Discovery of a druggable pocket in the SARS-CoV-2 spike protein | Hacker News Reader

Discovery of a druggable pocket in the SARS-CoV-2 spike protein

medicalxpress.com

87 pointsvirgilp61 comments

So, it's great that they're investigating, but this reads a bit too breathless to be taken entirely seriously: "stop virus in its tracks" "a potential game changer" "ground-breaking study"

...but further down: "The question now is how to turn this new knowledge against the virus itself and defeat the pandemic."

Yes, that would be the question, which is much closer to the beginning of the process of developing and producing and rolling out a new treatment, than it is to the end.

In other words, while it's great that this kind of thing is being investigated, it is extraordinarily unlikely to impact _this_ pandemic. No doubt Covid-19 will come back again, several years after this pandemic has run its course, and it's great to investigate how to be better prepared for that. But this is not going to result in a treatment fast enough that this pandemic will not already have spread to its (historically normal, typical, well-nigh-inevitable) conclusion.

COVID-19 (SARS-CoV-2) won't "come back again" because it will never go away in the first place. It has already spread so widely that it is now endemic in the worldwide human population (plus some other mammal species). Just like at least 4 other coronaviruses.

But we made no attempt to eliminate those corona viruses and we have, the last I checked, something like 20 vaccines in development for this one, if anything is a candidate for going the way of smallpox it’s sars cov 2, there’s too much inertia behind our efforts to stamp it out, it might just take a while.

We were able to eradicate smallpox because it had no significant animal host reservoir. That isn't the case with SARS-CoV-2. Are you going to vaccinate all the bats?

Technically, you're correct, but the 1957 flu "came back" in 1968, and again in 2009 if I recall correctly. Yes, I'm sure they were never completely gone, but at some point there are enough non-resistant people and it flares up again, and we would want to have better options for that "next time". But yes, technically, you're correct, it will never quite go away.

> In rhinovirus, a virus causing the common cold, a similar pocket was exploited to develop potent small molecules that bound tightly to the pocket distorting the structure of the rhinovirus, stopping its infectivity. These small molecules were successfully used as anti-viral drugs in human trials, defeating rhinovirus in the clinic.

So why isn't there a cure for the common cold? If there still isn't a cure for the common cold, why should we expect this drug target to be different?

Surely whoever wrote this article should have anticipated these questions?

(Edited to be less flippant)

On the other hand, Oracle is mentioned twice, in case we didn't spot it the first time.

The Oracle plugs are hilarious. "Enabled by Oracle high-performance cloud computing, a 3-D structure of SARS CoV-2 Spike protein was generated allowing the researchers to peer deep inside the Spike identifying its molecular composition." As if the cloud provider had anything to do with anything there.

I only trust IBM® Watson™️ Brand clinical research.

I heard Watson cured all of cancer!

Now it's quite clearly a plug - but serious amounts of compute power are required for protein analysis, and chucking cloud at it is a pretty good way of doing that.

Look at Folding@Home

Sure, but they intimate that only Oracle could get the job done.

Oracle cloud must be absolute garbage given how hard its being marketed. Didn't part of the US deal with Tiktok _force_ the application into using Oracle cloud as part of the deal? I have a lot of sympathies for the TikTok devs if its not very good as its now mandatory.

Its like instead of building something good or getting techie buy in they've instead tried to lobby governments to make it mandatory for some enslaved customers.

> Its like instead of building something good or getting techie buy in they've instead tried to lobby governments to make it mandatory for some enslaved customers

Haha, yes, you are literally describing the business model of SAP, Oracle, IBM, and most sales-led enterprise software or services companies. You can replace "lobby governments" with "lobby c-level of large corporations" as well

I do recall Microsoft using a similar tactic to try to get Windows Mobile more widely used.

Don't tug too hard at that thread, the entire narrative might unravel.

What does that mean?

Feels like an AD. No real author. Most of the paragraphs are long quotes with no follow up or analysis. Suspicious product placement.

