Background

The fruit fly Drosophila melanogaster possesses 152 muscle groups in the imago, as well as 54 in the larva (Bothe & Baylies, 2016). Each muscle is adapted to perform a specialized function, such as the extremely ordered indirect flight muscle (IFM) which is designed so as to enable the rapid oscillatory contractions needed to sustain flight. The marked phenotypic differences among these muscle types begs a molecular explanation. Alternative splicing is known to contribute to muscle type development and function with about 80% of sarcomere proteins being differentially expressed or spliced in Drosophila muscle (Nikonova et al., 2020). One such protein is Zasp52, a scaffolder, which contains numerous alternatively spliced exons generating a variety of isoforms.

All metazoan muscles share considerable structural homology (Steinmetz et al., 2012). They come in two types: striated, which appear striped given the presence of sarcomeres, and smooth, which do not possess sarcomeres. The sarcomere, a cylindrical structure, is the fundamental contractile unit of striated muscle and is bordered on either side by a dense circular structure called the Z-disc. Sarcomeres are arranged end-to-end in long arrays called myofibrils. Thin filaments composed of actin are anchored at and extend perpendicularly from the Z-disc into neighbouring sarcomeres, terminating at the H-zone (the region halfway between two Z-discs with no thin filaments). Thick filaments composed of muscle myosin are centred at the M-line (found at the centre of the H-zone) and interdigitate with thin filaments in a hexagonal lattice. The mechanism of sarcomere contraction is explained by the sliding filament theory in which thick filaments pull upon thin filaments in an ATP-dependent manner, bringing the Z-discs closer together. The simultaneous contraction of tandem sarcomeres comprising a myofibril is what leads to large-scale muscular contraction (Krans, 2010). The Z-disc is crucial for maintaining mechanical stability of the muscle and contains numerous proteins including Zasp52 which interacts with the thin filament cross-linker alpha-actinin (Liao et al., 2016) and actin (Liao et al., 2020) in Drosophila.

Zasp52, part of the three-membered Drosophila Alp/Enigma protein family along with Zasp66 and Zasp67, is crucial for maintaining structure at the Z-disc. It localizes to the Z-disc in all muscle types of Drosophila melanogaster (Katzemich et al., 2011) and is required for the assembly and maintenance of IFM sarcomeres (Katzemich et al., 2013; Liao et al., 2016). Developmentally, Zasp52 appears in the embryo first at myotendinous junctions and also at Z-bodies - clusters which later coalesce to form Z-discs (Katzemich et al., 2013). Zasp52 binds actin via its N-terminal PDZ domain plus a long C-terminal extension (Liao et al., 2020), and α-actinin via a region encompassing the PDZ domain plus a short C-terminal extension (Liao et al., 2016). In addition to the PDZ and Zasp motif (ZM) domain, Zasp52 contains four LIM domains which interact with the ZM domain resulting in auto-oligomerization; this mechanism is believed to drive radial growth of the Z-disc as well as pathological aggregate formation (González-Morales et al., 2019). Its primary human ortholog LDB3/ZASP is implicated in muscular dystrophies which are characterized by progressive adult-onset distal myofibrillar myopathy and sometimes cardiomyopathy (Griggs, et al., 2007).

Zasp52 is alternatively spliced into twenty-two different isoforms (according to FlyBase release FB2025_02, used throughout this paper), which leads to a variety of different domain and inter-domain combinations. The inter-domain linker regions connecting Zasp52’s structured domains (i.e., PDZ, ZM, and LIM domains) are less characterized but play an important role in protein function (Fisher & Schöck, 2022). Isoforms containing the first and second LIM domains contain a linker region between the two which can vary greatly in size. Encoded by several exons, this linker region may contain exon 15 which undergoes alternative 5’ splice site selection generating five exonic variants (15a-e) ranging from 130 nucleotides (15a) to 4429 nucleotides (15e) in length (Fig. 1A). The last 29 nucleotides of exon 15 contribute to the second LIM domain, but the remainder encodes the linker. When exon 15e is included, it constitutes the majority of the entire protein, encoding 1475 amino acids which total 160 kDa. In the canonical and longest isoform Zasp52-PF, the entire protein is 2194 amino acids long and the linker region between LIM1 and LIM2 is 1686 amino acids long. Exon 15e is referred to variously by other sources: as exon 9 in Flybase, exon 16c in Katzemich et al., 2011, and exon 18/19 in Benna et al., 2009; protein isoforms containing exon 15e were previously also referred to as Z(210) (Chechenova et al., 2013).

