The Microtubule Regulatory Protein Stathmin Is Required to Maintain the Integrity of Axonal Microtubules in Drosophila

Axonal transport, a form of long-distance, bi-directional intracellular transport that occurs between the cell body and synaptic terminal, is critical in maintaining the function and viability of neurons. We have identified a requirement for the stathmin (stai) gene in the maintenance of axonal microtubules and regulation of axonal transport in Drosophila . The stai gene encodes a cytosolic phosphoprotein that regulates microtubule dynamics by partitioning tubulin dimers between pools of soluble tubulin and polymerized microtubules, and by directly binding to microtubules and promoting depolymerization. Analysis of stai function in Drosophila , which has a single stai gene, circumvents potential complications with studies performed in vertebrate systems in which mutant phenotypes may be compensated by genetic redundancy of other members of the stai gene family. This has allowed us to identify an essential function for stai in the maintenance of the integrity of axonal microtubules. In addition to the severe disruption in the abundance and architecture of microtubules in the axons of stai mutant Drosophila , we also observe additional neurological phenotypes associated with loss of stai function including a posterior paralysis and tail-flip phenotype in third instar larvae, aberrant accumulation of transported membranous organelles in stai deficient axons, a progressive bang-sensitive response to mechanical stimulation reminiscent of the class of Drosophila mutants used to model human epileptic seizures, and a reduced adult lifespan. Reductions in the levels of Kinesin-1, the primary anterograde motor in axonal transport, enhance these phenotypes. Collectively, our results indicate that stai has an important role in neuronal function, likely through the maintenance of microtubule integrity in the axons of nerves of the peripheral nervous system necessary to support and sustain long-distance axonal transport.


Introduction
The organized transport of organelles, vesicles and macromolecular protein complexes is necessary to support cellular growth, function, and viability. The requirement for an efficient transport system is pronounced in neurons because their architecture typically comprises a long, narrow axon that extends to the synaptic terminal, and is orders of magnitude longer than the diameter of the cell body. The axon serves to transmit action potentials between the cell body and synaptic terminal, but also acts as a conduit for the long-distance transport of materials between these two discrete cellular compartments.
In axons, the motor proteins cytoplasmic dynein and kinesin bidirectionally transport cargo along tracks of microtubules (MTs), the main cytoskeletal component of axons. MTs are hollow polymers formed by the lateral association of linear, polarized protofilaments of heterodimers of α and β-tubulin joined head-to-tail. MTs are dynamic structures, the assemblyfavored 'plus' ends stochastically cycling between phases of growth and shrinkage, in a process known as 'dynamic instability' [1]. MTs are subject to intense regulation by a vast array of factors including MT-stabilizing proteins, MTpolymerizing and depolymerizing proteins, and MT-severing proteins that act coordinately to modulate MT dynamics by directly interacting with and modifying MTs (reviewed in [2]). Impairment of the axonal transport system is believed to cause or dramatically contribute to the development of human neurodegenerative diseases (reviewed in [3,4]). Indeed, mutations in genes that encode MT motor proteins that power axonal transport have been directly linked to human neurodegenerative diseases [5][6][7][8]. However, maintenance of the MT network that supports axonal transport is also critical. Mutations in genes encoding microtubule-associated proteins (MAPs) or proteins that regulate MT dynamics also impair axonal transport and cause human neurodegenerative disease. The neuronal MAP tau is the major constituent of insoluble intracellular inclusions called neurofibrillary tangles, a pathological hallmark of neurodegenerative tauopathies including Alzheimer's disease. Pathogenic forms of tau have been demonstrated to impair kinesin-dependent axonal transport [9,10], and reductions in kinesin transport exacerbate neurodegeneration in animal models of tauopathies [11]. In addition, mutations in the gene spastin, which encodes a MT severing protein, are the cause of the most common form of hereditary spastic paraplegia in humans [12]. Deletion or mutation of the mouse spastin gene impairs axonal transport, resulting in progressive axonal degeneration [13,14].
We have carried out genetic screens to identify novel genes required for axonal transport. Among the candidate genes isolated, we have identified mutant alleles in the gene that encodes the MT regulatory protein stathmin. Stathmin (also known as Oncoprotein18 (Op 18), Phosphoprotein p19, metablastin, Prosolin, and Leukemia Associated Protein 18) is a cytosolic phosphoprotein that regulates MT dynamics by partitioning tubulin dimers from pools of soluble tubulin thus preventing their assembly into polymerized MTs, and by directly binding to MTs and promoting their disassembly [15,16]. These functions are attributable to different regions of the stathmin protein. The N-terminus of stathmin promotes MT catastrophes whereas the C-terminus exhibits tubulinsequestering activity [17]. Four genes encode the vertebrate stathmin family of proteins; STMN1 encodes the ubiquitous, cytosolic stathmin protein, while STMN2, STMN3 and STMN4 encode the neuron-specific, membrane-associated stathminlike proteins SCG10, SCLIP and RB3/RB3'/RB3" respectively [18]. All share a homologous tubulin-binding stathmin-like domain, while SCG10, SCLIP and RB3 contain an N-terminal membrane-targeting domain [19]. The vertebrate stathmin proteins can bind two tubulin heterodimers and are subject to regulatory phosphorylation on four conserved serine residues that reduce their ability to bind tubulin and regulate MT polymerization [20,21].
In Drosophila, a single stathmin (stai) gene encodes four protein isoforms, representative of the entire vertebrate stathmin family [22]. The stai gene is differentially expressed during Drosophila embryogenesis, with transcripts encoding staiA isoforms maternally deposited to the embryo, whereas the staiB encoding isoforms are expressed at high levels in early neuroblasts, and later in the developing embryonic central and peripheral nervous system [22]. Neuronal enriched staiB isoforms have an N-terminal domain that has conserved cysteine amino acids that are potential palmitoylation sites thought to play a role in subcellular targeting of the protein [22].
Analysis of loss of stai function in Drosophila has allowed us to circumvent potential complications encountered in studies in vertebrate systems in which mutant phenotypes may be compensated by genetic redundancy of other member of the stai gene family.
In this study, we show through loss of stai activity, an essential function for stai in the maintenance of axonal MTs necessary to support axonal transport. The architecture of the axonal MT network appears severely disrupted in our stai mutants, implicating a critical role for stai in its regulation. Loss of stai activity results in a posterior paralysis and 'tail-flip' phenotype in third instar larvae, hallmarks of impaired axonal transport.
Immunohistochemistry reveals aberrant accumulation of synaptic vesicles in the axons of segmental nerves of stai mutant larvae. Despite the disruption in axonal transport, a significant number of mutant animals survive to the adult stage but have a reduced lifespan. In addition, stai mutant adults exhibit a progressive, late-onset bang-sensitive seizure in response to mechanical stimulation. This epilepsy-like behavioral deficit is characteristic of the class of mechanical shock-sensitive Drosophila mutants that have altered neuronal excitability and are used to model human epilepsy [23]. We demonstrate by mutant analysis and transgene rescue that the observed phenotypes are the result of loss of stai function. We also show that expression of the human stathmin gene, STMN1, can rescue these defects indicating that the Drosophila and human proteins are functional orthologs. Genetic reduction in the levels of Kinesin-1, the primary anterograde motor of axonal transport, enhances stai mutant phenotypes. Collectively, our results indicate that stai has an important function in the maintenance of the integrity of MTs in the axons of nerve cells of the peripheral nervous system necessary to support axonal transport.
The piggyBac insertion in PBac{5HPw + } stai B200 was precisely excised by crossing PBac{5HPw + } stai B200 /CyO against w 1118 ; CyO P{FRT(w + ) Tub-PBac\T} 2/wg Sp-1 and generating individual balanced stocks that were determined by eye color to have the piggyBac transposon excised. The nature of the excision for each was confirmed by PCR across the original insertion site and subsequent sequence analysis.
The stai B200 Khc 20 chromosome was generated by meiotic recombination, and balanced over CyO Act GFP. Recombinant stai B200 Khc 20 chromosomes were identified by outcrossing each potential recombinant line against w 1118 and following the inheritance of the mini w + , associated with stai B200 , in straightwinged progeny. The presence of Khc 20 was inferred by the absence of straight-winged flies in each recombinant stock, and confirmed by lethality in heterozygous combination with Df(2R) Jp8.

