2006: "Big Picture" Insights into the Synaptic Signaling Network
We undertook this study because we had observed that mutants lacking UNC-31 (CAPS) had phenotypes that bore a striking resemblance to mutants lacking a neuronal Gs pathway. The Gs pathway is one of the 3 major pathways in the synaptic signaling network. The pathways in this network work together to produce small molecule signals that are critical for activating synapses for behaviors. Two of the pathways (Gq and Go) regulate the production of diacyclglycerol (DAG) that stimulates neurotransmitter release for transmitting signals between neurons or between neurons and muscle cells. The 3rd pathway, Gs, is less well understood but is also critical for activating synapses for behavior. The Gs pathway produces microdomains of cyclic AMP (cAMP) that determines which synapses are activated. As we showed in Adventures>2000-2005B, it seems to interact with the Gq pathway downstream of DAG, and thus may allow synapses to respond to DAG in a spatially precise way to all organized behavior.
|
|
UNC-31 (known as CAPS in humans and other mammals) had recently been found to function in priming Dense Core Vesicles (DCVs), which meant it played some role in docking the DCVs tightly against the plasma membrane in preparation for secretion of its cargo in response to a calcium signal produced in response to a change in membrane voltage. DCVs often release neuropeptides at synapses in conjunction with the release of small molecule neurotransmitters by Synaptic Vesicles (SVs). However, it is important to remember that UNC-31 (CAPS) appears not to be absolutely necessary for DCV release since null unc-31 mutants can release neuropeptides at rates that are as much as half of wild type. Furthermore, UNC-31(CAPS) is a large protein with several conserved domains, so we can't rule out that it has other functions unrelated to neuropeptide release (such as release of a non-peptidergic signaling protein from DCVs) and perhaps even unrelated to DCV exocytosis. Some studies have found that it also regulates the release of neurotransmitters from SVs, although that has not been demonstrated to be a direct effect.
On culture plates, unc-31 null mutants, like mutants lacking a neuronal Gs pathway, are nearly paralyzed and have a somewhat "straight" posture as opposed to the sinusoidal posture of wild type. egl-30 (Gq) strong reduction-of-function mutants also shared this appearance.
To distinguish whether the near-paralysis and straight posture of mutants lacking UNC-31 was caused by defects in the Gs pathway or the Gq pathway, we assayed overall steady-state levels of neurotransmitter release in living animals by analyzing their sensitivity to the acetylcholinesterase inhibitor aldicarb, which causes secreted acetylcholine (ACh) to build up by preventing the breakdown of ACh after it is released. Since the secreted ACh that accumulates in the presence of aldicarb is toxic, mutations that decrease or increase the steady-state release of ACh confer resistance or hypersensitivity to aldicarb, respectively (Adventures>1993-1996). Mutants with an impaired Gq pathway are strongly resistant to aldicarb, indicated reduced neurotransmitter release. In contrast, mutants lacking an neuronal Gs pathway have an aldicarb sensitivity similar to wild type, indicating that the average motor neuron synapse in such mutants is releasing about wild type levels of ACh. Mutants lacking UNC-31 appear only slightly resistant to low concentrations of aldicarb, but still end up dying on the same concentration of aldicarb as wild type. Thus, based on this assay, unc-31 null mutants resemble mutants with an impaired Gs pathway more than mutants with an impaired Gq pathway. |
Epistasis Analysis
To investigate the relationship of UNC-31 (CAPS) to the Synaptic Signaling Network we once again turned to the genetic technique of epistasis analysis. This powerful method relies on the analysis of strategically-constructed double mutants to infer the upstream/ downstream order of proteins in a signaling pathway or, on a larger scale, the relationships of pathways in a network.
To be successful epistasis analysis requires the right set of mutants. Fortunately, the genetic studies of ourselves and others in C. elegans had produced a remarkable set of genetic mutations and tools, unique among model organisms in their scope, that are well-suited for epistasis studies of the Go, Gq, and Gs synaptic signaling pathways. These included mutations that completely knock out each of these 3 pathways, transgenic strains and native gain-of-function mutants in which each pathway is strongly hyperactivated, and transgenic and pharmacological (drug-based) tools that allow each pathway to be acutely activated in living animals.
To investigate the relationship of UNC-31 (CAPS) to the Synaptic Signaling Network we once again turned to the genetic technique of epistasis analysis. This powerful method relies on the analysis of strategically-constructed double mutants to infer the upstream/ downstream order of proteins in a signaling pathway or, on a larger scale, the relationships of pathways in a network.