The common cold isn't any specific virus, it's a huge range of viruses that cause "cold" symptoms. Some coronaviruses cause "colds" as well. Common cold viruses are research targets to discover knowledge that can be weaponized to prevent pandemics.

I almost never comment, but the answer here is simple: markets.

The common cold doesn't kill. There's almost no economic impetus for a cure. Note, I said "almost." There may be some, but there's still not enough to drive R&D in a corporate lab, and certainly not enough to drive policy in a government lab.

There are 100 other questions that would follow from science to logistic once you answer: Can we create a drug that kills infection in clinical tests?

Surely there's a Nobel Prize in it though?

NyQuil’s and Alka Seltzer’s sales figures beg to differ.

Those aren’t cures.

You'll make much more money selling symptoms relievers every years than fixing the issue once and for all.

You'd make a lot more money selling a yearly treatment that prevented the symptoms entirely. I probably spend $50/year on cold-symptom medications. I'd easily spend $100/year to just not get a cold in the first place.

NyQuil and Alka Seltzer would beg you not to kill the cash cows.

Plus NyQuil and AS can be produced and sold in such a way as to net a profit. We don't know that's the case for a cold cure.

Those both work against symptoms caused by an enormous range of pathologies though.

Compare that to a specific drug targeting a small percentage of the causative agents of the common cold, for which the duration is short enough that by the time you'd get a reliable test back, you'd have gotten over your cold anyway.

There's just not much of a practical use for such a drug.

Sounds like that is the prevailing opinion amongst software engineers on HN. “If the product doesn’t exist, then the market doesn’t exist” isn’t how biology works. I assure you that many in life sciences would kill to get their hands on a successful therapeutic program in this area. There is a wide world of markets available.

I'm not a software engineer and I used to work in several sectors of healthcare (emergency medicine and family practice). The drug target against rhinovirus already exists and I was explaining why it's not yet on the market. A drug (with potential side effects or allergies) against maybe 10% of 200 some odd common cold viruses (which typically have very little impact) isn't that useful, and I don't know any physicians that would consider it acceptable practice to order and incur the expense of a test and then write a prescription to address an issue that will resolve itself in just a few days anyway.

That's not to say it might not be worth it for someone with COPD, lung disease or serious immunocompromise or something, but you seem to be lacking a pragmatic understanding of the situation described, yet you're readily criticizing me on the basis of an (incorrect) ad hominem argument.

Tamiflu is a much better example than Nyquil (which you offered up earlier) where it does make sense. But in that case, influenza typically lasts longer, it's more serious, it's much easier to diagnose, and the potential complications are much more severe and likely. It's probably more prevalent as well.

Sorry, I was summarizing several replies and wasn’t trying to make a personal attack against you. My hope was to point out that there are other ways of looking at things.

You seem to be coming at things in terms of patient trade-offs. I’m saying that AbbVie loves selling more Humira, and some sort of hypothetical magical anti-infective would’ve been a way to do so.

Cold can kill, viral pneumonia is real as our current pandemic can attest to. There's certainly a market to cure all sorts of viral illness. Take tamiflu as an example. Most people if they get a flu will use it once a year to reduce symptoms. Imagine a drug that would reduce all symptoms of cold that you take every time you get cold symptoms. That's potentially billions of dollars in revenue each year.

I would actually argue that it should be the government's role to push the development of vaccines and drugs for treating even harmless diseases. We spend trillions of dollars on national defense in the US, when you look at the economic damage and cost in lives that the Covid-19 outbreak has cost us, the threat of disease outbreaks is every bit as damaging as a military attack, maybe even worse. If we had taken the time to develop and take all the way through the approval process a vaccine to previously encountered Coronaviruses such as SARS or MERS we would probably already have a vaccine against SARS-CoV-2 approved by the FDA and in distribution. Maybe we should be building a few less F-35s and instead be developing vaccines against otherwise harmless viruses, because you never know when a close relative of a harmless virus is going to turn into a deadly outbreak. In the US at least that seems like it should fall under the protective role of the federal government.

