Drosophila emerins control LINC complex localization and transcription to regulate myonuclear position

Skeletal muscle cells are characterized in part by the many nuclei that share a common cytoplasm. After the many nuclei are incorporated by iterative rounds of fusion, nuclei undergo a complex set of movements that leave them evenly spaced at the periphery of the cell. These movements are conserved throughout evolution (Folker and Baylies, 2013), and nuclei are mispositioned in the muscle cells of individuals with various muscle disorders (Dubowitz et al., 2007), suggesting that they are fundamental to muscle development.

One particular disease, Emery–Dreifuss muscular dystrophy (EDMD), is caused by mutations in a set of genes that encode proteins that mechanically link the nucleus to the cytoskeleton (Folker and Baylies, 2013; Crisp et al., 2006; Lombardi et al., 2011; Starr and Han, 2002). Many of the genes that are mutated in patients with EDMD encode for proteins that localize to the nucleus, including the inner nuclear membrane protein emerin, the structural nuclear protein lamin A/C, and proteins belonging to the linker of nucleoskeleton and cytoskeleton complex (LINC complex). The LINC complex is composed of SUN-domain proteins and KASH-domain proteins. SUN-domain proteins span the inner nuclear membrane and interact with the nucleoskeleton within the nucleus and with KASH-domain proteins in the lumen of the nuclear envelope. KASH-domain proteins span the outer nuclear membrane and interact with the cytoskeleton in the cytoplasm. Functionally, these proteins are critical for nuclear positioning in muscle cells (Roman and Gomes, 2018). Additionally, each gene is critical for nuclear movement in non-muscle cell types (Gundersen and Worman, 2013) indicating that regulation of nuclear movement is a fundamental function of EDMD-linked genes. However, how these individual components are coordinated to move nuclei is not known.

Emerin, the first identified cause of EDMD, is a LEM domain containing protein that is primarily localized to the inner nuclear membrane (Bione et al., 1994; Nagano et al., 1996; Manilal et al., 1996). LEM domain proteins interact with lamin and barrier-to-autointegration factor, and through these interactions can localize chromosomes to the nuclear periphery (Zheng et al., 2000; Cai et al., 2001; Laguri et al., 2001). Emerin also interacts with SUN1 and SUN2 (Haque et al., 2010) as well as short isoforms of KASH-domain proteins that localize to the inner nuclear envelope (Mislow et al., 2002; Wheeler et al., 2007). How each of these functions contributes to muscle development in general, or nuclear positioning during muscle development specifically, is not known. Drosophila provide an interesting system in which to study this mechanism as the Drosophila genome encodes only three LEM domain-containing proteins, MAN1, Bocksbeutel and Otefin (Ashery-Padan et al., 2015, 1997; Wagner et al., 2004, 2006; Pinto et al., 2008). Drosophila MAN1 is the homolog of mammalian LEM2 and MAN1 but both Bocksbeutel and Otefin are homologs of emerin (Wagner et al., 2006). Within the LEM domains Bocksbeutel and Otefin are 70% similar; however, outside of the LEM domain the similarity drops to 28% (Barton et al., 2014). Additionally, the expression patterns of these two emerin homologs differ, with uniform expression of Bocksbeutel throughout development, while Otefin is more highly expressed in embryos and first-instar larvae compared to later developmental stages (Wagner et al., 2006). The existence of two emerin-like proteins makes it possible that emerin functions are distributed between two separate proteins in Drosophila and therefore might simplify the process of understanding how each emerin function is coordinated and the contribution of each to muscle development.

We investigated the effects of emerin and other genes linked to EDMD during muscle development in Drosophila embryos and larvae (Collins et al., 2017). Consistent with previous reports, several EDMD-linked genes are critical for proper nuclear positioning (Elhanany-Tamir et al., 2012; Collins et al., 2017). Additionally, the Drosophila emerin homologs, bocksbeutel and otefin, both regulate nuclear position in embryonic and larval stages. However, the precise nuclear-positioning phenotypes that arise upon disruption of these genes differ. These differences are based on their distinct effects on the nuclear localization of the Drosophila KASH-domain protein, Klarsicht. Thus, nuclear level of Klarsicht is a critical regulator of nuclear positioning, which is differentially regulated by bocksbeutel and otefin.

