A mutation in SLC37A4 causes a dominantly inherited congenital disorder of glycosylation characterized by liver dysfunction

Abstract

SLC37A4 encodes an endoplasmic reticulum (ER)-localized multitransmembrane protein required for transporting glucose-6-phosphate (Glc-6P) into the ER. Once transported into the ER, Glc-6P is subsequently hydrolyzed by tissue-specific phosphatases to glucose and inorganic phosphate during times of glucose depletion. Pathogenic variants in SLC37A4 cause an established recessive disorder known as glycogen storage disorder 1b characterized by liver and kidney dysfunction with neutropenia. We report seven individuals who presented with liver dysfunction multifactorial coagulation deficiency and cardiac issues and were heterozygous for the same variant, c.1267C>T (p.Arg423^∗), in SLC37A4; the affected individuals were from four unrelated families. Serum samples from affected individuals showed profound accumulation of both high mannose and hybrid type N-glycans, while N-glycans in fibroblasts and undifferentiated iPSC were normal. Due to the liver-specific nature of this disorder, we generated a CRISPR base-edited hepatoma cell line harboring the c.1267C>T (p.Arg423^∗) variant. These cells replicated the secreted abnormalities seen in serum N-glycosylation, and a portion of the mutant protein appears to relocate to a distinct, non-Golgi compartment, possibly ER exit sites. These cells also show a gene dosage-dependent alteration in the Golgi morphology and reduced intraluminal pH that may account for the altered glycosylation. In summary, we identify a recurrent mutation in SLC37A4 that causes a dominantly inherited congenital disorder of glycosylation characterized by coagulopathy and liver dysfunction with abnormal serum N-glycans.

Keywords: Golgi pH; coagulopathy; congenital disordes of glycosylation; exome sequencing; glycosylation.

Figure 1

Identification of a recurrent SLC37A4 mutation in four unrelated families (A) Pedigrees showing segregation of the SLC37A4 c.1267C>T (p.Arg423^∗) mutation in seven affected individuals from four unrelated families. (B) LC-MS of serum transferrin from control and P7 serum with deconvoluted masses of intact serum TF from full scans showing the appearance of several peaks corresponding to distinctive peaks containing hybrid N-glycans. (C) Schematic of human SLC37A4 showing the p.Arg423^∗ localizing to the cytoplasmic tail (UniProt: O43826-1).

Figure 2

N-glycan abnormalities in serum from affected individuals MALDI-TOF MS spectra of serum protein-derived N-glycans from unrelated individuals (P1, P4, and P6) are abnormal. Specifically, both high mannose (peaks at m/z 1,579.8 and 1,783.9) and hybrid type N-glycans (peaks at m/z 1,981.9, 2,186.1, and 2,390.2) increases were seen in positive-ion mode as sodiated forms. Green circles, mannose; yellow circles, galactose; blue squares, N-acetyl glucosamine; red triangles, fucose; purple diamonds, sialic acid.

Figure 3

Characterization of N-glycans from p.Arg423^∗ base-edited Huh7 cells N-glycans released from secreted glycoproteins by PNGase F digestion showing the accumulation of both high mannose and hybrid type glycans in C21 with glycan abundances deduced from NSI-MSn measurements.

Figure 4

Localization of mutant SLC37A4 (A) Subcellular fractionations of extracts from control C9 and edited lines C21 and C71 showing SLC37A4 protein is absent from the GM130 Golgi-containing fractions, but with similar fractionation pattern as the ER marker, calnexin. Subcellular fractionations were performed with three biological replicates via the Nycodenz gradient method, and representative images are shown. (B) Immunofluorescence staining of Huh7 control and edited cells showing localization of SLC37A4 with the ER marker, KDEL (upper panel), the Golgi marker, GM130 (middle panel), and the ER exit site marker, SEC31 (lower panel). (C) Immunofluorescence staining for the ER marker, KDEL, and SLC37A4 in iPSC-derived hepatocytes from control and P7. Control (upper panel) showed strong colocalization (r = 0.98), while P7 (lower panel) showed reduced colocalization (r = 0.41).

Figure 5

Abnormal Golgi structure and function in p.Arg423^∗ base-edited Huh7 cells (A) Immunofluorescence staining of Huh7 control (C8 and C9) and edited cells (C21 and C71) with the Golgi marker, GM130, showing abnormalities Golgi morphology/area. Scale bar represents 20 µm. (B) Golgi area was quantified and then normalized to the area of the nucleus to provide a ratio showing both C21 (homozygous clone) and C71 (heterozygous clone) had significantly increased Golgi area. The effect is significantly more pronounced when both alleles are mutated as they are in C21. In three separate biological replicates, conducted within monthly intervals, each time with freshly thawed cells, n = 8 cells were analyzed, and the mean was taken. The graph represents an average of the means acquired over three different biological measurements. Total cells measured N = 24 over three biological replicates. Statistical significance ^∗p < 0.05, ^∗∗p < 0.005, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001 was calculated via one-way ANOVA.

Figure 6

Calibration and quantification of the Golgi pH in Huh7 control and edited cells (A) Calibration buffers were used to create a pH response curve for the GalT-mCherry-eGFP construct and allowed for the calculation of pH in Huh7 cells. Scale bar represents 20 µm. (B) Standard curve using calibration buffers to determine estimated luminal pH values. (C) Quantification of the Golgi luminal pH values in Huh7 controls (C8 and C9) and edited (C21, a homozygous clone, and C71, a heterozygous clone) cells showing acidification of the Golgi upon introduction of the c.1267C>T (p.Arg423^∗) mutation. The effect is significantly more pronounced when both alleles are mutated. Data were acquired in four (C8 and C71) or six (C9 and C21) biological replicates, conducted on different weeks, with cells freshly transfected with GalT-mCherry-eGFP construct. In each biological replicate, 10–15 cells were analyzed, and the mean was taken. The graph represents an average of the means acquired over the different biological measurements. Statistical significance ^∗p < 0.05, ^∗∗p < 0.005, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001 was calculated via one-way ANOVA.

Figure 7

Overview showing function of SLC37A4 in hepatocytes A schematic showing the function of wild-type SLC37A4 in hepatocytes during nourished or fasting conditions. Under nourished conditions, exogenous Glc provides ample Glc-6P for glycolysis, glycogenesis, and pentose phosphate pathways. Under fasting conditions, hepatocytes must generate Glc-6P from glycogenolysis (GL) and gluconeogenesis (GNG) for these pathways and also normalize plasma Glc. SLC37A4 imports Glc-6P into the ER and G6PC releases Pi+Glc so both can be returned to the cytoplasm and Glc to the circulation. Mutant SLC37A4 (p.Arg423^∗) is fully active and maintains normal glucose homeostasis under fasting conditions, but a portion of the active transporter becomes mislocalized to an undefined, spatially restricted pre-Golgi/post-ER compartment, possibly ERES. Glc-6P and/or Pi accumulates there, leading to a dose-dependent reduction of Golgi pH and Golgi architecture/homeostasis. Reduced pH is propagated in subsequent Golgi compartments, altering the activity and/or localization of multiple N- and O-glycan-modifying enzymes. Because Mn and Mg are critical co-factors for many of these reactions, reduced pH could affect their solubility and availability.