Cracking the ubiquitin code to unlock new drug targets

More than 25 years after the original concept, the first engineered protein-degrading drug, known as a Protac – proteolysis targeting chimera – was approval by the US Food and Drug Administration in May this year. Veppanu (vepdegestrant) was developed by biotech Arvinas and Pfizer for several types of hard-to-treat breast cancer. It allows the cancer cell to use its own protein-recycling machinery to destroy the mutant protein responsible for driving cancer proliferation. Hot on its heels, August saw the approval of multiple myeloma drug, Zenbexus (iberdomide) from Bristol Myers Squibb, the first molecular glue protein degrader, which hijacks the same protein-recycling machinery.

The key to the action of both these new classes of drugs is ubiquitin – a small 76-amino-acid protein which is, as its name suggests, ubiquitous. It’s found in all animal, plant and even yeast cells. In the normal life of a cell, ubiquitins, strung together in chains, are attached to other proteins, ‘often to signal that these proteins or biomolecules should be degraded’, says biochemist David Komander from WEHI in Melbourne, Australia.

Protacs can recruit both a target protein and the enzyme that tags that protein with a ubiquitin, leading to the protein’s degradation. With at least 40 Protacs in clinical trials, Komander predicts we will soon see ‘a complete shift in the pharmaceutical industry’ towards protein-degrading drugs and potentially drugs designed to stop protein degradation as well.

Ubiquitin tags are responsible for more than just protein degradation. In the 2000s researchers showed that multiple types of ubiquitin chains control cell programming. They signal when a protein should be moved to a different location or assemble a complex of proteins; they are involved in immune responses and responding to DNA damage. And it’s now apparent they are not even limited to proteins but have recently been found on lipids, sugars and other metabolites.

Why does ubiquitin matter?

The ubiquitin code is complex. Polyubiquitin chains come in linear and branched forms, and research teams are only just getting to grips with matching the huge variety found in cells to their specific function. Once this is understood, the rewards could be immense.

What makes ubiquitination so complex is the number of ways it can be attached to both a protein and to itself when it forms chains. This is via an amide bond from the carboxyl group at its glycine residue C-terminus to an amino group on a lysine side chain on the target protein. It can then bond to itself via any of the seven lysine amino acids residues within the molecule, as well as binding via its N-terminus, giving eight possible linear chains. That’s before you even consider the branching combinations as further ubiquitins get added – although we don’t yet know how many combinations actually exist in biology.

‘On a single protein you can have up to 30–40 ubiquitins, and they often form in a fraction of a second,’ says Benedikt Kessler, a biochemist at the University of Oxford in the UK who has developed novel methods to study the ubiquitin code using mass spectrometry. The tags are placed and removed from proteins via several classes of enzymes, including over 600 E3 ligases, which add a ubiquitin tag, and the near 100 deubiquitinating enzymes (Dubs) which detach them, reversing this very dynamic signalling.

One well-characterised polyubiquitin chain links ubiquitins via the lysine that sits at position 48 in its amino acid sequence, known as K48. ‘Once you have four of these lysine 48 ubiquitins on a substrate, that’s usually the signal that targets it to the proteasome for destruction,’ Kessler explains. The proteasome is the cell’s large barrel-shaped protein complex that chews up proteins for recycling.

Analysing the ubiquitinome

The method of choice to analyse the complete set of proteins in a cell that carry ubiquitin tags, known as the ubiquitinome, is liquid chromatography–tandem mass spectrometry (LC MSMS) which can identify which proteins are tagged and where. The reference database of the ubiquitinome now has over 100,000 unique entries.

As with all proteomics, analysis starts with trypsin digestion, the enzyme cleaving peptide bonds between lysine and arginine residues. For ubiquitin this means, when it is chopped up, a two-glycine residue is left behind on the ubiquinated proteins, known as a diGly tag. The protein itself is also broken up, but the tag adds an extra 114Da to a peptide fragment’s mass, providing the signature for ubiquitination, which can be identified when mapped to protein MS databases.

Although standard diGly proteomics tells you where a protein is ubiquitinated, trypsin also chops up the ubiquitin chains, meaning it destroys the architecture information, so you cannot tell how long it is or how it is connected to other ubiquitins – the complete ubiquitin code. Komander says the more we look, the more it is clear that the type and architecture of each chain programmes an individual cell signal.

In 2019 Komander developed a new method to read the code. ‘We discovered an enzyme which cleaves ubiquitin but doesn’t cleave any other proteins,’ he says. His method, known as Ub-clipping, uses an engineered viral protease that creates diGly tags but is not able to make further cuts within ubiquitin itself. It therefore separates each of the remaining ubiquitin units, leaving them almost intact, but carrying one or more diGly tags at whichever lysine they were attached to another ubiquitin.These masses can then be used to reconstruct some of the chain architecture – for example, a ubiquitin with two diGly tags would indicate a branching point.

