Molecular Structure of human NLRP3 solved

February 04, 2022

left: Prof. Matthias Geyer; right: Inga Hochheiser (picture: Johann Saba / UKB)


Elucidating the structure of a central inflammatory switch provides the basis for the development of powerful anti-inflammatory therapeutic tools


ImmunoSensation2 Member Prof. Matthias Geyer and his team, in collaboration with researchers at the University of Regensburg, have solved the structure of a central cellular inflammatory switch: NLRP3. Cryo Electron Microscopy (Cryo-EM) analysis revealed NLRP3 to form a decameric structure when incubated with the inhibitor CRID3 (Cytokine Release Inhibitory Drug 3). Identifying the CRID3 binding-site as well as solving the overall structure of human NLRP3 now provides the basis to further elucidate the molecular mechanisms leading to activation of NLRP3, as well es the development of potent anti-inflammatory pharmaceuticals. The results are published in the journal Nature.



We met Prof. Matthias Geyer and spoke about his finding.


Mr. Geyer, you published a new protein structure in the journal Nature today. What kind of protein are we talking about?

We have been working on a human protein from the immune system called NLRP3. The protein is able to recognize inflammatory signals, such as a bacterial infection or gout-causing uric acid crystals, and then trigger an immune response. When the NLRP3 protein is activated, this results in cytokines being released. These messengers of the inflammatory response activate surrounding cells, leading to a targeted immune response. In chronic inflammation, the threshold for activation of NLRP3 is lowered, so that the cells are permanently immune-reactive. However, the molecular basis of this mechanism is poorly understood to date. In our study, we investigated the inactive form of NLRP3: the state before the inflammatory response takes place and which one would like to strengthen to counteract chronic inflammation.


And what does this inactive form of NLRP3 look like?

In its inactive form, the NLRP3 protein forms a giant decamer. Ten subunits lay together, with two proteins interlocked at any given time, and then five of these pairs form a ring-like structure. In the middle, the ensemble is held together by two small domains that are completely absorbed into this sphere. When these so-called pyrin domains are presented on the outside of the molecule, this leads to the activation of the immune system, which eventually induces the release of cytokines. The inactive NLRP3 structure seems to form a perfect cage that controls the accessibility of the pyrin domain.


But your structure not only shows the NLRP3 decamer, but also contains the inhibitor CRID3. Was the inhibitor somehow key to your success?

Yes, that was the decisive breakthrough to success and is an exciting story. The pharmaceutical company Pfizer was already looking for a follow-up drug for a diabetes type II drug in the 1990s. In the process, they found a compound that was very effective in inhibiting the release of cytokines. In 2003, this compound was published under the name 'cytokine-release inhibitory drug' or CRID3, without knowing how it worked molecularly. The NLRP3 inflammasome had also not been discovered at that time. Then in 2015, under the leadership of an Irish research group, cell-biological experiments showed that the addition of CRID3 inhibited the activatability of NLRP3. We then found in protein-biochemistry experiments that the compound CRID3 significantly stabilizes the protein NLRP3. CRID3 can bind to NLRP3 only when the protein is in the ADP-bound inactive state. We found that to achieve this effect, it is best to add the drug while the protein is being expressed. The greatest effect was evident in the electron micrographs: The protein, which had always appeared blurry before, suddenly became clearly visible and we achieved much higher resolution, which allowed us to see structural details.


The technique of electron microscopy you used in order to elucidate the protein structure is still relatively new. How does it work and what challenges did you and your team face?

Electron microscopy (EM) has become a broad field in the life sciences, comparable to light microscopy in terms of the range of possible applications. We use the technique of "single-particle cryo-EM" in structural biology. The first challenge is to produce the protein in high purity and uniformity, and then to place it stably on the support material. This involves embedding individual molecules in an ice film. The molecules are then transilluminated with an electron beam and the image is captured with a camera. By combining a large number of views from different angles, it is possible to determine the shape of a molecule's three-dimensional structure. The next challenge is to place the amino acid chain into this shape in a coherent solution. Since molecular resolution is often not as high as in X-ray crystallography, especially for large protein complexes, this can be a time-consuming job where you have to be careful not to make mistakes. We performed our first measurements at the caesar research center in Bonn, then in collaboration with a research group in Regensburg, and the high-resolution data set was finally acquired at EMBL in Heidelberg. Since we don't have our own electron microscope and a lot of work has been done on the spike protein of SARS-CoV-2 in the last two years, we had to wait a long for measurement time. Finally, we were lucky and able to acquire a very good data set. But with access to our own electron microscope, we would have been much faster and could now do more advanced measurements.


The whole world is talking about AlphaFold, the AI software from Alphabet/Google for protein structure prediction. Did that help your team in their work?

No, when we solved the structure, AlphaFold didn't exist yet. But we put our work on a pre-print server the exact day the AlphaFold database was made publicly available. Indeed, AlphaFold's prediction agrees very well with our experimental data. But crucial sites, such as a long loop region that contributes to the formation of the dimer, or the assembly of the multimer, a conformation is not predicted by AlphaFold. For example, a Harvard research group that worked on the structure of the mouse NLRP3 protein at the same time as we did, could not achieve the resolution to see the loop region in the structure. The AlphaFold model was then incorporated into the structure. However, this created a node in the protein chain that never would occur in nature. From this we can see that AlphaFold can be a useful tool, but you have to know how to use it.

You said you were able to achieve the higher resolution by adding the inhibitor CRID3. How exactly does it act on NLRP3?

The CRID3 inhibitor binds in a pocket deep inside the protein and holds five different areas together there. As a result, the protein loses its dynamics and becomes frozen in this state, so to speak. Our current hypothesis is that this pocket exists only in the inactive state of the protein. When NLRP3 is activated, this cavity shears open and the protein adopts a different structure. Conversely, this also means that the protein cannot open when the drug is bound in the protein. NLRP3 is therefore held in the inactive state and the inflammatory response is inhibited. Taking a look back, what were the mayor milestones in pursuing this project and how do you assess your finding regarding future research on the inflammasomes? The project started for us in January 2017 when we first saw native NLRP3 take on a large, uniform shape. We then spent a very long time optimizing the protein preparation conditions for cryo-EM structure determination. When we saw that the inhibitor CRID3 improved the images of the structure so significantly, we knew we were on the right track. To resolve the structure of a protein bound to an agent is a great result. The fact that we have now even identified a new binding pocket that can be used to target inflammatory responses is simply fantastic. We assume that all 22 proteins in this family have a similar binding pocket, because all proteins have the same domain architecture and are activated in the same way by an ADP-to-ATP exchange. So that gives us a huge arsenal of possibilities to counter acute inflammation like sepsis or chronic inflammation like Crohn's disease. I think that will keep us busy for many years to come.

(Interview with Prof. Matthias Geyer from the Institute of Structural Biology at the University of Bonn, 03.02.2022, by Dr. David Fußhöller)



Participating institutions and funding:

The study was funded by the German Research Foundation (DFG) and by EU funds under the iNEXT-Discovery and Instruct-ERIC initiatives. The cryo-EM images for structure elucidation were recorded at EMBL in Heidelberg.



Inga V. Hochheiser et. al. (2022), Structure of the NLRP3 decamer bound to the cytokine release inhibitor CRID3; Nature; DOI: 10.1038/s41586-022-04467-w



Prof. Dr. Matthias Geyer Institute of Structural Biology at the University of Bonn Phone: +49-228/287-51400 E-mail: