The role of ubiquitin-mediated pathways in regulating synaptic development, axonal degeneration and regeneration: insights from fly and worm


  • This review was presented at The Journal of Physiology Symposium entitled Size matters: formation and function of GIANT synapses, which took place at the Annual meeting of the Society for Neuroscience, New Orleans, LA, USA on 12 October 2012. It was commissioned by the Editorial Board and reflects the views of the authors.

C. Wu: Email:


Abstract  The covalent attachment of the 76 amino acid peptide ubiquitin to target proteins is a rapid and reversible modification that regulates protein stability, activity and localization. As such, it is a potent mechanism for sculpting the synapse. Recent studies from two genetic model organisms, Caenorhabditis elegans and Drosophila, have provided mounting evidence that ubiquitin-mediated pathways play important roles in controlling the presynaptic size, synaptic elimination and stabilization, synaptic transmission, postsynaptic receptor abundance, axonal degeneration and regeneration. While the data supporting the requirement of ubiquitination/deubiquitination for normal synaptic development and repair are compelling, detailed analyses of signalling events up- and downstream of these ubiquitin modifications are often challenging. This article summarizes the related research conducted in worms and flies and provides insight into the fundamental questions facing this field.

  • image

[ Xiaolin Tian (right) is a postdoctoral research associate and Chunlai Wu (left) is an assistant professor in the Neuroscience Centre at Louisiana State University Health Sciences Centre. Both of them received their PhDs at Washington University School of Medicine. Dr. Tian was trained as a developmental biologist. After a postdoctoral fellowship with Dr. Jason Mills at Washington University, she joined Dr. Wu's laboratory where her major interest is to understand the role of ubiquitin pathway in normal and diseased brain. Dr. Wu received intensive training in the fields of biochemistry and developmental biology during his undergraduate and graduate works. Since his postdoctoral training with Dr. Aaron DiAntonio, Dr. Wu's research interests has centred on resolving the molecular mechanisms controlling presynaptic development, using an approach that combines the powerful fly genetics with transgenics-based biochemistry and proteomics.]


dual leucine-zipper-bearing kinase




neuromuscular junctions


Skp/Cullin/F-Box containing


ubiquitin proteasome system


Synaptic development is a fascinating process that starts with axon growth cone migration, target recognition, the formation of the right connection, and ultimately the establishment of synapses with appropriate pre- and postsynaptic size and strength that maintains activity-based and homeostatic plasticity. In reality this process is even more complicated given that synaptic competition, pruning and elimination also play a role in the process of synaptic maturation. Ubiquitin-mediated pathways play important roles in almost every phase of the synaptic development. Akin to the kinase/phosphatase pair, the ubiquitin ligases/deubiquitination enzymes dynamically control the ups and downs of the protein levels of the key regulators in the developmental processes. Through the precise control of protein abundance and activity at different developmental stages, cell types and cellular compartments, the ubiquitin pathways regulate synaptic development in a spatial- and temporal-specific manner.

Ubiquitin is a 76 amino acid long peptide that can be conjugated to target proteins by a cascade of enzymes called E1, E2 and E3 ubiquitin ligase. Ubiquitination, depending on the length and configuration of ubiquitin chains, can tag the target protein for degradation (polyubiquitination), direct their trafficking, or modify the function of the ubiquitinated proteins (monoubiquitination or multiubiquitination). Among the three enzymes (E1, E2 and E3), E3 ubiquitin ligase has the highest diversity, which allows for a broad spectrum of substrate specificity. E3 ligases can function as monomers or as multimeric E3 complexes, including the anaphase-promoting complex/cyclosome family E3s and Skp/Cullin/F-Box containing (SCF) superfamily E3s. Many recent studies reveal key roles of the ubiquitin system in regulating synaptic structure and function (reviewed in DiAntonio & Hicke, 2004; Patrick, 2006; Collins & DiAntonio, 2007; Ding & Shen, 2008; Haas & Broadie, 2008; Segref & Hoppe, 2009; Kawabe & Brose, 2011). In the present review, we focus on the upstream and downstream regulatory mechanism of the ubiquitination events and give our thoughts on how we can continue using the invertebrate model systems to understand the detailed molecular mechanisms through which ubiquitin pathways regulate synaptogenesis and axonal regeneration.

