Increased Apoptosis of Myoblasts in Drosophila Model for the Walker-Warburg Syndrome

Walker-Warburg syndrome, a progressive muscular dystrophy, is a severe disease with various kinds of symptoms such as muscle weakness and occasional seizures. The genes of protein O-mannosyltransferases 1 and 2 (POMT1 and POMT2), fukutin, and fukutin-related protein are responsible for this syndrome. In our previous study, we cloned Drosophila orthologs of human POMT1 and POMT2 and identified their activity. However, the mechanism of onset of this syndrome is not well understood. Furthermore, little is known about the behavioral properties of the Drosophila POMT1 and POMT2 mutants, which are called rotated abdomen (rt) and twisted (tw), respectively. First, we performed various kinds of behavioral tests and described in detail the muscle structures by using these mutants. The mutant flies exhibited abnormalities in heavy exercises such as climbing or flight but not in light movements such as locomotion. Defective motor function in mutants appeared immediately after eclosion and was exaggerated with aging. Along with motor function, muscle ultrastructure in the tw mutant was altered, as seen in human patients. We demonstrated that expression of RNA interference (RNAi) for the rt gene and the tw mutant was almost completely lethal and semi-lethal, respectively. Flies expressing RNAi had reduced lifespans. These findings clearly demonstrate that Drosophila POMT mutants are models for human muscular dystrophy. We then observed a high density of myoblasts with an enhanced degree of apoptosis in the tw mutant, which completely lost enzymatic activity. In this paper, we propose a novel mechanism for the development of muscular dystrophy: POMT mutation causes high myoblast density and position derangement, which result in apoptosis, muscle disorganization, and muscle cell defects.

Walker-Warburg Syndrome (WWS), the most severe CMD, is a rare recessive inherited disorder characterized by muscular dystrophy, severe brain malformations, and eye abnormalities [5][6][7][8][9]. Patients with WWS rarely survive to birth, and even if they do, the chances that they will survive to adulthood are low [10].
Recently, several mutations were found in the POMT1 and POMT2 genes of WWS patients [11][12][13][14][15]. These mutations cause a reduction in O-mannosylation of a-dystroglycan, which results in WWS. In fact, recombinant mutant forms of POMT1 coexpressed with wild-type POMT2 did not show any O-mannosyltransferase activity [22]. Although a mouse model for WWS has been generated by targeted disruption of the Pomt1 gene, the mouse ortholog of POMT1, the adult phenotype is unknown because Pomt1 knockout mice are embryonic lethal [23].
In Drosophila, few studies on muscular dystrophy have been reported [24][25][26]. The Drosophila genome also has the components of DGC [27,28]. The Drosophila orthologs of human POMT1 and POMT2 are called rotated abdomen (rt) and twisted (tw), respectively [29,30]. In our previous study, we cloned genes of these orthologs and identified their activity. Their enzymatic activities are similar to those of the human enzymes; when both RT and TW were coexpressed in cultured insect cells, O-mannosyltransferase activity was observed [29]. Moreover, defective muscles and/or thin muscles with large sarcomeres were observed in larvae of rt or tw mutants [25,31], and the synaptic transmission in larvae was also abnormal with changes in the subunit composition of the postsynaptic glutamate receptors at the neuromuscular junction [26].
In the current paper, we first analyze the behavioral properties and ultrastructure of adult muscles in rt and/or tw mutants and then provide evidence that these mutants are highly useful for elucidating the mechanism of muscular dystrophy. Finally, we report that the number of apoptotic myoblasts increases in tw mutants and propose a new mechanism for the development of muscular dystrophy, which involves an increase in the number of apoptotic myoblasts, thereby causing muscle disorganization.

Behavioral defects in rt and tw mutants
Patients with progressive muscular dystrophy show muscle weakness and motor dysfunction with age. Therefore, we evaluated the motor function in rt and tw mutant flies. We first examined age-related changes in climbing activities. In rt mutants, which showed specific reduction of rt transcripts (Fig. S1), the climbing abilities of flies homozygous for rt were significantly decreased compared to those of flies heterozygous for rt at all ages (Fig. 1A, Table 1). In tw mutants, the climbing abilities of flies homozygous for tw were also significantly decreased compared to those of flies heterozygous for tw at all ages, except at the age of 41 days between Df(1)/+ and Df(1)/tw (Fig. 1B, Table 2). These data showed reduced climbing abilities in rt and tw mutants at almost all ages.
To understand the response of these mutants to milder exercises, we also evaluated the general locomotive activities of flies. The number of times a fly crossed the center of the glass tube was in the range of approximately 500-1000 in the rt group at all ages and was almost equal (Fig. 1C). Likewise, in the tw group, the number of times a fly crossed the center was in the range of approximately 1000-1800 at all ages, and we could not find any differences in the numbers between the tw mutant and the wildtype fly (Fig. 1D). These data showed that locomotive activities in rt and tw mutants did not decrease compared to those in wild-type flies.
Climbing ability and locomotive activities reflect leg muscle function. In order to determine the status of another behavioral function involved in other muscles, we evaluated flying ability, which reflects the function of flight muscles in the thorax as well as that of the relevant nervous system. The flying index of rt EP /rt KG flies was significantly lower than that of rt EP /+ and rt KG /+ flies (p,0.001, Tukey test) (Fig. 1E). In the case of tw mutants, the flying index of tw/tw flies was approximately half that of tw/+ flies, and the flying index of tw/Df(1) flies was approximately threequarters that of Df(1)/+ flies. There were significant differences between tw/tw and tw/+, and between tw/Df(1) and tw/+ or Df(1)/+ (each p,0.001, Tukey test) (Fig. 1F). The above results showed that flying ability was reduced in rt and tw mutants.
Here, we showed 3 behavioral features of rt and tw mutants: (1) mutant flies have impaired motor function; (2) defective motor function is observed from just after eclosion to death; and (3) mutant flies show abnormalities in heavy exercises such as climbing or flying, but not in light movements such as locomotion. These results were compatible with the idea that these mutant flies had abnormal muscles and/or motor neurons.