Like mammalian skeletal muscles, Drosophila body wall muscles contain many nuclei that are precisely positioned to maximize the distance between nuclei. Two EDMD-linked genes, bocksbeutel (bocks, a Drosophila emerin) and klarsicht (klar, a Drosophila KASH-domain protein), are known regulators of nuclear positioning in both embryonic and larval muscle (Elhanany-Tamir et al., 2012; Collins et al., 2017). Whether the effect on nuclear position in muscle is a conserved consequence of disrupting EDMD-linked genes or specific to those two genes is not clear. Furthermore, the genetic mechanism(s) by which these genes regulate nuclear position is not known. As a first step toward answering both questions, we measured nuclear position in flies with mutations that had previously not been characterized, namely otefin (ote, a Drosophila emerin) and klaroid (koi, the Drosophila SUN protein).

In stage 16 control embryos, nuclei were positioned in two equal-sized clusters with one near the dorsal end of the muscle and the other near the ventral end of the muscle (separated; equal distribution) (Fig. 1A,B). In bocks^DP01391 homozygous mutants and klar¹ homozygous mutants, nuclei remained as a single cluster near the ventral end of the muscle (clustered phenotype), spread through the center of the muscle with no distinct dorsal or ventral cluster (spread phenotype) or separated into two clusters of unequal size (separated; unequal distribution) as previously described (Fig. 1A,B) (Collins et al., 2017). In animals with the ote^DB mutation, an amorphic allele caused by a nonsense mutation (Barton et al., 2013) that had not previously been investigated with respect to nuclear positioning, nuclei separated into two clusters. However there was an increase in the frequency of nuclei found in the center of the muscle (central phenotype) (Fig. 1A,B). In animals with the koi^EY03560 mutation, an allele with a p-element insertion in an early intron of koi (Technau and Roth, 2008) previously uninvestigated with respect to nuclear positioning, nuclei remained as a single cluster or separated into two clusters of unequal size as was seen in bocks^DP01391 and klar¹ mutants (Fig. 1A,B).

We also measured the distance from the dorsal and ventral ends of muscles to the nearest nucleus. Compared to controls, nuclei were positioned 63%, 59% and 48% further away from the dorsal muscle end in bocks^DP01391, klar¹ and koi^EY03560 embryos, respectively. Additionally, compared to controls, nuclei were positioned 11%, 12% and 15% closer to the ventral end of the muscle. Conversely, compared to controls, nuclei were positioned 11% closer to the dorsal muscle end in ote^DB embryos (Fig. 1C,D). Crossing each mutant allele to a deficiency replicated the nuclear positioning defects observed in homozygous mutants (Fig. S1A–D), except for bocks^DP01391/Df, which had a high frequency of fusion defects and missing muscles (Fig. S1A). Owing to the muscle defects in bocks^DP01391/Df embryos, only embryos with properly formed muscles were analyzed for nuclear positioning, leading to a bias in analysis of healthier embryos. Together, these data indicate that bocksbeutel, klarsicht and klaroid have a similar effect on nuclear positioning but that otefin has unique effects on nuclear position.

To test whether the disruptions in nuclear positioning persist through larval development, we measured nuclear distribution in bocks^DP01391, klar¹, ote^DB and koi^EY03560 mutant L3 larvae as previously described (Collins et al., 2017). In controls, nuclei were typically positioned in two parallel lines on the long axis of the muscle (Fig. 1E) with an internuclear distance ratio (actual distance as a proportion of maximal distance) of 0.75 (Fig. 1F). In bocks^DP01391 and klar¹ larvae, nuclei formed a single line in the center of the muscle as previously described (Fig. 1E) (Collins et al., 2017), which phenocopied what was seen in klar-null larvae (Ding et al., 2017). The internuclear distance ratio was 0.57 and 0.58 for bocks^DP01391 and klar¹ larvae, respectively (Fig. 1F). Nuclear positioning in koi^EY03560 larvae phenocopied bocks^DP01391 and klar¹ larvae with a single line of nuclei positioned in the center of the muscle (Fig. 1E) with an internuclear distance ratio of 0.6 (Fig. 1F). Additionally, nuclei in bocks^DP01391, klar¹ and koi^EY03560 larvae were all positioned further from the muscle edge compared to controls (Fig. 1G). It has been previously demonstrated that nuclear positioning relative to the muscle edge scales with cell width (Windner et al., 2019). Therefore, we measured the average width of VL3 muscles and found that there was no significant difference between the average width of control, bocks^DP01391, klar¹ and koi^EY03560 muscles (Fig. 2A) suggesting that the observed nuclear positioning defects are not a consequence of thinner muscles and instead are bona fide disruptions in nuclear positioning.

