CHAF1B (CAF-1 p60) - the one histone chaperone in this beat with a real leukaemia literature, and a mouse experiment that contradicts DepMap's verdict on its druggability
also called CHAF1B · CAF1A · CAF1P60 · MPHOSPH7 · MPP7 · CAF-1 p60 · chromatin assembly factor 1 subunit B · Q13112
CHAF1B is part of the machine that packages DNA right behind the copying fork, and it is the only protein in this group that leukaemia researchers have actually studied. Too much of it blocks blood cells from maturing and drives leukaemia; halving it in mice stops the leukaemia without harming normal blood.
The same thing, for a scientist
CHAF1B (Q13112, 559 aa) is the p60 subunit of CAF-1, the replication-coupled H3.1/H3.2 chaperone - the **contrast case** to the replication-independent HIRA route, and therefore the other half of the clearance-clock axis. Unlike every other gene in this beat it has a substantial, mechanistic leukaemia literature: - **CHAF1B is required for normal haematopoiesis and its overexpression promotes leukaemia**, acting by binding chromatin at discrete sites and interfering with the occupancy of transcription factors that promote myeloid differentiation, such as CEBPA [E1]. - **Reducing Chaf1b by heterozygous deletion or by a CAF-1 dominant-negative allele suppresses leukaemogenesis in vivo without impairing normal haematopoiesis** [E1]. That is a therapeutic window, stated as an in-vivo result. - A follow-up identifies **TRIM13** as a CHAF1B-repressed target and the RNA-seq that found it was run on **mouse MLL-AF9 leukaemic cells** plus paediatric AML aspirates [E2] - the only direct KMT2A-fusion connection any chaperone in this beat has. **And DepMap says the opposite about the window.** CHAF1B is the most pan-essential gene in this entire beat, **-2.314 across 1,208 lines**, with an index arm at -2.473 and comparator at -2.286 [E3]. A CRISPR knockout kills everything; a heterozygous mouse is healthy. **Both are true, and the disagreement is about dose, not about biology** - which is precisely the scope objection this graph's database-facts rule exists for.
The dose argument, because it is the most useful thing in this beat
A complete CRISPR knockout removes the protein. A heterozygous deletion halves it. For an abundant, processive assembly factor those are not the same experiment, and the gap between them is where a drug would live.
| observation | what it measures | verdict on druggability |
|---|---|---|
| DepMap Chronos -2.31 across 1,208 lines 3 | complete loss of function over ~2 weeks in proliferating lines | no window; pan-essential |
| Chaf1b heterozygous deletion suppresses leukaemogenesis in vivo, normal haematopoiesis intact 1 | 50% reduction, whole animal, physiological timescales | window exists at partial occupancy |
| CAF-1 dominant-negative allele, same result 1 | partial functional block rather than gene loss | window exists |
This is the cleanest worked example in the graph of why a pan-essentiality score is not a safety verdict, and it is the mirror of the gnomAD pLI objection already in the standing brief. It should be read alongside any entry in this project that dismisses a target as pan-essential - including gene-supt6h, gene-chaf1a and the FACT subunits in this same beat.
What it does not show. It is a mouse, the leukaemia models are myeloid, and no KMT2A::AFF1 model appears anywhere. It also does not show that a small molecule could achieve a stable 50% functional reduction, which is a much harder pharmacological ask than a germline heterozygote.
Why CAF-1 belongs in a beat about a mark with no eraser
H3K79 methylation is lost only when the histone carrying it is replaced (mechanism-h3k79-turnover-and-dilution). There are exactly two ways that happens, and CAF-1 runs one of them:
- Replication-coupled: at the fork, CAF-1 deposits new H3.1/H3.2-H4 behind the polymerase, diluting parental marks two-fold per division. Rate set by division rate.
- Replication-independent: HIRA and ASF1B deposit H3.3 at transcribed chromatin. Rate set by local turnover.
The clinical question the graph already poses - how many divisions of DOT1L inhibition are needed to clear H3K79me2 - assumes the first route dominates. If the second dominates at fusion target genes, the answer is not a number of divisions at all. mechanism-nucleosome-turnover-clearance-clock-machinery sets out what is known about which, and the answer is that a postmitotic cell still replaces its histones with a half-life of about two weeks, so the assumption is wrong in at least one well-measured tissue.
A cell-cycle-gated subnuclear address, and a second one that carries the leukaemia (Q87, 2026-09-21)
Subcellular. Nucleus in interphase, cytoplasm in M phase (UniProt Q13112, ECO:0000269 PubMed:9614144). HPA v25.1 IF sees only nucleoplasm, because it images asynchronous cells.
Subnuclear, and it is gated by the cycle. Sites of intranuclear DNA replication during S phase; predominantly dissociated from chromatin during mitosis. A single steady-state answer for CHAF1B is wrong.
A second chromatin address, and it is the leukaemic one. In AML, CHAF1B binds chromatin at discrete sites and blocks the occupancy of myeloid differentiation transcription factors including CEBPA (PMID 30423293) - this graph also holds a gene-cebpa entry, and the two are now linked.
Cell type and state. AML blasts, with a genuine in-vivo window against normal haematopoiesis - obtained by REDUCING Chaf1b, which bounds a dose rather than validating a target.
Shown for fusion. No. The single fusion-scoped localisation hit was retrieved and rejected, and the strong leukaemia genetics is AML without a fusion restriction.
| uniprot | Q13112 |
|---|---|
| gene symbol | CHAF1B |
| length AA | 559 |
| complex | CAF-1 (CHAF1A p150, CHAF1B p60, RBBP4 p48) |
| deposition mode | replication-coupled |
| cargo | H3.1/H3.2-H4 |
| therapeutic window evidence | heterozygous Chaf1b deletion or a CAF-1 dominant-negative allele suppresses leukaemogenesis in vivo without impairing normal haematopoiesis (PMID 30423293) |
| alias hazard | CAF1A is a curated alias symbol of CHAF1B and also the informal name of the other subunit CHAF1A (Q13111, 956 aa) |
| subnuclear localization | REPLICATION FOCI in S phase - CAF-1 p60 concentrates at sites of intranuclear DNA replication and predominantly dissociates from chromatin in mitosis (PMID 9614144) - plus, in AML specifically, DISCRETE CHROMATIN SITES where CHAF1B competes with myeloid differentiation transcription factors including CEBPA (PMID 30423293). The address is cell-cycle-gated, so any single steady-state answer is wrong; HPA's diffuse 'nucleoplasm' call is what an asynchronous population averages to. |
| cell state | TWO STATES, and they are different kinds of thing. (1) S PHASE - the replication-foci address exists only there, and CAF-1 dissociates from chromatin in mitosis (PMID 9614144). (2) THE UNDIFFERENTIATED, SELF-RENEWING AML BLAST STATE - CHAF1B is described as a master epigenetic regulator associated with self-renewal and the undifferentiated state of AML blasts, repressing differentiation factors and tumour suppressors (PMID 37205848), and mechanistically it blocks CEBPA occupancy (PMID 30423293). |
| shown for fusion | NO for localisation, and the AML work is NOT fusion-scoped. Measured 2026-09-21: (CHAF1B[tiab] OR "CAF-1 p60"[tiab] OR p60[tiab]) AND the eight-term fusion clause AND the ten-term localisation clause returns 1 record, PMID 37205848, which was retrieved and read: it is a CHAF1B/TRIM13 repression study in AML generally, contains no localisation experiment, and names no KMT2A fusion - the 'MLL' in reach of the query is not a fusion claim about this result. Screened count on topic: 0. The strong leukaemia genetics (PMID 30423293) is AML without a fusion restriction. |
depmap
{
"index_mean": -2.4728,
"other_leukaemia_mean": -2.2862,
"difference": -0.1866,
