By Gabriel C.S. Gavin
Most people agree that the brain is a pretty impressive engineering feat; the Aston Martin of biology.
Infinitely more impressive though than a fine British engine is the fact that the brain builds itself from scratch. From a single cell comes a vast number of different tissue types, transmitters and functions. For neural networks to exist, cells that are often far apart have to become intimately connected by means of short-range chemical synapses. This migratory process, often thousands of times the length of the cell, has to be tightly controlled for each and every neuronal projection to avoid ending up as a quivering mass of crossed wires.
To do this, the cell must first elongate its axon, which then is guided by molecules acting as signposts towards a target where, finally, it has to terminate pretty rapidly or risk overshooting and becoming dystrophic. Each of these steps provides its own problems, but understanding the final step in particular is essential not only for characterising the ways in which the developing brain connects up, but also for fuelling research into how it can be repaired after injury.
Connectivity problems have been of interests to researchers in the areas of schizophrenia, depression, autism and many other neuropsychiatric conditions. In spite of this, the signalling pathways that apply the brakes to axonal growth have remained elusive. Recent work by Baker et al. in the nematode worm C. elegans has shone light onto a highly-conserved protein, the so-called Regulator of Presynaptic Morphology (RPM-1), that would seem to link the way in which axons elongate and then terminate.
Using transgenic and proteomic analyses, they demonstrate that RPM-1 and its downstream pathways are essential for ensuring that axon terminals halt at the site of the prospective synapse. Their findings have profound implications for developmental neurobiology, despite highlighting the immense complexity of the system and the vast number of questions that remain unanswered.
C. elegans is one of the simplest organisms that has a nervous system and has been in- tensely studied for the past half-century. Needless to say, there is still a lot about it we don’t know, but it has been an invaluable tool for neuroscientists who want to study the connections between its relatively simple nerve bundles. In order to respond to touch and pressure, two mechanosensory Posterior Lateral Microtubule (PLM) neurons ascend from its tail. Each stretches out an axon to meet up with the cell body of the Anterior Lateral Microtubule (ALM) neuron (Image source: Baker et al., 2014. Modified from Worm Atlas). This basic network allows the nematode to navigate its environment and attempt to avoid predators. Despite how clumsy this single network sounds, its development has to be precise and highly reproducible between individual worms.
Any molecule that could cause the developing neuron to stop growing and form a synapse would likely be highly evolutionarily conserved, would have to respond to environmental cues to trigger intracellular signalling cascades and would have to be able to do so rapidly and absolutely. The Pam/Highwire/RPM-1 (PHR) family of proteins, of which RPM-1 is the nematode homolog, are found along the presynaptic terminals of developing neurons and were flagged as promising candidates for the role.
Scott Baker, Brock Gill and their colleagues at the Scripps Research Institute in Florida set out to identify proteins downstream of RPM-1. Using a transgenic nematode that produces green fluorescent RPM-1, they applied an antibody to precipitate out the protein, which was then purified and examined with mass spectrometry. They found a binding protein, Protein Phosphatase Magnesium/Manganese (PPM-2) attached to it in a range of diﬀerent cells. In order to show that PPM-2 had a role in axon termination, they then studied ppm-2 -/- mutants. On microscopic examination, it became obvious that the PLM axons had missed their targets, overshot and hooked around trying to get back to the target (Image source: Baker et al., 2014.). In other words, the neurons were still attracted towards the postsynaptic terminal, but simply could not stop elongating. The same is seen when RPM-1 is knocked out, but with a higher penetrance, suggesting that PPM-2 is one of a few key players in synapse formation downstream of RPM-1. Rather strangely, these synaptic defects do not seem to cause any real problems with the ability of the nematode to sense touch.
