CAPS in Search of a Lost Function

Neuron 2

and wild-type animals treated with paroxetine (Zhou et al., 2002). Serotonin uptake by catecholamine neurons was also observed in mice lacking one of the monoamine oxidase genes (MAO-A) (Cases et al., 1998). The evidence that antidepressants induce serotonin accumulation by dopamine neurons has now been significantly advanced by a study in this issue of Neuron from John Dani’s laboratory (Zhou et al., 2005). The investigators adapted rapid electrochemical measurements of evoked dopamine and serotonin release in acute “horizontal” slices that encompass a portion of the innervating axons from midbrain dopamine neurons, and their primary target area, the striatum, where both dopamine and DAT are present at far higher levels than serotonin and SERT. To record dopamine and serotonin release evoked by electrical stimuli, they used cyclic voltammetry, in which ramp voltages are applied to a carbon fiber electrode and that to some extent differentiates the transmitters on the basis of the I-V relationships of their oxidation and reduction peaks. They further noted that the serotonin component could be identified, due to the relatively greater adsorption of serotonin to the carbon surface, simply by analyzing a later point during the signal’s falling phase. When the investigators exposed striatal tissue to fluoxetine in the presence of nisoxetine, a specific NET inhibitor, evoked serotonin release increased, whereas dopamine release decreased. Serotonin may thus act as a false transmitter after exposure to SSRIs. The signal, however, represents the release of transmitter from hundreds of synaptic vesicle fusion events and could alternatively reflect an enhanced release of serotonin from its native terminals. They addressed this possibility by examining nonevoked spontaneous release events, which are much smaller and likely reflect transmitter release from dozens of synaptic vesicles at synaptic varicosities along the incoming dopamine fibers. Each of these smaller events likewise contained both dopamine and serotonin, further evidence consistent with corelease. Perhaps more convincingly, fluoxetine’s effect was inhibited by the selective DAT inhibitor GBR12909, as would be predicted if the serotonin was accumulated by DAT. The authors thus conclude that when SERT is blocked by SSRIs, serotonin acts as a false transmitter in dopamine neurons. Experiments in vivo suggest that the effect may require several days of administration, which could underlie the well-known delay in full therapeutic benefit of these drugs. And to bring the story full circle, they found that the MAO inhibitor clorgyline further promotes serotonin uptake and release by dopamine neurons. The data clearly indicate that, at least under some conditions, both major classes of antidepressants cause serotonin to act as false transmitter in dopamine neurons. It is not yet known if such serotonin release by dopamine neurons contributes to the therapeutic effect of these agents, and it would be very interesting to know whether inhibition of DAT blocks the antidepressant effects of SSRIs. The Dani lab, in the meantime, can take credit for an elegant proof of a phenomenon that may underlie the effects, and perhaps the delayed response, for the many patients who take these drugs. And well in time for the 50th anniversary of Hughes’ and Brodie’s seminal discovery.

David Sulzer1 and Robert H. Edwards2 Departments of Neurology, Psychiatry, and Pharmacology Columbia University Department of Neuroscience New York State Psychiatric Institute New York, New York 10032 2 Departments of Neurology and Physiology University of California, San Francisco School of Medicine San Francisco, California 94143 1

