these various phosphoproteins from the brains of normal individuals. Oligonucleotide probes based upon our knowledge of the sequences of these phosphoproteins could then be used to clone, sequence, and study the corresponding phosphoproteins in individuals with various mental illnesses. I have described our studies with cystic fibrosis in some detail because, by analogy, one or more psychiatric diseases could be associated with defects in ion channels or neurotransmitter receptors. From the viewpoint of mental health research, it is therefore of great importance to use molecular biological approaches to determine the structures of all known ion channels and neurotransmitter receptors in the normal nervous system and to use oligonucleotide probes based upon this knowledge of these channels and receptors to analyze their counterparts in various psychiatric diseases. Acknowledgements The work described in this article was supported by grants MH39327 and MH40899 from the National Institute of Mental Health. REFERENCES Horiuchi, A.; Kurihara, T.; Horiuchi, K.; Williams, K.R.; Nairn, A.C.; and Greengard, P. (untitled) Society for Neuroscience Abstracts. Submitted. Li, M.; McCann, J.D.; Liedtke, C.M.; Nairn, A.C.; Greengard, P.; and Welsh, M.J. Cyclic AMP-dependent protein kinase opens chloride channels in normal but not cystic fibrosis airway epithelium. Nature 331(6154):358-360, 1988. McPherson, M.A., and Dormer, R.L. Cystic fibrosis: A defect in stimulus-response coupling. Trends in Biochemical Sciences 13(1):10-13, 1988. Nestler, E.J.; Walaas, S.I.; and Greengard, P. Neuronal phosphoproteins: Physiological and clinical implications. Science 225(4668):1357-1364, 1984. Schoumacher, R.A.; Shoemaker, R.L.; Halm, D.R.; Tallant, E.A.; Wallace, R.W.; and Frizzell, R.A. Phosphorylation fails to activate chloride channels from cystic fibrosis airway cells. Nature 330(6150):752-754, 1987. Walaas, S.I.; Nairn, A.C.; and Greengard, P. Regional distribution of calcium- and cyclic adenosine 3':5'-monophosphate-regulated protein phosphorylation systems in mammalian brain. I. Particulate systems. Journal of Neuroscience 3(2):291-301, 1983a. Walaas, S.I.; Nairn, A.C.; and Greengard, P. Regional distribution of calcium- and cyclic adenosine 3':5'-monophosphate-regulated protein phosphorylation systems in mammalian brain. I. Soluble systems. Journal of Neuroscience 3(2):302311, 1983b. Peptides and R.B. Kelly, K. Buckley, E. Grote, A. Lowe, and L. Matsuuchi Department of Biochemistry and Biophysics Nerve terminals have synaptic vesicles, which contain neurotransmitter and dense core vesicles, which contain neuropeptides and neurotransmitters. Comparing the similarities and differences between the two vesicle types gives us clues to how they function and to how they are related to each other. Sorting Into Dense Core Secretory Vesicles Electron micrographs of neuronal, endocrine, and exocrine cells show characteristic spherical vesicles whose contents are electron dense. The electron density is associated with a very high protein content of neuropeptide, peptide hormone, or zymogen. The formation of the electron dense core can begin as early as the Golgi cisternae and becomes more pronounced in the trans Golgi network (TGN) and in immature secretory granules. Membrane proteins do not appear to reach the surface via the dense core secretory vesicle (DCSV) in either biochemical (Gumbiner and Kelly 1982) or immunoelectron microscopic experiments (Orci et al. 1987). Furthermore, some of the proteins secreted by the cell do not appear to exit via the DCSV (Moore et al. 1983; Burgess et al. 1985). The route they take is indistinguishable, so far, from the constitutive route taken by membrane proteins. Sorting of proteins destined for the constitutive pathway from proteins entering the DCSVS appears to occur at the TGN (Tooze et al. 1987). Cells can be made to secrete foreign proteins if the DNA encoding them is introduced either by DNA transfection or by generating a transgenic animal. When this is done, secreted proteins fall into two classes, those that enter DCSVs and those that do not. In our experiments with AtT-20 cells, proinsulin and human growth hormone fall in the first category, immunoglobulin chains in the second. The information that specifies sorting of proteins into secretory granules appears to be conserved during evolution and is decipherable by neuronal, endocrine, and exocrine cells. For example, when we transfect human growth hormone into the pheochromocytoma cell line PC12, it is released on stimulation with carbachol simultaneously with norepinephrine while the mouse kappa immunoglobulin light chain exits