{
  "version": "2026-10-23-nurd-1",
  "scope": "Qualitative literature model, separate from measured RNA. No kinetic fit, exact molecular stoichiometry or patient-specific HbF prediction.",
  "sources": [
    {
      "id": "borg2010",
      "year": 2010,
      "source": "Borg et al. · Nature Genetics",
      "title": "Haploinsufficiency for the erythroid transcription factor KLF1 causes hereditary persistence of fetal hemoglobin",
      "url": "https://doi.org/10.1038/ng.630",
      "scope": "Maltese family, primary erythroid progenitors and functional assays.",
      "finding": "KLF1 supports adult β-globin and activates BCL11A; p.K288X haploinsufficiency connects these routes to familial HPFH."
    },
    {
      "id": "suzuki2013",
      "year": 2013,
      "source": "Suzuki et al. · Molecular and Cellular Biology",
      "title": "Disruption of the Hbs1l-Myb locus causes hereditary persistence of fetal hemoglobin in a mouse model",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3624254/",
      "scope": "Mouse genetics plus separate human CD34-derived erythroid knockdown experiments.",
      "finding": "Human MYB knockdown lowered KLF1, BCL11A and TR2/TR4 RNA. This supports upstream regulation, with species and cell-state dependence."
    },
    {
      "id": "sher2019",
      "year": 2019,
      "source": "Sher et al. · Nature Genetics",
      "title": "Rational targeting of a NuRD subcomplex guided by comprehensive in situ mutagenesis",
      "url": "https://doi.org/10.1038/s41588-019-0453-4",
      "scope": "HUDEP-2 screening, proteomics, primary human erythroid cells and mouse validation.",
      "finding": "CHD4, GATAD2A, HDAC2, MBD2 and MTA2 were HbF-repression hits. RBBP4 had a distinct cell-fitness requirement. CHD4 uncoupling can preserve other NuRD components."
    },
    {
      "id": "yu2019",
      "year": 2019,
      "source": "Yu et al. · Haematologica",
      "title": "Disruption of the MBD2-NuRD complex but not MBD3-NuRD induces high level HbF expression in human adult erythroid cells",
      "url": "https://haematologica.org/article/view/9160",
      "scope": "HUDEP-2 knockout/rescue; adult human CD34-derived erythroid cells.",
      "finding": "MBD2 disruption relieved silencing. Its coiled-coil and disordered-region interfaces connect remodeling and deacetylase modules; MBD3 loss did not reproduce the MBD2 effect."
    },
    {
      "id": "shang2023",
      "year": 2023,
      "source": "Shang et al. · PNAS",
      "title": "MBD2a–NuRD binds to the methylated γ-globin gene promoter and uniquely forms a complex required for silencing of HbF expression",
      "url": "https://www.pnas.org/doi/10.1073/pnas.2302254120",
      "scope": "HUDEP-2 and primary CD34-derived erythroid cells; occupancy and chromatin assays.",
      "finding": "MBD2a, BCL11A and PRMT5 associate at the promoter. Loss of MBD2 reduces a promoter nucleosome and permits NF-Y binding."
    },
    {
      "id": "zheng2024",
      "year": 2024,
      "source": "Zheng et al. · Science",
      "title": "A tetramer of BCL11A is required for stable protein production and fetal hemoglobin silencing",
      "url": "https://doi.org/10.1126/science.adp3025",
      "scope": "Biochemical, structural and erythroid-cell experiments.",
      "finding": "BCL11A tetramerization supports protein stability and NuRD engagement. Its N-terminal sequence binds RBBP4; DNA binding and corepressor engagement are distinct functions."
    },
    {
      "id": "masuda2016",
      "year": 2016,
      "source": "Masuda et al. · Science",
      "title": "Transcription factors LRF and BCL11A independently repress expression of fetal hemoglobin",
      "url": "https://pubmed.ncbi.nlm.nih.gov/26816381/",
      "scope": "Human erythroid cells and mouse models.",
      "finding": "LRF/ZBTB7A supplies a parallel NuRD-associated γ-globin repression route, independent of BCL11A."
