The Diversity of Cadherins and Implications for a Synaptic Adhesive Code in the CNS

There are now, conservatively, at least 80 cadherins likely to be expressed in nervous tissue. These can be grouped into two broad categories: the classical cadherins and the CNR/protocadherins. There is already data showing that certain protocadherins are homophilically adhesive (Obata et al. 1995xObata, S., Sago, H., Mori, N., Rochelle, J.M., Seldin, M.F., Davidson, M., St. John, T., Taketani, S., and Suzuki, S.T. J. Cell Sci. 1995; 108: 3765–3773PubMedSee all ReferencesObata et al. 1995) and have a synaptic localization. Whether this is the full extent of their in vivo binding preference remains unclear. One of the notable structural features of the newly described protocadherins is an RGD (Arg-Gly-Asp) sequence conserved in the amino-terminal domain. This suggests the possibility that the protocadherins may also act as membrane-associated ligands for integrins. We have performed secondary structure modeling of some of these protocadherins, and the RGD sequence is predicted to protrude as a loop between two secondary structure elements, just as do functional RGD sequences from fibronectins (Shapiro and Colman 1998xShapiro, L. and Colman, D.R. Curr. Opin. Neurobiol. 1998; 8: 593–599Crossref | PubMed | Scopus (32)See all ReferencesShapiro and Colman 1998). Integrins are expressed early in neurogenesis and one—at least—is localized to synapses (Einheber et al. 1996xEinheber, S., Schnapp, L.M., Salzer, J.L., Cappiello, Z.B., and Milner, T.A. J. Comp. Neurol. 1996; 370: 105–134Crossref | PubMed | Scopus (98)See all ReferencesEinheber et al. 1996), but their partners are unknown. The precedent exists for integrin binding of transmembrane proteins (ICAM and VCAM) to function in cell–cell (rather than cell–matrix) adhesion. Also, a Drosophila learning mutant, volado, has been traced to an integrin mutation (Grotewiel et al. 1998xGrotewiel, M.S., Beck, C.D., Wu, K.H., Zhu, X.R., and Davis, R.L. Nature. 1998; 391: 455–460Crossref | PubMed | Scopus (221)See all ReferencesGrotewiel et al. 1998), suggesting that this mutation may have direct synaptic effects.What might be the functional implications of the remarkable genomic organization of the protocadherins uncovered by Wu and Maniatis? As suggested above, consolidation of these protocadherin genes in one region of a chromosome provides the potential for a central control mechanism for their expression. In this arrangement, the ectodomains are evolving their specificities separately from the cytoplasmic domain, which remains true to its common function for all ectodomains to which it is joined. The nearly perfect conservation of the constant cytoplasmic exons between mouse and human is evidence for evolutionary pressure to keep this domain very stable. In a conventional gene arrangement, a large number of complete genes would be needed to encode the different extracellular adhesive specificities, and each gene would of course contain those exons encoding the conserved cytoplasmic domain. The common cytoplasmic elements would be represented, then, multiple times throughout the genome. The novel and tightly integrated arrangement of the newly described protocadherin locus “streamlines” the conventional gene organization and may minimize potentially disadvantageous homologous recombination. It will be of great interest to identify all the intracellular binding partners of these conserved cytoplasmic segments, since they differ so markedly from their catenin-binding counterparts in the classical cadherins.Although there may be several adhesion mechanisms operative at the synapse that almost certainly involve members of other adhesion protein families, it is now clear that our broad knowledge of the cadherins provides a springboard for framing new and reframing old problems of profound importance to molecular neuroscience. Will topographic patterns of different adhesive molecules be found that delineate neural systems, like the olfactory and retinotectal systems? Will we find that the identity of synaptic adhesion molecules correlates with the type of neurotransmitter receptor at a particular synapse? Do morphogen gradients progressively restrict or activate the set of adhesion molecules that are expressed on the neurite surface as the target nears? How many specifiers of adhesivity are expressed on a single neuron's surface, and what is the distribution pattern of these adhesion molecules with respect to the thousands of synapses this neuron may form? Answers to these and related questions must yield a new and integrated view of synaptic adhesion and synaptic function, and the factors affecting their interrelationship, essential for a complete understanding of the molecular mechanisms underlying both the formation of nerve connections and the physiology of the mature synapse.‡To whom correspondence should be addressed (e-mail: colman@msvax.mssm.edu and shaprio@anguilla.physbio.mssm.edu).