To further investigate these nuclear positioning disruptions, we analyzed the relationship between the average distance of nuclei from the muscle edge as a function of muscle width. The slope of the line fit to the control data was nearly zero (Fig. 2B,C,F), indicating that the nuclei maintain a relatively constant distance from the edge of the muscle regardless of the muscle size. In bocks^DP01391, klar¹ and koi^EY03560 animals, nuclei were positioned further from the muscle edge as the muscle widened, indicating that nuclei had lost the ability to separate into two lines and maintain the proper distance from the muscle edge (Fig. 2C–F,H). The similarities in all measurements of nuclear position between bocksbeutel, klarsicht and klaroid mutants suggest that these genes regulate nuclear positioning through a common mechanism.

In contrast, the spacing between nuclei in ote^DB larvae was similar to controls. However, in some regions of the muscle there were three lines of offset nuclei (Fig. 1E, yellow box). This resulted in disrupted nuclear positioning relative to the muscle edge (Fig. 1G), but proper spacing of nuclei relative to other nuclei (Fig. 1F). Additionally, as muscle width increased, the distance between the muscle edge and the nuclei scaled similarly to controls (Fig. 2G). This indicates that although nuclei were closer to the muscle edge in ote^DB mutants, the effect is not caused by a difference in muscle width.

Crossing each mutant allele to a deficiency replicated the nuclear positioning defects observed in homozygous mutants (Fig. S1E–G), with the exception of ote^DB/Df. ote^DB/Df had more severe nuclear positioning defects compared to homozygous mutant larvae, possibly due to the additional genes affected by the deficiency. Therefore, to confirm that the ote mutation caused the nuclear positioning defects, we examined trans-heterozygotes of two distinct ote alleles. Similar to ote^DB homozygote larvae, the internuclear distance in ote^DB/B279 larvae was similar to controls, but nuclei were significantly closer to the edge of the muscle than in controls (Fig. S2A–C). That the disruption of otefin expression caused a distinct phenotype in homozygous and trans-heterozygous mutants further indicates that otefin regulates nuclear positioning differently from bocksbeutel, klarsicht and klaroid.

Since bocks^DP01391 mutants share nuclear-positioning phenotypes with klar¹ and koi^EY03560 mutants, we investigated whether bocksbeutel and the other Drosophila emerin homolog, otefin, genetically interact with other EDMD-linked genes to regulate nuclear positioning during embryonic and larval muscle development. In bocks^DP01391/+, klar¹/+ doubly heterozygous embryos, the frequencies of clustered nuclei, spread nuclei and two separate clusters of unequal size were increased compared to what was seen in either individual heterozygote (Fig. 3A,B). Additionally, the distance between the dorsal muscle end and the nearest nucleus was increased relative to that in each individual heterozygote (Fig. 3C), and the distance between the ventral end of the muscle and the nearest nucleus was decreased compared to each individual heterozygote (Fig. 3D). In bocks^DP01391/+, klar¹/+ doubly heterozygous larvae, nuclei formed a single line positioned in the center of the muscle, parallel to the long axis of the muscle. This phenotype was absent from each of the individual heterozygotes, although some regions of bock^DP01391/+ and klar¹/+ single heterozygote larval muscles contained single-file nuclei (Fig. 3E). Quantitatively, the internuclear distance ratio was significantly reduced in bocks^DP01391/+, klar¹/+ doubly heterozygous larval muscles compared to either individual heterozygote (Fig. 3F). Nuclei were also further from the muscle edge in bocks^DP01391/+, klar¹/+ double heterozygotes compared to each individual heterozygote (Fig. 3G). No genetic interactions were found between bocks^DP01391 and either ote^DB or koi^EY03560 in embryonic (Fig. S3A–F) or larval (Fig. S4A–E) muscles. Additionally, no genetic interactions were found between ote^DB and klar¹ and koi^EY03560 in embryonic (Fig. S3G–L) or larval (Fig. S4F–J) muscles. The genetic interaction between bocksbeutel and klarsicht suggests that Klarsicht may be a differentiating factor between the distinct mechanisms used to regulate nuclear positioning by Bocksbeutel and Otefin.