Behavioral defects of flies expressing RNAi for the rt gene
We examined in which tissue the expression of the rt and tw genes affects motor function. We used tissue-specific knockdown flies produced by the Gal4-UAS-IR technique [32], the tissuespecific gene knockdown technique that uses the Gal4-UAS system and RNA interference (RNAi) methods. Tissue-specific rt gene knockdown was induced when flies had both Gal4 driver and UASrt-IR, whereas gene knockdown was not induced when flies had only Gal4 driver or UAS-rt-IR (Fig. S2). The climbing ability of flies with ubiquitous expression of RNAi for the rt gene as driven by Act5C-Gal4 was significantly lower than that of control flies at all ages ( Fig. 2A, Table 3). We also examined the climbing ability of neuron-and glial cell-specific knockdown flies by using elav-Gal4 and repo-Gal4, respectively, since it has been reported that there are some defects in the efficacy of synaptic transmission and changes in the subunit composition of postsynaptic glutamate receptors at the larval neuromuscular junction of rt mutants [24]. Neither   neuron-nor glia-specific knockdown of the rt gene resulted in a distinct reduction in climbing ability at any age, except at the age of 1 day after eclosion in neuron-specific knockdown flies (Figs. 2B and C, Table 3). It was noted that the presence of elav-Gal4 or repo-Gal4 has a deteriorative effect on climbing ability for unknown reasons. These data indicated that the climbing ability of adult flies is not mainly influenced by knockdown of the rt gene in neurons or glial cells, and that the expression of this gene in muscles should be relevant to climbing ability.
Age-related abnormal patterning and ultrastructural abnormalities in muscles of rt and tw mutants The behavioral data strongly suggest that Drosophila POMT mutants exhibit the muscle defect. Muscles of WWS patients also show structural abnormalities. Hence, we examined the effect of rt or tw on the patterning of muscles by checking the larval body wall muscles in mutants and in flies expressing RNAi for the rt gene. The normal structure of wild-type muscles is shown in Fig. 3A. Deficient or thin muscles were observed in the abdominal segment of rt mutants (Figs. 3B and C) and in flies expressing RNAi for the tw gene (Fig. 3D). The frequency of abnormal patterning of muscles detected by FITCphalloidin staining in these mutants was approximately 10% (data not shown). In addition, we examined the frequency of abnormal patterning in larval muscles of tw mutants by using the MHC-tauGFP marker to observe live muscles. In MHC-tauGFP larvae, which did not show reduction of rt and tw transcripts (Fig. S3), the dorsal body wall muscles were structurally normal (Fig. 3E). A few muscles were absent in the mutant larvae (Figs. 3F and G), and many muscles were absent in rare cases of mutant larvae (Fig. 3H). Significantly higher frequencies of abnormal patterning were detected in the muscles of tw mutants (probabilities for all compared pairs were p,0.001, Fisher's exact test) (Fig. 3I).
We further examined the effect of the tw mutation on sarcomeric structure by performing detailed electron microscopic analysis on the leg and thoracic muscles from wild-type, tw mutant and the rescue flies. Similar changes in ultrastructure were observed in the leg and thoracic muscles of both male and female tw mutant flies ( Table 4). The normal sarcomeric structure of the muscles of wild-type flies is shown in Figs. 4A, 4C, and 5A. Sarcomeric disarray was frequently observed in the muscles of tw mutant flies (Figs. 4B, 4D, 5B, and 5C). In the mutant muscles, Zlines were irregular and often streaming (Figs. 4D, 5B, and 5C), nemaline bodies were observed in the muscle fiber (Fig. 4D), actin and myosin filaments were disorganized (Figs. 4E and 5B), and accumulated glycogen granules were seen (Fig. 4F). Enlarged mitochondria (Fig. 5F) and swollen sarcoplasmic reticulum (SR) were seen in the tw mutant muscles (Figs. 4H and 5D), while normal mitochondria and SR were observed between the muscle fibers (Figs. 4A, 4C, 4G, 5A, and 5E). The basement membrane was duplicated and multilayered in tw mutant muscles (Figs. 4J and 5H), while normal basement membrane was observed continuously along the sarcolemma in wild-type muscles (Figs. 4I and 5G). The abovementioned defective muscle phenotypes were observed both in 15-and 35-day-old tw mutants but were hardly detected in wild-type flies. The number of mutants with defective phenotypes was higher in 35-day-old tw mutants than in 15-dayold tw mutants. Moreover, these defective phenotypes could not be found among the rescue flies (Table 4), indicating that the defective phenotypes in the mutants were fully rescued. These results demonstrated that tw contributed to the maintenance of muscle ultrastructure. Next, we counted the number of sarcomeric disarray, irregular Z-line, and filament disorganization occurrences in a 590-mm 2 muscle area per individual and calculated the percentage of these abnormal structures (number of abnormal structures/number of observed). In thoracic muscles, these abnormal structures were observed in 35-day-old mutants but could not be detected in wild-type flies (Fig. 6A). In leg muscles, these abnormal structures were observed in both 15-and 35-dayold mutants but were hardly detected in wild-type flies (Fig. 6B). The percentages of abnormal structures in the tw mutant were significantly higher than those in wild-type flies except for filament disorganization in the leg muscles of 15-day-old flies, when abnormal structures were observed (Fig. 6). Moreover, these abnormal structures were significantly exaggerated with age in tw mutants but not significantly exaggerated with age in wild-type flies except for sarcomeric disarray in the leg muscles of 15-day-old flies ( Table 5). These changes, which are reminiscent of the progressive symptoms in WWS patients, were more frequently observed in 35-day aged mutant muscles than in 15-day aged mutant muscles. The abovementioned results clearly demonstrated several kinds of abnormalities in the muscles of tw mutants that become more severe with age.
Excessive apoptosis of myoblasts in the wing disc of the tw mutant Changing of the density of myoblasts is a result of an alteration in cell death or cell division. Therefore, we checked the number of apoptotic and dividing myoblasts. The nuclei of myoblasts were visualized by GFP (Figs. 8A and B), and we observed the apoptotic myoblasts by using the cleaved caspase-3 antibody (Figs. 8C and D). Both wild-type flies and tw mutants had apoptotic myoblasts (Figs. 8E and F); however, the number of apoptotic myoblasts in the wing discs of tw mutants was 2.4-fold higher than that in the wing discs of wild-type flies (Fig. 8G). We also observed dividing myoblasts by using phospho-histone H3 antibody. The number of dividing myoblasts in the wing disc did not differ between wild-type flies and tw mutants (Fig. S4). These results showed that apoptosis was enhanced in myoblasts of the tw mutant while the number of dividing cells was not altered. Excessive apoptosis of myoblasts during muscle differentiation should lead to muscle disorganization, including muscle cell defects.
Increased a-spectrin in myoblasts and bPS-integrin around myoblasts of the tw mutant Cytoskeletal and cell adhesion molecules, such as spectrins, cadherins, and integrins, proteolyze during apoptosis [33][34][35]. Thus, we examined the expression of a-spectrin, DE-cadherin, and bPS-integrin in myoblasts of the tw mutant. Surprisingly, the signals of a-spectrin and bPS-integrin excessively increased in the region of myoblasts compared to the lateral region of myoblasts in the tw mutant (Figs. 9E and M) although apoptosis increased in the myoblasts of tw mutant. These signals in wildtype flies did not change between the 2 regions (Figs. 9A and I). We could not find any difference in the DE-cadherin signal (Figs. 9B, F, J, and N). These data suggest that a-spectrin and bPS-integrin might be induced to protect myoblasts of the tw mutant from apoptosis.