Figure 1

Though not much is known about this linker region between LIM1 and LIM2, its predominant expression in IFM and unusually large length prompted us to investigate it. We were curious to see how deleting only exon 15e while retaining all conserved domains would affect muscle structure and function. This linker was found to be highly disordered; intrinsically disordered regions (IDRs) have mostly been analyzed biochemically, with only limited data on their in vivo function in model organisms (Jensen et al., 2025). Flight ability was impaired and myofibrils displayed a variety of phenotypes which were all indicative of a loss of thin filament anchoring at the Z-disc. Most phenotypes exhibited an age-related increase in severity. Put together, these data demonstrate a crucial role of this large IDR in maintaining sarcomere integrity by stabilizing thin filaments at the Z-disc in IFM.

Results

exon 15e-containing isoforms are expressed in a specific spatiotemporal manner

Western blotting using a polyclonal antibody capable of detecting all Zasp52 splice variants (hereinafter referred to as the Zasp52 full-length antibody, Jani & Schöck, 2007) was used to analyze the presence of Zasp52 in IFM, ovaries, whole larvae, heads, and whole male abdomens (Fig. 1B). IFM contains the second most Zasp52 protein of any isoform relative to total protein content compared to the other tissues analyzed and is the only tissue that significantly contains exon 15e-containing isoforms relative to total Zasp52 content. This agrees with previous studies (Chechenova et al., 2013, Benna et al., 2009) which found no higher weight isoforms in larvae, and with transcriptomics data. Whole organism RNA-seq data from the ENCODE project (Duff et al., 2015) reveals very little expression of exon 15e in all developmental stages up until four days past the white prepupal stage (Fig. S1). To verify that this is indeed a phenomenon of exon 15e inclusion and not merely Zasp52 expression, one can look at the surrounding exons which do not follow the same pattern as exon 15e. To gain a more precise perspective on IFM expression of exon 15e, we turned to an RNA-seq dataset from Spletter et al., 2018 which looked at total RNA content of IFM at various developmental time points (Fig. 1C). Expression of all Zasp52 transcripts in IFM surges from 48 to 72 h after puparium formation (APF), with most of these transcripts containing exon 15e, as opposed to including exon15 a-d only or no exon 15 at all.

We used various software to predict the level of disorder across Zasp52-PF (Fig. S2A). All identified the linker region between LIM1 and LIM2 as an intrinsically disordered region (IDR) to various degrees, although some regions within it appear more ordered. The actin-binding motif identified by Ashour et al., 2023 within exon 15e is predicted to be more structured than the surrounding regions. No significant sequence similarity was found when aligning Zasp52’s linker region to human or other Drosophila melanogaster ALP/Enigma protein family members. We then compared the lengths of the linker regions between LIM1 and LIM2 in all insect species possessing a Zasp52 ortholog as defined by NCBI’s Eukaryotic Genome Annotation Pipeline (Fig. S2B). Drosophilids generally have a longer linker than their taxonomical neighbours. Mayflies, dragonflies, and damselflies, which use direct flight muscle, have notably short linkers. Since linear sequence often does not fully describe an IDR’s function, we analyzed linear clustering or mixing of different residue types using NARDINI (Cohan et al., 2022). The z-score matrix of the linker region between LIM1 and LIM2 in Zasp52-PF showed a strong clustering/segregation of polar residues (μ), as well as alanine (A) and proline (P), and hydrophobic residues (h) (Fig. S2C). The z-score matrix of exon 15e alone displayed much of the same clustering, except with additional strong clustering of acidic residues (-) with respect to other residue classes (Fig. S2C’). This implies that the additional exons in the linker region dilute or mask the strong clustering of acidic residues in exon 15e alone. We attempted to compare z-score matrices across various arthropod species but were unable to detect any meaningful correlation indicating that the binary patterns uncovered by NARDINI do not reveal taxonomical differences in exon 15e composition.

A CRISPR deletion of exon 15e displays flight defects

To further investigate the function of exon 15e, we generated a CRISPR deletion spanning five intronic nucleotides five prime to exon 15e up to 149 nucleotides before the 3’ end of exon 15e (Fig. 1A). In total, 4285 nucleotides totaling 156 kDa were deleted. Thus, our mutants would affect all isoforms containing exons 15b-e. The deletion endpoints were sequenced to confirm effective CRISPR knockout. To confirm that our deletion worked and spliced as intended, we performed Western blots using the Zasp52 full-length antibody to compare our deletion with a control and the MiMIC line MI00979 (Venken et al., 2011) which truncates everything C-terminal to and including exon 15e (Fig. 2A). The largest protein that could possibly be present in our mutant would be 136 kDa (Zasp52-PI). Our deletion was devoid of all higher molecular weight isoforms containing exon 15e which were present in the control as expected. Our deletion also contained several isoforms which were heavier than the highest molecular weight isoforms of MI00979 which were also not present in control, suggesting that these are novel products containing the C-terminal LIM domains. To confirm that this was indeed the case, i.e., that the three C-terminal LIM domains were retained, we raised an antibody against them and performed a Western blot (Fig. 2A’). Our CRISPR deletion retained these LIM domains as expected. We refer to this CRISPR mutant as ex15e.

