- Email: [email protected]
Are Presynaptic Proteins Predisposed to Self-Assemble?
Neuron
Previews primary (as opposed to secondary or acquired) channelopathies as causes of pain. Thus at least a few pain disorders can be considered to be ‘‘channelopathic.’’ These disorders provide us with important model diseases in humans. The borders remain blurry, however, since ion channel genes may contain polymorphisms such as the R1140W NaV1.7 substitution that are present in control populations but render pain-signaling neurons hyperexcitable, lowering pain threshold and possibly enhancing the effect of environmental or epigenetic changes. Irrespective of these nosologic considerations, a growing list of channelopathies is helping us to make a translational leap in which we are beginning to unravel, molecule by molecule, the drivers of human pain. The list is still small, but each addition points toward a potential therapeutic target. Ultimately, this molec-
ular dissection of human pain may enable us to mute ‘‘God’s megaphone.’’
Herzog, R.I., Cummins, T.R., and Waxman, S.G. (2001). J. Neurophysiol. 86, 1351–1364.
REFERENCES
Holland, K.D., Kearney, J., Glauser, T., Buck, G., Keddache, M., Blankston, J., Glaaser, I., Kass, R.S., and Meisler, M.H. (2008). Neurosci. Lett. 433, 65–70.
Black, J.A., Nikolajsen, L., Kroner, K., Jensen, T.S., and Waxman, S.G. (2008). Ann. Neurol. 64, 644–653. Cox, J.J., Reimann, F., Nicholas, A.K., Thornton, G., Roberts, E., Springell, K., Karbani, G., Jafri, H., Mannan, J., Raashid, Y., et al. (2006). Nature 444, 894–898. Cummins, T.R., Dib-Hajj, S.D., and Waxman, S.G. (2004). J. Neurosci. 24, 8232–8236. Dib-Hajj, S.D., Cummins, T.R., Black, J.A., and Waxman, S.G. (2010). Annu. Rev. Neurosci., in press. Published online April 1, 2010. Estacion, M., Harty, T.P., Choi, J.S., Tyrrell, L., DibHajj, S.D., and Waxman, S.G. (2009). Ann. Neurol. 66, 862–866. Estacion, M., Gasser, A., Dib-Hajj, S.D., and Waxman, S.G. (2010). Exp. Neurol., in press. Published online April 24, 2010. 10.1016/j.expneurol.2010. 04.012.
Kremeyer, B., Lopera, F., Cox, J.J., Momin, A., Rugiero, F., Marsh, S., Woods, C.G., Jones, N.G., Paterson, K.J., Fricker, F.R., et al. (2010). Neuron 66, this issue, 671–680. Reimann, F., Cox, J.J., Belfer, I., Diatchenko, L., Zaykin, D., McHale, D., Drenth, J.P., Dai, F., Wheeler, J., Sander, F., et al. (2010). Proc. Natl. Acad. Sci. USA 107, 5148–5153. Renganathan, M., Cummins, T.R., and Waxman, S.G. (2001). J. Neurophysiol. 86, 629–640. Rush, A.M., Dib-Hajj, S.D., Liu, S., Cummins, T.R., Black, J.A., and Waxman, S.G. (2006). Proc. Natl. Acad. Sci. USA 103, 8245–8250. Waxman, S.G. (2006). Nat. Rev. Neurosci. 5, 932–942. Yang, Y., Wang, Y., Li, S., Xu, Z., Li, H., Ma, L., Fan, J., Bu, D., Liu, B., Fan, Z., et al. (2004). J. Med. Genet. 41, 171–174.
Are Presynaptic Proteins Predisposed to Self-Assemble? Ann Y.N. Goldstein1 and Thomas L. Schwarz1,* 1F.M. Kirby Neurobiology Center of Children’s Hospital, Boston, and Department of Neurobiology, Harvard Medical School, CLSB 12130, Boston, MA 02115, USA *Correspondence: [email protected] DOI 10.1016/j.neuron.2010.05.021
Tight control of synapse formation ensures that neurons connect to appropriate targets. In this issue of Neuron, Klassen et al. identify ARL-8 GTPase as a regulator of presynaptic assembly. Without ARL-8, presynaptic material aggregates en route to its destination, suggesting that ARL-8 acts like a dispersant to prevent premature synaptic assembly in the axon. Much attention has been paid to signals that initiate synaptogenesis. Contact between an axon and its proper target causes the postsynaptic membrane to accumulate receptors and scaffolding and causes the presynaptic varicosity to acquire an active zone and vesicle cluster (Jin and Garner, 2008; Owald and Sigrist, 2009). But focusing on signals for synapse building can overlook two equally important aspects of synaptogenesis: the infrastructure of the neuron that delivers the
building materials to their site of assembly and the negative control mechanisms that prevent synapses from assembling where they should not. In this issue of Neuron, Klassen et al. (2010) describe a C. elegans mutant that highlights these aspects of synaptogenesis and points to a mechanism for restricting presynaptic specializations to their proper positions. Presynaptic proteins are synthesized in the soma and transported along axons by specialized motors. The components
