Advertisement

Discovery of hit compounds for methyl-lysine reader proteins from a target class DNA-encoded library

  • Devan J. Shell
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Justin M. Rectenwald
    Affiliations
    School of Medicine, Department of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Peter H. Buttery
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Rebecca L. Johnson
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Caroline A. Foley
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Shiva K.R. Guduru
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Mélanie Uguen
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Juanita Sanchez Rubiano
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Xindi Zhang
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Fengling Li
    Affiliations
    Structural Genomics Consortium, University of Toronto, Toronto, ON M5G 1L7, Canada
    Search for articles by this author
  • Jacqueline L. Norris-Drouin
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Matthew Axtman
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • P. Brian Hardy
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Masoud Vedadi
    Affiliations
    Structural Genomics Consortium, University of Toronto, Toronto, ON M5G 1L7, Canada
    Search for articles by this author
  • Stephen V. Frye
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Lindsey I. James
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
  • Kenneth H. Pearce
    Correspondence
    Corresponding author.
    Affiliations
    UNC Eshelman School of Pharmacy, Center for Integrative Chemical Biology and Drug Discovery, Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
    Search for articles by this author
Open AccessPublished:October 19, 2022DOI:https://doi.org/10.1016/j.slasd.2022.10.003

      Highlights

      • Focused DNA-encoded libraries (DELs) are effective in target class hit discovery.
      • Reader domains of five methyl-lysine reader proteins screened using a focused DEL.
      • Hit compounds identified from qualitative and quantitative data interpretation.
      • Compounds validated using primary and secondary biochemical assay characterization.
      • Identified hit compounds are viable starting points for chemical probe development.

      Abstract

      Methyl-lysine (Kme) reader domains are prevalent in chromatin regulatory proteins which bind post-translational modification sites to recruit repressive and activating factors; therefore, these proteins play crucial roles in cellular signaling and epigenetic regulation. Proteins that contain Kme domains are implicated in various diseases, including cancer, making them attractive therapeutic targets for drug and chemical probe discovery. Herein, we report on expanding the utility of a previously reported, Kme-focused DNA-encoded library (DEL), UNCDEL003, as a screening tool for hit discovery through the specific targeting of Kme reader proteins. As an efficient method for library generation, focused DELs are designed based on structural and functional features of a specific class of proteins with the intent of novel hit discovery. To broadly assess the applicability of our library, UNCDEL003 was screened against five diverse Kme reader protein domains (53BP1 TTD, KDM7B JmjC-PHD, CDYL2 CD, CBX2 CD, and LEDGF PWWP) with varying structures and functions. From these screening efforts, we identified hit compounds which contain unique chemical scaffolds distinct from previously reported ligands. The selected hit compounds were synthesized off-DNA and confirmed using primary and secondary assays and assessed for binding selectivity. Hit compounds from these efforts can serve as starting points for additional development and optimization into chemical probes to aid in further understanding the functionality of these therapeutically relevant proteins.

      Graphical abstract

      Keywords

      Introduction

      DNA-encoded library (DEL) screening is an evolving high-throughput screening (HTS) technology that enables efficient small molecule hit discovery. Beginning as a concept in 1992, DELs are collections of compounds that are covalently attached to distinct DNA barcodes that serve as the compound's identifier [
      • Brenner S.
      • Lerner R.A.
      Encoded combinatorial chemistry.
      ]. This allows for affinity-based selections against targets of interest followed by DNA amplification and high-throughput sequencing to identify hit compounds. DELs have exhibited great advancements and successes in recent years [
      • Favalli N.
      • Bassi G.
      • Scheuermann J.
      • Neri D.
      DNA-encoded chemical libraries - achievements and remaining challenges.
      ,
      • Chamakuri S.
      • Lu S.
      • Ucisik M.N.
      • Bohren K.M.
      • Chen Y.-C.
      • Du H.-C.
      • Faver J.C.
      • Jimmidi R.
      • Li F.
      • Li J.-Y.
      • et al.
      DNA-encoded chemistry technology yields expedient access to SARS-CoV-2 Mpro inhibitors.
      ]. Synthesized through combinatorial chemistry, DELs are generated readily and often contain millions, or even billions, of compounds. As each compound in a DEL contains a unique DNA barcode for identification which can be decoded via next-generation sequencing (NGS), this provides a robust method for compound identification post-selection. With a research footprint in both industry and academia, DEL technology continues to be a flexible and economical technique for hit discovery.
      In comparison to traditional HTS, DELs are attractive for many reasons including: 1) they can cover a vast amount of chemical space, reaching billions of compounds, 2) they cost less to produce and maintain than conventional screening libraries, 3) the screening methods are more rapid and less expensive compared to traditional HTS, and 4) compound libraries can be stored within a single tube compared to the complex storage systems required by discrete compound HTS [
      • Goodnow R.A.
      • Dumelin C.E.
      • Keefe A.D.
      DNA-encoded chemistry: enabling the deeper sampling of chemical space.
      ]. This technology offers economical flexibility in combinatorial library design for both random diversity and focused libraries. Randomized libraries tend to be much larger, which allows for greater coverage of chemical space that often centers around a core scaffold to which library building blocks/synthons are added. These large libraries can be applied across multiple protein classes as the increased compound quantity and chemical diversity can potentially deliver hit compounds across a variety of target proteins without the need for structural knowledge. This library concept is popular in the pharmaceutical industry [
      • Clark M.A.
      • Acharya R.A.
      • Arico-Muendel C.C.
      • Belyanskaya S.L.
      • Benjamin D.R.
      • Carlson N.R.
      • Centrella P.A.
      • Chiu C.H.
      • Creaser S.P.
      • Cuozzo J.W.
      • et al.
      Design, synthesis and selection of DNA-encoded small-molecule libraries.
      ]. On the other hand, the focused DEL concept contains multiple differences from the randomized approach, but has also been shown to be an effective strategy [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ,
      • Yuen L.H.
      • Dana S.
      • Liu Y.
      • Bloom S.I.
      • Thorsell A.-G.
      • Neri D.
      • Donato A.J.
      • Kireev D.
      • Schüler H.
      • Franzini R.M.
      A focused DNA-encoded chemical library for the discovery of inhibitors of NAD+-dependent enzymes.
      ,
      • Dawadi S.
      • Simmons N.
      • Miklossy G.
      • Bohren K.M.
      • Faver J.C.
      • Ucisik M.N.
      • Nyshadham P.
      • Yu Z.
      • Matzuk M.M.
      Discovery of potent thrombin inhibitors from a protease-focused DNA-encoded chemical library.
      ]. As a strategy that is more resource-friendly, libraries that follow this concept tend to be much smaller in comparison. However, it is also becoming more apparent that compound library size does not directly correlate to hit rate [
      • Satz A.L.
      • Hochstrasser R.
      • Petersen A.C.
      Analysis of current DNA encoded library screening data indicates higher false negative rates for numerically larger libraries.
      ]. Although smaller in size, the possibility of hit discovery for this technique is greatly influenced within the design of the library. Focused libraries are designed with a specific target class in mind, where prior knowledge of protein structural information and native ligand binding interactions are utilized to select specific chemical moieties and synthon (i.e., fragment) combinations that can mimic/harness these known interactions and lead to efficient discovery of hit compounds. When deciding between focused and randomized library production, there are factors to consider when finding the balance between library size and makeup that are unique to each situation. These factors include available resources for production, the projected screening targets, and the prior knowledge of binding interactions within the target set. A well-known target class with extensive knowledge of binding interactions can present an advantageous position to produce a focused, designed library that is resource efficient.
      To exploit the advantages of a focused compound library, our lab has developed a methyl-lysine (Kme) reader targeted DEL (UNCDEL003) designed for multivalent chromatin reader proteins based on structural and fragment-screening information obtained through our target class platform for these proteins [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ]. Kme reader proteins bind to post-translationally modified (PTM) methylated lysine residues present on histone protein tails found within nucleosomes and other lysine methylated proteins. PTM binding then recruits regulatory complexes to carry out specific biological processes specified by the reader's binding profile and interaction partners [
      • Wagner T.
      • Robaa D.
      • Sippl W.
      • Jung M.
      Mind the methyl: methyllysine binding proteins in epigenetic regulation.
      ]. Selectivity of reader protein recognition is obtained by: 1) the positioning of the PTMs in reference to neighboring modifications, or the histone tail as a whole, 2) differing methylation states, 3) the specific surrounding residues to the PTM, and 4) the protein surface groove features with which the histone tails interact [
      • Ruthenburg A.J.
      • Li H.
      • Patel D.J.
      • Allis C.D.
      Multivalent engagement of chromatin modifications by linked binding modules.
      ]. As therapeutic targets, Kme reader proteins exhibit several characteristics that make the class appealing for drug discovery efforts: 1) there are over 200 Kme reader proteins in the human proteome, 2) readers are often altered with point mutations, translocations, amplifications, or deletions in many cancers, and 3) structural insights are known for many of these proteins and their chromatin binding interactions [
      • Mio C.
      • Bulotta S.
      • Russo D.
      • Damante G.
      Reading cancer: chromatin readers as druggable targets for cancer treatment.
      ,
      • Barnash K.D.
      • James L.I.
      • Frye S.V.
      Target class drug discovery.
      ]. Hit discovery and binding selectivity can be sought after by screening libraries enriched with chemical moieties containing basic amines that mimic the methylation states of lysine residues. These concepts, along with the common binding motifs shared across several proteins within this class, aid in the design and application of a target class drug discovery effort through our focused DEL.
      Here, we report an experimental study to assess the application of a Kme-focused DEL for hit discovery across diverse protein domains within the Kme reader class, including the P53 binding protein 1 tandem tudor domain (53BP1 TTD), PHD finger protein 8 jumonji-C-plant homeodomain dual-domains (PHF8/KDM7B JmjC-PHD), chromodomain Y-like protein 2 chromodomain (CDYL2 CD), chromobox protein homolog 2 chromodomain (CBX2 CD), and lens epithelium-derived growth factor Pro-Trp-Trp-Pro domain (LEDGF PWWP) [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ]. These proteins contain various reader domains and have been implicated in multiple human cancers and/or infectious diseases [
      • Tang J.
      • Cho N.W.
      • Cui G.
      • Manion E.M.
      • Shanbhag N.M.
      • Botuyan M.V.
      • Mer G.
      • Greenberg R.A.
      Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination.
      ,
      • Tong Q.
      • Mazur S.J.
      • Rincon-Arano H.
      • Rothbart S.B.
      • Kuznetsov D.M.
      • Cui G.
      • Liu W.H.
      • Gete Y.
      • Klein B.J.
      • Jenkins L.
      • et al.
      An acetyl-methyl switch drives a conformational change in p53.
      ,
      • Horton J.R.
      • Upadhyay A.K.
      • Qi H.H.
      • Zhang X.
      • Shi Y.
      • Cheng X.
      Enzymatic and structural insights for substrate specificity of a family of jumonji histone lysine demethylases.
      ,
      • Dong C.
      • Liu Y.
      • Lyu T.-J.
      • Beldar S.
      • Lamb K.N.
      • Tempel W.
      • Li Y.
      • Li Z.
      • James L.I.
      • Qin S.
      • et al.
      Structural basis for the binding selectivity of human CDY chromodomains.
      ,
      • Kaustov L.
      • Ouyang H.
      • Amaya M.
      • Lemak A.
      • Nady N.
      • Duan S.
      • Wasney G.A.
      • Li Z.
      • Vedadi M.
      • Schapira M.
      • et al.
      Recognition and specificity determinants of the human cbx chromodomains.
      ,
      • Eidahl J.O.
      • Crowe B.L.
      • North J.A.
      • McKee C.J.
      • Shkriabai N.
      • Feng L.
      • Plumb M.
      • Graham R.L.
      • Gorelick R.J.
      • Hess S.
      • et al.
      Structural basis for high-affinity binding of LEDGF PWWP to mononucleosomes.
      ]. Following affinity selection and sequencing, putative hit compounds were synthesized off-DNA and tested for binding using time-resolved fluorescence resonance energy transfer (TR-FRET) assays and/or surface plasmon resonance (SPR). Compounds of interest from initial hit validation efforts were further validated as true binders by differential scanning fluorimetry (DSF) and/or isothermal titration calorimetry (ITC). Although this effort employed a relatively small DEL screened against only five Kme readers, multiple hit compounds were identified and used in follow up validation assays for the respective reader domain(s) of both the 53BP1 (15 compounds with IC50 <100 μM) and KDM7B (6 compounds with IC50 <100 μM) proteins.

