The myopathy-causing mutation DNM2-S619L leads to defective tubulation in vitro and in developing zebrafish

DNM2 is a ubiquitously expressed GTPase that regulates multiple subcellular processes. Mutations in DNM2 are a common cause of centronuclear myopathy, a severe disorder characterized by altered skeletal muscle structure and function. The precise mechanisms underlying disease-associated DNM2 mutations are unresolved. We examined the common DNM2-S619L mutation using both in vitro and in vivo approaches. Expression of DNM2-S619L in zebrafish led to the accumulation of aberrant vesicular structures and to defective excitation-contraction coupling. Expression of DNM2-S619L in COS7 cells resulted in defective BIN1-dependent tubule formation. These data suggest that DNM2-S619L causes disease, in part, by interfering with membrane tubulation.

neuropathy (Züchner et al., 2005). DNM2 is a large GTPase that is a member of the dynamin family of membrane fission proteins. Numerous functions, including the regulation of endocytosis, cell migration and cytokinesis, are associated with the normal function of DNM2 (Jones et al., 1998;McNiven et al., 2000;Thompson et al., 2004). The mechanism(s) via which dominant mutations in DNM2 cause myopathy, however, are incompletely understood. Data from cell culture models has been conflicting; for example, some studies have identified abnormalities in endocytosis, whereas others have not (Bitoun et al., 2009;Koutsopoulos et al., 2011;Liu et al., 2011;Tanabe and Takei, 2009). Examination by immunohistochemistry of patient muscle biopsies revealed disorganization of the triad, and recent studies using viral expression in the mouse corroborated this observation (Cowling et al., 2011). However, the mechanism(s) by which DNM2 mutations impair triad formation is not known; furthermore, a functional effect on excitation-contraction coupling has yet to be documented.
In this study, we examine the S619L mutation in the pleckstrin homology domain of DNM2. The S619L mutation is one of the most common DNM2 mutations and is associated with a severe centronuclear myopathy muscle phenotype (Bitoun et al., 2007;Böhm et al., 2012). We determined the impact of DNM2-S619L expression in both the developing zebrafish and in a cell culture model of tubule formation (Bitoun et al., 2007). We show that the mutation results in defective membrane tubulation, and propose that one mechanism of disease is the interference of BIN1-mediated tubule formation and subsequent impairment of excitationcontraction coupling.

RESULTS AND DISCUSSION
Expression of DNM2-S619L in developing zebrafish disrupts triad structure To generate a model of centronuclear myopathy due to DNM2 mutation, we ubiquitously expressed either wild-type (DNM2-WT) or S619L-mutation-containing (DNM2-S619L) human DNM2 RNA in the developing zebrafish starting at the one-to two-cell stage (supplementary material Fig. S1). Expression of DNM2-S619L results in abnormalities in muscle structure (including abnormally located nuclei and perinuclear disorganization), impaired force generation and altered motor function. The general phenotypic features of DNM2-S619L zebrafish are recently published . In this study we specifically focused on the triad because of the mounting data suggesting that triad defects are a crucial aspect of centronuclear myopathy (Al-Qusairi et al., 2009;Dowling et al., 2009;Toussaint et al., 2011). We performed ultrastructural evaluation of the triad structure in zebrafish with confirmed expression of either DNM2-WT or DNM2-S619L.
Although there were rare regions of normal triads, the majority did not resemble the native structure but were instead characterized by disorganized and swollen membrane structures. The most striking aspect of the triad defects was the abundant accumulation of large, irregular vesicular structures that correspond in location and by general appearance to the terminal sarcoplasmic reticulum. Overall, these findings reflect substantial membrane abnormalities in the muscle of DNM2-S619L zebrafish.
To support our assertion that the abnormal structures seen by electron microscopy are, in fact, triads, we performed immunofluorescent analysis of triad markers on myofibers isolated from RNA-injected embryos. Myofibers from uninjected and DNM2-WT-injected embryos demonstrated the expected pattern of RyR1 (Fig. 1E) and DHPR (data not shown) staining. Conversely, myofibers from DNM2-S619L zebrafish had widespread abnormalities in RyR1 and DHPR staining, including aggregation of staining and interruption of the normal confluent linear pattern (Fig. 1F,G). Of note, α-actinin staining (highlighting Z-bands) was normal (Fig. 1H), suggesting that the changes in triad marker immunofluorescence were not due to non-specific dissolution of the entire myofiber structure. In all, these data corroborate the ultrastructural analysis and demonstrate severe disturbance of the triad structure with DNM2-S619L expression.

