What happens when proteins can't find their partners?
Cells have a lot of garbage disposal issues. There are lysosomes to digest large things like viruses, proteasomes to dispose of individual proteins, and lots of surveillance mechanisms to check that things are going as they should- that proteins coming off the ribosome are complete, that mRNAs are being spliced, that mitochondria are charged up as they should be, that the endoplasmic reticulum is making, folding, and secreting proteins as it should be, among many others. One basic problem that arises when cells have a lot of proteins that assemble and cooperate in the form of complexes, is that some of those subunits may be present in excess, or not join their intended complexes for other reasons such as misfolding. This can have very bad effects. Most protein binding makes use of hydrophobic surfaces, and having these floating around freely can lead to indiscriminate binding / agglomeration, like amyloid plaque formation, and cell death.
Bacteria have one partial solution, which is to encode proteins that are destined to the same complex from the same mRNA, made from what is called an "operon" of genes, like a train with successive gene-carriages. Each multi-protein-encoding message from such an operon is thus necessarily equally abundant, and, assuming simiar ribosomal rates of protein synthesis, the proteins should also be produced in equal quantities, providing at least one method to balance their abundance in the cell. But there are many other issues- proteins may have different life-spans, or different ribosomal production rates, or assembly into the complex may be slow and difficult, so bacteria still are not out of the woods. Eukaryotes do not use operons anyhow, so our more-finely regulated gene control mechanisms are called on to properly equalize (or adjust for) the ultimate subunit concentrations.
But when all this fails, and there is more of some complex subunit than needed, what happens then? When experimenters over-produce some complex component in cells, it is typically short-lived. And if they impair its production, the rest of the complex tends to be short-lived. This implies mechanisms in the cell to dispose of incomplete complexes and their components. It turns out that there are some specific chaperone proteins that detect such orphan subunits, and tag them to be destroyed. Several prominent complexes, such as ribosomes and proteasomes, even have specifically dedicated mop-up chaperones. A recent paper described a chaperone protein dedicated to mopping up the excess or misfolded subunits of another large and abundant complex - the chaperonin complex. That makes this protein, ZNRD2, a sort of metachaperone.
 |
| Some structural (though not dynamic) views of the CCT complex. A shows top and side views, respectively. C shows a layout of how the equator of the complex looks, as coded by each of the subunits. At the ring-ring interfaces are the ATP binding sites (d). And lastly (e) a cut-away view of the inside show where substrate proteins are enclosed and encouraged to fold correctly. |
The chaperonin complex, (also called CCT), is a large, hollow sphere that actively helps other proteins to fold correctly. The structural proteins actin and tubulin are the most prominent targets that need this help. When first synthesized, they are bound by adapters that ferry them to the chaperonin complex, which lifts its lid to allow the protein in. Then, ATP is used to induce dramatic cycling of the chaperonin structure, shifting from an internal hydrophobic structure to a more hydrophilic one. This allows the unfolded protein to alternately splay open over the hydrophobic surface, and then fold in piece-wise fashion, for as long as it takes till the barrel detects that it is fully folded and no longer sticking to the hydrophobic internal surfaces.
In the current work, the researchers drove the expression of several individual CCT subunits in cell lysates. Then they sent the products into a mass spectrometer to find out what was sticking to these "orphan" proteins. They found two major associated proteins, HERC2, and ZNRD2. HERC2 is known as a ubiquitin ligase, which is one of a large family of enzymes that tag proteins with ubiquitin, targeting them for disposal. But ZNRD2 was totally uncharacterized, known only as an auto-antigen reacted to by some people with Sjogren's syndrome or scleroderma. The question then was .. does HERC2 directly sense the presence of free-floating CCT subunits, or does it need a helper to do so, such as perhaps ZNRD2?
"... a sizable population of multiple CCT subunits are orphaned even under normal conditions, and the degradation of a subset of these can be stimulated by HERC2."
The researchers showed that deleting HERC2 strongly impaired the cleanup of most orphan CCT subunits. It is evident, however, that there are other chaperones not covered in this work that help clean up some of the other CCT subunits. Then they found that HERC2 interaction with the CCT proteins was dependent on ZNRD2, but that the reverse was not the case- ZNRD2 binds CCT subunits in any case. This, and other experiments, including mapping the location within the HERC2 protein that binds ZNRD2, showed that ZNRD2 is the adapter that does the detailed detection of orphaned CCT subunits. At only 199 amino acids, there is not much to it, and searches for domain signatures do not yield much. Its name reflects a structure that uses zinc ions for stabilization, but much of the protein is also disordered. It is notable for a high proportion of hydrophobic amino acids (alanine, leucine) and lots of prolines (15), which would contribute to a disordered structure.
Thankfully, with the advent of AI and alpha-fold, these researchers could also investigate and model how ZNRD2 interacts with both the HERC2 ubiquitin ligase and with one of the CCT subunits, CCT4- all without doing any laborious structure determinations.
 |
| AI-calculated structures of the complex of the ubiquitin ligase HERC2 with the adaptor ZNRD2 and the target subunit CCT4. At right, the hydrophobic residues of CCT4 are colored yellow, showing that the ZNRD2 orphan subunit detector and adaptor binds to a hydrophobic pocket which would otherwise be completely buried with the full CCT structure. The interacting domain of HERC4 in green is termed a 7-bladed beta propeller. |
"In the fully assembled CCT double ring, all potential ZNRD2 interaction sites are completely buried because they form the interface between the two individual rings."
They found that ZNRD2 binds to a hydrophobic pocket of CCT4, a pocket that is otherwise buried in the fully assembled CCT. This patch would also be exposed on partially assembled CCT complexes, indicating that this interaction is not only relevant for mopping up the individual subunit, but for several kinds of incomplete assembly of the entire complex, perhaps explaining why other subunits are also mopped up by this system.
This kind of work is a good example of normal science. A gene about which nothing was previously known (ZNDR2) is now given a function in the cell, and a process circumstantially known to exist is fleshed out with actors and structures that explain it. Of the ~20,000 human protein-coding genes, roughly ten percent still have no annotation, and many more have only tenuous annotation, perhaps only drawn from structural analogy, not direct study. So there is a great deal more work needed to evaluate our parts list, even on the most basic level, even before getting into the complexities of how these proteins act and interact in tissues and pathways.
