- Record: found
- Abstract: found
- Article: found
Vertebrate Fidgetin Restrains Axonal Growth by Severing Labile Domains of Microtubules
research-article
Author(s):
Lanfranco Leo
1 ,
Wenqian Yu
1 ,
Mitchell D’Rozario
2 ,
Edward A. Waddell
2 ,
Daniel R. Marenda
1
,
2 ,
Michelle A. Baird
3 ,
Michael W. Davidson
3 ,
Bin Zhou
4 ,
Bingro Wu
4 ,
Lisa Baker
5 ,
David J. Sharp
5 ,
Peter W. Baas
1
Publication date (Electronic):
03 September 2015
Journal:
Cell reports
Read this article at
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
SUMMARY
Individual microtubules (MTs) in the axon consist of a stable domain that is highly acetylated and a labile domain that is not. Traditional MT-severing proteins preferentially cut the MT in the stable domain. In Drosophila, fidgetin behaves in this fashion, with targeted knockdown resulting in neurons with a higher fraction of acetylated (stable) MT mass in their axons. Conversely, in a fidgetin knockout mouse, the fraction of MT mass that is acetylated is lower than in the control animal. When fidgetin is depleted from cultured rodent neurons, there is a 62% increase in axonal MT mass, all of which is labile. Concomitantly, there are more minor processes and a longer axon. Together with experimental data showing that vertebrate fidgetin targets unacetylated tubulin, these results indicate that vertebrate fidgetin (unlike its fly ortholog) regulates neuronal development by tamping back the expansion of the labile domains of MTs.
Graphical Abstract
Related collections
Most cited references17
- Record: found
- Abstract: found
- Article: not found
Characterization and spatial distribution of the ELAV protein during Drosophila melanogaster development.
S Robinow, Klane White (1991)
The embryonic lethal abnormal visual system (elav) gene of Drosophila melanogaster is required for the development and maintenance of the nervous system. Transcripts from this locus are distributed ubiquitously throughout the nervous system at all developmental stages. A product of this gene, the ELAV protein, has homology to known RNA binding proteins. The localization of the ELAV protein was studied in all developmental stages using antibodies that were generated against a hybrid protein made in Escherichia coli. In general, these data are consistent with previous results and demonstrate that (1) the ELAV protein is detected in the developing embryonic nervous system at a time coincident with the birth of the first neurons, (2) the ELAV protein is first detected in the majority of neurons of the central and peripheral nervous systems of embryos, larvae, pupae, and adults, (3) the ELAV protein appears to be localized to the nucleus, and (4) the ELAV protein is not detected in neuroblasts or identifiable glia. These data also provide new information concerning elav expression and show that (1) ELAV is not expressed in the ganglion mother cells (GMCs), (2) while the ELAV protein is localized to the nucleus, it is not uniformly distributed throughout this structure, and (3) other Drosophila species do express an ELAV-like antigen. We propose that the elav gene provides a neuronal-housekeeping function that is required for the successful posttranscriptional processing of transcripts from a set of genes the function of which is required for proper neuronal development and maintenance.
- Record: found
- Abstract: found
- Article: not found
The microtubule-severing proteins spastin and katanin participate differently in the formation of axonal branches.
L Qiang, Arzu Karabay, Wenqian Yu … (2008)
Neurons express two different microtubule-severing proteins, namely P60-katanin and spastin. Here, we performed studies on cultured neurons to ascertain whether these two proteins participate differently in axonal branch formation. P60-katanin is more highly expressed in the neuron, but spastin is more concentrated at sites of branch formation. Overexpression of spastin dramatically enhances the formation of branches, whereas overexpression of P60-katanin does not. The excess spastin results in large numbers of short microtubules, whereas the excess P60-katanin results in short microtubules intermingled with longer microtubules. We hypothesized that these different microtubule-severing patterns may be due to the presence of molecules such as tau on the microtubules that more strongly shield them from being severed by P60-katanin than by spastin. Consistent with this hypothesis, we found that axons depleted of tau show a greater propensity to branch, and that this is true whether or not the axons are also depleted of spastin. We propose that there are two modes by which microtubule severing is orchestrated during axonal branch formation, one based on the local concentration of spastin at branch sites and the other based on local detachment from microtubules of molecules such as tau that regulate the severing properties of P60-katanin.
- Record: found
- Abstract: found
- Article: not found
Drosophila Spastin Regulates Synaptic Microtubule Networks and Is Required for Normal Motor Function
Nina Tang Sherwood, Xiao-Qi Sun, Mingshan Xue … (2004)
Introduction “Pure” autosomal dominant hereditary spastic paraplegia (AD-HSP) is an inherited disease characterized by bilateral spasticity in the absence of other phenotypes (reviewed in Fink 2003; Reid 2003). Afflicted patients experience difficulty in walking and have a distinctive gait. Degeneration of the lateral corticospinal tracts, which contain the axons of cortical neurons that innervate primary limb motoneurons, is observed in the lumbar regions of the spinal cord in patients with AD-HSP. The distal segments of long dorsal root ganglion axons also display degeneration. No evidence is seen for cell death or for primary myelination defects, and the axons of primary motor neurons do not degenerate (Maia and Behan 1974; Wharton et al. 2003). AD-HSP thus appears to selectively affect the distal regions of the longest axons within the spinal cord. Because pathology is usually confined to long spinal cord axons, it has been suggested that the primary defect in pure AD-HSP is in axonal transport or some other process required for maintenance of axon terminals. Perturbation of anterograde or retrograde axonal transport might selectively affect the longest axons, because they would be most vulnerable to a reduction in efficiency of transport of material to or from their terminals. About 40% of cases of pure AD-HSP are caused by mutations in the SPG4 gene, which encodes an AAA ATPase called Spastin (Hazan et al. 1999). AAA ATPases are a large and diverse set of proteins