"pan_essentiality": -2.3136,
"myeloid_difference": -0.1812,
"lymphoid_difference": -0.1642,
"n_index": 11,
"interpretation": "floor-compressed; difference not interpretable as selectivity"
}depmap kmt2a selectivity
{
"chronos_mean_kmt2a_rearranged": -2.4728,
"chronos_mean_other_leukaemia": -2.2862,
"difference": -0.1866,
"pan_essentiality_all_lines": -2.3136,
"selectivity_rank": 291,
"rank_denominator": 17728,
"provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
"how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}localization
{
"subcellular": "NUCLEUS in interphase, CYTOPLASM in M phase - a cell-cycle-switched distribution, not a fixed one. UniProt Q13112 lists Nucleus and Cytoplasm, both ECO:0000269 PubMed:9614144, with the note 'DNA replication foci. Cytoplasmic in M phase'. The cited paper really is the paper for this fact (rule 91 satisfied): HeLa cells synchronised in G1, S, G2 and M show p150 and p60 nucleus-bound in interphase and predominantly dissociated from chromatin during mitosis. HPA v25.1 IF, reliability Supported, gives Nucleoplasm as the single main location - an asynchronous population, which is why it sees only the interphase answer.",
"subnuclear": "SITES OF INTRANUCLEAR DNA REPLICATION - replication foci - and only during S phase. CAF-1 p150 and p60 concentrate at replication foci in S phase and predominantly dissociate from chromatin during mitosis (PMID 9614144). This is a CELL-CYCLE-GATED subnuclear address: the compartment exists for part of the cycle and not at all for another part, so a single steady-state answer for CHAF1B would be wrong, and an asynchronous immunofluorescence population averages the two into a diffuse nucleoplasmic call - which is exactly what HPA reports.",
"the_second_chromatin_address_is_leukaemia_specific": "Beyond the replication fork, CHAF1B in AML binds chromatin at DISCRETE SITES and interferes with the occupancy of transcription factors that drive myeloid differentiation, CEBPA among them (PMID 30423293). That is a second, transcription-factor-competition address distinct from the replication-coupled one, and it is the one that carries the leukaemia phenotype.",
"how_shown": "cell-cycle synchronisation of HeLa cells into G1/S/G2/M with subcellular fractionation and nucleosome-assembly assays (PMID 9614144); chromatin occupancy plus genetic dose reduction in mouse haematopoiesis and leukaemia (PMID 30423293); HPA antibody IF on asynchronous cells",
"node": "localization-chromatin-bound-versus-soluble-fraction-kmt2a"
}cell type
{
"haematopoietic_answer": "AML BLASTS, with a genuine therapeutic window against normal haematopoiesis. CHAF1B is required for normal haematopoiesis, its overexpression promotes leukaemia, and reducing Chaf1b activity by heterozygous deletion or a CAF1 dominant-negative allele suppresses leukaemogenesis in vivo WITHOUT impairing normal haematopoiesis (PMID 30423293, mouse). CHAF1B is described as upregulated in almost all AML samples (PMID 37205848).",
"expression_atlas": "HPA v25.1: RNA tissue specificity 'Tissue enhanced'; IF reliability Supported (nucleoplasm)."
}- CAF-1 packages every newly replicated genome. A complete block is lethal to every dividing cell - DepMap -2.31 across 1,208 lines - so the entire therapeutic proposition rests on hitting a partial-occupancy window that has only ever been demonstrated as a germline heterozygote in a mouse.
- CHAF1B is required for NORMAL haematopoiesis as well as for leukaemia, so the window is between two doses of the same on-target effect rather than between tumour and host, and it will move with marrow reserve, infection and chemotherapy.
- Reducing replication-coupled deposition slows the dilution of parental histones, which by this graph's own clearance-clock argument PROLONGS H3K79 methylation - a CAF-1 inhibitor is predicted to antagonise a DOT1L inhibitor.