Lethality in tw mutants and flies expressing RNAi for the rt gene
It is known that WWS has high lethality rates during early development. Therefore, we determined whether Drosophila POMT genes play an important role in viability. We crossed females heterozygous for tw with male hemizygous for tw and checked the number of F 1 progenies (Fig. 10A). If all the eggs of F 1 progenies hatched and developed normally, the ratio of the number of individuals having a twisted abdomen, the tw phenotype, to those having a normal abdomen is expected to be 1. The ratio was found to be 0.2, significantly lower than the expected ratio (p,0.001, chi-square test) ( Table 6). In addition, in order to determine the necessity of the tw gene for viability, we performed a rescue experiment by expressing the tw gene in tw mutants (Fig. 10B). The ratio was found to be 1.25 and was significantly increased (p,0.001, chi-square test) ( Table 7). These data indicated that the tw gene plays an important role in viability and normal development. We also examined fly viability after knockdown of the rt gene under 3 different temperature conditions. Knockdown at high temperature is more efficient than that at low temperature because yeast transcriptional factor GAL4 binds strongly to UAS sequences. Ratios of the number of knockdown flies to the number of non-knockdown flies were 0.26, 0.12, and 0.00 at 18, 25, and 28uC, respectively (Fig. 10C, Table 8). Growing flies at a higher temperature resulted in higher lethality, indicating that the rt gene also contributed to normal development. The abovementioned results showed that the tw and rt genes were essential for the viability of the embryo, larva, and/or pupa. Elucidation of the reason for the lethal phenotype in these mutants could clarify the mechanism of low viability in human patients.  Table 3. The climbing ability of flies expressing RNAi for the rt gene driven by Act5C-Gal4 was significantly less than that of control flies at all ages. Neither neuron-nor glial cell-specific knockdown of the rt gene resulted in a distinct reduction in climbing ability at any age except at 1 day after eclosion in neuron-specific knockdown flies. doi:10.1371/journal.pone.0011557.g002

Shortened lifespan of flies expressing RNAi for the rt gene
Patients with WWS rarely survive to adulthood [10]. Therefore, we investigated whether rt knockdown mutants had shortened lifespans. Flies with ubiquitous expression of RNAi for the rt gene driven by Act5C-Gal4 had shorter lifespans than those of the control groups (both p,0.001, log-rank test, Table 9) (Fig. 11A). The median lifespan of Act5C-Gal4/rt-IR flies was 23 days, which was 251.1% of that of UAS-rt-IR/+ flies (47 days) and 266.7% of that of Act5C-Gal4/+ flies (69 days) (Tables 9 and 10). On the other hand, the lifespans of flies expressing RNAi for the rt gene in neurons and glial cells driven by elav-Gal4 and repo-Gal4, respectively, were not affected (Figs. 11B and C, Tables 9 and 10). These results indicated that knockdown of the rt gene in all tissues reduced the lifespan, while knockdown in neurons or glial cells did not influence lifespan. Together with the results of the muscle phenotype in Drosophila POMT mutants, these results suggest that age-related weakness of the muscles in the heart and/ or gastrointestinal tract may lead to early death in flies.
No enzymatic activity in the protein of the tw mutant form or in the larval extracts of rt and tw mutants Since there is no evidence that POMT2 mutation in WWS patients influences POMT activity, we examined the POMT activity in the Drosophila POMT2 mutant. We prepared recombinant wild-type RT (RT WT ), wild-type TW (TW WT ), and mutant TW (TW Mut ) in order to examine POMT activity in the tw mutant. After pVL1393-rt WT -HA and pVL1393-tw Mut or pVL1393-tw WT was co-transfected into Sf21 insect cells, microsomal membrane fractions were collected from each infected cell. The specific expression of recombinant proteins was confirmed by western blot analysis using anti-HA monoclonal antibody and anti-dPOMT2 antibody (Figs. 12A and B). POMT activity toward GST-a-DG was measured using each microsomal membrane fraction as described under ''Materials and Methods.'' We could not detect enzymatic activity when only RT WT -HA, TW WT , or TW Mut was expressed. POMT-specific activity was detected when RT WT -HA and TW WT were co-expressed (p,0.001, Tukey test). However, no POMT-specific activity was detected when RT WT -HA and TW Mut were co-expressed (n.s., Tukey test) (Fig. 12C). These data clearly demonstrated that the mutation involved in the tw mutant was a null mutation for POMT activity. Moreover, we examined the POMT activity of larval extracts of rt and tw mutants. POMT-specific activities of the rt and tw mutants were significantly reduced compared to those of wild-type flies (Fig. 12D). Taken together, the tw mutant was a null mutant and the rt mutant we used was an almost null or a strong hypomorphic mutant for POMT activity.