The role of ubiquitin proteasome system in regulating synaptic structure and function

At the fly larval neuromuscular junctions (NMJs), both ubiquitin-activating enzyme (E1) and the 20s proteasome subunit are present at the synaptic boutons (Speese et al. 2003). Fly dominant temperature-sensitive mutants in the proteasome subunits β2 and β6 were isolated and used to study the loss-of-function of ubiquitin proteasome system (UPS) in a tissue-specific manner (Holden & Suzuki, 1973; Smyth & Belote, 1999). Blocking the general UPS functions in the presynaptic or postsynaptic side using genetic (overexpressing the dominant temperature-sensitive mutant subunits) and pharmacological means drastically alters neurotransmitter release and glutamatergic synaptic strength (Speese et al. 2003; Haas et al. 2007). In fact, the endogenous ubiquitinated (both polyubiquitinated and monoubiquinated) proteins are highly enriched at both the presynaptic nerve terminal and the postsynaptic density of the fly larval NMJs, with the latter showing comparably stronger accumulation (Fig. 1AD). When a tagged ubiquitin was expressed in either neuronal or muscle tissue, the tagged ubiquitin and its ubiquitin conjugates were highly enriched in both the pre- and postsynaptic sites (Fig. 1EL). These data suggest that fly NMJs are well suited for studying ubiquitin pathways that regulate synaptic structure and function. The identities of some ubiquitinated proteins in the fly neuronal tissue were recently revealed through a proteomic approach. Among the 48 newly identified ubiquitinated proteins in neurons, many are key players in synaptogenesis (Franco et al. 2011). Similarly, work done in mammals by Ehlers and others (Ehlers, 2003; Mabb & Ehlers, 2010) demonstrates that UPS-mediated dynamic protein turnover at the postsynaptic density is crucial for normal postsynaptic function and plasticity.

Figure 1.

Ubiquitinated proteins are highly enriched at the Drosophila NMJs 
AD, representative confocal images of muscle 4 NMJs in wild-type third instar larvae stained with an antibody (FK2) that recognizes endogenous mono- and polyubiquitin moiety on ubiquitinated proteins (A), co-stained with an anti-Drosophila vesicular glutamate transporter (DVGlut) antibody (Daniels et al. 2004) (B), and merge (C). The boxed region in C is shown in D in higher magnification and demonstrates a predominant postsynaptic localization of mono- and polyubiquitinated proteins. Co-localizations of ubiquitinated proteins with DVGlut are also evident (arrowhead), suggesting the presence of ubiquitination events in the presynaptic terminal. E–L, Myc-tagged ubiquitin transgene was expressed in muscle (driven by 24B-Gal4) (EH) or neuron (driven by BG380-Gal4) (IL). Representative confocal images of muscle 4 NMJs in third instar larvae stained with a rabbit anti-Myc antibody (Bethyl Laboratories, Inc.) (E, I), co-stained with a mouse anti-BRP antibody (F), a mouse anti-DLG (Discs large, a postsynaptic marker) antibody (J) and Merge (G and K). The boxed region in G and K, shown in H and L in higher magnification, respectively, demonstrates that presynaptically expressed ubiquitin conjugates to areas surrounding the active zone (H), and that postsynaptically expressed ubiquitin conjugates to areas surrounding the postsynaptic densities. Scale bar in D (represents D, H, L) denotes 3 μm. Scale bar in A (represents all but D, H, L) denotes 10 μm.

In addition to loss-of-function of the UPS, perturbation of the ubiquitination process also leads to defective fly NMJ development. Mutation of an E2 ubiquitin conjugase, bendless, results in defects in axon guidance and target recognition (Thomas & Wyman, 1982, 1984; Muralidhar & Thomas, 1993; Oh et al. 1994; Uthaman et al. 2008). Mutations in Cullin2 or Cullin5, components of the SCF complex, cause defects in the organization of synaptic boutons at the NMJs (Ayyub, 2011). Promoting deubiquitination at the fly motoneuron drastically increases the synaptic terminal size (DiAntonio et al. 2001). Although the exact ubiquitination targets responsible for these defects remain unclear, these data suggest that balanced ubiquitination/deubiquitination and UPS function play important roles in regulating synaptic structure and function. Among the currently identified group of ubiquitin ligase complexes, a huge and highly conserved E3 ubiquitin ligase Highwire (Hiw)/RPM-1/Phr1 shows the most dramatic loss of function defects, suggesting a central role of this E3 complex in shaping the nervous system. As such it attracted a great amount of interest in the field. This ubiquitin ligase and its mediated pathways will be the main focus of the following sections.

