Modeling human PMM2-CDG in medaka to understand systemic effects of hypoglycosylation on development

Congenital disorders of glycosylation (CDG) are a group of rare metabolic diseases caused by mutations in the enzymes involved in glycosylation. To date, no conclusive pathogenic mechanism is linked to the disease and there are only limited therapeutic options available. The most common form of CDG is caused by compound heterozygous hypomorphic alleles of the cytosolic enzyme Phosphomannomutase 2 (PMM2), leading to global protein hypoglycosylation and a multiorgan phenotype in patients. Complete loss-of-function mutations in the essential protein are incompatible with life. Patients that survive carry mutations that lead to reduced enzyme activity of PMM2. PMM2 plays a crucial role at the basis of the glycosylation cascade in the endoplasmic reticulum by providing the essential mannose precursor required for the three glycosylation routes N-glycosylation, O- and C-mannosylation. The entire glycosylation machinery is evolutionarily highly conserved which allows generating translational models in other organisms. The small teleost medaka (Oryzias latipes) offers a great advantage as model organism to investigate the early embryogenesis through the extrauterine and transparent developing embryos. To understand acute effects of Pmm2 loss on development and disease progression, I have created translational models that mimic the reduced residual enzyme activity of PMM2-CDG patients. I followed two routes to investigate the role of PMM2 on development: establish a conditional knockdown system for acute Pmm2 interference at the protein level and patient-based genetic models by precision genome editing in medaka. For the conditional knockdown in medaka, I applied an inducible degron system to selectively degrade Pmm2-GFP, mimicking reduced enzyme activity. To generate patient-based pmm2 alleles, I used canonical base editing to generate the p.C139R mutation and a deletion variant. Further, I developed and carefully examined a novel two step one-shot base editing approach termed “inception”. Inception introduces a new binding site for a second base editing event that subsequently leads to the anticipated edit. With inception I created one of the common patient mutations p.F119L. Depending on the generated alleles, different pmm2 variants resulted in varying enzymatic activity in medaka, phenocopied the patient symptoms and caused hypoglycosylation of proteins. Bottom-up proteomics revealed molecular changes already present before the onset of multisystemic phenotypes, affecting particularly mRNA processing and eye development.