The economic damage is coming mainly from economical policies, namely lockdowns. The death rate has been quite low overall, even including the initial phase in which our incompetence in curing covid caused more deaths than it should have.

I agree the government wastes money on a lot of things (13bln per day!) and that's exactly why I'd rather have the market come up with what we should be spending our money on.

https://journals.plos.org/plosbiology/article?id=10.1371/jou...

It might be from 2019 so there may simply have not been enough time.

Definitely not enough time for traditional pharma to have actionable, public insights.

>> So why isn't there a cure for the common cold?

The "common cold" refers to a benign, self-limited upper respiratory viral infection by over 200 viral subtypes, including many strains of coronavirus (causing collectively 10-15% of colds), adenovirus, enteroviruses (echovirus, coxsackievirus) rhinovirus (over 100 serotypes, collectively causing 30-50% of colds), metapneumovirus, bocavirus, influenza and parainfluenza viruses (~5% of colds each), and respiratory syncytial virus (another 5%).

I think this is general lay knowledge. Certainly, any article that uses the words "glycoprotein" is aiming at the sort of audience that knows "common cold is a syndrome caused by a buncha different viruses."

When I go for the usual flu vaccine in November, how would I know which strains its protecting me against?

There is no "usual flu vaccine", they make a new one twice a year, meant to address whatever strain is in circulation.

I think they meant when they go for their regularly scheduled -"usual" flu vaccine, and their question was meant to address what strains would be targeted in a given year.

There's cross-reactivity, and they target both Influenza A and Influenza B strains (usually), so you'll get broader protection than strictly the strains most targeted by the vaccine. That said, you can ask the pharmacist for a copy of the FDA package insert, or google it yourself for the specific vaccine they're giving you. The package insert, under "DESCRIPTION" (Section 11), should tell you the precise strains that the antigens were developed from.

For example, google "FLUVIRIN fda package insert",and the one for the 2017/2018 flu season on page 8 you'll find "A/Singapore/GP1908/2015,IVR-180 (an A/Michigan/45/2015 (H1N1)pdm09- like virus; A/Hong Kong/4801/2014, NYMC X-263B (H3N2) (an A/Hong Kong/4801/2014- like virus); and B/Brisbane/60/2008, wild type (a B/Brisbane/60/2008-like virus)".

Google the same document for Fluzone quadrivalent, and you should find, for the 2020/2021 flu season, the strains "A/Guangdong-Maonan/SWL1536/2019 CNIC-1909 (H1N1), A/Hong Kong/2671/2019 IVR-208 (H3N2), B/Phuket/3073/2013 (B Yamagata lineage), and B/Washington/02/2019 (B Victoria lineage)" on pg 21.

There's also a summary maintained at Wikipedia

https://en.wikipedia.org/wiki/Historical_annual_reformulatio...

although it's presumably taken from the same sources you suggest.

In 2017, for example, the vaccine didn't work all that well, because of this exact problem. My dad, who lived in an eldercare facility, was confined to his room many days of the winter, which is something I sympathized with at the time, but sympathize a whole lot more with after this April's shelter-in-place order.

The deaths in the U.S. in 2017 were about 61,000, compared to the normal 40,000 or so.

Long story short: you don't know, but it's the best guess of those in charge of making that call.

I think if I was in your father's situation I would just go out and take my chances. At the age when I have to live in an eldercare facility I'll already be at a constant high risk of death from other causes. What's the point of avoiding the flu if it means I'm stuck in solitary confinement where a stroke might kill me tomorrow?

Not always benign. HCoV-OC43 may be roughly as deadly as SARS-CoV-2 (COVID-19) to immunologically naive patients. It is suspected to have caused a deadly worldwide pandemic in 1889. Now most people are infected during childhood and build up immunity which protects them later in life. But even today it kills immunocompromised patients.

>These small molecules were successfully used as anti-viral drugs in human trials, defeating rhinovirus in the clinic.