Regulation of synaptic development by ubiquitin ligase Highwire and RPM-1

A high molecular weight protein with a putative ubiquitin ligase domain was simultaneously identified in flies (Hiw) and worms (RPM-1) (Schaefer et al. 2000; Wan et al. 2000; Zhen et al. 2000). Loss of function of this multidomain protein in each organism has a profound impact on the development and regeneration of the nervous system. In Drosophila, hiw mutant larval NMJs show strikingly overgrown synaptic terminals but weakened synaptic transmission (Wan et al. 2000). At the adult stage, hiw mutants cause similar anatomical overgrowth and physiological dysfunction of the Giant Fibre-tergotrochanteral motoneuron synapse in the Drosophila CNS (Uthaman et al. 2008). More recent studies show that hiw also plays a positive role in the segregation of sister axons during mushroom body formation, in a non-cell autonomous fashion (Shin & DiAntonio, 2011). In Caenorhabditis elegans, rpm-1 mutants show altered distribution and density of GABAergic NMJs, and abnormal targeting and morphology of glutamatergic synapses (Schaefer et al. 2000; Zhen et al. 2000). These observations, together with studies of their mammalian homolog Phr1 (Burgess et al. 2004; Bloom et al. 2007; Lewcock et al. 2007; Culican et al. 2009; Hendricks & Jesuthasan, 2009) and zebrafish homolog Esrom (D'Souza et al. 2005), demonstrated that Hiw/RPM-1/Phr1 play crucial and conserved roles in synaptogenesis. The C-terminal RING finger domain shared by Hiw/RPM-1/Phr1 suggests that they may function as E3 ubiquitin ligase (Joazeiro & Weissman, 2000). Indeed genetic evidence supporting this notion first came from the identification of a strong genetic interaction between the deubiquitination enzyme fat facets (faf) and hiw in Drosophila (DiAntonio et al. 2001). Moreover, the isolated Ring Finger domain of RPM1 is sufficient to promote ubiquitination in vitro (Nakata et al. 2005), and mutations in this RING domain abolish the function of Hiw/Esrom in fly and fish (D'Souza et al. 2005; Wu et al. 2005).

Later studies from worms, flies and mice further demonstrated that Hiw/RPM-1/Phr1 associates with a highly conserved F-Box protein called FSN-1/DFsn/ Fbxo45 and functions as an SCF-like E3 ubiquitin ligase complex. Loss of function of this F-Box protein in flies or mice, respectively, leads to almost identical defects as hiw or phr1 mutants. Worm fsn-1 mutants also show similar phenotypes as rpm-1 mutants although the severity is less profound (Liao et al. 2004; Wu et al. 2007; Saiga et al. 2009). F-Box proteins are key conventional components of SCF ubiquitination complex. A typical SCF complex consists of four core subunits, Skp1, Cullin, F-Box protein and E3 ubiquitin ligase. In worms, FSN-1 forms SCF complex with RPM-1, SKR-1 and CUL-1 (Liao et al. 2004). However, in mammals the Fbxo45 binds to Phr1 and Skp1, but not to Cullin1 due to an amino acid substitution in the consensus motif in its F-Box domain (Saiga et al. 2009). In Drosophila, we were able to purify a Hiw/DFsn/SkpA complex in two independent in vivo tandem–affinity–purification experiments using Hiw or DFsn as bait, respectively (unpublished data). In both cases, Cullin1 was not present in the same complex (Wu et al. 2007; and unpublished data). Subsequent co-immunoprecipitation experiments also confirmed the lack of Cullin1 in the Hiw/DFsn complex (unpublished data). Taken together, these data suggest that fly Hiw/DFsn and mouse Phr1/Fbxo45 form a novel non-SCF complex that lacks the Cullin1 component, while the worm RPM-1/FSN-1/SKR-1/CUL-1 forms a classic SCF complex. It is not clear thus far whether the lack of Cullin protein in the Hiw/Phr-1 complex may impact its functions regarding the substrate selection as well as its own regulation.