PCR Amplification and Sequencing of the stai Genomic Region
Genomic DNA was isolated from homozygous third instar stai rdtp larvae according to standard protocol [25]. The coding region of the stai gene was amplified in four separate PCR reactions using HotStar HiFidelity DNA Polymerase (Qiagen, Valencia, CA) according to manufacturer's recommended protocol.

qRT-PCR
mRNA purified from third instar larvae using the Oligotext Direct mRNA Mini Kit (Qiagen) was used to synthesize cDNA with the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA). RT-PCR was performed with the SYBR Green PCR Master Mix (Applied Biosystems) on an ABI 7300 Real Time PCR System (Applied Biosystems). In order to distinguish between stai transcripts with different transcriptional start sites, stai message was quantified with 'forward' oligonucleotides complementary to coding sequence within exon 2 (5'-GGTGAACAACACTGTTGATACTG -3') shared by staiB isoforms, and sequence downstream of the copia insertion site in stai rdtp within exon 1' (5'-GATTAGTACGCGACTCGGTGA -3') shared by staiA isoforms, in separate amplifications with a 'reverse' oligonucleotide complementary to sequence in exon 3 (5'-CTCAATCTCCTCGACGCTAAC -3'), shared by all stai isoforms. Results were normalized against the signal from the GAPDH gene product (5'-AATTAAGGCCAAGGTTCAGGA-3', 5'-ACCAAGAGATCAGCTTCACGA-3').
All RT-PCR experiments were repeated five times for each genotype, and analyzed and quantified using the comparative C T method (ΔΔC T method). The results are presented as mean ± standard deviation (SD). Significance was determined by ANOVA of ΔC T values [26,27].

Quantification of the Posterior Paralysis 'Tail-Flip' Phenotype
The posterior paralysis phenotype of late third instar larvae was analyzed as described previously [28]. Briefly, third instar larvae of each genotype were collected, rinsed in ddH 2 O and placed in 60 x 15 mm culture dishes, and the crawling behavior of each larva observed. A minimum of one hundred larvae was analyzed for each genotype. The severity of the posterior paralysis phenotype was scored and quantified using the following criteria; larvae were determined to exhibit a robust tail-flip if the tail was raised greater than 40 o above horizontal when crawling; a mild tail-flip if the tail was raised less than 40 o above horizontal when crawling, and no-tail flip if the larva exhibited a normal crawling behavior. The crawling behavior of third instar larvae was recorded with a Canon PowerShot G10 coupled to an Olympus SZ61 stereomicroscope and the resultant movies processed in iMovie and QuickTime Pro.

Confocal Imaging and Analysis
All images were acquired on a Zeiss LSM-710 laser scanning confocal microscope. Imaging parameters, including signal gain, for a given experimental treatment were set such that the brightest treatment group produced images that were non-saturating. All imaging parameters were kept constant across a given experimental treatment.
Images of segmental nerve axons represent threedimensional maximum intensity projections of serial stacks, 10.4-14.2 µm thickness, acquired from the four to six medial segmental nerves spanning abdominal segments A4-A6. Disruptions in axonal transport were quantified from maximal intensity projection images of segmental nerve axons as they passed through abdominal segments A2 and A4 using the threshold and particle analysis functions in ImageJ (ver 1.47k) with a size detection minimum of 0.5 µm 2 . Results are reported as number of CSP accumulations/50 µm segmental nerve axon. Futsch and acetylated α-tubulin staining intensity was quantified using the histogram function of the Zen 2010 image analysis software package that reports a distribution of pixel intensities, as well as average pixel intensity, in a selected region of interest of an image. A lower threshold for the histogram analysis was applied to all images analyzed. Statistical analysis across the genotypes analyzed was performed using ANOVA of average pixel intensity of all segments analyzed within each experimental group, followed by Fisher's LSD post hoc analysis.
Images of larval muscle represent three-dimensional maximum intensity projections of serial stacks, ~14.8-16.2 µm thickness, through muscle 6 at abdominal segment A4. Perinuclear MT density was quantified with ImageJ 1.44o by measuring the ratio of tubulin positive: negative staining signal in an area that extends 5 µm around the nucleus. Average perinuclear MT density was determined for each genotype by measuring MT density surrounding at least twenty nuclei, recorded from six larvae. The average perinuclear MT density for each genotype was normalized against wildtype.

Analysis of Lifespan and Bang-Sensitive Phenotypes
The life span of male Drosophila was measured at 25 o C. Females were excluded from the life span analysis due to observed sex differences regarding the effect of dietary restriction on lifespan [31]. Adult flies were collected 24-hours post eclosion and maintained in standard culture vials (25 mm x 95 mm) with a maximum of 10 flies per vial. Flies were counted daily and transferred to fresh vials every other day. A minimum of one hundred flies per genotype was analyzed. The mean lifespan for each genotype was calculated by averaging individual lifespans for flies within a cohort. Significance was determined by ANOVA followed by Fisher's and Scheffe's post hoc analysis.
The bang-sensitive phenotype was quantified by the methodology of Ganetzky and Wu [32]. Briefly, male Drosophila of each genotype were aged at 25 o C. Forty-eight hours prior to testing, the aged flies were CO 2 anaesthetized and a maximum of five flies were placed into fresh vials. The flies were mechanically stimulated by vortexing at full-speed in culture vials for 10 seconds and the number of flies exhibiting paralysis counted. The penetrance and severity of the bangsensitive phenotype was quantified by testing a minimum of fifty aged adult male flies of each genotype for bang sensitivity following mechanical stimulation at 1, 7, 14, 21, 42 and 56 days post eclosion. In addition, the average time period required for recovery from the paralysis was also quantified for each fly by measuring the time from the end of the vortex period until the paralyzed flies regained the ability to stand and take their first steps.

Generation of Antisera against Drosophila Kinesin Light Chain
Rabbit polyclonal antisera was produced against recombinant Drosophila kinesin light chain protein following PCR amplification of the Klc gene from cDNA clone NM_079325. 3 with primers Klc(Xho1) 5'-ATCCGAGCTCGAGACGCAAATGTCGCAG-3' and Klc(Kpn1) 5'-CGAATTCCATGGTACCTTATGGTTTCGC-3'. The resultant PCR product was cut with Xho1 and Kpn1 and subcloned into the Xho1/Kpn1 site of the pRSET-A plasmid to generate pRSET-A-Klc. pRSET-A-Klc was sequenced to confirm a complete Klc open reading frame, transformed into BL21(DE3) pLysS, and expression of the His-tagged Klc protein was induced in a 100 ml culture at OD 600 =0.5 with IPTG to a final concentration of 1 mM. The cells were harvested four hours post induction and the recombinant Klc protein was purified using nickel-nitrilotriacetic acid (Ni-NTA) agarose resin (Qiagen), electrophoretically isolated and used to raise antisera in rabbits (Covance Research Products, Denver, PA). The specificity of the Klc antibody was confirmed by western blot against total protein extract isolated from third instar larvae. In wildtype Drosophila, the antibody recognized a single band of approximately 60 kDa, and a reduction in the intensity of the signal from this 60 kDa band was observed in protein from a heterozygous strain harboring the chromosomal deletion Df(3L) 8ex94 that removes the Klc gene [33].
The PCR product was cut with Not1 and Xho1 and directionally subcloned into the Not1/Xho1 site of the pCaSpeR-tub plasmid (gift of Pernille Rorth, IMCB, Singapore) to generate pCaSpeRtub-STMN1. The pCaSpeR-STMN1 rescue construct was sequenced to verify a complete STMN1 open reading frame. Standard germline transformation techniques [34] were employed to generate multiple independent transgenic lines that were mapped to identify third chromosome insertion lines used to generate homozygous stocks.

