For all those wondering how variant mutations can screw up antibody responses, have I got the preprint for you!

Buckle up for a long ride down epistasis & biochem road, thanks to this great study by @Dr_MattMcCallum and colleagues in the @veeslerlab & collabs at @Vir_Biotech.
First, some background. All the variants have different constellations of mutations in SARS-CoV-2 spike. This is the protein on the surface of the virus particle (virion) that bind the receptor ACE2 and allow the virus to enter & infect cells.

It looks like this (h/t @profvrr):
As you can see from the above virion, spike is a 3D structure on the surface of the virion. Antibodies bind all over the surface of the spike protein. Some of these bind to important parts of spike that render the virus non-infectious, or neutralize it.
We have an enormous repertoire of antibodies that bind specifically to different places (epitopes) on the surface of spike. Each epitope is a unique "lock" that will only fit with a specific antibody "key."

(H/t @dsgoodsell for this gorgeous image. Antibodies=yellow, spike=pink)
And to understand how mutations impact this, we need to understand what proteins are. Proteins are encoded by viral genes, and that code contains instructions to assemble a string of amino acids. Long strings of amino acids can fold into different 3D structures, like yarn.
Amino acids look like this. They have an amino side (N-terminus) and a carboxyl side (C-terminus). The N-terminus of one amino acid binds the C-terminus of its neighbor to make the chain, or peptide. That bond is called a peptide bond.…
See that R hanging off the carbon in between the N and C termini of the amino acid? That's called a side chain. These are different for each amino acid and can chemically interact with other amino acids to hold the 3D structure together.

(Bear with me, biochem part almost over)
If you liked this part of the thread, Khan Academy has a great explainer of all the different biochemical interactions that can contribute to the 3D shape of proteins.…
If you didn't like this part & didn't want intro to biochem, it's because this is how mutations can influence antibody binding.

A mutation in the genome can change the encoded amino acid, which can change the way amino acids interact, and change the structure of the protein.
And that's what this preprint is about!!!

Mutations in the B.1.427/B.1.429 variant discovered in California not only change protein structure, but they do so *in combination* with each other.…
This particular variant has 3 spike mutations: S13I, W152C, and L452R. To decode this:

Serine changed to isoleucine at position 13 in the chain
Tryptophan changed to cysteine at position 152
Leucine changed to arginine at position 452
We already knew that the B.1.429 variant is less effectively neutralized by antibodies in a pseudovirus assay compared to other wild-type (WT) SARS-CoV-2. It resembles the B.1.351 variant first detected in South Africa.

But B.1.429 has different mutations in spike than B.1.351.
So what's going on here? Well, the L452R mutation is in the receptor binding domain (RBD) of spike. This is the part of spike that binds ACE2 and allows it to get into a cell. ACE2 can't bind the RBD if there's already an antibody there and the virus can't get in.
And last September, we already had evidence that the L452R mutation resulted in reduced antibody neutralization based on this paper. It showed L452R reduced neutralization by both a panel of monoclonal antibodies & convalescent sera from COVID-19 patients.…
So that seems pretty straightforward and this paper found similar results. Some antibodies targeting the receptor binding domain didn't neutralize B.1.429 as well as WT virus. Yes, L452R disrupts binding of RBD-specific antibodies.
But what about those other 2 mutations, S13I and W152C? They aren't in the RBD. They are in a different part of spike called the N-terminal domain (NTD). It's at the "front" of the spike protein, pictured here in this structure from @McClellan_Lab.…
Spike on the virion surface is actually a trimer. That means it's 3 individual spike proteins (protomers) put together. The protomers can interact with each other & other proteins (like ACE2 or antibodies) and that can change their shape. This can impact function.
Spike has several functions besides just binding ACE2 & those functions are conformation-dependent. So antibody binding to other parts of spike besides the RBD (like the NTD) can still be neutralizing if it blocks a critical function.…
We know the NTD is important because of studies like this one by @florian_krammer and colleagues. Here it shows that antibodies to the NTD, not the RBD, have reduced neutralization against a variant with mutations in the NTD.…
And it looks like that's what's happening here too! NTD antibodies completely failed to neutralize this variant. How is that happening?
Well, S13I is in a part of the NTD that encodes the signal peptide. Most proteins have these. They are a short stretch of amino acids that tell the protein where to go in the cell, like a barcode used to sort & direct proteins to the nucleus, cell surface, or wherever.
Signal peptides are clipped off by proteases (enzyme that cuts proteins at a specific amino acid sequence). Most proteases are serine or cysteine proteases, and the "cleavage site" they recognize has either serine or cysteine.
In the wild-type, that's the serine at position 13. When this is mutated to an isoleucine, it shifts the cleavage site two doors down to the cysteine at position 15.

They used mass spectrometry to confirm that was indeed the case.
Okay but this is only a 2 amino acid shift. What's the big deal? Well, serine and cysteine both have another important quality. They both contain sulfur in their side chains, and are the only amino acids that do. They can also form disulfide bonds, where two sulfurs bind together
This is the biochemical equivalent of holding hands. Normally C15 "holds hands" with the cysteine at position 136, and this contributes to the 3D structure of the NTD. But when the signal peptide is cleaved at C15 instead of S13, it's like chopping the C15 hand off.
However, when W152C is in the mix, there's another cysteine down the road for the C136 hand to hold. That changes the overall 3D conformation of the NTD. Bad news for NTD-specific neutralizing antibodies.
And here's all that data, confirming that indeed this causes a big old reduction in antibody neutralization, due to the conformational rearrangement of the NTD caused by these two mutations working in tandem.
When individual mutations have a much more profound impact when they occur in combination it's called epistasis.