Selected Reading Blackwell, B., and Mabbitt, L.A. (1965). Lancet 62, 938–940. Cases, O., Lebrand, C., Giros, B., Vitalis, T., De Maeyer, E., Caron, M.G., Price, D.J., Gaspar, P., and Seif, I. (1998). J. Neurosci. 18, 6914–6927. Erickson, J.D., and Eiden, L.E. (1993). J. Neurochem. 61, 2314– 2317. Gu, H., Wall, S.C., and Rudnick, G. (1994). J. Biol. Chem. 269, 7124–7130. Hughes, F.B., and Brodie, B.B. (1959). J. Pharmacol. Exp. Ther. 127, 96–102. Hughes, F.B., Shore, P.A., and Brodie, B.B. (1958). Experientia 14, 178–180. Kopin, I.J. (1968). Annu. Rev. Pharmacol. 8, 377–394. Peter, D., Jimenez, J., Liu, Y., Kim, J., and Edwards, R.H. (1994). J. Biol. Chem. 269, 7231–7237. Rocha, B.A., Fumagalli, F., Gainetdinov, R.R., Jones, S.R., Ator, R., Giros, B., Miller, G.W., and Caron, M.G. (1998). Nat. Neurosci. 1, 132–137. Vanhatalo, S., and Soinila, S. (1995). Neurosci. Res. 22, 367–374. Zhou, F.C., Lesch, K.P., and Murphy, D.L. (2002). Brain Res. 942, 109–119. Zhou, F.-M., Liang, Y., Salas, R., Zhang, L., De Biasi, M., and Dani, J.A. (2005). Neuron 46, this issue, 65–74. DOI 10.1016/j.neuron.2005.03.013

CAPS in Search of a Lost Function Ca2+-dependent activator protein for secretion (CAPS) is an evolutionarily conserved secretory protein that was previously thought to mediate Ca2+-triggered fusion of dense-core vesicles. In an elegant study of CAPS1-deficient mice, Speidel et al. (this issue of Neuron) now show that CAPS function may have been misunderstood. CAPS appears to act upstream of fusion in the biogenesis or maintenance of mature secretory vesicles, raising the possibility of a completely new type of function for an essential component of the secretory machinery.

CAPS was discovered independently in C. elegans as the unc-31 gene that is required for synaptic transmission (Livingstone, 1991) and in PC12 cells as an essen-

Previews 3

tial component of the Ca2+-triggering machinery for exocytosis (Walent et al., 1992). CAPS is an w140 kDa protein that contains two major identifiable domains: a central PH domain that binds to phospholipids, and a C-terminal region that is homologous to Munc13s, proteins that are involved in synaptic vesicle priming (Speidel et al., 2003). CAPS is highly conserved in evolution; a single CAPS isoform is expressed in invertebrates, whereas two closely related isoforms are produced in vertebrates (referred to as CAPS1 and CAPS2; Speidel et al., 2003). CAPS forms dimers and interacts with phospholipids, suggesting a membrane-associated function (Walent et al., 1992). CAPS is primarily expressed in neurons and neuroendocrine cells, although nonneuronal tissues such as liver contain significant levels (Walent et al., 1992; Speidel et al., 2003). The seminal description of CAPS as an essential factor in the Ca2+-triggered release of norepinephrine from PC12 cells (Walent et al., 1992) spawned a series of studies on the role of CAPS in secretion in neuroendocrine cells. These studies led to the widely accepted conclusion that CAPS specifically mediates the Ca2+dependent exocytosis of large dense-core vesicles (LDCVs) and other dense-core vesicles (e.g., see Grishanin et al., 2004; Rupnik et al., 2000; Elhamdani et al., 1999; Tandon et al., 1998). A closer look, however, reveals that this conclusion may not be entirely certain. For example, the analysis of CAPS mutants in Drosophila (Renden et al., 2001) uncovered a selective decrease in evoked glutamatergic synaptic transmission at the neuromuscular junction of CAPS-deficient flies. In this study, the only evidence for a change in dense-core vesicle exocytosis was a moderate increase in the density of LDCVs in nerve terminals; however, the number of synaptic vesicles per active zone was also increased significantly in CAPS-deficient flies, casting doubt on a specific role for CAPS in LDCV exocytosis. The new results of Speidel et al. (2005) (this issue of Neuron), produced by a technical tour-de-force, throw fresh light on CAPS function. Speidel et al. generated KO mice lacking CAPS1. Homozygous mutant mice died within 30 min after birth, but exhibited no obvious developmental or biochemical abnormalities. The authors then performed an in-depth analysis of exocytosis in chromaffin cells at two developmental stages: embryonic homozygous mutant mice (at day E19, just before birth), and adolescent heterozygous mutant mice (at day P30). It should be stated at the outset that Speidel et al. do not propose to present a definitive function for CAPS. With the results of Speidel et al. available, we still don't know what CAPS does, but we know much more about what it does not do, and we have definitive ideas about what it might do. These intriguing results will stimulate the field and initiate many new studies on this fascinating protein. Speidel et al. show that in homozygous CAPS1 KO mice at E19, chromaffin cells exhibited no change in the amplitude or rate of Ca2+-triggered exocytosis. By itself, this results is not surprising, since CAPS2 is expressed at 8-fold higher levels than CAPS1 in the adrenals at this stage of development, and thus the lack of a phenotype in exocytosis could have been simply due to redundancy. Strikingly, however, using amperometry,

