Peters plus syndrome mutations affect the function and stability of human β1,3-glucosyltransferase

Abstract

Peters Plus Syndrome (PTRPLS OMIM #261540) is a severe congenital disorder of glycosylation where patients have multiple structural anomalies, including Peters anomaly of the eye (anterior segment dysgenesis), disproportionate short stature, brachydactyly, dysmorphic facial features, developmental delay, and variable additional abnormalities. PTRPLS patients and some Peters Plus-like (PTRPLS-like) patients (who only have a subset of PTRPLS phenotypes) have mutations in the gene encoding β1,3-glucosyltransferase (B3GLCT). B3GLCT catalyzes the transfer of glucose to O-linked fucose on thrombospondin type-1 repeats. Most B3GLCT substrate proteins belong to the ADAMTS superfamily and play critical roles in extracellular matrix. We sought to determine whether the PTRPLS or PTRPLS-like mutations abrogated B3GLCT activity. B3GLCT has two putative active sites, one in the N-terminal region and the other in the C-terminal glycosyltransferase domain. Using sequence analysis and in vitro activity assays, we demonstrated that the C-terminal domain catalyzes transfer of glucose to O-linked fucose. We also generated a homology model of B3GLCT and identified D421 as the catalytic base. PTRPLS and PTRPLS-like mutations were individually introduced into B3GLCT, and the mutated enzymes were evaluated using in vitro enzyme assays and cell-based functional assays. Our results demonstrated that PTRPLS mutations caused loss of B3GLCT enzymatic activity and/or significantly reduced protein stability. In contrast, B3GLCT with PTRPLS-like mutations retained enzymatic activity, although some showed a minor destabilizing effect. Overall, our data supports the hypothesis that loss of glucose from B3GLCT substrate proteins is responsible for the defects observed in PTRPLS patients, but not for those observed in PTRPLS-like patients.

Keywords: B3GLCT; O-fucose; Peters plus syndrome; enzyme catalysis; genetic disease; glycobiology; glycoprotein secretion; glycosyltransferase; thrombospondin type-1 repeats.

Figure 1

Human B3GLCT contains two putative GT domains, and PTRPLS/PTRPLS-like mutations are predominantly localized in the C-GT domain.A, domain map of human B3GLCT with schematic distribution of PTRPLS (black) and PTRPLS-like (blue) mutations throughout the protein. Two putative DxD motifs, ¹³²EEE¹³⁴ and ³⁴⁹DDD³⁵¹, are highlighted with arrows. B, common GT-A fold elements (6) identified from the multiple sequence alignment of the N-GT-like and C-GT domains from B3GLCT shown in Figure S1. Human (UNIPROT Q6Y288), mouse (UNIPROT Q8BHT6), zebrafish (UNIPROT A0A068F9P7), fruit fly (UNIPROT X2JDC2), and starfish (NCBI REFSEQ XP_038056415.1). Sequences were aligned with the MAFFT software/algorithm. Red background highlights regions of 100% sequence identity within the shown sequences; red letters indicate regions of 100% sequence similarity; purple boxes highlight DxD motifs; green boxes highlight G-loop; gray boxes highlight predicted xED motifs (note that two potential xED motifs are highlighted for N-GT-like domains); orange boxes highlight C-His.

Figure 2

The DxD motif of the C-GT domain is responsible for B3GLCT glucosyltransferase activity.A, in vitro enzyme activity assay (20 min) of WT, E132A mutant, and D349N mutant B3GLCTΔREEL with UDP-Glc as donor substrate (50 μM) and Fuc-O-TSR3 (25 μM) as acceptor substrate. Biological triplicates were performed with three different batches of enzymes. Error bars represent standard deviation, n = 3. Statistical analysis was performed with one-way ANOVA by comparing activities of the mutants to WT in Prism 7. ∗∗∗∗p < 0.0001; ns, not significant. B and C, plasmids encoding ADAMTS20 TSR2-8 (TSR2-8) and GFP were cotransfected into wild type (WT) and B3GLCT knockout (B3GLCT^−/−) HEK293T cells. Rescue experiments were performed by cotransfection with a plasmid encoding full-length B3GLCT WT (wtB3), E132A mutant (B) or D349N mutant (C). Serial dilutions of B3GLCT-FL plasmids were performed starting with 0.24 μg of plasmids diluted 5-, 10-, and 20-fold. Media and cell lysates were analyzed by western blot probed with anti-Myc (red) to detect ADAMTS20 TSR2-8, anti-B3GLCT to detect endogenous or transfected B3GLCT (green, 50 kDa), and for transfection and loading control anti-GFP (green, 25 kDa).