Figure 2

We wanted to see how ex15e affected the locomotory abilities of the fly. Since larvae do not express exon 15e, we do not expect them to be affected by our deletion. Therefore, we performed a larval crawling assay as a negative control to confirm that ex15e larval muscles were unaffected (Fig. 2B). As predicted, control and ex15e larvae demonstrated no differences in crawling ability. Next, we wanted to see how flight would be affected, given that exon 15e is present primarily in the adult IFM. ex15e flies of all age categories had impaired flight ability compared to control with a severe deterioration at three weeks of age (Fig. 2C). To ensure that no secondary-site mutations were affecting our mutant phenotypes, we performed a simple binary flight assay using the heterozygous ex15e deletion over a Zasp52 deficiency (Fig. S3A).

This impairment in flight could stem from a variety of causes, given the multimodal functionality of Zasp52. We decided to first look at gross muscle morphology before zooming in on our investigation. Thorax bisections revealed no noticeable difference in the IFMs of ex15e and control flies (Fig. S3B, C). Since Zasp52 was known to be involved in integrin interactions (Jani & Schöck, 2007) and junctional components (Ashour et al., 2023), we inspected the myotendinous junctions of IFM, which are called modified terminal Z-discs (MTZ), in three-week-old flies (Fig. S3D-G’’). βPS integrin localization appeared unaffected in ex15e; this agrees with integrins being upstream of Zasp52 (Jani & Schöck, 2007). However, the MTZ appeared disrupted and malformed, lacking the organized structure seen in controls. Furthermore, we also noticed myofibril defects which prompted us to further investigate IFM at the myofibrillar level.

ex15e mutant IFM phenotypes: bending at the Z-disc and thin filament intrusion into the H-zone

IFM tissue was analyzed by confocal microscopy with several phenotypes being observed. First, myofibrils in ex15e mutants were often found to be bent at the Z-disc (Fig. 3A-C). In control flies, sarcomeres across different age groups typically displayed the same proportion of degree of bentness. In ex15e, the proportion of bent sarcomeres increased with age, as well as the degree of bending, compared to control flies. Force-induced bending results in a sarcomere whose inner curve is bowed, much like a hose with a kink. The bending in ex15e appears to be a result of the improper anchoring of thin filaments which leads to a shear type angle shift in the sarcomere at the Z-disc. Bending was found exclusively at the Z-disc and not the M-line in all muscle analyzed.

Figure 3

Another phenotype observed in ex15e was the presence of a dense band of actin staining at the H-zone, where normally there would be no actin staining (Fig. 3D-G). Control flies have very few sarcomeres with this phenotype at all ages; ex15e are similar to control at one day of age but begin to accumulate a large number of these abnormal sarcomeres with time. At three weeks of age, a large proportion of sarcomeres have actin at the H-zone. Usually, groups of myofibrils in the same area would exhibit this phenotype rather than these abnormal myofibrils being randomly dispersed among the normal ones. This dense band of actin suggested that thin filaments were overlapping at the M-line. This could be a result of hypercontraction, where thin filaments are brought so close together that they overlap. However, the protocol used for these preparations includes a relaxing solution containing ATP, which induces cross-bridge detachment resulting in the adoption of a relaxed conformation. To see if these abnormal sarcomeres with actin in the H-zone were indeed contracted, we measured their length and compared them to normal sarcomeres from the same sample in three-week-old flies (Fig. 3G). Sarcomeres with actin in the H-zone had a smaller length than those without indicating that they were indeed contracted. Muscle decontraction relies on the elastic nature of the sarcomere to restore it to a relaxed conformation once myosin heads detach from the thin filaments. This requires stable Z-discs and myofilament attachments, the lack of which may be the cause of a permanent hypercontracted state in ex15e.

To gain a better understanding of this phenomenon, we performed transmission electron microscopy (TEM) of myofibrils and noticed phenotypes that concurred with our immunofluorescence confocal images (Fig. 4). Z-discs were highly disrupted and contained many holes and broken regions. M-lines were also affected in a similar manner. Gaps between filaments were abundant and in general appeared more disorganized than control. Filaments were often frayed at the edges of the sarcomeres.

Figure 4

Most ex15e defects are restored by rescue constructs

To determine if the phenotypes are truly due to the IDR or are caused by an altered isoform composition, we wanted to see if the phenotypes of ex15e could be rescued. First, we expressed either UAS-Zasp52-PF or UAS-Zasp52-PR using a UH3-GAL4 driver in a homozygous ex15e background and examined three-week-old flies (Fig. 5B-H). Zasp52-PF is the canonical full-length Zasp52 isoform which contains exon 15e, while Zasp52-PR contains all structured domains but not exon 15e (Fig. 5A). The degree of bending in Zasp52-PR rescue sarcomeres was statistically equivalent to that of ex15e mutants, while that of Zasp52-PF rescues was equivalent to control flies, indicating that Zasp52-PF but not Zasp52-PR can rescue this phenotype (Fig. 5F). The proportion of H-zone actin (i.e., hypercontracted) sarcomeres was not significantly different between ex15e and Zasp52-PR rescues as expected. Interestingly, no hypercontracted sarcomeres were observed upon re-expression of Zasp52-PF again indicating a successful rescue (Fig. 5G). Finally, we tested flight ability and observed a significant increase in flight strength in Zasp52-PF rescues compared to Zasp52-PR rescues, with the latter being not significantly different from ex15e (Fig. 5H). Zasp52-PF rescues were still weaker fliers than control flies suggesting a partial rescue of flight ability.

