Synapses and Protein Trafficking

Signal Transduction and Mental
Health Research

P. Greengard

The Rockefeller University, New York, NY 10021

A great deal of evidence shows that protein phosphorylation systems regulate a wide variety of physiological phenomena within the nervous system. Thus, it is now clear that many extracellular signals or first messengers, which in the nervous system can be neurotransmitters, hormones, or the nerve impulse itself, produce a wide variety of physiological responses through the sequence of steps depicted in figure 1.

EXTRACELLULAR SIGNALS

1st Messengers: Neurotransmitters, hormones, nerve impulses

INTRACELLULAR SIGNALS

2nd Messengers: Cyclic AMP, cyclic GMP, calcium, diacylglycerol

We now know of several types of intracellular signal or second messenger used by nerve cells. These include cyclic AMP, cyclic GMP, calcium, diacylglycerol, and inositol triphosphate. Four of these second messengers – cyclic AMP, cyclic GMP, calcium, and diacylglycerol-achieve many of their effects through activation of one or another subclass of protein kinase. These protein kinases include two subclasses of cyclic nucleotide-dependent protein kinases, namely, cyclic AMP-dependent protein kinase and cyclic GMPdependent protein kinase, as well as two subclasses of calcium-dependent protein kinase, namely, calcium/calmodulin-dependent protein kinase and calcium/diacylglycerol-dependent protein kinase. These protein kinases, in turn, achieve their effects through phosphorylation of substrate proteins. These substrate proteins, in turn, through one or more biochemical steps, produce the physiological response characteristic of the hormone or neurotransmitter and nerve cell type under investigation.

Studies carried out by Ivar Walaas and Angus Nairn in our laboratory clearly show that a large number of neuron-specific substrate proteins are within the mammalian nervous system (Walaas et al. 1983a, b). These substrate proteins can be divided into two general categories. The first category includes substrate proteins present in every nerve cell throughout the nervous system. Such substrate proteins appear to be involved in regulation of physiological processes common to all nerve cells. For example, synapsin I, a phosphoprotein present on virtually all small synaptic vesicles in all nerve terminals, appears to be involved in regulating a physiological process common to all nerve cells, namely, neurotransmitter release.

The second category includes substrate proteins that are present in specific types of nerve cells, where they appear to be involved in mediating physiological responses specific to that type of nerve cell. For example, DARPP-32 is a phosphoprotein localized to dopaminoceptive cells, that is to say, cells that are innervated by dopaminergic neurons. More specifically, DARPP-32 is localized to dopaminoceptive cells that contain the D-1 subclass of dopamine receptor.

Studies of Basal Ganglion Phosphoproteins

Since aberrations in dopaminergic transmission are generally thought to be involved in the etiology of schizophrenia, it is important to understand the mechanisms by which dopaminoceptive cells respond to dopamine. The D-1 subclass of dopaminoceptive cells has been found to contain approximately 10 phosphoproteins specific to that cell type. One of these phosphoproteins, DARPP-32, has been thoroughly studied. As indicated in figure 2, dopamine, acting on D-1 dopamine receptors on medium spiny neurons, the major nerve cell type in the caudate putamen, activates adenylate cyclase, leading to an

increase in cyclic AMP. The elevated cyclic AMP activates the enzyme cyclic AMP-dependent protein kinase, which then phosphorylates the substrate DARPP-32, converting it from an inactive to an active form. The active form of DARPP-32 is an extremely potent phosphatase inhibitor. More specifically, phospho-DARPP-32 inhibits an enzyme known as protein phosphatase-1, with a Ki of 109 M. Protein phosphatase-1 is an enzyme with broad substrate specificity that dephosphorylates a large number of substrate proteins. Therefore, inhibition of this phosphatase in the medium spiny neurons should result in an increase in the state of phosphorylation of a large number of substrate proteins in those nerve cells. Evidence reviewed elsewhere suggests that the effect of DARPP-32 on protein phosphatase-1 may be involved in the mechanisms by which dopamine regulates the glutamate excitability of medium spiny neurons (Nestler et al. 1984).

