Skip to main content

Capturing kinase-substrate pairs in intact mammalian cells by using unnatural amino acid mutagenesis and photocrosslinking

Final Report Summary - KINASE CROSSLINKING (Capturing kinase-substrate pairs in intact mammalian cells by using unnatural amino acid mutagenesis and photocrosslinking)

The basic concept of the project is using genetically encoded photo-crosslinking amino acids to capture protein-protein interactions in the live cell, in particular kinase-substrate pairs. The specific aim of the first year of the outgoing phase was to demonstrate the feasibility of the method using the well-known PKA/CREB pair as proof of principle. In first place we have addressed the issue of introducing photo-crosslinking amino acids into protein targets with the highest efficiency and minimal occurrence of side products. We have developed a novel versatile expression system that overcomes drawbacks of the existing ones, and allows incorporating photo-crosslinking amino acids derivatives of tyrosine into proteins expressed in mammalian cells with yields up to 30% of wild type.

Then, we have addressed the kinase-substrate interaction, and the specific challenges related to its transient nature. We decided to separate the problem in two parts. First we aimed at finding out in which positions of the kinase we should incorporate the crosslinker to capture an interacting partner. For this, we worked on the stable complex formed by PKA and its regulatory subunit RIIb, which reproduces the kinase-substrate interaction, but without the time variable intrinsic in the phosphorylation event. Once identified where to install the crosslinker, we would have aimed at capturing transiently interacting partners. However, the apparently simple first task turned out to be very challenging. We screened several positions throughout the protein kinase, both within the kinase catalytic pocket and in its proximity, using two different kinds of crosslinkers, but no obvious crosslinked products were formed. To better understand the potential and the limits of the use of genetically incorporated crosslinkers, we started in parallel investigating a peptide-protein interaction, namely the interaction between the class B G-protein coupled receptor (GPCR) corticotropin releasing factor receptor type 1 (CRF1R) and its native 40-mer peptide ligand Urocortin-I (Ucn1). From the methodological point of view the peptide-receptor interaction offers the big advantage that the position of the binding equilibrium can be influenced by applying different concentrations of ligand, so that the overall system is much easier to control. By introducing the photo-crosslinking amino acid “Azi” into CRF1R, we were indeed able to capture Ucn1 and clearly detect the covalent ligand-receptor complex in Western blot. To troubleshoot general crosslinking issues we then tested several crosslinking conditions. We came to the conclusion that for interactions having affinity constants in the nanomolar range the crosslinking yields with genetically incorporated amino acids are very low, even when the equilibrium is totally shifted toward the associated form. In the light of these results, at the end of the first year of work the chances of capturing the transient kinase-substrate crosslinking within the timeframe of the outgoing phase appeared extremely low. On the other hand, the methodology was lending itself to investigate the molecular determinants that lead to activation of a GPCR of class B by a natural peptide ligand. This is an exciting biological question and offers a quite broad open space of research, because interactions of GPCRs with long peptide ligands are hardly addressed by crystallography and no full-length 3D structure is available for any class B GPCR. Therefore, we have focused our effort in the CRF1R project. We have incorporated Azi systematically throughout the juxtamembrane domain of CRF1R, for a total of about 150 screened positions. Positions where the crosslinker captured the bound ligand identify receptor sites coming close to the ligand in the associated complex. We obtained a panoramic map of the peptide binding pocket in the activation domain of CRF1R at a single residue resolution, and identified at the same time hallmarks of structural elements of the receptor itself. This represents the first panoramic structural information derived from a fully post-translationally modified receptor in the native context of the cell membrane, and complements crystallographic data acquired in artificial environment. These exciting results prompted us to pinpoint intermolecular pairs of ligand and receptor residues proximal to each other, to understand how the ligand is oriented in the pocket. We have exploited a specific click reaction between the novel unnatural amino acid “Ffact” developed in Wang lab, which was genetically incorporated into the receptor, and cysteine residues chemically incorporated into the ligand at tolerant positions. The reaction is chemoselective and takes place only when the two amino acids are in close spatial proximity. By screening 115 combinations of FfactCRF1R-CysUcn1 variants we derived a number of ligand-receptor inter-molecular spatial constraints. In collaboration with the group of Ray Stevens at the Scripps Research Institute, the data were analysed in the context of the recently resolved crystal structure of the isolated TM domain of CRF1R and existing extracellular domain structures. We derived a complete conformational model for the peptide-receptor complex, which satisfies almost 50 independent experimental constraints. We discovered unique features of peptide ligand binding to class B GPCRs and gained new insights into the potential mechanism for receptor activation. Overall, we have developed a new method to map ligand-GPCR interactions directly in the native cellular environment, which complements biophysical reductionist approaches to understand functioning of GPCRs. As GPCRs are major pharmacological targets, the topic is interconnected with pharmacological research.

This experimental work was completed at the very end of the outgoing phase, so that the first months of the returning phase were dedicated to write and revise a manuscript, which has been published in the journal Cell. In the meantime, we have spotted in the lab of the returning host a molecular target that is a fantastic candidate to be investigated using the same crosslinking strategy. Recently, Scheidereit lab has identified a functionally crucial and possibly tight protein-protein interaction within the NF-kB signalling pathway, which involves the two NF-kB precursors. Not only the topology of the interaction is still unknown, but also evidence is missing to demonstrate that the interaction between the two proteins is direct, and the presence of third parties in the complex cannot be excluded. Based on the experience gained in the outgoing phase, we have genetically incorporated our crosslinking amino acid into one of the precursors and we found very clear evidence that specific crosslinked products are formed. We are planning further experiments to elucidate the nature of the crosslinked complexes and investigate interactions in which NF-kB precursors are involved. This will represent the base of a common project.

The fellow has now her independent position as a group leader at the University of Leipzig, supported by the prestigious Emmy-Noether award by the German Research Foundation. The work carried out during the outgoing phase at Salk and the knowledge acquired there were fundamental to enable her developing an articulate and outstanding proposal to start her independent research.