Separation-of-function RTF1 rescue in a KMT2A-fusion line: does the fusion need PAF1C for elongation, or for H2B ubiquitination?
also called RTF1 HMD mutant rescue · PAF1C ubiquitination-vs-elongation epistasis
Knock out RTF1 in a KMT2A-fusion leukaemia cell, then put back either the normal protein or a version that can still help transcription but cannot trigger the ubiquitin tag, and see which one rescues the leukaemia programme. It separates two explanations that the field currently runs together.
The same thing, for a scientist
The KMT2A-PAF1C contact is established and essential for MLL-AF9 immortalisation; PAF1C/RTF1 dependence of H2B ubiquitination is established; H2Bub dependence of DOT1L-mediated H3K79 methylation is established. What is not established is whether the first matters *because of* the second and third. This experiment asks it directly with a rescue design: degron-tagged endogenous RTF1 in a KMT2A-fusion line, acute depletion, and re-expression of either wild-type RTF1 or an RTF1 allele engineered to lose the H2B-ubiquitination-promoting activity while retaining Pol II association. Read out, in order: H2BK120ub by western and CUT&RUN; H3K79me2 by CUT&RUN at HOXA9/MEIS1 and genome-wide; fusion target-gene expression by RNA-seq; and proliferation.
Design
System. Two KMT2A-fusion lines with different partners to avoid a partner-specific artefact — SEM (KMT2A::AFF1, B-ALL, the disease-relevant genotype) and MV4;11 (KMT2A::AFF1) or MOLM-13 (KMT2A::MLLT3) — plus one non-KMT2A leukaemia line as the selectivity comparator. The comparator is not optional: this beat's central finding is that the H2Bub writers look selective in KMT2A papers and are not selective when a comparator is run.
Perturbation. dTAG or auxin degron knocked into endogenous RTF1, so depletion is acute and the cells are not given weeks to adapt. Rescue constructs expressed from a defined locus at near-endogenous level: (i) wild-type RTF1, (ii) the separation-of-function allele, (iii) empty vector.
The hard part, stated as such. No validated human separation-of-function RTF1 allele exists in the literature this beat searched. Rtf1's H2B-ubiquitination-promoting activity in yeast maps to a conserved region, and the human equivalent would have to be defined first — probably by a small alanine scan read out on H2Bub western, which is a month of work before the real experiment starts. An entry that hid that would be describing a cheaper experiment than the real one.
Read-outs, in causal order. H2BK120ub (western, then CUT&RUN for locus-level); H3K79me2 (CUT&RUN at HOXA9, MEIS1, and genome-wide, including the KMT2A::AFF1 enhancer elements the graph already carries); nascent transcription (PRO-seq or TT-seq, to separate elongation effects from steady-state RNA); proliferation and differentiation markers.
Controls. A DOT1L-inhibited arm as the positive control for what loss of the H3K79 arm looks like in the same cells; a catalytically dead RNF20 arm as the orthogonal way to remove H2Bub; and an RTF1-degron arm with no rescue as the floor.
Two objections that change the design, added in the 2026-09-19 coverage repair.
-
The handle may be the wrong protein in human cells. Cao et al. report human Rtf1 is not stably associated with PAF1C, that Rtf1 and PAF1C govern distinct gene subsets in HeLa, and that PAF1C recruitment was Rtf1-independent - the reverse of the yeast finding this design is built on (PMID 26217014, counter-evidence on 1). Before the degron is built, the human RTF1-PAF1C relationship has to be established in the actual leukaemia line. Where a separation-of-function allele for this function does exist, it is in RAD6/UBE2A, in yeast (PMID 37216505), which points at a UBE2A-side allele or a catalytically dead RNF20 knock-in as the cheaper route to the same question.
-
The elongation control runs the wrong way. Chen et al. showed PAF1 depletion INCREASES paused-Pol-II release via SEC recruitment (PMID 26279188, counter-evidence on 2), and SEC is what KMT2A::AFF1 recruits. The PRO-seq/TT-seq arm was meant to separate elongation from ubiquitination; with this effect present it measures both, so an SEC-occupancy arm (AFF1/AFF4/CDK9 CUT&RUN) is now part of the minimum design rather than an optional extra.
