Bi-allelic UGGT1 variants cause a congenital disorder of glycosylation

Abstract

Congenital disorders of glycosylation (CDGs) comprise a large heterogeneous group of metabolic conditions caused by defects in glycoprotein and glycolipid glycan assembly and remodeling, a fundamental molecular process with wide-ranging biological roles. Herein, we describe bi-allelic UGGT1 variants in fifteen individuals from ten unrelated families of various ethnic backgrounds as a cause of a distinctive CDG of variable severity. The cardinal clinical features of UGGT1-CDG involve developmental delay, intellectual disability, seizures, characteristic facial features, and microcephaly in the majority (9/11 affected individuals for whom measurements were available). The more severely affected individuals display congenital heart malformations, variable skeletal abnormalities including scoliosis, and hepatic and renal involvement, including polycystic kidneys mimicking autosomal recessive polycystic kidney disease. Clinical studies defined genotype-phenotype correlations, showing bi-allelic UGGT1 loss-of-function variants associated with increased disease severity, including death in infancy. UGGT1 encodes UDP-glucose:glycoprotein glucosyltransferase 1, an enzyme critical for maintaining quality control of N-linked glycosylation. Molecular studies showed that pathogenic UGGT1 variants impair UGGT1 glucosylation and catalytic activity, disrupt mRNA splicing, or inhibit endoplasmic reticulum (ER) retention. Collectively, our data provide a comprehensive genetic, clinical, and molecular characterization of UGGT1-CDG, broadening the spectrum of N-linked glycosylation disorders.

Keywords: MOGS; N-linked glycosylation; UDP-glucose:glycoprotein glucosyltransferase 1; UGGT1; autosomal recessive; congential disorder of glycosylation; microcephaly; monogenic disorder; neurodevelopmental disorder.

Graphical abstract

Figure 1

Carbohydrate-dependent quality control pathway Bold text indicates proteins with relevant enzymatic activity. Proteins highlighted in red indicate association with a congenital disorder of glycosylation (CDG) in OMIM. UGGT1 protein is highlighted in blue. Pathway is as follows: (1) protein is glycosylated during and after translation. (2) The resulting glycan attached to the glycoprotein is trimmed by glucosidase I encoded by MOGS (MIM: 601336) to form a di-glucosylated glycan. (3) The second glucose is trimmed by glucosidase II (alpha subunit encoded by GANAB [MIM: 104160] and beta subunit encoded by PRKCSH [MIM: 177060]) to form a monoglucosylated glycan. (4) The monoglucosylated glycan binds to calnexin (CNX), a membrane-bound lectin chaperone, or calreticulin (CRT), CNX’s soluble paralog. This promotes proper folding by preventing aggregation and premature export from the ER. (5) The final glucose is trimmed by glucosidase II to form a non-glucosylated glycan. (6) The resulting glycoprotein is released from CNX/CRT cycle. (7a) Natively folded proteins continue to traffic through the ER for release. (7b) Non-natively folded proteins with minor folding defects are recognized by the folding sensor UGGT1. (8) Terminally misfolded glycoproteins are extracted from the CNX/CRT cycle for degradation by the ER-associated degradation (ERAD) process. This starts by trimming B- and C-branch mannoses to create a degradation signal. These mannosidases include ER mannosidase 1 encoded by MAN1B1 (MIM: 604346) and EDEM 3 encoded by EDEM3 (MIM: 610214). (9) UGGT1 reglucosylates the protein to create a monoglucosylated glycan. (10) The glycan re-binds to CNX/CRT to be retained in the ER. The cycle continues from step 5. (11) Alongside this cycle, ER mannosidases trim an A-branch mannose from the glycan, precluding the ability of UGGT1 to reglucosylate, preventing slow folding glycoproteins from remaining in the cycle indefinitely. Once the A1 mannose has been trimmed, the glycoprotein is removed from the cycle for degradation by the ERAD process.