Genetic interaction between rt or tw and Dg in the wing
Dg is one of the putative core proteins that are O-mannosylated by RT and TW. Wild-type flies had normal-shaped wings (Fig. 13A); however, knockdown of Dg in the posterior region of the wing disc resulted in a blistered phenotype (Fig. 13B). The blistered phenotype results from cell adhesion failure in the 2 cell layers during wing development. Actually, Dg expression was dramatically decreased in the posterior region ( Fig. S5), indicating that Dg contributed the attachment of the 2 cell layers during wing development. We examined the genetic interaction between rt or tw and Dg by using this phenotype. At 25uC, the penetrances of the blistered phenotype in wings of single knockdown flies of rt, tw, and Dg were 0, 0, and 0.06, respectively. The penetrances of the phenotype in double knockdown flies of rt-Dg and tw-Dg were 0.14 and 0.08, respectively, but they were not significantly higher than those of the single knockdown flies (Fisher's exact test). As mentioned in the above section, ''Lethality in tw mutants and flies expressing RNAi for the rt gene,'' knockdown at 28uC is more efficient than that at 25uC. Thus, we performed the knockdown at 28uC. At 28uC, the penetrances of the blistered phenotype in wings of single knockdown flies of rt, tw, and Dg were 0, 0, and 0.34, respectively. The penetrances of double knockdown flies of rt-Dg and tw-Dg were 0.86 and 0.48, and they were significantly higher than those of single knockdown flies (p,0.001 and p,0.05, Fisher's exact test) (Fig. 13C). These data showed that rt or tw genetically interact with Dg to contribute to cell adhesion in the wings. Together with the high density of myoblasts observed in the tw mutant (Fig. 7F), these results suggest that both epidermal cells and muscle progenitor cells of Drosophila POMT mutants give rise to cell adhesion derangement.

Days after eclosion
The climbing abilities of rt and tw mutant flies were reduced compared to those of flies that were heterozygous for rt and tw, respectively (Figs. 1A and B, Tables 1 and 2). Reduced climbing ability was also observed in flies with ubiquitous expression of RNAi for the rt gene driven by Act5C-Gal4 (Fig. 2, Table 3). In addition, Dg and Dys (Dystrophin) mutant flies lack climbing ability [24]. These data indicate that the O-mannosyl glycan on Dg contributes to motor functions such as climbing. The mutation of rt or tw in Drosophila causes the behavioral defect like WWS, since the defect of O-mannosyl glycan on Dg leads to WWS [14,22]. Furthermore, the climbing abilities of mutants rapidly decreased with age ( Figs. 1A and B). These changes are similar to the behaviors seen in patients with WWS, such as difficulty in walking with age.
Both our study (Fig. 3) and other studies [25,31] have revealed structural defects in the larval body wall muscles in Dg, rt, and/or tw mutants. Moreover, we observed various kinds of defective ultrastructural phenotypes in the thoracic and leg muscles of adult tw mutant flies (Figs. 4 and 5, and Table 4) that have been reported in muscle biopsies of WWS patients [36]. A recent study revealed that expression of GluRIIB, a subunit of the postsynaptic glutamate receptor, and the efficacy of synaptic transmission decreases at the neuromuscular junctions of larval Dg and rt mutants [26]. Muscle contraction and membrane resistance in larval body wall muscles changed in flies expressing RNAi for the Dg gene [25]. These changes might cause decreased motor function in adult rt and tw mutants. On the other hand, we observed no difference in locomotive activity between mutant and control flies (Figs. 1C and D). The mutant flies showed abnormalities in heavy exercises, such as climbing or flying, but not in light movements, such as locomotion, probably because some muscles and/or neuromuscular junctions had normal functioning while other muscles had defective functioning.
The growing of flies with ubiquitous expression RNAi for the rt gene at a higher temperature resulted in higher lethality rates, and knockdown at 28uC was almost entirely lethal (Table 8). Some rt alleles that were hemizygous for the deficiency and entirely lacked the genomic region of rt showed partial lethality [31]. Thus, a null mutation in rt appears to be lethal. On the other hand, the tw 1 allele is semi-lethal (Table. 6). Davis [37] also reported that flies carrying a tw mutation showed reduced viability. Furthermore, we rescued the lethality of tw mutant flies by ubiquitous expression of the tw gene ( Table 7). The apparent incidence rate of WWS is very low [10] because of its high lethality rate in the embryonic stage. Indeed, defective development in the early embryonic stages causes embryonic lethality in Pomt1 mutant mice [23]. However, the mechanism underlying this lethality is unknown. Thus, the mechanism of the high lethality rate in the embryonic stage in mammals may be understood by elucidating the mechanisms of the high lethality rates in rt and tw mutant flies. Lethality in rt mutant flies is not associated with any particular developmental stage [31]. Several studies have revealed that Dg is involved in epithelial and oocyte polarity determination [38][39][40], so the Omannosyl glycans on Dg at least contribute to the viability through fly oocyte formation and motor function.
Even though WWS patients may survive to birth, their lifespan is generally short and they typically die before reaching adulthood. Flies with ubiquitous expression of RNAi for the rt gene have shortened lifespans as well, but those with neuron-or glial cellspecific expression of RNAi do not have shortened lifespans (Fig. 11, Tables 9 and 10). These results suggest that expression of POMT1 and POMT2 in the tissues other than neurons and glial cells plays a crucial role in longevity in humans.