- CAF-1 loss promotes alternative repair outcomes at forks and is linked to recombination-mediated instability in yeast, so partial inhibition in a paediatric patient carries a genome-instability risk on a timescale no leukaemia trial measures.
- Does the CHAF1B window hold in a KMT2A::AFF1 model? Every result is myeloid; the MLL-AF9 appearance is as an RNA-seq source, not as a tested genotype.
- Is CHAF1B overexpression present in infant KMT2A-rearranged B-ALL specifically? The 'almost all AML samples' claim has no published lymphoid counterpart.
- What is the relationship between CHAF1B-driven differentiation block and the CEBPA occupancy it interferes with, in a lymphoid cell where CEBPA is not the relevant differentiation factor?
- Could the partial-reduction window be reached pharmacologically at all? A germline heterozygote and a drug that occupies half the target for eight hours a day are not the same exposure.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CHAF1B, the p60 subunit of the CAF-1 complex responsible for assembling H3.1/H4 heterodimers at the replication fork in S phase, is required for normal haematopoiesis while its overexpression promotes leukaemia; it acts by binding chromatin at discrete sites and interfering with occupancy of transcription factors that promote myeloid differentiation such as CEBPA; and reducing Chaf1b activity by either heterozygous deletion or overexpression of a CAF-1 dominant-negative allele is sufficient to suppress leukaemogenesis in vivo without impairing normal haematopoiesis.
A CHAF1B-Dependent Molecular Switch in Hematopoiesis and Leukemia Pathogenesis. · Volk A, Liang K, Suraneni P, et al. · Cancer Cell 2018What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (PMC6235627 exists). **Mouse models and myeloid leukaemia.** The abstract does not name which leukaemia drivers were used, so 'suppress leukaemogenesis' is not resolved to a fusion here; the companion paper [E2] uses MLL-AF9. Graded strong rather than established because the load-bearing claim for this entry - a window at partial reduction - is an in-vivo genetic result being read as a pharmacological prediction, which it is not.
Evidence that argues against this (1)
- DepMap scores CHAF1B at a pan-essentiality of -2.314 across 1,208 cell lines, deeper than any other gene in this beat, so complete loss is lethal essentially everywhere. The mouse heterozygote and the CRISPR knockout disagree about the window, and only the mouse result supports intervention. Oncofusion knowledge graph (internal analysis) 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CHAF1B is upregulated in almost all AML samples and represses transcription of differentiation factors and tumour suppressors; analysis of RNA sequencing from mouse MLL-AF9 leukaemic cells and from paediatric AML bone-marrow aspirates identified the E3 ubiquitin ligase TRIM13 as a CHAF1B-repressed target, with CHAF1B binding the TRIM13 promoter; loss of TRIM13 or deletion of its catalytic domain enhanced leukaemogenesis in AML cell lines and patient-derived xenografts.
Repression of TRIM13 by chromatin assembly factor CHAF1B is critical for AML development. · Dean ST, Ishikawa C, Zhu X, et al. · Blood Advances 2023What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (PMC10469560 exists). **This is the only place in the entire histone-chaperone layer where a KMT2A fusion appears**, and it appears as the source of an RNA-seq dataset, not as the subject of the experiment - MLL-AF9 mouse cells were used to find a CHAF1B target, and no claim is made about CHAF1B being fusion-specific. MLL-AF9 is myeloid; the index disease here is infant B-ALL with KMT2A::AFF1. 'Upregulated in almost all AML samples' is the authors' framing of prior work.
Evidence that argues against this (1)
- A target found in one fusion model does not establish a fusion-specific mechanism. The companion paper's own framing is that CHAF1B is a general epigenetic regulator of the undifferentiated state across AML, which predicts the TRIM13 axis is not KMT2A-specific and would be found under any driver that keeps blasts undifferentiated. Volk A et al. 2018
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In DepMap CRISPR gene effect, CHAF1B's mean effect is -2.473 in 11 KMT2A-rearranged leukaemia lines against -2.286 in 43 other leukaemia lines, a difference of -0.187, at a pan-essentiality of -2.314 across 1,208 lines - the deepest pan-essentiality of any gene in this beat; the lineage split is -0.181 myeloid and -0.164 lymphoid. CHAF1A, the p150 subunit, sits at a pan-essentiality of -1.206 with a difference of +0.073.
