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

(A) A schematic of the Zasp52 locus. Exons are drawn to scale, introns are not. Coding exons are orange and untranslated regions in grey. Alternative start sites are also indicated. Below in green are mappings of structured domains according to SMART (Letunic & Bork, 2018); and also Katzemich et al, 2011 which defined LIM1 and LIM2 variants. The deletion in ex15e mutants is indicated in red. (B) Western blot using five different tissue extracts from control flies probed with a Zasp52 full-length antibody. Relative amounts of Zasp52 content as measured in each lane, over the total protein content as measured using the TGX Stain-free gel method, are quantified below for each tissue. Exon 15e content as measured using the protein content above the dotted blue line for each lane, over Zasp52 content is quantified below that. Molecular weights appear higher, perhaps due to retardation caused by non-specific interactions with the gel matrix, as also seen in Watts et al., 2017, Fig. S1. (C) RNA-seq data from Spletter et al., 2018 of IFM tissue across various timepoints after puparium formation; the last timepoint is of adults one day after eclosion. Gene expression levels in transcripts per million (TPM) are shown for Zasp52 isoforms classified according to their splicing of exon 15.

A CRISPR deletion of exon 15e displays flight defects

Western blots using either the Zasp52 full-length antibody (A) or antibody against LIM2-LIM4 (A’) of either w1118 (ctrl hereinafter), ex15e, or the MiMIC line MI00979 whole flies. Relative amounts of Zasp52 content as measured in each lane, over the total protein content, are quantified below for each genotype. (B) Larval crawling assay of either ctrl (n=21) or ex15e (n=21) third instar larvae showing no significant difference between the two (Unpaired t-test with Welch’s correction; p=0.6006). (C) Flight assays of ctrl vs ex15e flies of three different age categories: one-day-old (ctrl n=138 flies, ex15e n=123), five-days-old (ctrl n=125, ex15e n=119), and three-weeks-old (ctrl n=86, ex15e n=81). The y-axis indicates flight strength: flies were released into a tube and those that landed in the top segment (y=1) had the strongest flight strength while those that landed in the bottom dish (y=8) had the weakest. The x-axis indicates the proportion of flies that landed in that segment. Red arrows indicate the average position landed in control flies, yellow arrows are for ex15e flies. Difference in flight ability was statistically significant for all age categories (Fisher’s exact test; p<0.0001).

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

(A-C) Analysis of bent sarcomere phenotype. Actin is visualized by phalloidin in green. Control myofibrils (A) appear straight while ex15e (B) myofibrils contain a range of bending (representative three-week-old myofibrils shown). Red arrow indicates a representative sarcomere with a bend of <145°, yellow indicates 145-165°, and white >165°. Bend angles were statistically equivalent (≡) between ctrl and ex15e in one-day-olds and five-day-olds, but not equivalent (ne) in three-week-olds, where ex15e displayed greater bentness (C). Sample sizes: one-day-old (ctrl n=1242 sarcomeres, ex15e n=1851), five-days-old (ctrl n=1192, ex15e n=946), and three-weeks-old (ctrl n=1768, ex15e n=1125) (Two One-Sided Test (TOST) with ±4.5° equivalence margin, α=0.05; 1d.o. tL=12.95, tU=-12.07; 5d.o. tL=6.03, tU=-8.01; 3w.o. tL=-6.34, tU=-20.63). (D-F) Sarcomeres also exhibited actin in the H-zone phenotype (representative three-week-old myofibrils shown). Control (D) myofibrils display normal H-zones without actin with M-lines/H-zones indicated by blue arrows. ex15e (E) myofibrils display actin accumulation at the H-zone; arrows indicate same as above. Actin is visualized by phalloidin in green and Z-discs are identified by the full-length Zasp52 antibody in red. These sarcomeres with H-zone actin, which appear hypercontracted, are significantly overrepresented in ex15e compared to control across all age categories (F). Sample size were as follows: one-day-old (ctrl n=3255 sarcomeres, ex15e n=4056), five-days-old (ctrl n=3221, ex15e n=2677), and three-weeks-old (ctrl n=5394, ex15e n=4720 (Fisher’s exact test; p<0.0001). To confirm that H-zone actin sarcomeres were indeed contracted, the Z-to Z-disc length was measured in ex15e three-week-old flies of both sarcomeres with a bare H-zone (H-zone normal, n=89 sarcomeres) and those with actin in the H-zone (H-zone actin, n=117) (G). H-zone actin sarcomeres were significantly smaller (Unpaired t-test with Welch’s correction; p<0.0001).

TEM reveals defects in mutant sarcomere ultrastructure

TEM images reveal similar phenotypes as immunofluorescent ones (three-week-old shown). Control (A) sarcomeres are very structured and even. Longitudinal sections of ex15e (B,C) display broken Z-discs (red arrow), broken M-lines (yellow), bending at the Z-discs (green), splits in filaments (purple), and fraying peripheral filaments (blue). Transverse sections of ex15e (D) also display some of these abnormalities.