To be successful epistasis analysis requires the right set of mutants. Fortunately, the genetic studies of ourselves and others in C. elegans had produced a remarkable set of genetic mutations and tools, unique among model organisms in their scope, that are well-suited for epistasis studies of the Go, Gq, and Gs synaptic signaling pathways. These included mutations that completely knock out each of these 3 pathways, transgenic strains and native gain-of-function mutants in which each pathway is strongly hyperactivated, and transgenic and pharmacological (drug-based) tools that allow each pathway to be acutely activated in living animals.
In the first set of epistasis experiments, we combined a mutation that knocks out UNC-31 with a mutation that knocks out the Gs pathway and compared each single mutant with the double mutant in a quantitative locomotion assay. In all cases, whether the animal was missing UNC-31, the Gs pathway, or both, the locomotion rate was reduced to 3 - 5% of wild type, and the strains were not statistically different from each other. In genetic analysis, this suggests that UNC-31 could function in the same process as the Gs pathway.
In contrast, when we combined the unc-31 null mutation with mutations that reduce the activity of the Gq pathway, those double mutants were much more impaired for locomotion than either single mutant. In fact, those double mutants had locomotion rates that were close to that of a mutant lacking a Gq pathway, which is about 40-fold worse that a mutant lacking a Gs pathway. In genetic analysis, this is known as a "synthetic interaction", and it suggests that UNC-31 and Gq have distinct, largely non-overlapping functions, just as the Gs and Gq pathways have distinct functions. In the second set of epistasis experiments, we combined a mutation that knocks out UNC-31 with mutations that activate either the Gq or the Gs pathways. We had previously shown that when a Gq-activating mutation is crossed into a mutant lacking a neuronal Gs pathway, the near paralysis of the Gs mutant is rescued to about wild type levels, but the animals spend most of their time transitioning in and out of a tightly "knotted" posture. They still appear able to activate their Gq pathway but, in the absence of a Gs pathway, they appear unable to convert that neuronal activity into a sustained, organized behavior. Consistent with unc-31 null mutants having an impaired Gs pathway, we found that crossing a Gq-activating mutation into a mutant lacking UNC-31 also rescued the near-paralysis of the unc-31 null to about wild type levels but, at the same time, caused animals to spend some of their time in a tightly knotted posture. If unc-31 null mutants are unable to activate their Gs pathway, then it might be possible to rescue the paralysis of unc-31 null mutants by crossing in a mutation in which the Gs pathway is constitutively activated. We isolated several such dominant, gain-of-function Gs mutations in Adventures>2000-2005A. When we crossed one of those Gs gain-of-function mutations into unc-31, it restored hyperactive, fully-coordinated locomotion to unc-31 nulls that was nearly as hyperactive as the Gs gain-of-function single mutants. In genetic analysis, this suggests that UNC-31 functions upstream of the Gs pathway and promotes Gs pathway activation. In the absence of UNC-31, the Gs pathway cannot become significantly activated unless you introduce a mutation that causes Gs to become "stuck" in the ON state. In the third set of epistasis experiments, we used tried a different method of activating the Gq pathway in an unc-31 null mutant: we crossed a null mutation of the goa-1 (Go) pathway into an unc-31 null. Since GOA-1 (Go) normally inhibits the Gq pathway, knocking out Go causes too much Gq pathway activity. When we did this, we found that the unc-31 null was again partially rescued for locomotion, although not as much as when the Gq pathway was mutationally activated. However, when we tested neurotransmitter release in the unc-31; goa-1 double null mutant using aldicarb, we found that the double was just as extremely hypersensitive to aldicarb as goa-1 single mutants. This contrasts strongly with the egl-30; goa-1 double mutant, in which the goa-1 null mutation has no effect on a strong egl-30 (Gq) reduction-of-function mutation because GOA-1 (Go) functions upstream of EGL-30 (Gq) (Adventures>1996-1999). The fact that the strong aldicarb hypersensitivity the goa-1 null mutant is unaffected by the unc-31 null mutation suggests that the Gq pathway does not require UNC-31 to stimulate neurotransmitter release, again suggesting that the Gq pathway and the UNC-31/ Gs pathway have functionally distinct roles within the Synaptic Signaling Network. |
|
In the 4th set of epistasis experiments, we compared the effects of activating the Gs pathway in mutants lacking UNC-31 versus mutants with an impaired EGL-30 (Gq) pathway.