    },
    {
      "id": "liu2021",
      "year": 2021,
      "source": "Liu et al. · Nature Genetics",
      "title": "Transcription factor competition at the γ-globin promoters controls hemoglobin switching",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8038971/",
      "scope": "Human erythroid perturbations and promoter occupancy experiments.",
      "finding": "BCL11A binding near −115 opposes NF-Y access to the proximal CCAAT region. Promoter coordinates are relative to the transcription start site."
    },
    {
      "id": "bradner2010",
      "year": 2010,
      "source": "Bradner et al. · PNAS",
      "title": "Chemical genetic strategy identifies histone deacetylase 1 (HDAC1) and HDAC2 as therapeutic targets in sickle cell disease",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2906581/",
      "scope": "Chemical profiling and human erythroid experiments.",
      "finding": "HDAC1/2 inhibition can induce fetal globin. Catalytic inhibition alone does not establish physical disassembly of NuRD."
    },
    {
      "id": "architecture2020",
      "year": 2020,
      "source": "Low et al. · Cell Reports",
      "title": "The nucleosome remodeling and deacetylase complex has an asymmetric, dynamic, and modular architecture",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8908386/",
      "scope": "Purified mammalian complexes and structural/biochemical analyses.",
      "finding": "NuRD combines separable MTA–HDAC–RBBP and MBD–GATAD2–CHD modules. Subcomplexes can exist outside intact NuRD; no single assembly timeline follows from this."
    },
    {
      "id": "spruijt2021",
      "year": 2021,
      "source": "Spruijt et al. · The FEBS Journal",
      "title": "Cross-linking mass spectrometry reveals the structural topology of peripheral NuRD subunits relative to the core complex",
      "url": "https://doi.org/10.1111/febs.15650",
      "scope": "Cross-linking proteomics and biochemical mapping; not a family HbF experiment.",
      "finding": "CDK2AP1 and CDK2AP2 associate through CHD C-terminal regions. They extend the architecture beyond the six major subunit families."
    }
  ],
  "components": [
    {
      "id": "MYB",
      "name": "MYB",
      "role": "Upstream erythroid regulator",
      "text": "Supports the KLF1–BCL11A pathway in human erythroid experiments. MYB also influences proliferation and differentiation, so this is one branch of a wider programme.",
      "detail": "The arrow denotes regulation supported by knockdown. It does not assign a universal dose response or prove direct MYB binding at KLF1 in these family cells.",
      "genes": [
        "MYB"
      ],
      "refs": [
        "suzuki2013"
      ]
    },
    {
      "id": "KLF1",
      "name": "KLF1",
      "role": "Two arms of the adult programme",
      "text": "Activates HBB and supports BCL11A expression. Reduced functional KLF1 can therefore affect adult β-globin support and the fetal-globin repression pathway together.",
      "detail": "FamF p.K288X is represented as haploinsufficiency. Total KLF1 RNA cannot distinguish functional dosage, allelic output or DNA-binding activity.",
      "genes": [
        "KLF1"
      ],
      "refs": [
        "borg2010"
      ]
    },
    {
      "id": "BCL11A",
      "name": "BCL11A",
      "role": "DNA recognition and corepressor engagement",
      "text": "Binds the γ-promoter and engages NuRD. These are separate functions: its N-terminal NuRD-binding sequence contacts RBBP4, while C-terminal zinc fingers recognize DNA.",
      "detail": "Tetramerization via ZnF0 supports stable protein and effective repression. The single symbol here represents the factor, not one molecule. BCL11A loss is not equivalent to disassembling every NuRD complex.",
      "genes": [
        "BCL11A"
      ],
      "refs": [
        "zheng2024",
        "liu2021"
      ]
    },
    {
      "id": "MTA2",