An important step to define the ubiquitin pathway is to identify its ubiquitin targets that mediate its specific functions. To define the signalling pathway downstream of the ubiquitin ligase, modifier screens aimed for Hiw/RPM ubiquitination targets were conducted in both worms and flies, based on the rationale that mutations in the specific substrates of RPM-1/Hiw would attenuate the effects of rpm-1/hiw loss-of-function (Nakata et al. 2005; Collins et al. 2006). A MAP triple kinase, dual leucine-zipper-bearing kinase (DLK)-1 (worm)/Wallenda (Wnd, fly), was identified in both worms and flies. Evidence supporting that DLK/Wnd mediates RPM-1/Hiw action includes: (i) dlk-1/wnd mutants dominantly suppress the morphological phenotype caused by rpm-1/hiw mutants; (ii) DLK-1/Wnd protein levels are elevated in the absence of RPM-1/Hiw; and (iii) overexpression of DLK-1/Wnd mimics rpm-1/hiw mutant phenotype (Nakata et al. 2005; Collins et al. 2006). Subsequent works demonstrate that DLK-1/Wnd MAP kinase signalling is downstream of RPM-1/Hiw ubiquitin ligase complexes in many other neuronal processes in worms and flies, including synapse formation and or growth, segregation of sister axons during mushroom body formation (Shin & DiAntonio, 2011), modulation of AMPA receptor trafficking in worm postsynaptic interneurons (Park et al. 2009), and regulation of axonal regeneration (see below). These data suggest that DLK-1/Wnd is a key downstream target of RPM-1/Hiw, which plays a central role in invertebrate synaptic development. Interestingly, the signalling events downstream of DLK-1/Wnd diverge in worms and flies. To promote normal synaptic development, RPM-1 suppresses a DLK-1-activated p38 MAP kinase cascade in worms while Hiw suppresses Wnd-activated JNK MAP kinase signalling.

It is worth noting that knockout of DLK in mice fails to suppress the Phr1 mutant phenotypes, and no change of DLK protein levels is identified in the Phr1 mutant mouse brain, suggesting that DLK does not mediate Phr1's function in axon tract formation during mammalian CNS development (Bloom et al. 2007). Phr1 also plays an evolutionarily conserved role in regulating NMJ development in mouse; however, whether this function requires DLK remains to be determined. Identifying Phr1's ubiquitin targets will help address the question of how the Phr1 ubiquitin pathway plays its cell- and tissue-specific roles during neural development in mammals.

Regulation of synaptic development by other ubiquitin pathways

Synaptic growth during Drosophila development is regulated by another two ubiquitin pathways besides the Hiw/DFsn pathway. First, the deubiquitination enzyme Fat facets targets Liquid facets (Lqf), the fly homolog of Epsin 1, to promote synaptic growth and function (Bao et al. 2008). Although Lpf is not a Hiw ubiquitin target, it converges on the Hiw pathway in regulating neurotransmitter release, based on the fact that lqf mutation partially suppresses the synaptic transmission defect of hiw mutants (Bao et al. 2008). Second, the ubiquitin conjugatase (E2), Bendless, clearly regulates a distinct ubiquitination cascade to promote synaptic growth in the CNS (Uthaman et al. 2008), in contrast to Hiw whose function is required to restrain synaptic growth in the peripheral nervous system (Wan et al. 2000). Together these ubiquitin pathways may regulate synaptic development at different critical developmental stages, subcellular locations and/or different neuronal cell types.

