IS B4GALT5 DEFICIENCY A NEW CONGENITAL DISORDER OF GLYCOSYLATION?

Glycosphingolipids (GSLs) are relevant components of mammalian cell membranes where they play multiple roles. Many specific glycosyltransferase enzymes are responsible for the synthesis of the different oligosaccharide chains of GSLs. Knock-out mouse models lacking single or multiple glycosyltransferases were generated in the past elucidating the role of individual GSLs in mouse brain. Inactive variants of two specific glycosyltransferases were found responsible for two human congenital disorders of glycosylation (CDG) characterized by clinical pictures not really predicted by the corresponding mouse KO models. Previous studies showed that two similar galactosyltransferases, B4GALT5 and B4GALT6, are both responsible for the synthesis of lactosylceramide (LacCer), the precursor of many GSLs. B4GALT5 appeared more relevant, but B4GALT6 was still able to rescue B4GALT5 KO in mice. Genome analysis of a patient revealed a compound heterozygous mutation in B4GALT5 gene (DECIPHER Patient ID: 412221) opening the question if a novel human CDG exists not predicted by mouse models. The aim of this thesis is to investigate if the above B4GALT5 variants are actually inactive and if human B4GALT6 is unable to compensate for the lack of B4GALT5 activity, indicating the possibility of such a novel CDG. Cell models were constructed trough KO of either B4GALT5 or B4GALT6, or both together in HEK-293T cells using CRISPR/Cas9 genome editing. Transfections were performed with either coding cDNAs placed in expression vectors without or with N-terminal (Halo) or C-terminal (C-myc) tagging. Transcript levels by qPCR, protein expression by western blotting, galactosyltransferase activity by in vitro assay, and cell GSL biosynthesis by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) were then determined. Similar analyses were also performed on patient’s serum and dermal fibroblasts. Results showed that the mutations do not impair transcript levels or protein expression, while both variants are totally unable to use their substrates in vitro and to restore LacCer and GM3 ganglioside expression upon transfection in double KO HEK-293T cells. In silico studies of mutated proteins revealed pronounced conformational instability compared to the wildtype, affecting both protein backbone and UDP-Gal binding. ΔStability calculations confirmed strong destabilizing effects, particularly for G274R. Moreover, the GSL profile of patient serum and fibroblasts displayed an accumulation of GlcCer and a strong reduction of LacCer, GM3 and globoside Gb3. These findings together indicate that both variants have lost their functional activity. While B4GALT6 KO clones maintain almost unchanged the GSL pattern of HEK-293T cells, B4GALT5 KO clones retain less than 20% of the above GSLs, despite the transcript level of B4GLT6 is about 10-fold higher than that of B4GALT5. Moreover, upon transfection of B4GALT6 cDNA untagged or tagged at the C-terminus, no protein is detectable by western blotting, despite the high levels of B4GALT6 transcript achieved. Protein expression becomes evident only upon Halo tagging at the N-terminus, allowing detection of LacCer synthase activity in vitro and GSL rescuing in dKO clones. In particular, B4GALT6 ability to restore GSL expression is comparable to that of B4GALT5. However, detection of Halo-tagged B4GALT6 by western blotting is not comparable with that of Halo-tagged B4GALT5, since it requires much higher transcript levels to achieve comparable expression levels, as confirmed by transfecting the cDNAs placed under the control of weaker promoters. Hypothesizing a reduced efficiency of B4GALT6 translation, we constructed and transfected B4GALT6 chimeras where B4GALT5 N-terminal tail and transmembrane domain, or those plus the entire stem region, replaced the corresponding sequence of the native cDNA. Western blot analysis reveals detectable B4GALT6 protein with both chimeras, confirming the hypothesis. In conclusion, the evidence that: (1) B4GALT5 variants reported in a patient are totally inactive; (2) human B4GALT6 is not efficiently translated being able to drive relevant GSL biosynthesis only at extremely high levels of its transcripts, suggest that the amounts of brain GSLs, probably gangliosides, could not be sufficient, at least in some cell types, to assure proper brain development and function. It is possible that a novel CDG exists, although additional patients and/or studied on differentiated neural cells or organoids are necessary for definitive proof.