Fig. 3: LH2b is a collagen GGT.

a Amino acid sequence alignment of human LHs. Residues involved in Mn²⁺- and uridine diphosphate (UDP)-binding are indicated with arrows and asterisks, respectively. b and c LH1 galactosylhydroxylysyl glucosyltransferase (GGT) activity was assayed. Substrates were hyl-gal (b) or type IV collagen (c) that had been pre-treated with wild-type (+) or glucosidase-dead mutant (-) protein-glucosylgalactosylhydroxylysine glucosidase (PGGHG). GGT activity was measured by detecting UDP production with an ATP-based luciferase assay. d–f Wild type and mutant LH2b GGT activity was assayed using an ATP-based luciferase assay that detects UDP production. Substrates as described in (b) and (c). GGT activity was abolished by mutation of a Mn²⁺-binding residue (D115E) (f). g and h LH3 GGT assays. Substrates as described in (b) and (c). Residues involved in Mn²⁺-binding were mutated (D112A, D115A) (g). i LH2 protein levels were determined by western blot analysis of MC cells stably co-transfected with LH2 shRNA (shLH2) and empty vector (-) or vectors expressing Flag-tagged LH2a or LH2b. Tubulin used as loading control. j Quantification of collagen cross-links in matrices produced by MC cells in (i). HLCC-to-LCC ratio was calculated as (DHLNL + Pyr)/HHMD. k t-LH assay on recombinant LH proteins (LH2a or LH2b) using trimeric collagen telopeptide as substrate. LH activity was measured by detecting succinate production with an ATP-based luciferase assay. l and m LH2a and LH2b GGT activity was measured by an ATP-based luciferase assay that detects UDP production. Substrates used were hyl-gal (l) or PGGHG-treated type IV collagen (m). n Domain structures of LH2a and LH2b (top). Location of exon 13A-encoded sequences (red bar) in accessory domain (AC). LH2b homology model was generated using a homology-modelling server SWISS-MODEL (bottom). A recently determined LH3 structure was used as a template (PDB ID: 6FXT). Exon 13A-encoded loop (arrow). Mn²⁺ (magenta ball) and Fe²⁺ (orange balls) in GLT and LH domain active sites, respectively. o LH2’s UDP-glucose-binding affinity was determined by microscale thermophoresis. Fluorescein-conjugated UDP-Glucose (50 nM) was titrated with different concentrations of LH2a and LH2b recombinant proteins to generate the curves. Curves were used to calculate the K_d values for LH2a (red) and LH2b (cyan). Enzymatic activity assay results are mean values (±S.D.) from triplicate samples (n = 3) and microscale thermophoresis results are mean values from duplicate samples (n = 2). Error bars indicate ±S.D. p values, 2-tailed Student’s t test.