, J. A., Jacobi, A.M., and Haley, B. (2023). A versatile, high-efficiency platform for CRISPR-based gene activation. Nat. Commun. 14, 902.33. Wu, B., Miskolci, V., Sato, H., Tutucci, E., Kenworthy, C.A., Donnelly, S.K., Yoon, Y.J., Cox, D., Singer, R.H., and Hodgson, L. (2015). Synonymous modification results in high- fidelity gene expression of repetitive protein and nucleotide sequences. Genes Dev. 29, 876-886.34. Head, S.R., Komori, H.K., LaMere, S.A., Whisenant, T., Van Nieuwerburgh, F., Salomon, D.R., and Ordoukhanian, P. (2014). Library construction for next-generation sequencing: overviews and challenges. Biotechniques 56, 61-64, 66, 68, passim.35. Gordon, K.S., Kyung, T., Perez, C.R., Holec, P.V., Ramos, A., Zhang, A.Q., Agarwal, Y., Liu, Y., Koch, C.E., Starchenko, A., et al. (2022). Screening for CD19-specific chimaeric antigen receptors with enhanced signalling via a barcoded library of intracellular domains. Nat Biomed Eng 6, 855-866.36. Zhao, J., Wang, M., Chang, L., Yu, J., Song, A., Liu, C., Huang, W., Zhang, T., Wu, X., Shen, X., et al. (2020). RYBP / YAF2-PRC1 complexes and histone Hl-dependent chromatin compaction mediate propagation of H2AK119ubl during cell division. Nat. Cell Biol. 22, 439-452.37. Ogryzko, V.V., Schiltz, R.L., Russanova, V., Howard, B.H., and Nakatani, 2049 Y. (1996). The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87, 953-959.38. Lahn, B.T., Tang, Z.L., Zhou, J., Bamdt, R.J., Parvinen, M., Allis, C.D., and Page, D.C. (2002). Previously uncharacterized histone acetyltransferases implicated in mammalian spermatogenesis. Proc. Natl. Acad. Sci. U. S. A. 99, 8707-8712.39. Champagne, N., Pelletier, N., and Yang, X.J. (2001). The monocytic leukemia zinc finger protein MOZ is a histone acetyltransferase. Oncogene 20, 404-409.40. Guelman, S., Kozuka, K., Mao, Y., Pham, V., Solloway, M.J., Wang, J., Wu, J., Lili, J.R., and Zha, J. (2009). The double-histone-acetyltransferase complex ATAC is essential forACTIVEUS 211418665 106Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 mammalian development. Mol. Cell. Biol. 29, 1176-1188.41. Wei, W., Liu, X., Chen, J., Gao, S., Lu, L., Zhang, H., Ding, G., Wang, Z., Chen, Z., Shi, T., et al. (2017). Class I histone deacetylases are major histone decrotonylases: evidence for critical and broad function of histone crotonylation in transcription. Cell Res. 27, 898- 915.42. Gilbert, L.A., Larson, M.H., Morsut, L., Liu, Z., Brar, G.A., Torres, S.E., Stern-Ginossar, N., Brandman, O., Whitehead, E.H., Doudna, J.A., et al. (2013). CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell 154, 442-451.43. Cong, L., Zhou, R., Kuo, Y.-C., Cunniff, M., and Zhang, F. (2012). Comprehensive interrogation of natural TALE DNA-binding modules and transcriptional repressor domains. Nat. Commun. 3, 968.44. Knight, M.J., Leettola, C., Gingery, M., Li, EL, and Bowie, J.U. (2011). A human sterile alpha motif domain polymerizome. Protein Sci. 20, 1697-1706. Xu, T., Park, S.-S., Giaimo, B.D., Hall, D., Ferrante, F., Ho, D.M., Hori, K., Anhezini, L., Ertl,45. E, Bartkuhn, M., et al. (2017). RBPJ / CBF1 interacts with L3MBTL3 / MBT1 to promote repression of Notch signaling via histone demethylase KDM1A / LSD1. EMBO J. 36, 3232-3249.46. Schuetz, A., Min, J., Allali-Hassani, A., Schapira, M., Shuen, M., Loppnau, P., Mazitschek,mR., Kwiatkowski, N.P., Lewis, T.A., Maglathin, R.L., et al. (2008). Human HDAC7 harbors a class Ila histone deacetylase-specific zinc binding motif and cryptic deacetylase activity. J. Biol. Chem. 283, 11355-11363.47. Wang, A.H., Bertos, N.R., Vezmar, M., Pelletier, N., Crosato, M., Heng, H.H., Th’ng, J., Han, J., and Yang, X.J. (1999). HDAC4, a human histone deacetylase related to yeast HDA1, is a transcriptional corepressor. Mol. Cell. Biol. 19, 