Drosophila Torsin Protein Regulates Motor Control and Stress Sensitivity and Forms a Complex with Fragile-X Mental Retardation Protein

We investigated unknown in vivo functions of Torsin by using Drosophila as a model. Downregulation of Drosophila Torsin (DTor) by DTor-specific inhibitory double-stranded RNA (RNAi) induced abnormal locomotor behavior and increased susceptibility to H2O2. In addition, altered expression of DTor significantly increased the numbers of synaptic boutons. One important biochemical consequence of DTor-RNAi expression in fly brains was upregulation of alcohol dehydrogenase (ADH). Altered expression of ADH has also been reported in Drosophila Fragile-X mental retardation protein (DFMRP) mutant flies. Interestingly, expression of DFMRP was altered in DTor mutant flies, and DTor and DFMRP were present in the same protein complexes. In addition, DTor and DFMRP immunoreactivities were partially colocalized in several cellular organelles in larval muscles. Furthermore, there were no significant differences between synaptic morphologies of dfmrp null mutants and dfmrp mutants expressing DTor-RNAi. Taken together, our evidences suggested that DTor and DFMRP might be present in the same signaling pathway regulating synaptic plasticity. In addition, we also found that human Torsin1A and human FMRP were present in the same protein complexes, suggesting that this phenomenon is evolutionarily conserved.


Introduction
Dystonia is a phenomenological disorder encompassing various degrees of involuntary muscle contractions often initiated or worsened by voluntary actions [1]. Even though there are many different forms of heritable dystonia, Dystonia 1 (DYT1) dystonia has unique clinical characteristics such as early age of onset accompanied by generalized manifestation and progression of symptoms [1]. DYT1 dystonia is mainly caused by a 3 bp in-frame deletion in TOR1A resulting in a loss of glutamic acid (ΔE) in Tor1A [2]. Among humans with heterozygous Tor1A ΔE mutations, only ∼30% of them manifested severe symptoms in their adolescence, and the patient's life expectancy was not affected [3]. In addition, patients with homozygous Tor1A ΔE mutations have not yet been reported [2,3]. These inheritance patterns of DYT1 dystonia suggest that complete loss of Tor1A function may cause lethality during embryogenesis and that heterozygous Tor1A ΔE mutation may increase susceptibility to other genetic or environmental risk factors that caused severe symptoms [4,5].
To understand both the molecular and cellular basis of this disorder and the in vivo functions of Tor1A, various animal models have been generated. Because homozygous DYT1 global knock-in (Tor1A ΔE/ΔE ) or global knockout (Tor1A −/− ) mice died shortly after birth [6], homozygous mutant mice could not be used to study the abnormal behaviors. To overcome these problems, a mouse model in which Tor1A was knocked out specifically from the central nervous system was generated and showed severe dystonic 2 Neural Plasticity symptoms even though those mice died within 16 days after birth [7].
Other mouse models generated by overexpressing Tor1A ΔE in neurons showed very weak behavioral symptoms such as defects in beam walking [4], motor learning [8,9], or paw-print analysis [9,10]. Heterozygous knock-in mutant Tor1A (ΔE/+) mice exhibited defects in beam-walking test and hyperactivity in open field test [11]. These results suggested that Tor1A ΔE alone may be sufficient to alter neuronal circuits to cause mild behavioral defects, but not sufficient to evoke severe dystonic symptoms. To explain this possibility, the second hit models were suggested [5,8].
Drosophila models provide unique opportunities to investigate the cellular and molecular etiologies underlying various neurological disorders in humans [12]. In a previous study, we generated a Drosophila model of DYT1 dystonia by expressing human Tor1A ΔE (HTor1A ΔE ) [5]. We found that expression of Tor1A ΔE induced formation of protein aggregates and abnormalities at synapses and nuclear membranes. In addition, dystonic symptoms were observed when additional environmental stress was present. Furthermore, we found that expression of Tor1A ΔE activated unfolded protein responses and upregulated heat shock protein 22 to compensate for the toxic effects of mutant Tor1A proteins [13]. However, the in vivo function of Drosophila Torsin (DTor) in the maintenance and regulation of locomotor control and synaptic plasticity is still not clear.
Maintenance of normal motor control is one of the most important factors for the survival and development of animals. Recent studies have suggested that proper motor control is regulated by many different types of neural circuits in central and peripheral nervous systems acting in harmony [14]. Mutations that cause severe defects in motor control in humans are known to cause various defects in nervous systems [15]. To investigate the in vivo functions of Torsin proteins, we generated transgenic flies expressing fulllength DTor-cDNA or DTor-RNAi constructs. We performed various experiments to elucidate behavioral, biochemical, physiological, and proteomic consequences of altered Torsin protein levels in Drosophila. We found that Torsin proteins formed complexes with Fragile-X mental retardation protein (FMRP) in Drosophila and humans, and the neuromuscular junction (NMJ) morphology of Drosophila fmrp (dfmrp) mutant larvae expressing DTor-RNAi was similar to that of dfmrp mutant larvae, suggesting possible involvement of FMRP in DYT1 dystonia in regulation of motor control and synaptic plasticity.

