Team:Freiburg/Demonstrate
Our project focused on empowering the use of D-amino acids in synthetic biology and the identification of D-peptide therapeutics. To achieve this, we applied several different approaches. We successfully demonstrated that we reached the goal of our project in multiple ways.
We identified peptide ligands against PSMɑ3 via mirror-image phage display
D-peptides, made from D-amino acids, are resistant to protease degradation, as we demonstrated by a proteinase K digestion assay. Because of this, D-peptides have an increased half-life compared to their L-enantiomers, making them excellent candidates for therapeutics. We have set ourselves the goal to identify D-peptide ligands against the toxin of the Methicillin-resistant Staphylococcus aureus (MRSA), specifically against its exotoxin phenol soluble modulin alpha 3 (PSMα3). To reach this goal, we successfully established and performed a mirror-image phage display (MIPD).
To set up MIPD, we first had to chemically synthesize our target PSMɑ3 in its mirror-image D-version. Applying solid phase peptide synthesis (SPPS), we were able to successfully synthesize and formylate PSMα3 in L- and D-form in our lab. We verified the activity of our chemically synthesized PSMα3 with cellular assays. Notably, we observed comparable cytotoxicity on human Jurkat T-cells as with commercially available L- and D-PSMα3 pointing to a similar quality of synthesis.
By successfully performing MIPD against PSMα3, we could identify three peptide ligands that specifically bind to the D-form of PSMɑ3. Moreover, we could demonstrate the activity of these inhibitors towards PSMα3 by performing toxicity assays on Jurkat T cells The toxicity was reduced up to 35%.
We established a faithful computational model of PSMα3.
In order to model the D-version of this peptide toxin, we created an L-to-D converter which enables the conversion of any structure that can be found in the Protein Data Bank (PDB) or other sources, to its mirror image form. We proved the accuracy of the L-to-D converter using the protein monellin for verification, for which a crystal structure is available for both L- and D-form.
We modeled in silico the association of the L- and the D-version of PSMα3 by molecular docking. Successfully transferring our in silico predictions to the wet lab, we also demonstrated this by quantitative binding assays and cytotoxicity assays on Jurkat T-cells, thereby validating our modeling strategy.
We developed finDr, a software for the in silico identification of D-peptide ligands against a target of choice
To speed up the discovery of promising D-peptide binders against a protein of interest, we developed finDr, a software based on our modelling approach. We demonstrated its power by identifying, optimizing and testing D-peptide inhibitors against PSMα3.
finDr has a modality for performing mirror-image phage display in silico. To realize this we developed a toolset (PDBLibrary) for the creation of an up-to-date virtual peptide library to be used for computational screening against a target of choice.
Going beyond the straightforward screening approach we implemented the principles of Darwinian evolution to de novo create and optimize peptide ligands in silico with a genetic algorithm (GA). Through 14 generations of selection, recombination and mutation, two promising binders against PSMα3 were evolved. We chemically synthesized these GA-derived ligands and determined their binding kinetic to PSMα3 by using the biolayer interferometry (BLI) method.
We have combined two strategies to increase the efficiency of in vivo D-AAs incorporation into a protein of interest using E. coli: the previously established C321.DeltaA strain, where all amber stop codons and Release Factor 1 have been deleted from the genome, and the recently proposed compartmentalization of the orthogonal translation machinery into liquid droplets (1).
A significant increase of GFP fluorescence could be registered for our optimized translational system, especially in the C321.DeltaA strain, although it could not be definitively proven that solely D-phenylalanine incorporation is taking place. Within this subproject, we also successfully proved that the C.elegans protein SPD5 can form liquid droplets in E. coli.
References
- Reinkemeier et. al (2019): Designer membraneless organelles enable codon reassignment of selected mRNAs in eukaryotes. In: Science (New York, N.Y.) 363 (6434)
