futsch : Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References

Gene name - futsch

Synonyms - 22C10

Cytological map position - 1F-2A

Function - cytoskeletal component

Keywords - cytoskeleton, axon guidance

Symbol - futsch

FlyBase ID: FBgn0259108

Genetic map position -

Classification - Microtubule-associated protein

Cellular location - cytoplasmic



NCBI links: Entrez Gene | UniGene

Recent literature
Romano, M., Feiguin, F. and Buratti, E. (2016). TBPH/TDP-43 modulates translation of Drosophila futsch mRNA through an UG-rich sequence within its 5'UTR. Brain Res [Epub ahead of print]. PubMed ID: 26902497
Summary:
Nuclear factor TDP-43 is an evolutionarily conserved multifunctional RNA-binding protein associated with Frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS). In recent years, Drosophila models of ALS based on TDP-43 knockdown/overexpression have allowed to find several connections with disease. Among these, previous studies have described that silencing the expression of its fly ortholog (TBPH) can alter the expression of the neuronal microtubule-associated protein Futsch leading to alterations of neuromuscular junction (NMJ) organization. In particular, TBPH knocked out flies displayed a significant reduction of Futsch protein levels, although minimal variation in the futsch mRNA content was observed. These conclusions were recently validated in an independent study. Together, these observations strongly support the hypothesis that TBPH might regulate the translation of futsch mRNA. However, the mechanism of TBPH interference in futsch mRNA translation is still unknown. This work used EMSA experiments coupled with RNA-protein co-immunprecipitations and luciferase assays to show that TBPH interacts with a stretch of UG within the 5'UTR of futsch mRNA and translation is positively modulated by this binding. Most importantly, this function is also conserved in human TDP-43. This result can therefore represent the first step in elucidating the relationship between TDP-43, protein translation, and eventual disease onset or progression.
BIOLOGICAL OVERVIEW

Futsch protein is recognized by the long-known monoclonal antibody 22C10, identified in the laboratory of Seymor Benzer (Fujita, 1982; Zipursky, 1984). The 22C10 antigen is expressed by some CNS neurons as well as by all neurons in the PNS. The antigen is found in all cellular compartments, the dendrite, the soma, and the axon, where it is associated with the axonal cytoskeleton. In a large-scale EMS mutagenesis, five noncomplementing X-chromosomal mutations were identified that strongly reduce or eliminate the 22C10 antigen expression. The corresponding gene was termed futsch (German for 'gone'). Genetic analyses demonstrate that futsch is necessary for axonal growth and dendritic morphology. The deduced Futsch protein is 5327 amino acids in length. The N- and C-terminal domains of Futsch are homologous to the vertebrate MAP1B, whereas the central domain of the Futsch protein is highly repetitive. Futsch localizes to the microtubule compartment of the cell and can be precipitated with taxol-stabilized bovine microtubules in vitro. Thus, futsch encodes a novel cytoskeletal protein (Hummel, 2000).

Neuronal cell morphology requires a precisely regulated cytoskeleton, comprising intermediate filaments, F-actin, and microtubule fibers. In the axon, microtubules are organized in a polar fashion, with the plus end always pointing to the synapse, whereas in the dendrites, both orientations can be found. The regulation of microtubule dynamics depends on a large group of microtubule-associated proteins (MAPs) that have been classified into three classes (referring to mammalian proteins and in many cases to their Drosophila homologs). Kinesin-type proteins, dynamin, MAP2c, and tau comprise the low molecular weight group of MAPs. The intermediate group is comprised of the MAP3 and MAP4 proteins. The high molecular weight MAPs are MAP1A, MAP1B, MAP2, and cytoplasmic dynein (Hummel, 2000 and references therein).

A role for the nonmotor MAPs in the establishment of neuronal cell shape has been deduced from their distinct association with different neuronal cell compartments. MAP2 is preferentially found in dendrites and neuronal somata. In contrast, tau is localized in axons. The MAP1 proteins are found in dendrites, somata, and axons. The correlation of MAP expression and axonal growth in developing as well as in regenerating nerves suggests a role in axonal extension. The first evidence supporting this assumption stems from antisense experiments. Antisense RNA directed against tau blocks axonal outgrowth in cultured neurons. Similarly, depletion of MAP1B or MAP2 RNA by antisense RNA or by intracellular delivery of specific antibodies results in reduction or inhibition of neurite outgrowth in cultured cells. The unilateral ablation of phosphorylated MAP1B in growth cones using microscale chromophore-assisted laser inactivation (micro-CALI) induces changes in growth direction. The results indicate a functional role for P1-MAP1B in local growth cone stabilization and thus growth cone steering (Hummel, 2000 and references therein).