624 Neuron 66, June 10, 2010 ª2010 Elsevier Inc.
travel in at least two classes of transport vesicle. One contains the components of synaptic vesicles and a second, often called a piccolo/bassoon transport vesicle, contains components of the active zone (Jin and Garner, 2008; Owald and Sigrist, 2009). These components are delivered principally by the kinesin-3 motors, which are distinct from those that support axon outgrowth and pathfinding (Pack-Chung et al., 2007). Disruption of this transport can prevent synapse
Neuron
Previews formation much as transportamodel: the packets are intrinsition strikes halt building concally prone to aggregation and struction. The kinesins must self-assembly into a presyselect proper cargos, distinnapse and must be inhibited guish axons from dendrites, from doing so by an active and, through a process largely mechanism until targetunknown, halt and disengage induced signals are encounfrom their cargos at the right tered. spot. Cargo delivery is more In an elegant forward than a straightforward cruise genetic screen for disruptions to the end of the road. Axons in synapse formation in typically have many varicosiC. elegans, Klassen and colties in series and motors may leagues uncovered a mutation need to move past one in the small GTPase ARL-8 nascent synapse to deliver (Klassen et al., 2010). In the components down the line. absence of ARL-8, presynThe metaphor of ‘‘building aptic specializations didn’t a synapse’’ may be misleading arise properly in the distal if it invokes the image of an axon, but instead were shifted active process like the nailing proximally. This was eviof beams. The presynapse denced by the mislocalization might better be viewed as the of multiple synaptic vesicle self-assembly of proteins and markers, including rab3::GFP vesicles that have pre-existing and active zone components, affinities to bind one another. in all neuron classes examActive zones comprise proined. While there are multiple teins, including Ca2+ channels Figure 1. Aggregates of Presynaptic Components ways to interpret presynaptic At synapses, active zone proteins and synaptic vesicles are linked to one and structural proteins (Fig- another by filamentous connections. Synaptic precursors travel in similar proteins accumulating at inapure 1), that bind to one another assemblies and in arl-8 mutants, these packets aggregate prematurely and propriate locations, the arl-8 in regular arrays. Synaptic form presynapse-like assemblies without postsynaptic targets. phenotype pointed to a necesvesicles are tethered to active sary mechanism for enabling zones via cytomatrix proteins, synaptic cargoes to reach including RIM-1/UNC-10 and ELKS/ culture of their proper targets synapse their proper destination. The misplaced CAST/bruchpilot and, in mammals, onto one another, which never occurs accumulations were not due to defects piccolo and bassoon (Jin and Garner, in vivo. Drosophila motoneuron endings in axon guidance, growth, or neuronal 2008; Owald and Sigrist, 2009). Behind possess active zones and vesicle clusters polarity. Neither were the accumulations this first shell of vesicles, others are linked even when all muscle development has due to transport defects; unlike the axonal by filaments visible in EM. Some of these been genetically prevented (Jin and traffic jams seen with mutations in confilaments are likely to be synapsin; how- Garner, 2008; Owald and Sigrist, 2009). ventional kinesin (Duncan and Goldstein, ever, additional vesicle-associated pro- The predilection of synaptic components 2006), transport of other axonal material, teins likely serve this function, as some to bind one another is also evident during such as mitochondria, was unaffected in filaments and clustering persists in synap- transport: presynapse components travel arl-8. Ultrastructurally, the aggregates sin-1, -2, and -3 triple-knockout mice as armadas of synaptic vesicle precursors were surprisingly presynapse-like, with (Siksou et al., 2007). alongside vesicles with active zone a denser plasma membrane and a cloud If the proteins that form these structures proteins (Owald and Sigrist, 2009). It isn’t of vesicles. Although vesicle fusion and bind one another at synapses and create known why the components travel as recycling at these loci was not examined, a meshwork of vesicles, the proper ques- packets (Figure 1); plausibly, they are inter- the misplaced structures conceivably tion may not be ‘‘why do they assemble connected by the same filaments and inter- might function like proper presynapses. at synapses,’’ but ‘‘how do they avoid actions that link them at a mature synapse. Most, however, lacked any obvious postbinding one another and aggregating else- Indeed, cytomatrix proteins such as bas- synaptic partner. In this, they differ from where’’? Indeed, synapses have a strong soon, piccolo, and synapsin are present worms lacking the receptor-driven antipredilection to form even when proper in these transport packets. The association synaptogenic signals (Figure 2A) of Wnt/ signals for synaptogenesis are absent. of components in transit may promote fast frizzled and netrin/Unc-5. In the absence Examples include individual hippocampal synapse assembly upon target contact. of these spatial cues, full synapses neurons grown in isolation. Lacking proper At sites of synaptogenesis, are these trans- form at inappropriate axonal regions (Lu synaptic partners, the neuron makes ported cargos unpacked and then built into et al., 2009). Thus, arl-8 mutant neurons synapses (autapses) on its own dendrites. a functional synapse? The findings from seem to assemble ectopic presynapses Similarly, retinal ganglion cells deprived in Klassen et al. (2010) suggest a different in the absence of a target-contact signal.
