Metalloprotease-Disintegrins: Links to Cell Adhesion and Cleavage of TNFα and Notch

Why are such functionally distinct protein modules as a metalloprotease domain and a disintegrin domain combined in MDC proteins? With respect to the metalloprotease functions, one possibility is that the disintegrin domain might be used to target the metalloprotease to another cell in trans via an integrin (see Figure 1Figure 1). Alternatively, the disintegrin domain, or other protein domains such as the EGF repeat or cysteine-rich region, might be used to increase the efficiency of the protease by binding the substrate directly or indirectly in cis or in trans. It is also possible that removal of the metalloprotease domain may be used to regulate the function of the disintegrin domain, as suggested by two independent studies. In sperm fertilin β, removal of the noncatalytic metalloprotease-domain during sperm maturation correlates with the acquisition of fertilization competence and exposes an epitope that is recognized by a function-blocking monoclonal antibody. For meltrin α, which plays a role in muscle fusion, overexpression of a truncated form of the protein lacking the metalloprotease domain leads to an increase in muscle fusion, whereas overexpression of the full-length protein leads to a decrease in the observed fusion (Yagami-Hiromasa et al. 1995xYagami-Hiromasa, T., Sato, T., Kurisaki, T., Kamijo, K., Nabeshima, Y., and Fujisawa-Sehara, A. Nature. 1995; 377: 652–656Crossref | PubMedSee all ReferencesYagami-Hiromasa et al. 1995). Since only a small percentage of the detectable meltrin α lacks the metalloprotease domain in C2C12 mouse myoblasts, one interesting possibility is that only this small pool of processed protein may promote muscle cell binding and fusion. MDC proteins have been proposed to mediate cell–cell fusion directly (seeWolfsberg and White 1996xWolfsberg, T.G. and White, J.M. Dev. Biol. 1996; 180: 389–401Crossref | PubMed | Scopus (205)See all ReferencesWolfsberg and White 1996, for details), although it should be noted that the studies that point toward a role of fertilin and meltrin α in membrane fusion can also be explained by simply invoking a critical binding step via the disintegrin domain that is a prerequisite for fusion to occur.The concepts of targeting the metalloprotease domain via the disintegrin domain and modulation of the disintegrin domain function by removal of the metalloprotease domain illustrate two of several conceivable means of MDC protein regulation that are not necessarily mutually exclusive. Different MDC proteins may employ these protein modules in different ways, and any particular protein might also have distinct functions depending on the stage of development, the tissue, or even the subcellular localization that it is expressed in. The Drosophila metalloprotease-disintegrin KUZ, which is currently the only MDC protein for which a mutant phenotype has been reported, is required in the early embryo for neural inhibition, and is later involved in eye development, neural-promoting and -inhibiting processes (Rooke et al. 1996xRooke, J., Pan, D., Xu, T., and Rubin, G.M. Science. 1996; 273: 1227–1230Crossref | PubMedSee all ReferencesRooke et al. 1996), and axon extension (Fambrough et al. 1996xFambrough, D., Pan, D., Rubin, G.M., and Goodman, C.S. Proc. Natl. Acad. Sci. USA. 1996; 93: 13233–13238Crossref | PubMed | Scopus (141)See all ReferencesFambrough et al. 1996). A fascinating indication for a potentially distinct mechanism of the neural promoting and inhibiting activity mediated by KUZ is described by Rooke et al. 1996xRooke, J., Pan, D., Xu, T., and Rubin, G.M. Science. 1996; 273: 1227–1230Crossref | PubMedSee all ReferencesRooke et al. 1996. In the cuticle of adult mosaic Drosophila, clusters of sensory bristles appear at the boundary of kuz− and wildtype cells instead of the single sensory bristle that normally develops, whereas no sensory bristles are found in the interior of a mutant cell patch. Apparently a non-cell-autonomous neural-promoting signal can be supplied to the mutant cells by adjacent wildtype cells, and not by kuz− cells. Yet once kuz− cells have adopted a neural fate, they are unable to laterally inhibit the neural fate of other mutant cells, resulting in the formation of additional bristles only at the boundary of the mutant cell patch.Pan and Rubin 1997xPan, D. and Rubin, G.M. Cell. 1997; 90: 271–280Abstract | Full Text | Full Text PDF | PubMed | Scopus (299)See all ReferencesPan and Rubin 1997 have now provided strong genetic and biochemical evidence that the lateral inhibition mediated by KUZ involves a