Drosophila Genetics.
Drosophila used in this study were raised at 25 ± 1 ∘ C with 60% relative humidity. The open reading frame of Drosophila Torsin (DTor) was amplified from a Drosophila cDNA library and verified by sequencing both strands of cloned DNA. The DTor-cDNA was subcloned into pUAST germ line transfer vectors. The first 500 bp of the 1st exon and intron of DTor genomic DNA was amplified, cloned, and subcloned into pWIZ germ line transformation vectors designed to express double-stranded inhibitory RNA [18] after verifying its sequence. Transgenic flies were generated using typical germ line transfer techniques. Additional RNAi flies used in this study were obtained from the Vienna Drosophila RNAi Center (Stock No. 30985, VDRC, Vienna Biocenter Core Facilities, Vienna, Austria). Previously characterized UAS-HTor1A WT , Tubulin-Gal4 (Tub-Gal4), C155-Gal4, and C57-Gal4 flies were used [5,19]. dfmrp null mutants, fmr1 Δ50M /TM6b, Tb 1 were obtained from the Bloomington Drosophila Stock Center (Indiana University, Bloomington, IN, USA). fmr1 Δ50M and Tub-Gal4 chromosomes were recombined to generate dfmr1 Δ50M , Tub-Gal4/TM6b, Tb 1 . pWiz-DTor-RNAi; Tub-Gal4/TM6b, Tb 1 flies were generated by recombining the 2nd and 3rd chromosomes.

Antibodies and Immunohistochemistry.
To produce antibodies for DTor, His-tagged DTor recombinant peptides (residues 25-210) were expressed in Escherichia coli M15 carrying the pREP4 plasmid and then purified using Ni-NTA beads (Qiagen, Valencia, CA, USA). After SDS-PAGE separation, the fusion proteins were injected into rabbits for antibody generation.

PNGase F Treatment.
Proteins from larval body wall muscle preparations were extracted and then treated with PNGase F according to the manufacturer's protocol (New England Biolabs Inc., Ipswich, MA, USA). Briefly, 100 g of protein was mixed with glycogen denaturing buffer (New England Biolabs Inc.) and incubated for 10 min at 94 ∘ C. After adding NP-40, G7 buffer, BSA, and PNGase F (New England Biolabs Inc.), samples were incubated for 4 hr at 37 ∘ C. After boiling with SDS sample buffer, samples were loaded on a 12% SDS-PAGE gel. Rabbit anti-DTor antibodies were used to detect DTor proteins.
2.6. Behavioral Assays. One group of flies consisting of five males and five females of identical ages was tested for locomotor abilities as described in the literature [5]. After flies were transferred into a vial, the vials were transferred into a preheated 38 ∘ C water bath. The locomotor ability of each fly was determined, that is, whether it was climbing, walking, or staying at the bottom in vial at each time point. In total 120 flies were collected from three independent crosses and were raised at 30 ∘ C to accelerate the aging process. Gal4 driver flies were crossed with w1118, UAS-DTor-cDNA, or pWiz-DTor-RNAi flies and then compared. Twosample -tests performed using Minitab software (Minitab Inc.) were employed to determine differences in locomotor ability among experimental and control flies.

