A proposed pathway from D-glucose to D-arabinose in eukaryotes

Abstract

In eukaryotes, the D-enantiomer of arabinose (D-Ara) is an intermediate in the biosynthesis of D-erythroascorbate in yeast and fungi and in the biosynthesis of the nucleotide sugar GDP-α-D-arabinopyranose (GDP-D-Arap) and complex α-D-Arap-containing surface glycoconjugates in certain trypanosomatid parasites. Whereas the biosynthesis of D-Ara in prokaryotes is well understood, the route from D-glucose (D-Glc) to D-Ara in eukaryotes is unknown. In this paper, we study the conversion of D-Glc to D-Ara in the trypanosomatid Crithidia fasciculata using positionally labeled [¹³C]-D-Glc and [¹³C]-D-ribose ([¹³C]-D-Rib) precursors and a novel derivatization and gas chromatography-mass spectrometry procedure applied to a terminal metabolite, lipoarabinogalactan. These data implicate the both arms of pentose phosphate pathway and a likely role for D-ribulose-5-phosphate (D-Ru-5P) isomerization to D-Ara-5P. We tested all C. fasciculata putative sugar and polyol phosphate isomerase genes for their ability to complement a D-Ara-5P isomerase-deficient mutant of Escherichia coli and found that one, the glutamine fructose-6-phosphate aminotransferase (GFAT) of glucosamine biosynthesis, was able to rescue the E. coli mutant. We also found that GFAT genes of other trypanosomatid parasites, and those of yeast and human origin, could complement the E. coli mutant. Finally, we demonstrated biochemically that recombinant human GFAT can isomerize D-Ru-5P to D-Ara5P. From these data, we postulate a general eukaryotic pathway from D-Glc to D-Ara and discuss its possible significance. With respect to C. fasciculata, we propose that D-Ara is used not only for the synthesis of GDP-D-Arap and complex surface glycoconjugates but also in the synthesis of D-erythroascorbate.

Keywords: Crithidia fasciculata; D-arabinose; D-erythroascorbate; glucose metabolism; pentose phosphate pathway.

Figure 1

¹³C-labeling of D-Ara in C. fasciculata LAG. A modified methylation linkage analysis procedure (see Experimental procedures) was used to generate partially methylated alditol ethylates (PMAEs) of the constituent monosaccharides of LAG, including the [1-²H]-2,3,4-trimethyl-1,5-diethyl-D-arabitol derived from the nonreducing D-Arap residues of LAG, which was isolated and analyzed by GC-MS. The mass spectra of that derivative from LAG purified from C. fasciculata grown in unlabeled or positionally ¹³C-labeled Glc or Rib (as indicated in each panel) are shown in panels (A–H). The mass spectra shown here are details over the range m/z 100 to 200, where most of the key reporter ions are located. The complete spectra are shown in Fig. S1. The position(s) of the ¹³C-atoms in the [1-²H]-2,3,4-trimethyl-1,5-diethyl-D-arabitol derivatives inferred by the mass spectra are indicated by black dots in the insets of panels (B–H).

Figure 2

Eukaryotic GFAT genes can complement an E. coli API mutant. Agar plates containing ampicillin were used to select for bacteria containing the pT7 plasmid. The right-hand side (rhs) of the plate was overlaid with D-Ara-5-P and D-Glc-6P which rendered it permissive to growth by the Escherichia coli API mutant. The left-hand side (lhs) of the plate was nonpermissive to growth by the E. coli API mutant. Each side was segregated into three sectors: The top sectors were used to plate the E. coli API mutant transformed with an empty pT7 vector. These bacteria should grow on the rhs (permissive) sector, but not the lhs (nonpermissive) sector, acting as a negative control for each experiment. The bottom sectors were used to plate the E. coli API mutant transformed with the pT7-API (Bacteroides fragilis) vector. These bacteria should grow under permissive (rhs) and nonpermissive (lhs) conditions, acting as a positive control for each experiment. The middle sectors were used to plate the E. coli API mutant transformed with one of the pT7-isomerase vectors. All should grow under permissive (rhs) conditions, acting as an additional control for viability, but only under nonpermissive (lhs) if the isomerase gene possesses API activity. The isomerase genes tested are described in (Table 1) and were as follows: Panel A: Crithidia fasciculata glucose-6-phosphate isomerase (CfGPI). Panel B: C. fasciculata glucosamine-6-phosphate aminotransferase (CfGFAT). Panel C: Trypanosoma brucei glucosamine-6-phosphate aminotransferase (TbGFAT). Panel D: Leishmania donovani glucosamine-6-phosphate aminotransferase (LdGFAT). Panel E: Homo sapiens glucosamine-6-phosphate aminotransferase-2 (HsGFAT). Panel F: Saccharomyces cerevisiae glucosamine-6-phosphate aminotransferase (ScGFAT). Experiments where API mutant complementation by indicated the pT7-isomerase plasmid was successful and are ringed in green. Experiments where API mutant complementation by indicated the pT7-isomerase plasmid was unsuccessful and are ringed in red.

Figure 3

Human recombinant GFAT can convert D-Ru-5P to D-Ara-5P. High-pH anion exchange (HPAEC) chromatograms of D-Glc-6P (G6P), D-Fru-6P (F6P), D-Ara-5P (A5P), and D-Ru-5P (Ru5P) incubated without and with recombinant human GFAT (hGFAT) for the times indicated. The detector response (nC) was measured against the retention time. (R5P, D-Rib-5P).

Figure 4

C. fasciculata can make D-erythroascorbate from D-Ara. GC-MS total ion chromatograms of TMS-derivatized polar metabolite extracts of WT C. fasciculata grown in the absence (panel A) and presence (panel B) of 2 mM D-Ara. Panel C: The electron impact mass spectrum of the peak labeled D-erythro-ascorbic acid at 14.32 min in panel B. According to the National Institute of Standards and Technology (NIST) electron impact spectral database, other peaks in panels A and B are consistent with those of the TMS derivatives of glycyl-glutamate (10.7 min), proline and valine (11.7 min), glutamic acid (12.3 min), phenylalanine (12.5 min), ribose (15.3 min), phosphoglycerate (16.2 min), and myo-inositol (21.7 min).

Figure 5

Proposed pathway for the formation of D-Ara and its metabolites in eukaryotes. The bioconversion of D-glucose (D-Glc) to D-arabinose is proposed to be via D-glucose-6-phosphate (Glc-6P) and both the oxidative and nonoxidative branches of the pentose phosphate pathway (PPP) to D-ribulose-5-phosphate (D-Ru-5P). In Crithida fasciculata, the predominant route was the oxidative branch of the PPP. From the work in this paper, D-Ru-5P is proposed to be isomerized to D-arabinose-5-phosphate (D-Ara-5P) by the isomerase domain of glutamine fructose-6-phosphate aminotransferase (GFAT). An unidentified D-Ara-5P phosphatase is postulated (dotted arrow) to convert D-Ara-5P to D-Ara. The conversion of D-Ara via D-Ara-1,4-lactone to D-erythroascorbate is well described in yeast and fungal metabolism and also appears from this paper to occur in C. fasciculata. The conversion of D-Ara to GDP-α-D-Arap has been described in certain kinetoplastid organisms (C. fasciculata and Leishmania major), as has the incorporation of D-Arap residues from GDP-α-D-Arap, catalyzed by D-arabinosyltransferases, into complex cell surface glycoconjugates of those organisms.