Figure 5

ex15e and actin genetically interact

We next wanted to test for a genetic interaction between exon 15e and Act88F (the actin isoform present in IFM) by analyzing heterozygous ex15e, heterozygous Act88F^KM88 (an amorphic mutant of Act88F), and a di-heterozygous combination of both (ex15e/+; Act88F^KM88/+) (Fig. 6). ex15e/+ resembled control myofibrils and Act88F^KM88/+ has mild myofibrillar defects. ex15e/+; Act88F^KM88/+ exhibited supra-additive effects where we often noticed a novel phenotype of very large H-zones. A ratio of twice the thin filament length to the total sarcomere length was computed for all genotypes. Lower values represented sarcomeres whose thin filaments occupied a smaller proportion of the sarcomere and consequently had larger H-zones, whereas a value of one represented the absence of an H-zone. This ratio was not significantly different between control, ex15e/+, and Act88F^KM88/+, but was considerably smaller in di-heterozygotes (Fig. 6E), indicating a strong genetic interaction between Act88F and exon 15e.

Figure 6

FRAP reveals defects in ex15e protein dynamics

Given the differences we have thus observed in our mutants, as well as the great length of the IDR, we decided to investigate Zasp52 kinetics in ex15e sarcomeres by performing Fluorescence Recovery After Photobleaching (FRAP) (Fig. 7A, C). We expressed GFP-tagged Zasp52-PR (see Fig. 5A) which lacks exon 15e in either a wild-type background as our control (UH3>GFPZasp52-PR), or in an ex15e background (ex15e; UH3>GFPZasp52-PR). We demonstrate in Fig. 5 that Zasp52-PR cannot rescue ex15e. One-day-old IFM Z-discs were photobleached and recovery of fluorescence was monitored. Fluorescence recovery in the ex15e background was to a significantly higher level indicating a larger mobile fraction (Fig. 7B). A greater mobile fraction in our mutant background signifies that exon 15e is required to hold Zasp52 in place at the Z-disc, either by fastening itself, or keeping other molecules in place thereby contributing to overall Z-disc stability.

Figure 7

Most ex15e defects are rescued by immobilization

On a different note, we noticed that virtually all the ex15e phenotypes worsened with age, which suggests that use might contribute to deterioration. We therefore decided to prevent flies from using their IFMs by immobilizing them (Fig. 8). Flies were reared until three weeks of age in between glass plates separated by 1.5 mm spacers which prevented them from flying or even attempting to fly. The degree of bending in immobilized flies was equivalent to that of control flies (Fig. 8D). The proportion of hypercontracted sarcomeres was also not significantly different between immobilized and control flies (Fig. 8E). Intriguingly, even flight ability was restored in immobilized flies compared to age-matched controls with no significant difference between the two (Fig. 8F). These results suggest successful rescue by preventing IFM use via immobilization.

Figure 8

Discussion

Traditionally, IDRs were associated with the concept of amorphousness, finding themselves in transient structures such as condensates and membraneless organelles. As our understanding of disordered proteins has changed, their role in more disparate contexts has become increasingly appreciated, revealing them as modular hubs, flexible scaffolds, and more (Cortese et al., 2008). Furthermore, their role in disease has spurned great interest in their myriad functions. In this study, we demonstrate the role of an exceptionally large disordered region, Zasp52’s exon 15e, in maintaining stability and integrity in a highly rigid and force-bearing structure – the Z-disc.