Isolation and Molecular Characterization of stathmin Mutant Alleles
The redtape (rdtp) mutation was isolated in a genetic screen for recessive third chromosome P-element induced mutations that disrupt axonal transport in Drosophila melanogaster (unpublished data). The basis for the mutagenic screen was a posterior paralysis of third instar larvae, observed as a 'tail-flip', previously described for mutations in genes that encode proteins required for axonal transport [28,29,[35][36][37][38]. Initial analysis of the rdtp mutation led to the discovery that the posterior paralysis observed in homozygous mutant larvae was due to a second chromosome mutation and not the P-element insertion. The rdtp mutation was subsequently mapped to cytological region 26B5-26C1 on the left arm of the second chromosome by its failure to complement Df(2L) BSC5, Df(2L) ED384, Df(2L) ED385, Df(2L) Exel6015 and Df(2L) BSC239. Further analysis identified PBac{5HPw + } stai B200 , obtained from the Bloomington Drosophila Stock Center, as a noncomplementing gene mutation. Heteroallelic combination of rdtp and stai B200 recapitulated the larval 'tail flip' phenotype originally observed in rdtp homozygotes, indicating that rdtp is a mutant allele of the stai gene. The rdtp mutation was subsequently renamed stai rdtp .
PCR amplification of the stai gene from homozygous second instar stai rdtp larvae generated an ~5.6 kb PCR product for exon 1', larger than the predicted 510 bp PCR product observed in wildtype ( Figure 1B). The 5.6 kb PCR product was sequenced and found to encode a copia retrotransposon, a member of a broad class of natural genetic elements characterized by long direct terminal repeats [39]. Spontaneous copia-induced mutations have been previously reported for many genes in Drosophila [40][41][42][43][44]. The copia element in stai rdtp is inserted in the open reading frame of the stai gene ( Figure 1A). The noncomplementing stai B200 mutation is the result of a piggyBac transposon insertion in the 2.8 kb intron separating exons 1' and 2' of the stai gene ( Figure 1A), 1337bp downstream of exon1' and 1469bp upstream of exon 2' (genome sequence position 2L:6,111,561 -FlyBase version FB2013_01, released January 23 rd , 2013).

Expression of stai is Reduced in stai rdtp and stai B200 Larvae
We used qRT-PCR to determine if the copia insertion in stai rdtp and the piggyBac insertion in stai B200 altered the expression level of the stai gene in third instar larvae. Amplification with primer pairs specific for detection of staiA and staiB mRNA isoforms revealed significantly reduced levels of both transcripts in homozygous stai rdtp third instar larvae (P<0.01) ( Figure 1C). Transcripts encoding staiA isoforms were more strongly affected, reduced to as little as 12.0 ± 1.0% (mean ± SD) of wild type levels, compared to staiB encoding transcripts, which were reduced to 41.0 ± 4.0% of wild type levels ( Figure 1C). As expected, heterozygous combination of stai rdtp with Df(2L) Exel6015, a 170 kb chromosomal deletion that removes 14 genes including stai, reduces the abundance of staiA and staiB derived transcripts to 4.0 ± 1.0% and 15.0 ± 5.0% respectively, approximately 50% of the levels observed in stai rdtp homozygous larvae ( Figure 1C). The level of staiB message is also reduced in homozygous stai B200 mutant third instar larvae, to 1.0 ± 1.0% of wildtype (P<0.0001) ( Figure 1C). The levels of staiA encoding transcript in homozygous stai B200 larvae and message encoding both staiA and staiB isoforms in stai B200 /Df(2L) Exel6015 larvae are reduced to undetectable levels ( Figure 1C).

Lethal Phase Analysis of stai Mutants
Despite the dramatic reduction in stai message, the stai B200 allele causes semi-lethality, with approximately 40% of the expected homozygous progeny from a heterozygous cross surviving to the adult stage (n=1900), suggesting that stai function is not essential for adult viability. These results are supported by the previous observation that stai null Drosophila are adult viable [45]. Unexpectedly, homozygous stai rdtp larvae were inviable past the third instar stage with no homozygous adults recovered among the progeny of a heterozygous cross (n=1000). This was confounding given that greater stai message was detected in homozygous stai rdtp larvae than stai B200 larvae ( Figure 1C) and that stai null Drosophila are reported to be adult viable [45]. The mutagenic copia insertion in stai rdtp , however, is inserted in exon 1' of the stai open reading frame used to produce neuronal staiB splice variants. We considered the possibility that the copia sequence was incorporated in the staiB message produced by stai rdtp mutant larvae, resulting in a dominant negative protein. However, a significant number of viable adults were observed when the stai rdtp chromosome was placed in heterozygous combination with stai B200 or Df(2L) Exel6015. Therefore, it is likely that an additional mutation responsible for the homozygous lethality exists on the stai rdtp chromosome. Given the likely presence of a second, lethal mutation on the stai rdtp chromosome, and the observation by others that copia-induced gene mutations are subject to the action of suppressor or enhancer alleles mapping . Transcripts encoding staiA isoforms include exons 1 and 2 with alternative splicing that either includes (staiA1) or excludes (staiA2) exon 6. Transcripts encoding staiB isoforms include exons 1'and 2' with alternative splicing that either includes (staiB1) or excludes (staiB2) exon 6. The copia retrotransposon stai rdtp is inserted in the open reading frame of the stai gene in exon 1', ten base pairs downstream of the translational start site used to produce nervous system enriched staiB encoding transcripts. The piggyBac element PBac{5HPw + }stai B200 is inserted in the 2.8 kb intron separating exons 1' and 2', 1.3kb downstream of the splice junction of exon 1'. (B) The copia insertion in the stai gene was identified by PCR amplification across the open reading frame of stai exon 1' that resulted in an unexpectedly large 5.6 kb product from genomic DNA isolated from stai rdtp homozygous larvae. The black arrowheads in Figure 1A represent the relative position of PCR primers used. (C) qRT-PCR of staiA and staiB mRNA derived from third instar larvae is shown. The expression of staiA and staiB is significantly reduced in all stai mutant genotypes analyzed compared to wild type expression levels (P<0.01). The red and blue arrowheads in Figure 1A represent the relative position of primers used for qRT-PCR of staiA and staiB message. Results are normalized against the expression of the GAPDH gene product and are presented as mean ± SD. to other loci, resulting in complex phenotypes [46,47], the focus of this study is the stai B200 mutant allele and the use of the stai rdtp chromosome is limited to heterozygous analysis or in heteroallelic combination with stai B200 and Df(2L) Exel6015.