      "name": "MTA scaffold",
      "role": "Organizes the deacetylase core",
      "text": "MTA proteins help organize HDAC and RBBP subunits. The erythroid focus is MTA2; MTA1 and MTA3 are related alternatives.",
      "detail": "Paralogs are listed for architectural completeness, not as interchangeable HbF regulators. The figure shows subunit families rather than exact stoichiometry.",
      "genes": [
        "MTA2",
        "MTA1",
        "MTA3"
      ],
      "refs": [
        "architecture2020",
        "sher2019"
      ]
    },
    {
      "id": "HDAC2",
      "name": "HDAC enzymes",
      "role": "Histone deacetylation",
      "text": "HDAC1/2 remove acetyl groups from histones. This enzymatic activity is distinct from ATP-dependent nucleosome remodeling.",
      "detail": "HDAC2 was the HbF-repression hit in the 2019 dense screen. Other human erythroid experiments support effects of HDAC1/2 inhibition; assay context matters.",
      "genes": [
        "HDAC2",
        "HDAC1"
      ],
      "refs": [
        "bradner2010",
        "sher2019"
      ]
    },
    {
      "id": "RBBP4",
      "name": "RBBP histone binders",
      "role": "Core organization and regulator contacts",
      "text": "RBBP4/7 are histone-binding core components. RBBP4 can also bind the NuRD-interacting N-terminus of BCL11A.",
      "detail": "RBBP4 was a cell-fitness hit, not an equivalent HbF-repression hit, in the Sher screen. Reducing its RNA is not a specific readout of γ-globin derepression.",
      "genes": [
        "RBBP4",
        "RBBP7"
      ],
      "refs": [
        "zheng2024",
        "sher2019"
      ]
    },
    {
      "id": "MBD2",
      "name": "MBD connector",
      "role": "Methylated DNA reading and module coupling",
      "text": "MBD2 links the deacetylase core to GATAD2. MBD2a has an additional N-terminal region involved in the promoter-associated repressor complex.",
      "detail": "MBD2 and MBD3 define alternative NuRD complexes. MBD3 does not substitute for MBD2 in the cited adult erythroid HbF-silencing experiments; MBD2a and MBD2b also differ.",
      "genes": [
        "MBD2",
        "MBD3"
      ],
      "refs": [
        "yu2019",
        "shang2023"
      ]
    },
    {
      "id": "GATAD2A",
      "name": "GATAD2 bridge",
      "role": "Connects MBD to the remodeler",
      "text": "GATAD2A/B bridge MBD and CHD proteins. The erythroid focus is GATAD2A; its interfaces provide two distinct places to uncouple remodeling.",
      "detail": "MBD2–GATAD2 association involves a coiled-coil. GATAD2 contacts the CHD4 C-terminal region through a separate interface.",
      "genes": [
        "GATAD2A",
        "GATAD2B"
      ],
      "refs": [
        "yu2019",
        "spruijt2021"
      ]
    },
    {
      "id": "CHD4",
      "name": "CHD remodeler",
      "role": "ATP-dependent nucleosome remodeling",
      "text": "CHD proteins move or reorganize nucleosomes using ATP. CHD4 is the focus of the adult erythroid model; CHD3 and CHD5 are family alternatives.",
      "detail": "Sequestering CHD4 from NuRD can leave other subunits associated. That is a change in complex coupling, not necessarily CHD4 degradation.",
      "genes": [
        "CHD4",
        "CHD3",
        "CHD5"
      ],
      "refs": [
        "sher2019",
        "spruijt2021"
      ]
    },
    {
      "id": "CDK2AP1",
      "name": "CDK2AP partners",
      "role": "Peripheral remodeling-associated components",
      "text": "CDK2AP1 (DOC1) and CDK2AP2 associate with the CHD-containing side of NuRD. Their presence extends the six-family core map.",
      "detail": "Evidence here is structural and biochemical. No specific CDK2AP-dependent explanation of the FamF or Greek HU response is assigned.",
      "genes": [
        "CDK2AP1",
        "CDK2AP2"
      ],
      "refs": [
        "spruijt2021"
      ]
    },
    {
      "id": "PRMT5",