First, we showed that the expression of exon 15e is spatiotemporally regulated. Being confined to the adult IFM suggests that exon 15e’s function is uniquely required for the rapid oscillatory contractions in indirect flight. Supporting this is the fact that the linker region containing exon 15e is generally much longer in Drosophilids, and rather short in insects using direct flight. Although this linker region is predicted by multiple software programs to be disordered, within it are regions of greater order, one of which corresponds to a putative actin-binding motif (Ashour et al., 2023). This arrangement exists in other proteins such as Cno in Drosophila, where higher-ordered actin-binding regions are embedded in a large IDR (Jensen et al., 2025). Furthermore, a synthetic enhancement between ex15e and Act88F^KM88 heterozygotes indicates that exon 15e and actin genetically interact. There may be other important short linear motifs (SLiMs) contained within the linker region which serve as interacting sites with other sarcomere proteins. Indeed, multiple sequence alignment of the linker regions of various insects reveals several conserved motifs of unknown function. The presence of an actin-binding motif that may strengthen thin filament anchoring to the Z-disc would be in line with the phenotypes we observed in ex15e. Bending at the Z-disc is indicative of a loss of thin filament anchoring and stability at the Z-disc, allowing them to shift out of alignment. Normal sarcomeres bent due to force would exhibit a kinked-hose appearance, where the inner edge is sharply bent and the outer one bowed. This was occasionally observed in controls, whereas ex15e bends were always shifted uniformly. Actin in the H-zone is also indicative of loss of stable thin-filament anchoring. During preparation, IFMs are incubated overnight in skinning solution during which they enter a state of rigor. After washing them with an ATP-containing relaxing solution, the myosin heads detach from the thin filaments and the supple nature of the sarcomere which provides it with elastic recoil, causes it to resume a relaxed conformation. However, when thin filaments are weakly anchored at the Z-discs, the natural tension contained within the sarcomere is reduced and thus it will not “bounce back” into a relaxed conformation, resulting in a hypercontracted state which manifests as actin continuing to reside in the H-zone. Cheerio, a Drosophila filamin, was shown to stably attach actin thin filaments at the Z-disc and induces a similar actin incorporation into the H-zone when mutated (González-Morales et al., 2017). Thus, we also demonstrate the importance of Z-disc integrity in allowing sarcomeres to return to their relaxed state in rapidly contracting asynchronous IFM. Accordingly, synchronous muscles do not have such regularly structured Z-discs, and do not require Zasp52’s exon 15e. Indeed, exon 15e is absent in larval and other muscle types. We therefore propose that exon 15e is uniquely required to stabilize rapidly contracting asynchronous IFM.

The orthologous human ALP/Enigma family consists of seven members, with LDB3 (ZASP) considered the canonical Zasp52 ortholog. Zasp52 is more precisely a combined ALP/Enigma member; its first LIM domain is sequentially homologous to the single LIM domain of human ALP family members, and its last three LIM domains are homologous to the three Enigma family LIM domains. Though these human proteins do contain disordered linkers connecting their structured domains, they are much shorter than in Drosophila and are not sequentially homologous. The length of Drosophila’s linker was likely evolutionarily driven to be large, as we see insects using direct flight to have much shorter linkers. Several pathological mutations have been identified in the linker region between LDB3’s PDZ and LIM domains (Vatta et al., 2003), some of which are now known to lie in the actin-binding Zasp motif (ZM) (Lin et al., 2014). Interestingly, LDB3 has three skeletal muscle-specific splice isoforms which are differentially expressed and variably retain/omit exons 9 through 11 of a linker sequence (Lin et al., 2014). The prenatal long (L) isoform contains exons 10 and 11, post-natal long (LΔex10) has only exon 11, and short (S) has only 9. Post-natal LDB3-LΔex10 contains an actin-binding region in its linker region whose function is abolished by the inclusion of exon 10 in LDB3-L. Thus, human Zasp protein linkers also function in actin binding, regulated by developmentally differential expression. Linker regions in several human ALP/Enigma proteins were found to interact with both α-actinin and actin using AlphaFold (Healy & Collins, 2023). The mouse ortholog Cypher also contains tissue-specific splice variants which variably include or omit linker sequences (Huang et al., 2003). Like Zasp52’s exon 15e, Cypher is not required for sarcomerogenesis or Z-disc assembly but only maintenance (Zhou et al., 2001).

Our genetic rescues support our hypothesis of exon 15e involvement in stabilizing thin filaments and Z-discs. The rescues also demonstrate that the conserved domains (PDZ, ZM, LIM1-4) are not sufficient for proper Z-disc and thin filament stability in IFM, because Zasp52-PR could not rescue any of the defects. Interestingly, these phenotypes, as well as the flight defects we observed, worsened with age, with defects in one-day-old flies being very minor. This implies that exon 15e is not required developmentally, but rather for the maintenance of sarcomeres which naturally deteriorate with use. We originally hypothesized that since IDRs often lead to protein aggregation and phase separation, exon 15e might drive the formation of Z-bodies, primitive Z-discs which form through the coalescence of diffuse Z-disc proteins including Zasp52. However, given the lack of defects in early Z-discs and the fact that exon 15e expression does not temporally coincide with Z-body formation, we reject this hypothesis. The role of exon 15e in maintenance is supported by immobilization rescues, in which flies were prevented from flying. We opted for a low-ceiling apparatus which we believe would not only prevent flies from flying but discourage them from futilely attempting to utilize their IFM at all. Surprisingly, immobilization restored flight ability to wild type levels. Interestingly, human zaspopathies often manifest late in life, with one study citing onset at 44 to 73 years of age (Selcen & Engel, 2005), much like what we observed in flies. Our data suggest that avoiding strenuous exercise might delay the onset of zaspopathies. Given the maintenance-specific role of exon 15e, but also that as a whole, Zasp52 is indispensable for pupal myofibril development and not just maintenance (Katzemich et al., 2013), Zasp52 appears to have a multifaceted role more complex than previously considered.

