Novel interaction of myospryn with C-terminal titin and CAPN3

Myospryn (CMYA5, cardiomyopathy-associated 5), identified in the Y2H screens as an interaction partner for C-terminal titin and calpain 3, is a large multi functional protein related to the tripartite motif (TRIM) family. Its expression on transcript and protein levels was initially characterized as restricted to skeletal and cardiac muscles (Benson et al. 2004, Durham et al. 2006), but moderate myospryn expres-sion in the brain has been recently reported (Chen et al. 2011).

The official symbol CMYA5, suggesting an association with cardiomyopathy, was originally coined on hypothetical grounds, based on coexpression with known cardiomyopathy genes (Walker 2001). Aliases genethonin-3 (Tkatchenko et al.

2001) and stretch-responsive 553 or sr553 (McKoy et al. 2002) have been used in individual early studies. Although lacking the RING domain present in canonical TRIM proteins, myospryn is also classified as TRIM76 (HGNC Database).

Human myospryn, comprising 4069 amino acid residues, has a predicted mo-lecular weight of 449 kDa (Fig. 8). Its C-terminal region of ~600 amino acids has a domain architecture resembling the TRIM proteins: The region contains a B-Box’

domain (a variant of the B-Box zinc finger domain containing two instead of four Zn²⁺-coordinating residue pairs), a BBC (B-Box C-terminal coiled-coil), two FN3 domains, and a SPRY (SPlA and RYanodine receptor) domain (Benson et al. 2004).

According to Y2H and CoIP studies, the TRIM-like region is capable of self-associ-ation, suggesting that the protein may exist as a homodimer (Benson et al. 2004).

The TRIM-like region of myospryn is highly conserved in evolution, with ~90%

identity between the human and murine orthologues. Another region showing evo-lutionary conservation is a stretch of ~100 amino acid residues near myospryn N-terminus, but the domain structure and functions of this part of the protein are currently unknown. Most of myospryn is comprised of repetitive, glutamate-rich sequence without predictable domains (Benson et al. 2004); this low-complex-ity region shows variation in length and repeat composition between myospryn orthologues in different species, and could thus serve as a flexible linker between the conserved N-and C-terminal regions.

PGM1

FINmaj / FINmaj

α-actinin (Z-disc) PGM1 / α-actinin

Figure 7. Phosphoglucomutase 1 in muscle

Confocal microscopy of PGM1 and the Z-disc marker a-actinin in mouse muscle. PGM1 was mostly localized around the Z-disc, but it also exhibited variable degrees of M-band localization, ranging from undetectable to prominent (exemplified here), No obvious difference was seen between wild-type (not shown) and homozygous FINmaj KI (FINmaj / FINmaj) muscles. Scale bar, 10 mm.

5.1.3.1 Myospryn interacts in Y2H with wild-type but not FINmaj mutant M10 (I) In the M10 interaction screen, myospryn was represented by two recovered prey clones, spanning 259 (A3811–CT) and 210 (L3860–CT) C-terminal residues of the protein. Pairwise Y2H studies (I: Fig. 2) replicated the interaction of the two recovered myospryn prey clones with the wild-type M10 bait, while no reporter activation was seen with the empty bait vector. Moreover, both myospryn clones failed to cause reporter activation in combination with the FINmaj mutant M10 bait, suggesting that the interaction is specific and requires a correctly folded M10 domain. Additional Y2H experiments with deletion constructs showed that half of the C-terminal FN3 domain and the SPRY domain of myospryn are required for the interaction with titin M10. Progressively weaker reporter signals obtained with a series of deletion constructs starting in the region L3860–Y3875 suggested that this region may contain amino acid residues directly participating in titin binding (I: Fig. 1–2).

5.1.3.2 A larger C-terminal part of titin participates in myospryn binding (I) To confirm the interaction of myospryn with titin, CoIP experiments were carried out in COS-1 cells. The myospryn constructs utilized in these studies covered three different C-terminal regions: A3811–CT of human myospryn (corresponding to the longer prey clone recovered in the M10 screen), as well as G3039–CT (MD7) and Y2731–CT (MD9) of the murine protein (Benson et al. 2004). In contrast to

expec-C C N

BB BBC FN3 FN3 SPRY

self-association

dystrophin dysbindin

PKA RIIα

α-actinin

desmin calcineurin

human: 4069 aa / 449 kDa

mouse: 3739 aa / 413 kDa TRIM-like region ~600 aa

Figure 8. Structure and reported interactions of myospryn

Most of the myospryn (CMYA5) protein is composed of repetitive, acidic sequence with no recognizable domains. The C-terminal part of the protein contains a TRIM-like domain structure with BBox’ (BB), coiled-coil (BBC), FN3, and SPRY domains. Sites of reported protein interactions are indicated.

Adapted from Sarparanta J: Biology of Myospryn: What’s Known? Journal of Muscle Research and Cell Motility, 2008; 29:177–180 with kind permission from Springer Science and Business Media. © 2009 Springer Science and Business Media B.V.

tations from the Y2H studies, none of the myospryn constructs showed binding to a GFP-tagged M10 construct in the tested CoIP conditions. All were, however, pulled down with a longer (is6–M10-V5 is7⁺) titin construct, supporting the inter-action of the proteins but suggesting that myospryn binding is not limited to the M10 domain (I: Fig. 3).

Additional CoIP studies were performed with the myospryn construct MD9 and a series of V5-tagged titin constructs of different lengths. Either V5 or Myc antibodies were used for immunoprecipitating the titin or myospryn constructs, respectively. In these experiments—in addition to is6–M10-V5 is7⁺—also titin constructs spanning the regions is6–M10 (is7^–) and is6–is7 showed clear bind-ing to myospryn MD9, whereas for is6–M9 and is6–M8, the interaction was weak or undetectable (I: Fig. 3). The difference between is6–M9 and is6–is7 indicates that the is7 region participates in the interaction. The binding seen with is6–M10 is7^– but not with M10 alone, on the other hand, demonstrates the contribution of the M9 domain. In conclusion, the whole C-terminal region M9–is7–M10 in titin is involved in myospryn binding (Fig. 9).

In CoIP experiments the FINmaj mutation failed to disrupt the interaction of is6–M10 constructs with myospryn, but rather seemed to increase the binding.

This discrepancy between CoIP and Y2H results is likely due to preserved inter-action of M9–is7 with myospryn despite the mutation. The apparent increase in binding could be explained by unspecific adhering of proteins to the unfolded M10 domain.

Figure 9. Interactions between myospryn, C-terminal titin, and CAPN3

Summary of the novel and previously reported interactions between myospryn (CMYA5), C-terminal titin (TTN), and CAPN3. Regions participating in the interactions, as defined by Y2H and CoIP assays, are indicated by black bars.

5.1.3.3 The interaction of CAPN3 and myospryn is supported by CoIP (I)

The interaction of myospryn with CAPN3, identified in the Y2H screen for CAPN3 ligands, was confirmed by CoIP studies in COS-1 cells, where both myospryn con-structs MD7 and MD9 pulled down the proteolytically inactive CAPN3 C129S construct with comparable efficiencies (I: Fig. 3). Based on the regions covered by the Y2H bait construct and the recovered prey clones, the interaction involves the C2-like domain and IS2 of CAPN3, and a myospryn region encompassing the BBox’, coiled-coil, and the first FN3 domain (Fig. 9). As this was not refined further with deletion constructs, the minimal interacting regions could be smaller.