that include an approximately 250–amino acid (aa) conserved domain containing Walker A and B ATP-binding motif sequences (reviewed in Confalonieri and Duguet 1995; Patel and Latterich 1998; Neuwald et al. 1999). They use energy obtained from ATP hydrolysis to catalyze assembly or disassembly of a variety of protein complexes. AAA proteins are involved in many cellular processes, including vesicle trafficking, protein degradation, and microtubule dynamics. Many AAA ATPases form hexameric rings, and it is thought that the ring structures are required for catalytic activity (Vale 2000). Spastin is a member of the “meiotic” subgroup of AAA ATPases (Frohlich 2001; Frickey and Lupas 2004), which contains proteins involved in vesicle trafficking and microtubule dynamics. The only member of the subgroup whose activities have been biochemically characterized is Katanin-60, which is the catalytic subunit of a microtubule-severing protein (McNally and Vale 1993; Hartman et al. 1998). Katanin-60 and Spastin are homologous only within their AAA domains. However, cell culture studies have provided evidence that Spastin is also involved in microtubule dynamics. Expression of wild-type human Spastin in transfected cell lines and cortical neurons caused disassembly of the microtubule cytoskeleton, while a mutant Spastin lacking catalytic activity colocalized with tubulin (Errico et al. 2002; McDermott et al. 2003). The mechanisms by which spastin mutations produce dominant spasticity phenotypes in humans are controversial. A wide variety of nonsense and missense mutations, but no complete gene deletions, have been found in families with AD-HSP. It has been suggested that dominance arises from haploinsufficiency (Charvin et al. 2003). This model, however, would require that the processes in which Spastin participates are vulnerable to a 50% decrease in its enzymatic activity. Another possibility is that truncated or missense mutant Spastins function as dominant negatives. Hexameric AAA ATPase ring complexes might be especially vulnerable to the presence of nonfunctional (“poison”) subunits that assemble into rings but lack catalytic activity. Consistent with this idea, expression of a mutant Spastin that associates with microtubules but cannot catalyze severing altered organelle distribution in transfected cells (McDermott et al. 2003). In the dominant negative model, AD-HSP might only occur when Spastin activity is eliminated or greatly reduced. In this study, we describe the phenotypes arising from mutation of the Drosophila ortholog of human spastin. We initially identified this gene in a gain-of-function screen, in which we found that its overexpression in neurons causes axons in the embryonic central nervous system (CNS) to converge onto the midline (Sun 2000). Overexpression of Spastin in muscles erases their microtubule networks, consistent with the idea that Spastin is a microtubule-severing protein. We made loss-of-function (LOF) spastin mutations, and found that they produce recessive phenotypes affecting the larval neuromuscular system. The Drosophila neuromuscular junction (NMJ) uses glutamate as its neurotransmitter and employs ionotropic glutamate receptors homologous to vertebrate AMPA receptors (Schuster et al. 1991; Petersen et al. 1997; Marrus et al. 2004). It is organized into presynaptic boutons that are surrounded by a postsynaptic scaffold, and its synapses exhibit plastic behavior during development. These properties make the fly NMJ a useful genetic model system for the study of glutamatergic synapses in the mammalian brain and spinal cord (Keshishian et al. 1996; Koh et al. 2000). During the period from larval hatching through the third instar stage, the number of boutons at each NMJ increases by up to 10-fold in order to keep pace with the growth of its muscle target. New boutons are added by a process of budding (Zito et al. 1999). As these boutons mature, their microtubule cytoskeleton is thought to progress through a regulated series of alterations (Roos et al. 2000; Pennetta et al. 2002). In this paper, we show that synaptic growth and function are altered in spastin mutant larval NMJs. Boutons are more numerous than in wild-type larvae, and synaptic transmission is impaired. These changes could result from alterations in synaptic microtubule dynamics, because we find that microtubule bundles are depleted from the distal boutons of NMJs in spastin-null mutants. This is surprising, because the fact that Spastin overexpression destroys microtubule networks might lead one to expect that its removal would increase the number of microtubules. Morphological and microtubule phenotypes are seen only for a total gene deletion, indicating that complete loss of Spastin function is required to alter synaptic microtubules in the fly system. The phenotypes we see are quite different from those described in a recently published study of perturbation of Drosophila spastin using RNAi methods (Trotta et al. 2004). In particular, the changes in synaptic microtubules that occur in spastin LOF mutants are opposite to those reported in the RNAi perturbation paper. spastin is not an essential gene, but mutant adults have severely compromised motor behavior. Null mutants cannot fly or jump, they climb slowly, and they often drag their hind legs. While it is intriguing that spastin mutant flies display such movement phenotypes, further work will be required to determine whether Drosophila can provide a useful organismal model system for human AD-HSP. Nevertheless, insights into the cellular functions of Drosophila Spastin obtained from our work should be relevant to an understanding of Spastin's functions in human neurons. Results The Drosophila spastin Gene We identified spastin in an “EP” screen for genes involved in embryonic CNS development. EPs are P element derivatives with a block of “UAS” sites recognized by the yeast transcription factor GAL4 near one end (Rorth 1996). An EP element inserted in the proper orientation upstream of a gene will drive its expression in a cell-specific manner when the insertion line is crossed to the appropriate promoter-GAL4 “driver” line (Brand and Perrimon 1993). We generated approximately 6,000 new EP insertion lines and screened them by crossing to pan-neuronal (ElavC155 ) and pan-muscle (24B) GAL4 driver lines (Lin and Goodman 1994; Luo et al. 1994). Those lines for which crosses to either driver generated reduced numbers ( 0.1), and was 30% greater than in spastin5.75 mutants (p < 0.005; Figure 5E). We also observed that the electrophysiological phenotypes of the hypomorphic