We found that unc-31 nulls are strongly rescued by turning on the Gs pathway, up to 53-fold, even by mutations that turn on downstream parts of the Gs pathway. In contrast, the same mutations have much smaller effects on egl-30 (Gq) reduction-of-function mutants, despite the fact that one of the egl-30 reduction of function mutants is less paralyzed than the unc-31 null to start with.
We had previously shown that mutants lacking a neuronal Gs pathway could be rescued as adults by simply inducing the expression of a mutationally-activated Gs protein in adults (Adventures>2000-2005B). In the 5th set of epistasis experiments, we found that we could also acutely rescue the semi-paralysis of unc-31 null mutants in adults by simply turning ON the Gs pathway. This suggests that the activated Gs pathway's ability to rescue unc-31 null mutants is a consequence of functional changes in fully developed neurons rather than altered neuronal development.
Site-of-Action Studies
We showed in Adventures>2000-2005B that the Gs pathway functions in both neurons and muscle cells. Which Gs pathway is responsible for rescuing unc-31's paralysis? To test this, we analyzed strains carrying the unc-31 null mutation in combination with transgenes carrying mutations that activate either the neuronal or muscle Gs pathways. We found that the locomotion rate of the unc-31 null is greatly improved by the native acy-1(ce2) mutation, which activates ACY-1 (adenylyl cyclase) in both neurons and muscle cells. However, nearly the same level of rescue is caused by expressing the same mutation only in the nervous system, and there is no rescue when the same mutation is expressed in muscle cells. We next used cell-specific promoters to investigate where within the nervous system UNC-31 acts. We found that expressing the unc-31 cDNA either in all neurons or just in the ventral cord cholinergic motor neurons restores a near-wild-type locomotion rate to the unc-31 null. To determine if the Gs pathway can function in the same neurons as UNC-31, we expressed a Gs pathway activating mutation in the ventral cord cholinergic motor neurons and crossed this transgene into an unc-31 null background. We found that this transgene alone was sufficient to rescue the locomotion rate of the unc-31 null to a level that is about 50% of wild type. This suggests that UNC-31 and the Gs pathway can function in the same neurons to regulate locomotion rate, and that those neurons prominently include the ventral cord cholinergic motor neurons. |
|
As part of our site-of-action studies, we sought to determine where within neurons the UNC-31 protein is localized. To do this, we expressed a portion of the UNC-31 protein in bacteria, purified this region, and raised antibodies that recognized UNC-31 in fixed, whole mounts of adult worms. We also co-stained with other antibodies that recognize synaptic vesicle clusters and "active zones", which are the small subregions of synapses where synaptic vesicles fuse to release their neurotransmitter cargos.
We found that UNC-31 is sometimes concentrated in subsynaptic regions that overlap precisely with the active zone protein UNC-10 at the theoretical resolution limit of light microscopy. However, UNC-31 localization seemed to vary greatly at synapses. Some synapses were strongly enriched for UNC-31 but had much lower relative levels of UNC-10 (active zones). Others were very enriched for UNC-10, but had only low amounts of UNC-31. Others had less divergent UNC-31/ UNC-10 intensity ratios.
We found that UNC-31 is sometimes concentrated in subsynaptic regions that overlap precisely with the active zone protein UNC-10 at the theoretical resolution limit of light microscopy. However, UNC-31 localization seemed to vary greatly at synapses. Some synapses were strongly enriched for UNC-31 but had much lower relative levels of UNC-10 (active zones). Others were very enriched for UNC-10, but had only low amounts of UNC-31. Others had less divergent UNC-31/ UNC-10 intensity ratios.
Past studies have found that PDE-4 (cAMP phosphodiesterase), which is the component of the Gs pathway that breaks down the small molecule cAMP, can produced microdomains of cAMP at synapses, perhaps to provide positional information connecting neurotransmitter release to organized behavior. Our finding that PDE-4 is highly localized at synapses is consistent with this microdomain hypothesis (Adventures>2004-2006). The finding that UNC-31, which behaves genetically as another Gs pathway component, is also localized in variable amounts to spatially distinct domains, again suggests that the Gs pathway interprets spatial information, such as which synapses to turn ON or OFF, or which synaptic subdomains to locate the active zone. This would explain why the Gs pathway has little affect on overall steady state neurotransmitter release, but major effects on ensuring that organized behavior occurs.
Charlie NK, Schade MA, Thomure AM, and Miller KG. (2006). Presynaptic UNC-31 (CAPS) is required to activate the Gs pathway of the Caenorhabditis elegans synaptic signaling network. Genetics. 172(2): 943-961. PMCID: PMC1456257.