The in vivo function of the MAPs during neuronal development is not yet clear. MAP1B knockout experiments, carried out with mice, have yielded conflicting data. Genetic knockout data indicates that MAP1B is required for embryonic viability, and even heterozygous animals show mutant phenotypes, such as slower growth rates and motor system abnormalities (Edelmann, 1996). These results are very similar to the ones observed in the Drosophila mutation futschP158. Another MAP1B knockout mutation (Takei, 1997) does not affect all MAP1B splice variants, and residual MAP1B protein expression can still be detected (Kutschera, 1998). This hypomorphic MAP1B mutation results in homozygous viable mice with delayed nervous system development similar to the phenotype seen in mutant futschK68 or futschN94 animals. Another MAP1B gene knockout (Edelmann, 1996) also results in a truncation of the gene leaving a 571 amino acid N-terminal domain intact, for which a dominant-negative activity has been postulated (Takei, 1997). To test whether expression of a similar truncated Futsch protein leads a dominant-negative phenotype, a UAS-mini-futsch gene was constructed encoding a comparable Futsch protein consisting of only the first 527 amino acids. High levels of expression of this protein variant do not interfere with normal development, supporting the notion that the N-terminal domain of MAP1B has no dominant-negative effect. Rather, the N-terminal Futsch domain has weak rescuing abilities (Hummel, 2000 and Roos, 2000).

Evidence is presented that Futsch controls synaptic growth at the Drosophila neuromuscular junction (NMJ) through the regulation of the synaptic microtubule cytoskeleton. One mechanism for the addition of new synapses to a nerve terminal involves the sprouting of new synaptic boutons (Zito, 1999). Division of a synaptic bouton into thirds (or greater subdivisions) may be a mechanism for generating a branchpoint in the nerve terminal (Zito, 1999). Nerve-terminal plasticity achieved through processes such as nerve-terminal sprouting require precisely controlled and spatially regulated modifications to the nerve-terminal cytoskeleton. The cytoskeletal rearrangements that drive bouton division and the signaling events that control this process are not known. In principle, localized remodeling of the cytoskeleton could achieve the spatial precision required for plasticity that is both synapse-specific and activity-dependent. Specificity could be further refined if only a small subset of synapses, at any given time, has the capacity for cytoskeletal rearrangement (Roos, 2000 and references therein).

Regulation of microtubule organization within the growth cone may be an important determinant of growth cone motility. The macroscopic organization of microtubules within the growth cone is dynamic, as revealed by the formation and destruction of hairpin loop structures that incorporate a large portion of the growth cone microtubules (Tanaka, 1995; Dent, 1999). Of particular interest is the demonstration that the formation of microtubule hairpin loops is correlated with the cessation of growth cone motility: the opening (or destruction) of these loops is correlated with the reestablishment of motility (Dent, 1999). These data are strengthened by the observation that the transition of a motile growth cone to a stable synaptic contact is also correlated with the formation of a hairpin microtubule loop within the growth cone as it transforms into a synapse (Tsui, 1984). Thus, regulated microtubule architecture appears to be an essential element in the control of growth cone morphology and motility as well as synapse formation. The molecular regulation of these microtubule based structures, however, is not known (Roos, 2000 and references therein).

Futsch associates with microtubules in vitro, and a genetic analysis demonstrates that Futsch is necessary for axonal and dendritic growth. Futsch colocalizes with microtubules and identifies cytoskeletal loops that traverse the lateral margin of select synaptic boutons. An apparent rearrangement of microtubule loop architecture occurs during bouton division, and a genetic analysis indicates that Futsch is necessary for this process. futsch mutations disrupt synaptic microtubule organization, reduce bouton number, and increase bouton size. These deficits can be partially rescued by neuronal overexpression of a futsch MAP1B homology domain. Genetic manipulations that increase nerve-terminal branching correlate with increased synaptic microtubule loop formation, and both processes require normal Futsch function. These synaptic microtubule loops are highly reminiscent of microtubule loops present within the growth cone, implying a fundamental similarity between the mechanisms of growth cone motility and the mechanisms of synaptic growth and branching (Roos, 2000).