spastin17-7/spastin5.75 larvae were temperature sensitive. While the average QC in transheterozygotes was reduced to 78% of controls at 18 °C, this effect was exacerbated at higher temperatures. At 29 °C, QC was 54% of wild-type (Figure 5F). However, QC in control or in spastin5.75 null larvae was unaffected by temperature. These results suggest that the N-terminally truncated Spastin protein that is probably made from the spastin17-7 allele is temperature sensitive. At 29 °C, this hypomorphic allele behaves as a null with respect to NMJ electrophysiology. Finally, we examined escaper larvae (Elav-GAL4/+; T32/+) overexpressing Spastin in neurons, and found that synaptic transmission in these larvae was not significantly different from wild-type (data not shown). spastin Mutant Adults Have Severe Behavioral Phenotypes We observed that only approximately 20% of homozygous spastin5.75 pupae were able to eclose at room temperature compared to 94% of heterozygotes, and the adults that emerged had severe movement defects (see Video S1). They could not fly at all, and did not even appear to move their wings, although the wings inflated and straightened in a normal manner immediately after eclosion. They also did not jump spontaneously, but would jump if persistently prodded in the abdomen. Their legs were weak: when standing, the metathoracic legs often slipped out from underneath them, and during walking they often dragged these legs (see Video S1). They also had difficulty holding on to surfaces when they were upside down. These phenotypes were temperature dependent. Null mutant flies that developed at 18 °C eclosed at much higher rates (56%) than at higher temperatures and moved more normally. Flies homozygous for the hypomorphic mutations, spastin10-12 and spastin17-7, eclosed at normal frequencies at all temperatures. To evaluate these movement defects, we assayed flight and climbing ability in spastin-null and hypomorph flies (Figure 6). The flight assay could only be used for hypomorphs since null mutants were flightless. In this assay, flies were released into the top of a vertical cylinder that had been coated on the inside with oil (Benzer 1973; Atkinson et al. 2000). Poor fliers who took longer to fly fell to the bottom or collided with the lower walls of the cylinder, while good fliers who responded rapidly to being dropped collided with the upper walls. A histogram of the distribution of oil-trapped flies along the height of the cylinder showed that more than half of the spastin hypomorphs did not fly in time to avoid falling to the bottom of the cylinder (Figure 6A). In contrast, the majority (approximately 75%) of the controls, including w1118 (another Canton S-derived w− control) flies and flies homozygous for T32 (in the absence of a GAL4 driver; these have orange eyes), flew well enough to distribute themselves along the sides of the column. Interestingly, for those hypomorphs that did fly out to the sides, their distribution paralleled that of the controls, suggesting that flight responses in the column were relatively normal in this subpopulation of the mutants (Figure 6B). In the climbing assay, flies were tested for their ability to climb up the side of a vial within a limited time period. All WCS and almost all homozygous spastin10-12 and spastin17-7 flies, but only about 40% of spastin5.75 flies, climbed to the top of the vial within 30 s (Figure 6C). This difference did not reflect a loss of geotactic behavior, since spastin5.75 flies were typically found at the tops of their vials after several minutes. Overall, mean climbing velocity was approximately 9-fold slower for spastin-null mutants than for wild-type flies, while the hypomorphs were about 2-fold slower than controls (Figure 6D). We also measured the lifespan of the flies. Under our conditions, WCS flies lived an average of 46 d at 25 °C. spastin10-12 and spastin17-7 flies had somewhat shorter lifetimes, surviving an average of 35 d. Lifespan was dramatically reduced in spastin5.75 flies, which lived an average of only 8 d (Figure 6E and 6F). Examination of mortality curves (Figure 6E), however, revealed that, as in the case of flight ability in the hypomorphs, these flies had a bimodal lifespan distribution. Only 55% of flies were still alive at 4 d post-eclosion, but most of these (37% of the total) then remained alive until about 25 d post-eclosion. After this time, they rapidly died off, and no flies remained alive more than 33 d. Another spastin phenotype observed in adults was male-specific lethality (Figure 6G). For WCS and spastin hypomorphs, more than 40% of eclosed adults were males. However, only approximately 10% of eclosed spastin5.75 flies were male. We do not understand the origins of this phenotype. In summary, spastin-null adult flies had severely compromised movement behavior and were short-lived, while spastin hypomorphs displayed weaker movement and lifespan phenotypes. We also examined rescue for these behavioral phenotypes. When compared to their non-rescued siblings (spin-GAL4/CyOKr-GFP; spastin5.75) from the same cross, spin-GAL4/UAS-spastin; spastin5.75 flies climbed better, were more coordinated, and lived longer (Figure S2; Video S1), indicating that partial rescue was achieved. These flies were still very slow, and it is clear that spin-GAL4-driven Spastin expression did not restore behavior to the levels characteristic of control flies such as WCS. However, genetic background effects made the precise efficacy of rescue achieved in this experiment difficult to determine (the non-rescued spastin5.75 sibling flies bearing the driver and balancer chromosomes used in the rescue cross were much more unhealthy and slow-moving than spastin5.75 flies without these chromosomes). Spastin Overexpression Erases Microtubule Networks In Vivo To investigate whether Drosophila Spastin affects microtubule networks, we overexpressed it in embryonic muscles using the G14-GAL4 or 24B-GAL4 drivers, and then visualized muscle microtubules in late stage 16 embryos with an anti-β3-tubulin antibody that preferentially stains polymerized tubulin (Buttgereit et al. 1996). In wild-type embryos, a complex network of microtubules aligned along the muscle axes was observed (Figure 7A; two segments are shown). This was clearly seen in both vertically oriented (18 and 21–24) and diagonally oriented (5,11, and 19) muscles. When Spastin was overexpressed in muscles, the muscle microtubule network completely disappeared (Figure 7B). The muscles themselves appeared rounded, and were partially or totally detached from their insertion