7816-7827.48. Grazini, U., Zanardi, F., Citterio, E., Casola, S., Goding, C.R., and McBlane, F. (2010). The RING domain of RAG1 ubiquitylates histone H3: a novel activity in chromatin- mediated regulation of V(D)J joining. Mol. Cell 37, 282-293.49. Deng, Z., Liu, H., and Liu, X. (2015). RAGl-mediated ubiquitylation of histone H3 is required for chromosomal V(D)J recombination. Cell Res. 25, 181-192.50. Wheaton, K., Sarkari, F., Stanly Johns, B., Davarinejad, H., Egorova, O., Kaustov, L., Raught, B., Saridakis, V., and Sheng, Y. (2017). UbE2El / UBCH6 Is a Critical 2088 in Vivo E2 for the PRC 1 -catalyzed Ubiquitination of H2A at Lys-119. J. Biol. Chem. 292, 2893-2902.ACTIVEUS 211418665 107Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202551. Weirich, S., Kudithipudi, S., and Jeltsch, A. (2016). Specificity of the SUV4-20H1 and SUV4-20H2 protein lysine methyltransferases and methylation of novel substrates. J. Mol. Biol. 428, 2344-2358.52. Yurchenko, V., Xue, Z., and Sadofsky, M. (2003). The RAG1 N-terminal domain is an E3 ubiquitin ligase. Genes Dev. 17, 581-585.53. Sun, L., and Fang, J. (2016). E3 -Independent Constitutive Monoubiquitination Complements Histone Methyltransferase Activity of SETDB1. Mol. Cell 62, 958-966.54. Wu, H., Siarheyeva, A., Zeng, H., Lam, R., Dong, A., Wu, X.-H., Li, Y., Schapira, M., Vedadi, M., and Min, J. (2013). Crystal structures of the human histone H4K20 methyltransferases SUV420H1 and SUV420H2. FEBS Lett. 587, 3859-3868.55. Kang, T.-H., Park, D.-Y., Choi, Y.H., Kim, K.-I, Yoon, H.S., and Kim, K.-T. (2007). Mitotic histone H3 phosphorylation by vaccinia-related kinase 1 in mammalian cells. Mol. Cell. Biol. 27, 8533-8546.56. Kent, W.J., Sugnet, C.W., Furey, T.S., Roskin, K.M., Pringle, T.H., Zahler, A.M., and Haussler, D. (2002). The human genome browser at UCSC. Genome Res. 12, 996-1006.57. Conway, E., Jerman, E., Healy, E., Ito, S., Holoch, D., Oliviero, G., Deevy, O., Glancy, E., Fitzpatrick, D.J., Mucha, M., et al. (2018). A Family of Vertebrate-Specific Polycombs Encoded by the LCOR / LCORL Genes Balance PRC2 Subtype Activities. Mol. Cell 70, 408-42 Le8.58. Zhang, Q., Agius, S.C., Flanigan, S.F., Uckelmann, M., Levina, V., Owen, B.M., and Davidovich, C. (2021). PALI1 facilitates DNA and nucleosome binding by PRC2 and triggers an allosteric activation of catalysis. Nat. Commun. 12, 4592.59. Fong, K.-W., Zhao, J.C., Lu, X., Kim, J., Piunti, A., Shilatifard, A., and Yu, J. (2022). PALI1 promotes tumor growth through competitive recruitment of PRC2 to G9A-target chromatin for dual epigenetic silencing. Mol. Cell 82, 4611-4626. e7.60. Janzer, A., Stamm, K., Becker, A., Zimmer, A., Buettner, R., and Kirfel, J. (2012). The H3K4me3 histone demethylase FbxllO is a regulator of chemokine expression, cellular morphology, and the metabolome of fibroblasts. J. Biol. Chem. 287, 30984-30992.61. He, J., Kallin, E.M., Tsukada, Y.-E, and Zhang, Y. (2008). The H3K36 demethylase Jhdmlb / Kdm2b regulates cell proliferation and senescence through pl5(Ink4b). Nat. Struct. Mol. Biol. 15, 1169-1175.62. Kang, J.-Y., Kim, J.-Y., Kim, K.-B., Park, J.W., Cho, H., Hahm, J.Y., Chae, Y.-C., Kim, D., Kook, H., Rhee, S., et al. (2018). KDM2B is a histone H3K79 demethylase andACTIVEUS 211418665 108Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025 induces transcriptional repression via sirtuin-1 -mediated chromatin silencing. FASEB J.32, 5737-63. 5750. Yamane, K., Tateishi, K., Klose, R.J., Fang, J., Fabrizio, L.A., Erdjument- Bromage, H., Taylor-Papadimitriou, J., Tempst, P., and Zhang, Y. (2007). PLU-1 is an H3K4 demethylase involved in transcriptional repression and breast cancer cell proliferation. 