constitutively (Schweitzer and Kelly 1985; Lowe et al. 1988). We suggest that any neuron with DCSVs will package into them any neuropeptide, peptide hormone, or exocrine secretory product it expresses. Transport of DCSVS to Growing Tips The pheochromocytoma cell line PC12 can be induced to send out processes by the presence of NGF. When this is done, proinsulin-containing vesicles accumulate at the tips in transfected cell lines (Matsuuchi et al. 1988). Process extension in culture is not solely a property of neuronal cells; the endocrine cell line AtT-20, derived from a corticotroph precursor, can also extend processes and fill those processes with ACTH-containing DCSVs (Kelly et al. 1981; Tooze et al. 1987; Matsuuchi et al. 1988). Accumulation of DCSVS at the tips does not imply a specific transport mechanism. Constitutive secretory vesicles and even lysosomes will also accumulate there (Matsuuchi et al. 1988). It is likely that the organelles not tightly anchored in the cell body are sent down to the tips of process-forming cells and accumulate at the plus tips of microtubules (Tooze and Burke 1987; Matsuuchi et al. 1988). Conserved Properties of Synaptic Vesicles A synaptic vesicle in a nerve terminal should have membrane proteins specific for that class of nerve terminal, such as neurotransmitter packaging proteins. In addition, since all synaptic vesicles undergo exocytotic release and membrane recycling, some membrane proteins might be expected to be present in all types of synaptic vesicles. Evidence for the presence of some protein components in many, if not all, vesicle types came from early studies in which antibodies against electric fish synaptic vesicles cross-reacted with mammalian brain (Hooper et al. 1980). The cross-reactivity implied that the proteins were highly conserved during evolution. These early indications of strong conservation have been amply confirmed by the generation of monoclonal antibodies to the universal components (Matthew et al. 1981; Buckley and Kelly 1985; Wiedenmann and Franke 1985) and, more recently, by comparing cDNA sequences from different species (Sudhof et al. 1987). Antibodies to synaptic vesicles have been used effectively to study the maturation of synapses. Less successful has been the attempt to identify function. Expected are proteins that are involved in docking with the nerve terminal plasma membrane, in giving calcium-dependent fusion, and in clustering vesicles around the release site by interaction with cytoskeletal elements. Traffic of Vesicle Proteins While it would be very satisfying to know the functions of the integral vesicle proteins, the lack of such information does not hamper using these proteins to study the origin and life cycle of the synaptic vesicle. For example, the Golgi regions of immature neurons appear to be very rich in synaptic vesicle antigen while, conversely, the cell bodies of mature neurons are almost devoid of antigenicity (Buckley et al. 1985; Navone et al. 1986). Two possible explanations suggest themselves. Synaptic contact might switch off the biogensis of synaptic vesicle proteins, or it might induce their redistribution from cell body to periphery. A second issue that is now addressable is the endocytosis of synaptic vesicle components. Organelles carrying synaptic vesicle membrane proteins can be immunoadsorbed using antibodies to their cytoplasmic domains. We have asked whether the vesicle protein p38 is present in early endosomes, labeled using the receptor-mediated uptake of 1251-transferrin. We find that, of the transferrin internalized during 5 minutes at 37°C, up to 80 percent can be immunoadsorbed by anti-p38 attached to a magnetic bead. If the 1251-transferrin was bound but no internalization was allowed, the amount adsorbed was reduced twentyfold. During prolonged incubation at 37°C, the 1251-transferrin chases out of early endosome into the culture medium. The immunoadsorbed fraction of label decreases with the same kinetics. A marker of early and late endosomes, VLDL, chases out more slowly (figure 1). We conclude that some of the synaptic vesicle membrane proteins are internalized by the conventional early endosome pathway of PC-12 cells. We do not know if this is the only route of internalization, or if there is another one specific to regulated secretory cells. Presence of Synaptic Vesicle Antigens in DCSVS Recycling of synaptic vesicles at the nerve terminal shows that synaptic vesicles can be generated by endocytosis. In functional terms, it does not matter how synaptic vesicle proteins get to the nerve terminal plasma membrane as long as the endocytotic machinery is there to generate synaptic |