Anatomical similarities between the muscles of tw mutant flies and WWS patients
In the present study, we showed that the larval body wall muscles (Fig. 3A, muscle 5) were thin or missing in flies expressing RNAi for the rt and tw genes as in flies carrying other mutant alleles (Figs. 3B-D). Furthermore, we observed living larval body wall muscles in tw mutant flies by using the MHC-tauGFP reporter, a muscle marker (Figs. 3F-H). Studies have shown that larval body wall muscles of rt and tw mutant flies are sometimes thin and missing [25,31]. In our analysis, the frequency of abnormal patterning in the body wall muscles is approximately 50% in female and 40% in male tw mutant flies, both of which are higher than that reported in one study (10%) [25]. The discrepancies between the data seem to result from differences in the methods used: our methods did not involve dissection of larvae but, rather, observation of all muscles in living larvae using MHC-tauGFP reporter for muscle visualization; on the other hand, methods used    to obtain the previous data included dissection, fixation, FITCphalloidin staining for muscle visualization, and observation of particular muscles. The ultrastructure of muscles in adult mutant flies has not yet been reported, although age-dependent muscle degeneration and large sarcomeres in the larval body wall muscles of the flies in mutants have been reported based on light microscopic observations [24,25]. Here, we observed ultrastructural characteristics of tw mutant flies and discovered various phenotypes: sarcomeric disarray, irregular Z-lines, filament disorganization, swollen SR, accumulation of glycogen granules, enlargement of mitochondria, and duplication of basement membranes (Figs. 4 and 5). The ultrastructural defects observed in the thorax and leg muscles of tw mutant flies are consistent with muscle characteristics of patients with Duchenne muscular dystrophy, including WWS. Myofibril shearing, myofilament loss, wavy or disrupted Z-lines, changes in In thoracic muscles, these abnormal structures were observed in 35-day-old mutant flies but not in wild-type flies. In leg muscles, these abnormal structures were observed in both 15-and 35-day-old mutant flies but were hardly detected in wild-type flies. We compared data of each abnormal phenotype in 15-and 35-day-old wild-type, tw mutant, and rescued flies. Each bar represents the mean of 6 individuals. Error bars indicate standard error. **p,0.01, ***p,0.001 by Welch's two sample t test. n.s., not significant. doi:10.1371/journal.pone.0011557.g006 the size of SR and mitochondria, and clumped glycogen particles were observed in the muscles of patients in whom Duchenne muscular dystrophy was diagnosed clinically, histologically, and by serum creatine kinase assay [36]. In addition to the similarities in ultrastructural muscle defects seen between humans and flies, we demonstrated that the muscle phenotype increases in severity with age in flies ( Fig. 6 and Table 5). These facts suggest that both the muscle ultrastructure and the function of rt and tw in maintaining the muscle structure are conserved in flies and humans.

Similarities between POMT activities of WWS patients and tw mutant flies
POMT activity was clearly detected in Sf21 cells that coexpressed RT WT and TW WT but could not be detected in cells that expressed only RT WT or TW WT (Fig. 12). We obtained similar results previously [29]. In humans, coexistence of POMT1 and POMT2 is required for POMT activity [18], just as both RT and TW are required in flies. Mutations in the POMT1 gene considered to cause WWS lead to reduced POMT activity and a defect in protein O-mannosylation [22].
However, POMT activity in WWS patients carrying a POMT2 mutation has never been reported. We report here for the first time that a mutation in the tw gene, the Drosophila ortholog of POMT2, causes a reduction in O-mannosyltransferase activity (Fig. 12) and thus results in a defect in protein O-mannosylation. TW WT protein has 7 transmembrane helices, and the TW Mut protein (T59GS) contains a change in the first transmembrane region as predicted by SOSUI (http://bp.nuap.nagoya-u.ac.jp/ sosui/), a secondary structure prediction program. These results indicate that the first transmembrane region of TW may play an important role in O-mannosyltransferase activity. Other regions in the POMT2 gene important for POMT activity will be found through study of the mutations, enzymatic activity, and secondary structure of the protein in WWS patients, who have further POMT2 gene mutations.
New functional aspect of POMTs obtained from analyses of rt and tw mutant flies CMDs, including WWS, cause progressive muscle degeneration and necrosis [41]. These muscle changes result from a defect in the DGC, including the O-mannosyl glycan, which is synthesized by POMTs [22]. DGC plays an important role in binding between the cytoskeleton and basal membrane [42,43]. Because muscles are always subjected to severe physiological conditions, fragile binding between the cytoskeleton and basal membrane in patients with muscular dystrophies seems to result in weakness of the plasma membrane in muscles and degeneration of muscle fibers. Our results regarding the behavior and muscle structure of rt and tw mutant flies indicate that the rt and tw genes maintain muscle structure. Actually, the basal membrane was duplicated in muscles of the tw mutant (Figs. 4J and 5H). Further, O-mannosyl glycans on Dg contributed to attachment of the wing cells (Fig. 13). The above results suggest that cell adhesion in both muscle and epithelial cells, such as wing cells, which is involved in the Omannosyl glycans on Dg, contributes to tissue or organ organization in Drosophila POMT mutants.
Interestingly, we showed that apoptosis was increased in myoblasts of wing imaginal discs in tw mutant larvae, cells that otherwise differentiate into indirect flight muscles (Fig. 8G). Myoblast density was observed to be high (Fig. 7F), although no mitotic abnormality was observed (Fig. S4). Moreover, expression of a-spectrin and bPS-integrin in the myoblasts was increased (Fig. 9). These results suggest that myoblasts of the tw mutant might avoid enhanced apoptosis or compensate for a DCG defect by overexpressing cytoskeletal or cell adhesion molecules such as a-spectrin and bPS-integrin. Mutations in the tw gene will lead to a high density of myoblasts, cause disruption of myoblast intercellular interactions, and result in enhanced myoblast apoptosis. Thus, the tw gene also controls myoblast density or intercellular interaction.
Here, we propose a novel mechanism for the development of WWS. At first, POMT mutation causes cell adhesion abnormalities, myoblast position derangement, and a high density of myoblasts because O-mannosyl glycans do not form on core proteins, such as Dg. Since high densities of myoblasts do not develop into normal muscles, some of them are excluded by apoptosis. As a result, muscles in the POMT mutant have duplicated basal membranes and decreased motor function compared to those of wild-type flies. Apoptosis occurs not only in myogenesis but also in cell differentiation events such as neurogenesis. Thus, severe phenotypes will appear in several organs of patients with WWS.