      Materials and methods

      Materials

      Reagents and solvents were purchased from commercial sources. Fmoc protected building block synthons were obtained from Sigma-Aldrich (St. Louis, MO), Combi-Blocks (San Diego, CA), ChemBridge (San Diego, CA), Alfa Aesar (Ward Hill, MA), and Matrix Scientific (Columbia, SC). Dynabeads MyOne Streptavidin C1 (catalog number 65001), Dynabeads His-Tag Isolation and Pulldown (catalog number 10103D), and SYPRO Orange Protein Stain (catalog number S6651) were obtained from Invitrogen by ThermoFisher Scientific (Waltham, MA). A Phusion High-Fidelity Polymerase PCR Kit (catalog number E0553L) and Monarch DNA Gel Extraction Kit (catalog number T1020S) were obtained from New England Biolabs, Inc. (Ipswich, MA). SsoAdvanced Universal SYBR Green Supermix (catalog number 1725271) and Micro Bio-Spin P-30 columns (catalog number 7326250) were purchased from Bio-Rad Laboratories (Hercules, CA). The 300-cycle MiniSeq Mid Output Reagent Kit (catalog number FC-420-1004) was purchased from Illumina (San Diego, CA). LANCE Europium (Eu)-W1024 labeled Streptavidin (part number AD0063) and LANCE Ultra ULight-anti-6xHis labeled antibody (part number TRF0134-M) were obtained from PerkinElmer (Waltham, MA). Trizma hydrochloride (product number T3253), sodium chloride (product number S9888), β-mercaptoethanol (product number M6250), Tween-20 (product number P9416), and 1,4-Dithiothreitol (DTT, product number 1.11474) were obtained from Sigma-Aldrich (St. Louis, MO). Assay and dilution plates (#784904 and #781280 respectively) were obtained from Greiner Bio-One (Frickenhausen, Germany). Biotin-p53K381ac382me2 and H3K4me3K9me2-biotin were obtained from GenScript (Piscataway, NJ).

      Expression constructs

      The TTD of 53BP1 (residues 1484-1603 of NP_005648) was expressed with an N-terminal His-tag using a pET15 expression vector. The PHD and JmjC domains of PHF8/KDM7B (residues 1-447 of NP_055922) were expressed with either an N-terminal MBP/6x His-tag or an N-terminal MBP/C-terminal 10x His-tag. The CDs of CBX2 (residues 9-66 of NP_005180) and CDYL2 (residues 1-75 of NP_689555) were expressed with C-terminal His-tags using pET30 expression vectors. The PWWP domain of LEDGF (residues 1-135 of NP_066967) was expressed with an N-terminal His-tag using a pET28 expression vector.

      Protein expression and purification

      All expression constructs were transformed into Rosetta2 BL21(DE3)pLysS competent cells (Novagen, EMD Chemicals, San Diego, CA). Protein expression was induced by growing cells at 37°C with shaking until the OD600 reached ∼0.6-0.8, at which time the temperature was lowered to 18°C and expression was induced by adding 0.5 mM IPTG with continuous shaking overnight. Cells were harvested by centrifugation and pellets were stored at -80°C.
      His-tagged proteins were purified by resuspending thawed cell pellets in 30 mL of lysis buffer (50 mM sodium phosphate pH 7.2, 50 mM NaCl, 30 mM imidazole, 1X EDTA free protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN)) per liter of culture. Cells were lysed on ice by sonication with a Branson Digital 450 Sonifier (Branson Ultrasonics, Danbury, CT) at 40% amplitude for 12 cycles, with each cycle consisting of a 20 sec pulse followed by a 40 sec rest. The cell lysate was clarified by centrifugation and loaded onto a HisTrap FF column (GE Healthcare, Piscataway, NJ) that had been preequilibrated with 10 column volumes of binding buffer (50 mM sodium phosphate pH 7.2, 500 mM NaCl, 30 mM imidazole) using an AKTA FPLC (GE Healthcare, Piscataway, NJ). The column was washed with 15 column volumes of binding buffer and protein was eluted in a linear gradient to 100% elution buffer (50 mM sodium phosphate pH 7.2, 500 mM NaCl, 500 mM imidazole) over 20 column volumes. Peak fractions containing the desired protein were pooled and concentrated to 2 mL in Amicon Ultra-15 concentrators (Merck Millipore, Carrigtwohill Co. Cork IRL). Concentrated protein was loaded onto either a HiLoad 26/60 Superdex 75 prep grade or 26/60 Superdex 200 prep grade column (GE Healthcare, Piscataway, NJ) that had been preequilibrated with 1.2 column volumes of sizing buffer (25 mM Tris-HCl pH 7.5, 250 mM NaCl, 2 mM DTT, 5% glycerol) using an ATKA Purifier (GE Healthcare, Piscataway, NJ). Protein was eluted isocratically in sizing buffer over 1.3 column volumes at a flow rate of 2 mL/min collecting 3 mL fractions. Peak fractions were analyzed for purity by SDS-PAGE and those containing pure protein were pooled and concentrated using Amicon Ultra-15 concentrators (Merck Millipore, Carrigtwohill Co. Cork IRL). Proteins were dialyzed into a buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, and 2 mM β-mercaptoethanol prior to use for ITC.

      MBP removal

      The N-terminal MBP tag was removed from both versions of PHF8/KDM7B by TEV protease cleavage according to manufacturer's recommendations (Sigma-Aldrich, St. Louis, MO). Briefly, purified protein was incubated with GST/His tagged recombinant TEV at a concentration of 1 unit TEV per milligram of tagged protein for 16 hours at 4°C. The cleavage reaction was then passed over a HisTrap FF column to remove the TEV protease and any protein that still retained the tag. The column flow through was then collected and concentrated to 2 mL using Amico Ultra-15 concentrators, 10,000 molecular weight cut-off (Merck Millipore, Carrigtwohill Co. Cork IRL). Concentrated protein was loaded onto a HiLoad 26/60 Superdex 200 prep grade column (GE Healthcare, Piscataway, NJ) that had been preequilibrated with 1.2 column volumes of sizing buffer (25 mM Tris-HCl pH 7.5, 250 mM NaCl, 2 mM DTT, 5% glycerol) using an ATKA FPLC (GE Healthcare, Piscataway, NJ). Protein was eluted isocratically in sizing buffer over 1.3 column volumes at a flow rate of 2 mL/min collecting 3 mL fractions. Peak fractions were analyzed for purity by SDS-PAGE and those containing pure protein were pooled and concentrated using Amicon Ultra-15 concentrators, 10,000 molecular weight cut-off (Merck Millipore, Carrigtwohill Co. Cork IRL). C-terminal His-tagged protein retained the His-tag after MBP removal while the N-terminal His-tag was removed along with the MBP. Untagged protein was exchanged into buffer containing 25 mM Tris-HCl pH 7.5, 150 mM NaCl, and 2 mM β-mercaptoethanol prior to use for ITC.

      Biotinylated 53BP1 expression and purification

      Biotinylated 53BP1 TTD (residues 1484-1603 of NP_005648) was generated by cloning the coding sequence into a modified pET28 vector that includes both an N-terminal hexahistidine tag and Avi tag sequence (GLNDIFEAQKIEWHE). The resulting construct was co-transformed into Rosetta2 BL21(DE3)pLysS competent cells (Novagen, EMD Chemicals, San Diego, CA) along with an expression plasmid for E. coli biotin ligase, BirA (pET21a-BirA was a gift from Alice Ting (Addgene plasmid # 20857)142) [
      • Howarth M.
      • Takao K.
      • Hayashi Y.
      • Ting A.Y.
      Targeting quantum dots to surface proteins in living cells with biotin ligase.
      ]. A 2 L culture was grown to mid log phase at 37°C, at which time the temperature was lowered to 18°C and protein expression was induced by addition of 0.5 mM IPTG. D-biotin was added to the culture at this time to a final concentration of 50 μM. Expression was allowed to continue overnight. Cells were harvested by centrifugation and pellets were stored at -80°C.
      Biotinylated 53BP1 TTD was purified by resuspending thawed cell pellets in 30 mL of lysis buffer (50 mM sodium phosphate pH 7.2, 500 mM NaCl, 30 mM imidazole, 1X EDTA free protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN)) per liter of culture. Cells were lysed on ice by sonication with a Branson Digital 450 Sonifier (Branson Ultrasonics, Danbury, CT) at 40% amplitude for 12 cycles with each cycle consisting of a 20 sec pulse followed by a 40 sec rest. The cell lysate was clarified by centrifugation and loaded onto a HisTrap FF column (GE Healthcare, Piscataway, NJ) that had been pre-equilibrated with 10 column volumes of binding buffer (50 mM sodium phosphate pH 7.2, 500 mM NaCl, 30 mM imidazole) using an AKTA FPLC (GE Healthcare, Piscataway, NJ). The column was washed with 15 column volumes of binding buffer and protein was eluted in a linear gradient to 100% elution buffer (50 mM sodium phosphate pH 7.2, 500 mM NaCl, 500 mM imidazole) over 20 column volumes. Peak fractions containing the desired protein were pooled and concentrated to 1.5 mL in Amicon Ultra-15 concentrators 3,000 molecular weight cut-off (Merck Millipore, Carrigtwohill Co. Cork IRL). Biotinylated 53BP1 was isolated by passing concentrated protein over a 2 mL monomeric avidin column by gravity flow following manufacturer's recommendations (Pierce Biotechnology/Thermo Scientific, Rockford, IL). Bound protein was eluted in 12 mL PBS containing 2 mM D-biotin. Eluted protein was concentrated in Amicon Ultra-15 concentrators 3,000 molecular weight cut-off to 2 mL and further purified by loading onto a HiLoad 26/60 Superdex 75 prep grade column (GE Healthcare, Piscataway, NJ) that had been pre-equilibrated with 1.2 column volumes of sizing buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM DTT, 5% glycerol) using an ATKA Purifier (GE Healthcare, Piscataway, NJ). Protein was eluted isocratically in sizing buffer over 1.3 column volumes at a flow rate of 2 mL/min collecting 3 mL fractions. Peak fractions were analyzed for purity by SDS-PAGE and those containing pure protein were pooled and concentrated to 0.5 mL using Amicon Ultra-15 concentrators 3,000 molecular weight cut-off. The concentrated protein was aliquoted and stored at -80°C in sizing buffer.

      DNA-encoded library production

      UNCDEL003 was designed for multivalent Kme reader proteins and constructed as described previously [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ]. The library was synthesized using a traditional DNA-recording strategy to track three cycles of synthon incorporation. The DEL was stored as the third cycle product until ready for testing. Before completing selections, two equivalents of a closing primer were ligated to a 1000 pmol aliquot of library using T4 DNA Ligase with a final ligation volume of 50 µL in T4 Ligase Buffer (New England Biolabs, USA). A Micro Bio-Spin P-30 column was used to desalt the ligation product. The closing primer sequence was filled in by Klenow polymerization using DNA Polymerase I, Large (Klenow) overnight on the benchtop. The library was heat quenched at 80°C for 20 min and allowed to cool to room temperature. A Micro Bio-Spin P-30 column was used to desalt the final library product.

      DEL selections

      DEL selections and washes were performed in buffer containing 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.005% Tween-20 (vol/vol), and 2 mM DTT. 200 mM NaCl was used in LEDGF selections to mitigate non-specific DNA binding. On the day of the selection, the DTT was added fresh. 1 µM of biotinylated or His-tagged protein was immobilized on a previously tested and verified volume of Dynabeads MyOne Streptavidin C1 or Dynabeads His-Tag Isolation and Pulldown beads respectively. Protein was incubated with the associated bead type for 30 min (total volume 50 µL). The selection was completed manually using a magnetic tube rack. Three washes of buffer (200 µL) were used to wash excess protein away from the beads. 10 pmol of UNCDEL003 (in 100 µL) was added to the immobilized protein, then mixed gently for one hour on a rotating mixer. After one hour of mixing, the supernatant was removed. A single wash using buffer (200 µL) was performed to wash away any non-binding compounds. Then, fresh buffer (100 µL) was added to the protein/DEL mixture for a heat elution at 80°C for 10 min. The solution of eluted molecules was removed and added to a fresh batch of immobilized protein for another round of selection. This process was completed for a total of three cycles of incubation and wash steps. The last elution collected was used for amplification and sequencing.

      Amplification and sequencing

      An Analytik Jena qTower3 real-time PCR thermal cycler was used to amplify the eluted DNA from the DEL selections. The qPCR reactions contained 2 µL of elution, 500 nM of each complimentary forward and reverse primer, and Bio-Rad SsoAdvanced Universal SYBR Green Supermix containing dNTPs, Sso7d fusion polymerase, MgCl2, SYBR Green 1, and ROX normalization dyes diluted to 1X. The amplification protocol included 5 min at 95°C, then 40 cycles of 30 sec at 92°C, 30 sec at 61°C, and 15 sec at 72°C, followed by a final 10 min at 72°C. When the sample amplifications were two cycles post-plateau, they were stopped and removed. An agarose (1%) gel-based purification system was used to extract and purify the amplified DNA product. The amplified DNA products were cut from the agarose gel, dissolved, and extracted using a Monarch DNA Gel Extraction Kit.
      The standard Illumina MiniSeq library preparation protocol was followed to prepare the gel-purified DNA samples for sequencing. For quality control purposes, a randomized nucleotide region was incorporated into each individual DNA barcode during library construction to mitigate selection and sequencing bias. In addition, PhiX library control was added to our samples during final library preparation to increase sequence diversity and spatial distribution for the instrument. A 300-cycle MiniSeq Mid Output Reagent Cartridge was used on an Illumina MiniSeq instrument to sequence the DNA.