Furthermore, many other ubiquitin pathways also play important roles in a variety of aspects of synaptic development. To date, at least six different E3 ligase complexes were identified that regulate a variety of processes during synaptic development: (i) anaphase-promoting complex/cyclosome restrains synaptic size by controlling the presynaptic Liprin-α protein level at fly NMJs (van Roessel et al. 2004), and regulates synaptic transmission by controlling the number of glutamate receptors postsynaptically in both Drosophila and C. elegans (Juo & Kaplan, 2004; van Roessel et al. 2004), and controls axon growth and patterning in mice (Konishi et al. 2004); (ii) presynaptic-membrane-associated SCF E3 complex containing the F-Box protein Scrapper regulates synaptic vesicle release by targeting RIM1 in mice (Yao et al. 2007), and the fly Scrapper is expressed as a nerve-membrane-associated protein (unpublished observation), but whether it also downregulates RIM1 remains unclear; (iii) F-Box protein SEL-10 works in a SCF E3 complex to control synaptic elimination and stabilization in C. elegans (Ding et al. 2007); (iv) postsynaptic lin-23 containing SCF E3 and a KEL-8/Cullin-3 E3 regulate the abundance of the C. elegans AMPA-type glutamate receptor (Dreier et al. 2005; Schaefer & Rongo, 2006); (v) short isoform of fly Nedd4, a HECT E3 ubiquitin ligase, regulates NMJ formation by promoting ubiquitination and internalization of Commissureless (Comm), while a long isoform of Nedd4 plays an inhibitory role in NMJ formation (Ing et al. 2007; Zhong et al. 2011); and (vi) F-Box protein, Mec-15, functions in parallel with the RPM-1 pathway in worm touch receptor neurons to promote synaptic formation and development (Bounoutas et al. 2009).

The cross-talk between ubiquitin pathways and autophagy pathways during synaptic development

There are two major pathways that mediate protein degradation in eukaryotic cells: UPS pathway and autophagic pathway. It was long believed that ubiquitin pathways regulate target-specific degradations while autophagy serves as a bulk, non-selective degradative process. However, recent studies demonstrate that these two pathways not only interact with each other but also work together to achieve selective autophagy. Elevation of autophagy levels in fly motoneuron leads to downregulation of Hiw E3 ligase, which in turn causes synaptic terminal overgrowth, while knocking down autophagy levels leads to upregulation of Hiw and causes moderate synaptic terminal undergrowth (Shen & Ganetzky, 2009). A study on Rae1, an evolutionarily conserved Hiw/RPM-1/Phr1-interacting protein, sheds light on how the selective regulation of Hiw by autophagy is achieved in neurons. rae1 mutant flies or worms show loss-of-function phenotype that is reminiscent to hiw or rpm-1 mutants, respectively, suggesting that Rae1 is a positive regulator of the Hiw/RPM-1 ubiquitin ligase complexes (Tian et al. 2011; Grill et al. 2012). In flies, Rae1 is necessary and sufficient to promote Hiw protein abundance, and Rae1 plays such a role through physically binding to Hiw and protecting Hiw from autophagy-mediated degradation (Tian et al. 2011). In worms, genetic data indicate that Rae1 works downstream of RPM-1, together with Glo-4 and FSN-1 pathways (Grill et al. 2012).

Ubiquitin pathways interact with autophagy pathways beyond neural development. In flies, when UPS is impaired, autophagy is activated in a histone deacetylase 6 (HDAC6)-dependent fashion (Pandey et al. 2007). Emerging data suggest another important role of ubiquitin pathways is to tag proteins for selective autophagy (reviewed in Shaid et al. 2012). Cross-talk between ubiquitin pathways and autophagy is provided by autophagic adaptor proteins that simultaneously interact with ubiquitin and autophagosome markers (Shaid et al. 2012). It will be interesting to test whether the autophagy-dependent degradation of Hiw also requires ubiquitination and whether Rae1 may interfere with this process and in turn prevent Hiw from being degraded through the autophagic pathway.

Regulation of axonal degeneration and regeneration by Highwire-mediated ubiquitin pathways

In addition to regulating the development of nervous system, ubiquitin pathways also play important roles in the degeneration of distal axons after nerve injury, a process called Wallerian degeneration. Inhibition of UPS by pharmacological and genetic means leads to delayed Wallerian degeneration both in vitro and in vivo (Zhai et al. 2003). However, it is unclear whether the spectrum of or only few specific substrates are targeted by UPS in this process and which E3 ubiquitin ligases are affected by the pharmacological treatment. A recent study demonstrated that Hiw promotes Wallerian degeneration by downregulating a previously unknown target nicotinamide mononucleotide adenyltransferase (NMNAT) (Xiong et al. 2012). Research on Wallerian degeneration in the past two decades established a central role of NMNAT in protecting severed axons from degeneration (Zhai et al. 2008; Coleman & Freeman, 2010; Feng et al. 2010). Thus rapid degradation of NMNAT protein at the distal “stump” mediated by Hiw provides an example where a specific E3 ubiquitin pathway targets a key inhibitor of degeneration to achieve distal axon clearance (Xiong et al. 2012). This mechanism seems to be conserved from flies to mammals as Hiw can also target ectopically expressed mouse NMNAT2 in distal axons and synapses (Xiong et al. 2012). It will be interesting to test whether Phr1 null mice show slowed Wallerian degeneration. Interestingly, in addition to targeting NMNAT, Hiw downregulates Wnd/DLK in a parallel pathway to promote Wallerian degeneration (see below) (Xiong & Collins, 2012; Xiong et al. 2012).