Loss of stai Function Results in a Larval Posterior Paralytic Phenotype
Homozygous stai rdtp and stai B200 third instar larvae exhibited phenotypes characteristic of mutants with impaired axonal transport. Although stai mutant larvae were of comparable size to wild type larvae, their posterior regions were often weakly tapered and reduced in diameter as reported in other axonal transport mutants (data not shown) [48]. When homozygous stai rdtp and stai B200 larvae crawled along the surface of a substrate, they exhibited a crawling behavior in which the posterior body segments sharply flipped upward after each peristaltic wave of muscle contraction during the crawling cycle (compare Figure 2A with 2B and 2C, Movies S1-S3). Heterozygous combination of stai B200 with stai rdtp recapitulated the tail-flip phenotype (Movie S4). The larval posterior-paralysis or tail-flip phenotype is a hallmark of defective axonal transport in Drosophila and has been previously described for mutations in genes that encode components of MT motor protein complexes, such as kinesin and cytoplasmic dynein, as well as motor accessory proteins [28,29,33,[35][36][37]. The phenotype is thought to result from a dorsal-ventral gradient in posterior muscle paralysis that causes an imbalance in contraction of the larval body wall muscles [29]. Interestingly, stai null third instar larvae also exhibited a posterior paralysis phenotype (gift of Pernille Rorth, IMCB, Singapore) (Movie S5).

stai Deficient Axons Exhibit Defects in Axonal Transport
Mutations in genes required for axonal transport not only cause a similar posterior paralysis and tail-flip phenotype in third instar larvae, but they also result in the accumulation of anterograde and retrograde membranous axonal cargo in the segmental nerves of these animals [29]. To determine if the posterior paralysis phenotype observed in stai mutant larvae resulted from impaired axonal transport, the distribution of cysteine string protein (CSP) in the axonal compartment was assayed by immunohistochemistry. CSP is a synaptic vesicle protein transported in a microtubule-dependent manner within the axons of larval segmental nerves [49,50]. In wildtype segmental nerve axons, CSP exhibits a uniform, punctate staining pattern throughout the axoplasm ( Figure 2D). Heterozygous stai B200 /+ larvae exhibited an average of 1.64 ± 2.95 (mean ± SD) CSP accumulations/50 µm segmental nerve axon examined (n=71 axon segments examined, p>0.05) (arrowheads Figure 2E, 2H). stai B200 /+ larvae exhibit a normal crawling behavior and the absence of a posterior paralysis phenotype, indicating axonal transport is not significantly compromised in peripheral nerve axons. In contrast, homozygous stai B200 third instar larvae exhibited 13.55 ± 5.56 aggregations of CSP/50 µm segmental nerve axon (n=52, p<0.0001, Figure 2F, 2H), consistent with an impairment in axonal transport. The CSP aggregations were not significantly enhanced in stai B200 /Df(2L) Exel6015 larvae (15.10 ± 6.07 CSP accumulations/50 µm segmental nerve axon, n=57, p=0.06, Figure 2G, 2H). These results confirm that stai activity is required for efficient axonal transport in peripheral nerves.
To determine if there was an anterior to posterior gradient in the phenotypic severity of CSP aggregations in segmental nerve axons of stai mutant larvae, we quantified and compared the extent of axonal transport defects in regions of segmental nerve axons that passed through the more anterior abdominal segment A2 and the posterior abdominal segment A4 ( Figure  2H). The only regional difference in the abundance of axonal clogs in segmental nerve axons was observed in axons from stai B200 /Df(2L) Exel6015 larvae (p<0.0001).

Adult Viable stai Mutants Have Reduced Lifespans
Our observations indicate that Drosophila can tolerate significant reductions in stai levels. Despite a large percentage of third instar larvae exhibiting a posterior paralysis due to impairment of axonal transport, a significant percentage of stai mutant animals manage to survive to the adult stage but exhibit significant reductions in average adult lifespan at 25 o C ( Figure  3). The stai B200 /+ heterozygote has an average adult lifespan of 81.0 ± 1.2 days, not significantly different from the 77.5 ± 1.6 days observed in the wildtype (P≥0.05) ( Figure 3A, B) or the 80.2 ± 1.6 days in Df(2L) Exel6015/+ adults (P≥0.05) (data not shown). In contrast, a significant reduction in average adult lifespan was observed for stai B200 adults, averaging 46.6 ± 2.3 days (P<0.0001) ( Figure 3A, B). The average adult lifespan was reduced even further, to 34.8 ± 2.3 days (P<0.0001) in stai B200 /Df(2L) Exel6015 adult males ( Figure 3A, 3B).

Adult Viable stai Mutants Exhibit a Progressive Bang-Sensitive Seizure Phenotype
While performing the lifespan analysis on adult viable stai mutants, we observed they exhibit a temporary paralysis in response to the mechanical activity of being transferred to new food vials. This phenotype is similar to that observed in the "bang-sensitive" class of genes, mutant alleles of which cause neurological defects characterized by a paralysis or seizure in response to mechanical stimulation. Mutant alleles of a number of genes have been identified that exhibit a bang-sensitive phenotype including bang-sensitive (bas), bang-senseless (bss), technical knockout (tko), slam dance (sda) and easily shocked (eas). These genes are involved in diverse cellular processes including ion transport, Ca 2+ release and mitochondrial metabolism [51,52]. Typically, the seizures become easier to trigger, and the subsequent paralysis increases in duration, with increasing age [53].
We quantified the response of stai mutant animals to mechanical stimulation by performing a bang-sensitive assay on adult male flies of different ages. We observed a normal response to, and recovery from, mechanical stimulation for all genotypes tested at 1, 7 and 14 days post eclosion (Tables S2  and S3). However, 20.9% of stai B200 males tested at 21 days post eclosion (n=67) exhibited bang-sensitive paralysis with a mean recovery period of 48.6 ± 26.7 seconds (mean ± SD). During the bang-sensitive paralysis, flies laid motionless on the media at the bottom of the vial for periods ranging from 10 seconds to 3 minutes, depending on age. Upon recovery, flies exhibited a hyperactive phase, known as a recovery seizure, in which they vigorously flapped their wings and moved their legs for approximately 4-6 seconds, prior to righting themselves to a standing position and walking away. By 42 days, 75.8% of stai B200 adults exhibited bang-sensitive paralysis, phenotype In contrast, homozygous stai B200 larvae (B) and stai rdtp larvae (C) exhibit a posterior paralysis "tail flip" phenotype common for Drosophila mutants defective in axonal transport (Movies S1-S3). The tail-flip phenotype is also observed in heterozygous stai B200 /stai rdtp larvae (Movie S4). Confocal micrographs of sections of segmental nerve from wildtype (D), stai B200 /+ heterozygous (E) and stai B200 homozygous (F) third instar larvae stained with antibody against synaptic vesicle protein cysteine string protein (CSP). (D) The axons of wild type animals showed faint, punctate staining for CSP uniformly throughout the axons within the compound nerve bundle. (E) Although heterozygous stai B200 larvae showed no obvious sluggishness or tail-flip phenotype when they crawled, mild accumulations of CSP were observed in some segmental nerve axons. (F) In contrast, homozygous mutant stai B200 larvae exhibit abundant accumulations of CSP throughout the segmental nerve, confirming a disruption in axonal transport. (G) When stai B200 is placed in heterozygous combination with Df(2L) Exel6015, a chromosomal deletion that spans the stai locus, there is an enhancement in the abundance of axonal accumulation of CSP. (H) The severity of axonal transport defects was quantified by averaging the number of accumulations of CSP immunopositive anterograde and retrograde membranous axonal cargos observed in the segmental nerve axons of stai deficient third instar larvae. To determine if there was an anterior-to-posterior gradient in the severity of axonal transport defects, regions of segmental nerve axons were analyzed as they passed through abdominal segment A2 (light grey bars) and abdominal segment A4 (dark grey bars). Data is also presented as total number of accumulations observed in both abdominal regions (black bars). The only significant regional difference observed in the abundance of axonal clogs in segmental nerve axons between abdominal segment A2 and A4 was in the axons examined from stai B200 /Df(2L) Exel6015 larvae (p<0.0001). Wild type larvae exhibited 0.57 ± 1.21 axonal clogs/50 µm segmental nerve axon in all axons examined. Homozygous stai B200 larvae averaged 13.55 ± 5.56 axonal clogs/50 µm segmental nerve axon (p<0.0001). The average number of axonal clogs observed in the segmental nerve axons of stai B200 /Df(2L) Exel6015 larvae was 15.10 ± 6.07 axonal clogs/50 µm segmental nerve axon. In panels A-C, the scale bar = 1mm. In panels D-G, the scale bar = 10 µm.   Interestingly, the progressive bang-sensitive phenotype was also observed in heterozygous stai B200 /+ and stai rdtp /+ adults, but with a delayed onset and a marked reduction in the recovery period following paralysis compared to stai B200 adults (Tables S2 and S3). A bang-sensitive response to mechanical stimulation was first observed in heterozygous stai mutants at 42 days post eclosion where 25.2% of stai B200 /+ (n=111) and 18.0% of stai rdtp /+ adults (n=50) exhibited the phenotype. The mean recovery period for stai B200 /+ adults was 55.7 ± 27.4 seconds and for stai rdtp /+ adults was 55.4 ± 43.7 seconds. By 56 days post eclosion, the penetrance of the bang-sensitive phenotype had increased to 58.0% of stai B200 /+ and 52.9% of stai rdtp /+ adults, with an increase in the average recovery period to 85.6 ± 43.6 seconds and 124.4 ± 71.8 seconds respectively (Tables S2 and S3). We conclude that viable adult stai mutants exhibit a progressive, age-dependent bang-sensitive phenotype to mechanical stimulation, reflective of compromised nervous system function.