      Sequencing data visualization

      The bcl2fastq2 program was used to convert the Illumina MiniSeq .bcl files to fastq.gz files on the UNC Longleaf supercomputing cluster. Pipeline Pilot software was used to create a workflow that would refine the sequencing data. The sequences first were required to contain the DNA scaffold and closing primer sequences. Then, each cycle sequence was searched for, beginning with the A synthon cycle. If there were multiple reads of the same random region sequence in the closing primer, they were concatenated to one frequency count. The frequency of the read was then used to rank the compounds. Because many compounds share the same frequency, they were then provided a cluster rank (cRank) to prevent over ranking of one compound versus another. Microsoft Excel and TIBCO Spotfire were used to visualize data for compound synthon trends and selection frequency.

      Compound synthesis

      Selected hit molecule syntheses were obtained by solid-phase peptide synthesis techniques on Fmoc-Rink Amide MBHA resin. Compound syntheses were performed in a polypropylene syringe fitted with a porous polyethylene frit. Solvents and soluble reagents were removed by applying pressure on the syringe piston. NMP can be substituted for DMF in this protocol. Fmoc-Rink Amide MBHA (0.01 mmol) resin was soaked in 1 mL of DCM for 30 min on a tabletop shaker. After removing the solvent, resin was soaked in NMP solvent (1 mL) and allowed to shake for 15 min. Solvent was drained, and the syringe was filled with a 25% piperidine (or 2.5% pyrrolidine and DBU) solution in NMP (1 mL) and allowed to shake for 10 min. The deprotection solution was drained, and the resin was washed thrice with NMP solvent (3 × 2 mL). Then, the pre-activated Fmoc-amino acid (Fmoc-AA) in NMP was loaded into the syringe and shaken for 1 hour. [Pre-activation of Fmoc-AA: To the Fmoc-AA (0.03 mmol) dissolved in NMP (1 mL), DIPEA (0.05 mmol) was added and followed by HATU coupling reagent (0.03 mmol) in NMP (1 mL) and allowed to sit for 10 min.] Washings between deprotection, coupling, and subsequent deprotection steps were carried out with NMP or DMF (3 × 1 min) using 2 mL of solvent per 0.2 mmol of resin for each wash. After completion of the compound sequence, the resin was washed with NMP (2 × 3 mL) followed by DCM (2 × 3 mL). Then, the syringe was filled with a cleavage cocktail (1 mL) and shaken for 1 hour. [Cleavage cocktail preparation: TFA:H2O:TIPS-95:2.5:2.5 respectively.] Then, the TFA solution was collected in a glass vial and evaporated. The crude compound was purified using an Agilent Technologies 1260 Infinity preparative-HPLC [Solvent system: water:acetonitrile (90:10); Flow rate: 40 mL/min; UV detector wavelength: 254 nm] to obtain a pure final product. Final compounds were verified by LCMS prior to testing. All LCMS and NMR data of synthesized compounds are stored within the UNC CICBDD database.
      Analytical LCMS data for compounds were acquired on an Agilent 6110 Series system with UV detector set to 220 nm, 254 nm, and 280 nm. Samples were injected onto an Agilent ZORBAX Eclipse Plus 4.6 × 50 mm, 1.8 µm, C18 column at 25°C. Mobile phases A (H2O + 0.1% acetic acid) and B (ACN + 1% water and 0.1% acetic acid) were used in a linear gradient from 10% to 100% B in 5 min, followed by a flush at 100% B for another 2 min with a flow rate of 1.0 mL/min. Mass spectra (MS) data were acquired in positive ion mode using an Agilent 6110 single quadrupole mass spectrometer with an electrospray ionization (ESI) source.

      Time-resolved fluorescence resonance energy transfer (TR-FRET)

      TR-FRET assays were developed and performed following the workflow previously reported with adapted reagents and concentrations [
      • Rectenwald J.M.
      • Hardy P.B.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • James L.I.
      • Frye S.V.
      • Pearce K.H.
      A general TR-FRET assay platform for high-throughput screening and characterizing inhibitors of methyl-lysine reader proteins.
      ]. Briefly, assays were run using white, low-volume, flat-bottom, nonbinding, 384-well microplates (Greiner, 784904) containing a total assay volume of 10 μL per well. Test compounds were dispensed across a mother plate at 100X concentration to create a 10-point dose-response format with a 3-fold dilution between each data point using a TECAN Freedom EVO liquid handling workstation. A TTP Labtech Mosquito HTS liquid handling instrument was used to prepare assay plates by stamping 100 nL of control compound, compound dose-response curves from the mother plate, and DMSO control. Protein, biotinylated tracer ligand, and the TR-FRET reagents were combined prior to dispensing across the test plate. Promptly after the TR-FRET mixture was prepared, a Multidrop Combi was used to dispense 10 µL of the reagent mixture to each well of an assay ready plate. Compound inhibition was calculated on a scale of 0% (DMSO vehicle only) to 100% (100 µM control peptide/compound) in reference to the controls on each plate. Dose-response curves were fitted with a four-parameter non-linear regression analysis using ScreenAble software. Normalized data was then replotted using GraphPad Prism 9.0 software for final figure preparation.
      For the 53BP1 TTD TR-FRET assay, the assay buffer was composed of 20 mM Tris-HCl pH 7.5, 25 mM NaCl, 0.005% Tween-20 (vol/vol), and 2 mM DTT. LANCE Europium (Eu)-W1024 labeled Streptavidin (2 nM) and LANCE Ultra ULight-anti-6xHis labeled antibody (35 nM) were used as donor and acceptor fluorophores associated with the tracer ligand and protein, respectively. Final assay concentrations of 180 nM His-tagged 53BP1 and 180 nM of biotin-p53K381acK382me2 as a tracer ligand were used for compound testing.
      For the additional proteins tested using TR-FRET, the following differences from the above protocol were used with an assay buffer composed of 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.05% Tween-20 (vol/vol), and 2 mM DTT. In the KDM7B TR-FRET assay, 30 nM of His-tagged KDM7B and 50 nM of H3K4me3K9me2-biotin were used with LANCE Europium (Eu)-W8044 labeled Streptavidin (2 nM) and LANCE Ultra ULight-anti-6xHis labeled antibody (10 nM). In the CDYL2 TR-FRET assay, 10 nM His-tagged CDYL2 and 10 nM of UNC4195 (biotinylated chromodomain binding compound control) were used with LANCE Europium (Eu)-W1024 labeled Streptavidin (2 nM) and LANCE Ultra ULight-anti-6xHis labeled antibody (10 nM). In the CBX2 TR-FRET assay, 30 nM His-tagged CBX2 and 30 nM of UNC4195 were used with LANCE Europium (Eu)-W1024 labeled Streptavidin (2 nM) and LANCE Ultra ULight-anti-6xHis labeled antibody (10 nM).

      Differential scanning fluorimetry (DSF)

      Experiments were conducted using an Analytik Jena qTower3 real-time PCR thermal cycler. The buffer utilized for DSF was composed of 20 mM Tris-HCl pH 7.5, 75 mM NaCl, 0.005% Tween-20 (vol/vol), and 2 mM DTT. 53BP1 and KDM7B experiments were carried out using 10X compound over protein (5 µM protein and 50 µM compound) and 5X SYPRO Orange Protein Stain pre-incubated with compounds for 15 min before running the temperature gradient (1°C/min) from 25°C to 90°C to observe protein denaturation. The Analytik Jena software uses the first derivative of the 350/300 nm ratio to calculate the melting temperature.

      Isothermal titration calorimetry (ITC)

      All ITC measurements were recorded at 25°C with a MicroCal Auto-iTC200 isothermal titration calorimeter (Malvern Panalytical, UK). All protein and compound stock samples were stored in ITC buffer (25 mM Tris–HCl pH 7.5, 150 mM NaCl, and 2 mM β-mercaptoethanol) and then diluted to achieve the desired concentrations. Typically, 100 μM protein and 1.0 mM compound were used, and variations in these concentrations always maintained a 10:1 compound-to-protein ratio for all ITC experiments. Compound dilutions were made from 10 mM DMSO stocks and amounts of DMSO in the syringe and sample cell were constant in each experiment at 1% final DMSO. The experimental protocol included a single 0.2 µL compound injection into a 200 µL sample cell filled with protein, followed by 26 subsequent 1.5 µL injections of compound. Injections were performed with a spacing of 180 sec and a reference power of 8 cal/sec. The initial data point from the partial injection was routinely removed prior to analysis. The titration data was analyzed using Origin 7.0 software (Malvern Panalytical, UK) by the nonlinear least-squares method, fitting the heats of binding as a function of the compound-to-protein ratio to a one-site binding model.

      Surface plasmon resonance (SPR)

      LEDGF-PWWP SPR experiments were carried out on a Biacore 8K SPR system (Cytiva, USA). Biotinylated protein was immobilized on the flow cell of an SA sensor chip in 1X HBS-EP (Cytiva, USA) buffer, yielding approximately 4300 RU. Using the same buffer, with 0.5% (replicate 1) or 2% (replicate 2) of DMSO, single-cycle kinetics were used with a 60 sec contact time and a dissociation time of 120 sec at a flow rate of 75 µL/min. Compounds were tested at 100 µM (replicate 1) or 200 µM (replicate 2) as the highest concentration. A dilution factor of 0.25 was used to yield 5 concentrations for testing.

      Results and discussion

      In recent studies, our lab has worked towards developing a target class platform for the discovery and development of chemical probes for various Kme reader proteins [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ,
      • Barnash K.D.
      • James L.I.
      • Frye S.V.
      Target class drug discovery.
      ,
      • Rectenwald J.M.
      • Hardy P.B.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • James L.I.
      • Frye S.V.
      • Pearce K.H.
      A general TR-FRET assay platform for high-throughput screening and characterizing inhibitors of methyl-lysine reader proteins.
      ,
      • Barnash K.D.
      • Lamb K.N.
      • Stuckey J.I.
      • Norris J.L.
      • Cholensky S.H.
      • Kireev D.B.
      • Frye S.V.
      • James L.I.
      Chromodomain ligand optimization via target-class directed combinatorial repurposing.
      ,
      • Wigle T.J.
      • Herold J.M.
      • Senisterra G.A.
      • Vedadi M.
      • Kireev D.B.
      • Arrowsmith C.H.
      • Frye S.V.
      • Janzen W.P.
      Screening for inhibitors of low-affinity epigenetic peptide-protein interactions: an AlphaScreen-based assay for antagonists of methyl-lysine binding proteins.
      ,
      • Engelberg I.A.
      • Foley C.A.
      • James L.I.
      • Frye S.V.
      Improved methods for targeting epigenetic reader domains of acetylated and methylated lysine.
      ,
      • Barnash K.D.
      • Lamb K.N.
      • James L.I.
      • Frye S.V.
      Peptide technologies in the development of chemical tools for chromatin-associated machinery.
      ,
      • James L.I.
      • Frye S.V.
      Chemical probes for methyl lysine reader domains.
      ]. These efforts involved the development of a focused DEL platform, biophysical assays for Kme readers, and a chemical strategy based on pharmacophore and structure-based design and optimization. In development, the Kme-focused UNCDEL003 was validated within our lab by screening the library against another reader protein domain, the chromobox protein homolog 7 (CBX7) CD, using the UNCDEL003 with positive and negative control compounds spiked into the library [
      • Rectenwald J.M.
      • Guduru S.K.R.
      • Dang Z.
      • Collins L.B.
      • Liao Y.-E.
      • Norris-Drouin J.L.
      • Cholensky S.H.
      • Kaufmann K.W.
      • Hammond S.M.
      • Kireev D.B.
      • et al.
      Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
      ]. It was observed that the positive control compound was enriched during the selection process in comparison to the negative control, which was not observed. The positive control compound was highly enriched over beads alone as well as the input library baseline, where the spiked control compound was observed within the normal distribution of compounds in the library. These validation efforts enabled refinement of our selection process using the UNCDEL003 library for enriching compounds that bind to the target protein.
      To further validate and express the utility of our focused UNCDEL003 library, comprised of approximately 58,000 compounds, we screened multiple Kme reader protein domains, including the 53BP1 TTD, PHF8/KDM7B JmjC-PHD, CDYL2 CD, CBX2 CD, and LEDGF PWWP. These proteins are therapeutically relevant, represent multiple subfamilies within the Kme reader target class, and possess structurally diverse domains in which chromatin interaction is achieved (Table 1, Fig. 1). For the DEL screening process, 3 rounds of selections with fresh protein were completed as detailed in Materials and Methods. Following next-generation sequencing, we determined compound synthon prevalence through Spotfire 4-D visualizations and quantitative combination frequencies. Using these methods, we identified compounds for off-DNA synthesis and biochemical assay characterization. Top compound hits were further validated by DSF and/or ITC and evaluated in a binding selectivity panel against the other reader proteins within the study via a TR-FRET assay.
      Table 1Characterization of target Kme reader proteins and chromatin-interacting domains chosen for DEL screening.
      ProteinUniProt AccessionDomainsPTM RecognitionPDBReference
      53BP1Q12888Tandem TudorH4K20me2,
      Non-histone PTM.
      p53K381ac