Although NMNAT protects against injury-induced axonal degeneration, it has no effect on naturally occurring developmental axon degeneration in flies or mice (Hoopfer et al. 2006). It seems that developmental pruning of both axon and dendrite uses distinct molecular machinery from injury-induced degeneration. Unsurprisingly, the ubiquitin pathways again play important roles in developmental pruning of both axon and dendrite. Mutations in ubiquitin activating enzyme 1 (E1) or proteasome regulatory subunits inhibit axonal pruning of gamma neurons of the Drosophila mushroom bodies during metamorphosis (Watts et al. 2003). Activation of UbcD1, an E2 ubiquitin-conjugating enzyme, and deactivation of a caspase-antagonizing E3 ligase DIAP1 are required for proper pruning of the C4da dendritic arborization sensory neurons in Drosophila during metamorphosis (Kuo et al. 2006). Thus, UPS regulates axonal degeneration during development and after injuries through distinct ubiquitination components.

One common defect of the hiw/rpm-1/Phr1 loss of function is uncontrolled growth of presynaptic arbours, featured by overgrowth of NMJ presynaptic terminal in flies, abnormal axonal extension of glutamatergic neurons in worms and over-sprouting of motoneuron presynaptic terminals in mouse NMJs. These data show that one conserved function of the Hiw/RPM-1/Phr1 ubiquitin ligase is to restrain axonal terminal growth. It also suggests that the normal function of their ubiquitin targets, such as Wnd/DLK-1 in invertebrates, is to promote axonal terminal growth during neural development. In agreement with this notion, a series of studies demonstrated that DLK kinase plays an evolutionarily conserved role in promoting axonal regeneration after axonal injury. DLK-1 was first identified in a genetic screen in worms and was shown to be necessary and sufficient for robust motor neuron axonal regrowth after injury (Hammarlund et al. 2009). Meanwhile an elegant study demonstrated that DLK-1 promotes axonal regeneration through a signalling cascade involving DLK-1 and MAK-2 (a MAPKAP kinase). DLK-1 activates MAK-2 to stabilize the mRNA encoding CEBP-1, and elevated CEBP-1 protein level is required for axonal regrowth (Yan et al. 2009). Interestingly, loss of function in MAK-2, cebp-1 or uev-3 (ubiquitin E2 variant) can each suppress the rpm-1 mutant phenotype (Yan et al. 2009; Trujillo et al. 2010), suggesting the same DLK-1 downstream signalling functions during development and after axonal injury to promote growth and promote regeneration. Such a DLK-1 activity is also crucial to mediate the effect of elevated Ca2+ and cAMP in injured axonal regrowth (Ghosh-Roy et al. 2010). In support of this idea, independent studies using a nerve crush assay in Drosophila larvae demonstrate that axonal injury triggers rapid decrease of Hiw and simultaneous increase of Wnd/DLK protein levels (Xiong et al. 2010). The locally elevated Wnd kinase then induces a cell-autonomous injury signalling cascade, which requires axonal transport and JNK MAP kinase activation, to promote axonal growth. This injury-induced Wnd activation is also beneficial to the axon as it prevents degeneration of the distal axon from degeneration after future injury (Xiong & Collins, 2012).