Reversion and Rescue of stai Mutant Phenotypes
To demonstrate that all observed phenotypes are due to reduction in stai activity, we mobilized the piggyBac element in the stai B200 allele to generate the precise excision allele stai excision that restored the stai gene to its original structure and acted as a genotype matched control. All stai B200 mutant phenotypes were reverted to wildtype in stai excision animals. Homozygous stai excision third instar larvae did not exhibit the posterior paralysis and tail-flip phenotype observed in stai B200 larvae (n=0/120) (Table S1 and Movie S6). In addition, the aggregation of CSP immunoreactive vesicles in larval segmental nerves of stai excision larvae was not significantly different than wildtype (0.91 ± 1.37 CSP accumulations/50 µm segmental nerve axon, n=70, p=0.68, compare Figure 4A and B Figure 4E). The average adult lifespan was also reverted to wildtype (74.6 ± 1.2 days) (P≥0.05) ( Figure 3A, 3B). Finally, stai excision adults did not exhibit the progressive bang-sensitive phenotype observed in stai mutant adults (Tables S2 and S3).
Further, we were able to partially rescue all phenotypes observed in stai B200 third instar larvae and viable adults with the ubiquitous ectopic expression of staiB2, a neuronal specific Drosophila stai isoform (tub-staiB2 -gift of Pernille Rorth, IMCB, Singapore) [54]. The severity of the posterior paralysis phenotype was greatly reduced in stai B200 ; tub-staiB2/+ larvae compared to stai B200 larvae (Table S1 and Movie S7). Immunohistological analysis revealed that the tub-staiB2 transgene significantly reduced the severity of the aberrant accumulation of CSP observed in segmental nerves of homozygous stai B200 larvae (6.79 ± 6.43 CSP accumulations/50 µm segmental nerve axon, n=52, p<0.0001, compare Figure  4A and 4C Figure 4E). The tub-staiB2 transgene also delayed the onset and ameliorated the severity of the bang-sensitive phenotype exhibited by stai B200 males from 75.8% to 5.8% (n=52) at 42 days post eclosion, while significantly reducing the mean recovery period to 39.7 ± 12.5 seconds from 76.6 ± 41.4 seconds (Tables S2 and S3). A single copy of the tub-staiB2 transgene also significantly increased average adult lifespan to 68.5 ± 2.3 days in stai B200 ; tub-staiB2/+ adults, and to 64.8 ± 2.3 days in stai B200 /Df(2L) Exel6015; tub-staiB2/+ adults compared to the average lifespan of stai B200 and stai B200 /Df(2L) Exel6015 adults (P<0.001) ( Figure 3B). While the ectopic expression of staiB2 rescued the posterior paralysis and aberrant accumulation of CSP in segmental nerve axons of stai rdtp homozygous larvae (data not shown), it did not rescue lethality, supporting the idea that the stai rdtp chromosome carries a linked, lethal mutation. Finally, we were also able to rescue the stai mutant phenotypes with the ubiquitous ectopic expression of the human stathmin gene, STMN1. A single copy of the tub-STMN1 transgene significantly reduced the aberrant accumulation of CSP in the segmental nerve axons of stai B200 larvae (3.92 ± 3.87 CSP accumulations/50 µm segmental nerve axon, n=84, p<0.0001, compare Figure 4A and 4D Figure 4E). The tub-STMN1 transgene also reduced the percentage of stai B200 larvae that exhibited a posterior paralysis phenotype from 76.7% (n=92/120) to 39.6% (n=44/111) (Table S1). In addition, the severity of the posterior paralysis phenotype was reduced to only 9.1% (n=4/44) of stai B200 ; tub-STMN1 larvae that exhibited a robust tail-flip from 69.6% (n=64/92) of stai B200 larvae (Table S1 Movie S8). These data support previous observations that the Drosophila and human stathmin proteins are functionally equivalent orthologs.
Although all observed stai mutant phenotypes could be rescued with both Drosophila staiB2 and human STMN1 transgenes, rescue was incomplete. This was likely due to the fact that the rescuing Drosophila staiB2 transgene expresses only one of the four known stai protein isoforms, which may have been insufficient for complete rescue. In addition, expression of both transgenes was under the control of the tubulin promoter that may not have provided the correct spatial and temporal expression of the transgenes.

Loss of stai Function Reduces Tubulin and Kinesin Heavy Chain Protein Levels
Since stai functions as a tubulin-binding protein and is known to regulate MT dynamics, we used western blot analysis to examine and quantify tubulin levels in total protein extracts isolated from stai mutant third instar larvae and adults. Levels of α-tubulin were significantly reduced in stai B200 /+ larvae compared with wild type controls (p<0.0001, Figure 5A, 5B). A more pronounced reduction in the levels of α-tubulin protein was observed in homozygous stai B200 larvae ( Figure 5A). Third instar stai B200 larvae had levels of α-tubulin that were 15.0 ± 16.6% (mean ± SD) of that observed in wild type larvae (p<0.0001, Figure 5A, 5B). In order to assess the stability of polymerized MTs, we also assayed for levels of acetylated αtubulin in stai mutant larvae. The levels of acetylated α-tubulin were also significantly reduced in stai B200 /+ mutant larvae (p<0.0001, Figure 5A, 5B), and more greatly reduced in homozygous stai B200 mutant larvae to 36.9 ± 12.9% of the levels observed in wildtype (p<0.0001, Figure 5A, 5B). The posterior paralysis displayed by stai homozygotes is reminiscent of the 'tail flip' phenotype first described in Kinesin heavy chain (Khc) mutant larvae [29]. We therefore examined protein extracts from stai mutants to see if Khc levels were affected. We observed a significant reduction in the level of Khc, the force generating subunit of the MT motor protein kinesin-1, in stai mutant larvae ( Figure 5A). Heterozygous stai B200 /+ larvae had levels of Khc protein that were 91.2 ± 4.2% of that observed in wildtype (p<0.05) whereas stai B200 larvae had levels that were 53.8 ± 2.3% of wildtype (p<0.0001, Figure  5A, 5B). Collectively, these data suggest a general impairment of the MT-based transport system following loss of stai function in Drosophila.
Given the observed effect of loss of stai function on the levels of α-tubulin, acetylated α-tubulin and Khc protein in third instar larvae ( Figure 5A), and the observation that ectopic expression of staiB2 rescued the shortened lifespan of stai B200 adults ( Figure 3A), we analyzed the levels of α-tubulin and kinesin-1 motor components in adult stai mutant animals at 1 day and 21 days post-eclosion ( Figure 5C). These ages were chosen because they corresponded to the age at which a bang-sensitive response to mechanical stimulation was first observed in stai deficient animals (Table S2). At 1 day posteclosion, the levels of α-tubulin, Khc and Kinesin light chain (Klc) subunits were equivalent across all genotypes analyzed ( Figure 5C). However, by 21 days post-eclosion, there was a marked reduction in the levels of α-tubulin, Khc and Klc proteins in stai B200 adults. Surprisingly, ectopic expression of staiB2 not only restored the levels of α-tubulin in stai B200 ; tub-staiB2/+ adults as predicted, but also increased the levels of both the Khc and Klc protein subunits ( Figure 5C).

