A highly prevalent equine glycogen storage disease is explained by constitutive activation of a mutant glycogen synthase

Abstract

Background: Equine type 1 polysaccharide storage myopathy (PSSM1) is associated with a missense mutation (R309H) in the glycogen synthase (GYS1) gene, enhanced glycogen synthase (GS) activity and excessive glycogen and amylopectate inclusions in muscle.

Methods: Equine muscle biochemical and recombinant enzyme kinetic assays in vitro and homology modelling in silico, were used to investigate the hypothesis that higher GS activity in affected horse muscle is caused by higher GS expression, dysregulation, or constitutive activation via a conformational change.

Results: PSSM1-affected horse muscle had significantly higher glycogen content than control horse muscle despite no difference in GS expression. GS activity was significantly higher in muscle from homozygous mutants than from heterozygote and control horses, in the absence and presence of the allosteric regulator, glucose 6 phosphate (G6P). Muscle from homozygous mutant horses also had significantly increased GS phosphorylation at sites 2+2a and significantly higher AMPKα1 (an upstream kinase) expression than controls, likely reflecting a physiological attempt to reduce GS enzyme activity. Recombinant mutant GS was highly active with a considerably lower K_m for UDP-glucose, in the presence and absence of G6P, when compared to wild type GS, and despite its phosphorylation.

Conclusions: Elevated activity of the mutant enzyme is associated with ineffective regulation via phosphorylation rendering it constitutively active. Modelling suggested that the mutation disrupts a salt bridge that normally stabilises the basal state, shifting the equilibrium to the enzyme's active state.

General significance: This study explains the gain of function pathogenesis in this highly prevalent polyglucosan myopathy.

Keywords: Glycogen; Glycogen storage disease; Glycogen synthase; Muscle; PSSM1; Polyglucosan.

Figure 1

A) Glycogen content of wet weight muscle samples from WT (RR)(n=7), PSSM1-heterozygotes (RH)(n=8) and PSSM1-homozygotes (HH)(n=4) horses (median +/− min/max values, box represents interquartile range; RR vs. RH p=0.005, RR vs. HH p= 0.041, *p<0.05). B) Total glycogen synthase expression in muscle homogenates from each of the genotypes was not significantly different (p=0.98). Glycogen synthase activity shown as C) fractional velocity (FV) and D) l-form activity. There was an increase in %FV between PSSM1-homozygotes (HH) and WT controls (RR) (p=0.03) and an increase in %I-form activity between heterozygotes (RH) and PSSM1-homozygotes (HH) samples (p=0.04) (median +/− min/max values, box represents interquartile range, RR (n= 12), RH (n=13), HH (n=4) *p<0.05).

Figure 2. Western blot analysis of GS phosphorylation in skeletal muscle homogenates

There were no significant differences between the genotypes for phosphorylation of A) p1b isoform (p=0.48) and C) p3a+3b isoform (p=0.17). There was significantly increased phosphorylation at sites p2+2a (B) in PSSM1-homozygotes compared to control horses (RR) (p=0.009); (mean +/− SEM, RR (n=12) RH (n=13) HH (n=3) *p<0.05).

Figure 3. AMPK subunit expression and phosphorylation at Thr 172 from skeletal muscle extracts

There was a significant increase in AMPKα1 expression in PSSM1-homozygote samples (HH) compared to heterozygotes (RH) and WT controls (RR) (A; p=0.04) and a significant decrease in AMPKβ1 expression in the PSSM1-homozygote samples (HH) compared to the heterozygotes (C; p=0.03). There were no significant differences in AMPKα2 (p=0.15) or AMPKβ2 (p=0.28) expression between different genotypes (B and D). There was a significant and strong correlation between AMPKα1 and pAMPK expression (F; R²=0.66, p<0.0001). (mean +/− SEM, RR (n=12) RH (n=13) HH (n=3) *p<0.05). Data for A to E is normalised to the mean of the WT control samples.

Figure 4. A) Glucose-6-phosphate and B) UDPG titration (with or without 10mM G6P) curves for WT and R309H mutant enzyme

These curves show that the mutant enzyme has much higher enzyme activity at much lower concentrations of G6P and reaches its maximal activity at much lower G6P concentrations (A). The mutant enzyme also has a higher affinity for UDPG (B). (Mean +/− SEM, n=3 repeats)

Figure 5. Phosphorylation (A) and Activity ratio (B) of insect derived WT and mutant (R309H) equine and human WT and S7A glycogen synthases

Activity ratio is minimal in both equine and human WT enzymes that are untreated with protein phosphatase 1 (PP1), but each WT enzymes’ activity ratio increased when treated with the enzyme. In contrast, the equine mutant enzyme has a high ratio both with and without treatment with PP1 as also seen with the S7A human mutant GS, despite the equine mutant enzyme's phosphorylation. (n=3 replicates for each condition / enzyme; mean +/− SD).

Figure 6. Sequence alignment of equine GS (Genbank accession ACB14276) and yeast Gsy2 (sequence in PDB ID 3NB0)

Red star: R309 in equine GS and the location of the PSSM1-associated R309H mutation. Magenta circles: predicted phosphorylation sites. Blue triangles: binding residues of glucose-6-phosphate. Diamonds: binding residues of maltodextrin (yellow: sites 1 and 2; green: sites 3 and 4).

Figure 7. Homology models of equine GS

(A) Basal and (B) active-state equine WT GS model tetramers are shown as ribbons with monomers coloured different shades of grey. Residues in space-fill are R309 (red), G6P-binding contacts (blue) and maltodextrin-binding contacts (yellow: sites 1 and 2; green: sites 3 and 4). The 309 position and local charged residues are shown as sticks for WT GS [(C) basal; (D) active-state] and the PSSM1-associated R309H mutant [(E) basal; (F) active-state]. Also shown as sticks are the corresponding residues in the crystal structures of yeast Gsy2p [(G) basal; (H) active-state]. Salt bridge interactions are shown as green dashes.