Neuron 66, June 10, 2010 ª2010 Elsevier Inc. 625
Neuron
Previews Why does loss of ARL-8 lead Synapse formation is thus to accumulations of presynmore than a response to synaptic material? ARL-8 apparaptogenic signals; it is a ently resists the propensity of balance between those signals these components to selfand antisynaptogenic mechaassemble into aggregates. nisms (Figure 2). Some of Live imaging in arl-8 revealed these mechanisms, like ARLthat presynaptic material 8, act in axons as preventative stalled in transit. When addiagents. Another protein in this tional cargoes encountered class may be serine-arginine the stalled material, they were protein kinase 79D, whose likely to stably aggregate with loss causes aggregates of it rather than move past. The presynaptic densities in delivery of synaptic compoDrosophila axons (Johnson nents to the correct distal et al., 2009; Nieratschker targets was thwarted by this et al., 2009). Other factors aggregation (Figure 1). Overdefine axonal regions as inapexpression of the kinesin-3 propriate for synapses, as motor suppressed the ARL-8 exemplified by Wnt/Frz and phenotype, implying that the netrin/Unc-5 signaling that synaptic components can prevents synaptogenesis in arrive at their proper destinathe proximal region of a C. eletions if the propensity to stall gans neurite (Lu et al., 2009). Perhaps the best-studied en route is overcome. Conrestrictive factors (Figure 2B) sistent with a direct involveare E3 ubiquitin ligases that ment of ARL-8 with the moving Figure 2. Positive and Negative Signals Act Together to Determine limit the growth of synapses packets, ARL-8::YFP colocal- Sites of Synapse Formation ized with rab3::mCherry in (A) In developing neurons, restrictive mechanisms such as ARL-8 prevent pre- or the spread of synaptic moving puncta that presum- synapse assembly in inappropriate regions while target-derived and other components away from their proper sites (Jin and Garner, ably correspond to trans- environmental cues activate synaptogenesis at appropriate sites. (B) Local signals stabilize formed synapses, but their spread to surrounding 2008). Yet other proteins, ported synaptic precursors regions is checked. such as SKR-1, clear out (Figure 1). Overexpression of nascent synapses that are not ARL-8 decreased synaptic Rab3, consistent with the model in which move faster, and stall less frequently (Hof- stabilized (Jin and Garner, 2008) or, like ARL-8 disperses accumulations of syn- mann and Munro, 2006). Klassen and RSY-1, limit the assembly of the aptic material. Removing one copy of colleagues examined lysosome distribu- presynapse at synaptogenic sites (Patel SYD-1 or SYD-2/liprin, proteins that tion in their ARL-8 mutant and saw no and Shen, 2009). Together these inhibitory promote assembly, prevented the Rab3 change in lysosome localization nor did mechanisms sculpt the functional consnarls and shifted the assembling presy- lysosome disruption by other means cause nections that form in the developing napses distally. Thus, the proximally an arl-8-like phenotype. A more mecha- nervous system, holding in check the formed presynapses in arl-8 still depend nistic understanding of how ARL-8 propensity of synaptic components to on the intracellular apparatus needed for prevents presynapse assembly in axons self-assemble and of neurons to promisnormal synapse assembly. may reveal some commonality with these cuously hook up to one another. As in so many other systems, a combination of As a small Arf-like GTPase, ARL-8 has other cell-biological roles. the potential not only to associate with If, as suggested by the arl-8 phenotype, negative and positive signals may give membranes but also to act as a molecular the affinities of presynaptic components the most precise control of when and switch dependent on its nucleotide binding for one another give rise to a predisposi- where a synapse should form. state. Conceivably, this could allow ARL-8 tion for assembly, then the transport to resist presynaptic assembly in the axon packets in which they travel are not like and then to be switched off at the correct train cars filled with detached pieces but REFERENCES sites for synaptogenesis. Mammalian and rather like a spring-loaded umbrella ready Drosophila homologs of ARL-8 associate to open. ARL-8 is the clasp on the Bagshaw, R.D., Callahan, J.W., and Mahuran, D.J. with microtubules (Okai et al., 2004) and umbrella that keeps it from popping (2006). Biochem. Biophys. Res. Commun. 344, 1186–1191. lysosomes (Bagshaw et al., 2006; Hof- open while you are still on the bus or in mann and Munro, 2006). Previous studies the doorway. A propensity for self- Duncan, J.E., and Goldstein, L.S. (2006). PLoS of ARL-8 have not examined neurons, but assembly may enable synapses to form Genet. 2, e124. in heterologous cells overexpressing rapidly once correct contacts release the Hofmann, I., and Munro, S. (2006). J. Cell Sci. 119, ARL-8, lysosomes are more dispersed, clasp. 1494–1503.