      K382me2
      2LVM, 4X34[
      • Tang J.
      • Cho N.W.
      • Cui G.
      • Manion E.M.
      • Shanbhag N.M.
      • Botuyan M.V.
      • Mer G.
      • Greenberg R.A.
      Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination.
      ,
      • Tong Q.
      • Mazur S.J.
      • Rincon-Arano H.
      • Rothbart S.B.
      • Kuznetsov D.M.
      • Cui G.
      • Liu W.H.
      • Gete Y.
      • Klein B.J.
      • Jenkins L.
      • et al.
      An acetyl-methyl switch drives a conformational change in p53.
      ]
      PHF8/KDM7BQ9UPP1PHD, JumonjiH3K4me3K9me23KV4
      • Horton J.R.
      • Upadhyay A.K.
      • Qi H.H.
      • Zhang X.
      • Shi Y.
      • Cheng X.
      Enzymatic and structural insights for substrate specificity of a family of jumonji histone lysine demethylases.
      CDYL2Q8N8U2ChromodomainH3K9me3, H3K27me36V2H
      • Dong C.
      • Liu Y.
      • Lyu T.-J.
      • Beldar S.
      • Lamb K.N.
      • Tempel W.
      • Li Y.
      • Li Z.
      • James L.I.
      • Qin S.
      • et al.
      Structural basis for the binding selectivity of human CDY chromodomains.
      CBX2Q14781ChromodomainH3K27me33H91
      • Kaustov L.
      • Ouyang H.
      • Amaya M.
      • Lemak A.
      • Nady N.
      • Duan S.
      • Wasney G.A.
      • Li Z.
      • Vedadi M.
      • Schapira M.
      • et al.
      Recognition and specificity determinants of the human cbx chromodomains.
      LEDGF/PSIP1O75475PWWPH3K36me32M16
      • Eidahl J.O.
      • Crowe B.L.
      • North J.A.
      • McKee C.J.
      • Shkriabai N.
      • Feng L.
      • Plumb M.
      • Graham R.L.
      • Gorelick R.J.
      • Hess S.
      • et al.
      Structural basis for high-affinity binding of LEDGF PWWP to mononucleosomes.
      low asterisklow asterisk Non-histone PTM.
      Fig 1
      Fig. 1Kme reader protein domains selected for DEL screening via the UNCDEL003 library. These specific protein domains are responsible for the histone interactions of these target proteins. Protein residues that interact with histone marks are highlighted. Acronyms: tandem tudor domain (TTD), jumonji-C (JmjC), plant homeodomain (PHD), Pro-Trp-Trp-Pro (PWWP), and chromodomain (CD). PDB codes for these isolated domains are listed in .

      P53 binding protein 1 (53BP1)

      Although initially discovered through the association with tumor suppressor protein p53, 53BP1 is now well known for the role it plays in DNA double-strand break (DSB) repair [
      • Mirza-Aghazadeh-Attari M.
      • Mohammadzadeh A.
      • Yousefi B.
      • Mihanfar A.
      • Karimian A.
      • Majidinia M.
      53BP1: a key player of DNA damage response with critical functions in cancer.
      ]. 53BP1 recruits DSB repair proteins to the site of damaged DNA for subsequent repair signaling, promoting ataxia-telangiectasia mutated (ATM)-dependent checkpoint signaling, DSB repair pathway choice, and the synapsis of distal DNA ends during non-homologous end-joining (NHEJ) DSB repair [
      • Panier S.
      • Boulton S.J.
      Double-strand break repair: 53BP1 comes into focus.
      ]. Localization of the protein to DSB sites is heavily influenced by the interaction of the TTD with di-methylated lysine 20 of histone H4 (H4K20me2) and a ubiquitination-dependent recruitment (UDR) domain, which interacts with H2AK15ub and its surrounding residues on the H2A tail for recruitment to DSBs [
      • Fradet-Turcotte A.
      • Canny M.D.
      • Escribano-Díaz C.
      • Orthwein A.
      • Leung C.C.Y.
      • Huang H.
      • Landry M.-C.
      • Kitevski-LeBlanc J.
      • Noordermeer S.M.
      • Sicheri F.
      • et al.
      53BP1 is a reader of the DNA-damage-induced H2A Lys 15 ubiquitin mark.
      ,
      • Botuyan M.V.
      • Lee J.
      • Ward I.M.
      • Kim J.-E.
      • Thompson J.R.
      • Chen J.
      • Mer G.
      Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair.
      ]. Once localized to the appropriate site, 53BP1 serves as a scaffold for members in the NHEJ pathway. The TTD has also been observed to bind native p53 as well as a modified p53 peptide that is acetylated at lysine 381 and dimethylated at lysine 382 (p53K381ac382me2) [
      • Tong Q.
      • Mazur S.J.
      • Rincon-Arano H.
      • Rothbart S.B.
      • Kuznetsov D.M.
      • Cui G.
      • Liu W.H.
      • Gete Y.
      • Klein B.J.
      • Jenkins L.
      • et al.
      An acetyl-methyl switch drives a conformational change in p53.
      ]. In past studies, small molecules have been discovered that bind and/or inhibit 53BP1 via the TTD [
      • Sun Y.
      • Lu H.
      • Fang X.
      • Xiao S.
      • Yang F.
      • Chen Y.
      • Wang H.
      • Li X.
      • Lu J.
      • Lin H.
      • et al.
      Discovery of a novel 53BP1 inhibitor through AlphaScreen-based high-throughput screening.
      ,
      • Perfetti M.T.
      • Baughman B.M.
      • Dickson B.M.
      • Mu Y.
      • Cui G.
      • Mader P.
      • Dong A.
      • Norris J.L.
      • Rothbart S.B.
      • Strahl B.D.
      • et al.
      Identification of a fragment-like small molecule ligand for the methyl-lysine binding protein, 53BP1.
      ]. Here, we included the 53BP1 reader protein within our panel of targets to discover novel domain-specific inhibitors as a starting point to promote future investigation into the role of 53BP1 in DSB repair and protein recruitment.
      Following DEL selections with the 53BP1 TTD, analysis of the experimental sequencing data was completed using Spotfire, creating a 4D graph plotting the three various synthon groups and utilizing selection frequency counts to determine data point size. The C6 and C23 synthons showed a clear prominence in the affinity-selected compounds (Supplemental Fig. 1). The 20 compounds that were selected for synthesis without the DNA barcode were from the two putative hit series containing the fragments C6 and C23 (Supplemental Fig. 2). Accordingly, these features are also predominant in the compounds with the highest selection frequency counts (Table 2). As these B and C compound synthons appeared frequently within the top putative hits, and no clear trend in the A synthon was observed, two B-C only compounds were synthesized for testing (Table 2).
      Table 2Putative hit compounds chosen for synthesis off-DNA based on 53BP1 TTD DEL selections. Corresponding selection frequency, normalized enrichment percentage, UNC ID, TR-FRET (IC50 in µM), and DSF (ΔTm in °C) results are presented. Normalized enrichment percentages were calculated by the following: (compound selection frequency/total reads) x 100. All TR-FRET data values are reported as the mean ± SEM of the listed number of technical replicates. All DSF data values are reported as the mean ± SEM of three technical replicates.
      Synthon CombinationSelection

      Frequency
      Normalized Enrichment %
      22,562 total reads.
      UNC CodeTR-FRET IC50Rep.Δ Tm
      A12-B22-C6670.30UNC7605> 10060.5 ± 0.1
      A6-B20-C6540.24UNC758770.3 ± 3.831.3 ± 0.1
      A30-B25-C6530.23UNC758819.6 ± 2.353.3 ± 0.02
      A14-B22-C6480.21UNC7590> 10011-2.5 ± 0.02
      A21-B28-C6480.21UNC759158.5 ± 3.961.8 ± 0.4
      A26-B28-C6470.21UNC759221.5 ± 2.5172.8 ± 0.4
      A8-B22-C6470.21UNC764721.9 ± 1.772.8 ± 0.4
      A8-B25-C6450.20UNC76488.6 ± 1.082.8 ± 0.04
      A8-B38-C23450.20UNC765045.8 ± 6.441.8 ± 0.5
      A2-B20-C6400.18UNC765180 ± 9.021.1 ± 0.01
      A30-B22-C6390.17UNC765247.9 ± 13.872.8 ± 0.5
      A19-B1-C6370.16UNC768566.0 ± 1.021.1 ± 0.02
      A5-B20-C6370.16UNC768674.5 ± 1.521.1 ± 0.04
      A12-B7-C6340.15UNC760620.8 ± 1.251.8 ± 0.04
      A17-B1-C6320.14UNC768744.0 ± 7.851.2 ± 0.01
      A7-B16-C23290.13UNC7688> 10020.8 ± 0.03
      A17-B29-C23260.12UNC7689> 10020.1 ± 0.01
      A1-B9-C6230.10UNC7653> 10021.0 ± 0.01
      B25-C6--UNC764925.8 ± 3.052.2 ± 0.02
      A26-C6--UNC758924.7 ± 2.6102.7 ± 0.4
      low asterisk 22,562 total reads.
      After synthesis, compounds were tested using a TR-FRET competitive binding assay adapted for a 53BP1 TTD protein construct in dose-response format (Table 2). Most hits produced dose-response curves with IC50 values <100 µM. Seven of the synthesized compounds produced IC50 values <30 µM (Fig. 2A). These top seven compounds all shared the C6 synthon, and the top binding compound exhibited an IC50 value of 8.6 ± 1.0 µM (UNC7648), while one of the three synthesized compounds containing the C23 synthon produced an IC50 value of 45.8 ± 6.4 µM (UNC7650). Importantly, although UNC7648 did not possess the highest selection frequency count, this compound was observed with high enrichment (0.20%) in comparison to other compounds within the library (Table 2). In addition, it was encouraging to observe that 75% of the compounds selected for synthesis exhibited TR-FRET IC50 values <100 µM. Therefore, we can report with confidence that the screening methods and compound selection strategies used here were robust and successful in the discovery of small molecules that bind the 53BP1 TTD.
      Fig 2
      Fig. 2A) TR-FRET produced dose-response curves for the top seven binding compounds to the 53BP1 TTD with IC50 values <30 µM. B) DSF experimental data for the same seven hit compounds against the 53BP1 TTD, as well as protein and p53 peptide controls. C) Chemical structures and corresponding IC50 values (mean ± SEM) of the top three compounds identified from TR-FRET testing.
      DSF was utilized as a parallel strategy to validate compound binding to the 53BP1 TTD and to prioritize hits for ITC experiments (Fig. 2B, Supplemental Fig. 3). The controls included protein in buffer alone, and protein in buffer with 2% DMSO. These controls maintained similar protein melt temperatures with an average of 49.6 ± 0.1°C. The modified, unlabeled version of the p53K381acK382me2 peptide utilized for the TR-FRET assay was tested as well and was expected to be useful as a positive control for observing binding and protein stabilization. However, there was not a significant shift in the melting temperature with the peptide tested. In general, the compounds we synthesized stabilized the protein, as they increased the melting temperature, some compounds by more than 2°C. The two most potent compounds by TR-FRET, UNC7588 and UNC7648, had average thermal shift values of 3.3 ± 0.02°C and 2.8 ± 0.04°C, respectively (Table 2). Chemical structures of the top three binding compounds observed are shown in Fig. 2C.
      To ensure the validity of the top hit compound, ITC experiments were completed with the 53BP1 TTD construct and UNC7648. This compound was the most potent inhibitor in the TR-FRET assay and represents the most attractive compound for future SAR studies. The ITC experiments were completed in triplicate, revealing a Kd value of 14.5 ± 2.3 µM (Fig. 3). The average stoichiometric (n) value of the experiments was 0.7 ± 0.2 sites, suggesting the compound could be binding to, or inducing, a 53BP1 protein dimer, as observed for some prior small molecule ligands [
      • Perfetti M.T.
      • Baughman B.M.
      • Dickson B.M.
      • Mu Y.
      • Cui G.
      • Mader P.
      • Dong A.
      • Norris J.L.
      • Rothbart S.B.
      • Strahl B.D.
      • et al.
      Identification of a fragment-like small molecule ligand for the methyl-lysine binding protein, 53BP1.
      ]. Through the combination of DEL selections, TR-FRET, DSF, and ITC data, we have demonstrated that our focused DEL provided unique hit compounds for the 53BP1 TTD. These hit compounds are modest in potency but provide many new starting points for SAR studies and chemical probe development.
      Fig 3
      Fig. 3ITC analysis of UNC7648 and 53BP1 TTD binding, revealing a Kd of 14.5 ± 2.3 µM (blue). An experimental control observing the titration of UNC7648 into buffer is also shown (red). All ITC data values were analyzed and reported as the mean ± SEM of three technical replicates.

