Our work changed the picture on bacterial TPD, when studying how bacteria cope with adverse conditions. Our studies revealed a simplified version of the eukaryotic ubiquitin-proteasome system (UPS), in which pArg serves as ubiquitin like degradation tag but also regulates critical events of the bacterial stress response. We identified a unique mechanism of auto-stimulation of McsB, the pArg protein kinase. Further we showed that cellular concentration of pArg proteins regulates the oligomeric state of McsB which defines substrate specificity. Further, pArg labelled proteins act as inductor for the entire stress operon. Finally, binding of pArg proteins cause conversion the ClpC unfoldase from a resting to an active super-assembly, thereby revealing an unprecedented mechanism of how an AAA unfoldase is activated. In this super-assembly ClpC was caught in the act while unfolding proteins. To gain insight in general mechanisms on these HSP100 AAA unfoldases to which ClpC belongs, we extended our studies to eukaryotic Hsp104 proteins. We identified a conserved central domain as important regulatory device constraining the dynamicity and unfoldase activity. This effect is unleashed in the active ClpC super-assembly in which this domain connects ClpC hexamers. The outside of the four rings is poised to interact with the protease ClpP and the substrate receptor domain is located inside. The latter usually binds substrates, but we found that natural ClpC targeting antibiotics (e.g CymA) interfere with substrate binding by mimicking damaged proteins and hijacking this binding site. Most importantly, we found that proteotoxic stress induced by those antibiotics is counteracted in mycobacteria by upregulating two small Clp proteins that function as molecular scavengers in the mycobacterial PQC system.
All these findings culminated in the development of bifunctional molecules that allowed to purposely direct proteins to the ClpCP system. Similar molecules haven been developed for eukaryotic UPS, reprogramming the pathway to target neo-substrates. Proteolysis targeting chimeras (PROTACs) induce ubiquitination and degradation of selected proteins of interest (POI). These degraders act as bi-functional units redirecting an E3 ligase of the UPS against a POI. PROTACs exhibit mechanistic properties, providing them striking advantages over classic, occupancy-driven drugs, which reshaped many research programs in pharmaceutical industry and academia.
An equivalent approach in bacteria was unreported and the simple composition of an pArg tag, intrigued us to establish bacterial PROTACs (BacPROTACs). Forcing McsB to label neo-POIs thereby inducing degradation by ClpCP in vitro showed the feasibility of that approach. We linked the pArg tag to chemical moieties binding to POIs and such BacPROTACs induced degradation by ClpCP protease. When using a cell-permeable ClpC binder we succeeded in developing BacPROTACs that induced degradation in mycobacteria. Using this technology, we established a conditional knockdown technology that allowed removal of POIs upon genetic fusion. Ultimately, we developed potent BacPROTAC antibiotics that were able to kill Mycobacterium tuberculosis, residing in macrophages.
In the final part of our project, we moved back to the eukaryotic UPS, looking for functional homologs of double-ring unfoldases. We were intrigued by the giant ubiquitin ligase RNF213 that exhibits two AAA modules with similarity to ClpC. Reconstitution of this 600 kDa protein showed that, in contrast to HSP100 proteins, RNF213 contains six AAA modules within a single polypeptide chain. However, the functional ATPases of RNF213 do not generate mechanical force, but function as molecular switch coupling E3 activity to ATP binding. As such, RNF213 is unique as it senses ATP/AMP levels reacting to the energy state of the cell. RNF213 presents a novel type of ubiquitin ligases as it uses a Cys/His catalytic dyad, situated on a specialized zinc-finger domain, to promote the ubiquitin transfer from its E2 partner. Overall, our data highlight the (still) underestimated complexity of the ubiquitin and TPD network.