2-Dimensional Electrophoresis and Proteomic
Analysis. Proteins are differentially expressed between Tub-Gal/+ and pWIZ-DTor-RNAi/+; Tub-Gal4/+ were examined by 2-dimensional (2D) electrophoresis. Protein extracts were obtained from 10-day-old adult fly heads by homogenizing for 5 min using a manual pestle in 100 L lysis buffer [8.0 M urea, 18 mM DTT, 4% (w/v) CHAPS, 40 mM Tris-HCl (pH 8.0), 10 mM EDTA, 0.5% IPG buffer (pH 4∼7, GE Healthcare, Buckinghamshire, England)] with protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland). Then, 300 L lysis buffer was added and supernatant was harvested by spinning down at 15,000 ×g for 10 min at 4 ∘ C. After the protein concentration was quantified using the Bradford assay, 200 g protein was loaded on immobilized linear gradient strips (pH 4-7) using the IPGphor system (GE Healthcare). After the strips were rehydrated for 12 hr, focusing was accomplished in three steps: 200 V for 1 hr, 500 V for 1 hr, and 8,000 V for 8 hr. Before running the 2nd dimension on a 12% polyacrylamide SDS gel using an Ettan DALT electrophoresis kit (GE Healthcare), strips were incubated in 1. The spots for subsequent protein identification analysis were selected by comparing normalized volumes of protein spots from three different comparisons using a one-way analysis of variance (ANOVA; threshold of significance, < 0.05).
The proteins presented in spots were identified by MS/MS experiments performed with a nano-LC/MS system composed of a Surveyor HPLC system (Thermo Fisher Scientific) and an electrospray ionization-(ESI-) quadrupole ion trap (IT) mass spectrometer (MS) (LCQ Deca XP-Plus, Ther-moFisher Scientific) equipped with a nano-ESI source. A C18 trap column (ID 300 m, length 5 mm, particle size 5 m; DIONEX/LC Packings, Sunnyvale, CA) was used for desalting and concentrating 10 L digested peptide samples at a flow rate of 20 L/min. Trapped peptides were separated by back-flushing into a home-made microcapillary column (150 mm in length) packed with C18 resin (particle size 5 m) in 75 mm silica tubing (8 mm ID orifice) using mobile phase A (0% CH 3 CN, 0.02% HCO 2 H, 0.5% CH 3 COOH) and B (80% CH 3 CN, 0.02% HCO 2 H, and 0.5% CH 3 COOH). The gradient elution was started with 5% of the mobile phase B and 95% of the mobile phase A for 15 min. Mobile phase B was increased to 20% for 3 min, 50% for 32 min, 60% for 5 min, 80% for 5 min, and 100% for 10 min. When samples were changed, the column was equilibrated and cleaned with 5% CH 3 CN for 10 min. The operating conditions for obtaining the MS and MS/MS spectra was a capillary temperature of 220 ∘ C, an ESI voltage of 2.5 kV, and a collision energy setting of 35%. The three most abundant MS ions were selected as peaks and became the targets of MS/MS analysis with 180second dynamic exclusion. MS scan functions and HPLC solvent gradients were controlled by an Xcalibur data system (Thermo Fisher Scientific). A Drosophila nonredundant protein database at NCBI (ftp://ftp.ncbi.nlm.nih.gov/, May 2009 Released version, the number of entries were 220,666) was used for searching MS/MS spectra using SEQUEST (version 3.3.1, Thermo Fisher Scientific) with search algorithm from the Bioworks software package (version 3.3.1, Thermo Fisher Scientific). Search parameters were the following: a mass tolerance for parental ions was 2.0 Da or the MS/MS was 1.0 Da, one missed cleavage per peptide was allowed, and protein modifications were not considered. Matched peptide sequences were deduced based on the b and y ions and filtered using the cross-correlation score ( corr) and the normalized difference in cross-correlation scores (ΔCn) based on following thresholds: (1) the unique ΔCn > 0.1, (2) minimum corr of 1.9, 2.2, and 3.75 for charge states +1, +2, and +3, respectively. Proteins were only considered to be identified if they had at least two matched peptides.