In order to better understand the genetic interaction between bocksbeutel and klarsicht, we examined Klarsicht localization in bocks^DP01391 and klar¹ mutants. In both bocks^DP01391 and klar¹ mutants, nuclear Klarsicht levels were reduced compared to those in controls (Fig. 4A,B). Combined with the similar nuclear-positioning phenotype in both bocksbeutel and klarsicht mutants, these data suggest that Bocksbeutel contributes to nuclear position by regulating the levels of Klarsicht in the nuclear envelope. Similarly, in koi^EY03560 mutants, which exhibit nuclear positioning defects similar to those in bocks^DP01391 and klar¹ (Fig. 1A–G), nuclear Klarsicht levels were also reduced compared to controls (Fig. 4A,B). Conversely, in ote^DB mutants, which are phenotypically distinct from bocks^DP01391, klar¹ and koi^EY03560 (Fig. 1A–G), nuclear Klarsicht levels were increased compared to controls (Fig. 4A,B). These data, combined with the clustering phenotype in embryonic muscles and the single-file nuclear-positioning phenotype in larval muscles being phenocopied in our klar^1/Df mutant (Fig. S1E–G) and upon muscle-specific knockdown of klar by RNAi (Collins et al., 2017), suggest that these nuclear-positioning phenotypes result from a reduction in Klarsicht in the nuclear envelope, while the central nucleus phenotype in embryonic muscles and nuclei positioning closer to the edge in larval muscles results from an increase in Klarsicht in the nuclear envelope. To test the latter hypothesis, we overexpressed Klarsicht. Overexpressed Myc-tagged Klarsicht localized to the nuclear envelope and, similar to what was seen with the ote^DB mutation, caused nuclei to be positioned closer to the edge of the muscle compared to controls (Fig. 5A–D). These data further suggest that reduced Otefin levels disrupt nuclear position by causing an increase in Klarsicht in the nuclear envelope.

Because the LINC complex and emerin can either directly or indirectly regulate transcriptional activity (Lee et al., 2001; Wilkinson et al., 2003; Holaska and Wilson, 2007; Navarro et al., 2016), we investigated the transcript levels of EDMD-linked genes in each mutant to determine whether nuclear Klarsicht levels changed due to altered transcription or altered protein localization. All transcript levels were the same in bocks^DP01391 mutants and controls. In particular, there was no change in klar transcript levels (Fig. 4C), suggesting that the decreased Klarsicht immunofluorescence represented a change in its localization. Additionally, no changes in EDMD-linked genes were observed at the transcript level in klar¹ mutants compared to controls (Fig. 4D). However, in ote^DB mutants there was an increase in klar transcript levels compared to controls (Fig. 4E) suggesting that the increase in nuclear Klarsicht levels is caused by an increase in transcription of the klar gene. Additionally, in koi^EY03560 mutants, there was a significant increase in bocks transcript levels compared to controls (Fig. 4F) suggesting that SUN protein levels contribute to the regulation of emerin gene transcription.

Since bocks^DP01391 homozygous mutants and ote^DB homozygous mutants have different effects on nuclear Klarsicht levels, we investigated whether the functions of Bocksbeutel and Otefin counterbalance each other to regulate nuclear positioning. In bocks^DP01391 larvae (Figs 1E and 5E), nuclei formed a single line positioned in the center of the muscle, parallel to the long axis of the muscle. When a single copy of the ote^DB mutant allele was placed in a bocks^DP01391 homozygous mutant, there was a partial rescue of nuclear positioning with the internuclear distance ratio being 0.71 compared to 0.56 in the bocks^DP01391 homozygous mutant controls (Fig. 5E,F). Nuclear positioning was completely rescued in ote^DB; bocks^DP01391 larvae as nuclei formed two parallel lines along the long axis of the muscle (Fig. 5E) with an internuclear distance ratio of 0.78 (Fig. 5F). Additionally, the positioning relative to the muscle edge was rescued in ote^DB/+; bocks^DP01391 and ote^DB; bocks^DP01391 (Fig. 5G). Furthermore, the levels of nuclear-localized Klarsicht were rescued to control levels in ote^DB; bocks^DP01391 double mutants (Fig. 5H,I). These data suggest that Bocksbeutel and Otefin have opposing functions in regulating nuclear positioning and balancing these functions is necessary to properly position myonuclei and regulate nuclear Klarsicht levels.

We have used Drosophila musculature to elucidate the genetic network and cellular mechanisms that regulate myonuclear position in vivo. Consistent with previous work, disruption of EDMD-linked genes caused mispositioned nuclei. Deeper characterization revealed that disruption of bocksbeutel (a Drosophila emerin), klarsicht (a Drosophila KASH-domain protein) and klaroid (the Drosophila SUN) caused a similar phenotype. However, disruption of otefin, the other Drosophila emerin homolog caused a different nuclear-positioning phenotype. Furthermore, bocks, but not ote, genetically interacted with klar to regulate nuclear positioning. The distinct phenotypes and genetic interactions of bocks and ote suggest a division of emerin functions between the two Drosophila emerin homologs.

