Molecular Interactions of Shank Scaffolds Signaling Proteins

Several interactors of the Shank PDZ domains have been suggested to play a role in morphogenic signaling within spines. Most notable, the guanine nucleotide exchange factor PPIX is attached to Shank via this type of interaction, involving the C-terminal PDZ ligand motif of PIX and an additional leucine zipper, also located in the C-terminal region of PPIX (Park et al. 2003). PPIX is an exchange factor for rho GTPases, in particular for cdc42 and rac. Both GTPases have been implicated in the localized restructuring of actin networks in neurons and other cell types. Rac has been implicated in the generation of dendritic spines; one downstream effector for rac (and cdc42), the p21-activated protein kinase (PAK1-3), also interacts with PPIX and can be targeted to the postsynaptic density through the Shank/PPIX complex (Park et al. 2003). In addition, the interaction between IRSp53 and Shank (see above) is dependent on prior activation of IRSp53 by cdc42 (Soltau et al. 2002). Thus, it is conceivable that Shank and associated proteins form a signalosome complex for rho-type GTPases. Though it is not known what the final signaling output of this complex may be, a role in the structural maturation of spines is likely.

The concept that Shank establishes a signaling platform at the synapse is further supported by recent findings by Proepper et al. (2007), who identified an interaction of a proline-rich segment of Shank2 and Shank3 (but not Shank1) with the SH3 domain of the Abelson tyrosine kinase interacting protein 1 (Abi-1). In non-neuronal cell types, Abi-1 has been mainly implicated in signaling from receptor tyrosine kinases to the activation of rac, and furthermore to restructuring of the actin cytoskeleton. Abi-1 is localized to the PSD through interaction with Shank, and depletion of Abi-1 by RNA interference interferes with spine maturation and synapse formation, while in parallel allowing for excessive formation of dendritic branches. This may be ascribed to known actin-regulatory functions of Abi-1, which are tethered to locations of active synaptogenesis by Shank. Furthermore, once the PSD has been established, Abi-1 appears to assume a new type of signaling function as it travels from the PSD to the nucleus after synaptic stimulation of NMDA receptors. Within the nucleus, Abi-1 is involved in transcriptional regulation, suggesting that Abi-1 constitutes one of the cellular messengers turning a short-term synaptic stimulus into long-term changes in gene expression (Proepper et al. 2007).

We have identified an interaction of the SH3 domain of Shank1 and Shank3 with the PSD scaffold protein densin-180; in the PSD, densin-180 can interact additionally with the Ca-/calmodulin-dependent kinase II and a-actinin (S track et al. 2000; Robison et al. 2005). Interestingly, the N-terminal leucine-rich repeat region of densin-180 contributes to neuronal morphology by generating excessively branched dendrites in overexpression experiments. The effects of densin overexpression suggest that it is involved in signaling via 5-catenin, which also contributes strongly to dendrite branching when overexpressed (Martinez et al. 2003). Similar to Shank, 5-catenin interacts with the C-terminal region of densin (Izawa et al. 2002) and appears to act synergistically with densin on branching (unpublished data). One aspect of how Shank interferes with the effect of densin-180 on branching may be that it competitively blocks interaction of 5-catenin with densin. So far, it is however unclear how this is transduced to the N-terminal leucine-rich repeat region of densin, which is the main effector region regulating neuronal morphology. Densin is recruited into PSD-type clusters when Shank is coexpressed in neurons, again confirming the role of Shank in the maturation of the postsynapse (Quitsch et al. 2005).

For some interaction partners of Shank, their role at the synapse remains somewhat enigmatic at present. Sharpin binds to the ankyrin repeat region of Shank proteins and may add to the formation of Shank-containing protein clusters at the PSD, as it contains an N-terminal dimerization motif. However, sharpin expression is not limited to neurons, and even in neurons substantial amounts of the protein are found in non-synaptic compartments where it is likely to fulfill a so far unknown function (Lim et al. 2001).

Another potentially very interesting contact occurs between dynamin-2 and a sequence element within the proline-rich region of Shank1 and Shank2 containing multiple serine residues (Okamoto et al. 2001). Dynamin isoforms are GTPases that are known as essential elements of the endocytotic machinery. Dynamin-2 tethered to the PSD via Shank may play a role in the endocytosis of ionotropic glutamate receptors of the AMPA- or NMDA-type (Scott et al. 2004). However, no direct function for dynamin-2 in these processes has so far been demonstrated.

0 0

Post a comment