Figure 3

PTRPLS mutants are not catalytically active, but PTRPLS-like mutants are.A, top, reaction catalyzed by B3GLCT for transfer Glc to Fuc-O-TSR3. Solid blue circle, glucose; red oval, TSR3; red triangle, fucose. A, bottom, in vitro enzyme assay (20 min) of WT, PTRPLS mutants (black), and PTRPLS-like mutants (blue). Calculated UDP produced by the mutants was normalized to the UDP produced by WT. Experiment was repeated in biological triplicates with three batches of purified enzymes. Error bars represent standard deviation, n = 3. Statistical significance was analyzed by comparison of the mutants to WT with one-way ANOVA analysis in Prism 7. ∗∗∗∗p < 0.0001; ns, not significant. B and C, substrate concentration dependent kinetics of WT and PTRPLS-like mutants with various Fuc-O-TSR3 (B) and UDP-Glc concentrations (C) (n = 3). Technical replicates with error bar show standard deviation for each point.

Figure 4

PTRPLS, but not PTRPLS-like, mutations abolished the ability to rescue the secretion of ADAMTS20 TSR2-8 from B3GLCT^−/−cells. A, plasmids encoding ADAMTS20 TSR2-8-Myc-His₆ and GFP were cotransfected into WT or B3GLCT^−/− HEK293T cells. Rescue experiments were performed by cotransfection with plasmids encoding full-length WT, PTRPLS, or PTRPLS-like mutations in B3GLCT (B3GLCT). Media and cell lysates were analyzed by western blot probed with anti-Myc (red) to detect ADAMTS20 TSR2-8 (TSR2-8), anti-GFP for transfection and loading control (green, 25 kDa), and anti-B3GLCT for B3GLCT WT or mutants (green, 50 kDa). B, top, domain maps of N-His₆-B3GLCT and ADAMTS20 TSR2-8-Myc constructs used in this experiment. Bottom, plasmids encoding ADAMTS20 TSR2-8-Myc and GFP were cotransfected into WT or B3GLCT^−/− HEK293T cells. Rescue experiments were performed by cotransfection with serial dilutions of plasmids encoding N-His₆-B3GLCT, N-His₆-Y366∗ mutant, or N-His₆-R412∗ mutant. Serial dilutions of rescue plasmids were performed starting with 0.24 μg diluted 5-, 10-, and 20-fold. Media and cell lysates were analyzed by western blots probed with anti-Myc (green, 50 kDa) to detect ADAMTS20 TSR2-8-Myc (TSR2-8), anti-GFP (green, 25 kDa) for transfection and loading control, and anti-His₆ (red, 50 kDa or less) for B3GLCT WT, Y366∗, and R412∗.

Figure 5

PTRPLS G393E and PTRPLS-like Q457R mutants significantly destabilized B3GLCT. Thermo shift assays with SYPRO Orange were employed to analyze the protein stabilities by measuring the melting temperatures of the proteins. Melting temperatures of PTRPLS mutants (black) and PTRPLS-like mutants (blue) were compared with the melting temperature of the WT (Fig. S5). Statistical analysis was performed with one-way ANOVA analysis in Prism 7. Error bars, standard deviation, n = 6 from two batches of purified enzymes with three replicates for each batch. ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 6

Homology model revealed B3GLCT C-GT domain as a GT-A fold enzyme. Structure of B3GLCT C-GT domain was generated by homology modeling and represented in ribbon (light blue). Key conserved residues are highlighted in light blue. PTRPLS mutations are highlighted in salmon and PTRPLS-like mutations are highlighted in yellow.

Figure 7

Structure alignment of B3GLCT C-GT with other GT31 family proteins identified D421 as the catalytic base.A, the UDP ligands for MFNG (PDB 2J0B, blue) and B3GNT2 (PDB 7JHN, purple) are displayed along with catalytic base, D232 for MFNG and D333 for B3GNT2. D421 of B3GLCT overlayed with D232 in MFNG and D333 in B3GNT2. DxD motifs and G-loops of all three enzymes also overlayed. B, in vitro enzyme assay (20 min) of B3GLCT WT and D421A mutant with UDP-Glc and Fuc-O-TSR3. Statistical analysis was performed with paired t test, two-tailed in Prism 7, n = 3, ∗∗∗p < 0.001.