Mechanistically, the phenotypic differences between bocks and ote mutants correlate with distinct changes in nuclear-localized Klarsicht. Disruption in bocks leads to a decrease in nuclear-localized Klarsicht while disruption in ote leads to an increase in nuclear-localized Klarsicht. Disruption of bocks caused no effect on transcript levels of klar, suggesting that the decrease in nuclear Klarsicht is due to mislocalization of Klarsicht. Conversely, disruption in ote caused an increase in transcript levels of klar, suggesting the increase in nuclear-localized Klarsicht is due to an increase in transcription. Although the increase in transcript levels is modest, it is important to note that quantitative PCR (qPCR) was conducted on whole-larval lysates. Therefore, if the phenotype were specific to a subset of tissues, or perhaps muscle specific, this would explain the modest change in transcript levels we observed. Together, these data suggest that bocks and ote serve unique functions in Drosophila abdominal muscles, but that both functions are critical to the regulation of nuclear positioning.

Nesprins and emerin have previously been shown to interact physically. However, these interactions were demonstrated between shorter nesprin isoforms that localize to the inner nuclear membrane and function independently of the LINC complex (Mislow et al., 2002; Wheeler et al., 2007). In cell culture, emerin interacts with SUN proteins (Haque et al., 2010), but they do not rely on each other for nuclear envelope localization. If Bocksbeutel is not necessary for Klaroid localization, the decrease in nuclear-localized Klarsicht may be caused by LINC complex instability, possibly through lamins, which have been shown to be disrupted in large polytene nuclei that lack Bocksbeutel (Barton et al., 2014). In support of this, bocksbeutel expression was increased in koi mutants, perhaps to compensate for the loss of koi. Nevertheless, these data suggest that a loss of Klarsicht from the nuclear envelope is a driving factor of mispositioned myonuclei in bocks, klar and koi mutants. In support of this, the klar^1/Df mutant (Fig. S1A–G) and muscle-specific knockdown of klar by RNAi (Collins et al., 2017) phenocopy bocks^DP01391 mutants, klar¹ mutants and koi^EY03560 mutants with the clustering phenotype in embryonic muscles and the single file nuclear-positioning phenotype in larval muscles. Furthermore, KASH-domain protein levels, such as Klarsicht, at the nucleus being a driving factor of mispositioned nuclei may not be unique to EDMD as Nesprin-1 levels at the nucleus have also been found to be reduced in the MDX mouse model for Duchenne muscular dystrophy (Iyer et al., 2017).

The increased amount of Klar at the nuclear envelope in ote mutants, and the associated nuclear-positioning phenotype, suggest that any variations in Klarsicht abundance at the nuclear envelope will impact nuclear position in muscle. Consistent with this, overexpression of Klarsicht in a muscle-specific manner phenocopied larval nuclear positioning relative to the muscle edge (Fig. 5A–C). Additionally, as an increase in Klarsicht leads to mispositioned nuclei, an increase in Klaroid has also been demonstrated to disrupt nuclear positioning (Tan et al., 2018). These data suggest that misregulation of LINC complex components that lead to a change in protein levels at the nucleus may be a common mechanism through which nuclear position is disrupted.

Remarkably, we found that loss of Otefin was sufficient to rescue the nuclear-positioning phenotypes present in bocks^DP01391 homozygous mutants, including both nuclear-localized Klarsicht levels and nuclear positioning. Furthermore, even the introduction of a single ote^DB mutant allele was able to partially rescue nuclear positioning in bocks^DP01391 homozygous mutants, indicating the distinct emerin functions divided between the two Drosophila emerin homologs, bocksbeutel and otefin, must be balanced for proper nuclear positioning.

Taken together, these data suggest that nuclear positioning can be disrupted not only by the loss of LINC complex components but also increases in LINC complex components. Emerin is a critical regulator of LINC complex levels in the nucleus. Both the expression of Klarsicht and the localization of Klarsicht are regulated by emerin. However, we found here that in Drosophila, these two functions are divided among the two Drosophila emerin homologs, bocksbeutel and otefin. Thus, the specification of emerin activity may be the critical determinant of nuclear position and function. Given the functions of emerin in mechanosignaling (Guilluy and Burridge, 2015), genome organization (Boyle et al., 2001) and autophagy (Deroyer et al., 2014) among other functions the division of different emerin activities between bocksbeutel and otefin could serve as a valuable tool to further study emerin functions.