The DLK-mediated axonal regeneration in invertebrates seems to be conserved in mammals. DLK-deficient mice show reduced neurite growth of dorsal root ganglion neurons after axonal injury (Itoh et al. 2009). DLK promotes retrograde transport of an injury signal to the neuronal cell bodies where proregenerative signals are activated (Shin et al. 2012). Interestingly, while mammalian DLK promotes axonal regrowth at the proximal axons, it is also required for the degeneration of distal axons (Wallerian degeneration) after axotomy (Miller et al. 2009) and after growth factor withdrawal (Ghosh et al. 2011). It is tempting to argue that DLK plays dual roles in both the clearance of the distal axons and repair of proximal axons. Although the enzyme–substrate relationship between Phr1/Hiw/RPM-1 (PHR) and DLK/Wnd is not completely conserved across species, all the data suggest that PHR and DLK play important roles in synaptic development, maintenance and injury response (Fig. 2).

Figure 2.

RPM-1/Hiw/Phr1-mediated ubiquitin pathways in worm, fly or mammal 
The E3 ubiquitin ligases RPM-1, Hiw and Phr1 show highly conserved domain structures, binding partners, and all play a central role in neural development. They all form SCF/SCF-like ubiquitination complexes to target downstream substrates for degradation. While the Phr1's targets in mammalian brain are not yet identified, both RPM-1 and Hiw target a dual leucine-zipper-bearing kinase (DLK) to mediate an important part of their signalling events. These events include the development of presynaptic structures with appropriate size and strength, modulation of proximal axon regeneration and distal axon degeneration after axonal injury. Mammalian DLK itself also plays a key role in response to nerve injury, as it promotes both the degeneration of the distal axons and regeneration of the proximal axons.

Genetic and biochemical approaches to study the regulation and molecular action of ubiquitin pathways

How does the cell control the activity levels of a specific ubiquitin pathway? E3 ubiquitin ligases, which carry the capacity to achieve the precise control of substrate specificity, probably serve as targets for such a regulation. There are about 1000 E3 ligases in the mammalian genome (Glickman & Ciechanover, 2002). While the E3 ligases share similarities in their E3 ligase domains, they often contain distinct regulatory domains. Studying the regulatory domains of individual E3 ligase is key to understanding the mechanisms that regulate the activity of the ubiquitin pathway. In the case of Hiw/RPM-1, the huge E3 ligase bears a RCC-1 domain at the N-terminus, two PHR domains and a Myc-binding region in the middle, and a RING finger E3 ligase domain at the C-terminus. The fact that all the functional domains in the Hiw proteins are indispensible for Hiw's action demonstrates that these domains may all participate in the regulation of Hiw's E3 ubiquitin ligase activity (Tian et al. 2011). Indeed, the PHR domains of RPM-1 is both necessary and sufficient for targeting RPM-1 to the right subcellular location – the presynaptic sites (Abrams et al. 2008), and the region around the Myc-binding domains in Hiw/RPM-1 is required for the interaction between Rae1 and the E3 ubiquitin ligases, which regulates the protein abundance of the Hiw ligase (Tian et al. 2011; Grill et al. 2012).

Forward genetic screens in flies and worms identified many components of UPS pathways such as hiw, rpm-1 and fsn-1 that work at the E3 ligase level to regulate synaptogenesis (Schaefer et al. 2000; Wan et al. 2000; Zhen et al. 2000; Liao et al. 2004). Modifying screens looking for suppressors of hiw or rpm-1 mutant then helped to identify their downstream components, including hiw downstream target wnd, the MAP kinase signalling cascade dlk-1/mkk-4/pmk-3 downstream of rpm-1, and other components downstream of dlk-1 that mediate DLK-1's action in axonal regeneration (Yan et al. 2009). Future modifier screens looking for enhancers of hiw or rpm-1 mutant will probably identify genes that work upstream of, or in parallel with, E3 ligases. It is evident that genetic screens played and will continue to play essential roles in the studies of ubiquitin pathways using worms and flies. However, they also face the challenges of functional redundancy/compensation and lethality at early developmental stages.