At the Drosophila NMJ, Futsch protein is associated with a cytoskeletal core within the synaptic terminal that is continuous with the axonal cytoskeleton. This cytoskeletal core is composed of multiple fibers, consistent with Futsch binding the microtubule cytoskeleton of the axon and nerve terminal. Examination of Futsch staining at the nerve terminal using deconvolution confocal microscopy reveals periodic loop structures within a subset of synaptic boutons (loops are present within 24% of boutons at the wild-type synapse on muscles 6 and 7 and 22% of boutons at muscle 4. These loops appear at stereotypic locations within the synapse at every abdominal muscle, though the focus for this analysis is placed on muscle 4. Futsch-positive loops are present within all types of synaptic boutons, including type 1b, type 1s, and type II. This analysis focuses on type 1b boutons (Roos, 2000).

Synaptic microtubule loops are highly enriched at points of nerve-terminal branching (~90% of observed branchpoints include a loop), and loops are always present within the terminal bouton(s) at the end of each chain of synaptic boutons. The absence of loops at some branchpoints may reflect an inability to visualize these structures perfectly within every synapse. Microtubule loops are not present at the sites of muscle innervation where the axon initiates contact with the muscle surface. Two axons, both originating from the segmental nerve, form type 1b terminals at muscle 4. These axons reach muscle 4 and migrate on the muscle surface in opposite directions. The site of innervation is not a branchpoint but rather is the site of muscle innervation, and no loops are present. This morphology is consistently observed (Roos, 2000).

3D reconstructions of synaptic boutons that contain Futsch-positive loops, double stained with antisynaptotagmin (anti-syt; vesicle protein) to delineate the boundary of the synaptic bouton, demonstrate that these structures are indeed loops rather than an artifact of confocal sectioning through a spherical structure. Nerve terminals were double stained for anti-Futsch (22C10) and anti-syt, and optical sections were taken through the z plane of the nerve terminal of muscle 4 at 0.1 µm sections. Select synaptic boutons containing loop structures were serially reconstructed using Delta-Vision deconvolution algorithms. Following reconstruction, loop-containing synaptic boutons can be rotated in three dimensions, demonstrating that the Futsch staining is a loop that traverses the widest diameter of the synaptic bouton and is entirely contained within the top and bottom (z plane) boundaries of the bouton as defined by synaptotagmin staining. Futsch-positive loops often extend beyond the anti-syt staining in the x and y dimension, indicating that the loop is closer to the plasma membrane than are diffuse synaptic vesicles in some areas of the bouton (Roos, 2000).

The localization of microtubules at the third instar synapse has been followed by visualizing tubulin (antitubulin immunoreactivity) or by following the localization of neuronally overexpressed tau-GFP. Futsch colocalizes with synaptic microtubules as identified by anti-alpha-tubulin. Colocalization is observed throughout the axon and synapse and always occurs at microtubule loops. Near perfect colocalization of Futsch and tubulin persists even when the microtubule organization is disrupted in hypomorphic futsch mutations, implying an association of Futsch with the microtubules. There is no immunoreactivity of Futsch with muscle microtubules, demonstrating that there is not antibody cross-reactivity in these double-staining experiments. Overexpressed tau-GFP also colocalizes with Futsch in synaptic cytoskeletal loop structures (Roos, 2000). These data are consistent with Futsch being associated with the microtubule cytoskeleton as suggested by in vitro data (Hummel, 2000). The supernatant from adult head extracts was incubated with polymerized microtubules at 37°C. Futsch is enriched in the subsequent pellet fraction in the presence of taxol-stabilized bovine microtubules. Thus, Futsch cofractionates with microtubules, and soluble Futsch can be precipitated with exogenously added vertebrate microtubules to adult fly head supernatants (Hummel, 2000).

Both viable mutations identified by Hummel (2000) alter synaptic growth. There is no detectable Futsch immunoreactivity in the futschK68 mutation. futschN94 is a hypomorphic mutation that reduces protein expression to ~20% wild-type levels (based on reduced fluorescence intensity). The futschN94 mutation also disrupts the subcellular localization of Futsch. In futschN94, Futsch immunoreactivity fills up the volume of every synaptic bouton rather than being restricted to a cytoskeletal core. Microtubule localization is also disrupted in these futsch mutations. In both futschN94 and futschK68, synaptic microtubules no longer form a filamentous cytoskeletal shaft that runs through the nerve terminal, and microtubule loop formation is absent. Rather, tubulin staining in these mutants is punctate and diffuse, filling up the volume of every synaptic bouton within the NMJ. The diffuse, punctate tubulin staining in these mutants is identical to the diffuse anti-Futsch (22C10) staining observed in futschN94. Double staining for Futsch and alpha-tubulin demonstrates that colocalization persists even when the microtubules are dispersed in these mutations. Finally, there is a qualitative disruption of both anti-Futsch and tubulin staining in futschN94/+ heterozygous larvae; microtubule loops are not as clearly defined at the lateral margin of select synaptic boutons. Taken together, these data demonstrate that futsch is necessary for microtubule organization within synaptic boutons. These data also support the conclusion that futsch is necessary for the formation or stabilization of microtubule-based loop structures (Roos, 2000).