Cited by 51 papers as of April, 2018.
Cited by 51 papers as of April, 2018.
Remaining Important Questions
Perhaps the most pressing remaining question is Does the function of UNC-31 upstream of the Gs pathway have anything to do with neuropeptide release? In live animal studies, UNC-31's role in neuropeptide release remains unclear. One C. elegans study showed that essentially no Dense Core Vesicle docking occurs in unc-31 null mutants. However, it is not clear if docking, as defined by electron microscopy, is essential for neuropeptide secretion. One study reported an approximately 70% reduction in neuropeptide secretion in unc-31 null mutants, while our own studies and one other study find more like a 40 - 50% reduction in neuropeptide secretion in live animal assays.
Genetic analysis suggests that UNC-31's only role cannot be neuropeptide secretion because, by quantitative locomotion assays, unc-31 null mutants are about 6-fold more paralyzed than egl-3 null mutants, which cannot make neuropeptides. Furthermore, that locomotion analysis doesn't take into account the observation that unc-31 null mutants still release about half as much neuropeptide as wild type. Thus, UNC-31 must have at least one other function unrelated to neuropeptide release. UNC-31 could promote the regulated release of a non-neuropeptide signaling protein from dense core vesicles. Alternatively, UNC-31 may mediate the localized insertion of Gs protein-coupled receptors via dense core vesicle fusion with the plasma membrane. However, it also remains possible that UNC-31 has a Gs pathway-related function that is unrelated to dense core vesicles.
Genetic analysis suggests that UNC-31's only role cannot be neuropeptide secretion because, by quantitative locomotion assays, unc-31 null mutants are about 6-fold more paralyzed than egl-3 null mutants, which cannot make neuropeptides. Furthermore, that locomotion analysis doesn't take into account the observation that unc-31 null mutants still release about half as much neuropeptide as wild type. Thus, UNC-31 must have at least one other function unrelated to neuropeptide release. UNC-31 could promote the regulated release of a non-neuropeptide signaling protein from dense core vesicles. Alternatively, UNC-31 may mediate the localized insertion of Gs protein-coupled receptors via dense core vesicle fusion with the plasma membrane. However, it also remains possible that UNC-31 has a Gs pathway-related function that is unrelated to dense core vesicles.
Related Studies
Our studies of the C. elegans Gs pathway in Adventures>2000-2005B and Adventures>2004-2006 established the initial Gq-Gs pathway relationships that formed the basis for the current study.
Reynolds NK, Schade MA, and Miller KG (2005). Convergent, RIC-8 – dependent Ga signaling pathways in the Caenorhabditis elegans synaptic signaling network. Genetics. 169(2), 651-670. PMCID: PMC1449085.
Charlie NK, Thomure AM, Schade MA, and Miller KG (2006). The Dunce cAMP phosphodiesterase PDE-4 negatively regulates Gas – dependent and Gas – independent cAMP pools in the Caenorhabditis Elegans synaptic signaling network. Genetics 173(1), 111-130. PMCID: PMC1461419.
A joint study from the labs of Erik Jorgensen and Tom Martin showed that UNC-31 functions in Dense Core Vesicle, but not Synaptic Vesicle, exocytosis in C. elegans. However, the near blockade of neuropeptide secretion in cultured neurons was not supported by in vivo studies, where neuropeptide secretion still occurred at about 30% of wild type levels. A later study from Joshua Kaplan's lab using a similar in vivo neuropeptide secretion assay found neuropeptide secretion in unc-31 null mutants at 60% of wild type levels.
Speese, S., M. Petrie, K. Schuske, M. Ailion, K. Ann, K. Iwasaki, E.M. Jorgensen, and T.F. Martin. 2007. UNC-31 (CAPS) is required for dense-core vesicle but not synaptic vesicle exocytosis in Caenorhabditis elegans. J. Neurosci. 27:6150–6162. doi:10.1523/JNEUROSCI.1466-07.2007
Sieburth, D., J.M. Madison, and J.M. Kaplan. 2007. PKC-1 regulates secretion of neuropeptides. Nat. Neurosci. 10:49–57. doi:10.1038/nn1810
A beautiful high pressure freezing EM study from the Jorgensen lab showed that UNC-31 is absolutely required for Dense Core Vesicle docking. However, it is not clear if DCV docking, as defined by electron microscopy, is essential for neuropeptide secretion. A recent study found that eliminating the docking of synaptic vesicles does not eliminate neurotransmitter release by synaptic vesicles.