Although one-day-old ex15e myofibrils are morphologically indistinguishable from controls, we were already able to detect abnormalities in Zasp52 dynamics using FRAP. The presence of recovery suggests that Zasp52 is not immobile as might be assumed for an anchoring protein and does indeed possess kinetics on a sub-minute timeframe. This is supported by significant transcription of the gene, including exon 15e-containing isoforms, into adulthood (Fig. S1). The mobile fraction of Zasp52-PR was significantly increased in a system entirely lacking exon 15e (the PR isoform also lacks exon 15e) indicating a role of exon 15e in retaining Zasp52 at the Z-disc, whether by direct interaction or not. The FRAP data suggest that Zasp52’s IDR stabilizes the Z-disc, in line with the electron microscopy data. It suggests that Z-discs can be described as highly solid aggregates in part stabilized by IDRs, on the other end of the spectrum of the transient condensates initially described, thereby further expanding the importance of IDRs for establishing subcellular structures.

Overall, Zasp52 and its isoforms represent a complex toolbox of multifunctional proteins. There has been evidence suggesting that Zasp52 is not required for initial sarcomere assembly in embryos (Rui et al., 2010), important for Z-disc maintenance overall (Rui et al., 2010, Jani & Schöck, 2007, Katzemich et al., 2013), and required for pupal myofibril assembly (Katzemich et al., 2013). The complete picture of Zasp52’s varied function is still unclear; however, we have shown that a specific disordered region is required for maintaining IFM and stabilizing the rigidity of Z-discs. The long linker might not only contain protein-interacting regions but also provide the spacing needed for the neighboring structured domains to properly interact with their partners and perform their roles. We have seen in our genetic interaction experiment an inability of thin filaments to properly elongate when Act88F is partially depleted, which further implicates exon 15e not only in Act88F structure but dynamics. Most encouraging is the complete rescue of flight ability through immobilization, which may provide avenues for delaying the onset of human zaspopathies through reducing strenuous exercise.

Materials and Methods

Fly strains and husbandry

Flies were housed at room temperature (22°C) in standard glass vials and food unless indicated otherwise. Immobilized flies were reared in between 3.25×3.25 inch glass plates separated by 1.5 mm spacers; standard food was inserted into the gap. ex15e flies were created by Genetivision using CRISPR-mediated homology-directed repair. The gRNA sequences used were TTTGAACTTGAACTCTAATGTGG for the downstream target and CAGCCTCTCCACCTTTCCGATGG for upstream. The deletion was replaced with a cassette containing a 3xP3DsRed marker flanked by attP landing sites. It was validated by sequencing at the cut sites. The following genotypes were used: ex15e, w¹¹¹⁸ as control in all experiments, Df(2R)BSC427, Act88F^KM88, UAS-GFPZasp52-PR, UH3-Gal4, UAS-Zasp52-PF, UAS-Zasp52-PR, Zasp52^MI00979. The UAS constructs consist of the respective Zasp52 isoform with an N-terminal ALFA-tag and C-terminal Spot tag. UAS constructs were made by Genscript and injected into w¹¹¹⁸ by GenomeProlab.

Western blotting

Extracts were prepared by homogenizing ten flies/fly parts in 100 μL RIPA buffer with protease inhibitor plus an equal amount of 2× Laemmli buffer with β-mercaptoethanol. Samples were boiled at 95°C for five minutes and centrifuged at 12,000×rcf for 5 minutes. Supernatant was used for electrophoresis according to standard protocol using a 4-15% BioRad TGX Stain-free gel. Stain-free gels were imaged for total protein content on a BioRad ChemiDoc before protein was transferred onto a nitrocellulose membrane. Standard Western blotting procedures were followed. Primary antibodies used were rabbit anti-Zasp52 full-length (Jani & Schöck, 2007), and a rabbit antibody against the three C-terminal LIM domains of Zasp52 (FEKYLAPTCSKCAGKIKGDCLNAIGKHFHPECFTCGQCGKIFGNRPFFLEDGN AYCEADWNELFTTKCFACGFPVEAGDRWVEALNHNYHSQCFNCTFCKQNLEGQSFYNKGGRPF CKNHAR) made by Genscript; secondary antibody was goat anti-rabbit IgG (H+L) HRP (Invitrogen).