sites. This detachment may have been a consequence of the dissolution of the microtubule network. Oriented microtubule networks could still be seen within the cap cells of the chordotonal organs in each segment (Figure 7B, brackets); these cells did not express the GAL4 driver and therefore did not overexpress Spastin. Dissolution of the microtubule network was therefore specific to cells in which Spastin was overexpressed. We also overexpressed Spastin in larval muscles by crossing the muscle-specific MHC-GS-GAL4 driver line to UAS-spastin flies. MHC-GS-GAL4 is RU486-inducible, but we could not feed the larvae with RU486 as this was lethal. However, MHC-GS-GAL4 also confers late RU486-independent expression in third instar larvae, and we were able to obtain escaper larvae from the cross and double-stain these for Spastin and α-tubulin. As shown in Figure 7C–7E, larvae that lacked detectable Spastin expression had a dense network of muscle microtubules. In contrast, MHC-GS-GAL4/UAS-spastin larvae that had high levels of muscle Spastin displayed a dramatic reduction in microtubules, so that only faint and sparse microtubules were observed in the muscle fibers (Figure 7F–7H). The strongly staining microtubules still visible in these larvae are those of the neurons and tracheae, which do not overexpress Spastin (Figure 7G–7H). These results also show that this α-tubulin antibody preferentially recognizes polymerized tubulin, since the total amount of tubulin dimers would be the same in both sets of muscles (tubulin dimers are very stable proteins and are unlikely to be proteolyzed after severing). Interestingly, when we overexpressed Spastin in the embryonic or larval CNS, we did not observe an obvious alteration of the axonal microtubule architecture (see also Figure 8). This suggests that Spastin may be unable to disassemble stable axonal microtubule bundles. Nevertheless, the dramatic effects of Spastin overexpression on muscle microtubules suggest that the embryonic CNS collapse phenotype conferred by neuronal overexpression (see Figure 3) may also arise from breakdown of key neuronal microtubules during the axonal growth phase. Microtubule Bundles Are Depleted in Distal NMJ Boutons of spastin LOF Mutants The finding that Spastin overexpression erases the microtubule network in muscles suggested that the spastin LOF NMJ phenotypes could arise from alterations in microtubule networks. To investigate this, we first examined the distribution of Futsch, a microtubule-associated protein related to vertebrate MAP1B (Hummel et al. 2000; Roos et al. 2000). Futsch staining is restricted to stable neuronal microtubule bundles. Because Futsch is not expressed in the underlying muscle, Futsch antibody staining provides the optimal method for quantitatively evaluating stable microtubules within NMJ boutons. At muscle 4 NMJs in wild-type larvae (Figure 8A) we observed continuous microtubule bundles stained by anti-Futsch (mAb 22C10; green) within axons and along the axis of each branch of the NMJ (delineated by anti-HRP, which labels neuronal membranes; magenta). The intensity of Futsch staining weakens in the distal portions of the branches. Consistent with earlier findings, we also observed distinctive “loops” of Futsch staining within some boutons (Roos et al. 2000; Packard et al. 2002; Pennetta et al. 2002). Loops were typically observed in terminal boutons at the ends of branches. In some terminal boutons, however, we detected only punctate staining or no staining at all. This last case may reflect the limits of detection rather than the complete absence of Futsch protein in a bouton. We quantified Futsch distribution by dividing the patterns of Futsch staining in boutons into three classes: continuous (bundles or splayed bundles), looped, and diffuse or undetectable. In spastin5.75 larvae (Figure 8B), there was a shift in the Futsch pattern toward less organized morphologies (i.e., diffuse/undetectable). At the muscle 4 NMJ of spastin mutant larvae, 74% and 54% as many boutons contained continuous and looped Futsch, respectively, as compared to WCS. In contrast, 63% more boutons in spastin mutants displayed only diffuse or no staining (Figure 8D). These differences were most pronounced at the distal ends of the synaptic branches (Figure 8B, arrows and arrowheads). 65% of terminal boutons in mutants had no detectable Futsch staining, as compared to 8% in WCS (Figure 8E; p < 2 × 10−6). The Futsch distribution phenotypes were rescued by expression of Spastin from the spin-GAL4 or the RU486-induced, neural-specific Elav-GS-GAL4 drivers. Rescued larvae had Futsch staining patterns very similar to those seen in WCS (Figure 8C–8E). These results show that the reduction in stable synaptic microtubules seen in spastin LOF mutants is due to loss of Spastin from neurons during larval development. We also examined Futsch staining in Elav-GS-GAL4/UAS-spastin larvae grown on RU486 food, but saw no difference from the pattern in wild-type controls. Thus, stable microtubules at the NMJ do not break down when Spastin is overexpressed at the levels induced by this driver. Having demonstrated statistically significant differences in Futsch localization between control and spastin mutant NMJs, we then directly examined tubulin in muscle 4 NMJ boutons using fixation conditions that reduce muscle microtubule staining (see Materials and Methods). The pattern of NMJ microtubules is complex, and is difficult to quantitatively analyze because of residual signal from microtubules in the underlying muscle. However, using the α-tubulin antibody described above, we were able to clearly visualize looped microtubule structures (Figure 9A, green) within anti-HRP-labeled boutons (magenta). These loops were present both along the branches and in the terminal boutons (inset). spastin5.75 NMJs exhibited weaker α-tubulin staining than wild-type controls, particularly in terminal boutons (Figure 9B). Looped microtubule structures could be seen in some boutons in mutants. However, boutons at the ends of NMJ branches or at the outer edges of the bouton clumps that are characteristic of spastin NMJs often lacked any tubulin staining (inset). Thus, our results indicate that microtubule bundles are selectively depleted from the distal boutons of NMJs in larvae lacking Spastin protein. Discussion Mutations in the human spastin gene, which encodes an AAA ATPase, are the most common cause of pure AD-HSP. We identified the Drosophila Spastin ortholog (see Figure 1) in a gain-of-function screen (see Figure 3). Drosophila Spastin is a cytoplasmic protein that