626 Neuron 66, June 10, 2010 ª2010 Elsevier Inc.
Neuron
Previews Jin, Y., and Garner, C.C. (2008). Annu. Rev. Cell Dev. Biol. 24, 237–262. Johnson, E.L., 3rd, Fetter, R.D., and Davis, G.W. (2009). PLoS Biol. 7, e1000193. Klassen, M.P., Wu, Y.E., Maeder, C.I., Nakae, I., Cueva, J.G., Lehrman, E.K., Tada, M., GengyoAndo, K., Wang, G.J., Goodman, M., et al. (2010). Neuron 66, this issue, 710–723. Lu, B., Wang, K.H., and Nose, A. (2009). Curr. Opin. Neurobiol. 19, 162–167.
Nieratschker, V., Schubert, A., Jauch, M., Bock, N., Bucher, D., Dippacher, S., Krohne, G., Asan, E., Buchner, S., and Buchner, E. (2009). PLoS Genet. 5, e1000700.
Okai, T., Araki, Y., Tada, M., Tateno, T., Kontani, K., and Katada, T. (2004). J. Cell Sci. 117, 4705–4715.
Owald, D., and Sigrist, S.J. (2009). Curr. Opin. Neurobiol. 19, 311–318.
Pack-Chung, E., Kurshan, P.T., Dickman, D.K., and Schwarz, T.L. (2007). Nat. Neurosci. 10, 980–989. Patel, M.R., and Shen, K. (2009). Science 323, 1500–1503. Siksou, L., Rostaing, P., Lechaire, J.P., Boudier, T., Ohtsuka, T., Fejtova´, A., Kao, H.T., Greengard, P., Gundelfinger, E.D., Triller, A., and Marty, S. (2007). J. Neurosci. 27, 6868–6877.
BRAGging about Mechanisms of Long-Term Depression Stephen M. Fitzjohn1 and Zafar I. Bashir1,* 1MRC Centre for Synaptic Plasticity, Department of Anatomy, University of Bristol, Bristol, BS8 1TD, UK *Correspondence: [email protected] DOI 10.1016/j.neuron.2010.05.027
The mechanisms of long-term depression (LTD) underlie various aspects of normal brain function. Therefore, it is important to understand the signaling that underpins LTD. The study by Scholz et al. in this issue of Neuron describes how BRAG2, mGluRs, and AMPARs come together to produce LTD through AMPAR internalization.
Picture the scene: the London underground—one of the world’s busiest subway systems—at rush hour on a Friday evening. As is probably the case in most similar transport systems, entry into or out of the underground is controlled by electronic gates. At rush hour it takes a LONG time to get the thousands and thousands of commuters through the limited numbers of gates. Wouldn’t it be great if it were possible to increase rapidly the number of gates when needed and reduce the number of gates when not needed? However, one cannot even begin to imagine the mechanical, electrical, and computer engineering and design needed to make this possible. How would the gates be brought into the station lobby and how would they be removed? Where would they be stored? Would the station lobby have to increase and decrease in size to accommodate these changes? What multitude of different control mechanisms would need to be in place to ensure that everything happened in a controlled and regulated manner?
Remarkably, however, the central nervous system deals with a similar problem at synaptic junctions. At most synapses fast chemical transmission is mediated by release of glutamate acting on AMPA receptors. One of the most impressive things about synapses is that their strength can be increased and decreased very rapidly, a property known as synaptic plasticity. These changes can last a long time, if required (LTP, longterm potentiation; LTD, long-term depression). One of the most well-studied mechanisms responsible for synaptic plasticity is alterations in the numbers of AMPA receptors on the receiving neuron. In many ways, this is akin to the problem of the underground—in response to particular demands the synapse increases or decreases the number of AMPA receptors, thus providing almost instantaneously greater or reduced capacity to cope with the demands thrown at the synapse. The mechanisms that control these changes in synaptic strength are turning out to be hugely complex and are subject to a level
of fine tuning that could not have been imagined even a few short years ago. Currently, there is pretty good consensus concerning the processes that initiate or trigger synaptic plasticity, generally a rise in intracellular calcium resulting from activation of particular classes of receptors (e.g., NMDA or mGluRs). We are also fairly confident that insertion or removal of AMPARs at the synapse is one of the key final steps that bring about the change in synaptic strength. However, the details of the precise mechanisms between the initial trigger and the final insertion or removal of AMPARs is still the subject of intense investigation and evidence exists for a variety of different intracellular processes that are likely to be involved in some way. Every so often in the investigation of such mechanisms, an exciting, new, and controversial discovery is put forward, such as in this issue of Neuron, in which Scholz et al. (2010) describe a novel signaling mechanism that controls the removal of synaptic AMPARs, thereby controlling LTD.