To overcome these obstacles, relatively large-scale biochemical/proteomic screens coupled with powerful genetic manipulations have been conducted in both flies and worms to identify the interacting partners of Hiw and RPM-1, respectively. In flies, the Tandem Affinity Purification (TAP) technique was combined with the UAS/Gal-4 system to identify Hiw-interacting proteins from the fly brains. The TAP tag contains two distinct affinity binding domains separated by a highly specific protease cleavage site, thus allowing enrichment of specific protein–protein binding through a two-step purification process (Rigaut et al. 1999). To purify the Hiw ubiquitination complex, an N-terminal TAP-tagged functional Hiw transgene (NTAP-Hiw) was generated (Wu et al. 2005). A two-step affinity purification (illustrated in Fig. 3) was performed to isolate the Hiw-interacting complex from adult fly heads (Wu et al. 2007; Tian et al. 2011). In worms, a similar proteomic screen was also conducted to identify RPM-1-interacting proteins. A function GFP-tagged RPM-1 transgene was generated and RPM-1-interacting proteins were isolated from whole worm lysates by immunoprecipitation using an anti-GFP antibody (Grill et al. 2007). Two positive regulators of the Hiw/RPM-1 pathways, FSN-1/DFsn and Rae1, were independently identified in both screens (Wu et al. 2007; Tian et al. 2011; Grill et al. 2012), demonstrating a great deal of conservation between flies and worms, as well as the feasibility of the biochemical assay. In addition to FSN-1 and RAE-1, GLO-4 (a guanine nucleotide exchange factor) was identified as an RPM-1-interacting protein in worms. GLO-4 and its target GLO-1 act in a linear pathway downstream of RPM-1 to regulate axonal termination and synaptogenesis (Grill et al. 2007). In summary, these biochemical analyses performed in flies and worms demonstrate a potentially powerful approach to identify and study previously unknown components of a given ubiquitin pathway.

Figure 3.

Tandem-Affinity-Purification (TAP) to identify Hiw-interacting proteins 
In a standard TAP experiment, the gene of interest is expressed with a TAP tag. Total protein lysate containing the tagged complex is loaded to the first column (IgG beads). The tagged complex is then released by a TEV protease cleavage and loaded to the second column (calmodulin beads). After two rounds of purification, the final elution was analysed by direct mass spectrometry or SDS-PAGE followed by mass spectrometry of individual bands.

Conclusion and future directions

Neuronal development and plasticity not only rely on programmed gene expression and cell–cell signalling-induced trafficking and post-translational modification (such as phosphorylation), but also the tightly regulated control of protein levels at the synapses. The ubiquitin–proteasome degradation system is one of the key mechanisms that control protein abundance at the synapse. What we learned from current studies on ubiquitin pathways in neurons is that many ubiquitin pathways are present at the synapse to target specific key factors for synaptogenesis (Ding & Shen, 2008; Haas & Broadie, 2008; Kawabe & Brose, 2011). The same ubiquitin pathway is often involved in distinct cellular processes depending on different developmental stages, cell types and species. These UPS pathways are often tightly regulated to suit the temporal and cell type-specific requirement of their functions. The ubiquitination/deubiquitination processes are sometimes coupled with phosphorylation/dephosphorylation events to fine tune the levels of signal transduction (Kawabe & Brose, 2011). Furthermore, cross-talks between ubiquitination and autophagy add another layer of regulation to control signalling outcome precisely.

Despite much progress in understanding the role of ubiquitin pathways in synaptic development, we still know relatively little about detailed molecular mechanisms underlying local ubiquitination events at the synapses. A number of key questions remain to be answered: What are the exact molecular components and co-factors of each ubiquitination complex? How is the activity of ubiquitination complexes controlled? And particularly, what are their ubiquitination targets? How do the deubiquitination enzymes antagonize the ubiquitin pathways and how are these enzymes regulated? Addressing these questions requires a combination of multidisciplinary approaches, including genetic, proteomic and computational analyses to systematically identify and characterize previously unknown components of ubiquitination complexes, specific ubiquitin targets, and upstream and downstream signalling cascades of a given ubiquitin pathway. We believe that Drosophila and C. elegans model systems hold great potential to tackle these challenges in the context of neural development and regeneration. Both systems feature amenable in vivo preparation to genetic and biochemical manipulation, where it is easy to assay both synaptic structure and function, as well as axonal degeneration and regeneration. Both systems allow large-scale genetic and proteomic screens aimed at defining any given ubiquitin-mediated molecular pathways. With available tools to study the loss-of- function of almost every gene, research using flies and worms could once again make a huge impact on our understanding of the biology of ubiquitination in the neural system.


Additional information


We thank NIH/NINDS (NS070962) for supporting our research.


We are grateful to Dr. Ian D. Forsythe for valuable discussion about the manuscript.