What each outcome would mean
| Result | Interpretation |
|---|---|
| SoF allele rescues H3K79me2 and growth | the ubiquitin arm is dispensable in KMT2A-fusion leukaemia; this entire beat is a chromatin footnote and the graph should say so |
| SoF allele fails to rescue H3K79me2 but rescues growth | H3K79me2 is not the load-bearing output of PAF1C in this disease, which would be a substantial revision to mechanism-dot1l-h3k79-methylation |
| SoF allele fails to rescue either | PAF1C's role in KMT2A leukaemia is substantially a ubiquitination role; the H2Bub layer moves from footnote to mechanism |
| Non-KMT2A comparator behaves identically | whatever the answer, it is not KMT2A-specific - the outcome this beat's DepMap results predict |
| design | Degron-mediated acute depletion of endogenous RTF1 in KMT2A-fusion lines with rescue by wild-type versus a separation-of-function RTF1 allele that retains Pol II association but not H2B-ubiquitination-promoting activity, plus a non-KMT2A leukaemia comparator. |
|---|---|
| readout | H2BK120ub (western and CUT&RUN); H3K79me2 (CUT&RUN, HOXA9/MEIS1 and genome-wide); nascent transcription (PRO-seq/TT-seq); fusion target-gene expression; proliferation and differentiation markers. |
| controls | DOT1L inhibitor arm; catalytically dead RNF20 arm; RTF1-degron with no rescue; non-KMT2A leukaemia line throughout. Added 2026-09-19: SEC-occupancy arm (AFF1/AFF4/CDK9 CUT&RUN), because PAF1 depletion increases paused-Pol-II release through SEC (PMID 26279188) and would otherwise confound the nascent-transcription readout. |
| cost tier | high |
| timeline | 3-4 months to map and validate a human separation-of-function RTF1 allele, then 4-6 months for the rescue experiment in three lines. |
| prerequisite | No validated human separation-of-function RTF1 allele exists; it must be built first, and if it cannot be built the experiment cannot be run as designed. |
| premise at risk | human RTF1 may not be the PAF1C handle at all - PAF1C recruitment was Rtf1-independent in HeLa (PMID 26217014); the only separation-of-function precedent is in yeast RAD6 (PMID 37216505) |
tests hypotheses
[ "hyp-h2bub-required-but-not-selective" ]
- If the separation-of-function allele is even slightly hypomorphic for elongation, the experiment reads out elongation loss and is uninterpretable. The mapping phase has to demonstrate normal Pol II association and normal nascent transcription at control genes before the leukaemia read-outs mean anything.
- PAF1C perturbation increases SEC-driven pause release (PMID 26279188) while removing the ubiquitination route, so both the positive and the negative outcome have a second, opposite-signed explanation unless SEC occupancy is measured alongside.
- Is human RTF1 even in PAF1C in a KMT2A-fusion line? PMID 26217014 says it is not stably associated in human cells and that PAF1C recruitment is Rtf1-independent, which would make this experiment a test of RTF1 rather than of the fusion's PAF1C contact.
- Can a human separation-of-function RTF1 allele be built at all? If RTF1's H2Bub function is inseparable from its Pol II association in human cells, this experiment is not executable and the question needs a different handle - most likely a catalytically dead RNF20 knock-in instead.
- Should the comparator be a germline-KMT2A B-ALL line or a different fusion? The first controls for lineage, the second for the fusion mechanism, and they answer different questions.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Rtf1 is required for H2B ubiquitination and for global H3K4 and H3K79 methylation but not H3K36 methylation, Dot1 association with active genes is largely Rtf1-independent, and the authors suggest Rtf1 affects H2B ubiquitination by a mechanism distinct from its elongation-factor function - the premise that a separation-of-function allele should be constructible.