Impaired excitation-contraction coupling in the muscle of DNM2-S619L larvae
The abnormalities in triads in our zebrafish model correspond well with the immunohistological abnormalities previously reported in DNM2 patient biopsies and the ultrastructural changes described in the viral overexpression model of DNM2-associated centronuclear myopathy (Cowling et al., 2011;Toussaint et al., 2011). The potential impact of structural triad changes on triad function, however, has yet to be examined for any DNM2 model. In order to determine the functional effect(s) of DNM2-S619L expression on excitation-contraction coupling, we examined calcium transients in spontaneously contracting embryos using a genetically encoded calcium indicator, GCaMP3. Embryos co-injected with DNM2 RNA and GCaMP3 cDNA were screened for variegated GCaMP expression in skeletal muscle ( Fig. 2A). Both DNM2-WT and DNM2-S619L larvae expressed GCaMP3 at a similar level. At 24 hours post fertilization (hpf), embryos were embedded in

Clinical issue
Congenital myopathies are a heterogeneous group of childhood-onset muscle diseases associated with significant morbidity and early mortality. To date, no therapies are available for any congenital myopathy. Dominant mutations in DNM2 (dynamin-2) are associated with a common congenital myopathy subtype called autosomal dominant centronuclear myopathy. Much remains to be understood regarding the pathogenesis of this condition, with one key barrier being the relative lack of adequate models of the disease. The goal of this study was to further characterize zebrafish and cell culture models of DNM2-related myopathy and to use these models to better understand disease pathogenesis and potentially inform the development of suitable therapies.

Results
The skeletal muscle phenotype of early-stage zebrafish embryos injected with wild-type or mutant DNM2 RNA was characterized. Larvae expressing mutant DNM2 (DNM2-S619L) had structural abnormalities in the triad, the skeletal muscle substructure that represents the intersection of the T-tubule and terminal sarcoplasmic reticulum. Subsequent functional studies revealed that contraction-related calcium release was also defective in DNM2-S619L embryos. The authors also subjected the larvae to ultrastructural analysis, which, combined with the functional data, indicated a defect in excitation-contraction coupling. This led to a hypothesis that the DNM2-S619L mutation promotes defective membrane tubulation, which was tested using an in vitro tubulation assay. Exogenous expression of DNM2-S619L cDNA (but not wild-type DNM2) into COS7 cells altered the production of normal tubules, thus supporting the idea that DNM2 mutation, at least in part, causes disease by preventing proper tubulation.

Implications and future directions
The data in this short report demonstrate that a common congenitalmyopathy-associated mutation disturbs tubule formation, both in vitro and in a zebrafish model of the disease. In turn, this abnormal tubulation prevents normal excitation-contraction coupling, which likely leads to muscle weakness and impaired muscle force generation. This new knowledge of underlying disease pathogenesis addresses an unanswered aspect of DNM2-related myopathy, and will serve as a springboard for rational therapy development for this devastating childhood disorder and related congenital myopathies. agarose and fluorescence was measured during spontaneous muscle contractions. Both DNM2-WT and DNM2-S619L embryos displayed abundant spontaneous muscle contractions at this point. In DNM2-WT embryos, a pulse of fluorescence corresponding to sarcoplasmic reticulum (SR)-based calcium release was temporally linked to each contraction. In DNM2-S619L embryos, however, there was minimal or no visible increase in fluorescence during muscle contraction in 10 of 12 embryos examined (Fig. 2C,D). In order to directly compare the fluorescence changes between embryos, we calculated the relative change in fluorescent recordings before versus during a muscle contraction (ΔF/F). We observed a significant difference in this dynamic ratio of calcium transients in DNM2-WT as compared with DNM2-S619L embryos ( Fig. 2B and supplementary material Fig. S2). This indicated that, in addition to altering triad organization in zebrafish muscle, DNM2-S619L expression impairs stimulus-associated intracellular calcium release, a triad-mediated muscle function.