Synaptic morphology is severely altered in viable futsch mutant backgrounds, indicating that Futsch-dependent microtubule organization is necessary for normal synaptic growth and development. In futschN94 and futschK68, there is a reduction in bouton number and an increase in bouton size. Bouton number is reduced from an average of 64.8 boutons at muscle 4 in wild type to 37.9 in futschN94 and 40.8 in futschK68 (p < 0.001). No change in muscle size is observed in these mutations. There is also a modest but statistically significant reduction in bouton number in futschN94/+ heterozygous larvae (51.2 boutons at muscle 4). This correlates with a qualitative disruption in microtubule organization at the NMJ in these heterozygous larvae. In parallel with the observed decrease in bouton number, the average bouton size and the distribution of bouton sizes within the synapse are dramatically increased in both futsch alleles. Mutant nerve-terminals at muscle 4 rarely have branchpoints. However, this may be a consequence of the drastic reduction in bouton number (Roos, 2000).

A remarkable feature of microtubule organization at the nerve terminal is the formation of loop structures within select synaptic boutons. A small number of microtubule loops appear periodically along the nerve terminal. These intraterminal loops are closer together the further out they are along a chain of boutons. The average interloop distance in the proximal half of the nerve terminal (closer to the site of innervation) is 12.1 µm, and this decreases to 4.4 µm in the distal half of the nerve terminal. In addition, examination of microtubule loops at the second instar synapse (~2 days earlier in development) reveals that loops are on average closer together and that the loops are composed of a narrower gauge filament of Futsch. This is consistent with loops forming during synaptic growth and then becoming stabilized structures within the nerve terminal. As the nerve terminal grows, more cytoskeleton is predicted to be added to the synapse, and the gauge within the loops will be increased. In addition, boutons that are formed early in development will be separated by larger distances due to the expansion of the synapse with the growth of the muscle. Thus, loops formed early in development will be present in the proximal half of the NMJ and are predicted to be separated by larger distances, as observed (Roos, 2000).

An analysis of loop morphology demonstrates that microtubules within a loop do not rejoin the major cytoskeletal strand within the nerve terminal. The diameter of the Futsch immunoreactivity 0.5 µm before and 0.5 µm after loop structures was measured. These measurements were compared with diameters taken at two points separated by 4 µm (larger than the average diameter of a loop) at various positions along the nerve terminal without an interposed loop. The diameter of the main shaft of Futsch immunoreactivity is significantly reduced at the distal side of a loop compared to the proximal side, whereas there is no change in the diameter of Futsch immunoreactivity over a similar distance of nerve terminal without an interposed loop. One essential role for Futsch as a microtubule-associated protein (MAP) could be to stabilize the microtubules at the free end of such a loop. The disruption of microtubule organization and loop formation in the futsch mutations support this hypothesis (Roos, 2000).

Live visualization of Drosophila neuromuscular synapses demonstrates that synaptic bouton division is a mechanism for bouton addition and branchpoint addition to the NMJ during development (Zito, 1999). This process is termed division as opposed to spouting because previously existing active zones are partitioned between newly formed boutons (Zito, 1999). Boutons that are predicted to be undergoing division can be identified by an irregular, hourglass-like shape (Zito, 1999). Examination of such bouton profiles demonstrates that these boutons not only contain microtubule loops, but these loops appear to be undergoing rearrangement. While dynamic rearrangement can only be proven by in vivo live observation, the irregular and highly variable branched microtubule structure is suggestive of an active process. Examination of Futsch staining within numerous putative dividing boutons suggests a sequence for the process of bouton division. At the earliest stages of division, Futsch strands reach across the center of a loop within a bouton. The strands of Futsch become more elaborate within boutons that have an hourglass-like shape, indicative of boutons that are midway through the process of division. In many cases, strands of Futsch protein cross the middle of the dividing bouton precisely at the site where cleavage is predicted. At the final stages of bouton division, two adjacent loops are produced in the neighboring newly divided boutons by the division of the original loop. As the synapse grows, these new adjacent loops stabilize and later separate from one another as the synapse expands on the muscle surface, generating the stereotyped periodicity of loops within the synapse (Roos, 2000).