      Nε-methyl lys-demethylase 7B/ PHD finger protein 8 (KDM7B/PHF8)

      The addition of a methyl mark through methyl transferases, or removal of a methyl mark by a demethylase, can alter gene expression in healthy or diseased tissues [
      • Loenarz C.
      • Ge W.
      • Coleman M.L.
      • Rose N.R.
      • Cooper C.D.O.
      • Klose R.J.
      • Ratcliffe P.J.
      • Schofield C.J.
      PHF8, a gene associated with cleft lip/palate and mental retardation, encodes for an Nepsilon-dimethyl lysine demethylase.
      ]. In this study, the KDM7B protein was chosen for screening. KDM7B, also known as PHF8, contains a PHD Kme reader domain that aids in histone mark detection of H3K4me3 and enhances demethylation. Adjacent to the PHD domain, KDM7B also contains a JmjC domain that can bind and demethylate H3K9me2 [
      • Horton J.R.
      • Upadhyay A.K.
      • Qi H.H.
      • Zhang X.
      • Shi Y.
      • Cheng X.
      Enzymatic and structural insights for substrate specificity of a family of jumonji histone lysine demethylases.
      ]. Located between these two domains is a linker region in which the flexibility and conformation both play a role in chromatin binding selectivity along with the PHD and JmjC domains [
      • Chaturvedi S.S.
      • Ramanan R.
      • Waheed S.O.
      • Ainsley J.
      • Evison M.
      • Ames J.M.
      • Schofield C.J.
      • Karabencheva-Christova T.G.
      • Christov C.Z.
      Conformational dynamics underlies different functions of human KDM7 histone demethylases.
      ]. This protein has been reported to play a significant role in neuronal differentiation and craniofacial development, as mutations in KDM7B are often connected to X-linked mental retardation and cleft lip/cleft palate [
      • Abidi F.E.
      • Miano M.G.
      • Murray J.C.
      • Schwartz C.E.
      A novel mutation in the PHF8 gene is associated with X-linked mental retardation with cleft lip/cleft palate.
      ,
      • Qi H.H.
      • Sarkissian M.
      • Hu G.-Q.
      • Wang Z.
      • Bhattacharjee A.
      • Gordon D.B.
      • Gonzales M.
      • Lan F.
      • Ongusaha P.P.
      • Huarte M.
      • et al.
      Histone H4K20/H3K9 demethylase PHF8 regulates zebrafish brain and craniofacial development.
      ]. In addition, this protein is overexpressed in numerous human cancers, with expression levels often in correlation with cancer stages, as the presence of KDM7B is necessary for cancer cell survival, colonization, and growth [
      • Zhu G.
      • Liu L.
      • She L.
      • Tan H.
      • Wei M.
      • Chen C.
      • Su Z.
      • Huang D.
      • Tian Y.
      • Qiu Y.
      • et al.
      Elevated expression of histone demethylase PHF8 associates with adverse prognosis in patients of laryngeal and hypopharyngeal squamous cell carcinoma.
      ]. Furthermore, depletion of KDM7B in diseased cells induces apoptotic cell death and inhibits cancer colony formation and invasion, supporting the presentation of KDM7B as a therapeutic target [
      • Sun X.
      • Qiu J.J.
      • Zhu S.
      • Cao B.
      • Sun L.
      • Li S.
      • Li P.
      • Zhang S.
      • Dong S.
      Oncogenic features of PHF8 histone demethylase in esophageal squamous cell carcinoma.
      ]. Reported hit finding efforts for this protein predominantly target the JmjC domain with varying potency and selectivity [
      • Kaniskan H.Ü.
      • Martini M.L.
      • Jin J.
      Inhibitors of protein methyltransferases and demethylases.
      ]. Importantly, we included the KDM7B JmjC-PHD dual-domains within our panel of targets to identify compounds that bind at the interface of these domains. A compound that achieves this mode of binding would not rely on iron chelation, or need to mimic alpha-ketoglutarate like standard KDM inhibitors [
      • Suzuki T.
      • Ozasa H.
      • Itoh Y.
      • Zhan P.
      • Sawada H.
      • Mino K.
      • Walport L.
      • Ohkubo R.
      • Kawamura A.
      • Yonezawa M.
      • et al.
      Identification of the KDM2/7 histone lysine demethylase subfamily inhibitor and its antiproliferative activity.
      ,
      • Rose N.R.
      • Woon E.C.Y.
      • Tumber A.
      • Walport L.J.
      • Chowdhury R.
      • Li X.S.
      • King O.N.F.
      • Lejeune C.
      • Ng S.S.
      • Krojer T.
      • et al.
      Plant growth regulator daminozide is a selective inhibitor of human KDM2/7 histone demethylases.
      ]. This is also an attractive target due to the prevalence of PHD domains in Kme readers (more than 100), and the discovery of novel inhibitors for the JmjC-PHD dual-domains would aid in further understanding protein binding and function.
      Unlike with 53BP1, observation of the sequencing results from screening the KDM7B JmjC-PHD dual-domains using the Spotfire software failed to yield chemical moieties that were as prominent as the C6 and C23 synthons within 53BP1 screening. Therefore, to visualize the sequencing data and select compounds for follow-up synthesis, a combination of two different strategies was implemented. These strategies included observing compounds with the highest frequency count across selections and identifying singular synthons or di-synthon combinations that were most prevalent/attractive when summed across all bound compounds. It was observed that the B26 synthon was the most prevalent singular moiety, and the B28-C19 di-synthon was the most prevalent synthon combination (Supplemental Table 1). Compounds that contained these synthons were given priority in selecting compounds for off-DNA synthesis. Using this method, 29 compounds were selected and synthesized for the KDM7B protein (Supplemental Fig. 4). Although selection frequencies differ from 53BP1 and KDM7B, the top compounds were similar when observing enrichment percentages that are normalized against the total reads across each respective protein sample (Table 2 and 3).
      Table 3Putative hit compounds chosen for synthesis off-DNA based on KDM7B JmjC-PHD DEL selections. Corresponding selection frequency, normalized enrichment percentage, UNC ID, TR-FRET (IC50 in µM), and DSF (ΔTm in °C) results are presented. Normalized enrichment percentages were calculated by the following: (compound selection frequency/total reads) x 100. All TR-FRET and DSF data values are reported as the mean ± SEM of three technical replicates.
      Synthon CombinationSelection

      Frequency
      Normalized Enrichment %
      1,744 total reads, (d) = derivative that was not present in the original library.
      UNC CodeTR-FRET IC50Δ Tm
      A6-B15-C1580.46UNC919320.9 ± 2.41.9 ± 0.5
      A30-B26-C1070.40UNC9217> 100
      A25-B28-C1970.40UNC919095.6 ± 4.90.2 ± 0.3
      A23-B28-C1960.34UNC9198> 100
      A1-B21-C2460.34UNC9195> 100
      A7-B26-C3050.29UNC9208> 100
      A14-B16-C1350.29UNC9199> 100
      A18-B17-C3850.29UNC9196> 100
      A9-B24-C1850.29UNC9216> 100
      A22-B36-C2050.29UNC9202> 100
      A12-B34-C3150.29UNC9218> 100
      A8-B20-C3350.29UNC9213> 100
      A5-B28-C1940.23UNC9192> 100
      A7-B26-C3340.23UNC9209> 100
      A8-B18-C1840.23UNC921164.1 ± 10.40.8 ± 0.2
      A17-B20-C1140.23UNC9214> 100
      A1-B26-C1830.17UNC9194> 100
      A7-B26-C3530.17UNC9210> 100
      A20-B26-C3530.17UNC920058.1 ± 10.30.5 ± 0.3
      A24-B26-C3530.17UNC9207> 100
      A17-B25-C2030.17UNC9205> 100
      A17-B25-C3330.17UNC9206> 100
      A23-B28-C130.17UNC9197> 100
      A17-B21-C1130.17UNC9215> 100
      A33-B11-C3130.17UNC9191> 100
      A8-B26-C1820.11UNC921294.6 ± 7.01.0 ± 0.1
      A20-B26-C620.11UNC920130.0 ± 2.21.3 ± 0.5
      A17-B11-C11(d)--UNC9203> 100
      A17-B21-C11(d)--UNC9204> 100
      low asterisk 1,744 total reads, (d) = derivative that was not present in the original library.
      Using a TR-FRET assay, the synthesized compounds were tested for competitive binding against a KDM7B JmjC-PHD construct (Table 3). Of the 29 compounds tested, six compounds exhibited binding IC50 values <100 µM (Fig. 4A). Four of the six binding compounds contained one or more of the most prevalent synthons in B26, B28, or C19. The most potent binding compound was found to be UNC9193, with an IC50 value of 20.9 ± 0.2 µM. Interestingly, this compound did not contain one of the most prevalent synthons from the sequencing analysis but possessed the highest frequency count of all selected compounds. This data supports the methods implemented for KDM7B selection data as effective strategies for identifying hit compounds.
      Fig 4
      Fig. 4A) TR-FRET produced dose-response curves for the six binding compounds to the KDM7B JmjC-PHD dual-domains with IC50 values <100 µM. B) DSF experimental data for the same six hit compounds against the KDM7B JmjC-PHD, as well as protein and H3K4me3K9me2 peptide controls. C) Chemical structures and corresponding IC50 values (mean ± SEM) of the top three compounds identified from TR-FRET testing.
      In addition to TR-FRET, DSF was completed for all compounds that gave a binding IC50 <100 µM in TR-FRET (Fig. 4B). The controls included protein in buffer alone, and protein in buffer with 2% DMSO. These controls maintained similar protein melt temperatures with an average of 44.3 ± 0.1°C. The modified, unlabeled version of the H3K4me3K9me2 peptide utilized for the TR-FRET assay was tested as well and stabilized the protein with a thermal shift of 2.0 ± 0.4°C, more than all compounds tested. As expected, the top binding compounds from TR-FRET characterization, UNC9193 and UNC9201, were the top performing compounds with thermal shift values of 1.9 ± 0.5°C and 1.3 ± 0.5°C, respectively. Chemical structures of the top three binding compounds observed are shown in Fig. 4C. In looking at the structures of the top compounds, it was observed that none of the top compounds contained pharmacophores present in iron chelators that are common hit compounds within this protein family. Although modest potency was observed through TR-FRET and DSF characterization, screening our in-house designed UNCDEL003 yielded hit compounds for the KDM7B JmjC-PHD dual-domains for which potent small molecule hit discovery is evasive.

      Additional screening proteins: chromodomain Y-like protein 2 (CDYL2), chromobox protein homolog 2 (CBX2), and lens epithelium-derived growth factor (LEDGF)