Certain ex15e defects are restored by rescue constructs

Zasp52-PF or Zasp52-PR was heterozygously overexpressed with a UH3-GAL4 driver over a homozygous ex15e background in a rescue attempt. In this figure, all myofibrils had actin visualized with phalloidin in green, all experiments were done at three weeks of age, and ctrl and ex15e data are same as previous. (A) Schematic showing Zasp52-PF which is the canonical full-length isoform, and Zasp52-PR which contains all structured domains but lacks exon 15e. PDZ domain (blue), ZM (red), LIM domains (green). (B-E) Myofibrils display bending, Z-disc disruption, and overall myofibrillar disorganization in Zasp52-PR overexpressions (D), while Zasp52-PF overexpression (E) appears similar to ctrl (B). Bend angles were equivalent (≡) between ex15e and ex15e; UH3>Zasp52-PR, and between ctrl and ex15e; UH3>Zasp52-PF (F). Sample sizes: ex15e; UH3>Zasp52-PR n=2049 sarcomeres, ex15e; UH3>Zasp52-PF n=2221. (Two One-Sided Test (TOST) with ±4.5° equivalence margin, α=0.05; ex15e-ex15e; UH3>Zasp52-PR tL=1.92, tU=-10.16; ctrl-ex15e; UH3>Zasp52-PF tL=17.65, tU=-8.51). The proportion of sarcomeres that were contracted (i.e., actin in H-zone) was not significantly different between ex15e and ex15e-ex15e; UH3>Zasp52-PR, whereas no such sarcomeres were found in ex15e; UH3>Zasp52-PF (G). Sample sizes: ex15e; UH3>Zasp52-PR n=2843 sarcomeres, ex15e; UH3>Zasp52-PF n=2480. (Fisher’s exact test p=0.0521). (H) Flight ability between ex15e; UH3>Zasp52-PR (n=66) and ex15e; UH3>Zasp52-PF (n=85) was significantly improved, with ex15e; UH3>Zasp52-PR showing no significant difference from ex15e but ex15e; UH3>Zasp52-PF showing significant difference from ctrl. The y-axis indicates flight strength: flies were released into a tube and those that landed in the top segment (y=1) had the strongest flight strength while those that landed in the bottom dish (y=8) had the weakest. The x-axis indicates the proportion of flies that landed in that segment. Red arrows indicate the average position landed in control flies, yellow are ex15e, green are ex15e; UH3>Zasp52-PR, and blue are ex15e; UH3>Zasp52-PF. (Fisher’s exact test; ex15e; UH3>Zasp52-PR-ex15e; UH3>Zasp52-PF p<0.0001; ex15e-ex15e; UH3>Zasp52-PR p=0.2668; ctrl-ex15e; UH3>Zasp52-PF p=0.0129).

Large H-zones demonstrate a genetic interaction between exon 15e and Act88F

(A-D) Myofibrils of one-day-old flies of various genotypes. Actin was visualized with phalloidin in green. Control and heterozygous ex15e myofibrils appear indifferentiable; heterozygous Act88FKM88 are slightly disrupted with fraying and slight narrowing. Representative ex15e/+; Act88FKM88/+ myofibrils with large H-zones are shown. The ratio of twice the thin filament length denoted distance “A” to the corresponding sarcomere’s length denoted as “B” for all four genotypes is quantified in (E). A schematic illustrates the measurement of these distances. Sample sizes were as follows: ctrl n=119 sarcomeres, ex15e/+ n=122, Act88FKM88/+ n=119, ex15e/+; Act88FKM88/+ n=119. The ratio A/B represents the proportion of the sarcomere length occupied by thin filaments and is inversely correlated with H-zone length; ex15e/+; Act88FKM88/+ sarcomeres have significantly larger H-zones than all others. All sarcomeres with an Act88FKM88 allele were radially narrower than control, but a synthetic enhancement could not be concluded from this observation. (Unpaired t-test with Welch’s correction; ctrl-ex15e/+ p=0.0502; ctrl-Act88FKM88/+ p=0.2770; ctrl-ex15e/+; Act88FKM88/+ p<0.0001; ex15e/+-Act88FKM88 p=0.4938; ex15e/+-ex15e/+; Act88FKM88 p<0.0001; Act88FKM88/+-ex15e/+; Act88FKM88 p<0.0001).