Decoding ubiquitin chain architecture

It was this method that showed that between 10–20% of ubiquitin chains are branched – more than previously thought and ‘quite a significant proportion’, says Leo Kiss, a biochemist at the University of Duisburg–Essen in Germany. To relate structure to the function of each chain is still difficult because of the heterogeneity of different chains inside living cells, with multiple architecture present at once.

In 2025, Kiss was a postdoc in Brenda Schulman’s lab at the Max Planck Institute of Biochemistry in Martinsried, Germany, and developed a method called UbiRead – ubiquitinated reporter evaluation after intracellular delivery – which uses enzymes to create a protein tagged with a known ubiquitin chain that has a fluorescent label attached. This is delivered to a cell using electroporation – short electrical pulses that temporarily open up the cell membrane – and its fate can be monitored to see if the labelled chain caused the protein to be degraded and if so, how fast. ‘It’s a very simple way of looking at these things,’ says Kiss. ‘The cell essentially translates the code for us.’

So far the method has provided some striking results. For example, Kiss has shown that when a K48 chain grows to three ubiquitin units, protein degradation is triggered within minutes. The result came as a surprise, he says: previous work suggested they had to be longer but he also observed a fraction of the protein where ubiquitins were removed, showing how the system dynamically balances itself.

Targeting undruggable proteins

Despite the limited understanding of the ubiquitin code, there are already ways ubiquitin tagging is being exploited to create therapies. Only about 20% of proteins can be drugged by small molecule inhibitors. The rest lack the binding pockets that conventional drugs can block, so harnessing the ubiquitin systems offers a different approach for reaching the undruggable 80%. If the right ubiquitin tag could be added, the cell’s own machinery would degrade the whole protein altogether.

Protein degrader drugs are now ‘a really big field’ says Komander. ‘The whole scope of what kind of proteins you can now start to think about removing from the cell is pretty unlimited, and that’s very exciting.’ Cancer therapies are dominating new drug development, because the uncontrolled growth of cancer cells is sometimes linked to failure of the ubiquitin system, explains Kessler.

Biochemist Alessio Ciulli from the University of Dundee in the UK was one of the earliest Protac designers. The drugs are made up of a molecule that binds the target protein with another that binds a ubiquitin ligase via a bridging linker. The proximity allows the enzyme to tag the protein with ubiquitin – a process that would not have otherwise occurred. ‘We can now make thousands of compounds very quickly with a range of linkers, a range of binders, and test them very quickly,’ says Ciuli.

He is now trying to understand why two Protacs that bind to both target and enzyme don’t always have the same ability to degrade. In 2024 he developed the idea of ‘ubiquitinability’ based on cryo-EM studies of a successful Protac system. His team discovered that binding a protein and ligase in close proximity via a Protac is not enough. The protein needs to be in the right orientation and contain a lysine residue in the right position for the enzymatic transfer of the ubiquitin tag. ‘There’s a [protein] face where the ubiquitinated enzyme is almost shining light [on it],’ says Ciulli, which he has named the ‘ubiquitination zone’. Ciulli’s model for ubiquitinability is providing a blueprint for more rational Protac design.

But another type of protein degrader is also making waves. ‘It turns out that we can actually do this with molecules that are much smaller than the Protacs, that don’t have the two heads,’ says Ciulli. The principle is the same but molecular glue degraders do the job in one small molecule, which simply creates a new improved protein–protein interface between the E3 ligase and target protein so they directly interact, leading to ubiquitination. Smaller molecular glue degraders more readily pass into cells and thus would make better drugs.

There are now about 20 molecular glues in clinical trials, treating cancers, and a first, Zenbexus (iberdomide), has crossed the finishing line, in securing approval to treat multiple myeloma in August. It targets cereblon E3 ligase, one of the E3 ligases also investigated for Protac design, and degrades two transcription factors responsible for driving the cancer.

Designing these molecules is still a challenge and so far they have typically been discovered serendipitously, but Ciulli says we do now understand a lot more about how to design them. The process often starts by finding complementarities or possible modes of interaction between the ligase and target protein that could be stabilised by the molecular glue. ‘Once we define that, we’re off to the races, because we can now systematically develop some screening cascade and be very rational about how we approach our design,’ he says

Deubiquitinating strategies for Parkinson’s disease

Kessler has been developing another way to exploit the ubiquitin code. Rather than recruiting the enzymes responsible for ubiquitination, he has been developing novel drugs to inhibit deubiquitinating enzymes, the Dubs, that remove ubiquitin from proteins to stop their degradation. ‘For about 20 of the 100 Dubs, we know in which biological process they play a role,’ he says. These can be simple small molecule inhibitors, which Kessler says is a more conventional and less risky strategy than other protein degrader drugs.