A previously known function of the PHR proteins is their ability to block the Mitogen-activated Protein Kinases (MAPK) signalling that is essential in many aspects of development and cell functioning. Specifically, they cause F-box Synaptic protein 1 (FSN-1) to add ubiquitin groups to MAP3K, which marks it for enzymatic degradation in the proteasome, preventing further elongation. If both PPM-2 and FSN-1 have roles in synapse formation, knocking out both should produce a much higher rate of synapse failure. Indeed, the ppm-2 -/-; fsn-1 -/- double mutant has even more penetrant synaptic defects, with over-elongated hooking axons found in almost all of the mutant nematodes studied.
When PPM-2 was expressed only in the neurons studied by means of a cell-specific promoter, the double mutants were rescued and only a third showed the synaptic defects. This not only demonstrates that PPM-2 and FSN-1 have key roles in terminating the growth of a developing neuron, but that PPM-2 functions as a phosphatase in a cell-autonomous manner to achieve this. Knocking out both MAP3K and PPM-2 lead to double mutants that had no synaptic defects, eliminating both the repressor and the repressed and preventing the system from going unregulated.
Precipitating out the equivalent of MAP3K, they found that PPM-2 was frequently bound to it. Using further transgenic nematodes, they showed that PPM-2 acted on a specific serine residue on MAP3K to dephosphorylate the protein and halt its signalling pathway. Dephosphorylation, like ubiquitination, arrests MAPK signalling, but in a far less permanent way. The protein can still be phosphorylated by other enzymes to set it back on the road to axonal elongation.
Working in this way, it would seem that RPM-1 can control MAPK function, and hence synapse formation, in the short-term by means of the PPM-2 phosphatase and in the long-term by FSN-1 mediated ubiquitination (Image source: Modified from Baker et al., 2014.). This is particularly exciting as it represents what would seem to be the first known incidence of a single signalling molecule bringing about both changes in a MAPK signalling pathway. It presents an almost uniquely sophisticated example of a potential way that axonal elongation and synapse formation could be regulated.
That said, there are, as always, a few caveats. C. elegans have no myelin or invasive glia, making their neuronal environments profoundly diﬀerent to ours. Quite how this aﬀects the reproducibility of their observations in other models remains to be seen.
The group seem to have found a fascinating pathway to respond to positional information, but, as of yet, have no idea where that information could come from. Historically, the suggestion of an internal clock for growth termination was popular, but there is little evidence to substantiate its existence. Much more likely, extracellular signalling molecules such as morphogens found at diﬀerent concentration gradients are responsible for starting up this cascade. Alternatively, membrane-bound adhesion molecules on the postsynaptic membrane may play a role. It is even conceivable that this pathway is actually a part of a much, much larger web of interacting signalling molecules.
Already, the complexity and diversity of these biologically important signalling proteins has led to calls from many suggesting that analysis may only be possible with a computational approach to provide an overall view of the system. Certainly, delineating developmental processes one knockout at a time does very quickly begin to feel like trying to understand the engineering of a sports car by analysing the function of each tyre bolt in turn.
Abrams B. et al. “Cellular and Molecular DeterminantsTargeting the Caenorhabditis elegans PHR Protein RPM-1 to Perisynaptic Regions,” Developmental Dynamics, January 2008 237:630 DOI: 10.1002/dvdy.21446
Baker S.T. et al. “RPM-1 Uses Both Ubiquitin Ligase and Phosphatase-Based Mechanisms to Regulate
DLK-1 during Neuronal Development,” PLOS Genetics, May 2014 10:5 DOI: 10.1371/journal.pgen.1004297
Tessier-Lavigne M. and Goodman C.S. “The molecular biology of axon guidance,” Science, November 1996, 274:5290.
Nguyen L.K. et al. “When ubiquitination meets phosphorylation: a systems biology perspective of EGFR/ MAPK signalling,” Cell Communications and Signalling, July 2013 11:52 DOI: 10.1186/1478-811X-11-52
Gabriel Gavin is a neuroscientist at University College London (UCL) with a particular interest in neural networks, cell signalling and the role of glia in these processes. https://twitter.com/GabrielCSGavin