can also localize to axons (see Figure 2). spastin-null larvae have altered NMJs in which presynaptic boutons are more numerous and smaller than in wild-type, and are organized in dense clusters (see Figure 4). These changes in bouton number and organization are rescued by expression of Spastin in neurons (see Figures 4 and S1). QC, a measure of the number of vesicles of neurotransmitter released in response to an action potential, is reduced at NMJs in both null and hypomorphic spastin mutants (see Figure 5). spastin-null flies have severe movement defects. They cannot fly at all, and do not jump. They climb and walk very slowly, often drag their hind legs when walking (see Video S1), and have greatly reduced lifespans. spastin hypomorphs have milder phenotypes, displaying flying defects and a decrease in climbing speed (see Figure 6). Regulation of Synaptic Microtubule Networks by Spastin The AAA domain of Spastin is quite similar to that of Katanin-60, which is a microtubule-severing protein. To determine whether Spastin might also sever or otherwise alter microtubules in vivo, we overexpressed the protein in embryonic and larval muscles. Strikingly, this overexpression erases or greatly reduces the microtubule network (see Figure 7). These data are consistent with the finding that overexpression of human Spastin in transfected mammalian cells causes microtubule disassembly (Errico et al. 2002; McDermott et al. 2003). Having demonstrated that Spastin can cause disassembly of microtubules in vivo, we then examined how its absence affects the synaptic microtubule cytoskeleton. Based on the overexpression phenotype, one might have expected that microtubules would be more stable or more numerous in spastin LOF mutants. However, our observations indicate the opposite: microtubule bundles are depleted in NMJ boutons when Spastin is absent. At the wild-type muscle 4 NMJ, boutons are arranged along linear axes. Continuous microtubule bundles run along the axes and connect to larger bundles within the innervating axon. Microtubules within boutons are typically arranged in loops and swirls. In spastin-null mutants, boutons are arranged in clumps, and the distal boutons of these clumps often lack any detectable tubulin staining (see Figure 9). Looped microtubule structures are present within some proximal boutons, however, and the bundles connecting the NMJ to the axon are still present. These results suggest that the absence of Spastin selectively affects the construction of the presynaptic microtubule cytoskeleton, and that the severity of the microtubule defects in a bouton are correlated with its distance from the NMJ's axonal branchpoint. We quantitated these defects using an antibody against the microtubule-associated Futsch protein, which defines a subpopulation of stable neuronal microtubule bundles. In wild-type larvae, Futsch staining forms continuous lines along the main branches of the NMJ. Some individual boutons have Futsch loops, while others display only diffuse staining. A comparison of wild-type and spastin-null larvae shows that the distribution of Futsch within boutons shifts from organized structures (bundles and loops) toward diffuse patterns or the absence of detectable staining. This effect is most pronounced at terminal boutons, and is rescued by neuronal expression of Spastin (see Figure 8). If Spastin's function in vivo is to disassemble microtubules, as suggested by our overexpression experiments (see Figure 7), why does its absence produce a paradoxical reduction in microtubules within the NMJ (see Figures 8 and 9)? One possibility is that microtubule severing is required for movement of microtubules into or within the presynaptic region. Some evidence for this idea has been published. In one study, injection of function-blocking anti-Katanin-60 antibody into cultured sympathetic neurons reduced process outgrowth, and microtubules were 4- to 5-fold longer in antibody-injected neurons than in control cells (Ahmad et al. 1999). More recent work demonstrated that expression of dominant-negative Katanin-60 reduces axonal outgrowth (Karabay et al. 2004). These results were interpreted as indicating that Katanin is required for severing microtubules to a length that allows their transport along the axon to its growing tip. When Katanin is inhibited, microtubule segments may be too long to be efficiently transported, and this results in a reduction in axon outgrowth. Based on these findings, we suggest that the depletion of microtubules in the distal boutons of spastin mutant NMJs arises because severing of axonal microtubules by Spastin is necessary to generate microtubule polymers that are short enough to be efficiently moved into and through the presynaptic terminals. Perhaps Spastin normally excises sections of microtubules at branchpoints where NMJ branches leave the axon trunk, and these severed microtubule segments (or individual tubulin dimers) are then moved distally into the boutons of the NMJ as it grows. Is Spastin also involved in axon outgrowth or guidance, as suggested by its embryonic gain-of-function phenotype (see Figure 3)? Clearly loss of Spastin activity in Drosophila does not strongly affect outgrowth, since the embryonic CNS axon ladder develops in a normal manner and motor axons reach their appropriate targets in spastin mutants. Furthermore, axonal and muscle microtubules are not detectably altered in spastin-null embryos. Severing of microtubules in vivo, however, may usually involve the actions of multiple severing proteins. In addition to Spastin, the Drosophila genome encodes three AAA ATPases whose AAA domains are closely related to that of vertebrate Katanin-60. These are Katanin-60, CG1193, and an ortholog of mammalian Fidgetins, CG3326 (see Figure 1C). None of these proteins have been genetically characterized. Of these four proteins, Spastin is most distant from vertebrate Katanin-60, yet we have shown that Spastin overexpression causes microtubule disassembly in vivo (see Figure 7). Thus, our results suggest that all four fly proteins are microtubule-severing enzymes or proteins that otherwise facilitate disassembly of microtubule networks. Perhaps each is dedicated to severing microtubules in particular cellular and subcellular contexts, and their functions may be partially redundant. If so, generation of severe phenotypes in which microtubule networks are disrupted might require loss of two or more of these AAA ATPases. In mammals, Katanin and Spastin are both expressed in CNS neurons (Wharton et al. 2003; Karabay et al. 2004), consistent with the idea