Genetic Reductions in the Kinesin Heavy Chain Gene Enhance stai Phenotypes
Because loss of stai function resulted in decreased levels of Khc protein in third instar larvae, and ectopic expression of staiB2 restored reduced levels of conventional kinesin motor protein subunits in adult flies, we assayed for a genetic interaction between stai and kinesin by combining a genetic reduction in the gene that encodes the conventional kinesin heavy chain protein, Khc, with mutations in the stai gene. Loss  Representative western blots of total protein extracted from stai mutant third instar larvae. Heterozygous stai/+ mutant larvae exhibit mild reductions in the levels of α-tubulin and acetylated α-tubulin compared with wild type larvae. Homozygous stai mutant larvae have dramatic reductions in the levels of α-tubulin, acetylated α-tubulin and the heavy chain subunit of the microtubule motor protein kinesin compared to wildtype. Df(2L) BSC239 represents a second chromosomal deficiency that excludes the stai gene. GAPDH is used as a loading control. (B) Quantification of western blots of total protein extracts isolated from third instar larvae probed with antibodies against α-tubulin (mAB DM1A), acetylated α-tubulin (mAB 611B1), and kinesin heavy chain (AKIN01). Quantification was performed as described in the Materials and Methods. Results are normalized against wildtype and presented as mean ± S.D. Statistical comparison reveals that the levels of α-tubulin, acetylated α-tubulin, and kinesin heavy chain proteins for all genotypes examined differ significantly from wildtype (p<0.001), with the exception of levels of acetylated α-tubulin in stai rdtp /+ larvae (**p<0.01) and kinesin heavy chain in stai B200 /+ (*p<0.05). (C) Western blot analysis of total protein extracted from stai mutant adult Drosophila aged 1 day and 21 days post eclosion. The levels of α-tubulin, and conventional kinesin motor heavy and light chain subunits are present at equal levels in wildtype, stai B200 /+, stai B200 and stai B200 ; tub-staiB2 adults 1 day post eclosion. At 21 days post eclosion, the age we observe stai B200 adults exhibiting a bang-sensitive response to mechanical stimulation, the levels of α-tubulin, and conventional kinesin motor heavy and light chain subunits are noticeably reduced. Introduction of an exogenous copy of the rescuing transgene tub-staiB2, not only increases the levels of α-tubulin, but also increases levels of the conventional kinesin motor heavy and light chain subunits. of a single copy of Khc did not cause a posterior paralysis phenotype or enhance the axonal transport defects observed in stai B200 /+ larvae (data not shown). However, introduction of the Khc 20 null allele into the homozygous stai B200 genetic background increased the penetrance of the posterior paralysis phenotype observed in stai B200 Khc 20 /stai B200 third instar larvae to 97.1% (n=101/104) compared with only 76.7% (n=92/120) of stai B200 homozygous larvae (Table S1). In addition, we observed a small, but significant, increase in the severity of aberrant accumulation of CSP in the segmental nerve axons of stai B200 Khc 20 /stai B200 compared to stai B200 third instar larvae (15.89 ± 3.80 CSP accumulations/50 µm segmental nerve axon, n=48, p<0.01, Figure 4E).

Loss of stai Function Alters the Integrity of Axonal MTs
MTs are the main cytoskeletal component of axons. The motor proteins kinesin and cytoplasmic dynein mediate organelle transport on MT tracks within the axon. Given the known function of stai as a regulator of MT dynamics and our demonstration that loss of stai function disrupts axonal transport and reduces the levels of α-tubulin and acetylated αtubulin in third instar larvae, we assayed the integrity of MTs in segmental nerve axons by immunostaining stai mutant larvae with antibody against the neuronal specific MAP Futsch (mAb 22c10), the Drosophila homolog of the vertebrate MAP1B protein that reveals stabilized MTs in neurons [55].
Wild type larvae have a well-defined MT network in segmental nerve axons, as determined by the even distribution of Futsch protein along the length of individual axons within the compound nerve ( Figure 6A). In contrast, the distribution of Futsch protein appeared to be disrupted and less consistent along the length of segmental nerve axons of stai mutant larvae, intense and prominent in some axonal regions (arrows Figure 6B), yet noticeably reduced and absent in other regions of the same axon (small arrowheads Figure 6B). We interpret this observation as a disruption in Futsch-stabilized MTs within stai deficient segmental nerve axons.
We also observed a dosage dependent reduction in the intensity of Futsch staining in the axons of segmental nerve bundles of stai deficient larvae. We quantified and compared the average pixel intensity of Futsch staining in axons of third instar larvae segmental nerves (Table S4). The axons of stai B200 /+ third instar larvae exhibited a significant reduction in the average pixel intensity of Futsch immunostained segmental nerve axons compared to that observed in wildtype axons (WT 19000 ± 5200 arbitrary units (a.u.), n=43; stai B200 /+ 12600 ± 4800 a.u, n=42, p<0.0001, Figure 6G, Table S4). The average intensity of Futsch staining was further reduced in the axons of stai B200 homozygous third instar larvae (9000 ± 3000 a.u, n=32 p<0.0001, compare Figure 6A and B) and greater yet in the axons of stai B200 /Df(2L) Exel6015 larvae (7600 ± 2200 a.u, n=30, p<0.0001, compare Figure 6A and C) compared to wildtype. In addition to acting as tracks to support transport through the axon, MTs also provide internal structural support that influences and maintains the morphology of the axon. Individual axons that comprise the compound segmental nerve of wild type larvae are arranged in linear, ordered bundles ( Figure 6A). In contrast, stai deficient larvae exhibited alterations in the morphology of individual axons within the segmental nerve that are not observed in wildtype or stai/+ larvae (large arrowheads Figure 6B). Individual axons were often bent and non-linear in stai deficient segmental nerves. This observed alteration in axon morphology was more pronounced and affects a greater number of segmental nerve axons of stai B200 /Df(2L) Exel6015 larvae ( Figure 6C).