2127 Mol. Cell 25, 801-812.64. Biggar, K.K., Charih, F., Liu, H., Ruiz-Blanco, Y.B., Stalker, L., Chopra, A., Connolly, J., Adhikary, H., Frensemier, K., Hoekstra, M., et al. (2020). Proteome-wide Prediction of Lysine Methylation Leads to Identification of H2BK43 Methylation and Outlines the Potential Methyllysine Proteome. Cell Rep. 32, 107896.65. Zalatan, J.G., Lee, M.E., Almeida, R., Gilbert, L.A., Whitehead, E.H., La Russa, M., Tsai, J.C., Weissman, J.S., Dueber, J.E., Qi, L.S., et al. (2015). Engineering complex synthetic transcriptional programs with CRISPR RNA scaffolds. Cell 160, 339-350.66. Martella, A., Firth, M., Taylor, B.J.M., Goppert, A., Cuomo, E.M., Roth, R.G., Dickson, A. J., and Fisher, D.I. (2019). Systematic Evaluation of CRISPRa and CRISPRi Modalities Enables Development of a Multiplexed, Orthogonal Gene Activation and Repression System. ACS Synth. Biol. 8, 1998-2006.67. Liang, Y., Xu, H., Cheng, T., Fu, Y., Huang, H., Qian, W., Wang, J., Zhou, Y., Qian, P., Yin, Y., et al. (2022). Gene activation guided by nascent RNA-bound transcription factors. Nat. Commun. 13, 7329.68. Boetcher, M., Tian, R., Blau, J. A., Markegard, E., Wagner, R.T., Wu, D., Mo, X., Biton, A., Zaitlen, N., Fu, H., et al. (2018). Dual gene activation and knockout screen reveals directional dependencies in genetic networks. Nat. Biotechnol. 36, 170-178.69. Pacalin, N.M., Steinhart, Z., Shi, Q., Belk, J. A., Dorovskyi, D., Kraft, K., Parker, K.R., Shy, B.R., Marson, A., and Chang, H.Y. (2024). Bidirectional epigenetic editing reveals hierarchies in gene regulation. Nat. Biotechnol. 10.1038 / s41587-024-02213-3.70. Gowher, H., and Jeltsch, A. (2001). Enzymatic properties of recombinant Dnmt3a DNA methyltransferase from mouse: the enzyme modifies DNA in a non-processive manner and also methylates non-CpA sitesl lEdited by J. Karn. J. Mol. Biol. 309, 1201-1208.71. Jurkowska, R.Z., Anspach, N., Urbanke, C., Jia, D., Reinhardt, R., Nellen, W., Cheng, X., and Jeltsch, A. (2008). Formation of nucleoprotein filaments by mammalian DNA methyltransferase Dnmt3a in complex with regulator Dnmt3L. Nucleic Acids Res. 36, 6656-6663.ACTIVEUS 211418665 109Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202572. Liu, X.S., Wu, H., Ji, X., Stelzer, Y., Wu, X., Czauderna, S., Shu, J., Dadon, D., Young,R.A., and Jaenisch, R. (2016). Editing DNA Methylation in the Mammalian Genome. Cell 167, 233-247. el7.73. Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y., Agarwal,S., Iyer, L.M., Liu, D.R., Aravind, L., et al. (2009). Conversion of 5-methylcytosine to 5- hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324, SO- SS.74. Liu, X.S., Wu, H., Krzisch, M., Wu, X., Graef, J., Muffat, J., Hnisz, D., Li, C.H., Yuan,B., Xu, C., et al. (2018). Rescue of Fragile X Syndrome Neurons by DNA Methylation Editing of the FMRI Gene. Cell 172, 979-992.e6.75. Xu, X., Tao, Y., Gao, X., Zhang, L., Li, X., Zou, W., Ruan, K., Wang, F., Xu, G.-L., and Hu, R. (2016). A CRISPR-based approach for targeted DNA demethylation. Cell Discov 2,16009.76. Choudhury, S.R., Cui, Y., Lubecka, K., Stefanska, B., and Irudayaraj, J. (2016). CRISPR2169 dCas9 mediated TET1 targeting for selective DNA demethylation at BRCA1 promoter. Oncotarget 7, 46545-46556.77. Jenkins, Y., Markovtsov, V., Lang, W., Sharma, P., Pearsall, D., Warner, J., Franci, C., Huang, B., Huang, J., Yam, G.C., et al. (2005). Critical role of the ubiquitin ligase activity of UHRF1, a nuclear RING finger protein, in tumor cell growth. Mol. Biol. Cell 16, 5621-5629.78. Blazer, L.L., Lima-Fernandes, E., Gibson, E., Eram, M.S., Loppnau, P., Arrowsmith,C.H., Schapira, M., and Vedadi, M. (2016). PR Domain-containing Protein 7 (PRDM7) Is a Histone 3 Lysine 4 Trimethyltransferase. J. Biol. Chem. 291, 13509-13519.79. Liu, X., Gao, Q., Li, P., Zhao, Q., Zhang, J., Li, J., Koseki, H., and Wong, J. (2013). UHRF1 targets DNMT1 for DNA methylation through cooperative binding of hemimethylated DNA and methylate...