Because it's still in trials, and not widely available, due to the long process of approving such medicines.

It still seems possible that there will be a drug that can be used to combat rhinovirus, as well as coronavirus someday, possibly sooner rather than later.

"Enabled by Oracle high-performance cloud computing, ..."

Anyone know why that phrase is in there? It sounds very much like product marketing, but I can't see why the researchers or this story's author would be motivated to promote Oracle.

> Funding: ... This work received generous support from the Oracle Higher Education and Research program to enable cryo-EM data processing using Oracle’s high-performance public cloud infrastructure (https://cloud.oracle.com/en_US/cloud-infrastructure) ...

https://science.sciencemag.org/content/early/2020/09/18/scie...

Why this summary of the work includes it prominantly, rather than in the footnotes like in the original paper, I'm not sure. Perhaps it was a condition the authors had for contributing to the article. Or perhaps it was Oracle's PR that got them the article in the first place. There is a comment from Oracle later in the article.

Thanks. I didn't think to look in the original sciencemag.org article.

Also possible Oracle donated the resources with the stipulation that they get mentioned in any articles about the research.

"Should I acknowledge Oracle support for my research project work, for example, in publications or on websites? Please acknowledge the “Oracle for Research” program in any publications or on websites."

https://www.oracle.com/oracle-for-research/faq.html

Apparently pushed as a requirement for the "free" Oracle cloud resources researchers can apply for.

I think the medicalxpress article might have been pushed by Oracle PR?

The article itself includes Oracle in the acknowledgements, together with the various University of Bristol HPC systems used, and Archer.

Indeed, 'Medicalexpress' appears to be nothing more than reprints of press-releases about medical research. The PR teams of Bristol & Oracle probably coordinated on the authorship. This is nearly pure promotion, with a thin veneer of reporting-like language.

Does this mean a diet rich in linoleic acid could lessen the symptoms?

My impression is that they've only just begun to study the correlation.

Following my curiosity, I found a list of dietary sources of linoleic acid:

https://en.wikipedia.org/wiki/Linoleic_acid#Dietary_sources

---

Here are the relavant passages from the article:

> Unexpectedly, the research team's analysis revealed the presence of a small molecule, linoleic acid (LA), buried in a tailor-made pocket within the Spike protein.

> LA is a free fatty acid, which is indispensable for many cellular functions. The human body cannot produce LA. Instead, the body absorbs this essential molecule through diet.

> Intriguingly, LA plays a vital role in inflammation and immune modulation, which are both key elements of COVID-19 disease progression. LA is also needed to maintain cell membranes in the lungs so that we can breathe properly.

> From other diseases we know that tinkering with LA metabolic pathways can trigger systemic inflammation, acute respiratory distress syndrome and pneumonia. These pathologies are all observed in patients suffering from severe COVID-19.

> A recent study of COVID-19 patients showed markedly reduced LA levels in their sera.

I'm left wondering, can AWS generate 3-D structures of SARS CoV-2 Spike proteins?

I’m wondering if a long strip of tape, the ability to move left or right, and the ability to write a symbol to the tape could generate 3-D structures of SARS CoV-2 Spike proteins.

There are plenty of covid-19 vaccine trackers, but I can't find trackers for antivirals, especially novel, non-repurposed ones. E.g. what happened to compound 11a from this paper? https://science.sciencemag.org/content/368/6497/1331

The best I could find was an overview in this survey paper: https://www.sciencedirect.com/science/article/pii/S0960894X2...

The best I have a Lancet review for repurposed drugs found via computational methods (published last week). Ignore the "AI;" most of these are just graph algorithms or vanilla neural networks. I haven't seen anything this comprehensive for novel therapeutics.

https://www.thelancet.com/action/showPdf?pii=S2589-7500%2820...

Like drag and drop isn't complicated enough, they have to add a Druggable event?

Oracle's cloud API really is expanding.

Reply on news.ycombinator.com