Design an anti-inflammatory protein for COVID-19! We'd like players to keep focusing on making closely-packed interfaces with no extra Buried Unsats. This puzzle prohibits SER and THR in helices, which may make it harder to satisfy BUNS, but should help the helices to fold correctly. Remember, if your designed protein creates Buried Unsats, then it will be less likely to fold and bind to the target. (Note that this target protein includes 15 buried unsats that players may be unable to fix.) See the blog for more details about buried unsats, and for helpful tips to make a successful protein binder! Players will be unable to load solutions from previous puzzles.
Many COVID-19 complications are caused indirectly by the virus, and result from a severe over-stimulation of the human immune system. This kind of immune over-stimulation is commonly called a "cytokine storm." During a viral infection, immune cells normally release signaling proteins called cytokines, which inform the rest of the immune system about the infection and trigger inflammation. The inflammation is supposed to help the immune system fight off the infection, but too much inflammation can result in sepsis and organ failure.
One proposed strategy for treating serious COVID-19 cases is to prevent the "cytokine storm" by blocking certain cytokine signals. We want to design a protein that could block cytokine IL6, by binding to the IL6 receptor (IL6R). For more details, see our YouTube video about blocking the cytokine storm.
In this puzzle, players are presented with the binding site of IL6R, which receives cytokine signals and triggers inflammation. The backbone and most of the sidechains are completely frozen, except for sidechains at the cytokine binding site. In order to bind the IL6R target, designs will need to make lots of contacts and H-bonds with the target protein at this binding site. But designs will also need to have lots of secondary structure (helices or sheets) and a large core, so that they fold up correctly! See the puzzle comments for Objective details.
Buried Unsats (max +100)
Penalizes 60 points for each polar atom that cannot make any H-bonds. Note that the frozen target includes 15 buried unsats that may be impossible to satisfy.
Residue Count (max +275)
Penalizes extra residues inserted beyond the starting 151, at a cost of 55 points per residue. Players may use up to 156 residues in total.
Core Existence (max +1000)
Ensures that at least 25 percent of residues are buried in the core of the monomer unit.
Ideal Loops (max +500)
Penalizes any loop region that does not match one of the Building Blocks in the Blueprint tool. Use "Auto Structures" to see which regions of your protein count as loops.
SS Design (max +500)
Penalizes all CYS residues. Penalizes GLY, ALA residues in sheets. Penalizes GLY, ALA, SER, THR in helices.
It's likely this question has been asked before, but why not use Hydrogen Bond Networks in these types of puzzles? Maybe its a limitation in Foldit but it seems they'd be useful in making the binding specific.
So when are we going to get to start designing proteins with alanines in the helices, like in real life?
If you really don't want to see alanine-rich helices, aan't you add a penalty function that allows no more than, say, ten percent of helix residues to be alanines?
We absolutely could reward H-bond Networks in these puzzles, but I think the BUNS objective is more appropriate.
Ultimately, the specificity is driven by polar atoms at the interface. The polar atoms fend off unintended binding partners, because these atoms will become BUNS if they are not satisfied by complementary polar atoms on the target.
Specificity is a big problem in symmetry puzzles, because we can design both sides of the interface. This means we have the power to design interfaces that are entirely hydrophobic (and these typically have the best Foldit score). We specifically want to encourage the addition of polar atoms at the interface, and this is the role of the H-bond Network objective. The BUNS objective alone does not encourage new polar atoms.
This IL6R binder puzzle is different. The target IL6R protein already has lots of polar atoms at the interface. It is basically impossible to design an entirely hydrophobic interface – any top-scoring solution will have to bury some polar atoms on the target, so we are very worried about creating BUNS. The BUNS objective will force us to add polar atoms at the interface, and we think this alone should create the specificity we want. There is no need to encourage more polar atoms with the H-bond Network objective.
That's exactly what we'd like to do, and it is on the to-do list. Currently it's taken a back-seat to other priorities for protein design. That penalty shouldn't be too difficult to add (although we do need to make sure it plays nicely with the Mutate tool); hopefully we will be able to tackle it soon.
I currently have the high soloist score on this puzzle. I have no sidechains from my designed protein binding to any sidechains of the target protein. Yet I have thirteen BUNS (and there are supposedly fifteen we can't fix).
Is this a flaw in the scoring? Isn't the goal to create a protein that will bind with this target? I don't think I've created one, yet my score implies otherwise.
I'll have to take a look at your specific solution, but you're right that the Foldit score alone is probably not sufficient for binder design. There are several binder metrics that other researchers have found useful for binder design, and these will be available in Foldit soon.
It sounds like you've found a way to bind to the target without burying any polar atoms at the interface (which is great!). There is a possibility that the BUNS objective is missing some buried polar atoms – the BUNS objective relies on some approximations that trade accuracy for speed. But, most likely, if the BUNS objective fails to flag any atoms at your binder interface, that means all of the polar atoms are satisfied or able to make H-bonds with the surrounding water.
