A novel GDP-D-glucose phosphorylase involved in quality control of the nucleoside diphosphate sugar pool in Caenorhabditis elegans and mammals

Abstract

The plant VTC2 gene encodes GDP-L-galactose phosphorylase, a rate-limiting enzyme in plant vitamin C biosynthesis. Genes encoding apparent orthologs of VTC2 exist in both mammals, which produce vitamin C by a distinct metabolic pathway, and in the nematode worm Caenorhabditis elegans where vitamin C biosynthesis has not been demonstrated. We have now expressed cDNAs of the human and worm VTC2 homolog genes (C15orf58 and C10F3.4, respectively) and found that the purified proteins also display GDP-hexose phosphorylase activity. However, as opposed to the plant enzyme, the major reaction catalyzed by these enzymes is the phosphorolysis of GDP-D-glucose to GDP and D-glucose 1-phosphate. We detected activities with similar substrate specificity in worm and mouse tissue extracts. The highest expression of GDP-D-glucose phosphorylase was found in the nervous and male reproductive systems. A C. elegans C10F3.4 deletion strain was found to totally lack GDP-D-glucose phosphorylase activity; this activity was also found to be decreased in human HEK293T cells transfected with siRNAs against the human C15orf58 gene. These observations confirm the identification of the worm C10F3.4 and the human C15orf58 gene expression products as the GDP-D-glucose phosphorylases of these organisms. Significantly, we found an accumulation of GDP-D-glucose in the C10F3.4 mutant worms, suggesting that the GDP-D-glucose phosphorylase may function to remove GDP-D-glucose formed by GDP-D-mannose pyrophosphorylase, an enzyme that has previously been shown to lack specificity for its physiological D-mannose 1-phosphate substrate. We propose that such removal may prevent the misincorporation of glucosyl residues for mannosyl residues into the glycoconjugates of worms and mammals.

FIGURE 1.

GDP-d-Glc phosphorylase activity in C. elegans C10F3.4 deletion strain and in HEK293T cells treated with C15orf58 siRNA. Extracts were prepared from wild-type (N2) or C10F3.4 mutant (tm2679) C. elegans strains (A) and from siRNA-transfected HEK293T cells (B), and GDP-d-Glc phosphorylase activity was measured by the HPLC assay described under “Experimental Procedures.” GDP-d-Glc was added to the reaction mixtures at a final concentration of 50 μm. The worm and HEK293T cell extracts were diluted in the reaction mixtures to final protein concentrations of 0.5–1.4 and 0.3–0.4 mg/ml, respectively. Control reactions were run in the absence of sodium phosphate, and GDP-d-Glc phosphorylase activities were calculated based on the P_i-dependent formation of GDP. A, worms were grown for 4–7 days at 20 °C in liquid culture and harvested for protein extraction. B, HEK293T cell cultures were stopped 48–72 h after siRNA transfection. In addition to GDP-d-Glc phosphorylase activity assays (black bars), quantitative RT-PCR was performed on cDNA derived from these cells. C15orf58 mRNA levels were normalized to those of ACTB (β-actin; gray bars) or GAPDH (white bars), and -fold changes in C15orf58 expression in knockdown versus control cells were calculated using the 2^−ddCt method. The results shown represent the means ± S.D. of two to three (A) or four (B) biological replicates.

FIGURE 2.

C10F3.4 expression pattern in C. elegans. A transgenic C. elegans strain (UZ119) expressing a C10F3.4::GFP fusion protein was generated as described under “Experimental Procedures.” The differential interference contrast images are presented in the left panels, GFP images are in the middle panels, and merged images are in the right panels. C10F3.4::GFP expression is seen in head neurons (arrows) (A), in neuronal cells throughout the ventral nerve cord and in tail (lumbar) ganglia (arrow) (B), in the spermatheca (arrow) (C), and in anterior hypodermis cells (arrow in inset representing a magnified region) (D). The bottom panels show a worm that lost the extrachromosomal array containing the C10F3.4::GFP transgene (E, negative control). Transgenic animals were analyzed for GFP expression at 40× power magnification.

FIGURE 3.

Tissue distribution of GDP-d-Glc phosphorylase activity in mouse. Mouse tissue extracts were prepared, and GDP-d-Glc phosphorylase activities were measured by HPLC as described under “Experimental Procedures.” GDP-d-Glc and mouse tissue extracts were added to the reaction mixtures at final concentrations of 50 μm and 0.3–1.5 mg of protein/ml, respectively. Control reactions were run in the absence of sodium phosphate, and GDP-d-Glc phosphorylase activities were calculated based on the P_i-dependent formation of GDP. The results shown represent the means ± S.D. of three to five biological replicates.

FIGURE 4.

Accumulation of GDP-d-Glc in C10F3.4 mutant C. elegans strain (tm2679). GDP-d-Man and GDP-d-Glc standards (A) and wild-type N2 and mutant tm2679 small molecule extracts (B) were analyzed by the IP-LC-ESI-MS method described under “Experimental Procedures.” A, extracted ion chromatograms of product ions m/z 604 → 158 (middle panel, white area) and m/z 604 → 362 (bottom panel, black area) obtained after injection of a solution containing 0.2 pmol of GDP-d-Man and 0.2 pmol of GDP-d-Glc. GDP-d-Man and GDP-d-Glc produced the same precursor ion, m/z 604, and the same product ions, m/z 604 → 158 and m/z 604 → 362. These GDP-hexoses, however, displayed different retention times with GDP-d-Man eluting first and different relative peak intensities for the product ions. The top panel shows an overlay of the extracted ion chromatograms of product ions m/z 604 → 158 (white area) and m/z 604 → 362 (black area) presented separately in the middle and bottom panels. B, extracted ion chromatograms of product ions m/z 604 → 158 (middle panels, white areas) and m/z 604 → 362 (bottom panels, black areas) after injection of equal volumes of C. elegans N2 and tm2679 perchloric acid extracts. The top panels show overlays of the corresponding middle and bottom panels. The arrows indicate the peak identified as GDP-d-Glc and accumulating in C10F3.4 mutant worms. Representative chromatograms obtained with one set of wild-type and mutant extracts are shown. Four other sets of wild-type and tm2679 mutant extracts were analyzed with similar results. The GDP-d-Man and GDP-d-Glc levels found in these worm extracts were quantified, and the values obtained are given in Table 3. cps, counts/s.

FIGURE 5.

Possible role of GDP-d-Glc phosphorylase in preventing misincorporation of glucose in place of mannose residues into glycoconjugates. In the presence of GTP and d-Man-1-P, GDP-d-Man pyrophosphorylase forms GDP-d-Man, a major mannosyl donor for glycoconjugation. The native pyrophosphorylase is composed of α- and β-subunits (GMPPA and GMPPB) and has been shown to also catalyze the formation of GDP-d-Glc when d-Man-1-P is replaced by d-Glc-1-P. However, we show in this study that mammalian tissues contain only very low levels of GDP-d-Glc. In addition, this nucleotide sugar accumulates in mutant worms lacking GDP-d-Glc phosphorylase, the enzyme described in the present work. We therefore propose that GDP-d-Glc phosphorylase functions to compensate for the lack of specificity of GDP-d-Man pyrophosphorylase by degrading the apparently useless GDP-d-Glc. This may be important to prevent misincorporation of glucose residues for mannose residues into oligosaccharide chains such as those linked to dolichol (this figure) and eventually transferred to proteins. GDP-d-Glc phosphorylase might thus be necessary to preserve functional protein N-glycosylation or other glycoconjugation processes. ER, endoplasmic reticulum.