Claims

Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 2025CLAIMSWhat is claimed is:

1. A nucleic acid comprising a sequence encoding an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1 or comprising at least two epigenetic effectors of Table 2.

2. The nucleic acid of claim 1, wherein the first epigenetic effector comprises a Kruppel associated box (KRAB) domain.

3. The nucleic acid of claims 1-2, wherein the second epigenetic effector comprises L3MBTL3 sterile alpha motif (SAM) domain.

4. The nucleic acid of claims 2-3, wherein the epigenetic effector fusion comprises a KRAB domain of ZNF10 and a SAM domain of L3MBTL3, a KRAB domain of ZNF705 and a SAM domain of L3MBTL3, or a KRAB domain of ZIM3 and a SAM domain of L3MBTL3.

5. The nucleic acid of claim 1, wherein the first epigenetic effector comprises p65.

6. The nucleic acid of claims 1 or 5, wherein the second epigenetic effector comprisesHSF1.

7. The nucleic acid of claim 1, wherein the first epigenetic effector comprises TET1.

8. The nucleic acid of claim 7, wherein the second epigenetic effector comprises theRING finger domain of RNF20, the SET domain of SMYD1, or the PR / SET domain ofPRDMl.

9. The nucleic acid of claims 1-8, wherein the first and the second epigenetic effectors are connected by a linker.

10. The nucleic acid of claims 1-9, wherein the epigenetic effector fusion further comprises at least one RNA binding protein.

11. The nucleic acid of claim 10, wherein the RNA binding protein comprises MS2 bacteriophage coat protein (MCP) or PP7 bacteriophage coat protein (PCP).ACTIVEUS 211418665 127Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202512. The nucleic acid of claims 1-11, wherein the epigenetic effector fusion further comprises a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain.

13. The nucleic acid of claim 12, wherein the Cas protein comprises a Cas9, Cpfl, Cas 12b, Cas 12c, Cas 12d, Casl2e, Casl2f, Casl2h, Casl2i, or Cas 12g that lacks nuclease and / or nickase activity.

