The COMPASS complex (Complex of Proteins Associated with Set1) is a multi-protein molecular machine that attaches methyl groups to a specific spot on histone proteins, the spools around which DNA is wound in every cell. First identified in baker’s yeast more than two decades ago, COMPASS was the first enzyme shown to place a methyl mark on lysine 4 of histone H3, a chemical tag now recognized as a universal signal that helps switch genes on or keep them primed for activation.1PubMed Central. The COMPASS family of histone H3K4 methylases: mechanisms of regulation in development and disease pathogenesis Since that discovery, researchers have found versions of COMPASS in organisms from yeast to plants to humans, and mutations in its components turn up in cancers, developmental syndromes, and problems with how stem cells decide their fate.
From Yeast to Human Cells
The original COMPASS was discovered in the yeast Saccharomyces cerevisiae, built around a catalytic subunit called Set1. Since then, studies have confirmed that both the complex and the methyl mark it produces are deeply conserved across the tree of life.2PubMed Central. The COMPASS family of histone H3K4 methylases: mechanisms of regulation in development and disease pathogenesis Yeast gets by with a single Set1-based COMPASS, but mammals have expanded the toolkit. Humans carry six related complexes, each assembled around a different catalytic subunit: SET1A, SET1B, MLL1, MLL2, MLL3, and MLL4. These belong to the KMT2 family of methyltransferases and share a common set of accessory proteins, yet each version has distinct jobs in the cell.3PubMed Central. Resetting the Epigenetic Balance of Polycomb and COMPASS Function at Enhancers for Cancer Therapy Structural work comparing yeast and human forms has shown that the core architecture is remarkably similar, reinforcing the idea that this machinery was already in place long before multicellular life evolved.4PubMed Central. Structural analysis of the core COMPASS family of histone H3K4 methylases from yeast to human
How the Machine Is Built
At its core, every COMPASS complex contains a catalytic SET-domain subunit surrounded by a set of shared accessory proteins often called WRAD (WDR5, RbBP5, ASH2L, and DPY-30 in mammals, or their yeast equivalents Swd3, Swd1, Bre2, and Sdc1). These accessory subunits are not just along for the ride. A crystal structure of the yeast catalytic module revealed that Swd1 acts as the organizational backbone: its conserved tail nucleates the other subunits and joins them to Set1, constructing a regulatory pocket right next to the active site.5PubMed Central. Crystal Structure of the COMPASS H3K4 Methyltransferase Catalytic Module Within that pocket, a motif from Swd3 acts like a doorstop, physically tuning which version of the methyl mark the enzyme can produce. This doorstop mechanism helps explain why different members of the COMPASS family add different numbers of methyl groups to the same histone position.
When COMPASS docks onto a nucleosome (a histone-DNA unit), it spans the entire face of the nucleosome disc and uses four of its six subunits to make direct contacts with histones and DNA. The catalytic SET domain sits at the center of this interface, positioned to reach the flexible tail of histone H3 that sticks out from the nucleosome. The total buried surface area at the interface is roughly 3,000 square angstroms, which accounts for the tight grip COMPASS has once it latches on.6Cell. Cryo-EM Structure of the COMPASS Histone H3K4 Methyltransferase Complex Bound to a Nucleosome Cryo-EM imaging has also captured the complex in several partially assembled states, suggesting that it goes through a stepwise assembly process rather than arriving at the nucleosome fully formed.7PubMed Central. Multistate structures of the MLL1-WRAD complex bound to H2B-ubiquitinated nucleosome
Three Flavors of a Methyl Mark
COMPASS family complexes can add one, two, or three methyl groups to lysine 4 on histone H3, producing marks abbreviated as H3K4me1, H3K4me2, and H3K4me3. These are not interchangeable. Each mark sits in different parts of the genome and signals something different to the cell’s gene-reading machinery.
H3K4me3, the trimethyl version, concentrates near gene promoters and is strongly associated with active transcription. SET1A, SET1B, MLL1, and MLL2 are the main complexes that produce this mark. H3K4me1, the monomethyl version, has a different geography: it marks enhancers, the distant regulatory switches that fine-tune when and where genes are turned on. MLL3 and MLL4 are the primary monomethylases at enhancers and are among the most frequently mutated histone-modifying enzymes in human cancers.8PubMed Central. Resetting the Epigenetic Balance of Polycomb and COMPASS Function at Enhancers for Cancer Therapy
The monomethyl mark is not always a green light, though. Work in muscle precursor cells found that MLL3 and MLL4 place H3K4me1 at promoters of certain muscle and inflammatory genes, and that mark is actually associated with keeping those genes quiet. When muscle cells begin to mature, MLL3 leaves those promoters, a different COMPASS complex moves in, and the mark shifts from mono- to trimethylation as the genes switch on.9PubMed Central. A role for H3K4 monomethylation in gene repression and partitioning of chromatin readers So the same histone position can carry opposing messages depending on how many methyl groups sit on it and which COMPASS variant put them there.