that they could have overlapping functions. After this manuscript was submitted for initial review, a paper appeared on perturbation of Drosophila spastin using transgenic RNAi techniques (Trotta et al. 2004). In direct contrast to our results, this paper concluded that (1) spastin is an essential gene (since crossing spastin RNAi flies to a ubiquitous GAL4 driver line was reported to produce lethality), (2) spastin RNAi larvae have reduced NMJs and an increase in synaptic transmission, and (3) loss of Spastin from neurons produces an increase rather than a decrease in stable microtubules in the NMJ. The conclusions in our paper are based on phenotypic analysis of spastin mutations that delete part or all of the coding region and on rescue of null mutant phenotypes by neuronal expression from a transgene. Our results show that spastin is not an essential gene: even spastin-null flies can eclose and live for several days, and spastin hypomorphs, which would be expected to more closely resemble most RNAi-perturbed flies, eclose at normal rates and have lifespans and behavior that do not greatly differ from wild-type (see Figure 6). We also determined that the 17-7 mutation, which removes more than one-third of the coding region, produces no detectable alterations in bouton number or NMJ microtubules and slightly decreases synaptic transmission, while spastin-null mutants have more boutons than wild-type larvae, a reduction in NMJ microtubule bundles, and more severely reduced transmission (see Figures 4, 5, 8, and 9). In most transgenic RNAi work in Drosophila, different transgenic lines yield phenotypes that range from hypomorphic to near-null, and transgenic RNAi does not completely eliminate expression of the target protein (e.g., Billuart et al. 2001; Kalidas and Smith 2002). In the Trotta et al. paper, it is unclear whether more than one transgenic RNAi line was analyzed, but RNAi is described as reducing the level of Spastin protein expression by less than 4-fold. Our findings on Spastin hypomorphic phenotypes imply that such RNAi larvae would not have morphological or microtubule bouton phenotypes and that adult flies would be relatively healthy. We do not understand the origin of the discrepancies between the two sets of results. Implications of Studies of Drosophila Spastin for the Understanding of Human AD-HSP spastin-null adult flies have severe movement defects. Their hind legs are particularly weak (see Video S1); this is interesting in light of the restriction of symptoms to the legs in most AD-HSP patients. Other aspects of the spastin mutant adult phenotypes also resemble observations made in human AD-HSP patients. The penetrance of the human spasticity phenotype is highly variable, so that some individuals carrying a spastin mutation appear unaffected, while others with the same mutation are confined to wheelchairs. In our experiments, we observed that some spastin hypomorphs exhibit normal flying behavior in a cylinder assay, while most fail to fly and crash into the cylinder base (see Figure 6A and 6B). Half of the spastin-null adults die within 4 d, but most of the survivors then live more than 25 d (see Figure 6E). The selective male lethality we observed (see Figure 6G) is also interesting in light of the discovery of a large SPG4 pedigree in which only males exhibit AD-HSP phenotypes (Starling et al. 2002). Despite these apparent parallels, there is no evidence at present that the fly behavioral phenotypes arise through mechanisms related to those that cause human AD-HSP. The anatomy of the Drosophila nervous system is quite different from that of the mammalian spinal cord. Furthermore, AD-HSP is thought to be a neurodegenerative disease that progresses over a period of years, and it is unclear whether neurodegeneration as a result of spastin mutations could occur during the short lifespan of Drosophila. Further work will be required to determine to what extent the Drosophila system can provide an organismal model for AD-HSP pathology. The clearest implications of our work for AD-HSP emerge from the analysis of the cellular phenotypes arising from loss of Spastin. We show that the absence of Spastin alters the microtubule network at nerve terminals. Microtubule bundles are depleted in the distal boutons of the NMJ, which is a glutamatergic synapse that resembles excitatory synapses within the mammalian spinal cord. These results suggest that microtubules within the terminals of neurons in the human spinal cord could also be disordered or absent in patients with AD-HSP. The terminals of these neurons might eventually degenerate as a consequence of these microtubule defects, leading to a selective loss of distal axon segments within the spinal cord. Materials and Methods Genetics and molecular biology For the overexpression screen, approximately 6,000 new EP insertion lines were generated by crossing an X chromosome EP line, EP55, to a SbΔ2-3 transposase line. EP insertions on Chromosome II or III were crossed to the pan-neuronal driver line ElavC155-GAL4 and the muscle driver line 24B-GAL4. Embryos from EP lines that exhibited less than 20% viability in combination with either driver were immunostained with mAb 1D4. Eighteen lines had embryonic CNS and/or motor axon defects when crossed to ElavC155-GAL4, including line T32. The flanking genomic region of T32 was cloned and sequenced by plasmid rescue. This sequence matched three overlapping EST's from BDGP, one of which, GH11184, contained the complete ORF of the gene downstream of the T32 insertion. This cDNA was sequenced in its entirety. Unrooted trees for Spastin and its closest relatives were constructed using six different algorithms (fitch, kitsch, neighbor, upgma, protein maximum likelihood, and parsimony) from the Phylip package, based on alignment to the PFAM AAA consensus. For the spastin excision lines, alleles 10-12, 17-7, and 5.75 were generated via imprecise excision of EP T32 using SbΔ2-3. All alleles were homozygous viable, and their deletions were mapped by PCR and sequencing of larval or adult genomic DNA. Allele 5.75 causes sterility in both sexes. For the spastin rescue construct, the UAS-spastin cDNA construct was made by subcloning a 2.9-kb BglII fragment from GH11184 into the BglII site of pUAST (Brand and Perrimon 1993). This fragment contains the spastin cDNA up to 350 bp after the stop codon (excluding 681 bp of the 3′ UTR) and including 28 bp of polylinker sequence from the pOT2 plasmid at the 5′ end. The construct was injected at approximately 300 ng/μl into KiΔ2-3 embryos and several transgenic lines