Speidel et al. found that only 40% of the granules undergoing exocytosis were actually filled with catecholamines! This unexpected phenotype was not due to a developmental defect, because the phenotype could be rescued by expression of recombinant CAPS1. Quantitative electron microscopy of the CAPS1-deficient cells uncovered no changes—in particular, the number and appearance of chromaffin granules were not altered. Biochemically, the levels of adrenaline and noradrenaline were normal, but the concentration of the metabolite DOPEG that is derived from adrenaline and noradrenaline was increased 3- to 4-fold, suggesting that noradrenaline and adrenaline turn over much more rapidly in the CAPS1-deficient cells than in control cells. Overall, the results obtained by Speidel et al. in mutant embryonic chromaffin cells point toward an essential role for CAPS1 in maintaining stable secretory vesicles that are filled with neurotransmitters. Empty secretory vesicles were previously observed at a low frequency (7%) in wild-type chromaffin cells, but increased to >40% after treatment of cells with the catecholamine uptake inhibitor reserpine (Tabares et al., 2001). Similar to the results reported by Speidel et al., the vesicles characterized by Tabares et al. were also either completely full or completely empty. Thus, Speidel et al. conclusively show that in embryonic cells, CAPS1 in some manner is essential for the normal filling of vesicles with catecholamines. The phenotype observed in young adult heterozygous CAPS1-deficient chromaffin cells was quite different from that of embryonic homozygous mutant cells. The heterozygous mutant cells (which have approximately half the wild-type levels of CAPS1) exhibited a moderate decrease (w30%) in the amplitude of the fast and slow component of the exocytosis, but no change in the kinetics of release. Amperometry showed that all granules were filled with catecholamines at wild-type levels in the heterozygous mutant cells. Electron microscopy demonstrated that the mutant chromaffin cells were overall normal, but uncovered a single change, namely a modest but significant decrease (w30%) in the density of vesicles close (<200 nm) to the plasma membrane. At first glance, the phenotypes of the homozygous embryonic and the heterozygous adolescent chromaffin cells appear incompatible, but they can be reconciled by the simple assumption that CAPS normally functions in filling or maintaining filled chromaffin granules and that in mature cells empty granules are largely eliminated by a quality control mechanism, whereas in embryonic cells such a quality control mechanism does not yet operate. The analysis of the CAPS1-deficient chromaffin cells thus confirms that CAPS1 has a function in secretion, although it provides no evidence that this function is related to Ca2+ triggering. Normal CAPS1 levels clearly are not essential for normal fusion, and CAPS1 is unlikely to play a role in the fusion pore. This result highlights the sensitivity of fusion pores to overexpression, because CAPS1 is yet another protein in a long list of molecules that when overexpressed alter fusion pores without directly participating in fusion (e.g., see synaptotagmin 1, which alters fusion-pore