Genetic interaction between Dg and rt and/or tw in the wing
Several studies have revealed the genetic interaction between Dg and rt and/or tw. One study demonstrated that Dg and rt contribute to the promotion of synaptic vesicle release and regulate glutamate receptor subunit composition at the neuromuscular junction [26]; moreover, it showed that muscle attachment formation and sarcomere size determination in third instar larvae require functional Dg and rt or tw [26].
In the present study, we showed that basement membranes are duplicated and multilayered in tw mutant muscles (Figs. 4J, 5H). This change in the basement membrane is due to the muscle attachment failure that results from mediation of Dg and laminin by O-mannosyl glycan [43]. In addition, knockdown of Dg in the wing resulted in the blistered phenotype (Fig. 13B), which was caused by cell adhesion failure of the 2 cell layers during wing development. Moreover, double knockdown of Dg and rt or tw enhanced this phenotype (Fig. 13C). Consequently, we demonstrated that the effect of O-mannosyl glycan on Dg contributes to epidermal cell attachment in the wing. Incidentally, a mutant of wb (wing blister), a Drosophila laminin, also showed blistered wings [44]. Thus, it is suggested that O-mannosyl glycan mediates the binding  of Dg with molecules such as laminin in the extracellular matrix, and that failure of this binding leads to the blistered wing phenotype observed in mutants. Laminin and adhesion molecules play important roles in muscle attachment [45][46][47]. Thus, the binding of Dg to laminin mediated by O-mannosyl glycan might contribute to adhesion between the 2 epithelial cell layers of the wing as well as muscle attachment.

Conclusion
Analyses of rt and tw mutants, the Drosophila models for WWS, help in understanding not only the symptoms of this human disease but also the mechanisms of muscular dystrophy. We proposed a new mechanism for the development of muscular dystrophy involving increased apoptosis in developing muscles that causes muscle disorganization. Further studies with these mutants will provide additional insight into the mechanisms of muscular dystrophies and will help in the development of useful drugs for WWS patients.

Fly stocks
All stocks were raised at room temperature (23-25uC) using a cornmeal-yeast-glucose medium. All stocks except those for rt EP571 , MHC-tauGFP, 1151-Gal4, and Canton-S were obtained from Bloomington Stock Center (http://flystocks.bio.indiana.edu/). We refer to the rt KG04772 line (stock number: 14434) as rt KG . rt EP571 was obtained from Szeged Stock Center. In this paper, we refer to this line as rt EP . Df(1)Exel6223, which disrupts the tw gene, is referred to as Df(1). The tw stock was backcrossed onto the Canton-S background for 20 generations. MHC-tauGFP [48], a muscle marker, and 1151-Gal4 [49], a driver for marking the myoblasts in wing imaginal discs, were supplied by Dr. T. Maqbool. Canton-S was a gift from Dr. D. Yamamoto. UAS-tw-IR and UAS-rt-IR have been referred to by Ichimiya et al. [29] as UAS-dPOMT2-IR and UAS-dPOMT1-IR, respectively. UAS-Dg-IR and UAS-tw were generated by the following methods.
Merged image of (A) and (B). (D) High-magnification image of the notum plane region in the wing disc. Myoblast nuclei are seen as white or gray spots. (E) The percentage of wing imaginal disc area occupied by myoblasts in wild-type and tw mutant (tw, 1151.GFPnls) larvae. (F) The number of myoblasts in a 2500 mm 2 area of the wing imaginal disc in wild-type and tw mutant larvae. (G) Total number of myoblasts in a single notum plane region in the wing imaginal disc of wild-type and tw mutant larvae. Error bars in E-G indicate standard error. The wing disc area occupied by myoblasts in tw mutant flies was less than that in the wild-type flies. The number of myoblasts in the constant area of the mutant fly was more than that of the wild-type fly. The total number of myoblasts in a single wing disc was equal in the tw mutant and wild-type flies. The density of myoblasts in tw mutant larvae was higher than that in wild-type larvae. **p,0.01 by t test. n.s., not significant. doi:10.1371/journal.pone.0011557.g007

UAS-Dg-IR flies
The UAS-IR fly line was obtained as described in previous reports [29,32]. The cDNA fragment of Dg (nucleotide positions 327-826 of the coding sequence of Dg-RA) was amplified by PCR using a cDNA library derived from Drosophila melanogaster.
The amplified fragment was inserted as an inverted repeat (IR) sequence into the pSC1 vector. The IR-containing fragment was then subcloned into the transformation vector pUAST, and this vector was introduced into Drosophila embryos of the w 1118 mutant stock, which was used as the host to construct the UAS-IR fly line according to the procedure reported by Spradling and Rubin [50].