recovered; experiments described here used the Chromosome II insertion line 8-3-5. Rescue of spastin-null phenotypes by spin-GAL4-driven expression was assayed by crossing UAS-spastin/CyOKr-GFP; spastin5.75/TM3SerAct-GFP to spin-GAL4/CyOKr-GFP; spastin5.75/TM3SerAct-GFP. The numbers of Ib boutons in rescued larvae (UAS-spastin/spin-GAL4; spastin5.75/spastin5.75) were compared to those in unrescued sibling mutants (spin-GAL4/CyOKr-GFP; spastin5.75/spastin5.75) to calculate the ratios used to determine the efficacy of rescue (see Figure 4G). This was done because we observed that the presence of the driver chromosome in the background increased the absolute number of Ib boutons, so that the appropriate control was to the sibling mutants also bearing this chromosome. Rescue of spastin-null phenotypes by postembryonic Elav-GS-GAL4-driven expression was assayed by crossing UAS-spastin/CyOKr-GFP; spastin5.75/TM3SerAct-GFP to Elav-GS-GAL4, spastin5.75/TM6B and raising the larvae on RU486-containing food as described (Osterwalder et al. 2001; McGuire et al. 2004). The numbers of Ib boutons in rescued larvae (UAS-spastin/+; Elav-GS-GAL4, spastin5.75/spastin5.75) were compared to those in unrescued sibling mutants (CyOKr-GFP/+; Elav-GS-GAL4, spastin5.75/spastin5.75) as described above for spin-GAL4 rescue. The effect of postembryonic neuronal overexpression of Spastin in a wild-type background was assayed by counting Ib bouton numbers in UAS-spastin/+; Elav-GS-GAL4, spastin5.75/TM3SerAct-GFP larvae from this same cross and comparing them to the numbers for UAS-spastin/+; Elav-GS-GAL4, spastin5.75/spastin5.75 . The wild-type control lines used were Oregon R, w1118, and WCS, a Canton S line backcrossed ten times to white (gift of Anne Simon). For larval overexpression experiments, the MHC-GS-GAL4213-3 driver was used to induce muscle-specific expression of UAS-spastin at 23 °C in the absence of RU486. Expression induced by this driver was confirmed in a separate cross using UAS-EGFP. Generation of Spastin antibodies Regions of the spastin cDNA encoding aa 136–416 (pGEX-T32PvuII), 1–167 (pAcG2T-T32BamRIa), and 380–758 (pAcG2T-T32BBA) were subcloned from GH11184 into bacterial (pGEX) or baculovirus (pAcG2T) expression vectors. Expressed protein (in the form of inclusion bodies for T32PvuII) was injected into guinea pigs (Covance, Princeton, New Jersey, United States) and the antiserum tested on en-GAL4/+2/+ embryos. Antisera against T32PvuII (86EX) and T32BBA (1239) both showed strong staining in the Engrailed pattern; thus, both recognized overexpressed Spastin. 86EX was purified by incubation with membrane-bound pGEX-T32PvuII protein and subsequent elution with 100 mM glycine (pH 2.5) followed by neutralization with 3M Tris (pH 8.8). pAb1239 was affinity purified using the immunogen bound to Affi-gel10 beads (Bio-Rad Laboratories, Hercules, California, United States), followed by preabsorption with spastin5.75 larval fillets. Embryonic and larval immunocytochemistry Stage 16 embryos were fixed and stained using standard methods (Patel 1994) with mAb 1D4 (1:5) or anti-β3-tubulin (1:500; a gift of R. Renkawitz-Pohl and D. Buttgereit). Third instar larvae were live-dissected in room temperature HL3 (see below) or PBS and fixed in 4% paraformaldehyde or Bouin's fix (for staining of NMJs with anti-tubulin) for 25 min. Primary antibodies used on larvae were mouse anti-Syt at 1:400 (Menon and Zinn 1998), rabbit anti-Dlg (1:2,000; a gift of D. Woods and P. Bryant), mAb 22C10 (anti-Futsch, 1:50; Developmental Studies Hybridoma Bank, Iowa City, Iowa, United States), rabbit anti-HRP (1:250, Cappel, MP Biomedicals, Irvine, California, United States), mouse anti-α-tubulin DM1A (1:500; Sigma, St. Louis, Missouri, United States). Anti-Spastin pAb 1239 was used at 1:300. Staining was visualized with HRP-conjugated goat anti-mouse secondary (1:200; Jackson Laboratory, Bar Harbor, Maine, United States) or Alexa-Fluor 488 and 568 anti-mouse, -rabbit, or -guinea-pig secondaries (1:200; Molecular Probes, Eugene, Oregon, United States). All fluorescently labeled samples were imaged by acquiring z-series projections with a Zeiss (Oberkochen, Germany) 510 inverted confocal microscope and 63×/1.4 n.a. or 100×/1.2 n.a. PlanApo objectives. Only larval segments A2 and A3 were analyzed. Individual boutons were defined as a Syt-positive area encircled by Dlg-positive staining (see Figure 4), or a Syt- or HRP-positive varicosity in the synaptic arbor (see Figure 8). All type Ib boutons were scored for each muscle 4 NMJ. Average bouton numbers per muscle 4 (see Figure 4) were as follows (mean ± s.e.): WCS, 44 ± 2.6 (n = 26 NMJs); spastin5.75 mutant, 68 ± 2.4 (38); spin rescue, 50 ± 4.9 (26); mutant sibling with spin driver chromosome, 86 ± 4.6 (8); Elav-GS rescue, 70 ± 3.0 (19); and mutant sibling with Elav-GS driver chromosome, 110 ± 9.3 (21). p < 0.02 by one-way ANOVA for all paired comparisons. Electrophysiology Intracellular recordings were obtained at 18 °C, using sharp microelectrodes (boroscilicate glass, 1.0 mm OD; 18–35 MΩ resistance; World Precision Instruments Sarasota, Florida, United States) filled with 3M KCl, from body wall muscle 6 (segments A3 or A4) of filleted third instar larvae, following standard methods (Jan and Jan 1976). Larvae were bathed in HL3 solution (Stewart et al. 1994), in mM: NaCl, 70 (EM Science, Gibbstown, New Jersey, United States); KCl, 5; MgCl2, 20; NaHCO3, 10; HEPES, 5; Sucrose, 115; Trehalose, 5; and CaCl2, 1 (Sigma). Larvae were visualized with a 5×/0.10 n.a. Olympus (Tokyo, Japan) objective on an Olympus BX50WI microscope. EJPs were evoked by pulling the cut end of the innervating segmental nerve into a heat-polished suction electrode and passing a depolarizing pulse sufficient to depolarize both motoneurons (Grass SD9 stimulator). For each experiment, 10–15 single EJPs evoked at 0.2 Hz were recorded, and then spontaneous mEJPs recorded for 1 min afterwards. Only recordings with resting membrane potential below −60 mV were acquired. The average resting membrane potential for control (WCS) larvae was −72.2 mV, and did not differ significantly from any of the experimental groups. Average muscle input resistance in control larvae was 8.9 MΩ, and differed significantly only from the input resistance determined for spastin 5.75 /spastin17-7 transheterozygotes (7.5 MΩ p < 0.04). Recordings were performed using an Axon Instruments (Foster City, California, United States) Axopatch 200B amplifier with CV203BU headstage operating in current clamp mode. The signal was low-pass filtered at 5 kHz, digitized through an