Neuron 4

opening upon overexpression, but does not normally participate in determining the kinetics of fusion-pore opening [reviewed in Su¨dhof, 2004]). The effects of CAPS1 overexpression are most likely due to the fact that any change in membrane composition or tension, be it ever so slight, will change fusion pores, and so, many overexpressed transfected proteins, even proteins that are unrelated to secretion, are likely to alter fusion pores. In addition, the data by Speidel et al. (2005) exclude a function for CAPS in diverse processes such as Ca2+ triggering of exocytosis or endocytosis. It is thus safe to say at this point that CAPS is important for secretion, but it is unsafe to say what exactly it is important for. Clearly, the original notion of a clearcut function in the Ca2+-dependent exocytosis of LDCVs is untenable. However, a specific role in catecholamine storage as such is also unlikely. Although such a role would explain why vesicles in the CAPS1 KO mice are empty (Speidel et al., 2005) and why recombinant CAPS1 is required for norepinephrine release from semi-intact PC12 cells (since empty vesicles would appear not to exocytose), such a function is inconsistent with the widespread presence of CAPS in all synapses—most of which do not secrete catecholamines—and even in tissues like liver, which is not known for Ca2+-triggered exocytosis (Speidel et al., 2003). Can a unifying hypothesis of CAPS function that satisfies all available data be postulated? A general role in the stabilization of exocytic compartments seems improbable because one would then expect that the number and/or shape of secretory vesicles should be altered. However, it is possible that CAPS functions in maintaining the pH gradient across the membrane of secretory vesicles as they wait to be exocytosed. Another possibility is that CAPS is important for stabilizing the phospholipid bilayers of secretory vesicles, although in such a case a change in the ultrastructure of chromaffin cells would have been expected. A third possibility is that CAPS functions in the trafficking of secretory vesicles that have undergone endocytosis and are being recycled for exocytosis. For example, empty chromaffin granules may accumulate if vesicles are not properly prepared for refilling after endocytosis. Solving CAPS function will remain a major challenge that is certain to generate many additional surprises and important new insights. Thomas C. Su¨dhof Center for Basic Neuroscience Department of Molecular Genetics and Howard Hughes Medical Institute UT Southwestern Medical Center 6000 Harry Hines Boulevard NA4.118 Dallas, Texas 75390

Selected Reading Elhamdani, A., Martin, T.F., Kowalchyk, J.A., and Artalejo, C.R. (1999). J. Neurosci. 19, 7375–7383.

Grishanin, R.N., Kowalchyk, J.A., Klenchin, V.A., Ann, K., Earles, C.A., Chapman, E.R., Gerona, R.R., and Martin, T.F. (2004). Neuron 43, 551–562. Livingstone, D. (1991). Studies on the UNC-31 gene of Caenorhabditis elegans. PhD thesis, University of Cambridge, Cambridge, United Kingdom. Renden, R., Berwin, B., Davis, W., Ann, K., Chin, C., Kreber, R., Ganetzky, B., Martin, T.F., and Broadie, K. (2001). Neuron 31, 421– 437. Rupnik, M., Kreft, M., Sikdar, S.K., Grilc, S., Romih, R., Zupancic, G., Martin, T.F., and Zorec, R. (2000). Proc. Natl. Acad. Sci. USA 97, 5627–5632. Speidel, D., Varoqueaux, F., Enk, C., Nojiri, M., Grishanin, R.N., Martin, T.F., Hofmann, K., Brose, N., and Reim, K. (2003). J. Biol. Chem. 278, 52802–52809. Speidel, D., Bruederle, C.E., Enk, C., Voets, T., Varoqueaux, F., Reim, K., Becherer, U., Fornai, F., Ruggieri, S., Holighaus, Y., et al. (2005). Neuron 46, this issue, 75–88. Su¨dhof, T.C. (2004). Annu. Rev. Neurosci. 27, 509–547. Tabares, L., Ales, E., Lindau, M., and Alvarez de Toledo, G. (2001). J. Biol. Chem. 276, 39974–39979. Tandon, A., Bannykh, S., Kowalchyk, J.A., Banerjee, A., Martin, T.F., and Balch, W.E. (1998). Neuron 21, 147–154. Walent, J.H., Porter, B.W., and Martin, T.F. (1992). Cell 70, 765–775. DOI 10.1016/j.neuron.2005.03.017