Locomotory assays

For flight assays, flies were reared in vials (unless immobilized) until the desired age, with vials flipped whenever pupae began to appear. Flight assays were conducted in a manner similar to that described in Babcock & Ganetzky, 2014. A tube 22 cm long and 3.2 cm in diameter was placed in a funnel; the bottom of the tube rested 4 cm above the bottom spout of the funnel. The funnel rested in a 60 cm long tube which was 13.5 cm in diameter and covered with TAD Shake & Spray (T-141) on the inside. The bottom of the funnel was 12 cm below the top of the larger tube. The larger tube was divided vertically into seven 8.5 cm segments, with segment 1 being at the top, and segment 7 being at the bottom. A dish containing water was placed below the tube and was considered segment 8. Vials containing flies were dropped down the narrow tube so that they would be ejected into the larger tube and land somewhere along its length. Flies naturally fly upwards and those with stronger flight would land higher up the tube. Binary flight assays were conducted by releasing flies one at a time and recording whether they could generate any upwards lift. They were raised in the same conditions as above.

For larval crawling assays, third instar larvae of the desired genotype were collected from bottles and assayed according to Post & Paululat, 2018. In brief, videos of larvae on a slightly moistened petri dish were taken. Time segments where larvae were crawling in a straight line continuously for at least ten seconds were analyzed – contractions per distance were measured and recorded.

Muscle preparation and immunofluorescence

Primary antibodies used were rabbit anti-Zasp52 full-length (Jani & Schöck, 2007), rabbit anti-LIM234 (this study), and mouse anti-βPS-integrin (CF.6G11; obtained from Developmental Studies Hybridoma Bank, Brower et al., 1984) at a concentration of 1:200 for all. Secondary antibodies used were Alexa Fluor 647 goat anti-rat (Life Technologies) and Alexa Fluor 488 goat anti-rabbit (Invitrogen) at a concentration of 1:400 for all.

For IFM tissue, a protocol similar to Xiao et al., 2017 was followed (refer to for recipes). Briefly, thoraces with bisected and incubated in Relaxing-Glycerol solution at -20°C overnight. IFM were isolated and fixed in 4% paraformaldehyde for 15 minutes. Fixed IFM were incubated in primary antibodies at 4°C overnight. IFM were incubated in Acti-stain 555 fluorescent phalloidin (Cytoskeleton Inc) and secondary antibodies, at room temperature for one hour. Samples were mounted in Mowiol mounting medium.

For thorax hemisections including MTZ, a scalpel was used to sagittally bisect thoraces. Samples were fixed in 4% paraformaldehyde for 15 minutes and incubated in primary antibodies for one hour at room temperature, and then Acti-stain 555 fluorescent phalloidin and secondary antibodies for one hour at room temperature. Samples were mounted in Mowiol mounting medium using three 0.12 mm spacers.

For TEM, bisected thoraces were treated with 5 mM MOPS, 150 mM KCl, 5 mM EGTA, 5 mM MgCl₂, 5 mM ATP, 1% Triton X-100, and protease inhibitor for two hours. then the same solution with 50% glycerol instead of Triton X-100 overnight at 4°C. This was repeated at -20°C. Samples were washed in 5 mM MOPS, 40 mM KCl, 5 mM EGTA, 5 mM MgCl₂, 5 mM NaN₃, protease inhibitor, and then fixed in 3% glutaraldehyde and 0.2% tannic acid for two hours. Samples were washed in 20 mM MOPS, 5 mM EGTA, 5 mM MgCl₂, 5 mM NaN₃, and then in 100 mM sodium phosphate buffer, 5 mM MgCl₂, 5 mM NaN₃. Secondary fixation was performed in 1% osmium tetroxide in the previous sodium phosphate buffer for one hour. Samples were stained in 2% uranyl acetate, washed with water, and underwent a serial dehydration in ethanol. Samples were washed with propylene oxide and serially infiltrated with Mollenhauer Epon-Araldite. After curing in Beem capsules, 70 nm sections were cut and stained in 4% uranyl acetate followed by SATO lead stain.

Fluorescence Recovery After Photobleaching

One-day-old female flies were anaesthetized and their abdomens, wings, and legs removed. Carcasses were pinned down by the head on a silicone dish and immersed in hemolymph-like 3 (HL3) saline (Stewart et al., 1994). Thoraces were sagittally bisected, washed thrice in HL3 saline, and mounted with a 0.12 mm spacer. Imaging was performed immediately using a HC PL APO 63×/1.40 OIL CS2 objective on a Leica SP8 point-scanning confocal microscope. Five pre-bleach frames (1.29 seconds each) were recorded before bleaching with a 488-nm laser to achieve a bleach depth averaging 76%±7% (SD). Seventy-five post-bleach frames (1.29 seconds each) were then acquired. ROIs (1.25×3 μm) were placed over individual Z-discs and those with drift were discarded. For each fly, three Z-discs on different myofibrils were bleached, and intensities of adjacent unbleached Z-discs and unbleached background regions were recorded. Fluorescence was normalized as (bleached–background)/(unbleached– background), averaged across the three ROIs, and scaled such that the post-bleach frame equaled 0 and the pre-bleach mean equaled 100. Recovery curves were presented as mean±SEM. Each fly’s curve was fitted to a one-phase exponential association model to obtain the plateau (mobile fraction). Fixed samples were prepared in the same manner except with an additional ten-minute fixation in 4% paraformaldehyde in HL3 saline.