      CDYL2 is a member of the chromodomain on Y (CDY) gene family that encodes three proteins containing unique and functional CDs: CDY1, CDYL1b, and CDYL2. These proteins contain a canonical N-terminal CD, which serves to recognize and bind methylated lysine residues present on histone proteins, as well as a crotonase-like catalytic domain on the C-terminus. Specifically, CDYL2 has been observed to recognize testis-specific H3tK27me3, canonical H3K27me3, and H3K9me2/3 and can act to recruit methyltransferases such as G9a and EZH2 [
      • Dong C.
      • Liu Y.
      • Lyu T.-J.
      • Beldar S.
      • Lamb K.N.
      • Tempel W.
      • Li Y.
      • Li Z.
      • James L.I.
      • Qin S.
      • et al.
      Structural basis for the binding selectivity of human CDY chromodomains.
      ,
      • Siouda M.
      • Dujardin A.D.
      • Barbollat-Boutrand L.
      • Mendoza-Parra M.A.
      • Gibert B.
      • Ouzounova M.
      • Bouaoud J.
      • Tonon L.
      • Robert M.
      • Foy J.-P.
      • et al.
      CDYL2 Epigenetically Regulates MIR124 to Control NF-κB/STAT3-Dependent Breast Cancer Cell Plasticity.
      ]. The protein plays a prominent role in spermatogenesis and has also been reported as a potential therapeutic target of interest due to its upregulation in various cancers such as breast and colorectal cancer [
      • Siouda M.
      • Dujardin A.D.
      • Barbollat-Boutrand L.
      • Mendoza-Parra M.A.
      • Gibert B.
      • Ouzounova M.
      • Bouaoud J.
      • Tonon L.
      • Robert M.
      • Foy J.-P.
      • et al.
      CDYL2 Epigenetically Regulates MIR124 to Control NF-κB/STAT3-Dependent Breast Cancer Cell Plasticity.
      ,
      • Kim S.T.
      • Sohn I.
      • Do I.-G.
      • Jang J.
      • Kim S.H.
      • Jung I.H.
      • Park J.O.
      • Park Y.S.
      • Talasaz A.
      • Lee J.
      • et al.
      Transcriptome analysis of CD133-positive stem cells and prognostic value of survivin in colorectal cancer.
      ]. Small molecule hit discovery for this protein class is limited, so successful hit compound discovery for CDYL2 would greatly benefit future inhibitor development and optimization [
      • Yang L.
      • Liu Y.
      • Fan M.
      • Zhu G.
      • Jin H.
      • Liang J.
      • Liu Z.
      • Huang Z.
      • Zhang L.
      Identification and characterization of benzo[d]oxazol-2(3H)-one derivatives as the first potent and selective small-molecule inhibitors of chromodomain protein CDYL.
      ].
      The CBX2 protein is a member of the chromodomain containing chromobox (CBX) protein family. The eight mammalian CBX protein homologs (CBX1-8) are related to the Drosophila HP1 (dHP1; CBX1, -3, -5 and Pc (dPc; CBX2, -4, -6, -7, -8) proteins that recognize and bind H3K9me3 and H3K27me3 PTM marks [
      • Fischle W.
      • Wang Y.
      • Jacobs S.A.
      • Kim Y.
      • Allis C.D.
      • Khorasanizadeh S.
      Molecular basis for the discrimination of repressive methyl-lysine marks in histone H3 by Polycomb and HP1 chromodomains.
      ]. The CBX2 protein homolog recognizes and binds the H3K27me3 via the CD domain [
      • Kaustov L.
      • Ouyang H.
      • Amaya M.
      • Lemak A.
      • Nady N.
      • Duan S.
      • Wasney G.A.
      • Li Z.
      • Vedadi M.
      • Schapira M.
      • et al.
      Recognition and specificity determinants of the human cbx chromodomains.
      ]. H3K27me3 is a transcriptionally repressive signal that is recognized by the Kme reader component of the multiprotein polycomb repressive complex 1 (PRC1) [
      • Stuckey J.I.
      • Dickson B.M.
      • Cheng N.
      • Liu Y.
      • Norris J.L.
      • Cholensky S.H.
      • Tempel W.
      • Qin S.
      • Huber K.G.
      • Sagum C.
      • et al.
      A cellular chemical probe targeting the chromodomains of Polycomb repressive complex 1.
      ]. CBX2 is one of the PRC1 CDs that can facilitate PRC1-mediated transcriptional repression by binding the complex to H3K27me3, allowing for monoubiquitylation of lysine 19 on histone 2A (H2AK119ub), which is associated with downstream gene repression, specifically in developmental genes [
      • Eskeland R.
      • Leeb M.
      • Grimes G.R.
      • Kress C.
      • Boyle S.
      • Sproul D.
      • Gilbert N.
      • Fan Y.
      • Skoultchi A.I.
      • Wutz A.
      • et al.
      Ring1B compacts chromatin structure and represses gene expression independent of histone ubiquitination.
      ,
      • Jin B.
      • Ernst J.
      • Tiedemann R.L.
      • Xu H.
      • Sureshchandra S.
      • Kellis M.
      • Dalton S.
      • Liu C.
      • Choi J.-H.
      • Robertson K.D.
      Linking DNA methyltransferases to epigenetic marks and nucleosome structure genome-wide in human tumor cells.
      ]. This protein has also been reported to have elevated expression levels in various cancers, and mutations within the protein have also been associated with human sexual development deficits [
      • Clermont P.L.
      • Sun L.
      • Crea F.
      • Thu K.L.
      • Zhang A.
      • Parolia A.
      • Lam W.L.
      • Helgason C.D.
      Genotranscriptomic meta-analysis of the Polycomb gene CBX2 in human cancers: initial evidence of an oncogenic role.
      ,
      • Sproll P.
      • Eid W.
      Biason-Lauber, A. CBX2-dependent transcriptional landscape: implications for human sex development and its defects.
      ]. Reported hit discovery for this protein includes both peptidomimetic and small molecule ligands that inhibit CBX2 binding to chromatin [
      • Wang S.
      • Alpsoy A.
      • Sood S.
      • Ordonez-Rubiano S.C.
      • Dhiman A.
      • Sun Y.
      • Jiao G.
      • Krusemark C.J.
      • Dykhuizen E.C.
      A potent, selective CBX2 chromodomain ligand and its cellular activity during prostate cancer neuroendocrine differentiation.
      ,
      • Lercher L.
      • Simon N.
      • Bergmann A.
      • Tauchert M.
      • Bochmann D.
      • Bashir T.
      • Neuefeind T.
      • Riley D.
      • Danna B.
      • Krawczuk P.
      • et al.
      Identification of Two Non-Peptidergic Small Molecule Inhibitors of CBX2 Binding to K27 Trimethylated Oligonucleosomes.
      ].
      The LEDGF protein (also known as p75, or PC4 and SFRS1-interacting protein 1 (PSIP1)), is a member of the hepatoma-derived growth factor (HDGF) related protein (HRP) family. LEDGF is tightly associated with chromatin at H3K36me throughout the cell cycle due to the N-terminal nuclear localization signal (NLS), a dual copy of the AT-hook DNA binding motif, and the PWWP domain that is conserved among this family of proteins [
      • Turlure F.
      • Maertens G.
      • Rahman S.
      • Cherepanov P.
      • Engelman A.
      A tripartite DNA-binding element, comprised of the nuclear localization signal and two AT-hook motifs, mediates the association of LEDGF/p75 with chromatin in vivo.
      ,
      • Llano M.
      • Vanegas M.
      • Hutchins N.
      • Thompson D.
      • Delgado S.
      • Poeschla E.M.
      Identification and characterization of the chromatin-binding domains of the HIV-1 integrase interactor LEDGF/p75.
      ]. This ubiquitous nuclear protein has been highlighted for the crucial role it plays in lentiviral integration, specifically human immunodeficiency virus (HIV) integration [
      • Botbol Y.
      • Raghavendra N.K.
      • Rahman S.
      • Engelman A.
      • Lavigne M.
      Chromatinized templates reveal the requirement for the LEDGF/p75 PWWP domain during HIV-1 integration in vitro.
      ,
      • Engelman A.
      • Cherepanov P.
      The lentiviral integrase binding protein LEDGF/p75 and HIV-1 replication.
      ]. In addition to its virological relevance, the PWWP domain of LEDGF is critical in rearranged mixed lineage leukemia (MLL) leukemogenesis, but dispensable for healthy hematopoiesis [
      • El Ashkar S.
      • Schwaller J.
      • Pieters T.
      • Goossens S.
      • Demeulemeester J.
      • Christ F.
      • Van Belle S.
      • Juge S.
      • Boeckx N.
      • Engelman A.
      • et al.
      LEDGF/p75 is dispensable for hematopoiesis but essential for MLL-rearranged leukemogenesis.
      ]. Currently, reported inhibitors of LEDGF to date target the integrase-binding domain (IBD), so hit discovery for the PWWP domain would be beneficial for the protein space [
      • Christ F.
      • Shaw S.
      • Demeulemeester J.
      • Desimmie B.A.
      • Marchand A.
      • Butler S.
      • Smets W.
      • Chaltin P.
      • Westby M.
      • Debyser Z.
      • et al.
      Small-molecule inhibitors of the LEDGF/p75 binding site of integrase block HIV replication and modulate integrase multimerization.
      ].
      Here, we included these three additional protein reader domains within our panel of targets to discover novel binders of the CD and PWWP domains to further understand the binding interactions of these domains and to promote future chemical probe development for functional analysis. As the LEDGF PWWP domain is a well-known binding partner of double stranded DNA, previously reported buffer conditions were adopted to mitigate non-specific DNA binding [
      • Botbol Y.
      • Raghavendra N.K.
      • Rahman S.
      • Engelman A.
      • Lavigne M.
      Chromatinized templates reveal the requirement for the LEDGF/p75 PWWP domain during HIV-1 integration in vitro.
      ]. In addition to demonstrating breadth across Kme readers of various classes, our DEL library screening approach has expanded utility in that it is amenable to the screening of more challenging targets, as these protein domains possess limited reported binders.
      In common with the KDM7B screening efforts, analysis of the experimental sequencing data for these protein domains using Spotfire failed to yield prominent chemical moieties for compound selection. Therefore, to identify compounds for follow-up synthesis, the selection strategies applied to KDM7B sequencing data were also applied to these proteins. Strategies involved selecting compounds from among those with the top frequency counts from sequencing as well as mining the data to look for synthons, or synthon combinations, that exhibited high cumulative frequency. Compounds that were synthetically tractable and that contained recognizable or enriched lysine mimetics were given priority over compounds with similar frequency counts (Supplemental Tables 2, 4, and 6). In total, 29, 22, and 12 compounds were selected and synthesized off-DNA using these selection strategies for CDYL2, CBX2, and LEDGF, respectively (Supplemental Figs. 5-7).
      The compounds were then tested against the respective CDs of CDYL2 and CBX2 using a TR-FRET competitive binding assay developed for each protein construct (Supplemental Tables 3 and 5). SPR experiments were utilized for LEDGF PWWP compound testing, as the native H3K36me 14-mer peptide binds the PWWP domain of LEDGF with a Kd of 17 mM, so a TR-FRET assay could not be established [
      • van Nuland R.
      • van Schaik F.M.
      • Simonis M.
      • van Heesch S.
      • Cuppen E.
      • Boelens R.
      • Timmers H.M.
      • van Ingen H.
      Nucleosomal DNA binding drives the recognition of H3K36-methylated nucleosomes by the PSIP1-PWWP domain.
      ]. Since SPR is a lower throughput assay, less compounds were selected for synthesis and testing (Supplemental Table 7). Unfortunately, none of the synthesized compounds produced binding curves with IC50 <100 µM for any of the three protein domains tested. Since no significant hits were obtained from both CDYL2 CD and CBX2 CD selections, this could suggest a difficulty in targeting CD pockets with our designed DEL. Although the UNCDEL003 allowed us to screen multiple challenging targets in which reported inhibitors are limited, these efforts yielded no binding compounds for these protein domains.

      Hit compound cross-screening for binding selectivity

      In addition to testing compounds chosen from DEL selections against their specific proteins, the top binding compounds from 53BP1 and KDM7B validation efforts were put through a binding selectivity panel against the other reader proteins used in this study. To assess compound binding, TR-FRET assays were utilized for all protein domains in which an assay was developed (53BP1 TTD, KDM7B JmjC-PHD, CDYL2 CD, and CBX2 CD). The results from the binding selectivity analysis are summarized in Table 4. Upon cross-screening, the compounds chosen from selections using the UNCDEL003 library were observed to be inactive against all other proteins used in this study.
      Table 4Binding selectivity analysis (IC50 in µM) for top hit compounds from 53BP1 and KDM7B DEL screening and characterization. All TR-FRET data values are reported as the mean ± SEM of three technical replicates (KDM7B) and the previously reported number of replicates (53BP1).
      Compound ID53BP1KDM7BCDYL2CBX2
      UNC76488.6 ± 1.0>100>100>100
      UNC760620.8 ± 1.2>100>100>100
      UNC758819.6 ± 2.3>100>100>100
      UNC764721.9 ± 1.7>100>100>100
      UNC759221.5 ± 2.5>100>100>100
      UNC758924.7 ± 2.6>100>100>100
      UNC764925.8 ± 3.0>100>100>100
      UNC9193>10020.9 ± 2.4>100>100
      UNC9201>10030.0 ± 2.2>100>100
      UNC9211>10064.1 ± 10.4>100>100
      UNC9200>10058.1 ± 10.3>100>100
      UNC9190>10095.6 ± 4.9>100>100
      UNC9212>10094.6 ± 7.0>100>100

      Conclusions

      In this study, we tested and reported on the applicability of our previously synthesized UNDEL003 library designed for chromatin reader protein hit discovery. For assessment, we screened the UNCDEL003 against five different Kme reader domain types (TTD, JmjC-PHD, PWWP, CD) from five chromatin reader proteins. All screened proteins exhibit a level of therapeutic relevance with varying human diseases and cancers. For all screened proteins, putative hit compounds were selected using methods including multi-dimensional graphics, individual compound frequency counts, and synthon combination prevalence. TR-FRET assays, or SPR experiments, were used for hit validation. As a result, prevalent hit compound sets were identified and validated with top compound IC50 values of 8.6 ± 1.0 µM and 20.9 ± 0.2 µM for the 53BP1 and KDM7B proteins, respectively. For these two targets, hits were further characterized by DSF and ITC, and confirmed to be selective when tested for binding against the other reader proteins used in the study. In addition to validating library applicability, these reported compounds can serve as viable starting points for subsequent hit optimization efforts.
      Although hit compounds were not discovered for all proteins tested herein, we report the focused UNCDEL003 as an accessible and facile HTS platform for screening against many additional chromatin reader proteins. As a validated screening method, the utility of our focused Kme library can be harnessed to aid in the expansion and development of ligand diversity within the Kme chemical space. This can be accomplished through further optimization of current hit compounds, as well as the construction of future DEL libraries based on previous selection data. In addition to the advancement of screening strategies, machine learning algorithms can be applied to DEL output data as a tool for selecting putative hit compounds for validation post-sequencing [
      • McCloskey K.
      • Sigel E.A.
      • Kearnes S.
      • Xue L.
      • Tian X.
      • Moccia D.
      • Gikunju D.
      • Bazzaz S.
      • Chan B.
      • Clark M.A.
      • et al.
      Machine Learning on DNA-Encoded Libraries: A New Paradigm for Hit Finding.
      ]. Developed algorithms can also greatly expand the observed chemical space through the screening of virtual compound libraries. In conjunction, these strategies may increase chemical diversity, expand screening space, and improve observed hit rates in future hit discovery campaigns.

