Just this week, we had the first promising report of a drug that appears to improve the recovery time of patients suffering from COVID-19. Hot on the heels of that announcement, a scientific journal has released a paper that describes how the drug interferes with the virus. While there's no real surprises in what has been revealed, it provides key details of how SARS-CoV-2 can be blocked.
Copying machine
Targeting a virus with a drug is a challenge. Viruses make their living by using their host's proteins to do most of the work of making new viruses. That means a drug has to target some of the few proteins encoded by the virus while not interfering with any of the far more prevalent host cell proteins. In the case of the coronavirus, biologists have identified a number of distinct features of the virus that may be targeted without an obvious risk of causing severe side effects.
Remdesivir, which saw a large clinical trial produce promising results, is a drug that's designed to target one of these virus-specific vulnerabilities. The coronavirus genome is encoded using the chemical RNA, as opposed to the DNA used for our genome. In fact, there's nothing about our cells that requires them to make an RNA copy of an RNA molecule. As a result, the coronavirus genome encodes proteins that do this RNA-to-RNA copying, called an RNA-dependent RNA polymerase. Remdesivir was designed to look like one of the building blocks of RNA in the hope that it would bind to an RNA virus' polymerase and inhibit it.
That said, this drug was designed with the intention of inhibiting the polymerase of a different virus (Ebola), so it wasn't guaranteed to work against coronavirus. And our cells need to make RNA copies of DNA, a process that's similar enough that remdesivir could interfere with that, too.
Still, tests in cells had been promising enough to drive testing in humans. While that testing was starting, a group of Chinese scientists decided to look into how remdesivir actually works. To do so, they decided to figure out how the drug interacted with the coronavirus RNA polymerase at the atomic level. And that requires a technique to determine where all the atoms in the protein and drug are.
A few decades ago, figuring out atomic-level details of proteins would have required many months of laboriously trying to get the drug and protein to form neat, orderly crystals. But we've since developed a combination of hardware and algorithms that now allow us to take what are essentially electron microscope images of individual proteins and combine them with enough precision to figure out where all the atoms are. The technique, called cryo electron microscopy, has been so revolutionary that it earned the people who developed it a Nobel Prize.
These scientists also benefited from earlier work on other coronaviruses, which had identified three different proteins that were critical for copying the virus' genome. One of these is the enzyme that actually strings together individual units called "bases" to form a new RNA molecule. The other proteins in question simply help it clamp down on and move along the RNA it makes a copy of. So the researchers produced these three proteins, put in an RNA template and a partial copy, and then added remdesivir.
In the atoms
Well, they didn't technically add remdesivir. Bases are added to RNA in a form with three phosphates attached. That form is very negatively charged, and it won't make it across enzymes very easily. Instead, the drug is provided in a form with minimal charge, which can transit across cell membranes. Once inside the cell, the cell's own enzymes convert it into the charged form, which then gets used by the RNA-copying enzyme. Since they weren't working in cells, the researchers had to do this conversion themselves. It's a bit of an aside, but it illustrates some of the challenges that the people doing drug development face.
In any case, the imaging technique requires taking thousands of electron micrographs of individual protein-RNA-drug complexes, all of them oriented randomly. While none of these is, on its own, enough to figure out where the atoms are, computer algorithms are able to combine all of these images and figure out what the underlying structure that's compatible with all of these different images.
In the end, after over 80,000 images were combined, the resolution of the structure they determined had a precision of about 2.5 angstroms (10-10 meters). For comparison, a carbon atom in these same molecules has a width of about 1.5 angstroms. But given the fact that many configurations consistent with this resolution make no sense—they place two carbon atoms on top of each other or something similar—the pictRead More – Source
[contf] [contfnew] 
arstechnica
[contfnewc] [contfnewc]