Microscopy

Confocal images were acquired using a HC PL APO 63×/1.40 OIL CS2 objective on a Leica SP8 point-scanning confocal microscope. The position along the z-axis was adjusted until perfectly in the centre of the myofibril, i.e., the widest point. Polarized light images were acquired using a 5X N-ACHROPLAN, NA=0.13 objective on a Zeiss Axio Observer microscope. TEM images were acquired using an FEI Tecnai G2 Spirit Twin 120 kV Cryo-TEM using an AMT NanoSprint15 MK2 CMOS camera.

Computational analyses

Schematics were illustrated with IBS 2.0 (Xie et al., 2022). All protein domains were defined per SMART (Letunic & Bork, 2018). Graphing and statistical analysis was performed using GraphPad Prism 10. Statistical analyses were performed in GraphPad Prism 10, where ns=p≥0.05, *=p<0.05; **=p<0.01, and ****=p<0.0001, and calculated to four decimal places. Equivalence of bend angles was assessed using the Two One-Sided Test (TOST) procedure. Model sarcomeres with bend angles ranging from 180° (straight) to 95° and blurring were generated. Each was hand measured ten times and the average absolute measurement error (4.5°) was used as the equivalence margin for all TOST analyses.

Minus strand RNA-seq bigWig files from the ENCODE project (Duff et al., 2015) with the following accessions were retrieved and loaded in IGV 2.19.4 (Robinson et al., 2011): ENCFF289ZZT, ENCFF982OGI, ENCFF430HWM, ENCFF716DNU, ENCFF829OGQ, ENCFF936YJM, ENCFF987YVI, ENCFF033MIY, ENCFF845HII, ENCFF021VXY, ENCFF792FNT. Signal tracks were aligned to the dm3 assembly and the data range was adjusted uniformly for all tracks.

SRA files with the following accession numbers were downloaded from the Bioproject PRJNA419412 (Spletter et al., 2018): SRR6314253, SRR6314256, SRR6314275, SRR6314259, SRR6314277, SRR6314262, SRR6314273. Salmon 1.10.0 (Patro et al., 2017) was used to quantify expression of the following transcripts which correspond to identified Zasp52 isoforms: FBtr0392902, FBtr0329914, FBtr0329912, FBtr0100387, FBtr0111048, FBtr0087315, FBtr0329913, FBtr0310086, FBtr0329916, FBtr0329915, FBtr0310088, FBtr0310087, FBtr0392901, FBtr0329911, FBtr0342711, FBtr0100388, FBtr0302163, FBtr0301313, FBtr0479884, FBtr0329909, FBtr0392918, FBtr0329910. The first six were classified as containing only exon 15a, the next six containing exons 15b-d, the next four containing exon 15e, and the last six containing no exon 15.

ImageJ was used to evaluate bending manually using the angle tool, and sarcomere length using the line measuring tool. For measuring patchiness, Niblack auto local thresholding of the actin channel using a radius of 15 units was performed, and outliers removed. The area of background within the myofibrils was divided by the area of the foreground thresholded selection to compute a patchiness metric.

For disorder prediction, the entire Zasp52-PF sequence was entered into AIUPred v0.1, DISOPRED3 (as two sequences, the first covering the first 1200 amino acids and the second, the remaining), or metapredict online v3.0. Raw values for each residue were obtained and smoothed using a Savitzky-Golay filter with fourth order smoothing and 11 neighbours on each side. AlphaFold v2.0 was used to obtain pLDDT scores for the alpha carbon of each residue. pLDDT scores were min-max normalized and smoothed in the same manner as above.

Zasp52 orthologs were obtained for members of the class Insecta using NCBI’s Eukaryotic Genome Annotation Pipeline. Canonical protein sequences for each were downloaded and those which did not contain the classical PDZ and four LIM domains were removed. The number of residues between the first and second LIM domains was calculated. A phylogenetic tree of these species was created using NCBI’s Common Tree (Schoch et al., 2020) and a heatmap created using iTOL v.6 (Letunic & Bork, 2024).

NARDINI (Cohan et al., 2022) was used for the sequences indicated to generate a z-score matrix, with 100,000 scrambled sequence variants.

Figure supplements

Supplementary Figure 1

Supplementary Figure 2

Supplementary Figure 3

Acknowledgements

This work was funded by the Canadian Institutes of Health Research project grant PJT-178114. Confocal images were collected in the McGill University Advanced BioImaging Facility (ABIF). Electron microscopy images were collected and analyzed in the McGill University Facility for Electron Microscopy Research (FEMR).

Additional information

Funding

Canadian Institutes of Health Research (CIHR) (PJT-178114)

• Frieder Schöck