Interestingly, microtubule loops are always observed to lie in the same plane as the muscle fiber surface. If Futsch is involved in the process of bouton division, then the plane of the microtubule loop may also determine the plane of bouton division. This would prevent bouton division from occurring into the volume of the muscle fiber, which is never observed at the wild-type synapse (Roos, 2000).

To support the conclusion that bouton division occurs at microtubule-based loops and is a mechanism of synaptic growth, advantage was taken of a genetic background that increases the number of microtubule loops at the synapse. flexin is a novel Drosophila muscle protein. Overexpression of flexin in muscle during postembryonic development causes a dramatic increase in nerve-terminal branching. There is a two-fold increase in the occurrence of microtubule loops at flexin overexpressing synapses (50% of boutons contain a loop at flexin overexpressing synapses as compared to 22% at wild-type synapses). This correlates with an approximate two-fold increase in nerve-terminal branch formation. Thus, flexin appears to act as a muscle-derived signal to increase nerve-terminal branching. flexin-induced branching is correlated with increased organization of presynaptic microtubule loop structures. Since Futsch is expected to act cell autonomously, elevated branching induced by flexin overexpression in muscle ought to be suppressed in a futsch mutant background. This was confirmed by the demonstration that muscle overexpression of flexin does not alter the futschN94 mutant phenotype (overexpression of flexin being driven by the strong muscle promoter). There remains a reduction in bouton number (31.6 boutons per synapse compared to wild type = 64.8) and an increase in the average bouton size (26.6 µm2 compared to wild type = 9.3 µm2) in futschN94; flexin/24B-GAL4 larvae. Thus, normal futsch expression is necessary for the increased loop formation and increased branching observed when flexin is overexpressed in muscle (Roos, 2000).

A model for synaptic growth requiring regulated microtubule loop architecture is presented. The subsynaptic microtubule loops, identified by Futsch and tubulin immunoreactivity, are sites of bouton division within the neuromuscular synapse. Loops occur at stereotypic locations that are, or once were, sites of active bouton division, including branchpoints and terminal boutons (Zito, 1999). In addition, examination of numerous putative dividing bouton profiles supports the conclusion that there is a progressive alteration in the microtubule-based loop architecture that is involved in the process of bouton division. Ultimately this model will have to be addressed by live, in vivo, visualization of synaptic microtubules during synaptic growth (Roos, 2000).

Genetic analysis of futsch function demonstrates that both microtubule organization and bouton division require wild-type Futsch. There are fewer and larger boutons as well as impaired microtubule organization in futsch mutants. By analogy with cell division, this is the expected phenotype if bouton division were impaired. Genetic rescue experiments indicate that futsch can drive the processes of microtubule organization. Bouton size and number as well as microtubule organization are partially rescued by overexpression of the N-terminal MAP1B homology domain of futsch. Further genetic analysis indicates that bouton division and subsequent nerve-terminal branching can be promoted by exogenous factors but require wild-type Futsch. There is a remarkable correlation between increased branching due to postsynaptic flexin overexpression and an increase in the number of presynaptic microtubule loops. Since the division of preexisting boutons occurs at loop-bearing boutons and since bouton division can generate nerve-terminal branching, the data present a correlation between increased loop formation and increased bouton division. Both elevated loop formation and increased branching require wild-type futsch, since flexin overexpression does not alter the futsch mutant phenotype. Although Futsch-dependent regulation of microtubule architecture predicts sites of apparent bouton division and appears necessary for this process, it remains to be determined whether microtubule rearrangement can drive this process (Roos, 2000).

The hypotheses concerning the role of synaptic microtubule loops during bouton division is supported by analysis of similar structures observed within the growth cones of neurons in cell culture. Nearly identical microtubule loops (observed by injection of fluorescently labeled tubulin into cultured neurons) are observed within growth cones that have paused during their migration (Tsui, 1984; Tanaka and Kirschner, 1991; Dent, 1999). Resumed growth cone motility correlates with the disruption of the loop structure into fan-like conformations (Dent, 1999). Thus, highly organized loops within growth cones and within synaptic boutons are correlated with stability or lack of change, and the disruption of these loops is associated with motility and plasticity. Therefore, a switch from a stable mode to the active process of bouton division could be controlled by the regulated destabilization of microtubule loops (Roos, 2000).