Is any histone chaperone or variant-exchange gene selectively required in KMT2A-rearranged leukaemia lines, and does the answer survive a lineage split? · Oncofusion knowledge graph (internal analysis) 2026What this evidence cannot show
Run 2026-09-20. **The -0.187 difference is uninterpretable and is reported only to show that it is.** At a pan-essentiality of -2.31 every arm is at or past the Chronos floor, so a group difference measures distance to the floor rather than selectivity - this project's documented floor-compression failure mode, worked out on HCFC1 and KAT8. The number must not be used as a selectivity claim in either direction.
Evidence that argues against this (1)
- The pan-essential reading is contradicted at a different dose by mouse genetics: heterozygous Chaf1b deletion suppresses leukaemogenesis without impairing normal haematopoiesis. A knockout screen cannot see a window that exists only at partial reduction. Volk A et al. 2018
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
CAF-1 is the H3.1-carrying chaperone of the DNA-synthesis-dependent nucleosome assembly pathway, distinct from the HIRA/H3.3 DNA-synthesis-independent pathway, and CAF-1 is stabilised at replication forks through a DNA-binding winged-helix domain functioning with PCNA.
Histone H3.1 and H3.3 complexes mediate nucleosome assembly pathways dependent or independent of DNA synthesis. · Tagami H, Ray-Gallet D, Almouzni G, et al. · Cell 2004What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read. The PCNA/winged-helix half of this claim comes from a separate study (PMID 26908650, 2016 Nucleic Acids Research) which is cited as the counter-evidence-side source below and whose record was fetched the same day; it is stated here because the two together define what 'replication-coupled' means mechanically. Tagged-histone purification from HeLa cells.
Evidence that argues against this (1)
- Replication coupling is a property of CAF-1's recruitment, not an absolute restriction: the winged-helix/PCNA interaction stabilises CAF-1 at forks, which implies CAF-1 activity away from forks is possible and simply less favoured. A strict two-pathway model is a simplification. Zhang K et al. 2016
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
CHAF1B is UniProt Q13112 (CAF1B_HUMAN), chromatin assembly factor 1 subunit B / CAF-1 subunit B, 559 amino acids, with alias symbols CAF1A, CAF1P60, MPHOSPH7 and MPP7.
UniProtKB Q13112 (CAF1B_HUMAN), Chromatin assembly factor 1 subunit B, 559 aa · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026What this evidence cannot show
Fetched 2026-09-20 with `gene_primary` requested and read back per rule 62: exactly one reviewed human record with primary gene name CHAF1B. **Rule 63 applies with unusual force here: CHAF1B carries the alias symbol `CAF1A`, which reads as the OTHER subunit.** The p150 subunit is CHAF1A (Q13111, 956 aa). A lookup matching alias fields can resolve 'CAF1A' to the 559-residue p60 when the 956-residue p150 was meant. The accession and the length are asserted so a length check catches the swap.
Evidence that argues against this (1)
- The alias collision is exactly the CSNK2A1/CSNK2A2 shape this project logged: same complex, similar naming, and the wrong answer does not read as an error. Any literature count for CAF-1 subunits run over alias fields is suspect until title-checked. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CHAF1B's compartment is cell-cycle-switched, not fixed. In HeLa cells synchronised into G1, S, G2 and M, CAF-1 p150 and p60 are nucleus-bound in interphase, concentrate at sites of intranuclear DNA replication during S phase, and predominantly dissociate from chromatin during mitosis. UniProt Q13112 records exactly this - Nucleus and Cytoplasm on ECO:0000269 PubMed:9614144, note 'DNA replication foci. Cytoplasmic in M phase' - and here the coded citation genuinely is the founding observation.