14. The nucleic acid of claim 13, wherein the Cas protein comprises spdCas9.

15. The nucleic acid of claims 1-14, further comprising an inducible promoter.

16. A vector comprising any of the nucleic acids of claims 1-15.

17. A host cell comprising the nucleic acid of claims 1-15 or the vector of claim 16, wherein the host cell expresses the epigenetic effector fusion.

18. The host cell of claim 17, wherein the transcription of a target DNA locus is reduced by about 5% to about 95% relative to the transcription of the target DNA locus before expression of the nucleic acid.

19. The host cell of claim 17, wherein the transcription of a target DNA locus is increased by about 5% to about 95% relative to the transcription of the target DNA locus before expression of the nucleic acid.

20. The host cell of claim 18, wherein the first epigenetic effector comprises TET1 and the transcription of a target DNA locus is decreased for at least 5 days.

21. The host cell of claim 18, wherein the first epigenetic effector comprises TET1 and the transcription of a target DNA locus is decreased for at least 60 days.

22. The host cell of claims 20-21, wherein the epigenetic effector fusion is expressed in the host cell transiently.

23. An epigenetic editing system comprising a nucleic acid of any of claims 1-15.ACTIVEUS 211418665 128Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202524. The epigenetic editing system comprising a nucleic acid of claims 10 or 11 and a nucleic acid encoding a gRNA comprising an RNA aptamer sequence that binds to the RNA binding protein.

25. The epigenetic editing system of claim 24, wherein the RNA binding protein is MCP and the RNA aptamer sequence comprises MS2.

26. The epigenetic editing system of claim 24, wherein the RNA binding protein is PCP and the RNA aptamer sequence comprises PP7.

27. The epigenetic editing system of claims 24-26, wherein gRNA comprises a spacer sequence that targets the gRNA to a gene of interest.

28. The epigenetic editing system of claims 24-27, further comprising a nucleic acid encoding a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g lacks nuclease and / or nickase activity that binds to the gRNA.

29. The epigenetic editing system of claim 28, further comprising a nucleic acid encoding a spdCas9.

30. A host cell comprising the epigenetic editing system of claims 23-29, wherein the host cell expresses the epigenetic editing system.

31. A host cell comprising the epigenetic editing system of claims 23-27, wherein the host cell expresses spdCas9.

32. A method of regulating gene expression, the method comprising introducing into a cell: the epigenetic modulation system of claims 23-29.

33. The method of claim 32, wherein the epigenetic modulation system reduces or inhibits the transcription of a target DNA locus.

34. A nucleic acid comprising a sequence encoding a first epigenetic effector that reduces expression of a first target DNA locus and a first RNA binding protein and a second epigenetic effector that increases expression of a second target DNA locus and second RNA binding protein.ACTIVEUS 211418665 129Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202535. The nucleic acid of claim 34, wherein the first epigenetic effector comprises KRAB domain.

36. The nucleic acid of claims 34-35, wherein the second epigenetic effector comprises HSF1.

37. The nucleic acid of claims 34-36, wherein the first epigenetic effector comprises an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1.

38. The nucleic acid of claim 37, wherein the first epigenetic effector comprises L3MBTL3 sterile alpha motif (SAM) domain.

39. The nucleic acid of claims 34-38, wherein the second epigenetic effector comprises an epigenetic effector fusion comprising at least two epigenetic effectors of Table 1.

40. The nucleic acid of claim 39, wherein the second epigenetic effector comprises p65.

41. The nucleic acid of claims 37-40, wherein the epigenetic effectors of the first epigenetic fusion and / or second epigenetic fusion are connected by a linker.

42. The nucleic acid of claims 34-41, wherein the first RNA binding protein comprises MS2 bacteriophage coat protein (MCP) and the second PP7 bacteriophage coat protein (PCP), or vice-versa.

43. The nucleic acid of claims 34-42, further comprising an inducible promoter.

44. A nucleic acid comprising a sequence encoding a first epigenetic effector fusion that reduces expression of a first target DNA locus fused to a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain, and a second epigenetic effector fusion that increases expression of a second target DNA locus fused to a Cas protein, a zinc finger DNA binding domain, or a TALE DNA binding domain.

45. The nucleic acid of claim 44, wherein the first epigenetic effector fusion comprises at least two epigenetic effectors of Table 1 and the second epigenetic effector fusion comprises at least two epigenetic effectors of Table 1.

