Skip to main content

Chemical Biology

Watching Protein Degradation Happen

You know that a technique has attained wide currency when the vendors start selling reagents and tools around it. Here’s an example in the protein degradation field: a team at Promega (well-known vendors of assay reagents and kits) report a new system to evaluate the extent and time course of protein degradation compounds.

For those outside the field, the idea behind these is to hijack the cell’s normal machinery to recycle proteins when they’re damaged or no longer needed. The ubiquitous small protein ubiquitin is stapled onto lysine residues of the protein to be degraded, via a ubiquitin ligase complex, and this serves as a marker for transport to the proteasome. That organelle is like one of those giant shredders you see at a car-recycling facility, and it rips whatever protein it gets ahold of into its constituent amino acids for re-use. That’s the normal state of affairs.

For targeted degradation, you assemble a molecule (often rather odd-looking by the time you’re finished) that has a ligand for your desired marked-for-destruction protein at one end, a linker group, and a ligand for one of the ligase enzymes at the other. This brings the two into artificial proximity, whereupon the ubiquitin ligase decides to do what it does best and start slapping ubiquitins on this protein that’s suddenly been brought within range. And that lead to said protein being hauled off to the proteasome and demolished, and if proteins were capable of surprised angry protest, it would be shouting all the way. It’s very much like E. O. Wilson’s experiments when a live ant gets dabbed with a bit of “dead ant” smell and keeps getting dragged off to the dump heap until the smell finally wears off. In this case, though, the protein is destroyed in the first pass, and the degrader molecule is released to go set off another cycle of destruction.

Now, this is an exciting technique, because wiping out a protein like this can be a very different thing than just blocking some small-molecule binding site it might have. But more tools are definitely needed. As it stands now, protein degradation seems to have rather too high a Voodoo Quotient. It can be difficult to figure out, for example, what kind of linker to put between your two binding ligands – what it should look like and how long it should be. And different ligands can have different effects, even when they’re binding the same target protein, and (naturally enough) recruitment of different ubiquitin ligases can make a big difference, too. Overlaid on these are the problems of cell penetration and compound stability, which in practice can mean a fair amount of hunt-and-peck as you try to find a system that does what you want it to do. There are surely rules, but we’re still learning them.

Promega’s technique is based on their widely-used NanoBit technology. They use CRISPR/Cas9 to append the (11-amino-acid) HiBiT peptide to a protein of interest (in this case, the three bromodomain proteins BRD2, BRD3, and BRD4). That small peptide is engineered to bind very tightly to the “Large BiT” protein, and together they form a functional, stable, and rather bright luciferase enzyme. If there is some substrate around for it (added by you, the experimenter), this binding event lights up in a very easily detectable manner. In this case, they engineer the expression of the LgBiT protein in the cells as well as the altered BRD proteins.

That gives you a tracker for the levels of all three of those proteins, since they are now luminescent. You’ll want to make sure that things are working correctly (which this paper naturally did) by checking expression levels of the proteins involved, making sure that they’re the only things lighting up, doing some microscopy to make sure that they’re showing up where you’d expect them to be, and so on. But once you’ve sorted these things out, you’re ready to go. In this case, they’re using the well-characterized MZ1 degrader, which has a bromodomain ligand (JQ1) on one end, and a ligand for the VHL degrading complex on the other, and the dBET1 degrader, which has the same bromodomain ligand end, but recruits a different ubiquitin ligase complex (CRBN).

What you see are that (1) the more degrader molecule you add, the more BRD protein degradation you get, but only up to a saturation point, (2) BRD2 and BRD4 degrade much more quickly than BRD3 if you’re using MZ1, but not if you’re using dBET1, and (3) the time needed to reach maximum degradation varies quite a bit depending on the protein and degrader. And after that maximum has been reached, the protein concentrations start to fight their way back up as the degrader molecules themselves get cleared out. These recovery rates vary as well; some proteins have more compensatory mechanisms for resynthesis than others. The paper makes the point that you have to pick your time points carefully in comparing degraders, because you can get completely different profiles (or miss activity altogether if you were to wait too long before taking the first points). The authors suggest determining maximum degradation, no matter the time course, as one comparison, and the time to reach that Dmax as another.

Interestingly, they also tried the experiment where BRD4 had the complete intact luciferase (NanoLuc) fused to it instead of relying on the two pieces assembling. Several other degradation studies have used such modified proteins, but it turns out that this gives you a completely different profile. Everything looks worse: less of the protein gets degraded, it takes longer to happen, and the recovery rate is faster. So that’s definitely something to consider when you’re evaluating these things. Assay development is full of this sort of thing, and unless you have several different approaches going, you have no way of knowing what part of the landscape you’re in.

The paper also describes use of a related technique (NanoBRET) to study the kinetics of these processes (that one’s better for transient interactions). Among the things revealed by this system are that both degrader compounds do indeed appear to be catalytic – their cellular penetration is not great, and substoichiometric concentrations are all that can be achieved, but they’re still very efficacious. And the team also monitored the levels of a key downstream event of BRD inhibition/degradation, the loss of the well-known oncogenic protein cMyc. This illustrates yet again the power of protein degradation: treatment of tumor cells with the JQ1 ligand alone has only small effects on cMyc levels. But treatment with dBET1 showed a 50% decline at 24 hours, and MZ1 knocked it down 80%.

Making a protein disappear completely on demand (non-genetically) is something we haven’t been able to do until this technique came along, and there are going to be a lot of such studies needed to figure out what it’s doing for us and how best to do it. Bring on the assays!


11 comments on “Watching Protein Degradation Happen”

  1. ChrisP says:

    Do we know how much biological feedback is done by degrading proteins compared to inhibiting them? In-vivo, is a cell going to generate more than usual of the degraded proteins over time?

    1. Derek Lowe says:

      Good question. There’s a huge range of half-lives for proteins, ranging from weeks/months/years down to about 20 minutes or so. So turnover can definitely be a regulatory factor, although you’d have to figure that its importance varies just as widely as those half-lives.

  2. Some Dude says:

    “Making a protein disappear completely on demand (non-genetically) is something we haven’t been able to do until this technique came along”

    Have you seen this method using Trim21?

    1. Derek Lowe says:

      I have indeed. It’s very nice, but it suffers from having to use microinjection to get the antibodies into cells, IMO.

      1. Joe Breuer says:

        Thoughts on using liposomes for delivery?

  3. William Koesters says:

    Arvinas just filed an IPO based on a protein degradation platform…

    first candidate prostate cancer…

    Androgen receptor degradation by the proteolysis-targeting chimera ARCC-4 outperforms enzalutamide in cellular models of prostate cancer drug resistance

  4. Thoryke says:

    Is there a risk of a cell-level equivalent of tumor lysis syndrome, where there are suddenly too many degradation products for the system to handle?

    1. Derek Lowe says:

      Hadn’t thought about that! My guess is that this would have the best chance of happening if you targeted a very abundant protein and hit it very hard, and in that case, the cell would have plenty of other problems as well. . .

  5. Scott says:

    Very interesting stuff (and Thank You, Derek for keeping the descriptions of this down to ‘close enough overhead that I need to duck’ 😀 )

    I think I can see the next grant applications: “This is a foundational-level study. Now that we can physically see our proteins degrade in real-time, we’re going to apply the technique to every single protein we can think of and see what happens. We hope to develop additional drug targets and new methods of action this way, but acknowledge that further studies will be needed to get there.”

  6. Joy says:

    Would the dead-ant thing work on politicians? asking for an endangered democracy.

Comments are closed.