Costameric integrin and sarcoglycan protein levels are altered in a Drosophila model for Limb-girdle muscular dystrophy type 2H

Abstract

Mutations in two different domains of the ubiquitously expressed TRIM32 protein give rise to two clinically separate diseases, one of which is Limb-girdle muscular dystrophy type 2H (LGMD2H). Uncovering the muscle-specific role of TRIM32 in LGMD2H pathogenesis has proven difficult, as neurogenic phenotypes, independent of LGMD2H pathology, are present in TRIM32 KO mice. We previously established a platform to study LGMD2H pathogenesis using Drosophila melanogaster as a model. Here we show that LGMD2H disease-causing mutations in the NHL domain are molecularly and structurally conserved between fly and human TRIM32. Furthermore, transgenic expression of a subset of myopathic alleles (R394H, D487N, and 520fs) induce myofibril abnormalities, altered nuclear morphology, and reduced TRIM32 protein levels, mimicking phenotypes in patients afflicted with LGMD2H. Intriguingly, we also report for the first time that the protein levels of βPS integrin and sarcoglycan δ, both core components of costameres, are elevated in TRIM32 disease-causing alleles. Similarly, murine myoblasts overexpressing a catalytically inactive TRIM32 mutant aberrantly accumulate α- and β-dystroglycan and α-sarcoglycan. We speculate that the stoichiometric loss of costamere components disrupts costamere complexes to promote muscle degeneration.

FIGURE 1:

Muscle degeneration is neuron-independent. (A, B) L3 larval ventral longitudinal muscles 3 (VL3) and 4 (VL4) stained with phalloidin to visualize F-actin (green). (A, A’) The stereotypical morphology of WT muscle with normal sarcomeric organization. (B, B’) In addition to sarcomeric disorganization, the VL3 and VL4 muscles are noticeably smaller in tn–/– larvae. (C–F) Phalloidin labeling of VL3 muscles in L3 larvae. Muscle morphology is unaffected in mef2-Gal4, C, or elav-Gal4, E, driver controls. Dystrophic muscle is apparent upon muscle-specific tn RNAi, D, but not upon a decrease of TRIM32 in neurons, F. (G) Quantification shows that induction of tn RNAi in muscles, but not neurons, results in muscle defects. (H) Locomotion assay performed on L3 larvae shows a decrease in mobility upon loss of TRIM32 (tn–/–) or the induction of tn RNAi in muscle tissue (mef2>tn RNAi). The reduction in locomotion is minor upon neuronal-specific RNAi knockdown of TRIM32 (elav>tn RNAi). Mean ± SEM. ****, p < 0.001; *, p < 0.05; n.s., not significant. Scale bars: 50 μm, A, B; 25 μm, A’, B’, C–F.

FIGURE 2:

RING and NHL domain deletion mutants limit muscle growth and induce degeneration. (A) Schematic representation of the RING, B-box, coiled-coil, and NHL domains in TRIM32FL and the TRIM32 deletion mutants. (B–E) VL3 and VL4 muscles stained for F-actin to visualize muscle fiber structure in L3 larvae. (B) tn–/– mutant muscles appear thinner with prominent sarcomere patterning defects. (C) Overexpression of TRIM32FL under control of the mef2 promoter in a tn–/– background (tn–/–; mef2>TRIM32FL) rescued muscle morphology defects. (D, E) Deletion of either the RING, D, or the NHL domain, E, mimics the loss of TRIM32 function phenotype. In addition to muscle morphology defects, ∆RING and ∆NHL deletion mutants exhibited noticeably thinner VL3 and VL4 muscles similar to tn mutants. (F) A scatterplot showing VL3 muscle diameter in the indicated genotypes. Note that TRIM32FL, TRIM32_∆RING, and TRIM32_∆NHL are all expressed in a tn–/– background. Deletion of RING or NHL domains reduces VL3 muscle diameter, while muscle expression of TRIM32FL rescues this muscle loss. (G) Box and whisker plot of the pupal case axial ratios show ∆RING and ∆NHL deletion mutants are defective in muscle contraction, although not quite as severe as tn mutants. Note that TRIM32FL, TRIM32_∆RING, and TRIM32_∆NHL are all expressed in a tn–/– background. Mean ± SEM. ****, p < 0.001; ***, p < 0.005; **, p < 0.01; *, p < 0.05; n.s., not significant. Scale bars: 50 μm, B–E.

FIGURE 3:

Expression of LGMD2H disease-causing alleles causes dystrophic muscle and reduces TRIM32 protein levels. (A) Relative location of human TRIM32 mutations in NHL1 (R394H), NHL3 (D487N), or NHL4 (fs520). (B–E) Phalloidin-labeled body wall muscles in L3 larvae. (B) TRIM32FL control animals (tn-/;mef2>TRIM32FL) show WT muscle morphology. (C–E) Expression of human pathogenic mutations in a tn mutant background cause defects in muscle morphology. (C) Muscle expression of the R394H mutation (tn–/–;mef2>TRIM32_R394H) resulted in extensive muscle damage. Dotted lines indicate absence of sarcomeric structure. (D, E) The severity of the D487N, D, or 520fs, E, mutations in larval musculature was milder than that of R394H mutant muscles, C, with myofibrillar abnormalities represented by white arrowheads. (F–I) TRIM32 protein localization in L3 muscle tissue upon transgene expression in a tn–/– background. (F) Z-disk expression of TRIM32 in tn–/–; mef2>TRIM32FL muscle. (G–I) Z-disk localization appears normal, yet is reduced in protein amounts upon expression of pathogenic alleles. (J) Western blot analysis of whole larvae shows decreased TRIM32 protein levels in R394H, D487N or 520fs mutants compared with TRIM32FL controls. ATP5α is used as a loading control. (K) Bar graph depicts the relative intensity of TRIM32 protein levels normalized to ATP5α in the indicated genotypes. N = 3. (L) qPCR reveals that tn transcripts are indeed present in muscle carcasses upon expression of all transgenes, with mutant transgenes increased over control (mef2/+) or I. mRNA was normalized to rp49 transcripts. N = 3 biological replicates and 3 technical replicates for each genotype. Mean ± SEM. ***, p <0.005; *, p <0.05). Scale bars: 50 μm, A–D; 25 μm, E–H.