That's why it's hard to tell what properties the different variants will have just by looking at the sequences and defining the mutations.
Some mutations can have no effect. Some have a minor impact when they occur by themselves. Some can have an outsized impact when they occur with others.

I really liked @K_G_Andersen's "hanging out with the wrong crowd" explanation of epistasis:
The bottom line is that as new variants emerge, we need to do studies to understand their significance to public health. And in doing so, we might just learn some cool new mechanisms of immune evasion by SARS-CoV-2.

Congrats and thanks to the authors for such fantastic work!
One quick addendum: antibody neutralization is not the totality of the immune response, or even the antibody response. Thus reduced neutralization is not an indicator of how well vaccines work.
This describes just one mechanism of immune evasion. The immune system is complex and multifaceted, and can't be escaped by one mutation or even a handful of them. Clinical and real-world data also shows that vaccines still protect against this variant (and others).
Dammit, there's a typo. I meant to tag Dr. Jason @McLellan_Lab!
Addendum #2: I am embarrassed to make such a basic mistake, but I mixed up my AAs. Methionine (not serine) & cysteine are the amino acids with sulfur. I was thinking too much about serine & cysteine proteases and got carried away. Also it's Friday. Forgive me, biochemists!
Reposting here: I am embarrassed to make such a basic mistake, but I mixed up my AAs. Methionine (not serine) & cysteine are the amino acids with sulfur. I was thinking too much about serine & cysteine proteases and got carried away. Also it's Friday. Forgive me, biochemists!
And yes, I know methionine can't make disulfide bonds. If it could, it would be...cysteine.

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Dr. Angela Rasmussen

Dr. Angela Rasmussen Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @angie_rasmussen

31 Mar
BREAKING: Major and Champ Biden are dogs who do normal dog things
And all the people saying normal dogs don't poop in the house has never owned a senior dog or a pug of any age.
Also BREAKING: Ripley thinks Major and Champ seem cool and she’d like to compare indoor pooping techniques with them at some point
Read 4 tweets
31 Mar
I’ll always make time for @apoorva_nyc, especially about great news like this: top line trial results show safety, robust antibody production, & 100% efficacy in 12-15 year olds!
W/ the wise & wonderful @VirusesImmunity, @JenniferNuzzo, & Kristin Oliver.…
Approach the efficacy data cautiously, because it was based on just 18 cases, all in the placebo group, but this is what we’d expect to see: the vaccine works in adolescents at least as well as in adults (maybe better, at least in terms of antibody responses).
While this data still needs to be evaluated by FDA and we need to see the full trial data, parents of teens can rest easier knowing their kids will likely be able to get vaccinated before school starts in the fall.
Read 5 tweets
30 Mar
Barely out and already my feed is filling up with (some pretty racist) complaints that this report is incomplete and dissatisfying.

But did you really expect a 2 week mission to yield a definitive answer about SARS-CoV-2 origins?

Origin investigations take years, even decades.
The purpose of this mission was really to lay the groundwork for collaborative studies moving forward.

Like it or not, that requires working cooperatively with China.

Like it or not, @WHO isn't equipped to conduct an audit of WIV's freezers or records or interrogate its staff.
This report contains some new information: about excess mortality, ILI, environmental testing, retrospective sample testing, animal testing. It also acknowledges there is a lot more to do.
Read 9 tweets
21 Mar
2 things have been troubling me lately:

1. The notion that recognizing the nuances & uncertainties of SARS-CoV-2 transmission is somehow "denying" that respiratory transmission occurs.

2. This is all @WHO's fault for giving bad guidance.

Are these true? Let's look at the data.
1. SARS-CoV-2 is a respiratory virus. That means it infects cells of the respiratory tract in both the upper (nose) and lower (lungs) airway. Naturally, this means you become infected after exposure to virus via inhalation, direct contact (droplets), or indirect contact (fomites)
Infection occurs not from exposure to purified virus, but from virus that's emitted in particles of saliva and mucus that an infected person exhales (or speaks, sings, etc).

Go outside on a cold day and breathe out. That cloud of steamy breath? Those are respiratory particles.
Read 48 tweets
19 Mar
No, you can’t. Kids being lower risk does not mean “same as an already vaccinated grandparent.”

I value other perspectives, like economics, in vaccine discussions. All are stakeholders in public health. But have the humility to know when you’ve hit the limits of your expertise.
This is insufficient. The problem wasn't just the headline. The problem is that two of the fundamental arguments are wrong.

1. Risk of infection is just as relevant as risk of disease.

2. Vaccinating only adults won't get us to herd immunity by summer.
And I should add there's another risk that wasn't addressed at all: the vaccines are not 100% effective (though they are very good). Failing to mitigate transmission risk by relying on flawed risk assessments still presents a danger, especially to more vulnerable adults.
Read 4 tweets
18 Mar
So, here's another great paper out today from my friend and colleague @DrSimonAnthony.

It will be of particular interest to those pondering possible SARS-CoV-2 origins, and viral evolution enthusiasts in general.

Brace yourself: it's about how different sarbecoviruses use ACE2.
In short: Dr. Anthony and his PREDICT colleagues found 3 isolates of a novel sarbecovirus (SARS-like betacoronavirus) in bats...from Uganda and Rwanda. Hey, that's nowhere near China, so what does that have to do with SARS-CoV-2?
A lot, actually. Sarbecoviruses are not exclusive to China or East Asia. That's where they've been studied most extensively, but they aren't restricted to that part of the world. Here's how these viruses and host bat species relate to each other and where they were discovered.
Read 19 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!