Coimmunoprecipitation.
To examine whether Torsin and FMRP were present in the same protein complexes, rabbit anti-DTor antibodies, rabbit anti-HTor1A antibodies [5,19], mouse anti-DFMRP antibodies, and mouse anti-FMR1 antibodies were used for coimmunoprecipitation. Protein extracts from adult heads or larval body wall muscle preparations with brains were precleared with normal rabbit or mouse IgG agarose beads (Santa Cruz Biotech., Santa Cruz, CA, USA) and protein A/G plus agarose beads (Santa Cruz Biotech.). After mixing with 20 L mouse anti-DFMRP antibodies at 4 ∘ C overnight, protein A/G agarose beads preblocked with 1% bovine serum albumin (Amresco Inc., Solon, OH, USA) in RIPA buffer were added. We also performed control experiments without primary antibodies. After incubation at 4 ∘ C for 2 hr with shaking, agarose beads were precipitated by centrifuging them at 15,000 rpm for 10 minutes. After the beads were washed 3 times with RIPA buffer, sample buffers were added, and the samples were boiled before being loaded on a 12% SDS-PAGE gel. Proteins were transferred onto a nitrocellulose membrane and then probed with appropriate antibodies.
Two independent human plasma samples purchased from Innovative Research (Novi, MI, USA) were used for HTor1A and FMR1 coimmunoprecipitation experiments. Ten microliter of rabbit anti-HTor1A antibodies or mouse anti-FMR1 antibodies was added to 50 L of precleaned human plasma as described above. After mixing with protein A/G agarose beads for 2 hr at 4 ∘ C, beads were precipitated by centrifuging at 13,000 ×g for 10 min. We also performed control experiments without primary antibodies. After being washed 3 times with RIPA buffer, beads were boiled with SDS sample buffer and loaded on 12% SDS-PAGE gels. Proteins were transferred onto a nitrocellulose membrane and then probed with antibodies. The heavy chains of mouse IgG or rabbit IgG-specific HRP conjugated secondary antibodies (Thermo Fisher Scientific) were used to avoid detection of the IgG light chains.

Two Different Isoforms of Drosophila Torsin Proteins in
Drosophila. The expected molecular weight of Drosophila Torsin (DTor) proteins based on the deduced amino acid sequence is 38.17 kDa. The size of endogenous DTor protein detected by western blot analysis was approximately 32∼35 kDa (Figure 1(a)). To investigate the in vivo function of DTor, two independent UAS-DTor-cDNA lines were generated and crossed with Tub-Gal4 driver flies to produce flies that ubiquitously expressed DTor proteins. The two UAS-DTor-cDNA lines had significantly increased DTor protein levels compared to those of control genetic background flies (Tub-Gal4/+; Figures 1(a) and 1(b)). In addition, expression of two independent DTor-specific double-stranded inhibitory RNA (DTor-RNAi) constructs driven by Tub-Gal4 promoters significantly reduced expression of endogenous DTor to less than 30% of DTor proteins that were detected in the controls. Combined expression of UAS-DTor-cDNA and DTor-RNAi constructs caused DTor expression to be 1.5 times higher than the expression seen in flies of control genetic background.
The cDNA of DTor is known to consist of two exons (http://www.flybase.org/). There are several methionine (Met) residues at the beginning of DTor suggesting that the Met at 2nd, 6th, or 125th positions may be used to initiate translation of DTor protein, which may contribute to the several different sizes of DTor bands observed. Another possible cause of several sizes of DTor proteins is glycosylation. Thus, we examined the possible glycosylation of DTor using N-, O-, and C-glycosylation prediction servers [20,21]. Even though two potential N-glycosylation sites (Asparagine (Asn) positions at 96 and 311) were present in DTor, only the Asn at 96 was predicted to be N-glycosylated (Supplementary Figure 1). To confirm whether some DTor proteins in Drosophila were N-glycosylated, protein extracts were treated with PNGase F before performing anti-DTor western blot analysis. Band 1 appeared in nontreated samples, and Band 3 was clearly present in treated samples ( Figure 2). However, band 2 was unchanged by treatment. These results suggested that DTor may be modified by N-linked glycosylation and other posttranslational modifications of DTor proteins, which are not yet known.

