I guess nobody has missed the fact that we are in the middle of a new Coronavirus outbreak. In the beginning, I followed the updates from far, without going in too much detail. That was before I saw a LinkedIn post on a homology model of a 2019-nCoV protease last week.

Being highly influenced (read “damaged”) by my job, my mind immediately started popping out questions. Questions like… What does that protein actually do for the virus? Can we target that protein with a drug? How similar is the binding site to other druggable sites? What other 2019-nCoV proteins could we target? Are the 3D structures resolved yet? I want to see protein structures!

In short, my curiosity got triggered and I started investigating. Since I’ve never worked on viruses, I learned a lot of new cool stuff. I, therefore, thought I’d share some of my findings here. I’m also pretty sure I have at least a couple of people in my network who know as little as I did (and a couple of protein fans too).

I added a bonus at the very bottom of the article, so be sure to read to the very end.

What proteins are expressed by this virus?

The complete genome of the Wuhan seafood market pneumonia virus, aka 2019-nCoV, has been sequenced and published in GenBank, so this was my first stop. 2019-nCOV’s genome consists of one unique RNA molecule. To turn these genes into proteins, the virus uses the host cell’s machinery. 

More precisely, when 2019-nCOV arrives at a new host cell, one of its surface proteins, Spike glycoprotein (S), binds to a receptor protein on the host cell's surface - the Angiotensin-converting enzyme 2 (ACE2). Another protein on the host cell's surface, a protease, then comes along and cuts off a piece of the S protein. This releases the Spike fusion that helps in the virus entry. Edit: It's next unclear how exactly the 2019-nCoV enters the cell. It could either fuse with the cell membrane or enter the host cell as a "membrane bubble" (endocytosis).

The 2019-nCoV is a single positive-stranded RNA virus (Edit: not a retro-virus), which means it can serve directly as a messenger RNA (mRNA). The virus can, therefore, directly hijack the host cell's Ribosome to synthesize its proteins. One of the first proteins to be synthesized is RNA polymerase, whose job is to make a negative-stranded "copy" of the viral RNA. This complementary RNA strand has two purposes - allow the cell to make numerous copies of its original RNA, and transcribe smaller and more specific mRNA strands that can be fed into the protein synthesis machinery.

So, what proteins are encoded from the 2019-nCOV genome? To start with, we have a long polyprotein chain (orf1ab) that is cut into 16 different non-structural proteins (nsp1-16). This is again a smart virus trick. One single ride through the gene expression machinery results in 16 proteins. That’s pretty efficient! Among these 16 proteins, you have proteinase, cofactors, the RNA polymerase and proteins with unknown function.

The genome also encodes for five other non-structural proteins - ORF3a, ORF6, ORF7a, ORF8, ORF10 (I won’t go into the detail here) and four structural proteins - S, M, N and E. As their names state, the structural proteins build up and shape the virus membrane envelope, together with lipids stolen from the host cell (see image below).

So there you have it. A bunch of proteins is expressed. But which of these proteins are relevant when it comes to killing this virus? Let’s look at the current treatment strategies.

How do you kill a virus?

Several strategies are currently pursued to stop this viral outbreak: vaccines, antibody treatment and repurposing of small molecule inhibitors.

Vaccines

To develop a vaccine, you need a part of the virus that the human immune system can easily detect, “kill” and remember in order to create an immunity. 

For 2019-nCoV, we already have the sequence of these proteins and a lot of knowledge from previous Coronavirus outbreaks (SARS-CoV and MERS-CoV). Several pharmaceutical companies and research groups are therefore optimistic in bringing a vaccine to patients this year. Different strategies are used for triggering the wanted immune response. 

Moderna is developing a messenger RNA (mRNA) vaccine derived from the viral RNA. 
Inovio Pharmaceuticals is developing a DNA plasmid vaccine that expresses the structural S protein – (spike glycoprotein). 
Novavax is working on a nanoparticle covered by the 2019-nCoV-specific surface protein (protein S?). 

