Simon Erlendsson:
Molecular Recognition within Synaptic Scaffolds

Date: 15-06-2015    Supervisor: Kaare Teilum

Scaffolding proteins are abundant participants and regulators of the extensive intracellular framework required for maintaining cellular functions such as cellular adhesion and signal transduction cascades. In excitatory neuronal synapses these scaffolding proteins often contain one or more PDZ domains, responsible for tethering their respective synaptic protein ligands. Therefore, understanding the specificity and binding mechanisms of PDZ domain proteins is essential to understand regulation of synaptic plasticity. PICK1 is a PDZ domain-containing scaffolding protein predominantly expressed and characterized in the postsynaptic neurons, where it is involved in regulating processes underlying LTP and LTD. However, PICK1 has also been found to interact with a wide range of other regulatory proteins, receptors and transporters, which implicates PICK1 in several processes important for proper synaptic function.

At the molecular level PICK1 contains both a BAR and a PDZ domain making it quite unique. Especially the specificity and promiscuity of the PICK1 PDZ domain seems to be more complicated than normally seen for PDZ domains. Also, the ability of PICK1 to form dimeric structures via its central BAR domain has been fairly broadly accepted but both structurally and functionally sparsely described in literature. Moreover, extrapolating findings in vitro to functionality in a cellular context has been problematic, because the interactions are often membrane proximal and probably also affected by the spatial architecture of the synapse itself. In this thesis, the molecular scaffolding mechanisms of PICK1 have been investigated in both isolated and near native conditions. Our findings have significantly benefitted the general understanding of how PICK1 and PDZ domain scaffolding works.

In the first study, we investigated the PICK1 PDZ domain at a very detailed structural level, and characterized various binding modes responsible for coordinating such structurally diverse ligands. We were able to solve a high resolution NMR structure of the PICK1 PDZ DAT interaction as well as identifying two non-canonical ligand binding modes including an upstream binding motif outside the confined PDZ binding pocket for class I ligands, and an upper insert allowing for binding of internal PDZ binding motifs for unclassified PDZ ligands. Furthermore, our results suggest that these non-canonical binding motifs have evolved later in evolution to accommodate increasingly diverse PDZ domain ligands. Our findings provide basis for development of new and more specific peptide inhibitors.

In the second study, we utilized SAXS, NMR spectroscopy, MD simulations and various other biochemical methods, to construct a full-length structural model of the PICK1 dimer in-solution. We found the PICK1 BAR dimer to resemble an elongated crescent-shaped structure, spanning ~160 Å, with the PICK1 PDZ domains loosely attached to the BAR domain. This finding is in contrast to previous findings for other BAR domain proteins, where adjacent domains are rigidly attached to the BAR domain. We were also able to characterize a tetrameric assembly where the central PDZ domains of adjacent BAR domain dimers would come into immediate proximity. Lastly, we were able to show that the PICK1 LKV construct possibly alleviates an auto-inhibitory mechanism of PICK1 and allows the N-BAR domains or the PDZ domains themselves to cluster and shape membranes.

Finally, we utilized our in-solution structural knowledge to investigate the scaffolding events in context of a native cell membrane. We initially showed that we were able to qualitatively assess and quantify scaffolding protein mediated binding events on the intracellular side of the plasma membrane using confocal microscopy. We observed a large gain in functional affinity compared to measured in-solution intrinsic affinities, which could be related to bivalent PDZ domain binding and general membrane avidity. Surprisingly, we were able to demonstrate that such kinetics have large functional consequences for the exact mode of scaffolding, as well as the ability to regulate and accumulate in the synapse. We propose that this finding can extend to cover a vast majority of other scaffolding proteins, and therefore introduces a new framework for understanding cellular scaffolding in general.