All stocks were grown under standard condition at 25°C. Stocks used were apRed which expresses DsRed fused to a nuclear localization signal downstream of the apterous mesodermal enhancer (Richardson et al., 2007), bocks^DP01391 (21846; Bloomington Drosophila Stock Center), klar¹ (3256; Bloomington Drosophila Stock Center), ote^DB (5092; Bloomington Drosophila Stock Center), koi^EY03560 (20000; Bloomington Drosophila Stock Center), bocks deficiency Df(3R)Exel6153 (7632; Bloomington Drosophila Stock Center), klar deficiency Df(3L)BSC247 (9721, Bloomington Drosophila Stock Center), ote deficiency Df(2R)BSC337 (24361; Bloomington Drosophila Stock Center), koi deficiency Df(2R)Exel6050 (7532, Bloomington Drosophila Stock Center and UAS-klar.6Xmyc; derived from stock 25668; Bloomington Drosophila Stock Center) and ote^B279 (16189, Bloomington Drosophila Stock Center). Mutants were balanced and identified using CyO, Dfd-GMR-nvYFP and TM6b, Dfd-GMR-nvYFP. UAS-klar.6Xmyc was driven specifically in muscle using DMEf2-GAL4, apRed. The twist-GAL4, apRed and Dmef2-GAL4, apRed Drosophila lines were both used as controls. Third chromosome alleles have twist-GAL4, apRed on the second chromosome and second chromosome alleles have Dmef2-GAL4, apRed on the third chromosome to allow visualization of nuclei within the LT muscles during embryonic stages. Because there are slight variations between these two genotypes, each was used as a control. The twist-GAL4, apRed and Dmef2-GAL4, apRed Drosophila lines were made by recombining the apRed transgene and the specific GAL4 driver.

Embryos were collected at 25°C and then dechorionated by submersion in 50% bleach for 4 min. Embryos were then washed with water and then fixed in 50% formalin (HT501128; Sigma-Aldrich) diluted in 1:1 with heptane and placed on an orbital shaker that rotated at a rate of 250 rpm for 20 min. In all cases, embryos were devitellinized by vortexing in a 1:1 methanol:heptane solution. Embryos were stored in methanol at −20°C until immunostaining.

Larvae were dissected as previously described (Metzger et al., 2012; Louis et al., 2008) with minor modifications. Larvae were dissected in ice-cold 1,4-piperazinediethanesulfonic acid (PIPES) dissection buffer containing 100 mM PIPES (P6757; Sigma-Aldrich), 115 mM D-sucrose (BP220-1; Thermo Fisher Scientific), 5 mM trehalose (182550250; Acros Organics), 10 mM sodium bicarbonate (BP328-500; Thermo Fisher Scientific), 75 mM potassium chloride (P333-500; Thermo Fisher Scientific), 4 mM magnesium chloride (M1028; Sigma-Aldrich) and 1 mM ethylene glycol tetraacetic acid (28-071-G; Thermo Fisher Scientific) and then fixed with 10% formalin for 20 min.

Antibodies for embryo staining were used at the following final dilutions: rabbit anti-dsRed, 1:400 (632496; Clontech); rat anti-tropomyosin, 1:200 (ab50567; Abcam), and mouse anti-green fluorescent protein, 1:50 (GFP-G1; Developmental Studies Hybridoma Bank). Antibodies for larval staining were used at the following final dilutions: mouse anti-Klar, 1:25 (KLAR-C 9E10; Developmental Studies Hybridoma Bank), mouse anti-LamC, 1:20 (LC28.26, Developmental Studies Hybridoma Bank), mouse anti-Myc, 1:200 (9B11, Cell Signaling Technology). Conjugated fluorescent secondary antibodies used for embryo staining were Alexa Fluor 555 donkey anti-rabbit-IgG (1:200), Alexa Fluor 488 donkey anti-rat-IgG (1:200) and Alexa Fluor 647 donkey anti-mouse-IgG (1:200; all Life Technologies). Alexa Fluor donkey anti-mouse-IgG (1:200; Life Technologies), Acti-stain 555-conjugated phalloidin (1:400; PHDH1-A; Cytoskeleton) and Hoechst 33342 (1 μg/ml; H3570; Life Technologies) were used for larval staining. Embryos and larvae were mounted in ProLong Gold (P36930; Life Technologies) and imaged with an Apochromat 40×/1.4 numerical aperture (NA) objective with a 1.0× optical zoom for all embryo images on a Zeiss 700 LSM. Larvae were imaged using the same microscope and objective lens at 0.5× optical zoom for nuclear positioning analysis and 2.0× optical zoom for Klarsicht localization, Hoechst intensity analysis and Myc-tagged Klarsicht localization.

