Genetic studies and molecular structures: the dystrophin associated complex

Genetic studies are usually used to infer the genomic localisation of a locus of interest. Sometimes, they can also reveal the structural properties of the encoded proteins. In the past 16 months the genes coding for the α- (1), β- (2,3) and γ-sarcoglycans (4) have been identified as being involved in the etiology of three clinically similar, yet genetically distinct forms of autosomal recessive (AR) progressive muscular dystrophy (MD), namely LGMD2D, 2E and 2C (reviewed in ref. 5). All three glycoproteins are known to be part of a transmembrane complex, involving a number of additional proteins, and linking the intracellular actin to the extracellular matrix (reviewed in refs 6–8). These three sarcoglycans can be co-purified as part of a dystrophin-associated glycoprotein complex (DGC), and mutations in any one of these loci were reported to lead to severely reduced amounts of all sarcoglycans in muscles of patients (7). Therefore, there is no doubt that they form a complex oligomeric structure, symbolised here, for the sake of simplicity, as (αi, βi, γi) (although by default the DGC is usually represented as consisting of monomers). Ervasti and Campbell (9) provided biochemical evidence suggesting that the sarcoglycans were in equimolar ratios with dystrophin. Yet, their exact stoichiometry within the DGC is not known. We also do not yet know whether dystrophin is present in muscle fibers as a monomer or a dimer, nor whether it is intertwined in an antiparallel fashion to another dystrophin molecule, as initially suggested (10) based on its similarity with spectrin. Both null type and mis sense mutations were identified in all three sarcoglycan genes. In general, missense mutations were found to lead to less severe phenotypes than null mutations (11,12). Also, as mentioned earlier, all three sarcoglycan genes are associated with AR progressive MDs. Yet, considering the oligomeric structure of the sarcoglycan complex, one might expect particular missense mutations in one or another of these genes to cause structural alterations having a dominant-negative effect. This however was not observed. Therefore, in view of their AR inheritance, I propose that genetic evidence suggests that the stoichiometric index ‘i’ of the (αi, βi, γi) complex is equal to 1. Let us focus on α-sarcoglycan, although the same rationale applies to both the β- and γ-sarcoglycans, as well as the extracellular Iaminin-α2 gene, involved in congenital MD (13), and consider two alternative situations, one where i = 1 versus one where i > 1, say 2 (Fig. 1). Furthermore, the present discussion is restricted to those missense mutations that do not induce a drastic reduction of the sarcoglycan complex, as the others are, for all practical purposes, equivalent to null mutations. Schematic representation of a αβγ protein complex, with normal and mutant α chains. Let us examine what would happen if i = 1. In a heterozygous carrier of a α-sarcoglycan mutation (α*), in theory, there should be—depending on the relative stability of the mutant polypeptide—equal amounts of α and α* and of the (α,β,γ) and α*,β,γ) derived complexes. Thus for mutations that do not abolish the formation of this complex and its functional addressing into the sarcolemmal membrane, the muscle cell membranes will only have half of the normal amount of functional sarcoglycan complex. This we know is not pathogenic, as heterozygous carriers of null mutations do not show clinical symptoms. If these complexes are interconnected through dystrophin, the α* mutation will only decrease the amounts of functional paired complexes to one quarter of the normal amount (each complex has a 50% chance to connect to either a normal or mutated complex). Whereas these residual amounts could be sufficient to prevent the development of the pathology, they may nevertheless impair the muscle at the cellular level, leading to occasional leakages, thus explaining the elevated levels of serum CPK observed in some heterozygous carriers. In contrast, three types of structures can be formed when i = 2; the normal α2, the chimeric αα* and the inactive α*2 (with relative proportions of 1:2:1; Fig. 1B). If a single subunit suffices to destroy the function of the complex, only 25% of the formed complexes would be active since the αα* and α*α* complexes would not be functional [in general, the residual amounts of functional complex (αi) are given by 1/2i]. Furthermore, if these complexes are connected intracellularly by two antiparallel dystrophin molecules (10), in some instances as little as 1 in 16 linked pairs of complexes will be functional (1/4 ×l/4). Depending on the tolerance level, i.e., the buffering capacity of the muscle to decreasing dystrophin-associated complexes, we expect different modes of inheritance: a recessive one if the buffering capacity is high, a dominant one if low. An additional category of non-lethal missense mutations ought to be considered, namely those that change the function of the sarcoglycan complex. Indeed, given that the latter is part of a more complex network, linking the extracellular matrix to the cytoskeleton, a fraction of these mutations can be expected to exert a poisonous effect ‘in trans’ on the remaining quarter of normal α2-based complexes. These mutations are thus likely to dramatically perturb the amount of functional complexes, destabilising neighbouring DGC or the sarcolemma. Evidently the impact of such mutations will be drastically different from those described above, in that they will have a dominant-negative mode of inheritance. Examples of such dominant-negative effects have been documented, a.o., collagen mutations in osteogenesis imperfecta type II (14), myosin-dependent dominant familial hypertrophic cardiomyopathy (15), or myotonia congenita (16). As a matter of fact, Steinmeyer et ah (16) used this observation to infer the homo-oligomeric structure of the C1C-1 chloride channel. However, no dominant sarcoglycanopathy has been proven so far. We propose that this argues against complex oligomeric structures for the DGC, where i > 1. It could instead be interpreted as evidence in favor of a simple one-to-one structure for the different components of this complex. Furthermore, this argument could be extrapolated to the dystrophin molecule itself. As already mentioned, the latter is conventionally represented as a pair of antiparallel intertwined molecules, a structure thought to provide the basis for the regular lattice of DGC complexes along the sarcolemma. This view is, however, contested following the observation of the lack of sequence conservation among the surface residues along the spectrin-like domain of the dystrophin (and utrophin) coiled coil region(s) (17). In addition to this, female carriers of X-linked Becker mutations are clinically asymptomatic. Furthermore, the absence of dominant-negative effects in wild type transgenic mice overexpressing various truncated or mutated versions of the dystrophin molecule (J. Chamberlain, pers. comm.) can also be taken as additional evidence in support of the monomeric configuration. The same argument could be extrapolated to question the existence of the proposed utrophin/dystrophin heterodimer at the neuromuscular junction (18). It should be noted that the intricate structure of the sarcoglycan complex raises the possibility that compound heterozygotes (i.e., carrying a non-allelic mutation in genes for two different dystrophin-associated proteins) could in particular cases manifest clinical MD symptoms. This is in agreement with the observation by Beggs et al. (19) of a pseudo-dominant phenotype in patients suspected to be heterozygous carriers of a Fukuyama congenital muscular dystrophy mutation while also having a mutation in the X-linked dystrophin locus. These situations do, however, represent rare exceptions. To sum up, the recessive inheritance, or rather the absence of dominance of these muscular dystrophies, suggests that the various proteins involved are present in the DGC complex as monomers. Whereas genetic reasoning can be used to infer the stoichiometry of various polypeptides of the DGC complex, final demonstration of its precise molecular organisation still awaits further biochemical characterisation. I would like to thank Drs K.P. Campbell, J. Chamberlain, F. Duclos, M. Jeanpierre, J.-C. Kaplan andE. Ozawafor stimulating discussions and the Association Française contre les Myopathies (AFM) for support.