UAS-tw flies
The DNA fragment containing the tw gene was amplified by two-step PCR. For the first PCR, we used the plasmid DNA from an EST clone (LP01681) as a template and the following primer set for the amplification of the tw coding region: forward primer 59-AAAAAGCAGGCTTGGCAGCAAGTGTTGTTA-39; and reverse primer 59-AGAAAGCTGGGTCTAGAACTCCCAGG-TAGAAAG-39. For the second PCR, we used the first PCR product as a template; a forward primer that included the attB1 site, 59-GGGGACAAGTTTGTACAAAAAAGCAGGCT-39; and a reverse primer that included the attB2 site, 59-GGGGACCACTTTGTACAAGAAAGCTGGGT-39. The amplified fragments were subcloned into the pDONR201 TM vector (Invitrogen, http://www.invitrogen.com). The inserted fragment was then recombined between the attR1 and attR2 sites in the multi-cloning site of the modified pUAST vector to yield pUASTtw. The transgenic fly was generated by the method described in the ''UAS-Dg-IR flies'' section.  Table 6. (B) The cross scheme for the rescue experiment of the tw mutant. Female Act5C-Gal4 driver flies with tw mutation were crossed with male UAS-tw flies with tw mutation. In the F 1 progeny, rescued and non-rescued individuals are born and are described as [Cy 2 , GFP 2 ] and [Cy + , GFP + ], respectively. The number of F 1 progeny with phenotypes [Cy 2 , GFP 2 ] or [Cy + , GFP + ] is shown in Table 7. (C) The cross scheme for examining the viability of flies expressing RNAi for the rt gene. Female Act5C-Gal4 driver flies were crossed with male UAS-rt-IR flies. In the F 1 progeny, flies expressing RNAi for the rt gene, Act5C-Gal4 driver, and UAS-rt-IR individuals were born. Flies expressing RNAi for the rt gene do not have CyO GFP balancer; thus, the phenotype is described as [Cy 2 , GFP 2 ]. Act5C-Gal4 driver and UAS-rt-IR flies have CyO GFP balancer; thus, the phenotype is described as [Cy + , GFP + ]. The number of F 1 progeny with ([Cy 2 , GFP 2 ]) or without ([Cy + , GFP + ]) ubiquitous expression of RNAi for the rt gene is shown in Table 8. doi:10.1371/journal.pone.0011557.g010 Table 7. Viability of rescued tw mutant flies.  Climbing assay We used a modified version of a previously described assay to assess climbing ability [51]. Twenty individual flies were gently introduced into a glass vial height 240 mm tall and 25 mm in diameter. After a 5-minute rest, the bottom of the vial was gently tapped and the maximum height reached in 10 seconds was recorded by a digital camera. Five trials were performed in each experiment. Climbing ability was then calculated from the 5 trials.

Locomotion assay
We used the procedure described in Kaneuchi et al. to assess locomotive ability [52]. One day after eclosion, flies were individually introduced into a glass tube 65 mm long and 3 mm in diameter with a medium containing 60% swelling SP-Sephadex C-50 (weight/volume), 10% glucose, 0.6% propionic acid, 2% yeast extract, and 1.2% agar. After the tubes were placed in the Drosophila Activity Monitor (TriKinetics Inc., http://www. trikinetics.com/), the medium was changed every week until the fly died. The number of times a fly crossed the center of the glass tube was automatically recorded every 30 minutes. At least 16 individuals were tested in each group.

Flight assay
We used a modified version of the method by Stockinger et al. to assess flying ability [53]. Twenty individual flies (30-35 days of age) were dumped through a plastic funnel into a glass cylinder 450 mm in height and 80 mm in diameter whose inside surface was coated with mineral oil. Numbers on the y-axis represented the height marks on the glass cylinder: 1, 0-200 ml; 2, 200-400 ml; 3, 400-600 ml…10, 1800-2000 ml. Scores were individually recorded, and the score average was calculated from 5 independent experiments.

Longevity assay
About 30 test flies were maintained in the standard medium after eclosion for assessment of longevity. Every 2 or 3 days, flies were transferred to a fresh food vial and the number of dead flies was determined until all flies died. More than 4 replicate vials were tested in each group and data were pooled for statistical analysis.
Staining of muscles and examination of the frequencies of abnormal muscle patterning Third instar larvae were dissected along the dorsal midline in ice-cold phosphate-buffered saline (PBS: 130 mM NaCl, 7 mM Na 2 HPO 4 , and 3 mM NaH 2 PO 4 ). After removal of the digestive organs, fat bodies, and main trachea, the preparations were fixed with 4% formaldehyde in PBS for 30 minutes at room temperature. The tissue was then washed in PBS with 0.1% Triton X-100 (PBT) and stained with phalloidin conjugated with FITC. Stained samples were mounted in 90% glycerol in PBS and observed under a Zeiss LSM5 Pascal confocal microscope. We counted the number of individuals that had more than 1 abnormal muscle and calculated the frequencies of abnormal muscle patterning.

