New bifunctional molecules as tools for targeted protein degradation in bacteria

Targeted protein degradation (TPD) was developed as an innovative approach to regulating protein levels in cells for research and therapeutic applications. The TPD enables to deplete the levels of endogenous proteins by selectively directing them for the degradation by endogenous proteases. This degradation is mediated by specifically designed molecules such as PROTACs (Proteolysis-Targeting Chimeras) which are composed of two ligand moieties joined by a linker. One of the PROTACs ligands binds to the target protein and the other ligand recruits the E3 ubiquitin ligase to induce the ubiquitination of the target protein which is then recognised by the 26S proteasome as a degradation signal. The TPD methods enable to irreversibly eliminate protein molecules from the cells which can be applied in research for validating potential drug targets. The degradation of the disease-related proteins can be also developed as a new therapeutic strategy which is supported by the approval of numerous PROTAC molecules for clinical trials. The TPD methods are well established in eukaryotes, however, they are not currently applied in bacteria. One of the primary limitations of applying TPD in bacteria is the lack of eukaryotic-like protein degradation pathways. Adapting the method for bacterial cells would require catering the TPD for different proteolytic machineries. Nonetheless developing similar methods for bacteria could become a promising strategy for creating new antimicrobials. This is particularly relevant in the times of a rapidly developing antimicrobial crisis which is recognised as one of the most relevant challenges in global healthcare. The dissertation presents the research process of establishing the fundamentals for developing bacterial TPD. The experiments presented in the thesis provide a proof-of-concept for the ClpXP-mediated inducible protein degradation in Escherichia coli. Establishing the new method for the bacterial TPD required careful design and optimisation of the new degrader molecules. The degrader molecules were designed to tether the target proteins directly to the bacterial ClpXP protease and activate the proteolytic complex to induce the target proteins' degradation. The molecules were designed based on the known peptide ligands of the target proteins (“baits”) and peptide ligands of the ClpXP-SspB protease complex (“anchors”). The “baits” and “anchors” were joined with a peptide linker modality. To choose the optimal “anchor” which could interact with the ClpXP protease and activate it, I tested the potential “anchor” peptides using the eGFP reporter system. Monitoring the stability of the eGFP fusions with different peptides in E. coli allowed me to select the XB peptide as the promising “anchor” candidate and confirm its dependence on the ClpXP protease. The sequences encoding the putative degraders, composed of the XB “anchor”, linker, and target-specific “bait” peptides, were cloned to the arabinose-inducible expression plasmids. The system was designed to enable studying the degradation effects in cells without encountering the cell permeability issue. The targets of degraders were carefully selected to enable monitoring of the degradation by easy phenotype readouts: cell viability (GroEL, DnaK, replisome proteins), cell shape (divisome proteins), or fluorescence (GFP-tagged proteins). The degrader against the GFP-tagged proteins was created by fusing the XB “anchor” with the GFP nanobody. Expression of this construct did not allow me to observe the expected growth inhibition of the bacteria expressing the RNase E-GFP fusion protein caused by RNase E depletion. Although the MST measurement detected that the degrader bound to the eGFP protein with an affinity comparable with the affinity of the “bait” alone, the effect of the target degradation could not be observed even in the microscopic fluorescence measurements. The degrader against cell division proteins (divisome) exploited the C-terminal peptide of the FtsZ protein. Since the divisome regulates the division of cells the expected phenotype of the degraders was the alteration of the cell length. The microscopic measurement has shown that the degrader-expressing cells were on average 20% shorter than the cells expressing the control peptides. Although the differences were reproducible, detailed proteomic studies are necessary to unveil the exact mechanism of action of these degraders. The degraders against the replisome proteins were composed of the XB in fusion with different linkers and the “bait” peptide derived from the C-terminus of the SSB protein. These degraders were designed to bind and degrade the DNA-binding proteins which are cellular partners of the SSB protein. Although the expression of such peptides was inducing the changes in bacterial growth in a linker-dependent manner, the proteomic changes in the level of replication proteins could not be detected despite multiple optimisation attempts. The chaperones DnaK and GroEL were chosen as promising degraders targets since they are involved in the essential protein folding process. The fusion of the XB “anchor” with DnaK binding “baits” did not exhibit the suppressive effect of such fusions on bacterial growth. This could probably be caused by the high toxicity of the chosen “baits”. The degraders against GroEL (“GroTACs”) composed of the XB “anchor”, a flexible linker, and the SBP “bait” caused the temperature-dependent growth inhibition of E. coli (up to 50% inhibition). The inhibition of growth was complemented by enzymatic viability tests. The phenotype tests of the degraders were followed by proteomic studies by immunoblotting and the comparative TMT-MS proteomics. The experiments confirmed a reduced level of the GroEL and multiple essential GroEL substrates in the GroTAC-expressing cells. The results obtained in vivo were complemented and validated by in vitro studies on the binding between peptides and their targets. The biophysical characterisation of the kinetics of the peptide-protein interactions by biolayer interferometry (BLI) showed that the GroTACs can bind both to the GroEL target and the ClpX component of the ClpXP protease with the parameters comparable to the individual protein ligands. The obtained 3.3 Å electrostatic potential density map of the GroEL with the GroTAC peptide confirmed the binding and allowed me to observe the binding orientation which explains the limited effect of the GroTACs on the GroEL level. The final experimental part of the thesis presents the microbial susceptibility tests on the small molecule degraders of bacterial DHFR protein. The custom-synthesised PROTAC-like molecules were tested in model E. coli and Neisseria gonorrhoeae strains. The chimeric molecules did not result in the observable bactericidal effect on tested strains. This highlights the limitations in the permeability of many antibacterial compounds. The results obtained in the presented experiments provide general guidelines for the future design and development of TPD-inducing compounds for bacteria. The final part of the thesis discusses the limitations of the presented degraders and concludes the general rules for optimisation of “anchors” and “baits” as well as the targets of the degraders. Finally, the results provide the first evidence of the ClpXP-mediated TPD in bacteria and validate it as a potential protease for the future development of degradation-inducing compounds.