Reduced Expression of DTor Protein Was Involved in Locomotor
Control. Next, we tested whether altered expression of DTor in Drosophila might cause any locomotor defect. When DTor-RNAi constructs were expressed ubiquitously using Tub-Gal4 drivers, flies showed early death compared with control flies or DTor-cDNA expressing flies. From 9 days of age, significantly fewer flies (6.50 ± 0.529) survived compared with genetic control flies (9.417 ± 0.229) or DTor-cDNA expressing flies (9.417 ± 0.229; Figure 4(a)). The loss of motor control in DTor-RNAi expressing flies manifested at 3 days of age. Compared with control or DTor-cDNA expressing flies, significantly fewer flies maintained their motor control (Figure 4(a)). Those motor-defects grew more severe as the flies aged. Bay day 12, 2 or fewer flies maintained their motor control even though at least 8 control or DTor-cDNA expressing flies still retained their motor control (Figure 4(a)). We further characterized whether loss of motor control in DTor-RNAi expressing flies was caused by reduced expression of DTor proteins in the nervous system or in the muscles (Figures 4(b) and 4(c)). When DTor-RNAi was expressed in the nervous system, the significant differences in the numbers of live flies were not observed until Day 12  Day 12 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * UAS-DTor-cDNA Gal4 pWiz-DTor-RNAi  (Figure 4(c)). These behavioral analysis data suggested that the early death of flies required the loss of DTor protein from both the neurons and muscles, but reduction of muscular expression of DTor proteins may contribute to the loss of motor control. Torsin, we treated flies with two oxidative stressors. When flies were exposed to H 2 O 2 , the hazardous ratio (HR) of flies ubiquitously expressing DTor-RNAi ( Figure 5(a)) was 1.924 (95% confidence interval (CI), 1.449∼2.550) compared with control flies. However, the HRs of other flies were not significantly different from those of control flies.

Resistance to Chemical Stressors Was Changed in Flies with
Flies expressing DTor-cDNA or HTor1A-cDNA showed increased resistance to paraquat. The HRs of flies expressing DTor-or HTor1A-cDNA were 0.4866 (95% CI, 0.3655∼0.6480) and 0.3328 (95% CI, 0.2456∼0.4511), respectively, compared with control flies (Figure 5(b)). However, flies expressing DTor-RNAi were not significantly different (HR, 0.9956; 95% CI, 0.7564∼1.3340) from control flies. These chemical stressor experiments suggested that an increased total amount of Torsin proteins might contribute to increase resistance to paraquat, but not H 2 O 2 .

Altered Expression of ADH in DTor Mutant Flies.
To gain further insight into the role of DTor in fly neuronal systems, unbiased 2D proteomic approaches were employed. A total of 6 spots from DTor-RNAi gels had significantly different volumes compared with those from control gels (Figures 6(a)-6(c)). The list of proteins identified from dysregulated spots was summarized in Table 1 and categorized using the PAN-THER Classification System (http://www.pantherdb.org/) [22] according to biological functions and pathways (Figures 7(a)-7(b)). Of the 15 dysregulated proteins, 10 proteins were allocated into metabolic processes. Two proteins were allocated into biological regulation, cellular processes, or localization. One protein was related to response to stimuli. Signaling pathways that those dysregulated proteins were categorized into were de novo purine biosynthesis, de novo pyrimidine deoxyribonucleotide biosynthesis, de novo pyrimidine ribonucleotide biosynthesis, glutamine glutamate conversion, and leucine biosynthesis. In addition, the protein-protein interactions among 15 dysregulated proteins were examined using the Search Tool for the Retrieval of Interacting Genes (STRING; http://www.string-db.org/) database [16]. Slo, ADH, CG2767, Fkbp13, Nurf-38, ATPsynd, Ef1 , and HSC70-4 showed strong association among those dysregulated proteins in DTor mutant brains. Because ADH was detected from several spots and is also known to function in metabolism, we further investigated by performing anti-ADH western blot analysis. Consistent with the 2D proteomic results, the normalized amount of ADH in DTor-RNAi expressing brains was increased 3.1-fold (3.11 ± 0.38, < 0.05; Figures 6(d) and 6(e)). These data suggested that reduced DTor significantly altered metabolism in adult Drosophila brains.