The position of nuclei was measured in stage 16 embryos, which is the latest stage before cuticle development blocks the ability to perform immunofluorescence microscopy. Embryos were staged primarily by gut morphology as previously described (Folker et al., 2012). Images acquired as described above were processed as maximum intensity projections of confocal z-stacks using ImageJ. The positioning of the nuclei was measured using the line function in ImageJ to determine the distance between either the dorsal end of the muscle and the nearest nucleus or the ventral end of the muscle and the nearest nucleus. All four LT muscles were measured in three or four hemisegments from each embryo. At least 20 embryos from at least two independent experiments were measured for each genotype, with the exception of experiments investigating mutants crossed to deficiencies, with each data point representing the average for all muscles measured within a single embryo. For experiments investigating mutants crossed to deficiencies, at least seven embryos from at least two independent experiments were measured. Statistical analysis was performed with Prism 4.0. Student's t-test was used to assess the statistical significance of differences in measurements between experimental genotypes and controls.

For qualitative nuclear-positioning phenotype analysis, embryos were scored on how nuclei positioned themselves within the first three LT muscles of three or four hemisegments in at least 20 embryos from at least two independent experiments, with the exception of experiments investigating mutants crossed to deficiencies. For experiments investigating mutants crossed to deficiencies at least seven embryos from at least two independent experiments were measured. LT4 was excluded for this analysis due to its variable muscle morphology. Nuclei were categorized as ‘separated (equal distribution)’ to indicate that nuclei were properly segregated into two distinct even clusters with a dorsal:ventral cluster area ratio >0.6 and <1.4, ‘separated (unequal distribution)’ to indicate that nuclei were separated into two distinct clusters that were uneven in size with a dorsal:ventral cluster area ratio of <0.6 or >1.4, ‘central’ to indicate that a nucleus or small cluster of nuclei was located in the middle of the myofiber that is not associated with either the dorsal or ventral cluster, ‘clustered’ to indicate that nuclei remain in a single cluster toward the ventral end of the myofiber, or ‘spread’ to indicate that nuclei are distributed along the myofiber with no distinct dorsal or ventral cluster. For the distinction of separated (equal distribution) and separated (unequal distribution), the areas of dorsal and ventral clusters were measured from each LT muscle using ImageJ. The nuclear distribution ratio was calculated by dividing the dorsal areas by the ventral areas. Statistical analysis was performed with Prism 4.0. Student's t-test was used to assess the statistical significance of differences in measurements between experimental genotypes and controls.

We measured nuclear position in larvae with our previously described method (Collins et al., 2017; Auld et al., 2018). First, the area and length of the muscle were measured. Next, the position and number of nuclei were calculated using the multipoint tool in ImageJ to place a point in the center of each nucleus. The position of each nucleus was used to calculate the actual internuclear distance. The maximal internuclear distance was then determined by taking the square root of the muscle area divided by the nuclear number. This value represents the distance between nuclei if their internuclear distance is fully maximized. The ratio between the actual internuclear distance and the maximal internuclear distance was then used to determine how evenly nuclei were distributed. This method normalizes the internuclear distance to both the nuclear count and the muscle area, which leads to a more representative means of comparison between muscles, larvae and genotypes. In addition, the distance of each nucleus from the lengthwise edge of the muscle was determined by measuring the shortest distance from the center of the nucleus to the nearest lengthwise edge of the muscle. A total of 24 ventral longitudinal (VL3) muscles were measured from at least six larvae with at least three VL3 muscles measured from each larva from at least two independent experiments. Statistical analysis was performed with Prism 4.0. Student's t-test was used to assess the statistical significance of differences in measurements between experimental genotypes and controls. The slope of the linear regression and R² values for the distance from muscle edge versus the average muscle width were determined using the linear regression function in Prism 4.0.

Nuclear Klarsicht localization was measured in VL3 muscles of L3 larvae. Z-stack maximum projection images that extended through the entire nucleus were analyzed. The fluorescence intensity of Klarsicht and Hoechst were measured for the nucleus and the cytoplasm. The ratio between the background-subtracted average nuclear Klarsicht and maximum Hoechst fluorescence intensity was then used to determine the degree of Klarsicht localization at the nucleus while normalizing to Hoechst intensity to control for any staining variation between experiments (Fig. S5B,C). Since the EDMD-linked genes tested have an effect on nuclear size (Fig. S5A), the Hoechst normalized average nuclear Klarsicht fluorescence intensity ratio was also normalized to nuclear size. The Hoechst and size normalized intensity ratios were also normalized to the intensity ratios of control larvae that were dissected and stained on the same day using the same materials. A total of at least 20 nuclei were measured from at least six larvae from at least two independent experiments.