One Dub strategy developed to treat cancer is also not based on protein degradation but another part of the ubiquitin code. The USP1 deubiquitinating enzyme normally removes single K63 ubiquitin tags that have been added to repair proteins to activate a quick, often error-prone, temporary DNA repair mechanism. When the tags are removed by USP1 it’s a signal to turn off the DNA repair machinery. In cancer cells carrying the BRCA mutation, there are high levels of DNA damage and the mechanism becomes important for cell survival. But if USP1 is prevented from removing the ubiquitin tags, the repair mechanism keeps going and causes errors to pile up, eventually killing the cell. Several USP1 inhibitors are now in early clinical trials.

Mission Therapeutics, based in Cambridge UK, have started clinical trials to treat Parkinson’s disease using another Dub inhibitor. The Dub USP30 usually removes ubiquitin tags from the mitochondrial membrane, which slows the removal of dysfunctional mitochondria. The mechanism seems to be a back-up precaution to balance any over-zealous destruction of mitochondria – one of many checks and balances in the ubiquitin code.

Normally, when damaged, ubiqutin tags will keep coating the mitochondria so it becomes degraded, but this mechanism is weaker in people with Parkinson’s disease, and with the additional action of the USP30 removing ubiquitin signals to destroy, the damaged mitochondria accumulate in neurons, triggering cell death. ‘What our compounds do is potently and specifically bind to USP30 and prevent it from removing ubiquitin,’ says Suhail Nurbhai, chief medical officer at Mission Therapeutics. This seems to be enough to encourage the cell to degrade damaged mitochondria and create new more functional ones.

It’s taken them 10 years; ‘the chemistry is not straightforward’, says Nurbhai, but Mission’s small molecule drug candidate MTX 325 has now successfully completed a phase 1 safety trial and the company is about to recruit Parkinson’s disease patients to test its efficacy. There is excitement about this potentially disease-modifying treatment and future prospects for USP30 inhibition in other neurodegenerative conditions.

As with many new areas of drug discovery Dub therapeutics have also had their failures. ‘We need to see the first clinical successes in order to really get an appreciation of what these enzymes can do,’ says Komander. He is also working on another approach which more directly inverts the Protac concept – the Dubtac, which uses a ligand that recruits a Dub to remove a ubiquitin chain from a target protein to stop it being degraded. The proof-of-concept came in 2022 from Daniel Nomura and colleagues at the University of California, Berkeley, in the US who showed they could stabilise a mutant protein found in human cystic fibrosis respiratory cells. The protein regulates the mucus lining in the lungs. ‘Because it’s not quite functional, it’s recognised as being broken and it’s removed by the ubiquitin system. If we keep the protein free of ubiquitin, more of the protein is generated and, although it’s not 100% OK, it’s [functional] enough to help your lungs to breathe,’ Komander explains.

Ubiquitination beyond proteins

The opportunities to translate the ubiquitin code into life-saving therapies are immense but may be the tip of an iceberg. A discovery by Komander published in April could open up even more avenues. ‘We are able to detect ubiquitination beyond proteins, including on sugars,’ he says. This follows a 2021 discovery that bacterial lipids from salmonella were ubiquitinated in human cells as part of an immune response, but Komander’s work is the first example of the mechanism being applied to a cells own non-protein molecules.

These results suggest that ubiquination of sugars also provides a signal for degradation. When mice were fasted and in need of glucose, the presence of ubiquitin tags on glycogen increased and glycogen levels depleted. The discovery of this new pathway for regulating glycogen breakdown on demand could perhaps one day be therapeutically exploited.

His team also uncovered ubiquitinated metabolites like glycerol and spermine. ’We were entirely blind to anything that was beyond proteins, and it was shocking to see how many ubiquitinated species we had missed in a normal cell lysate,’ says Komander. ‘It’s a bit mind blowing, really, it’s a level of metabolism that [we did not know] to exist.’

These findings are part of an even more complex ubiquitin code being uncovered. Ubiquitins have also recently been found to bind to the amino acids serine and theronine. Plus, it’s clear that ubiquitin isn’t the only type of tag. From the 1990s a whole family of ubiquitin-like proteins were discovered that share a structural similarity with ubiquitin and work in similar ways. ‘To some extent, you cannot avoid the ubiquitin system because it’s such a powerful, fast, beautiful and complex [system],’ says Ciulli.

Rachel Brazil is a science writer based in London, UK