Nucleosome assembly activity and intracellular localization of human CAF-1 changes during the cell division cycle. · Marheineke K, Krude T · J Biol Chem 1998What this evidence cannot show
HeLa cells, 1998, with chemical synchronisation - synchronising agents perturb replication themselves, so the S-phase enrichment is measured in cells that were manipulated into S phase. The paper reports p150 and p60 together and does not separate CHAF1B's behaviour from CHAF1A's. Nothing haematopoietic and nothing about a fusion.
Evidence that argues against this (1)
- Treating replication foci as CHAF1B's subnuclear address is incomplete in the direction that matters for leukaemia, and an entry stopping there would predict the wrong mechanism. In AML, CHAF1B exerts its pro-leukaemic effect by binding chromatin at DISCRETE SITES and interfering with the occupancy of transcription factors that promote myeloid differentiation - CEBPA specifically - which is a transcription-factor-competition address, not a replication-coupled one. The therapeutic argument rests on that second address: reducing Chaf1b activity suppresses leukaemogenesis in vivo without impairing normal haematopoiesis, which would be hard to explain if the only thing CHAF1B did were assemble nucleosomes behind every replication fork. Volk A et al. 2018
- E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In haematopoiesis and leukaemia CHAF1B occupies a second, non-replicative chromatin address: it binds chromatin at discrete sites and interferes with the occupancy of transcription factors that promote myeloid differentiation, such as CEBPA. CHAF1B is required for normal haematopoiesis while its overexpression promotes leukaemia, and reducing Chaf1b activity by heterozygous deletion or a CAF1 dominant-negative allele suppresses leukaemogenesis in vivo without impairing normal haematopoiesis.
A CHAF1B-Dependent Molecular Switch in Hematopoiesis and Leukemia Pathogenesis. · Volk A, Liang K, Suraneni P, Li X, Zhao J, Bulic M, et al. · Cancer Cell 2018What this evidence cannot show
Mouse, and AML in general rather than any KMT2A fusion - the system is not the index disease and the claim must not be carried to infant B-lineage KMT2A::AFF1 without saying so (rule 73). 'Binds chromatin at discrete sites' is an occupancy statement from ChIP, not an imaging result, so it is a genomic address rather than a subnuclear one in the microscopy sense.
Evidence that argues against this (1)
- The therapeutic-window claim is the load-bearing part and it is also the part most likely to be system-specific, because CHAF1B's other job is obligatory: it is the p60 subunit of the complex that assembles H3.1/H4 behind the replication fork in every dividing cell, and that function has no leukaemia specificity at all. The reason the window exists in the mouse experiment is that Chaf1b was reduced rather than removed - heterozygous deletion, or a dominant-negative allele - so the result bounds a DOSE, not a target. A complete inhibitor would be predicted to hit the replication-coupled function in every proliferating tissue, and nothing in this work tests that. Marheineke K et al. 1998
- related to →Histone H3.3 - one protein, two genes, and the reason this graph should not file it as gene-h3f3a
CAF-1 carries H3.1/H3.2, the replication-coupled contrast case to H3.3
p150 and p60 subunits of the same CAF-1 complex
CHAF1B-repressed TRIM13 found from mouse MLL-AF9 RNA-seq; only fusion link in the chaperone layer
replication-coupled arm: per-division dilution of parental histones
replication-coupled arm of the clock
floor-compressed at -2.31; the mouse heterozygote shows the screen cannot see the window
a screen cannot see the partial-loss window its mouse data show
CAF-1 p60 concentrates at S-phase replication foci and dissociates from chromatin in mitosis (PMID 9614144); a cell-cycle-gated chromatin address (Q87 2026-09-21)
In AML, CHAF1B binds chromatin at discrete sites and interferes with the occupancy of myeloid differentiation transcription factors including CEBPA (PMID 30423293, mouse)