      Funding

      Funding for this work was supported by the UNC Eshelman Institute for Innovation and the UNC Lineberger Comprehensive Cancer Center. The Structural Genomics Consortium is a registered charity (no. 1097737) that receives funds from Bayer AG, Boehringer Ingelheim, Bristol Myers Squibb, Genentech, Genome Canada through Ontario Genomics Institute (OGI-196), Janssen, Merck KGaA (aka EMD in Canada and USA), Pfizer, Takeda, and the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement no. 875510. The JU receives support from the European Union's Horizon 2020 research and innovation program, EFPIA, Ontario Institute for Cancer Research, Royal Institution for the Advancement of Learning McGill University, Kungliga Tekniska Hoegskolan, and Diamond Light Source Limited.

      CRediT authorship contribution statement

      Devan J. Shell: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization. Justin M. Rectenwald: Methodology, Validation, Formal analysis, Investigation, Resources. Peter H. Buttery: Methodology, Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing. Rebecca L. Johnson: Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing. Caroline A. Foley: Conceptualization, Methodology, Formal analysis, Investigation, Resources. Shiva K.R. Guduru: Methodology, Resources. Mélanie Uguen: Methodology, Resources, Writing – review & editing. Juanita Sanchez Rubiano: Methodology, Resources. Xindi Zhang: Methodology, Resources. Fengling Li: Methodology, Validation, Investigation, Resources. Jacqueline L. Norris-Drouin: Methodology, Resources, Writing – original draft. Matthew Axtman: Methodology, Resources. P. Brian Hardy: Methodology, Resources. Masoud Vedadi: Methodology, Resources, Supervision. Stephen V. Frye: Conceptualization, Methodology, Validation, Resources, Writing – review & editing, Supervision, Project administration, Funding acquisition. Lindsey I. James: Conceptualization, Methodology, Validation, Resources, Writing – review & editing, Supervision, Project administration, Funding acquisition. Kenneth H. Pearce: Conceptualization, Methodology, Validation, Formal analysis, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition.

      Declaration of Conflicting Interests

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgements

      We thank the Structural Genomics Consortium (CBX2 and LEDGF) and Nate Wesley (KDM7B) for providing these protein constructs used in this study.