Neuron 66, June 10, 2010 ª2010 Elsevier Inc. 627
Previews primary (as opposed to secondary or acquired) channelopathies as causes of pain. Thus at least a few pain disorders can be considered to be ‘‘channelopathic.’’ These disorders provide us with important model diseases in humans. The borders remain blurry, however, since ion channel genes may contain polymorphisms such as the R1140W NaV1.7 substitution that are present in control populations but render pain-signaling neurons hyperexcitable, lowering pain threshold and possibly enhancing the effect of environmental or epigenetic changes. Irrespective of these nosologic considerations, a growing list of channelopathies is helping us to make a translational leap in which we are beginning to unravel, molecule by molecule, the drivers of human pain. The list is still small, but each addition points toward a potential therapeutic target. Ultimately, this molec-
ular dissection of human pain may enable us to mute ‘‘God’s megaphone.’’
Herzog, R.I., Cummins, T.R., and Waxman, S.G. (2001). J. Neurophysiol. 86, 1351–1364.
REFERENCES
Holland, K.D., Kearney, J., Glauser, T., Buck, G., Keddache, M., Blankston, J., Glaaser, I., Kass, R.S., and Meisler, M.H. (2008). Neurosci. Lett. 433, 65–70.
Black, J.A., Nikolajsen, L., Kroner, K., Jensen, T.S., and Waxman, S.G. (2008). Ann. Neurol. 64, 644–653. Cox, J.J., Reimann, F., Nicholas, A.K., Thornton, G., Roberts, E., Springell, K., Karbani, G., Jafri, H., Mannan, J., Raashid, Y., et al. (2006). Nature 444, 894–898. Cummins, T.R., Dib-Hajj, S.D., and Waxman, S.G. (2004). J. Neurosci. 24, 8232–8236. Dib-Hajj, S.D., Cummins, T.R., Black, J.A., and Waxman, S.G. (2010). Annu. Rev. Neurosci., in press. Published online April 1, 2010. Estacion, M., Harty, T.P., Choi, J.S., Tyrrell, L., DibHajj, S.D., and Waxman, S.G. (2009). Ann. Neurol. 66, 862–866. Estacion, M., Gasser, A., Dib-Hajj, S.D., and Waxman, S.G. (2010). Exp. Neurol., in press. Published online April 24, 2010. 10.1016/j.expneurol.2010. 04.012.
Kremeyer, B., Lopera, F., Cox, J.J., Momin, A., Rugiero, F., Marsh, S., Woods, C.G., Jones, N.G., Paterson, K.J., Fricker, F.R., et al. (2010). Neuron 66, this issue, 671–680. Reimann, F., Cox, J.J., Belfer, I., Diatchenko, L., Zaykin, D., McHale, D., Drenth, J.P., Dai, F., Wheeler, J., Sander, F., et al. (2010). Proc. Natl. Acad. Sci. USA 107, 5148–5153. Renganathan, M., Cummins, T.R., and Waxman, S.G. (2001). J. Neurophysiol. 86, 629–640. Rush, A.M., Dib-Hajj, S.D., Liu, S., Cummins, T.R., Black, J.A., and Waxman, S.G. (2006). Proc. Natl. Acad. Sci. USA 103, 8245–8250. Waxman, S.G. (2006). Nat. Rev. Neurosci. 5, 932–942. Yang, Y., Wang, Y., Li, S., Xu, Z., Li, H., Ma, L., Fan, J., Bu, D., Liu, B., Fan, Z., et al. (2004). J. Med. Genet. 41, 171–174.
Are Presynaptic Proteins Predisposed to Self-Assemble? Ann Y.N. Goldstein1 and Thomas L. Schwarz1,* 1F.M. Kirby Neurobiology Center of Children’s Hospital, Boston, and Department of Neurobiology, Harvard Medical School, CLSB 12130, Boston, MA 02115, USA *Correspondence: [email protected] DOI 10.1016/j.neuron.2010.05.021
Tight control of synapse formation ensures that neurons connect to appropriate targets. In this issue of Neuron, Klassen et al. identify ARL-8 GTPase as a regulator of presynaptic assembly. Without ARL-8, presynaptic material aggregates en route to its destination, suggesting that ARL-8 acts like a dispersant to prevent premature synaptic assembly in the axon. Much attention has been paid to signals that initiate synaptogenesis. Contact between an axon and its proper target causes the postsynaptic membrane to accumulate receptors and scaffolding and causes the presynaptic varicosity to acquire an active zone and vesicle cluster (Jin and Garner, 2008; Owald and Sigrist, 2009). But focusing on signals for synapse building can overlook two equally important aspects of synaptogenesis: the infrastructure of the neuron that delivers the
building materials to their site of assembly and the negative control mechanisms that prevent synapses from assembling where they should not. In this issue of Neuron, Klassen et al. (2010) describe a C. elegans mutant that highlights these aspects of synaptogenesis and points to a mechanism for restricting presynaptic specializations to their proper positions. Presynaptic proteins are synthesized in the soma and transported along axons by specialized motors. The components