Figure 2

Family pedigrees and bi-allelic UGGT1 variants associated with UGGT1-CDG (A) Pedigrees of the families investigated depicting autosomal recessive segregation of the pathogenic UGGT1 variants identified. Co-segregation was confirmed in other family members as indicated; in each case, the plus symbol indicates the variant allele, and the minus sign indicates the wild-type allele. Blue/orange text has been used to differentiate between variants in a pedigree when the UGGT1 genotype in affected individuals is compound heterozygous. # indicates variants that have been shown to impact splicing. (B) (i) Simplified schematic depicting UGGT1 protein structure showing location of each UGGT1 variant in relation to the predicted domain architecture. Pink shaded box indicates the REEL endoplasmic reticulum retrieval sequence, gray shaded boxes indicate predicted nonsense-mediated decay escape regions, gray circles indicate glycosylation sites, and colored boxes represent protein domains. TRXL1, thioredoxin-like domain 1 (orange); TRXL2, thioredoxin-like domain 2 (purple); TRXL3, thioredoxin-like domain 3 (yellow); TRXL4, thioredoxin-like domain 4 (red); βS1, beta sheet-1 (green); βS2, beta sheet-2 (purple); GT24, glycotransferase 24 (blue). (ii) Schematic of the UGGT1 MANE select transcript (GenBank: NM_020120.4) depicting intron-exon organization and the location of each UGGT1 variant.

Figure 3

Facial features and neuroimaging findings seen in individuals with UGGT1-CDG (A and B) Individual II-1 from family 1 showing coarse facial features, micrognathia, microcephaly with a sloping forehead, prominent widely spaced eyes/narrow palpebral fissures, broad nasal bridge with upturned bulbous nose, and low set and posteriorly rotated ears. (C) Individual II-1 from family 3 has features that include bilateral cleft lip, high nasal root with broad nasal bridge, narrow palpebral fissures, and arched eyebrows. (D and E) Individuals II-1 (D) and II-2 (E) from family 6 showing microcephaly, bulbous nasal tip, and triangular facies. (F) Individual II-1 from family 7 shows hypertelorism, wide and prominent central incisors, mild bifid nasal tip, and strabismus. (G–I) Individuals II-1 (G), II-2 (H), and II-3 (I) from family 9 showing coarse facies, long face, arched eyebrows, high forehead, broad nasal root, prominent nasal bridge, barrel nose, long smooth philtrum, everted lower lip, macrostomia, protruded tongue, and low set ears. (J) Family 1, II:1. MRI was performed at 2 weeks of age. Axial T2 (i), T1 (ii), and DWI (iii) and coronal T2 (iv) showing mildly abnormal white matter, which may be immaturity or edema in both corona radiata (i and ii, blue asterisk) with no restricted diffusion (iii). Periventricular heterotopic gray matter (i, ii, and iv, orange arrows). (K) Family 7, individual II:1. Axial T2 (i) and T1 (ii), sagittal T2 (iii), and coronal T1 (iv) showing lack of frontal volume (i, ii, and iii, orange arrows). Bilateral high T2/low T1 signal in the putamina bilaterally (i and ii, green arrowheads). Hippocampi appear normal (iii and iv, blue asterisk). (L) Family 8, individual II:2. MRI coronal T2 (i) and FLAIR (ii), axial T2 (iii), and proton density (iv) performed at 3 years of age demonstrating left mesial temporal sclerosis (i and ii, green arrowhead), impaired myelination temporal white matter (iv, orange arrows), and left cerebral hemisphere atrophy (all images). (M) Family 9, individuals II:3 (i–iv) and II:2 (v and vi). II:3: axial T2 (i) and T1 (ii), sagittal T2 (iii), and coronal FLAIR (iv). Abnormal T2 signal in anterior temporal lobes (i and iii, orange arrow), without low T1 signal (ii, orange arrow) suggesting insult. Possible increased T2/FLAIR signal in hippocampi (iv, blue asterisk). II:2: coronal FLAIR (v) and axial T2 (vi). Bilateral increased T2/FLAIR signal in hippocampi (v and vi, blue asterisk) without other associated abnormalities.