      Appendix. Supplementary materials

      References

        • Brenner S.
        • Lerner R.A.
        Encoded combinatorial chemistry.
        Proc Natl Acad Sci USA. 1992; 89: 5381-5383https://doi.org/10.1073/pnas.89.12.5381
        • Favalli N.
        • Bassi G.
        • Scheuermann J.
        • Neri D.
        DNA-encoded chemical libraries - achievements and remaining challenges.
        FEBS Lett. 2018; 592: 2168-2180https://doi.org/10.1002/1873-3468.13068
        • Chamakuri S.
        • Lu S.
        • Ucisik M.N.
        • Bohren K.M.
        • Chen Y.-C.
        • Du H.-C.
        • Faver J.C.
        • Jimmidi R.
        • Li F.
        • Li J.-Y.
        • et al.
        DNA-encoded chemistry technology yields expedient access to SARS-CoV-2 Mpro inhibitors.
        Proc Natl Acad Sci USA. 2021; 118https://doi.org/10.1073/pnas.2111172118
        • Goodnow R.A.
        • Dumelin C.E.
        • Keefe A.D.
        DNA-encoded chemistry: enabling the deeper sampling of chemical space.
        Nat Rev Drug Discov. 2017; 16: 131-147https://doi.org/10.1038/nrd.2016.213
        • Clark M.A.
        • Acharya R.A.
        • Arico-Muendel C.C.
        • Belyanskaya S.L.
        • Benjamin D.R.
        • Carlson N.R.
        • Centrella P.A.
        • Chiu C.H.
        • Creaser S.P.
        • Cuozzo J.W.
        • et al.
        Design, synthesis and selection of DNA-encoded small-molecule libraries.
        Nat Chem Biol. 2009; 5: 647-654https://doi.org/10.1038/nchembio.211
        • Rectenwald J.M.
        • Guduru S.K.R.
        • Dang Z.
        • Collins L.B.
        • Liao Y.-E.
        • Norris-Drouin J.L.
        • Cholensky S.H.
        • Kaufmann K.W.
        • Hammond S.M.
        • Kireev D.B.
        • et al.
        Design and construction of a focused DNA-encoded library for multivalent chromatin reader proteins.
        Molecules. 2020; 25https://doi.org/10.3390/molecules25040979
        • Yuen L.H.
        • Dana S.
        • Liu Y.
        • Bloom S.I.
        • Thorsell A.-G.
        • Neri D.
        • Donato A.J.
        • Kireev D.
        • Schüler H.
        • Franzini R.M.
        A focused DNA-encoded chemical library for the discovery of inhibitors of NAD+-dependent enzymes.
        J Am Chem Soc. 2019; 141: 5169-5181https://doi.org/10.1021/jacs.8b08039
        • Dawadi S.
        • Simmons N.
        • Miklossy G.
        • Bohren K.M.
        • Faver J.C.
        • Ucisik M.N.
        • Nyshadham P.
        • Yu Z.
        • Matzuk M.M.
        Discovery of potent thrombin inhibitors from a protease-focused DNA-encoded chemical library.
        Proc Natl Acad Sci USA. 2020; 117: 16782-16789https://doi.org/10.1073/pnas.2005447117
        • Satz A.L.
        • Hochstrasser R.
        • Petersen A.C.
        Analysis of current DNA encoded library screening data indicates higher false negative rates for numerically larger libraries.
        ACS Comb Sci. 2017; 19: 234-238https://doi.org/10.1021/acscombsci.7b00023
        • Wagner T.
        • Robaa D.
        • Sippl W.
        • Jung M.
        Mind the methyl: methyllysine binding proteins in epigenetic regulation.
        ChemMedChem. 2014; 9: 466-483https://doi.org/10.1002/cmdc.201300422
        • Ruthenburg A.J.
        • Li H.
        • Patel D.J.
        • Allis C.D.
        Multivalent engagement of chromatin modifications by linked binding modules.
        Nat Rev Mol Cell Biol. 2007; 8: 983-994https://doi.org/10.1038/nrm2298
        • Mio C.
        • Bulotta S.
        • Russo D.
        • Damante G.
        Reading cancer: chromatin readers as druggable targets for cancer treatment.
        Cancers (Basel). 2019; 11https://doi.org/10.3390/cancers11010061
        • Barnash K.D.
        • James L.I.
        • Frye S.V.
        Target class drug discovery.
        Nat Chem Biol. 2017; 13: 1053-1056https://doi.org/10.1038/nchembio.2473
        • Tang J.
        • Cho N.W.
        • Cui G.
        • Manion E.M.
        • Shanbhag N.M.
        • Botuyan M.V.
        • Mer G.
        • Greenberg R.A.
        Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination.
        Nat Struct Mol Biol. 2013; 20: 317-325https://doi.org/10.1038/nsmb.2499
        • Tong Q.
        • Mazur S.J.
        • Rincon-Arano H.
        • Rothbart S.B.
        • Kuznetsov D.M.
        • Cui G.
        • Liu W.H.
        • Gete Y.
        • Klein B.J.
        • Jenkins L.
        • et al.
        An acetyl-methyl switch drives a conformational change in p53.
        Structure. 2015; 23: 322-331https://doi.org/10.1016/j.str.2014.12.010
        • Horton J.R.
        • Upadhyay A.K.
        • Qi H.H.
        • Zhang X.
        • Shi Y.
        • Cheng X.
        Enzymatic and structural insights for substrate specificity of a family of jumonji histone lysine demethylases.
        Nat Struct Mol Biol. 2010; 17: 38-43https://doi.org/10.1038/nsmb.1753
        • Dong C.
        • Liu Y.
        • Lyu T.-J.
        • Beldar S.
        • Lamb K.N.
        • Tempel W.
        • Li Y.
        • Li Z.
        • James L.I.
        • Qin S.
        • et al.
        Structural basis for the binding selectivity of human CDY chromodomains.
        Cell Chem Biol. 2020; 27 (e7): 827-838https://doi.org/10.1016/j.chembiol.2020.05.007
        • Kaustov L.
        • Ouyang H.
        • Amaya M.
        • Lemak A.
        • Nady N.
        • Duan S.
        • Wasney G.A.
        • Li Z.
        • Vedadi M.
        • Schapira M.
        • et al.
        Recognition and specificity determinants of the human cbx chromodomains.
        J Biol Chem. 2011; 286: 521-529https://doi.org/10.1074/jbc.M110.191411
        • Eidahl J.O.
        • Crowe B.L.
        • North J.A.
        • McKee C.J.
        • Shkriabai N.
        • Feng L.
        • Plumb M.
        • Graham R.L.
        • Gorelick R.J.
        • Hess S.
        • et al.
        Structural basis for high-affinity binding of LEDGF PWWP to mononucleosomes.
        Nucleic Acids Res. 2013; 41: 3924-3936https://doi.org/10.1093/nar/gkt074
        • Howarth M.
        • Takao K.
        • Hayashi Y.
        • Ting A.Y.
        Targeting quantum dots to surface proteins in living cells with biotin ligase.
        Proc Natl Acad Sci USA. 2005; 102: 7583-7588https://doi.org/10.1073/pnas.0503125102
        • Rectenwald J.M.
        • Hardy P.B.
        • Norris-Drouin J.L.
        • Cholensky S.H.
        • James L.I.
        • Frye S.V.
        • Pearce K.H.
        A general TR-FRET assay platform for high-throughput screening and characterizing inhibitors of methyl-lysine reader proteins.
        SLAS Discov. 2019; 24: 693-700https://doi.org/10.1177/2472555219844569
        • Barnash K.D.
        • Lamb K.N.
        • Stuckey J.I.
        • Norris J.L.
        • Cholensky S.H.
        • Kireev D.B.
        • Frye S.V.
        • James L.I.
        Chromodomain ligand optimization via target-class directed combinatorial repurposing.
        ACS Chem Biol. 2016; 11: 2475-2483https://doi.org/10.1021/acschembio.6b00415
        • Wigle T.J.
        • Herold J.M.
        • Senisterra G.A.
        • Vedadi M.
        • Kireev D.B.
        • Arrowsmith C.H.
        • Frye S.V.
        • Janzen W.P.
        Screening for inhibitors of low-affinity epigenetic peptide-protein interactions: an AlphaScreen-based assay for antagonists of methyl-lysine binding proteins.
        J Biomol Screen. 2010; 15: 62-71https://doi.org/10.1177/1087057109352902
        • Engelberg I.A.
        • Foley C.A.
        • James L.I.
        • Frye S.V.
        Improved methods for targeting epigenetic reader domains of acetylated and methylated lysine.
        Curr Opin Chem Biol. 2021; 63: 132-144https://doi.org/10.1016/j.cbpa.2021.03.002
        • Barnash K.D.
        • Lamb K.N.
        • James L.I.
        • Frye S.V.
        Peptide technologies in the development of chemical tools for chromatin-associated machinery.
        Drug Dev Res. 2017; 78: 300-312https://doi.org/10.1002/ddr.21398
        • James L.I.
        • Frye S.V.
        Chemical probes for methyl lysine reader domains.
        Curr Opin Chem Biol. 2016; 33: 135-141https://doi.org/10.1016/j.cbpa.2016.06.004
        • Mirza-Aghazadeh-Attari M.
        • Mohammadzadeh A.
        • Yousefi B.
        • Mihanfar A.
        • Karimian A.
        • Majidinia M.
        53BP1: a key player of DNA damage response with critical functions in cancer.
        DNA Repair (Amst). 2019; 73: 110-119https://doi.org/10.1016/j.dnarep.2018.11.008
        • Panier S.
        • Boulton S.J.
        Double-strand break repair: 53BP1 comes into focus.
        Nat Rev Mol Cell Biol. 2014; 15: 7-18https://doi.org/10.1038/nrm3719
        • Fradet-Turcotte A.
        • Canny M.D.
        • Escribano-Díaz C.
        • Orthwein A.
        • Leung C.C.Y.
        • Huang H.
        • Landry M.-C.
        • Kitevski-LeBlanc J.
        • Noordermeer S.M.
        • Sicheri F.
        • et al.
        53BP1 is a reader of the DNA-damage-induced H2A Lys 15 ubiquitin mark.
        Nature. 2013; 499: 50-54https://doi.org/10.1038/nature12318
        • Botuyan M.V.
        • Lee J.
        • Ward I.M.
        • Kim J.-E.
        • Thompson J.R.
        • Chen J.
        • Mer G.
        Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair.
        Cell. 2006; 127: 1361-1373https://doi.org/10.1016/j.cell.2006.10.043
        • Sun Y.
        • Lu H.
        • Fang X.
        • Xiao S.
        • Yang F.
        • Chen Y.
        • Wang H.
        • Li X.
        • Lu J.
        • Lin H.
        • et al.
        Discovery of a novel 53BP1 inhibitor through AlphaScreen-based high-throughput screening.
        Bioorg Med Chem. 2021; 34116054https://doi.org/10.1016/j.bmc.2021.116054
        • Perfetti M.T.
        • Baughman B.M.
        • Dickson B.M.
        • Mu Y.
        • Cui G.
        • Mader P.
        • Dong A.
        • Norris J.L.
        • Rothbart S.B.
        • Strahl B.D.
        • et al.
        Identification of a fragment-like small molecule ligand for the methyl-lysine binding protein, 53BP1.
        ACS Chem Biol. 2015; 10: 1072-1081https://doi.org/10.1021/cb500956g
        • Loenarz C.
        • Ge W.
        • Coleman M.L.
        • Rose N.R.
        • Cooper C.D.O.
        • Klose R.J.
        • Ratcliffe P.J.
        • Schofield C.J.
        PHF8, a gene associated with cleft lip/palate and mental retardation, encodes for an Nepsilon-dimethyl lysine demethylase.
        Hum Mol Genet. 2010; 19: 217-222https://doi.org/10.1093/hmg/ddp480
        • Chaturvedi S.S.
        • Ramanan R.
        • Waheed S.O.
        • Ainsley J.
        • Evison M.
        • Ames J.M.
        • Schofield C.J.
        • Karabencheva-Christova T.G.
        • Christov C.Z.
        Conformational dynamics underlies different functions of human KDM7 histone demethylases.
        Chem Eur J. 2019; 25: 5422-5426https://doi.org/10.1002/chem.201900492
        • Abidi F.E.
        • Miano M.G.
        • Murray J.C.
        • Schwartz C.E.
        A novel mutation in the PHF8 gene is associated with X-linked mental retardation with cleft lip/cleft palate.
        Clin Genet. 2007; 72: 19-22https://doi.org/10.1111/j.1399-0004.2007.00817.x
        • Qi H.H.
        • Sarkissian M.
        • Hu G.-Q.
        • Wang Z.
        • Bhattacharjee A.
        • Gordon D.B.
        • Gonzales M.
        • Lan F.
        • Ongusaha P.P.
        • Huarte M.
        • et al.
        Histone H4K20/H3K9 demethylase PHF8 regulates zebrafish brain and craniofacial development.
        Nature. 2010; 466: 503-507https://doi.org/10.1038/nature09261
        • Zhu G.
        • Liu L.
        • She L.
        • Tan H.
        • Wei M.
        • Chen C.
        • Su Z.
        • Huang D.
        • Tian Y.
        • Qiu Y.
        • et al.
        Elevated expression of histone demethylase PHF8 associates with adverse prognosis in patients of laryngeal and hypopharyngeal squamous cell carcinoma.
        Epigenomics. 2015; 7: 143-153https://doi.org/10.2217/epi.14.82
        • Sun X.
        • Qiu J.J.
        • Zhu S.
        • Cao B.
        • Sun L.
        • Li S.
        • Li P.
        • Zhang S.
        • Dong S.
        Oncogenic features of PHF8 histone demethylase in esophageal squamous cell carcinoma.
        PLoS One. 2013; 8: e77353https://doi.org/10.1371/journal.pone.0077353
        • Kaniskan H.Ü.
        • Martini M.L.
        • Jin J.
        Inhibitors of protein methyltransferases and demethylases.
        Chem Rev. 2018; 118: 989-1068https://doi.org/10.1021/acs.chemrev.6b00801
        • Suzuki T.
        • Ozasa H.
        • Itoh Y.
        • Zhan P.
        • Sawada H.
        • Mino K.
        • Walport L.
        • Ohkubo R.
        • Kawamura A.
        • Yonezawa M.
        • et al.
        Identification of the KDM2/7 histone lysine demethylase subfamily inhibitor and its antiproliferative activity.
        J Med Chem. 2013; 56: 7222-7231https://doi.org/10.1021/jm400624b
        • Rose N.R.
        • Woon E.C.Y.
        • Tumber A.
        • Walport L.J.
        • Chowdhury R.
        • Li X.S.
        • King O.N.F.
        • Lejeune C.
        • Ng S.S.
        • Krojer T.
        • et al.
        Plant growth regulator daminozide is a selective inhibitor of human KDM2/7 histone demethylases.
        J Med Chem. 2012; 55: 6639-6643https://doi.org/10.1021/jm300677j
        • Siouda M.
        • Dujardin A.D.
        • Barbollat-Boutrand L.
        • Mendoza-Parra M.A.
        • Gibert B.
        • Ouzounova M.
        • Bouaoud J.
        • Tonon L.
        • Robert M.
        • Foy J.-P.
        • et al.
        CDYL2 Epigenetically Regulates MIR124 to Control NF-κB/STAT3-Dependent Breast Cancer Cell Plasticity.
        iScience. 2020; 23101141https://doi.org/10.1016/j.isci.2020.101141
        • Kim S.T.
        • Sohn I.
        • Do I.-G.
        • Jang J.
        • Kim S.H.
        • Jung I.H.
        • Park J.O.
        • Park Y.S.
        • Talasaz A.
        • Lee J.
        • et al.
        Transcriptome analysis of CD133-positive stem cells and prognostic value of survivin in colorectal cancer.
        Cancer Genomics Proteomics. 2014; 11: 259-266
        • Yang L.
        • Liu Y.
        • Fan M.
        • Zhu G.
        • Jin H.
        • Liang J.
        • Liu Z.
        • Huang Z.
        • Zhang L.
        Identification and characterization of benzo[d]oxazol-2(3H)-one derivatives as the first potent and selective small-molecule inhibitors of chromodomain protein CDYL.
        Eur J Med Chem. 2019; 182111656https://doi.org/10.1016/j.ejmech.2019.111656
        • Fischle W.
        • Wang Y.
        • Jacobs S.A.
        • Kim Y.
        • Allis C.D.
        • Khorasanizadeh S.
        Molecular basis for the discrimination of repressive methyl-lysine marks in histone H3 by Polycomb and HP1 chromodomains.
        Genes Dev. 2003; 17: 1870-1881https://doi.org/10.1101/gad.1110503
        • Stuckey J.I.
        • Dickson B.M.
        • Cheng N.
        • Liu Y.
        • Norris J.L.
        • Cholensky S.H.
        • Tempel W.
        • Qin S.
        • Huber K.G.
        • Sagum C.
        • et al.
        A cellular chemical probe targeting the chromodomains of Polycomb repressive complex 1.
        Nat Chem Biol. 2016; 12: 180-187https://doi.org/10.1038/nchembio.2007
        • Eskeland R.
        • Leeb M.
        • Grimes G.R.
        • Kress C.
        • Boyle S.
        • Sproul D.
        • Gilbert N.
        • Fan Y.
        • Skoultchi A.I.
        • Wutz A.
        • et al.
        Ring1B compacts chromatin structure and represses gene expression independent of histone ubiquitination.
        Mol Cell. 2010; 38: 452-464https://doi.org/10.1016/j.molcel.2010.02.032
        • Jin B.
        • Ernst J.
        • Tiedemann R.L.
        • Xu H.
        • Sureshchandra S.
        • Kellis M.
        • Dalton S.
        • Liu C.
        • Choi J.-H.
        • Robertson K.D.
        Linking DNA methyltransferases to epigenetic marks and nucleosome structure genome-wide in human tumor cells.
        Cell Rep. 2012; 2: 1411-1424https://doi.org/10.1016/j.celrep.2012.10.017
        • Clermont P.L.
        • Sun L.
        • Crea F.
        • Thu K.L.
        • Zhang A.
        • Parolia A.
        • Lam W.L.
        • Helgason C.D.
        Genotranscriptomic meta-analysis of the Polycomb gene CBX2 in human cancers: initial evidence of an oncogenic role.
        Br J Cancer. 2014; 111: 1663-1672https://doi.org/10.1038/bjc.2014.474
        • Sproll P.
        • Eid W.
        Biason-Lauber, A. CBX2-dependent transcriptional landscape: implications for human sex development and its defects.
        Sci Rep. 2019; 9: 16552https://doi.org/10.1038/s41598-019-53006-7
        • Wang S.
        • Alpsoy A.
        • Sood S.
        • Ordonez-Rubiano S.C.
        • Dhiman A.
        • Sun Y.
        • Jiao G.
        • Krusemark C.J.
        • Dykhuizen E.C.
        A potent, selective CBX2 chromodomain ligand and its cellular activity during prostate cancer neuroendocrine differentiation.
        Chembiochem. 2021; 22: 2335-2344https://doi.org/10.1002/cbic.202100118
        • Lercher L.
        • Simon N.
        • Bergmann A.
        • Tauchert M.
        • Bochmann D.
        • Bashir T.
        • Neuefeind T.
        • Riley D.
        • Danna B.
        • Krawczuk P.
        • et al.
        Identification of Two Non-Peptidergic Small Molecule Inhibitors of CBX2 Binding to K27 Trimethylated Oligonucleosomes.
        SLAS Discov. 2022; 27: 306-313https://doi.org/10.1016/j.slasd.2022.04.003
        • Turlure F.
        • Maertens G.
        • Rahman S.
        • Cherepanov P.
        • Engelman A.
        A tripartite DNA-binding element, comprised of the nuclear localization signal and two AT-hook motifs, mediates the association of LEDGF/p75 with chromatin in vivo.
        Nucleic Acids Res. 2006; 34: 1653-1665https://doi.org/10.1093/nar/gkl052
        • Llano M.
        • Vanegas M.
        • Hutchins N.
        • Thompson D.
        • Delgado S.
        • Poeschla E.M.
        Identification and characterization of the chromatin-binding domains of the HIV-1 integrase interactor LEDGF/p75.
        J Mol Biol. 2006; 360: 760-773https://doi.org/10.1016/j.jmb.2006.04.073
        • Botbol Y.
        • Raghavendra N.K.
        • Rahman S.
        • Engelman A.
        • Lavigne M.
        Chromatinized templates reveal the requirement for the LEDGF/p75 PWWP domain during HIV-1 integration in vitro.
        Nucleic Acids Res. 2008; 36: 1237-1246https://doi.org/10.1093/nar/gkm1127
        • Engelman A.
        • Cherepanov P.
        The lentiviral integrase binding protein LEDGF/p75 and HIV-1 replication.
        PLoS Pathog. 2008; 4e1000046https://doi.org/10.1371/journal.ppat.1000046
        • El Ashkar S.
        • Schwaller J.
        • Pieters T.
        • Goossens S.
        • Demeulemeester J.
        • Christ F.
        • Van Belle S.
        • Juge S.
        • Boeckx N.
        • Engelman A.
        • et al.
        LEDGF/p75 is dispensable for hematopoiesis but essential for MLL-rearranged leukemogenesis.
        Blood. 2018; 131: 95-107https://doi.org/10.1182/blood-2017-05-786962
        • Christ F.
        • Shaw S.
        • Demeulemeester J.
        • Desimmie B.A.
        • Marchand A.
        • Butler S.
        • Smets W.
        • Chaltin P.
        • Westby M.
        • Debyser Z.
        • et al.
        Small-molecule inhibitors of the LEDGF/p75 binding site of integrase block HIV replication and modulate integrase multimerization.
        Antimicrob Agents Chemother. 2012; 56: 4365-4374https://doi.org/10.1128/AAC.00717-12
        • van Nuland R.
        • van Schaik F.M.
        • Simonis M.
        • van Heesch S.
        • Cuppen E.
        • Boelens R.
        • Timmers H.M.
        • van Ingen H.
        Nucleosomal DNA binding drives the recognition of H3K36-methylated nucleosomes by the PSIP1-PWWP domain.
        Epigenetics Chromatin. 2013; 6: 12https://doi.org/10.1186/1756-8935-6-12
        • McCloskey K.
        • Sigel E.A.
        • Kearnes S.
        • Xue L.
        • Tian X.
        • Moccia D.
        • Gikunju D.
        • Bazzaz S.
        • Chan B.
        • Clark M.A.
        • et al.
        Machine Learning on DNA-Encoded Libraries: A New Paradigm for Hit Finding.
        J Med Chem. 2020; 63: 8857-8866https://doi.org/10.1021/acs.jmedchem.0c00452