Nuclear size, Hoechst integrated density, and mean and maximum fluorescence intensities were measured as previously described with minor modifications (Xiang et al., 2017; Wang et al., 2018). Briefly, individual nuclei in VL3 muscles were imaged as Z-stacks with 0.25 μm steps so as to image the entire nucleus. A low laser power was used to avoid saturation of the detectors, and imaging settings were kept constant throughout all nuclear size, Hoechst integrated density, and mean and maximum fluorescence intensity experiments. The nucleus was identified in ImageJ by converting the signals from the Lamin C fluorescence channel into a binary image, applying the ‘fill holes’ function and using the ‘analyze particles’ function with a size threshold set at >25 pixels resulting in selected regions of interest (ROI). The area of the ROI was recorded as the nuclear area. All slices from the Hoechst fluorescence channel were summed to create a projection of the nucleus, and a ROI from the Lamin C channel was selected in the Hoechst fluorescence channel using the restore selection function in ImageJ. The ROI in the Hoechst fluorescence channel was then measured for the mean and maximum Hoechst fluorescence intensities as well as the Hoechst integrated density.

Gene expression was quantified by RT-qPCR. RNA was extracted and isolated from five L3 larvae by crushing in an Eppendorf tube in 1 ml of TRIzol according to the manufacturer's instructions (15596026, Invitrogen). A DNase I (04716728001; Sigma-Aldrich) digest was performed on the isolated RNA at 37°C for 30 min according to manufacturer's instructions. DNase I was inactivated through the addition of EDTA to a final concentration of 8 mM and heating to 75°C for 10 min. RNA integrity and concentrations were determined using the NanoDrop2000 system (Thermo Fisher Scientific). The cDNA library was established by performing reverse transcription using the SuperScript VILO cDNA synthesis kit (11-754-050; Invitrogen), according to the manufacturer's protocol. Briefly, purified RNA was incubated with SuperScript III reverse transcriptase at 42°C for 2 h and then reactions were terminated at 85°C for 5 min. The resulting cDNA was used as the template for qPCR using an ABI 7500 Fast real-time PCR system (Applied Biosystems) and Power SYBR green PCR Master Mix (4367659, Applied Biosystems) for detection. For each genotype, biological and technical triplicates were performed. Gene transcript levels were quantified using gene-specific primers designed using FlyPrimerBank (Hu et al., 2013) and primers were validated according to Applied Biosystems' instructions. The primers used were: RP49 forward, 5′-GCCCAAGGGTATCGACAACA-3′; RP49 reverse, 5′-GCGCTTGTTCGATCCGTAAC-3′; GAPDH forward, 5′-TAAATTCGACTCGACTCACGGT-3′; GAPDH reverse, 5′-CTCCACCACATACTCGGCTC-3′; αTub84b forward, 5′-GATCGTGTCCTCGATTACCGC-3′; αTub84b reverse, 5′-GGGAAGTGAATACGTGGGTAGG-3′; bocks forward, 5′-AGGACCAGCAGCCTAGACG-3′; bocks reverse, 5′-TCAACTTCGCGTGTGTAAGATG-3′; klar forward, 5′-GCGTGGGACAACTACCAAGA-3′; klar reverse, 5′-AATTCCAAGAGACGCCGGG-3′; ote forward, 5′-GATTCTCTGTCCAATGCTGAGTT-3′; ote reverse, 5′-TAGAACCTTCCGGCTGCTATC-3′; and koi forward, 5′-CTGACCTCGGACTATTCGAGC-3′; koi reverse, 5′-GGTGAGAATCGACGTGACTGT-3′. To confirm the effective removal of contaminating DNA and specificity of the primers, experiments were also conducted with reactions lacking reverse transcriptase. The differences in gene expression were calculated using the ΔΔC_t method. Rp49, GAPDH and αTub84b were used as the reference genes for comparison to the gene of interest for ΔC_t values for each sample. Fold changes were expressed as 2^−ΔΔCt and plotted in Log₂ for graphical representation. Statistical analysis was performed with Prism 4.0. Student's t-test was used to assess the statistical significance of differences in ΔC_t measurements between experimental genotypes and controls.

The authors would like to thank all members of the Folker lab for the helpful discussions in completing and interpreting these experiments.