The Ubiquitin Switch
One of the more striking findings about COMPASS is that it does not work alone. It depends on a chemical tag on a completely different histone, a phenomenon sometimes called trans-histone crosstalk. Specifically, COMPASS’s ability to methylate histone H3 depends on a single ubiquitin molecule attached to histone H2B at a neighboring position on the same nucleosome.
In yeast, the link was traced to a subunit called Cps35, which joins the complex only when H2B is ubiquitinated. Without that ubiquitin signal, Cps35 stays away and COMPASS is catalytically dead in living cells. Adding purified Cps35 back to a COMPASS isolated from cells that lack the ubiquitin mark was enough to restore methylation activity.10PubMed. Histone crosstalk between H2B monoubiquitination and H3 methylation mediated by COMPASS Cps35 also turned out to be required for trimethylation at a different histone position (H3K79), hinting that it serves as a broader translator of the ubiquitin signal across multiple marks.
Structural studies have shown how this works at the atomic level. COMPASS straddles the nucleosome face with ubiquitin on one side and the H3 tail on the other. The catalytic subunit Set1 contains an arginine-rich motif that normally inhibits the enzyme, essentially acting as a built-in safety lock. When ubiquitin is present, it packs against the Swd1 scaffold subunit and pins that inhibitory motif to the nucleosome’s acidic patch, relieving the autoinhibition and allowing methylation to proceed.11Molecular Cell. Cryo-EM Structure of the COMPASS Histone H3K4 Methyltransferase Complex Bound to a Nucleosome This means COMPASS is essentially a coincidence detector: it only fires when both the right nucleosome context and the upstream ubiquitin signal are in place.
Linking Methylation to Active Genes
Placing a methyl mark in the right spot requires COMPASS to arrive at genes that are being actively read by the cell’s transcription machinery. In yeast, this recruitment happens through a physical interaction between the Set1 protein and RNA polymerase II, the enzyme that copies DNA into messenger RNA. Specifically, the N-terminal region of Set1 and the accessory subunit Swd2 cooperate to grab onto the tail of RNA polymerase II, dragging COMPASS along as the polymerase moves through the early part of a gene.12Nature Communications. The Set1 N-terminal domain and Swd2 interact with RNA polymerase II CTD to recruit COMPASS This piggyback mechanism explains why H3K4me3 accumulates near gene beginnings rather than being spread evenly across the genome: COMPASS gets on early in the transcription cycle and drops off as the polymerase moves farther along.
Enhancer Priming and Cell Fate Decisions
While SET1-containing complexes ride the polymerase at promoters, MLL3 and MLL4 have a different specialty. These two COMPASS family members are the dominant enzymes that place H3K4me1 and H3K4me2 at enhancers, preparing them for activation during cell differentiation. MLL4 localizes to enhancers in a cell-type-specific and differentiation-stage-specific manner, co-localizing with transcription factors that define cell identity. Deleting MLL4 causes a collapse in the signals that mark active enhancers, including not just H3K4me1 but also the acetylation mark H3K27ac and the recruitment of the Mediator co-activator complex.13PubMed Central. H3K4 mono- and di-methyltransferase MLL4 is required for enhancer activation during cell differentiation
MLL3 and MLL4 are partially redundant with each other, so losing one is often compensated by the other. But their priming activity controls something important about cell fate transitions: while cells can maintain their existing identity without MLL4, they struggle to change into a new cell type. Enhancer priming by MLL4 orchestrates the activation of new enhancers needed for a cell to acquire a different identity, acting as a gatekeeper for fate transitions.14PubMed Central. Enhancer priming by H3K4 methyltransferase MLL4 controls cell fate transition In embryonic stem cells, MLL4 and a demethylase called LSD1 form an opposing regulatory axis: MLL4 activates gene targets at enhancers while LSD1 removes those same methyl marks to keep them in check. Depleting LSD1 can even rescue the differentiation defects caused by losing MLL4, demonstrating how tightly balanced this system is.15PubMed Central. An Mll4/COMPASS-Lsd1 epigenetic axis governs enhancer function and pluripotency transition in embryonic stem cells
Disease Connections
Given COMPASS’s central role in gene regulation, it is no surprise that mutations in its components surface across a range of diseases. The connections fall broadly into two categories: cancer and developmental syndromes.
On the cancer side, the MLL1 gene (KMT2A) is a frequent target of chromosomal rearrangements in leukemia. These rearrangements fuse MLL1 to dozens of different partner genes, producing chimeric proteins that hijack gene activation programs and drive blood cancers. MLL3 and MLL4, meanwhile, are among the most commonly mutated genes across many solid tumor types, with loss-of-function mutations that disrupt their enhancer-priming activity.16PubMed Central. Resetting the Epigenetic Balance of Polycomb and COMPASS Function at Enhancers for Cancer Therapy
On the developmental side, mutations in five COMPASS-associated genes have been linked to three congenital syndromes: Kabuki syndrome (caused by mutations in KMT2D or the associated demethylase KDM6A), Rubinstein-Taybi syndrome (caused by mutations in the co-activators CBP or EP300 that work alongside COMPASS at enhancers), and Kleefstra syndrome type 2 (caused by KMT2C mutations).17PubMed Central. KMT2C/D COMPASS complex-associated diseases [K(CD)COM-ADs]: an emerging class of congenital regulopathies These syndromes share features such as intellectual disability, distinctive facial characteristics, and growth delays, consistent with the idea that disrupting COMPASS-mediated enhancer activation during embryonic development has sweeping consequences.