Hammarlund, M., S. Watanabe, K. Schuske, and E.M. Jorgensen. 2008. CAPS and syntaxin dock dense core vesicles to the plasma membrane in neurons. J. Cell Biol. 180:483–491. doi:10.1083/jcb.200708018.
Wang, S. S., R. G. Held, M. Y. Wong, C. Liu, A. Karakhanyan et al., 2016 Fusion Competent Synaptic Vesicles Persist upon Active Zone Disruption and Loss of Vesicle Docking. Neuron 91: 777-79.
At least two studies in mammals have found that CAPS can regulate the release of neurotransmitter from synaptic vesicles, although the directness of that effect has not been carefully addressed.
Jockusch WJ, Speidel D, Sigler A, Sorensen JB, Varoqueaux F, et al. (2007) CAPS-1 and CAPS-2 are essential synaptic vesicle priming proteins. Cell 131: 796–808.
Shinoda, Y., Ishii, C., Fukazawa, Y., Sadakata, T., Ishii, Y., Sano, Y., Iwasato, T., Itohara, S., and Furuichi, T. (2016). CAPS1 stabilizes the state of readily releasable synaptic vesicles to fusion competence at CA3-CA1 synapses in adult hippocampus. Sci Rep 6, 31540.
Reynolds NK, Schade MA, and Miller KG (2005). Convergent, RIC-8 – dependent Ga signaling pathways in the Caenorhabditis elegans synaptic signaling network. Genetics. 169(2), 651-670. PMCID: PMC1449085.
Charlie NK, Thomure AM, Schade MA, and Miller KG (2006). The Dunce cAMP phosphodiesterase PDE-4 negatively regulates Gas – dependent and Gas – independent cAMP pools in the Caenorhabditis Elegans synaptic signaling network. Genetics 173(1), 111-130. PMCID: PMC1461419.
A joint study from the labs of Erik Jorgensen and Tom Martin showed that UNC-31 functions in Dense Core Vesicle, but not Synaptic Vesicle, exocytosis in C. elegans. However, the near blockade of neuropeptide secretion in cultured neurons was not supported by in vivo studies, where neuropeptide secretion still occurred at about 30% of wild type levels. A later study from Joshua Kaplan's lab using a similar in vivo neuropeptide secretion assay found neuropeptide secretion in unc-31 null mutants at 60% of wild type levels.
Speese, S., M. Petrie, K. Schuske, M. Ailion, K. Ann, K. Iwasaki, E.M. Jorgensen, and T.F. Martin. 2007. UNC-31 (CAPS) is required for dense-core vesicle but not synaptic vesicle exocytosis in Caenorhabditis elegans. J. Neurosci. 27:6150–6162. doi:10.1523/JNEUROSCI.1466-07.2007
Sieburth, D., J.M. Madison, and J.M. Kaplan. 2007. PKC-1 regulates secretion of neuropeptides. Nat. Neurosci. 10:49–57. doi:10.1038/nn1810
A beautiful high pressure freezing EM study from the Jorgensen lab showed that UNC-31 is absolutely required for Dense Core Vesicle docking. However, it is not clear if DCV docking, as defined by electron microscopy, is essential for neuropeptide secretion. A recent study found that eliminating the docking of synaptic vesicles does not eliminate neurotransmitter release by synaptic vesicles.
Hammarlund, M., S. Watanabe, K. Schuske, and E.M. Jorgensen. 2008. CAPS and syntaxin dock dense core vesicles to the plasma membrane in neurons. J. Cell Biol. 180:483–491. doi:10.1083/jcb.200708018.
Wang, S. S., R. G. Held, M. Y. Wong, C. Liu, A. Karakhanyan et al., 2016 Fusion Competent Synaptic Vesicles Persist upon Active Zone Disruption and Loss of Vesicle Docking. Neuron 91: 777-79.
At least two studies in mammals have found that CAPS can regulate the release of neurotransmitter from synaptic vesicles, although the directness of that effect has not been carefully addressed.
Jockusch WJ, Speidel D, Sigler A, Sorensen JB, Varoqueaux F, et al. (2007) CAPS-1 and CAPS-2 are essential synaptic vesicle priming proteins. Cell 131: 796–808.
Shinoda, Y., Ishii, C., Fukazawa, Y., Sadakata, T., Ishii, Y., Sano, Y., Iwasato, T., Itohara, S., and Furuichi, T. (2016). CAPS1 stabilizes the state of readily releasable synaptic vesicles to fusion competence at CA3-CA1 synapses in adult hippocampus. Sci Rep 6, 31540.