DTor Proteins Formed Protein Complexes with Fragile-X-Mental Retardation
Proteins. Because the expression of ADH was reduced in dfmrp mutants [23], we further examined whether expression of DFMRP in DTor-cDNA or DTor-RNAi expressing brains was altered. There was 5.1-fold or 2.7fold increase in the normalized amount of DFMRP in DTor-cDNA or DTor-RNAi expressing brains, respectively, compared with control brains (Figures 8(a) and 8(b)). Because this result raised the possibility of an interaction between DTor and DFMRP, we performed coimmunoprecipitation (Co-IP) experiments to examine whether DTor and DFMRP existed in the same protein complex. When mouse anti-DFMRP antibodies were used for IP, a strong DTor band was detected from DTor-cDNA expressing brains. A weak but definite band was detected from the control brains. However, a DTor band was not detected from DTor-RNAi expressing  brains (Figure 8(c)). In addition, without primary antibodies, DTor bands were not detected (Supplementary Figure 2(A)).
We further tested the interaction between HTor1A with FMR1 by performing Co-IP experiments using human plasma. Strong HTor1A and FMR1 bands were detected from two independent human plasma samples. When anti-HTor1A antibodies were used for IP, strong FMR1 bands were detected. In addition, when mouse anti-FMR1 antibodies were used for IP, strong anti-HTor1A bands were detected. Without primary antibodies, no HTor1A or FMR1 bands were detected (Supplementary Figure 2(B)). These results suggested that coexistence of Torsin proteins and FMR1 proteins in the same protein complex might be an evolutionarily conserved phenomenon.

The NMJ Phenotype of DTor-RNAi Expressing dfmrp
Mutants Was Similar to That of dfmrp Mutants. The presence of DTor and DFMRP in the same protein complexes raised the possibility that DTor and DFMRP might act via the same signal transduction pathway regulating synaptic plasticity. First, we examined whether DTor and DFMRP localized to the same cellular compartments by performing doublelabeling immunohistochemistry using anti-DTor and anti-DFMRP antibodies (Figures 9(a)-9(f)). Most of the DTor and DFMRP immunoreactivity was colocalized throughout muscles (Figures 9(a)-9(c)) with strong signals outside of nuclear membranes (Figures 9(a)-9(d)) or muscle fibers (Figure 9(e)). Even though strong DTor signals (red) were detected from synaptic boutons, DFMRP signals (green) did   not colocalize (Figure 9(f)). Double-labeling with mouse anti-Synapsin antibodies revealed that only a small portion of DTor was colocalized with Synapsin at the presynaptic terminals. The colocalization of DTor and DFMRP in muscle nuclei and fibers suggested that they might regulate synaptic plasticity by controlling signaling from NMJs to nuclei. Therefore, we examined synaptic bouton numbers in DTor-RNAi expressing dfmrp mutant NMJs (pWiz-DTor-RNAi/+; fmr1 Δ50m , Tub-Gal4/fmr1 Δ50m ) and compared them with dfmrp mutant (fmr1 Δ50m , Tub-Gal4/fmr1 Δ50m ) or control genetic background (fmr1 Δ50m , Tub-Gal4/+) NMJs (Figures 9(h)-9(j)). We found that expression of DTor-RNAi in dfmrp mutant did not produce a synergistic phenotype. The numbers of synaptic boutons in double mutant NMJs (129.7± 4.3) were similar to those of dfmrp mutant NMJs (129.6 ± 3.6, value > 0.5; Figure 9(k)). Taken together, DTor and DFMRP might be present in the same signaling pathway regulating NMJ development and extensions.