624 Neuron 66, June 10, 2010 ª2010 Elsevier Inc.
travel in at least two classes of transport vesicle. One contains the components of synaptic vesicles and a second, often called a piccolo/bassoon transport vesicle, contains components of the active zone (Jin and Garner, 2008; Owald and Sigrist, 2009). These components are delivered principally by the kinesin-3 motors, which are distinct from those that support axon outgrowth and pathfinding (Pack-Chung et al., 2007). Disruption of this transport can prevent synapse
Neuron
Previews formation much as transportamodel: the packets are intrinsition strikes halt building concally prone to aggregation and struction. The kinesins must self-assembly into a presyselect proper cargos, distinnapse and must be inhibited guish axons from dendrites, from doing so by an active and, through a process largely mechanism until targetunknown, halt and disengage induced signals are encounfrom their cargos at the right tered. spot. Cargo delivery is more In an elegant forward than a straightforward cruise genetic screen for disruptions to the end of the road. Axons in synapse formation in typically have many varicosiC. elegans, Klassen and colties in series and motors may leagues uncovered a mutation need to move past one in the small GTPase ARL-8 nascent synapse to deliver (Klassen et al., 2010). In the components down the line. absence of ARL-8, presynThe metaphor of ‘‘building aptic specializations didn’t a synapse’’ may be misleading arise properly in the distal if it invokes the image of an axon, but instead were shifted active process like the nailing proximally. This was eviof beams. The presynapse denced by the mislocalization might better be viewed as the of multiple synaptic vesicle self-assembly of proteins and markers, including rab3::GFP vesicles that have pre-existing and active zone components, affinities to bind one another. in all neuron classes examActive zones comprise proined. While there are multiple teins, including Ca2+ channels Figure 1. Aggregates of Presynaptic Components ways to interpret presynaptic At synapses, active zone proteins and synaptic vesicles are linked to one and structural proteins (Fig- another by filamentous connections. Synaptic precursors travel in similar proteins accumulating at inapure 1), that bind to one another assemblies and in arl-8 mutants, these packets aggregate prematurely and propriate locations, the arl-8 in regular arrays. Synaptic form presynapse-like assemblies without postsynaptic targets. phenotype pointed to a necesvesicles are tethered to active sary mechanism for enabling zones via cytomatrix proteins, synaptic cargoes to reach including RIM-1/UNC-10 and ELKS/ culture of their proper targets synapse their proper destination. The misplaced CAST/bruchpilot and, in mammals, onto one another, which never occurs accumulations were not due to defects piccolo and bassoon (Jin and Garner, in vivo. Drosophila motoneuron endings in axon guidance, growth, or neuronal 2008; Owald and Sigrist, 2009). Behind possess active zones and vesicle clusters polarity. Neither were the accumulations this first shell of vesicles, others are linked even when all muscle development has due to transport defects; unlike the axonal by filaments visible in EM. Some of these been genetically prevented (Jin and traffic jams seen with mutations in confilaments are likely to be synapsin; how- Garner, 2008; Owald and Sigrist, 2009). ventional kinesin (Duncan and Goldstein, ever, additional vesicle-associated pro- The predilection of synaptic components 2006), transport of other axonal material, teins likely serve this function, as some to bind one another is also evident during such as mitochondria, was unaffected in filaments and clustering persists in synap- transport: presynapse components travel arl-8. Ultrastructurally, the aggregates sin-1, -2, and -3 triple-knockout mice as armadas of synaptic vesicle precursors were surprisingly presynapse-like, with (Siksou et al., 2007). alongside vesicles with active zone a denser plasma membrane and a cloud If the proteins that form these structures proteins (Owald and Sigrist, 2009). It isn’t of vesicles. Although vesicle fusion and bind one another at synapses and create known why the components travel as recycling at these loci was not examined, a meshwork of vesicles, the proper ques- packets (Figure 1); plausibly, they are inter- the misplaced structures conceivably tion may not be ‘‘why do they assemble connected by the same filaments and inter- might function like proper presynapses. at synapses,’’ but ‘‘how do they avoid actions that link them at a mature synapse. Most, however, lacked any obvious postbinding one another and aggregating else- Indeed, cytomatrix proteins such as bas- synaptic partner. In this, they differ from where’’? Indeed, synapses have a strong soon, piccolo, and synapsin are present worms lacking the receptor-driven antipredilection to form even when proper in these transport packets. The association synaptogenic signals (Figure 2A) of Wnt/ signals for synaptogenesis are absent. of components in transit may promote fast frizzled and netrin/Unc-5. In the absence Examples include individual hippocampal synapse assembly upon target contact. of these spatial cues, full synapses neurons grown in isolation. Lacking proper At sites of synaptogenesis, are these trans- form at inappropriate axonal regions (Lu synaptic partners, the neuron makes ported cargos unpacked and then built into et al., 2009). Thus, arl-8 mutant neurons synapses (autapses) on its own dendrites. a functional synapse? The findings from seem to assemble ectopic presynapses Similarly, retinal ganglion cells deprived in Klassen et al. (2010) suggest a different in the absence of a target-contact signal.