It is likely that the dynamics of microtubule loops are controlled by a MAP. The small diameter of microtubule loops observed in growth cones and at the Drosophila synapse indicates that a MAP is necessary to hold the microtubules in such a conformation. The predicted force necessary to bend polymerized tubulin into a loop with a diameter of ~3 µm is greater that the predicted buckling force of a microtubule. That vertebrate MAP1B may be involved in microtubule loop formation is supported by the demonstration that MAP1B overexpression in vitro induces the formation of wavy microtubule conformations. Vertebrate MAPs including MAP1B are regulated by phosphorylation. Thus, phosphorylation could represent a switch capable of inducing rapid changes in microtubule loop conformation within a growth cone or synaptic bouton (Roos, 2000 and references therein).

In vertebrates, phosphorylated MAP1B is enriched in growth cones. The phosphorylation-dependent regulation of MAP1B and the subsequent effects on microtubule function are complex. Phosphorylation by casein kinase II appears to increase the affinity of MAP1B for microtubules. Phosphorylation of MAP1B by glycogen synthase kinase 3-beta (GSK3beta) appears to maintain microtubules in a state of dynamic instability that is considered necessary for growth cone motility and migration (Goold, 1999). It has been suggested that phosphorylation of MAP1B by GSK3beta could act as a molecular switch to confer dynamic instability to microtubules, thereby promoting growth cone dynamics (Goold, 1999). In one model, phosphorylated Futsch could promote bouton division by promoting the dynamic instability of the microtubules within synaptic loops. Dephosphorylation of Futsch could decrease microtubule dynamics, promote loop formation, and switch synaptic boutons into a stable mode. If phosphorylation of Futsch is regulated by activity-dependent signaling, then the phosphorylation of Futsch could act as a permissive switch for activity-dependent plasticity at specific subsynaptic locations (Roos, 2000).

Recent results from studies of synaptic plasticity in the vertebrate brain indicate that the sprouting of new dendritic spines may be correlated with the consolidation of long-term synaptic plasticity. Interestingly, MAP1B is enriched in areas of the vertebrate brain that show substantial activity-dependent synaptic plasticity. Cytoskeletal rings have been observed in similar areas of the brain and in vertebrate hippocampal cell culture. It is speculated that a MAP regulation of microtubule loops may participate in the process of synaptic morphological change within the vertebrate central nervous system during activity-dependent plasticity (Roos, 2000 and references therein).

Thus, Drosophila futsch is essential for both axon elongation and synaptic growth. Futsch is implicated in the regulation of microtubule dynamics through the formation of microtubule loop structures at the synapse. Nearly identical microtubule structures have been previously demonstrated to regulate growth cone morphology and motility. The control of microtubule organization by MAPs may represent a common mechanism for regulated growth cone motility as well as synaptic growth and plasticity (Roos, 2000).

Coordinated regulation of axonal microtubule organization and transport by Drosophila Neurexin and BMP pathway

Neurexins are well known trans-synaptic cell adhesion molecules that are required for proper synaptic development and function across species. Beyond synapse organization and function, little is known about other roles Neurexins might have in the nervous system. This study reports novel phenotypic consequences of mutations in Drosophila neurexin (dnrx), which alters axonal microtubule organization and transport. dnrx mutants display phenotypic similarities with the BMP receptor wishful thinking (wit) and one of the downstream effectors, futsch, which is a known regulator of microtubule organization and stability. dnrx has genetic interactions with wit and futsch. Loss of dnrx also results in reduced levels of other downstream effectors of BMP signaling, phosphorylated-Mad and Trio. Interestingly, postsynaptic overexpression of the BMP ligand, Glass bottom boat, in dnrx mutants partially rescues the axonal transport defects but not the synapse undergrowth at the neuromuscular junctions. These data suggest that dnrx and BMP signaling are involved in many diverse functions and that regulation of axonal MT organization and transport might be distinct from regulation of synaptic growth in dnrx mutants. Together, this work uncovers a novel function of Drosophila Neurexin and may provide insights into functions of Neurexins in vertebrates (Banerjee, 2018).

Neurexins are some of the most extensively studied synaptic cell adhesion molecules for their role in synapse formation and function. In Drosophila, dnrx mutants have reduced synaptic growth. This reduced synaptic growth is similar to what was observe in the BMP signaling mutants: wit, gbb, mad and trio. Recent studies have shown that dnrx regulates components of the BMP signaling pathway and directs synaptic growth and cytoarchitecture in conjunction with dnlg and wit. The data presented in this study shows that dnrx engages in a complex molecular machinery with the BMP receptor Wit, the ligand Gbb and other downstream effectors to allow smooth axonal transport and proper MT organization (Banerjee, 2018).