Discussion
The major goal of this investigation was to elucidate the in vivo function of Torsin family proteins in the regulation of locomotor ability and structural synaptic plasticity. The most significant finding in this study was that the downregulation of DTor expression was sufficient to increase the fly's susceptibility to environmental stress, which evoked locomotor disabilities. However, our finding that reduced expression of DTor in all tissues or specifically in muscles, but not specifically in neurons, was sufficient to cause increased susceptibility to environmental stress ( Figure 4) was contrary to our expectation, because it has been shown that the expression of mutant HTor1A in Drosophila increased the fly's susceptibility [5,13,19]. DTor proteins are abundantly expressed in larval and adult brains [24], and the DTor null mutant was semilethal in flies, with the few surviving homozygous adults being sterile and exhibiting locomotor defects [25]. How can we solve discrepancy between our data and previous results? One possible interpretation could be that fly neurons that regulate motor controls might have resistance to the reduced amount of DTor. This interpretation might also explain why only 30∼40% of ΔE HTor1A heterozygote mutant carriers in humans manifested the severe behavioral symptoms. Similar to fly neurons showing resistance to lower level of DTor, neurons in human brains involved with DYT1 dystonia might be resistant to a reduced level of normal HTor1A. Thus, 50% of normal HTor1A could be sufficient to maintain proper motor control in humans unless another mutation that could trigger manifestation of DYT1 dystonia is present.
Another important finding in this study is the biochemical and genetic interaction between DTor and DFMRP in fly brains and NMJs (Figures 8 and 9). Interestingly, DTor and DFMRP were present in the same protein complexes, and their colocalization in various cellular organelles such as muscle nuclear membrane and muscle fibers is very striking. In addition, the phenotype of dfmrp null mutants expressing DTor-RNAi was similar to those of dfmrp null mutant or DTor-RNAi NMJs, suggesting that DTor and DFMRP act through the same signal transduction pathway regulating synaptic plasticity. Even though we confirmed the biochemical interaction between HTor1A and FMR1 in human plasma in this study (Figure 8), the possible genetic interaction between these two proteins in humans could be supported by the fact that the age of onset for DTY1 dystonia is in adolescence. FMRP family proteins are known to play pivotal roles in nucleocytoplasmic trafficking of mRNA, dendritic localization of mRNA, and negative translational regulation [26]. In contrast to Fragile-X mental retardation, which is mainly caused by reduced expression of FMRP in human brains [26], the amount of FMR1 in DYT1 dystonia patients might be increased and results in reduced translation of mRNAs that might be important for proper development and maintenance of neural circuits and synaptic architecture. Thus, it will be intriguing to test this hypothesis in mouse transgenic models or human patients in the near future.
Unknown signal transduction pathways or components are often revealed by unbiased proteomic profiling experiments [27]. In this study, we found that the amount of ADH was altered in DTor-RNAi expressing fly brains compared with those of control brains ( Figure 6). ADH is an evolutionarily conserved enzyme responsible for breaking down toxic alcohols to generate various metabolites such as aldehydes and ketones with the reduction of nicotinamide adenine dinucleotide (NAD + ) to NADH in humans and animals [28]. Thus, an increased amount of ADH in fly brains might increase the amount of aldehydes or ketones. Recent studies in human neurological disorders suggested that many neurological disorders accompany metabolic changes in brains. For example, significant metabolic changes were reported from human brains with Alzheimer's disease and other related neurodegenerative disorders [29], Parkinson's disease [30], motor neuron disorders [31], major depressive disorder [32], and so forth. Interestingly, increased glucose metabolism was reported from the putamen, anterior cingulate, and cerebellar hemispheres in DYT1 dystonia patients and carriers [33]. In addition, DTor null mutant flies had significantly reduced brain dopamine levels in and showed strong genetic interaction with GTP cyclohydrolase [25]. Even though we have not investigated whether other enzymes present in ADHrelated metabolic pathway such as ALDH were increased in DTor mutant fly brains, recent studies of ALDH showed that it regulates a GABA synthesis pathway in human midbrain dopaminergic neurons that play key roles in locomotor control [34]. The key characteristic of DYT1 dystonia is excessive cocontraction of agonist and antagonist muscles during movements and impairment of inhibitory functions in several parts of patients brains has been reported [35]. Taken together, altered metabolism in DTor-RNAi-expressing fly brains could cause defects in inhibitory mechanisms in neuronal circuits that are responsible for motor controls.

Conclusion
In this study, we provide genetic, biochemical, molecular, and cellular evidence that DTor and FMRP are present in the same protein complexes and may act via the same signaling pathway regulating synaptic plasticity. In addition, by performing unbiased proteomic profiling and biochemical analyses, we found that the amount of ADH in DTor-RNAi-expressing fly brains was altered. Further studies using various Drosophila models for DYT1 dystonia will reveal unknown pathophysiological mechanisms underlying DTY1 dystonia that will provide new insight into this enigmatic disorder.