I didn’t find any details on J&J’s strategy, but if they are using the same as for their Ebola vaccine, they are using a genetically modified 2019-nCoV which is unable to replicate in human cells.
Even though some of these vaccine developments could go very fast (under a year), it might still be too long to have an impact on this outbreak. As a reminder, the entire SARS-CoV outbreak lasted “only” 6 months. 
Also, vaccines are preventive, that is, they can protect you from future viral infections, but they are to no use for the person that is already infected. 

Antibodies

One way of stopping the infection is to treat the patient with antibodies that bind to the virus and hinder it from docking to the host cell's receptor protein. If it can’t bind to the host cell, it can’t reproduce and spread. 

Neutralizing monoclonal antibodies for MERS-CoV were already under development when the 2019-nCoV outbreak started but have not yet been tested clinically. Antibodies are highly specific, though, which means that if these antibodies bind to MERS-CoV, they probably won’t work as well on 2019-CoV. 

Vir Biotechnology has announced that they are working on 2019-nCoV-specific neutralizing antibodies.

Small molecules

In order to treat people that are already infected, hospitals have turned to drugs that are already on the market. If they are already on the market it means that these drugs were developed for other indications and that their efficacity on 2019-CoV is uncertain to say the least. But they are approved and available so why not try them, right? 

Chinese researchers have screened 30 different compounds (a mix of anti-HIV drugs, an immunosuppressant, an anti-schizophrenia drug, and also natural products and Chinese traditional medicines), for their efficacity on 2019-nCoV. According to the Chinese Academy of Sciences (CAS), three of these molecules have fairly good inhibitory effects:

• Remdesivir – initially designed to combat the Ebola virus, developed by Gilead Sciences. Works as an RNA polymerase inhibitor.

• Chloroquine – an antimalaria drug initially discovered by Bayer, later repurposed as an anti-HIV in clinical trials. The anti-viral mechanism of action is still unclear to me. 

• Ritonavir – an anti-HIV drug developed by Abbott (now Abbvie). Works as a protease inhibitor.

The Chinese researchers are now trying to get approval for a clinical trial to validate their efficacy on 2019-nCoV in patients. 

Show me the structures! 

According to CAS, researchers at ShanghaiTech University have resolved a high-resolution crystal structure of 2019-nCoV’s main proteinase (Mpro). Unfortunately, I haven’t found any trace of the structure coordinates. Has anybody seen them somewhere? Edit: it was released by the RCSB PDB on February 5, 2020. The structure is available here.

On the modeling side, Gruber’s group at Innophore has built homology models of the 2019-nCoV main protease (putative target of Ritonavir) by using SARS homolog structures as templates. They’ve also docked the anti-HIV drug Lopinavir into the active site and run MD simulation of a set of poses.

The Zhang Lab has published homology models of 24 different 2019-nCov proteins, namely the main proteinase, the RNA-directed RNA polymerase (putative target of Remedesivir) and the spike (S) glycoprotein.

On our side, in 3decision, we’ve registered the 2019-nCoV polyprotein sequence and the 7 different 2019-nCov protease homology models (6 from Innophore and 1 from the Zhang Lab). I’m currently setting up a project in order to compare the binding sites in the different models. We'll also use 3decison to compare the 2019-nCOV protease to the HIV-protease (for which Ritonavir was originally developed). 

Bonus

We want to contribute to the world-wide effort to finding a treatment for this viral outbreak. We, therefore, offer free access to our cloud-based platform 3decision®.

3decision® is a web-based and collaborative platform for structural analytics. It has a fully annotated and searchable database integrated (the entire PDB + modeling results + pocketome). The analytics and visualization tools are designed to facilitate tasks like druggability assessment, selectivity analysis, and 3D design idea generation. The collaborative setup (common projects and share user-session), allow researchers all over the world to work together and share ideas with each other.

For everybody who is interested and wants to work on this project, just send us an email on 2019-nCoV@discngine.com and we’ll activate an account for you. You are also welcome to upload additional homology models, docking results or mutagenesis data.

3decision’s blog has been moved!