Both dnrx and wit mutants display phenotypic similarities with axonal MT organization and transport of various cargo along the segmental nerves. It is interesting to note that axonal MT and transport phenotypes resulting from gain of dnrx function are similar to loss of function mutants. This suggests that a fine balance in levels of dnrx is important, and either too little or too much of it is detrimental for MT organization and the transport machinery. Defects in axonal transport resulting from cell type specific reduction in motor neurons using dnrx-RNAi confirmed that the axon transport defect is autonomous to motor neurons. The defects resulting from the dnrx knockdown, however, were milder than dnrx loss-of-function possibly due to penetrance or expressivity factors. The Gal4-only controls (elav-Gal4 and OK6-Gal4) did not display any axonal transport or MT organization defects further confirming the specificity of the dnrx phenotypes. It is also worth noting that gain of wit did not disrupt MT organization or axonal transport (Banerjee, 2018).

The genetic interaction between dnrx and wit suggest that these proteins regulate axonal MT organization and transport. Recent studies have also showed that dnrx and wit coordinate synaptic organization and growth, which suggests that these proteins are needed for multiple processes during nervous system development. The ultrastructural phenotypes in dnrx and wit mutants reveal a structural disorganization in the arrangement of MT filaments along the axons. These datasets showed that there is indeed disorganization in MT morphology in the mutants and these findings are not artifacts from the altered localization of Futsch (Banerjee, 2018).

Both dnrx and wit are presynaptic proteins and their loss shows a multitude of defects at the axonal, cytoskeletal and synaptic levels. It was therefore important to examine the range of behavioral deficits in these mutant larvae. dnrx and wit mutants displayed defects both in peristalsis and locomotor behaviors. These behaviors are attributable, most likely at least in part, to the defects in axon transport, MT organization and possible dysfunction of the NMJ synapses. Behavior in animals could be a composite of defects at many levels of integration at the CNS, PNS, NMJ and circuits responsible for producing rhythmic patterns of movement and general locomotion. Defects in both peristalsis and wandering behavior in cysteine string protein (csp) mutant larvae, for example, are thought to be due to motor defects at the NMJ and at the levels of CNS output generation. The locomotor behaviors observed in dnrx and wit mutants also indicate that these proteins might play a boarder role in organization of motor circuits (Banerjee, 2018).

Recent studies have implicated BMP signaling in the regulation of synaptic strength, axonal MT organization and transport. While the current studies point to the reliance of MT organization and axonal transport on BMP signaling, there are other studies that show disruption of axonal transport perturbing BMP signaling contributes to synaptic abnormalities in neurodegenerative diseases. It is important to note that mutations that fall under the categories of both positive regulators of BMP, like dnrx8 and negative regulators of BMP such as, Spicthyin and Spartin both lead to MT disorganization. It is, therefore, interesting that while Spichthyin and Spartin function as negative regulators of the BMP signaling by regulating BMP receptor traffic and modulate regulation of MT cytoskeleton, loss of these proteins does not show impaired axonal transport. dnrx loss results in disorganized MTs, but unlike Spichthyin and Spartin, loss of dnrx shows transport defects. These findings suggest that MT cytoskeletal organization could be compromised by both positive and negative regulation of BMP signaling (Banerjee, 2018).

The regulation of microtubule stability is essential for axonal transport. The microtubule-associated protein, Futsch, recently implicated in the BMP pathway, has a role in the stabilization of MTs. futsch mutants display axonal transport defects consistent with altered MT organization similar to dnrx and wit mutants. A previous report, however, showed that futsch mutants do not show defects in axonal transport of Brp. This study examined futsch mutants for accumulation of Brp puncta along axons. Mild, yet significant, increase in accumulation of Brp was found along the axons in futsch mutants. These studies are not completely in agreement with the previous report. The discrepancies between the studies could be possibly due to: (1) differences in quantification parameters (Brp area/Hrp area as opposed to number of Brp puncta along nerve length); (2) difference in segmental nerves analyzed (the current analysis was restricted to segmental nerves 5-8); and (3) the region of nerve fiber analyzed (proximal or distal). It is also worth noting that light microscopy studies showed Syt and DCSP2 aggregation along mutant axons did not always co-localize with Brp puncta. These observations are consistent with recently published report that showed Syt co-traffic less efficiently with active zone scaffold proteins Brp and Basoon. Collectively, these findings suggest the possibility of the presence of distinct transport mechanisms for different synaptic cargoes. Future studies will be aimed at addressing the dynamics of axonal membrane traffic in mutants of dnrx and BMP pathway, and whether some synaptic proteins are co-transported and others are not, or whether transport mechanisms are able to discriminate synaptic cargo as it is transported along the axons (Banerjee, 2018).