FRAP reveals defects in ex15e protein dynamics

FRAP experiments visualizing recovery of GFPZasp52-PR were performed on one-day-old female IFM Z-discs either in our mutant background (ex15e;UH3>GFPZasp52-PR, n=9 flies) or wild-type background which served as a control (UH3>GFPZasp52-PR, n=12) (A). Control UH3>GFPZasp52-PR Z-discs fixed with 4% paraformaldehyde for 10 minutes (labelled “Fixed”, n=7) displayed no recovery as expected. Plateaus of fitted curves were significantly different, with a greater mobile fraction in the mutant background (Unpaired t-test with Welch’s correction; p<0.0001) (B). Representative snapshots of bleached Z-disc recovery illustrate this difference between genotypes (C).

Certain ex15e defects are rescued by immobilization

Rescues were attempted by immobilizing flies thereby preventing IFM use. ex15e myofibrils in flies that were immobilized (C) resembled those of ctrl (A). Bend angles were equivalent between ctrl and ex15e immobilized flies (n=2119) (D). (Two One-Sided Test (TOST) with ±4.5° equivalence margin, α=0.05; tL=15.36, tU=-12.00). The proportion of sarcomeres that were contracted was not significantly different between ctrl and immobilized flies (n=2246) (E). (Fisher’s exact test p=0.2026). Flight ability was rescued with immobilized flies (n=67 flies) having no significant flight defect compared to control (n=86) but being significantly improved from their non-immobilized counterparts (n=81) (F). (Fisher’s exact test; ctrl-immobilized p=0.0640; ctrl-ex15e and ex15e-immobilized p<0.0001).

MODENCODE RNA-seq data showing expression levels at the Zasp52 locus across various developmental stages.

Exon 15e is highlighted in blue.

(A) Various disorder prediction algorithms were used to assess the level of per-residue disorder at the Zasp52 locus. The corresponding Zasp52 locus is schematized below, with PDZ domain (blue), ZM (red), LIM domains (green), and the ABM identified in Ashour et al., 2023 (yellow). For disorder prediction algorithms (AiUPRED, Disopred, and metapredict), the higher y-values represent higher levels of disorder while low values close to zero represent low levels of disorder (structure). AlphaFold pLDDT scores are plotted alongside in black. pLDDT scores can be used to inform about disorder with higher scores indicating higher confidence that a region is ordered, and lower scores suggesting disorder (Ruff & Pappu, 2021; Emenecker et al., 2021). (B) Dendrogram of various insects with Zasp52 orthologs that contain a PDZ domain and four LIM domains. Canonical isoforms which usually are also the longest ones were used to compute the number of residues comprising the linker between the first and second LIM domains, with data represented in a heatmap. The purple bar indicates all Drosophilids, and the teal bar indicates species of the order Ephemeroptera and Odonata which use direct flight. Though some Drosophilids contain shorter linkers, no other species contain linkers of comparable length to Drosophilids and other closely related Dipterans. NARDINI Z-score matrices of the linker region between LIM1 and LIM2 in Zasp52-PF (C) and of only the sequence encoded by exon 15e. The diagonal squares represent Ω-values which quantify the linear mixing versus segregation of all of residue type x with respect to all other residues. The off-diagonal squares represent δ-values which quantify the linear mixing versus segregation of pairs of residue types. A high z-score thus indicates segregation/blocking of a certain residue type with respect to itself/others compared to a null model; a low score indicates even dispersion and intermixing. Polar residues μ = {S,T,N,Q,C,H}; hydrophobic residues h = {I,L,V,M}; basic residues + = {R,K}; acidic residues - = {E,D}; aromatic residues π = {F,W,Y}; alanine = A; proline = P; glycine = G.

(A) A binary flight assay across three age categories of control, homozygous ex15e, and ex15e over a Zasp52 deficiency BSC427. Sample size were as follows: one-day-old (ctrl n=44 flies, ex15e n=42, ex15e/BSC427 n=45), five-days-old (ctrl n=48, ex15e n=39, ex15e/BSC427 n=28), and three-weeks-old (ctrl n=61, ex15e n=30, ex15e/BSC427 n=23). Flight was scored manually on whether they could generate upwards lift or not. There was no significant difference between the homozygous mutant or over the deficiency (Fisher’s exact test; 1d.o. p=1.0000; 5d.o. p=0.7843; 3w.o. p=0.5729). Polarized light images of sagittally bisected thoraces of three-week-old control (B) and ex15e (C) revealed no obvious differences in gross morphology; females shown. (D-G’’) MTZ were affected in ex15e. Close up views revealed a disruption in the MTZ and adjacent myofibrils in three-week-old ex15e mutants (E) compared to control (D). The regular “crown”-shaped structure of the control MTZ is replaced with a disordered structure of variable shape in ex15e mutants. βPS integrin and Zasp52 localization appeared unaffected in ex15e (F-G’’). Thin filaments were visualized with phalloidin in green, Z-discs with a Zasp52 full-length antibody in red, and integrin adhesion sites at MTZs with a βPS integrin antibody in blue.