Protein kinases talk to lipid phosphatases at the synapse
The ability to police membrane traffic in space and time reaches an especially high level of sophistication at presynaptic nerve terminals, where synaptic vesicles (SVs) containing nonpeptide neurotransmitters are found in a state of flux between the presynaptic plasma membrane and various internal pools. Commonly termed the SV cycle, this movement of SV contents and membrane components at axon terminals comprises three stages, namely exocytosis, endocytosis, and vesicle recycling. At the center of this dynamic picture is the lipid–protein interface, where physical and biochemical interactions between lipids and proteins determine the course of membrane traffic (1). These interactions are, in turn, regulated by a cycle of lipid and protein phosphorylation/dephosphorylation in response to appropriate cues (2). In a recent issue of PNAS, Lee et al. (3) reported findings that shed light on a key aspect of synaptic membrane traffic. By using purified synaptojanin 1, they demonstrated that this lipid phosphatase, implicated in the uncoating of retrieved SVs during clathrin-mediated endocytosis, is subject to regulation by phosphorylation/dephosphorylation like other proteins of the endocytic machinery (4).
Protein phosphorylation plays a central role in the control of the SV cycle. Synapsin, whose phosphorylation promotes the release SVs from the reserve pool in stimulated nerve terminals, is arguably the best-characterized phosphoprotein in this regard (5). Evidence continues to accumulate for the phosphorylation-dependent regulation of endocytic machinery components as well. Here protein kinases appear to function primarily as negative modulators, providing tonic inhibition of SV endocytosis in resting neurons. This inhibition is reversed by nerve terminal stimulation, which triggers endocytosis through the coordinate Ca2+-dependent dephosphorylation of a heterogeneous group of phosphoproteins known as dephosphins (4). Synaptojanin 1, a polyphosphoinositide phosphatase, belongs to this group.
While phosphoproteins have understandably garnered attention as key regulators of the SV cycle, it is becoming clear that membrane phospholipids also play an active role in this process. Members of the phosphoinositide (PI) family have received increasing consideration in this regard (6). For example, phosphotidylinositol 4,5-bisphosphate [PtdIns(4,5)P 2], the familiar precursor of diacylglycerol and inositol 1,4,5-trisphosphate, participates directly in a number of transport phenomena, including cytoskeletal organization and membrane trafficking (7). At the synapse, it is implicated in regulating the exocytosis of a number of neurotransmitters. In endocytosis, it plays a role in clathrin nucleation and the generation of an actin-based cytoskeletal scaffold. Moreover, the dephosphorylation of PtdIns(4,5)P 2 is associated with the release of endocytosed SVs from these interactions.
Synaptojanin 1 is subject to regulation by phosphorylation like other proteins of the endocytic machinery.
Synaptojanin 1 is one of several lipid phosphatases involved in PI metabolism. Concentrated in axon terminals, it dephosphorylates the inositol ring of various PIs, including PtdIns(4,5)P 2, and has been implicated in the uncoating of endocytosed SVs. Like other dephosphins, synaptojanin 1 is constitutively phosphorylated under resting conditions. Lee et al. (3) have identified the neuronal kinase Cdk5 (cyclin-dependent kinase 5) as one of the protein kinases responsible for resting levels of synaptojanin 1 phosphorylation. They find that synaptojanin 1 is less active in this phosphorylated form and becomes more active upon dephosphorylation by the Ca2+/calmodulin-dependent protein phosphatase calcineurin (CaN). They extend these in vitro findings to synaptosomes, where depolarization results in increased synaptojanin 1 activity, and the reversal of this CaN-dependent effect requires Cdk5. The authors also report the identity of the major Cdk5 phosphorylation site in the molecule and suggest that the phosphorylation of this residue not only reduces synaptojanin 1 activity but also inhibits the membrane recruitment of the protein during endocytosis, perhaps by controlling physical and regulatory interactions with endophilin 1 and amphiphysin 1. These results are consistent with a model in which nerve terminal stimulation removes the tonic inhibition exerted by protein kinases on synaptojanin 1, thus allowing it to dephosphorylate PtdIns(4,5)P 2 for clathrin coat disassembly. Although direct in vivo evidence for the dependence of vesicle uncoating on the phosphorylation state of synaptojanin 1 is lacking, site-directed mutants and phosphorylation state-specific antibodies should help to resolve this issue, now that at least one important site of phosphorylation in synaptojanin 1 has been identified.
