In-frame exon deletions of the Duchenne muscular dystrophy (DMD) gene produce internally truncated proteins that typically lead to Becker muscular dystrophy (BMD), a milder allelic disorder of DMD. We hypothesized that differences in the structure of mutant dystrophin may be responsible for the clinical heterogeneity observed in Becker patients and we studied four prevalent in-frame exon deletions, i.e. Δ45-47, Δ45-48, Δ45-49 and Δ45-51. Molecular homology modelling revealed that the proteins corresponding to deletions Δ45-48 and Δ45-51 displayed a similar structure (hybrid repeat) than the wild-type dystrophin, whereas deletions Δ45-47 and Δ45-49 lead to proteins with an unrelated structure (fractional repeat). All four proteins in vitro expressed in a fragment encoding repeats 16-21 were folded in α-helices and remained highly stable. Refolding dynamics were slowed and molecular surface hydrophobicity were higher in fractional repeat containing Δ45-47 and Δ45-49 deletions compared with hybrid repeat containing Δ45-48 and Δ45-51 deletions. By retrospectively collecting data for a series of French BMD patients, we showed that the age of dilated cardiomyopathy (DCM) onset was delayed by 11 and 14 years in Δ45-48 and Δ45-49 compared with Δ45-47 patients, respectively. A clear trend toward earlier wheelchair dependency (minimum of 11 years) was also observed in Δ45-47 and Δ45-49 patients compared with Δ45-48 patients. Muscle dystrophin levels were moderately reduced in most patients without clear correlation with the deletion type. Disease progression in BMD patients appears to be dependent on the deletion itself and associated with a specific structure of dystrophin at the deletion site.

Dystrophin is a cytoskeletal protein that confers resistance to the sarcolemma against the stress of contraction-relaxation cycles by interacting with cytoskeletal and membrane partners. Apart from several proteins, membrane phospholipids are a partner of the central rod domain made up of 24 spectrin-like repeats, separated into sub-domains by four hinges. We previously showed that repeats 1 to 3 bind to membrane anionic phospholipids, while repeats 20 to 24 are not able to do so. We focus here on the phospholipid-binding properties of the major part of the central rod domain, namely, the sub-domain delineated by hinges 2 and 3 comprising 16 repeats ranging from repeat 4 to 19 (R4-19). We designed and produced multirepeat proteins comprising three to five repeats and report their lipid-binding properties as well as their thermal stabilities. When these proteins are mixed with liposomes including the anionic lipid phosphatidylserine, they form stable protein-vesicle complexes as determined by gel-filtration chromatography. The absence of an anionic lipid precludes the formation of such complexes. Spectroscopic analyses by circular dichroism and tryptophan fluorescence show that, while the alpha-helical secondary structures are not modified by the binding, protein trans conformation leads to the movement of tryptophan residues into more hydrophobic environments. In addition, the decrease in the molar ellipticity ratio at 222/208 nm as observed by circular dichroism indicates that lipid binding reduces the inter-helical interactions of multirepeat proteins, thus suggesting partly "opened" coiled-coil structures. Combining these results with data from our previous studies, we propose a new model of the dystrophin molecule lying along the membrane bilayer, in which the two sub-domains R1-3 and R4-19 interact with lipids and F-actin, while the distal sub-domain R20-24 does not exhibit any interaction. These lipid-binding domains should thus maintain a structural link between cytoskeletal actin and sarcolemma via the membrane phospholipids.

Annexin 1 (An 1), a phospholipid and calcium binding protein, is strongly expressed in differentiated U 937 cells. In attempting to correlate the expression of An 1 with phospholipase A2 (PLA2) activity, U 937 cells were stably transfected both with a Sense and Antisense cDNA for An 1. PLA2 activity was measured by Flow cytometry analysis utilizing the bis‐Bodipy‐C11‐PC fluorescent probe. U 937 cells stably transfected with the sense or antisense vectors were differentiated for 24 h with phorbol 12‐myristate 13‐acetate (PMA, 6 ng ml−1). Both in undifferentiated and differentiated cells, the Antisense clone (36.4 AS) showed consistently higher PLA2 activity than the control Sense clone (15 S). Since the fluorescent probe measures the total PLA2 activity, we used two different stimuli, PMA: (100 ng ml−1) or lipopolysaccharide (LPS, 10 ng ml−1), and two different inhibitors, to discriminate the PLA2 involved (namely arachidonyl trifluoromethyl ketone or AACOCF3, which is specific for the cytosolic PLA2, and SB 203347 specific for the secretory PLA2). In the Antisense clone the inhibitory effect of AACOCF was stronger [68%, P<0.025] than in the Sense, which may reflect the lower endogenous level of An 1 present in the cells. On the contrary, the inhibitory effect of SB 203347 [60% of inhibition] was identical in both clones. Since cPLA2 activity is correlated with its phosphorylation, Western and shift blot analysis were performed. They did not show any significative difference between the phosphorylated and non phosphorylated form of the enzyme in both the differentiated or not, Sense and Antisense clones. Furthermore the tyrosine phosphorylation analysis of An 1 showed that less than 10% of An 1 was phosphorylated irrespective of PMA presence or absence. From the pattern of inhibition observed, we propose that the endogenous unphosphorylated form of An 1 may act intracellularly to block the activity of a cytosolic PLA2. British Journal of Pharmacology (1998) 124, 1675–1683; doi: