How synapses are formed and maintained is a fundamental question in neuroscience that is assuming increasing importance for our understanding of numerous neurological disorders. Recently, proteins containing complement factor C1Q-related domains (C1q-domains) have been attracting interest because of their potential role in synapse formation and/or maintenance. C1q-domains are small globular domains found in the eponymous complement factor C1Q and in a multitude of other proteins, including both small secreted proteins such as cerebellins and C1ql proteins, and large multidomain proteins such as emilins and multimerins (Ghai et al., 2007). Accumulating evidence is now implicating several C1q-domain proteins in synapse formation and/or elimination: cerebellin-1 was revealed to function in synapse maturation (Yuzaki, 2009) (Hirai et al., 2005), complement factor C1Q was implicated in synapse elimination (Stevens et al., 2007), and members of a third C1q-domain protein family, C1qls (for C1q-like), were demonstrated to reduce synapse numbers in cultured neurons (Bolliger et al., 2011). New data now published in this issue of EJN (Matsuda and Yuzaki, 2011), complementing earlier reports published in Cell and Science (Matsuda et al., 2010; Uemura et al., 2010), reveal that cerebellin-1 functions in synapse formation by binding to the presynaptic cell-adhesion neurexin molecules, and that other cerebellin isoforms may also do so.
Cerebellin-1 is secreted from cerebellar granule cells, and acts as a ligand for the postsynaptic glutamate receptor (GluR)δ2 on Purkinje cells (Matsuda et al., 2010). Cerebellin-1 additionally binds to presynaptic neurexin-1β, resulting in a trimeric trans-synaptic complex composed of cerebellin-1, GluRδ2, and neurexin-1β (Uemura et al., 2010). In the present EJN article, Matsuda and Yuzaki use cell-based binding assays to demonstrate that cerebellin-1 also binds to neurexin-1α, as well as to all three β-neurexins. This binding interaction requires the inclusion of an alternatively spliced exon known as ‘splice site #4’ (Ushkaryov et al., 1992), and appears to be Ca2+-independent. The cerebellin-1 binding properties for neurexins are thus distinct from those of neuroligins (Ichtchenko et al., 1995) or LRRTMs (Ko et al., 2009; de Wit et al., 2009). This result suggests that alternative splicing of neurexins at splice site #4 influences whether they bind to either cerebellin-1 or to neuroligins and LRRTM2.
Matsuda and Yuzaki further demonstrate that cerebellin-2 (but, interestingly, not cerebellin-4) also binds to neurexin-1β. The authors provide evidence that cerebellin-1 and cerebellin-2 have similar synaptogenic activities in cultured hippocampal and cortical neurons, not just in cerebellar neurons. Thus, the new article provides further evidence that cerebellin-1 acts as a bi-directional synapse organizer, initiating the recruitment of both presynaptic and postsynaptic proteins to the points of contact with both neurexin-1β and GluRδ2.
The picture of cerebellins emerging from these studies suggests that they function as ‘connectors’, linking presynaptic neurexins to postsynaptic GluRδ-type receptors, and that, in doing so, they activate synapse formation. This attractive hypothesis raises a slew of new questions. For example, what is the mechanism by which such binding activates synapse formation – possibly by multimerization of the binding partners – and does this binding initiate synapse formation, or stabilize transient synapses established by other mechanisms? On a higher level, these results raise the questions of whether cerebellin binding is the primary function of neurexin proteins containing splice site #4, how important such binding is for the formation of brain circuits, and whether other postsynaptic cerebellin receptors exist. Answers to these questions will yield a greater understanding of how synapses are formed and maintained in the cerebellum and the numerous other brain regions where cerebellin family members are expressed.