Loss of stai Function Alters the Integrity of Muscle MTs
The observed phenotypes following loss of stai function are consistent with a reduction in the level of polymerized MTs in the axons of stai mutant animals, contradictory to the function of stai as a MT depolymerizing protein. Because the integrity of MTs within individual segmental nerve axons is difficult to assess by direct visualization, we also looked at the effect of loss of stai function on the architecture of the MT network in body wall muscle 6 from abdominal segment A4 of our stai mutant larvae.
The MT network in muscle cells from wild type larvae is very well defined, enriched around the nucleus and extending into the cytoplasm (Figure S1 A). The MT network of stai/+ muscle cells appears indistinguishable from wildtype (data not shown). The MT network is noticeably less dense, however, both around the nucleus and within the cytoplasm of muscle cells from stai B200 mutant larvae (compare Figure S1 A and B). The reduction in polymerized MTs is even more pronounced in stai B200 /Df(2L) Exel6015 larvae (compare Figure S1 A, B and C) confirming that the stai B200 allele retains residual stai activity. This observation directly supports our results from western blot data regarding the observed reduction in levels of acetylated αtubulin in stai and stai/Df(2L) Exel6015 third instar larvae. Interestingly, we detect an abundance of fragmented MTs within the cytoplasm of stai deficient cells (arrowheads Figure  S1 B, C).
We quantified the density of perinuclear MTs in muscle cells of our stai mutant animals and found that MT density was significantly reduced in stai B200 (P<0.05) and stai B200 /Df(2L) Exel6015 (P<0.01) larvae (Figure S1 G). The architecture of the MT network was restored to wildtype in stai excision muscle cells (Figure S1 D) and rescued with expression of both the tub-staiB2 (Figure S1 E), and the human tub-STMN1 transgenes ( Figure S1 F). Rescue of the MT network in stai B200 larvae with the tub-staiB2 transgene resulted in a perinuclear MT density greater than wildtype (P<0.01) (Figure S1 G).