      "name": "PRMT5 / MEP50",
      "role": "Associated repression machinery",
      "text": "The MBD2a-associated promoter complex can include PRMT5 and its partner MEP50 (WDR77). They are associated factors, rather than one of the six major NuRD families.",
      "detail": "The atlas separates these partners from core NuRD architecture. Their occupancy and catalytic state were not measured in the supplied expression matrices.",
      "genes": [
        "PRMT5",
        "WDR77"
      ],
      "refs": [
        "shang2023"
      ]
    },
    {
      "id": "ZBTB7A",
      "name": "LRF / ZBTB7A",
      "role": "A parallel repression route",
      "text": "LRF can recruit a NuRD-associated repressive programme independently of BCL11A. Removing one route does not erase every source of γ-globin repression.",
      "detail": "The LRF promoter region is near −200; the BCL11A region is near −115. The schematic does not combine them into a single obligatory complex.",
      "genes": [
        "ZBTB7A"
      ],
      "refs": [
        "masuda2016",
        "liu2021"
      ]
    }
  ],
  "scenarios": [
    {
      "id": "intact",
      "label": "Intact repression",
      "category": "REFERENCE",
      "title": "Coupling, recruitment and activity cooperate",
      "summary": "The assembled complex supports the adult γ-globin repression route. Other activators and repressors also shape the final output.",
      "assembly": "Modules coupled",
      "recruitment": "Promoter-associated route",
      "activity": "Deacetylation + remodeling",
      "gamma": "Repression supported",
      "beta": "KLF1 → HBB support retained",
      "refs": [
        "architecture2020",
        "shang2023"
      ]
    },
    {
      "id": "myb",
      "label": "Reduce MYB",
      "category": "UPSTREAM",
      "title": "An upstream change reaches more than one branch",
      "summary": "Human MYB knockdown reduced KLF1/BCL11A and TR2/TR4 expression. NuRD assembly itself was not established by that result.",
      "assembly": "Not determined",
      "recruitment": "BCL11A route may weaken",
      "activity": "Not determined",
      "gamma": "Release of repression is supported in the tested cells",
      "beta": "KLF1-mediated support may change",
      "refs": [
        "suzuki2013"
      ]
    },
    {
      "id": "klf1",
      "label": "KLF1 insufficiency",
      "category": "FAMILY CONNECTION",
      "title": "One regulator connects adult activation and fetal repression",
      "summary": "Reduced functional KLF1 can weaken BCL11A activation and adult β-globin support. This does not assign the cause of variation between individual carriers.",
      "assembly": "Not determined",
      "recruitment": "Indirect BCL11A route",
      "activity": "Not determined",
      "gamma": "Reduced repression is mechanistically supported",
      "beta": "Direct KLF1 → HBB support reduced",
      "refs": [
        "borg2010"
      ]
    },
    {
      "id": "bcl11a",
      "label": "Remove BCL11A",
      "category": "RECRUITMENT",
      "title": "Remove the recruiter, retain the assembly distinction",
      "summary": "BCL11A depletion removes its DNA-bound repression route and permits NF-Y access. NuRD is not assumed to disassemble globally; independent LRF-associated repression remains.",
      "assembly": "No disassembly assumed",
      "recruitment": "BCL11A-dependent route lost",
      "activity": "Core enzymes not assumed inactive",
      "gamma": "This repressive route is weakened",
      "beta": "No direct KLF1 perturbation",
      "refs": [
        "liu2021",
        "zheng2024",
        "masuda2016"
      ]
    },
    {
      "id": "mbd_bridge",
      "label": "Break MBD–GATAD2",
      "category": "UNCOUPLING",
      "title": "Separate the remodeling arm at its MBD interface",
      "summary": "A disrupted MBD2 coiled-coil fails to support the GATAD2/CHD4 connection and normal silencing in rescue experiments.",