Vector construction and expression of mutant form of TW
The full-length ORFs of the wild-type rt form, wild-type tw form, and mutant tw form were expressed in insect cells as described previously [29]. One mutant in tw, tw 1 , has been reported. Sequencing of the tw gene in tw 1 mutant flies revealed 3 alterations. Two of these alterations do not cause amino acid substitutions. The third alteration, a 2-base substitution and 3-base insertion, is predicted to affect the translated protein sequence: the 59th threonine residue from the initiating methionine changes to glycine and serine residues [30]. Hereafter, we represent the wildtype rt form, wild-type tw form, and mutant tw 1 form as rt WT , tw WT , and tw Mut , respectively. The coding region of tw Mut was amplified from the cDNA of tw mutant flies using the same primer sets as that for tw WT [29], and the amplified fragment was inserted into the vector pVL1393 g. pVL1393-rt WT -HA, pVL1393-tw WT , and pVL1393-tw Mut were co-transfected with BD BaculoGold Linearized Baculovirus DNA (BD Biosciences, http://www. bdbiosciences.com) into Sf21 insect cells, and the cells were incubated for 7 days at 25uC to produce recombinant viruses. Sf21 Figure 11. Lifespans of flies expressing RNAi for the rt gene. The lifespan of whole-body knockdown flies using Act5C-Gal4 (A), that of neuronspecific knockdown flies using elav-Gal4 (B), and that of glial cell-specific knockdown flies using repo-Gal4 (C). The results of statistical analyses in (A), (B), and (C) are shown in Table 9. The median lifespan for each genotype is shown in Table 10. Flies with ubiquitous expression of RNAi for the rt gene driven by Act5C-Gal4 (Act5C-Gal4/rt-IR) had a shorter lifespan than that of the UAS-rt-IR/+ and Act5C-Gal4/+ control groups (both p,0.001, log-rank test, Table 9). The median lifespan of Act5C-Gal4/rt-IR flies was 23 days, which was 251.1% of that of UAS-rt-IR/+ flies (47 days) and 266.7% of that of Act5C-Gal4/+ flies (69 days) (Tables 9 and 10). On the other hand, the lifespan of [elav-Gal4/+; UAS-rt-IR/+] flies and [UAS-rt-IR/+; repo-Gal4/+] flies was longer than that of [UAS-rt-IR/+; +/+] flies, one control group, and was shorter than that of [elav-Gal4/+; +/+] flies and [+/+; repo-Gal4/+] flies, the other control groups. The median lifespans of [elav-Gal4/+; UAS-rt-IR/+] flies and [UAS-rt-IR/+; repo-Gal4/+] flies were 57 and 59 days, respectively (Table 10). Both lifespans were longer than that of [UAS-rt-IR/+; +/+] flies and shorter than that of [elav-Gal4/+; +/+] flies and [+/+; repo-Gal4/+] flies (Table 9). doi:10.1371/journal.pone.0011557.g011 Table 10. Median lifespan of flies expressing RNAi for the rt gene. Preparation of cellular microsomal membrane fraction and assay of POMT activity The microsomal membrane fraction was prepared and the POMT activity was assayed as described previously [29,55] with some modification. The infected cells were first collected and washed twice in PBS. They were then suspended in a solution containing 10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 250 mM sucrose, 1 mM DTT, and protease inhibitors (5 mg/ml pepstatin A, 2 mg/ml leupeptin, 2 mg/ml aproptin, 1 mM benzamidine-HCl, and 1 mM PMSF) and homogenized using a 1-ml Dounce homogenizer. After centrifugation at 9006g for 10 minutes, the supernatant was subjected to ultracentrifugation at 100,0006g for 1 hr. The pellet was suspended in 20 mM Tris-HCl (pH 8.0), 10 mM EDTA, 2 mM 2-mercaptoethanol, and 0.5% n-octyl-b-D- Figure 12. POMT activity of recombinant wild-type RT and wild-type or mutant TW flies. Western blot analysis of HA-tagged wild-type RT using anti-HA monoclonal antibody (A) and of TW using anti-TW antibody (B). The prepared microsomal membrane fractions of infected cells were applied to 7% SDS-PAGE at 20 mg for western blot analysis with anti-HA monoclonal antibody and 5 mg for that with anti-TW antibody. The arrowhead and arrow indicate HA-tagged RT and TW, respectively. (C) POMT activity for GST-a-DG of recombinant mutant TW (TW Mut ). POMT-specific activity was detected when RT WT -HA and TW WT were co-expressed as we reported. However, POMT-specific activity was not detected when RT WT -HA and TW Mut were co-expressed. Each bar represents the mean of 8 replicates. Error bars indicate standard error. ***p,0.001 by the Tukey test. n.s, not significant. (D) POMT activity for GST-a-DG of larval extract of rt and tw mutant. POMT-specific activities of rt and tw mutant were significantly reduced. Each bar represents the mean of 6 replicates. Error bars indicate standard error. ***p,0.001 by t test. doi:10.1371/journal.pone.0011557.g012 thioglucoside, and this suspension was used as the microsomal membrane fraction. The reaction mixture contained 20 mM Tris-HCl (pH 9.0), 10 mM EDTA, 2 mM 2-mercaptoethanol, 0.5% noctyl-b-D-thioglucoside, 100 nM Dol-P-[ 3 H]Man (133,200 dpm/ pmol) (American Radiolabeled Chemicals, http://www.arcincusa. com/), 2.5 mg GST-a-hDG, and 80 mg of microsomal membrane fraction as the enzyme source in a total volume of 20 ml. After 1-hr incubation at 18uC, the reaction was stopped by addition of 100 ml of PBS containing 1% Triton X-100, and the reaction mixture was centrifuged at 10,0006g for 5 minutes. The supernatant was transferred, mixed with 400 ml of PBS and 20 ml of Glutathione-Sepharose TM 4B beads, rotated for 1 hr, and washed 3 times with PBS containing 0.2% Triton X-100. The radioactivity of the beads was measured using a liquid scintillation counter.

Western blot analysis
Each microsomal membrane fraction was subjected to 7% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The membrane to which the separated proteins were transferred was probed with anti-HA mouse monoclonal antibody (1:2000 dilution) (Santa Cruz Biotechnology, Inc., http://www. scbt.com/) or anti-dPOMT2 rabbit polyclonal antibody (1:400 dilution) [29]. Each membrane was then reacted with HRPconjugated secondary antibody and stained with Amersham TM ECL TM Plus (GE Healthcare).

Preparation of larval extract and assay of POMT activity
Fourty third instar larvae of rt EP /rt KG , tw/Y, and wild-type (Canton-S) flies were homogenized in 20 mM Tris-HCl (pH 8.0), 10 mM EDTA, 2 mM 2-mercaptoethanol, and 0.5% n-octyl-b-Dthioglucoside with protease inhibitors (5 mg/ml pepstatin A, 2 mg/ ml leupeptin, 2 mg/ml aproptin, 1 mM benzamidine-HCl, and 1 mM PMSF) (300 ml for every 40 larvae). The supernatant was obtained by centrifugation at 9,0006g for 10 minutes and used as larval extract. The same reaction mixture and conditions described in the ''Preparation of cellular microsomal membrane fraction and assay of POMT activity'' section above were used except for the following 2 modifications: (1) the amount of GST-a-hDG in the reaction mixture was 10 mg; and (2) the incubation time was 2 hr.

Statistical analyses
All statistical analyses were performed using the public domain R program (http://www.r-project.org/). Figure S1 Quantitative analysis of rt and tw mRNAs in rt mutant fly. rt (A)and tw (B) transcript levels of rt EP /rt KG mutant flies of third instar larvae were determined by real-time PCR. Error bars indicate standard error. Lines above the bars show compared groups by one-way ANOVA. *p,0.05 by the one-way ANOVA. n.s., not significant. * above the bar of rt EP /rt KG means p,0.05 by Tukey test. levels of the flies with the expression of RNAi for rt gene of third instar larvae were determined by real-time PCR. The expression level of rt in Act5C.rt-IR was significantly reduced, while that in other lines was not significantly reduced. The expression level of tw was no different among all lines. Error bars indicate standard error. Lines above the bars show compared groups by one-way ANOVA. ***p,0.001 by the one-way ANOVA. n.s., not significant. * above the bar of Act5C.rt-IR means p,0.05 by Tukey test.