Discussion
The development and function of the nervous system is dependent on a highly dynamic, yet tightly regulated, microtubule (MT) cytoskeleton (reviewed in [2,57]). The stathmin (stai) family of proteins comprise a group of MT destabilizing proteins that indirectly function to regulate microtubule dynamics by binding and sequestering free tubulin and stimulating MT depolymerization [16]. It has been previously demonstrated in Drosophila that stai is maternally contributed to the oocyte and is required for the development of the embryonic nervous system [58]. The recessive, hypomorphic alleles we have isolated in the Drosophila stai gene have allowed us to circumvent the requirement for stai during embryonic development, and identify a function in the maintenance of axonal MTs in the mature nervous system, necessary to support and sustain axonal transport.
Collectively, our results indicate that stai function is required for axonal transport. Phenotypes consistent with impaired axonal transport, including a posterior paralysis tail-flip and accumulation of synaptic proteins in peripheral segmental nerve axons, were observed in stai mutant larvae. Furthermore, genetic perturbations that independently affect axonal transport (such as a 50% reduction in Kinesin heavy chain (Khc) copy number) enhance the severity of the stai mutant phenotype including the accumulation of synaptic proteins in segmental nerve axons. We also observe reductions in the levels of tubulin, acetylated tubulin and kinesin heavy and light chain protein subunits in our stai mutants. Importantly, Khc levels were reduced by less than 50%. Since Khc hemizygotes are phenotypically normal, our data argue that the observed reduction in Khc levels contribute to the stai mutant phenotype, but are not its primary cause. These phenotypes are ameliorated by ectopic expression of a Drosophila stai transgene encoding a neuronal specific stai isoform. In addition, we are able to rescue the observed phenotypes with ectopic expression of the human gene STMN1. Mutations in cytoplasmic dynein subunits and motor protein accessory factors also result in axonal transport defects. A potential direction for future studies could be to comprehensively examine how microtubule motor protein expression is affected in Drosophila stai mutants. A comparison of STMN1 +/+ vs. STMN1 -/-mouse embryonic fibroblasts identified motor proteins as the largest group of genes with altered expression profiles. Interestingly, while differential expression of 13 kinesin motor subunits was detected alteration in dynein motor subunit expression was not observed, and only 5 genes encoding dynein light or intermediate chains showed altered expression [59].
The severity of axonal transport blockages observed in stai mutant axons is significantly less than that observed in microtubule motor mutants. It has previously been reported that Drosophila third instar larvae null for the microtubule motor proteins Dynein heavy chain (Dhc) and Khc exhibit 250 ± 25 clogs per 50 µm length of segmental nerve axon [60]. Third instar larvae of the genotype stai B200 /Df(2L) Exel6015 exhibit on average 15.10 ± 6.07 clogs per 50 µm length of segmental nerve, significantly less than reported in Dhc and Khc null third instar larvae. The severity of axonal transport blockages observed in stai mutant larvae, however, is equivalent to levels observed in larvae null for the amyloid precursor protein-like gene Appl [61]. Interestingly, Appl null larvae are also homozygous viable, suggesting that the severity of the disruption of axonal transport may determine viability and in part explain why a significant percentage of stai B200 and stai B200 /Df(2L) Exel6015 mutants are homozygous viable.
We routinely recover viable stai B200 /Df(2L) Exel6015 adult flies, despite undetectable levels of stai transcript as measured by qPCR analysis, suggesting that stai function is not necessary for viability. This observation is consistent with a recent report that stai knockout flies are viable, indicating that stai is a non-essential gene [45]. The stai mutant adult flies exhibit a significantly reduced lifespan and a progressive, agedependent epileptic-like seizure and paralysis in response to mechanical stimulation. To our knowledge, the bang-sensitive phenotype has not been previously reported in motor-protein mutants, and is not specific to defective axonal transport. However mutations in the Khc gene have been shown to exhibit a temperature-sensitive paralysis, a phenotype found in bang-sensitive paralytic (para) and maleless (mle) mutants, and interact with para and mle to cause lethality [62]. Khc mutants also suppressed the Shaker and ether-a-go-go leg shaking phenotype, which is similarly suppressed by para and mle alleles [62]. Hurd et al. conclude that Khc is likely required for anterograde transport of vesicles bearing sodium membrane channels and that reduced transport of these channels results in reduced sodium current. Likewise, the identification of a progressive bang-sensitive phenotype in the stai mutants we have analyzed, suggests that the transport of ion channels to the axon membrane is impaired.
Our observation of age-dependent phenotypes in Drosophila stai mutants parallels the observation that STMN1 knockout mice initially develop normally and display no overt phenotypes [63], but when aged exhibit a minor axonopathy [64] and behavioral deficits that include a lack of both learned and innate fear responses [65]. Interestingly, altered levels of STMN1 expression in human brain tissue has recently been linked to intractable temporal lobe epilepsy [66].
It has recently been reported that stai is required for stability of the neuromuscular junction in Drosophila, a phenotype that likely arises as a result of impaired axonal transport [67]. While Graf et al., present evidence of perturbed axonal transport, analysis of the integrity of axonal MTs did not reveal alterations in MT architecture. Our analysis supports the observation that stai function is necessary for axonal transport. In contrast to the studies of Graf et al., however, we also present evidence that the MT cytoskeleton of the axon is disrupted in stai mutant larvae and likely the cause of impaired axonal transport. We demonstrate this in three ways. First, we observe a reduction in the intensity of Futsch and acetylated α-tubulin staining in the segmental nerve axon that correlates with a reduction in stai dosage, and an increase in the severity of CSP accumulation in the axon. Second, we observe alterations in the organization of Futsch staining in individual segmental nerve axons lacking stai function. Finally, within the segmental nerve, individual axons exhibit an altered, distorted morphology in which they undulate along the length of the compound nerve bundle. We interpret this phenotype as a consequence of reduced axonal MTs that result in a loss of internal structural support for the axon. Interestingly, this phenotype is similar to the morphology of axons that are subjected to dynamic stretch injury, resulting in the breakage of MTs within the axon that correlate with undulations in axonal morphology [68]. One possible explanation for the observed differences in axonal MT morphology reported by each group is perdurance of maternally contributed stai product in differently aged third instar larvae examined in each study.
A wealth of evidence indicates that stai family members are negative regulators of MT dynamics. Decreasing the levels of stai in vivo increases polymerized MTs, as demonstrated by microinjection of neutralizing anti-stathmin antibody into newt lung cells [69], shRNA depletion of endogenous stathmin in human interphase cells [70,71], and knockout of stathmin function in MEFs [27], and amygdala slices from whole brains of stathmin knockout mice [65]. Conversely, microinjection of wildtype stathmin protein into newt lung cells results in massive loss of MT polymer [72]. Surprisingly, in contrast to vertebrate systems, loss of stai function results in dramatic reduction in the levels of both α-tubulin and acetylated α-tubulin. This reduction was visible on western blots as well as by immunohistochemical visualization that showed that polymerized MTs were reduced in our stai mutant Drosophila.
It is not clear why we see a reduction in the levels of polymerized tubulin in our stai mutant Drosophila. It has been previously demonstrated that the addition of recombinant Drosophila stai to purified bovine brain tubulin in vitro results in a decrease in MT polymerization in a dose dependent manner, while the expression of three different Drosophila stai isoforms, staiA1, staiB1 and staiB2 in HeLa cells in vivo depolymerizes the MT network in a subset of cells [22]. These observations indicate that Drosophila stai isoforms also act as MT depolymerizers in mammalian cells and mammalian cell extracts. It is possible that the MTs within the axonal compartment of a Drosophila nerve cell are subject to extrinsic signals, regulatory factors and MAPs that are distinct and not present in other subcellular regions of a neuron, or in different cell types. While this may be true, it is an unsatisfactory explanation because we also observe a reduction in polymerized MTs in muscle cells that comprise the body wall musculature.
Another possibility is that reduction in stai could alter the expression of a gene that encodes a MT polymer stabilizer. However, a more likely explanation for the observed reduction of the MT network in stai mutant Drosophila, is that the phenotype is the consequence of a complete depletion of all stai protein activity. Unlike vertebrate model systems, the entire stai family of proteins are encoded by a single gene in Drosophila that produces four stai protein isoforms [22]. Therefore, mutation in the stai gene removes or greatly reduces the activity of all stai proteins. In contrast, in studies performed in vertebrate model systems the function of only one member of the stai family is altered, and genetic redundancy by the remaining stathmin-like proteins may compensate and account for the observed phenotypic differences. Indeed, it has been reported in STMN1 knockout MEFs that the expression of STMN2 is decreased and STMN4 is increased [59], while in STMN1 knockout mice, an increase in the expression of STMN3 and STMN4 is observed [64,73]. The stai protein has two known activities; a MT depolymerizing activity as well as a tubulin binding activity. Depleting stai not only removes the MT depolymerizing activity of the protein, which is expected to result in an increase in the polymerized MT network, but it also abolishes its tubulin binding activity as well, which may be required to maintain the polymerized MT network. The net effect of removing both functions entirely may be a gross reduction in the MT network.
Observation of the MT network in muscle cells of third instar larvae may provide a clue as to how the MT network becomes reduced in stai mutant larvae. The distinct architecture allowed us to observe an abundance of fragmented or severed MTs in the peripheral cytoplasm of cells lacking stai activity. The MT cytoskeleton of a cell is subject to both stabilizing and destabilizing factors that act coordinately to maintain the MT architecture. For example, an antagonistic relationship in maintaining the MT network has been demonstrated between tubulin-specific chaperone E (TBCE), required for the formation of tubulin heterodimers, and the microtubule severing protein spastin [74]. It is possible that stai acts downstream of TBCE, and its removal results in a diminished MT network, as observed in TBCE mutant larvae, suggesting an antagonistic relationship with MT severing proteins spastin and/or katanin. It is interesting to note that mutation of TBCE in mice results in a progressive motor neuropathy [75,76] while mutations in spastin are the cause of the most common form of hereditary spastic paraplegia in humans, likely through disruption of axonal transport [12][13][14].
Stai is a cytosolic phosphoprotein whose function is regulated by many different signaling pathways [77]. The ability of stai to bind tubulin is dependent on its phosphorylation state. In Xenopus metaphase extracts, the phosphorylation of stathmin is in part regulated by the serine/threonine type-2A phosphatase PP2A [78]. PP2A is known to associate with MTs and is thought to regulate MTs through phosphorylation of the MAP Tau [79,80]. Interestingly, PP2A is a component of the striatin-interacting phosphatase and kinase (STRIPAK) complex, of which the Monopolar spindle-one-binder protein 3 (Mob3/Phocein) is a known component [81]. DMob4, the Drosophila Phocein homolog has recently been demonstrated to regulate microtubule organization and axonal transport [82].
Defects in axonal transport are known to cause or contribute to neurodegenerative disease and it is possible that stai may play a role in some cases. In mammals, loss of stai function results in mild, progressive phenotypes consistent with neurodegenerative disease [64]. In addition, a link has been established between the altered expression of stathmin and the abnormal architecture of MTs and organelle transport in motor axons of a mouse model of the motor neuron degeneration disorder spinal muscular atrophy [83]. Our in vivo analysis in Drosophila has identified a role for stai in regulating the architecture of MTs in the axon of peripheral nerves, necessary to support and sustain axonal transport.  The penetrance and severity of the posterior paralysis phenotype was scored and quantified by measuring the severity of the taiflip, as determined by the angle that the tail was raised above the substrate on which the larvae crawled. If the tail was raised greater than 40 o above the horizontal crawling plane during the crawling cycle, larvae were scored as having a robust tail-flip. If the tail was raised less than 40 o above horizontal during the crawling cycle, larvae were scored as having a mild tail-flip. If larvae exhibited a normal crawling behavior, they were scored as having no tail-flip. The crawling behavior of at least one hundred larvae was analyzed, for a minimum of one minute each, for each genotype tested. (XLSX )   Table S2. Percentage of Adult Drosophila Exhibiting a Bang-Sensitive Phenotype. Adult male Drosophila were assayed for paralytic behavior in response to mechanical stimulation at six different age periods post eclosion. In wild type controls, a paralysis in response to mechanical stimulation was observed with a very low penetrance of 1.5% at 42 days of age, and 3.3% at 56 days of age. This response is attributed to the general effects of aging on the adult nervous system. A pronounced bang-sensitive phenotype is observed in 20.9% of homozygous stai B200 mutant animals tested at 21 days of age, increasing to 75.8% by 42 days of age and 91.7% at 56 days of age. The bang-sensitive phenotype is also observed at appreciable levels in stai B200 /+ and stai rdtp /+ flies, but with a later onset than observed in stai B200 homozygous adults. At 42 days of age, 25.2% of stai B200 /+ and 18.0% of stai rdtp /+ adults exhibit a bang sensitive phenotype, increasing to 58.0% of stai B200 /+ and 52.9% of stai rdtp /+ flies by 56 days. Ubiquitous expression of a rescuing transgene, tub-staiB2, delays the onset of the bang-sensitive response to mechanical stimulation to 24.6% of stai B200 /+; tub-staiB2/+ animals aged 56 days. A minimal response to mechanical stimulation, not significantly different that wildtype, was observed in stai rdtp /+; tub-staiB2/+ animals aged 56 days. The tub-staiB2 transgene also delayed and ameliorated the bang-sensitive phenotype in stai B200 homozygous adults. The response to mechanical stimulation by genotype matched stai excision adults is not significantly different than the response of wild type control animals. A minimum of fifty adult Drosophila were tested at each age for each genotype analyzed.