These findings are notable for several reasons. First, they call attention to the emerging role of phospholipids in cell biology, not only as structural elements but also as functional signaling molecules. Second, they bring to light an essential link between protein and lipid metabolism at the membrane. Third, they recap the universal role of Ca2+ signaling in seemingly disparate stages of the SV cycle. Last but not least, they make a landmark contribution to our understanding of Cdk5 as a major player in the presynaptic and postsynaptic regulation of neurotransmission (8). In the presynaptic compartment, Cdk5 has already been suggested to phosphorylate synapsin 1 (5) and voltage-gated Ca2+ channels (9). Furthermore, synaptojanin 1 is clearly not the only dephosphin that is subject to Cdk5-dependent phosphorylation. Dynamin 1 and amphiphysin 1 are similarly Cdk5 substrates (10–12). By the same token, lipidmetabolizing enzymes other than synaptojanin 1 may also be regulated by Cdk5 in this compartment, as there is reason to believe that Cdk5 may coordinately control multiple pathways of PI metabolism, and perhaps different stages of the SV cycle, by talking to lipid kinases as well as lipid phosphatases at the synapse. In addition to these presynaptic roles, Cdk5 has also been implicated in the regulation of several mediators of postsynaptic neurotransmission (8). Future studies may show that neurotransmitter receptor trafficking, which relies on some of the same lipid–protein interactions as SV endocytosis (13), is likewise subject to regulation by Cdk5. Thus, together with other emerging evidence, the results of Lee et al. (3) help to paint a more cohesive picture of the role of Cdk5 at the synapse.
The article also highlights the need for exquisite precision in the timing and localization of discrete events during SV endocytosis (Fig. 1). After neurotransmitter release, endocytosis is initiated by clathrin coat nucleation and membrane invagination. The subsequent budding and fission of clathrin-coated vesicles is followed quickly by vesicle uncoating. The disassembly of the clathrin coat is accomplished in part by the CaN-dependent activation of synaptojanin 1, which degrades PtdIns(4,5)P 2. Because PtdIns(4,5)P 2 is essential for clathrin coat assembly, however, synaptojanin 1 activity must be tightly regulated while the earlier stages of clathrin-mediated endocytosis are in progress. After all, uncontrolled activation of this enzyme could result in the premature loss of important membrane phospholipids. One possible mechanism of control is localization. For example, the membrane recruitment of synaptojanin 1 may depend on the successful completion of coat assembly, directing enzyme activity only to fully retrieved SVs and sparing intermediate forms. A complementary mechanism may involve the active exclusion of Cdk5 from the newly formed clathrin coat, thereby facilitating coat disassembly. A reduction of Cdk5 activity in stimulated nerve terminals would be consistent with such a model.
A simplified schematic diagram of the SV cycle. Sequential steps of SV endocytosis (15) are highlighted. Dephosphins are maintained in phosphorylated form by high Cdk5 and low CaN activity. This relationship is reversed by excitation-dependent Ca2+ influx, thereby triggering endocytosis. dyn, dynamin 1; amph, amphipysin 1; synptjn, synaptojanin 1; P, phosphorylated forms; dP, dephosphorylated forms.
Synaptic transmission in central neurons is a function of the speed with which SVs are endocytosed and recycled. These kinetic features of SV endocytosis may, in turn, emerge from a more thorough understanding of its dynamic elements, like synaptojanin 1. Several questions come to mind in this regard. For example, is the phosphorylation/dephosphorylation of endocytic machinery components an all-or-none molecular switch, or are these biochemical reactions subject to a higher order of modulation that ultimately controls the rate of SV endocytosis and, thereby, the efficiency of synaptic transmission? What is the role of Cdk5 in the phosphorylation of other dephosphins in the presynaptic compartment (12)? What, in turn, regulates Cdk5 activity here, and what implications might this have for synaptic plasticity? Like many discoveries in biology, the findings of Lee et al. (3) seem to open more avenues of inquiry than they close. In the foreseeable future, lipids will undoubtedly continue making headlines in the scientific literature (14), as will Cdk5. Moreover, the emerging partnership between lipid- and protein-metabolizing enzymes in the SV cycle is likely to illuminate some common principles of intracellular traffic and general cell function. In this respect, studies such as the one by Lee et al. can only be a step in the right direction.