DNM2-S619L expression alters the formation of BIN1induced tubules
The above data reveal that DNM2-S619L expression can disrupt triad formation and function, but do not provide a potential mechanism via which it may do so. It has been previously established that DNM2-WT can interact with BIN1 (also called amphyphysin-2), a protein that can sense and regulate membrane curvature (Takei et al., 1999). A muscle-specific isoform of BIN1 (BIN1 isoform 8) has been shown to be required for T-tubule biogenesis, and recessive mutations in BIN1 cause an autosomalrecessive form of centronuclear myopathy associated with triad defects (Lee et al., 2002;Razzaq et al., 2001). Based on these known functions of BIN1, we hypothesized that DNM2-S619L mutations function to impair triad structure by interfering with BIN1-mediated tubulogenesis. To test this hypothesis, we used a well-established in vitro model of BIN1dependent tubule formation (Nicot and Laporte, 2007;. Expression of BIN1 isoform 8 (iso8) can induce the formation of tubular membrane structures when expressed in non-muscle cells, thus enabling an in vitro model of T-tubule biogenesis. To examine the impact of DNM2-S619L expression in this system, we coexpressed in COS7 cells GFP-tagged BIN1 iso8 and either DNM-WT cDNA or DNM2-S619L cDNA tagged with mCherry. Consistent with previously published results, we found by confocal microscopy that co-expression of BIN1 iso8 and DNM2-WT resulted in extensive tubulation (Fig. 3A). When DNM2-S619L was co-expressed with BIN1 iso8, however, we observed predominantly the formation of short punctate structures instead of tubules (Fig. 3B). To quantify these effects on membrane organization, tubulation in each cell was classified as short, long or intermediate. Whereas DNM2-WT-expressing cells exhibit a phenotype consisting of long tubules, only 19% of DNM2-S619L-expressing cells had a similar phenotype ( Fig. 3D; DNM2-WT n=68, DNM2-S619L n=73, cells from three independent experiments). Thus, compared with DNM2-WT, expression of DNM2-S619L results in a significant impairment in BIN1-dependent tubule formation.

Conclusions
We demonstrate that the common myopathy-causing S619L mutation in DNM2 disrupts triad structure and function in vivo and BIN1induced tubulation in vitro. These data represent the first direct evidence that a DNM2 mutation can interfere with tubulation and disrupt triadic function. Based on these novel findings, we hypothesize that muscle disease related to DNM2 mutations is caused by the production of dominantly active defective DNM2 proteins that interfere with the tubulation process. Disrupted tubulation, in turn, impairs triad formation and excitation contraction, which results in abnormal muscle force generation and ultimately muscle weakness . Future experimentation, particularly with the testing of additional DNM2-associated centronuclear myopathy mutations, will be required to support this hypothesis. These data, however, provide an important step toward establishing the pathomechanisms related to DNM2 mutations.

Animal care
Zebrafish (AB strain) were bred and raised according to established protocols approved by the University of Michigan Animal Care and Use Committee (UCUCA protocol #09835).

Plasmid construction and RNA synthesis
Wild-type human DNM2 plasmid was purchased from Invitrogen (ORF Gateway ® Entry IOH53617). The S619L point mutation was introduced using the QuikChange Lightning Site-directed Mutagenesis kit (Stratagene). Expression vectors were generated using the Gateway system and p5E-CMV/SP6, p3E-polyA and pDestTol2pA2 cassettes from the Tol2kit v1.2 (kind gift of Dr Chi-Bin Chien, University of Utah, UT). Gateway recombination reactions were performed using LR Clonase II Plus Enzyme Mix (Invitrogen). For RNA injection, plasmids were linearized with NotI and transcribed using the SP6 mMessage Machine kit (Ambion).

RNA injection of zebrafish embryos
Zebrafish embryos were injected as previously described (Dowling et al., 2009). Briefly, fertilized eggs were injected at the one-to two-cell stage using a Nanoject II injector (Drummond Scientific). Embryos were injected with DNM2 RNA (30 ng/μl) in a 4.6 nl volume.

Muscle ultrastructure
Zebrafish at 3 dpf were fixed overnight in Karnovsky's fixative and then processed for electron microscopy by the Microscopy and Imaging Laboratory (MIL) core facility at the University of Michigan. Electron microscopy was performed using a Phillips CM-100 transmission electron microscope.

Measurements of calcium transients
The construct containing GCaMP under a muscle actin promoter was a kind gift of Dr John Kuwada (University of Michigan, MI). Plasmid DNA was co-injected with DNM2 RNA at a concentration of 10 ng/μl in one-to twocell embryos. At 24 hours, embryos were embedded and immobilized in 1% low-melt agarose. Fluorescent intensity during 10 seconds of spontaneous muscle contractions was recorded using a Nikon AZ-100 microscope. For image processing, ImageJ was used to measure total fluorescent intensity in a region of the larval trunk over the recorded interval. Baseline fluorescent levels were used to calculate relative fluorescence (ΔF/F).

Ex vivo tubulation in COS cells
Ex vivo tubulation was performed following the basic methods described previously   exhibited punctate tubules with lengths no greater than four times the tubule width, whereas cells categorized as 'long' displayed tubules that extended uninterrupted to the plasma membrane. Cells with tubules longer than four times the width that did not fully extend to the plasma membrane were classified as 'intermediate'.

Statistical analysis
Statistical analysis was performed using the GraphPad Prism 5 software package. Significance was determined using Student's t-test or Fisher's exact test.