FIGURE 4:

Human pathological mutations cause IFM defects and alter TRIM32 localization. IFMs labeled with phalloidin (magenta) and TRIM32 (green) to visualize myofiber architecture. (A₁–A₄) Overexpression of TRIM32FL in a tn–/– mutant background results in flies with normal adult flight muscles. TRIM32 is localized at the Z-disk and M-line within the myofibers at all timepoints examined. (B₁) Induction of the D487N mutation results in IFM’s with WT morphology at day1. Myofibers show mild deterioration at day 3 (B₂) that continues to degenerate through day 6 (B₃) and day 9 (B₄). TRIM32 shows its normal sarcomere association at days 1–6, B_1’–B_3’. By day 9, B_4’, TRIM32 is no longer localized to sarcomeres, but appears in puncta. (C_1’–D_4’) In contrast, expression of the R394H and 520fs mutations induced myofiber defects by day 1, C₁, D₁, that progressively worsened as the flies aged, C₂–D₄. In addition to myofibrillar anomalies, TRIM32 was not localized in the myofibers of R394H, C_1’–C_4’, or 520fs mutants, D_1’–D_4’, and its expression appeared reduced compared with that of TRIM32FL control IFM. (E, F) Stacked bar graphs depicting flight ability of adult flies of the indicated genotypes in a tn–/– background at day 1, E, or day 12, F. Expression of all three mutations reduced flight ability. Scale bars: 10 μm, A₁–D_4’.

FIGURE 5:

The R394H point mutation destabilizes the NHL structure. (A, B) The average energy of each residue in the top 100 lowest-energy members of each ensemble was calculated. The difference between the energies of the WT and D487N, A, or R394H, B, are shown. The site of mutation is shown in spheres. The mutated residues had at least a +2 REU increase in energy (magenta), while amino acids surrounding those residues have a small (0.2–0.4 REU) increase in energy (yellow), especially in D487N, A. Several other surrounding residues find lower-energy conformations of 0.2–0.4 REU (cyan), 0.4–1 REU (violet), or even 1–2 REU (dark blue), evident in NHL domain 4 of R394H, B. (C) Coomassie-stained nickel-column purified fractions of His₍₈₎MBP, His₍₈₎MBP _TRIM32NHL, and the His₍₈₎MBP_TRIM32NHL_R394H mutant protein. (D) Thermal unfolding transitions measured for His₍₈₎MBP (black), His₍₈₎MBP _TRIM32NHL (red), and His₍₈₎MBP _TRIM32NHL_R394H (green) using DSC.

FIGURE 6:

Costamere proteins accumulate abnormally upon loss of TRIM32 function in Drosophila or murine C2C12 cells. (A–B’, D–G) L3 larval muscles stained with phalloidin (green) and βPS integrin (magenta) to visualize muscle morphology. Horizontal panels are XZ confocal scans to show the sarcolemmal association of βPS integrin. (A, A’) βPS integrin localizes normally at the sarcolemma in WT larval muscle (white arrowheads). (B, B’) In tn mutants, βPS accumulates abnormally and appears diffuse at the sarcolemma (white arrowheads). (C) Immunoblot analysis of L3 whole larvae reveal elevated βPS integrin protein levels in tn mutants. Bar graph representing quantification of βPS-integrin/ATP5α protein levels. (D, D’) Muscle rescue with TRIM32FL reverts the βPS localization defect to WT. (E, E’) Transgenic expression of the R394H mutation causes βPS clustering at the sarcolemma (white arrowheads). (F, F’) βPS integrin is mostly normal at the sarcolemma, but increased along discrete regions, in D487N mutants. (G, G’) Abnormal distribution of βPS is prominent in degenerative myofibers in 520fs mutants. (H) Increased βPS integrin protein levels are prominent in R394H, D487N, and 520fs mutants expressed in a tn–/– background. Quantification of βPS-integrin/ATP5α protein levels represented by a bar graph. (I) Western blot depicts significantly high levels of Scgδ protein in tn mutants. Bar graph depicting the relative intensity of Scgδ/ATP5α. (J) Elevated Scgδ protein levels are prominent in all TRIM32 alleles expressed in a tn–/– background as compared with TRIM32FL control animals. Bar graph projecting Scgδ/ATP5α protein levels. (K) Total lysates from C2C12 myoblasts transfected for 48 h with the indicated plasmids were separated by SDS–PAGE. Immunoblotting was performed with antibodies against the indicated costamere protein components. Vinculin was used as loading control. (L) Bar graphs reporting an increase in α−dystroglycan, β−dystroglycan, and α−sarcoglycan protein levels upon expression of catalytically inactive TRIM32. N = 3, Mean ± SEM. **, p < 0.01; *, p < 0.05, n.s., nonsignificant. Scale bars: 10 μm.