Figure 4

UGGT1 variants disrupt UGGT1-mediated glucosylation and glucosyltransferase activity (A) Wild-type (WT) UGGT1 (lanes 1–6), p.Ala711Val (lanes 7–12), p.Arg1272His (lanes 13–18), and p.Arg1546^∗ (lanes 19–24) were expressed individually or co-transfected with AAT-Z in ALG6/UGGT1/2^−/− cells. Once prepared, the cellular lysates were split between whole-cell lysates (WCLs) (20%) to detect total protein, 35% for CRT to isolate monoglucosylated UGGT1 or AAT-Z, and 35% to CRT^∗ to detect any background binding to the recombinant GST-CRT protein. The samples were resolved by 9% SDS-PAGE gel before being probed by western blot. Immunoblots were probed with an anti-FLAG polyclonal antibody to measure the expression and reglucosylation of exogenous UGGT1 (top) and AAT-Z (bottom). The immunoblots were also probed with an anti-UGGT1 polyclonal antibody to ensure bands were specific to overexpressed UGGT1 (middle). (B) Western blot measuring the activity of WT UGGT1 and heterozygous UGGT1 mutants. UGGT1 p.Tyr127^∗ (lanes 1–6), p.Asp390-Gly397del (lanes 7–12), p.Phe723Ser (lanes 13–18), and p.Gln1361Profs^∗27 (lanes 19–24) were expressed along with AAT-Z, prepared, and analyzed as in (C)–(E). The endogenous UGGT1 antibody was raised against a fusion protein containing the amino acid sequence 1,456–1,555 for human UGGT1 and thus is unable to detect mutants missing this sequence. (C and D) Quantification of UGGT1 (C) and AAT-Z (D) glucosylation from (A) and (B). Percentage of reglucosylation was determined by dividing the amount of quantified CRT, subtracting any background from the CRT^∗, and dividing it by the WCL before multiplying it by 100. Error bars represent standard deviation. ^∗∗p < 0.01 and ^∗∗∗∗p < 0.0001, respectively. Colored circles represent replicates. All western blots are representative of three separate biological replicates. (E) Glucosyltransferase activity of recombinant UGGT1 proteins harboring each variant as a percentage of WT glucosyltransferase activity. p.Tyr127^∗ variant protein could not be detected by western blot and so has been omitted. Error bars represent standard deviation. ^∗p < 0.05 and ^∗∗∗p < 0.001.

Figure 5

Molecular consequences of UGGT1 variants identified in families with UGGT1-CDG (A) Description of wild-type (WT) and homozygous UGGT1 mutants. Gray indicates signal sequence cleaved upon entry into the ER, and blue denotes mature soluble protein. The orange line in the WT indicates sites of potential N-glycosylation sites found in UGGT1. All potential sites are preserved in all of the mutants except for the UGGT1 c.381_384del (p.Tyr127^∗) mutant. A 3xFLAG tag was inserted before the C-terminal ER retrieval sequence (REEL) to allow for detection by western blot. UGGT1 variants are noted with an asterisk, indicating the insertion of a premature stop codon. (B) Immunoblot measuring the expression and secretion of WT UGGT1 and homozygous UGGT1 mutants from (A) expressed in HEK cells. The levels of UGGT1 expressed within the cell (whole-cell lysate [WCL], lanes 1, 3, 5, and 7) was compared to levels secreted into the media (media [M], lanes 2, 4, 6, and 8) using a mouse monoclonal anti-FLAG antibody for immunoprecipitation. Detection was performed with an anti-FLAG polyclonal antibody. (C) Description of heterozygous UGGT1 mutants, similar to (A). The locations of the variants are similarly noted. (D) Immunoblot measuring expression and secretion of heterozygous UGGT1 mutants as described in (B). (E) Quantification of the levels of UGGT1 secretion into the media from (B) and (D). The percentage of UGGT1 in media was calculated by dividing the amount of UGGT1 quantified in the media by the total amount of protein in the WCL and media combined. Error bars represent standard deviation. ^∗∗∗p < 0.001. Colored circles represent replicates. All western blots are representative of three independent biological replicates. (F) Western blot analysis of UGGT1 protein in WCL (left) and cell culture supernatants (right) from healthy control subjects’ fibroblasts (GM1652, GM5381, and GM8447, colored blue) and affected individuals with the homozygous UGGT1 variant p.Arg1546^∗ (family 9: II-1 and II-2, labeled as BAB15130 and BAB15131, respectively, colored red), demonstrating a severe reduction in UGGT1 protein levels in patient WCLs with a concomitant increase in secreted UGGT1 protein (UGGT1 molecular weight = ∼177 kDa).