[1] is not generally supportive of your very broad statement. You must qualify you statement with some indication of this stated limitation: "carrying “single letter” changes in a key portion of their spike protein"
No, there are not solved structures for these variants (in general). They're using computer modeling.
Even if there were, changes of a single amino acid that don't induce significant conformational (shape) change are the sort of thing that you'd need a very good structure to discern, since they're mostly on the surface of the spike, and probably spin around a lot.
The most relevant thing to know is what a mutation / mutations looks like in the context of antibody binding, but that's even harder to get a good structure. It's not something that can be done quickly.
This is not very convincing, though of course it should be followed up on. Considerably more is known about the structure of the virus than the article seems to imply.
I'd like to see more about what they mean by simulations "in water". If they didn't simulate intermolecular forces at all, these results seem unlikely to be relevant.
The details really matter here. Many early visualizations of the spike protein only showed the protein amino acid residues, not any attached sugars (glycosylation). Turns out the sugars act as a reconfigurable armour that was preventing targeting of the spike protein by meds.
It would be great if there was an "expert" mode on this, with a bit more detail on the coat protein structures. Modern molecular biology is so abstracted, you often lose sense of the actual scales of things. Being able to look at things as a whole like this really enable intuition to kick in.
It looks like they've tried to get surface density of proteins right, but I'd love to know if those densities are about right. This is all important for understanding viral receptor binding, and the importance of multi-valent interactions.
Yes there is. It's the most easy way to verify - see if the protein matches with the problem receptor. It's pretty much black or white if something binds to a receptor.
I think the perspective just needs to change. Maybe a little handheld "protein receptor match" kit that you plug your numbers into, and it will "make" your particle to try to match in a small lab kit that is portable.
Fascinating that they have to physically screen individual molecules one at a time to see if they can bind with a specific protein. Even at a Princeton lab, it took years. Are there other institutions/companies working in this space?
“We could search for a new protein that matches the melody and rhythm of an antibody capable of binding to the spike protein, interfering with its ability to infect,” said Professor Buehler.
Interesting. Encode the info, in potentially, an easier format for comparison.
This is a slightly different result, actually - they're suggesting that these different variants are actually proteins of unrelated function. I'm not sure I buy this off a vast interaction study, which is a pretty artificial and noisy way to scan protein function, but it's a novel claim.
>"2) Looking at this seqence has shown several reasons why proteins don't often go viral"
Yes, but to clarify a bit: It looks like all proteins have this ability, but not all are just as likely to get into that state for the reasons mentioned. IE, it should be possible to find some way to make amyloids/prions out of any protein in the lab. It just requires figuring out the necessary conditions for that protein.
>"The elucidation of this code has enabled the identification of factors that determine the intrinsic aggregation propensity of these molecules"
Translation: Being able to get the amino acid sequence of proteins and predict what type of structures such a sequence leads to has allowed researchers to also predict which are likely to form aggregates.
Sorry, I realize those excerpts in my earlier comment don't really convey what I intended. The paper does talk about about how the various improvements over the years were changes in just a few (or even just one) amino acids in the molecule. It's a large molecule, nearly 6000 daltons (for reference, the limit for what we call small molecules is around 1500 da), so there's been a lot of opportunity for tinkering.
I'm sorry, that is simply not the case, neither at the protein level nor at the DNA level. Furin cleavage sites have diversity.
The canonical minimal furin cleavage site sequence is RX(R/K)R, which had a highly variable amino acid and a choice between the two basic amino acids in another position. This variability is sampled across the coronavirus furin cleavage site sequences.
Thanks for the info! That ability to retune sounds excellent. I wonder if you could spin this out to have an EWS for influenzas too (although hopefully it’ll never get as severe!)
I guess for the glycosylation when you say it’s implicitly handled, there’s a N-linked sequon pattern somewhere in the language model, which I guess covers a good deal of info :)
Jury still out on effects of O-linked glyco on spike, but give me a shout if you’re interested in it!
Anyway, cool work, and all the best in where you’re taking the work next!
[1] https://jbloomlab.github.io/SARS-CoV-2-RBD_DMS/
reply