The Rtf1 component of the Paf1 transcriptional elongation complex is required for ubiquitination of histone H2B. · Ng HH, Dole S, Struhl K · J Biol Chem 2003What this evidence cannot show
Abstract read in full; full text not read. Yeast. The separability is a suggestion in the discussion, not a demonstrated allele, and the human RTF1 region responsible is not defined by this paper - which is why this experiment carries a mapping phase.
Evidence that argues against this (2)
- The design's premise is yeast-specific and is contradicted in human cells. Cao et al. showed human Rtf1 is not stably associated with PAF1C, that it acts as an elongation factor apparently independent of PAF1C, that Rtf1 and PAF1C regulate distinct subsets of genes in HeLa, and - directly opposite to the yeast result - that PAF1C was recruited to the genes examined in an Rtf1-INDEPENDENT manner. If that holds in a leukaemia line, an RTF1 degron does not perturb PAF1C and the experiment answers a question about RTF1 rather than about the KMT2A-PAF1C contact it was designed to interrogate. A published correction to this article exists (PMID 26518638) whose content this repair did not establish. Cao QF et al. 2015
- Where a separation-of-function allele for this function actually exists, it is in the wrong protein and the wrong organism. Fetian et al. mapped the Rtf1 histone-modification-domain contact to the conserved N-terminal helix of Rad6 and built separation-of-function mutants that impair the Rad6-HMD interaction and H2BK123 ubiquitylation while sparing other Rad6 functions - in Saccharomyces cerevisiae, in RAD6, not in RTF1 and not in human cells. The precedent supports a UBE2A-side allele rather than the RTF1-side allele this design assumes, which changes the mapping phase and the interpretation. Fetian T et al. 2023
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
MLL interacts directly with PAFc through sequences flanking the CxxC domain, PAFc knockdown disrupts MLL-fusion-mediated transcriptional activation and MLL recruitment to target loci, and MLL deletions that abolish PAFc interaction eliminate MLL-AF9-mediated immortalisation.
The PAF complex synergizes with MLL fusion proteins at HOX loci to promote leukemogenesis. · Muntean AG, Tan J, Sitwala K, Huang Y, Bronstein J, Connelly JA, Basrur V, Elenitoba-Johnson KS, Hess JL · Cancer Cell 2010What this evidence cannot show
PMID fetched and title read back by this beat; abstract read in full, full text not read. MLL-AF9 mouse model; the B-ALL/KMT2A::AFF1 case is assumed rather than shown.
Evidence that argues against this (1)
- The readout is confounded by an effect running the other way through the fusion's own machinery. Chen et al. showed that reducing PAF1 INCREASES release of paused Pol II at thousands of genes, with higher nascent transcription and Ser2P, because the super elongation complex is recruited. KMT2A::AFF1 recruits SEC. A PAF1C-directed perturbation in a KMT2A-fusion cell therefore changes elongation in the opposite direction from the one this design implicitly assumes, and the PRO-seq/TT-seq arm that was meant to separate elongation effects from ubiquitination effects would be measuring both at once. The design needs an SEC-occupancy arm to be interpretable at all. Chen FX et al. 2015
- ← related toRTF1 — the PAF1C subunit the graph was missing, on a complex that does bind the retained KMT2A fragment
Design objection (Cao 2015, PMID 26217014): if human Rtf1 is not stably in PAF1C and PAF1C recruitment is Rtf1-independent in human cells, an RTF1 degron tests RTF1 rather than the fusion's PAF1C contact; the separation-of-function precedent that does exist is in yeast RAD6 (PMID 37216505), pointing at a UBE2A-side allele or a catalytically dead RNF20 knock-in instead
Design objection (Chen 2015, PMID 26279188): PAF1 depletion increases paused-Pol-II release via SEC recruitment, and SEC is what KMT2A::AFF1 recruits, so the PRO-seq/TT-seq arm meant to separate elongation from ubiquitination measures both at once unless an SEC-occupancy arm is added
separates PAF1C's elongation role from its H2B ubiquitination role in a KMT2A-fusion cell
separates the elongation and ubiquitination explanations of the KMT2A-PAF1C contact