specific cleavage event in the extracellular domain of the transmembrane receptor Notch. Cleavage of Notch receptors in the extracellular domain appears to be an evolutionarily conserved feature, and the subcellular location of human Notch2 processing has been narrowed down to the trans-Golgi network (Blaumueller et al. 1997xBlaumueller, C.M., Qi, H., Zagouras, P., and Artavanis-Tsakonas, S. Cell. 1997; 90: 281–291Abstract | Full Text | Full Text PDF | PubMed | Scopus (416)See all ReferencesBlaumueller et al. 1997). Processing of the full-length ∼300 kDa human Notch2 yields a 110 kDa fragment containing the transmembrane domain and cytoplasmic tail, and a disulfide-linked 180 kDa fragment that most likely corresponds to the extracellular domain. Cell surface labeling experiments indicate that only cleaved but not full-length Notch2 emerges on the cell surface. The cleaved 110 kDa membrane-anchored fragment of human Notch2 resembles the ∼100 kDa processed form of Drosophila Notch. Since the ∼100 kDa form of Notch is not detectable in kuz− embryosPan and Rubin 1997xPan, D. and Rubin, G.M. Cell. 1997; 90: 271–280Abstract | Full Text | Full Text PDF | PubMed | Scopus (299)See all ReferencesPan and Rubin 1997 propose that KUZ mediates the extracellular cleavage of Notch and that this cleavage is necessary for Notch to mediate lateral inhibition. Taken together, the two papers support a model where KUZ or its mammalian homolog MDC/ADAM10 is responsible for maturation and functional activation of Notch receptors in the secretory pathway (4xBlaumueller, C.M., Qi, H., Zagouras, P., and Artavanis-Tsakonas, S. Cell. 1997; 90: 281–291Abstract | Full Text | Full Text PDF | PubMed | Scopus (416)See all References, 11xPan, D. and Rubin, G.M. Cell. 1997; 90: 271–280Abstract | Full Text | Full Text PDF | PubMed | Scopus (299)See all References). It should be noted that the extracellular domain cleavage of Notch discussed here is distinct from a putative cytoplasmic cleavage that might allow the cytoplasmic domain to enter the nucleus (Blaumueller et al. 1997xBlaumueller, C.M., Qi, H., Zagouras, P., and Artavanis-Tsakonas, S. Cell. 1997; 90: 281–291Abstract | Full Text | Full Text PDF | PubMed | Scopus (416)See all ReferencesBlaumueller et al. 1997).If the processing of Notch receptors is evolutionarily conserved, then one might expect the Notch processing protease to be functionally conserved as well. IndeedPan and Rubin 1997xPan, D. and Rubin, G.M. Cell. 1997; 90: 271–280Abstract | Full Text | Full Text PDF | PubMed | Scopus (299)See all ReferencesPan and Rubin 1997 show that expression of a mouse dominant negative KUZ lacking the metalloprotease domain (KUZDN) in both Drosophila and Xenopus laevis results in an increased number of neurogenic cells, presumably due to a lack of lateral inhibition. Furthermore, expression of a Drosophila KUZDN in Drosophila neurons mimics the defect in axon extension reported by Fambrough et al. 1996xFambrough, D., Pan, D., Rubin, G.M., and Goodman, C.S. Proc. Natl. Acad. Sci. USA. 1996; 93: 13233–13238Crossref | PubMed | Scopus (141)See all ReferencesFambrough et al. 1996, confirming the idea that the metalloprotease domain of KUZ, as opposed to other domains, is responsible for axon extension. The substrates of KUZ during axonal extension have not been identified but could be different from Notch, such as matrix proteins or cytokines.In conclusion, it is clear that metalloprotease-disintegrins are involved in a remarkably diverse set of tasks, ranging from a role in fertilization and muscle fusion, TNFα release from the plasma membrane, intracellular cleavage and activation of Notch, and other essential functions in Drosophila development. The nature of these diverse tasks further suggests that MDC proteins may function at different subcellular locations, such as on the cell surface (fertilin, TACE?), or intracellularly in the secretory pathway (KUZ?). It seems likely that the proteins discussed here, and the more than 20 family members of presently unknown function, which have been found in organisms ranging from C. elegans to mammals (references can be found inWolfsberg and White 1996xWolfsberg, T.G. and White, J.M. Dev. Biol. 1996; 180: 389–401Crossref | PubMed | Scopus (205)See all ReferencesWolfsberg and White 1996) will unveil further exciting secrets. Due to an increasing interest in MDC proteins, a better understanding should soon begin to emerge about the mechanism of MDC protein function, of the specific functions of different family members in development and disease, and of the interactions with other proteins that govern MDC protein activity.