Neuron 66, June 10, 2010 ª2010 Elsevier Inc. 625
Neuron
Previews Why does loss of ARL-8 lead Synapse formation is thus to accumulations of presynmore than a response to synaptic material? ARL-8 apparaptogenic signals; it is a ently resists the propensity of balance between those signals these components to selfand antisynaptogenic mechaassemble into aggregates. nisms (Figure 2). Some of Live imaging in arl-8 revealed these mechanisms, like ARLthat presynaptic material 8, act in axons as preventative stalled in transit. When addiagents. Another protein in this tional cargoes encountered class may be serine-arginine the stalled material, they were protein kinase 79D, whose likely to stably aggregate with loss causes aggregates of it rather than move past. The presynaptic densities in delivery of synaptic compoDrosophila axons (Johnson nents to the correct distal et al., 2009; Nieratschker targets was thwarted by this et al., 2009). Other factors aggregation (Figure 1). Overdefine axonal regions as inapexpression of the kinesin-3 propriate for synapses, as motor suppressed the ARL-8 exemplified by Wnt/Frz and phenotype, implying that the netrin/Unc-5 signaling that synaptic components can prevents synaptogenesis in arrive at their proper destinathe proximal region of a C. eletions if the propensity to stall gans neurite (Lu et al., 2009). Perhaps the best-studied en route is overcome. Conrestrictive factors (Figure 2B) sistent with a direct involveare E3 ubiquitin ligases that ment of ARL-8 with the moving Figure 2. Positive and Negative Signals Act Together to Determine limit the growth of synapses packets, ARL-8::YFP colocal- Sites of Synapse Formation ized with rab3::mCherry in (A) In developing neurons, restrictive mechanisms such as ARL-8 prevent pre- or the spread of synaptic moving puncta that presum- synapse assembly in inappropriate regions while target-derived and other components away from their proper sites (Jin and Garner, ably correspond to trans- environmental cues activate synaptogenesis at appropriate sites. (B) Local signals stabilize formed synapses, but their spread to surrounding 2008). Yet other proteins, ported synaptic precursors regions is checked. such as SKR-1, clear out (Figure 1). Overexpression of nascent synapses that are not ARL-8 decreased synaptic Rab3, consistent with the model in which move faster, and stall less frequently (Hof- stabilized (Jin and Garner, 2008) or, like ARL-8 disperses accumulations of syn- mann and Munro, 2006). Klassen and RSY-1, limit the assembly of the aptic material. Removing one copy of colleagues examined lysosome distribu- presynapse at synaptogenic sites (Patel SYD-1 or SYD-2/liprin, proteins that tion in their ARL-8 mutant and saw no and Shen, 2009). Together these inhibitory promote assembly, prevented the Rab3 change in lysosome localization nor did mechanisms sculpt the functional consnarls and shifted the assembling presy- lysosome disruption by other means cause nections that form in the developing napses distally. Thus, the proximally an arl-8-like phenotype. A more mecha- nervous system, holding in check the formed presynapses in arl-8 still depend nistic understanding of how ARL-8 propensity of synaptic components to on the intracellular apparatus needed for prevents presynapse assembly in axons self-assemble and of neurons to promisnormal synapse assembly. may reveal some commonality with these cuously hook up to one another. As in so many other systems, a combination of As a small Arf-like GTPase, ARL-8 has other cell-biological roles. the potential not only to associate with If, as suggested by the arl-8 phenotype, negative and positive signals may give membranes but also to act as a molecular the affinities of presynaptic components the most precise control of when and switch dependent on its nucleotide binding for one another give rise to a predisposi- where a synapse should form. state. Conceivably, this could allow ARL-8 tion for assembly, then the transport to resist presynaptic assembly in the axon packets in which they travel are not like and then to be switched off at the correct train cars filled with detached pieces but REFERENCES sites for synaptogenesis. Mammalian and rather like a spring-loaded umbrella ready Drosophila homologs of ARL-8 associate to open. ARL-8 is the clasp on the Bagshaw, R.D., Callahan, J.W., and Mahuran, D.J. with microtubules (Okai et al., 2004) and umbrella that keeps it from popping (2006). Biochem. Biophys. Res. Commun. 344, 1186–1191. lysosomes (Bagshaw et al., 2006; Hof- open while you are still on the bus or in mann and Munro, 2006). Previous studies the doorway. A propensity for self- Duncan, J.E., and Goldstein, L.S. (2006). PLoS of ARL-8 have not examined neurons, but assembly may enable synapses to form Genet. 2, e124. in heterologous cells overexpressing rapidly once correct contacts release the Hofmann, I., and Munro, S. (2006). J. Cell Sci. 119, ARL-8, lysosomes are more dispersed, clasp. 1494–1503.