Axon Instruments Digidata 1322A 16-bit acquisition system, and recorded using Axon Instruments Clampex 8.2 software. Mean EJP amplitude was determined by averaging all single EJPs with Axon Instruments Clampfit 8.2, and corrected for nonlinear summation according to McLachlan and Martin (1981) and Feeney et al. (1998). mEJPs were measured using Mini Analysis Program (Synaptosoft, Decatur, Georgia, United States). mEJPs with a slow time course arising from neighboring electrically coupled muscle cells were excluded from analysis (Zhang et al. 1998). QC for a given NMJ was estimated by dividing the average EJP amplitude by the average mEJP amplitude. Statistics were calculated using one-way ANOVA. Adult phenotypes Eclosion rates were determined by counting numbers of empty versus full (dead) pupae on the sides of bottles in which flies had been allowed to lay for comparable time periods. The flight test assay was performed at room temperature using an opaque cylinder (a 52-cm-tall pipette washer, 18 cm in diameter) coated on the inside with fresh mineral oil. Flies of a given genotype were dumped through a hole in the center of a lid at the top. The cylinder was divided into bins along its height, and the number of flies per bin counted. Flies of different genotypes were age-matched; more than 200 flies were counted for each. The climbing assay was performed on 4–5 d old flies maintained individually in vials. Climbing velocity for each fly was measured by transferring it to an empty vial, banging it to the bottom, and then measuring either the time required to reach the top of the vial or the maximum distance it climbed in 30 s, whichever came first. Three trials were performed, and the best speed was used. For lifespan tests, flies were maintained at 25 °C, transferred every 3 d to fresh food vials, and their lifespan noted. Supporting Information Figure S1 Spastin Expression in Neurons Rescues the spastin Mutant Morphology Representative muscle 4 NMJs stained with antibodies against Dlg (green) and Syt (magenta) are shown for (A) spastin5.75 mutant (genotype +/CyOKr-GFP; Elav-GS-GAL4,spastin5.75 /spastin5.75 ), (B) neuronally rescued (UAS-spastin /CyOKr-GFP; Elav-GS-GAL4,spastin5.75 /spastin5.75), and (C) neuronally overexpressing (UAS-spastin /+; Elav-GS-GAL4,spastin5.75 /TM3Ser-ActGFP) larvae. The clustered, smaller, and more numerous boutons observed in mutant NMJs (A, arrowhead) are absent in neuronally rescued larvae, which resemble controls (WCS; see Figure 4D). Spastin overexpression in neurons produces an opposite morphological phenotype compared to the loss of function: boutons appear slightly larger than in wild-type, and bouton counts show that they are reduced in number (83% of control; see text). Scale bar, 5 μm. (1.4 MB TIF). Click here for additional data file. Figure S2 Adult Behavior Is Partially Rescued by spin-GAL4-Driven Expression of Spastin in the spastin-Null Background Behavioral tests were performed on flies from the four genotypes arising from the spin-GAL4 rescue crosses, raised at 18 oC. These genotypes were (1) spin-GAL4/UAS-spastin; spastin5.75 (spin Rescue), (2) spin-GAL4/CyOKr-GFP; spastin5.75 (non-rescued spastin mutant, denoted 5.75[R]), (3) spin-GAL4/UAS-spastin; spastin5.75 /TM3SerAct-GFP (Cy+ Ctrl; heterozygous for the spastin mutation), and spin-GAL4/CyOKr-GFP; spastin5.75/TM3SerAct-GFP (Cy Ctrl; heterozygous for the spastin mutation). (A) Climbing behavior. None of the spastin mutants (0%) from these crosses (5.75[R]; n = 21) reached the top of the vial in the prescribed 30 s time limit, compared to 8% for Rescue flies (n = 75), and 100% for both spastin/+ controls (n = 39 and 21). Twenty-seven percent of mutants (5.75[R]) did not climb at all, compared to only 4% of the Rescue flies and 0% of the spastin/+ controls. Thus, although both genotypes in the mutant background (homozygous for spastin5.75 ) were much weaker than either spastin5.75 heterozygous control, Rescue flies showed improved climbing ability compared to the mutants. (B) Similar to the results in (A), mean lifespan in spastin mutants (10 ± 1.3 d, n = 32) was significantly rescued by spin-driven expression of spastin (16 ± 1, n = 95, p < 0.004), although lifespans were much shorter in spastin5.75 homozygotes than in heterozygous spastin/+ controls (43 ± 3.1 and 44 ± 2.7; n = 20 each). (218 KB PDF). Click here for additional data file. Video S1 Motor Behavior in Control, spastin5.75 Mutant, and spin-GAL4-Rescued Flies Flies are shown moving in a vial. Segment 1: Wild-type. One female and one male WCS fly are shown. Note the rapid rate at which they walk, as well as exhibiting climbing, jumping and flying behaviors. When still, their legs are controlled, and they are able to walk upside-down (out-of-focus fly near end of segment) for prolonged periods without falling. Segment 2: Mutant. One spastin5.75 female is shown. Leg weakness is obvious, particularly for the mesothoracic and metathoracic legs, both when walking and standing still. She climbs poorly, and when rotated so that she is upside-down, is unable to maintain a hanging position. No wing movement or jumping is observed. Segment 3: Rescue. In spin-GAL4/UAS-spastin; spastin5.75 flies, Spastin expression via the spin-GAL4 driver partially rescues the movement defects seen in spastin5.75 mutants. Two males are shown, followed by one female. Note their improved leg steadiness, velocity, and hanging ability. These flies can also jump spontaneously. The female appears to be less fully rescued; however, she is able to walk upside-down for prolonged periods, and exhibits wing movement. (7.4 MB MOV). Click here for additional data file.
Author and article information
Journal
Journal ID (nlm-journal-id): 101573691
Journal ID (pubmed-jr-id): 39703
Journal ID (nlm-ta): Cell Rep
Journal ID (iso-abbrev): Cell Rep
Title:
Cell reports
ISSN
(Electronic):
2211-1247
Publication date Nihms-submitted: 11
April
2016
Publication date
(Electronic):
03
September
2015
Publication date
(Print):
22
September
2015
Publication date PMC-release: 20
April
2016
Volume: 12
Issue: 11
Pages: 1723-1730
Affiliations
[1
]Department of Neurobiology and Anatomy, Drexel University College of Medicine, Philadelphia,
PA 19129, USA
Author notes
Article
Manuscript ID: NIHMS763086DOI: 10.1016/j.celrep.2015.08.017
PMC ID: 4837332
PubMed ID: 26344772
SO-VID: a70d0af6-a702-470c-82a3-aff817657b21
License:
This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/).
History
Categories
Subject:
Article
ScienceOpen disciplines: Cell biology
Data availability:
ScienceOpen disciplines: Cell biology