Human skeletal myopathy myosin mutations disrupt myosin head sequestration

Myosin heavy chains encoded by MYH7 and MYH2 are abundant in human skeletal muscle and important for muscle contraction. However, it is unclear how mutations in these genes disrupt myosin structure and function leading to skeletal muscle myopathies termed myosinopathies. Here, we used multiple approaches to analyze the effects of common MYH7 and MYH2 mutations in the light meromyosin (LMM) region of myosin. Analyses of expressed and purified MYH7 and MYH2 LMM mutant proteins combined with in silico modeling showed that myosin coiled coil structure and packing of filaments in vitro are commonly disrupted. Using muscle biopsies from patients and fluorescent ATP analog chase protocols to estimate the proportion of myosin heads that were super-relaxed, together with x-ray diffraction measurements to estimate myosin head order, we found that basal myosin ATP consumption was increased and the myosin super-relaxed state was decreased in vivo. In addition, myofiber mechanics experiments to investigate contractile function showed that myofiber contractility was not affected. These findings indicate that the structural remodeling associated with LMM mutations induces a pathogenic state in which formation of shutdown heads is impaired, thus increasing myosin head ATP demand in the filaments, rather than affecting contractility. These key findings will help design future therapies for myosinopathies.


Table S2
The FiberSim 2.1.0software has been thoroughly described (1).Briefly, FiberSim tracks the position and status of actin and myosin molecules within a network of compliant thick and thin filaments.For all the simulations, the half-sarcomere lattice was composed of 100 thick filaments and 200 thin filaments.These filaments were arranged in a hexagonal lattice to mimic the architecture of human myofibres.Filaments located at the edge of the lattice were "mirrored" on the opposite side to minimize edge effects (2).Each thin filament was composed of two actin strands.Each strand contained 27 regulatory units, and each regulatory unit contained 7 binding sites.Each thick filament was composed of 54 myosin crowns with each crown consisting of 3 pairs of myosin dimers.Each binding site on actin could be in an inactive (unavailable for myosin binding) or active (available for myosin binding) state.All 7 binding sites from a regulatory unit switched simultaneously between those two states, depending on the Ca 2+ concentration and on the transition rate constants kon and koff which we set.A cooperative mechanism was also implemented such that the transition probability for a regulatory unit was influenced by the states of its neighbours.Finally, a regulatory unit was prevented from deactivating if one or more myosin head(s) were bound.Although this model only has two explicit thin filament states, it mimics an important feature of the three state thin filament model described by McKillop and Geeves (3) in that bound myosin heads inhibited relaxation.Heads in SRX switched to DRX at a rate k1 that is assumed to be force-dependent (1).DRX heads could then attach to available binding sites on actin.The attachment and detachment rates depended on x, where x is the distance to the binding site measured parallel to the filaments.The rate functions are provided in Table S2.The detachment rate function (k4) was updated for the present study so that it had an exponential strain-dependence similar to that measured for single myosin heads via optical trapping (4).Model parameters were chosen to reproduce physiological values for: maximal isometric force (≈ 150-200 kPa at a sarcomere length of 2.2 µm), passive force (≈ 1-2% of the maximal isometric force) and Ca 2+ sensitivity (pCa50 ≈ 5.7).Here are the parameters used.

Figure S1
Myosin filament length.Distributed deconvolution (DDecon) was applied from the acquired images with a specific plugin for ImageJ (National Institutes of Health, Bethesda, MD) (5).Note that DDecon is a super-resolution light microscopy technique that allows the computation of filament lengths with a precision of 10.00-20.00nm (5).All line scans were background-corrected. Distances (and myosin filament lengths) were finally calculated by converting pixel sizes into micrometer using the magnification factor for each image (5). A. displays two typical images obtained with confocal microscopy using the A4.951 antibody and allowing the measurement of myosin filament lengths (scale bar: 5 µm).B. shows measurements for individual myofibres expressing the β/slow myosin heavy chain isoform in every single subject (n=161) whilst C has data relating to muscle fibres expressing the type IIA myosin heavy chain isoform (n=94).Means and standard deviations also appear on these graphs.The one-way ANOVA with Dunnett's test post-correction was used but no significant differences were seen.

Figure S2
Figure S2Myosin head order.A. depicts typical X-ray diffraction patterns.B. shows equatorial intensity ratio (IR) and the main myosin meridional reflections, namely M3 and M6 (M3 and M6 intensities were normalised to the 6 th actin-layer line, ALL6).To ensure reliable results and avoid misinterpretation, we pooled all the patients' data (n=11) together and compared these to images acquired for controls (n=12).Means and standard deviations also appear on these graphs.* The one-way ANOVA with Dunnett's test post-correction was used with p < 0.05.
Figure S3Charge plot of MYH7 LMM region.The diagram shows the positions of mutations in GST-LMM, and the plot shows the alternating regions of positive and negative charges important for filament formation, together with the positions of the mutations.

Table S1
Patient and control muscle biopsy samples used.

Table S3
Summary of the results from Figures 2 to 5. Mean values are presented (NDnot done, NFno filaments).

Table S4
Summary of the results from Figures 6 to 7. Mean values are presented.