Dissecting Semaphorin Signaling

Previews 935

D., Kitamura, M., Hardy, S., Nicoll, R.A., Malenka, R.C., and Von Zastrow, M. (1998). Proc. Natl. Acad. Sci. USA 95, 7097–7102. Mammen, A.L., Huganir, R.L., and O’Brien, R.J. (1997). J. Neurosci. 17, 7351–7358. Miller, K.D. (1996). Neuron 17, 371–374. Nusser, Z., Cull-Candy, S.G., and Farrant, M. (1997). Neuron 19, 697–709. O’Brien, R.J., Kambol, S., Ehlers, M.D., Rosen, K.R., Kischback, G.D., and Huganir, R.L. (1998). Neuron 21, this issue, 1067–1078. Rutherford, L.C., Nelson, S.B., and Turrigiano, G.G. (1998). Neuron 21, 521–530. Turrigiano, G.G. (1998). Trends Neurosci., in press. Turrigiano, G.G., Leslie, K.R., Desai, N.S., Rutherford, L.C., and Nelson, S.B. (1998). Nature 391, 892–896.

Dissecting Semaphorin Signaling The precise wiring of neural architecture requires numerous signals for governing axon targeting. While chemotropic effects have been known for some time, only in the last decade has the importance of repulsive chemical cues become apparent. Several families of molecules have now been identified that are involved in inhibiting or repelling axon growth. Members of the Class III semaphorin family are secreted molecules that have been shown to act as repulsive factors for specific axonal populations. The first identified, SemaIII (also known as Collapsin-1 or SemaD), causes growth cone collapse and axonal retraction and repulsion in sensory and sympathetic axons in culture. The receptor for SemaD was identified last year as neuropilin-1, a transmembrane protein expressed in specific cell populations (see Kolodkin and Ginty, 1997). The importance of the Sema3/neuropilin-1 interaction for proper nervous system development in vivo was demonstrated by studies showing that both SemaD and neuropilin-1 mutations resulted in identical axonal projection defects (Kitsukawa et al., 1997; Taniguchi et al., 1997). Other semaphorins with repulsive activity in specific neuronal populations have also been identified. For instance, the related molecules Sema A, SemaE, and SemaIV repel sympathetic axons but have no effect on sensory axons. Three recent studies begin to clarify the mechanisms for the biological specificity of the semaphorins. Takahashi et al. (1998) and Giger et al. (1998 [this issue of Neuron]) report that neuropilin-2, identified last year in a homology screen, is the functional receptor for SemaA and SemaE (Takahashi et al., 1998) and SemaIV (Giger et al., 1998). Giger et al. show that neuropilin-2 is present in postganglionic sympathetic neurons, neuronal populations that respond to SemaIV, and they present evidence that SemaIV and neuropilin-2 are present during development in specific complementary patterns. They also demonstrate that expression of neuropilin-2 is necessary and sufficient to produce a collapse response to SemaIV. These studies begin to paint a picture in which repulsion of axons of specific neuronal subtypes is mediated by the interaction of specific semaphorin and neuropilin

family members. However, we know that semaphorin/ neuropilin binding alone is not sufficient for the repulsive activity. Neuropilin-1 binds SemaD, SemaA, SemaE, and SemaIV with high affinity, yet only SemaD leads to growth cone collapse in neuropilin-1-expressing neurons (Chen et al., 1997; Koppel et al., 1997; Giger et al., 1998). Two studies in this issue of Neuron begin to dissect the molecular basis of the specificity of semaphorin/neuropilin interactions. Secreted class III semaphorins are known to have a conserved semaphorin domain, an Ig domain, and a basic tail (Chen et al., 1998; Kolodkin and Ginty, 1997). The semaphorin domain has been shown to be responsible for the binding specificities of the semaphorins in situ (Feiner et al., 1997), while the Ig-basic domain also exhibits binding properties. The neuropilins have two N-terminal domains similar to complement binding domains, called CUB domains; two coagulation factor domains; a C-terminal MAM domain; a transmembrane domain; and a short cytoplasmic tail (see figure). Both Giger et al. (1998) and Nakamura et al. (1998) made constructs lacking specific portions of the semaphorin and neuropilin proteins. Ligands had an alkaline phosphatase moiety fused to portions of chick SemaD, or human or rat SemaIII. Neuropilin-1 receptor constructs were made lacking a, b, or c domains, singly and in combination. Two types of assays were used to determine the specificity of the ligand/receptor interactions. Physical interactions were determined through binding studies performed when these molecules were expressed in COS cells, while the functional specificity of responses was assayed by monitoring growth cone collapse in neurons expressing the receptor constructs. Binding studies clearly demonstrated that the CUB domain is necessary for physical binding of the sema domain (Giger et al., 1998; Nakamura et al., 1998). Strittmatter and colleagues (Nakamura et al., 1998) also tested whether the CUB domain alone was sufficient to confer binding specificity

Semaphorin and Neuropilin Domains

Neuron 936

of semaphorins. Neuropilin-2 CUB domains were fused to the neuropilin-1 coagulation factor, MAM, transmembrane, and cytoplasmic domains. The chimeric receptor demonstrated neuropilin-2-like binding (Nakamura et al., 1998). Some discrepancies in the binding data, which may be due to the particular constructs used in the two studies, make the role of the coagulation factor domains for binding more difficult to determine. Binding studies alone, however, are not proof of functional specificity. Giger et al. addressed the issue of functional specificity by fusing the CUB and coagulation factor domains of neuropilin-2 to the MAM, transmembrane, and C-terminal domains of neuropilin-1 and expressing the chimeric receptor in sensory neurons. The neurons, normally repelled by SemaIII but unaffected by SemaIV, now grew away from a SemaIV source (Giger et al., 1998). This experiment demonstrated that the CUB and coagulation factor domains are sufficient for conferring functional specificity of the receptor. Further studies showed that binding of the Ig-basic ligand to the coagulation factor domain was not sufficient to confer a collapse response, suggesting that sema binding to the CUB domain is necessary for a functional response (Giger et al., 1998). Taken together, the binding and functional assays show that the CUB domain is necessary for specific semaphorin binding, and that the CUB and coagulation factor domains together are sufficient to mediate response specificity. However, the possibility remains that the CUB domain alone is sufficient for functional specificity; this needs to be tested directly. Binding studies with MAM domain deletion constructs expressed in COS cells clearly show that the MAM domain is not required for semaphorin binding (Giger et al., 1998; Nakamura et al., 1998). However, transfection of MAM deletion constructs into retinal explants fails to result in growth cone collapse in response to semaphorin, indicating that the MAM domain is necessary for signal transduction (Nakamura et al., 1998). This is consistent with the fact that the blocking antibodies used by Giger et al. are directed against the neuropilin-2 MAM domain. It is known that neuropilins dimerize to form functional receptors (Mark et al., 1997). Coimmunoprecipitation studies with neuropilins containing MAM domain deletions implicate the MAM domain in neuropilin oligomerization (Giger et al., 1998; Nakamura et al., 1998). The function of the cytoplasmic domain of the neuropilins has remained somewhat perplexing, as it is short and has no similarity to known signal sequences. This has raised the question of whether neuropilin multimers are sufficient to transduce repulsive signals, or whether they form part of a larger receptor complex (Feiner et al., 1997; Chen et al., 1998). Takahashi et al. (1998) have shown that expression of neuropilin-1 in E8 retinal ganglion cells, which normally lack neuropilin-1, results in a repulsive response to SemaD, suggesting that neuropilin dimers do not require accessory molecules to function. However, further tests of this hypothesis have led to surprising results. Nakamura et al. fused neuropilin-1 ectodomains to the transmembrane and cytoplasmic domains of neuropilin-2 and L1, a cell adhesion molecule unrelated to the neuropilins. Both constructs produced a collapse response in retinal ganglion cells presented with SemaD. Taking this approach to the limit, they removed transmembrane and cytoplasmic sequences,

GPI linking the ectodomain to the cell surface. This construct, too, produced a functional response in retinal ganglion cells, although the repulsion appeared to not be as potent as in cells expressing wild-type neuropilin-1. Thus, the extracellular domain of neuropilin-1 alone is sufficient to mediate a repulsive response, strong evidence that neuropilin-1 multimers form a receptor complex with an unidentified but promiscuous accessory molecule that transduces the collapse signal. As a recent study has shown that neuropilin-1 serves as a nonsignaling coreceptor with a receptor tyrosine kinase for the vascular endothelial growth factor VEGF (Soker et al., 1998), one is tempted to speculate whether the mystery molecule is another receptor tyrosine kinase. Future studies will focus upon identifying this missing partner, and on further elucidating the signaling cascade downstream of neuropilin binding, which ultimately leads to the depolymerization of the actin cytoskeleton. Adina L. Roskies Neuron Selected Reading Chen, H., Chedotal, A., He, Z., Goodman, C.S., and Tessier-Lavigne, M. (1997). Neuron 19, 547–559. Chen, H., He, Z., and Tessier-Lavigne, M. (1998). Nat. Neurosci. 1, 436–439. Feiner, L., Koppel, A.M., Kobayashi, H., and Raper, J.A. (1997). Neuron 19, 539–545. Giger, R.J., Urquhart, E.R., Gillespie, S.K.H., Levengood, D.V., Ginty, D.D., and Kolodkin, A.L. (1998). Neuron 21, this issue, 1079–1092. Kitsukawa, T., Shimizu, M., Sanbo, M., Hirata, T., Taniguchi, M., Bekku, Y., Yagi, T., and Fujisawa, H. (1997). Neuron 19, 995–1005. Kolodkin, A.L., and Ginty, D.D. (1997). Neuron 19, 1159–1162. Koppel, A.M., Feiner, L., Kobayashi, H., and Raper, J.A. (1997). Neuron 19, 531–537. Mark, M.D., Lohrum, M., and Puschel, A.W. (1997). Cell Tissue Res. 290, 299–306. Nakamura, F., Tanaka, M., Takahashi, T., Kalb, R.G., and Strittmatter, S.M. (1998). Neuron 21, this issue, 1093–1100. Soker, S., Takashima, S., Miao, H.Q., Neufeld, G., and Klagsbrun, M. (1998). Cell 92, 735–745. Takahashi, T., Nakamura, F., Jin, Z., Kalb, R.G., and Strittmatter, S.M. (1998). Nat. Neurosci. 1, 487–493. Taniguchi, M., Yuasa, S., Fujisawa, H., Naruse, I., Saga, S., Mishina, M., and Yagi, T. (1997). Neuron 19, 519–530.

Making Proteins at the Synapse: Activity-Regulated Translation and CPEB Cytoplasmic polyadenylation element binding protein (CPEB) is a sequence-specific RNA binding protein that was originally identified as playing a role in the translational control of c-Mos kinase in oocyte maturation (de Moor and Richter, 1997). Wu et al. (1998), in this issue of Neuron, present compelling observations that indicate a