46. The nucleic acid of claim 45, wherein the first epigenetic effector fusion comprises a KRAB domain.ACTIVEUS 211418665 130Atorney Docket No.: 2220476.0013 OWOl (Arc-0019-WO1) Date of Electronic Filing: September 18, 202547. The nucleic acid of claims 44-46, wherein the second epigenetic effector comprises HSF1.

48. The nucleic acid of claims 44-47, wherein the first epigenetic effector fusion comprises L3MBTL3 SAM domain.

49. The nucleic acid of claims 47-48, wherein the second epigenetic effector fusion comprises p65.

50. The nucleic acid of claims 44-49, wherein the epigenetic effectors of the first epigenetic fusion and / or second epigenetic fusion are connected by a linker.

51. The nucleic acid of claims 44-50, wherein the first and second epigenetic fusions are fused to a Cas protein.

52. The nucleic acid of claim 51, wherein the Cas protein comprises a Cas9, Cpfl, Cas 12b, Cas 12c, Cas 12d, Casl2e, Casl2f, Casl2h, Casl2i, or Cas 12g that lacks nuclease and / or nickase activity or a Cas9, Cpfl, Casl2b, Casl2c, Casl2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g with nuclease and / or nickase activity.

53. The nucleic acid of claim 52, wherein the Cas protein comprises spdCas9.

54. The nucleic acid of claims 34-43, wherein the first epigenetic effector and second epigenetic effector are on separate nucleic acid molecules.

55. The nucleic acid of claims 44-53, wherein the first epigenetic effector fusion and second epigenetic effector fusion are on separate nucleic acid molecules.

56. A vector comprising any of the nucleic acids of claims 34-55.

57. A host cell comprising the nucleic acid of claims 34-55 or the vector of claim 56, wherein the host cell expresses the epigenetic effectors or epigenetic effector fusions.

58. The host cell of claim 57, wherein expression of the first target DNA locus is reduced by about 5% to about 95% relative to the expression of the first target DNA locus before expression of the nucleic acid and wherein expression of the second target DNA locus is increased by about 5% to about 95% relative to the expression of the second target DNA locus before expression of the nucleic acid.ACTIVEUS 211418665 131Attorney Docket No.: 2220476.00130WO1 (Arc-0019-WO1) Date of Electronic Filing: September 18, 202559. An epigenetic editing system comprising a nucleic acid of any of claims 34-55.

60. A epigenetic editing system comprising a nucleic acid of claim 34-42, a nucleic acid encoding a first gRNA comprising an RNA aptamer sequence that binds to the first RNA binding protein and a nucleic acid encoding a second gRNA comprising an RNA aptamer sequence that binds to the second RNA binding protein.

61. The epigenetic editing system of claim 60, wherein the first RNA binding protein is MCP and the RNA aptamer sequence comprises MS2 and wherein the second RNA binding protein is PCP and the RNA aptamer sequence comprises PP7.

62. The epigenetic editing system of claim 60, wherein the second RNA binding protein is MCP and the RNA aptamer sequence comprises MS2 and wherein the first RNA binding protein is PCP and the RNA aptamer sequence comprises PP7.

63. The epigenetic editing system of claims 60-62, wherein the first gRNA comprises a spacer sequence that targets the gRNA to a first gene of interest and the second gRNA comprises a spacer sequence that targets the gRNA to a second gene of interest.

64. The epigenetic editing system of claims 60-63, further comprising a nucleic acid encoding a Cas9, Cpfl, Cast 2b, Cast 2c, Cast 2d, Casl2e, Casl2f, Casl2h, Casl2i, or Casl2g lacks nuclease and / or nickase activity that binds to the gRNA.

65. The epigenetic editing system of claim 64, further comprising a nucleic acid encoding a spdCas9.

66. A host cell comprising the epigenetic editing system of claims 60-65, wherein the host cell expresses the epigenetic editing system.

67. A host cell comprising the epigenetic editing system of claims 60-65, wherein the host cell expresses spdCas9.

68. A method of regulating gene expression of at least two genes, the method comprising introducing into a cell: the epigenetic modulation system of claims 60-65.

69. The method of claim 68, wherein the epigenetic modulation system reduces or inhibits expression of a first target DNA locus and increase expression of a second DNA locus gene.ACTIVEUS 211418665 132

Images