A Crystal Clear View of the SARS-CoV-2 Protease

This article was reviewed by Dr. Daniel Kneller and Dr. Andrey Kovalevsky. Dr. Kneller is the first author on this paper and Dr. Kovalevsky is the principle investigator. Drs Kneller and Kovalevsky are also two of our subject matter experts.

The paper we’re demystifying today can be found here, if you’d like to follow along.


As the COVID19 pandemic continues across the globe, scientific efforts to develop vaccines and new antiviral drugs are pushing on. 

(The  Moderna, and Pfizer vaccines are now approved for use. You can check out our summaries of the related trials, as well as of the AstraZeneca (Oxford) trial at the links)

As we said in this sidebar about how SARS-CoV-2 infects cells, the virus depends on a protease (a protein ‘scissors’) called 3CL Mpro to produce more infectious virus particles. Since this protease is so critical to viral function, scientists are working to develop drugs that prevent the function of the protease- which would stop the virus’ ability toreproduce. Here’s the catch- developing new drugs for a particular target is hard to do. Structure-assisted drug design is a popular method of developing new drugs, where scientists look at the structure of their target (in this case, 3CL Mpro) and use that to guide their drug design. Scientists have used X-ray crystallography, a very powerful technique for finding out detailed 3D structures of proteins, to solve the structure of 3CL Mpro. However, X-ray crystallography can’t detect hydrogen atoms, which is a limitation of the technique. Why are hydrogen atoms important?

In our sidebar on the central dogma of molecular biology, you learned about how proteins are made up of amino acids. Some amino acids can exist in protonated or non-protonated state. This means that amino acids have certain chemical groups that can be either positively charged, negatively charged or neutral- based on whether or not a hydrogen atom (proton) is bound there. The electric charge of an amino acid or a protein region is very important for drug development because like charges will repel and opposite charges will attract. For example, if there’s a positively charged amino acid in the area you’re trying to target with a drug, you would have trouble properly docking a positively charged drug in that area- it simply will bounce off! On top of that, hydrogen atoms can form hydrogen bonds, a special type of chemical bonding which stabilizes the binding between parts of the protease or between the protease and its substrate (the thing it cuts). Disrupting these hydrogen bonds is another way that drugs can mess with the protease. 

Since X-ray crystallography can’t tell us where hydrogen atoms are located, that makes it difficult to understand the biological function and the drug binding principles of the protease. In today’s study, researchers at Oak Ridge National Laboratory used a highly sensitive technique called neutron crystallography to determine a structure of the 3CL Mpro protease. Neutron crystallography is sensitive enough to determine the locations of hydrogen atoms and can even give us information about hydrogen bonding!

The Technique

X-ray and neutron crystallography are both very, very cool techniques that would probably take an entire course to properly explain. If you want to find out more about the techniques, this video on X-ray crystallography is a great place to start. Here, I will briefly explain the basic principles of neutron crystallography.

In neutron crystallography, the protein of interest is crystallized (very difficult to do) and the crystal is placed in heavy water. Heavy water has the chemical formula D2O and contains the isotope deuterium in place of hydrogen. When the protein crystal is placed in heavy water, exposed hydrogen atoms in the crystal are ‘exchanged’ for deuterium from the water. This whole process is a method used by crystallographers to make their images more ‘sharp.’ These deuterium-exchanged crystals are then bombarded with neutrons. The neutrons are scattered by protein atoms, including deuterium, and the scattering pattern (diffraction pattern) can be used to construct a 3D model of the protein. Often, data from X-ray crystallography experiments can be used to improve (refine) the model.

Using this technique, the researchers obtained a neutron structure of 3CL Mpro at room-temperature and a pH normally present in the body (pH 7.0). This model allowed them to figure out where exchangeable hydrogen atoms were located.

Figure 1: A side-by-side comparison of three models of the 3CL Mpro protease. A shows a cartoon model of the protease. B shows the model obtained from X-ray crystallography experiments, with water molecules (and attached hydrogens) shown in red spheres. C shows the neutron structure, with deuterium atoms in yellow and hydrogen atoms in grey. The neutron structure has so many more spheres because neutron crystallography can visualize all atoms- including hydrogen and deuterium. In contrast, X-ray can only visualize hydrogen atoms in water molecules

Interesting Findings

This new structure allowed the researchers to learn a lot of new, interesting things about the 3CL Mpro protease.  I will list (and briefly explain) a few here:

  1. There are charged residues in the active site 

The active site is the part of the protease where the substrate cutting happens- think of it as the spot between the blades of scissors. The active site is a major target for drug development because that’s where a drug molecule will bind, so knowing whether or not there are charged residues here is very important! The researchers found that there is a negatively charged cysteine (Cys145) and a positively charged histidine (His41) amino acids. Previous structures suggested that this histidine was in the neutral state, so this is a significant finding. This makes the active site zwitterionic, which means that it has an overall 0 charge (+1 -1 = 0), but there are charged components within it.

2. Extensive hydrogen bonding in six substrate-binding subsites

The substrate-binding subsites are portions of the active site site that bind specific amino acids of the thing that the protease is cutting (the substrate). The neutron crystallographic structure showed the researchers that there are several important hydrogen bonds in and between the substrate-binding subsites. Disrupting or competing for those hydrogen bonds with a drug may be a way to impact 3CL Mpro function, a significant finding for drug design.

3. Hydrogen bonds at the dimer interface

3CL Mpro is a dimer- which means the actual, working protease consists of two identical halves (like a scissors needs two blades). The dimer interface is the area where these two halves come together. The researchers found that there are numerous hydrogen bonds at this interface, with most of the hydrogen bonds made by a protein stretch from Ser1 to Cys16 of each half with other parts of the protein. Based on this, the researchers have suggested that drugs should target amino acids 1-16 of the protein halves. This is supported both by their neutron crystallography model and previous research showing that the protein does not function without forming a dimer.

Conclusions and Limitations

This structural study aimed to develop the most complete model of 3CL Mpro that showed the locations of important hydrogen atoms. The researchers succeeded in this aim and gathered some critical insights into the protease structure and function. While this structure contains a lot of information about the protease, the main limitation of structural studies is that structural predictions don’t always carry into biological function. This structure is a great starting point for future drug development experiments, but more research is needed before we can create a drug to treat people COVID-19.

This is a big step forward by researchers at Oak Ridge National Laboratory. Thanks to their efforts, we have a better understanding of the structure of 3CL Mpro. This understanding will now drive future drug development efforts as we continue to fight back against this pandemic. It will greatly advance computer simulations because the model now is complete and correct. Just another way that science is fighting for you on all fronts- from structural to biochemical to clinical. Stay safe and keep helping. 

Leave a Reply