      "assembly": "Bridge to remodeling arm disrupted",
      "recruitment": "Not equivalent to BCL11A loss",
      "activity": "Coupled remodeling impaired",
      "gamma": "Silencing weakened in tested cells",
      "beta": "No direct KLF1 perturbation",
      "refs": [
        "yu2019"
      ]
    },
    {
      "id": "mbd_core",
      "label": "Break MBD–core",
      "category": "UNCOUPLING",
      "title": "Detach the connector from the deacetylase core",
      "summary": "MBD2 disordered-region mutations impair association with the histone deacetylase core. Protein presence alone does not rescue the missing connection.",
      "assembly": "Core–connector interface disrupted",
      "recruitment": "Promoter occupancy not computed",
      "activity": "Coordinated core function impaired",
      "gamma": "Silencing weakened in tested cells",
      "beta": "No direct KLF1 perturbation",
      "refs": [
        "yu2019"
      ]
    },
    {
      "id": "chd",
      "label": "Detach CHD4",
      "category": "UNCOUPLING",
      "title": "The remodeler separates; other subunits can remain together",
      "summary": "CHD4-interface perturbations and sequestration experiments support selective uncoupling of the ATP-dependent remodeling arm.",
      "assembly": "CHD4 separated from remaining subunits",
      "recruitment": "Remaining core need not disappear",
      "activity": "Coupled remodeling impaired",
      "gamma": "HbF derepression supported experimentally",
      "beta": "No direct KLF1 perturbation",
      "refs": [
        "sher2019"
      ]
    },
    {
      "id": "hdac",
      "label": "Inhibit HDAC activity",
      "category": "CATALYSIS",
      "title": "Turn down an enzyme without declaring the complex dismantled",
      "summary": "The model suppresses deacetylase activity while retaining the assembly drawing. This is a conceptual distinction; a particular inhibitor may have additional effects.",
      "assembly": "No disassembly assumed",
      "recruitment": "No occupancy change assumed",
      "activity": "Deacetylation inhibited",
      "gamma": "Fetal-globin induction supported in HDAC studies",
      "beta": "No direct KLF1 perturbation",
      "refs": [
        "bradner2010"
      ]
    },
    {
      "id": "hu",
      "label": "Inspect HU evidence",
      "category": "OBSERVED RNA",
      "title": "The recovered HU arrays ask which genes respond",
      "summary": "PC → PH changes measure RNA. They do not establish NuRD protein abundance, assembly, promoter occupancy or enzymatic activity. Use the paired plots below to inspect each gene.",
      "assembly": "Not measured",
      "recruitment": "Not measured",
      "activity": "Not measured",
      "gamma": "Inspect probe-level changes",
      "beta": "Inspect probe-level changes",
      "refs": []
    }
  ],
  "steps": [
    {
      "title": "Meet the modules",
      "text": "Start with separable deacetylase and remodeling components. This arrangement is a teaching view, not a measured assembly intermediate."
    },
    {
      "title": "Connect the machinery",
      "text": "MBD and GATAD2 couple the core to CHD. Interactions are reversible; this is not a claim that every complex assembles in this order."
    },
    {
      "title": "Engage the promoter",
      "text": "BCL11A and MBD2a-associated interactions help localize a repressive complex. NuRD assembly and locus recruitment are distinct layers."
    },
    {
      "title": "Support repression",
      "text": "Histone deacetylation and nucleosome remodeling help maintain a repressive promoter state. NF-Y access is constrained in the cited adult erythroid model."
    },
    {
      "title": "Change one layer",
      "text": "Choose an evidence-linked perturbation. The view separates upstream regulation, recruitment, physical coupling and catalytic activity."
    }
  ]
}