Another phosphoprotein localized to medium spiny neurons, and therefore of potential interest in the study of schizophrenia, is ARPP-16. ARPP-16, and a closely related phosphoprotein referred to as ARPP-19, have both been cloned and sequenced (Horiuchi et al. 1988 and unpublished results). In contrast to ARPP-16, which appears to be restricted in its distribution to D-1 dopaminoceptive neurons, ARPP-19 is widespread, being present in almost every area of the nervous system and in many nonneuronal tissues. It is of

particular interest, therefore, that ARPP-16 and ARPP-19 manifest a high degree of structural homology. ARPP-16 is 96 amino acids long, and ARPP-19 is 112 amino acids long, based on the deduced amino acid sequences obtained from the experimentally determined complementary DNA sequences of these two phosphoproteins. The N-terminal 16 amino acids of ARPP-19 are not shared by ARPP-16. However, amino acids 17 through 112 of ARPP-19 show 100-percent homology with amino acids 1 through 96 of ARPP-16, i.e., with the entire ARPP-16 molecule. The 5' flanking regions of ARPP-19 and ARPP-16 show virtually no homology. In contrast, the 3′ flanking regions of ARPP-19 and ARPP-16 immediately adjacent to the coding region show a high degree of homology. These results strongly suggest that ARPP-19 and ARPP-16 are made by alternate splicing of the same primary transcript. Therefore, the specific localization of ARPP-16 to D-1 dopaminoceptive neurons may be a specific splicing mechanism characteristic of ARPP-16 in these neurons. The specificity of location of various phosphoproteins to medium spiny neurons, and the elucidation of the enzymological basis for this specificity, as exemplified in the case of ARPP-16, provide strong encouragement concerning the possibilities of developing powerful and selective new antipsychotic drugs. Thus, it should be possible to develop drugs specifically directed toward regulating the phosphorylation of these substrate proteins of the basal ganglia. Another approach would be to develop drugs specifically directed toward regulating the synthesis of these phosphoproteins.

Speaking in more general terms, the demonstration of a wide variety of nerve cell-type specific phosphoproteins raises the possibility of a new generation of drugs directed at nerve cell-type specific molecules. The development of drugs that act in this way should, in principle, lead to improved therapeutic agents because of the increased selectivity of action of such agents.

Studies of Cystic Fibrosis

Another approach to understanding the etiology of mental illness and to developing improved therapeutic agents for the treatment of mental illness can be illustrated by recent studies of the biological basis of cystic fibrosis. Michael Welsh and his colleagues, in collaboration with Angus Nairn and myself, have gained an improved understanding, at the molecular level, of this disease (Li et al. 1988). Similar results were obtained independently by Schoumacher et al. (1987). Cystic fibrosis is the leading fatal genetic disease in Western Europe and North America. More than 90 percent of children who have this disease die of respiratory infections. It now seems clear, based on studies in several laboratories (see McPherson and Dormer 1988 for review), that the secretion of chloride ions from the airway epithelial cells is impaired in cystic fibrosis.

In normal individuals, upon activation of ß-receptors in the serosal membrane of the airway epithelial cells, chloride ions are released from these cells through the apical membrane into the lumen. The movement of chloride ions into the lumen is followed by movement of sodium ions and water. The water causes a liquification of the mucus, which is then removed from the airways by cilia on the epithelial cells. In cystic fibrosis, the secretion of chloride ions in response to activation of the ß-receptors on the epithelial cells is defective, leading to a diminished release of sodium ions and water. The mucus remains viscous, the cilia are unable to remove the mucus, and the children develop the respiratory infections that account for a large percentage of the fatalities in this disease.

In experiments using excised patches of apical membrane from airway epithelial cells of healthy individuals, the application of ATP plus the catalytic subunit of cyclic AMP-dependent protein kinase caused an opening of chloride channels. In contrast, excised patches from the apical membranes of children with cystic fibrosis did not respond to ATP plus the catalytic subunit of cyclic AMP-dependent protein kinase with an opening of chloride channels. Since purified cyclic AMP-dependent protein kinase prepared from bovine brain was used in both series of experiments, our results cannot be explained by a deficiency of the catalytic subunit of cyclic AMP-dependent protein kinase in cystic fibrosis. Rather, the results pinpoint the defect either to the chloride channel itself or to a modulator of the chloride channel present in the apical membrane. Since a good deal of evidence in the literature shows that cystic fibrosis represents a point mutation, our results suggest that the defect in cystic fibrosis is either an inability of the mutated chloride channel (or its modulator) to be phosphorylated or a defect in the ability of the phosphorylated chloride channel to undergo the conformational change necessary for the passage of chloride ions. To distinguish between these two possibilities, we plan to use a molecular biological approach. More specifically, we plan to clone this chloride channel from normal individuals to determine the amino acid sequence of the normal chloride channel, to prepare oligonucleotide probes based on this information about the structure of the normal chloride channel, with the help of these probes to clone chloride channels from children with cystic fibrosis, and finally, to compare the deduced amino acid sequences. In this way, it should be possible to determine precisely the molecular basis for this disease.

Conclusion

In view of the potential importance of the phosphoproteins of the medium spiny neurons and of other nerve cell-specific phosphoproteins in the etiology of mental diseases, it is of considerable importance to clone and sequence