Targeting COMPASS in the Clinic
The involvement of MLL1 fusion proteins in leukemia has made the COMPASS system a target for drug development. A key vulnerability in MLL-rearranged leukemia is the interaction between MLL1 and menin, a scaffold protein that helps recruit the fusion protein to its target genes. Blocking that interaction effectively cuts the fusion protein off from the genome.
Researchers have developed highly potent small-molecule inhibitors of the menin-MLL interaction, including compounds designated MI-463 and MI-503, which bind menin with low-nanomolar affinity and are orally bioavailable. In mouse models of MLL leukemia, these compounds provided a substantial survival benefit.18Cancer Cell. Blockade of the Menin-MLL Interaction Leads to Reversal of Differentiation Block and Effective Inhibition of MLL Leukemia Since then, the field has expanded: menin-MLL inhibitors have also shown tumor-suppressive effects in prostate cancer, breast cancer, liver cancer, and lung cancer cell models, suggesting the approach may have applications beyond blood cancers.19PubMed Central. Menin-MLL1 Interaction Small Molecule Inhibitors: A Potential Therapeutic Strategy for Leukemia and Cancers Several menin inhibitors have entered clinical trials in recent years, making this one of the more active areas of epigenetic drug development.
Roles Beyond the Methyl Mark
A growing theme in COMPASS research is that some of its subunits do things the methyl mark alone cannot explain. In the roundworm C. elegans, the accessory subunit WDR-5 was found to have substantial effects on gene expression that persist even when the methylation mark itself is completely abolished. Researchers compared worms missing WDR-5 alone to worms missing both WDR-5 and another COMPASS subunit, RBBP-5. The double mutant, which lacks all H3K4 methylation, had a gene expression profile nearly identical to the WDR-5-only mutant, indicating that the transcriptional changes caused by losing WDR-5 are largely independent of its catalytic role in methylation.20PubMed Central. WDR-5 exhibits H3K4 methylation-independent activity during embryonic development in C. elegans This suggests that COMPASS subunits moonlight as scaffolds or interaction platforms for other regulatory complexes, contributing to gene control in ways that go beyond their canonical enzyme activity.
Separately, research in the worm has connected COMPASS to DNA damage survival. Removal of the WRAD subcomplex worsened developmental growth problems and accelerated aging under ultraviolet-induced DNA damage, while depleting the enzymes that erase the H3K4me2 mark had the opposite effect, promoting growth and extending lifespan under genotoxic stress. The H3K4me2 mark induced by DNA damage was linked to activation of genes involved in RNA processing, ribosome building, and protein quality control, pointing to a role for COMPASS in the recovery of protein production after genomic insults.21Nature Structural & Molecular Biology. H3K4me2 regulates the recovery of protein biosynthesis and homeostasis following DNA damage
Long Noncoding RNA and COMPASS Recruitment
COMPASS does not always find its targets through protein-protein interactions alone. Long noncoding RNAs, stretches of RNA that are not translated into protein, can serve as molecular guides. In humans, a long noncoding RNA called HOTTIP, transcribed from one end of the HOXA gene cluster, physically associates with WDR5 and helps recruit the trimethylation complex to nearby HOXA genes to keep them active. A parallel system exists in the plant Arabidopsis, where an intronic RNA derived from the FLC gene reinforces the binding of the COMPASS catalytic subunit ATX1 to FLC chromatin, boosting H3K4me3 and gene expression.22Scientific Reports. Exogenously overexpressed intronic long noncoding RNAs activate host gene expression by affecting histone modification in Arabidopsis The conservation of this RNA-guided targeting strategy across kingdoms suggests it is an ancient feature of COMPASS regulation, not a one-off trick in a particular organism.
Phase Separation and the Assembly Problem
Getting all the COMPASS subunits together in the right place at the right time is itself a regulatory challenge. Under physiological conditions, the fully assembled complex actually tends to fall apart, making spontaneous assembly thermodynamically unfavorable. One model proposes that the cell solves this by corralling the subunits into biomolecular condensates, droplet-like compartments that form within the nucleus through a process called phase separation. The MLL1 protein contains large stretches of intrinsically disordered sequence, a feature strongly associated with the ability to enter such condensates. Concentrating the subunits inside these droplets could push the assembly equilibrium toward the complete complex, allowing it to form, act on nearby nucleosomes, and then disassemble when the condensate dissolves.23bioRxiv. Hierarchical assembly of the MLL1 core complex within a biomolecular condensate regulates H3K4 methylation If this “swinging-domain” model holds up, it would mean that COMPASS activity is controlled not just by what signals are on the chromatin, but by the physical organization of the nucleus itself. The idea is still being tested, but it offers an elegant explanation for how such an unstable complex manages to function reliably in living cells.