626 Neuron 66, June 10, 2010 ª2010 Elsevier Inc.
Neuron
Previews Jin, Y., and Garner, C.C. (2008). Annu. Rev. Cell Dev. Biol. 24, 237–262. Johnson, E.L., 3rd, Fetter, R.D., and Davis, G.W. (2009). PLoS Biol. 7, e1000193. Klassen, M.P., Wu, Y.E., Maeder, C.I., Nakae, I., Cueva, J.G., Lehrman, E.K., Tada, M., GengyoAndo, K., Wang, G.J., Goodman, M., et al. (2010). Neuron 66, this issue, 710–723. Lu, B., Wang, K.H., and Nose, A. (2009). Curr. Opin. Neurobiol. 19, 162–167.
Nieratschker, V., Schubert, A., Jauch, M., Bock, N., Bucher, D., Dippacher, S., Krohne, G., Asan, E., Buchner, S., and Buchner, E. (2009). PLoS Genet. 5, e1000700.
Okai, T., Araki, Y., Tada, M., Tateno, T., Kontani, K., and Katada, T. (2004). J. Cell Sci. 117, 4705–4715.
Owald, D., and Sigrist, S.J. (2009). Curr. Opin. Neurobiol. 19, 311–318.
Pack-Chung, E., Kurshan, P.T., Dickman, D.K., and Schwarz, T.L. (2007). Nat. Neurosci. 10, 980–989. Patel, M.R., and Shen, K. (2009). Science 323, 1500–1503. Siksou, L., Rostaing, P., Lechaire, J.P., Boudier, T., Ohtsuka, T., Fejtova´, A., Kao, H.T., Greengard, P., Gundelfinger, E.D., Triller, A., and Marty, S. (2007). J. Neurosci. 27, 6868–6877.
BRAGging about Mechanisms of Long-Term Depression Stephen M. Fitzjohn1 and Zafar I. Bashir1,* 1MRC Centre for Synaptic Plasticity, Department of Anatomy, University of Bristol, Bristol, BS8 1TD, UK *Correspondence: [email protected] DOI 10.1016/j.neuron.2010.05.027
The mechanisms of long-term depression (LTD) underlie various aspects of normal brain function. Therefore, it is important to understand the signaling that underpins LTD. The study by Scholz et al. in this issue of Neuron describes how BRAG2, mGluRs, and AMPARs come together to produce LTD through AMPAR internalization.
Picture the scene: the London underground—one of the world’s busiest subway systems—at rush hour on a Friday evening. As is probably the case in most similar transport systems, entry into or out of the underground is controlled by electronic gates. At rush hour it takes a LONG time to get the thousands and thousands of commuters through the limited numbers of gates. Wouldn’t it be great if it were possible to increase rapidly the number of gates when needed and reduce the number of gates when not needed? However, one cannot even begin to imagine the mechanical, electrical, and computer engineering and design needed to make this possible. How would the gates be brought into the station lobby and how would they be removed? Where would they be stored? Would the station lobby have to increase and decrease in size to accommodate these changes? What multitude of different control mechanisms would need to be in place to ensure that everything happened in a controlled and regulated manner?
Remarkably, however, the central nervous system deals with a similar problem at synaptic junctions. At most synapses fast chemical transmission is mediated by release of glutamate acting on AMPA receptors. One of the most impressive things about synapses is that their strength can be increased and decreased very rapidly, a property known as synaptic plasticity. These changes can last a long time, if required (LTP, longterm potentiation; LTD, long-term depression). One of the most well-studied mechanisms responsible for synaptic plasticity is alterations in the numbers of AMPA receptors on the receiving neuron. In many ways, this is akin to the problem of the underground—in response to particular demands the synapse increases or decreases the number of AMPA receptors, thus providing almost instantaneously greater or reduced capacity to cope with the demands thrown at the synapse. The mechanisms that control these changes in synaptic strength are turning out to be hugely complex and are subject to a level
of fine tuning that could not have been imagined even a few short years ago. Currently, there is pretty good consensus concerning the processes that initiate or trigger synaptic plasticity, generally a rise in intracellular calcium resulting from activation of particular classes of receptors (e.g., NMDA or mGluRs). We are also fairly confident that insertion or removal of AMPARs at the synapse is one of the key final steps that bring about the change in synaptic strength. However, the details of the precise mechanisms between the initial trigger and the final insertion or removal of AMPARs is still the subject of intense investigation and evidence exists for a variety of different intracellular processes that are likely to be involved in some way. Every so often in the investigation of such mechanisms, an exciting, new, and controversial discovery is put forward, such as in this issue of Neuron, in which Scholz et al. (2010) describe a novel signaling mechanism that controls the removal of synaptic AMPARs, thereby controlling LTD.
Neuron 66, June 10, 2010 ª2010 Elsevier Inc. 627