futsch and dnrx interact genetically as double heterozygotes display axonal transport defects consistent with dose-dependent genetic interactions. This data further corroborates the idea that dnrx is not only necessary for trans-synaptic adhesion and apposition of the pre- and post-synaptic compartments, but also for maintaining the stability of the BMP receptor wit for optimal BMP signaling. Future studies will be aimed at addressing the link between dnrx and Futsch. dnrx is a transmembrane protein with a PDZ domain binding motif and Futsch is a MT-binding protein. Both of these proteins are highly expressed along the axons and could potentially exist as a complex. It is also possible that dnrx might interact with MT motor proteins such as Kinesin or Dynein. The Kinesin superfamily protein, KIF 17, was previously shown to interact with PDZ domain of mLin-10 (Mint1/X11) to transport NMDA receptor 2B. mLin-10 is part of a larger protein complex that includes mLin-2 (Cask) and mLin-7 (Veli). It is likely that the intracellular PDZ domain of dnrx is capable of recruiting protein complexes such as these to form a link between axonal membranes to the axonal MT cytoskeleton (Banerjee, 2018).

While it was not a surprise that levels of the downstream effectors of the BMP cascade, pMad and Trio, would be reduced in dnrx mutant BL/VNC given that previous studies showed reduced levels of Wit, Tkv and pMad at the dnrx mutant NMJ synapses. It is interesting that postsynaptic overexpression of the BMP ligand, Gbb, partially rescues the axonal transport phenotypes in dnrx mutant background but not the NMJ bouton undergrowth phenotype. This indicates that distinct regulatory mechanisms might be involved in proper axonal MT organization and transport and growth during NMJ development, which are still not well understood. While this finding may suggest that Wit/Gbb signaling does not require dnrx to regulate axon transport, it is important to note that the rescue of the axonal transport phenotype was partial and did not reach wild type levels, suggesting that dnrx is required to allow optimal levels of axonal transport to occur. This finding also indicates: (1) that possibly it is the retrograde transport that gets somewhat corrected while there might still be defects in anterograde transport of cargoes in dnrx mutants; (2) that BMP signaling solely is not responsible for trafficking of cargoes and overall MT health in dnrx mutants; and (3) that there might be additional proteins involved in forming a cascade to link dnrx from axonal membrane to MT cytoskeleton. Given the findings reported in this study on Drosophila Nrx, it would be of immense interest to investigate whether mammalian Neurexins will have similar functions separate from their role in synapse formation and their functional modulation, and also axonal microtubular organization. Future studies will further address other aspects of Neurexin functions in the nervous system (Banerjee, 2018).


PROTEIN STRUCTURE

Amino Acids - 5327

Structural Domains

The futsch ORF shows high homology to MAP1B and MAP1A at both the N terminus as well as at the C terminus. The large middle domain of the deduced Futsch protein shows a highly repetitive structure. In total, 66 direct repeats of a 37 amino acid sequence motif have been found. Two blocks of 11 and 20 highly conserved units are flanked by less conserved repeat units (homology >40% identity). Database searches using the Futsch repeat unit reveal some homology to the neurofilament protein family (probability e-14). Interestingly, the sequence KSPXXXP, which is frequently found in the Futsch repeats and in neurofilament proteins, has been described as a target of Erk2 protein kinases (Veeranna, 1998). Whether Futsch is phosphorylated at these positions remains to be determined (Hummel, 2000).

The MAP1B proteins and especially the N- and C-terminal protein domains have been highly conserved during vertebrate evolution. MAP1A and MAP1B sequences are more distantly related, and again, the strongest sequence conservation is observed at the ends of the proteins (N-terminal domain = 70%, C-terminal domain = 80% identity). The same regions show high homology to the Futsch protein, suggesting that Futsch may represent a new member of this MAP family. MAP1B differs from MAP1A by the addition of about 220 amino acids at the N terminus. Since the homology to Futsch starts within this region, Futsch has been designated as a MAP1B homolog. The microtubule binding domains identified in the N and C parts of MAP1B are not conserved in the Futsch protein (Hummel, 2000).


futsch : Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References

date revised: 3 August 2000

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.